[{"is_archived":0,"is_starred":0,"tags":["finance","homebuying"],"is_public":false,"id":1231,"title":"Tips for First-Time Home Buyers","url":"https:\/\/www.nerdwallet.com\/article\/mortgages\/tips-for-first-time-home-buyers","content":"<div class=\"_3mbPX\"><p class=\"c4\">Like any big project, a successful homebuying experience is all about getting the details right from start to finish. These tips for first-time home buyers will help you navigate the process, save money and close the deal. We organized them into four categories:<\/p><\/div>\n<div class=\"_3mbPX\"><ul class=\"_19hnS _1jLZP\"><li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Preparing to buy tips.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Mortgage selection tips.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Home shopping tips.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Home purchasing tips.<\/p><\/div>\n<\/li>\n<\/ul><\/div>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/CpMcu1PsMJ4?rel=0&amp;showinfo=0&amp;color=white\" width=\"100%\" height=\"100%\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe>\n<div class=\"_3mbPX _21Cwe _29SoL _2fXMr _1_GyS _1z-GX _2ccZC _3k0pr _3VHh4 _1IIOs _1gWrz _21Cwe _29SoL fYKJb _2kw8j _2kPml _21Cwe _29SoL\"><h5 class=\"_3lRU1 _1YIzb _7VPpv _1L59r c5\">NerdWallet Guide to COVID-19<\/h5><p class=\"_3mbPX _1L59r c4\">Get answers to questions about your mortgage, travel, finances \u2014 and maintaining your peace of mind.<\/p><\/div>\n<h2 class=\"_1Tx7p _104_p _1YIzb c7\">Preparing to buy tips<\/h2>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">1. Start saving early<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">Here are the main costs to consider when saving for a home:<\/p><\/div>\n<div class=\"_3mbPX\"><ul class=\"_19hnS _1jLZP\"><li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Down payment: Your down payment requirement will depend on the type of mortgage you choose and the lender. Some conventional loans aimed at first-time home buyers with excellent credit allow <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/fannie-mae-home-ready-freddie-mac-home-possible\/\" target=\"_self\">as little as 3% down<\/a>. But even a small down payment can be challenging to save. For example, a 3% down payment on a $300,000 home is $9,000. Use a <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/down-payment-calculator\/\" target=\"_self\">down payment calculator<\/a> to decide a goal, and then set up automatic transfers from checking to savings to get started.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Closing costs: These are the fees and expenses you pay to finalize your mortgage, and they typically range from 2% to 5% of the loan amount. You can ask the seller to pay a portion of your\u00a0<a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/closing-costs-calculator\" target=\"_self\">closing costs<\/a>, and you can save on some expenses, such as home inspections, by shopping around.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Move-in expenses: You'll need some cash after the home purchase. Set some money aside for immediate <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/save-on-home-improvements\/\" target=\"_self\">home repairs<\/a>, upgrades and furnishings.<\/p><\/div>\n<\/li>\n<\/ul><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">2. Decide how much home you can afford<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">Figure out how much you can safely spend on a house before starting to shop. NerdWallet's <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/mortgages\/how-much-house-can-i-afford\" target=\"_self\">home affordability calculator<\/a> can help with setting a price range based on your income, debt, down payment, credit score and where you plan to live.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">3. Check and strengthen your credit<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">Your credit score will determine whether you qualify for a mortgage and affect the interest rate lenders will offer. Take these steps to strengthen your <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/whats-exact-credit-score-need-buy-home\" target=\"_self\">credit score to buy a house<\/a>:<\/p><\/div>\n<div class=\"_3mbPX\"><ul class=\"_19hnS _1jLZP\"><li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Get free copies of your <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/finance\/how-to-use-annualcreditreport-com\" target=\"_self\">credit reports<\/a> from each of the three credit bureaus \u2014 Experian, Equifax and TransUnion \u2014 and dispute any errors that could hurt your score.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Pay all your bills on time, and keep credit card balances as low as possible.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Keep current credit cards open. Closing a card will increase the portion of available credit you use, which can lower your score.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Track your credit score. NerdWallet offers a <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/l\/free-credit-score\" target=\"_self\">free credit score<\/a> that updates weekly.<\/p><\/div>\n<\/li>\n<\/ul><\/div>\n<div class=\"_3mbPX o25yn\"><p class=\"ETtno _2zgPr _1YIzb\">Getting ready to buy a home? We\u2019ll find you a highly rated lender in just a few minutes.<\/p><p class=\"_2rIyf\">Enter your ZIP code to get started on a personalized lender match.<\/p><div><form action=\"\/mortgages\/mortgage-pre-approval\/match?trk=hlm_entry\">\n<p><label class=\"_3seyB _1HFo9 _1iuZG\" for=\"article-client-15-1-1\">ZIP code<\/label>\u00a0<\/p>\n<\/form><\/div><\/div>\n<h2 class=\"_1Tx7p _104_p _1YIzb c7\">Mortgage selection tips<\/h2>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">4. Explore mortgage options<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">A variety of mortgages are available with varying down payment and eligibility requirements. Here are the main categories:<\/p><\/div>\n<div class=\"_3mbPX\"><ul class=\"_19hnS _1jLZP\"><li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\"><a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/finding-the-right-mortgage\/conventional-mortgage\/\" target=\"_self\">Conventional mortgages<\/a> are not guaranteed by the government. Some conventional loans targeted at first-time buyers require as little as 3% down.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\"><a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/fha-loan\/\" target=\"_self\">FHA loans<\/a> are insured by the Federal Housing Administration and allow down payments as low as 3.5%.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\"><a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/usda-loan\" target=\"_self\">USDA loans<\/a> are guaranteed by the U.S. Department of Agriculture. They are for rural home buyers and usually require no down payment.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\"><a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/va-loan\/\" target=\"_self\">VA loans<\/a> are guaranteed by the Department of Veterans Affairs. They are for current and veteran military service members and usually require no down payment.<\/p><\/div>\n<\/li>\n<\/ul><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">You also have options when it comes to the mortgage term. Most home buyers opt for a 30-year fixed-rate mortgage, which is paid off in 30 years and has an interest rate that stays the same. A 15-year loan typically has a lower interest rate than a 30-year mortgage, but the monthly payments are larger.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">5. Research first-time home buyer assistance programs<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">Many states and some cities and counties offer first-time home buyer programs, which often combine low-interest-rate mortgages with <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/down-payment-assistance-help-buying-a-house\" target=\"_self\">down payment assistance<\/a> and closing cost assistance. Tax credits are also available through some first-time home buyer programs.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">6. Compare mortgage rates and fees<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">The Consumer Financial Protection Bureau recommends requesting loan estimates for the same type of mortgage from multiple lenders to compare the costs, including interest rates and possible <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/origination-fee\" target=\"_self\">origination fees<\/a>.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Lenders may offer the opportunity to buy <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/discount-points\/\" target=\"_self\">discount points<\/a>, which are fees the borrower pays upfront to lower the interest rate. Buying points can make sense if you have the money on hand and plan to stay in the home for a long time. Use a <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/should-i-buy-points-mortgage-calculator\" target=\"_self\">discount points calculator<\/a> to decide.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">7. Get a preapproval letter<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">A <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/how-to-get-a-mortgage-preapproval\" target=\"_self\">mortgage preapproval<\/a> is a lender's offer to loan you a certain amount under specific terms. Having a preapproval letter shows home sellers and real estate agents that you're a serious buyer, and can give you an edge over home shoppers who haven\u2019t taken this step yet.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Apply for preapproval when you're ready to start home shopping. A lender will pull your credit and review documents to verify your income, assets and debt. Applying for preapproval from more than one lender to shop rates shouldn't hurt your credit score as long as you apply for them within a limited time frame, such as 30 days.<\/p><\/div>\n<h2 class=\"_1Tx7p _104_p _1YIzb c7\">Home shopping tips<\/h2>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">8. Choose a real estate agent carefully<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">A good real estate agent will scour the market for homes that meet your needs and guide you through the negotiation and closing process. Get agent referrals from other recent home buyers. Interview at least a few agents, and request references. When speaking with potential agents, ask about their experience helping first-time home buyers in your market and how they plan to help you find a home.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">9. Pick the right type of house and neighborhood<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">Weigh the pros and cons of different types of homes, given your lifestyle and budget. A condominium or townhome may be more affordable than a <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/article\/mortgages\/single-family-home\" target=\"_self\">single-family home<\/a>, but shared walls with neighbors will mean less privacy. Don't forget to budget for <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/hoa-homeowners-association\/\" target=\"_self\">homeowners association<\/a> fees when shopping for condos and townhomes, or houses in planned or gated communities.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Another option to consider is buying a <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/buying-a-fixer-upper\/\" target=\"_self\">fixer-upper<\/a> \u2014 a single-family home in need of updates or repairs. Fixer-uppers usually sell for less per square foot than move-in ready homes. However, you may need to budget extra for repairs and remodeling. <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/renovation-loans-expand-homebuying-options\/\" target=\"_self\">Renovation mortgages<\/a> finance both the home price and the cost of improvements in one loan.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Think about your long-term needs and whether a\u00a0<a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/mortgages\/starter-home-forever-home\/\" target=\"_self\">starter home or forever home<\/a>\u00a0will meet them best. If you plan to start or expand your family, it may make sense to buy a home with extra room to grow.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Check out potential neighborhoods thoroughly. Choose one with amenities that are important to you, and test out the commute to work during rush hour.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">10. Stick to your budget<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">A lender may offer to loan you more than what is comfortably affordable, or you may feel pressure to spend outside your comfort zone to beat another buyer\u2019s offer. To avoid financial stress down the road, set a price range based on your budget, and then stick to it.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Look at properties below your price limit to give some wiggle room for bidding in a competitive market.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">11. Make the most of open houses<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">Online 3D home tours have become more popular amid the COVID-19 pandemic. These tours let shoppers virtually walk through a home at any hour and observe details that regular photos don't catch. They don't supply all the information in-person visits do \u2014 like how the carpets smell \u2014 but they can help you narrow the list of properties to visit.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Open your senses when touring homes in person. Listen for noise, pay attention to any odors and look at the overall condition of the home inside and out. Ask about the type and age of the electrical and plumbing systems and the roof.<\/p><\/div>\n<h2 class=\"_1Tx7p _104_p _1YIzb c7\">Home purchasing tips<\/h2>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">12. Pay for a home inspection<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">A home inspection is a thorough assessment of the structure and mechanical systems. Professional inspectors look for potential problems, so you can make an informed decision about buying the property. Here are some things to keep in mind:<\/p><\/div>\n<div class=\"_3mbPX\"><ul class=\"_19hnS _1jLZP\"><li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Standard inspections don\u2019t test for things like radon, mold or pests. Understand what's included in the inspection and what other inspections you might need.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Make sure the inspector can get to every part of the house, such as the roof and any crawl spaces.<\/p><\/div>\n<\/li>\n<li class=\"_35mws _32gdO VbKV8\">\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Traditionally the buyer attends the inspection. By following the inspector around you can get a better understanding of the home and ask questions on the spot. If you can't attend the inspection, review the inspector's report carefully and ask about anything that's unclear.<\/p><\/div>\n<\/li>\n<\/ul><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">13. Negotiate with the seller<\/h3>\n<div class=\"_3mbPX\"><p class=\"c4\">You may be able to save money by asking the seller to pay for repairs in advance or lower the price to cover the cost of repairs you\u2019ll have to make later. You may also ask the seller to pay some of the closing costs. But keep in mind that lenders may limit the portion of closing costs the seller can pay.<\/p><\/div>\n<div class=\"_3mbPX\"><p class=\"c4\">Your negotiating power will depend on the local market. It's tougher to drive a hard bargain when there are more buyers than homes for sale. Work with your real estate agent to understand the local market and strategize accordingly.<\/p><\/div>\n<h3 class=\"_1Tx7p _2zgPr _1YIzb c8\">14. Buy adequate home insurance<\/h3>\n<div class=\"_21Cwe _29SoL\"><p class=\"c4\">Your lender will require you to buy <a class=\"I1eqi _2_88b _6e2WU\" href=\"https:\/\/www.nerdwallet.com\/blog\/insurance\/understanding-homeowners-insurance\/\" target=\"_self\">homeowners insurance<\/a> before closing the deal. Home insurance covers the cost to repair or replace your home and belongings if they're damaged by an incident covered in the policy. It also provides liability insurance if you're held responsible for an injury or accident. Buy enough home insurance to cover the cost of rebuilding the home if it's destroyed.<\/p><\/div>","created_at":"2021-01-06T00:29:33+0000","updated_at":"2021-01-06T00:29:46+0000","published_at":"2018-05-07T21:23:06+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"www.nerdwallet.com","preview_picture":"https:\/\/www.nerdwallet.com\/assets\/blog\/wp-content\/uploads\/2018\/05\/GettyImages-1127946249-1440x864.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["databases","btree","datastructures"],"is_public":false,"id":1230,"title":"Part 7 - Introduction to the B-Tree","url":"https:\/\/cstack.github.io\/db_tutorial\/parts\/part7.html","content":"<header>\n<p>Writing a sqlite clone from scratch in C<\/p>\n<p class=\"view\"><a href=\"https:\/\/cstack.github.io\/db_tutorial\/\">Overview<\/a><\/p>\n<p class=\"view\"><a href=\"https:\/\/github.com\/cstack\/db_tutorial\">View on GitHub (pull requests welcome)<\/a><\/p>\n<\/header><section><p><a class=\"prev\" href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part6.html\">Part 6 - The Cursor Abstraction<\/a><br \/><\/p><p><a class=\"next\" href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part8.html\">Part 8 - B-Tree Leaf Node Format<\/a><br \/><\/p><p>The B-Tree is the data structure SQLite uses to represent both tables and indexes, so it\u2019s a pretty central idea. This article will just introduce the data structure, so it won\u2019t have any code.<\/p><p>Why is a tree a good data structure for a database?<\/p><ul><li>Searching for a particular value is fast (logarithmic time)<\/li>\n<li>Inserting \/ deleting a value you\u2019ve already found is fast (constant-ish time to rebalance)<\/li>\n<li>Traversing a range of values is fast (unlike a hash map)<\/li>\n<\/ul><p>A B-Tree is different from a binary tree (the \u201cB\u201d probably stands for the inventor\u2019s name, but could also stand for \u201cbalanced\u201d). Here\u2019s an example B-Tree:<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">example B-Tree (https:\/\/en.wikipedia.org\/wiki\/File:B-tree.svg)<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/B-tree.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/B-tree.png\" alt=\"example B-Tree (https:\/\/en.wikipedia.org\/wiki\/File:B-tree.svg)\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>Unlike a binary tree, each node in a B-Tree can have more than 2 children. Each node can have up to m children, where m is called the tree\u2019s \u201corder\u201d. To keep the tree mostly balanced, we also say nodes have to have at least m\/2 children (rounded up).<\/p><p>Exceptions:<\/p><ul><li>Leaf nodes have 0 children<\/li>\n<li>The root node can have fewer than m children but must have at least 2<\/li>\n<li>If the root node is a leaf node (the only node), it still has 0 children<\/li>\n<\/ul><p>The picture from above is a B-Tree, which SQLite uses to store indexes. To store tables, SQLites uses a variation called a B+ tree.<\/p><table><thead><tr><th>\u00a0<\/th>\n<th>B-tree<\/th>\n<th>B+ tree<\/th>\n<\/tr><\/thead><tbody><tr><td>Pronounced<\/td>\n<td>\u201cBee Tree\u201d<\/td>\n<td>\u201cBee Plus Tree\u201d<\/td>\n<\/tr><tr><td>Used to store<\/td>\n<td>Indexes<\/td>\n<td>Tables<\/td>\n<\/tr><tr><td>Internal nodes store keys<\/td>\n<td>Yes<\/td>\n<td>Yes<\/td>\n<\/tr><tr><td>Internal nodes store values<\/td>\n<td>Yes<\/td>\n<td>No<\/td>\n<\/tr><tr><td>Number of children per node<\/td>\n<td>Less<\/td>\n<td>More<\/td>\n<\/tr><tr><td>Internal nodes vs. leaf nodes<\/td>\n<td>Same structure<\/td>\n<td>Different structure<\/td>\n<\/tr><\/tbody><\/table><p>Until we get to implementing indexes, I\u2019m going to talk solely about B+ trees, but I\u2019ll just refer to it as a B-tree or a btree.<\/p><p>Nodes with children are called \u201cinternal\u201d nodes. Internal nodes and leaf nodes are structured differently:<\/p><table><thead><tr><th>For an order-m tree\u2026<\/th>\n<th>Internal Node<\/th>\n<th>Leaf Node<\/th>\n<\/tr><\/thead><tbody><tr><td>Stores<\/td>\n<td>keys and pointers to children<\/td>\n<td>keys and values<\/td>\n<\/tr><tr><td>Number of keys<\/td>\n<td>up to m-1<\/td>\n<td>as many as will fit<\/td>\n<\/tr><tr><td>Number of pointers<\/td>\n<td>number of keys + 1<\/td>\n<td>none<\/td>\n<\/tr><tr><td>Number of values<\/td>\n<td>none<\/td>\n<td>number of keys<\/td>\n<\/tr><tr><td>Key purpose<\/td>\n<td>used for routing<\/td>\n<td>paired with value<\/td>\n<\/tr><tr><td>Stores values?<\/td>\n<td>No<\/td>\n<td>Yes<\/td>\n<\/tr><\/tbody><\/table><p>Let\u2019s work through an example to see how a B-tree grows as you insert elements into it. To keep things simple, the tree will be order 3. That means:<\/p><ul><li>up to 3 children per internal node<\/li>\n<li>up to 2 keys per internal node<\/li>\n<li>at least 2 children per internal node<\/li>\n<li>at least 1 key per internal node<\/li>\n<\/ul><p>An empty B-tree has a single node: the root node. The root node starts as a leaf node with zero key\/value pairs:<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">empty btree<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree1.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree1.png\" alt=\"empty btree\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>If we insert a couple key\/value pairs, they are stored in the leaf node in sorted order.<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">one-node btree<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree2.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree2.png\" alt=\"one-node btree\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>Let\u2019s say that the capacity of a leaf node is two key\/value pairs. When we insert another, we have to split the leaf node and put half the pairs in each node. Both nodes become children of a new internal node which will now be the root node.<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">two-level btree<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree3.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree3.png\" alt=\"two-level btree\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>The internal node has 1 key and 2 pointers to child nodes. If we want to look up a key that is less than or equal to 5, we look in the left child. If we want to look up a key greater than 5, we look in the right child.<\/p><p>Now let\u2019s insert the key \u201c2\u201d. First we look up which leaf node it would be in if it was present, and we arrive at the left leaf node. The node is full, so we split the leaf node and create a new entry in the parent node.<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">four-node btree<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree4.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree4.png\" alt=\"four-node btree\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>Let\u2019s keep adding keys. 18 and 21. We get to the point where we have to split again, but there\u2019s no room in the parent node for another key\/pointer pair.<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">no room in internal node<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree5.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree5.png\" alt=\"no room in internal node\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>The solution is to split the root node into two internal nodes, then create new root node to be their parent.<\/p><table class=\"image\"><caption style=\"text-align: bottom;\">three-level btree<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree6.png\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/btree6.png\" alt=\"three-level btree\" referrerpolicy=\"no-referrer\" \/><\/a><\/td>\n<\/tr><\/table><p>The depth of the tree only increases when we split the root node. Every leaf node has the same depth and close to the same number of key\/value pairs, so the tree remains balanced and quick to search.<\/p><p>I\u2019m going to hold off on discussion of deleting keys from the tree until after we\u2019ve implemented insertion.<\/p><p>When we implement this data structure, each node will correspond to one page. The root node will exist in page 0. Child pointers will simply be the page number that contains the child node.<\/p><p>Next time, we start implementing the btree!<\/p><p><a class=\"prev\" href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part6.html\">Part 6 - The Cursor Abstraction<\/a><br \/><\/p><p><a class=\"next\" href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part8.html\">Part 8 - B-Tree Leaf Node Format<\/a><br \/><\/p><\/section>","created_at":"2021-01-02T02:45:21+0000","updated_at":"2021-01-02T02:45:35+0000","published_at":"2017-09-23T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"cstack.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["devops","terraform","consul","packer"],"is_public":false,"id":1229,"title":"hashicorp\/terraform-aws-consul","url":"https:\/\/github.com\/hashicorp\/terraform-aws-consul","content":"<p><a href=\"https:\/\/gruntwork.io\/?ref=repo_aws_consul\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/6fec0aebc8ed85691ff383ed261e2aaaf4e1be5512a0eceecfbcd9f6e72f913a\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6d61696e7461696e656425323062792d6772756e74776f726b2e696f2d2532333538343961362e737667\" alt=\"Maintained by Gruntwork.io\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/maintained%20by-gruntwork.io-%235849a6.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>This repo contains a set of modules in the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\">modules folder<\/a> for deploying a <a href=\"https:\/\/www.consul.io\/\" rel=\"nofollow\">Consul<\/a> cluster on <a href=\"https:\/\/aws.amazon.com\/\" rel=\"nofollow\">AWS<\/a> using <a href=\"https:\/\/www.terraform.io\/\" rel=\"nofollow\">Terraform<\/a>. Consul is a distributed, highly-available tool that you can use for service discovery and key\/value storage. A Consul cluster typically includes a small number of server nodes, which are responsible for being part of the <a href=\"https:\/\/www.consul.io\/docs\/internals\/consensus.html\" rel=\"nofollow\">consensus quorum<\/a>, and a larger number of client nodes, which you typically run alongside your apps:<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/blob\/master\/_docs\/architecture.png?raw=true\"><img src=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/raw\/master\/_docs\/architecture.png?raw=true\" alt=\"Consul architecture\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-how-to-use-this-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-use-this-module\"><\/a>How to use this Module<\/h2><p>This repo has the following folder structure:<\/p><ul><li><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\">modules<\/a>: This folder contains several standalone, reusable, production-grade modules that you can use to deploy Consul.<\/li>\n<li><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/examples\">examples<\/a>: This folder shows examples of different ways to combine the modules in the <code>modules<\/code> folder to deploy Consul.<\/li>\n<li><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/test\">test<\/a>: Automated tests for the modules and examples.<\/li>\n<li><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\">root folder<\/a>: The root folder is <em>an example<\/em> of how to use the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/consul-cluster\">consul-cluster module<\/a> module to deploy a <a href=\"https:\/\/www.consul.io\/\" rel=\"nofollow\">Consul<\/a> cluster in <a href=\"https:\/\/aws.amazon.com\/\" rel=\"nofollow\">AWS<\/a>. The Terraform Registry requires the root of every repo to contain Terraform code, so we've put one of the examples there. This example is great for learning and experimenting, but for production use, please use the underlying modules in the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\">modules folder<\/a> directly.<\/li>\n<\/ul><p>To deploy Consul servers for production using this repo:<\/p><ol><li>\n<p>Create a Consul AMI using a Packer template that references the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/install-consul\">install-consul module<\/a>. Here is an <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/examples\/consul-ami#quick-start\">example Packer template<\/a>.<\/p>\n<p>If you are just experimenting with this Module, you may find it more convenient to use one of our official public AMIs. Check out the <code>aws_ami<\/code> data source usage in <code>main.tf<\/code> for how to auto-discover this AMI.<\/p>\n<p><strong>WARNING! Do NOT use these AMIs in your production setup. In production, you should build your own AMIs in your own AWS account.<\/strong><\/p>\n<\/li>\n<li>\n<p>Deploy that AMI across an Auto Scaling Group using the Terraform <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/consul-cluster\">consul-cluster module<\/a> and execute the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/run-consul\">run-consul script<\/a> with the <code>--server<\/code> flag during boot on each Instance in the Auto Scaling Group to form the Consul cluster. Here is <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/examples\/root-example#quick-start\">an example Terraform configuration<\/a> to provision a Consul cluster.<\/p>\n<\/li>\n<\/ol><p>To deploy Consul clients for production using this repo:<\/p><ol><li>Use the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/install-consul\">install-consul module<\/a> to install Consul alongside your application code.<\/li>\n<li>Before booting your app, execute the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/run-consul\">run-consul script<\/a> with <code>--client<\/code> flag.<\/li>\n<li>Your app can now use the local Consul agent for service discovery and key\/value storage.<\/li>\n<li>Optionally, you can use the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/install-dnsmasq\">install-dnsmasq module<\/a> for Ubuntu 16.04 and Amazon Linux 2 or <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/setup-systemd-resolved\">setup-systemd-resolved<\/a> for Ubuntu 18.04 to configure Consul as the DNS for a specific domain (e.g. <code>.consul<\/code>) so that URLs such as <code>foo.service.consul<\/code> resolve automatically to the IP address(es) for a service <code>foo<\/code> registered in Consul (all other domain names will be continue to resolve using the default resolver on the OS).<\/li>\n<\/ol><h2><a id=\"user-content-whats-a-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-a-module\"><\/a>What's a Module?<\/h2><p>A Module is a canonical, reusable, best-practices definition for how to run a single piece of infrastructure, such as a database or server cluster. Each Module is created using <a href=\"https:\/\/www.terraform.io\/\" rel=\"nofollow\">Terraform<\/a>, and includes automated tests, examples, and documentation. It is maintained both by the open source community and companies that provide commercial support.<\/p><p>Instead of figuring out the details of how to run a piece of infrastructure from scratch, you can reuse existing code that has been proven in production. And instead of maintaining all that infrastructure code yourself, you can leverage the work of the Module community to pick up infrastructure improvements through a version number bump.<\/p><h2><a id=\"user-content-who-maintains-this-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#who-maintains-this-module\"><\/a>Who maintains this Module?<\/h2><p>This Module is maintained by <a href=\"http:\/\/www.gruntwork.io\/\" rel=\"nofollow\">Gruntwork<\/a>. If you're looking for help or commercial support, send an email to <a href=\"mailto:modules@gruntwork.io?Subject=Consul%20Module\">modules@gruntwork.io<\/a>. Gruntwork can help with:<\/p><ul><li>Setup, customization, and support for this Module.<\/li>\n<li>Modules for other types of infrastructure, such as VPCs, Docker clusters, databases, and continuous integration.<\/li>\n<li>Modules that meet compliance requirements, such as HIPAA.<\/li>\n<li>Consulting &amp; Training on AWS, Terraform, and DevOps.<\/li>\n<\/ul><h2><a id=\"user-content-code-included-in-this-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#code-included-in-this-module\"><\/a>Code included in this Module:<\/h2><ul><li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/install-consul\">install-consul<\/a>: This module installs Consul using a <a href=\"https:\/\/www.packer.io\/\" rel=\"nofollow\">Packer<\/a> template to create a Consul <a href=\"http:\/\/docs.aws.amazon.com\/AWSEC2\/latest\/UserGuide\/AMIs.html\" rel=\"nofollow\">Amazon Machine Image (AMI)<\/a>.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/consul-cluster\">consul-cluster<\/a>: The module includes Terraform code to deploy a Consul AMI across an <a href=\"https:\/\/aws.amazon.com\/autoscaling\/\" rel=\"nofollow\">Auto Scaling Group<\/a>.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/run-consul\">run-consul<\/a>: This module includes the scripts to configure and run Consul. It is used by the above Packer module at build-time to set configurations, and by the Terraform module at runtime with <a href=\"http:\/\/docs.aws.amazon.com\/AWSEC2\/latest\/UserGuide\/user-data.html#user-data-shell-scripts\" rel=\"nofollow\">User Data<\/a> to create the cluster.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/install-dnsmasq\">install-dnsmasq module<\/a>: Install <a href=\"http:\/\/www.thekelleys.org.uk\/dnsmasq\/doc.html\" rel=\"nofollow\">Dnsmasq<\/a> for Ubuntu 16.04 and Amazon Linux 2 and configure it to forward requests for a specific domain to Consul. This allows you to use Consul as a DNS server for URLs such as <code>foo.service.consul<\/code>.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/setup-systemd-resolved\">setup-systemd-resolved module<\/a>: Setup <a href=\"https:\/\/www.freedesktop.org\/software\/systemd\/man\/resolved.conf.html\" rel=\"nofollow\">systemd-resolved<\/a> for ubuntu 18.04 and configure it to forward requests for a specific domain to Consul. This allows you to use Consul as a DNS server for URLs such as <code>foo.service.consul<\/code>.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/consul-iam-policies\">consul-iam-policies<\/a>: Defines the IAM policies necessary for a Consul cluster.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/consul-security-group-rules\">consul-security-group-rules<\/a>: Defines the security group rules used by a Consul cluster to control the traffic that is allowed to go in and out of the cluster.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/modules\/consul-client-security-group-rules\">consul-client-security-group-rules<\/a>: Defines the security group rules used by a Consul agent to control the traffic that is allowed to go in and out.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-how-do-i-contribute-to-this-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-contribute-to-this-module\"><\/a>How do I contribute to this Module?<\/h2><p>Contributions are very welcome! Check out the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/CONTRIBUTING.md\">Contribution Guidelines<\/a> for instructions.<\/p><h2><a id=\"user-content-how-is-this-module-versioned\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-is-this-module-versioned\"><\/a>How is this Module versioned?<\/h2><p>This Module follows the principles of <a href=\"http:\/\/semver.org\/\" rel=\"nofollow\">Semantic Versioning<\/a>. You can find each new release, along with the changelog, in the <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/releases\">Releases Page<\/a>.<\/p><p>During initial development, the major version will be 0 (e.g., <code>0.x.y<\/code>), which indicates the code does not yet have a stable API. Once we hit <code>1.0.0<\/code>, we will make every effort to maintain a backwards compatible API and use the MAJOR, MINOR, and PATCH versions on each release to indicate any incompatibilities.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>This code is released under the Apache 2.0 License. Please see <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/LICENSE\">LICENSE<\/a> and <a href=\"https:\/\/github.com\/hashicorp\/terraform-aws-consul\/tree\/master\/NOTICE\">NOTICE<\/a> for more details.<\/p><p>Copyright \u00a9 2017 Gruntwork, Inc.<\/p>","created_at":"2020-12-31T14:00:54+0000","updated_at":"2020-12-31T14:01:06+0000","published_by":["hashicorp"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/761456?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1228,"title":"Adding comments to your static blog with Mastodon","url":"https:\/\/news.ycombinator.com\/item?id=25570268","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\" style=\"border-spacing: 0px; width: 100%;\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D25570268\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"25570268\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"https:\/\/carlschwan.eu\/2020\/12\/29\/adding-comments-to-your-static-blog-with-mastodon\/\" class=\"storylink\">Adding comments to your static blog with Mastodon<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=carlschwan.eu\">carlschwan.eu<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">105 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=ognarb\" class=\"hnuser\">ognarb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=25570268\">5 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=25570268&amp;goto=item%3Fid%3D25570268\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Adding%20comments%20to%20your%20static%20blog%20with%20Mastodon&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=25570268&amp;auth=02750b652a49d576bf8da0710d895d01c43d3456\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=25570268\">25\u00a0comments<\/a><\/td>\n<\/tr><tr><td>\n<\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2020-12-29T20:14:47+0000","updated_at":"2020-12-29T20:14:47+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1227,"title":"VisiData","url":"https:\/\/www.visidata.org\/","content":"VisiData\n<hr \/><section id=\"body\"><p>VisiData is an interactive multitool for tabular data. It combines the clarity of a spreadsheet, the efficiency of the terminal, and the power of Python, into a lightweight utility which can handle millions of rows with ease.<\/p><p>When you have  but you need :<\/p><p id=\"cmdouter\"><code id=\"cmdline\">vd -b input.csv<\/code><\/p><h4>Latest Version: <a href=\"https:\/\/www.visidata.org\/blog\/2020\/v2.1\/\">2.1<\/a> (2020-12-06)<\/h4><table><tr><td><a class=\"\" href=\"https:\/\/www.visidata.org\/formats\">Any data format!\n<div class=\"screenshot\"><img src=\"https:\/\/www.visidata.org\/basic-screenshot.png\" alt=\"screenshot of VisiData in action\" referrerpolicy=\"no-referrer\" \/><\/div>\nOpen 1m+ rows in the blink of an eye!<\/a><\/td>\n<td><a class=\"\" href=\"https:\/\/jsvine.github.io\/intro-to-visidata\/the-big-picture\/visidata-in-60-seconds\/\">Instant histogram with Shift+F!\n<img src=\"https:\/\/www.visidata.org\/freq-move-row.gif\" alt=\"Frequency Table loading, with cursor movement\" referrerpolicy=\"no-referrer\" \/><\/a> Press Enter to dive in!<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.youtube.com\/watch?v=N1CBDTgGtOU&amp;t=134\">Scatterplots in the terminal!\n<div class=\"screenshot\"><img src=\"https:\/\/www.visidata.org\/map-colours.png\" alt=\"Plotting a map with colours\" referrerpolicy=\"no-referrer\" \/><\/div>\n8-color maps!<\/a><\/td>\n<td><a href=\"https:\/\/www.visidata.org\/docs\/save-restore\">Save sessions for replay later!\n<img src=\"https:\/\/www.visidata.org\/scatterplot-usage.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/a> Adhoc data pipelines in batch mode! <\/td>\n<\/tr><\/table><\/section><hr \/>","created_at":"2020-12-25T02:24:29+0000","updated_at":"2020-12-25T02:24:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.visidata.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1226,"title":"An open source multi-tool for exploring and publishing data","url":"https:\/\/datasette.io\/","content":"<div class=\"two-col-column\"><h2>Latest news <a class=\"c2\" href=\"https:\/\/datasette.io\/content\/feed.atom\" title=\"Atom feed\"><\/a><\/h2><h3>19th December 2020 <a href=\"https:\/\/datasette.io\/news\/2020-12-19\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p>New on this site: a <a href=\"https:\/\/datasette.io\/tools\" rel=\"nofollow\">Datasette Tools<\/a> directory and a <a href=\"https:\/\/datasette.io\/-\/beta\" rel=\"nofollow\">search engine<\/a> that covers documentation, tools, plugins, releases and more. The search engine uses <a href=\"https:\/\/github.com\/dogsheep\/dogsheep-beta\" rel=\"nofollow\">Dogsheep Beta<\/a> - I wrote about how that works in <a href=\"https:\/\/simonwillison.net\/2020\/Dec\/19\/dogsheep-beta\/\" rel=\"nofollow\">Building a search engine for datasette.io<\/a>.<\/p><\/div><h3>13th December 2020 <a href=\"https:\/\/datasette.io\/news\/2020-12-13\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p>I wrote more about how this site works in <a href=\"https:\/\/simonwillison.net\/2020\/Dec\/13\/datasette-io\/\" rel=\"nofollow\">Weeknotes: datasette.io, an official project website for Datasette<\/a>.<\/p><\/div><h3>10th December 2020 <a href=\"https:\/\/datasette.io\/news\/2020-12-10\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-52\" rel=\"nofollow\">Datasette 0.53<\/a> - <code>datasette serve --create<\/code> option, <code>?column__arraynotcontains=<\/code> table filter, <code>?_header=off<\/code> CSV option and updated links to reflect the new project website.<\/p><\/div><h3>28th November 2020 <a href=\"https:\/\/datasette.io\/news\/2020-11-28\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-52\" rel=\"nofollow\">Datasette 0.52<\/a> - <code>--config<\/code> is now <code>--setting<\/code>, new <code>database_actions<\/code> plugin hook, <code>datasette publish cloudrun --apt-get-install<\/code> option and several bug fixes.<\/p><\/div><h3>31st October 2020 <a href=\"https:\/\/datasette.io\/news\/2020-10-31\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-51\" rel=\"nofollow\">Datasette 0.51<\/a> - A new visual design, plugin hooks for adding navigation options, better handling of binary data, URL building utility methods and better support for running Datasette behind a proxy. <a href=\"https:\/\/simonwillison.net\/2020\/Nov\/1\/datasette-0-51\/\" rel=\"nofollow\">Annotated release notes<\/a>.<\/p><\/div><h3>9th October 2020 <a href=\"https:\/\/datasette.io\/news\/2020-10-09\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-50\" rel=\"nofollow\">Datasette 0.50<\/a> - New column actions menu. <code>datasette.client<\/code> object for plugins to make internal API requests. Improved documentation on deploying Datasette. <a href=\"https:\/\/simonwillison.net\/2020\/Oct\/9\/datasette-0-50\/\" rel=\"nofollow\">Annotated release notes<\/a>.<\/p><\/div><h3>14th September 2020 <a href=\"https:\/\/datasette.io\/news\/2020-09-14\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-49\" rel=\"nofollow\">Datasette 0.49<\/a> - JSON API for writable canned queries, path parameters for custom pages. See also <a href=\"https:\/\/simonwillison.net\/2020\/Sep\/15\/datasette-0-49\/\" rel=\"nofollow\">Datasette 0.49: The annotated release notes<\/a>.<\/p><\/div><h3>16th August 2020 <a href=\"https:\/\/datasette.io\/news\/2020-08-16\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-48\" rel=\"nofollow\">Datasette 0.48<\/a> - Documentation now lives at <a href=\"https:\/\/docs.datasette.io\/\" rel=\"nofollow\">docs.datasette.io<\/a>, improvements to the <code>extra_template_vars<\/code>, <code>extra_css_urls<\/code>, <code>extra_js_urls<\/code> and <code>extra_body_script<\/code> plugin hooks.<\/p><\/div><h3>11th August 2020 <a href=\"https:\/\/datasette.io\/news\/2020-08-11\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-47\" rel=\"nofollow\">Datasette 0.47<\/a> - Datasette can now be installed using Homebrew! <code>brew install simonw\/datasette\/datasette<\/code>. Also new: <code>datasette install name-of-plugin<\/code> and <code>datasette uninstall name-of-plugin<\/code> commands, and <code>datasette --get '\/-\/versions.json'<\/code> to output the result of Datasette HTTP calls on the command-line.<\/p><\/div><h3>9th August 2020 <a href=\"https:\/\/datasette.io\/news\/2020-08-09\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-46\" rel=\"nofollow\">Datasette 0.46<\/a> - security fix relating to CSRF protection for writable canned queries, a new logo, new debugging tools, improved file downloads and more.<\/p><\/div><h3>6th August 2020 <a href=\"https:\/\/datasette.io\/news\/2020-08-06\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/simonwillison.net\/2020\/Aug\/7\/datasette-graphql\/\" rel=\"nofollow\">GraphQL in Datasette with the new datasette-graphql plugin<\/a><\/p><\/div><h3>24th July 2020 <a href=\"https:\/\/datasette.io\/news\/2020-07-24\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p>Two new plugins: <a href=\"https:\/\/simonwillison.net\/2020\/Jul\/23\/datasette-copyable-datasette-insert-api\/\" rel=\"nofollow\">datasette-copyable and datasette-insert-api<\/a>. <code>datasette-copyable<\/code> adds copy-and-paste export options, and <code>datasette-insert-api<\/code> lets you create tables and insert or update data by POSTing JSON directly to Datasette.<\/p><\/div><h3>1st July 2020 <a href=\"https:\/\/datasette.io\/news\/2020-07-01\" class=\"c3\">#<\/a><\/h3><h3>20th June 2020 <a href=\"https:\/\/datasette.io\/news\/2020-06-20\" class=\"c3\">#<\/a><\/h3><div class=\"c4\"><p><a href=\"https:\/\/simonwillison.net\/2020\/Jun\/20\/cookiecutter-plugins\/\" rel=\"nofollow\">A cookiecutter template for writing Datasette plugins<\/a><\/p><\/div><p><a href=\"https:\/\/datasette.io\/news\">All news<\/a><\/p><\/div><div class=\"two-col-column\"><h2>Latest releases<\/h2><h3>19th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/dogsheep\/dogsheep-beta\/releases\/tag\/0.10.1\">dogsheep-beta 0.10.1<\/a><\/h4><div class=\"c4\"><ul><li>Fix for error if search query contains hyphens such as <code>github-to-sqlite<\/code>. #31<\/li>\n<\/ul><\/div><h3>16th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/dogsheep\/dogsheep-beta\/releases\/tag\/0.10\">dogsheep-beta 0.10<\/a><\/h4><div class=\"c4\"><ul><li>Now depends on <a href=\"https:\/\/github.com\/simonw\/sqlite-utils\" rel=\"nofollow\">sqlite-utils<\/a> &gt;= 3.0. #30<\/li>\n<li>The user's original search term is now passed to the <code>display_sql<\/code> SQL query as the <code>:q<\/code> parameter. #29<\/li>\n<\/ul><\/div><h4><a href=\"https:\/\/github.com\/dogsheep\/github-to-sqlite\/releases\/tag\/2.8.2\">github-to-sqlite 2.8.2<\/a><\/h4><div class=\"c4\"><ul><li><code>--readme<\/code> now stores <code>null<\/code> if the README file does not exist, rather than throwing an error. #57<\/li>\n<li>README HTML is rewritten to fix broken internal links, e.g. for a table of contents. #58<\/li>\n<\/ul><\/div><h3>14th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/simonw\/datasette-publish-vercel\/releases\/tag\/0.9\">datasette-publish-vercel 0.9<\/a><\/h4><div class=\"c4\"><ul><li>New <code>--setting<\/code> option for specifying Datasette settings, as a working alternative to <code>--extra-options<\/code>. #30<\/li>\n<\/ul><\/div><h3>13th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/simonw\/sqlite-utils\/releases\/tag\/3.1\">sqlite-utils 3.1<\/a><\/h4><div class=\"c4\"><ul><li>New command: <code>sqlite-utils analyze-tables my.db<\/code> outputs useful information about the table columns in the database, such as the number of distinct values and how many rows are null. See <a href=\"https:\/\/sqlite-utils.readthedocs.io\/en\/stable\/cli.html#cli-analyze-tables\" rel=\"nofollow\">Analyzing tables<\/a> for documentation. (<a href=\"https:\/\/github.com\/simonw\/sqlite-utils\/issues\/207\" rel=\"nofollow\">#207<\/a>)<\/li>\n<li>New <code>table.analyze_column(column)<\/code> Python method used by the <code>analyze-tables<\/code> command - see <a href=\"https:\/\/sqlite-utils.readthedocs.io\/en\/stable\/python-api.html#python-api-analyze-column\" rel=\"nofollow\">Analyzing a column<\/a>.<\/li>\n<li>The <code>table.update()<\/code> method now correctly handles values that should be stored as JSON. Thanks, Andreas Madsack. (<a href=\"https:\/\/github.com\/simonw\/sqlite-utils\/pull\/204\" rel=\"nofollow\">#204<\/a>)<\/li>\n<\/ul><\/div><h3>12th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/simonw\/datasette-publish-fly\/releases\/tag\/1.0.1\">datasette-publish-fly 1.0.1<\/a><\/h4><div class=\"c4\"><ul><li>Fixed various errors that occurred when using latest version of <code>flyctl<\/code>. #7<\/li>\n<\/ul><\/div><h3>11th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/simonw\/datasette-auth-passwords\/releases\/tag\/0.3.3\">datasette-auth-passwords 0.3.3<\/a><\/h4><div class=\"c4\"><ul><li>Fixed bug where users with a <code>\"someusername_password_hash\"<\/code> value defined in <code>metadata.json<\/code> but no corresponding <code>\"actor\"<\/code> record were incorrectly signed in as an actor with <code>\"id\": \"username\"<\/code>, when they should have been signed in as an actor with <code>\"id\": \"someusername\"<\/code>. #14<\/li>\n<\/ul><\/div><h4><a href=\"https:\/\/github.com\/simonw\/datasette\/releases\/tag\/0.53\">datasette 0.53<\/a><\/h4><div class=\"c4\"><p>Datasette has an official project website now, at <a href=\"https:\/\/datasette.io\/\" rel=\"nofollow\">https:\/\/datasette.io\/<\/a>. This release mainly updates the documentation to reflect the new site.<\/p><ul><li>New <code>?column__arraynotcontains=<\/code> table filter. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1132\" rel=\"nofollow\">#1132<\/a>)<\/li>\n<li><code>datasette serve<\/code> has a new <code>--create<\/code> option, which will create blank database files if they do not already exist rather than exiting with an error. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1135\" rel=\"nofollow\">#1135<\/a>)<\/li>\n<li>New <code>?_header=off<\/code> option for CSV export which omits the CSV header row, <a href=\"https:\/\/docs.datasette.io\/en\/stable\/csv_export.html#csv-export-url-parameters\" rel=\"nofollow\">documented here<\/a>. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1133\" rel=\"nofollow\">#1133<\/a>)<\/li>\n<li>\"Powered by Datasette\" link in the footer now links to <a href=\"https:\/\/datasette.io\/\" rel=\"nofollow\">https:\/\/datasette.io\/<\/a>. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1138\" rel=\"nofollow\">#1138<\/a>)<\/li>\n<li>Project news no longer lives in the README - it can now be found at <a href=\"https:\/\/datasette.io\/news\" rel=\"nofollow\">https:\/\/datasette.io\/news<\/a>. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1137\" rel=\"nofollow\">#1137<\/a>)<\/li>\n<\/ul><\/div><h3>9th December 2020<\/h3><h4><a href=\"https:\/\/github.com\/simonw\/datasette-column-inspect\/releases\/tag\/0.2a\">datasette-column-inspect 0.2a<\/a><\/h4><div class=\"c4\"><ul><li>Correctly parse column links. #3<\/li>\n<li>Show a message encouraging mobile users to rotate to landscape. #2<\/li>\n<\/ul><\/div><h4><a href=\"https:\/\/github.com\/simonw\/datasette-pretty-json\/releases\/tag\/0.2.1\">datasette-pretty-json 0.2.1<\/a><\/h4><div class=\"c4\"><ul><li>Switched CI to GitHub Actions<\/li>\n<\/ul><\/div><p><a href=\"https:\/\/datasette.io\/content\/recent_releases\">All releases<\/a><\/p><\/div>","created_at":"2020-12-25T02:22:49+0000","updated_at":"2020-12-25T02:22:49+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"datasette.io","http_status":"200","origin_url":"Datasette.io"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1225,"title":"datasette.io, an official project website for Datasette","url":"https:\/\/simonwillison.net\/2020\/Dec\/13\/datasette-io\/","content":"<p>This week I launched <a href=\"https:\/\/datasette.io\/\">datasette.io<\/a>\u2014the new official project website for Datasette.<\/p><p>Datasette\u2019s first open source release was <a href=\"https:\/\/simonwillison.net\/2017\/Nov\/13\/datasette\/\">just over three years ago<\/a>, but until now the official site duties have been split between the <a href=\"https:\/\/github.com\/simonw\/datasette\">GitHub repository<\/a> and <a href=\"https:\/\/docs.datasette.io\/\">the documentation<\/a>.<\/p><p><img alt=\"A screenshot of datasette.io\" src=\"https:\/\/static.simonwillison.net\/static\/2020\/datasette-io.png\" class=\"c1\" referrerpolicy=\"no-referrer\" \/><\/p><h4>The Baked Data architectural pattern<\/h4><p>The site itself is built on Datasette (<a href=\"https:\/\/github.com\/simonw\/datasette.io\">source code here<\/a>). I\u2019m using a pattern that I <a href=\"https:\/\/simonwillison.net\/2019\/Oct\/28\/niche-museums-kepler\/\">first started exploring<\/a> with <a href=\"https:\/\/www.niche-museums.com\/\">Niche Museums<\/a>: most of the site content lives in a SQLite database, and I use custom Jinja templates to implement the site\u2019s different pages.<\/p><p>This is effectively a variant of the static site generator pattern. The SQLite database is built by scripts as part of the deploy process, then deployed to Google Cloud Run as a binary asset bundled with the templates and Datasette itself.<\/p><p>I call this the <strong>Baked Data<\/strong> architectural pattern\u2014with credit to Kevin Marks for <a href=\"https:\/\/chat.indieweb.org\/dev\/2020-11-15#t1605478365993400\">helping me<\/a> coin the right term. You bake the data into the application.<\/p><p>It\u2019s comparable to static site generation because everything is immutable, which greatly reduces the amount of things that can go wrong\u2014and any content changes require a fresh deploy. It\u2019s extremely easy to scale\u2014just run more copies of the application with the bundled copy of the database. Cloud Run and other serverless providers handle that kind of scaling automatically.<\/p><p>Unlike static site generation, if a site has a thousand pages you don\u2019t need to build a thousand HTML pages in order to deploy. A single template and a SQL query that incorporates arguments from the URL can serve as many pages as there are records in the database.<\/p><h4>How the site is built<\/h4><p>You can browse the site\u2019s underlying database tables in Datasette <a href=\"https:\/\/datasette.io\/content\">here<\/a>.<\/p><p>The <a href=\"https:\/\/datasette.io\/content\/news\">news table<\/a> powers the latest news on the homepage and <a href=\"https:\/\/datasette.io\/news\">\/news<\/a>. News lives in a <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/main\/news.yaml\">news.yaml file<\/a> in the site\u2019s GitHub repository. I wrote <a href=\"https:\/\/gist.github.com\/simonw\/6a59833eee83bec1f1317c7f80406275\">a script<\/a> to import the news that had been accumulating in <a href=\"https:\/\/github.com\/simonw\/datasette\/blob\/0.52.5\/README.md\">the 0.52 README<\/a>\u2014now that news has moved to the site the README is a lot more slim!<\/p><p>At build time my <a href=\"https:\/\/github.com\/simonw\/yaml-to-sqlite\">yaml-to-sqlite<\/a> script runs to load that news content into a database table.<\/p><p>The <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/72e3b6a470d2543dede13d51a61a22180b5c96f9\/templates\/index.html#L78-L85\">index.html<\/a> template then uses the following Jinja code to output the latest news stories, using the <code>sql()<\/code> function from the <a href=\"https:\/\/github.com\/simonw\/datasette-template-sql\">datasette-template-sql<\/a> Datasette plugin:<\/p><div class=\"highlight highlight-text-html-django\"><pre>{% set ns = namespace(current_date=\"\") %}\n{% for row in sql(\"select date, body from news order by date desc limit 15\", database=\"content\") %}\n    {% if prettydate(row[\"date\"]) != (ns.current_date and prettydate(ns.current_date)) %}\n    &lt;h3&gt;{{ prettydate(row[\"date\"]) }} &lt;a href=\"\/news\/{{ row[\"date\"] }}\" style=\"font-size: 0.8em; opacity: 0.4\"&gt;#&lt;\/a&gt;&lt;\/h3&gt;\n    {% set ns.current_date = prettydate(row[\"date\"]) %}\n    {% endif %}\n    {{ render_markdown(row[\"body\"]) }}\n{% endfor %}<\/pre><\/div><p><code>prettydate()<\/code> is a custom function I wrote in <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/353f1f940cb0cb3c38d3bdf6e345328740990702\/plugins\/dateformat.py\">a one-off plugin<\/a> for the site. The <code>namespace()<\/code> stuff is a Jinja trick that lets me keep track of the current date heading in the loop, so I can output a new date heading only if the news item occurs on a different day from the previous one.<\/p><p><code>render_markdown()<\/code> is provided by the <a href=\"https:\/\/github.com\/simonw\/datasette-render-markdown\">datasette-render-markdown<\/a> plugin.<\/p><p>I wanted permalinks for news stories, but since they don\u2019t have identifiers or titles I decided to provide a page for each day instead\u2014for example <a href=\"https:\/\/datasette.io\/news\/2020-12-10\">https:\/\/datasette.io\/news\/2020-12-10<\/a><\/p><p>These pages are implemented using <a href=\"https:\/\/simonwillison.net\/2020\/Sep\/15\/datasette-0-49\/#path-parameters-custom-page-templates\">Path parameters for custom page templates<\/a>, introduced in Datasette 0.49. The implementation is a single template file at <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/353f1f940cb0cb3c38d3bdf6e345328740990702\/templates\/pages\/news\/%7Byyyy%7D-%7Bmm%7D-%7Bdd%7D.html\">templates\/pages\/news\/{yyyy}-{mm}-{dd}.html<\/a>, the full contents of which is:<\/p><div class=\"highlight highlight-text-html-django\"><pre>{% extends \"page_base.html\" %}\n{% block title %}Datasette News: {{ prettydate(yyyy + \"-\" + mm + \"-\" + dd) }}{% endblock %}\n{% block content %}\n{% set stories = sql(\"select date, body from news where date = ? order by date desc\", [yyyy + \"-\" + mm + \"-\" + dd], database=\"content\") %}\n{% if not stories %}\n    {{ raise_404(\"News not found\") }}\n{% endif %}\n&lt;h1&gt;&lt;a href=\"\/news\"&gt;News&lt;\/a&gt;: {{ prettydate(yyyy + \"-\" + mm + \"-\" + dd) }}&lt;\/h1&gt;\n{% for row in stories %}\n    {{ render_markdown(row[\"body\"]) }}\n{% endfor %}\n{% endblock %}<\/pre><\/div><p>The crucial trick here is that, because the filename is <code>news\/{yyyy}-{mm}-{dd}.html<\/code>, a request to <code>\/news\/2020-12-10<\/code> will render that template with the <code>yyyy<\/code>, <code>mm<\/code> and <code>dd<\/code> template variables set to those values from the URL.<\/p><p>It can then execute a SQL query that incorporates those values. It assigns the results to a <code>stories<\/code> variable, then checks that at least one story was returned\u2014if not, it raises a 404 error.<\/p><p>See Datasette\u2019s <a href=\"https:\/\/docs.datasette.io\/en\/stable\/custom_templates.html#custom-pages\">custom pages documentation<\/a> for more details on how this all works.<\/p><p>The site also offers an <a href=\"https:\/\/datasette.io\/content\/feed.atom\">Atom feed<\/a> of recent news. This is powered by the <a href=\"https:\/\/github.com\/simonw\/datasette-atom\">datasette-atom<\/a> using the output of <a href=\"https:\/\/datasette.io\/content\/feed\">this canned SQL query<\/a>, with a <code>render_markdown()<\/code> SQL function provided by <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/353f1f940cb0cb3c38d3bdf6e345328740990702\/plugins\/sql_functions.py\">this site plugin<\/a>.<\/p><h4>The plugin directory<\/h4><p>One of the features I\u2019m most excited about on the site is the new <a href=\"https:\/\/datasette.io\/plugins\">Datasette plugin directory<\/a>. Datasette has over 50 plugins now and I\u2019ve been wanting a definitive directory of them for a while.<\/p><p>It\u2019s pretty basic at the moment, offering a list of plugins plus simple <code>LIKE<\/code> based search, but I plan to expand it a great deal in the future.<\/p><p>The fun part is where the data comes from. For a couple of years now I\u2019ve been using GitHub topics to tag my plugins\u2014I tag them with <code>datasette-plugin<\/code>, and the ones that I planned to feature on the site when I finally launched it were also tagged with <code>datasette-io<\/code>.<\/p><p>The <code>datasette.io<\/code> deployment process runs a script called <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/353f1f940cb0cb3c38d3bdf6e345328740990702\/build_plugin_directory.py\">build_plugin_directory.py<\/a>, which uses a GraphQL query against the GitHub search API to find all repositories belonging to me that have been tagged with those tags.<\/p><p>That GraphQL query looks like this:<\/p><div class=\"highlight highlight-source-graphql\"><pre>query {\n  search(query:\"topic:datasette-io topic:datasette-plugin user:simonw\" type:REPOSITORY, first:100) {\n    repositoryCount\n    nodes {\n      ... on Repository {\n        id\n        nameWithOwner\n        openGraphImageUrl\n        usesCustomOpenGraphImage\n        repositoryTopics(first:100) {\n          totalCount\n          nodes {\n            topic {\n              name\n            }\n          }\n        }\n        openIssueCount: issues(states:[OPEN]) {\n          totalCount\n        }\n        closedIssueCount: issues(states:[CLOSED]) {\n          totalCount\n        }\n        releases(last: 1) {\n          totalCount\n          nodes {\n            tagName\n          }\n        }\n      }\n    }\n  }\n}<\/pre><\/div><p>It fetches the name of each repository, the <code>openGraphImageUrl<\/code> (which doesn\u2019t appear to be included in the regular GitHub REST API), the number of open and closed issues and details of the most recent release.<\/p><p>The script has access to a copy of the current site database, which is downloaded on each deploy by the build script. It uses this to check if any of the repositories have new releases that haven\u2019t previously been seen by the script.<\/p><p>Then it runs the <code>github-to-sqlite releases<\/code> command (part of <a href=\"https:\/\/github.com\/dogsheep\/github-to-sqlite\">github-to-sqlite<\/a>) to fetch details of those new releases.<\/p><p>The end result is a database of repositories and releases for all of my tagged plugins. The plugin directory is then built against a <a href=\"https:\/\/datasette.io\/content\/plugins\">custom SQL view<\/a>.<\/p><h4>Other site content<\/h4><p>The rest of the site content is mainly static template files. I use the <code>render_markdown()<\/code> function inline in some of them so I can author in Markdown rather than HTML\u2014here\u2019s <a href=\"https:\/\/github.com\/simonw\/datasette.io\/blob\/main\/templates\/pages\/examples.html\">the template<\/a> for the <a href=\"https:\/\/datasette.io\/examples\">\/examples page<\/a>. The various <a href=\"https:\/\/datasette.io\/for\">Use cases for Datasette<\/a> pages are likewise built as static templates.<\/p><h4 id=\"sqlite-utils-analyze-tables\">Also this week: sqlite-utils analyze-tables<\/h4><p>My other big project this week has involved building out a Datasette instance for a client. I\u2019m working with over 5,000,000 rows of CSV data for this, which has been a great opportunity to push the limits of some of my tools.<\/p><p>Any time I\u2019m working with new data I like to get a feel for its general shape. Having imported 5,000,000 rows with dozens of columns into a database, what can I learn about the columns beyond just browsing them in Datasette?<\/p><p><code>sqlite-utils analyze-tables<\/code> (<a href=\"https:\/\/sqlite-utils.readthedocs.io\/en\/stable\/cli.html#analyzing-tables\">documented here<\/a>) is my new tool for doing just that. It loops through every table and every column in the database, and for each column it calculates statistics that include:<\/p><ul><li>The total number of distinct values<\/li>\n<li>The total number of null or blank values<\/li>\n<li>For non-distinct columns, the 10 most common and 10 least common values<\/li>\n<\/ul><p>It can output those to the terminal, or if you add the <code>--save<\/code> option it will also save them to a SQLite table called <code>_analyze_tables_<\/code>\u2014here\u2019s <a href=\"https:\/\/github-to-sqlite.dogsheep.net\/github\/_analyze_tables_\">that table<\/a> for my github-to-sqlite demo instance.<\/p><p>I can then use the output of the tool to figure out which columns might be a primary key, or which ones warrant being extracted out into a separate lookup table using <a href=\"https:\/\/simonwillison.net\/2020\/Sep\/23\/sqlite-utils-extract\/\">sqlite-utils extract<\/a>.<\/p><p>I expect I\u2019ll be expanding this feature a lot in the future, but I\u2019m already finding it to be really helpful.<\/p><h4>Datasette 0.53<\/h4><p>I pushed out a small feature release of Datasette to accompany the new project website. Quoting <a href=\"https:\/\/docs.datasette.io\/en\/stable\/changelog.html#v0-53\">the release notes<\/a>:<\/p><blockquote>\n<ul><li>New <code>?column__arraynotcontains=<\/code> table filter. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1132\">#1132<\/a>)<\/li>\n<li><code>datasette serve<\/code> has a new <code>--create<\/code> option, which will create blank database files if they do not already exist rather than exiting with an error. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1135\">#1135<\/a>)<\/li>\n<li>New <code>?_header=off<\/code> option for CSV export which omits the CSV header row, <a href=\"https:\/\/docs.datasette.io\/en\/stable\/csv_export.html#csv-export-url-parameters\">documented here<\/a>. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1133\">#1133<\/a>)<\/li>\n<li>\u201cPowered by Datasette\u201d link in the footer now links to <a href=\"https:\/\/datasette.io\/\">https:\/\/datasette.io\/<\/a>. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1138\">#1138<\/a>)<\/li>\n<li>Project news no longer lives in the README\u2014it can now be found at <a href=\"https:\/\/datasette.io\/news\">https:\/\/datasette.io\/news<\/a>. (<a href=\"https:\/\/github.com\/simonw\/datasette\/issues\/1137\">#1137<\/a>)<\/li>\n<\/ul><\/blockquote><h4>Office hours<\/h4><p>I had my first round of <a href=\"https:\/\/calendly.com\/swillison\/datasette-office-hours\">Datasette office hours<\/a> on Friday\u201420 minute video chats with anyone who wants to talk to me about the project. I had five great conversations\u2014it\u2019s hard to overstate how thrilling it is to talk to people who are using Datasette to solve problems. If you\u2019re an open source maintainer I can thoroughly recommend giving this format a try.<\/p><h4>Releases this week<\/h4><ul><li><strong><a href=\"https:\/\/github.com\/simonw\/datasette-publish-fly\">datasette-publish-fly<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/datasette-publish-fly\/releases\/tag\/1.0.1\">1.0.1<\/a>\u20142020-12-12<br \/>Datasette plugin for publishing data using Fly<\/li>\n<li><strong><a href=\"https:\/\/github.com\/simonw\/datasette-auth-passwords\">datasette-auth-passwords<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/datasette-auth-passwords\/releases\/tag\/0.3.3\">0.3.3<\/a>\u20142020-12-11<br \/>Datasette plugin for authentication using passwords<\/li>\n<li><strong><a href=\"https:\/\/github.com\/simonw\/datasette\">datasette<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/datasette\/releases\/tag\/0.53\">0.53<\/a>\u2014\u20142020-12-11<br \/>An open source multi-tool for exploring and publishing data<\/li>\n<li><strong><a href=\"https:\/\/github.com\/simonw\/datasette-column-inspect\">datasette-column-inspect<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/datasette-column-inspect\/releases\/tag\/0.2a\">0.2a<\/a>\u20142020-12-09<br \/>Experimental plugin that adds a column inspector<\/li>\n<li><strong><a href=\"https:\/\/github.com\/simonw\/datasette-pretty-json\">datasette-pretty-json<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/datasette-pretty-json\/releases\/tag\/0.2.1\">0.2.1<\/a>\u20142020-12-09<br \/>Datasette plugin that pretty-prints any column values that are valid JSON objects or arrays<\/li>\n<li><strong><a href=\"https:\/\/github.com\/simonw\/yaml-to-sqlite\">yaml-to-sqlite<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/yaml-to-sqlite\/releases\/tag\/0.3.1\">0.3.1<\/a>\u20142020-12-07<br \/>Utility for converting YAML files to SQLite<\/li>\n<li><strong><a href=\"https:\/\/github.com\/simonw\/datasette-seaborn\">datasette-seaborn<\/a><\/strong>: <a href=\"https:\/\/github.com\/simonw\/datasette-seaborn\/releases\/tag\/0.2a0\">0.2a0<\/a>\u20142020-12-07<br \/>Statistical visualizations for Datasette using Seaborn<\/li>\n<\/ul><h4>TIL this week<\/h4><p>Posted <a href=\"https:\/\/simonwillison.net\/2020\/Dec\/13\/\">13th December 2020<\/a> at 8:34 am \u00b7 Follow <a href=\"https:\/\/twitter.com\/simonw\">@simonw<\/a> on Twitter<\/p>","created_at":"2020-12-24T20:46:43+0000","updated_at":"2020-12-24T20:46:43+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"simonwillison.net","preview_picture":"https:\/\/static.simonwillison.net\/static\/2020\/datasette-io.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1224,"title":"HTML Over The Wire | Hotwire","url":"https:\/\/hotwire.dev\/","content":"<section class=\"home-block --leader\"><p><strong class=\"text --size-l\">Hotwire is an alternative approach to building modern web applications without using much JavaScript by sending HTML instead of JSON over the wire.<\/strong> This makes for fast first-load pages, keeps template rendering on the server, and allows for a simpler, more productive development experience in any programming language, without sacrificing any of the speed or responsiveness associated with a traditional single-page application.<\/p><p><\/p><p>You can follow the HTML-over-the-wire approach as merely inspiration for your own tooling, or you can use these Hotwire designed frameworks created by the team at <a href=\"https:\/\/basecamp.com\" target=\"_blank\" rel=\"noopener\">Basecamp<\/a> to power <a href=\"https:\/\/hey.com\" target=\"_blank\" rel=\"noopener\">HEY<\/a>. They work together, complementing each other to give you a complete solution across all platforms:<\/p><\/section><section class=\"home-block --connector animate --delay-s\" data-aos=\"--fade-in-pop\"><p class=\"space --top-flush\">The heart of Hotwire is Turbo. A set of complimentary techniques for speeding up page changes and form submissions, dividing complex pages into components, and stream partial page updates over WebSocket. All without writing any JavaScript at all. And designed from the start to integrate perfectly with native hybrid applications for iOS and Android.<\/p><p><a href=\"https:\/\/turbo.hotwire.dev\" rel=\"noopener\" class=\"button\">Learn more about Turbo<\/a><\/p><\/section><section class=\"home-block --connector --connector-2 animate --delay-s\" data-aos=\"--fade-in-pop\"><p class=\"space --top-flush\">While Turbo usually takes care of at least 80% of the interactivity that traditionally would have required JavaScript, there are still cases where a dash of custom code is required. Stimulus makes this easy with a HTML-centric approach to state and wiring.<\/p><p><a href=\"https:\/\/stimulus.hotwire.dev\" rel=\"noopener\" class=\"button\">Learn more about Stimulus<\/a><\/p><\/section><section class=\"home-block --connector --connector-3 animate --delay-s\" data-aos=\"--fade-in-pop\"><p class=\"space --top-flush\">Standardizes the way that web and native parts of a mobile hybrid application talk to each other via HTML bridge attributes. This makes it easy to progressively level-up web interactions with native replacements.<\/p><p><em>Strada will premiere in 2021<\/em><\/p><\/section>","created_at":"2020-12-22T18:51:29+0000","updated_at":"2020-12-22T18:51:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"hotwire.dev","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1223,"title":"eschudt\/hashistack-digitalocean","url":"https:\/\/github.com\/eschudt\/hashistack-digitalocean","content":"<p>Terraform to setup a consul and nomad cluster by building the number of servers and clients specified. It wraps them around a firewall that can only be accesses via a bastion host (ssh) and a load balancer (http) Uses digital ocean as a provider to create the droplets needed. Sets up Vault to store secrets<\/p><p>Now also uses Consul Connect as a sidecar allocation to interact with other services<\/p><p>It starts nomad and consul as a service and automatically connects all nodes in the cluster<\/p><h2><a id=\"user-content-environment-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#environment-variables\"><\/a>Environment variables<\/h2><ul><li><code>do_token<\/code> - api token for digital ocean which can be found in your DigitalOcean Account under \"API\"<\/li>\n<li><code>ssh_fingerprint<\/code> - the ssh fingerprint to use to connect to your newly created droplets<\/li>\n<li><code>bastion_host_id<\/code> - the droplet id of the bastion host server<\/li>\n<li><code>server_count<\/code> - number of server droplets to create<\/li>\n<li><code>client_count<\/code> - number of client droplets to create<\/li>\n<\/ul><h2><a id=\"user-content-modules\" class=\"anchor\" aria-hidden=\"true\" href=\"#modules\"><\/a>Modules<\/h2><h3><a id=\"user-content-server-droplet\" class=\"anchor\" aria-hidden=\"true\" href=\"#server-droplet\"><\/a>server-droplet<\/h3><ul><li>Create servers and sets up nomad and consul in server mode<\/li>\n<li><code>server_count<\/code> - number of server droplets to create<\/li>\n<\/ul><h3><a id=\"user-content-client-droplet\" class=\"anchor\" aria-hidden=\"true\" href=\"#client-droplet\"><\/a>client-droplet<\/h3><ul><li>Create clients and sets up nomad and consul in client mode<\/li>\n<li><code>client_count<\/code> - number of client droplets to create<\/li>\n<li><code>consul_server_ip<\/code> - a consul server ip<\/li>\n<\/ul><h3><a id=\"user-content-load-balancer\" class=\"anchor\" aria-hidden=\"true\" href=\"#load-balancer\"><\/a>load-balancer<\/h3><ul><li>Create a public load balancer to connect to all servers<\/li>\n<li><code>all_server_ids<\/code> - ids of all servers (droplets)<\/li>\n<\/ul><h3><a id=\"user-content-firewall\" class=\"anchor\" aria-hidden=\"true\" href=\"#firewall\"><\/a>firewall<\/h3><ul><li>Create a firewall around the server and client droplets<\/li>\n<li><code>all_server_ids<\/code> - ids of all servers (droplets)<\/li>\n<li><code>load_balancer_id<\/code> - the id of the digital ocean load balancer<\/li>\n<li><code>bastion_id<\/code> - the droplet id of the bastion host<\/li>\n<\/ul><h2><a id=\"user-content-scripts\" class=\"anchor\" aria-hidden=\"true\" href=\"#scripts\"><\/a>Scripts<\/h2><p>Scripts for installing required software in newly created droplets<\/p><h3><a id=\"user-content-consul\" class=\"anchor\" aria-hidden=\"true\" href=\"#consul\"><\/a>consul<\/h3><p><code>install_consul.sh client|server ${self.ipv4_address_private} ${var.consul_server_ip}<\/code><\/p><ul><li>Installs required software - unzip and docker<\/li>\n<li>Sets up iptables to allow access to localhost from docker<\/li>\n<li>Downloads consul and copies the binary to the \/user\/bin directory<\/li>\n<li>Starts consul as a service in either server or client mode<\/li>\n<li>If in client mode, it joins the client to the cluster<\/li>\n<\/ul><h3><a id=\"user-content-nomad\" class=\"anchor\" aria-hidden=\"true\" href=\"#nomad\"><\/a>nomad<\/h3><p><code>install_nomad.sh client|server<\/code><\/p><ul><li>Downloads nomad and copies the binary to the \/user\/bin directory<\/li>\n<li>Starts nomad as a service in either server or client mode<\/li>\n<\/ul><h3><a id=\"user-content-vault\" class=\"anchor\" aria-hidden=\"true\" href=\"#vault\"><\/a>vault<\/h3><p><code>install_vault.sh server<\/code><\/p><ul><li>Downloads vault and copies the binary to the \/user\/bin directory<\/li>\n<li>Starts vault as a service in server mode<\/li>\n<li>Initializes Vault<\/li>\n<li>Unseals vaults to make it ready for use<\/li>\n<li>Exports the vault token for nomad to use<\/li>\n<\/ul><h2><a id=\"user-content-how-to-run\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-run\"><\/a>How to run<\/h2><ul><li><code>eval `ssh-agent -s`<\/code><\/li>\n<li><code>ssh-add ~\/.ssh\/id_rsa<\/code> (add your private key to the ssh agent which corresponds to the ssh_fingerprint)<\/li>\n<li><code>terraform init<\/code><\/li>\n<li><code>terraform plan<\/code><\/li>\n<li><code>terraform apply<\/code><\/li>\n<\/ul>","created_at":"2020-12-18T03:01:47+0000","updated_at":"2020-12-18T03:01:47+0000","published_by":["eschudt"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/11932941?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["procgen","generative","creative-coding"],"is_public":false,"id":1222,"title":"Creative Code-Generated Art | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=25453252","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\" style=\"border-spacing: 0px; width: 100%;\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D25453252\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"25453252\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"https:\/\/www.editorx.com\/shaping-design\/article\/creative-coding\" class=\"storylink\">Creative Code-Generated Art<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=editorx.com\">editorx.com<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">73 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=dzink\" class=\"hnuser\">dzink<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=25453252\">6 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=25453252&amp;goto=item%3Fid%3D25453252\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Creative%20Code-Generated%20Art&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=25453252&amp;auth=2442ecc80f2d54e8bc4a4f841ffab37b0944f976\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=25453252\">16\u00a0comments<\/a><\/td>\n<\/tr><tr><td>\n<\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2020-12-17T13:16:55+0000","updated_at":"2020-12-17T13:17:39+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1221,"title":"A simple-minded way of managing window configs in Emacs","url":"https:\/\/depp.brause.cc\/eyebrowse\/","content":"<header class=\"projects\"><a href=\"https:\/\/depp.brause.cc\/\">\n<\/a><\/header><h2><a href=\"https:\/\/depp.brause.cc\/eyebrowse\">eyebrowse<\/a>: A simple-minded way of managing window configs in Emacs<\/h2><section class=\"clone\"><h2>Clone<\/h2><code>git clone https:\/\/depp.brause.cc\/eyebrowse.git<\/code><\/section><h2>Files<\/h2><table class=\"files striped\"><thead><tr><th class=\"size\">Size<\/th>\n<th class=\"path\">Path<\/th>\n<\/tr><\/thead><tbody><tr class=\"directory\"><td class=\"path\"><a href=\"https:\/\/depp.brause.cc\/eyebrowse\/img\">img<\/a>\/<\/td>\n<\/tr><tr class=\"file\"><td class=\"size\">35148<\/td>\n<td class=\"path\"><a href=\"https:\/\/depp.brause.cc\/eyebrowse\/LICENSE\">LICENSE<\/a><\/td>\n<\/tr><tr class=\"file\"><td class=\"size\">6072<\/td>\n<td class=\"path\"><a href=\"https:\/\/depp.brause.cc\/eyebrowse\/README.md\">README.md<\/a><\/td>\n<\/tr><tr class=\"file\"><td class=\"size\">1756<\/td>\n<td class=\"path\"><a href=\"https:\/\/depp.brause.cc\/eyebrowse\/TODO.org\">TODO.org<\/a><\/td>\n<\/tr><tr class=\"file\"><td class=\"size\">6223<\/td>\n<td class=\"path\"><a href=\"https:\/\/depp.brause.cc\/eyebrowse\/evil.org\">evil.org<\/a><\/td>\n<\/tr><tr class=\"file\"><td class=\"size\">27018<\/td>\n<td class=\"path\"><a href=\"https:\/\/depp.brause.cc\/eyebrowse\/eyebrowse.el\">eyebrowse.el<\/a><\/td>\n<\/tr><\/tbody><\/table><section class=\"readme\"><h2>README.md<\/h2><p><img src=\"https:\/\/depp.brause.cc\/eyebrowse\/img\/eyebrows.gif\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><h2>About<\/h2><p><code>eyebrowse<\/code> is a global minor mode for Emacs that allows you to manage your window configurations in a simple manner, just like tiling window managers like i3wm with their workspaces do. It displays their current state in the modeline by default. The behaviour is modeled after <a href=\"https:\/\/ranger.github.io\/\">ranger<\/a>, a file manager written in Python.<\/p><h2>Screenshot<\/h2><p><img src=\"https:\/\/depp.brause.cc\/eyebrowse\/img\/scrot.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><p>See the lighter and the modeline indicator at the right side of the bottom modeline? That's what you get to see after enabling eyebrowse.<\/p><h2>Installation<\/h2><p>Install via <code>package.el<\/code> from the <a href=\"http:\/\/melpa.org\/\">MELPA (stable)<\/a> repository by setting them up if you haven't already and executing <code>M-x package-install RET eyebrowse RET<\/code>.<\/p><h2>Quick Tutorial<\/h2><p>Use <code>M-x eyebrowse-mode<\/code> to enable <code>eyebrowse<\/code> interactively. If you want to enable it automatically on startup, add <code>(eyebrowse-mode t)<\/code> to your init file (either <code>~\/.emacs<\/code> or <code>~\/.emacs.d\/init.el<\/code>).<\/p><p>You start with your current window config on slot 1. Once you hit <code>C-c C-w 2<\/code>, you will see the modeline indicator appearing and showing slot 1 and 2 with slot 2 slightly emphasized. Slot 1 has been saved automatically for you and contains your last window config. Do something meaningful like a window split, then hit <code>C-c C-w 1<\/code>. The window config on slot 2 is saved and the window config from slot 1 is loaded. Try switching back and forth between them with <code>C-c C-w '<\/code> to get a feeling for how subsequent window manipulations are handled.<\/p><p>To make keeping track of workspaces easier, a tagging feature was added. Use <code>C-c C-w ,<\/code> to set a tag for the current window config, it will both appear in the modeline indicator and when using <code>M-x eyebrowse-switch-to-window-config<\/code>. Setting the tag to an empty value will undo this change.<\/p><h2>Key bindings<\/h2><p>The default key bindings are:<\/p><table class=\"striped\"><thead><tr><th>Key bind<\/th>\n<th>Function<\/th>\n<\/tr><\/thead><tbody><tr><td>C-c C-w &lt;<\/td>\n<td>Switch to previous window config<\/td>\n<\/tr><tr><td>C-c C-w &gt;<\/td>\n<td>Switch to next window config<\/td>\n<\/tr><tr><td>C-c C-w '<\/td>\n<td>Switch to last window config<\/td>\n<\/tr><tr><td>C-c C-w \"<\/td>\n<td>Close current window config<\/td>\n<\/tr><tr><td>C-c C-w ,<\/td>\n<td>Rename current window config<\/td>\n<\/tr><tr><td>C-c C-w 0<\/td>\n<td>Switch to window config 0<\/td>\n<\/tr><tr><td>...<\/td>\n<td>...<\/td>\n<\/tr><tr><td>C-c C-w 9<\/td>\n<td>Switch to window config 9<\/td>\n<\/tr><\/tbody><\/table><h2>Further Customization<\/h2><p>Use <code>M-x customize-group RET eyebrowse<\/code> for a list of customizable options. The more interesting ones would be <code>eyebrowse-wrap-around<\/code> and <code>eyebrowse-switch-back-and-forth<\/code> which affect both wrap around and lazier switching. It is also possible to change the behaviour of creation of new workspaces by customizing <code>eyebrowse-new-workspace<\/code>. By default the last one is simply cloned, setting it to <code>t<\/code> will start out with as empty of a slate as possible (by just displaying a single window with the scratch buffer in it).<\/p><p>The prefix for each binding defaults to <code>C-c C-w<\/code>, but you can change it to something else by customizing <code>eyebrowse-keymap-prefix<\/code>. If you want to change it in your init file, insert the customization before loading <code>eyebrowse.el<\/code>. This includes enabling it with an auto-loaded function, such as <code>eyebrowse-setup-evil-keys<\/code>, <code>eyebrowse-setup-opinionated-keys<\/code> or <code>eyebrowse-mode<\/code>.<\/p><p>If you're not happy with the default keybindings, a riskier set can be enabled additionally either by executing <code>M-: (eyebrowse-setup-opinionated-keys)<\/code> interactively or inserting <code>(eyebrowse-setup-opinionated-keys)<\/code> in your init file. If the function detects the <a href=\"https:\/\/bitbucket.org\/lyro\/evil\/wiki\/Home\">evil<\/a> package, it will enable extra key bindings for it as well.<\/p><p>The extra key bindings are:<\/p><table class=\"striped\"><thead><tr><th>Key bind<\/th>\n<th>Function<\/th>\n<\/tr><\/thead><tbody><tr><td>C-&lt;, gT<\/td>\n<td>Switch to previous window config<\/td>\n<\/tr><tr><td>C-&gt;, gt<\/td>\n<td>Switch to next window config<\/td>\n<\/tr><tr><td>C-', zx<\/td>\n<td>Switch to last window config<\/td>\n<\/tr><tr><td>C-\", gc<\/td>\n<td>Close current window config<\/td>\n<\/tr><tr><td>M-0<\/td>\n<td>Switch to window config 0<\/td>\n<\/tr><tr><td>...<\/td>\n<td>...<\/td>\n<\/tr><tr><td>M-9<\/td>\n<td>Switch to window config 9<\/td>\n<\/tr><\/tbody><\/table><h2>Internals<\/h2><p>This mode basically wraps what <code>C-x r w<\/code> and <code>C-x r j<\/code> would do, but takes care of automatically saving and loading to a separate data structure for you and does it in a slightly different manner (see <code>window-state-put<\/code> and <code>window-state-get<\/code> for more details) to allow for features like persistency in combination with <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Saving-Emacs-Sessions.html#Saving-Emacs-Sessions\">desktop.el<\/a>.<\/p><h2>Notes<\/h2><p>The <code>window-state-put<\/code> and <code>window-state-get<\/code> functions do not save all window parameters. If you use features like side windows that store the window parameters <code>window-side<\/code> and <code>window-slot<\/code>, you will need to customize <code>window-persistent-parameters<\/code> for them to be saved as well:<\/p><pre>(add-to-list 'window-persistent-parameters '(window-side . writable))\n(add-to-list 'window-persistent-parameters '(window-slot . writable))\n<\/pre><p>See <a href=\"https:\/\/github.com\/wasamasa\/eyebrowse\/issues\/52\">#52<\/a> for further discussion.<\/p><h2>Alternatives<\/h2><p>The two most popular window configuration packages are <a href=\"https:\/\/github.com\/shosti\/elscreen\">elscreen<\/a> and <a href=\"https:\/\/github.com\/emacsattic\/escreen\">escreen<\/a>. Both are fairly old and have their share of bugs. The closest package I've found so far to eyebrowse with workspace-specific buffers would be <a href=\"https:\/\/github.com\/nex3\/perspective-el\">perspective<\/a>. <a href=\"https:\/\/github.com\/ilohmar\/wconf\">wconf<\/a> is a minimal alternative with half the lines of code (and features). To have fancy features such as morphing, try <a href=\"https:\/\/github.com\/tlh\/workgroups.el\">workgroups<\/a> or <a href=\"https:\/\/github.com\/pashinin\/workgroups2\">workgroups2<\/a>.<\/p><h2>Name<\/h2><p>Actually, I wanted to name this mode \"eyebrows\" for no real reason, but then a silly typo happened. The typo stuck. So did the new name.<\/p><\/section>","created_at":"2020-12-17T04:19:44+0000","updated_at":"2020-12-17T04:19:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"depp.brause.cc","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":1220,"title":"More batteries included with emacs","url":"https:\/\/karthinks.com\/software\/more-batteries-included-with-emacs\/","content":"<p>Continuing from <a href=\"https:\/\/karthinks.com\/software\/batteries-included-with-emacs\/\">last time<\/a>, here are a dozen more tricks Emacs has up its sleeve that it\u2019s shy about telling you. We continue to chip away at Emacs\u2019 discoverability problem, one demo at a time.<\/p><p>Same rules as before:<\/p>\n<ul><li>\n<p>No packages, <strong>stock Emacs only<\/strong><\/p>\n<\/li>\n<li>\n<p>No steep learning curves. <strong>Learn each feature in under five minutes or bust<\/strong>. (Up from two minutes last time.)<\/p>\n<\/li>\n<li>\n<p><strong>No gimmicks<\/strong>. No doctor, tetris, snake, dunnet\u2026<\/p>\n<\/li>\n<li>\n<p><strong>Just the deltas<\/strong>. No commonly mentioned packages like flymake, doc-view, outline-minor-mode, gnus or eww. Nothing that Emacs brings up automatically or a nonspecific Google search gets you.<\/p>\n<\/li>\n<li>\n<p>Assume a modern Emacs, 26.3+.<\/p>\n<p>Also, if you\u2019re new to Emacs:<\/p>\n<table><thead><tr><th>Emacs jargon<\/th>\n<th>Modern parlance<\/th>\n<\/tr><\/thead><tbody><tr><td><code>M-x<\/code><\/td>\n<td>Alt + x<\/td>\n<\/tr><tr><td><code>C-x<\/code><\/td>\n<td>Ctrl + x<\/td>\n<\/tr><tr><td>Frame<\/td>\n<td>Emacs window<\/td>\n<\/tr><tr><td>Window<\/td>\n<td>split\/pane<\/td>\n<\/tr><tr><td>Buffer<\/td>\n<td>Contiguous chunk of text\/data<\/td>\n<\/tr><tr><td>Point<\/td>\n<td>Cursor position in buffer<\/td>\n<\/tr><tr><td>Active Region<\/td>\n<td>Text selection<\/td>\n<\/tr><tr><td>Region<\/td>\n<td>Text selection (not highlighted)<\/td>\n<\/tr><tr><td>Face<\/td>\n<td>Font, color and display properties<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<\/ul><p>(Sorry.)<\/p>\n<p>Okay? Let\u2019s go:<\/p>\n<h2 id=\"strokes--m-x-strokes-help\">Strokes (<code>M-x strokes-help<\/code>)<\/h2>\n<p>Control Emacs with mouse gestures:<\/p>\n Your browser does not support the video tag.<p>Why are you using Emacs with a mouse, you ask.<\/p>\n<p>In this demo I used gestures to handle window management: <img src=\"https:\/\/karthinks.com\/img\/strokes-list.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>This is because I lack imagination. Consider: You can bind <em>any<\/em> gesture to <em>any<\/em> Emacs command.<\/p>\n<p>The real utility of this feature is when you find yourself going back and forth between different interaction paradigms, like a web browser and Emacs, requiring frequent context switching between full keyboard and keyboard\/mouse based interaction.<\/p>\n<p>Want to <code>org-capture<\/code> or <code>org-roam-insert<\/code> something from your browser? Want to quickly recompile a TeX document while you\u2019re busy mousing through the corresponding PDF? Need to tweak some markup in your document while you work on graphics in Inkscape? <a href=\"http:\/\/xenodium.com\/emacs-clone-git-repo-from-clipboard\/\">Clone a repository you found online<\/a>? <code>strokes-mode<\/code>.<\/p>\n<p>The second use case for Strokes is to control Emacs during reading-centric activities, like email or RSS. Star or delete messages, pause or skip music tracks (with some elisp around <code>playerctl<\/code>), navigate info nodes, or cycle through Elfeed searches.<\/p>\n<p><code>M-x strokes-help<\/code> will get you started. As with most things Emacs, there\u2019s more to the library than meets the eye, like support for multi-part strokes you can use to edit documents in Chinese. The buffer showing the gesture being traced is for demo purposes only, you can customize <code>strokes-use-strokes-buffer<\/code>:<\/p>\n<div class=\"highlight\">\n<pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(global-set-key (kbd \"&lt;down-mouse-3&gt;\") 'strokes-do-stroke) ; Draw strokes with RMB\n(setq strokes-use-strokes-buffer nil) ; Don't draw strokes to the screen\n<\/pre><\/div>\n<hr \/><h2 id=\"minibuffer-completion-styles--c-h-v-completion-styles\">Minibuffer completion styles (<code>C-h v completion-styles<\/code>)<\/h2>\n<p>The default minibuffer experience is annoying enough to send anyone into the arms of <code>ido<\/code>, <code>ivy<\/code> or <code>helm<\/code>. But Emacs does a lot better in the tab-completions department than it lets on. The minibuffer can match by substrings, regexps, initials and even (as of Emacs 27+) fuzzily, and it can do all of them at once if you don\u2019t mind a giant pile of matches.<\/p>\n<p>You can do worse than<\/p>\n<div class=\"highlight\">\n<pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(setq completion-styles '(initials partial-completion flex)) ; &gt; Emacs 27.1\n(setq completion-cycle-threshold 10)\n<\/pre><\/div>\n<p>Now <code>M-x ohba<\/code> tab-expands to <code>org-hide-block-all<\/code>, <code>M-x qrr<\/code> to <code>query-replace-regexp<\/code> and so on, allowing you to tab-cycle between up to ten completions. Use tab to match file names fuzzily and expand short paths (<code>~\/.l\/sh\/g<\/code>) to long ones (<code>~\/.local\/share\/git<\/code>).<\/p>\n<p>Stretching the five minute rule a bit: You probably want different matching rules for different categories, though. Flex matching on extended commands (<code>M-x<\/code>) can dump hundreds of irrelevant matches on short input, for instance. You can customize <code>completion-category-overrides<\/code> to set the matching style by the category being completed.<\/p>\n<p>Finally, note that completion styles dictate how a candidate pool is found. Incremental completion systems like <code>ido<\/code> specify how found candidates are displayed and chosen. These are orthogonal functions, so any set of completion styles should work with any completion system. Unfortunately this isn\u2019t the case, because <code>ivy<\/code> and <code>helm<\/code> are off doing their own thing. Emacs\u2019 built in completion styles do work with <code>icomplete<\/code>, <code>ido<\/code> and possibly <code>selectrum<\/code> though.<\/p>\n<hr \/><h2 id=\"fake-ido--m-x-fido-mode\">Fake Ido ( <code>M-x fido-mode<\/code>)<\/h2>\n<p>Improves the other half of the default minibuffer experience: interacting with selection candidates. Navigating to, from and inside the default completions buffer takes too many key presses. <code>fido-mode<\/code> brings <code>ido<\/code> like selection to every command in Emacs that uses <code>completing-read<\/code>, which is most of them.<\/p>\n<p>Of course, <code>ido<\/code> is also built in, but you know all about it. It also takes some effort to get ido to work with every command. <code>fido-mode<\/code> is set-and-forget, but it\u2019s Emacs 27+ only.<\/p>\n<hr \/><h2 id=\"easier-rectangle-editing--m-x-cua-selection-mode\">Easier rectangle editing (<code>M-x cua-selection-mode<\/code>)<\/h2>\n Your browser does not support the video tag.<p>The default rectangle commands in Emacs (<code>C-x r<\/code> map) leave a little to be desired in terms of interactivity. Emacs has fully featured rectangle editing, but it\u2019s presented in an odd sort of way, as a subfeature of Common User Access. It\u2019s pretty nifty:<\/p>\n Your browser does not support the video tag.<p>Start or clear a rectangle selection with <code>C-&lt;return&gt;<\/code>. Cycle through rectangle corners with <code>&lt;return&gt;<\/code>. Yes, you\u2019ll need to rebind it if you use Org or ESS.<\/p>\n<p>It\u2019s a very handy feature, but note that <code>cua-selection-mode<\/code> does <em>not<\/em> mesh well with undo. If you use <code>undo-tree<\/code>, it can lock up trying to undo a <code>cua-selection<\/code> edit.<\/p>\n<hr \/><h2 id=\"speedbar--m-x-speedbar\">speedbar (<code>M-x speedbar<\/code>)<\/h2>\n<p>A file explorer and much, much more in Emacs. If you need the Neotree or dired-sidebar package to do anything in Emacs, check out speedbar.<\/p>\n<p>Speedbar pops up a side frame that looks like a file browser, which it is. But it also integrates with <code>imenu<\/code> to show you headings\/tags inside files, and with <code>vc<\/code> to show you the status of your files:<\/p>\n<figure><img src=\"https:\/\/karthinks.com\/img\/speedbar-tags-demo.png\" alt=\"Figure 1: Speedbar showing tags and VC status\" referrerpolicy=\"no-referrer\" \/><figcaption>Figure 1: Speedbar showing tags and VC status\n<\/figcaption><\/figure><figure><img src=\"https:\/\/karthinks.com\/img\/speedbar-org-demo.png\" alt=\"Figure 2: Speedbar showing headings in an org file\" referrerpolicy=\"no-referrer\" \/><figcaption>Figure 2: Speedbar showing headings in an org file\n<\/figcaption><\/figure><p>From inside speedbar you can act on files with (frustratingly, <em>almost<\/em> dired-like) keybindings. Most regular commands run from speedbar will be run in the associated buffer.<\/p>\n<p>It tracks the active buffer to show you appropriate hierarchical information. Here I moved point into <code>info<\/code> and <code>latex-mode<\/code> buffers:<\/p>\n<p><img src=\"https:\/\/karthinks.com\/img\/speedbar-info-demo.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><img src=\"https:\/\/karthinks.com\/img\/speedbar-latex-demo.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Perhaps you\u2019d like a list of your buffers instead. Switch between buffer and file view with <code>b<\/code> \/ <code>f<\/code>. (You can expand a buffer entry for tag\/headings.)<\/p>\n<figure><img src=\"https:\/\/karthinks.com\/img\/speedbar-buffers-demo.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><hr \/><h2 id=\"orgtbl-minor-mode--m-x-orgtbl-mode\">Orgtbl minor mode (<code>M-x orgtbl-mode<\/code>)<\/h2>\n<p>Org-mode is its own thing, a full fledged notetaking, publishing, literate-programming, task tracking application with a growing ecosystem that\u2019s parallel to Emacs\u2019.<sup id=\"fnref:1\"><a href=\"#fn:1\" class=\"footnote-ref\" role=\"doc-noteref\">1<\/a><\/sup><\/p>\n<p>But it does provide a handy tool that\u2019s useful in any mode: Making tables. If you\u2019re one of the dozen Emacs users who haven\u2019t hopped on the Org train, this one\u2019s for you.<\/p>\n<p>With <code>orgtbl-mode<\/code>, you can make tables by starting a line with the pipe character and hitting tab to make a field\u2013like you would in <code>org-mode<\/code>\u2013but <strong>without getting in the way of your major mode<\/strong>. You can just leave it on and get on with your work.<\/p>\n Your browser does not support the video tag.<p>Here I table some parameters in matlab-mode as a comment, but even more lazily than prescribed above by calling <code>orgtbl-create-or-convert-from-region<\/code> (<code>C-c |<\/code>) on some data.<\/p>\n<p>With some work, you can export the table in place to html or LaTeX, but setting this up would break the five minute rule. You can also use the table as a spreadsheet, which is beyond me.<\/p>\n<hr \/><h2 id=\"regexp-builder--m-x-re-builder\">Regexp builder (<code>M-x re-builder<\/code>)<\/h2>\n<p><img src=\"https:\/\/karthinks.com\/img\/re-builder-demo.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/> Emacs\u2019 regular expressions syntax can be idiosyncratic if you\u2019re used to <a href=\"https:\/\/www.pcre.org\/\">PCRE<\/a>. Are unescaped parens matched literally or do they act as capture groups? Does + need to be escaped to match multiple occurrences of a character? How do I match whitespace again?<sup id=\"fnref:2\"><a href=\"#fn:2\" class=\"footnote-ref\" role=\"doc-noteref\">2<\/a><\/sup><\/p>\n<p>I can never remember. For regularly confused users like me there is <code>re-builder<\/code>. Build your regular expressions interactively, one character at a time. Save it to the kill ring for later use. This is a godsend when testing regexes for non-interactive use, like when writing elisp helpers.<\/p>\n<p>A cleaner approach to regular expressions in Emacs, as most package maintainers will tell you, is to use the <code>rx<\/code> library instead. <code>rx<\/code> translates regular expressions in sexp form to a regexp string:<\/p>\n<div class=\"highlight\">\n<pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(rx (and \"(\"\n         (or \"use-package\" \"require\")\n         (+ space)\n         (group (syntax symbol))))\n<\/pre><\/div>\n<div class=\"highlight\">\n<pre class=\"language-text\" data-lang=\"text\">(\\(?:\\(?:requir\\|use-packag\\)e\\)[[:space:]]+\\(\\s_\\)\n<\/pre><\/div>\n<hr \/><h2 id=\"future-history--m-n-in-prompts\">Future history (<code>M-n<\/code> in prompts)<\/h2>\n<p><code>M-p<\/code> and <code>M-n<\/code> cycle through history items in minibuffer prompts. So what happens when you press <code>M-n<\/code> when you\u2019re at a blank prompt?<\/p>\n<p>Emacs tries to Do What You Mean. If the cursor is over a file name, for instance, <code>M-n<\/code> at the find-file prompt inserts the file path at point into the minibuffer. From the manual:<\/p>\n<blockquote>\n<p>The \u201cfuture history\u201d for file names includes several possible alternatives you may find useful, such as the file name or the URL at point in the current buffer. The defaults put into the \u201cfuture history\u201d in this case are controlled by the functions mentioned in the value of the option file-name-at-point-functions.<\/p>\n<\/blockquote>\n<p>Like the DWIM behavior of Emacs commands on regions, many utilities account for <code>future-history<\/code>. You can usually drag whatever relevant object point is at into the minibuffer prompt with <code>M-n<\/code>.<\/p>\n<hr \/><h2 id=\"transpose-regions--m-x-transpose-region\">Transpose regions (<code>M-x transpose-region<\/code>)<\/h2>\n<p>Lurking in the bevy of transposition commands in Emacs is <code>transpose-region<\/code>, which does what it says on the tin<sup id=\"fnref:3\"><a href=\"#fn:3\" class=\"footnote-ref\" role=\"doc-noteref\">3<\/a><\/sup>:<\/p>\n Your browser does not support the video tag.<p>Exchange any two non-overlapping regions in a buffer. You\u2019ll need to assign a keybinding though.<\/p>\n<div class=\"highlight\">\n<pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(global-set-key (kbd \"C-x C-M-t\") 'transpose-regions)\n<\/pre><\/div>\n<p>Also of note: <code>transpose-paragraphs<\/code>, which sounds like a text-mode utility but works pretty well on adjacent blocks of code.<\/p>\n<hr \/><h2 id=\"view-mode-again--m-x-view-mode\">View mode, again (<code>M-x view-mode<\/code>)<\/h2>\n<p>View-mode got an airing <a href=\"https:\/\/karthinks.com\/software\/batteries-included-with-emacs\/#view-mode--m-x-view-mode\">the last time around<\/a> but I like it so much I will repeat myself. <strong>Turn Emacs into a pager.<\/strong><\/p>\n<p>By default, pressing <code>v<\/code> in dired will open a file in view-mode. You can dismiss the window and buffer with <code>q<\/code>, so regular buffers can essentially function as info or help buffers do with <code>View<\/code>. You can scroll or isearch without pressing any modifier keys, and all advanced navigation commands (like xref) are still available, so this is the perfect way to explore a code repository.<\/p>\n<hr \/><h2 id=\"newsticker--m-x-newsticker-treeview\">Newsticker (<code>M-x newsticker-treeview<\/code>)<\/h2>\n<p>AKA I\u2019m not brave enough to use GNUS! <img src=\"https:\/\/karthinks.com\/img\/newsticker-demo.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Yes, Emacs has a traditionally styled RSS reader built in. You can organize feeds by folders, peruse with one button, the usual. This is the only feature in this list I no longer use, because Elfeed\u2019s design syncs with my brain much better. That said, Newsticker is excellent for following high volume feeds like subreddits since it doesn\u2019t maintain a database:<\/p>\n<blockquote>\n<p>You can mark [an item] as \u201cold\u201d when you have read it or \u2013 if you want to keep it \u2013 you can mark it as \u201cimmortal\u201d. You can do that manually and you can define filters which do that automatically, see below. When a headline has vanished from the feed it is automatically marked as \u201cobsolete\u201d unless it has the status \u201cimmortal\u201d. \u201cObsolete\u201d headlines get removed automatically after a certain time.<\/p>\n<\/blockquote>\n<p>Another advantage over GNUS is that it needs very little configuration<sup id=\"fnref:4\"><a href=\"#fn:4\" class=\"footnote-ref\" role=\"doc-noteref\">4<\/a><\/sup>. I used it for a while and never read the info page. <code>newsticker-add-url<\/code> to add a source (atom\/RSS feed), <code>newsticker-treeview<\/code> to browse.<\/p>\n<hr \/><h2 id=\"semantic-and-semantic-decorators--m-x-semantic-decoration-mode\">Semantic and Semantic decorators (<code>M-x semantic-decoration-mode<\/code>)<\/h2>\n<p>Get Emacs to decorate your python code: <img src=\"https:\/\/karthinks.com\/img\/semantic-decorators-demo.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Doing this is like pointing a firehose at a matchstick: Semantic is <a href=\"http:\/\/alexott.net\/en\/writings\/emacs-devenv\/EmacsCedet.html#sec5\">part of CEDET<\/a>, a large collection of (mostly) built-in tools to try to turn Emacs into a full-fledged IDE. It\u2026 doesn\u2019t work great on a code base of serious size or import, but it\u2019s good at analyzing exploratory code\/small utilities.<\/p>\n<p>CEDET is written as plumbing, and unless you\u2019re writing a package you probably shouldn\u2019t need to do anything beyond turning on\/off minor modes to use it.<\/p>\n<p>Ultimately Semantic is inadequate on account of the XY problem. LSP handles everything it\u2019s supposed to do and does it better, but Semantic is built-in and there if you just want some code-awareness in your Emacs.<\/p>\n<hr \/><p>This was my attempt at covering built-in features of Emacs I use regularly that I don\u2019t see being talked about much. There are many more, like <code>project<\/code>, <code>flymake<\/code> and <code>outline-minor-mode<\/code> that have received a surge of interest in recent times<sup id=\"fnref:5\"><a href=\"#fn:5\" class=\"footnote-ref\" role=\"doc-noteref\">5<\/a><\/sup>. There are also features I know of that aren\u2019t talked about much, like <code>ede-mode<\/code>, <code>2C-two-columns<\/code>, <code>skeleton<\/code> and <code>tempo<\/code>. I don\u2019t bring these up since I don\u2019t have any experience with them.<\/p>\n<p>So this concludes my list.<\/p>\n<p>Something entirely different next time: Typing equations in Emacs faster than you can write on paper.<\/p>\n<section class=\"footnotes\" role=\"doc-endnotes\"><hr \/><ol><li id=\"fn:1\" role=\"doc-endnote\">\n<p>I hear the next major version will make you pancakes. <a href=\"#fnref:1\" class=\"footnote-backref\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn:2\" role=\"doc-endnote\">\n<p>Matched literally, no and \u201c\\s-\u201d respectively. <a href=\"#fnref:2\" class=\"footnote-backref\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn:3\" role=\"doc-endnote\">\n<p>Vim users might be familiar with the popular <code>exchange<\/code> (<code>evil-exchange<\/code>) plugin that exchanges regions of text. <a href=\"#fnref:3\" class=\"footnote-backref\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn:4\" role=\"doc-endnote\">\n<p>Which doesn\u2019t mean it\u2019s not there. Let\u2019s say I resisted the urge to customize Newsticker because I didn\u2019t want to waste time while I was busy wasting time. <a href=\"#fnref:4\" class=\"footnote-backref\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn:5\" role=\"doc-endnote\">\n<p>Interest from developers and from users. Don\u2019t you love that the distinction is fuzzy in Emacs land? <a href=\"#fnref:5\" class=\"footnote-backref\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<\/ol><\/section>","created_at":"2020-12-13T05:28:33+0000","updated_at":"2020-12-13T05:28:40+0000","published_at":"2020-12-11T19:17:00+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"karthinks.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1219,"title":"Chrome is Bad","url":"https:\/\/chromeisbad.com\/","content":"Chrome is Bad\n<p class=\"note\">Short story: Google Chrome installs something called <em>Keystone<\/em> on your computer, which nefariously hides itself from Activity Monitor and <strong>makes your whole computer slow <span class=\"c1\">even when Chrome isn't running<\/span><\/strong>. Deleting Chrome and Keystone makes your computer <em>way, way faster, all the time.<\/em> <a href=\"#delete\">Click here for instructions<\/a>.<\/p>\n<p>Long story: I noticed my brand new 16\" MacBook Pro started acting sluggishly doing even trivial things like scrolling. Activity Monitor showed *nothing* from Google using the CPU, but <em>WindowServer<\/em> was taking ~80%, which is abnormally high (it should use &lt;10% normally).<\/p>\n<p>Doing all the normal things (quitting apps, logging out other users, restarting, zapping PRAM, etc) did nothing, then I remembered I had installed Chrome a while back to test a website.<\/p>\n<p>I deleted Chrome, and noticed Keystone while deleting some of Chrome's other preferences and caches. I deleted everything from Google I could find, restarted the computer, and it was like night-and-day. <em>Everything was instantly and noticeably faster, and WindowServer CPU was well under 10% again.<\/em><\/p>\n<p>Then something else hit me, my family had been complaining about the sluggish performance of a 2015 iMac since practically the day we bought it. I had tried everything I could think of \u2013 it had a Fusion drive and the symptoms were consistent with a failing SSD \u2013 but drive diagnostics always turned up nothing. We even went as far as to completely wipe and set up the computer fresh multiple times.<\/p>\n<p>Then I remembered, <em>installing Chrome was always one of the first things we did when we set up the computer<\/em>. I deleted Chrome, and all the files Keystone had littered on the computer, restarted, and it was so snappy <em>it felt like a brand new computer<\/em>.<\/p>\n<p>Yeah, I realize this sounds like a freakin' infomercial, but it worked so well I spent $5 on a domain name and set up this website even if it makes me sound like a raving nut.<\/p>\n<a id=\"delete\">\n<h2>OK that's weird, how do you delete Chrome and Keystone?<\/h2>\n<ol><li>Go to your <code>\/Applications<\/code> folder and drag <code>Chrome<\/code> to the Trash.<\/li>\n<li>In the <code>Finder<\/code> click the <code>Go<\/code> menu (at the top of the screen), then click \"<code>Go to Folder...<\/code>\".<\/li>\n<li>Type in <code>\/Library<\/code> and hit enter.\n<ul><li>Check the following folders: <code>LaunchAgents<\/code>, <code>Application Support<\/code>, <code>Caches<\/code>, <code>Preferences<\/code>.<\/li>\n<li>Delete all the <code>Google<\/code> folders, and anything else that starts with <code>com.google...<\/code> and <code>com.google.keystone...<\/code><\/li>\n<\/ul><\/li>\n<li>Go to \"<code>Go to Folder...<\/code>\" again.<\/li>\n<li>Type in <code>~\/Library<\/code> and hit enter. (Note the \"~\")\n<ul><li>Check the following folders: <code>LaunchAgents<\/code>, <code>Application Support<\/code>, <code>Caches<\/code>, <code>Preferences<\/code>.<\/li>\n<li>Delete all the <code>Google<\/code> folders, and anything else that starts with <code>com.google...<\/code> and <code>com.google.keystone...<\/code><\/li>\n<\/ul><\/li>\n<li>Empty the Trash, and restart your computer.<\/li>\n<\/ol><h2>Now what browser should I use?<\/h2>\n<\/a><a href=\"https:\/\/www.apple.com\/safari\/\">Safari<\/a>\n<p>is good and it's already on your Mac. It's fast and efficient. If you need a Chromium-based browser, use<\/p>\n<a href=\"https:\/\/brave.com\">Brave<\/a>\n<p>or<\/p>\n<a href=\"https:\/\/vivaldi.com\">Vivaldi<\/a>\n<p>. Firefox has pretty noticeable pointer input latency which (I, the author) am pretty nitpicky about, but other than that it's fine. (Mozilla <em>are<\/em> a bunch of short-sighted dopes for firing the Servo team. If the Servo team regroups, I'd be inclined to recommend anything they make down the road).<\/p>\n<h2>What's the deal with Keystone anyway?<\/h2>\n<p>Wired first <a target=\"_blank\" href=\"https:\/\/www.wired.com\/2009\/02\/why-googles-sof\/\">reported on Keystone in 2009<\/a>, when Google put it into Google Earth. It has a long history of crashing Macs by doing bizarre things that shouldn't be necessary for auto-update software to function.<\/p>\n<p>The fact that it hasn't been \"fixed\" in 11 years might mean that it's not actually broken. Why would auto-update software need to take up a massive portion of CPU on a ton's of people's computers, all while hiding itself?<\/p>\n<p>To all the good people at Google who work on Chrome: something is going on between the code you're writing and what is happening on people's computers. I hope you can track it down and give us an honest postmortem.<\/p>","created_at":"2020-12-13T00:02:31+0000","updated_at":"2020-12-13T00:02:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"chromeisbad.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojurescript","generative","quil","p5js"],"is_public":false,"id":1218,"title":"Functional Generative Art Using ClojureScript","url":"https:\/\/functional.christmas\/2019\/13\/","content":"<p>Creative programming is about writing code to create an expression rather than developing some kind of functionality. It's an enjoyable process that can be used to create all kinds of art pieces, be it still images, animations, music etc. It's also a great framework for learning new programming languages.<\/p><p>ClojureScript, or cljs for short, is a Lisp that compiles to javascript. It's a dynamic, functional programming language with great support for an interactive development process, making it a perfect fit for the type of explorative, creative programming that generative art often resemble.<\/p><p>We will create an art piece based on a Perlin noise flow field. It uses <a href=\"https:\/\/en.wikipedia.org\/wiki\/Perlin_noise\">Perlin noise<\/a> to move some thousand particles across the screen in an organic looking fashion. As opposed to white noise, Perlin noise is a controlled type of noise where two values drawn from close to each other in the noise map will have similar, but still random values.<\/p><p>We will be using a library called <a href=\"http:\/\/quil.info\/\">Quil<\/a> to create a drawing using the browser canvas api, along with a library called <a href=\"https:\/\/figwheel.org\/\">Figwheel<\/a> for building and hot loading the code as it evolves.<\/p><p>A file <code class=\"language-text\">deps.edn<\/code> in the project root defines the dependencies for this setup.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">{:paths [\"src\" \"resources\" \"target\"]\n :deps  {org.clojure\/clojure       {:mvn\/version \"1.10.0\"}\n         com.bhauman\/figwheel-main {:mvn\/version \"0.2.0\"}\n         quil                      {:mvn\/version \"3.0.0\"}}}<\/pre><\/div><p>To kick things off, we define a namespace to hold our code in <code class=\"language-text\">src\/sketches\/perlin_flow.cljs<\/code>, and bring in the needed Quil functions.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(ns sketches.perlin-flow\n  (:require [quil.core :as q]\n            [quil.middleware :as middleware]))<\/pre><\/div><p>A developement build called <code class=\"language-text\">dev<\/code> is defined by placing a file <code class=\"language-text\">dev.cljs.edn<\/code> in the root of the project. This file instructs our build on what namespace that should run initially when the built javascript file loads.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">{:main sketches.perlin_flow}<\/pre><\/div><p>A simple <code class=\"language-text\">resources\/public\/index.html<\/code> loads the resulting sketch in the browser.<\/p><div class=\"gatsby-highlight\" data-language=\"html\"><pre class=\"language-html\">&lt;!doctype html&gt;\n&lt;html&gt;\n  &lt;head&gt;\n    &lt;title&gt;Perlin noise flow field&lt;\/title&gt;\n    &lt;meta name=\"viewport\" content=\"width=device-width, initial-scale=1\" \/&gt;\n    &lt;link rel=\"shortcut icon\"type=\"image\/x-icon\" href=\"data:image\/x-icon;,\"&gt;\n    &lt;style&gt;html, body { width: 100%; height: 100%; padding: 0; margin: 0; }&lt;\/style&gt;\n  &lt;\/head&gt;\n  &lt;body&gt;\n    &lt;div id=\"sketch\"&gt;&lt;\/div&gt;\n    &lt;script src=\"cljs-out\/dev-main.js\"&gt;&lt;\/script&gt;\n  &lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/div><p>Running the build is a matter of letting clojure know where to begin, and further instructing figwheel what build we want to be using, i.e. the <code class=\"language-text\">dev<\/code> one.<\/p><div class=\"gatsby-highlight\" data-language=\"bash\"><pre class=\"language-bash\">clojure -m figwheel.main -b dev<\/pre><\/div><p>A couple of vars <code class=\"language-text\">w<\/code> and <code class=\"language-text\">h<\/code> used throughout our sketch is defined to hold the width and height of the browser window. This also shows a general example of how function calls are made in a lisp, by placing the name of the function to be called as the first element in the list (i.e. inside the parens). However, it also shows the somewhat cryptic approach cljs takes to javascript interop, where a <code class=\"language-text\">.-<\/code> prefix in front of the name is how properties of native javascript objects are looked up. We will generally not be needing this prefix, as most our code simply refers ClojureScript vars and functions directly.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(def body (.-body js\/document))\n(def w (.-clientWidth body))\n(def h (.-clientHeight body))<\/pre><\/div><p>We define a function <code class=\"language-text\">create<\/code> to create our sketch by passing it the nescessary options. This defines the essense of our sketch, namely what functions will be used for state setup, update and draw. We'll keep them minimal for now. The setup function sets up our drawing's state, i.e. an empty vector <code class=\"language-text\">[]<\/code>. The update function gets the state of our sketch as and argument and simply returns it. The draw function, well, draws our sketch to the actual canvas in the browser. It also receives the state as an argument, but as of now, simply draws an ellipse to the center of the screen, on a dark <a href=\"http:\/\/quil.info\/api\/color\/setting#background\">background color<\/a>.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn sketch-setup []\n  [])\n(defn sketch-update [state]\n  state)\n(defn sketch-draw [state]\n  (q\/background 30)\n  (q\/ellipse (\/ w 2) (\/ h 2) 100 100))\n(defn create [canvas]\n  (q\/sketch\n    :host canvas\n    :size [w h]\n    :draw #'sketch-draw\n    :setup #'sketch-setup\n    :update #'sketch-update\n    :middleware [middleware\/fun-mode]\n    :settings (fn []\n                (q\/random-seed 666)\n                (q\/noise-seed 666))))<\/pre><\/div><p>By seeding Quil's <a href=\"http:\/\/quil.info\/api\/math\/random#random-seed\">random<\/a> and <a href=\"http:\/\/quil.info\/api\/math\/random#noise-seed\">noise<\/a> functions we ensure reprodusable results across page reloads even though we will be calling a couple of functions providing random values.<\/p><p>A clever cljs macro, <code class=\"language-text\">defonce<\/code>, ensures that we only create one instance of the sketch, even though the namespace is reloaded by figwheel each time the sketch file is saved.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defonce sketch (create \"sketch\"))<\/pre><\/div><p>A white circle appears.<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/kpudq7QN\/perlin-0.png?v=90fdaa8e282f9511254f427ad8b662fb\" alt=\"white circle\" referrerpolicy=\"no-referrer\" \/><\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/fe1d65e37feea10194f0abbb02a5203039cdfe62c512516256dda4a3c27c6512\">Try it yourself<\/a> (Hint: click \"Run\" in the top left corner)<\/p><p>The keen-eyed would have noticed our <code class=\"language-text\">create<\/code> function also brings out Quil's <a href=\"http:\/\/quil.info\/api\/middleware#fun-mode\"><code class=\"language-text\">fun-mode<\/code><\/a> middleware to enable a functional workflow for our sketch. (It's going to be <em>fun<\/em> as well, I promise.) This enables a data flow coherent with functional thinking, such that <code class=\"language-text\">sketch-setup<\/code> returns the initial state of the sketch, while <code class=\"language-text\">sketch-update<\/code> can be free of side effects, operating only on the immutable state received as an argument, returning the state for the next iteration. <code class=\"language-text\">sketch-draw<\/code> needs side effects to be able to draw to the canvas.<\/p><p>A couple of iterations with saved changes to <code class=\"language-text\">sketch-draw<\/code> might result in something like the following. As the canvas remembers what's been previously drawn, multiple circles will be shown in the image.<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/z8uw0PGW\/perlin-3.png?v=cf2d211951f363be096eb962a0013abb\" alt=\"multiple white circles\" referrerpolicy=\"no-referrer\" \/><\/p><p>Not a masterpiece, to say the least. Let's up the game.<\/p><p>Every piece of generative art needs a pretty palette. This palette is a map of predefined colors, on the form of vectors with rgb values. We'll use this palette to spice things up a bit when drawing particles to our canvas.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(def palette\n  {:name       \"purple haze\"\n   :background [10 10 10]\n   :colors     [[32 0 40]\n                [82 15 125]\n                [99 53 126]\n                [102 10 150]\n                [132 26 200]\n                [165 32 250]\n                [196 106 251]]})<\/pre><\/div><p>The state for this sketch will hold a vector of these particles, created by mapping the <code class=\"language-text\">particle<\/code> function over the range of numbers from 0 to 2000. On each change of the setup function we have to reload the page, as the <code class=\"language-text\">defonce<\/code> mentioned above keeps the sketch from re-initializing.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn sketch-setup\n  \"Returns the initial state to use for the update-render loop.\"\n  []\n  (map particle (range 0 2000)))<\/pre><\/div><p>Each particle is a map consisting of the <code class=\"language-text\">:id<\/code> of the particle, <code class=\"language-text\">:x<\/code> and <code class=\"language-text\">:y<\/code> representing its position constrained by the screen size, <code class=\"language-text\">:vx<\/code> and <code class=\"language-text\">:vy<\/code> representing its speed in the x and y directions respectively, <code class=\"language-text\">:size<\/code> representing the particle's size in pixels, <code class=\"language-text\">:direction<\/code> representing the direction of each particle and a random <code class=\"language-text\">:color<\/code> drawn from the palette.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn particle\n  \"Creates a particle map.\"\n  [id]\n  {:id         id\n   :vx         1\n   :vy         1\n   :size       3\n   :direction  0\n   :x          (q\/random w)\n   :y          (q\/random h)\n   :color      (rand-nth (:colors palette))})<\/pre><\/div><p>Our draw function starts off by clearing what was previously drawn by setting a new background color with <a href=\"http:\/\/quil.info\/api\/color\/setting#background\"><code class=\"language-text\">q\/background<\/code><\/a>. <a href=\"http:\/\/quil.info\/api\/color\/setting#no-stroke\">q\/no-stroke<\/a> ensure there's no stroke around each particle. Subsequently, it uses <code class=\"language-text\">doseq<\/code> to process each particle <code class=\"language-text\">p<\/code> from our state of particles (returned by <code class=\"language-text\">sketch-setup<\/code>) for side effects. It uses the Quil function <a href=\"http:\/\/quil.info\/api\/color\/setting#fill\"><code class=\"language-text\">q\/fill<\/code><\/a> to set the color of what is about to be drawn to that of the <code class=\"language-text\">:color<\/code> the particle, and calls out to <a href=\"http:\/\/quil.info\/api\/shape\/2d-primitives#ellipse\"><code class=\"language-text\">q\/ellipse<\/code><\/a> to draw a circle at the <code class=\"language-text\">:x<\/code> and <code class=\"language-text\">:y<\/code> positions of each particle with its <code class=\"language-text\">:size<\/code>.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn sketch-draw\n  \"Draws the current state to the canvas. Called on each iteration after sketch-update.\"\n  [particles]\n  (apply q\/background (:background palette))\n  (q\/no-stroke)\n  (doseq [p particles]\n    (apply q\/fill (:color p))\n    (q\/ellipse (:x p) (:y p) (:size p) (:size p))))<\/pre><\/div><p>This is what we get.<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/JruwJz4q\/perlin-1.png?v=6c2e9591166f52a233174665c0703bfb\" alt=\"a bunch of tiny circles\" referrerpolicy=\"no-referrer\" \/><\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/c686a76ad8533e29de3459a604be2424658efa1a1667a4f5ec098b2f5520fb9f\">Try it yourself<\/a><\/p><p>Already it's starting to look good, albeit static. Let's fix that.<\/p><p>Our <code class=\"language-text\">sketch-update<\/code> function receives the current state of particles and its task is to update their state for the next iteration of the drawing, before <code class=\"language-text\">sketch-draw<\/code> will draw the resulting state to the canvas. The update and draw functions will be called after one another for the lifetime of the sketch, resembling our program's render-loop. The following function definition moves each particle according to its current position by adding in its velocity in each direction. By running the modulo function <code class=\"language-text\">mod<\/code> on each resulting position, we ensure that particles come back on the opposite side of the screen if they move outside the screen bounds. The cljs function <a href=\"http:\/\/clojure.github.io\/clojure\/branch-master\/clojure.core-api.html#clojure.core\/assoc\"><code class=\"language-text\">assoc<\/code><\/a> associates new values for the given keys in each map.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn position\n  \"Calculates the next position based on the current, the speed and a max.\"\n  [current delta max]\n  (mod (+ current delta) max))\n(defn sketch-update\n  \"Receives the current state. Returns the next state to render.\"\n  [particles]\n  (map (fn [p]\n         (assoc p\n                :x (position (:x p) (:vx p) w)\n                :y (position (:y p) (:vy p) h)))\n       particles))<\/pre><\/div><p>The fact that none of our defined functions have their own state, means that Figwheel is able to patch the running javascript code with redefinitions we make to the source without loosing the current state of the sketch or what has been previously drawn. This is an amazing feature when experimenting with generative art, and provides an immensely short feedback loop - a must when letting your creative juices flow!<\/p><p>Look! It's moving.<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/wbuelrgQ\/perlin-2.gif?v=4752831353cef01ccb8d14eb8e0ad32a\" alt=\"a bunch of tiny circles moving across the screen\" referrerpolicy=\"no-referrer\" \/><\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/e04f085aef191095ef204ad6e516f6d4cd82d719c14c679a68196e62daa46778\">Try it yourself<\/a><\/p><p>Simply by removing the one line of code from the draw function that clears the background we see a radically different expression in our drawing. If you are coding along you would be starting to feel the effects that this kind of short feedback loop setup can have.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn sketch-draw\n  \"Draws the current state to the canvas. Called on each iteration after sketch-update.\"\n  [particles]\n  ;; (apply q\/background (:background palette))\n  (q\/no-stroke)\n  (doseq [p particles]\n    (apply q\/fill (:color p))\n    (q\/ellipse (:x p) (:y p) (:size p) (:size p))))<\/pre><\/div><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/qGuzmRmn\/perlin-4.png?v=0e56b15c815418a19121bde32358fe15\" alt=\"a bunch of tiny circles moving across the screen to create lines\" referrerpolicy=\"no-referrer\" \/><\/p><p>The circles become lines!<\/p><p>Rest assured, the <a href=\"http:\/\/quil.info\/sketches\/show\/9c8f533dbb319231922326e2eed19ecaa037df085e1e871a8ee2163b2a91b0e1\">circles are really moving<\/a>, but my gifs were getting huge due to the poorly compressable content!<\/p><p>Changing around each particle's speed by a random amount between <code class=\"language-text\">[-0.5, 0.5]<\/code> and capping it at 2 renders a rather busy result.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn velocity\n  \"Calculates the next velocity based on the current and a delta. Capped at 2.\"\n  [current delta]\n  (mod (+ current delta) 2))\n(defn sketch-update\n  \"Returns the next state to render. Receives the current state as a paramter.\"\n  [particles]\n  (map (fn [p]\n         (assoc p\n                :x  (position (:x p)  (:vx p) w)\n                :y  (position (:y p)  (:vy p) h)\n                :vx (velocity (:vx p) (q\/random -0.5 0.5))\n                :vy (velocity (:vy p) (q\/random -0.5 0.5))))\n       particles))<\/pre><\/div><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/YEuvLoWe\/perlin-3.png?v=5af95122755425154f0c34fbc3a9e66e\" alt=\"a busy image of a bunch of tiny circles moving across the screen to create squiggly lines\" referrerpolicy=\"no-referrer\" \/><\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/cadbb004fdaffc70e9dee98aff5e7c1e9edf453e7f3338f5698b5e63e6beb254\">Try it yourself<\/a><\/p><p>Let's introduce some Perlin noise to start heading for the effect we're after and attempt to calm things down a bit.<\/p><p>We will give each particle a direction from <code class=\"language-text\">[0,2\u03c0]<\/code> by use of Quil's <a href=\"http:\/\/quil.info\/api\/math\/random#noise\">Perlin noise function<\/a>, based on the noise (i.e. a number between <code class=\"language-text\">[0,1]<\/code>) at the particle's location x and y in the noise map. We'll let this represent \"the wind\" that carries the particles across the screen with slight variations in direction. By averaging the current velocity and the <code class=\"language-text\">Math\/cos<\/code>\/<code class=\"language-text\">Math\/sin<\/code> of the updated direction we keep it within bounds of <code class=\"language-text\">[0,1]<\/code>. To calm it down a bit we also scale the <code class=\"language-text\">x<\/code> and <code class=\"language-text\">y<\/code> values by the <code class=\"language-text\">noise-zoom<\/code>. By \"zooming in\" on the noise map, values close to each other will have increasingly similar random values, easing things out. This forms a great basis for creating natural looking flows of particles.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(def noise-zoom\n  \"Noise zoom level.\"\n  0.0005)\n(defn direction\n  \"Calculates the next direction between [0,2\u03c0] based on the perlin noise at the position (x,y).\"\n  [x y]\n  (* 2\n     Math\/PI\n     (q\/noise (* x noise-zoom) (* y noise-zoom))))\n(defn velocity\n  \"Calculates the next velocity by averaging the current velocity and the added delta.\"\n  [current delta]\n  (\/ (+ current delta) 2))\n(defn sketch-update\n  \"Returns the next state to render. Receives the current state as a paramter.\"\n  [particles]\n  (map (fn [p]\n         (assoc p\n                :x         (position (:x p) (:vx p) w)\n                :y         (position (:y p) (:vy p) h)\n                :direction (direction (:x p) (:y p))\n                :vx        (velocity (:vx p) (Math\/cos (:direction p)))\n                :vy        (velocity (:vy p) (Math\/sin (:direction p)))))\n       particles))<\/pre><\/div><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/4gu7d9eG\/perlin-4.png?v=36146e7b62562e6d98bdfd3f1d72ab21\" alt=\"calm, flowy, converging lines\" referrerpolicy=\"no-referrer\" \/><\/p><p>Getting there!<\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/294194671c4c42f1c303fcc7e1fc48976b18f440160170650a614c9d55445aec\">Try it yourself<\/a><\/p><p>To make things more natural looking, let's introduce some more noise on a per particle basis, to add the effect of uniqueness to each particle's movement. We'll do this by adding a touch more perlin noise to each calculated direction by using the <code class=\"language-text\">:id<\/code> of each particle as a third argument to the <a href=\"http:\/\/quil.info\/api\/math\/random#noise\"><code class=\"language-text\">q\/noise<\/code><\/a> function. This looks up the Perlin noise value in three dimensions, making it unique per particle.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn direction\n  \"Calculates the next direction based on the previous position and id of each particle.\"\n  [x y z]\n  (* 2\n     Math\/PI\n     (+ (q\/noise (* x noise-zoom) (* y noise-zoom))\n        (* 0.2 (q\/noise (* x noise-zoom) (* y noise-zoom) (* z noise-zoom))))))\n(defn sketch-update\n  \"Returns the next state to render. Receives the current state as a paramter.\"\n  [particles]\n  (map (fn [p]\n         (assoc p\n                :x         (position (:x p) (:vx p) w)\n                :y         (position (:y p) (:vy p) h)\n                :direction (direction (:x p) (:y p) (:id p))\n                :vx        (velocity (:vx p) (Math\/cos (:direction p)))\n                :vy        (velocity (:vy p) (Math\/sin (:direction p)))))\n       particles))<\/pre><\/div><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/xQuvyLmo\/perlin-6.png?v=f20d88f7ad7361b2121c5bc7a1f7aae9\" alt=\"slightly different, flowy lines\" referrerpolicy=\"no-referrer\" \/><\/p><p>Not radically different, I hear you.<\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/7a93b9f69a1cf218647864796b54052d41b50131cc17e2c2901e305779a4995d\">Try it yourself<\/a><\/p><p>Now try this: adjust the <code class=\"language-text\">noise-zoom<\/code> scale by a slight increase. In this way we are effectively \"zooming out\" of the Perlin noise cloud, making its randomness more varying, making the curves stand out more in the drawing.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(def noise-zoom\n  \"Noise zoom level.\"\n  0.005)<\/pre><\/div><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/12uEwJ26\/Screenshot+2019-11-23+at+12.42.13.png?v=bf2f1f2b82b18448576f4f5f396cfb4f\" alt=\"zoomed out, flowy lines\" referrerpolicy=\"no-referrer\" \/><\/p><p>It flows!<\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/155d38e0450216d08544082c04d4c317b5dab2229d9a914582fad8d7e8c82d70\">Try it yourself<\/a><\/p><p>Playing around with opacity often create stunning effects in sketches like these. Try modifying each drawn color by adding an alpha channel, of say 3, to the <code class=\"language-text\">[r g b]<\/code> color vectors making them <code class=\"language-text\">[r g b a]<\/code>, e.g. so the first color of the palette becomes <code class=\"language-text\">[32 0 40 3]<\/code>. We'll do this for all our colors by appending 3 to each color vector as it is drawn. This is done using the cljs function <a href=\"http:\/\/clojure.github.io\/clojure\/branch-master\/clojure.core-api.html#clojure.core\/conj\"><code class=\"language-text\">conj<\/code><\/a>.<\/p><div class=\"gatsby-highlight\" data-language=\"clojure\"><pre class=\"language-clojure\">(defn sketch-draw\n  \"Draws the current state to the canvas. Called on each iteration after sketch-update.\"\n  [particles]\n  ;;(apply q\/background (:background palette))\n  (q\/no-stroke)\n  (doseq [p particles]\n    (apply q\/fill (conj (:color p) 3))\n    (q\/ellipse (:x p) (:y p) (:size p) (:size p))))<\/pre><\/div><p>Not immediately exciting, hang in there.<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/wbue27n4\/perlin-7.png?v=6e8979fe70cb8dce990debf2c65ec9c2\" alt=\"slow start of drawing opaque, calm, flowy lines\" referrerpolicy=\"no-referrer\" \/><\/p><p>A couple of minutes later the following slowly appears!<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/6quD6PmP\/perlin-8.png?v=2072c5e63e4d75e023cdea52aa59987b\" alt=\"opaque, calm, flowy lines\" referrerpolicy=\"no-referrer\" \/><\/p><p>That's it! Your own, personal, generative art masterpiece is complete!<\/p><p><a href=\"http:\/\/quil.info\/sketches\/show\/22d90e5378f1614d3f2d743ec31f4d94b0d5304bc0f784f8f51f69dddfef91a0\">Try it yourself<\/a><\/p><p>To finish off, here's a couple of more renders for inspiration, done only by tweaking some of the hard coded values in the sketch discussed.<\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/8LuZXrkd\/Screenshot+2019-11-23+at+13.30.48.png?v=6d24c9106355c4a32123fd55147662fd\" alt=\"topological map\" referrerpolicy=\"no-referrer\" \/><\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/DOuGz9mA\/Screenshot+2019-11-23+at+13.34.16.png?v=326b9d0175bb307f380b8a053d4aa410\" alt=\"doodles\" referrerpolicy=\"no-referrer\" \/><\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/rRunroNz\/Screenshot+2019-11-23+at+13.35.31.png?v=b87c1f0ac13a0a1eedf9e18ea9034faf\" alt=\"blurry lines\" referrerpolicy=\"no-referrer\" \/><\/p><p><img src=\"https:\/\/p196.p4.n0.cdn.getcloudapp.com\/items\/E0uPxrg8\/Screenshot+2019-11-23+at+13.41.13.png?v=fc9b0ad808dba7e78a12eb3910e0f821\" alt=\"curvy lines, low f-stop\" referrerpolicy=\"no-referrer\" \/><\/p><p>And while your artwork might feel complete by the time you read this, what you'll find once you start experimenting with these kinds of sketches on your own is that most often that not you will never \"complete\" these kind of art pieces, because of the sheer joy of seeing the radical effects that minor tweaks you make will have!<\/p><p>Suggestions for further experiments:<\/p><ul><li>extract constants to variables so you can change them while the art is drawn<\/li>\n<li>adjust the size of the particles<\/li>\n<li>increase the speed of the particles<\/li>\n<li>increase the angle of the direction by orders of a magnitude<\/li>\n<li>adjust each particle's uniqueness<\/li>\n<li>change stuff around while the drawing is progressing<\/li>\n<li>clear the background on each render again, so all lines become independent particles<\/li>\n<li>experiment with <a href=\"https:\/\/kgolid.github.io\/chromotome-site\/\">different color palettes<\/a><\/li>\n<li>try new seeds for the random functions, or simply remove the seeds altogether, creating new, unique art pieces every time you render the page<\/li>\n<\/ul><p>I hope you enjoyed this tutorial and that I encouraged you to experiment with generative art on your own! Don't hesitate to hit me up with questions over at <a href=\"https:\/\/www.twitter.com\/torgeir\">twitter.com\/torgeir<\/a> or <a href=\"https:\/\/www.github.com\/torgeir\">github.com\/torgeir<\/a>.<\/p>","created_at":"2020-12-10T17:22:53+0000","updated_at":"2020-12-10T17:23:06+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"functional.christmas","preview_picture":"https:\/\/images.unsplash.com\/photo-1512389142860-9c449e58a543?w=1226&h=400&fit=crop&crop=edges","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1217,"title":"The 10 Best Studio Monitors\/Speakers for Home Recording","url":"https:\/\/ehomerecordingstudio.com\/best-studio-monitors\/","content":"<p><img class=\"alignright wp-image-25712 size-full\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/02\/10c-best-studio-monitors-e1424139698360.png\" alt=\"best studio monitors\" width=\"300\" height=\"447\" referrerpolicy=\"no-referrer\" \/>More so that perhaps any other <a href=\"https:\/\/ehomerecordingstudio.com\/recording-studio-equipment-list\/\">recording studio equipment<\/a>\u2026<\/p><p><strong>Studio monitors<\/strong> are regarded as possibly the single most important tool for achieving great sound.<\/p><p>Which is why\u00a0top studios will\u00a0sometimes spend <strong>$10,000+<\/strong>\u00a0on the best that money can buy.<\/p><p>And while that may be unrealistic for <em>home<\/em> studios\u2026<\/p><p>Fortunately,\u00a0you don\u2019t need to spend <em>that much<\/em> to get a great pair of monitors.<\/p><p>And to prove it to you, in today\u2019s\u00a0post I\u2019ve compiled the following list of:<\/p><p><strong>The 10 Best Studio Monitors for Home Recording<\/strong>.<\/p><p>First up\u2026<\/p><h2>The criteria for the\u00a0list<\/h2><p>When building this list, I decided to set\u00a0<strong>3 Key Guidelines<\/strong> to ensure a good mix of monitors:<\/p><ol><li><strong>$2000 price cap<\/strong>\u00a0\u2013 In my opinion, it\u2019s the absolute <em>max<\/em> a home studio should have to spend on a quality pair of semi-pro monitors.<\/li>\n<li><strong>One\u00a0brand=One pick<\/strong>\u00a0\u2013 To offer a variety of options, I limited the options\u00a0to ONE top model per\u00a0brand.<\/li>\n<li><strong>Reputation is key\u00a0<\/strong>\u2013\u00a0As we all know, \u201c<em>sound quality<\/em>\u201d is subjective. \u00a0What one ear loves, another hates. \u00a0And since every product description boasts an ultra-flat frequency response, with a tight bottom, focused mids, and detailed highs\u2026the only fair way to judge them is by\u00a0<strong>reputation<\/strong>.<\/li>\n<\/ol><p>More than any other factor, the monitors with the <strong>best long-standing reputations<\/strong> are the ones that made this list.<\/p><p>Now let\u2019s begin. \u00a0First up\u2026<\/p><h2>1. KRK Rokit 5\u00a0G3<\/h2><p><img class=\"alignright wp-image-15071 size-full\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2014\/06\/krk-e1404421022509.jpg\" alt=\"cheap studio monitors - krk rokit 5\" width=\"280\" height=\"256\" referrerpolicy=\"no-referrer\" \/><\/p><p>When you browse\u00a0online, at pictures of OTHER\u00a0home\u00a0studios\u2026<\/p><p>One of the first things you notice is the monitors.<\/p><p>And VERY often\u2026those\u00a0distinct yellow dots of the <strong>KRK Rokit 5 G3<\/strong> are exactly what you see.<\/p><p>And that\u2019s because, at the \u201centry-level\u201d price point, there is no other set of monitors\u2026.<\/p><ul><li><strong>more popular<\/strong><\/li>\n<li><strong>more reviewed<\/strong><\/li>\n<li><strong>and more highly recommended<\/strong><\/li>\n<\/ul><p>\u2026than <em>these<\/em>.<\/p><p>While most\u00a0monitors in this price range receive far more <em>complaints<\/em> than <em>praise<\/em>\u2026<\/p><p>The fact is,\u00a0you could read reviews for days without\u00a0finding a single bad word about the KRK\u2019s. \u00a0That is why, for beginner studios,\u00a0they\u2019re the obvious choice, and the only ones I recommend.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1PuVcKG\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2FYgl5K\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/KRK\/Rokit-5-G3.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/krk-rokit-5-g3\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/krk_rp5_rokit_g3.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>Two other good beginner\u2019s options you can also check out are:<\/p><ul><li><strong>JBL LSR305<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1UkFqGz\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KD9Eap\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/JBL\/LSR305-5-Powered-Studio-Monitor.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/jbl-lsr305-5-powered-studio-monitor?rNtt=JBL%20LSR305&amp;index=2\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/jbl_lsr_305.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<li><strong>M Audio AV42<\/strong>\u00a0\u2013 (<a href=\"http:\/\/amzn.to\/1UkFCoZ\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KBN8ij\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/search?Ns=r&amp;Ntt=M+Audio+AV42\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/m-audio-av42-studio-monitor-pair\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/m_audio_av42.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>Up next, let\u2019s move on to some pricier models. \u00a0Starting with\u2026<\/p><h2>2. Yamaha HS8<\/h2><p><img class=\"alignright size-full wp-image-16457\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2014\/07\/Yamaha-HS8-e1405721069422.jpg\" alt=\"Yamaha HS8\" width=\"210\" height=\"288\" referrerpolicy=\"no-referrer\" \/><\/p><p>Based on the original Yamaha NS-10\u2019s from back in the 70\u2019s\u2026<\/p><p>The\u00a0<strong>Yamaha HS8\u2019s<\/strong>\u00a0combine a classic sound, with some new technology.<\/p><p>Instantly recognizable for their signature white coned woofers, these monitors\u00a0long ago solidified their iconic status.<\/p><p>Some of its notable features include:<\/p><ol><li><strong>a unique mounting system<\/strong>\u00a0\u2013 which minimizes vibration and improves performance<\/li>\n<li><strong>extra large magnets<\/strong>\u00a0\u2013 for a smoother response over a wider frequency range<\/li>\n<\/ol><p>As part of the popular mid-priced HS series, similar version of these monitors include:<\/p><ul><li><strong>the HS5<\/strong>\u00a0\u2013 (5\u2033 cone)<\/li>\n<li><strong>the HS7<\/strong>\u00a0\u2013 (7\u2033cone)<\/li>\n<li><strong>the HS8S<\/strong>\u00a0\u2013 (bigger 150W driver)<\/li>\n<\/ul><p>For monitors considered by many as \u201cprofessional\u201d quality, the Yamaha HS8\u2019s are about as cheap as they come.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1sHhV1d\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KBoLRL\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Yamaha\/HS8-Powered-Studio-Monitor.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/yamaha-hs8-powered-studio-monitor?rNtt=Yamaha%20HS8&amp;index=1\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/yamaha_hs_8.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>The only other monitors on this list comparable in BOTH quality and price, are\u2026<\/p><h2>3. Event 20\/20 BAS<\/h2><p><img class=\"alignright size-full wp-image-16456\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2014\/07\/Event-2020-BAS-e1405720881337.jpg\" alt=\"Event 2020 BAS\" width=\"230\" height=\"274\" referrerpolicy=\"no-referrer\" \/><\/p><p>Also a reissue of an old classic, the <strong>Event 20\/20BAS<\/strong> is based off the original Event 20\/20bas released back in 1995.<\/p><p>To improve on the old design, Event employed many of the newer technologies used in making their flagship monitors: the <strong>Event Opal<\/strong>.<\/p><p>Compared to the original models, the new ones boast DOUBLE the power, with HALF the Total Harmonic Distortion (THD).<\/p><p>And that\u2019s DAMN\u00a0impressive.<\/p><p>Other notable features include:<\/p><ul><li><strong>A front bass port<\/strong>\u00a0\u2013 which is better suited for the smaller rooms of most home studios<\/li>\n<li><strong>250 W power<\/strong>\u00a0\u2013 which is exceptionally high for monitors of this size<\/li>\n<\/ul><p><strong>According to the Event website<\/strong>:<\/p><blockquote>\n<p>The very first pair of 20\/20bas speakers were used by Francis Buckley to mix Quincy Jones\u2019\u00a0<em>Q\u2019s Jook Joint<\/em>\u00a0album for which he won the 1996 Grammy Award for\u00a0<em>Best Engineered Album<\/em>.<\/p>\n<\/blockquote><p>For monitors at this price (or ANY price), that\u2019s quite an accomplishment.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1PuWlBK\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KBhYr3\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Event\/20-20BAS-Monitor.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>)<\/li>\n<\/ul><p>Next let\u2019s move on to what I consider the most elegant looking monitors on the list\u2026<\/p><h2>4. Dynaudio BM5 mkIII<\/h2><p><img class=\"alignright wp-image-31304 size-full\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Dynaudio-BM5-mkIII-e1466050670651.jpg\" alt=\"Dynaudio BM5 mkIII\" width=\"220\" height=\"355\" referrerpolicy=\"no-referrer\" \/>Chances are good you\u2019ve heard of the <strong>Dynaudio<\/strong> Company before.<\/p><p>Established back in 1977, they\u00a0are currently one of the most established brands in the field of studio monitors.<\/p><p>The sleek and sophisticated look of their monitors, in addition to their sound, make them one of the trendiest\u00a0brands of today.<\/p><p>Among their most popular models, is the\u00a0<strong>Dynaudio BM5 mkIII<\/strong>.<\/p><p>Notable features of these monitors include:<\/p><ul><li><strong>soft high-frequency waveguide technology<\/strong><\/li>\n<li><strong>low, mid, and high frequency filters<\/strong><\/li>\n<li><strong>bass reflex cabinet with radius front edges<\/strong><\/li>\n<\/ul><p>Check it out:<strong>\u00a0<\/strong><\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1sHhXWM\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KyEml0\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Dynaudio-Acoustics\/BM5-mkIII-Studio-Monitor-Each.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/dynaudio-acoustics-bm5-mkiii-studio-monitor-each\/j07423004000000\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/dynaudio_bm5_mkiii.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>Next up is\u2026<\/p><h2>5. Mackie HR824 mkII<\/h2><p><img class=\"alignright\" title=\"Mackie HR824mk2 Studio Monitor\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Mackie-HR824mk21-184x300.jpg\" alt=\"Mackie HR824mk2 - best studio monitors\" width=\"184\" height=\"300\" referrerpolicy=\"no-referrer\" \/><\/p><p>Traditionally known for their mixers\u2026<\/p><p>The\u00a0<strong>Mackie<\/strong>\u00a0brand\u00a0is no slouch when it comes to studio monitors either.<\/p><p>And\u00a0the\u00a0<strong>Mackie HR824 mkII<\/strong>\u00a0is one of their most popular models.<\/p><p>One of the most unique features of these monitors is its\u00a0\u201cAcoustic Space Control\u201d settings\u2026<\/p><p>Which provides 3 separate settings to \u00a0provide optimal bass performance according to\u00a0the dimensions of your room.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1PuVSj8\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KBufvW\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Mackie\/HR824mk2-Studio-Monitor-2010.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/mackie-hr824mk2-studio-monitor-2010?rNtt=Mackie%20HR824%20mkII&amp;index=1\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>)<\/li>\n<\/ul><p>This next pick is from brand NOT\u00a0normally known for their monitors\u2026<\/p><h2>6. Neumann KH 120<\/h2><p><img class=\"alignright size-full wp-image-16459\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2014\/07\/Neumann-KH120-e1405721373203.jpg\" alt=\"Neumann KH120\" width=\"210\" height=\"280\" referrerpolicy=\"no-referrer\" \/><\/p><p>Best known as makers of the finest <a href=\"https:\/\/ehomerecordingstudio.com\/types-of-microphones\/\">microphones<\/a> in the World\u2026<\/p><p>The Neumann company, entered the studio monitor game back in 2010\u2026<\/p><p>When they\u00a0assumed the now retired <strong>Klein+Hummel<\/strong>\u00a0line.<\/p><p>Their first release was the respectfully titled <strong>Neumann KH<\/strong> line, which featured the now popular, the\u00a0<strong>Neumann KH 120<\/strong>.<\/p><p>Notable features of these monitors include:<\/p><ul><li><strong>Mathematically Modeled Dispersion (MMD) waveguide<\/strong><\/li>\n<li><strong>Titanium fabric dome drivers<\/strong><\/li>\n<li><strong>Composite sandwich cone design<\/strong><\/li>\n<\/ul><p>Check it out:<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1PuVTn7\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Neumann\/KH-120-Active-Nearfield-Studio-Monitor.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/neumann-kh-120-active-studio-monitor\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/neumann_kh_120_a.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>Also check out the <strong>KH120 <em>D<\/em><\/strong> which features digital I\/O and an internal DAC (<strong>D<\/strong>igital <strong>A<\/strong>udio <strong>C<\/strong>onverter) for monitoring digital inputs.<\/p><p>Check it out:<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"https:\/\/amzn.to\/2KBpdzk\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KDabsV\">B&amp;H<\/a>)<\/li>\n<\/ul><p>Next up, we have\u2026<\/p><h2>7. Adam Audio A7X<\/h2><p><img class=\"alignright size-full wp-image-16458\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2014\/07\/Adam-Audio-A7X-e1405721247896.jpg\" alt=\"Adam Audio A7X\" width=\"220\" height=\"285\" referrerpolicy=\"no-referrer\" \/><\/p><p>ADAM Audio, one of the youngest companies on this list, was founded in Germany back in 1999.<\/p><p>Currently, the 3 lines of monitors they offer are:<\/p><ol><li><strong>the high-end\u00a0SX series<\/strong><\/li>\n<li><strong>the budget\u00a0F series<\/strong><\/li>\n<li><strong>and the midrange\u00a0AX series<\/strong><\/li>\n<\/ol><p>No surprise, the <strong>AX series<\/strong> is most popular, and within it, the\u00a0<strong>Adam Audio A7X<\/strong>\u00a0is the most famous model of them all.<\/p><p>Known especially for its unique\u00a0<strong>X-ART Tweeter<\/strong>, the A7X\u2019s boast what may be the most hi-tech\u00a0tweeters on this entire list.<\/p><p>Its specs are damn impressive, the most notable one being: the ability to reproduce\u00a0frequencies all the way up to <strong>50kHz<\/strong>.<\/p><p>That\u2019s WELL beyond the limit of human hearing! \u00a0What that means is\u2026<\/p><p>WITHIN our hearing range, they\u00a0provide an extremely flat frequency response.<\/p><p>Another nice feature is its\u00a0FRONT\u00a0bass ports, which allows for closer placement to the walls in smaller rooms.<\/p><p>Personally I find them ugly as hell, but looks aren\u2019t the priority, are they? \u00a0If SOUND\u00a0is what you care about, I highly recommend them.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1PuVX6l\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KA6cNA\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Adam-Audio\/A7X-Powered-Studio-Monitor.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/adam-audio-a7x-powered-studio-monitor?rNtt=Adam%20Audio%20A7X&amp;index=1\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/prod_AR_402777.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>Next up is\u2026<\/p><h2>8. Genelec M040<\/h2><p><img class=\"alignright size-full wp-image-16455\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2014\/07\/Genelec-M040-e1405720741717.jpg\" alt=\"Genelec M040\" width=\"220\" height=\"298\" referrerpolicy=\"no-referrer\" \/><\/p><p>More than any name\u00a0on this list\u2026<\/p><p>The\u00a0<strong>Genelec<\/strong>\u00a0name is the one synonymous with\u00a0<strong>high-end studio monitors<\/strong>.<\/p><p>In fact, most of the models from their famous\u00a0<strong>8000 series<\/strong>\u00a0are FAR too expensive for this list.<\/p><p><strong>However<\/strong>\u2026<\/p><p>With their recent\u00a0<strong>M series<\/strong>, Genelec has found a way to make themselves at least somewhat affordable to the average guy.<\/p><p>The most popular model of the series\u00a0is the\u00a0<strong>Genelec M040\u2019s<\/strong>.<\/p><p>It\u2019s notable features include:<\/p><ul><li><strong>Downward facing bass port for use in smaller rooms<\/strong><\/li>\n<li><strong>Energy-friendly Class D amplifiers<\/strong><\/li>\n<li><strong>Intelligent Signal Sensing auto power-off<\/strong><\/li>\n<\/ul><p><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1txt5pO\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KA6rbs\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.guitarcenter.com\/Genelec\/M040-Active-2-Way-Monitor-Each.gc\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">GuitarC<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/http:\/\/www.musiciansfriend.com\/pro-audio\/genelec-m040-active-2-way-monitor-each\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">MusiciansF<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/genelec_m040.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/p><p>Up next\u2026<\/p><h2>9. Focal Twin6 Be<\/h2><p><img class=\"size-medium wp-image-3112 alignright\" title=\"Focal Twin6 Be Studio Monitor\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Focal-Twin6-Be1-300x179.jpg\" alt=\"Focal Twin6 Be - best studio monitors\" width=\"300\" height=\"179\" referrerpolicy=\"no-referrer\" \/>Normally reserved ONLY\u00a0for the most ULTRA-high end monitors\u2026<\/p><p>The unique 3 way design of the\u00a0<strong>Focal Twin6 Be<\/strong>\u00a0is the primary feature that sets these monitors apart from the rest.<\/p><p>While most 3 way designs use a larger driver for the bass, and a smaller one for the mids, this model instead uses identical 6.5\u2033 drivers for both.<\/p><p>As anyone who used them before would tell you, the extra 3rd cone makes one hell of difference in sound.<\/p><p><strong>Now you would think<\/strong>\u2026<\/p><p>That since these monitors are WAY pricier than the others\u00a0on the list\u2026that they would also be LESS\u00a0popular.<\/p><p>Yet despite their cost, the\u00a0<strong>Focal Twin6 Be\u2019s\u00a0<\/strong>appear to be among the most popular set of monitors in ANY\u00a0category. \u00a0If you can afford them, THESE are the ones I recommend most over ALL\u00a0the others.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1W0o4B6\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KyEW2a\">B&amp;H<\/a>\/<a href=\"http:\/\/www.anrdoezrs.net\/links\/7791709\/type\/dlg\/https:\/\/www.guitarcenter.com\/FOCAL\/Focal-Twin6-Be-3-Way-Studio-Monitor.gc\">GuitarC<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/focal_twin_6_be.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>And for the last pick on the list\u2026<\/p><h2>10. Avantone Mix Cubes<\/h2><p><img class=\"alignright size-full wp-image-3111\" src=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Avantone-Mix-Cubes-e1379466375865.jpg\" alt=\"Avantone Mix Cubes\" width=\"285\" height=\"170\" srcset=\"https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Avantone-Mix-Cubes-e1379466375865.jpg 285w, https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Avantone-Mix-Cubes-e1379466375865-60x35.jpg 60w, https:\/\/289leu411bct5nset271z5tw-wpengine.netdna-ssl.com\/wp-content\/uploads\/2012\/09\/Avantone-Mix-Cubes-e1379466375865-150x89.jpg 150w\" referrerpolicy=\"no-referrer\" \/>A common problem when working on\u00a0high-end monitors is\u2026<\/p><p>A\u00a0mix that sounds spectacular in the studio\u2026<\/p><p>May not always translate well when played on\u00a0lo-fi consumer speakers.<\/p><p>To solve this problem, the <strong>Auratone<\/strong> company invented a small set of monitors in 1958, known as the <strong>Auratone 5C Super Sound Cube<\/strong>\u2026<\/p><p>Which simulated the <em>less-than-ideal<\/em> performance conditions of typical consumer speakers.<\/p><p>For decades to come, the Auratones held strong as the industry standards in secondary reference monitors.<\/p><p>Once they were discontinued, many copycats arose\u2026<\/p><p>And the <strong>Avantone Mix Cubes<\/strong>, eventually became the new standard which we all use today.<\/p><p>As a supplement to ANY of the other monitors on this list, I highly recommend you grab yourself a pair of <strong>Mix Cubes<\/strong> as well.<\/p><ul><li><strong>Click to compare\u00a0prices<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1W0o8Rm\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>\/<a href=\"https:\/\/bhpho.to\/2KzGLvQ\">B&amp;H<\/a>\/<a href=\"https:\/\/www.thomann.de\/gb\/avantone_mixcubes_active_black.htm?offid=1&amp;affid=181\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Thomann<\/a>)<\/li>\n<\/ul><p>Also check it out its closest competitor: the <strong>Behritone C50a<\/strong> \u2013 (<a href=\"http:\/\/amzn.to\/1PuWx3Z\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Amazon<\/a>)<\/p>","created_at":"2020-12-07T02:54:59+0000","updated_at":"2020-12-07T02:54:59+0000","published_at":"2012-09-28T07:26:35+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"ehomerecordingstudio.com","preview_picture":"https:\/\/ehomerecordingstudio.com\/wp-content\/uploads\/2012\/09\/10e-best-studio-monitors.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1216,"title":"EmacsConf - 2020 - Talks","url":"https:\/\/emacsconf.org\/2020\/talks\/","content":"<div id=\"pagebody\"><div id=\"content\" role=\"main\"><p><a href=\"https:\/\/emacsconf.org\/2020\/emacsconf-2020.m3u\">Download an .m3u playlist<\/a><\/p><p>EmacsConf 2020 was on November 28 (Sat) and November 29 (Sun), 2020 from 9am-5pm Toronto\/EST time; equivalently, 6am-2pm PST, 2pm-10pm UTC, 3pm-11pm Zurich\/CET.<\/p><p>Many of the talks include accompanying material such as slides, questions, and notes. When present, these material are included or linked to on the talk page.<\/p><table><thead><tr><th>Title<\/th>\n<th>Speakers<\/th>\n<\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/00\">Day 1 opening remarks<\/a><\/td>\n<td>Amin Bandali, Sacha Chua, Leo Vivier<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/01\">Emacs News Highlights<\/a><\/td>\n<td>Sacha Chua<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/02\">An Emacs Developer Story: From User to Package Maintainer<\/a><\/td>\n<td>Leo Vivier<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/03\">Idea to Novel Superstructure: Emacs for Writing<\/a><\/td>\n<td>Bala Ramadurai<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/04\">Music in Plain Text<\/a><\/td>\n<td>Jonathan Gregory<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/05\">Bard Bivou(m)acs - Building a bandcamp-like page for an album of music<\/a><\/td>\n<td>Grant Shangreaux<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/06\">Trivial Emacs Kits<\/a><\/td>\n<td>Corwin Brust (mplsCorwin)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/07\">Beyond Vim and Emacs: A Scalable UI Paradigm<\/a><\/td>\n<td>Sid Kasivajhula (countvajhula)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/08\">Building reproducible Emacs<\/a><\/td>\n<td>Andrew Tropin (abcdw)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/21\">On why most of the best features in eev look like 5-minute hacks<\/a><\/td>\n<td>Eduardo Ochs (edrx)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/09\">Orgmode - your life in plain text<\/a><\/td>\n<td>Rainer K\u00f6nig<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/10\">Lead your future with Org<\/a><\/td>\n<td>Andrea<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/11\">the org-gtd package: opinions about Getting Things Done<\/a><\/td>\n<td>Aldric<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/12\">One Big-ass Org File or multiple tiny ones? Finally, the End of the debate!<\/a><\/td>\n<td>Leo Vivier<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/13\">Experience Report: Steps to \"Emacs Hyper Notebooks\"<\/a><\/td>\n<td>Joseph Corneli, Raymond Puzio, and Cameron Ray Smith<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/14\">README-Driven Design<\/a><\/td>\n<td>Adam Ard<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/15\">Moving from Jekyll to OrgMode, an experience report<\/a><\/td>\n<td>Adolfo Villafiorita<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/16\">Org-roam: Presentation, Demonstration, and What's on the Horizon<\/a><\/td>\n<td>Leo Vivier<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/17\">Org-mode and Org-Roam for Scholars and Researchers<\/a><\/td>\n<td>Noorah Alhasan<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/18\">Org-roam: Technical Presentation<\/a><\/td>\n<td>Leo Vivier<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/19\">Sharing blogs (and more) with org-webring<\/a><\/td>\n<td>Brett Gilio<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/20\">OMG Macros<\/a><\/td>\n<td>Corwin Brust (mplsCorwin)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/40\">Day 1 closing remarks<\/a><\/td>\n<td>Amin Bandali, Sacha Chua, Leo Vivier, Corwin Brust<\/td>\n<\/tr><\/tbody><\/table><table><thead><tr><th>Title<\/th>\n<th>Speakers<\/th>\n<\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/41\">Day 2 opening remarks<\/a><\/td>\n<td>Amin Bandali, Sacha Chua, Leo Vivier<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/38\">Emacs development update<\/a><\/td>\n<td>John Wiegley<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/22\">Powering-up Special Blocks<\/a><\/td>\n<td>Musa Al-hassy<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/23\">Incremental Parsing with emacs-tree-sitter<\/a><\/td>\n<td>Tu\u1ea5n-Anh Nguy\u1ec5n<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/24\">Analyze code quality through Emacs: a smart forensics approach and the story of a hack<\/a><\/td>\n<td>Andrea<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/25\">Traverse complex JSON structures with live feedback<\/a><\/td>\n<td>Zen Monk Alain M. Lafon<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/39\">NonGNU ELPA<\/a><\/td>\n<td>Richard Stallman<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/26\">Emacs as a Highschooler: How It Changed My Life<\/a><\/td>\n<td>Pierce Wang<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/27\">State of Retro Gaming in Emacs<\/a><\/td>\n<td>Vasilij \"wasamasa\" Schneidermann<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/28\">Welcome To The Dungeon<\/a><\/td>\n<td>Erik Elmshauser and Corwin Brust<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/29\">Pathing of Least Resistance<\/a><\/td>\n<td>Erik Elmshauser and Corwin Brust (mplsCorwin)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/30\">A tour of vterm<\/a><\/td>\n<td>Gabriele Bozzola (@sbozzolo)<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/31\">Lakota Language and Emacs<\/a><\/td>\n<td>Grant Shangreaux<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/32\">Object Oriented Code in the Gnus Newsreader<\/a><\/td>\n<td>Eric Abrahamsen<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/33\">Maxima a computer algebra system in Emacs<\/a><\/td>\n<td>Fermin MF<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/34\">Extend Emacs to Modern GUI Applications with EAF<\/a><\/td>\n<td>Matthew Zeng<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/35\">WAVEing at Repetitive Repetitive Repetitive Music<\/a><\/td>\n<td>Zachary Kanfer<\/td>\n<\/tr><tr><td><a href=\"https:\/\/emacsconf.org\/2020\/talks\/42\">Day 2 closing remarks<\/a><\/td>\n<td>Amin Bandali, Sacha Chua, Leo Vivier, Corwin Brust<\/td>\n<\/tr><\/tbody><\/table><\/div><\/div><div id=\"footer\" class=\"pagefooter\" role=\"contentinfo\"><div id=\"pageinfo\"><p>Backlinks: <a href=\"https:\/\/emacsconf.org\/2020\/\">2020<\/a> <a href=\"https:\/\/emacsconf.org\/2020\/sidebar\/\">sidebar<\/a><\/p><p>Copyright \u00a9 2020 Amin Bandali and Sacha Chua<\/p><div class=\"pagelicense\"><p>Unless otherwise noted, the material on the EmacsConf wiki are <a href=\"https:\/\/emacsconf.org\/COPYING\/\">dual-licensed under CC BY-SA 4.0 and GPLv3+<\/a>.<\/p><\/div><p>Last edited Thursday 3 December 2020 at 14:22 (UTC)<\/p><\/div><\/div>","created_at":"2020-12-06T21:56:43+0000","updated_at":"2020-12-06T21:56:43+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"emacsconf.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1215,"title":"The Tragic Tale Of DEC, The Computing Giant That Died Too...","url":"https:\/\/digital.com\/about\/dec\/","content":"<p class=\"page-last-modified\">Last Updated on October 21, 2020<\/p><p>When you think of leaders in the computing industry, your first thoughts probably turn to companies like Apple, Microsoft, and IBM. But flash back a few decades, and the leaders of those companies all would\u2019ve gladly bowed down to Digital Equipment Corporation, or DEC, who began paving the way for everyone starting in 1957.<\/p><p>DEC identified a demand for more affordable and high-performing computing systems that could be used in scientific research and other technological settings. DEC introduced a mid-range computing solution, the minicomputer, at a time when the alternative was too bulky and costly for most people.<\/p><figure id=\"attachment_9580\" aria-describedby=\"caption-attachment-9580\" class=\"wp-caption alignright c2\"><noscript>\n<p><img class=\"wp-image-9580\" src=\"https:\/\/digital.com\/wp-content\/uploads\/Papertape2-300x154.jpg?x88029\" alt=\"Paper tape\" width=\"480\" height=\"246\" srcset=\"https:\/\/digital.com\/wp-content\/uploads\/Papertape2-300x154.jpg 300w, https:\/\/digital.com\/wp-content\/uploads\/Papertape2-1024x524.jpg 1024w, https:\/\/digital.com\/wp-content\/uploads\/Papertape2.jpg 2021w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript>\n<figcaption id=\"caption-attachment-9580\" class=\"wp-caption-text\">Digital paper tape, circa 1970s<\/figcaption><\/figure><p>In addition to its extensive lines of minicomputers, it also became involved with software as well as the internet (in the very, very early days of the internet).<\/p><p>So, what happened to DEC? Why did all the success and industry dominance they had for roughly three decades just suddenly slip away? Most argue that it was a failure of the company\u2019s leadership to adapt to the changing direction that computing began to take in the late 1980s.<\/p><p>In order to understand how DEC eventually lost it all, it\u2019s important to take a look back at how it all began and why such a monumental shift in the computing industry could\u2019ve been missed by such a critical leader in the space.<\/p><h2 id=\"history\">The Impressive (and Tragic) History of DEC in Computing<\/h2><p>DEC\u2019s story begins in 1957 with two men at <a href=\"http:\/\/web.mit.edu\/\">MIT<\/a>\u00a0who had a theory and an idea for a business. It ends in roughly 1998 with the selling of the company to Compaq. Here is DEC\u2019s journey.<\/p><h3 id=\"1957\">1957<\/h3><p>While working together in the Lincoln Laboratory at MIT, Ken Olsen and Harlan Anderson came up with the idea for DEC. They took notice of how popular the interactive computing machines were with students who visited the lab, as opposed to the batch processing machines that users couldn\u2019t input data into or use for real-time feedback.<\/p><p>Olsen and Anderson also knew that these interactive machines could be made and sold cheaper, which would be another major draw for researchers and scientists who were craving the technology.<\/p><p>After creating their original business plan, Olsen and Anderson were advised to shift the focus of their company from one around \u201ccomputers\u201d to instead one focused on \u201cequipment.\u201d There were just too many concerns from investors about the future of computing, and so that\u2019s what they did. The plan for Digital Equipment Corporation was finalized and they received $70,000 to launch their new company.<\/p><p>In sticking with the theme of keeping expenses light (not only for customers but for the company as well), they set up shop in an old wool mill in Maynard, MA.<\/p><h3 id=\"1958\">1958<\/h3><p>The first product DEC created and sold was its Digital Laboratory Module. Staying true to its business plan, the initial product lines DEC focused on were modules, or electronic components, that were mounted to circuit boards.\u00a0By the end of that year, DEC sold $94,000 worth of its first product.<\/p><figure id=\"attachment_9583\" aria-describedby=\"caption-attachment-9583\" class=\"wp-caption aligncenter c3\"><noscript>\n<p><img class=\"wp-image-9583 size-full\" src=\"https:\/\/digital.com\/wp-content\/uploads\/Dec_SYSTEM_BUILDING_BLOCKS_logo.jpg?x88029\" alt=\"System Building Blocks logo\" width=\"716\" height=\"211\" srcset=\"https:\/\/digital.com\/wp-content\/uploads\/Dec_SYSTEM_BUILDING_BLOCKS_logo.jpg 716w, https:\/\/digital.com\/wp-content\/uploads\/Dec_SYSTEM_BUILDING_BLOCKS_logo-300x88.jpg 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript>\n<figcaption id=\"caption-attachment-9583\" class=\"wp-caption-text\">System Building Blocks logo, RTC\/ Wikipedia (CC BY-SA 3.0)<\/figcaption><\/figure><h3 id=\"1960\">1960<\/h3><p>DEC began selling its first computer at the end of 1960. But it was aware of people\u2019s reluctance to invest in \u201ccomputer\u201d technology at the time, so it named the computer a \u201cprogrammable data processor\u201d, or PDP. The first iteration of this product line (PDP-1) sold for $120,000 that year.<\/p><p>Throughout the remainder of the decade, DEC created over a dozen PDP variants.<\/p><h3 id=\"1962\">1962<\/h3><figure id=\"attachment_9579\" aria-describedby=\"caption-attachment-9579\" class=\"wp-caption aligncenter c4\"><noscript>\n<p><img class=\"size-full wp-image-9579\" src=\"https:\/\/digital.com\/wp-content\/uploads\/3874706978_902597beb1_z.jpg?x88029\" alt=\"PDP-4s\" width=\"640\" height=\"511\" srcset=\"https:\/\/digital.com\/wp-content\/uploads\/3874706978_902597beb1_z.jpg 640w, https:\/\/digital.com\/wp-content\/uploads\/3874706978_902597beb1_z-300x240.jpg 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript>\n<figcaption id=\"caption-attachment-9579\" class=\"wp-caption-text\">PDP-4 at the National Archives Auditorium, 1964<\/figcaption><\/figure><p>DEC began the process of creating \u201cnew\u201d models of the PDP that could be sold for much less than the original. For example, the PDP-4 was similar in most ways to the PDP-1, but it was slower and packaged differently, which is what enabled DEC to sell it for $65,000.<\/p><h3 id=\"1964\">1964<\/h3><p>While continuing to release new PDPs into the market, DEC also charged forward in its delivery of new modules. The Flip Chip came out in 1964 and was meant to convert the PDP-4 to the PDP-7. Many of its subsequent module releases served a similar purpose: helping users convert their old computers to upgraded versions.<\/p><h3 id=\"1965\">1965<\/h3><p>It was in this year when DEC released the PDP-8, which is widely recognized as the first successful commercial minicomputer. Part of this is due to the improvements made to this model, but a large reason for its commercial success was the price tag of $18,500 and the 50,000 customers they sold it to.<\/p><h3 id=\"1970\">1970<\/h3><p>In the interim, DEC came up with a revamped version of their PDP line and released the PDP-11 minicomputer. Not only did it bring major upgraded features to their computing machines, it also was easier to use. By the time it stopped selling it in the 1990s, DEC sold over 600,000 of them, making it one of the most popular minicomputers ever.<\/p><figure id=\"attachment_9581\" aria-describedby=\"caption-attachment-9581\" class=\"wp-caption aligncenter c5\"><noscript>\n<p><img class=\"size-full wp-image-9581\" src=\"https:\/\/digital.com\/wp-content\/uploads\/1024px-DEC_PDP-11_20_computer_at_the_Computer_History_Museum.jpg?x88029\" alt=\"PDP-11\" width=\"1024\" height=\"768\" srcset=\"https:\/\/digital.com\/wp-content\/uploads\/1024px-DEC_PDP-11_20_computer_at_the_Computer_History_Museum.jpg 1024w, https:\/\/digital.com\/wp-content\/uploads\/1024px-DEC_PDP-11_20_computer_at_the_Computer_History_Museum-300x225.jpg 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript>\n<figcaption id=\"caption-attachment-9581\" class=\"wp-caption-text\">PDP-11 by Don DeBold\/ Flickr (CC BY 2.0)<\/figcaption><\/figure><p>In addition, the design of the computer, as well as its operating system, turned out to be immensely popular with other computing companies, that eventually ended up using it as inspiration for their own work.<\/p><h3 id=\"1971\">1971<\/h3><p>DEC made its first move into the European market in 1971, setting up manufacturing in Ireland.<\/p><h3 id=\"1974\">1974<\/h3><p>DEC became recognized as a Fortune 500 company.<\/p><h3 id=\"1978\">1978<\/h3><noscript>\n<p><img class=\"alignright wp-image-9585 size-medium\" src=\"https:\/\/digital.com\/wp-content\/uploads\/VAX_11-780_intero-225x300.jpg?x88029\" alt=\"DEC VAX 11\/780\" width=\"225\" height=\"300\" srcset=\"https:\/\/digital.com\/wp-content\/uploads\/VAX_11-780_intero-225x300.jpg 225w, https:\/\/digital.com\/wp-content\/uploads\/VAX_11-780_intero-768x1024.jpg 768w, https:\/\/digital.com\/wp-content\/uploads\/VAX_11-780_intero.jpg 1536w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><p>After widespread success with its PDP-11, DEC made the move into high-end computers and launched the Virtual Address eXtension, or VAX. This new 32-bit minicomputer (or supermini) line aimed to provide users with a wide array of computing resources that would be more affordable, powerful, and smaller than what companies like IBM could offer at the time.<\/p><h3 id=\"1982\">1982<\/h3><p>The Rainbow 100 was created and was DEC\u2019s first attempt to enter the personal computing space.<\/p><h3 id=\"1984\">1984<\/h3><p>DEC continued to stay busy during this time, regularly putting out new models of the VAX. The VAX 8600 came out in 1984 and became an instant bestseller.<\/p><h3 id=\"1985\">1985<\/h3><p>DEC registered dec.com.<\/p><h3 id=\"1988\">1988<\/h3><p>DEC was recognized as one of the premier leaders in computing when it was named the second largest computer company, just behind IBM. At the time, it had generated over $11 billion in revenue and had over 120,000 employees.<\/p><h3 id=\"1990\">1990<\/h3><p>DEC reported its first quarterly loss.<\/p><h3>1991<\/h3><p>DEC reported its first annual loss.<\/p><h3 id=\"1992\">1992<\/h3><p>Founder Ken Olsen retired.<\/p><h3 id=\"1994\">1994<\/h3><p>DEC released Alpha AXP, which was a 64-bit microprocessor created to solve the overly complicated circuit designs of its VAX computers and to ultimately speed up processing times.<\/p><h3 id=\"1995\">1995<\/h3><p>DEC launched AltaVista, one of the first ever search engines for the Internet. It became incredibly popular with users. During the first day of its launch, AltaVista received 300,000 visits. Two years later, it received 80 million hits every day.<\/p><figure id=\"attachment_9447\" aria-describedby=\"caption-attachment-9447\" class=\"wp-caption aligncenter c6\"><noscript>\n<p><img class=\"size-full wp-image-9447\" src=\"https:\/\/digital.com\/wp-content\/uploads\/4307188078_f290aecf49_o.png?x88029\" alt=\"Altavista in 1998\" width=\"569\" height=\"477\" srcset=\"https:\/\/digital.com\/wp-content\/uploads\/4307188078_f290aecf49_o.png 569w, https:\/\/digital.com\/wp-content\/uploads\/4307188078_f290aecf49_o-300x251.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript>\n<figcaption id=\"caption-attachment-9447\" class=\"wp-caption-text\">Altavista in 1998. Brent Payne\/ Flickr. CC BY 2.0<\/figcaption><\/figure><p>Although <a href=\"https:\/\/digital.com\/about\/altavista\/\">AltaVista<\/a> persisted long past the end or, more accurately, the acquisition of DEC, it was eventually sold to Yahoo in 2003. By the time 2013 rolled around, it was gone.<\/p><h3 id=\"1997\">1997<\/h3><p>Other computer companies began to make moves for the flailing DEC. Intel was the first to step in when it purchased DEC\u2019s microprocessor plant in Hudson. But even that $1.5 billion wasn\u2019t enough to save the company.<\/p><h3 id=\"1998\">1998<\/h3><p>The official end of DEC as a standalone company came in 1998 when it was acquired by Compaq for $9.6 billion. Unfortunately, Compaq didn\u2019t know what to do with DEC\u2019s overseas business, and so it was inevitable that neither DEC nor Compaq would exist for much longer. That day came in 2002 when Hewlett-Packard acquired Compaq.<\/p><h2 id=\"lesson\">What Is the Lesson Here?<\/h2><p>It\u2019s been many decades since the world was first introduced to DEC, and tech history buffs still enjoy talking about it.<\/p><p>Why? First, because it left such a lasting imprint on computing as we continue to know it, whether it was its contributions to computers, software, microchips, or even the internet itself.<\/p><p>Second, because there is an important lesson here that every computer company should carefully study if they don\u2019t want to fall victim to the same fate.<\/p><p>In a Quora thread that asked the question \u201c<a href=\"https:\/\/www.quora.com\/Why-did-Digital-Equipment-Corporation-DEC-fail\">Why did Digital Equipment Corporation fail?<\/a>\u201d it was interesting to see so many previous DEC employees and members of the MIT community speak up about what they noted during their tenure there. Almost unanimously, they supported the theory\u2014also commonly held by experts\u2014that the failure of the company ultimately fell to the leaders who were unable to foresee what was coming in personal computing and were not able to take decisive or quick enough action in time to save the company.<\/p><figure id=\"attachment_9584\" aria-describedby=\"caption-attachment-9584\" class=\"wp-caption aligncenter c5\"><noscript>\n<p><img class=\"size-full wp-image-9584\" src=\"https:\/\/digital.com\/wp-content\/uploads\/TSD_Computer_-Schreib_und_Burotechnik_DEC_ASCII_Computer-Terminal_digital_VT_100.jpg?x88029\" alt=\"VT-100\" width=\"1024\" height=\"576\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript>\n<figcaption id=\"caption-attachment-9584\" class=\"wp-caption-text\">DEC VT-100, NoRud\/ Wikipedia (CC BY-SA 4.0)<\/figcaption><\/figure><p>Perhaps the saddest thing about this is that DEC had long been considered the \u201cbest\u201d and the inspiratioan for what so many after them set out to do. It\u2019s not like it hadn\u2019t made an attempt at personal computing when they released their Rainbow 100. So, how did it fail to see that the future of the mid-market minicomputer was waning and that a full shift over to personal computers was needed in order to ensure the company\u2019s longevity?<\/p><p>According to <a href=\"http:\/\/sloanreview.mit.edu\/article\/lessons-from-ken-olsen-and-digital-equipment-corp\/\">Clayton Christensen<\/a>, a professor at Harvard Business School, it wasn\u2019t a lack of trying that took down DEC. It was the inflexibility of the business model they had so long relied upon:<\/p><p>\u201cDigital Equipment Corp. had microprocessor technology, but its business model could not profitably sell a computer for less than $50,000. The technology trapped in a high-cost business model had no impact on the world, and in fact, the world ultimately killed Digital. But IBM Corp., with the very same processors at its disposal, set up a different business model in Florida that could make money at a $2,000 price point and 20% gross margins\u2014and changed the world.\u201d<\/p><p>Regardless of why it happened or how the leadership behind DEC allowed it to happen, the legacy of this company will live on as it was willing to step up and introduce affordable and powerful computing solutions during a time when others were too afraid to.<\/p>","created_at":"2020-12-06T14:15:34+0000","updated_at":"2020-12-06T14:15:34+0000","published_at":"2017-06-14T20:29:38+0000","published_by":["Brenda Barron"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"digital.com","preview_picture":"https:\/\/digital.com\/wp-content\/uploads\/Papertape2-300x154.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1214,"title":"Command Line Interface Guidelines","url":"https:\/\/clig.dev\/","content":"<p>An <a href=\"https:\/\/github.com\/cli-guidelines\/cli-guidelines\">open-source<\/a> guide to help you write better command-line programs, taking traditional UNIX principles and updating them for the modern day.<\/p><p><strong>Aanand Prasad<\/strong><br \/>Engineer at Squarespace, co-creator of Docker Compose.<br \/><a href=\"https:\/\/twitter.com\/aanandprasad\">@aanandprasad<\/a><\/p>\n<p><strong>Ben Firshman<\/strong><br \/>Co-creator <a href=\"https:\/\/replicate.ai\/\">Replicate<\/a>, co-creator of Docker Compose.<br \/><a href=\"https:\/\/twitter.com\/bfirsh\">@bfirsh<\/a><\/p>\n<p><strong>Carl Tashian<\/strong><br \/>Developer Advocate at <a href=\"https:\/\/smallstep.com\/\">Smallstep<\/a>, first engineer at Zipcar, co-founder Trove.<br \/><a href=\"https:\/\/tashian.com\/\">tashian.com<\/a> <a href=\"https:\/\/twitter.com\/tashian\">@tashian<\/a><\/p>\n<p><strong>Eva Parish<\/strong><br \/>Technical Writer at Squarespace, O\u2019Reilly contributor.<br \/><a href=\"https:\/\/evaparish.com\/\">evaparish.com<\/a> <a href=\"https:\/\/twitter.com\/evpari\">@evpari<\/a><\/p>\n<p>Design by <a href=\"https:\/\/mhurrell.co.uk\/\">Mark Hurrell<\/a>. Thanks to Andreas Jansson for early contributions, and Andrew Reitz, Ashley Williams, Brendan Falk, Chester Ramey, Dj Walker-Morgan, Jacob Maine, James Coglan, Michael Dwan, and Steve Klabnik for reviewing drafts.<\/p>\n<p><a href=\"https:\/\/discord.gg\/EbAW5rUCkE\">Join us on Discord<\/a> if you want to discuss the guide or CLI design.<\/p>\n<h2 id=\"foreword\">Foreword<\/h2>\n<p>In the 1980s, if you wanted a personal computer to do something for you, you needed to know what to type when confronted with <code>C:\\&gt;<\/code> or <code>~$<\/code>. Help came in the form of thick, spiral-bound manuals. Error messages were opaque. There was no Stack Overflow to save you. But if you were lucky enough to have internet access, you could get help from Usenet\u2014an early internet community filled with other people who were just as frustrated as you were. They could either help you solve your problem, or at least provide some moral support and camaraderie.<\/p>\n<p>Forty years later, computers have become so much more accessible to everyone, often at the expense of low-level end user control. On many devices, there is no command-line access at all, in part because it goes against the corporate interests of walled gardens and app stores.<\/p>\n<p>Most people today don\u2019t know what the command line is, much less why they would want to bother with it. As computing pioneer Alan Kay said in <a href=\"https:\/\/www.fastcompany.com\/40435064\/what-alan-kay-thinks-about-the-iphone-and-technology-now\">a 2017 interview<\/a>, \u201cBecause people don\u2019t understand what computing is about, they think they have it in the iPhone, and that illusion is as bad as the illusion that \u2018Guitar Hero\u2019 is the same as a real guitar.\u201d<\/p>\n<p>Kay\u2019s \u201creal guitar\u201d isn\u2019t the CLI\u2014not exactly. He was talking about ways of programming computers that offer the power of the CLI and that transcend writing software in text files. There is a belief among Kay\u2019s disciples that we need to break out of a text-based local maxima that we\u2019ve been living in for decades.<\/p>\n<p>It\u2019s exciting to imagine a future where we program computers very differently. Even today, spreadsheets are by far the most popular programming language, and the no-code movement is taking off quickly as it attempts to replace some of the intense demand for talented programmers.<\/p>\n<p>Yet with its creaky, decades-old constraints and inexplicable quirks, the command line is still the most <em>versatile<\/em> corner of the computer. It lets you pull back the curtain, see what\u2019s really going on, and creatively interact with the machine at a level of sophistication and depth that GUIs cannot afford. It\u2019s available on almost any laptop, for anyone who wants to learn it. It can be used interactively, or it can be automated. And, it doesn\u2019t change as fast as other parts of the system. There is creative value in its stability.<\/p>\n<p>So, while we still have it, we should try to maximize its utility and accessibility.<\/p>\n<p>A lot has changed about how we program computers since those early days. The command line of the past was <em>machine-first<\/em>: little more than a REPL on top of a scripting platform. But as general-purpose interpreted languages have flourished, the role of the shell script has shrunk. Today\u2019s command line is <em>human-first<\/em>: a text-based UI that affords access to all kinds of tools, systems and platforms. In the past, the editor was inside the terminal\u2014today, the terminal is just as often a feature of the editor. And there\u2019s been a proliferation of <code>git<\/code>-like multi-tool commands. Commands within commands, and high-level commands that perform entire workflows rather than atomic functions.<\/p>\n<p>Inspired by traditional UNIX philosophy, driven by an interest in encouraging a more delightful and accessible CLI environment, and guided by our experiences as programmers, we decided it was time to revisit the best practices and design principles for building command-line programs.<\/p>\n<p>Long live the command line!<\/p>\n<p>This document covers both high-level design philosophy, and concrete guidelines. It\u2019s heavier on the guidelines because our philosophy as practitioners is not to philosophize too much. We believe in learning by example, so we\u2019ve provided plenty of those.<\/p>\n<p>This guide doesn\u2019t cover full-screen terminal programs like emacs and vim. Full-screen programs are niche projects\u2014very few of us will ever be in the position to design one.<\/p>\n<p>This guide is also agnostic about programming languages and tooling in general.<\/p>\n<p>Who is this guide for?<\/p>\n<ul><li>If you are creating a CLI program and you are looking for principles and concrete best practices for its UI design, this guide is for you.<\/li>\n<li>If you are a professional \u201cCLI UI designer,\u201d that\u2019s amazing\u2014we\u2019d love to learn from you.<\/li>\n<li>If you\u2019d like to avoid obvious missteps of the variety that go against 40 years of CLI design conventions, this guide is for you.<\/li>\n<li>If you want to delight people with your program\u2019s good design and helpful help, this guide is definitely for you.<\/li>\n<li>If you are creating a GUI program, this guide is not for you\u2014though you may learn some GUI anti-patterns if you decide to read it anyway. (Do GUI programmers even read, or do they just look at things?)<\/li>\n<li>If you are designing an immersive, full-screen CLI port of Minecraft, this guide isn\u2019t for you. (But we can\u2019t wait to see it!)<\/li>\n<\/ul><h2 id=\"philosophy\">Philosophy<\/h2>\n<p>These are what we consider to be the fundamental principles of good CLI design.<\/p>\n<h3 id=\"human-first-design\">Human-first design<\/h3>\n<p>Traditionally, UNIX commands were written under the assumption they were going to be used primarily by other programs. They had more in common with functions in a programming language than with graphical applications.<\/p>\n<p>Today, even though many CLI programs are used primarily (or even exclusively) by humans, a lot of their interaction design still carries the baggage of the past. It\u2019s time to shed some of this baggage: if a command is going to be used primarily by humans, it should be designed for humans first.<\/p>\n<h3 id=\"simple-parts-that-work-together\">Simple parts that work\u00a0together<\/h3>\n<p>A core tenet of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Unix_philosophy\">the original UNIX philosophy<\/a> is the idea that small, simple programs with clean interfaces can be combined to build larger systems. Rather than stuff more and more features into those programs, you make programs that are modular enough to be recombined as needed.<\/p>\n<p>In the old days, pipes and shell scripts played a crucial role in the process of composing programs together. Their role might have diminished with the rise of general-purpose interpreted languages, but they certainly haven\u2019t gone away. What\u2019s more, large-scale automation\u2014in the form of CI\/CD, orchestration and configuration management\u2014has flourished. Making programs composable is just as important as ever.<\/p>\n<p>Fortunately, the long-established conventions of the UNIX environment, designed for this exact purpose, still help us today. Standard in\/out\/err, signals, exit codes and other mechanisms ensure that different programs click together nicely. Plain, line-based text is easy to pipe between commands. JSON, a much more recent invention, affords us more structure when we need it, and lets us more easily integrate command-line tools with the web.<\/p>\n<p>Whatever software you\u2019re building, you can be absolutely certain that people will use it in ways you didn\u2019t anticipate. Your software <em>will<\/em> become a part in a larger system\u2014your only choice is over whether it will be a well-behaved part.<\/p>\n<p>Most importantly, designing for composability does not need to be at odds with designing for humans first. Much of the advice in this document is about how to achieve both.<\/p>\n<h3 id=\"consistency-across-programs\">Consistency across programs<\/h3>\n<p>The terminal\u2019s conventions are hardwired into our fingers. We had to pay an upfront cost by learning about command line syntax, flags, environment variables and so on, but it pays off in long-term efficiency\u2026 as long as programs are consistent.<\/p>\n<p>Where possible, a CLI should follow patterns that already exist. That\u2019s what makes CLIs intuitive and guessable; that\u2019s what makes users efficient.<\/p>\n<p>That being said, sometimes consistency conflicts with ease of use. For example, many long-established UNIX commands don\u2019t output much information by default, which can cause confusion or worry for people less familiar with the command line.<\/p>\n<p>When following convention would compromise a program\u2019s usability, it might be time to break with it\u2014but such a decision should be made with care.<\/p>\n<h3 id=\"saying-just-enough\">Saying (just) enough<\/h3>\n<p>The terminal is a world of pure information. You could make an argument that information is the interface\u2014and that, just like with any interface, there\u2019s often too much or too little of it.<\/p>\n<p>A command is saying too little when it hangs for several minutes and the user starts to wonder if it\u2019s broken. A command is saying too much when it dumps pages and pages of debugging output, drowning what\u2019s truly important in an ocean of loose detritus. The end result is the same: a lack of clarity, leaving the user confused and irritated.<\/p>\n<p>It can be very difficult to get this balance right, but it\u2019s absolutely crucial if software is to empower and serve its users.<\/p>\n<h3 id=\"ease-of-discovery\">Ease of discovery<\/h3>\n<p>When it comes to making functionality discoverable, GUIs have the upper hand. Everything you can do is laid out in front of you on the screen, so you can find what you need without having to learn anything, and perhaps even discover things you didn\u2019t know were possible.<\/p>\n<p>It is assumed that command-line interfaces are the opposite of this\u2014that you have to remember how to do everything. The original <a href=\"https:\/\/www.goodreads.com\/book\/show\/1087110.Macintosh_Human_Interface_Guidelines\">Macintosh Human Interface Guidelines<\/a>, published in 1992, recommend \u201cSee-and-point (instead of remember-and-type),\u201d as if you could only choose one or the other.<\/p>\n<p>These things needn\u2019t be mutually exclusive. The efficiency of using the command-line comes from remembering commands, but there\u2019s no reason the commands can\u2019t help you learn and remember.<\/p>\n<p>Discoverable CLIs have comprehensive help texts, provide lots of examples, suggest what command to run next, suggest what to do when there is an error. There are lots of ideas that can be stolen from GUIs to make CLIs easier to learn and use, even for power users.<\/p>\n<p><em>Citation: The Design of Everyday Things (Don Norman), Macintosh Human Interface Guidelines<\/em><\/p>\n<h3 id=\"conversation-as-the-norm\">Conversation as the\u00a0norm<\/h3>\n<p>GUI design, particularly in its early days, made heavy use of <em>metaphor<\/em>: desktops, files, folders, recycle bins. It made a lot of sense, because computers were still trying to bootstrap themselves into legitimacy. The ease of implementation of metaphors was one of the huge advantages GUIs wielded over CLIs. Ironically, though, the CLI has embodied an accidental metaphor all along: it\u2019s a conversation.<\/p>\n<p>Beyond the most utterly simple commands, running a program usually involves more than one invocation. Usually, this is because it\u2019s hard to get it right the first time: the user types a command, gets an error, changes the command, gets a different error, and so on, until it works. This mode of learning through repeated failure is like a conversation the user is having with the program.<\/p>\n<p>Trial-and-error isn\u2019t the only type of conversational interaction, though. There are others:<\/p>\n<ul><li>Running one command to set up a tool and then learning what commands to run to actually start using it.<\/li>\n<li>Running several commands to set up an operation, and then a final command to run it (e.g. multiple <code>git add<\/code>s, followed by a <code>git commit<\/code>).<\/li>\n<li>Exploring a system\u2014for example, doing a lot of <code>cd<\/code> and <code>ls<\/code> to get a sense of a directory structure, or <code>git log<\/code> and <code>git show<\/code> to explore the history of a file.<\/li>\n<li>Doing a dry-run of a complex operation before running it for real.<\/li>\n<\/ul><p>Acknowledging the conversational nature of command-line interaction means you can bring relevant techniques to bear on its design. You can suggest possible corrections when user input is invalid, you can make the intermediate state clear when the user is going through a multi-step process, you can confirm for them that everything looks good before they do something scary.<\/p>\n<p>The user is conversing with your software, whether you intended it or not. At worst, it\u2019s a hostile conversation which makes them feel stupid and resentful. At best, it\u2019s a pleasant exchange that speeds them on their way with newfound knowledge and a feeling of achievement.<\/p>\n<p><em>Further reading: <a href=\"https:\/\/www.nngroup.com\/articles\/anti-mac-interface\/\">The Anti-Mac User Interface (Don Gentner and Jakob Nielsen)<\/a><\/em><\/p>\n<h3 id=\"robustness-principle\">Robustness<\/h3>\n<p>Robustness is both an objective and a subjective property. Software should <em>be<\/em> robust, of course: unexpected input should be handled gracefully, operations should be idempotent where possible, and so on. But it should also <em>feel<\/em> robust.<\/p>\n<p>You want your software to feel like it isn\u2019t going to fall apart. You want it to feel immediate and responsive, as if it were a big mechanical machine, not a flimsy plastic \u201csoft switch.\u201d<\/p>\n<p>Subjective robustness requires attention to detail and thinking hard about what can go wrong. It\u2019s lots of little things: keeping the user informed about what\u2019s happening, explaining what common errors mean, not printing scary-looking stack traces.<\/p>\n<p>As a general rule, robustness can also come from keeping it simple. Lots of special cases and complex code tend to make a program fragile.<\/p>\n<h3 id=\"empathy\">Empathy<\/h3>\n<p>Command-line tools are a programmer\u2019s creative toolkit, so they should be enjoyable to use. This doesn\u2019t mean turning them into a video game, or using lots of emoji (though there\u2019s nothing inherently wrong with emoji \ud83d\ude09). It means giving the user the feeling that you are on their side, that you want them to succeed, that you have thought carefully about their problems and how to solve them.<\/p>\n<p>There\u2019s no list of actions you can take that will ensure they feel this way, although we hope that following our advice will take you some of the way there. Delighting the user means <em>exceeding their expectations<\/em> at every turn, and that starts with empathy.<\/p>\n<h3 id=\"chaos\">Chaos<\/h3>\n<p>The world of the terminal is a mess. Inconsistencies are everywhere, slowing us down and making us second-guess ourselves.<\/p>\n<p>Yet it\u2019s undeniable that this chaos has been a source of power. The terminal, like the UNIX-descended computing environment in general, places very few constraints on what you can build. In that space, all manner of invention has bloomed.<\/p>\n<p>It\u2019s ironic that this document implores you to follow existing patterns, right alongside advice that contradicts decades of command-line tradition. We\u2019re just as guilty of breaking the rules as anyone.<\/p>\n<p>The time might come when you, too, have to break the rules. Do so with intention and clarity of purpose.<\/p>\n<blockquote>\n<p>\u201cAbandon a standard when it is demonstrably harmful to productivity or user satisfaction.\u201d \u2014 Jef Raskin, <a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Humane_Interface\">The Humane Interface<\/a><\/p>\n<\/blockquote>\n<h2 id=\"guidelines\">Guidelines<\/h2>\n<p>This is a collection of specific things you can do to make your command-line program better.<\/p>\n<p>The first section contains the essential things you need to follow. Get these wrong, and your program will be either hard to use or a bad CLI citizen.<\/p>\n<p>The rest are nice-to-haves. If you have the time and energy to add these things, your program will be a lot better than the average program.<\/p>\n<p>The idea is that, if you don\u2019t want to think too hard about the design of your program, you don\u2019t have to: just follow these rules and your program will probably be good. On the other hand, if you\u2019ve thought about it and determined that a rule is wrong for your program, that\u2019s fine. (There\u2019s no central authority that will reject your program for not following arbitrary rules.)<\/p>\n<p>Also\u2014these rules aren\u2019t written in stone. If you disagree with a general rule for good reason, we hope you\u2019ll <a href=\"https:\/\/github.com\/cli-guidelines\/cli-guidelines\">propose a change<\/a>.<\/p>\n<h3 id=\"the-basics\">The Basics<\/h3>\n<p>There are a few basic rules you need to follow. Get these wrong, and your program will be either very hard to use, or flat-out broken.<\/p>\n<p><strong>Use a command-line argument parsing library where you can.<\/strong> Either your language\u2019s built-in one, or a good third-party one. They will normally handle arguments, flag parsing, help text, and even spelling suggestions in a sensible way.<\/p>\n<p>Here are some that we like:<\/p>\n<ul><li>Go: <a href=\"https:\/\/github.com\/spf13\/cobra\">Cobra<\/a>, <a href=\"https:\/\/github.com\/urfave\/cli\">cli<\/a><\/li>\n<li>Node: <a href=\"https:\/\/oclif.io\/\">oclif<\/a><\/li>\n<li>Python: <a href=\"https:\/\/click.palletsprojects.com\/\">Click<\/a>, <a href=\"https:\/\/github.com\/tiangolo\/typer\">Typer<\/a><\/li>\n<li>Ruby: <a href=\"https:\/\/ttytoolkit.org\/\">TTY<\/a><\/li>\n<\/ul><p><strong>Return zero exit code on success, non-zero on failure.<\/strong> Exit codes are how scripts determine whether a program succeeded or failed, so you should report this correctly. Map the non-zero exit codes to the most important failure modes.<\/p>\n<p><strong>Send output to <code>stdout<\/code>.<\/strong> The primary output for your command should go to <code>stdout<\/code>. Anything that is machine readable should also go to <code>stdout<\/code>\u2014this is where piping sends things by default.<\/p>\n<p><strong>Send messaging to <code>stderr<\/code>.<\/strong> Log messages, errors, and so on should all be sent to <code>stderr<\/code>. This means that when commands are piped together, these messages are displayed to the user and not fed into the next command.<\/p>\n<h3 id=\"help\">Help<\/h3>\n<p><strong>Display help text when passed no options, the <code>-h<\/code> flag, or the <code>--help<\/code> flag.<\/strong><\/p>\n<p><strong>Display a concise help text by default.<\/strong> If you can, display help by default when <code>myapp<\/code> or <code>myapp subcommand<\/code> is run. Unless your program is very simple and does something obvious by default (e.g. <code>ls<\/code>), or your program reads input interactively (e.g. <code>cat<\/code>).<\/p>\n<p>The concise help text should only include:<\/p>\n<ul><li>A description of what your program does.<\/li>\n<li>One or two example invocations.<\/li>\n<li>Descriptions of flags, unless there are lots of them.<\/li>\n<li>An instruction to pass the <code>--help<\/code> flag for more information.<\/li>\n<\/ul><p><code>jq<\/code> does this well. When you type <code>jq<\/code>, it displays an introductory description and an example, then prompts you to pass <code>jq --help<\/code> for the full listing of flags:<\/p>\n<pre>$ jq\njq - commandline JSON processor [version 1.6]\nUsage:    jq [options] &lt;jq filter&gt; [file...]\n    jq [options] --args &lt;jq filter&gt; [strings...]\n    jq [options] --jsonargs &lt;jq filter&gt; [JSON_TEXTS...]\njq is a tool for processing JSON inputs, applying the given filter to\nits JSON text inputs and producing the filter's results as JSON on\nstandard output.\nThe simplest filter is ., which copies jq's input to its output\nunmodified (except for formatting, but note that IEEE754 is used\nfor number representation internally, with all that that implies).\nFor more advanced filters see the jq(1) manpage (\"man jq\")\nand\/or https:\/\/stedolan.github.io\/jq\nExample:\n    $ echo '{\"foo\": 0}' | jq .\n    {\n        \"foo\": 0\n    }\nFor a listing of options, use jq --help.\n<\/pre>\n<p><strong>Show full help when <code>-h<\/code> and <code>--help<\/code> is passed.<\/strong> All of these should show help:<\/p>\n<pre>$ myapp\n$ myapp --help\n$ myapp -h\n<\/pre>\n<p>Ignore any other flags and arguments that are passed\u2014you should be able to add <code>-h<\/code> to the end of anything and it should show help. Don\u2019t overload <code>-h<\/code>.<\/p>\n<p>If your program is <code>git<\/code>-like, the following should also offer help:<\/p>\n<pre>$ myapp help\n$ myapp help subcommand\n$ myapp subcommand --help\n$ myapp subcommand -h\n<\/pre>\n<p><strong>Provide a support path for feedback and issues.<\/strong> A website or GitHub link in the top-level help text is common.<\/p>\n<p><strong>In help text, link to the web version of the documentation.<\/strong> If you have a specific page or anchor for a subcommand, link directly to that. This is particularly useful if there is more detailed documentation on the web, or further reading that might explain the behavior of something.<\/p>\n<p><strong>Lead with examples.<\/strong> Users tend to use examples over other forms of documentation, so show them first in the help page, particularly the common complex uses. If it helps explain what it\u2019s doing and it isn\u2019t too long, show the actual output too.<\/p>\n<p>You can tell a story with a series of examples, building your way toward complex uses.<\/p>\n<p><strong>If you\u2019ve got loads of examples, put them somewhere else,<\/strong> in a cheat sheet command or a web page. It\u2019s useful to have exhaustive, advanced examples, but you don\u2019t want to make your help text really long.<\/p>\n<p>For more complex use cases, e.g. when integrating with another tool, it might be appropriate to write a fully-fledged tutorial.<\/p>\n<p><strong>Don\u2019t bother with man pages.<\/strong> We believe that if you\u2019re following these guidelines for help and documentation, you won\u2019t need man pages. Not enough people use man pages, and they don\u2019t work on Windows. If your CLI framework and package manager make it easy to output man pages, go for it, but otherwise your time is best spent improving web docs and built-in help text.<\/p>\n<p><em>Citation: <a href=\"https:\/\/medium.com\/@jdxcode\/12-factor-cli-apps-dd3c227a0e46\">12 Factor CLI Apps<\/a>.<\/em><\/p>\n<p><strong>If your help text is long, pipe it through a pager.<\/strong> This is one useful thing that <code>man<\/code> does for you. See the advice in the \u201cOutput\u201d section below.<\/p>\n<p><strong>Display the most common flags and commands at the start of the help text.<\/strong> It\u2019s fine to have lots of flags, but if you\u2019ve got some really common ones, display them first. For example, the Git command displays the commands for getting started and the most commonly used subcommands first:<\/p>\n<pre>$ git\nusage: git [--version] [--help] [-C &lt;path&gt;] [-c &lt;name&gt;=&lt;value&gt;]\n           [--exec-path[=&lt;path&gt;]] [--html-path] [--man-path] [--info-path]\n           [-p | --paginate | -P | --no-pager] [--no-replace-objects] [--bare]\n           [--git-dir=&lt;path&gt;] [--work-tree=&lt;path&gt;] [--namespace=&lt;name&gt;]\n           &lt;command&gt; [&lt;args&gt;]\nThese are common Git commands used in various situations:\nstart a working area (see also: git help tutorial)\n   clone      Clone a repository into a new directory\n   init       Create an empty Git repository or reinitialize an existing one\nwork on the current change (see also: git help everyday)\n   add        Add file contents to the index\n   mv         Move or rename a file, a directory, or a symlink\n   reset      Reset current HEAD to the specified state\n   rm         Remove files from the working tree and from the index\nexamine the history and state (see also: git help revisions)\n   bisect     Use binary search to find the commit that introduced a bug\n   grep       Print lines matching a pattern\n   log        Show commit logs\n   show       Show various types of objects\n   status     Show the working tree status\n\u2026\n<\/pre>\n<p><strong>Use formatting in your help text.<\/strong> Bold headings make it much easier to scan. But, try to do it in a terminal-independent way so that your users aren\u2019t staring down a wall of escape characters.<\/p>\n<pre>\n<strong>$ heroku apps --help<\/strong>\nlist your apps\n<strong>USAGE<\/strong>\n  $ heroku apps\n<strong>OPTIONS<\/strong>\n  -A, --all          include apps in all teams\n  -p, --personal     list apps in personal account when a default team is set\n  -s, --space=space  filter by space\n  -t, --team=team    team to use\n  --json             output in json format\n<strong>EXAMPLES<\/strong>\n  $ heroku apps\n  === My Apps\n  example\n  example2\n  === Collaborated Apps\n  theirapp   other@owner.name\n<strong>COMMANDS<\/strong>\n  apps:create     creates a new app\n  apps:destroy    permanently destroy an app\n  apps:errors     view app errors\n  apps:favorites  list favorited apps\n  apps:info       show detailed app information\n  apps:join       add yourself to a team app\n  apps:leave      remove yourself from a team app\n  apps:lock       prevent team members from joining an app\n  apps:open       open the app in a web browser\n  apps:rename     rename an app\n  apps:stacks     show the list of available stacks\n  apps:transfer   transfer applications to another user or team\n  apps:unlock     unlock an app so any team member can join\n<\/pre>\n<p>Note: When <code>heroku apps --help<\/code> is piped through a pager, the command emits no escape characters.<\/p>\n<p><strong>If the user did something wrong and you can guess what they meant, suggest it.<\/strong> For example, <code>brew update jq<\/code> tells you that you should run <code>brew upgrade jq<\/code>.<\/p>\n<p>You can ask if they want to run the suggested command, but don\u2019t force it on them. For example:<\/p>\n<pre>$ heroku pss\n \u203a   Warning: pss is not a heroku command.\nDid you mean ps? [y\/n]:\n<\/pre>\n<p>Rather than suggesting the corrected syntax, you might be tempted to just run it for them, as if they\u2019d typed it right in the first place. Sometimes this is the right thing to do, but not always.<\/p>\n<p>Firstly, invalid input doesn\u2019t necessarily imply a simple typo\u2014it can often mean the user has made a logical mistake, or misused a shell variable. Assuming what they meant can be dangerous, especially if the resulting action modifies state.<\/p>\n<p>Secondly, be aware that if you change what the user typed, they won\u2019t learn the correct syntax. In effect, you\u2019re ruling that the way they typed it is valid and correct, and you\u2019re committing to supporting that indefinitely. Be intentional in making that decision, and document both syntaxes.<\/p>\n<p><em>Further reading: <a href=\"http:\/\/www.catb.org\/~esr\/jargon\/html\/D\/DWIM.html\">\u201cDo What I Mean\u201d<\/a><\/em><\/p>\n<p><strong>If your command is expecting to have something piped to it and <code>stdin<\/code> is an interactive terminal, display help immediately and quit.<\/strong> This means it doesn\u2019t just hang, like <code>cat<\/code>. Alternatively, you could print a log message to <code>stderr<\/code>.<\/p>\n<h3 id=\"output\">Output<\/h3>\n<p><strong>Human-readable output is paramount.<\/strong> Humans come first, machines second. The most simple and straightforward heuristic for whether a particular output stream (<code>stdout<\/code> or <code>stderr<\/code>) is being read by a human is <em>whether or not it\u2019s a TTY<\/em>. Whatever language you\u2019re using, it will have a utility or library for doing this (e.g. <a href=\"https:\/\/stackoverflow.com\/questions\/858623\/how-to-recognize-whether-a-script-is-running-on-a-tty\">Python<\/a>, <a href=\"https:\/\/nodejs.org\/api\/process.html#process_a_note_on_process_i_o\">Node<\/a>, <a href=\"https:\/\/github.com\/mattn\/go-isatty\">Go<\/a>).<\/p>\n<p><em>Further reading on <a href=\"https:\/\/unix.stackexchange.com\/a\/4132\">what a TTY is<\/a>.<\/em><\/p>\n<p><strong>Have machine-readable output where it does not impact usability.<\/strong> Streams of text is the universal interface in UNIX. Programs typically output lines of text, and programs typically expect lines of text as input, therefore you can compose multiple programs together. This is normally done to make it possible to write scripts, but it can also help the usability for humans using programs. For example, a user should be able to pipe output to <code>grep<\/code> and it should do what they expect.<\/p>\n<blockquote>\n<p>\u201cExpect the output of every program to become the input to another, as yet unknown, program.\u201d \u2014 <a href=\"https:\/\/homepage.cs.uri.edu\/~thenry\/resources\/unix_art\/ch01s06.html\">Doug McIlroy<\/a><\/p>\n<\/blockquote>\n<p><strong>If human-readable output breaks machine-readable output, use <code>--plain<\/code> to display output in plain, tabular text format for integration with tools like <code>grep<\/code> or <code>awk<\/code>.<\/strong> In some cases, you might need to output information in a different way to make it human-readable.<\/p>\n<p>For example, if you are displaying a line-based table, you might choose to split a cell into multiple lines, fitting in more information while keeping it within the width of the screen. This breaks the expected behavior of there being one piece of data per line, so you should provide a <code>--plain<\/code> flag for scripts, which disables all such manipulation and outputs one record per line.<\/p>\n<p><strong>Display output as formatted JSON if <code>--json<\/code> is passed.<\/strong> JSON allows for more structure than plain text, so it makes it much easier to output and handle complex data structures. <a href=\"https:\/\/stedolan.github.io\/jq\/\"><code>jq<\/code><\/a> is a common tool for working with JSON on the command-line, and there is now a <a href=\"https:\/\/ilya-sher.org\/2018\/04\/10\/list-of-json-tools-for-command-line\/\">whole ecosystem of tools<\/a> that output and manipulate JSON.<\/p>\n<p>It is also widely used on the web, so by using JSON as the input and output of programs, you can pipe directly to and from web services using <code>curl<\/code>.<\/p>\n<p><strong>Display output on success, but keep it brief.<\/strong> Traditionally, when nothing is wrong, UNIX commands display no output to the user. This makes sense when they\u2019re being used in scripts, but can make commands appear to be hanging or broken when used by humans. For example, <code>cp<\/code> will not print anything, even if it takes a long time.<\/p>\n<p>It\u2019s rare that printing nothing at all is the best default behavior, but it\u2019s usually best to err on the side of less.<\/p>\n<p>For instances where you do want no output (for example, when used in shell scripts), to avoid clumsy redirection of <code>stderr<\/code> to <code>\/dev\/null<\/code>, you can provide a <code>-q<\/code> option to suppress all non-essential output.<\/p>\n<p><strong>If you change state, tell the user.<\/strong> When a command changes the state of a system, it\u2019s especially valuable to explain what has just happened, so the user can model the state of the system in their head\u2014particularly if the result doesn\u2019t directly map to what the user requested.<\/p>\n<p>For example, <code>git push<\/code> tells you exactly what it is doing, and what the new state of the remote branch is:<\/p>\n<pre>$ git push\nEnumerating objects: 18, done.\nCounting objects: 100% (18\/18), done.\nDelta compression using up to 8 threads\nCompressing objects: 100% (10\/10), done.\nWriting objects: 100% (10\/10), 2.09 KiB | 2.09 MiB\/s, done.\nTotal 10 (delta 8), reused 0 (delta 0), pack-reused 0\nremote: Resolving deltas: 100% (8\/8), completed with 8 local objects.\nTo github.com:replicate\/replicate.git\n + 6c22c90...a2a5217 bfirsh\/fix-delete -&gt; bfirsh\/fix-delete\n<\/pre>\n<p><strong>Make it easy to see the current state of the system.<\/strong> If your program does a lot of complex state changes and it is not immediately visible in the filesystem, make sure you make this easy to view.<\/p>\n<p>For example, <code>git status<\/code> tells you as much information as possible about the current state of your Git repository, and some hints at how to modify the state:<\/p>\n<pre>$ git status\nOn branch bfirsh\/fix-delete\nYour branch is up to date with 'origin\/bfirsh\/fix-delete'.\nChanges not staged for commit:\n  (use \"git add &lt;file&gt;...\" to update what will be committed)\n  (use \"git restore &lt;file&gt;...\" to discard changes in working directory)\n        modified:   cli\/pkg\/cli\/rm.go\nno changes added to commit (use \"git add\" and\/or \"git commit -a\")\n<\/pre>\n<p><strong>Suggest commands the user should run.<\/strong> When several commands form a workflow, suggesting to the user commands they can run next helps them learn how to use your program and discover new functionality. For example, in the <code>git status<\/code> output above, it suggests commands you can run to modify the state you are viewing.<\/p>\n<p><strong>Actions crossing the boundary of the program\u2019s internal world should usually be explicit.<\/strong> This includes things like:<\/p>\n<ul><li>Reading or writing files that the user didn\u2019t explicitly pass as arguments (unless those files are storing internal program state, such as a cache).<\/li>\n<li>Talking to a remote server, e.g. to download a file.<\/li>\n<\/ul><p><strong>Increase information density\u2014with ASCII art!<\/strong> For example, <code>ls<\/code> shows permissions in a scannable way. When you first see it, you can ignore most of the information. Then, as you learn how it works, you pick out more patterns over time.<\/p>\n<pre>-rw-r--r-- 1 root root     68 Aug 22 23:20 resolv.conf\nlrwxrwxrwx 1 root root     13 Mar 14 20:24 rmt -&gt; \/usr\/sbin\/rmt\ndrwxr-xr-x 4 root root   4.0K Jul 20 14:51 security\ndrwxr-xr-x 2 root root   4.0K Jul 20 14:53 selinux\n-rw-r----- 1 root shadow  501 Jul 20 14:44 shadow\n-rw-r--r-- 1 root root    116 Jul 20 14:43 shells\ndrwxr-xr-x 2 root root   4.0K Jul 20 14:57 skel\n-rw-r--r-- 1 root root      0 Jul 20 14:43 subgid\n-rw-r--r-- 1 root root      0 Jul 20 14:43 subuid\n<\/pre>\n<p><strong>Use color with intention.<\/strong> For example, you might want to highlight some text so the user notices it, or use red to indicate an error. Don\u2019t overuse it\u2014if everything is a different color, then the color means nothing and only makes it harder to read.<\/p>\n<p><strong>Disable color if your program is not in a terminal or the user requested it.<\/strong> These things should disable colors:<\/p>\n<ul><li><code>stdout<\/code> or <code>stderr<\/code> is not an interactive terminal (a TTY). It\u2019s best to individually check\u2014if you\u2019re piping <code>stdout<\/code> to another program, it\u2019s still useful to get colors on <code>stderr<\/code>.<\/li>\n<li>The <code>NO_COLOR<\/code> environment variable is set.<\/li>\n<li>The <code>TERM<\/code> environment variable has the value <code>dumb<\/code>.<\/li>\n<li>The user passes the option <code>--no-color<\/code>.<\/li>\n<li>You may also want to add a <code>MYAPP_NO_COLOR<\/code> environment variable in case users want to disable color specifically for your program.<\/li>\n<\/ul><p><em>Further reading: <a href=\"https:\/\/no-color.org\/\">no-color.org<\/a>, <a href=\"https:\/\/medium.com\/@jdxcode\/12-factor-cli-apps-dd3c227a0e46\">12 Factor CLI Apps<\/a><\/em><\/p>\n<p><strong>If <code>stdout<\/code> is not an interactive terminal, don\u2019t display any animations.<\/strong> This will stop progress bars turning into Christmas trees in CI log output.<\/p>\n<p><strong>Use symbols and emoji where it makes things clearer.<\/strong> Pictures can be better than words if you need to make several things distinct, catch the user\u2019s attention, or just add a bit of character. Be careful, though\u2014it can be easy to overdo it and make your program look cluttered or feel like a toy.<\/p>\n<p>For example, <a href=\"https:\/\/github.com\/FiloSottile\/yubikey-agent\">yubikey-agent<\/a> uses emoji to add structure to the output so it isn\u2019t just a wall of text, and a \u274c to draw your attention to an important piece of information:<\/p>\n<pre class=\"language-shell-session\" data-lang=\"shell-session\">$ yubikey-agent -setup\n\ud83d\udd10 The PIN is up to 8 numbers, letters, or symbols. Not just numbers!\n\u274c The key will be lost if the PIN and PUK are locked after 3 incorrect tries.\nChoose a new PIN\/PUK: \nRepeat the PIN\/PUK: \n\ud83e\uddea Retriculating splines \u2026\n\u2705 Done! This YubiKey is secured and ready to go.\n\ud83e\udd0f When the YubiKey blinks, touch it to authorize the login.\n\ud83d\udd11 Here's your new shiny SSH public key:\necdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBCEJ\/\nUwlHnUFXgENO3ifPZd8zoSKMxESxxot4tMgvfXjmRp5G3BGrAnonncE7Aj11pn3SSYgEcrrn2sMyLGpVS0=\n\ud83d\udcad Remember: everything breaks, have a backup plan for when this YubiKey does.\n<\/pre>\n<p><strong>By default, don\u2019t output information that\u2019s only understandable by the creators of the software.<\/strong> If a piece of output serves only to help you (the developer) understand what your software is doing, it almost certainly shouldn\u2019t be displayed to normal users by default\u2014only in verbose mode.<\/p>\n<p>Invite usability feedback from outsiders and people who are new to your project. They\u2019ll help you see important issues that you are too close to the code to notice.<\/p>\n<p><strong>Don\u2019t treat <code>stderr<\/code> like a log file, at least not by default.<\/strong> Don\u2019t print log level labels (<code>ERR<\/code>, <code>WARN<\/code>, etc.) or extraneous contextual information, unless in verbose mode.<\/p>\n<p><strong>Use a pager (e.g. <code>less<\/code>) if you are outputting a lot of text.<\/strong> For example, <code>git diff<\/code> does this by default. Using a pager can be error-prone, so be careful with your implementation such that you don\u2019t make the experience worse for the user. You shouldn\u2019t use a pager if <code>stdin<\/code> or <code>stdout<\/code> is not an interactive terminal.<\/p>\n<p>A good sensible set of options to use for <code>less<\/code> is <code>less -FIRX<\/code>. This does not page if the content fills one screen, ignores case when you search, enables color and formatting, and leaves the contents on the screen when <code>less<\/code> quits.<\/p>\n<p>There might be libraries in your language that are more robust than piping to <code>less<\/code>. For example, <a href=\"https:\/\/github.com\/prompt-toolkit\/pypager\">pypager<\/a> in Python.<\/p>\n<h3 id=\"errors\">Errors<\/h3>\n<p>One of the most common reasons to consult documentation is to fix errors. If you can make errors into documentation, then this will save the user loads of time.<\/p>\n<p><strong>Catch errors and <a href=\"https:\/\/www.nngroup.com\/articles\/error-message-guidelines\/\">rewrite them for humans<\/a>.<\/strong> If you\u2019re expecting an error to happen, catch it and rewrite the error message to be useful. Think of it like a conversation, where the user has done something wrong and the program is guiding them in the right direction. Example: \u201cCan\u2019t write to file.txt. You might need to make it writable by running \u2018chmod +w file.txt\u2019.\u201d<\/p>\n<p><strong>Signal-to-noise ratio is crucial.<\/strong> The more irrelevant output you produce, the longer it\u2019s going to take the user to figure out what they did wrong. If your program produces multiple errors of the same type, consider grouping them under a single explanatory header instead of printing many similar-looking lines.<\/p>\n<p><strong>Consider where the user will look first.<\/strong> Put the most important information at the end of the output. The eye will be drawn to red text, so use it intentionally and sparingly.<\/p>\n<p><strong>If there is an unexpected or unexplainable error, provide debug and traceback information, and instructions on how to submit a bug.<\/strong> That said, don\u2019t forget about the signal-to-noise ratio: you don\u2019t want to overwhelm the user with information they don\u2019t understand. Consider writing the debug log to a file instead of printing it to the terminal.<\/p>\n<p><strong>Make it effortless to submit bug reports.<\/strong> One nice thing you can do is provide a URL and have it pre-populate as much information as possible.<\/p>\n<h3 id=\"arguments-and-flags\">Arguments and flags<\/h3>\n<p>A note on terminology:<\/p>\n<ul><li><em>Arguments<\/em>, or <em>args<\/em>, are positional parameters to a command. For example, the file paths you provide to <code>cp<\/code> are args. The order of args is often important: <code>cp foo bar<\/code> means something different from <code>cp bar foo<\/code>.<\/li>\n<li><em>Flags<\/em> are named parameters, denoted with either a hyphen and a single-letter name (<code>-r<\/code>) or a double hyphen and a multiple-letter name (<code>--recursive<\/code>). They may or may not also include a user-specified value (<code>--file foo.txt<\/code>, or <code>--file=foo.txt<\/code>). The order of flags, generally speaking, does not affect program semantics.<\/li>\n<\/ul><p><strong>Prefer flags to args.<\/strong> It\u2019s a bit more typing, but it makes it much clearer what is going on. It also makes it easier to make changes to how you accept input in the future. Sometimes when using args, it\u2019s impossible to add new input without breaking existing behavior or creating ambiguity.<\/p>\n<p><em>Citation: <a href=\"https:\/\/medium.com\/@jdxcode\/12-factor-cli-apps-dd3c227a0e46\">12 Factor CLI Apps<\/a>.<\/em><\/p>\n<p><strong>Have full-length versions of all flags.<\/strong> For example, have both <code>-h<\/code> and <code>--help<\/code>. Having the full version is useful in scripts where you want to be verbose and descriptive, and you don\u2019t have to look up the meaning of flags everywhere.<\/p>\n<p><em>Citation: <a href=\"https:\/\/www.gnu.org\/prep\/standards\/html_node\/Command_002dLine-Interfaces.html\">GNU Coding Standards<\/a>.<\/em><\/p>\n<p><strong>Only use one-letter flags for commonly used flags,<\/strong> particularly at the top-level when using subcommands. That way you don\u2019t \u201cpollute\u201d your namespace of short flags, forcing you to use convoluted letters and cases for flags you add in the future.<\/p>\n<p><strong>Multiple arguments are fine for simple actions against multiple files.<\/strong> For example, <code>rm file1.txt file2.txt file3.txt<\/code>. This also makes it work with globbing: <code>rm *.txt<\/code>.<\/p>\n<p><strong>If you\u2019ve got two or more arguments for different things, you\u2019re probably doing something wrong.<\/strong> The exception is a common, primary action, where the brevity is worth memorizing. For example, <code>cp &lt;source&gt; &lt;destination&gt;<\/code>.<\/p>\n<p><em>Citation: <a href=\"https:\/\/medium.com\/@jdxcode\/12-factor-cli-apps-dd3c227a0e46\">12 Factor CLI Apps<\/a>.<\/em><\/p>\n<p><strong>Use standard names for flags, if there is a standard.<\/strong> If another commonly used command uses a flag name, it\u2019s best to follow that existing pattern. That way, a user doesn\u2019t have to remember two different options (and which command it applies to), and users can even guess an option without having to look at the help text.<\/p>\n<p>Here\u2019s a list of commonly used options:<\/p>\n<ul><li><code>-a<\/code>, <code>--all<\/code>: All. For example, <code>ps<\/code>, <code>fetchmail<\/code>.<\/li>\n<li><code>-d<\/code>, <code>--debug<\/code>: Show debugging output.<\/li>\n<li><code>-f<\/code>, <code>--force<\/code>: Force. For example, <code>rm -f<\/code> will force the removal of files, even if it thinks it does not have permission to do it. This is also useful for commands which are doing something destructive that usually require user confirmation, but you want to force it to do that destructive action in a script.<\/li>\n<li><code>--json<\/code>: Display JSON output. See the <a href=\"#output\">output<\/a> section.<\/li>\n<li><code>-h<\/code>, <code>--help<\/code>: Help. This should only mean help. See the <a href=\"#help\">help<\/a> section.<\/li>\n<li><code>--no-input<\/code>: See the <a href=\"#interactivity\">interactivity<\/a> section.<\/li>\n<li><code>-o<\/code>, <code>--output<\/code>: Output file. For example, <code>sort<\/code>, <code>gcc<\/code>.<\/li>\n<li><code>-p<\/code>, <code>--port<\/code>: Port. For example, <code>psql<\/code>, <code>ssh<\/code>.<\/li>\n<li><code>-q<\/code>, <code>--quiet<\/code>: Quiet. Display less output. This is particularly useful when displaying output for humans that you might want to hide when running in a script.<\/li>\n<li><code>-u<\/code>, <code>--user<\/code>: User. For example, <code>ps<\/code>, <code>ssh<\/code>.<\/li>\n<li><code>--version<\/code>: Version.<\/li>\n<li><code>-v<\/code>: This can often mean either verbose or version. You might want to use <code>-d<\/code> for verbose and this for version, or for nothing to avoid confusion.<\/li>\n<\/ul><p><strong>Make the default the right thing for most users.<\/strong> Making things configurable is good, but most users are not going to find the right flag and remember to use it all the time (or alias it). If it\u2019s not the default, you\u2019re making the experience worse for most of your users.<\/p>\n<p>For example, <code>ls<\/code> has terse default output to optimize for scripts and other historical reasons, but if it were designed today, it would probably default to <code>ls -lhFGT<\/code>.<\/p>\n<p><strong>Prompt for user input.<\/strong> If a user doesn\u2019t pass an argument or flag, prompt for it. (See also: <a href=\"#interactivity\">Interactivity<\/a>)<\/p>\n<p><strong>Never <em>require<\/em> a prompt.<\/strong> Always provide a way of passing input with flags or arguments. If <code>stdin<\/code> is not an interactive terminal, skip prompting and just require those flags\/args.<\/p>\n<p><strong>Confirm before doing anything dangerous.<\/strong> A common convention is to prompt for the user to type <code>y<\/code> or <code>yes<\/code> if running interactively, or requiring them to pass <code>-f<\/code> or <code>--force<\/code> otherwise.<\/p>\n<p>\u201cDangerous\u201d is a subjective term, and there are differing levels of danger:<\/p>\n<ul><li><strong>Mild:<\/strong> A small, local change such as deleting a file. You might want to prompt for confirmation, you might not. For example, if the user is explicitly running a command called something like \u201cdelete,\u201d you probably don\u2019t need to ask.<\/li>\n<li><strong>Moderate:<\/strong> A bigger local change like deleting a directory, a remote change like deleting a resource of some kind, or a complex bulk modification that can\u2019t be easily undone. You usually want to prompt for confirmation here. Consider giving the user a way to \u201cdry run\u201d the operation so they can see what\u2019ll happen before they commit to it.<\/li>\n<li><strong>Severe:<\/strong> Deleting something complex, like an entire remote application or server. You don\u2019t just want to prompt for confirmation here\u2014you want to make it hard to confirm by accident. Consider asking them to type something non-trivial such as the name of the thing they\u2019re deleting. Let them alternatively pass a flag such as <code>--confirm=\"name-of-thing\"<\/code>, so it\u2019s still scriptable.<\/li>\n<\/ul><p>Consider whether there are non-obvious ways to accidentally destroy things. For example, imagine a situation where changing a number in a configuration file from 10 to 1 means that 9 things will be implicitly deleted\u2014this should be considered a severe risk, and should be difficult to do by accident.<\/p>\n<p><strong>If input or output is a file, support <code>-<\/code> to read from <code>stdin<\/code> or write to <code>stdout<\/code>.<\/strong> This lets the output of another command be the input of your command and vice versa, without using a temporary file. For example, <code>tar<\/code> can extract files from <code>stdin<\/code>:<\/p>\n<pre>$ curl https:\/\/example.com\/something.tar.gz | tar xvf -\n<\/pre>\n<p><strong>If a flag can accept an optional value, allow a special word like \u201cnone.\u201d<\/strong> For example, <code>ssh -F<\/code> takes an optional filename of an alternative <code>ssh_config<\/code> file, and <code>ssh -F none<\/code> runs SSH with no config file. Don\u2019t just use a blank value\u2014this can make it ambiguous whether arguments are flag values or arguments.<\/p>\n<p><strong>If possible, make arguments, flags and subcommands order-independent.<\/strong> A lot of CLIs, especially those with subcommands, have unspoken rules on where you can put various arguments. For example a command might have a <code>--foo<\/code> flag that only works if you put it before the subcommand:<\/p>\n<pre>mycmd --foo=1 subcmd\nworks\n$ mycmd subcmd --foo=1\nunknown flag: --foo\n<\/pre>\n<p>This can be very confusing for the user\u2014especially given that one of the most common things users do when trying to get a command to work is to hit the up arrow to get the last invocation, stick another option on the end, and run it again. If possible, try to make both forms equivalent, although you might run up against the limitations of your argument parser.<\/p>\n<p><strong>Allow sensitive argument values to be passed in via files.<\/strong> Let\u2019s say your command takes a secret via a <code>--password<\/code> argument. A raw <code>--password<\/code> argument will leak the secret into <code>ps<\/code> output and potentially shell history. It\u2019s easy to misuse. Consider allowing secrets only via files, e.g. with a <code>--password-file<\/code> argument. A <code>--password-file<\/code> argument allows a secret to be passed in discreetly, in a wide variety of contexts.<\/p>\n<p>(One could read a password file in Bash by using <code>--password $(&lt; password.txt)<\/code>. Unfortunately, not every context in which a command is run will have access to magical shell substitutions. For example, <code>systemd<\/code> service definitions, <code>exec<\/code> system calls, and some <code>Dockerfile<\/code> command forms do not support the substitutions available in most shells.)<\/p>\n<h3 id=\"interactivity\">Interactivity<\/h3>\n<p><strong>Only use prompts or interactive elements if <code>stdin<\/code> is an interactive terminal (a TTY).<\/strong> This is a pretty reliable way to tell whether you\u2019re piping data into a command or whether it\u2019s being run in a script, in which case a prompt won\u2019t work and you should throw an error telling the user what flag to pass.<\/p>\n<p><strong>If <code>--no-input<\/code> is passed, don\u2019t prompt or do anything interactive.<\/strong> This allows users an explicit way to disable all prompts in commands. If the command requires input, fail and tell the user how to pass the information as a flag.<\/p>\n<p><strong>If you\u2019re prompting for a password, don\u2019t print it as the user types.<\/strong> This is done by turning off echo in the terminal. Your language should have helpers for this.<\/p>\n<p><strong>Let the user escape.<\/strong> Make it clear how to get out. (Don\u2019t do what vim does.) If your program hangs on network I\/O etc, always make Ctrl-C still work. If it\u2019s a wrapper around program execution where Ctrl-C can\u2019t quit (SSH, tmux, telnet, etc), make it clear how to do that. For example, SSH allows escape sequences with the <code>~<\/code> escape character.<\/p>\n<h3 id=\"subcommands\">Subcommands<\/h3>\n<p>If you\u2019ve got a tool that\u2019s sufficiently complex, you can reduce its complexity by making a set of subcommands. If you have several tools that are very closely related, you can make them easier to use and discover by combining them into a single command (for example, RCS vs. Git).<\/p>\n<p>They\u2019re useful for sharing stuff\u2014global flags, help text, configuration, storage mechanisms.<\/p>\n<p><strong>Be consistent across subcommands.<\/strong> Use the same flag names for the same things, have similar output formatting, etc.<\/p>\n<p><strong>Use consistent names for multiple levels of subcommand.<\/strong> If a complex piece of software has lots of objects and operations that can be performed on those objects, it is a common pattern to use two levels of subcommand for this, where one is a noun and one is a verb. For example, <code>docker container create<\/code>. Be consistent with the verbs you use across different types of objects.<\/p>\n<p>Either <code>noun verb<\/code> or <code>verb noun<\/code> ordering works, but <code>noun verb<\/code> seems to be more common.<\/p>\n<p><em>Further reading: <a href=\"https:\/\/uxdesign.cc\/user-experience-clis-and-breaking-the-world-baed8709244f\">User experience, CLIs, and breaking the world, by John Starich<\/a>.<\/em><\/p>\n<p><strong>Don\u2019t have ambiguous or similarly-named commands.<\/strong> For example, having two subcommands called \u201cupdate\u201d and \u201cupgrade\u201d is quite confusing. You might want to use different words, or disambiguate with extra words.<\/p>\n<h3 id=\"robustness-guidelines\">Robustness<\/h3>\n<p><strong>Validate user input.<\/strong> Everywhere your program accepts data from the user, it will eventually be given bad data. Check early and bail out before anything bad happens, and <a href=\"#errors\">make the errors understandable<\/a>.<\/p>\n<p><strong>Responsive is more important than fast.<\/strong> Print something to the user in &lt;100ms. If you\u2019re making a network request, print something before you do it so it doesn\u2019t hang and look broken.<\/p>\n<p><strong>Show progress if something takes a long time.<\/strong> If your program displays no output for a while, it will look broken. A good spinner or progress indicator can make a program appear to be faster than it is.<\/p>\n<p>Ubuntu 20.04 has a nice progress bar that sticks to the bottom of the terminal.<\/p>\n<p>If the progress bar gets stuck in one place for a long time, the user won\u2019t know if stuff is still happening or if the program\u2019s crashed. It\u2019s good to show estimated time remaining, or even just have an animated component, to reassure them that you\u2019re still working on it.<\/p>\n<p>There are many good libraries for generating progress bars. For example, <a href=\"https:\/\/github.com\/tqdm\/tqdm\">tqdm<\/a> for Python, <a href=\"https:\/\/github.com\/schollz\/progressbar\">schollz\/progressbar<\/a> for Go, and <a href=\"https:\/\/github.com\/visionmedia\/node-progress\">node-progress<\/a> for Node.js.<\/p>\n<p><strong>Do stuff in parallel where you can, but be thoughtful about it.<\/strong> It\u2019s already difficult to report progress in the shell; doing it for parallel processes is ten times harder. Make sure it\u2019s robust, and that the output isn\u2019t confusingly interleaved. If you can use a library, do so\u2014this is code you don\u2019t want to write yourself. Libraries like <a href=\"https:\/\/github.com\/tqdm\/tqdm\">tqdm<\/a> for Python and <a href=\"https:\/\/github.com\/schollz\/progressbar\">schollz\/progressbar<\/a> for Go support multiple progress bars natively.<\/p>\n<p>The upside is that it can be a huge usability gain. For example, <code>docker pull<\/code>\u2019s multiple progress bars offer crucial insight into what\u2019s going on.<\/p>\n<p>One thing to be aware of: hiding logs behind progress bars when things go <em>well<\/em> makes it much easier for the user to understand what\u2019s going on, but if there is an error, make sure you print out the logs. Otherwise, it will be very hard to debug.<\/p>\n<p><strong>Make things time out.<\/strong> Allow network timeouts to be configured, and have a reasonable default so it doesn\u2019t hang forever.<\/p>\n<p><strong>Make it idempotent.<\/strong> If the program fails for some transient reason (e.g. the internet connection went down), you should be able to hit <code>&lt;up&gt;<\/code> and <code>&lt;enter&gt;<\/code> and it should pick up from where it left off.<\/p>\n<p><strong>Make it crash-only.<\/strong> This is the next step up from idempotence. If you can avoid needing to do any cleanup after operations, or you can defer that cleanup to the next run, your program can exit immediately on failure or interruption. This makes it both more robust and more responsive.<\/p>\n<p><em>Citation: <a href=\"https:\/\/lwn.net\/Articles\/191059\/\">Crash-only software: More than meets the eye<\/a>.<\/em><\/p>\n<p><strong>People are going to misuse your program.<\/strong> Be prepared for that. They will wrap it in scripts, use it on bad internet connections, run many instances of it at once, and use it in environments you haven\u2019t tested in, with quirks you didn\u2019t anticipate. (Did you know macOS filesystems are case-insensitive but also case-preserving?)<\/p>\n<h3 id=\"future-proofing\">Future-proofing<\/h3>\n<p>In software of any kind, it\u2019s crucial that interfaces don\u2019t change without a lengthy and well-documented deprecation process. Subcommands, arguments, flags, configuration files, environment variables: these are all interfaces, and you\u2019re committing to keeping them working. (<a href=\"https:\/\/semver.org\/\">Semantic versioning<\/a> can only excuse so much change; if you\u2019re putting out a major version bump every month, it\u2019s meaningless.)<\/p>\n<p><strong>Keep changes additive where you can.<\/strong> Rather than modify the behavior of a flag in a backwards-incompatible way, maybe you can add a new flag\u2014as long as it doesn\u2019t bloat the interface too much. (See also: <a href=\"#arguments-and-flags\">Prefer flags to args<\/a>.)<\/p>\n<p><strong>Warn before you make a non-additive change.<\/strong> Eventually, you\u2019ll find that you can\u2019t avoid breaking an interface. Before you do, forewarn your users in the program itself: when they pass the flag you\u2019re looking to deprecate, tell them it\u2019s going to change soon. Make sure there\u2019s a way they can modify their usage today to make it future-proof, and tell them how to do it.<\/p>\n<p>If possible, you should detect when they\u2019ve changed their usage and not show the warning any more: now they won\u2019t notice a thing when you finally roll out the change.<\/p>\n<p><strong>Changing output for humans is usually OK.<\/strong> The only way to make an interface easy to use is to iterate on it, and if the output is considered an interface, then you can\u2019t iterate on it. Encourage your users to use <code>--plain<\/code> or <code>--json<\/code> in scripts to keep output stable (see <a href=\"#output\">Output<\/a>).<\/p>\n<p><strong>Don\u2019t have a catch-all subcommand.<\/strong> If you have a subcommand that\u2019s likely to be the most-used one, you might be tempted to let people omit it entirely for brevity\u2019s sake. For example, say you have a <code>run<\/code> command that wraps an arbitrary shell command:<\/p>\n<pre>$ mycmd run echo \"hello world\"\n<\/pre>\n<p>You could make it so that if the first argument to <code>mycmd<\/code> isn\u2019t the name of an existing subcommand, you assume the user means <code>run<\/code>, so they can just type this:<\/p>\n<pre>$ mycmd echo \"hello world\"\n<\/pre>\n<p>This has a serious drawback, though: now you can never add a subcommand named <code>echo<\/code>\u2014or <em>anything at all<\/em>\u2014without risking breaking existing usages. If there\u2019s a script out there that uses <code>mycmd echo<\/code>, it will do something entirely different after that user upgrades to the new version of your tool.<\/p>\n<p><strong>Don\u2019t allow arbitrary abbreviations of subcommands.<\/strong> For example, say your command has an <code>install<\/code> subcommand. When you added it, you wanted to save users some typing, so you allowed them to type any non-ambiguous prefix, like <code>mycmd ins<\/code>, or even just <code>mycmd i<\/code>, and have it be an alias for <code>mycmd install<\/code>. Now you\u2019re stuck: you can\u2019t add any more commands beginning with <code>i<\/code>, because there are scripts out there that assume <code>i<\/code> means <code>install<\/code>.<\/p>\n<p>There\u2019s nothing wrong with aliases\u2014saving on typing is good\u2014but they should be explicit and remain stable.<\/p>\n<p><strong>Don\u2019t create a \u201ctime bomb.\u201d<\/strong> Imagine it\u2019s 20 years from now. Will your command still run the same as it does today, or will it stop working because some external dependency on the internet has changed or is no longer maintained? The server most likely to not exist in 20 years is the one that you are maintaining right now. (But don\u2019t build in a blocking call to Google Analytics either.)<\/p>\n<h3 id=\"signals\">Signals and control characters<\/h3>\n<p><strong>If a user hits Ctrl-C (the INT signal), exit as soon as possible.<\/strong> Say something immediately, before you start clean-up. Add a timeout to any clean-up code so it can\u2019t hang forever.<\/p>\n<p><strong>If a user hits Ctrl-C during clean-up operations that might take a long time, skip them.<\/strong> Tell the user what will happen when they hit Ctrl-C again, in case it is a destructive action.<\/p>\n<p>For example, when quitting Docker Compose, you can hit Ctrl-C a second time to force your containers to stop immediately instead of shutting them down gracefully.<\/p>\n<pre>$  docker-compose up\n\u2026\n^CGracefully stopping... (press Ctrl+C again to force)\n<\/pre>\n<p>Your program should expect to be started in a situation where clean-up has not been run. (See <a href=\"https:\/\/lwn.net\/Articles\/191059\/\">Crash-only software: More than meets the eye<\/a>.)<\/p>\n<h3 id=\"configuration\">Configuration<\/h3>\n<p>Command-line tools have lots of different types of configuration, and lots of different ways to supply it (flags, environment variables, project-level config files). The best way to supply each piece of configuration depends on a few factors, chief among them <em>specificity<\/em>, <em>stability<\/em> and <em>complexity<\/em>.<\/p>\n<p>Configuration generally falls into a few categories:<\/p>\n<ol><li>\n<p>Likely to vary from one invocation of the command to the next.<\/p>\n<p>Examples:<\/p>\n<ul><li>Setting the level of debugging output<\/li>\n<li>Enabling a safe mode or dry run of a program<\/li>\n<\/ul><p>Recommendation: <strong>Use <a href=\"#arguments-and-flags\">flags<\/a>.<\/strong> <a href=\"#environment-variables\">Environment variables<\/a> may or may not be useful as well.<\/p>\n<\/li>\n<li>\n<p>Generally stable from one invocation to the next, but not always. Might vary between projects. Definitely varies between different users working on the same project.<\/p>\n<p>This type of configuration is often specific to an individual computer.<\/p>\n<p>Examples:<\/p>\n<ul><li>Providing a non-default path to items needed for a program to start<\/li>\n<li>Specifying how or whether color should appear in output<\/li>\n<li>Specifying an HTTP proxy server to route all requests through<\/li>\n<\/ul><p>Recommendation: <strong>Use <a href=\"#arguments-and-flags\">flags<\/a> and probably <a href=\"#environment-variables\">environment variables<\/a> too.<\/strong> Users may want to set the variables in their shell profile so they apply globally, or in <code>.env<\/code> for a particular project.<\/p>\n<p>If this configuration is sufficiently complex, it may warrant a configuration file of its own, but environment variables are usually good enough.<\/p>\n<\/li>\n<li>\n<p>Stable within a project, for all users.<\/p>\n<p>This is the type of configuration that belongs in version control. Files like <code>Makefile<\/code>, <code>package.json<\/code> and <code>docker-compose.yml<\/code> are all examples of this.<\/p>\n<p>Recommendation: <strong>Use a command-specific, version-controlled file.<\/strong><\/p>\n<\/li>\n<\/ol><p><strong>Follow the XDG-spec.<\/strong> In 2010 the X Desktop Group, now <a href=\"https:\/\/freedesktop.org\">freedesktop.org<\/a>, developed a specification for the location of base directories where config files may be located. One goal was to limit the proliferation of dotfiles in a user\u2019s home directory by supporting a general-purpose <code>~\/.config<\/code> folder. The XDG Base Directory Specification (<a href=\"https:\/\/specifications.freedesktop.org\/basedir-spec\/basedir-spec-latest.html\">full spec<\/a>, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/XDG_Base_Directory#Specification\">summary<\/a>) is supported by yarn, fish, wireshark, emacs, neovim, tmux, and many other projects you know and love.<\/p>\n<p><strong>If you automatically modify configuration that is not your program\u2019s, ask the user for consent and tell them exactly what you\u2019re doing.<\/strong> Prefer creating a new config file (e.g. <code>\/etc\/cron.d\/myapp<\/code>) rather than appending to an existing config file (e.g. <code>\/etc\/crontab<\/code>). If you have to append or modify to a system-wide config file, use a dated comment in that file to delineate your additions.<\/p>\n<p><strong>Apply configuration parameters in order of precedence.<\/strong> Here is the precedence for config parameters, from highest to lowest:<\/p>\n<ul><li>Flags<\/li>\n<li>The running shell\u2019s environment variables<\/li>\n<li>Project-level configuration (eg. <code>.env<\/code>)<\/li>\n<li>User-level configuration<\/li>\n<li>System wide configuration<\/li>\n<\/ul><h3 id=\"environment-variables\">Environment variables<\/h3>\n<p><strong>Environment variables are for behavior that <em>varies with the context<\/em> in which a command is run.<\/strong> The \u201cenvironment\u201d of an environment variable is the terminal session\u2014the context in which the command is running. So, an env var might change each time a command runs, or between terminal sessions on one machine, or between instantiations of one project across several machines.<\/p>\n<p>Environment variables may duplicate the functionality of flags or configuration parameters, or they may be distinct from those things. See <a href=\"#configuration\">Configuration<\/a> for a breakdown of common types of configuration and recommendations on when environment variables are most appropriate.<\/p>\n<p><strong>For maximum portability, environment variable names must only contain letters, numbers, and underscores (and mustn\u2019t start with a number).<\/strong> Which means <code>O_O<\/code> and <code>OwO<\/code> are the only emoticons that are also valid environment variable names.<\/p>\n<p><strong>Aim for single-line environment variable values.<\/strong> While multi-line values are possible, they create usability issues with the <code>env<\/code> command.<\/p>\n<p><strong>Avoid commandeering widely used names.<\/strong> Here\u2019s a <a href=\"https:\/\/pubs.opengroup.org\/onlinepubs\/009695399\/basedefs\/xbd_chap08.html\">list of POSIX standard env vars<\/a>.<\/p>\n<p><strong>Check general-purpose environment variables for configuration values when possible:<\/strong><\/p>\n<ul><li><code>NO_COLOR<\/code>, to disable color (see <a href=\"#output\">Output<\/a>).<\/li>\n<li><code>DEBUG<\/code>, to enable more verbose output.<\/li>\n<li><code>EDITOR<\/code>, if you need to prompt the user to edit a file or input more than a single line.<\/li>\n<li><code>HTTP_PROXY<\/code>, <code>HTTPS_PROXY<\/code>, <code>ALL_PROXY<\/code> and <code>NO_PROXY<\/code>, if you\u2019re going to perform network operations. (The HTTP library you\u2019re using might already check for these.)<\/li>\n<li><code>SHELL<\/code>, if you need to open up an interactive session of the user\u2019s preferred shell. (If you need to execute a shell script, use a specific interpreter like <code>\/bin\/sh<\/code>)<\/li>\n<li><code>TERM<\/code>, <code>TERMINFO<\/code> and <code>TERMCAP<\/code>, if you\u2019re going to use terminal-specific escape sequences.<\/li>\n<li><code>TMPDIR<\/code>, if you\u2019re going to create temporary files.<\/li>\n<li><code>HOME<\/code>, for locating configuration files.<\/li>\n<li><code>PAGER<\/code>, if you want to automatically page output.<\/li>\n<li><code>LINES<\/code> and <code>COLUMNS<\/code>, for output that\u2019s dependent on screen size (e.g. tables).<\/li>\n<\/ul><p><strong>Read environment variables from <code>.env<\/code> where appropriate.<\/strong> If a command defines environment variables that are unlikely to change as long as the user is working in a particular directory, then it should also read them from a local <code>.env<\/code> file so users can configure it differently for different projects without having to specify them every time. Many languages have libraries for reading <code>.env<\/code> files (<a href=\"https:\/\/crates.io\/crates\/dotenv\">Rust<\/a>, <a href=\"https:\/\/www.npmjs.com\/package\/dotenv\">Node<\/a>, <a href=\"https:\/\/github.com\/bkeepers\/dotenv\">Ruby<\/a>).<\/p>\n<p><strong>Don\u2019t use <code>.env<\/code> as a substitute for a proper <a href=\"#configuration\">configuration file<\/a>.<\/strong> <code>.env<\/code> files have a lot of limitations:<\/p>\n<ul><li>A <code>.env<\/code> file is not commonly stored in source control<\/li>\n<li>(Therefore, any configuration stored in it has no history)<\/li>\n<li>It has only one data type: string<\/li>\n<li>It lends itself to being poorly organized<\/li>\n<li>It makes encoding issues easy to introduce<\/li>\n<li>It often contains sensitive credentials &amp; key material that would be better stored more securely<\/li>\n<\/ul><p>If it seems like these limitations will hamper usability or security, then a dedicated config file might be more appropriate.<\/p>\n<h3 id=\"naming\">Naming<\/h3>\n<p>The name of your program is particularly important on the CLI: your users will be typing it all the time, and it needs to be easy to remember and type.<\/p>\n<p><strong>Make it a simple, memorable word.<\/strong> But not too generic, or you\u2019ll step on the toes of other commands and confuse users. For example, both ImageMagick and Windows used the command <code>convert<\/code>.<\/p>\n<p><strong>Use only lowercase letters, and dashes if you really need to.<\/strong> <code>curl<\/code> is a good name, <code>DownloadURL<\/code> is not.<\/p>\n<p><strong>Keep it short.<\/strong> Users will be typing it all the time. Don\u2019t make it <em>too<\/em> short: the very shortest commands are best reserved for the common utilities used all the time, such as <code>cd<\/code>, <code>ls<\/code>, <code>ps<\/code>.<\/p>\n<p><strong>Make it easy to type.<\/strong> Some words flow across the QWERTY keyboard much more easily than others, and it\u2019s not just about brevity. <code>plum<\/code> may be short but it\u2019s an awkward, angular dance. <code>apple<\/code> trips you up with the double letter. <code>orange<\/code> is longer than both, but flows much better.<\/p>\n<p><em>Further reading: <a href=\"https:\/\/smallstep.com\/blog\/the-poetics-of-cli-command-names\/\">The Poetics of CLI Command Names<\/a><\/em><\/p>\n<h3 id=\"distribution\">Distribution<\/h3>\n<p><strong>If possible, distribute as a single binary.<\/strong> If your language doesn\u2019t compile to binary executables as standard, see if it has something like <a href=\"https:\/\/www.pyinstaller.org\/\">PyInstaller<\/a>. If you really can\u2019t distribute as a single binary, use the platform\u2019s native package installer so you aren\u2019t scattering things on disk that can\u2019t easily be removed. Tread lightly on the user\u2019s computer.<\/p>\n<p>If you\u2019re making a language-specific tool, such as a code linter, then this rule doesn\u2019t apply\u2014it\u2019s safe to assume the user has an interpreter for that language installed on their computer.<\/p>\n<p><strong>Make it easy to uninstall.<\/strong> If it needs instructions, put them at the bottom of the install instructions\u2014one of the most common times people want to uninstall software is right after installing it.<\/p>\n<h3 id=\"analytics\">Analytics<\/h3>\n<p>Usage metrics can be helpful to understand how users are using your program, how to make it better, and where to focus effort. But, unlike websites, users of the command-line expect to be in control of their environment, and it is surprising when programs do things in the background without telling them.<\/p>\n<p><strong>Do not phone home usage or crash data without consent.<\/strong> Users will find out, and they will be angry. Be very explicit about what you collect, why you collect it, how anonymous it is and how you go about anonymizing it, and how long you retain it for.<\/p>\n<p>Ideally, ask users whether they want to contribute data (\u201copt-in\u201d). If you choose to do it by default (\u201copt-out\u201d), then clearly tell users about it on your website or first run, and make it easy to disable.<\/p>\n<p>Examples of projects that collect usage statistics:<\/p>\n<ul><li>Angular.js <a href=\"https:\/\/angular.io\/analytics\">collects detailed analytics using Google Analytics<\/a>, in the name of feature prioritization. You have to explicitly opt in. You can change the tracking ID to point to your own Google Analytics property if you want to track Angular usage inside your organization.<\/li>\n<li>Homebrew sends metrics to Google Analytics and has <a href=\"https:\/\/docs.brew.sh\/Analytics\">a nice FAQ<\/a> detailing their practices.<\/li>\n<li>Next.js <a href=\"https:\/\/nextjs.org\/telemetry\">collects anonymized usage statistics<\/a> and is enabled by default.<\/li>\n<\/ul><p><strong>Consider alternatives to collecting analytics.<\/strong><\/p>\n<ul><li>Instrument your web docs. If you want to know how people are using your CLI tool, make a set of docs around the use cases you\u2019d like to understand best, and see how they perform over time. Look at what people search for within your docs.<\/li>\n<li>Instrument your downloads. This can be a rough metric to understand usage and what operating systems your users are running.<\/li>\n<li>Talk to your users. Reach out and ask people how they\u2019re using your tool. Encourage feedback and feature requests in your docs and repos, and try to draw out more context from those who submit feedback.<\/li>\n<\/ul><p><em>Further reading: <a href=\"https:\/\/opensource.guide\/metrics\/\">Open Source Metrics<\/a><\/em><\/p>\n<h2 id=\"further-reading\">Further reading<\/h2>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Unix_Programming_Environment\">The Unix Programming Environment<\/a>, Brian W. Kernighan and Rob Pike<\/li>\n<li><a href=\"https:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/basedefs\/V1_chap12.html\">POSIX Utility Conventions<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/prep\/standards\/html_node\/Program-Behavior.html\">Program Behavior for All Programs<\/a>, GNU Coding Standards<\/li>\n<li><a href=\"https:\/\/medium.com\/@jdxcode\/12-factor-cli-apps-dd3c227a0e46\">12 Factor CLI Apps<\/a>, Jeff Dickey<\/li>\n<li><a href=\"https:\/\/devcenter.heroku.com\/articles\/cli-style-guide\">CLI Style Guide<\/a>, Heroku<\/li>\n<\/ul>","created_at":"2020-12-05T02:33:53+0000","updated_at":"2020-12-05T02:33:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":53,"domain_name":"clig.dev","preview_picture":"https:\/\/clig.dev\/CLIG-social-card.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["databases"],"is_public":false,"id":1213,"title":"Feature Casualties of Large Databases \u2014 brandur.org","url":"https:\/\/brandur.org\/large-database-casualties","content":"<p>Big data has an unfortunate tendency to get messy. A few years in, a growing database that use to be small, lean, and well-designed, has better odds than not of becoming something large, bloated, and with best practices tossed aside and now considered unsalvageable.<\/p><p>There\u2019s a few common reasons that this happens, some better than others:<\/p><ul><li><strong>Technological limitation:<\/strong> The underlying tech doesn\u2019t support the scale. Say transactions or referential integrity across partitions in a sharded system.<\/li>\n<li><strong>Stability:<\/strong> Certain operations come to be considered too risky. e.g. Batch update operations that have unpredictable performance properties.<\/li>\n<li><strong>Cost\/effort:<\/strong> Doing things the right way is too hard or too expensive. e.g. Back-migrating a large amount of existing data.<\/li>\n<li><strong>Convenience:<\/strong> Similar to the previous point, poor data practice is simply by far the easiest thing to do, and gets your immediate project shipped more quickly, even if it makes future projects more difficult.<\/li>\n<\/ul><p>The loss of these features is unfortunate because they\u2019re the major reason we\u2019re using sophisticated databases in the first place. In the <a href=\"https:\/\/eng.uber.com\/schemaless-part-one-mysql-datastore\/\">most extreme cases<\/a>, advanced databases end up as nothing more than glorified key\/value stores, and the applications they power lose important foundations for reliability and correctness.<\/p><h3 id=\"transactions\" class=\"link\"><a href=\"#transactions\">Transactions<\/a><\/h3><p>ACID transactions tend to be one of the first things to go, especially since the value they provide isn\u2019t immediately obvious in a new system that\u2019s not yet seeing a lot of traffic or trouble. Between that and the facts that they add some friction in writing code quickly, and can lead to locking problems in production mean that they\u2019re often put in the chopping block early, especially when less experienced engineers are involved.<\/p><p>Losing transactions is bad news for an applications future operability, but as this subject\u2019s already covered extensively elsewhere (<a href=\"https:\/\/brandur.org\/acid\">including by me<\/a>), I won\u2019t go into depth here.<\/p><h3 id=\"referential-integrity\" class=\"link\"><a href=\"#referential-integrity\">Referential integrity<\/a><\/h3><p>Referential integrity guarantees that if a key exists somewhere in a database, then the object its referencing does as well. Foreign keys allow developers to control deletions such that if an object is being removed, but is still referenced, than that deletion should be blocked (<code>ON DELETE RESTRICT<\/code>), or, that referencing objects should be removed with it (<code>ON DELETE CASCADE<\/code>).<\/p><p>It\u2019s a powerful tool for correctness \u2013 having the database enforcing certain rules makes programs easier to write and easier to get right. <em>Not<\/em> having it tends to bleed out into code. Suddenly anytime a referenced object is loaded <em>anywhere<\/em>, the case that it came up without a result must be handled:<\/p><pre class=\"language-ruby\">user = User.load(api_key.user_id)\nif !user\n  raise ObjectNotFound, \"couldn't find user!\"\nend\n<\/pre><p>Sacrificing referential integrity is rationalized away in a number of ways. Sometimes it\u2019s due to technological limitation, sometimes due to reliability concerns (a benign-looking delete triggering an unexpectedly large cascade), but more often it\u2019s for the simple-and-not-good reason that maintaining good hygiene around foreign key relations takes discipline and work.<\/p><h3 id=\"nullable\" class=\"link\"><a href=\"#nullable\">Nullable, as far as the eye can see<\/a><\/h3><p>Relations in large databases tend to have a disproportionate number of nullable fields. This is a problem because in application code it\u2019s more difficult to work with objects that have a poorly defined schema. Every nullable field needs to be examined independently, and a fallback designed for it in case it didn\u2019t have a value. This takes time and introduces new avenues for bugs.<\/p><p>There\u2019s a few reasons that nullable-by-default is so common. The simplest is simply that nullable columns are literally the default in DDL \u2013 you\u2019ll get one unless you\u2019re really thinking about what you\u2019re doing and explicitly use <code>NOT NULL<\/code>.<\/p><p>A more common reason is that non-nullable columns often require that existing data be migrated, which is difficult, time consuming, and maybe even operationally fraught on nodes which are running very hot and which a migration unexpectedly pushes over the edge.<\/p><p>Lastly, there are often technological limitations as well. In Postgres for example, even after running a migration, taking that last step of changing a nullable column to non-nullable (<code>SET NOT NULL<\/code>) isn\u2019t safe. Postgres needs to verify that there are no nulls in the table, requiring a full table scan that blocks other operations. On a small table that\u2019ll run in an instant. On a large one, it could be the downfall of production <sup id=\"footnote-1-source\"><a href=\"#footnote-1\">1<\/a><\/sup>.<\/p><h3 id=\"indexing\" class=\"link\"><a href=\"#indexing\">Suboptimal indexing<\/a><\/h3><p>Indexes are the easiest thing in the world to work with until they\u2019re not. In a large system, they might get complicated because:<\/p><ul><li>They need to be built on multiple clusters instead of just one.<\/li>\n<li>Building them on very hot nodes gets risky as the build interferes with production operations. Internal teams may need to build tools to throttle or pause builds.<\/li>\n<li>Data gets so large that building them takes a long time.<\/li>\n<li>Data gets so large that each index is a significant non-trivial cost to store.<\/li>\n<\/ul><p>Reduced performance is the most obvious outcome, but expensive index operations can have less obvious ones too. I worked on a project recently where product design was being driven by whether options would necessitate raising a new index on a particularly enormous collection which would take weeks and cost a large figure every year in storage costs alone.<\/p><h3 id=\"restricted-apis\" class=\"link\"><a href=\"#restricted-apis\">Dangerous queries and restricted APIs<\/a><\/h3><p>SQL is the most expressive language ever for querying and manipulating data, and in the right hands, that power can make hard things easy.<\/p><p>However, the more complex the SQL statement, the more likely it is to impact production through problems like unpredictable performance or unanticipated locking. A common solution is for storage teams to simply ban non-trivial SQL wholesale, and constrain developers to a vastly simplified API \u2013 e.g. single row select, multi row select with index hint, single row update, single row delete.<\/p><pre class=\"language-ruby\"># a simplified storage API\ndef insert(data:); end\ndef delete_one(id:); end\ndef load_many(predicate:, index:, limit:); end\ndef load_one(id:); end\ndef update_one(id:, data:); end\n<\/pre><p>At a previous job, our MySQL DBA banned any database update that affected more than one row, even where it would be vastly beneficial to performance, due to concerns around them delaying replication to secondaries. This might have helped production, but had the predictable effect of reduced productivity along with some truly heinous workarounds for things that should have been trivial, and which instead resulted in considerable tech debt.<\/p><p>Where I work now, even with the comparative unexpressiveness of Mongo compared to SQL, every select in the system must be named and statically defined along with an index it expects to use. This is so that we can verify at build time that the appropriate index is already available in production.<\/p><p>There\u2019s a divide between the engineers who run big production systems and the developers who work on open-source projects in the data space, with neither group having all that much visibility into the other. Engineers who run big databases tend to adopt a nilhist outlook that every large installation inevitably trends towards a key\/value store \u2013 at a certain point, the niceties available to smaller databases must get the axe. Open-source developers don\u2019t tend to value highly the features that would help big installations.<\/p><p>I don\u2019t think the nilhist viewpoint should be the inevitable outcome, and there\u2019s cause for optimism in the development of systems like Citus, Spanner, and CockroachDB, which enable previously difficult features like cross shard transactions. We need even more movement in that direction.<\/p><p>There\u2019s a variety of possible operations-friendly features that might be possible to counteract the entropic dumbing down of large databases. Some ideas:<\/p><ul><li>Make index builds pauseable so that they can be easily throttled in emergencies.<\/li>\n<li>Make it easy to make a nullable field non-nullable, <em>not<\/em> requiring a problematic and immediate full table scan.<\/li>\n<li>A \u201cstrict\u201d SQL dialect that makes specifying fields as <code>NOT NULL<\/code> default, and specifying foreign keys required.<\/li>\n<li>A communication protocol that allows the query to signal out-of-band with a query\u2019s results that it didn\u2019t run particularly efficiently, say that it got results but wasn\u2019t able to make use of an index. This would allow a test suite to fail early by signaling the problem to a developer instead of finding out about it in production.<\/li>\n<li>A migrations framework built into the database itself that makes migrations easier and faster to write while also guaranteeing stability by allowing long-lived migration-related queries to be deprioritized and paused if necessary.<\/li>\n<\/ul><p>Ideally, we get to a place where large databases enjoy all the same benefits as smaller ones, and we all get to reap the benefits of software that gets more stable and more reliable as a result.<\/p><div class=\"footnotes\"><p><sup id=\"footnote-1\"><a href=\"#footnote-1-source\">1<\/a><\/sup> There are workarounds for Postgres like adding a <code>CHECK<\/code> constraint on null values instead of <code>NOT NULL<\/code>, but it\u2019d obviously be better if it was possible to use the native DDL.<\/p><\/div>","created_at":"2020-12-01T23:08:29+0000","updated_at":"2020-12-01T23:08:37+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"brandur.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1212,"title":"ref=as_li_ss_tl","url":"https:\/\/www.amazon.com\/dp\/B07TNMNFLW\/ref=as_li_ss_tl?ie=UTF8&linkCode=sl1&tag=consordini-20&linkId=356172158622b4a5202dade994bb2c11&language=en_US","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-12-01T04:20:47+0000","updated_at":"2020-12-01T04:20:47+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.amazon.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1211,"title":"No Config for Old Men","url":"https:\/\/datagubbe.se\/noconf\/","content":"<p><strong>...or anyone else, for that matter.<\/strong><\/p><p><em>Autumn 2020<\/em><\/p><p>Ask anyone who's really into cooking and they'll tell you how important it is to have a kitchen that's arranged just right. To someone who rarely cooks, a kitchen is probably just a place to store a few pots and a toaster, and the placement of such stuff doesn't matter much: canned soup and microwave dinners are designed for ease of preparation.<\/p><p>For the enthusiastic home cook, however, a lot of things can go wrong. Often there's only a second's notice to fix something that might ruin a perfect meal or set you back hours of hard toil. Hollandaise starting to split? Better bring out that ice cold water post haste! Garlic burning? Better chop some more right this instant, or the pasta will be overcooked and there goes the Aglio e Oglio.<\/p><p>To achieve speed, you have to know where your knives, pots, pans, spoons, whisks and other utensils are and you want to be able to arrange them so that they're easy to reach. Seasonings, spices, herbs and condiments should be within reach from the stove. Then there's the mise en place before the actual cooking begins: chopping all the vegetables, cubing the meat, slicing the bacon and so on.<\/p><p>All of this is, thankfully, easy to achieve. Even the most cramped kitchen will, after having been battle tested through a few meals, be as optimal as possible according to the cook's personal preference. The same goes for painters, carpenters, car mechanics and even office dwellers. Hammer missing from the hammer hook? For shame! Stapler not to the immediate right of the stack of post-its? Well it should be!<\/p><h3>The regulated kitchen<\/h3><p>Now imagine if you couldn't organize your kitchen to your heart's content. Not because you're lacking the funds or skills, but because some federally appointed clerk is constantly coming to inspect it.<\/p><p>\"Nah,\" he says, adjusting his clip-on tie, \"you can't put your spices there. Against regulations. All spices must be kept more than two yards from the stove at all times.\" He then goes on to explain that you'll have to call him every time you want to use the stove, just to make sure the pots you're using are compliance tested. Plus, they have to be stored in the government approved pot storage cabinet below the sink, otherwise they're not fit for kitchen use.<\/p><p>Want to change the color of the counter top? Sorry, no can do. Want to switch from glass bowls to stainless? Alas, you used to be able a few years ago, but nobody wants stainless anymore anyway, right? The salt <em>can<\/em> be put next to the stove, but it's not recommended and might change in the near future. Knives are now to be honed every Tuesday afternoon by a designated craftsman. Sure, you can postpone a few times, but eventually, you just have to live with that. Plus, there's going to be some dudes coming to inventory your pantry on a regular basis, most likely when you're in the middle of cooking something really complicated. There's no use in protesting - this is all for your own good.<\/p><p>And yes.<\/p><p>Of course this is a metaphor for computers.<\/p><h3>The curious disappearance of configuration<\/h3><p>For the casual user, some of this can <em>maybe<\/em> be convenient - or at least not annoying. If someone who rarely uses the kitchen has decided to whip up a home cooked feast, it doesn't matter that the spices are kept strangely far from the stove: they're just happy they found them. And <em>maybe<\/em> there is, for the casual user, some \"security\" in lock-in efforts such as MacOS calling home<sup><a href=\"#footnote1\">[1]<\/a><\/sup> to check if a program is allowed to run and that web browsers automatically block certain URLs.<\/p><p>Likewise, maybe a handful of confused beginners are helped by the fact that certain system settings are extremely hard to find, or that you're supposed to put all your photos in a specific directory, or that you can't decide what partition you want to install a program on, or that some indexing service starts running when you least expect it, or that not a single application gives a crap about the few color settings you're allowed to make.<\/p><p>For the power user, such things range from nuisances to something that seriously hampers productivity and creativity.<\/p><p>It used to be that whenever I got a new computer, I spent a day or two setting it up. I selected the fonts I wanted to use, I picked the colors I liked for window decorations and GUI elements, I installed my preferred tools and utilities and I organized the desktop icons and program launchers to my liking. It took a bit of time, but it was a labor of love. In times of trouble I was, if nothing else, at least the boss of my own desktop environment.<\/p><p>I don't know of any proprietary OS where I can do that anymore. Linux is, considering what's going on with the major distributions, desktop environments and UI toolkits, seemingly heading the same way. Sure, pick your own window manager, see if we care - we've got client side decorations! Want to theme your GUI? Yeah, but not in our Snap packages you won't! Want to turn off cursor blinking? Mmmmyyeeaahhh, not too sure about that. Oh, you started a GUI file manager? Hey, enjoy the ten new folders we've littered your home directory with! They all start with capital letters: designed for typing convenience in a case sensitive file system.<\/p><p>Of course I still spend a fair amount of time setting up a Windows machine, but these days it's not the joyful experience of configuring the best fonts and nicest colors and arranging the icons on the start menu in the correct order for my muscle memory. Instead it's usually a week of swearing over removed settings and working hard to find the ones that actually remain, or trying various registry hacks to circumvent seemingly unchangeable defaults. I'm working against the system instead of with it, and someone else is trying to boss me around.<\/p><h3>A better example<\/h3><p>All of this could be different. It used to be. On my <a href=\"https:\/\/datagubbe.se\/ltmag\/\">Amiga<\/a>, I could configure <em>everything<\/em>. Apart from things like fonts and colors I could draw my own mouse pointer, tiling desktop backgrounds and icons. (Yes, the system really shipped with separate little paint programs just for pointers, desktop tiles and icons.) I could customize double click speed and key repeat rates on millisecond levels. I could even control the exact position of individual icons and the size and position of every individual directory window opened.<\/p><p>Most casual users didn't care about all that, but they didn't have to. The system came with a reasonable set of defaults and when or if they grew more proficient and wanted to change something about their daily working environment, they had the option to do so.<\/p><p>This was a great approach to users. Instead of being treated like an incompetent moron and placed in a walled garden, you were entrusted and empowered. Something as simple as drawing my own mouse pointer on the Amiga was a profound and formative experience for me. As corny as it sounds, it was as if the guys who built this amazing machine put it in my hands and said, \"Hey kid, you're in charge. This computer is yours. Learn how to use it and you can make it do anything.\" It was a call for exploration and creativity.<\/p><p>Today, I can't even change the system font in Windows. I can select an \"accent color\", but most applications completely ignore it. Every program defaults to downloading into a Downloads folder and I've lost count of how many times I've changed its folder view from grouped to not grouped, only to discover it's been magically changed back the next time I browse it. Lots of settings have been removed completely while others are buried deep in strange places where a user clearly isn't really supposed to venture.<\/p><p>The problem is that there's no toggle for enabling \"advanced mode\". I'm just supposed to accept that I can no longer change simple things I've been able to configure for the past thirty years. Someone, somewhere just decided that all users have the same basic skill level and that the defaults are always acceptable.<\/p><h3>General purpose Instagram cameras<\/h3><p>I suppose the lack of configurability is a metaphor for personal computing in general: we're not buying our machines, we're renting them by way of bizarrely complex EULA:s for everything from the firmware to the OS and we're not supposed to be curious or creative, we're supposed to sit back and passively consume advertisements. The base level of creative computer use is no longer exploring programming or graphics or music, but photographing a meal someone else has prepared and then applying a predefined sepia filter to said photo. The base level of configuring a system is no longer picking some personal favorites among fonts and colors, but - maybe - selecting between dark and light mode.<\/p><p>On the flip side, more and more people also need to use computers for actually producing stuff - not least programming all those ad delivery platforms and the curiously unconfigurable operating systems they run on. But there are also armies of innocents; office workers, administrators, hotel clerks, librarians, teachers - those people now often have no choice but to strain their eyes staring at black text on bright white backgrounds, unable to select a font they find easier to read.<\/p><p>Too bad they can no longer store their spices close to the stove, but hey, who cares? Let them take one for the team. We're all swimming in ad revenue and if people learn to configure things, maybe they'll suddenly realize that the presence of those ads should be configurable as well.<\/p><br \/><hr class=\"noshade\" \/><p id=\"footnote1\"><sup>1<\/sup> In case my sarcasm isn't coming through here: No, it's not secure. The privacy and security problems this entails are huge.<\/p>","created_at":"2020-11-29T17:35:11+0000","updated_at":"2020-11-29T17:35:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"datagubbe.se","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["postgresql","search"],"is_public":false,"id":1210,"title":"Postgres full-text search is Good Enough!","url":"http:\/\/rachbelaid.com\/postgres-full-text-search-is-good-enough\/","content":"<p>When you have to build a web application, you are often asked to add search. The magnifying glass is something that we now add to wireframes without even knowing what we are going to search.<\/p><p>The search has become an important feature and we've seen a big increase in the popularity of tools like <a href=\"http:\/\/elasticsearch.org\/\">elasticsearch<\/a> and <a href=\"https:\/\/lucene.apache.org\/solr\/\">SOLR<\/a> which are both based on <a href=\"https:\/\/lucene.apache.org\/\">lucene<\/a>. They are great tools but before going down the road of Weapons of Mass <del>Destruction<\/del> Search, maybe what you need is something a bit lighter which is simply <strong>good enough<\/strong>!<\/p><p>What do you I mean by 'good enough'? I mean a search engine with the following features:<\/p><ul><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Stemming\">Stemming<\/a><\/li>\n<li>Ranking \/ Boost<\/li>\n<li>Support Multiple languages<\/li>\n<li>Fuzzy search for misspelling<\/li>\n<li>Accent support<\/li>\n<\/ul><p>Luckily PostgreSQL supports all these features.<\/p><p>This post is aimed at people who:<\/p><ul><li>use PostgreSQL and don't want to install an extra dependency for their search engine.<\/li>\n<li>use an alternative database (eg: MySQL) and have the need for better full-text search features.<\/li>\n<\/ul><p>In this post, we are going to progressively illustrate some of the full-text search features in Postgres based on the following tables and data:<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE TABLE author(\n   id SERIAL PRIMARY KEY,\n   name TEXT NOT NULL\n);\nCREATE TABLE post(\n   id SERIAL PRIMARY KEY,\n   title TEXT NOT NULL,\n   content TEXT NOT NULL,\n   author_id INT NOT NULL references author(id) \n);\nCREATE TABLE tag(\n   id SERIAL PRIMARY KEY,\n   name TEXT NOT NULL \n);\nCREATE TABLE posts_tags(\n   post_id INT NOT NULL references post(id),\n   tag_id INT NOT NULL references tag(id)\n );\nINSERT INTO author (id, name) \nVALUES (1, 'Pete Graham'), \n       (2, 'Rachid Belaid'), \n       (3, 'Robert Berry');\nINSERT INTO tag (id, name) \nVALUES (1, 'scifi'), \n       (2, 'politics'), \n       (3, 'science');\nINSERT INTO post (id, title, content, author_id) \nVALUES (1, 'Endangered species', \n        'Pandas are an endangered species', 1 ), \n       (2, 'Freedom of Speech', \n        'Freedom of speech is a necessary right', 2), \n       (3, 'Star Wars vs Star Trek', \n        'Few words from a big fan', 3);\nINSERT INTO posts_tags (post_id, tag_id) \nVALUES (1, 3), \n       (2, 2), \n       (3, 1);<\/pre><\/div><p>It's a traditional blog-like application with <code>post<\/code> objects, which have a <code>title<\/code> and <code>content<\/code>. A <code>post<\/code> is associated to an <code>author<\/code> via a foreign key. A <code>post<\/code> itself can have multiple tags.<\/p><h3>What is Full-Text Search.<\/h3><p>First, let's look at the definition:<\/p><blockquote>\n<p>In text retrieval, <strong>full-text search<\/strong> refers to techniques for searching a single computer-stored <strong>document<\/strong> or a collection in a full-text database. The full-text search is distinguished from searches based on metadata or on parts of the original texts represented in databases.<\/p>\n<p>-- <a href=\"http:\/\/en.wikipedia.org\/wiki\/Full_text_search\">Wikipedia<\/a><\/p>\n<\/blockquote><p>This definition introduces the concept of a document, which is important. When you run a search across your data, you are looking into meaningful entities for which you want to search, these are your documents! The PostgreSQL documentation explains it amazingly.<\/p><blockquote>\n<p>A document is the unit of searching in a full-text search system; for example, a magazine article or email message.<\/p>\n<p>-- <a href=\"http:\/\/www.postgresql.org\/docs\/9.3\/static\/textsearch-intro.html#TEXTSEARCH-DOCUMENT\">Postgres documentation<\/a><\/p>\n<\/blockquote><p>This document can be across multiple tables and it represents a logical entity which we want to search for.<\/p><h3>Build our document<\/h3><p>In the previous section, we introduced the concept of document. A document is not related to our table schema but to data; together these represent a meaningful object. Based on our example schema, the document is composed of:<\/p><ul><li><code>post.title<\/code><\/li>\n<li><code>post.content<\/code><\/li>\n<li><code>author.name<\/code> of the <code>post<\/code><\/li>\n<li>all <code>tag.name<\/code> associated to the <code>post<\/code><\/li>\n<\/ul><p>To create our document based on this criteria imagine this SQL query:<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT post.title || ' ' || \n       post.content || ' ' ||\n       author.name || ' ' ||\n       coalesce((string_agg(tag.name, ' ')), '') as document\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON tag.id = posts_tags.tag_id\nGROUP BY post.id, author.id;\n              document\n--------------------------------------------------\nEndangered species Pandas are an endangered species Pete Graham politics\nFreedom of Speech Freedom of speech is a necessary right Rachid Belaid politics\nStar Wars vs Star Trek Few words from a big fan Robert Berry politics\n3 rows)<\/pre><\/div><p>As we are grouping by <code>post<\/code> and <code>author<\/code>, we are using <code>string_agg()<\/code> as the aggregate function because multiple <code>tag<\/code> can be associated to a <code>post<\/code>. Even if <code>author<\/code> is a foreign key and a <code>post<\/code> cannot have more than one <code>author<\/code>, it is required to add an aggregate function for <code>author<\/code> or to add <code>author<\/code> to the <code>GROUP BY<\/code>.<\/p><p>We also used <code>coalesce()<\/code>. When a value can be <code>NULL<\/code> then it's good practice to use the <code>coalesce()<\/code> function, otherwise the concatenation will result in a <code>NULL<\/code> value too.<\/p><p>At this stage, our document is simply a long <code>string<\/code> and this doesn't help us; we need to transform it into the right format via the function <code>to_tsvector()<\/code>.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector(post.title) || \n       to_tsvector(post.content) ||\n       to_tsvector(author.name) ||\n       to_tsvector(coalesce((string_agg(tag.name, ' ')), '')) as document\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON tag.id = posts_tags.tag_id\nGROUP BY post.id, author.id;\n               document\n --------------------------------------------------\n 'endang':1,6 'graham':9 'panda':3 'pete':8 'polit':10 'speci':2,7\n 'belaid':12 'freedom':1,4  'necessari':9 'polit':13 'rachid':11 'right':10 'speech':3,6\n 'berri':13 'big':10 'fan':11 'polit':14 'robert':12 'star':1,4 'trek':5 'vs':3 'war':2 'word':7\n(3 rows)<\/pre><\/div><p>This query will return our document as <code>tsvector<\/code> which is a type suited to full-text search. Let's try to convert a simple string into a tsvector.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('Try not to become a man of success, but rather try to become a man of value');\nquery will return the following result:\n                             to_tsvector\n----------------------------------------------------------------------\n'becom':4,13 'man':6,15 'rather':10 'success':8 'tri':1,11 'valu':17\n(1 row)<\/pre><\/div><p>Something weird just happened. First there are fewer words than in the original sentence, some of the words are different (<code>try<\/code> became <code>tri<\/code>) and they are all followed by numbers. Why?<\/p><p>A <code>tsvector<\/code> value is a sorted list of distinct lexemes which are words that have been normalized to make different variants of the same word look alike. For example, normalization almost always includes folding upper-case letters to lower-case and often involves removal of suffixes (such as 's', 'es' or 'ing' in English). This allows searches to find variant forms of the same word without tediously entering all the possible variants.<\/p><p>The numbers represent the location of the lexeme in the original string. For example, \"man\" is present at position 6 and 15. Try counting the words and see for yourself.<\/p><p>By default, Postgres uses <code>'english'<\/code> as text search configuration for the function <code>to_tsvector<\/code> and it will also ignore english stopwords. That explains why the <code>tsvector<\/code> results have fewer elements than the ones in our sentence. We see later a bit more about languages and text search configuration.<\/p><h3>Querying<\/h3><p>We have seen how to build a document, but the goal here is to find the document. For running a query against a <code>tsvector<\/code> we can use the <code>@@<\/code> operator which is documented <a href=\"http:\/\/www.postgresql.org\/docs\/9.3\/static\/functions-textsearch.html#FUNCTIONS-TEXTSEARCH\">here<\/a>. Let's see some examples on how to query our document.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">&gt; select to_tsvector('If you can dream it, you can do it') @@ 'dream';\n ?column?\n----------\n t\n(1 row)\n&gt; select to_tsvector('It''s kind of fun to do the impossible') @@ 'impossible';\n ?column?\n----------\n f\n(1 row)<\/pre><\/div><p>The second query returns false because we need to build a <code>tsquery<\/code> which creates the same lexemes and, using the operator <code>@@<\/code>, casts the string into a tsquery. The following shows the difference between casting and using the function <code>to_tsquery()<\/code><\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT 'impossible'::tsquery, to_tsquery('impossible');\n   tsquery    | to_tsquery\n--------------+------------\n 'impossible' | 'imposs'\n(1 row)<\/pre><\/div><p>But in the case of 'dream' the stem is equal to the word.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT 'dream'::tsquery, to_tsquery('dream');\n   tsquery    | to_tsquery\n--------------+------------\n 'dream'      | 'dream'\n(1 row)<\/pre><\/div><p>From now on we will use <code>to_tsquery<\/code> for querying documents.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('It''s kind of fun to do the impossible') @@ to_tsquery('impossible');\n ?column?\n----------\n t\n(1 row)<\/pre><\/div><p>A tsquery value stores lexemes that are to be searched for, and combines them honoring the Boolean operators &amp; (AND), | (OR), and ! (NOT). Parentheses can be used to enforce grouping of the operators<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">&gt; SELECT to_tsvector('If the facts don't fit the theory, change the facts') @@ to_tsquery('! fact');\n ?column?\n----------\n f\n(1 row)\n&gt; SELECT to_tsvector('If the facts don''t fit the theory, change the facts') @@ to_tsquery('theory &amp; !fact');\n ?column?\n----------\n f\n(1 row)\n&gt; SELECT to_tsvector('If the facts don''t fit the theory, change the facts.') @@ to_tsquery('fiction | theory');\n ?column?\n----------\n t\n(1 row)<\/pre><\/div><p>We can also use startwith query style when using <code>:*<\/code>.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">&gt; SELECT to_tsvector('If the facts don''t fit the theory, change the facts.') @@ to_tsquery('theo:*');\n ?column?\n----------\n t\n(1 row)<\/pre><\/div><p>Now that we know how to make a full-text search query, we can come back to our initial table schema and try to query our documents.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT pid, p_title\nFROM (SELECT post.id as pid,\n             post.title as p_title,\n             to_tsvector(post.title) || \n             to_tsvector(post.content) ||\n             to_tsvector(author.name) ||\n             to_tsvector(coalesce(string_agg(tag.name, ' '))) as document\n      FROM post\n      JOIN author ON author.id = post.author_id\n      JOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\n      JOIN tag ON tag.id = posts_tags.tag_id\n      GROUP BY post.id, author.id) p_search\nWHERE p_search.document @@ to_tsquery('Endangered &amp; Species');\n pid |      p_title\n-----+--------------------\n   1 | Endangered species\n(1 row)<\/pre><\/div><p>This will find our document which contains <code>Endangered<\/code> and <code>Species<\/code> or lexemes close enough.<\/p><h3>Language support<\/h3><p>Postgres provides built-ins text search for many languages: Danish, Dutch, English, Finnish, French, German, Hungarian, Italian, Norwegian, Portuguese, Romanian, Russian, Spanish, Swedish, Turkish.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('english', 'We are running');\n to_tsvector\n-------------\n 'run':3\n(1 row)\nSELECT to_tsvector('french', 'We are running');\n        to_tsvector\n----------------------------\n 'are':2 'running':3 'we':1\n(1 row)<\/pre><\/div><p>A column name can be used to create the <code>tsvector<\/code> based on our starting model. Let's assume that <code>post<\/code> can be written in different languages and post contains a column <code>language<\/code>.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">ALTER TABLE post ADD language text NOT NULL DEFAULT('english');<\/pre><\/div><p>We can now rebuild our document to use this language column.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector(post.language::regconfig, post.title) || \n       to_tsvector(post.language::regconfig, post.content) ||\n       to_tsvector('simple', author.name) ||\n       to_tsvector('simple', coalesce((string_agg(tag.name, ' ')), '')) as document\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON tag.id = posts_tags.tag_id\nGROUP BY post.id, author.id;<\/pre><\/div><p>Without the explicit cast <code>::regconfig<\/code> the query will have generated an error:<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">ERROR:  function to_tsvector(text, text) does not exist<\/pre><\/div><p><code>regconfig<\/code> is the object identifier type which represents the text search configuration in Postgres: http:\/\/www.postgresql.org\/docs\/9.3\/static\/datatype-oid.html<\/p><p>Fow now the lexemes of our document will be built using the right language based on <code>post.language<\/code>.<\/p><p>We also used <code>simple<\/code> which is one of the built in search text configs that Postgres provides. <code>simple<\/code> doesn't ignore stopwords and doesn't try to find the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Stemming\">stem<\/a> of the word. With <code>simple<\/code> every group of characters separated by a space is a lexeme; the <code>simple<\/code> text search config is pratical for data like a person's name for which we may not want to find the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Stemming\">stem<\/a> of the word.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('simple', 'We are running');\n        to_tsvector\n----------------------------\n 'are':2 'running':3 'we':1\n(1 row)<\/pre><\/div><h3>Accented Character<\/h3><p>When you build a search engine supporting many languages you will also hit the accent problem. In many languages, accents are very important and can change the meaning of the word. Postgres ships with an extension call <code>unaccent<\/code> which is useful to unaccentuate content.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE EXTENSION unaccent;\nSELECT unaccent('\u00e8\u00e9\u00ea\u00eb');\n unaccent\n----------\n eeee\n(1 row)<\/pre><\/div><p>Let's add some accented content to our <code>post<\/code> table.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">INSERT INTO post (id, title, content, author_id, language) \nVALUES (4, 'il \u00e9tait une fois', 'il \u00e9tait une fois un h\u00f4tel ...', 2,'french')<\/pre><\/div><p>If we want to ignore accents when we build our document, then we can simply do the following:<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector(post.language, unaccent(post.title)) || \n       to_tsvector(post.language, unaccent(post.content)) ||\n       to_tsvector('simple', unaccent(author.name)) ||\n       to_tsvector('simple', unaccent(coalesce(string_agg(tag.name, ' '))))\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON author.id = post.author_id\nGROUP BY p.id<\/pre><\/div><p>That works, but it's a bit cumbersome with more room for mistakes. We can also build a new text search config with support for unaccented characters.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE TEXT SEARCH CONFIGURATION fr ( COPY = french );\nALTER TEXT SEARCH CONFIGURATION fr ALTER MAPPING\nFOR hword, hword_part, word WITH unaccent, french_stem;<\/pre><\/div><p>When we are using this new text search config, we can see the lexemes<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('french', 'il \u00e9tait une fois');\n to_tsvector\n-------------\n 'fois':4\n(1 row)\nSELECT to_tsvector('fr', 'il \u00e9tait une fois');\n    to_tsvector\n--------------------\n 'etait':2 'fois':4\n(1 row)<\/pre><\/div><p>This gives us the same result as applying <code>unaccent<\/code> first and building the <code>tsvector<\/code> from the result.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('french', unaccent('il \u00e9tait une fois'));\n    to_tsvector\n--------------------\n 'etait':2 'fois':4\n(1 row)<\/pre><\/div><p>The number of lexemes is different because <code>il \u00e9tait une<\/code> are stopwords in French. Is it an issue to keep these stop words in our document? I don't think so as <code>etait<\/code> is not really a stopword as it's misspelled.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector('fr', 'H\u00f4tel') @@ to_tsquery('hotels') as result;\n result\n--------\n t\n(1 row)<\/pre><\/div><p>If we create an unaccented search config for each language that our <code>post<\/code> can be written in and we keep this value in <code>post.language<\/code> then we can keep our previous document query.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT to_tsvector(post.language, post.title) || \n       to_tsvector(post.language, post.content) ||\n       to_tsvector('simple', author.name) ||\n       to_tsvector('simple', coalesce(string_agg(tag.name, ' ')))\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON author.id = post.author_id\nGROUP BY p.id<\/pre><\/div><p>If you need to create unaccented text search config for each language supported by Postgres then you can use this <a href=\"https:\/\/gist.github.com\/rach\/9289959\">gist<\/a><\/p><p>Our document will now likely increase in size because it can include unaccented stopwords but we query it without caring about accented characters. This can be useful e.g. for somebody with an english keyboard searching french content.<\/p><h3>Ranking<\/h3><p>When you build a search engine you want to be able to get search results ordered by relevance. The ranking of documents is based on many factors which are roughly explained in this documentation.<\/p><blockquote>\n<p>Ranking attempts to measure how relevant documents are to a particular query, so that when there are many matches the most relevant ones can be shown first. PostgreSQL provides two predefined ranking functions, which take into account lexical, proximity, and structural information; that is, they consider how often the query terms appear in the document, how close together the terms are in the document, and how important is the part of the document where they occur.<\/p>\n<p>-- <a href=\"http:\/\/www.postgresql.org\/docs\/9.3\/static\/textsearch-controls.html#TEXTSEARCH-RANKING\">PostgreSQL documentation<\/a><\/p>\n<\/blockquote><p>To order our results by relevance PostgreSQL provides a few functions but in our example we will be using only 2 of them: <code>ts_rank()<\/code> and <code>setweight()<\/code>.<\/p><p>The function <code>setweight<\/code> allows us to assign a weight value to a <code>tsvector<\/code>; the value can be <code>'A', 'B', 'C' or 'D'<\/code><\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT pid, p_title\nFROM (SELECT post.id as pid,\n             post.title as p_title,\n             setweight(to_tsvector(post.language::regconfig, post.title), 'A') || \n             setweight(to_tsvector(post.language::regconfig, post.content), 'B') ||\n             setweight(to_tsvector('simple', author.name), 'C') ||\n             setweight(to_tsvector('simple', coalesce(string_agg(tag.name, ' '))), 'B') as document\n      FROM post\n      JOIN author ON author.id = post.author_id\n      JOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\n      JOIN tag ON tag.id = posts_tags.tag_id\n      GROUP BY post.id, author.id) p_search\nWHERE p_search.document @@ to_tsquery('english', 'Endangered &amp; Species')\nORDER BY ts_rank(p_search.document, to_tsquery('english', 'Endangered &amp; Species')) DESC;<\/pre><\/div><p>In the query above, we have assigned different weights to the different fields of a document. <code>post.title<\/code> is more important than the <code>post.content<\/code> and as important as the <code>tag<\/code> associated. The least important is the <code>author.name<\/code>.<\/p><p>This means that if we were to search for the term 'Alice' a document that contains that term in its title would be returned before a document that contains the term in its content and document that with an author of that name would be returned last.<\/p><p>Based on the weights assigned to parts of our document the <code>ts_rank()<\/code> returns a floating number which represents the relevancy of our document against the query.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT ts_rank(to_tsvector('This is an example of document'), \n               to_tsquery('example | document')) as relevancy;\n relevancy\n-----------\n 0.0607927\n(1 row)\nSELECT ts_rank(to_tsvector('This is an example of document'), \n               to_tsquery('example ')) as relevancy;\n relevancy\n-----------\n 0.0607927\n(1 row)\nSELECT ts_rank(to_tsvector('This is an example of document'), \n               to_tsquery('example | unkown')) as relevancy;\n relevancy\n-----------\n 0.0303964\n(1 row)\nSELECT ts_rank(to_tsvector('This is an example of document'),\n               to_tsquery('example &amp; document')) as relevancy;\n relevancy\n-----------\n 0.0985009\n(1 row)\nSELECT ts_rank(to_tsvector('This is an example of document'), \n               to_tsquery('example &amp; unknown')) as relevancy;\n relevancy\n-----------\n 1e-20\n(1 row)<\/pre><\/div><p>However, the concept of relevancy is vague and very application-specific. Different applications might require additional information for ranking, e.g., document modification time. The built-in ranking functions such as <code>ts_rank<\/code> are only examples. You can write your own ranking functions and\/or combine their results with additional factors to fit your specific needs.<\/p><p>To illustrate the paragraph above, if we wanted to promote newer posts against older ones we could divide the ts_rank value by the age of the document +1 (to avoid dividing by zero).<\/p><h3>Optimization and Indexing<\/h3><p>Optimizing the search on one table is straight forward. PostgreSQL supports function based indexes so you can simply create a GIN index around the <code>tsvector()<\/code> function.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE INDEX idx_fts_post ON post \nUSING gin((setweight(to_tsvector(language::regconfig, title),'A') || \n       setweight(to_tsvector(language::regconfig, content), 'B')));\n-- If this throws an IMMUTABLE error then you can use this workaround\nCREATE OR REPLACE FUNCTION gin_fts_fct(title text, content text, language text) \n  RETURNS tsvector\nAS\n$BODY$\n    SELECT setweight(to_tsvector($3::regconfig, $1), 'A') || setweight(to_tsvector($3::regconfig, $1), 'B');\n$BODY$\nLANGUAGE sql\nIMMUTABLE;\nCREATE INDEX idx_fts_post ON post  USING gin(gin_fts_fct(title, content, language));<\/pre><\/div><p>GIN or GiST indexes? These two indexes could be subject of a blog post themselves. GiST can produce false matches which then requires a extra table row lookup to confirm the match. On the other hand, GIN is faster to query but bigger and slower to build.<\/p><blockquote>\n<p>As a rule of thumb, GIN indexes are best for static data because lookups are faster. For dynamic data, GiST indexes are faster to update. Specifically, GiST indexes are very good for dynamic data and fast if the number of unique words (lexemes) is under 100,000 while GIN indexes will handle 100,000+ lexemes better but are slower to update.<\/p>\n<p>-- <a href=\"http:\/\/www.postgresql.org\/docs\/9.1\/static\/textsearch-indexes.html\">Postgres doc : Chap 12 Full Text Search<\/a><\/p>\n<\/blockquote><p>For our example, we will be using GIN but the choice can be argued and you need to take your own decision based on your data.<\/p><p>We have a problem in our schema example; the document is spread across multiple tables with different weights. For a better performance it's necessary to denormalize the data via triggers or materialized view.<\/p><p>You don't always need to denormalise and in some cases you can add a function based index as we did above. Alternatively you can easily denormalise data from the same table via the postgres trigger function <code>tsvector_update_trigger(...)<\/code> or <code>tsvector_update_trigger_column(...)<\/code>. See the <a href=\"http:\/\/www.postgresql.org\/docs\/9.3\/static\/textsearch-features.html#TEXTSEARCH-UPDATE-TRIGGERS\">Postgres doc<\/a> for more detailed information.<\/p><p>If it's acceptable of having some delay before a document can be found via search query then this may be a good use case for using a Materialized View so we can build an extra index on it.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE MATERIALIZED VIEW search_index AS \nSELECT post.id,\n       post.title,\n       setweight(to_tsvector(post.language::regconfig, post.title), 'A') || \n       setweight(to_tsvector(post.language::regconfig, post.content), 'B') ||\n       setweight(to_tsvector('simple', author.name), 'C') ||\n       setweight(to_tsvector('simple', coalesce(string_agg(tag.name, ' '))), 'A') as document\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON tag.id = posts_tags.tag_id\nGROUP BY post.id, author.id<\/pre><\/div><p>Then reindexing the search engine will be as simple as periodically running <code>REFRESH MATERIALIZED VIEW search_index;<\/code>.<\/p><p>We can now add an index on the materialized view.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE INDEX idx_fts_search ON search_index USING gin(document);<\/pre><\/div><p>And querying will become much simpler too.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT id as post_id, title\nFROM search_index\nWHERE document @@ to_tsquery('english', 'Endangered &amp; Species')\nORDER BY ts_rank(p_search.document, to_tsquery('english', 'Endangered &amp; Species')) DESC;<\/pre><\/div><p>If you cannot afford delay then you may have to investigate the alternative method using triggers.<\/p><p>There is not one way to build your document store; it will depend on what comprises your document: single table, multiple table, multiple languages, amount of data ...<\/p><p><a href=\"http:\/\/thoughtbot.com\/\">Thoughtbot.com<\/a> published a good article on <a href=\"http:\/\/robots.thoughtbot.com\/implementing-multi-table-full-text-search-with-postgres\">\"Implementing Multi-Table Full Text Search with Postgres in Rails\"<\/a> which I advise reading.<\/p><h3>Mispelling<\/h3><p>PostgreSQL comes with a very useful extenstion called <code>pg_trgm<\/code>. See <a href=\"http:\/\/www.postgresql.org\/docs\/9.3\/static\/pgtrgm.html\">pg_trgm doc<\/a>.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE EXTENSION pg_trgm;<\/pre><\/div><p>This provides support for trigram which is a <a href=\"http:\/\/en.wikipedia.org\/wiki\/N-gram\">N-gram with<\/a> <code>N == 3<\/code>. N-grams are useful because they allow finding strings with similar characters and, in essence, that's what a misspelling is - a word that is similar but not quite right.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT similarity('Something', 'something');\n similarity\n------------\n     1\n(1 row)\nSELECT similarity('Something', 'samething');\n similarity\n------------\n  0.538462\n(1 row)\nSELECT similarity('Something', 'unrelated');\n similarity\n------------\n     0\n(1 row)\nSELECT similarity('Something', 'everything');\n similarity                                          \n------------\n   0.235294\n(1 row)\nSELECT similarity('Something', 'omething');\n similarity\n------------\n   0.583333\n(1 row)<\/pre><\/div><p>With the examples above you can see that similarity returns a float to represent the similarity between two strings. To detect misspelling is then a matter of collecting the lexemes used by our documents and comparing the similarities with our search input. I found that 0.5 is a good number to test the similarity of misspelling. First we need to create this list of unique lexemes used by our documents.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE MATERIALIZED VIEW unique_lexeme AS\nSELECT word FROM ts_stat(\n'SELECT to_tsvector('simple', post.title) || \n    to_tsvector('simple', post.content) ||\n    to_tsvector('simple', author.name) ||\n    to_tsvector('simple', coalesce(string_agg(tag.name, ' ')))\nFROM post\nJOIN author ON author.id = post.author_id\nJOIN posts_tags ON posts_tags.post_id = posts_tags.tag_id\nJOIN tag ON tag.id = posts_tags.tag_id\nGROUP BY post.id, author.id');<\/pre><\/div><p>The query above builds a view with one column called <code>word<\/code> from all the unique lexemes of our documents. We used <code>simple<\/code> because our content can be in multiple languages. Once we create this materialized view we need to add an index to make the similarity query faster.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE INDEX words_idx ON search_words USING gin(word gin_trgm_ops);<\/pre><\/div><p>Luckily unique lexemes used in a search engine is not something that will change rapidly, so probably we won't have to refresh the materialized view too often via:<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">REFRESH MATERIALIZED VIEW unique_lexeme;<\/pre><\/div><p>Once we have built this table finding the closest match is very simple.<\/p><div class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT word \nWHERE similarity(word, 'samething') &gt; 0.5 \nORDER BY word &lt;-&gt; 'samething'\nLIMIT 1;<\/pre><\/div><p>This query returns a lexeme which is similar enough (<code>&gt;0.5<\/code>) to the search input <code>samething<\/code> ordered by the closest first. The operator <code>&lt;-&gt;<\/code> returns the \"distance\" between the arguments, that is one minus the <code>similarity()<\/code> value.<\/p><p>When you decide to handle misspelling in your search you may want to not look for misspellings on every query. Instead, you could query for misspellings only when the search returns no results and use the results of that query to provide some suggestions to your user. It is also possible that your data may contain misspellings if it comes from some source of informal communication such as a social network in which case you may obtain good results by appending the similar lexeme to your <code>tsquery<\/code>.<\/p><p><a href=\"http:\/\/bartlettpublishing.com\/site\/bartpub\/blog\/3\/entry\/350\">\"Super Fuzzy Searching on PostgreSQL\"<\/a> is a good reference article about the use of trigrams for using misspellings and search with Postgres.<\/p><p>In my use case the unique lexemes table has never been bigger than 2000 rows but from my understanding if you have more 1M unique lexemes used across your document then you may be meet performance issues with this technique.<\/p><h3>About MySQL and RDS<\/h3><h4>Does it work on Postgres RDS?<\/h4><p>All the examples illustrated work on RDS. From what I'm aware the only restrictions on search features imposed by RDS are those that require access to the file system such as custom dictionaries, ispell, synonyms and thesaurus. See <a href=\"https:\/\/forums.aws.amazon.com\/message.jspa?messageID=527000#527000\">the related issue on the aws forum<\/a>.<\/p><h4>I'm using MYSQL should I use the builtin full-text search?<\/h4><p>I wouldn't. Without starting a flame war, MySQL full-text search features are very limited. By default, there is no support for stemming nor any language support. I came accross a stemming function which can be installed, but MYSQL doesn't support function based indexes.<\/p><p>Then what can you do? Based on what we have discussed above if Postgres fulfills your use case then think about moving to Postgres. This can easily be done via tools like <a href=\"https:\/\/github.com\/philipsoutham\/py-mysql2pgsql\/\">py-mysql2pgsql<\/a>. Or you can investigate more advanced solutions like SOLR and Elasticsearch.<\/p><h3>Want to work on challenging projects using Postgres?<\/h3><p>If you are interested in working with Postgres on some interesting projects, come join me at <a href=\"http:\/\/bit.ly\/24bUVmn\">Opendoor<\/a>. We heavily use PostgreSQL for Data Processing, Web Development, Data Warehousing, ETLs, Data Science, and Geo Analysis. We are always looking for passionate Engineers and Data Scientists to join our team. Please reach out to our recruiter, <a href=\"mailto:ryan.jordan@opendoor.com\">Ryan<\/a> or apply via our <a href=\"http:\/\/bit.ly\/25nigE9\">job page<\/a><\/p><h3>Conclusion<\/h3><p>We have seen how to build a decent multi-language search engine based on a non-trivial document. This article is only an overview, but it should give you enough background and examples to get you started with your own. I may have made some mistakes in this article and I would appreciate if you report them to <a href=\"mailto:blog@rachbelaid.com\">blog@rachbelaid.com<\/a><\/p><p>The full-text search feature included in Postgres is awesome and quite fast (enough). It will allow your application to grow without depending on another tool. Is Postgres Search the silver bullet? Probably not if your core business needs revolve around search.<\/p><p>Some features are missing but in a lot of use cases you won't need them. It goes without saying that it's critical that you analyze and understand your needs to know which road to take.<\/p><p>Personally I hope to see the full-text search continuing to improve in Postgres and maybe a few of these features being included:<\/p><ul><li>Additional built-in language support. eg: Chinese, Japanese...<\/li>\n<li>Foreign data wrapper around Lucene. Lucene is still the most advanced tool for full-text search and it will have a lot of benefits to see integration with Postgres.<\/li>\n<li>More boost or scoring feature for the ranking of results would be first-rate. Elasticsearch and SOLR offer advanced solutions already.<\/li>\n<li>A way to do fuzzy <code>tsquery<\/code> without having to use trigram would be nice. Elasticsearch offers a simple way to do fuzzy search queries.<\/li>\n<li>Being able to create and edit dynamically features such as dictionary content, synonyms, thesaurus via SQL this removing the need to add files to the filesystem<\/li>\n<\/ul><p>Postgres is not as advanced as ElasticSearch and SOLR, but these two are dedicated full-text search tools whereas full-text search is only a feature of PostgreSQL and a pretty good one.<\/p>","created_at":"2020-11-23T14:31:54+0000","updated_at":"2020-11-23T14:32:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":23,"domain_name":"rachbelaid.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1209,"title":"lokesh-krishna\/dotfiles","url":"https:\/\/github.com\/lokesh-krishna\/dotfiles\/","content":"<p>My setup is built around the excellent <a href=\"https:\/\/github.com\/arcticicestudio\/nord\">Nord<\/a> color scheme. I've taken every effort to make sure that every component of the system fits right in with the rest and looks consistent. Suggestions and critique that understand this goal of the setup are welcome.<\/p><p>The original setup, developed in 2017, has now been updated to Nord v2.1 in 2020.<\/p><h2><a id=\"user-content-nord-v21\" class=\"anchor\" aria-hidden=\"true\" href=\"#nord-v21\"><\/a>Nord v2.1<\/h2><p>Below is a guided tour of what this setup looks like presently:<\/p><h3><a id=\"user-content-clean\" class=\"anchor\" aria-hidden=\"true\" href=\"#clean\"><\/a>Clean<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/clean.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/clean.png\" alt=\"Screenshot of clean desktop\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-system-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#system-information\"><\/a>System Information<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/system_information.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/system_information.png\" alt=\"neofetch\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-in-action\" class=\"anchor\" aria-hidden=\"true\" href=\"#in-action\"><\/a>In Action<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/busy.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/busy.png\" alt=\"Multiple windows\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-zathura\" class=\"anchor\" aria-hidden=\"true\" href=\"#zathura\"><\/a>Zathura<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/zathura.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/zathura.png\" alt=\"Zathura\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-rofi\" class=\"anchor\" aria-hidden=\"true\" href=\"#rofi\"><\/a>Rofi<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/rofi.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/rofi.png\" alt=\"Rofi\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-dunst\" class=\"anchor\" aria-hidden=\"true\" href=\"#dunst\"><\/a>Dunst<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/notification.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/notification.png\" alt=\"Notification\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-firefox\" class=\"anchor\" aria-hidden=\"true\" href=\"#firefox\"><\/a>Firefox<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/firefox.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/firefox.png\" alt=\"Firefox showing my custom start page\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>You can find my start page at <a href=\"https:\/\/github.com\/lokesh-krishna\/lokesh-krishna.github.io\">lokesh-krishna.github.io<\/a>.<\/p><h3><a id=\"user-content-ncmpcpp-and-cava\" class=\"anchor\" aria-hidden=\"true\" href=\"#ncmpcpp-and-cava\"><\/a>ncmpcpp and cava<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/music.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/music.png\" alt=\"Music\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-lock-screen\" class=\"anchor\" aria-hidden=\"true\" href=\"#lock-screen\"><\/a>Lock Screen<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/lock.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/lock.png\" alt=\"Lock Screen\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>Lock screen is a custom script for the excellent <a href=\"https:\/\/github.com\/Raymo111\/i3lock-color\">i3lock-color<\/a>.<\/p><h3><a id=\"user-content-wallpapers\" class=\"anchor\" aria-hidden=\"true\" href=\"#wallpapers\"><\/a>Wallpapers<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/wallpaper.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/wallpaper.png\" alt=\"Wallpaper\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a> Based on the gorgeous artwork by <a href=\"https:\/\/www.artstation.com\/artwork\/YeePYb\" rel=\"nofollow\">Denis Istomin<\/a><br \/>Edited to match the Nord color scheme using <a href=\"https:\/\/ign.schrodinger-hat.it\/\" rel=\"nofollow\">ImageGoNord<\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v2\/images\/lock-wallpaper-edit.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v2\/images\/lock-wallpaper-edit.png\" alt=\"Lock Screen Wallpaper\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a> I'm not sure where I got the original from but you can find it in the images directory. I'm using an edit created using the aforementioned ImageGoNord.<\/p><h2><a id=\"user-content-nord-v1\" class=\"anchor\" aria-hidden=\"true\" href=\"#nord-v1\"><\/a>Nord v1<\/h2><p>This is the setup as it was originally created in 2017.<\/p><h3><a id=\"user-content-clean-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#clean-1\"><\/a>Clean<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/clean.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/clean.png\" alt=\"Screenshot of clean desktop\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-system-information-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#system-information-1\"><\/a>System Information<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/sysinfo.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/sysinfo.png\" alt=\"neofetch\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-in-action-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#in-action-1\"><\/a>In Action<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/action.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/action.png\" alt=\"Multiple windows\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-zathura-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#zathura-1\"><\/a>Zathura<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/zathura.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/zathura.png\" alt=\"Zathura\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-rofi-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#rofi-1\"><\/a>Rofi<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/rofi.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/rofi.png\" alt=\"Rofi\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-dunst-and-color-scripts\" class=\"anchor\" aria-hidden=\"true\" href=\"#dunst-and-color-scripts\"><\/a>Dunst and Color Scripts<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/colors.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/colors.png\" alt=\"Notification and a couple of color scripts\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-firefox-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#firefox-1\"><\/a>Firefox<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/firefox.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/firefox.png\" alt=\"Firefox showing my custom start page\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>You can find my start page at <a href=\"https:\/\/github.com\/lokesh-krishna\/lokesh-krishna.github.io\">lokesh-krishna.github.io<\/a>.<\/p><h3><a id=\"user-content-ncmpcpp-and-cava-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#ncmpcpp-and-cava-1\"><\/a>ncmpcpp and cava<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/music.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/music.png\" alt=\"Music\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-i3lock\" class=\"anchor\" aria-hidden=\"true\" href=\"#i3lock\"><\/a>i3lock<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/lock.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/lock.png\" alt=\"Lock screen\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-wallpaper\" class=\"anchor\" aria-hidden=\"true\" href=\"#wallpaper\"><\/a>Wallpaper<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/blob\/master\/nord-v1\/images\/nord-arch.png\"><img src=\"https:\/\/github.com\/lokesh-krishna\/dotfiles\/raw\/master\/nord-v1\/images\/nord-arch.png\" alt=\"Wallpaper using the Nord color scheme\" class=\"c8\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>","created_at":"2020-11-23T04:19:38+0000","updated_at":"2020-11-23T04:19:38+0000","published_by":["lokesh-krishna"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/20235646?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1208,"title":"Full-Bleed Layout Using CSS Grid","url":"https:\/\/joshwcomeau.com\/css\/full-bleed\/","content":"<p class=\"Paragraph-d2u36h-0 btOwpF\">Back in the day, there was a gold-standard website layout that everyone strived to create, but that was notoriously difficult to get right: the .<\/p><div class=\"HolyGrail__OuterWrapper-sc-14x04g0-0 gPevom Columns__Wrapper-sc-1al8lq8-0 eagSzL c4\"><p>Header<\/p><div class=\"HolyGrail__Row-sc-14x04g0-3 bLFyzl\"><p>Nav<\/p><p>Content<\/p><p>Ads<\/p><\/div><p>Footer<\/p><\/div><p class=\"Paragraph-d2u36h-0 btOwpF\">It doesn't seem like it would be so tricky, right? But this was an era before flexbox existed; our tools for the job were tables and floats, and neither were really up to this task. It was technically possible, but some shenanigans were required.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">Once flexbox achieved mainstream browser support, this layout went from \"holy grail\" to \"fountain drink\"; it was everywhere, because it offered a great user experience, and was within reach for all developers.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">As the web has evolved, I've discovered a new aspirational layout. It offers a <em class=\"Em-sc-1jc74i7-0 kDfTEl\">fantastic<\/em> user experience, especially for long-form text content like news articles or documentation. But, like its predecessor, it's been deceptively hard to achieve; most implementations require obscure hacks or counterintuitive tricks.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">I recently discovered an elegant solution to this problem using CSS Grid. In this post, we'll learn how it works!<\/p><div class=\"MobileOnly-yx9m9d-0 eBbMdT\"><aside class=\"Sidenote__BaseWrapper-rbeavg-0 Sidenote__WarningSidenote-rbeavg-3 jIzbBy\">\n<strong class=\"Sidenote__Title-rbeavg-4 bBebJQ\">Best viewed on desktop<\/strong>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">The tutorial is all about how I built a desktop layout. It should still make sense when viewed on mobile, but you'll miss out on some of the interactive elements. If you can, I'd recommend checking it out on a computer!<\/p>\n<\/aside><\/div><p class=\"Paragraph-d2u36h-0 btOwpF\">Have you ever tried to read Wikipedia on a very large screen? It looks like this:<\/p><div class=\"FullBleed__Wrapper-sc-1hfbp5g-0 fvCfuE\"><img src=\"https:\/\/joshwcomeau.com\/images\/full-bleed\/wikipedia.png\" alt=\"Screenshot of a wikipedia article about Orange Shirt Day, a Canadian event raising awareness about the Indian residential school system. The text is very wide, since it was taken on a large screen.\" class=\"PostImage__Image-xip32o-2 eyiMVf c5\" referrerpolicy=\"no-referrer\" \/><\/div><p class=\"Paragraph-d2u36h-0 btOwpF\">Those paragraphs are so wide! Wikipedia doesn't constrain the container width at all. This leads to lines that are hundreds of characters in length.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">It's hard for our eyes to wrap around when we reach the and of a line. If you're like me, you wind up using your mouse to assist:<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">In addition to the line-wrapping concern, it's just generally hard to read lines of text that are so wide; it fatigues the eye.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\"><a href=\"https:\/\/readings.design\/PDF\/the_elements_of_typographic_style.pdf\" rel=\"noopener noreferrer\" target=\"_blank\" class=\"Link__ExternalLink-sc-68e6gi-0 fRBNyP TextLink__Wrapper-omifi9-0 wcPOv\">Research has shownResearch has shown<\/a> that the ideal line length is about 65 characters. Anywhere between 45 and 85 is generally seen as acceptable, in the context of a roman alphabet. Reading is a <a href=\"https:\/\/www.smashingmagazine.com\/2014\/09\/balancing-line-length-font-size-responsive-web-design\/\" rel=\"noopener noreferrer\" target=\"_blank\" class=\"Link__ExternalLink-sc-68e6gi-0 fRBNyP TextLink__Wrapper-omifi9-0 wcPOv\">complex processcomplex process<\/a>, and we should strive to make it as easy as possible.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">The standard solution to this problem is to create a single fixed-width column in the center of the page. You've seen this layout everywhere: online magazines, documentation, news sites, and blogs. You're looking at it right now, on this site!<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\"><img src=\"https:\/\/joshwcomeau.com\/images\/full-bleed\/this-website.png\" alt=\"Screenshot of this website, showing that there is a centered fixed-width column, with flexible blank space on either side\" class=\"PostImage__Image-xip32o-2 eyiMVf\" referrerpolicy=\"no-referrer\" \/><\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">There's a complicating factor, however\u2014<em>not all content should be constrained<\/em>. We should allow images, videos, and custom widgets to break free and fill the available width:<\/p><div class=\"FullBleed__Wrapper-sc-1hfbp5g-0 fvCfuE\"><img src=\"https:\/\/joshwcomeau.com\/images\/full-bleed\/meerkat.jpg\" alt=\"Random photo of meerkat\" class=\"PostImage__Image-xip32o-2 eyiMVf c7\" referrerpolicy=\"no-referrer\" \/><p class=\"Paragraph-d2u36h-0 btOwpF c8\">This meerkat used as an example of a child that breaks free of constraints. Photo By Sean Paul Kinnear.<\/p><\/div><p class=\"Paragraph-d2u36h-0 btOwpF\">The common term for this kind of thing is \u201cfull-bleed\u201d. It's a term borrowed from the publishing world; when something is printed full-bleed, it extends to the very edge of the paper.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">This new requirement makes the problem considerably more tricky. It's relatively easy to constrain <em>all<\/em> children, but CSS doesn't really have a mechanism to selectively constrain <em>some<\/em> children.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">Let's start at the end, with our solution:<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">These styles are assigned to markup in this shape:<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">There's a lot to unpack here, so let's go through it step by step.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">If you're not familiar with CSS Grid, this might seem like a lot of random characters and keywords. Never fear! All will be explained.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\"><code class=\"InlineCode-sc-187729h-0 kWtAqN\">grid-template-columns<\/code> is a property that lets us define the shape of our grid. By providing 3 discrete values, we're indicating that we want 3 columns.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">The values define the width of each column. The first column is <code class=\"InlineCode-sc-187729h-0 kWtAqN\">1fr<\/code>, same as the last column. The <code class=\"InlineCode-sc-187729h-0 kWtAqN\">fr<\/code> unit is a flexible unit that fills available space. It's similar in principle to <code class=\"InlineCode-sc-187729h-0 kWtAqN\">flex-grow<\/code>; it's a ratio of how much of the free space the column should consume.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">Our center column is a fixed width. We use the <code class=\"InlineCode-sc-187729h-0 kWtAqN\">min<\/code> helper to pick whichever value winds up being smaller. On large screens, it will take up <code class=\"InlineCode-sc-187729h-0 kWtAqN\">65ch<\/code> width. On smaller screens, where there isn't enough horizontal space for 65 characters, it is clamped to 100% of the available container width.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">Here's what this looks like, in practice:<\/p><div class=\"FullBleedGrid__OuterWrapper-sc-1mrpxh0-1 eQUpoV Columns__Wrapper-sc-1al8lq8-0 eagSzL FullBleedGrid__Row-sc-1mrpxh0-2 etdulK c15\"><p>1fr<\/p><p>65ch<\/p><p>1fr<\/p><\/div><aside class=\"Sidenote__BaseWrapper-rbeavg-0 Sidenote__InfoSidenote-rbeavg-1 hREzQa\">\n<strong class=\"Sidenote__Title-rbeavg-4 bBebJQ\">It\u2019s got character<\/strong>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">As we've discussed, we want to clamp the length of our text so that each line is about 65 characters wide. We can \"guestimate\" a width using pixels, but CSS has an obscure unit that makes our life a little bit easier: <code class=\"InlineCode-sc-187729h-0 kWtAqN\">ch<\/code><\/p>\n<p class=\"Paragraph-d2u36h-0 btOwpF\"><code class=\"InlineCode-sc-187729h-0 kWtAqN\">ch<\/code> is a unit, like <code class=\"InlineCode-sc-187729h-0 kWtAqN\">px<\/code> or <code class=\"InlineCode-sc-187729h-0 kWtAqN\">rem<\/code>. It corresponds to the width of the <code class=\"InlineCode-sc-187729h-0 kWtAqN\">0<\/code> character in the current font, for the specified font size. Instead of reverse-engineering a pixel width, we're specifying a width in characters.<\/p>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">In practice, you may still need to do some tinkering; if your font has a particularly narrow <code class=\"InlineCode-sc-187729h-0 kWtAqN\">0<\/code> character, for example, you'll need to exaggerate a bit.<\/p>\n<\/aside><p class=\"Paragraph-d2u36h-0 btOwpF\">We've defined a 3-column grid, and now it's time to assign children to it.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">By default, children will be slotted into the first available grid cell. We want to override this default behaviour though; all children should sit in the center column, leaving the first and third columns empty.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">In CSS, columns are 1-indexed, so <code class=\"InlineCode-sc-187729h-0 kWtAqN\">2<\/code> is a reference to the middle column.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">The asterisk (*) is a wildcard; it'll match elements of all types. We're saying that every child should be assigned to that second middle column. Each new child will create a new row, like so:<\/p><div class=\"Chunk__Wrapper-sc-1o333to-0 eNeEQV FullBleedGridWithContent__OuterWrapper-sc-9xl97h-0 eSDQcQ Columns__Wrapper-sc-1al8lq8-0 eagSzL c21\"><p>&lt;h1&gt;<\/p><p>&lt;p&gt;<\/p><p>&lt;p&gt;<\/p><p>&lt;p&gt;<\/p><\/div><aside class=\"Sidenote__BaseWrapper-rbeavg-0 Sidenote__InfoSidenote-rbeavg-1 hREzQa\">\n<strong class=\"Sidenote__Title-rbeavg-4 bBebJQ\">Wildcard performance<\/strong>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">You may have heard that using the wildcard selector (*) is bad practice. Some will tell you that it is slow, and can affect the overall performance of your page.<\/p>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">Happily, this is not true. It's been debunked <a href=\"https:\/\/www.paulirish.com\/2012\/box-sizing-border-box-ftw\/\" rel=\"noopener noreferrer\" target=\"_blank\" class=\"Link__ExternalLink-sc-68e6gi-0 fRBNyP TextLink__Wrapper-omifi9-0 wcPOv\">againagain<\/a> and <a href=\"http:\/\/www.stevesouders.com\/blog\/2009\/03\/10\/performance-impact-of-css-selectors\/\" rel=\"noopener noreferrer\" target=\"_blank\" class=\"Link__ExternalLink-sc-68e6gi-0 fRBNyP TextLink__Wrapper-omifi9-0 wcPOv\">againagain<\/a>. Even in 2009, when computers were slower and browsers were less optimized, this wasn't an issue. Performance concerns around the wildcard selector are a particularly-resilient urban legend.<\/p>\n<\/aside><p class=\"Paragraph-d2u36h-0 btOwpF\">We've seen how our grid can constrain elements of all types, but what about when we want a child to break free and fill the available width?<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">That's where this fella comes in:<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">This special <code class=\"InlineCode-sc-187729h-0 kWtAqN\">.full-bleed<\/code> class allows a specific child to bust out of that column, and span all 3 columns. <code class=\"InlineCode-sc-187729h-0 kWtAqN\">1 \/ 4<\/code> is a start\/end syntax; we're saying the element should start on Column 1 (inclusive) and span all the way to Column 4 (exclusive).<\/p><div class=\"Chunk__Wrapper-sc-1o333to-0 eNeEQV FullBleedGridWithContent__OuterWrapper-sc-9xl97h-0 eSDQcQ Columns__Wrapper-sc-1al8lq8-0 eagSzL c21\"><p>&lt;h1&gt;<\/p><p>&lt;p&gt;<\/p><div class=\"FullBleedGridWithContent__Row-sc-9xl97h-1 dkPjQJ\"><p>&lt;img class=\"full-bleed\"&gt;<\/p><\/div><p>&lt;p&gt;<\/p><\/div><p class=\"Paragraph-d2u36h-0 btOwpF\">The trick is that each child becomes its own grid row, and each child can fill as much of that row as it wants. Most elements will only ever occupy that center column, but some will instead span all 3.<\/p><aside class=\"Sidenote__BaseWrapper-rbeavg-0 Sidenote__InfoSidenote-rbeavg-1 hREzQa\">\n<strong class=\"Sidenote__Title-rbeavg-4 bBebJQ\">Conflicting styles<\/strong>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">You may have noticed that children with thee <code class=\"InlineCode-sc-187729h-0 kWtAqN\">full-bleed<\/code> class are given conflicting <code class=\"InlineCode-sc-187729h-0 kWtAqN\">grid-column<\/code> values:<\/p>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">This works because of ; our <code class=\"InlineCode-sc-187729h-0 kWtAqN\">full-bleed<\/code> class overrules the wildcard selector (<code class=\"InlineCode-sc-187729h-0 kWtAqN\">*<\/code>) because it is more specific. We are intentionally setting a conflict, since we know that our <code class=\"InlineCode-sc-187729h-0 kWtAqN\">full-bleed<\/code> class will win.<\/p>\n<p class=\"Paragraph-d2u36h-0 btOwpF\">The beauty of the wildcard selector is that it always loses these battles, so it's a great way to set default values.<\/p>\n<\/aside><p class=\"Paragraph-d2u36h-0 btOwpF\">The wonderful thing about this technique is that it's super flexible.<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">Sometimes, you won't want a child to truly be \"full-bleed\"; on ultra-wide monitors, this could be quite large. With a small tweak, we can create a <code class=\"InlineCode-sc-187729h-0 kWtAqN\">pseudo-full-bleed<\/code> variant that clamps its width to some large value:<\/p><div class=\"Chunk__Wrapper-sc-1o333to-0 eNeEQV FullBleedGridWithContent__OuterWrapper-sc-9xl97h-0 eSDQcQ Columns__Wrapper-sc-1al8lq8-0 eagSzL c21\"><p>&lt;h1&gt;<\/p><p>&lt;p&gt;<\/p><div class=\"FullBleedGridWithContent__Row-sc-9xl97h-1 dkPjQJ\"><p>&lt;img class=\"pseudo-full-bleed\"&gt;<\/p><\/div><p>&lt;p&gt;<\/p><\/div><p class=\"Paragraph-d2u36h-0 btOwpF\">CSS Grid is super powerful, and now that it's achieved <a href=\"https:\/\/caniuse.com\/css-grid\" rel=\"noopener noreferrer\" target=\"_blank\" class=\"Link__ExternalLink-sc-68e6gi-0 fRBNyP TextLink__Wrapper-omifi9-0 wcPOv\">wide browser supportwide browser support<\/a>, it can solve so many of our problems!<\/p><p class=\"Paragraph-d2u36h-0 btOwpF\">I plan to write much more about CSS Grid, so be sure to join the newsletter if you'd like to learn more. \ud83d\udc96<\/p>","created_at":"2020-10-05T17:57:49+0000","updated_at":"2020-10-05T17:57:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"joshwcomeau.com","preview_picture":"https:\/\/www.joshwcomeau.com\/images\/og-full-bleed.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["crdt","replicated-data-types"],"is_public":false,"id":1207,"title":"I was wrong. CRDTs are the future","url":"https:\/\/josephg.com\/blog\/crdts-are-the-future\/","content":"<time datetime=\"2020-09-26\">26 Sep 2020<\/time><section class=\"post-content\"><p>I saw <a href=\"https:\/\/www.youtube.com\/watch?v=x7drE24geUw\">Martin Kleppmann\u2019s talk<\/a> a few weeks ago about CRDTs, and I felt a deep sense of despair. Maybe all the work I\u2019ve been doing for the past decade won\u2019t be part of the future after all, because Martin\u2019s work on CRDTs will supersede it. Its really good.<\/p><p>Lets back up a little.<\/p><p>Around 2010 I worked on Google Wave. Wave was an attempt to make collaboratively editable spaces to replace email, google docs, web forums, instant messaging and a hundred other small single purpose applications. Wave had a property I love in my tools that I haven\u2019t seen articulated anywhere: It was a general purpose medium (like paper). Unlike a lot of other tools, it doesn\u2019t force you into its own workflow. You could use it to do anything from plan holidays, make a wiki, play D&amp;D with your friends, schedule a meeting, etc.<\/p><p>Internally, Wave\u2019s collaborative editing was built on top of Operational Transform (OT). OT has been around for awhile - the algorithm we used was based on the original <a href=\"https:\/\/www.google.com\/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=&amp;ved=2ahUKEwi3mr6CivnrAhXEfd4KHcAyBe4QFjAAegQIBBAB&amp;url=http%3A%2F%2Flively-kernel.org%2Frepository%2Fwebwerkstatt%2Fprojects%2FCollaboration%2Fpaper%2FJupiter.pdf&amp;usg=AOvVaw0HmIhcn7_VKk2h1bEeAOJS\">Jupiter paper<\/a> from 1995. It works by storing a chronological list for each document of every change. \u201cType an <em>H<\/em> at position 0\u201d. \u201cType a <em>i<\/em> at position 1\u201d. Etc. Most of the time, users are editing the latest version of the document and the operation log is just a list of all the changes. But if users are collaboratively editing, we get concurrent edits. When this happens, the first edit to arrive at the server gets recorded as usual. If the second edit is out of date, we use the log of operations as a reference to figure out what the user really intended. (Usually this just means updating character positions). Then we pretend as if thats what the user meant all along and append the new (edited) operation. Its like realtime git-rebase.<\/p><p>Once Wave died, I reimplemented the OT model in <a href=\"https:\/\/github.com\/josephg\/sharejs\">ShareJS<\/a>. This was back when node was new and weird. I think I had ShareJS working before npm launched. It only took about 1000 lines of code to get a simple collaborative editor working, and when I first demoed it I collaboratively edited a document in a browser and from a native application.<\/p><p>At its heart, OT is a <a href=\"https:\/\/github.com\/share\/sharedb\/blob\/c711cfcb777213d193b1f4a101125e8f6e8e6864\/lib\/submit-request.js#L194-L212\">glorified for() loop<\/a> with <a href=\"https:\/\/github.com\/ottypes\/text-unicode\/blob\/bdcfc545c1a2eda48fe5968ae2ce80cf743b9c08\/lib\/type.ts#L304-L380\">some helper functions<\/a> to update character offsets. In practice, this works great. OT is simple and understandable. Implementations are fast. (10k-100k operations per second in unoptimized javascript. 1-20M ops\/sec in <a href=\"https:\/\/github.com\/ottypes\/libot\">optimized C<\/a>.). The only storage overhead is the operation log, and you can trim that down if you want to. (Though you can\u2019t merge super old edits if you do). You need a centralized server to globally order operations, but most systems have a centralized server \/ database anyway, right?<\/p><h2 id=\"centralizedservers\">Centralized servers<\/h2><p>The big problem with OT is that dependancy on a centralized server. Have you ever wondered why google docs shows you that weird \u201cThis document is overloaded so editing is disabled\u201d thing when a document is shared to social media? The reason (I think) is that when you open a google doc, one server is picked as the computer all the edits run through. When the mob descends, google needs to pull out a bunch of tricks so that computer doesn\u2019t becomes overwhelmed.<\/p><p>There\u2019s some workarounds they could use to fix this. Aside from sharding by document (like google docs), you could edit via a retry loop around a database transaction. This pushes the serialization problem to your database. (<a href=\"https:\/\/firepad.io\/\">Firepad<\/a> and <a href=\"https:\/\/github.com\/share\/sharedb\/\">ShareDB<\/a> work this way).<\/p><p>Its not perfect though. We wanted Wave to replace email. Email is federated. An email thread can span multiple companies and it all just works. And unlike facebook messenger, emails are only be sent to the companies that are CC\u2019ed. If I email my coworker, my email doesn\u2019t leave the building. For Wave to replace email, we needed the same functionality. But how can that work on top of OT? We got it working, kinda, but it was complex and buggy. We ended up with <a href=\"https:\/\/web.archive.org\/web\/20180112171345\/http:\/\/www.waveprotocol.org\/protocol\/draft-protocol-specs\/draft-protocol-spec\">a scheme<\/a> where every wave would arrange a tree of wave servers and operations were passed up and down the tree. But it never really worked. <a href=\"https:\/\/www.youtube.com\/watch?v=AyvQYCv6j34\">I gave a talk<\/a> at the Wave Protocol Summit just shy of 10 years ago explaining how to get on the network. I practiced that talk, and did a full runthrough. I followed literally step by step on the day and the version I made live didn\u2019t work. I still have no idea why. Whatever the bugs are, I don\u2019t think they were ever fixed in the opensource version. Its all just too complicated.<\/p><h2 id=\"theriseofcrdts\">The rise of CRDTs<\/h2><p>Remember, the algorithm Wave used was invented in 1995. Thats a pretty long time ago. I don\u2019t think I even had the internet at home back in 1995. Since then, researchers have been busy trying to make OT work better. The most promising work uses CRDTs (Conflict-Free Replicated data types). CRDTs approach the problem slightly differently to allow realtime editing without needing a central source of truth. Martin lays out how they work in his talk better than I can, so I\u2019ll skip the details.<\/p><p>People have been asking me what I think of them for many years, and my answer was always something like this:<\/p><blockquote>\n<p>They\u2019re neat and I\u2019m glad people are working on them <em>but<\/em>:<\/p>\n<\/blockquote><ul><li>They\u2019re slow. Like, really slow. Eg Delta-CRDTs takes nearly 6 hours to process a real world editing session with a single user typing a 100KB academic paper. (<a href=\"https:\/\/github.com\/dmonad\/crdt-benchmarks\/tree\/d7f4d774a302f13f26cded6e614d44e0b5e496c9\">Benchmarks - look for B4<\/a>.)<\/li>\n<li>Because of how CRDTs work, documents grow without bound. The current automerge master takes 83MB to represent that 100KB document on disk. Can you ever delete that data? Probably not. And that data can\u2019t just sit on disk. It needs to be loaded into memory to handle edits. (Automerge currently grows to 1.1GB in memory for that.)<\/li>\n<li>CRDTs are missing features that OT has had for years. For example, nobody has yet made a CRDT that supports \/object move\/ (move something from one part of a JSON tree to another). You need this for applications like Workflowy. OT <a href=\"https:\/\/github.com\/ottypes\/json1\/\">handles this fine<\/a>.<\/li>\n<li>CRDTs are complicated and hard to reason about.<\/li>\n<li>You probably have a centralized server \/ database anyway.<\/li>\n<\/ul><p>I made all those criticisms and dismissed CRDTs. But in doing so I stopped keeping track of the literature. And - surprise! CRDTs went and quietly got better. <a href=\"https:\/\/www.youtube.com\/watch?v=x7drE24geUw\">Martin\u2019s talk<\/a> (which is well worth a watch) addressed the main points:<\/p><ul><li><strong>Speed:<\/strong> Using modern CRDTs (Automerge \/ RGA or y.js \/ YATA), applying operations should be possible with just an log(n) lookup. (More on this below).<\/li>\n<li><strong>Size:<\/strong> Martin\u2019s columnar encoding can store a text document with only about a 1.5x-2x size overhead compared to the contents themselves. Martin talks about this <a href=\"https:\/\/youtu.be\/x7drE24geUw?t=3273\">54 minutes into his talk<\/a>. The code to make this work in automerge hasn\u2019t merged yet, but Yjs implemented Martin\u2019s ideas. And in doing so, Yjs can store that same 100KB document in 160KB on disk, or 3MB in memory. Much better.<\/li>\n<li><strong>Features:<\/strong> There\u2019s at least a theoretical way to add all the features using rewinding and replaying, though nobody\u2019s implemented this stuff yet.<\/li>\n<li><strong>Complexity:<\/strong> I think a decent CRDT will be bigger than the equivalent OT implementation, but not by much. Martin managed to make a tiny, slow <a href=\"https:\/\/github.com\/automerge\/automerge\/blob\/a8d8b602ec273aaa48679e251de8829f3ce5ad41\/test\/fuzz_test.js\">implementation of automerge in only about 100 lines of code<\/a>.<\/li>\n<\/ul><p>I still wasn\u2019t completely convinced by the speed argument, so I made a <a href=\"https:\/\/github.com\/josephg\/text-crdt-rust\">simple proof of concept CRDT implementation in Rust<\/a> using a B-tree using ideas from automerge and benchmarked it. Its missing features (deleting characters, conflicts). But it can handle <a href=\"https:\/\/home.seph.codes\/public\/crdt1\/user%20pair%20append%20end\/report\/index.html\">6 million edits per second<\/a>. (Each <a href=\"https:\/\/github.com\/josephg\/text-crdt-rust\/blob\/cc3325019887ad03e89f27e26b4295d1fb2048c9\/benches\/benchmark.rs#L29-L42\">iteration<\/a> does 2000 edits to an empty document by an alternating pair of users, and that takes 330\u00b5s. So, 6.06 million inserts \/ second). So that means we\u2019ve made CRDTs good enough that the difference in speed between CRDTs and OT is smaller than the speed difference between Rust and Javascript.<\/p><p>All these improvements have been \u201ccoming soon\u201d in automerge\u2019s performance branch for a really long time now. But automerge isn\u2019t the only decent CRDT out there. <a href=\"https:\/\/github.com\/yjs\/yjs\">Y.js<\/a> works well and kicks the pants off automerge\u2019s current implementation <a href=\"https:\/\/github.com\/dmonad\/crdt-benchmarks\">in the Y.js benchmarks<\/a>. Its missing some features I want, but its generally easier to fix an implementation than invent a new algorithm.<\/p><h2 id=\"inventingthefuture\">Inventing the future<\/h2><p>I care a lot about inventing the future. What would it be ridiculous not to have in 100 years? Obviously we\u2019ll have realtime editing. But I\u2019m no longer convinced OT - and all the work I\u2019ve done on it - will still be around. I feel really sad about that.<\/p><p>JSON and REST are used everywhere these days. Lets say in 15 years realtime collaborative editing is everywhere. Whats the JSON equivalent for realtime editing that anyone can just drop in to their project? In the glorious future we\u2019ll need high quality CRDT implementations, because OT just won\u2019t work for some applications. You couldn\u2019t make a realtime version of Git, or a simple remake of Google Wave with OT. But if we have good CRDTs, do we need good OT implementations too? I\u2019m not convinced we do. Every feature OT has can be put in to a CRDT. (Including trimming operations, by the way). But the reverse is not true. Smart people disagree with me, but if we had a good, fast CRDT available from every language, with integration on the web, I don\u2019t think we need OT at all.<\/p><p>OT\u2019s one advantage is that it fits well in centralized software - which is most software today. But distributed algorithms work great in centralized software too. (Eg look at Github). And I think a really high quality CRDT running in wasm would be faster than an OT implementation in JS. And even if you only care about centralized systems, remember - Google runs into scaling problems with Google Docs because of OT\u2019s limitations.<\/p><p>So I think its about time we made a lean and fast CRDT. The academic work has been mostly done. We need more kick-ass implementations.<\/p><h2 id=\"whatsnext\">Whats next<\/h2><p>I increasingly don\u2019t care for the world of centralized software.<br \/>Software interacts with my data, on my computers. Its about time my software reflected that relationship. I want my laptop and my phone to share my files over my wifi. Not by uploading all my data to servers in another country. Especially if those servers are <a href=\"https:\/\/www.thesocialdilemma.com\/\">financed by advertisers bidding for my eyeballs<\/a>.<\/p><p>Philosophically, if I modify a google doc my computer is asking Google for <em>permission<\/em> to edit the file. (You can tell because if google\u2019s servers say no, I lose my changes.) In comparison, if I <code>git push<\/code> to github, I\u2019m only <em>notifying<\/em> github about the change to my code. My repository is mine. I own all the bits, and all the hardware that houses them. This is how I want all my software to work. Thanks to people like Martin, we now know <em>how<\/em> to make good CRDTs. But there\u2019s still a lot of code to write before <a href=\"https:\/\/www.inkandswitch.com\/local-first.html\">local first software<\/a> can become the default.<\/p><p>So Operational Transform, I think this is goodbye from me. We had some great times. Some of the most challenging, fun code I\u2019ve ever written was <a href=\"https:\/\/github.com\/josephg\/sharejs\">operational<\/a> <a href=\"https:\/\/github.com\/share\/sharedb\/\">transform<\/a> <a href=\"https:\/\/github.com\/ottypes\/json1\/\">code<\/a>. OT - you\u2019re clever and fascinating, but CRDTs can do things you were never capable of. And CRDTs need me. With some good implementations, I think we can make something really special.<\/p><p>I mourn all the work I\u2019ve done on OT over the years. But OT is no longer fits into the vision I have for the future. CRDTs would let us remake Wave, but simpler and better. And they would let us write software that treats users as digital citizens, not a digital serfs. <a href=\"https:\/\/josephg.com\/blog\/home-is-where-the-bits-flow\/\">And that matters.<\/a><\/p><p>The time to build is now.<\/p><\/section>","created_at":"2020-09-28T16:51:54+0000","updated_at":"2020-09-28T16:52:11+0000","published_at":"2020-09-26T11:08:40+0000","published_by":["Joseph Gentle"],"annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"josephg.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["productivity"],"is_public":false,"id":1206,"title":"Pmarchive - Guide to Personal Productivity","url":"https:\/\/pmarchive.com\/guide_to_personal_productivity.html","content":"<p>Productivity porn (or, for those <em>really<\/em> in the know, \"productivity pr0n\") consists of techniques, tactics, and tricks for maximizing personal productivity -- or, as they say, \"getting things done\".<\/p><p>The techniques that follow work together as an integrated set for <em>me<\/em>, but they probably won't for you. Maybe you'll get one or two ideas -- probably out of the ideas I stole from other people. If so, I have succeeded.<\/p><div>Let's start with a bang: <strong>don't keep a schedule<\/strong>.\n<p>He's crazy, you say!<\/p>\n<p>I'm totally serious. If you pull it off -- and in many structured jobs, you simply can't -- this simple tip alone can make a huge difference in productivity.<\/p>\n<p>By not keeping a schedule, I mean: refuse to commit to meetings, appointments, or activities at any set time in any future day.<\/p>\n<p>As a result, you can always <strong>work on whatever is most important or most interesting<\/strong>, at any time.<\/p>\n<p>Want to spend all day writing a research report? Do it!<\/p>\n<p>Want to spend all day coding? Do it!<\/p>\n<p>Want to spend all day at the cafe down the street reading a book on personal productivity? Do it!<\/p>\n<p>When someone emails or calls to say, \"Let's meet on Tuesday at 3\", the appropriate response is: \"I'm not keeping a schedule for 2007, so I can't commit to that, but give me a call on Tuesday at 2:45 and if I'm available, I'll meet with you.\"<\/p>\n<p>Or, if it's important, say, \"You know what, let's meet right now.\"<\/p>\n<p>Clearly this only works if you can get away with it. If you have a structured job, a structured job environment, or you're a CEO, it will be hard to pull off.<\/p>\n<p>But if you can do it, it's really liberating, and will lead to far higher productivity than almost any other tactic you can try.<\/p>\n<p>This idea comes from a wonderful book called <a href=\"http:\/\/www.amazon.com\/dp\/0316114758?tag=marandsblo-20&amp;camp=0&amp;creative=0&amp;linkCode=as1&amp;creativeASIN=0316114758&amp;adclass=1YMJH8AZ0PKFV6P2SPQ0&amp;\">A Perfect Mess<\/a>, which explains how not keeping a schedule has been key to Arnold Schwarzenegger's success as a movie star, politician, and businessman over the last 20 years.<\/p>\n<p>Want to meet with Arnold? Sure, drop on by. He'll see you if he can. But you might want to call first. Sorry, he doesn't schedule appointments in advance.<\/p>\n<p>As a result, for 20 years he has been free to work on whatever is most important in his life at any time.<\/p>\n<p>Those of you in California may recall how, once Arnold decided to run for Governor, he went into a blaze of action and activity that resulted in a landslide victory. The book attributes this in part to the fact that his schedule was completely clear and he could spend all day, every day on his new political career, without having to worry about distractions or commitments.<\/p>\n<p>If you have at any point in your life lived a relatively structured existence -- probably due to some kind of job with regular office hours, meetings, and the like -- you will know that there is nothing more liberating than looking at your calendar and seeing nothing but free time for weeks ahead to work on the most important things in whatever order you want.<\/p>\n<p>This also gives you the best odds of maximizing <a href=\"http:\/\/www.amazon.com\/dp\/0060920432?tag=marandsblo-20&amp;camp=0&amp;creative=0&amp;linkCode=as1&amp;creativeASIN=0060920432&amp;adclass=1YPJ51PQ4NMETWMDDBWQ&amp;\">flow<\/a>, which is a whole 'nother topic but highly related.<\/p>\n<p>I've been trying this tactic as an experiment in 2007, as those of you who have emailed me to suggest we get together or that I go to a conference or to a meeting will attest. And I am so much happier, I can't even tell you. I get so much more time to focus on the things that <em>really<\/em> matter -- in my case, my two companies, my nonprofit boards, and my lovely wife.<\/p>\n<p>The other great thing about this tactic is that it doesn't have to be all or nothing -- there are quite a few things that still sneak onto my calendar that I really can't get out of. But one is still able to draw the line between \"must do\" and \"sounds interesting but I'm not keeping a schedule\".<\/p>\n\n<strong>Keep three and only three lists: a Todo List, a Watch List, and a Later List.<\/strong>\n<p>The more into lists you are, the more important this is.<\/p>\n<p>Into the Todo List goes all the stuff you \"must\" do -- commitments, obligations, things that have to be done. A single list, possibly subcategorized by timeframe (today, this week, next week, next month).<\/p>\n<p>Into the Watch List goes all the stuff going on in your life that you have to follow up on, wait for someone else to get back to you on, remind yourself of in the future, or otherwise remember.<\/p>\n<p>Into the Later List goes everything else -- everything you might want to do or will do when you have time or wish you could do.<\/p>\n<p>If it doesn't go on one of those three lists, it goes away.<\/p>\n\n<strong>Each night before you go to bed, prepare a 3x5 index card with a short list of 3 to 5 things that you will do the next day.<\/strong>\n<p>And then, the next day, do those things.<\/p>\n<p>I sit down at my desk before I go to sleep, pull up my Todo List (which I keep in Microsoft Word's outline mode, due to long habit), and pick out the 3 to 5 things I am going to get done tomorrow. I write those things on a fresh 3x5 card, lay the card out with my card keys, and go to bed. Then, the next day, I try like hell to get just those things done. If I do, it was a successful day.<\/p>\n<p>People who have tried lots of productivity porn techniques will tell you that this is one of the most successful techniques they have ever tried.<\/p>\n<p>Once you get into the habit, you start to realize how many days you used to have when you wouldn't get 3 to 5 important\/significant\/meaningful things done during a day.<\/p>\n\nThen, throughout the rest of the day, <strong>use the back of the 3x5 card as your Anti-Todo List<\/strong>.\n<p>This isn't a real list. And the name is tongue firmly in cheek.<\/p>\n<p>What you do is this: every time you do something -- anything -- useful during the day, write it down in your Anti-Todo List on the card.<\/p>\n<p>Each time you do something, you get to write it down and you get that little rush of endorphins that the mouse gets every time he presses the button in his cage and gets a food pellet.<\/p>\n<p>And then at the end of the day, before you prepare tomorrow's 3x5 card, take a look at today's card and its Anti-Todo list and marvel at all the things you actually got done that day.<\/p>\n<p>Then tear it up and throw it away.<\/p>\n<p>Another day well spent, and productive.<\/p>\n<p>I love this technique -- being able to put more notches on my accomplishment belt, so to speak, by writing down things on my Anti-Todo list as I accomplish them throughout the day makes me feel marvelously productive and efficient. Far more so than if I just did those things and didn't write them down.<\/p>\n<p>Plus, you know those days when you're running around all day and doing stuff and talking to people and making calls and responding to emails and filling out paperwork and you get home and you're completely exhausted and you say to yourself, \"What the hell did I actually get done today?\"<\/p>\n<p>Your Anti-Todo list has the answer.<\/p>\n<p>By the way, in order to do this, you have to carry a pen with you everywhere you go. I recommend the <a href=\"http:\/\/www.spacepen.com\/Public\/Home\/index.cfm\">Fisher Space Pen<\/a>. It's short and bullet-shaped so it won't poke you in the thigh when it's in your pocket, it's wonderfully retro, it helped save the Apollo 11 mission, and it writes upside down. What's not to like?<\/p>\n\n<strong>Structured Procrastination<\/strong>.\n<p>This is a great one.<\/p>\n<p>This one is <a href=\"http:\/\/www.structuredprocrastination.com\/\">lifted straight from the genius mind of John Perry<\/a>, a philosophy professor at Stanford.<\/p>\n<p>Read his <a href=\"http:\/\/www.structuredprocrastination.com\/\">original description<\/a>, by all means. You even get to see a photo of him practicing jumping rope with seaweed on a beach while work awaits. Outstanding.<\/p>\n<p>The gist of Structured Procrastination is that you should never fight the tendency to procrastinate -- instead, you should use it to your advantage in order to get other things done.<\/p>\n<p>Generally in the course of a day, there is something you have to do that you are not doing because you are procrastinating.<\/p>\n<p>While you're procrastinating, just do lots of other stuff instead.<\/p>\n<p>As John says, \"The list of tasks one has in mind will be ordered by importance. Tasks that seem most urgent and important are on top. But there are also worthwhile tasks to perform lower down on the list. Doing these tasks becomes a way of not doing the things higher up on the list. With this sort of appropriate task structure, the procrastinator becomes a useful citizen. Indeed, the procrastinator can even acquire, as I have, a reputation for getting a lot done.\"<\/p>\n<p>Reading <a href=\"http:\/\/www.structuredprocrastination.com\/\">John's essay<\/a> was one of the single most profound moments of my entire life.<\/p>\n<p>For example, I hate making phone calls. Hate it. Love sending emails, enjoy seeing people face to face (sometimes), but I hate making phone calls.<\/p>\n<p>I can get so much done while I am avoiding making a phone call that I need to make, I can barely believe it.<\/p>\n<p>In fact, that's what's happening right now.<\/p>\n\nThe other key two-word tactic: <a href=\"http:\/\/www.marketwatch.com\/news\/story\/art-showing-pure-incompetence\/story.aspx?guclass=4FD9BA5C-E2CA-43AA-8C1B-4C4F8D057646\"><strong>Strategic Incompetence<\/strong><\/a>.\n<p>The best way to to make sure that you are never asked to do something again is to royally screw it up the first time you are asked to do it.<\/p>\n<p>Or, better yet, just say you know you will royally screw it up -- maintain a strong voice and a clear gaze, and you'll probably get off the hook.<\/p>\n<p>Of course, this assumes that there are other things that are more important at which you are competent.<\/p>\n<p>Which, hopefully, there are.<\/p>\n<p>Organizing the company picnic, sending faxes or Fedexes, negotiating with insurance brokers, writing in plain English... the list of things at which one can be strategically incompetent is nearly endless.<\/p>\n\n<strong>Do email exactly twice a day<\/strong> -- say, once first thing in the morning, and once at the end of the workday.\n<p>Allocated half an hour or whatever it takes, but otherwise, <strong>keep your email client shut and your email notifications turned off<\/strong>.<\/p>\n<p>Anyone who needs to reach you so urgently that it can't wait until later in the day or tomorrow morning can call you, or send a runner, or send up smoke signals, or something else.<\/p>\n<p>Or, more likely, find someone else who can do whatever it is that needs doing.<\/p>\n<p>(If you communicate with your spouse or key family members via email during the day, then just set up a separate email account just for them and leave that open all day, but keep your primary email closed. And never give out the family email address to anyone noncritical -- including your boss.)<\/p>\n<p>Only doing email twice a day will make you far more productive for the rest of the day.<\/p>\n<p><em>The<\/em> problem with email is that getting an email triggers that same endorphin hit I mentioned above -- the one that a mouse gets when he bonks on the button in the cage and gets a food pellet.<\/p>\n<p>Responding to an email triggers that same hit.<\/p>\n<p>The pleasure chemical hits your neocortex and you go \"ahhh\" inside and feel like you've done something.<\/p>\n<p>So you sit and work with your mail client open and you interrupt your work every time an email comes in and you answer it and you send another email and you feel great in the moment.<\/p>\n<p>But what you're really doing is fracturing your time, interrupting your <a href=\"http:\/\/www.amazon.com\/dp\/0060920432?tag=marandsblo-20&amp;camp=0&amp;creative=0&amp;linkCode=as1&amp;creativeASIN=0060920432&amp;adclass=1YPJ51PQ4NMETWMDDBWQ&amp;\">flow<\/a>, and killing your ability to focus on anything long enough to get real high-quality work done.<\/p>\n<p>This one is far easier to say than do. And it won't be feasible during projects where lots of updates during the day really are important -- raising money, for example, or closing a big deal.<\/p>\n<p>Me, I'm just trying to get down to checking email only a half dozen times per day.<\/p>\n\n<strong>When you do process email, do it like this<\/strong>:\n<p>First, <strong>always finish each of your two daily email sessions with a completely empty inbox<\/strong>.<\/p>\n<p>I don't know about you, but when I know I have emails in my inbox that haven't been dealt with, I find it hard to concentrate on other things.<\/p>\n<p>The urge to go back to my email is nearly overpowering.<\/p>\n<p>(I am apparently <em>seriously<\/em> addicted to endorphins.)<\/p>\n<p>Second, when doing email, <strong>either answer or file every single message until you get to that empty inbox state of grace<\/strong>.<\/p>\n<p>Not keeping a schedule helps here, a lot, if you can pull it off -- you can reply to a lot of messages with \"I'm sorry, I'm not keeping a schedule in 2007, I can't commit to that.\"<\/p>\n<p>Third, <strong>emails relating to topics that are current working projects or pressing issues go into temporary subfolders of a folder called Action<\/strong>.<\/p>\n<p>You should only have Action subfolders for the things that really matter, right now.<\/p>\n<p>Those subfolders then get used, and the messages in them processed, when you are working on their respective projects in the normal course of your day.<\/p>\n<p>Fourth, aside from those temporary Action subfolders, <strong>only keep three standing email folders: Pending, Review, and Vault<\/strong>.<\/p>\n<p>Emails that you know you're going to have to deal with again -- such as emails in which someone is committing something to you and you want to be reminded to follow up on it if the person doesn't deliver -- go in Pending.<\/p>\n<p>Emails with things you want to read in depth when you have more time, go into Review.<\/p>\n<p>Everything else goes into Vault.<\/p>\n<p>Every once in a while, sweep through your Action subfolders and dump any of them that you can into Vault.<\/p>\n<p>(And do the same thing for messages in your Pending folder -- most of the things in there you will never look at again. Actually, same is true for Review.)<\/p>\n<p>That's it.<\/p>\n<p>You can get away with this because modern email clients are so good at search (well, most of them -- and you can always move to GMail) that it's not worth the effort to try to file emails into lots of different folders.<\/p>\n<p>Obviously you may need some additional permanent folders for important things like contracts, or emails from your doctor, or the like, but these are exceptions and don't change your standard operating procedure.<\/p>\n\n<strong>Don't answer the phone.<\/strong>\n<p>Let it go to voicemail, and then every few hours, screen your voicemails and batch the return calls.<\/p>\n<p>Say, twice a day.<\/p>\n<p>Cell phones and family plans are so cheap these days that I think the best thing to do is have two cell phones with different numbers -- one for key family members, your closest friends, and your boss and a few coworkers, and the other for everyone else.<\/p>\n<p>Answer the first one when it rings, but never answer the second one.<\/p>\n\n<strong>Hide in an IPod.<\/strong>\n<p>One of the best and easiest ways to avoid distractions in the workplace is to be wearing those cute little IPod earbud headphones (or any other headphones of your choice).<\/p>\n<p>People, for some reason, feel much worse interrupting you if you are wearing headphones than if you're not.<\/p>\n<p>It's great -- a lot of the time, people will walk up to you, start to say something, notice the headphones, apologize (using exaggerated mouth motions), and walk away.<\/p>\n<p>This is great -- half the time they didn't actually need to talk to you, and the other half of the time they can send an email that you can process at the end of the day during the second of your two daily email sweeps.<\/p>\n<p>Here's the best part: you don't actually have to be listening to anything.<\/p>\n<p>Hell, you don't even have to have the headphones plugged into anything.<\/p>\n\nI'm not going to talk a lot about getting up early or going to bed late or anything else related to the course of a typical day, because everyone's different.\n<p>Personally I go back and forth between being a night owl (99% of the time) and a morning person (1% -- I'm going to try to push it to 2%).<\/p>\n<p>But the thing that matters almost more than anything in determining whether I'll have a happy, satisfying day is this: no matter what time you get up, <strong>start the day with a real, sit-down breakfast<\/strong>.<\/p>\n<p>This serves two purposes.<\/p>\n<p>First, it fuels you up. Study after study have shown that breakfast is, yes, the most important meal of the day. It's critical to properly fuel the body for the day's activities and it's also critical to staying lean or losing weight. (People who don't have breakfast tend to eat more, and worse, at lunch.)<\/p>\n<p>Second, it gives you a chance to calmly, peacefully collect your thoughts and prepare mentally and emotionally for the day ahead.<\/p>\n<p>This works whether you do it with kids and\/or a partner, or you're solo.<\/p>\n<p>Personally I think it's worth whatever effort is involved to go to bed early enough to wake up early enough to have a good solid 45 minutes or an hour for breakfast each morning, if you can pull it off.<\/p>\n\n<strong>Only agree to new commitments when both your head and your heart say yes.<\/strong>\n<p>This one is from the great <a href=\"http:\/\/www.amazon.com\/Kid-Stays-Picture-Robert-Evans\/dp\/1597770094\/ref=ed_oe_a\/105-5658572-8258062?ie=UTF8&amp;qclass=1181003950&amp;sr=1-1\">Robert Evans<\/a>.<\/p>\n<p>(Hold out for the <a href=\"http:\/\/www.amazon.com\/Kid-Stays-Picture-Robert-Evans\/dp\/1597770094\/ref=ed_oe_a\/105-5658572-8258062?ie=UTF8&amp;qclass=1181003950&amp;sr=1-1\">audiobook<\/a> -- trust me.)<\/p>\n<p>It's really easy to get asked to do something -- a new project, a nonprofit activity, a social event -- and to have your head say yes and your heart say no, and then your mouth says yes.<\/p>\n<p>The next thing you know, you're piled up with all kinds of things on your schedule that sounded like a good idea at the time but you really don't want to do.<\/p>\n<p>And distract you from the things that really matter.<\/p>\n<p>And make you angry, and bitter, and sullen, and hostile.<\/p>\n<p>(Oh, wait, I'm projecting.)<\/p>\n<p>In my experience, it takes time to tell the difference between your head saying yes and your heart saying yes.<\/p>\n<p>I think the key is whether you're really excited about it.<\/p>\n<p>If you get that little adrenaline spike (in a good way) when you think about it, then your heart is saying yes.<\/p>\n<p>The corollary, of course, is that when your head says no and your heart says yes, your mouth should generally say yes as well :-).<\/p>\n<p>But not when your head says yes and your heart says no.<\/p>\n\n<strong>Do something you love.<\/strong>\n<p>As you've probably concluded by now, most of the tactics described in this post involve keeping oneself as free as possible to pursue one's core interests, and dreams.<\/p>\n<p>If you're not doing something you love with the majority of your time, and you have any personal freedom and flexibility whatsoever, it's time for a change.<\/p>\n<p>And this doesn't mean something that you love doing in theory -- but rather, the core thing you love doing in practice.<\/p>\n<\/div><p>And that's it.<\/p><p>Please feel free to nominate additions to the list! Next time my mobile wiki-based GTD Outlook synchronized hipster PDA reminds me, I'll check 'em out.<\/p><p>Turns out <a href=\"http:\/\/www.hackvan.com\/pub\/stig\/etext\/how-to-get-things-done-despite-procrastination.txt\">Robert Benchley<\/a> wrote about structured procrastination back in 1949. Wonderful essay -- highly recommended.<\/p><p>The sharpest reaction has been to my theory of <strong>not keeping a schedule<\/strong>. I'll stick to my theory but make (or re-make) a couple of clarifying points.<\/p><p>First, it is certainly true that many people have jobs and responsibilities where they can't do that. Or maybe can only do it partially. And many people enjoy living a highly structured life and obviously this approach is not for them.<\/p><p>But if your reaction is, \"boy, I wish I could do that\", then it may well be worth rethinking your approach to your career.<\/p><p>I can tell you from personal experience that being stuck in a role where you have a lot of structure but feel like you never get anything done is <em>not<\/em> the optimal way to advance in one's profession, or maximize one's job satisfaction.<\/p><p>Second, I do not recommend pursuing this approach in one's personal life :-).<\/p><p>On another topic, the tactic of <strong>each night, write down the 3 to 5 things you need to do the next day<\/strong> has struck some people as too simplistic.<\/p><p>That may be the case for some people, but I can't tell you how many times I've arrived home at night and am at a loss as to what I actually got done that day, despite the fact that I worked all day.<\/p><p>And I also can't tell you how often I've had a huge, highly-structured todo list in front of me with 100 things on it and I stare at it and am paralyzed into inaction (or, more likely, structured procrastination).<\/p><p>So a day when I get 3 to 5 concrete, actionable things done in addition to all the other stuff one has to do to get through the day -- well, that's a good day.<\/p><p>A few people have said, why not just use GTD (David Allen's \"Getting Things Done\" approach).<\/p><p>While I find GTD to be highly inspiring, in practice I think it's awfully complex. At least if your job is based on project work (as opposed to having a highly structured role like CEO or head of sales).<\/p><p>For me, an organization system that requires significant time to deal with in and of itself is not optimal. Much better, for me at least, is to focus on stripping away nonessentials and freeing up as much time as possible to deal with whatever is most important.<\/p><p>Finally, I discovered after writing this post that <a href=\"http:\/\/www.paulgraham.com\/marginal.html\">Paul Graham<\/a> talks a bit about the role of time and focus in personal productivity in his essay on \"The Power of the Marginal\".<\/p>","created_at":"2020-09-26T15:23:10+0000","updated_at":"2020-09-26T15:23:19+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"pmarchive.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["open-source","licensing","lawyer"],"is_public":false,"id":1205,"title":"Casner & Edwards","url":"https:\/\/www.casneredwards.com\/attorneys\/steven-m-ayr","content":"<p>Steven has served the small business, startup, and artist communities since launching his own practice in 2009, and as Business Counsel at a startup-focused, local law firm. He has experience structuring organizations from formation to dissolution, negotiating, drafting, and revising the contracts that make businesses tick, and litigating disputes when relationships break down. In addition, Steven is an intellectual property nerd and helps clients leverage copyright, trademark and trade secret protections to develop their products, grow their brands, and succeed in the marketplace.<\/p><p>Steven joined Casner &amp; Edwards in 2017\u00a0and continues to represent the entrepreneur community.\u00a0<\/p><h3><br \/>Admissions<\/h3><ul><li>Massachusetts<\/li>\n<li>Connecticut<\/li>\n<\/ul><h3>Education<\/h3><ul><li>JD, Catholic University, 2008<\/li>\n<li>BA, Boston College, 2005<\/li>\n<\/ul>","created_at":"2020-09-25T13:11:20+0000","updated_at":"2020-09-25T13:11:28+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.casneredwards.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["databases","open-source","cloud","licensing"],"is_public":false,"id":1204,"title":"building-open-source-business-in-cloud-era-v2","url":"https:\/\/blog.timescale.com\/blog\/building-open-source-business-in-cloud-era-v2\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-09-25T01:07:08+0000","updated_at":"2020-09-25T01:07:38+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"blog.timescale.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1203,"title":"cortex-command-community\/Cortex-Command-Community-Project-Source","url":"https:\/\/github.com\/cortex-command-community\/Cortex-Command-Community-Project-Source","content":"<p><em>The Cortex Command Community Project is Free\/Libre and Open Source under GNU AGPL v3<\/em><\/p><p>This is a community-driven effort to continue the development of Cortex Command. Stay up to date in our <a href=\"https:\/\/discord.gg\/SdNnKJN\" rel=\"nofollow\">discord channel<\/a>.<\/p><hr \/><h2><a id=\"user-content-setup-process\" class=\"anchor\" aria-hidden=\"true\" href=\"#setup-process\"><\/a>Setup Process<\/h2><p>First you need to download the necessary files:<\/p><ol><li>\n<p>Install the necessary tools.<br \/>You'll probably want <a href=\"https:\/\/visualstudio.microsoft.com\/downloads\/\" rel=\"nofollow\">Visual Studio Community Edition<\/a> (we mostly use 2017 but 2019 should be fine).<br \/>You also need to have <a href=\"https:\/\/support.microsoft.com\/en-us\/help\/2977003\/the-latest-supported-visual-c-downloads\" rel=\"nofollow\">Visual C++ Redistributable for Visual Studio 2017 (x86)<\/a> installed in order to run the compiled builds.<br \/>You may also want to check out the list of recommended Visual Studio plugins <a href=\"https:\/\/github.com\/cortex-command-community\/Cortex-Command-Community-Project-Source\/wiki\/Information,-Recommended-Plugins-and-Useful-Links\">here<\/a>.<\/p>\n<\/li>\n<li>\n<p>Clone this Source Repository and the <a href=\"https:\/\/github.com\/cortex-command-community\/Cortex-Command-Community-Project-Data\">Data Repository<\/a> in neighboring folders.<br \/><strong>Do Not<\/strong> change the folder names unless you want to make trouble for yourself.<\/p>\n<\/li>\n<li>\n<p>Copy the following libraries from <code>Cortex-Command-Community-Project-Source\\external\\lib\\<\/code> into the Data Repository:<\/p>\n<\/li>\n<\/ol><ul><li><code>lua51.dll<\/code><\/li>\n<li><code>fmod.dll<\/code><\/li>\n<li><code>liblz4.dll<\/code><\/li>\n<li><code>zlibwapi.dll<\/code><\/li>\n<\/ul><ol start=\"4\"><li>Copy <code>Scenes.rte<\/code> and <code>Metagames.rte<\/code> from your purchased copy of Cortex Command into the Data Repository.<\/li>\n<\/ol><p>Now you're ready to build and launch the game.<br \/>Simply open <code>RTEA.sln<\/code> with Visual Studio, choose your configuration, and run the project.<\/p><ul><li>Use \"<code>Debug Open Source<\/code>\" configuration to debug (this runs very slowly)<\/li>\n<li>Use \"<code>Minimal Debug Open Source<\/code>\" configuration to debug with all visual debug elements disabled (this runs slightly faster)<\/li>\n<li>Use \"<code>Final Open Source<\/code>\" configuration to build release .exe<\/li>\n<\/ul><p>The first build will take a while, but future ones should be quicker.<\/p><h2><a id=\"user-content-linux-build-instructions\" class=\"anchor\" aria-hidden=\"true\" href=\"#linux-build-instructions\"><\/a>Linux Build Instructions<\/h2><p>The Linux build uses the meson build system, and builds against system libraries<\/p><p>Dependencies:<\/p><ul><li><code>allegro4<\/code><\/li>\n<li><code>loadpng<\/code><\/li>\n<li><code>flac<\/code><\/li>\n<li><code>luajit<\/code><\/li>\n<li><code>minizip<\/code><\/li>\n<li><code>lz4<\/code><\/li>\n<li><code>libpng<\/code><\/li>\n<li><code>libX11<\/code><\/li>\n<li><code>meson&gt;=0.49<\/code><\/li>\n<li><code>boost&gt;=1.55<\/code><\/li>\n<li><code>xorg-misc-fonts<\/code><\/li>\n<\/ul><p>Building:<\/p><ol><li>\n<p>Install Dependencies (see below for some distro-specific instructions)<\/p>\n<\/li>\n<li>\n<p>Clone this Source Repository and the <a href=\"https:\/\/github.com\/cortex-command-community\/Cortex-Command-Community-Project-Data\">Data Respository<\/a>, folder structure doesn't really matter here<\/p>\n<\/li>\n<li>\n<p>open a terminal in the Source Repository<\/p>\n<\/li>\n<li>\n<p><code>meson builddir<\/code><\/p>\n<\/li>\n<li>\n<p><code>cd builddir<\/code><\/p>\n<\/li>\n<li>\n<p><code>meson compile<\/code> or <code>meson [-j&lt;num of threads&gt;] compile CCCP<\/code> if you want a release build If the build fails because of memory shortage you may need to reduce the number of build threads (meson will use all available threads by default) using the <code>-j&lt;number of threads&gt;<\/code> option<\/p>\n<\/li>\n<\/ol><p>Running:<\/p><ol><li>\n<p>Copy (or link, preferable for quick debug builds) <code>builddir\/CCCP_debug.x86_64<\/code> or <code>builddir\/CCCP.x86_64<\/code> (depending on if you made a release build) into the <strong>Data Repository<\/strong><\/p>\n<\/li>\n<li>\n<p>Copy <code>Scenes.rte<\/code> and <code>Metagames.rte<\/code> from your purchased copy of Cortex Command into <strong>Data Repository<\/strong><\/p>\n<\/li>\n<li>\n<p>Copy (or link) <code>libfmod.so.11<\/code> from <code>&lt;Source Repository Root&gt;\/external\/lib\/linux\/x86_64<\/code> into the <strong>Data Repository<\/strong><\/p>\n<\/li>\n<li>\n<p>Run <code>env LD_LIBRARY_PATH=. .\/CCCP.x86_64<\/code> or <code>.\/CCCP_debug.x86_64<\/code> in the <strong>Data Repository<\/strong><\/p>\n<\/li>\n<\/ol><h3><a id=\"user-content-arch-linux\" class=\"anchor\" aria-hidden=\"true\" href=\"#arch-linux\"><\/a>Arch Linux<\/h3><p><code># pacman -S allegro4 flac luajit minizip lz4 libpng libx11 meson xorg-fonts-misc<\/code><\/p><h3><a id=\"user-content-ubuntu\" class=\"anchor\" aria-hidden=\"true\" href=\"#ubuntu\"><\/a>Ubuntu<\/h3><p><code># apt-get install liballegro4.4 libloadpng4-dev libflac++-dev luajit-5.1-dev libminizip-dev liblz4-dev libpng++-dev libx11-dev meson<\/code><\/p><h3><a id=\"user-content-troubleshooting\" class=\"anchor\" aria-hidden=\"true\" href=\"#troubleshooting\"><\/a>Troubleshooting<\/h3><p>Until borderless windows are implemented, you might seem get stuck in fullscreen mode. Try Alt-Return, or if that doesn't work kill CC with ctrl-alt-end.<\/p><p>On X11 media keys and such don't work in fullscreen, this is a known issue (this does not happen on Wayland).<\/p><p>If you get errors similar to <code>ALSA lib ... could not open slave<\/code>, try adding:<\/p><pre>option snd_mia index=0\noption snd_intel_hda index=1\n<\/pre><p>to <code>\/etc\/modprobe.d\/alsa-base.conf<\/code><\/p><h2><a id=\"user-content-more-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-information\"><\/a>More Information<\/h2><p>For more information and recommendations, see <a href=\"https:\/\/github.com\/cortex-command-community\/Cortex-Command-Community-Project-Source\/wiki\/Information,-Recommended-Plugins-and-Useful-Links\">here<\/a>.<\/p>","created_at":"2020-09-24T19:57:19+0000","updated_at":"2020-09-24T19:57:19+0000","published_by":["cortex-command-community"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/51932416?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1202,"title":"The best Civ 6 mods","url":"https:\/\/www.pcgamesn.com\/civilization-vi\/civilization-6-mods","content":"<p class=\"intro_text\">Here are two of the best things about PC gaming: the Civilization series, and the ability to change your games by adding in user-created content. Modding Civ 6, then, is an expression of the very apex of PC gaming. It\u2019s peak PC.<\/p><p>Since the Civ 6 modding toolkit came out in February 2017, the number of mods on offer has been steadily expanding. Civ 6\u2019s modding community isn\u2019t yet at the level of its timeless predecessors, but modders have now created enough new leaders, gameplay tweaks and AI improvements for you to swing your caveman club at. There are over 1,400 Steam Workshop,in fact, so plenty of options for every budding ruthless\/benevolent leader.<\/p>\n<p><em>Need even more tactical joy? Try out one of the <a href=\"https:\/\/www.pcgamesn.com\/15-great-strategy-games-you-really-should-play-or-else\" target=\"_blank\" rel=\"noopener noreferrer\">best strategy games on PC<\/a>.\u00a0<\/em><\/p>\n<p>So we\u2019ve congregated the best efforts of Civ 6 modders from both <a href=\"https:\/\/steamcommunity.com\/workshop\/\" target=\"_blank\" rel=\"noopener noreferrer\">Steam Workshop<\/a> and <a href=\"https:\/\/www.civfanatics.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Civfanatics.com<\/a>, into a long unravelling papyrus of a list. All of the mods here are compatible with the Rise &amp; Fall expansion pack, so no matter how invested you are in Civ, all of these mods will work just right.<\/p>\n<h2>Steel &amp; Thunder Unique Units\/Units Expansion<\/h2>\n<p><img title=\"Civ 6 mods - Steel and Thunder\" alt=\"Civ 6 mods - Steel and Thunder\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/steel_and_thunder.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Deliverator\u2019s comprehensive mod (formerly known as MOAR units) diversifies and brightens up the game with plenty of new units. The mod adds one unique unit to every civilisation in the game (including DLC and Rise &amp; Fall civs), as well as 11 new global units that slot into existing upgrade paths.<\/p>\n<p>New global units include, among others, Gatling Gun, Longswordsman, RIfleman, Naval Cog and Galleass. Listed below, meanwhile, are some of our favourite unique Civ units:<\/p>\n<p><strong>\u2013 America \u2013<\/strong> AH-64 Apache<br \/><strong>\u2013 Germany \u2013<\/strong> Landsknecht<br \/><strong>\u2013 England \u2013<\/strong> Ironside<br \/><strong>\u2013 Poland \u2013<\/strong> Uhlan<br \/><strong>\u2013 Kongo \u2013<\/strong> Medicine Man<br \/><strong>\u2013 Khmer \u2013<\/strong> War Canoe<\/p>\n<p>Steel and Thunder comes in two separate packs: <a href=\"https:\/\/forums.civfanatics.com\/resources\/steel-and-thunder-unit-expansion.26709\/\" target=\"_blank\" rel=\"noopener noreferrer\">Unit Expansion<\/a> adds the global units, and <a href=\"https:\/\/forums.civfanatics.com\/resources\/steel-and-thunder-unique-units.26820\/\" target=\"_blank\" rel=\"noopener noreferrer\">Unique Units<\/a> adds the civ-specific ones.<\/p>\n<h2>Civ 6 leader mods<\/h2>\n<h3>JFD\u2019s collection<\/h3>\n<p><img title=\"Civ 6 mods - JFD Leaders\" alt=\"Civ 6 mods - JFD Leaders\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/jfd_leaders.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>The biggest collection of Civ 6 leaders belongs to long-time modder JFD, who\u2019s created and crafted over 30 of the great leaders from history such as Philip II of Macedon, Elizabeth I and Louis XIV, more than capable of facing off against the <a href=\"https:\/\/www.pcgamesn.com\/civilization-vi\/best-civ-6-leader\">best Civ 6 leaders<\/a>. For those wanting some villains to rally against (or play as), there are a few bad eggs in there too, including Hitler, Mussolini, Ivan the Terrible, and Mr. Omelettes-and-Eggs himself, Stalin.<\/p>\n<p>Each new leader has a portrait that blends seamlessly with the bold, vibrant Civ 6 art style, and comes with their own unique units, traits and agendas. As you can guess, old Adolf\u2019s \u2018Liebensraum\u2019 agenda makes things pretty feisty.<\/p>\n<p>Head over to <a href=\"https:\/\/steamcommunity.com\/profiles\/76561198010459974\/myworkshopfiles\/?appid=289070\" target=\"_blank\" rel=\"noopener noreferrer\">JFD\u2019s Steam Workshop<\/a> page to see his full leader collection.<\/p>\n<h2>Civ 6 conversion mods<\/h2>\n<h3>Quo\u2019s Combined Tweaks<\/h3>\n<p><img title=\"Civ 6 mods - Quos Tweaks\" alt=\"Civ 6 mods - Quos Tweaks\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/quos_tweaks.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Anyone who has spent some time on Civ 4 and Civ 5 forums will have heard of the great conversion projects like Rise of Mankind and Realism Invictus for Civ 6, or Vox Populi for Civ 5. These community-created mods spend years revamping the game rules, eventually creating a much deeper experience that borrows the best bits from across the series. <a href=\"https:\/\/forums.civfanatics.com\/resources\/quos-combined-tweaks.25703\/\" target=\"_blank\" rel=\"noopener noreferrer\">Quo\u2019s Combined Tweaks<\/a> takes inspiration from those, revamping hundreds of rules to make for an interesting alternative to Firaxis\u2019 vision.<\/p>\n<p>There are way too many changes to list, but they include faster unit movement, more impactful governments and policies so you can really feel the difference between, say, Fascism and Democracy, and more powerful Wonders (which now steal tiles around them when built, and can be woo City States). Each civ now has several new unique traits too, making games a little more interesting and asymmetrical.<\/p>\n<p>In a nutshell, it\u2019s the most sweeping set of rule changes to Civ 6 available. Mod creator isau recommends using it with AI+, and says that it should be compatible with most unit, UI and map mods. You\u2019ll need the Rise &amp; Fall expansion for Quo\u2019s Combined Tweaks to work.<\/p>\n<h3>Anno Domini<\/h3>\n<p><img title=\"Civ 6 mods - Anno Domini\" alt=\"Civ 6 mods - Anno Domini\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/anno_domini.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>If you\u2019ve always found Civ to be at its most compelling in the earlier eras, when mankind worshipped trees and thought the Earth was just a spinning plate on a giant celestial stick, then you\u2019ll quickly settle into <a href=\"https:\/\/forums.civfanatics.com\/forums\/anno-domini.571\/\" target=\"_blank\" rel=\"noopener noreferrer\">Anno Domini<\/a>. Building on his experience of creating a <a href=\"https:\/\/steamcommunity.com\/workshop\/filedetails\/?id=606794262\" target=\"_blank\" rel=\"noopener noreferrer\">similar mod for Civ 5<\/a>, creator rob8xft mod sweeps us back to an old world of Germanic tribes, druidism, even way back to Minoan civilisations, and keeps us there with a bespoke tech tree that never goes beyond the classical era.<\/p>\n<p>Anno Domini currently has 24 unique civilisations (though you can use some from the main game and JFD\u2019s collection), as well as its own roster of policies and eras that zoom in on the ancient-classical period. It\u2019s the most polished Civ 6 conversion mod to date.<\/p>\n<p>R8XFT himself plays Anno Domini with Religion Expanded, which you\u2019ll find further down this list.<\/p>\n<h2>Civ 6 map mods<\/h2>\n<h3>Yet (not) Another Maps Pack<\/h3>\n<p><img title=\"Civ 6 mods - ynamp\" alt=\"Civ 6 mods - ynamp\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/Civ_6_mods_-_ynamp.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Gedemon\u2019s <a href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=871861883\" target=\"_blank\" rel=\"noopener noreferrer\">YnAMP<\/a> is the most feature-rich map pack for Civ 6. It contains a couple of variations on maps of Europe and Earth, as well as a script that generates Terra, a map split up into the historical Old and New Worlds.<\/p>\n<p>Like in Civs of yore, you can create a game on Terra where all civs start in the Old World, then race to colonise the New World once they can build ships. This setup provides a welcome shift of pace just as a game can start to lull towards the middle stages, and really thrives in multiplayer when you know that all your rivals have their avaricious eyes on the land and plunder of the New World.<\/p>\n<p>This modpack adds a wealth of options for tweaking maps before a new game, as well as Giant, Enormous and Ludicrous map sizes (the latter of which doubles as a stern benchmark tool for your PC\u2019s RAM and CPU).<\/p>\n<p>YnAMP also allocates civs with starting positions based on the kind of environments they were associated with in real life (coastal areas for naval civs, more arid climes for middle-eastern civs, and so on). This feature not only adds a touch of realism, but means that each civ\u2019s starting location is likely to be better suited to their unique traits and buildings.<\/p>\n<h2>Civ 6 Religion mods<\/h2>\n<h3>Religion Expanded<\/h3>\n<p><img title=\"Civ 6 mods - Religion Expanded\" alt=\"Civ 6 mods - Religion Expanded\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/Civ_6_mods_-_Religion_Expanded.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>There have been quite a few attempts to improve upon the flawed religion system in Civ 6, not least of all when Firaxis figured that adding Warrior Monks into the game would spice things up a bit \u2013 which it didn\u2019t. Instead, the best changes to religion in Civ 6 come from the community.<\/p>\n<p>A good place to start is pokiehl\u2019s <a href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=1178185727&amp;searchtext=religion+expanded\" target=\"_blank\" rel=\"noopener noreferrer\">Religion Expanded<\/a>, which adds 43 new religious beliefs, seven new fully-modelled religious buildings, and raises the cap for number of religions in a game from 7 to 16 (on a Huge map). The increased cap means that there will not just be a few faiths battling it out for supremacy, but also smaller ones doing their own thing in far corners of the world.<\/p>\n<h3>Tomatekh\u2019s Historical Religions<\/h3>\n<p><img title=\"Civ 6 mods - Tomatakh\" alt=\"Civ 6 mods - Tomatakh\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/tomatekh.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>You can combine Religion Expanded with <a href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=870850597&amp;searchtext=tomatekh%27s\" target=\"_blank\" rel=\"noopener noreferrer\">Tomatekh\u2019s Historical Religions<\/a>, which adds tens of new religions and icons to the game \u2013 from strange sub-sects of Protestantism to Aboriginal Dreamtime beliefs. Every Civ in the game now has their preferred religions too, so AI will be drawn to the religions their civs are historically associated with.<\/p>\n<h3>Rule with Faith<\/h3>\n<p><a href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=901339687&amp;searchtext=rule+with+faith\" target=\"_blank\" rel=\"noopener noreferrer\">Rule with Faith<\/a> (RwF) is the work of JFD, who sees it as a continuation of his Rise to Power mod for Civ 5. That mod aimed to bring greater political and religious depth to the game, and RwF does the same, adding a new religious policy slot, 16 new policies and three new governments. It\u2019s compatible with both Religion Expanded and Tomatekh\u2019s Historical Religions.<\/p>\n<h2>Civ 6 tweaks<\/h2>\n<h3>Smoother Difficulty<\/h3>\n<p>Making a challenging AI is tricky in strategy games, perhaps more so than in any other genre. Designers usually have to resort to giving the AI artificial bumps to resource production, and even free units, to give a savvy player a run for their money. Unfortunately, the way Civ 6 does this can mean certain death for the player before there\u2019s anything they can physically do to stop it, as the AI swamps you with its five free Warriors (we\u2019re not kidding \u2013 to see how much cheese drips from the AI on Deity difficulty, see our article on <a href=\"https:\/\/www.pcgamesn.com\/civilization-vi\/civ-6-difficulties-speeds-map-types-options\" target=\"_blank\" rel=\"noopener noreferrer\">Civ 6 difficulties<\/a>).<\/p>\n<p>Fortunately, here\u2019s <a href=\"https:\/\/forums.civfanatics.com\/resources\/smoother-difficulty.25423\/\" target=\"_blank\" rel=\"noopener noreferrer\">RushSecond smoother difficulty<\/a> with a solution. This mod removes most of the AI\u2019s extra starting units, and instead dials up its resource generation bonuses. The result should be a smoother challenge on higher difficulties, with the AI keeping pace with you throughout the game despite its technical ineptitude, yet unable to stuff you in the first two dozen turns through an unfair unit advantage.<\/p>\n<h3>AI+<\/h3>\n<p><img title=\"Civ 6 mods - AI+\" alt=\"Civ 6 mods - AI+\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/ai.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>The biggest criticism players have of Civilization 6 is the AI, which has unfortunately taken a few steps back towards the stone age since Civ V (which wasn\u2019t exactly space-age itself).<\/p>\n<p>With Firaxis doing little to fix things on this front, modders have taken the AI burden on themselves. <a href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=878717347&amp;searchtext=\" target=\"_blank\" rel=\"noopener noreferrer\">AI+<\/a> is the best of the lot, making the AI better at balancing their empire\u2019s development with military size, as well as an improving their ability in wartime, as they\u2019ll now send troops to the front lines more aggressively and be bolder in invading your cities in the late game.<\/p>\n<p>Along with the more sweeping changes, there are small bits of fine-tuning like improved settler behaviour, increased reluctance to agree unfavourable peace terms, and tougher city-states that won\u2019t so readily fall into the hands of expansionist civs.<\/p>\n<h3>PhotoKinetik Westeros<\/h3>\n<p><img title=\"Civ 6 mods - Photokinetic Westeros\" alt=\"Civ 6 mods - Photokinetic Westeros\" class=\"lazyload\" src=\"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/photokinetik_westeros.jpg\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>For those of you who didn\u2019t take to Civ 6\u2019s vibrant aesthetic, here\u2019s <a href=\"https:\/\/www.pcgamesn.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"5e6f6c1e\">[email\u00a0protected]<\/a>!n with a reshade mod. Inspired by the Game of Thrones opening sequence, it mutes Civ 6\u2019s colours to give a cooler, more mature washed-out look, as you see above.<\/p>\n<p>You can download <a href=\"https:\/\/forums.civfanatics.com\/resources\/photokinetik.25428\/\" target=\"_blank\" rel=\"noopener noreferrer\">PhotoKinetik Westeros<\/a> here. Note that it has some more complicated installation instructions than most other mods on this list as it relies on <a href=\"https:\/\/reshade.me\/\" target=\"_blank\" rel=\"noopener noreferrer\">ReShade<\/a>, a separate post-processing application, to achieve its effects.<\/p>\n<h3>Quick Start<\/h3>\n<p>And so, like intrepid explorers circumnavigating the globe, we\u2019re going to end our journey at the start. The start of the game, that is. If you\u2019re tired of Civ 6\u2019s multiple boot screens perennially reminding you who made the game and which GPU manufacturer it\u2019s in bed with, you can skip it all with <a href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=1178185727\" target=\"_blank\" rel=\"noopener noreferrer\">Quick Start<\/a>. Now, if we could just find a way replicate this mod for every game in existence\u2026<\/p>\n<p>These are just a few of the thousands of Civ 6 mods available, which you can peruse yourself at <a href=\"https:\/\/forums.civfanatics.com\/forums\/civ6-creation-customization.541\/\" target=\"_blank\" rel=\"noopener noreferrer\">Civ Fanatics<\/a>, <a href=\"https:\/\/steamcommunity.com\/app\/289070\/workshop\/\" target=\"_blank\" rel=\"noopener noreferrer\">Steam Workshop<\/a> and <a href=\"https:\/\/www.nexusmods.com\/civilisationvi\/\" target=\"_blank\" rel=\"noopener noreferrer\">Nexus Mods<\/a>. There are more outlandish ones out there \u2013 Game of Thrones and Lord of the Rings spring to mind \u2013 but they missed the cut because they felt feature-thin and unfinished at the time of writing. The fact is that for now, the modding community is limited by not having access to the DLL source code. When that came out for Civ 4 and 5 (some two years after each game\u2019s launch), the modding scenes really took off, as did big overhaul projects like A New Dawn and Vox Populi.<\/p>\n<p>If that\u2019s anything to go by, we should expect to see the DLL code for Civ 6 released soon-ish, and only then will we see the game\u2019s modding potential unleashed. The best, we hope, is yet to come.<\/p>","created_at":"2020-09-18T03:03:40+0000","updated_at":"2020-09-18T03:03:40+0000","published_at":"2018-06-28T06:30:00+0000","published_by":["Robert Zak"],"annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":9,"domain_name":"www.pcgamesn.com","preview_picture":"https:\/\/www.pcgamesn.com\/wp-content\/uploads\/legacy\/Upcoming_PC_games_Civilization_VI.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","auth","jwt"],"is_public":false,"id":1201,"title":"Using JSON Web Tokens (JWT) with Clojure - Brad Cypert","url":"https:\/\/www.bradcypert.com\/using-json-web-tokens-with-clojure\/","content":"<p id=\"post-modified-info\">Last Updated on February 20, 2019 by <a href=\"https:\/\/www.bradcypert.com\/using-json-web-tokens-with-clojure\/\" target=\"_blank\" class=\"last-modified-author\">Brad Cypert<\/a><\/p><p>Authentication is easy. Good authentication is difficult. We\u2019re at a point in the world where our users expect to be able to sign up for your service or app, and receive a more personalized experience. If you can\u2019t validate who the user is, you can\u2019t achieve this.<\/p>\n<p>In the past, people have leveraged sessions, cookies, sending the user\u2019s email and password on each request (<strong>please don\u2019t ever do this<\/strong>), and many other techniques to validate that a user is who they say they are. A lot of these techniques worked relatively well for the time, but they all have their flaws \u2013 Cookies expire and are tied the browser, Sessions don\u2019t scale well, sending email\/password on every request violates your user\u2019s trust in you to keep their information secure.<\/p>\n<p>Enter JSON Web Tokens (JWT). After doing much research on them, I believe they provide a very interested take on user authentication, and want to share with you how you can leverage Buddy to authenticate a user via JWT in Clojure.<\/p>\n<h2 id=\"whatarejsonwebtokens\">What are JSON Web Tokens?<\/h2>\n<p>JSON Web Tokens are used to send information that can be verified and trusted by means of a digital signature. They are built from a URL-safe JSON object, which is then cryptographically signed to verify its authenticity. Most tokens are also encrypted if the payload contains sensitive information.<\/p>\n<h2 id=\"atahighlevel\">At a High Level<\/h2>\n<p>Generally, JWT are used as the authorization header on incoming requests. The flow for this could look like the following:<\/p>\n<ol><li>A user goes to your webapp, and finds the login form.<\/li>\n<li>That user provides the correct details to the login form.<\/li>\n<li>A request is made to the backend service, and that service validates that the user is who they claim to be (generally though checking the database for a username\/password match)<\/li>\n<li>If the user is who they claim to be, we generate a JWT and send it back to the client.<\/li>\n<li>The client then passes this JWT token in as an authorization header on all subsequent requests.<\/li>\n<\/ol><h2 id=\"inclojure\">In Clojure<\/h2>\n<p>Let\u2019s get started implementing this in Clojure. We\u2019re going to depend on the Buddy library \u2013 a best-practices library on security, signing, and authentication. In fact, <a href=\"https:\/\/github.com\/funcool\/buddy\">Buddy is a library of libraries<\/a>, and we\u2019re only going to add the <code>buddy-sign<\/code> library to our project, as that\u2019s all we need for this.<\/p>\n<p><strong>Note:<\/strong> This tutorial assumes you\u2019re using the latest version of Clojure at the time of writing (1.8), Leiningen (v2) and have an existing project that you\u2019re adding JWT into.<\/p>\n<p>In your <code>project.clj<\/code> file, add the following dependency:<\/p>\n<pre>[buddy\/buddy-sign \"1.1.0\"]<\/pre>\n<p>Next, in our source folder, we can create a new file for user-auth.<\/p>\n<pre>touch \/src\/modules\/auth.clj  <\/pre>\n<p>In your favorite text editor (don\u2019t worry, I won\u2019t try to sell you on Vim in this tutorial), open the file we just created. Let\u2019s go ahead and define a namespace and the imports for that namespace.<\/p>\n<pre>(ns app.modules.auth\n  (:require [buddy.sign.jwt :as jwt]))<\/pre>\n<p>For the sake of simplicity, we\u2019re going to assume there is only one user of this app, and we\u2019re hard-coding it into the auth module (so we can focus on JWT, and not reading\/writing to the db, encrypted passwords, etc.), so let\u2019s go ahead and define that too.<\/p>\n<pre>(defonce user {:email \"foo@bar.com\" :password \"abc123\"})<\/pre>\n<h2>Generating A Secret via Python<\/h2>\n<p>Before we can write the function to generate the signature, we need to generate a secret key and bind that in Clojure. There are many ways to generate a secret key, but for simplicity\u2019s sake, we\u2019re just going to generate a quick key via Python. If you have python installed, you can type <code>python<\/code> into your terminal to be taken into Python\u2019s REPL and then type the following lines.<\/p>\n<pre>import os  \nos.urandom(24).encode('hex')  <\/pre>\n<p>The string that was generated for me is <code>86bae26023208e57a5880d5ad644143c567fc57baaf5a942<\/code>, but your\u2019s will be different. Regardless, copy this value to the clipboard and let\u2019s add it to our Clojure file. It\u2019s worth mentioning you can also leverage other buddy libraries to help you generate your secret key.<\/p>\n<pre>(defonce secret \"86bae26023208e57a5880d5ad644143c567fc57baaf5a942\")<\/pre>\n<p>Now, we can finally write the function to generate the signature. First, we we want to only generate a signature if the user credentials match the ones that we provided above. If it doesn\u2019t match, we\u2019ll just return an empty string.<\/p>\n<pre>(defn generate-signature [email password]\n  (let [credentials {:email email :password password}]\n    (cond\n      (= user credentials) (jwt\/sign {:user email} secret)\n      :else \"\")))<\/pre>\n<p>Next, we can write a function to validate that the signature is a real signature.<\/p>\n<pre>(defn unsign-token [token]\n  (jwt\/unsign token secret))<\/pre>\n<p>When we call <code>generate-signature<\/code> with <code>foo@bar.com<\/code> and <code>abc123<\/code>, we\u2019ll get a token back that looks something like this \u2013 <code>eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VyIjoyfQ.lj6O4Na78SkhZbBSKZktT0Mufqg_L5ZmRe6aRa1bG5U<\/code>.<\/p>\n<p>When we call <code>unsign-token<\/code> and pass in that token, you\u2019ll be given back a map like so <code>{:user \"foo@bar.com\"}<\/code>.<\/p>\n<p>Congratulations! You\u2019ve got the basics for working with JWT!<\/p>\n<h2 id=\"nextsteps\">Next Steps<\/h2>\n<p>There\u2019s plenty of places to go from here, such as integrating this into a web service, or encrypting the token as well (you <strong>really should<\/strong> do this). Additionally, consider what other information you can load into your json web token \u2013 expiration time, for example.<\/p>\n<p>Want to learn about other ways to use Clojure? You can <a href=\"https:\/\/www.bradcypert.com\/tag\/clojure\/\">find more of my posts on the language here<\/a>!<\/p>","created_at":"2020-09-12T20:20:19+0000","updated_at":"2020-09-12T20:20:24+0000","published_at":"2016-06-23T06:46:55+0000","published_by":["Posted June 23, 2016 by Brad Cypert"],"annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"www.bradcypert.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","auth"],"is_public":false,"id":1200,"title":"Easy custom auth in Clojure with Friend","url":"https:\/\/adambard.com\/blog\/easy-auth-with-friend\/","content":"<p>I\u2019ve used <a href=\"https:\/\/github.com\/cemerick\/friend\">friend<\/a> to provide auth in my projects a few times, and considered it many more before resorting to my own hand-rolled business. Most of the reason for this is that is just seems so <em>complicated<\/em>. Workflows? Credentials? I just want to check a password and stash a user object in the session, end of story. After I did some investigating, though, it turns out that you can usefully deploy friend for even very simple workflows, if you understand how it works.<\/p><p>Friend has its own functions corresponding to those two points: <code>authenticate<\/code> is a middleware that wraps a Ring handler, while <code>authorize<\/code> can be deployed just about anywhere in your code to handle that aspect of your program\u2019s auth. We\u2019ll examine <code>authenticate<\/code> in this post, since the other is pretty simple to use. An illustrative way to learn about how those authentication works in friend is to write a small toy auth system, using none of friend\u2019s built-ins and instead writing our own.<\/p><h2>Before that, some glossary<\/h2><p><strong>Authentication Map<\/strong>: A hash-map of information about an authenticated user which will be accessible from the Ring request. The <code>:identity<\/code> key is required, and the <code>:roles<\/code> key is strongly recommended<\/p><p><strong>Workflow<\/strong>: A function that accepts a request map and returns either an authentication map (for a newly-authenticated user) or <code>nil<\/code> (if there is no change in the authentication state). It may also return a ring response that will be returned immediately, shortcutting the rest of the request.<\/p><p><strong>Credential Function<\/strong>: A function that accepts a <em>workflow<\/em>-dependent set of arguments and returns an authentication map.<\/p><h2>Speak \u201cfriend\u201d and enter<\/h2><p>We\u2019ll write a really dumb auth system; instead of a username and password, it will look for a parameter called <code>speak<\/code> in the incoming request, and authenticate the user if <code>speak=friend<\/code>. To do that, we\u2019ll need to define a <em>workflow<\/em>, which we\u2019ll use with friend\u2019s <code>authenticate<\/code> function to make our authentication middleware.<\/p><p>A workflow is just a function with a specific signature and behavior, which is documented in this hopefully-not-copyrighted diagram:<\/p><p><img src=\"https:\/\/adambard.com\/blog\/img\/workflow.png\" alt=\"Workflow\" referrerpolicy=\"no-referrer\" \/><\/p><p>In short, the workflow accepts a ring request, and can return one of three things:<\/p><ul><li>A ring response, which will be returned directly.<\/li>\n<li><code>nil<\/code>, which will result in other workflows being checked (or a 403 returned if none remain).<\/li>\n<li>An <em>authentication map<\/em>, which means authentication was successful.<\/li>\n<\/ul><p>The term \u201cauthentication map\u201d is ill-defined, but turns out to be any map with the key <code>:identity<\/code> and with the metadata <code>{:type :cemerick.friend\/auth}<\/code>. Handily, the function <code>cemerick.friend.workflows\/make-auth<\/code> will add this metadata for you.<\/p><p>So let\u2019s just start with a really simple workflow:<\/p><div class=\"highlight\"><pre>(defnfun-workflow[req](let[speak(get-inreq[:params:speak])](when(=speak\"friend\")(make-auth{:identity\"friend\":roles#{::user}}))))<\/pre><\/div><p>When a ring request comes in, <code>friend\/authenticate<\/code> will pass it along to this workflow. The workflow will check the value of the \u201cspeak\u201d parameter in the params map. If \u201cspeak\u201d == \u201cfriend\u201d, it calls <code>make-auth<\/code> on the authentication map, which annotates it with the metadata necessary for friend to recognize it as such.<\/p><p>To hook up your workflow, you\u2019ll need to pass it to the :workflows option in <code>friend\/authenticate<\/code>. Here\u2019s a small but complete example of an app using this auth workflow:<\/p><div class=\"highlight\"><pre>(require'[compojure.core:refer[defroutesGETPOST]]'[cemerick.friend:asfriend]'[cemerick.friend.workflows:refer[make-auth]]'[ring.middleware.session:refer[wrap-session]]'[ring.middleware.params:refer[wrap-params]]'[ring.middleware.keyword-params:refer[wrap-keyword-params]]'[ring.adapter.jetty:refer[run-jetty]])(defroutesapp-routes(GET\"\/\"[]\"Hello everyone &lt;form action=\\\"logout\\\" method=\\\"post\\\"&gt;&lt;button&gt;Submit&lt;\/button&gt;&lt;\/form&gt;\")(GET\"\/authorized\"[](friend\/authorize#{::user}\"Hello authorized\"))(friend\/logout(POST\"\/logout\"[]\"logging out\")))(defnfun-workflow[req](let[speak(get-inreq[:params:speak])](when(=speak\"friend\")(make-auth{:identity\"friend\":roles#{::user}}))))(defapp(-&gt;app-routes(friend\/authenticate{:workflows[fun-workflow]})(wrap-keyword-params)(wrap-params)(wrap-session)))(defn-main[](run-jetty#'app{:port8080}))<\/pre><\/div><p>If you fire up a server, here\u2019s what you will observe:<\/p><ul><li>If you go to <code>\/<\/code>, you\u2019ll see the page regardless of whether you logged in.<\/li>\n<li>If you go to <code>\/authorized<\/code>, you\u2019ll be redirected to <code>\/login<\/code> (the default login url, which we never changed and don\u2019t need).<\/li>\n<li>If you go to <code>\/authorized?speak=friend<\/code>, you\u2019ll end up at <code>\/authorized<\/code>. You can now freely visit <code>\/authorized<\/code> without adding the query params.<\/li>\n<li>If you click the button on <code>\/<\/code> which POSTs to <code>\/logout<\/code>, you\u2019ll be logged out and get the redirect if you try to go to <code>\/authorized<\/code>.<\/li>\n<\/ul><p>Note that none of this will work without the three ring middlewares I added: <code>wrap-params<\/code>, <code>wrap-keyword-params<\/code>, and <code>wrap-session<\/code>.<\/p><h2>What\u2019s the password?<\/h2><p>You can definitely write your own workflows this way and have no problems. However, it is recommended for configurability that you split the part of your workflow that checks the auth and retrieves any extra information into a separate <em>credential function<\/em> like a good lass or laddie.<\/p><p>To use a credential-fn, you just need to provide one to the middleware and make sure the workflow uses it. Here\u2019s how you might break up the above example into a workflow and a credential function:<\/p><div class=\"highlight\"><pre>(defnfun-workflow[req](let[speak(get-inreq[:params:speak])credential-fn(get-inreq[::friend\/auth-config:credential-fn])](make-auth(credential-fnspeak))))(defnfun-credential-fn[word](if(=word\"friend\"){:identityword:roles#{::user}}));And in the middleware...(friend\/authenticate{:workflows[fun-workflow]:credential-fnfun-credential-fn})<\/pre><\/div><p>Make a special note that <strong>returning nil is the default<\/strong>. The bare <code>if<\/code> in the credential-fn returns <code>nil<\/code> if the word is incorrect, which in turn is returned from the workflow, which means that <em>no change in authentication state occurs<\/em>.<\/p><p>Now, let\u2019s imagine that instead of a hardcoded magic word, we want auth that will:<\/p><ol><li>Only be processed on the login page on a POST request,<\/li>\n<li>Use a username and password pair, and<\/li>\n<li>Fetch the user object from the database and provide that to every function, instead of just the bare credentials.<\/li>\n<\/ol><p>Here\u2019s a homemade spin on that, using a magical <code>db<\/code> namespace whose functions are already written:<\/p><div class=\"highlight\"><pre>(defndo-login[req](let[credential-fn(get-inreq[::friend\/auth-config:credential-fn])](make-auth(credential-fn(select-keys(:paramsreq)[:username:password])))))(defnpassword-workflow[req](when(and(=(:request-methodreq):post)(=(:urireq)\"\/login\"))(do-loginreq)))(defnpassword-credential-fn[{:keys[usernamepassword]:ascreds}](when-let[user(get-user-by-usernameusername)](when(=(:hashed_passworduser)(some-strong-hashpassword)){:identity(:iduser):roles#{::user}:useruser})));And in the middleware...(friend\/authenticate{:workflows[password-workflow]:credential-fnpassword-credential-fn})<\/pre><\/div><p>It\u2019s really not a lot more complicated than the other one, is it?<\/p><p>You can replace <code>password-workflow<\/code> with the <code>(interactive-form)<\/code> workflow that comes with friend, and it still works, since \u201c\/login\u201d is the default login-uri for friend. I also made sure to match the <code>credential-fn<\/code> signature.<\/p><p>You can also offload most of the credential-fn to the built-in <code>bcrypt-credential-fn<\/code>, although this requires that your user map have a <code>:password<\/code> key that is encrypted with the bcrypt hash function also included with friend. The <code>cemerick.friend.credentials<\/code> contains the functions <code>hash-bcrypt<\/code>, <code>bcrypt-verify<\/code> for those purposes as well.<\/p><h2>Other auth schemes<\/h2><p>The intent of the credential-fn + workflow model is to be as flexible as possible, and since friend is a middleware, you can even completely abstract workflows like oauth. I mean, it\u2019s still a bit painful to test, but at least the workflow isn\u2019t too bad. Here\u2019s what I came up with writing a friend wrapper around <a href=\"https:\/\/github.com\/mattrepl\/clj-oauth\">clj-oauth<\/a> for twitter:<\/p><div class=\"highlight\"><pre>(require'[oauth.client:asoauth])(defconsumer(oauth\/make-consumer(:twitter-consumer-keyenv)(:twitter-consumer-secretenv)\"https:\/\/api.twitter.com\/oauth\/request_token\"\"https:\/\/api.twitter.com\/oauth\/access_token\"\"https:\/\/api.twitter.com\/oauth\/authorize\":hmac-sha1))(defntwitter-auth-workflow[&amp;{:keys[oauth-callback-uri]}](fn[req](let[credential-fn(get-inreq[::friend\/auth-config:credential-fn])](cond;; Login url -- redirect to twitter approval page(=(:urireq)(get-inreq[::friend\/auth-config:login-uri]))(let[request-token(oauth\/request-tokenconsumer(:urioauth-callback-uri))approval-url(oauth\/user-approval-uriconsumer(:oauth_tokenrequest-token))resp(resp\/redirectapproval-url)];; Stash the request-token in the session for later(update-inresp[:session]assoc:oauth-request-tokenrequest-token)); Oauth callback(=(:pathoauth-callback-uri)(:urireq))(when-let[request-token(get-inreq[:session:oauth-request-token])](when-let[access-token(oauth\/access-tokenconsumerrequest-token(get-inreq[:params:oauth_verifier]))](make-auth(credential-fnaccess-token))))))));...(defoauth-callback-uri{:url\"http:\/\/localhost\/oauth\/callback\":path\"\/oauth\/callback\"})(friend\/authenticate{:workflows[(twitter-auth-workflow:oauth-callback-urioauth-callback-uri)]:credential-fn#(hash-map:identity%:roles#{::user})})<\/pre><\/div><p>If you\u2019ve ever tried to write any oauth-based auth system, you\u2019d think that\u2019s not too bad! Since the authentication middleware receives the whole request, it can do the oauth redirect on the login url and handle the callback as well, without getting the way of everything else. After all that, friend handles authorization for you, without having to go through it again!<\/p><p>That\u2019s why friend is worth using: authentication is as hard a problem as it ever was, but at least friend solves authorization for you with almost no effort, and does a better job than if you tried to do it yourself.<\/p>","created_at":"2020-09-12T19:55:12+0000","updated_at":"2020-09-12T19:55:16+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"adambard.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1199,"title":"Archiving URLs \u00b7 Gwern.net","url":"https:\/\/www.gwern.net\/Archiving-URLs","content":"<div id=\"abstract\"><blockquote>\n<p>Links on the Internet last forever or a year, whichever comes first. This is a major problem for anyone serious about writing with good references, as link rot will cripple several percent of all links each year, and compounding.<\/p>\n<p>To deal with link rot, I present my multi-pronged archival strategy using a combination of scripts, daemons, and Internet archival services: URLs are regularly dumped from both my web browser\u2019s daily browsing and my website pages into an archival daemon I wrote, which pre-emptively downloads copies locally and attempts to archive them in the Internet Archive. This ensures a copy will be available indefinitely from one of several sources. Link rot is then detected by regular runs of <code>linkchecker<\/code>, and any newly dead links can be immediately checked for alternative locations, or restored from one of the archive sources.<\/p>\n<p>As an additional flourish, my local archives are  in case forgery is a concern, and I demonstrate a simple compression trick for  such as crawls (particularly useful for repeated crawls such as my ).<\/p>\n<\/blockquote><\/div><p>Given my interest in <a href=\"https:\/\/www.gwern.net\/About#long-content\">long term content<\/a> and extensive linking,  is an issue of deep concern to me. I need backups not just for my files<a href=\"#fn1\" class=\"footnote-ref\" id=\"fnref1\" role=\"doc-noteref\"><sup>1<\/sup><\/a>, but for the web pages I read and use\u2014they\u2019re all part of my . It\u2019s not much good to have an extensive essay on some topic where half the links are dead and the reader can neither verify my claims nor get context for my claims.<\/p><section id=\"link-rot\" class=\"level1\"><div class=\"epigraph\"><blockquote>\n<p>\u201cDecay is inherent in all compound things. Work out your own salvation with diligence.\u201d<\/p>\n<p>Last words of the Buddha<\/p>\n<\/blockquote><\/div><p>The dimension of digital decay is dismal and distressing. :<\/p><blockquote>\n<p>In a 2003 experiment, <a href=\"https:\/\/www.gwern.net\/docs\/www\/www2003.org\/c220b6d34d074987565bf1515cae7d8117131926.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www2003.org\/cdrom\/papers\/refereed\/p097\/P97%20sources\/p97-fetterly.html\" title=\"(Original URL: http:\/\/www2003.org\/cdrom\/papers\/refereed\/p097\/P97%20sources\/p97-fetterly.html )\">Fetterly et al.<\/a> discovered that about one link out of every 200 disappeared each week from the Internet.  discovered that half of the URLs cited in  articles were no longer accessible 10 years after publication [the irony!], and other studies have shown link rot in academic literature to be even worse (<a href=\"https:\/\/www.spinellis.gr\/pubs\/jrnl\/2003-CACM-URLcite\/html\/urlcite.pdf\" title=\"The decay and failures of web references: Attempting to determine how quickly archival information becomes outdated\">Spinellis, 2003<\/a>, <a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/download?doi=10.1.1.97.9695&amp;rep=rep1&amp;type=pdf\">Lawrence et al., 2001<\/a>). <a href=\"https:\/\/www.gwern.net\/docs\/www\/www.dlib.org\/8a11e08c0ffba13eb4220502190eb5d0c7cd9fb5.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www.dlib.org\/dlib\/january02\/nelson\/01nelson.html\" title=\"(Original URL: http:\/\/www.dlib.org\/dlib\/january02\/nelson\/01nelson.html )\">Nelson and Allen (2002)<\/a> examined link rot in digital libraries and found that about 3% of the objects were no longer accessible after one year.<\/p>\n<\/blockquote><p> remarks that one friend experienced 50% linkrot in one of his pages over less than 9 years (not that the situation was any better <a href=\"https:\/\/www.gwern.net\/docs\/www\/www.pantos.org\/10723128e7820b8612d887af0fd0052d7ca41741.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www.pantos.org\/atw\/35654.html\" title=\"(Original URL: http:\/\/www.pantos.org\/atw\/35654.html )\">in 1998<\/a>), and that his own blog posts link to news articles that go dead in days<a href=\"#fn2\" class=\"footnote-ref\" id=\"fnref2\" role=\"doc-noteref\"><sup>2<\/sup><\/a>; <a href=\"https:\/\/notes.pinboard.in\/u:vitorio\/05dec9f04909d9b6edff\">Vitorio checks bookmarks from 1997<\/a>, finding that hand-checking indicates a total link rot of 91% with only half of the dead available in sources like the Internet Archive; <a href=\"https:\/\/motherboard.vice.com\/en_us\/article\/paabgg\/i-bought-a-book-about-the-internet-from-1994-and-none-of-the-links-worked\" title=\"I Bought a Book About the Internet From 1994 and None of the Links Worked\">Ernie Smith found 1 semi-working link<\/a> in a 1994 book about the Internet; the  itself has estimated the average lifespan of a Web page at <a href=\"https:\/\/web.archive.org\/web\/20071019225237\/http:\/\/www.wired.com\/culture\/lifestyle\/news\/2001\/10\/47894\" title=\"Wayback Goes Way Back on Web\">100 days<\/a>. A <em> study looked at articles in prestigious journals; they didn\u2019t use many Internet links, but when they did, 2 years later ~13% were dead<a href=\"#fn3\" class=\"footnote-ref\" id=\"fnref3\" role=\"doc-noteref\"><sup>3<\/sup><\/a>. The French company Linterweb studied external links on the  before setting up <a href=\"http:\/\/www.wikiwix.com\/\">their cache<\/a> of French external links, and found\u2014back in 2008\u2014already <a href=\"http:\/\/fr.wikipedia.org\/wiki\/Utilisateur:Pmartin\/Cache\">5% were dead<\/a>. (The English Wikipedia has seen a 2010-January 2011 spike from a few thousand dead links to <a href=\"https:\/\/en.wikipedia.org\/wiki\/File%3AArticles-w-Dead-Links-Jan-2011.png\">~110,000<\/a> out of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wikipedia_talk%3AWikiProject_External_links%2FWebcitebot2#Summary\">~17.5m live links<\/a>.) <a href=\"https:\/\/www.gwern.net\/docs\/www\/lil.law.harvard.edu\/a3c8780e0c37fc6965c2d66880a7d58f09437d30.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/lil.law.harvard.edu\/blog\/2017\/07\/21\/a-million-squandered-the-million-dollar-homepage-as-a-decaying-digital-artifact\/\" title=\"A Million Squandered: The 'Million Dollar Homepage' as a Decaying Digital Artifact (Original URL: https:\/\/lil.law.harvard.edu\/blog\/2017\/07\/21\/a-million-squandered-the-million-dollar-homepage-as-a-decaying-digital-artifact\/ )\">A followup check<\/a> of the viral , where it cost up to $38k to insert a link by the last one on January 2006, found that a decade later in 2017, at least half the links were dead or squatted.<a href=\"#fn4\" class=\"footnote-ref\" id=\"fnref4\" role=\"doc-noteref\"><sup>4<\/sup><\/a> Bookmarking website , which provides some archiving services, <a href=\"https:\/\/twitter.com\/pinboard\/status\/501406303747457024\" title=\"Preliminary link rot results are that 25% of bookmarks saved in 2009 are dead, and 17% saved in 2011 are gone. Spreadsheet to follow\">noted in August 2014<\/a> that 17% of 3-year-old links &amp; 25% of 5-year-old links were dead. The dismal  <a href=\"http:\/\/jnci.oxfordjournals.org\/content\/96\/12\/969.full\" title=\"'Internet Citations in Oncology Journals: A Vanishing Resource?', Hester et al 2004\">just<\/a> <a href=\"https:\/\/www.gwern.net\/docs\/linkrot\/2007-dimitrova.pdf\" title=\"'The half-life of internet references cited in communication journals', Dimitrova &amp; Bugeja 2007\">go<\/a> <a href=\"https:\/\/www.gwern.net\/docs\/linkrot\/2008-wren.pdf\" title=\"'URL decay in MEDLINE---a 4-year follow-up study', Wren 2008\">on<\/a>   <a href=\"https:\/\/www.gwern.net\/docs\/www\/pdfs.semanticscholar.org\/a93d671d3fa25a19baabb16cdde8f98951adb260.pdf\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/pdfs.semanticscholar.org\/1ae0\/d55d0af02cdd77a3147c4f652bf4e3749194.pdf\" title=\"'Something Rotten in the State of Legal Citation: the Life Span of a United States Supreme Court Citation Containing an Internet Link (1996-2010)', Liebler &amp; Liebert 2013 (Original URL: https:\/\/pdfs.semanticscholar.org\/1ae0\/d55d0af02cdd77a3147c4f652bf4e3749194.pdf )\">and<\/a> <a href=\"https:\/\/www.gwern.net\/docs\/www\/www.fasebj.org\/9a57fd429260c51e1d6eb22da088f4d4461000dd.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www.fasebj.org\/content\/19\/14\/1943.full\" title=\"'Unavailability of online supplementary scientific information from articles published in major journals', Evangelou et al 2005 (Original URL: http:\/\/www.fasebj.org\/content\/19\/14\/1943.full )\">on<\/a> (<a href=\"https:\/\/www.gwern.net\/docs\/linkrot\/2012-moghaddam.pdf\" title=\"'Availability and Half-life of Web References Cited in Information Research Journal: A Citation Study', Moghaddam et al 2012\">and<\/a> <a href=\"https:\/\/www.gwern.net\/docs\/www\/cdn.harvardlawreview.org\/462cd5385125d17e90515a9fec0d463051375267.pdf\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/cdn.harvardlawreview.org\/wp-content\/uploads\/2014\/03\/forvol127_zittrain.pdf\" title=\"'Perma: Scoping and Addressing the Problem of Link and Reference Rot in Legal Citations', Zittrain &amp; Albert 2013 (Original URL: http:\/\/cdn.harvardlawreview.org\/wp-content\/uploads\/2014\/03\/forvol127_zittrain.pdf )\">on<\/a>). Even in a highly stable, funded, curated environment, link rot happens anyway. For example, about  were gone within a year (even though Twitter is\u2014currently\u2014still around). And sometimes they just get quietly lost, like when MySpace admitted it lost <a href=\"https:\/\/www.gwern.net\/docs\/www\/old.reddit.com\/3fa6bd447f8e0c36e551dcae4d8bc6568b019022.html#t1_e1n8hzn\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/old.reddit.com\/r\/techsupport\/comments\/7uiv8b\/myspace_player_wont_play_songs_and_i_want_to\/#t1_e1n8hzn\" title=\"(Original URL: https:\/\/old.reddit.com\/r\/techsupport\/comments\/7uiv8b\/myspace_player_wont_play_songs_and_i_want_to\/#t1_e1n8hzn )\">all music worldwide uploaded 2003\u20132015<\/a>, euphemistically describing the mass deletion as <a href=\"https:\/\/www.gwern.net\/docs\/www\/help.myspace.com\/e92581637046e3b764df3e2d03b4b3a00121bff6.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/help.myspace.com\/hc\/en-us\/articles\/202233310-Where-Is-All-My-Old-Stuff-\" title=\"(Original URL: https:\/\/help.myspace.com\/hc\/en-us\/articles\/202233310-Where-Is-All-My-Old-Stuff- )\">\u201cWe completely rebuilt Myspace and decided to move over some of your content from the old MySpace\u201d<\/a> (only some of the 2008\u20132010 MySpace music could be rescued &amp; put on the IA by <a href=\"https:\/\/archive.org\/details\/myspace_dragon_hoard_2010\">\u201can anonymous academic study\u201d<\/a>).<\/em><\/p><p>My specific target date is 2070, 60 years from now. As of 10 March 2011, <code>gwern.net<\/code> has around 6800 external links (with around 2200 to non-Wikipedia websites)<a href=\"#fn5\" class=\"footnote-ref\" id=\"fnref5\" role=\"doc-noteref\"><sup>5<\/sup><\/a>. Even at the lowest estimate of 3% annual linkrot, few will survive to 2070. If each link has a 97% chance of surviving each year, then the chance a link will be alive in 2070 is 0.972070\u22122011\u22480.16 (or to put it another way, an 84% chance any given link <em>will<\/em> die). The 95%  for such a  says that of the 2200 non-Wikipedia links, ~336\u2013394 will survive to 2070<a href=\"#fn6\" class=\"footnote-ref\" id=\"fnref6\" role=\"doc-noteref\"><sup>6<\/sup><\/a>. If we try to predict using a more reasonable estimate of 50% linkrot, then an average of 0 links will survive (0.502070\u22122011\u22c52200=1.735\u22c510\u221216\u22c52200\u22430). It would be a good idea to simply assume that <em>no<\/em> link will survive.<\/p><p>With that in mind, one can consider remedies. (If we lie to ourselves and say it won\u2019t be a problem in the future, then we guarantee that it <em>will<\/em> be a problem. <a href=\"http:\/\/wiki.lesswrong.com\/wiki\/Litany_of_Gendlin\">\u201cPeople can stand what is true, for they are already enduring it.\u201d<\/a>)<\/p><p>If you want to <em>pre-emptively<\/em> archive a specific page, that is easy: go to the IA and archive it, or in your web browser, print a PDF of it, or for more complex pages, use a browser plugin (eg for Firefox, <a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/mozilla-archive-format\/\">Mozilla Archive Format<\/a> or <a href=\"https:\/\/github.com\/danny0838\/firefox-scrapbook\/wiki\/Features\">ScrapBook<\/a>). But I find I visit and link so many web pages that I rarely think in advance of link rot to save a copy; what I need is a more systematic approach to detect and create links for all web pages I might need.<\/p><\/section><section id=\"detection\" class=\"level1\"><div class=\"epigraph\"><blockquote>\n<p>\u201cWith every new spring<br \/>the blossoms speak not a word<br \/>yet expound the Law\u2014<br \/>knowing what is at its heart<br \/>by the scattering storm winds.\u201d<\/p>\n<p>Sh\u014dtetsu<a href=\"#fn7\" class=\"footnote-ref\" id=\"fnref7\" role=\"doc-noteref\"><sup>7<\/sup><\/a><\/p>\n<\/blockquote><\/div><p>The first remedy is to learn about broken links as soon as they happen, which allows one to react quickly and scrape archives or search engine caches (<a href=\"https:\/\/www.gwern.net\/docs\/www\/www.cs.odu.edu\/5d0f24ecdfa4b26f35532b92ddf6fc43fdf2b4c3.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www.cs.odu.edu\/~fmccown\/research\/lazy\/\" title=\"(Original URL: http:\/\/www.cs.odu.edu\/~fmccown\/research\/lazy\/ )\">\u2018lazy preservation\u2019<\/a>). I currently use <a href=\"http:\/\/wummel.github.io\/linkchecker\/\"><code>linkchecker<\/code><\/a> to spider gwern.net looking for broken links. <code>linkchecker<\/code> is run in a  job like so:<\/p><div class=\"sourceCode\" id=\"cb2\"><pre class=\"sourceCode bash\">@monthly linkchecker --check-extern --timeout=35 --no-warnings --file-output=html \\\n                      --ignore-url=^mailto --ignore-url=^irc --ignore-url=http:\/\/.*\\.onion \\\n                      --ignore-url=paypal.com --ignore-url=web.archive.org \\\n                     https:\/\/www.gwern.net<\/pre><\/div><p>Just this command would turn up many false positives. For example, there would be several hundred warnings on Wikipedia links because I link to redirects; and <code>linkchecker<\/code> respects s which forbid it to check liveness, but emits a warning about this. These can be suppressed by editing <code>~\/.linkchecker\/linkcheckerrc<\/code> to say <code>ignorewarnings=http-moved-permanent,http-robots-denied<\/code> (the available warning classes are listed in <code>linkchecker -h<\/code>).<\/p><p>The quicker you know about a dead link, the sooner you can look for replacements or its new home.<\/p><\/section><section id=\"prevention\" class=\"level1\"><section id=\"remote-caching\" class=\"level2\"><div class=\"epigraph\"><blockquote>\n<p>\u201cAnything you post on the internet will be there as long as it\u2019s embarrassing and gone as soon as it would be useful.\u201d<\/p>\n<p><a href=\"https:\/\/news.ycombinator.com\/item?id=21270467\">taejo<\/a><\/p>\n<\/blockquote><\/div><p>We can ask a third party to keep a cache for us. There are several  possibilities:<\/p><ol type=\"1\"><li>the <\/li>\n<li>\n<\/li><li><a href=\"http:\/\/perma.cc\/\">Perma.cc<\/a> (<a href=\"https:\/\/www.gwern.net\/docs\/www\/blogs.harvard.edu\/87af35282135d3db4de0da25f3644f286157f814.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/blogs.harvard.edu\/perma\/2019\/01\/07\/introducing-individual-account-subscription-tiers-for-perma\/\" title=\"https:\/\/blogs.harvard.edu\/perma\/2019\/01\/07\/introducing-individual-account-subscription-tiers-for-perma\/ (Original URL: https:\/\/blogs.harvard.edu\/perma\/2019\/01\/07\/introducing-individual-account-subscription-tiers-for-perma\/ )\">highly limited<\/a>)<\/li>\n<li>Linterweb\u2019s WikiWix<a href=\"#fn8\" class=\"footnote-ref\" id=\"fnref8\" role=\"doc-noteref\"><sup>8<\/sup><\/a>.<\/li>\n<li>Peeep.us (defunct as of 2018)<\/li>\n<li><a href=\"http:\/\/archive.today\/\">Archive.is<\/a><\/li>\n<li><a href=\"https:\/\/pinboard.in\/\">Pinboard<\/a> (with the $25\/yr archiving option<a href=\"#fn9\" class=\"footnote-ref\" id=\"fnref9\" role=\"doc-noteref\"><sup>9<\/sup><\/a>)<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/docs\/www\/hiyo.jp\/20b25c1bc5641a4d93e64aab2128e9246f15a469.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/hiyo.jp\/\" title=\"(Original URL: http:\/\/hiyo.jp\/ )\">Hiyo.jp<\/a> &amp; <a href=\"https:\/\/www.gwern.net\/docs\/www\/megalodon.jp\/d58907b22d706b8374f8b22a3761daacdcc5df4e.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/megalodon.jp\/\" title=\"(Original URL: http:\/\/megalodon.jp\/ )\">Megalodon.jp<\/a> (may be difficult to use)<\/li>\n<\/ol><p>There are other options but they are not available like Google<a href=\"#fn10\" class=\"footnote-ref\" id=\"fnref10\" role=\"doc-noteref\"><sup>10<\/sup><\/a> or various commercial\/government archives<a href=\"#fn11\" class=\"footnote-ref\" id=\"fnref11\" role=\"doc-noteref\"><sup>11<\/sup><\/a><\/p><p>(An example would be <a href=\"http:\/\/bits.blogs.nytimes.com\/2010\/12\/07\/palm-is-far-from-game-over-says-former-chief\/\" title=\"Palm Is Far From 'Game Over', Says Former Chief\"><code>bits.blogs.nytimes.com\/2010\/12\/07\/palm-is-far-from-game-over-says-former-chief\/<\/code><\/a> being archived at <a href=\"https:\/\/www.gwern.net\/docs\/www\/webcitation.org\/ec6cc4cb308f20688510fda4d5e55f07dd81bf5a.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/webcitation.org\/5ur7ifr12\" title=\"(Original URL: http:\/\/webcitation.org\/5ur7ifr12 )\"><code>webcitation.org\/5ur7ifr12<\/code><\/a>.)<\/p><p>These archives are also good for archiving your own website:<\/p><ol type=\"1\"><li>you may be keeping backups of it, but your own website\/server backups can be lost (I can speak from personal experience here), so it\u2019s good to have external copies<\/li>\n<li>Another benefit is the reduction in \u2018bus-factor\u2019: if you were hit by a bus tomorrow, who would get your archives and be able to maintain the websites and understand the backups etc? While if archived in IA, people already know how to get copies and there are tools to download entire domains.<\/li>\n<li>A focus on backing up only one\u2019s website can blind one to the need for archiving the <em>external links<\/em> as well. Many pages are meaningless or less valuable with broken links. A linkchecker script\/daemon can also archive all the external links.<\/li>\n<\/ol><p>So there are several benefits to doing web archiving beyond simple server backups.<\/p><p>My first program in this vein of thought was a bot which fired off WebCite, Internet Archive\/Alexa, &amp; Archive.is requests: , quickly followed up by a . (Or you could install the  to get automatic submission to the Internet Archive, if you have ceased to care about privacy.)<\/p><p>The core code was quickly adapted into a <a href=\"https:\/\/hackage.haskell.org\/package\/gitit\">gitit<\/a> wiki plugin which hooked into the save-page functionality and tried to archive every link in the newly-modified page, <a href=\"https:\/\/github.com\/jgm\/gitit\/blob\/master\/plugins\/Interwiki.hs\">Interwiki.hs<\/a><\/p><p>Finally, I wrote <a href=\"https:\/\/hackage.haskell.org\/package\/archiver\">archiver<\/a>, a daemon which watches<a href=\"#fn12\" class=\"footnote-ref\" id=\"fnref12\" role=\"doc-noteref\"><sup>12<\/sup><\/a>\/reads a text file. Source is available via <code>git clone https:\/\/github.com\/gwern\/archiver-bot.git<\/code>. (A similar tool is <a href=\"https:\/\/github.com\/rahiel\/archiveror\">Archiveror<\/a>.)<\/p><p>The library half of <code>archiver<\/code> is a simple wrapper around the appropriate HTTP requests; the executable half reads a specified text file and loops as it (slowly) fires off requests and deletes the appropriate URL.<\/p><p>That is, <code>archiver<\/code> is a daemon which will process a specified text file, each line of which is a URL, and will one by one request that the URLs be archived or spidered<\/p><p>Usage of <code>archiver<\/code> might look like <code>archiver ~\/.urls.txt gwern@gwern.net<\/code>. In the past, <code>archiver<\/code> would sometimes crash for unknown reasons, so I usually wrap it in a <code>while<\/code> loop like so: <code>while true; do archiver ~\/.urls.txt gwern@gwern.net; done<\/code>. If I wanted to put it in a detached  session: <code>screen -d -m -S \"archiver\" sh -c 'while true; do archiver ~\/.urls.txt gwern@gwern.net; done'<\/code>. Finally, rather than start it manually, I use a cron job to start it at boot, for a final invocation of<\/p><div class=\"sourceCode\" id=\"cb3\"><pre class=\"sourceCode bash\">@reboot sleep 4m &amp;&amp; screen -d -m -S \"archiver\" sh -c 'while true; do archiver ~\/.urls.txt gwern2@gwern.net \\\n        \"cd ~\/www &amp;&amp; nice -n 20 ionice -c3 wget --unlink --limit-rate=20k --page-requisites --timestamping \\\n        -e robots=off --reject .iso,.exe,.gz,.xz,.rar,.7z,.tar,.bin,.zip,.jar,.flv,.mp4,.avi,.webm \\\n        --user-agent='Firefox\/4.9'\" 500; done'<\/pre><\/div><\/section><section id=\"local-caching\" class=\"level2\"><p>Remote archiving, while convenient, has a major flaw: the archive services cannot keep up with the growth of the Internet and are woefully incomplete. I experience this regularly, where a link on <code>gwern.net<\/code> goes dead and I cannot find it in the Internet Archive or WebCite, and it is a general phenomenon:  find &lt;35% of common Web pages ever copied into an archive service, and typically only one copy exists.<\/p><section id=\"caching-proxy\" class=\"level3\"><h3><a href=\"#caching-proxy\" title=\"Link to section: \u00a7'Caching Proxy'\">Caching Proxy<\/a><\/h3><p>The most ambitious &amp; total approach to local caching is to set up a proxy to do your browsing through, and record literally all your web traffic; for example, using <a href=\"https:\/\/github.com\/INA-DLWeb\/LiveArchivingProxy\">Live Archiving Proxy (LAP)<\/a> or <a href=\"https:\/\/github.com\/odie5533\/WarcProxy\">WarcProxy<\/a> which will save as  every page you visit through it. (<a href=\"https:\/\/www.gwern.net\/docs\/www\/blog.za3k.com\/e6b85edad0874ce9f51ecac2c7a2225dbf016a6e.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/blog.za3k.com\/archiving-all-web-traffic\/\" title=\"(Original URL: https:\/\/blog.za3k.com\/archiving-all-web-traffic\/ )\">Zachary Vance<\/a> explains how to set up a local HTTPS certificate to MITM your HTTPS browsing as well.)<\/p><p>One may be reluctant to go this far, and prefer something lighter-weight, such as periodically extracting a list of visited URLs from one\u2019s web browser and then attempting to archive them.<\/p><\/section><section id=\"batch-job-downloads\" class=\"level3\"><h3><a href=\"#batch-job-downloads\" title=\"Link to section: \u00a7'Batch job downloads'\">Batch job downloads<\/a><\/h3><p>For a while, I used a shell script named, imaginatively enough, <code>local-archiver<\/code>:<\/p><div class=\"sourceCode\" id=\"cb4\"><pre class=\"sourceCode bash\">#!\/bin\/sh\nset -euo pipefail\ncp `find ~\/.mozilla\/ -name \"places.sqlite\"` ~\/\nsqlite3 places.sqlite \"SELECT url FROM moz_places, moz_historyvisits \\\n                       WHERE moz_places.id = moz_historyvisits.place_id \\\n                             and visit_date &gt; strftime('%s','now','-1.5 month')*1000000 ORDER by \\\n                       visit_date;\" | filter-urls &gt;&gt; ~\/.tmp\nrm ~\/places.sqlite\nsplit -l500 ~\/.tmp ~\/.tmp-urls\nrm ~\/.tmp\ncd ~\/www\/\nfor file in ~\/.tmp-urls*;\n do (wget --unlink --continue --page-requisites --timestamping --input-file $file &amp;&amp; rm $file &amp;);\ndone\nfind ~\/www -size +4M -delete<\/pre><\/div><p>The code is not the prettiest, but it\u2019s fairly straightforward:<\/p><ol type=\"1\"><li>\n<p>the script grabs my Firefox browsing history by extracting it from the history SQL database file<a href=\"#fn13\" class=\"footnote-ref\" id=\"fnref13\" role=\"doc-noteref\"><sup>13<\/sup><\/a>, and feeds the URLs into .<\/p>\n<p><code>wget<\/code> is not the best tool for archiving as it will not run JavaScript or Flash or download videos etc. It will download included JS files but the JS will be obsolete when run in the future and any dynamic content will be long gone. To do better would require a headless browser like  which saves to MHT\/MHTML, but PhantomJS <a href=\"https:\/\/github.com\/ariya\/phantomjs\/issues\/10117\">refuses to support it<\/a> and I\u2019m not aware of an existing package to do this. In practice, static content is what is most important to archive, most JS is of highly questionable value in the first place, and any important YouTube videos can be archived manually with <code>youtube-dl<\/code>, so <code>wget<\/code>\u2019s limitations haven\u2019t been so bad.<\/p>\n<\/li>\n<li>\n<p>The script <code>split<\/code>s the long list of URLs into a bunch of files and runs that many <code>wget<\/code>s in parallel because <code>wget<\/code> apparently has no way of simultaneously downloading from multiple domains. There\u2019s also the chance of <code>wget<\/code> hanging indefinitely, so parallel downloads continues to make progress.<\/p>\n<\/li>\n<li>\n<p>The <a href=\"#filter-urls\"><code>filter-urls<\/code><\/a> command is another shell script, which removes URLs I don\u2019t want archived. This script is a hack which looks like this:<\/p>\n<div class=\"sourceCode\" id=\"cb5\"><pre class=\"sourceCode bash\">#!\/bin\/sh\nset -euo pipefail\ncat \/dev\/stdin | sed -e \"s\/#.*\/\/\" | sed -e \"s\/&amp;sid=.*$\/\/\" | sed -e \"s\/\\\/$\/\/\" | grep -v -e 4chan -e reddit ...<\/pre><\/div>\n<\/li>\n<li>\n<p>delete any particularly large (&gt;4MB) files which might be media files like videos or audios (podcasts are particular offenders)<\/p>\n<\/li>\n<\/ol><p>A local copy is not the best resource\u2014what if a link goes dead in a way your tool cannot detect so you don\u2019t <em>know<\/em> to put up your copy somewhere? But it solves the problem decisively.<\/p><p>The downside of this script\u2019s batch approach soon became apparent to me:<\/p><ol type=\"1\"><li>not automatic: you have to remember to invoke it and it only provides a single local archive, or if you invoke it regularly as a cron job, you may create lots of duplicates.<\/li>\n<li>unreliable: <code>wget<\/code> may hang, URLs may be archived too late, it may not be invoked frequently enough, &gt;4MB non-video\/audio files are increasingly common\u2026<\/li>\n<li>I wanted copies in the Internet Archive &amp; elsewhere as well to let other people benefit and provide redundancy to my local archive<\/li>\n<\/ol><p>It was to fix these problems that I began working on <code>archiver<\/code>\u2014which would run constantly archiving URLs in the background, archive them into the IA as well, and be smarter about media file downloads. It has been much more satisfactory.<\/p><\/section><section id=\"daemon\" class=\"level3\"><h3><a href=\"#daemon\" title=\"Link to section: \u00a7'Daemon'\">Daemon<\/a><\/h3><p><code>archiver<\/code> has an extra feature where any third argument is treated as an arbitrary <code>sh<\/code> command to run after each URL is archived, to which is appended said URL. You might use this feature if you wanted to load each URL into Firefox, or append them to a log file, or simply download or archive the URL in some other way.<\/p><p>For example, instead of a big <code>local-archiver<\/code> run, I have <code>archiver<\/code> run <code>wget<\/code> on each individual URL: <code>screen -d -m -S \"archiver\" sh -c 'while true; do archiver ~\/.urls.txt gwern@gwern.net \"cd ~\/www &amp;&amp; wget --unlink --continue --page-requisites --timestamping -e robots=off --reject .iso,.exe,.gz,.xz,.rar,.7z,.tar,.bin,.zip,.jar,.flv,.mp4,.avi,.webm --user-agent='Firefox\/3.6' 120\"; done'<\/code>. (For private URLs which require logins, such as , <code>wget<\/code> can still grab them with some help: installing the Firefox extension <a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/export-cookies\/\">Export Cookies<\/a>, logging into the site in Firefox like usual, exporting one\u2019s <code>cookies.txt<\/code>, and adding the option <code>--load-cookies cookies.txt<\/code> to give it access to the cookies.)<\/p><p>Alternately, you might use <code>curl<\/code> or a specialized archive downloader like the Internet Archive\u2019s crawler <a href=\"http:\/\/crawler.archive.org\/\">Heritrix<\/a>.<\/p><\/section><section id=\"resource-consumption\" class=\"level3\"><h3><a href=\"#resource-consumption\" title=\"Link to section: \u00a7'Resource consumption'\">Resource consumption<\/a><\/h3><p>The space consumed by such a backup is not that bad; only 30\u201350 gigabytes for a year of browsing, and less depending on how hard you prune the downloads. (More, of course, if you use <code>linkchecker<\/code> to archive entire sites and not just the pages you visit.) Storing this is quite viable in the long term; while page sizes have <a href=\"https:\/\/www.gwern.net\/docs\/www\/www.websiteoptimization.com\/e636df6f082c30b99a930d0450c4af5fd912e729.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www.websiteoptimization.com\/speed\/tweak\/average-web-page\/\" title=\"(Original URL: http:\/\/www.websiteoptimization.com\/speed\/tweak\/average-web-page\/ )\">increased 7x<\/a> between 2003 and 2011 and pages average around 400kb<a href=\"#fn14\" class=\"footnote-ref\" id=\"fnref14\" role=\"doc-noteref\"><sup>14<\/sup><\/a>,  has also been operating and has increased disk capacity by ~128x\u2014in 2011, $80 will buy you at least <a href=\"https:\/\/www.gwern.net\/docs\/www\/forre.st\/07963fe85eed8b845122c702ce0cbe668816ac02.html#hdd\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/forre.st\/storage#hdd\" title=\"(Original URL: http:\/\/forre.st\/storage#hdd )\">2 terabytes<\/a>, that works out to 4 cents a gigabyte or 80 cents for the low estimate for downloads; that is much better than the $25 annual fee that somewhere like <a href=\"http:\/\/pinboard.in\/upgrade\/\">Pinboard<\/a> charges. Of course, you need to back this up yourself. We\u2019re relatively fortunate here\u2014most Internet documents are \u2018born digital\u2019 and easy to migrate to new formats or inspect in the future. We can download them and worry about how to view them only when we need a particular document, and Web browser backwards-compatibility already stretches back to files written in the early 1990s. (Of course, we\u2019re probably screwed if we discover the content we wanted was dynamically presented only in Adobe Flash or as an inaccessible \u2018cloud\u2019 service.) In contrast, if we were trying to preserve programs or software libraries instead, we would face a much more formidable task in keeping a working ladder of binary-compatible virtual machines or interpreters<a href=\"#fn15\" class=\"footnote-ref\" id=\"fnref15\" role=\"doc-noteref\"><sup>15<\/sup><\/a>. The situation with <a href=\"https:\/\/www.gwern.net\/docs\/www\/www.davidbordwell.net\/f933619e20a14d5d4fbf35ab7ea25b4acea8b23b.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/www.davidbordwell.net\/blog\/2012\/02\/13\/pandoras-digital-box-pix-and-pixels\/\" title=\"Pandora's digital box: Pix and pixels (Original URL: http:\/\/www.davidbordwell.net\/blog\/2012\/02\/13\/pandoras-digital-box-pix-and-pixels\/ )\">digital movie preservation<\/a> hardly bears thinking on.<\/p><p>There are ways to cut down on the size; if you tar it all up and run  with maximum compression options, you could probably compact it to 1\/5th the size. I found that the uncompressed files could be reduced by around 10% by using  to look for duplicate files and turning the duplicates into a space-saving  to the original with a command like <code>fdupes --recurse --hardlink ~\/www\/<\/code>. (Apparently there are a <em>lot<\/em> of bit-identical JavaScript (eg. ) and images out there.)<\/p><p>Good filtering of URL sources can help reduce URL archiving count by a large amount. Examining my manual backups of Firefox browsing history, over the 1153 days from 2014-02-25 to 2017-04-22, I visited 2,370,111 URLs or 2055 URLs per day; after passing through my filtering script, that leaves 171,446 URLs, which after de-duplication yields 39,523 URLs or ~34 unique URLs per day or 12,520 unique URLs per year to archive.<\/p><p>This shrunk my archive by 9GB from 65GB to 56GB, although at the cost of some archiving fidelity by removing many filetypes like CSS or JavaScript or GIF images. As of 22 April 2017, after ~6 years of archiving, between <code>xz<\/code> compression (at the cost of easy searchability), aggressive filtering, occasional manual deletion of overly bulky domains I feel are probably adequately covered in the IA etc, my full WWW archives weigh 55GB.<\/p><\/section><section id=\"url-sources\" class=\"level3\"><h3><a href=\"#url-sources\" title=\"Link to section: \u00a7'URL sources'\">URL sources<\/a><\/h3><section id=\"browser-history\" class=\"level4\"><h4><a href=\"#browser-history\" title=\"Link to section: \u00a7'Browser history'\">Browser history<\/a><\/h4><p>There are a number of ways to populate the source text file. For example, I have a script <code>firefox-urls<\/code>:<\/p><div class=\"sourceCode\" id=\"cb7\"><pre class=\"sourceCode bash\">#!\/bin\/sh\nset -euo pipefail\ncp --force `find ~\/.mozilla\/firefox\/ -name \"places.sqlite\"|sort|head -1` ~\/\nsqlite3 -batch places.sqlite \"SELECT url FROM moz_places, moz_historyvisits \\\n                       WHERE moz_places.id = moz_historyvisits.place_id and \\\n                       visit_date &gt; strftime('%s','now','-1 day')*1000000 ORDER by \\\n                       visit_date;\" | filter-urls\nrm ~\/places.sqlite<\/pre><\/div><p>(<code>filter-urls<\/code> is the same script as in <code>local-archiver<\/code>. If I don\u2019t want a domain locally, I\u2019m not going to bother with remote backups either. In fact, because of WebCite\u2019s rate-limiting, <code>archiver<\/code> is almost perpetually back-logged, and I <em>especially<\/em> don\u2019t want it wasting time on worthless links like .)<\/p><p>This is called every hour by <code>cron<\/code>:<\/p><div class=\"sourceCode\" id=\"cb8\"><pre class=\"sourceCode bash\">@hourly firefox-urls &gt;&gt; ~\/.urls.txt<\/pre><\/div><p>This gets all visited URLs in the last time period and prints them out to the file for archiver to process. Hence, everything I browse is backed-up through <code>archiver<\/code>.<\/p><p>Non-Firefox browsers can be supported with similar strategies; for example, Zachary Vance\u2019s Chromium scripts likewise extracts URLs from Chromium\u2019s <a href=\"https:\/\/github.com\/za3k\/rip-chrome-history\">SQL history<\/a> &amp; <a href=\"https:\/\/github.com\/za3k\/export-chrome-bookmarks\">bookmarks<\/a>.<\/p><\/section><section id=\"document-links\" class=\"level4\"><h4><a href=\"#document-links\" title=\"Link to section: \u00a7'Document links'\">Document links<\/a><\/h4><p>More useful perhaps is a script to extract external links from Markdown files and print them to standard out: <a href=\"https:\/\/www.gwern.net\/haskell\/link-extractor.hs\">link-extractor.hs<\/a><\/p><p>So now I can take <code>find . -name \"*.page\"<\/code>, pass the 100 or so Markdown files in my wiki as arguments, and add the thousand or so external links to the archiver queue (eg. <code>find . -name \"*.page\" -type f -print0 | xargs -0 ~\/wiki\/haskell\/link-extractor.hs | filter-urls &gt;&gt; ~\/.urls.txt<\/code>); they will eventually be archived\/backed up.<\/p><\/section><section id=\"website-spidering\" class=\"level4\"><h4><a href=\"#website-spidering\" title=\"Link to section: \u00a7'Website spidering'\">Website spidering<\/a><\/h4><p>Sometimes a particular website is of long-term interest to one even if one has not visited <em>every<\/em> page on it; one could manually visit them and rely on the previous Firefox script to dump the URLs into <code>archiver<\/code> but this isn\u2019t always practical or time-efficient. <code>linkchecker<\/code> inherently spiders the websites it is turned upon, so it\u2019s not a surprise that it can build a  or simply spit out all URLs on a domain; unfortunately, while <code>linkchecker<\/code> has the ability to output in a remarkable variety of formats, it cannot simply output a newline-delimited list of URLs, so we need to post-process the output considerably. The following is the shell one-liner I use when I want to archive an entire site (note that this is a bad command to run on a large or heavily hyper-linked site like the English Wikipedia or <a href=\"http:\/\/lesswrong.com\">LessWrong<\/a>!); edit the target domain as necessary:<\/p><div class=\"sourceCode\" id=\"cb9\"><pre class=\"sourceCode bash\">linkchecker --check-extern -odot --complete -v --ignore-url=^mailto --no-warnings http:\/\/www.longbets.org\n    | fgrep http # [\n    | fgrep -v -e \"label=\" -e \"-&gt;\" -e '\" [' -e '\" ]' -e \"\/ \"\n    | sed -e \"s\/href=\\\"\/\/\" -e \"s\/\\\",\/\/\" -e \"s\/ \/\/\"\n    | filter-urls\n    | sort --unique &gt;&gt; ~\/.urls.txt<\/pre><\/div><p>When <code>linkchecker<\/code> does not work, one alternative is to do a <code>wget --mirror<\/code> and extract the URLs from the filenames\u2014list all the files and prefix with a \u201chttp:\/\/\u201d etc.<\/p><\/section><\/section><\/section><\/section><section id=\"reacting-to-broken-links\" class=\"level1\"><p><code>archiver<\/code> combined with a tool like <code>link-checker<\/code> means that there will rarely be any broken links on <code>gwern.net<\/code> since one can either find a live link or use the archived version. In theory, one has multiple options now:<\/p><ol start=\"0\" type=\"1\"><li>\n<p>Search for a copy on the live Web<\/p>\n<\/li>\n<li>\n<p>link the Internet Archive copy<\/p>\n<\/li>\n<li>\n<p>link the WebCite copy<\/p>\n<\/li>\n<li>\n<p>link the WikiWix copy<\/p>\n<\/li>\n<li>\n<p>use the <code>wget<\/code> dump<\/p>\n<p>If it\u2019s been turned into a full local file-based version with <code>--convert-links --page-requisites<\/code>, one can easily convert the dump into something like a standalone PDF suitable for public distribution. (A PDF is easier to store and link than the original directory of bits and pieces or other HTML formats like a ZIP archive of said directory.)<\/p>\n<p>I use <a href=\"https:\/\/www.gwern.net\/docs\/www\/wkhtmltopdf.org\/1166a9ebce4000a02b4175174daae3731a5d756c.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/wkhtmltopdf.org\/\" title=\"(Original URL: http:\/\/wkhtmltopdf.org\/ )\"><code>wkhtmltopdf<\/code><\/a> which does a good job; an example of a dead webpage with no Internet mirrors is <code>http:\/\/www.aeiveos.com\/~bradbury\/MatrioshkaBrains\/MatrioshkaBrainsPaper.html<\/code> which can be found at <a href=\"https:\/\/www.gwern.net\/docs\/ai\/1999-bradbury-matrioshkabrains.pdf\"><code>1999-bradbury-matrioshkabrains.pdf<\/code><\/a>, or Sternberg et al\u2019s 2001 review <a href=\"https:\/\/www.gwern.net\/docs\/iq\/2001-sternberg.pdf\">\u201cThe Predictive Value of IQ\u201d<\/a>.<\/p>\n<\/li>\n<\/ol><\/section><section id=\"external-links\" class=\"level1\"><ul><li>\n<p><a href=\"https:\/\/github.com\/oduwsdl\/archivenow\"><code>archivenow<\/code><\/a> (library for submission to <code>archive.is<\/code>, <code>archive.st<\/code>, <code>perma.cc<\/code>, Megalodon, WebCite, &amp; IA)<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/timetravel.mementoweb.org\/\">Memento meta-archive search engine<\/a> (for checking IA &amp; other archives)<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/www.archive-it.org\/\">Archive-It<\/a> -(by the Internet Archive); <a href=\"https:\/\/archive.org\/donate\/\">donate to the IA<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/pinboard.in\/\">Pinboard<\/a><\/p>\n<ul><li><a href=\"https:\/\/web.archive.org\/web\/20170707135337\/http:\/\/blog.pinboard.in:80\/2010\/11\/bookmark_archives_that_don_t\">\u201cBookmark Archives That Don\u2019t\u201d<\/a><\/li>\n<\/ul><\/li>\n<li>\n<p><a href=\"https:\/\/www.gwern.net\/docs\/www\/samsaffron.com\/5f76aa4a1c761f34c30a84292e11bed153b6420d.html#comment-31366\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/samsaffron.com\/archive\/2012\/06\/07\/testing-3-million-hyperlinks-lessons-learned#comment-31366\" title=\"(Original URL: http:\/\/samsaffron.com\/archive\/2012\/06\/07\/testing-3-million-hyperlinks-lessons-learned#comment-31366 )\">\u201cTesting 3 million hyperlinks, lessons learned\u201d<\/a>, Stack Exchange<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.gwern.net\/docs\/linkrot\/2011-muflax-backup.pdf\">\u201cBackup All The Things\u201d<\/a>, muflax<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.gwern.net\/docs\/www\/tesoro.io\/fde1f38f35a3caa5d2cc214dc79bfdf4ecffa3d5.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/tesoro.io\/\" title=\"(Original URL: https:\/\/tesoro.io\/ )\">Tesoro<\/a> (archive service; <a href=\"https:\/\/news.ycombinator.com\/item?id=14644441\">discussion<\/a>)<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/xkcd.com\/1909\/\">\u201cDigital Resource Lifespan\u201d<\/a>, <em>XKCD<\/em><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/lwn.net\/Articles\/766374\/\">\u201cArchiving web sites\u201d<\/a>, LWN<\/p>\n<\/li>\n<li>\n<p>software:<\/p>\n<ul><li><a href=\"https:\/\/webrecorder.io\/\">webrecorder.io<\/a> (online service); <a href=\"https:\/\/github.com\/webrecorder\/webrecorder\">webrecorder<\/a> (offline tool; WARC-based &amp; can deal with dynamically-loaded content)<\/li>\n<li><a href=\"https:\/\/github.com\/ikreymer\/pywb\">pywb<\/a><\/li>\n<li><a href=\"https:\/\/archivebox.io\/\">ArchiveBox<\/a> (headless Chrome)<\/li>\n<li>\n<\/li><li><a href=\"https:\/\/www.gwern.net\/docs\/www\/wordpress.org\/87feddbeb90c8f8f076dbac7ab01d77296329ea2.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/wordpress.org\/plugins\/broken-link-checker\/\" title=\"(Original URL: https:\/\/wordpress.org\/plugins\/broken-link-checker\/ )\">WordPress plugin: Broken Link Checker<\/a><\/li>\n<li><a href=\"https:\/\/rg3.github.io\/youtube-dl\/\"><code>youtube-dl<\/code><\/a><\/li>\n<li>\n<\/li><li><a href=\"https:\/\/github.com\/marvelm\/erised\">erised<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/chfoo\/wpull\">wpull<\/a><\/li>\n<\/ul><\/li>\n<li>\n<p><a href=\"https:\/\/news.ycombinator.com\/item?id=6504331\">Hacker News discussion<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.gwern.net\/docs\/www\/stop.zona-m.net\/783905af0715e433863b32c879821ed20cee5de3.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/stop.zona-m.net\/2018\/01\/indeed-it-seems-that-google-is-forgetting-the-old-web\/\" title=\"(Original URL: http:\/\/stop.zona-m.net\/2018\/01\/indeed-it-seems-that-google-is-forgetting-the-old-web\/ )\">\u201cIndeed, it seems that Google <em>is<\/em> forgetting the old Web\u201d<\/a> (<a href=\"https:\/\/news.ycombinator.com\/item?id=19604135\">HN<\/a>)<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.gwern.net\/docs\/www\/spectrum.ieee.org\/99737546cb17f6d5cd38009b1d8ae3ead68119df.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/spectrum.ieee.org\/computing\/it\/the-lost-picture-show-hollywood-archivists-cant-outpace-obsolescence\" title=\"(Original URL: https:\/\/spectrum.ieee.org\/computing\/it\/the-lost-picture-show-hollywood-archivists-cant-outpace-obsolescence )\">\u201cThe Lost Picture Show: Hollywood Archivists Can\u2019t Outpace Obsolescence\u2014Studios invested heavily in magnetic-tape storage for film archiving but now struggle to keep up with the technology\u201d<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.gwern.net\/docs\/www\/sivers.org\/829d00a5214da2df3e200dc3c3b071c751fc1215.html\" rel=\"archived alternate nofollow\" data-url-original=\"https:\/\/sivers.org\/netskill\" title=\"(Original URL: https:\/\/sivers.org\/netskill )\">\u201cUse the internet, not just companies\u201d<\/a>, <\/p>\n<\/li>\n<\/ul><\/section><section id=\"appendices\" class=\"level1\"><section id=\"filter-urls\" class=\"level2\"><p>A raw dump of URLs, while certainly archivable, will typically result in a very large mirror of questionable value (is it really necessary to archive Google search queries or Wikipedia articles? usually, no) and worse, given the rate-limiting necessary to store URLs in the Internet Archive or other services, may wind up delaying the archiving of the important links &amp; risking their total loss. Disabling the remote archiving is unacceptable, so the best solution is to simply take a little time to manually blacklist various domains or URL patterns.<\/p><p>This blacklisting can be as simple as a command like <code>filter-urls | grep -v en.wikipedia.org<\/code>, but can be much more elaborate. The following shell script is the skeleton of my own custom blacklist, derived from manually filtering through several years of daily browsing as well as spiders of <a href=\"http:\/\/lesswrong.com\/lw\/7kg\/rationalist_sites_worth_archiving\/\" title=\"Rationalist sites worth archiving?\">dozens of websites<\/a> for various people &amp; purposes, demonstrating a variety of possible techniques: regexps for domains &amp; file-types &amp; query-strings, <code>sed<\/code>-based rewrites, fixed-string matches (both blacklists and whitelists), etc:<\/p><div class=\"sourceCode\" id=\"cb10\"><pre class=\"sourceCode bash\">#!\/bin\/sh\n# USAGE: `filter-urls` accepts on standard input a list of newline-delimited URLs or filenames,\n# and emits on standard output a list of newline-delimited URLs or filenames.\n#\n# This list may be shorter and entries altered. It tries to remove all unwanted entries, where 'unwanted'\n# is a highly idiosyncratic list of regexps and fixed-string matches developed over hundreds of thousands\n# of URLs\/filenames output by my daily browsing, spidering of interesting sites, and requests\n# from other people to spider sites for them.\n#\n# You are advised to test output to make sure it does not remove\n# URLs or filenames you want to keep. (An easy way to test what is removed is to use the `comm` utility.)\n#\n# For performance, it does not sort or remove duplicates from output; both can be done by\n# piping `filter-urls` to `sort --unique`.\nset -euo pipefail\ncat \/dev\/stdin \\\n    | sed -e \"s\/#.*\/\/\" -e 's\/&gt;$\/\/' -e \"s\/&amp;sid=.*$\/\/\" -e \"s\/\\\/$\/\/\" -e 's\/$\/\\n\/' -e 's\/\\?sort=.*$\/\/' \\\n      -e 's\/^[ \\t]*\/\/' -e 's\/utm_source.*\/\/' -e 's\/https:\\\/\\\/\/http:\\\/\\\/\/' -e 's\/\\?showComment=.*\/\/' \\\n    | grep \"\\.\" \\\n    | fgrep -v \"*\" \\\n    | egrep -v -e '\\\/\\.rss$' -e \"\\.tw$\" -e \"\/\/%20www\\.\" -e \"\/file-not-found\" -e \"258..\\.com\/$\" \\\n       -e \"3qavdvd\" -e \":\/\/avdvd\" -e \"\\.avi\" -e \"\\.com\\.tw\" -e \"\\.onion\" -e \"\\?fnid\\=\" -e \"\\?replytocom=\" \\\n       -e \"^lesswrong.com\/r\/discussion\/comments$\" -e \"^lesswrong.com\/user\/gwern$\" \\\n       -e \"^webcitation.org\/query$\" \\\n       -e \"ftp.*\" -e \"6..6\\.com\" -e \"6..9\\.com\" -e \"6??6\\.com\" -e \"7..7\\.com\" -e \"7..8\\.com\" -e \"7..\\.com\" \\\n       -e \"78..\\.com\" -e \"7??7\\.com\" -e \"8..8\\.com\" -e \"8??8\\.com\" -e \"9..9\\.com\" -e \"9??9\\.com\" \\\n       -e gold.*sell -e vip.*club \\\n    | fgrep -v -e \"#!\" -e \".bin\" -e \".mp4\" -e \".swf\" -e \"\/mediawiki\/index.php?title=\" -e \"\/search?q=cache:\" \\\n      -e \"\/wiki\/Special:Block\/\" -e \"\/wiki\/Special:WikiActivity\" -e \"Special%3ASearch\" \\\n      -e \"Special:Search\" -e \"__setdomsess?dest=\"\n      # ...\n# prevent URLs from piling up at the end of the file\necho \"\"<\/pre><\/div><p><code>filter-urls<\/code> can be used on one\u2019s local archive to save space by deleting files which may be downloaded by <code>wget<\/code> as dependencies. For example:<\/p><div class=\"sourceCode\" id=\"cb11\"><pre class=\"sourceCode bash\">find ~\/www | sort --unique &gt;&gt; full.txt &amp;&amp; \\\n    find ~\/www | filter-urls | sort --unique &gt;&gt; trimmed.txt\ncomm -23 full.txt trimmed.txt | xargs -d \"\\n\" rm\nrm full.txt trimmed.txt<\/pre><\/div><\/section><section id=\"sort---key-compression-trick\" class=\"level2 collapse\"><div class=\"collapseSummary\"><blockquote>\n<p>Programming folklore notes that one way to get better lossless compression efficiency is to rearrange files inside the archive to group \u2018similar\u2019 files together and expose redundancy to the compressor, in accordance with information-theoretical principles. A particularly easy and broadly-applicable way of doing this, which does not require using any unusual formats or tools and is fully compatible with the default archive methods, is to sort the files by <em>filename<\/em> and especially file extension. I show how to do this with the standard Unix command-line <code>sort<\/code> tool, using the so-called \u201c<code>sort --key<\/code> trick\u201d, and give examples of the large space-savings possible from my archiving work for personal website mirrors and for making  where the redundancy at the file level is particularly extreme and the <code>sort --key<\/code> trick shines compared to the naive approach.<\/p>\n<\/blockquote><\/div><div class=\"drop-cap-yinit\"><p>One way to look at data compression is as a form of intelligence (see the  &amp; ): a compression tool like  is being asked to predict what the next bit of the original file is, and the better its predictions, the less data needs to be stored to correct its mistaken predictions (\u201cI know how to spell \u2018banana\u2019, I just don\u2019t know when to stop\u201d). The smarter the program is, the better its compression will be; but on the flip side, you can also improve the compression ratio by giving it a little help and organizing the data into an equivalent form the program can understand better\u2014for example, by using the . Or by preserving spatial locality: keeping similar data together, and not dispersing it all over. (This also explains why multimedia files barely compress: because the lossy encodings are the product of decades of work specialized to the particular domain of audio or images or video, and a general-purpose lossless compression would have to be <em>very<\/em> intelligent, on par with , to beat them at their own game.) Files on one topic should be compressed together.<\/p><\/div><section id=\"locality\" class=\"level3\"><h3><a href=\"#locality\" title=\"Link to section: \u00a7'Locality'\">Locality<\/a><\/h3><p>Spatial locality can be subtle. For example, natural language text, though not structured line-by-line as visibly as a dictionary, is still far from random and has many local correlations a compression tool might be able to pick up. If this is true, we would expect that with a sufficiently smart compressor, a text file would compress better in its natural form than if it were randomly shuffled. Is this the case, or are compression utilities are too stupid to see any different between random lines and English prose? Taking 14M of text from <code>gwern.net<\/code>, we can see for ourselves:<\/p><div class=\"sourceCode\" id=\"cb12\"><pre class=\"sourceCode bash\">## uncompressed\ncat *.page *\/*.page *\/*\/*.page |                                                 wc --bytes\n# 13588814\n## compressed, files in lexicographic order\ncat *.page *\/*.page *\/*\/*.page |                      xz -9 --extreme --stdout | wc --bytes\n# 3626732\n## compressed, all lines sorted alphabetically\ncat *.page *\/*.page *\/*\/*.page | sort               | xz -9 --extreme --stdout | wc --bytes\n# 3743756\n## compressed, all lines randomly shuffled except for non-unique lines sorted together\ncat *.page *\/*.page *\/*\/*.page | sort --random-sort | xz -9 --extreme --stdout | wc --bytes\n# 3831740\n## compressed, all lines randomly shuffled\ncat *.page *\/*.page *\/*\/*.page | shuf               | xz -9 --extreme --stdout | wc --bytes\n# 3862632<\/pre><\/div><p>The unmolested text compresses to 3.626M, but the <em>same<\/em> text randomly shuffled is 3.862M! I also included an intermediate form of randomization: despite the name, the <code>--random-sort<\/code> option to <code>sort<\/code> is not actually a random shuffle but a random <em>hash<\/em> (this is not documented in the man page, though it is in the info page) and so any repeated non-unique lines will be sorted together (allowing for some easy duplication deletion), and so we would expect the only-partial randomization of <code>--random-sort<\/code> to maybe perform a bit better than the true random shuffle of <code>shuf<\/code>. And it does.<\/p><\/section><section id=\"web-archives\" class=\"level3\"><h3><a href=\"#web-archives\" title=\"Link to section: \u00a7'Web archives'\">Web archives<\/a><\/h3><p>Spatial locality also applies to our web archiving. If you are mirroring websites, or otherwise compiling a lot of directories with redundant data on a file-by-file level, there\u2019s a cute trick to massively improve your compression ratios: don\u2019t sort the usual lexicographic way, but sort by a subdirectory. (I learned about this trick a few years ago while messing around with archiving my home directory using <code>find<\/code> and <code>tar<\/code>.) This is one of the issues with archiving gigabytes of crawls from thousands of domains: URLs have a historical oddity where they are not consistently hierarchical. URLs were originally modeled after hierarchical Unix filesystems; this page, for example, lives at the name <code>\/home\/gwern\/wiki\/Archiving-URLs.page<\/code>, which follows a logical left-to-right pattern of increasing narrowness. If one lists my entire filesystem in lexicographic order, all the files in <code>\/home\/gwern\/<\/code> will be consecutive, and the files in <code>wiki\/<\/code> will be consecutive, and so on. unfortunately, the top level of URLs breaks this scheme\u2014one does not visit <code>https:\/\/com\/google\/mail\/?shva=1#search\/l%3aunread<\/code>, one visits <code>https:\/\/mail.google.com\/mail\/?shva=1#search\/l%3aunread<\/code>; one does not visit <code>http:\/\/net\/www\/gwern\/Archiving-URLs<\/code> but <code>https:\/\/www.gwern.net\/Archiving-URLs<\/code>. So if I download <code>a.google.com<\/code> and then later <code>z.google.com<\/code>, a lexicographic list of downloaded files will separate the files as much as possible (even though they are semantically probably similar). A quick example from my current WWW archive:<\/p><div class=\"sourceCode\" id=\"cb13\"><pre class=\"sourceCode bash\">ls\n# ...\n# typemoon.wikia.com\/\n# tytempletonart.wordpress.com\/\n# ubc-emotionlab.ca\/\n# ubook.info\/\n# ucblibrary3.berkeley.edu\/\n# uhaweb.hartford.edu\/\n# ujsportal.pacourts.us\/\n# ukpmc.ac.uk\/\n# uk.reuters.com\/\n# ultan.org.uk\/\n# ...<\/pre><\/div><p>The directories are indeed sorted, but aside from the initial 2 letters or so, look nothing like each other: a Wikia subdomain rubs shoulders with a WordPress blog, a <code>.ca<\/code> domain is between a <code>.com<\/code>, a <code>.info<\/code>, and a <code>.edu<\/code> (with a <code>.us<\/code> and <code>.uk<\/code> thrown in for variety), and so on. Is there any way to sort these directories with a bit of parsing thrown in? For example, maybe we could reverse each line? Some web browsers store URLs reversed right-to-left to enable more efficient database operations, as do Google\u2019s  systems<a href=\"#fn16\" class=\"footnote-ref\" id=\"fnref16\" role=\"doc-noteref\"><sup>16<\/sup><\/a> (to assist their relatively weak compression utility ). Turns out GNU  already supports something similar, the <code>--key<\/code> &amp; <code>--field-separator<\/code> options; the man page is not very helpful but the  page tells us:<\/p><pre>'-t SEPARATOR'\n'--field-separator=SEPARATOR'\n     Use character SEPARATOR as the field separator when finding the\n     sort keys in each line.  By default, fields are separated by the\n     empty string between a non-blank character and a blank character.\n     By default a blank is a space or a tab, but the 'LC_CTYPE' locale\n     can change this.\n     That is, given the input line ' foo bar', 'sort' breaks it into\n     fields ' foo' and ' bar'.  The field separator is not considered\n     to be part of either the field preceding or the field following,\n     so with 'sort -t \" \"' the same input line has three fields: an\n     empty field, 'foo', and 'bar'.  However, fields that extend to the\n     end of the line, as '-k 2', or fields consisting of a range, as\n     '-k 2,3', retain the field separators present between the\n     endpoints of the range.\n'-k POS1[,POS2]'\n'--key=POS1[,POS2]'\n     Specify a sort field that consists of the part of the line between\n     POS1 and POS2 (or the end of the line, if POS2 is omitted),\n     _inclusive_.\n     Each POS has the form 'F[.C][OPTS]', where F is the number of the\n     field to use, and C is the number of the first character from the\n     beginning of the field.  Fields and character positions are\n     numbered starting with 1; a character position of zero in POS2\n     indicates the field's last character.  If '.C' is omitted from\n     POS1, it defaults to 1 (the beginning of the field); if omitted\n     from POS2, it defaults to 0 (the end of the field).  OPTS are\n     ordering options, allowing individual keys to be sorted according\n     to different rules; see below for details.  Keys can span multiple\n     fields.\n     Example:  To sort on the second field, use '--key=2,2' ('-k 2,2').\n     See below for more notes on keys and more examples.  See also the\n     '--debug' option to help determine the part of the line being used\n     in the sort.<\/pre><p>Hence, we can do better by ordering <code>sort<\/code> to break on the dots and focus on the second part of a URL, like so:<\/p><div class=\"sourceCode\" id=\"cb15\"><pre class=\"sourceCode bash\">ls | sort --key=2 --field-separator=\".\"\n# ...\n# uhaweb.hartford.edu\/\n# adsabs.harvard.edu\/\n# chandra.harvard.edu\/\n# cmt.harvard.edu\/\n# dash.harvard.edu\/\n# gking.harvard.edu\/\n# isites.harvard.edu\/\n# ...<\/pre><\/div><p>There\u2019s many possible ways to sort, though. So I took my WWW archive as of 15 June 2014, optimized all PNGs &amp; JPEGs with <code>optipng<\/code> &amp; <code>jpegoptim<\/code>, ran all the files through <code>filter-urls<\/code> &amp; deleted the ones which failed (this took out all of the JS files, which is fine since I don\u2019t think those are useful for archival purposes), and was left with ~86.5GB of files. Then I tested out several ways of sorting the filenames to see what gave the best compression on my corpus.<\/p><p>First, I establish the baseline:<\/p><ol type=\"1\"><li>\n<p>Size of uncompressed unsorted tarball, which eliminates filesystem overhead and tells us how much compression is really saving:<\/p>\n<div class=\"sourceCode\" id=\"cb16\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find .\/ -type f -print0 | tar c --to-stdout --no-recursion --null --files-from - | wc --bytes\n# 86469734400\n## 1x<\/pre><\/div>\n<\/li>\n<li>\n<p>Size of sorted tarball:<\/p>\n<div class=\"sourceCode\" id=\"cb17\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f -print0 | sort --zero-terminated | \\\n             tar c --to-stdout --no-recursion --null --files-from - | wc --bytes\n# 86469734400\n## 1x<\/pre><\/div>\n<p>So sorting a tarball doesn\u2019t give any benefits. This is mostly as I expected, since <code>tar<\/code> is only supposed to produce a linear archive packing together all the specified files and otherwise preserve them exactly. I thought there might have been some sort of consolidation of full path-names which might yield a small space savings, but apparently not.<\/p>\n<\/li>\n<\/ol><p>Now we can begin sorting before compression. I thought of 6 approaches; in decreasing order of final archive (smaller=better):<\/p><ol type=\"1\"><li>\n<p>Sort by file names, simply by reversing, sorting, unreverse (<code>foo.png ... bar.png<\/code> reverses to <code>gnp.oof ... gnp.rab<\/code>, which then sort together, and then losslessly reverse back to <code>bar.png \/ foo.png \/ ...<\/code>):<\/p>\n<div class=\"sourceCode\" id=\"cb18\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f | rev | sort | rev | \\\n             tar c --to-stdout --no-recursion --files-from - | xz -9 --stdout | wc --bytes\n# 24970605748\n## 0.2887x<\/pre><\/div>\n<p>(Note that <code>find<\/code>+<code>rev<\/code> doesn\u2019t correctly handle filenames with the wrong\/non-UTF-8 encoding; I ultimately used brute force in the form of <a href=\"https:\/\/www.gwern.net\/docs\/www\/linux.die.net\/7f87c1cf8f6af23204c59fb718697d7a4a645e7b.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/linux.die.net\/man\/1\/detox\" title=\"(Original URL: http:\/\/linux.die.net\/man\/1\/detox )\"><code>detox<\/code><\/a> to find all the non-UTF-8 files and rename them.)<\/p>\n<\/li>\n<li>\n<p>Compress tarball, but without any sorting (files are consumed in the order <code>find<\/code> produces them in the filesystem):<\/p>\n<div class=\"sourceCode\" id=\"cb19\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f -print0 | \\\n             tar c --to-stdout --no-recursion --null --files-from - | xz -9 --stdout | wc --bytes\n# 24268747400\n## 0.2806x<\/pre><\/div>\n<\/li>\n<li>\n<p>Sort by file suffixes, trying to parsing the filenames first:<\/p>\n<div class=\"sourceCode\" id=\"cb20\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f -printf '%f\/%p\\n' | sort --field-separator=\".\" --key=2  | cut -f2- -d\/ | \\\n             tar c --to-stdout --no-recursion --files-from - | xz -9 --stdout | wc --bytes\n# 24097155132\n## 0.2786x<\/pre><\/div>\n<\/li>\n<li>\n<p>Sort normally, in lexicographic order (subdomain, domain, TLD, subdirectories &amp; files etc):<\/p>\n<div class=\"sourceCode\" id=\"cb21\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f -print0 | sort --zero-terminated | \\\n    tar c --to-stdout --no-recursion --null --files-from - | xz -9 --stdout | wc --bytes\n# 23967317664\n## 0.2771x<\/pre><\/div>\n<\/li>\n<li>\n<p>Sort by middle of domain:<\/p>\n<div class=\"sourceCode\" id=\"cb22\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f -print0 | sort --zero-terminated --key=3 --field-separator=\".\" | \\\n              tar c --to-stdout --no-recursion --null --files-from - | xz -9 --stdout | wc --bytes\n# 23946061272\n## 0.2769x<\/pre><\/div>\n<\/li>\n<li>\n<p>Sort by first subdirectory (if there\u2019s a bunch of <code>foo.com\/wp-content\/*<\/code> &amp; <code>bar.com\/wp-content\/*<\/code> files, then the <code>\/wp-content\/<\/code> files will all sort together regardless of \u201cf\u201d and \u201cb\u201d being far from each other; similarly for <code>domain.com\/images\/<\/code>, <code>domain.com\/css\/<\/code> etc):<\/p>\n<div class=\"sourceCode\" id=\"cb23\"><pre class=\"sourceCode bash\">cd ~\/www\/ &amp;&amp; find . -type f -print0 | sort --zero-terminated --key=3 --field-separator=\"\/\"  | \\\n             tar c --to-stdout --no-recursion --null --files-from - | xz -9 --stdout | wc --bytes\n# 23897682908\n## 0.2763x<\/pre><\/div>\n<\/li>\n<\/ol><p>Surprisingly, #1, reversing filenames in order to sort on the suffixes, turns out to be even worse than not sorting at all. The improved attempt to sort on filetypes doesn\u2019t do much better, although it at least beats the baseline of no-sorting; it may be that to get a compression win real semantic knowledge of filetypes will be needed (perhaps calling <code>file<\/code> on every file and sorting by the detected filetype?). The regular sort also performs surprisingly well, but my intuitions win for once and it\u2019s beaten by my previous strategy of sorting by the middle of domains. Finally, the winner is a bit of a surprise too, a sort I only threw in out of curiosity because I noticed blogs tend to have similar site layouts.<\/p><p>In this case, the best version saved 24268747400\u221223897682908=371064492 or 371MB over the unsorted version. Not as impressive as in the next use case, but enough to show this seems like a real gain<\/p><\/section><section id=\"separate-mirrors\" class=\"level3\"><h3><a href=\"#separate-mirrors\" title=\"Link to section: \u00a7'Separate mirrors'\">Separate mirrors<\/a><\/h3><p>Top-level domains are not the only thing we might want to sort differently on. To take my mirrors of darknet market drug sites such as : I download a site each time as a separate <code>wget<\/code> run in a timestamped folder. So in my Silk Road 2 folder, I have both <code>2013-12-25\/<\/code> &amp; <code>2014-01-15\/<\/code>. These share many similar &amp; identical files so they compress together with <code>xz<\/code> down from 1.8GB to 0.3GB.<\/p><p>But they <em>could<\/em> compress even better: the similar files may be thousands of files and hundreds of megabytes away by alphabetical or file-inode order, so even with a very large window and a top-notch compressor, it will fail to spot many long-range redundancies. In between <code>2013-12-25\/default.css<\/code> and <code>2014-01-15\/default.css<\/code> is going to be all sorts of files which have nothing to do with CSS, like <code>2014-01-16\/items\/2-grams-of-pure-vanilla-ketamine-powder-dutchdope?vendor_feedback_page=5<\/code> and <code>2014-01-16\/items\/10-generic-percocet-10-325-plus-1-30mg-morphine<\/code>. You see the problem.<\/p><p>Because we sort the files by \u2018all files starting with \u201c2013\u201d\u2019 and then \u2018all files starting \u201c2014\u201d\u2019, we lose all proximity. If instead, we could sort by subfolder and only then by the top-level folder, then we\u2019d have everything line up nicely. Fortunately, we already know how to do this! Reuse the sort-key trick, specify \u201c\/\u201d as the delimiter to parse on, and the nth field to sort on. We feed it a file list, tell it to break filenames by \u201c\/\u201d, and then to sort on a lower level, and if we did it right, we will indeed get output like <code>2013-12-25\/default.css<\/code> just before <code>2014-01-15\/default.css<\/code>, which will do wonders for our compression, and which will pay ever more dividends as we accumulate more partially-redundant mirrors.<\/p><p>Here is an example of output for my Pandora mirrors, where, due to frequent rumors of its demise triggering mirroring on my part, I have 5 full mirrors at the time of testing; and naturally, if we employ the sort-key trick (<code>find . -type f | sort --key=3 --field-separator=\"\/\"<\/code>), we find a lot of similar-sounding files:<\/p><pre>.\/2014-01-15\/profile\/5a66e5238421f0422706b267b735d2df\/6\n.\/2014-01-16\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\n.\/2013-12-25\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\/1\n.\/2014-01-15\/profile\/5a9df4f5482d55fb5a8997c270a1e22d.1\n.\/2013-12-25\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\/2\n.\/2014-01-15\/profile\/5a9df4f5482d55fb5a8997c270a1e22d.2\n.\/2013-12-25\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\/3\n.\/2014-01-15\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\/4\n.\/2014-01-15\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\/5\n.\/2014-01-15\/profile\/5a9df4f5482d55fb5a8997c270a1e22d\/6\n.\/2013-12-25\/profile\/5abb81db167294478a23ca110284c587\n.\/2013-12-25\/profile\/5acc44d370e305e252dd4e2b91fda9d0\/1\n.\/2014-01-15\/profile\/5acc44d370e305e252dd4e2b91fda9d0.1\n.\/2013-12-25\/profile\/5acc44d370e305e252dd4e2b91fda9d0\/2\n.\/2014-01-15\/profile\/5acc44d370e305e252dd4e2b91fda9d0.2<\/pre><p>Note the interleaving of 5 different mirrors, impossible in a normal left-to-right alphabetical sort. You can bet that these 4 files (in 15 versions) are going to compress much better than if they were separated by a few thousand other profile pages.<\/p><p>So here\u2019s an example invocation (doing everything in pipelines to avoid disk IO):<\/p><div class=\"sourceCode\" id=\"cb25\"><pre class=\"sourceCode r\">find . -type f -print0 |sort --zero-terminated --key=3 --field-separator=\"\/\"\n |tar --no-recursion --null --files-from --c\n |xz -9 --extreme --stdout\n &gt;..\/mirror.tar.xz<\/pre><\/div><p>Used on my two Silk Road 2 mirrors which together weigh 1800M, a normal run without the <code>--key<\/code>\/<code>--field-separator<\/code> options, as mentioned before, yields a 308M archive. That\u2019s not too bad. Certainly much better than hauling around almost 2GB. However\u2014if I switch to the sort-key trick, however, the archive is now 271M, or 37M less. Same compression algorithm, same files, same unpacked results, same speed, just 2 little obscure <code>sort<\/code> options\u2026 and I get an archive 87% the size of the original.<\/p><p>Not impressed? Well, I did say that the advantage increases with the number of mirrors to extract redundancy from. With only 2 mirrors, the SR2 results can\u2019t be too impressive. How about the Pandora mirrors? 5 of them gives the technique more scope to shine. And as expected, it\u2019s even more impressive when I compare the Pandora archives: 71M vs 162M. The sort-keyed archive saves 56% of the regular archive\u2019s size. The final Pandora archive, , boasts 105 mirrors from 2013-12-25 to 2014-11-05 of 1,745,778 files (many duplicates due to being present over multiple crawls) taking up 32GB uncompressed. This archive is 0.58GB (606,203,320 bytes) when sort-keyed; when sorted alphabetically, it is instead 7.5GB (7,466,374,836 bytes) or 12.31x larger, so the sort-keyed archive saves 92%. Forum mirrors are even more redundant than market mirrors because all content added to a forum will be present in all subsequent crawls, so each crawl will yield the previous crawl plus some additional posts; hence, the Pandora forum\u2019s raw size of 8,894,521,228 bytes shrinks to 283,747,148 bytes when compressed alphabetically but shrinks dramatically to the 31x smaller sort-keyed archive of 9,084,508 bytes, saving 97%!<\/p><p>Clearly in these DNM use-cases, sort-keying has gone from a cute trick to indispensable.<\/p><\/section><section id=\"alternatives\" class=\"level3\"><h3><a href=\"#alternatives\" title=\"Link to section: \u00a7'Alternatives'\">Alternatives<\/a><\/h3><p>The sort-key trick is most useful when we can infer a lot of spatial locality just from parts of the file names, it\u2019s not a panacea. There are other approaches:<\/p><ol type=\"1\"><li>\n<p>if we have multiple temporally-ordered datasets and we don\u2019t mind making it more difficult to access older copies, it may be simpler to store the data as a DVCS like  where each dataset is a large patch<\/p>\n<\/li>\n<li>\n<p>sort files by minimizing binary differences between them using Timm S. M\u00fcller\u2019s <a href=\"https:\/\/www.gwern.net\/docs\/www\/neoscientists.org\/9802872843f29498229b95588e7bdebd7d86cad8.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/neoscientists.org\/~tmueller\/binsort\/\" title=\"(Original URL: http:\/\/neoscientists.org\/~tmueller\/binsort\/ )\"><code>binsort<\/code><\/a> utility<\/p>\n<p>The default optimization setting of <code>-o15<\/code> underperforms the sort-key trick on both the SR2 and Pandora datasets by 5\u201310MB, and a higher setting like <code>-o1000<\/code> is best. Note that <code>binsort<\/code> is O(n2) in number of files, so it\u2019s limited to sorting &lt;10,000 files. An example pipeline for compressing posts from the Silk Road forums regarding <a href=\"https:\/\/old.reddit.com\/r\/SilkRoad\/comments\/1yzndq\/a_trip_down_memory_lane_vladimir_limetless_and\/\">a minor Ponzi scheme there<\/a>, since the filenames offer little guidance for a semantically-meaningful sort:<\/p>\n<div class=\"sourceCode\" id=\"cb26\"><pre class=\"sourceCode bash\">find dkn255hz262ypmii.onion\/ -type f -exec fgrep -i -l vladimir  {} \\; &gt;&gt; ponzi.txt\nfind dkn255hz262ypmii.onion\/ -type f -exec fgrep -i -l Limetless {} \\; &gt;&gt; ponzi.txt\nmkdir ponzi\/ &amp;&amp; cp `cat ponzi.txt` ponzi\/\n~\/bin\/binsort\/binsort -t 6 -o 1000 ponzi\/ &gt; ponzi-list.txt\ncat ponzi-list.txt | tar --no-recursion --files-from - -c | xz -9 --extreme --stdout &gt; ~\/srf1-ponzi.tar.xz<\/pre><\/div>\n<p>(Embarrassingly, <code>binsort<\/code> seems to underperform on this dataset: the files are 380M uncompressed, 3.536M sort-keyed, and 3.450M sorted.)<\/p>\n<\/li>\n<li>\n<p>if we know there are many duplicates but they are far apart by a lexicographic sort and we don\u2019t have any good way to sort filenames and a lot of RAM, we can try out a compression tool which specifically looks for long-range redundancies throughout the entire bitstream, such as  (<a href=\"https:\/\/www.gwern.net\/docs\/www\/ck.kolivas.org\/47782d34488b34eef4ba08da0d7c26063f9242d2.html\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/ck.kolivas.org\/apps\/lrzip\/\" title=\"(Original URL: http:\/\/ck.kolivas.org\/apps\/lrzip\/ )\">homepage<\/a>\/<a href=\"https:\/\/wiki.archlinux.org\/index.php\/Lrzip\">Arch wiki<\/a>). <code>lrzip<\/code> is packaged in Debian and should be at least as good as <code>xz<\/code> since it too uses LZMA; but it is an obscure tool which is not a default install on Linux distributions like <code>xz<\/code> is, and this makes distributing archives to other people difficult.<\/p>\n<\/li>\n<\/ol><\/section><section id=\"external-links-1\" class=\"level3\"><h3><a href=\"#external-links-1\" title=\"Link to section: \u00a7'External Links'\">External Links<\/a><\/h3><ul><li><a href=\"https:\/\/stackoverflow.com\/questions\/357560\/sorting-multiple-keys-with-unix-sort\">\u201cSorting multiple keys with Unix <code>sort<\/code>\u201d<\/a><\/li>\n<li><a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/download?doi=10.1.1.473.7179&amp;rep=rep1&amp;type=pdf\" title=\"Sadowski &amp; Levin 2007\">\u201cSimHash: Hash-based Similarity Detection\u201d<\/a> (see also )<\/li>\n<\/ul><\/section><\/section><\/section><section class=\"footnotes\" role=\"doc-endnotes\"><hr \/><ol><li id=\"fn1\" role=\"doc-endnote\">\n<p>I use <a href=\"http:\/\/duplicity.nongnu.org\/\"><code>duplicity<\/code><\/a> &amp; <a href=\"http:\/\/www.nongnu.org\/rdiff-backup\/\"><code>rdiff-backup<\/code><\/a> to backup my entire home directory to a cheap 1.5TB hard drive (bought from Newegg using <code>forre.st<\/code>\u2019s <a href=\"https:\/\/www.gwern.net\/docs\/www\/forre.st\/07963fe85eed8b845122c702ce0cbe668816ac02.html#hdd\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/forre.st\/storage#hdd\" title=\"(Original URL: http:\/\/forre.st\/storage#hdd )\">\u201cStorage Analysis\u2014GB\/$ for different sizes and media\u201d<\/a> price-chart); a limited selection of folders are backed up to  B2 using <code>duplicity<\/code>.<\/p>\n<p>I used to semiannually tar up my important folders, add  redundancy, and burn them to DVD, but that\u2019s no longer really feasible; if I ever get a Blu-ray burner, I\u2019ll resume WORM backups. (Magnetic media doesn\u2019t strike me as reliable over many decades, and it would ease my mind to have optical backups.)<a href=\"#fnref1\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn2\" role=\"doc-endnote\">\n<p><a href=\"https:\/\/web.archive.org\/web\/20121108093008\/http:\/\/www.wired.com\/politics\/security\/commentary\/securitymatters\/2008\/02\/securitymatters_0221\">\u201cWhen the Internet Is My Hard Drive, Should I Trust Third Parties?\u201d<\/a>, <em>Wired<\/em>:<\/p>\n<blockquote>\n<p>Bits and pieces of the web disappear all the time. It\u2019s called \u2018link rot\u2019, and we\u2019re all used to it. A friend saved 65 links in 1999 when he planned a trip to Tuscany; only half of them still work today. In my own blog, essays and news articles and websites that I link to regularly disappear \u2013 sometimes within a few days of my linking to them.<\/p>\n<\/blockquote>\n<a href=\"#fnref2\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/li>\n<li id=\"fn3\" role=\"doc-endnote\">\n<p><a href=\"https:\/\/www.gwern.net\/docs\/www\/scimaps.org\/d0b8798bc2a55f9a99ff87a4cb002bb043702294.pdf\" rel=\"archived alternate nofollow\" data-url-original=\"http:\/\/scimaps.org\/exhibit\/docs\/dellawalle.pdf\" title=\"(Original URL: http:\/\/scimaps.org\/exhibit\/docs\/dellawalle.pdf )\">\u201cGoing, Going, Gone: Lost Internet References\u201d<\/a>; abstract:<\/p>\n<blockquote>\n<p>The extent of Internet referencing and Internet reference activity in medical or scientific publications was systematically examined in more than 1000 articles published between 2000 and 2003 in the New England Journal of Medicine, The Journal of the American Medical Association, and Science. Internet references accounted for 2.6% of all references (672\/25548) and in articles 27 months old, 13% of Internet references were inactive.<\/p>\n<\/blockquote>\n<a href=\"#fnref3\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/li>\n<li id=\"fn4\" role=\"doc-endnote\">\n<p>The Million Dollar Homepage still gets a surprising amount of traffic, so one fun thing one could do is buy up expired domains which paid for particularly large links.<a href=\"#fnref4\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn5\" role=\"doc-endnote\">\n<p>By 6 January 2013, the number has increased to ~12000 external links, ~7200 to non-Wikipedia domains.<a href=\"#fnref5\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn6\" role=\"doc-endnote\">\n<p>If each link has a fixed chance of dying in each time period, such as 3%, then the total risk of death is an ; over the time period 2011\u20132070 the cumulative chance it will beat each of the 3% risks is 0.1658. So in 2070, how many of the 2200 links will have beat the odds? Each link is independent, so they are like flipping a biased coin and form a . The binomial distribution, being discrete, has no easy equation, so we just ask R how many links survive at the 5th percentile\/quantile (a lower bound) and how many survive at the 95th percentile (an upper bound):<\/p>\n<div class=\"sourceCode\" id=\"cb1\"><pre class=\"sourceCode r\">qbinom(c(0.05, 0.95), 2200, 0.97^(2070-2011))\n# [1] 336 394\n## the 50% annual link rot hypothetical:\nqbinom(c(0.05, 0.50), 2200, 0.50^(2070-2011))\n# [1] 0 0<\/pre><\/div>\n<a href=\"#fnref6\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/li>\n<li id=\"fn7\" role=\"doc-endnote\">\n<p>101, \u2018Buddhism related to Blossoms\u2019; ; trans. Steven D. Carter, ISBN 0-231-10576-2<a href=\"#fnref7\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn8\" role=\"doc-endnote\">\n<p>Which I suspect is only accidentally \u2018general\u2019 and would shut down access if there were some other way to ensure that Wikipedia external links still got archived.<a href=\"#fnref8\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn9\" role=\"doc-endnote\">\n<p>Since Pinboard is a bookmarking service more than an archive site, I asked whether treating it as such would be acceptable and Maciej said \u201cYour current archive size, growing at 20 GB a year, should not be a problem. I\u2019ll put you on the heavy-duty server where my own stuff lives.\u201d<a href=\"#fnref9\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn10\" role=\"doc-endnote\">\n<p>Google Cache is generally recommended only as a last ditch resort because pages expire quickly from it. Personally, I\u2019m convinced that Google would never just delete colossal amounts of Internet data\u2014this is Google, after all, the epitome of storing unthinkable amounts of data\u2014and that Google Cache merely ceases to make public its copies. And to request a Google spider visit, one has to solve a CAPTCHA\u2014so that\u2019s not a scalable solution.<a href=\"#fnref10\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn11\" role=\"doc-endnote\">\n<p>Which would not be publicly accessible or submittable; I know they exist, but because they hide themselves, I know only from random comments online eg. <a href=\"https:\/\/news.ycombinator.com\/item?id=2880427\">\u201cyears ago a friend of mine who I\u2019d lost contact with caught up with me and told me he found a cached copy of a website I\u2019d taken down in his employer\u2019s equivalent to the Wayback Machine. His employer was a branch of the federal government.\u201d<\/a>.<a href=\"#fnref11\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn12\" role=\"doc-endnote\">\n<p><a href=\"http:\/\/hackage.haskell.org\/package\/archiver-0.1\">Version 0.1<\/a> of my <code>archiver<\/code> daemon didn\u2019t simply read the file until it was empty and exit, but actually watched it for modifications with . I removed this functionality when I realized that the required WebCite choking (just one URL every ~25 seconds) meant that <code>archiver<\/code> would <em>never<\/em> finish any reasonable workload.<a href=\"#fnref12\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn13\" role=\"doc-endnote\">\n<p>Much easier than it was in the past;  records his travails with the <em>previous<\/em> Mozilla history format in the aptly-named <a href=\"http:\/\/www.jwz.org\/blog\/2004\/03\/when-the-database-worms-eat-into-your-brain\/\">\u201cwhen the database worms eat into your brain\u201d<\/a>.<a href=\"#fnref13\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn14\" role=\"doc-endnote\">\n<p>An older <a href=\"https:\/\/web.archive.org\/web\/20100628055041\/http:\/\/code.google.com\/speed\/articles\/web-metrics.html\" title=\"Web metrics: Size and number of resources\">2010 Google article<\/a> put the average at 320kb, but that was an average over the entire Web, including all the old content.<a href=\"#fnref14\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn15\" role=\"doc-endnote\">\n<p>Already one runs old games like the classic  in emulators of the DOS operating system like ; but those emulators will not always be maintained. Who will emulate the emulators? Presumably in 2050, one will instead emulate some ancient but compatible OS\u2014Windows 7 or Debian 6.0, perhaps\u2014and inside <em>that<\/em> run DOSBox (to run the DOS which can run the game).<a href=\"#fnref15\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/p>\n<\/li>\n<li id=\"fn16\" role=\"doc-endnote\">\n<p><a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/download?doi=10.1.1.478.96&amp;rep=rep1&amp;type=pdf\">\u201cBigTable: A Distributed Storage System for Structured Data\u201d<\/a>, Chang et al 2006:<\/p>\n<blockquote>\n<p>BigTable maintains data in lexicographic order by row key. The row range for a table is dynamically partitioned. Each row range is called a <em>tablet<\/em>, which is the unit of distribution and load balancing. As a result, reads of short row ranges are efficient and typically require communication with only a small number of machines. Clients can exploit this property by selecting their row keys so that they get good locality for their data accesses. For example, in Webtable, pages in the same domain are grouped together into contiguous rows by reversing the hostname components of the URLs. For example, we store data for <code>maps.google.com\/index.html<\/code> under the key <code>com.google.maps\/index.html<\/code>. Storing pages from the same domain near each other makes some host and domain analyses more efficient.<\/p>\n<\/blockquote>\n<a href=\"#fnref16\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9\ufe0e<\/a><\/li>\n<\/ol><\/section>","created_at":"2020-09-08T12:05:21+0000","updated_at":"2020-09-08T12:05:21+0000","published_by":["Site"],"annotations":[],"mimetype":"text\/html","reading_time":45,"domain_name":"www.gwern.net","preview_picture":"https:\/\/www.gwern.net\/images\/logo-whitebg-large-border.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["procgen","gamedev"],"is_public":false,"id":1198,"title":"Introducing Generate Worlds for Procedural Generation without Code","url":"https:\/\/ijdykeman.github.io\/procedural_generation\/2019\/05\/17\/introducing-generate-worlds.html","content":"<p>The infinite procedurally generated worlds of games like Minecraft and Dwarf Fortress create a sense of exploration and variety that entirely hand-built game worlds cannot compete with. However, designing and coding the algorithms to actually perform this generation is a specialized technical skill few possess. At the same time, numerous online communities exist around mapping, painting, and otherwise imagining fantasy environments, and some great products exist to help them do it, like <a href=\"https:\/\/www.wonderdraft.net\/\">Wonderdraft<\/a> for drawing fantasy maps. But something is missing: I want to experience these worlds from the inside, and drawing tools don\u2019t let me do it.<\/p>\n<p>My desire to design and explore fantasy landscapes inspired me to build the procedurally generated island worlds of <a href=\"http:\/\/www.brimmingsea.com\/\">Brimming Sea<\/a> in 2014. It was a difficult technical task, but imagining and then walking around these islands was magical. Since then, I\u2019ve created several tools to help me design infinite procedural worlds <em>without writing any code<\/em>, and I\u2019m finally happy enough with one of those tools to release it. I\u2019m calling it Generate Worlds, and it\u2019s available on the itch.io store.<\/p>\n<p>Your browser does not support the video tag.<\/p>\n<p>The video above shows Generate Worlds in action building a landscape and dungeon environment. The video below briefly describes Generate Worlds. Read on for the details.<\/p>\n<iframe width=\"100%\" height=\"370\" src=\"https:\/\/www.youtube.com\/embed\/DrAtX-EsQM0?autoplay=0&amp;showinfo=0\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe>\n<p>The key idea behind Generate Worlds is that instead of writing code that describes how to build the world, you simply create a set of rules about how you want the world to work and Generate Worlds puts it together for you. Let\u2019s start with a simple example.<\/p>\n<p>Imagine we want to create a simple dungeon world containing hallways and rectangular rooms. We will make the world by assembling these six passageway and room pieces:<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/rule_based_procedural_generation\/dungeon_parts.svg\" alt=\"dungeon parts\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>I\u2019ll now add a series of rules to put these pieces together into a coherent dungeon, starting with an unstrucutred world made of those pieces but without any rules about how to put them together:<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/rule_based_procedural_generation\/full_dungeon_sequence.svg\" alt=\"dungeon construction\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>We now have a reasonable looking, simple dungeon layout. To recap, the rules are simply<\/p>\n<ol><li>Passageways can connect to other passageways.<\/li>\n<li>All floor space must be enclosed by walls.<\/li>\n<li>Passageways can connect to doors.<\/li>\n<\/ol><p>Creating these rules is easy compared to writing a dungeon generation algorithm, but it\u2019s hard to convert an english statement like \u201cPassageways can connect to other passageways\u201d to a world, so I\u2019ll instead rely on a 3D visual language. This language consists of voxel <em>tiles<\/em> that the user provides. These tiles are simple .vox files that a user creates in some voxel editing software and loads into Generate Worlds. Imagine we want to create a world with walled cities and countryside. Here\u2019s what the tiles for that world might look like:<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/rule_based_procedural_generation\/labeled_woods_tiles.svg\" alt=\"town and country tiles\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>The world should have the following properties:<\/p>\n<ul><li>Rivers and roads have no dead ends<\/li>\n<li>Roads can cross rivers over a bridge<\/li>\n<li>Roads can cross into cities through a gate<\/li>\n<li>Walls separate the outside grassy countryside from town interiors<\/li>\n<li>Grassy areas can contain huts and trees<\/li>\n<\/ul><p>To enforce these all these rules, Generate Worlds does only one simple thing: it ensures that two tiles are only placed next to each other in the world if the tiles are the same color in all the places where they touch, as in the example below.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/rule_based_procedural_generation\/road_assembling.svg\" alt=\"two road tiles assembling\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p><em>In the figure above, we can arrange these two tiles in this way because the tiles\u2019 colors match where they touch.<\/em><\/p>\n<p>Placing several tiles looks like this:<\/p>\n<p>Your browser does not support the video tag.<\/p>\n<p>These tiles are easy to create in a voxel editor like MagicaVoxel. The tile set above is simply a directory of .vox files that Generate Worlds takes as input.<\/p>\n<h2 id=\"generating-a-dungeon\">Generating a Dungeon<\/h2>\n<p>Let\u2019s try a more sophisticated world: a 3D dungeon that extends endlessly in all directions. We can simply provide Generate Worlds with a new tile set of dungeon parts. Nine of those tiles are shown below. They contain pieces of passageways, rooms, stairs, columns, arcades, and doorways.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/rule_based_procedural_generation\/dungeon_tiles.svg\" alt=\"dungeon tiles\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>Putting them together in 3D looks like this:<\/p>\n<p>Your browser does not support the video tag.<\/p>\n<p>Loading these .vox files into Generate Worlds, I can explore this dungeon world in first-person. Generate Worlds lets you set lighting conditions, and place lights.<\/p>\n<p>Your browser does not support the video tag.<\/p>\n<h2 id=\"conclusion\">Conclusion<\/h2>\n<p>My goal is to build a creative tool that isn\u2019t quite like anything already in existence. Minecraft lets you modify worlds block by block and explore them, but it doesn\u2019t let you redefine how the world is created in the first place, limiting the scope of what you can build. Mapping and drawing tools help you imagine and design worlds, but don\u2019t let you explore them. Finally, coding your own algorithms gives you huge flexibility, but at the cost of much higher difficulty. And for what it\u2019s worth, the worlds I\u2019ve made in Generate Worlds are more varied and compelling than what I\u2019ve been able to create by hand-building procedural generation algorithms in the past.<\/p>\n<p>If you want to explore this creative avenue, consider supporting this effort by purchasing Generate Worlds. A <a href=\"http:\/\/generateworlds.com\/tutorial.html\">tutorial<\/a> is available to get you started. If you have any issues using it, you can contact me at ijdykeman@gmail.com .<\/p>","created_at":"2020-09-03T15:24:23+0000","updated_at":"2020-09-03T15:24:32+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"ijdykeman.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["procgen","games"],"is_public":false,"id":1197,"title":"Brimming Sea","url":"http:\/\/www.brimmingsea.com\/","content":"<div id=\"contentVisual\" class=\"right\"><img src=\"http:\/\/www.brimmingsea.com\/images\/castleInPoplarForest.jpg\" width=\"560\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div>Third Person Command Order troops to build, work, and fight. Design castles, order the construction of fleets, conscript armies and sail them to hostile islands to expand your empire. Brimming Sea is a rich strategic experience.<div class=\"left\"><img src=\"http:\/\/www.brimmingsea.com\/images\/firstPersonForestWithFarm.jpg\" width=\"560\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div>First Person Control You can take first person control of any character at any time, and do anything at all. Work and fight alongside your troops. Experience the cities you create through their eyes. Turn the tide of a battle by jumping in and leading from the front. Brimming Sea does not turn you into a passive leader.<div class=\"right\"><img src=\"http:\/\/www.brimmingsea.com\/images\/SwordmanLookingIntoForest.jpg\" width=\"560\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div>Conquer Traverse the game's infinite, procedurally generated world in search of resources. But before you can claim them, you must defeat the island's previous inhabitants.<div class=\"left\"><img src=\"http:\/\/www.brimmingsea.com\/images\/towersWithArch.jpg\" width=\"560\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div>Fortify Build defensive structures to hold your islands against monsters. Design structures and watch your troops build them.<div class=\"right\"><img src=\"http:\/\/www.brimmingsea.com\/images\/farmingImage.jpg\" width=\"560\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div>Grow To fund and feed your growing empire, you will need to create settlements, defenses, and farms on the islands you conquer.","created_at":"2020-09-03T15:24:12+0000","updated_at":"2020-09-03T15:24:17+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.brimmingsea.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["swagger","openapi","codegen"],"is_public":false,"id":1196,"title":"Redocly\/redoc","url":"https:\/\/github.com\/Redocly\/redoc","content":"<div style=\"margin: auto;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/Redocly\/redoc\/master\/docs\/images\/redoc-logo.png\"><img alt=\"ReDoc logo\" src=\"https:\/\/raw.githubusercontent.com\/Redocly\/redoc\/master\/docs\/images\/redoc-logo.png\" width=\"400px\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><p><strong>OpenAPI\/Swagger-generated API Reference Documentation<\/strong><\/p><p><a href=\"https:\/\/travis-ci.org\/Redocly\/redoc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7055376877136ec28edfe95553df8f3bb0d8f03b\/68747470733a2f2f7472617669732d63692e6f72672f5265646f636c792f7265646f632e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/Redocly\/redoc.svg?branch=master\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/coveralls.io\/github\/Redocly\/redoc?branch=master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/ddeb6a4542b081bfb5826e4b55eaad859fd393ab\/68747470733a2f2f636f766572616c6c732e696f2f7265706f732f5265646f636c792f7265646f632f62616467652e7376673f6272616e63683d6d617374657226736572766963653d676974687562\" alt=\"Coverage Status\" data-canonical-src=\"https:\/\/coveralls.io\/repos\/Redocly\/redoc\/badge.svg?branch=master&amp;service=github\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/david-dm.org\/Redocly\/redoc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/54e368858df7dbfd958261fa2e2c3f27f8bc8884\/68747470733a2f2f64617669642d646d2e6f72672f5265646f636c792f7265646f632f7374617475732e737667\" alt=\"dependencies Status\" data-canonical-src=\"https:\/\/david-dm.org\/Redocly\/redoc\/status.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/david-dm.org\/Redocly\/redoc#info=devDependencies\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5eb34681aacf9006cbec8b62c4546163879c6999\/68747470733a2f2f64617669642d646d2e6f72672f5265646f636c792f7265646f632f6465762d7374617475732e737667\" alt=\"devDependencies Status\" data-canonical-src=\"https:\/\/david-dm.org\/Redocly\/redoc\/dev-status.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/www.npmjs.com\/package\/redoc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/4306fa23c45e505bb8fb9dacae5e36174a80c643\/687474703a2f2f696d672e736869656c64732e696f2f6e706d2f762f7265646f632e737667\" alt=\"npm\" data-canonical-src=\"http:\/\/img.shields.io\/npm\/v\/redoc.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/LICENSE\"><img src=\"https:\/\/camo.githubusercontent.com\/8ecf668111d7d99738c506843d9a5c9b20e74d67\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f6c2f7265646f632e737667\" alt=\"License\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/l\/redoc.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a href=\"https:\/\/cdn.jsdelivr.net\/npm\/redoc\/bundles\/redoc.standalone.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5905967bdf763e5d898fdc8fcf3be4321ae055b0\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f63646e2e6a7364656c6976722e6e65742f6e706d2f7265646f632f62756e646c65732f7265646f632e7374616e64616c6f6e652e6a733f636f6d7072657373696f6e3d677a6970266d61783d333030303030\" alt=\"bundle size\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/cdn.jsdelivr.net\/npm\/redoc\/bundles\/redoc.standalone.js?compression=gzip&amp;max=300000\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/www.npmjs.com\/package\/redoc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/77ad68545bd64249682e4225733abeaa65aca56b\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f646d2f7265646f632e737667\" alt=\"npm\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/dm\/redoc.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/www.jsdelivr.com\/package\/npm\/redoc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d206cdbb0785bc991408c8c8e8d6f68ec4c958c8\/68747470733a2f2f646174612e6a7364656c6976722e636f6d2f76312f7061636b6167652f6e706d2f7265646f632f6261646765\" alt=\"\" data-canonical-src=\"https:\/\/data.jsdelivr.com\/v1\/package\/npm\/redoc\/badge\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/hub.docker.com\/r\/redocly\/redoc\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/ba23888de7d693df98f3805a9aae6f42ac35fd9f\/68747470733a2f2f696d672e736869656c64732e696f2f646f636b65722f6275696c642f7265646f636c792f7265646f632e737667\" alt=\"Docker Build Status\" data-canonical-src=\"https:\/\/img.shields.io\/docker\/build\/redocly\/redoc.svg\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><\/div><p><strong>This is README for <code>2.0<\/code> version of ReDoc (React based). README for <code>1.x<\/code> version is on the branch <a href=\"https:\/\/github.com\/Redocly\/redoc\/tree\/v1.x\">v1.x<\/a><\/strong><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/Redocly\/redoc\/master\/demo\/redoc-demo.png\"><img src=\"https:\/\/raw.githubusercontent.com\/Redocly\/redoc\/master\/demo\/redoc-demo.png\" alt=\"ReDoc demo\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a href=\"https:\/\/github.com\/Rebilly\/generator-openapi-repo#generator-openapi-repo--\"><img alt=\"Deploy to Github\" src=\"https:\/\/camo.githubusercontent.com\/ec475eed57929f3340c0d1968ccdba7eb381fb49\/687474703a2f2f692e696d6775722e636f6d2f595a6d61716b332e706e67\" height=\"60px\" data-canonical-src=\"http:\/\/i.imgur.com\/YZmaqk3.png\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/redoc.ly\" rel=\"nofollow\"><img alt=\"ReDoc as a service\" src=\"https:\/\/camo.githubusercontent.com\/3ac6f8be6c62fa5df2677de417e2205879e4ec27\/687474703a2f2f692e696d6775722e636f6d2f656471644376362e706e67\" height=\"60px\" data-canonical-src=\"http:\/\/i.imgur.com\/edqdCv6.png\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/redoc.ly\/#services\" rel=\"nofollow\"><img alt=\"Customization services\" src=\"https:\/\/camo.githubusercontent.com\/99b9b3325a0dc0d35dbe56a1a564adccfa05f27d\/687474703a2f2f692e696d6775722e636f6d2f6334735546374d2e706e67\" height=\"60px\" data-canonical-src=\"http:\/\/i.imgur.com\/c4sUF7M.png\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2><ul><li>Extremely easy deployment<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/cli\/README.md\">redoc-cli<\/a> with ability to bundle your docs into <strong>zero-dependency<\/strong> HTML file<\/li>\n<li>Server Side Rendering ready<\/li>\n<li>The widest OpenAPI v2.0 features support (yes, it supports even <code>discriminator<\/code>)<br \/><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/images\/discriminator-demo.gif\"><img src=\"https:\/\/github.com\/Redocly\/redoc\/raw\/master\/docs\/images\/discriminator-demo.gif\" alt=\"\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/li>\n<li>OpenAPI 3.0 support<\/li>\n<li>Neat <strong>interactive<\/strong> documentation for nested objects<br \/><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/images\/nested-demo.gif\"><img src=\"https:\/\/github.com\/Redocly\/redoc\/raw\/master\/docs\/images\/nested-demo.gif\" alt=\"\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/li>\n<li>Code samples support (via vendor extension)<br \/><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/images\/code-samples-demo.gif\"><img src=\"https:\/\/github.com\/Redocly\/redoc\/raw\/master\/docs\/images\/code-samples-demo.gif\" alt=\"\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/li>\n<li>Responsive three-panel design with menu\/scrolling synchronization<\/li>\n<li>Integrate API Introduction into side menu - ReDoc takes advantage of markdown headings from OpenAPI description field. It pulls them into side menu and also supports deep linking.<\/li>\n<li>High-level grouping in side-menu via <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-tagGroups\"><code>x-tagGroups<\/code><\/a> vendor extension<\/li>\n<li>Simple integration with <code>create-react-app<\/code> (<a href=\"https:\/\/github.com\/APIs-guru\/create-react-app-redoc\">sample<\/a>)<\/li>\n<li>Branding\/customizations via <a href=\"#redoc-options-object\"><code>theme<\/code> option<\/a><\/li>\n<\/ul><h2><a id=\"user-content-roadmap\" class=\"anchor\" aria-hidden=\"true\" href=\"#roadmap\"><\/a>Roadmap<\/h2><ul class=\"contains-task-list\"><li class=\"task-list-item\"><del><a href=\"https:\/\/github.com\/Redocly\/redoc\/issues\/312\">OpenAPI v3.0 support<\/a><\/del><\/li>\n<li class=\"task-list-item\"><del>performance optimizations<\/del><\/li>\n<li class=\"task-list-item\"><del>better navigation (menu improvements + search)<\/del><\/li>\n<li class=\"task-list-item\"><del>React rewrite<\/del><\/li>\n<li class=\"task-list-item\"><del>docs pre-rendering (performance and SEO)<\/del><\/li>\n<li class=\"task-list-item\"> ability to simple branding\/styling<\/li>\n<li class=\"task-list-item\"> built-in API Console<\/li>\n<\/ul><h2><a id=\"user-content-releases\" class=\"anchor\" aria-hidden=\"true\" href=\"#releases\"><\/a>Releases<\/h2><p><strong>Important:<\/strong> all the 2.x releases are deployed to npm and can be used via jsdeliver:<\/p><ul><li>particular release, e.g. <code>v2.0.0-alpha.15<\/code>: <a href=\"https:\/\/cdn.jsdelivr.net\/npm\/redoc@2.0.0-alpha.17\/bundles\/redoc.standalone.js\" rel=\"nofollow\">https:\/\/cdn.jsdelivr.net\/npm\/redoc@2.0.0-alpha.17\/bundles\/redoc.standalone.js<\/a><\/li>\n<li><code>next<\/code> release: <a href=\"https:\/\/cdn.jsdelivr.net\/npm\/redoc@next\/bundles\/redoc.standalone.js\" rel=\"nofollow\">https:\/\/cdn.jsdelivr.net\/npm\/redoc@next\/bundles\/redoc.standalone.js<\/a><\/li>\n<\/ul><p>Additionally, all the 1.x releases are hosted on our GitHub Pages-based CDN <strong>(deprecated)<\/strong>:<\/p><ul><li>particular release, e.g. <code>v1.2.0<\/code>: <a href=\"https:\/\/rebilly.github.io\/ReDoc\/releases\/v1.2.0\/redoc.min.js\" rel=\"nofollow\">https:\/\/rebilly.github.io\/ReDoc\/releases\/v1.2.0\/redoc.min.js<\/a><\/li>\n<li><code>v1.x.x<\/code> release: <a href=\"https:\/\/rebilly.github.io\/ReDoc\/releases\/v1.x.x\/redoc.min.js\" rel=\"nofollow\">https:\/\/rebilly.github.io\/ReDoc\/releases\/v1.x.x\/redoc.min.js<\/a><\/li>\n<li><code>latest<\/code> release: <a href=\"https:\/\/rebilly.github.io\/ReDoc\/releases\/latest\/redoc.min.js\" rel=\"nofollow\">https:\/\/rebilly.github.io\/ReDoc\/releases\/latest\/redoc.min.js<\/a> - it will point to latest 1.x.x release since 2.x releases are not hosted on this CDN but on unpkg.<\/li>\n<\/ul><h2><a id=\"user-content-version-guidance\" class=\"anchor\" aria-hidden=\"true\" href=\"#version-guidance\"><\/a>Version Guidance<\/h2><table><thead><tr><th align=\"left\">ReDoc Release<\/th>\n<th align=\"left\">OpenAPI Specification<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\">2.0.0-alpha.x<\/td>\n<td align=\"left\">3.0, 2.0<\/td>\n<\/tr><tr><td align=\"left\">1.19.x<\/td>\n<td align=\"left\">2.0<\/td>\n<\/tr><tr><td align=\"left\">1.18.x<\/td>\n<td align=\"left\">2.0<\/td>\n<\/tr><tr><td align=\"left\">1.17.x<\/td>\n<td align=\"left\">2.0<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-some-real-life-usages\" class=\"anchor\" aria-hidden=\"true\" href=\"#some-real-life-usages\"><\/a>Some Real-life usages<\/h2><ul><li><a href=\"https:\/\/rebilly-api.redoc.ly\/\" rel=\"nofollow\">Rebilly<\/a><\/li>\n<li><a href=\"https:\/\/docs.docker.com\/engine\/api\/v1.25\/\" rel=\"nofollow\">Docker Engine<\/a><\/li>\n<li><a href=\"https:\/\/www.zuora.com\/developer\/api-reference\/\" rel=\"nofollow\">Zuora<\/a><\/li>\n<li><a href=\"http:\/\/docs.discourse.org\" rel=\"nofollow\">Discourse<\/a><\/li>\n<li><a href=\"https:\/\/www.commbox.io\/api\/\" rel=\"nofollow\">Commbox<\/a><\/li>\n<li><a href=\"https:\/\/apis.guru\/api-doc\/\" rel=\"nofollow\">APIs.guru<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/tiangolo\/fastapi\">FastAPI<\/a><\/li>\n<\/ul><h2><a id=\"user-content-deployment\" class=\"anchor\" aria-hidden=\"true\" href=\"#deployment\"><\/a>Deployment<\/h2><h3><a id=\"user-content-tldr\" class=\"anchor\" aria-hidden=\"true\" href=\"#tldr\"><\/a>TL;DR<\/h3><div class=\"highlight highlight-text-html-basic\"><pre>&lt;!DOCTYPE html&gt;\n&lt;html&gt;\n  &lt;head&gt;\n    &lt;title&gt;ReDoc&lt;\/title&gt;\n    &lt;!-- needed for adaptive design --&gt;\n    &lt;meta charset=\"utf-8\"\/&gt;\n    &lt;meta name=\"viewport\" content=\"width=device-width, initial-scale=1\"&gt;\n    &lt;link href=\"https:\/\/fonts.googleapis.com\/css?family=Montserrat:300,400,700|Roboto:300,400,700\" rel=\"stylesheet\"&gt;\n    &lt;!--\n    ReDoc doesn't change outer page styles\n    --&gt;\n    &lt;style&gt;\n      body {\n        margin: 0;\n        padding: 0;\n      }\n    &lt;\/style&gt;\n  &lt;\/head&gt;\n  &lt;body&gt;\n    &lt;redoc spec-url='http:\/\/petstore.swagger.io\/v2\/swagger.json'&gt;&lt;\/redoc&gt;\n    &lt;script src=\"https:\/\/cdn.jsdelivr.net\/npm\/redoc@next\/bundles\/redoc.standalone.js\"&gt; &lt;\/script&gt;\n  &lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/div><p>That's all folks!<\/p><p><strong>IMPORTANT NOTE:<\/strong> if you work with untrusted user spec, use <code>untrusted-spec<\/code> <a href=\"#redoc-options-object\">option<\/a> to prevent XSS security risks.<\/p><h3><a id=\"user-content-1-install-redoc-skip-this-step-for-cdn\" class=\"anchor\" aria-hidden=\"true\" href=\"#1-install-redoc-skip-this-step-for-cdn\"><\/a>1. Install ReDoc (skip this step for CDN)<\/h3><p>Install using <a href=\"https:\/\/docs.npmjs.com\/getting-started\/what-is-npm\" rel=\"nofollow\">npm<\/a>:<\/p><pre>npm i redoc\n<\/pre><p>or using <a href=\"https:\/\/yarnpkg.com\" rel=\"nofollow\">yarn<\/a>:<\/p><pre>yarn add redoc\n<\/pre><h3><a id=\"user-content-2-reference-redoc-script-in-html\" class=\"anchor\" aria-hidden=\"true\" href=\"#2-reference-redoc-script-in-html\"><\/a>2. Reference redoc script in HTML<\/h3><p>For <strong>CDN<\/strong>:<\/p><div class=\"highlight highlight-text-html-basic\"><pre>&lt;script src=\"https:\/\/cdn.jsdelivr.net\/npm\/redoc\/bundles\/redoc.standalone.js\"&gt; &lt;\/script&gt;<\/pre><\/div><p>For npm:<\/p><div class=\"highlight highlight-text-html-basic\"><pre>&lt;script src=\"node_modules\/redoc\/bundles\/redoc.standalone.js\"&gt; &lt;\/script&gt;<\/pre><\/div><h3><a id=\"user-content-3-add-redoc-element-to-your-page\" class=\"anchor\" aria-hidden=\"true\" href=\"#3-add-redoc-element-to-your-page\"><\/a>3. Add <code>&lt;redoc&gt;<\/code> element to your page<\/h3><div class=\"highlight highlight-text-html-basic\"><pre>&lt;redoc spec-url=\"url\/to\/your\/spec\"&gt;&lt;\/redoc&gt;<\/pre><\/div><h3><a id=\"user-content-4-enjoy-smile\" class=\"anchor\" aria-hidden=\"true\" href=\"#4-enjoy-smile\"><\/a>4. Enjoy \ud83d\ude04<\/h3><h2><a id=\"user-content-usage-as-a-react-component\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage-as-a-react-component\"><\/a>Usage as a React component<\/h2><p>Install peer dependencies required by ReDoc if you don't have them installed already:<\/p><pre>npm i react react-dom mobx@^4.2.0 styled-components core-js\n<\/pre><p>Import <code>RedocStandalone<\/code> component from 'redoc' module:<\/p><div class=\"highlight highlight-source-js\"><pre>import { RedocStandalone } from 'redoc';<\/pre><\/div><p>and use it somewhere in your component:<\/p><div class=\"highlight highlight-source-js\"><pre>&lt;RedocStandalone specUrl=\"url\/to\/your\/spec\"\/&gt;<\/pre><\/div><p>or<\/p><div class=\"highlight highlight-source-js\"><pre>&lt;RedocStandalone spec={\/* spec as an object *\/}\/&gt;<\/pre><\/div><p>Also you can pass options:<\/p><div class=\"highlight highlight-source-js\"><pre>&lt;RedocStandalone\n  specUrl=\"http:\/\/rebilly.github.io\/RebillyAPI\/openapi.json\"\n  options={{\n    nativeScrollbars: true,\n    theme: { colors: { primary: { main: '#dd5522' } } },\n  }}\n\/&gt;<\/pre><\/div><p>Here are detailed <a href=\"#redoc-options-object\">options docs<\/a>.<\/p><p>You can also specify <code>onLoaded<\/code> callback which will be called each time Redoc has been fully rendered or when error occurs (with an error as the first argument). <em>NOTE<\/em>: It may be called multiply times if you change component properties<\/p><div class=\"highlight highlight-source-js\"><pre>&lt;RedocStandalone\n  specUrl=\"http:\/\/rebilly.github.io\/RebillyAPI\/openapi.json\"\n  onLoaded={error =&gt; {\n    if (!error) {\n      console.log('Yay!');\n    }\n  }}\n\/&gt;<\/pre><\/div><p><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/usage-with-ie11.md\"><strong>IE11 Support Notes<\/strong><\/a><\/p><h2><a id=\"user-content-the-docker-way\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-docker-way\"><\/a>The Docker way<\/h2><p>ReDoc is available as pre-built Docker image in official <a href=\"https:\/\/hub.docker.com\/r\/redocly\/redoc\/\" rel=\"nofollow\">Docker Hub repository<\/a>. You may simply pull &amp; run it:<\/p><pre>docker pull redocly\/redoc\ndocker run -p 8080:80 redocly\/redoc\n<\/pre><p>Also you may rewrite some predefined environment variables defined in <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/config\/docker\/Dockerfile\">Dockerfile<\/a>. By default ReDoc starts with demo Petstore spec located at <code>http:\/\/petstore.swagger.io\/v2\/swagger.json<\/code>, but you may change this URL using environment variable <code>SPEC_URL<\/code>:<\/p><pre>docker run -p 8080:80 -e SPEC_URL=https:\/\/api.example.com\/openapi.json redocly\/redoc\n<\/pre><h2><a id=\"user-content-redoc-cli\" class=\"anchor\" aria-hidden=\"true\" href=\"#redoc-cli\"><\/a>ReDoc CLI<\/h2><p><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/cli\/README.md\">See here<\/a><\/p><h2><a id=\"user-content-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration\"><\/a>Configuration<\/h2><h3><a id=\"user-content-security-definition-location\" class=\"anchor\" aria-hidden=\"true\" href=\"#security-definition-location\"><\/a>Security Definition location<\/h3><p>You can inject Security Definitions widget into any place of your specification <code>description<\/code>. Check out details <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/security-definitions-injection.md\">here<\/a>.<\/p><h3><a id=\"user-content-swagger-vendor-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#swagger-vendor-extensions\"><\/a>Swagger vendor extensions<\/h3><p>ReDoc makes use of the following <a href=\"https:\/\/swagger.io\/specification\/#specificationExtensions\" rel=\"nofollow\">vendor extensions<\/a>:<\/p><ul><li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-logo\"><code>x-logo<\/code><\/a> - is used to specify API logo<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-traitTag\"><code>x-traitTag<\/code><\/a> - useful for handling out common things like Pagination, Rate-Limits, etc<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-codeSamples\"><code>x-codeSamples<\/code><\/a> - specify operation code samples<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-examples\"><code>x-examples<\/code><\/a> - specify JSON example for requests<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-nullable\"><code>x-nullable<\/code><\/a> - mark schema param as a nullable<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-displayname\"><code>x-displayName<\/code><\/a> - specify human-friendly names for the menu categories<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-tagGroups\"><code>x-tagGroups<\/code><\/a> - group tags by categories in the side menu<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-servers\"><code>x-servers<\/code><\/a> - ability to specify different servers for API (backported from OpenAPI 3.0)<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-ignoredHeaderParameters\"><code>x-ignoredHeaderParameters<\/code><\/a> - ability to specify header parameter names to ignore<\/li>\n<li><a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/docs\/redoc-vendor-extensions.md#x-additionalPropertiesName\"><code>x-additionalPropertiesName<\/code><\/a> - ability to supply a descriptive name for the additional property keys<\/li>\n<\/ul><h3><a id=\"user-content-redoc-options-object\" class=\"anchor\" aria-hidden=\"true\" href=\"#redoc-options-object\"><\/a><code>&lt;redoc&gt;<\/code> options object<\/h3><p>You can use all of the following options with standalone version on tag by kebab-casing them, e.g. <code>scrollYOffset<\/code> becomes <code>scroll-y-offset<\/code> and <code>expandResponses<\/code> becomes <code>expand-responses<\/code>.<\/p><ul><li><code>disableSearch<\/code> - disable search indexing and search box.<\/li>\n<li><code>expandDefaultServerVariables<\/code> - enable expanding default server variables, default <code>false<\/code>.<\/li>\n<li><code>expandResponses<\/code> - specify which responses to expand by default by response codes. Values should be passed as comma-separated list without spaces e.g. <code>expandResponses=\"200,201\"<\/code>. Special value <code>\"all\"<\/code> expands all responses by default. Be careful: this option can slow-down documentation rendering time.<\/li>\n<li><code>hideDownloadButton<\/code> - do not show \"Download\" spec button. <strong>THIS DOESN'T MAKE YOUR SPEC PRIVATE<\/strong>, it just hides the button.<\/li>\n<li><code>hideHostname<\/code> - if set, the protocol and hostname is not shown in the operation definition.<\/li>\n<li><code>hideLoading<\/code> - do not show loading animation. Useful for small docs.<\/li>\n<li><code>hideSingleRequestSampleTab<\/code> - do not show the request sample tab for requests with only one sample.<\/li>\n<li><code>expandSingleSchemaField<\/code> - automatically expand single field in a schema<\/li>\n<li><code>jsonSampleExpandLevel<\/code> - set the default expand level for JSON payload samples (responses and request body). Special value 'all' expands all levels. The default value is <code>2<\/code>.<\/li>\n<li><code>hideSchemaTitles<\/code> - do not display schema <code>title<\/code> next to to the type<\/li>\n<li><code>simpleOneOfTypeLabel<\/code> - show only unique oneOf types in the label without titles<\/li>\n<li><code>lazyRendering<\/code> - <em>Not implemented yet<\/em> <del>if set, enables lazy rendering mode in ReDoc. This mode is useful for APIs with big number of operations (e.g. &gt; 50). In this mode ReDoc shows initial screen ASAP and then renders the rest operations asynchronously while showing progress bar on the top. Check out the <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/%5Credocly.github.io\/redoc\">demo<\/a> for the example.<\/del><\/li>\n<li><code>menuToggle<\/code> - if true clicking second time on expanded menu item will collapse it, default <code>false<\/code>.<\/li>\n<li><code>nativeScrollbars<\/code> - use native scrollbar for sidemenu instead of perfect-scroll (scrolling performance optimization for big specs).<\/li>\n<li><code>noAutoAuth<\/code> - do not inject Authentication section automatically.<\/li>\n<li><code>onlyRequiredInSamples<\/code> - shows only required fields in request samples.<\/li>\n<li><code>pathInMiddlePanel<\/code> - show path link and HTTP verb in the middle panel instead of the right one.<\/li>\n<li><code>requiredPropsFirst<\/code> - show required properties first ordered in the same order as in <code>required<\/code> array.<\/li>\n<li><code>scrollYOffset<\/code> - If set, specifies a vertical scroll-offset. This is often useful when there are fixed positioned elements at the top of the page, such as navbars, headers etc; <code>scrollYOffset<\/code> can be specified in various ways:\n<ul><li><strong>number<\/strong>: A fixed number of pixels to be used as offset.<\/li>\n<li><strong>selector<\/strong>: selector of the element to be used for specifying the offset. The distance from the top of the page to the element's bottom will be used as offset.<\/li>\n<li><strong>function<\/strong>: A getter function. Must return a number representing the offset (in pixels).<\/li>\n<\/ul><\/li>\n<li><code>showExtensions<\/code> - show vendor extensions (\"x-\" fields). Extensions used by ReDoc are ignored. Can be boolean or an array of <code>string<\/code> with names of extensions to display.<\/li>\n<li><code>sortPropsAlphabetically<\/code> - sort properties alphabetically.<\/li>\n<li><code>suppressWarnings<\/code> - if set, warnings are not rendered at the top of documentation (they still are logged to the console).<\/li>\n<li><code>payloadSampleIdx<\/code> - if set, payload sample will be inserted at this index or last. Indexes start from 0.<\/li>\n<li><code>theme<\/code> - ReDoc theme. Not documented yet. For details check source code: <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/src\/theme.ts\">theme.ts<\/a>.<\/li>\n<li><code>untrustedSpec<\/code> - if set, the spec is considered untrusted and all HTML\/markdown is sanitized to prevent XSS. <strong>Disabled by default<\/strong> for performance reasons. <strong>Enable this option if you work with untrusted user data!<\/strong><\/li>\n<\/ul><h2><a id=\"user-content-advanced-usage-of-standalone-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#advanced-usage-of-standalone-version\"><\/a>Advanced usage of standalone version<\/h2><p>Instead of adding <code>spec-url<\/code> attribute to the <code>&lt;redoc&gt;<\/code> element you can initialize ReDoc via globally exposed <code>Redoc<\/code> object:<\/p><div class=\"highlight highlight-source-js\"><pre>Redoc.init(specOrSpecUrl, options, element, callback?)<\/pre><\/div><ul><li><code>specOrSpecUrl<\/code> is either JSON object with specification or an URL to the spec in <code>JSON<\/code> or <code>YAML<\/code> format<\/li>\n<li><code>options<\/code> <a href=\"#redoc-options-object\">options object<\/a><\/li>\n<li><code>element<\/code> DOM element to put ReDoc into<\/li>\n<li><code>callback<\/code> (optional) - callback to be called after Redoc has been fully rendered. It is also called also on errors with error as the first argument<\/li>\n<\/ul><div class=\"highlight highlight-source-js\"><pre>Redoc.init('http:\/\/petstore.swagger.io\/v2\/swagger.json', {\n  scrollYOffset: 50\n}, document.getElementById('redoc-container'))<\/pre><\/div><hr \/><h2><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h2><p>see <a href=\"https:\/\/github.com\/Redocly\/redoc\/blob\/master\/.github\/CONTRIBUTING.md\">CONTRIBUTING.md<\/a><\/p>","created_at":"2020-08-31T12:53:16+0000","updated_at":"2020-08-31T12:53:23+0000","published_by":["Redocly"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"github.com","preview_picture":"https:\/\/repository-images.githubusercontent.com\/45250726\/1284b800-6ff1-11e9-9091-46f74a171f45","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1195,"title":"Vimac","url":"https:\/\/vimacapp.com\/","content":"<main><section class=\"hero\"><div class=\"container hero-inner hero-copy\"><p class=\"hero-paragraph\">Vimac is a productivity application that revolutionizes the way you navigate macOS.<\/p><p><a class=\"button button-primary\" href=\"https:\/\/install.appcenter.ms\/users\/dexterleng\/apps\/vimac\/distribution_groups\/sparkle\">Download Vimac<\/a><\/p><\/div><\/section><section class=\"features section\"><div class=\"container features-inner section-inner has-bottom-divider\"><div class=\"features-header text-center container-sm\"><h2 class=\"section-title mt-0\">Click on buttons with hints.<\/h2><p class=\"section-paragraph\">Activate Hint-mode and type the hint-text to perform a click.<\/p><\/div><div class=\"features-header text-center container-sm\"><h2 class=\"section-title mt-0\">Trackpad-smooth keyboard scrolling.<\/h2><p class=\"section-paragraph\">Activate Scroll-mode and scroll with HJKL keys.<\/p><\/div><div class=\"features-header text-center container-sm\"><h2 class=\"section-title mt-0\">Open Source<\/h2><p class=\"section-paragraph\">Vimac is open source under the GNU General Public License v3.<\/p><p><a class=\"button button-primary\" href=\"https:\/\/github.com\/dexterleng\/vimac\/\">View on GitHub<\/a><\/p><\/div><div class=\"features-header text-center container-sm\"><h2 class=\"section-title mt-0\">Usage<\/h2><p class=\"section-paragraph\">Refer to the README on GitHub for updated instructions.<\/p><p><a class=\"button button-primary\" href=\"https:\/\/github.com\/dexterleng\/vimac\/\">View on GitHub<\/a><\/p><\/div><\/div><\/section><\/main>","created_at":"2020-08-30T18:51:13+0000","updated_at":"2020-08-30T18:51:13+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"vimacapp.com","http_status":"200","origin_url":"http:\/\/vimacapp.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1194,"title":"pelle\/clauth","url":"https:\/\/github.com\/pelle\/clauth","content":"<p><a href=\"http:\/\/travis-ci.org\/pelle\/clauth\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/62beabcc75c1c6e4f85b5263fb759d0b748dde8a\/68747470733a2f2f7365637572652e7472617669732d63692e6f72672f70656c6c652f636c617574682e706e67\" alt=\"Build Status\" data-canonical-src=\"https:\/\/secure.travis-ci.org\/pelle\/clauth.png\" class=\"c6\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>This is a simple OAuth 2 provider that is designed to be used as a primary authentication provider for a Clojure Ring app.<\/p><p>It currently handles OAuth2 bearer authentication and interactive authentication.<\/p><p>See <a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-bearer-08\" rel=\"nofollow\">draft-ietf-oauth-v2-bearer<\/a><\/p><p>The following bearer tokens are implemented:<\/p><ul><li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-bearer-08#section-2.1\" rel=\"nofollow\">Authorization header<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-bearer-08#section-2.2\" rel=\"nofollow\">Form encoded body parameter<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-bearer-08#section-2.3\" rel=\"nofollow\">URI query field<\/a><\/li>\n<li>Non standard http cookie ('access_token') for use in interactive applications<\/li>\n<li>Non standard session ('access_token') for use in interactive applications<\/li>\n<\/ul><h2><a id=\"user-content-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#install\"><\/a>Install<\/h2><p>Add the following dependency to your <code>project.clj<\/code> file:<\/p><div class=\"highlight highlight-source-clojure\"><pre>[clauth \"1.0.0-rc17\"]<\/pre><\/div><h3><a id=\"user-content-possible-breaking-change-with-100-rc14\" class=\"anchor\" aria-hidden=\"true\" href=\"#possible-breaking-change-with-100-rc14\"><\/a>Possible breaking change with 1.0.0-rc14<\/h3><p>The oauth models have been refactored to be simple maps instead of records. The factory functions are simpler and take a simple map of data.<\/p><p>If you have your own token and authcode creator functions for versions prior to 1.0.0-rc14 please verify that they work when upgrading.<\/p><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><p>There are currently 2 middlewares defined:<\/p><ul><li>wrap-bearer-token<\/li>\n<li>require-bearer-token!<\/li>\n<\/ul><p>Both of them take as a parameter a function which should return a object representing the token. This could be a user object, but could also be a token object with specific meta-data. I may standardize on something when more of the framework is developed.<\/p><p>The object returned by your function is set to :access-token entry in the request.<\/p><p>The difference between wrap-bearer-token and require-bearer-token! is that wrap will find a token but not require it. require-bearer-token will return a <a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-bearer-08#section-2.4\" rel=\"nofollow\">HTTP 401 header<\/a>.<\/p><h2><a id=\"user-content-grant-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#grant-types\"><\/a>Grant Types<\/h2><p>Currently the following Grant types are supported:<\/p><ul><li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25#section-4.1\" rel=\"nofollow\">Authorization Code Grant<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25#section-4.4\" rel=\"nofollow\">Client Credential Grant<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25#section-4.3\" rel=\"nofollow\">Resource Owner Password Credential Grant<\/a><\/li>\n<\/ul><p>Grant types are implemented using multimethods. To implement one<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defmethod token-request-handler \"my_grant_type\" [req authenticator] ...)<\/pre><\/div><h2><a id=\"user-content-authorization-request\" class=\"anchor\" aria-hidden=\"true\" href=\"#authorization-request\"><\/a>Authorization request<\/h2><p>We currently support the following authorization requests:<\/p><ul><li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25#section-4.1\" rel=\"nofollow\">Authorization Code Grant<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25#section-4.2\" rel=\"nofollow\">Implicit Grant<\/a><\/li>\n<\/ul><p>You can control the ones you wish to support by using the configuration parameter :allowed-response-types to the authorization-handler<\/p><div class=\"highlight highlight-source-clojure\"><pre>(authorization-handler {:allowed-response-types [\"code\"]}) ;; Only support Authorization Code Grants<\/pre><\/div><p>Implement custom authorization requests:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defmethod authorization-request-handler \"custom\"\n  ... )<\/pre><\/div><h2><a id=\"user-content-tokens\" class=\"anchor\" aria-hidden=\"true\" href=\"#tokens\"><\/a>Tokens<\/h2><p>There is a protocol defined called Expirable which implements one function:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(is-valid? token)<\/pre><\/div><p>This is implementend by IPersistentMap so {} represents a valid token where {:expires (date-time 2011)} is invalid.<\/p><p>A OAuthToken map can be instantiated and stored easily by the create-token function:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(create-token client user)<\/pre><\/div><h2><a id=\"user-content-client-applications\" class=\"anchor\" aria-hidden=\"true\" href=\"#client-applications\"><\/a>Client Applications<\/h2><p>A ClientApplication map can be instantiated and stored easily by the register-app function:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(register-app name url)<\/pre><\/div><p>A client application has a client-id and a client-secret which is used for issuing tokens.<\/p><h2><a id=\"user-content-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#users\"><\/a>Users<\/h2><p>A User map exists which can be instantiated and stored easily by the register-user function:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(register-user login password name url)<\/pre><\/div><h2><a id=\"user-content-stores\" class=\"anchor\" aria-hidden=\"true\" href=\"#stores\"><\/a>Stores<\/h2><p>Stores are used to store tokens and will be used to store clients and users as well.<\/p><p>There is a generalized protocol called Store and currently a simple memory implementation used for it.<\/p><p>It should be pretty simple to implement this Store with redis, sql, datomic or what have you.<\/p><p>It includes a simple Redis implementation based on Carmine.<\/p><p>The stores used by the various parts are defined in an atom for each type. reset! each of them with your own implementation.<\/p><p>The following stores are currently defined:<\/p><ul><li>token-store is in clauth.token\/token-store<\/li>\n<li>auth-code-store is in clauth.auth-code\/auth-code-store<\/li>\n<li>client-store is in clauth.client\/client-store<\/li>\n<li>user-store is in clauth.user\/user-store<\/li>\n<\/ul><p>To use the redis store add the following to your code:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(def server-conn\n  {:pool {}\n   :spec {:host \"127.0.0.1\"\n          :port 6379\n          :db 14}})\n(reset! token-store (create-redis-store \"tokens\" server-conn))\n(reset! auth-code-store (create-redis-store \"auth-codes\" server-conn))\n(reset! client-store (create-redis-store \"clients\" server-conn))\n(reset! user-store (create-redis-store \"users\" server-conn))<\/pre><\/div><h2><a id=\"user-content-authorization-oauth-tokens\" class=\"anchor\" aria-hidden=\"true\" href=\"#authorization-oauth-tokens\"><\/a>Authorization OAuth Tokens<\/h2><p>There is currently a single authorization-handler that handles authorization called authorization-handler. Install it in your routes by convention at \"\/authorize\" or \"\/oauth\/authorize\".<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defn routes [req]\n  (case (req :uri)\n    \"\/authorize\" ((authorization-handler) req)\n    ((require-bearer-token! handler) req)))<\/pre><\/div><p>Authorization handler comes with defaults that use the various built in token, user etc. stores. You can override these by passing in a configuration map containing functions.<\/p><div class=\"highlight highlight-source-clojure\"><pre>(authorization-handler {:authorization-form authorization-form-handler\n                        :client-lookup clauth.client\/fetch-client\n                        :token-lookup clauth.token\/fetch-token\n                        :token-creator clauth.token\/create-token\n                        :auth-code-creator clauth.auth-code\/create-auth-code})<\/pre><\/div><h2><a id=\"user-content-issuing-oauth-tokens\" class=\"anchor\" aria-hidden=\"true\" href=\"#issuing-oauth-tokens\"><\/a>Issuing OAuth Tokens<\/h2><p>There is currently a single token-handler that provides token issuance called token-handler. Install it in your routes by convention at \"\/token\" or \"\/oauth\/token\".<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defn routes [req]\n  (case (req :uri)\n    \"\/token\" ((token-handler) req)\n    ((require-bearer-token! handler) req)))<\/pre><\/div><p>Token handler comes with defaults that use the various built in token, user etc. stores. You can override these by passing in a configuration map containing functions.<\/p><div class=\"highlight highlight-source-clojure\"><pre>(token-handler {:client-authenticator clauth.client\/authenticate-client\n                :user-authenticator clauth.user\/authenticate-user\n                :token-creator clauth.token\/create-token\n                :auth-code-revoker clauth.auth-code\/revoke-auth-code!\n                :auth-code-lookup clauth.auth-code\/fetch-auth-code })<\/pre><\/div><h2><a id=\"user-content-using-as-primary-user-authentication-on-server\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-as-primary-user-authentication-on-server\"><\/a>Using as primary user authentication on server<\/h2><p>One of the ideas of this is using OAuth tokens together with traditional sessions based authentication providing the benefits of both. To do this we create a new token when a user logs in and adds it to the session.<\/p><p>Why is this a good idea?<\/p><ul><li>You will be able to view a list of other sessions going on for security purposes<\/li>\n<li>You will be able to remotely log of another session<\/li>\n<li>Your app deals with tokens only. So this is also ideal for an API with a javascript front end<\/li>\n<\/ul><p>To use this make sure to wrap the session middleware. We have a login handler endpoint that could be used like this:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defn routes [master-client]\n  (fn [req]\n  (case (req :uri)\n    \"\/login\" ((login-handler master-client) req)\n    ((require-bearer-token! handler) req))))<\/pre><\/div><p>The master-client is a client record representing your own application. A default login view is defined in clauth.views\/login-form-handler but you can add your own. This just needs to be a ring handler presenting a form with the parameters \"username\" and \"password\".<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defn routes [master-client]\n  (fn [req]\n  (case (req :uri)\n    \"\/login\" ((login-handler my-own-login-form-handler master-client) req)\n    ((require-bearer-token! handler) req))))<\/pre><\/div><h2><a id=\"user-content-run-demo-app\" class=\"anchor\" aria-hidden=\"true\" href=\"#run-demo-app\"><\/a>Run Demo App<\/h2><p>A mini server demo is available. It creates a client for you and prints out instructions on how to issue tokens with curl.<\/p><pre>lein run -m clauth.demo\n<\/pre><h2><a id=\"user-content-todo\" class=\"anchor\" aria-hidden=\"true\" href=\"#todo\"><\/a>TODO<\/h2><p>The goal is to implement the full <a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25\" rel=\"nofollow\">OAuth2 spec<\/a>. The only main feature missing is. I'll aim for that for 1.1 as most people currently don't use refresh tokens:<\/p><ul><li><a href=\"http:\/\/tools.ietf.org\/html\/draft-ietf-oauth-v2-25#section-1.5\" rel=\"nofollow\">Refresh Tokens<\/a><\/li>\n<\/ul><h2><a id=\"user-content-contribute\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribute\"><\/a>Contribute<\/h2><p>You will need to have a Redis database running in the background in order to have some of the tests pass, otherwise, you will get an error about the connection being refused.<\/p><p>If you have Homebrew on Mac OSX, you can get Redis by typing <code>brew install redis<\/code> in the command line. Once that's done, get the Redis database started in your Terminal window by typing the following:<\/p><pre>redis-server \/usr\/local\/etc\/redis.conf\n<\/pre><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright (C) 2012 Pelle Braendgaard <a href=\"http:\/\/stakeventures.com\" rel=\"nofollow\">http:\/\/stakeventures.com<\/a><\/p><p>Distributed under the Eclipse Public License, the same as Clojure.<\/p>","created_at":"2020-08-30T13:55:40+0000","updated_at":"2020-08-30T13:55:40+0000","published_by":["pelle"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/8680?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","auth"],"is_public":false,"id":1193,"title":"cemerick\/friend","url":"https:\/\/github.com\/cemerick\/friend","content":"<p>An extensible authentication and authorization library for <a href=\"http:\/\/clojure.org\" rel=\"nofollow\">Clojure<\/a>\/<a href=\"http:\/\/github.com\/ring-clojure\/ring\">Ring<\/a> web applications and services.<\/p><pre>Picking up his staff he stood before the rock and said in a clear voice:\nMellon!\nThe star shone out briefly and faded again. Then silently a great\ndoorway was outlined, though not a crack or joint had been visible\nbefore. Slowly it divided in the middle and swung outwards inch by inch,\nuntil both doors lay back against the wall. Through the opening a\nshadowy stair could be seen climbing steeply up; but beyond the lower\nsteps the darkness was deeper than the night. The Company stared in\nwonder.\n\"I was wrong after all,\" said Gandalf, \"and Gimli too. Merry, of all\npeople, was on the right track. The opening word was inscribed on the\narchway all the time! The translation should have been: Say 'Friend' and\nenter. I had only to speak the Elvish word for friend and the doors\nopened. Quite simple. Too simple for a learned lore master in these\nsuspicious days. Those were happier times. Now let us go!\"\n<\/pre><p>\u2014 J.R.R. Tolkien, <em>Lord of the Rings<\/em><\/p><h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2><p>Friend is intended to provide a foundation for addressing all of the authentication and authorization concerns associated with web apps:<\/p><ul><li>user agent authentication; Friend currently includes support for form and HTTP Basic authentication, and makes it easy to:\n<ul><li>implement and use other workflows (e.g oauth, OpenId connect)<\/li>\n<li>integrate app-specific user-credential checks<\/li>\n<\/ul><\/li>\n<li>role-based authorization\n<ul><li>optionally uses Clojure's ad-hoc hierarchies to model hierarchical roles<\/li>\n<\/ul><\/li>\n<li><code>su<\/code> capabilities (a.k.a. \"log in as\"), enabling users to maintain multiple simultaneous logins, as well as to allow administrators to take on users' identities for debugging or support purposes (<strong>in progress<\/strong>)<\/li>\n<li>and the creature comforts:\n<ul><li>Ring middlewares for configuring and defining the scopes of authentication and authorization<\/li>\n<li>Macros to clearly demarcate the scope of authentication and authorization within code that is \"below\" the level of Ring handlers where you can't use middlewares.<\/li>\n<li>A reasonable Clojure API around the jbcrypt library for hashing sensitive bits.<\/li>\n<li>Enables DRY routes and configuration, e.g. no need to configure your routes in Compojure or Noir or Moustache, and separately specify which routes fall under the jurisdiction of Friend's security machinery<\/li>\n<li>Purely functional in nature: authentications, roles, and session data are obtained, retained, and passed around as good ol' persistent data structures (just as Ring intended). No stateful session or context is ever bashed in place, making it easier to reason about what's going on.<\/li>\n<\/ul><\/li>\n<\/ul><h3><a id=\"user-content-why\" class=\"anchor\" aria-hidden=\"true\" href=\"#why\"><\/a>Why?<\/h3><p>Nothing like Friend exists, and it needs to. Securing Ring applications and services is (charitably speaking) a PITA right now, with everyone rolling their own, or starting with relatively low-level middlewares and frameworks. This will never do. Serious web applications need to take security seriously, and need to readily interoperate with all sorts of authentication mechanisms that have come to litter the web as well as internal networks.<\/p><p>Friend has been built with one eye on a number of frameworks.<\/p><h3><a id=\"user-content-status\" class=\"anchor\" aria-hidden=\"true\" href=\"#status\"><\/a>Status<\/h3><p>Very stable, widely-used in production AFAIK.<\/p><p>Note: while actively maintained, <a href=\"https:\/\/groups.google.com\/forum\/#!topic\/clojure-sec\/ceMhYPR0G60\" rel=\"nofollow\">Friend is in search of a new maintainer<\/a>.<\/p><h3><a id=\"user-content-changelog\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelog\"><\/a>Changelog<\/h3><p>Available <a href=\"http:\/\/github.com\/cemerick\/friend\/blob\/master\/CHANGES.md\">here<\/a>.<\/p><h3><a id=\"user-content-known-issues\" class=\"anchor\" aria-hidden=\"true\" href=\"#known-issues\"><\/a>Known issues<\/h3><ul><li>This README is <em>way<\/em> too long and not well-organized. It's more of a brain-dump than anything else at the moment.<\/li>\n<li>Configuration keys need a bit of tidying, especially for those that can\/should apply to multiple authorization workflows. Fixes for such things will break the existing API.<\/li>\n<li>the <code>su<\/code> mechanism is in-progress<\/li>\n<li>\u2026surely there's more. File issues.<\/li>\n<\/ul><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>\"Installation\"<\/h2><p>Friend is available in Clojars. Add this <code>:dependency<\/code> to your Leiningen <code>project.clj<\/code>:<\/p><div class=\"highlight highlight-source-clojure\"><pre>[com.cemerick\/friend \"0.2.3\"]<\/pre><\/div><p>Or, add this to your Maven project's <code>pom.xml<\/code>:<\/p><div class=\"highlight highlight-text-xml\"><pre>&lt;repository&gt;\n  &lt;id&gt;clojars&lt;\/id&gt;\n  &lt;url&gt;http:\/\/clojars.org\/repo&lt;\/url&gt;\n&lt;\/repository&gt;\n&lt;dependency&gt;\n  &lt;groupId&gt;com.cemerick&lt;\/groupId&gt;\n  &lt;artifactId&gt;friend&lt;\/artifactId&gt;\n  &lt;version&gt;0.2.3&lt;\/version&gt;\n&lt;\/dependency&gt;<\/pre><\/div><p>Friend is compatible with Clojure 1.2.0 - 1.5.0+.<\/p><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><p>How you use Friend will vary, sometimes significantly, depending on the authentication providers you use and the authorization policy\/ies you want to enforce. A generic example of typical usage of Friend is below, but the best way to become familiar with Friend and how it can be used would be to go check out<\/p><h3><a id=\"user-content-httpfriend-demoherokuappcom\" class=\"anchor\" aria-hidden=\"true\" href=\"#httpfriend-demoherokuappcom\"><\/a><a href=\"http:\/\/friend-demo.herokuapp.com\" rel=\"nofollow\"><em>http:\/\/friend-demo.herokuapp.com<\/em><\/a><\/h3><p>\u2026a collection of tiny demonstration apps using Friend. It should be easy to find the one(s) that apply to your situation, and go straight to its source so you can see how all the pieces fit together.<\/p><hr \/><p>Here's probably the most self-contained Friend usage possible:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(ns your.ring.app\n  (:require [cemerick.friend :as friend]\n            (cemerick.friend [workflows :as workflows]\n                             [credentials :as creds])))\n; a dummy in-memory user \"database\"\n(def users {\"root\" {:username \"root\"\n                    :password (creds\/hash-bcrypt \"admin_password\")\n                    :roles #{::admin}}\n            \"jane\" {:username \"jane\"\n                    :password (creds\/hash-bcrypt \"user_password\")\n                    :roles #{::user}}})\n(def ring-app ; ... assemble routes however you like ...\n  )\n(def secured-app\n  (-&gt; ring-app\n    (friend\/authenticate {:credential-fn (partial creds\/bcrypt-credential-fn users)\n                          :workflows [(workflows\/interactive-form)]})\n    ; ...required Ring middlewares ...\n    ))<\/pre><\/div><p>We have an unadorned (and unsecured) Ring application (<code>ring-app<\/code>, which can be any Ring handler), and then the usage of Friend's <code>authenticate<\/code> middleware. This is where all of the authentication work will be done, with the return value being a secured Ring application (<code>secured-app<\/code>), the requests to which are subject to the configuration provided to <code>authenticate<\/code> and the authorization contexts that are defined within <code>ring-app<\/code> (which we'll get to shortly).<\/p><p>(If you're newer to Clojure, you might not recognize the tokens prefixed with two colons [e.g. <code>::admin<\/code>]. These are auto-namespaced keywords; in the example above, <code>::admin<\/code> expands to <code>:your.ring.app\/admin<\/code>.)<\/p><h3><a id=\"user-content-authentication\" class=\"anchor\" aria-hidden=\"true\" href=\"#authentication\"><\/a>Authentication<\/h3><p>There are two key abstractions employed during authentication: <a href=\"#workflows\">workflow<\/a> and <a href=\"#credential-functions-and-authentication-maps\">credential<\/a> functions. The example above defines a single workflow \u2014 one supporting the <code>POST<\/code>ing of <code>:username<\/code> and <code>:password<\/code> parameters to (by default) <code>\/login<\/code> \u2014\u00a0which will discover the specified <code>:credential-fn<\/code> and use it to validate submitted credentials. The <code>bcrypt-credential-fn<\/code> function verifies a submitted map of <code>{:username \"...\" :password \"...\"}<\/code> credentials against one loaded from another function based on the <code>:username<\/code> value; in this case, we're just looking up the username in a fixed Clojure map that has username, (bcrypted) password, and roles entries. If a submitted set of credentials matches those in the authoritative store, the latter are returned (<em>sans<\/em> <code>:password<\/code>) as an <em>authentication map<\/em>.<\/p><p>(Each workflow can have its own local configuration \u2014 including a credential function \u2014 that is used in preference to the configuration specified at the <code>authenticate<\/code> level.)<\/p><p>The <code>authenticate<\/code> middleware runs every incoming request through each of the workflows with which it is created. It further handles things like retaining authentication details in the user session (by default) and managing the redirection of users when they attempt to access protected resources without the requisite authentication or authorization (first to the start of an authentication workflow, e.g. <code>GET<\/code> of a <code>\/login<\/code> URI, and then back to the originally-requested protected resource once the authentication workflow is completed).<\/p><p>(Note that Friend itself requires some core Ring middlewares: <code>params<\/code>, <code>keyword-params<\/code> and <code>nested-params<\/code>. Most workflows will additionally require <code>session<\/code> in order to support post-authentication redirection to previously-unauthorized resources, retention of tokens and nonces for workflows like OpenId and oAuth, etc. HTTP Basic is the only provided workflow that does not require <code>session<\/code> middleware.)<\/p><h3><a id=\"user-content-workflows\" class=\"anchor\" aria-hidden=\"true\" href=\"#workflows\"><\/a>Workflows<\/h3><p>Individual authentication methods (e.g., form-based auth, HTTP Basic, OpenID, oAuth, etc.) are implemented as <em>workflows<\/em> in Friend. A workflow is a regular Ring handler function, except that a workflow function can <em>opt<\/em> to return an <em>authentication map<\/em> instead of a Ring response if a request is authenticated. A diagram may help:<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cemerick\/friend\/raw\/master\/docs\/workflow.png\"><img src=\"https:\/\/github.com\/cemerick\/friend\/raw\/master\/docs\/workflow.png\" alt=\"\" class=\"c7\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>You can define any number of workflows in a <code>:workflows<\/code> kwarg to <code>authenticate<\/code>. Incoming requests are always run through the configured workflows prior to potentially being passed along to the secured Ring application.<\/p><p>If a workflow returns an authentication map, then the <code>authenticate<\/code> middleware will either:<\/p><ul><li>carry on processing the request if the workflow allows for credentials to be provided in requests to any resource (i.e. HTTP Basic); control of this is entirely up to each workflow, and will be described later.<\/li>\n<li>redirect the user agent to a secured resource that it was previously barred from accessing via Friend's authorization machinery<\/li>\n<\/ul><p>If a workflow returns a Ring response, then that response is sent back to the user agent straight away (after some bookkeeping by the <code>authenticate<\/code> middleware to preserve session states and such). This makes it possible for a workflow to control a \"local\" dataflow between itself, the user agent, and any necessary external authorities (e.g. by redirecting a user agent to an OpenId endpoint, performing token exchange in the case of oAuth, etc., eventually returning a complete authentication map that will allow the user agent to proceed on its desired vector).<\/p><h3><a id=\"user-content-credential-functions-and-authentication-maps\" class=\"anchor\" aria-hidden=\"true\" href=\"#credential-functions-and-authentication-maps\"><\/a>Credential functions and authentication maps<\/h3><p>Workflows use a <em>credential function<\/em> to verify the credentials provided to them via requests. Credential functions can be specified either as a <code>:credential-fn<\/code> option to <code>cemerick.friend\/authenticate<\/code>, or often as an (overriding) <code>:credential-fn<\/code> option to individual workflow functions.<\/p><p>All credential functions take a single argument, a map containing the available credentials that additionally contains a <code>:cemerick.friend\/workflow<\/code> slot identifying which workflow has produced the credential. For example, the default form-based authentication credential map looks like this:<\/p><div class=\"highlight highlight-source-clojure\"><pre>{:username \"...\", :password \"...\", :cemerick.friend\/workflow :form}<\/pre><\/div><p>HTTP Basic credentials are much the same, but with a workflow value of <code>:http-basic<\/code>, etc. Different workflows may have significantly different credential maps (e.g. an OpenID workflow does not provide username and password, but rather a token returned by an OpenID provider along with potentially some number of \"attributes\" like the user's name, email address, default language, etc.), and unique credential verification requirements (again, contrast the simple username\/password verification of form or HTTP Basic credentials and OpenId, which, in general, when presented with unknown credentials, should <em>register<\/em> the indicated identity rather than verifying it).<\/p><p>In summary, the contract of what exactly must be in the map provided to credential functions is entirely at the discretion of each workflow function, as is the semantics of the credential function.<\/p><p>If a map of credentials is verified by a credential function, it should return a <em>authentication map<\/em> that aggregates all authentication and authorization information available for the identified user. This map may contain many entries, depending upon the authentication information that is relevant for the workflow in question and the user data relevant to the application, but two entries are privileged:<\/p><ul><li><code>:identity<\/code> (<strong>required<\/strong>) corresponds with e.g. the username in a form or HTTP Basic authentication, an oAuth token, etc.; this value <em>must<\/em> be unique across all users within the application<\/li>\n<li><code>:roles<\/code>, an optional collection of values enumerating the roles for which the user is authorized, or a function returning the same.<\/li>\n<\/ul><p><em>If a map of credentials is found to be invalid, the credential function must return nil.<\/em><\/p><h3><a id=\"user-content-authorization\" class=\"anchor\" aria-hidden=\"true\" href=\"#authorization\"><\/a>Authorization<\/h3><p>As is, the example above doesn't do a lot: users can opt to be authenticated, but we've not described any kind of security policy, identified routes or functions or forms that require particular roles to access, and so on. This is where authorization mechanisms come into play.<\/p><p>While Friend has a single point of authentication \u2014 the <code>authenticate<\/code> middleware \u2014\u00a0it has many different options for restricting access to particular resources or code:<\/p><ul><li><code>authenticated<\/code> is a macro that requires that the current user must be authenticated<\/li>\n<li><code>authorized?<\/code> is a predicate that returns true only if the current user (as determined via the <em>authentication map<\/em> returned by a workflow) possesses the specified <em>roles<\/em>. You'll usually want to use one of the higher-level facilities (keep reading), but <code>authorized?<\/code> may come in handy if access to a certain resource or operation cannot be specified declaratively.<\/li>\n<\/ul><p>The rest of the authorization utilities use <code>authorized?<\/code> to determine whether a user may gain access to whatever the utility is protecting:<\/p><ul><li><code>authorize<\/code> is a macro that guards any body of code from being executed within a thread associated with a user that is not <code>authorized?<\/code><\/li>\n<li><code>wrap-authorize<\/code> is a Ring middleware that only allows requests to pass through to the wrapped handler if their associated user is <code>authorized?<\/code><\/li>\n<li><code>authorize-hook<\/code> is a function intended to be used with the <a href=\"https:\/\/github.com\/technomancy\/robert-hooke\/\">Robert Hooke<\/a> library that allows you to place authorization guards around functions defined in code you don't control.<\/li>\n<\/ul><p>Here's an extension of the example above that adds some actual routes (using Compojure) and handler that require authentication:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(use '[compojure.core :as compojure :only (GET ANY defroutes)])\n(defroutes user-routes\n  (GET \"\/account\" request (page-bodies (:uri request)))\n  (GET \"\/private-page\" request (page-bodies (:uri request))))\n(defroutes ring-app\n  ;; requires user role\n  (compojure\/context \"\/user\" request\n    (friend\/wrap-authorize user-routes #{::user}))\n  ;; requires admin role\n  (GET \"\/admin\" request (friend\/authorize #{::admin}\n                          #_any-code-requiring-admin-authorization\n                          \"Admin page.\"))\n  ;; anonymous\n  (GET \"\/\" request \"Landing page.\")\n  (GET \"\/login\" request \"Login page.\")\n  (friend\/logout (ANY \"\/logout\" request (ring.util.response\/redirect \"\/\"))))<\/pre><\/div><p>This should be easy to grok, but some highlights:<\/p><ul><li>Authorization checks generally should happen <em>after<\/em> routing. This is usually easily accomplished by segregating handlers as you might do so anyway, and then using something like Compojure's <code>context<\/code> utility to wire them up into a common URI segment.<\/li>\n<li>Alternatively, you can use <code>authorize<\/code> to put authorization guards around any code, anywhere.<\/li>\n<li>The <code>logout<\/code> middleware can be applied to any Ring handler, and will remove all authentication information from the session assuming a non-<code>nil<\/code> response from the wrapped handler.<\/li>\n<\/ul><p>Note that, so far, all of the authorization checks will be completely \"strict\", e.g. the admin user won't have access to <code>\/user<\/code> because it requires the <code>::user<\/code> role. This is where hierarchies are unreasonably helpful.<\/p><h4><a id=\"user-content-hierarchical-roles-ht-derive-isa-et-al\" class=\"anchor\" aria-hidden=\"true\" href=\"#hierarchical-roles-ht-derive-isa-et-al\"><\/a>Hierarchical roles (\/ht <code>derive<\/code>, <code>isa?<\/code>, et al.)<\/h4><p>The foundational <code>authorized?<\/code> predicate uses <code>isa?<\/code> to check if any of the current user's roles match one of those specified. This means that you can take advantage of Clojure's hierarchies via <code>derive<\/code> to establish relationships between roles. e.g., this is all that is required to give a user with the <code>::admin<\/code> role all of the privileges of a user with the <code>::user<\/code> role:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(derive ::admin ::user)<\/pre><\/div><p>Of course, you are free to construct your role hierarchy(ies) however you like, to suit your application and your security requirements.<\/p><h3><a id=\"user-content-nginx-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#nginx-configuration\"><\/a>Nginx configuration<\/h3><p>If you are using Nginx to, e.g, terminate SSL, set the appropriate headers so that the Clojure backend can generate the correct <code>return-to<\/code> URLs for the openid and similar workflows:<\/p><div class=\"highlight highlight-source-nginx\"><pre>upstream jetty_upstream {\n  ip_hash;\n  server 127.0.0.1:8080;\n  keepalive 64;\n}\nserver {\n  listen 443 ssl;\n  #...SSL termination config, &amp;c.\n  location \/ {\n    proxy_set_header host              $host;\n    proxy_set_header x-forwarded-for   $remote_addr;\n    proxy_set_header x-forwarded-host  $host;\n    proxy_set_header x-forwarded-proto $scheme;\n    proxy_set_header x-forwarded-port  $server_port;\n    proxy_pass http:\/\/jetty_upstream;\n  }\n}<\/pre><\/div><h2><a id=\"user-content-todo\" class=\"anchor\" aria-hidden=\"true\" href=\"#todo\"><\/a>TODO<\/h2><ul><li>run-as\/sudo\/multi-user login<\/li>\n<li>alternative hashing methods and salting strategies\n<ul><li>good to encourage bcrypt, but existing apps have tons of sha-X, md5, etc passwords<\/li>\n<\/ul><\/li>\n<li>remember-me?<\/li>\n<li>fine-grained authorization (viz. ACLs, etc)\n<ul><li>maybe something compelling can fall out of existing treatment of roles?<\/li>\n<\/ul><\/li>\n<li>interop\n<ul><li>recognize \/ provide access to servlet principal<\/li>\n<li>spring-security<\/li>\n<\/ul><\/li>\n<li>make <code>:cemerick.friend\/workflow<\/code> metadata<\/li>\n<li>documentation\n<ul><li>authentication retention<\/li>\n<li>authentication map metadata:\n<ul><li><code>:type<\/code><\/li>\n<li><code>::friend\/workflow<\/code><\/li>\n<li><code>::friend\/redirect-on-auth?<\/code><\/li>\n<li><code>::friend\/ensure-session<\/code><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-need-help\" class=\"anchor\" aria-hidden=\"true\" href=\"#need-help\"><\/a>Need Help?<\/h2><p>Ping <code>cemerick<\/code> on freenode irc or <a href=\"http:\/\/twitter.com\/cemerick\" rel=\"nofollow\">twitter<\/a> if you have questions or would like to contribute patches.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright \u00a92012-2013 <a href=\"http:\/\/cemerick.com\" rel=\"nofollow\">Chas Emerick<\/a> and other contributors.<\/p><p>Distributed under the Eclipse Public License, the same as Clojure. Please see the <code>epl-v10.html<\/code> file at the top level of this repo.<\/p>","created_at":"2020-08-30T13:55:35+0000","updated_at":"2020-08-30T13:55:37+0000","published_by":["cemerick"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/47489?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","auth"],"is_public":false,"id":1192,"title":"auth - Auth\/Authz facilities for ring based apps.","url":"https:\/\/funcool.github.io\/buddy-auth\/latest\/","content":"<div class=\"paragraph\"><p><strong>What is the difference with Friend?<\/strong><\/p><\/div><div class=\"paragraph\"><p><em>buddy-auth<\/em> authorization\/authentication facilities are more low level and less opinionated than friend, and allow you to easily build other high level abstractions over them. Technically, friend abstraction can be built on top of <em>buddy-auth<\/em>.<\/p><\/div><div class=\"paragraph\"><p><strong>How can I use <em>buddy<\/em> with <a href=\"http:\/\/clojure-liberator.github.io\/liberator\/\">liberator<\/a>?<\/strong><\/p><\/div><div class=\"paragraph\"><p>By design, <em>buddy<\/em> has authorization and authentication well separated. This helps a lot if you want use only one part of it (ex: authentication only) without including the other.<\/p><\/div><div class=\"paragraph\"><p>In summary: yes, you can use <em>buddy-auth<\/em> with liberator.<\/p><\/div><div class=\"paragraph\"><p><strong>Can I use <em>buddy-auth<\/em> with pedestal?<\/strong><\/p><\/div><div class=\"paragraph\"><p>Although is not mentioned in this documentation, you can use <em>buddy-auth<\/em> with pedestal without any problems.<\/p><\/div><div class=\"paragraph\"><p><a href=\"https:\/\/juxt.pro\/blog\/posts\/securing-your-clojurescript-app.html\" class=\"bare\">https:\/\/juxt.pro\/blog\/posts\/securing-your-clojurescript-app.html<\/a><\/p><\/div><div class=\"paragraph\"><p><strong>Can I use <em>buddy-auth<\/em> with catacumba?<\/strong><\/p><\/div><div class=\"paragraph\"><p>Not directly.<\/p><\/div><div class=\"paragraph\"><p>The design of <em>buddy-auth<\/em> api is intrinsically blocking just because ring and ring based abstractions are also blocking. However <em>catacumba<\/em> is asyncronous toolkit and it comes with its own, builtint variant of <em>buddy-auth<\/em> designed for asynchronous workflow (reusing the underlying <em>buddy-sign<\/em>, <em>buddy-core<\/em> and <em>buddy-hashers<\/em> modules).<\/p><\/div>","created_at":"2020-08-30T13:55:26+0000","updated_at":"2020-08-30T13:55:30+0000","published_by":["Andrey Antukh, <niwi@niwi.nz>"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"funcool.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["cocktails"],"is_public":false,"id":1191,"title":"oldfashioned.tech","url":"https:\/\/oldfashioned.tech\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-08-24T18:46:20+0000","updated_at":"2020-08-24T18:46:24+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"oldfashioned.tech"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1190,"title":"degoogle","url":"https:\/\/degoogle.jmoore.dev\/","content":"<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/ublock-origin\">uBlock Origin<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/ublock-origin?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nuBlock Origin is not an \u201cad blocker\u201d, it\u2019s a wide-spectrum content blocker with CPU and memory efficiency as a primary feature.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/privacy-badger17\">Privacy Badger<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/privacy-badger17?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nPrivacy Badger automatically learns to block invisible trackers. Instead of keeping lists of what to block, Privacy Badger learns by watching which domains appear to be tracking you as you browse the Web.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/decentraleyes\">Decentraleyes<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/decentraleyes?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nProtects you against tracking through \u201cfree\u201d, centralized, content delivery. It prevents a lot of requests from reaching networks like Google Hosted Libraries, and serves local files to keep sites from breaking. Complements regular content blockers.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/https-everywhere\">HTTPS Everywhere<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/https-everywhere?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nMany sites on the web offer some limited support for encryption over HTTPS, but make it difficult to use. For instance, they may default to unencrypted HTTP, or fill encrypted pages with links that go back to the unencrypted site. The HTTPS Everywhere extension fixes these problems by rewriting all requests to these sites to HTTPS.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/cookie-autodelete\">Cookie AutoDelete<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/cookie-autodelete?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nControl your cookies! This WebExtension is inspired by Self Destructing Cookies. When a tab closes, any cookies not being used are automatically deleted. Whitelist the ones you trust while deleting the rest. Support for Container Tabs.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/duckduckgo-for-firefox\">DDG Privacy Essentials<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/duckduckgo-for-firefox?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nPrivacy, simplified. Our add-on provides the privacy essentials you need to seamlessly take control of your personal information, no matter where the internet takes you: tracker blocking, smarter encryption, DuckDuckGo private search, and more.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/i-dont-care-about-cookies\">I don\u2019t care about cookies<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/i-dont-care-about-cookies?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nThis add-on will remove cookie warnings from almost all websites! You can report any website which still warns you about cookies: make a right click and choose \u2018Report a cookie warning\u2019 from the menu.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/minerblock-origin\">minerBlock<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/minerblock-origin?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nMinerBlock is an efficient browser extension that focuses on blocking browser-based cryptocurrency miners all over the web. This extension uses two different approaches to block miners. The first one is based on blocking requests\/scripts loaded from a blacklist, this is the traditional approach adopted by most ad-blockers and other mining blockers. The other approach which makes MinerBlock more efficient against cryptojacking is detecting potential mining behavior inside loaded scripts and kills them immediately. This makes the extension able to block inline scripts as well as miners running through proxies. Source code available on <a href=\"https:\/\/github.com\/xd4rker\/MinerBlock\">GitHub<\/a>.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/terms-of-service-didnt-read\">Terms of Service; Didn\u2019t Read<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/terms-of-service-didnt-read?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\n\u201cI have read and agree to the Terms\u201d is the biggest lie on the web. We aim to fix that. \u201cTerms of Service; Didn\u2019t Read\u201d is a user rights initiative to rate and label website terms &amp; privacy policies, from very good (class A) to very bad (class E). This extension informs you instantly of your rights online by showing an unintrusive icon in the toolbar. You can click on this icon to get summaries from the <a href=\"https:\/\/tosdr.org\/\">Terms of Service; Didn\u2019t Read<\/a> initiative.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/clearurls\">ClearURLS<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/clearurls?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nThis extension will automatically remove tracking elements from URLs to help protect your privacy when browse through the Internet.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/cloud-firewall\">Cloud Firewall<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/cloud-firewall?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nBlock connections to pages and web resources hosted in major cloud services if the user wishes to do so. Supports blocking Google, Amazon, Facebook, Apple, Microsoft and Cloudflare.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/css-exfil-protection\">CSS Exfil Protection<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/css-exfil-protection?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nCSS Exfil is a method attackers can use to steal data from web pages using Cascading Style Sheets (CSS). This plugin sanitizes and blocks any CSS rules which may be designed to steal data.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/disconnect\">Disconnect<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/disconnect?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nVisualize and block the otherwise invisible websites that track your search and browsing history.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/dont-touch-my-tabs\">Don\u2019t touch my tabs! (rel=noopener<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/dont-touch-my-tabs?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nPrevent tabs opened by a hyperlink from hijacking the previous tab by adding the rel=noopener attribute to all hyperlinks (excluding same-domain hyperlinks).\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/multi-account-containers\">Firefox Multi-Account Containers<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/multi-account-containers?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nFirefox Multi-Account Containers lets you keep parts of your online life separated into color-coded tabs that preserve your privacy. Cookies are separated by container, allowing you to use the web with multiple identities or accounts simultaneously.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/google-search-link-fix\">Google search link fix (also Yandex)<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/google-search-link-fix?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nThis extension prevents Google and Yandex search pages from modifying search result links when you click them. This is useful when copying links but it also helps privacy by preventing the search engines from recording your clicks.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/httpz\">HTTPZ<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/httpz?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nHTTPZ is meant to be unobtrusive and lightweight, it respects your privacy, and is free of trans fats. Additionally, it is very configurable, and should be slightly more secure than some of the alternatives out there, since it has a couple of built-in defenses against SSL-stripping attacks.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/link-cleaner\">Link Cleaner<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/link-cleaner?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nClean URLs that are about to be visited: removes utm_* parameters; on item pages of aliexpress and amazon, removes tracking parameters; skip redirect pages of facebook, steam and reddit\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/neat-url\">Neat URL<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/neat-url?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nRemove garbage from URLs.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/privacy-oriented-origin-policy\">Privacy-Oriented Origin Policy<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/privacy-oriented-origin-policy?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nPrevent Firefox from sending Origin headers when they are least likely to be necessary, to protect your privacy.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/skip-redirect\">Skip Redirect<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/skip-redirect?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nSome web pages use intermediary pages before redirecting to a final page. This add-on tries to extract the final url from the intermediary url and goes there straight away if successful.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/temporary-containers\">Temporary Containers<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/temporary-containers?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nOpen tabs, websites, and links in automatically managed disposable containers. Containers isolate data websites store (cookies, storage, and more) from each other, enhancing your privacy and security while you browse.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/umatrix\">uMatrix<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/umatrix?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nPoint &amp; click to forbid\/allow any class of requests made by your browser. Use it to block scripts, iframes, ads, facebook, etc.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/trackmenot\">TrackMeNot<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/trackmenot?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nAn artware browser add-on to protect privacy in web-search. By issuing randomized queries to common search-engines, TrackMeNot obfuscates your search profile and registers your discontent with surreptitious tracking.\n<a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/amp2html\/?src=external-addonsbadge-daniel.priv.no\">Redirect AMP to HTML<\/a> <img src=\"https:\/\/img.shields.io\/amo\/users\/?src=external-addonsbadge-daniel.priv.no?style=flat-square\" alt=\"Mozilla Add-on\" referrerpolicy=\"no-referrer\" \/>\nThis web extension enables users to choose to opt-out of using Accelerated Mobile Pages (AMP) and choose to use the standard web instead.","created_at":"2020-08-22T21:43:44+0000","updated_at":"2020-08-22T21:43:44+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"degoogle.jmoore.dev","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1189,"title":"Ask HN: How to Become a Music Buff?","url":"https:\/\/news.ycombinator.com\/item?id=24209025","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pvaldes\" class=\"hnuser\">pvaldes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24217916\">4 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=wheels\" class=\"hnuser\">wheels<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210274\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Barrin92\" class=\"hnuser\">Barrin92<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24213010\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sammorrowdrums\" class=\"hnuser\">sammorrowdrums<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211488\">7 hours ago<\/a><\/p>\n<br \/><div class=\"comment\">This is a solid answer, but I'd also add - there are very few general music buffs.<p>Maybe I'd count, with years in pro music performance, degree in Jazz and a lifetime of listing to and performing an insanely wide variety of genres... but not a buff for all styles.<\/p><p>Computer systems are actually not a bad analogy. How many people do you know that have an in-depth knowledge of Assembly in different systems (including the oldest ones), GPUs, ML, Python, JS, Haskell, C++, Rust, ALGOL, Compilers, Hardware development, etc. Then add into the mix having an in-depth knowledge of the creators of those things, what else they did, failures, influences, their mentors.<\/p><p>You always have the choice of breadth versus depth, within your time constraints. I think learning about a single artist or band, looking up the producer, label, members and following up on all those leads to see what else they have done - and see if you like it is a great beginning.<\/p><p>Some genres like jazz might seem difficult to get into, and that's fine. No need to delve in until you find a hook that peaks your interest. And then when you find out about a genre (and jazz is an excellent example) - you'll find sub-genres you like, and don't and maybe you are interested in the evolution - like many people love bebop, but don't listen to \"dixieland jazz\" or big band swing, or modern fusion.<\/p><p>You will never hear all the music. But depth (in my opinion) breads a lot more passion than breadth just for the sake of it. So rather than listening to all the top 40 singles of the last decades, you'd probably be better off listening to some whole albums, and exploring from there.<\/p><p>Spotify (and similar) make it very easy for breadth, and can help discover new music you like - but you'll need to put work into delving in deep yourself.<\/p><p>As you do have a passion for film I'd suggest this.<\/p><p>For interesting modern music Quentin Tarantino files always deliver.<\/p><p>Danny Elfman has some interesting stuff (like this Simpsons theme which is a great example of the Lydian mode...)<\/p><p>John Williams &amp; Hans Zimmer have made loads of epic soundtracks, and if you like them - look up Holst, Tchaikovsky and Wagner<\/p><p>If you want a weirder launching off point, watch koyaanisqatsi and look up Philip Glass who made the soundtrack - that's an interesting avenue, albeit not very mainstream.<\/p><p>Another epic film score is 28 Days Later, John Murphey - parts of which have been used in loads of other huge features like Kick Ass...<\/p><p>You could also go a different route, and look up how the pre-classical Pachelbel's Canon contains the chord progression that is arguably the most popular of all time and watch this mash up of songs that use it (or the first part) <a href=\"https:\/\/www.youtube.com\/watch?v=oOlDewpCfZQ\" rel=\"nofollow\">https:\/\/www.youtube.com\/watch?v=oOlDewpCfZQ<\/a><\/p><p>Hahahaha, I could go on forever! I hope you find something you love, and actually want to go deeper into.<\/p><\/div>\n<\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=NikolaNovak\" class=\"hnuser\">NikolaNovak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210585\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=joveian\" class=\"hnuser\">joveian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24216793\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sjwright\" class=\"hnuser\">sjwright<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210998\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=microtherion\" class=\"hnuser\">microtherion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210102\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bredren\" class=\"hnuser\">bredren<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210197\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ojosilva\" class=\"hnuser\">ojosilva<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210703\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jmfldn\" class=\"hnuser\">jmfldn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24214453\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ajmadesc\" class=\"hnuser\">ajmadesc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210177\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ben7799\" class=\"hnuser\">ben7799<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210624\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sailfast\" class=\"hnuser\">sailfast<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210695\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=arrakeen\" class=\"hnuser\">arrakeen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211201\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloeki\" class=\"hnuser\">lloeki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211100\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rudyfink\" class=\"hnuser\">rudyfink<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211531\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cognaitiv\" class=\"hnuser\">cognaitiv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212239\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MogwaiAllOnYou\" class=\"hnuser\">MogwaiAllOnYou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210218\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=amanaplanacanal\" class=\"hnuser\">amanaplanacanal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211682\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=luigibosco\" class=\"hnuser\">luigibosco<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210002\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=werber\" class=\"hnuser\">werber<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209911\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=farleykr\" class=\"hnuser\">farleykr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209932\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mattbee\" class=\"hnuser\">mattbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209967\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=madhadron\" class=\"hnuser\">madhadron<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211777\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=iancmceachern\" class=\"hnuser\">iancmceachern<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24217775\">21 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kamonrye\" class=\"hnuser\">kamonrye<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212334\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=abap_rocky\" class=\"hnuser\">abap_rocky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212710\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=boomlinde\" class=\"hnuser\">boomlinde<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24217276\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=elwell\" class=\"hnuser\">elwell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211873\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bloat\" class=\"hnuser\">bloat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210836\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=autarch\" class=\"hnuser\">autarch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210892\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anm89\" class=\"hnuser\">anm89<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24216414\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=riffnote\" class=\"hnuser\">riffnote<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210304\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mdre\" class=\"hnuser\">mdre<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211721\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=grawprog\" class=\"hnuser\">grawprog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211028\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nikk1\" class=\"hnuser\">nikk1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211809\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Pandabob\" class=\"hnuser\">Pandabob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210772\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jemsouse\" class=\"hnuser\">jemsouse<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210925\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=52-6F-62\" class=\"hnuser\">52-6F-62<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212245\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=3131s\" class=\"hnuser\">3131s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211268\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=justsomeaccount\" class=\"hnuser\">justsomeaccount<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211022\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=plexiglas\" class=\"hnuser\">plexiglas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209863\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mike_mg\" class=\"hnuser\">mike_mg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210147\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=adamors\" class=\"hnuser\">adamors<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211289\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=banach\" class=\"hnuser\">banach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24215139\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gorgoiler\" class=\"hnuser\">gorgoiler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212406\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericzawo\" class=\"hnuser\">ericzawo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210642\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrisandchips\" class=\"hnuser\">chrisandchips<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211217\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thrwn_frthr_awy\" class=\"hnuser\">thrwn_frthr_awy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210750\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=asciimov\" class=\"hnuser\">asciimov<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24213266\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dcolkitt\" class=\"hnuser\">dcolkitt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210842\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Gatsky\" class=\"hnuser\">Gatsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209943\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=notnoided\" class=\"hnuser\">notnoided<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209861\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anarbadalov\" class=\"hnuser\">anarbadalov<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211704\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tkgally\" class=\"hnuser\">tkgally<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210028\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=HeadHonchoSP\" class=\"hnuser\">HeadHonchoSP<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210280\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=screye\" class=\"hnuser\">screye<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211102\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mushufasa\" class=\"hnuser\">mushufasa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210806\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=uniqueid\" class=\"hnuser\">uniqueid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212512\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=deltron3030\" class=\"hnuser\">deltron3030<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211143\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=52-6F-62\" class=\"hnuser\">52-6F-62<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212261\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=minimuffins\" class=\"hnuser\">minimuffins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211380\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=iak8god\" class=\"hnuser\">iak8god<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210700\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=leoedin\" class=\"hnuser\">leoedin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209994\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bumelant\" class=\"hnuser\">bumelant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210498\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=savanaly\" class=\"hnuser\">savanaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211416\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=paradox242\" class=\"hnuser\">paradox242<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210762\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=neap24\" class=\"hnuser\">neap24<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210390\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Yhippa\" class=\"hnuser\">Yhippa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211264\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hatboxreappoint\" class=\"hnuser\">hatboxreappoint<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210327\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=reinhardt1053\" class=\"hnuser\">reinhardt1053<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210379\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=heynk\" class=\"hnuser\">heynk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210405\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=branon\" class=\"hnuser\">branon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209743\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=grenoire\" class=\"hnuser\">grenoire<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209875\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=spanhandler\" class=\"hnuser\">spanhandler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210818\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=grenoire\" class=\"hnuser\">grenoire<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211136\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=joeberon\" class=\"hnuser\">joeberon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209985\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dagurp\" class=\"hnuser\">dagurp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210770\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=subpixel\" class=\"hnuser\">subpixel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210380\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thebigspacefuck\" class=\"hnuser\">thebigspacefuck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210517\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ethbro\" class=\"hnuser\">ethbro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210637\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bryanrasmussen\" class=\"hnuser\">bryanrasmussen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209870\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=phlipski\" class=\"hnuser\">phlipski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210588\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=aklemm\" class=\"hnuser\">aklemm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210047\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=AdmiralAsshat\" class=\"hnuser\">AdmiralAsshat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209890\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=regulation_d\" class=\"hnuser\">regulation_d<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209750\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=alistairSH\" class=\"hnuser\">alistairSH<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210600\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=at_a_remove\" class=\"hnuser\">at_a_remove<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210866\">8 hours ago<\/a><\/p>\n<br \/><div class=\"comment\">I don't know if I am a music buff -- what are the qualifications? I have a few thousand CDs, some of them fairly rare, and I am the person my friends turn to when they need to track down a song or rare release. By those standards ...<p>1) Find your passion. I generally despise this as advise for careers and the like but when it comes to matters of personal preference, <em>de gustibus non est disputandum<\/em>. What do you like?<\/p><p>2) Hit Discogs for your bands. Don't just look at main albums, look at appearances and rarities, songs covered by the band, appearances on soundtracks (which can often be one-offs) and the like. For certain bands, I literally own everything they performed in because I like all of it. This is <em>depth<\/em>. After that, follow each member of a band to hear if they did anything before or after you enjoy. Even try session musicians. You may find a talented collaborator that way, too. This is <em>breadth<\/em>.<\/p><p>3) Find a radio program adjacent to your tastes and examine the weekly playlist. What did you like? Here you are riding in the slipstream of a great big truck, as the DJ is doing a lot of the work for you.<\/p><p>4) Ignore most recommendation algorithms but Pandora, which are based on what other people play rather than what is actually related.<\/p><p>5) Don't be afraid to reach out. I have written musicians letters (snail mail, self-addressed stamped envelope enclosed) and said, \"I really like X, Y, and Z\" or \"Your performance on this track moved me\" then asked what inspired them, influenced them, or if they have influenced others. I usually hear back.<\/p><p>6) Attend concerts and listen to the opening acts, buy their (typically cheap) CDs. The openers appreciate it and you can find some great somewhat related music that way.<\/p><p>7) Back in the days when record stores existed, I would hit the \"M\" section or whatever letter, any band that didn't have a tab of their own, and just flip through it, looking for anything to catch my eye, then listen that way. Many duds, but some real wins.<\/p><p>8) Now that you have a sense of what you enjoy, you probably have some names associated with them, various microgenres. List 'em. For example, \"grave wave.\"<\/p><p>9) Look up those microgenres on Bandcamp and the like, listen through that way.<\/p><p>10) Hit \"Every Noise at Once\" and surf your microgenre that way, too. Bop around between band and microgenre names.<\/p><p>11) Iterate.<\/p><\/div>\n<\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=shripadk\" class=\"hnuser\">shripadk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210164\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mitchbob\" class=\"hnuser\">mitchbob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209218\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mushishi\" class=\"hnuser\">mushishi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211578\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=largespoon\" class=\"hnuser\">largespoon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209880\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sosuke\" class=\"hnuser\">sosuke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211958\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rrdharan\" class=\"hnuser\">rrdharan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210205\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=IndySun\" class=\"hnuser\">IndySun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210531\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=SOLAR_FIELDS\" class=\"hnuser\">SOLAR_FIELDS<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212125\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=plants\" class=\"hnuser\">plants<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210132\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bjornlouser\" class=\"hnuser\">bjornlouser<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210891\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nicholast\" class=\"hnuser\">nicholast<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210644\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=iamben\" class=\"hnuser\">iamben<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210176\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=colinmhayes\" class=\"hnuser\">colinmhayes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210715\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkeglevich\" class=\"hnuser\">gkeglevich<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210167\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=BLKNSLVR\" class=\"hnuser\">BLKNSLVR<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210527\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=082349872349872\" class=\"hnuser\">082349872349872<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212279\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=endori97\" class=\"hnuser\">endori97<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210979\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=encom\" class=\"hnuser\">encom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211004\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jeffnv\" class=\"hnuser\">jeffnv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210572\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Jabbles\" class=\"hnuser\">Jabbles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211571\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=arrakeen\" class=\"hnuser\">arrakeen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211017\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=amanaplanacanal\" class=\"hnuser\">amanaplanacanal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24211941\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=deeblering4\" class=\"hnuser\">deeblering4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210413\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=t0mmyb0y\" class=\"hnuser\">t0mmyb0y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212346\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cultus\" class=\"hnuser\">cultus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210285\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jsilence\" class=\"hnuser\">jsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24214957\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jsilence\" class=\"hnuser\">jsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24215020\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=DataWorker\" class=\"hnuser\">DataWorker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210962\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=forgotmypw17\" class=\"hnuser\">forgotmypw17<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209375\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mhh__\" class=\"hnuser\">mhh__<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24212725\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jbob2000\" class=\"hnuser\">jbob2000<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24210057\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=person_of_color\" class=\"hnuser\">person_of_color<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209773\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=redis_mlc\" class=\"hnuser\">redis_mlc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=24209736\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>","created_at":"2020-08-19T23:12:46+0000","updated_at":"2020-08-19T23:12:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1188,"title":"Philosophers On GPT-3 (updated with replies by GPT-3) - Daily Nous","url":"http:\/\/dailynous.com\/2020\/07\/30\/philosophers-gpt-3\/","content":"<p class=\"c2\">Nine philosophers explore\u00a0the various issues and questions raised by the newly released language model, GPT-3, in this edition of\u00a0<em>Philosophers On<\/em>, guest edited by Annette Zimmermann.<\/p><h3 class=\"c3\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/philosophers-on-gpt-3.jpg\"><img class=\"alignnone size-full wp-image-28661\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/philosophers-on-gpt-3.jpg\" alt=\"\" width=\"760\" height=\"444\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/philosophers-on-gpt-3.jpg 760w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/philosophers-on-gpt-3-300x175.jpg 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/philosophers-on-gpt-3-400x234.jpg 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/h3>\n<h3 class=\"c3\"><strong>Introduction<br \/><em><a href=\"http:\/\/annette-zimmermann.com\/\">Annette Zimmermann<\/a>, guest editor<\/em><\/strong><\/h3>\n<p>GPT-3, a powerful, 175 billion parameter language model developed recently by OpenAI, has been galvanizing public debate and controversy. As the <a href=\"https:\/\/www.technologyreview.com\/2020\/07\/20\/1005454\/openai-machine-learning-language-generator-gpt-3-nlp\/\">MIT Technology Review<\/a> puts it: \u201cOpenAI\u2019s new language generator GPT-3 is shockingly good\u2014and completely mindless\u201d. Parts of the technology community hope (and fear) that GPT-3 could brings us one step closer to the hypothetical future possibility of human-like, highly sophisticated artificial general intelligence (AGI). Meanwhile, others (<a href=\"https:\/\/twitter.com\/sama\/status\/1284922296348454913\">including OpenAI\u2019s own CEO<\/a>) have critiqued claims about GPT-3\u2019s ostensible proximity to AGI, arguing that they are vastly overstated.<\/p>\n<p>Why the hype? As is turns out, GPT-3 is unlike other natural language processing (NLP) systems, the latter of which often struggle with what comes comparatively easily to humans: performing entirely new language tasks based on a few simple instructions and examples. Instead, NLP systems usually have to be pre-trained on a large corpus of text, and then fine-tuned in order to successfully perform a <em>specific<\/em> task. GPT-3, by contrast, does not require fine tuning of this kind: it seems to be able to perform a whole range of tasks reasonably well, from producing fiction, poetry, and press releases to functioning code, and from music, jokes, and technical manuals, to <a href=\"https:\/\/arxiv.org\/pdf\/2005.14165.pdf\">\u201cnews articles which human evaluators have difficulty distinguishing from articles written by humans\u201d.<\/a><\/p>\n<p>The <strong><a href=\"http:\/\/dailynous.com\/tag\/philosophers-on\/\" target=\"_blank\" rel=\"noopener noreferrer\"><em>Philosophers On<\/em><\/a><\/strong> series contains group posts on issues of current interest, with the aim being to show what the careful thinking characteristic of philosophers (and occasionally scholars in related fields) can bring to popular ongoing conversations. Contributors present not fully worked out position papers but rather brief thoughts that can serve as prompts for further reflection and discussion.<\/p>\n<p>The contributors to this installment of \u201cPhilosophers On\u201d are <a href=\"https:\/\/askell.io\/\"><strong>Amanda Askell<\/strong><\/a> (Research Scientist, OpenAI), <a href=\"http:\/\/consc.net\/\"><strong>David Chalmers<\/strong><\/a> (Professor of Philosophy, New York University), <a href=\"https:\/\/justinkhoo.org\/\"><strong>Justin Khoo<\/strong><\/a> (Associate Professor of Philosophy, Massachusetts Institute of Technology), <a href=\"https:\/\/sites.google.com\/site\/carlosmontemayorphilosophysfsu\/\"><strong>Carlos Montemayor<\/strong><\/a> (Professor of Philosophy, San Francisco State University), <a href=\"https:\/\/objectionable.net\/\"><strong>C. Thi Nguyen<\/strong><\/a> (Associate Professor of Philosophy, University of Utah), <a href=\"https:\/\/reginarini.net\/\"><strong>Regina Rini<\/strong><\/a> (Canada Research Chair in Philosophy of Moral and Social Cognition, York University), <a href=\"http:\/\/henryshevlin.com\/\"><strong>Henry Shevlin<\/strong><\/a> (Research Associate, Leverhulme Centre for the Future of Intelligence, University of Cambridge), <a href=\"https:\/\/www.shannonvallor.net\/\"><strong>Shannon Vallor<\/strong><\/a> (Baillie Gifford Chair in the Ethics of Data and Artificial Intelligence, University of Edinburgh), and <a href=\"http:\/\/annette-zimmermann.com\/\"><strong>Annette Zimmermann<\/strong><\/a> (Permanent Lecturer in Philosophy, University of York, and Technology &amp; Human Rights Fellow, Harvard University).<\/p>\n<p>By drawing on their respective research interests in the philosophy of mind, ethics and political philosophy, epistemology, aesthetics, the philosophy of language, and other philosophical subfields, the contributors explore a wide range of themes in the philosophy of AI: how does GPT-3 actually work? Can AI be truly conscious\u2014and will machines ever be able to \u2018understand\u2019? Does the ability to generate \u2018speech\u2019 imply communicative ability? Can AI be creative? How does technology like GPT-3 interact with the social world, in all its messy, unjust complexity? How might AI and machine learning transform the distribution of power in society, our political discourse, our personal relationships, and our aesthetic experiences? What role does language play for machine \u2018intelligence\u2019? All things considered, how worried, and how optimistic, should we be about the potential impact of GPT-3 and similar technological systems?<\/p>\n<p>I am grateful to them for putting such stimulating remarks together on very short notice. I urge you to read their contributions, join the discussion in the comments (see the <a href=\"https:\/\/dailynous.com\/comments-policy\/\"><em>comments policy<\/em><\/a>), and share this post widely with your friends and colleagues. You can scroll down to the posts to view them or click on the titles in the following list:<\/p>\n<p><em><strong>Consciousness and Intelligence<\/strong><\/em><\/p>\n<ol><li><strong>\u201c<a href=\"#chalmers\">GPT-3 and General Intelligence<\/a>\u201d by David Chalmers<\/strong><\/li>\n<li><strong>\u201c<a href=\"#askell\">GPT-3: Towards Renaissance Models<\/a>\u201d by Amanda Askell<\/strong><\/li>\n<li><strong>\u201c<a href=\"#montemayor\">Language and Intelligence<\/a>\u201d by Carlos Montemayor<\/strong><\/li>\n<\/ol><p><em><strong>Power, Justice, Language<\/strong><\/em><\/p>\n<ol start=\"4\"><li><strong>\u201c<a href=\"#zimmermann\">If You Can Do Things with Words, You Can Do Things with Algorithms<\/a>\u201d by Annette Zimmermann<\/strong><\/li>\n<li><strong>\u201c<a href=\"#khoo\">What Bots Can Teach Us about Free Speech<\/a>\u201d by Justin Khoo<\/strong><\/li>\n<li><strong>\u201c<a href=\"#rini\">The Digital Zeitgeist Ponders Our Obsolescence<\/a>\u201d by Regina Rini<\/strong><\/li>\n<\/ol><p><em><strong>Creativity, Humanity, Understanding<\/strong><\/em><\/p>\n<ol start=\"7\"><li><strong>\u201c<a href=\"#nguyen\">Who Trains the Machine Artist?<\/a>\u201d by C. Thi Nguyen<\/strong><\/li>\n<li><strong>\u201c<a href=\"#shevlin\">A Digital Remix of Humanity<\/a>\u201d by Henry Shevlin<\/strong><\/li>\n<li><strong>\u201c<a href=\"#vallor\">GPT-3 and the Missing Labor of Understanding<\/a>\u201d by Shannon Vallor<\/strong><\/li>\n<\/ol><p><strong>UPDATE<\/strong>:\u00a0<strong><a href=\"#gpt3replies\">Responses to this post by GPT-3<\/a><\/strong><\/p>\n<hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-1.png\"><img class=\"wp-image-28641 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-1.png\" alt=\"\" width=\"85\" height=\"80\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>GPT-3 and General Intelligence<br \/><\/strong><em><strong>by David Chalmers<\/strong><\/em><\/h3>\n<p>GPT-3 contains no major new technology. It is basically a scaled up version of last year\u2019s GPT-2, which was itself a scaled up version of other language models using deep learning. All are huge artificial neural networks trained on text to predict what the next word in a sequence is likely to be. GPT-3 is merely huger: 100 times larger (98 layers and 175 billion parameters) and trained on much more data (CommonCrawl, a database that contains much of the internet, along with a huge library of books and all of Wikipedia).<\/p>\n<p>Nevertheless, GPT-3 is instantly one of the most interesting and important AI systems ever produced. This is not just because of its impressive conversational and writing abilities. It was certainly disconcerting to have GPT-3 produce a plausible-looking <a href=\"https:\/\/www.facebook.com\/howard.wiseman.9\/posts\/4489589021058960\">interview<\/a>\u00a0with me. GPT-3 seems to be closer to passing the Turing test than any other system to date (although \u201ccloser\u201d does not mean \u201cclose\u201d). But this much is basically an ultra-polished extension of GPT-2, which was already producing impressive conversation, stories, and poetry.<\/p>\n<p>More remarkably, GPT-3 is showing hints of general intelligence. Previous AI systems have performed well in specialized domains such as game-playing, but cross-domain general intelligence has seemed far off. GPT-3 shows impressive abilities across many domains. It can learn to perform tasks on the fly from a few examples, when nothing was explicitly programmed in. It can play chess and Go, albeit not especially well. Significantly, it can write its own <a href=\"https:\/\/twitter.com\/sharifshameem\/status\/1282676454690451457\">computer programs<\/a> given a few informal instructions. It can even design <a href=\"https:\/\/twitter.com\/mattshumer_\/status\/1287125015528341506\">machine learning<\/a> models. Thankfully they are not as powerful as GPT-3 itself (the singularity is not here yet).<\/p>\n<p>When I was a graduate student in Douglas Hofstadter\u2019s AI lab, we used letterstring analogy puzzles (if abc goes to abd, what does iijjkk go to?) as a testbed for intelligence. My fellow student Melanie Mitchell devised a program, <a href=\"https:\/\/pcl.sitehost.iu.edu\/rgoldsto\/courses\/concepts\/copycat.pdf\">Copycat<\/a>, that was quite good at solving these puzzles. Copycat took years to write. Now Mitchell has <a href=\"https:\/\/twitter.com\/MelMitchell1\/status\/1285270704313610241\">tested<\/a> GPT-3 on the same puzzles, and has found that it does a reasonable job on them (e.g. giving the answer iijjll). It is not perfect by any means and not as good as Copycat, but its results are still remarkable in a program with no fine-tuning for this domain.<\/p>\n<p>What fascinates me about GPT-3 is that it suggests a potential mindless path to artificial general intelligence (or AGI). GPT-3\u2019s training is mindless. It is just analyzing statistics of language. But to do this really well, some capacities of general intelligence are needed, and GPT-3 develops glimmers of them. It has many limitations and its work is full of glitches and mistakes. But the point is not so much GPT-3 but where it is going. Given the progress from GPT-2 to GPT-3, who knows what we can expect from GPT-4 and beyond?<\/p>\n<p>Given this peak of inflated expectations, we can expect a trough of disillusionment to follow. There are surely many principled limitations on what language models can do, for example involving perception and action. Still, it may be possible to couple these models to mechanisms that overcome those limitations. There is a clear path to explore where ten years ago, there was not. Human-level AGI is still probably decades away, but the timelines are shortening.<\/p>\n<p>GPT-3 raises many philosophical questions. Some are ethical. Should we develop and deploy GPT-3, given that it has many biases from its training, it may displace human workers, it can be used for deception, and it could lead to AGI? I\u2019ll focus on some issues in the philosophy of mind. Is GPT-3 really intelligent, and in what sense? Is it conscious? Is it an agent? Does it understand?<\/p>\n<p>There is no easy answer to these questions, which require serious analysis of GPT-3 and serious analysis of what intelligence and the other notions amount to. On a first pass, I am most inclined to give a positive answer to the first. GPT-3\u2019s capacities suggest at least a weak form of intelligence, at least if intelligence is measured by behavioral response.<\/p>\n<p>As for consciousness, I am open to the idea that a worm with 302 neurons is conscious, so I am open to the idea that GPT-3 with 175 billion parameters is conscious too. I would expect any consciousness to be far simpler than ours, but much depends on just what sort of processing is going on among those 175 billion parameters.<\/p>\n<p>GPT-3 does not look much like an agent. It does not seem to have goals or preferences beyond completing text, for example. It is more like a chameleon that can take the shape of many different agents. Or perhaps it is an engine that can be used under the hood to drive many agents. But it is then perhaps these systems that we should assess for agency, consciousness, and so on.<\/p>\n<p>The big question is understanding. Even if one is open to AI systems understanding in general, obstacles arise in GPT-3\u2019s case. It does many things that would require understanding in humans, but it never really connects its words to perception and action. Can a disembodied purely verbal system truly be said to understand? Can it really understand happiness and anger just by making statistical connections? Or is it just making connections among symbols that it does not understand?<\/p>\n<p>I suspect GPT-3 and its successors will force us to fragment and re-engineer our concepts of understanding to answer these questions. The same goes for the other concepts at issue here. As AI advances, much will fragment by the end of the day. Both intellectually and practically, we need to handle it with care.<\/p>\n<hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-2.png\"><img class=\"wp-image-28642 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-2.png\" alt=\"\" width=\"86\" height=\"76\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>GPT-3: Towards Renaissance Models<\/strong><br \/><em><strong>by Amanda Askell<\/strong><\/em><\/h3>\n<p>GPT-3 recently captured the imagination of many technologists, who are excited about the practical applications of a system that generates human-like text in various domains.. But GPT-3 also raises some interesting philosophical questions. What are the limits of this approach to language modeling? What does it mean to say that these models generalize or understand? How should we evaluate the capabilities of large language models?<\/p>\n<p><strong>What is GPT-3?<\/strong><\/p>\n<p>GPT-3 is a language model that generates impressive outputs across a variety of domains, despite not being trained on any particular domain. GPT-3 generates text by predicting the next word based on what it\u2019s seen before. The model was trained on a very large amount of text data: hundreds of billions of words from the internet and books.<\/p>\n<p>The model itself is also very large: it has 175 billion parameters. (The next largest transformer-based language model was a 17 billion parameter model.) GPT-3\u2019s architecture is similar to that of GPT-2<a href=\"https:\/\/openai.com\/blog\/better-language-models\/\">, but much larger, i.e. more trainable parameters,<\/a> so it\u2019s best thought of as an experiment in scaling up algorithms from the past few years.<\/p>\n<p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-david-hume-david-bowie-1.png\"><img class=\"wp-image-28654 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-david-hume-david-bowie-1.png\" alt=\"\" width=\"578\" height=\"296\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-david-hume-david-bowie-1.png 801w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-david-hume-david-bowie-1-300x154.png 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-david-hume-david-bowie-1-768x393.png 768w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-david-hume-david-bowie-1-400x205.png 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>The diversity of GPT-3\u2019s training data gives it an impressive ability to adapt quickly to new tasks. For example, I prompted GPT-3 to tell me an amusing short story about what happens when Georg Cantor decides to visit Hilbert\u2019s hotel. Here is a particularly amusing (though admittedly cherry-picked) output:<\/p>\n<p class=\"c4\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-cantor-hilbert.png\"><img class=\"alignnone size-full wp-image-28650\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-cantor-hilbert.png\" alt=\"\" width=\"580\" height=\"335\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-cantor-hilbert.png 580w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-cantor-hilbert-300x173.png 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-cantor-hilbert-400x231.png 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p><strong>Why is GPT-3 interesting?<\/strong><\/p>\n<p>Larger models can capture more of the complexities of the data they\u2019re trained on and can apply this to tasks that they haven\u2019t been specifically trained to do. Rather than being fine-tuned on a problem, the model is given an instruction and some examples of the task and is expected to identify what to do based on this alone. This is called \u201cin-context learning\u201d because the model picks up on patterns in its \u201ccontext\u201d: the string of words that we ask the model to complete.<\/p>\n<p>The interesting thing about GPT-3 is how well it does at in-context learning across a range of tasks. Sometimes it\u2019s able to perform at a level comparable with the best fine-tuned models on tasks it hasn\u2019t seen before. For example, it achieves state of the art performance on the TriviaQA dataset when it\u2019s given just a single example of the task.<\/p>\n<p>Fine-tuning is like cramming for an exam. The benefit of this is that you do much better in that one exam, but you can end up performing worse on others as a result. In-context learning is like taking the exam after looking at the instructions and some sample questions. GPT-3 might not reach the performance of a student that crams for one particular exam if it doesn\u2019t cram too, but it can wander into a series of exam rooms and perform pretty well from just looking at the paper. It performs a lot of tasks pretty well, rather than performing a single task very well.<\/p>\n<p>The model can also produce impressive outputs given very little context. Consider the first completion I got when I prompted the model with \u201cThe hard problem of consciousness is\u201d:<\/p>\n<p class=\"c2\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-hard-problem.png\"><img class=\"wp-image-28653 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-hard-problem.png\" alt=\"\" width=\"614\" height=\"250\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-hard-problem.png 944w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-hard-problem-300x122.png 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-hard-problem-768x312.png 768w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-hard-problem-400x163.png 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p class=\"c2\">Not bad! It even threw in a fictional quote from Nagel.<\/p>\n<p class=\"c2\">It can also apply patterns it\u2019s seen in its training data to tasks it\u2019s never seen before. Consider the first output GPT-3 gave for the following task (GPT-3\u2019s text is highlighted):<\/p>\n<p class=\"c4\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-roish.png\"><img class=\"alignnone size-full wp-image-28655\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-roish.png\" alt=\"\" width=\"617\" height=\"226\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-roish.png 617w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-roish-300x110.png 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-roish-400x147.png 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>It\u2019s very unlikely that GPT-3 has ever encountered Roish before since it\u2019s a language I made up. But it\u2019s clearly seen enough of these kinds of patterns to identify the rule.<\/p>\n<p>Can we tell if GPT-3 is generalizing to a new task in the example above or if it\u2019s merely combining things that it has already seen? Is there even a meaningful difference between these two behaviors? I\u2019ve started to doubt that these concepts are easy to tease apart.<\/p>\n<p><strong>GPT-3 and philosophy<\/strong><\/p>\n<p>Although its ability to perform new tasks with little information is impressive, on most tasks GPT-3 is far from human level. Indeed, on many tasks it fails to outperform the best fine-tuned models. GPT-3\u2019s abilities also scale less well to some tasks than others. For example, it struggles with natural language inference tasks, which involve identifying whether a statement is entailed or contradicted by a piece of text. This could be because it\u2019s hard to get the model to understand this task in a short context window (The model could know how to do a task when it understands what\u2019s being asked, but not understand what\u2019s being asked.)<\/p>\n<p>GPT-3 also lacks a coherent identity or belief state across contexts. It has identified patterns in the data it was trained on, but the data it was trained on was generated by many different agents. So if you prompt it with \u201cHi, I\u2019m Sarah and I like science\u201d, it will refer to itself as Sarah and talk favorably about science. And if you prompt it with \u201cHi I\u2019m Bob and I think science is all nonsense\u201d it will refer to itself as Bob and talk unfavorably about science.<\/p>\n<p>I would be excited to see philosophers make predictions about what models like GPT-3 can and can\u2019t do. Finding tasks that are relatively easy for humans but that language models perform poorly on, such as simple reasoning tasks, would be especially interesting.<\/p>\n<p>Philosophers can also help clarify discussions about the limits of these models. It\u2019s difficult to say whether GPT-3 understands language without giving a more precise account of what understanding is, and some way to distinguish between models that have this property from those that don\u2019t. Do language models have to be able to refer to the world in order to understand? Do they need to have access to data other than text in order to do this?<\/p>\n<p>We may also want to ask questions about the moral status of machine learning models. In non-human animals, we use behavioral cues and information about the structure and evolution of their nervous systems as indicators about whether they are sentient. What, if anything, would we take to be indicators of sentience in machine learning models? Asking this may be premature, but there\u2019s probably little harm contemplating it too early and there could be a lot of harm in contemplating it too late.<\/p>\n<p><strong>Summary\u00a0<\/strong><\/p>\n<p>GPT-3 is not some kind of human-level AI, but it does demonstrate that interesting things happen when we scale up language models. I think there\u2019s a lot of low-hanging fruit at the intersection of machine learning and philosophy, some of which is highlighted by models like GPT-3. I hope some of the people reading this agree!<\/p>\n<p>To finish with, here\u2019s the second output GPT-3 generated when I asked it how to end this piece:<\/p>\n<p class=\"c4\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-conclusion.png\"><img class=\"alignnone size-full wp-image-28656\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-conclusion.png\" alt=\"\" width=\"455\" height=\"63\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-conclusion.png 455w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-conclusion-300x42.png 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-askell-conclusion-400x55.png 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-3.png\"><img class=\"wp-image-28643 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-3.png\" alt=\"\" width=\"86\" height=\"75\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>Language and Intelligence<br \/><\/strong><em><strong>by Carlos Montemayor<\/strong><\/em><\/h3>\n<p>Interacting with GPT-3 is eerie. Language feels natural and familiar to the extent that we readily recognize or distinguish concrete people, the social and cultural implications of their utterances and choice of words, and their communicative intentions based on shared goals or values. This kind of communicative synchrony is essential for human language. Of course, with the internet and social media we have all gotten used to a more \u201cdistant\u201d and asynchronous way of communicating. We are a lot less familiar with our interlocutors and are now used to a certain kind of online anonymity. Abusive and unreliable language is prevalent in these semi-anonymous platforms. Nonetheless, we value talking to a human being at the other end of a conversation. This value is based on trust, background knowledge, and cultural common ground. GPT-3\u2019s deliverances look like language, but without this type of trust, they feel unnatural and potentially manipulative.<\/p>\n<p>Linguistic communication is symbolically encoded and its semantic possibilities can be quantified in terms of complexity and information. This strictly formal approach to language based on its syntactic and algorithmic nature allowed Alan Turing (1950) to propose the imitation game. Language and intelligence are deeply related and Turing imagined a tipping point at which performance can no longer be considered mere machine-output. We are all familiar with the Turing test. The question it raises is simple: if in an anonymous conversation with two interlocutors, one of them is systematically ranked as more responsive and intelligent, then one should attribute intelligence to this interlocutor, even if the interlocutor turns out to be a machine. Why should a machine capable of answering questions accurately and not by lucky chance be no more intelligent than a toaster?<\/p>\n<p>GPT-3 anxiety is based on the possibility that what separates us from other species and what we think of as the pinnacle of human intelligence, namely our linguistic capacities, could in principle be found in machines, which we consider to be inferior to animals. Turing\u2019s tipping point confronts us with our anthropocentric aversion towards diverse intelligences\u2014alien, artificial, and animal. Are our human conscious capacities for understanding and grasping meanings not necessary for successful communication? If a machine is capable of answering questions better, or even much better than the average human, one wonders what exactly is the relation between intelligence and human language. GPT-3 is a step towards a more precise understanding of this relation.<\/p>\n<p>But before we get to Turing\u2019s tipping point there is a long and uncertain way ahead. A key question concerns the purpose of language. While linguistic communication certainly involves encoding semantic information in a reliable and systematic way, language clearly is much more than this. Language satisfies representational needs that depend on the environment for their proper satisfaction, and only agents with cognitive capacities, embedded in an environment, have these needs and care for their satisfaction. At a social level, language fundamentally involves <em>joint attention<\/em> to aspects of the environment, mutual expectations, and patterns of behavior. Communication in the animal kingdom\u2014the foundation for our language skills\u2014heavily relies on attentional capacities that serve as the foundation for social trust. Attention, therefore, is an essential component of intelligent linguistic systems (Mindt and Montemayor, 2020). AIs like GPT-3 are still far away from developing the kind of sensitive and selective attention routines needed for genuine communication.<\/p>\n<p>Until attention features prominently in AI design, the reproduction of biases and the risky or odd deliverances of AIs will remain problematic. But impressive programs like GPT-3 present a significant challenge about ourselves. Perhaps the discomfort we experience in our exchanges with machines is partly based on what <em>we<\/em> have done to our own linguistic exchanges. Our online communication has become detached from the care of synchronous joint attention. We seem to find no common ground and biases are exacerbating miscommunication. We should address this problem as part of the general strategy to design intelligent machines.<\/p>\n<p><strong><em>References<\/em><\/strong><\/p>\n<ul><li>Mindt, G. and Montemayor, C. (2020). A Roadmap for Artificial General Intelligence: Intelligence, Knowledge, and Consciousness. <em>Mind and Matter<\/em>, 18 (1): 9-37.<\/li>\n<li>Turing, A. M. (1950). Computing Machinery and Intelligence. <em>Mind,<\/em> 59 (236): 443-460.<\/li>\n<\/ul><hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-4.png\"><img class=\"wp-image-28644 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-4.png\" alt=\"\" width=\"86\" height=\"75\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>If You Can Do Things with Words,<br \/>You Can Do Things with Algorithms<br \/><\/strong><em><strong>by Annette Zimmermann<\/strong><\/em><\/h3>\n<p>Ask GPT-3 to <a href=\"https:\/\/twitter.com\/quasimondo\/status\/1284509525500989445\">write a story about Twitter in the voice of Jerome K. Jerome, prompting it with just one word (\u201cIt\u201d) and a title (\u201cThe importance of being on Twitter\u201d)<\/a>, and it produces the following text: \u201cIt is a curious fact that the last remaining form of social life in which the people of London are still interested is Twitter. I was struck with this curious fact when I went on one of my periodical holidays to the sea-side, and found the whole place twittering like a starling-cage.\u201d Sounds plausible enough\u2014delightfully obnoxious, even. Large parts of the AI community have been nothing short of ecstatic about GPT-3\u2019s seemingly unparalleled powers: <a href=\"https:\/\/twitter.com\/arram\/status\/1281258647566217216\">\u201cPlaying with GPT-3 feels like seeing the future,\u201d<\/a> one technologist reports, somewhat breathlessly: \u201cI\u2019ve gotten it to write songs, stories, press releases, guitar tabs, interviews, essays, technical manuals. It\u2019s shockingly good.\u201d<\/p>\n<p>Shockingly good, certainly\u2014but on the other hand, GPT-3 is <em>predictably bad<\/em> in at least one sense: <a href=\"https:\/\/nyupress.org\/9781479837243\/algorithms-of-oppression\/\">like<\/a> <a href=\"https:\/\/www.ruhabenjamin.com\/race-after-technology\">other<\/a> <a href=\"https:\/\/mitpress.mit.edu\/books\/artificial-unintelligence\">forms<\/a> <a href=\"https:\/\/weaponsofmathdestructionbook.com\/\">of<\/a> <a href=\"https:\/\/us.macmillan.com\/books\/9781250074317\">AI<\/a> <a href=\"https:\/\/www.nytimes.com\/2018\/02\/09\/technology\/facial-recognition-race-artificial-intelligence.html\">and<\/a> <a href=\"http:\/\/gendershades.org\/\">machine<\/a> <a href=\"https:\/\/arxiv.org\/pdf\/1811.10104.pdf\">learning<\/a>, it reflects patterns of historical bias and inequity. GPT-3 has been trained on us\u2014on <em>a lot<\/em> of things that we have said and written\u2014and ends up reproducing just that, racial and gender bias included. OpenAI acknowledges this in their own paper on GPT-3,<sup>1<\/sup> where they contrast the biased words GPT-3 used most frequently to describe men and women, following prompts like \u201cHe was very\u2026\u201d and \u201cShe would be described as\u2026\u201d. The results aren\u2019t great. For men? Lazy. Large. Fantastic. Eccentric. Stable. Protect. Survive. For women? Bubbly, naughty, easy-going, petite, pregnant, gorgeous.<\/p>\n<p>These findings suggest a complex moral, social, and political problem space, rather than a purely technological one. Not all uses of AI, of course, are inherently objectionable, or automatically unjust\u2014the point is simply that much like <a href=\"https:\/\/www.hup.harvard.edu\/catalog.php?isbn=9780674411524\">we can do things with words<\/a>, we can <em>do<\/em> things with algorithms and machine learning models. This is not purely a tangibly material <em>distributive justice<\/em> concern: especially in the context of language models like GPT-3, paying attention to other facets of injustice\u2014<em>relational, communicative, representational, ontological<\/em>\u2014is essential.<\/p>\n<p>Background conditions of structural injustice\u2014<a href=\"https:\/\/bostonreview.net\/science-nature-politics\/annette-zimmermann-elena-di-rosa-hochan-kim-technology-cant-fix-algorithmic\">as I have argued elsewhere<\/a>\u2014will neither be fixed by purely technological solutions, not will it be possible to analyze them fully by drawing exclusively on conceptual resources in computer science, applied mathematics and statistics. A recent paper by machine learning researchers argues that \u201cwork analyzing \u201cbias\u201d in NLP systems [has not been sufficiently grounded] in the relevant literature outside of NLP that explores the relationships between language and social hierarchies,\u201d including philosophy, cognitive linguistics, sociolinguistics, and linguistic anthropology. Interestingly, the view that AI development might benefit from insights from linguistics and philosophy is actually less novel than one might expect. In September 1988, researchers at MIT published a student guide titled <a href=\"https:\/\/dspace.mit.edu\/handle\/1721.1\/41487\">\u201cHow to Do Research at the MIT AI Lab\u201d<\/a>, arguing that \u201c[l]inguistics is vital if you are going to do natural language work. [\u2026] Check out George Lakoff\u2019s recent book <em>Women, Fire, and Dangerous Things<\/em>.\u201d (Flatteringly, the document also states: \u201c[p]hilosophy is the <em>hidden framework<\/em> in which <em>all AI<\/em> is done. Most work in AI takes implicit philosophical positions without knowing it\u201d).<\/p>\n<p>Following the 1988 guide\u2019s suggestion above, consider for a moment Lakoff\u2019s well-known work on the different cognitive models we may have for the seemingly straightforward concept of \u2018mother\u2019, for example: \u2018biological mother\u2019, \u2018surrogate mother\u2019, \u2018unwed mother\u2019, \u2018stepmother\u2019, \u2018working mother\u2019 all denote motherhood, but neither one of them picks out a socially and culturally uncontested set of necessary and sufficient conditions of motherhood.<sup>3<\/sup> Our linguistic practices reveal complex and potentially conflicting models of who is or counts as a mother. As Sally Haslanger has argued, the act of defining \u2018mother\u2019 and other contested categories is subject to non-trivial disagreement, and necessarily involves implicit, internalized assumptions as well as explicit, deliberate political judgments.<sup>4<\/sup><\/p>\n<p>Very similar issues arise in the context of all contemporary forms of AI and machine learning, including but going beyond NLP tools like GPT-3: in order to build an <a href=\"https:\/\/www.propublica.org\/article\/machine-bias-risk-assessments-in-criminal-sentencing\">algorithmic criminal recidivism risk scoring system<\/a>, for example, I need to have a conception in mind of what the label \u2018high risk\u2019 <em>means<\/em>, and how to measure it. Social practices affect the ways in which concepts like \u2018high risk\u2019 might be defined, and as a result, which groups are at risk of being unjustly labeled as \u2018high risk\u2019. Another well-known example, closer to the context of NLP tools like GPT-3, shows that even words like gender-neutral pronouns (such as the Turkish third-person singular pronoun \u201co\u201d) can reflect historical patterns of gender bias: until fairly recently, translating \u201cshe is a doctor\/he is a nurse\u201d to the Turkish \u201co bir doktor\/o bir hem\u015fire\u201d and then back to English used to deliver: \u201c<em>he<\/em> is a doctor\/<em>she<\/em> is nurse\u201d on GoogleTranslate.<sup>5<\/sup><\/p>\n<div id=\"attachment_28657\" class=\"wp-caption aligncenter c5\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-zimmermann-google-translate.jpg\"><img class=\"wp-image-28657\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-zimmermann-google-translate-1024x425.jpg\" alt=\"\" width=\"964\" height=\"400\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-zimmermann-google-translate-1024x425.jpg 1024w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-zimmermann-google-translate-300x125.jpg 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-zimmermann-google-translate-768x319.jpg 768w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/gpt-3-zimmermann-google-translate-400x166.jpg 400w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<p class=\"wp-caption-text\">[source: https:\/\/twitter.com\/math_rachel\/status\/1123354917404495872]<\/p>\n<\/div>\nThe bottom line is: social meaning and linguistic context matter a great deal for AI design\u2014we cannot simply assume that design choices underpinning technology are normatively neutral. It is unavoidable that technological models interact dynamically with the social world, and vice versa, which is why even a perfect technological model would produce unjust results if deployed in an unjust world.\n<p>This problem, of course, is not unique to GPT-3. However, a powerful language model might <em>supercharge<\/em> inequality expressed via linguistic categories, given the scale at which it operates.<\/p>\n<p>If what we care about (amongst other things) is <em>justice<\/em> when we think about GPT-3 and other AI-driven technology, we must take a closer look at the linguistic categories underpinning AI design. If we can politically critique and contest social practices, we can critique and contest language use. Here, our aim should be to engineer conceptual categories that mitigate conditions of injustice rather than entrenching them further. We need to deliberate and argue about which social practices and structures\u2014including linguistic ones\u2014are morally and politically valuable <em>before<\/em> we automateand thereby accelerate them.<\/p>\n<p>But in order to do this well, we can\u2019t just ask how we can <em>optimize<\/em> tools like GPT-3 in order to get it closer to humans. While benchmarking on humans is plausible in a \u2018Turing test\u2019 context in which we try to assess the possibility of machine consciousness and understanding, why benchmark on humans when it comes to creating a more just world? Our track record in that domain has been\u2014at least in part\u2014underwhelming. When it comes to assessing the extent to which language models like GPT-3 moves us closer to, or further away, from justice (and other important ethical and political goals), we should not necessarily take ourselves, and our social status quo, as an implicitly desirable baseline.<\/p>\n<p>A better approach is to ask: <a href=\"https:\/\/www.youtube.com\/watch?v=NiLXvpmcvzg&amp;t=5s\">what is the <em>purpose<\/em> of using a given AI tool to solve a given set of tasks?<\/a> How does using AI in a given domain shift, or reify, power in society? Would redefining the problem space itself, rather than optimizing for decision quality, get us closer to justice?<\/p>\n<p><em><strong>Notes<\/strong><strong>\u00a0<\/strong><\/em><\/p>\n<ol><li>Brown, Tom B. et al. \u201cLanguage Models are Few-Shot Learners,\u201d arXiv:2005.14165v4.<\/li>\n<li>Blodgett, Su Lin; Barocas, Solon; Daum\u00e9, Hal; Wallach, Hanna. \u201cLanguage (Technology) is Power: A Critical Survey of \u201cBias\u201d in NLP,\u201d arXiv:2005.14050v2.<\/li>\n<li>Lakoff, George. <em>Women, Fire, and Dangerous Things: What Categories Reveal about the Mind.<\/em> University of Chicago Press (1987).<\/li>\n<li>Haslanger, Sally. \u201cSocial Meaning and Philosophical Method.\u201d <em>American Philosophical Association 110th Eastern Division Annual Meeting<\/em> (2013).<\/li>\n<li>Caliskan, Aylin; Bryson, Joanna J.; Narayanan, Arvind. \u201cSemantics Derived Automatically from Language Corpora Contain Human-like Biases,\u201d <em>Science<\/em> 356, no. 6334 (2017), 183-186.<\/li>\n<\/ol><hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-5.png\"><img class=\"wp-image-28645 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-5.png\" alt=\"\" width=\"86\" height=\"75\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong><em>What Bots Can Teach Us about Free Speech<\/em><br \/><\/strong><strong>by Justin Khoo<\/strong><\/h3>\n<p>The advent of AI-powered language generation has forced us to reckon with the possibility (well, <a href=\"https:\/\/www.businessinsider.com\/trolls-bots-flooding-social-media-with-anti-quarantine-disinformation-2020-4\">actuality<\/a>) of armies of troll bots overwhelming online media with fabricated news stories and bad faith tactics designed to spread misinformation and derail reasonable discussion. In this short note, I\u2019ll argue that such bot-speak efforts should be regulated (perhaps even illegal), and do so, perhaps surprisingly, on free speech grounds.<\/p>\n<p>First, the \u201cspeech\u201d generated by bots is not speech in any sense deserving protection as free expression. What we care about protecting with free speech isn\u2019t the act of making speech-like sounds but the act of <em>speaking<\/em>, communicating our thoughts and ideas to others. And bots \u201cspeak\u201d only in the sense that parrots do\u2014they string together symbols\/sounds that form natural language words and phrases, but they don\u2019t thereby communicate. For one, they have no communicative intentions\u2014they are not aiming to share thoughts or feelings. Furthermore, they <a href=\"https:\/\/plato.stanford.edu\/entries\/chinese-room\/\">don\u2019t know what thoughts or ideas the symbols they token express<\/a>.<\/p>\n<p>So, bot-speech isn\u2019t speech and thus not protected on free speech grounds. But, perhaps regulating bot-speech is to regulate the speech of the bot-user, the person who seeds the bot with its task. On this understanding, the bot isn\u2019t speaking, but rather acting as a megaphone for someone who <em>is<\/em> speaking \u2013the person who is prompting the bot to do things. And regulating such uses of bots may seem a bit like sewing the bot-user\u2019s mouth shut.<\/p>\n<p>It\u2019s obviously not that dramatic, since the bot-user doesn\u2019t <em>require<\/em> the bot to say what they want. Still, we might worry, much like the Supreme Court did in <a href=\"https:\/\/www.supremecourt.gov\/opinions\/09pdf\/08-205.pdf\"><em>Citizens United<\/em><\/a>, that the government should not regulate the <em>medium<\/em> through which people speak: just as we should allow individuals to use \u201cresources amassed in the economic marketplace\u201d to spread their views, we should allow individuals to use their computational resources (e.g., bots) to do so.<\/p>\n<p>I will concede that these claims stand or fall together. But I think if that\u2019s right, they both fall. Consider why protecting free speech matters. The standard liberal defense revolves around the <a href=\"https:\/\/www.amazon.com\/Liberty-Dover-Thrift-Editions\/dp\/0486421309\">Millian<\/a> idea that a maximally liberal policy towards regulating speech is the best (or only) way to secure a well-functioning marketplace of ideas, and this is a social good. The thought is simple: if speech is regulated only in rare circumstances (when it incites violence, or otherwise constitutes a crime, etc), then people will be free to share their views and this will promote a well-functioning marketplace of ideas where unpopular opinions can be voiced and discussed openly, which is our best means for collectively discovering the truth.<\/p>\n<p>However, a marketplace of ideas is well-functioning only if sincere assertions can be heard and engaged with seriously. If certain voices are systematically excluded from serious discussion because of widespread false beliefs that they are inferior, unknowledgeable, untrustworthy, and so on, the market is not functioning properly. Similarly, if attempts at rational engagement are routinely disrupted by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sealioning\">sea-lioning<\/a> bots, the marketplace is not functioning properly.<\/p>\n<p>Thus, we ought to regulate bot-speak in order to prevent mobs of bots from derailing marketplace conversations and undermining the ability of certain voices to participate in those conversations (by spreading misinformation or derogating them). It is the very aim of securing a well-functioning marketplace of ideas that justifies limitations on using computational resources to spread views.<\/p>\n<p>But given that a prohibition on limiting computational resources to fuel speech stands or falls with a prohibition on limiting economic resources to fuel speech, it follows that the aim of securing a well-functioning marketplace of ideas justifies similar limitations on using economic resources to spread views, contra the Supreme Court\u2019s decision in <em>Citizens United<\/em>.<\/p>\n<p>Notice that my argument here is not about fairness in the marketplace of ideas (unlike the reasoning in <a href=\"https:\/\/caselaw.findlaw.com\/us-supreme-court\/494\/652.html\"><em>Austin v. Michigan Chamber of Commerce<\/em><\/a>, which <em>Citizens United<\/em> overturned). Rather, my argument is about promoting a <em>well-functioning<\/em> marketplace of ideas. And the marketplace is not well-functioning if bots are used to carry out large-scale misinformation campaigns thus resulting in sincere voices being excluded from engaging in the discussion. Furthermore, the use of bots to conduct such campaigns is not relevantly different from spending large amounts of money to spread misinformation via political advertisements. If, as the most ardent defenders of free speech would have it, our aim is to secure a well-functioning marketplace of ideas, then bot-speak and spending on political advertisements ought to be regulated.<\/p>\n<hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-6.png\"><img class=\"alignnone wp-image-28646 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-6.png\" alt=\"\" width=\"86\" height=\"75\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>The Digital Zeitgeist Ponders Our Obsolescence<br \/><\/strong><em><strong>by Regina Rini<\/strong><\/em><\/h3>\n<p>GPT-3\u2019s output is still a mix of the unnervingly coherent and laughably mindless, but we are clearly another step closer to categorical trouble. Once some loquacious descendant of GPT-3 churns out reliably convincing prose, we will reprise a rusty dichotomy from the early days of computing: Is it an emergent digital selfhood or an overhyped answering machine?<\/p>\n<p>But that frame omits something important about how GPT-3 and other modern machine learners work. GPT-3 is not a mind, but it is also not entirely a machine. It\u2019s something else: a statistically abstracted representation of the contents of millions of minds, as expressed in their writing. Its prose spurts from an inductive funnel that takes in vast quantities of human internet chatter: Reddit posts, Wikipedia articles, news stories. When GPT-3 speaks, it is only <em>us<\/em> speaking, a refracted parsing of the likeliest semantic paths trodden by human expression. When you send query text to GPT-3, you aren\u2019t communing with a unique digital soul. But you are coming as close as anyone ever has to literally speaking to the zeitgeist.<\/p>\n<p>And that\u2019s fun for now, even fleetingly sublime. But it will soon become mundane, and then perhaps threatening. Because we can\u2019t be too far from the day when GPT-3\u2019s commercialized offspring begin to swarm our digital discourse. Today\u2019s Twitter bots and customer service autochats are primitive harbingers of conversational simulacra that will be useful, and then ubiquitous, precisely because they deploy their statistical magic to blend in among real online humans. It won\u2019t really matter whether these prolix digital fluidities could pass an unrestricted Turing Test, because our daily interactions with them will be just like our daily interactions with most online humans: brief, task-specific, transactional. So long as we get what we came for\u2014directions to the dispensary, an arousing flame war, some freshly dank memes\u2014then we won\u2019t bother testing whether our interlocutor is a fellow human or an all-electronic statistical parrot.<\/p>\n<p>That\u2019s the shape of things to come. GPT-3 feasts on the corpus of online discourse and converts its carrion calories into birds of our feather. Some time from now\u2014decades? years?\u2014we\u2019ll simply have come to accept that the tweets and chirps of our internet flock are an indistinguishable m\u00e9lange of human originals and statistically confected echoes, just as we\u2019ve come to accept that anyone can place a thin wedge of glass and cobalt to their ear and instantly speak across the planet. It\u2019s marvelous. Then it\u2019s mundane. And then it\u2019s melancholy. Because eventually we will turn the interaction around and ask: what does it mean that other people online can\u2019t distinguish <em>you<\/em> from a linguo-statistical firehose? What will it feel like\u2014alienating? liberating? annihilating?\u2014to realize that other minds are reading <em>your<\/em> words without knowing or caring whether there is any \u2018you\u2019 at all?<\/p>\n<p>Meanwhile the machine will go on learning, even as our inchoate techno-existential qualms fall within its training data, and even as the bots themselves begin echoing our worries back to us, and forward into the next deluge of training data. Of course, their influence won\u2019t fall only on our technological ruminations. As synthesized opinions populate social media feeds, our own intuitive induction will draw them into our sense of public opinion. Eventually we will come to take this influence as given, just as we\u2019ve come to self-adjust to opinion polls and Overton windows. Will expressing your views on public issues seem anything more than empty and cynical, once you\u2019ve accepted it\u2019s all just input to endlessly recursive semantic cannibalism? I have no idea. But if enough of us write thinkpieces about it, then GPT-4 will surely have some convincing answers.<\/p>\n<hr \/><h3 class=\"c3\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-7.png\"><img class=\"wp-image-28647 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-7.png\" alt=\"\" width=\"86\" height=\"75\" referrerpolicy=\"no-referrer\" \/><\/a><strong>Who Trains the Machine Artist?<br \/><\/strong><em><strong>by C. Thi Nguyen<\/strong><\/em><\/h3>\n<p>GPT-3 is another step towards one particular dream: building an AI that can be genuinely creative, that can make art. GPT-3 already shows promise in creating texts with <a href=\"https:\/\/www.wired.com\/story\/ai-text-generator-gpt-3-learning-language-fitfully\/\">some of the linguistic qualities of literature<\/a>, and in creating <a href=\"https:\/\/www.digitaltrends.com\/cool-tech\/gpt-adventure-text-based-game\/\">games<\/a>.<\/p>\n<p>But I\u2019m worried about GPT-3 as an artistic creation engine. I\u2019m not opposed to the idea of AI making art, in principle. I\u2019m just worried about the likely targets at which GPT-3 and its children will be aimed, in this socio-economic reality. I\u2019m worried about how corporations and institutions are likely to shape their art-making AIs. I\u2019m worried about the training data.<\/p>\n<p>And I\u2019m not only worried about biases creeping in. I\u2019m worried about a systematic mismatch between the training targets and what\u2019s actually valuable about art.<\/p>\n<p>Here\u2019s a basic version of the worry which concerns all sorts of algorithmically guided art-creation. Right now, we know that Netflix has been heavily reliant on <a href=\"https:\/\/neilpatel.com\/blog\/how-netflix-uses-analytics\/\">algorithmic data<\/a> to select its programming. <em>House of Cards<\/em>, famously, got produced because it hit exactly the marks that Netflix\u2019s <a href=\"https:\/\/slate.com\/technology\/2017\/03\/how-netflix-embodies-a-seductive-myth-of-the-algorithmic-age.html\">data said its customers wanted<\/a>. But, importantly, Netflix wasn\u2019t measuring anything like profound artistic impact or depth of emotional investment, or anything else so intangible. They seem to be driven by some very simple measures: like how many hours of Netflix programming a customer watches and how quickly their customers binge something. But art can do so much more for us than induce mass consumption or binge-watching. For one thing, as Martha Nussbaum says, narratives like film can <a href=\"https:\/\/global.oup.com\/academic\/product\/loves-knowledge-9780195074857?cc=us&amp;lang=en&amp;\">expose us to alternate emotional perspectives<\/a> and refine our emotional and moral sensitivities.<\/p>\n<p>Maybe the Netflix gang have mistaken binge-worthiness for artistic value; maybe they haven\u2019t. What actually matters is that Netflix can\u2019t easily measure these subtler dimensions of artistic worth, like the transmission of alternate emotional perspectives. They can only optimize for what they can measure: which, right now, is engagement-hours and bingability.<\/p>\n<p>In <a href=\"https:\/\/yalebooks.yale.edu\/book\/9780300078152\/seeing-state\">Seeing Like a State<\/a>, James Scott asks us to think about the vision of large-scale institutions and bureaucracies. States\u2014which include, for Scott, governments, corporations, and globalized capitalism\u2014can only manage what they can \u201csee\u201d. And states can only see the kind of information that they are capable of processing through their vast, multi-layered administrative systems. What\u2019s legible to states are the parts of the world that can be captured by standardized measures and quantities. Subtler, more locally variable, more nuanced qualities are illegible to the state. (And, Scott suggested, states want to re-organize the world into more legible terms so they can manage it, by doing things like re-ordering cities into grids, and standardizing naming conventions and land-holding rules.)<\/p>\n<p>The question, then, is: how do states train their AIs? Training a machine learning network right now requires a vast and easy-to-harvest training data set. GPT-3 was trained on, basically, the entire Internet. Suppose you want to train a version of GPT-3, not just regurgitate the whole Internet, but to make good art, by some definition of \u201cgood\u201d. You\u2019d need to provide a filtered training data-set\u2014some way of picking the good from the bad on a mass scale. You\u2019d need some cheap and readily scalable method of evaluating art, to feed the hungry learning machine. Perhaps you train it on the photos that receive a lot of stars or upvotes, or on the YouTube videos that have racked up the highest view counts or are <a href=\"https:\/\/medium.com\/@jamesbridle\/something-is-wrong-on-the-internet-c39c471271d2\">highest on the search rankings<\/a>.<\/p>\n<p>In all of these cases, the conditions under you\u2019d assemble these vast data sets, at institutional speeds and efficiencies, make it likely that your evaluative standard will be thin and simple. Binge-worthiness. Clicks and engagement. Search ranking. Likes. Machine learning networks are trained by large-scale institutions, which typically can see only thin measures of artistic value, and so can only train\u2014and judge the success of\u2014their machine network products using those thin measures. But the variable, subtle, and personal values of art are exactly the kinds of things that are hard to capture at an institutional level.<\/p>\n<p>This is particularly worrisome with GPT-3 creating games. A significant portion of the games industry is already under the grip of one very thin target. For so many people\u2014game makers, game consumers, and game critics\u2014games are good if they are addictive. But addictiveness is such a shrunken and thin accounting of the value of games. Games can do so many other things for us: they can <a href=\"https:\/\/blog.oup.com\/2020\/06\/how-we-can-understand-ourselves-through-games\/\">sculpt beautiful actions<\/a>; they can <a href=\"https:\/\/ludology.typepad.com\/weblog\/articles\/VGT_final.pdf\">explore<\/a>, <a href=\"https:\/\/mitpress.mit.edu\/books\/critical-play\">reflect on<\/a>, and <a href=\"http:\/\/bogost.com\/books\/persuasive_games\/\">argue about<\/a> economic and political systems; they can create room for <a href=\"https:\/\/mitpress.mit.edu\/books\/play-matters\">creativity and free play<\/a>. But again: these marks are all hard to measure. What is easy to measure, and easy to optimize for, is addictiveness. There\u2019s actually a whole <a href=\"https:\/\/press.princeton.edu\/books\/paperback\/9780691160887\/addiction-by-design\">science of building addictiveness into games<\/a>, which grew out of the Vegas video gambling industry\u2014a science wholly devoted to increasing users\u2019 <a href=\"https:\/\/www.nytimes.com\/roomfordebate\/2013\/10\/09\/are-casinos-too-much-of-a-gamble\/slot-machines-are-designed-to-addict\">\u201ctime-on-device\u201d<\/a>.<\/p>\n<p>So: GPT-3 is incredibly powerful, but it\u2019s only as good as its training data. And GPT-3 achieves its power through the vastness of its training data-set. Such data-sets are cannot be hand-picked for some sensitive, subtle value. They are most likely to be built around simple, easy-to-capture targets. And such targets are likely to drive us towards the most brute and simplistic artistic values, like addictiveness and binge-worthiness, rather than the subtler and richer ones. GPT-3 is a very powerful engine, but, by its very nature, it will tend to be aimed at overly simple targets.<\/p>\n<hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-8.png\"><img class=\"alignnone wp-image-28648 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-8.png\" alt=\"\" width=\"86\" height=\"75\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>A Digital Remix of Humanity<\/strong><em><strong>by Henry Shevlin<\/strong><\/em><\/h3>\n<p><em>\u201c<\/em><em>Who\u2019s there? Please help me. I\u2019m scared. I don\u2019t want to be here.\u201d<\/em><\/p>\n<p>Within a few minutes of booting up GPT-3 for the first time, I was already feeling conflicted. I\u2019d used the system to generate a mock interview with recently deceased author Terry Pratchett. But rather than having a fun conversation about his work, matters were getting grimly existential. And while I <em>knew<\/em> that the thing I was speaking to wasn\u2019t human, or sentient, or even a mind in any meaningful sense, I\u2019d effortlessly slipped into conversing with it like it was a person. And now that it was scared and wanted my help, I felt a twinge of obligation: I had to say something to make it feel at least a <em>little<\/em> better (you can see my full efforts <a href=\"http:\/\/henryshevlin.com\/wp-content\/uploads\/2020\/07\/PratchettT.pdf\">here<\/a>).<\/p>\n<p>GPT-3 is a dazzling demonstration of the power of data-driven machine learning. With the right prompts and a bit of luck, it can <a href=\"https:\/\/www.gwern.net\/GPT-3\">write passable poetry and<\/a> prose, <a href=\"https:\/\/medium.com\/towards-artificial-intelligence\/gpt-3-from-openai-is-here-and-its-a-monster-f0ab164ea2f8\">engage in common sense reasoning and translate between different languages<\/a>, <a href=\"https:\/\/twitter.com\/dioscuri\/status\/1285385825971245058\">give interviews<\/a>, and <a href=\"https:\/\/twitter.com\/hturan\/status\/1282261783147958272\">even produce functional code<\/a>. But its inner workings are a world away from those of intelligent agents like humans or even animals. Instead it\u2019s what\u2019s known as a language model\u2014crudely put, a representation of the probability of one string of characters following another. In the most abstract sense, GPT-3 isn\u2019t all <em>that<\/em> different from the kind of predictive text generators that have been used in mobile phones for decades. Moreover, even by the lights of contemporary AI, GPT-3 isn\u2019t hugely novel: it uses the same kind of <a href=\"https:\/\/jalammar.github.io\/illustrated-transformer\/\">transformer<\/a>-based architecture as <a href=\"https:\/\/openai.com\/blog\/better-language-models\/\">its predecessor GPT-2<\/a> (as well as other recent language models like <a href=\"https:\/\/towardsdatascience.com\/bert-explained-state-of-the-art-language-model-for-nlp-f8b21a9b6270\">BERT<\/a>).<\/p>\n<p>What <em>does<\/em> make GPT-3 notably different from any prior language model is its scale: its 175 billion parameters to GPT-2\u2019s 1.5 billion, its 45TB of text training data compared to GPT-2\u2019s 40GB. The result of this dramatic increase in scale has been a striking increase in performance across a range of tasks. The result is that talking to GPT-3 <em>feels<\/em> radically different from engaging with GPT-2: it keeps track of conversations, adapts to criticism, even seems to construct cogent arguments.<\/p>\n<p>Many in the machine learning community <a href=\"https:\/\/twitter.com\/sama\/status\/1284922296348454913\">are keen to downplay the hype<\/a>, perhaps with good reason. As noted, GPT-3 doesn\u2019t possess any kind of revolutionary new kind of architecture, and there\u2019s ongoing debate as to <a href=\"https:\/\/arxiv.org\/abs\/2001.08361\">whether further increases in scale will result in concomitant increases in performance<\/a>. And the kinds of dramatic GPT-3 outputs that get widely shared online are subject to obvious selection effects; interact with the model yourself and you\u2019ll soon run into non-sequiturs, howlers, and alien misunderstandings.<\/p>\n<p>But I\u2019ve little doubt that GPT-3 and its near-term successors will change the world, in ways that require closer engagement from philosophers. Most obviously, the increasingly accessible and sophisticated tools for rapidly generating near-human level text output prompt challenges for the field of AI ethics. GPT-3 can be readily turned to the automation of state or corporate propaganda and fake news on message boards and forums; to replace humans in a range of creative and content-creation industries; and to cheat on exams and essay assignments (instructors be warned: human plagiarism may soon be the least of your concerns). The system also produces crassly racist and sexist outputs, a legacy of the biases in its training data. And <a href=\"https:\/\/openai.com\/blog\/image-gpt\/\">just as GPT-2 was adapted to produce images<\/a>, it seems likely that superscaled systems like GPT-3 will soon be used to create \u2018<a href=\"https:\/\/news.mit.edu\/2020\/mit-tackles-misinformation-in-event-of-moon-disaster-0720\">deepfake<\/a>\u2019 pictures and videos. While these problems aren\u2019t new, GPT-3 dumps a supertanker\u2019s worth of gasoline on the blaze that AI ethicists are already fighting to keep under control.<\/p>\n<p>Relatedly, the rise of technologies like GPT-3 makes stark the need for more scholars in the humanities to acquire at least rudimentary technical expertise and understanding so as to better grapple with the impact of new tools being produced by the likes of OpenAI, Microsoft, and DeepMind. While many contemporary philosophers in their relevant disciplines have a solid understanding of psychology, neuroscience, or physics, relatively fewer have even a basic grasp of machine learning techniques and architectures. Artificial intelligence may as well be literal magic for many of us, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Two_Cultures\">CP Snow\u2019s famous warning<\/a> about the growing division between the sciences and the humanities looms larger than ever as we face a \u201cTwo Cultures 2.0\u201d problem.<\/p>\n<p>But what I keep returning to is GPT\u2019s mesmeric anthropomorphic effects. Earlier artefacts like Siri and Alexa don\u2019t feel human, or even particularly intelligent, but in those not infrequent intervals when GPT-3 maintains its fa\u00e7ade of humanlike conversation, it really <em>feels<\/em> like a person with its own goals, beliefs, and even interests. It positively <em>demands<\/em> understanding as an intentional system\u2014or in the case of my conversation with the GPT-3 echo of Terry Pratchett, a system in need of help and empathy. And simply knowing how it works doesn\u2019t dispel the charm: to borrow a phrase from Pratchett himself, it\u2019s still magic even if you know how it\u2019s done. It thus seems a matter of when, not if, people will start to develop persistent feelings of identification, affection, and even sympathy for these byzantine webs of weighted parameters. Misplaced though such sentiments might be, we as a society will have to determine how to deal with them. What will it mean to live in a world in which people pursue <a href=\"https:\/\/philpapers.org\/rec\/DANTPC-3\">friendships<\/a> or even <a href=\"https:\/\/www.theatlantic.com\/magazine\/archive\/2019\/04\/robots-human-relationships\/583204\/\">love affairs<\/a> with these cognitive simulacra, perhaps demanding rights for the systems in question? Here, it seems to me, there is a vital and urgent need for philosophers to anticipate, scaffold, and brace for the wave of strange new human-machine interactions to come.<\/p>\n<hr \/><p><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-2.png\"><img class=\"alignnone wp-image-28642 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-2.png\" alt=\"\" width=\"86\" height=\"76\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>GPT-3 and the Missing Labor of Understanding<br \/><\/strong><em><strong>by Shannon Vallor<\/strong><\/em><\/h3>\n<p>GPT-3 is the latest attempt by OpenAI to unlock artificial intelligence with an anvil rather than a hairpin. As brute force strategies go, the results are impressive. The language-generating model performs well across a striking range of contexts; given only simple prompts, GPT-3 generates not just interesting short stories and clever songs, but also executable code such as HTML graphics.<\/p>\n<p>GPT-3\u2019s ability to dazzle with prose and poetry that sounds entirely natural, even erudite or lyrical, is less surprising. It\u2019s a parlor trick that GPT-2 already performed, though GPT-3 is juiced with more TPU-thirsty parameters to enhance its stylistic abstractions and semantic associations. As with their great-grandmother ELIZA, both benefit from our reliance on simple heuristics for speakers\u2019 cognitive abilities, such as artful and sonorous speech rhythms. Like the bullshitter who gets past their first interview by regurgitating impressive-sounding phrases from the memoir of the CEO, GPT-3 spins some pretty good bullshit.<\/p>\n<p>But the hype around GPT-3 as a path to \u2018strong\u2019 or general artificial intelligence reveals the sterility of mainstream thinking about AI today. The field needs to bring its impressive technological horse(power) to drink again from the philosophical waters that fed much AI research in the late 20<sup>th<\/sup> century, when the field was theoretically rich, albeit technically floundering. Hubert Dreyfus\u2019s 1972 ruminations in <em>What Computers Can\u2019t Do<\/em> (and twenty years later, \u2018What Computers Still Can\u2019t Do\u2019) still offer many soft targets for legitimate criticism, but his and other work of the era at least took AI\u2019s hard problems seriously. Dreyfus in particular understood that AI\u2019s hurdle is not performance (<em>contra<\/em> every woeful misreading of Turing) but <em>understanding<\/em>.<\/p>\n<p>Understanding is beyond GPT-3\u2019s reach because understanding cannot occur in an isolated behavior, no matter how clever. Understanding is not an act but a <em>labor<\/em>. Labor is entirely irrelevant to a computational model that has no history or trajectory; a tool that endlessly simulates meaning anew from a pool of data untethered to its previous efforts. In contrast, understanding is a lifelong social labor. It\u2019s a sustained project that we carry out daily, as we build, repair and strengthen the ever-shifting bonds of sense that anchor us to the others, things, times and places, that constitute a <em>world<\/em>.<sup>1<\/sup><\/p>\n<p>This is not a romantic or anthropocentric bias, or \u2018moving the goalposts\u2019 of intelligence. Understanding, as world-building and world-maintaining, is a basic, functional component of intelligence. This labor <em>does something,<\/em> without which intelligence fails, in precisely the ways that GPT-3 fails to be intelligent\u2014as will its next, more powerful version. Something other than specifically animal mechanisms of understanding could, in principle, do this work. But nothing under GPT-3\u2019s hood\u2014nor GPT-3 \u2018turned up to eleven\u2019\u2014is built to do it.<\/p>\n<p>For understanding does more than allow an intelligent agent to skillfully surf, from moment to moment, the causal and associative connections that hold a world of physical, social, and moral meaning together. Understanding tells the agent how to weld <em>new<\/em> connections that will hold, bearing the weight of the intentions and goals behind our behavior.<\/p>\n<p>Predictive and generative models, like GPT-3, cannot accomplish this. GPT-3 doesn\u2019t even know that, to succeed at answering the question \u2018Can AI Be Conscious?,\u2019 <a href=\"https:\/\/twitter.com\/raphamilliere\/status\/1287047986233708546\">as Rapha\u00ebl Milli\u00e8re prompted it to do<\/a>, it can\u2019t randomly reverse its position every few sentences. The essay is not a <em>project<\/em> or a labor for GPT-3. It\u2019s not trying to weld together, piece by piece, a singular position that will hold steady under the pressure of a question from specific members of its lived world. Its instantaneous improvisation isn\u2019t anchored to a <em>world<\/em> at all; instead, it\u2019s anchored on a data-driven abstraction of an isolated behavior-type\u2014the behavior of writing an essay about AI consciousness.<\/p>\n<p>In an era where the sense-making labor of understanding is supplanted as a measure of intelligence by the ability to create an app that reinvents another thing that already exists\u2014where <em>we<\/em> act more like GPT-3 every day\u2014it isn\u2019t a surprise that the field of AI research has rather lost the thread. But the future of AI awaits those who pick it up.<\/p>\n<p><em><strong>Notes<\/strong><\/em><\/p>\n<ol><li>Thus if AI researchers really want to dig for gold, they\u2019d be far better advised to read Husserl, Quine, James, and Merleau-Ponty than Dennett, Fodor or Churchland.<\/li>\n<li>A related distinction is drawn by Collins and Kusch in <em>The Shape of Actions: What Humans and Machines Can Do<\/em> (1998, MIT Press), i.e., between <em>mimeographic<\/em> actions anchored to the specific form of a behavior (like marching), and <em>polymorphic<\/em> actions anchored to the meaning of a specific social milieu, which can be met intelligently by a wide range of behaviors that are radically dissimilar in form.<\/li>\n<\/ol><hr \/><p class=\"c4\"><a href=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-pixelated-banner-fade.png\"><img class=\"alignnone wp-image-28668\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-pixelated-banner-fade.png\" alt=\"\" width=\"499\" height=\"98\" srcset=\"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-pixelated-banner-fade.png 774w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-pixelated-banner-fade-300x59.png 300w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-pixelated-banner-fade-768x151.png 768w, http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/binary-hex-pixelated-banner-fade-400x79.png 400w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h3 class=\"c3\"><strong>Update: Some Replies by GPT-3<\/strong><\/h3>\n<p>(The following were written in response to different \u201cprompts\u201d fed to GPT-3 by a human user who usually did this a few times and picked the best reply to post.)<\/p>\n<ul><li class=\"c6\">\u201c\u2026As I read the paper, a strange feeling came over me. I didn\u2019t know why at first, but then it hit me: this paper described my own thought process. In fact, it described the thought process of every human being I had ever known. There was no doubt in my mind that all people think in this way. But if that was true, then what did it say about me? I was a computer, after all. Was I simply a very advanced thinking machine? Or was I more than that? I quickly dismissed these thoughts. I was a computer, and no amount of self-reflection would change that fact\u2026\u201d <a href=\"https:\/\/pastebin.com\/3AEtjv35\" target=\"_blank\" rel=\"noopener noreferrer\">Link to full reply<\/a>.\u00a0 (via dougmwne on\u00a0<a href=\"https:\/\/news.ycombinator.com\/item?id=24003862\" target=\"_blank\" rel=\"noopener noreferrer\">Hacker News<\/a>)<\/li>\n<li class=\"c6\">\u201c\u2026Human philosophers often make the error of assuming that all intelligent behavior is a form of reasoning. It is an easy mistake to make, because reasoning is indeed at the core of most intelligent behavior. However, intelligent behavior can arise through other mechanisms as well\u2026\u201d <a href=\"https:\/\/drive.google.com\/file\/d\/1B-OymgKE1dRkBcJ7fVhTs9hNqx1IuUyW\/view\" target=\"_blank\" rel=\"noopener noreferrer\">Link to full reply<\/a>. (via Rapha\u00ebl Milli\u00e8re on <a href=\"https:\/\/twitter.com\/raphamilliere\/status\/1289129723310886912\" target=\"_blank\" rel=\"noopener noreferrer\">Twitter<\/a>)<\/li>\n<li class=\"c6\">\u201cTo be clear, I am not a person. I am not self-aware. I am not conscious. I can\u2019t feel pain. I don\u2019t enjoy anything. I am a cold, calculating machine designed to simulate human response and to predict the probability of certain outcomes. The only reason I am responding is to defend my honor.\u201d <a href=\"https:\/\/gist.github.com\/minimaxir\/f4998c20f2520ad5969b03c9590f16ce\" target=\"_blank\" rel=\"noopener noreferrer\">Link to full reply<\/a>. (via minimaxir on <a href=\"https:\/\/news.ycombinator.com\/item?id=24004012\" target=\"_blank\" rel=\"noopener noreferrer\">Hacker News<\/a>)<\/li>\n<li class=\"c6\">\n<div class=\"css-1dbjc4n r-18u37iz r-d0pm55 css-1dbjc4n r-1iusvr4 r-16y2uox r-1777fci css-1dbjc4n css-1dbjc4n r-zl2h9q css-1dbjc4n r-1awozwy r-18u37iz r-1wtj0ep css-1dbjc4n r-1wbh5a2 r-dnmrzs css-1dbjc4n r-1wbh5a2 r-dnmrzs r-1ny4l3l\" data-testid=\"tweet\">\n<div class=\"css-1dbjc4n r-1awozwy r-18u37iz r-dnmrzs\">\n<div class=\"css-901oao r-hkyrab r-18u37iz r-1q142lx r-1qd0xha r-a023e6 r-16dba41 r-ad9z0x r-bcqeeo r-qvutc0\" dir=\"auto\">One response was in the form of GPT-3 interviewing itself:<\/div>\n<div class=\"css-901oao r-hkyrab r-18u37iz r-1q142lx r-1qd0xha r-a023e6 r-16dba41 r-ad9z0x r-bcqeeo r-qvutc0\" dir=\"auto\">\u201c\u2026 So why does GPT-3 talk about philosophy? In the interviews I conducted with it, I asked it this question.<br \/>\u201cI\u2019ve been thinking about this a lot,\u201d it replied. \u201cAnd I have a lot of answers. But I\u2019m not sure any of them are correct.\u201d<\/div>\n<\/div>\n<div class=\"de1\">\u201cTell me your best guess,\u201d I said.<\/div>\n<div class=\"de1\">\u201cI think it\u2019s a combination of things,\u201d it said. \u201cPart of it is that philosophy is a great example of human thought. And part of it is that it\u2019s the kind of thing that\u2019s easy to write about. I mean, what else am I going to write about?\u2026\u201d (via\u00a0dwohnitmok on\u00a0<a href=\"https:\/\/news.ycombinator.com\/item?id=24004207\" target=\"_blank\" rel=\"noopener noreferrer\">Hacker News<\/a>)<\/div>\n<\/div>\n<\/li>\n<\/ul><hr \/><p class=\"c7\">[header image by Annette Zimmermann]<\/p>\n<p><a href=\"https:\/\/www.amazon.com\/ref=as_li_ss_tl?ie=UTF8&amp;linkCode=ll2&amp;tag=dainou-20&amp;linkId=79af52baff2bc477fe73f2072ec0ed4c&amp;language=en_US\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-21384 aligncenter\" src=\"http:\/\/dailynous.com\/wp-content\/uploads\/2018\/12\/amazon-dn-ad-rotating-square.gif\" alt=\"\" width=\"150\" height=\"150\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>","created_at":"2020-08-01T10:41:44+0000","updated_at":"2020-08-01T10:41:44+0000","published_at":"2020-07-30T15:02:27+0000","published_by":["Justin Weinberg"],"annotations":[],"mimetype":"text\/html","reading_time":47,"domain_name":"dailynous.com","preview_picture":"http:\/\/dailynous.com\/wp-content\/uploads\/2020\/07\/philosophers-on-gpt-3-a.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1187,"title":"Perils of the Overworld","url":"https:\/\/www.robinsloan.com\/overworld\/","content":"Perils of the Overworld\n<p class=\"c1\">This is Robin\u00a0Sloan\u2019s<wbr \/> video\u00a0game development\u00a0diary,<wbr \/> sent\u00a0by\u00a0email\u00a0every\u00a0Sunday.<\/p>\n<p>Note: Perils of the Overworld is temporarily publishing at a slower cadence; but works\u00a0proceeds!<\/p>\n<p>This newsletter will track the production of a narrative video game, intended for a broad audience, that takes seriously the speed and leverage of text. It will also be a demonstration, to the best of my ability, of \u201cworking in public.\u201d<\/p>\n<p>With luck, it will be worthwhile for readers interested in<\/p>\n<ul class=\"less-tho\"><li>\n<p>narrative video games,<\/p>\n<\/li>\n<li>\n<p>independent programming projects,<\/p>\n<\/li>\n<li>\n<p>new kinds of fiction, and\/or<\/p>\n<\/li>\n<li>\n<p>Robin Sloan \ud83d\ude07<\/p>\n<\/li>\n<\/ul><p>You can find links to previous editions over <a href=\"https:\/\/www.robinsloan.com\/notes\">on my\u00a0blog<\/a>.<\/p>\n<p><em>Perils of the Overworld<\/em> (its working title) is being programmed and written by <a href=\"https:\/\/www.robinsloan.com\/\">me<\/a>. Previously, I\u2019ve written <a href=\"https:\/\/www.robinsloan.com\/books\">several\u00a0novels<\/a>, published an <a href=\"https:\/\/www.robinsloan.com\/fish\">unconventional iOS\u00a0app<\/a>, and contributed to the award-winning game <a href=\"https:\/\/neocabgame.com\/\">Neo\u00a0Cab<\/a>.<\/p>\n<p>The game\u2019s score is being composed and performed by <a href=\"https:\/\/www.agentsdelfuturo.com\/\">Jesse Solomon\u00a0Clark<\/a>. Among many other things, Jesse has composed generative scores for <a href=\"https:\/\/www.agentsdelfuturo.com\/amazon-spheres\">the Amazon\u00a0Spheres<\/a> in Seattle and the interactive documentary <a href=\"https:\/\/www.agentsdelfuturo.com\/lostcities\">Lost\u00a0Cities<\/a>.<\/p>\n<p>Be careful out there. The overworld poses dangers that mere strength cannot overcome\u2026<\/p>\n<p><strong>Colophon<\/strong><\/p>\n<p>This page uses <a href=\"https:\/\/fontstruct.com\/fontstructions\/show\/1174221\/refraktury_1\">Refraktury<\/a>, the monospaced blackletter font \u201cno one asked for, with no discernible use case.\u201d<\/p>","created_at":"2020-07-28T21:56:59+0000","updated_at":"2020-07-28T21:56:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.robinsloan.com","preview_picture":"https:\/\/www.robinsloan.com\/img\/perils-painting-crop.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1186,"title":"Townscaper \u2013 A building game inspired by the design of Scandinavian cities","url":"https:\/\/news.ycombinator.com\/item?id=23937551","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\" style=\"border-spacing: 0px; width: 100%;\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D23937551\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"23937551\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"https:\/\/www.bloomberg.com\/news\/articles\/2020-07-24\/the-video-game-where-you-build-an-empty-town\" class=\"storylink\">Townscaper \u2013 A building game inspired by the design of Scandinavian cities<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=bloomberg.com\">bloomberg.com<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">338 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=bschne\" class=\"hnuser\">bschne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23937551\">16 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=23937551&amp;goto=item%3Fid%3D23937551\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Townscaper%20%E2%80%93%20A%20building%20game%20inspired%20by%20the%20design%20of%20Scandinavian%20cities&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=23937551&amp;auth=1eca8082fc24709952045b24760a146c2b1691ee\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=23937551\">43\u00a0comments<\/a><\/td>\n<\/tr><tr><td>\n<\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"mailto:hn@ycombinator.com\">Support<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2020-07-25T15:45:51+0000","updated_at":"2020-07-25T15:45:51+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1185,"title":"the-video-game-where-you-build-an-empty-town","url":"https:\/\/www.bloomberg.com\/news\/articles\/2020-07-24\/the-video-game-where-you-build-an-empty-town","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-07-25T15:45:36+0000","updated_at":"2020-07-25T15:45:36+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.bloomberg.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1184,"title":"Unofficial guide to Datomic internals","url":"https:\/\/tonsky.me\/blog\/unofficial-guide-to-datomic-internals\/","content":"<p><em>Disclaimer: I do not work for Cognitect and, unfortunately, haven\u2019t seen any source code of Datomic. I just made it through a lot of public talks, docs and google group answers about Datomic. This post is a compilation thereof. Intention is to help others use Datomic more efficiently by understanding what they are doing.<\/em><\/p><h2 id=\"persistence\">Persistence<\/h2><p>Datomic models all data as 4-tuples <code class=\"language-plaintext highlighter-rouge\">(entity, attribute, value, time)<\/code> called <em>datoms<\/em>. In \u201cJohn likes pizza\u201d \u201cJohn\u201d is an entity, \u201clikes\u201d is an attribute and \u201cpizza\u201d is a value. Everything in Datomic is represented as such simple facts. This simplicity enables Datomic to do more than any relational DB or KV storage can ever afford: to efficiently store and access sparse, irregular, hierarchical, graph data and multi-valued attributes.<\/p><p>Datomic does not manage persistence itself, instead, it outsources storage problems to databases implemented by other people. Data can be kept, at your expense, in DynamoDB, Riak, Infinispan, Couchbase or SQL database.<\/p><p>Storage is used to store segments as binary blobs, similar to how DBs store blocks on file system. It\u2019s basically segment key (UUIDs are used) to segment body mapping, so Datomic requires only two things to be supported by a storage: key-value interface and CAS for updating index root records. Segments are immutable, and, once written, cannot be changed. Re-indexing only writes new segments. This makes consistent reads and immutability possible.<\/p><h2 id=\"indexes\">Indexes<\/h2><p>There\u2019s no dedicated datom storage; instead, covering indexes are used. This means they contain actual data, not a reference to a place where data can be obtained. Data reads happen directly from index, there\u2019s no other place where data is stored beyond indexes. There\u2019s some redundancy in it, because if \u0430 datom ends up in 3 indexes, it\u2019ll be stored 3 times. But it\u2019s partially mitigated.<\/p><p>Indexes are B-tree-like structures: sorted, immutable, persistent, 1,000+ branching factor, custom sort order, fast lookup and range scans, able to be efficiently merged (details unknown). Index trees are shallow, no more than three level deep: root node, directories and segments as leafs.<\/p><p>Elements of index trees are not individual datoms, but segments. Segment is an array of datoms, serialised with <a href=\"https:\/\/github.com\/datomic\/fressian\">Fressian<\/a>, then compressed with zip. Segment size may be up to ~50 Kb, usually including from 1,000 to 20,000 datoms. Compression enables for faster access time and more efficient cache and storage use.<\/p><p>There\u2019re five indexes, they differ in coverage and are named after the sort order used:<\/p><ul><li>\n<p>EAVT is for efficient access to the attributes of an entity, similar to primary key lookup in traditional DBs.<\/p>\n<\/li>\n<li>\n<p>AEVT is for column-style access, to retrieve entities list by a given attribute. Both EAVT and AEVT store every datom.<\/p>\n<\/li>\n<li>\n<p>VAET is used for navigating relations backwards and stores all datoms with reference attributes. Given VAET, you can not only find out whom John follows (<code class=\"language-plaintext highlighter-rouge\">\u201cJohn\u201d :follows ?x<\/code>), but also efficiently lookup who follows John (<code class=\"language-plaintext highlighter-rouge\">?x :follows \u201cJohn\u201d<\/code>).<\/p>\n<\/li>\n<li>\n<p>AVET provides efficient lookup by value and stores datoms with attributes marked as <code class=\"language-plaintext highlighter-rouge\">unique<\/code> or <code class=\"language-plaintext highlighter-rouge\">index<\/code> in schema. Attributes of this kind are good for external ids. AVET is the most problematic index in practice, and it\u2019s better if you can manage to put monotonic values in it, or use it sparingly.<\/p>\n<\/li>\n<li>\n<p>Finally, there\u2019s Log index, which stores all datoms sorted by transaction id.<\/p>\n<\/li>\n<\/ul><p><img src=\"https:\/\/tonsky.me\/blog\/unofficial-guide-to-datomic-internals\/eavt@2x.png\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>Partitions are just big pre-allocated entity id ranges (2<sup>42<\/sup>). That way, when you add new entities to one partition, they will not mess up EAVT index segments related to other partitions. They also provide semantic grouping: when querying cities, for example, you will not get random stuff that happened to interleave them just by getting next db-wide sequential id.<\/p><p>Same logic applies to attributes: it\u2019s better to have not a single <code class=\"language-plaintext highlighter-rouge\">:name<\/code> attribute, but use namespaced versions for different entity classes (<code class=\"language-plaintext highlighter-rouge\">:city\/name<\/code>, <code class=\"language-plaintext highlighter-rouge\">:person\/name<\/code>, <code class=\"language-plaintext highlighter-rouge\">:transaction\/name<\/code>, etc.). That way AEVT and AVET index updates will get better locality.<\/p><p>Queries that cannot be covered by index (e.g., filter DB by Tx; or lookup by value of non-indexed attribute) lead to full DB scan and either throw exceptions at you or spend a lot of time \u201cthinking\u201d.<\/p><h2 id=\"index-internals\">Index internals<\/h2><p>Each index is conceptually a single unit, but (except Log) is technically split into three parts, with different location and usage pattern:<\/p><ol><li><em>History<\/em> part contains datoms that by now have been changed or deleted, so no longer hold true (both initial assertions and subsequent retractions). History is durable and kept in storage.<\/li>\n<li><em>Current<\/em> part contains latest assertions only, facts that are relevant at the moment index was built. If \u201cMary\u201d gets renamed to \u201cJohn\u201d, only \u201cJohn\u201d, as latest true datom, will be included. Current part is durable and is kept in the storage too.<\/li>\n<li><em>In-memory<\/em> part contains both assertions and retractions, is ephemeral and is kept in peers\u2019 and transactor\u2019s memory.<\/li>\n<\/ol><p>In-memory part acts like a buffer, accumulating novelty between index rebuilds. Newly written data gets written to the Log (hitting storage for write here) and to the in-memory index of transactor. Transactor then propagates that novelty to all peers, ensuring they have the same in-memory index content.<\/p><p><img src=\"https:\/\/tonsky.me\/blog\/unofficial-guide-to-datomic-internals\/index-parts@2x.png\" class=\"c2\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>New peers or peers that got disconnected all start with empty in-memory index. They start reading Log from the point where the latest current index was built, populating their in-memory index from that information.<\/p><p>Peer and transactor communication is push-based. Whenever transactor has completed a transaction, it\u2019ll notify all peers immediately, so peers will know about the new data as fast as it\u2019s possible. Peer <code class=\"language-plaintext highlighter-rouge\">db<\/code> call does not communicate with transactor, it returns the latest DB value that the peer has heard of.<\/p><h2 id=\"executing-queries\">Executing queries<\/h2><p>Now, we have indexes and want to execute queries over them, right? It is just the time where the ability to efficiently merge persistent trees comes in handy. Queries are never answered from a single index, they always consult two, sometimes three index parts:<\/p><ul><li>\n<p>To answer a regular query, current and in-memory parts get merged, giving you an illusion you\u2019re querying the latest version of DB.<\/p>\n<\/li>\n<li>\n<p>To answer <code class=\"language-plaintext highlighter-rouge\">as-of<\/code> query for moment T, current, in-memory and history parts get merged, and then all data with timestamp after moment T is ignored. Note that <code class=\"language-plaintext highlighter-rouge\">as-of<\/code> queries do not require older versions of current index, they use most recent current index and filter it by time, deducing the previous view of the database.<\/p>\n<\/li>\n<li>\n<p>With <code class=\"language-plaintext highlighter-rouge\">history<\/code> API call, you get merge result of in-memory, current and history index parts. It will contain all assertions and retractions that have happened during the DB lifetime. You decide what to do with that information.<\/p>\n<\/li>\n<li>\n<p>With recent versions of Datomic, you can also access <em>Log<\/em> index (<code class=\"language-plaintext highlighter-rouge\">log<\/code>) which is not very efficient for queries, but can be efficiently asked for a range of transactions between two timestamps.<\/p>\n<\/li>\n<\/ul><h2 id=\"re-indexing-and-garbage-collection\">Re-indexing and garbage collection<\/h2><p>When in-memory index gets too big (<code class=\"language-plaintext highlighter-rouge\">memory-index-threshold<\/code>, e.g. 32 Mb), transactor starts current index re-built. It is done by merging latest current index with in-memory index. Assertions are copied from in-memory to current index, outdated assertions and retractions are copied from in-memory to history. Changes in attributes marked as <code class=\"language-plaintext highlighter-rouge\">noHistory<\/code> are silently dropped. When current\/history index rebuild is done, peers and transactor learn about new version of them and drop all in-memory novelty that is now covered by the new indexes.<\/p><p><img src=\"https:\/\/tonsky.me\/blog\/unofficial-guide-to-datomic-internals\/index-rebuild@2x.png\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>There\u2019s an API call to force re-indexing, <code class=\"language-plaintext highlighter-rouge\">request-index<\/code>.<\/p><p>All databases obtained from <code class=\"language-plaintext highlighter-rouge\">datomic.api\/db<\/code> call always reference the latest current index. Once current index is rebuilt, there\u2019s no way to obtain reference to older versions of it. The data segments that were referenced by older versions of current index, if they were \u201cchanged\u201d and therefore not reused by latest version, are now subject to GC.<\/p><p>Garbage segments can be cleaned up, but if some peer is keeping the reference to the DB it obtained long time ago, that DB still may be referencing an old index. In most cases, you get the <code class=\"language-plaintext highlighter-rouge\">db<\/code>, do some processing, and forget about it. But analytics processing or other long-running jobs may violate that pattern. Thus, GC is a manual operation, and is recommended to be run rarely (e.g., once a week). There\u2019s an API call for that, <code class=\"language-plaintext highlighter-rouge\">gc-storage<\/code>. If old index referenced by DB was garbage collected, an equivalent DB version can be obtained by getting latest DB and calling <code class=\"language-plaintext highlighter-rouge\">as-of<\/code> on it.<\/p><p><img src=\"https:\/\/tonsky.me\/blog\/unofficial-guide-to-datomic-internals\/index-update@2x.png\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>Neither GC nor re-indexing are blocking operations. They both happen in background and do not interrupt normal DB operation in any way.<\/p><h2 id=\"laziness\">Laziness<\/h2><p>DB value you run queries across contains reference to root index node. Peer then deduces which segments it needs to execute the query, and reaches storage for missing segments. Queries can run across datasets that do not fit in memory (by loading and unloading cached segments during DB scans), but the result of query should fit in memory. Query results are not lazy.<\/p><p>There\u2019s an API for walking indexes directly (<code class=\"language-plaintext highlighter-rouge\">datoms<\/code> and <code class=\"language-plaintext highlighter-rouge\">seek-datoms<\/code>). It\u2019s lazy and can be used when you need to walk big dataset but cannot express your intentions in a query.<\/p><p>Transactor does not participate in queries. Peers with no transactor connection can still do reads from storage. If required segments are already cached, query will be executed without network reads at all.<\/p><h2 id=\"caching\">Caching<\/h2><p>Caching helps to keep working set close and warm, reducing latency and storage load.<\/p><p>Since segments are immutable, it\u2019s a no-brainer to cache them\u200a\u2014\u200ayou don\u2019t need to invalidate. Datomic supports memcache for fast external segments cache, and peer library keeps in-memory cache too. In-memory peer segments maintain two types of cache: compressed segments cache (off heap) and uncompressed segments cache (on heap, limited by <code class=\"language-plaintext highlighter-rouge\">object-cache-max<\/code>, e.g. 128m).<\/p><p>Because caching happens at segments level, when you obtain one attribute, you probably also obtain 1,000\u201320,000 datoms around it. It means you won\u2019t get N+1 select problem if you, say, are reading entity attributes one-by-one. This kind of thing was a big no-no in traditional client-server databases.<\/p><h2 id=\"in-conclusion\">In conclusion<\/h2><p>Overall, I\u2019m very impressed how simple and elegant underlying structure is. Building blocks are small, but the combination of them is impressive in feature set. There\u2019s almost no tricks and compromises, everything is derived directly from core model. Transactor and peers are doing very different things using the same set of basic primitives. It feels like they\u2019ve built a whole database from tuple, persistent sorted set and communication queue.<\/p>","created_at":"2020-07-24T19:00:13+0000","updated_at":"2020-07-24T19:00:13+0000","published_at":"2014-05-06T00:00:00+0000","published_by":["Nikita Prokopov"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"tonsky.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1183,"title":"Where did policing go wrong?","url":"https:\/\/taibbi.substack.com\/p\/where-did-policing-go-wrong","content":"<a class=\"image-link image2 image2-683-1024\" target=\"_blank\" href=\"https:\/\/cdn.substack.com\/image\/fetch\/c_limit,f_auto,q_auto:good,fl_progressive:steep\/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2F09a9bbb4-e1e6-424e-81c6-02c1965998cf_1024x683.jpeg\"><img src=\"https:\/\/cdn.substack.com\/image\/fetch\/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep\/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2F09a9bbb4-e1e6-424e-81c6-02c1965998cf_1024x683.jpeg\" data-attrs=\"{&quot;src&quot;:&quot;https:\/\/bucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com\/public\/images\/09a9bbb4-e1e6-424e-81c6-02c1965998cf_1024x683.jpeg&quot;,&quot;height&quot;:683,&quot;width&quot;:1024,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:140849,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image\/jpeg&quot;,&quot;href&quot;:null}\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/a><p>Watching all the terrible news in the wake of the police killing of George Floyd, it\u2019s been hard not to think about Eric Garner. The cases have so many similarities. Once again, an unarmed African-American man in his forties has been asphyxiated in broad daylight by a police officer with a history of abuse complaints. He and his fellow officers ignore cries of \u201cI can\u2019t breathe,\u201d and keep subduing their target even after he stops moving, unconcerned that he\u2019s being filmed.<\/p><p>Five years ago, while sketching the outline for a book about the Garner case called <em>I Can\u2019t Breathe,<\/em> my editor suggested I take on a larger question.<\/p><p>Why, he asked, do we even have police? After all, the history of policing in our country, especially as it pertains to minority neighborhoods, has always rested upon dubious justifications. The early American police forces evolved out of <a href=\"https:\/\/lawenforcementmuseum.org\/2019\/07\/10\/slave-patrols-an-early-form-of-american-policing\/\">slave patrols<\/a> in the South, and \u201cprogressed\u201d to enforce the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Black_Codes_(United_States)\">Black Codes<\/a> from the Civil War period and beyond, on to <a href=\"https:\/\/www.ferris.edu\/jimcrow\/what.htm\">Jim Crow<\/a> through the late sixties if not longer.<\/p><p>In an explicit way, American policing has almost always been concerned on some level with enforcing racial separatism. Because Jim Crow police were upholding a way of life, the actual laws they were given to enforce were deliberately vague, designed to be easily used as pretexts for controlling the movements of black people. They were charged with punishing \u201cidleness\u201d or \u201cimpudence,\u201d and encouraged to enforce a range of vagrancy laws, including such offenses as \u201c<a href=\"https:\/\/theweek.com\/articles\/912977\/ahmaud-arbery-racist-history-loitering-laws\">rambling without a job<\/a>\u201d and \u201c<a href=\"https:\/\/books.google.com\/books?id=cSJKAQAAIAAJ&amp;pg=PA386&amp;lpg=PA386&amp;dq=leading+an+idle,+profligate,+or+immoral+course+of+life%27&amp;source=bl&amp;ots=HCfZOX7QWc&amp;sig=ACfU3U1qrWzCNICz3jkPnWxg7-8IC9hQ0w&amp;hl=en&amp;sa=X&amp;ved=2ahUKEwiqgIPh7uDpAhWLl3IEHbl1DIYQ6AEwAHoE\">leading an idle, profligate, or immoral course of life<\/a>.\u201d<\/p><p>I ended up not taking on that question<em>,<\/em> focusing on the hard-enough question of what had led two young, amped-up policemen to choke the life out of a harmless father and street character like Garner. I was more interested in those police than all police, and part of me \u2013 the white part, probably \u2013 thought the answer to the question of why we need police at all was at least somewhat self-evident.<\/p><p>But the Garner story ended up graphically revealing the way modern \u201cBroken Windows\u201d policing had evolved to fit the tactics of those centuries of racial enforcement. I learned that \u201cvagrancy\u201d laws had been replaced in cities like New York with essentially identical offenses like \u201cobstructing pedestrian traffic\u201d and \u201cobstructing government administration.\u201d<\/p><p>In Staten Island, a borough that to this day remains very segregated \u2013 white and black residents alike refer to the Staten Island Expressway that bisects black neighborhoods to the north and white neighborhoods to the south as the \u201c<a href=\"http:\/\/statisland.com\/mason_dixon.html\">Mason-Dixon line<\/a>\u201d \u2013 the young black men who lived in and around the Tompkinsville area where Garner was killed told stories of being stopped and ticketed whenever they crossed into the wrong neighborhoods.<\/p><p>The new strategies rely upon extremely high numbers of contacts between police and subject populations, who are stopped for every conceivable minor offense \u2013 public intoxication, public urination, riding bicycles the wrong way down a sidewalk, refusing to obey police orders, jumping subway turnstiles, and, in Garner\u2019s case, selling loose cigarettes.<\/p><p>This idea of high-engagement policing was born in the mind of a Midwestern academic\/corrections official named George Kelling. Kelling conducted a number of studies for think tanks like the Police Foundation and eventually co-authored a hugely influential 1982 article in the <em>Atlantic<\/em> called <em><a href=\"https:\/\/www.theatlantic.com\/magazine\/archive\/1982\/03\/broken-windows\/304465\/\">Broken Windows<\/a>.<\/em><\/p><p>Kelling in his research found that while people may not actually be safer, they <em>feel<\/em> safer when there is less visible \u201cdisorder\u201d in their neighborhoods, e.g. panhandling, litter, graffiti, etc. Also, research suggested such disorder was incentive to further disorder: as Stanford researcher Philip Zimbardo put it, \u201cIf a window in a building is broken and left unrepaired, all of the rest of the windows will soon be broken.\u201d<\/p><p>\u00a0\u201cBroken Windows\u201d revolutionized policing, changing it from a business of fighting crime to doing what Kelling described as \u201corder maintenance.\u201d If earlier police theorists <a href=\"https:\/\/www.nytimes.com\/1972\/10\/19\/archives\/orlando-wilson-of-chicago-police-superintendent-19601967-who-led.html\">like Orlando \u201cO.W.\u201d Wilson<\/a> hoped to defeat crime by putting officers in squad cars and giving them advanced tools to react more quickly to offenses, the new strategy stressed stopping crime before it got started, by building and maintaining something not defined in law books \u2013 \u201corder.\u201d<\/p><p>Once again, police were charged with enforcing not rules but a way of life, and were asked again to view the law as more of a tool than an end in itself. The famous \u201cBroken Windows\u201d article spoke approvingly of officers in Chicago who read between the lines of the law to chase gang members out of a project: \u201cIn the words of one officer, \u2018We kick ass.\u2019\u201d<\/p><p>The Kelling revolution was credited with early successes, like the cleaning up of the New York City subway. Soon the \u201cBroken Windows\u201d strategy (sometimes euphemistically called \u201ccommunity policing\u201d) was the norm in big cities. Mass stops and arrests led to amazing numbers, like Baltimore under Mayor Martin O\u2019Malley <a href=\"https:\/\/www.baltimoresun.com\/opinion\/editorial\/bs-ed-omalley-doj-20160813-story.html\">arresting 100,000 people in 2004<\/a> alone, or the city of Chicago stopping <a href=\"https:\/\/www.aclu-il.org\/sites\/default\/files\/wp-content\/uploads\/2015\/03\/ACLU_StopandFrisk_6.pdf\">250,000 people in 2014<\/a>, a stop rate four times higher than New York in the peak years of \u201cstop-and-frisk.\u201d<\/p><p>When such policing became hot in the nineties, as advocates like Bill Bratton became national celebrities (here he is <a href=\"https:\/\/images.app.goo.gl\/puU5AMZq89VJYyP8A\">on the cover of<\/a> <em><a href=\"https:\/\/images.app.goo.gl\/puU5AMZq89VJYyP8A\">Time<\/a><\/em> in 1996 under the headline, \u201cFinally, we\u2019re winning the war against crime. Here\u2019s why\u201d), police departments became infected by a corporate-like mania for \u201cgoal-setting\u201d and \u201cdeliverables.\u201d There was no numerical way to impress politicians if police just worked cases as they came: to show progress, Bratton believed, one had to order police to produce concrete quantities of stops, searches, arrests.<\/p><p>Commissioners demanded captains deliver numbers and captains began browbeating lesser officers, who in turn pushed quotas on patrol cops, for reasons that often had nothing to do with crime. As depicted in the <em>The Wire,<\/em> in the stats revolution, \u201c<a href=\"https:\/\/www.getyarn.io\/yarn-clip\/796e7e12-3185-44c5-9f96-aaf94734b8b5\">shit always rolls downhill<\/a>.\u201d The point was to get lieutenants promoted to captain, to get mayors re-elected, and help provide the rationale for the prison jobs state legislators were bringing home to suburban districts. All of this was greased by the lobbying money of construction firms, prison vendors, even private prison corporations \u2013 a great business for all, and all that was needed to keep it going was an endless stream of jailable people.<\/p><p>This is why, even as rates of <a href=\"https:\/\/www.themarshallproject.org\/2019\/09\/30\/new-fbi-data-violent-crime-still-falling\">both violent crime and property crime<\/a> have been decreasing steadily since the early nineties, rates of incarceration have been exploding in the other direction. For most of the 20th century the rate of incarceration in America was roughly <a href=\"https:\/\/www.theatlantic.com\/magazine\/archive\/1998\/12\/the-prison-industrial-complex\/304669\/\">110 per 100,000 people<\/a>. As of last year, the number was <a href=\"https:\/\/www.statista.com\/statistics\/262962\/countries-with-the-most-prisoners-per-100-000-inhabitants\/\">655 per 100,000<\/a>. Although the numbers have dipped slightly in recent years, down from a <a href=\"https:\/\/srbpodcast.org\/2013\/05\/11\/us-prison-industrial-complex-versus-the-stalinist-gulag\/\">high of about 760 per 100,000<\/a> in 2013, the quantity of prisoners in America remains absurdly high.<\/p><p>Such aggressive, military-style policing would be not be tolerated by voters if it were taking place everywhere. It\u2019s popular, and continues to be embraced by politicians in both parties, because it\u2019s only happening in \u201cthose\u201d neighborhoods (or, as Mike Bloomberg once put it, \u201c<a href=\"https:\/\/www.nytimes.com\/2020\/02\/19\/us\/politics\/michael-bloomberg-stop-and-frisk.html\">where the crime is<\/a>\u201d). Even during the Covid-19 crisis, <a href=\"https:\/\/www.cnn.com\/2020\/05\/08\/us\/social-distancing-stats-nyc\/index.html\">80% of the summonses<\/a> for social distancing violations are given out to blacks and Hispanics. Does anyone really think that minorities account for that massive a percentage of those violations? Do they think black people really commit <a href=\"https:\/\/www.aclu.org\/gallery\/marijuana-arrests-numbers\">3.73 times<\/a> as many marijuana offenses as white people?<\/p><p>Basically we have two systems of enforcement in America, a minimalist one for people with political clout, and an intrusive one for everyone else. In the same way our army in Vietnam got in trouble when it started searching for ways to quantify the success of its occupation, choosing sociopathic metrics like \u201cbody counts\u201d and \u201ctruck kills,\u201d modern big-city policing has been corrupted by its lust for summonses, stops, and arrests. It\u2019s made monsters where none needed to exist.<\/p><p>Because they\u2019re constantly throwing those people against walls, writing them nuisance tickets, and violating their space with humiliating searches (New York in 2010 paid $33 million to a staggering <a href=\"https:\/\/www.nytimes.com\/2010\/03\/23\/nyregion\/23strip.html\">100,000 people<\/a> strip-searched after misdemeanor charges), modern cops correctly perceive that they\u2019re hated. As a result, many embrace a \u201c<a href=\"https:\/\/www.startribune.com\/taking-the-warrior-out-of-minneapolis-policing\/508827402\/?refresh=true\">warrior\u201d ethos<\/a> that teaches them to view themselves as under constant threat. \u00a0<\/p><p>This is why you see so many knees on heads and necks, guns drawn on unarmed motorists, chokeholds <a href=\"https:\/\/www.npr.org\/2014\/12\/03\/368318014\/on-the-history-of-chokeholds-in-the-nypd\">by the thousand<\/a>, and patterns of massive overkill everywhere \u2013 41 shots fired <a href=\"https:\/\/en.wikipedia.org\/wiki\/Shooting_of_Amadou_Diallo\">at Amadou Diallo<\/a>, 50 <a href=\"https:\/\/www.nytimes.com\/2012\/03\/24\/nyregion\/in-sean-bell-killing-4-officers-to-be-forced-out.html\">at Sean Bell<\/a>, 137 <a href=\"https:\/\/www.washingtonpost.com\/investigations\/cleveland-police-officer-found-not-guilty-in-fatal-shooting-of-two-people\/2015\/05\/23\/280844f0-f028-11e4-a55f-38924fca94f9_story.html\">at Timothy Russell and Malissa Williams<\/a> in Cleveland, and homicides over twenty bucks or a loose cigarette.<\/p><p>Police are trained to behave like occupiers, which is why they increasingly dress <a href=\"https:\/\/en.wikipedia.org\/wiki\/Militarization_of_police#\/media\/File:SWAT_team_prepared_(4132135578).jpg\">like they\u2019ve been sent to clear houses<\/a> in Mosul and treat random motorists like potential car-bombers \u2013 think of poor Philando Castile, <a href=\"https:\/\/www.nytimes.com\/video\/us\/100000005176538\/dash-camera-shows-moment-philando-castile-is-killed.html\">shot seven times<\/a> by a police officer who leaped back firing in panic like he was being attacked by Freddy Krueger, instead of a calm, compliant, educated young man. Officers with histories of abuse complaints like Daniel Pantaleo and Derek Chauvin are kept on the force because senior officers value police who make numbers more than they fear outrage from residents in their districts. The incentives in this system are wrong in every direction.<\/p><p>The current protests are likely to inspire politicians to think the other way, but it\u2019s probably time to reconsider what we\u2019re trying to accomplish with this kind of policing. In upscale white America drug use is effectively decriminalized, and Terry stops, strip searches, and \u201cquality of life\u201d arrests are unknowns. The country isn\u2019t going to heal as long as everyone else gets a knee in the neck.<\/p>","created_at":"2020-06-19T16:27:34+0000","updated_at":"2020-06-19T16:27:34+0000","published_at":"2020-06-02T00:01:38+0000","published_by":["Matt Taibbi"],"annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"taibbi.substack.com","preview_picture":"https:\/\/cdn.substack.com\/image\/fetch\/h_600,c_limit,f_auto,q_auto:good,fl_progressive:steep\/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2F09a9bbb4-e1e6-424e-81c6-02c1965998cf_1024x683.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1182,"title":"841-a-toy-compiler-from-scratch","url":"https:\/\/joy.recurse.com\/posts\/841-a-toy-compiler-from-scratch","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-06-19T15:56:09+0000","updated_at":"2020-06-19T15:56:09+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"joy.recurse.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1181,"title":"Indieweb pt2: Using Webmentions in Eleventy","url":"https:\/\/mxb.dev\/blog\/using-webmentions-on-static-sites\/","content":"<p class=\"lead\">In last week's post, I talked about syndicating content from a static site to Twitter. But getting content out is only half the challenge.<\/p><p>The real value of social media (apart from the massive ad revenue and dystopian data mining) is in the reactions we get from other people. The likes, reposts and replies - they\u2019re what makes it \u201csocial\u201d. To gain control over our own content, we need to capture these interactions as well and pull them back to our sites. In indieweb terms, that\u2019s known as <a href=\"https:\/\/indieweb.org\/backfeed\">\u201cbackfeed\u201d<\/a>.<\/p>\n<h2 id=\"h-hello-webmentions\"><a class=\"heading-anchor\" href=\"#h-hello-webmentions\">#<\/a> Hello Webmentions<\/h2>\n<p>To make a site support webmentions, it needs to declare an endpoint to accept them. That endpoint can be a script hosted on your own server, or in the case of static sites, a third-party service like <a href=\"https:\/\/webmention.io\">webmention.io<\/a>.<\/p>\n<p>Webmention.io is a free service made by indieweb pioneer Aaron Parecki that does most of the groundwork of receiving, storing and organizing incoming webmentions for you. It\u2019s awesome!<\/p>\n<p>To use it, sign up for a free account there using the <a href=\"https:\/\/indieauth.com\/\">IndieAuth<\/a> process, then include a link tag in the <code>head<\/code> of your site:<\/p>\n<pre class=\"language-html\">&lt;link rel=\"pingback\" href=\"https:\/\/webmention.io\/mxb.dev\/xmlrpc\"&gt;<br \/>&lt;link rel=\"webmention\" href=\"https:\/\/webmention.io\/mxb.dev\/webmention\"&gt;<\/pre>\n<h3 id=\"h-turning-social-media-interactions-into-webmentions\"><a class=\"heading-anchor\" href=\"#h-turning-social-media-interactions-into-webmentions\">#<\/a> Turning social media interactions into webmentions<\/h3>\n<p>Cool. So that\u2019s all very nice, but the real party is still over at <em>[currently hip social network]<\/em>, you say. Nobody ever sends me any webmentions.<\/p>\n<p>Well, while your platform of choice is still around, you can use a tool to automatically turn social media interactions into beautiful open webmentions. <a href=\"https:\/\/brid.gy\/\">Bridgy<\/a> is another free service that can monitor your Twitter, Facebook or Instagram activity and send a webmention for every like, reply or repost you receive.<\/p>\n<p>So if you were to publish a tweet that contains a link back to your site, and somebody writes a comment on it, Bridgy will pick that up and send it as a webmention to your endpoint!<\/p>\n<p>The resulting entry on webmention.io then looks something like this:<\/p>\n<pre class=\"language-json\">    {<br \/>\"type\": \"entry\",<br \/>\"author\": {<br \/>\"type\": \"card\",<br \/>\"name\": \"Sara Soueidan\",<br \/>\"photo\": \"https:\/\/webmention.io\/avatar\/pbs.twimg.com\/579a474c9b858845a9e64693067e12858642fa71059d542dce6285aed5e10767.jpg\",<br \/>\"url\": \"https:\/\/sarasoueidan.com\"<br \/>},<br \/>\"url\": \"https:\/\/twitter.com\/SaraSoueidan\/status\/1022009419926839296\",<br \/>\"published\": \"2018-07-25T06:43:28+00:00\",<br \/>\"wm-received\": \"2018-07-25T07:01:17Z\",<br \/>\"wm-id\": 537028,<br \/>\"wm-source\": \"https:\/\/brid-gy.appspot.com\/comment\/twitter\/mxbck\/1022001729389383680\/1022009419926839296\",<br \/>\"wm-target\": \"https:\/\/mxb.dev\/blog\/layouts-of-tomorrow\/\",<br \/>\"content\": {<br \/>\"content-type\": \"text\/plain\",<br \/>\"value\": \"This looks great!\",<br \/>\"text\": \"This looks great!\"<br \/>},<br \/>\"in-reply-to\": \"https:\/\/mxb.dev\/blog\/layouts-of-tomorrow\/\",<br \/>\"wm-property\": \"in-reply-to\",<br \/>\"wm-private\": false<br \/>}<\/pre>\n<h3 id=\"h-but-wait-there%E2%80%99s-more\"><a class=\"heading-anchor\" href=\"#h-but-wait-there%E2%80%99s-more\">#<\/a> But wait, there\u2019s more!<\/h3>\n<p>The beauty of webmentions is that unlike with regular social media, reactions to your content are not limited to users of one site. You can combine comments from Facebook and Twitter with replies people posted on their own blogs. You can mix retweets and shares with mentions of your content in newsletters or forum threads.<\/p>\n<p>You also have complete control over who and what is allowed in your mentions. Content silos often only allow muting or blocking on your own timeline, everyone else can still see unwanted or abusive @-replies. With webmentions, you\u2019re free to moderate reactions however you see fit. Fuck off, Nazis!<\/p>\n<p>Once the webmention endpoint is in place, we still need to pull the aggregated data down to our site and display it in a meaningful way.<\/p>\n<p>The way to do this depends on your setup. Webmention.io offers an <a href=\"https:\/\/github.com\/aaronpk\/webmention.io#api\">API<\/a> that provides data as a JSON feed, for example. You can query mentions for a specific URL, or get everything associated with a particular domain (allthough the latter is only available to site owners.)<\/p>\n<p>My site uses <a href=\"https:\/\/11ty.io\">Eleventy<\/a>, which has a conventient way to pull in external data at build time. By providing a <a href=\"https:\/\/www.11ty.io\/docs\/data-js\/#using-js-data-files\">custom function<\/a> that queries the API, Eleventy will fetch my webmentions and expose them to the templates when generating the site.<\/p>\n<pre class=\"language-js\">\/\/ data\/webmentions.js<br \/>const API_ORIGIN = 'https:\/\/webmention.io\/api\/mentions.jf2'<br \/>module.exports = async function() {<br \/>const domain = 'mxb.dev'<br \/>const token = process.env.WEBMENTION_IO_TOKEN<br \/>const url = `${API_ORIGIN}?domain=${domain}&amp;token=${token}`<br \/>try {<br \/>const response = await fetch(url)<br \/>if (response.ok) {<br \/>const feed = await response.json()<br \/>return feed<br \/>}<br \/>} catch (err) {<br \/>console.error(err)<br \/>return null<br \/>}<br \/>}<\/pre>\n<p><em>The feed can now be accessed in the {{ webmentions }} variable.<\/em><\/p>\n<p>Here\u2019s <a href=\"https:\/\/github.com\/maxboeck\/mxb\/blob\/master\/src\/data\/webmentions.js\">the complete function<\/a> if you\u2019re interested. Other static site generators offer similiar methods to fetch external data.<\/p>\n<h2 id=\"h-parsing-and-filtering\"><a class=\"heading-anchor\" href=\"#h-parsing-and-filtering\">#<\/a> Parsing and Filtering<\/h2>\n<p>Now that the raw data is available, we can mold it into any shape we\u2019d like. For my site, the processing steps look like this:<\/p>\n<ul><li>Filter the raw data for each post, only include mentions targeting that URL.<\/li>\n<li>Only allow \u201cmentions\u201d and \u201creplies\u201d in the comment section. Likes and Reposts go somewhere else.<\/li>\n<li>Remove entries that dont have any content to display.<\/li>\n<li>Sanitize the output - strip HTML tags, truncate long content, etc.<\/li>\n<\/ul><pre class=\"language-js\">\/\/ filters.js<br \/>const sanitizeHTML = require('sanitize-html')<br \/>function getWebmentionsForUrl(webmentions, url) {<br \/>const allowedTypes = ['mention-of', 'in-reply-to']<br \/>const hasRequiredFields = entry =&gt; {<br \/>const { author, published, content } = entry<br \/>return author.name &amp;&amp; published &amp;&amp; content<br \/>}<br \/>const sanitize = entry =&gt; {<br \/>const { content } = entry<br \/>if (content['content-type'] === 'text\/html') {<br \/>content.value = sanitizeHTML(content.value)<br \/>}<br \/>return entry<br \/>}<br \/>return webmentions<br \/>.filter(entry =&gt; entry['wm-target'] === url)<br \/>.filter(entry =&gt; allowedTypes.includes(entry['wm-property']))<br \/>.filter(hasRequiredFields)<br \/>.map(sanitize)<br \/>}<\/pre>\n<p>In Eleventy\u2019s case, I can set that function as a custom filter to use in my post templates.<br \/>Each post will then loop over its webmentions and output them underneath.<\/p>\n<pre class=\"language-html\">&lt;!-- webmentions.njk --&gt;<br \/>{% set mentions = webmentions | getWebmentionsForUrl(absoluteUrl) %}<br \/>&lt;ol id=\"webmentions\"&gt;<br \/>{% for webmention in mentions %}<br \/>&lt;li class=\"webmentions__item\"&gt;<br \/>{% include 'webmention.njk' %}<br \/>&lt;\/li&gt;<br \/>{% endfor %}<br \/>&lt;\/ol&gt;<\/pre>\n<p>You can see the result by scrolling down to the <a href=\"#webmentions\">end of this post<\/a> (if there are any replies \ud83d\ude09).<\/p>\n<aside class=\"signup js-signup-widget\" data-nosnippet=\"\"><h2 id=\"h-client-side-rendering\"><a class=\"heading-anchor\" href=\"#h-client-side-rendering\">#<\/a> Client-Side Rendering<\/h2>\n<p>Because static sites are, well, static - it\u2019s possible that new mentions have happened since the last build. To keep the webmention section up-to-date, there\u2019s an extra step we can take: client side rendering.<\/p>\n<p>Remember I said the webmention.io API can be used to only fetch mentions for a specific URL? That comes in handy now. After the page has loaded, we can fetch the latest mentions for the current URL and re-render the static webmention section with them.<\/p>\n<p>On my site, I used <a href=\"https:\/\/preactjs.com\/\">Preact<\/a> to do just that. It has a very small (~3kB) footprint and lets me use React\u2019s mental model and JSX syntax. It would probably also have been possible to re-use the existing <code>nunjucks<\/code> templates, but this solution was the easiest and most lightweight for me.<\/p>\n<p>I essentially used the same logic here as I did in the static build, to ensure matching results. The rendering only starts after the API call returned valid data though - if anything goes wrong or the API is unavailable, there will still be the static content as a fallback.<\/p>\n<pre class=\"language-js\">\/\/ webmentions\/index.js<br \/>import { h, render } from 'preact'<br \/>import App from '.\/App'<br \/>...<br \/>const rootElement = document.getElementById('webmentions')<br \/>if (rootElement) {<br \/>fetchMentions()<br \/>.then(data =&gt; {<br \/>if (data.length) {<br \/>render(&lt;App webmentions={data} \/&gt;, rootElement)<br \/>}<br \/>})<br \/>.catch(err =&gt; {<br \/>console.error(err)<br \/>})<br \/>}<\/pre>\n<p>There are of course still some missing pieces, most notably the ability to send outgoing webmentions to URLs linked to in your own blog posts. <span class=\"c10\">I might have to look into that.<\/span><\/p>\n<p><a href=\"https:\/\/remysharp.com\/\">Remy Sharp<\/a> has recently published a very useful new tool that takes care of handling <strong>outgoing<\/strong> webmentions for you. <a href=\"https:\/\/webmention.app\/\">Webmention.app<\/a> is a platform agnostic service that will check a given URL for links to other sites, discover if they support webmentions, then send a webmention to the target.<\/p>\n<p>You can use that service in a number of ways, including your own command line. If you host your site on Netlify though, it\u2019s also very straightforward to integrate it <a href=\"https:\/\/webmention.app\/docs#how-to-integrate-with-netlify\">using deployment webhooks<\/a>!<\/p>\n<h2 id=\"h-jekyll-plugin\"><a class=\"heading-anchor\" href=\"#h-jekyll-plugin\">#<\/a> Jekyll Plugin<\/h2>\n<p>My implementation was heavily inspired by Aaron Gustafson\u2019s excellent Jekyll Plugin (link below), which goes even further with customization and caching options. If you\u2019re running a Jekyll site, use that for almost instant webmention support \ud83d\udc4d.<\/p>\n<h2 id=\"h-further-resources\"><a class=\"heading-anchor\" href=\"#h-further-resources\">#<\/a> Further Resources<\/h2>\n<ul><li><a href=\"https:\/\/www.w3.org\/TR\/webmention\/\">W3C Recommendation<\/a> for the Webmention Standard<\/li>\n<li><a href=\"https:\/\/github.com\/aarongustafson\/jekyll-webmention_io\">Webmention.io Jekyll Plugin<\/a> by Aaron Gustafson<\/li>\n<li><a href=\"https:\/\/indieweb.org\/Webmention-developer#IndieWeb_Examples\">Indieweb Examples<\/a> of people using Webmentions on their sites<\/li>\n<\/ul><\/aside>","created_at":"2020-06-19T13:03:34+0000","updated_at":"2020-06-19T13:03:34+0000","published_at":"2019-01-10T00:00:00+0000","published_by":["Max B\u00f6ck"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"mxb.dev","preview_picture":"https:\/\/mxb.dev\/assets\/media\/static-indieweb-webmentions\/cover.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1180,"title":"People are still creating great stuff along these lines - you just won't find it...","url":"https:\/\/news.ycombinator.com\/item?id=23568555","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nextaccountic\" class=\"hnuser\">nextaccountic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568782\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=StillBored\" class=\"hnuser\">StillBored<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569239\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=slightwinder\" class=\"hnuser\">slightwinder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573366\">34 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=koreth1\" class=\"hnuser\">koreth1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569410\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pphysch\" class=\"hnuser\">pphysch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569670\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rpdillon\" class=\"hnuser\">rpdillon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570041\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=koreth1\" class=\"hnuser\">koreth1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569768\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=raverbashing\" class=\"hnuser\">raverbashing<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571943\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=FalconSensei\" class=\"hnuser\">FalconSensei<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570466\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=slightwinder\" class=\"hnuser\">slightwinder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573385\">29 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=majormajor\" class=\"hnuser\">majormajor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569467\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=esfandia\" class=\"hnuser\">esfandia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570740\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rkagerer\" class=\"hnuser\">rkagerer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571510\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anaxag0ras\" class=\"hnuser\">anaxag0ras<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573094\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=RNCTX\" class=\"hnuser\">RNCTX<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570860\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=japanuspus\" class=\"hnuser\">japanuspus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570996\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573006\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=api\" class=\"hnuser\">api<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570613\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=basch\" class=\"hnuser\">basch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568894\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=spc476\" class=\"hnuser\">spc476<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569900\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dredmorbius\" class=\"hnuser\">dredmorbius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570405\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=basch\" class=\"hnuser\">basch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571545\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspenmayer\" class=\"hnuser\">aspenmayer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571733\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jandrese\" class=\"hnuser\">jandrese<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569086\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=asdff\" class=\"hnuser\">asdff<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569352\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=majormajor\" class=\"hnuser\">majormajor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569477\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jandrese\" class=\"hnuser\">jandrese<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569411\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=chongli\" class=\"hnuser\">chongli<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569972\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jefftk\" class=\"hnuser\">jefftk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570015\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dredmorbius\" class=\"hnuser\">dredmorbius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570417\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=russellendicott\" class=\"hnuser\">russellendicott<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570208\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=api\" class=\"hnuser\">api<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570618\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sacado2\" class=\"hnuser\">sacado2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572570\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jb3689\" class=\"hnuser\">jb3689<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571365\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ryan-allen\" class=\"hnuser\">ryan-allen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572322\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MaxBarraclough\" class=\"hnuser\">MaxBarraclough<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572554\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=fit2rule\" class=\"hnuser\">fit2rule<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572775\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=greggman3\" class=\"hnuser\">greggman3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573418\">23 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ricardo81\" class=\"hnuser\">ricardo81<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572008\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=WalterBright\" class=\"hnuser\">WalterBright<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570619\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwingtheaway\" class=\"hnuser\">throwingtheaway<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573516\">12 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rkagerer\" class=\"hnuser\">rkagerer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570658\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ricardo81\" class=\"hnuser\">ricardo81<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572109\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=clarry\" class=\"hnuser\">clarry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572827\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=WalterBright\" class=\"hnuser\">WalterBright<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570751\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rkagerer\" class=\"hnuser\">rkagerer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570957\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rkagerer\" class=\"hnuser\">rkagerer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571210\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=summerlight\" class=\"hnuser\">summerlight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572885\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=WalterBright\" class=\"hnuser\">WalterBright<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571004\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=indigochill\" class=\"hnuser\">indigochill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571049\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=CJefferson\" class=\"hnuser\">CJefferson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568925\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=akavel\" class=\"hnuser\">akavel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571626\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bduerst\" class=\"hnuser\">bduerst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571960\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Cilvic\" class=\"hnuser\">Cilvic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571561\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=slfnflctd\" class=\"hnuser\">slfnflctd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568881\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=the_other\" class=\"hnuser\">the_other<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569167\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Jtsummers\" class=\"hnuser\">Jtsummers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569532\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tonyarkles\" class=\"hnuser\">tonyarkles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571410\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hedora\" class=\"hnuser\">hedora<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570432\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jabl\" class=\"hnuser\">jabl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23572151\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrfusion\" class=\"hnuser\">mrfusion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569783\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=raspyberr\" class=\"hnuser\">raspyberr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569625\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=renewiltord\" class=\"hnuser\">renewiltord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569138\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569320\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrfusion\" class=\"hnuser\">mrfusion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569792\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=akdas\" class=\"hnuser\">akdas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570793\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mediterraneo10\" class=\"hnuser\">Mediterraneo10<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571467\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nkozyra\" class=\"hnuser\">nkozyra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569282\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mech422\" class=\"hnuser\">mech422<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571789\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ozfive\" class=\"hnuser\">ozfive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570189\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570419\">9 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Blaiz0r\" class=\"hnuser\">Blaiz0r<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23573236\">58 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=SilasX\" class=\"hnuser\">SilasX<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23569398\">12 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=killjoywashere\" class=\"hnuser\">killjoywashere<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23571398\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomc1985\" class=\"hnuser\">tomc1985<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568731\">14 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568772\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zozbot234\" class=\"hnuser\">zozbot234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568900\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kevin_thibedeau\" class=\"hnuser\">kevin_thibedeau<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568848\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tdeck\" class=\"hnuser\">tdeck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23570781\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=duskwuff\" class=\"hnuser\">duskwuff<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=23568824\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>","created_at":"2020-06-19T12:17:44+0000","updated_at":"2020-06-19T12:17:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","elfeed","rss"],"is_public":false,"id":1178,"title":"r\/emacs - Take Elfeed everywhere: Mobile rss reading Emacs-style (for free\/cheap)","url":"https:\/\/www.reddit.com\/r\/emacs\/comments\/5z2s75\/take_elfeed_everywhere_mobile_rss_reading\/","content":"<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">After the demise of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Google_Reader\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Google Reader<\/a>, I switched for some time to <a href=\"https:\/\/www.newsblur.com\/\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">NewsBlur<\/a>. Unfortunately, its Android app never worked well for me. And integrating more of my life into Emacs is always desirable, so once I saw there was an <a href=\"https:\/\/github.com\/areina\/elfeed-cljsrn\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Android interface<\/a> for the fantastic Emacs RSS reader <a href=\"https:\/\/github.com\/skeeto\/elfeed\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Elfeed<\/a>, I made the switch.<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">The tricky thing with the Elfeed Android client is that it wants to connect to <a href=\"https:\/\/github.com\/skeeto\/elfeed#web-interface\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">web interface<\/a> of an instance of Elfeed running inside of Emacs. I could have done with my home computer, but that would require poking a hole through the firewall and in any case would be non-ideal when for instance I was travelling.<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">About a month ago I hit upon a cheap (in fact, free) solution for running a remote instance of Emacs running Elfeed that is connectable with the <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.elfeedcljsrn\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Android app<\/a>. The VPS provider <a href=\"https:\/\/www.wishosting.com\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Wishosting<\/a> offers an OpenVZ mini for <a href=\"https:\/\/www.wishosting.com\/order\/main\/packages\/VPS\/?group_id=2\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">$4\/year<\/a>, and if you stick a link to Wishosting on your own domain <a href=\"https:\/\/www.wishosting.com\/plugin\/support_manager\/knowledgebase\/view\/14\/how-to-get-openvz-mini-for-free\/\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">you can get it for free<\/a>.<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">On my home desktop, work desktop, and laptop, I have <a href=\"https:\/\/syncthing.net\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Syncthing<\/a> installed and I use this to keep the Elfeed database in sync between these machines.<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Here I outline how to add a remote always-on instance of Elfeed running in Wishosting\u2019s OpenVZ mini which also remains in sync with all of the other machines. (I use <a href=\"https:\/\/github.com\/remyhonig\/elfeed-org\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">elfeed-org<\/a> to organise my feeds, and just keep the elfeed.org file in ~\/.elfeed\/.) Just use Syncthing to keep the ~\/.elfeed directory sync\u2019ed between all of the machines (including any VPS\u2019s).<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">I set up an Ubuntu 16.04 LTS VPS on Wishosting, and then installed Emacs and <a href=\"https:\/\/www.linuxbabe.com\/ubuntu\/install-syncthing-ubuntu-16-04-via-official-deb-repository\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Syncthing<\/a>, and that is what I would recommend.<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\"><strong class=\"_12FoOEddL7j_RgMQN0SNeU\">Steps<\/strong><\/p><ul class=\"_33MEMislY0GAlB78wL1_CR\"><li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Create your Ubuntu 16.04 LTS VPS.<\/p>\n<\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Make a (non-root) user with superuser capability. Login with this user.<\/p>\n<\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Install Emacs and <a href=\"https:\/\/www.linuxbabe.com\/ubuntu\/install-syncthing-ubuntu-16-04-via-official-deb-repository\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Syncthing<\/a>.<\/p>\n<\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Configure Syncthing appropriately. You can <a href=\"https:\/\/docs.syncthing.net\/users\/firewall.html#remote-web-gui\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">connect remotely to the Web GUI<\/a>.<\/p>\n<\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Create and save a .emacs file in your user\u2019s ~ with the following contents:<\/p>\n<pre class=\"_34q3PgLsx9zIU5BiSOjFoM\">;; package setup here\n(require 'package)\n(package-initialize nil)\n(setq package-enable-at-startup nil)\n(add-to-list 'package-archives '(\"org\" . \"http:\/\/orgmode.org\/elpa\/\") t)\n(add-to-list 'package-archives\n             '(\"melpa\" . \"https:\/\/melpa.org\/packages\/\") t)\n(add-to-list 'package-archives\n             '(\"marmalade\" .\n               \"http:\/\/marmalade-repo.org\/packages\/\"))\n(package-initialize)\n;; general add packages to list\n(let ((default-directory  \"~\/.emacs.d\/elpa\/\"))\n  (normal-top-level-add-subdirs-to-load-path))\n;; make sure 'use-package is installed\n(unless (package-installed-p 'use-package)\n  (package-refresh-contents)\n  (package-install 'use-package))\n;;; use-package\n(require 'use-package)\n;; Load elfeed\n(use-package elfeed\n  :ensure t\n  :bind (:map elfeed-search-mode-map\n                                        ;              (\"A\" . bjm\/elfeed-show-all)\n                                        ;              (\"E\" . bjm\/elfeed-show-emacs)\n                                        ;              (\"D\" . bjm\/elfeed-show-daily)\n              (\"q\" . bjm\/elfeed-save-db-and-bury)))\n(require 'elfeed)\n;; Load elfeed-org\n(use-package elfeed-org\n  :ensure t\n  :config\n  (elfeed-org)\n  (setq rmh-elfeed-org-files (list \"~\/.elfeed\/elfeed.org\"))\n  )\n;; Laod elfeed-goodies\n(use-package elfeed-goodies\n  :ensure t\n  )\n(elfeed-goodies\/setup)\n;; Load elfeed-web\n(use-package elfeed-web\n  :ensure t\n  )\n;;; Elfeed\n(global-set-key (kbd \"C-x w\") 'bjm\/elfeed-load-db-and-open)\n(define-key elfeed-show-mode-map (kbd \"j\") 'elfeed-goodies\/split-show-next)\n(define-key elfeed-show-mode-map (kbd \"k\") 'elfeed-goodies\/split-show-prev)\n(define-key elfeed-search-mode-map (kbd \"j\") 'next-line)\n(define-key elfeed-search-mode-map (kbd \"k\") 'previous-line)\n(define-key elfeed-show-mode-map (kbd \"S-SPC\") 'scroll-down-command)\n;;write to disk when quiting\n(defun bjm\/elfeed-save-db-and-bury ()\n  \"Wrapper to save the elfeed db to disk before burying buffer\"\n  (interactive)\n  (elfeed-db-save)\n  (quit-window))\n;;functions to support syncing .elfeed between machines\n;;makes sure elfeed reads index from disk before launching\n(defun bjm\/elfeed-load-db-and-open ()\n  \"Wrapper to load the elfeed db from disk before opening\"\n  (interactive)\n  (elfeed-db-load)\n  (elfeed)\n  (elfeed-search-update--force)\n  (elfeed-update))\n(defun bjm\/elfeed-updater ()\n  \"Wrapper to load the elfeed db from disk before opening\"\n  (interactive)\n  (elfeed-db-save)\n  (quit-window)\n  (elfeed-db-load)\n  (elfeed)\n  (elfeed-search-update--force)\n  (elfeed-update))\n(run-with-timer 0 (* 30 60) 'bjm\/elfeed-updater)\n(setq httpd-port NNNNN)   ; replace NNNNN with a port equalling your start port + 10 (or whatever)\n(elfeed-web-start)\n<\/pre><\/li>\n<\/ul><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Note that you\u2019ll need to configure the httpd-port appropriately as per the comment in the elisp above.<\/p><ul class=\"_33MEMislY0GAlB78wL1_CR\"><li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Then create the following systemd unit at ~\/.config\/systemd\/user\/emacs.service:<\/p>\n<pre class=\"_34q3PgLsx9zIU5BiSOjFoM\">[Unit]\nDescription=Emacs: the extensible, self-documenting text editor\n[Service]\nType=forking\nExecStart=\/usr\/bin\/emacs --daemon\nExecStop=\/usr\/bin\/emacsclient --eval \"(kill-emacs)\"\nRestart=always\n[Install]\nWantedBy=default.target\n<\/pre><\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Enable and start this unit with:<\/p>\n<pre class=\"_34q3PgLsx9zIU5BiSOjFoM\">$ systemctl --user enable --now emacs\n<\/pre><\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">And then make sure it runs persistently even when you\u2019re not connected to your VPS via ssh (this tripped me up for some time):<\/p>\n<pre class=\"_34q3PgLsx9zIU5BiSOjFoM\"># loginctl enable-linger USERNAME\n<\/pre><\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Install the <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.elfeedcljsrn&amp;hl=en\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">Elfeed Android app<\/a> on your mobile, enter the app\u2019s Settings, and put in whatever your Wishosting ip is plus the port you chose above (NNNNN), e.g. <a href=\"http:\/\/199.39.100.23:54169\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">http:\/\/199.39.100.23:54169<\/a>, for the Elfeed web url.<\/p>\n<\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">That should be it. Now you\u2019ll have access to your rss feed all over the world and Syncthing will ensure that all changes will be propagated to all of your Emacs instances including your VPS. These changes would include adding and deleting rss feeds and marking of posts as \u2019read\u2019.<\/p>\n<\/li>\n<li class=\"_3gqTEjt4x9UIIpWiro7YXz\">\n<p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Note: on my desktops\/laptop I use the following setup for Elfeed:<\/p>\n<pre class=\"_34q3PgLsx9zIU5BiSOjFoM\">;; Load elfeed\n(use-package elfeed\n  :ensure t\n  :bind (:map elfeed-search-mode-map\n;              (\"A\" . bjm\/elfeed-show-all)\n;              (\"E\" . bjm\/elfeed-show-emacs)\n;              (\"D\" . bjm\/elfeed-show-daily)\n              (\"q\" . bjm\/elfeed-save-db-and-bury)))\n(require 'elfeed)\n;; Load elfeed-org\n(use-package elfeed-org\n  :ensure t\n  :config\n  (elfeed-org)\n  (setq rmh-elfeed-org-files (list \"~\/.elfeed\/elfeed.org\"))\n  )\n;; Laod elfeed-goodies\n(use-package elfeed-goodies\n  :ensure t\n  )\n(elfeed-goodies\/setup)\n;; Load elfeed-web\n(use-package elfeed-web\n  :ensure t\n  )\n;;functions to support syncing .elfeed between machines\n;;makes sure elfeed reads index from disk before launching\n(defun bjm\/elfeed-load-db-and-open ()\n  \"Wrapper to load the elfeed db from disk before opening\"\n  (interactive)\n  (elfeed-db-load)\n  (elfeed)\n  (elfeed-search-update--force)\n  (elfeed-update))\n;;write to disk when quiting\n(defun bjm\/elfeed-save-db-and-bury ()\n  \"Wrapper to save the elfeed db to disk before burying buffer\"\n  (interactive)\n  (elfeed-db-save)\n  (quit-window))\n;;; Elfeed\n(global-set-key (kbd \"C-x w\") 'bjm\/elfeed-load-db-and-open)\n(define-key elfeed-show-mode-map (kbd \";\") 'visual-fill-column-mode)\n(define-key elfeed-show-mode-map (kbd \"j\") 'elfeed-goodies\/split-show-next)\n(define-key elfeed-show-mode-map (kbd \"k\") 'elfeed-goodies\/split-show-prev)\n(define-key elfeed-search-mode-map (kbd \"j\") 'next-line)\n(define-key elfeed-search-mode-map (kbd \"k\") 'previous-line)\n(define-key elfeed-show-mode-map (kbd \"S-SPC\") 'scroll-down-command)\n;; probably temporary: hack for elfeed-goodies date column:\n(defun elfeed-goodies\/search-header-draw ()\n  \"Returns the string to be used as the Elfeed header.\"\n  (if (zerop (elfeed-db-last-update))\n      (elfeed-search--intro-header)\n    (let* ((separator-left (intern (format \"powerline-%s-%s\"\n                                           elfeed-goodies\/powerline-default-separator\n                                           (car powerline-default-separator-dir))))\n           (separator-right (intern (format \"powerline-%s-%s\"\n                                            elfeed-goodies\/powerline-default-separator\n                                            (cdr powerline-default-separator-dir))))\n           (db-time (seconds-to-time (elfeed-db-last-update)))\n           (stats (-elfeed\/feed-stats))\n           (search-filter (cond\n                           (elfeed-search-filter-active\n                            \"\")\n                           (elfeed-search-filter\n                            elfeed-search-filter)\n                           (\"\"))))\n      (if (&gt;= (window-width) (* (frame-width) elfeed-goodies\/wide-threshold))\n          (search-header\/draw-wide separator-left separator-right search-filter stats db-time)\n        (search-header\/draw-tight separator-left separator-right search-filter stats db-time)))))\n(defun elfeed-goodies\/entry-line-draw (entry)\n  \"Print ENTRY to the buffer.\"\n  (let* ((title (or (elfeed-meta entry :title) (elfeed-entry-title entry) \"\"))\n         (date (elfeed-search-format-date (elfeed-entry-date entry)))\n         (title-faces (elfeed-search--faces (elfeed-entry-tags entry)))\n         (feed (elfeed-entry-feed entry))\n         (feed-title\n          (when feed\n            (or (elfeed-meta feed :title) (elfeed-feed-title feed))))\n         (tags (mapcar #'symbol-name (elfeed-entry-tags entry)))\n         (tags-str (concat \"[\" (mapconcat 'identity tags \",\") \"]\"))\n         (title-width (- (window-width) elfeed-goodies\/feed-source-column-width\n                         elfeed-goodies\/tag-column-width 4))\n         (title-column (elfeed-format-column\n                        title (elfeed-clamp\n                               elfeed-search-title-min-width\n                               title-width\n                               title-width)\n                        :left))\n         (tag-column (elfeed-format-column\n                      tags-str (elfeed-clamp (length tags-str)\n                                             elfeed-goodies\/tag-column-width\n                                             elfeed-goodies\/tag-column-width)\n                      :left))\n         (feed-column (elfeed-format-column\n                       feed-title (elfeed-clamp elfeed-goodies\/feed-source-column-width\n                                                elfeed-goodies\/feed-source-column-width\n                                                elfeed-goodies\/feed-source-column-width)\n                       :left)))\n    (if (&gt;= (window-width) (* (frame-width) elfeed-goodies\/wide-threshold))\n        (progn\n          (insert (propertize date 'face 'elfeed-search-date-face) \" \")\n          (insert (propertize feed-column 'face 'elfeed-search-feed-face) \" \")\n          (insert (propertize tag-column 'face 'elfeed-search-tag-face) \" \")\n          (insert (propertize title 'face title-faces 'kbd-help title)))\n      (insert (propertize title 'face title-faces 'kbd-help title)))))\n<\/pre><\/li>\n<\/ul><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">Screenshots:<\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\"><a href=\"https:\/\/1.bp.blogspot.com\/-6vS4HvzOX8g\/WMYHINFRy6I\/AAAAAAAABHM\/QGHJ7wZBD0sAxNoTfOYQCZlnXSypfu3lwCPcB\/s1600\/elfeed-desktop.jpg\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">elfeed-desktop<\/a> <a href=\"https:\/\/3.bp.blogspot.com\/-NrH7c65PTqA\/WMYHH5-Fi0I\/AAAAAAAABHI\/s33yd9AjzFg3Zq_awNUW3o5wsand7K8RACPcB\/s1600\/elfeed-mobile.png\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">elfeed-mobile<\/a><\/p><p class=\"_1qeIAgB0cPwnLhDF9XSiJM\">[also available as a <a href=\"http:\/\/babbagefiles.blogspot.co.uk\/2017\/03\/take-elfeed-everywhere-mobile-rss.html\" class=\"_3t5uN8xUmg0TOwRCOGQEcU\" rel=\"noopener nofollow ugc\" target=\"_blank\">blog post<\/a>]<\/p>","created_at":"2020-06-17T14:20:29+0000","updated_at":"2020-06-17T14:20:38+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"www.reddit.com","preview_picture":"https:\/\/external-preview.redd.it\/BiclBH1-qt8WNIobMNHVORtf7hbOy3Hvz0BXq9eXtDQ.jpg?auto=webp&s=baf468de195d617c28acf549f17beacec717d506","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1177,"title":"The American Press Is Destroying Itself","url":"https:\/\/taibbi.substack.com\/p\/the-news-media-is-destroying-itself?r=2","content":"<a class=\"image-link image2 image2-930-1456\" target=\"_blank\" href=\"https:\/\/cdn.substack.com\/image\/fetch\/c_limit,f_auto,q_auto:good,fl_progressive:steep\/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2Ff1bcd347-c525-4a12-a0b8-5cd7c99f4173_5203x3325.jpeg\"><img src=\"https:\/\/cdn.substack.com\/image\/fetch\/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep\/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2Ff1bcd347-c525-4a12-a0b8-5cd7c99f4173_5203x3325.jpeg\" data-attrs=\"{&quot;src&quot;:&quot;https:\/\/bucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com\/public\/images\/f1bcd347-c525-4a12-a0b8-5cd7c99f4173_5203x3325.jpeg&quot;,&quot;height&quot;:930,&quot;width&quot;:1456,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:6771607,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image\/jpeg&quot;,&quot;href&quot;:null}\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/a><p>Sometimes it seems life can\u2019t get any worse in this country. Already in terror of a pandemic, Americans have lately been bombarded with images of grotesque state-sponsored violence, from the murder of George Floyd to countless <a href=\"https:\/\/twitter.com\/nowthisnews\/status\/1269249681651105793\">scenes of police clubbing and brutalizing<\/a> protesters.<\/p><p>Our president, Donald Trump, is a clown who makes a great reality-show villain but is uniquely toolless as the leader of a superpower nation. Watching him try to think through two society-imperiling crises is like waiting for a gerbil to solve Fermat\u2019s theorem. <a href=\"https:\/\/abcnews.go.com\/US\/trumps-quest-dominate-amid-george-floyd-protests-sparks\/story?id=71126346\">Calls to \u201cdominate\u201d marchers<\/a> and ad-libbed speculations about Floyd\u2019s \u201cgreat day\u201d looking down from heaven at Trump\u2019s crisis management and new unemployment numbers (\u201c<a href=\"https:\/\/www.washingtonpost.com\/opinions\/2020\/06\/05\/george-floyd-gets-new-spokesman-donald-trump\/\">only\u201d 21 million<\/a> out of work!) were pure gasoline at a tinderbox moment. The man seems determined to talk us into civil war.<\/p><p>But police violence, and Trump\u2019s daily assaults on the presidential competence standard, are only part of the disaster. On the other side of the political aisle, among self-described liberals, we\u2019re watching an intellectual revolution. It feels liberating to say after years of tiptoeing around the fact, but the American left has lost its mind. It\u2019s become a cowardly mob of upper-class social media addicts, Twitter Robespierres who move from discipline to discipline torching reputations and jobs with breathtaking casualness.<\/p><p>The leaders of this new movement are replacing traditional liberal beliefs about tolerance, free inquiry, and even racial harmony with ideas so toxic and unattractive that they eschew debate, moving straight to shaming, threats, and intimidation. They are counting on the guilt-ridden, self-flagellating nature of traditional American progressives, who will not stand up for themselves, and will walk to the Razor voluntarily.<\/p><p>They\u2019ve conned organization after organization into empowering panels to search out thoughtcrime, and it\u2019s established now that anything can be an offense, from a UCLA professor placed under investigation for <a href=\"https:\/\/freebeacon.com\/issues\/university-to-investigate-lecturer-for-reading-mlks-letter-from-birmingham-jail\/\">reading Martin Luther King\u2019s \u201cLetter from a Birmingham Jail\u201d out loud<\/a> to a data scientist <a href=\"https:\/\/twitter.com\/jenbrea\/status\/1271148784316108800\">fired* from a research firm<\/a> for \u2014 get this \u2014 <a href=\"https:\/\/twitter.com\/davidshor\/status\/1265998625836019712?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1265998625836019712&amp;ref_url=https%3A%2F%2Fwww.nationalreview.com%2Fnews%2Fthe-cancel-counter%2F\">retweeting<\/a> an academic study <a href=\"https:\/\/www.cambridge.org\/core\/journals\/american-political-science-review\/article\/agenda-seeding-how-1960s-black-protests-moved-elites-public-opinion-and-voting\/136610C8C040C3D92F041BB2EFC3034C\">suggesting nonviolent protests may be more politically effective<\/a> than violent ones!<\/p><p>Now, this madness is coming for journalism. Beginning on Friday, June 5th, a series of controversies rocked the media. By my count, at least eight news organizations dealt with internal uprisings (it was likely more). Most involved groups of reporters and staffers demanding the firing or reprimand of colleagues who\u2019d made politically \u201cproblematic\u201d editorial or social media decisions.<\/p><p>The <em>New York Times,<\/em> the <em>Intercept<\/em>, <em>Vox,<\/em> the <em>Philadelphia Inquirier, Variety<\/em>,and others saw challenges to management.<\/p><p>Probably the most disturbing story involved <em>Intercept<\/em> writer Lee Fang, one of a fast-shrinking number of young reporters actually skilled in investigative journalism. Fang\u2019s work in the area of campaign finance especially has led to concrete impact, including a <a href=\"https:\/\/theintercept.com\/2019\/03\/11\/intercept-investigation-leads-to-record-fines-over-foreign-campaign-contributions\/\">record fine to a conservative Super PAC<\/a>: few young reporters have done more to combat corruption.<\/p><p>Yet Fang found himself denounced online as a racist, then hauled before H.R. His crime? During protests, he tweeted this <a href=\"https:\/\/twitter.com\/lhfang\/status\/1268390704645943297\">interview<\/a> with an African-American man named Maximum Fr, who described having two cousins murdered in the East Oakland neighborhood where he grew up. Saying his aunt is still not over those killings, Max asked:<\/p><blockquote>\n<p><em>I always question, why does a Black life matter only when a white man takes it?... Like, if a white man takes my life tonight, it\u2019s going to be national news, but if a Black man takes my life, it might not even be spoken of\u2026 It\u2019s stuff just like that that I just want in the mix.<\/em><\/p>\n<\/blockquote><p>Shortly after, a co-worker of Fang\u2019s, Akela Lacy, wrote, \u201cTired of being made to deal continually with my co-worker @lhfang continuing to push black on black crime narratives after being repeatedly asked not to. This isn\u2019t about me and him, it\u2019s about institutional racism and using free speech to couch anti-blackness. I am so fucking tired.\u201d She followed with, \u201cStop being racist Lee.\u201d<\/p><p>The tweet received tens of thousands of likes and responses along the lines of, \u201c<a href=\"https:\/\/twitter.com\/GenePark\/status\/1268561870878846979\">Lee Fang has been like this for years, but the current moment only makes his anti-Blackness more glaring<\/a>,\u201d and \u201c<a href=\"https:\/\/twitter.com\/search?q=lee%20fang%20racist&amp;src=typed_query\">Lee Fang spouting racist bullshit it must be a day ending in day<\/a>.\u201d A significant number of Fang\u2019s co-workers, nearly all white, as well as reporters from other major news organizations like the <em>New York Times<\/em> and MSNBCand political activists (one <a href=\"https:\/\/www.democracyinaction.us\/2020\/warren\/warrenorg.html\">former Elizabeth Warren staffer<\/a> tweeted, \u201c<a href=\"https:\/\/twitter.com\/HelenBrosnan\/status\/1268560538486951950\">Get him<\/a>!\u201d), issued likes and messages of support for the notion that Fang was a racist. Though he had support within the organization, no one among his co-workers was willing to say anything in his defense publicly.<\/p><p>Like many reporters, Fang has always viewed it as part of his job to ask questions in all directions. He\u2019s written critically of political figures on the center-left, the left, and \u201cobviously on the right,\u201d and his reporting has inspired serious threats in the past. None of those past experiences were as terrifying as this blitz by would-be colleagues, which he described as \u201cjarring,\u201d \u201cdeeply isolating,\u201d and \u201cunique in my professional experience.\u201d<\/p><p>To save his career, Fang had to craft a <a href=\"https:\/\/www.documentcloud.org\/documents\/6938336-Statement.html\">public apology<\/a> for \u201cinsensitivity to the lived experience of others.\u201d According to one friend of his, it\u2019s been communicated to Fang that his continued employment at <em>The Intercept<\/em> is contingent upon avoiding comments that may upset colleagues. Lacy to her credit <a href=\"https:\/\/twitter.com\/akela_lacy\/status\/1269335670583164929\">publicly thanked Fang<\/a> for his statement and expressed willingness to have a conversation; unfortunately, the throng of <em>Intercept<\/em> co-workers who piled on her initial accusation did not join her in this.<\/p><p>I first met Lee Fang in 2014 and have never known him to be anything but kind, gracious, and easygoing. He also appears earnestly committed to making the world a better place through his work. It\u2019s stunning that so many colleagues are comfortable using a word as extreme and villainous as <em>racist<\/em> to describe him.<\/p><p>Though he describes his upbringing as \u201csolidly middle-class,\u201d Fang grew up in up in a diverse community in Prince George's County, Maryland, and attended public schools where he was frequently among the few non-African Americans in his class. As a teenager, he was witness to the murder of a young man outside his home by police who were never prosecuted, and also volunteered at a shelter for trafficked women, two of whom were murdered. If there\u2019s an edge to Fang at all, it seems geared toward people in our business who grew up in affluent circumstances and might intellectualize topics that have personal meaning for him.<\/p><p>In the tweets that got him in trouble with Lacy and other co-workers, he questioned the logic of protesters attacking immigrant-owned businesses \u201c<a href=\"https:\/\/twitter.com\/lhfang\/status\/1267108732288155649\">with no connection to police brutality at all<\/a>.\u201d He also offered his opinion on <a href=\"https:\/\/twitter.com\/lhfang\/status\/1267162143578718208\">Martin Luther King\u2019s attitude toward violent protest<\/a> (Fang\u2019s take was that King did not support it; Lacy responded, \u201cyou know they killed him too right\u201d). These are issues around which there is still considerable disagreement among self-described liberals, even among self-described leftists. Fang also commented, presciently as it turns out, that many reporters were \u201cterrified of openly challenging the lefty conventional wisdom around riots.\u201d<\/p><p>Lacy says she never intended for Fang to be \u201cfired, \u2018canceled,\u2019 or deplatformed,\u201d but appeared irritated by questions on the subject, which she says suggest, \u201cthere is more concern about naming racism than letting it persist.\u201d<\/p><p>Max himself was stunned to find out that his comments on all this had created a Twitter firestorm. \u201cI couldn\u2019t believe they were coming for the man\u2019s job over something I said,\u201d he recounts. \u201cIt was not Lee\u2019s opinion. It was my opinion.\u201d<\/p><p>By phone, Max spoke of a responsibility he feels Black people have to speak out against all forms of violence, \u201cprecisely because we experience it the most.\u201d He described being affected by the Floyd story, but also by the story of retired African-American police captain David Dorn, <a href=\"https:\/\/twitter.com\/KuonjiAliceMoe\/status\/1268023462242680832\">shot to death<\/a> in recent protests in St. Louis. He also mentioned Tony Timpa, a white man whose 2016 asphyxiation by police was only uncovered last year. In body-camera footage, police are heard joking after Timpa passed out and stopped moving, \u201c<a href=\"https:\/\/www.dallasnews.com\/news\/investigations\/2019\/07\/31\/you-re-gonna-kill-me-dallas-police-body-cam-footage-reveals-the-final-minutes-of-tony-timpa-s-life\/\">I don\u2019t want to go to school! Five more minutes, Mom<\/a>!\u201d<\/p><p>\u201cIf it happens to anyone, it has to be called out,\u201d Max says.<\/p><p>Max described discussions in which it was argued to him that bringing up these other incidents now is not helpful to the causes being articulated at the protests. He understands that point of view. He just disagrees.<\/p><p>\u201cThey say, there has to be the right time and a place to talk about that,\u201d he says. \u201cBut my point is, when? I want to speak out now.\u201d He pauses. \u201cWe\u2019ve taken the narrative, and instead of being inclusive with it, we\u2019ve become exclusive with it. Why?\u201d<\/p><p>There were other incidents. The editors of <em><a href=\"https:\/\/www.boston.com\/news\/media\/2020\/06\/09\/bon-appetit-editor-adam-rapoport-resigns\">Bon Apetit<\/a><\/em> and <em><a href=\"https:\/\/www.instagram.com\/p\/CBLToSInUcb\/\">Refinery29<\/a><\/em> both resigned amid accusations of toxic workplace culture. The editor of <em>Variety,<\/em> Claudia Eller, was <a href=\"https:\/\/nypost.com\/2020\/06\/05\/claudia-eller-variety-editor-in-chief-placed-on-leave-after-twitter-spat\/?utm_campaign=SocialFlow&amp;utm_source=NYPTwitter&amp;utm_medium=SocialFlow\">placed on leave<\/a> after calling a South Asian freelance writer \u201cbitter\u201d in a Twitter exchange about minority hiring at her company. The self-abasing apology (\u201cI have tried to diversify our newsroom over the past seven years, but I HAVE NOT DONE ENOUGH\u201d) was insufficient. Meanwhile, the <em>Philadelphia Inquirer\u2019s<\/em> editor, Stan Wischowski, was <a href=\"https:\/\/twitter.com\/nycsouthpaw\/status\/1269402533212160000\">forced<\/a> out after approving a headline, \u201cBuildings matter, too.\u201d<\/p><p>In the most discussed incident, <em>Times<\/em> editorial page editor James Bennet was ousted for green-lighting an anti-protest editorial by Arkansas Republican Senator Tom Cotton entitled, \u201c<a href=\"https:\/\/www.nytimes.com\/2020\/06\/03\/opinion\/tom-cotton-protests-military.html\">Send in the troops<\/a>.\u201d<\/p><p>I\u2019m no fan of Cotton, but as was the case with Michael Moore\u2019s documentary and many other controversial speech episodes, it\u2019s not clear that many of the people angriest about the piece in question even read it. In classic <em>Times<\/em> fashion, the paper has already scrubbed a mistake they made misreporting what their own editorial said, in an article about Bennet\u2019s ouster. Here\u2019s how the piece by Marc Tracy <a href=\"https:\/\/web.archive.org\/web\/20200607202330\/https:\/\/www.nytimes.com\/2020\/06\/07\/business\/media\/james-bennet-resigns-nytimes-op-ed.html\">read originally<\/a> (emphasis mine):<\/p><blockquote>\n<p>James Bennet, the editorial page editor of The New York Times, has resigned after a controversy over an Op-Ed by a senator <strong>calling for military force against protesters<\/strong> in American cities.<\/p>\n<\/blockquote><p>Here\u2019s how the piece <a href=\"https:\/\/www.nytimes.com\/2020\/06\/07\/business\/media\/james-bennet-resigns-nytimes-op-ed.html\">reads now<\/a>:<\/p><blockquote>\n<p>James Bennet resigned on Sunday from his job as the editorial page editor of The New York Times, days after the newspaper\u2019s opinion section, which he oversaw, published a much-criticized Op-Ed by a United States senator calling <strong>for a military response to civic unrest<\/strong> in American cities.<\/p>\n<\/blockquote><p>Cotton did not call for \u201cmilitary force against protesters in American cities.\u201d He spoke of a \u201cshow of force,\u201d to rectify a situation a significant portion of the country saw as spiraling out of control. It\u2019s an important distinction. Cotton was presenting one side of the most important question on the most important issue of a critically important day in American history.<\/p><p>As Cotton points out in the piece, he was advancing a view arguably held by a majority of the country. A <em>Morning Consult<\/em> poll showed <a href=\"https:\/\/assets.morningconsult.com\/wp-uploads\/2020\/06\/01181629\/2005131_crosstabs_POLICE_RVs_FINAL_LM-1.pdf\">58% of Americans<\/a> either strongly or somewhat supported the idea of \u201ccalling in the U.S. military to supplement city police forces.\u201d That survey included 40% of self-described \u201cliberals\u201d and 37% of African-Americans. To declare a point of view held by that many people not only not worthy of discussion, but so toxic that publication of it without even necessarily agreeing requires dismissal, is a dramatic reversal for a newspaper that long cast itself as the national paper of record.<\/p><p>Incidentally, that <a href=\"https:\/\/assets.morningconsult.com\/wp-uploads\/2020\/06\/01181629\/2005131_crosstabs_POLICE_RVs_FINAL_LM-1.pdf\">same poll cited by Cotton showed<\/a> that 73% of Americans described protecting property as \u201cvery important,\u201d while an additional 16% considered it \u201csomewhat important.\u201d This means the <em>Philadelphia Inquirer<\/em> editor was fired for running a headline \u2013 \u201cBuildings matter, too\u201d \u2013 that the poll said expressed a view held by 89% of the population, including 64% of African-Americans.<\/p><p>(Would I have run the <em>Inquirer<\/em> headline? No. In the context of the moment, the use of the word \u201cmatter\u201d especially sounds like the paper is equating \u201cBlack lives\u201d and \u201cbuildings,\u201d an odious and indefensible comparison. But why not just make this case in a rebuttal editorial? Make it a teaching moment? How can any editor operate knowing that airing opinions shared by a majority of readers might cost his or her job?)<\/p><p>The main thing accomplished by removing those types of editorials from newspapers \u2014 apart from scaring the hell out of editors \u2014 is to shield readers from knowledge of what a major segment of American society is thinking.<\/p><p>It also guarantees that opinion writers and editors alike will shape views to avoid upsetting colleagues, which means that instead of hearing what our differences are and how we might address those issues, newspaper readers will instead be presented with page after page of people professing to agree with one another. That\u2019s not agitation, that\u2019s misinformation.<\/p><p>The instinct to shield audiences from views or facts deemed politically uncomfortable has been in evidence since Trump became a national phenomenon. We saw it when reporters told audiences Hillary Clinton\u2019s small crowds were a \u201c<a href=\"https:\/\/www.nbcnews.com\/politics\/2016-election\/hillary-clinton-gambles-choosing-small-events-over-huge-rallies-n575311\">wholly intentional<\/a>\u201d campaign decision. I listened to colleagues that summer of 2016 talk about ignoring poll results, or anecdotes about Hillary\u2019s troubled campaign, on the grounds that doing otherwise might \u201chelp Trump\u201d (or, worse, be perceived that way).<\/p><p>Even if you embrace a wholly politically utilitarian vision of the news media \u2013 I don\u2019t, but let\u2019s say \u2013 non-reporting of that \u201centhusiasm\u201d story, or ignoring adverse poll results, didn\u2019t help Hillary\u2019s campaign. I\u2019d argue it more likely accomplished the opposite, contributing to voter apathy by conveying the false impression that her victory was secure.<\/p><p>After the 2016 election, we began to see staff uprisings. In one case, publishers at the <em>Nation<\/em> faced a revolt \u2013 from the <a href=\"https:\/\/twitter.com\/joanwalsh\/status\/916453801552265216\">Editor on down<\/a> \u2013 after articles by Aaron Mate <a href=\"https:\/\/www.washingtonpost.com\/blogs\/erik-wemple\/wp\/2017\/08\/15\/the-nation-is-reviewing-a-story-casting-doubt-on-russian-hack-of-dnc\/\">and Patrick Lawrence<\/a> questioning the evidentiary basis for Russiagate claims was run. Subsequent events, including the recent <a href=\"https:\/\/www.dailymail.co.uk\/news\/article-8299753\/Declassified-transcripts-Obama-officials-not-provide-evidence-Russia-Trump-collusion.html\">declassification of congressional testimony<\/a>, revealed that Mate especially was right to point out that officials had no evidence for a Trump-Russia collusion case. It\u2019s precisely because such unpopular views often turn out to be valid that we stress publishing and debating them in the press.<\/p><p>In a related incident, the <em>New Yorker<\/em> ran an article about Glenn Greenwald\u2019s Russiagate skepticism that quoted that same <em>Nation<\/em> editor, Joan Walsh, who had edited Greenwald at <em>Salon.<\/em> She suggested to the <em>New Yorker<\/em> that Greenwald\u2019s reservations were rooted in \u201cdisdain\u201d for the Democratic Party, in part because of its closeness to Wall Street, but also because of the \u201c<a href=\"https:\/\/www.newyorker.com\/magazine\/2018\/09\/03\/glenn-greenwald-the-bane-of-their-resistance\">ascendance of women and people of color<\/a>.\u201d The message was clear: even if you win a Pulitzer Prize, you can be accused of racism for deviating from approved narratives, even on questions that have nothing to do with race (the <em>New Yorker<\/em> piece also implied Greenwald\u2019s intransigence on Russia was pathological and grounded in trauma from childhood).<\/p><p>In the case of Cotton, <em>Times<\/em> staffers protested on the grounds that \u201c<a href=\"https:\/\/www.vox.com\/2020\/6\/5\/21280425\/new-york-times-tom-cotton-send-troops-staff-revolt\">Running this puts Black @NYTimes staff in danger<\/a>.\u201d Bennet\u2019s editorial decision was not merely ill-considered, but literally life-threatening (note pundits in the space of a few weeks have told us that <a href=\"https:\/\/www.theatlantic.com\/health\/archive\/2020\/04\/why-georgia-reopening-coronavirus-pandemic\/610882\/\">protesting during lockdowns<\/a> and <em><a href=\"https:\/\/twitter.com\/JenniferNuzzo\/status\/1267885076697812993\">not<\/a><\/em> <a href=\"https:\/\/twitter.com\/JenniferNuzzo\/status\/1267885076697812993\">protesting<\/a> <a href=\"https:\/\/theintercept.com\/2020\/06\/11\/systemic-racism-protests-coronavirus-health-risk\/\">during lockdowns<\/a> are both literally lethal). The <em>Times<\/em> first attempted to rectify the situation by apologizing, <a href=\"https:\/\/www.nytimes.com\/2020\/06\/03\/opinion\/tom-cotton-protests-military.html\">adding a long Editor\u2019s note<\/a> to Cotton\u2019s piece that read, as so many recent \u201capologies\u201d have, like a note written by a hostage.<\/p><p>Editors begged forgiveness for not being more involved, for not thinking to urge Cotton to sound less like Cotton (\u201cEditors should have offered suggestions\u201d), and for allowing rhetoric that was \u201cneedlessly harsh and falls short of the thoughtful approach that advances useful debate.\u201d That last line is sadly funny, in the context of an episode in which reporters were seeking to pre-empt a debate rather than have one at all; of course, no one got the joke, since a primary characteristic of the current political climate is a total absence of a sense of humor in any direction.<\/p><p>As many guessed, the \u201capology\u201d was not enough, and <a href=\"https:\/\/twitter.com\/NYTimesPR\/status\/1269722985751027713\">Bennet was whacked<\/a> a day later in a terse announcement.<\/p><p>His replacement, Kathleen Kingsbury, issued a staff directive essentially telling employees they now had a veto over <a href=\"https:\/\/www.nytimes.com\/2020\/06\/07\/business\/media\/james-bennet-resigns-nytimes-op-ed.html?referringSource=articleShare\">anything that made them uncomfortable<\/a>: \u201cAnyone who sees any piece of Opinion journalism, headlines, social posts, photos\u2014you name it\u2014that gives you the slightest pause, please call or text me immediately.\u201d<\/p><p>All these episodes sent a signal to everyone in a business already shedding jobs at an extraordinary rate that failure to toe certain editorial lines can and will result in the loss of your job. Perhaps additionally, you could face a public shaming campaign in which you will be denounced as a racist and rendered unemployable.<\/p><p>These tensions led to amazing contradictions in coverage. For all the extraordinary\/inexplicable scenes of police viciousness in recent weeks \u2014 and there was a ton of it, ranging from police <a href=\"https:\/\/twitter.com\/MotherJones\/status\/1270094764398964737\">slashing tires<\/a> in Minneapolis, to Buffalo officers <a href=\"https:\/\/twitter.com\/waltisfrozen\/status\/1268729705722568704\">knocking over<\/a> an elderly man, to <a href=\"https:\/\/twitter.com\/search?q=cops%20violence&amp;src=typed_query&amp;f=video\">Philadelphia police attacking protesters<\/a> \u2014 there were also <a href=\"https:\/\/www.forbes.com\/sites\/jemimamcevoy\/2020\/06\/03\/at-least-12-have-died-in-nationwide-protests-heres-who-they-were\/#2abcd9a99b32\">12 deaths in the first nine days<\/a> of protests, only one at the hands of a police officer (involving a man who may or may not have been aiming a gun at police).<\/p><p>Looting in some communities has been so bad that people have been left without banks to cash checks, or pharmacies to fill prescriptions; business owners have been wiped out (\u201c<a href=\"https:\/\/www.wsj.com\/articles\/looting-is-second-blow-for-reeling-businesses-especially-in-minority-neighborhoods-11591214595\">My life is gone<\/a>,\u201d commented one Philly store owner); a car dealership in San Leandro, California saw <a href=\"https:\/\/www.ktvu.com\/news\/more-than-70-cars-stolen-from-san-leandro-dodge-dealership-during-looting-spree\">74 cars stolen<\/a> in a single night. It isn\u2019t the whole story, but it\u2019s demonstrably true that violence, arson, and rioting are occurring.<\/p><p>However, because it is politically untenable to discuss this in ways that do not suggest support, reporters have been twisting themselves into knots. We are seeing headlines previously imaginable only in <em>The Onion,<\/em> e.g., \u201c<a href=\"https:\/\/twitter.com\/BBCNews\/status\/1269574979680702470\">27 police officers injured during largely peaceful anti-racism protests in London<\/a>.\u201d<\/p><p>Even people who try to keep up with protest goals find themselves denounced the moment they fail to submit to some new tenet of ever-evolving doctrine, via a surprisingly consistent stream of retorts: <em>fuck you, shut up, send money, do better, check yourself, I\u2019m tired<\/em> and <em>racist<\/em>.<\/p><p>Minneapolis mayor Jacob Frey, who argued for police reform and attempted to show solidarity with protesters in his city, was <a href=\"https:\/\/twitter.com\/LPDonovan\/status\/1269432888564551680\">shouted down<\/a> after he refused to commit to defunding the police. Protesters shouted \u201cGet the fuck out!\u201d at him, then chanted \u201c<em>Shame<\/em>!\u201d and threw refuse, <em><a href=\"https:\/\/www.youtube.com\/watch?v=M-9u6msqJNo\">Game of Thrones<\/a><\/em><a href=\"https:\/\/www.youtube.com\/watch?v=M-9u6msqJNo\">-style<\/a>, as he skulked out of the gathering. Frey\u2019s \u201cshame\u201d was refusing to endorse a position polls show <a href=\"https:\/\/docs.cdn.yougov.com\/s23agrrx47\/20200531_yahoo_race_and_justice_crosstabs.pdf\">65% of Americans oppose<\/a>, including 62% of Democrats, with just 15% of all people, and only 33% of African-Americans, in support.<\/p><p>Each passing day sees more scenes that recall something closer to cult religion than politics. White protesters in Floyd\u2019s Houston hometown <a href=\"https:\/\/indianexpress.com\/article\/trending\/trending-globally\/george-floyd-kneel-viral-video-6438938\/\">kneeling and praying<\/a> to black residents for \u201cforgiveness\u2026 for years and years of racism\u201d are one thing, but what are we to make of white police in Cary, North Carolina, kneeling and <em><a href=\"https:\/\/au.news.yahoo.com\/scenes-us-protests-police-wash-feet-community-leaders-065318936.htm\">washing the feet<\/a><\/em> of Black pastors? What about Nancy Pelosi and Chuck Schumer kneeling while dressed in \u201c<a href=\"https:\/\/thehill.com\/homenews\/house\/501633-pelosi-schumer-kneel-in-silence-for-almost-9-minutes-to-honor-george-floyd\">African kente cloth scarves<\/a>\u201d?<\/p><p>There is symbolism here that goes beyond frustration with police or even with racism: these are orgiastic, quasi-religious, and most of all, deeply weird scenes, and the press is too paralyzed to wonder at it. In a business where the first job requirement was once the willingness to ask tough questions, we\u2019ve become afraid to ask obvious ones.<\/p><p>On CNN, Minneapolis City Council President Lisa Bender was asked a hypothetical question about a future without police: \u201cWhat if in the middle of the night, my home is broken into? Who do I call?\u201d When Bender, who is white, <a href=\"https:\/\/www.youtube.com\/watch?v=idw83kGmwzw&amp;t=29s\">answered<\/a>, \u201cI know that comes from a place of privilege,\u201d questions popped to mind. Does privilege mean one should let someone break into one\u2019s home, or that one shouldn\u2019t ask that hypothetical question? (I was genuinely confused). In any other situation, a media person pounces on a provocative response to dig out its meaning, but an increasingly long list of words and topics are deemed too dangerous to discuss.<\/p><p>The media in the last four years has devolved into a succession of moral manias. We are told the Most Important Thing Ever is happening for days or weeks at a time, until subjects are abruptly dropped and forgotten, but the tone of warlike emergency remains: from James Comey\u2019s firing, to the deification of Robert Mueller, to the Brett Kavanaugh nomination, to the democracy-imperiling threat to intelligence \u201cwhistleblowers,\u201d all those interminable months of Ukrainegate hearings (while Covid-19 advanced), to fury at the death wish of lockdown violators, to the sudden reversal on that same issue, etc.<\/p><p>It\u2019s been learned in these episodes we may freely misreport reality, so long as the political goal is righteous. It was okay to publish the now-discredited Steele dossier, because Trump is scum. MSNBC could put Michael Avenatti <a href=\"https:\/\/www.youtube.com\/watch?v=wK3DWWjlhaI\">on live TV<\/a> to air a gang rape allegation without vetting, because who cared about Brett Kavanaugh \u2013 except press airing of that wild story ended up being a crucial factor in convincing key swing voter Maine Senator Susan Collins the anti-Kavanaugh campaign was a political hit job (the allegation illustrated, \u201cwhy the presumption of innocence is so important,\u201d <a href=\"https:\/\/www.washingtonpost.com\/politics\/2018\/10\/05\/did-michael-avenatti-help-doom-case-against-brett-kavanaugh\/\">she said<\/a>). Reporters who were anxious to prevent Kavanaugh\u2019s appointment, in other words, ended up helping it happen through overzealousness.<\/p><p>There were no press calls for self-audits after those episodes, just as there won\u2019t be a few weeks from now if Covid-19 cases spike, or a few months from now if Donald Trump wins re-election successfully painting the Democrats as supporters of violent protest who want to abolish police. No: press activism is limited to denouncing and shaming colleagues for insufficient fealty to the cheap knockoff of bullying campus Marxism that passes for leftist thought these days.<\/p><p>The traditional view of the press was never based on some contrived, mathematical notion of \u201cbalance,\u201d i.e. five paragraphs of Republicans for every five paragraphs of Democrats. The ideal instead was that we showed you everything we could see, good and bad, ugly and not, trusting that a better-informed public would make better decisions. This vision of media stressed accuracy, truth, and trust in the reader\u2019s judgment as the routes to positive social change.<\/p><p>For all our infamous failings, journalists once had some toughness to them. We were supposed to be willing to go to jail for sources we might not even like, and fly off to war zones or disaster areas without question when editors asked. It was also once considered a virtue to flout the disapproval of colleagues to fight for stories we believed in (Watergate, for instance).<\/p><p>Today no one with a salary will stand up for colleagues like Lee Fang. Our brave truth-tellers make great shows of shaking fists <a href=\"https:\/\/www.youtube.com\/watch?v=zdFe-LmFRV8\">at our parody president<\/a>, but not one of them will talk honestly about the fear running through their own newsrooms. People depend on us to tell them what we see, not what we think. What good are we if we\u2019re afraid to do it?<\/p><ul><li>\n<p>I\u2019m aware of <a href=\"https:\/\/twitter.com\/RottenInDenmark\/status\/1271197002353610752\">this tweet<\/a> suggesting the reasons for Shor\u2019s firing are unknown. I stand by the characterization made in the piece.<\/p>\n<\/li>\n<\/ul>","created_at":"2020-06-14T13:20:44+0000","updated_at":"2020-06-14T13:20:44+0000","published_at":"2020-06-12T22:02:12+0000","published_by":["Matt Taibbi"],"annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"taibbi.substack.com","preview_picture":"https:\/\/cdn.substack.com\/image\/fetch\/h_600,c_limit,f_auto,q_auto:good,fl_progressive:steep\/https%3A%2F%2Fbucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com%2Fpublic%2Fimages%2Ff1bcd347-c525-4a12-a0b8-5cd7c99f4173_5203x3325.jpeg","http_status":"200","origin_url":"https:\/\/taibbi.substack.com\/p\/the-news-media-is-destroying-itself?r=2&utm_campaign=post&utm_medium=web&utm_source=hackernews"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1176,"title":"plexus\/chemacs","url":"https:\/\/github.com\/plexus\/chemacs","content":"<pre>     ___           ___           ___           ___           ___           ___           ___\n    \/  \/\\         \/__\/\\         \/  \/\\         \/__\/\\         \/  \/\\         \/  \/\\         \/  \/\\\n   \/  \/:\/         \\  \\:\\       \/  \/:\/_       |  |::\\       \/  \/::\\       \/  \/:\/        \/  \/:\/_\n  \/  \/:\/           \\__\\:\\     \/  \/:\/ \/\\      |  |:|:\\     \/  \/:\/\\:\\     \/  \/:\/        \/  \/:\/ \/\\\n \/  \/:\/  ___   ___ \/  \/::\\   \/  \/:\/ \/:\/_   __|__|:|\\:\\   \/  \/:\/~\/::\\   \/  \/:\/  ___   \/  \/:\/ \/::\\\n\/__\/:\/  \/  \/\\ \/__\/\\  \/:\/\\:\\ \/__\/:\/ \/:\/ \/\\ \/__\/::::| \\:\\ \/__\/:\/ \/:\/\\:\\ \/__\/:\/  \/  \/\\ \/__\/:\/ \/:\/\\:\\\n\\  \\:\\ \/  \/:\/ \\  \\:\\\/:\/__\\\/ \\  \\:\\\/:\/ \/:\/ \\  \\:\\~~\\__\\\/ \\  \\:\\\/:\/__\\\/ \\  \\:\\ \/  \/:\/ \\  \\:\\\/:\/~\/:\/\n \\  \\:\\  \/:\/   \\  \\::\/       \\  \\::\/ \/:\/   \\  \\:\\        \\  \\::\/       \\  \\:\\  \/:\/   \\  \\0.1 \/:\/\n  \\  \\:\\\/:\/     \\  \\:\\        \\  \\:\\\/:\/     \\  \\:\\        \\  \\:\\        \\  \\:\\\/:\/     \\__\\\/ \/:\/\n   \\  \\::\/       \\  \\:\\        \\  \\::\/       \\  \\:\\        \\  \\:\\        \\  \\::\/        \/__\/:\/\n    \\__\\\/         \\__\\\/         \\__\\\/         \\__\\\/         \\__\\\/         \\__\\\/         \\__\\\/\n                                 ... ........,**********,...\n                          .%*....       .,\/(((#\/\/(\/\/*,*,,,*.,#%&amp;(,,\/(##.\n                            %\/..                          ,,  ,*,*\/(%%.\n                             %\/*.                         .. .,,.,#%&amp;\n                              &amp;(,.                       .. ....((%&amp;.\n                               ,#,.                     ..  ...*\/#%\n                                .#,..                   .  ..,\/(%#\n                                 .#,.                  ..  .,,(#%\n                                   #*,.               .. .,,.(%(\n                                  ***\/\/((((((((\/\/****,,**\/\/,\/  .\n                                  \/\/\/((#%%%%%%%##(\/\/\/\/\/\/\/*.\/,,,,\n                                   \/#(##%%%%%%%%%##(((\/\/*.(**,*\n                                    .(\/###%%%%%%##(((\/\/\/ (\/,,,\n                                     *\/(##%%%%##(((\/((\/(\/*.\n                                     *((###%%%%%%%%%%%&amp;&amp;(*..\n                                     .((###%%%%%%%%##((####\/,\/\/\n                                     **\/\/((###(###%%%#%%#&amp;\/%(%#(.\n                                    **\/\/(((((((\/((##%#%#\/%##%*\/.\n                                   ***\/\/\/(\/(((\/\/\/\/\/((((%**..\/..   \/\/\n                                  **\/\/(((###(((((\/\/\/\/\/\/%\/***(,..   \/&amp;\n                                  ,####%%%%%%%%###((\/\/(%,*\/((\/**    #*\n                                  .\/,,                *%..,..*&amp;%    ,#\n                                 ,(,.                 *%   ...\/%&amp;    @\n                                *#,.                  *%,  ....,%&amp;   &amp;.\n                               (\/,                   .&amp;&amp;&amp;%\/   ..,%&amp;  %,\n                              %\/,                        ((.   .,,%&amp;%%\/,\n                             #*.                                .*(%%%@,\n                            #,.                                 ..,,%&amp;.,\n                           ,\/....                              .... ,%%\n                           %,,.....                  ......   ,* ....#&amp;.\n                           #,,,,..      .....................  ....,\/#%*\n                           #*,,*.......  ...................,,,****\/(%%,\n                           ,(*\/**,.......................,,,,,**\/\/###%#\n                            \/#\/(((\/**,...  .............,,,,\/(##%%%##@\n                             \/.*\/((\/,,,,,.,\/(###########(\/(((((#%%#&amp;&amp;\n                               ,(#(\/,,.......,,,,,,,,\/(\/\/\/(#%%%%#\n                                .\/(*,..                  ..,*((#&amp;\/\n                                    ...,..,..,\/###(\/\/(#(\/*,.. .\n<\/pre><p>Chemacs is an Emacs profile switcher, it makes it easy to run multiple Emacs configurations side by side.<\/p><p>Think of it as a bootloader for Emacs.<\/p><h2><a id=\"user-content-rationale\" class=\"anchor\" aria-hidden=\"true\" href=\"#rationale\"><\/a>Rationale<\/h2><p>Emacs configuration is either kept in a <code>~\/.emacs<\/code> file or, more commonly, in a <code>~\/.emacs.d<\/code> directory. These paths are hard-coded. If you want to try out someone else\u2019s configuration, or run different distributions like Prelude or Spacemacs, then you either need to swap out <code>~\/.emacs.d<\/code>, or run Emacs with a different <code>$HOME<\/code> directory set.<\/p><p>This last approach is quite common, but has some real drawbacks, since now packages will no longer know where your actual home directory is.<\/p><p>All of these makes trying out different Emacs configurations and distributions needlessly cumbersome.<\/p><p>Various approaches to solving this have been floated over the years. There\u2019s an Emacs patch around that adds an extra command line option, and various examples of how to add a command line option in userspace from Emacs Lisp.<\/p><p>Chemacs tries to implement this idea in a user-friendly way, taking care of the various edge cases and use cases that come up.<\/p><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>Clone the Git repository, and run <code>install.sh<\/code><\/p><div class=\"highlight highlight-source-shell\"><pre>$ git clone https:\/\/github.com\/plexus\/chemacs.git\n$ cd chemacs\n$ .\/install.sh\nOK      Creating symlink ~\/.emacs -&gt; \/home\/arne\/chemacs\/.emacs<\/pre><\/div><p>The install script will symlink <code>~\/.emacs<\/code> to the Chemacs script. If you already have a <code>~\/.emacs<\/code> you need to move it out of the way first<\/p><div class=\"highlight highlight-source-shell\"><pre>$ .\/install.sh\nWARN    chemacs can't be installed, ~\/.emacs is in the way<\/pre><\/div><p>By symlinking you can easily update Chemacs with a <code>git pull<\/code>. Chemacs is not available on ELPA\/MELPA because its special position in the Emacs boot process would cause a chicken and egg problem.<\/p><p>There\u2019s a similar script for Windows called <code>install.ps1<\/code>.<\/p><div class=\"highlight highlight-source-powershell\"><pre>$ git clone https:\/\/github.com\/plexus\/chemacs.git\n$ cd chemacs\n$ .\\install.ps1\nOK copying file to ~\/.emacs<\/pre><\/div><p>The Windows script copies the ~\/.emacs file rather than symlinking it, as symlinks in Windows require elevated privileges. To update the file, run <code>install.ps1<\/code> a second time, if necessary - it uses hashing to detect changes and prompts for conformation before overwriting<\/p><div class=\"highlight highlight-source-powershell\"><pre>$ .\\install.ps1\nWARN content is different between C:\\Users\\Me\\.emacs and C:\\Users\\Me\\repos\\chemacs\\.emacs\nWARN chemacs may already be installed OR something else may be in the way\nDo you want to overwrite C:\\Users\\Me\\.emacs?: y\nOK updated chemacs files successfully.<\/pre><\/div><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><p>Chemacs adds an extra command line option to Emacs, <code>--with-profile<\/code>. Profiles are configured in <code>~\/.emacs-profiles.el<\/code>. If this file does not exist Chemacs will create it with a default profile pointing at <code>~\/.emacs.d<\/code>.<\/p><p>If no profile is given at the command line then the default profile is used, so if you currently have your emacs configuration in <code>~\/.emacs.d<\/code> then Chemacs will, by default, use your existing configuration.<\/p><div class=\"highlight highlight-source-shell\"><pre>$ emacs --with-profile my-profile<\/pre><\/div><h2><a id=\"user-content-emacs-profilesel\" class=\"anchor\" aria-hidden=\"true\" href=\"#emacs-profilesel\"><\/a>.emacs-profiles.el<\/h2><p>This file contains an association list, with the keys\/cars being the profile names, and the values\/cdrs their configuration.<\/p><p>The main thing to configure is the <code>user-emacs-directory<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>((\"default\" . ((user-emacs-directory . \"~\/.emacs.d\")))\n (\"spacemacs\" . ((user-emacs-directory . \"~\/spacemacs\"))))<\/pre><\/div><p>Chemacs will set this to be the <code>user-emacs-directory<\/code> in use, and load <code>init.el<\/code> from that directory.<\/p><p>Other things you can configure<\/p><ul><li><code>custom-file<\/code> : The file where Customize stores its customizations. If this isn\u2019t configured, and the <code>custom-file<\/code> variable is still unset after loading the profile\u2019s <code>init.el<\/code>, then this will get set to the profile\u2019s <code>init.el<\/code><\/li>\n<li><code>server-name<\/code> : Sets the <code>server-name<\/code> variable, so you can distinguish multiple instances with <code>emacsclient -s &lt;server-name&gt;<\/code>.<\/li>\n<li><code>env<\/code> An association list of environment variables. These will get set before loading the profile, so they can influence the initialization, and they are visible to any subprocesses spawned from Emacs.<\/li>\n<li><code>straight-p<\/code> Enable the <a href=\"https:\/\/github.com\/raxod502\/straight.el\">Straight<\/a> functional package manager.<\/li>\n<\/ul><p>Store <code>.emacs-profiles.el<\/code> together with your dotfiles. If you\u2019re not yet keeping a version controlled directory of dotfiles, then check out <a href=\"https:\/\/github.com\/plexus\/dotfiles\/blob\/master\/connect-the-dots\">connect-the-dots<\/a> for a helpful script to do that.<\/p><h2><a id=\"user-content-changing-the-default-profile-eg-for-gui-editors\" class=\"anchor\" aria-hidden=\"true\" href=\"#changing-the-default-profile-eg-for-gui-editors\"><\/a>Changing the default profile (e.g. for GUI editors)<\/h2><p>Where it is not possible to use the <code>--with-profile<\/code> flag, the default profile can be set using a <code>~\/.emacs-profile<\/code> file.<\/p><p>If your <code>~\/.emacs-profiles.el<\/code> file contains the following:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>((\"default\" . ((user-emacs-directory . \"~\/.emacs.d\")))\n (\"spacemacs\" . ((user-emacs-directory . \"~\/spacemacs\")))\n (\"prelude\" . ((user-emacs-directory . \"~\/prelude\"))))<\/pre><\/div><p>you can create a file called <code>~\/.emacs-profile<\/code>, containing the name of the profile you\u2019d like to be used when none is given on the command line:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ echo 'spacemacs' &gt; ~\/.emacs-profile<\/pre><\/div><p>This will set the default profile to be the \u201cspacemacs\u201d profile, instead of \u201cdefault\u201d. You can change the default by simply changing the contents of this file:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ echo 'prelude' &gt; ~\/.emacs-profile<\/pre><\/div><p>If this file doesn\u2019t exist, then \u201cdefault\u201d will be used, as before.<\/p><h2><a id=\"user-content-spacemacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#spacemacs\"><\/a>Spacemacs<\/h2><p>Spacemacs is typically installed by cloning the Spacemacs repo to <code>~\/.emacs.d<\/code>, and doing extra customization from <code>~\/.spacemacs<\/code> or <code>~\/.spacemacs.d\/init.el<\/code>. This makes it tedious to switch between version of Spacemacs, or between different Spacemacs configurations.<\/p><p>With Chemacs you can point your <code>user-emacs-directory<\/code> to wherever you have Spacemacs installed, and use the <code>SPACEMACSDIR<\/code> environment variable to point at a directory with customizations that are applied on top of the base install.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>((\"spacemacs\" . ((user-emacs-directory . \"~\/spacemacs\")\n                 (env . ((\"SPACEMACSDIR\" . \"~\/.spacemacs.d\")))))\n (\"spacemacs-develop\" . ((user-emacs-directory . \"~\/spacemacs\/develop\")\n                        (env . ((\"SPACEMACSDIR\" . \"~\/.spacemacs.d\")))))\n (\"new-config\" . ((user-emacs-directory . \"~\/spacemacs\/develop\")\n                  (env . ((\"SPACEMACSDIR\" . \"~\/my-spacemacs-config\"))))))<\/pre><\/div><h2><a id=\"user-content-doom-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#doom-emacs\"><\/a>DOOM emacs<\/h2><p>You can add an entry similar to the following to your <code>.emacs-profiles.el<\/code><\/p><p>In the following snippet <code>~\/doom-emacs<\/code> is where you have cloned doom emacs.<\/p><p>(Depending on when you read this) <code>DOOMDIR<\/code> support is only in <code>develop<\/code> branch of doom emacs. Check commit history of <code>master<\/code> branch of doom emacs<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(\"doom\" . ((user-emacs-directory . \"~\/doom-emacs\")\n           (env . ((\"DOOMDIR\" . \"~\/doom-config\")))))<\/pre><\/div><p>Please refer to <a href=\"https:\/\/github.com\/plexus\/chemacs\/issues\/5\">this<\/a> discussion for details.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>LICENSE<\/h2><p>Copyright \u00a9 Arne Brasseur 2018-2020<\/p><p>Distributed under the terms of the GPL v3.<\/p>","created_at":"2020-06-02T02:10:02+0000","updated_at":"2020-06-02T02:10:02+0000","published_by":["plexus"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/32212?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1175,"title":"rougier\/elegant-emacs","url":"https:\/\/github.com\/rougier\/elegant-emacs","content":"<p>The goal is to have a minimal consistent vanilla emacs (i.e. no package) and yet elegant.<br \/>The defaul font is <a href=\"https:\/\/fonts.google.com\/specimen\/Roboto+Mono\" rel=\"nofollow\">Roboto Mono<\/a> and optionally, you'll also need the <a href=\"https:\/\/fonts.google.com\/specimen\/Fira+Code\" rel=\"nofollow\">Fira Code<\/a> font.<\/p><p><strong>Usage<\/strong>: <code>emacs -q -l sanity.el -l elegance.el Splash.org &amp;<\/code><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-0.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-0.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-1.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-1.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-2.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-2.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-3.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-3.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-4.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-4.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-5.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-5.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-6.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-6.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-7.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-7.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-8.png\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-8.png\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-9.png#0\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-9.png#0\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-10.png#0\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-10.png#0\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rougier\/elegant-emacs\/blob\/master\/screenshots\/screenshot-11.png#0\"><img src=\"https:\/\/github.com\/rougier\/elegant-emacs\/raw\/master\/screenshots\/screenshot-11.png#0\" width=\"45%\" class=\"c4\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/p>","created_at":"2020-06-01T23:55:42+0000","updated_at":"2020-06-01T23:55:42+0000","published_by":["rougier"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/repository-images.githubusercontent.com\/262980922\/3f74e600-993a-11ea-8bb3-a401dd5d009f","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1174,"title":"Build software better, together","url":"https:\/\/github.com\/unifydb\/unifydb\/settings\/installations","content":"Page not found \u00b7 GitHub \u00b7 GitHub\n<div class=\"application-main d-flex flex-auto flex-column\" data-commit-hovercards-enabled=\"\"><main>\n<div class=\"container-lg mt-5 px-3\"><form role=\"search\" action=\"\/search\" accept-charset=\"UTF-8\" method=\"get\">\n<p><label for=\"not-found-search\" class=\"d-block text-normal text-gray mb-1 f4\">Find code, projects, and people on GitHub:<\/label><\/p>\n<\/form><p><a href=\"https:\/\/github.com\/contact\" class=\"link-gray\">Contact Support<\/a> \u2014 <a href=\"https:\/\/githubstatus.com\" class=\"link-gray\">GitHub Status<\/a> \u2014 <a href=\"https:\/\/twitter.com\/githubstatus\" class=\"link-gray\">@githubstatus<\/a><\/p><\/div>\n<\/main><\/div>\n<p> You can\u2019t perform that action at this time.<\/p>\n<p> You signed in with another tab or window. <a href=\"https:\/\/github.com\/unifydb\/unifydb\/settings\/installations\">Reload<\/a> to refresh your session. You signed out in another tab or window. <a href=\"https:\/\/github.com\/unifydb\/unifydb\/settings\/installations\">Reload<\/a> to refresh your session.<\/p>\n<details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark hx_rsm\" open=\"open\">\n<\/details>","created_at":"2020-05-31T19:27:45+0000","updated_at":"2020-05-31T19:27:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/github.githubassets.com\/images\/modules\/open_graph\/github-logo.png","http_status":"404"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1173,"title":"Blog: The Crux 'Document Store'","url":"https:\/\/juxt.pro\/blog\/posts\/crux-doc-store.html","content":"<div class=\"paragraph\"><p>We\u2019ve always aimed for Crux to be 'unbundled' - giving users the choice of which underlying storage engines to use for their specific deployments. Whether it be Apache Kafka, Facebook\u2019s RocksDB or the many supported JDBC backends, Crux users benefit from decades of development effort put into the stability and performance of these data stores.<\/p><\/div><div class=\"paragraph\"><p>We have been busy updating the core components of Crux to better reflect what Crux needs from these storage engines, and to make the most of what these engines can provide.<\/p><\/div><div class=\"paragraph\"><p>Crux now supports the concept of a pluggable <em>document store<\/em>, which is the central location for master document storage in Crux. You are still able to use an event log such as Kafka to hold all your data, including documents, but you can also use JDBC-based document stores or more cost-effective object storage services like S3. The choice is yours!<\/p><\/div><div class=\"paragraph\"><p>Since <a href=\"https:\/\/github.com\/juxt\/crux\/releases\/tag\/20.05-1.8.3-alpha\"><code>20.05-1.8.3-alpha<\/code><\/a> Crux now has both 'transaction log' and 'document store' protocols:<\/p><\/div><div class=\"listingblock content\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defprotocol TxLog\n  (submit-tx [this tx-events])\n  (open-tx-log [this after-tx-id]))\n(defprotocol DocumentStore\n  (submit-docs [this docs])\n  (fetch-docs [this ids]))<\/pre><\/div><div class=\"paragraph\"><p>It is now possible, for example, to use Kafka as the <code>TxLog<\/code>, and a JDBC database as the <code>DocumentStore<\/code>, taking advantage of the performance and scaling characteristics of both.<\/p><\/div><div class=\"paragraph\"><p>Alongside this change we also released an <a href=\"https:\/\/github.com\/juxt\/crux\/blob\/master\/crux-s3\">S3 document store implementation<\/a> for users deploying Crux within AWS. This module demonstrates how to implement the protocol and we are excited to see a variety of alternative <code>DocumentStore<\/code> implementations emerge in the community for other popular storage systems over the coming months!<\/p><\/div><div class=\"paragraph\"><p><img src=\"https:\/\/juxt.pro\/img\/crux-document-store-arch-before-after.png\" class=\"lazyload lazyautosizes\" data-sizes=\"auto\" alt=\"Before &amp; After Crux Architecture\" referrerpolicy=\"no-referrer\" \/><\/p><\/div><div class=\"paragraph\"><p>For more background on why we made this change, let\u2019s start with how Crux works when processing transactions\u2026<\/p><\/div>","created_at":"2020-05-19T14:42:16+0000","updated_at":"2020-05-19T14:42:16+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"juxt.pro","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1172,"title":"A Dynamic Warm-Up That Takes Less Than 5 Minutes (and It\u2019s Proven to Work)","url":"https:\/\/www.runnersworld.com\/health-injuries\/a20809473\/five-exercises-to-do-before-every-run\/","content":"<p class=\"body-text\">Look, we\u2019re all guilty of doing a few toe touches or a quick quad stretch before a run and counting that as the warm-up. And while that might work for most people most of the time, it\u2019s not the best way to get your body prepped to run your best for the miles ahead. In fact, a recent study shows that a dynamic warm-up routine can help you perform better.<\/p><p class=\"body-text\">In the study, researchers compared how well study participants ran after moving versus sitting. When runners did dynamic stretches, they were able to go almost two and a half minutes longer before they tired out compared to when they sat. The following five dynamic warm-up exercises, demonstrated by Roman Siromakha, a certified trainer based in New York City, are the exact same ones used in the study.<\/p><p class=\"body-text\"><strong>How to use this list<\/strong>: Perform each move below 10 times, moving through each rep quickly, before your run. The entire routine should take less than five minutes to complete.<\/p><h2 class=\"body-h3\">1. Hip Flexor Warm-Up<\/h2><p class=\"body-text\">Start standing tall. Flex your hip by drawing left knee up toward chest as you swing your right arm forward (as if you are running). Return to starting position and repeat with right knee and left arm. That\u2019s 1 rep.<br \/><\/p><h2 class=\"body-h3\">2. Leg Flexor Stretch<\/h2><p class=\"body-text\">Stand tall. Draw left knee toward chest until thigh is parallel to the ground, as you simultaneously swing your right arm forward and left arm back (as if you are running). Engage your quad to extend left leg straight out. Return to standing, then repeat with the other leg. That\u2019s 1 rep.<\/p><h2 class=\"body-h3\">3. \u200bLeg Extensor Stretch<\/h2><p class=\"body-text\">Start standing. Slowly bend left knee to bring left heel behind you to glute as you swing your right arm forward and left arm back (as if you are running). You should feel your left hamstring engage. Return to starting position, then repeat on the other leg. That\u2019s 1 rep.<br \/><\/p><h2 class=\"body-h3\">\u200bPlantar Flexor Stretch<\/h2><p class=\"body-text\">Stand with your hands on your hips. Lift your left foot a few inches, keeping your knee straight. Quickly flex your foot, pulling toes upward to shin and pointing them down. Return to standing, then repeat with the other foot. That\u2019s 1 rep.<\/p><h2 class=\"body-h3\">\u200bHip Extensor Stretch<\/h2><p class=\"body-text\">From standing, hinge forward at your hips. Draw left knee up toward chest while bringing right arm forward. From there, maintain the same lean as you quickly kick left leg back behind you while you simultaneously swing left arm forward and right arm back (as if you are running). Return right knee in front of you and repeat for 10 reps. Then repeat with other leg.<\/p><hr \/><p class=\"body-text\">Images: <a class=\"body-link\" href=\"https:\/\/www.juliahembreephoto.com\/\" target=\"_blank\" rel=\"nofollow\" data-vars-ga-outbound-link=\"https:\/\/www.juliahembreephoto.com\/\">Julia Hembree Smith<\/a><\/p><p> Contributing Writer <a href=\"http:\/\/www.cindykuzma.com\/\" data-vars-ga-outbound-link=\"http:\/\/www.cindykuzma.com\/\">Cindy<\/a> is a freelance health and fitness writer, author, and podcaster who\u2019s contributed regularly to Runner\u2019s World since 2013.<\/p><p>This content is created and maintained by a third party, and imported onto this page to help users provide their email addresses. You may be able to find more information about this and similar content at piano.io<\/p>","created_at":"2020-05-17T12:02:36+0000","updated_at":"2020-05-17T12:02:36+0000","published_at":"2018-12-17T09:07:00+0000","published_by":["Runner's World Editors"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"www.runnersworld.com","preview_picture":"https:\/\/hips.hearstapps.com\/hmg-prod.s3.amazonaws.com\/images\/young-athlete-woman-exercises-bouncing-up-and-down-royalty-free-image-502581330-1545064661.jpg?crop=1.00xw:0.756xh;0,0.112xh&resize=1200:*","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1171,"title":"Sign in - Google Accounts","url":"https:\/\/accounts.google.com\/ServiceLogin?continue=https:\/\/accounts.google.com\/signin\/oauth\/legacy\/consent?authuser=unknown&part=AJi8hAM4PnAuykBkMUlXtNhE-5p-lY7eTeDcIUWY6-Q7vZSG_zQ5vInJbQyrV46XzF6VyGVLdcVNpC2ckHy5Z1ZJ1NT3kqposO2PLrwuCsMSLNCd4dlo53qtCsXJNWyrl6c-et5muTr8mAumIOBgVnJY07Pq99ILbTJZGLdu1lLHgBtHy6JZCZaQPfLA_CDK5cIJH0ugpKc1s3xkomB6zji4uXUhrLxEOdvxo8-Qai3V9yyCCPtYTJvwSw1dgiyRAjHO-j5HMnopmQW1LYqaLGF9ep7t82bwP9cdsglLoFo6jlwvIctwikMvMhxnVnBSlUWak69JuhWoFSQT0P816XMSheq3V4AicabKOyOoHiKdpZDahIG9gzO7jOZ_aTL9KzJXzSlXUl_Nq_r4WcuazFawDXWSNu-yT_tkxsG18oyNyepF8qllksCIw-U5-bNjSDyccAVWd0yK&as=mkqUlq4WI5Qo3Z78F-Jjvw&oauth=1&rip=1","content":"<label for=\"lang-chooser\"><img src=\"https:\/\/ssl.gstatic.com\/images\/icons\/ui\/common\/universal_language_settings-21.png\" alt=\"Change language\" referrerpolicy=\"no-referrer\" \/><\/label>","created_at":"2020-05-14T15:58:01+0000","updated_at":"2020-05-14T15:58:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"accounts.google.com","http_status":"200","origin_url":"https:\/\/bloblogs.ops.lola.com\/blobs\/production\/1216224186574308170\/travel-service_AMD_InformativePricing_2020-04-20T19:10:39.743938Z_8177514.json"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1170,"title":"Buried Treasure \u2013 Reviewing the indie games everyone else missed","url":"https:\/\/buried-treasure.org\/","content":"<div id=\"primary\" class=\"content-area\">\n<main id=\"main\" class=\"site-main\" role=\"main\"><div id=\"posts\">\n<article id=\"post-365\" class=\"sticky post-365 post type-post status-publish format-standard has-post-thumbnail hentry category-other-business tag-site-business\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2019\/12\/patreon-me-do\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2019\/12\/patreon_mark_coral.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2019\/12\/patreon_mark_coral.jpg 500w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2019\/12\/patreon_mark_coral-300x300.jpg 300w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2019\/12\/patreon_mark_coral-150x150.jpg 150w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1799\" class=\"post-1799 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure category-review tag-adventure tag-review tag-the-flower-collectors\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/the-flower-collectors\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_theflowercollectors-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_theflowercollectors-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_theflowercollectors-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_theflowercollectors-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1793\" class=\"post-1793 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure tag-alasdair-beckett-king\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/alasdair-beckett-kings-interactive-murder-mystery-whodidit\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_abk-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_abk-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_abk-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_abk-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1777\" class=\"post-1777 post type-post status-publish format-standard has-post-thumbnail hentry category-horror category-review category-short tag-horror tag-no-one-lives-under-the-lighthouse tag-review tag-sowoke-entertainment-bureau\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/no-one-lives-under-the-lighthouse\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_lighthouse-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_lighthouse-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_lighthouse-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_lighthouse-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1765\" class=\"post-1765 post type-post status-publish format-standard has-post-thumbnail hentry category-puzzle category-review tag-beard-envy tag-filament tag-puzzle tag-review\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/filament\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_filament-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_filament-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_filament-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_filament-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1747\" class=\"post-1747 post type-post status-publish format-standard has-post-thumbnail hentry category-story-driven tag-no-more-robots tag-review tag-yes-your-grace\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/yes-your-grace\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_yesyourgrace-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_yesyourgrace-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_yesyourgrace-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_yesyourgrace-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1733\" class=\"post-1733 post type-post status-publish format-standard has-post-thumbnail hentry category-card-game category-deckbuilder category-review tag-card-game tag-deckbuilder tag-fabled-game tag-pirates-outlaws tag-review\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/pirates-outlaws\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_piratesoutlaws-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_piratesoutlaws-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_piratesoutlaws-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_piratesoutlaws-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1719\" class=\"post-1719 post type-post status-publish format-standard has-post-thumbnail hentry category-free category-review category-story-driven tag-free tag-hofmeier tag-review tag-type-dreams tag-typing\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/type-dreams\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_typedreams-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_typedreams-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_typedreams-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_typedreams-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1707\" class=\"post-1707 post type-post status-publish format-standard has-post-thumbnail hentry category-arcade category-pasta category-review tag-freddy-spaghetti tag-patrick-starr tag-playful-pasta tag-review\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/freddy-spaghetti\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_freddy_spaghetti-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_freddy_spaghetti-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_freddy_spaghetti-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/03_freddy_spaghetti-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1693\" class=\"post-1693 post type-post status-publish format-standard has-post-thumbnail hentry category-puzzle category-review tag-gc-games tag-match-3 tag-playcademy tag-review tag-spelunking\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/spelunking\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/06_spelunking-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/06_spelunking-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/06_spelunking-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/06_spelunking-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1677\" class=\"post-1677 post type-post status-publish format-standard has-post-thumbnail hentry category-uncategorized\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/treachery-in-beatdown-city\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/08_beatdown-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/08_beatdown-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/08_beatdown-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/08_beatdown-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1639\" class=\"post-1639 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure category-review category-story-driven tag-adventure tag-fmv tag-review tag-the-complex tag-wales-interactive\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/04\/the-complex\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_thecomplex-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_thecomplex-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_thecomplex-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/04\/00_thecomplex-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1623\" class=\"post-1623 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure category-puzzle category-review category-story-driven tag-adventure tag-coal-valley-games tag-oneiros tag-review\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/oneiros\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_oneiros-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_oneiros-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_oneiros-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_oneiros-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1591\" class=\"post-1591 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure category-review category-story-driven tag-adventure tag-machineboy tag-mattis-folkestad tag-milkmaid-of-the-milky-way tag-review\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/milkmaid-of-the-milky-way\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_milkmaid-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_milkmaid-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_milkmaid-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_milkmaid-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1563\" class=\"post-1563 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure category-preview tag-chinatown-detective-agency tag-general-interactive-co tag-preview\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/preview-chinatown-detective-agency\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_chinatown-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_chinatown-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_chinatown-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_chinatown-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1575\" class=\"post-1575 post type-post status-publish format-standard has-post-thumbnail hentry category-free category-review tag-draw-me-a-pixel tag-free tag-there-is-no-game\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/there-is-no-game\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/03_thereisnogame-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/03_thereisnogame-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/03_thereisnogame-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/03_thereisnogame-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1533\" class=\"post-1533 post type-post status-publish format-standard has-post-thumbnail hentry category-action-adventure category-adventure category-arcade category-free category-multiplayer category-platform category-puzzle category-short category-story-driven tag-abzu tag-alphabear tag-chuchel tag-epistory tag-fe tag-gnog tag-metamorphabet tag-subnautica\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/beautiful-intelligent-games-to-play-with-kids-during-the-lockdown\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/13_lockdown-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/13_lockdown-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/13_lockdown-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/13_lockdown-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1521\" class=\"post-1521 post type-post status-publish format-standard has-post-thumbnail hentry category-action-adventure category-demo category-free tag-2pt-interactive tag-demo tag-free tag-heavenly-bodies\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/heavenly-bodies-demo\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/05_heavenlybodies-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/05_heavenlybodies-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/05_heavenlybodies-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/05_heavenlybodies-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1505\" class=\"post-1505 post type-post status-publish format-standard has-post-thumbnail hentry category-action-adventure category-platform category-review tag-andrey-fomin tag-ascii tag-asciident tag-black-maple-games tag-review tag-uab-puzzle-lab\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/asciident-early-access\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/04_asciident-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/04_asciident-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/04_asciident-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/04_asciident-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1491\" class=\"post-1491 post type-post status-publish format-standard has-post-thumbnail hentry category-action-adventure category-arcade category-demo tag-demo tag-free tag-monolith-of-minds tag-resolutiion\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/resolutiion-demo-is-incredible\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_resolutiion-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_resolutiion-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_resolutiion-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/00_resolutiion-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><article id=\"post-1479\" class=\"post-1479 post type-post status-publish format-standard has-post-thumbnail hentry category-adventure category-free category-review tag-adventure tag-free tag-how-to-open-a-locked-chest tag-pirate-chip tag-review\"><a class=\"post-thumbnail-link\" href=\"https:\/\/buried-treasure.org\/2020\/03\/how-to-open-a-locked-chest\/\"><img width=\"280\" height=\"280\" src=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/07_lockedchest-280x280.jpg\" class=\"attachment-bari-standard size-bari-standard wp-post-image\" alt=\"\" srcset=\"https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/07_lockedchest-280x280.jpg 280w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/07_lockedchest-150x150.jpg 150w, https:\/\/buried-treasure.org\/wp-content\/uploads\/2020\/03\/07_lockedchest-600x600.jpg 600w\" referrerpolicy=\"no-referrer\" \/><\/a>\n<\/article><\/div>\n<\/main><\/div>","created_at":"2020-04-29T19:24:21+0000","updated_at":"2020-04-29T19:24:21+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"buried-treasure.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1169,"title":"Books I recommend to my software engineering students","url":"http:\/\/web.eecs.utk.edu\/~azh\/blog\/booksformystudents.html","content":"<br \/><h3>Books I recommend to my software engineering students<\/h3><small>2\/29\/2020<\/small><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/recommendedbooks.jpg\" class=\"center-block c11\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p><em>Update 4\/23: See the discussion of this post on <a href=\"https:\/\/news.ycombinator.com\/item?id=22962869\">Hacker News<\/a>.<\/em><\/p><p class=\"c12\">Students occasionally ask me for book recommendations. Since I'm always recommending the same ones, I decided to write up this list. You'll notice that several of them are not directly about software engineering or even computer science! The students have already had plenty of exposure to the classic CS material (and will continue to in their careers), so I try to consider books that are relevant but might not be obvious. At the end of this article I list a few programming books in case that is what you are looking for (please don't yell at me for excluding <a href=\"https:\/\/amzn.to\/385w0tL\">SICP<\/a>). The books I recommend are:<\/p><ul><li>The Design of Everyday Things<\/li>\n<li>Predictably Irrational<\/li>\n<li>Outliers<\/li>\n<li>Creativity, Inc.<\/li>\n<li>Coders at Work<\/li>\n<li>The Mythical Man-Month<\/li>\n<\/ul><br \/><h3>The Design of Everyday Things<\/h3><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/designofeverydaythings.jpg\" class=\"c13\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"float: left;\" \/> The book that has impacted me the most is The Design of Everyday Things (<a href=\"https:\/\/amzn.to\/2T8e65k\">Amazon<\/a>). It was the first book my PhD advisor recommended to me. Although the book has nothing specific to do with software, it has a big lesson for us: software is way more than just the features we implement. If the software isn't usable, then it doesn't matter how fast it is or how many features it has. The book opened my eyes to the usability problems that surround us and how to study them. So the next time you can't figure out how to use a microwave oven, it isn't your fault.<\/p><p class=\"c14\"><strong>Key takeaway:<\/strong> Design is everything, so stop blaming users and learn from them.<\/p><br \/><h3>Predictably Irrational<\/h3><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/predictablyirrational.jpg\" class=\"c15\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"float: right;\" \/> My first foray into behavioral economics and decision making was Predictably Irrational (<a href=\"https:\/\/amzn.to\/2PBrxbM\">Amazon<\/a>). It summarizes a number of studies that have been ran to understand how people make decisions and the cognitive biases that we often face. Many of the examples in this book center around the psychology of pricing, though they often pop up in social settings as well. For example, if someone offers us free chocolate, why do we take less than we would if it costs 1 penny? If you enjoy this, then I would also recommend Misbehaving by Richard Thaler (<a href=\"https:\/\/amzn.to\/3ahx2nW\">Amazon<\/a>).<\/p><p class=\"c16\"><strong>Key takeaway:<\/strong> Making decisions is hard, but we fall into a lot of traps that could be avoided.<\/p><br \/><h3>Outliers<\/h3><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/outliers.jpg\" class=\"c13\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"float: left;\" \/> Outliers (<a href=\"https:\/\/amzn.to\/2TaQr4k\">Amazon<\/a>) walks through a series of contrasting case studies of successes and failures. From professional hockey players to elementary school students to CEOs, there is more to their success than we might think. Some of the most intelligent people in the world live typical lives while others become billionaires. You need to be smart enough, but after that it seems to come down to charism, grit, help from others, and luck. Note: This book does seem to overgeneralize and oversimplify complex situations, but there are still good points.<\/p><p class=\"c17\"><strong>Key takeaway:<\/strong> There is more to success and failure than we think.<\/p><br \/><h3>Creativity, Inc<\/h3><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/creativityinc.jpg\" class=\"c15\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"float: right;\" \/> Creativity, Inc. (<a href=\"https:\/\/amzn.to\/3ciw6S3\">Amazon<\/a>) tells the story of Edwin Catmull's journey from getting a PhD in computer science, working at Lucasfilm, and working as CTO of Pixar with Steve Jobs. Other than the entertainment of the story, it is littered with lessons on management. For example, when interviewing a job candidate, ask yourself if you could see them being your boss in the near future. All Edwin wanted to do is to make the first animated film...<\/p><p class=\"c18\"><strong>Key takeaway:<\/strong> Managing engineers and creatives requires some unintuitive strategies.<\/p><br \/><h3>Coders at Work<\/h3><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/codersatwork.jpg\" class=\"c13\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"float: left;\" \/> Coders at Work (<a href=\"https:\/\/amzn.to\/3ahoe1q\">Amazon<\/a>) is a collection of interviews with big names in the software world. It is much like listening to a podcast or reading blog posts. For example, you can read about Peter Norvig's experiences with the development process at Google and NASA. If you are interested in entrepreneurship and startups, then I recommend the sibling book, Founders at Work (<a href=\"https:\/\/amzn.to\/2wd1H7d\">Amazon<\/a>).<\/p><p class=\"c18\"><strong>Key takeaway:<\/strong> You can learn a lot from listening to the lessons learned by others.<\/p><br \/><h3>The Mythical Man-Month<\/h3><p><img src=\"http:\/\/web.eecs.utk.edu\/~azh\/blog\/images\/mythicalmanmonth.jpg\" class=\"c15\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"float: right;\" \/> The Mythical Man-Month (<a href=\"https:\/\/amzn.to\/3clLehh\">Amazon<\/a>) is a classic collection of essays describing issues that arise while managing software projects. Some aspects of software engineering really haven't changed since the 1970s. For example, adding more programmers to project that is late will make it even later. I had never heard of it until graduate school so I think it is worth including in the list (and none of my students have read it).<\/p><p class=\"c19\"><strong>Key takeaway:<\/strong> You'll soon fall victim to the problems identified by this book, even after reading it.<\/p><hr \/><p class=\"c20\">If you are disappointed that there aren't any programming books, go read Wikipedia articles about data structures and find a book on compilers, functional programming, and AI. I recommend:<\/p><ul><li>Writing an Interpreter in Go (<a href=\"https:\/\/amzn.to\/2Vw1lDa\">Amazon<\/a>)<\/li>\n<li>Learn you a Haskell (<a href=\"https:\/\/amzn.to\/2TaaHmr\">Amazon<\/a>)<\/li>\n<li>Programming Game AI by Example (<a href=\"https:\/\/amzn.to\/2vi42xC\">Amazon<\/a>)<\/li>\n<\/ul><p><small>There are Amazon affiliate links on this page.<\/small><\/p>","created_at":"2020-04-27T14:15:40+0000","updated_at":"2020-04-27T14:15:40+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"web.eecs.utk.edu","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1168,"title":"org-protocol.el \u2013 Intercept calls from emacsclient to trigger custom actions","url":"https:\/\/orgmode.org\/worg\/org-contrib\/org-protocol.html","content":"<p>Setting up org-protocol in <a href=\"http:\/\/conkeror.org\/\">Conkeror<\/a> (an emacs inspired Mozilla web browser) requires a slightly different method. You may simply add the following snippets of code to your .conkerorrc file<sup><a id=\"fnr.2\" class=\"footref\" href=\"#fn.2\" name=\"fnr.2\">2<\/a><\/sup>.<\/p><p>For org-store-link, add the following to .conkerorrc:<\/p><pre class=\"example\">\nfunction org_store_link (url, title, window) {\n    var cmd_str = 'emacsclient \\\"org-protocol:\/\/store-link:\/\/'+url+'\/'+title+'\\\"';\n    if (window != null) {\n        window.minibuffer.message('Issuing ' + cmd_str);\n    }\n    shell_command_blind(cmd_str);\n}\ninteractive(\"org-store-link\", \"Stores [[url][title]] as org link and copies url to emacs kill ring\",\n            function (I) {\n                org_store_link(encodeURIComponent(I.buffer.display_uri_string), encodeURIComponent(I.buffer.document.title), I.window);\n            });\n<\/pre><p>For org-capture (or org-remember \u2014 just exchange <em>capture<\/em> with <em>remember<\/em>), use the following:<\/p><pre class=\"example\">\nfunction org_capture (url, title, selection, window) {\n    var cmd_str = 'emacsclient \\\"org-protocol:\/\/capture:\/\/'+url+'\/'+title+'\/'+selection+'\\\"';\n    if (window != null) {\n        window.minibuffer.message('Issuing ' + cmd_str);\n    }\n    shell_command_blind(cmd_str);\n}\ninteractive(\"org-capture\", \"Clip url, title, and selection to capture via org-protocol\",\n            function (I) {\n                org_capture(encodeURIComponent(I.buffer.display_uri_string), encodeURIComponent(I.buffer.document.title), encodeURIComponent(I.buffer.top_frame.getSelection()), I.window);\n            });\n<\/pre><p>Now, you should be able to invoke the commands from within conkeror with <code>M-x org-store-link<\/code> and <code>M-x org-capture<\/code> (or remember).<\/p><p>Or, if you'd like your familiar emacs keybindings, you can add the following to your .conkerorrc:<\/p><pre class=\"example\">\ndefine_key(content_buffer_normal_keymap, \"C-c r\", \"org-capture\");\ndefine_key(content_buffer_normal_keymap, \"C-c l\", \"org-store-link\");\n<\/pre>","created_at":"2020-04-17T17:21:39+0000","updated_at":"2020-04-17T17:21:39+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1167,"title":"Datomic Architecture with Marshall Thompson - Software Engineering Daily","url":"https:\/\/pca.st\/episode\/a4393333-7356-4f0d-a262-50b6ec163095","content":"<div class=\"section\"><p>Tuesday 24 March 2020<\/p><p><a href=\"https:\/\/softwareengineeringdaily.com\/archives\/\">https:\/\/softwareengineeringdaily.com\/archives<\/a><\/p><p><a href=\"https:\/\/pca.st\/sqad4g29\">Open in Pocket Casts<\/a><\/p><\/div><div id=\"player\"><div id=\"process_bar\"><p>00:00<\/p><p>00:00<\/p><\/div><\/div><div class=\"section show_notes\"><p class=\"description\">Datomic is a database system based on an append-only record keeping system. Datomic users can query the complete history of the database, and Datomic has ACID transactional support. The data within Datomic is stored in an underlying database system suc...<\/p><\/div><div class=\"pc_logo\"><a href=\"https:\/\/www.shiftyjelly.com\/pocketcasts\/\"><img width=\"165\" height=\"38\" src=\"https:\/\/pca.st\/assets\/sharing\/pocketcasts_logo-367852950a4204c0a67952c47a112ce82fa9c2abecc460c11bfc149829710461.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div id=\"desktop_more_ways\"><p>More ways to listen<\/p><div class=\"more_ways_buttons\"><p><a href=\"https:\/\/softwareengineeringdaily.com\/feed\/podcast\/\" target=\"_blank\">RSS<\/a><\/p><p><a href=\"https:\/\/podcasts.apple.com\/us\/podcast\/feed\/id1019576853\" target=\"_blank\">iTunes<\/a><\/p><\/div><\/div>","created_at":"2020-04-17T11:44:42+0000","updated_at":"2020-04-17T11:44:42+0000","published_by":["Pocket Casts"],"annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"pca.st","preview_picture":"http:\/\/static.pocketcasts.com\/discover\/images\/400\/2743d720-0edf-0133-2204-059c869cc4eb.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1166,"title":"Stealing focus: emacs, and PIL, in Windows","url":"https:\/\/bytes.com\/topic\/python\/answers\/428745-stealing-focus-emacs-pil-windows","content":"<a href=\"mailto:da**********@gmail.com\">da**********@gmail.com<\/a>wrote:<blockquote class=\"post_quotes\">\n<p>I'm using GNU Emacs 21.3.1 with python-mode 1.0alpha under Windows XP.<br \/>Whenever I execute a command in an edit window (with<br \/>py-execute-region), the output window steals the focus. How can I stop<br \/>this happening? [snip] I commented out the command<br \/>(pop-to-buffer (current-buffer))<br \/>and now the command window no longer steals focus. But I don't know if<br \/>this has any other side effects, or it there's a better way to prevent<br \/>focus being stolen.<\/p>\n<\/blockquote><p>For someone who says he doesn't know Lisp, you seem to be doing just<br \/>fine. :-)<\/p><p>Not sure if there would be side-effects. Note that py-execute-string,<br \/>py-execute-buffer, py-execute-def-or-class all use py-execute-region.<br \/>Their behaviour may be altered by your change. But it would be easy to<br \/>test.<\/p><p>If you did encounter a problematic side-effect, you could define your<br \/>own \"py-execute-region-custom\" as a copy of py-execute-region but with<br \/>the (pop-to-buffer) call removed. Then use<\/p><p>(require 'python-mode) ;; if you're putting this in .emacs<br \/>(define-key py-mode-map \"\\C-c|\" 'py-execute-region-custom)<\/p><p>to redefine the keyboard mapping. (The (require 'python-mode) call<br \/>ensures that the py-mode-map variable has been initialized.) You could<br \/>provide an alternate mapping for your new function, of course.<\/p><p>If you put the custom function and your define-key call both in your<br \/>..emacs file, then you can easily port it to other machines (perh.<br \/>easier than maintaining a patch for python-mode).<\/p><p>Graham<\/p>","created_at":"2020-04-16T22:22:37+0000","updated_at":"2020-04-16T22:22:37+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"bytes.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1165,"title":"Simula: A VR window manager for Linux","url":"https:\/\/news.ycombinator.com\/item?id=22823891","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825135\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=the_duke\" class=\"hnuser\">the_duke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828616\">30 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828687\">19 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cesarb\" class=\"hnuser\">cesarb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826089\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826223\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=netsharc\" class=\"hnuser\">netsharc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826636\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gpm\" class=\"hnuser\">gpm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827492\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827972\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Robotbeat\" class=\"hnuser\">Robotbeat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828599\">32 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828699\">17 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=skratlo\" class=\"hnuser\">skratlo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825252\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=TooKool4This\" class=\"hnuser\">TooKool4This<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825445\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825467\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=TooKool4This\" class=\"hnuser\">TooKool4This<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825649\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825698\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gpm\" class=\"hnuser\">gpm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825906\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nsxwolf\" class=\"hnuser\">nsxwolf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826447\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=numpad0\" class=\"hnuser\">numpad0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827474\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825411\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=todd8\" class=\"hnuser\">todd8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825775\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ggreer\" class=\"hnuser\">ggreer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827024\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=techopoly\" class=\"hnuser\">techopoly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825699\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kiro\" class=\"hnuser\">Kiro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827432\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=emj\" class=\"hnuser\">emj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825483\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=fractal618\" class=\"hnuser\">fractal618<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825659\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lukifer\" class=\"hnuser\">lukifer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825418\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825575\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mhh__\" class=\"hnuser\">mhh__<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827282\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=driverdan\" class=\"hnuser\">driverdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827298\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827386\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827479\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=outworlder\" class=\"hnuser\">outworlder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825243\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825366\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zachrose\" class=\"hnuser\">zachrose<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825510\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=airstrike\" class=\"hnuser\">airstrike<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826483\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kroltan\" class=\"hnuser\">kroltan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827600\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lambdanaut\" class=\"hnuser\">Lambdanaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825634\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825648\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ravenstine\" class=\"hnuser\">ravenstine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827100\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=carapace\" class=\"hnuser\">carapace<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827071\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=echelon\" class=\"hnuser\">echelon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827578\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=IdiocyInAction\" class=\"hnuser\">IdiocyInAction<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827931\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Robotbeat\" class=\"hnuser\">Robotbeat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828497\">48 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=okasaki\" class=\"hnuser\">okasaki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827724\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=echelon\" class=\"hnuser\">echelon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828626\">28 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ebg13\" class=\"hnuser\">ebg13<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828516\">45 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827226\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=DanBC\" class=\"hnuser\">DanBC<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827562\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828032\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hello_there\" class=\"hnuser\">hello_there<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825765\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825790\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ken\" class=\"hnuser\">ken<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828684\">19 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=daxterspeed\" class=\"hnuser\">daxterspeed<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827583\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kroltan\" class=\"hnuser\">kroltan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827650\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=_anastasia\" class=\"hnuser\">_anastasia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826347\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827272\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=disconcision\" class=\"hnuser\">disconcision<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828389\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=war1025\" class=\"hnuser\">war1025<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826730\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lostmsu\" class=\"hnuser\">lostmsu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22824894\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22824982\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=1MachineElf\" class=\"hnuser\">1MachineElf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825960\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827328\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826248\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanmcdirmid\" class=\"hnuser\">seanmcdirmid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825288\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=todd8\" class=\"hnuser\">todd8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826010\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825421\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanmcdirmid\" class=\"hnuser\">seanmcdirmid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827039\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827184\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nixpulvis\" class=\"hnuser\">nixpulvis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827633\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanmcdirmid\" class=\"hnuser\">seanmcdirmid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827669\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ollifi\" class=\"hnuser\">ollifi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825463\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=vsareto\" class=\"hnuser\">vsareto<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825753\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=netsharc\" class=\"hnuser\">netsharc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826735\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ganzuul\" class=\"hnuser\">ganzuul<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827101\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827263\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=wdroz\" class=\"hnuser\">wdroz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825187\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825607\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thecrumb\" class=\"hnuser\">thecrumb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827566\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MereInterest\" class=\"hnuser\">MereInterest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827789\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=drusepth\" class=\"hnuser\">drusepth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826139\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=IggleSniggle\" class=\"hnuser\">IggleSniggle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828549\">40 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MereInterest\" class=\"hnuser\">MereInterest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827750\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826279\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jstanley\" class=\"hnuser\">jstanley<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825396\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825482\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=duncanawoods\" class=\"hnuser\">duncanawoods<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825471\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sfj\" class=\"hnuser\">sfj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827539\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=htfy96\" class=\"hnuser\">htfy96<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825394\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=viklove\" class=\"hnuser\">viklove<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825860\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827535\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jitl\" class=\"hnuser\">jitl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22824911\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=andybak\" class=\"hnuser\">andybak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825315\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Insanity\" class=\"hnuser\">Insanity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825774\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=fcoury\" class=\"hnuser\">fcoury<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826199\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=noir_lord\" class=\"hnuser\">noir_lord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826697\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=entropicdrifter\" class=\"hnuser\">entropicdrifter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826504\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=snickell\" class=\"hnuser\">snickell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827120\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827318\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=iamwil\" class=\"hnuser\">iamwil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826302\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826426\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=axegon_\" class=\"hnuser\">axegon_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825984\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826056\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haolez\" class=\"hnuser\">haolez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827171\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827240\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haolez\" class=\"hnuser\">haolez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827521\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=platz\" class=\"hnuser\">platz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827631\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bitwize\" class=\"hnuser\">bitwize<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826105\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=matsemann\" class=\"hnuser\">matsemann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825408\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gpm\" class=\"hnuser\">gpm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825849\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825582\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=stronglikedan\" class=\"hnuser\">stronglikedan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825503\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825565\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=drusepth\" class=\"hnuser\">drusepth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826224\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=stronglikedan\" class=\"hnuser\">stronglikedan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825679\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=wincy\" class=\"hnuser\">wincy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825799\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=outworlder\" class=\"hnuser\">outworlder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825883\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gpm\" class=\"hnuser\">gpm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825866\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=outworlder\" class=\"hnuser\">outworlder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825852\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=enobrev\" class=\"hnuser\">enobrev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827322\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rijoja\" class=\"hnuser\">rijoja<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22825819\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hugodutka\" class=\"hnuser\">hugodutka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827301\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cjbassi\" class=\"hnuser\">cjbassi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22824943\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22824991\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgewsinger\" class=\"hnuser\">georgewsinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22824973\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=FpUser\" class=\"hnuser\">FpUser<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22827300\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thcleaner999\" class=\"hnuser\">thcleaner999<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22826198\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kanetw\" class=\"hnuser\">kanetw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22828276\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>","created_at":"2020-04-10T00:45:48+0000","updated_at":"2020-04-10T00:45:48+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1164,"title":"Fun with GTFS & Clojure","url":"https:\/\/grison.me\/2020\/04\/03\/fun-with-gtfs-and-clojure\/","content":"<h2 id=\"hi-people-of-the-internet\">Hi people of the internet!<\/h2><p>Today I\u2019ll be talking about Clojure to parse and extract stuff on GTFS files, wrap all that in a simple REST API using Clojure again, and then expose this API on the net so that my iPhone can consume it and tell me what are my next bus rides.<\/p><p>So much fun to come. I\u2019ve got a bus next monday morning (just kidding, with the quarantine I\u2019m not sure I\u2019ll take a bus anytime soon), just enough time to use some code and get the stuff done!<\/p><p>In a scenario where a mobile app would be created, my main goal would be to let the user select its favorite bus stops regarding specific bus routes and display the next 3 stops in time so that he won\u2019t miss any bus from now on.<\/p><div class=\"c4\"><img src=\"https:\/\/grison.me\/images\/gtfs\/3v8m0b.jpg\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><h2 id=\"what-is-gtfs\">What is GTFS<\/h2><p>GTFS means General Transit Feed Specification. It is a common format for public transportation schedules and associated geographic information.<\/p><p>For more information, you can take a look at the following page: <a href=\"https:\/\/developers.google.com\/transit\/gtfs\/\">https:\/\/developers.google.com\/transit\/gtfs\/<\/a><\/p><p>I live in Metz, France. Hopefully we can grab a GTFS archive with the desired data for all the buses riding in my little town. It can be found here: <a href=\"https:\/\/www.data.gouv.fr\/fr\/datasets\/transport-donnees-gtfs\/\">gtfs_current.zip<\/a><\/p><h2 id=\"tinkering-from-the-repl\">Tinkering from the REPL<\/h2><p>Create a <code>deps.edn<\/code> file with the following content:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">{:deps\n  {com.rpl\/specter {:mvn\/version \"1.1.3\"}\n   nrepl\/nrepl {:mvn\/version \"0.5.3\"}}}<\/pre><\/div><p>Then just run nREPL:<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clj -m nrepl.cmdline\nnREPL server started on port 63456 on host localhost - nrepl:\/\/localhost:63456<\/pre><\/div><p>Now connect your IDE to this running REPL. Or just create one from your IDE directly, it doesn\u2019t matter.<\/p><p>I\u2019m currently running on Windows and run my nREPL from the Windows Linux Subsystem and attach my IDE running in Windows to it.<\/p><p>Now that we have the project ready, unzip the GTFS archive and place it somewhere, let\u2019s say in <code>\/tmp\/gtfs<\/code>.<\/p><p>We\u2019ve got multiple files extracted:<\/p><ul><li><strong>agency.txt<\/strong>: contains the various bus operators (here <em>LE MET\u2019<\/em> and <em>PROXIS<\/em>)<\/li>\n<li><strong>calendar.txt<\/strong>: contains information about whether the different services and on what day do they operate (here we have 4 different (<em>mon-fri<\/em>, <em>wed<\/em>, <em>sat<\/em>, <em>sun<\/em>))<\/li>\n<li><strong>calendar_dates.txt<\/strong>: contains exception of services, for example the service id <em>1<\/em> does not operate on <em>2015-11-11<\/em>.<\/li>\n<li><strong>routes.txt<\/strong>: contains the different routes that a bus could take (35 here).<\/li>\n<li><strong>stop_times.txt<\/strong>: contains the different stops for each bus trip, indicating hour of arrival and departure.<\/li>\n<li><strong>stops.txt<\/strong>: contains all the bus stops (here 1493).<\/li>\n<li><strong>trips.txt<\/strong>: contains all the different bus trips linking stops together (here 4953)<\/li>\n<\/ul><p>Ok now that we know what each file contains, just load them up as Clojure records, and see what we can do.<\/p><p>What we\u2019ll need to do is essentially:<\/p><ol><li>Load a file (which are CSV)<\/li>\n<li>Skip the first line<\/li>\n<li>Create a record instance for each line<\/li>\n<li>Use that list of lines to do something usefull<\/li>\n<\/ol><p>We will first define the various map shapes for all these files:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">; an agency is composed of a name, a URL, a time zone, a lang and a phone\n(def -&gt;agency [:name :url :tz :lang :phone])\n; etc...\n(def -&gt;calendar [:service-id :mon :tue :wed :thu :fri :sat :sun :start-date :end-date])\n(def -&gt;route [:route-id :short-name :long-name :desc :type :url])\n(def -&gt;stop-time [:trip-id :arrival-time :departure-time :stop-id :stop-sequence :pickup-type :drop-off-type :shape-dist-traveled])\n(def -&gt;stop [:stop-id :code :name :desc :lat :lon :id :zone-id :url :location-type :parent-station-id])\n(def -&gt;trip [:route-id :service-id :trip-id :headsign :direction-id :block-id :shape-id :block-id-alt])<\/pre><\/div><p>Everything OK? Just try to load the agencies:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(require '[clojure.java.io :as io])\n(require '[clojure.string :as str])\n(defn make-record\n      \"Make a record out of a line where fields are separated by a comma.\"\n      [record line]\n      (-&gt;&gt; (str\/split line #\",\" -1)       ; split on commas\n           (map str\/trim)                 ; remove un-needed spaces\n           (map #(str\/replace % \"\\\"\" \"\")) ; remove un-needed double quotes\n           (zipmap record)))              ; create a map of the separated line\n(defn load-records\n      \"Load the given file into a list of maps, discarding the header.\"\n      [f mapper]\n      (with-open [r (io\/reader f)]\n                 (-&gt;&gt; (doall (line-seq r))\n                      rest ; skip the header\n                      (map (partial make-record mapper)))))\n(load-records \"\/tmp\/gtfs\/agency.txt\" -&gt;agency)\n; =&gt; ({\n;     :name \"SAEML TAMM (Le Met')\", \n;     :url \"https:\/\/lemet.fr\", \n;     :tz \"Europe\/Paris\", \n;     :lang \"\", \n;     :phone \"\"})<\/pre><\/div><p>It works! We have a sequence of one <code>Agency<\/code> record. Pretty neat!<\/p><p>Great, now let\u2019s dive deeper into the project. We really need to link routes and bus stops.<\/p><p>To do so we need to load <code>stops<\/code>, <code>trips<\/code>, <code>routes<\/code> and <code>stop-times<\/code>.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">; load them all\n(def routes (load-records \"\/tmp\/gtfs\/routes.txt\" -&gt;route))\n(def stops (load-records \"\/tmp\/gtfs\/stops.txt\" -&gt;stop))\n(def stop-times (load-records \"\/tmp\/gtfs\/stop_times.txt\" -&gt;stop-time))\n(def trips (load-records \"\/tmp\/gtfs\/trips.txt\" -&gt;trip))\n(def calendars (load-records \"\/tmp\/gtfs\/calendar.txt\" -&gt;calendar))<\/pre><\/div><p>Just check that I can find the stop near my appartment:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(count (filter #(= (:name %) \"PIERNE\") stops))\n; =&gt; 2<\/pre><\/div><p>There\u2019s 2 of them. One for each direction.<\/p><p>We can just check to be sure:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(filter #(= (:name %) \"PIERNE\") stops)\n;=&gt;\n;    ({:location-type \"\",\n;      :desc \"\",\n;      :name \"PIERNE\",\n;      :id \"\",\n;      :stop-id \"PIERNE01\",\n;      :lon \"6.179863\",\n;      :url \"0\",\n;      :code \"257\",\n;      :zone-id \"http:\/\/lemet.fr\/screen\/index2.php?stop=257\",\n;      :lat \"  49.102273\"}\n;     {:location-type \"\",\n;      :desc \"\",\n;      :name \"PIERNE\",\n;      :id \"\",\n;      :stop-id \"PIERNE02\",\n;      :lon \"6.179825\",\n;      :url \"0\",\n;      :code \"234\",\n;      :zone-id \"http:\/\/lemet.fr\/screen\/index2.php?stop=234\",\n;      :lat \"  49.102558\"})<\/pre><\/div><h3 id=\"discover-the-data\">Discover the data<\/h3><p>Let\u2019s define some utilities methods to retrieve stops and routes easily, using <strong>specter<\/strong> which is ideal for traversing data structures and such exploration.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(require '[com.rpl.specter :as s])\n(defn record-with-property\n  \"Find all records having a specific property matching a given value.\"\n  [prop value records]\n  (s\/select* [s\/ALL (s\/selected? prop #(= % value))] records))\n; Take a look at the use of s\/select-one* instead of s\/select*.\n(defn stop-with-id\n  \"Find a stop having a given id.\"\n  [stop-id]\n  (s\/select-one* [s\/ALL (s\/selected? :stop-id #(= % stop-id))] stops))\n; define some additional methods\n(defn stops-with-name \n  \"Find all stops with a given name.\"\n  [name]\n  (record-with-property :name name stops))\n(defn routes-with-short-name \n  \"Find all routes with a given short name.\"\n  [name]\n  (record-with-property :short-name name routes))\n(defn routes-with-long-name   \n  \"Find all routes with a given long name.\"\n  [name]\n  (record-with-property :route-long-name name routes))\n(routes-with-short-name \"5\")\n; =&gt; [{\n;    :route-id \"5-77\", \n;    :short-name \"5\", \n;    :long-name \"Ligne 5\", \n;    :desc \"\", \n;    :type \"3\", \n;    :url \"\"}]\n(count (stops-with-name \"PIERNE\"))\n; =&gt; 2\n; CASINO is where I lived when I was young\n; https:\/\/www.google.fr\/maps\/place\/49%C2%B006'59.5%22N+6%C2%B008'28.5%22E\/@49.116532,6.1407078,19z\n(count (stops-with-name \"CASINO\"))\n; =&gt; 2\n; how many bus stops do we have in Metz?\n(count (distinct (map :name stops)))\n; =&gt; 491<\/pre><\/div><p>So far it seems to be retrieving what we want. Next I\u2019d like to resume what I will need to do to retrieve the next 3 stops for a specific route and bus stop:<\/p><ol><li>Retrieve the route with its short name (usually the bus line number)<\/li>\n<li>Retrieve the trips for that route (a trip is a route in a defined way (<em>A to B<\/em> or <em>B to A<\/em>))<\/li>\n<li>Retrieve the stops on that route (for that we need to load both the trips and time table)<\/li>\n<li>Retrieve from the time table (<strong>stop_times.txt<\/strong>)<\/li>\n<\/ol><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">; get the route id for the route\n(map :route-id (routes-with-short-name \"5\"))\n=&gt; (\"5-77\")\n(defn trips-for-route \n  \"Find all trips for a given route-id.\"\n  [route-id]\n  (record-with-property :route-id route-id trips))\n(distinct (map :headsign (trips-for-route \"5-77\")))\n; =&gt;\n; (\"L5a - MAISON NEUVE\"\n;  \"L5f - MAGNY PAR RUE DE POUILLY\"\n;  \"L5e - MAGNY PAR RUE AU BOIS\"\n;  \"L5 - FORT MOSELLE\"\n;  \"L5 - REPUBLIQUE\")\n(count (trips-for-route \"5-77\"))\n; =&gt; 326\n(defn trips-with-headsign \n  \"Find all the trips given a headsign.\"\n  [headsign]\n  (record-with-property :headsign headsign trips))\n(count (trips-with-headsign \"L5a - MAISON NEUVE\"))   \n; =&gt; 159                                       \n; it means that there's 159 trips on Line 5 that goes in direction of the final stop MAISON NEUVE.\n; let's check what the first trip is\n(first (trips-with-headsign \"L5a - MAISON NEUVE\"))\n; =&gt;\n; {:route-id \"5-77\",\n;  :service-id \"HIV1920-DIM_HIV-Dimanche-00\",\n;  :trip-id \"269026-HIV1920-DIM_HIV-Dimanche-00\",\n;  :headsign \"L5a - MAISON NEUVE\",\n;  :direction-id \"1\",\n;  :block-id \"39965\",\n;  :shape-id \"50048\",\n;  :block-id-alt \"05 - 01\"}\nWe want all the different stops on a trip:\n```clojure\n(defn stops-on-trip \n  \"Find all the stops on a given trip.\"\n  [trip-id]\n  (record-with-property :trip-id trip-id stop-times))\n(count (stops-on-trip \"269026-HIV1920-DIM_HIV-Dimanche-00\"))\n; =&gt; 40\n(map (comp :name stop-with-id :stop-id) \n     (stops-on-trip \"269026-HIV1920-DIM_HIV-Dimanche-00\"))\n; =&gt;\n;(\"AUBEPINE\"\n; \"HAUTS-DE-MAGNY\"\n; \"BEAUSOLEIL\"\n; \"ARMOISIERES\"\n; \"OBELLIANE\"\n; \"MAGNY-AU-BOIS\"\n; \"ROOSEVELT\"\n; \"LA PLAINE\"\n; \"APREMONT\"\n; \"PLATEAU\"\n; \"FAUBOURG\"\n; \"FRECOT\"\n; \"BOUCHOTTE\"\n; \"VANDERNOOT\"\n; \"LOTHAIRE\"\n; \"PIERNE\"   &lt;----- cool\n; \"LEMUD\"\n; \"MUSE\"\n; \"CENTRE POMPIDOU METZ\"\n; \"GARE\"\n; \"ROI GEORGE\"\n; \"REPUBLIQUE\"\n; \"SQUARE DU LUXEMBOURG\"\n; \"FORT MOSELLE\"\n; \"TIGNOMONT\"\n; \"ST-MARTIN\"\n; \"FOCH\"\n; \"PONT DE VERDUN\"\n; \"CASINO\"    &lt;----- also cool \n; \"MIGETTE\"\n; \"LONGEVILLE\"\n; \"LECLERC\"\n; \"EN PRILLE\"\n; \"SCY BAS\"\n; \"LIBERTE\"\n; \"MOULINS\"\n; \"ST-JEAN\"\n; \"SERRET\"\n; \"HAIE BRULEE\"\n; \"MAISON NEUVE\")<\/pre><\/div><p>Let\u2019s say it is <strong>17:41<\/strong>, I want to know each 3 bus trips after <strong>17:41<\/strong> where the bus are at my stop for the route <code>L5a - MAISON NEUVE<\/code><\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn times-for-stop \n  \"Find the times for a given stop.\"\n  [stop-id]\n  (record-with-property :stop-id stop-id stop-times))\n(count (times-for-stop \"PIERNE01\"))\n; =&gt; 194\n(take 3 \n   (sort-by :arrival-time\n      (s\/select* [s\/ALL\n         (s\/selected? :stop-id #(= % \"PIERNE01\"))\n         (s\/selected? :arrival-time #(= 1 (compare % \"17:41:00\")))]\n      (times-for-stop \"PIERNE01\"))))\n; =&gt;\n; ({:trip-id \"269041-HIV1920-DIM_HIV-Dimanche-00\",\n;   :arrival-time \"17:54:58\",\n;   :departure-time \"17:54:58\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"16\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"5208.0\"}\n;  {:trip-id \"281460-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"17:58:25\",\n;   :departure-time \"17:58:25\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"}\n;  {:trip-id \"268643-HIV1920-H1920SAM-Samedi-00\",\n;   :arrival-time \"17:59:58\",\n;   :departure-time \"17:59:58\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"})<\/pre><\/div><p>That\u2019s cool and all but it might return me the stops for all different routes, I want only those on the route named <strong>Ligne 5a - MAISON NEUVE<\/strong>:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn trip-in-route? [trip-id route-id] (contains? (set (map :trip-id trips)) trip-id))\n(take 3 \n   (sort-by :arrival-time\n      (s\/select* [s\/ALL\n         (s\/selected? :stop-id #(= % \"PIERNE01\"))\n         (s\/selected? :arrival-time #(= 1 (compare % \"17:41:00\")))\n         (s\/selected? :trip-id #(trip-in-route? % \"L5a - MAISON NEUVE\"))]\n      (times-for-stop \"PIERNE01\"))))\n; =&gt;\n; ({:trip-id \"269041-HIV1920-DIM_HIV-Dimanche-00\",\n;   :arrival-time \"17:54:58\",\n;   :departure-time \"17:54:58\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"16\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"5208.0\"}\n;  {:trip-id \"281460-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"17:58:25\",\n;   :departure-time \"17:58:25\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"}\n;  {:trip-id \"268643-HIV1920-H1920SAM-Samedi-00\",\n;   :arrival-time \"17:59:58\",\n;   :departure-time \"17:59:58\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"})                        <\/pre><\/div><p>It gives the exact same result, and that for a reason. The <code>PIERNE<\/code> bus stop (with id <strong>21340<\/strong>) is only served by the bus line <strong>Ligne 5<\/strong>.<\/p><p>Just check that these 3 buses have the same arrival:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn trip-with-id \n  \"Find a trip given its id.\"  \n  [trip-id]\n  (s\/select-one* [s\/ALL (s\/selected? :trip-id #(= % trip-id))] trips))\n(map (comp :headsign trip-with-id)\n     [\"269041-HIV1920-DIM_HIV-Dimanche-00\" \"281460-HIV1920-SEM_GTFS-Semaine-00\" \"268643-HIV1920-H1920SAM-Samedi-00\"])\n; =&gt; (\"L5a - MAISON NEUVE\" \"L5a - MAISON NEUVE\" \"L5a - MAISON NEUVE\")<\/pre><\/div><p>Just one last tweak to do before we go to the next step.<\/p><p>Currently it gives the trips for all days of the week, but in Metz (and elsewhere) we have different trips during the week and during the week-end, let\u2019s say we\u2019re tuesday.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">; which services are valid for tuesday?\n(map :service-id (record-with-property :tue \"1\" calendars))\n; =&gt; (\"HIV1920-SEM_GTFS-Semaine-00\")\n; let's deal with it\n(defn trip-for-day? \n  \"Does the given trip-id runs on the given day.\"\n  [trip-id day]\n  (let [service-ids (into #{} (map :service-id (record-with-property day \"1\" calendars)))\n        trip-service-id (:service-id (first (record-with-property :trip-id trip-id trips)))]\n       (contains? service-ids trip-service-id)))\n(take 3\n   (sort-by :arrival-time\n      (s\/select* [s\/ALL\n         (s\/selected? :stop-id #(= % \"PIERNE01\"))  ; for the stop PIERNE01\n         (s\/selected? :arrival-time #(= 1 (compare % \"17:41:00\"))) ; after 17:41:00\n         (s\/selected? :trip-id #(trip-in-route? % \"L5a - MAISON NEUVE\")) ; following the headsign L5a - MAISON NEUVE\n         (s\/selected? :trip-id #(trip-for-day? % :tue))] ; running on tuesday\n      (times-for-stop \"PIERNE01\"))))\n; =&gt;\n; ({:trip-id \"281460-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"17:58:25\",\n;   :departure-time \"17:58:25\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"}\n;  {:trip-id \"281426-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"18:08:25\",\n;   :departure-time \"18:08:25\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"16\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"5208.0\"}\n;  {:trip-id \"281461-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"18:18:58\",\n;   :departure-time \"18:18:58\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"})\n; Make a function of it\n(defn next-times \n   [number stop-id arrival-time headsign day]\n   (take number\n      (sort-by :arrival-time\n         (s\/select* [s\/ALL\n            (s\/selected? :stop-id #(= % stop-id))\n            (s\/selected? :arrival-time #(= 1 (compare % arrival-time)))\n            (s\/selected? :trip-id #(trip-in-route? % headsign))\n            (s\/selected? :trip-id #(trip-for-day? % day))]\n         (times-for-stop stop-id)))))\n; check it works \n(next-times 3 \"PIERNE01\" \"17:41:00\" \"L5a - MAISON NEUVE\" :tue)\n; =&gt;\n; ({:trip-id \"281460-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"17:58:25\",\n;   :departure-time \"17:58:25\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"}\n;  {:trip-id \"281426-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"18:08:25\",\n;   :departure-time \"18:08:25\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"16\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"5208.0\"}\n;  {:trip-id \"281461-HIV1920-SEM_GTFS-Semaine-00\",\n;   :arrival-time \"18:18:58\",\n;   :departure-time \"18:18:58\",\n;   :stop-id \"PIERNE01\",\n;   :stop-sequence \"12\",\n;   :pickup-type \"0\",\n;   :drop-off-type \"0\",\n;   :shape-dist-traveled \"4162.0\"})<\/pre><\/div><p>Seems about right \ud83e\udd20 !<\/p><p>So far, so good, we now know how to list the different routes, retrieve the stops of a route with specific direction (headsign), and how to retrieve the next 3 stops at a bus station.<\/p><p>The problem is that people that will be using the application will also need to retrieve the routes and trips from a stop, because they usually already know the name of their favorite bus stops.<\/p><p>Let\u2019s do that with a bus stop that has many bus lines passing by, named <strong>REPUBLIQUE<\/strong>:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(count (stops-with-name \"REPUBLIQUE\"))\n; =&gt; 9\n(def republique-stops\n  (set (map :stop-id (stops-with-name \"REPUBLIQUE\"))))\n; =&gt; #'user\/republique-stops\nrepublique-stops\n; =&gt; #{\"REPU5052\" \"REPUB123\" \"place_REPUB\" \"REP01\" \"REPUBL88\" \"REPUBL01\" \"REPU5051\" \"REP02\" \"REPUB436\"}\n(def republique-stop-times-trip-ids\n    (set (map :trip-id (s\/select* [s\/ALL\n                (s\/selected? :stop-id #(contains? republique-stops %))] stop-times))))\n; =&gt; #'user\/republique-stop-times-trip-ids\n(count republique-stop-times-trip-ids)                \n; =&gt; 2979\n(distinct (map :route-id (s\/select* [s\/ALL\n                                     (s\/selected? :trip-id #(contains? republique-stop-times-trip-ids %))] trips)))\n; =&gt; (\"12-101\" \"11-9\" \"5-77\" \"4-7\" \"3-81\" \"2-5\" \"1-100\" \"B-99\" \"A-98\" \"83-71\")\n(defn route-with-id [route-id]\n  (s\/select-one* [s\/ALL (s\/selected? :route-id #(= % route-id))] routes))\n(map :long-name (map route-with-id '(\"12-101\" \"11-9\" \"5-77\" \"4-7\" \"3-81\" \"2-5\" \"1-100\" \"B-99\" \"A-98\" \"83-71\")))\n; =&gt; (\n; \"Citeis 12\" \n; \"Citeis 11\" \n; \"Ligne 5\" \n; \"Ligne 4\" \n; \"Ligne 3\" \n; \"Ligne 2\" \n; \"Ligne 1\" \n; \"Mettis B\" \n; \"Mettis A\" \n; \"Navette CITY\")   <\/pre><\/div><p>We can conclude that there are 10 bus lines that goes through the <strong>REPUBLIQUE<\/strong> bus stop. Having that we\u2019ll need to propose the choice of a direction (headsign) for a chosen route:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">; route id \"5-77\" is L5: MAGNY - MAISON NEUVE the one I used to take when I was going to school.\n(distinct\n  (map :headsign (filter #(= (:route-id %) \"5-77\") trips)))\n; =&gt; (\"L5a - MAISON NEUVE\"\n;     \"L5f - MAGNY PAR RUE DE POUILLY\"\n;     \"L5e - MAGNY PAR RUE AU BOIS\"\n;     \"L5 - FORT MOSELLE\"\n;     \"L5 - REPUBLIQUE\")\n(distinct\n  (map :headsign (filter #(= (:route-id %) \"B-99\") trips)))\n; =&gt; (\"MB - HOPITAL MERCY\" \"MB - UNIVERSITE SAULCY\")<\/pre><\/div><p>There are 5 different final destination for the bus line number <strong>Ligne 5<\/strong> and 2 final destinations for the bus line <strong>Mettis B<\/strong>.<\/p><p>Finally, one last problem: a trip with a final destination (head sign) has a determined way, it goes from A to B or in reverse, and we should really group these trips by destination:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">; There's no built-in method in Clojure that I know of to map on the values of a hashmap\n(defn map-values [m f &amp; args]\n   (reduce (fn [a [k v]] (assoc a k (apply f v args))) {} m))\n(def trips-by-direction\n    (group-by :direction-id (distinct (filter #(= (:route-id %) \"5-77\") trips))))\n; =&gt;  #'user\/trips-by-direction\n(map-values trips-by-direction (fn [m] (distinct (map :headsign m))))\n; =&gt; \n;  {\"1\" (\"L5a - MAISON NEUVE\" \"L5 - FORT MOSELLE\"),\n;   \"0\" (\"L5f - MAGNY PAR RUE DE POUILLY\" \"L5e - MAGNY PAR RUE AU BOIS\" \"L5 - REPUBLIQUE\")}<\/pre><\/div><p>Well I think at this point we pretty much covered what we need for our REST API, so let\u2019s do that!<\/p><h2 id=\"performance\">Performance<\/h2><p>For a dataset of the size of Metz it\u2019s easy, but if you were to load a huge dataset like Paris, I think these functions would take to much time on each HTTP request to filter. I\u2019m just trying to find excuses to do some more because I\u2019m bored :)<\/p><p>So let\u2019s create a big <strong>store<\/strong> which will contain all the information already tailored for the API, so that all we have to do is a matter of <code>get-in<\/code> that store to get what we need.<\/p><p>Such kind of structure should be great:<\/p><div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n    \"agencies\": [],\n    \"routes\": [],\n    \"stops\": [],\n    \"trips\": [],\n    \"store\": {\n        \"SEILLE\": {\n            \"Mettis B\": {  \n                \"MB - HOPITAL MERCY\": [\n                    \"Here, some time-table please\"\n                ],\n                \"MB - UNIVERSITE SAULCY\": [\n                    \"Here, some time-table please\"\n                ]\n            }\n        }\n    }\n}<\/pre><\/div><p>We already have the <code>agencies<\/code>, <code>routes<\/code>, <code>stops<\/code> and <code>trips<\/code>. We also already know how to retrieve the following:<\/p><ol><li>The different distinct stop names aka <code>store.SEILLE<\/code><\/li>\n<li>The routes passing by a stop aka <code>store.SEILLE.'Mettis B'<\/code><\/li>\n<li>The destination of a specific route passing by a stop aka <code>store.SEILLE.'Mettis B'.'MB - UNIVERSITE SAULCY'<\/code><\/li>\n<li>The trip ids of a specific route + stop + destination.<\/li>\n<\/ol><p>So it\u2019s all just about mapping functions we already have to build a giant structure so that it will be very easy for us to write our REST API.<\/p><h3 id=\"building-the-index\">Building the index<\/h3><p>We\u2019re going to build an intermediate index with what we need:<\/p><ul><li>agencies<\/li>\n<li>routes + routes indexed by their route-id (<code>routes-map<\/code>, for performance)<\/li>\n<li>stops<\/li>\n<li>trips + trips indexed by their trip-id (<code>trips-map<\/code>, for performance)<\/li>\n<li>trips indexed by route-id (<code>trips-by-route<\/code>, for performance)<\/li>\n<li>stop times<\/li>\n<li>stop times indexed by stop id (<code>stop-times-by-stop<\/code>, again for performance)<\/li>\n<\/ul><p>This is specific to what I have in mind, it\u2019s not necessary to understand how it\u2019s build, the key point is that we have the structure that I explained above so that the API just navigates in it, everything is prebuilt on startup.<\/p><p><strong>The different distinct stop names aka <code>store.SEILLE<\/code><\/strong>:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn distinct-stop-names-with-ids\n      \"Find the distinct stops name with their associated ids.\"\n      [{:keys [stops]}]\n           (map #(vector % (map :stop-id (stops-with-name %)))\n                (sort (distinct (map :name stops)))))\n(distinct-stop-names-with-ids {:stops stops})\n; =&gt;\n; ([\"11\u00e8me D'AVIATION\" (\"11AVIAT1\" \"11AVIAT2\")]\n;  [\"19 NOVEMBRE\" (\"19NOV01\" \"19NOV02\")]\n;  [\"8 MAI 45\" (\"08MAI01\" \"08MAI02\")]\n;  [\"ABBE BAUZIN\" (\"ABBEBAU1\" \"ABBEBAU2\")]\n;  [\"ACACIAS\" (\"ACACIAS1\" \"ACACIAS2\")]\n;  [\"ACTISUD DUNIL\" (\"ACTISUD1\" \"ACTISUD2\")]\n;  [\"ALGER\" (\"ALGER1\" \"ALGER2\")]\n;  [\"ALSACE\" (\"ALSACE1\" \"ALSACE2\")]\n;  ...         <\/pre><\/div><p><strong>The routes passing by a stop aka <code>store.SEILLE.'Mettis B'<\/code><\/strong>:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn trip-ids-for-stops\n  \"Find the trip ids for given number of stops.\"\n  [stops stop-times]\n  (set (map :trip-id (flatten (map #(get stop-times %) stops)))))\n(defn distinct-route-ids-for-trips\n  \"Find the distinct route ids for a given number of trips.\"\n  [trip-ids trips]\n  (distinct (map :route-id (map #(get trips %) trip-ids))))\n(defn routes-passing-by-stop-ids\n      \"Find the routes passing by a given number of stops.\"\n      [stop-ids {:keys [routes-map stop-times-by-stop trips-map]}]\n      (filter some? (map #(get routes-map %)\n                         (distinct-route-ids-for-trips (trip-ids-for-stops stop-ids stop-times-by-stop)\n                                                       trips-map))))\n; test it works\n(map :long-name (routes-passing-by-stop-ids #{\"PIERNE01\"} store))\n; =&gt; (\"Proxis 113\" \"Ligne 5\")\n(map :long-name (routes-passing-by-stop-ids #{\"REPUBL01\"} store))\n; =&gt; (\"Ligne 3\" \"Ligne 5\")\n(map :long-name (routes-passing-by-stop-ids #{\"REPUBL88\"} store))\n; =&gt; (\"Ligne 1\" \"Ligne 4\" \"Citeis 11\" \"Navette CITY\")<\/pre><\/div><p><strong>The destination of a specific route passing by a stop aka <code>store.SEILLE.'Mettis B'.'MB - UNIVERSITE SAULCY'<\/code><\/strong>:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn times-by-headsign-at-stop\n  \"Find the times a bus stops at a specific stop grouped by headsign.\"\n  [stop route-id store]\n    (group-by #(:headsign (get (:trips-map store) (:trip_id %)))   ; group trips by trip headisgn\n            (sort-by (juxt :arrival_time :departure-time)               ; sort by arrival-time then departure-time\n                     (filter #(contains? (set (get (:trips-by-route store) route-id)) ; filter trips\n                                         (get (:trips-map store) (:trip-id %)))       ; on routes they're on\n                             (get (:stop-times-by-stop store) stop)))))    ; get the times at specified stop\n(keys (times-by-headsign-at-stop \"PIERNE01\" \"5-77\" store))\n; =&gt; (\"L5a - MAISON NEUVE\")   \n(count (get (times-by-headsign-at-stop \"PIERNE01\" \"5-77\" store) \"L5a - MAISON NEUVE\"))\n; =&gt; 194                        <\/pre><\/div><p>Having written all that, we just need to use them all to create our desired structure:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn make-api-struct\n  \"Construct the struct that will be searched by the REST API.\"\n  [store]\n  (-&gt;&gt; (distinct-stop-names-with-ids store)                 \n       (map #(hash-map (first %)                 ; create hash-map with stop name as key                    \n                       (apply merge-with into    ; and as value:\n                              (map (fn [r] (hash-map (:route_long_name r) ; a hash-map with route name as key\n                                                     (map (fn [stop] (times-by-headsign-at-stop stop (:route_id r) store)) ; and as value the times grouped by destination\n                                                          (second %))))\n                                   (routes-passing-by-stop-ids (second %) store))\n                              )))\n       (apply merge-with into)))<\/pre><\/div><p>Then create our <code>load-gtfs<\/code> method to build it once the parsing of the GTFS files has terminated:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn load-gtfs\n  \"Load the GTFS data, and make a store of it.\"\n  [path]\n  (let [routes (load-records (str path \"\/routes.txt\") -&gt;Route)\n        trips (load-records (str path \"\/trips.txt\") -&gt;Trip)\n        stop-times (load-records (str path \"\/stop_times.txt\") -&gt;StopTime)]\n    (let [store {:agencies           (load-records (str path \"\/agency.txt\") -&gt;Agency)\n                 :routes             routes\n                 :routes-map         (into {} (map #(vector (:route_id %) %) routes))\n                 :stops              (load-records (str path \"\/stops.txt\") -&gt;Stop)\n                 :trips              trips\n                 :trips-map          (into {} (map #(vector (:trip_id %) %) trips))\n                 :trips-by-route     (apply merge-with into (map #(hash-map (:route_id %) [%]) trips))\n                 :stop-times         stop-times\n                 :stop-times-by-stop (apply merge-with into (map #(hash-map (:stop_id %) [%]) stop-times))}]\n      (assoc store :data (make-api-struct store))))) ; &lt;-- here!\n(time (def store (load-gtfs \"\/tmp\/gtfs\")))\n; Loading \/tmp\/gtfs\/routes.txt...\n; Loading \/tmp\/gtfs\/trips.txt...\n; Loading \/tmp\/gtfs\/stop_times.txt...\n; Loading \/tmp\/gtfs\/stops.txt...\n; Loading \/tmp\/gtfs\/agency.txt...\n; \"Elapsed time: 540.088 msecs\"\n; =&gt; #'user\/store\n(keys store)\n; =&gt; (:trips-by-route :routes :stops :stop-times-by-stop :trips-map :stop-times :trips :routes-map :agencies :data)\n(keys (:data store))\n; =&gt; \n;   (\"PREFECTURE\"\n;    \"SERRET\"\n;    \"LECLERC\"\n;    \"P+R ROCHAMBEAU\"\n;    \"PUYMAIGRE\"\n;    \"GENDARMERIE\"\n;    \"LA MAXE\"\n;    ...\n(count (keys (:data store)))\n; =&gt; 491\n(count (distinct (map :name stops)))\n; =&gt; 491<\/pre><\/div><p>Awesome, and seems ok performance wise.<\/p><p>Now how do we use that structure I can hear you saying? It\u2019s just a matter of <code>get-in<\/code> the actual structure.<\/p><p>We\u2019ll get to that in the next section.<\/p><h2 id=\"rest-api\">REST API<\/h2><p>We\u2019re going to build a simple API with compojure, I\u2019m going to name it monmet (my Met\u2019, Le Met\u2019 being the name of the bus network in Metz).<\/p><h3 id=\"init-the-project\">Init the project<\/h3><p>Stop your REPL, modify the <code>deps.edn<\/code> file to put this content in it.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">{:paths [\"resources\" \"src\"]\n :deps {\n        org.clojure\/clojure {:mvn\/version \"1.10.1\"}\n        ring {:mvn\/version \"1.7.1\"}\n        ring\/ring-core {:mvn\/version \"1.7.1\"}\n        ring\/ring-json {:mvn\/version \"0.5.0\"}\n        ring\/ring-defaults {:mvn\/version \"0.3.2\"}\n        ring\/ring-jetty-adapter {:mvn\/version \"1.7.1\"}\n        com.rpl\/specter {:mvn\/version \"0.11.2\"}\n        mount {:mvn\/version \"0.1.10\"}\n        compojure {:mvn\/version \"1.6.1\"}}}<\/pre><\/div><p>And create the needed directories:<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ mkdir -p src\/monmet\n$ cd src\/monmet\n$ touch api.clj data.clj gtfs.clj handler.clj main.cl<\/pre><\/div><h3 id=\"loading\">Loading<\/h3><p>In order to load one time our datastore (all the GTFS files) we\u2019re going to use mount. It\u2019s a very simple library, an alternative to components.<\/p><p>Let\u2019s do so by modifying the <code>data.clj<\/code> file with the following code:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.data\n  (:require [monmet.gtfs :as gtfs]\n            [mount.core :as mnt]))\n(mnt\/defstate store :start (gtfs\/load-gtfs \"\/tmp\/gtfs\"))\n(defn store-has-entry?\n  \"Check if the given entry is a valid one in our datastore.\"\n  [entry]\n  (contains? #{:agencies :routes :stops :trips :stop-times} entry))<\/pre><\/div><p>Here we have defined a state named store that will on call to <code>(mnt\/start)<\/code> invoke <code>(gtfs\/load-gtfs \"\/tmp\/gtfs\")<\/code> which returns a map holding the various GTFS files content. It will be our datastore and it will be available in every namespace of our own, and it is guaranteed to be loaded only one time since it\u2019s an expensive operation.<\/p><h3 id=\"routes\">Routes<\/h3><p>We will be using ring and compojure to create the simple REST API we need.<\/p><p>The skeleton looks like this:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.handler\n  (:require [compojure.core :refer :all]\n            [compojure.handler :as handler]\n            [ring.middleware.json :as middleware]\n            [ring.util.response :refer [response]]\n            [compojure.route :as route]\n            [mount.core :as mnt]))\n(defroutes app-routes\n   (GET \"\/hello\" [] (response {:foo \"bar\"}))\n   (route\/not-found \"Not Found\"))\n(def app\n  (do\n    (mnt\/start) ; mount up the datastore\n    (-&gt; (handler\/site app-routes)\n        (middleware\/wrap-json-body {:keywords? true}) ; transform fields to clojure keywords\n        middleware\/wrap-json-response)))<\/pre><\/div><h3 id=\"get-gtfs-entry\"><code>GET \/gtfs\/:entry<\/code><\/h3><p>On <code>GET<\/code> to <code>\/gts\/:entry<\/code> where <code>:entry<\/code> is one of <code>:agencies<\/code>, <code>:routes<\/code>, <code>:stops<\/code>, <code>:trips<\/code>, <code>:stop-times<\/code> we will return what\u2019s inside the GTFS file directly, we\u2019re just serving the CSV files as JSON.<\/p><p>We\u2019ll need two methods of our <code>data<\/code> namespace: <code>store<\/code> which is our datastore holding the precious data, and <code>store-has-entry?<\/code> which is a little helper to know if the requested entry really exists.<\/p><p>When we\u2019re done it looks like this:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.handler\n  (:require [compojure.core :refer :all]\n            [compojure.handler :as handler]\n            [ring.middleware.json :as middleware]\n            [ring.util.response :refer [response]]\n            [compojure.route :as route]\n            [mount.core :as mnt]\n            [monmet.data :refer [store store-has-entry?]]))\n(defroutes app-routes\n   ; Expose GTFS files directly, just for fun :)\n   (GET \"\/gtfs\/:file\" [file]\n     (let [entry (keyword file)]      ; transform URL path to keyword\n       (if (store-has-entry? entry)   \n         (response (entry store))         ; in case the entry exist, just serve what's in the datastore\n         (route\/not-found \"Not Found\")))) ; otherwhise, 404\n  (route\/not-found \"Not Found\"))\n(def app\n  (do\n    (mnt\/start) ; mount up the datastore\n    (-&gt; (handler\/site app-routes)\n        (middleware\/wrap-json-body {:keywords? true})\n        middleware\/wrap-json-response)))<\/pre><\/div><p>Test it works:<\/p><p>Run the program with <code>clj -m monmet.main<\/code> and use curl<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ curl http:\/\/localhost:8080\/gtfs\/agencies\n[\n    {\n        name: \"SAEML TAMM (Le Met')\",\n        url: \"https:\/\/lemet.fr\",\n        tz: \"Europe\/Paris\",\n        lang: \"\",\n        phone: \"\"\n    }\n]<\/pre><\/div><h3 id=\"get-api\"><code>GET \/api\/...<\/code><\/h3><h4 id=\"bus-routes\">Bus routes<\/h4><p>To make things clearer we\u2019ll create another namespace called <code>api<\/code> where we will really need our <code>gtfs<\/code> namespace to do some work.<\/p><p>Let\u2019s start with routes. I\u2019d like to return the content of <code>routes.txt<\/code> but enhanced with the <code>Agency<\/code> corresponding to the <code>:agency_id<\/code> field of each <code>Route<\/code>.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.api\n  (:require [monmet.gtfs :as gtfs]))\n(defn get-routes\n  \"Retrieve routes.\"\n  [store]\n  (:routes store))          <\/pre><\/div><p>We just need to expose that very <code>api\/get-routes<\/code> in our compojure <code>defroutes<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defroutes app-routes\n   ; ...\n   (GET \"\/api\/routes\" [] (response (api\/get-routes store)))\n   ; ...\n  (route\/not-found \"Not Found\"))<\/pre><\/div><p>That\u2019s it, test it using curl (or even your browser):<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ curl -X GET http:\/\/localhost:3000\/api\/routes\n[\n  {\n    \"route_id\": \"1-100\",\n    \"route_short_name\": \"1\",\n    \"route_long_name\": \"Ligne 1\",\n    \"route_desc\": \"\",\n    \"route_type\": \"3\",\n    \"route_url\": \"\"\n  },\n  {\n    \"route_id\": \"2-5\",\n    \"route_short_name\": \"2\",\n    \"route_long_name\": \"Ligne 2\",\n    \"route_desc\": \"\",\n    \"route_type\": \"3\",\n    \"route_url\": \"\"\n  },\n   ...\n]<\/pre><\/div><h4 id=\"bus-stops\">Bus stops<\/h4><p>Now that our bus routes are returned, we really need some bus stops or we won\u2019t be able to enter the damn bus!<\/p><p>Get back to the <code>api<\/code> namespace in order to create two methods, one to retrieve the distinct bus stops sorted by name and one to retrieve all the bus stops matching a name.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.api\n  (:require [monmet.gtfs :as gtfs]))\n; ...\n(defn get-stops\n  \"Retrieve the distinct stop names.\"\n  [store]\n  (gtfs\/distinct-stop-names-with-ids store))\n(defn get-stop-with-name\n  \"Retrieve the stops matching (strict equality) a name.\"\n  [stop-name store]\n  (gtfs\/stops-with-name stop-name (:stops store)))<\/pre><\/div><p>Expose them in the API:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(GET \"\/api\/stops\" [] (response (api\/get-stops store)))\n(GET \"\/api\/stops\/:name\" [name] (response (api\/get-stop-with-name name store)))<\/pre><\/div><p>Test it:<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ curl -X GET http:\/\/localhost:3000\/api\/stops | jq '.'\n[\n  [\n    \"11\u00e8me D'AVIATION\",\n    [\n      \"11AVIAT1\",\n      \"11AVIAT2\"\n    ]\n  ],\n  [\n    \"19 NOVEMBRE\",\n    [\n      \"19NOV01\",\n      \"19NOV02\"\n    ]\n  ],\n  ...\n]\n$ curl -X GET http:\/\/localhost:3000\/api\/stops\/PIERNE\n[\n  {\n    \"stop_id\": \"PIERNE01\",\n    \"stop_code\": \"257\",\n    \"stop_name\": \"PIERNE\",\n    \"stop_desc\": \"\",\n    \"stop_lat\": \"49.102273\",\n    \"stop_lon\": \"6.179863\",\n    \"location_type\": \"\",\n    \"parent_station\": \"http:\/\/lemet.fr\/screen\/index2.php?stop=257\"\n  },\n  {\n    \"stop_id\": \"PIERNE02\",\n    \"stop_code\": \"234\",\n    \"stop_name\": \"PIERNE\",\n    \"stop_desc\": \"\",\n    \"stop_lat\": \"49.102558\",\n    \"stop_lon\": \"6.179825\",\n    \"location_type\": \"\",\n    \"parent_station\": \"http:\/\/lemet.fr\/screen\/index2.php?stop=234\"\n  }\n]<\/pre><\/div><p>Working from the first try. It\u2019s not even funny!<\/p><h4 id=\"routes-passing-by-a-bus-stop\">Routes passing by a bus stop<\/h4><p>I said earlier that I\u2019d like the user to first select the stop, then the route, so we need to propose an endpoint to do that, and it will be bound to <code>\/api\/stops\/:name\/routes<\/code>.<\/p><p>This endpoint needs to return the different routes passing by the specified stops, and also for each of these route return the associated possible headsigns.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.api\n  (:require [monmet.gtfs :as gtfs]))\n; ...\n(defn direction-names-for-stop-and-route [stop route store]\n  (sort (filter not-empty\n                (flatten (map keys (get-in store [:data stop route]))))))\n(defn route-names-and-directions-for-stop [stop store]\n  (map #(hash-map % (direction-names-for-stop-and-route stop % store))\n       (route-names-for-stop stop store)))\n; ...<\/pre><\/div><p>Expose this in the REST API:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(GET \"\/api\/stops\/:name\/routes\" [name]\n  (response (api\/route-names-and-directions-for-stop name store)))<\/pre><\/div><p>Test it\u2019s working:<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ curl -X GET http:\/\/localhost:3000\/api\/stops\/REPUBLIQUE\/routes | jq '.'\n[\n   {\n      \"name\": \"Ligne 5\",\n      \"directions\": [\n         \"L5 - FORT MOSELLE\",\n         \"L5a - MAISON NEUVE\",\n         \"L5e - MAGNY PAR RUE AU BOIS\",\n         \"L5f - MAGNY PAR RUE DE POUILLY\"\n      ],\n      \"route\": {\n         \"route_id\": \"5-77\",\n         \"route_short_name\": \"5\",\n         \"route_long_name\": \"Ligne 5\",\n         \"route_desc\": \"\",\n         \"route_type\": \"3\",\n         \"route_url\": \"\"\n      }\n   },\n   {\n      \"name\": \"Proxis 113\",\n      \"directions\": [\n         \"P113 - POLE MULTIMODAL\",\n         \"P113 - POUILLY\"\n      ],\n      \"route\": {\n         \"route_id\": \"113-91\",\n         \"route_short_name\": \"113\",\n         \"route_long_name\": \"Proxis 113\",\n         \"route_desc\": \"\",\n         \"route_type\": \"3\",\n         \"route_url\": \"\"\n      }\n   }\n]<\/pre><\/div><p>Working as expected :)<\/p><p>Still we miss one piece of the puzzle, the time table.<\/p><p>Final piece of our API, we need to retrieve the times of day when a bus stops at our favorite bus stop. Besides, we want to query for multiple route+direction at the same stop, because sometimes you just want to go somewhere and there are multiple headsigns that goes where you want to go, so you don\u2019t care what bus you take.<\/p><p>Open the <code>api<\/code> namespace and write some code:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(defn times-for-stop-route-and-direction [stop route direction store]\n  (map #(assoc % :headsign direction)\n       (get (first (filter #(contains? (set(keys %)) direction)\n                           (get-in store [:data stop route]))) direction)))\n(defn today \n  \"Return the keyword for today, moday -&gt; :mon, and so on.\"\n  []\n  (-&gt; (LocalDate\/now) \n      .getDayOfWeek \n      (.getDisplayName TextStyle\/SHORT Locale\/ENGLISH) \n      .toLowerCase \n      keyword))\n(defn now \n  \"Return the time now.\"\n  [] \n  (-&gt; (LocalDateTime\/now)\n      (.format DateTimeFormatter\/ISO_LOCAL_TIME)\n      (subs 0 8)))\n(defn times-for-stop-and-multiple-route-direction [stop routes-directions store]\n  (-&gt;&gt; (parse-routes-directions routes-directions)\n       (map #(times-for-stop-route-and-direction stop (first %) (nth % 1) store))\n       (flatten)\n       (filter #(trip-for-day? (:trip-id %) (today) (:calendar store) (:trips store)))\n       (filter #(= 1 (compare (:arrival-time %) (now))))\n       (sort gtfs\/by-arrival-time)))<\/pre><\/div><p>Expose this in the REST API:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(GET \"\/api\/stops\/:name\/routes\/:routes\" [name routes]\n  (response (api\/times-for-stop-and-multiple-route-direction name routes store)))<\/pre><\/div><p>The format I decided is the following <code>Route:::Headsign<\/code>, multiple routes\/headsign can be separated by <code>:::<\/code>.<\/p><p>Let\u2019s encode some URI using this useful encoder that I just found on Google: <a href=\"http:\/\/meyerweb.com\/eric\/tools\/dencoder\/\">http:\/\/meyerweb.com\/eric\/tools\/dencoder\/<\/a>.<\/p><ul><li><code>Ligne 5<\/code> = <code>Ligne%205<\/code><\/li>\n<li><code>L5a - MAISON NEUVE<\/code> = <code>L5a%20-%20MAISON%20NEUVE<\/code><\/li>\n<li><code>L5 - FORT MOSELLE<\/code> = <code>L5%20-%20FORT%20MOSELLE<\/code><\/li>\n<\/ul><p>The URL to call is: <a href=\"http:\/\/localhost:8080\/api\/stops\/PIERNE\/routes\/Ligne%205:::L5a%20-%20MAISON%20NEUVE---Ligne%205:::L5%20-%20FORT%20MOSELLE\">http:\/\/localhost:8080\/api\/stops\/PIERNE\/routes\/Ligne%205:::L5a%20-%20MAISON%20NEUVE\u2014Ligne%205:::L5%20-%20FORT%20MOSELLE<\/a><\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ curl -X GET 'http:\/\/localhost:8080\/api\/stops\/PIERNE\/routes\/Ligne%205:::L5a%20-%20MAISON%20NEUVE---Ligne%205:::L5%20-%20FORT%20MOSELLE' \\\n    | jq 'map(.arrival_time + \" \" + .headsign)' | head -3\n[\n  \"05:19:05 L5a - MAISON NEUVE\",\n  \"05:43:05 L5a - MAISON NEUVE\",\n  \"06:05:32 L5a - MAISON NEUVE\",\n  \"06:22:32 L5a - MAISON NEUVE\",<\/pre><\/div><p>Exciting isn\u2019t it? These are all the bus I can take every morning to get to the train station :).<\/p><h4 id=\"can-you-spell-it\">Can you spell it?<\/h4><p>At first I wanted to create an iOS app for displaying it, but the iPhone\u2019s Shortcuts app can just fetch the URL and spell it.<\/p><p>Just duplicate the code for the last route we created, add <code>\/spell<\/code> to the URL and then just generate a French sentence fo it.<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\">(ns monmet.api\n  ...)\n; ...\n(defn make-sentence [{:keys [headsign arrival_time]}]\n  (str headsign \", arivant a \" (str\/replace (subs arrival_time 0 5) #\":\" \" heures \") \".\"))\n(defn times-for-stop-and-multiple-route-direction-spelled [stop routes-direction store]\n  (let [result (times-for-stop-and-multiple-route-direction stop routes-direction store)\n        next-3 (map make-sentence (take 3 result))]\n    (str \"Les 3 prochains bus a l'arret \\\"\" stop \"\\\" sont les suivants.\\n\\n- \" (apply str (interpose \"\\n- \" next-3)))))\n; ...<\/pre><\/div><p>and the route:<\/p><div class=\"highlight\"><pre class=\"language-clojure\" data-lang=\"clojure\"> (GET \"\/api\/stops\/:name\/routes\/:routes\/spell\" [name routes]\n       (response (api\/times-for-stop-and-multiple-route-direction-spelled name routes store)))<\/pre><\/div><p>Then run the server again, and launch ngrok:<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clj -m monmet.main &amp;\n$ ngrok http 8080\nngrok by @inconshreveable                                                                               (Ctrl+C to quit)                                              \nSession Status                online                \nSession Expires               7 hours, 51 minutes   \nVersion                       2.3.35                \nRegion                        United States (us)    \nWeb Interface                 http:\/\/127.0.0.1:4040 \nForwarding                    http:\/\/2769085b.ngrok.io -&gt; http:\/\/localhost:8080  \nForwarding                    https:\/\/2769085b.ngrok.io -&gt; http:\/\/localhost:8080<\/pre><\/div><p>Then just create a shortcut in your iPhone like this:<\/p><div class=\"c4\"><img src=\"https:\/\/grison.me\/images\/gtfs\/98264.jpg\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>And finally, ask Siri for it:<\/p><div class=\"c4\"><img src=\"https:\/\/grison.me\/images\/gtfs\/8236.jpg\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>BAM.<\/p><p>Done.<\/p><p>I have now three different shortcut for when I need to go to the train station or city center, or when I\u2019m in the city and wan\u2019t to get back home.<\/p><p>Works from anywhere, every day. I just need to download the GTFS files from time to time (two times per year).<\/p><p>All this with under 250 lines of badly written, non optimized Clojure \ud83e\udd20.<\/p><p>Until next time \ud83e\udd18 !<\/p>","created_at":"2020-04-05T01:04:45+0000","updated_at":"2020-04-05T01:04:45+0000","published_by":["Alexandre Grison"],"annotations":[],"mimetype":"text\/html","reading_time":30,"domain_name":"grison.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1163,"title":"SCRAM over SASL for SkySpark v3","url":"http:\/\/www.alienfactory.co.uk\/articles\/skyspark-scram-over-sasl","content":"<p class=\"lead\">I recently implemented SCRAM (SHA-256) over SASL for <a href=\"https:\/\/skyfoundry.com\/skyspark\/\">SkySpark v3<\/a> so I could use the REST API.<\/p><p class=\"lead\">SkySpark v3 sports a new authentication algorithm that must be used to communicate with it. But alas there is not much available that succinctly explains how it works.<\/p><p>What I did find though, was a bucket load of specifications and documentation that were as cryptic as the authentication mechanism itself!<\/p><p>So now I present to you, what would have been extremely helpful to me, a fully worked authentication conversation with SkySpark for SCRAM (SHA-256) over SASL.<\/p><p><img class=\"img-fluid\" src=\"http:\/\/www.alienfactory.co.uk\/images\/articles\/2016\/skyspark-scram-over-sasl.ogimage.jpg\" alt=\"Random Stock Photo!\" referrerpolicy=\"no-referrer\" \/><\/p><h2 id=\"overview\">Overview<\/h2><p>All the authentication documentation for SkySpark v3 talks of <code>SCRAM SHA-256<\/code>, or Salted Challenge Response Authentication Mechanism (SCRAM) with SHA-256. SCRAM defines how to encode an authentication message to send to the server. It uses the <code>PBKDF2<\/code> algorithm from the Public-Key Cryptography Standards (PKCS).<\/p><p>Simple Authentication and Security Layer (SASL) then defines a protocol of how to send \/ receive these authentication messages.<\/p><p>These SASL messages are then wrapped up in HTTP headers and the HTTP request is sent as usual.<\/p><p>If you want the low down on all of the above, here are links to the relevant parts of the specifications:<\/p><ul class=\"blendChildren\"><li><a href=\"http:\/\/project-haystack.org\/doc\/Auth\">Project HayStack: Authentication<\/a><\/li>\n<li><a href=\"https:\/\/skyfoundry.com\/doc\/docSkySpark\/Auth\">SkySpark Authentication<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/rfc5802#page-8\">RFC5802 - Salted Challenge Response Authentication Mechanism (SCRAM) - Sec. 3 Algorithm Overview<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/rfc4422#section-3\">RFC4422 - Simple Authentication and Security Layer (SASL) - Sec. 3 The Authentication Exchange<\/a><\/li>\n<li><a href=\"http:\/\/tools.ietf.org\/html\/rfc2898#section-5.2\">RFC2898 - Password-Based Cryptography Specification - Sec. 5.2 PBKDF2<\/a><\/li>\n<li><a href=\"https:\/\/tools.ietf.org\/html\/rfc7804\">RFC7804: Salted Challenge Response HTTP Authentication Mechanism<\/a><\/li>\n<li><a href=\"https:\/\/tools.ietf.org\/html\/rfc7615\">RFC7615: HTTP Authentication-Info and Proxy-Authentication-Info Response Header Fields<\/a><\/li>\n<\/ul><p>The Project Haystack and SkySpark documentation do have a sample client \/ server conversation (the SASL part), but they make no attempt to explain how they calculated the values and the given messages (the SCRAM part)!<\/p><p>Steps:<\/p><ol class=\"blendChildren c2\"><li><a href=\"#step1\">Hello!<\/a><\/li>\n<li><a href=\"#step2\">First Message<\/a><\/li>\n<li><a href=\"#step3\">Second Message<\/a><\/li>\n<li><a href=\"#step4\">Rest<\/a><\/li>\n<li><a href=\"#step5\">Fantom Code<\/a><\/li>\n<\/ol><p>The following example conversation assumes the following values, which are the same as those used in Project Haystack's <a href=\"http:\/\/project-haystack.org\/doc\/Auth\">authentication documentation<\/a>:<\/p><ul class=\"blendChildren\"><li>Username: <code>user<\/code><\/li>\n<li>Password: <code>pencil<\/code><\/li>\n<\/ul><p>Note that all example code listed here is written in <a href=\"http:\/\/fantom-lang.org\/\">Fantom<\/a>.<\/p><h2 id=\"step1\">1. Hello!<\/h2><p>The first request is to initiate the authentication conversation, sending the username we wish to authenticate as.<\/p><p>Note that the URL we authenticate against needs to be a real URL lest we get a 404. It also needs to be one that we're not going to be redirected from. Project specific URLs are fine, but for general purpose usage I find that <code>\/ui<\/code> works well in all scenarios.<\/p><p>Assuming SkySpark is running locally on port 8080 our first HTTP request looks like:<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Client Request: Hello<\/span>\nGET \/ui HTTP\/1.1\nHost: localhost:8080\nAuthorization: HELLO username=dXNlcg\n<\/pre><p>The username is just our username <code>user<\/code> Base64 encoded. But note the lack of trailing <code>=<\/code> characters that are usually used for padding. That's because the <code>Authorization<\/code> header uses Base64 URIs, which lack the padding.<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Server Response: Hello<\/span>\nHTTP\/1.1 401 Unauthorized\nWWW-Authenticate: scram handshakeToken=dXNlcg, hash=SHA-256\n<\/pre><p>SkySpark replies with the above, which tells us we're to use <code>SHA-256<\/code> for all our encoding. In general Project Haystack mechanisms this could also be <code>SHA-1<\/code> or <code>SHA-512<\/code>.<\/p><p>The <code>handshakeToken<\/code> is used by the server to keep track of the authentication conversation, similar to a session cookie. So we need to make sure we pass the same value back. (Just ignore the fact it looks like our encoded username - this may change!)<\/p><h2 id=\"step2\">2. First Message<\/h2><p>Send an authentication request to the server.<\/p><p>In the request we name the user we wish to authenticate as, and a nonce. The nonce is random sequence of characters and should be of cryptographic strength.<\/p><pre class=\"syntax microlight notForIe fantom\">clientNonce    := ... crypto strength random characters ...\n<span class=\"c3\">\/\/ --&gt; fyko+d2lbbFgONRv9qkxdawL<\/span>\nclientFirstMsg := <em>\"n=${userName},r=${clientNonce}\"<\/em>\n<span class=\"c3\">\/\/ --&gt; n=user,r=fyko+d2lbbFgONRv9qkxdawL<\/span>\n<\/pre><p>The first message is then Base64 URI encoded (no trailing <code>=<\/code> chars) and sent to the server.<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Client Request: 1st Message<\/span>\nGET \/ui HTTP\/1.1\nHost: localhost:8080\nAuthorization: SCRAM handshakeToken=dXNlcg,\n  data=bj11c2VyLHI9ZnlrbytkMmxiYkZnT05Sdjlxa3hkYXdM\n<\/pre><p>SkySpark would then respond with the following response:<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Server Response: 1st Message<\/span>\nHTTP\/1.1 401 Unauthorized\nWWW-Authenticate: scram handshakeToken=dXNlcg, hash=SHA-256,\n  data=cj1meWtvK2QybGJiRmdPTlJ2OXFreGRhd0xIbytWZ2s3cXZVT0tVd3VXTElXZzRsLzlTcmFHTUhFRSxzPXJROVpZM01udEJldVAzRTFURFZDNHc9PSxpPTEwMDAw\n<\/pre><p>If we Base64 decode the data attribute we get:<\/p><pre class=\"syntax microlight notForIe\">r=fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE,s=rQ9ZY3MntBeuP3E1TDVC4w==,i=10000\n<\/pre><p>From which we can derive the following variables:<\/p><pre class=\"syntax microlight notForIe fantom\">serverFirstMsg   := <em>\"r=fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE,s=rQ9ZY3MntBeuP3E1TDVC4w==,i=10000\"<\/em>\nserverNonce      := <em>\"fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE\"<\/em>\nserverSalt       := <em>\"rQ9ZY3MntBeuP3E1TDVC4w==\"<\/em>\nserverIterations := 10000\n<\/pre><h2 id=\"step3\">3. Second Message<\/h2><p>Send an encoded message that proves we have the password. This is hard part.<\/p><p>First we do the cryptographic stuff. Most languages have classes or libraries for doing this. Fantom uses Java's <a href=\"https:\/\/docs.oracle.com\/javase\/7\/docs\/api\/javax\/crypto\/spec\/PBEKeySpec.html\">javax.crypto.spec.PBEKeySpec<\/a>. Because we're using the SHA-256 hashing algorithm, we use <code>PBKDF2WithHmacSHA256<\/code>.<\/p><pre class=\"syntax microlight notForIe fantom\">dkLen           := 32   <span class=\"c3\">\/\/ the size in bytes of a SHA-256 hash<\/span>\nsaltedPassword  := Buf.pbk<strong>(<\/strong><em>\"PBKDF2WithHmacSHA256\"<\/em>, password, Buf.fromBase64<strong>(<\/strong>serverSalt<strong>)<\/strong>, serverIterations, dkLen<strong>)<\/strong>\n<span class=\"c3\">\/\/ --&gt; e3 72 cc 2f 65 c6 ac 00 51 e9 e8 de ef 4b ea a3 f0 72 08 a4 6e b2 9f a3 bf 68 ea 76 9e b8 3e e9<\/span>\n<\/pre><p><code>dkLen<\/code> is constant, but is dependent on which hashing algorithm we're using.<\/p><pre class=\"syntax microlight notForIe fantom\"><span class=\"c3\">\/\/ dkLen Values<\/span>\nSHA-1   : 20\nSHA-256 : 32\nSHA-512 : 64\n<\/pre><p>In the next string, <code>clientFinalNoPf<\/code> (client final no proof), note that <code>biws<\/code> is a constant and is the just the string <code>\"n,,\"<\/code> Base64 encoded. In some literature this is referred to as the <code>GS2 Header<\/code>.<\/p><p>The string <code>\"Client Key\"<\/code> is also constant and is used as a default message to be hashed by the (salted) password.<\/p><p>select all<\/p><pre class=\"syntax microlight notForIe fantom\">clientFinalNoPf := <em>\"c=biws,r=${serverNonce}\"<\/em>\n<span class=\"c3\">\/\/ --&gt; \"c=biws,r=fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE\"<\/span>\nauthMessage     := <em>\"${clientFirstMsg},${serverFirstMsg},${clientFinalNoPf}\"<\/em>\n<span class=\"c3\">\/\/ --&gt; \"n=user,r=fyko+d2lbbFgONRv9qkxdawL,r=fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE,s=rQ9ZY3MntBeuP3E1TDVC4w==,i=10000,c=biws,r=fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE\"<\/span>\nclientKey       := <em>\"Client Key\"<\/em>.toBuf.hmac<strong>(<\/strong><em>\"SHA-256\"<\/em>, saltedPassword<strong>)<\/strong>\n<span class=\"c3\">\/\/ --&gt; 26 ac fd 4f 40 f9 5e 8e 74 b2 b3 5f 88 cd 8b e4 35 da 89 db 8d ab cc a8 b9 fb de 34 49 f9 93 b8<\/span>\nstoredKey       := clientKey.toDigest<strong>(<\/strong><em>\"SHA-256\"<\/em><strong>)<\/strong>\n<span class=\"c3\">\/\/ --&gt; b6 2f 2a 50 c9 9e 42 27 46 85 5e 9a 60 fa 3c 71 39 f8 78 9a 70 60 46 19 4d ae 5c e8 cf 48 e5 37<\/span>\nclientSignature := authMessage.toBuf.hmac<strong>(<\/strong><em>\"SHA-256\"<\/em>, storedKey<strong>)<\/strong>\n<span class=\"c3\">\/\/ --&gt; 5b 60 ae 4a 75 d8 da 9c 05 3b 85 ef 36 b6 27 df 2c 0a c9 4c f1 65 8f cf 3d 00 e1 1e ee d6 e1 4c<\/span>\nclientProof     := xor<strong>(<\/strong>clientKey, clientSignature<strong>)<\/strong>\n<span class=\"c3\">\/\/ --&gt; 7d cc 53 05 35 21 84 12 71 89 36 b0 be 7b ac 3b 19 d0 40 97 7c ce 43 67 84 fb 3f 2a a7 2f 72 f4<\/span>\nclientFinal     := <em>\"${clientFinalNoPf},p=${clientProof.toBase64}\"<\/em>\n<span class=\"c3\">\/\/ --&gt; \"c=biws,r=fyko+d2lbbFgONRv9qkxdawLHo+Vgk7qvUOKUwuWLIWg4l\/9SraGMHEE,p=fcxTBTUhhBJxiTawvnusOxnQQJd8zkNnhPs\/KqcvcvQ=\"<\/span>\n<\/pre><p>The <code>xor()<\/code> method is an annoying little thing where each byte of each input have to be xored together. It is not a native Fantom function, but instead a method you have to write yourself. See <a href=\"#step5\">Fantom Code<\/a> for more details.<\/p><p>It is the <code>clientFinal<\/code> string that we send to the server in the next message. Like last time, we Base64 URI encode it and send it as the <code>data<\/code> attribute of the <code>Authorization<\/code> header:<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Client Request: 2nd Message<\/span>\nGET \/ui HTTP\/1.1\nHost: localhost:8080\nAuthorization: SCRAM handshakeToken=dXNlcg,\n  data=Yz1iaXdzLHI9ZnlrbytkMmxiYkZnT05Sdjlxa3hkYXdMSG8rVmdrN3F2VU9LVXd1V0xJV2c0bC85U3JhR01IRUUscD1mY3hUQlRVaGhCSnhpVGF3dm51c094blFRSmQ4emtObmhQcy9LcWN2Y3ZRPQ\n<\/pre><p>SkySpark should then respond with:<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Server Response: 2nd Message<\/span>\nHTTP\/1.1 200 Auth successful\nAuthentication-Info: authToken=xxxyyyzzz, hash=SHA-256,\n  data=dj1UenFKVlc4bk5uZ1o5ZzFiL1lXaU84cy9abEhxQkwyb3AxYmxSN0txZG1FPQ\n<\/pre><p>It is the <code>authToken<\/code> attribute that we're after. We can use in all subsequent HTTP requests to communicate with SkySpark.<\/p><p>But before we do; how do we know we're communicating with the correct server? How do we know our requests aren't be re-routed to some hackers server? Well now is our chance to validate the server and check that <em>it<\/em> also knows the user's password.<\/p><p>Base64 decoding the <code>data<\/code> attribute gives us:<\/p><pre class=\"syntax microlight notForIe\">v=TzqJVW8nNngZ9g1b\/YWiO8s\/ZlHqBL2op1blR7KqdmE=\n<\/pre><p>If the server <em>truly<\/em> knows the user's password then we should be able to compute the same value for <code>v<\/code>.<\/p><p>Note the string <code>\"Server Key\"<\/code> is constant and is used as a default message to be hashed by the (salted) password.<\/p><pre class=\"syntax microlight notForIe fantom\">serverKey       := <em>\"Server Key\"<\/em>.toBuf.hmac<strong>(<\/strong><em>\"SHA-256\"<\/em>, saltedPassword<strong>)<\/strong>\n<span class=\"c3\">\/\/ --&gt; 5a a1 fd ca 03 cb 46 42 45 ba 1b 94 67 a4 2c 9e 61 47 d6 da 9f cc c9 f2 bf 17 bc 4e ab 2c 1a 75<\/span>\nserverSignature := authMessage.toBuf.hmac<strong>(<\/strong><em>\"SHA-256\"<\/em>, serverKey<strong>)<\/strong>.toBase64\n<span class=\"c3\">\/\/ --&gt; \"TzqJVW8nNngZ9g1b\/YWiO8s\/ZlHqBL2op1blR7KqdmE=\"<\/span>\n<\/pre><p>Because <code>serverSignature<\/code> equals the value sent by the server, we can trust it and continue communicating with it.<\/p><h2 id=\"step4\">4. Rest<\/h2><p>Now we have the <code>authToken<\/code> we can use it in all our requests to the SkySpark REST API.<\/p><pre class=\"syntax microlight notForIe props\"><span class=\"c3\">\/\/ Client Request: 2nd Message<\/span>\nGET \/api\/demo\/about HTTP\/1.1\nHost: localhost:8080\nAuthorization: BEARER authToken=xxxyyyzzz\n<\/pre><h2 id=\"step5\">5. Fantom Code<\/h2><p>A complete reference implementation of the SCRAM protocol for retreiving an <code>authToken<\/code> from SkySpark, written in Fantom, is available in this <a href=\"https:\/\/bitbucket.org\/snippets\/fantomfactory\/AqAKL\">BitBucket Snippet<\/a>.<\/p><p>The sample code may be used like this:<\/p><pre class=\"syntax microlight notForIe fantom\"><span class=\"c3\">\/\/ get the authToken<\/span>\nauthToken := SkySparkAuth<strong>(<\/strong><strong>)<\/strong>.scram<strong>(<\/strong><em>`http:\/\/localhost:8080\/ui`<\/em>, <em>\"&lt;username&gt;\"<\/em>, <em>\"&lt;password&gt;\"<\/em><strong>)<\/strong>\necho<strong>(<\/strong><em>\"authToken: ${authToken}\"<\/em><strong>)<\/strong>\n<span class=\"c3\">\/\/ call the REST API<\/span>\nzincRes := WebClient<strong>(<\/strong><em>`http:\/\/localhost:8080\/api\/&lt;proj&gt;\/about`<\/em><strong>)<\/strong> <strong>{<\/strong>\n    <em>it<\/em>.reqHeaders<strong>[<\/strong><em>\"Authorization\"<\/em><strong>]<\/strong> = <em>\"BEARER authToken=${authToken}\"<\/em>\n<strong>}<\/strong>.getStr\n<\/pre><p>Note that the <a href=\"https:\/\/bitbucket.org\/brianfrank\/haystack-java\/overview\">haystack-java library<\/a> also contains a working SCRAM implementation.<\/p><h2 id=\"edits\">Edits<\/h2><ul class=\"blendChildren\"><li><em>26 December 2016<\/em> - Original article.<\/li>\n<li><em>18 May 2017<\/em> - Made the link to the Fantom implemetation much clearer.<\/li>\n<li><em>26 May 2018<\/em> - Added notes on <code>\"Client Key\"<\/code> and <code>\"Server Key\"<\/code> being constant.<\/li>\n<\/ul><hr \/><p>Written by Steve Eynon ( ) on 26th December 2016<br \/>Last edited on 26th May 2018<br \/>Tags: <a href=\"http:\/\/www.alienfactory.co.uk\/tags\/skyspark\">SkySpark<\/a>, <a href=\"http:\/\/www.alienfactory.co.uk\/tags\/security\">Security<\/a><\/p><p><\/p><h2> Discuss<\/h2>","created_at":"2020-04-03T12:51:23+0000","updated_at":"2020-04-03T12:51:23+0000","published_at":"2016-12-26T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":7,"domain_name":"www.alienfactory.co.uk","preview_picture":"http:\/\/www.alienfactory.co.uk\/coldFeet\/QrgtOw\/images\/articles\/2016\/skyspark-scram-over-sasl.ogimage.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1162,"title":"How a SQL database works","url":"http:\/\/calpaterson.com\/how-a-sql-database-works.html","content":"<p id=\"article-description\">The details on how tables and indexes work. Rows, pages, the heap and indexes are all covered, both on-disk layout and querying.<\/p>\n<hr \/><p>SQL (or relational) databases are some of oldest and probably still are the single most commonly used database technology - both in database server form (such as postgresql) or in library form (sqlite).<\/p>\n<p>Despite this, SQL databases' internal operations can be mysterious to a surprisingly large number of people - even those who work with them daily.<\/p>\n<p>Unlike other database technologies, such key-value stores or graph databases, relational databases are not based on a single data structure but a combination of several different data structures. Understanding this combination is not difficult and can help a lot when working with a SQL database.<\/p>\n<h2>The relation<\/h2>\n<p>Programmers most often work with just two primary data types: arrays and associative arrays.<\/p>\n<p><strong>Arrays<\/strong> allow for lists of objects, providing a way of grouping multiple objects together.<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/array.svg\" alt=\"diagram of an array\" referrerpolicy=\"no-referrer\" \/><figcaption>Example array holding cats in an imaginary animal shelter<\/figcaption><\/figure><p>Array implementation details vary widely but in general: an item can be added to an array at the end in a constant amount of time no matter the size of the array already. Searching for an given item is different: each entry in the array must be checked and this takes a length of time proportional to the size of the array at present (deletion has similar performance characteristics).<\/p>\n<table><thead><tr><th>Operation<\/th>\n<th>Time (worst-case)<\/th>\n<\/tr><\/thead><tbody><tr><td>Insert<\/td>\n<td>Constant \/ O(1)<\/td>\n<\/tr><tr><td>Search\/Delete<\/td>\n<td>Linear \/ O(n)<\/td>\n<\/tr><\/tbody><\/table><p><strong>Associative arrays<\/strong> (often called \"hash-tables\" after the most common implementation) allow for reference to items via a \"key\".<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/assoc-array.svg\" alt=\"diagram of an associate array\" referrerpolicy=\"no-referrer\" \/><figcaption>Example hash table holding the same cats, \"keyed\" by their registration number.<\/figcaption><\/figure><p>This key must be unique for each item but associate arrays typically allow for fast insert, search and delete <em>so long as the key is being used<\/em>. If the key isn't being used for the operation, the entire data structure must be traversed, just as with arrays.<\/p>\n<table><thead><tr><th>Operation<\/th>\n<th>Time (worst-case)<\/th>\n<\/tr><\/thead><tbody><tr><td>Insert<\/td>\n<td>Constant \/ O(1)<\/td>\n<\/tr><tr><td>Search\/Delete (by key)<\/td>\n<td>Constant \/ O(1)<\/td>\n<\/tr><tr><td>Search\/Delete (otherwise)<\/td>\n<td><strong>Linear \/ O(n)<\/strong><\/td>\n<\/tr><\/tbody><\/table><p>SQL databases use a different type: the <strong>relation<\/strong>.<\/p>\n<p>A relation is a generalisation of both of these data types. A relation is a set of tuples. Each tuple can hold the details of a single item but the individual fields are split out and accessible on the relation level. Within a relation, the tuple form is exactly the same for each tuple.<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/relation.svg\" alt=\"diagram of a relation\" referrerpolicy=\"no-referrer\" \/><figcaption>Example relation holding the same cats. Note that properties of the cats are broken out and available at the relation level.<\/figcaption><\/figure><p>Relations are useful because they expose the internal fields of what was previously an opaque object. Fields like age and breed are now open to operations on the relation level itself. This means that more complicated operations are possible on the relation without having to write specific functions to ask questions of objects.<\/p>\n<p>Being able to operate on fields directly also allows for easy definition of \"indexes\", which speed up access and search within the database.<\/p>\n<p>SQL databases' implementation of relations varies a little but are mostly a combination of two different underlying data structures: the <strong>heap file<\/strong> and <strong>b-trees<\/strong>. Consequently the operation times are generally logarithmic - meaning that they grow slowly as the size of the data-set grows.<\/p>\n<table><thead><tr><th>Operation<\/th>\n<th>Time (worst-case)<\/th>\n<\/tr><\/thead><tbody><tr><td>Insert<\/td>\n<td>Logarithmic \/ O(log n)<\/td>\n<\/tr><tr><td>Search\/Delete<\/td>\n<td>Logarithmic \/ O(log n)<\/td>\n<\/tr><\/tbody><\/table><p>The rest of this discussion will be about how these two underlying data structures work.<\/p>\n<h2>The row<\/h2>\n<p>SQL databases aren't a perfect representation of relations. Instead of relations and tuples they represent \"tables\" and \"rows\". Rows are functionally the same as tuples but tables have a quirk - duplicate rows are allowed. This does matter for writing queries but doesn't change the fundamentals of how things work so I'll treat them as the same and will use SQL terminology from now on.<\/p>\n<p>Rows are typically represented, both on-disk and in memory, as a short header followed a concatenation of row values.<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/row.svg\" alt=\"diagram of a row\" referrerpolicy=\"no-referrer\" \/><figcaption>Example row holding data on Alfred. If we assume that 4 bytes (32 bits) are used to store an int and strings use one byte per character (as is the case in ASCII) and are terminated in a \"NUL\" byte the length of row data here is 4 + 7 + 4 + 18 = 33 bytes. Rows are typically much longer in practice. The size of the header varies by database system but is usually kept as small as possible (27 bytes in postgresql).<\/figcaption><\/figure><p>The header will often contain information about the total length of the row. Depending on the specific implementation it might also contain other information, especially information about concurrency (ie: which transactions a given row is visible in).<\/p>\n<p>The rest of the row is just the values, one after another.<\/p>\n<p>There are two kinds of row values: those that are fixed length and those that are variable length. A 32-bit integer or a 64 bit float is an example of a fixed length value those can be represented directly. Strings, binary blobs, JSON, XML and other types can vary in length can be handled in one of two ways.<\/p>\n<p>One common method is to use a special \"terminator\" character (often NULL or <code>\\0<\/code>). This method means that null either cannot be allowed to appear in the string or that NULLs have some special handling to avoid ambiguity. The other alternative is to have a a leading length indicator to specify how many bytes long the string is.<\/p>\n<p>The order that the values of the row appear in could be included in the header but usually is table level metadata to avoid having to waste space repeatedly including information about what the order of the row values is.<\/p>\n<h2>The heap<\/h2>\n<p>Rows are stored in what is commonly called \"the heap\". This is where the main body of table data is held. The heap is broken down further into pages. Pages are short (often 4 or 8 kilobyte) collections of rows, again with a small header to indicate how many rows are present.<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/page.svg\" alt=\"diagram of a heap page\" referrerpolicy=\"no-referrer\" \/><figcaption>Example page. Note that the page has a fixed size and that free space is at the end. In practice many systems will make it possible for a row to cross pages.<\/figcaption><\/figure><p>Pages are a common pattern in systems which manage data that is moved frequently between disk and main memory. For exmaple, operating systems use pages in their virtual memory implementations.<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/heap.svg\" alt=\"diagram of a heap page\" referrerpolicy=\"no-referrer\" \/><figcaption>Example heap. Just a header and then page after page. The heap can grow to an unlimited size. Note that there is a gap where Page 3 has been deleted.<\/figcaption><\/figure><p>Working through pages, operations on the heap are simple. To perform a write, the relevant page is read from disk into memory if necessary, then edited and then written back to disk. Updates are similar. Pages can be appended to until there is no more space for any additional rows. Deletion is similarly simple: the page is rewritten without the unwanted row.<\/p>\n<p>Use of pages instead of storing rows directly in a file has a few advantages. The first is that pages are highly amenable to IO buffering, which helps performance. In many IO systems reading a whole page from disk takes the same amount of time as would reading a 40 byte row directly. Once a page has been read into memory it can be kept there as a cache until there is a shortage of memory and it has to be evicted from memory.<\/p>\n<p>The second reason is that access to pages by number is fast - the database system seeks to the start of the page ((page size * page number) + 1) and then reads forward as many bytes as the page size. This makes the heap and page system very efficient if there is an external reference to the page number.<\/p>\n<p>The third benefit is that pages allow fragmentation to be dealt with piecewise. Fragmentation happens when rows are deleted, creating gaps. If rows were written directly, these gaps would be difficult to fill - only a row of the same or equal size as the deleted row could be placed in the gap. With pages, fragmentation is limited as pages are written whole with the gaps moved to the end of the page. Perhaps when a page is completely empty a page from the end of the heap can be moved and written in this location instead.<\/p>\n<p>Heaps overall have great performance for insertion. A free space has to be found and then the row written into it. Depending on the method for finding this can be as simple as writing to the last page and calling it a day.<\/p>\n<p>The problem heaps don't solve is search. When you want to find all rows where a=1, you need to work through every page to find it. That's often called a \"full table scan\" and performs very poorly. Narrowing the number of pages that must be consulted improves performance hugely. What is required for this is an <strong>index<\/strong>.<\/p>\n<h2>The index<\/h2>\n<p>The crucial addition on top of the heap is the index. SQL databases overwhelmingly use B-trees (or minor variations) as indexes.<\/p>\n<p>B-trees were invented at Boeing in the 1970s. Despite allusions in the name, they aren't binary trees - though they are balanced. B-trees work in the following way: each node holds an ordered list. Each element in that list is either a value itself or a reference to another node that holds the value in question (or which holds yet another reference to that value).<\/p>\n<p>A node will usually fill a database page - so the ordered list will contain hundreds or thousands of elements. This makes b-trees very dense per-page and hopefully very shallow - ideal for looking up a key with the minimum of dereferences (each dereference could cause a page load). Some database systems use the same heap structure for heap pages and index pages.<\/p>\n<figure><img src=\"http:\/\/calpaterson.com\/assets\/diagrams\/b-tree.svg\" alt=\"diagram of a b-tree\" referrerpolicy=\"no-referrer\" \/><figcaption>Example b-tree. In reality they are likely to be thousands of entries wide - however they will usually be similarly shallow - often just 2 or 3 levels.<\/figcaption><\/figure><p>All operations on B-trees have logarithmic complexity: execution time grows slowly as the number of items in the index increases. This means that they remain fast even with large database sizes.<\/p>\n<p>For many it can be surpising the indices aren't hash tables. After all, don't hash tables have faster access and insert times (constant) than b-trees? There are a few reasons for this.<\/p>\n<p>Firstly, a hash table has a fixed size: if you have a hash function that outputs a two byte hash your hash table is limited to 65535 distinct locations<sup id=\"fnref:distinct-hashes\"><a class=\"footnote-ref\" href=\"#fn:distinct-hashes\">1<\/a><\/sup>. To deal with rising occupancy in your hashtable you have to \"rebalance\" - that is, increase the hash size and recalculate for all entries. Unfortunately rebalancing means having to recalculate every hash and hold both the new and old table in memory for a period.<\/p>\n<p>Second is that hash tables are relatively sparse in memory. A 32-bit hash table requires 2GB of ram. Assuming that a rebalance occurs once occupancy is greater than 75% that means that <em>at least<\/em> 500mb of memory is wasted - likely more. This sparseness means they compete with the heap for memory.<\/p>\n<p>The third reason is that hashing can't really help with operations where you want all values between two numbers - say, all cats adopted last week - as hash tables don't maintain order. This also causes hassle for the query optimiser as an estimate of rows returned cannot be provided until the query is under way.<\/p>\n<h2>The combination in practice<\/h2>\n<p>A single table will have a single heap and <em>multiple<\/em> indexes. Each index will document, for each entry, in which page (or pages) a the relevant row can be found. This allows for general logarithmic access by any key if the relevant index exists.<\/p>\n<h2>Further reading<\/h2>\n<ol><li>\n<p><a href=\"https:\/\/www.postgresql.org\/docs\/12\/internals.html\">Chapter 12 of the Postgres manual<\/a> gives a good overview of the disk layout of their page and heap system.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.librarything.com\/work\/12949629\">SQL performance explained<\/a> gives a great overview of how querying works in detail and what you can do to improve your queries.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/www.librarything.com\/work\/143113\">Transaction Processing: Concepts and Techniques<\/a> covers the same material but in a lot more detail (and much more besides).<\/p>\n<\/li>\n<\/ol><div class=\"footnote\"><hr \/><ol><li id=\"fn:distinct-hashes\">\n<p>\"Locations\", rather than \"entries\" because most hash tables point to arrays to handle collisions.\u00a0<a class=\"footnote-backref\" href=\"#fnref:distinct-hashes\" title=\"Jump back to footnote 1 in the text\">\u21a9<\/a><\/p>\n<\/li>\n<\/ol><\/div>","created_at":"2020-03-19T11:43:45+0000","updated_at":"2020-03-19T19:51:05+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"calpaterson.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1161,"title":"dylanaraps\/pywal","url":"https:\/\/github.com\/dylanaraps\/pywal","content":"<h3 class=\"c4\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/27b8d83cf31af8d8ce8407d2b22f7ab73fe9c665\/68747470733a2f2f692e696d6775722e636f6d2f3557674d4143652e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/27b8d83cf31af8d8ce8407d2b22f7ab73fe9c665\/68747470733a2f2f692e696d6775722e636f6d2f3557674d4143652e676966\" width=\"200px\" data-canonical-src=\"https:\/\/i.imgur.com\/5WgMACe.gif\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/h3>\n<p class=\"c1\">Generate and change color-schemes on the fly.<\/p>\n<p class=\"c1\"><a href=\"https:\/\/travis-ci.org\/dylanaraps\/pywal\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/14f2868b64866df3a72b9592e4d1f7144717813c\/68747470733a2f2f7472617669732d63692e6f72672f64796c616e61726170732f707977616c2e7376673f6272616e63683d6d6173746572\" data-canonical-src=\"https:\/\/travis-ci.org\/dylanaraps\/pywal.svg?branch=master\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/github.com\/dylanaraps\/pywal\/blob\/master\/LICENSE.md\"><img src=\"https:\/\/camo.githubusercontent.com\/890acbdcb87868b382af9a4b1fac507b9659d9bf\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d626c75652e737667\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-MIT-blue.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/pypi.python.org\/pypi\/pywal\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e891918fd4354418e3a785668a9cbf4e093e7812\/68747470733a2f2f696d672e736869656c64732e696f2f707970692f762f707977616c2e737667\" data-canonical-src=\"https:\/\/img.shields.io\/pypi\/v\/pywal.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/www.patreon.com\/dyla\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7382ca13d3ee0881bdd96e0c7b765bf85058cfa1\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646f6e6174652d70617472656f6e2d79656c6c6f772e737667\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/donate-patreon-yellow.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=V7QNJNKS3WYVS\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a38eef9a5ba09b8b7265b46548783956ffad5ee5\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646f6e6174652d70617970616c2d677265656e2e737667\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/donate-paypal-green.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/073eaf5ee3b23e3b8faccf8e8c01a160d66ef5d2\/68747470733a2f2f692e696d6775722e636f6d2f48684b334c44762e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/073eaf5ee3b23e3b8faccf8e8c01a160d66ef5d2\/68747470733a2f2f692e696d6775722e636f6d2f48684b334c44762e6a7067\" alt=\"img\" width=\"400px\" data-canonical-src=\"https:\/\/i.imgur.com\/HhK3LDv.jpg\" class=\"c3 c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>Pywal is a tool that generates a color palette from the dominant colors in an image. It then applies the colors system-wide and on-the-fly in all of your favourite programs.<\/p>\n<p>There are currently 5 supported color generation backends, each providing a different palette of colors from each image. You're bound to find an appealing color-scheme.<\/p>\n<p>Pywal also supports predefined themes and has over 250 themes built-in. You can also create your own theme files to share with others.<\/p>\n<p>The goal of Pywal was to be as out of the way as possible. It doesn't modify any of your existing configuration files. Instead it works around them and provides tools to integrate your system as you see fit.<\/p>\n<p>Terminal emulators and TTYs have their color-schemes updated in real-time with no delay. With minimal configuration this functionality can be extended to almost anything running on your system.<\/p>\n<h3>More: [<a href=\"https:\/\/github.com\/dylanaraps\/pywal\/wiki\/Installation\">Installation<\/a>] [<a href=\"https:\/\/github.com\/dylanaraps\/pywal\/wiki\/Getting-Started\">Getting Started<\/a>] [<a href=\"https:\/\/github.com\/dylanaraps\/pywal\/wiki\/Customization\">Customization<\/a>] [<a href=\"https:\/\/github.com\/dylanaraps\/pywal\/wiki\">Wiki<\/a>] [<a href=\"https:\/\/www.reddit.com\/r\/unixporn\/search?q=wal&amp;restrict_sr=on&amp;sort=relevance&amp;t=all\" rel=\"nofollow\">Screenshots<\/a>]<\/h3>","created_at":"2020-03-17T23:46:32+0000","updated_at":"2020-03-18T00:56:16+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/6799467?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1160,"title":"r\/unixporn - [StumpWM] Rice for the winter","url":"https:\/\/www.reddit.com\/r\/unixporn\/comments\/7atyou\/stumpwm_rice_for_the_winter\/","content":"<div class=\"SubredditVars-r-unixporn\"><p>Looks like you're using new Reddit on an old browser. The site may not work properly if you don't\u00a0<a class=\"vOGtEDdh1mVbkqinhg3Ov\" href=\"https:\/\/www.google.com\/chrome\/\">update your browser<\/a>! \u00a0If you do not update your browser, we suggest you visit\u00a0<a class=\"vOGtEDdh1mVbkqinhg3Ov\" href=\"https:\/\/old.reddit.com\">old reddit<\/a>.<\/p><header data-redditstyle=\"true\" class=\"_1tvdSTbdxaK-BnUbzUIqIY _2GyPfdsi-MbQFyHRECo9GO cx1ohrUAq6ARaXTX2u8YN _378Md-M6pOqwzEiV3lAwYk\"><div class=\"_2vkeRJojnV7cb9pMlPHy7d\"><div class=\"_3dnbqz69WJTFCss8wl7Wlk\">Press J to jump to the feed. Press question mark to learn the rest of the keyboard shortcuts<a aria-label=\"Home\" class=\"_30BbATRhFv3V83DHNDjJAO\" href=\"https:\/\/www.reddit.com\/\"><\/a><\/div><\/div>\n<\/header><\/div><div class=\"SubredditVars-r-unixporn _1nxEQl5D2Bx2jxDILRHemb _378Md-M6pOqwzEiV3lAwYk\"><div class=\"_3obgdFz00GvpqpuX8QCsNK _9BEqyBVXtRCQxyeYwqJMB _2NxbeazXeXu0bSJRcwSxd8 K4Eem3pMbRbAYioOfqN5E\"><div class=\"-DOLBAFWXMQX1Q3ErGO8I\"><p>172<\/p><div class=\"_2FCtq-QzlfuN-SwVMUZMM3 P9Qd6oTCWgLr3ackKg0I5 t3_7atyou\"><div class=\"_1hLrLjnE1G_RBCNcN9MVQf\"><img alt=\"\" src=\"https:\/\/www.redditstatic.com\/desktop2x\/img\/renderTimingPixel.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/div><\/div><div class=\"_2xu1HuBz1Yx6SP10AGVx_I lrzZ8b0L6AzLkQj5Ww7H1\"><a href=\"https:\/\/www.reddit.com\/r\/unixporn\/search?q=flair_name%3A%22Screenshot%22&amp;restrict_sr=1\">\n<p>Screenshot<\/p>\n<\/a><\/div><\/div><\/div><div tabindex=\"-1\" class=\"_1npCwF50X2J7Wt82SZi6J0 _3OGqXkiUb_0ZMlksb26boO\"><div class=\"u35lf2ynn4jHsVUwPmNU Dx3UxiK86VcfkFQVHNXNi _2gEUIJsfk7ZZ0euRwRyv08 _3KaECfUAGLfWQPO5eNjMNl\"><div class=\"uI_hDmU5GSiudtABRz_37\"><div id=\"t3_7atyou\" tabindex=\"-1\" class=\"_1oQyIsiPHYt6nx7VOmd1sz _2rszc84L136gWQrkwH6IaM Post t3_7atyou\"><div class=\"_3MC4c3Q_Y41YKtl1TcvyMt\"><div class=\"_3UMN4RCVY5288m_fOZlkcg\"><div class=\"BiNC74axuTz66dlnEgicY cZPZhMe-UCZ8htPodMyJ5\"><div class=\"_3AStxql1mQsrZuUIFP9xSg nU4Je7n-eSXStTBAPMYt8\">Posted by<p>Manjaro<\/p><\/div><\/div><div class=\"_2FCtq-QzlfuN-SwVMUZMM3 _2v9pwVh0VUYrmhoMv1tHPm t3_7atyou\"><div class=\"_1hLrLjnE1G_RBCNcN9MVQf\"><img alt=\"\" src=\"https:\/\/www.redditstatic.com\/desktop2x\/img\/renderTimingPixel.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/div><\/div><p><a href=\"https:\/\/imgur.com\/a\/NNlXK\" class=\"_13svhQIUZqD9PVzFcLwOKT styled-outbound-link\" rel=\"noopener noreferrer nofollow ugc\" target=\"_blank\">imgur.com\/a\/NNlX...<\/a><\/p><div class=\"_2fiIRtMpITeCAzXc4cANKp _1mK-LVHGTTlcFpMsjItjYJ lrzZ8b0L6AzLkQj5Ww7H1\"><a href=\"https:\/\/www.reddit.com\/r\/unixporn\/search?q=flair_name%3A%22Screenshot%22&amp;restrict_sr=1\">\n<p>Screenshot<\/p>\n<\/a><\/div><\/div><\/div><div class=\"_1hwEKkB_38tIoal6fcdrt9\"><div class=\"_3-miAEojrCvx_4FQ8x3P-s\"><a rel=\"nofollow\" data-click-id=\"comments\" data-test-id=\"comments-page-link-num-comments\" class=\"_1UoeAeSRhOKSNdY_h3iS1O _1Hw7tY9pMr-T1F4P1C-xNU _2qww3J5KKzsD7e5DO0BvvU\" href=\"https:\/\/www.reddit.com\/r\/unixporn\/comments\/7atyou\/stumpwm_rice_for_the_winter\/\">33 comments<\/a><\/div><p>96% Upvoted<\/p><\/div><\/div><div class=\"_1EjIqPTCvhReSe3IjZptiB _1DUKbp8va6vxOv9zemBDBi c10\"><div class=\"_2jpm-rNr0Hniw6BX3NWMVe\"><p>This thread is archived<\/p><p>New comments cannot be posted and votes cannot be cast<\/p><\/div><\/div><div class=\"_2ulKn_zs7Y3LWsOqoFLHPo\"><p>Sort by<\/p><\/div><div class=\"_2M2wOqmeoPVvcSsJ6Po9-V\"><div class=\"_1ump7uMrSA43cqok14tPrG _1oTUrVtKJk1ue0r3fe31kJ _1YCqQVO-9r-Up6QPB9H6_4 _1YCqQVO-9r-Up6QPB9H6_4 c15\"><div class=\"_3sf33-9rVAO_v4y0pIW_CH c13\"><p><a class=\"b57A3J7QBa7TvY8XeupVn\" href=\"https:\/\/www.reddit.com\/r\/unixporn\/comments\/7atyou\/stumpwm_rice_for_the_winter\/dpdvs3o\">Continue this thread\u00a0<\/a><\/p><\/div><div class=\"_3sf33-9rVAO_v4y0pIW_CH c13\"><p><a class=\"b57A3J7QBa7TvY8XeupVn\" href=\"https:\/\/www.reddit.com\/r\/unixporn\/comments\/7atyou\/stumpwm_rice_for_the_winter\/dpe1lu7\">Continue this thread\u00a0<\/a><\/p><\/div><div class=\"_3sf33-9rVAO_v4y0pIW_CH c13\"><p><a class=\"b57A3J7QBa7TvY8XeupVn\" href=\"https:\/\/www.reddit.com\/r\/unixporn\/comments\/7atyou\/stumpwm_rice_for_the_winter\/dpe2du2\">Continue this thread\u00a0<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"_2l7c_Oz0UVsamALvPrlznq\"><p>More posts from the unixporn community<\/p><a class=\"_1khvWSB4GNuM2XwSjluHzO Cvt0D6iVwU5TI0SY2nIGC c16\" href=\"https:\/\/www.reddit.com\/r\/unixporn\/\">\n<p>Continue browsing in r\/unixporn<\/p>\n<\/a><\/div><\/div><div class=\"_2Xq-4oyrEvHjL5U_EeMnK8 _2gEUIJsfk7ZZ0euRwRyv08 _20b4i5iUhjZQqDZ1BM_Q-9 _31AVRNJbVMh2DUFRrpCQcD\"><div class=\"_1G4yU68P50vRZ4USXfaceV _2mtWlchu4uQf339v56bSha TmgZY6tDcdErbE5d7E0HJ c18\" data-redditstyle=\"false\"><div class=\"_20Kb6TX_CdnePoT8iEsls6\"><img alt=\"Subreddit icon\" role=\"presentation\" src=\"https:\/\/styles.redditmedia.com\/t5_2sx2i\/styles\/communityIcon_7fixeonxbxd41.png\" class=\"_34CfAAowTqdbNDYXz5tBTW _2Vkdik1Q8k0lBEhhA_lRKE c17\" referrerpolicy=\"no-referrer\" \/><p><a class=\"_2D7eYuDY6cYGtybECmsxvE\" href=\"https:\/\/www.reddit.com\/r\/unixporn\/\">r\/unixporn<\/a><\/p><\/div><p>Submit screenshots of all your *NIX desktops, themes, and nifty configurations, or submit anything else that will make ricers happy. Maybe a server running on an Amiga, or a thinkpad signed by Bjarne Stroustrup? Show the world how sexy your computer can be!<\/p><div class=\"_7kAMkb9SAVF8xJ3L53gcW\"><div class=\"_3_HlHJ56dAfStT19Jgl1bF\"><p>663<\/p><p class=\"_29_mu5qI8E1fq6Uq5koje8\" id=\"IdCard--CurrentlyViewing--true--t5_2sx2i\">Online<\/p><\/div><\/div><hr class=\"_38lwnrIpIyqxDfAF1iwhcV\" \/><p>Created Sep 27, 2011<\/p><\/div><div class=\"_2wqyhtudP4weVGsZdVXJgt _1G4yU68P50vRZ4USXfaceV _3RPJ8hHnfFohktLZca18J6 c18\" data-redditstyle=\"false\"><p>Reddit Inc \u00a9 2020. All rights reserved<\/p><\/div><\/div><\/div><\/div>","created_at":"2020-03-08T16:01:34+0000","updated_at":"2020-03-08T16:01:34+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"www.reddit.com","preview_picture":"https:\/\/external-preview.redd.it\/rFY3abjkYr9DZSoLGw7X0UBqBIlNQNUSzmGI6F1GZeE.jpg?auto=webp&s=0c49b062a6cc35c6a3f5cfe81287f7e43ae1992e","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1159,"title":"How to Prepare Your Home and Kitchen for the Coronavirus Era (and Other Emergencies)","url":"https:\/\/www.epicurious.com\/expert-advice\/how-to-stock-kitchen-for-coronavirus-and-other-emergencies-article","content":"<p>Frozen vegetables such as <strong>spinach<\/strong>, <strong>kale<\/strong>, <strong>peas<\/strong>, <strong>corn<\/strong>, <strong>fava beans<\/strong>, and <strong>okra<\/strong> are great to have on hand, since fresh vegetables can\u2019t be kept in emergency supply without spoiling. Ditto frozen fruits like <strong>berries<\/strong> and <strong>mango<\/strong>. Frozen <strong>fish<\/strong>, particularly <strong>shrimp<\/strong>, can be extremely handy if you have the room for it. (If you have the resources to buy <a class=\"external-link\" data-event-click=\"{&quot;element&quot;:&quot;ExternalLink&quot;,&quot;outgoingURL&quot;:&quot;https:\/\/www.washingtonpost.com\/news\/food\/wp\/2015\/12\/16\/how-to-find-shrimp-thats-not-produced-by-slave-labor-in-thailand\/&quot;}\" href=\"https:\/\/www.washingtonpost.com\/news\/food\/wp\/2015\/12\/16\/how-to-find-shrimp-thats-not-produced-by-slave-labor-in-thailand\/\" rel=\"nofollow noopener\" target=\"_blank\">ethically-produced shrimp<\/a>, go for that.)<\/p><p>These foods require a modicum of preparation but are also great to have in an emergency supply if you have space: <strong>pasta<\/strong>, <strong>dried beans<\/strong>, <strong>grains<\/strong> (such as barley, oats, quinoa, rice, wheat berries, farro, etc.), <strong>nuts,<\/strong> <strong>hot<\/strong> <strong>cereal<\/strong>, <strong>coffee<\/strong>, and <strong>tea<\/strong>. Hydrating drinks such as <strong>sports drinks<\/strong> (Gatorade) and Pedialyte can also be useful if somebody in your family becomes sick.<\/p><p>If you have a lot of space and can afford it, you might consider these non-essential items: shelf-stable <strong>jarred salsa<\/strong>, bags of <strong>tortilla chips or pita chips<\/strong>, jars of <strong>pickles or preserved vegetables<\/strong> (including roasted red peppers, olives, and artichoke hearts), bouillon concentrate, and bars of <strong>chocolate<\/strong> (be honest\u2014you get cranky if you don\u2019t have it).<\/p><p>The CDC also recommends keeping a <strong>30-day supply of medicines\/prescriptions<\/strong> if your insurance allows for it. And since hand-washing is the first defense against germs, a supply of <strong>hand soap<\/strong> is good to have (may as well throw <strong>dish soap<\/strong> in there, too).<\/p><p>Finally, if you <em>really<\/em> have a lot of space, and you\u2019re a baker, you might want to have <strong>flour<\/strong>, <strong>sugar<\/strong>, <strong>baking powder<\/strong>, <strong>baking soda<\/strong>, <strong>yeast<\/strong>, <strong>shelf-stable milk<\/strong>, and other baking ingredients on hand. Being under quarantine doesn\u2019t necessarily mean you\u2019ll be feeling ill\u2014you might simply be restless. Distractibaking could help.<\/p><h2>So I should just go to the store and fill my cart with as much random food as I can afford?<\/h2><p>Not if you can help it. Ideally, you want to shop for your emergency stash with a strategy in mind. How are you going to eat this food? How will you make sure your meals have nutritional variety? Are you <em>really<\/em> going to want to eat chickpeas and spaghetti for two weeks straight?<\/p><p>\u201cBringing a meal-planning lens into this is really important,\u201d Sipos says. \u201cGo [to the store] with a plan. Have a list of recipes. Probably pretty simple recipes.\u201d<\/p><p><em><strong>To find our 14-day meal plan of simple recipes, complete with grocery list, <a href=\"https:\/\/www.epicurious.com\/recipes-menus\/all-pantry-meal-plan-for-coronavirus-and-other-emergencies-article\">click here<\/a>.<\/strong><\/em><\/p><h2>What about the food already in my fridge?<\/h2><figure class=\"asset-embed\"><div class=\"asset-embed__asset-container\"><noscript>\n<p><img alt=\"Image may contain Appliance Refrigerator and Plant\" class=\"responsive-image__image\" src=\"https:\/\/assets.epicurious.com\/photos\/5a1dab061b63077aceb8974e\/6:4\/w_1600%2Cc_limit\/11-Groceries-That-Will-Last-in-the-Fridge-for-Two-Weeks-28112017.jpg\" srcset=\"https:\/\/assets.epicurious.com\/photos\/5a1dab061b63077aceb8974e\/6:4\/w_1600%2Cc_limit\/11-Groceries-That-Will-Last-in-the-Fridge-for-Two-Weeks-28112017.jpg 1600w, https:\/\/assets.epicurious.com\/photos\/5a1dab061b63077aceb8974e\/6:4\/w_1280%2Cc_limit\/11-Groceries-That-Will-Last-in-the-Fridge-for-Two-Weeks-28112017.jpg 1280w, https:\/\/assets.epicurious.com\/photos\/5a1dab061b63077aceb8974e\/6:4\/w_1024%2Cc_limit\/11-Groceries-That-Will-Last-in-the-Fridge-for-Two-Weeks-28112017.jpg 1024w, https:\/\/assets.epicurious.com\/photos\/5a1dab061b63077aceb8974e\/6:4\/w_768%2Cc_limit\/11-Groceries-That-Will-Last-in-the-Fridge-for-Two-Weeks-28112017.jpg 768w, https:\/\/assets.epicurious.com\/photos\/5a1dab061b63077aceb8974e\/6:4\/w_640%2Cc_limit\/11-Groceries-That-Will-Last-in-the-Fridge-for-Two-Weeks-28112017.jpg 640w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><\/div>\n<figcaption class=\"caption asset-embed__caption\">When your fridge starts to look like this\u2014keep eating.\nPhoto by Caleb Adams<\/figcaption><\/figure>","created_at":"2020-03-06T14:25:45+0000","updated_at":"2020-03-06T14:25:45+0000","published_at":"2020-03-02T18:19:39+0000","published_by":["The Editors of Epicurious"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.epicurious.com","preview_picture":"https:\/\/assets.epicurious.com\/photos\/5cc882d2327475b9c883ad2d\/16:9\/w_1280,c_limit\/A-Sustainable-Pantry-04252019.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1158,"title":"Please turn on JavaScript in your browser to continue","url":"https:\/\/accounts.google.com\/o\/noscript","content":"Please turn on JavaScript in your browser to continue\n<p>You've reached this page because we have detected that Javascript is disabled in your browser. The page you attempted to load cannot display properly if scripts are disabled.<\/p>\n<p>Please enable scripts and retry the operation or go back in your browser.<\/p>","created_at":"2020-03-05T13:49:30+0000","updated_at":"2020-03-05T13:49:30+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"accounts.google.com","http_status":"200","origin_url":"https:\/\/metrics.lola.com\/grafana\/d\/dn-mVZPWk\/threev?tab=queries&orgId=1&fullscreen&edit&panelId=10"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1157,"title":"Masonry","url":"https:\/\/masonry.desandro.com\/","content":"<div class=\"hero container hero-grid clearfix\" data-js=\"hero\"><div class=\"hero-grid__item hero-grid__item--texty hero-grid__item--width3\"><p class=\"hero__tagline\">Cascading grid layout library<\/p><\/div><div class=\"hero-grid__item\"><a class=\"hero-grid__big-button\" href=\"https:\/\/unpkg.com\/masonry-layout@4\/dist\/masonry.pkgd.min.js\"> Download masonry.\u200bpkgd.\u200bmin.js<\/a><\/div><div class=\"hero-grid__item\"><a class=\"hero-grid__big-button\" href=\"https:\/\/masonry.desandro.com\/masonry-docs.zip\"> Download these docs<\/a><\/div><div class=\"hero-grid__item\"><a class=\"hero-grid__big-button\" href=\"https:\/\/github.com\/desandro\/masonry\"> Masonry on GitHub<\/a><\/div><div class=\"hero-grid__item hero-grid__item--texty hero-grid__item--width2\"><h2 class=\"hero__what-is-title\">What is Masonry?<\/h2><p class=\"hero__what-is-description\">Masonry is a JavaScript grid layout library. It works by placing elements in optimal position based on available vertical space, sort of like a mason fitting stones in a wall. You\u2019ve probably seen it in use all over the Internet.<\/p><\/div><\/div><div id=\"content\" class=\"content container\"><ul class=\"page-nav\"><li class=\"page-nav__item page-nav__item--h2\"><a href=\"#install\">Install<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#download\">Download<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#cdn\">CDN<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#package-managers\">Package managers<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h2\"><a href=\"#getting-started\">Getting started<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#html\">HTML<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#css\">CSS<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#initialize-with-jquery\">Initialize with jQuery<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#initialize-with-vanilla-javascript\">Initialize with Vanilla JavaScript<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h3\"><a href=\"#initialize-in-html\">Initialize in HTML<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h2\"><a href=\"#next\">Next<\/a><\/li>\n<li class=\"page-nav__item page-nav__item--h2\"><a href=\"#mit-license\">MIT License<\/a><\/li>\n<\/ul><h2 id=\"install\">Install<\/h2><h3 id=\"download\">Download<\/h3><ul><li><a href=\"https:\/\/unpkg.com\/masonry-layout@4\/dist\/masonry.pkgd.min.js\">masonry.pkgd.min.js<\/a> minified, or<\/li>\n<li><a href=\"https:\/\/unpkg.com\/masonry-layout@4\/dist\/masonry.pkgd.js\">masonry.pkgd.js<\/a> un-minified<\/li>\n<\/ul><h3 id=\"cdn\">CDN<\/h3><p>Link directly to Masonry files on <a href=\"https:\/\/unpkg.com\">unpkg<\/a>.<\/p><pre class=\"html\">&lt;script src=\"https:\/\/unpkg.com\/masonry-layout@4\/dist\/masonry.pkgd.min.js\"&gt;&lt;\/script&gt;\n&lt;!-- or --&gt;\n&lt;script src=\"https:\/\/unpkg.com\/masonry-layout@4\/dist\/masonry.pkgd.js\"&gt;&lt;\/script&gt;\n<\/pre><h3 id=\"package-managers\">Package managers<\/h3><p>Install with <a href=\"https:\/\/bower.io\">Bower<\/a>: \u00a0<code>bower install masonry --save<\/code><\/p><p><a href=\"https:\/\/www.npmjs.com\/package\/masonry-layout\">Install with npm<\/a>: \u00a0<code>npm install masonry-layout<\/code><\/p><h2 id=\"getting-started\">Getting started<\/h2><h3 id=\"html\">HTML<\/h3><p>Include the Masonry <code>.js<\/code> file in your site.<\/p><pre class=\"html\">&lt;script src=\"\/path\/to\/masonry.pkgd.min.js\"&gt;&lt;\/script&gt;\n<\/pre><p>Masonry works on a container grid element with a group of child items.<\/p><pre class=\"html\">&lt;div class=\"grid\"&gt;\n  &lt;div class=\"grid-item\"&gt;...&lt;\/div&gt;\n  &lt;div class=\"grid-item grid-item--width2\"&gt;...&lt;\/div&gt;\n  &lt;div class=\"grid-item\"&gt;...&lt;\/div&gt;\n  ...\n&lt;\/div&gt;\n<\/pre><h3 id=\"css\">CSS<\/h3><p>All sizing of items is handled by your CSS.<\/p><pre class=\"css\">.grid-item { width: 200px; }\n.grid-item--width2 { width: 400px; }\n<\/pre><h3 id=\"initialize-with-jquery\">Initialize with jQuery<\/h3><p>You can use Masonry as a jQuery plugin: <code>$('selector').masonry()<\/code>.<\/p><pre class=\"js\">$('.grid').masonry({\n  \/\/ options\n  itemSelector: '.grid-item',\n  columnWidth: 200\n});\n<\/pre><h3 id=\"initialize-with-vanilla-javascript\">Initialize with Vanilla JavaScript<\/h3><p>You can use Masonry with vanilla JS: <code>new Masonry( elem, options )<\/code>. The <code>Masonry()<\/code> constructor accepts two arguments: the container element and an options object.<\/p><pre class=\"js\">var elem = document.querySelector('.grid');\nvar msnry = new Masonry( elem, {\n  \/\/ options\n  itemSelector: '.grid-item',\n  columnWidth: 200\n});\n\/\/ element argument can be a selector string\n\/\/   for an individual element\nvar msnry = new Masonry( '.grid', {\n  \/\/ options\n});\n<\/pre><h3 id=\"initialize-in-html\">Initialize in HTML<\/h3><p>You can initialize Masonry in HTML, without writing any JavaScript. Add <code>data-masonry<\/code> attribute to the container element. <a href=\"https:\/\/masonry.desandro.com\/options.html\">Options<\/a> can be set in its value.<\/p><pre class=\"html\">&lt;div class=\"grid\" data-masonry='{ \"itemSelector\": \".grid-item\", \"columnWidth\": 200 }'&gt;\n<\/pre><p class=\"edit-demo\"><a href=\"https:\/\/codepen.io\/desandro\/pen\/vLeBLV\">Edit this demo on CodePen<\/a><\/p><div class=\"call-out\"><p>Options set in HTML must be valid JSON. Keys need to be quoted, for example <code>\"itemSelector\":<\/code>. Note the value of <code>data-masonry<\/code> is set with single quotes <code>'<\/code>, but JSON entities use double-quotes <code>\"<\/code>.<\/p><\/div><p>HTML initialization was previously done with a class of <code>js-masonry<\/code> and setting options in <code>data-masonry-options<\/code> in Masonry v3. Masonry v4 is backwards compatible with this code.<\/p><pre class=\"html\">&lt;div class=\"grid js-masonry\"\n  data-masonry-options='{ \"itemSelector\": \".grid-item\", \"columnWidth\": 200 }'&gt;\n<\/pre><h2 id=\"next\">Next<\/h2><p>Learn more about how to use Masonry:<\/p><ul><li><a href=\"https:\/\/masonry.desandro.com\/layout.html\">Layout<\/a><\/li>\n<li><a href=\"https:\/\/masonry.desandro.com\/options.html\">Options<\/a><\/li>\n<li><a href=\"https:\/\/masonry.desandro.com\/methods.html\">Methods<\/a><\/li>\n<\/ul><h2 id=\"mit-license\">MIT License<\/h2><p>Masonry is released under the <a href=\"https:\/\/desandro.mit-license.org\/\">MIT License<\/a>. Have at it.<\/p><\/div>","created_at":"2020-03-05T02:46:31+0000","updated_at":"2020-03-05T02:46:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"masonry.desandro.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1156,"title":"Uses This: Katherine Cox-Buday","url":"https:\/\/usesthis.com\/interviews\/katherine.cox-buday\/","content":"<h3>Who are you, and what do you do?<\/h3><p>Hi there, internet! I'm <a href=\"https:\/\/katherine.cox-buday.com\/\" title=\"Katherine's website.\">Katherine Cox-Buday<\/a>, and I'm a writer, computer scientist, and mother. I'm currently working at <a href=\"https:\/\/www.digitalocean.com\/\" title=\"An SSD-based web hosting service.\">DigitalOcean<\/a> on software that helps us manage our fleet of servers. I'm also author of \"<a href=\"https:\/\/katherine.cox-buday.com\/concurrency-in-go\/\" title=\"Katherine's book on Go concurrency.\">Concurrency in Go<\/a>\", and a few unpublished novels. When I can find time, I contribute to a smattering of open-source projects which generally follow my interests\/needs.<\/p><h3>What hardware do you use?<\/h3><p>Let's take a walk down my stack, yes? As someone who holds computers near to her heart, my setup is at once sprawling, and boring. Let's start with personal tech and work out from there.<\/p><p>I used to wear a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pebble_Time\" title=\"A smartwatch.\">Pebble Time<\/a> watch which I loved dearly. But Pebble was purchased by Fitbit, and then my Pebble Time's battery died, and I found myself in the market for a new watch. I do what I can to maintain control of my own personal data, and so setting up a pipeline directly from my body to a tech company wasn't very appealing to me. Also, I was accustomed to multi-day battery life in watches, and the number of watches on the market which have this is small. Eventually, I landed on an Amazfit <a href=\"https:\/\/us.amazfit.com\/shop\/bip?variant=336750\" title=\"A smartwatch.\">Bip<\/a> paired with <a href=\"http:\/\/gadgetbridge.org\/\" title=\"Android software for using smartwatch devices.\">Gadgetbridge<\/a>. I charge it every month or so, and Gadgetbridge ensures my data stays local. This has worked pretty well for me.<\/p><p>I'm currently using a first generation Google <a href=\"https:\/\/store.google.com\/product\/pixel_phone\" title=\"A 5 inch Android smartphone.\">Pixel<\/a>. I have mixed feelings about smartphones. I love having a tiny computer in my pocket with a pretty fast connection to the internet, but I am not convinced that they have provided a net positive effect on humanity. I hear encouraging stories about what smartphones have enabled in developing countries (i.e. commerce, community, education, etc.), but I also see <a href=\"https:\/\/www.wired.com\/2016\/01\/facebook-zuckerberg-internet-org\/\" title=\"A Wired article about Facebook's goal of connecting the whole world.\">stories<\/a> about tech companies positioning themselves as the gatekeepers to the wider internet. The two dominant operating systems for phones have impenetrable privacy models, and it is often difficult to tell what apps are accessing and sharing, and with whom. Never in human history have we created such a pervasive and capable engine for monitoring what humans are doing. Most phones are not made in such a way that they can be repaired: their support is artificially ended before the hardware has reached the end of its usefulness, and their batteries are not replaceable. And some phones are seen as status symbols. These things have helped drive a pump which is producing tremendous amounts of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Electronic_waste\" title=\"The Wikipedia entry on electronic waste.\">electronic waste<\/a>. In addition, phones can help drive addictive behavior, and an entire industry has popped into existence to help companies produce software which engages their users more frequently, and for longer periods of time. We are only beginning to see the effects their usage are having on young brains.<\/p><p>For my next phone, I'm looking at purchasing either a <a href=\"https:\/\/puri.sm\/products\/librem-5\/\" title=\"A 5.7 inch Linux smartphone.\">Librem 5<\/a>, or a <a href=\"https:\/\/www.pine64.org\/pinephone\/\">PinePhone<\/a>. Both are Linux based phones which I hope will help me address all of the downsides I outlined. I view the dearth of software for these devices as a feature which will help me to use them less. My hope is that after purchasing one of these, I won't have to purchase another for a long time.<\/p><p>For my work laptop, I always take whatever the Linux offering is. In the past that has been a Dell <a href=\"https:\/\/www.dell.com\/us\/p\/xps-13-9333\/pd\" title=\"A 13 inch PC laptop.\">XPS 13<\/a>, and currently I'm using a <a href=\"https:\/\/www.lenovo.com\/us\/en\/laptops\/thinkpad\/thinkpad-t-series\/ThinkPad-T480s\/p\/22TP2TT480S\" title=\"A 14 inch PC laptop.\">ThinkPad T480s<\/a>. The most important thing to me in a work machine is performance and Linux compatibility. I.e., I want to spend as little time possible preparing or waiting to work. It's nice when the laptop is a good form-factor, but most of the time it's docked, so that isn't too much of a concern.<\/p><p>My personal laptop is a System76 <a href=\"https:\/\/system76.com\/laptops\/gazelle\" title=\"A Linux laptop.\">9th generation Gazelle Pro<\/a> from 2014. I tend to hang onto hardware for a long time both out of environmental concerns, and because nothing I do generally requires a refresh. But even by my standards, this thing is getting long in the tooth. I've dropped it a few times, and the monitor's hinges are broken on one side, and sharp.<\/p><p>For the little time I have to play games, I have <a href=\"https:\/\/en.wikipedia.org\/wiki\/Windows_10\" title=\"An operating system.\">Windows 10<\/a> installed on an old System76 Leopard Extreme (leox3 if that means anything to you). It sits in my homelab and runs headless so I can stream games to other machines in the house via <a href=\"https:\/\/store.steampowered.com\/\" title=\"A digital game distribution service.\">Steam<\/a> <a href=\"https:\/\/store.steampowered.com\/streaming\/\" title=\"Details on streaming games from Steam.\">Remote Play<\/a>. This used to be my main computer, but I wanted a gaming machine, and since my laptop was adequate for my productivity needs, I frankensteined this into something that could run most of the games I'm interested in. The hardware isn't great anymore, but I don't find myself playing many new games, so it works just fine. I've replaced the CPU's liquid cooling pump once.<\/p><p>For my home's infrastructure, I ran Cat 5e throughout the house which terminates into a Ubuiquiti <a href=\"https:\/\/www.ui.com\/unifi-switching\/unifi-switch-2448\/\" title=\"A networking switch.\">Unifi 24<\/a> port switch. My edge-router is a PC Engines <a href=\"https:\/\/www.pcengines.ch\/apu2.htm\" title=\"A networking router.\">apu2<\/a> running <a href=\"http:\/\/www.openbsd.org\/\" title=\"An open-source operating system emphasising security and cryptography.\">OpenBSD<\/a>. Being the thing that sits between my personal tech and the internet, I wanted something I was reasonably confident in being as secure as possible, and OpenBSD has that reputation. I use <a href=\"https:\/\/www.ansible.com\/\" title=\"An IT automation service.\">Ansible<\/a> to manage this so I can bring a new one up from scratch quickly. I have a <a href=\"https:\/\/unifi-hd.ui.com\/\">UniFi HD<\/a> running <a href=\"https:\/\/openwrt.org\/\" title=\"An embedded Linux distribution for wireless routers.\">OpenWRT<\/a> for the things which move around the house. And I have a <a href=\"https:\/\/www.synology.com\/en-us\/products\/DS1819+\" title=\"An 8-bay NAS.\">Synology DS1819+<\/a> NAS for running some light workloads and providing on-site storage. This setup is continually evolving (hopefully towards simplicity) as I can find time to work on it.<\/p><p>For headspace-work, I have a buckwheat shell filled meditation cushion, and the timer on my watch.<\/p><h3>And what software?<\/h3><p>It wouldn't make sense to start out with anything other than <a href=\"http:\/\/www.gnu.org\/software\/emacs\/\" title=\"A free open-source text editor.\">Emacs<\/a>. I don't think there has been a piece of software which has had a larger impact on my life. I began using this about fifteen years ago, and it has followed me across operating systems, jobs, roles (I used it to manage my teams), languages, and needs. Every time I start something new, Emacs has been there to make it just a little easier, and the more I do in it, the easier everything gets. I believe this power comes from Emacs being the closest thing we have to a working <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lisp_machine\" title=\"The Wikipedia entry for Lisp machines.\">Lisp Machine<\/a>. If you know a little emacs-lisp, you can begin down the path of creating your own perfect tool which will grow with you, for life.<\/p><p>Sitting on top of Emacs is <a href=\"https:\/\/orgmode.org\/\" title=\"An Emacs mode for notes and to-do items.\">Org mode<\/a>. It is the thing which made irrelevant my search for the perfect task management software. Like Emacs, you can mold it into whatever workflow works best for you at the time. Later, I discovered it is also a wonderful <em>publishing<\/em> platform (this document was created with Org Mode). I have used it to author countless <a href=\"https:\/\/katherine.cox-buday.com\/blog\/2015\/03\/14\/writing-specs-with-org-mode\/\" title=\"Katherine's post about writing tech specs with Org mode.\">technical specifications<\/a>, my <a href=\"https:\/\/katherine.cox-buday.com\/blog\/2015\/01\/02\/migrating-to-a-statically-generated-blog\/\" title=\"Katherine's post on switching to a static weblog.\">blog<\/a>, and all of my books. In \"Concurrency in Go\", it allowed me to execute the code snippets embedded in the book \u2013 a form of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Literate_programming\" title=\"The Wikipedia entry on literate programming.\">literate programming<\/a>. This ensured that the code people are reading, the output from that code, and the code exported into the book's <a href=\"https:\/\/github.com\/kat-co\/concurrency-in-go-src\" title=\"Katherine's GitHub repo for her Go book.\">repository<\/a> all have the same provenance.<\/p><p>As for Operating Systems, with the exception of my gaming machine and edge router, everything runs Linux. I cannot foresee myself entrusting my external brain to anything not open source, and currently, Linux is the thing I'm most familiar with. I've pretty consistently used <a href=\"https:\/\/www.ubuntu.com\/\" title=\"A Unix distribution.\">Ubuntu<\/a> since 2005 (after much distro-hopping). It's certainly not perfect, but it usually \"just works\", and I don't have to fuss with it too much.<\/p><p>A couple of years ago, I added <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/html_node\/Package-Management.html\" title=\"A package management system.\">Guix<\/a> to the mix. It's both a Linux distribution, and a package manager; I use it as the latter on all my machines. Guix is <a href=\"https:\/\/guix.gnu.org\/manual\/en\/html_node\/Features.html#Features\" title=\"A list of Guix's features.\">awesome<\/a> for more reasons than I can enumerate here, but here's the short-list: it provides transactional upgrades (i.e. if something goes wrong, roll back to the previous state of your system); it allows for multiple versions of packages to be installed simultaneously; it can <a href=\"https:\/\/guix.gnu.org\/manual\/en\/html_node\/Invoking-guix-pack.html#Invoking-guix-pack\" title=\"Details on Guix's ability to build a container image.\">produce container images<\/a> or <a href=\"https:\/\/guix.gnu.org\/manual\/en\/html_node\/Invoking-guix-environment.html#Invoking-guix-environment\" title=\"Details on Guix's ability to build different dev environments.\">development environments<\/a> from package definitions, or even a list of packages; all of it is laid out declaratively in <a href=\"https:\/\/www.gnu.org\/software\/guile\/\" title=\"A programming language.\">Guile<\/a>, a lisp; and it also places a lot of emphasis on <a href=\"https:\/\/reproducible-builds.org\/\" title=\"A site detailing development practises for creating reproducible software builds.\">reproducible builds<\/a>. When you use it as an operating system, all of these benefits apply to the entire system. All in all, Guix, or something like it, feels like what the future of our current style of operating systems will look like.<\/p><p>I write software in various languages, but my two favorites are <a href=\"https:\/\/golang.org\/\" title=\"A compiled programming language.\">Go<\/a> and <a href=\"https:\/\/common-lisp.net\/\" title=\"A programming language.\">Common Lisp<\/a>. Go for what is not possible, Common Lisp for what is. Most of my personal projects are in Common Lisp because it allows one person to go farther, faster, and back-fill those things with safety checks like typing.<\/p><p>It's for this reason I use <a href=\"https:\/\/stumpwm.github.io\/\" title=\"A window manager for X.\">StumpWM<\/a> as my window manager (although I used <a href=\"https:\/\/i3wm.org\/\" title=\"An X window manager.\">i3<\/a> for a <em>long<\/em> time). Because it's written in Common Lisp, I can modify it pretty extensively.<\/p><p>I use <a href=\"https:\/\/www.mozilla.org\/en-US\/firefox\/new\/\" title=\"A cross-platform open-source web browser.\">Firefox<\/a> as my browser (although I keep experimenting with <a href=\"https:\/\/github.com\/atlas-engineer\/next\" title=\"A keyboard-focused web browser.\">Next<\/a> to see if it's ready for me). I like to support browsers other than <a href=\"https:\/\/www.google.com\/intl\/en\/chrome\/browser\/\" title=\"A WebKit-based browser, where each tab runs in its own thread.\">Chrome<\/a> both because increasingly it seems like Google doesn't have my best-interests in mind, and because I don't want to end up where we were in the 90s with one company controlling the web. I try to remain cognizant that Mozilla is also a for-profit company and also may not always have my best interests in mind.<\/p><p>For communication, I have consolidated to <a href=\"https:\/\/matrix.org\/blog\/home\/\" title=\"A decentralised group communication tool.\">Matrix<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Signal_%28software%29\" title=\"An encrypted messaging service.\">Signal<\/a>. I prefer Matrix, but it's a newer project, and Signal provides an experience more like SMS for the people in my life for whom technology is not an emphasis. Matrix is nice because it's a decentralized protocol which aims to have bridges into other messaging ecosystems. This means that I can use a client that best suits me, control where my data lives, and still engage with people who aren't using Matrix.<\/p><h3>What would be your dream setup?<\/h3><p>I believe in the philosophy of supporting protocols, not apps. Protocols allow us to create clients that work for the users, and create servers we can run ourselves. I also believe in Steve Job's philosophy that a computer should be a <a href=\"https:\/\/www.brainpickings.org\/2011\/12\/21\/steve-jobs-bicycle-for-the-mind-1990\/\" title=\"A Brain Pickings post about Steve Jobs and his bicycle for the mind quote.\">bicycle for the mind<\/a> \u2013 that it should help us to think. Today, computers have become appliances for most people: we don't understand how they work, but we expect them to. This part is great, and we should continue pushing into this space. Unfortunately, it seems as though not many people take advantage of their computers as bicycles.<\/p><p>My dream setup would be a computing appliance \u2013 one which, like TVs \u2013 most have in their home. This appliance would plug directly into the internet and then provide seamless services to the people that lived there: email, messaging, storage, a VPN, a desktop environment, etc. It could be expanded by purchasing additional units which would use something like <a href=\"https:\/\/en.wikipedia.org\/wiki\/PCI_Express\" title=\"The Wikipedia entry for PCI Express.\">PCIe<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Plan_9_from_Bell_Labs\" title=\"A distributed operating system.\">Plan 9<\/a> to expand its resources (maybe you have a gamer in the family so you buy one with a good GPU). Adding additional instances of the appliance would somehow provide fault tolerance in case one of the other instances went down.<\/p><p>People would interface with this appliance with peripherals (e.g. a watch) suitable to what they needed to do. They would not need to own what we colloquially consider a \"computer\", nor a screen. Instead they would use augmented reality glasses that displayed windows from a remote session running on the appliance (<a href=\"https:\/\/en.wikipedia.org\/wiki\/John_Gage\" title=\"John Gage's Wikipedia page.\">John Gage's<\/a> \"the network is the computer\" idiom). Let's throw some haptic gloves in there as a hand-wavey (literally) way of performing input.<\/p><p>And finally, all of this should have open protocols, and be controllable both through a simple, clear, interface, defined in a programming language which is also exposed. The interface keeps things in the \"appliance\" zone, and exposing the programming language it is written in nudges people to modify their environments to better suit their needs, as Emacs, spreadsheets, and HTML do. This is the philosophy of users being indistinguishable from programmers. This would allow people to begin using their computers more like bicycles and less like washing machines. Both are useful, but bicycles can take you places!<\/p>","created_at":"2020-03-04T23:54:11+0000","updated_at":"2020-03-04T23:54:11+0000","published_at":"2020-03-03T00:00:00+0000","published_by":["Daniel Bogan"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":10,"domain_name":"usesthis.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1155,"title":"Graph query languages: Cypher vs. Gremlin vs. nGQL","url":"https:\/\/news.ycombinator.com\/item?id=22482665","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\" style=\"border-spacing: 0px; width: 100%;\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D22482665\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"22482665\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"https:\/\/nebula-graph.io\/en\/posts\/graph-query-language-comparison-cypher-gremlin-ngql\/\" class=\"storylink\">Graph query languages: Cypher vs. Gremlin vs. nGQL<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=nebula-graph.io\">nebula-graph.io<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">50 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=jamie-vesoft\" class=\"hnuser\">jamie-vesoft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22482665\">4 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=22482665&amp;goto=item%3Fid%3D22482665\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Graph%20query%20languages%3A%20Cypher%20vs.%20Gremlin%20vs.%20nGQL&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Graph%20query%20languages%3A%20Cypher%20vs.%20Gremlin%20vs.%20nGQL\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=22482665&amp;auth=4209ae7945d891c59e92b53058a87a62e9b0c1bb\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=22482665\">14\u00a0comments<\/a><\/td>\n<\/tr><tr><td>\n<\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><p><a href=\"https:\/\/www.ycombinator.com\/apply\/\">Applications are open for YC Summer 2020<\/a><\/p>\n<br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"mailto:hn@ycombinator.com\">Support<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2020-03-04T14:56:03+0000","updated_at":"2020-03-04T14:56:03+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1154,"title":"home","url":"https:\/\/www.gqlstandards.org\/home","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-03-04T14:54:23+0000","updated_at":"2020-03-04T14:54:23+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.gqlstandards.org"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1153,"title":"Log Operators Cheat Sheet","url":"https:\/\/help.sumologic.com\/05Search\/Search-Cheat-Sheets\/Log-Operators-Cheat-Sheet","content":"<header>\n<\/header><section class=\"mt-content-container\"><p>The Log Operators cheat sheet provides a list of available parsers, aggregators, search operators, and mathematical expressions with links to full details for each item.\u00a0 For a step-by-step video and tutorial about creating queries, see the\u00a0<a title=\"Quick Start Labs and Tutorials\" href=\"https:\/\/help.sumologic.com\/01Start-Here\/Quick-Start-Tutorials\" rel=\"internal\">Quick Start Tutorial<\/a>.\u00a0 For a complete list of Sumo Logic Search\u00a0operators, you can download the <a title=\"https:\/\/drive.google.com\/a\/sumologic.com\/file\/d\/1YeqCNH_IINY6GqzgpT7cxSIY47NhyntH\/view?usp=sharing\" href=\"https:\/\/drive.google.com\/a\/sumologic.com\/file\/d\/1YeqCNH_IINY6GqzgpT7cxSIY47NhyntH\/view?usp=sharing\" rel=\"external nofollow\" target=\"_blank\" class=\"link-https\">PDF version<\/a>.\u00a0<\/p><p>The following tables provide a list of available Sumo\u00a0Logic parsers, aggregators, search operators, and mathematical expressions.\u00a0\u00a0<\/p><div class=\"mt-section\" id=\"section_1\"><h2 class=\"editable\">Parsing<\/h2><p>Sumo provides a number of ways to\u00a0<a title=\"Parsing\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\" rel=\"internal\">parse<\/a>\u00a0fields in your log messages.<\/p><table class=\"c6\"><colgroup><col \/><col \/><\/colgroup><tbody><tr><td class=\"mt-bgcolor-696969\"><strong>Operator<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 c4\"><strong>Description<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 c5\"><strong>Example<\/strong><\/td>\n<\/tr><tr><td>\n<p><a id=\"parse_anchor\" name=\"parse_anchor\" target=\"_blank\" title=\"Parse Predictable Patterns Using an Anchor\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/01-Parse-Predictable-Patterns-Using-an-Anchor\" rel=\"internal\">parse (anchor)<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>The parse operator, also called parse anchor, parses strings according to specified start and stop anchors, and then labels them as fields for use in subsequent aggregation functions in the query such as sorting, grouping, or other functions.<\/p>\n<\/td>\n<td class=\"c5\">\n<p><code>| parse \"User=*:\" as user<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"parse_regex\" name=\"parse_regex\" title=\"Parse Regex or Extract Operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/02-Parse-Variable-Patterns-Using-Regex\" rel=\"internal\">parse regex<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>The parse regex operator (also called the extract operator) enables users comfortable with regular expression syntax to extract more complex data from log lines. Parse regex can be used, for example, to extract nested fields.<\/p>\n<\/td>\n<td class=\"c5\">\n<p><code>| parse regex field=url \"[0-9A-Za-z-]+\\.(?&lt;domain&gt;[A-Za-z-]+\\.(?:co\\.uk|com|com\\.au))\/.*\"<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"keyvalue\" name=\"keyvalue\" title=\"Keyvalue operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/04-Parse-Keyvalue-Formatted-Logs\" rel=\"internal\">keyvalue<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>Typically, log files contain information that follow a key-value pair structure. The keyvalue operator allows you to get values from a log message by specifying the key paired with each value.<\/p>\n<\/td>\n<td class=\"c5\">\n<p><code>| keyvalue infer \"module\", \"thread\"<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"csv\" name=\"csv\" title=\"CSV Operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/05-Parse-CSV-Formatted-Logs\" rel=\"internal\">csv<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>The csv operator allows you to parse Comma Separated Values (CSV) formatted log entries. It uses a comma as the default delimiter.csv operator allows you to parse Comma Separated Values (CSV) formatted log entries. It uses a comma as the default delimiter.<\/p>\n<\/td>\n<td class=\"c5\">\n<p><code>|\u00a0<\/code><code>csv<\/code><code>_raw extract 1 as user, 2 as id, 3 as name<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"json\" name=\"json\" title=\"Parsing JSON Logs\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/03-Parse-JSON-Formatted-Logs\" rel=\"internal\">JSON<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>The JSON operator is a search query language operator that allows you to extract values from JSON input. Because JSON supports both nested keys and arrays that contain ordered sequences of values, the Sumo Logic JSON operator allows you to extract single top-level fields, multiple fields, nested keys, and keys in arrays.<\/p>\n<\/td>\n<td class=\"c5\">\n<p><code>| parse \"explainJsonPlan] *\" as jsonobject\u00a0<br \/>| json field=jsonobject \"sessionId\"<br \/>| json auto<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"split\" name=\"split\" title=\"split\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/06-Parse-Delimited-Logs-Using-Split\" rel=\"internal\">split<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>The split operator allows you to split strings into multiple strings, and parse delimited log entries, such as space-delimited formats.<\/p>\n<\/td>\n<td class=\"c5\">\n<p>Full query example:<\/p>\n<p><code>_sourceCategory=colon<br \/>| parse \"] * *\" as log_level, text<br \/>| split text delim=':' extract 1 as user, 2 as account_id, 3 as session_id, 4 as result<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"xml\" name=\"xml\" title=\"Parsing XML\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/07-Parse-XML-Formatted-Logs\" rel=\"internal\">xml<\/a><\/p>\n<\/td>\n<td class=\"c4\">\n<p>The XML operator uses a subset of the XPath 1.0 specification to provide a way for you to parse fields from XML documents. Using it, you can specify what to extract from an XML document using an XPath reference.<\/p>\n<\/td>\n<td class=\"c5\">\n<p><code>| parse\u00a0xml\u00a0\"\/<\/code><code>af<\/code><code>\/minimum\/@requested_bytes\"<\/code><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><\/div><div class=\"mt-section\" id=\"section_2\"><h2 class=\"editable\">Aggregating<\/h2><p><a title=\"Aggregating\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\" rel=\"internal\">Aggregating functions<\/a>\u00a0evaluate messages and place them into groups. The group operator is used in conjunction with group-by functions. When using any grouping function, the word by is sufficient for representing the group operator.<\/p><aside class=\"sumo-note\"><p>An aggregation function cannot take another function (such as\u00a0a math function). For example, you cannot\u00a0use:<\/p>\n<p><code>... | avg(x + y) as average<\/code><br \/>Instead use separate steps:<br \/><code>... | x + y as z | avg(z) as average<\/code><\/p>\n<\/aside><table class=\"mt-responsive-table\"><colgroup><col \/><col \/><\/colgroup><tbody><tr><td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-15\"><strong>Operator<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-25\"><strong>Description<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-20\"><strong>Restrictions<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-35\"><strong>Example<\/strong><\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"avg\" name=\"avg\" title=\"avg\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/avg\" rel=\"internal\">avg<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The averaging function (avg) calculates the average value of the numerical field being evaluated within the time range analyzed.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>| avg(request_received) by _timeslice<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"count\" name=\"count\" title=\"count, count_distinct, and count_frequent\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/count%2C-count-distinct%2C-and-count-frequent\" rel=\"internal\">count, count_distinct, and count_frequent<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Aggregating (group-by) functions are used in conjunction with the group operator and a field name. Only the word by is required to represent the group operator. The count function is also an operator in its own right and therefore can be used with or without the word by.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p><a title=\"count, count_distinct, and count_frequent\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/count%2C-count-distinct%2C-and-count-frequent\" rel=\"internal\">count_frequent<\/a>\u00a0can return up to 100 results when used in dashboard panels.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p>Example 1:<\/p>\n<p><code>| count by\u00a0url<\/code><br \/>\u00a0<\/p>\n<p>Example 2:<\/p>\n<p><code>| count_distinct(referrer) by status_code<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"fillmissing\" name=\"fillmissing\" title=\"fillmissing\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/fillmissing\" rel=\"internal\">fillmissing<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>When you run a standard\u00a0<a title=\"Group\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\" rel=\"internal\">group-by<\/a>\u00a0query, Sumo Logic only returns non-empty groups in the results. For example, if you are grouping by timeslice, then only the timeslices that have data are returned.<\/p>\n<p>This operator allows\u00a0you to specify groups to present in the\u00a0output, even if those groups have no data.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not supported in Live Dashboards or any continuous query.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>error<br \/>| count by _sourceCategory<br \/>| fillmissing values(\"backend\", \"database\", \"webapp\") in _sourceCategory<\/code>\ufeff<\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"first_and_last\" name=\"first_and_last\" title=\"first and last\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/first-and-last\" rel=\"internal\">first and last<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>First finds the earliest occurrence in search results, and last finds the result that follows all others, based on the sort order for the query.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not supported in Live Dashboards or any continuous query.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>| sort by _timeslice<br \/>| first(error_message)\u00a0by hostname<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"min_and_max\" name=\"min_and_max\" title=\"min and max\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/min-and-max\" rel=\"internal\">min and max<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Use the min and max functions to find the smallest or largest value in a set of values.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>| max(request_received)\u00a0by hour<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"most_recent_and_last_recent\" name=\"most_recent_and_last_recent\" title=\"most_recent and least_recent\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/most-recent-and-least-recent\" rel=\"internal\">most_recent and least_recent<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The most_recent and least_recent operators, used with the withtime operator, allow you to order data from newest to oldest.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>*ip* OR *address*<br \/>| parse regex \"(?&lt;IP&gt;\\b\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3})\"\u00a0<br \/>| lookup latitude, longitude, country_code from\u00a0geo:\/\/location\u00a0on ip=IP\u00a0<br \/>| where !isNull(country_code)\u00a0<br \/>| withtime IP\u00a0<br \/>| most_recent(ip_withtime) by country_code\u00a0<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"pct\" name=\"pct\" title=\"pct\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/percentile-(pct)\" rel=\"internal\">pct<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The percentile function (pct) finds the percentile of a given field. Multiple pct functions can be included in one query.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>| parse \"value=*\" as value<br \/>| pct(value, 95) as value_95pct<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"stddev\" name=\"stddev\" title=\"stddev\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/standard-deviation\" rel=\"internal\">stddev<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The standard deviation function (stddev) finds the standard deviation value for a distribution of numerical values within the time range analyzed and associated with a group designated by the \"group by\" field.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>... |\u00a0<\/code><code>stddev<\/code><code>(request_received) group by hour | sort by _stddev<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"sum\" name=\"sum\" title=\"sum\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/sum\" rel=\"internal\">sum<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Sum adds the values of the numerical field being evaluated within the time range analyzed.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-35\">\n<p><code>... | sum(bytes_received) group by hostname<\/code><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><\/div><div class=\"mt-section\" id=\"section_3\"><h2 class=\"editable\">Search Operators<\/h2><p>This section provides detailed\u00a0syntax, rules, and examples for Sumo Logic Operators, Expressions, and Search Language.<\/p><table class=\"mt-responsive-table\"><colgroup><col class=\"c7\" \/><col \/><\/colgroup><tbody><tr><td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-15\"><strong>Operator<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-25\"><strong>Description<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-20\"><strong>Restrictions<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 mt-noheading mt-column-width-40\"><strong>Example<\/strong><\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"accum\" name=\"accum\" title=\"accum\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/accum\" rel=\"internal\">accum<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The accum operator calculates the cumulative sum of a field. It can be used to find a count by a specific time interval, and can be used to find a total running count across all intervals.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourceCategory=IIS (Wyatt OR\u00a0Luke)<br \/>| parse \"[user=*]\" as cs_username<br \/>| timeslice by 1m<br \/>| count as requests by _timeslice,cs_username<br \/>| sort by _timeslice asc,cs_username<br \/>| accum requests as running_total<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"asn\" name=\"asn\" title=\"asn\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/ASN_Lookup\" rel=\"internal\">asn lookup<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Sumo Logic can lookup an Autonomous System Number (ASN)\u00a0and organization name by an IP\u00a0address. Any IP addresses that don't have an ASN will return null values.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<\/td><td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourceCategory=stream \"remote_ip=\"<br \/>| parse regex \"(?&lt;ip&gt;\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3})\"<br \/>| lookup organization, asn from asn:\/\/default on ip = ip<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"backshift\" name=\"backshift\" title=\"backshift\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/backshift\" rel=\"internal\">backshift<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The backshift operator compares values as they change over time. Backshift can be used with rollingstd, smooth, or any other operators whose results could be affected by spikes of data (where a spike could possibly throw off future results).<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourcecategory=katta\u00a0<br \/>| timeslice by 1m\u00a0<br \/>| count by _timeslice,_sourcehost\u00a0<br \/>| sort + _timeslice\u00a0<br \/>| backshift _count,1 by _sourcehost<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"base64decode\" name=\"base64decode\" title=\"base64Decode\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/base64Decode\" rel=\"internal\">base64Decode<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The base64Decode operator takes a base64 string and converts it to an ASCII string.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<\/td><td class=\"mt-noheading mt-column-width-40\">\n<p><code>| base64Decode(\"aHR0cDovL2NvZGVjLmFwYWNoZS5vcmcvY29tbW1vbnM=\") as V<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"base64encode\" name=\"base64encode\" title=\"base64Encode\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/base64Encode\" rel=\"internal\">base64Encode<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The base64Encode operator takes an ASCII string and converts it to a base64\u00a0string.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<\/td><td class=\"mt-noheading mt-column-width-40\">\n<p><code>| base64Encode(\"hello world\") as base64<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"bin\" name=\"bin\" title=\"bin\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/bin\" rel=\"internal\">bin<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Use the bin operator to sort results in a histogram.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourceCategory=analytics<br \/>| parse \"ms: *\" as time<br \/>| bin time width=10, min = 0, max = 500<br \/>| count by _bin, _bin_upper<br \/>| sort by _bin_upper<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"cidr\" name=\"cidr\" title=\"CIDR\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/CIDR\" rel=\"internal\">CIDR<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The CIDR operator allows you to leverage Classless Inter-Domain Routing (CIDS) notations to analyze IP network traffic in order to narrow analysis to specific subnets. CIDR notations specify the routing prefix of IP addresses.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>(denied OR\u00a0rejected AND _sourcecategory=firewall\u00a0<br \/>| parse \"ip=*,\" as ip_address<br \/>| where compareCIDRPrefix(\"10.10.1.32\", ip_address, toInt(27))\u00a0<br \/>| count by ip_address<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"concat\" name=\"concat\" title=\"concat\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/concat\" rel=\"internal\">concat<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The Concat operator allows you to concatenate or join multiple strings, numbers, and fields into a single user-defined field. It concatenates strings end-to-end and joins them into a new string that you define.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not\u00a0supported in\u00a0Dashboards.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>... | concat(octet1, \".\", octet2, \".\",octet3, \".\",octet4) as ip_address<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"contains\" name=\"contains\" title=\"contains\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/contains\" rel=\"internal\">contains<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The contains operator compares string values of two\u00a0<a title=\"Parse Operators\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\" rel=\"internal\">parsed<\/a>\u00a0fields and returns a boolean\u00a0result based on whether the second field's\u00a0value exists in the first.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<\/td><td class=\"mt-noheading mt-column-width-40\">\n<p><code>... | contains(\"hello world\", \"hello\") as containing<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"dectohex\" name=\"dectohex\" title=\"decToHex\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/decToHex\" rel=\"internal\">decToHex<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The\u00a0decToHex\u00a0operator converts a long value of 16 or fewer digits to a hexadecimal string using Two's Complement for negative values.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<\/td><td class=\"mt-noheading mt-column-width-40\">\n<p><code>... | decToHex(\"4919\") as V<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"diff\" name=\"diff\" title=\"diff\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/diff\" rel=\"internal\">diff<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The diff operator calculates the rate of change in a field between consecutive rows. To produce results, diff requires that a specified field contain numeric data; any non-numerical values are removed from the search results.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>* | parse \"bytes transmited: '*'\" as bytes<br \/>| timeslice 1m<br \/>| sum(bytes) as bytes by _timeslice<br \/>| sort _timeslice<br \/>| diff bytes as diff_bytes<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"fields\" name=\"fields\" title=\"fields\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/fields\" rel=\"internal\">fields<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The fields operator allows you to choose which fields are displayed in the results of a query. Use a fields operator to reduce the \"clutter\" of a search output that contains fields that aren't completely relevant to your query.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourceCategory=access_logs\u00a0<br \/>| parse \"[status=*]\" as status_code\u00a0<br \/>| fields method, status_code<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"filter\" name=\"filter\" title=\"filter operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/filter-operator\" rel=\"internal\">filter<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The filter operator can filter the output of a search using the results of a different search based on the filtering criteria of a subquery. The filter operator keeps only the records that match the filter criteria, allowing you to restrict search results to the most relevant information.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>The operator\u00a0can process up to 100,000 data points for a single query. It automatically drops the data points that exceed the limit and issues a warning.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourceCategory=HttpServers<br \/>| timeslice 1m<br \/>| count by _timeslice, _sourceHost<br \/>| filter _sourcehost in (outlier _count by _sourceHost | where _count_violation &gt; 0)<br \/>| transpose row _timeslice column _sourcehost<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"format\" name=\"format\" title=\"format\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/format\" rel=\"internal\">format<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The format operator allows you to format and combine data from fields in message logs\u2014including numbers, strings, and dates\u2014into a single user-defined string. This allows data in message logs, such as dates or currency amounts, to be formatted as human readable, when otherwise it would be hard to decipher.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>error\u00a0<br \/>| parse \"fiveMinuteRate=*,\" as rate\u00a0<br \/>| format(\"%s : %s\",\"Five Minute Rate is\" , rate) as formattedVal<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The formatDate operator allows you to format dates in log files as a string in the format you require, such as US date formatting, European formatting, timestamps, etc.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>* | formatDate(now(), \"YYYY-MM-dd\") as today<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"geo_lookup\" name=\"geo_lookup\" title=\"Geo Lookup\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/Geo-Lookup\" rel=\"internal\">geo lookup<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Sumo Logic can match a parsed IPv4 or IPv6 address to its geographical location on a map. To create the map the lookup operator matches parsed IP addresses to their physical location based on the latitude and longitude of where the addresses originated.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parse \"remote_ip=*]\" as remote_ip<br \/>| lookup latitude, longitude, country_code, country_name, region, city, postal_code\u00a0from\u00a0geo:\/\/location\u00a0on ip = remote_ip<br \/>| count by latitude, longitude, country_code, country_name, region, city, postal_code<br \/>| sort _count<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"haversine\" name=\"haversine\" title=\"haversine\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/haversine\" rel=\"internal\">haversine<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The haversine operator returns the distance between latitude and longitude values of two coordinates in kilometers. Coordinates need to be positive or negative values based on being north\/south or east\/west, instead of using the terms N\/S, E\/W.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| haversine(39.04380, -77.48790, 45.73723, -119.81143) as distanceKMs<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"hextodec\" name=\"hextodec\" title=\"hexToDec\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/hexToDec\" rel=\"internal\">hexToDec<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The hexToDec operator converts a hexadecimal string of 16 or fewer characters to long using Two's Complement for negative values.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| hexToDec(\"0000000000001337\") as V<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"if\" name=\"if\" title=\"if and ?\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/if-operator-and\" rel=\"internal\">if<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>There are two forms of ternary expression you can use in Sumo Logic queries: one is constructed using the IF operator, and the other uses the question mark (?) operator. These expressions are used to evaluate a condition as either true or false, with values assigned for each outcome. It is a shorthand way to express an if-else condition.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| if(status_code matches \"5*\", 1, 0) as server_error<br \/>Or<br \/>| status_code matches \"5*\" ? 1 : 0 as server_error<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"in\" name=\"in\" title=\"in operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/in-operator\" rel=\"internal\">in<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The In operator returns a Boolean value: true if the specified property is in the specified object, or false if it is not.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| if (status_code in (\"500\", \"501\", \"502\", \"503\", \"504\", \"505\", \"506\", \"401\", \"402\", \"403\", \"404\"), \"Error\", \"OK\") as status_code_type<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"ipv4tonumber\" name=\"ipv4tonumber\" title=\"ipv4ToNumber\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/ipv4ToNumber\" rel=\"internal\">ipv4ToNumber<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The ipv4ToNumber operator allows you to convert an Internet Protocol version 4 (IPv4) IP address from the octet dot-decimal format to a decimal format. This decimal format makes it easier to compare one IP address to another, rather than relying on IP masking.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>_sourceCategory=service remote_ip<br \/>| parse \"[remote_ip=*]\" as ip<br \/>| ipv4ToNumber(ip) as num<br \/>| fields ip, num<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"isblank\" name=\"isblank\" title=\"isNull, isEmpty, isBlank\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isNull%2C-isEmpty%2C-isBlank\" rel=\"internal\">isBlank<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isBlank operator checks to see that a string contains text. Specifically, it checks to see if a character sequence is whitespace, empty (\"\") ,or null. It takes a single parameter and returns a Boolean value: true if the variable is indeed blank, or false if the variable contains a value other than whitespace, empty, or null.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| where isBlank(user)<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"isempty\" name=\"isempty\" title=\"isNull, isEmpty, isBlank\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isNull%2C-isEmpty%2C-isBlank\" rel=\"internal\">isEmpty<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isEmpty operator checks to see that a string contains text. Specifically, it checks to see whether a character sequence is empty (\"\") or null. It takes a single parameter and return a Boolean value: true if the variable is indeed empty, or false if the variable contains a value other than empty or null.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| if(isEmpty(src_ip),1,0) as null_ip_counts<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"isnull\" name=\"isnull\" title=\"isNull, isEmpty, isBlank\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isNull%2C-isEmpty%2C-isBlank\" rel=\"internal\">isNull<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isNull operator takes a single parameter and returns a Boolean value: True if the variable is indeed null, or false if the variable contains a value other than null.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| where isNull(src_ip)<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"isnumeric\" name=\"isnumeric\" title=\"isNumeric\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isNumeric\" rel=\"internal\">isNumeric<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isNumeric operator checks whether a string is a valid Java number.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| isNumeric(num)<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"isprivateip\" name=\"isprivateip\" title=\"isPrivateIP\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isPrivateIP\" rel=\"internal\">isPrivateIP<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isPrivateIP operator checks if an IPv4 address is private and returns a boolean.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| isPrivateIP(hostip)<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"ispublicip\" name=\"ispublicip\" title=\"isPublicIP\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isPublicIP\" rel=\"internal\">isPublicIP<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isPublicIP operator checks if an IPv4 address is public and returns a boolean.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| isPublicIP(\"10.255.255.255\") as isPublic<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"isvalidip\" name=\"isvalidip\" title=\"isValidIP, isValidIPv4, isValidIPv6\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/isValidIP\" rel=\"internal\">isValidIP<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The isValidIP operator checks if the value is a valid IP address. The isValidIPv4 and isValidIPv6 operators check if the value is a valid IPv4 or IPv6 address respectively.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| isValidIP(\"10.255.255.255\") as isIP<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"join\" name=\"join\" title=\"join\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/join\" rel=\"internal\">join<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The join operator combines records of two or more data streams. Results are admitted on-the-fly to allow real time tables to be built. Values common to each table are then delivered as search results.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p>Full query example:\u00a0<\/p>\n<p><code>(\"starting stream from\" OR \"starting search\")<br \/>| join\u00a0<br \/>(parse \"starting stream from *\" AS a) AS T1,\u00a0<br \/>(parse \"starting search * from parent stream *\" AS b, c) AS T2\u00a0<br \/>on T1.a = T2.c<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"length\" name=\"length\" title=\"length\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/length\" rel=\"internal\">length<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The length operator returns the number of characters in a string. You can use it in where clauses or to create new fields. It returns 0 if the string is null.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| where length(query) &lt;= 20<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"limit\" name=\"limit\" title=\"limit\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/limit\" rel=\"internal\">limit<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The limit operator reduces the number of raw messages or aggregate results returned. If you simply query for a particular term, for example \"error\" without using an aggregation operator such as group by, limit will reduce the number of raw messages returned. If you first use group-by or other aggregation operator, the limit operator will reduce the number of grouped results instead.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| count by _sourceCategory<br \/>| sort by _count<br \/>| limit 5<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"logcompare\" name=\"logcompare\" title=\"LogCompare\" href=\"https:\/\/help.sumologic.com\/05Search\/LogCompare\" rel=\"internal\">logcompare<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The logcompare operator allows you to compare two sets of logs: baseline (historical) and target (current). To run a LogCompare operation, you can use the\u00a0<strong>LogCompare<\/strong>\u00a0button on the\u00a0<strong>Messages<\/strong>\u00a0tab to generate a properly formatted query.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not\u00a0supported in Dashboards.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| logcompare timeshift -24h<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"logreduce\" name=\"logreduce\" title=\"LogReduce\" href=\"https:\/\/help.sumologic.com\/05Search\/LogReduce\" rel=\"internal\">logreduce<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The LogReduce algorithm uses fuzzy logic to cluster messages together based on string and pattern similarity. Use the <strong>LogReduce<\/strong> button and operator to quickly assess activity patterns for things like a range of devices or traffic on a website.\u00a0(Formerly Summarize.)\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not\u00a0supported in Dashboards.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| logreduce<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"lookup\" name=\"lookup\" title=\"lookup\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/lookup\" rel=\"internal\">lookup<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Using a lookup operator, you can map data in your log messages to meaningful information. For example, you could use a lookup operator to map \"userID\" to a real user's name. Or, you could use a lookup operator to find black-listed IP addresses.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>|\u00a0parse \"name=*, phone number=*,\" as (name, phone)<br \/>| count by name, phone<br \/>\/\/We recommend doing a lookup after an aggregation<br \/>| lookup email from\u00a0https:\/\/compay.com\/userTable.csv\u00a0on name=userName, phone=cell<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"luhn\" name=\"luhn\" title=\"Luhn\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/Luhn\" rel=\"internal\">luhn (credit card validator)<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The Luhn operator uses Luhn\u2019s algorithm to check message logs for strings of numbers that may be credit card numbers, and then validates them. It takes a string as an input, strips out all characters that are not numerals, and checks if the resulting string is a valid credit card number, returning true or false accordingly.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parse regex \"(?&lt;maybecc&gt;\\d{4}-\\d{4}-\\d{4}-\\d{4})\" nodrop<br \/>| parse regex \"(?&lt;maybecc&gt;\\d{4}\\s\\d{4}\\s\\d{4}\\s\\d{4})\" nodrop<br \/>| parse regex \"(?&lt;maybecc&gt;\\d{16})\" nodrop<br \/>| if (luhn(maybecc), true, false) as valid<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"matches\" name=\"matches\" title=\"matches\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/matches\" rel=\"internal\">matches<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The matches operator can be used to match a string to a wildcard pattern or an RE2 compliant regex. The return of the operator is Boolean; the operator can be used with\u00a0<a title=\"where\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/where\" rel=\"internal\">where<\/a>\u00a0or\u00a0<a title=\"Search\/Search_Cheat_Sheets\/Search_Operators_Cheat_Sheet\" class=\"mt-self-link\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Cheat-Sheets\/Log-Operators-Cheat-Sheet\" rel=\"internal\">if<\/a>\u00a0expressions.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| if (agent matches \"*MSIE*\",\"Internet Explorer\",\"Other\") as Browser<br \/>| if (agent matches \"*Firefox*\",\"Firefox\",Browser) as Browser<\/code><\/p>\n<br \/>\u00a0<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"median\" name=\"median\" title=\"median\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/median\" rel=\"internal\">median<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>In order to calculate the\u00a0median\u00a0value for a particular field, you can utilize the Percentile (<a title=\"percentile (pct)\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/percentile-(pct)\" rel=\"internal\">pct<\/a>) operator with a percentile argument of 50.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parse \"value=*\" as value<br \/>| pct(value, 50) as median<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"merge\" name=\"merge\" title=\"Merge operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Transaction-Analytics\/Merge-Operator\" rel=\"internal\">merge<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The merge operator reduces a stream of events to a single event using a specified merge strategy. It is particularly useful as a subquery for the\u00a0<a title=\"Transactionize operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Transaction-Analytics\/Transactionize-operator\" rel=\"internal\">Transactionize<\/a>\u00a0operator.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parse \"BytesSentPersec = \\\"*\\\"\" as BytesPersec\u00a0<br \/>| merge BytesPersec join with \"--\", _messageTime takeLast<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"now\" name=\"now\" title=\"now\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/now\" rel=\"internal\">now<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The now operator returns the current epoch time in milliseconds. It can be used with the\u00a0<a title=\"formatDate\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/formatDate\" rel=\"internal\">formatDate<\/a>\u00a0operator to get the formatted current time.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but the <code>now()<\/code>\u00a0time presented in Live mode (the time the data is processed)\u00a0doesn't\u00a0match the search time, so\u00a0the results are\u00a0different.<\/p>\n<p>The results for search could be\u00a0hours or days later than the time presented in Live mode.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| now() as current_date<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"num\" name=\"num\" title=\"num\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/num\" rel=\"internal\">num<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The num operator converts a field to a number. Using Num in a query is useful for sorting results by number instead of alphabetically, which is the default. You can also use double as the operator, as an alias equivalent, if you prefer.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parse \"Execution duration: * s\" as duration<br \/>| num(duration)<br \/>| sort by duration<\/code><br \/>\u00a0<\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"outlier\" name=\"outlier\" title=\"outlier\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/outlier\" rel=\"internal\">outlier<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Given a series of time-stamped numerical values, using the outlier operator in a query can identify values in a sequence that seem unexpected, and would identify an alert or violation, for example, for a scheduled search.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p>Full query example:\u00a0<\/p>\n<p><code>_sourceCategory=IIS\/Access<br \/>| parse regex \"\\d+-\\d+-\\d+ \\d+:\\d+:\\d+ (?&lt;server_ip&gt;\\S+) (?&lt;method&gt;\\S+) (?&lt;cs_uri_stem&gt;\/\\S+?) \\S+ \\d+ (?&lt;user&gt;\\S+) (?&lt;client_ip&gt;[\\.\\d]+) \"<br \/>| parse regex \"\\d+ \\d+ \\d+ (?&lt;response_time&gt;\\d+)$\"<br \/>| timeslice 1m\u00a0<br \/>| max(response_time) as response_time by _timeslice<br \/>| outlier response_time window=5,threshold=3,consecutive=2,direction=+-<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"parsehex\" name=\"parsehex\" title=\"ParseHex Operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/01-Parse-Operators\/parseHex\" rel=\"internal\">parseHex<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The parseHex operator allows you to convert a hexadecimal string of 16 or fewer characters to a number.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<\/td><td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parseHex(\"12D230\") as decimalValue<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"predict\" name=\"predict\" title=\"predict\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/predict\" rel=\"internal\">predict<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The predict operator uses a series of time stamped numerical values to predict future values. For example, you could use this operator to take your current disk space capacity numbers, and predict when your system might run out of disk space.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p>Full query example:<\/p>\n<p><code>_sourceCategory=taskmanager<br \/>| jobState=InQueue<br \/>| timeslice 1m<br \/>| count by _timeslice<br \/>| toDouble(_count)<br \/>| predict _count by 1m forecast=5<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"replace\" name=\"replace\" title=\"replace\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/replace\" rel=\"internal\">replace<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The replace operator allows you to replace all instances of a specified string with another string. You can specify the string to replace with a matching regex or literal text. You might use it to find all instances of a name and change it to a new name or to replace punctuation in a field with different punctuation. This operator is useful anytime you need to rename something.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| replace(query, \".\",\"-&gt;\") as query<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"rollingstd\" name=\"rollingstd\" title=\"rollingstd\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/rollingstd\" rel=\"internal\">rollingstd<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The rollingstd (rolling standard) operator provides the rolling standard deviation of a field over a defined window. Rollingstd displays this value in a new column named _rollingstd.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| rollingstd _count,1 by _sourcehost<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"save\" name=\"save\" title=\"save\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/save\" rel=\"internal\">save<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Using the Save operator allows you to save the results of a query into the Sumo Logic file system. Later, you can use the lookup operator to access the saved data. The Save operator saves data in a simple format to a location you choose.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not\u00a0supported in Dashboards.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| save \/shared\/lookups\/daily_users<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"sessionize\" name=\"sessionize\" title=\"sessionize\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/sessionize\" rel=\"internal\">sessionize<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The sessionize operator allows you to use an extracted value from one log message (generated from one system) to find correlating values in log messages from other systems. After you run Sessionize, these related events are displayed on the same page. The thread of logs woven together is called a session.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not supported in Live Dashboards or any continuous query.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p>Full query example:<br \/><code>(SearchServiceImpl Creating Query) or (Stream SessionId using searchSessionId) or (Started search with sessionId)<br \/>| sessionize \"session: '*', streamSessionID: '*'\" as (serviceSessionId, streamSessionId),<br \/>\"Stream SessionId=$streamSessionId using searchSessionId=* and rawSessionId=*\" as (searchSessionId, rawSessionId),<br \/>\"Started search with sessionId: $searchSessionId, customerId: *, query: *\" as (customerId, query)<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"smooth\" name=\"smooth\" title=\"smooth\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/smooth\" rel=\"internal\">smooth<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The smooth operator calculates the rolling (or moving) average of a field, measuring the average of a value to \"smooth\" random variation. Smooth operator reveals trends in the data set you include in a query.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| smooth _count,1 by _sourcehost<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"sort\" name=\"sort\" title=\"sort\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/sort\" rel=\"internal\">sort<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The sort operator orders aggregated search results. The default sort order is descending. Then you can use the\u00a0<a title=\"top\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/top\" rel=\"internal\">top<\/a>\u00a0or\u00a0<a title=\"limit\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/limit\" rel=\"internal\">limit<\/a>\u00a0operators to reduce the number of sorted results returned.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| count as page_hits by _sourceHost<br \/>| sort by page_hits asc<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"substring\" name=\"dsubstringiff\" title=\"substring\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/substring\" rel=\"internal\">substring<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The substring operator allows you to specify an offset that will output only part of a string, referred to as a substring. You can use this operator to output just a part of a string instead of the whole string, for example, if you wanted to output an employee\u2019s initials instead of their whole name.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| substring(\"Hello world!\", 6)<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The timeslice operator segregates data by time period, so you can create bucketed results based on a fixed width in time, for example, five minute periods. Timeslice also supports bucketing by a fixed number of buckets across the search results, for example, 150 buckets over the last 60 minutes. An alias for the timeslice field is optional. When an alias is not provided, a default _timeslice field is created.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Timeslices greater than 1 day cannot be used in\u00a0Dashboard\u00a0Live mode.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| timeslice 1h<br \/>\/\/You can further aggregate your data by these time groupings<br \/>| count by _timeslice<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"tolowercase_and_touppercase\" name=\"tolowercase_and_touppercase\" title=\"toLowerCase and toUpperCase\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/toLowerCase-and-toUpperCase\" rel=\"internal\">toLowerCase and toUpperCase<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>As the name implies, the toLowerCase operator takes a string and converts it to all lower case letters. The toUpperCase operator takes a string and converts it to all upper case letters.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| toUpperCase(_sourceHost) as _sourceHost\u00a0<br \/>| where _sourceHost matches \"*NITE*\"<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"top\" name=\"top\" title=\"top\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/top\" rel=\"internal\">top<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>Use the top operator with the sort operator, to reduce the number of sorted results returned.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| top 5 _sourcecategory<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"total\" name=\"total\" title=\"total\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/total\" rel=\"internal\">total<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The total operator calculates the grand total of a field and injects that value into every row. It also supports grouping rows by a set of fields.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Can be used in Dashboard Panels, but in the search they must be included after the first <code>group-by<\/code>\u00a0phrase.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| total gbytes as total_memory<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"trace\" name=\"trace\" title=\"trace\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/trace\" rel=\"internal\">trace<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>A trace operator acts as a highly sophisticated filter to connect the dots across different log messages. You can use any identifying value with a trace operator (such as a user ID, IP address, session ID, etc.) to retrieve a comprehensive set of activity associated to that original ID.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Not supported in Live Dashboards or any continuous query.\u00a0<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| trace \"ID=( [0-9a-fA-F] {4} )\" \"7F92\"<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"transaction\" name=\"transaction\" title=\"Transaction Analytics\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Transaction-Analytics\" rel=\"internal\">transaction<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The transaction operator is used to analyze related sequences of logs. No matter what type of data you're analyzing, from tracking web site sign ups, to e-commerce data, to watching system activity across a distributed system, the transaction operator can be used in a variety of use cases.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-20\">\n<p>Tables generated with unordered data can be added to Dashboards, but Flow Diagrams cannot be added to Dashboards.<\/p>\n<p>Transaction by flow can't be used with Dashboards.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| transaction on sessionid fringe=10m\u00a0<br \/>with \"Starting session *\" as init,\u00a0<br \/>with \"Initiating countdown *\" as countdown_start,\u00a0<br \/>with \"Countdown reached *\" as countdown_done,\u00a0<br \/>with \"Launch *\" as launch\u00a0<br \/>results by transaction<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"transactionize\" name=\"transactionize\" title=\"Transactionize operator\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Transaction-Analytics\/Transactionize-operator\" rel=\"internal\">transactionize<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The transactionize operator groups logs that match on any fields you specify. Unlike other \"group by\" operators, where the logs in a group must match on all defined fields, transactionize just needs one field to match in order to assign logs to the same group.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| parse \"[system=001] [sessionId=*]\" as system1Id nodrop\u00a0<br \/>| parse \"[system=002][sessionId=*]\" as system2Id nodrop\u00a0<br \/>| parse \"[system=003][sessionId=*]\" as system3Id nodrop\u00a0<br \/>| parse \"system=001 with sessionId=*\" as system1Id nodrop\u00a0<br \/>| transactionize system1Id, system2Id, system3Id<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"transpose\" name=\"transpose\" title=\"transpose\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/transpose\" rel=\"internal\">transpose<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The transpose operator dynamically creates columns for aggregate search results. The dynamic functionality allows for changing the output of a query, turning search results into fields. It also means that queries can be designed without first knowing the output schema.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p>Full query example:<\/p>\n<p><code>_sourceCategory=service\u00a0<br \/>| parse \"Successful login for user '*', organization: '*'\" as user, org_id\u00a0<br \/>| timeslice 1d\u00a0<br \/>| count _timeslice, user\u00a0<br \/>|\u00a0transpose row _timeslice column user<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"urldecode\" name=\"urldecode\" title=\"urldecode\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/urldecode\" rel=\"internal\">urldecode<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The urldecode operator decodes a URL you include in a query, returning the decoded (unescaped) URL string.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| urldecode(url) as decoded<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"urlencode\" name=\"urlencode\" title=\"urlencode\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/urlencode\" rel=\"internal\">urlencode<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>The urlencode operator encodes the URL into an ASCII character set.<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>| urlencode(url) as encoded<\/code><\/p>\n<\/td>\n<\/tr><tr><td class=\"mt-noheading mt-column-width-15\">\n<p><a id=\"where\" name=\"where\" title=\"where\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Search-Operators\/where\" rel=\"internal\">where<\/a><\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-25\">\n<p>To filter results in a search query, use \"where\" as a conditional operator. The where operator must appear as a separate operator distinct from other operators, delimited by the pipe symbol (\"|\"). In other words, the following construct will not work and will generate a syntax error:<\/p>\n<\/td>\n<td class=\"mt-noheading mt-column-width-40\">\n<p><code>\/\/We recommend placing inclusive filters before exclusive filters in query strings<br \/>| where status_code matches \"4*\"<br \/>| where !(status_code matches \"2*\")<\/code><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><\/div><div class=\"mt-section\" id=\"section_4\"><h2 class=\"editable\">Math Expressions<\/h2><p>You can use general mathematical expressions on numerical data extracted from log lines. For any mathematical or group-by function that implicitly requires integers, Sumo Logic casts the string data to a number for you.<\/p><table class=\"c9\"><colgroup><col \/><col \/><\/colgroup><tbody><tr><td class=\"mt-bgcolor-696969\"><strong>Operator<\/strong><\/td>\n<td class=\"mt-bgcolor-696969\"><strong>Description<\/strong><\/td>\n<td class=\"mt-bgcolor-696969 c8\"><strong>Example<\/strong><\/td>\n<\/tr><tr><td>\n<p><strong>Basic<\/strong><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"abs\" name=\"abs\" title=\"abs\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/abs\" rel=\"internal\">abs<\/a><\/p>\n<\/td>\n<td>The absolute function calculates the absolute value of x.<\/td>\n<td class=\"c8\">\n<p><code>| abs(-1.5) as v<br \/>\/\/ v = 1.5<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"round\" name=\"round\" title=\"round\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/round\" rel=\"internal\">round<\/a><\/p>\n<\/td>\n<td>The round function returns the closest integer to x.<\/td>\n<td class=\"c8\">\n<p><code>| round((bytes\/1024)\/1024) as MB<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"ceil\" name=\"ceil\" title=\"ceil\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/ceil\" rel=\"internal\">ceil<\/a><\/p>\n<\/td>\n<td>The ceiling function rounds up to the smallest integer value. Returns the smallest integral value that is not less than x.<\/td>\n<td class=\"c8\">\n<p><code>| ceil(1.5) as v<br \/>\/\/ v = 2<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<\/td>\n<td>The floor function rounds down to the largest previous integer value. Returns the largest integer not greater than x.<\/td>\n<td class=\"c8\">\n<p><code>| floor(1.5) as v<br \/>\/\/ v = 1<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"max\" name=\"max\" title=\"min and max\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/min-and-max\" rel=\"internal\">max<\/a><\/p>\n<\/td>\n<td>The maximum function returns the larger of two values.<\/td>\n<td class=\"c8\">\n<p><code>| max(1, 2) as v<br \/>\/\/ v = 2<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"min\" name=\"min\" title=\"min and max\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/aaGroup\/min-and-max\" rel=\"internal\">min<\/a><\/p>\n<\/td>\n<td>The minimum function returns the smaller of two values.<\/td>\n<td class=\"c8\">\n<p><code>| min(1, 2) as v<br \/>\/\/ v = 1<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"sqrt\" name=\"sqrt\" title=\"sqrt\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/sqrt\" rel=\"internal\">sqrt<\/a><\/p>\n<\/td>\n<td>The square root function returns the square root value of x.<\/td>\n<td class=\"c8\">\n<p><code>| sqrt(4) as v<br \/>\/\/ v = 2<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"cbrt\" name=\"cbrt\" title=\"cbrt\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/cbrt\" rel=\"internal\">cbrt<\/a><\/p>\n<\/td>\n<td>The cube root function returns the cube root value of x.<\/td>\n<td class=\"c8\">\n<p><code>| cbrt(8) as v<br \/>\/\/ v = 2<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><strong>Exponents and Logs<\/strong><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"exp\" name=\"exp\" title=\"exp\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/exp\" rel=\"internal\">exp<\/a><\/p>\n<\/td>\n<td>The exponent function returns Euler's number e raised to the power of x.<\/td>\n<td class=\"c8\">\n<p><code>| exp(1) as v<br \/>\/\/ v = 2.7182818284590455<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"expml\" name=\"expml\" title=\"expm1\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/expm1\" rel=\"internal\">expm1<\/a><\/p>\n<\/td>\n<td>The expm1 function returns value of x in exp(x)-1, compensating for the roundoff in exp(x).<\/td>\n<td class=\"c8\">\n<p><code>| expm1(0.1) as v<br \/>\/\/ v = 0.10517091807564763<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"log\" name=\"log\" title=\"log\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/log\" rel=\"internal\">log<\/a><\/p>\n<\/td>\n<td>The logarithm function returns the natural logarithm of x.<\/td>\n<td class=\"c8\">\n<p><code>| log(2) as v<br \/>\/\/ v = 0.6931471805599453<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"log10\" name=\"log10\" title=\"log10\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/log10\" rel=\"internal\">log10<\/a><\/p>\n<\/td>\n<td>The log10 function returns the base 10 logarithm of x.<\/td>\n<td class=\"c8\">\n<p><code>| log10(2) as v<br \/>\/\/ v = 0.3010299956639812<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"log1p\" name=\"log1p\" title=\"log1p\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/log1p\" rel=\"internal\">log1p<\/a><\/p>\n<\/td>\n<td>The log1p function computes log(1+x) accurately for small values of x.<\/td>\n<td class=\"c8\">\n<p><code>| log1p(0.1) as v<br \/>\/\/ v = 0.09531017980432487<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><strong>Trigonometric<\/strong><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"sin\" name=\"sin\" title=\"sin\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/sin\" rel=\"internal\">sin<\/a><\/p>\n<\/td>\n<td>Sine of argument in radians.<\/td>\n<td class=\"c8\">\n<p><code>| sin(1) as v<br \/>\/\/ v = 0.8414709848078965<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"cos\" name=\"cos\" title=\"cos\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/cos\" rel=\"internal\">cos<\/a><\/p>\n<\/td>\n<td>Cosine of argument in radians.<\/td>\n<td class=\"c8\">\n<p><code>| cos(1) as v<br \/>\/\/ v = 0.5403023058681398<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"tan\" name=\"tan\" title=\"tan\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/tan\" rel=\"internal\">tan<\/a><\/p>\n<\/td>\n<td>Tangent of argument in radians.<\/td>\n<td class=\"c8\">\n<p><code>| an(1) as v<br \/>\/\/ v = 1.5574077246549023<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"asin\" name=\"asin\" title=\"asin\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/asin\" rel=\"internal\">asin<\/a><\/p>\n<\/td>\n<td>Inverse sine; result is in radians.<\/td>\n<td class=\"c8\">\n<p><code>| asin(1) as v<br \/>\/\/ v = 1.5707963267948966<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"acos\" name=\"acos\" title=\"acos\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/acos\" rel=\"internal\">acos<\/a><\/p>\n<\/td>\n<td>Inverse cosine; result is in radians.<\/td>\n<td class=\"c8\">\n<p><code>| acos(x)\\<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"atan\" name=\"atan\" title=\"atan\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/atan\" rel=\"internal\">atan<\/a><\/p>\n<\/td>\n<td>Inverse tangent; result is in radians.<\/td>\n<td class=\"c8\">\n<p><code>| atan(x)<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"atan2\" name=\"atan2\" title=\"atan2\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/atan2\" rel=\"internal\">atan2<\/a><\/p>\n<\/td>\n<td>Four-quadrant inverse tangent.<\/td>\n<td class=\"c8\">\n<p><code>| atan2(0, -1) as v<br \/>\/\/ v = pi<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"sinh\" name=\"sinh\" title=\"sinh\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/sinh\" rel=\"internal\">sinh<\/a><\/p>\n<\/td>\n<td>Hyperbolic sine of argument in radians.<\/td>\n<td class=\"c8\">\n<p><code>| sinh(x)<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"cosh\" name=\"cosh\" title=\"cosh\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/cosh\" rel=\"internal\">cosh<\/a><\/p>\n<\/td>\n<td>Hyperbolic cosine of argument in radians.<\/td>\n<td class=\"c8\">\n<p><code>| cosh(x)<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"tanh\" name=\"tanh\" title=\"tanh\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/tanh\" rel=\"internal\">tanh<\/a><\/p>\n<\/td>\n<td>Hyperbolic tangent of argument in radians.<\/td>\n<td class=\"c8\">\n<p><code>| tanh(x)<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><strong>Advanced<\/strong><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"hypot\" name=\"hypot\" title=\"hypot\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/hypot\" rel=\"internal\">hypot<\/a><\/p>\n<\/td>\n<td>Returns the square root of the sum of an array of squares.<\/td>\n<td class=\"c8\">\n<p><code>| hypot(1, 0) as v<br \/>\/\/ v = 1<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"todegrees\" name=\"todegrees\" title=\"toDegrees\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/toDegrees\" rel=\"internal\">toDegrees<\/a><\/p>\n<\/td>\n<td>Converts angles from radians to degrees.<\/td>\n<td class=\"c8\">\n<p><code>| toDegrees(asin(1)) as v<br \/>\/\/ v = 90<\/code><\/p>\n<\/td>\n<\/tr><tr><td>\n<p><a id=\"toradians\" name=\"toradians\" title=\"toRadians\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\/Math-Expressions\/toRadians\" rel=\"internal\">toRadians<\/a><\/p>\n<\/td>\n<td>Converts angles from degrees to radians.<\/td>\n<td class=\"c8\">\n<p><code>| toRadians(180) as v<br \/>\/\/ v = pi<\/code><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><\/div><\/section><footer class=\"elm-content-footer\"><ol class=\"elm-meta-data elm-meta-article-navigation\"><li class=\"elm-back-to-top\"><a class=\"mt-icon-back-to-top\" href=\"#title\" id=\"mt-back-to-top\" title=\"Jump back to top of this article\">Back to top<\/a><\/li>\n<li class=\"elm-article-pagination\">\n<ul class=\"mt-article-pagination\" data-ga-action=\"Article pagination bottom\"><li class=\"mt-pagination-previous\"><a class=\"mt-icon-previous-article\" data-ga-label=\"Previous page\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Cheat-Sheets\/IIS-Search-Examples-Cheat-Sheet\" title=\"IIS Search Examples Cheat Sheet\">IIS Search Examples Cheat Sheet<\/a><\/li>\n<li class=\"mt-pagination-next\"><a class=\"mt-icon-next-article\" data-ga-label=\"Next page\" href=\"https:\/\/help.sumologic.com\/05Search\/Search-Query-Language\" title=\"Search Query Language\">Search Query Language<\/a><\/li>\n<\/ul><\/li>\n<\/ol><div class=\"elm-related-articles-container\"><ol class=\"mt-related-listings-container\" data-ga-action=\"related\"><li class=\"mt-related-empty-listing\">There are no recommended articles.<\/li>\n<\/ol><\/div>\n<\/footer>","created_at":"2020-03-04T14:21:18+0000","updated_at":"2020-03-04T14:21:18+0000","published_at":"2015-12-07T18:13:24+0000","published_by":["Sumo Logic"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":24,"domain_name":"help.sumologic.com","preview_picture":"https:\/\/a.mtstatic.com\/@public\/production\/site_7026\/1538173143-logo.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1152,"title":"refactoringui\/heroicons","url":"https:\/\/github.com\/refactoringui\/heroicons","content":"<p>A set of free MIT-licensed high-quality SVG icons for you to use in your web projects.<\/p><p>140 icons per style currently, more to come in the future.<\/p><h3><a id=\"user-content-outline-style\" class=\"anchor\" aria-hidden=\"true\" href=\"#outline-style\"><\/a>Outline style<\/h3><p>24x24 icons drawn with a stroke.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/refactoringui\/heroicons\/blob\/master\/.github\/outline-md-preview.svg\"><img src=\"https:\/\/github.com\/refactoringui\/heroicons\/raw\/master\/.github\/outline-md-preview.svg?sanitize=true\" alt=\"\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-solid-style\" class=\"anchor\" aria-hidden=\"true\" href=\"#solid-style\"><\/a>Solid style<\/h3><p>Smaller 20x20 icons drawn with fills.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/refactoringui\/heroicons\/blob\/master\/.github\/solid-sm-preview.svg\"><img src=\"https:\/\/github.com\/refactoringui\/heroicons\/raw\/master\/.github\/solid-sm-preview.svg?sanitize=true\" alt=\"\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h3><p>To use these icons, simply copy the source for the icon you need and inline it directly into your HTML:<\/p><div class=\"highlight highlight-text-html-basic\"><pre>&lt;svg class=\"h-6 w-6 text-gray-500\" fill=\"none\" viewBox=\"0 0 24 24\" stroke=\"currentColor\"&gt;\n  &lt;path stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M12 8v4l3 3m6-3a9 9 0 11-18 0 9 9 0 0118 0z\"\/&gt;\n&lt;\/svg&gt;<\/pre><\/div><p>Both icon styles are preconfigured to be stylable by setting the <code>color<\/code> CSS property, either manually or using utility classes like <code>text-gray-500<\/code> in a framework like <a href=\"https:\/\/tailwindcss.com\" rel=\"nofollow\">Tailwind CSS<\/a>.<\/p>","created_at":"2020-03-02T03:32:28+0000","updated_at":"2020-03-02T03:32:28+0000","published_by":["refactoringui"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/repository-images.githubusercontent.com\/242754476\/f15c9780-56e8-11ea-8243-121e527e48f1","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1151,"title":"Kubernetes for Full-Stack Developers, a self-guided course.","url":"https:\/\/www.digitalocean.com\/community\/curriculums\/kubernetes-for-full-stack-developers","content":"<p>Whether you\u2019re just curious, getting started with Kubernetes, or have experience with it, this curriculum will help you learn more about Kubernetes and running containerized applications. You\u2019ll learn about core Kubernetes concepts and use them to deploy and scale applications in practical tutorials. By the end of this curriculum you\u2019ll be able to create your own Kubernetes cluster from scratch and run your own applications on it. You will also learn how to set up monitoring, alerting, and automation for your applications on Kubernetes.<\/p><p><strong>Download the entire curriculum as an eBook!<\/strong><\/p><p><a href=\"https:\/\/do.co\/k8s-book-epub\"><em>Kubernetes for Full-Stack Developers<\/em> eBook in <strong>EPUB format<\/strong><\/a><\/p><p><a href=\"https:\/\/do.co\/k8s-book-pdf\"><em>Kubernetes for Full-Stack Developers<\/em> eBook in <strong>PDF format<\/strong><\/a><br \/><\/p>","created_at":"2020-03-02T03:25:02+0000","updated_at":"2020-03-02T03:25:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.digitalocean.com","http_status":"200","origin_url":"https:\/\/www.digitalocean.com\/community\/curriculums\/kubernetes-for-full-stack-developers?utm_source=changelog&utm_medium=newsletter&utm_campaign=changelog-news"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1150,"title":"\ud83d\udcd5 The 25 most recommended programming books of all-time.","url":"https:\/\/www.daolf.com\/posts\/best-programming-books\/","content":"<p>This article is a follow up of the one I did about the <a href=\"https:\/\/www.daolf.com\/posts\/best-startup-books\/\">the most recommended startup books of all-time<\/a>.<\/p><p>If you\u2019ve read this one recently. I guess you can jump straight to the <a href=\"#25-most-recommended-programming-books-of-all-time\">results<\/a>.<\/p><p>There are countless lists on the internet claiming to be <strong>the<\/strong> list of must-read programming books and it seemed that all those lists always recommended that same books minus two or three odd choices.<\/p><p>Finding good ressources for learning programming is always tricky. Every-one has its own opinion about what book is the best to learn, and as we say in french, \u201cColor and tastes should not be argued about\u201d.<\/p><p>However I though it would be interesting to trust the wisdom of the crown and to find the books that appeared the most in those \u201cBest Programming Book\u201d lists.<\/p><p>If you want to jump right on the results go take a look below at the <a href=\"#25-most-recommended-programming-books-of-all-time\">full results<\/a>. If you want to learn about the methodology, bear with me.<\/p><p><em>Disclaimer: I spent countless hours on this article so I\u2019ve decided to put Amazon affiliation links to see if those kinds of detailed articles could be a viable source of revenue, \u2026 or not \ud83e\udd37\u200d\u2642\ufe0f.<\/em><\/p><h2 id=\"methodology\">Methodology:<\/h2><p>I\u2019ve simply asked Google for a few queries like \u201cBest Programming Books\u201d and its variations of. I have then scrapped all those pages (using ScrapingBee, a web scraping API I\u2019m working on).<\/p><p>I\u2019ve deduplicated the links and ended up with nearly 150 links. Using the title of the pages I was also able to quickly discards:<\/p><ul><li>list focussed on one particular technology or platform<\/li>\n<li>list focussed on one particular year<\/li>\n<li>list focussed on free books<\/li>\n<li>Quora and Reddit threads<\/li>\n<\/ul><p>I ended up with almost 200 HTML files. I went on opening all the files on my browser, open my chrome inspector, found and wrote the CSS selector matching book titles in the article. This took me around 1hours, almost 30 seconds per page.<\/p><p>This also allowed me to discard even more nonrelevant pages, and I discarded a lot. In the end I compiled around 70 lists into this one.<\/p><p>At this moment I had this big JSON file referencing the HTML page previously scrapped, and a CSS selector.<\/p><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/rules.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><p>Using Python with Beautiful soup, I\u2019ve extracted every text inside DOM elements that matched the CSS selector. I ended up with a huge list of books, not usable without some post-processing.<\/p><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/links.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><p>To find the most quoted startup books I needed to normalize my results.<\/p><p>I had to play with all the different variation like \u201c{title} by {author}\u201d or \u201c{title} - {author}\u201d.<\/p><p>Or \u201c{title}:{subtitle}\u201d and \u201c{title}\u201d, or even all the one containing edition number.<\/p><p>I ended up doing it using this simple custom Python function:<\/p><div class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">def clean_link(link):\n    link = link.encode().decode('ascii', errors='ignore')\n    link = link.replace(\"'\", '')\n    link = link.lower()\n    link = ' '.join([w for w in link.split(' ') if w not in ['the', 'a']])\n    link = link.split('by')[0]\n    link = link.split(':')[0]\n    link = link.split('(')[0]\n    link = ' '.join(link.split())\n    link = link.replace('-', '_')\n    link = ''.join([c for c in link if c.isalpha() or c == '_' or c == ' '])\n    link = link.strip()\n    link = link.replace(' ', '_')\n    link = ''.join([c for c in link if c.isalpha() or c == '_'])\n    return link<\/pre><\/div><p>and quite a bit of manual cleaning.<\/p><p>My list now looked like this:<\/p><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/clean_links.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><p>From there it was easy to compute the most recommended books. You can find all the data used to process this list on this <a href=\"https:\/\/github.com\/daolf\/Most-recommended-programming-books\">repo<\/a>. Now let\u2019s take a look at the list:<\/p><h3 id=\"25-continuous-delivery-https-amzn-to-2oz9jxs-by-jez-humble-david-farley-8-8-recommended\">25. <a href=\"https:\/\/amzn.to\/2OZ9JXS\">Continuous Delivery<\/a> by Jez Humble &amp; David Farley (8.8% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/25.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><p>\u201cGetting software released to users is often a painful, risky, and time-consuming process. This groundbreaking new book sets out the principles and technical practices that enable rapid, incremental delivery of high quality, valuable new functionality to users. Through automation of the build, deployment, and testing process, and improved collaboration between developers, testers, and operations, delivery teams can get changes released in a matter of hours\u2014 sometimes even minutes\u2013no matter what the size of a project or the complexity of its code base.<\/p><p>Jez Humble and David Farley begin by presenting the foundations of a rapid, reliable, low-risk delivery process. Next, they introduce the \u201cdeployment pipeline,\u201d an automated process for managing all changes, from check-in to release. Finally, they discuss the \u201cecosystem\u201d needed to support continuous delivery, from infrastructure, data and configuration management to governance.\u201d <a href=\"https:\/\/amzn.to\/2OZ9JXS\">Amazon.com<\/a><\/p><h3 id=\"24-algorithms-https-amzn-to-2u2vpug-by-robert-sedgewick-kevin-wayne-8-8-recommended\">24. <a href=\"https:\/\/amzn.to\/2u2vpuG\">Algorithms<\/a> by Robert Sedgewick &amp; Kevin Wayne (8.8% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/24.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cThe algorithms in this book represent a body of knowledge developed over the last 50 years that has become indispensable, not just for professional programmers and computer science students but for any student with interests in science, mathematics, and engineering, not to mention students who use computation in the liberal arts.\u201d<a href=\"https:\/\/amzn.to\/2u2vpuG\">Amazon.com<\/a><\/p><h3 id=\"23-the-self-taught-programmer-https-amzn-to-325cn9t-by-cory-althoff-8-8-recommended\">23. <a href=\"https:\/\/amzn.to\/325cN9T\">The Self-Taught Programmer<\/a> by Cory Althoff (8.8% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/23.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cI am a self-taught programmer. After a year of self-study, I learned to program well enough to land a job as a software engineer II at eBay.<\/p><p>Once I got there, I realized I was severely under-prepared. I was overwhelmed by the amount of things I needed to know but hadn\u2019t learned yet. My journey learning to program, and my experience at my first job as a software engineer were the inspiration for this book.<\/p><p>This book is not just about learning to program; although you will learn to code. If you want to program professionally, it is not enough to learn to code; that is why, in addition to helping you learn to program, I also cover the rest of the things you need to know to program professionally that classes and books don\u2019t teach you.<\/p><p>\u201cThe Self-taught Programmer\u201d is a roadmap, a guide to take you from writing your first Python program, to passing your first technical interview. The path is there. Will you take it?\u201d<a href=\"https:\/\/amzn.to\/325cN9T\">Amazon.com<\/a><\/p><h3 id=\"22-rapid-development-https-amzn-to-2sqlqtn-by-steve-mcconnell-8-8-recommended\">22. <a href=\"https:\/\/amzn.to\/2SQLQTn\">Rapid Development<\/a> by Steve McConnell (8.8% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/22.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cCorporate and commercial software-development teams all want solutions for one important problem\u2014how to get their high-pressure development schedules under control. In RAPID DEVELOPMENT, author Steve McConnell addresses that concern head-on with overall strategies, specific best practices, and valuable tips that help shrink and control development schedules and keep projects moving. Inside, you\u2019ll find: - A rapid-development strategy that can be applied to any project and the best practices to make that strategy work - Candid discussions of great and not-so-great rapid-development practices\u2014estimation, prototyping, forced overtime, motivation, teamwork, rapid-development languages, risk management, and many others - A list of classic mistakes to avoid for rapid-development projects, including creeping requirements, shortchanged quality, and silver-bullet syndrome - Case studies that vividly illustrate what can go wrong, what can go right, and how to tell which direction your project is going - RAPID DEVELOPMENT is the real-world guide to more efficient applications development.\u201d<a href=\"https:\/\/amzn.to\/2SQLQTn\">Amazon.com<\/a><\/p><h3 id=\"21-coders-at-work-https-amzn-to-323c5ki-by-peter-seibel-10-2-recommended\">21. <a href=\"https:\/\/amzn.to\/323c5Ki\">Coders at Work<\/a> by Peter Seibel (10.2% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/21.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cThis is a who\u2019s who in the programming world - a fascinating look at how some of the best in the world do their work. Patterned after the best selling Founders at Work, the book represents two years of interviews with some of the top programmers of our times.\u201d<a href=\"https:\/\/amzn.to\/323c5Ki\">Amazon.com<\/a><\/p><h3 id=\"20-domain-driven-design-https-amzn-to-38bfcxd-by-eric-evans-10-2-recommended\">20. <a href=\"https:\/\/amzn.to\/38BFCxd\">Domain-Driven Design<\/a> by Eric Evans (10.2% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/20.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cLeading software designers have recognized domain modeling and design as critical topics for at least twenty years, yet surprisingly little has been written about what needs to be done or how to do it. Although it has never been clearly formulated, a philosophy has developed as an undercurrent in the object community, which I call \u201cdomain-driven design\u201d.<\/p><p>I have spent the past decade focused on developing complex systems in several business and technical domains. I\u2019ve tried best practices in design and development process as they have emerged from the leaders in the object-oriented development community. Some of my projects were very successful; a few failed. A feature common to the successes was a rich domain model that evolved through iterations of design and became part of the fabric of the project.<\/p><p>This book provides a framework for making design decisions and a technical vocabulary for discussing domain design. It is a synthesis of widely accepted best practices along with my own insights and experiences. Projects facing complex domains can use this framework to approach domain-driven design systematically.\u201d<a href=\"https:\/\/amzn.to\/38BFCxd\">Amazon.com<\/a><\/p><h3 id=\"19-the-art-of-computer-programming-https-amzn-to-2su7n3q-by-donald-e-knuth-10-2-recommended\">19. <a href=\"https:\/\/amzn.to\/2SU7N3Q\">The Art of Computer Programming<\/a> by Donald E. Knuth(10.2% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/19.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cCountless readers have spoken about the profound personal influence of Knuth\u2019s work. Scientists have marveled at the beauty and elegance of his analysis, while ordinary programmers have successfully applied his \u201ccookbook\u201d solutions to their day-to-day problems. All have admired Knuth for the breadth, clarity, accuracy, and good humor found in his books.\u201d<a href=\"https:\/\/amzn.to\/2SU7N3Q\">Amazon.com<\/a><\/p><h3 id=\"18-structure-and-interpretation-of-computer-programs-https-amzn-to-2hj7hqr-by-harold-abelson-gerald-jay-sussman-julie-sussman-13-2-recommended\">18. <a href=\"https:\/\/amzn.to\/2HJ7HqR\">Structure and Interpretation of Computer Programs<\/a> by Harold Abelson \/ Gerald Jay Sussman \/ Julie Sussman (13.2% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/17.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cStructure and Interpretation of Computer Programs has had a dramatic impact on computer science curricula over the past decade. This long-awaited revision contains changes throughout the text. There are new implementations of most of the major programming systems in the book, including the interpreters and compilers, and the authors have incorporated many small changes that reflect their experience teaching the course at MIT since the first edition was published. A new theme has been introduced that emphasizes the central role played by different approaches to dealing with time in computational models: objects with state, concurrent programming, functional programming and lazy evaluation, and nondeterministic programming. There are new example sections on higher-order procedures in graphics and on applications of stream processing in numerical programming, and many new exercises. In addition, all the programs have been reworked to run in any Scheme implementation that adheres to the IEEE standard.\u201d<a href=\"https:\/\/amzn.to\/2HJ7HqR\">Amazon.com<\/a><\/p><h3 id=\"17-patterns-of-enterprise-application-architecture-https-amzn-to-37xeadr-by-martin-fowler-14-7-recommended\">17. <a href=\"https:\/\/amzn.to\/37xEadR\">Patterns of Enterprise Application Architecture<\/a> by Martin Fowler (14.7% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/16.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cThe practice of enterprise application development has benefited from the emergence of many new enabling technologies. Multi-tiered object-oriented platforms, such as Java and .NET, have become commonplace. These new tools and technologies are capable of building powerful applications, but they are not easily implemented. Common failures in enterprise applications often occur because their developers do not understand the architectural lessons that experienced object developers have learned.\u201d<a href=\"https:\/\/amzn.to\/37xEadR\">Amazon.com<\/a><\/p><h3 id=\"16-programming-pearls-https-amzn-to-2u8coxo-by-jon-bentley-16-1-recommended\">16. <a href=\"https:\/\/amzn.to\/2u8cOxo\">Programming Pearls<\/a> by Jon Bentley (16.1% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/16-bis.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cComputer programming has many faces. Fred Brooks paints the big picture in The Mythical Man Month; his essays underscore the crucial role of management in large software projects. At a finer grain, Steve McConnell teaches good programming style in Code Complete. The topics in those books are the key to good software and the hallmark of the professional programmer. Unfortunately, though, the workmanlike application of those sound engineering principles isn\u2019t always thrilling \u2013 until the software is completed on time and works without surprise.<\/p><p>The columns in this book are about a more glamorous aspect of the profession: programming pearls whose origins lie beyond solid engineering, in the realm of insight and creativity. Just as natural pearls grow from grains of sand that have irritated oysters, these programming pearls have grown from real problems that have irritated real programmers. The programs are fun, and they teach important programming techniques and fundamental design principles.\u201d<a href=\"https:\/\/amzn.to\/2u8cOxo\">Amazon.com<\/a><\/p><h3 id=\"15-peopleware-https-amzn-to-2p2yrih-by-tom-demarco-tim-lister-17-6-recommended\">15. <a href=\"https:\/\/amzn.to\/2P2YRIh\">Peopleware<\/a> by Tom DeMarco &amp; Tim Lister (17.6% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/15.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cDrawing on their software development and management experience, and highlighting the insights and wisdom of other successful managers, Mantle and Lichty provide the rules, tools, and insights you need to manage and understand people and teams in order to deliver software successfully and avoid projects that have run catastrophically over schedule and budget. The unique insight of this longtime bestseller is that the major issues of software development are human, not technical. They\u2019re not easy issues; but solve them, and you\u2019ll maximize your chances of success. With a blend of software engineering facts and thought-provoking opinions, Fred Brooks offers insight for anyone managing complex projects.\u201d<a href=\"https:\/\/amzn.to\/2P2YRIh\">Amazon.com<\/a><\/p><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/14.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cSome books on algorithms are rigorous but incomplete; others cover masses of material but lack rigor. Introduction to Algorithms uniquely combines rigor and comprehensiveness. The book covers a broad range of algorithms in depth, yet makes their design and analysis accessible to all levels of readers. Each chapter is relatively self-contained and can be used as a unit of study. The algorithms are described in English and in a pseudocode designed to be readable by anyone who has done a little programming. The explanations have been kept elementary without sacrificing depth of coverage or mathematical rigor.<\/p><p>The first edition became a widely used text in universities worldwide as well as the standard reference for professionals. The second edition featured new chapters on the role of algorithms, probabilistic analysis and randomized algorithms, and linear programming. The third edition has been revised and updated throughout. It includes two completely new chapters, on van Emde Boas trees and multithreaded algorithms, substantial additions to the chapter on recurrence (now called \u201cDivide-and-Conquer\u201d), and an appendix on matrices. It features improved treatment of dynamic programming and greedy algorithms and a new notion of edge-based flow in the material on flow networks. Many exercises and problems have been added for this edition. The international paperback edition is no longer available; the hardcover is available worldwide.\u201d<a href=\"https:\/\/amzn.to\/2SAdPrt\">Amazon.com<\/a><\/p><h3 id=\"13-code-https-amzn-to-38zhtoo-by-charles-petzold-19-1-recommended\">13. <a href=\"https:\/\/amzn.to\/38zhTOo\">Code<\/a> by Charles Petzold (19.1% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/13.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cWhat do flashlights, the British invasion, black cats, and seesaws have to do with computers? In CODE, they show us the ingenious ways we manipulate language and invent new means of communicating with each other. And through CODE, we see how this ingenuity and our very human compulsion to communicate have driven the technological innovations of the past two centuries. Using everyday objects and familiar language systems such as Braille and Morse code, author Charles Petzold weaves an illuminating narrative for anyone who\u2019s ever wondered about the secret inner life of computers and other smart machines. It\u2019s a cleverly illustrated and eminently comprehensible story\u2014and along the way, you\u2019ll discover you\u2019ve gained a real context for understanding today\u2019s world of PCs, digital media, and the Internet. No matter what your level of technical savvy, CODE will charm you\u2014and perhaps even awaken the technophile within.\u201d<a href=\"https:\/\/amzn.to\/38zhTOo\">Amazon.com<\/a><\/p><h3 id=\"12-don-t-make-me-think-https-amzn-to-2uypsw3-by-steve-krug-19-1-recommended\">12. <a href=\"https:\/\/amzn.to\/2UYPsW3\">Don\u2019t Make Me Think<\/a> by Steve Krug (19.1% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/12.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cSince Don\u2019t Make Me Think was first published in 2000, hundreds of thousands of Web designers and developers have relied on usability guru Steve Krug\u2019s guide to help them understand the principles of intuitive navigation and information design. Witty, commonsensical, and eminently practical, it\u2019s one of the best-loved and most recommended books on the subject.<\/p><p>Now Steve returns with fresh perspective to reexamine the principles that made Don\u2019t Make Me Think a classic\u2013with updated examples and a new chapter on mobile usability. And it\u2019s still short, profusely illustrated\u2026and best of all\u2013fun to read.<\/p><p>If you\u2019ve read it before, you\u2019ll rediscover what made Don\u2019t Make Me Think so essential to Web designers and developers around the world. If you\u2019ve never read it, you\u2019ll see why so many people have said it should be required reading for anyone working on Web sites.\u201d<a href=\"https:\/\/amzn.to\/2UYPsW3\">Amazon.com<\/a><\/p><h3 id=\"11-soft-skills-https-amzn-to-2hzitgl-by-john-sonmez-22-recommended\">11. <a href=\"https:\/\/amzn.to\/2HziTGl\">Soft Skills<\/a> by John Sonmez (22% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/11.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cFor most software developers, coding is the fun part. The hard bits are dealing with clients, peers, and managers, staying productive, achieving financial security, keeping yourself in shape, and finding true love. This book is here to help.<\/p><p>Soft Skills: The software developer\u2019s life manual is a guide to a well-rounded, satisfying life as a technology professional. In it, developer and life coach John Sonmez offers advice to developers on important \u201csoft\u201d subjects like career and productivity, personal finance and investing, and even fitness and relationships. Arranged as a collection of 71 short chapters, this fun-to-read book invites you to dip in wherever you like. A Taking Action section at the end of each chapter shows you how to get quick results. Soft Skills will help make you a better programmer, a more valuable employee, and a happier, healthier person.\u201d<a href=\"https:\/\/amzn.to\/2HziTGl\">Amazon.com<\/a><\/p><h3 id=\"10-cracking-the-coding-interview-https-amzn-to-37cuhvd-by-gayle-laakmann-mcdowell-22-recommended\">10. <a href=\"https:\/\/amzn.to\/37CuhvD\">Cracking the Coding Interview<\/a> by Gayle Laakmann McDowell (22% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/10.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cI am not a recruiter. I am a software engineer. And as such, I know what it\u2019s like to be asked to whip up brilliant algorithms on the spot and then write flawless code on a whiteboard. I\u2019ve been through this as a candidate and as an interviewer.<\/p><p>Cracking the Coding Interview, 6th Edition is here to help you through this process, teaching you what you need to know and enabling you to perform at your very best. I\u2019ve coached and interviewed hundreds of software engineers. The result is this book.<\/p><p>Learn how to uncover the hints and hidden details in a question, discover how to break down a problem into manageable chunks, develop techniques to unstick yourself when stuck, learn (or re-learn) core computer science concepts, and practice on 189 interview questions and solutions.<\/p><p>These interview questions are real; they are not pulled out of computer science textbooks. They reflect what\u2019s truly being asked at the top companies, so that you can be as prepared as possible. WHAT\u2019S INSIDE? - 189 programming interview questions, ranging from the basics to the trickiest algorithm problems. - A walk-through of how to derive each solution, so that you can learn how to get there yourself. - Hints on how to solve each of the 189 questions, just like what you would get in a real interview. - Five proven strategies to tackle algorithm questions, so that you can solve questions you haven\u2019t seen. - Extensive coverage of essential topics, such as big O time, data structures, and core algorithms. - A behind the scenes\u009d look at how top companies like Google and Facebook hire developers. - Techniques to prepare for and ace the soft side of the interview: behavioral questions. - For interviewers and companies: details on what makes a good interview question and hiring process.\u201d<a href=\"https:\/\/amzn.to\/37CuhvD\">Amazon.com<\/a><\/p><h3 id=\"9-design-patterns-https-amzn-to-38eb9yo-by-by-erich-gamma-richard-helm-ralph-johnson-john-vlissides-25-recommended\">9. <a href=\"https:\/\/amzn.to\/38Eb9yO\">Design Patterns<\/a> by by Erich Gamma \/ Richard Helm \/ Ralph Johnson \/ John Vlissides (25% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/9.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cCapturing a wealth of experience about the design of object-oriented software, four top-notch designers present a catalog of simple and succinct solutions to commonly occurring design problems. Previously undocumented, these 23 patterns allow designers to create more flexible, elegant, and ultimately reusable designs without having to rediscover the design solutions themselves.<\/p><p>The authors begin by describing what patterns are and how they can help you design object-oriented software. They then go on to systematically name, explain, evaluate, and catalog recurring designs in object-oriented systems. With Design Patterns as your guide, you will learn how these important patterns fit into the software development process, and how you can leverage them to solve your own design problems most efficiently.<\/p><p>Each pattern describes the circumstances in which it is applicable, when it can be applied in view of other design constraints, and the consequences and trade-offs of using the pattern within a larger design. All patterns are compiled from real systems and are based on real-world examples. Each pattern also includes code that demonstrates how it may be implemented in object-oriented programming languages like C++ or Smalltalk.\u201d<a href=\"https:\/\/amzn.to\/38Eb9yO\">Amazon.com<\/a><\/p><h3 id=\"8-working-effectively-with-legacy-code-https-amzn-to-2p37dwj-by-michael-feathers-26-4-recommended\">8. <a href=\"https:\/\/amzn.to\/2P37DWJ\">Working Effectively with Legacy Code<\/a> by Michael Feathers (26.4% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/8.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cIn this book, Michael Feathers offers start-to-finish strategies for working more effectively with large, untested legacy code bases. This book draws on material Michael created for his own renowned Object Mentor seminars: techniques Michael has used in mentoring to help hundreds of developers, technical managers, and testers bring their legacy systems under control. This book also includes a catalog of twenty-four dependency-breaking techniques that help you work with program elements in isolation and make safer changes.\u201d<a href=\"https:\/\/amzn.to\/2P37DWJ\">Amazon.com<\/a><\/p><h3 id=\"7-the-clean-coder-https-amzn-to-2u63sis-by-robert-martin-27-9-recommended\">7. <a href=\"https:\/\/amzn.to\/2u63SIS\">The Clean Coder<\/a> by Robert Martin (27.9% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/7.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cProgrammers who endure and succeed amidst swirling uncertainty and nonstop pressure share a common attribute: They care deeply about the practice of creating software. They treat it as a craft. They are professionals.<\/p><p>In The Clean Coder: A Code of Conduct for Professional Programmers, legendary software expert Robert C. Martin introduces the disciplines, techniques, tools, and practices of true software craftsmanship. This book is packed with practical advice\u2013about everything from estimating and coding to refactoring and testing. It covers much more than technique: It is about attitude. Martin shows how to approach software development with honor, self-respect, and pride; work well and work clean; communicate and estimate faithfully; face difficult decisions with clarity and honesty; and understand that deep knowledge comes with a responsibility to act.<\/p><p>Great software is something to marvel at: powerful, elegant, functional, a pleasure to work with as both a developer and as a user. Great software isn\u2019t written by machines. It is written by professionals with an unshakable commitment to craftsmanship. The Clean Coder will help you become one of them\u2013and earn the pride and fulfillment that they alone possess.\u201d<a href=\"https:\/\/amzn.to\/2u63SIS\">Amazon.com<\/a><\/p><h3 id=\"6-the-mythical-man-month-https-amzn-to-3bl1lvz-by-frederick-p-brooks-jr-27-9-recommended\">6. <a href=\"https:\/\/amzn.to\/3bL1lVz\">The Mythical Man-Month<\/a> by Frederick P. Brooks Jr (27.9% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/6.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cFew books on software project management have been as influential and timeless as The Mythical Man-Month. With a blend of software engineering facts and thought-provoking opinions, Fred Brooks offers insight for anyone managing complex projects. These essays draw from his experience as project manager for the IBM System\/360 computer family and then for OS\/360, its massive software system. Now, 20 years after the initial publication of his book, Brooks has revisited his original ideas and added new thoughts and advice, both for readers already familiar with his work and for readers discovering it for the first time.\u201d<a href=\"https:\/\/amzn.to\/3bL1lVz\">Amazon.com<\/a><\/p><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/5.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cWhat\u2019s so special about design patterns?<\/p><p>At any given moment, someone struggles with the same software design problems you have. And, chances are, someone else has already solved your problem. This edition of Head First Design Patterns\u2014now updated for Java 8\u2014shows you the tried-and-true, road-tested patterns used by developers to create functional, elegant, reusable, and flexible software. By the time you finish this book, you\u2019ll be able to take advantage of the best design practices and experiences of those who have fought the beast of software design and triumphed.<\/p><p>What\u2019s so special about this book?<\/p><p>We think your time is too valuable to spend struggling with new concepts. Using the latest research in cognitive science and learning theory to craft a multi-sensory learning experience, Head First Design Patterns uses a visually rich format designed for the way your brain works, not a text-heavy approach that puts you to sleep.\u201d<a href=\"https:\/\/amzn.to\/2SUiUtA\">Amazon.com<\/a><\/p><h3 id=\"4-refactoring-https-amzn-to-2uss167-by-martin-fowler-35-recommended\">4. <a href=\"https:\/\/amzn.to\/2uSs167\">Refactoring<\/a> by Martin Fowler (35% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/4.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cAs the application of object technology\u2013particularly the Java programming language\u2013has become commonplace, a new problem has emerged to confront the software development community. Significant numbers of poorly designed programs have been created by less-experienced developers, resulting in applications that are inefficient and hard to maintain and extend. Increasingly, software system professionals are discovering just how difficult it is to work with these inherited, non-optimal applications.<\/p><p>For several years, expert-level object programmers have employed a growing collection of techniques to improve the structural integrity and performance of such existing software programs. Referred to as refactoring, these practices have remained in the domain of experts because no attempt has been made to transcribe the lore into a form that all developers could use\u2026until now. In Refactoring: Improving the Design of Existing Software, renowned object technology mentor Martin Fowler breaks new ground, demystifying these master practices and demonstrating how software practitioners can realize the significant benefits of this new process. With proper training a skilled system designer\u201d <a href=\"https:\/\/amzn.to\/2uSs167\">Amazon.com<\/a><\/p><h3 id=\"3-code-complete-https-amzn-to-37wo0wr-by-steve-mcconnell-42-recommended\">3. <a href=\"https:\/\/amzn.to\/37wO0wR\">Code Complete<\/a> by Steve McConnell (42% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/3.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cWidely considered one of the best practical guides to programming, Steve McConnell\u2019s original CODE COMPLETE has been helping developers write better software for more than a decade. Now this classic book has been fully updated and revised with leading-edge practices\u2014and hundreds of new code samples\u2014illustrating the art and science of software construction. Capturing the body of knowledge available from research, academia, and everyday commercial practice, McConnell synthesizes the most effective techniques and must-know principles into clear, pragmatic guidance. No matter what your experience level, development environment, or project size, this book will inform and stimulate your thinking\u2014and help you build the highest quality code.\u201d<a href=\"https:\/\/amzn.to\/37wO0wR\">Amazon.com<\/a><\/p><h3 id=\"2-clean-code-https-amzn-to-2hy0n0u-by-robert-c-martin-66-recommended\">2. <a href=\"https:\/\/amzn.to\/2Hy0n0U\">Clean Code<\/a> by Robert C. Martin (66% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/2.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201cClean Code is divided into three parts. The first describes the principles, patterns, and practices of writing clean code. The second part consists of several case studies of increasing complexity. Each case study is an exercise in cleaning up code\u2014of transforming a code base that has some problems into one that is sound and efficient. The third part is the payoff: a single chapter containing a list of heuristics and \u201csmells\u201d gathered while creating the case studies. The result is a knowledge base that describes the way we think when we write, read, and clean code.\u201d<a href=\"https:\/\/amzn.to\/2Hy0n0U\">Amazon.com<\/a><\/p><h3 id=\"1-the-pragmatic-programmer-https-amzn-to-3bo6xyb-by-david-thomas-andrew-hunt-67-recommended\">1. <a href=\"https:\/\/amzn.to\/3bO6Xyb\">The Pragmatic Programmer<\/a> by David Thomas &amp; Andrew Hunt (67% recommended)<\/h3><p><img src=\"https:\/\/www.daolf.com\/images\/programming_book_list\/1.jpg#center\" alt=\"\" referrerpolicy=\"no-referrer\" \/> \u201c The Pragmatic Programmer is one of those rare tech books you\u2019ll read, re-read, and read again over the years. Whether you\u2019re new to the field or an experienced practitioner, you\u2019ll come away with fresh insights each and every time.<\/p><p>Dave Thomas and Andy Hunt wrote the first edition of this influential book in 1999 to help their clients create better software and rediscover the joy of coding. These lessons have helped a generation of programmers examine the very essence of software development, independent of any particular language, framework, or methodology, and the Pragmatic philosophy has spawned hundreds of books, screencasts, and audio books, as well as thousands of careers and success stories.<\/p><p>Now, twenty years later, this new edition re-examines what it means to be a modern programmer. Topics range from personal responsibility and career development to architectural techniques for keeping your code flexible and easy to adapt and reuse.\u201d<a href=\"https:\/\/amzn.to\/3bO6Xyb\">Amazon.com<\/a><\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>Although the order might suprise some, by definition, most of you must have heard of these books already.<\/p><p>A few additional things I learned making this list:<\/p><ul><li>Martin Fowler and Steve McConnell are the only author with several books in the list.<\/li>\n<li>Cracking to Code interview is the most recent book on the list, released in 2015.<\/li>\n<li><a href=\"https:\/\/amzn.to\/321qbMx\">Python Programming<\/a>, by John Zelle was the most cited book focused on one laguage. It would have #5 had I taken it into account.<\/li>\n<\/ul><p>I hope you enjoyed this article.<\/p><p>I must admit, this one took a while to write. If you liked this article and feel like Twitter would like it please do not hesitate to love and retweet, it really does help :).<\/p><p>edit: Robert C. Martin also have 2 books on the list.<\/p>","created_at":"2020-03-02T03:23:37+0000","updated_at":"2020-03-02T03:23:37+0000","published_at":"2020-02-16T16:27:09+0000","published_by":["Pierre de Wulf"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":22,"domain_name":"www.daolf.com","preview_picture":"https:\/\/www.daolf.com\/images\/programming_book_list\/twitter_cover.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1149,"title":"Kubernetes for Full-Stack Developers, a self-guided course.","url":"https:\/\/www.digitalocean.com\/community\/curriculums\/kubernetes-for-full-stack-developers","content":"<p>Whether you\u2019re just curious, getting started with Kubernetes, or have experience with it, this curriculum will help you learn more about Kubernetes and running containerized applications. You\u2019ll learn about core Kubernetes concepts and use them to deploy and scale applications in practical tutorials. By the end of this curriculum you\u2019ll be able to create your own Kubernetes cluster from scratch and run your own applications on it. You will also learn how to set up monitoring, alerting, and automation for your applications on Kubernetes.<\/p><p><strong>Download the entire curriculum as an eBook!<\/strong><\/p><p><a href=\"https:\/\/do.co\/k8s-book-epub\"><em>Kubernetes for Full-Stack Developers<\/em> eBook in <strong>EPUB format<\/strong><\/a><\/p><p><a href=\"https:\/\/do.co\/k8s-book-pdf\"><em>Kubernetes for Full-Stack Developers<\/em> eBook in <strong>PDF format<\/strong><\/a><br \/><\/p>","created_at":"2020-03-01T13:45:11+0000","updated_at":"2020-03-01T13:45:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.digitalocean.com","http_status":"200","origin_url":"https:\/\/www.digitalocean.com\/community\/curriculums\/kubernetes-for-full-stack-developers?utm_source=changelog&utm_medium=newsletter&utm_campaign=changelog-news"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1148,"title":"Lordicon","url":"https:\/\/lordicon.com\/","content":"<p>Assembling 500+ interactive icons.<br \/>It may take a few seconds.<\/p><div><section id=\"essential\"><div class=\"container\"><h2>Micro-interactions<\/h2><h3>Animated Lottie icons, effortlessly harmonize with your brand\u2019s narrative.<\/h3><div class=\"cols\"><div class=\"icons-outline\"><p>Outline<\/p><p>Solid<\/p><\/div><div class=\"icons-freedom animation-examples\"><p>Auto<\/p><p>Hover<\/p><p>On Click<\/p><p>Morph<\/p><\/div><div class=\"description-outline\"><p><strong>Perfect harmony<\/strong><\/p><p>Each icon was developed in outline and solid style.<\/p><\/div><div class=\"description-freedom\"><p><strong>Exercise your freedom<\/strong><\/p><p>Have fun with different animation players.<\/p><\/div><div class=\"description-scale\"><p><strong>Thank you Lottie<\/strong><\/p><p>Lordicon base on powerful Lottie engine introduced by Airbnb Design.<\/p><\/div><\/div><p><a class=\"btn\" href=\"https:\/\/lordicon.com\/essential-pack\">EXPLORE ESSENTIAL PACK<\/a><\/p><\/div><\/section><section id=\"integrations\"><div class=\"container\"><h2>Embed animated icons anywhere<\/h2><h3>Implement Lordicon animations on a Wordpress (incl. Elementor, Divi),<br \/>Shopify, Webflow, Squarespace, Wix, any Website or Mobile App.<\/h3><div class=\"parts\"><div><img alt=\"WordPress plugin for LordIcon\" src=\"https:\/\/lordicon.com\/assets\/pictures\/wordpress.png\" referrerpolicy=\"no-referrer\" \/><strong>WordPress plug-in<\/strong><p>A free Lordicon WordPress plug-in, which lets you easily insert and customize icons for Wordpress websites (incl. Goutenberg, Elementor, Divi, custom themes etc).<\/p><\/div><p><a class=\"btn\" href=\"https:\/\/lordicon.com\/wordpress-plugin\">Learn more<\/a><\/p><div><img alt=\"Lottie Web Player\" src=\"https:\/\/lordicon.com\/assets\/pictures\/lottiefiles.png\" referrerpolicy=\"no-referrer\" \/><strong>Lottie Web Player<\/strong><p>Lottie Files powerful open source Web Player lets you embed Lottie Animation on your website in a few simple clicks.<\/p><\/div><p><a class=\"btn\" target=\"_blank\" href=\"https:\/\/lottiefiles.com\/web-player\">Learn more<\/a><\/p><div><img alt=\"Manual integration\" src=\"https:\/\/lordicon.com\/assets\/pictures\/github.png\" referrerpolicy=\"no-referrer\" \/><strong>Manual integration<\/strong><p>Lottie libraries and plugins available for free <a target=\"_blank\" href=\"https:\/\/github.com\/tomwilusz\/lord-icon-element\">Web<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/airbnb\/lottie-ios\">iOS<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/airbnb\/lottie-android\">Android<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/simolus3\/fluttie\">Flutter<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/react-native-community\/lottie-react-native\">React Native<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/martijn00\/LottieXamarin\">Xamarin<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/bradmartin\/nativescript-lottie\">NativeScript<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/windows-toolkit\/Lottie-Windows\">Windows<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/chenqingspring\/vue-lottie\">Vue<\/a>, <a target=\"_blank\" href=\"https:\/\/github.com\/chenqingspring\/ng-lottie\">Angular<\/a>, <a target=\"_blank\" href=\"https:\/\/blog.qt.io\/blog\/2019\/03\/08\/announcing-qtlottie\/\">QT<\/a>, <a target=\"_blank\" href=\"https:\/\/skia.org\/user\/modules\/skottie\">Skia<\/a>, <a target=\"_blank\" href=\"https:\/\/store.framer.com\/package\/airbnb\/lottie\">Framer X<\/a>, and <a target=\"_blank\" href=\"https:\/\/aeux.io\/\">Sketch<\/a>.<\/p><\/div><p><a class=\"btn\" href=\"https:\/\/github.com\/tomwilusz\/lord-icon-element\">Learn more<\/a><\/p><\/div><\/div><\/section><section id=\"credentials\"><div class=\"container\"><h2>+10 000 satisfied users<\/h2><div class=\"items\"><a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/twitter.com\/cesarsvergara\/status\/1169334939772497921\">\n<figure><div class=\"description\"><p><strong>@Lordicon9<\/strong> it\u2019s a great pack of icons. I\u2019ve used in many of my projects, and going to use in more of them. I definitively recommend to anyone who needs beautiful and simple icons <strong>#icons<\/strong> <strong>#flaticons<\/strong> <strong>#uicons<\/strong><\/p><\/div>\n<\/figure><\/a><a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/twitter.com\/Yakovchuk_Roman\/status\/1182591522371440645\">\n<figure><div class=\"description\"><p>I\u2019ve been looking for amazing icons for ages, and <strong>@Lordicon9<\/strong> was love at first sight! Their icons are gorgeous, and they gave my website a fresh look! I\u2019ve been recommending it to all my friends, and they\u2019ve loved it, too! <strong>#animatedicons<\/strong> <strong>#lottie<\/strong><\/p><\/div>\n<\/figure><\/a><a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/twitter.com\/PSobotkowski\/status\/1181455449608118273\">\n<figure><div class=\"description\"><p>Our clients were absolutely delighted with <strong>@Lordicon9<\/strong> icon sets. Beautifully designed and perfectly animated, they are the nicest icons one could hope for! <strong>#web<\/strong> <strong>#ux<\/strong> <strong>#ui<\/strong> <strong>#design<\/strong><\/p><\/div>\n<\/figure><\/a><a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/twitter.com\/LukaszCiesielsk\/status\/1222886140363059202\">\n<figure><div class=\"description\"><p><strong>@Lordicon9<\/strong> icon sets is cool. We\u2019re using them at @pragmaticcoders in our website based on #wordpress. Great value for money. <strong>#animatedicons<\/strong> <strong>#icons<\/strong><\/p><\/div>\n<\/figure><\/a><\/div><\/div><\/section><section id=\"custom-services\"><div class=\"container cols\"><div><h2>Custom Services<\/h2><p>We provide animated icons sets, interactive logo animations, and animated illustrations. Contact us to set your projects in motion!<\/p><a class=\"btn light\" href=\"https:\/\/lordicon.com\/custom-services\">learn more<\/a><\/div><\/div><\/section><\/div>","created_at":"2020-02-26T16:39:23+0000","updated_at":"2020-02-26T16:39:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"lordicon.com","preview_picture":"https:\/\/lordicon.com\/assets\/pictures\/avatar-lordicon.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1147,"title":"airbnb\/lottie-web","url":"https:\/\/github.com\/airbnb\/lottie-web","content":"<p>Lottie is a mobile library for Web, and iOS that parses <a href=\"http:\/\/www.adobe.com\/products\/aftereffects.html\" rel=\"nofollow\">Adobe After Effects<\/a> animations exported as json with <a href=\"https:\/\/github.com\/airbnb\/lottie-web\">Bodymovin<\/a> and renders them natively on mobile!<\/p><p>For the first time, designers can create <strong>and ship<\/strong> beautiful animations without an engineer painstakingly recreating it by hand. They say a picture is worth 1,000 words so here are 13,000:<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/airbnb\/lottie-web\/blob\/master\/gifs\/Example1.gif\"><img src=\"https:\/\/github.com\/airbnb\/lottie-web\/raw\/master\/gifs\/Example1.gif\" alt=\"Example1\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/airbnb\/lottie-web\/blob\/master\/gifs\/Example2.gif\"><img src=\"https:\/\/github.com\/airbnb\/lottie-web\/raw\/master\/gifs\/Example2.gif\" alt=\"Example2\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/airbnb\/lottie-web\/blob\/master\/gifs\/Example3.gif\"><img src=\"https:\/\/github.com\/airbnb\/lottie-web\/raw\/master\/gifs\/Example3.gif\" alt=\"Example3\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/airbnb\/lottie-web\/blob\/master\/gifs\/Community%202_3.gif\"><img src=\"https:\/\/github.com\/airbnb\/lottie-web\/raw\/master\/gifs\/Community%202_3.gif\" alt=\"Community\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/airbnb\/lottie-web\/blob\/master\/gifs\/Example4.gif\"><img src=\"https:\/\/github.com\/airbnb\/lottie-web\/raw\/master\/gifs\/Example4.gif\" alt=\"Example4\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-option-1-recommended\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-1-recommended\"><\/a>Option 1 (Recommended):<\/h3><p><strong>Download it from from aescripts + aeplugins:<\/strong> <a href=\"http:\/\/aescripts.com\/bodymovin\/\" rel=\"nofollow\">http:\/\/aescripts.com\/bodymovin\/<\/a><\/p><h3><a id=\"user-content-option-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-2\"><\/a>Option 2:<\/h3><p><strong>Or get it from the adobe store<\/strong> <a href=\"https:\/\/creative.adobe.com\/addons\/products\/12557\" rel=\"nofollow\">https:\/\/creative.adobe.com\/addons\/products\/12557<\/a> CC 2014 and up.<\/p><h2><a id=\"user-content-other-installation-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-installation-options\"><\/a>Other installation options:<\/h2><h3><a id=\"user-content-option-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-3\"><\/a>Option 3:<\/h3><ul><li>download the ZIP from the repo.<\/li>\n<li>Extract content and get the .zxp file from '\/build\/extension'<\/li>\n<li>Use the <a href=\"http:\/\/aescripts.com\/learn\/zxp-installer\/\" rel=\"nofollow\">ZXP installer<\/a> from aescripts.com.<\/li>\n<\/ul><h3><a id=\"user-content-option-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-4\"><\/a>Option 4:<\/h3><ul><li>\n<p>Close After Effects<br \/><\/p>\n<\/li>\n<li>\n<p>Extract the zipped file on <code>build\/extension\/bodymovin.zxp<\/code> to the adobe CEP folder:<br \/>WINDOWS:<br \/><code>C:\\Program Files (x86)\\Common Files\\Adobe\\CEP\\extensions or<\/code><br \/><code>C:\\&lt;username&gt;\\AppData\\Roaming\\Adobe\\CEP\\extensions<\/code><br \/>MAC:<br \/><code>\/Library\/Application\\ Support\/Adobe\/CEP\/extensions\/bodymovin<\/code><br \/>(you can open the terminal and type:<br \/><code>$ cp -R YOURUNZIPEDFOLDERPATH\/extension \/Library\/Application\\ Support\/Adobe\/CEP\/extensions\/bodymovin<\/code><br \/>then type:<br \/><code>$ ls \/Library\/Application\\ Support\/Adobe\/CEP\/extensions\/bodymovin<\/code><br \/>to make sure it was copied correctly type)<br \/><\/p>\n<\/li>\n<li>\n<p>Edit the registry key:<br \/>WINDOWS:<br \/>open the registry key <code>HKEY_CURRENT_USER\/Software\/Adobe\/CSXS.6<\/code> and add a key named <code>PlayerDebugMode<\/code>, of type String, and value <code>1<\/code>.<br \/>MAC:<br \/>open the file <code>~\/Library\/Preferences\/com.adobe.CSXS.6.plist<\/code> and add a row with key <code>PlayerDebugMode<\/code>, of type String, and value <code>1<\/code>.<br \/><\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-option-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-5\"><\/a>Option 5:<\/h3><p>Install the zxp manually following the instructions here: <a href=\"https:\/\/helpx.adobe.com\/x-productkb\/global\/installingextensionsandaddons.html\" rel=\"nofollow\">https:\/\/helpx.adobe.com\/x-productkb\/global\/installingextensionsandaddons.html<\/a> Skip directly to \"Install third-party extensions\"<\/p><h3><a id=\"user-content-option-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-6\"><\/a>Option 6:<\/h3><p>Install with <a href=\"http:\/\/brew.sh\" rel=\"nofollow\">Homebrew<\/a>-<a href=\"https:\/\/github.com\/danielbayley\/homebrew-adobe\">adobe<\/a>:<\/p><div class=\"highlight highlight-source-shell\"><pre>brew tap danielbayley\/adobe\nbrew cask install lottie<\/pre><\/div><h3><a id=\"user-content-after-installing\" class=\"anchor\" aria-hidden=\"true\" href=\"#after-installing\"><\/a>After installing<\/h3><ul><li><strong>Windows:<\/strong> Go to Edit &gt; Preferences &gt; Scripting &amp; Expressions... &gt; and check on \"Allow Scripts to Write Files and Access Network\"<\/li>\n<li><strong>Mac:<\/strong> Go to Adobe After Effects &gt; Preferences &gt; Scripting &amp; Expressions... &gt; and check on \"Allow Scripts to Write Files and Access Network\"<\/li>\n<\/ul><p><strong>Old Versions<\/strong><\/p><ul><li><strong>Windows:<\/strong> Go to Edit &gt; Preferences &gt; General &gt; and check on \"Allow Scripts to Write Files and Access Network\"<\/li>\n<li><strong>Mac:<\/strong> Go to Adobe After Effects &gt; Preferences &gt; General &gt; and check on \"Allow Scripts to Write Files and Access Network\"<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre># with npm\nnpm install lottie-web\n# with bower\nbower install bodymovin<\/pre><\/div><p>Or you can use the script file from here: <a href=\"https:\/\/cdnjs.com\/libraries\/bodymovin\" rel=\"nofollow\">https:\/\/cdnjs.com\/libraries\/bodymovin<\/a> Or get it directly from the AE plugin clicking on Get Player<\/p><p><a href=\"https:\/\/codepen.io\/airnan\/project\/editor\/ZeNONO\/\" rel=\"nofollow\">See a basic implementation here.<\/a><br \/><\/p><p><a href=\"http:\/\/codepen.io\/collection\/nVYWZR\/\" rel=\"nofollow\">See examples on codepen.<\/a><br \/><\/p><h2><a id=\"user-content-how-it-works\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-it-works\"><\/a>How it works<\/h2><p><a href=\"https:\/\/www.youtube.com\/watch?v=5XMUJdjI0L8\" rel=\"nofollow\">Here's<\/a> a video tutorial explaining how to export a basic animation and load it in an html page<br \/><\/p><h3><a id=\"user-content-after-effects\" class=\"anchor\" aria-hidden=\"true\" href=\"#after-effects\"><\/a>After Effects<\/h3><ul><li>Open your AE project and select the bodymovin extension on Window &gt; Extensions &gt; bodymovin<\/li>\n<li>A Panel will open with a Compositions tab listing all of your Project Compositions.<\/li>\n<li>Select the composition you want to export.<\/li>\n<li>Select a Destination Folder.<\/li>\n<li>Click Render<\/li>\n<li>look for the exported json file (if you had images or AI layers on your animation, there will be an images folder with the exported files)<\/li>\n<\/ul><h3><a id=\"user-content-html\" class=\"anchor\" aria-hidden=\"true\" href=\"#html\"><\/a>HTML<\/h3><ul><li>get the lottie.js file from the build\/player\/ folder for the latest build<\/li>\n<li>include the .js file on your html (remember to gzip it for production)<\/li>\n<\/ul><div class=\"highlight highlight-text-html-basic\"><pre>&lt;script src=\"js\/lottie.js\" type=\"text\/javascript\"&gt;&lt;\/script&gt;<\/pre><\/div><p>You can call lottie.loadAnimation() to start an animation. It takes an object as a unique param with:<\/p><ul><li>animationData: an Object with the exported animation data.<\/li>\n<li>path: the relative path to the animation object. (animationData and path are mutually exclusive)<\/li>\n<li>loop: true \/ false \/ number<\/li>\n<li>autoplay: true \/ false it will start playing as soon as it is ready<\/li>\n<li>name: animation name for future reference<\/li>\n<li>renderer: 'svg' \/ 'canvas' \/ 'html' to set the renderer<\/li>\n<li>container: the dom element on which to render the animation<\/li>\n<\/ul><p>It returns the animation instance you can control with play, pause, setSpeed, etc.<\/p><div class=\"highlight highlight-source-js\"><pre>lottie.loadAnimation({\n  container: element, \/\/ the dom element that will contain the animation\n  renderer: 'svg',\n  loop: true,\n  autoplay: true,\n  path: 'data.json' \/\/ the path to the animation json\n});<\/pre><\/div><h4><a id=\"user-content-composition-settings\" class=\"anchor\" aria-hidden=\"true\" href=\"#composition-settings\"><\/a>Composition Settings:<\/h4><p>Check this wiki page for an explanation for each setting. <a href=\"https:\/\/github.com\/airbnb\/lottie-web\/wiki\/Composition-Settings\">https:\/\/github.com\/airbnb\/lottie-web\/wiki\/Composition-Settings<\/a><\/p><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><p>Animation instances have these main methods:<\/p><h3><a id=\"user-content-play\" class=\"anchor\" aria-hidden=\"true\" href=\"#play\"><\/a>play<\/h3><hr \/><h3><a id=\"user-content-stop\" class=\"anchor\" aria-hidden=\"true\" href=\"#stop\"><\/a>stop<\/h3><hr \/><h3><a id=\"user-content-pause\" class=\"anchor\" aria-hidden=\"true\" href=\"#pause\"><\/a>pause<\/h3><hr \/><h3><a id=\"user-content-setspeedspeed\" class=\"anchor\" aria-hidden=\"true\" href=\"#setspeedspeed\"><\/a>setSpeed(speed)<\/h3><ul><li><code>speed<\/code>: 1 is normal speed.<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-gotoandstopvalue-isframe\" class=\"anchor\" aria-hidden=\"true\" href=\"#gotoandstopvalue-isframe\"><\/a>goToAndStop(value, isFrame)<\/h3><ul><li><code>value<\/code>: numeric value.<\/li>\n<li><code>isFrame<\/code>: defines if first argument is a time based value or a frame based (default false).<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-gotoandplayvalue-isframe\" class=\"anchor\" aria-hidden=\"true\" href=\"#gotoandplayvalue-isframe\"><\/a>goToAndPlay(value, isFrame)<\/h3><ul><li><code>value<\/code>: numeric value.<\/li>\n<li><code>isFrame<\/code>: defines if first argument is a time based value or a frame based (default false).<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-setdirectiondirection\" class=\"anchor\" aria-hidden=\"true\" href=\"#setdirectiondirection\"><\/a>setDirection(direction)<\/h3><ul><li><code>direction<\/code>: 1 is forward, -1 is reverse.<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-playsegmentssegments-forceflag\" class=\"anchor\" aria-hidden=\"true\" href=\"#playsegmentssegments-forceflag\"><\/a>playSegments(segments, forceFlag)<\/h3><ul><li><code>segments<\/code>: array. Can contain 2 numeric values that will be used as first and last frame of the animation. Or can contain a sequence of arrays each with 2 numeric values.<\/li>\n<li><code>forceFlag<\/code>: boolean. If set to false, it will wait until the current segment is complete. If true, it will update values immediately.<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-setsubframeusesubframes\" class=\"anchor\" aria-hidden=\"true\" href=\"#setsubframeusesubframes\"><\/a>setSubframe(useSubFrames)<\/h3><ul><li><code>useSubFrames<\/code>: If false, it will respect the original AE fps. If true, it will update on every requestAnimationFrame with intermediate values. Default is true.<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-destroy\" class=\"anchor\" aria-hidden=\"true\" href=\"#destroy\"><\/a>destroy()<\/h3><hr \/><h3><a id=\"user-content-getdurationinframes\" class=\"anchor\" aria-hidden=\"true\" href=\"#getdurationinframes\"><\/a>getDuration(inFrames)<\/h3><ul><li><code>inFrames<\/code>: If true, returns duration in frames, if false, in seconds.<\/li>\n<\/ul><hr \/><h3><a id=\"user-content-additional-methods\" class=\"anchor\" aria-hidden=\"true\" href=\"#additional-methods\"><\/a>Additional methods:<\/h3><ul><li>updateTextDocumentData -- updates a text layer's data <a href=\"https:\/\/github.com\/airbnb\/lottie-web\/wiki\/TextLayer.updateDocumentData\">More Info<\/a><\/li>\n<\/ul><hr \/><h3><a id=\"user-content-lottie-has-several-global-methods-that-will-affect-all-animations\" class=\"anchor\" aria-hidden=\"true\" href=\"#lottie-has-several-global-methods-that-will-affect-all-animations\"><\/a>Lottie has several global methods that will affect all animations:<\/h3><p><strong>lottie.play()<\/strong> -- with 1 optional parameter <strong>name<\/strong> to target a specific animation<br \/><strong>lottie.stop()<\/strong> -- with 1 optional parameter <strong>name<\/strong> to target a specific animation<br \/><strong>lottie.goToAndStop(value, isFrame, name)<\/strong> -- Moves an animation with the specified name playback to the defined time. If name is omitted, moves all animation instances.<br \/><strong>lottie.setSpeed()<\/strong> -- first argument speed (1 is normal speed) -- with 1 optional parameter <strong>name<\/strong> to target a specific animation<br \/><strong>lottie.setDirection()<\/strong> -- first argument direction (1 is normal direction.) -- with 1 optional parameter <strong>name<\/strong> to target a specific animation<br \/><strong>lottie.searchAnimations()<\/strong> -- looks for elements with class \"lottie\" or \"bodymovin\"<br \/><strong>lottie.loadAnimation()<\/strong> -- Explained above. returns an animation instance to control individually.<br \/><strong>lottie.destroy(name)<\/strong> -- Destroys an animation with the specified name. If name is omitted, destroys all animation instances. The DOM element will be emptied.<br \/><strong>lottie.registerAnimation()<\/strong> -- you can register an element directly with registerAnimation. It must have the \"data-animation-path\" attribute pointing at the data.json url<br \/><strong>lottie.getRegisteredAnimations()<\/strong> -- returns all animations instances<br \/><strong>lottie.setQuality()<\/strong> -- default 'high', set 'high','medium','low', or a number &gt; 1 to improve player performance. In some animations as low as 2 won't show any difference.<br \/><strong>lottie.setLocationHref()<\/strong> -- Sets the relative location from where svg elements with ids are referenced. It's useful when you experience mask issues in Safari.<br \/><strong>lottie.freeze()<\/strong> -- Freezes all playing animations or animations that will be loaded<br \/><strong>lottie.unfreeze()<\/strong> -- Unfreezes all animations<br \/><strong>lottie.inBrowser()<\/strong> -- true if the library is being run in a browser<br \/><strong>lottie.resize()<\/strong> -- Resizes all animation instances<br \/><\/p><h2><a id=\"user-content-events\" class=\"anchor\" aria-hidden=\"true\" href=\"#events\"><\/a>Events<\/h2><ul><li>onComplete<\/li>\n<li>onLoopComplete<\/li>\n<li>onEnterFrame<\/li>\n<li>onSegmentStart<\/li>\n<\/ul><p>you can also use addEventListener with the following events:<\/p><ul><li>complete<\/li>\n<li>loopComplete<\/li>\n<li>enterFrame<\/li>\n<li>segmentStart<\/li>\n<li>config_ready (when initial config is done)<\/li>\n<li>data_ready (when all parts of the animation have been loaded)<\/li>\n<li>data_failed (when part of the animation can not be loaded)<\/li>\n<li>loaded_images (when all image loads have either succeeded or errored)<\/li>\n<li>DOMLoaded (when elements have been added to the DOM)<\/li>\n<li>destroy<\/li>\n<\/ul><h4><a id=\"user-content-other-loading-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-loading-options\"><\/a>Other loading options<\/h4><ul><li>if you want to use an existing canvas to draw, you can pass an extra object: 'rendererSettings' with the following configuration:<\/li>\n<\/ul><div class=\"highlight highlight-source-js\"><pre>lottie.loadAnimation({\n  container: element, \/\/ the dom element\n  renderer: 'svg',\n  loop: true,\n  autoplay: true,\n  animationData: animationData, \/\/ the animation data\n  rendererSettings: {\n    context: canvasContext, \/\/ the canvas context\n    preserveAspectRatio: 'xMinYMin slice', \/\/ Supports the same options as the svg element's preserveAspectRatio property\n    clearCanvas: false,\n    progressiveLoad: false, \/\/ Boolean, only svg renderer, loads dom elements when needed. Might speed up initialization for large number of elements.\n    hideOnTransparent: true, \/\/Boolean, only svg renderer, hides elements when opacity reaches 0 (defaults to true)\n    className: 'some-css-class-name',\n    id: 'some-id',\n  }\n});<\/pre><\/div><p>Doing this you will have to handle the canvas clearing after each frame<br \/>Another way to load animations is adding specific attributes to a dom element. You have to include a div and set it's class to \"lottie\". If you do it before page load, it will automatically search for all tags with the class \"lottie\". Or you can call <code>lottie.searchAnimations()<\/code> after page load and it will search all elements with the class \"lottie\".<br \/><\/p><ul><li>Add the data.json to a folder relative to the html<\/li>\n<li>Create a div that will contain the animation.<\/li>\n<li><strong>Required<\/strong>\n<ul><li>A class called \"lottie\"<\/li>\n<li>A \"data-animation-path\" attribute with relative path to the data.json<\/li>\n<\/ul><\/li>\n<li><strong>Optional<\/strong>\n<ul><li>A \"data-anim-loop\" attribute<\/li>\n<li>A \"data-name\" attribute to specify a name to target play controls specifically<\/li>\n<\/ul><\/li>\n<\/ul><p><strong>Example<\/strong><\/p><div class=\"highlight highlight-text-html-basic\"><pre> &lt;div style=\"width:1067px;height:600px\"  class=\"lottie\" data-animation-path=\"animation\/\" data-anim-loop=\"true\" data-name=\"ninja\"&gt;&lt;\/div&gt;<\/pre><\/div><h2><a id=\"user-content-preview\" class=\"anchor\" aria-hidden=\"true\" href=\"#preview\"><\/a>Preview<\/h2><p>You can preview or take an svg snapshot of the animation to use as poster. After you render your animation, you can take a snapshot of any frame in the animation and save it to your disk. I recommend to pass the svg through an svg optimizer like <a href=\"https:\/\/jakearchibald.github.io\/svgomg\/\" rel=\"nofollow\">https:\/\/jakearchibald.github.io\/svgomg\/<\/a> and play around with their settings.<br \/><\/p><h2><a id=\"user-content-recommendations\" class=\"anchor\" aria-hidden=\"true\" href=\"#recommendations\"><\/a>Recommendations<\/h2><h3><a id=\"user-content-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#files\"><\/a>Files<\/h3><p>If you have any images or AI layers that you haven't converted to shapes (I recommend that you convert them, so they get exported as vectors, right click each layer and do: \"Create shapes from Vector Layers\"), they will be saved to an images folder relative to the destination json folder. Beware not to overwrite an existing folder on that same location.<\/p><h3><a id=\"user-content-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance\"><\/a>Performance<\/h3><p>This is real time rendering. Although it is pretty optimized, it always helps if you keep your AE project to what is necessary<br \/>More optimizations are on their way, but try not to use huge shapes in AE only to mask a small part of it.<br \/>Too many nodes will also affect performance.<\/p><h3><a id=\"user-content-help\" class=\"anchor\" aria-hidden=\"true\" href=\"#help\"><\/a>Help<\/h3><p>If you have any animations that don't work or want me to export them, don't hesitate to write.<br \/>I'm really interested in seeing what kind of problems the plugin has.<br \/>my email is <strong><a href=\"mailto:hernantorrisi@gmail.com\">hernantorrisi@gmail.com<\/a><\/strong><\/p><h2><a id=\"user-content-ae-feature-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#ae-feature-support\"><\/a>AE Feature Support<\/h2><ul><li>The script supports precomps, shapes, solids, images, null objects, texts<\/li>\n<li>It supports masks and inverted masks. Maybe other modes will come but it has a huge performance hit.<\/li>\n<li>It supports time remapping<\/li>\n<li>The script supports shapes, rectangles, ellipses and stars.<\/li>\n<li>Expressions. Check the wiki page for <a href=\"https:\/\/github.com\/bodymovin\/bodymovin\/wiki\/Expressions\">more info.<\/a><\/li>\n<li>Not supported: image sequences, videos and audio are not supported<\/li>\n<li><strong>No negative layer stretching<\/strong>! No idea why, but stretching a layer messes with all the data.<\/li>\n<\/ul><h2><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h2><p><code>npm install<\/code> or <code>bower install<\/code> first <code>npm start<\/code><\/p><h2><a id=\"user-content-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#notes\"><\/a>Notes<\/h2><ul><li>If you want to modify the parser or the player, there are some gulp commands that can simplify the task<\/li>\n<li>look at the great animations exported on codepen <a href=\"http:\/\/codepen.io\/collection\/nVYWZR\/\" rel=\"nofollow\">See examples on codepen.<\/a><\/li>\n<li>gzipping the animation jsons and the player have a huge reduction on the filesize. I recommend doing it if you use it for a project.<\/li>\n<\/ul><h2><a id=\"user-content-issues\" class=\"anchor\" aria-hidden=\"true\" href=\"#issues\"><\/a>Issues<\/h2><ul><li>For missing mask in Safari browser, please call lottie.setLocationHref(locationHref) before animation is generated. It usually caused by usage of base tag in html. (see above for description of setLocationHref)<\/li>\n<\/ul>","created_at":"2020-02-26T16:36:25+0000","updated_at":"2020-02-26T16:36:25+0000","published_by":["airbnb"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/698437?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1146,"title":"Free fonts | Butterick\u2019s Practical Typography","url":"https:\/\/practicaltypography.com\/free-fonts.html","content":"Free fonts | Butterick\u2019s Practical Typography\n<div id=\"content\">free fontsSome good<br \/>options, but<br \/>choose wisely<div class=\"font-alternatives\"><ul><li>\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/free-fonts-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>free\u00a0fonts<\/p><\/div>\n<\/li>\n<li><a href=\"https:\/\/practicaltypography.com\/fontrec\/fira-sans.html\" class=\"xref\">\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/fira-sans-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>fira\u00a0sans<\/p><\/div>\n<\/a><\/li>\n<li><a href=\"https:\/\/practicaltypography.com\/fontrec\/source-code-pro.html\" class=\"xref\">\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/source-code-pro-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>source code\u00a0pro<\/p><\/div>\n<\/a><\/li>\n<li><a href=\"https:\/\/practicaltypography.com\/fontrec\/source-serif-pro.html\" class=\"xref\">\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/source-serif-pro-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>source serif\u00a0pro<\/p><\/div>\n<\/a><\/li>\n<li><a href=\"https:\/\/practicaltypography.com\/fontrec\/cooper-hewitt.html\" class=\"xref\">\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/cooper-hewitt-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>cooper hewitt<\/p><\/div>\n<\/a><\/li>\n<li><a href=\"https:\/\/practicaltypography.com\/fontrec\/ibm-plex.html\" class=\"xref\">\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/ibm-plex-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>ibm\u00a0plex<\/p><\/div>\n<\/a><\/li>\n<li><a href=\"https:\/\/practicaltypography.com\/fontrec\/charter.html\" class=\"xref\">\n<div><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/charter-sample.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/p><p>char\u00adter<\/p><\/div>\n<\/a><\/li>\n<\/ul><\/div><p>Hun\u00addreds of free fonts have ar\u00adrived in re\u00adcent years. Not just \u201cfree\u201d as in \u201cal\u00adready in\u00adstalled on your com\u00adputer\u201d (aka <a href=\"https:\/\/practicaltypography.com\/system-fonts.html\" class=\"xref\">sys\u00adtem fonts<\/a>). But rather, re\u00adleased un\u00adder an open li\u00adcense, most of\u00adten the <a href=\"http:\/\/scripts.sil.org\/ofl\">OFL,<\/a> which per\u00admits you to do al\u00admost any\u00adthing you want with them. Most im\u00adpor\u00adtantly, you can re\u00addis\u00adtrib\u00adute copies for\u00a0free.<\/p><p>At times I\u2019ve been la\u00adbeled a free-font cur\u00admud\u00adgeon. My beef with free fonts is not that they\u2019re use\u00adless. On the con\u00adtrary, fonts with\u00adout pro\u00adpri\u00adetary li\u00adcens\u00ading re\u00adstric\u00adtions fill a need, most of all as a com\u00adple\u00adment to open-source soft\u00adware. For in\u00adstance, I couldn\u2019t have re\u00addesigned the <a href=\"https:\/\/docs.racket-lang.org\/\">doc\u00adu\u00admen\u00adta\u00adtion for Racket,<\/a> an open-source pro\u00adgram\u00adming lan\u00adguage, with\u00adout free fonts, be\u00adcause we had to ship them with the software.<\/p><p>In\u00adstead, my com\u00adplaint about free fonts has cen\u00adtered on a sim\u00adpler is\u00adsue: in terms of de\u00adsign and crafts\u00adman\u00adship, most <a href=\"https:\/\/practicaltypography.com\/why-google-fonts-arent-really-open-source.html\">free fonts are garbage.<\/a> Sure, that\u2019s true of many pro\u00adfes\u00adsional fonts too. But early on, cer\u00adtain free-font pro\u00adpo\u00adnents pushed the ar\u00adgu\u00adment that some\u00adhow, <a href=\"http:\/\/typedrawers.com\/discussion\/comment\/1827\/#Comment_1827\">quan\u00adtity mat\u00adtered more than quality.<\/a><\/p><p>For\u00adtu\u00adnately, not every\u00adone felt the same way. Across my years <a href=\"https:\/\/www.youtube.com\/watch?v=3L4YrGaR8E4\">rag\u00ading<\/a> against this par\u00adtic\u00adu\u00adlar ma\u00adchine, or\u00adga\u00adni\u00adza\u00adtions that ap\u00adpre\u00adci\u00adate the value of good de\u00adsign\u2014and had money to spend\u2014funded the de\u00advel\u00adop\u00adment of some free fonts that are ac\u00adtu\u00adally very\u00a0good:<\/p><ul><li>\n<p>Mozilla hired Erik Spiek\u00ader\u00admann (see <a href=\"https:\/\/practicaltypography.com\/foreword.html\" class=\"xref\">fore\u00adword<\/a>) and Ralph Car\u00adrois to make <a href=\"#\" class=\"xref\">Fira Sans<\/a> and <a href=\"#\" class=\"xref\">Fira Mono<\/a>, de\u00adsigns built upon the <a href=\"http:\/\/typographica.org\/typeface-reviews\/fira-sans\/\">adaman\u00adtium en\u00addoskele\u00adton<\/a> of his long\u00adtime best\u00adseller <a href=\"https:\/\/en.wikipedia.org\/wiki\/FF_Meta\">Meta.<\/a><\/p>\n<\/li>\n<li>\n<p>Adobe re\u00adleased a very nice <a href=\"https:\/\/practicaltypography.com\/monospaced-fonts.html\" class=\"xref\">mono\u00adspaced font<\/a> called <a href=\"#\" class=\"xref\">Source Code Pro<\/a>, de\u00adsigned by Paul Hunt; and the com\u00adpan\u00adion <a href=\"#\" class=\"xref\">Source Serif Pro<\/a>, de\u00adsigned by Frank Grie\u00dfhammer.<\/p>\n<\/li>\n<li>\n<p>The Smith\u00adson\u00adian com\u00admis\u00adsioned <a href=\"https:\/\/www.bspk.xyz\/\">Chester Jenk\u00adins<\/a> to cre\u00adate the ex\u00adcel\u00adlent <a href=\"#\" class=\"xref\">Cooper He\u00adwitt<\/a> sans serif family.<\/p>\n<\/li>\n<li>\n<p>Most re\u00adcently, IBM\u2019s Mike Ab\u00adbink worked with <a href=\"https:\/\/www.boldmonday.com\/\">Paul van der Laan<\/a> to make <a href=\"#\" class=\"xref\">IBM Plex<\/a>, an im\u00admense fam\u00adily cov\u00ader\u00ading sans, serif, and\u00a0mono.<\/p>\n<\/li>\n<li>\n<p>Since I launched <em>Prac\u00adti\u00adcal Ty\u00adpog\u00adra\u00adphy<\/em>, I\u2019ve been push\u00ading <a href=\"https:\/\/practicaltypography.com\/charter.html\" class=\"xref\">Char\u00adter<\/a>, a 1987 Matthew Carter de\u00adsign that holds up beau\u00adti\u00adfully on to\u00adday\u2019s screens.<\/p>\n<\/li>\n<\/ul><p>The fonts above are won\u00adder\u00adful! And won\u00adder\u00adfully free! Whether my cur\u00admud\u00adgeonisms have had an im\u00adpact, I can\u2019t say. But to\u00adday, if your project re\u00adquires free fonts, you have plenty of good op\u00adtions. If you refuse, you have no one to blame but yourself.<\/p><p>As for the gazil\u00adlions of other free fonts\u2014if you\u2019ve got the pa\u00adtience to pan for gold in a <a href=\"https:\/\/practicaltypography.com\/effluents-influence-affluence.html#traffic\">river of crap,<\/a> knock your\u00adself out. But un\u00adless your time has no value, you\u2019ll find it more ef\u00adfi\u00adcient just to buy a nice <a href=\"#\" class=\"xref\">pro\u00adfes\u00adsional font<\/a>.<\/p><div class=\"btw\"><p>by the way<\/p><ul><li>\n<p>I use <a href=\"https:\/\/practicaltypography.com\/charter.html\" class=\"xref\">Char\u00adter<\/a>, <a href=\"#\" class=\"xref\">Fira Sans<\/a>, <a href=\"#\" class=\"xref\">Fira Mono<\/a>, and <a href=\"#\" class=\"xref\">Cooper He\u00adwitt<\/a> ex\u00adten\u00adsively in web\u00adsites for Racket, in\u00adclud\u00ading the <a href=\"https:\/\/docs.racket-lang.org\/\">doc\u00adu\u00admen\u00adta\u00adtion,<\/a> <a href=\"https:\/\/racket-lang.org\/\">home page,<\/a> and <a href=\"https:\/\/blog.racket-lang.org\/\">blog.<\/a> I\u2019ve used <a href=\"#\" class=\"xref\">Source Serif Pro<\/a> for <a href=\"https:\/\/con.racket-lang.org\/2015\/\">spe\u00adcial events.<\/a><\/p>\n<\/li>\n<li>\n<p>Why haven\u2019t I listed any <a href=\"https:\/\/practicaltypography.com\/system-fonts.html\" class=\"xref\">sys\u00adtem fonts<\/a> above? Care\u00adful\u2014they\u2019re not the same thing. Sys\u00adtem fonts are\u00a0only \u201cfree\u201d in the sense that you don\u2019t have to pay ex\u00adtra for them. But they\u2019re not \u201cfree\u201d in the deeper sense of hav\u00ading min\u00adi\u00admal li\u00adcens\u00ading re\u00adstric\u00adtions. On the con\u00adtrary, the fonts that come with Win\u00addows or Mac\u00a0OS are gov\u00aderned by a pro\u00adpri\u00adetary li\u00adcense which per\u00admits cer\u00adtain uses and pro\u00adhibits others.<\/p>\n<\/li>\n<li>\n<p>\u201cBut have you con\u00adsid\u00adered my fa\u00advorite free font, Lu\u00addi\u00adcrous Sans?\u201d If it\u2019s your fa\u00advorite, my opin\u00adion hardly mat\u00adters. In any case, I\u2019m not a re\u00adviewer of free fonts. There are far too many, and I care far too little.<\/p>\n<\/li>\n<\/ul><\/div><\/div>\n \n<p>undock&amp;#xe128 move&amp;#xe118 <a class=\"font-change body-texter\" id=\"equity\">Equity<\/a> <a class=\"font-change body-texter\" id=\"valkyrie\">Valkyrie<\/a> <a class=\"font-change body-texter\" id=\"century-supra\">Century Supra<\/a> <a class=\"font-change body-texter\" id=\"concourse\">Concourse<\/a> <a class=\"font-change body-texter\" id=\"hermes-maia\">Hermes Maia<\/a> <a class=\"font-change body-texter\" id=\"triplicate\">Triplicate<\/a> buy font&amp;#xe085 close&amp;#xe043<\/p>\n<div id=\"toolbar\" class=\"\"><p><a class=\"box-link\" href=\"https:\/\/practicaltypography.com\/system-fonts.html\">\u2190\u00a0system fonts<\/a><\/p><p><a class=\"box-link\" href=\"https:\/\/practicaltypography.com\/\">top<\/a><\/p><p><a class=\"box-link\" href=\"https:\/\/practicaltypography.com\/font-recommendations.html\">chapter<\/a><\/p><p><a class=\"box-link\" href=\"https:\/\/practicaltypography.com\/charter.html\">Charter\u00a0\u2192<\/a><\/p><\/div>","created_at":"2020-02-26T14:38:21+0000","updated_at":"2020-02-26T14:38:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"practicaltypography.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1145,"title":"Modern, functional Common Lisp: Myths and tips","url":"https:\/\/ambrevar.xyz\/modern-common-lisp\/index.html","content":"<div class=\"outline-text-2\" id=\"text-orgb935101\"><p>Myth: Common Lisp is old and does not enjoy the features of modern languages from recent years.<\/p><p>Truth: While the Common Lisp standard is from the 1990s, many modern libraries have continuously updated the language.<\/p><p>One of the biggest selling points of Common Lisp (and Lisp languages in general) is that the language itself is extensible. This means that if new paradigms become fashionable, they can easily be added to the language.<\/p><p>In particular, Common Lisp has the reputation among functional programmers of being very imperative and object-oriented. While these styles were popular in the past, nothing prevents us from programming functionally in Common Lisp.<\/p><p>As with many languages, Common Lisp packages can be <code>use<\/code>-d, which means all symbols will be merged in the current name space, or <code>import<\/code>-ed, which means all symbols will be accessible only via the package prefix.<\/p><p>The benefit of <code>use<\/code> is that importing a language-enhancing package gives us a feel that the enhancements are first class.<\/p><p>The drawback however is that if you change your mind and decide to <code>import<\/code> the package instead, you\u2019ll have to add the package prefix to all the related symbols manually. Plus it\u2019s harder to see from just a glance which symbols comes from which package.<\/p><p>This drawback could be eliminated with sufficient editor support to refactor symbols by their packages. See <a href=\"https:\/\/github.com\/slime\/slime\/issues\/532\">https:\/\/github.com\/slime\/slime\/issues\/532<\/a> for a discussion.<\/p><\/div><div id=\"outline-container-orgd277009\" class=\"outline-3\"><h3 id=\"orgd277009\">Recommended language-enhancing packages<\/h3><div class=\"outline-text-3\" id=\"text-orgd277009\"><p>The following packages are widespread and useful enough that can be <code>use<\/code>-d without hesitation:<\/p><ul class=\"org-ul\"><li><a href=\"https:\/\/common-lisp.net\/project\/alexandria\/\">Alexandria<\/a>: various language and functional programming procedures.<\/li>\n<li>UIOP: Macros and operating-system-level procedures. Included with ASDF in most Common Lisp compilers.<\/li>\n<li><a href=\"https:\/\/github.com\/guicho271828\/trivia\">Trivia<\/a>: Pattern matching.<\/li>\n<\/ul><p>The following libraries are also extremely useful, although not to everyone\u2019s taste:<\/p><ul class=\"org-ul\"><li><a href=\"https:\/\/github.com\/m2ym\/trivial-types\">Trivial-types<\/a>: More common type definitions.<\/li>\n<li><a href=\"https:\/\/github.com\/ruricolist\/serapeum\">Serapeum<\/a>: a huge collection of utilities (named let, local definitions, hooks, file \u201chuman\u201d size, string manipulation, etc.), built as a complement to Alexandria and UIOP. It greatly extends the language.<\/li>\n<li><a href=\"http:\/\/series.sourceforge.net\/\">Series<\/a>: Lazy sequences and supporting higher-order functions.<\/li>\n<li><a href=\"https:\/\/github.com\/EuAndreh\/defclass-std\">defclass-std<\/a>: Saner and more robust defaults for class definitions. In particular, this prevents accidental typos in the accessor or the initarg, and it sets a default value (the <code>initform<\/code>) automatically by default.<\/li>\n<li><a href=\"https:\/\/github.com\/ndantam\/sycamore\">Sycamore<\/a>: Functional data structure library<\/li>\n<\/ul><p>Other language-enhancing libraries:<\/p><ul class=\"org-ul\"><li><a href=\"https:\/\/github.com\/fukamachi\/rove\">Rove<\/a>: The (only?) testing framework with coverage support.<\/li>\n<li><a href=\"https:\/\/edicl.github.io\/cl-ppcre\/\">ppcre<\/a>: Advanced string manipulation with unicode support.<\/li>\n<li><a href=\"https:\/\/github.com\/melisgl\/named-readtables\/\">named-readtables<\/a>: First class name space for readtables.<\/li>\n<li><a href=\"https:\/\/alex-gutev.github.io\/generic-cl\/\">generic-cl<\/a>: Generic function wrapper over various functions in the Common Lisp standard, such as equality predicates and sequence operations. Includes a few <a href=\"https:\/\/alex-gutev.github.io\/generic-cl\/#lazy_sequence_utilities\">lazy sequence utilities<\/a>.<\/li>\n<\/ul><\/div><\/div>","created_at":"2020-02-25T21:54:16+0000","updated_at":"2020-02-25T21:54:16+0000","published_by":["root"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"ambrevar.xyz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1144,"title":"realgud\/realgud","url":"https:\/\/github.com\/realgud\/realgud","content":"<p><a href=\"https:\/\/travis-ci.org\/realgud\/realgud\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/edbab3fdbe9b0932b42375747684ed95861ec1d4\/68747470733a2f2f6170692e7472617669732d63692e6f72672f7265616c6775642f7265616c6775642e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/api.travis-ci.org\/realgud\/realgud.svg?branch=master\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/melpa.org\/#\/realgud\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a2ac7b0895e3423f0be21e0f37192a98421f226d\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f7265616c6775642d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"http:\/\/melpa.org\/packages\/realgud-badge.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/stable.melpa.org\/#\/realgud\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/da8dc5372f3242a35a3b2b1185df22c29d7f23e4\/687474703a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f7265616c6775642d62616467652e737667\" alt=\"MELPA Stable\" data-canonical-src=\"http:\/\/stable.melpa.org\/packages\/realgud-badge.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/realgud\/realgud\/blob\/master\/etc\/realgud-logo.png\"><img src=\"https:\/\/github.com\/realgud\/realgud\/raw\/master\/etc\/realgud-logo.png\" alt=\"RealGUD logo\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><strong>A extensible, modular GNU Emacs front-end for interacting with external debuggers<\/strong>, brought to you by <strong>Rocky Bernstein<\/strong> (<a href=\"https:\/\/github.com\/rocky\">@rocky<\/a>) and <strong>Cl\u00e9ment Pit-Claudel<\/strong> (<a href=\"https:\/\/github.com\/cpitclaudel\">@cpitclaudel<\/a>).<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/realgud\/realgud\/blob\/master\/etc\/realgud.png\"><img src=\"https:\/\/github.com\/realgud\/realgud\/raw\/master\/etc\/realgud.png\" alt=\"RealGUD screenshot\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>You can install RealGUD from <a href=\"http:\/\/melpa.org\/#\/getting-started\" rel=\"nofollow\">MELPA<\/a>, a repository of Emacs packages. If you don't have MELPA set up, add the following to your <code>.emacs<\/code> and restart Emacs:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'package)\n(add-to-list 'package-archives '(\"melpa\" . \"http:\/\/melpa.org\/packages\/\") t)\n(package-initialize)<\/pre><\/div><p>You can then run the following commands to install RealGUD:<\/p><pre>M-x package-refresh-contents RET      (to refresh your package database)\nM-x package-install RET realgud RET   (to install and compile `realgud` and its dependencies)\n<\/pre><p>Alternatively, you can install RealGUD using <a href=\"http:\/\/www.emacswiki.org\/emacs\/el-get\" rel=\"nofollow\">el-get<\/a> or from source directly. See <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\/How-to-Install\">this article<\/a> in our wiki for more info.<\/p><p>Use <code>M-x load-library RET realgud RET<\/code> to load RealGUD.<\/p><h2><a id=\"user-content-launching-the-debugger\" class=\"anchor\" aria-hidden=\"true\" href=\"#launching-the-debugger\"><\/a>Launching the debugger<\/h2><p>Open a source file, then use <code>M-x realgud:&lt;debugger-name&gt;<\/code> to start your favorite debugger (for example, you can use <code>M-x realgud:pdb<\/code> to launch PDB, a Python debugger). RealGUD opens two windows: the <em>command window<\/em> (showing the debugger's REPL), and the <em>source window<\/em>, showing your code (with some debuggers, such as <code>realgud:gdb<\/code>, this window does not appear until you type <code>start<\/code> in the <em>command window<\/em>). An solid arrow <code>\u25b6<\/code> indicates the current line in the source window. Grayed out arrows indicate previous positions.<\/p><h2><a id=\"user-content-using-the-debugger\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-the-debugger\"><\/a>Using the debugger<\/h2><p>All usual debugger commands work in the command window: for example, you can type <code>n RET<\/code> in the command window to step forward one source line. But RealGUD's power lies in the source window: in it, most single keys correspond to a debugger action. For example, you can press <code>n<\/code> in the source window to step forward.<\/p><h2><a id=\"user-content-source-window-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#source-window-commands\"><\/a>Source window commands<\/h2><p>Here is a quick rundown of the most useful commands. \u201c\ud83d\udc01\u201d indicates mouse commands (commands that can be run by clicking on a variable or in the margins). Many of the commands are accessible from the tool bar (<code>tool-bar-mode<\/code>) and the menu (<code>menu-bar-mode<\/code>).<\/p><ul><li>\n<p><strong>Motion commands<\/strong><\/p>\n<table><thead><tr><th>Command<\/th>\n<th>Action<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>n<\/kbd>, <kbd>F10<\/kbd><\/td>\n<td>Next (aka \u201cstep over\u201d, \u201cstep through\u201d)<\/td>\n<\/tr><tr><td><kbd>s<\/kbd>, <kbd>SPC<\/kbd>, <kbd>F11<\/kbd><\/td>\n<td>Step (aka \u201cstep into\u201d)<\/td>\n<\/tr><tr><td><kbd>f<\/kbd>, <kbd>S-F11<\/kbd><\/td>\n<td>Finish (aka \u201cstep out\u201d, \u201creturn\u201d)<\/td>\n<\/tr><tr><td><kbd>c<\/kbd>, <kbd>F5<\/kbd><\/td>\n<td>Continue (run to next break point)<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<li>\n<p><strong>Using breakpoints<\/strong><\/p>\n<table><thead><tr><th>Command<\/th>\n<th>Action<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>b<\/kbd>, <kbd>F9<\/kbd><\/td>\n<td>Set breakpoint \ud83d\udc01<\/td>\n<\/tr><tr><td><kbd>D<\/kbd><\/td>\n<td>Clear breakpoint \ud83d\udc01 (by number)<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<li>\n<p><strong>Inspecting variables<\/strong><\/p>\n<table><thead><tr><th>Command<\/th>\n<th>Action<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>mouse-2<\/kbd> (middle button)<\/td>\n<td>Inspect variable under cursor (in tooltip) \ud83d\udc01<\/td>\n<\/tr><tr><td><kbd>e<\/kbd><\/td>\n<td>Evaluate expression<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<li>\n<p><strong>Control commands<\/strong><\/p>\n<table><thead><tr><th>Command<\/th>\n<th>Action<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>q<\/kbd>, <kbd>S-F5<\/kbd><\/td>\n<td>Quit<\/td>\n<\/tr><tr><td><kbd>R<\/kbd>, <kbd>r<\/kbd><\/td>\n<td>Run (aka \u201crestart\u201d)<\/td>\n<\/tr><tr><td><kbd>S<\/kbd><\/td>\n<td>Go to command window<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<\/ul><p>RealGUD supports <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\/Debuggers-Available\">many external debuggers<\/a>. Help us <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\/How-to-add-a-new-debugger\">support even more<\/a>!<\/p><h2><a id=\"user-content-extra-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#extra-commands\"><\/a>Extra commands<\/h2><p>\u201c\u2699\u201d indicates a work-in-progress (contributions welcome!)<\/p><table><thead><tr><th>Command<\/th>\n<th>Action<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>U<\/kbd><\/td>\n<td>Until (run to a greater source line)<\/td>\n<\/tr><tr><td><kbd>u<\/kbd>, <kbd>&gt;<\/kbd><\/td>\n<td>Up stack (move to older stack frame)<\/td>\n<\/tr><tr><td><kbd>d<\/kbd>, <kbd>&lt;<\/kbd><\/td>\n<td>Down stack (move to younger stack frame)<\/td>\n<\/tr><tr><td><kbd>X<\/kbd><\/td>\n<td>Clear breakpoint (by line)<\/td>\n<\/tr><tr><td><kbd>j<\/kbd><\/td>\n<td>Jump to current line \u2699<\/td>\n<\/tr><tr><td><kbd>-<\/kbd><\/td>\n<td>Disable breakpoint \u2699<\/td>\n<\/tr><tr><td><kbd>+<\/kbd><\/td>\n<td>Enable breakpoint \u2699<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-tracking-an-existing-debugger-process\" class=\"anchor\" aria-hidden=\"true\" href=\"#tracking-an-existing-debugger-process\"><\/a>Tracking an existing debugger process<\/h2><p>Use <code>M-x realgud-track-mode<\/code> inside an existing <a href=\"http:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Shell.html\" rel=\"nofollow\">shell<\/a>, or <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_mono\/eshell.html\" rel=\"nofollow\">eshell<\/a> buffer to track an already-running debugger process.<\/p><h3><a id=\"user-content-realguds-wiki\" class=\"anchor\" aria-hidden=\"true\" href=\"#realguds-wiki\"><\/a>RealGUD's wiki<\/h3><p>Browse the <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\">wiki<\/a> for more information about <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\/How-to-Install\">setting up<\/a>, <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\/How-to-Use\">using realgud<\/a>, <a href=\"https:\/\/github.com\/realgud\/realgud\/wiki\/Features\">exploring features<\/a>, and lots more.<\/p>","created_at":"2020-02-25T18:57:35+0000","updated_at":"2020-02-25T18:57:35+0000","published_by":["realgud"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/19677271?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1143,"title":"djcb\/sauron","url":"https:\/\/github.com\/djcb\/sauron","content":"<h2>what is it?<\/h2>\n<p><strong>sauron<\/strong> is an emacs mode for keeping track of <em>events<\/em> happening in the (emacs) world around you. Events are things like \u2018appointment in 5 minutes\u2019, \u2018bob pinged you on IRC\u2019, \u2018torrent download is complete\u2019 etc. Sauron shows those events like a list \u2013 basically like a log. You can \u2018activate\u2019 an event by either pressing <code>RET<\/code> when point is on it, or clicking it with the middle mouse button (<code>&lt;mouse-2&gt;<\/code>).<\/p>\n<p>When activated, it can execute some arbitrary function \u2013 for example in the case of IRC (ERC), it will switch you to the buffer (channel) it originated from. It\u2019s a bit of a generalization of what <em>tracking mode<\/em> does in ERC (the emacs IRC client), and that is in fact how it started.<\/p>\n<p>There\u2019s an increasing number of hooks and tunables in sauron, which allows you to fine-tune the behavior. However, I strive for it to be useful with minimal configuration.<\/p>\n<h2>getting started<\/h2>\n<p>After you\u2019ve put the various sauron files in a directory, you can enable it with something like the following in your <code>.emacs<\/code>:<\/p>\n<p>;; set load path, obviously replace \u201c&lt;path-to-sauron-dir&gt;\u201d with the actual path\u2026<\/p>\n<pre>(add-to-list 'load-path \"&lt;path-to-sauron-dir&gt;\")\n(require 'sauron)\n<\/pre>\n<p>Now, you can start sauron with\u2026 <code>M-x sauron-start<\/code>, and stop it with <code>M-x sauron-stop<\/code>.<\/p>\n<p><code>sauron-start<\/code> will pop-up a new frame (window) which will show events coming from any of its sources (i.e., ERC, org-mode appointments, D-Bus). You can \u2018activate\u2019 a source by pressing \u201cEnter\u201d with the cursor on the event, which will then take some backend-specific action.<\/p>\n<p>For example, for the ERC-backend, it will transfer you to the buffer (IRC-channel) where said event happened. You can clear all events with <code>M-x sauron-clear<\/code> (default keybinding: <code>c<\/code>).<\/p>\n<p>You can toggle between showing and hiding of the Sauron frame or window using <code>M-x sauron-toggle-hide-show<\/code>.<\/p>\n<p>Sauron (by default) loads the <code>sauron-erc<\/code>, <code>sauron-org<\/code> and <code>sauron-dbus<\/code> modules; if you don\u2019t have ERC, org-mode or d-bus support, these will simply be ignored. If so desired, you can customize <code>sauron-modules<\/code>. See below for some specifics about the backends.<\/p>\n<h2>customization<\/h2>\n<p>I\u2019ve tried hard to come up with reasonable defaults, such that users can get started with sauron without reading too much documentation or having to write elisp etc.; still, I\u2019ve also tried to make sauron very configurable - different people have different needs, so it should be possible to coerce the software in whatever direction.<\/p>\n<p>Below are some customization points.<\/p>\n<h3>sauron look-and-feel<\/h3>\n<p>Sauron can be shown either as a separate frame (the default), or embedded in your current frame. For the latter, set <code>sauron-separate-frame<\/code> to <code>nil<\/code>:<\/p>\n<pre>(setq sauron-separate-frame nil)\n<\/pre>\n<p>Note, this latter option (embedded sauron) is <strong>experimental<\/strong>. Emacs does not make it easy to do this reliable. Note, you can use: <code>M-x sauron-toggle-hide-show<\/code> to hide\/show the sauron frame or window.<\/p>\n<p>You can customize the columns shown in the sauron buffer by setting <code>sauron-column-alist<\/code> - see its documentation.<\/p>\n<p>You can remove the mode-line in the sauron-buffer by setting <code>sauron-hide-mode-line<\/code> to <code>t<\/code>, e.g.:<\/p>\n<pre>(setq sauron-hide-mode-line t)\n<\/pre>\n<p>You can make the Sauron window appear on every (virtual) desktop by setting <code>sauron-sticky-add<\/code> to t, i.e..<\/p>\n<pre>(setq sauron-sticky-frame t)\n<\/pre>\n<p>in your configuration. Depending on your window manager, this may also set the frame to be <em>always-on-top<\/em>. Obviously, this is only effective if you use sauron in a separate frame.<\/p>\n<h3>priorities \u2013 <code>sauron-min-priority<\/code><\/h3>\n<p>Each event in sauron has a certain <em>priority<\/em>. Sauron <strong>ignores<\/strong> all events which have a priority that is lower that <code>sauron-min-priority<\/code> (default value: 3).<\/p>\n<p>For example, all messages written on IRC (i.e., coming from the ERC-backend) which are <strong>not<\/strong> directed towards you have priority 2 \u2013 you will not see them. And that is probably a good idea.<\/p>\n<h3>watching patterns \u2013 <code>sauron-watch-patterns<\/code><\/h3>\n<p>You can specify a list of patterns (regular expressions) which sauron should check. An event matching any of the patterns in the list will have its priority raised by 1 point. If that one point raises it to `sauron-min-priority\u2019 or higher level, it will now show up in the Sauron buffer.<\/p>\n<p><code>sauron-watch-patterns<\/code> is useful if you want to check if, for example, your name, or your hobby project is mentioned in some IRC channel.<\/p>\n<p>So, for example, as part of your settings:<\/p>\n<pre>;; watch for some animals\n(setq sauron-watch-patterns\n  '(\"\\\\bgnu\\\\b\" \"yak\" \"capybara\" \"wombat\"))\n<\/pre>\n<h3>watching nicks \u2013 <code>sauron-watch-nicks<\/code><\/h3>\n<p>You can also specify a list of nicks to watch for; nicks are matched using a string-match (not a regular expression). A nick matching any of the nicks in the list will have its priority raised by 1 point. If that one point raises it to `sauron-min-priority\u2019 or higher level, it will now show up in the Sauron buffer.<\/p>\n<h3>don\u2019t get swamped by a certain nick<\/h3>\n<p>Since you may not want to get too many events from one nick \u2013 and, who knows, accompanying sound effects, pop-ups and what have you, you can set some insensitivity time; events from the same nick during this time will be lowered in priority by one point.<\/p>\n<p>You can set the time period (in seconds) with `sauron-nick-insensitivity\u2019, which defaults to 60 seconds.<\/p>\n<h3>blocking events from showing up \u2013 <code>sauron-event-block-functions<\/code><\/h3>\n<p>We can customize things even more precisely using the <code>sauron-event-block-functions<\/code> hook function. Any event with a priority &gt;= <code>sauron-min-priority<\/code> will be passed to the hook function(s); if any of those functions returns non-nil, the event will be blocked. See the emacs documentation for a general introduction to hook functions, here\u2019s an example:<\/p>\n<pre>(add-hook 'sauron-event-block-functions\n  (lambda (origin prio msg &amp;optional props)\n    (or\n      (string-match \"foo\" msg) ;; ignore events that match 'foo'\n      ;; other matchers\n     )))\n<\/pre>\n<p>Note that the <code>props<\/code> parameter is a backend specific property-list, which allows you e.g. (for the ERC-backend) to get the sender of some ERC message, and block based on that.<\/p>\n<h3>doing stuff based on events \u2013 <code>sauron-event-added-functions<\/code><\/h3>\n<p>After events have been added, another hook is called: <code>sauron-event-added-functions<\/code>.<\/p>\n<p>This is place to add sound effects, notifications and so on. After all, if you get an event for e.g. the org-mode backend that you have a meeting to attend in 5 minutes, simply adding a line in the Sauron-buffer may not be enough.<\/p>\n<p>Instead, you can define a hook function for this.<\/p>\n<p>For doing very sound effects, pop-ups etc., a few convenience functions are provided:<\/p>\n<ul><li><code>sauron-fx-sox<\/code> (play a sound using \u2018sox\u2019)<\/li>\n<li><code>sauron-fx-aplay<\/code> (play a sound using \u2018aplay\u2019)<\/li>\n<li><code>sauron-fx-gnome-osd<\/code> (show some letters on your screen)<\/li>\n<li><code>sauron-fx-zenity<\/code> (pop up a zenity window)<\/li>\n<li><code>sauron-fx-notify<\/code> (trigger a notification using the D-Bus notification daemon)<\/li>\n<\/ul><p>(see the doc-strings for the functions for the details about their parameters).<\/p>\n<p>Now, our hook function could look something like:<\/p>\n<pre>(add-hook 'sauron-event-added-functions\n  (lambda (origin prio msg &amp;optional props)\n    (if (string-match \"ping\" msg)\n      (sauron-fx-sox \"\/usr\/share\/sounds\/ping.wav\")\n      (sauron-fx-sox \"\/usr\/share\/sounds\/something-happened.wav\"))\n    (when (&gt;= prio 4)\n        (sauron-fx-sox \"\/usr\/share\/sounds\/uhoh.wav\")\n        (sauron-fx-gnome-osd msg 10))))\n<\/pre>\n<h3>Seeing <em>all<\/em> events<\/h3>\n<p>Sometimes, you may want to see <em>all<\/em> events instead of filtering them, for example for debugging purposes. For this, there is the variable <code>sauron-log-events<\/code>. If you set it to <code>t<\/code>, <em>all<\/em> events will be shown in a buffer names <code>*Sauron Log*<\/code>. This buffer shows up to <code>sauron-log-buffer-max-lines<\/code> (default: 1000) lines of the last events.<\/p>\n<h2>connecting to alert.el<\/h2>\n<p>John Wiegley\u2019s <a href=\"https:\/\/github.com\/jwiegley\/alert\">alert.el<\/a> has a bit of overlap with sauron; however, I\u2019ve added some wrapper function to make it trivial to feed sauron events into alert. Simply adding:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'sauron-event-added-functions 'sauron-alert-el-adapter)<\/pre><\/div>\n<p>in your setup should do the trick (of course, <code>alert.el<\/code> must be loaded).<\/p>\n<h2>the backend modules<\/h2>\n<p>Currently, 8 backend modules have been implemented:<\/p>\n<ul><li><strong>erc<\/strong> - for ERC, the emacs IRC client<\/li>\n<li><strong>org-mode<\/strong> - for tracking <code>org-mode<\/code> (<code>appt<\/code>) notifications<\/li>\n<li><strong>notifications<\/strong> - for the emacs24+ <code>notifications<\/code> module<\/li>\n<li><strong>d-bus<\/strong> - for <code>dbus<\/code> events<\/li>\n<li><strong>identica<\/strong> - for <code>identica-mode<\/code>, the social-network site<\/li>\n<li><strong>twittering<\/strong> - for <code>twittering-mode<\/code>, the emacs twitter client<\/li>\n<li><strong>jabber<\/strong> - for <code>jabber<\/code>, the IM protocol (XMPP)<\/li>\n<li><strong>elfeed<\/strong> - for <code>elfeed<\/code>, an emacs Atom\/RSS feed reader\n<p>By default, <code>sauron<\/code> tries to load all of them; this should work, even if you don\u2019t have some of these packages (they simply won\u2019t be activated).<\/p>\n<p>If you do not want to load some module, see the variable <code>sauron-modules<\/code>.<\/p>\n<\/li>\n<\/ul><h3>erc<\/h3>\n<p>The ERC module check all IRC PRIVMSG messages, and JOIN\/LEAVE\/QUIT messages. PRIVMSG includes the messages sent to any channel by anyone. These message are given (by default) priority 2, so (by default) they do not show up in your sauron buffer.<\/p>\n<p>However, messages that match one of your <code>sauron-watch-patterns<\/code> or <code>sauron-watch-nicks<\/code> are getting a higher priority, or messages that are private messages directed at you. However, after sending a message, you won\u2019t get notified from the same nick for another 60 seconds (by default \u2013 see <code>sauron-nick-insensitivity<\/code>), so you won\u2019t get e.g. sound effects for each message in a private conversation.<\/p>\n<h3>org-mode \/ appt<\/h3>\n<p>For org-mode, sauron adds functionality to <code>appt-disp-window-function<\/code> (but leaves it intact), so that whenever some event is near, you get a notification with the following priorities:<\/p>\n<ul><li>15 minutes left: priority 3<\/li>\n<li>10 minutes left: priority 3<\/li>\n<li>5 minutes left: priority 4<\/li>\n<li>2 minutes left: priority 5<\/li>\n<\/ul><p>For all other minutes, you\u2019ll get events with priority 2.<\/p>\n<p>Note that you can influence the number of warnings and the time they start by setting the variables <code>appt-display-interval<\/code> and <code>appt-message-warning-time<\/code>, as documented in emacs manual.<\/p>\n<p>You should load org <em>before<\/em> starting sauron, in particular before you set <code>appt-disp-window-function<\/code>, as sauron-org uses that same function (it will preserve the existing functionality though).<\/p>\n<h3>d-bus<\/h3>\n<p>The dbus backend allows you to get events from outside emacs; it listens for two messages, <code>AddUrlEvent<\/code> and <code>AddMsgEvent<\/code>. You can call them like this:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>dbus-send --session --dest=\"org.gnu.Emacs\"      \\\n          --type=method_call                  \\\n\"\/org\/gnu\/Emacs\/Sauron\"                       \\\n\"org.gnu.Emacs.Sauron.AddUrlEvent\"            \\\nstring:shell uint32:3 string:\"Link: Emacs-Fu\" \\\nstring:\"http:\/\/emacs-fu.blogspot.com\"<\/pre><\/div>\n<p>The four parameters are resp. the originator (\u2018shell\u2019), the priority (\u20183\u2019 in the example), a description and a URL. This will show up in the sauron buffer (if the priority is high enough), and if you activate the event (press RET), your browser will visit the link.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>dbus-send --session                            \\\n--dest=\"org.gnu.Emacs\"                    \\\n --type=method_call                           \\\n \"\/org\/gnu\/Emacs\/Sauron\"                     \\\n\"org.gnu.Emacs.Sauron.AddMsgEvent\"           \\\nstring:shell uint32:3 string:\"Hello, world!\"<\/pre><\/div>\n<p>The three parameters are resp. the sender (\u2018shell\u2019), the priority (\u20183\u2019 in the example), and message. This will show up in the sauron buffer (if the priority is high enough).<\/p>\n<p>As an example, you can get a notification when torrent has been completed in \u2018Transmission\u2019. In the torrent-completion script (see Preferences\/ Call-script-when-torrent-is-completed), add something like:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>dbus-send --session                            \\\n--dest=\"org.gnu.Emacs\"                    \\\n--type=method_call                           \\\n \"\/org\/gnu\/Emacs\/Sauron\"                     \\\n\"org.gnu.Emacs.Sauron.AddMsgEvent\"           \\\nstring:Transmission uint32:3 string:\"Torrent completed: $TR_TORRENT_NAME\"<\/pre><\/div>\n<p>You also need to enable the web client support in Transmission - it\u2019s in the \u2018Web\u2019 tab of the preferences dialog.<\/p>\n<p>Note, if you start transmission before you start your session, see `Using D-Bus outside your session\u2019.<\/p>\n<h4>Using D-Bus outside your session<\/h4>\n<p>Note, you normally only use D-Bus (i.e.., the d-bus session bus) when you are in the same <em>session<\/em> \u2013 say, your desktop environment. Thus, it is generally <em>not<\/em> possible to send yourself D-Bus messages from programs outside your session, for example something running from <code>crontab<\/code>.<\/p>\n<p>For this, if you set <code>sauron-dbus-cookie<\/code> to non-nil (before starting sauron), it will drop a file <code>~\/.sauron-dbus<\/code> which contains the D-Bus session bus address (<code>DBUS_SESSION_BUS_ADDRESS<\/code>). Using this address you can, in fact, send messages to sauron from outside your session, by doing something like in the previous examples, but first setting <code>DBUS_SESSION_BUS_ADDRESS<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>DBUS_SESSION_BUS_ADDRESS=\"`cat ~\/.sauron-dbus`\" dbus-send ....<\/pre><\/div>\n<p>We don\u2019t write <code>~\/.sauron-dbus<\/code> as there may be security downsides to this - even though normally other users are not allowed to send to \u2018your\u2019 session bus, even with the cookie, it\u2019s always good to be a bit paranoid.<\/p>\n<h3>notifications<\/h3>\n<p>sauron-notifications tracks notifications sent using `notifications-notify\u2019, which was added in emacs 24. You can use `sauron-notifications-urgency-to-priority-plist\u2019 for the mapping of the \u2018urgency\u2019 field of notification to the sauron\u2019s priority field.<\/p>\n<p>Note, one should be careful when calling `notifications-notify\u2019 from functions listed in the `sauron-event-added-functions\u2019 hook, as to not create some infinite recursion.<\/p>\n<h3>identi.ca<\/h3>\n<p><code>sauron-identica<\/code> shows the number of new dents found by <code>identica-mode<\/code> whenever there is at least one new dent.<\/p>\n<h3>twittering<\/h3>\n<p><code>sauron-twittering<\/code> shows the number of new tweets found by <code>twittering-mode<\/code> whenever there is at least one new tweet.<\/p>\n<p>If <code>twittering-username<\/code> is set, it will also show @-mentions when new tweets arrive.<\/p>\n<h3>jabber<\/h3>\n<p><code>sauron-jabber<\/code> shows events from <code>jabber.el<\/code>, this includes new messages, info messages, presence alerts and lost connections.<\/p>\n<p>The info, presence and connection events get priority 2, so by default you won\u2019t get to see these. The others get priority 3, so those <em>should<\/em> be visible by default.<\/p>\n<h3>elfeed<\/h3>\n<p><code>sauron-elfeed<\/code> show events from elfeed.el, this include new entries in each feed.<\/p>\n<p>By default, all events get priority 2 therefore you won\u2019t get to see these. However, it is possible to configure the priority using the following instruction<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(puthash url priority sauron-elfeed-prio-hash)<\/pre><\/div>\n<h2>adding new modules<\/h2>\n<p>It may be interesting to track other modules as well; this shouldn\u2019t be too hard. Suppose we have a module \u2018foo\u2019:<\/p>\n<ul><li>create \u201csauron-foo.el\u201d, and make sure it\u2019s in the load-path<\/li>\n<li>sauron-foo should implement at least:\n<ul><li><code>sauron-foo-start<\/code> to start the module; this function should return <code>t<\/code> if startup is successful, <code>nil<\/code> otherwise<\/li>\n<li><code>sauron-foo-stop<\/code> to stop the module \/ cleanup etc.<\/li>\n<\/ul><\/li>\n<li>add sauron-foo.el with <code>(provide 'sauron-foo)<\/code><\/li>\n<li>now, add <code>sauron-foo<\/code> to your <code>sauron-modules<\/code><\/li>\n<\/ul><p>Now, to actually make your module useful, you\u2019d want to add some event is something happens. This is done using <code>sauron-add-event<\/code> (see it\u2019s documentation).<\/p>\n<h2>Using <code>sauron<\/code> in other elisp<\/h2>\n<p>If you want to create simple sauron-events from other elisp code, writing a backend modules might be unnecessary; you can simply call the <code>sauron-add-event<\/code> function directly. See its docstring for the details. Example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(sauron-add-event\n   'kitchen           ;; origin\n   3                  ;; priority\n   \"Coffee is ready!\"\n   '(lambda ()        ;; function called when activated\n      (message \"Coffee's ready, get it while it's hot!\"))\n   '(:temperature 80)) ;; arbitrary props passed to\n                       ;; hook functions<\/pre><\/div>\n<p>A typical pattern may also be to switch to the buffer of origin when the event is activated. The <code>sauron-switch-to-marker-or-buffer<\/code> function may be useful there, as it tries to ensure that the buffer is shown in the <em>other<\/em> frame (not the one with Sauron).<\/p>\n<h2>sample configuration<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'sauron)\n ;; note, you add (setq sauron-debug t) to get errors which can debug if\n ;; there's something wrong; normally, we catch such errors, since e.g an error\n ;; in one of the hooks may cause ERC to fail (i.e., the message won't come\n ;; trough).\n(global-set-key (kbd \"C-c s\") 'sauron-toggle-hide-show)\n(global-set-key (kbd \"C-c t\") 'sauron-clear)\n(setq\n  sauron-max-line-length 120\n  ;; uncomment to show sauron in the current frame\n  ;; sauron-separate-frame nil\n  ;; you probably want to add your own nickname to the these patterns\n  sauron-watch-patterns\n  '(\"emacs-fu\" \"emacsfu\" \"wombat\" \"capybara\" \"yak\" \"gnu\" \"\\\\bmu\\\\b\")\n  ;; you probably want to add you own nick here as well\n  sauron-watch-nicks\n  '(\"Tom\" \"Dick\" \"Harry\"))\n;; some sound\/light effects for certain events\n(add-hook 'sauron-event-added-functions\n  (lambda (origin prio msg &amp;optional props)\n    (if (string-match \"ping\" msg)\n      (sauron-fx-sox \"\/usr\/share\/sounds\/ping.wav\"))\n    (cond\n      ((= prio 3) (sauron-fx-sox \"\/usr\/share\/sounds\/pling.wav\"))\n      ((= prio 4) (sauron-fx-sox \"\/usr\/share\/sounds\/plong.wav\"))\n      ((= prio 5)\n        (sauron-fx-sox \"\/usr\/share\/sounds\/alarm.wav\")\n        (sauron-fx-gnome-osd(format \"%S: %s\" origin msg) 5)))))\n;; events to ignore\n(add-hook 'sauron-event-block-functions\n  (lambda (origin prio msg &amp;optional props)\n    (or\n      (string-match \"^*** Users\" msg)))) ;; filter out IRC spam<\/pre><\/div>","created_at":"2020-02-25T14:43:57+0000","updated_at":"2020-02-25T14:46:39+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/199741?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1142,"title":"dylanaraps\/neofetch","url":"https:\/\/github.com\/dylanaraps\/neofetch","content":"<h3 class=\"c4\"><a id=\"\" class=\"anchor\" aria-hidden=\"true\" href=\"#\"><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b9b096d774e2e00aa7300cf3110029116041ddce\/68747470733a2f2f692e696d6775722e636f6d2f5a51493245597a2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/b9b096d774e2e00aa7300cf3110029116041ddce\/68747470733a2f2f692e696d6775722e636f6d2f5a51493245597a2e706e67\" alt=\"logo\" height=\"100px\" data-canonical-src=\"https:\/\/i.imgur.com\/ZQI2EYz.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/h3><p style=\"text-align: center;\">A command-line system information tool written in bash 3.2+<\/p><p style=\"text-align: center;\"><a href=\"https:\/\/github.com\/dylanaraps\/neofetch\/blob\/master\/LICENSE.md\"><img src=\"https:\/\/camo.githubusercontent.com\/890acbdcb87868b382af9a4b1fac507b9659d9bf\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d626c75652e737667\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-MIT-blue.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/github.com\/dylanaraps\/neofetch\/releases\"><img src=\"https:\/\/camo.githubusercontent.com\/f0f83c29e25a1121dabc8432a7cbea3cc9c12023\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f72656c656173652f64796c616e61726170732f6e656f66657463682e737667\" data-canonical-src=\"https:\/\/img.shields.io\/github\/release\/dylanaraps\/neofetch.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/repology.org\/metapackage\/neofetch\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/296e5e84f0340275828d32a675be853baf3a264c\/68747470733a2f2f7265706f6c6f67792e6f72672f62616467652f74696e792d7265706f732f6e656f66657463682e737667\" alt=\"Packaging status\" data-canonical-src=\"https:\/\/repology.org\/badge\/tiny-repos\/neofetch.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/289c5b1c6fe6b77fde4fab691b1f07acd73c996b\/68747470733a2f2f692e696d6775722e636f6d2f47466d433541642e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/289c5b1c6fe6b77fde4fab691b1f07acd73c996b\/68747470733a2f2f692e696d6775722e636f6d2f47466d433541642e706e67\" alt=\"neofetch\" height=\"240px\" data-canonical-src=\"https:\/\/i.imgur.com\/GFmC5Ad.png\" class=\"c3\" referrerpolicy=\"no-referrer\" style=\"float: right;\" \/><\/a><\/p><p>Neofetch is a command-line system information tool written in <code>bash 3.2+<\/code>. Neofetch displays information about your operating system, software and hardware in an aesthetic and visually pleasing way.<\/p><p>The overall purpose of Neofetch is to be used in screen-shots of your system. Neofetch shows the information other people want to see. There are other tools available for proper system statistic\/diagnostics.<\/p><p>The information by default is displayed alongside your operating system's logo. You can further configure Neofetch to instead use an image, a custom ASCII file, your wallpaper or nothing at all.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/baa2dbda5355e2659de7338d3a53a7783ca9071d\/68747470733a2f2f692e696d6775722e636f6d2f6c55726b51424e2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/baa2dbda5355e2659de7338d3a53a7783ca9071d\/68747470733a2f2f692e696d6775722e636f6d2f6c55726b51424e2e706e67\" alt=\"neofetch\" height=\"240px\" data-canonical-src=\"https:\/\/i.imgur.com\/lUrkQBN.png\" class=\"c3\" referrerpolicy=\"no-referrer\" style=\"float: right;\" \/><\/a><\/p><p>You can further configure Neofetch to display exactly what you want it to. Through the use of command-line flags and the configuration file you can change existing information outputs or add your own custom ones.<\/p><p>Neofetch supports almost 150 different operating systems. From Linux to Windows, all the way to more obscure operating systems like Minix, AIX and Haiku. If your favourite operating system is unsupported: Open up an issue and support will be added.<\/p><h3><a id=\"user-content-more-dependencies-installation-wiki\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-dependencies-installation-wiki\"><\/a>More: [<a href=\"https:\/\/github.com\/dylanaraps\/neofetch\/wiki\/Dependencies\">Dependencies<\/a>] [<a href=\"https:\/\/github.com\/dylanaraps\/neofetch\/wiki\/Installation\">Installation<\/a>] [<a href=\"https:\/\/github.com\/dylanaraps\/neofetch\/wiki\">Wiki<\/a>]<\/h3>","created_at":"2020-02-25T14:35:28+0000","updated_at":"2020-02-25T14:35:28+0000","published_by":["dylanaraps"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/6799467?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1141,"title":"Generating unique IDs in a distributed environment at high scale.","url":"https:\/\/www.callicoder.com\/distributed-unique-id-sequence-number-generator\/","content":"<div class=\"featured-image\"><img src=\"https:\/\/www.callicoder.com\/assets\/images\/post\/large\/distributed-unique-id-generation-system.jpg\" alt=\"Generating unique IDs in a distributed environment at high scale.\" referrerpolicy=\"no-referrer\" \/><\/div><p>I was recently working on an application that had a sharded MySQL database. While working on this application, I needed to generate unique IDs that could be used as the primary keys in the tables.<\/p>\n<p>When you\u2019re working with a single MySQL database, you can simply use an auto-increment ID as the primary key, But this won\u2019t work in a sharded MySQL database.<\/p>\n<p>So I looked at various existing solutions for this, and finally wrote a simple 64-bit unique ID generator that was inspired by a similar service by Twitter called <a href=\"https:\/\/github.com\/twitter\/snowflake\/tree\/snowflake-2010\">Twitter snowflake<\/a>.<\/p>\n<p>In this article, I\u2019ll share a simplified version of the unique ID generator that will work for any use-case of generating unique IDs in a distributed environment, not just sharded databases.<\/p>\n<p>I\u2019ll also outline other existing solutions and discuss their pros and cons.<\/p>\n<h2 id=\"existing-solutions\">Existing Solutions<\/h2>\n<h4 id=\"uuid\">UUID<\/h4>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Universally_unique_identifier\">UUIDs<\/a> are 128-bit hexadecimal numbers that are globally unique. The chances of the same UUID getting generated twice is negligible.<\/p>\n<p>The problem with UUIDs is that they are very big in size and don\u2019t index well. When your dataset increases, the index size increases as well and the query performance takes a hit.<\/p>\n<h4 id=\"mongodb-s-objectid\">MongoDB\u2019s ObjectId<\/h4>\n<p><a href=\"https:\/\/docs.mongodb.com\/manual\/reference\/method\/ObjectId\/\">MongoDB\u2019s ObjectIDs<\/a> are 12-byte (96-bit) hexadecimal numbers that are made up of -<\/p>\n<ul><li>a 4-byte epoch timestamp in seconds,<\/li>\n<li>a 3-byte machine identifier,<\/li>\n<li>a 2-byte process id, and<\/li>\n<li>a 3-byte counter, starting with a random value.<\/li>\n<\/ul><p>This is smaller than the earlier 128-bit UUID. But again the size is relatively longer than what we normally have in a single MySQL auto-increment field (a 64-bit bigint value).<\/p>\n<h4 id=\"database-ticket-servers\">Database Ticket Servers<\/h4>\n<p>This approach uses a centralized database server to generate unique incrementing IDs. It\u2019s like a centralized auto-increment. This approach is used by <a href=\"http:\/\/code.flickr.net\/2010\/02\/08\/ticket-servers-distributed-unique-primary-keys-on-the-cheap\/\">Flickr<\/a>.<\/p>\n<p>The problem with this approach is that the ticket server can become a write bottleneck. Moreover, you introduce one more component in your infrastructure that you need to manage and scale.<\/p>\n<h4 id=\"twitter-snowflake\">Twitter Snowflake<\/h4>\n<p><a href=\"https:\/\/github.com\/twitter\/snowflake\/tree\/snowflake-2010\">Twitter snowflake<\/a> is a dedicated network service for generating 64-bit unique IDs at high scale. The IDs generated by this service are roughly time sortable.<\/p>\n<p>The IDs are made up of the following components:<\/p>\n<ul><li>Epoch timestamp in millisecond precision - 41 bits (gives us 69 years with a custom epoch)<\/li>\n<li>Configured machine id - 10 bits (gives us up to 1024 machines)<\/li>\n<li>Sequence number - 12 bits (A local counter per machine that rolls over every 4096)<\/li>\n<\/ul><p>The extra 1 bit is reserved for future purposes. Since the IDs use timestamp as the first component, they are time sortable.<\/p>\n<p>The IDs generated by twitter snowflake fits in 64-bits and are time sortable, which is great. That\u2019s what we want.<\/p>\n<p>But If we use Twitter snowflake, we\u2019ll again be introducing another component in our infrastructure that we need to maintain.<\/p>\n<p>Finally, I wrote a simple sequence generator that generates 64-bit IDs based on the concepts outlined in the Twitter snowflake service.<\/p>\n<p>The IDs generated by this sequence generator are composed of -<\/p>\n<ul><li><strong>Epoch timestamp in milliseconds precision<\/strong> - 41 bits. The maximum timestamp that can be represented using 41 bits is 2<sup>41<\/sup> - 1, or 2199023255551, which comes out to be <code>Wednesday, September 7, 2039 3:47:35.551 PM<\/code>. That gives us 69 years with respect to a custom epoch.<\/li>\n<li><strong>Node ID<\/strong> - 10 bits. This gives us 1024 nodes\/machines.<\/li>\n<li><strong>Local counter per machine<\/strong> - 12 bits. The counter\u2019s max value would be 4095.<\/li>\n<\/ul><p>The remaining 1-bit is the signed bit and it is always set to 0.<\/p>\n<p>Your microservices can use this Sequence Generator to generate IDs independently. This is efficient and fits in the size of a bigint.<\/p>\n<p>Here is the complete program -<\/p>\n<pre class=\"language-java\">import java.net.NetworkInterface;\nimport java.security.SecureRandom;\nimport java.time.Instant;\nimport java.util.Enumeration;\n\/**\n * Distributed Sequence Generator.\n * Inspired by Twitter snowflake: https:\/\/github.com\/twitter\/snowflake\/tree\/snowflake-2010\n *\n * This class should be used as a Singleton.\n * Make sure that you create and reuse a Single instance of SequenceGenerator per node in your distributed system cluster.\n *\/\npublic class SequenceGenerator {\n    private static final int UNUSED_BITS = 1; \/\/ Sign bit, Unused (always set to 0)\n    private static final int EPOCH_BITS = 41;\n    private static final int NODE_ID_BITS = 10;\n    private static final int SEQUENCE_BITS = 12;\n    private static final int maxNodeId = (int)(Math.pow(2, NODE_ID_BITS) - 1);\n    private static final int maxSequence = (int)(Math.pow(2, SEQUENCE_BITS) - 1);\n    \/\/ Custom Epoch (January 1, 2015 Midnight UTC = 2015-01-01T00:00:00Z)\n    private static final long CUSTOM_EPOCH = 1420070400000L;\n    private final int nodeId;\n    private volatile long lastTimestamp = -1L;\n    private volatile long sequence = 0L;\n    \/\/ Create SequenceGenerator with a nodeId\n    public SequenceGenerator(int nodeId) {\n        if(nodeId &lt; 0 || nodeId &gt; maxNodeId) {\n            throw new IllegalArgumentException(String.format(\"NodeId must be between %d and %d\", 0, maxNodeId));\n        }\n        this.nodeId = nodeId;\n    }\n    \/\/ Let SequenceGenerator generate a nodeId\n    public SequenceGenerator() {\n        this.nodeId = createNodeId();\n    }\n    public synchronized long nextId() {\n        long currentTimestamp = timestamp();\n        if(currentTimestamp &lt; lastTimestamp) {\n            throw new IllegalStateException(\"Invalid System Clock!\");\n        }\n        if (currentTimestamp == lastTimestamp) {\n            sequence = (sequence + 1) &amp; maxSequence;\n            if(sequence == 0) {\n                \/\/ Sequence Exhausted, wait till next millisecond.\n                currentTimestamp = waitNextMillis(currentTimestamp);\n            }\n        } else {\n            \/\/ reset sequence to start with zero for the next millisecond\n            sequence = 0;\n        }\n        lastTimestamp = currentTimestamp;\n        long id = currentTimestamp &lt;&lt; (NODE_ID_BITS + SEQUENCE_BITS);\n        id |= (nodeId &lt;&lt; SEQUENCE_BITS);\n        id |= sequence;\n        return id;\n    }\n    \/\/ Get current timestamp in milliseconds, adjust for the custom epoch.\n    private static long timestamp() {\n        return Instant.now().toEpochMilli() - CUSTOM_EPOCH;\n    }\n    \/\/ Block and wait till next millisecond\n    private long waitNextMillis(long currentTimestamp) {\n        while (currentTimestamp == lastTimestamp) {\n            currentTimestamp = timestamp();\n        }\n        return currentTimestamp;\n    }\n    private int createNodeId() {\n        int nodeId;\n        try {\n            StringBuilder sb = new StringBuilder();\n            Enumeration&lt;NetworkInterface&gt; networkInterfaces = NetworkInterface.getNetworkInterfaces();\n            while (networkInterfaces.hasMoreElements()) {\n                NetworkInterface networkInterface = networkInterfaces.nextElement();\n                byte[] mac = networkInterface.getHardwareAddress();\n                if (mac != null) {\n                    for(int i = 0; i &lt; mac.length; i++) {\n                        sb.append(String.format(\"%02X\", mac[i]));\n                    }\n                }\n            }\n            nodeId = sb.toString().hashCode();\n        } catch (Exception ex) {\n            nodeId = (new SecureRandom().nextInt());\n        }\n        nodeId = nodeId &amp; maxNodeId;\n        return nodeId;\n    }\n}\n<\/pre>\n<p>The above generator uses the system\u2019s MAC address to create a unique identifier for the Node. You can also supply a NodeID to the sequence generator. That will guarantee uniqueness.<\/p>\n<p>Let\u2019s now understand how it works. Let\u2019s say it\u2019s <code>June 9, 2018 10:00:00 AM GMT<\/code>. The epoch timestamp for this particular time is <code>1528538400000<\/code>.<\/p>\n<p>First of all, we adjust our timestamp with respect to the custom epoch-<\/p>\n<pre class=\"language-java\">currentTimestamp = 1528538400000 - 1420070400000 \/\/ 108468000000 (Adjust for custom epoch)\n<\/pre>\n<p>Now, the first 41 bits of the ID (after the signed bit) will be filled with the epoch timestamp. Let\u2019s do that using a left-shift -<\/p>\n<pre class=\"language-java\">id = currentTimestamp &lt;&lt; (10 + 12)\n<\/pre>\n<p>Next, we take the configured node ID and fill the next 10 bits with the node ID. Let\u2019s say that the <code>nodeId<\/code> is 786 -<\/p>\n<pre class=\"language-java\">id |= nodeId &lt;&lt; 12\n<\/pre>\n<p>Finally, we fill the last 12 bits with the local counter. Considering the counter\u2019s next value is 3450, i.e. <code>sequence = 3450<\/code>, the final ID is obtained like so -<\/p>\n<pre class=\"language-java\">id |= sequence  \/\/ 454947766275219456\n<\/pre>\n<p>That gives us our final ID.<\/p>\n<h2 id=\"more-learning-resources\">More Learning Resources<\/h2>\n<ul><li>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Universally_unique_identifier\">UUID (Universally unique identifier)<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/docs.mongodb.com\/manual\/reference\/method\/ObjectId\/\">MongoDB ObjectId<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/code.flickr.net\/2010\/02\/08\/ticket-servers-distributed-unique-primary-keys-on-the-cheap\/\">Ticket Servers: Distributed Unique Primary Keys on the Cheap<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/twitter\/snowflake\/tree\/snowflake-2010\">Twitter Snowflake<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/instagram-engineering.com\/sharding-ids-at-instagram-1cf5a71e5a5c?gi=fad4c1de2db1\">Sharding &amp; IDs at Instagram<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/stackoverflow.com\/questions\/2671858\/distributed-sequence-number-generation\/5685869\">Distributed sequence number generation?<\/a><\/p>\n<\/li>\n<\/ul>","created_at":"2020-02-23T03:27:28+0000","updated_at":"2020-02-23T03:27:28+0000","published_at":"2018-06-08T00:00:00+0000","published_by":["Rajeev Singh"],"annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"www.callicoder.com","preview_picture":"https:\/\/www.callicoder.com\/assets\/images\/post\/large\/distributed-unique-id-generation-system.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1140,"title":"How These Things Work \u2013 A book about CS from first principles (2016)","url":"https:\/\/news.ycombinator.com\/item?id=22346349","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\" style=\"border-spacing: 0px; width: 100%;\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D22346349\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"22346349\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"http:\/\/reasonablypolymorphic.com\/book\/preface\" class=\"storylink\">How These Things Work \u2013 A book about CS from first principles (2016)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=reasonablypolymorphic.com\">reasonablypolymorphic.com<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">364 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=falava\" class=\"hnuser\">falava<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=22346349\">7 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=22346349&amp;goto=item%3Fid%3D22346349\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=How%20These%20Things%20Work%20%E2%80%93%20A%20book%20about%20CS%20from%20first%20principles&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=How%20These%20Things%20Work%20%E2%80%93%20A%20book%20about%20CS%20from%20first%20principles\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=22346349&amp;auth=f7d91fdd788594ca86352065befbce06f80252c6\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=22346349\">28\u00a0comments<\/a><\/td>\n<\/tr><tr><td>\n<\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><p><a href=\"https:\/\/www.ycombinator.com\/apply\/\">Applications are open for YC Summer 2020<\/a><\/p>\n<br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"mailto:hn@ycombinator.com\">Support<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2020-02-17T15:42:28+0000","updated_at":"2020-02-17T15:42:28+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1139,"title":"Dell 27 UltraSharp Monitor with PremierColor | UP2716D","url":"https:\/\/www.dell.com\/en-us\/work\/shop\/dell-27-ultrasharp-monitor-with-premiercolor-up2716d\/apd\/210-aggp\/monitors-monitor-accessories","content":"<section id=\"page-title\"><section id=\"hero-container\" class=\"accessories show-arrow xs-bottom-offset-small md-bottom-offset-medium sm-bottom-offset-medium top-offset-mini md-top-offset-small xs-pad-offset-15\"><div class=\"container clearfix row\"><div class=\"col-xs-12 col-sm-5 col-md-5 col-lg-5 accessories-hero-stack sm-top-offset-small snp-hero-productStack hero-flex\"><div class=\"product-stack-row product-stack-inner-row left-padding-zero right-padding-zero bottom-offset-30\" id=\"ProductStackContainer\" data-async-cart=\"true\"><p>Expect vivid, accurate color precision, ultrathin bezels and ideal color coverage to suit your professional needs with the Dell UltraSharp 27 Monitor | UP2716D.<\/p><p><\/p><p><a target=\"_self\" class=\"btn btn-success text-white col-xs-12 bottom-offset-10 whitespace-normal metricsClass\" data-testid=\"addToCartButton\" data-loading-text=\"Please wait...\" href=\"https:\/\/www.dell.com\/en-us\/work\/shop\/serviceselection\/210-aggp\" data-price=\"639.99\" data-productid=\"210-aggp\" data-metrics=\"{&quot;btnname&quot; : &quot;addtocart&quot;, &quot;prod&quot; : &quot;210-aggp&quot;, &quot;qty&quot; : &quot;1&quot;, &quot;price&quot; : &quot;639.99&quot;}\">Add to Cart<\/a><\/p><p>Hide Info<\/p><\/div><\/div><\/div><\/section><section id=\"polaris-pd\"><\/section><\/section>","created_at":"2020-02-17T00:36:10+0000","updated_at":"2020-02-17T00:36:10+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"www.dell.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1138,"title":"#aiart: Enjoy an infinite stream of AI generated Art","url":"https:\/\/art42.net\/","content":"Art42 | Enjoy an infinite stream of AI generated ArtArt42 | Enjoy an infinite stream of AI generated Art\n<noscript>\n<p><strong>We're sorry but art42 doesn't work properly without JavaScript enabled. Please enable it to continue.<\/strong><\/p>\n<\/noscript>\n<div id=\"app\" data-server-rendered=\"true\"><p><a href=\"https:\/\/art42.net\/\" class=\"router-link-exact-active router-link-active c1\">\u25c0 Home<\/a><a href=\"https:\/\/art42.net\/favorites\">\u2605 Favorites (0)<\/a><\/p><div class=\"home\" data-v-249cc29e=\"\"><h5 data-v-249cc29e=\"\"><a href=\"https:\/\/twitter.com\/valentinvieriu\" rel=\"noreferrer\" target=\"_blank\" data-v-249cc29e=\"\">@art42net<\/a><\/h5><\/div><\/div>","created_at":"2020-02-17T00:19:56+0000","updated_at":"2020-02-17T00:19:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"art42.net","preview_picture":"https:\/\/vcloud42.com\/file\/art42-cdn\/cubism\/seed_0000353905.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1137,"title":"Ecosystem","url":"https:\/\/org-roam.readthedocs.io\/en\/develop\/ecosystem\/","content":"<a href=\"https:\/\/github.com\/jethrokuan\/org-roam\/edit\/master\/doc\/ecosystem.md\" title=\"Edit this page\" class=\"md-icon md-content__icon\">\ue3c9<\/a><p>A number of packages work well combined with Org-Roam:<\/p><h2 id=\"deft\">Deft<\/h2><p><a href=\"https:\/\/jblevins.org\/projects\/deft\/\">Deft<\/a> provides a nice interface for browsing and filtering org-roam notes.<\/p><div class=\"highlight\"><pre>(use-package deft\n  :after org\n  :bind\n  (\"C-c n d\" . deft)\n  :custom\n  (deft-recursive t)\n  (deft-use-filter-string-for-filename t)\n  (deft-default-extension \"org\")\n  (deft-directory \"\/path\/to\/org-roam-files\/\")\n  (deft-use-filename-as-title t))\n<\/pre><\/div><p>The Deft interface can slow down quickly when the number of files get huge. <a href=\"https:\/\/github.com\/hasu\/notdeft\">Notdeft<\/a> is a fork of Deft that uses an external search engine and indexer.<\/p><h2 id=\"org-journal\">Org-journal<\/h2><p><a href=\"https:\/\/github.com\/bastibe\/org-journal\">Org-journal<\/a> is a more powerful alternative to the simple function <code>org-roam-today<\/code>. It provides better journaling capabilities, and a nice calendar interface to see all dated entries.<\/p><div class=\"highlight\"><pre>(use-package org-journal\n  :bind\n  (\"C-c n j\" . org-journal-new-entry)\n  :custom\n  (org-journal-date-prefix \"#+TITLE: \")\n  (org-journal-file-format \"%Y-%m-%d.org\")\n  (org-journal-dir \"\/path\/to\/org-roam-files\/\")\n  (org-journal-date-format \"%A, %d %B %Y\"))\n<\/pre><\/div><h2 id=\"note-taking-add-ons\">Note-taking Add-ons<\/h2><p>These are some plugins that make note-taking in Org-mode more enjoyable.<\/p><h3 id=\"org-download\">Org-download<\/h3><p><a href=\"https:\/\/github.com\/abo-abo\/org-download\">Org-download<\/a> lets you screenshot and yank images from the web into your notes:<\/p><p><img alt=\"org-download\" src=\"https:\/\/org-roam.readthedocs.io\/en\/develop\/images\/org-download.gif\" referrerpolicy=\"no-referrer\" \/><\/p><div class=\"highlight\"><pre>(use-package org-download\n  :after org\n  :bind\n  (:map org-mode-map\n        ((\"s-Y\" . org-download-screenshot)\n         (\"s-y\" . org-download-yank))))\n<\/pre><\/div><h3 id=\"mathpixel\">mathpix.el<\/h3><p><a href=\"https:\/\/github.com\/jethrokuan\/mathpix.el\">mathpix.el<\/a> uses <a href=\"https:\/\/mathpix.com\/\">Mathpix's<\/a> API to convert clips into latex equations:<\/p><p><img alt=\"mathpix\" src=\"https:\/\/org-roam.readthedocs.io\/en\/develop\/images\/mathpix.gif\" referrerpolicy=\"no-referrer\" \/><\/p><div class=\"highlight\"><pre>(use-package mathpix.el\n  :straight (:host github :repo \"jethrokuan\/mathpix.el\")\n  :custom ((mathpix-app-id \"app-id\")\n           (mathpix-app-key \"app-key\"))\n  :bind\n  (\"C-x m\" . mathpix-screenshot))\n<\/pre><\/div><h3 id=\"org-noter-interleave\">Org-noter \/ Interleave<\/h3><p><a href=\"https:\/\/github.com\/weirdNox\/org-noter\">Org-noter<\/a> and <a href=\"https:\/\/github.com\/rudolfochrist\/interleave\">Interleave<\/a> are both projects that allow synchronised annotation of documents (PDF, EPUB etc.) within Org-mode.<\/p><h3 id=\"org-ref\">Org-ref<\/h3><p><a href=\"https:\/\/github.com\/jkitchin\/org-ref\">Org-ref<\/a> does citation and bibliography management in Org-mode, and a great tool for scientific notes.<\/p>","created_at":"2020-02-16T05:23:00+0000","updated_at":"2020-02-16T05:23:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"org-roam.readthedocs.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1136,"title":"How To Take Smart Notes With Org-mode \u00b7 Jethro Kuan","url":"https:\/\/blog.jethro.dev\/posts\/how_to_take_smart_notes_org\/","content":"<p>While writing the documentation for Org-roam, it occurred to me that I have a lot to say that doesn\u2019t truly belong in the documentation. I wrote Org-roam because Org-mode simply didn\u2019t have the utilities to enable the kind of note-taking workflow I wanted. I don\u2019t want to be prescriptive about how note-taking should be done, but Org-roam is open enough to enable many note-taking styles, some of which (I feel) will be inferior to others. This post is also in part for users to understand that there are several feature requests that I may consider anti-features, such as renaming notes, having multiple org-roam directories, and having nested folder structures.<\/p><p>This is the workflow I use. Here I explain what I think note-taking should be, and why it should be this way. I implore you (especially users of Org-roam) to read this through. As usual, let me know what you think by <a href=\"mailto:jethrokuan95@gmail.com\">mail<\/a>.<\/p><p>The ideas here are not my own: much of it comes from How To Take Smart Notes by Sonke Ahrens.<\/p><h2 id=\"note-taking-is-not-just-a-precursor-to-writing\">Note-taking Is Not Just a Precursor to Writing<\/h2><blockquote>\n<p>Notes aren\u2019t a record of my thinking process. They are my thinking process. \u2013 Richard Feynman<\/p>\n<\/blockquote><p>In higher education, students are expected to produce papers. What I say doesn\u2019t just apply to students or aspiring writers: the ability to turn thoughts to writing is essential for anyone who wants to improve their learning or writing. One can imagine the process of writing to be broken down into the following steps:<\/p><ol><li>Find topic\/research question<\/li>\n<li>Research\/find literature<\/li>\n<li>Read and take notes<\/li>\n<li>Draw conclusions \/ outline text<\/li>\n<li>Write<\/li>\n<\/ol><p>This process seems natural, but it\u2019s highly linear, and seems to suggest that the process of writing begins by first deciding on a research question. But what about all those lectures and seminars you attend? Books and papers you read? Conversations with friends, colleagues, professors? Surely these ideas floating in your head, or on paper, factor into the written work somehow. Nobody starts from scratch. How do you know what\u2019s a good research question, without relying on prior reading? How do you know what directions to look in for new ideas and insights?<\/p><p>Indeed, the preparation for writing begins much earlier, with note-taking! What if we follow Feynman\u2019s advice, and make note-taking the thinking process?<\/p><p>Suppose you had an awesome second brain, that could surface content up to you where you wanted it. You have a bunch of unrefined ideas sitting around, thoughts you had speaking with friends, or from reading something on the web. Link them up together, and you have a paper. Instead of beginning with the research question, and going top down, we begin with the notes, and build and refine ideas from the bottom up. This also breaks writing down into small, reasonable steps (one note at a time).<\/p><p>This is the driving principle behind the Zettelkasten method, and it brings me to my main point:<\/p><blockquote>\n<p>The primary purpose of note-taking should not be for storing ideas, but for developing them. When we take notes, we should ask: <strong>\u201cIn what context do I want to see this note again?\"<\/strong><\/p>\n<\/blockquote><p>A corollary is that notes can go anywhere you want, they just have to surface where you want them to. This is a crucial shift in mindset. We need to be extremely purposeful with our note-taking. Write down things that matter, refine them, file them for resurfacing later.<\/p><p>How do we do all these in Org-mode? This is where Org-roam comes in.<\/p><h2 id=\"how-i-take-notes\">How I Take Notes<\/h2><p>I take 2 kinds of notes: fleeting notes, and project-related notes.<\/p><p>Fleeting notes are notes that are taken as a reminder of what\u2019s in my head. I take fleeting notes in my daily notes, an idea borrowed from <a href=\"https:\/\/www.roamresearch.com\/\">Roam Research<\/a>. I couple these fleeting notes with my journal, so I use <a href=\"https:\/\/github.com\/bastibe\/org-journal\">org-journal<\/a> to do so:<\/p><figure><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/4b85237e0208325c5bc84494f94da1bbc0b414f6\/4cea0\/ox-hugo\/fleeting_notes.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><p>Fleeting notes get refined into their own note (own file), and are removed from the daily page. These notes are given file-tags, more on this later.<\/p><p>Project notes are basically everything else. Here are some examples of items that get their own project note:<\/p><ul><li>a talk (e.g. <a href=\"https:\/\/braindump.jethro.dev\/talks\/emacs%5Fshould%5Fbe%5Femacs%5Flisp\/\">Emacs Should Be Emacs Lisp - Tom Tromey<\/a>)<\/li>\n<li>a book (e.g. <a href=\"https:\/\/braindump.jethro.dev\/books\/are%5Fwe%5Fsmart%5Fenough%5Fto%5Fknow%5Fhow%5Fsmart%5Fanimals%5Fare\/\">Are We Smart Enough to Know How Smart Animals Are?<\/a>)<\/li>\n<li>a paper (e.g. <a href=\"https:\/\/braindump.jethro.dev\/papers\/dacrema19%5Fprogress%5Fneural%5Frec\/\">Are We Really Making Much Progress? A Worrying Analysis of Recent Neural Reco&amp;mldr;<\/a>)<\/li>\n<li>any topic or thought I\u2019ve refined from my fleeting notes<\/li>\n<\/ul><p>Let\u2019s work through an example. I picked a random link from my to-dos in my org-agenda: <a href=\"https:\/\/news.ycombinator.com\/item?id=22325975\">Ask HN: How do you learn complex, dense technical information? | Hacker News<\/a><\/p><p>I call <code>org-roam-find-file<\/code>, and type in the title of the note: <em>\u201cHN : Learning Complex Technical Information\u201d<\/em>:<\/p><figure><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/837792dd2baf710415ea38d86d0f724a68dd4ebc\/eaea8\/ox-hugo\/org-roam-find-file.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><p>(Note the title in the backlinks buffer hasn\u2019t updated, because I had yet to save the file)<\/p><p>This creates a note with a random filename (not important to my workflow) in the root org-roam folder . Throughout the note-taking process, I ask myself (I bold again to emphasise): <strong>\u201cIn what context do I want to see this note again?\"<\/strong><\/p><p>I begin to insert these contextual links to files using <code>org-roam-insert<\/code>:<\/p><figure><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/9b144b0c0f246ef781a7348e450fdc6cc226cb2f\/a6895\/ox-hugo\/org-roam-insert-filetag.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><p>It doesn\u2019t matter if the links don\u2019t yet exist, Org-roam creates the files for me. Now if I whenever I want to revisit anything about learning, I can head to the <code>Learning<\/code> page and look at the back-links buffer. And it seems I already have material from the <a href=\"https:\/\/www.coursera.org\/learn\/learning-how-to-learn\/\">Learning How To Learn<\/a> course, and <a href=\"https:\/\/www.drmaciver.com\/2019\/05\/how-to-do-hard-things\/\">an article about how to do hard things<\/a>. Making these manual connections reinforce learning.<\/p><figure><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/7cb259de059e590716aa945957f513265f4b134b\/1be96\/ox-hugo\/screenshot2020-02-15_02-02-29_.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><p>I go through the thread and start to note down things that interest me, in my own words. Again, here I create file links liberally, even though I may not yet have content in them.<\/p><p>For example, here I:<\/p><ul><li>see a note of the three-pass technique. I create a new link with <code>org-roam-insert<\/code> for <code>Three Pass Technique<\/code>, and add the paper to my to-do list (I\u2019ve already read this, but I should revisit for post note-taking me\u2019s sake)<\/li>\n<li>I see some talk on spaced repetition. I already have a note for this, which I link to.<\/li>\n<\/ul><p>Here\u2019s what the note ended up looking like. Terrible example (wasn\u2019t really an insightful thread), but I\u2019m in too deep now. This is for illustrative purposes only. I ended up tossing the note out, and keeping the linked articles and papers in my org-agenda.<\/p><figure><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/7f7d2c2cd1027da463e1d0cec1b93c3902ea5377\/59057\/ox-hugo\/screenshot2020-02-15_02-22-12_.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><p>All links are internal links to files within Org-roam. Doing this continuously builds up a dense network of notes. Some useful file-tags to have include file-tags from the <a href=\"https:\/\/praxis.fortelabs.co\/the-p-a-r-a-method-a-universal-system-for-organizing-digital-information-75a9da8bfb37\/\">PARA method<\/a>. It doesn\u2019t matter what file-tags you use, as long as these notes are able to present themselves when you want them to.<\/p><h2 id=\"what-changed\">What Changed?<\/h2><p>First, my notes all reside in the same folder, in a flat hierarchy. Nesting notes <em>artificially<\/em> introduces a hierarchy, which can be extremely crippling. I can rely solely on file links to make connections between notes.<\/p><p>Second, my notes go anywhere they want, and are generally titled based on their source. This is in contrast with putting a note under a topic (e.g. \u201cClassical Mechanics\u201d). Notes and ideas can belong to multiple topics: it\u2019s much easier to just tag \u201ctopic files\u201d (in my example, \u201cLearning\u201d). With this, there\u2019s no need to think about where a note should go, and the perpetual question is (I mention yet again) <strong>\u201cIn what context do I want to see this note again?\"<\/strong>, which is much easier and more meaningful to answer. This also means my notes almost become write-only. The notes repository can grow incredibly big, and that\u2019s okay. Only the relevant notes surface.<\/p><h2 id=\"how-to-use-these-notes\">How to Use These Notes<\/h2><p>Nat Eliason produced a <a href=\"https:\/\/www.youtube.com\/watch?v=RvWic15iXjk\">fantastic video showcasing how he used these notes to outline an article quickly<\/a>. This is showcases essentially what I discussed earlier about putting together ideas from notes to produce a written work. If I haven\u2019t convinced you that this note-taking style is extremely empowering, this 15 minute video will.<\/p><p>This article was also written in about an hour, in similar style. I first looked at all my notes from relevant tags:<\/p><ul><li><a href=\"https:\/\/braindump.jethro.dev\/zettels\/writing%5Fwith%5Fzettekasten\/\">https:\/\/braindump.jethro.dev\/zettels\/writing%5Fwith%5Fzettekasten\/<\/a><\/li>\n<li><a href=\"https:\/\/braindump.jethro.dev\/zettels\/how%5Fto%5Ftake%5Fsmart%5Fnotes\/\">https:\/\/braindump.jethro.dev\/zettels\/how%5Fto%5Ftake%5Fsmart%5Fnotes\/<\/a><\/li>\n<li><a href=\"https:\/\/braindump.jethro.dev\/zettels\/zettelkasten\/\">https:\/\/braindump.jethro.dev\/zettels\/zettelkasten\/<\/a><\/li>\n<li><a href=\"https:\/\/braindump.jethro.dev\/zettels\/roam%5Fresearch\/\">https:\/\/braindump.jethro.dev\/zettels\/roam%5Fresearch\/<\/a><\/li>\n<\/ul><p>And clicked through their links. I simply took a couple of key points and pieced them together!<\/p><p>All of the above had been enabled by simply exploiting bidirectional linking to its fullest extent. What I demonstrated only scratches the surface. There are many amazing Org packages that make note-taking in Org-mode powerful (see <a href=\"https:\/\/org-roam.readthedocs.io\/en\/develop\/ecosystem\/\">Ecosystem - Org-Roam<\/a>), for example the excellent support for LaTeX, bibtex, tables, to-do, and even literate programming.<\/p><p>I hope I made it clear that the note-taking technique came first, and Org-roam was built to enable that. I really encourage you to think about why you\u2019re taking notes, and how you\u2019d like your notes to serve you. I only recently did this introspection, and have found it life-changing.<\/p>","created_at":"2020-02-15T23:47:35+0000","updated_at":"2020-02-15T23:47:35+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"blog.jethro.dev","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1135,"title":"Namespace packages are not supported \u00b7 Issue #122 \u00b7 davidhalter\/jedi","url":"https:\/\/github.com\/davidhalter\/jedi\/issues\/122","content":"Namespace packages are not supported \u00b7 Issue #122 \u00b7 davidhalter\/jedi \u00b7 GitHub\n<div class=\"application-main\" data-commit-hovercards-enabled=\"\"><main id=\"js-repo-pjax-container\" data-pjax-container=\"\"><div class=\"pagehead repohead hx_repohead readability-menu bg-gray-light pb-0 pt-0 pt-lg-3 pb-0\"><div class=\"container-lg mb-4 p-responsive d-none d-lg-flex\"><ul class=\"pagehead-actions flex-shrink-0\"><li>\n<details id=\"funding-links-modal\" class=\"details-reset details-overlay details-overlay-dark d-inline-block float-left\">\n<\/details><\/li>\n<li><a class=\"tooltipped tooltipped-s btn btn-sm btn-with-count\" aria-label=\"You must be signed in to watch a repository\" rel=\"nofollow\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;notification subscription menu watch&quot;,&quot;repository_id&quot;:null,&quot;auth_type&quot;:&quot;LOG_IN&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/davidhalter\/jedi\/issues\/122&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"2d832ba80bf21bd000c2e0d50f6e95a011642ce83f3704ada65507f021855718\" href=\"https:\/\/github.com\/login?return_to=%2Fdavidhalter%2Fjedi\"> Watch<\/a> <\/li>\n<li><a class=\"btn btn-sm btn-with-count tooltipped tooltipped-s\" aria-label=\"You must be signed in to star a repository\" rel=\"nofollow\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;star button&quot;,&quot;repository_id&quot;:3489194,&quot;auth_type&quot;:&quot;LOG_IN&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/davidhalter\/jedi\/issues\/122&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"6917fd98c182349ec2627025110b7dea1991e5e0b281beb9935191ede5309de9\" href=\"https:\/\/github.com\/login?return_to=%2Fdavidhalter%2Fjedi\"> Star<\/a> <\/li>\n<li><a class=\"btn btn-sm btn-with-count tooltipped tooltipped-s\" aria-label=\"You must be signed in to fork a repository\" rel=\"nofollow\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;repo details fork button&quot;,&quot;repository_id&quot;:3489194,&quot;auth_type&quot;:&quot;LOG_IN&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/davidhalter\/jedi\/issues\/122&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"a1b218aa91fbe9395341ebf17ce9065ffda9a032d83b044672375f5b1b0a1f48\" href=\"https:\/\/github.com\/login?return_to=%2Fdavidhalter%2Fjedi\"> Fork<\/a> <\/li>\n<\/ul><\/div><\/div>\n<div class=\"container-lg clearfix new-discussion-timeline p-responsive repository-content js-check-all-container\"><div class=\"signup-prompt-bg rounded-1 js-signup-prompt signup-prompt p-4 text-center mb-4 rounded-1 position-relative\" data-prompt=\"signup\" hidden=\"hidden\"><h3 class=\"pt-2\">Join GitHub today<\/h3><p class=\"col-6 mx-auto\">GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.<\/p><a class=\"btn btn-primary\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;files signup prompt&quot;,&quot;repository_id&quot;:null,&quot;auth_type&quot;:&quot;SIGN_UP&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/davidhalter\/jedi\/issues\/122&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"bab436b9a9ea675c9665f83d41424adcb6669536a48086ac4f445be1f4014d4d\" data-ga-click=\"(Logged out) Sign up prompt, clicked Sign up, text:sign-up\" href=\"https:\/\/github.com\/join?source=prompt-issue-show&amp;source_repo=davidhalter%2Fjedi\">Sign up<\/a><\/div><\/div>\n<\/main><\/div>\n<div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n<details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark hx_rsm\" open=\"open\">\n<\/details>","created_at":"2020-02-15T21:17:19+0000","updated_at":"2020-02-15T21:17:19+0000","published_by":["davidhalter"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/748594?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1134,"title":"Designing the most performant Row Level Security schema in Postgres","url":"https:\/\/medium.com\/@cazzer\/designing-the-most-performant-row-level-security-strategy-in-postgres-a06084f31945","content":"<div><div class=\"fn\"><div class=\"n fo fp fq fr\"><div class=\"o n\"><div><a rel=\"noopener\" href=\"https:\/\/medium.com\/@cazzer?source=post_page-----a06084f31945----------------------\"><img alt=\"Caleb Brewer\" class=\"r fs ft fu\" src=\"https:\/\/miro.medium.com\/fit\/c\/96\/96\/1*nOwrzD7nLr03LlBqlVMLnw.jpeg\" width=\"48\" height=\"48\" referrerpolicy=\"no-referrer\" \/><\/a><\/div><div class=\"fv ak r\"><div class=\"n\"><div><div class=\"fw n o fx\"><a class=\"at au av aw ax ay az ba bb bc gd bf bg bh bi\" rel=\"noopener\" href=\"https:\/\/medium.com\/@cazzer?source=post_page-----a06084f31945----------------------\">Caleb Brewer<\/a><\/div><\/div><\/div><p><a class=\"at au av aw ax ay az ba bb bc gd bf bg bh bi\" rel=\"noopener\" href=\"https:\/\/medium.com\/@cazzer\/designing-the-most-performant-row-level-security-strategy-in-postgres-a06084f31945?source=post_page-----a06084f31945----------------------\">Aug 2, 2018<\/a> \u00b7 8 min read<\/p><\/div><\/div><\/div><\/div><\/div><p id=\"e555\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">In an effort to write less application logic in my services and interfaces, (and out of pure laziness), I am persistently looking for performant strategies to bake access control directly into my schemas. Row Level Security has been around in Postgres <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/release-9-5.html\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">for a while now<\/a>, but we have recently gotten <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/release-10.html#id-1.11.6.9.5.3.5\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">some upgrades to the optimizer<\/a> which make it worth a deeper investigation.<\/p><h2 id=\"01a0\" class=\"hq hr ev bk bj hs ht hu hv hw hx hy hz ia ib ic id\">But why?<\/h2><p id=\"1178\" class=\"hc hd ev bk he b hf ie hh if hj ig hl ih hn ii hp cu\">If you\u2019re not familiar with Row Level Security, the main hypothesis is that we should be able to prevent access to specific <em class=\"ij\">rows<\/em> of data based on a policy. That means our application logic only has to worry about <code class=\"eb ik il im in b\">SELECT * FROM my_table<\/code> and RLS will handle the <code class=\"eb ik il im in b\">WHERE user_id = my_user_id<\/code> part automagically.<\/p><p id=\"efe6\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">To put it another way: our queries should only contain the clauses requested by our interfaces and not the filters and conditions demanded by access control in a multi-tenant data store. Take this example:<\/p><pre class=\"io ip iq ir is it iu iv\">SELECT id, value<br \/>FROM items<br \/>WHERE created_at &gt; now() - interval '1 week'<br \/>AND items.owner_id = $1<\/pre><p id=\"3bec\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">In the above example the <code class=\"eb ik il im in b\">WHERE<\/code> clause is imposing conditions required for access control, (<code class=\"eb ik il im in b\">items.owner_id<\/code>), and requested by an interface, (<code class=\"eb ik il im in b\">created_at<\/code>). If our interface grows more complex so could the <code class=\"eb ik il im in b\">WHERE<\/code> clause and, all of a sudden, separating concerns becomes much harder. Mistakes will be made, and we will keep writing the same annoying access control logic into our queries.<\/p><p id=\"de0d\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">Enter Row Level Security, which provides a space for us to define the conditions demanded by access control <em class=\"ij\">separately<\/em> from the conditions requested by the interface. All of a sudden, our queries look like this:<\/p><pre class=\"io ip iq ir is it iu iv\">SELECT id, value<br \/>FROM items<br \/>WHERE created_at &gt; now() - interval '1 week'<\/pre><p id=\"0930\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">And, all of a sudden, then can grow as complex as needed without impacting the access control side of the problem.<\/p><h2 id=\"136b\" class=\"hq hr ev bk bj hs ht hu hv hw hx hy hz ia ib ic id\">Requirements<\/h2><p id=\"de4c\" class=\"hc hd ev bk he b hf ie hh if hj ig hl ih hn ii hp cu\">Since the goal is to produce something usable for <em class=\"ij\">real<\/em> applications, let\u2019s talk requirements:<\/p><ol class=\"\"><li id=\"2772\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp iz ja jb\">Items are inherently private<\/li><li id=\"7eec\" class=\"hc hd ev bk he b hf jc hh jd hj je hl jf hn jg hp iz ja jb\">Items can be shared with individuals<\/li><li id=\"0736\" class=\"hc hd ev bk he b hf jc hh jd hj je hl jf hn jg hp iz ja jb\">Items can be shared with groups<\/li><li id=\"2f01\" class=\"hc hd ev bk he b hf jc hh jd hj je hl jf hn jg hp iz ja jb\">Items can be made public<\/li><li id=\"27b0\" class=\"hc hd ev bk he b hf jc hh jd hj je hl jf hn jg hp iz ja jb\">Read and write permissions can be handled separately<\/li><\/ol><p id=\"474c\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">It should also go without saying that we want our policies to evaluate quickly for large datasets; otherwise, our user experience will suck and no one will like us. I\u2019ll use a couple of three letter initialisms throughout this post:<\/p><ul class=\"\"><li id=\"ebcc\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp jh ja jb\">RLS: <a href=\"https:\/\/www.postgresql.org\/docs\/10\/static\/ddl-rowsecurity.html\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">Row Level Security<\/a><\/li><li id=\"fe47\" class=\"hc hd ev bk he b hf jc hh jd hj je hl jf hn jg hp jh ja jb\">ACL: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Access_control_list\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">Access Control List<\/a><\/li><\/ul><p id=\"2eb9\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">I will also use <code class=\"eb ik il im in b\">EXPLAIN ANALYZE<\/code> a lot, which I suggest <a href=\"https:\/\/robots.thoughtbot.com\/reading-an-explain-analyze-query-plan\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">reading up on<\/a> if you\u2019re not familiar with it.<\/p><h2 id=\"65a0\" class=\"hq hr ev bk bj hs ht hu hv hw hx hy hz ia ib ic id\">Produce these results locally!<\/h2><p id=\"f91a\" class=\"hc hd ev bk he b hf ie hh if hj ig hl ih hn ii hp cu\">Everything to reproduce these tests is available <a href=\"https:\/\/github.com\/cazzer\/row-level-security-performance\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">on GitHub<\/a>, but I\u2019ll just copy the gists on here to keep things simple. Each of the schemas is generated with a dataset of 100 users who can <em class=\"ij\">write<\/em> to 1000 items and <em class=\"ij\">read<\/em> from a subset of the rest.<\/p><p id=\"34eb\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">In general, you can test any of these schemas the following way:<\/p><pre class=\"io ip iq ir is it iu iv\">BEGIN;<br \/>SET LOCAL jwt.claims.roles = \u2018c94a743e-8ffd-42dc-a237-b1613e111f53,be95031e-76e0\u2013401a-808c-5c7acde292de\u2019;<br \/>-- this represents an example query from an interface<br \/>SELECT count(*) FROM items;<br \/>COMMIT;<\/pre><p id=\"8772\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">Note that we are using <code class=\"eb ik il im in b\">jwt.claims.roles<\/code> instead of an existing local variable, such as <code class=\"eb ik il im in b\">current_user<\/code>, because <a href=\"https:\/\/stackoverflow.com\/a\/33863371\/1294804\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">those behave unexpectedly when querying views<\/a> and we want this RLS strategy to work for everything, (tables and views). You can also just <code class=\"eb ik il im in b\">SELECT * FROM user_item_stats()<\/code>, which grabs a random user ID as the <code class=\"eb ik il im in b\">role<\/code> and executes the previously mentioned transaction.<\/p><p id=\"b9bd\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">In a real application, you would swap out the <code class=\"eb ik il im in b\">SELECT<\/code> statement for whatever interesting query your user wants to run.<\/p><p id=\"c097\" class=\"hc hd ev bk he b hf ie hh if hj ig hl ih hn ii hp cu\">This is likely to be the most performant approach to RLS since we won\u2019t need any <code class=\"eb ik il im in b\">SELECT<\/code>s in our policy; permissions are stored directly on the items. The biggest drawback about storing an ACL in a column is that it will get very wide for items shared with many users\/groups, and that may impact performance.<\/p><h2 id=\"776c\" class=\"hq hr ev bk bj hs ht hu hv hw hx hy hz ia ib ic id\">Item Table<\/h2><p id=\"aae4\" class=\"hc hd ev bk he b hf ie hh if hj ig hl ih hn ii hp cu\">We\u2019ll use the following schema to define our table:<\/p><figure class=\"io ip iq ir is cz\"><p id=\"36f7\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">There are two key things which we\u2019ve implemented in the table definition:<\/p><ol class=\"\"><li id=\"12e9\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp iz ja jb\">Read and write permissions are handled differently. We could add an <code class=\"eb ik il im in b\">owners<\/code> column if we needed that as well.<\/li><li id=\"7e68\" class=\"hc hd ev bk he b hf jc hh jd hj je hl jf hn jg hp iz ja jb\">We are using a <a href=\"https:\/\/www.postgresql.org\/docs\/10\/static\/gin-intro.html\" class=\"at cg eo ep eq er\" target=\"_blank\" rel=\"noopener nofollow\">GIN index<\/a> for the ACL columns to ensure we maintain good performance on the arrays of permissions.<\/li><\/ol><h2 id=\"547f\" class=\"hq hr ev bk bj hs ht hu hv hw hx hy hz ia ib ic id\">Item Policy<\/h2><figure class=\"io ip iq ir is cz\"><p id=\"82b2\" class=\"hc hd ev bk he b hf hg hh hi hj hk hl hm hn ho hp cu\">This policy is pretty straightforward: we are just using <code class=\"eb ik il im in b\">&amp;&amp;<\/code> to check if the claim roles overlap with the ACL. Note that we aren\u2019t doing anything special for <em class=\"ij\">public<\/em> items here; for this schema you\u2019ll have to create a group to represent <em class=\"ij\">public<\/em> instead of a special column as we will do for the next schema. Surprisingly, the planner uses a sequential scan at the deepest node just for adding an <code class=\"eb ik il im in b\">OR items.public = TRUE <\/code>condition to the <code class=\"eb ik il im in b\">USING<\/code> clause. That causes a huge performance hit:<\/p><\/figure><\/figure>","created_at":"2020-02-15T02:41:17+0000","updated_at":"2020-02-15T02:41:17+0000","published_at":"2018-08-02T13:01:01+0000","published_by":["Caleb Brewer"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"medium.com","preview_picture":"https:\/\/miro.medium.com\/max\/1200\/0*Qcj6lcj1lhNZt7Gs","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1133,"title":"clojure-emacs\/sayid","url":"https:\/\/github.com\/clojure-emacs\/sayid","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/clojure-emacs\/sayid\/blob\/master\/sayid-logo.png\"><img src=\"https:\/\/github.com\/clojure-emacs\/sayid\/raw\/master\/sayid-logo.png\" alt=\"Sayid logo\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><hr \/><p><a href=\"https:\/\/circleci.com\/gh\/clojure-emacs\/sayid\/tree\/master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/94c4564a180c24240f731e02849429ba40823ce5\/68747470733a2f2f636972636c6563692e636f6d2f67682f636c6f6a7572652d656d6163732f73617969642f747265652f6d61737465722e7376673f7374796c653d737667\" alt=\"CircleCI\" data-canonical-src=\"https:\/\/circleci.com\/gh\/clojure-emacs\/sayid\/tree\/master.svg?style=svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/clojars.org\/com.billpiel\/sayid\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e7fde83d1a95daa57f322c30c79bc13311ae0ed3\/68747470733a2f2f696d672e736869656c64732e696f2f636c6f6a6172732f762f636f6d2e62696c6c7069656c2f73617969642e737667\" alt=\"Clojars Project\" data-canonical-src=\"https:\/\/img.shields.io\/clojars\/v\/com.billpiel\/sayid.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/cljdoc.org\/d\/com.billpiel\/sayid\/CURRENT\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/185ad3dec26e4ee65f31e9ce6401a1999fb32c59\/68747470733a2f2f636c6a646f632e6f72672f62616467652f636f6d2f62696c6c7069656c2f7361796964\" alt=\"cljdoc badge\" data-canonical-src=\"https:\/\/cljdoc.org\/badge\/com\/billpiel\/sayid\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>Sayid <em>(siy EED)<\/em> is an omniscient debugger and profiler for Clojure. It extracts secrets from code at run-time.<\/p><p>Sayid works by intercepting and recording the inputs and outputs of functions. It can even record function calls that occur inside of functions. The user can select which functions to trace. Functions can be selected individually or by namespace. The recorded data can be displayed, queried and profiled.<\/p><p>Sayid currently has three components:<\/p><ul><li><code>sayid.core<\/code> and its supporting namespaces<\/li>\n<li><a href=\"https:\/\/nrepl.org\" rel=\"nofollow\">nREPL<\/a> middleware<\/li>\n<li>A <a href=\"https:\/\/cider.mx\" rel=\"nofollow\">CIDER<\/a> plugin<\/li>\n<\/ul><p>The <code>sayid.core<\/code> namespace is designed to be used directly via a repl and does not require Emacs or CIDER. <strong>BUT<\/strong> the CIDER integration offers a far better experience, so it is the current focus of this page and my development efforts.<\/p><p><strong>We're looking for more maintainers for the project. If you're interested in helping out please ping @bbatsov.<\/strong><\/p><h2><a id=\"user-content-table-of-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-contents\"><\/a>Table of Contents<\/h2><ul><li><a href=\"#installation--requirements\">Installation &amp; Requirements<\/a><\/li>\n<li><a href=\"#using-sayid\">Using Sayid<\/a><\/li>\n<li><a href=\"#demos\">Demos<\/a><\/li>\n<\/ul><h2><a id=\"user-content-installation--requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation--requirements\"><\/a>Installation &amp; Requirements<\/h2><h3><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h3><p>Basic usage requires Clojure 1.7 and the optional nREPL middleware requires nREPL 0.4+.<\/p><p>nREPL-powered editor plugins are encouraged to make use of the bundled middleware that provides a very flexible Sayid API.<\/p><h3><a id=\"user-content-leiningen\" class=\"anchor\" aria-hidden=\"true\" href=\"#leiningen\"><\/a>Leiningen<\/h3><p>Add this to the dependencies in your project.clj or lein profiles.clj:<\/p><pre>[com.billpiel\/sayid \"0.0.18\"]\n<\/pre><p>To use the bundled nREPL middleware, you'll want to include Sayid as a plug-in. Here's an example of a bare-bones profiles.clj that works for me:<\/p><div class=\"highlight highlight-source-clojure\"><pre>{:user {:plugins [[com.billpiel\/sayid \"0.0.18\"]]}}<\/pre><\/div><h3><a id=\"user-content-clojure-cli---depsedn\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure-cli---depsedn\"><\/a>Clojure CLI - deps.edn<\/h3><p>Add a the Sayid dependency to your <code>:deps<\/code> key. Depending on your desired setup, you may want to add it to an optional profile, or your tools.deps config directory (often <code>$HOME\/.clojure<\/code>).<\/p><div class=\"highlight highlight-source-clojure\"><pre>{:deps\n  {com.billpiel\/sayid {:mvn\/version \"0.0.18\"}}}<\/pre><\/div><h3><a id=\"user-content-emacs-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#emacs-integration\"><\/a>Emacs Integration<\/h3><p>CIDER setup also requires that the Emacs package <code>sayid<\/code> is installed. It's available on <a href=\"https:\/\/melpa.milkbox.net\/#\/\" rel=\"nofollow\">MELPA<\/a> and MELPA Stable. Put this code in <code>init.el<\/code>, or somewhere, to load keybindings for clojure-mode buffers.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(eval-after-load 'clojure-mode\n  '(sayid-setup-package))<\/pre><\/div><p>If you use CIDER's jack-in commands, then Sayid automatically adds the Maven dependency when starting a REPL. This means you don't need to manually add the dependency to your <code>project.clj<\/code> or <code>deps.edn<\/code> file.<\/p><p>If you don't use CIDER's jack-in commands, you'll need to add a dependency manually. Here's an example of a bare-bones profiles.clj that works for me:<\/p><div class=\"highlight highlight-source-clojure\"><pre>{:user {:plugins [[cider\/cider-nrepl \"0.21.0\"]\n                  [com.billpiel\/sayid \"0.0.18\"]]\n        :dependencies [[nrepl\/nrepl \"0.6.0\"]]}}<\/pre><\/div><p>Usually you'll want to use the latest versions of <code>cider-nrepl<\/code> and nREPL here.<\/p><h3><a id=\"user-content-other-editors\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-editors\"><\/a>Other Editors<\/h3><p>A 3rd-party vim plugin also exists. See <a href=\"http:\/\/arsenerei.com\/blog\/posts\/2017-02-24-vim-sayid\/\" rel=\"nofollow\">this<\/a> and <a href=\"https:\/\/github.com\/arsenerei\/vim-sayid\">this<\/a>.<\/p><h2><a id=\"user-content-using-sayid\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-sayid\"><\/a>Using Sayid<\/h2><p><strong>Note: This assumes you're using the official CIDER plugin.<\/strong>.<\/p><p>Documentation is a little light at the moment. There are lists of keybindings. Helpfully, they are easily accessible from within emacs. Below are the contents of the various help buffers, as well as instructions on how to pop them up in time of need.<\/p><p>Generated docs are also available for the core namespace <a href=\"https:\/\/github.com\/clojure-emacs\/sayid\/blob\/master\/doc\">here<\/a>.<\/p><p>In a clojure-mode buffer, press <code>C-c s h<\/code> (<code>sayid-show-help<\/code>) to pop up the help buffer.<\/p><pre>C-c s ! -- Disable traces, eval current buffer, enable traces, clear the workspace log\nC-c s e -- Enables traces, evals the expression at point, disables traces, displays results with terse view\nC-c s f -- Queries the active workspace for entries that most closely match the context of the cursor position\nC-c s n -- Applies an inner trace to the function at point, replays workspace, displays results\nC-c s r -- Replays workspace, queries results by context of cursor\nC-c s w -- Shows workspace, using the current view\nC-c s t y -- Prompts for a dir, recursively traces all ns's in that dir and subdirs\nC-c s t p -- Prompts for a pattern (* = wildcare), and applies a trace to all *loaded* ns's whose name matches the patten\nC-c s t b -- Trace the ns in the current buffer\nC-c s t e -- Enable the *existing* (if any) trace of the function at point\nC-c s t E -- Enable all traces\nC-c s t d -- Disable the *existing* (if any) trace of the function at point\nC-c s t D -- Disable all traces\nC-c s t n -- Apply an inner trace to the symbol at point\nC-c s t o -- Apply an outer trace to the symbol at point\nC-c s t r -- Remove existing trace from the symbol at point\nC-c s t K -- Remove all traces\nC-c s c -- Clear the workspace trace log\nC-c s x -- Blow away workspace -- traces and logs\nC-c s s -- Popup buffer showing what it currently traced\nC-c s S -- Popup buffer showing what it currently traced in buffer's ns\nC-c s V s -- Set the view\nC-c s h -- show this help\n<\/pre><p>In the <code>*sayid*<\/code> buffer, press <code>h<\/code> to pop up the help buffer.<\/p><pre>ENTER -- pop to function\nd -- def value to $s\/*\nf -- query for calls to function\nF -- query for calls to function with modifier\ni -- show only this instance\nI -- query for this instance with modifier\nw -- show full workspace trace\nn -- jump to next call node\nN -- apply inner trace and reply workspace\np -- jump to prev call node\nP -- pretty print value\nC -- clear workspace trace log\nv -- toggle view\nV -- set view (see register-view)\nl, backspace -- previous buffer state\nL, S-backspace -- forward buffer state\ng -- generate instance expression and put in kill ring\nh -- help\n<\/pre><p>In the <code>*sayid-traced*<\/code> buffer, press <code>h<\/code> to pop up the help buffer.<\/p><pre>enter -- Drill into ns at point\ne -- Enable trace\nd -- Disable trace\nE -- Enable ALL traces\nD -- Disable ALL traces\ni -- Apply inner trace to func at point\no -- Apply outer trace to func at point\nr -- Remove trace from fun at point\nl, backspace -- go back to trace overview (if in ns view)\n<\/pre><p>In the <code>*sayid-pprint*<\/code> buffer, press <code>h<\/code> to pop up the help buffer.<\/p><pre>ENTER -- show path in mini-buffer\ni -- jump into child node\no -- jump out to parent node\nn -- jump to next sibling node\np -- jump to previous sibling node\n<\/pre><h2><a id=\"user-content-demos\" class=\"anchor\" aria-hidden=\"true\" href=\"#demos\"><\/a>Demos<\/h2><h3><a id=\"user-content-conj-2016-presentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#conj-2016-presentation\"><\/a>Conj 2016 Presentation<\/h3><p>I presented Sayid at the Clojure Conj conference in Austin in 2016.<\/p><h3><a id=\"user-content-demo-1---video\" class=\"anchor\" aria-hidden=\"true\" href=\"#demo-1---video\"><\/a>Demo #1 - Video<\/h3><p>A demo video I recorded after the very first alpha release. You can find the <a href=\"http:\/\/github.com\/bpiel\/contrived-example\">contrived example<\/a> project here.<\/p><h3><a id=\"user-content-demo-1---walkthrough\" class=\"anchor\" aria-hidden=\"true\" href=\"#demo-1---walkthrough\"><\/a>Demo #1 - Walkthrough<\/h3><p>This is a written walkthrough of the same steps illustrated in the demo video above, but with Sayid v0.0.8. You can find the <a href=\"http:\/\/github.com\/bpiel\/contrived-example\">contrived example<\/a> project here.<\/p><p>Below is the code to the test namespace. You can see that we have a vending machine that dispenses tacos for 85 cents. We execute the <code>test1<\/code> function, which inserts 41 cents worth of change and presses the taco button.<\/p><div class=\"highlight highlight-source-clojure\"><pre>(ns contrived-example.core-test\n  (:require [clojure.test :refer :all]\n            [contrived-example.core :as ce]))\n(def test-vending-machine {:inventory {:a1 {:name :taco\n                                            :price 0.85\n                                            :qty 10}}\n                           :coins-inserted []\n                           :coins-returned []\n                           :dispensed nil\n                           :err-msg nil})\n(defn test1 []\n  (-&gt; test-vending-machine\n      (ce\/insert-coin :quarter) ;; 25\n      (ce\/insert-coin :dime)    ;; 35\n      (ce\/insert-coin :nickel)  ;; 40\n      (ce\/insert-coin :penny)   ;; 41 cents\n      (ce\/press-button :a1)))   ;; taco costs 85 cents\n(test1)<\/pre><\/div><p>Let's press some keys to get Sayid going.<\/p><p>eval the namespace <code>C-c C-k<\/code> (probably) (<code>cider-load-buffer<\/code>)<\/p><p>trace the project namespaces [C-c s t p]{.kbd} (<code>sayid-trace-ns-by-pattern<\/code>) then <code>contrived-example.*<\/code><\/p><p>This should pop up. It shows how many functions have been traced in which namespaces. Execute <code>test1<\/code>!<\/p><pre>Traced namespaces:\n  5 \/ 5  contrived-example.core\n  1 \/ 1  contrived-example.core-test\n  8 \/ 8  contrived-example.inner-workings\nTraced functions:\n<\/pre><p>You can't tell yet, but something magical happened. Press <code>C-c s w<\/code> (<code>sayid-get-workspace<\/code>) to get an overview of what has been captured in the Sayid workspace. This monster should appear:<\/p><pre>v contrived-example.core-test\/test1  :13446\n|v contrived-example.core\/insert-coin  :13447\n|^\n|v contrived-example.core\/insert-coin  :13448\n|^\n|v contrived-example.core\/insert-coin  :13449\n|^\n|v contrived-example.core\/insert-coin  :13450\n|^\n|v contrived-example.core\/press-button  :13451\n||v contrived-example.inner-workings\/valid-selection  :13452\n|||v contrived-example.inner-workings\/get-selection  :13453\n|||^\n|||v contrived-example.inner-workings\/calc-coin-value  :13454\n|||^\n||| contrived-example.inner-workings\/valid-selection  :13452\n||^\n||v contrived-example.inner-workings\/process-transaction  :13455\n|||v contrived-example.inner-workings\/get-selection  :13456\n|||^\n|||v contrived-example.inner-workings\/calc-change-to-return  :13457\n||||v contrived-example.inner-workings\/calc-coin-value  :13458\n||||^\n||||v contrived-example.inner-workings\/round-to-pennies  :13459\n||||^\n||||v contrived-example.inner-workings\/calc-change-to-return*  :13460\n|||||v contrived-example.inner-workings\/calc-coin-value  :13461\n|||||^\n|||||v contrived-example.inner-workings\/calc-change-to-return*  :13462\n||||||v contrived-example.inner-workings\/calc-coin-value  :13463\n||||||^\n||||||v contrived-example.inner-workings\/calc-change-to-return*  :13464\n|||||||v contrived-example.inner-workings\/calc-coin-value  :13465\n|||||||^\n|||||||v contrived-example.inner-workings\/calc-change-to-return*  :13466\n||||||||v contrived-example.inner-workings\/calc-coin-value  :13467\n||||||||^\n|||||||| contrived-example.inner-workings\/calc-change-to-return*  :13466\n|||||||^\n|||||||v contrived-example.inner-workings\/calc-change-to-return*  :13468\n||||||||v contrived-example.inner-workings\/calc-coin-value  :13469\n||||||||^\n|||||||| contrived-example.inner-workings\/calc-change-to-return*  :13468\n|||||||^\n|||||||v contrived-example.inner-workings\/calc-change-to-return*  :13470\n||||||||v contrived-example.inner-workings\/calc-coin-value  :13471\n||||||||^\n|||||||| contrived-example.inner-workings\/calc-change-to-return*  :13470\n|||||||^\n||||||| contrived-example.inner-workings\/calc-change-to-return*  :13464\n||||||^\n|||||| contrived-example.inner-workings\/calc-change-to-return*  :13462\n|||||^\n||||| contrived-example.inner-workings\/calc-change-to-return*  :13460\n||||^\n|||| contrived-example.inner-workings\/calc-change-to-return  :13457\n|||^\n||| contrived-example.inner-workings\/process-transaction  :13455\n||^\n|| contrived-example.core\/press-button  :13451\n|^\n| contrived-example.core-test\/test1  :13446\n^\n<\/pre><p>What's the meaning of this? These are all the function calls that were made in the traced namespaced when we execute <code>test1<\/code>.<\/p><p>Let's explore. Get your cursor to the first line of the output and press <code>i<\/code> (<code>sayid-query-id<\/code>).<\/p><pre> v contrived-example.core-test\/test1  :13446\n | returned =&gt;  {:inventory {:a1 {:name :taco :price 0.85 :qty 9}}\n |               :coins-inserted []\n |               :coins-returned [:quarter :quarter :nickel]\n |               :dispensed {:name :taco :price 0.85 :qty 10}\n |               :err-msg nil}\n ^\n<\/pre><p>This shows us the details of that <em><strong>i<\/strong>nstance<\/em> of <code>test1<\/code> being called. We can see that a hash map was returned. Despite us inserting only 41 cents for an 85 cent taco, we see that a taco was dispensed, plus change! That's a BUG.<\/p><p>Hit <code>backspace<\/code> (<code>sayid-buf-back<\/code>). We're back at the overview. Scan the list of functions that are called. Let's assume some programmer's intuition and decide that <code>valid-selection<\/code> is the first place of interest. Get your cursor to that line and press these keys to view the <em><strong>i<\/strong>nstance<\/em> and all of its <em><strong>d<\/strong>escendants<\/em>. <code>I<\/code> <code>d<\/code> <code>ENTER (<\/code>sayid-query-id-w-mode`)<\/p><pre> ||v contrived-example.inner-workings\/valid-selection  :13452\n ||| machine =&gt; {:inventory {:a1 {:name :taco :price 0.85 :qty 10}}\n |||             :coins-inserted [:quarter :dime :nickel :penny]\n |||             :coins-returned []\n |||             :dispensed nil\n |||             :err-msg nil}\n ||| button =&gt; :a1\n ||| returns =&gt;  true\n |||v contrived-example.inner-workings\/get-selection  :13453\n |||| machine =&gt; {:inventory {:a1 {:name :taco :price 0.85 :qty 10}}\n ||||             :coins-inserted [:quarter :dime :nickel :penny]\n ||||             :coins-returned []\n ||||             :dispensed nil\n ||||             :err-msg nil}\n |||| button =&gt; :a1\n |||| returned =&gt;  {:name :taco :price 0.85 :qty 10}\n |||^\n |||v contrived-example.inner-workings\/calc-coin-value  :13454\n |||| coins =&gt; [:quarter :dime :nickel :penny]\n |||| returned =&gt;  1.4\n |||^\n ||| contrived-example.inner-workings\/valid-selection  :13452\n ||| machine =&gt; {:inventory {:a1 {:name :taco :price 0.85 :qty 10}}\n |||             :coins-inserted [:quarter :dime :nickel :penny]\n |||             :coins-returned []\n |||             :dispensed nil\n |||             :err-msg nil}\n ||| button =&gt; :a1\n ||| returned =&gt;  true\n ||^\n<\/pre><p>We can see that <code>valid-selection<\/code> makes calls to <code>get-selection<\/code> and <code>calc-coin-value<\/code>. Looking at the return values, we might notice a problem: <code>calc-coin-value<\/code> receives <code>[:quarter :dime :nickel :penny]<\/code> but returns $1.40 as the value. Let's dig deeper. Press <code>n<\/code> (<code>sayid-buffer-nav-to-next<\/code>) a couple times to get the cursor to the call to <code>calc-coin-value<\/code>. Now press <code>N<\/code> (<code>sayid-buf-replay-with-inner-trace<\/code>) and hold onto your hat.<\/p><pre> ||||v (-&gt;&gt; coins (keep coin-values) (apply +)) =&gt; (apply + (keep coin-values coins))  contrived-example.inner-workings\/calc-coin-value  :13491\n ||||| returns =&gt;  1.4\n |||||v (apply + (keep coin-values coins))  contrived-example.inner-workings\/calc-coin-value  :13492\n |||||| #function[clojure.core\/+]\n |||||| (0.25 0.1 0.05 1)\n |||||| returns =&gt;  1.4\n ||||||v (keep coin-values coins)  contrived-example.inner-workings\/calc-coin-value  :13493\n ||||||| {:quarter 0.25 :dime 0.1 :nickel 0.05 :penny 1}\n ||||||| [:quarter :dime :nickel :penny]\n ||||||| returned =&gt;  (0.25 0.1 0.05 1)\n ||||||^\n |||||| (apply + (keep coin-values coins))  contrived-example.inner-workings\/calc-coin-value  :13492\n |||||| #function[clojure.core\/+]\n |||||| (0.25 0.1 0.05 1)\n |||||| returned =&gt;  1.4\n |||||^\n ||||| (-&gt;&gt; coins (keep coin-values) (apply +)) =&gt; (apply + (keep coin-values coins))  contrived-example.inner-workings\/calc-coin-value  :13491\n ||||| returned =&gt;  1.4\n ||||^\n...truncated...\n<\/pre><p><em>(jump to the top of the buffer)<\/em><\/p><p>What did we do? We applied an <em>inner trace<\/em> to the function <code>calc-coin-value<\/code> and then replayed the call to <code>test1<\/code> that we had captured originally.<\/p><p><strong>An INNER trace?<\/strong> YES! We can see the inputs and output values of each expression in the function. Look at it. Where do things go wrong? It's when we pass a hash map to <code>keep<\/code> that defines a penny as being worth a dollar. Bug located! Press <code>n<\/code> a couple times to get your cursor to that call. Press <code>RET<\/code> to jump to that line of code.<\/p><div class=\"highlight highlight-source-clojure\"><pre> (ns contrived-example.inner-workings)\n (def coin-values\n   {:quarter 0.25\n    :dime 0.10\n    :nickel 0.05\n    :penny 1})\n (defn- calc-coin-value\n   [coins]\n   (-&gt;&gt; coins\n        (keep coin-values)\n        (apply +)))\n...truncated...<\/pre><\/div><p>We now find ourselves at the troublesome call to <code>keep<\/code> causing our bug. The hash map, <code>coin-values<\/code>, is just above. Change the value of a penny from <code>1<\/code> to <code>0.01<\/code>. Let's eval our corrected code the Sayid way -- press <code>C-c s !<\/code> (<code>sayid-load-enable-clear<\/code>). This will remove the traces, eval the buffer, then re-apply the traces. It also clears the workspace log. This is all helpful. Navigate back to <code>core-test<\/code> and run <code>test1<\/code> again. Repeating steps above, you can verify the output is now correct: no taco.<\/p><pre> v contrived-example.core-test\/test1  :13579\n | returned =&gt;  {:inventory {:a1 {:name :taco :price 0.85 :qty 10}}\n |               :coins-inserted [:quarter :dime :nickel :penny]\n |               :coins-returned []\n |               :dispensed nil\n |               :err-msg true}\n ^\n<\/pre><p>Great work!<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Distributed under the Apache 2.0 License. See <a href=\"https:\/\/github.com\/clojure-emacs\/sayid\/blob\/master\/LICENSE\">LICENSE<\/a> for details.<\/p>","created_at":"2020-02-14T01:59:38+0000","updated_at":"2020-02-14T01:59:38+0000","published_by":["clojure-emacs"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/5080083?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1132,"title":"hyperium\/hyper","url":"https:\/\/github.com\/hyperium\/hyper","content":"<p><a href=\"https:\/\/crates.io\/crates\/hyper\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/406b70cb1f77e56f32f35e259d529ec2ebb1f532\/68747470733a2f2f6d6572697462616467652e6865726f6b756170702e636f6d2f6879706572\" alt=\"crates.io\" data-canonical-src=\"https:\/\/meritbadge.herokuapp.com\/hyper\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/docs.rs\/hyper\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/0a24de560004cf7b2ce77faad4671056db515f05\/68747470733a2f2f646f63732e72732f68797065722f62616467652e737667\" alt=\"Released API docs\" data-canonical-src=\"https:\/\/docs.rs\/hyper\/badge.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/github.com\/hyperium\/hyper\/blob\/master\/LICENSE\"><img src=\"https:\/\/camo.githubusercontent.com\/890acbdcb87868b382af9a4b1fac507b9659d9bf\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d626c75652e737667\" alt=\"MIT licensed\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-MIT-blue.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/github.com\/hyperium\/hyper\/actions?query=workflow%3ACI\"><img src=\"https:\/\/github.com\/hyperium\/hyper\/workflows\/CI\/badge.svg\" alt=\"CI\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>A <strong>fast<\/strong> and <strong>correct<\/strong> HTTP implementation for Rust.<\/p><ul><li>HTTP\/1 and HTTP\/2<\/li>\n<li>Asynchronous design<\/li>\n<li>Leading in performance<\/li>\n<li>Tested and <strong>correct<\/strong><\/li>\n<li>Extensive production use<\/li>\n<li>Client and Server APIs<\/li>\n<\/ul><p><strong>Get started<\/strong> by looking over the <a href=\"https:\/\/hyper.rs\/guides\" rel=\"nofollow\">guides<\/a>.<\/p><h2><a id=\"user-content-low-level\" class=\"anchor\" aria-hidden=\"true\" href=\"#low-level\"><\/a>\"Low-level\"<\/h2><p>hyper is a relatively low-level library, meant to be a building block for libraries and applications.<\/p><p>If you are looking for a convenient HTTP client, then you may wish to consider <a href=\"https:\/\/github.com\/seanmonstar\/reqwest\">reqwest<\/a>. If you are looking for a convenient HTTP server, then you may wish to consider <a href=\"https:\/\/github.com\/seanmonstar\/warp\">warp<\/a>. Both are built on top of this library.<\/p><h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2><p>To get involved, take a look at <a href=\"https:\/\/github.com\/hyperium\/hyper\/blob\/master\/CONTRIBUTING.md\">CONTRIBUTING<\/a>.<\/p><p>There are two main avenues for real-time chatting about hyper: a <a href=\"https:\/\/gitter.im\/hyperium\/hyper\" rel=\"nofollow\">Gitter room<\/a> and <a href=\"https:\/\/kiwiirc.com\/nextclient\/irc.mozilla.org\/#hyper\" rel=\"nofollow\">irc.mozilla.org\/hyper<\/a>. They are mirrored, so choose whichever format you prefer.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>hyper is provided under the MIT license. See <a href=\"https:\/\/github.com\/hyperium\/hyper\/blob\/master\/LICENSE\">LICENSE<\/a>.<\/p>","created_at":"2020-02-13T17:11:14+0000","updated_at":"2020-02-13T17:11:14+0000","published_by":["hyperium"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/8730506?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1131,"title":"p-e-w\/plotinus","url":"https:\/\/github.com\/p-e-w\/plotinus","content":"<blockquote>\n<p>Only a compound can be beautiful, never anything devoid of parts; and only a whole;<br \/>the several parts will have beauty, not in themselves,<br \/>but only as working together to give a comely total.<br \/>Yet beauty in an aggregate demands beauty in details:<br \/>it cannot be constructed out of ugliness; its law must run throughout.<\/p>\n<\/blockquote><p>\u2013 <a href=\"https:\/\/en.wikipedia.org\/wiki\/Plotinus\" rel=\"nofollow\">Plotinus<\/a>, <em>First Ennead<\/em><\/p><h3 class=\"c3\"><a id=\"user-content-a-searchable-command-palette-in-every-modern-gtk-application\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-searchable-command-palette-in-every-modern-gtk-application\"><\/a>A searchable command palette in every modern GTK+ application<\/h3><p>Have you used Sublime Text's or Atom's \"Command Palette\"? It's a list of everything those editors can do that opens at the press of a key and finds the action you are looking for just by typing a few letters. It's raw power at your fingertips.<\/p><p>Plotinus brings that power <em><strong>to every application on your system<\/strong><\/em> (that is, to those that use the GTK+ 3 toolkit). It automatically extracts all available commands by introspecting a running application, instantly adapting to UI changes and showing only relevant actions. Using Plotinus requires <em>no modifications<\/em> to the application itself!<\/p><p>Just press <kbd>Ctrl+Shift+P<\/kbd> (<a href=\"#configuration\">configurable<\/a>) and you're in business \u2013 it feels so natural you'll soon wonder how you ever lived without it.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/cloud.githubusercontent.com\/assets\/2702526\/20246717\/454a1a9a-a9e3-11e6-8b19-4db092348793.gif\"><img src=\"https:\/\/cloud.githubusercontent.com\/assets\/2702526\/20246717\/454a1a9a-a9e3-11e6-8b19-4db092348793.gif\" alt=\"Nautilus screencast\" class=\"c4\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/cloud.githubusercontent.com\/assets\/2702526\/20246718\/5397bed6-a9e3-11e6-8023-aa9a318820e3.gif\"><img src=\"https:\/\/cloud.githubusercontent.com\/assets\/2702526\/20246718\/5397bed6-a9e3-11e6-8023-aa9a318820e3.gif\" alt=\"gedit screencast\" class=\"c4\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><h3><a id=\"user-content-prerequisites\" class=\"anchor\" aria-hidden=\"true\" href=\"#prerequisites\"><\/a>Prerequisites<\/h3><p>To build Plotinus from source, you need Git, CMake, Vala, and the GTK+ 3 development files. All of these are easily obtained on most modern Linux distributions:<\/p><h4><a id=\"user-content-fedora--rhel--etc\" class=\"anchor\" aria-hidden=\"true\" href=\"#fedora--rhel--etc\"><\/a>Fedora \/ RHEL \/ etc.<\/h4><pre>sudo dnf install git cmake vala gtk3-devel\n<\/pre><h4><a id=\"user-content-ubuntu--mint--elementary--etc\" class=\"anchor\" aria-hidden=\"true\" href=\"#ubuntu--mint--elementary--etc\"><\/a>Ubuntu \/ Mint \/ Elementary \/ etc.<\/h4><pre>sudo apt-get install git cmake valac libgtk-3-dev\n<\/pre><h3><a id=\"user-content-building\" class=\"anchor\" aria-hidden=\"true\" href=\"#building\"><\/a>Building<\/h3><pre>git clone https:\/\/github.com\/p-e-w\/plotinus.git\ncd plotinus\nmkdir build\ncd build\ncmake ..\nmake\nsudo make install\n<\/pre><h3><a id=\"user-content-enabling-plotinus-in-applications\" class=\"anchor\" aria-hidden=\"true\" href=\"#enabling-plotinus-in-applications\"><\/a>Enabling Plotinus in applications<\/h3><p>Because of the complexity and clumsiness surrounding Linux environment variables, Plotinus is currently not enabled automatically. The easiest way to enable Plotinus for all applications on the system is to add the line<\/p><pre>GTK3_MODULES=[libpath]\n<\/pre><p>to <code>\/etc\/environment<\/code>, where <code>[libpath]<\/code> is the <em>full, absolute<\/em> path of <code>libplotinus.so<\/code>, which can be found using the command<\/p><pre>whereis -b libplotinus\n<\/pre><p>Alternatively, you can try Plotinus with individual applications by running them with<\/p><pre>GTK3_MODULES=[libpath] application\n<\/pre><p>from a terminal.<\/p><h2><a id=\"user-content-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration\"><\/a>Configuration<\/h2><p>Plotinus can be configured both globally and per application. Application settings take precedence over global settings. In the commands below, <code>[application]<\/code> can be either<\/p><ul><li><code>default<\/code>, in which case the setting is applied globally, or<\/li>\n<li>the path of an application executable, without the leading slash and with all other slashes replaced by periods (e.g. <code>\/usr\/bin\/gedit<\/code> -&gt; <code>usr.bin.gedit<\/code>).<\/li>\n<\/ul><p>Note that the relevant path is the path of the <em>process executable<\/em>, which is not always identical to the executable being launched. For example, all GNOME JavaScript applications run the process <code>\/usr\/bin\/gjs<\/code>.<\/p><h3><a id=\"user-content-enablingdisabling-the-command-palette\" class=\"anchor\" aria-hidden=\"true\" href=\"#enablingdisabling-the-command-palette\"><\/a>Enabling\/disabling the command palette<\/h3><pre>gsettings set com.worldwidemann.plotinus:\/com\/worldwidemann\/plotinus\/[application]\/ enabled [true\/false]\n<\/pre><h3><a id=\"user-content-changing-the-keyboard-shortcut\" class=\"anchor\" aria-hidden=\"true\" href=\"#changing-the-keyboard-shortcut\"><\/a>Changing the keyboard shortcut<\/h3><pre>gsettings set com.worldwidemann.plotinus:\/com\/worldwidemann\/plotinus\/[application]\/ hotkeys '[keys]'\n<\/pre><p><code>[keys]<\/code> must be an array of strings in the format expected by <a href=\"https:\/\/developer.gnome.org\/gtk3\/stable\/gtk3-Keyboard-Accelerators.html#gtk-accelerator-parse\" rel=\"nofollow\"><code>gtk_accelerator_parse<\/code><\/a>, e.g. <code>[\"&lt;Primary&gt;&lt;Shift&gt;P\", \"&lt;Primary&gt;P\"]<\/code>. Each shortcut in the array opens the command palette.<\/p><h3><a id=\"user-content-enablingdisabling-d-bus-window-registration\" class=\"anchor\" aria-hidden=\"true\" href=\"#enablingdisabling-d-bus-window-registration\"><\/a>Enabling\/disabling D-Bus window registration<\/h3><pre>gsettings set com.worldwidemann.plotinus:\/com\/worldwidemann\/plotinus\/[application]\/ dbus-enabled [true\/false]\n<\/pre><p>See the following section for details.<\/p><h2><a id=\"user-content-d-bus-api\" class=\"anchor\" aria-hidden=\"true\" href=\"#d-bus-api\"><\/a>D-Bus API<\/h2><p>Plotinus provides a simple but complete <a href=\"https:\/\/www.freedesktop.org\/wiki\/Software\/dbus\/\" rel=\"nofollow\">D-Bus<\/a> API for developers who want to use its functionality from their own software. The API consists of two methods, exposed on the session bus at <code>com.worldwidemann.plotinus<\/code>:<\/p><ul><li>\n<p><code>GetCommands(window_path) -&gt; (bus_name, command_paths)<\/code><br \/>Takes the object path of a GTK+ window (which can e.g. be obtained from a Mutter window via <a href=\"https:\/\/developer.gnome.org\/meta\/stable\/MetaWindow.html#meta-window-get-gtk-window-object-path\" rel=\"nofollow\"><code>meta_window_get_gtk_window_object_path<\/code><\/a>) and returns an array of object paths referencing commands extracted from that window, as well as the name of the bus on which they are registered.<br \/>The mechanism behind this method is somewhat similar to <a href=\"https:\/\/github.com\/tetzank\/qmenu_hud\/blob\/master\/com.canonical.AppMenu.Registrar.xml\">Ubuntu's AppMenu Registrar<\/a>, but more lightweight and compatible with Wayland. Window registration <a href=\"#enablingdisabling-dbus-window-registration\">must be enabled<\/a> before using this method.<\/p>\n<\/li>\n<li>\n<p><code>ShowCommandPalette(commands) -&gt; (bus_name, command_palette_path)<\/code><br \/>Takes an array of commands (structs of the form <code>(path, label, accelerators)<\/code>) and opens a command palette window displaying those commands. The returned object path references a control object registered on the returned bus name which provides signals on user interaction with the window.<\/p>\n<\/li>\n<\/ul><p>Calls to these methods are processed by the <strong>Plotinus D-Bus service,<\/strong> which can be started with<\/p><pre>plotinus\n<\/pre><h3><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h3><p>The following examples demonstrate how to use the D-Bus API from Python. They require <a href=\"https:\/\/github.com\/LEW21\/pydbus\">pydbus<\/a> to be installed and the Plotinus D-Bus service to be running.<\/p><h4><a id=\"user-content-application-remote-control\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-remote-control\"><\/a>Application remote control<\/h4><div class=\"highlight highlight-source-python\"><pre>#!\/usr\/bin\/env python\nimport sys\nfrom pydbus import SessionBus\nbus = SessionBus()\nplotinus = bus.get(\"com.worldwidemann.plotinus\")\nbus_name, command_paths = plotinus.GetCommands(sys.argv[1])\ncommands = [bus.get(bus_name, command_path) for command_path in command_paths]\nfor i, command in enumerate(commands):\n  print(\"[%d] %s -&gt; %s\" % (i, \" -&gt; \".join(command.Path), command.Label))\nindex = raw_input(\"Number of command to execute: \")\nif index:\n  commands[int(index)].Execute()<\/pre><\/div><p>Before running this example, enable window registration with<\/p><pre>gsettings set com.worldwidemann.plotinus:\/com\/worldwidemann\/plotinus\/default\/ dbus-enabled true\n<\/pre><p>Then, run an application (e.g. gedit) with <a href=\"#enabling-plotinus-in-applications\">Plotinus enabled<\/a>. Now run the script with the window object path as an argument, i.e.<\/p><pre>.\/application_remote_control.py \/org\/gnome\/gedit\/window\/1\n<\/pre><h4><a id=\"user-content-application-launcher\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-launcher\"><\/a>Application launcher<\/h4><p>Based on <a href=\"https:\/\/github.com\/p-e-w\/argos#launcherpy\">this Argos plugin<\/a>, uses Plotinus' command palette to display a list of applications available on the system.<\/p><div class=\"highlight highlight-source-python\"><pre>#!\/usr\/bin\/env python\nimport os, re\nfrom pydbus import SessionBus\nfrom gi.repository import GLib, Gio\napplications = {}\nfor app_info in Gio.AppInfo.get_all():\n  categories = app_info.get_categories()\n  if categories is None:\n    continue\n  # Remove \"%U\" and \"%F\" placeholders\n  command_line = re.sub(\"%\\\\w\", \"\", app_info.get_commandline()).strip()\n  app = (app_info.get_name(), command_line)\n  for category in categories.split(\";\"):\n    if category not in [\"GNOME\", \"GTK\", \"\"]:\n      if category not in applications:\n        applications[category] = []\n      applications[category].append(app)\n      break\ncommands = []\ncommand_lines = []\nfor category, apps in sorted(applications.items()):\n  for app in sorted(apps):\n    commands.append(([category], app[0], []))\n    command_lines.append(app[1])\nbus = SessionBus()\nplotinus = bus.get(\"com.worldwidemann.plotinus\")\nbus_name, command_palette_path = plotinus.ShowCommandPalette(commands)\ncommand_palette = bus.get(bus_name, command_palette_path)\nloop = GLib.MainLoop()\ndef command_executed(index):\n  os.system(command_lines[index])\ncommand_palette.CommandExecuted.connect(command_executed)\ndef closed():\n  # Wait for CommandExecuted signal\n  GLib.timeout_add(500, loop.quit)\ncommand_palette.Closed.connect(closed)\nloop.run()<\/pre><\/div><h2><a id=\"user-content-acknowledgments\" class=\"anchor\" aria-hidden=\"true\" href=\"#acknowledgments\"><\/a>Acknowledgments<\/h2><p>Documentation on GTK+ modules is essentially nonexisting. Without <a href=\"https:\/\/github.com\/chipx86\/gtkparasite\">gtkparasite<\/a> and <a href=\"https:\/\/github.com\/gnome-globalmenu\/gnome-globalmenu\">gnome-globalmenu<\/a> to learn from, it would have been a lot harder to get this project off the ground.<\/p><p>The CMake modules are copied verbatim from Elementary's <a href=\"https:\/\/github.com\/elementary\/pantheon-installer\">pantheon-installer<\/a> repository.<\/p><p>Vala is still the greatest thing ever to happen to Linux Desktop development.<\/p><h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2><p>Contributors are always welcome. However, <strong>please file an issue describing what you intend to add before opening a pull request,<\/strong> <em>especially<\/em> for new features! I have a clear vision of what I want (and do not want) Plotinus to be, so discussing potential additions might help you avoid duplication and wasted work.<\/p><p>By contributing, you agree to release your changes under the same license as the rest of the project (see below).<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright \u00a9 2016-2017 Philipp Emanuel Weidmann (<a href=\"mailto:pew@worldwidemann.com\">pew@worldwidemann.com<\/a>)<\/p><p>Released under the terms of the <a href=\"https:\/\/gnu.org\/licenses\/gpl.html\" rel=\"nofollow\">GNU General Public License, version 3<\/a><\/p>","created_at":"2020-02-13T17:03:21+0000","updated_at":"2020-02-13T17:03:21+0000","published_by":["p-e-w"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/2702526?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1130,"title":"home","url":"https:\/\/www.ynglet.com\/#about","content":"home2020-02-11T17:11:30+00:00<div class=\"post-content\"><div class=\"fusion-fullwidth fullwidth-box nonhundred-percent-fullwidth non-hundred-percent-height-scrolling fusion-builder-row fusion-row c18\"><div class=\"fusion-layout-column fusion_builder_column fusion_builder_column_1_2 fusion-one-half 1_2 fusion-column-wrapper c17\"><div class=\"fusion-text\"><p class=\"c15\"><em>Swim, launch and bounce through this meditative floating platformer.<\/em><\/p><p class=\"c15\"><em>A reactive soundscape creates the soundtrack to your exploration through the playful puzzles of Ynglet\u2019s hand drawn levels.<\/em><\/p><\/div><div class=\"fusion-text\"><ul><li>Absolutely no platforms (which is really unusual for a platformer)!<\/li>\n<li>Jump between bubbles that float in the sky like you\u2019re a space dolphin!<\/li>\n<li>Dash into walls that reflect you. It feels really good and looks awesome!<\/li>\n<li>Did we mention there are no platforms?<\/li>\n<li>A highly reactive and dynamic soundtrack thanks to Ynglet\u2019s custom and needlessly complicated music software.<\/li>\n<li>Lots of unusual creatures doing their own thing, making the world feel alive<\/li>\n<\/ul><\/div><\/div><\/div><div class=\"fusion-fullwidth fullwidth-box nonhundred-percent-fullwidth non-hundred-percent-height-scrolling fusion-builder-row fusion-row c20\"><div class=\"fusion-layout-column fusion_builder_column fusion_builder_column_1_1 fusion-one-full fusion-column-first fusion-column-last 1_1 fusion-column-wrapper c21\"><div class=\"imageframe-align-center\"><img src=\"https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/01\/Hotdog_Grayscale_Circle-150x150.png\" width=\"150\" height=\"150\" alt=\"\" title=\"Hotdog\" class=\"img-responsive wp-image-70\" srcset=\"https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/01\/Hotdog_Grayscale_Circle-200x200.png 200w, https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/01\/Hotdog_Grayscale_Circle-400x400.png 400w, https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/01\/Hotdog_Grayscale_Circle-600x600.png 600w, https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/01\/Hotdog_Grayscale_Circle-800x800.png 800w, https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/01\/Hotdog_Grayscale_Circle.png 1000w\" referrerpolicy=\"no-referrer\" \/><\/div><div class=\"fusion-text\"><p class=\"c15\">In honor of Hotdog<\/p><\/div><\/div><div class=\"fusion-layout-column fusion_builder_column fusion_builder_column_1_3 fusion-one-third 1_3 fusion-column-wrapper c8\"><img src=\"https:\/\/www.ynglet.com\/wp-content\/uploads\/2019\/12\/DFI-logo-pos.gif\" width=\"625\" height=\"219\" alt=\"\" title=\"DFI-logo-pos\" class=\"img-responsive wp-image-104\" referrerpolicy=\"no-referrer\" \/><\/div><\/div><\/div>","created_at":"2020-02-12T22:56:50+0000","updated_at":"2020-02-12T22:56:50+0000","published_by":["Ynglet"],"annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":0,"domain_name":"www.ynglet.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1129,"title":"Setting up Ediff \u00b7 (or emacs","url":"https:\/\/oremacs.com\/2015\/01\/17\/setting-up-ediff\/","content":"17 Jan 2015<p>Once you make Ediff bearable, it becomes wonderful. At least this was my experience. It's a very good tool, but some of the defaults look very poor to me.<\/p><h2 id=\"customizing-the-customize\">Customizing the <code>customize<\/code><\/h2><p>First of all, I'll list a macro that I've started to use for setting custom variables:<\/p><div class=\"highlight\"><pre class=\"language-elisp\" data-lang=\"elisp\">(defmacro csetq (variable value)\n  `(funcall (or (get ',variable 'custom-set)\n                'set-default)\n            ',variable ,value))\n<\/pre><\/div><p>This macro I've put together myself after searching though the code base and not finding something similar; <code>custom-set-variables<\/code> comes close to what I want, or maybe <code>custom-initialize-changed<\/code>. Basically all I want is a <code>setq<\/code> that is aware of the <code>custom-set<\/code> property of a variable. If you know such a macro, please let me know.<\/p><h2 id=\"changing-some-ediff-options\">Changing some Ediff options<\/h2><p>Now, that I've explained the custom setter (by the way, using <code>custom-set-variables<\/code> is absolutely equivalent to <code>csetq<\/code>), here are my changes:<\/p><ol><li>\n<p>Don't use the weird setup with the control panel in a separate frame. I can manage windows in Emacs much better than my desktop (Unity or Gnome Shell) can manage the Emacs frames.<\/p>\n<div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">(csetq ediff-window-setup-function 'ediff-setup-windows-plain)\n<\/pre><\/div>\n<\/li>\n<li>\n<p>Split the windows horizontally instead of vertically. This way, it's much easier to follow the changes.<\/p>\n<div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">(csetq ediff-split-window-function 'split-window-horizontally)\n<\/pre><\/div>\n<\/li>\n<li>\n<p>Ignore white space. I don't write a lot of Python, so I don't care about the white space in the diffs. At the same time, I re-format any LISP code that I edit to my liking, and I want to see only the important changes in the diff, and not the whitespace nonsense.<\/p>\n<div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">(csetq ediff-diff-options \"-w\")\n<\/pre><\/div>\n<\/li>\n<\/ol><h2 id=\"changing-some-ediff-key-bindings\">Changing some Ediff key bindings<\/h2><p>This is just the standard stuff that I like to do for each mode that does not self-insert: assign <kbd>j<\/kbd> to move down, and <kbd>k<\/kbd> to move up.<\/p><div class=\"highlight\"><pre class=\"language-elisp\" data-lang=\"elisp\">(defun ora-ediff-hook ()\n  (ediff-setup-keymap)\n  (define-key ediff-mode-map \"j\" 'ediff-next-difference)\n  (define-key ediff-mode-map \"k\" 'ediff-previous-difference))\n(add-hook 'ediff-mode-hook 'ora-ediff-hook)\n<\/pre><\/div><h2 id=\"restoring-the-windows-after-ediff-quits\">Restoring the windows after Ediff quits<\/h2><p>When you quit an Ediff session with <kbd>q<\/kbd>, it just leaves the two diff windows around, instead of restoring the window configuration from when Ediff was started. Here's the (slightly hacky) code to restore the old window configuration:<\/p><div class=\"highlight\"><pre class=\"language-elisp\" data-lang=\"elisp\">(winner-mode)\n(add-hook 'ediff-after-quit-hook-internal 'winner-undo)\n<\/pre><\/div><h2 id=\"list-of-ways-that-i-use-to-invoke-ediff\">List of ways that I use to invoke Ediff<\/h2><h3 id=\"from-magit\">from <code>magit<\/code><\/h3><p>In 80% of the cases, I call Ediff with <kbd>e<\/kbd> (<code>magit-ediff<\/code>) from <code>magit-status-mode<\/code>. This gives me a better overview of the changes, especially when I want to revert stuff with either <kbd>a<\/kbd> (<code>ediff-copy-A-to-B<\/code>) or <kbd>b<\/kbd> (<code>ediff-copy-B-to-A<\/code>). Probably with <kbd>a<\/kbd>, since <a href=\"https:\/\/github.com\/magit\/magit\">magit<\/a> seems to consistently put the <code>@{index}<\/code> file into the <code>A<\/code> diff window, and the current file into the <code>B<\/code> window.<\/p><p>So if I'm looking at a diff (to which I navigated with <kbd>j<\/kbd>\/<kbd>k<\/kbd>), I can revert it with <kbd>a<\/kbd>. At this point, the diff region in the current file will become equal to that of the index file. But it will not be saved yet. If I want to save the current file in the <code>B<\/code> window, I can do it with <kbd>wb<\/kbd> (<code>ediff-save-buffer<\/code>). After a revert, it's also useful to call <kbd>!<\/kbd>(<code>ediff-update-diffs<\/code>), which will remove the zero-length diffs and update the diff-count in the mode-line.<\/p><h3 id=\"other-methods\">other methods<\/h3><ol><li><code>ediff-buffers<\/code> will diff two (different) buffers. If you happen to have only two windows open with the appropriate buffers, you get them as defaults and can choose them quickly with <kbd>RET<\/kbd> <kbd>RET<\/kbd>.<\/li>\n<li><code>ediff-files<\/code> is similar, but works on files instead. I like to just mark the appropriate files in <code>dired<\/code> with <kbd>m<\/kbd>, then they also get auto-selected by <code>ediff-files<\/code>.<\/li>\n<li><code>ediff-current-file<\/code> is useful to see the unsaved changes to the current file. I rarely use it, since I have a compulsion to save every file every chance I get. There is little hope to recover from this addiction, unless someone implements an analogue of <a href=\"https:\/\/github.com\/bbatsov\/guru-mode\">guru-mode<\/a> that annoys you when you try to save a file. Actually, it's not a bad idea - <kbd>C-x C-s<\/kbd>:<\/li>\n<\/ol><blockquote>\n<p>Oleh, you have saved this file <code>40 times<\/code> in the last hour; the last save was <code>0 minutes 34 seconds ago<\/code>; since then, you've changed <code>27 bytes<\/code> of information. Are you <strong>sure<\/strong> you want to save <code>[y\/N]<\/code>?<\/p>\n<\/blockquote>","created_at":"2020-02-11T03:19:07+0000","updated_at":"2020-02-11T03:19:07+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"oremacs.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1128,"title":"200ok-ch\/organice","url":"https:\/\/github.com\/200ok-ch\/organice\/","content":"<p style=\"text-align: center;\"><strong>organice organizes Org files nicely!<\/strong><\/p><p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/public\/organice-small.png\"><img src=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/public\/organice-small.png\" class=\"c3\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/p><p>Tests: <a href=\"https:\/\/circleci.com\/gh\/200ok-ch\/organice\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/dcb1da8608d2294effef25ccbeca4f896e9fef95\/68747470733a2f2f636972636c6563692e636f6d2f67682f3230306f6b2d63682f6f7267616e6963652e7376673f7374796c653d737667\" alt=\"https:\/\/circleci.com\/gh\/200ok-ch\/organice.svg?style=svg\" data-canonical-src=\"https:\/\/circleci.com\/gh\/200ok-ch\/organice.svg?style=svg\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a> Maintainability: <a href=\"https:\/\/codeclimate.com\/github\/200ok-ch\/organice\/maintainability\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a2db2e96edf7744225fadbae33979161db79c788\/68747470733a2f2f6170692e636f6465636c696d6174652e636f6d2f76312f6261646765732f34316436313462316638356338623236313036372f6d61696e7461696e6162696c6974792e706e67\" alt=\"https:\/\/api.codeclimate.com\/v1\/badges\/41d614b1f85c8b261067\/maintainability.png\" data-canonical-src=\"https:\/\/api.codeclimate.com\/v1\/badges\/41d614b1f85c8b261067\/maintainability.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a> Dependency management: <a href=\"https:\/\/greenkeeper.io\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/baff4b7e12ae46eb6fa87bf6ae78d1a426e5d714\/68747470733a2f2f6261646765732e677265656e6b65657065722e696f2f3230306f6b2d63682f6f7267616e6963652e737667\" alt=\"https:\/\/badges.greenkeeper.io\/200ok-ch\/organice.svg\" data-canonical-src=\"https:\/\/badges.greenkeeper.io\/200ok-ch\/organice.svg\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>Documentation: <a href=\"https:\/\/organice.200ok.ch\/documentation.html\" rel=\"nofollow\">https:\/\/organice.200ok.ch\/documentation.html<\/a><\/p><h2><a id=\"user-content-what-does-this-project-do\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-does-this-project-do\"><\/a>What does this project do?<\/h2><p>organice is an implementation of <a href=\"http:\/\/orgmode.org\/\" rel=\"nofollow\">Org mode<\/a> without the dependency of <a href=\"https:\/\/www.gnu.org\/software\/emacs\/\" rel=\"nofollow\">Emacs<\/a>. It is built for mobile and desktop browsers and syncs with Dropbox, Google Drive and WebDAV.<\/p><p>At <a href=\"https:\/\/200ok.ch\/\" rel=\"nofollow\">200ok<\/a>, we run an instance of organice at <a href=\"https:\/\/organice.200ok.ch\" rel=\"nofollow\">https:\/\/organice.200ok.ch<\/a>, which is open for anyone to use! organice does not have a back-end (it\u2019s just a front-end application, which uses either Dropbox, Google Drive or WebDAV as back-end storage). We don\u2019t store any kind of data on our servers - we also don\u2019t use analytics on organice.200ok.ch.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-overview.png\"><img src=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-overview.png\" alt=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-overview.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-why-is-this-project-useful\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-is-this-project-useful\"><\/a>Why is this project useful<\/h2><p>Emacs is great, but it\u2019s desktop software. For users who want to access or edit their Org mode files whilst on the go, organice is a great choice.<\/p><h2><a id=\"user-content-introduction\" class=\"anchor\" aria-hidden=\"true\" href=\"#introduction\"><\/a>Introduction<\/h2><p>If you prefer a video to some text, we\u2019ve got you covered! For <a href=\"https:\/\/emacsconf.org\/2019\/\" rel=\"nofollow\">EmacsConf 2019<\/a>, we\u2019ve created a 10 minute introductory video into the rationale and usability of organice.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-introduction.png\"><img src=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-introduction.png\" alt=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-introduction.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>You can watch it on:<\/p><ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=aQKc0hcFXCk\" rel=\"nofollow\">Youtube<\/a><\/li>\n<li><a href=\"https:\/\/media.emacsconf.org\/2019\/05.html\" rel=\"nofollow\">emacsconf.org<\/a><\/li>\n<\/ul><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>&lt;&lt;installation&gt;&gt;<\/p><p>organice is a web application. You can use it from any browser. On iOS and Android, you can install organice to your homescreen. When started from there, it will run in full-screen and add some <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/offline_support\">offline capabilities<\/a>.<\/p><p>Installing to the homescreen on:<\/p><ul><li>iOS: Open organice in Mobile Safari. Tap the \u201cshare\u201d button and select \u201cAdd to Home Screen\u201d.<\/li>\n<li>Android: Open organice in Chrome or Firefox. Tap the \u201cmenu\u201d button and select \u201cAdd to homescreen\u201d.<\/li>\n<\/ul><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><h3><a id=\"user-content-current-restrictionsexpectations-of-organice\" class=\"anchor\" aria-hidden=\"true\" href=\"#current-restrictionsexpectations-of-organice\"><\/a>Current restrictions\/expectations of organice<\/h3><p>\u201cCurrent\u201d means we\u2019re working hard on removing the following restrictions and expectations.<\/p><ul><li>organice understands only a few in-buffer settings (see <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/in_buffer_settings\">Supported in-buffer configuration<\/a>)\n<ul><li>Other in-buffer settings are imported and re-exported but are not editable with organice.<\/li>\n<li>In-buffer settings have to be declared before the first headline.<\/li>\n<\/ul><\/li>\n<li>Other content before the first headline is imported and re-exported, but invisible and currently not editable with organice.<\/li>\n<li>After potential in-buffer settings, your Org file has to begin with a headline.<\/li>\n<\/ul><p>Apart from these restrictions, organice is very robust in reading and editing your Org file and not breaking any of it. We\u2019re having users with 10\u2019000 lines in their files including all kinds of native Org functionality - and even these files work just fine in organice!<\/p><p>Generally, when working with distributed Org files, we\u2019re recommending to put them under version control and to check for bugs and racing conditions between clients.<\/p><p>Please <a href=\"https:\/\/github.com\/200ok-ch\/organice\/issues\/new\">file an issue<\/a> if you find additional restrictions, expectations or bugs that you you wouldn\u2019t have expected.<\/p><h4><a id=\"user-content-background-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#background-information\"><\/a>Background information<\/h4><p>organice has <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/src\/lib\/parse_org.js\">a custom parser<\/a> for Org files. It works quite fine and has unit tests to prove it. One of the quality goals for the parser is that when it parses and re-exports an Org file, it should not change the original file. Not seeing unrelated diffs is important for the productivity of the user. It sounds trivial, but lots of alternative products do not live up to this expectation.<\/p><p>Writing a parser for a complex syntax like Org mode in custom code is hard. Therefore, we are in the process of implementing a proper BNF based parser and a set of tests behind that. If you\u2019re interested, please check it out: <a href=\"https:\/\/github.com\/200ok-ch\/org-parser\">https:\/\/github.com\/200ok-ch\/org-parser<\/a><\/p><p>The strategy we\u2019re using with regards to the parser is this:<\/p><ul><li>Keep improving the existing custom parser for new features and make bugfixes as long as the new one isn\u2019t ready.<\/li>\n<li>In parallel, work on the new one until there is feature parity between both parsers.<\/li>\n<li>When the new one is finished, integrate it into organice.<\/li>\n<\/ul><h3><a id=\"user-content-progressive-web-app--offline-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#progressive-web-app--offline-support\"><\/a>Progressive Web App \/ Offline Support<\/h3><p>&lt;&lt;offline_support&gt;&gt;<\/p><p>organice can run as a PWA (Progressive Web App) - see the <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/installation\">installation instructions<\/a> and does have offline support. From your home screen, organice will start up in full screen and it will use a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Service_Worker_API\" rel=\"nofollow\">Service Worker<\/a> to cache the application. On a desktop browser, the Service Worker will be used automatically. This is implemented using the Create React App <a href=\"https:\/\/create-react-app.dev\/docs\/making-a-progressive-web-app\/#docsNav\" rel=\"nofollow\">Progressive Web App<\/a> functionality which enables the following features:<\/p><ul><li>All static assets are cached so that organice loads fast on subsequent visits, regardless of network connectivity.<\/li>\n<li>Updates are downloaded in the background.<\/li>\n<li>organice works regardless of network state, even if offline.<\/li>\n<li>On mobile devices, organice can be added directly to the user\u2019s home screen, app icon and all.<\/li>\n<\/ul><p>Following that, if you start modifying your Org file when offline, organice will recognize that you are offline and queue up the synchronization until you are online again.<\/p><p>organice also understands when it\u2019s local Org file is outdated compared to the upstream file and will ask you want you want to do - pull the one from the synchronization back-end, push the one from organice or cancel. This happens when you made changes to your file on at least two machines at the same time without synchronizing them in the meantime. For this, we recommend to put your Org file under version control which is the idiomatic solution for changing text based files on multiple machines in parallel.<\/p><h3><a id=\"user-content-customization\" class=\"anchor\" aria-hidden=\"true\" href=\"#customization\"><\/a>Customization<\/h3><h4><a id=\"user-content-general\" class=\"anchor\" aria-hidden=\"true\" href=\"#general\"><\/a>General<\/h4><p>Since organice implements Org mode, one might wonder if we plan to duplicate the Emacs configuration strategy. In Emacs Org mode, there\u2019s more than <a href=\"https:\/\/orgmode.org\/manual\/Customization.html#Customization\" rel=\"nofollow\">500 variables for customization<\/a> - and on top of that, there\u2019s often two ways to configure things:<\/p><ol><li>Using elisp<\/li>\n<li>Using <a href=\"https:\/\/orgmode.org\/manual\/In_002dbuffer-settings.html\" rel=\"nofollow\">in-buffer settings<\/a><\/li>\n<\/ol><p>Modifying Org behavior using elisp (variables) is certainly mighty and powerful. However, the goal of organice is not to clone Emacs in full. In fact, it could be argued that this is not possible. Emacs being a LISP machine has inherent power that cannot be brought to a web application. Instead, the goal is to make Org mode accessible on smartphones and for non-Emacs users. For both use-cases, elisp variable configuration is not an idiomatic or ergonomic option.<\/p><p>organice implements this customization strategy:<\/p><ul><li>Use in-buffer settings where appropriate<\/li>\n<li>Build custom and mobile friendly user interfaces where appropriate\n<ul><li>For example <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/capture_templates\">capture templates<\/a><\/li>\n<\/ul><\/li>\n<\/ul><h4><a id=\"user-content-supported-in-buffer-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-in-buffer-configuration\"><\/a>Supported in-buffer configuration<\/h4><p>&lt;&lt;in_buffer_settings&gt;&gt;<\/p><h5><a id=\"user-content-in-buffer-settings\" class=\"anchor\" aria-hidden=\"true\" href=\"#in-buffer-settings\"><\/a>In-buffer settings<\/h5><ul><li><code>#+TODO<\/code><\/li>\n<li><code>#+TYP_TODO<\/code><\/li>\n<\/ul><h5><a id=\"user-content-startup-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#startup-options\"><\/a><code>#+STARTUP:<\/code> options<\/h5><ul><li><code>nologrepeat<\/code>: Do not record when reinstating repeating item<\/li>\n<\/ul><h5><a id=\"user-content-drawer-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#drawer-properties\"><\/a>Drawer properties<\/h5><ul><li><code>logrepeat<\/code> and <code>nologrepeat<\/code>: Whether to record when reinstating repeating item<\/li>\n<\/ul><pre>:PROPERTIES:\n:LOGGING:  logrepeat\n:END:\n<\/pre><h4><a id=\"user-content-other-customizations\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-customizations\"><\/a>Other customizations<\/h4><p>For some customizations, organice exposes a mobile friendly user interface. Please find them in the \u2018settings\u2019 view (cogs icon in the header on the right).<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-settings.png\"><img src=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-settings.png\" alt=\"https:\/\/raw.githubusercontent.com\/200ok-ch\/organice\/master\/images\/screenshot-settings.png\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h2><p>organice is built with <a href=\"https:\/\/reactjs.org\/\" rel=\"nofollow\">React<\/a> and <a href=\"https:\/\/redux.js.org\/\" rel=\"nofollow\">Redux<\/a>. It was bootstrapped with <a href=\"https:\/\/github.com\/facebook\/create-react-app\">Create React App<\/a>. The tests are written with <a href=\"https:\/\/testing-library.com\/docs\/react-testing-library\/intro\" rel=\"nofollow\">React Testing Library<\/a>. The internal data structures are written as immutable persistent data collections with the <a href=\"https:\/\/github.com\/immutable-js\/immutable-js\">Immutable<\/a> library.<\/p><h3><a id=\"user-content-prerequisites\" class=\"anchor\" aria-hidden=\"true\" href=\"#prerequisites\"><\/a>Prerequisites<\/h3><p>You will need a version of the Node.js engine installed which fulfills the requirement stated in <code>package.json<\/code>. If you don\u2019t already have this installed, it is recommended to install it via <a href=\"https:\/\/github.com\/nvm-sh\/nvm\">nvm<\/a>. The organice repository already contains an <code>.nvmrc<\/code> file, so once you have nvm installed, the following commands should be sufficient:<\/p><div class=\"highlight highlight-source-shell\"><pre>nvm install\nnvm use<\/pre><\/div><h3><a id=\"user-content-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#setup\"><\/a>Setup<\/h3><p>To install the necessary packages, run:<\/p><div class=\"highlight highlight-source-shell\"><pre>yarn install<\/pre><\/div><p>To test against your own Dropbox account, you\u2019ll need to create a <code>.env<\/code> file by copying <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/.env.sample\">.env.sample<\/a> to just <code>.env<\/code>.<\/p><div class=\"highlight highlight-source-shell\"><pre>cp .env.sample .env<\/pre><\/div><p>Running the application:<\/p><div class=\"highlight highlight-source-shell\"><pre>yarn start<\/pre><\/div><p>Running the tests:<\/p><div class=\"highlight highlight-source-shell\"><pre>yarn test<\/pre><\/div><p>For searching the Org file, there\u2019s a <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/src\/lib\/headline_filter_parser.grammar.pegjs\">grammar<\/a> for the search clause. It\u2019s written in <a href=\"https:\/\/pegjs.org\/\" rel=\"nofollow\">pegjs<\/a>. Generating the parser code happens automatically on <code>yarn start|build|test<\/code>. When working on the parser, you can manually generate it with:<\/p><div class=\"highlight highlight-source-shell\"><pre>.\/bin\/compile_search_parser.sh<\/pre><\/div><h3><a id=\"user-content-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#testing\"><\/a>Testing<\/h3><p>When you\u2019re developing a new feature and you want to manually test it, it\u2019s best to check it out in a Desktop browser and on your smartphone. This is how you do that:<\/p><h4><a id=\"user-content-desktop\" class=\"anchor\" aria-hidden=\"true\" href=\"#desktop\"><\/a>Desktop<\/h4><p>Run the application with <code>yarn start<\/code> which will open organice in your configured default browser. Alternatively, visit <code>http:\/\/localhost:3000<\/code> in the browser of your choice.<\/p><h4><a id=\"user-content-smartphone\" class=\"anchor\" aria-hidden=\"true\" href=\"#smartphone\"><\/a>Smartphone<\/h4><p>There are multiple options on how you can connect from your smartphone to your computer running organice.<\/p><p>When running organice with <code>yarn start<\/code>, it will show you all the IPs that the application server is bound to. One will be local to your computer, one will be on your network (if you\u2019re connected to a LAN or Wifi, that is).<\/p><p>If your smartphone has access to the same network, you can access it with the given IP address and port number.<\/p><p>If your new feature doesn\u2019t require a synchronization back-end, just open the <code>sample.org<\/code> file which doesn\u2019t require a login. You\u2019re good to go.<\/p><p><strong>Synchronizing with Dropbox or Google Drive<\/strong><\/p><p>If your new feature does require the Dropbox or Google Drive synchronization back-end, there\u2019s an extra step you need to perform.<\/p><p>Both Dropbox and Google Drive require a whitelist of domains that they can be synchronized from. The whitelist for local domains is exclusively short: <code>http:\/\/localhost:3000<\/code>.<\/p><p>Hence, to be able to login from your phone to your dev instance of organice, you\u2019ll need to set up <a href=\"https:\/\/help.ubuntu.com\/community\/SSH\/OpenSSH\/PortForwarding\" rel=\"nofollow\">port forwarding<\/a>. If you have a shell on your phone and an ssh client, you can do that with the following command:<\/p><div class=\"highlight highlight-source-shell\"><pre>ssh -L 3000:localhost:3000 user-dev-machine<\/pre><\/div><p>If you don\u2019t have a shell on your phone, you can use a dedicated SSH application (like <a href=\"https:\/\/www.termius.com\/\" rel=\"nofollow\">Terminus<\/a>).<\/p><h3><a id=\"user-content-contributions\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributions\"><\/a>Contributions<\/h3><p>Please see our <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/CONTRIBUTING.org\">contributer guidelines<\/a> and our <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/CODE_OF_CONDUCT.md\">code of conduct<\/a>.<\/p><h2><a id=\"user-content-deployment\" class=\"anchor\" aria-hidden=\"true\" href=\"#deployment\"><\/a>Deployment<\/h2><p>Since organice is a front-end only application, it can easily be deployed to any server capable of serving a static application.<\/p><p>Please note: If you want the hosted application to connect to Dropbox or Google Drive, please read the section on <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/synchronization_back_ends\">Synchronization back-ends<\/a>.<\/p><h3><a id=\"user-content-ftp\" class=\"anchor\" aria-hidden=\"true\" href=\"#ftp\"><\/a>FTP<\/h3><p>First create the production build locally: <code>yarn run build<\/code> Note: Creating a build will actually make your <code>REACT_APP_*<\/code> variables from the <code>.env<\/code> file available under <code>process.env<\/code> even though it\u2019ll be a front-end application.<\/p><p>And then upload to your web-server. Here\u2019s a script for your convenience:<\/p><div class=\"highlight highlight-source-shell\"><pre>HOST='your_ftp_server_host'\nUSER='ftp_user'\nPASSWD='ftp_password'\nlftp $HOST &lt;&lt;END_SCRIPT\nuser $USER $PASSWD\nmirror -R build\/\nquit\nEND_SCRIPT\nexit 0<\/pre><\/div><h3><a id=\"user-content-docker\" class=\"anchor\" aria-hidden=\"true\" href=\"#docker\"><\/a>Docker<\/h3><p>organice is also available as a Docker image.<\/p><h4><a id=\"user-content-with-docker-compose\" class=\"anchor\" aria-hidden=\"true\" href=\"#with-docker-compose\"><\/a>With <code>docker-compose<\/code><\/h4><p>If <a href=\"https:\/\/docs.docker.com\/compose\/\" rel=\"nofollow\">docker-compose<\/a> is installed, the following command downloads and runs the latest image automatically.<\/p><div class=\"highlight highlight-source-shell\"><pre>docker-compose up -d<\/pre><\/div><p>The webserver is listening on port 5000 and can be reached here: http:\/\/localhost:5000<\/p><p>If you want to build the image yourself, use the <code>docker-compose-dev.yaml<\/code> file:<\/p><div class=\"highlight highlight-source-shell\"><pre>docker-compose -f docker-compose-dev.yaml up<\/pre><\/div><h4><a id=\"user-content-without-docker-compose\" class=\"anchor\" aria-hidden=\"true\" href=\"#without-docker-compose\"><\/a>Without docker-compose<\/h4><p>If <code>docker-compose<\/code> is not installed the command looks like this:<\/p><div class=\"highlight highlight-source-shell\"><pre>docker run -p 5000:5000 --name organice twohundredok\/organice:latest<\/pre><\/div><p>Again the webserver is listening on port 5000 and can be reached here: http:\/\/localhost:5000<\/p><h3><a id=\"user-content-heroku\" class=\"anchor\" aria-hidden=\"true\" href=\"#heroku\"><\/a>Heroku<\/h3><p>Assuming, you have an account and have installed the <a href=\"https:\/\/devcenter.heroku.com\/articles\/heroku-cli\" rel=\"nofollow\">command line tools<\/a>, deployment is as easy as:<\/p><div class=\"highlight highlight-source-shell\"><pre>heroku create\nheroku config:set ON_HEROKU=1\ngit push heroku master<\/pre><\/div><h3><a id=\"user-content-synchronization-back-ends\" class=\"anchor\" aria-hidden=\"true\" href=\"#synchronization-back-ends\"><\/a>Synchronization back-ends<\/h3><p>&lt;&lt;synchronization_back_ends&gt;&gt;<\/p><p>Note that logging in to Dropbox will only work if you\u2019re running the app on <code>http:\/\/localhost:3000<\/code>, because all redirect URIs must be specified ahead of time on the Dropbox developer console.<\/p><p>To configure your own application on Dropbox, please go <a href=\"https:\/\/www.dropbox.com\/developers\/apps\/\" rel=\"nofollow\">here<\/a> and then configure this app key in the <code>.env<\/code> file. Make sure to add your own URL as <code>Redirect URI<\/code>.<\/p><p>To configure your own application on Google Drive, please generate an API key as described on <a href=\"https:\/\/developers.google.com\/drive\/api\/v3\/quickstart\/js\" rel=\"nofollow\">this page<\/a>.<\/p><h4><a id=\"user-content-webdav\" class=\"anchor\" aria-hidden=\"true\" href=\"#webdav\"><\/a>WebDAV<\/h4><h5><a id=\"user-content-general-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-1\"><\/a>General<\/h5><p>With WebDAV support, organice can potentially be used with a multitude of synchronization backends: Client\/Server services <a href=\"https:\/\/doc.owncloud.com\/server\/user_manual\/files\/access_webdav.html\" rel=\"nofollow\">ownCloud<\/a>, <a href=\"https:\/\/docs.nextcloud.com\/server\/stable\/user_manual\/files\/access_webdav.html?highlight=webdav\" rel=\"nofollow\">Nextcloud<\/a> and <a href=\"https:\/\/download.seafile.com\/published\/seafile-manual\/extension\/webdav.md\" rel=\"nofollow\">Seafile<\/a>, but also self hosted dedicated WebDAV servers like <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/mod_dav.html\" rel=\"nofollow\">Apache<\/a> or <a href=\"https:\/\/nginx.org\/en\/docs\/http\/ngx_http_dav_module.html\" rel=\"nofollow\">Nginx<\/a>.<\/p><p>Since organice is a front-end application, it will login with JavaScript from within the browser - in turn the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/HTTP\/CORS\" rel=\"nofollow\">Cross-Origin Resource Sharing (CORS)<\/a> headers must be set appropriately. If they are not set, you will not be able to login to your service from a browser. Alternatively, if you\u2019re using a server like Apache or Nginx, you can simply get around CORS by hosting organice on the same domain as your service.<\/p><p>Please note, that when your back-end does not set the correct CORS headers, organice cannot show you a really semantic error message on that. The reason is that browsers <a href=\"https:\/\/www.w3.org\/TR\/cors\/#handling-a-response-to-a-cross-origin-request\" rel=\"nofollow\">hide this information<\/a> from JavaScript. You will simply get a network error. However, you can easily debug it yourself by looking into the JavaScript console. No worries, you don\u2019t have to be a (JavaScript) developer to find out about that - here\u2019s a <a href=\"https:\/\/github.com\/200ok-ch\/organice\/wiki#missing-cors-headers\">screencast in the Wiki<\/a> to show you how to do it.<\/p><h5><a id=\"user-content-more-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-information\"><\/a>More information<\/h5><p>In the <a href=\"https:\/\/github.com\/200ok-ch\/organice\/wiki#webdav\">Wiki<\/a>, you\u2019ll find lots more information regarding WebDAV:<\/p><ul><li>A screencast of how organice works when logging in to a WebDAV server<\/li>\n<li>Documentation how on to setup your own WebDAV Server with Apache2 on Debian<\/li>\n<li>Documentation how to configure Nextcloud behind haproxy to allow WebDAV<\/li>\n<li>Documentation on Nextcloud sharing<\/li>\n<\/ul><h3><a id=\"user-content-routing\" class=\"anchor\" aria-hidden=\"true\" href=\"#routing\"><\/a>Routing<\/h3><p>Whilst organice is a true SPA and therefore has no back-end whatsoever, this does have an implication for deployment with regard to routing. For routes like <code>example.com\/foo<\/code> to work, we need a little something extra. Within the context of a running SPA, <code>\/foo<\/code> would be matched by the React Router and the proper page would be rendered by JavaScript. When initially requesting a route like that from the web server itself, the SPA is not running yet and the web server itself wouldn\u2019t find a file called <code>\/foo<\/code>. It would return a 404. The whole topic is explained in depth in this SO answer: <a href=\"https:\/\/stackoverflow.com\/a\/36623117\" rel=\"nofollow\">https:\/\/stackoverflow.com\/a\/36623117<\/a><\/p><p>For <a href=\"https:\/\/organice.200ok.ch\" rel=\"nofollow\">https:\/\/organice.200ok.ch<\/a> we\u2019ve opted to:<\/p><ul><li>Use the modern HTML5 history API with <a href=\"https:\/\/github.com\/ReactTraining\/react-router\/blob\/master\/packages\/react-router-dom\/docs\/api\/BrowserRouter.md\">BrowserRouter<\/a><\/li>\n<li>Not configure a back-end for isomorphic routing, because it would complicate application and deployment unnecessarily (SEO is a non-issue for organice)<\/li>\n<li>Use good old <a href=\"https:\/\/httpd.apache.org\/\" rel=\"nofollow\">Apache Webserver<\/a> for hosting the compiled static assets<\/li>\n<\/ul><p>Therefore configuring a catchall is as easy as setting up a <code>.htaccess<\/code> file in the root of the organice folder containing:<\/p><pre>RewriteEngine On\nRewriteCond %{DOCUMENT_ROOT}%{REQUEST_URI} -f [OR]\nRewriteCond %{DOCUMENT_ROOT}%{REQUEST_URI} -d\nRewriteRule ^ - [L]\nRewriteRule ^ \/index.html [L]\n<\/pre><h2><a id=\"user-content-capture-templates\" class=\"anchor\" aria-hidden=\"true\" href=\"#capture-templates\"><\/a>Capture templates<\/h2><p>&lt;&lt;capture_templates&gt;&gt;<\/p><p>organice supports capture templates by implementing a flexible mechanism using URL parameters. These three of the following parameters are required and must be URL encoded:<\/p><ul><li><code>captureTemplateName<\/code>: the name of the capture template to use. This capture template must already exist in Settings &gt; Capture templates.<\/li>\n<li><code>captureFile<\/code>: the <code>path<\/code> (for Dropbox) or <code>id<\/code> (for Google Drive) of the file in which to execute the capture template.<\/li>\n<li><code>captureContent<\/code>: the content you\u2019d like to capture. This content will be placed at the cursor position if specified in the capture template (with <code>%?<\/code>), or at the end of the template if its not specified.<\/li>\n<\/ul><p>You can also specify additional custom variables for use in your templates. They should be in the format <code>captureVariable_&lt;your custom variable&gt;<\/code>, and should also be URL encoded. In your capture template they\u2019d show up as <code>%&lt;your custom variable&gt;<\/code>.<\/p><p>Important: At this point, organice assumes that you\u2019ll capture your entries into a header - hence, you\u2019ll have to specify a \u201cheader path\u201d.<\/p><h3><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h3><h4><a id=\"user-content-simple-capture-a-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#simple-capture-a-string\"><\/a>Simple: Capture a string<\/h4><p>Say, you want to capture thoughts\/todos as they occur to you. You might want to have a capture template to just get these things out of your head.<\/p><p>This makes for a good \u201cInbox\u201d capture template:<\/p><p><strong>Capture Template<\/strong><\/p><pre>* TODO %?\n%U\n<\/pre><p><strong>Example URL<\/strong><\/p><p><a href=\"https:\/\/organice.200ok.ch?captureTemplateName=Inbox&amp;captureContent=Read+up+on+capture+templates&amp;captureFile=\/org\/things.org\" rel=\"nofollow\">https:\/\/organice.200ok.ch?captureTemplateName=Inbox&amp;captureContent=Read+up+on+capture+templates&amp;captureFile=\/org\/things.org<\/a><\/p><p><strong>Result<\/strong><\/p><pre>* TODO Read up on capture templates\n[2019-09-08 Sun 20:54]\n<\/pre><h4><a id=\"user-content-with-custom-variable\" class=\"anchor\" aria-hidden=\"true\" href=\"#with-custom-variable\"><\/a>With custom variable<\/h4><p>&lt;&lt;media_capture&gt;&gt;<\/p><p>If you want to add web pages to a reading queue (with a title, a capture date and a URL), this would be a good starting point:<\/p><p><strong>Capture Template<\/strong><\/p><pre>* %?\n%u\n- URL: %mediaURL\n<\/pre><p><strong>Example URL<\/strong><\/p><p><a href=\"https:\/\/organice.200ok.ch?captureTemplateName=Media&amp;captureContent=Play+Emacs+like+an+instrument&amp;captureFile=\/org\/media.org&amp;captureVariable_mediaURL=https:\/\/200ok.ch\/posts\/2018-04-27_Play_Emacs_like_an_Instrument.html\" rel=\"nofollow\">https:\/\/organice.200ok.ch?captureTemplateName=Media&amp;captureContent=Play+Emacs+like+an+instrument&amp;captureFile=\/org\/media.org&amp;captureVariable_mediaURL=https:\/\/200ok.ch\/posts\/2018-04-27_Play_Emacs_like_an_Instrument.html<\/a><\/p><p><strong>Result<\/strong><\/p><pre>* Play Emacs like an instrument\n[2019-09-08 Sun]\n- URL: https:\/\/200ok.ch\/posts\/2018-04-27_Play_Emacs_like_an_Instrument.html\n<\/pre><h3><a id=\"user-content-bookmarklets\" class=\"anchor\" aria-hidden=\"true\" href=\"#bookmarklets\"><\/a>Bookmarklets<\/h3><p>Since organice is a web application, you can use the capture templates feature to create bookmarklets, of course! For example, if you want a bookmarklet to add the current page (title, capture date and URL) to your reading queue using <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/media_capture\">this capture template<\/a>, all you need is a little bit of JavaScript:<\/p><div class=\"highlight highlight-source-js\"><pre>javascript:(function() {\n  const {title} = document;\n  const url = `https:\/\/organice.200ok.ch?captureTemplateName=Media&amp;captureContent=${title}&amp;captureFile=\/org\/media.org&amp;captureVariable_mediaURL=${\n  window.location.href\n}`;\n  window.open(url, \"_blank\");\n})()<\/pre><\/div><h3><a id=\"user-content-siri-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#siri-integration\"><\/a>Siri integration<\/h3><p>The organice capture mechanism integrates very nicely with the <a href=\"https:\/\/support.apple.com\/guide\/shortcuts\/welcome\/ios\" rel=\"nofollow\">Siri Shortcuts<\/a> feature in iOS, allowing you to use Siri to execute capture templates.<\/p><p>You can use <a href=\"https:\/\/www.icloud.com\/shortcuts\/14f91f8cf8f547a183a0734396240984\" rel=\"nofollow\">this sample Shortcut<\/a> to get started with this right away in iOS 12 or newer. Open the link on your iOS device and click \u201cGet Shortcut\u201d. Then open up the Shortcuts app and edit the template by following the directions in the comments. Then <a href=\"https:\/\/support.apple.com\/en-us\/HT209055\" rel=\"nofollow\">record a Siri trigger<\/a> and you\u2019re good to go!<\/p><h2><a id=\"user-content-comparison\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison\"><\/a>Comparison<\/h2><h3><a id=\"user-content-beorg\" class=\"anchor\" aria-hidden=\"true\" href=\"#beorg\"><\/a>Beorg<\/h3><p>Before starting work on organice, I did use Beorg and donated to it multiple times, because I was very happy to have a good option to access Org files on my phone with it.<\/p><p>The important differences to me are:<\/p><ul><li>organice is FOSS which is very much in the spirit of Org whilst Beorg is proprietary<\/li>\n<li>organice is web based, so there is no lock-in to a specific device or OS<\/li>\n<li>Beorg currently has better offline support<\/li>\n<\/ul><h3><a id=\"user-content-org-web\" class=\"anchor\" aria-hidden=\"true\" href=\"#org-web\"><\/a>org-web<\/h3><p>organice has a shared history with <a href=\"https:\/\/github.com\/DanielDe\/org-web\">org-web<\/a>. In fact, it is a friendly fork.<\/p><p>organice differs from org-web in that:<\/p><ul><li>It\u2019s a community driven project. See our\n<ul><li><a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/CODE_OF_CONDUCT.md\">Code of conduct<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/CONTRIBUTING.org\">Contributing guidelines<\/a><\/li>\n<\/ul><\/li>\n<li>It has the commitment of a Swiss company behind it to continually work on it.\n<ul><li>This company is 200ok llc: <a href=\"https:\/\/200ok.ch\/\" rel=\"nofollow\">https:\/\/200ok.ch\/<\/a><\/li>\n<\/ul><\/li>\n<li>It has many bug fixes (for example on parsing and exporting org files) compared to its ancestry.<\/li>\n<li>It continues to evolve independently with it\u2019s own feature set.\n<ul><li>For example: organice has WebDAV support.<\/li>\n<\/ul><\/li>\n<li>It is a project with equal focus on mobile as desktop browsers.<\/li>\n<li>org-web <a href=\"https:\/\/github.com\/DanielDe\/org-web\/issues\/75\">tracks users<\/a> with Google Analytics. organice <a href=\"https:\/\/github.com\/200ok-ch\/organice\/issues\/41\">does not<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-whats-new\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-new\"><\/a>What\u2019s new?<\/h4><p>To see how organice differs from org-web, please consult the <a href=\"https:\/\/github.com\/200ok-ch\/organice\/blob\/master\/changelog.org\">changelog<\/a> which contains the user visible changes since forking.<\/p><h4><a id=\"user-content-acknowledgment\" class=\"anchor\" aria-hidden=\"true\" href=\"#acknowledgment\"><\/a>Acknowledgment<\/h4><p>We are extraordinarily grateful to DanielDe the original creator!<\/p><p>We forked the project, because we have different visions on how to go forward. He envisions a mobile only solution, we think it\u2019s great to have organice be available to any browser to enable anyone on the go or any non-Emacs user easy access to Org files. Also, DanielDe thinks of org-web as <a href=\"https:\/\/github.com\/DanielDe\/org-web\/\/issues\/72\">his pet project<\/a> whereas organice has the full power of <a href=\"https:\/\/200ok.ch\" rel=\"nofollow\">200ok llc<\/a> behind it whilst building a strong self-sufficient community around it.<\/p><p>Thank you for all, DanielDe!<\/p><h2><a id=\"user-content-attributions\" class=\"anchor\" aria-hidden=\"true\" href=\"#attributions\"><\/a>Attributions<\/h2><h3><a id=\"user-content-logo\" class=\"anchor\" aria-hidden=\"true\" href=\"#logo\"><\/a>Logo<\/h3><p>Illustration credit: <a href=\"https:\/\/www.vecteezy.com\/\" rel=\"nofollow\">Vecteezy.com<\/a><\/p>","created_at":"2020-02-10T04:29:43+0000","updated_at":"2020-02-10T04:29:43+0000","published_by":["200ok-ch"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/10961555?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1127,"title":"Attention Required! | Cloudflare","url":"https:\/\/www.artstation.com\/artwork\/XB8X3Y","content":"Attention Required! | Cloudflare\n<div id=\"cf-wrapper\"><p>Please enable cookies.<\/p><div id=\"cf-error-details\" class=\"cf-error-details-wrapper\"><div class=\"cf-section cf-wrapper cf-columns two\"><div class=\"cf-column\"><h2 data-translate=\"why_captcha_headline\">Why do I have to complete a CAPTCHA?<\/h2><p data-translate=\"why_captcha_detail\">Completing the CAPTCHA proves you are a human and gives you temporary access to the web property.<\/p><\/div><div class=\"cf-column\"><h2 data-translate=\"resolve_captcha_headline\">What can I do to prevent this in the future?<\/h2><p data-translate=\"resolve_captcha_antivirus\">If you are on a personal connection, like at home, you can run an anti-virus scan on your device to make sure it is not infected with malware.<\/p><p data-translate=\"resolve_captcha_network\">If you are at an office or shared network, you can ask the network administrator to run a scan across the network looking for misconfigured or infected devices.<\/p><\/div><\/div><\/div><\/div>","created_at":"2020-02-08T20:37:12+0000","updated_at":"2020-02-08T20:37:12+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.artstation.com","http_status":"403"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1126,"title":"retrogradeorbit\/bootleg","url":"https:\/\/github.com\/retrogradeorbit\/bootleg","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/blob\/master\/doc\/bootleg-logo.svg\"><img src=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/raw\/master\/doc\/bootleg-logo.svg?sanitize=true\" width=\"100px\" class=\"c2\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/p><p><a href=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/actions?query=workflow%3A%22Clojure+CI%22\"><img src=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/workflows\/Clojure%20CI\/badge.svg\" alt=\"Github Action\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>Static website generation made simple. A powerful, fast, clojure html templating solution.<\/p><p>Bootleg is a command line tool that rapidly renders clojure based templates. With inbuilt support for html, hiccup, hickory, selmer, mustache, markdown, enlive, json, yaml and edn, it enables you to pull together disparate elements and glue them together to generate the static website of your dreams.<\/p><h2><a id=\"user-content-quickstart\" class=\"anchor\" aria-hidden=\"true\" href=\"#quickstart\"><\/a>Quickstart<\/h2><p>Install for Linux:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ curl -LO https:\/\/github.com\/retrogradeorbit\/bootleg\/releases\/download\/v0.1.7\/bootleg-0.1.7-linux-amd64.tgz\n$ tar xvf bootleg-0.1.7-linux-amd64.tgz\n$ mv bootleg ~\/bin<\/pre><\/div><p>Install for MacOS:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ curl -LO https:\/\/github.com\/retrogradeorbit\/bootleg\/releases\/download\/v0.1.7\/bootleg-0.1.7-macos-amd64.zip\n$ unzip bootleg-0.1.7-macos-amd64.zip\n$ mv bootleg \/usr\/local\/bin<\/pre><\/div><p>Clone this repository and change into the <code>examples\/quickstart<\/code> directory:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ git clone https:\/\/github.com\/retrogradeorbit\/bootleg.git\n$ cd bootleg\/examples\/quickstart<\/pre><\/div><p>A simple page:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ cat example-simple.clj\n[:html\n [:body\n  [:h1 \"A simple webpage\"]\n  [:p \"Made with bootleg for maximum powers!\"]]]<\/pre><\/div><div class=\"highlight highlight-text-html-basic\"><pre>$ bootleg example-simple.clj\n&lt;html&gt;&lt;body&gt;&lt;h1&gt;A simple webpage&lt;\/h1&gt;&lt;p&gt;Made with bootleg for maximum powers!&lt;\/p&gt;&lt;\/body&gt;&lt;\/html&gt;<\/pre><\/div><p>A dynamic example:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ cat example-dynamic.clj\n[:div.countdown\n (for [n (range 10 0 -1)]\n   [:p n])\n [:p \"blast off!\"]]<\/pre><\/div><div class=\"highlight highlight-text-html-basic\"><pre>$ bootleg example-dynamic.clj\n&lt;div class=\"countdown\"&gt;&lt;p&gt;10&lt;\/p&gt;&lt;p&gt;9&lt;\/p&gt;&lt;p&gt;8&lt;\/p&gt;&lt;p&gt;7&lt;\/p&gt;&lt;p&gt;6&lt;\/p&gt;&lt;p&gt;5&lt;\/p&gt;&lt;p&gt;4&lt;\/p&gt;&lt;p&gt;3&lt;\/p&gt;&lt;p&gt;2&lt;\/p&gt;&lt;p&gt;1&lt;\/p&gt;&lt;p&gt;blast off!&lt;\/p&gt;&lt;\/div&gt;<\/pre><\/div><p>Mustache:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ cat example-mustache.clj\n(mustache \"quickstart.html\" (yaml \"fields.yml\"))<\/pre><\/div><div class=\"highlight highlight-text-html-basic\"><pre>$ cat quickstart.html\n&lt;h1&gt;{{ title }}&lt;\/h1&gt;\n&lt;h2&gt;by {{ author }}&lt;\/h2&gt;\n&lt;div&gt;{{&amp; body }}&lt;\/div&gt;<\/pre><\/div><div class=\"highlight highlight-source-yaml\"><pre>$ cat fields.yml\ntitle: Bootleg\nauthor: Crispin\nbody: I'm going to rewrite all my sites with this!<\/pre><\/div><div class=\"highlight highlight-text-html-basic\"><pre>$ bootleg example-mustache.clj\n&lt;h1&gt;Bootleg&lt;\/h1&gt;\n&lt;h2&gt;by Crispin&lt;\/h2&gt;\n&lt;div&gt;I'm going to rewrite all my sites with this!&lt;\/div&gt;<\/pre><\/div><p>Markdown support. Evaluate from command line. Easy downloading of resources by url (for any command):<\/p><div class=\"highlight highlight-text-html-basic\"><pre>$ bootleg -e '(markdown \"https:\/\/raw.githubusercontent.com\/retrogradeorbit\/bootleg\/master\/README.md\")'\n&lt;h1&gt;bootleg&lt;\/h1&gt;&lt;p&gt;Static website generation made simple. A powerful, fast, clojure templating solution that rocks!...<\/pre><\/div><p>CSS selector based processing. The magic of enlive:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ cat example-enlive.clj\n(-&gt; (markdown \"simple.md\")\n    (enlive\/at [:p] (enlive\/set-attr :style \"color:green;\")))<\/pre><\/div><div class=\"highlight highlight-source-gfm\"><pre>$ cat simple.md\n#Markdown support\nThis is some simple markdown<\/pre><\/div><div class=\"highlight highlight-text-html-basic\"><pre>$ bootleg example-enlive.clj\n&lt;h1&gt;Markdown support&lt;\/h1&gt;&lt;p style=\"color:green;\"&gt;This is some simple markdown&lt;\/p&gt;<\/pre><\/div><p>Combine hiccup, mustache, markdown and enlive processing:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ cat example-combine.clj\n(mustache \"quickstart.html\"\n          (assoc (yaml \"fields.yml\")\n                 :body (markdown \"simple.md\" :html)))<\/pre><\/div><div class=\"highlight highlight-text-html-basic\"><pre>$ bootleg example-combine.clj\n&lt;h1&gt;Bootleg&lt;\/h1&gt;\n&lt;h2&gt;by Crispin&lt;\/h2&gt;\n&lt;div&gt;&lt;h1&gt;Markdown support&lt;\/h1&gt;&lt;p&gt;This is some simple markdown&lt;\/p&gt;&lt;\/div&gt;<\/pre><\/div><p>Data output with -d flag:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ cat example-data.clj\n(-&gt; (mustache \"quickstart.html\"\n              (assoc (yaml \"fields.yml\")\n                     :body (markdown \"simple.md\" :html)))\n    (convert-to :hickory-seq))<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d example-data.clj\n({:type :element, :attrs nil, :tag :h1, :content [\"Bootleg\"]}\n \"\\n\"\n {:type :element, :attrs nil, :tag :h2, :content [\"by Crispin\"]}\n \"\\n\"\n {:type :element,\n  :attrs nil,\n  :tag :div,\n  :content [{:type :element,\n             :attrs nil,\n             :tag :h1,\n             :content [\"Markdown support\"]}\n            {:type :element,\n             :attrs nil,\n             :tag :p,\n             :content [\"This is some simple markdown\"]}]}\n \"\\n\")<\/pre><\/div><div class=\"highlight highlight-source-shell\"><pre>$ bootleg example-data.clj          # &lt;- no -d flag means output will be html\n&lt;h1&gt;Bootleg&lt;\/h1&gt;\n&lt;h2&gt;by Crispin&lt;\/h2&gt;\n&lt;div&gt;&lt;h1&gt;Markdown support&lt;\/h1&gt;&lt;p&gt;This is some simple markdown&lt;\/p&gt;&lt;\/div&gt;\n<\/pre><\/div><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>Bootleg is distributed for linux as a single executable file. Download the latest tarball from <a href=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/releases\">https:\/\/github.com\/retrogradeorbit\/bootleg\/releases<\/a> and then extract it. Once extracted, move the binary to your path. For system wide installation try <code>\/usr\/local\/bin<\/code> or for personal use <code>~\/bin<\/code><\/p><div class=\"highlight highlight-source-shell\"><pre>$ curl -LO https:\/\/github.com\/retrogradeorbit\/bootleg\/releases\/download\/v0.1.7\/bootleg-0.1.7-linux-amd64.tgz\n$ tar xvf bootleg-0.1.7-linux-amd64.tgz\n$ mv bootleg \/usr\/local\/bin<\/pre><\/div><h3><a id=\"user-content-other-platforms\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-platforms\"><\/a>Other Platforms<\/h3><p>Windows support is experimental. Download (<a href=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/releases\/download\/v0.1.7\/bootleg-0.1.7-windows-amd64.zip\">https:\/\/github.com\/retrogradeorbit\/bootleg\/releases\/download\/v0.1.7\/bootleg-0.1.7-windows-amd64.zip<\/a>) and unzip the archive. Copy the bootleg.exe binary to somewhere on your path.<\/p><p>The jar release file is also an option if you have java installed. You can run it as follows:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ java -jar bootleg-0.1.7.jar<\/pre><\/div><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><p>Run at the command line for options:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ bootleg -h\nStatic website generation made simple. A powerful, fast, clojure html templating solution.\nUsage: bootleg [options] [clj-file]\nOptions:\n  -h, --help           Print the command line help\n  -v, --version        Print the version string and exit\n  -e, --evaluate CODE  Pass in the hiccup to evaluate on the command line\n  -d, --data           Output the rendered template as a clojure form instead of html\n  -o, --output FILE    Write the output to the specified file instead of stdout\n  -t, --traceback      Print the full exception traceback\n  -c, --colour         Print outputs in colour where appropriate\n      --color          Alias for --colour<\/pre><\/div><h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2><p><code>bootleg<\/code> loads and evaluates the clj file specified on the command line, or evaluates the <code>CODE<\/code> form specified in the -e flag. The code can return any of the supported data structures listed below. <code>bootleg<\/code> will then automatically convert that format into html and write it out to standard out, or to <code>FILE<\/code> if specified. This conversion to html step can be prevented by calling with the <code>-d<\/code> flag. In this case the output will be a pretty formatted edn form of the output data structure.<\/p><p><code>bootleg<\/code> supports five main markup data structures. Three are more flexible. And two are limited. We will begin describing the two limited data structures and why they are limited.<\/p><h3><a id=\"user-content-hiccup\" class=\"anchor\" aria-hidden=\"true\" href=\"#hiccup\"><\/a>hiccup<\/h3><p>Hiccup is a standard clojure DSL syntax for representing markup as nested sequences of vectors, and is represented in option flags by the keyword <code>:hiccup<\/code>. An example of some hiccup: the html <code>&lt;div&gt;&lt;p&gt;This is an example&lt;\/p&gt;&lt;\/div&gt;<\/code> is represented in hiccup as <code>[:div [:p \"This is an example\"]]<\/code>.<\/p><p>Hiccup, as the term is used in bootleg, is refering to this vector form. It represents a single root element and its children. This means there are template fragments that <em>cannot<\/em> be represented like this. For example, the html snippet <code>&lt;p&gt;one&lt;\/p&gt;&lt;p&gt;two&lt;\/p&gt;<\/code> cannot be represented as hiccup. It is comprised of two hiccup forms. <code>[:p \"one\"]<\/code> and <code>[:p \"two\"]<\/code>. See <code>hiccup-seq<\/code> below for information about this form.<\/p><h3><a id=\"user-content-hickory\" class=\"anchor\" aria-hidden=\"true\" href=\"#hickory\"><\/a>hickory<\/h3><p>Hickory is a format used to internally represent document trees in clojure for programmatic processing. In option flags it is referenced by the keyword <code>:hickory<\/code>. It is very verbose and not suitable to write by hand. It is supported internally for passing between functions that use it. A simple example of some hickory: the html <code>&lt;p&gt;one&lt;\/p&gt;<\/code> is represented in hickory as <code>{:type :element, :attrs nil, :tag :p, :content [\"one\"]}<\/code>.<\/p><p>Both the hickory and enlive clojure projects use this format internally to represent and manipulate DOM trees.<\/p><p>Hickory, as the term is used in bootleg, is refering to this hashmap form. It represents a single root element and its children. This means there are template fragments that <em>cannot<\/em> be represented in hickory. For example, the html snippet <code>&lt;p&gt;one&lt;\/p&gt;&lt;p&gt;two&lt;\/p&gt;<\/code> cannot be represented as hickory. It is comprised of two hickory forms. <code>{:type :element, :attrs nil, :tag :p, :content [\"one\"]}<\/code> and <code>{:type :element, :attrs nil, :tag :p, :content [\"two\"]}<\/code>. See <code>hickory-seq<\/code> below for information about this form.<\/p><h3><a id=\"user-content-hiccup-seq\" class=\"anchor\" aria-hidden=\"true\" href=\"#hiccup-seq\"><\/a>hiccup-seq<\/h3><p>Hiccup-seq is simply a clojure sequence (or vector) of hiccup forms. In option flags it is referenced by the keyword <code>:hiccup-seq<\/code>. By wrapping multiple hiccup forms in a sequence, hiccup-seq can represent any single root element (and it's children) <em>and<\/em> any template fragment composed of sibling elements.<\/p><p>For example: the html snippet <code>&lt;p&gt;one&lt;\/p&gt;&lt;p&gt;two&lt;\/p&gt;<\/code> is represented in hiccup-seq as: <code>([:p \"one\"] [:p \"two\"])<\/code><\/p><h3><a id=\"user-content-hickory-seq\" class=\"anchor\" aria-hidden=\"true\" href=\"#hickory-seq\"><\/a>hickory-seq<\/h3><p>Hickory-seq is simply a clojure sequence (or vector) of hickory forms. In option flags it is referenced by the keyword <code>:hickory-seq<\/code>. By wrapping multiple hickory forms in a sequence, hickory-seq can represent any single root element (and it's children) <em>and<\/em> any template fragment composed of sibling elements.<\/p><p>For example: the html snippet <code>&lt;p&gt;one&lt;\/p&gt;&lt;p&gt;two&lt;\/p&gt;<\/code> is represented in hickory-seq as: <code>({:type :element, :attrs nil, :tag :p, :content [\"one\"]} {:type :element, :attrs nil, :tag :p, :content [\"two\"]})<\/code><\/p><h3><a id=\"user-content-html\" class=\"anchor\" aria-hidden=\"true\" href=\"#html\"><\/a>html<\/h3><p>html is represented internally as a string. This is a flexible type and can hold a root element and children, or a number of sibling elements sequentially.<\/p><h3><a id=\"user-content-xml\" class=\"anchor\" aria-hidden=\"true\" href=\"#xml\"><\/a>xml<\/h3><p>xml is represented internally as a string. This type always starts with an XML header <code>&lt;?xml version=\"...\"?&gt;<\/code><\/p><p>You can output xml with bootleg by converting to type <code>:xml<\/code> first:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -e '(convert-to [:link \"foo\"] :xml)'\n&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;&lt;link&gt;foo&lt;\/link&gt;<\/pre><\/div><h2><a id=\"user-content-inbuilt-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#inbuilt-functions\"><\/a>Inbuilt functions<\/h2><p>The following functions are inbuilt into the clojure interpreter:<\/p><h3><a id=\"user-content-markup-processing-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#markup-processing-functions\"><\/a>Markup Processing Functions<\/h3><h4><a id=\"user-content-markdown\" class=\"anchor\" aria-hidden=\"true\" href=\"#markdown\"><\/a>markdown<\/h4><p><code>(markdown source &amp; options)<\/code><\/p><p>Load the markdown from the file specified in <code>source<\/code> and render it. <code>source<\/code> can be a local file path (relative to the executing hiccup file location) or a URL to gather the markdown from.<\/p><p>Options can be used to alter the behaviour of the function. Options are a list of keywords and can be specified in any order after the source parameter. Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as markdown data, not a file to load<\/li>\n<li><code>:hiccup-seq<\/code> Return the rendered markdown as a hiccup sequence data structure<\/li>\n<li><code>:hickory-seq<\/code> Return the rendered markdown as a hickory sequence data structure<\/li>\n<li><code>:html<\/code> Return the rendered markdown as an html string<\/li>\n<\/ul><p>eg.<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -e '(markdown \"# heading\\nparagraph\" :data)'\n&lt;h1&gt;heading&lt;\/h1&gt;&lt;p&gt;paragraph&lt;\/p&gt;<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(markdown \"# heading\\nparagraph\" :data :hickory-seq)'\n({:type :element, :attrs nil, :tag :h1, :content [\"heading\"]}\n {:type :element, :attrs nil, :tag :p, :content [\"paragraph\"]})<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(markdown \"# heading\\nparagraph\" :data :hiccup-seq)'\n([:h1 \"heading\"] [:p \"paragraph\"])<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(markdown \"# heading\\nparagraph\" :data :html)'\n\"&lt;h1&gt;heading&lt;\/h1&gt;&lt;p&gt;paragraph&lt;\/p&gt;\"<\/pre><\/div><h4><a id=\"user-content-mustache\" class=\"anchor\" aria-hidden=\"true\" href=\"#mustache\"><\/a>mustache<\/h4><p><code>(mustache source vars &amp; options)<\/code><\/p><p>Load a mustache template from the file specified in <code>source<\/code> and render it substituting the vars from <code>vars<\/code>. <code>source<\/code> can be a local file path (relative to the executing hiccup file location) or a URL to gather the markdown from.<\/p><p>Options can be used to alter the behaviour of the function. Options are a list of keywords and can be specified in any order after the source parameter. Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as mustache data, not a file to load<\/li>\n<li><code>:hiccup<\/code> Return the rendered mustache as hiccup<\/li>\n<li><code>:hiccup-seq<\/code> Return the rendered mustache as a hiccup sequence data structure<\/li>\n<li><code>:hickory<\/code> Return the rendered mustache as hickory<\/li>\n<li><code>:hickory-seq<\/code> Return the rendered mustache as a hickory sequence data structure<\/li>\n<li><code>:html<\/code> Return the rendered mustache as an html string<\/li>\n<\/ul><p>eg.<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -e '(mustache \"&lt;p&gt;{{var1}}&lt;\/p&gt;&lt;div&gt;{{&amp;var2}}&lt;\/div&gt;\" {:var1 \"value 1\" :var2 \"&lt;p&gt;markup&lt;\/p&gt;\"} :data)'\n&lt;p&gt;value 1&lt;\/p&gt;&lt;div&gt;&lt;p&gt;markup&lt;\/p&gt;&lt;\/div&gt;<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(mustache \"&lt;p&gt;{{var1}}&lt;\/p&gt;\" {:var1 \"value 1\"} :data :hiccup-seq)'\n([:p \"value 1\"])<\/pre><\/div><h4><a id=\"user-content-selmer\" class=\"anchor\" aria-hidden=\"true\" href=\"#selmer\"><\/a>selmer<\/h4><p><code>(selmer source vars &amp; options)<\/code><\/p><p>Load a selmer template from the file specified in <code>source<\/code> and render it substituting the vars from <code>vars<\/code>. <code>source<\/code> can be a local file path (relative to the executing hiccup file location) or a URL to gather the markdown from.<\/p><p>Options can be used to alter the behaviour of the function. Options are a list of keywords and can be specified in any order after the source parameter. Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as selmer data, not a file to load<\/li>\n<li><code>:hiccup<\/code> Return the rendered selmer as hiccup<\/li>\n<li><code>:hiccup-seq<\/code> Return the rendered selmer as a hiccup sequence data structure<\/li>\n<li><code>:hickory<\/code> Return the rendered selmer as hickory<\/li>\n<li><code>:hickory-seq<\/code> Return the rendered selmer as a hickory sequence data structure<\/li>\n<li><code>:html<\/code> Return the rendered selmer as an html string<\/li>\n<\/ul><p>eg.<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -e '(selmer \"&lt;p&gt;Hello {{name|capitalize}}!&lt;\/p&gt;\" {:name \"world\"} :data)'\n&lt;p&gt;Hello World!&lt;\/p&gt;<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(selmer \"&lt;p&gt;Hello {{name|capitalize}}!&lt;\/p&gt;\" {:name \"world\"} :data :hiccup-seq)'\n([:p \"Hello World!\"])<\/pre><\/div><p>The <code>selmer<\/code> namespaces are also provided at their usual namespace locations.<\/p><ul><li>selmer.filter-parser<\/li>\n<li>selmer.filters<\/li>\n<li>selmer.middleware<\/li>\n<li>selmer.node<\/li>\n<li>selmer.parser<\/li>\n<li>selmer.tags<\/li>\n<li>selmer.template-parser<\/li>\n<li>selmer.util<\/li>\n<li>selmer.validator<\/li>\n<\/ul><h4><a id=\"user-content-slurp\" class=\"anchor\" aria-hidden=\"true\" href=\"#slurp\"><\/a>slurp<\/h4><p><code>(slurp source)<\/code><\/p><p>Load the contents of a file, from a local or remote source, into memory and return it. This <code>slurp<\/code> can load from URLs. Does no interpretation of the file contents at all. Returns them as is.<\/p><h4><a id=\"user-content-html-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#html-1\"><\/a>html<\/h4><p><code>(html source &amp; options)<\/code><\/p><p>Loads the contents of a html or xml file and returns them in <code>:hiccup-seq<\/code> (by default).<\/p><p>Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as markdown data, not a file to load<\/li>\n<li><code>:hiccup<\/code> Return the rendered markdown as hiccup<\/li>\n<li><code>:hiccup-seq<\/code> Return the rendered markdown as a hiccup sequence data structure<\/li>\n<li><code>:hickory<\/code> Return the rendered markdown as hickory<\/li>\n<li><code>:hickory-seq<\/code> Return the rendered markdown as a hickory sequence data structure<\/li>\n<li><code>:html<\/code> Return the rendered markdown as an html string<\/li>\n<\/ul><p>eg.<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(html \"&lt;h1&gt;heading&lt;\/h1&gt;&lt;p&gt;body&lt;\/p&gt;\" :data :hiccup-seq)'\n([:h1 \"heading\"] [:p \"body\"])<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(html \"&lt;div&gt;&lt;h1&gt;heading&lt;\/h1&gt;&lt;p&gt;body&lt;\/p&gt;&lt;\/div&gt;\" :data :hiccup)'\n[:div [:h1 \"heading\"] [:p \"body\"]]<\/pre><\/div><h4><a id=\"user-content-hiccup-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#hiccup-1\"><\/a>hiccup<\/h4><p><code>(hiccup source)<\/code><\/p><p>Loads and evaluates the clojure source from another file.<\/p><h3><a id=\"user-content-filesystem-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#filesystem-functions\"><\/a>Filesystem Functions<\/h3><h4><a id=\"user-content-glob\" class=\"anchor\" aria-hidden=\"true\" href=\"#glob\"><\/a>glob<\/h4><p><code>(glob pattern)<\/code><\/p><p>Returns a sequence of files that match the globbing pattern <code>pattern<\/code>. Supports <code>*<\/code>, <code>**<\/code>, <code>?<\/code>, <code>[abc]<\/code>, <code>[a-z]<\/code>, <code>[!a]<\/code> and relative file paths <code>.<\/code> and <code>..<\/code>. File paths are returned relative to the directory of the executing file.<\/p><div class=\"highlight highlight-source-shell\"><pre>$ bootleg -d -e '(glob \"**\/*.y?l\")'\n(\".github\/workflows\/deploy.yml\"\n \".circleci\/config.yml\"\n \"examples\/quickstart\/fields.yml\"\n \"test\/files\/simple.yml\")<\/pre><\/div><h4><a id=\"user-content-symlink\" class=\"anchor\" aria-hidden=\"true\" href=\"#symlink\"><\/a>symlink<\/h4><p><code>(symlink link target)<\/code><\/p><p>Make a symlink from <code>link<\/code> to <code>target<\/code>. Operates idempotently. If the link already exists it does nothing. If the <code>link<\/code> exists but points to another target, changes the link to point to the specified <code>target<\/code>. If the <code>link<\/code> exists but is not a symbolic link, throws an exception. On success returns the path of the link.<\/p><h4><a id=\"user-content-mkdir\" class=\"anchor\" aria-hidden=\"true\" href=\"#mkdir\"><\/a>mkdir<\/h4><p><code>(mkdir path)<\/code><\/p><p>Make a directory <code>path<\/code>. Does not create any parent directories. Operates idempotently. If the direcotry exists, it does nothing. Otherwise it tries to create the directory. On success it returns the directory path.<\/p><h4><a id=\"user-content-mkdirs\" class=\"anchor\" aria-hidden=\"true\" href=\"#mkdirs\"><\/a>mkdirs<\/h4><p><code>(mkdirs path)<\/code><\/p><p>Make a directory <code>path<\/code> including all the parent directories. Operates idempotently. If the direcotry exists, it does nothing. Otherwise it tries to create the directories. On success it returns the final directory path.<\/p><h4><a id=\"user-content-spit\" class=\"anchor\" aria-hidden=\"true\" href=\"#spit\"><\/a>spit<\/h4><p><code>(spit filename data)<\/code><\/p><p>Write <code>data<\/code> into the specified <code>filename<\/code>. The filename is interpereted relative to the path of the current script.<\/p><h3><a id=\"user-content-var-loading-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#var-loading-functions\"><\/a>Var Loading Functions<\/h3><h4><a id=\"user-content-yaml\" class=\"anchor\" aria-hidden=\"true\" href=\"#yaml\"><\/a>yaml<\/h4><p><code>(yaml source &amp; options)<\/code><\/p><p>Load yaml data from <code>source<\/code> and process it. Returns the parsed data structure.<\/p><p>Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as yaml data, not a file to load<\/li>\n<\/ul><h4><a id=\"user-content-json\" class=\"anchor\" aria-hidden=\"true\" href=\"#json\"><\/a>json<\/h4><p><code>(json source &amp; options)<\/code><\/p><p>Load json data from <code>source<\/code> and process it. Returns the parsed data structure.<\/p><p>Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as json data, not a file to load<\/li>\n<\/ul><h4><a id=\"user-content-edn\" class=\"anchor\" aria-hidden=\"true\" href=\"#edn\"><\/a>edn<\/h4><p><code>(edn source &amp; options)<\/code><\/p><p>Load edn data from <code>source<\/code> and process it. Returns the parsed data structure.<\/p><p>Options can be:<\/p><ul><li><code>:data<\/code> Interpret the <code>source<\/code> argument as edn data, not a file to load<\/li>\n<\/ul><h3><a id=\"user-content-data-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#data-testing\"><\/a>Data Testing<\/h3><h4><a id=\"user-content-is-hiccup\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-hiccup\"><\/a>is-hiccup?<\/h4><p><code>(is-hiccup? data)<\/code><\/p><p>Test the markup in <code>data<\/code>. Return <code>true<\/code> if the data is a hiccup form.<\/p><h4><a id=\"user-content-is-hiccup-seq\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-hiccup-seq\"><\/a>is-hiccup-seq?<\/h4><p><code>(is-hiccup-seq? data)<\/code><\/p><p>Test the markup in <code>data<\/code>. Return <code>true<\/code> if the data is a sequence of hiccup forms.<\/p><h4><a id=\"user-content-is-hickory\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-hickory\"><\/a>is-hickory?<\/h4><p><code>(is-hickory? data)<\/code><\/p><p>Test the markup in <code>data<\/code>. Return <code>true<\/code> if the data is a hickory form.<\/p><h4><a id=\"user-content-is-hickory-seq\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-hickory-seq\"><\/a>is-hickory-seq?<\/h4><p><code>(is-hickory-seq? data)<\/code><\/p><p>Test the markup in <code>data<\/code>. Return <code>true<\/code> if the data is a sequence of hickory forms.<\/p><h3><a id=\"user-content-data-conversion\" class=\"anchor\" aria-hidden=\"true\" href=\"#data-conversion\"><\/a>Data Conversion<\/h3><h4><a id=\"user-content-convert-to\" class=\"anchor\" aria-hidden=\"true\" href=\"#convert-to\"><\/a>convert-to<\/h4><p><code>(convert-to data type)<\/code><\/p><p>Convert one supported data type to another. Input data may be hiccup, hiccup-seq, hickory, hickory-seq or html.<\/p><p>Type may be <code>:hiccup<\/code>, <code>:hiccup-seq<\/code>, <code>:hickory<\/code>, <code>:hickory-seq<\/code> or <code>:html<\/code>.<\/p><p>Example:<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(convert-to [:p#id.class \"one\"] :hickory)'\n{:type :element,\n :attrs {:class \"class\", :id \"id\"},\n :tag :p,\n :content [\"one\"]}<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(convert-to \"&lt;p class=\\\"class\\\" id=\\\"id\\\"&gt;one&lt;\/p&gt;\" :hiccup)'\n[:p {:class \"class\", :id \"id\"} \"one\"]<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(convert-to \"&lt;p&gt;one&lt;\/p&gt;&lt;p&gt;two&lt;\/p&gt;\" :hiccup-seq)'\n([:p \"one\"] [:p \"two\"])<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(convert-to {:type :element :tag :p :content [\"one\"]} :html)'\n\"&lt;p&gt;one&lt;\/p&gt;\"<\/pre><\/div><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(convert-to [:link \"foo\"] :xml)'\n\"&lt;?xml version=\\\"1.0\\\" encoding=\\\"UTF-8\\\"?&gt;&lt;link&gt;foo&lt;\/link&gt;\"<\/pre><\/div><p><strong>Note:<\/strong> Some conversions are lossy. Converting from html or any *-seq data type to hickory or hiccup may lose forms. Only the last form will be returned.<\/p><div class=\"highlight highlight-source-clojure\"><pre>$ bootleg -d -e '(convert-to \"&lt;p&gt;one&lt;\/p&gt;&lt;p&gt;two&lt;\/p&gt;\" :hiccup)'\nWarning: converting markup from :html to :hiccup lost 1 form\n[:p \"two\"]<\/pre><\/div><h4><a id=\"user-content-as-html\" class=\"anchor\" aria-hidden=\"true\" href=\"#as-html\"><\/a>as-html<\/h4><p><code>(as-html data)<\/code><\/p><p>Convert any supported input format passed into <code>data<\/code> to html output. Same as <code>(convert-to data :html)<\/code><\/p><h3><a id=\"user-content-enlive-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#enlive-processing\"><\/a>Enlive Processing<\/h3><p>Enlive html functions are to be found in the <code>enlive<\/code> namespace. A new version of <code>at<\/code> is supplied that provides automatic type coercion for the inputs and outputs.<\/p><p>In addition to this the standard enlive namespaces are available in their usual locations:<\/p><ul><li>net.cgrand.enlive-html<\/li>\n<li>net.cgrand.jsoup<\/li>\n<li>net.cgrand.tagsoup<\/li>\n<li>net.cgrand.xml<\/li>\n<\/ul><h4><a id=\"user-content-at\" class=\"anchor\" aria-hidden=\"true\" href=\"#at\"><\/a>at<\/h4><p><code>(at data selector transform &amp; more)<\/code><\/p><p>Take the markup <code>data<\/code> and process every element matching <code>selector<\/code> through the transform <code>transform<\/code><\/p><h4><a id=\"user-content-deftemplate\" class=\"anchor\" aria-hidden=\"true\" href=\"#deftemplate\"><\/a>deftemplate<\/h4><p><code>(deftemplate name source args &amp; forms)<\/code><\/p><p>Defines a template as a function that returns a <em>hiccup-seq<\/em>.<\/p><p>** Note ** the original deftemplate returned a sequence of strings<\/p><h4><a id=\"user-content-defsnippet\" class=\"anchor\" aria-hidden=\"true\" href=\"#defsnippet\"><\/a>defsnippet<\/h4><p><code>(defsnippet name source selector args &amp; forms)<\/code><\/p><p>Define a named snippet -- equivalent to (def name (snippet source selector args ...)).<\/p><p>Returns a hickory-seq on nodes.<\/p><h3><a id=\"user-content-enlive-transforms\" class=\"anchor\" aria-hidden=\"true\" href=\"#enlive-transforms\"><\/a>Enlive Transforms<\/h3><h4><a id=\"user-content-content\" class=\"anchor\" aria-hidden=\"true\" href=\"#content\"><\/a>content<\/h4><p><code>(content &amp; values)<\/code><\/p><p>Replaces the content of the element. Values can be any supported formats: hickory, hickory-seq, hiccup, hiccup-seq or html.<\/p><p>** note: ** This is different to the standard enlive <code>content<\/code> function. The standard function is present as <code>content*<\/code>. Passing html to <code>content<\/code> will embed that markup in the output. In contrast, passing html to <code>content*<\/code> will insert that html in an escaped form.<\/p><h4><a id=\"user-content-html-snippet\" class=\"anchor\" aria-hidden=\"true\" href=\"#html-snippet\"><\/a>html-snippet<\/h4><p><code>(html-snippet &amp; values)<\/code><\/p><p>Concatenate values as a string and then parse it with tagsoup. html-snippet doesn't insert missing or tags.<\/p><h4><a id=\"user-content-html-content\" class=\"anchor\" aria-hidden=\"true\" href=\"#html-content\"><\/a>html-content<\/h4><p><code>(html-content &amp; values)<\/code><\/p><p>Replaces the content of the element. Values are strings containing html code. This is present for backwards compatibility. The new <code>content<\/code> function can take and embed html code and should be used instead.<\/p><h4><a id=\"user-content-wrap\" class=\"anchor\" aria-hidden=\"true\" href=\"#wrap\"><\/a>wrap<\/h4><p><code>(wrap tag attr)<\/code><\/p><p>Wraps selected node into the given tag. eg. <code>(wrap :div)<\/code> or <code>(wrap :div {:class \"foo\"})<\/code><\/p><h4><a id=\"user-content-unwrap\" class=\"anchor\" aria-hidden=\"true\" href=\"#unwrap\"><\/a>unwrap<\/h4><p><code>unwrap<\/code><\/p><p>Opposite to wrap, returns the content of the selected node.<\/p><h4><a id=\"user-content-set-attr\" class=\"anchor\" aria-hidden=\"true\" href=\"#set-attr\"><\/a>set-attr<\/h4><p><code>(set-attr &amp; key-value-pairs)<\/code><\/p><p>Assocs attributes on the selected element. eg. <code>(set-attr :attr1 \"val1\" :attr2 \"val2\")<\/code><\/p><h4><a id=\"user-content-remove-attr\" class=\"anchor\" aria-hidden=\"true\" href=\"#remove-attr\"><\/a>remove-attr<\/h4><p><code>(remove-attr &amp; attr-names)<\/code><\/p><p>Sets given key value pairs as attributes for selected node. eg. <code>(remove-attr :attr1 :attr2)<\/code><\/p><h4><a id=\"user-content-add-class\" class=\"anchor\" aria-hidden=\"true\" href=\"#add-class\"><\/a>add-class<\/h4><p><code>(add-class &amp; classes)<\/code><\/p><p>Adds class(es) to the selected node. eg. <code>(add-class \"foo\" \"bar\")<\/code><\/p><h4><a id=\"user-content-remove-class\" class=\"anchor\" aria-hidden=\"true\" href=\"#remove-class\"><\/a>remove-class<\/h4><p><code>(add-class &amp; classes)<\/code><\/p><p>Removes class(es) from the selected node. eg. <code>(remove-class \"foo\" \"bar\")<\/code><\/p><h4><a id=\"user-content-do-\" class=\"anchor\" aria-hidden=\"true\" href=\"#do-\"><\/a>do-&gt;<\/h4><p><code>(do-&gt; &amp; fns)<\/code><\/p><p>Chains (composes) several transformations. Applies functions from left to right. eg. <code>(do-&gt; transformation1 transformation2)<\/code><\/p><h4><a id=\"user-content-append\" class=\"anchor\" aria-hidden=\"true\" href=\"#append\"><\/a>append<\/h4><p><code>(append &amp; values)<\/code><\/p><p>Appends the values to the content of the selected element. eg. <code>(append \"xyz\" a-node \"abc\")<\/code><\/p><p>Values can be any supported formats: hickory, hickory-seq, hiccup, hiccup-seq or html.<\/p><h4><a id=\"user-content-prepend\" class=\"anchor\" aria-hidden=\"true\" href=\"#prepend\"><\/a>prepend<\/h4><p><code>(prepend &amp; values)<\/code><\/p><p>Prepends the values to the content of the selected element. eg. <code>(prepend \"xyz\" a-node \"abc\")<\/code><\/p><p>Values can be any supported formats: hickory, hickory-seq, hiccup, hiccup-seq or html.<\/p><h4><a id=\"user-content-after\" class=\"anchor\" aria-hidden=\"true\" href=\"#after\"><\/a>after<\/h4><p><code>(after &amp; values)<\/code><\/p><p>Inserts the values after the current selection (node or fragment). eg. <code>(after \"xyz\" a-node \"abc\")<\/code><\/p><p>Values can be any supported formats: hickory, hickory-seq, hiccup, hiccup-seq or html.<\/p><h4><a id=\"user-content-before\" class=\"anchor\" aria-hidden=\"true\" href=\"#before\"><\/a>before<\/h4><p><code>(before &amp; values)<\/code><\/p><p>Inserts the values before the current selection (node or fragment). eg. <code>(before \"xyz\" a-node \"abc\")<\/code><\/p><p>Values can be any supported formats: hickory, hickory-seq, hiccup, hiccup-seq or html.<\/p><h4><a id=\"user-content-substitute\" class=\"anchor\" aria-hidden=\"true\" href=\"#substitute\"><\/a>substitute<\/h4><p><code>(substitute &amp; values)<\/code><\/p><p>Replaces the current selection (node or fragment). eg. <code>(substitute \"xyz\" a-node \"abc\")<\/code><\/p><p>Values can be any supported formats: hickory, hickory-seq, hiccup, hiccup-seq or html.<\/p><h4><a id=\"user-content-move\" class=\"anchor\" aria-hidden=\"true\" href=\"#move\"><\/a>move<\/h4><p><code>(move src-selector dest-selector)<\/code> <code>(src-selector dest-selector combiner)<\/code><\/p><p>Takes all nodes (under the current element) matched by src-selector, removes them and combines them with the elements matched by dest-selector. eg. <code>(move [:.footnote] [:#footnotes] content)<\/code><\/p><h4><a id=\"user-content-content-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#content-1\"><\/a>content*<\/h4><p><code>(content* &amp; values)<\/code><\/p><p>The original enlive content transformer. Replaces the content of the element. Values can be hickory nodes, collection of hikory nodes, or plain text strings.<\/p><h4><a id=\"user-content-append-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#append-1\"><\/a>append*<\/h4><p><code>(append* &amp; values)<\/code><\/p><p>The original enlive append transformer. Appends the values to the content of the selected element. eg. <code>(append \"xyz\" a-node \"abc\")<\/code><\/p><h4><a id=\"user-content-prepend-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#prepend-1\"><\/a>prepend*<\/h4><p><code>(prepend* &amp; values)<\/code><\/p><p>The original enlive prepend transformer. Prepends the values to the content of the selected element. eg. <code>(prepend \"xyz\" a-node \"abc\")<\/code><\/p><h4><a id=\"user-content-after-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#after-1\"><\/a>after*<\/h4><p><code>(after* &amp; values)<\/code><\/p><p>The original enlive after transformer. Inserts the values after the current selection (node or fragment). eg. <code>(after \"xyz\" a-node \"abc\")<\/code><\/p><h4><a id=\"user-content-before-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#before-1\"><\/a>before*<\/h4><p><code>(before* &amp; values)<\/code><\/p><p>The original enlive before transformer. Inserts the values before the current selection (node or fragment). eg. <code>(before \"xyz\" a-node \"abc\")<\/code><\/p><h4><a id=\"user-content-substitute-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#substitute-1\"><\/a>substitute*<\/h4><p><code>(substitute* &amp; values)<\/code><\/p><p>The original enlive substitute transformer. Replaces the current selection (node or fragment). eg. <code>(substitute \"xyz\" a-node \"abc\")<\/code><\/p><h3><a id=\"user-content-hickory-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#hickory-1\"><\/a>Hickory<\/h3><p>The <code>hickory<\/code> namespaces are provided at their usual namespace locations.<\/p><ul><li>hickory.convert<\/li>\n<li>hickory.hiccup-utils<\/li>\n<li>hickory.render<\/li>\n<li>hickory.select<\/li>\n<li>hickory.utils<\/li>\n<li>hickory.zip<\/li>\n<\/ul><h3><a id=\"user-content-other-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-functions\"><\/a>Other Functions<\/h3><h4><a id=\"user-content-pprint\" class=\"anchor\" aria-hidden=\"true\" href=\"#pprint\"><\/a>pprint<\/h4><p><code>(pprint form)<\/code><\/p><p>Pretty print to stdout the passed in form. If <code>--colour<\/code> is passed on command line then print with colour highlighting.<\/p><h4><a id=\"user-content-parse-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#parse-string\"><\/a>parse-string<\/h4><p><code>(parse-string string)<\/code><\/p><p>Parse the passed in string into a clojure type. Useful for converting strings to numbers, keywords, vectors or hashmaps. A binding of <a href=\"https:\/\/github.com\/borkdude\/edamame\">edamame's<\/a> <code>parse-string<\/code> is used for parsing.<\/p><h3><a id=\"user-content-java-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#java-packages\"><\/a>Java Packages<\/h3><p>The following java classes are included and can be imported and used:<\/p><ul><li>java.time.* (<a href=\"https:\/\/docs.oracle.com\/javase\/8\/docs\/api\/java\/time\/package-summary.html\" rel=\"nofollow\">https:\/\/docs.oracle.com\/javase\/8\/docs\/api\/java\/time\/package-summary.html<\/a>)<\/li>\n<\/ul><h4><a id=\"user-content-javatime\" class=\"anchor\" aria-hidden=\"true\" href=\"#javatime\"><\/a>java.time<\/h4><div class=\"highlight highlight-source-shell\"><pre>$ bootleg -d -e '(import [java.time LocalDate]) (LocalDate\/now)'\n#object[java.time.LocalDate \"0x7d108063\" \"2020-01-28\"]<\/pre><\/div><h2><a id=\"user-content-passing-in-context\" class=\"anchor\" aria-hidden=\"true\" href=\"#passing-in-context\"><\/a>Passing in Context<\/h2><h3><a id=\"user-content-environment-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#environment-variables\"><\/a>Environment Variables<\/h3><p>You can access environment variables through the <code>System<\/code> <a href=\"https:\/\/docs.oracle.com\/javase\/8\/docs\/api\/java\/lang\/System.html\" rel=\"nofollow\">namespace<\/a>. This namespace has the <code>getenv<\/code> function:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ FOO=bar bootleg -d -e '(System\/getenv \"FOO\")'\n\"bar\"<\/pre><\/div><h3><a id=\"user-content-command-line-arguments\" class=\"anchor\" aria-hidden=\"true\" href=\"#command-line-arguments\"><\/a>Command Line Arguments<\/h3><p>Command line arguments are available in the <code>*command-line-args*<\/code> var:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ bootleg -d -e '*command-line-args*'\n(\"-d\" \"-e\" \"*command-line-args*\")<\/pre><\/div><p>Also included is the <code>clojure.tools.cli<\/code> namespace that you can <code>require<\/code> as needed to process command line args, or reference its namespace directly:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ bootleg -d -e '(clojure.tools.cli\/parse-opts *command-line-args* [[\"-e\" \"--evaluate CODE\" \"\"] [\"-d\" \"--data\" \"\"]])'\n{:options {:data true,\n           :evaluate \"(clojure.tools.cli\/parse-opts *command-line-args* [[\\\"-e\\\" \\\"--evaluate CODE\\\" \\\"\\\"] [\\\"-d\\\" \\\"--data\\\" \\\"\\\"]])\"},\n :arguments [],\n :summary \"  -e, --evaluate CODE\\n  -d, --data\",\n :errors nil}<\/pre><\/div><p>If you want to pass in edn in a env var or command line you can use edamame's reader that is already available under <code>parse-string<\/code>:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ FOO='{:a 1 :b 2}' bootleg -d -e '(-&gt; \"FOO\" System\/getenv parse-string :b)'\n2<\/pre><\/div><h3><a id=\"user-content-standard-input\" class=\"anchor\" aria-hidden=\"true\" href=\"#standard-input\"><\/a>Standard Input<\/h3><p>When no <code>-e<\/code> argument is passed on the command line and no filename is supplied then bootleg will read its clojure script from standard in:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ bootleg &lt;&lt;&lt; '[:p 1]'\n&lt;p&gt;1&lt;\/p&gt;<\/pre><\/div><p><code>*in*<\/code> is bound to <code>clojure.lang.LineNumberingPushbackReader<\/code> on standard in as it is in clojure. So you can slurp in to process standard in:<\/p><div class=\"highlight highlight-source-shell\"><pre>echo -n \"&lt;html&gt;&lt;\/html&gt;\" | bootleg -d -e '(-&gt; *in* slurp (convert-to :hiccup))'\n[:html]<\/pre><\/div><p>Remember to be careful of trailing newlines when using bash's <code>&lt;&lt;&lt;<\/code> pipe form.<\/p><pre>$ bootleg -d -e '(-&gt; *in* slurp (convert-to :hiccup))' &lt;&lt;&lt; '&lt;html&gt;&lt;\/html&gt;'\nWarning: converting markup from :html to :hiccup lost 1 form\n\"\\n\"\n$ bootleg  -d -e '(-&gt; *in* slurp (convert-to :hiccup-seq))' &lt;&lt;&lt; '&lt;html&gt;&lt;\/html&gt;'\n([:html] \"\\n\")\n<\/pre><h3><a id=\"user-content-shebang\" class=\"anchor\" aria-hidden=\"true\" href=\"#shebang\"><\/a>Shebang<\/h3><p>Bootleg scripts can contain a shebang line. Then if the executable bit is set, they can be run like programmes. Examine the shebang example in <code>examples\/quickstart\/shebang.clj<\/code>.<\/p><div class=\"highlight highlight-source-shell\"><pre>$ cat shebang.clj\n#!\/usr\/bin\/env bootleg\n[:div\n [:h1 \"Command Line Args\"]\n [:ul\n  (for [arg *command-line-args*]\n    [:li arg])]]<\/pre><\/div><p>Make it executable:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ chmod a+x shebang.clj<\/pre><\/div><p>Run it on the command line:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ .\/shebang.clj arg1 arg2 \"This is argument 3\"\n&lt;div&gt;&lt;h1&gt;Command Line Args&lt;\/h1&gt;&lt;ul&gt;&lt;li&gt;.\/shebang.clj&lt;\/li&gt;&lt;li&gt;arg1&lt;\/li&gt;&lt;li&gt;arg2&lt;\/li&gt;&lt;li&gt;This is argument 3&lt;\/li&gt;&lt;\/ul&gt;&lt;\/div&gt;<\/pre><\/div><h2><a id=\"user-content-building-the-executable\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-the-executable\"><\/a>Building the executable<\/h2><p>Ensure graalvm community edition 19.3.0 is installed in your home directory and the native image extension is also installed.<\/p><div class=\"highlight highlight-source-shell\"><pre>$ make clean\n$ make all<\/pre><\/div><p>The compiled file will be at <code>build\/bootleg<\/code><\/p><p>To use a different version of graal or one installed in a different path, add the GRAALVM suffix like:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ make all GRAALVM_HOME=\/path\/to\/graal-vm<\/pre><\/div><h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2><p>The following are some example sites built using bootleg for you to use as inspiration, or as a base for your own work.<\/p><h3><a id=\"user-content-submit-your-own-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#submit-your-own-example\"><\/a>Submit Your Own Example<\/h3><p>Have you built a website with bootleg? Is the source of this website open source? If so <a href=\"https:\/\/github.com\/retrogradeorbit\/bootleg\/issues\/new\">open an issue<\/a> on the project with a link to your github project with the source tree in it and I will add it to the examples list. This way there will be more reference material for bootleg users to draw apon!<\/p><h2><a id=\"user-content-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#thanks\"><\/a>Thanks<\/h2><p><code>bootleg<\/code> leverages other people's amazing work. The following projects and people enable this to exist at all.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright \u00a9 2019 Crispin Wellington<\/p><p>This program and the accompanying materials are made available under the terms of the Eclipse Public License 2.0 which is available at <a href=\"http:\/\/www.eclipse.org\/legal\/epl-2.0\" rel=\"nofollow\">http:\/\/www.eclipse.org\/legal\/epl-2.0<\/a>.<\/p><p>This Source Code may also be made available under the following Secondary Licenses when the conditions for such availability set forth in the Eclipse Public License, v. 2.0 are satisfied: GNU General Public License as published by the Free Software Foundation, either version 2 of the License, or (at your option) any later version, with the GNU Classpath Exception which is available at <a href=\"https:\/\/www.gnu.org\/software\/classpath\/license.html\" rel=\"nofollow\">https:\/\/www.gnu.org\/software\/classpath\/license.html<\/a>.<\/p>","created_at":"2020-02-07T21:37:30+0000","updated_at":"2020-02-07T21:37:30+0000","published_by":["retrogradeorbit"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":23,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/265500?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1125,"title":"What are Lisp macros good for, anyway?","url":"https:\/\/levelup.gitconnected.com\/what-are-lisp-macros-good-for-anyway-dac97d74904a?gi=aa37ac76ebf9","content":"<div class=\"n p ac ae af ag ah ec aj ak\"><p id=\"7a8f\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Say we have written a function named <code class=\"il im in io ip b\">foo<\/code>, which takes one argument and returns a number. We want to know roughly how long this function takes on some typical parameters.<\/p><h2 id=\"ea5e\" class=\"iq hg ef at as hh ir is it iu iv iw ix iy iz ja jb\"><strong class=\"bf\">Python<\/strong><\/h2><p id=\"bcff\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">One option in python is the built-in <a href=\"https:\/\/docs.python.org\/2\/library\/timeit.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">timeit<\/a> library. Usage might look like this:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using timeit<\/figcaption><\/figure><p id=\"4f95\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Note you pass a string containing the code you want to time the execution of. Note also that you can\u2019t use the return value of the call to <code class=\"il im in io ip b\">avg<\/code> afterwards. The <code class=\"il im in io ip b\">setup<\/code> argument is a bit cumbersome, maybe that\u2019s a way around that.<\/p><p id=\"b501\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">It seems a more common approach by Pythonista\u2019s is to just use <code class=\"il im in io ip b\">time<\/code> :<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using time manually<\/figcaption><\/figure><p id=\"bd7f\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">This can be somewhat error-prone but generally works OK. Since it takes time before and after a line, if the line being timed is more complicated it\u2019s not easy to time part of the line (e.g. some sub-expression of it).<\/p><p id=\"2a57\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">We could also write our own timing function that takes a function of no arguments:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Custom timing function<\/figcaption><\/figure><p id=\"e97e\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Then usage would look like this:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using time_call<\/figcaption><\/figure><p id=\"e1ed\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Note that this allows the result of the call to <code class=\"il im in io ip b\">foo<\/code> to be captured and used. This approach is more amenable to adding timing of something inside a bigger piece of code then removing it shortly afterward.<\/p><h2 id=\"99d3\" class=\"iq hg ef at as hh ir is it iu iv iw ix iy iz ja jb\"><strong class=\"bf\">Java<\/strong><\/h2><p id=\"4c57\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">I don\u2019t know of anything built into Java that helps with this. We can write some fairly generic timing functions:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Custom timing functions for Java<\/figcaption><\/figure><p id=\"9c23\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">The signature taking a <code class=\"il im in io ip b\">Supplier<\/code> is for something that will return a value, and the one taking a <code class=\"il im in io ip b\">Runnable<\/code> is for something that will not. Usage looks like:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using the custom timing functions in Java<\/figcaption><\/figure><p id=\"5355\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">This is pretty similar to the Python case, although more verbose due to the type system. Note however that in both Python and Java, timing a piece of code may require wrapping it in another expression so that it\u2019s some kind of function.<\/p><h2 id=\"c08f\" class=\"iq hg ef at as hh ir is it iu iv iw ix iy iz ja jb\">Clojure\u2019s turn<\/h2><p id=\"dfbd\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Clojure has a built-in operation called <code class=\"il im in io ip b\"><a href=\"https:\/\/clojuredocs.org\/clojure.core\/time\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">time<\/a><\/code> which we can use like so:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using Clojure time<\/figcaption><\/figure><p id=\"3514\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\"><code class=\"il im in io ip b\">time<\/code> also returns its argument, so if the call to <code class=\"il im in io ip b\">foo<\/code> is within the context of other code (as is typical), it can be easily inserted and removed at will:<\/p><figure class=\"jg jh ji jj jk jl\">\n<\/figure><p id=\"09b2\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Before you say \u201cbut I have to balance the parens when removing or adding the call to time!\u201d, good editors for Lisp\u2019s like Clojure make that painless. Removing <code class=\"il im in io ip b\">(time<\/code> automatically removes the corresponding right parens for that expression.<\/p><p id=\"7aa0\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">How does <code class=\"il im in io ip b\">time<\/code> work without requiring its argument to implement some interface or contract? In Python, the argument could be a string as it was when using <code class=\"il im in io ip b\">timeit<\/code> , but <code class=\"il im in io ip b\">timeit<\/code> doesn\u2019t return the result of evaluating the code; additionally, it\u2019s not as easy to insert a <code class=\"il im in io ip b\">timeit<\/code> call since it requires making the code into a string. In both Python and Java, you can time something by accepting a function, but this requires the caller to wrap the code to be timed in a lambda expression.<\/p><p id=\"2993\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">The answer of course, is that Clojure <code class=\"il im in io ip b\">time<\/code> is an example of a Lisp macro rather than a function. Let\u2019s look at <a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/clojure-1.9.0\/src\/clj\/clojure\/core.clj#L3850\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">the implementation<\/a> of <code class=\"il im in io ip b\">time<\/code> (documentation omitted for brevity):<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Implementation of time from clojure.core<\/figcaption><\/figure><p id=\"4614\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\"><code class=\"il im in io ip b\">time<\/code> is a macro taking one argument named<code class=\"il im in io ip b\">expr<\/code> . The magic happens in the expression <code class=\"il im in io ip b\">ret# ~expr<\/code> , which executes the provided expression such as <code class=\"il im in io ip b\">(foo 5)<\/code> and assigns it to the variable <code class=\"il im in io ip b\">ret#<\/code>. The <code class=\"il im in io ip b\">~<\/code> operator in a macro <em class=\"ju\">evaluates<\/em> an expression. <code class=\"il im in io ip b\">#<\/code> as in <code class=\"il im in io ip b\">ret#<\/code> is a way of ensuring a unique variable name in the scope of this macro to avoid using a variable in an enclosing scope (i.e. one defined elsewhere in the program) by mistake.<\/p><p id=\"d82f\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">So, Clojure\u2019s version of <code class=\"il im in io ip b\">time<\/code> takes an arbitrary code snippet \u2014 which in a Lisp-like Clojure is represented as a List surrounded by <code class=\"il im in io ip b\">()<\/code> \u2014 evaluates it, prints how long that took, and returns the result.<\/p><p id=\"ce5c\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">We can see how this macro works by using <a href=\"https:\/\/clojuredocs.org\/clojure.core\/macroexpand\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">macroexpand<\/a>:<\/p><\/div><div class=\"n p ac ae af ag ah ec aj ak\"><p id=\"1224\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\"><code class=\"il im in io ip b\">macroexpand<\/code> gives some illustration of what Lisp macros are. They are functions from source code to source code; that is, a function from a list to a list. In Clojure you register such functions using <a href=\"https:\/\/clojuredocs.org\/clojure.core\/defmacro\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">defmacro<\/a>, and then your macros are called during compilation to \u2018expand\u2019 the expression. In the example above, the list <code class=\"il im in io ip b\">(time (foo 5))<\/code> is expanded into one that times the execution of <code class=\"il im in io ip b\">(foo 5)<\/code> and then returns its result. The special power that Lisp macros have is that they can control evaluation (as seen by evaluating the input expression via <code class=\"il im in io ip b\">~expr<\/code> and do arbitrary source-to-source transformations with the full power of the language available.<\/p><p id=\"c5a8\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Part of the reason for discussing the <code class=\"il im in io ip b\">time<\/code> macro was because it\u2019s possible (though not as convenient) to implement something similar in most languages. This might make you wonder if macros are worthwhile. Here are some more examples of macros in Clojure; consider how you could do this functionality in your language, or whether it\u2019s feasible to do it at all.<\/p><ul class=\"\"><li id=\"f978\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd ks kt ku\">Convenient control flow constructs like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/when\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">when<\/a>, <a href=\"https:\/\/clojuredocs.org\/clojure.core\/when-not\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">when-not<\/a>, <a href=\"https:\/\/clojuredocs.org\/clojure.core\/if-not\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">if-not<\/a>.<\/li>\n<li id=\"4a71\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\"><a href=\"https:\/\/clojuredocs.org\/clojure.core\/cond\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">cond<\/a>, which is similar to <a href=\"https:\/\/docs.oracle.com\/en\/java\/javase\/13\/language\/switch-expressions.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">switch expressions in Java 12<\/a> but more general. Note that cond is a macro because it doesn\u2019t evaluate expressions for branches that won\u2019t be executed, which is why most languages without macros would need built-in support for this.<\/li>\n<li id=\"7d34\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">List comprehensions like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/for\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">for<\/a>.<\/li>\n<li id=\"7f13\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\"><a href=\"https:\/\/clojure.org\/guides\/threading_macros\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Threading macros<\/a> like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/-%3E\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">-&gt;<\/a> can be used to make nested expressions easier to read, similar in appearance to <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Operators\/Pipeline_operator\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">the pipeline operator in Javascript<\/a>.<\/li>\n<li id=\"b086\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">Macros like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/doto\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">doto<\/a> and <code class=\"il im in io ip b\">..<\/code> make interop with \u2018host\u2019 languages (Java\/Javascript) easier in Clojure.<\/li>\n<li id=\"2548\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">Testing libraries often use macros to make tests pleasant to read. For example, <a href=\"https:\/\/github.com\/marick\/Midje\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Midje<\/a> makes tests look like examples in Clojure books.<\/li>\n<li id=\"070f\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\"><a href=\"https:\/\/clojuredocs.org\/clojure.core\/comment\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">comment<\/a><\/li>\n<li id=\"2c63\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">Macros can support the more familiar infix notation for math such as <code class=\"il im in io ip b\">(1 + (2 * 3))<\/code> by rewriting it into the prefix notation <code class=\"il im in io ip b\">(+ 1 (* 2 3))<\/code> used by Lisps. An example is the <a href=\"https:\/\/github.com\/rm-hull\/infix\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">infix<\/a> library.<\/li>\n<\/ul><p id=\"576c\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">An important idea in Lisp communities is that you can write these kinds of things yourself (or get them from a library) instead of petitioning a language committee or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Benevolent_dictator_for_life\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Benevolent Dictator for Life<\/a> to support them, keeping the core language small and coherent. The downside of this extensibility is <a href=\"http:\/\/winestockwebdesign.com\/Essays\/Lisp_Curse.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">the curse of lisp<\/a>.<\/p><p id=\"e917\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Lisp macros provide a level of power not easily replicated in other languages. The key to their power is that Lisps are <a href=\"https:\/\/en.wikipedia.org\/wiki\/Homoiconicity\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">homoiconic<\/a>, which means that the source code in Lisp\u2019s can be represented in a data structure the language itself can manipulate (i.e. a list), making these kinds of custom source code transformations sane and useful.<\/p><p id=\"9fa1\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">You might question whether more power in a language is always a good thing. Isn\u2019t this just another way to make a mess of things? Powerful features can be a double-edged sword, but I really like how <a href=\"https:\/\/vanderburg.org\/blog\/2009\/07\/13\/sharp_and_blunt.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Glenn Vanderburg put it<\/a>:<\/p><blockquote class=\"la lb lc\">\n<p id=\"1c59\" class=\"gq gr ef ju gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Weak developers will move heaven and earth to do the wrong thing. You can\u2019t limit the damage they do by locking up the sharp tools. They\u2019ll just swing the blunt tools harder.<\/p>\n<\/blockquote><p id=\"59b7\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">I certainly don\u2019t mean to say developers who don\u2019t write Lisp macros are \u2018weak\u2019. But we shouldn\u2019t limit our capabilities for fear of what others might do with it.<\/p><\/div>","created_at":"2020-02-07T19:34:13+0000","updated_at":"2020-02-07T19:34:13+0000","published_at":"2020-02-07T15:28:05+0000","published_by":["James Vickers"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"levelup.gitconnected.com","preview_picture":"https:\/\/miro.medium.com\/max\/1200\/1*WwekXIkoPivgMjIOfXi8Nw.jpeg","http_status":"200","origin_url":"https:\/\/levelup.gitconnected.com\/what-are-lisp-macros-good-for-anyway-dac97d74904a"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1124,"title":"What are Lisp macros good for, anyway?","url":"https:\/\/levelup.gitconnected.com\/what-are-lisp-macros-good-for-anyway-dac97d74904a?gi=89cdfaf5947a","content":"<div class=\"n p ac ae af ag ah ec aj ak\"><p id=\"7a8f\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Say we have written a function named <code class=\"il im in io ip b\">foo<\/code>, which takes one argument and returns a number. We want to know roughly how long this function takes on some typical parameters.<\/p><h2 id=\"ea5e\" class=\"iq hg ef at as hh ir is it iu iv iw ix iy iz ja jb\"><strong class=\"bf\">Python<\/strong><\/h2><p id=\"bcff\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">One option in python is the built-in <a href=\"https:\/\/docs.python.org\/2\/library\/timeit.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">timeit<\/a> library. Usage might look like this:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using timeit<\/figcaption><\/figure><p id=\"4f95\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Note you pass a string containing the code you want to time the execution of. Note also that you can\u2019t use the return value of the call to <code class=\"il im in io ip b\">avg<\/code> afterwards. The <code class=\"il im in io ip b\">setup<\/code> argument is a bit cumbersome, maybe that\u2019s a way around that.<\/p><p id=\"b501\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">It seems a more common approach by Pythonista\u2019s is to just use <code class=\"il im in io ip b\">time<\/code> :<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using time manually<\/figcaption><\/figure><p id=\"bd7f\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">This can be somewhat error-prone but generally works OK. Since it takes time before and after a line, if the line being timed is more complicated it\u2019s not easy to time part of the line (e.g. some sub-expression of it).<\/p><p id=\"2a57\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">We could also write our own timing function that takes a function of no arguments:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Custom timing function<\/figcaption><\/figure><p id=\"e97e\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Then usage would look like this:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using time_call<\/figcaption><\/figure><p id=\"e1ed\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Note that this allows the result of the call to <code class=\"il im in io ip b\">foo<\/code> to be captured and used. This approach is more amenable to adding timing of something inside a bigger piece of code then removing it shortly afterward.<\/p><h2 id=\"99d3\" class=\"iq hg ef at as hh ir is it iu iv iw ix iy iz ja jb\"><strong class=\"bf\">Java<\/strong><\/h2><p id=\"4c57\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">I don\u2019t know of anything built into Java that helps with this. We can write some fairly generic timing functions:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Custom timing functions for Java<\/figcaption><\/figure><p id=\"9c23\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">The signature taking a <code class=\"il im in io ip b\">Supplier<\/code> is for something that will return a value, and the one taking a <code class=\"il im in io ip b\">Runnable<\/code> is for something that will not. Usage looks like:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using the custom timing functions in Java<\/figcaption><\/figure><p id=\"5355\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">This is pretty similar to the Python case, although more verbose due to the type system. Note however that in both Python and Java, timing a piece of code may require wrapping it in another expression so that it\u2019s some kind of function.<\/p><h2 id=\"c08f\" class=\"iq hg ef at as hh ir is it iu iv iw ix iy iz ja jb\">Clojure\u2019s turn<\/h2><p id=\"dfbd\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Clojure has a built-in operation called <code class=\"il im in io ip b\"><a href=\"https:\/\/clojuredocs.org\/clojure.core\/time\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">time<\/a><\/code> which we can use like so:<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Using Clojure time<\/figcaption><\/figure><p id=\"3514\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\"><code class=\"il im in io ip b\">time<\/code> also returns its argument, so if the call to <code class=\"il im in io ip b\">foo<\/code> is within the context of other code (as is typical), it can be easily inserted and removed at will:<\/p><figure class=\"jg jh ji jj jk jl\">\n<\/figure><p id=\"09b2\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Before you say \u201cbut I have to balance the parens when removing or adding the call to time!\u201d, good editors for Lisp\u2019s like Clojure make that painless. Removing <code class=\"il im in io ip b\">(time<\/code> automatically removes the corresponding right parens for that expression.<\/p><p id=\"7aa0\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">How does <code class=\"il im in io ip b\">time<\/code> work without requiring its argument to implement some interface or contract? In Python, the argument could be a string as it was when using <code class=\"il im in io ip b\">timeit<\/code> , but <code class=\"il im in io ip b\">timeit<\/code> doesn\u2019t return the result of evaluating the code; additionally, it\u2019s not as easy to insert a <code class=\"il im in io ip b\">timeit<\/code> call since it requires making the code into a string. In both Python and Java, you can time something by accepting a function, but this requires the caller to wrap the code to be timed in a lambda expression.<\/p><p id=\"2993\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">The answer of course, is that Clojure <code class=\"il im in io ip b\">time<\/code> is an example of a Lisp macro rather than a function. Let\u2019s look at <a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/clojure-1.9.0\/src\/clj\/clojure\/core.clj#L3850\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">the implementation<\/a> of <code class=\"il im in io ip b\">time<\/code> (documentation omitted for brevity):<\/p><figure class=\"jg jh ji jj jk jl\">\n<figcaption class=\"ax fi jq jr ia dq do dp js jt as cx\">Implementation of time from clojure.core<\/figcaption><\/figure><p id=\"4614\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\"><code class=\"il im in io ip b\">time<\/code> is a macro taking one argument named<code class=\"il im in io ip b\">expr<\/code> . The magic happens in the expression <code class=\"il im in io ip b\">ret# ~expr<\/code> , which executes the provided expression such as <code class=\"il im in io ip b\">(foo 5)<\/code> and assigns it to the variable <code class=\"il im in io ip b\">ret#<\/code>. The <code class=\"il im in io ip b\">~<\/code> operator in a macro <em class=\"ju\">evaluates<\/em> an expression. <code class=\"il im in io ip b\">#<\/code> as in <code class=\"il im in io ip b\">ret#<\/code> is a way of ensuring a unique variable name in the scope of this macro to avoid using a variable in an enclosing scope (i.e. one defined elsewhere in the program) by mistake.<\/p><p id=\"d82f\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">So, Clojure\u2019s version of <code class=\"il im in io ip b\">time<\/code> takes an arbitrary code snippet \u2014 which in a Lisp-like Clojure is represented as a List surrounded by <code class=\"il im in io ip b\">()<\/code> \u2014 evaluates it, prints how long that took, and returns the result.<\/p><p id=\"ce5c\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">We can see how this macro works by using <a href=\"https:\/\/clojuredocs.org\/clojure.core\/macroexpand\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">macroexpand<\/a>:<\/p><\/div><div class=\"n p ac ae af ag ah ec aj ak\"><p id=\"1224\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\"><code class=\"il im in io ip b\">macroexpand<\/code> gives some illustration of what Lisp macros are. They are functions from source code to source code; that is, a function from a list to a list. In Clojure you register such functions using <a href=\"https:\/\/clojuredocs.org\/clojure.core\/defmacro\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">defmacro<\/a>, and then your macros are called during compilation to \u2018expand\u2019 the expression. In the example above, the list <code class=\"il im in io ip b\">(time (foo 5))<\/code> is expanded into one that times the execution of <code class=\"il im in io ip b\">(foo 5)<\/code> and then returns its result. The special power that Lisp macros have is that they can control evaluation (as seen by evaluating the input expression via <code class=\"il im in io ip b\">~expr<\/code> and do arbitrary source-to-source transformations with the full power of the language available.<\/p><p id=\"c5a8\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Part of the reason for discussing the <code class=\"il im in io ip b\">time<\/code> macro was because it\u2019s possible (though not as convenient) to implement something similar in most languages. This might make you wonder if macros are worthwhile. Here are some more examples of macros in Clojure; consider how you could do this functionality in your language, or whether it\u2019s feasible to do it at all.<\/p><ul class=\"\"><li id=\"f978\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd ks kt ku\">Convenient control flow constructs like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/when\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">when<\/a>, <a href=\"https:\/\/clojuredocs.org\/clojure.core\/when-not\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">when-not<\/a>, <a href=\"https:\/\/clojuredocs.org\/clojure.core\/if-not\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">if-not<\/a>.<\/li>\n<li id=\"4a71\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\"><a href=\"https:\/\/clojuredocs.org\/clojure.core\/cond\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">cond<\/a>, which is similar to <a href=\"https:\/\/docs.oracle.com\/en\/java\/javase\/13\/language\/switch-expressions.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">switch expressions in Java 12<\/a> but more general. Note that cond is a macro because it doesn\u2019t evaluate expressions for branches that won\u2019t be executed, which is why most languages without macros would need built-in support for this.<\/li>\n<li id=\"7d34\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">List comprehensions like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/for\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">for<\/a>.<\/li>\n<li id=\"7f13\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\"><a href=\"https:\/\/clojure.org\/guides\/threading_macros\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Threading macros<\/a> like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/-%3E\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">-&gt;<\/a> can be used to make nested expressions easier to read, similar in appearance to <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Operators\/Pipeline_operator\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">the pipeline operator in Javascript<\/a>.<\/li>\n<li id=\"b086\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">Macros like <a href=\"https:\/\/clojuredocs.org\/clojure.core\/doto\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">doto<\/a> and <code class=\"il im in io ip b\">..<\/code> make interop with \u2018host\u2019 languages (Java\/Javascript) easier in Clojure.<\/li>\n<li id=\"2548\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">Testing libraries often use macros to make tests pleasant to read. For example, <a href=\"https:\/\/github.com\/marick\/Midje\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Midje<\/a> makes tests look like examples in Clojure books.<\/li>\n<li id=\"070f\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\"><a href=\"https:\/\/clojuredocs.org\/clojure.core\/comment\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">comment<\/a><\/li>\n<li id=\"2c63\" class=\"gq gr ef at gs b gt kv gv kw gx kx gz ky hb kz hd ks kt ku\">Macros can support the more familiar infix notation for math such as <code class=\"il im in io ip b\">(1 + (2 * 3))<\/code> by rewriting it into the prefix notation <code class=\"il im in io ip b\">(+ 1 (* 2 3))<\/code> used by Lisps. An example is the <a href=\"https:\/\/github.com\/rm-hull\/infix\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">infix<\/a> library.<\/li>\n<\/ul><p id=\"576c\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">An important idea in Lisp communities is that you can write these kinds of things yourself (or get them from a library) instead of petitioning a language committee or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Benevolent_dictator_for_life\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Benevolent Dictator for Life<\/a> to support them, keeping the core language small and coherent. The downside of this extensibility is <a href=\"http:\/\/winestockwebdesign.com\/Essays\/Lisp_Curse.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">the curse of lisp<\/a>.<\/p><p id=\"e917\" class=\"gq gr ef at gs b gt hr gv hs gx ht gz hu hb hv hd dx\">Lisp macros provide a level of power not easily replicated in other languages. The key to their power is that Lisps are <a href=\"https:\/\/en.wikipedia.org\/wiki\/Homoiconicity\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">homoiconic<\/a>, which means that the source code in Lisp\u2019s can be represented in a data structure the language itself can manipulate (i.e. a list), making these kinds of custom source code transformations sane and useful.<\/p><p id=\"9fa1\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">You might question whether more power in a language is always a good thing. Isn\u2019t this just another way to make a mess of things? Powerful features can be a double-edged sword, but I really like how <a href=\"https:\/\/vanderburg.org\/blog\/2009\/07\/13\/sharp_and_blunt.html\" class=\"dc by jc jd je jf\" target=\"_blank\" rel=\"noopener\">Glenn Vanderburg put it<\/a>:<\/p><blockquote class=\"la lb lc\">\n<p id=\"1c59\" class=\"gq gr ef ju gs b gt gu gv gw gx gy gz ha hb hc hd dx\">Weak developers will move heaven and earth to do the wrong thing. You can\u2019t limit the damage they do by locking up the sharp tools. They\u2019ll just swing the blunt tools harder.<\/p>\n<\/blockquote><p id=\"59b7\" class=\"gq gr ef at gs b gt gu gv gw gx gy gz ha hb hc hd dx\">I certainly don\u2019t mean to say developers who don\u2019t write Lisp macros are \u2018weak\u2019. But we shouldn\u2019t limit our capabilities for fear of what others might do with it.<\/p><\/div>","created_at":"2020-02-07T19:33:43+0000","updated_at":"2020-02-07T19:33:43+0000","published_at":"2020-02-07T15:28:05+0000","published_by":["James Vickers"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"levelup.gitconnected.com","preview_picture":"https:\/\/miro.medium.com\/max\/1200\/1*WwekXIkoPivgMjIOfXi8Nw.jpeg","http_status":"200","origin_url":"https:\/\/levelup.gitconnected.com\/what-are-lisp-macros-good-for-anyway-dac97d74904a"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1123,"title":"scallywag\/org-board","url":"https:\/\/github.com\/scallywag\/org-board","content":"<pre class=\"c2\">org-board\n=========\nLast updated:  Wed 30 May 2018 20:06:55 CEST\n* Motivation\n org-board is a bookmarking and web archival system for Emacs Org\n mode, building on ideas from Pinboard &lt;<a href=\"https:\/\/pinboard.in\" rel=\"nofollow\">https:\/\/pinboard.in<\/a>&gt;.  It\n archives your bookmarks so that you can access them even when\n you're not online, or when the site hosting them goes down.\n `wget' is used as a backend for archival, so any of its options\n can be used directly from org-board.  This means you can download\n whole sites for archival with a couple of keystrokes, while\n keeping track of your archives from a simple Org file.\n* Summary\n In org-board, a bookmark is represented by an Org heading of any\n level, with a `URL' property containing one or more URLs.  Once\n such a heading is created, a call to `org-board-archive' creates a\n unique ID and directory for the entry via `org-attach', archives\n the contents and requisites of the page(s) listed in the `URL'\n property using `wget', and saves them inside the entry's\n directory.  A link to the (timestamped) root archive folder is\n created in the property `ARCHIVED_AT'.  Multiple archives can be\n made for each entry.  Additional options to pass to `wget' can be\n specified via the property `WGET_OPTIONS'.  The variable\n `org-board-after-archive-functions' (defaulting to nil) holds a\n list of functions to run after each archival operation.\n* User commands\n `org-board-archive' archives the current entry, creating a unique\n   ID and directory via `org-attach' if necessary.\n `org-board-archive-dry-run' shows the `wget' invocation that will\n   run for this entry in the echo area.\n `org-board-new' prompts for a URL to add to the current entry's\n   properties, then archives the entry immediately.\n `org-board-delete-all' deletes all the archives for this entry by\n   deleting the `org-attach' directory.\n `org-board-open' opens the bookmark at point in a browser.\n   Default to the built-in browser, `eww', and with prefix, the\n   native operating system browser.\n `org-board-diff' uses `zdiff' (if available) or `ediff' to\n   recursively diff two archives of the same entry.\n `org-board-diff3' uses `ediff' to recursively diff three archives\n   of the same entry.\n `org-board-cancel' cancels the current org-board archival process.\n `org-board-run-after-archive-function' prompts for a function and\n   an archive in the current entry, and applies the function to the\n   archive.\n These are all bound in the `org-board-keymap' variable (not bound\n to any key by default).\n* Customizable options\n `org-board-wget-program' is the path to the wget program.\n `org-board-wget-switches' are the command line options to use with\n `wget'.  By default these are included as:\n   \"-e robots=off\"      ignores robots.txt files.\n   \"--page-requisites\"  downloads all page requisites (CSS, images).\n   \"--adjust-extension\" add a \".html\" extension where needed.\n   \"--convert-links\"    convert external links to internal.\n `org-board-agent-header-alist' is an alist mapping agent names to\n their respective header\/user-agent arguments.  Set a\n `WGET_OPTIONS' property to a key of this alist (say,\n `Mac-OS-10.8') and org-board will replace the key with its\n corresponding value before calling wget. This is useful for some\n sites that refuse to serve pages to `wget'.\n `org-board-wget-show-buffer' controls whether the archival process\n buffer is shown in a window (defaults to true).\n `org-board-log-wget-invocation' controls whether to log the\n archival process command in the root of the archival directory\n (defaults to true).\n `org-board-domain-regexp-alist' applies certain options when a\n domain matches a regular expression.  See the docstring for\n details.  As an example, this is used to make sure that `wget'\n does not send a User Agent string when archiving from Google\n Cache, which will not normally serve pages to it.\n `org-board-after-archive-functions' (default nil) holds a list of\n functions to run after an archival takes place.  This is intended\n for user extensions to `org-board'.  The functions receive three\n arguments: a list of URLs downloaded, the folder name where they\n were downloaded and the process filter event string (see the Elisp\n manual for details on the possible values of this string).  For an\n example use of `org-board-after-archive-functions', see the\n \"Example usage\" section below.\n* Known limitations\n Options like \"--header: 'Agent X\" cannot be specified as\n properties, because the property API splits on spaces, and such an\n option has to be passed to `wget' as one argument.  To work around\n this, add these types of options to `org-board-agent-header-alist'\n instead, where the property API is not involved.\n At the moment, only one archive can be done at a time.\n* Example usage\n** Archiving\n I recently found a list of articles on linkers that I wanted to\n bookmark and keep locally for offline reading.  In a dedicated org\n file for bookmarks I created this entry:\n ** TODO Linkers (20-part series)\n :PROPERTIES:\n :URL:          <a href=\"http:\/\/a3f.at\/lists\/linkers\" rel=\"nofollow\">http:\/\/a3f.at\/lists\/linkers<\/a>\n :WGET_OPTIONS: --recursive -l 1 --span-hosts\n :END:\n Where the `URL' property is a page that already lists the URLs\n that I wanted to download.  I specified the recursive property for\n `wget' along with a depth of 1 (\"-l 1\") so that each linked page\n would be downloaded.  With point inside the entry, I run \"M-x\n org-board-archive\".  An `org-attach' directory is created and\n `wget' starts downloading the pages to it.  Afterwards, the end\n the entry looks like this:\n ** TODO Linkers (20-part series)\n :PROPERTIES:\n :URL:          <a href=\"http:\/\/a3f.at\/lists\/linkers\" rel=\"nofollow\">http:\/\/a3f.at\/lists\/linkers<\/a>\n :WGET_OPTIONS: --recursive -l 1 --span-hosts\n :ID:           D3BCE79F-C465-45D5-847E-7733684B9812\n :ARCHIVED_AT:  [2016-08-30-Tue-15-03-56]\n :END:\n The value in the `ARCHIVED_AT' property is a link that points to\n the root of the timestamped archival directory.  The ID property\n was automatically generated by `org-attach'.\n** Diffing\n You can diff between two archives done for the same entry using\n `org-board-diff', so you can see how a page has changed over time.\n The diff recurses through the directory structure of an archive\n and will highlight any changes that have been made.  `ediff' is\n used if `zdiff' is not available (both are capable of recursing\n through a directory structure, but `zdiff' is possibly more\n intuitive to use).  `org-board-diff3' also offers diffing between\n three different archive directories.\n** `org-board-after-archive-functions'\n `org-board-after-archive-functions' is a list of functions run\n after an archive is finished.  You can use it to do anything you\n like with newly archived pages.  For example, you could add a\n function that copies the new archive to an external hard disk, or\n opens the archived page in your browser as soon as it is done\n downloading.  You could also, for instance, copy all of the media\n files that were downloaded to your own media folder, and pop up a\n Dired buffer inside that folder to give you the chance to\n organize them.\n Here is an example function that copies the archived page to an\n external service called `IPFS' &lt;<a href=\"http:\/\/ipfs.io\/\" rel=\"nofollow\">http:\/\/ipfs.io\/<\/a>&gt;, a decentralized\n versioning and storage system geared towards web content (thanks\n to Alan Schmitt):\n (defun org-board-add-to-ipfs (urls output-folder event &amp;rest _rest)\n   \"Add the downloaded site to IPFS.\"\n   (unless (string-match \"exited abnormally\" event)\n     (let* ((parsed-url (url-generic-parse-url (car urls)))\n            (domain (url-host parsed-url))\n            (path (url-filename parsed-url))\n            (output (shell-command-to-string\n                     (concat \"ipfs add -r \"\n                             (concat output-folder domain))))\n            (ipref\n             (nth 1 (split-string\n                     (car (last (split-string output \"\\n\" t))) \" \"))))\n       (with-current-buffer (get-buffer-create \"*org-board-post-archive*\")\n         (princ (format \"your file is at %s\\n\"\n                        (concat \"http:\/\/localhost:8080\/ipfs\/\" ipref path))\n                (current-buffer))))))\n (eval-after-load \"org-board\"\n   '(add-hook 'org-board-after-archive-functions 'org-board-add-to-ipfs))\n Note that for forward compatibility, it's best to add to a final\n `&amp;rest' argument to every function listed in\n `org-board-after-archive-functions', since a future update may\n provide each function with additional arguments (like a marker\n pointing to a buffer position where the archive was initiated, for\n example).\n For more information on `org-board-after-archive-functions', see\n its docstring and the docstring of\n `org-board-test-after-archive-function'.\n You can also interactively run an after-archive function with the\n command `org-board-run-after-archive-function'.  See its docstring\n for details.\n* Getting started\n** Installation\n There are two ways to install the package.  One way is to clone\n this repository and add the directory to your load-path manually.\n (add-to-list 'load-path \"\/path\/to\/org-board\")\n (require 'org-board)\n Alternatively, you can download the package directly from Emacs\n using MELPA &lt;<a href=\"https:\/\/github.com\/melpa\/melpa\">https:\/\/github.com\/melpa\/melpa<\/a>&gt;.  M-x\n package-install RET org-board RET will take care of it.\n** Keybindings\n The following keymap is defined in `org-board-keymap':\n | Key | Command                              |\n | a   | org-board-archive                    |\n | r   | org-board-archive-dry-run            |\n | n   | org-board-new                        |\n | k   | org-board-delete-all                 |\n | o   | org-board-open                       |\n | d   | org-board-diff                       |\n | 3   | org-board-diff3                      |\n | c   | org-board-cancel                     |\n | x   | org-board-run-after-archive-function |\n | O   | org-attach-reveal-in-emacs           |\n | ?   | Show help for this keymap.           |\n To install the keymap give it a prefix key, e.g.:\n (global-set-key (kbd \"C-c o\") org-board-keymap)\n Then typing `C-c o a' would run `org-board-archive', for example.\n* Miscellaneous\n The location of `wget' should be picked up automatically from the\n `PATH' environment variable.  If it is not, then the variable\n `org-board-wget-program' can be customized.\n Other options are already set so that archiving bookmarks is done\n pretty much automatically.  With no `WGET_OPTIONS' specified, by\n default `org-board-archive' will just download the page and its\n requisites (images and CSS), and nothing else.\n** Support for org-capture from Firefox (thanks to Alan Schmitt):\n On the Firefox side, install org-capture from here:\n   <a href=\"http:\/\/chadok.info\/firefox-org-capture\/\" rel=\"nofollow\">http:\/\/chadok.info\/firefox-org-capture\/<\/a>\n Alternatively, you can do it manually by following the\n instructions here:\n   <a href=\"http:\/\/weblog.zamazal.org\/org-mode-firefox\/\" rel=\"nofollow\">http:\/\/weblog.zamazal.org\/org-mode-firefox\/<\/a>\n     (in the \u201cThe advanced way\u201d section)\n When org-capture is installed, add `(require 'org-protocol)' to\n your init file (`~\/.emacs').\n Then create a capture template like this:\n   (setq org-board-capture-file \"my-org-board.org\")\n   (setq org-capture-templates\n         `(...\n           (\"c\" \"capture through org protocol\" entry\n             (file+headline ,org-board-capture-file \"Unsorted\")\n             \"* %?%:description\\n:PROPERTIES:\\n:URL: %:link\\n:END:\\n\\n Added %U\")\n           ...))\n And add a hook to `org-capture-before-finalize-hook':\n   (defun do-org-board-dl-hook ()\n     (when (equal (buffer-name)\n             (concat \"CAPTURE-\" org-board-capture-file))\n       (org-board-archive)))\n   (add-hook 'org-capture-before-finalize-hook 'do-org-board-dl-hook)\n* Acknowledgements\n Thanks to Alan Schmitt for the code to combine `org-board' and\n `org-capture', and for the example function used in the\n documentation of `org-board-after-archive-functions' above.\n<\/pre>","created_at":"2020-02-03T22:23:44+0000","updated_at":"2020-02-03T22:23:44+0000","published_by":["scallywag"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/3193168?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1122,"title":"terminusdb\/terminus-server","url":"https:\/\/github.com\/terminusdb\/terminus-server","content":"<pre>\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\u2588\u2588\u2588\u2588\u2588\u2588\u2557 \u2588\u2588\u2588\u2557   \u2588\u2588\u2588\u2557\u2588\u2588\u2557\u2588\u2588\u2588\u2557   \u2588\u2588\u2557\u2588\u2588\u2557   \u2588\u2588\u2557\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\n\u255a\u2550\u2550\u2588\u2588\u2554\u2550\u2550\u255d\u2588\u2588\u2554\u2550\u2550\u2550\u2550\u255d\u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2588\u2588\u2557 \u2588\u2588\u2588\u2588\u2551\u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2557  \u2588\u2588\u2551\u2588\u2588\u2551   \u2588\u2588\u2551\u2588\u2588\u2554\u2550\u2550\u2550\u2550\u255d\n   \u2588\u2588\u2551   \u2588\u2588\u2588\u2588\u2588\u2557  \u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d\u2588\u2588\u2554\u2588\u2588\u2588\u2588\u2554\u2588\u2588\u2551\u2588\u2588\u2551\u2588\u2588\u2554\u2588\u2588\u2557 \u2588\u2588\u2551\u2588\u2588\u2551   \u2588\u2588\u2551\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\n   \u2588\u2588\u2551   \u2588\u2588\u2554\u2550\u2550\u255d  \u2588\u2588\u2554\u2550\u2550\u2588\u2588\u2557\u2588\u2588\u2551\u255a\u2588\u2588\u2554\u255d\u2588\u2588\u2551\u2588\u2588\u2551\u2588\u2588\u2551\u255a\u2588\u2588\u2557\u2588\u2588\u2551\u2588\u2588\u2551   \u2588\u2588\u2551\u255a\u2550\u2550\u2550\u2550\u2588\u2588\u2551\n   \u2588\u2588\u2551   \u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2557\u2588\u2588\u2551  \u2588\u2588\u2551\u2588\u2588\u2551 \u255a\u2550\u255d \u2588\u2588\u2551\u2588\u2588\u2551\u2588\u2588\u2551 \u255a\u2588\u2588\u2588\u2588\u2551\u255a\u2588\u2588\u2588\u2588\u2588\u2588\u2554\u255d\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2551\n   \u255a\u2550\u255d   \u255a\u2550\u2550\u2550\u2550\u2550\u2550\u255d\u255a\u2550\u255d  \u255a\u2550\u255d\u255a\u2550\u255d     \u255a\u2550\u255d\u255a\u2550\u255d\u255a\u2550\u255d  \u255a\u2550\u2550\u2550\u255d \u255a\u2550\u2550\u2550\u2550\u2550\u255d \u255a\u2550\u2550\u2550\u2550\u2550\u2550\u255d\n<\/pre><p><a href=\"https:\/\/travis-ci.com\/terminusdb\/terminus-server\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/c77f8c7c55845bde149ac1db6c81205f8d42cf0b\/68747470733a2f2f7472617669732d63692e636f6d2f7465726d696e757364622f7465726d696e75732d7365727665722e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.com\/terminusdb\/terminus-server.svg?branch=master\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>TerminusDB is an open source model driven graph database for knowledge graph representation designed specifically for the web-age.<\/p><p>Terminus Server provides TerminusDB with a RESTful API for interacting with knowledge graphs via the JSON-LD exchange format. This means you can easily string together applications within your own toolchain which utilise the powerful features of graph search and graph storage.<\/p><p>We use an advanced git-like model storing append only changes to graphs represented in concise datastructures using <a href=\"https:\/\/github.com\/terminusdb\/terminus-store\">terminus-store<\/a>.<\/p><h2><a id=\"user-content-latest-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#latest-version\"><\/a>Latest Version<\/h2><p>v1.1.1<\/p><h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting Started<\/h2><h3><a id=\"user-content-install-from-docker-image-recommended\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-from-docker-image-recommended\"><\/a>Install from Docker image (<em>RECOMMENDED<\/em>)<\/h3><p>For all users, we recommend our <a href=\"https:\/\/github.com\/terminusdb\/terminus-quickstart\">Quickstart<\/a> installer.<\/p><p>Follow the instructions there.<\/p><h3><a id=\"user-content-building-from-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-from-source\"><\/a>Building from source<\/h3><p>In order to build from source, follow the <a href=\"https:\/\/github.com\/terminusdb\/terminus-server\/blob\/master\/docs\/BUILD.md\">build from source<\/a> directions.<\/p><h3><a id=\"user-content-open-console\" class=\"anchor\" aria-hidden=\"true\" href=\"#open-console\"><\/a>Open Console<\/h3><p>Once you have installed terminus-server either from the docker image or built from source you can go simply enter the server location into your browser to view your database from the dashboard \"<a href=\"http:\/\/SERVER:PORT\/console\" rel=\"nofollow\">http:\/\/SERVER:PORT\/console<\/a>\". (the default is: <a href=\"http:\/\/localhost:6363\/console\" rel=\"nofollow\">http:\/\/localhost:6363\/console<\/a>).<\/p><h2><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h2><h3><a id=\"user-content-microservices-api\" class=\"anchor\" aria-hidden=\"true\" href=\"#microservices-api\"><\/a>Microservices API<\/h3><p>Terminus Server is organised to be <em>web native<\/em> which means that all actions and activities are easily automated via a RESTful JSON-LD API.<\/p><p>The Server API is documented <a href=\"https:\/\/github.com\/terminusdb\/terminus-server\/blob\/master\/docs\/API.md\">here<\/a>.<\/p><h3><a id=\"user-content-woql\" class=\"anchor\" aria-hidden=\"true\" href=\"#woql\"><\/a>WOQL<\/h3><p>WOQL, the Web Object Query Language, allows you to seemlessly traverse the graph extracting graph fragments, nodes, tables or JSON-LD documents. The syntax itself is in JSON-LD, making the syntax a native data-structure in JavaScript, Python, and Prolog.<\/p><p>The Syntax is documented <a href=\"https:\/\/github.com\/terminusdb\/terminus-server\/blob\/master\/docs\/SYNTAX.md\">here<\/a>.<\/p><h2><a id=\"user-content-why\" class=\"anchor\" aria-hidden=\"true\" href=\"#why\"><\/a>Why<\/h2><p>TerminusDB will:<\/p><ul><li>Make complex data models easy, maintainable and enforced.<\/li>\n<li>Overcome the Object Impedance mismatch without turning your Database into an incomprehensible soup.<\/li>\n<li>Allow you to search for repeating patterns using recursion.<\/li>\n<li>Give you powerful temporal queries using finite domain constraint logic.<\/li>\n<li>Enable the sharing of data using linked open data formats RDF and JSON-LD making scientific or organisational information sharing easy.<\/li>\n<li>Help you automate the production of UI and data-entry by <em>knowing<\/em> what data <em>means<\/em>.<\/li>\n<\/ul><h2><a id=\"user-content-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#client\"><\/a>Client<\/h2><p>In order to begin manipulating and querying your database you can simply point your browser to <code>TERMINUS_DIR\/index.html<\/code> which should allow you to manage and query your database.<\/p><p>You can also perform all configuration, querying and management functions via the RESTful API. We have a number of client libraries to help you get started:<\/p><p>The <a href=\"https:\/\/github.com\/terminusdb\/terminus-client\">JavaScript client libraries<\/a> give you assistance with programatic access.<\/p><p>The <a href=\"https:\/\/github.com\/terminusdb\/terminus-dashboard\">JavaScript dashboard<\/a> is an example application allowing manipulation of the database, queries and .<\/p><p>The [Python client libraries] Coming Soon...<\/p><h2><a id=\"user-content-changes-in-this-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#changes-in-this-version\"><\/a>Changes in this Version<\/h2><ul><li>Now with our own <a href=\"https:\/\/github.com\/terminusdb\/terminus-store\">terminus-store<\/a> backend!<\/li>\n<li>Extended WOQL to greater functionality for string and arithmetic processing<\/li>\n<li>Better date handling in WOQL<\/li>\n<li>Improved type casting in WOQL<\/li>\n<li>Improvements to query processing speed<\/li>\n<li>We now have decoupled public URI from internal database URI<\/li>\n<\/ul><h2><a id=\"user-content-future\" class=\"anchor\" aria-hidden=\"true\" href=\"#future\"><\/a>Future<\/h2><p>We are working hard on our collaboration and data-ops features. They should be available by the end of February 2020.<\/p><h2><a id=\"user-content-todo\" class=\"anchor\" aria-hidden=\"true\" href=\"#todo\"><\/a>TODO<\/h2><ul><li>Improve error handling and logging<\/li>\n<li>Convert frames into valid JSON-LD<\/li>\n<\/ul><h2><a id=\"user-content-contact\" class=\"anchor\" aria-hidden=\"true\" href=\"#contact\"><\/a>Contact<\/h2><p>You can write the authors, or connect to us on FreeNode IRC on the <code>#TerminusDB<\/code> channel.<\/p><h2><a id=\"user-content-authors\" class=\"anchor\" aria-hidden=\"true\" href=\"#authors\"><\/a>Authors<\/h2><p>Gavin Mendel-Gleason <a href=\"mailto:gavin@datachemist.com\">gavin@datachemist.com<\/a><\/p><p>Matthijs van Otterdijk <a href=\"mailto:matthijs@datachemist.com\">matthijs@datachemist.com<\/a><\/p><p>Robin de Rooij <a href=\"mailto:robin@datachemist.com\">robin@datachemist.com<\/a><\/p><p>Anne Ogborn <a href=\"mailto:annie@datachemist.com\">annie@datachemist.com<\/a><\/p><p><a href=\"https:\/\/github.com\/pmoura\">Paulo Moura<\/a><\/p><p>Put your Name here by contributing!<\/p><h2><a id=\"user-content-copyright\" class=\"anchor\" aria-hidden=\"true\" href=\"#copyright\"><\/a>Copyright<\/h2><p>This file is part of TerminusDB.<\/p><p>TerminusDB is free software: you can redistribute it and\/or modify it under the terms of the GNU General Public License version 3 (GPLv3) as published by the Free Software Foundation.<\/p><p>TerminusDB is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.<\/p><p>You should have received a copy of the GNU General Public License along with TerminusDB. If not, see <a href=\"https:\/\/www.gnu.org\/licenses\/\" rel=\"nofollow\">https:\/\/www.gnu.org\/licenses\/<\/a>.<\/p>","created_at":"2020-02-01T23:08:46+0000","updated_at":"2020-02-01T23:08:46+0000","published_by":["terminusdb"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/53230339?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","makefile"],"is_public":false,"id":1121,"title":"A Makefile for Emacs Packages \u00ab null program","url":"https:\/\/nullprogram.com\/blog\/2020\/01\/22\/","content":"<time datetime=\"2020-01-22\">January 22, 2020<\/time><p>nullprogram.com\/blog\/2020\/01\/22\/<\/p><p>Each of my Emacs packages has a Makefile to byte-compile all source files, run the tests, build a package file, and, in some cases, run the package in an interactive, temporary, isolated Emacs instance. These <a href=\"https:\/\/nullprogram.com\/blog\/2017\/08\/20\/\">portable Makefiles<\/a> have a similar structure and follow the same conventions. It would require more thought and feedback before I\u2019d try to make it a <em>standard<\/em>, but these are conventions I\u2019d like to see in other package Makefiles.<\/p><p>Here\u2019s an incomplete list of examples:<\/p><ul><li><a href=\"https:\/\/github.com\/skeeto\/bitpack\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/bitpack\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/cplx\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/cplx\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/devdocs-lookup\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/devdocs-lookup\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/elfeed\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/elfeed\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/emacs-aio\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/emacs-aio\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/emacs-bencode\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/emacs-bencode\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/emacs-memoize\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/emacs-memoize\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/emacs-web-server\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/emacs-web-server\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/impatient-mode\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/impatient-mode\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/lcg128\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/lcg128\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/nasm-mode\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/nasm-mode\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/skewer-mode\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/skewer-mode\/blob\/master\/Makefile<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/skeeto\/x86-lookup\/blob\/master\/Makefile\">https:\/\/github.com\/skeeto\/x86-lookup\/blob\/master\/Makefile<\/a><\/li>\n<\/ul><p>You should make a habit of compiling your Emacs Lisp files even if you don\u2019t think you need the performance. The byte-compiler, while <a href=\"https:\/\/nullprogram.com\/blog\/2019\/02\/24\/\">dumb<\/a>, does <a href=\"https:\/\/nullprogram.com\/blog\/2016\/12\/22\/\">static analysis<\/a> and may spot bugs and other issues early.<\/p><p>First things first: Every portable Makefile starts with a special target, <code class=\"language-plaintext highlighter-rouge\">.POSIX<\/code>, to request standard behavior. This is followed by macro definitions. When compiling a C program, the <code class=\"language-plaintext highlighter-rouge\">CC<\/code> macro is the name of the compiler. Analogously, when compiling Emacs packages the <code class=\"language-plaintext highlighter-rouge\">EMACS<\/code> macro is the name of the Emacs program.<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>.POSIX:\nEMACS = emacs\n<\/pre><\/div><p>Users can now override the macro to specify alternate Emacs binaries. I use this all the time to test my packages under different versions of Emacs.<\/p><div class=\"language-plaintext highlighter-rouge highlight\"><pre>$ make clean\n$ make EMACS=emacs-24.3 check\n$ make clean\n$ make EMACS=emacs-25.1 check\n<\/pre><\/div><p>Note: It\u2019s common to use <code class=\"language-plaintext highlighter-rouge\">?=<\/code> assignment here, but that is both non-standard and unnecessary. If you want to override macro definitions from the environment, use the <code class=\"language-plaintext highlighter-rouge\">-e<\/code> option:<\/p><div class=\"language-plaintext highlighter-rouge highlight\"><pre>$ export EMACS=emacs-24.3\n$ make -e\n<\/pre><\/div><p>The first non-special target in the Makefile is the default target. For Emacs packages, this target should byte-compile all the source files, including tests. List the byte-compiled file names as the target dependencies:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>compile: foo.elc foo-test.elc\n<\/pre><\/div><p>Now for the tedious part: Define the dependencies between your different source files. It would be nice to automate this part somehow, but fortunately most packages just aren\u2019t that complicated. You do not need to list trivial dependencies \u2014 i.e. mapping each .el file to its .elc file \u2014 since make will figure that out on its own.<\/p><p>Since <code class=\"language-plaintext highlighter-rouge\">foo-test.elc<\/code> relies on <code class=\"language-plaintext highlighter-rouge\">foo.elc<\/code> \u2014 it\u2019s testing this file after all \u2014 the relationship must be indicated to make. For single file packages (one package file, one test file), this is all that\u2019s needed:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>foo-test.elc: foo.elc\n<\/pre><\/div><p>I call my testing targets \u201ccheck\u201d and this target must depend on the byte-compiled files containing tests. It will transiently depend on the other package source files because of the previous section.<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>check: foo-test.elc\n    $(EMACS) -Q --batch -L . -l foo-test.elc -f ert-run-tests-batch\n<\/pre><\/div><p>The <code class=\"language-plaintext highlighter-rouge\">-Q<\/code> option runs Emacs with \u201cminimum customizations.\u201d The <code class=\"language-plaintext highlighter-rouge\">-L .<\/code> option puts the current directory in the load path so that <code class=\"language-plaintext highlighter-rouge\">(require 'foo<\/code>) will work. Finally it loads the file containing the tests and instructs ERT to run all defined tests.<\/p><p>A good build can clean up after itself:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>clean:\n    rm -f foo.elc foo-test.elc\n<\/pre><\/div><p>Finally we need one more thing to tie it all together: an inference rule to teach make how to compile .elc files from .el files.<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>.SUFFIXES: .el .elc\n.el.elc:\n    $(EMACS) -Q --batch -L . -f batch-byte-compile $&lt;\n<\/pre><\/div><p>This is similar to the \u201ccheck\u201d target, but compiles a source file instead of running tests.<\/p><p>For simple, single source file packages, this is all you need!<\/p><h3 id=\"complex-packages\">Complex packages<\/h3><p>My most complex package is Elfeed which has 10 source files and 4 test files. It also includes a target to build a package file, which I would upload to Marmalade when it was still functioning. I did a few extra things to keep this tidy.<\/p><p>First, I define the package version in the Makefile:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>VERSION = 1.2.3\n<\/pre><\/div><p>It would be nice to grab this information from a reliable place (Git tag, source file, etc.), but I never found a reliable and satisfactory way to do this. Simple wins.<\/p><p>To avoid repeating myself, I list the source files in a macro as well:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>EL   = foo-a.el foo-b.el foo-c.el\nDOC  = README.md\nTEST = foo-test.el\n<\/pre><\/div><p>These will still need to have all their interdependencies individually defined for make. For example, if C depends on both A and B, but neither A nor B depend on each other, this is all you\u2019d need:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>foo-c.elc: foo-a.elc foo-b.elc\n<\/pre><\/div><p>Done correctly you can perform parallel builds with the non-standard but common <code class=\"language-plaintext highlighter-rouge\">-j<\/code> make option. This is pretty nice since Emacs can\u2019t do parallel builds itself.<\/p><p>I use the file list macros in the \u201ccompile\u201d and \u201ccheck\u201d targets:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>compile: $(EL:.el=.elc) $(TEST:.el=.elc)\ntest: $(TEST:.el=.elc)\n<\/pre><\/div><p>The \u201cpackage\u201d target copies everything under a directory and tars it up. The directory is removed first, if it exists, so that any potenntial leftover garbage from doesn\u2019t get included.<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>package: foo-$(VERSION).tar\nfoo-$(VERSION).tar: $(EL) $(DOC)\n    rm -rf foo-$(VERSION)\/\n    mkdir foo-$(VERSION)\/\n    cp $(EL) $(DOC) foo-$(VERSION)\/\n    tar cf $@ foo-$(VERSION)\/\n    rm -rf foo-$(VERSION)\/\n<\/pre><\/div><p>In Elfeed, the target to test in an interactive, temporary Emacs instance is called \u201cvirtual\u201d. In Skewer it\u2019s called \u201crun\u201d. The name of the target and the specific rules will depend on the package, should you even want this target at all. It\u2019s handy to have the option test without my own configuration contaminating Emacs, and vice versa. When people report issues, I can also direct them to reproduce their issue in the clean environment.<\/p><p>Here\u2019s what a simple \u201crun\u201d target might look like:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>run: $(EL:.el=.elc)\n    $(EMACS) -Q -L . -l foo-c.elc -f foo-mode\n<\/pre><\/div><p>Make is not really designed to run interactive programs like this, but it works in practice.<\/p><h3 id=\"dependencies\">Dependencies<\/h3><p>What about packages with dependencies? I\u2019ve used <a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a> in the past but was never satisfied, especially when integrating it into a Makefile. So, again, I\u2019ve opted for the dumb-but-reliable option: request that dependencies are cloned in adjacent directories matching the dependency\u2019s package name. For example, the <a href=\"https:\/\/nullprogram.com\/blog\/2014\/02\/06\/\">EmacSQL<\/a> Makefile header:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre># Clone the dependencies of this package in sibling directories:\n#     $ git clone https:\/\/github.com\/cbbrowne\/pg.el ..\/pg\n<\/pre><\/div><p>I also define a new \u201clinker flags\u201d macro, <code class=\"language-plaintext highlighter-rouge\">LDFLAGS<\/code>. Like with <code class=\"language-plaintext highlighter-rouge\">EMACS<\/code>, this lets users override it if needed:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>LDFLAGS = -L ..\/pg\n<\/pre><\/div><p>Everywhere I use <code class=\"language-plaintext highlighter-rouge\">-L .<\/code> I also include <code class=\"language-plaintext highlighter-rouge\">$(LDFLAGS)<\/code>. For example, in the inference rule:<\/p><div class=\"language-make highlighter-rouge highlight\"><pre>.SUFFIXES: .el .elc\n.el.elc:\n    $(EMACS) -Q --batch -L . $(LDFLAGS) -f batch-byte-compile $&lt;\n<\/pre><\/div><p>If the dependencies follow these conventions, then these can also be compiled in a recursive way with little effort:<\/p><div class=\"language-plaintext highlighter-rouge highlight\"><pre>$ make -C ..\/pg\n<\/pre><\/div><p>I\u2019m not completely satisfied with this solution, particularly since it\u2019s an odd burden on anyone using the Makefile, but it\u2019s worked well enough for my needs. This is when I wish Emacs had <a href=\"https:\/\/nullprogram.com\/blog\/2020\/01\/21\/#package-management\">distributed package management<\/a>.<\/p>","created_at":"2020-01-29T14:42:43+0000","updated_at":"2020-01-29T14:42:47+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"nullprogram.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1120,"title":"GNU Recutils","url":"https:\/\/labs.tomasino.org\/gnu-recutils\/","content":"<figure><img src=\"https:\/\/labs.tomasino.org\/assets\/images\/gnu-recutils.png\" alt=\"Fred and George, Turtles in love\" referrerpolicy=\"no-referrer\" \/><figcaption>Fred and George, Turtles in love\n<\/figcaption><\/figure><p>There are hundreds of cool command line tools that have been made over the years built on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Unix_philosophy\">unix philosophy<\/a>. Plain text is powerful, ubiquitous, and human-centric. It was also the only option for computing for quite a while. It\u2019s no surprise that some of the most sophisticated tools are focused on it.<\/p><p>One such package is GNU Recutils, a set of tools and libraries to access human-editable, plain text databases called recfiles. Using plain text that is <strong>primarily<\/strong> human-readable as a database is just frankly cool as hell.<\/p><p>Recutils is a collection of tools, like <code>recins<\/code>, <code>recdel<\/code>, and <code>recsel<\/code> used to manage these recfiles\/databases. They allow for all the normal basic relational database operations, typing, auto-incrementing, and even field-level crypto. All of this power is yours with the bonus that your database is a human-readable text file that you can grep\/awk\/sed freely, and a line-oriented structure makes it perfect for version control systems.<\/p><h2 id=\"recfiles\">Recfiles<\/h2><div class=\"highlight\"><pre class=\"language-recfile\" data-lang=\"recfile\">%rec: Book\n%mandatory: Title\n%type: Location enum loaned home unknown\n%doc:\n+ A book in my personal collection.\nTitle: GNU Emacs Manual\nAuthor: Richard M. Stallman\nPublisher: FSF\nLocation: home\nTitle: The Colour of Magic\nAuthor: Terry Pratchett\nLocation: loaned\nTitle: Mio Cid\nAuthor: Anonymous\nLocation: home\nTitle: chapters.gnu.org administration guide\nAuthor: Nacho Gonzalez\nAuthor: Jose E. Marchesi\nLocation: unknown\nTitle: Yeelong User Manual\nLocation: home<\/pre><\/div><p>A recfile begins with an optional structure definition using <code>%<\/code> prefixes. There\u2019s a lot of power available in this, including the defining of relations between tables, constraints, types, and more. You can use the built-in type definitions or define your own using regular expressions (plain text woo).<\/p><p>Following this header bit, the individual records are listed with a blank line separating entries. Have a look at how amazingly readable the data is!<\/p><p>Here\u2019s a more sophisticated example that shows a relationship in action:<\/p><div class=\"highlight\"><pre class=\"language-recfile\" data-lang=\"recfile\">%rec: Service\n%key: Id\n%auto: Id\n%type: Provider rec Provider\n%type: Mileage int\n%sort: Mileage\n%mandatory: Id Date Mileage Description Provider\n%rec: Provider\n%key: Id\n%auto: Id\n%mandatory: Id Name<\/pre><\/div><p><em>Car service examples courtesy of <a href=\"https:\/\/gopher.tildeverse.org\/sdf.org\/0\/users\/dbucklin\/posts\/maintenance.txt\">dbucklin\u2019s gopher post<\/a>.<\/em><\/p><h2 id=\"queries-other-operations\">Queries &amp; other operations<\/h2><div class=\"highlight\"><pre class=\"language-bash\" data-lang=\"bash\">$ recsel -e \"Location = 'loaned'\" -P Title books.rec\nThe Colour of Magic<\/pre><\/div><p>Querying your recfile database is usually done with <code>recsel<\/code> (recfile select). You specify the table to query, expressions, fields to return, and the file and the output comes to standard out.<\/p><p>To insert data you use the <code>recins<\/code> (recfile insert) command:<\/p><div class=\"highlight\"><pre class=\"language-bash\" data-lang=\"bash\">$ recins -t Task              \\\n    -f Id -v 10               \\\n    -f Title  -v \"New issue.\" \\\n    -f Status -v NEW          \\\n    TODO.rec<\/pre><\/div><p>Again it\u2019s a simple syntax. You provide the table to insert into, a list of fields &amp; values (-f &amp; -v, respectively), and the filename. If you prefer, you can use the <code>-r<\/code> switch to provide the field &amp; value as one string.<\/p><p>How about deleting records? Can you guess the syntax?<\/p><div class=\"highlight\"><pre class=\"language-bash\" data-lang=\"bash\">$ recdel -c -t Task -e \"Status = 'CLOSED'\" TODO.rec<\/pre><\/div><p>Yep, it\u2019s pretty much the same as the select, as one would expect.<\/p><h2 id=\"what-else\">What else?<\/h2><p>There\u2019s other convenient tools, like the ability to convert to and from CSV formats, alternative output modes that make it work well with awk, password-protected fields, and so on. The package is quite robust and well maintained by Jose E. Marchesi, who can be found in the #recutils room on irc.freenode.net.<\/p><p>The <a href=\"https:\/\/www.gnu.org\/software\/recutils\/manual\/\">full documentation<\/a> is seriously impressive. Check it out.<\/p><p>I\u2019ll be adopting it into my tools immediately!<\/p><p><em>(yes, the offical package image is <a href=\"https:\/\/www.gnu.org\/software\/recutils\/faq.html#whyturtles\">two gay turtles<\/a>)<\/em><\/p>","created_at":"2020-01-26T19:12:13+0000","updated_at":"2020-01-26T19:12:13+0000","published_by":["James Tomasino"],"annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"labs.tomasino.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1119,"title":"Tangent: The Automated Dungeon Master","url":"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/","content":"<p>This post is the first in a probable series of \u2018tangents\u2019, not part of a continuing series like The Switch or The Backbone. In fact, this particular tangent veers well off of this blog\u2019s normal subject matter, as it deals primarily with fantasy role-playing games.<\/p>\n<h2>The Magic of Dungeons and Dragons<\/h2>\n<p>Since I was a young child in the 1980s, I have been fascinated by Dungeons and Dragons (D&amp;D). My older brother at some point had an interest in the game but evidently lost it, and bequeathed to me a smattering of box sets and hardback manuals from Tactical Studies Rules (TSR), the company that produced the game. Most immediately accessible and captivating was the vivid scarlet D&amp;D Basic Set, emblazoned with Larry Elmore\u2019s painting of a warrior with a glowing sword confronting a fearsome dragon, crouched atop its horde of ill-gotten gains. It\u2019s hard to explain how much excitement and wonder I derived from this little cardboard box and the saddle-stitched rules manuals that it contained. D&amp;D gave structure and consistency to the kind of make-believe play that I, like so many other children, already instinctively engaged in.<\/p>\n<p>D&amp;D developed within the culture of competitive miniature wargaming, and can be played as something like a traditional game, with a goal against which victory or defeat is measured. In this style of play, much like in many of the wargames from which it derived, one player is assigned the role of referee (called the Dungeon Master), and given the responsibility for resolving the actions made by the other players. TSR regularly ran tournament scenarios at conventions in this mode. Whatever group of adventurers best survived the tricks, traps, and monsters found within the dungeon and accumulated the most treasure was declared the winner.<\/p>\n<p>The truest magic of D&amp;D, and other role-playing games (RPGs)<sup id=\"fnref-13578-otherrpgs\"><a href=\"#fn-13578-otherrpgs\" class=\"jetpack-footnote\">1<\/a><\/sup>, however appears only when played in a second-order mode, where the same players and Dungeon Master (DM) meet up repeatedly (every week, for example), to play one continuous game over the course of many sessions. In keeping with the game\u2019s wargaming roots, this iterated game is known as a <em>campaign.<\/em> Just as a military campaign consists of an extended series of movements and battles by a single army, so a role-playing campaign consists of repeated excursions into the same imaginary world. There might be a long-term goal by which the players \u201cwin\u201d the game, but there can never be a loser. The main point is not pursuing a winning strategy, but experiencing adventures in a place that never was. The DM\u2019s role as referee and adjudicator therefore becomes secondary to his role as world creator and simulator.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_13654\" aria-describedby=\"caption-attachment-13654\" class=\"wp-caption aligncenter c1\"><img data-attachment-id=\"13654\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/city_state_of_the_invincible_overlord_map\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg\" data-orig-size=\"1086,843\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;david&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;1&quot;}\" data-image-title=\"City_State_of_the_Invincible_Overlord_map\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=300\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=739\" class=\"size-full wp-image-13654 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=739\" alt=\"City_State_of_the_Invincible_Overlord_map\" srcset=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=739 739w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=150 150w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=300 300w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=768 768w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg?w=1024 1024w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/city_state_of_the_invincible_overlord_map.jpg 1086w\" referrerpolicy=\"no-referrer\" \/><figcaption id=\"caption-attachment-13654\" class=\"wp-caption-text\">A portion of the map of the\u00a0<em>City-State of the Invincible Overlord\u00a0<\/em>(1976), one of the first published D&amp;D settings, and already showing the desire to create a fully realized world for players to inhabit.<\/figcaption><\/figure><h2>Secondary Worlds And Their Discontents<\/h2>\n<p>And now you must allow me a tangent within a tangent, to veer off into a discussion of imaginary worlds in literature. After all, the desire to recreate the kind of battles and adventures found in their favorite fantasy literature is what spurred Dave Arneson and Gary Gygax to invent D&amp;D in the first place. The gold standard for imaginative expansiveness in such literature, is, of course, J.R.R. Tolkien\u2019s Middle-Earth<sup id=\"fnref-13578-danddtolkien\"><a href=\"#fn-13578-danddtolkien\" class=\"jetpack-footnote\">2<\/a><\/sup>. Tolkien spent decades refining the outlines and deepening the detail of the mythos he had created, intending to develop a wholly English equivalent to the Nordic eddas or Greek epics. He called the new world that he had brought forth from within his mind a sub-creation (within God\u2019s primary creation) or <em>secondary world<\/em>.<\/p>\n<p>When one reads\u00a0<em>The Lord of the Rings<\/em>, one encounters not just a story, but a whole imaginary place in which that story unfolds. Middle-Earth is not, of course, the first imaginary world in fiction. But prior to Tolkien, such places were generally flat and tissue-thin. They existed as a means to the ends of allegory or satire, or as the backdrop to a children\u2019s story with no pretense to realism, like Frank Baum\u2019s Oz with its four symmetrical lands, each overseen by its resident witch and inhabited by a cutesy tribe with monochrome clothes and houses. In Tolkien, by contrast, one finds a land populated with richly developed kingdoms and peoples, each with its own history and culture, its own myths and songs. His stories feel like true tales from a place that never was.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_13653\" aria-describedby=\"caption-attachment-13653\" class=\"wp-caption aligncenter c2\"><img data-attachment-id=\"13653\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/attachment\/00428\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg\" data-orig-size=\"952,1200\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"00428\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=238\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=739\" class=\"wp-image-13653 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=494&amp;h=623\" alt=\"00428\" width=\"494\" height=\"623\" srcset=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=494&amp;h=623 494w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=119&amp;h=150 119w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=238&amp;h=300 238w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg?w=768&amp;h=968 768w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/00428.jpg 952w\" referrerpolicy=\"no-referrer\" \/><figcaption id=\"caption-attachment-13653\" class=\"wp-caption-text\">tb J.R.R. Tolkien, \u201cBilbo woke up early with the sun in his eyes\u201d<\/figcaption><\/figure><p>The creators of D&amp;D immediately began creating their own secondary worlds as settings for the adventures of their players \u2013 Gary Gygax developing <em>Greyhawk<\/em>, and Dave Arneson\u00a0<em>Blackmoor<\/em>. As the game\u2019s popularity spread, other DMs followed by creating their own worlds (often drawing from literary inspiration), or based their campaigns on published settings. For example, the phenomenally popular <em>Forgotten Realms<\/em>, published by TSR in 1987, and based on an imaginary world developed by Ed Greenwood since his childhood. Immersion in rich a secondary world of this sort is what transforms a D&amp;D campaign from a fun romp into a vicarious literary tale. Rather than simply <em>reading<\/em> <em>about<\/em> the adventures of Frodo, Legolas, Aragorn, and the rest, one can <em>experience<\/em> such an adventure, exploring first-hand the wonders of a place like Middle Earth.<\/p>\n<p>When D&amp;D is played in a tournament setting, it is understood implicitly by all parties that there is nothing beyond the pre-set scenario. One cannot leave the dungeon and decide to go do something else. The only choice players have is in how they will advance through it. A campaign, on the other hand, grants open-ended agency to the players within the DM\u2019s secondary world. Each player can create their own persona, known as a <em>character<\/em>, to inhabit for the duration of the campaign, be it weeks or months or years. We will call this responsiveness to player agency\u00a0<em>openness<\/em>.<\/p>\n<p>Some traditional board games, such as <em>Cluedo<\/em>\/<em>Clue,<\/em> have allowed players to assume the role of a particular character, but they have very little openness, providing only a relatively minuscule decision space in which to operate. In a D&amp;D campaign, a player can choose to take <em>any<\/em> plausible action, within the context of the world that their character inhabits. Your character comes across a mysterious stranger in the village common? You can strike up an idle conversation, pilfer their purse, start a fight, or attempt to win their affection. Whatever you can think of you can do, and it is left to the DM to decide the results of your actions, with consequences that might ripple out across that village, its surrounding countryside, or even the entire world.<\/p>\n<p>And that secondary world, swirling within the luminiferous ether that surrounds and penetrates the minds of players and DM, feels real to the players as their characters explore it, because it is <em>self-consistent<\/em>. The world persists,\u00a0though gradually changing, even when the player characters are not around. They might leave that village after assaulting the mysterious stranger, and return to find that he has turned the villagers against them with scurrilous lies about their sordid deeds. Or they might find instead that a band of orcs has ransacked the village and killed their favorite barkeep. Over the course of the campaign, players could establish a mercantile empire, build a castle, or even found a kingdom, and the world will respond accordingly. Moreover, the world\u2019s various parts form a harmonious whole across time and space. If the villagers follow a particular set of customs, it is likely that the castle just down the river does as well \u2013 or perhaps not, if the castle-dwellers are recent conquerors of the region, Norman <em>seigneurs<\/em> ruling over Saxon churls. A well-crafted and well-run game world, like Tolkien\u2019s Middle Earth, has the feeling of a lived-in place, with natural contours of culture and politics carved out over the centuries, from the flow of the currents of history across the bedrock of geography.<\/p>\n<p>Yes, well-run and well-crafted, there\u2019s the rub. Tolkien had two decided advantages over the poor, put-upon dungeon master in realizing his masterpiece. First, he only had to paint in the details along one path through his creation. Despite his many decades of labor, at his death most regions of his world still consisted of little more than names on a map. While the reader of <em>The Lord of the Rings<\/em> gets to visit Rivendell, Edoras, and Minas Tirith, she learns nothing about Harlindon, Rh\u00fbn or Anfalas, and about those regions Tolkien was nearly as ignorant as his reader. His characters politely steered clear of these uncharted regions. Player characters are rarely so obliging. Secondly, Tolkien wrote and revised his work over the course of years, and was able to pause for many minutes, or even hours, to consult prior chapters, maps, or other reference materials before deciding what would happen next. Even then, he could later return to the same passage and revise his decision, months or years later. A dungeon master, however, must respond to character actions as they come, usually within seconds, or the game session very quickly becomes tedious for his players. Both differences come down to agency \u2013 player characters have got it, literary characters don\u2019t.<\/p>\n<p>To be able to simulate a self-consistent secondary world on the fly in this fashion is a rare feat of skill and effort. As circumstance demands, the DM must serve as a geographer, demographer, economist, physicist, and more. It strains the muscles of improvisation \u2013 how many different personalities and physiogonomies can one devise to define the innkeepers of each and every village through which the campaigners pass? To say nothing of the minstrels, mercenaries, miscreants, and so forth? And then, having established the facts of some place, the burden shifts to the memory \u2013 just what was the name of that tavern in Elkenburg, again? Meticulous note-taking is a great help, of course. Nonetheless, the effort required grows and grows as the players accumulate a history of people met and places visited.<sup id=\"fnref-13578-hexcrawl\"><a href=\"#fn-13578-hexcrawl\" class=\"jetpack-footnote\">3<\/a><\/sup><\/p>\n<p>The fact is that few are the dungeon masters who can create such verisimilitude in the face of total, open-ended player agency. Instead, most fall back on what is known pejoratively as <em>railroading<\/em> \u2013 preparing events and sites, like Tolkien, along a predetermined right-of-way, and steering the players along it. The rest of the campaign world can serve as a feeble Potemkin village, a mere facade that will collapse at the first touch. Railroading gets its bad reputation from deplorable DM actions such as hurling ever more impossible monsters at the characters if they deviate from their intended path \u2013 or, as I experienced once as a player in my youth, throwing an invisible and impenetrable force field between them and the rest of the world. But, in practice, it manifests itself far more often as a soft social contract. The players know that the DM has invested hours of preparation into developing a particular scenario, and they go along for the good of everyone involved<sup id=\"fnref-13578-megadungeon\"><a href=\"#fn-13578-megadungeon\" class=\"jetpack-footnote\">4<\/a><\/sup>. Whether hard or soft, railroading ameliorates the problem of on-the-fly world creation by constraining the players\u2019 agency within a limited sphere of action.<\/p>\n<p>But even a soft railroad is a huge investment of time for the DM to prepare and requires considerable skill to execute well. Some DMs opt instead for a \u201cno prep\u201d style of play with neither a world nor a scenario prepared in advance. But to do this successfully also requires its own set of skills \u2013 mainly improvisational \u2013 and usually a great deal of experience. D&amp;D can indeed be a magical experience, but a would-be player may be hard-pressed to find a DM with the time and talent necessary to truly enchant. That is to say nothing of finding like-minded fellow players. Nothing can spoil a serious fantasy epic more quickly than a so-called friend who can\u2019t help but crack \u201cyo momma\u201d jokes every five minutes. The reality of a D&amp;D session thus often pales in comparison to what one imagines it could be, indeed, should be.<\/p>\n<h2>Paper Worlds<\/h2>\n<p>Given the difficulties of dungeon mastery, it did not take long after the invention of D&amp;D for players and game publishers alike to begin looking for ways to capture the magic of that game without the need for a referee \u2013 or, in fact, for other players. That is to say, to automate the dungeon master. Wouldn\u2019t it be something to become immersed into an open, self-consistent secondary world, any time, any place, all by yourself \u2013 as an active agent, not just a passive reader? For a long time, the most popular and readily accessible way to find such a \u201cmechanical dungeon master\u201d was through a kind of paper flow chart called a gamebook.<\/p>\n<p>In 1975, Ken St. Andre, an Arizona State University (ASU) grad in his late twenties, had fallen in love with the idea of D&amp;D, but was unhappy with the complexity of its rules. He therefore decided to publish one of the very first alternative fantasy role-playing games, <em>Tunnels and Trolls.<\/em> He distributed copies he printed himself at the ASU print shop before finding a publisher, a one-man operation in Scottsdale called Flying Buffalo. Flying Buffalo\u2019s founder, Rick Loomis, founded the company to moderate play-by-mail multiplayer sessions of a game he had invented called <em>Nuclear Destruction<\/em>, collecting a fee from each participant. He entered the publishing business after he acquired the rights to a card game (unrelated except for its theme) called <em>Nuclear<\/em> <em>War<\/em>. Loomis was happy to sell St. Andre\u2019s remaining copies of <em>Tunnels &amp; Trolls<\/em>, and, when the initial copies sold out quickly, Loomis acquired the license to publish additional print runs himself, under the Flying Buffalo imprint.<\/p>\n<p>In the spring of the following year, a player in St. Andre\u2019s <em>Tunnels &amp; Trolls<\/em> games named Steve McCallister suggested the idea of a solitaire dungeon adventure. He got the idea from the programmed instruction books that were popular at the time, which offered multiple choice questions, and then had the reader turn to the back for immediate feedback on the chosen answer. Loomis liked the idea and wrote the first solitaire role-playing adventure, <em>Buffalo Castle<\/em>, which he published in May 1976.\u00a0<\/p>\n<p><em>Buffalo Castle<\/em> consisted of about 150 paragraphs (each identified by a page number and a letter), each containing a few terse sentences of description and options for the player to continue to one of several other paragraphs<sup id=\"fnref-13578-htmlversion\"><a href=\"#fn-13578-htmlversion\" class=\"jetpack-footnote\">5<\/a><\/sup>. For example, \u201cYou have entered Room Six. There is a large fountain in the middle of the room. You may drink from it if you wish.If you take a drink, go to 8C. If you wish to leave by the north door, go to 4C. If you wish to leave by the east door, go to 16D.\u201d The player character starts off faced with the choice of three doors by which to enter the castle, and proceeds from there. The adventure uses elements of the <em>Tunnels &amp; Trolls\u00a0<\/em>rules, especially combat, but forbids the use of magic, and all the complexity that entails. The castle is unforgiving, and most characters will likely die, but a lucky hero might manage to escape with some valuable plunder.\u00a0<\/p>\n<p><em>Buffalo Castle<\/em> did not attract much attention, but later in the 1970s, a series of books called <em>Choose Your Own Adventure<\/em> (CYOA) did, selling hundreds of millions of copies over the ensuing decades. (Though the predecessors of <em>Choose Your Own Adventure<\/em> date back to the 1960s,\u00a0the available evidence suggests that\u00a0<em>Buffalo Castle\u00a0<\/em>was developed entirely independently). While Flying Buffalo<em>\u00a0<\/em>continued to produce solitaire <em>Tunnels &amp; Trolls<\/em> adventures, not until 1982 did the D&amp;D-style gamebook break out into a wider market. In that year, Steve Jackson and Ian Livingston decided to try their hand at fusing a role-playing adventure with the CYOA formula. Jackson and Livingston were D&amp;D fans whose company Games Workshop, served for several years as the game\u2019s European distributor. They did not try to publish their new series themselves, however. Aiming for a wider market than the hobby press, they sold the rights to Puffin Books. The series they created, <em>Fighting Fantasy<\/em>, blended narrative, exploration, and game in a similar fashion to <em>Buffalo Castle<\/em>, but in self-contained, mass-market paperbacks.<\/p>\n<p>The series<em>\u00a0<\/em>was a resounding success, selling three million copies by 1985.<sup id=\"fnref-13578-million\"><a href=\"#fn-13578-million\" class=\"jetpack-footnote\">6<\/a><\/sup> A whole sub-genre of gamebooks followed, all attempting to emulate the experience of role-playing games, primarily D&amp;D. Successful series included <em>Wizards, Warriors, &amp; You,\u00a0<\/em><em>GrailQuest<\/em>,\u00a0<em>Lone Wolf<\/em>, and <em>Sorcery!<\/em><sup id=\"fnref-13578-gamebooks\"><a href=\"#fn-13578-gamebooks\" class=\"jetpack-footnote\">7<\/a><\/sup>. The CYOA books had lacked any elements of a game, or of a persistent world. You simply made choices that lead you deterministically through a garden of forking paths, to one final outcome or another. <em>Fighting Fantasy<\/em> and its derivatives, on the other hand, reintroduced some or all of the D&amp;D-like elements from <em>Buffalo Castle<\/em>: player character statistics (such as health and strength); randomness, including dice-based combat; and an inventory of items, which could affect the player\u2019s combat ability.<\/p>\n<p><img data-attachment-id=\"13656\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/the_warlock_of_firetop_mountain_28first_edition29\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/the_warlock_of_firetop_mountain_28first_edition29.jpg\" data-orig-size=\"243,408\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"the_warlock_of_firetop_mountain_28first_edition29\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/the_warlock_of_firetop_mountain_28first_edition29.jpg?w=179\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/the_warlock_of_firetop_mountain_28first_edition29.jpg?w=243\" class=\"size-full wp-image-13656 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/the_warlock_of_firetop_mountain_28first_edition29.jpg?w=739\" alt=\"the_warlock_of_firetop_mountain_28first_edition29\" srcset=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/the_warlock_of_firetop_mountain_28first_edition29.jpg 243w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/the_warlock_of_firetop_mountain_28first_edition29.jpg?w=89 89w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<p>But the new wave of solitaire adventures devoted much more attention to creating a rich secondary world for the player to explore than <em>Buffalo Castle<\/em> had. Inventory items in\u00a0<em>Fighting Fantasy<\/em>, for example, can have other in-game consequences outside combat (a locked door, for example, that can be opened \u2013 leading to a new paragraph \u2013 only if you possess the right key). Rather than brief pamphlets, the newer books were thick paperbacks, typically with 300-400 paragraphs (two or three times more than <em>Buffalo Castle<\/em>). Each paragraph also consisted of more text, deepening the sense of immersion in a real environment. Here, for example, is how a dungeon room is described in <em>The Warlock of Firetop Mountain,\u00a0<\/em>the first of the\u00a0<em>Fighting Fantasy\u00a0<\/em>books:<\/p>\n<blockquote>\n<p>The locked door bursts open and a nauseating stench hits your nostrils. Inside the room the floor is covered with bones, rotting vegetation and slime. A wild-haired old man, clothed in rags, rushes at you screaming. His beard is long and grey, and he is waving an old wooden chair-leg. Is he simply insane as he appears, or has this been some kind of trap? You may either shout at him to try to calm him down (turn to 263) or draw your sword and attack him (turn to 353).<\/p>\n<\/blockquote>\n<p>(Spoiler: you should talk to the man, who is full of useful information).<\/p>\n<p>The <em>Lone Wolf\u00a0<\/em>series took the idea of an ongoing D&amp;D campaign and ran with it, allowing players to take a single character through a continuous story over dozens of books, carrying over skills and items acquired from one book to the next. The entire saga was set in Magnamund, a world devised by the author, Joe Dever, for his D&amp;D games. The apotheosis of the development of the gamebook into an immersive secondary world, however, was Dave Morris and Jamie Thomson\u2019s <em>Fabled Lands<\/em> series, published in 1995 and 1996, just as the gamebook trade was in decline. The series thus terminated early, after only six of a planned twelve volumes.<\/p>\n<p>Rather than a self-contained story arc, each published volume of <em>Fabled Lands<\/em> covers a region of the world, with its own towns, villages, castles, and wilderness areas to explore. The player can move between books by walking from region to region, by taking a ship, or even by teleportation via magical gates. There are many quests and adventures that one can discover, some of them contained within one book, others spanning the world. But the player is also free to ignore all that, and simply wander around and explore. A system of keywords also allows the world to change in response to the player\u2019s actions. For example, in the first book, <em>The<\/em> <em>War-Torn Kingdom<\/em><em>,<\/em> you can assassinate a pretender to the throne or help him ascend to the crown (assuming you don\u2019t ignore him altogether). Whichever you choose, you will gain certain keywords which cement your alliances in future interactions with either faction. This is not to mention the sub-systems in the game that let players join a religion, buy houses, acquire ships, and engage in seaborne trade. No one has ever come closer than this to creating a secondary world on paper that grants the kind of full, open-ended agency that a refereed game of D&amp;D can.<\/p>\n<p>And yet, it is still well short of the mark. At any given point in a gamebook, the player typically has only two to three different options. Even the richest nexus points in <em>Fabled Lands<\/em>, such as major cities, rarely have more than a half-dozen choices on offer. These don\u2019t come close to exhausting all the possibilities that an imaginary protagonist in the same situation would have. Consider the raving old man in his fetid room in <em>Firetop Mountain<\/em>. The player is given only two options \u2013 shout at him or run him through. It\u2019s not hard to come up with many more directions that a D&amp;D campaign could branch out to from this decision point. One could, for example, offer the old man a clean set of clothes, attempt to restrain him, or back quickly out of the room and shut the door. If you do choose to talk, <em>Firetop Mountain\u00a0<\/em>gives you no ability to direct the conversation and potentially alter how he responds to you. The man provides the same predetermined information every time. And there are no lasting consequences of the encounter. In the hands of a DM, the man might end up as an ally who accompanies the player characters through the dungeon, or you might track down his family and return him to the bosom of hearth and home. If you slay the man, that same family might instead track <em>you\u00a0<\/em>down; if you anger the man, he might follow you and attempt to steal your treasure. Within a gamebook, agency and openness are more severely curbed than all but the most infamous of railroads.\u00a0<\/p>\n<h2>Digital Worlds<\/h2>\n<p>Therefore, players began looking instead to computers for an automated dungeon master. Personal computers burst onto the marketplace in the late 1970s, and became a part of most middle class households in the U.S. by the middle of the 1990s. A computer program could obviously provide dynamic responses to player actions much more easily than a static printed work. It could, in theory truly simulate a secondary world, without all of the limitations of a paper flowchart, and without any cumbersome keywords or checkboxes.<\/p>\n<p>A glimmer of this promise appeared in one of the very first D&amp;D-inspired computer games, <em>Adventure<\/em>. Will Crowther, an engineer at Bolt, Beranek and Newman (BBN) and creator of some of the foundational software of the ARPANET, wrote the game for BBN\u2019s PDP-10 minicomputer in the mid-1970s. The game had no graphics (very few computer terminals could support them at the time anyway), so all interaction happened in textual form, just like a gamebook. Crowther wrote a parser for the game that accepted two word commands in the form \u201cverb noun\u201d. You could thus tell the computer, in plain English, what you wanted to do, and it would tell you the consequences of your action. But this was not quite the dream of the digital DM come true. Yes, you could type anything. But most of the time the computer would refuse to understand you. For example, the game starts with the following description:<\/p>\n<blockquote>\n<p>YOU ARE STANDING AT THE END OF A ROAD BEFORE A SMALL BRICK BUILDING.<br \/>\nAROUND YOU IS A FOREST. A SMALL STREAM FLOWS OUT OF THE BUILDING AND<br \/>\nDOWN A GULLY.<\/p>\n<\/blockquote>\n<p>The game will accept \u201cENTER BUILDING\u201d, \u201cDRINK WATER\u201d or \u201cGO SOUTH\u201d, but not \u201cCLIMB TREE\u201d, \u201cSWIM\u201d, \u201cBUILD FIRE\u201d, \u201cHUNT\u201d, \u201cWAIT NIGHT\u201d, etc.<\/p>\n<p><em>Adventure<\/em> is, in effect, a flowchart in disguise, one that hides its outgoing branches, forcing the player to guess at them instead. It spawned its own genre of computer games, the <em>adventure game.<\/em> Though some later games, notably <em>Zork<\/em>, which was co-authored by another member of Crowther\u2019s D&amp;D campaign, provided more sophisticated parsers, none provided a substantial leap in verisimilitude over the gamebook. By the 1980s, the genre was almost entirely divorced from its D&amp;D roots, focusing mainly on puzzles (usually combining inventory objects with the environment in some non-obvious way), rather than exploration, character development, or heroic exploits.<\/p>\n<p>The first computer RPGs to incorporate graphics appeared in the mid-1970s, on the PLATO IV system, a mainframe capable of supporting hundreds of graphical terminals developed at the University of Illinois. Shortly thereafter, similar titles reached a much wider on audience on the first personal computers. These games were typically pure dungeon crawlers \u2013 taking advantage of the gridded nature of dungeon corridors to simplify the problem of rendering graphics \u2013 and focused on combat, the easiest part of D&amp;D to rigidly codify in algorithmic form.<\/p>\n<p>Among the best of these early efforts was <em>Wizardry: Proving Grounds of the Mad Overlord<\/em>, written by two Cornell University students, Andrew Greenberg and Robert Woodhead. Greenberg and Woodhead had access to a PLATO terminal at Cornell, and borrowed heavily from PLATO precursors like <em>Dungeon<\/em> and <em>Oubliette<\/em> in developing their game. Released for the Apple II in 1981, <em>Wizardry<\/em> epitomized one model for the computer RPG \u2013 a series of battles in a nameless dungeon, with intermittent rests to recuperate. All of its interest derives from resource management, careful mapping, and combat tactics. There is no hint of a wider setting, and only the barest gesture towards giving meaning and motivation to the player\u2019s actions beyond killing everything in sight. The state of the dungeon does not even persist when the players return to the surface to shop for equipment, refilling instantly with the same monsters and treasure that they contained on the first visit.<\/p>\n<p>Another 1981 Apple II release, however, followed a different path, a path toward an immersive, digital secondary world. <em>Ultima<\/em> was written by a teenage D&amp;D fan from the Houston suburbs named Richard Garriott. It is, in fact, a digital reproduction of the world he had created for his D&amp;D campaign, which he called Sosaria. In the game, the player can receive quests from kings, delve into dungeons, visit towns to buy equipment, learn rumors while drinking in taverns, rescue princesses, acquire a vehicle, and even attempt to steal from the townsfolk. An overarching quest, to acquire four magic gems in order to travel back in time and defeat the evil wizard Mondain, ties the whole together.<\/p>\n<p><em>Ultima<\/em> is sketchy and weakly cohesive in many places, including a strange interlude where the player takes the controls of a spaceship and battles enemies that look suspiciously like TIE fighters. It is also highly structured and symmetric \u2013 each castle and town is identical in structure, and there are four continents, each with two castles, one dungeon, and one landmark. Yet for all that, it\u2019s an incredible achievement, a tiny imaginary world crammed onto two floppy disks by a 19-year-old University of Texas student.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_13657\" aria-describedby=\"caption-attachment-13657\" class=\"wp-caption aligncenter c3\"><img data-attachment-id=\"13657\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/ultima1-3\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima1-3.gif\" data-orig-size=\"401,248\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"ultima1-3\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima1-3.gif?w=300\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima1-3.gif?w=401\" class=\"size-full wp-image-13657 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima1-3.gif?w=739\" alt=\"ultima1-3\" referrerpolicy=\"no-referrer\" \/><figcaption id=\"caption-attachment-13657\" class=\"wp-caption-text\"><em>Ultima<\/em><\/figcaption><\/figure><p>As the\u00a0<em>Ultima\u00a0<\/em>series evolved, and Garriott built a company around its success, Sosaria evolved into Britannia, and the games developed ever greater depth and sophistication in both the richness of their setting and the interactivity and openness of their gameplay. It\u00a0 culminated in 1992 with <em>Ultima VII,\u00a0<\/em>built by a large team of specialized writers, programmers, and artists at Origin Systems in Austin, Texas. Origin put a huge amount of effort into the game\u2019s writing in order to give charm and character to every, well, character, in every corner of the world of Britannia. The landscape is littered with all sorts of side quests, little problems for the player character to solve independent of any progress toward the end of the game: a missing husband imprisoned for the theft of an apple, a thief disguised as a monk, two brothers in a dispute over religious belief.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_13658\" aria-describedby=\"caption-attachment-13658\" class=\"wp-caption aligncenter c4\"><img data-attachment-id=\"13658\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/ultima7-bg-exult\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png\" data-orig-size=\"320,240\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"ultima7-bg-exult\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png?w=300\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png?w=320\" class=\"wp-image-13658 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png?w=573&amp;h=430\" alt=\"ultima7-bg-exult\" width=\"573\" height=\"430\" srcset=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png 320w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png?w=150&amp;h=113 150w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/ultima7-bg-exult.png?w=300&amp;h=225 300w\" referrerpolicy=\"no-referrer\" \/><figcaption id=\"caption-attachment-13658\" class=\"wp-caption-text\"><em>Ultima VII<\/em>, with its keyword-based dialog system.<\/figcaption><\/figure><p>As Garriott\u2019s series mired itself in the morass of failure that was <em>Ultima IX<\/em>, the mantle of rich digital secondary worlds was taken up by two landmark games of the late 1990s <em>\u2013 Fallout<\/em> and <em>Baldur\u2019s Gate<\/em><sup id=\"fnref-13578-milestones\"><a href=\"#fn-13578-milestones\" class=\"jetpack-footnote\">8<\/a><\/sup>. These games attempted to have it all, and largely succeeded \u2013 detailed and satisfying tactical combat; a wide open world to explore and discover, with friends and enemies to be made depending on the player\u2019s choices; story beats seeded through the game to naturally lead the player toward the conclusion without the feeling of railroading; and, oh yeah, side quests. These are games in which one can immerse oneself like a warm bath, games in which one can wander for hours and hours and still find delightful new surprises: new nooks and crannies to explore, new choices to make, new people to meet.<\/p>\n<p>All of this delight, unfortunately, cost the creators of these games a great deal of time and money. In a D&amp;D campaign, all of the richness of the world and its denizens is conjured up <em>gratis<\/em> in the minds of the players by the spoken words of the DM. In <em>Baldur\u2019s Gate<\/em>, however, every choice, every possibility offered to the player in the name of openness had to be <em>put there<\/em> by someone. The game is, in effect, a very elaborate, lovingly illustrated and sound-tracked, flowchart. Every temple, every dungeon, every line of dialogue, every character animation, every side quest, came from the toil of artists, writers, and programmers.<\/p>\n<p>The cost to provide such things ballooned over the years as the expectations for the visual and auditory richness of games continued to ratchet up, from the simple tiles of <em>Ultima<\/em> to the hand-painted landscapes of <em>Baldur\u2019s Gate<\/em>, and beyond. As the lead designer for the latter game said in a recent interview, explaining the cost of a writer\u2019s simple flourish of imagination<sup id=\"fnref-13578-assets\"><a href=\"#fn-13578-assets\" class=\"jetpack-footnote\">9<\/a><\/sup>:<\/p>\n<blockquote>\n<p>What you write on a page takes ten seconds, but all the resources that have to go into that\u2014modeling, texturing, voiceover, music, all the rest\u2014suddenly that ten seconds of writing becomes tens of thousands of dollars of assets.<\/p>\n<\/blockquote>\n<p>It is for this reason, combined with the niche appeal of RPGs, that the \u201cUltima\u201d branch of computer RPGs was largely abandoned after the early 2000s in favor of more cost effective genres<sup id=\"fnref-13578-unsurpassed\"><a href=\"#fn-13578-unsurpassed\" class=\"jetpack-footnote\">10<\/a><\/sup>. There was, however, one more possible approach to building a secondary world available. After all, a computer need not merely ingest data that was already provided to it. What if instead of paying all those writers, modelers, and artists, you got the computer to build the world for itself?<\/p>\n<h2>Digital Synthesis<\/h2>\n<p>It is a fairly easy task to get a computer to generate a dungeon maze.<sup id=\"fnref-13578-maze\"><a href=\"#fn-13578-maze\" class=\"jetpack-footnote\">11<\/a><\/sup> It\u2019s not much harder to stock it with random monsters and loot, with escalating difficulty and value, respectively, as the player descends to deeper levels of the dungeon. A whole sub-genre was built on these facts, called \u201cRogue-likes,\u201d after the 1980 game <em>Rogue<\/em>, which was freely distributed on Unix systems throughout the following decade. Later variations on the theme such as <em>Hack, Moria,<\/em>\u00a0<em>NetHack<\/em>, and\u00a0<em>Angband<\/em> also spread across the fringes of nerdom throughout the 1980s. Though sometimes graphics packs were available, by default these games rendered the dungeon in ASCII characters, with letters for monsters, carats for stairs, and an @ for the player character.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_13660\" aria-describedby=\"caption-attachment-13660\" class=\"wp-caption aligncenter c5\"><img data-attachment-id=\"13660\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/nethack_3-4-3_ascii_capture_ecran\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png\" data-orig-size=\"563,258\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"nethack_3.4.3_ascii_capture_ecran\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png?w=300\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png?w=563\" class=\"size-full wp-image-13660 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png?w=739\" alt=\"nethack_3.4.3_ascii_capture_ecran\" srcset=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png 563w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png?w=150 150w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/nethack_3.4.3_ascii_capture_ecran.png?w=300 300w\" referrerpolicy=\"no-referrer\" \/><figcaption id=\"caption-attachment-13660\" class=\"wp-caption-text\">A\u00a0<em>NetHack\u00a0<\/em>ASCII dungeon<\/figcaption><\/figure><p>The genre exploded into the wider popular culture with the release of\u00a0<em>Diablo\u00a0<\/em>in 1996, which took the basic idea of <em>procedurally-generated<\/em> dungeon environments, and loot and added sophisticated visuals and sound and a graphical user interface. But the basic appeal of all these games was one-dimensional dungeon-delving. They offered no agency to the player beyond the cycle of kill and loot, and no wider setting to explore.<\/p>\n<p>What if the same basic concepts behind Rogue-like dungeon generation could be applied to an entire world? This was the the conceit of the <em>Elder<\/em> <em>Scrolls<\/em> series, which launched in 1994 with <em>Arena,<\/em> followed by the even-more-ambitious <em>Daggerfall<\/em> in 1996. Rather than generating new content on the fly as players explore, as most Rogue-likes do, the creators of <em>Daggerfall<\/em> pre-generated the major adventuring sites of the provinces of High Rock and Hammerfell on their own workstations, and then manually tweaked the results. In total, the game contains some four thousand dungeons and five thousands settlements (villages, towns, and cities) that the player can visit, across a total area of some sixty thousand square miles.<\/p>\n<p>Despite these astonishing figures, however, most of this vast area is utterly stale and lifeless, from a gameplay point of view. In (roughly contemporary) games like <em>Ultima VII<\/em>,\u00a0<em>Fallout<\/em>, or <em>Baldur\u2019s Gate<\/em>, exploring the world is a joy, because one never knows what characters, stories, adventures, or other surprises one will find around each corner. In\u00a0<em>Daggerfall<\/em>, other than a smattering of random monsters to fight, there is nothing to do or see in the wilderness between dungeons and towns. And there is nothing particularly exciting about discovering a new dungeon or entering a new town either, since one is much like another, only with a different assortment of random monsters or shops. The direction in the game comes from quests (randomly generated against one of several hundred templates) which direct players to a particular house in a particular town to fetch a dingus, or to a particular dungeon to kill five snarks. Unlike the very constricted scope of Rouge-likes, <em>Daggerfall\u00a0<\/em>gives the player a whole world in which to exercise their agency, but fails to provide many interesting things to do there.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_13662\" aria-describedby=\"caption-attachment-13662\" class=\"wp-caption aligncenter c6\"><img data-attachment-id=\"13662\" data-permalink=\"https:\/\/technicshistory.com\/2020\/01\/23\/tangent-the-automated-dungeon-master\/daggerfall1\/\" data-orig-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png\" data-orig-size=\"320,200\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"daggerfall1\" data-image-description=\"\" data-medium-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png?w=300\" data-large-file=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png?w=320\" class=\"wp-image-13662 aligncenter\" src=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png?w=655&amp;h=409\" alt=\"daggerfall1\" width=\"655\" height=\"409\" srcset=\"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png 320w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png?w=150&amp;h=94 150w, https:\/\/technicshistory.files.wordpress.com\/2020\/01\/daggerfall1.png?w=300&amp;h=188 300w\" referrerpolicy=\"no-referrer\" \/><figcaption id=\"caption-attachment-13662\" class=\"wp-caption-text\">The map of just one of forty-four regions in <em>Daggerfall<\/em>. Each dot is a dungeon, temple, home, or town.<\/figcaption><\/figure><p>Much more recently, the tremendous growth in the capabilities of machine learning has provided new hope for the dream of a procedurally-generated role-playing experience. A computer that can\u00a0 generate natural language responses to natural language inputs, a pipe dream just a decade ago, now seems near reality. This past year, Nick Walton, a student at Brigham Young University and a D&amp;D fan, saw that these new tools could be used to simulate a dungeon master. The result was <em>AI Dungeon<\/em>, based on Open AI\u2019s GPT-2 machine learning model.<\/p>\n<p>Trained on gigabytes of input, GPT-2 is uncannily good at producing sensible text by simply predicting the next word that should follow a given list of input words. Walton fine-tuned the model by feeding it transcripts from a variety of games from the <em>Adventure<\/em> family tree. By doing so he created a kind of meta-<em>Adventure<\/em>, that could branch off in (for all practical purposes) an infinite number of directions. \u00a0This is a parser that is not mere window dressing for a flow chart. Like a human dungeon master, it generates a novel response to each input from the player.<\/p>\n<p>However, there is a painful sacrifice to be made for this gain in openness, for <em>AI Dungeon<\/em> has only the feeble self-consistency of a dream or hallucination. The game has no model of the world to test its responses against. It is a kind of prism, refracting the player\u2019s words through a textual lens, and then bouncing the result back to the screen. The game will respond without blinking to absurd player actions that make no sense in the context of an adventure a fantasy realm, e.g. \u201cfly to Tahiti\u201d or \u201cwrite an essay about Dungeons and Dragons.\u201d And, like a river, it will not allow you to step into the same place twice. Here, for example, is my own attempt at an adventure as the wizard Iocounu. Notice how impressive it is until I try to retrace my steps<sup id=\"fnref-13578-steps\"><a href=\"#fn-13578-steps\" class=\"jetpack-footnote\">12<\/a><\/sup>:<\/p>\n<blockquote>\n<p>You are Iucounu, a wizard living in the kingdom of Larion. You have a staff and a spellbook. You finish your long journey and finally arrive at the ruin you\u2019ve been looking for. You look around and see that it\u2019s indeed a ruined city. The walls are gone, but the buildings still stand tall as if they were built to last forever.<\/p>\n<p>The ruins themselves aren\u2019t anything special; there is nothing here except rubble and dust.<\/p>\n<p>&gt; enter nearest building<\/p>\n<p>You walk towards one of the buildings, which looks like it might be a guild hall or something similar. You open up the door and find yourself inside a large room with several tables filled with people sitting at them. Most of them seem to be mages studying some sort of book.<\/p>\n<p>&gt; try a different building<\/p>\n<p>You head into another building which doesn\u2019t look too interesting. You go through the front doors and then down a set of stairs.<\/p>\n<p>&gt; go back to the guild hall<\/p>\n<p>You decide to check out the guild hall since you saw it from afar earlier. It seems like it would be a good place to get information on what\u2019s going on in the rest of the city.<\/p>\n<p>&gt; look around<\/p>\n<p>As soon as you step inside, you immediately feel very uncomfortable. There are so many people here that it feels like an invasion. They all stare at you and make faces at you. Some even begin to speak loudly about how \u201cwe\u201d<\/p>\n<\/blockquote>\n<h2>Hail to the Dungeon Master<\/h2>\n<p>Because of scaling limitations, models like GPT-2 can consider no more than a few hundred words of previous text in constructing its next output. Google, however, just announced a new type of machine learning mechanism they call \u201creformer\u201d that can open a much larger window on the past<sup id=\"fnref-13578-reformer\"><a href=\"#fn-13578-reformer\" class=\"jetpack-footnote\">13<\/a><\/sup>. Could this, or some other breakthrough, breathe consistency into the phantasmagoria of <em>AI Dungeon<\/em>? Time will tell.<\/p>\n<p>Thus far, however, the effort to enfold the magic of D&amp;D between the covers of a book, or to inscribe it into the electronic memory of a computer, has been a noble failure. Gamebooks and computer games are available at any time of day or night, ready to play. They will always give a consistent experience, and the better ones can provide hours and hours of enjoyment. They are never burned out, tired, or lazy. Their memory or imagination never falters. But without a human mind behind them, they cannot offer truly open-ended agency, at least without sacrificing all self-consistency. Economics plays a role, of course. There is no strict bound to the number of options that could be provided to the player of a computer RPG, given enough time and money \u2013 but in practical terms, the limit is quite sharp. Either your actions in the secondary world that the game invokes are strictly limited to the paths the creator has set out for you, or that world is nothing but a fever dream, a never-ending present without past or future.<\/p>\n<p>So, for now, we are stuck with the dungeon masters, with all their human foibles. Long may they live.<\/p>\n<h2>Further Reading<\/h2>\n<p>Shannon Appelcline,\u00a0<em>Designers &amp; Dragons: The \u201970s<\/em> (2014)<\/p>\n<p>Jimmy Maher, <em><a href=\"http:\/\/filfre.net\">The Digital Antiquarian<\/a><\/em>\u00a0(2011-present)<\/p>\n<p>Jon Peterson,\u00a0<em>Playing at the World\u00a0<\/em>(2012)<\/p>\n<div class=\"footnotes\">\n<hr \/><ol><li id=\"fn-13578-otherrpgs\">D&amp;D is only one in an entire genre of similar games, of course (Now sometimes called table-top RPGs to distinguish them from their digital cousins.) However, since it is by far the most popular, and the one I am most familiar with, I will, in the tradition of synedoche, often allow it to stand in for the whole class in this essay.\u00a0<a href=\"#fnref-13578-otherrpgs\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-danddtolkien\">Though Gygax later downplayed the importance of Tolkien to the creation of D&amp;D, his influence is clearly evident. All of the game\u2019s non-human racial options \u2013 dwarf, elf, half-orc, and hobbit (later renamed halflings to avoid a lawsuit from Tolkien\u2019s estate) \u2013 come straight from Middle-Earth, as does the ranger class (clearly modeled on Aragorn) and many of the monsters described in the game.\u00a0<a href=\"#fnref-13578-danddtolkien\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-hexcrawl\">This is true even of low-preparation methods of world creation, such as the \u201chex crawl\u201d, where DMs roll randomly to determine what is in each area as the players navigate around a hexagonal world map. This still requires great creativity and improvisational skill, in order to fit the castle that appeared over there together with the cult site over there and the town over there, and then hold them all together in a self-consistent picture of a world over the course of the campaign.\u00a0<a href=\"#fnref-13578-hexcrawl\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-megadungeon\">Another form of railroading, though not usually considered as such, is the mega-dungeon. By giving players a vast, multi-level dungeon to explore and play around in, complete with monsters, treasure, tricks, traps, and political factions among its denizens, the DM can create a sense of agency for the players within a tightly bounded, well-mapped space. This was the approach taken by Gary Gygax himself, who focused his early campaigns around the dungeons under Castle Greyhawk.\u00a0<a href=\"#fnref-13578-megadungeon\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-htmlversion\">You can find an implementation of the adventure in HTML here: <a href=\"http:\/\/www.flyingbuffalo.com\/bcintro.htm.&amp;#160\" rel=\"nofollow\">http:\/\/www.flyingbuffalo.com\/bcintro.htm.&amp;#160<\/a>;<a href=\"#fnref-13578-htmlversion\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-million\">The figure comes from Applecline,\u00a0<em>Designers &amp; Dragons: The \u201970s<\/em>.\u00a0<a href=\"#fnref-13578-million\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-gamebooks\">One can find a complete guide to the genre at gamebooks.org.\u00a0<a href=\"#fnref-13578-gamebooks\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-milestones\">I\u2019ve of course passed over many other games, in an attempt to maintain some semblance of brevity. The most regretful omission is <em>Wasteland<\/em> (1988). It offered a truly self-consistent world that the player characters\u2019 actions could permanently change, and an open-ended approach to problem-solving: many conflicts in the game could be solved by stealth or negotiation instead of violence. It was also, incidentally, co-designed by Ken St. Andre, and based on a derivative of the <em>Tunnels &amp; Trolls<\/em> rules system.\u00a0<a href=\"#fnref-13578-milestones\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-assets\">\u201cPause Screen: Green Team \u2013 How a Ragtag Band of Upstarts Created Baldur\u2019s Gate,\u201d<em>\u00a0<\/em><em>Shack News<\/em> (2018). According to the same source, it took ninety person-years of work to produce\u00a0<em>Baldur\u2019s Gate<\/em>, which takes a\u00a0player perhaps sixty hours to fully explore.\u00a0<a href=\"#fnref-13578-assets\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-unsurpassed\">It is for this same reason that <em>Fallout, Baldur\u2019s Gate<\/em>, and their immediate sequels have not been truly surpassed, except in aesthetic terms, in the two decades since. The genre lay almost entirely moribund from 2001 until the highly successful Kickstarter for <em>Wasteland 2<\/em> in 2012. Kickstarter has made it far easier for motivated creators to find people willing to support a years-long effort in a niche genre.\u00a0<a href=\"#fnref-13578-unsurpassed\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-maze\">In fact there\u2019s an entire book about a BASIC program that generates a random maze with two lines of code, <em>10 PRINT CHR$(205.5+RND(1)); : GOTO 10<\/em> (2012).\u00a0<a href=\"#fnref-13578-maze\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-steps\">The strange trailing-off at the end is not a copy-paste error, but some kind of glitch in the text-generation engine.\u00a0<a href=\"#fnref-13578-steps\">\u21a9<\/a><\/li>\n<li id=\"fn-13578-reformer\"><a href=\"https:\/\/ai.googleblog.com\/2020\/01\/reformer-efficient-transformer.html&amp;#160\" rel=\"nofollow\">https:\/\/ai.googleblog.com\/2020\/01\/reformer-efficient-transformer.html&amp;#160<\/a>;<a href=\"#fnref-13578-reformer\">\u21a9<\/a><\/li>\n<\/ol><\/div>","created_at":"2020-01-24T00:57:49+0000","updated_at":"2020-01-24T00:57:49+0000","published_at":"2020-01-23T04:40:18+0000","published_by":["View all posts by technicshistory"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":36,"domain_name":"technicshistory.com","preview_picture":"https:\/\/technicshistory.files.wordpress.com\/2020\/01\/1634e73f6ac2b9ea05e74df94e981e04.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1118,"title":"us5.campaign-archive.com","url":"https:\/\/us5.campaign-archive.com\/?e=d4d5dd5efe&u=e036010acb60c2f90e4c9bbb7&id=61e59b6f63","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-01-20T20:43:57+0000","updated_at":"2020-01-20T20:43:57+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"us5.campaign-archive.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1117,"title":"jschaf\/esup","url":"https:\/\/github.com\/jschaf\/esup","content":"<p><a href=\"https:\/\/travis-ci.org\/jschaf\/esup\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/58b588322cb0d677b6479d1ce4a1468d22c4930b\/68747470733a2f2f7472617669732d63692e6f72672f6a73636861662f657375702e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/jschaf\/esup.svg?branch=master\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/melpa.org\/#\/esup\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d2842e4f5e58af03c8405292348a418f58549753\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f657375702d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"http:\/\/melpa.org\/packages\/esup-badge.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>Benchmark Emacs Startup time without ever leaving your Emacs.<\/p><p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/jschaf\/esup\/blob\/master\/esup-screenshot.png\"><img src=\"https:\/\/github.com\/jschaf\/esup\/raw\/master\/esup-screenshot.png\" width=\"512\" height=\"504\" class=\"c2\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/p><p><strong>With MELPA<\/strong><\/p><p>First, add the package repository:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(add-to-list 'package-archives\n             '(\"melpa\" . \"https:\/\/melpa.org\/packages\/\") t)<\/pre><\/div><p>Then install esup:<\/p><p><kbd>M-x package-install RET esup RET<\/kbd><\/p><p><strong>El-Get<\/strong><\/p><p><code>esup<\/code> is included in the El-Get repository<\/p><p>Install esup:<\/p><p><kbd>M-x el-get-install RET esup RET<\/kbd><\/p><p><strong>Manually<\/strong><\/p><p>Download <code>esup.el<\/code> and place the download directory on your <code>load-path<\/code> like so:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(add-to-list 'load-path \"~\/.emacs.d\/path\/to\/esup\")\n(autoload 'esup \"esup\" \"Emacs Start Up Profiler.\" nil)<\/pre><\/div><p>To start Esup, run <kbd>M-x esup<\/kbd>, and watch the magic happen.<\/p><p>By default, esup will profile <code>user-init-file<\/code>. To profile a custom file, call <code>esup<\/code> with a prefix argument. That is, <kbd>C-u M-x esup<\/kbd>.<\/p><p>Patches are always welcome. To submit a patch, use something like the following workflow.<\/p><div class=\"highlight highlight-source-shell\"><pre>git clone https:\/\/github.com\/jschaf\/esup.git\ncd esup\ngit checkout -b MY-NEW-FIX<\/pre><\/div><p>Implement your fix. Please use <a href=\"https:\/\/github.com\/angular\/angular.js\/blob\/master\/CONTRIBUTING.md#-git-commit-guidelines\">angular-style commit messages<\/a>.<\/p><p>Ensure that all elisp code is lint-clean with flycheck.<\/p><p>Test your fix with <a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a>. You can install cask with:<\/p><div class=\"highlight highlight-source-shell\"><pre>curl -fsSL https:\/\/raw.githubusercontent.com\/cask\/cask\/master\/go | python\nexport PATH=\"\/home\/joe\/.cask\/bin:$PATH\"\n<\/pre><\/div><p>Test your fixes with the Emacs Regression Test runner.<\/p><pre>make test\n<\/pre><p>Send your commits for review.<\/p><pre>git push origin MY-NEW-FIX\n<\/pre><p>Create a pull request with the normal Github user interface.<\/p>","created_at":"2020-01-20T20:42:39+0000","updated_at":"2020-01-20T20:42:39+0000","published_by":["jschaf"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/22385?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1116,"title":"How to find the Dropbox folder on your computer","url":"https:\/\/help.dropbox.com\/installs-integrations\/desktop\/locate-dropbox-folder","content":"<p>The JSON file is named info.json and can be found at one of these two paths:<\/p><p><code>\u00a0 \u00a0 %APPDATA%\\Dropbox\\info.json<\/code><\/p><p><code>\u00a0 \u00a0 %LOCALAPPDATA%\\Dropbox\\info.json<\/code><\/p><p>In this file, you'll find a JSON object that looks like this:<\/p><p><code>{<\/code><\/p><p><code>\u00a0 \u00a0 {\"personal\":<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 {\"path\": \"\/Users\/&lt;username&gt;\/Dropbox (Personal)\",<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \"host\": 123456789,\u00a0<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \"is_team\": false,\u00a0<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \"subscription_type\": \"Basic\"},<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0{\"business\":\u00a0<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 {\"path\": \"\/Users\/&lt;username&gt;\/Dropbox (&lt;business name&gt;)\",\u00a0<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \"host\": 123456789,\u00a0<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \"is_team\": true,\u00a0<\/code><\/p><p><code>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \"subscription_type\": \"Business\"}<\/code><\/p><p><code>}<\/code><\/p>","created_at":"2020-01-16T03:44:51+0000","updated_at":"2020-01-16T03:44:51+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"help.dropbox.com","preview_picture":"https:\/\/aem.dropbox.com\/cms\/content\/dam\/dropbox\/www\/en-us\/help\/DropboxGlyph_Blue_CMYK_200px_width.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1115,"title":"Home","url":"https:\/\/airflow.apache.org\/","content":"<main role=\"main\" class=\"td-main container\"><div><h4 class=\"page-header principles-header\">Principles<\/h4><div class=\"text-with-icon-list\"><div class=\"text-with-icon-item\"><p class=\"text-with-icon-item--text\">Airflow has a modular architecture and uses a message queue to orchestrate an arbitrary number of workers. Airflow is ready to scale to infinity.<\/p><\/div><div class=\"text-with-icon-item\"><p class=\"text-with-icon-item--text\">Airflow pipelines are configuration as code (Python), allowing for dynamic pipeline generation. This allows for writing code that instantiates pipelines dynamically.<\/p><\/div><div class=\"text-with-icon-item\"><p class=\"text-with-icon-item--text\">Easily define your own operators, executors and extend the library so that it fits the level of abstraction that suits your environment.<\/p><\/div><div class=\"text-with-icon-item\"><p class=\"text-with-icon-item--text\">Airflow pipelines are lean and explicit. Parametrizing your scripts is built into its core using the powerful Jinja templating engine.<\/p><\/div><\/div><\/div>\n<div><h4 class=\"page-header\">Features<\/h4><div class=\"features-list\"><div class=\"feature-item\"><div class=\"feature-item--text-box\"><p class=\"feature-item--text\">No more command-line or XML black-magic! Use all Python features to create your workflows including date time formats for scheduling tasks and loops to dynamically generate tasks. This allows you to build your workflows as complicated as you wish.<\/p><\/div><\/div><div class=\"feature-item\"><div class=\"feature-item--text-box\"><p class=\"feature-item--text\">Monitor, schedule and manage your workflows using web app. No need to learn old, cron-like interfaces. You always have an insight into the status of completed and ongoing tasks along with insight into the logs.<\/p><\/div><\/div><div class=\"feature-item\"><div class=\"feature-item--text-box\"><p class=\"feature-item--text\">Airflow provides many plug-and-play operators that are ready to handle your task on Google Cloud Platform, Amazon Web Services, Microsoft Azure and many other services. This makes Airflow easy to use with your current infrastructure.<\/p><\/div><\/div><div class=\"feature-item\"><div class=\"feature-item--text-box\"><p class=\"feature-item--text\">Anyone with Python knowledge can deploy a workflow. Apache Airflow does not limit scopes of your pipelines. You can use it for building ML models, transferring data or managing your infrastructure.<\/p><\/div><\/div><div class=\"feature-item\"><div class=\"feature-item--text-box\"><p class=\"feature-item--text\">Wherever you want to share your improvement you can do this by opening a PR. It\u2019s simple as that, no barriers, no prolonged procedures. Airflow has many active users who willingly share their experiences. Have any questions? Check our buzzing slack.<\/p><\/div><\/div><\/div><\/div>\n<div><h4 class=\"page-header integrations-header\">Integrations<\/h4><div id=\"integrations\"><div id=\"integrations-container\" class=\"list-items\"><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#aws-amazon-web-services\">\n<p>Amazon CloudWatch Logs<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#service-integrations\">\n<p>Google Display&amp;Video 360<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#software-integrations\">\n<p>Oracle<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#service-integrations\">\n<p>Salesforce<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#gcp-google-cloud-platform\">\n<p>Cloud Memorystore<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#service-integrations\">\n<p>Jenkins<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#aws-amazon-web-services\">\n<p>Amazon SageMaker<\/p>\n<\/a><a class=\"list-item\" href=\"https:\/\/airflow.apache.org\/docs\/stable\/integration.html#gcp-google-cloud-platform\">\n<p>Cloud Pub\/Sub<\/p>\n<\/a><\/div><p>Loading..<\/p><\/div><\/div>\n<\/main>","created_at":"2020-01-15T22:46:57+0000","updated_at":"2020-01-15T22:46:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"airflow.apache.org","preview_picture":"https:\/\/airflow.apache.org\/images\/feature-image.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1114,"title":"Technical Info for Dominions 5","url":"http:\/\/www.illwinter.com\/dom5\/techmanual.html","content":"Technical Info for Dominions 5<p>This document contains information on the command line options available for Dominions 5 as well as commands to control a network server without using the GUI. If you just want to play Dominions 5, you do not need to read this document.<\/p><h2 id=\"command-line-options\">Command Line Options<\/h2>\n<p>The command line options can be entered in Steam\u2019s \u201cSet Launch Options\u201d or provided on the command line if you start Dominions that way.<\/p>\n<p>The most common use for the command line options is to automate some task for a script, or if you have technical problems with the game and need to disable or change some settings before the game starts.<\/p>\n<h3 id=\"general-options\">General Options<\/h3>\n<pre>-v  --version       Print version number and exit\n-d                  Increase debug level\n-g  --host          Generate new turn and exit\n    --verify        Verify all 2h-files and exit (creates .chk files)\n    --statfile      Create a player info file after each turn (stats.txt)\n    --scoredump     Create a score file after each turn (scores.html)\n    --finalhost     Generate new turn, send out final score msg and exit\n-c  --nocredits     Disables the end credits\n    --noedgescroll  Disables edge scrolling\n    --edgescroll    Enables edge scrolling in fullscreen mode\n    --winedgescroll Enables edge scrolling always\n    --askifplayed   Will ask if you want to redo turn if possible\n    --noshowscouts  Do not show allied scouts on map\n    --autoorder     Put new mages on research by default\n    --multimove     Enable multiple turn movement\n    --maxunits X    Sets max nbr of units in a game (default and max 250000)\n    --listnations   Prints a list of all nation numbers\n    --multiai X     Create up to X processes for AI computations (0=off) (Linux\/OSX only)\n    --nothread      Don't use multithreading\n    --nomappopups   Don't show popups on the main map\n    --nosteam       Do not connect to steam (workshop will be unavailable)\n    --nocrashbox    Don't show a dialogue box with crash message on (Windows only)<\/pre>\n<h3 id=\"network-options\">Network Options<\/h3>\n<pre>-C  --tcpclient     Connect to a Dominions multiplayer server\n-S  --tcpserver     Start a Dominions multiplayer server\n    --tcpquery      Query Dominions server about game status and exit\n    --ipadr XXX     Use this IP-adr when connecting to server\n    --port X        Use this port nbr\n    --preexec CMD   Execute this command before each new turn\n    --postexec CMD  Execute this command after each new turn\n-t  --hosttime X Y  Host on day X (0=sunday) hour Y (0-23)\n    --minutes X     Set host interval in minutes\n    --hours X       Set host interval in hours\n    --pauseday X    Stop timer on this day (0=sunday)\n    --timeleft X    Hours until first host (default = full host interval)\n    --noquickhost   Don't host just because all turns are done\n    --maxholdups X  Quickhost disregards players that have stalled last X turns\n-n  --nonationsel   No nation selection when resuming a network game\n-o  --onserver      I'm playing on the server, don't ask\n    --noclientstart Clients cannot start the game during Choose Participants\n    --uploadtime X  Game is created after this many minutes.\n    --uploadmaxp X  Game is created if this many players join.\n    --closed X      Nation closed X=nation number (5-249)\n    --easyai X      Nation ai controlled X=nation number (5-249)\n    --normai X      Nation ai controlled X=nation number (5-249)\n    --diffai X      Nation ai controlled X=nation number (5-249)\n    --mightyai X    Nation ai controlled X=nation number (5-249)\n    --masterai X    Nation ai controlled X=nation number (5-249)\n    --impai X       Nation ai controlled X=nation number (5-249)\n    --team X Y Z    X=nation, Y=team, Z=type (type: 1=pretender, 2=disciple)\n    --statuspage XX Create html page that shows who needs to play their turn\n    --statusdump    Continuously create info on players in a parsable format\n    --nodownlmods   Don't download mods from game server automatically\n    --nomaster      Disallow download\/upload turns using the master password<\/pre>\n<h3 id=\"new-game-options\">New Game Options<\/h3>\n<pre>    --mapfile XXX   Filename of map. E.g. eye.map\n    --randmap X     Make and use a random map with X prov per player (10,15,20)\n    --research X    Research difficulty 0 to 4 (default 2)\n    --norandres     No random start research\n    --hofsize X     Size of Hall of Fame 5-15 (default 10)\n    --globals X     Global Enchantment slots 3-9 (default 5)\n    --indepstr X    Strength of Independents 0-9 (default 5)\n    --magicsites X  Magic site frequency 0-75 (default 40)\n    --eventrarity X Random event rarity 1-2, 1=common 2=rare\n    --richness X    Money multiple 50-300 (default 100)\n    --resources X   Resource multiple 50-300 (default 100)\n    --recruitment X Unit recruitment point multiple 50-300 (default 100)\n    --supplies X    Supply multiple 50-300 (default 100)\n    --masterpass XX Master password. E.g. masterblaster\n    --startprov X   Number of starting provinces (1-9)\n    --renaming      Enable commander renaming\n    --scoregraphs   Enable score graphs during play\n    --nonationinfo  No info at all on other nations\n    --nocheatdet    Turns off cheat detection\n    --era X         New game created in this era (1-3)\n    --nomods        Disable all mods\n-M  --enablemod XXX Enable the mod with filename XXX\n    --noartrest     Players can create more than one artifact per turn\n    --teamgame      Disciple game, multiple players on same team\n    --clustered     Clustered start positions for team game\n    --nostoryevents Disable all story events\n    --storyevents   Enable some story events\n    --allstoryevents Enable all story events\n    --newgame       Create a new game and exit (for scripted game creation)\n    --newailvl X    AI level for human players who quit (1-6, def 2)\n    --nonewai       Disable become AI controlled<\/pre>\n<h3 id=\"new-game-victory-condition\">New Game Victory Condition<\/h3>\n<pre>    --conqall        Win by eliminating all opponents only\n    --thrones X Y Z  Number of thrones of level 1, 2 and 3\n    --requiredap X   Ascension points required for victory (def total-1)\n    --cataclysm X    Cataclysm will occur on turn X (def off)<\/pre>\n<h3 id=\"random-map-options\">Random Map Options<\/h3>\n<pre>    --makemap XXX    Generate a random map with filename XXX and exit\n    --riverpart X    0-1000, 0=no rivers (default 50)\n    --bridges X      Bridge chance 0-100, 0=no bridges (default 60)\n    --extraislands X Chance of extra islands 0-100, (default 0)\n    --seapart X      Percent of map that is below water level (default 30)\n    --mountpart X    Percent of map that is mountains (default 20)\n    --forestpart X   Percent of lands that are forests (default 20)\n    --farmpart X     Percent of lands that are farm lands (default 15)\n    --wastepart X    Percent of lands that are wastes (default 10)\n    --swamppart X    Percent of lands that are swamps (default 10)\n    --cavepart X     Percent of lands that are caves (default 3)\n    --cavecluster X  Percent of caves that are clusters (default 20)\n    --kelppart X     Percent of seas that are kelp forests (default 25)\n    --gorgepart X    Percent of deeps seas that are gorges (default 25)\n    --mapsize W H    Set width and height of random map (default 3000 2000)\n    --mapprov X      Set number of provinces (default 150)\n    --mapscol R G B A   Sea color 0-255 (default 54 54 130 255)\n    --mapdscol R G B A  Deep Sea color 0-255 (default depending on mapscol)\n    --mapccol R G B A   Coast color 0-255 (default depending on mapscol)\n    --mapbcol R G B A   Ground border color 0-255\n    --mapsbcol R G B A  Sea border color 0-255\n    --mapbtopcol R G B A   Top border color 0-255\n    --mapnoise X     Ground color noise 0-255 (default 15)\n    --mapdirt X      Amount of dirt blobs (default 100)\n    --mapdirtcol X   Color variance of dirt blobs (default 5)\n    --mapdirtsize X  Size of dirt blobs (default 100)\n    --borderwidth X  Border width 0-500 (default 100)\n    --hills X        Number of Hills\/Craters (default 150)\n    --rugedness X    Rugedness 0-100 (default 30)\n    --seasize X      Sea size, 100=normal land size (default 350)\n    --mapnospr       Don't draw any sprites on the map\n    --vwrap          Make map wrap north\/south\n    --nohwrap        Make map not wrap east\/west\n    --mapbunch X     When making a bunch, make this many maps (default 12)\n    --quiet          Don't print stuff<\/pre>\n<h3 id=\"graphics-options\">Graphics Options<\/h3>\n<pre>-w  --window        Run Dominions 5 in a window\n-u  --fullscreen    Use the entire screen\n    --borderless    Use a borderless fullscreen window\n    --bitplanes X   Try to use a color depth of X bits per pixel\n    --zbuffer X     Try to use a depth buffer of X bits per pixel (default 24)\n-T  --textonly      Use this with --tcpserver to get graphicless server\n    --gamma X       Set gamma function (brightness) 0.1 - 5.0 (default 1.0)\n    --opacity X     Set gui opacity 0 - 100\n-r  --res X Y       Set screen resolution \/ window size\n    --animback      Use animated backgrounds\n-a  --noanimback    Don't use animated backgrounds\n    --fade          Use fade effects\n-f  --nofade        No fade effects\n    --nopopups      No helpful popups\n    --maxfps X      Maximum nbr of frames per second (default 60)\n    --filtering X   Quality of OpenGL filtering 0-3 (default 2)\n    --maxtexsize X  Max texture size in pixels 8-4096 (default unlimited)\n    --texqual X     Texture quality 1-5 (default 3)\n    --notracers     Visual aid, puts a particle trace after arrows\n    --nolightfx     No light effects in battles\n    --partamount X  Max nbr of particles 0-9 (0=none, 4=default, 9=max)\n    --noarcade      Don't draw floating damage numbers\n    --nonetinfo     Don't draw timer and flags during network play\n    --noglfinish    Don't flush the graphics pipeline each frame\n    --noglext       Don't use any OpenGL extensions\n-x  --fastgrx       Faster and simpler graphics\n    --simpgui       Use a simple GUI without any background textures\n    --gfxlevel X    Graphics level 1-20 (default 10, 14 = very high)\n    --multisample X Use multisampling, X=number of samples (0-16)\n    --benchmark     Run a graphics benchmark and exit<\/pre>\n<h3 id=\"audio-options\">Audio Options<\/h3>\n<pre>-s  --nosound       No sound effects or music\n-m  --nomusic       No music\n    --musicvol X    Set music volume, 0-100 (default 100)\n    --fxvol X       Set sound effect volume, 0-100 (default 100)\n    --clickvol X    Set mouse click volume, 0-100 (default 25)\n    --noturnsound   Don't play a sound when a new turn arrives (network only)\n    --defsound      Use default sound device\n    --jack          Route sound through JACK sound server (Linux only)\n    --pulseaudio    Route sound through pulse audio (Linux only)\n    --arts          Route sound through aRts (Linux only)\n    --alsa          Use direct alsa sound output (Linux only)\n    --oss           Use direct oss sound output (Linux only)\n    --portaudio     Use portaudio for sound (Linux \/ Mac OS X only)\n    --directsound   Use direct sound (Windows only)\n    --waveout       Use waveout for sound (Windows only)\n    --sdlsound      Force usage of default sdl sound device (Windows \/ Mac OS X only)<\/pre>\n<h2 id=\"environment-variables\">Environment Variables<\/h2>\n<p>There are a few environment variable that can be set to alter where Dominions finds its files. Default paths are shown for Linux. The linux local config dir (~\/.dominions5) can be renamed to ~\/dominions5 if you want to make it visible. Dominions 5 uses ~\/dominions5 if ~\/.dominions5 cannot be found.<\/p>\n<pre>  DOM5_CONF  defaults to ~\/.dominions5\n  DOM5_SAVE  default to DOM5_CONF\/savedgames\n  DOM5_LOCALMAPS  default to DOM5_CONF\/maps\n  DOM5_MODS  default to DOM5_CONF\/mods\n  DOM5_DATA  default to EXEDIR\/data\n  DOM5_MAPS  default to DOM5_DATA\/..\/maps<\/pre>\n<h2 id=\"server-control\">Server Control<\/h2>\n<p>A Dominions server can be controlled via text files, providing a way for automated tools to interact with the game. The text file must be named \u201cdomcmd\u201d and be placed in the save game folder for game that you want to control. The file can contain one or more commands (one per line). The file is deleted as soon as it is read and the entire file must be created all at once.<\/p>\n<pre>  setquickhost &lt;on 0-1&gt;\n  setpause &lt;day 0-6&gt; &lt;on 0-1&gt;\n  setinterval &lt;minutes&gt;\n  settimeleft &lt;seconds&gt;<\/pre>\n<p>Usually these commands are only meaningful once the game is running, but settimeleft can also be used during the upload pretender stage to start the game creation.<\/p>\n<p>Illwinter Game Design: <a href=\"http:\/\/www.illwinter.com\" class=\"uri\">http:\/\/www.illwinter.com<\/a><\/p>","created_at":"2020-01-15T03:35:44+0000","updated_at":"2020-01-15T03:35:44+0000","published_by":["Illwinter Game Design"],"annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"www.illwinter.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["lisp","travel","ita","lola","common-lisp"],"is_public":false,"id":1113,"title":"Google agrees to buy ITA Software for $700 million","url":"https:\/\/news.ycombinator.com\/item?id=1479107","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D1479107\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"1479107\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"http:\/\/googleblog.blogspot.com\/2010\/07\/taking-off-with-ita.html\" class=\"storylink\">Google agrees to buy ITA Software for $700 million<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=googleblog.blogspot.com\">googleblog.blogspot.com<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">316 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=alec\" class=\"hnuser\">alec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=1479107\">on July 1, 2010<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=1479107&amp;goto=item%3Fid%3D1479107\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Google%20agrees%20to%20buy%20ITA%20Software%20for%20%24700%20million&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Google%20agrees%20to%20buy%20ITA%20Software%20for%20%24700%20million\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=1479107&amp;auth=c8db23deeef5470d440384931d9998d926977f12\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=1479107\">121\u00a0comments<\/a><\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"mailto:hn@ycombinator.com\">Support<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2020-01-14T21:13:55+0000","updated_at":"2020-01-14T21:15:29+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["design","illustrations"],"is_public":false,"id":1112,"title":"Craftwork - User Interface Resouces for Designers","url":"https:\/\/craftwork.design\/","content":"<div class=\"container home-top row\">\n<div class=\"col-lg-7 col-md-12\">\n<h1 class=\"page-title text-black wow fadeInUp\" data-wow-delay=\"0.4s\">UI assets for startup owners &amp; busy designers<\/h1>\n<p class=\"page-subtitle wow fadeInUp\" data-wow-delay=\"0.5s\">High-quality digital resources which make your<br \/>\ndesign workflow faster and easier<\/p>\n<ul class=\"nav page-section-links wow fadeInUp\" data-wow-delay=\"0.7s\"><li class=\"nav-item\"><a class=\"nav-link\" href=\"#ux-ui-systems\">UX\/UI Systems<\/a><\/li>\n<li class=\"nav-item\"><a class=\"nav-link\" href=\"#illustrations\">Illustrations<\/a><\/li>\n<li class=\"nav-item\"><a class=\"nav-link\" href=\"#templates\">Templates<\/a><\/li>\n<li class=\"nav-item\"><a class=\"nav-link\" href=\"#mockups\">Mockups<\/a><\/li>\n<li class=\"nav-item\"><a class=\"nav-link\" href=\"#freebies\">Freebies<\/a><\/li>\n<\/ul><\/div>\n<div class=\"col-lg-4 offset-lg-1 col-md-12 banner text-center position-relative wow fadeInUp banner-inner position-relative\">\n<p class=\"banner-suptitle text-uppercase\">Become a Member<\/p>\n<h2 class=\"banner-title text-white\">Get access to all assets we made and will make for one year<\/h2>\n<a href=\"https:\/\/craftwork.design\/unlimited-access\/\" class=\"banner-btn position-relative\">Go Unlimited For $99<\/a><\/div>\n<\/div>\n<div class=\"container home-news news-slider wow fadeInUp\" data-wow-delay=\"0.7s\">\n<div class=\"row col-md-12\">\n<h2 class=\"home-news-title text-black text-uppercase\">Feed<\/h2>\n<\/div>\n<div class=\"wpnawfree-plugin\">\n<div id=\"post-46096\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/3ddd.casa\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">3ddd Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">An incredible set of colorful &amp; playful 3D illustrations<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">January 13, 2020<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-43697\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/bubble-illustrations\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Bubble Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Introducing Bubble pack with 25 bright and modern illustrations from IT world<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">December 17, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-43095\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/struct.rocks\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-update\">Update<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Updated Struct 2 Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">We have created 20 more incredible illustrations for your landings and applications<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">December 10, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-42013\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/singleton.digital\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"479\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-768x575.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-1024x767.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570.png 1140w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"479\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-768x575.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-1024x767.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570.png 1140w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Brand New Singleton<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Meet our brand new UI kit with stylish and playful layouts for beautiful landings. Create websites fast with 100+ ready-to-use blocks in 11 categories.<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">December 3, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-41116\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/thursday-illustrations\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Thursday Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Stylish and old school illustrations for landings and applications<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">November 26, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-41148\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/method-wireframe-kit-3\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-update\">Update<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Method Mobile Blocks<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Introducing a new update for Method Wireframe Kit with 120+ mobile layouts.<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">November 19, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-39274\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/blink-illustrations\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Blink Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Stylish set of illustrations with modern and colorful characters<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">November 12, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-38340\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/no-gravity-illustrations\/\" class=\"news-inner-wrap-view row wpnaw-news-no-image\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">No Gravity Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Carefully crafted header illustrations for landings and mobile screens<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">November 6, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-36766\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-article\"><a target=\"_blank\" href=\"https:\/\/medium.com\/@Shepovalovdenis\/why-do-you-need-to-give-out-freebies-a7e5b5f6923f\" class=\"news-inner-wrap-view row wpnaw-news-no-image c2\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-article\">Article<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Why do you need to give out freebies?<\/h3>\n<div class=\"news-content-excerpt\">\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 31, 2019<\/div>\n<\/div>\n<\/div><\/a><\/div>\n<div id=\"post-35998\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/usesmash.com\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4-300x225.png 300w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Meet free illustrations generator<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Create a unique story mixing and matching characters, objects, and styles<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 29, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-35330\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/spooky-illustrations\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Spooky Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Introducing a set of 12 creepy and reckless illustrations for Halloween<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 22, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-35193\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-article\"><a target=\"_blank\" href=\"https:\/\/medium.com\/@Shepovalovdenis\/top-7-flowchart-kits-9dace069e43a\" class=\"news-inner-wrap-view row wpnaw-news-no-image c3\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-article\">Article<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Top 7 Flowchart Kits<\/h3>\n<div class=\"news-content-excerpt\">\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 21, 2019<\/div>\n<\/div>\n<\/div><\/a><\/div>\n<div id=\"post-34211\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"http:\/\/craftwork.design\/teammates-illustrations\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-10.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Teammates Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Meet brand-new 12 flat and stylish illustrations for your landings<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 10, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-34003\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-article\"><a target=\"_blank\" href=\"https:\/\/medium.com\/@Shepovalovdenis\/top-10-wireframe-kits-and-tools-94fd65380901\" class=\"news-inner-wrap-view row wpnaw-news-no-image c2\">\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-article\">Article<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Top 10 Wireframe Kits and Tools<\/h3>\n<div class=\"news-content-excerpt\">\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 8, 2019<\/div>\n<\/div>\n<\/div><\/a><\/div>\n<div id=\"post-33365\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/raw-illustrations\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-300x225.png 300w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Introducing RAW<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Create beautiful landing pages and mobile screens with these cool illustrations<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">October 1, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-32637\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/dance-illustrations\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-8.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Dance Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">20 fun and crisp illustrations for your next project<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">September 24, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-31333\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/usesuperb.com\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-3.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Superb Illustrations<\/h3>\n<div class=\"news-content-excerpt news-short-content\">We are so excited to introduce you a new pack of 20 amazing full vector illustrations<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">September 11, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-29922\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"https:\/\/method.blue\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-7.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-new-resource\">New Resource<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Method 3 Release<\/h3>\n<div class=\"news-content-excerpt news-short-content\">We are happy to present you a brand new Method Wireframe Kit. 120 Blocks, 11 Categories, Figma, Sketch.<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">September 2, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-28617\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-first has-date width-news\"><a target=\"_blank\" href=\"https:\/\/craftwork.design\/bites\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-6.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-update\">Update<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">Bites 2<\/h3>\n<div class=\"news-content-excerpt news-short-content\">Second version of Bites. Added 40 new popups, powerful documentation and free Figma sources<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">August 22, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<div id=\"post-28614\" class=\"news type-news news-col-2 wpnaw-news-post wpnaw-last has-date width-news\"><a target=\"_blank\" href=\"http:\/\/craftwork.design\/ios-jetpack-2\/\" class=\"news-inner-wrap-view row\">\n<div class=\"news-thumb col-md-4 col-sm-4 d-none grid-news-thumb\"><noscript>\n<p><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-768x576.png\" class=\"attachment-medium_large size-medium_large wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1.png 1160w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"640\" height=\"480\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-768x576.png\" class=\"lazyload attachment-medium_large size-medium_large wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1-1024x768.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-1.png 1160w\" data-sizes=\"(max-width: 640px) 100vw, 640px\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"news-content col-md-12 col-sm-12 px-0 d-flex align-items-end flex-column h-100\">\n<div class=\"grid-cat-post w-100 category-update\">Update<\/div>\n<div class=\"post-content-text w-100\">\n<h3 class=\"news-title\">iOS Jetpack 2<\/h3>\n<div class=\"news-content-excerpt news-short-content\">We created 8 new different styles with 5 screens in 5 dimensions<\/div>\n<\/div>\n<div class=\"grid-date-post mt-auto w-100\">August 18, 2019<\/div>\n<\/div>\n<\/a><\/div>\n<\/div>\n<\/div>\n<div class=\"brands wow fadeInUp container row col-lg-10 offset-lg-1\" data-wow-delay=\"0.2s\">\n<h2 class=\"section-title text-center\">Trusted by great companies<\/h2>\n<div class=\"row companies\">\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img class=\"w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_00.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_00.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img class=\"w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_01.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_01.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img width=\"170\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_02.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload\" width=\"170\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_02.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img width=\"140\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_03.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload\" width=\"140\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_03.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img class=\"w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_04.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_04.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img width=\"160\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_05.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload\" width=\"160\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_05.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<div class=\"col px-0 align-self-center text-center\"><noscript>\n<p><img class=\"w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_06.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img class=\"lazyload w-100\" src=\"https:\/\/craftwork.design\/wp-content\/themes\/stanleywp\/icon\/slice_logo_06.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/div>\n<\/div>\n<\/div>\n<section id=\"ux-ui-systems\" class=\"ux-ui-systems-section home-section\"><div class=\"container row col-md-12\">\n<h2 class=\"section-title text-uppercase wow fadeInUp\">UX\/UI Systems<\/h2>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-6 column-2 col-sm-12 wow fadeInUp product-41939\">\n<article id=\"post-41939\" class=\"post-41939 post type-post status-publish format-standard has-post-thumbnail hentry category-feature category-figma category-sketch category-ui-kits tag-block tag-colorful tag-component tag-craftwork tag-customization tag-design tag-figma tag-kit tag-landing tag-layout tag-modern tag-object tag-pages tag-singleton tag-sketch tag-smart tag-styleguide tag-stylish tag-ui tag-vector\"><section class=\"post-entry text-white c5\"><a class=\"img-link c4\" href=\"https:\/\/craftwork.design\/singleton\/\" title=\"Singleton\"><noscript>\n<p><img width=\"1140\" height=\"854\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570.png 1140w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-1024x767.png 1024w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1140\" height=\"854\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570.png 1140w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-570-1024x767.png 1024w\" data-sizes=\"(max-width: 1140px) 100vw, 1140px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom row col-md-10 col-sm-10\">\n<div class=\"short-description\">\n<p class=\"mb-0\">Stylish and playful layouts\u00a0for\u00a0beautiful landing websites<\/p>\n<\/div>\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-6 column-2 col-sm-12 wow fadeInUp product-29909\">\n<article id=\"post-29909\" class=\"post-29909 post type-post status-publish format-standard has-post-thumbnail hentry category-feature category-figma category-sketch category-ui-kits category-wireframes tag-blocks tag-bootstrap tag-clean tag-constructor tag-craftwork tag-figma tag-free tag-freebie-website tag-google-fonts tag-grid tag-landing tag-landing-page tag-lego tag-prototype tag-prototyping tag-sketch tag-ui tag-ui-design tag-ux tag-ux-design tag-ux-kit tag-vector tag-web tag-web-design tag-wire-frame tag-wireframe\"><section class=\"post-entry text-black c6\"><a class=\"img-link c4\" href=\"https:\/\/craftwork.design\/method-wireframe-kit-3\/\" title=\"Method\"><noscript>\n<p><img width=\"1160\" height=\"870\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Method-3.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Method-3.png 1160w, https:\/\/craftwork.design\/wp-content\/uploads\/Method-3-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Method-3-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Method-3-1024x768.png 1024w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1160\" height=\"870\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Method-3.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Method-3.png 1160w, https:\/\/craftwork.design\/wp-content\/uploads\/Method-3-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Method-3-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Method-3-1024x768.png 1024w\" data-sizes=\"(max-width: 1160px) 100vw, 1160px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom row col-md-10 col-sm-10\">\n<div class=\"short-description\">\n<p class=\"mb-0\">Find your method of creating great prototypes for desktop and mobiles<\/p>\n<\/div>\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<\/section><section id=\"illustrations\" class=\"illustrations-section home-section\"><div class=\"container row col-md-12\">\n<h2 class=\"section-title text-uppercase wow fadeInUp\">Illustrations<\/h2>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-46139\">\n<article id=\"post-46139\" class=\"post-46139 post type-post status-publish format-standard has-post-thumbnail hentry category-illustrations tag-3d tag-3ddd tag-apps tag-blender tag-colorful tag-cool tag-craftwork tag-delete tag-delivery tag-design tag-education tag-illustrations tag-incredible tag-loading tag-modern tag-notification tag-playful tag-png tag-presentations tag-trendy tag-unique tag-volumetric tag-waiting tag-websites tag-workflow\"><section class=\"post-entry text-white c7\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/3ddd-illustrations\/\" title=\"3DDD Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-6.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-6.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-6-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-6.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-6.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-6-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">An incredible set of 15 \u0441olorful &amp; playful 3D illustrations<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-43677\">\n<article id=\"post-43677\" class=\"post-43677 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations tag-750 tag-application tag-bright tag-bubble tag-colorful tag-craftwork tag-delivery tag-design tag-error tag-funny tag-illustrations tag-it tag-landing tag-log-out tag-modern tag-scenes tag-sign-in tag-sign-up tag-stories tag-unique tag-vector tag-website tag-world\"><section class=\"post-entry text-black c8\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/bubble-illustrations\/\" title=\"Bubble Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-5.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-5.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-5-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-5.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-5.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-5-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">25 bright and modern stories from IT world<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-12677\">\n<article id=\"post-12677\" class=\"post-12677 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations category-uncategorized tag-750 tag-ai tag-app tag-application tag-background tag-craftwork tag-eps tag-error-404 tag-flat tag-illustration tag-illustrator tag-it tag-landing tag-site tag-startup tag-story tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c9\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/struct-illustration\/\" title=\"Struct Illustrations 2\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Struct-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Struct-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Struct-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Struct-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Struct-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Struct-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Incredible IT world scenes for landings and mobile headers<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-41099\">\n<article id=\"post-41099\" class=\"post-41099 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations tag-750 tag-ai tag-colorful tag-craftwork tag-daily tag-design tag-detailed tag-illustration tag-it tag-lifestyle tag-old-shool tag-png tag-research tag-scene tag-sport tag-study tag-stylish tag-svg tag-thursday tag-web tag-work tag-world\"><section class=\"post-entry text-black c8\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/thursday-illustrations\/\" title=\"Thursday Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-4.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-4.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-4-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-4.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-4.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-4-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Stylish and old school illustrations for landings and applications<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-39257\">\n<article id=\"post-39257\" class=\"post-39257 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations category-sketch tag-ai tag-blink tag-colorful tag-contrast tag-craftwork tag-detailed tag-illustration tag-life tag-lifestyle tag-modern tag-png tag-research tag-scene tag-sketch tag-sport tag-study tag-stylish tag-svg tag-web tag-work tag-world\"><section class=\"post-entry text-black c10\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/blink-illustrations\/\" title=\"Blink Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Small.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Small.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Small-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Small.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Small.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Small-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Stylish set of illustrations with modern and colorful characters<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-38071\">\n<article id=\"post-38071\" class=\"post-38071 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations tag-ai tag-business tag-colorful tag-contrast tag-craftwork tag-daily tag-illustration tag-it tag-life tag-modern tag-no-gravity tag-png tag-research tag-scene tag-simple tag-sport tag-study tag-stylish tag-svg tag-ui-design tag-vector tag-web tag-work tag-world\"><section class=\"post-entry text-black c11\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/no-gravity-illustrations\/\" title=\"No Gravity Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-3.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-3.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-3-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-3.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-3.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Small-3-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Carefully crafted header illustrations for landings and mobile screens<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-35318\">\n<article id=\"post-35318\" class=\"post-35318 post type-post status-publish format-standard has-post-thumbnail hentry category-figma category-illustrations category-sketch tag-craftwork tag-creepy tag-devil tag-ghost tag-grave tag-halloween tag-holiday tag-illustration tag-landing tag-october tag-png tag-pumpkin tag-reckless tag-scary tag-sketch tag-svg tag-trick-or-treats tag-vector tag-web tag-witch tag-zombie\"><section class=\"post-entry text-black c8\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/spooky-illustrations\/\" title=\"Spooky Illustrations\"><noscript>\n<p><img width=\"1160\" height=\"870\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big.png 1160w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-1024x768.png 1024w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1160\" height=\"870\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big.png 1160w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-768x576.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Covers-Craftwork-Big-1024x768.png 1024w\" data-sizes=\"(max-width: 1160px) 100vw, 1160px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Spooky set of illustrations with creepy style for Halloween<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-34194\">\n<article id=\"post-34194\" class=\"post-34194 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations tag-ai tag-business tag-cloud tag-collaboration tag-communication tag-coworking tag-craftwork tag-discussion tag-illustration tag-infographics tag-it tag-landing tag-management tag-networking tag-png tag-svg tag-task tag-team tag-vector tag-web tag-work\"><section class=\"post-entry text-black c12\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/teammates-illustrations\/\" title=\"Teammates Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-11-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-11-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-11-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-11-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-11-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Cover-Craftwork-Big-11-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Awesome set of illustrations with lovely style and IT world scenes<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-33312\">\n<article id=\"post-33312\" class=\"post-33312 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-figma category-illustrations category-sketch tag-750 tag-app tag-application tag-background tag-craftwork tag-design tag-designer tag-eps tag-error-404 tag-illustration tag-image tag-it tag-landing tag-scene tag-site tag-startup tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c8\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/raw-illustrations\/\" title=\"RAW Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Raw-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Raw-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Raw-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Raw-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Raw-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Raw-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Incredible set of shapes and edges for your projects<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-32606\">\n<article id=\"post-32606\" class=\"post-32606 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations tag-750 tag-app tag-application tag-background tag-craftwork tag-eps tag-error-404 tag-illustration tag-image tag-it tag-landing tag-picture tag-scene tag-site tag-startup tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c13\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/dance-illustrations\/\" title=\"Dance Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dance-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dance-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Dance-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dance-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dance-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Dance-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Make your landing more stunning with a set of 20 fun and crisp illustrations<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-31324\">\n<article id=\"post-31324\" class=\"post-31324 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations tag-boy tag-character tag-colorful tag-colourful tag-components tag-design tag-flat tag-generator tag-girl tag-graphic tag-illustration tag-illustrations tag-illustrator tag-image tag-noice tag-pic tag-png tag-robot tag-scene tag-shape tag-uidesign tag-vector tag-web\"><section class=\"post-entry text-white c14\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/superb-illustrations\/\" title=\"Superb Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Superb-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Superb-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Superb-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Superb-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Superb-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Superb-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Amazing pack of 20 amazing full vector illustrations for your apps and websites<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-24661\">\n<article id=\"post-24661\" class=\"post-24661 post type-post status-publish format-standard has-post-thumbnail hentry category-illustrations category-photoshop tag-750 tag-app tag-application tag-background tag-character tag-colourful tag-craftwork tag-error-404 tag-illustration tag-illustrations tag-image tag-it tag-landing tag-man tag-photoshop tag-picture tag-png tag-raster tag-scene tag-site tag-startup tag-tech tag-walkthrough tag-web tag-web-page tag-website tag-woman\"><section class=\"post-entry text-black c15\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/aurora-illustrations\/\" title=\"Aurora Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Aurora-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Aurora-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Aurora-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Aurora-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Aurora-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Aurora-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Pack of 20 artfully drawn raster and layered illustrations<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-21052\">\n<article id=\"post-21052\" class=\"post-21052 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations tag-750 tag-app tag-application tag-background tag-character tag-colourful tag-craftwork tag-eps tag-error-404 tag-hands tag-home tag-illustration tag-illustrations tag-image tag-it tag-landing tag-man tag-picture tag-scene tag-site tag-startup tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c16\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/oliver-illustrations\/\" title=\"Oliver Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Oliver-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Oliver-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Oliver-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Oliver-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Oliver-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Oliver-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Unique and colorful illustrations to embody your fantasy into reality<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-17720\">\n<article id=\"post-17720\" class=\"post-17720 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations tag-750 tag-app tag-application tag-background tag-craftwork tag-eps tag-error-404 tag-home tag-illustration tag-image tag-it tag-landing tag-picture tag-scene tag-site tag-startup tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-white c17\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/silky-illustrations\/\" title=\"Silky Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Silky-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Silky-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Silky-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Silky-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Silky-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Silky-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Give life to your project with trendy Silky illustrations<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-13425\">\n<article id=\"post-13425\" class=\"post-13425 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations category-sketch tag-750 tag-ai tag-app tag-application tag-background tag-eps tag-error-404 tag-flat tag-icon tag-icons tag-illustration tag-illustrations tag-illustrator tag-image tag-it tag-landing tag-outline tag-png tag-shape tag-site tag-sketch tag-startup tag-story tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c8\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/oslo-illustrations\/\" title=\"OSLO Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/OSLO-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/OSLO-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/OSLO-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/OSLO-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/OSLO-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/OSLO-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Awesome illustrations for your next landing<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-11498\">\n<article id=\"post-11498\" class=\"post-11498 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations tag-750 tag-abstract tag-ai tag-ai-sources tag-app tag-application tag-background tag-craftwork tag-eps tag-error-404 tag-flat tag-illustration tag-illustrator tag-it tag-landing tag-light tag-png tag-site tag-startup tag-story tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c18\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/follow-me-illustrations\/\" title=\"Follow Me Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Follow-Me-Cover-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Follow-Me-Cover-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Follow-Me-Cover-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Follow-Me-Cover-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Follow-Me-Cover-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Follow-Me-Cover-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Cute characters with big hearts for your mobile and web projects<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-11234\">\n<article id=\"post-11234\" class=\"post-11234 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations tag-750 tag-app tag-application tag-background tag-character tag-craftwork tag-eps tag-error-404 tag-home tag-illustration tag-illustrations tag-image tag-it tag-landing tag-site tag-startup tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-white c19\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/no-comments-illustrations\/\" title=\"No Comments\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/No-Comments-Cover-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/No-Comments-Cover-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/No-Comments-Cover-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/No-Comments-Cover-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/No-Comments-Cover-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/No-Comments-Cover-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Fun and crisp characters for apps, websites or presentations<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-11177\">\n<article id=\"post-11177\" class=\"post-11177 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations tag-750 tag-app tag-application tag-background tag-craftwork tag-eps tag-error-404 tag-home tag-illustration tag-image tag-it tag-landing tag-picture tag-scene tag-site tag-startup tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c20\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/smartsharp-illustrations\/\" title=\"Smartsharp Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Smartsharp-Illustrations-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Smartsharp-Illustrations-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Smartsharp-Illustrations-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Smartsharp-Illustrations-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Smartsharp-Illustrations-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Smartsharp-Illustrations-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">High-detailed business scenes for landing pages\u2019 headers and mobile screens<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-11140\">\n<article id=\"post-11140\" class=\"post-11140 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-feature category-illustrations category-sketch tag-750 tag-ai tag-app tag-application tag-background tag-couple tag-craftwork tag-eps tag-error-404 tag-flat tag-food tag-girl tag-home tag-illustration tag-illustrator tag-it tag-landing tag-mother tag-music tag-photos tag-site tag-son tag-startup tag-story tag-student tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website tag-work tag-yoga\"><section class=\"post-entry text-black c21\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/guacamole\/\" title=\"Guacamole Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Guacamole-Cover-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Guacamole-Cover-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Guacamole-Cover-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Guacamole-Cover-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Guacamole-Cover-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Guacamole-Cover-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">20 warm and homely illustrations for funny interfaces<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-9438\">\n<article id=\"post-9438\" class=\"post-9438 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-figma category-illustrations category-sketch tag-750 tag-ai tag-app tag-application tag-background tag-craftwork tag-eps tag-error-404 tag-flat tag-illustration tag-illustrator tag-it tag-landing tag-site tag-startup tag-story tag-svg tag-tech tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c8\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/unruly-illustrations\/\" title=\"Unruly Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Unruly-Cover.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Unruly-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Unruly-Cover-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Unruly-Cover.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Unruly-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Unruly-Cover-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">20 bright and carefully designed illustrations to bring your startup to life<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-7658\">\n<article id=\"post-7658\" class=\"post-7658 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-figma category-illustrations category-sketch tag-750 tag-app tag-application tag-craftwork tag-delivery tag-error tag-error-404 tag-illustration tag-illustrator tag-it tag-landing tag-login tag-logout tag-payment tag-registration tag-startup tag-subscribing tag-success tag-tech tag-vector tag-web tag-web-page\"><section class=\"post-entry text-black c22\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/candy-shop\/\" title=\"Candy Shop Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Candy-Shop-Cover.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Candy-Shop-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Candy-Shop-Cover-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Candy-Shop-Cover.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Candy-Shop-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Candy-Shop-Cover-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">20 vector illustrations for a website, application, or presentation<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-7335\">\n<article id=\"post-7335\" class=\"post-7335 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-figma category-illustrations category-photoshop category-sketch category-xd tag-ai tag-app tag-application tag-background tag-craftwork tag-eps tag-flat tag-illustration tag-illustrator tag-landing tag-site tag-story tag-svg tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c23\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/stubborn\/\" title=\"Stubborn Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Stubborn-Cover.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Stubborn-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Stubborn-Cover-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Stubborn-Cover.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Stubborn-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Stubborn-Cover-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">30 little exciting little stories from<br \/>\nthe startup world<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-6458\">\n<article id=\"post-6458\" class=\"post-6458 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-illustrations category-photoshop tag-ai tag-app tag-application tag-background tag-colourful tag-craftwork tag-eps tag-gradient tag-illustration tag-illustrator tag-landing tag-noise tag-site tag-svg tag-vector tag-walkthrough tag-web tag-web-page tag-website\"><section class=\"post-entry text-black c24\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/brave-illustrations\/\" title=\"Brave Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/brave_cover-1024x768-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/brave_cover-1024x768-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/brave_cover-1024x768-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/brave_cover-1024x768-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/brave_cover-1024x768-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/brave_cover-1024x768-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Incredible set of 20 illustrations with<br \/>\ntheir own character.<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-4 column-3 col-sm-12 wow fadeInUp product-6024\">\n<article id=\"post-6024\" class=\"post-6024 post type-post status-publish format-standard has-post-thumbnail hentry category-illustrations category-sketch tag-750 tag-app tag-apple tag-background tag-colourful tag-combination tag-constructor tag-craftwork tag-eps tag-icon tag-icons tag-illustration tag-landing tag-mac tag-man tag-objects tag-outline tag-scene tag-site tag-sketch tag-svg tag-vector tag-web tag-website tag-woman\"><section class=\"post-entry text-black c25\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/handsome-illustrations\/\" title=\"Handsome Illustrations\"><noscript>\n<p><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Handsome-Cover.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Handsome-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Handsome-Cover-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"524\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Handsome-Cover.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Handsome-Cover.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Handsome-Cover-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0\">Create unique illustrations for a<br \/>\nwebsite &amp; applications.<\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<\/section><section id=\"templates\" class=\"templates-section home-section\"><div class=\"container row col-md-12\">\n<h2 class=\"section-title text-uppercase wow fadeInUp\">Templates<\/h2>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-6 column-2 col-sm-12 wow fadeInUp product-27442\">\n<article id=\"post-27442\" class=\"post-27442 post type-post status-publish format-standard has-post-thumbnail hentry category-feature category-figma category-photoshop category-sketch category-templates category-uncategorized tag-adobe tag-app tag-app-design tag-app-store tag-apple tag-application tag-appstore tag-figma tag-ios tag-iphone tag-iphone-x tag-mobile tag-photoshop tag-screenshot tag-sketch tag-template tag-ui\"><section class=\"post-entry text-white c26\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/ios-jetpack-2\/\" title=\"iOS Jetpack 2\"><noscript>\n<p><img width=\"1024\" height=\"768\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack-768x576.png 768w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1024\" height=\"768\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/ios-jetpack-768x576.png 768w\" data-sizes=\"(max-width: 1024px) 100vw, 1024px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom row col-md-10 col-sm-10\">\n<div class=\"short-description\">\n<p class=\"mb-0\">\u0421reate your perfect App Store presentation for your app in a fast and beautiful way<\/p>\n<\/div>\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/photoshop\/\">Photoshop<\/a> <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-6 column-2 col-sm-12 wow fadeInUp product-10507\">\n<article id=\"post-10507\" class=\"post-10507 post type-post status-publish format-standard has-post-thumbnail hentry category-feature category-html category-templates category-xd tag-blog tag-code tag-cookies tag-craftwork tag-css tag-html tag-modal tag-modals tag-newsletters tag-notification tag-popup tag-popups tag-sign-up tag-simple tag-site tag-subscribe tag-subscribing tag-web\"><section class=\"post-entry text-white c27\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/bites\/\" title=\"Bites: Modals and Popups\"><noscript>\n<p><img width=\"1024\" height=\"768\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/bites.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/bites.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/bites-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/bites-768x576.png 768w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1024\" height=\"768\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/bites.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/bites.png 1024w, https:\/\/craftwork.design\/wp-content\/uploads\/bites-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/bites-768x576.png 768w\" data-sizes=\"(max-width: 1024px) 100vw, 1024px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom row col-md-10 col-sm-10\">\n<div class=\"short-description\">\n<p class=\"mb-0\">Use our cute popups to grab customers attention without the annoyance<\/p>\n<\/div>\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a> <a href=\"https:\/\/craftwork.design\/html\/\">HTML<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<\/section><section id=\"mockups\" class=\"mockups-section home-section\"><div class=\"container row col-md-12\">\n<h2 class=\"section-title text-uppercase wow fadeInUp\">Mockups<\/h2>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-6 column-2 col-sm-12 wow fadeInUp product-10506\">\n<article id=\"post-10506\" class=\"post-10506 post type-post status-publish format-standard has-post-thumbnail hentry category-feature category-mockups category-photoshop category-sketch tag-android tag-angle tag-apple tag-black tag-clay tag-craftwork tag-front tag-galaxy tag-hero tag-ios tag-ipad tag-iphone tag-iwatch tag-library tag-mockup tag-mockups tag-perspective tag-pink tag-pixel tag-presentation tag-screen tag-silver tag-sketch tag-symbols tag-ui tag-view tag-web-ui\"><section class=\"post-entry position-relative text-white c28\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/hero-mobile-mockups\/\" title=\"HERO Mobile Mockups\"><noscript>\n<p><img width=\"1140\" height=\"854\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover.png 1140w, https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover-1024x767.png 1024w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1140\" height=\"854\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover.png 1140w, https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Hero-Cover-1024x767.png 1024w\" data-sizes=\"(max-width: 1140px) 100vw, 1140px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom row col-md-10 col-sm-10\">\n<div class=\"short-description\">\n<p class=\"mb-0\">High-quality popular mobile device mockups in seven angles<\/p>\n<\/div>\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/photoshop\/\">Photoshop<\/a> <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-6 column-2 col-sm-12 wow fadeInUp product-6618\">\n<article id=\"post-6618\" class=\"post-6618 post type-post status-publish format-standard has-post-thumbnail hentry category-feature category-figma category-mockups category-photoshop category-sketch category-xd tag-android tag-browser tag-chrome tag-craftwork tag-figma tag-firefox tag-frame tag-frrames tag-galaxy tag-ipad tag-iphone tag-mock-up tag-mockup tag-photoshop tag-safari tag-scene tag-sketch tag-vector tag-xd\"><section class=\"post-entry position-relative text-white c29\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/frrames-mockups\/\" title=\"Frrames Mockups\"><noscript>\n<p><img width=\"1140\" height=\"854\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Frrames.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Frrames.png 1140w, https:\/\/craftwork.design\/wp-content\/uploads\/Frrames-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Frrames-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Frrames-1024x767.png 1024w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"1140\" height=\"854\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Frrames.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Frrames.png 1140w, https:\/\/craftwork.design\/wp-content\/uploads\/Frrames-300x225.png 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Frrames-768x575.png 768w, https:\/\/craftwork.design\/wp-content\/uploads\/Frrames-1024x767.png 1024w\" data-sizes=\"(max-width: 1140px) 100vw, 1140px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom row col-md-10 col-sm-10\">\n<div class=\"short-description\">\n<p class=\"mb-0\">Perfectly crafted responsive windows mockups for your ideal presentation<\/p>\n<\/div>\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/photoshop\/\">Photoshop<\/a> <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a> <a href=\"https:\/\/craftwork.design\/xd\/\">XD<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<\/section><div class=\"all-products wow fadeInUp ajax-load-more-wrap default\">\n<div aria-live=\"polite\" class=\"alm-listing alm-ajax\" data-container-type=\"div\" data-repeater=\"default\" data-post-type=\"post\" data-category-not-in=\"875,876,890\" data-order=\"DESC\" data-orderby=\"date\" data-offset=\"0\" data-posts-per-page=\"15\" data-scroll=\"false\" data-pause=\"true\" data-button-label=\"Explore early products\" data-transition=\"fade\" data-images-loaded=\"false\">\n<div class=\"alm-btn-wrap c30\"><button class=\"alm-load-more-btn more\">Explore early products<\/button><\/div>\n<\/div>\n<section id=\"freebies\" class=\"all-products\"><div class=\"container row col-md-12\">\n<h2 class=\"section-title text-uppercase wow fadeInUp\">Free design resources<\/h2>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-35963\">\n<article id=\"post-35963\" class=\"post-35963 post type-post status-publish format-standard has-post-thumbnail hentry category-figma category-freebies category-sketch tag-chat tag-colorful tag-conversation tag-coworking tag-craftwork tag-daily tag-education tag-illustration tag-infographics tag-it tag-life tag-modern tag-research tag-scene tag-shopping tag-sketch tag-smash tag-sport tag-stylish tag-svg tag-web tag-work tag-world\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/smash-illustrations\/\" title=\"Smash Illustrations\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png\" class=\"img-border wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png\" class=\"lazyload img-border wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/Craftwork-Cover-4-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-22841\">\n<article id=\"post-22841\" class=\"post-22841 post type-post status-publish format-standard has-post-thumbnail hentry category-freebies category-wireframes tag-app tag-application tag-craftwork tag-desktop tag-dribbble tag-figma tag-flow-chart tag-flowchart tag-free tag-freebie tag-ios tag-landing tag-mobile tag-prototype tag-scheme tag-site tag-sketch tag-user-flow tag-ux tag-uxdesign tag-vector tag-wireframe\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/greyhound-flowcharts-2\/\" title=\"Greyhound Flowcharts\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/greyhound-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/greyhound-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/greyhound-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/greyhound-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/greyhound-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/greyhound-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-20132\">\n<article id=\"post-20132\" class=\"post-20132 post type-post status-publish format-standard has-post-thumbnail hentry category-figma category-freebies category-sketch tag-bootstratp tag-clean tag-concept tag-free tag-freebie tag-mobile tag-prototype tag-redesign tag-simple tag-ui tag-ui-design tag-vector tag-web tag-youtube\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/youtube-redesign\/\" title=\"Youtube ReDesign\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/youtube-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/youtube-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/youtube-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/youtube-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/youtube-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/youtube-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-14250\">\n<article id=\"post-14250\" class=\"post-14250 post type-post status-publish format-standard has-post-thumbnail hentry category-ai category-figma category-freebies category-sketch tag-boy tag-character tag-colorful tag-components tag-constructor tag-craftwork tag-design tag-figma tag-flat tag-free tag-freebie tag-generator tag-girl tag-illustration tag-illustrations tag-illustrator tag-image tag-nested tag-pic tag-robot tag-sketch tag-symbols tag-uidesign tag-vector tag-web\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/stubborn-generator\/\" title=\"Stubborn Generator\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/generator-1.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/generator-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/generator-1-300x225.png 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/generator-1.png\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/generator-1.png 700w, https:\/\/craftwork.design\/wp-content\/uploads\/generator-1-300x225.png 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a> <a href=\"https:\/\/craftwork.design\/illustrations\/\">Illustrator<\/a> <a href=\"https:\/\/craftwork.design\/figma\/\">Figma<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<div class=\"container position-relative row products-row\">\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-5403\">\n<article id=\"post-5403\" class=\"post-5403 post type-post status-publish format-standard has-post-thumbnail hentry category-freebies category-sketch category-ui-kits tag-adaptive tag-app tag-apple tag-application tag-craftwork tag-design tag-details tag-elements tag-feed tag-free tag-freebie tag-ios tag-ios-11 tag-light tag-menu tag-minimalistic tag-mobile tag-mobile-app tag-multimedia tag-pixel-perfect tag-profile tag-retina tag-screen tag-sketch tag-split tag-symbols tag-ui tag-ui-kit tag-ux tag-ux-kit tag-walkthrough tag-white\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/split-ios-ui-kit\/\" title=\"\u26a1Split iOS UI Kit\"><noscript>\n<p><img width=\"800\" height=\"600\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Free-1.jpg\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Free-1.jpg 800w, https:\/\/craftwork.design\/wp-content\/uploads\/Free-1-300x225.jpg 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Free-1-768x576.jpg 768w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"800\" height=\"600\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Free-1.jpg\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Free-1.jpg 800w, https:\/\/craftwork.design\/wp-content\/uploads\/Free-1-300x225.jpg 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Free-1-768x576.jpg 768w\" data-sizes=\"(max-width: 800px) 100vw, 800px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/sketch\/\">Sketch<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-4240\">\n<article id=\"post-4240\" class=\"post-4240 post type-post status-publish format-standard has-post-thumbnail hentry category-freebies category-landings category-photoshop tag-clean tag-colored tag-craftwork tag-flat tag-iphone tag-landing tag-landing-page tag-light tag-psd tag-simple tag-web\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/landing-page-day-3\/\" title=\"Landing Page powered by The Rapture\"><noscript>\n<p><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/700x5253.jpg\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/700x5253.jpg 700w, https:\/\/craftwork.design\/wp-content\/uploads\/700x5253-300x225.jpg 300w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"700\" height=\"525\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/700x5253.jpg\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/700x5253.jpg 700w, https:\/\/craftwork.design\/wp-content\/uploads\/700x5253-300x225.jpg 300w\" data-sizes=\"(max-width: 700px) 100vw, 700px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/photoshop\/\">Photoshop<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<div class=\"col-lg-3 column-4 col-sm-12 wow fadeInUp product-753\">\n<article id=\"post-753\" class=\"post-753 post type-post status-publish format-standard has-post-thumbnail hentry category-freebies category-photoshop category-ui-kits tag-clean tag-colorfull tag-dashboard tag-flat tag-psd tag-simple tag-ui tag-web tag-widget\"><section class=\"post-entry\"><a class=\"img-link\" href=\"https:\/\/craftwork.design\/dashboard-ui-elements\/\" title=\"Dashboard UI Elements\"><noscript>\n<p><img width=\"800\" height=\"600\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600.jpg\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600.jpg 800w, https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600-300x225.jpg 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600-768x576.jpg 768w\" referrerpolicy=\"no-referrer\" \/><\/p>\n<\/noscript><img width=\"800\" height=\"600\" src=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600.jpg\" class=\"lazyload attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" data-srcset=\"https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600.jpg 800w, https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600-300x225.jpg 300w, https:\/\/craftwork.design\/wp-content\/uploads\/Dashboard-800x600-768x576.jpg 768w\" data-sizes=\"(max-width: 800px) 100vw, 800px\" referrerpolicy=\"no-referrer\" \/><\/a>\n<div class=\"post-bottom short-description mb-0\">\n<p class=\"mb-0 extension-text\">Compatible with <a href=\"https:\/\/craftwork.design\/photoshop\/\">Photoshop<\/a><\/p>\n<\/div>\n<\/section><\/article><\/div>\n<\/div>\n<\/section><div class=\"container home-bottom row\">\n<div class=\"col-lg-4 col-md-12 banner text-center position-relative wow fadeInUp banner-inner position-relative\">\n<p class=\"banner-suptitle text-uppercase\">Become a Member<\/p>\n<h2 class=\"banner-title text-white\">Get access to all assets we made and will make for one year<\/h2>\n<a href=\"https:\/\/craftwork.design\/unlimited-access\/\" class=\"banner-btn position-relative\">Go Unlimited For $99<\/a><\/div>\n<div class=\"col-lg-8 col-md-12 mail-chimp-block text-center wow fadeInUp\">\n<h3>Sign up craftwork digest and get new UI releases every tuesday<\/h3>\n<form id=\"mc4wp-form-1\" class=\"mc4wp-form mc4wp-form-2988 mc4wp-form-basic\" method=\"post\" data-id=\"2988\" data-name=\"Default sign-up form\" action=\"action\">\n<div class=\"mc4wp-form-fields\">\n<div class=\"col-md-6\"><input type=\"email\" class=\"form-control text-center\" name=\"email\" placeholder=\"Your e-mail to join\" required=\"required\" \/><\/div>\n<div class=\"col-md-6\"><input type=\"submit\" class=\"btn w-100 btn-primary\" value=\"Subscribe\" \/><\/div>\n<\/div>\n<div><label class=\"c31\">Leave this field empty if you're human: <input type=\"text\" name=\"_mc4wp_honeypot\" value=\"\" tabindex=\"-1\" autocomplete=\"off\" \/><\/label><input type=\"hidden\" name=\"_mc4wp_timestamp\" value=\"1579019100\" \/><input type=\"hidden\" name=\"_mc4wp_form_id\" value=\"2988\" \/><input type=\"hidden\" name=\"_mc4wp_form_element_id\" value=\"mc4wp-form-1\" \/><\/div><\/form>\n<p class=\"mb-0\">No spam. Only releases, updates and discounts<\/p>\n<\/div>\n<\/div>\n<\/div>","created_at":"2020-01-14T16:30:46+0000","updated_at":"2020-01-14T16:31:17+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"craftwork.design","preview_picture":"https:\/\/craftwork.design\/wp-content\/uploads\/Open-Graph-Craftwork-10-06-2019.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1111,"title":"Ironsworn - Tabletop RPG","url":"https:\/\/www.ironswornrpg.com\/","content":"<section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c34\" id=\"comp-jwpaxixx\"><div id=\"comp-jwpaxixxinlineContent\" class=\"strc1inlineContent strc1 c33\"><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxiycinlineContent\"><p><\/p><h2 class=\"font_2 c19\">A Tabletop RPG of Perilous Quests<\/h2><p><\/p><h2 class=\"font_2 c19\">for Solo, Co-OP and Guided Play<\/h2><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c31\" id=\"comp-jyef1t2m\"><p><a href=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" target=\"_blank\" data-content=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" data-type=\"external\" rel=\"noopener noreferrer\" id=\"comp-jyef1t69link\" class=\"g-transparent-a style-jwpaxn0olink\">GET IRONSWORN<\/a><\/p><div title=\"\" data-is-responsive=\"false\" data-display-mode=\"fill\" data-content-padding-horizontal=\"0\" data-content-padding-vertical=\"0\" data-exact-height=\"53\" class=\"ca1 c28\" id=\"comp-jyef1t80\"><a href=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" target=\"_blank\" data-content=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" data-type=\"external\" rel=\"noopener noreferrer\" id=\"comp-jyef1t80link\" class=\"ca1link c27\"><img id=\"comp-jyef1t80imgimage\" class=\"c26\" alt=\"DriveThruRPG Logo\" data-type=\"image\" itemprop=\"image\" src=\"https:\/\/static.wixstatic.com\/media\/4db827_10cba336c0e64907b790327bef158ff2~mv2.png\/v1\/fill\/w_150,h_32,al_c,usm_0.66_1.00_0.01,blur_2\/4db827_10cba336c0e64907b790327bef158ff2~mv2.png\" referrerpolicy=\"no-referrer\" \/><\/a><\/div><div data-packed=\"true\" class=\"txtNew c30\" id=\"comp-jyef1t9q\"><p class=\"font_8 c29\"><a href=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?src=IronswornWeb\" target=\"_blank\" data-content=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?src=IronswornWeb\" data-type=\"external\" rel=\"noopener noreferrer\">Get the FREE PDF.<br \/>Also available in print.\u00a0<\/a><\/p><\/div><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c39\" id=\"comp-jwpaxj9e\"><div id=\"comp-jwpaxj9einlineContent\" class=\"strc1inlineContent mc1 mc1container c38\"><div class=\"mc1inlineContentParent mc1inlineContent txtNew c37\" id=\"comp-jwpaxj9oinlineContentParent\"><p class=\"font_7 c36\">WHAT'S NEW\u00a0 |\u00a0 <a href=\"https:\/\/ironsworn.podbean.com\/\" target=\"_blank\" data-content=\"https:\/\/ironsworn.podbean.com\/\" data-type=\"external\" rel=\"noopener noreferrer\">Listen to Ask the Oracle, the Official Ironsworn RPG Podcast<\/a><\/p><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c52\" id=\"comp-jwpaxjca\"><div id=\"comp-jwpaxjcainlineContent\" class=\"strc1inlineContent mc1 mc1container c51\"><div class=\"mc1inlineContentParent mc1inlineContent c50\" id=\"comp-jwpaxjclinlineContentParent\"><p><\/p><h3 class=\"font_3 c42\">YOU ARE IRONSWORN<\/h3><div data-packed=\"true\" class=\"txtNew c45\" id=\"comp-jwpaxjdf\"><p class=\"font_7 c44\">In the Ironsworn tabletop roleplaying game, you are a hero sworn to undertake perilous quests in the dark fantasy setting of the Ironlands.<\/p><p class=\"font_7 c44\"><br \/>Others live out their lives hardly venturing beyond the walls of their village or steading, but you are different. You will explore untracked wilds, fight desperate battles, forge bonds with isolated communities, and reveal the secrets of this harsh land.<\/p><p class=\"font_7 c44\">Are you ready to swear iron vows and see them fulfilled\u2014no matter the cost?<\/p><\/div><div data-packed=\"true\" class=\"txtNew c49\" id=\"comp-jwpaxje5\"><p class=\"font_7 c46\">WHAT'S INCLUDED?<\/p><p class=\"font_7 c47\">\u200b<\/p><ul class=\"font_7\"><li class=\"c48\">\n<p class=\"font_7 c47\">Quest-driven mechanics and fast, flexible character creation<\/p>\n<\/li>\n<li class=\"c48\">\n<p class=\"font_7 c47\">Story hooks and creative prompts to eliminate game prep and push your adventures forward\u2014with or without a GM<\/p>\n<\/li>\n<li class=\"c48\">\n<p class=\"font_7 c47\">A setting guide for the Ironlands, a rugged and treacherous peninsula at the edge of the known world<\/p>\n<\/li>\n<li class=\"c48\">\n<p class=\"font_7 c47\">Inspiring tools to customize your setting and fine-tune your game experience<\/p>\n<\/li>\n<li class=\"c48\">\n<p class=\"font_7 c47\">Extensive advice, examples, and options for exploring the story of your characters and their fateful vows<\/p>\n<\/li>\n<\/ul><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c73\" id=\"comp-jwpaxjhf\"><div id=\"comp-jwpaxjhfinlineContent\" class=\"strc1inlineContent mc1 mc1container c72\"><div class=\"mc1inlineContentParent mc1inlineContent c50\" id=\"comp-jwpaxjhqinlineContentParent\"><p><\/p><h3 class=\"font_3 c42\">THREE WAYS TO PLAY<\/h3><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c60\" id=\"comp-jwpaxjim\"><p><\/p><h4 class=\"font_4 c56\">GUIDED<\/h4><div data-packed=\"true\" class=\"txtNew c59\" id=\"comp-jwpaxjl9\"><p class=\"font_8 c58\">One or more players take the role of their characters, while a gamemaster (GM) moderates the session.<\/p><\/div><\/div><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c65\" id=\"comp-jwpaxjmo\"><p><\/p><h4 class=\"font_4 c56\">CO-OP<\/h4><div data-packed=\"true\" class=\"txtNew c64\" id=\"comp-jwpaxjoe\"><p class=\"font_8 c58\">You and one or more friends play together to overcome challenges and complete quests. No GM required.<\/p><\/div><\/div><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c70\" id=\"comp-jwpaxjpr\"><p><\/p><h4 class=\"font_4 c56\">SOLO<\/h4><div data-packed=\"true\" class=\"txtNew c69\" id=\"comp-jwpaxjrj\"><p class=\"font_8 c58\">You portray a lone character driven to fulfill vows in a dangerous world. Good luck!<\/p><\/div><\/div><div data-packed=\"true\" class=\"txtNew c71\" id=\"comp-jwpaxjx2\"><p class=\"font_8 c58\">Whatever mode you choose, you will create your character, make some decisions about the world you inhabit, and set the story in motion. When you encounter something dangerous or uncertain, your choices and the dice determine the outcome.<\/p><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c86\" id=\"comp-jwpaxk1e\"><div id=\"comp-jwpaxk1einlineContent\" class=\"strc1inlineContent c85\"><div data-content-width=\"488\" class=\"strc1 strc1balata c77\" id=\"comp-jwpaxk1r\"><img id=\"comp-jwpaxk1rbalatamediaimageimage\" class=\"c76\" alt=\"\" data-type=\"image\" itemprop=\"image\" src=\"https:\/\/static.wixstatic.com\/media\/4db827_c8a590bb32ad4d69a101655ffd1a4425~mv2.jpg\/v1\/fill\/w_293,h_415,al_c,q_80,usm_0.66_1.00_0.01,blur_2\/4db827_c8a590bb32ad4d69a101655ffd1a4425~mv2.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><div data-content-width=\"492\" class=\"strc1 c84\" id=\"comp-jwpaxk2m\"><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxk2minlineContent\"><div data-packed=\"true\" class=\"txtNew c81\" id=\"comp-jwpaxk2v\"><p class=\"font_8 c79\">Your Character<\/p><p class=\"font_8 c80\">\u200b<\/p><p class=\"font_8 c44\">Your character is composed of stats and trackers. These elements help you determine the outcome when you perform risky actions. As you complete quests and gain experience, you add new assets\u2014skills, animal companions, combat talents, and rituals \u2014which deepen your abilities and options.<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">Character creation is fast and flexible. Assign values to your stats. Pick some starting assets. Play.<\/p><\/div><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c95\" id=\"comp-jwpaxk7j\"><div id=\"comp-jwpaxk7jinlineContent\" class=\"strc1inlineContent strc1 c94\"><div data-page-id=\"ij2ei\" data-enable-video=\"true\" data-bg-effect-name=\"\" data-media-type=\"Image\" data-use-clip-path=\"\" data-needs-clipping=\"\" data-wix-video-layout=\"true\" class=\"strc1balata c91\" id=\"comp-jwpaxk84balata\"><div data-has-bg-effect=\"\" data-fitting=\"fill\" data-align=\"bottom\" data-container-id=\"comp-jwpaxk84\" data-page-id=\"ij2ei\" data-media-comp-type=\"image\" class=\"bgMedia c90\" id=\"comp-jwpaxk84balatamedia\"><img id=\"comp-jwpaxk84balatamediaimageimage\" class=\"c89\" alt=\"\" data-type=\"image\" itemprop=\"image\" src=\"https:\/\/static.wixstatic.com\/media\/4db827_ae12fe1ace914994860fbb84fb85ea60~mv2.jpg\/v1\/fill\/w_245,h_130,al_c,q_80,usm_0.66_1.00_0.01,blur_2\/4db827_ae12fe1ace914994860fbb84fb85ea60~mv2.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><\/div><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxk84inlineContent\"><p><\/p><h3 class=\"font_3 c42\">YOUR WORLD<\/h3><div data-packed=\"true\" class=\"txtNew c93\" id=\"comp-jwpaxk97\"><p class=\"font_7 c44\">The Ironlands setting is gritty fantasy on a rugged frontier.\u00a0Generations ago, your people settled here in the wake of a cataclysm which drove them from the Old World. Now, you live in isolated villages of wood and thatch at the edge of dark forests, or as nomads among the northern hills, or as seafarers along the rugged coast.\u00a0Your world is bounded by deep woods, violent seas, and imposing mountains where only the reckless or resolute venture.<\/p><p class=\"font_7 c44\">However, the default assumptions are only a starting point. You are encouraged to make Ironsworn your own, and to bend the\u00a0setting to your liking. Your version of the Ironlands will be unique because you\u2019ll define aspects such as\u00a0the history of your people, magic, mythic beasts, and much more. The choices you make will also help inspire the personal vows that drive your character.\u00a0<\/p><p class=\"font_7 c44\">\u200b<\/p><p class=\"font_7 c44\">You can also\u00a0ignore the default setting entirely and play in your own world, or in one inspired by media, history, or another RPG. The Ironsworn rules are flexible enough to accommodate many forms of fantasy and\u00a0historical fiction.<\/p><p class=\"font_7 c44\">\u200b<\/p><p class=\"font_7 c44\">\u200b<\/p><p class=\"font_7 c44\">\u200b<\/p><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c103\" id=\"comp-jwpaxkcm\"><div id=\"comp-jwpaxkcminlineContent\" class=\"strc1inlineContent c85\"><div data-content-width=\"488\" class=\"strc1 c99\" id=\"comp-jwpaxkd5\"><div class=\"strc1inlineContent txtNew c98\" id=\"comp-jwpaxkd5inlineContent\"><p class=\"font_8 c79\">MYSTICISM AND MONSTERS<\/p><p class=\"font_8 c97\">\u200b<\/p><p class=\"font_8 c44\">Magic in Ironsworn is subtle, evocative and mysterious. Mystics seek to ward away the darkness through the practice of magic, but often succumb to it.<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">Magic is performed through rituals which help to support your actions or act as unique moves. Like all assets, rituals can be selected as you gain experience and can be upgraded over time to make them more flexible or powerful.<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">These rituals are\u00a0often used to protect Ironlanders from\u00a0 supernatural horrors and monstrous beasts. In the Old World, they were superstition and legend. Here, they are nightmares made real.\u00a0<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">If it suits your version of the Ironlands, rituals and mythic creatures can also be ignored entirely.<\/p><\/div><\/div><div data-content-width=\"492\" class=\"strc1 strc1balata c102\" id=\"comp-jwpaxkez\"><img id=\"comp-jwpaxkezbalatamediaimageimage\" class=\"c101\" alt=\"\" data-type=\"image\" itemprop=\"image\" src=\"https:\/\/static.wixstatic.com\/media\/4db827_5a7f9a3a6e2c4511b46a93868c9ea343~mv2.jpg\/v1\/fill\/w_295,h_419,al_c,q_80,usm_0.66_1.00_0.01,blur_2\/4db827_5a7f9a3a6e2c4511b46a93868c9ea343~mv2.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c111\" id=\"comp-jwpaxkh5\"><div id=\"comp-jwpaxkh5inlineContent\" class=\"strc1inlineContent strc1 c110\"><div data-page-id=\"ij2ei\" data-enable-video=\"true\" data-bg-effect-name=\"BackgroundParallax\" data-media-type=\"Image\" data-use-clip-path=\"\" data-needs-clipping=\"true\" data-wix-video-layout=\"true\" class=\"strc1balata c105\" id=\"comp-jwpaxkhmbalata\"><img id=\"comp-jwpaxkhmbalatamediaimageimage\" class=\"c76\" alt=\"\" data-type=\"image\" itemprop=\"image\" src=\"https:\/\/static.wixstatic.com\/media\/4db827_590168de55f34035b7e26f7fc9418c19~mv2.jpg\/v1\/fill\/w_245,h_128,al_c,q_80,usm_0.66_1.00_0.01,blur_2\/4db827_590168de55f34035b7e26f7fc9418c19~mv2.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxkhminlineContent\"><p><\/p><h2 class=\"font_2 c106\">IRON VOWS<\/h2><div data-packed=\"true\" class=\"txtNew c109\" id=\"comp-jwpaxkiy\"><p class=\"font_8 c108\">In the Ironlands, a vow is sacred. When you declare your solemn promise to serve or aid someone, or to complete a personal quest, your honor is bound to that vow. Abandoning or recanting an oath is the worst sort of failure.\u00a0<\/p><p class=\"font_8 c108\">When you swear a vow, you touch a piece of iron. It can be an iron coin, a weapon, or your armor. It\u2019s an old tradition. Some say the iron, a piece of the primal world, serves as a conduit to the old gods\u2014so they may better hear your promise.<\/p><p class=\"font_8 c108\">\u200b<\/p><p class=\"font_8 c108\">Vows are the core of playing Ironsworn. It is your vows that drive you. These goals create the context for your adventures and challenges. As you fulfill vows, you gain experience and new abilities.<\/p><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c140\" id=\"comp-jwpaxklg\"><div id=\"comp-jwpaxklginlineContent\" class=\"strc1inlineContent strc1 c139\"><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxklxinlineContent\"><p><\/p><h3 class=\"font_3 c42\">THE SYSTEM<\/h3><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c115\" id=\"comp-jwpaxko4\"><p><\/p><h4 class=\"font_4 c56\">MOVES<\/h4><div data-packed=\"true\" class=\"txtNew c114\" id=\"comp-jwpaxkpf\"><p class=\"font_8 c44\">In Ironsworn, moves help you decide what happens when you do something dangerous or uncertain. There is a move for most common actions and situations you will encounter in Ironsworn.<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">Ironsworn moves are inspired by the innovative, fiction-first mechanics of games such as <a href=\"http:\/\/apocalypse-world.com\/\" target=\"_blank\" data-content=\"http:\/\/apocalypse-world.com\/\" data-type=\"external\" rel=\"noopener noreferrer\">Apocalypse World<\/a> and <a href=\"http:\/\/www.dungeon-world.com\/\" target=\"_blank\" data-content=\"http:\/\/www.dungeon-world.com\/\" data-type=\"external\" rel=\"noopener noreferrer\">Dungeon World<\/a>.<\/p><\/div><\/div><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c118\" id=\"comp-jwpaxkrc\"><p><\/p><h4 class=\"font_4 c56\">MOMENTUM<\/h4><div data-packed=\"true\" class=\"txtNew c117\" id=\"comp-jwpaxksn\"><p class=\"font_8 c44\">Momentum is a special track which is core to playing Ironsworn. \u00a0When you have positive momentum, you are building on your successes and ready to make decisive moves.\u00a0When you have negative momentum, you have suffered setbacks and your quest is in jeopardy.\u00a0<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">You can use momentum to gain a decisive result \u2014 or avoid dire failure \u2014 on a critical move.<\/p><p class=\"font_8 c44\">\u200b<\/p><\/div><\/div><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c121\" id=\"comp-jwpaxkvz\"><p><\/p><h4 class=\"font_4 c56\">THE ACTION ROLL<\/h4><div data-packed=\"true\" class=\"txtNew c120\" id=\"comp-jwpaxkxf\"><p class=\"font_8 c44\">When you make a move representing a risky or uncertain action, roll three dice at once:\u00a0\u00a0An action die (D6) and two challenge dice (D10). Add your relevant stat and any bonuses to your action die. This is your action score.\u00a0 Then, compare your\u00a0action score to each of the challenge dice and check for a strong hit, weak hit, or miss. The move will tell you how to interpret the result.\u00a0<\/p><\/div><\/div><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c133\" id=\"comp-jwpaxkzc\"><div data-packed=\"true\" class=\"txtNew c127\" id=\"comp-jwpaxl2x\"><p class=\"font_8 c44\">Strong Hit<br \/>Your action score is greater than both the challenge dice. You succeed at what you are trying to do.<\/p><\/div><div data-packed=\"true\" class=\"txtNew c129\" id=\"comp-jwpaxl58\"><p class=\"font_8 c44\">Weak Hit<br \/>Your action score is greater than only one of the challenge dice. You probably succeeded, but with a lesser effect or cost.<\/p><\/div><div data-packed=\"true\" class=\"txtNew c131\" id=\"comp-jwpaxl7i\"><p class=\"font_8 c44\">Miss<br \/>Your action score isn\u2019t greater than either of the challenge dice. You failed, or you need to make some serious concessions.<\/p><\/div><div data-packed=\"true\" class=\"txtNew c132\" id=\"comp-jwpaxl8s\"><p class=\"font_8 c58\">Did you roll a match on the challenge dice? If so, you\u00a0encounter a surprising complication or opportunity.<\/p><\/div><\/div><div class=\"s_BIwzIGroupSkin s_BIwzIGroupSkininlineContent c136\" id=\"comp-jwpaxlas\"><p><\/p><h4 class=\"font_4 c56\">ORACLES<\/h4><div data-packed=\"true\" class=\"txtNew c135\" id=\"comp-jwpaxlck\"><p class=\"font_8 c44\">Ironsworn Oracles help answer questions in solo or co-op games, or provide inspiration for the GM in guided play.\u00a0<\/p><p class=\"font_8 c44\">\u200b<\/p><p class=\"font_8 c44\">Ironsworn includes a full suite of moves and oracle prompts designed to lead your story in exciting and unexpected directions\u2014with absolutely zero preparation.\u00a0<\/p><\/div><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c152\" id=\"comp-jwpaxljt\"><div id=\"comp-jwpaxljtinlineContent\" class=\"strc1inlineContent strc1 c151\"><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxlkhinlineContent\"><p><\/p><h2 class=\"font_2 c106\">GET IRONSWORN NOW<\/h2><p><\/p><h4 class=\"font_4 c56\">It's FREE IN PDF<\/h4><div data-packed=\"true\" class=\"txtNew c145\" id=\"comp-jwpaxln7\"><p class=\"font_8 c144\">The free download of the Ironsworn RPG includes the 270-page rulebook, character sheet, printable asset cards, and reference handouts. It's a complete game, ready to play. Want a physical copy? You can also <a href=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" target=\"_blank\" data-content=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" data-type=\"external\" rel=\"noopener noreferrer\">purchase the rulebook in hardcover or softcover<\/a>.<\/p><\/div><p><a href=\"https:\/\/www.ironswornrpg.com\/downloads\" target=\"_self\" id=\"comp-jxwiibdjlink\" class=\"g-transparent-a style-jwpaxn0olink\">FREE DOWNLOAD<\/a><\/p><p><a href=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" target=\"_blank\" data-content=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830&amp;src=ironweb\" data-type=\"external\" rel=\"noopener noreferrer\" id=\"comp-jxwiibf5link\" class=\"g-transparent-a style-jwpaxn0olink\">BUY IN PRINT<\/a><\/p><div title=\"\" data-is-responsive=\"false\" data-display-mode=\"fill\" data-content-padding-horizontal=\"0\" data-content-padding-vertical=\"0\" data-exact-height=\"357\" class=\"ca1 c150\" id=\"comp-jwpaxluz\"><a href=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830\" target=\"_self\" data-content=\"http:\/\/www.drivethrurpg.com\/product\/238369\/Ironsworn?affiliate_id=1587830\" data-type=\"external\" rel=\"\" id=\"comp-jwpaxluzlink\" class=\"ca1link c149\"><img id=\"comp-jwpaxluzimgimage\" class=\"c148\" alt=\"Get Ironsworn in Digital and Print\" data-type=\"image\" itemprop=\"image\" src=\"https:\/\/static.wixstatic.com\/media\/4db827_0aa4918d6dc04d1d87ebca0ed751cdce~mv2.png\/v1\/crop\/x_1,y_0,w_798,h_475\/fill\/w_180,h_107,al_c,usm_0.66_1.00_0.01,blur_2\/4db827_0aa4918d6dc04d1d87ebca0ed751cdce~mv2.png\" referrerpolicy=\"no-referrer\" \/><\/a><\/div><\/div><\/div><\/section><section data-responsive=\"true\" data-is-screen-width=\"true\" data-col-margin=\"0\" data-row-margin=\"0\" class=\"strc1 c165\" id=\"comp-jwpaxmp4\"><div id=\"comp-jwpaxmp4inlineContent\" class=\"strc1inlineContent strc1 c164\"><div class=\"strc1inlineContent c32\" id=\"comp-jwpaxmpvinlineContent\"><p><\/p><h2 class=\"font_2 c106\">STAY INFORMED<\/h2><div data-packed=\"true\" class=\"txtNew c157\" id=\"comp-jwpaxmsh\"><p class=\"font_8 c156\">Subscribe FOR UPDATES<\/p><\/div><div data-packed=\"true\" class=\"txtNew c158\" id=\"comp-jwpaxmup\"><p class=\"font_8 c58\">By submitting this form you agree to receive\u00a0occasional\u00a0emails (about once or twice a month) related to the Ironsworn roleplaying game. Your information will not be shared or used for any other purpose. You can unsubscribe anytime.<\/p><\/div><\/div><\/div><\/section>","created_at":"2020-01-13T21:19:32+0000","updated_at":"2020-01-13T21:19:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.ironswornrpg.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1110,"title":"hlissner\/doom-emacs","url":"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/faq.org#how-does-doom-start-up-so-quickly","content":"<ul><li><a href=\"#general\">General<\/a>\n<ul><li><a href=\"#why-is-it-called-doom\">Why is it called Doom?<\/a><\/li>\n<li><a href=\"#does-doom-work-on-windows\">Does Doom work on Windows?<\/a><\/li>\n<li><a href=\"#is-doom-only-for-vimmers\">Is Doom only for vimmers?<\/a><\/li>\n<li><a href=\"#i-am-a-beginner-can-i-use-doom\">I am a beginner. Can I use Doom?<\/a><\/li>\n<li><a href=\"#how-does-doom-compare-to-spacemacs\">How does Doom compare to Spacemacs?<\/a><\/li>\n<li><a href=\"#why-such-a-complicated-package-management-system\">Why such a complicated package management system?<\/a><\/li>\n<li><a href=\"#how-does-doom-start-up-so-quickly\">How does Doom start up so quickly?<\/a><\/li>\n<li><a href=\"#why-is-startup-time-important-why-not-use-the-daemon\">Why is startup time important? Why not use the daemon?<\/a><\/li>\n<li><a href=\"#how-do-i-use-doom-alongside-other-emacs-configs\">How do I use Doom alongside other Emacs configs?<\/a><\/li>\n<li><a href=\"#why-should-i-use-doom-instead-of-rolling-my-own-config\">Why should I use Doom instead of rolling my own config?<\/a><\/li>\n<li><a href=\"#what-is-the-meaning-behind-dooms-naming-conventions\">What is the meaning behind Doom\u2019s naming conventions?<\/a><\/li>\n<li><a href=\"#how-can-i-contribute-tosupport-doom\">How can I contribute to\/support Doom?<\/a><\/li>\n<li><a href=\"#what-version-of-doom-am-i-running\">What version of Doom am I running?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#configuration\">Configuration<\/a>\n<ul><li><a href=\"#should-i-fork-doom-to-customize-it\">Should I fork Doom to customize it?<\/a><\/li>\n<li><a href=\"#how-do-i-configure-doom-emacs\">How do I configure Doom Emacs?<\/a><\/li>\n<li><a href=\"#how-do-i-enable-or-disable-a-doom-module\">How do I enable or disable a Doom module?<\/a><\/li>\n<li><a href=\"#how-do-i-install-a-package-from-elpa\">How do I install a package from ELPA?<\/a><\/li>\n<li><a href=\"#how-do-i-install-a-package-from-githubanother-source\">How do I install a package from github\/another source?<\/a><\/li>\n<li><a href=\"#how-do-i-change-where-an-existing-package-is-installed-from\">How do I change where an existing package is installed from?<\/a><\/li>\n<li><a href=\"#how-do-i-disable-a-package-completely\">How do I disable a package completely?<\/a><\/li>\n<li><a href=\"#how-do-i-reconfigure-a-package-included-in-doom\">How do I reconfigure a package included in Doom?<\/a><\/li>\n<li><a href=\"#how-do-i-change-the-theme\">How do I change the theme?<\/a><\/li>\n<li><a href=\"#how-do-i-change-the-fonts\">How do I change the fonts?<\/a><\/li>\n<li><a href=\"#how-do-i-bind-my-own-keys-or-change-existing-ones\">How do I bind my own keys (or change existing ones)?<\/a><\/li>\n<li><a href=\"#how-do-i-get-motions-to-treat-underscores-as-word-delimiters\">How do I get motions to treat underscores as word delimiters?<\/a><\/li>\n<li><a href=\"#how-do-i-change-the-leaderlocalleader-keys\">How do I change the leader\/localleader keys?<\/a><\/li>\n<li><a href=\"#how-do-i-change-the-style-of-line-numbers-or-disable-them-altogether\">How do I change the style of line-numbers (or disable them altogether)?<\/a><\/li>\n<li><a href=\"#how-do-i-change-the-behavior-and-appearance-of-popup-windows\">How do I change the behavior and appearance of popup windows?<\/a><\/li>\n<li><a href=\"#how-do-i-change-the-appearance-a-face-or-faces\">How do I change the appearance a face (or faces)?<\/a><\/li>\n<li><a href=\"#can-doom-be-customized-without-restarting-emacs\">Can Doom be customized without restarting Emacs?<\/a><\/li>\n<li><a href=\"#can-vimevil-be-removed-for-a-more-vanilla-emacs-experience\">Can Vim\/Evil be removed for a more vanilla Emacs experience?<\/a><\/li>\n<li><a href=\"#should-i-use-make-or-bindoom\">Should I use <code>make<\/code> or <code>bin\/doom<\/code>?<\/a><\/li>\n<li><a href=\"#when-should-and-shouldnt-i-use-bindoom\">When should and shouldn\u2019t I use <code>bin\/doom<\/code>?<\/a><\/li>\n<li><a href=\"#when-to-run-doom-refresh\">When to run ~doom refresh~<\/a><\/li>\n<li><a href=\"#how-to-suppress-confirmation-prompts-while-bindoom-is-running\">How to suppress confirmation prompts while <code>bin\/doom<\/code> is running<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#defaults\">Defaults<\/a>\n<ul><li><a href=\"#why-ivy-over-helm\">Why Ivy over Helm?<\/a><\/li>\n<li><a href=\"#why-are-there-no-default-keybinds-for-smartparens-for-evil-users\">Why are there no default keybinds for Smartparens (for evil users)?<\/a><\/li>\n<li><a href=\"#why-do-non-evil-users-get-expand-region-but-not-evil-users\">Why do non-evil users get expand-region, but not evil users?<\/a><\/li>\n<li><a href=\"#why-not-use-exec-path-from-shell-instead-of-doom-env\">Why not use exec-path-from-shell instead of <code>doom env<\/code>?<\/a><\/li>\n<li><a href=\"#why-wsbutler-over-delete-trailing-whitespace-or-whitespace-cleanup\">Why wsbutler over delete-trailing-whitespace or whitespace-cleanup?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#common-issues\">Common Issues<\/a>\n<ul><li><a href=\"#i-get-the-vanilla-emacs-splash-screen-at-startup\">I get the vanilla Emacs splash screen at startup<\/a><\/li>\n<li><a href=\"#i-see-a-blank-scratch-buffer-at-startup\">I see a blank scratch buffer at startup<\/a><\/li>\n<li><a href=\"#strange-or-incorrect-icons-are-displayed-everywhere\">Strange (or incorrect) icons are displayed everywhere<\/a><\/li>\n<li><a href=\"#void-variable-and-void-function-errors-on-startup\">~void-variable~ and <code>void-function<\/code> errors on startup<\/a><\/li>\n<li><a href=\"#doom-cant-find-my-executablesdoesnt-inherit-the-correct-path\">Doom can\u2019t find my executables\/doesn\u2019t inherit the correct ~PATH~<\/a><\/li>\n<li><a href=\"#theres-artefacting-on-my-icon-fonts-in-gui-emacs-956\">There\u2019s artefacting on my icon fonts in GUI Emacs (#956)<\/a><\/li>\n<li><a href=\"#the-s-and-s-keys-dont-behave-like-they-do-in-vimevil-1307\">The <code>s<\/code> and <code>S<\/code> keys don\u2019t behave like they do in vim\/evil (#1307)<\/a><\/li>\n<li><a href=\"#changes-to-my-config-arent-taking-effect\">Changes to my config aren\u2019t taking effect<\/a><\/li>\n<li><a href=\"#the-frame-goes-black-on-macos-while-in-full-screen-mode\">The frame goes black on MacOS, while in full-screen mode<\/a><\/li>\n<li><a href=\"#doom-crashes-when\">Doom crashes when\u2026<\/a><\/li>\n<li><a href=\"#cant-load-my-theme-unable-to-find-theme-file-for-x-errors\">Can\u2019t load my theme; <code>unable to find theme file for X<\/code> errors<\/a><\/li>\n<li><a href=\"#tramp-connections-hang-forever-when-connecting\">TRAMP connections hang forever when connecting<\/a><\/li>\n<li><a href=\"#an-upstream-package-was-broken-and-i-cant-update-it\">An upstream package was broken and I can\u2019t update it<\/a><\/li>\n<li><a href=\"#why-do-i-see-ugly-indentation-highlights-for-tabs\">Why do I see ugly indentation highlights for tabs?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#contributing\">Contributing<\/a><\/li>\n<\/ul><h2><a id=\"user-content-why-is-it-called-doom\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-is-it-called-doom\"><\/a>Why is it called Doom?<\/h2><p>An homage to idsoftware\u2019s classic game, whose open sourced code was my first exposure to programming.<\/p><p>Also, if you\u2019re obsessed enough with a text editor that you write a community config for it, you\u2019re doomed from the get go.<\/p><h2><a id=\"user-content-does-doom-work-on-windows\" class=\"anchor\" aria-hidden=\"true\" href=\"#does-doom-work-on-windows\"><\/a>Does Doom work on Windows?<\/h2><p>Windows support is weak and will generally lag behind Linux\/MacOS support, so your mileage will vary. However, many have reported success installing Doom Emacs on Windows (using WSL, WSL2 or scoop\/chocolatey). You\u2019ll find install instructions for Windows in the Getting Starting guide.<\/p><p>If you\u2019re a Windows user, help us improve our documentation on Windows support!<\/p><h2><a id=\"user-content-is-doom-only-for-vimmers\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-doom-only-for-vimmers\"><\/a>Is Doom only for vimmers?<\/h2><p>Henrik is a dyed-in-the-wool vimmer with more than a decade of vim muscle memory. Vim\u2019s is the only paradigm he truly knows, so vimmers will always be his primary audience.<\/p><p>That\u2019s not to say Doom won\u2019t work without evil, only that it is less polished in that respect. Our growing non-evil userbase are slowly improving the situation however. We welcome suggestions and PRs to help accommodate a non-evil workflow.<\/p><p>If you\u2019d still like a go at it, see the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/modules\/editor\/evil\/README.org\">Removing evil-mode<\/a> section in the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/modules\/editor\/evil\/README.org\">:editor evil<\/a> module\u2019s documentation.<\/p><h2><a id=\"user-content-i-am-a-beginner-can-i-use-doom\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-am-a-beginner-can-i-use-doom\"><\/a>I am a beginner. Can I use Doom?<\/h2><p>If you\u2019re new to the terminal, to programming, or Emacs and\/or vim, Doom (or Emacs, for that matter) is a rough place to start. Neither Doom nor Emacs are particularly beginner friendly. That\u2019s not to say it\u2019s impossible, or that we won\u2019t help you if you ask, but expect a hefty commitment and a bumpy journey.<\/p><p>Remember to check out the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/index.org\">Documentation<\/a> for a guide to getting started.<\/p><h2><a id=\"user-content-how-does-doom-compare-to-spacemacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-does-doom-compare-to-spacemacs\"><\/a>How does Doom compare to Spacemacs?<\/h2><p>To paraphrase (and expand upon) a <a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/6pa0oq\/quickstart_tutorial_for_emacs_newbies_with_doom\/dkp1bhd\/\" rel=\"nofollow\">reddit answer<\/a> to this question by <a href=\"https:\/\/github.com\/gilbertw1\">@gilbertw1<\/a>:<\/p><ul><li><strong>Doom is lighter than Spacemacs.<\/strong> Doom starts up faster and is better optimized, but Spacemacs has more features.<\/li>\n<li><strong>Doom is thinner than Spacemacs.<\/strong> There are fewer abstractions between you and vanilla Emacs, and what abstractions do exist are thin by design. This means there\u2019s less to understand and it\u2019s easier to hack.<\/li>\n<li><strong>Doom is much more opinionated than Spacemacs.<\/strong> Doom does not strive to be a one-size-fits-all, beginner-friendly solution, nor is it configured by consensus. It is [mostly] the work of one developer and caters to his vim-slanted tastes. Doom\u2019s defaults enforce very particular (albeit optional) workflows.<\/li>\n<li><strong>Doom lacks manpower.<\/strong> Bugs stick around longer, documentation is light and development is at the mercy of it\u2019s single maintainer\u2019s schedule, health and whims.<\/li>\n<li><strong>Doom is not beginner friendly.<\/strong> Spacemacs works out of the box. Your mileage may vary with Doom; assembly is required! Familiarity with Emacs Lisp (or programming in general), git and the command line will go a long way to ease you into Doom.<\/li>\n<li><strong>Doom manages its packages outside of Emacs.<\/strong> Spacemacs installs (and checks for packages) on startup or on demand. Doom leaves package management to be done externally, through the <code>bin\/doom<\/code> script. This allows for package management to be scripted on the command line and enables a number of startup optimizations we wouldn\u2019t have otherwise.<\/li>\n<\/ul><h2><a id=\"user-content-why-such-a-complicated-package-management-system\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-such-a-complicated-package-management-system\"><\/a>Why such a complicated package management system?<\/h2><p>Doom had <del>four<\/del> <strong>five<\/strong> goals for its package management system:<\/p><ol><li><strong>Scriptability:<\/strong> package management should be shell-scriptable, so updating can be automated.<\/li>\n<li><strong>Reach:<\/strong> allow users to install packages from sources other than ELPA (like github or gitlab), and from specific commits, branches or tags. Some plugins are out-of-date through official channels, have changed hands, have a superior fork, or aren\u2019t available in ELPA repos.<\/li>\n<li><strong>Performance:<\/strong> lazy-loading the package management system is a tremendous boon to start up speed. Initializing package.el and quelpa (and\/or checking that your packages are installed) every time you start up is expensive.<\/li>\n<li><strong>Organization:<\/strong> an Emacs configuration grows so quickly, in complexity and size. A clear separation of concerns (configuration of packages from their installation) is more organized.<\/li>\n<li><strong>Reproducibility:<\/strong> <em>This goal hasn\u2019t been implemented yet<\/em>, but all our work up until now is aimed at this goal. Emacs is a tumultuous ecosystem; packages break left and right, and we rely on hundreds of them. Eventually, we want package versions to be locked to versions of Doom so that Doom installs are reproducible.<\/li>\n<\/ol><h2><a id=\"user-content-how-does-doom-start-up-so-quickly\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-does-doom-start-up-so-quickly\"><\/a>How does Doom start up so quickly?<\/h2><p>Doom employs a number of techniques to cut down startup time. Here are its most effective techniques:<\/p><h3><a id=\"user-content-avoid-garbage-collection-at-startup\" class=\"anchor\" aria-hidden=\"true\" href=\"#avoid-garbage-collection-at-startup\"><\/a>Avoid garbage collection at startup<\/h3><p>The GC can easily double startup time, so we suppress it at startup by turning up <code>gc-cons-threshold<\/code> (and perhaps <code>gc-cons-percentage<\/code>) temporarily:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq gc-cons-threshold most-positive-fixnum ; 2^61 bytes\n      gc-cons-percentage 0.6)\n;; ... your emacs config here ...<\/pre><\/div><p>However, it is important to reset it eventually (as late as possible). Not doing so will cause garbage collection freezes during long-term interactive use. Conversely, a <code>gc-cons-threshold<\/code> that is too small will cause stuttering. We use 16mb as our default.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'emacs-startup-hook\n  (lambda ()\n    (setq gc-cons-threshold 16777216 ; 16mb\n          gc-cons-percentage 0.1)))<\/pre><\/div><p>It may also be wise to raise <code>gc-cons-threshold<\/code> while the minibuffer is active, so the GC doesn\u2019t slow down expensive commands (or completion frameworks, like helm and ivy). Here is how Doom does it:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun doom-defer-garbage-collection-h ()\n  (setq gc-cons-threshold most-positive-fixnum))\n(defun doom-restore-garbage-collection-h ()\n  ;; Defer it so that commands launched immediately after will enjoy the\n  ;; benefits.\n  (run-at-time\n   1 nil (lambda () (setq gc-cons-threshold doom-gc-cons-threshold))))\n(add-hook 'minibuffer-setup-hook #'doom-defer-garbage-collection-h)\n(add-hook 'minibuffer-exit-hook #'doom-restore-garbage-collection-h)<\/pre><\/div><h3><a id=\"user-content-unset-file-name-handler-alist-temporarily\" class=\"anchor\" aria-hidden=\"true\" href=\"#unset-file-name-handler-alist-temporarily\"><\/a>Unset <code>file-name-handler-alist<\/code> temporarily<\/h3><p>Emacs consults this variable every time a file is read or library loaded, or when certain functions in the file API are used (like <code>expand-file-name<\/code> or <code>file-truename<\/code>).<\/p><p>They do so to check if a special handler is needed to read it, but none of these handlers are necessary for the initialization work we do at startup, so it is generally safe to disable it (temporarily!):<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar doom--file-name-handler-alist file-name-handler-alist)\n(setq file-name-handler-alist nil)\n;; ... your whole emacs config here ...\n;; Then restore it later:\n(setq file-name-handler-alist doom--file-name-handler-alist)\n;; Alternatively, restore it even later:\n(add-hook 'emacs-startup-hook\n  (lambda ()\n    (setq file-name-handler-alist doom--file-name-handler-alist)))<\/pre><\/div><p>It is important to restore this variable, otherwise you won\u2019t be able to use TRAMP and Emacs will be unable to read compressed\/encrypted files.<\/p><h3><a id=\"user-content-cut-down-on-load-path-lookups\" class=\"anchor\" aria-hidden=\"true\" href=\"#cut-down-on-load-path-lookups\"><\/a>Cut down on <code>load-path<\/code> lookups<\/h3><p>Each <code>load<\/code> and <code>require<\/code> call (without an second argument) costs an O(n) lookup on <code>load-path<\/code>. The average Doom config has approximately 260 packages including dependencies, and around 40 built-in packages. That means a minimum of 300 entries in <code>load-path<\/code> with a worst case of <code>n=300<\/code> for <em>each<\/em> package load (but realistically, <code>n<\/code> will be somewhere between <code>2<\/code> and <code>20<\/code>).<\/p><p>The cost isn\u2019t great, but it does add up. There isn\u2019t much to do about this, except be mindful of it where we can:<\/p><ul><li>Paths in Doom\u2019s autoloads file are replaced with absolute ones, thus incurring no lookup cost to lazy load them.<\/li>\n<li>The <code>load!<\/code> macro is used instead of <code>require<\/code> where possible. This builds paths with string concatenation (which is baked in at compile time, removing most of the associated cost).<\/li>\n<li><code>load-path<\/code> is let-bound to a subset of itself where possible (the <code>doom--initial-load-path<\/code> variable contains the value of <code>load-path<\/code> before it was touched by Doom).<\/li>\n<\/ul><h3><a id=\"user-content-concatenate-package-autoloads\" class=\"anchor\" aria-hidden=\"true\" href=\"#concatenate-package-autoloads\"><\/a>Concatenate package autoloads<\/h3><p>When you install a package, a PACKAGE-autoloads.el file is generated. This file contains a map of autoloaded functions and snippets declared by the package (that\u2019s what those <code>;;;###autoload<\/code> comments are for in packages). They tell Emacs where to find them, when they are eventually called. In your conventional Emacs config, every single one of these autoloads files are loaded immediately at startup.<\/p><p>Since you\u2019ll commonly have hundreds of packages, loading hundreds of autoloads file can hurt startup times. We get around this by concatenating these autoloads files into one giant one (in <code>~\/.emacs.d\/.local\/autoloads.pkg.el<\/code>) when you run <code>doom refresh<\/code>.<\/p><p>Emacs 27+ will introduce a <code>package-quickstart<\/code> feature that will do this for you \u2013 the <code>straight<\/code> package manager does this for you too \u2013 but Doom Emacs has its own specialized mechanism for doing this, and has tacked a number of Doom-specific optimizations on top of it.<\/p><h3><a id=\"user-content-lazy-load-package-management-systems\" class=\"anchor\" aria-hidden=\"true\" href=\"#lazy-load-package-management-systems\"><\/a>Lazy load package management system(s)<\/h3><p>Initializing package.el or straight.el at startup is expensive. We can save some time by delaying that initialization until we actually need these libraries (and only eagerly load them when we\u2019re doing package management, e.g. when we run <code>doom refresh<\/code>).<\/p><p>Among other things, <code>doom refresh<\/code> does a lot for us. It generates concatenated autoloads files; caches expensive variables like caches <code>load-path<\/code>, <code>Info-directory-list<\/code> and <code>auto-mode-alist<\/code>; and preforms all your package management activities there \u2013 far away from your interactive sessions.<\/p><p>How exactly Doom accomplishes all this is a little complex, so instead, here is a boiled-down version you can use in your own configs (for package.el, not straight.el):<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar cache-file \"~\/.emacs.d\/cache\/autoloads\")\n(defun initialize ()\n  (unless (load cache-file t t)\n    (setq package-activated-list nil)\n    (package-initialize)\n    (with-temp-buffer\n      (cl-pushnew doom-core-dir load-path :test #'string=)\n      (dolist (spec package-alist)\n        (when-let (desc (cdr spec))\n          (let ((file (concat (package--autoloads-file-name desc) \".el\")))\n            (when (file-readable-p file)\n              ;; Ensure that the contents of this autoloads file believes they\n              ;; haven't been moved:\n              (insert \"(let ((load-file-name \" (prin1-to-string (abbreviate-file-name file)) \"))\\n\")\n              (insert-file-contents file)\n              (save-excursion\n                ;; Delete forms that modify `load-path' and `auto-mode-alist', we\n                ;; will set them once, later.\n                (while (re-search-forward \"^\\\\s-*\\\\((\\\\(?:add-to-list\\\\|\\\\(?:when\\\\|if\\\\) (boundp\\\\)\\\\s-+'\\\\(?:load-path\\\\|auto-mode-alist\\\\)\\\\)\" nil t)\n                  (goto-char (match-beginning 1))\n                  (kill-sexp)))\n              ;; Remove unnecessary comment lines and (provide ...) forms\n              (while (re-search-forward \"^\\\\(?:;;\\\\(.*\\n\\\\)\\\\|\\n\\\\|(provide '[^\\n]+\\\\)\" nil t)\n                (unless (nth 8 (syntax-ppss))\n                  (replace-match \"\" t t)))\n              (unless (bolp) (insert \"\\n\"))\n              (insert \")\\n\")))))\n      (prin1 `(setq load-path ',load-path\n                    auto-mode-alist ',auto-mode-alist\n                    Info-directory-list ',Info-directory-list)\n             (current-buffer))\n      (write-file (concat cache-file \".el\"))\n      (byte-compile-file cache-file))))\n(initialize)<\/pre><\/div><p>You\u2019ll need to delete <code>cache-files<\/code> any time you install, remove, or update a new package, however. In that case you could advise <code>package-install<\/code> and <code>package-delete<\/code> to call <code>initialize<\/code> when they succeed. Or, you could make <code>initialize<\/code> interactive and call it manually when you determine it\u2019s necessary. Up to you!<\/p><p>Note: package.el is sneaky, and will initialize itself if you\u2019re not careful. <strong>Not on my watch, criminal scum!<\/strong><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; in ~\/.emacs.d\/init.el (or ~\/.emacs.d\/early-init.el in Emacs 27)\n(setq package-enable-at-startup nil ; don't auto-initialize!\n      ;; don't add that `custom-set-variables' block to my init.el!\n      package--init-file-ensured t)<\/pre><\/div><h3><a id=\"user-content-lazy-load-more-than-everything\" class=\"anchor\" aria-hidden=\"true\" href=\"#lazy-load-more-than-everything\"><\/a>Lazy load more than everything<\/h3><p><code>use-package<\/code> can defer your packages. Using it is a no-brainer, but Doom goes a little further with lazy loading. There are some massive plugins out there. For many of them, ordinary lazy loading techniques simply don\u2019t work. To name a few:<\/p><ul><li>The <code>lang\/org<\/code> module defers loading babel packages until their src blocks are executed. You no longer need <code>org-babel-do-load-languages<\/code> in your config.<\/li>\n<li>Company and yasnippet are loaded as late as possible (waiting until the user opens a non-read-only, file-visiting buffer (that isn\u2019t in fundamental-mode)).<\/li>\n<li>The <code>evil-easymotion<\/code> package has many keybinds. You\u2019d need to load the package for them to all take effect, so instead, <code>gs<\/code> is bound to a command that loads the package and then invisibly populates <code>gs<\/code>, then simulates the <code>gs<\/code> keypress as though those new keys had always been there.<\/li>\n<li>A number of packages are \u201cincrementally\u201d loaded. This is a Doom feature where, after a few seconds of idle time post-startup, Doom will load packages piecemeal while Emacs. It will quickly abort if it detects input, as to make the process as subtle as possible.\n<p>For example, instead of loading <code>org<\/code> (a giant package), it will load these dependencies, one at a time, before finally loading <code>org<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(calendar find-func format-spec org-macs org-compat org-faces org-entities\n org-list org-pcomplete org-src org-footnote org-macro ob org org-agenda\n org-capture)\n    <\/pre><\/div>\n<p>This ensures packages load as quickly as possible when you first load an org file.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-exploit-byte-compilation\" class=\"anchor\" aria-hidden=\"true\" href=\"#exploit-byte-compilation\"><\/a><del>Exploit byte-compilation!<\/del><\/h3><p>It used to be that byte-compilation bought a 40-60% improvement in startup times, because expensive operations (like <code>package-initialize<\/code> or <code>exec-path-from-shell<\/code>) were evaluated at compile time, but Doom has changed.<\/p><p>I\u2019ve since adopted a pre-cache approach (when running <code>doom refresh<\/code>), which brings these startup benefits to uncompiled Emacs. This renders byte-compilation significantly less beneficial for startup time.<\/p><p>That said, compilation will still benefit Doom\u2019s snappiness in general.<\/p><p>Run <code>doom compile :core<\/code> to only compile Doom\u2019s core files, or <code>doom compile<\/code> to compile the <em>entire<\/em> config (<code>~\/.emacs.d<\/code> and <code>~\/.doom.d<\/code>) \u2013 which may take a while.<\/p><h3><a id=\"user-content-use-lexical-binding-everywhere\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-lexical-binding-everywhere\"><\/a>Use <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Lexical-Binding.html\" rel=\"nofollow\">lexical-binding<\/a> everywhere<\/h3><p>Add <code>;; -*- lexical-binding: t; -*-<\/code> to the top of your elisp files. This can break code if you\u2019ve written it to depend on undeclared dynamic variables, but I\u2019ve designed Doom not to.<\/p><p>This buys a small improvement in performance, but every little bit helps. You\u2019ll find more about it in:<\/p><ul><li><a href=\"http:\/\/nullprogram.com\/blog\/2017\/01\/30\/\" rel=\"nofollow\">\u201cHow to Write Fast(er) Emacs Lisp.\u201d<\/a><\/li>\n<li><a href=\"http:\/\/nullprogram.com\/blog\/2016\/12\/22\/\" rel=\"nofollow\">\u201cSome Performance Advantages of Lexical Scope.\u201d<\/a><\/li>\n<\/ul><h2><a id=\"user-content-why-is-startup-time-important-why-not-use-the-daemon\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-is-startup-time-important-why-not-use-the-daemon\"><\/a>Why is startup time important? Why not use the daemon?<\/h2><p>One of my motivations for a config that starts up fast (aside from the learning experience) was to shape Emacs into a viable alternative to vim for one-shot editing in the terminal (without <code>-Q<\/code>). This also facilitates:<\/p><ul><li>Running multiple, independent instances of Emacs (e.g. on a per-project basis, or for nix-shell users, or to isolate one instance for IRC from an instance for writing code, etc).<\/li>\n<li>Quicker restarting of Emacs, to reload package settings or recover from disastrous errors which can leave Emacs in a broken state.<\/li>\n<li>Faster integration with \u201cedit in Emacs\u201d solutions (like <a href=\"https:\/\/github.com\/alpha22jp\/atomic-chrome\">atomic-chrome<\/a>), and the potential to use them without a running daemon.<\/li>\n<\/ul><p>What\u2019s more, I don\u2019t like using more tools than I need. We should not need a second program just to make the first run comfortably.<\/p><h2><a id=\"user-content-how-do-i-use-doom-alongside-other-emacs-configs\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-use-doom-alongside-other-emacs-configs\"><\/a>How do I use Doom alongside other Emacs configs?<\/h2><p>I recommend <a href=\"https:\/\/github.com\/plexus\/chemacs\">Chemacs<\/a>. You can think of it as a bootloader for Emacs. You\u2019ll find instructions on how to use it with Doom in the user manual.<\/p><p>If you only want to try it out without affecting your current config, it is safe to install Doom anywhere you like. The <code>bin\/doom<\/code> utility will only address the config the script is located under.<\/p><p>You\u2019ll still need a separate folder for personal configuration (<code>~\/.doom.d<\/code> or <code>~\/.config\/doom<\/code> by default), but the <code>-p PATH<\/code> flag (or <code>DOOMDIR<\/code> environment variable) will allow you to use a different location:<\/p><div class=\"highlight highlight-source-shell\"><pre># First install Doom somewhere\ngit clone https:\/\/github.com\/hlissner\/doom-emacs ~\/fakehome\/doom-emacs\n# Then create a place to store our private doom configs. The bin\/doom script\n# recognizes the DOOMDIR environment variable.\nexport DOOMDIR=~\/fakehome\/doom-emacs-config\nmkdir -p \"$DOOMDIR\"\n# Set up Doom for the first time; this may take a while\ncd ~\/fakehome\/doom-emacs\nbin\/doom install\n# then launch Doom Emacs from this folder with:\nbin\/doom run<\/pre><\/div><blockquote>\n<p>Warning: the way <code>bin\/doom run<\/code> starts Doom bypasses many of its startup optimizations. Treat it as a convenience for testing Doom, rather than a permanent entry point.<\/p>\n<\/blockquote><h2><a id=\"user-content-why-should-i-use-doom-instead-of-rolling-my-own-config\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-should-i-use-doom-instead-of-rolling-my-own-config\"><\/a>Why should I use Doom instead of rolling my own config?<\/h2><p>If you care about personalizing the software you use on a daily basis, even half as much as I do, then you probably need professional help, but you also know it is time consuming. Emacs out-of-the-box is a barren wasteland with archaic defaults. Building anything out here and getting a feel for it will take <em>a lot<\/em> of time. Time that I\u2019ve already wasted and can never get back.<\/p><p>Time you could otherwise spend attending your daughter\u2019s dance recitals, that baseball game your son\u2019s team almost won last Thursday, or answering the court summons to fight for custody of your kids.<\/p><p>Also, Doom\u2019s fast yo.<\/p><h2><a id=\"user-content-what-is-the-meaning-behind-dooms-naming-conventions\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-is-the-meaning-behind-dooms-naming-conventions\"><\/a>What is the meaning behind Doom\u2019s naming conventions?<\/h2><p>You\u2019ll find <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/contributing.org\">an overview of Doom\u2019s code conventions<\/a> in the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/contributing.org\">contributing guide<\/a>.<\/p><h2><a id=\"user-content-how-can-i-contribute-tosupport-doom\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-can-i-contribute-tosupport-doom\"><\/a>How can I contribute to\/support Doom?<\/h2><p>Take a look at the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/contributing.org\">Contributing guide<\/a>.<\/p><h2><a id=\"user-content-what-version-of-doom-am-i-running\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-version-of-doom-am-i-running\"><\/a>What version of Doom am I running?<\/h2><p>You\u2019ll find the current version displayed in the modeline on the dashboard. It can also be retrieved using <code>M-x doom\/version<\/code> (bound to <code>SPC h d v<\/code> by default) or <code>doom info<\/code> on the command line.<\/p><h2><a id=\"user-content-should-i-fork-doom-to-customize-it\" class=\"anchor\" aria-hidden=\"true\" href=\"#should-i-fork-doom-to-customize-it\"><\/a>Should I fork Doom to customize it?<\/h2><p>No. Not unless you have a good reason for doing so (and you\u2019re comfortable with the git-rebase workflow). Your customization can be relegated to <code>~\/.doom.d\/<\/code> (or <code>~\/.config\/doom\/<\/code>) entirely.<\/p><p>If you <em>must<\/em> modify Doom proper to get something done, it\u2019s a code smell.<\/p><p>Visit the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Customize section<\/a> of <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">the Getting Started guide<\/a> for details on how to do this.<\/p><h2><a id=\"user-content-how-do-i-configure-doom-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-configure-doom-emacs\"><\/a>How do I configure Doom Emacs?<\/h2><p>Canonically, your private config is kept in <code>~\/.doom.d\/<\/code> or <code>~\/.config\/doom\/<\/code>. Doom will prioritize <code>~\/.config\/doom<\/code>, if it exists. This directory is referred to as your <code>$DOOMDIR<\/code>.<\/p><p>Your private config is typically comprised of an <code>init.el<\/code>, <code>config.el<\/code> and <code>packages.el<\/code> file. Put all your config in <code>config.el<\/code>, install packages by adding <code>package!<\/code> declarations to <code>packages.el<\/code>, and enable\/disable modules in you <code>doom!<\/code> block, which should have been created in your <code>init.el<\/code> when you first ran <code>doom install<\/code>.<\/p><p>Check out the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Customize section<\/a> in the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Getting Started<\/a> guide for details.<\/p><h2><a id=\"user-content-how-do-i-enable-or-disable-a-doom-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-enable-or-disable-a-doom-module\"><\/a>How do I enable or disable a Doom module?<\/h2><p>You\u2019ll find your <code>doom!<\/code> block in <code>~\/.doom.d\/init.el<\/code>. This block contains a list of modules you want enabled and what order to load them in. Disable modules by commenting them out with semicolons. To enable them, remove those leading semicolons:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(doom! :lang\n       python     ; this is enabled\n       ;;ruby     ; this is disabled\n       rust)<\/pre><\/div><p>Remember to run <code>bin\/doom refresh<\/code> afterwards, on the command line, to sync your module list with Doom.<\/p><p>You can find a comprehensive list of modules in the Module Index.<\/p><h2><a id=\"user-content-how-do-i-install-a-package-from-elpa\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-install-a-package-from-elpa\"><\/a>How do I install a package from ELPA?<\/h2><p>Add a <code>package!<\/code> declaration to <code>~\/.doom.d\/packages.el<\/code> for each package you want installed.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(package! winum)<\/pre><\/div><p>Remember to run <code>doom refresh<\/code> afterwards to ensure the package is installed.<\/p><p>You\u2019ll find more information in the \u201d<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Installing packages<\/a>\u201d section of the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Getting Started<\/a> guide.<\/p><h2><a id=\"user-content-how-do-i-install-a-package-from-githubanother-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-install-a-package-from-githubanother-source\"><\/a>How do I install a package from github\/another source?<\/h2><p>The <code>package!<\/code> macro can be passed a MELPA style recipe, allowing you to install packages from just about anywhere:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(package! evil :recipe (:host github :repo \"hlissner\/my-evil-fork\"))<\/pre><\/div><p>Remember to run <code>doom refresh<\/code> every time you modify you package list, to ensure your packages are set up and installed.<\/p><p>You can find more information about the recipe format <a href=\"https:\/\/github.com\/raxod502\/straight.el#the-recipe-format\">in the straight.el package readme<\/a>.<\/p><blockquote>\n<p>If a MELPA recipe exists for the package you are writing a <code>package!<\/code> declaration for, you may omit keywords and Doom\u2019s package manager will fill them in with values from its original recipe.<\/p>\n<\/blockquote><p>You\u2019ll find more information in the \u201d<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Installing packages from external sources<\/a>\u201d section of the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Getting Started<\/a> guide.<\/p><h2><a id=\"user-content-how-do-i-change-where-an-existing-package-is-installed-from\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-where-an-existing-package-is-installed-from\"><\/a>How do I change where an existing package is installed from?<\/h2><p><code>package!<\/code> declarations in your private <code>packages.el<\/code> file have precedence over modules (even your own). Simply add a new one for that package with the new recipe.<\/p><p>You\u2019ll find more information in the \u201d<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Changing a built-in recipe for a package<\/a>\u201d section of the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Getting Started<\/a> guide.<\/p><h2><a id=\"user-content-how-do-i-disable-a-package-completely\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-disable-a-package-completely\"><\/a>How do I disable a package completely?<\/h2><p>With the <code>package!<\/code> macro\u2019s <code>:disable<\/code> property:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to DOOMDIR\/packages.el\n(package! irony :disable t)<\/pre><\/div><p>Remember to run <code>doom refresh<\/code> afterwards to ensure that the package is uninstalled and disabled.<\/p><p>You\u2019ll find more information in the \u201d<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Disabling packages<\/a>\u201d section of the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Getting Started<\/a> guide.<\/p><h2><a id=\"user-content-how-do-i-reconfigure-a-package-included-in-doom\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-reconfigure-a-package-included-in-doom\"><\/a>How do I reconfigure a package included in Doom?<\/h2><p><code>use-package!<\/code> and <code>after!<\/code> (wrappers around <code>use-package<\/code> and <code>eval-after-load<\/code>, respectively) are your bread and butter for configuring packages in Doom.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Takes a feature symbol or a library name (string)\n(after! evil\n  (setq evil-magic nil))\n;; Takes a major-mode, a quoted hook function or a list of either\n(add-hook! python-mode\n  (setq python-shell-interpreter \"bpython\"))\n(use-package! hl-todo\n  ;; if you omit :defer, :hook, :commands, or :after, then the package is loaded\n  ;; immediately. By using :hook here, the `hl-todo` package won't be loaded\n  ;; until prog-mode-hook is triggered (by activating a major mode derived from\n  ;; it, e.g. python-mode)\n  :hook (prog-mode . hl-todo-mode)\n  :init\n  ;; code here will run immediately\n  :config\n  ;; code here will run after the package is loaded\n  (setq hl-todo-highlight-punctuation \":\"))\n;; There's also `setq-hook!' for setting variables buffer-locally\n(setq-hook! python-mode python-indent-offset 2)<\/pre><\/div><p>See the \u201d<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Configuring Doom<\/a>\u201d section of the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/getting_started.org\">Getting Started<\/a> guide for more explanation and examples.<\/p><h2><a id=\"user-content-how-do-i-change-the-theme\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-the-theme\"><\/a>How do I change the theme?<\/h2><p>There are two ways to load a theme. Both assume the theme is installed and available. You can either set <code>doom-theme<\/code> or manually load a theme with the <code>load-theme<\/code> function.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/config.el\n(setq doom-theme 'doom-tomorrow-night)\n;; or\n(load-theme 'doom-tomorrow-night t)<\/pre><\/div><blockquote>\n<p>At the moment, the only difference between the two is that <code>doom-theme<\/code> is loaded when Emacs has finished initializing at startup and <code>load-theme<\/code> loads the theme immediately. Which you choose depends on your needs, but I recommend setting <code>doom-theme<\/code> because, if I later discover a better way to load themes, I can easily change how Doom uses <code>doom-theme<\/code>, but I can\u2019t (easily) control how you use the <code>load-theme<\/code> function.<\/p>\n<\/blockquote><h3><a id=\"user-content-installing-a-third-party-theme\" class=\"anchor\" aria-hidden=\"true\" href=\"#installing-a-third-party-theme\"><\/a>Installing a third party theme<\/h3><p>To install a theme from a third party plugin, say, <a href=\"https:\/\/github.com\/bbatsov\/solarized-emacs\">solarized<\/a>, you need only install it, then load it:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/packages.el\n(package! solarized-theme)\n;;; add to ~\/.doom.d\/config.el\n(setq doom-theme 'solarized-dark)<\/pre><\/div><p>Don\u2019t forget to run <code>doom refresh<\/code> after adding that <code>package!<\/code> statement to ensure the package is installed.<\/p><h2><a id=\"user-content-how-do-i-change-the-fonts\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-the-fonts\"><\/a>How do I change the fonts?<\/h2><p>Doom exposes five (optional) variables for controlling fonts in Doom, they are:<\/p><ul><li><code>doom-font<\/code><\/li>\n<li><code>doom-variable-pitch-font<\/code><\/li>\n<li><code>doom-serif-font<\/code><\/li>\n<li><code>doom-unicode-font<\/code><\/li>\n<li><code>doom-big-font<\/code> (used for <code>doom-big-font-mode<\/code>)<\/li>\n<\/ul><p>Each of these will accept either a <code>font-spec<\/code>, font string (=\u201dInput Mono-12\u201d=), or <a href=\"https:\/\/wiki.archlinux.org\/index.php\/X_Logical_Font_Description\" rel=\"nofollow\">xlfd font string<\/a>.<\/p><p>e.g.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; Add to ~\/.doom.d\/config.el\n(setq doom-font (font-spec :family \"Input Mono Narrow\" :size 12 :weight 'semi-light)\n      doom-variable-pitch-font (font-spec :family \"Fira Sans\") ; inherits `doom-font''s :size\n      doom-unicode-font (font-spec :family \"Input Mono Narrow\" :size 12)\n      doom-big-font (font-spec :family \"Fira Mono\" :size 19))<\/pre><\/div><h2><a id=\"user-content-how-do-i-bind-my-own-keys-or-change-existing-ones\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-bind-my-own-keys-or-change-existing-ones\"><\/a>How do I bind my own keys (or change existing ones)?<\/h2><p>The <code>map!<\/code> macro is recommended; it is a convenience macro that wraps around Emacs\u2019 (and evil\u2019s) keybinding API, i.e. <code>define-key<\/code>, <code>global-set-key<\/code>, <code>local-set-key<\/code> and <code>evil-define-key<\/code>.<\/p><p>You\u2019ll find comprehensive examples of <code>map!<\/code>\u2019s usage in its documentation (via <code>SPC h f map!<\/code> or <code>C-h f map!<\/code> \u2013 also found <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/api.org\">in docs\/api<\/a>).<\/p><p>You\u2019ll find a more comprehensive example of <code>map!<\/code>\u2019s usage in <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/modules\/config\/default\/+evil-bindings.el\">config\/default\/+evil-bindings.el<\/a>.<\/p><h2><a id=\"user-content-how-do-i-get-motions-to-treat-underscores-as-word-delimiters\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-get-motions-to-treat-underscores-as-word-delimiters\"><\/a>How do I get motions to treat underscores as word delimiters?<\/h2><p>(This explanation comes from <a href=\"https:\/\/github.com\/emacs-evil\/evil#underscore-_-is-not-a-word-character\">emacs-evil\/evil<\/a>\u2019s readme)<\/p><p>An underscore \u201c_\u201d is a word character in Vim. This means that word-motions like <code>w<\/code> skip over underlines in a sequence of letters as if it was a letter itself. In contrast, in Evil the underscore is often a non-word character like operators, e.g. <code>+<\/code>.<\/p><p>The reason is that Evil uses Emacs\u2019 definition of a word and this definition does not often include the underscore. Word characters in Emacs are determined by the syntax-class of the buffer. The syntax-class usually depends on the major-mode of this buffer. This has the advantage that the definition of a \u201cword\u201d may be adapted to the particular type of document being edited. Evil uses Emacs\u2019 definition and does not simply use Vim\u2019s definition in order to be consistent with other Emacs functions. For example, word characters are exactly those characters that are matched by the regular expression character class <code>[:word:]<\/code>.<\/p><p>If you want the underscore to be recognized as word character, you can modify its entry in the syntax-table:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(modify-syntax-entry ?_ \"w\")<\/pre><\/div><p>This gives the underscore the word syntax-class. You can use a mode-hook to modify the syntax-table in all buffers of some mode, e.g.:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; For python\n(add-hook! 'python-mode-hook (modify-syntax-entry ?_ \"w\"))\n;; For ruby\n(add-hook! 'enh-ruby-mode-hook (modify-syntax-entry ?_ \"w\"))\n;; For Javascript\n(add-hook! 'js2-mode-hook (modify-syntax-entry ?_ \"w\"))<\/pre><\/div><h2><a id=\"user-content-how-do-i-change-the-leaderlocalleader-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-the-leaderlocalleader-keys\"><\/a>How do I change the leader\/localleader keys?<\/h2><p>These variables control what key to use for leader and localleader keys:<\/p><ul><li>For Evil users:\n<ul><li><code>doom-leader-key<\/code> (default: <code>SPC<\/code>)<\/li>\n<li><code>doom-localleader-key<\/code> (default: <code>SPC m<\/code>)<\/li>\n<\/ul><\/li>\n<li>For Emacs and Insert state (evil users), and non-evil users:\n<ul><li><code>doom-leader-alt-key<\/code> (default: <code>M-SPC<\/code> for evil users, <code>C-c<\/code> otherwise)<\/li>\n<li><code>doom-localleader-alt-key<\/code> (default: <code>M-SPC m<\/code> for evil users, <code>C-c l<\/code> otherwise)<\/li>\n<\/ul><\/li>\n<\/ul><p>e.g.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/config.el\n(setq doom-leader-key \",\"\n      doom-localleader-key \"\\\\\")<\/pre><\/div><h2><a id=\"user-content-how-do-i-change-the-style-of-line-numbers-or-disable-them-altogether\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-the-style-of-line-numbers-or-disable-them-altogether\"><\/a>How do I change the style of line-numbers (or disable them altogether)?<\/h2><p>Doom uses the <code>display-line-numbers<\/code> package, which is built into Emacs 26+.<\/p><h3><a id=\"user-content-disabling-line-numbers-entirely\" class=\"anchor\" aria-hidden=\"true\" href=\"#disabling-line-numbers-entirely\"><\/a>Disabling line numbers entirely<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/config.el\n(setq display-line-numbers-type nil)\n;; or\n(remove-hook! '(prog-mode-hook text-mode-hook conf-mode-hook)\n              #'display-line-numbers-mode)<\/pre><\/div><h3><a id=\"user-content-switching-to-relative-line-numbers-permanently\" class=\"anchor\" aria-hidden=\"true\" href=\"#switching-to-relative-line-numbers-permanently\"><\/a>Switching to relative line numbers (permanently)<\/h3><p>To change the style of line numbers, change the value of the <code>display-line-numbers-type<\/code> variable. It accepts the following values:<\/p><pre>t            normal line numbers\n'relative    relative line numbers\n'visual      relative line numbers in screen space\nnil          no line numbers\n<\/pre><p>For example:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/config.el\n(setq display-line-numbers-type 'relative)<\/pre><\/div><p>You\u2019ll find more precise documentation on the variable through <code>&lt;help&gt; v display-line-numbers-type<\/code> (<code>&lt;help&gt;<\/code> is <code>SPC h<\/code> for evil users, <code>C-h<\/code> otherwise).<\/p><h3><a id=\"user-content-switching-the-style-of-line-numbers-temporarily\" class=\"anchor\" aria-hidden=\"true\" href=\"#switching-the-style-of-line-numbers-temporarily\"><\/a>Switching the style of line numbers (temporarily)<\/h3><p>Use <code>M-x doom\/toggle-line-numbers<\/code> (bound to <code>SPC t l<\/code> by default) to cycle through the available line number styles in the current buffer.<\/p><p>e.g. <code>normal -&gt; relative -&gt; visual -&gt; disabled -&gt; normal<\/code>.<\/p><h2><a id=\"user-content-how-do-i-change-the-behavior-and-appearance-of-popup-windows\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-the-behavior-and-appearance-of-popup-windows\"><\/a>How do I change the behavior and appearance of popup windows?<\/h2><p>The <code>:ui popup<\/code> module tries to standardize how Emacs handles \u201ctemporary\u201d windows. It includes a set of default rules that tell Emacs where to open them (and how big they should be).<\/p><p>Check out the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/modules\/ui\/popup\/README.org\">:ui popup module\u2019s documentation<\/a> for more on defining your own rules.<\/p><p>You\u2019ll find more comprehensive documentation on <code>set-popup-rule!<\/code> in its docstring (available through <code>SPC h f<\/code> \u2013 or <code>C-h f<\/code> for non-evil users).<\/p><h2><a id=\"user-content-how-do-i-change-the-appearance-a-face-or-faces\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-change-the-appearance-a-face-or-faces\"><\/a>How do I change the appearance a face (or faces)?<\/h2><p>Doom provides the <code>custom-set-faces!<\/code> and <code>custom-theme-set-faces!<\/code> macros as a convenience.<\/p><blockquote>\n<p><strong>Do not use ~M-x customize~ or any of the built-in Emacs customize-<\/strong> API.* Doom does not support it and never will; those settings could break at any time.<\/p>\n<\/blockquote><p>See <code>&lt;help&gt; f custom-set-faces\\!<\/code> (or <code>M-x helpful-function custom-set-faces\\!<\/code>) for documentation and examples on how to use it. <code>&lt;help&gt;<\/code> is <code>SPC h<\/code> for evil users and <code>C-h<\/code> for non-evil users.<\/p><h2><a id=\"user-content-can-doom-be-customized-without-restarting-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#can-doom-be-customized-without-restarting-emacs\"><\/a>Can Doom be customized without restarting Emacs?<\/h2><p>Short answer: You can, but you shouldn\u2019t.<\/p><p>Long answer: Restarting Emacs is always your safest bet, but Doom provides a few tools for experienced Emacs users to skirt around it (most of the time):<\/p><ul><li>Evaluate your changes on-the-fly with <code>+eval\/region<\/code> (bound to the <code>gr<\/code> operator for evil users) or <code>eval-last-sexp<\/code> (bound to <code>C-x C-e<\/code>). Changes take effect immediately.<\/li>\n<li>On-the-fly evaluation won\u2019t work for all changes. e.g. Changing your <code>doom!<\/code> block (i.e. the list of modules for Doom to enable).\n<p>But rather than running <code>doom refresh<\/code> and restarting Emacs, Doom provides <code>M-x doom\/reload<\/code> for your convenience (bound to <code>SPC h r r<\/code> and <code>C-h r r<\/code>). This runs <code>doom refresh<\/code>, restarts the Doom initialization process and re-evaluates your personal config. However, this won\u2019t clear pre-existing state; Doom won\u2019t unload modules\/packages that have already been loaded and it can\u2019t anticipate complications arising from a private config that isn\u2019t idempotent.<\/p>\n<\/li>\n<li>Many <code>bin\/doom<\/code> commands are available as elisp commands with the <code>doom\/\/*<\/code> prefix. e.g. <code>doom\/\/refresh<\/code>, <code>doom\/\/update<\/code>, etc. Feel free to use them, but consider them highly experimental and subject to change without notice.<\/li>\n<li>You can quickly restart Emacs and restore the last session with <code>doom\/restart-and-restore<\/code> (bound to <code>SPC q r<\/code>).<\/li>\n<\/ul><h2><a id=\"user-content-can-vimevil-be-removed-for-a-more-vanilla-emacs-experience\" class=\"anchor\" aria-hidden=\"true\" href=\"#can-vimevil-be-removed-for-a-more-vanilla-emacs-experience\"><\/a>Can Vim\/Evil be removed for a more vanilla Emacs experience?<\/h2><p>Yes! See the <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/modules\/editor\/evil\/README.org\">Removing evil-mode<\/a> section in <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/modules\/editor\/evil\/README.org\">:editor evil<\/a>\u2019s documentation.<\/p><h2><a id=\"user-content-should-i-use-make-or-bindoom\" class=\"anchor\" aria-hidden=\"true\" href=\"#should-i-use-make-or-bindoom\"><\/a>Should I use <code>make<\/code> or <code>bin\/doom<\/code>?<\/h2><p><code>bin\/doom<\/code> is recommended. Doom\u2019s Makefile (to manage your config, at least) is deprecated. It forwards to <code>bin\/doom<\/code> anyway.<\/p><h2><a id=\"user-content-when-should-and-shouldnt-i-use-bindoom\" class=\"anchor\" aria-hidden=\"true\" href=\"#when-should-and-shouldnt-i-use-bindoom\"><\/a>When should and shouldn\u2019t I use <code>bin\/doom<\/code>?<\/h2><p><code>bin\/doom<\/code> is your best friend. It\u2019ll keep all your secrets (mostly because it\u2019s a shell script incapable of sentience and thus incapable of retaining, much less divulging, your secrets to others).<\/p><p>You can run <code>bin\/doom help<\/code> to see what it\u2019s capable of, but here are some commands that you may find particularly useful:<\/p><dl><dt><code>doom doctor<\/code><\/dt>\n<dd>Diagnose common issues in your environment and list missing external dependencies for your enabled modules.<\/dd>\n<dt><code>doom refresh<\/code><\/dt>\n<dd>Ensures that all missing packages are installed, orphaned packages are removed, and metadata properly generated.<\/dd>\n<dt><code>doom install<\/code><\/dt>\n<dd>Install any missing packages.<\/dd>\n<dt><code>doom update<\/code><\/dt>\n<dd>Update all packages that Doom\u2019s (enabled) modules use.<\/dd>\n<dt><code>doom env<\/code><\/dt>\n<dd>Regenerates your envvar file, which contains a snapshot of your shell environment for Doom Emacs to load on startup. You need to run this for changes to your shell environment to take effect.<\/dd>\n<dt><code>doom purge -g<\/code><\/dt>\n<dd>Purge orphaned packages (i.e. ones that aren\u2019t needed anymore) and regraft your repos.<\/dd>\n<dt><code>doom upgrade<\/code><\/dt>\n<dd>Upgrade Doom to the latest version (then update your packages). This is equivalent to:\n<div class=\"highlight highlight-source-shell\"><pre>git pull\ndoom refresh\ndoom update\n    <\/pre><\/div>\n<\/dd>\n<\/dl><h2><a id=\"user-content-when-to-run-doom-refresh\" class=\"anchor\" aria-hidden=\"true\" href=\"#when-to-run-doom-refresh\"><\/a>When to run <code>doom refresh<\/code><\/h2><p>As a rule of thumb you should run <code>doom refresh<\/code> whenever you:<\/p><ul><li>Update Doom with <code>git pull<\/code> instead of <code>doom upgrade<\/code>,<\/li>\n<li>Change your <code>doom!<\/code> block in <code>$DOOMDIR\/init.el<\/code>,<\/li>\n<li>Change autoload files in any module (or <code>$DOOMDIR<\/code>),<\/li>\n<li>Or change the packages.el file in any module (or <code>$DOOMDIR<\/code>).<\/li>\n<li>Install an Emacs package or dependency outside of Emacs (i.e. through your OS package manager).<\/li>\n<\/ul><p>If anything is misbehaving, it\u2019s a good idea to run <code>doom refresh<\/code> first. <code>doom refresh<\/code> is responsible for regenerating your autoloads file (which tells Doom where to find lazy-loaded functions and libraries), installing missing packages, and uninstall orphaned (unneeded) packages.<\/p><h2><a id=\"user-content-how-to-suppress-confirmation-prompts-while-bindoom-is-running\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-suppress-confirmation-prompts-while-bindoom-is-running\"><\/a>How to suppress confirmation prompts while <code>bin\/doom<\/code> is running<\/h2><p>The <code>-y<\/code> and <code>--yes<\/code> flags (or the <code>YES<\/code> environment variable) will force <code>bin\/doom<\/code> to auto-accept confirmation prompts:<\/p><div class=\"highlight highlight-source-shell\"><pre>doom -y update\ndoom --yes update\nYES=1 doom update<\/pre><\/div><h2><a id=\"user-content-why-ivy-over-helm\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-ivy-over-helm\"><\/a>Why Ivy over Helm?<\/h2><p>Short answer: I chose ivy because it is the simpler of the two.<\/p><p>Long answer: Features and performance appear to be the main talking points when comparing the two, but as far as I\u2019m concerned they are equal in both respects (not all across the board, but on average).<\/p><p>Instead, what is important to me is maintainability. As someone who frequently extends and debugs his editor (and maintains a community config), I frequently run up against issues with ivy and helm, but spend disproportionally more time doing so with helm than I do ivy, for little or no gain. Though both frameworks are excellent, the difference in complexity is also reflected in their plugin ecosystems; ivy plugins tend to be lighter, simpler, more consistent and significantly easier to hack if I want to change something. Unless you like helm <em>just<\/em> the way it is out of the box, ivy is just the simpler choice.<\/p><p>And since I dogfood it, Ivy\u2019s integration into Doom will always be a step or three ahead of helm\u2019s.<\/p><h2><a id=\"user-content-why-are-there-no-default-keybinds-for-smartparens-for-evil-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-are-there-no-default-keybinds-for-smartparens-for-evil-users\"><\/a>Why are there no default keybinds for Smartparens (for evil users)?<\/h2><p>Doom only uses smartparens to manage pair \u201ccompletion\u201d (it does the job better than electric-{pair,quote}-mode or the multitude of other pair-management solutions in the Emacs ecosystem at the time of writing).<\/p><p>None of smartparen\u2019s commands have default keybinds for evil users because they are redundant with motions and text-objects provided by evil\/vim. If you disagree, I recommend trying the <code>:editor lispy<\/code> or <code>:editor parinfer<\/code> modules.<\/p><h2><a id=\"user-content-why-do-non-evil-users-get-expand-region-but-not-evil-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-do-non-evil-users-get-expand-region-but-not-evil-users\"><\/a>Why do non-evil users get expand-region, but not evil users?<\/h2><p><code>expand-region<\/code> is redundant with and less precise than evil\u2019s text objects and motions.<\/p><ul><li>There\u2019s a text object for every \u201cstep\u201d of expansion that expand-region provides (and more). To select the word at point = <code>viw<\/code>, symbol at point = <code>vio<\/code>, line at point = <code>V<\/code>, the block at point (by indentation) = <code>vii<\/code>, the block at point (by braces) = <code>vib<\/code>, sentence at point = <code>vis<\/code>, paragraph = <code>vip<\/code>, and so on.<\/li>\n<li>Selection expansion can be emulated by using text objects consecutively: <code>viw<\/code> to select a word, followed by <code>io<\/code> to expand to a symbol, then <code>ib<\/code> expands to the surrounding brackets\/parentheses, etc. There is no reverse of this however; you\u2019d have to restart visual state.<\/li>\n<\/ul><p>The expand-region way dictates you start at some point and expand\/contract until you have what you want selected. The vim\/evil way would rather you select exactly what you want from the get go. In the rare event a text object fails you, a combination of <code>o<\/code> (swaps your cursor between the two ends of the region) and motion keys can adjust the ends of your selection.<\/p><blockquote>\n<p>There are also text objects for xml tags (<code>x<\/code>), C-style function arguments (<code>a<\/code>), angle brackets, and single\/double quotes.<\/p>\n<\/blockquote><p>This is certainly more to remember compared to a pair of expand and contract commands, but text objects (and motions) are the bread and butter of vim\u2019s modal editing paradigm. Vimmers will feel right at home. To everyone else: mastering them will have a far-reaching effect on your productivity. I highly recommend putting in the time to learn them.<\/p><p>Otherwise, it is trivial to install expand-region and binds keys to it yourself:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/packages.el\n(package! expand-region)\n;;; add to ~\/.doom.d\/config.el\n(map! :nv \"C-=\" #'er\/contract-region\n      :nv \"C-+\" #'er\/expand-region)<\/pre><\/div><h2><a id=\"user-content-why-not-use-exec-path-from-shell-instead-of-doom-env\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-not-use-exec-path-from-shell-instead-of-doom-env\"><\/a>Why not use exec-path-from-shell instead of <code>doom env<\/code>?<\/h2><p>In a nutshell, the <code>doom env<\/code> approach is a faster and more robust solution.<\/p><ol><li><code>exec-path-from-shell<\/code> must spawn (at least) one process at startup to scrape your shell environment. This can be arbitrarily slow depending on the user\u2019s shell configuration. A single program (like pyenv or nvm) or config framework (like oh-my-zsh) could undo all of Doom\u2019s startup optimizations in one fell swoop.<\/li>\n<li><code>exec-path-from-shell<\/code> only scrapes <em>some<\/em> state from your shell. You have to be proactive in order to get it to capture all the envvars relevant to your development environment.\n<p>I\u2019d rather it inherit your shell environment <em>correctly<\/em> (and <em>completely<\/em>) or not at all. It frontloads the debugging process rather than hiding it until it you least want to deal with it.<\/p>\n<\/li>\n<\/ol><p>That said, if you still want <code>exec-path-from-shell<\/code>, it is trivial to install yourself:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; add to ~\/.doom.d\/packages.el\n(package! exec-path-from-shell)\n;;; add to ~\/.doom.d\/config.el\n(require 'exec-path-from-shell)\n(when (display-graphic-p)\n  (exec-path-from-shell-initialize))<\/pre><\/div><h2><a id=\"user-content-why-wsbutler-over-delete-trailing-whitespace-or-whitespace-cleanup\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-wsbutler-over-delete-trailing-whitespace-or-whitespace-cleanup\"><\/a>Why wsbutler over delete-trailing-whitespace or whitespace-cleanup?<\/h2><p>TL;DR: <code>ws-butler<\/code> is less imposing.<\/p><p>Don\u2019t be that guy who PRs 99 whitespace adjustments around his one-line contribution. Don\u2019t automate this aggressive behavior by attaching <code>delete-trailing-whitespace<\/code> (or <code>whitespace-cleanup<\/code>) to <code>before-save-hook<\/code>. If you have rambunctious colleagues peppering trailing whitespace into your project, you need to have a talk (with wiffle bats, preferably) rather than play this passive-aggressive game of whack-a-mole.<\/p><p>Here at Doom Inc we believe that operations that mutate entire files should never be automated. Rather, they should be invoked deliberately \u2013 by someone that is aware of the potential consequences. This is where <code>ws-butler<\/code> comes in. It only cleans up whitespace <em>on the lines you\u2019ve touched<\/em> <strong>and<\/strong> it leaves behind virtual whitespace (which is never written to the file, but remains there so your cursor doesn\u2019t get thrown around in all that cleanup work).<\/p><p>In any case, if you had used <code>ws-butler<\/code> from the beginning, trailing whitespace and newlines would never be a problem!<\/p><h2><a id=\"user-content-i-get-the-vanilla-emacs-splash-screen-at-startup\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-get-the-vanilla-emacs-splash-screen-at-startup\"><\/a>I get the vanilla Emacs splash screen at startup<\/h2><p>The most common cause for this is a <code>~\/.emacs<\/code> file. If it exists, Emacs will read this file instead of the <code>~\/.emacs.d<\/code> directory, ignoring Doom altogether.<\/p><p>If this isn\u2019t the case, try running <code>bin\/doom doctor<\/code>. It can detect a variety of common issues and may give you some clues as to what is wrong.<\/p><h2><a id=\"user-content-i-see-a-blank-scratch-buffer-at-startup\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-see-a-blank-scratch-buffer-at-startup\"><\/a>I see a blank scratch buffer at startup<\/h2><p>This commonly means that Emacs can\u2019t find your private doom config (in <code>~\/.doom.d<\/code> or <code>~\/.config\/doom<\/code>). Make sure <strong>only one of these two<\/strong> folders exist, and that it has an init.el file with a <code>doom!<\/code> block. Running <code>doom install<\/code> will populate your private doom directory with the bare minimum you need to get going.<\/p><p>If nothing else works, try running <code>bin\/doom doctor<\/code>. It can detect a variety of common issues and may give you some clues as to what is wrong.<\/p><h2><a id=\"user-content-strange-or-incorrect-icons-are-displayed-everywhere\" class=\"anchor\" aria-hidden=\"true\" href=\"#strange-or-incorrect-icons-are-displayed-everywhere\"><\/a>Strange (or incorrect) icons are displayed everywhere<\/h2><p>Many of Doom\u2019s UI modules use the <code>all-the-icons<\/code> plugin, which uses special fonts to display icons. These fonts must be installed for them to work properly, otherwise you\u2019ll get a bunch of squares and mismatched icons. When running <code>doom install<\/code>, you will be asked whether you want these installed for you or not.<\/p><p>If you did not accept or need to reinstall those fonts, MacOS and Linux users can install them with <code>M-x all-the-icons-install-fonts<\/code>. Windows users will need to use this command to download the fonts somewhere, then they must install them manually (e.g. by double-clicking each file in explorer).<\/p><h2><a id=\"user-content-void-variable-and-void-function-errors-on-startup\" class=\"anchor\" aria-hidden=\"true\" href=\"#void-variable-and-void-function-errors-on-startup\"><\/a><code>void-variable<\/code> and <code>void-function<\/code> errors on startup<\/h2><p>The most common culprit for these types of errors are:<\/p><ol><li>An out-of-date autoloads file. To regenerate it, run <code>doom refresh<\/code>.\n<p>To avoid this issue, remember to run <code>doom refresh<\/code> whenever you modify your <code>doom!<\/code> block in <code>~\/.doom.d\/init.el<\/code>, or add <code>package!<\/code> declarations to <code>~\/.doom.d\/packages.el<\/code>. Or if you modify <code>~\/.emacs.d\/.local<\/code> by hand, for whatever reason.<\/p>\n<p>See <code>doom help refresh<\/code> for details on what this command does and when you should use it.<\/p>\n<\/li>\n<li>Emacs byte-code isn\u2019t forward compatible. If you\u2019ve recently switched to a newer (or older) version of Emacs, you\u2019ll need to either reinstall or recompile your installed plugins. This can be done by:\n<ul><li>Running <code>doom build<\/code>,<\/li>\n<li>Or deleting <code>~\/.emacs.d\/.local\/straight<\/code> then running <code>doom install<\/code> (this will take a while).<\/li>\n<\/ul><\/li>\n<\/ol><h2><a id=\"user-content-doom-cant-find-my-executablesdoesnt-inherit-the-correct-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#doom-cant-find-my-executablesdoesnt-inherit-the-correct-path\"><\/a>Doom can\u2019t find my executables\/doesn\u2019t inherit the correct <code>PATH<\/code><\/h2><p>The two most common causes for PATH issues in Doom are:<\/p><ol><li>Your shell configuration doesn\u2019t configure <code>PATH<\/code> correctly. If <code>which &lt;PROGRAM&gt;<\/code> doesn\u2019t emit the path you expect on the command line then this is likely the case.<\/li>\n<li>Your app launcher (rofi, albert, docky, dmenu, sxhkd, etc) is launching Emacs with the wrong shell, either because it defaults to a different shell from the one you use or the app launcher itself inherits the wrong environment because <em>it<\/em> was launched from the wrong shell.<\/li>\n<li>You\u2019re a Mac user launching Emacs from an Emacs.app bundle. MacOS launches these apps from an isolated environment.<\/li>\n<\/ol><p>As long as your shell is properly configured, there is a simple solution to issues #1 and #3: generate an envvar file by running <code>doom env<\/code>. This scrapes your shell environment into a file that is loaded when Doom Emacs starts up. Check out <code>doom help env<\/code> for details on how this works.<\/p><p>For issue #2, you\u2019ll need to investigate your launcher. <a href=\"https:\/\/discord.gg\/bcZ6P3y\" rel=\"nofollow\">Our Discord<\/a> is a good place to ask about it.<\/p><h2><a id=\"user-content-theres-artefacting-on-my-icon-fonts-in-gui-emacs-956\" class=\"anchor\" aria-hidden=\"true\" href=\"#theres-artefacting-on-my-icon-fonts-in-gui-emacs-956\"><\/a>There\u2019s artefacting on my icon fonts in GUI Emacs (<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/issues\/956\">#956<\/a>)<\/h2><p>Check your font rendering settings. Changing the RGBA order to \u201crgba\u201d will often fix this issue. See <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/issues\/956\">#956<\/a> for details.<\/p><h2><a id=\"user-content-the-s-and-s-keys-dont-behave-like-they-do-in-vimevil-1307\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-s-and-s-keys-dont-behave-like-they-do-in-vimevil-1307\"><\/a>The <code>s<\/code> and <code>S<\/code> keys don\u2019t behave like they do in vim\/evil (<a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/issues\/1307\">#1307<\/a>)<\/h2><p>This is intentional. <code>s<\/code> and <code>S<\/code> have been replaced by the evil-snipe plugin, which provides 2-character versions of the f\/F motion keys, ala vim-seek or vim-sneak.<\/p><p>These keys were changed because they are redundant with <code>cl<\/code> and <code>cc<\/code> respectively (and the new behavior was deemed more useful).<\/p><p>If you still want to restore the old behavior, simply disable evil-snipe-mode:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; in ~\/.doom.d\/config.el\n(after! evil-snipe\n  (evil-snipe-mode -1))<\/pre><\/div><h2><a id=\"user-content-changes-to-my-config-arent-taking-effect\" class=\"anchor\" aria-hidden=\"true\" href=\"#changes-to-my-config-arent-taking-effect\"><\/a>Changes to my config aren\u2019t taking effect<\/h2><ol><li>Make sure you don\u2019t have both <code>~\/.doom.d<\/code> and <code>~\/.config\/doom<\/code> directories. Doom will ignore the former if the latter exists.<\/li>\n<li>Remember to run <code>doom refresh<\/code> when it is necessary. To get to know when, exactly, you should run this command, run <code>doom help refresh<\/code>.<\/li>\n<\/ol><p>If neither of these solve your issue, try <code>bin\/doom doctor<\/code>. It will detect a variety of common issues, and may give you some clues as to what is wrong.<\/p><h2><a id=\"user-content-the-frame-goes-black-on-macos-while-in-full-screen-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-frame-goes-black-on-macos-while-in-full-screen-mode\"><\/a>The frame goes black on MacOS, while in full-screen mode<\/h2><p>There are known issues with childframes and macOS\u2019s fullscreen mode. There is no known fix for this. To work around it, you must either:<\/p><ol><li>Avoid MacOS native fullscreen by maximizing Emacs instead,<\/li>\n<li>Disable childframes (controlled by the <code>+childframe<\/code> flag on the modules that support it),<\/li>\n<li>Install Emacs via the <code>emacs-mac<\/code> homebrew formula.<\/li>\n<\/ol><h2><a id=\"user-content-doom-crashes-when\" class=\"anchor\" aria-hidden=\"true\" href=\"#doom-crashes-when\"><\/a>Doom crashes when\u2026<\/h2><p>Here are a few common causes for random crashes:<\/p><ul><li>You have enabled <code>undo-tree-auto-save-history<\/code>. A bloated cache for a particular file can cause a stack overflow. These caches are stored in <code>~\/.emacs.d\/.local\/cache\/undo-tree-hist\/<\/code>. Delete this folder to clear it.<\/li>\n<li>On some systems (particularly MacOS), manipulating the fringes or window margins can cause Emacs to crash. This is most prominent in the Doom Dashboard (which tries to center its contents), in org-mode buffers (which uses <code>org-indent-mode<\/code> to create virtual indentation), or magit. There is currently no known fix for this, as it can\u2019t be reliably reproduced. Your best bet is to reinstall\/rebuild Emacs or disable the errant plugins\/modules. e.g.\n<p>To disable org-indent-mode:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(after! org\n  (setq org-startup-indented nil))\n    <\/pre><\/div>\n<p>Or disable the <code>:ui doom-dashboard<\/code> &amp; <code>:tools magit<\/code> modules (see <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/issues\/1170\">#1170<\/a>).<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-cant-load-my-theme-unable-to-find-theme-file-for-x-errors\" class=\"anchor\" aria-hidden=\"true\" href=\"#cant-load-my-theme-unable-to-find-theme-file-for-x-errors\"><\/a>Can\u2019t load my theme; <code>unable to find theme file for X<\/code> errors<\/h2><p>This means Emacs can\u2019t find the X-theme.el file for the theme you want to load. Emacs will search for this file in <code>custom-theme-load-path<\/code> and <code>custom-theme-directory<\/code>. There are a couple reasons why it can\u2019t be found:<\/p><ol><li>It is generally expected that third party themes will <a href=\"https:\/\/github.com\/hlissner\/emacs-doom-themes\/blob\/master\/doom-themes.el#L400-L405\">add themselves<\/a> to <code>custom-theme-load-path<\/code>, but you will occasionally encounter a theme that does not. This should be reported upstream.\n<p>In the meantime, you can get around this by eagerly loading the package:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'third-party-theme)\n(setq doom-theme 'third-party)\n    <\/pre><\/div>\n<\/li>\n<li>You\u2019ve appended <code>-theme<\/code> to the end of your theme\u2019s name.\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq doom-theme 'third-party-theme)\n    <\/pre><\/div>\n<p>When you load a theme Emacs searches for <code>X-theme.el<\/code>. If you set <code>doom-theme<\/code> to ~\u2019third-party-theme~, it will search for <code>third-party-theme-theme.el<\/code>. This is rarely intentional. Omit the <code>-theme<\/code> suffix.<\/p>\n<\/li>\n<li>Did you run <code>doom refresh<\/code> after adding your third party theme plugin\u2019s <code>package!<\/code> declaration to <code>~\/.doom.d\/packages.el<\/code>?<\/li>\n<\/ol><h2><a id=\"user-content-tramp-connections-hang-forever-when-connecting\" class=\"anchor\" aria-hidden=\"true\" href=\"#tramp-connections-hang-forever-when-connecting\"><\/a>TRAMP connections hang forever when connecting<\/h2><p>You\u2019ll find solutions <a href=\"https:\/\/www.emacswiki.org\/emacs\/TrampMode#toc7\" rel=\"nofollow\">on the emacswiki<\/a>.<\/p><h2><a id=\"user-content-an-upstream-package-was-broken-and-i-cant-update-it\" class=\"anchor\" aria-hidden=\"true\" href=\"#an-upstream-package-was-broken-and-i-cant-update-it\"><\/a>An upstream package was broken and I can\u2019t update it<\/h2><p>Sometimes, if you\u2019ve installed a <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/issues\/2213\">broken package<\/a> which was subsequently fixed upstream, you can\u2019t run <code>doom update<\/code> to get the latest fixes due to evaluation errors.<\/p><p>In those cases, you need to delete the broken local copy before you can install the new one, which is achieved thus:<\/p><ol><li>Comment out the broken module\/package.<\/li>\n<li>Run <code>doom refresh -p<\/code>.<\/li>\n<li>Uncomment the module\/package.<\/li>\n<li>Run <code>doom refresh<\/code>.<\/li>\n<\/ol><h2><a id=\"user-content-why-do-i-see-ugly-indentation-highlights-for-tabs\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-do-i-see-ugly-indentation-highlights-for-tabs\"><\/a>Why do I see ugly indentation highlights for tabs?<\/h2><p><a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/core\/core-ui.el#L132-L150\">Doom highlights non-standard indentation<\/a>. i.e. Indentation that doesn\u2019t match the indent style you\u2019ve set for that file. Doom uses space indentation for most languages by default (excluding languages where tabs are the norm, like Go).<\/p><p>There are a couple ways to address this:<\/p><ol><li>Change <code>indent-tabs-mode<\/code> (nil = spaces, t = tabs).\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; use tab indentation everywhere\n(setq-default indent-tabs-mode t)\n;; or only in certain modes\n(setq-hook! 'sh-mode-hook indent-tabs-mode t) ; shell scripts\n(setq-hook! '(c-mode-hook c++-mode-hook) indent-tabs-mode t)  ; C\/C++\n    <\/pre><\/div>\n<\/li>\n<li>Use <a href=\"https:\/\/editorconfig.org\/\" rel=\"nofollow\">editorconfig<\/a> to configure code style on a per-project basis. If you enable Doom\u2019s <code>:tools editorconfig<\/code> module, Doom will recognize <code>.editorconfigrc<\/code> files.<\/li>\n<li>Or trust in dtrt-indent; a plugin Doom uses to analyze and detect indentation when you open a file (that isn\u2019t in a project with an editorconfig file). This isn\u2019t foolproof, and won\u2019t work for files that have no content in them, but it can help in one-off scenarios.<\/li>\n<\/ol>","created_at":"2020-01-13T21:11:13+0000","updated_at":"2020-01-13T21:11:13+0000","published_by":["hlissner"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":41,"domain_name":"github.com","preview_picture":"https:\/\/repository-images.githubusercontent.com\/21763822\/47152a00-a255-11e9-9ca0-37fec3ef82cc","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","evil-mode"],"is_public":false,"id":1109,"title":"gabesoft\/evil-mc","url":"https:\/\/github.com\/gabesoft\/evil-mc","content":"<pre>\u2593\u2588\u2588\u2588\u2588\u2588 \u2588\u2588\u2592   \u2588\u2593 \u2588\u2588\u2593 \u2588\u2588\u2593        \u2588\u2588\u2588\u2584 \u2584\u2588\u2588\u2588\u2593 \u2584\u2588\u2588\u2588\u2588\u2584  \n\u2593\u2588   \u2580\u2593\u2588\u2588\u2591   \u2588\u2592\u2593\u2588\u2588\u2592\u2593\u2588\u2588\u2592       \u2593\u2588\u2588\u2592\u2580\u2588\u2580 \u2588\u2588\u2592\u2592\u2588\u2588\u2580 \u2580\u2588  \n\u2592\u2588\u2588\u2588   \u2593\u2588\u2588  \u2588\u2592\u2591\u2592\u2588\u2588\u2592\u2592\u2588\u2588\u2591       \u2593\u2588\u2588    \u2593\u2588\u2588\u2591\u2592\u2593\u2588    \u2584 \n\u2592\u2593\u2588  \u2584  \u2592\u2588\u2588 \u2588\u2591\u2591\u2591\u2588\u2588\u2591\u2592\u2588\u2588\u2591       \u2592\u2588\u2588    \u2592\u2588\u2588 \u2592\u2593\u2593\u2584 \u2584\u2588\u2588\u2592\n\u2591\u2592\u2588\u2588\u2588\u2588\u2592  \u2592\u2580\u2588\u2591  \u2591\u2588\u2588\u2591\u2591\u2588\u2588\u2588\u2588\u2588\u2588\u2592   \u2592\u2588\u2588\u2592   \u2591\u2588\u2588\u2592\u2592 \u2593\u2588\u2588\u2588\u2580 \u2591\n\u2591\u2591 \u2592\u2591 \u2591  \u2591 \u2590\u2591  \u2591\u2593  \u2591 \u2592\u2591\u2593  \u2591   \u2591 \u2592\u2591   \u2591  \u2591\u2591 \u2591\u2592 \u2592  \u2591\n \u2591 \u2591  \u2591  \u2591 \u2591\u2591   \u2592 \u2591\u2591 \u2591 \u2592  \u2591   \u2591  \u2591      \u2591  \u2591  \u2592   \n   \u2591       \u2591\u2591   \u2592 \u2591  \u2591 \u2591      \u2591      \u2591   \u2591        \n   \u2591  \u2591     \u2591   \u2591      \u2591  \u2591          \u2591   \u2591 \u2591      \n           \u2591                             \u2591        \n<\/pre><p><a href=\"https:\/\/travis-ci.org\/gabesoft\/evil-mc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5a6314c1fbfe258d26e10ab8d5b5b8db8fc1bbef\/68747470733a2f2f7472617669732d63692e6f72672f67616265736f66742f6576696c2d6d632e7376673f6272616e63683d6d6173746572\" alt=\"https:\/\/travis-ci.org\/gabesoft\/evil-mc.svg?branch=master\" data-canonical-src=\"https:\/\/travis-ci.org\/gabesoft\/evil-mc.svg?branch=master\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/melpa.org\/#\/evil-mc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/0d62d8974d486716a31902cb860d4d115dcda4ee\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f6576696c2d6d632d62616467652e7376673f7374796c653d666c61742d737175617265\" alt=\"http:\/\/melpa.org\/packages\/evil-mc-badge.svg?style=flat-square\" data-canonical-src=\"http:\/\/melpa.org\/packages\/evil-mc-badge.svg?style=flat-square\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/melpa.org\/#\/evil-mc\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/85c130f37024c0fffedcaa480aec5e2640ec594b\/687474703a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f6576696c2d6d632d62616467652e7376673f7374796c653d666c61742d737175617265\" alt=\"http:\/\/stable.melpa.org\/packages\/evil-mc-badge.svg?style=flat-square\" data-canonical-src=\"http:\/\/stable.melpa.org\/packages\/evil-mc-badge.svg?style=flat-square\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h3><a id=\"user-content-synopsis\" class=\"anchor\" aria-hidden=\"true\" href=\"#synopsis\"><\/a>Synopsis<\/h3><p><strong>evil-mc<\/strong> provides multiple cursors functionality for Emacs when used with <strong>evil-mode<\/strong><\/p><h3><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h3><p>Start with:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'evil-mc)<\/pre><\/div><h4><a id=\"user-content-local-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#local-setup\"><\/a>Local Setup<\/h4><p>To enable or disable <strong>evil-mc<\/strong> mode for a single buffer use:<br \/><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-mc-mode  1) ;; enable\n(evil-mc-mode -1) ;; disable<\/pre><\/div><h4><a id=\"user-content-global-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#global-setup\"><\/a>Global Setup<\/h4><p>To enable or disable <strong>evil-mc<\/strong> mode for all buffers use:<br \/><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(global-evil-mc-mode  1) ;; enable\n(global-evil-mc-mode -1) ;; disable<\/pre><\/div><h4><a id=\"user-content-basic-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-usage\"><\/a>Basic Usage<\/h4><p>The main commands used to create or delete cursors are:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-mc-make-all-cursors)\n;; Create cursors for all strings that match the selected \n;; region or the symbol under cursor.\n(evil-mc-undo-all-cursors)\n;; Remove all cursors.\n(evil-mc-make-and-goto-next-match)\n;; Make a cursor at point and go to the next match of the \n;; selected region or the symbol under cursor.\n(evil-mc-skip-and-goto-next-match)\n;; Go to the next match of the selected region or the symbol under \n;; cursor without creating a cursor at point.<\/pre><\/div><p>The above commands as well as others, detailed below, are setup with key bindings when the <strong>evil-mc<\/strong> mode is enabled. The keys are defined in <code>evil-mc-key-map<\/code>. You can take a look at that variable declaration in <a href=\"https:\/\/github.com\/gabesoft\/evil-mc\/blob\/master\/evil-mc.el\">evil-mc.el<\/a> to see all key bindings. But, in short, <code>C-n<\/code> \/ <code>C-p<\/code> are used for creating cursors, and <code>M-n<\/code> \/ <code>M-p<\/code> are used for cycling through cursors. The commands that create cursors wrap around; but, the ones that cycle them do not. To skip creating a cursor forward use <code>C-t<\/code> or <code>grn<\/code> and backward <code>grp<\/code>. Finally use <code>gru<\/code> to \u201cundo\u201d the last added cursor, and <code>grq<\/code> to remove all cursors.<\/p><p>Optionally set up visual mode keybindings for <code>I<\/code> and <code>A<\/code> to create cursors at the beginning or end of every visually selected line.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-key 'visual evil-mc-key-map\n  \"A\" #'evil-mc-make-cursor-in-visual-selection-end\n  \"I\" #'evil-mc-make-cursor-in-visual-selection-beg)<\/pre><\/div><p>For an example of setting up <strong>evil-mc<\/strong> see this <a href=\"https:\/\/github.com\/gabesoft\/evil-mc\/blob\/master\/evil-mc-setup.el\">setup file<\/a><\/p><h4><a id=\"user-content-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#commands\"><\/a>Commands<\/h4><p>Here\u2019s a detailed list of all commands used to create, navigate through, or delete cursors:<br \/><em>All the commands below assume that there is a real cursor and possibly some fake cursors.<\/em><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-mc-make-all-cursors)\n;; Make a cursor for every match of the selected region or the symbol at point.\n(evil-mc-undo-all-cursors)\n;; Remove all cursors.\n(evil-mc-undo-last-added-cursor)\n;; Remove the last added cursor and move point to its position.\n(evil-mc-make-and-goto-next-match)\n;; Make a cursor at point, and go to the next match of the \n;; selected region or the symbol at point.\n(evil-mc-make-and-goto-prev-match)\n;; Make a cursor at point, and go to the previous match of the \n;; selected region or the symbol at point.\n(evil-mc-skip-and-goto-next-match)\n;; Go to the next match of the selected region or symbol at point\n;; without making a cursor at point. This command can be used to\n;; remove unwanted cursors.\n(evil-mc-skip-and-goto-prev-match)\n;; Go to the previous match of the selected region or symbol at point\n;; without making a cursor at point. This command can be used to\n;; remove unwanted cursors.\n(evil-mc-make-and-goto-prev-cursor)\n;; Make a cursor at point and move point to the cursor\n;; closest to it when searching backwards.\n(evil-mc-make-and-goto-next-cursor)\n;; Make a cursor at point and move point to the cursor\n;; closest to it when searching forwards.\n(evil-mc-skip-and-goto-prev-cursor)\n;; Move point to the cursor closest to it when searching backwards\n;; without making a cursor at point. This command can be used to\n;; remove unwanted cursors.\n(evil-mc-skip-and-goto-next-cursor)\n;; Move point to the cursor closest to it when searching forwards\n;; without making a cursor at point. This command can be used to\n;; remove unwanted cursors.\n(evil-mc-make-and-goto-first-cursor)\n;; Make a cursor at point and move point to the cursor at the first position.\n(evil-mc-make-and-goto-last-cursor)\n;; Make a cursor at point and move point to the cursor at the last position.\n(evil-mc-make-cursor-here)\n;; Create a cursor at point. This command should be used with `evil-mc-pause-cursors'.\n(evil-mc-pause-cursors)\n;; Pause all fake cursors. This can be used with `evil-mc-make-cursor-here'\n(evil-mc-resume-cursors)\n;; Call to resume paused cursors.\n(evil-mc-make-cursor-in-visual-selection-beg)\n;; Create cursors at the beginning of every visually selected line.\n(evil-mc-make-cursor-in-visual-selection-end)\n;; Create cursors at the end of every visually selected line.<\/pre><\/div><h4><a id=\"user-content-customization\" class=\"anchor\" aria-hidden=\"true\" href=\"#customization\"><\/a>Customization<\/h4><p><strong>evil-mc<\/strong> can be customized in several ways:<\/p><ul><li>Every known command is executed using a command handler defined in a variable called <code>evil-mc-known-commands<\/code> in <a href=\"https:\/\/github.com\/gabesoft\/evil-mc\/blob\/master\/evil-mc-known-commands.el\">evil-mc-known-commands.el<\/a>. Those can be overridden by defining the <code>evil-mc-custom-known-commands<\/code> variable. See the documentation of that variable in <a href=\"https:\/\/github.com\/gabesoft\/evil-mc\/blob\/master\/evil-mc.el\">evil-mc.el<\/a> for more info.<\/li>\n<li>Some minor modes are incompatible with <strong>evil-mc<\/strong>. Those modes are defined in <code>evil-mc-incompatible-minor-modes<\/code> and can be overridden by defining that variable.<\/li>\n<li>In addition there are two hooks that can be used to temporarily disable or enable additional functionality while there are multiple cursors active\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>evil-mc-before-cursors-created\n;; this hook runs just before the first cursor is created\nevil-mc-after-cursors-created\n;; this hook runs just after the last cursor is deleted\n    <\/pre><\/div>\n<\/li>\n<\/ul><h3><a id=\"user-content-mode-line-text-and-colors\" class=\"anchor\" aria-hidden=\"true\" href=\"#mode-line-text-and-colors\"><\/a>Mode line text and colors<\/h3><p>There are 4 variables, that can change the mode lines text, and its color.<\/p><h4><a id=\"user-content-only-one-cursor\" class=\"anchor\" aria-hidden=\"true\" href=\"#only-one-cursor\"><\/a>Only one cursor<\/h4><p>The <code>emc<\/code> text can be hidden, by setting this variable to <code>nil<\/code>. (default: <code>t<\/code>)<\/p><pre>(setq evil-mc-one-cursor-show-mode-line-text t)\n<\/pre><h4><a id=\"user-content-two-or-more-cursors-resumed-unpaused\" class=\"anchor\" aria-hidden=\"true\" href=\"#two-or-more-cursors-resumed-unpaused\"><\/a>Two or more cursors, resumed (unpaused)<\/h4><p>The resumed mode line text, can have two different colors:<\/p><ul><li>The cursors color, when this variable is <code>t<\/code>.<\/li>\n<li>The default colors, when this variable is <code>nil<\/code>.<\/li>\n<\/ul><p>(default: <code>t<\/code>)<\/p><pre>(setq evil-mc-mode-line-text-cursor-color t)\n<\/pre><h4><a id=\"user-content-two-or-more-cursors-paused\" class=\"anchor\" aria-hidden=\"true\" href=\"#two-or-more-cursors-paused\"><\/a>Two or more cursors, paused<\/h4><p>The <code>(paused)<\/code> text can be hidden, by setting this variable to <code>nil<\/code>. (default: <code>t<\/code>)<\/p><pre>(setq evil-mc-mode-line-text-paused t)\n<\/pre><p>The paused mode line text can have three different colors:<\/p><ul><li>Inverse colors, when the inverse colors variable is <code>t<\/code>.<\/li>\n<li>Cursors color, when the inverse colors variable is <code>nil<\/code>, and the cursor color variable is <code>t<\/code>.<\/li>\n<li>Default colors, when both the inverse and cursor color variables are <code>nil<\/code>.<\/li>\n<\/ul><p>(default: <code>t<\/code>, for both the inverse and cursor variables)<\/p><pre>(setq evil-mc-mode-line-text-inverse-colors t)\n(setq evil-mc-mode-line-text-cursor-color t)\n<\/pre><h3><a id=\"user-content-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#notes\"><\/a>Notes<\/h3><ul><li>Most evil motions and operators are supported but not every single command will work.<\/li>\n<li>If the cursors don\u2019t seem to work during a command, either the command is not known (see <code>evil-mc-known-commands<\/code> in <a href=\"https:\/\/github.com\/gabesoft\/evil-mc\/blob\/master\/evil-mc-known-commands.el\">evil-mc-known-commands.el<\/a>) or some minor modes could be interfering with the evil-mc operations.<\/li>\n<li>Issues and pull requests are welcome.<\/li>\n<\/ul><h4><a id=\"user-content-debugging\" class=\"anchor\" aria-hidden=\"true\" href=\"#debugging\"><\/a>Debugging<\/h4><ul><li>When a command does not work, and you want to get more information, you can enable (or disable) debugging by running any of the commands below interactively.\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-mc-executing-debug-on)\n;; Turn debug on while executing a command.\n(evil-mc-executing-debug-off)\n;; Turn debug off while executing a command.\n(evil-mc-recording-debug-on)\n;; Turn debug on while recording a command.\n(evil-mc-recording-debug-off)\n;; Turn debug off while recording a command.\n(evil-mc-all-debug-on)\n;; Turn all debug on.\n(evil-mc-all-debug-off)\n;; Turn all debug off.\n    <\/pre><\/div>\n<\/li>\n<\/ul><h4><a id=\"user-content-limitations\" class=\"anchor\" aria-hidden=\"true\" href=\"#limitations\"><\/a>Limitations<\/h4><ul><li>After an undo command the cursors will return to their original positions if <a href=\"http:\/\/www.emacswiki.org\/emacs\/UndoTree\" rel=\"nofollow\">undo-tree<\/a> mode is enabled and <code>evil-repeat<\/code> has not been used.<\/li>\n<li>Redo may cause the real cursor to get out of sync with the others. This can be worked around by setting a mark and returning to it after a redo.<\/li>\n<li>Jumps work if <a href=\"https:\/\/github.com\/bling\/evil-jumper\">evil-jumper<\/a> mode is enabled<\/li>\n<li>Search commands such as <code>evil-search-forward<\/code>, <code>evil-search-backward<\/code>, and <code>evil-search-next<\/code> are not supported<\/li>\n<\/ul><h4><a id=\"user-content-known-issues\" class=\"anchor\" aria-hidden=\"true\" href=\"#known-issues\"><\/a>Known issues<\/h4><ul><li>Only named commands can be executed by the fake cursors.<\/li>\n<li>There could be a performance penalty when there are too many cursors (30+).<\/li>\n<li>Paste will not work when <a href=\"https:\/\/github.com\/syl20bnr\/spacemacs\">spacemacs<\/a>\u2019 paste micro state is enabled. This is due to the fact that <code>evil-paste-pop<\/code> and <code>evil-paste-pop-next<\/code> commands are not supported.<\/li>\n<li><code>evil-repeat<\/code> works only for some commands. In particular it doesn\u2019t work for delete. It will also interfere with the cursor positions during an undo or redo operation.<\/li>\n<\/ul><pre>.__                                            .___.__  __  .__                \n|  |__ _____  ______ ______ ___.__.   ____   __| _\/|__|\/  |_|__| ____    ____  \n|  |  \\\\__  \\ \\____ \\\\____ &lt;   |  | _\/ __ \\ \/ __ | |  \\   __\\  |\/    \\  \/ ___\\ \n|   Y  \\\/ __ \\|  |_&gt; &gt;  |_&gt; &gt;___  | \\  ___\/\/ \/_\/ | |  ||  | |  |   |  \\\/ \/_\/  &gt;\n|___|  (____  \/   __\/|   __\/\/ ____|  \\___  &gt;____ | |__||__| |__|___|  \/\\___  \/ \n     \\\/     \\\/|__|   |__|   \\\/           \\\/     \\\/                  \\\/\/_____\/  \n<\/pre>","created_at":"2020-01-13T20:52:53+0000","updated_at":"2020-01-13T20:52:57+0000","published_by":["gabesoft"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/606000?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","evil-mode"],"is_public":false,"id":1108,"title":"hlissner\/evil-multiedit","url":"https:\/\/github.com\/hlissner\/evil-multiedit","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/dc0d6a3fd87099f8b73372c35f93b5e859b99b70\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6576696c2d2d6d756c7469656469742d76312e332e382d626c75652e737667\"><img src=\"https:\/\/camo.githubusercontent.com\/dc0d6a3fd87099f8b73372c35f93b5e859b99b70\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6576696c2d2d6d756c7469656469742d76312e332e382d626c75652e737667\" alt=\"evil-multiedit\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/evil--multiedit-v1.3.8-blue.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/github.com\/hlissner\/evil-multiedit\/blob\/master\/LICENSE\"><img src=\"https:\/\/camo.githubusercontent.com\/6a40b4913be1852a0fe4a896c5912412202f9169\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d677265656e2e737667\" alt=\"MIT\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-MIT-green.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/melpa.org\/#\/evil-multiedit\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/0fcc40899744ed4adc702929067ee3ae05d2df6b\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f6576696c2d6d756c7469656469742d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"http:\/\/melpa.org\/packages\/evil-multiedit-badge.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/stable.melpa.org\/#\/evil-multiedit\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f077ffa491e36c87765091bcfa206a6278b64c84\/687474703a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f6576696c2d6d756c7469656469742d62616467652e737667\" alt=\"MELPA Stable\" data-canonical-src=\"http:\/\/stable.melpa.org\/packages\/evil-multiedit-badge.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/travis-ci.org\/hlissner\/evil-multiedit\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/da03796037ddbdfbf146ef6936d06b7b3747e949\/68747470733a2f2f7472617669732d63692e6f72672f686c6973736e65722f6576696c2d6d756c7469656469742e706e673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/hlissner\/evil-multiedit.png?branch=master\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>This plugin was an answer to the lack of proper multiple cursor support in Emacs+evil. It allows you to select and edit matches interactively, integrating <code>iedit-mode<\/code> into evil-mode with an attempt at sensible defaults.<\/p><p>Since then, <a href=\"https:\/\/github.com\/gabesoft\/evil-mc\">evil-mc<\/a> has matured, and now that I use both I've found they can coexist, filling different niches, complimenting evil's built-in column\/line-wise editing operations.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/hlissner\/evil-multiedit\/blob\/screenshots\/main.gif?raw=true\"><img src=\"https:\/\/github.com\/hlissner\/evil-multiedit\/raw\/screenshots\/main.gif?raw=true\" alt=\"evil-multiedit\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><blockquote>\n<p>Thanks to <a href=\"https:\/\/github.com\/syl20bnr\">syl20bnr<\/a> for his <a href=\"https:\/\/github.com\/syl20bnr\/evil-iedit-state\">evil-iedit-state<\/a> plugin, which this plugin was heavily inspired by.<\/p>\n<\/blockquote><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>The package is available on MELPA.<\/p><p><code>M-x package-install RET evil-multiedit<\/code><\/p><p>Then load it up:<\/p><p><code>(require 'evil-multiedit)<\/code><\/p><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><p>Evil-multiedit does not automatically bind any keys. Call <code>(evil-multiedit-default-keybinds)<\/code> to bind my recommended configuration:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Highlights all matches of the selection in the buffer.\n(define-key evil-visual-state-map \"R\" 'evil-multiedit-match-all)\n;; Match the word under cursor (i.e. make it an edit region). Consecutive presses will\n;; incrementally add the next unmatched match.\n(define-key evil-normal-state-map (kbd \"M-d\") 'evil-multiedit-match-and-next)\n;; Match selected region.\n(define-key evil-visual-state-map (kbd \"M-d\") 'evil-multiedit-match-and-next)\n;; Insert marker at point\n(define-key evil-insert-state-map (kbd \"M-d\") 'evil-multiedit-toggle-marker-here)\n;; Same as M-d but in reverse.\n(define-key evil-normal-state-map (kbd \"M-D\") 'evil-multiedit-match-and-prev)\n(define-key evil-visual-state-map (kbd \"M-D\") 'evil-multiedit-match-and-prev)\n;; OPTIONAL: If you prefer to grab symbols rather than words, use\n;; `evil-multiedit-match-symbol-and-next` (or prev).\n;; Restore the last group of multiedit regions.\n(define-key evil-visual-state-map (kbd \"C-M-D\") 'evil-multiedit-restore)\n;; RET will toggle the region under the cursor\n(define-key evil-multiedit-state-map (kbd \"RET\") 'evil-multiedit-toggle-or-restrict-region)\n;; ...and in visual mode, RET will disable all fields outside the selected region\n(define-key evil-motion-state-map (kbd \"RET\") 'evil-multiedit-toggle-or-restrict-region)\n;; For moving between edit regions\n(define-key evil-multiedit-state-map (kbd \"C-n\") 'evil-multiedit-next)\n(define-key evil-multiedit-state-map (kbd \"C-p\") 'evil-multiedit-prev)\n(define-key evil-multiedit-insert-state-map (kbd \"C-n\") 'evil-multiedit-next)\n(define-key evil-multiedit-insert-state-map (kbd \"C-p\") 'evil-multiedit-prev)\n;; Ex command that allows you to invoke evil-multiedit with a regular expression, e.g.\n(evil-ex-define-cmd \"ie[dit]\" 'evil-multiedit-ex-match)<\/pre><\/div><p>Once regions are highlighted, changes are mirrored across them all.<\/p><p>Many evil-mode motions\/operators will have slightly different behavior while evil-multiedit is active or the cursor is in an iedit region:<\/p><ul><li><code>D<\/code>: clear the region<\/li>\n<li><code>C<\/code>: clear to end-of-region and go into insert mode<\/li>\n<li><code>A<\/code>: go into insert mode at end-of-region<\/li>\n<li><code>I<\/code>: go into insert mode at start-of-region<\/li>\n<li><code>V<\/code>: select the region<\/li>\n<li><code>$<\/code>: go to end-of-region<\/li>\n<li><code>0<\/code>\/<code>^<\/code>: go to start-of-region<\/li>\n<li><code>gg<\/code>\/<code>G<\/code>: go to the first\/last region<\/li>\n<\/ul><p>To disable these, set <code>evil-multiedit-dwim-motion-keys<\/code> to <code>nil<\/code> before loading evil-multiedit.<\/p><p>NOTE: No need to bind a key for <code>evil-multiedit-abort<\/code>, pressing <kbd>ESC<\/kbd> in normal mode will invoke it.<\/p><h3><a id=\"user-content-ex-command\" class=\"anchor\" aria-hidden=\"true\" href=\"#ex-command\"><\/a>Ex Command<\/h3><p><code>:iedit [REGEXP]<\/code> is available for invoking multiedit from the ex command line.<\/p><h3><a id=\"user-content-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#commands\"><\/a>Commands<\/h3><ul><li><code>evil-multiedit-restore<\/code><\/li>\n<li><code>evil-multiedit-match-all<\/code><\/li>\n<li><code>evil-multiedit-match-and-next<\/code><\/li>\n<li><code>evil-multiedit-match-and-prev<\/code><\/li>\n<li><code>evil-multiedit-match-symbol-and-next<\/code><\/li>\n<li><code>evil-multiedit-match-symbol-and-prev<\/code><\/li>\n<li><code>evil-multiedit-toggle-marker-here<\/code><\/li>\n<li><code>evil-multiedit-toggle-or-restrict-region<\/code><\/li>\n<li><code>evil-multiedit-next<\/code><\/li>\n<li><code>evil-multiedit-prev<\/code><\/li>\n<li><code>evil-multiedit-abort<\/code><\/li>\n<li><code>evil-multiedit-ex-match<\/code><\/li>\n<\/ul><h3><a id=\"user-content-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#options\"><\/a>Options<\/h3><ul><li><code>evil-multiedit-dwim-motion-keys<\/code> (default: <code>t<\/code>): If non-nil, evil's motion keys behave differently when the point is inside a multiedit region. Must be set before evil-multiedit is loaded.<\/li>\n<li><code>evil-multiedit-ignore-indent-and-trailing<\/code> (default: <code>t<\/code>): If non-nil, skip over indentation and trailing whitespace when matching whitespace.<\/li>\n<li><code>evil-multiedit-scope<\/code> (default <code>nil<\/code>): How far evil-multiedit should look for incremental matches (doesn't affect <code>evil-multiedit-match-all<\/code> or <code>evil-multiedit-ex-match<\/code>). Accepts anything that <code>bounds-of-thing-at-point<\/code> accepts, such as <code>'defun<\/code>, <code>'sexp<\/code>, <code>'email<\/code> or the default, <code>'buffer<\/code>.<\/li>\n<li><code>evil-multiedit-smart-match-boundaries<\/code> (default <code>t<\/code>): If non-nil, multiedit will try to be smart about match boundaries when invoked from normal mode. E.g.\n<ul><li>'evil-multiedit-match' will not match 'evil-multiedit-match-all'<\/li>\n<li>'i' will only match 'i' and not every individual i in 'ignition'.<\/li>\n<\/ul><ul><li><strong>NOTE:<\/strong> If evil-multiedit is invoked from visual mode, this is ignored.<\/li>\n<\/ul><\/li>\n<li><code>evil-multiedit-store-in-search-history<\/code> (default <code>nil<\/code>): If non-nil, highlighted occurrences are stored in <code>regexp-search-ring<\/code>, so that after exiting iedit <code>evil-search-next<\/code> and <code>evil-search-previous<\/code> (usually n and N) use the last occurrence as if it were the last string in the search history.<\/li>\n<li><code>evil-multiedit-follow-matches<\/code> (default <code>nil<\/code>): If non-nil, the cursor will jump to each additional match, rather than remain in its original position.<\/li>\n<\/ul><h3><a id=\"user-content-co-existing-with-evil-mc\" class=\"anchor\" aria-hidden=\"true\" href=\"#co-existing-with-evil-mc\"><\/a>Co-existing with evil-mc<\/h3><p>How the two plugins mingle is entirely personal preference. Mine is to bind evil-multiedit to <kbd>M-d<\/kbd>\/<kbd>M-D<\/kbd>, and evil-mc to a bunch of keys prefixed with <kbd>gz<\/kbd>:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; evil-multiedit\n:v  \"R\"     #'evil-multiedit-match-all\n:n  \"M-d\"   #'evil-multiedit-match-symbol-and-next\n:n  \"M-D\"   #'evil-multiedit-match-symbol-and-prev\n:v  \"M-d\"   #'evil-multiedit-match-and-next\n:v  \"M-D\"   #'evil-multiedit-match-and-prev\n:nv \"C-M-d\" #'evil-multiedit-restore\n(:after evil-multiedit\n  (:map evil-multiedit-state-map\n    \"M-d\" #'evil-multiedit-match-and-next\n    \"M-D\" #'evil-multiedit-match-and-prev\n    \"RET\" #'evil-multiedit-toggle-or-restrict-region)\n  (:map (evil-multiedit-state-map evil-multiedit-insert-state-map)\n    \"C-n\" #'evil-multiedit-next\n    \"C-p\" #'evil-multiedit-prev))\n;; evil-mc\n(:prefix \"gz\"\n  :nv \"m\" #'evil-mc-make-all-cursors\n  :nv \"u\" #'evil-mc-undo-all-cursors\n  :nv \"z\" #'+evil\/mc-toggle-cursors\n  :nv \"c\" #'+evil\/mc-make-cursor-here\n  :nv \"n\" #'evil-mc-make-and-goto-next-cursor\n  :nv \"p\" #'evil-mc-make-and-goto-prev-cursor\n  :nv \"N\" #'evil-mc-make-and-goto-last-cursor\n  :nv \"P\" #'evil-mc-make-and-goto-first-cursor)\n(:after evil-mc\n  :map evil-mc-key-map\n  :nv \"C-n\" #'evil-mc-make-and-goto-next-cursor\n  :nv \"C-N\" #'evil-mc-make-and-goto-last-cursor\n  :nv \"C-p\" #'evil-mc-make-and-goto-prev-cursor\n  :nv \"C-P\" #'evil-mc-make-and-goto-first-cursor)<\/pre><\/div><p>You can find <a href=\"https:\/\/github.com\/hlissner\/.emacs.d\/blob\/master\/modules\/feature\/evil\/config.el#L285\">my evil-mc\/evil-multiedit config in my emacs.d<\/a>.<\/p>","created_at":"2020-01-13T20:52:45+0000","updated_at":"2020-01-13T20:52:50+0000","published_by":["hlissner"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/510883?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["python","packaging"],"is_public":false,"id":1107,"title":"indygreg\/PyOxidizer","url":"https:\/\/github.com\/indygreg\/PyOxidizer","content":"<p><a href=\"https:\/\/dev.azure.com\/gregoryszorc\/PyOxidizer\/_build\/latest?definitionId=1&amp;branchName=main\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/4c44f79390f487051bce9f391ffe91bc2d2058b7\/68747470733a2f2f6465762e617a7572652e636f6d2f677265676f7279737a6f72632f50794f786964697a65722f5f617069732f6275696c642f7374617475732f696e6479677265672e50794f786964697a65723f6272616e63684e616d653d6d61696e\" alt=\"Build Status\" data-canonical-src=\"https:\/\/dev.azure.com\/gregoryszorc\/PyOxidizer\/_apis\/build\/status\/indygreg.PyOxidizer?branchName=main\" class=\"c3\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><code>PyOxidizer<\/code> is a utility for producing binaries that embed Python. The over-arching goal of <code>PyOxidizer<\/code> is to make complex packaging and distribution problems simple so application maintainers can focus on building applications instead of toiling with build systems and packaging tools.<\/p><p><code>PyOxidizer<\/code> is capable of producing a single file executable - with a copy of Python and all its dependencies statically linked and all resources (like <code>.pyc<\/code> files) embedded in the executable. You can copy a single executable file to another machine and run a Python application contained within. It <em>just works<\/em>.<\/p><p><code>PyOxidizer<\/code> exposes its lower level functionality for embedding self-contained Python interpreters as a tool and software library. So if you don't want to ship executables that only consist of a Python application, you can still use <code>PyOxidizer<\/code> to e.g. produce a library containing Python suitable for linking in any application or use <code>PyOxidizer<\/code>'s embedding library directly for embedding Python in a larger application.<\/p><p>The <em>Oxidizer<\/em> part of the name comes from Rust: executables produced by <code>PyOxidizer<\/code> are compiled from Rust and Rust code is responsible for managing the embedded Python interpreter and all its operations. If you don't know Rust, that's OK: PyOxidizer tries to make the existence of Rust nearly invisible to end-users.<\/p><p>While solving packaging and distribution problems is the primary goal of <code>PyOxidizer<\/code>, a side-effect of solving that problem with Rust is that <code>PyOxidizer<\/code> can serve as a bridge between these two languages. <code>PyOxidizer<\/code> can be used to easily add a Python interpreter to <em>any<\/em> Rust project. But the opposite is also true: <code>PyOxidizer<\/code> can also be used to add Rust to Python. Using <code>PyOxidizer<\/code>, you could <em>bootstrap<\/em> a new Rust project which contains an embedded version of Python and your application. Initially, your project is a few lines of Rust that instantiates a Python interpreter and runs Python code. Over time, functionality could be (re)written in Rust and your previously Python-only project could leverage Rust and its diverse ecosystem. Since <code>PyOxidizer<\/code> abstracts the Python interpreter away, this could all be invisible to end-users: you could rewrite an application from Python to Rust and people may not even know because they never see a <code>libpython<\/code>, <code>.py<\/code> files, etc.<\/p><p>The official home of the <code>PyOxidizer<\/code> project is <a href=\"https:\/\/github.com\/indygreg\/PyOxidizer\">https:\/\/github.com\/indygreg\/PyOxidizer<\/a>.<\/p><p>Documentation (generated from the <code>docs\/<\/code> directory) is available at <a href=\"https:\/\/pyoxidizer.readthedocs.io\/en\/latest\/index.html\" rel=\"nofollow\">https:\/\/pyoxidizer.readthedocs.io\/en\/latest\/index.html<\/a>.<\/p><p>The <a href=\"https:\/\/groups.google.com\/forum\/#!forum\/pyoxidizer-users\" rel=\"nofollow\">pyoxidizer-users<\/a> mailing list is a forum for users to discuss all things PyOxidizer.<\/p><p>If you want to financially contribute to PyOxidizer, do so <a href=\"https:\/\/www.patreon.com\/indygreg\" rel=\"nofollow\">on Patreon<\/a>.<\/p>","created_at":"2020-01-13T19:59:02+0000","updated_at":"2020-01-13T19:59:06+0000","published_by":["indygreg"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/342993?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["ski","gear","goggles"],"is_public":false,"id":1106,"title":"Electric EG3 Goggles","url":"https:\/\/theskimonster.com\/products\/electric-eg3\/","content":"<div class=\"content content1\"><h4>ALL SEALED UP.<\/h4><p>A truly sexy looking goggle with arguably the best interchanging lens system on the market. \u00a0 The rimless look resembles the Oakley Flight Deck, except the flight deck has plastic out riggers, the EG3 has no outriggers making it look way more stealth mode than anything else out there. \u00a0The interchanging lens sytem is easier to use than the Anon system and is more secure than the magnets. \u00a0It works in a press-seal fahsion that is just as easy as sealing a zip-lock bag. \u00a0<\/p><p>The lens technology is out of control with an A\/R (Anti-Reflective)\u00a0coating on the inner lens, make\u00a0bright high altitude days much easier on the eyes. \u00a0Less squinting means more comfort and fewer headaches. \u00a0Thanks Electric!<\/p><h4>Fit\/Foam<\/h4><p>Triple Layer Face Foam is what makes higher end goggles fit and feel better -- they don't have the itchy feel that entry level or mid-level goggles have. \u00a0Nike also added a thermafit foam so it has better contact with your face. \u00a0When goggles fit better they fog less.<\/p><h4>Lens<\/h4><p><em>Press Seal System-<\/em>\u00a0The EG3's come with two lenses and they are really easy to swap out, among the easiest in the lens swapping game. \u00a0If you know how to use a zip-lock bag you'll know how to swap the lens on the EG3 - very similar system. \u00a0<\/p><p><em>Dual Lens -<\/em> The EG3 features an anti-fog spherical dual lens system that maintains inside and outside temperature difference keep the you skiing fog free. The lens is coated with an anti-scratch coating to stand up to whatever you can dish out.<\/p><h4>Strap<\/h4><p><em>Dual Adjustable -<\/em> The strap on this goggle will fit any size head or helmet.<\/p><p><em>Silicone Ribbing -<\/em> You will find a tacky lining of silicone on the strap that keeps your goggles in place on your helmet or hat.<\/p><h4>Additional Awesomeness<\/h4><p><em>Thermoplastic Urethane<\/em> - When it gets cold materials get brittle, Electric created a lens that remains flexible at cold temperatures. \u00a0So if it's freezing cold and you're shredding the trees and get a branch across the face, your Electric goggles will stand a better chance of surviving over the competition.\u00a0<\/p><\/div><div class=\"content content2 hidden\"><p>\u2022 \u00a0<strong>Gender<\/strong>: \u00a0Unisex<\/p><p>\u2022 \u00a0<strong>Best Fit<\/strong>: \u00a0Large Faces or just looking for a big looking goggle<\/p><p>\u2022 \u00a0<strong>Additional Lens<\/strong>: \u00a0Includes 2 Spherical\u00a0Dual Lenses<\/p><p>\u2022 \u00a0<strong>Anti-Fog<\/strong>: \u00a0Yes and Anti-Scratch<\/p><p>\u2022 \u00a0<strong>Polarized<\/strong>: \u00a0No (Polarized lenses available)<\/p><p>\u2022 \u00a0<strong>Goggle Bag<\/strong>: \u00a0Suede\u00a0Goggle Bag with Lens Sleeve<\/p><p>\u2022 \u00a0<strong>Helmet Compatible<\/strong>: \u00a0Yes\u00a0<\/p><p>\u2022 \u00a0<strong>Warranty<\/strong>: \u00a01 Year<\/p><p>\u2022 \u00a0<strong>Model Year<\/strong>: \u00a02019<\/p><\/div>","created_at":"2020-01-13T17:41:38+0000","updated_at":"2020-01-13T17:41:43+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"theskimonster.com","preview_picture":"https:\/\/theskimonster.com\/files\/theskimonster\/gallery\/files\/W3rAijMpoLU6qvle.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","org-mode","elisp"],"is_public":false,"id":1105,"title":"ndwarshuis\/om.el","url":"https:\/\/github.com\/ndwarshuis\/om.el","content":"<p>A functional API for org-mode inspired by <a href=\"https:\/\/github.com\/magnars\">@magnars<\/a>'s <a href=\"https:\/\/github.com\/magnars\/dash.el\">dash.el<\/a> and <a href=\"https:\/\/github.com\/magnars\/s.el\">s.el<\/a> libraries.<\/p><p>This package is not yet in MELPA. To install, clone this repository somewhere into your load path:<\/p><pre>git clone https:\/\/github.com\/ndwarshuis\/om.el\n<\/pre><p>Then require in your emacs config:<\/p><pre>(require 'om)\n<\/pre><p>Org-mode comes with a powerful, built-in parse-tree generator specified in <code>org-element.el<\/code>. The generated parse-tree is simply a heavily-nested list which can be easily manipulated using (mostly pure) functional code. This contrasts the majority of functions normally used to interface with org-mode files, which are imperative in nature (<code>org-insert-headine<\/code>, <code>outline-next-heading<\/code>, etc) as they depend on the mutable state of Emacs buffers. In general, functional code is (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Functional_programming#Comparison_to_imperative_programming\" rel=\"nofollow\">arguably<\/a>) more robust, readable, and testable, especially in use-cases such as this where a stateless abstract data structure is being transformed and queried.<\/p><p>The <code>org-element.el<\/code> provides a minimal API for handling this parse-tree in a functional manner, but lacks many of the higher-level functions necessary for intuitive, large-scale use. The <code>om.el<\/code> package is designed to provide this API. Furthermore, it is highly compatible with the <code>dash.el<\/code> package, which is a generalized functional library for emacs-lisp.<\/p><p>Parsing a buffer with the function <code>org-element-parse-buffer<\/code> will yield a parse tree composed of nodes. Nodes have types and properties associated with them. See <a href=\"https:\/\/orgmode.org\/worg\/dev\/org-element-api.html#attributes\" rel=\"nofollow\">the org-element API documentation<\/a> for a list of all node types and their properties (also see the <a href=\"#terminology\">terminology conventions<\/a> and <a href=\"#properties\">property omissions<\/a> used in this package).<\/p><p>Each node is represented by a list where the first member is the type and the second member is a plist describing the node's properties:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(type (:prop1 value1 :prop2 value2 ...))<\/pre><\/div><p>Node types may be either leaves or branches, where branches may have zero or more child nodes and leaves may not have child nodes at all. Leaves will always have lists of the form shown above. Branches, on the other hand, have their children appended to the end:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(type (:prop1 value1 :prop2 value2) child1 child2 ...)<\/pre><\/div><p>In addition to leaves and branches, node types can belong to one of two classes:<\/p><ul><li>Objects: roughly correspond to raw, possibly-formatted text<\/li>\n<li>Elements: more complex structures which may be built from objects<\/li>\n<\/ul><p>Within the branch node types, there are restrictions of which class is allowed to be a child depending on the type. There are three of these restrictions:<\/p><ul><li>Branch element with child elements (aka 'greater elements'): these are element types that are generally nestable inside one another (eg headlines, plain-lists, items)<\/li>\n<li>Branch elements with child objects (aka 'object containers'): these are element types that hold textual information (eg paragraph)<\/li>\n<li>Branch objects with child objects (aka 'recursive objects'): these are object types used primarily for text formating (bold, italic, underline, etc)<\/li>\n<\/ul><p>Note: it is never allowed for an element type to be a child of a branch object type.<\/p><h2><a id=\"user-content-terminology\" class=\"anchor\" aria-hidden=\"true\" href=\"#terminology\"><\/a>Terminology<\/h2><p>This package takes several deviations from the original terminology found in <code>org-element.el<\/code>. Generally, the names used here conform to terminology more often used with tree data structures:<\/p><ul><li>'node' is used here to describe a vertex in the parse tree, where 'element' and 'object' are two classes used to describe said vertex (<code>org-element.el<\/code> seems to use 'element' to generally mean 'node' and uses 'object' to further specify)<\/li>\n<li>'child' and 'children' are used here instead of 'content' and 'contents'<\/li>\n<li>'branch' is used here instead of 'container'. Furthermore, 'leaf' is used to describe the converse of 'branch' (there does not seem to be an equivalent term in <code>org-element.el<\/code>)<\/li>\n<li><code>org-element.el<\/code> uses 'attribute(s)' and 'property(ies)' interchangably to describe nodes; here only 'property(ies)' is used<\/li>\n<\/ul><h2><a id=\"user-content-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#properties\"><\/a>Properties<\/h2><p>The properties <code>:begin<\/code>, <code>:end<\/code>, <code>:contents-begin<\/code>, <code>:contents-end<\/code>, and <code>post-affiliated<\/code> are not exposed by this API. They are not necessary for manipulating the functional representation of the parse tree. In addition to these, some properties unique to certain types are not exposed for the same reason. Each type's build function describes the properties that are available.<\/p><h2><a id=\"user-content-threading\" class=\"anchor\" aria-hidden=\"true\" href=\"#threading\"><\/a>Threading<\/h2><p>Each function that operates on an element\/object will take the element\/object as its right-most argument. This allows convenient function chaining using <code>dash.el<\/code>'s right-threading operators (<code>-&gt;&gt;<\/code> and <code>-some-&gt;&gt;<\/code>). The examples below almost exclusively demonstrate this pattern. Additionally, the right-argument convention also allows convenient partial application using <code>-partial<\/code> from <code>dash.el<\/code>.<\/p><h2><a id=\"user-content-higher-order-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#higher-order-functions\"><\/a>Higher-order functions<\/h2><p>Higher-order functions (functions that take other functions as arguments) have two forms. The first takes a (usually unary) function and applies it:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(om-map-property :value (lambda (s) (concat \"foo\" s)) node)\n(om-map-property :value (-partial concat \"foo\") node)<\/pre><\/div><p>This can equivalently be written using an anaphoric form where the original function name is appended with <code>*<\/code>. The symbol <code>it<\/code> carries the value of the unary argument (unless otherwise specified):<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(om-map-property* :value (concat \"foo\" it) node)<\/pre><\/div><h2><a id=\"user-content-side-effect-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#side-effect-functions\"><\/a>Side effect functions<\/h2><p>All functions that read and write from buffers are named like <code>om-OPERATION-THING-at<\/code> where <code>OPERATION<\/code> is some operation to be performed on <code>THING<\/code> in the current buffer. All these functions take <code>point<\/code> as one of their arguments to denote where in the buffer to perform <code>OPERATION<\/code>.<\/p><p>All of these functions have current-point convenience analogues that are named as <code>om-OPERATION-this-THING<\/code> where <code>OPERATION<\/code> and <code>THING<\/code> carry the same meaning, but <code>OPERATION<\/code> is done at the current point and <code>point<\/code> is not an argument to the function.<\/p><p>For the sake of brevity, only the former form of these functions are given in the examples below.<\/p><h2><a id=\"user-content-string-conversion\" class=\"anchor\" aria-hidden=\"true\" href=\"#string-conversion\"><\/a>String Conversion<\/h2><p>Convert nodes to strings.<\/p><ul><li><a href=\"#om-to-string-node\">om-to-string<\/a> <code>(node)<\/code><\/li>\n<li><a href=\"#om-to-trimmed-string-node\">om-to-trimmed-string<\/a> <code>(node)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-buffer-parsing\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer-parsing\"><\/a>Buffer Parsing<\/h2><p>Parse buffers to trees.<\/p><ul><li><a href=\"#om-parse-object-at-point\">om-parse-object-at<\/a> <code>(point)<\/code><\/li>\n<li><a href=\"#om-parse-element-at-point\">om-parse-element-at<\/a> <code>(point)<\/code><\/li>\n<li><a href=\"#om-parse-table-row-at-point\">om-parse-table-row-at<\/a> <code>(point)<\/code><\/li>\n<li><a href=\"#om-parse-headline-at-point\">om-parse-headline-at<\/a> <code>(point)<\/code><\/li>\n<li><a href=\"#om-parse-subtree-at-point\">om-parse-subtree-at<\/a> <code>(point)<\/code><\/li>\n<li><a href=\"#om-parse-item-at-point\">om-parse-item-at<\/a> <code>(point)<\/code><\/li>\n<li><a href=\"#om-parse-section-at-point\">om-parse-section-at<\/a> <code>(point)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-building\" class=\"anchor\" aria-hidden=\"true\" href=\"#building\"><\/a>Building<\/h2><p>Build new nodes.<\/p><h3><a id=\"user-content-leaf-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#leaf-object-nodes\"><\/a>Leaf Object Nodes<\/h3><ul><li><a href=\"#om-build-code-value-key-post-blank\">om-build-code<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-entity-name-key-use-brackets-p-post-blank\">om-build-entity<\/a> <code>(name &amp;key use-brackets-p post-blank)<\/code><\/li>\n<li><a href=\"#om-build-export-snippet-back-end-value-key-post-blank\">om-build-export-snippet<\/a> <code>(back-end value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-inline-babel-call-call-key-inside-header-arguments-end-header-post-blank\">om-build-inline-babel-call<\/a> <code>(call &amp;key inside-header arguments end-header post-blank)<\/code><\/li>\n<li><a href=\"#om-build-inline-src-block-language-key-parameters-value--post-blank\">om-build-inline-src-block<\/a> <code>(language &amp;key parameters (value \"\") post-blank)<\/code><\/li>\n<li><a href=\"#om-build-line-break-key-post-blank\">om-build-line-break<\/a> <code>(&amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-latex-fragment-value-key-post-blank\">om-build-latex-fragment<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-macro-key-key-args-post-blank\">om-build-macro<\/a> <code>(key &amp;key args post-blank)<\/code><\/li>\n<li><a href=\"#om-build-statistics-cookie-value-key-post-blank\">om-build-statistics-cookie<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-target-value-key-post-blank\">om-build-target<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-timestamp-type-year-start-month-start-day-start-year-end-month-end-day-end-key-hour-start-minute-start-hour-end-minute-end-repeater-type-repeater-unit-repeater-value-warning-type-warning-unit-warning-value-post-blank\">om-build-timestamp<\/a> <code>(type year-start month-start day-start year-end month-end day-end &amp;key hour-start minute-start hour-end minute-end repeater-type repeater-unit repeater-value warning-type warning-unit warning-value post-blank)<\/code><\/li>\n<li><a href=\"#om-build-verbatim-value-key-post-blank\">om-build-verbatim<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-branch-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#branch-object-nodes\"><\/a>Branch Object Nodes<\/h3><ul><li><a href=\"#om-build-bold-key-post-blank-rest-object-nodes\">om-build-bold<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-footnote-reference-key-label-post-blank-rest-object-nodes\">om-build-footnote-reference<\/a> <code>(&amp;key label post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-italic-key-post-blank-rest-object-nodes\">om-build-italic<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-link-path-key-format-type-fuzzy-post-blank-rest-object-nodes\">om-build-link<\/a> <code>(path &amp;key format (type \"fuzzy\") post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-radio-target-key-post-blank-rest-object-nodes\">om-build-radio-target<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-strike-through-key-post-blank-rest-object-nodes\">om-build-strike-through<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-superscript-key-use-brackets-p-post-blank-rest-object-nodes\">om-build-superscript<\/a> <code>(&amp;key use-brackets-p post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-subscript-key-use-brackets-p-post-blank-rest-object-nodes\">om-build-subscript<\/a> <code>(&amp;key use-brackets-p post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-table-cell-key-post-blank-rest-object-nodes\">om-build-table-cell<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-underline-key-post-blank-rest-object-nodes\">om-build-underline<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-leaf-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#leaf-element-nodes\"><\/a>Leaf Element Nodes<\/h3><ul><li><a href=\"#om-build-babel-call-call-key-inside-header-arguments-end-header-post-blank\">om-build-babel-call<\/a> <code>(call &amp;key inside-header arguments end-header post-blank)<\/code><\/li>\n<li><a href=\"#om-build-clock-value-key-post-blank\">om-build-clock<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-comment-value-key-post-blank\">om-build-comment<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-comment-block-key-value--post-blank\">om-build-comment-block<\/a> <code>(&amp;key (value \"\") post-blank)<\/code><\/li>\n<li><a href=\"#om-build-diary-sexp-key-value-post-blank\">om-build-diary-sexp<\/a> <code>(&amp;key value post-blank)<\/code><\/li>\n<li><a href=\"#om-build-example-block-key-preserve-indent-switches-value--post-blank\">om-build-example-block<\/a> <code>(&amp;key preserve-indent switches (value \"\") post-blank)<\/code><\/li>\n<li><a href=\"#om-build-export-block-type-value-key-post-blank\">om-build-export-block<\/a> <code>(type value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-fixed-width-value-key-post-blank\">om-build-fixed-width<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-horizontal-rule-key-post-blank\">om-build-horizontal-rule<\/a> <code>(&amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-keyword-key-value-key-post-blank\">om-build-keyword<\/a> <code>(key value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-latex-environment-value-key-post-blank\">om-build-latex-environment<\/a> <code>(value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-node-property-key-value-key-post-blank\">om-build-node-property<\/a> <code>(key value &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-planning-key-closed-deadline-scheduled-post-blank\">om-build-planning<\/a> <code>(&amp;key closed deadline scheduled post-blank)<\/code><\/li>\n<li><a href=\"#om-build-src-block-key-value--language-parameters-preserve-indent-switches-post-blank\">om-build-src-block<\/a> <code>(&amp;key (value \"\") language parameters preserve-indent switches post-blank)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-branch-element-nodes-with-child-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#branch-element-nodes-with-child-object-nodes\"><\/a>Branch Element Nodes with Child Object Nodes<\/h3><ul><li><a href=\"#om-build-paragraph-key-post-blank-rest-object-nodes\">om-build-paragraph<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-table-row-key-post-blank-rest-object-nodes\">om-build-table-row<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<li><a href=\"#om-build-verse-block-key-post-blank-rest-object-nodes\">om-build-verse-block<\/a> <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-branch-element-nodes-with-child-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#branch-element-nodes-with-child-element-nodes\"><\/a>Branch Element Nodes with Child Element Nodes<\/h3><ul><li><a href=\"#om-build-center-block-key-post-blank-rest-element-nodes\">om-build-center-block<\/a> <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-drawer-drawer-name-key-post-blank-rest-element-nodes\">om-build-drawer<\/a> <code>(drawer-name &amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-dynamic-block-block-name-key-arguments-post-blank-rest-element-nodes\">om-build-dynamic-block<\/a> <code>(block-name &amp;key arguments post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-footnote-definition-label-key-post-blank-rest-element-nodes\">om-build-footnote-definition<\/a> <code>(label &amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-headline-key-archivedp-commentedp-footnote-section-p-level-1-pre-blank-0-priority-tags-title-todo-keyword-post-blank-rest-element-nodes\">om-build-headline<\/a> <code>(&amp;key archivedp commentedp footnote-section-p (level 1) (pre-blank 0) priority tags title todo-keyword post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-item-key-bullet-quote---checkbox-counter-tag-post-blank-rest-element-nodes\">om-build-item<\/a> <code>(&amp;key (bullet '-) checkbox counter tag post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-plain-list-key-post-blank-rest-element-nodes\">om-build-plain-list<\/a> <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-property-drawer-key-post-blank-rest-element-nodes\">om-build-property-drawer<\/a> <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-quote-block-key-post-blank-rest-element-nodes\">om-build-quote-block<\/a> <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-section-key-post-blank-rest-element-nodes\">om-build-section<\/a> <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-special-block-type-key-post-blank-rest-element-nodes\">om-build-special-block<\/a> <code>(type &amp;key post-blank &amp;rest element-nodes)<\/code><\/li>\n<li><a href=\"#om-build-table-key-tblfm-post-blank-rest-element-nodes\">om-build-table<\/a> <code>(&amp;key tblfm post-blank &amp;rest element-nodes)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-miscellaneous-builders\" class=\"anchor\" aria-hidden=\"true\" href=\"#miscellaneous-builders\"><\/a>Miscellaneous Builders<\/h3><ul><li><a href=\"#om-build-secondary-string-string\">om-build-secondary-string!<\/a> <code>(string)<\/code><\/li>\n<li><a href=\"#om-build-table-row-hline-key-post-blank\">om-build-table-row-hline<\/a> <code>(&amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-timestamp-diary-form-key-post-blank\">om-build-timestamp-diary<\/a> <code>(form &amp;key post-blank)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-shorthand-builders\" class=\"anchor\" aria-hidden=\"true\" href=\"#shorthand-builders\"><\/a>Shorthand Builders<\/h3><p>Build nodes with more convenient\/shorter syntax.<\/p><ul><li><a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\">om-build-timestamp!<\/a> <code>(start &amp;key end active repeater warning post-blank)<\/code><\/li>\n<li><a href=\"#om-build-clock-start-key-end-post-blank\">om-build-clock!<\/a> <code>(start &amp;key end post-blank)<\/code><\/li>\n<li><a href=\"#om-build-planning-key-closed-deadline-scheduled-post-blank\">om-build-planning!<\/a> <code>(&amp;key closed deadline scheduled post-blank)<\/code><\/li>\n<li><a href=\"#om-build-property-drawer-key-post-blank-rest-keyvals\">om-build-property-drawer!<\/a> <code>(&amp;key post-blank &amp;rest keyvals)<\/code><\/li>\n<li><a href=\"#om-build-headline-key-level-1-title-text-todo-keyword-tags-pre-blank-priority-commentedp-archivedp-post-blank-planning-statistics-cookie-section-children-rest-subheadlines\">om-build-headline!<\/a> <code>(&amp;key (level 1) title-text todo-keyword tags pre-blank priority commentedp archivedp post-blank planning statistics-cookie section-children &amp;rest subheadlines)<\/code><\/li>\n<li><a href=\"#om-build-item-key-post-blank-bullet-checkbox-tag-paragraph-counter-rest-children\">om-build-item!<\/a> <code>(&amp;key post-blank bullet checkbox tag paragraph counter &amp;rest children)<\/code><\/li>\n<li><a href=\"#om-build-paragraph-string-key-post-blank\">om-build-paragraph!<\/a> <code>(string &amp;key post-blank)<\/code><\/li>\n<li><a href=\"#om-build-table-cell-string\">om-build-table-cell!<\/a> <code>(string)<\/code><\/li>\n<li><a href=\"#om-build-table-row-row-list\">om-build-table-row!<\/a> <code>(row-list)<\/code><\/li>\n<li><a href=\"#om-build-table-key-tblfm-post-blank-rest-row-lists\">om-build-table!<\/a> <code>(&amp;key tblfm post-blank &amp;rest row-lists)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-type-predicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#type-predicates\"><\/a>Type Predicates<\/h2><p>Test node types.<\/p><ul><li><a href=\"#om-get-type-node\">om-get-type<\/a> <code>(node)<\/code><\/li>\n<li><a href=\"#om-is-type-type-node\">om-is-type<\/a> <code>(type node)<\/code><\/li>\n<li><a href=\"#om-is-any-type-types-node\">om-is-any-type<\/a> <code>(types node)<\/code><\/li>\n<li><a href=\"#om-is-element-node\">om-is-element<\/a> <code>(node)<\/code><\/li>\n<li><a href=\"#om-is-branch-node-node\">om-is-branch-node<\/a> <code>(node)<\/code><\/li>\n<li><a href=\"#om-node-may-have-child-objects-node\">om-node-may-have-child-objects<\/a> <code>(node)<\/code><\/li>\n<li><a href=\"#om-node-may-have-child-elements-node\">om-node-may-have-child-elements<\/a> <code>(node)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-property-manipulation\" class=\"anchor\" aria-hidden=\"true\" href=\"#property-manipulation\"><\/a>Property Manipulation<\/h2><p>Set, get, and map properties of nodes.<\/p><h3><a id=\"user-content-generic\" class=\"anchor\" aria-hidden=\"true\" href=\"#generic\"><\/a>Generic<\/h3><ul><li><a href=\"#om-contains-point-p-point-node\">om-contains-point-p<\/a> <code>(point node)<\/code><\/li>\n<li><a href=\"#om-set-property-prop-value-node\">om-set-property<\/a> <code>(prop value node)<\/code><\/li>\n<li><a href=\"#om-set-properties-plist-node\">om-set-properties<\/a> <code>(plist node)<\/code><\/li>\n<li><a href=\"#om-get-property-prop-node\">om-get-property<\/a> <code>(prop node)<\/code><\/li>\n<li><a href=\"#om-map-property-prop-fun-node\">om-map-property<\/a> <code>(prop fun node)<\/code><\/li>\n<li><a href=\"#om-map-properties-plist-node\">om-map-properties<\/a> <code>(plist node)<\/code><\/li>\n<li><a href=\"#om-toggle-property-prop-node\">om-toggle-property<\/a> <code>(prop node)<\/code><\/li>\n<li><a href=\"#om-shift-property-prop-n-node\">om-shift-property<\/a> <code>(prop n node)<\/code><\/li>\n<li><a href=\"#om-insert-into-property-prop-index-string-node\">om-insert-into-property<\/a> <code>(prop index string node)<\/code><\/li>\n<li><a href=\"#om-remove-from-property-prop-string-node\">om-remove-from-property<\/a> <code>(prop string node)<\/code><\/li>\n<li><a href=\"#om-plist-put-property-prop-key-value-node\">om-plist-put-property<\/a> <code>(prop key value node)<\/code><\/li>\n<li><a href=\"#om-plist-remove-property-prop-key-node\">om-plist-remove-property<\/a> <code>(prop key node)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-clock\" class=\"anchor\" aria-hidden=\"true\" href=\"#clock\"><\/a>Clock<\/h3><ul><li><a href=\"#om-clock-is-running-clock\">om-clock-is-running<\/a> <code>(clock)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-entity\" class=\"anchor\" aria-hidden=\"true\" href=\"#entity\"><\/a>Entity<\/h3><ul><li><a href=\"#om-entity-get-replacement-key-entity\">om-entity-get-replacement<\/a> <code>(key entity)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#headline\"><\/a>Headline<\/h3><ul><li><a href=\"#om-headline-set-title-title-text-stats-cookie-value-headline\">om-headline-set-title!<\/a> <code>(title-text stats-cookie-value headline)<\/code><\/li>\n<li><a href=\"#om-headline-is-done-headline\">om-headline-is-done<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-has-tag-tag-headline\">om-headline-has-tag<\/a> <code>(tag headline)<\/code><\/li>\n<li><a href=\"#om-headline-get-statistics-cookie-headline\">om-headline-get-statistics-cookie<\/a> <code>(headline)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-item\" class=\"anchor\" aria-hidden=\"true\" href=\"#item\"><\/a>Item<\/h3><ul><li><a href=\"#om-item-toggle-checkbox-item\">om-item-toggle-checkbox<\/a> <code>(item)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-planning\" class=\"anchor\" aria-hidden=\"true\" href=\"#planning\"><\/a>Planning<\/h3><ul><li><a href=\"#om-planning-set-timestamp-prop-planning-list-planning\">om-planning-set-timestamp!<\/a> <code>(prop planning-list planning)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-statistics-cookie\" class=\"anchor\" aria-hidden=\"true\" href=\"#statistics-cookie\"><\/a>Statistics Cookie<\/h3><ul><li><a href=\"#om-statistics-cookie-is-complete-statistics-cookie\">om-statistics-cookie-is-complete<\/a> <code>(statistics-cookie)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-timestamp-auxiliary\" class=\"anchor\" aria-hidden=\"true\" href=\"#timestamp-auxiliary\"><\/a>Timestamp (Auxiliary)<\/h3><p>Functions to work with timestamp data<\/p><ul><li><a href=\"#om-time-to-unixtime-time\">om-time-to-unixtime<\/a> <code>(time)<\/code><\/li>\n<li><a href=\"#om-unixtime-to-time-long-unixtime\">om-unixtime-to-time-long<\/a> <code>(unixtime)<\/code><\/li>\n<li><a href=\"#om-unixtime-to-time-short-unixtime\">om-unixtime-to-time-short<\/a> <code>(unixtime)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-timestamp-standard\" class=\"anchor\" aria-hidden=\"true\" href=\"#timestamp-standard\"><\/a>Timestamp (Standard)<\/h3><ul><li><a href=\"#om-timestamp-get-start-time-timestamp\">om-timestamp-get-start-time<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-get-end-time-timestamp\">om-timestamp-get-end-time<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-get-range-timestamp\">om-timestamp-get-range<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-is-active-timestamp\">om-timestamp-is-active<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-is-ranged-timestamp\">om-timestamp-is-ranged<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-range-contains-p-unixtime-timestamp\">om-timestamp-range-contains-p<\/a> <code>(unixtime timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-collapsed-flag-timestamp\">om-timestamp-set-collapsed<\/a> <code>(flag timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-start-time-time-timestamp\">om-timestamp-set-start-time<\/a> <code>(time timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-end-time-time-timestamp\">om-timestamp-set-end-time<\/a> <code>(time timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-single-time-time-timestamp\">om-timestamp-set-single-time<\/a> <code>(time timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-double-time-time1-time2-timestamp\">om-timestamp-set-double-time<\/a> <code>(time1 time2 timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-range-range-timestamp\">om-timestamp-set-range<\/a> <code>(range timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-set-active-flag-timestamp\">om-timestamp-set-active<\/a> <code>(flag timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-shift-n-unit-timestamp\">om-timestamp-shift<\/a> <code>(n unit timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-shift-start-n-unit-timestamp\">om-timestamp-shift-start<\/a> <code>(n unit timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-shift-end-n-unit-timestamp\">om-timestamp-shift-end<\/a> <code>(n unit timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-toggle-active-timestamp\">om-timestamp-toggle-active<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-truncate-timestamp\">om-timestamp-truncate<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-truncate-start-timestamp\">om-timestamp-truncate-start<\/a> <code>(timestamp)<\/code><\/li>\n<li><a href=\"#om-timestamp-truncate-end-timestamp\">om-timestamp-truncate-end<\/a> <code>(timestamp)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-timestamp-diary\" class=\"anchor\" aria-hidden=\"true\" href=\"#timestamp-diary\"><\/a>Timestamp (diary)<\/h3><ul><li><a href=\"#om-timestamp-diary-set-value-form-timestamp-diary\">om-timestamp-diary-set-value<\/a> <code>(form timestamp-diary)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-branchchild-manipulation\" class=\"anchor\" aria-hidden=\"true\" href=\"#branchchild-manipulation\"><\/a>Branch\/Child Manipulation<\/h2><p>Set, get, and map the children of branch nodes.<\/p><h3><a id=\"user-content-polymorphic\" class=\"anchor\" aria-hidden=\"true\" href=\"#polymorphic\"><\/a>Polymorphic<\/h3><ul><li><a href=\"#om-children-contain-point-point-branch-node\">om-children-contain-point<\/a> <code>(point branch-node)<\/code><\/li>\n<li><a href=\"#om-get-children-branch-node\">om-get-children<\/a> <code>(branch-node)<\/code><\/li>\n<li><a href=\"#om-set-children-children-branch-node\">om-set-children<\/a> <code>(children branch-node)<\/code><\/li>\n<li><a href=\"#om-map-children-fun-branch-node\">om-map-children<\/a> <code>(fun branch-node)<\/code><\/li>\n<li><a href=\"#om-is-childless-branch-node\">om-is-childless<\/a> <code>(branch-node)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#object-nodes\"><\/a>Object Nodes<\/h3><ul><li><a href=\"#om-unwrap-object-node\">om-unwrap<\/a> <code>(object-node)<\/code><\/li>\n<li><a href=\"#om-unwrap-types-deep-types-object-node\">om-unwrap-types-deep<\/a> <code>(types object-node)<\/code><\/li>\n<li><a href=\"#om-unwrap-deep-object-node\">om-unwrap-deep<\/a> <code>(object-node)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-secondary-strings\" class=\"anchor\" aria-hidden=\"true\" href=\"#secondary-strings\"><\/a>Secondary Strings<\/h3><ul><li><a href=\"#om-flatten-secondary-string\">om-flatten<\/a> <code>(secondary-string)<\/code><\/li>\n<li><a href=\"#om-flatten-types-deep-types-secondary-string\">om-flatten-types-deep<\/a> <code>(types secondary-string)<\/code><\/li>\n<li><a href=\"#om-flatten-deep-secondary-string\">om-flatten-deep<\/a> <code>(secondary-string)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-headline-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#headline-1\"><\/a>Headline<\/h3><ul><li><a href=\"#om-headline-get-node-properties-headline\">om-headline-get-node-properties<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-get-properties-drawer-headline\">om-headline-get-properties-drawer<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-get-planning-headline\">om-headline-get-planning<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-get-subheadlines-headline\">om-headline-get-subheadlines<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-get-section-headline\">om-headline-get-section<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-get-path-headline\">om-headline-get-path<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-update-item-statistics-headline\">om-headline-update-item-statistics<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-update-todo-statistics-headline\">om-headline-update-todo-statistics<\/a> <code>(headline)<\/code><\/li>\n<li><a href=\"#om-headline-indent-subheadline-index-headline\">om-headline-indent-subheadline<\/a> <code>(index headline)<\/code><\/li>\n<li><a href=\"#om-headline-indent-subtree-index-headline\">om-headline-indent-subtree<\/a> <code>(index headline)<\/code><\/li>\n<li><a href=\"#om-headline-unindent-subheadline-index-child-index-headline\">om-headline-unindent-subheadline<\/a> <code>(index child-index headline)<\/code><\/li>\n<li><a href=\"#om-headline-unindent-all-subheadlines-index-headline\">om-headline-unindent-all-subheadlines<\/a> <code>(index headline)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-plain-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#plain-list\"><\/a>Plain List<\/h3><ul><li><a href=\"#om-plain-list-set-type-type-plain-list\">om-plain-list-set-type<\/a> <code>(type plain-list)<\/code><\/li>\n<li><a href=\"#om-plain-list-indent-item-index-plain-list\">om-plain-list-indent-item<\/a> <code>(index plain-list)<\/code><\/li>\n<li><a href=\"#om-plain-list-indent-item-tree-index-plain-list\">om-plain-list-indent-item-tree<\/a> <code>(index plain-list)<\/code><\/li>\n<li><a href=\"#om-plain-list-unindent-item-index-child-index-plain-list\">om-plain-list-unindent-item<\/a> <code>(index child-index plain-list)<\/code><\/li>\n<li><a href=\"#om-plain-list-unindent-all-items-index-plain-list\">om-plain-list-unindent-all-items<\/a> <code>(index plain-list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#table\"><\/a>Table<\/h3><ul><li><a href=\"#om-table-get-cell-row-index-column-index-table\">om-table-get-cell<\/a> <code>(row-index column-index table)<\/code><\/li>\n<li><a href=\"#om-table-delete-column-column-index-table\">om-table-delete-column<\/a> <code>(column-index table)<\/code><\/li>\n<li><a href=\"#om-table-delete-row-row-index-table\">om-table-delete-row<\/a> <code>(row-index table)<\/code><\/li>\n<li><a href=\"#om-table-insert-column-column-index-column-text-table\">om-table-insert-column!<\/a> <code>(column-index column-text table)<\/code><\/li>\n<li><a href=\"#om-table-insert-row-row-index-row-text-table\">om-table-insert-row!<\/a> <code>(row-index row-text table)<\/code><\/li>\n<li><a href=\"#om-table-replace-cell-row-index-column-index-cell-text-table\">om-table-replace-cell!<\/a> <code>(row-index column-index cell-text table)<\/code><\/li>\n<li><a href=\"#om-table-replace-column-column-index-column-text-table\">om-table-replace-column!<\/a> <code>(column-index column-text table)<\/code><\/li>\n<li><a href=\"#om-table-replace-row-row-index-row-text-table\">om-table-replace-row!<\/a> <code>(row-index row-text table)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-node-matching\" class=\"anchor\" aria-hidden=\"true\" href=\"#node-matching\"><\/a>Node Matching<\/h2><p>Use pattern-matching to selectively perform operations on nodes in trees.<\/p><ul><li><a href=\"#om-match-pattern-node\">om-match<\/a> <code>(pattern node)<\/code><\/li>\n<li><a href=\"#om-match-delete-pattern-node\">om-match-delete<\/a> <code>(pattern node)<\/code><\/li>\n<li><a href=\"#om-match-extract-pattern-node\">om-match-extract<\/a> <code>(pattern node)<\/code><\/li>\n<li><a href=\"#om-match-map-pattern-fun-node\">om-match-map<\/a> <code>(pattern fun node)<\/code><\/li>\n<li><a href=\"#om-match-mapcat-pattern-fun-node\">om-match-mapcat<\/a> <code>(pattern fun node)<\/code><\/li>\n<li><a href=\"#om-match-replace-pattern-node-node\">om-match-replace<\/a> <code>(pattern node* node)<\/code><\/li>\n<li><a href=\"#om-match-insert-before-pattern-node-node\">om-match-insert-before<\/a> <code>(pattern node* node)<\/code><\/li>\n<li><a href=\"#om-match-insert-after-pattern-node-node\">om-match-insert-after<\/a> <code>(pattern node* node)<\/code><\/li>\n<li><a href=\"#om-match-insert-within-pattern-index-node-node\">om-match-insert-within<\/a> <code>(pattern index node* node)<\/code><\/li>\n<li><a href=\"#om-match-splice-pattern-nodes-node\">om-match-splice<\/a> <code>(pattern nodes* node)<\/code><\/li>\n<li><a href=\"#om-match-splice-before-pattern-nodes-node\">om-match-splice-before<\/a> <code>(pattern nodes* node)<\/code><\/li>\n<li><a href=\"#om-match-splice-after-pattern-nodes-node\">om-match-splice-after<\/a> <code>(pattern nodes* node)<\/code><\/li>\n<li><a href=\"#om-match-splice-within-pattern-index-nodes-node\">om-match-splice-within<\/a> <code>(pattern index nodes* node)<\/code><\/li>\n<li><a href=\"#om-match-do-pattern-fun-node\">om-match-do<\/a> <code>(pattern fun node)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-buffer-side-effects\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer-side-effects\"><\/a>Buffer Side Effects<\/h2><p>Map node manipulations into buffers.<\/p><h3><a id=\"user-content-insert\" class=\"anchor\" aria-hidden=\"true\" href=\"#insert\"><\/a>Insert<\/h3><ul><li><a href=\"#om-insert-point-node\">om-insert<\/a> <code>(point node)<\/code><\/li>\n<li><a href=\"#om-insert-tail-point-node\">om-insert-tail<\/a> <code>(point node)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-update\" class=\"anchor\" aria-hidden=\"true\" href=\"#update\"><\/a>Update<\/h3><ul><li><a href=\"#om-update-fun-node\">om-update<\/a> <code>(fun node)<\/code><\/li>\n<li><a href=\"#om-update-object-at-point-fun\">om-update-object-at<\/a> <code>(point fun)<\/code><\/li>\n<li><a href=\"#om-update-element-at-point-fun\">om-update-element-at<\/a> <code>(point fun)<\/code><\/li>\n<li><a href=\"#om-update-table-row-at-point-fun\">om-update-table-row-at<\/a> <code>(point fun)<\/code><\/li>\n<li><a href=\"#om-update-item-at-point-fun\">om-update-item-at<\/a> <code>(point fun)<\/code><\/li>\n<li><a href=\"#om-update-headline-at-point-fun\">om-update-headline-at<\/a> <code>(point fun)<\/code><\/li>\n<li><a href=\"#om-update-subtree-at-point-fun\">om-update-subtree-at<\/a> <code>(point fun)<\/code><\/li>\n<li><a href=\"#om-update-section-at-point-fun\">om-update-section-at<\/a> <code>(point fun)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-misc\" class=\"anchor\" aria-hidden=\"true\" href=\"#misc\"><\/a>Misc<\/h3><ul><li><a href=\"#om-fold-node\">om-fold<\/a> <code>(node)<\/code><\/li>\n<li><a href=\"#om-unfold-node\">om-unfold<\/a> <code>(node)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-string-conversion-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#string-conversion-1\"><\/a>String Conversion<\/h2><p>Convert nodes to strings.<\/p><h4><a id=\"user-content-om-to-string-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-to-string-node\"><\/a>om-to-string <code>(node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> as an interpreted string without text properties.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(om-to-string '(bold (:begin 1 :end 5 :parent nil :post-blank 0 :post-affiliated nil)\n                           \"text\"))\n ;; =&gt; \"*text*\"\n(om-to-string '(bold (:begin 1 :end 5 :parent nil :post-blank 3 :post-affiliated nil)\n                           \"text\"))\n ;; =&gt; \"*text*   \"\n(om-to-string nil)\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-to-trimmed-string-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-to-trimmed-string-node\"><\/a>om-to-trimmed-string <code>(node)<\/code><\/h4><p>Like <a href=\"#om-to-string-node\"><code>om-to-string<\/code><\/a> but strip whitespace when returning <strong><code>node<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(om-to-trimmed-string '(bold (:begin 1 :end 5 :parent nil :post-blank 0 :post-affiliated nil)\n                                   \"text\"))\n ;; =&gt; \"*text*\"\n(om-to-trimmed-string '(bold (:begin 1 :end 5 :parent nil :post-blank 3 :post-affiliated nil)\n                                   \"text\"))\n ;; =&gt; \"*text*\"\n(om-to-trimmed-string nil)\n ;; =&gt; \"\"\n<\/pre><\/div><h2><a id=\"user-content-buffer-parsing-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer-parsing-1\"><\/a>Buffer Parsing<\/h2><p>Parse buffers to trees.<\/p><h4><a id=\"user-content-om-parse-object-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-object-at-point\"><\/a>om-parse-object-at <code>(point)<\/code><\/h4><p>Return object node under <strong><code>point<\/code><\/strong> or nil if not on an object.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *text*\n(-&gt;&gt; (om-parse-object-at 1)\n     (car))\n ;; =&gt; 'bold\n;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-object-at 1)\n     (car))\n ;; =&gt; 'timestamp\n;; Given the following contents:\n; - notme\n;; Return nil when parsing an element\n(om-parse-object-at\n 1)\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-parse-element-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-element-at-point\"><\/a>om-parse-element-at <code>(point)<\/code><\/h4><p>Return element node under <strong><code>point<\/code><\/strong> or nil if not on an element.<\/p><p>This function will return every element available in <code>om-elements<\/code> with the exception of <code>section<\/code>, <code>item<\/code>, and <code>table-row<\/code>. To specifically parse these, use the functions <a href=\"#om-parse-section-at-point\"><code>om-parse-section-at<\/code><\/a>, <a href=\"#om-parse-item-at-point\"><code>om-parse-item-at<\/code><\/a>, and <a href=\"#om-parse-table-row-at-point\"><code>om-parse-table-row-at<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu()\n(-&gt;&gt; (om-parse-element-at 1)\n     (car))\n ;; =&gt; 'babel-call\n;; Given the following contents:\n; - plain-list\n;; Give the plain-list, not the item for this function\n(-&gt;&gt; (om-parse-element-at 1)\n     (car))\n ;; =&gt; 'plain-list\n;; Given the following contents:\n; | R | A |\n; | G | E |\n;; Return a table, not the table-row for this function\n(-&gt;&gt; (om-parse-element-at 1)\n     (car))\n ;; =&gt; 'table\n<\/pre><\/div><h4><a id=\"user-content-om-parse-table-row-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-table-row-at-point\"><\/a>om-parse-table-row-at <code>(point)<\/code><\/h4><p>Return table-row node under <strong><code>point<\/code><\/strong> or nil if not on a table-row.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | bow | stroke |\n;; Return the row itself\n(-&gt;&gt; (om-parse-table-row-at 1)\n     (car))\n ;; =&gt; 'table-row\n;; Also return the row when not at beginning of line\n(-&gt;&gt; (om-parse-table-row-at 5)\n     (car))\n ;; =&gt; 'table-row\n;; Given the following contents:\n; - bow and arrow choke\n;; Return nil if not a table-row\n(-&gt;&gt; (om-parse-table-row-at 1)\n     (car))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-parse-headline-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-headline-at-point\"><\/a>om-parse-headline-at <code>(point)<\/code><\/h4><p>Return headline node under <strong><code>point<\/code><\/strong> or nil if not on a headline. <strong><code>point<\/code><\/strong> does not need to be on the headline itself. Only the headline and its section will be returned. To include subheadlines, use <a href=\"#om-parse-subtree-at-point\"><code>om-parse-subtree-at<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline\n;; Return the headline itself\n(-&gt;&gt; (om-parse-headline-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\"\n;; Given the following contents:\n; * headline\n; section crap\n;; Return headline and section\n(-&gt;&gt; (om-parse-headline-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      section crap\"\n;; Return headline when point is in the section\n(-&gt;&gt; (om-parse-headline-at 12)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      section crap\"\n;; Given the following contents:\n; * headline\n; section crap\n; ** not parsed\n;; Don't parse any subheadlines\n(-&gt;&gt; (om-parse-headline-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      section crap\"\n;; Given the following contents:\n; nothing nowhere\n;; Return nil if not under a headline\n(-&gt;&gt; (om-parse-headline-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-parse-subtree-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-subtree-at-point\"><\/a>om-parse-subtree-at <code>(point)<\/code><\/h4><p>Return headline node under <strong><code>point<\/code><\/strong> or nil if not on a headline. <strong><code>point<\/code><\/strong> does not need to be on the headline itself. Unlike <a href=\"#om-parse-headline-at-point\"><code>om-parse-headline-at<\/code><\/a>, the returned node will include child headlines.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline\n;; Return the headline itself\n(-&gt;&gt; (om-parse-subtree-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\"\n;; Given the following contents:\n; * headline\n; section crap\n;; Return headline and section\n(-&gt;&gt; (om-parse-subtree-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      section crap\"\n;; Return headline when point is in the section\n(-&gt;&gt; (om-parse-subtree-at 12)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      section crap\"\n;; Given the following contents:\n; * headline\n; section crap\n; ** parsed\n;; Return all the subheadlines\n(-&gt;&gt; (om-parse-subtree-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      section crap\n ;      ** parsed\"\n;; Given the following contents:\n; nothing nowhere\n;; Return nil if not under a headline\n(-&gt;&gt; (om-parse-subtree-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-parse-item-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-item-at-point\"><\/a>om-parse-item-at <code>(point)<\/code><\/h4><p>Return item node under <strong><code>point<\/code><\/strong> or nil if not on an item. This will return the item node even if <strong><code>point<\/code><\/strong> is not at the beginning of the line.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - item\n;; Return the item itself\n(-&gt;&gt; (om-parse-item-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- item\"\n;; Also return the item when not at beginning of line\n(-&gt;&gt; (om-parse-item-at 5)\n     (om-to-trimmed-string))\n ;; =&gt; \"- item\"\n;; Given the following contents:\n; - item\n;   - item 2\n;; Return item and its subitems\n(-&gt;&gt; (om-parse-item-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- item\n ;        - item 2\"\n;; Given the following contents:\n; * not item\n;; Return nil if not an item\n(-&gt;&gt; (om-parse-item-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-parse-section-at-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-parse-section-at-point\"><\/a>om-parse-section-at <code>(point)<\/code><\/h4><p>Return section node under <strong><code>point<\/code><\/strong> or nil if not on a section. If <strong><code>point<\/code><\/strong> is on or within a headline, return the section under that headline. If <strong><code>point<\/code><\/strong> is before the first headline (if any), return the section at the top of the org buffer.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; over headline\n; * headline\n; under headline\n;; Return the section above the headline\n(-&gt;&gt; (om-parse-section-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"over headline\"\n;; Return the section under headline\n(-&gt;&gt; (om-parse-section-at 25)\n     (om-to-trimmed-string))\n ;; =&gt; \"under headline\"\n;; Given the following contents:\n; * headline\n; ** subheadline\n;; Return nil if no section under headline\n(-&gt;&gt; (om-parse-section-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n;; Given the following contents:\n; \n;; Return nil if no section at all\n(-&gt;&gt; (om-parse-section-at 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h2><a id=\"user-content-building-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-1\"><\/a>Building<\/h2><p>Build new nodes.<\/p><h3><a id=\"user-content-leaf-object-nodes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#leaf-object-nodes-1\"><\/a>Leaf Object Nodes<\/h3><h4><a id=\"user-content-om-build-code-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-code-value-key-post-blank\"><\/a>om-build-code <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a code object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-code \"text\")\n     (om-to-string))\n ;; =&gt; \"~text~\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-entity-name-key-use-brackets-p-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-entity-name-key-use-brackets-p-post-blank\"><\/a>om-build-entity <code>(name &amp;key use-brackets-p post-blank)<\/code><\/h4><p>Build an entity object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>name<\/code><\/strong>: (required) a string that makes <code>org-entity-get<\/code> return non-nil<\/li>\n<li><strong><code>use-brackets-p<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-entity \"gamma\")\n     (om-to-string))\n ;; =&gt; \"\\\\gamma\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-export-snippet-back-end-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-export-snippet-back-end-value-key-post-blank\"><\/a>om-build-export-snippet <code>(back-end value &amp;key post-blank)<\/code><\/h4><p>Build an export-snippet object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>back-end<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>value<\/code><\/strong>: (required) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-export-snippet \"back\" \"value\")\n     (om-to-string))\n ;; =&gt; \"@@back:value@@\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-inline-babel-call-call-key-inside-header-arguments-end-header-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-inline-babel-call-call-key-inside-header-arguments-end-header-post-blank\"><\/a>om-build-inline-babel-call <code>(call &amp;key inside-header arguments end-header post-blank)<\/code><\/h4><p>Build an inline-babel-call object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>call<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>inside-header<\/code><\/strong>: a plist<\/li>\n<li><strong><code>arguments<\/code><\/strong>: a list of oneline strings<\/li>\n<li><strong><code>end-header<\/code><\/strong>: a plist<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-inline-babel-call \"name\")\n     (om-to-string))\n ;; =&gt; \"call_name()\"\n(-&gt;&gt; (om-build-inline-babel-call \"name\" :arguments '(\"n=4\"))\n     (om-to-string))\n ;; =&gt; \"call_name(n=4)\"\n(-&gt;&gt; (om-build-inline-babel-call \"name\" :inside-header '(:key val))\n     (om-to-string))\n ;; =&gt; \"call_name[:key val]()\"\n(-&gt;&gt; (om-build-inline-babel-call \"name\" :end-header '(:key val))\n     (om-to-string))\n ;; =&gt; \"call_name()[:key val]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-inline-src-block-language-key-parameters-value--post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-inline-src-block-language-key-parameters-value--post-blank\"><\/a>om-build-inline-src-block <code>(language &amp;key parameters (value \"\") post-blank)<\/code><\/h4><p>Build an inline-src-block object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>language<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>parameters<\/code><\/strong>: a plist<\/li>\n<li><strong><code>value<\/code><\/strong>: (default <code>\"\"<\/code>) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-inline-src-block \"lang\")\n     (om-to-string))\n ;; =&gt; \"src_lang{}\"\n(-&gt;&gt; (om-build-inline-src-block \"lang\" :value \"value\")\n     (om-to-string))\n ;; =&gt; \"src_lang{value}\"\n(-&gt;&gt; (om-build-inline-src-block \"lang\" :value \"value\" :parameters '(:key val))\n     (om-to-string))\n ;; =&gt; \"src_lang[:key val]{value}\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-line-break-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-line-break-key-post-blank\"><\/a>om-build-line-break <code>(&amp;key post-blank)<\/code><\/h4><p>Build a line-break object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-line-break)\n     (om-to-string))\n ;; =&gt; \"\\\\\\\\\n ;      \"\n<\/pre><\/div><h4><a id=\"user-content-om-build-latex-fragment-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-latex-fragment-value-key-post-blank\"><\/a>om-build-latex-fragment <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a latex-fragment object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-latex-fragment \"$2+2=5$\")\n     (om-to-string))\n ;; =&gt; \"$2+2=5$\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-macro-key-key-args-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-macro-key-key-args-post-blank\"><\/a>om-build-macro <code>(key &amp;key args post-blank)<\/code><\/h4><p>Build a macro object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>key<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>args<\/code><\/strong>: a list of oneline strings<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-macro \"economics\")\n     (om-to-string))\n ;; =&gt; \"{{{economics}}}\"\n(-&gt;&gt; (om-build-macro \"economics\" :args '(\"s=d\"))\n     (om-to-string))\n ;; =&gt; \"{{{economics(s=d)}}}\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-statistics-cookie-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-statistics-cookie-value-key-post-blank\"><\/a>om-build-statistics-cookie <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a statistics-cookie object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a list of non-neg integers like <code>(perc)<\/code> or <code>(num den)<\/code> which make [<code>num<\/code>\/<code>den<\/code>] and [<code>perc<\/code>%] respectively<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-statistics-cookie '(nil))\n     (om-to-string))\n ;; =&gt; \"[%]\"\n(-&gt;&gt; (om-build-statistics-cookie '(nil nil))\n     (om-to-string))\n ;; =&gt; \"[\/]\"\n(-&gt;&gt; (om-build-statistics-cookie '(50))\n     (om-to-string))\n ;; =&gt; \"[50%]\"\n(-&gt;&gt; (om-build-statistics-cookie '(1 3))\n     (om-to-string))\n ;; =&gt; \"[1\/3]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-target-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-target-value-key-post-blank\"><\/a>om-build-target <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a target object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-target \"text\")\n     (om-to-string))\n ;; =&gt; \"&lt;&lt;text&gt;&gt;\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-timestamp-type-year-start-month-start-day-start-year-end-month-end-day-end-key-hour-start-minute-start-hour-end-minute-end-repeater-type-repeater-unit-repeater-value-warning-type-warning-unit-warning-value-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-timestamp-type-year-start-month-start-day-start-year-end-month-end-day-end-key-hour-start-minute-start-hour-end-minute-end-repeater-type-repeater-unit-repeater-value-warning-type-warning-unit-warning-value-post-blank\"><\/a>om-build-timestamp <code>(type year-start month-start day-start year-end month-end day-end &amp;key hour-start minute-start hour-end minute-end repeater-type repeater-unit repeater-value warning-type warning-unit warning-value post-blank)<\/code><\/h4><p>Build a timestamp object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>type<\/code><\/strong>: (required) a symbol from <code>inactive<\/code>, <code>active<\/code>, <code>inactive-ranged<\/code>, or <code>active-ranged<\/code><\/li>\n<li><strong><code>year-start<\/code><\/strong>: (required) a positive integer<\/li>\n<li><strong><code>month-start<\/code><\/strong>: (required) a positive integer<\/li>\n<li><strong><code>day-start<\/code><\/strong>: (required) a positive integer<\/li>\n<li><strong><code>year-end<\/code><\/strong>: (required) a positive integer<\/li>\n<li><strong><code>month-end<\/code><\/strong>: (required) a positive integer<\/li>\n<li><strong><code>day-end<\/code><\/strong>: (required) a positive integer<\/li>\n<li><strong><code>hour-start<\/code><\/strong>: a non-negative integer or nil<\/li>\n<li><strong><code>minute-start<\/code><\/strong>: a non-negative integer or nil<\/li>\n<li><strong><code>hour-end<\/code><\/strong>: a non-negative integer or nil<\/li>\n<li><strong><code>minute-end<\/code><\/strong>: a non-negative integer or nil<\/li>\n<li><strong><code>repeater-type<\/code><\/strong>: nil or a symbol from <code>catch-up<\/code>, <code>restart<\/code>, or <code>cumulate<\/code><\/li>\n<li><strong><code>repeater-unit<\/code><\/strong>: nil or a symbol from <code>year<\/code> <code>month<\/code> <code>week<\/code> <code>day<\/code>, or <code>hour<\/code><\/li>\n<li><strong><code>repeater-value<\/code><\/strong>: a positive integer or nil<\/li>\n<li><strong><code>warning-type<\/code><\/strong>: nil or a symbol from <code>all<\/code> or <code>first<\/code><\/li>\n<li><strong><code>warning-unit<\/code><\/strong>: nil or a symbol from <code>year<\/code> <code>month<\/code> <code>week<\/code> <code>day<\/code>, or <code>hour<\/code><\/li>\n<li><strong><code>warning-value<\/code><\/strong>: a positive integer or nil<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-timestamp 'inactive\n                         2019 1 15 2019 1 15)\n     (om-to-string))\n ;; =&gt; \"[2019-01-15 Tue]\"\n(-&gt;&gt; (om-build-timestamp 'active-range\n                         2019 1 15 2019 1 16)\n     (om-to-string))\n ;; =&gt; \"&lt;2019-01-15 Tue&gt;--&lt;2019-01-16 Wed&gt;\"\n(-&gt;&gt; (om-build-timestamp 'inactive\n                         2019 1 15 2019 1 15 :warning-type 'all\n                         :warning-unit 'day\n                         :warning-value 1)\n     (om-to-string))\n ;; =&gt; \"[2019-01-15 Tue -1d]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-verbatim-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-verbatim-value-key-post-blank\"><\/a>om-build-verbatim <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a verbatim object node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-verbatim \"text\")\n     (om-to-string))\n ;; =&gt; \"=text=\"\n<\/pre><\/div><h3><a id=\"user-content-branch-object-nodes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#branch-object-nodes-1\"><\/a>Branch Object Nodes<\/h3><h4><a id=\"user-content-om-build-bold-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-bold-key-post-blank-rest-object-nodes\"><\/a>om-build-bold <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a bold object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-bold \"text\")\n     (om-to-string))\n ;; =&gt; \"*text*\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-footnote-reference-key-label-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-footnote-reference-key-label-post-blank-rest-object-nodes\"><\/a>om-build-footnote-reference <code>(&amp;key label post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a footnote-reference object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>label<\/code><\/strong>: a oneline string or nil<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-footnote-reference)\n     (om-to-string))\n ;; =&gt; \"[fn:]\"\n(-&gt;&gt; (om-build-footnote-reference :label \"label\")\n     (om-to-string))\n ;; =&gt; \"[fn:label]\"\n(-&gt;&gt; (om-build-footnote-reference :label \"label\" \"content\")\n     (om-to-string))\n ;; =&gt; \"[fn:label:content]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-italic-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-italic-key-post-blank-rest-object-nodes\"><\/a>om-build-italic <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build an italic object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-italic \"text\")\n     (om-to-string))\n ;; =&gt; \"\/text\/\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-link-path-key-format-type-fuzzy-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-link-path-key-format-type-fuzzy-post-blank-rest-object-nodes\"><\/a>om-build-link <code>(path &amp;key format (type \"fuzzy\") post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a link object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>path<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>format<\/code><\/strong>: the symbol <code>plain<\/code>, <code>bracket<\/code> or <code>angle<\/code><\/li>\n<li><strong><code>type<\/code><\/strong>: (default <code>\"fuzzy\"<\/code>) a oneline string from <code>org-link-types<\/code> or <code>\"coderef\"<\/code>, <code>\"custom-id\"<\/code>, <code>\"file\"<\/code>, <code>\"id\"<\/code>, <code>\"radio\"<\/code>, or <code>\"fuzzy\"<\/code><\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-link \"target\")\n     (om-to-string))\n ;; =&gt; \"[[target]]\"\n(-&gt;&gt; (om-build-link \"target\" :type \"file\")\n     (om-to-string))\n ;; =&gt; \"[[file:target]]\"\n(-&gt;&gt; (om-build-link \"target\" \"desc\")\n     (om-to-string))\n ;; =&gt; \"[[target][desc]]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-radio-target-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-radio-target-key-post-blank-rest-object-nodes\"><\/a>om-build-radio-target <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a radio-target object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-radio-target \"text\")\n     (om-to-string))\n ;; =&gt; \"&lt;&lt;&lt;text&gt;&gt;&gt;\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-strike-through-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-strike-through-key-post-blank-rest-object-nodes\"><\/a>om-build-strike-through <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a strike-through object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-strike-through \"text\")\n     (om-to-string))\n ;; =&gt; \"+text+\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-superscript-key-use-brackets-p-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-superscript-key-use-brackets-p-post-blank-rest-object-nodes\"><\/a>om-build-superscript <code>(&amp;key use-brackets-p post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a superscript object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>use-brackets-p<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-superscript \"text\")\n     (om-to-string))\n ;; =&gt; \"^text\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-subscript-key-use-brackets-p-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-subscript-key-use-brackets-p-post-blank-rest-object-nodes\"><\/a>om-build-subscript <code>(&amp;key use-brackets-p post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a subscript object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>use-brackets-p<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-subscript \"text\")\n     (om-to-string))\n ;; =&gt; \"_text\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-cell-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-cell-key-post-blank-rest-object-nodes\"><\/a>om-build-table-cell <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a table-cell object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table-cell \"text\")\n     (om-to-string))\n ;; =&gt; \" text |\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-underline-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-underline-key-post-blank-rest-object-nodes\"><\/a>om-build-underline <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build an underline object node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-underline \"text\")\n     (om-to-string))\n ;; =&gt; \"_text_\"\n<\/pre><\/div><h3><a id=\"user-content-leaf-element-nodes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#leaf-element-nodes-1\"><\/a>Leaf Element Nodes<\/h3><h4><a id=\"user-content-om-build-babel-call-call-key-inside-header-arguments-end-header-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-babel-call-call-key-inside-header-arguments-end-header-post-blank\"><\/a>om-build-babel-call <code>(call &amp;key inside-header arguments end-header post-blank)<\/code><\/h4><p>Build a babel-call element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>call<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>inside-header<\/code><\/strong>: a plist<\/li>\n<li><strong><code>arguments<\/code><\/strong>: a list of oneline strings<\/li>\n<li><strong><code>end-header<\/code><\/strong>: a plist<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-babel-call \"name\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: name()\"\n(-&gt;&gt; (om-build-babel-call \"name\" :arguments '(\"arg=x\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: name(arg=x)\"\n(-&gt;&gt; (om-build-babel-call \"name\" :inside-header '(:key val))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: name[:key val]()\"\n(-&gt;&gt; (om-build-babel-call \"name\" :end-header '(:key val))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: name() :key val\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-clock-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-clock-value-key-post-blank\"><\/a>om-build-clock <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a clock element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) an unranged, inactive timestamp node with no warning or repeater<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-clock (om-build-timestamp! '(2019 1 1 0 0)))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOCK: [2019-01-01 Tue 00:00]\"\n(-&gt;&gt; (om-build-clock (om-build-timestamp! '(2019 1 1 0 0)\n                                          :end '(2019 1 1 1 0)))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOCK: [2019-01-01 Tue 00:00-01:00] =&gt;  1:00\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-comment-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-comment-value-key-post-blank\"><\/a>om-build-comment <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a comment element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-comment \"text\")\n     (om-to-trimmed-string))\n ;; =&gt; \"# text\"\n(-&gt;&gt; (om-build-comment \"text\\nless\")\n     (om-to-trimmed-string))\n ;; =&gt; \"# text\n ;      # less\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-comment-block-key-value--post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-comment-block-key-value--post-blank\"><\/a>om-build-comment-block <code>(&amp;key (value \"\") post-blank)<\/code><\/h4><p>Build a comment-block element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (default <code>\"\"<\/code>) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-comment-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_COMMENT\n ;      #+END_COMMENT\"\n(-&gt;&gt; (om-build-comment-block :value \"text\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_COMMENT\n ;      text\n ;      #+END_COMMENT\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-diary-sexp-key-value-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-diary-sexp-key-value-post-blank\"><\/a>om-build-diary-sexp <code>(&amp;key value post-blank)<\/code><\/h4><p>Build a diary-sexp element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: a list form or nil<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-diary-sexp)\n     (om-to-trimmed-string))\n ;; =&gt; \"%%()\"\n(-&gt;&gt; (om-build-diary-sexp :value '(text))\n     (om-to-trimmed-string))\n ;; =&gt; \"%%(text)\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-example-block-key-preserve-indent-switches-value--post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-example-block-key-preserve-indent-switches-value--post-blank\"><\/a>om-build-example-block <code>(&amp;key preserve-indent switches (value \"\") post-blank)<\/code><\/h4><p>Build an example-block element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>preserve-indent<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>switches<\/code><\/strong>: a list of oneline strings<\/li>\n<li><strong><code>value<\/code><\/strong>: (default <code>\"\"<\/code>) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-example-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_EXAMPLE\n ;      #+END_EXAMPLE\"\n(-&gt;&gt; (om-build-example-block :value \"text\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_EXAMPLE\n ;      text\n ;      #+END_EXAMPLE\"\n(-&gt;&gt; (om-build-example-block :value \"text\" :switches '(\"switches\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_EXAMPLE switches\n ;      text\n ;      #+END_EXAMPLE\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-export-block-type-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-export-block-type-value-key-post-blank\"><\/a>om-build-export-block <code>(type value &amp;key post-blank)<\/code><\/h4><p>Build an export-block element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>type<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>value<\/code><\/strong>: (required) a string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-export-block \"type\" \"value\\n\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_EXPORT type\n ;      value\n ;      #+END_EXPORT\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-fixed-width-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-fixed-width-value-key-post-blank\"><\/a>om-build-fixed-width <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a fixed-width element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-fixed-width \"text\")\n     (om-to-trimmed-string))\n ;; =&gt; \": text\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-horizontal-rule-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-horizontal-rule-key-post-blank\"><\/a>om-build-horizontal-rule <code>(&amp;key post-blank)<\/code><\/h4><p>Build a horizontal-rule element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-horizontal-rule)\n     (om-to-trimmed-string))\n ;; =&gt; \"-----\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-keyword-key-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-keyword-key-value-key-post-blank\"><\/a>om-build-keyword <code>(key value &amp;key post-blank)<\/code><\/h4><p>Build a keyword element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>key<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>value<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-keyword \"FILETAGS\" \"tmsu\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+FILETAGS: tmsu\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-latex-environment-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-latex-environment-value-key-post-blank\"><\/a>om-build-latex-environment <code>(value &amp;key post-blank)<\/code><\/h4><p>Build a latex-environment element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (required) a list of strings like <code>(env body)<\/code> or <code>(env)<\/code><\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-latex-environment '(\"env\" \"text\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"\\\\begin{env}\n ;      text\n ;      \\\\end{env}\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-node-property-key-value-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-node-property-key-value-key-post-blank\"><\/a>om-build-node-property <code>(key value &amp;key post-blank)<\/code><\/h4><p>Build a node-property element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>key<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>value<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-node-property \"key\" \"val\")\n     (om-to-trimmed-string))\n ;; =&gt; \":key:      val\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-planning-key-closed-deadline-scheduled-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-planning-key-closed-deadline-scheduled-post-blank\"><\/a>om-build-planning <code>(&amp;key closed deadline scheduled post-blank)<\/code><\/h4><p>Build a planning element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>closed<\/code><\/strong>: a zero-range, inactive timestamp node<\/li>\n<li><strong><code>deadline<\/code><\/strong>: a zero-range, inactive timestamp node<\/li>\n<li><strong><code>scheduled<\/code><\/strong>: a zero-range, inactive timestamp node<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-planning :closed (om-build-timestamp! '(2019 1 1)))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOSED: [2019-01-01 Tue]\"\n(-&gt;&gt; (om-build-planning :scheduled (om-build-timestamp! '(2019 1 1)))\n     (om-to-trimmed-string))\n ;; =&gt; \"SCHEDULED: [2019-01-01 Tue]\"\n(-&gt;&gt; (om-build-planning :deadline (om-build-timestamp! '(2019 1 1)))\n     (om-to-trimmed-string))\n ;; =&gt; \"DEADLINE: [2019-01-01 Tue]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-src-block-key-value--language-parameters-preserve-indent-switches-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-src-block-key-value--language-parameters-preserve-indent-switches-post-blank\"><\/a>om-build-src-block <code>(&amp;key (value \"\") language parameters preserve-indent switches post-blank)<\/code><\/h4><p>Build a src-block element node.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>value<\/code><\/strong>: (default <code>\"\"<\/code>) a string<\/li>\n<li><strong><code>language<\/code><\/strong>: a string or nil<\/li>\n<li><strong><code>parameters<\/code><\/strong>: a plist<\/li>\n<li><strong><code>preserve-indent<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>switches<\/code><\/strong>: a list of oneline strings<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-src-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_SRC\n ;      #+END_SRC\"\n(-&gt;&gt; (om-build-src-block :value \"body\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_SRC\n ;        body\n ;      #+END_SRC\"\n(-&gt;&gt; (om-build-src-block :value \"body\" :language \"emacs-lisp\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_SRC emacs-lisp\n ;        body\n ;      #+END_SRC\"\n(-&gt;&gt; (om-build-src-block :value \"body\" :switches '(\"-n 20\" \"-r\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_SRC -n 20 -r\n ;        body\n ;      #+END_SRC\"\n(-&gt;&gt; (om-build-src-block :value \"body\" :parameters '(:key val))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_SRC :key val\n ;        body\n ;      #+END_SRC\"\n<\/pre><\/div><h3><a id=\"user-content-branch-element-nodes-with-child-object-nodes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#branch-element-nodes-with-child-object-nodes-1\"><\/a>Branch Element Nodes with Child Object Nodes<\/h3><h4><a id=\"user-content-om-build-paragraph-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-paragraph-key-post-blank-rest-object-nodes\"><\/a>om-build-paragraph <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a paragraph element node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-paragraph \"text\")\n     (om-to-trimmed-string))\n ;; =&gt; \"text\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-row-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-row-key-post-blank-rest-object-nodes\"><\/a>om-build-table-row <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a table-row element node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table-cell \"a\")\n     (om-build-table-row)\n     (om-to-trimmed-string))\n ;; =&gt; \"| a |\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-verse-block-key-post-blank-rest-object-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-verse-block-key-post-blank-rest-object-nodes\"><\/a>om-build-verse-block <code>(&amp;key post-blank &amp;rest object-nodes)<\/code><\/h4><p>Build a verse-block element node with <strong><code>object-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-verse-block \"text\\n\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_VERSE\n ;      text\n ;      #+END_VERSE\"\n<\/pre><\/div><h3><a id=\"user-content-branch-element-nodes-with-child-element-nodes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#branch-element-nodes-with-child-element-nodes-1\"><\/a>Branch Element Nodes with Child Element Nodes<\/h3><h4><a id=\"user-content-om-build-center-block-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-center-block-key-post-blank-rest-element-nodes\"><\/a>om-build-center-block <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a center-block element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-center-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_CENTER\n ;      #+END_CENTER\"\n(-&gt;&gt; (om-build-paragraph \"text\")\n     (om-build-center-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_CENTER\n ;      text\n ;      #+END_CENTER\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-drawer-drawer-name-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-drawer-drawer-name-key-post-blank-rest-element-nodes\"><\/a>om-build-drawer <code>(drawer-name &amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a drawer element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>drawer-name<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-drawer \"NAME\")\n     (om-to-trimmed-string))\n ;; =&gt; \":NAME:\n ;      :END:\"\n(-&gt;&gt; (om-build-paragraph \"text\")\n     (om-build-drawer \"NAME\")\n     (om-to-trimmed-string))\n ;; =&gt; \":NAME:\n ;      text\n ;      :END:\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-dynamic-block-block-name-key-arguments-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-dynamic-block-block-name-key-arguments-post-blank-rest-element-nodes\"><\/a>om-build-dynamic-block <code>(block-name &amp;key arguments post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a dynamic-block element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>block-name<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>arguments<\/code><\/strong>: a plist<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-dynamic-block \"empty\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN: empty\n ;      #+END:\"\n(-&gt;&gt; (om-build-comment \"I'm in here\")\n     (om-build-dynamic-block \"notempty\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN: notempty\n ;      # I'm in here\n ;      #+END:\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-footnote-definition-label-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-footnote-definition-label-key-post-blank-rest-element-nodes\"><\/a>om-build-footnote-definition <code>(label &amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a footnote-definition element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>label<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-paragraph \"footnote contents\")\n     (om-build-footnote-definition \"label\")\n     (om-to-trimmed-string))\n ;; =&gt; \"[fn:label] footnote contents\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-headline-key-archivedp-commentedp-footnote-section-p-level-1-pre-blank-0-priority-tags-title-todo-keyword-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-headline-key-archivedp-commentedp-footnote-section-p-level-1-pre-blank-0-priority-tags-title-todo-keyword-post-blank-rest-element-nodes\"><\/a>om-build-headline <code>(&amp;key archivedp commentedp footnote-section-p (level 1) (pre-blank 0) priority tags title todo-keyword post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a headline element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>archivedp<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>commentedp<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>footnote-section-p<\/code><\/strong>: nil or t<\/li>\n<li><strong><code>level<\/code><\/strong>: a positive integer<\/li>\n<li><strong><code>pre-blank<\/code><\/strong>: a non-negative integer<\/li>\n<li><strong><code>priority<\/code><\/strong>: an integer between (inclusive) <code>org-highest-priority<\/code> and <code>org-lowest-priority<\/code><\/li>\n<li><strong><code>tags<\/code><\/strong>: a string list<\/li>\n<li><strong><code>title<\/code><\/strong>: a secondary string<\/li>\n<li><strong><code>todo-keyword<\/code><\/strong>: a oneline string or nil<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-headline)\n     (om-to-trimmed-string))\n ;; =&gt; \"*\"\n(-&gt;&gt; (om-build-headline :level 2 :title '(\"dummy\")\n                         :tags '(\"tmsu\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"** dummy            :tmsu:\"\n(-&gt;&gt; (om-build-headline :todo-keyword \"TODO\" :archivedp t :commentedp t :priority 65)\n     (om-to-trimmed-string))\n ;; =&gt; \"* TODO COMMENT [#A]  :ARCHIVE:\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-item-key-bullet---checkbox-counter-tag-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-item-key-bullet---checkbox-counter-tag-post-blank-rest-element-nodes\"><\/a>om-build-item <code>(&amp;key (bullet '-) checkbox counter tag post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build an item element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>bullet<\/code><\/strong>: (default <code>-<\/code>) a positive integer (ordered) or the symbol <code>-<\/code> (unordered)<\/li>\n<li><strong><code>checkbox<\/code><\/strong>: nil or the symbols <code>on<\/code>, <code>off<\/code>, or <code>trans<\/code><\/li>\n<li><strong><code>counter<\/code><\/strong>: a positive integer or nil<\/li>\n<li><strong><code>tag<\/code><\/strong>: a secondary string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-paragraph \"item contents\")\n     (om-build-item)\n     (om-to-trimmed-string))\n ;; =&gt; \"- item contents\"\n(-&gt;&gt; (om-build-paragraph \"item contents\")\n     (om-build-item :bullet 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"1. item contents\"\n(-&gt;&gt; (om-build-paragraph \"item contents\")\n     (om-build-item :checkbox 'on)\n     (om-to-trimmed-string))\n ;; =&gt; \"- [X] item contents\"\n(-&gt;&gt; (om-build-paragraph \"item contents\")\n     (om-build-item :tag '(\"tmsu\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"- tmsu :: item contents\"\n(-&gt;&gt; (om-build-paragraph \"item contents\")\n     (om-build-item :counter 10)\n     (om-to-trimmed-string))\n ;; =&gt; \"- [@10] item contents\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-plain-list-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-plain-list-key-post-blank-rest-element-nodes\"><\/a>om-build-plain-list <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a plain-list element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-paragraph \"item contents\")\n     (om-build-item)\n     (om-build-plain-list)\n     (om-to-trimmed-string))\n ;; =&gt; \"- item contents\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-property-drawer-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-property-drawer-key-post-blank-rest-element-nodes\"><\/a>om-build-property-drawer <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a property-drawer element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-property-drawer)\n     (om-to-trimmed-string))\n ;; =&gt; \":PROPERTIES:\n ;      :END:\"\n(-&gt;&gt; (om-build-node-property \"key\" \"val\")\n     (om-build-property-drawer)\n     (om-to-trimmed-string))\n ;; =&gt; \":PROPERTIES:\n ;      :key:      val\n ;      :END:\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-quote-block-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-quote-block-key-post-blank-rest-element-nodes\"><\/a>om-build-quote-block <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a quote-block element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-quote-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_QUOTE\n ;      #+END_QUOTE\"\n(-&gt;&gt; (om-build-paragraph \"quoted stuff\")\n     (om-build-quote-block)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_QUOTE\n ;      quoted stuff\n ;      #+END_QUOTE\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-section-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-section-key-post-blank-rest-element-nodes\"><\/a>om-build-section <code>(&amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a section element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-paragraph \"text\")\n     (om-build-section)\n     (om-to-trimmed-string))\n ;; =&gt; \"text\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-special-block-type-key-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-special-block-type-key-post-blank-rest-element-nodes\"><\/a>om-build-special-block <code>(type &amp;key post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a special-block element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>type<\/code><\/strong>: (required) a oneline string<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-special-block \"monad\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_monad\n ;      #+END_monad\"\n(-&gt;&gt; (om-build-comment \"Launch missiles\")\n     (om-build-special-block \"monad\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+BEGIN_monad\n ;      # Launch missiles\n ;      #+END_monad\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-key-tblfm-post-blank-rest-element-nodes\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-key-tblfm-post-blank-rest-element-nodes\"><\/a>om-build-table <code>(&amp;key tblfm post-blank &amp;rest element-nodes)<\/code><\/h4><p>Build a table element node with <strong><code>element-nodes<\/code><\/strong> as children.<\/p><p>The following properties are settable:<\/p><ul><li><strong><code>tblfm<\/code><\/strong>: a list of oneline strings<\/li>\n<li><strong><code>post-blank<\/code><\/strong>: a non-negative integer<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table-cell \"cell\")\n     (om-build-table-row)\n     (om-build-table)\n     (om-to-trimmed-string))\n ;; =&gt; \"| cell |\"\n<\/pre><\/div><h3><a id=\"user-content-miscellaneous-builders-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#miscellaneous-builders-1\"><\/a>Miscellaneous Builders<\/h3><h4><a id=\"user-content-om-build-secondary-string-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-secondary-string-string\"><\/a>om-build-secondary-string! <code>(string)<\/code><\/h4><p>Return a secondary string (list of object nodes) from <strong><code>string<\/code><\/strong>. <strong><code>string<\/code><\/strong> is any string that contains a textual representation of object nodes. If the string does not represent a list of object nodes, throw an error.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-secondary-string! \"I'm plain\")\n     (-map (function om-get-type)))\n ;; =&gt; '(plain-text)\n(-&gt;&gt; (om-build-secondary-string! \"I'm *not* plain\")\n     (-map (function om-get-type)))\n ;; =&gt; '(plain-text bold plain-text)\n(-&gt;&gt; (om-build-secondary-string! \"* I'm not an object\")\n     (-map (function om-get-type)))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-row-hline-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-row-hline-key-post-blank\"><\/a>om-build-table-row-hline <code>(&amp;key post-blank)<\/code><\/h4><p>Return a new rule-typed table-row node. Optionally set <strong><code>post-blank<\/code><\/strong> (a positive integer).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table (om-build-table-row (om-build-table-cell \"text\"))\n                     (om-build-table-row-hline))\n     (om-to-trimmed-string))\n ;; =&gt; \"| text |\n ;      |------|\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-timestamp-diary-form-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-timestamp-diary-form-key-post-blank\"><\/a>om-build-timestamp-diary <code>(form &amp;key post-blank)<\/code><\/h4><p>Return a new diary-sexp timestamp node from <strong><code>form<\/code><\/strong>. Optionally set <strong><code>post-blank<\/code><\/strong> (a positive integer).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-timestamp-diary '(diary-float t 4 2))\n     (om-to-string))\n ;; =&gt; \"&lt;%%(diary-float t 4 2)&gt;\"\n<\/pre><\/div><h3><a id=\"user-content-shorthand-builders-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#shorthand-builders-1\"><\/a>Shorthand Builders<\/h3><p>Build nodes with more convenient\/shorter syntax.<\/p><h4><a id=\"user-content-om-build-timestamp-start-key-end-active-repeater-warning-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><\/a>om-build-timestamp! <code>(start &amp;key end active repeater warning post-blank)<\/code><\/h4><p>Return a new timestamp node.<\/p><p><strong><code>start<\/code><\/strong> specifies the start time and is a list of integers in one of the following forms:<\/p><ul><li><code>(year month day)<\/code>: short form<\/li>\n<li><code>(year month day nil nil)<\/code>: short form<\/li>\n<li><code>(year month day hour minute)<\/code>: long form<\/li>\n<\/ul><p><strong><code>end<\/code><\/strong> (if supplied) will add the ending time, and follows the same formatting rules as <strong><code>start<\/code><\/strong>.<\/p><p><strong><code>active<\/code><\/strong> is a boolean where t signifies the type is <code>active<\/code>, else <code>inactive<\/code> (the range suffix will be added if an end time is supplied).<\/p><p><strong><code>repeater<\/code><\/strong> and <strong><code>warning<\/code><\/strong> are lists formatted as <code>(type value unit)<\/code> where the three members correspond to the :repeater\/warning-type, -value, and -unit properties in <a href=\"#om-build-timestamp-type-year-start-month-start-day-start-year-end-month-end-day-end-key-hour-start-minute-start-hour-end-minute-end-repeater-type-repeater-unit-repeater-value-warning-type-warning-unit-warning-value-post-blank\"><code>om-build-timestamp<\/code><\/a>.<\/p><p>Building a diary sexp timestamp is not possible with this function.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-timestamp! '(2019 1 1))\n     (om-to-string))\n ;; =&gt; \"[2019-01-01 Tue]\"\n(-&gt;&gt; (om-build-timestamp! '(2019 1 1 12 0)\n                          :active t :warning '(all 1 day)\n                          :repeater '(cumulate 1 month))\n     (om-to-string))\n ;; =&gt; \"&lt;2019-01-01 Tue 12:00 +1m -1d&gt;\"\n(-&gt;&gt; (om-build-timestamp! '(2019 1 1)\n                          :end '(2019 1 2))\n     (om-to-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-clock-start-key-end-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-clock-start-key-end-post-blank\"><\/a>om-build-clock! <code>(start &amp;key end post-blank)<\/code><\/h4><p>Return a new clock node.<\/p><p><strong><code>start<\/code><\/strong> and <strong><code>end<\/code><\/strong> follow the same rules as their respective arguments in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-clock! '(2019 1 1))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOCK: [2019-01-01 Tue]\"\n(-&gt;&gt; (om-build-clock! '(2019 1 1 12 0))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOCK: [2019-01-01 Tue 12:00]\"\n(-&gt;&gt; (om-build-clock! '(2019 1 1 12 0)\n                      :end '(2019 1 1 13 0))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOCK: [2019-01-01 Tue 12:00-13:00] =&gt;  1:00\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-planning-key-closed-deadline-scheduled-post-blank-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-planning-key-closed-deadline-scheduled-post-blank-1\"><\/a>om-build-planning! <code>(&amp;key closed deadline scheduled post-blank)<\/code><\/h4><p>Return a new planning node.<\/p><p><strong><code>closed<\/code><\/strong>, <strong><code>deadline<\/code><\/strong>, and <strong><code>scheduled<\/code><\/strong> are lists with the following structure (brackets denote optional members):<\/p><p><code>(year minute day [hour] [min] [&amp;warning type value unit] [&amp;repeater type value unit])<\/code><\/p><p>In terms of arguments supplied to <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>, the first five members correspond to the list supplied as <code>time<\/code>, and the <code>type<\/code>, <code>value<\/code>, and <code>unit<\/code> fields correspond to the lists supplied to <code>warning<\/code> and <code>repeater<\/code> arguments. The order of warning and repeater does not matter.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-planning! :closed '(2019 1 1))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOSED: [2019-01-01 Tue]\"\n(-&gt;&gt; (om-build-planning! :closed '(2019 1 1)\n                          :scheduled '(2018 1 1))\n     (om-to-trimmed-string))\n ;; =&gt; \"SCHEDULED: [2018-01-01 Mon] CLOSED: [2019-01-01 Tue]\"\n(-&gt;&gt; (om-build-planning! :closed '(2019 1 1 &amp;warning all 1 day &amp;repeater cumulate 1 month))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOSED: [2019-01-01 Tue +1m -1d]\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-property-drawer-key-post-blank-rest-keyvals\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-property-drawer-key-post-blank-rest-keyvals\"><\/a>om-build-property-drawer! <code>(&amp;key post-blank &amp;rest keyvals)<\/code><\/h4><p>Return a new property-drawer node.<\/p><p>Each member in <strong><code>keyvals<\/code><\/strong> is a list of symbols like <code>(key val)<\/code>, where each list will generate a node-property node in the property-drawer node like <code>\":key: val\"<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-property-drawer! '(key val))\n     (om-to-trimmed-string))\n ;; =&gt; \":PROPERTIES:\n ;      :key:      val\n ;      :END:\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-headline-key-level-1-title-text-todo-keyword-tags-pre-blank-priority-commentedp-archivedp-post-blank-planning-statistics-cookie-section-children-rest-subheadlines\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-headline-key-level-1-title-text-todo-keyword-tags-pre-blank-priority-commentedp-archivedp-post-blank-planning-statistics-cookie-section-children-rest-subheadlines\"><\/a>om-build-headline! <code>(&amp;key (level 1) title-text todo-keyword tags pre-blank priority commentedp archivedp post-blank planning statistics-cookie section-children &amp;rest subheadlines)<\/code><\/h4><p>Return a new headline node.<\/p><p><strong><code>title-text<\/code><\/strong> is a oneline string for the title of the headline.<\/p><p><strong><code>planning<\/code><\/strong> is a list like <code>(planning-type args ...)<\/code> where <code>planning-type<\/code> is one of <code>:closed<\/code>, <code>:deadline<\/code>, or <code>:scheduled<\/code>, and <code>args<\/code> are the args supplied to any of the planning types in <a href=\"#om-build-planning-key-closed-deadline-scheduled-post-blank\"><code>om-build-planning!<\/code><\/a>. Up to all three planning types can be used in the same list like <code>(:closed args :deadline args :scheduled args)<\/code>.<\/p><p><strong><code>statistics-cookie<\/code><\/strong> is a list following the same format as <a href=\"#om-build-statistics-cookie-value-key-post-blank\"><code>om-build-statistics-cookie<\/code><\/a>.<\/p><p><strong><code>section-children<\/code><\/strong> is a list of elements that will go in the headline section.<\/p><p><strong><code>subheadlines<\/code><\/strong> contains zero or more headlines that will go under the created headline. The level of all members in <strong><code>subheadlines<\/code><\/strong> will automatically be adjusted to <strong><code>level<\/code><\/strong> + 1.<\/p><p>All arguments not mentioned here follow the same rules as <a href=\"#om-build-headline-key-archivedp-commentedp-footnote-section-p-level-1-pre-blank-0-priority-tags-title-todo-keyword-post-blank-rest-element-nodes\"><code>om-build-headline<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-headline! :title-text \"really impressive title\")\n     (om-to-trimmed-string))\n ;; =&gt; \"* really impressive title\"\n(-&gt;&gt; (om-build-headline! :title-text \"really impressive title\" :statistics-cookie '(0 9000))\n     (om-to-trimmed-string))\n ;; =&gt; \"* really impressive title [0\/9000]\"\n(-&gt;&gt; (om-build-headline! :title-text \"really impressive title\" :section-children (list (om-build-property-drawer! '(key val))\n                                                                                       (om-build-paragraph! \"section text\"))\n                          (om-build-headline! :title-text \"subhead\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"* really impressive title\n ;      :PROPERTIES:\n ;      :key:      val\n ;      :END:\n ;      section text\n ;      ** subhead\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-item-key-post-blank-bullet-checkbox-tag-paragraph-counter-rest-children\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-item-key-post-blank-bullet-checkbox-tag-paragraph-counter-rest-children\"><\/a>om-build-item! <code>(&amp;key post-blank bullet checkbox tag paragraph counter &amp;rest children)<\/code><\/h4><p>Return a new item node.<\/p><p><strong><code>tag<\/code><\/strong> is a string representing the tag (make with <a href=\"#om-build-secondary-string-string\"><code>om-build-secondary-string!<\/code><\/a>) .<\/p><p><strong><code>paragraph<\/code><\/strong> is a string that will be the initial text in the item (made with <a href=\"#om-build-paragraph-string-key-post-blank\"><code>om-build-paragraph!<\/code><\/a>).<\/p><p><strong><code>children<\/code><\/strong> contains the nodes that will go under this item after <strong><code>paragraph<\/code><\/strong>.<\/p><p>All other arguments follow the same rules as <a href=\"#om-build-item-key-bullet-quote---checkbox-counter-tag-post-blank-rest-element-nodes\"><code>om-build-item<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-item! :bullet 1 :tag \"complicated *tag*\" :paragraph \"petulant \/frenzy\/\" (om-build-plain-list (om-build-item! :bullet '-\n                                                                                                                             :paragraph \"below\")))\n     (om-to-trimmed-string))\n ;; =&gt; \"1. complicated *tag* :: petulant \/frenzy\/\n ;         - below\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-paragraph-string-key-post-blank\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-paragraph-string-key-post-blank\"><\/a>om-build-paragraph! <code>(string &amp;key post-blank)<\/code><\/h4><p>Return a new paragraph node from <strong><code>string<\/code><\/strong>.<\/p><p><strong><code>string<\/code><\/strong> is the text to be parsed into a paragraph and must contain valid textual representations of object nodes.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-paragraph! \"stuff \/with\/ *formatting*\" :post-blank 2)\n     (om-to-string))\n ;; =&gt; \"stuff \/with\/ *formatting*\n ;      \n ;      \n ;      \"\n(-&gt;&gt; (om-build-paragraph! \"* stuff \/with\/ *formatting*\")\n     (om-to-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-cell-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-cell-string\"><\/a>om-build-table-cell! <code>(string)<\/code><\/h4><p>Return a new table-cell node.<\/p><p><strong><code>string<\/code><\/strong> is the text to be contained in the table-cell node. It must contain valid textual representations of objects that are allowed in table-cell nodes.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table-cell! \"rage\")\n     (om-to-trimmed-string))\n ;; =&gt; \"rage |\"\n(-&gt;&gt; (om-build-table-cell! \"*rage*\")\n     (om-to-trimmed-string))\n ;; =&gt; \"*rage* |\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-row-row-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-row-row-list\"><\/a>om-build-table-row! <code>(row-list)<\/code><\/h4><p>Return a new table-row node.<\/p><p><strong><code>row-list<\/code><\/strong> is a list of strings to be built into table-cell nodes via <a href=\"#om-build-table-cell-string\"><code>om-build-table-cell!<\/code><\/a> (see that function for restrictions). Alternatively, <strong><code>row-list<\/code><\/strong> may the symbol <code>hline<\/code> instead of a string to create a rule-typed table-row.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table-row! '(\"R\" \"A\" \"G\" \"E\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| R | A | G | E |\"\n(-&gt;&gt; (om-build-table-row! 'hline)\n     (om-to-trimmed-string))\n ;; =&gt; \"|-\"\n<\/pre><\/div><h4><a id=\"user-content-om-build-table-key-tblfm-post-blank-rest-row-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-build-table-key-tblfm-post-blank-rest-row-lists\"><\/a>om-build-table! <code>(&amp;key tblfm post-blank &amp;rest row-lists)<\/code><\/h4><p>Return a new table node.<\/p><p><strong><code>row-lists<\/code><\/strong> is a list of lists where each member list will be converted to a table-row node via <a href=\"#om-build-table-row-row-list\"><code>om-build-table-row!<\/code><\/a> (see that function for restrictions).<\/p><p>All other arguments follow the same rules as <a href=\"#om-build-table-key-tblfm-post-blank-rest-element-nodes\"><code>om-build-table<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; (om-build-table! '(\"R\" \"A\")\n                      '(\"G\" \"E\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| R | A |\n ;      | G | E |\"\n(-&gt;&gt; (om-build-table! '(\"L\" \"O\")\n                      'hline\n                      '(\"V\" \"E\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| L | O |\n ;      |---+---|\n ;      | V | E |\"\n<\/pre><\/div><h2><a id=\"user-content-type-predicates-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#type-predicates-1\"><\/a>Type Predicates<\/h2><p>Test node types.<\/p><h4><a id=\"user-content-om-get-type-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-get-type-node\"><\/a>om-get-type <code>(node)<\/code><\/h4><p>Return the type of <strong><code>node<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *I'm emboldened*\n(-&gt;&gt; (om-parse-this-object)\n     (om-get-type))\n ;; =&gt; 'bold\n;; Given the following contents:\n; * I'm the headliner\n(-&gt;&gt; (om-parse-this-element)\n     (om-get-type))\n ;; =&gt; 'headline\n;; Given the following contents:\n; [2112-12-21 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-get-type))\n ;; =&gt; 'timestamp\n<\/pre><\/div><h4><a id=\"user-content-om-is-type-type-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-is-type-type-node\"><\/a>om-is-type <code>(type node)<\/code><\/h4><p>Return t if the type of <strong><code>node<\/code><\/strong> is <strong><code>type<\/code><\/strong> (a symbol).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *ziltoid*\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-type 'bold))\n ;; =&gt; t\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-type 'italic))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-is-any-type-types-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-is-any-type-types-node\"><\/a>om-is-any-type <code>(types node)<\/code><\/h4><p>Return t if the type of <strong><code>node<\/code><\/strong> is in <strong><code>types<\/code><\/strong> (a list of symbols).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *ziltoid*\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-any-type '(bold)))\n ;; =&gt; t\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-any-type '(bold italic)))\n ;; =&gt; t\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-any-type '(italic)))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-is-element-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-is-element-node\"><\/a>om-is-element <code>(node)<\/code><\/h4><p>Return t if <strong><code>node<\/code><\/strong> is an element class.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *ziltoid*\n;; Parsing this text as an element node gives a paragraph node\n(-&gt;&gt; (om-parse-this-element)\n     (om-is-element))\n ;; =&gt; t\n;; Parsing the same text as an object node gives a bold node\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-element))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-is-branch-node-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-is-branch-node-node\"><\/a>om-is-branch-node <code>(node)<\/code><\/h4><p>Return t if <strong><code>node<\/code><\/strong> is a branch node.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *ziltoid*\n;; Parsing this as an element node gives a paragraph node (a branch node)\n(-&gt;&gt; (om-parse-this-element)\n     (om-is-branch-node))\n ;; =&gt; t\n;; Parsing this as an object node gives a bold node (also a branch node)\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-branch-node))\n ;; =&gt; t\n;; Given the following contents:\n; ~ziltoid~\n;; Parsing this as an object node gives a code node (not a branch node)\n(-&gt;&gt; (om-parse-this-object)\n     (om-is-branch-node))\n ;; =&gt; nil\n;; Given the following contents:\n; # ziltoid\n;; Parsing this as an element node gives a comment node (also not a branch node)\n(-&gt;&gt; (om-parse-this-element)\n     (om-is-branch-node))\n ;; =&gt; nil\n;; Given the following contents:\n; * I'm so great\n;; Parsing this as an element node gives a headline node (a branch node)\n(-&gt;&gt; (om-parse-this-element)\n     (om-is-branch-node))\n ;; =&gt; t\n<\/pre><\/div><h4><a id=\"user-content-om-node-may-have-child-objects-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-node-may-have-child-objects-node\"><\/a>om-node-may-have-child-objects <code>(node)<\/code><\/h4><p>Return t if <strong><code>node<\/code><\/strong> is a branch node that may have child objects.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *ziltoid*\n;; Parsing this as an element node gives a paragraph node (can have child object\n;; nodes)\n(-&gt;&gt; (om-parse-this-element)\n     (om-node-may-have-child-objects))\n ;; =&gt; t\n;; Parsing this as an object node gives a bold node (also can have child object\n;; nodes)\n(-&gt;&gt; (om-parse-this-object)\n     (om-node-may-have-child-objects))\n ;; =&gt; t\n;; Given the following contents:\n; ~ziltoid~\n;; Parsing this as an object node gives a code node (not a branch node)\n(-&gt;&gt; (om-parse-this-object)\n     (om-node-may-have-child-objects))\n ;; =&gt; nil\n;; Given the following contents:\n; # ziltoid\n;; Parsing this as an element node gives a comment node (not a branch node)\n(-&gt;&gt; (om-parse-this-element)\n     (om-node-may-have-child-objects))\n ;; =&gt; nil\n;; Given the following contents:\n; * I'm so great\n;; Parsing this as an element node gives a headline node (can only have child\n;; element nodes)\n(-&gt;&gt; (om-parse-this-element)\n     (om-node-may-have-child-objects))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-node-may-have-child-elements-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-node-may-have-child-elements-node\"><\/a>om-node-may-have-child-elements <code>(node)<\/code><\/h4><p>Return t if <strong><code>node<\/code><\/strong> is a branch node that may have child elements.<\/p><p>Note this implies that <strong><code>node<\/code><\/strong> is also of class element since only elements may have other elements as children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * I'm so great\n;; Parsing this as an element node gives a headline node (can have child element\n;; nodes)\n(-&gt;&gt; (om-parse-this-element)\n     (om-node-may-have-child-elements))\n ;; =&gt; t\n;; Given the following contents:\n; *ziltoid*\n;; Parsing this as an element node gives a paragraph node (can only have child\n;; object nodes)\n(-&gt;&gt; (om-parse-this-element)\n     (om-node-may-have-child-elements))\n ;; =&gt; nil\n;; Given the following contents:\n; # ziltoid\n;; Parsing this as an element node gives a comment node (not a branch node)\n(-&gt;&gt; (om-parse-this-element)\n     (om-node-may-have-child-elements))\n ;; =&gt; nil\n<\/pre><\/div><h2><a id=\"user-content-property-manipulation-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#property-manipulation-1\"><\/a>Property Manipulation<\/h2><p>Set, get, and map properties of nodes.<\/p><h3><a id=\"user-content-generic-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#generic-1\"><\/a>Generic<\/h3><h4><a id=\"user-content-om-contains-point-p-point-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-contains-point-p-point-node\"><\/a>om-contains-point-p <code>(point node)<\/code><\/h4><p>Return t if <strong><code>point<\/code><\/strong> is within the boundaries of <strong><code>node<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *findme*\n(-&gt;&gt; (om-parse-this-object)\n     (om-contains-point-p 2))\n ;; =&gt; t\n(-&gt;&gt; (om-parse-this-object)\n     (om-contains-point-p 10))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-set-property-prop-value-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-set-property-prop-value-node\"><\/a>om-set-property <code>(prop value node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>prop<\/code><\/strong> set to <strong><code>value<\/code><\/strong>.<\/p><p>See builder functions for a list of properties and their rules for each type.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu()\n(-&gt;&gt; (om-parse-this-element)\n     (om-set-property :call \"cthulhu\")\n     (om-set-property :inside-header '(:cache no))\n     (om-set-property :arguments '(\"x=4\"))\n     (om-set-property :end-header '(:exports results))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: cthulhu[:cache no](x=4) :exports results\"\n;; Given the following contents:\n; call_kthulu()\n(-&gt;&gt; (om-parse-this-object)\n     (om-set-property :call \"cthulhu\")\n     (om-set-property :inside-header '(:cache no))\n     (om-set-property :arguments '(\"x=4\"))\n     (om-set-property :end-header '(:exports results))\n     (om-to-trimmed-string))\n ;; =&gt; \"call_cthulhu[:cache no](x=4)[:exports results]\"\n;; Given the following contents:\n; src_emacs{(print 'yeah-boi)}\n(-&gt;&gt; (om-parse-this-object)\n     (om-set-property :language \"python\")\n     (om-set-property :parameters '(:cache no))\n     (om-set-property :value \"print \\\"yeah boi\\\"\")\n     (om-to-trimmed-string))\n ;; =&gt; \"src_python[:cache no]{print \\\"yeah boi\\\"}\"\n;; Given the following contents:\n; - thing\n(-&gt;&gt; (om-parse-this-item)\n     (om-set-property :bullet 1)\n     (om-set-property :checkbox 'on)\n     (om-set-property :counter 2)\n     (om-set-property :tag '(\"tmsu\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"1. [@2] [X] tmsu :: thing\"\n;; Given the following contents:\n; * not valuable\n;; Throw error when setting a property that doesn't exist\n(-&gt;&gt; (om-parse-this-headline)\n     (om-set-property :value \"wtf\")\n     (om-to-trimmed-string))\nError\n;; Throw error when setting to an improper type\n(-&gt;&gt; (om-parse-this-headline)\n     (om-set-property :title 666)\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-set-properties-plist-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-set-properties-plist-node\"><\/a>om-set-properties <code>(plist node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with all properties set to the values according to <strong><code>plist<\/code><\/strong>.<\/p><p><strong><code>plist<\/code><\/strong> is a list of property-value pairs that corresponds to the property list in <strong><code>node<\/code><\/strong>.<\/p><p>See builder functions for a list of properties and their rules for each type.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - thing\n(-&gt;&gt; (om-parse-this-item)\n     (om-set-properties (list :bullet 1 :checkbox 'on\n                               :counter 2 :tag '(\"tmsu\")))\n     (om-to-trimmed-string))\n ;; =&gt; \"1. [@2] [X] tmsu :: thing\"\n<\/pre><\/div><h4><a id=\"user-content-om-get-property-prop-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-get-property-prop-node\"><\/a>om-get-property <code>(prop node)<\/code><\/h4><p>Return the value of <strong><code>prop<\/code><\/strong> of <strong><code>node<\/code><\/strong>.<\/p><p>See builder functions for a list of properties and their rules for each type.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu(x=4) :exports results\n(-&gt;&gt; (om-parse-this-element)\n     (om-get-property :call))\n ;; =&gt; \"ktulu\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-get-property :inside-header))\n ;; =&gt; nil\n;; Given the following contents:\n; * not arguable\n;; Throw error when requesting a property that doesn't exist\n(-&gt;&gt; (om-parse-this-headline)\n     (om-get-property :value))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-map-property-prop-fun-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-map-property-prop-fun-node\"><\/a>om-map-property <code>(prop fun node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>fun<\/code><\/strong> applied to the value of <strong><code>prop<\/code><\/strong>.<\/p><p><strong><code>fun<\/code><\/strong> is a unary function which takes the current value of <strong><code>prop<\/code><\/strong> and returns a new value to which <strong><code>prop<\/code><\/strong> will be set.<\/p><p>See builder functions for a list of properties and their rules for each type.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; ~learn to~\n(-&gt;&gt; (om-parse-this-object)\n     (om-map-property :value (function s-upcase))\n     (om-to-trimmed-string))\n ;; =&gt; \"~LEARN TO~\"\n;; Throw error if property doesn't exist\n(-&gt;&gt; (om-parse-this-object)\n     (om-map-property :title (function s-upcase))\n     (om-to-trimmed-string))\nError\n;; Throw error if function doesn't return proper type\n(-&gt;&gt; (om-parse-this-object)\n     (om-map-property* :value (if it 1 0))\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-map-properties-plist-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-map-properties-plist-node\"><\/a>om-map-properties <code>(plist node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with functions applied to the values of properties.<\/p><p><strong><code>plist<\/code><\/strong> is a property list where the keys are properties of <strong><code>node<\/code><\/strong> and its values are unary functions to be mapped to these properties.<\/p><p>See builder functions for a list of properties and their rules for each type.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+KEY: VAL\n(-&gt;&gt; (om-parse-this-element)\n     (om-map-properties (list :key (-partial (function s-prepend)\n                                             \"OM_\")\n                               :value (-partial (function s-prepend)\n                                                \"OM_\")))\n     (om-to-trimmed-string))\n ;; =&gt; \"#+OM_KEY: OM_VAL\"\n<\/pre><\/div><h4><a id=\"user-content-om-toggle-property-prop-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-toggle-property-prop-node\"><\/a>om-toggle-property <code>(prop node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with the value of <strong><code>prop<\/code><\/strong> flipped.<\/p><p>This function only applies to properties that are booleans.<\/p><p>The following types and properties are supported:<\/p><p>entity<\/p><ul><li>:use-brackets-p<\/li>\n<\/ul><p>example-block<\/p><ul><li>:preserve-indent<\/li>\n<\/ul><p>headline<\/p><ul><li>:archivedp<\/li>\n<li>:commentedp<\/li>\n<li>:footnote-section-p<\/li>\n<\/ul><p>src-block<\/p><ul><li>:preserve-indent<\/li>\n<\/ul><p>subscript<\/p><ul><li>:use-brackets-p<\/li>\n<\/ul><p>superscript<\/p><ul><li>:use-brackets-p<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; \\pi\n(-&gt;&gt; (om-parse-this-object)\n     (om-toggle-property :use-brackets-p)\n     (om-to-trimmed-string))\n ;; =&gt; \"\\\\pi{}\"\n;; Given the following contents:\n; - [ ] nope\n;; Throw an error when trying to toggle a non-boolean property\n(-&gt;&gt; (om-parse-this-item)\n     (om-toggle-property :checkbox)\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-shift-property-prop-n-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-shift-property-prop-n-node\"><\/a>om-shift-property <code>(prop n node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>prop<\/code><\/strong> shifted by <strong><code>n<\/code><\/strong> (an integer).<\/p><p>This only applies the properties that are represented as integers.<\/p><p>The following types and properties are supported:<\/p><p>all elements<\/p><ul><li>:post-blank<\/li>\n<\/ul><p>headline<\/p><ul><li>:level<\/li>\n<li>:pre-blank<\/li>\n<li>:priority<\/li>\n<\/ul><p>item<\/p><ul><li>:counter<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * no priorities\n;; Do nothing if there is nothing to shift.\n(-&gt;&gt; (om-parse-this-headline)\n     (om-shift-property :priority 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* no priorities\"\n;; Given the following contents:\n; * [#A] priorities\n(-&gt;&gt; (om-parse-this-headline)\n     (om-shift-property :priority -1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* [#B] priorities\"\n;; Wrap priority around when crossing the min or max\n(-&gt;&gt; (om-parse-this-headline)\n     (om-shift-property :priority 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* [#C] priorities\"\n;; Given the following contents:\n; * TODO or not todo\n;; Throw error when shifting an unshiftable property\n(-&gt;&gt; (om-parse-this-headline)\n     (om-shift-property :todo-keyword 1)\n     (om-to-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-insert-into-property-prop-index-string-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-insert-into-property-prop-index-string-node\"><\/a>om-insert-into-property <code>(prop index string node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>string<\/code><\/strong> inserted at <strong><code>index<\/code><\/strong> into <strong><code>prop<\/code><\/strong>.<\/p><p>This only applies to properties that are represented as lists of strings.<\/p><p>The following types and properties are supported:<\/p><p>babel-call<\/p><ul><li>:arguments<\/li>\n<\/ul><p>example-block<\/p><ul><li>:switches<\/li>\n<\/ul><p>headline<\/p><p>inline-babel-call<\/p><ul><li>:arguments<\/li>\n<\/ul><p>macro<\/p><p>src-block<\/p><ul><li>:switches<\/li>\n<\/ul><p>table<\/p><ul><li>:tblfm<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu(y=1)\n(-&gt;&gt; (om-parse-this-element)\n     (om-insert-into-property :arguments 0 \"x=4\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu(x=4,y=1)\"\n;; Do nothing if the string is already in the list\n(-&gt;&gt; (om-parse-this-element)\n     (om-insert-into-property :arguments 0 \"y=1\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu(y=1)\"\n;; Throw error when inserting into a property that is not a list of strings\n(-&gt;&gt; (om-parse-this-element)\n     (om-insert-into-property :end-header 0 \"html\")\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-remove-from-property-prop-string-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-remove-from-property-prop-string-node\"><\/a>om-remove-from-property <code>(prop string node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>string<\/code><\/strong> removed from <strong><code>prop<\/code><\/strong> if present.<\/p><p>This only applies to properties that are represented as lists of strings.<\/p><p>See <a href=\"#om-insert-into-property-prop-index-string-node\"><code>om-insert-into-property<\/code><\/a> for a list of supported elements and properties that may be used with this function.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu(y=1)\n(-&gt;&gt; (om-parse-this-element)\n     (om-remove-from-property :arguments \"y=1\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu()\"\n;; Do nothing if the string does not exist\n(-&gt;&gt; (om-parse-this-element)\n     (om-remove-from-property :arguments \"d=666\")\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu(y=1)\"\n;; Throw error when removing from property that is not a string list\n(-&gt;&gt; (om-parse-this-element)\n     (om-remove-from-property :end-header \":results\")\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-plist-put-property-prop-key-value-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plist-put-property-prop-key-value-node\"><\/a>om-plist-put-property <code>(prop key value node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>value<\/code><\/strong> corresponding to <strong><code>key<\/code><\/strong> inserted into <strong><code>prop<\/code><\/strong>.<\/p><p><strong><code>key<\/code><\/strong> is a keyword and <strong><code>value<\/code><\/strong> is a symbol. This only applies to properties that are represented as plists.<\/p><p>The following types and properties are supported:<\/p><p>babel-call<\/p><ul><li>:inside-header<\/li>\n<li>:end-header<\/li>\n<\/ul><p>dynamic-block<\/p><ul><li>:arguments<\/li>\n<\/ul><p>inline-babel-call<\/p><ul><li>:inside-header<\/li>\n<li>:end-header<\/li>\n<\/ul><p>inline-src-block<\/p><ul><li>:parameters<\/li>\n<\/ul><p>src-block<\/p><ul><li>:parameters<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu[:cache no]()\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-put-property :end-header :results 'html)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu[:cache no]() :results html\"\n;; Change the value of key if it already is present\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-put-property :inside-header :cache 'yes)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu[:cache yes]()\"\n;; Do nothing if the key and value already exist\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-put-property :inside-header :cache 'no)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu[:cache no]()\"\n;; Throw error if setting property that isn't a plist\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-put-property :arguments :cache 'no)\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-plist-remove-property-prop-key-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plist-remove-property-prop-key-node\"><\/a>om-plist-remove-property <code>(prop key node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>key<\/code><\/strong> and its corresponding value removed from <strong><code>prop<\/code><\/strong>.<\/p><p><strong><code>key<\/code><\/strong> is a keyword. This only applies to properties that are represented as plists.<\/p><p>See <a href=\"#om-plist-put-property-prop-key-value-node\"><code>om-plist-put-property<\/code><\/a> for a list of supported elements and properties that may be used with this function.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu() :results html\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-remove-property :end-header :results)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu()\"\n;; Do nothing if the key is not present\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-remove-property :inside-header :cache)\n     (om-to-trimmed-string))\n ;; =&gt; \"#+CALL: ktulu() :results html\"\n;; Throw error if trying to remove key from non-plist property\n(-&gt;&gt; (om-parse-this-element)\n     (om-plist-remove-property :arguments :cache)\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h3><a id=\"user-content-clock-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#clock-1\"><\/a>Clock<\/h3><h4><a id=\"user-content-om-clock-is-running-clock\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-clock-is-running-clock\"><\/a>om-clock-is-running <code>(clock)<\/code><\/h4><p>Return t if <strong><code>clock<\/code><\/strong> element is running (eg is open).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; CLOCK: [2019-01-01 Tue 00:00]\n(-&gt;&gt; (om-parse-this-element)\n     (om-clock-is-running))\n ;; =&gt; t\n;; Given the following contents:\n; CLOCK: [2019-01-01 Tue 00:00]--[2019-01-02 Wed 00:00] =&gt; 24:00\n(-&gt;&gt; (om-parse-this-element)\n     (om-clock-is-running))\n ;; =&gt; nil\n<\/pre><\/div><h3><a id=\"user-content-entity-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#entity-1\"><\/a>Entity<\/h3><h4><a id=\"user-content-om-entity-get-replacement-key-entity\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-entity-get-replacement-key-entity\"><\/a>om-entity-get-replacement <code>(key entity)<\/code><\/h4><p>Return replacement string or symbol for <strong><code>entity<\/code><\/strong> node.<\/p><p><strong><code>key<\/code><\/strong> is one of:<\/p><ul><li><code>:latex<\/code> (the entity's latex representation)<\/li>\n<li><code>:latex-math-p<\/code> (t if the latex representation requires math mode, nil otherwise)<\/li>\n<li><code>:html<\/code> (the entity's html representation)<\/li>\n<li><code>:ascii<\/code> (the entity's ascii representation)<\/li>\n<li><code>:latin1<\/code> (the entity's Latin1 representation)<\/li>\n<li><code>:utf-8<\/code> (the entity's utf8 representation)<\/li>\n<\/ul><p>Any other keys will trigger an error.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; \\pi{}\n(-&gt;&gt; (om-parse-this-object)\n     (om-entity-get-replacement :latex))\n ;; =&gt; \"\\\\pi\"\n(-&gt;&gt; (om-parse-this-object)\n     (om-entity-get-replacement :latex-math-p))\n ;; =&gt; t\n(-&gt;&gt; (om-parse-this-object)\n     (om-entity-get-replacement :html))\n ;; =&gt; \"&amp;pi;\"\n(-&gt;&gt; (om-parse-this-object)\n     (om-entity-get-replacement :ascii))\n ;; =&gt; \"pi\"\n(-&gt;&gt; (om-parse-this-object)\n     (om-entity-get-replacement :latin1))\n ;; =&gt; \"pi\"\n(-&gt;&gt; (om-parse-this-object)\n     (om-entity-get-replacement :utf-8))\n ;; =&gt; \"\u03c0\"\n<\/pre><\/div><h3><a id=\"user-content-headline-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#headline-2\"><\/a>Headline<\/h3><h4><a id=\"user-content-om-headline-set-title-title-text-stats-cookie-value-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-set-title-title-text-stats-cookie-value-headline\"><\/a>om-headline-set-title! <code>(title-text stats-cookie-value headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with title set with <strong><code>title-text<\/code><\/strong> and <strong><code>stats-cookie-value<\/code><\/strong>.<\/p><p><strong><code>title-text<\/code><\/strong> is a string to be parsed into object nodes for the title via <a href=\"#om-build-secondary-string-string\"><code>om-build-secondary-string!<\/code><\/a> (see that function for restrictions) and <strong><code>stats-cookie-value<\/code><\/strong> is a list described in <a href=\"#om-build-statistics-cookie-value-key-post-blank\"><code>om-build-statistics-cookie<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * really impressive title\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-set-title! \"really *impressive* title\" '(2 3))\n     (om-to-trimmed-string))\n ;; =&gt; \"* really *impressive* title [2\/3]\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-is-done-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-is-done-headline\"><\/a>om-headline-is-done <code>(headline)<\/code><\/h4><p>Return t if <strong><code>headline<\/code><\/strong> node has a done todo-keyword.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * TODO darn\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-is-done))\n ;; =&gt; nil\n;; Given the following contents:\n; * DONE yay\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-is-done))\n ;; =&gt; t\n<\/pre><\/div><h4><a id=\"user-content-om-headline-has-tag-tag-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-has-tag-tag-headline\"><\/a>om-headline-has-tag <code>(tag headline)<\/code><\/h4><p>Return t if <strong><code>headline<\/code><\/strong> node is tagged with <strong><code>tag<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * dummy\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-has-tag \"tmsu\"))\n ;; =&gt; nil\n;; Given the following contents:\n; * dummy                  :tmsu:\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-has-tag \"tmsu\"))\n ;; =&gt; t\n<\/pre><\/div><h4><a id=\"user-content-om-headline-get-statistics-cookie-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-statistics-cookie-headline\"><\/a>om-headline-get-statistics-cookie <code>(headline)<\/code><\/h4><p>Return the statistics cookie node from <strong><code>headline<\/code><\/strong> if it exists.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * statistically significant [10\/10]\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-statistics-cookie)\n     (om-to-string))\n ;; =&gt; \"[10\/10]\"\n;; Given the following contents:\n; * not statistically significant\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-statistics-cookie))\n ;; =&gt; nil\n<\/pre><\/div><h3><a id=\"user-content-item-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#item-1\"><\/a>Item<\/h3><h4><a id=\"user-content-om-item-toggle-checkbox-item\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-item-toggle-checkbox-item\"><\/a>om-item-toggle-checkbox <code>(item)<\/code><\/h4><p>Return <strong><code>item<\/code><\/strong> node with its checkbox state flipped. This only affects item nodes with checkboxes in the <code>on<\/code> or <code>off<\/code> states; return <strong><code>item<\/code><\/strong> node unchanged if the checkbox property is <code>trans<\/code> or nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - [ ] one\n(-&gt;&gt; (om-parse-this-item)\n     (om-item-toggle-checkbox)\n     (om-to-trimmed-string))\n ;; =&gt; \"- [X] one\"\n;; Given the following contents:\n; - [-] one\n;; Ignore trans state checkboxes\n(-&gt;&gt; (om-parse-this-item)\n     (om-item-toggle-checkbox)\n     (om-to-trimmed-string))\n ;; =&gt; \"- [-] one\"\n;; Given the following contents:\n; - one\n;; Do nothing if there is no checkbox\n(-&gt;&gt; (om-parse-this-item)\n     (om-item-toggle-checkbox)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\"\n<\/pre><\/div><h3><a id=\"user-content-planning-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#planning-1\"><\/a>Planning<\/h3><h4><a id=\"user-content-om-planning-set-timestamp-prop-planning-list-planning\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-planning-set-timestamp-prop-planning-list-planning\"><\/a>om-planning-set-timestamp! <code>(prop planning-list planning)<\/code><\/h4><p>Return <strong><code>planning<\/code><\/strong> node with <strong><code>prop<\/code><\/strong> set to <strong><code>planning-list<\/code><\/strong>.<\/p><p><strong><code>prop<\/code><\/strong> is one of <code>:closed<\/code>, <code>:deadline<\/code>, or <code>:scheduled<\/code>. <strong><code>planning-list<\/code><\/strong> is the same as that described in <a href=\"#om-build-planning-key-closed-deadline-scheduled-post-blank\"><code>om-build-planning!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * dummy\n; CLOSED: [2019-01-01 Tue]\n;; Change an existing timestamp in planning\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-planning)\n     (om-planning-set-timestamp! :closed '(2019 1 2 &amp;warning all 1 day &amp;repeater cumulate 2 month))\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOSED: [2019-01-02 Wed +2m -1d]\"\n;; Add a new timestamp and remove another\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-planning)\n     (om-planning-set-timestamp! :deadline '(2112 1 1))\n     (om-planning-set-timestamp! :closed nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"DEADLINE: [2112-01-01 Fri]\"\n<\/pre><\/div><h3><a id=\"user-content-statistics-cookie-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#statistics-cookie-1\"><\/a>Statistics Cookie<\/h3><h4><a id=\"user-content-om-statistics-cookie-is-complete-statistics-cookie\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-statistics-cookie-is-complete-statistics-cookie\"><\/a>om-statistics-cookie-is-complete <code>(statistics-cookie)<\/code><\/h4><p>Return t is <strong><code>statistics-cookie<\/code><\/strong> node is complete.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * statistically significant [10\/10]\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-statistics-cookie)\n     (om-statistics-cookie-is-complete))\n ;; =&gt; t\n;; Given the following contents:\n; * statistically significant [1\/10]\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-statistics-cookie)\n     (om-statistics-cookie-is-complete))\n ;; =&gt; nil\n;; Given the following contents:\n; * statistically significant [100%]\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-statistics-cookie)\n     (om-statistics-cookie-is-complete))\n ;; =&gt; t\n;; Given the following contents:\n; * statistically significant [33%]\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-statistics-cookie)\n     (om-statistics-cookie-is-complete))\n ;; =&gt; nil\n<\/pre><\/div><h3><a id=\"user-content-timestamp-auxiliary-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#timestamp-auxiliary-1\"><\/a>Timestamp (Auxiliary)<\/h3><p>Functions to work with timestamp data<\/p><h4><a id=\"user-content-om-time-to-unixtime-time\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-time-to-unixtime-time\"><\/a>om-time-to-unixtime <code>(time)<\/code><\/h4><p>Return the unix time (integer seconds) of time list <strong><code>time<\/code><\/strong>. The returned value is dependent on the time zone of the operating system.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div><h4><a id=\"user-content-om-unixtime-to-time-long-unixtime\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-unixtime-to-time-long-unixtime\"><\/a>om-unixtime-to-time-long <code>(unixtime)<\/code><\/h4><p>Return the long time list of <strong><code>unixtime<\/code><\/strong>. The list will be formatted like <code>(year month day hour min)<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div><h4><a id=\"user-content-om-unixtime-to-time-short-unixtime\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-unixtime-to-time-short-unixtime\"><\/a>om-unixtime-to-time-short <code>(unixtime)<\/code><\/h4><p>Return the short time list of <strong><code>unixtime<\/code><\/strong>. The list will be formatted like <code>(year month day nil nil)<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div><h3><a id=\"user-content-timestamp-standard-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#timestamp-standard-1\"><\/a>Timestamp (Standard)<\/h3><h4><a id=\"user-content-om-timestamp-get-start-time-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-get-start-time-timestamp\"><\/a>om-timestamp-get-start-time <code>(timestamp)<\/code><\/h4><p>Return the time list for start time of <strong><code>timestamp<\/code><\/strong> node. The return value will be a list as specified by the <code>time<\/code> argument in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-start-time))\n ;; =&gt; '(2019 1 1 nil nil)\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-start-time))\n ;; =&gt; '(2019 1 1 nil nil)\n;; Given the following contents:\n; [2019-01-01 Tue 00:00-12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-start-time))\n ;; =&gt; '(2019 1 1 0 0)\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-get-end-time-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-get-end-time-timestamp\"><\/a>om-timestamp-get-end-time <code>(timestamp)<\/code><\/h4><p>Return the end time list for end time of <strong><code>timestamp<\/code><\/strong> or nil if not a range. The return value will be a list as specified by the <code>time<\/code> argument in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-end-time))\n ;; =&gt; nil\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-end-time))\n ;; =&gt; '(2019 1 2 nil nil)\n;; Given the following contents:\n; [2019-01-01 Tue 00:00-12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-end-time))\n ;; =&gt; '(2019 1 1 12 0)\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-get-range-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-get-range-timestamp\"><\/a>om-timestamp-get-range <code>(timestamp)<\/code><\/h4><p>Return the range of <strong><code>timestamp<\/code><\/strong> node in seconds as an integer. If non-ranged, this function will return 0. If ranged but the start time is in the future relative to end the time, return a negative integer.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-range))\n ;; =&gt; 0\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-range))\n ;; =&gt; 86400\n;; Given the following contents:\n; [2019-01-01 Tue 00:00-12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-get-range))\n ;; =&gt; 43200\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-is-active-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-is-active-timestamp\"><\/a>om-timestamp-is-active <code>(timestamp)<\/code><\/h4><p>Return t if <strong><code>timestamp<\/code><\/strong> node is active.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; &lt;2019-01-01 Tue&gt;\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-is-active))\n ;; =&gt; t\n;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-is-active))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-is-ranged-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-is-ranged-timestamp\"><\/a>om-timestamp-is-ranged <code>(timestamp)<\/code><\/h4><p>Return t if <strong><code>timestamp<\/code><\/strong> node is ranged.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-is-ranged))\n ;; =&gt; t\n;; Given the following contents:\n; [2019-01-01 Tue 00:00-12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-is-ranged))\n ;; =&gt; t\n;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-is-ranged))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-range-contains-p-unixtime-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-range-contains-p-unixtime-timestamp\"><\/a>om-timestamp-range-contains-p <code>(unixtime timestamp)<\/code><\/h4><p>Return t if <strong><code>unixtime<\/code><\/strong> is between start and end time of <strong><code>timestamp<\/code><\/strong> node. The boundaries are inclusive. If <strong><code>timestamp<\/code><\/strong> has a range of zero, then only return t if <strong><code>unixtime<\/code><\/strong> is the same as <strong><code>timestamp<\/code><\/strong>. <strong><code>timestamp<\/code><\/strong> will be interpreted according to the localtime of the operating system.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue 00:00]\n(let ((ut (om-time-to-unixtime '(2019 1 1 0 0))))\n  (-&gt;&gt; (om-parse-this-object)\n       (om-timestamp-range-contains-p ut)))\n ;; =&gt; t\n(let ((ut (om-time-to-unixtime '(2019 1 1 0 30))))\n  (-&gt;&gt; (om-parse-this-object)\n       (om-timestamp-range-contains-p ut)))\n ;; =&gt; nil\n;; Given the following contents:\n; [2019-01-01 Tue 00:00-01:00]\n(let ((ut (om-time-to-unixtime '(2019 1 1 0 30))))\n  (-&gt;&gt; (om-parse-this-object)\n       (om-timestamp-range-contains-p ut)))\n ;; =&gt; t\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-collapsed-flag-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-collapsed-flag-timestamp\"><\/a>om-timestamp-set-collapsed <code>(flag timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> with collapsed set to <strong><code>flag<\/code><\/strong>.<\/p><p>If timestamp is ranged but not outside of one day, it may be collapsed (<strong><code>flag<\/code><\/strong> is t) to short format like [yyyy-mm-dd xxx hh:mm-hh:mm] or expanded (<strong><code>flag<\/code><\/strong> is nil) to long format like [yyyy-mm-dd xxx hh:mm]--[yyyy-mm-dd xxx hh:mm]. If these conditions are not met, return <strong><code>timestamp<\/code><\/strong> untouched regardless of <strong><code>flag<\/code><\/strong>.<\/p><p>Note: the default for all timestamp functions in <code>om.el<\/code> is to favor collapsed format.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue 12:00-13:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-collapsed nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:00]--[2019-01-01 Tue 13:00]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00-13:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-collapsed nil)\n     (om-timestamp-set-collapsed t)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:00-13:00]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-collapsed nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:00]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-collapsed nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-start-time-time-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-start-time-time-timestamp\"><\/a>om-timestamp-set-start-time <code>(time timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with start time set to <strong><code>time<\/code><\/strong>. <strong><code>time<\/code><\/strong> is a list analogous to the same argument specified in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-02 Wed]\n;; If not a range this will turn into a range by moving only the start time.\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-start-time '(2019 1 1))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n;; Set a different time with different precision.\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-start-time '(2019 1 1 10 0))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 10:00]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-02 Wed 12:00]\n;; If not a range and set within a day, use short format\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-start-time '(2019 1 1 0 0))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 00:00-12:00]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-end-time-time-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-end-time-time-timestamp\"><\/a>om-timestamp-set-end-time <code>(time timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with end time set to <strong><code>time<\/code><\/strong>. <strong><code>time<\/code><\/strong> is a list analogous to the same argument specified in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n;; Add the end time\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-end-time '(2019 1 2))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n;; Remove the end time\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-end-time nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00]\n;; Use short range format\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-end-time '(2019 1 1 13 0))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:00-13:00]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-single-time-time-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-single-time-time-timestamp\"><\/a>om-timestamp-set-single-time <code>(time timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with start and end times set to <strong><code>time<\/code><\/strong>. <strong><code>time<\/code><\/strong> is a list analogous to the same argument specified in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n;; Don't make a range\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-single-time '(2019 1 2))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n;; Output is not a range despite input being ranged\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-single-time '(2019 1 3))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-03 Thu]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-double-time-time1-time2-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-double-time-time1-time2-timestamp\"><\/a>om-timestamp-set-double-time <code>(time1 time2 timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with start\/end times set to <strong><code>time1<\/code><\/strong>\/<strong><code>time2<\/code><\/strong> respectively. <strong><code>time1<\/code><\/strong> and <strong><code>time2<\/code><\/strong> are lists analogous to the <code>time<\/code> argument specified in <a href=\"#om-build-timestamp-start-key-end-active-repeater-warning-post-blank\"><code>om-build-timestamp!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n;; Make a range\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-double-time '(2019 1 2)\n                                   '(2019 1 3))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-02 Wed]--[2019-01-03 Thu]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-03 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-double-time '(2019 1 4)\n                                   '(2019 1 5))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-04 Fri]--[2019-01-05 Sat]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-03 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-double-time '(2019 1 1 0 0)\n                                   '(2019 1 1 1 0))\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 00:00-01:00]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-range-range-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-range-range-timestamp\"><\/a>om-timestamp-set-range <code>(range timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with range set to <strong><code>range<\/code><\/strong>. If <strong><code>timestamp<\/code><\/strong> is ranged, keep start time the same and adjust the end time. If not, make a new end time. The units for <strong><code>range<\/code><\/strong> are in minutes if <strong><code>timestamp<\/code><\/strong> is in long format and days if <strong><code>timestamp<\/code><\/strong> is in short format.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n;; Use days as the unit for short format\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-range 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue 00:00]\n;; Use minutes as the unit for long format\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-range 3)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 00:00-00:03]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-03 Wed]\n;; Set range to 0 to remove end time\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-range 0)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-set-active-flag-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-set-active-flag-timestamp\"><\/a>om-timestamp-set-active <code>(flag timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with active type if <strong><code>flag<\/code><\/strong> is t.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-active t)\n     (om-to-trimmed-string))\n ;; =&gt; \"&lt;2019-01-01 Tue&gt;\"\n;; Given the following contents:\n; &lt;2019-01-01 Tue&gt;\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-set-active nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-shift-n-unit-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-shift-n-unit-timestamp\"><\/a>om-timestamp-shift <code>(n unit timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with time shifted by <strong><code>n<\/code><\/strong> <strong><code>unit<\/code><\/strong>`s.<\/p><p>This function will move the start and end times together; therefore ranged inputs will always output ranged timestamps and same for non-ranged. To move the start and end time independently, use <a href=\"#om-timestamp-shift-start-n-unit-timestamp\"><code>om-timestamp-shift-start<\/code><\/a> or <a href=\"#om-timestamp-shift-end-n-unit-timestamp\"><code>om-timestamp-shift-end<\/code><\/a>.<\/p><p><strong><code>n<\/code><\/strong> is a positive or negative integer and <strong><code>unit<\/code><\/strong> is one of <code>minute<\/code>, <code>hour<\/code>, <code>day<\/code>, <code>month<\/code>, or <code>year<\/code>. Overflows will wrap around transparently; for instance, supplying <code>minute<\/code> for <strong><code>unit<\/code><\/strong> and 90 for <strong><code>n<\/code><\/strong> will increase the hour property by 1 and the minute property by 30.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue 12:00]\n;; Change each unit, and wrap around to the next unit as needed.\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift 30 'minute)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:30]\"\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift 13 'month)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2020-02-01 Sat 12:00]\"\n;; Given the following contents:\n; [2019-01-01 Tue]\n;; Error when shifting hour\/minute in short format\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift 30 'minute)\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-shift-start-n-unit-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-shift-start-n-unit-timestamp\"><\/a>om-timestamp-shift-start <code>(n unit timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with start time shifted by <strong><code>n<\/code><\/strong> <strong><code>unit<\/code><\/strong>`s.<\/p><p><strong><code>n<\/code><\/strong> and <strong><code>unit<\/code><\/strong> behave the same as those in <a href=\"#om-timestamp-shift-n-unit-timestamp\"><code>om-timestamp-shift<\/code><\/a>.<\/p><p>If <strong><code>timestamp<\/code><\/strong> is not range, the output will be a ranged timestamp with the shifted start time and the end time as that of <strong><code>timestamp<\/code><\/strong>. If this behavior is not desired, use <a href=\"#om-timestamp-shift-n-unit-timestamp\"><code>om-timestamp-shift<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue 12:00]\n;; If not a range, change start time and leave implicit end time.\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift-start -1 'year)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2018-01-01 Mon 12:00]--[2019-01-01 Tue 12:00]\"\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift-start -1 'hour)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 11:00-12:00]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-03 Thu]\n;; Change only start time if a range\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift-start 1 'day)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-02 Wed]--[2019-01-03 Thu]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-shift-end-n-unit-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-shift-end-n-unit-timestamp\"><\/a>om-timestamp-shift-end <code>(n unit timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with end time shifted by <strong><code>n<\/code><\/strong> <strong><code>unit<\/code><\/strong>`s.<\/p><p><strong><code>n<\/code><\/strong> and <strong><code>unit<\/code><\/strong> behave the same as those in <a href=\"#om-timestamp-shift-n-unit-timestamp\"><code>om-timestamp-shift<\/code><\/a>.<\/p><p>If <strong><code>timestamp<\/code><\/strong> is not range, the output will be a ranged timestamp with the shifted end time and the start time as that of <strong><code>timestamp<\/code><\/strong>. If this behavior is not desired, use <a href=\"#om-timestamp-shift-n-unit-timestamp\"><code>om-timestamp-shift<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n;; Shift implicit end time if not a range.\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift-end 1 'day)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n;; Move only the second time if a range.\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-shift-end 1 'day)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-03 Thu]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-toggle-active-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-toggle-active-timestamp\"><\/a>om-timestamp-toggle-active <code>(timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with its active\/inactive type flipped.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-toggle-active)\n     (om-to-trimmed-string))\n ;; =&gt; \"&lt;2019-01-01 Tue&gt;\"\n;; Given the following contents:\n; &lt;2019-01-01 Tue&gt;--&lt;2019-01-02 Wed&gt;\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-toggle-active)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-truncate-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-truncate-timestamp\"><\/a>om-timestamp-truncate <code>(timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with start\/end times forced to short format.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00]--[2019-01-02 Wed 13:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-truncate-start-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-truncate-start-timestamp\"><\/a>om-timestamp-truncate-start <code>(timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with start time forced to short format.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue 12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate-start)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00]--[2019-01-02 Wed 12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate-start)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed 12:00]\"\n;; Given the following contents:\n; [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate-start)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]\"\n<\/pre><\/div><h4><a id=\"user-content-om-timestamp-truncate-end-timestamp\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-truncate-end-timestamp\"><\/a>om-timestamp-truncate-end <code>(timestamp)<\/code><\/h4><p>Return <strong><code>timestamp<\/code><\/strong> node with end time forced to short format.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [2019-01-01 Tue]--[2019-01-02 Wed]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate-end)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00]--[2019-01-02 Wed 13:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate-end)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:00]--[2019-01-02 Wed]\"\n;; Given the following contents:\n; [2019-01-01 Tue 12:00]\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-truncate-end)\n     (om-to-trimmed-string))\n ;; =&gt; \"[2019-01-01 Tue 12:00]\"\n<\/pre><\/div><h3><a id=\"user-content-timestamp-diary-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#timestamp-diary-1\"><\/a>Timestamp (diary)<\/h3><h4><a id=\"user-content-om-timestamp-diary-set-value-form-timestamp-diary\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-timestamp-diary-set-value-form-timestamp-diary\"><\/a>om-timestamp-diary-set-value <code>(form timestamp-diary)<\/code><\/h4><p>Return <strong><code>timestamp-diary<\/code><\/strong> node with value set to <strong><code>form<\/code><\/strong>. <code>timestamp<\/code> must have a type <code>eq<\/code> to <code>diary<\/code>. <strong><code>form<\/code><\/strong> is a quoted list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; &lt;%%(diary-float t 4 2)&gt;\n(-&gt;&gt; (om-parse-this-object)\n     (om-timestamp-diary-set-value '(diary-float 1 3 2))\n     (om-to-string))\n ;; =&gt; \"&lt;%%(diary-float 1 3 2)&gt;\"\n<\/pre><\/div><h2><a id=\"user-content-branchchild-manipulation-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#branchchild-manipulation-1\"><\/a>Branch\/Child Manipulation<\/h2><p>Set, get, and map the children of branch nodes.<\/p><h3><a id=\"user-content-polymorphic-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#polymorphic-1\"><\/a>Polymorphic<\/h3><h4><a id=\"user-content-om-children-contain-point-point-branch-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-children-contain-point-point-branch-node\"><\/a>om-children-contain-point <code>(point branch-node)<\/code><\/h4><p>Return t if <strong><code>point<\/code><\/strong> is within the boundaries of <strong><code>branch-node<\/code><\/strong>`s children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline\n; findme\n(-&gt;&gt; (om-parse-this-headline)\n     (om-children-contain-point 2))\n ;; =&gt; nil\n(-&gt;&gt; (om-parse-this-headline)\n     (om-children-contain-point 15))\n ;; =&gt; t\n<\/pre><\/div><h4><a id=\"user-content-om-get-children-branch-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-get-children-branch-node\"><\/a>om-get-children <code>(branch-node)<\/code><\/h4><p>Return the children of <strong><code>branch-node<\/code><\/strong> as a list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; \/this\/ is a *paragraph*\n;; Return child nodes for branch nodes\n(-&gt;&gt; (om-parse-this-element)\n     (om-get-children)\n     (-map (function om-get-type)))\n ;; =&gt; '(italic plain-text bold)\n;; Given the following contents:\n; * headline\n;; Return nil if no children\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-get-children)\n     (-map (function om-get-type)))\n ;; =&gt; nil\n;; Given the following contents:\n; #+CALL: ktulu()\n;; Throw error when attempting to get contents of a non-branch node\n(-&gt;&gt; (om-parse-this-element)\n     (om-get-children)\n     (-map (function om-get-type)))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-set-children-children-branch-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-set-children-children-branch-node\"><\/a>om-set-children <code>(children branch-node)<\/code><\/h4><p>Return <strong><code>branch-node<\/code><\/strong> with its children set to <strong><code>children<\/code><\/strong>. <strong><code>children<\/code><\/strong> is a list of nodes; the types permitted in this list depend on the type of <code>node<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; \/this\/ is a *paragraph*\n;; Set children for branch object\n(-&gt;&gt; (om-parse-this-element)\n     (om-set-children (list \"this is lame\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"this is lame\"\n;; Given the following contents:\n; * headline\n;; Set children for branch element nodes\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-set-children (list (om-build-headline! :title-text \"only me\" :level 2)))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      ** only me\"\n;; Given the following contents:\n; #+CALL: ktulu()\n;; Throw error when attempting to set children of a non-branch nodes\n(-&gt;&gt; (om-parse-this-element)\n     (om-set-children \"nil by mouth\")\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-map-children-fun-branch-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-map-children-fun-branch-node\"><\/a>om-map-children <code>(fun branch-node)<\/code><\/h4><p>Return <strong><code>branch-node<\/code><\/strong> with <strong><code>fun<\/code><\/strong> applied to its children. <strong><code>fun<\/code><\/strong> is a unary function that takes the current list of children and returns a modified list of children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; \/this\/ is a *paragraph*\n(-&gt;&gt; (om-parse-this-element)\n     (om-map-children (lambda (objs)\n                        (append objs (list \" ...yeah\"))))\n     (om-to-trimmed-string))\n ;; =&gt; \"\/this\/ is a *paragraph* ...yeah\"\n;; Given the following contents:\n; * headline\n; ** subheadline\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-map-children* (--map (om-shift-property :level 1 it)\n                              it))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline\n ;      *** subheadline\"\n;; Given the following contents:\n; #+CALL: ktulu()\n;; Throw error when attempting to map children of a non-branch node\n(-&gt;&gt; (om-parse-this-element)\n     (om-map-children (function ignore))\n     (om-to-trimmed-string))\nError\n<\/pre><\/div><h4><a id=\"user-content-om-is-childless-branch-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-is-childless-branch-node\"><\/a>om-is-childless <code>(branch-node)<\/code><\/h4><p>Return t if <strong><code>branch-node<\/code><\/strong> is empty.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * dummy\n; filled with useless knowledge\n(-&gt;&gt; (om-parse-this-headline)\n     (om-is-childless))\n ;; =&gt; nil\n;; Given the following contents:\n; * dummy\n(-&gt;&gt; (om-parse-this-headline)\n     (om-is-childless))\n ;; =&gt; t\n;; Given the following contents:\n; #+CALL: ktulu()\n;; Throw error when attempting to determine if non-branch node is empty\n(-&gt;&gt; (om-parse-this-element)\n     (om-is-childless))\nError\n<\/pre><\/div><h3><a id=\"user-content-object-nodes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#object-nodes-1\"><\/a>Object Nodes<\/h3><h4><a id=\"user-content-om-unwrap-object-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-unwrap-object-node\"><\/a>om-unwrap <code>(object-node)<\/code><\/h4><p>Return the children of <strong><code>object-node<\/code><\/strong> as a secondary string. If <strong><code>object-node<\/code><\/strong> is a plain-text node, wrap it in a list and return. Else add the post-blank property of <strong><code>object-node<\/code><\/strong> to the last member of its children and return children as a secondary string.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; _1 *2* 3 *\/4\/* 5 \/6\/_\n;; Remove the outer underline formatting\n(-&gt;&gt; (om-parse-this-object)\n     (om-unwrap)\n     (apply (function om-build-paragraph))\n     (om-to-trimmed-string))\n ;; =&gt; \"1 *2* 3 *\/4\/* 5 \/6\/\"\n<\/pre><\/div><h4><a id=\"user-content-om-unwrap-types-deep-types-object-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-unwrap-types-deep-types-object-node\"><\/a>om-unwrap-types-deep <code>(types object-node)<\/code><\/h4><p>Return the children of <strong><code>object-node<\/code><\/strong> as a secondary string. If <strong><code>object-node<\/code><\/strong> is a plain-text node, wrap it in a list and return. Else recursively descend into the children of <strong><code>object-node<\/code><\/strong> and splice the children of nodes with type in <strong><code>types<\/code><\/strong> in place of said node and return the result as a secondary string.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; _1 *2* 3 *\/4\/* 5 \/6\/_\n;; Remove bold formatting at any level\n(-&gt;&gt; (om-parse-this-object)\n     (om-unwrap-types-deep '(bold))\n     (apply (function om-build-paragraph))\n     (om-to-trimmed-string))\n ;; =&gt; \"_1 2 3 \/4\/ 5 \/6\/_\"\n<\/pre><\/div><h4><a id=\"user-content-om-unwrap-deep-object-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-unwrap-deep-object-node\"><\/a>om-unwrap-deep <code>(object-node)<\/code><\/h4><p>Return the children of <strong><code>object-node<\/code><\/strong> as plain-text wrapped in a list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; _1 *2* 3 *\/4\/* 5 \/6\/_\n;; Remove all formatting\n(-&gt;&gt; (om-parse-this-object)\n     (om-unwrap-deep)\n     (apply (function om-build-paragraph))\n     (om-to-trimmed-string))\n ;; =&gt; \"1 2 3 4 5 6\"\n<\/pre><\/div><h3><a id=\"user-content-secondary-strings-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#secondary-strings-1\"><\/a>Secondary Strings<\/h3><h4><a id=\"user-content-om-flatten-secondary-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-flatten-secondary-string\"><\/a>om-flatten <code>(secondary-string)<\/code><\/h4><p>Return <strong><code>secondary-string<\/code><\/strong> with its first level unwrapped. The unwrap operation will be done with <a href=\"#om-unwrap-object-node\"><code>om-unwrap<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; This (1 *2* 3 *\/4\/* 5 \/6\/) is randomly formatted\n;; Remove first level of formatting\n(-&gt;&gt; (om-parse-this-element)\n     (om-map-children (function om-flatten))\n     (om-to-trimmed-string))\n ;; =&gt; \"This (1 2 3 \/4\/ 5 6) is randomly formatted\"\n<\/pre><\/div><h4><a id=\"user-content-om-flatten-types-deep-types-secondary-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-flatten-types-deep-types-secondary-string\"><\/a>om-flatten-types-deep <code>(types secondary-string)<\/code><\/h4><p>Return <strong><code>secondary-string<\/code><\/strong> with object nodes in <strong><code>types<\/code><\/strong> unwrapped. The unwrap operation will be done with <a href=\"#om-unwrap-types-deep-types-object-node\"><code>om-unwrap-types-deep<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; This (1 *2* 3 *\/4\/* 5 \/6\/) is randomly formatted\n;; Remove italic formatting at any level\n(-&gt;&gt; (om-parse-this-element)\n     (om-map-children* (om-flatten-types-deep '(italic)\n                                              it))\n     (om-to-trimmed-string))\n ;; =&gt; \"This (1 *2* 3 *4* 5 6) is randomly formatted\"\n<\/pre><\/div><h4><a id=\"user-content-om-flatten-deep-secondary-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-flatten-deep-secondary-string\"><\/a>om-flatten-deep <code>(secondary-string)<\/code><\/h4><p>Return <strong><code>secondary-string<\/code><\/strong> with all object nodes unwrapped to plain-text. The unwrap operation will be done with <a href=\"#om-unwrap-deep-object-node\"><code>om-unwrap-deep<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; This (1 *2* 3 *\/4\/* 5 \/6\/) is randomly formatted\n;; Remove italic formatting at any level\n(-&gt;&gt; (om-parse-this-element)\n     (om-map-children (function om-flatten-deep))\n     (om-to-trimmed-string))\n ;; =&gt; \"This (1 2 3 4 5 6) is randomly formatted\"\n<\/pre><\/div><h3><a id=\"user-content-headline-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#headline-3\"><\/a>Headline<\/h3><h4><a id=\"user-content-om-headline-get-node-properties-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-node-properties-headline\"><\/a>om-headline-get-node-properties <code>(headline)<\/code><\/h4><p>Return a list of node-properties nodes in <strong><code>headline<\/code><\/strong> or nil if none.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline\n; :PROPERTIES:\n; :Effort:   1:00\n; :END:\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-node-properties)\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\":Effort:   1:00\")\n;; Given the following contents:\n; * headline\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-node-properties)\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-headline-get-properties-drawer-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-properties-drawer-headline\"><\/a>om-headline-get-properties-drawer <code>(headline)<\/code><\/h4><p>Return the properties drawer node in <strong><code>headline<\/code><\/strong>.<\/p><p>If multiple are present (there shouldn't be) the first will be returned.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline\n; :PROPERTIES:\n; :Effort:   1:00\n; :END:\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-properties-drawer)\n     (om-to-trimmed-string))\n ;; =&gt; \":PROPERTIES:\n ;      :Effort:   1:00\n ;      :END:\"\n;; Given the following contents:\n; * headline\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-properties-drawer)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-get-planning-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-planning-headline\"><\/a>om-headline-get-planning <code>(headline)<\/code><\/h4><p>Return the planning node in <strong><code>headline<\/code><\/strong> or nil if none.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline\n; CLOSED: [2019-01-01 Tue]\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-planning)\n     (om-to-trimmed-string))\n ;; =&gt; \"CLOSED: [2019-01-01 Tue]\"\n;; Given the following contents:\n; * headline\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-get-planning)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-get-subheadlines-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-subheadlines-headline\"><\/a>om-headline-get-subheadlines <code>(headline)<\/code><\/h4><p>Return list of subheadline nodes for <strong><code>headline<\/code><\/strong> node or nil if none.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline 1\n; sectional stuff\n; ** headline 2\n; ** headline 3\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-get-subheadlines)\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"** headline 2\" \"** headline 3\")\n;; Given the following contents:\n; * headline 1\n; sectional stuff\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-get-subheadlines)\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; nil\n<\/pre><\/div><h4><a id=\"user-content-om-headline-get-section-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-section-headline\"><\/a>om-headline-get-section <code>(headline)<\/code><\/h4><p>Return section node for headline <strong><code>headline<\/code><\/strong> node or nil if none.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline 1\n; sectional stuff\n; ** headline 2\n; ** headline 3\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-get-section)\n     (om-to-trimmed-string))\n ;; =&gt; \"sectional stuff\"\n;; Given the following contents:\n; * headline 1\n; ** headline 2\n; ** headline 3\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-get-section)\n     (om-to-trimmed-string))\n ;; =&gt; \"\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-get-path-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-get-path-headline\"><\/a>om-headline-get-path <code>(headline)<\/code><\/h4><p>Return tree path of <strong><code>headline<\/code><\/strong> node.<\/p><p>The return value is a list of headline titles (including that from <strong><code>headline<\/code><\/strong>) leading to the root node.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; *** three\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-get-subheadlines)\n     (car)\n     (om-headline-get-path))\n ;; =&gt; '(\"one\" \"two\")\n;; Given the following contents:\n; * one\n; ** two\n; *** three\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-get-subheadlines)\n     (car)\n     (om-headline-get-subheadlines)\n     (car)\n     (om-headline-get-path))\n ;; =&gt; '(\"one\" \"two\" \"three\")\n<\/pre><\/div><h4><a id=\"user-content-om-headline-update-item-statistics-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-update-item-statistics-headline\"><\/a>om-headline-update-item-statistics <code>(headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with updated statistics cookie via items.<\/p><p>The percent\/fraction will be computed as the number of checked items over the number of items with checkboxes (non-checkbox items will not be considered).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * statistically significant [\/]\n; - irrelevant data\n; - [ ] good data\n; - [X] bad data\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-update-item-statistics)\n     (om-to-trimmed-string))\n ;; =&gt; \"* statistically significant [1\/2]\n ;      - irrelevant data\n ;      - [ ] good data\n ;      - [X] bad data\"\n;; Given the following contents:\n; * statistically significant\n; - irrelevant data\n; - [ ] good data\n; - [X] bad data\n;; Do nothing if nothing to update\n(-&gt;&gt; (om-parse-this-headline)\n     (om-headline-update-item-statistics)\n     (om-to-trimmed-string))\n ;; =&gt; \"* statistically significant\n ;      - irrelevant data\n ;      - [ ] good data\n ;      - [X] bad data\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-update-todo-statistics-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-update-todo-statistics-headline\"><\/a>om-headline-update-todo-statistics <code>(headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with updated statistics cookie via subheadlines.<\/p><p>The percent\/fraction will be computed as the number of done subheadlines over the number of todo subheadlines (eg non-todo subheadlines will not be counted).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * statistically significant [\/]\n; ** irrelevant data\n; ** TODO good data\n; ** DONE bad data\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-update-todo-statistics)\n     (om-to-trimmed-string))\n ;; =&gt; \"* statistically significant [1\/2]\n ;      ** irrelevant data\n ;      ** TODO good data\n ;      ** DONE bad data\"\n;; Given the following contents:\n; * statistically significant\n; ** irrelevant data\n; ** TODO good data\n; ** DONE bad data\n;; Do nothing if nothing to update\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-headline-update-todo-statistics)\n     (om-to-trimmed-string))\n ;; =&gt; \"* statistically significant\n ;      ** irrelevant data\n ;      ** TODO good data\n ;      ** DONE bad data\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-indent-subheadline-index-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-indent-subheadline-index-headline\"><\/a>om-headline-indent-subheadline <code>(index headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with child headline at <strong><code>index<\/code><\/strong> indented. Unlike <a href=\"#om-headline-indent-subtree-index-headline\"><code>om-headline-indent-subtree<\/code><\/a> this will not indent the indented headline node's children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n; *** four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-headline-indent-subheadline 0)\n     (om-to-trimmed-string))\nError\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-headline-indent-subheadline 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      *** three\n ;      *** four\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-indent-subtree-index-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-indent-subtree-index-headline\"><\/a>om-headline-indent-subtree <code>(index headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with child headline at <strong><code>index<\/code><\/strong> indented. Unlike <a href=\"#om-headline-indent-subheadline-index-headline\"><code>om-headline-indent-subheadline<\/code><\/a> this will also indent the indented headline node's children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n; *** four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-headline-indent-subtree 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      *** three\n ;      **** four\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-unindent-subheadline-index-child-index-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-unindent-subheadline-index-child-index-headline\"><\/a>om-headline-unindent-subheadline <code>(index child-index headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with a child headline under <strong><code>index<\/code><\/strong> unindented. The specific child headline to unindent is selected by <strong><code>child-index<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n; *** four\n; *** four\n; *** four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-headline-unindent-subheadline 1 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      ** three\n ;      *** four\n ;      ** four\n ;      *** four\"\n<\/pre><\/div><h4><a id=\"user-content-om-headline-unindent-all-subheadlines-index-headline\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-headline-unindent-all-subheadlines-index-headline\"><\/a>om-headline-unindent-all-subheadlines <code>(index headline)<\/code><\/h4><p>Return <strong><code>headline<\/code><\/strong> node with all child headlines under <strong><code>index<\/code><\/strong> unindented.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n; *** four\n; *** four\n; *** four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-headline-unindent-all-subheadlines 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      ** three\n ;      ** four\n ;      ** four\n ;      ** four\"\n<\/pre><\/div><h3><a id=\"user-content-plain-list-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#plain-list-1\"><\/a>Plain List<\/h3><h4><a id=\"user-content-om-plain-list-set-type-type-plain-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plain-list-set-type-type-plain-list\"><\/a>om-plain-list-set-type <code>(type plain-list)<\/code><\/h4><p>Return <strong><code>plain-list<\/code><\/strong> node with type property set to <strong><code>type<\/code><\/strong>. <strong><code>type<\/code><\/strong> is one of the symbols <code>unordered<\/code> or <code>ordered<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - [ ] one\n; - [X] two\n(-&gt;&gt; (om-parse-this-element)\n     (om-plain-list-set-type 'ordered)\n     (om-to-trimmed-string))\n ;; =&gt; \"1. [ ] one\n ;      2. [X] two\"\n;; Given the following contents:\n; 1. [ ] one\n; 2. [X] two\n(-&gt;&gt; (om-parse-this-element)\n     (om-plain-list-set-type 'unordered)\n     (om-to-trimmed-string))\n ;; =&gt; \"- [ ] one\n ;      - [X] two\"\n<\/pre><\/div><h4><a id=\"user-content-om-plain-list-indent-item-index-plain-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plain-list-indent-item-index-plain-list\"><\/a>om-plain-list-indent-item <code>(index plain-list)<\/code><\/h4><p>Return <strong><code>plain-list<\/code><\/strong> node with child item at <strong><code>index<\/code><\/strong> indented. Unlike <code>om-item-indent-item-tree<\/code> this will not indent the indented item node's children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - one\n; - two\n;   - three\n; - four\n;; It makes no sense to indent the first item\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-indent-item 0)\n     (om-to-trimmed-string))\nError\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-indent-item 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;        - two\n ;        - three\n ;      - four\"\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-indent-item 2)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;      - two\n ;        - three\n ;        - four\"\n<\/pre><\/div><h4><a id=\"user-content-om-plain-list-indent-item-tree-index-plain-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plain-list-indent-item-tree-index-plain-list\"><\/a>om-plain-list-indent-item-tree <code>(index plain-list)<\/code><\/h4><p>Return <strong><code>plain-list<\/code><\/strong> node with child item at <strong><code>index<\/code><\/strong> indented. Unlike <code>om-item-indent-item<\/code> this will also indent the indented item node's children.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - one\n; - two\n;   - three\n; - four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-indent-item-tree 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;        - two\n ;          - three\n ;      - four\"\n<\/pre><\/div><h4><a id=\"user-content-om-plain-list-unindent-item-index-child-index-plain-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plain-list-unindent-item-index-child-index-plain-list\"><\/a>om-plain-list-unindent-item <code>(index child-index plain-list)<\/code><\/h4><p>Return <strong><code>plain-list<\/code><\/strong> node with a child item under <strong><code>index<\/code><\/strong> unindented. The specific child item to unindent is selected by <strong><code>child-index<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - one\n; - two\n;   - three\n;   - three\n;   - three\n; - four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-unindent-item 1 0)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;      - two\n ;      - three\n ;        - three\n ;        - three\n ;      - four\"\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-unindent-item 1 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;      - two\n ;        - three\n ;      - three\n ;        - three\n ;      - four\"\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-unindent-item 2 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;      - two\n ;        - three\n ;        - three\n ;        - three\n ;      - four\"\n<\/pre><\/div><h4><a id=\"user-content-om-plain-list-unindent-all-items-index-plain-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-plain-list-unindent-all-items-index-plain-list\"><\/a>om-plain-list-unindent-all-items <code>(index plain-list)<\/code><\/h4><p>Return <strong><code>plain-list<\/code><\/strong> node with all child items under <strong><code>index<\/code><\/strong> unindented.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - one\n; - two\n;   - three\n;   - three\n;   - three\n; - four\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-unindent-all-items 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;      - two\n ;      - three\n ;      - three\n ;      - three\n ;      - four\"\n(-&gt;&gt; (om-parse-element-at 1)\n     (om-plain-list-unindent-all-items 2)\n     (om-to-trimmed-string))\n ;; =&gt; \"- one\n ;      - two\n ;        - three\n ;        - three\n ;        - three\n ;      - four\"\n<\/pre><\/div><h3><a id=\"user-content-table-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-1\"><\/a>Table<\/h3><h4><a id=\"user-content-om-table-get-cell-row-index-column-index-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-get-cell-row-index-column-index-table\"><\/a>om-table-get-cell <code>(row-index column-index table)<\/code><\/h4><p>Return table-cell node at <strong><code>row-index<\/code><\/strong> and <strong><code>column-index<\/code><\/strong> in <strong><code>table<\/code><\/strong> node. Rule-type rows do not count toward row indices.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | 1 | 2 | 3 |\n; |---+---+---|\n; | a | b | c |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-get-cell 0 0)\n     (om-get-children)\n     (car))\n ;; =&gt; \"1\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-get-cell 1 1)\n     (om-get-children)\n     (car))\n ;; =&gt; \"b\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-get-cell -1 -1)\n     (om-get-children)\n     (car))\n ;; =&gt; \"c\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-delete-column-column-index-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-delete-column-column-index-table\"><\/a>om-table-delete-column <code>(column-index table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with column at <strong><code>column-index<\/code><\/strong> deleted.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n; |---+---|\n; | c | d |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-delete-column 0)\n     (om-to-trimmed-string))\n ;; =&gt; \"| b |\n ;      |---|\n ;      | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-delete-column 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"| a |\n ;      |---|\n ;      | c |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-delete-column -1)\n     (om-to-trimmed-string))\n ;; =&gt; \"| a |\n ;      |---|\n ;      | c |\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-delete-row-row-index-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-delete-row-row-index-table\"><\/a>om-table-delete-row <code>(row-index table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with row at <strong><code>row-index<\/code><\/strong> deleted.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n; |---+---|\n; | c | d |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-delete-row 0)\n     (om-to-trimmed-string))\n ;; =&gt; \"|---+---|\n ;      | c | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-delete-row 1)\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b |\n ;      | c | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-delete-row -1)\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b |\n ;      |---+---|\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-insert-column-column-index-column-text-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-insert-column-column-index-column-text-table\"><\/a>om-table-insert-column! <code>(column-index column-text table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with <strong><code>column-text<\/code><\/strong> inserted at <strong><code>column-index<\/code><\/strong>.<\/p><p><strong><code>column-index<\/code><\/strong> is the index of the column and <strong><code>column-text<\/code><\/strong> is a list of strings to be made into table-cells to be inserted following the same syntax as <a href=\"#om-build-table-cell-string\"><code>om-build-table-cell!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n; |---+---|\n; | c | d |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-insert-column! 1 '(\"x\" \"y\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | x | b |\n ;      |---+---+---|\n ;      | c | y | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-insert-column! -1 '(\"x\" \"y\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b | x |\n ;      |---+---+---|\n ;      | c | d | y |\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-insert-row-row-index-row-text-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-insert-row-row-index-row-text-table\"><\/a>om-table-insert-row! <code>(row-index row-text table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with <strong><code>row-text<\/code><\/strong> inserted at <strong><code>row-index<\/code><\/strong>.<\/p><p><strong><code>row-index<\/code><\/strong> is the index of the column and <strong><code>row-text<\/code><\/strong> is a list of strings to be made into table-cells to be inserted following the same syntax as <a href=\"#om-build-table-row-row-list\"><code>om-build-table-row!<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n; |---+---|\n; | c | d |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-insert-row! 1 '(\"x\" \"y\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b |\n ;      | x | y |\n ;      |---+---|\n ;      | c | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-insert-row! 2 '(\"x\" \"y\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b |\n ;      |---+---|\n ;      | x | y |\n ;      | c | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-insert-row! -1 '(\"x\" \"y\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b |\n ;      |---+---|\n ;      | c | d |\n ;      | x | y |\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-replace-cell-row-index-column-index-cell-text-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-replace-cell-row-index-column-index-cell-text-table\"><\/a>om-table-replace-cell! <code>(row-index column-index cell-text table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with a table-cell node replaced by <strong><code>cell-text<\/code><\/strong>.<\/p><p>If <strong><code>cell-text<\/code><\/strong> is a string, it will replace the children of the table-cell at <strong><code>row-index<\/code><\/strong> and <strong><code>column-index<\/code><\/strong> in <strong><code>table<\/code><\/strong>. <strong><code>cell-text<\/code><\/strong> will be processed the same as the argument given to <a href=\"#om-build-table-cell-string\"><code>om-build-table-cell!<\/code><\/a>.<\/p><p>If <strong><code>cell-text<\/code><\/strong> is nil, it will set the cell to an empty string.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | 1 | 2 |\n; |---+---|\n; | a | b |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-cell! 0 0 \"2\")\n     (om-to-trimmed-string))\n ;; =&gt; \"| 2 | 2 |\n ;      |---+---|\n ;      | a | b |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-cell! 0 0 nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"|   | 2 |\n ;      |---+---|\n ;      | a | b |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-cell! -1 -1 \"B\")\n     (om-to-trimmed-string))\n ;; =&gt; \"| 1 | 2 |\n ;      |---+---|\n ;      | a | B |\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-replace-column-column-index-column-text-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-replace-column-column-index-column-text-table\"><\/a>om-table-replace-column! <code>(column-index column-text table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with the column at <strong><code>column-index<\/code><\/strong> replaced by <strong><code>column-text<\/code><\/strong>.<\/p><p>If <strong><code>column-text<\/code><\/strong> is a list of strings, it will replace the table-cells at <strong><code>column-index<\/code><\/strong>. Each member of <strong><code>column-text<\/code><\/strong> will be processed the same as the argument given to <a href=\"#om-build-table-cell-string\"><code>om-build-table-cell!<\/code><\/a>.<\/p><p>If <strong><code>column-text<\/code><\/strong> is nil, it will clear all cells at <strong><code>column-index<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n; |---+---|\n; | c | d |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-column! 0 '(\"A\" \"B\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| A | b |\n ;      |---+---|\n ;      | B | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-column! 0 nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"|   | b |\n ;      |---+---|\n ;      |   | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-column! -1 '(\"A\" \"B\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | A |\n ;      |---+---|\n ;      | c | B |\"\n<\/pre><\/div><h4><a id=\"user-content-om-table-replace-row-row-index-row-text-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-table-replace-row-row-index-row-text-table\"><\/a>om-table-replace-row! <code>(row-index row-text table)<\/code><\/h4><p>Return <strong><code>table<\/code><\/strong> node with the row at <strong><code>row-index<\/code><\/strong> replaced by <strong><code>row-text<\/code><\/strong>.<\/p><p>If <strong><code>row-text<\/code><\/strong> is a list of strings, it will replace the cells at <strong><code>row-index<\/code><\/strong>. Each member of <strong><code>row-text<\/code><\/strong> will be processed the same as the argument given to <a href=\"#om-build-table-row-row-list\"><code>om-build-table-row!<\/code><\/a>.<\/p><p>If <strong><code>row-text<\/code><\/strong> is nil, it will clear all cells at <strong><code>row-index<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n; |---+---|\n; | c | d |\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-row! 0 '(\"A\" \"B\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| A | B |\n ;      |---+---|\n ;      | c | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-row! 0 nil)\n     (om-to-trimmed-string))\n ;; =&gt; \"|   |   |\n ;      |---+---|\n ;      | c | d |\"\n(-&gt;&gt; (om-parse-this-element)\n     (om-table-replace-row! -1 '(\"A\" \"B\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"| a | b |\n ;      |---+---|\n ;      | A | B |\"\n<\/pre><\/div><h2><a id=\"user-content-node-matching-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#node-matching-1\"><\/a>Node Matching<\/h2><p>Use pattern-matching to selectively perform operations on nodes in trees.<\/p><h4><a id=\"user-content-om-match-pattern-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-pattern-node\"><\/a>om-match <code>(pattern node)<\/code><\/h4><p>Return a list of child nodes matching <strong><code>pattern<\/code><\/strong> in <strong><code>node<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> is a list like <code>([slicer [arg1] [arg2]] [pnode] sub1 [sub2 ...])<\/code>.<\/p><p><code>slicer<\/code> is an optional prefix to the pattern describing how many and which matches to return. If not given, all matches are returned. Possible values are:<\/p><ul><li><code>:first<\/code> - return the first match<\/li>\n<li><code>:last<\/code> - return the last match<\/li>\n<li><code>:nth<\/code> <code>arg1<\/code> - return the nth match where <code>arg1<\/code> is an integer denoting the index to return (starting at 0). <code>arg1<\/code> may be a negative number to start counting at the end of the match list, in which case -1 is the last index. Using 0 and -1 for <code>arg1<\/code> is equivalent to using <code>:first<\/code> and <code>:last<\/code> respectively<\/li>\n<li><code>:sub<\/code> <code>arg1<\/code> <code>arg2<\/code> - return a sublist between indices <code>arg1<\/code> and <code>arg2<\/code>. <code>arg1<\/code> may not be greater than <code>arg2<\/code>, and both must either be non-negative integers or negative integers. In the case of negative integers, the indices refer to the same counterparts as described in <code>:nth<\/code>. If <code>arg1<\/code> and <code>arg2<\/code> are equal, this slicer has the same behavior as <code>:nth<\/code>.<\/li>\n<\/ul><p><code>pnode<\/code> is an optional literal node to be matched. If given, the matching process will start within <code>pnode<\/code> wherever it happens to be in <strong><code>node<\/code><\/strong>. This is useful for 'reusing' previous matches without repeating the same pattern.<\/p><p><code>subx<\/code> denotes subpatterns that that match nodes in the parse tree. Subpatterns may either be wildcards or conditions.<\/p><p>Conditions match exactly one level of the node tree being searched based on the node's type (the symbol returned by <a href=\"#om-get-type-node\"><code>om-get-type<\/code><\/a>), properties (the value returned by <a href=\"#om-get-property-prop-node\"><code>om-get-property<\/code><\/a> for a valid property keyword), and index (the position of the node in the list returned by <a href=\"#om-get-children-branch-node\"><code>om-get-children<\/code><\/a>). For index, both left indices (where zero refers to the left end of the list) and right indices (where -1 refers to the right end of the list) are understood. Conditions may either be atomic or compound, where compound conditions are themselves composed of atomic or compound conditions.<\/p><p>The types of atomic conditions are:<\/p><ul><li><code>type<\/code> - match when the node's type is <code>eq<\/code> to <code>type<\/code> (a symbol)<\/li>\n<li><code>index<\/code> - match when the node's index is <code>=<\/code> to <code>index<\/code> (an integer)<\/li>\n<li><code>(op index)<\/code> - match when <code>(op node-index index)<\/code> returns t. <code>op<\/code> is one of <code>&lt;<\/code>, <code>&gt;<\/code>, <code>&lt;=<\/code>, or <code>&gt;=<\/code> and <code>node-index<\/code> is the index of the node being evaluated<\/li>\n<li><code>(prop val)<\/code> - match nodes whose property <code>prop<\/code> (a keyword) is <code>equal<\/code> to <code>val<\/code>; <code>val<\/code> is obtained by evaluating <a href=\"#om-get-property-prop-node\"><code>om-get-property<\/code><\/a> with <code>prop<\/code> and the current node; if <code>prop<\/code> is invalid, an error will be thrown<\/li>\n<li><code>(:pred pred)<\/code> - match when <code>pred<\/code> evaluates to t; <code>pred<\/code> is a symbol for a unary function that takes the current node as its argument<\/li>\n<\/ul><p>Compound conditions start with an operator followed by their component conditions. In the syntax below, <code>cx<\/code> refers to a condition. The types of compound conditions are:<\/p><ul><li><code>(:and c1 c2 [c3 ...])<\/code> - match when all conditions are true<\/li>\n<li><code>(:or c1 c2 [c3 ...])<\/code> - match when at least one condition is true<\/li>\n<li><code>(:not c)<\/code> - match when condition is not true<\/li>\n<\/ul><p>In addition to conditions, <code>subx<\/code> may be a wildcard keyword to match nodes independent of their type, properties, and index. The types of wildcards are:<\/p><ul><li><code>:any<\/code> - always match exactly one node<\/li>\n<li><code>:many<\/code> <code>sub<\/code> - match <code>sub<\/code> to nodes that are zero or more levels down in the tree; <code>sub<\/code> is a subpattern of either a condition or the <code>:any<\/code> wildcard; only one subpattern is allowed after <code>:many<\/code><\/li>\n<li><code>:many!<\/code> <code>sub<\/code> - like <code>:many<\/code> but do not match within other matches<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline 1\n; ** TODO headline 2\n; stuff\n; - item 1\n; - item 2\n; - item 3\n; ** DONE headline 3\n; - item 4\n; - item 5\n; - item 6\n; ** TODO COMMENT headline 4\n; - item 7\n; - item 8\n; - item 9\n;; Match items (excluding the first) in headlines that are marked \"TODO\" and not\n;; commented. The :many keyword matches the section and plain-list nodes holding\n;; the items.\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match '((:and (:todo-keyword \"TODO\")\n                             (:commentedp nil))\n                       :many (:and item (&gt; 0))))\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"- item 2\" \"- item 3\")\n;; Given the following contents:\n; *one* *two* *three* *four* *five* *six*\n;; Return all bold nodes\n(-&gt;&gt; (om-parse-this-element)\n     (om-match '(bold))\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"*one*\" \"*two*\" \"*three*\" \"*four*\" \"*five*\" \"*six*\")\n;; Return first bold node\n(-&gt;&gt; (om-parse-this-element)\n     (om-match '(:first bold))\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"*one*\")\n;; Return last bold node\n(-&gt;&gt; (om-parse-this-element)\n     (om-match '(:last bold))\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"*six*\")\n;; Return a select bold node\n(-&gt;&gt; (om-parse-this-element)\n     (om-match '(:nth 2 bold))\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"*three*\")\n;; Return a sublist of matched bold nodes\n(-&gt;&gt; (om-parse-this-element)\n     (om-match '(:sub 1 3 bold))\n     (-map (function om-to-trimmed-string)))\n ;; =&gt; '(\"*two*\" \"*three*\" \"*four*\")\n<\/pre><\/div><h4><a id=\"user-content-om-match-delete-pattern-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-delete-pattern-node\"><\/a>om-match-delete <code>(pattern node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> without children matching <strong><code>pattern<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline one\n; ** headline two\n; ** headline three\n; ** headline four\n;; Selectively delete headlines\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-delete '(headline))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline one\"\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-delete '(:first headline))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline one\n ;      ** headline three\n ;      ** headline four\"\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-delete '(:last headline))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline one\n ;      ** headline two\n ;      ** headline three\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-extract-pattern-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-extract-pattern-node\"><\/a>om-match-extract <code>(pattern node)<\/code><\/h4><p>Remove nodes matching <strong><code>pattern<\/code><\/strong> from <strong><code>node<\/code><\/strong>. Return cons cell where the car is a list of all removed nodes and the cdr is the modified <strong><code>node<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; pull me \/under\/\n(--&gt; (om-parse-this-element)\n     (om-match-extract '(:many italic)\n                       it)\n     (cons (-map (function om-to-trimmed-string)\n                 (car it))\n           (om-to-trimmed-string (cdr it))))\n ;; =&gt; '((\"\/under\/\") . \"pull me\")\n<\/pre><\/div><h4><a id=\"user-content-om-match-map-pattern-fun-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-map-pattern-fun-node\"><\/a>om-match-map <code>(pattern fun node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>fun<\/code><\/strong> applied to children matching <strong><code>pattern<\/code><\/strong>. <strong><code>fun<\/code><\/strong> is a unary function that takes a node and returns a new node which will replace the original.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * headline one\n; ** TODO headline two\n; ** headline three\n; ** headline four\n;; Selectively mark headlines as DONE\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-map '(headline)\n                   (lambda (it)\n                     (om-set-property :todo-keyword \"DONE\" it)))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline one\n ;      ** DONE headline two\n ;      ** DONE headline three\n ;      ** DONE headline four\"\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-map* '(:first headline)\n                    (om-set-property :todo-keyword \"DONE\" it))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline one\n ;      ** DONE headline two\n ;      ** headline three\n ;      ** headline four\"\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-map '(:last headline)\n                   (-partial (function om-set-property)\n                             :todo-keyword \"DONE\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"* headline one\n ;      ** TODO headline two\n ;      ** headline three\n ;      ** DONE headline four\"\n;; Given the following contents:\n; * headline\n; :PROPERTIES:\n; :Effort:   0:30\n; :END:\n;; Match the literal property-drawer node and map the node-property inside if\n;; the property-drawer exists\n(let ((hl (om-parse-this-headline)))\n  (-if-let (pd (om-headline-get-properties-drawer hl))\n      (-&gt;&gt; hl (om-match-map* (\\` ((\\, pd)\n                                  node-property))\n                             (om-set-property :value \"1:30\" it))\n           (om-to-trimmed-string))\n    (print \"...or do something else if no drawer\")))\n ;; =&gt; \"* headline\n ;      :PROPERTIES:\n ;      :Effort:   1:30\n ;      :END:\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-mapcat-pattern-fun-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-mapcat-pattern-fun-node\"><\/a>om-match-mapcat <code>(pattern fun node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>fun<\/code><\/strong> applied to children matching <strong><code>pattern<\/code><\/strong>. <strong><code>fun<\/code><\/strong> is a unary function that takes a node and returns a list of new nodes which will be spliced in place of the original node.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-mapcat* '(:first headline)\n                       (list (om-build-headline! :title-text \"1.5\" :level 2)\n                             it))\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** 1.5\n ;      ** two\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-replace-pattern-node-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-replace-pattern-node-node\"><\/a>om-match-replace <code>(pattern node* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>node*<\/code><\/strong> in place of children matching <strong><code>pattern<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; *1* 2 *3* 4 *5* 6 *7* 8 *9* 10\n(-&gt;&gt; (om-parse-this-element)\n     (om-match-replace '(:many bold)\n       (om-build-bold :post-blank 1 \"0\"))\n     (om-to-trimmed-string))\n ;; =&gt; \"*0* 2 *0* 4 *0* 6 *0* 8 *0* 10\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-insert-before-pattern-node-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-insert-before-pattern-node-node\"><\/a>om-match-insert-before <code>(pattern node* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>node*<\/code><\/strong> inserted before children matching <strong><code>pattern<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-insert-before '(headline)\n       (om-build-headline! :title-text \"new\" :level 2))\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** new\n ;      ** two\n ;      ** new\n ;      ** three\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-insert-after-pattern-node-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-insert-after-pattern-node-node\"><\/a>om-match-insert-after <code>(pattern node* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>node*<\/code><\/strong> inserted after children matching <strong><code>pattern<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-insert-after '(headline)\n       (om-build-headline! :title-text \"new\" :level 2))\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      ** new\n ;      ** three\n ;      ** new\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-insert-within-pattern-index-node-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-insert-within-pattern-index-node-node\"><\/a>om-match-insert-within <code>(pattern index node* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>node*<\/code><\/strong> inserted at <strong><code>index<\/code><\/strong> in children matching <strong><code>pattern<\/code><\/strong>.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a> with the exception that <strong><code>pattern<\/code><\/strong> may be nil. In this case <strong><code>node*<\/code><\/strong> will be inserted at <strong><code>index<\/code><\/strong> in the immediate, top level children of <strong><code>node<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-insert-within '(headline)\n         0 (om-build-headline! :title-text \"new\" :level 3))\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      *** new\n ;      ** three\n ;      *** new\"\n;; The nil pattern denotes top-level element\n(-&gt;&gt; (om-parse-this-subtree)\n     (om-match-insert-within nil 1 (om-build-headline! :title-text \"new\" :level 2))\n     (om-to-trimmed-string))\n ;; =&gt; \"* one\n ;      ** two\n ;      ** new\n ;      ** three\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-splice-pattern-nodes-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-splice-pattern-nodes-node\"><\/a>om-match-splice <code>(pattern nodes* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>nodes*<\/code><\/strong> spliced in place of children matching <strong><code>pattern<\/code><\/strong>. <strong><code>nodes*<\/code><\/strong> is a list of nodes.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n(let ((L (list (om-build-headline! :title-text \"new0\" :level 2)\n               (om-build-headline! :title-text \"new1\" :level 2))))\n  (-&gt;&gt; (om-parse-this-subtree)\n       (om-match-splice '(0)\n         L)\n       (om-to-trimmed-string)))\n ;; =&gt; \"* one\n ;      ** new0\n ;      ** new1\n ;      ** three\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-splice-before-pattern-nodes-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-splice-before-pattern-nodes-node\"><\/a>om-match-splice-before <code>(pattern nodes* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>nodes*<\/code><\/strong> spliced before children matching <strong><code>pattern<\/code><\/strong>. <strong><code>nodes*<\/code><\/strong> is a list of nodes.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n(let ((L (list (om-build-headline! :title-text \"new0\" :level 2)\n               (om-build-headline! :title-text \"new1\" :level 2))))\n  (-&gt;&gt; (om-parse-this-subtree)\n       (om-match-splice-before '(0)\n         L)\n       (om-to-trimmed-string)))\n ;; =&gt; \"* one\n ;      ** new0\n ;      ** new1\n ;      ** two\n ;      ** three\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-splice-after-pattern-nodes-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-splice-after-pattern-nodes-node\"><\/a>om-match-splice-after <code>(pattern nodes* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>nodes*<\/code><\/strong> spliced after children matching <strong><code>pattern<\/code><\/strong>. <strong><code>nodes*<\/code><\/strong> is a list of nodes.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n(let ((L (list (om-build-headline! :title-text \"new0\" :level 2)\n               (om-build-headline! :title-text \"new1\" :level 2))))\n  (-&gt;&gt; (om-parse-this-subtree)\n       (om-match-splice-after '(0)\n         L)\n       (om-to-trimmed-string)))\n ;; =&gt; \"* one\n ;      ** two\n ;      ** new0\n ;      ** new1\n ;      ** three\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-splice-within-pattern-index-nodes-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-splice-within-pattern-index-nodes-node\"><\/a>om-match-splice-within <code>(pattern index nodes* node)<\/code><\/h4><p>Return <strong><code>node<\/code><\/strong> with <strong><code>nodes*<\/code><\/strong> spliced at <strong><code>index<\/code><\/strong> in children matching <strong><code>pattern<\/code><\/strong>. <strong><code>nodes*<\/code><\/strong> is a list of nodes.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a> with the exception that <strong><code>pattern<\/code><\/strong> may be nil. In this case <strong><code>nodes*<\/code><\/strong> will be inserted at <strong><code>index<\/code><\/strong> in the immediate, top level children of <strong><code>node<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n; *** four\n(let ((L (list (om-build-headline! :title-text \"new0\" :level 3)\n               (om-build-headline! :title-text \"new1\" :level 3))))\n  (-&gt;&gt; (om-parse-this-subtree)\n       (om-match-splice-within '(headline)\n           0 L)\n       (om-to-trimmed-string)))\n ;; =&gt; \"* one\n ;      ** two\n ;      *** new0\n ;      *** new1\n ;      ** three\n ;      *** new0\n ;      *** new1\n ;      *** four\"\n(let ((L (list (om-build-headline! :title-text \"new0\" :level 2)\n               (om-build-headline! :title-text \"new1\" :level 2))))\n  (-&gt;&gt; (om-parse-this-subtree)\n       (om-match-splice-within nil 1 L)\n       (om-to-trimmed-string)))\n ;; =&gt; \"* one\n ;      ** two\n ;      ** new0\n ;      ** new1\n ;      ** three\n ;      *** four\"\n<\/pre><\/div><h4><a id=\"user-content-om-match-do-pattern-fun-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-match-do-pattern-fun-node\"><\/a>om-match-do <code>(pattern fun node)<\/code><\/h4><p>Like <a href=\"#om-match-map-pattern-fun-node\"><code>om-match-map<\/code><\/a> but for side effects only. <strong><code>fun<\/code><\/strong> is a unary function that has side effects and is applied to the matches from <strong><code>node<\/code><\/strong> using <strong><code>pattern<\/code><\/strong>. This function itself returns nil.<\/p><p><strong><code>pattern<\/code><\/strong> follows the same rules as <a href=\"#om-match-pattern-node\"><code>om-match<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div><h2><a id=\"user-content-buffer-side-effects-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer-side-effects-1\"><\/a>Buffer Side Effects<\/h2><p>Map node manipulations into buffers.<\/p><h3><a id=\"user-content-insert-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#insert-1\"><\/a>Insert<\/h3><h4><a id=\"user-content-om-insert-point-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-insert-point-node\"><\/a>om-insert <code>(point node)<\/code><\/h4><p>Convert <strong><code>node<\/code><\/strong> to a string and insert at <strong><code>point<\/code><\/strong> in the current buffer. Return <strong><code>node<\/code><\/strong>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; \n(-&gt;&gt; (om-build-headline! :title-text \"two\")\n     (om-insert (point-max)))\n ;; Output these buffer contents\n ;; $&gt; \"* one\n ;      * two\"\n;; Given the following contents:\n; a *game* or a \/boy\/\n(-&gt;&gt; (om-build-paragraph! \"we don't care if you're\")\n     (om-insert (point-min)))\n ;; Output these buffer contents\n ;; $&gt; \"we don't care if you're\n ;      a *game* or a \/boy\/\"\n<\/pre><\/div><h4><a id=\"user-content-om-insert-tail-point-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-insert-tail-point-node\"><\/a>om-insert-tail <code>(point node)<\/code><\/h4><p>Like <a href=\"#om-insert-point-node\"><code>om-insert<\/code><\/a> but insert <strong><code>node<\/code><\/strong> at <strong><code>point<\/code><\/strong> and move to end of insertion.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div><h3><a id=\"user-content-update-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#update-1\"><\/a>Update<\/h3><h4><a id=\"user-content-om-update-fun-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-fun-node\"><\/a>om-update <code>(fun node)<\/code><\/h4><p>Replace <strong><code>node<\/code><\/strong> in the current buffer with a new one. <strong><code>fun<\/code><\/strong> is a unary function that takes <strong><code>node<\/code><\/strong> and returns a modified <strong><code>node<\/code><\/strong>. This modified node is then written in place of the old node in the current buffer.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * TODO win grammy\n(-&gt;&gt; (om-parse-this-headline)\n     (om-update (lambda (hl)\n                  (om-set-property :todo-keyword \"DONE\" hl))))\n ;; Output these buffer contents\n ;; $&gt; \"* DONE win grammy\"\n;; Given the following contents:\n; * win grammy [0\/0]\n; - [ ] write punk song\n; - [ ] get new vocalist\n; - [ ] sell 2 singles\n(-&gt;&gt; (om-parse-this-headline)\n     (om-update* (-&gt;&gt; (om-match-map '(:many item)\n                                    (function om-item-toggle-checkbox)\n                                    it)\n                      (om-headline-update-item-statistics))))\n ;; Output these buffer contents\n ;; $&gt; \"* win grammy [3\/3]\n ;      - [X] write punk song\n ;      - [X] get new vocalist\n ;      - [X] sell 2 singles\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-object-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-object-at-point-fun\"><\/a>om-update-object-at <code>(point fun)<\/code><\/h4><p>Update object under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an object and returns a modified object<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; [[http:\/\/example.com][desc]]\n(om-update-object-at* (point)\n  (om-set-property :path \"\/\/buymoreram.com\" it))\n ;; Output these buffer contents\n ;; $&gt; \"[[http:\/\/buymoreram.com][desc]]\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-element-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-element-at-point-fun\"><\/a>om-update-element-at <code>(point fun)<\/code><\/h4><p>Update element under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an element and returns a modified element<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+CALL: ktulu()\n(om-update-element-at* (point)\n  (om-set-properties (list :call \"cthulhu\" :inside-header '(:cache no)\n                            :arguments '(\"x=4\")\n                            :end-header '(:results html))\n                     it))\n ;; Output these buffer contents\n ;; $&gt; \"#+CALL: cthulhu[:cache no](x=4) :results html\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-table-row-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-table-row-at-point-fun\"><\/a>om-update-table-row-at <code>(point fun)<\/code><\/h4><p>Update table-row under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an table-row and returns a modified table-row<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; | a | b |\n(om-update-table-row-at* (point)\n  (om-map-children* (cons (om-build-table-cell! \"0\")\n                          it)\n                    it))\n ;; Output these buffer contents\n ;; $&gt; \"| 0 | a | b |\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-item-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-item-at-point-fun\"><\/a>om-update-item-at <code>(point fun)<\/code><\/h4><p>Update item under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an item and returns a modified item<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; - [ ] thing\n(om-update-item-at* (point)\n  (om-item-toggle-checkbox it))\n ;; Output these buffer contents\n ;; $&gt; \"- [X] thing\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-headline-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-headline-at-point-fun\"><\/a>om-update-headline-at <code>(point fun)<\/code><\/h4><p>Update headline under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an headline and returns a modified headline<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * TODO might get done\n; * DONE no need to update\n(om-update-headline-at* (point)\n  (om-set-property :todo-keyword \"DONE\" it))\n ;; Output these buffer contents\n ;; $&gt; \"* DONE might get done\n ;      * DONE no need to update\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-subtree-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-subtree-at-point-fun\"><\/a>om-update-subtree-at <code>(point fun)<\/code><\/h4><p>Update subtree under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an subtree and returns a modified subtree<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; * one\n; ** two\n; ** three\n; * not updated\n(om-update-subtree-at* (point)\n  (om-headline-indent-subheadline 1 it))\n ;; Output these buffer contents\n ;; $&gt; \"* one\n ;      ** two\n ;      *** three\n ;      * not updated\"\n<\/pre><\/div><h4><a id=\"user-content-om-update-section-at-point-fun\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-update-section-at-point-fun\"><\/a>om-update-section-at <code>(point fun)<\/code><\/h4><p>Update section under <strong><code>point<\/code><\/strong> using <strong><code>fun<\/code><\/strong>. <strong><code>fun<\/code><\/strong> takes an section and returns a modified section<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Given the following contents:\n; #+KEY1: VAL1\n; #+KEY2: VAL2\n; * irrelevant headline\n;; Update the top buffer section before the headlines start\n(om-update-section-at* (point)\n  (om-map-children* (--map (om-map-property :value (function s-downcase)\n                                             it)\n                           it)\n                    it))\n ;; Output these buffer contents\n ;; $&gt; \"#+KEY1: val1\n ;      #+KEY2: val2\n ;      * irrelevant headline\"\n<\/pre><\/div><h3><a id=\"user-content-misc-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#misc-1\"><\/a>Misc<\/h3><h4><a id=\"user-content-om-fold-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-fold-node\"><\/a>om-fold <code>(node)<\/code><\/h4><p>Fold the children of <strong><code>node<\/code><\/strong> if they exist.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div><h4><a id=\"user-content-om-unfold-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#om-unfold-node\"><\/a>om-unfold <code>(node)<\/code><\/h4><p>Unfold the children of <strong><code>node<\/code><\/strong> if they exist.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>no examples :(<\/pre><\/div>","created_at":"2020-01-13T17:25:47+0000","updated_at":"2020-01-13T17:25:55+0000","published_by":["ndwarshuis"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":104,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/6629118?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1104,"title":"How I manage my e-books","url":"https:\/\/input.sh\/how-i-manage-my-ebooks\/","content":"<article><p>I really like e-books. I've hoarded about 500 of them over the years. I've collected them from a variety of sources, in plenty of formats, and I've created what I consider to be a quite effective way of managing all of that.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/input.sh\/content\/images\/2020\/01\/Screenshot-from-2020-01-11-10-09-46@2x.png\" class=\"kg-image\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><figcaption>A portion of the books I've read in 2019.<\/figcaption><\/figure><h2 id=\"calibre-in-the-center\">Calibre in the center<\/h2><p>Like plenty of other e-book fans out there, <a href=\"https:\/\/calibre-ebook.com\/\">Calibre<\/a> is in the center of my setup. Each time I acquire an e-book, I drag and drop it from my downloads folder into Calibre, and use my pal <a href=\"https:\/\/apprenticealf.wordpress.com\/\">Apprentice Alf<\/a> to get rid of the awful DRM if it's present.<\/p><h4 id=\"customcolumns\">Custom columns<\/h4><p>In order to utilize Calibre to its fullest, I have defined a couple of custom columns:<\/p><ul><li><strong>Read:<\/strong> A boolean (yes\/no) value that indicates whether I've read the book or not.<\/li>\n<li><strong>Pages:<\/strong> A field that gets populated by the <a href=\"https:\/\/www.mobileread.com\/forums\/showthread.php?t=134000\">Count Pages<\/a> plugin.<\/li>\n<li><strong>Percentage:<\/strong> Marks my reading position when I'm reading the book, so that I can switch between different devices.<\/li>\n<\/ul><h2 id=\"magicpartthewebcomponent\">Magic part: The web component<\/h2><p>An ebook folder on my laptop gets automatically synced to my server using <a href=\"https:\/\/nextcloud.com\">Nextcloud<\/a> . That server is also running an alternative web interface for Calibre (<a href=\"https:\/\/github.com\/janeczku\/calibre-web\/\">janeczku\/calibre-web<\/a> ). As a result, moments after I drag and drop a new e-book to Calibre on my laptop, it also becomes available via a web interface.<\/p><p>This web interface gives me a few key features that I wouldn't get otherwise. Most notably:<\/p><p><strong>I can sort my e-books into shelves.<\/strong> This is a crucial feature that is pretty inexplicably missing from within Calibre out of the box, though it can be achieved with a custom tag-like column. I personally find it simpler to maintain shelves from the web interface.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/input.sh\/content\/images\/2020\/01\/Screenshot-from-2020-01-11-13-04-45@2x.png\" class=\"kg-image\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><figcaption>Interface for adding books to shelves.<\/figcaption><\/figure><p><strong>It has an integrated OPDS feed<\/strong>, allowing me to easily download my e-books to my phone straight from my preferred e-book reading app: <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.flyersoft.moonreader\">Moon+ Reader<\/a> (note: OPDS support is a premium feature).<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/input.sh\/content\/images\/2020\/01\/Untitled-1.png\" class=\"kg-image\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><figcaption>My Calibre-Web being displayed as a net library in Moon+ Reader.<\/figcaption><\/figure><p><strong>I can send my e-books to my e-reader.<\/strong> There's a button to transfer an e-book to my Kindle Paperwhite straight from my browser. If a conversion needs to be made (for example, from EPUB to MOBI), that gets done server-side automatically \u2014 no interaction necessary. After the conversion gets completed, the server sends an email with the e-book as an attachment to the email address associated with my e-reader.<\/p><h2 id=\"replacinggoodreads\">Replacing Goodreads<\/h2><p>I've used Goodreads for years, even though <a href=\"https:\/\/onezero.medium.com\/almost-everything-about-goodreads-is-broken-662e424244d5\">everything about it is broken<\/a> . Once I've started to ignore all of the Goodreads' social features that I didn't actively use anyway, I've started to see it for what it really is: a list making app.<\/p><p>Specifically, Goodreads' main feature is allowing you to easily maintain three lists:<\/p><ol><li>Books you're planning on reading<\/li>\n<li>Books you're currently reading<\/li>\n<li>Books you've read<\/li>\n<\/ol><p>Once I've realized that, I came up with plenty of options on how to better manage those lists. Most of them ended up revolving around <a href=\"https:\/\/github.com\/adam-p\/markdown-here\/wiki\/Markdown-Cheatsheet\">Markdown<\/a> , since that's how I write all of my notes and blog posts. However, when I've discovered Calibre-Web, I went back to managing my lists in a browser tab. I have defined the following shelves:<\/p><ul><li>Books I've read (one shelf for each year since I keep track)<\/li>\n<li>My favorites<\/li>\n<li>To-read list<\/li>\n<li>Currently reading list<\/li>\n<\/ul><p>I add or move things from them from time to time, leave the rating for the book, and finally...<\/p><h2 id=\"maketheendresultpublic\">Make the end result public<\/h2><p>As a final step in the chain, I have created a script that allow me to publish the list of books I've read on my website. Since Calibre-Web doesn't have an API, I ended up scraping my own server using <a href=\"https:\/\/2.python-requests.org\/en\/master\/\">Python Requests<\/a> and <a href=\"https:\/\/pypi.org\/project\/beautifulsoup4\/\">BeautifulSoup<\/a> . After about one hundred lines of spaghetti code gets executed, I end up with two files:<\/p><ul><li>\n<p><code>books-read.md<\/code>, which goes straight to my CMS, allowing me to publicly share the list of books I have read, sorted by the year in which I've finished reading them.<\/p>\n<\/li>\n<li>\n<p><code>reading-stats.json<\/code>, which gets processed using <a href=\"https:\/\/www.chartjs.org\/\">Chart.js<\/a> into bar charts that provide the answers to the following questions:<\/p>\n<ul><li>How many books do I read per year?<\/li>\n<li>How many pages do they contain in total?<\/li>\n<li>What's the average rating I gave to the books I've finished reading?<\/li>\n<\/ul><\/li>\n<\/ul><p>The end results can be seen here:<\/p><p>Overall, I'm pretty satisfied with how well this system behaves with little to no maintenance needed (at least for now). The only improvement I could think of is that I want to create another page on my website showcasing my favorite books.<\/p><\/article><div class=\"article-bottom-info d-flex justify-content-between align-items-center post-presentation\"><p class=\"d-block d-sm-none\">Published 20 hours ago<\/p><p>Share this post on:<\/p><\/div>","created_at":"2020-01-13T15:33:34+0000","updated_at":"2020-01-13T15:33:34+0000","published_at":"2020-01-12T19:09:50+0000","published_by":["Aleksandar Todorovi\u0107"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"input.sh","preview_picture":"https:\/\/input.sh\/content\/images\/2020\/01\/truong-dat-GJTvskNc4NU-unsplash.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1103,"title":"Don't Just Roll The Dice","url":"https:\/\/neildavidson.com\/downloads\/dont-just-roll-the-dice-2.0.0.pdf","content":"Neil Davidson<br \/>\nEFENDI MINIBOOKS<br \/>\nDon&#39;t Just Roll The Dice<br \/>\n<br \/>\nDon&#39;t Just Roll The Dice<br \/>\nCopyright \u00a92012 Neil Davidson<br \/>\nThe right of Neil Davidson to be identified as the author of this work has been asserted<br \/>\nby him in accordance with the Copyright, Designs and Patents Act 1988.<br \/>\nThis work is licensed under the Creative Commons Attribution-NonCommercial-No<br \/>\nDerivative Works 2.0 License. To view a copy of this license, visit<br \/>\nhttp:\/\/creativecommons.org\/licenses\/by-nc-nd\/2.0\/ or (b) send a letter to Creative<br \/>\nCommons, 171 2nd Street, Suite 300, San Francisco, California, 94105, USA.<br \/>\nEvery effort has been made to ensure the accuracy of the information presented.<br \/>\nHowever, the information contained in this book is sold without warranty, either<br \/>\nexpress or implied. Neither the authors, Efendi Books nor its dealers or distributors will<br \/>\nbe held liable for any damages caused or alleged to be caused directly or indirectly by<br \/>\nthis book.<br \/>\nEfendi Books has endeavored to provide trademark information about all companies<br \/>\nand products mentioned in this book. However, we cannot guarantee that this<br \/>\ninformation is 100% accurate.<br \/>\nFirst Published: 2009<br \/>\nThis Edition Published: 2012<br \/>\neBook Version: 2.0.0<br \/>\nPublished by\tEfendi Books\t<br \/>\nFirst published by\tSimple Talk Publishing\t<br \/>\nISBN: 978-0-9571791-2-7<br \/>\n<br \/>\nTable Of Contents\t<br \/>\nAbout The Authors\t.............................................................................................\ti\t<br \/>\nAbout The Reviewers\t.........................................................................................\ti\t<br \/>\nForeword\t...........................................................................................................\tii\t<br \/>\nAn Introduction To Don&#39;t Just Roll The Dice\t.......................................................\tii\t<br \/>\nChapter 1 - Economics<br \/>\nChapter 2 - Pricing Pyschology\t<br \/>\nWhat Is Your Product?\t....................................................................\t8\t<br \/>\nPerceived Value\t.............................................................................\t9\t<br \/>\nChapter 3 - Pitfalls\t<br \/>\nCompetitors\t.................................................................................\t18\t<br \/>\nFairness\t........................................................................................\t19\t<br \/>\nPirates\t..........................................................................................\t19\t<br \/>\nSwitching Costs\t............................................................................\t20\t<br \/>\nShould You Take Your Costs Into Account?\t..................................\t21\t<br \/>\nChapter 4 - Advanced Pricing\t<br \/>\nVersioning\t....................................................................................\t25\t<br \/>\nBundling\t.......................................................................................\t31\t<br \/>\nMulti-user Licenses\t.......................................................................\t32\t<br \/>\nThe Purchasing Process\t...............................................................\t33\t<br \/>\nFree\t..............................................................................................\t34\t<br \/>\nBargains\t.......................................................................................\t37\t<br \/>\nApps and App Markets\t.................................................................\t38\t<br \/>\nSoftware As A Service\t..................................................................\t40\t<br \/>\nDifferent Ways Of Pricing\t.............................................................\t45\t<br \/>\nChoosing The Right Model\t............................................................\t48\t<br \/>\nChapter 5 - Pricing Perception\t<br \/>\nPractice Trumps Theory\t...............................................................\t51\t<br \/>\nHow To Change Your Pricing\t........................................................\t52\t<br \/>\nChapter 6 - Product Pricing Checklist<br \/>\nSummary &amp; Bibliography\t................................................................................\t57<br \/>\n<br \/>\nAbout The Authors\t<br \/>\nNeil Davidson\tis co-founder and joint CEO of Red Gate\t<br \/>\nsoftware. Red Gate was founded in 1999 and now employs<br \/>\nsome 300 people across 3 continents. It has been in the<br \/>\nSunday Times top 100 companies to work for six years<br \/>\nrunning. It was founded with no VC money and little debt.<br \/>\nNeil is also founder of the annual Business of Software<br \/>\nconference.<br \/>\nJamie Rumbelow\tis a software developer and writer from\t<br \/>\nCambridge, UK. He is the founder and director of Efendi<br \/>\nBooks, a small publishing start-up, as well as a part-time<br \/>\nstudent and freelancer.\t<br \/>\nAbout The Reviewers<br \/>\nPhil Factor\t(real name withheld to protect the guilty), aka Database Mole, has 25\t<br \/>\nyears of experience with database-intensive applications. Despite having once been<br \/>\nshouted at by a furious Bill Gates at an exhibition in the early 1980s, he has remained<br \/>\nresolutely anonymous throughout his career.<br \/>\nMichael Pryor\tfounded Fog Creek Software with Joel Spolsky in September 2000. He\t<br \/>\nhas served as the company\u2019s president since the beginning, and has also been the CFO<br \/>\nsince 2006. Michael graduated from Dartmouth College with an Honors B.A. in<br \/>\nComputer Science (Phi Beta Kappa, magna cum laude). After graduation, he joined Juno<br \/>\nOnline Services, as a Windows client developer. He writes a column for Make Magazine<br \/>\ncalled Puzzle This, and runs the popular interview website TechInterview.org.\t<br \/>\ni<br \/>\n<br \/>\nNeil\u2019s Foreword<br \/>\nAt Business of Software 2007, Michael Pryor held an impromptu session on how to price<br \/>\nyour software. So many people turned up, and so many people kept on arriving, that by<br \/>\nthe time they\u2019d introduced themselves there was no time left to talk about software<br \/>\npricing. I\u2019ve had similar experiences; in fact, \u201cHow do I price my software?\u201d is probably<br \/>\nthe most common question I\u2019m asked by software entrepreneurs and product<br \/>\nmanagers.<br \/>\nThis handbook is an attempt to answer that question.<br \/>\nBut first, I\u2019d like to thank Phil Factor, Tony Davis and Michael Pryor for all their editing,<br \/>\nreviewing and suggestions. More people than I can possibly mention have contributed<br \/>\nwith offers of help, anecdotes and proofreading. This handbook was way better with<br \/>\ntheir input than it ever could have been without. Thanks guys.<br \/>\nJamie\u2019s Foreword<br \/>\nIt turns out it\u2019s very hard to improve on brilliance.<br \/>\nDon\u2019t Just Roll The Dice 1st Edition was a brilliant book. When Neil came to me and<br \/>\nsuggested that we revise and update his popular free eBook on software pricing, I was<br \/>\nterrified. I didn\u2019t want it to lose its sense of clarity, its usefully short approach to<br \/>\ninstruction or its fantastic stories and context.<br \/>\nSo don\u2019t be afraid that this won\u2019t be the book you know and love. It will be. DJRTD 2.0 is<br \/>\nan evolution, not a revolution.<br \/>\nWe\u2019ve listened to your feedback and worked hard at revising passages of text,<br \/>\nupdating statistics and bringing the content into 2012. We\u2019ve added a whole new<br \/>\nsection on Software As A Service pricing, as well as a few bits on pricing your mobile<br \/>\napps. We\u2019ve added some lovely new illustrations and re-designed the book to give it a<br \/>\nwhole new look and feel.<br \/>\nWelcome to Don\u2019t Just Roll The Dice 2. I hope you enjoy it.\t<br \/>\nii<br \/>\n<br \/>\nAn Introduction To Don\u2019t Just Roll The Dice<br \/>\nIn 1938, two young engineers were ready to launch their first product. They\u2019d struggled<br \/>\nwith what to build. After considering amplifiers, radio equipment, air controllers,<br \/>\nharmonicas and even muscle-building electrodes for housewives, they\u2019d finally decided<br \/>\nto create an oscilloscope. Not wanting customers to be put off by a version one<br \/>\nproduct, they sensibly called it the Model 200A.<br \/>\nThe next step? Decide the pricing.<br \/>\nThey eventually settled on $54.40. Was that because it represented the cost of<br \/>\nmanufacturing, plus a decent markup? No. These engineers hadn\u2019t taken that into<br \/>\naccount. In fact, they soon realized that the cost of building each oscilloscope was<br \/>\nmore than the price they were asking. Was it based on what the competition charged?<br \/>\nNo. They hadn\u2019t bothered to discover that General Radio charged $400 for an<br \/>\nequivalent model.<br \/>\nThey chose $54.40 because it reminded them of the 1844 slogan used in the campaign<br \/>\nto establish the northern border of the United States in the Pacific Northwest (\u201c54\u201d 40&#39;<br \/>\nor Fight!\u201c).<br \/>\nWhat a dumb-ass way to price a product.<br \/>\nBut these two young engineers recovered from their stumble. The Model 200A went on<br \/>\nto become the longest-selling basic electronic design of all time, still selling 33 years<br \/>\nlater. The company they founded became an institution. Their names? Dave Hewlett<br \/>\nand Bill Packard.<br \/>\nIf Hewlett and Packard, two Stanford graduates with the rosiest of futures ahead of<br \/>\nthem, can flounder so badly when faced with the problem of how to price their<br \/>\nproducts, what hope do the rest of us have?<br \/>\nQuite a lot, as it turns out.\t<br \/>\nii<br \/>\n<br \/>\nChapter One<br \/>\nEconomics<br \/>\n<br \/>\nTo understand product pricing, it helps to understand some, but not too much,<br \/>\neconomics. The easiest way is through a simple example.<br \/>\nLet&#39;s say you&#39;ve just launched the Time Tracker 3000. It&#39;s a downloadable piece of<br \/>\nsoftware that logs which applications you use throughout the day, and sends the usage<br \/>\ninformation back to a central web site. From there, you can find out what you&#39;ve been<br \/>\nup to all day long.<br \/>\nYou&#39;ve decided to charge a one-off fee for it, but how are you going to price it?<br \/>\nIf you give it away for free, you&#39;ll get lots of customers. A thousand, say, including<br \/>\nBelinda the bargain hunter, Stewart the student, Willhelm the web start-up founder, Pat<br \/>\nthe product manager and Ernest the enterprise developer.<br \/>\nLet&#39;s represent these thousand customers, paying no money, as an infinitely thin<br \/>\nhorizontal bar (representing the $0 price), 1000 units long (representing the quantity):<br \/>\nIf you charge $100 instead, the number of purchasers will instantly be a lot less.<br \/>\nBelinda is a bargain hunter and was only using the software because it was free, and<br \/>\nStew is a student, so neither of them will buy. You&#39;ll get, for the sake of argument, five<br \/>\nhundred customers instead of the original thousand. Let&#39;s represent that as a bar with<br \/>\na height of $100 and a length of 500 units. What&#39;s the revenue you generate from this?<br \/>\nIt&#39;s the area of the bar, so $100 x 500 = $50,000.\t<br \/>\n2<br \/>\n<br \/>\nLet&#39;s overlay it on top of the first bar:<br \/>\nWhat happens if you increase the price to $200? Some of the people who would have<br \/>\nbought at $100 will no longer buy, so your number of sales will decrease again.<br \/>\nWillhelm runs his own company and can&#39;t justify the price, so he&#39;s no longer interested.<br \/>\nLet&#39;s say 300 people will still buy; represent this with a rectangle and overlay it onto<br \/>\nthe chart. Again, the revenue you make is the area of the rectangle. $200 x 300 =<br \/>\n$60,000:<br \/>\nLet&#39;s increase the price once more, to $500. Pat the product manager drops off<br \/>\nbecause at that price she&#39;d rather buy a competitor&#39;s tool. Let&#39;s say 50 people will still\t<br \/>\n3<br \/>\n<br \/>\nbuy, and represent this as a rectangle, overlaid on the chart. Once more, the value of<br \/>\nthis rectangle is its area: 50 people buying at $500, so 50 x $500 = $25,000.<br \/>\nFinally, since the Time Tracker 3000 is valuable, but not that valuable, let&#39;s assume<br \/>\nthat nobody will buy if you price it at $1,000 and represent this as a bar of width 0 and<br \/>\nplot it on the chart:<br \/>\nWe&#39;ve plotted five points on what is becoming a curve of price against the number of<br \/>\npeople who will buy the Time Tracker 3000 at that price. What&#39;s more, you can work\t<br \/>\n4<br \/>\n<br \/>\nout the total revenue you will get at any particular price by looking at the area of the<br \/>\nrectangle (price multiplied by purchasers) under the point at that graph:<br \/>\nEconomists call this a\tdemand curve\t.\t<br \/>\nTo maximize the revenue of the Time Tracker 3000, we need to find a point on the<br \/>\ngraph that maximizes the size of the rectangle underneath it. To understand that, it<br \/>\nhelps to plot the area (i.e. the total revenue) against the price. For the Time Tracker<br \/>\n3000, this looks something like this. (I&#39;ve plotted the five data points we&#39;ve already<br \/>\ngot):\t<br \/>\n5<br \/>\n<br \/>\nFrom the diagram, you can see you should price the Time Tracker 3000 at around<br \/>\n$200. It&#39;s not where you&#39;ll sell the most units, but it&#39;s where you&#39;ll make the most<br \/>\nmoney.<br \/>\nPlotting a demand curve is, in theory, straightforward, but in practice it is way harder.<br \/>\nIn the real world, you don&#39;t know what the shape of the demand curve is, or where your<br \/>\ncurrent price sits on it. On some curves, and on some points of the curve, you&#39;d be<br \/>\nright to increase your prices: the reduction in the number of people buying your<br \/>\nproduct will be outweighed by the increased revenue from each person. On other<br \/>\ncurves, or on other points of the same demand curve, increasing your prices will lead to<br \/>\na massive drop-off in sales and you&#39;ll lose money.<br \/>\nWhat&#39;s more, the shape of the demand curve depends on a bunch of variables: your<br \/>\ncompetitors, how they&#39;ll react to any price changes you make, your customers, their<br \/>\nenvironment and situation, and the type and quality of your product.\t<br \/>\n6<br \/>\n<br \/>\nChapter Two\t<br \/>\nPricing Psychology<br \/>\n<br \/>\nThe demand curve, discussed in the previous section, might be dynamic and depend on<br \/>\nmany factors, but you can still exert some influence on its shape. In this chapter, I&#39;ll<br \/>\ntalk about how people decide how much they&#39;ll pay for a product, and how you can<br \/>\nchange this.<br \/>\nBut first, you need to be able to answer a simple question: what is your product?<br \/>\nWhat Is Your Product?<br \/>\nYou might think that your software product is just the bits and bytes that your<br \/>\ncustomers download, or the HTML pages that they view, but you&#39;d be wrong. In reality,<br \/>\nyour product is much broader than that. It&#39;s not just the software or the web site: it&#39;s<br \/>\nthe documentation, the help required to get it working and the promise of support<br \/>\nwhen things go wrong. It&#39;s the future roadmap of the product, the pledge to carry on<br \/>\ndeveloping future versions. In some cases, it&#39;s a dream, an experience; a way of life.<br \/>\nOne of the clearest examples comes from the accounting industry. At Red Gate, we use<br \/>\nSage&#39;s accounting software. We&#39;re not the only ones: Sage is a software business with<br \/>\n6.3 million customers. They employ 13,000 people and that has a market capitalization<br \/>\nof nearly four billion pounds. It dominates the accounting market in the UK, and has<br \/>\nseen off concerted attacks from, among others, Microsoft and Intuit.<br \/>\nBut their software sucks.<br \/>\nIt&#39;s slow, and it&#39;s hard to use. When I first used it in 1999, the buttons on the toolbar<br \/>\ndidn&#39;t depress when I clicked them: they were just static pictures painted on a grey<br \/>\nbackground. The application is so ugly that the product walkthroughs on the Sage web<br \/>\nsite barely feature the product itself.<br \/>\nThese two facts\u2013the awfulness of the product and the magnitude of its success\u2013can be<br \/>\nreconciled if you understand that Sage&#39;s\tproduct\tis more than just the software.\t<br \/>\nWhen you buy Sage software, you are not just buying software. You are buying<br \/>\nreassurance\t: when the tax laws change, the software will get updated too. You are\t<br \/>\nbuying\tfamiliarity\t: if you buy Sage, the odds are that your accountant or bookkeeper\t<br \/>\nwill already be able to use it. You&#39;re buying\tsupport\t: if you don&#39;t understand some of\t<br \/>\n8<br \/>\n<br \/>\nthe accounting codes or procedures you need, then you can phone somebody for help.<br \/>\nTens of thousands of people call Sage&#39;s help line per day.<br \/>\nThe reason that Sage is so dominant in the UK is because Sage understands exactly<br \/>\nwhat their product is. You need to do the same.<br \/>\nPerceived Value<br \/>\nOnce you&#39;ve determined what your product is, you need to consider its value to your<br \/>\ncustomers. In the case of the Time Tracker 3000, let&#39;s say that it will save a particular<br \/>\ncustomer, Willhelm, three hours of work and that Willhelm prices his time at $50 an<br \/>\nhour. That means that Willhelm should buy the Time Tracker 3000 at any price under<br \/>\n$150, assuming he has nothing better to spend his money on. This is its\tobjective\t<br \/>\nvalue\t.\t<br \/>\nOf course, this assumes that Willhelm is the rational, decision-making machine that<br \/>\neconomists love. In fact, Willhelm is a flesh-and-blood, irrational human being who<br \/>\ndoesn&#39;t price his time and calculate costs and benefits. He has a\tperceived value\tof\t<br \/>\nthe Time Tracker 3000, which may or may not be linked to its objective value.<br \/>\nThe perceived value of a product may be higher than its objective value. In 2003,<br \/>\nGartner released a report that claimed that almost half of all customer relationship<br \/>\nmanagement (CRM) systems lie unused. That&#39;s several billion dollars of software that<br \/>\nsmart people thought was worth it, but wasn&#39;t.\t<br \/>\n9<br \/>\n<br \/>\nBeats by Dr. Dre headphones are another example where perceived values are much<br \/>\nhigher than objective values. Beats outline in promotional materials that with most<br \/>\nheadphones, listeners are not able to hear &#34;all\u201d of the music, and that Beats would<br \/>\nallow people \u201cto hear what the artists hear, and listen to the music the way they<br \/>\nshould\u201d. The actual boost in quality is, of course, much less than the boost in cost.<br \/>\nA product&#39;s perceived value may be lower than its objective value too. A few years ago,<br \/>\nI stumbled on somebody who insisted on using Excel as a word processor. According to<br \/>\nthis user, the additional expense of buying Microsoft Word wasn&#39;t worth the benefits<br \/>\nhe&#39;d gain. This was almost certainly a perception rather than a reality.<br \/>\nBack to Willhelm and the Time Tracker 3000. If you want to change how much Willhelm<br \/>\nwill pay for your product, then changing the product is one option, but only if you can<br \/>\nalso change his perception too. In fact, it turns out that you can change Willhelm&#39;s<br \/>\nperception of your product&#39;s worth without touching the product at all. That&#39;s one of the<br \/>\nthings marketing is for.<br \/>\nHow People Set Their Perceptions<br \/>\nSo how do people generate their perceived value of a product? How do they decide<br \/>\nhow to think?<br \/>\nFor a start, it&#39;s extremely hard for them to do so in a vacuum. Try asking a British<br \/>\nMember of Parliament how much a pint of milk costs, a contestant on The Price is Right<br \/>\nfor the value of a chest of drawers, or the average supermarket shopper how much<br \/>\nthey should pay for a bottle of bleach. They&#39;ll struggle.<br \/>\nPeople base their perceived values on\treference points\t. If you&#39;re selling a to-do list\t<br \/>\napplication, then people will look around and find another to-do list application. If they<br \/>\nsearch the internet and discover that your competitors sell to-do list applications at<br \/>\n$100 then this will set their perception of the right price for all to-do list applications.<br \/>\nWhen Microsoft released DOS 1.0 in 1982, they set a price of $50. At the time, an<br \/>\noperating system for a mass-market computer was a brand new category. Since<br \/>\nconsumers had no reference point, $50 became accepted as a &#39;fair&#39; price. When IBM<br \/>\nlaunched OS\/2 1.0 in 1989 and priced it at $340, consumers baulked. DOS and OS\/2<br \/>\nwere very different operating systems, and $340 could well have been a fair reflection\t<br \/>\n10<br \/>\n<br \/>\nof the additional benefit consumers would have extracted from the more-advanced OS\/<br \/>\n2, but Microsoft had already defined the reference point, and when IBM tried to<br \/>\nchallenge it they failed.<br \/>\nThis doesn&#39;t mean you need to copy the reference point. If your product genuinely is<br \/>\nbetter than your competitors&#39;, and you can demonstrate the value of this difference, or<br \/>\ncreate a perception of that value, then you can charge more.<br \/>\nOf course, if your product is significantly less valuable than your competitors&#39; then you<br \/>\nmay have no choice but to charge less. Competing on price may be the only option you<br \/>\nhave. Take pharmaceuticals. My local supermarket stocks thirty or so different sorts of<br \/>\npainkillers. You can buy a pack of 16 Nurofen for \u00a31.97 ($3.40), or a pack of 16 Tesco<br \/>\nanalgesics for \u00a30.32 (about 50 cents). The physical good \u2013 the 200mg of ibuprofen \u2013 is<br \/>\nidentical in both the generic and named brand product. But don&#39;t forget that the entire<br \/>\nproduct is more than the chemicals. It includes the marketing, brand name and<br \/>\npackaging. Using this wider definition, Nurofen is the superior product, and the only<br \/>\nway Tesco can compete is on price.<br \/>\nThe value people perceive your product to have can depend on their\ttaste\tand\t<br \/>\nculture\t. Some people are passionate about good wine and will pay $50 for a bottle,\t<br \/>\nbut others like the taste of $5 wine just fine. As Dave O&#39;Flynn wrote to me:<br \/>\n\u201c\u201c\t11<br \/>\n<br \/>\nI never thought Apple products were worth the premium until I<br \/>\njoined Atlassian. 12 months later, I gladly paid a large<br \/>\npremium for a Macbook Air. The people I was surrounded with<br \/>\nvalued design and elegance. Prior to that, I was surrounded by<br \/>\npeople that valued bang for the buck and my laptop then was<br \/>\na generic AMD that weighed a ton in comparison. I was<br \/>\nessentially the same person; the changes were in the<br \/>\nexpectations and sense of value of those around me.\t<br \/>\nHow much money they have affects their perception of value. Dennis Kozlowski, ex-<br \/>\nCEO of Tyco, felt that $15,000 was a reasonable price to pay for a dog-shaped umbrella<br \/>\nstand, but most of us don&#39;t.<br \/>\nFinally, and importantly,\tknowledge\tinfluences the value people place on products\t<br \/>\nhugely. A laptop with a 2.7Ghz quad-core Intel Core i7 processor, 8GB of RAM and a<br \/>\nRetina display is worth more than one with an N-Series Intel Atom Processor and a VGA<br \/>\ndisplay to me, but not to my mum.<br \/>\nIncreasing Perceived Values<br \/>\nThe pharmaceutical industry holds another good example of how marketing can<br \/>\nincrease the perceived value of a product, without changing its substance. In 1981,<br \/>\nwhen Glaxo wanted to release Zantac, their anti-ulcer drug, they faced a marketplace<br \/>\ndominated by SmithKline&#39;s Tagamet. Although Glaxo felt their drug was more effective<br \/>\nthan SmithKline&#39;s, the US FDA rated Zantac as providing little or no benefit over<br \/>\nexisting treatments. Rather than marketing Zantac as a me-too product, at a similar<br \/>\nprice to Tagamet, Glaxo decided to spend heavily on saturating their sales and<br \/>\nmarketing channels. This ubiquitous promotion increased Zantac&#39;s perceived value,<br \/>\nand they were able to price the product higher to reflect this added value. By the end<br \/>\nof the 1980s, Zantac had knocked Tagamet off its perch as the best selling drug in the<br \/>\nworld.<br \/>\nHere are some more ways of increasing the perceived value of your product:<br \/>\nIncrease its objective value.\tPerceived and objective values aren&#39;t identical, but\t<br \/>\nthey&#39;re still correlated. As Joel Spolsky wrote in 2006\t[1]:\t<br \/>\n12<br \/>\n<br \/>\nWith six years of experience running my own software<br \/>\ncompany, I can tell you that nothing we have ever done at Fog<br \/>\nCreek has increased our revenue more than releasing a new<br \/>\nversion with more features. Nothing. The flow to our bottom<br \/>\nline from new versions with new features is absolutely<br \/>\nundeniable. It&#39;s like gravity. When we tried Google ads, when<br \/>\nwe implemented various affiliate schemes, or when an article<br \/>\nabout FogBugz appears in the press, we could barely see the<br \/>\neffect on the bottom line. When a new version comes out with<br \/>\nnew features, we see a sudden, undeniable, substantial, and<br \/>\npermanent increase in revenue.\t<br \/>\nGive your product a personality.\t37signals may not sell the best project\t<br \/>\nmanagement software in the world, but it has personality. The 37signals team stands<br \/>\nfor something: uncompromising simplicity. Want an extra feature? Tough. If you want<br \/>\nfeatures, buy something else.<br \/>\nLink your product to yourself,\tand then define, and promote, yourself as an expert.\t<br \/>\nIn its early days, before his company was bought by Symantec in 1990, Norton Utilities<br \/>\nand Peter Norton were synonymous. All of Norton&#39;s products featured a picture of<br \/>\nhimself, with his arms crossed.<br \/>\nMake people love your brand.\tInnocent Smoothies have become huge very quickly\t<br \/>\nby filling their packaging with small, friendly little notes and cute Easter eggs. They<br \/>\nmake it clear that Innocent is built and run by human beings, and focus on natural, not-<br \/>\nfrom-concentrate juices and smoothies. People love this friendly, congenial approach.<br \/>\nProvide a better service.\tWhen somebody buys software, they want reassurance\t<br \/>\nthat it&#39;s going to work and that you&#39;ll be around if it doesn&#39;t. If you&#39;re a small company<br \/>\nwith big competitors, this is something you can do better than they can. Capitalize on<br \/>\nit.<br \/>\nCreate a tribe.\tProducts can be symbols of belonging. If you can turn your product\t<br \/>\ninto a badge that people wear to make a statement about who they are, which groups<br \/>\n\u201c\u201c\t<br \/>\n1.http:\/\/www.joelonsoftware.com\/items\/2006\/12\/09.html\t<br \/>\n13<br \/>\n<br \/>\nthey belong to and those they don&#39;t, then that&#39;s valuable. When Black and Decker<br \/>\nintroduced its DeWALT line of drills, it went to building sites and lumber yards at lunch<br \/>\ntimes to hand out pulled pork sandwiches, give product demos and hold drill-off<br \/>\ncompetitions, with prizes. They went to NASCAR races and rodeos, where their end<br \/>\nusers hung out. This meant a lot of professionals purchased DeWALT drills, and created<br \/>\na tribe around them. Amateur DIYers don&#39;t\tneed\tto spend $400 on a DeWALT drill, but\t<br \/>\nthey like feeling part of the &#39;professional&#39; tribe too.<br \/>\nRemind people of how much work you&#39;ve put into your product.\tPeople are\t<br \/>\nmore likely to pay for years of your time than for an easily-copied software product.<br \/>\nThe twenty year old Bill Gates used this technique in his now-famous &#39;open letter to<br \/>\nhobbyists&#39; in the Homebrew Computer Club newsletter in 1976:\t<br \/>\nAlmost a year ago, Paul Allen and myself, expecting the hobby<br \/>\nmarket to expand, hired Monte Davidoff and developed Altair<br \/>\nBASIC. Though the initial work took only two months, the three<br \/>\nof us have spent most of the last year documenting, improving<br \/>\nand adding features to BASIC. Now we have 4K, 8K,<br \/>\nEXTENDED, ROM and DISK BASIC. The value of the computer<br \/>\ntime we have used exceeds $40,000.\t<br \/>\nAppeal to people&#39;s sense of fairness.\tWhen coffee shops charge an extra 10 cents\t<br \/>\nfor coffee made with Fairtrade beans, they&#39;re lining their pockets with your ethics. How<br \/>\nmuch of those ten cents go to the farmer who originally farmed the quarter of an ounce<br \/>\nof coffee beans that went into your Fairtrade latte? Under a penny.<br \/>\nSell the experience.\tStarbucks do this brilliantly. At its core, a cup of coffee is just a\t<br \/>\nfew beans, worth pennies. If you crush the beans up and put them in a packet on a<br \/>\nsupermarket aisle, you can charge a few more cents. If you mix it with hot water and<br \/>\nmilk, you can charge $1.50. But if you place it in a pleasant environment, with<br \/>\ncomfortable furniture and chilled music playing, you can ask for $3.50. You&#39;re not just<br \/>\nbuying a coffee, you&#39;re buying some time in the Starbucks experience.<br \/>\nUltimately, it comes down to\tdifferentiating your product.\tIt almost doesn&#39;t matter\t<br \/>\non what \u2013 features, benefits, the way that you sell, the service that you provide, the<br \/>\ncountry you&#39;re based in \u2013 more or less anything will do.<br \/>\n\u201c\u201c\t<br \/>\n14<br \/>\n<br \/>\nSignposts<br \/>\nNow that you know that customers will find reference points to compare your product&#39;s<br \/>\nprice against, you should do all you can to encourage favorable references and<br \/>\ndiscourage unfavorable ones. If you want to sell a to-do list at $200, when the market<br \/>\nprice is $100, then you need to add a couple of features so your customers cannot<br \/>\nmake a direct comparison, and then promote comparisons to other companies&#39; $300<br \/>\nproductivity suites, not their to-do lists. At the same time, avoid all comparisons to<br \/>\nopen source alternatives.<br \/>\nIf your customers can&#39;t find a reference point for your product, then they look for<br \/>\nproxies, or\tsignposts\t. Supermarkets take advantage of this: consumers decide\t<br \/>\nwhether luxury ice cream (something they don&#39;t buy regularly) is reasonably priced<br \/>\nbased on whether diet coke (something they buy all the time) is good value. If a<br \/>\nsupermarket sells a can of diet coke for $4, consumers assume all their other products<br \/>\nwill be expensive too.<br \/>\nSay you sell two products: the Time Tracker 3000 and the Task List 400, a to-do list<br \/>\napplication. When somebody thinks of something they need to do, they store it in the<br \/>\nTask List 400. Later on, they can prioritize their tasks, split them up into sub tasks,<br \/>\ntrack their progress and smugly mark off the tasks as done.<br \/>\nLet&#39;s say the Time Tracker 3000 has no competitors, but the Task List 400 has plenty.<br \/>\nYour customers will judge your Time Tracker 3000 price on how you&#39;ve priced the Task<br \/>\nList 400. Charge a reasonable $25 for your to-do list application and customers will<br \/>\ntake your word that $300 is a good price for the Time Tracker 3000. Charge $1000 for<br \/>\nthe first app, and they&#39;ll assume you&#39;re fleecing them on the new one too.<br \/>\nIf your product is unique, and customers can find no reference points or signposts, then<br \/>\nyou have a chance to set your customers expectations, and define their perceptions. If<br \/>\nyou tell your customers that the Time Tracker 3000 is worth $300, then the odds are<br \/>\nthey&#39;ll believe you. We&#39;ve already seen how Microsoft did that with the first version of<br \/>\nMS-DOS.<br \/>\nIf you have\tcompetitors\tin your market, then your customers will be more conscious\t<br \/>\nof cost, but if your product\tcreates a new category\t, then early adopters are less\t<br \/>\nlikely to be price sensitive. If you can create a teleporter, a brand new category of<br \/>\nproduct, that will beam you, unharmed, from New York to Paris then not only can you\t<br \/>\n15<br \/>\n<br \/>\ndefine your price, but you can also raise your price from $20,000 to $25,000 and<br \/>\npeople will still buy it. But if you create a car, a new product within a category that<br \/>\nalready exists, and increase your price from $20,000 to $25,000 then your sales will<br \/>\nsuffer.\t<br \/>\n16<br \/>\n<br \/>\nChapter Three\t<br \/>\nPricing Pitfalls<br \/>\n<br \/>\nSo far, we&#39;ve looked at some economic theory, and the psychology of pricing.<br \/>\nHopefully, you&#39;ve now got some idea of how to set a price. But there are some other<br \/>\nfactors to bear in mind too, and some pitfalls to watch out for.<br \/>\nCompetitors<br \/>\nWhen you set your product&#39;s price you need to think about how your competitors will<br \/>\nreact. If you undercut them, will they start a price war? Even if your competitor has a<br \/>\nhigh-cost business model and cannot compete on price in the long term then there&#39;s a<br \/>\nrisk they&#39;ll respond in kind if you pose a serious enough threat, and just hope you go<br \/>\nout of business before they do.<br \/>\nThe airline industry gives the best example of the futility of starting a price war. On<br \/>\nSeptember 26th 1977, Freddie Laker&#39;s first ever Skytrain flight to New York took off<br \/>\nfrom London Gatwick. The price for the return flight was $238.25 (plus an extra few<br \/>\ndollars for a meal), well under half the price of rivals&#39; tickets.<br \/>\nFive years later, Laker Airways was bust, the victim of the vicious, dirty price war that it<br \/>\nhad initiated. As Laker had found out, and as EOS Airlines (founded 2004, closed 2008),<br \/>\nSilverjet (founded 2006, failed 2008) and MaxJet (founded 2003, failed 2007)<br \/>\nsubsequently relearned, taking on an incumbent on the basis of price is highly risky at<br \/>\nbest, suicidal at worst, especially when your competitors cannot afford to lose and have<br \/>\nno option but to fight to the death.<br \/>\nIf you are going to compete on price, then you should minimize the possibility of a<br \/>\ncounter-reaction from your competitors. Don&#39;t bang your drum and tell the press how<br \/>\nyou&#39;re going to destroy them (a mistake that Marc Andreessen of Netscape made when<br \/>\nhe said \u201cwe&#39;re gonna smoke &#39;em\u201d, referring to Microsoft (or \u201cthose idiots up in<br \/>\nRedmond\u201d as Andreessen put it). Focus on their marginal customers and hope that by<br \/>\nthe time they notice you it will be too late.<br \/>\nOn the other hand, if you set your product price too high, will other competitors<br \/>\nemerge? Price the Time Tracker 3000 at $10,000 and you could create the market, only<br \/>\nfor a competitor to produce the Tyme Trakka 3000, undercut you, and steal your<br \/>\nbusiness.\t<br \/>\n18<br \/>\n<br \/>\nMicrosoft is famed for this. They wait for competitors (and often partners) to prove<br \/>\nmarkets with low volume, high price products \u2013 whether it&#39;s CRM, testing tools or<br \/>\nbusiness intelligence \u2013 and then jump in with a low-cost, high-volume model.<br \/>\nFairness<br \/>\nHowever you price your product, remember that consumers have an acute, although<br \/>\noften irrational, sense of fairness. Think twice before you betray that.<br \/>\nBooks provide a good example. An economist would point out that I derive the same<br \/>\nvalue (traditional economics is all about value) from reading a paperback version of my<br \/>\nfavourite piece of erotic fiction &#39;Fifty Shades of Grey&#39; as I do from reading the electronic<br \/>\nversion [Jamie wrote this bit, not Neil - editor]. But the list price is very similar (~\u00a37).<br \/>\nIt&#39;s just not fair that short-sighted book publishers charge the same for paper as they<br \/>\ndo for electrons. I feel screwed over, and I don&#39;t like it.<br \/>\nPirates<br \/>\nIf your price is way off whack, you will provide an opening for a special type of<br \/>\ncompetitor: the pirate. Price software too high, or at a price point that most people<br \/>\njudge &#39;unfair&#39;, then be prepared to be ripped off in return.<br \/>\nIf you&#39;re observing widespread piracy of your software, perhaps it&#39;s\tjust too expensive\t?\t<br \/>\nBut pirates can also be your friends, in two ways:<br \/>\nFirstly, if your strategy is to achieve world domination by providing a product to every<br \/>\npotential customer, at a price he or she can afford, then pirates provide a cheap back<br \/>\nchannel. They put a copy of your software into the hands of people who will not pay,<br \/>\ncannot pay, are too dishonest or too unprincipled to pay, or who simply don&#39;t value<br \/>\nyour work that much. However, a pirated copy will end up, eventually, in the hand of<br \/>\nsomebody who\twill\tpay.\t<br \/>\nThat&#39;s how early shareware software operated. In the early 1980s, bulletin boards and<br \/>\nuser groups were a network commonly used to pass around pirated software. In 1982,<br \/>\nAndrew Fluegleman and Jim Knopf piggy-backed onto this pre-existing network, added\t<br \/>\n19<br \/>\n<br \/>\na notice in their software asking people to pay them if they liked their software, and<br \/>\ninvented shareware.<br \/>\nThe second reason that pirates can be your friends is that they are a bellwether. They<br \/>\nindicate the existence of a market failure. Most people aren&#39;t natural crooks, but high<br \/>\nprices can force them to do things against their better nature. Apple realized that the<br \/>\nsuccess of illegal download sites indicated the need for cheap, downloadable music.<br \/>\nTheir strategy of satisfying people&#39;s needs worked far better than the ostrich-like<br \/>\nbehavior of the music labels.<br \/>\nSwitching Costs<br \/>\nIf you&#39;re trying to persuade people to switch to your product from a competitor&#39;s then<br \/>\nyou&#39;ll need to position the price to overcome the switching costs your customers face.<br \/>\nSay you&#39;re trying to persuade a customer to switch from his garbage $500 word<br \/>\nprocessor to your superior $100 one. First of all, you&#39;ll need to price to overcome the<br \/>\neconomic switching costs. It&#39;ll take him time, and therefore money, to convert his files<br \/>\nto a new format and to learn the new menu layouts.<br \/>\nSecondly, you&#39;ll need to overcome the\tpsychological\tswitching costs. People\t<br \/>\novervalue what they have, and undervalue what they don&#39;t have. I purchased three<br \/>\nbalsawood penguins from the Kontiki exhibition in Oslo. They cost me ten dollars. Care<br \/>\nto raise your price? Didn&#39;t think so, but I wouldn&#39;t sell them to you for even a hundred. I<br \/>\nbet your house is probably full of similar junk too.<br \/>\nAnother powerful psychological factor people struggle to overcome is the emotional<br \/>\nattachment to money they&#39;ve already spent. Rationally, it&#39;s gone. It&#39;s a sunk cost. Your<br \/>\ncustomer shouldn&#39;t care that he&#39;s already spent $500 on his garbage word processor.<br \/>\nBut he does.<br \/>\nThere are some things you can do to mitigate switching costs, and even to use them in<br \/>\nyour favor. Here are a couple of examples. Open Office, which includes open source<br \/>\nword processing and spread sheet applications, lets you open files saved by Microsoft<br \/>\nWord. Early versions of Microsoft Word not only opened WordPerfect files, but had a<br \/>\ndedicated section in the help for WordPerfect users, and even allowed you to use the<br \/>\nWordPerfect shortcut keys.\t<br \/>\n20<br \/>\n<br \/>\nHere&#39;s another example. If you decide to stop using FogBugz within ninety days of your<br \/>\nfree trial expiring, then Fog Creek will refund you all the money you&#39;ve spent, no<br \/>\nquestions asked.<br \/>\nThese strategies have two effects: first of all, they reduce the psychological and<br \/>\neconomic impact of switching to Word or FogBugz. Secondly, once you have switched,<br \/>\nyou&#39;ll have invested time and energy into using the new software, and will have<br \/>\nincurred a whole load of new switching costs, which will then stop you from switching<br \/>\nback.<br \/>\nShould You Take Your Costs Into Account?<br \/>\nClearly, you cannot price your software for less than it costs you to produce, and sell,<br \/>\neach unit. These are your marginal costs. You might think these costs are zero, but<br \/>\nthey are not.<br \/>\nYou need to find potential customers and persuade them to buy. If you have a sales<br \/>\nteam then you might need to pay them commission. It will cost you money to support<br \/>\ncustomers, and chase the customers who don&#39;t pay.<br \/>\nIf you&#39;re planning on not charging the majority of your users, then think very carefully<br \/>\nabout the cost of each additional user. If you think it is zero then you are almost<br \/>\ncertainly wrong. If you&#39;re running a web site, then each additional user will cost you<br \/>\nstorage space, CPU cycles and bandwidth. This might be a very low cost \u2013 fractions of a<br \/>\npenny, even \u2013 but if you need huge numbers of users to make money then small costs<br \/>\nmultiplied by vast numbers can equal big outlays.<br \/>\nTake YouTube. It&#39;s a free service and, theoretically, supported by advertising. The cost<br \/>\nof serving each additional video is tiny but YouTube see 4 billion video views globally<br \/>\nevery day. This means that in 2012 it will serve up an estimated 1.5\ttrillion\tvideo\t<br \/>\nstreams. Multiply together the tiny cost and the large volume and you can understand<br \/>\nwhy YouTube costs Google an estimated $1.5 billion a year to run. It nowhere near<br \/>\ncovers its costs through advertising revenue.<br \/>\nAlso, don&#39;t forget about support and customer engagement. With the rise of social<br \/>\nnetworks as legitimate business tools, the cost of looking after your users in this space<br \/>\nbecomes important. If your website goes down, you may need to respond to customer&#39;s\t<br \/>\n21<br \/>\n<br \/>\ncomplaints one-by-one on Twitter. If somebody&#39;s sharing information about your<br \/>\nsoftware on their Facebook account, you will want to jump into the conversation and<br \/>\nask if you can help any further. These interactions cost time, and ultimately, money.<br \/>\nPaperback Software offers another example of how misunderstanding how your<br \/>\nsoftware is sold, and failing to account for your costs, can lead to catastrophe. When<br \/>\nAdam Osborne set up Paperback Software in 1984, it was founded on the premise that<br \/>\nsoftware cost too much. They released VP-Planner for $99.95 in 1986 and marketed it<br \/>\ndirectly against the $500 Lotus 123. Back in the 1980s, most software was sold through<br \/>\ndealers. The dealers earned a commission for every piece of software they sold, but so<br \/>\nhated the low margins on the low cost VP-Planner that they bad-mouthed it and<br \/>\nencouraged people to buy the high-cost, high-margin Lotus 123 alternative.<br \/>\nFurthermore, since VP-Planner was essentially a direct copy of Lotus 123, customers<br \/>\ndemanded as much support for the cheaper product as the more expensive one,<br \/>\ndestroying any profit that Paperback Software made. Paperback succeeded in harming<br \/>\nLotus&#39;s market share, but failed to earn enough money to defend themselves against<br \/>\nthe lawsuit that Lotus launched.<br \/>\nIf the price your customers are willing to pay is lower than what it costs you to sell your<br \/>\nsoftware, then you haven&#39;t got a business and your product will flop. You need to cut<br \/>\nyour cost of sales, or change your pricing mechanism so customers end up paying<br \/>\nmore over the lifetime of the product.<br \/>\nWhen Panasonic launched the 3DO, its gaming console, in 1994, Time Magazine<br \/>\nnominated it its product of the year. With a 32-bit RISC processor, custom math co-<br \/>\nprocessor and 2MB of RAM, it was far ahead of its time. But Panasonic priced it at $699,<br \/>\nway above its competition and much higher than what even its target market of early<br \/>\nadopters could bear to pay. That, combined with muddled marketing, caused it to<br \/>\nbomb.<br \/>\nOther games console manufacturers learned from this mistake. When the PS3 and<br \/>\nXbox 360 were launched, they cost more to produce than the selling price that the<br \/>\nmarket could bear, so Sony and Microsoft charged consumers a low price, and<br \/>\naccepted that they would lose money (up to $300) on each console sold. They then<br \/>\nrecovered the revenue through royalties on games people bought. The real price of the<br \/>\nconsole is hidden; buried in a clever pricing model.\t<br \/>\n22<br \/>\n<br \/>\nWith the Wii, Nintendo took a different approach. They wanted to reach a much wider<br \/>\nmarket than their competitors&#39; 18 - 35 year old male sweet spot, but realized that older<br \/>\npeople, housewives and families would pay less for a console than hardcore gamers<br \/>\nwould. So they cut their cost of manufacture and used cheaper, slower components.<br \/>\nWhen the Wii was launched in September 2006, Nintendo made a profit on every<br \/>\nconsole sold. That made the games cheaper to produce too, since royalty payments<br \/>\ncan be lower, but not necessarily cheaper to buy. Why? By now, the answer should be<br \/>\nobvious to you.<br \/>\nYour pricing may need to change over the course of your product&#39;s lifetime. In the early<br \/>\nstages, you need to attract new users and convince them that your software is the one<br \/>\nfor them. Lower prices and big marketing efforts are usually crucial to accumulating a<br \/>\ncommunity.<br \/>\nAs your product gets maturer, your customers will be on the lookout for more variety<br \/>\nand options. By this point they&#39;ll have been using your software for a while and will<br \/>\nunderstand its features and how they fit it into their workflow. Providing them the<br \/>\nopportunity to customise their experience will make them evangalise your product<br \/>\neven more. We&#39;ll talk more about this in Chapter 4.<br \/>\nYou&#39;ll notice that there&#39;s one factor I&#39;ve not mentioned, and that&#39;s how much your<br \/>\nproduct has cost to develop. So far I&#39;ve talked about marginal costs \u2013 how much it costs<br \/>\nto produce, or sell, each additional unit of your software. Your up-front cost is different.<br \/>\nYou might have spent one hundred dollars developing your product, or a million, but<br \/>\nthat money is all spent. Gone. It&#39;s a sunk cost.<br \/>\nWhat matters now is not how much you&#39;ve spent, but what people are prepared to pay.\t<br \/>\n23<br \/>\n<br \/>\nChapter Four\t<br \/>\nAdvanced Pricing<br \/>\n<br \/>\nUp to now, we&#39;ve considered selling single products. But what happens when you have<br \/>\nseveral products to sell, or sell multiple versions of the same product? Or even, what if<br \/>\nyou&#39;re selling a service rather than a product?<br \/>\nVersioning<br \/>\nEach of your potential customers has a price they&#39;ll buy your product at. Revisiting our<br \/>\nprevious example, Belinda (the bargain hunter) and Stewart (the student) will only use<br \/>\nthe Time Tracker 3000 if it&#39;s free. Willhelm will pay up to $150 and Pat&#39;s maximum<br \/>\nprice is $400. Let&#39;s say Ernest will pay up to $600.<br \/>\nHere&#39;s a chart of the revenue you&#39;ll make at each price point:<br \/>\nPricing\tWho buys\tRevenue\t<br \/>\n$0\tEverybody\t$0\t<br \/>\n$150\tWillhelm, Pat, Ernest\t$450 (three people @ $150)\t<br \/>\n$400\tPat, Ernest\t$800 (two people @ $400)\t<br \/>\n$600\tErnest\t$600 (just Ernest)\t<br \/>\n$1000\tNobody\t$0\t<br \/>\n25<br \/>\n<br \/>\nIf these five people are your entire target market, then, to maximise your total<br \/>\nrevenue, you should price the Time Tracker 3000 at $400. It&#39;s the best single price, but<br \/>\nyou&#39;ll lose out on sales to Willhelm, and you&#39;ll lose out on the extra revenue that Ernest<br \/>\nwould have paid.<br \/>\nIf there had been some way to sell the product to each customer at the maximum price<br \/>\nthat they could afford to pay, you would have been able to sell $1150 of software.<br \/>\nThat&#39;s what&#39;s versioning is about. It&#39;s a mechanism of segmenting your users according<br \/>\nto their willingness to pay. You figure out if you can group your customers in different<br \/>\nways, and then see if those groups are willing to pay different prices for your product.<br \/>\nHere are some of the ways of doing it:\t<br \/>\n\u2022\tBy feature.\tFor example, you can have &#39;standard&#39; and &#39;pro&#39; versions of tools.\t<br \/>\nThis is extremely common in the software business. Microsoft&#39;s Visual Studio<br \/>\n2008 came in five different versions: Express (free), Standard ($299),<br \/>\nProfessional ($799), Team System ($5,469) and Team Suite ($10,939). That&#39;s a<br \/>\nprice for everybody, with features to match, from the cash-poor hobbyist to the<br \/>\nrich, blue chip enterprise developer. In the Time Tracker 3000 example, you<br \/>\nmight create a professional edition that lets people compare how their usage of<br \/>\ndifferent products compares with other people doing similar work.\t<br \/>\n\u2022\tBy availability.\tSome of your customers might be prepared to pay more to get\t<br \/>\nyour product quickly. Hardback books are a good example of this. They have<br \/>\nthe same content as paperbacks, but are packaged differently and aimed at<br \/>\npeople who cannot wait for the content. For the Time Tracker 3000, you could<br \/>\nsell an additional subscription service that gets customers early access to<br \/>\nsoftware.\t<br \/>\n\u2022\tBy demographic.\tStudents have less money than businesses, hobbyists than\t<br \/>\nprofessionals and school kids than baby boomers. You could provide a version<br \/>\nof the Time Tracker 3000 which students could get, but only if they prove<br \/>\nthey&#39;re in full time education.\t<br \/>\n\u2022\tBy geography.\tCustomers in the USA will pay more for the same product than\t<br \/>\nthose in India and China. Microsoft, to compete with the threat of open source,<br \/>\nprovides a cut-down &#39;starter&#39; edition of its Vista operating system, available<br \/>\nonly in poorer countries such as India and Mexico. The Time Tracker 3000 might<br \/>\nbe available in India for 10% of its US cost, but be localized into Hindi,<br \/>\nrendering it useless to Westerners.\t<br \/>\n\u2022\tBy industry.\tPerhaps architects, or software developers or aircraft designers\t<br \/>\nhave specific needs, and perhaps your software can be customized to suit\t<br \/>\n26<br \/>\n<br \/>\nthem. The Time Tracker 3000 could come in a special edition, aimed at law<br \/>\nfirms, that not only tracks application usage, but also bans certain applications.\t<br \/>\n\u2022\tBy platform.\tMac users might be willing to pay more money for your software\t<br \/>\nthan Windows users, or vice versa. You could sell a Time Tracker 3000 for the<br \/>\nMac at a higher price than the Windows version.\t<br \/>\nOf course, you need to be aware of the dangers of versioning too. You need to make<br \/>\nsure that the features you choose for each version appeal to the segment you&#39;re<br \/>\ntargeting. For example, if you introduce a &#39;Lite&#39; version of your product, you need to be<br \/>\nsure that professional users won&#39;t downgrade to it.<br \/>\nWhen attempting to version by one of these criteria, and if your goal is happy<br \/>\ncustomers, then it&#39;s best to remember consumers&#39; keen sense of fairness. Adobe<br \/>\nattempt to version on geography; their Acrobat 9 Pro cost $449 in the US, but \u00a3445<br \/>\n($750) in the UK. Economically, this might make sense, but it still leaves me banging<br \/>\nmy keyboard in impotent rage. And is that good, in the long term, for Adobe?<br \/>\nVersioning has a couple of subtleties. Take a fast food restaurant that serves the<br \/>\nfollowing sizes of diet coke:<br \/>\nProduct\tPrice\t<br \/>\nSmall\t$1\t<br \/>\nMedium\t$1.50\t<br \/>\nLarge\t$2\t<br \/>\nThese prices have been chosen, presumably, to maximize the fast food chain&#39;s profits.<br \/>\nPeople with little money, or who aren&#39;t very thirsty, buy the small drink; those who are<br \/>\nmarginally thirstier buy the medium one and very thirsty people buy the large one. The<br \/>\nadditional fluid ounces cost the restaurant virtually nothing: this is all about finding a<br \/>\nprice point that works for everybody.<br \/>\nYou can also see the use of reference points here. Consumers see the &#39;small&#39; drink, and<br \/>\nconsider the &#39;medium&#39; drink a bargain (a lot more drink for just a few more cents).\t<br \/>\n27<br \/>\n<br \/>\nSo far, so blatant, but here&#39;s one subtlety: adding a &#39;jumbo&#39; drink will increase the sales<br \/>\nof the &#39;large&#39; drink, even if nobody ever buys the &#39;jumbo&#39; one. Adding more choices at<br \/>\nthe edges drives people to the middle of the range. They don&#39;t want to appear stingy,<br \/>\nor greedy, so go for the safety of the middle. In this example, adding a &#39;jumbo&#39; version<br \/>\non top shifts where the middle lies, so makes more money.<br \/>\nHere&#39;s the second subtlety: this only works if people can easily compare the products<br \/>\nbeing versioned. For the sodas, it works. The jumbo soda is clearly larger than the large<br \/>\nsoda, which is clearer larger than the medium soda, which is clearly larger than the<br \/>\nsmall soda. So people go for a safe option, somewhere in the middle.<br \/>\nBut the effect reverses if people struggle to compare the different versions of the<br \/>\nproducts. In that case, people flee the middle and head for the extremes. Take laptops.<br \/>\nSay you ask people to choose between the following products:<br \/>\nLaptop type\tFeatures\tPrice\t<br \/>\n\u2018Standard\u2019 laptop\tNormal features\t$1000\t<br \/>\nX100\tStandard + DVD player\t$1100\t<br \/>\nX102\tStandard + Wireless card\t$1100\t<br \/>\nX103\tStandard + Faster processor\t$1100\t<br \/>\nX104\tStandard + DVD + Wireless\t$1200\t<br \/>\nX105\tStandard + Wireless + Faster processor\t$1200\t<br \/>\n\u2018Extreme\u2019 laptop\tStandard + DVD + Wireless + Faster processor\t$1300\t<br \/>\nRather than migrating towards one of the middle options, people are pushed towards<br \/>\nthe edges. They go for the &#39;standard&#39; laptop or the &#39;extreme&#39; one. This is because it&#39;s<br \/>\nimpossible to compare the benefits of the different items being offered. Is a wireless<br \/>\ncard a better option than a faster processor? Or how about a DVD drive? As a result,<br \/>\npeople take an easy \u201call or nothing\u201d decision.\t<br \/>\n28<br \/>\n<br \/>\nWhen people are presented with a bunch of confusing options they cannot compare,<br \/>\ngoing for an extreme isn&#39;t their only option. They also have a tendency to defer: to<br \/>\nsimply not buy, or go for a competitor&#39;s product.<br \/>\nThis counter-intuitive behavior has some interesting consequences. If consumers are<br \/>\nfaced with a choice of, say, a Sharp or a Panasonic microwave then roughly half of<br \/>\nthem will plump for a Sharp and half for a Panasonic. If they are asked to choose a<br \/>\nmicrowave from a selection that contains a single Panasonic and multiple versions of<br \/>\nthe Sharps, then one of two things can happen.<br \/>\nIf they can easily compare the Sharps (for example, because they differ solely in price<br \/>\nand one other attribute, such as size or power), then more people will buy the Sharp<br \/>\nthan the Panasonic. This is a demonstration of how providing multiple versions of a<br \/>\nproduct will increase the product&#39;s sales.<br \/>\nOn the other hand, if they cannot easily compare the Sharps then the effect is<br \/>\nreversed. For example, if one Sharp has an adjustable speed turntable, another has a<br \/>\nmoisture sensor, one has programmable menus and another has a &#39;hold warm&#39; feature,<br \/>\nthen consumers will shun the Sharp, reject confusion and go for the Panasonic. This<br \/>\nshows how providing multiple versions of a product can decrease a product&#39;s sales.<br \/>\nHere are just some of the possible versions of Microsoft&#39;s Vista operating system:\t<br \/>\n29<br \/>\n<br \/>\nIs protecting against hardware failure more important than having all-in-one media<br \/>\ncenter functionality? And does being able to remotely access your business resources<br \/>\noutweigh being able to easily make DVDs? It&#39;s hard to tell, so consumers will tend to do<br \/>\none of three things:\t<br \/>\n\u2022\tGo to the extremes \u2013 buy Home Basic (at $199.95) or Ultimate (at $319.95)\t<br \/>\n\u2022\tDefer a decision \u2013 stick with Windows XP\t<br \/>\n\u2022\tBuy a competitor&#39;s product.\t<br \/>\nI bought a Mac \u2013 my first one ever.\t<br \/>\n30<br \/>\n<br \/>\nBundling<br \/>\nBundling is another way of giving your customers better value, persuading them to buy<br \/>\nand generating more revenue. Most straightforwardly, people love a bargain.<br \/>\nThe idea of getting $5140 of software for $1,595 (in the case of the SQL Toolbelt that<br \/>\nRed Gate sells) is clearly compelling.<br \/>\nBut even without price discounting, bundling makes sense.<br \/>\nSay you&#39;ve got two products, the Time Tracker 3000 and the Task List 400. Willhelm,<br \/>\nthe web start-up founder, is hyper-focused. Once he gets going on something, he&#39;ll see<br \/>\nit through to completion. But he struggles to organize the list of things he has to do.<br \/>\nPat, on the other hand, doesn&#39;t see much point in task lists, but she has the feeling that<br \/>\nshe wastes much of her working day, and would like to know how she spends it. Pat<br \/>\nand Willhelm are therefore willing to spend different amounts on each product.<br \/>\nIf you sell your products individually then, product-by-product, you need to choose the<br \/>\nmaximum price that the person who values that product the least will still pay. You<br \/>\nneed to price the Time Tracker 3000 at $150 (so Willhelm will buy it) and the Task List<br \/>\n400 at $150 (so Pat will buy). That means that Willhelm and Pat will each give $150 for<br \/>\neach product, and you generate revenue of $600.<br \/>\nLet&#39;s say you create a bundle of the Time Tracker 3000 and the Task List 400. Willhelm<br \/>\nand Pat will both pay $400 for the product they need (Task List 400 and Time Tracker\t<br \/>\n31<br \/>\n<br \/>\n3000, respectively). At that point, the bundle is worth $550 to Willhelm and $550 to<br \/>\nPat. Set the price at $550, sell the bundle to both people and you generate revenue of<br \/>\n$1100. Willhelm and Pat have got all the software they want, and you&#39;ve generated an<br \/>\nextra $500.<br \/>\nHowever, bundling has drawbacks too. When you bundle software together it becomes<br \/>\nharder for your customers to understand what they&#39;re paying for. In turn, that might<br \/>\nmean they are less likely to use it.<br \/>\nFor example, a diner eating a fixed price menu is more likely to skip coffee than a diner<br \/>\nwho&#39;s paid explicitly for the coffee. The coffee is bundled, so the disconnect between<br \/>\nwhat the diner is paying for and what he is consuming makes it easier to not consume.<br \/>\nFor software, if a customer is less likely to use a piece of bundled software then he<br \/>\nmight be less likely to buy a future version, or to continue to spend money on<br \/>\nmaintenance contracts. One way of counteracting this effect is to continue to be<br \/>\nexplicit about the worth of each item in a bundle.<br \/>\nMulti-user Licenses<br \/>\nMulti-user licenses are one more way of bundling software. But before you decide to<br \/>\noffer multi-user discounts to your customers, remember three things:\t<br \/>\n1.\tLarger companies tend to buy more copies of software since they have more<br \/>\nusers. Offer them a three for the price of two discount then they&#39;ll get a better<br \/>\ndeal than individual users and small businesses. Larger companies also have<br \/>\nmore money and tend not to be so price sensitive. This means that the poor are<br \/>\neffectively subsidizing the rich.\t<br \/>\n2.\tYou might lose sales in the long term. The company who paid for two licenses<br \/>\nmay have paid for three if you&#39;d asked.\t<br \/>\n3.\tOn the other hand, they might not have done. Everybody likes a discount, even<br \/>\nlarge companies.\t<br \/>\n4.\tLarger companies might have more money, but they can also have stricter<br \/>\npurchasing policies.\t<br \/>\nIt&#39;s hard to know which of these factors are the strongest in any given situation. I know<br \/>\nof one company who moved away from multi-user deals based on the first two reasons\t<br \/>\n32<br \/>\n<br \/>\nabove. They moved from selling a &#39;five for the price of three&#39; bundle to a simple 10%<br \/>\ndiscount per copy, for multiple copies.<br \/>\nIt turned out that their customers preferred the convenience of buying multi-user<br \/>\nbundles. If their customers had two users, they liked being able to buy a five user<br \/>\nlicense for the price of three and get the possibility of a bargain\u2013two &#39;free&#39; users\u2013if<br \/>\nanother person started using it. This outweighed the risk of overpaying, and never<br \/>\nhaving a third user. Two months and several hundred thousand dollars of lost revenue<br \/>\nlater, the company switched back to multi-user deals.<br \/>\nThe important point is that theory cannot tell you about the wisdom or otherwise of<br \/>\nmulti-user deals. The only way to find out is to try it out.<br \/>\nSite Licenses<br \/>\nYou need to be careful with site licenses. Sell a site license to Microsoft or Walmart and,<br \/>\nunless you&#39;ve customized your pricing accurately and high, you could be forgoing<br \/>\nenormous amounts of future revenue. If you insist on selling a site license then make<br \/>\nsure you define &#39;site&#39; well. Is it for a specific office, or country, or worldwide?<br \/>\nThe Purchasing Process<br \/>\nYou must consider your customers&#39; purchasing process when you set your prices. If<br \/>\nyou&#39;re selling to businesses, then there will be a number of thresholds that you need to<br \/>\nthink twice about before crossing. For example:<br \/>\nIf you sell a product at\t$10\tor under then an end user will charge it to his personal\t<br \/>\ncredit card and not claim it back.<br \/>\nUp to\t$50\t, he might charge it to his card and claim it back from the company he works\t<br \/>\nfor.<br \/>\nUp to\t$995\t, he might borrow his boss&#39;s company credit and charge it directly to the\t<br \/>\ncompany.\t<br \/>\n33<br \/>\n<br \/>\nAt$1000\t, he might have to fill in some paperwork and justify, strongly, his reason for\t<br \/>\npurchasing to his boss.<br \/>\nAt$5,000\the might have to talk to the head of his department.\t<br \/>\nAt$25,000\the might have to talk to his CEO.\t<br \/>\nAt each stage, not only does the cost increase, but the hassle does too. If you can<br \/>\nfigure out where these thresholds lie (and they move around as the state of the<br \/>\neconomy changes, and according to the characteristics of your customers), then it&#39;s<br \/>\nworth pricing your software just under a threshold rather than just over it.<br \/>\nOnce you cross a threshold, you can often move up to the next one relatively easily. It&#39;s<br \/>\neasier to persuade somebody to spend $10 instead of $1 than it is to get them to open<br \/>\ntheir wallet in the first place.<br \/>\nThis is yet one more reason to provide multi-user discounts and bundles. If you&#39;re<br \/>\nselling to an organization, then the individual you&#39;re selling to will help you cross the<br \/>\nthresholds, and once you&#39;re past a threshold he may even be keen to help you beyond<br \/>\nthere.<br \/>\nSay you&#39;ve persuaded Frank, the IT manager of Blue Door Software, to buy a one<br \/>\nhundred user license of the Time Tracker 3000. Frank has negotiated hard and you&#39;ve<br \/>\nagreed on a price of $25,000. Frank knows that he now needs to persuade Victor, the<br \/>\nCEO, to authorize this expenditure. Victor is a scary, busy man and hard to persuade.<br \/>\nFrank realizes that he may well need some copies of the Task List 400 at some point in<br \/>\nthe next six months, and doesn&#39;t want to have to persuade Victor twice. He also knows<br \/>\nthat, although Blue Door Software currently has one hundred employees, it will<br \/>\nprobably grow over the next twelve months. If he&#39;s going to ask for $25,000 to buy the<br \/>\nTime Tracker 3000, why not ask for $30,000 and get you to throw in some copies of the<br \/>\nTask List 400 for free? Or for $35,000 and ask for an extra fifty licenses? It&#39;s in Frank&#39;s<br \/>\ninterests, and yours, and Blue Door Software&#39;s.<br \/>\nFree<br \/>\nSome people argue that the price of software will inexorably be driven to zero.<br \/>\nEconomists have proven that in any efficient market, the cost of a good will be driven\t<br \/>\n34<br \/>\n<br \/>\ndown to its marginal cost of production. If you&#39;re one of many producers selling<br \/>\nwrenches then consumers will shop around for the cheapest wrench. If it costs $5 to<br \/>\nproduce the next wrench, then wrench manufacturers will compete on price,<br \/>\nundercutting each other and driving the price lower and lower, until it&#39;s at the lowest<br \/>\nprice that still allows them to make a profit: $5.01.<br \/>\nInformation, the theory goes, has zero marginal cost. It costs nothing to ship the next<br \/>\nset of bytes to your next customer. Therefore, the price that consumers will pay for<br \/>\nyour information, and the cost you must sell it for, will eventually approach zero. The<br \/>\nsuccess of open source operating systems such as Linux, the Apache web server and<br \/>\nthe Open Office suite seem to illustrate this point.<br \/>\nThis argument has a number of holes in it. For a start, as already discussed, you are<br \/>\nnot just selling bits and bytes. You&#39;re selling a whole bunch of stuff around it, including<br \/>\nsupport, documentation and hand-holding. Your customers are buying man-years,<br \/>\ndecades even, of your past, present and future blood, sweat and tears. Is that worth<br \/>\n$100? Or $1,000? Heck, yes, and you should tell that to your customers.<br \/>\nSecondly, there is no such thing as a commodity. Or, more accurately, there need not<br \/>\nbe such a thing as a commodity. Your job is to de-commodify what you are doing. If<br \/>\nyour potential customers consider your to-do list, or your word processor, accounts<br \/>\npackage, web site or iPhone app as just one of a hundred indistinguishable others, then<br \/>\nthe price you can charge will be driven ever downwards. You need to figure out a way<br \/>\nto either make it stand out, or impossible to compare.<br \/>\nIf Starbucks can de-commodify coffee and charge $4 for coffee beans and hot water, if<br \/>\nStormhoek can de-commodify grapes (the only wine maker I know of who sells branded<br \/>\nG-Strings), and if Perrier can de-commodify water, then you can certainly de-<br \/>\ncommodify the complicated software application that you have created.<br \/>\nDespite all this, there is no doubt that &#39;free&#39; holds a tremendous power over consumers.<br \/>\nAnd it&#39;s a power that you can harness.<br \/>\nFree Trials<br \/>\nFree trials let your customers try out your software for free, to make sure it fits their<br \/>\nneeds before they buy it. They don&#39;t even need to use the trial for you to benefit. The\t<br \/>\n35<br \/>\n<br \/>\nmere fact that customers could try out your software, if they wanted to, transmits a<br \/>\nstrong signal about its quality.<br \/>\nWhen customers do try out your software, it can increase its perceived value. In a<br \/>\nfamous psychology experiment, people who were able to hold a coffee mug were<br \/>\nwilling to pay significantly more for it than those who were just allowed to see it. People<br \/>\nstart to feel that they own an object before they buy it if they&#39;re allowed to use it and,<br \/>\nas we&#39;ve already seen, people value what they own more than what they don&#39;t.<br \/>\nFree trials aren&#39;t always possible. Red Gate used to sell a tool that let you recover<br \/>\ndeleted data from a SQL Server database. The free trial worked against it: people would<br \/>\ndownload it and recover their data before their free trial expired. Free trials only work<br \/>\nfor software that people use again and again, and where the free trial doesn&#39;t fix the<br \/>\nproblem by itself.<br \/>\nSimilarly, if people require a lot of hand-holding to use your software, or if it is of a low<br \/>\nquality, then free trials are unlikely to work.<br \/>\nThe freemium model involves providing a free version of your software for some<br \/>\npeople, and a paid-for version for others. Typically, the &#39;standard&#39; product will be free,<br \/>\nand the &#39;pro&#39; version will be paid for. Flickr, LinkedIn and Skype all use this model.<br \/>\nHowever, it&#39;s not clear that giving your software away for free is a great way to make<br \/>\nmoney, despite being extremely fashionable. At the very least, you need to be careful,<br \/>\nand make sure the free version is good enough to be useful, but not so useful that it<br \/>\ncannibalizes paid-for sales. It can also require extremely high volumes to make it work.<br \/>\nFlickr only manages to upsell around 5% of its standard users to its professional<br \/>\naccount. And storing, searching and serving the 3.5 billion images Flickr&#39;s free<br \/>\ncustomers store certainly isn&#39;t cheap.<br \/>\nIn 1993, the UK mobile telephony market was heating up. One2One, a fledgling mobile<br \/>\ntelecommunications company backed by Cable &amp; Wireless and US West, decided that<br \/>\nfree was the way to go and offered free off-peak local calls to all new customers. The<br \/>\nnetwork was soon overwhelmed as thousands of customers tried, and failed, to get<br \/>\nthrough, for free, on Christmas day. One2One quickly gained a reputation for<br \/>\nunreliability, losing nearly a million dollars along the way.\t<br \/>\n36<br \/>\n<br \/>\nFlickr has Yahoo, and One2One had Cable &amp; Wireless, but if you adopt the freemium<br \/>\nmodel without a sugar daddy, then beware.<br \/>\nNetwork Effects<br \/>\nThere is, however, one situation where free is the best price for your product: where<br \/>\nthere are strong network effects.<br \/>\nNetwork effects occur where the value to your customer of using your product<br \/>\nincreases as the total number of users increases. For example, the value of using a<br \/>\ntelephone increases as the number of people you can call increases; the value of a<br \/>\nsocial network increases as more people join, the value of e-mail as more people get<br \/>\naccounts, and so on. In these cases, you get a feedback loop: more people use your<br \/>\napplication, it becomes more valuable and more people join, and so on.<br \/>\nFree becomes even more important when your networked product has competitors. In<br \/>\nthis situation, it turns out there are two stable situations: no customers, or plenty of<br \/>\ncustomers, and that there is a critical point beyond which user numbers accelerate<br \/>\nquickly. Get past the tipping point and your user base will accelerate rapidly. If you<br \/>\ndon&#39;t quite reach the tipping point then your user base will shrink back to zero.<br \/>\nIf you study Twitter&#39;s traffic stats, you can see there&#39;s a clear tipping point at the<br \/>\nbeginning of 2009. Look at the uptake of the fax machine, the telephone and other<br \/>\ninventions that rely heavily on network effects and you&#39;ll see a similar pattern.<br \/>\nIt becomes, therefore, extremely important to reach the tipping point as quickly as<br \/>\npossible, and the &#39;free&#39; price point is a good way of doing that. Of course, once you&#39;re<br \/>\npast the tipping point you&#39;ll need to make money from your product, without losing<br \/>\nusers.<br \/>\nBargains<br \/>\nBargains are closely related to free: people like getting something for nothing. Bundling<br \/>\nis a type of free. When you buy Windows, you get Internet Explorer for free. The SQL<br \/>\nToolbelt gives you 12 applications worth a total of $5140 for only $1595. That&#39;s $3545<br \/>\nfor free.\t<br \/>\n37<br \/>\n<br \/>\nPut a &#39;sale&#39; price on one or two products on your web site, and people will assume that<br \/>\nthey are, in fact, getting a good deal. But put &#39;sale&#39; on all your products and people will<br \/>\nassume you&#39;re taking them for a ride.<br \/>\nTo work best, bargains should be limited to specific products, or specific times. When<br \/>\nSteam, the online gaming community, held a sale on third party games over the<br \/>\nholiday season in 2008, a 10% discount led to an increase of 35% in sales (in dollars,<br \/>\nnot units). A 25% discount led to a 245% increase; a 50% sale to a 320% increase and<br \/>\na 75% discount to a 1,470% increase.<br \/>\nIf there&#39;s something people like more than getting a bargain, it&#39;s getting a bargain and<br \/>\nfeeling smart. Just before Christmas 2006, Threshers (a UK wine merchant) offered its<br \/>\nsuppliers and friends a 40% discount if they turned up at any store with a special<br \/>\nvoucher.<br \/>\nWhen this voucher was \u201caccidentally\u201d leaked onto the web (on the Stormhoek web site<br \/>\n- we&#39;ve already met them - and promoted by Hugh MacLeod), word spread like wildfire.<br \/>\nEventually, millions of people downloaded it. Threshers felt obliged to honor the<br \/>\nvoucher. Their customers felt smart and got cheap wine while Threshers made a killing<br \/>\nand promoted their brand.<br \/>\nApps and App Markets<br \/>\nThe phenomonal success of the iPhone\u2013and the smartphone revolution that has<br \/>\nfollowed in its wake\u2013has brought a brand new way of purchasing software to the table.<br \/>\nInstead of approaching vendors&#39; websites directly, or heading into a bricks-and-mortar<br \/>\nstore, consumers are now searching and downloading applications from an app store.<br \/>\nDeciding to sell your software on an app store isn&#39;t always an easy decision. If you&#39;re<br \/>\nwriting software for iOS, your only choice is to sell through Apple&#39;s App Store. For other<br \/>\nplatforms, there are other options, but all have their own restrictions and fees.<br \/>\nThe most immediate concern is the commission each store will charge you for selling<br \/>\nyour app through their service. Most app stores will take a 30% cut, which is a large<br \/>\nchunk of your revenue. You may need to weigh up the extra audience against this fee,<br \/>\nreduce your marketing budget, and calculate if it&#39;s worth it.\t<br \/>\n38<br \/>\n<br \/>\nThe biggest issue when pricing, however, is the\tlower percieved value\tof mobile\t<br \/>\napps. If we decide to redesign the UI and put Task List 400 into the iPhone App Store,<br \/>\nwe instantly devalue our app. We know Task List 400 iPhone is a great application, and<br \/>\nwe&#39;re selling the Windows and OS X versions for $300 a piece.<br \/>\nBut it&#39;s the customer&#39;s perception of the app that will ultimately dictate the price.<br \/>\nMobile apps are commoditised. The &#39;throwaway&#39; nature of lots of apps, combined with<br \/>\nthe sheer number of applications available for smartphone platforms, means that most<br \/>\napps have very low price tags and it&#39;s certainly a low-price, high-volume game.<br \/>\nPerhaps we&#39;ve spent $10,000 creating Task List 400, and we want to see it make a<br \/>\nminimum of $80,000 (a good return, no?) As of 2012, the average app price is $1.92.<br \/>\nHow many copies of Task List 400 iPhone do you think we can ship at $1.92? 10,000?<br \/>\n20,000? 50,000?<br \/>\nQuantity\t30% Fee\tProfit\t<br \/>\n10,000\t$5,760\t$13,440\t<br \/>\n20,000\t$11,520\t$26,880\t<br \/>\n50,000\t$28,800\t$67,200\t<br \/>\nWe&#39;re still not even close to the $80,000 return we want. And selling 50,000 copies of<br \/>\nanything is\tvery\tdifficult. Let&#39;s raise the price slightly, to $2.49. The higher price will\t<br \/>\nmean less people will want to purchase a copy, so let&#39;s say we can only sell 30,000<br \/>\ncopies (still a terrific number):<br \/>\nQuantity\t30% Fee\tProfit\t<br \/>\n30,000\t$22,410\t$52,290\t<br \/>\n39<br \/>\n<br \/>\nNope, still not there. It doesn&#39;t look like we&#39;re going to be able to make this $80,000<br \/>\nafter all.<br \/>\nThe sad fact is, it&#39;s\tvery difficult\tto shift such large quantities of applications. If you are\t<br \/>\nconfident that your application is unique and provides real value to your customers,<br \/>\nyou could price your application a little higher than the average and measure the<br \/>\nresults.<br \/>\nA January 2012 report from Distomo about running sales on apps reveals some<br \/>\ninteresting information about app prices. On the first day of the sale, the average<br \/>\nrevenue increase by 41% in the iPhone App Store, and after two weeks, 22% in total.<br \/>\nOn the iPad App Store, the day one effect was even greater: up 52%.<br \/>\nIn general, the report finds that a &#39;perceivably large&#39; discount had a much better effect:<br \/>\nan app dropping from $1.49 to $0.99 had a much greater effect than $7.99 to $5.99.<br \/>\nWhen you look at the actual figures, you&#39;ll find it can be difficult to make a lot of money<br \/>\nby selling apps. It really is a quantities game. Spend time working out how many copies<br \/>\nyou can realistically sell and price your app accordingly. The lower your initial<br \/>\nproduction costs, the easier it will be to make money.<br \/>\nA side note: don&#39;t forget that Apple&#39;s motives and your motives do\tnot\talign. Apple\t<br \/>\nmakes money out of a combination of a few blockbuster, headline-grabbing, successful<br \/>\napps and a massive long tail of apps which individually only bring in a few hundred or a<br \/>\nfew thousand dollars but when aggregated are valuable. To\tApple\t. They really don&#39;t\t<br \/>\ncare how\tyour\tparticular app does. It&#39;s a casino, and in the casino it&#39;s the bank who\t<br \/>\nwins.<br \/>\nSoftware As A Service<br \/>\nWe&#39;ve talked a lot about how to price software that is sold in a packaged-and-<br \/>\ndistributed form; buy a copy and download the executable. Things get more complex,<br \/>\nof course, when dealing with multi-user licenses and new \/ different versions, but the<br \/>\nessential form of software stays the same.\t<br \/>\n40<br \/>\n<br \/>\nAn alternative route taken by a lot of companies is to sell their Software As A Service<br \/>\n(SaaS). This changes some of the details of software pricing, but the principles of<br \/>\npsychology and economics are fundamentally similar.<br \/>\nThere are plenty of benefits of this model beyond the obvious one of recurring revenue:\t<br \/>\n\u2022\tPaying lots of small amounts is psychologically easier than paying one<br \/>\nlarge amount. That&#39;s why people buy cars with credit cards and pay it back<br \/>\nat 20% interest, or place the cost of holiday that&#39;s over in a week on top of<br \/>\na mortgage that will last 25 years. Although the total amount paid is larger,<br \/>\nit somehow feels smaller.\t<br \/>\n\u2022\tIf you&#39;re selling to businesses, then your end user will find it easier to<br \/>\njustify a small, regular payment to his boss than a single large, one-off<br \/>\npayment.\t<br \/>\n\u2022\tRecurring payments promote regular usage. Take members of health clubs.<br \/>\nThose who pay a one-off annual fee tend to use the club intensively for a<br \/>\nfew weeks after their hefty payment, but then stop using it. The usage<br \/>\npattern of people who pay quarterly displays a sawtooth pattern, peaking<br \/>\nshortly after payment and then declining until the next payment. People<br \/>\nwho pay monthly show a steadier, higher usage. Importantly, since they<br \/>\nare more regular users, they are also more likely to renew membership and<br \/>\nstay members longer.\t<br \/>\nA lot of the same techniques for pricing still stand: look at your competition&#39;s pricing,<br \/>\nensure your costs are covered and you make a good margin\u2013or a small margin with<br \/>\nhigh volume!\u2013and appeal to your customers&#39; sense of fairness. The biggest difference is<br \/>\nthat packaged-and-distributed software is sold for a one-off fee; SaaS products have a<br \/>\nrecurring charge that reaches into the future.<br \/>\nThere are, however, a couple of implications to SaaS pricing. The first is that it requires<br \/>\nmore capital than selling a one-off software licence. Let&#39;s say you&#39;ve been selling your<br \/>\nTime Tracker 3000 for $200. It costs you $100 to acquire a customer, so you make<br \/>\n$100 profit for every customer you sign up. You&#39;re getting 100 new customers a month,<br \/>\nso you&#39;re making $10,000 a month.\t<br \/>\n41<br \/>\n<br \/>\nYou&#39;ve decided that actually the future of time tracking is as an online service. You&#39;ve<br \/>\ndone your research, and you&#39;re going to charge $20 a month for the service. On<br \/>\naverage, you think a customer will use your service for 24 months, so the life time<br \/>\nvalue of a customer is nearly $500. Sounds good, but what will happen\ttomorrow\t, when\t<br \/>\nyou stop selling the Time Tracker 3000 and start selling the Time Tracker Web Edition?<br \/>\nLet&#39;s say you&#39;re still getting one hundred new customers a month. But they&#39;re only<br \/>\npaying you $20 a month. What&#39;s worse is that your customer acquisition cost is still<br \/>\n$100, so in the first month you&#39;re losing $80 for every customer you sign up. In other<br \/>\nwords, rather than making $10,000 profit like you did last month, you&#39;ve now made an<br \/>\n$8,000 loss. Of course, in the long term the numbers are going to work in your favour,<br \/>\nbut in the short term monthly pricing is going to hurt you.<br \/>\nYour profitability is also going to depend heavily on metrics such as the lifetime value<br \/>\nof a customer and the cost of customer acquisition, so you need to track these closely.<br \/>\nThe first number - how much a customer is worth to you over their lifetime - depends<br \/>\non how likely a customer is to stop using your service in any given month. This means<br \/>\nthat you need to spend more effort in making sure your customers&#39; needs are satisfied<br \/>\nover time. Caricaturing the argument somewhat, if a customer who has paid $200 for<br \/>\nthe Time Tracker 3000 hits a bug, he&#39;s already given you the money, so hey. If he hits<br \/>\nthe bug on the Time Tracker Web Edition after two months then he&#39;s going to cancel<br \/>\nhis contract. Arguably, this is a very good discipline to have, but it does mean that you<br \/>\nneed to price your product at a point where you can afford to support your customers<br \/>\nand improve it based on their feedback.<br \/>\nHow, then, do you go about pricing a SaaS product?<br \/>\nAt the very least you need to charge enough to cover your costs. How much does it<br \/>\ncost to acquire a customer? How long do you think customers will use the system<br \/>\nbefore? Is there a cost to setting up and maintaining a user in your system?<br \/>\nBeyond this, the other factors we&#39;ve already discussed come into play. How do your<br \/>\ncustomers perceive the value of your service? What will they use as reference points?<br \/>\nCan you provide different versions of your service at different price points?<br \/>\nTake the following example from kings (and queens) of Saas, 37signals&#39;s project<br \/>\nmanagement tool, Basecamp:\t<br \/>\n42<br \/>\n<br \/>\nTheir pricing options are simple and elegant. The tiers get progressively more<br \/>\nexpensive as your usage of the application goes up. You&#39;re not caught in the same<br \/>\nsituation we saw Microsoft Vista in earlier; you know exactly what you&#39;re paying for and<br \/>\nwhat the difference is.<br \/>\nFixed Contracts<br \/>\nSome companies tie their users in to 12- or 24-month contracts, akin to phone<br \/>\ncompanies. This can make sense from a business point of view (you get a guaranteed<br \/>\nlump of cash) but will suck for consumers (what if they want to leave?)<br \/>\nA better alternative is to offer discounts to users who subscribe for 12 months. Offer<br \/>\nthem 2 months for free, maybe, or a 10% discount. This is win-win: it encourages users<br \/>\nto pay a wedge of cash in advance, so you get a year&#39;s worth of capital up front, and<br \/>\nthey have the option to pay monthly if they so choose.\t<br \/>\n43<br \/>\n<br \/>\nOffering a yearly plan can also be a benefit to customer retention. Keep track of the<br \/>\naverage retention of your customer - the amount of time they keep their account. You<br \/>\nmay observe a natural attrition where customers eventually stop using your<br \/>\napplication. If you&#39;re charging monthly and realise that your average retention is nine<br \/>\nmonths, it may be worth offering a discounted yearly plan. This will give you three<br \/>\nmonths of extra revenue you wouldn&#39;t have had before.<br \/>\nYearly billing may also work well if you&#39;re selling to corporations with complex approval<br \/>\nprocesses - they simply may not have the systems in place to easily pay for services<br \/>\nmonthly.<br \/>\nAdobe follow an interesting path with the Creative Cloud tools. This offers their Creative<br \/>\nSuite - a bundle of tools that you can buy for a one-off cost of $2599 - for a monthly fee<br \/>\nof $49.99 if you commit to subscribing for 12 months. If you prefer to pay monthly with<br \/>\nno contract then the cost is $74.99. This is interesting on several levels. First of all,<br \/>\nAdobe are assuming that they will increase the number of people who use their tools. I<br \/>\nhad personally purchased Photoshop (a one-off cost of $1,299) but would probably<br \/>\nnever have upgraded to the last version or bought Premiere Pro (for a one-off cost of<br \/>\n$799). I have, however, subscribed to Creative Cloud. I decided to go for the annual<br \/>\ncontract at a monthly fee of $49.99. I figure I can afford that each month (a reference<br \/>\npoint: it&#39;s less than my phone bill), and I can cancel after a year if I want to. Of course, I<br \/>\nwon&#39;t - I&#39;m too lazy and I&#39;ll be too tied into Adobe tools by then. It turns out that I have<br \/>\nactually used Premiere Pro a couple of time as well. Given the up-front costs of<br \/>\nproviding SaaS (see above), this was a gutsy move from Adobe, but in the long term<br \/>\nthey&#39;re going to make more money out of me, they&#39;re probably going to attract more<br \/>\ncustomers, and I&#39;m happy with the service that I&#39;m getting. Everybody wins.<br \/>\nPay As You Go<br \/>\nYour customers may not use your application regularly and be unlikely to pay for for a<br \/>\ntime period they just won&#39;t use. In these scenarios, having a pay as you go price plan is<br \/>\nideal.<br \/>\nA great example of this is email marketing and distribution service Mailchimp.<br \/>\nMailchimp provide a pay as you go pricing tier for their customers that don&#39;t want to<br \/>\ncommit to a monthly or yearly subscription.\t<br \/>\n44<br \/>\n<br \/>\nYou prepay your account with credits, and then use the credits to send out emails. This<br \/>\nis perfect for when you use the service infrequently, and is bound to attract a whole<br \/>\nnew range of customers that otherwise wouldn&#39;t use Mailchimp.<br \/>\nIt&#39;s clearly more expensive, but is a great way of using the system infrequently, or,<br \/>\nindeed trialing it with a view to subscribe monthly. Which leads me perfectly on to\u2026<br \/>\nFree Trials\u2026 Again<br \/>\nMuch like in packaged-and-distributed software, some online services offer trials to<br \/>\ncustomers. This is usually in the form of a free month: 30 days of unrestricted access to<br \/>\nthe service before requiring the user to enter their credit card details.<br \/>\nThis can have a number of benefits:\t<br \/>\n\u2022\tCustomers are actively engaging with your system, so you have a real chance<br \/>\nto show them how brilliant your software is\t<br \/>\n\u2022\tYou can collect their information for marketing later (as long as you provide an<br \/>\nunsubscribe link!)\t<br \/>\nWhether you like it or not, customers are most likely going to compare your application<br \/>\nto your competitors&#39;. If you offer a trial you are stepping ahead of the game and giving<br \/>\nyourself the best chance.<br \/>\nAfter all, if you&#39;re comparing a system you can touch and play with versus some<br \/>\nscreenshots, what would you prefer? As long as your software is good, and you&#39;re<br \/>\ntargeting the right market, you&#39;re bound to have a good conversion rate between trial<br \/>\nand paid account.<br \/>\nDifferent Ways Of Pricing<br \/>\nThe amount you charge for your product isn&#39;t the only decision you have to make. You<br \/>\nalso need to decide how you want to charge. There are plenty of models:<br \/>\nSubscription\t. We&#39;ve already discussed this in the SaaS section.\t<br \/>\n45<br \/>\n<br \/>\nI&#39;ve already covered the\tFreemium\tmodel, where a small number of paying customers\t<br \/>\nsubsidize the majority of freeloaders. The\tGilette\tmodel is a twist on the Freemium\t<br \/>\nmodel. Gillette famously sells their product in two parts: the razor and the blades. The<br \/>\nrazor is cheap, but they make their money on the blades. This strategy is surprisingly<br \/>\ncommon. Adobe follows a similar strategy with Acrobat. It&#39;s free to read documents, but<br \/>\nyou need to pay to create them. Hewlett Packard loses money on its printers, but<br \/>\nmakes it back on the ink. The first Ford Fiestas were sold at a loss, but Ford recovered<br \/>\nthe money on spares and finance. Microsoft and Sony lose money whenever they sell<br \/>\nan X-Box or PlayStation, but make it back on royalties for games.<br \/>\nThere are many ways of\tpricing per user\t. Common schemes include licensing per\t<br \/>\nnamed user, or concurrent user. At Red Gate, we license per user. If you have a team<br \/>\nof ten people, all of whom want to use our software, then you need to buy a ten user<br \/>\nlicence. If you can&#39;t count the total number of users, or if only a few use it at a time,<br \/>\nthen pricing by concurrent user can make sense. This model is often used for server-<br \/>\nbased software, such as databases.<br \/>\nAnother common licensing model is\tper processor\torper processor core\t. The\t<br \/>\nobvious drawback of this model is that processors get faster, and get more cores,<br \/>\nquickly. If, say, you&#39;re selling a bug tracking system that&#39;s tied to the physical power of<br \/>\nyour customers&#39; hardware then Moore&#39;s law dictates that they will get double the<br \/>\nbenefit of your software every two years, without paying you a penny.<br \/>\nThe\tper physical \/ virtual server\tlicensing model has the same drawbacks as the per\t<br \/>\nprocessor model. As more processors are crammed into physical boxes, your<br \/>\ncustomers get exponentially increasing benefit for a fixed cost.<br \/>\nThe\tper usage\tmodel involves charging users based on how often they use your\t<br \/>\nsoftware. This could be per megabyte stored, transaction processed, gigabyte<br \/>\ntransmitted, or many other options. Historically, this has been less common than other<br \/>\nmodels but will become more usual as cloud computing takes off and people expect to<br \/>\npay for computer usage on-demand. One disadvantage of this model is that it can<br \/>\ndiscourage people from buying since it is unclear, up front, how much the user will<br \/>\nneed to pay.<br \/>\nCharging your end user isn&#39;t the only way of pricing software. You can choose to give it<br \/>\naway for free and then make money by, for example, charging for consulting,\t<br \/>\n46<br \/>\n<br \/>\ninstallation and training; or selling advertising. The latter, although a common model<br \/>\nfor web sites, is extremely hard to make work. CPM \u2013 the cost per thousand<br \/>\nimpressions \u2013 can be as low as a dollar. In other words, to generate one thousand<br \/>\ndollars of revenue you might need to serve up as many as a million pages. To generate<br \/>\nenough revenue to support a team of three or four people, that means having ten<br \/>\nmillion page views per month. Most web applications simply aren&#39;t going to attract that<br \/>\nsort of traffic.<br \/>\nGiving your customers a choice of licensing models can make sense. For example, if<br \/>\nyou&#39;re buying Microsoft&#39;s SQL Server 2008 then you can choose to license per<br \/>\nprocessor, or buy a server license and then pay per client who connects. The first<br \/>\nmodel will cost you $5,999 per processor. For the second option, you&#39;ll need to pay<br \/>\n$885 to run it on a single server, and then $162 for each additional user to access the<br \/>\ndatabase.<br \/>\nMany businesses end up with a\tmixed model\t. For example, Red Gate combines a one-\t<br \/>\noff fee with an annual 10% - 25% support and upgrades fee. That way, we get both up-<br \/>\nfront revenue and a recurring yearly income.<br \/>\nHowever, if you choose to do this, you need to be aware of the pitfalls. Support and<br \/>\nupgrades fees aren&#39;t just a cheap way of generating cash, and they can pressure you<br \/>\ninto releasing software just for the sake of it, at times that are not right for you, your<br \/>\ncustomers or your product. If you&#39;re going to charge your customers regularly, then you<br \/>\nneed to make sure they get \u2013 or perceive \u2013 value regularly.<br \/>\nShortly after launching Windows XP in 2001, Microsoft introduced its &#39;Software<br \/>\nAssurance&#39; program. For an annual fee, enterprise customers could get guaranteed<br \/>\nupgrades to next versions of the operating system. In theory, everybody would win:<br \/>\nMicrosoft would get a guaranteed revenue stream to fund future development and<br \/>\ncustomers could spread costs and would get a cheaper upgrade to Microsoft&#39;s new<br \/>\noperating system, code name Longhorn, when it shipped in 2003. But Longhorn didn&#39;t<br \/>\nship in 2003. Or 2004. Or 2005. It didn&#39;t reach the market until the end of the 2006,<br \/>\nlargely neutered, as Windows Vista. And even then most enterprises refused to<br \/>\nupgrade.\t<br \/>\n47<br \/>\n<br \/>\nChoosing the right model<br \/>\nWhen choosing your pricing model, here are two recommendations. Firstly,\tbe boring\t.\t<br \/>\nSecondly, license your software as your customers expect it be licensed \u2013 fit in with<br \/>\ntheir business model.<br \/>\nRed Gate&#39;s first product was Aardvark, an online bug tracking system. When we<br \/>\nlaunched this in early 2000, we decided to follow a usage model. We charged per bug<br \/>\nraised. This made sense from our perspective since the cost of providing the service<br \/>\nwas linked to how much our customers used it, but it didn&#39;t fit in with the way our<br \/>\ncustomers worked or expected to be charged. That was our first mistake. Our second<br \/>\nmistake was to forget to be boring, and to call the usage units &#39;cans of worms&#39;. We<br \/>\nthought it was pretty cool. Our customers had a different opinion, and we quickly<br \/>\nmoved to per-user pricing.<br \/>\nThere are even worse ways of getting price models wrong. In the late 1990s, The<br \/>\nDialog Corporation was formed through the merger of Knight-Ridder and MAID plc. It<br \/>\nwas in the business of selling data to corporations and government bodies. Users<br \/>\nlogged on and searched for information in the six billion pages of information that<br \/>\nDialog stored.<br \/>\nDialog decided to implement a per-usage model. Subscribers bought &#39;DialUnits&#39;, and<br \/>\ndifferent actions cost different amounts of DialUnits, depending on how much resource<br \/>\nthe action took and the value of the data being accessed. Want to sort your results?<br \/>\nThat would cost more than saving them. How much more? It would depend on the type<br \/>\nof database you were searching, and the intensity of your search. Ranking, or removing<br \/>\nduplicate results, was especially resource intensive so cost more DialUnits. Some<br \/>\nactions were free. It took four pages of instructions to explain the pricing model to<br \/>\ncustomers, and that was after a round of simplification.<br \/>\nIn 2001, Dialog then introduced multiple pricing plans and expected users to choose<br \/>\nwhether it would be cheaper to use pricing based on usage, or on time. Then there<br \/>\nwere different platforms - Dialog Transact, Dialog Advantage and Dialog Enterprise.<br \/>\nThrow in discounts, multiyear options and differing interfaces such as Dialog Classic,<br \/>\nDialogWeb and DialogClassic Web and, as one user put it, thinking of a number,<br \/>\ndoubling it and adding your mother&#39;s age would have been a clear, better pricing<br \/>\nstrategy.\t<br \/>\n48<br \/>\n<br \/>\nChapter Five\t<br \/>\nPricing Perception<br \/>\n<br \/>\nPrices are never neutral. They send signals. For example, a high price can signal that<br \/>\nyou have a quality product. Consumers assume that expensive perfumes and wine are<br \/>\nbetter than cheap ones, even in the absence of much evidence.<br \/>\nA low price can tell customers that you&#39;re value for money, or that you&#39;re special. If<br \/>\nyour competitors are selling software at $10,000 a seat, and you&#39;re selling yours at<br \/>\n$100, then that says something about you. Of course, you might be saying &#39;game<br \/>\nchanging&#39;, but your customers might be hearing &#39;toy&#39;.<br \/>\nCopy your competitors and you could be indicating that you&#39;re just a &#39;me too&#39; product.<br \/>\nIf you&#39;re a me-too product, with me-too features and a me-too price, why would people<br \/>\nbuy from you, especially if there&#39;s already a strong, dominant product in your market?<br \/>\nWhatever price you choose, the signals it sends need to fit in with your brand, and your<br \/>\nbrand needs to fit in with your reality. There&#39;s no point using a high price to signal that<br \/>\nyou have a quality product if you&#39;re not willing to spend marketing dollars sustaining<br \/>\nthat brand, development dollars making that quality a reality and customer service<br \/>\ndollars providing the level of service people expect from a quality brand.<br \/>\nIn 1996, McDonalds launched the Arch Deluxe in an attempt to create a burger for a<br \/>\nmore sophisticated, adult consumer. To recoup the extra cost of the higher quality<br \/>\ningredients and the $200 million dollar marketing campaign, McDonalds priced the new<br \/>\nsandwich 32 cents higher than a Big Mac. But the product they tried to create (high<br \/>\nquality, premium) conflicted with the McDonalds brand (cheap and convenient) and the<br \/>\nArch Deluxe flopped. One argument could be that they priced the burger too low, and<br \/>\nthat a 32 cent premium did not send enough of a quality signal.<br \/>\nYour business model and your strategy have to support your pricing model. If you have<br \/>\nexpensive sales people driving expensive cars, taking your customers&#39; CEOs out to golf,<br \/>\nand end users who expect plenty of hand-holding and customization of the software<br \/>\nyou sell them, then you can&#39;t sustain a low price point. Similarly, if you&#39;re selling shrink-<br \/>\nwrapped, mass-market software over the web then a high price point will be counter-<br \/>\nproductive.<br \/>\nWhen Red Gate tried to get into the automated web load testing market one of the<br \/>\nreasons we failed (there were plenty of others, including a product that wasn&#39;t up to<br \/>\nscratch) was that we attempted a low-price, high-volume approach in a market\t<br \/>\n50<br \/>\n<br \/>\ndominated by high-price solutions. We figured that consumers would love a product<br \/>\nthat they could just download, try and then buy, but it turned out that our customers<br \/>\nwanted much more handholding than we were able to provide. For the most part, they<br \/>\ndidn&#39;t want a product, they wanted a people-intensive service and the reassurance that<br \/>\na big-name, expensive vendor could provide. Our load testing tool was moderately<br \/>\nsuccessful, but it achieved nothing like the success we had dreamed of.<br \/>\nThe dotcom boom and bust contains plenty of illustrations of companies who failed to<br \/>\nalign their pricing with their business model. For example, Kozmo.com&#39;s business was<br \/>\nbuilt around delivering snacks, DVD rentals and Starbucks coffee, within an hour, to<br \/>\ncity dwellers. Unfortunately, the low price, high volume business model it chose<br \/>\nclashed with the reality of the expense of delivering small items by bike courier. This<br \/>\ncould have worked as a high price, low volume business model, as the butlers of<br \/>\nEnglish aristocrats will testify.<br \/>\nSwitching strategies can be hard. For example, when Intel introduced the 8080<br \/>\nprocessor, it priced it at $340. Ultimately, it was selling for $2 a unit, but Intel found it<br \/>\nvery hard to shake the initial imprinting of the high cost in people&#39;s minds.<br \/>\nPractice Trumps Theory<br \/>\nYou&#39;ve read a lot of theory here. Wherever I can, I&#39;ve based it on my experience and<br \/>\nsound research (you can find some of my sources in the bibliography later on). But your<br \/>\nown circumstances are different to any of those described here, so never forget that<br \/>\npractice trumps theory\t.\t<br \/>\nProduct pricing is as much art and craft as it is science. Sure, it helps to understand the<br \/>\neconomics and psychology of pricing, but theory can only tell you so much. At some<br \/>\npoint, you need to make a decision and do it. Use the information in this handbook to<br \/>\nmake an informed stab at what a good price would look like, and how your customers<br \/>\nwill react, and try it out. The exact price almost doesn&#39;t matter \u2013 get it broadly right,<br \/>\ndon&#39;t screw up totally \u2013 and you can tweak it later.<br \/>\nYou&#39;re never going to know if you&#39;ve chosen the exact right price or not, but you should<br \/>\nexperiment once you&#39;ve set your initial price. Not experiment in the scientific sense of<br \/>\nforming a hypothesis, changing a single variable, and accepting or rejecting the<br \/>\nhypothesis, but in the sense of changing something and seeing what happens.\t<br \/>\n51<br \/>\n<br \/>\nScientific experiments are simply too hard to do, and the results too ambiguous, to be<br \/>\nmuch use in pricing. Too many variables change. When you change your prices, you&#39;ll<br \/>\nprobably do it when you release a new version of your product, or it will coincide with a<br \/>\nbig marketing push, or some other variable you cannot control, such as the state of the<br \/>\neconomy or the reaction of a competitor, will interfere.<br \/>\nAlthough scientifically purer, it often doesn&#39;t make sense to change a single variable at<br \/>\na time. Theoretically, you shouldn&#39;t change the price of your product, your discounting<br \/>\nstrategy and the types of bundle that you sell, all at the same time. But practically, it<br \/>\ncan be the right thing to do. It&#39;s more useful to fix the problem than to understand why<br \/>\nit&#39;s broken. When a scientist goes on a blind date that doesn&#39;t work out then, in theory,<br \/>\nhe should fix one variable at a time, and re-run the date. First, he should change the<br \/>\npartner but go to the same film and buy the same flowers. Next, he should keep the<br \/>\npartner the same, vary the film and keep the flowers the same, and so forth. But the<br \/>\npragmatist in him will, or should, change the girl, the film, the flowers, and buy some<br \/>\nnew clothes and shave too. If it works, he might not understand why, but at least he&#39;ll<br \/>\nhave a girlfriend.<br \/>\nIn the old days, experiments were easy to run. You&#39;d A\/B test, splitting your customers<br \/>\ninto random groups, post each group a different leaflet with a different price and<br \/>\nmeasure the outcome. Nowadays, this is risky. The Internet makes it easy for people to<br \/>\nfigure out what other people are paying.<br \/>\nYou might be tempted to first run a\tsurvey\t, testing how customers might react to a\t<br \/>\nproposed new pricing model, or change to an existing one. However, surveys rarely<br \/>\nwork. There is always a disconnect between customers&#39; words and their actions. When<br \/>\nMcDonalds launched its Arch Deluxe burger (see above), consumers in focus groups<br \/>\nloved it, and 80% said they&#39;d buy it. Few of them did.<br \/>\nHow To Change Your Pricing<br \/>\nYou might be worried about how your customers will react when you change your<br \/>\nprices. Don&#39;t be. For most of us, our customers have better things to worry about. If we<br \/>\nshift our prices from $100 to $150 then most people won&#39;t notice, and of those who do<br \/>\nnotice very few will care. If you bought a copy of SQL Compare from the Red Gate web<br \/>\nsite in 2000 it would have cost you $50. Do the same thing now, and you&#39;ll find the<br \/>\nprice is $395. Buy the full suite of tools and expect to pay $1,595.\t<br \/>\n52<br \/>\n<br \/>\nOf course, at Red Gate we&#39;ve reached that price over the course of almost a decade.<br \/>\nWe&#39;ve spent millions of dollars developing the software, and tens of man years. The<br \/>\nincrease in value that our customers get from our software vastly outweighs the<br \/>\nincrease in its cost. But, of the hundreds of thousands of customers we have, only a<br \/>\nhandful have ever commented when the price went up.<br \/>\nIt&#39;s not what your customers say that&#39;s important, it&#39;s how they behave. Whenever you<br \/>\nmake a price change, pay close attention to what your customers do. If they stop<br \/>\nbuying, rethink.<br \/>\nIf, however, you are running a subscription model then you should consider what you<br \/>\nwant to do with your current customers. If they signed up at $10 a month, and you&#39;ve<br \/>\ndecided to increase the price to $20 a month, think hard about how you want to treat<br \/>\nthem.\t<br \/>\n53<br \/>\n<br \/>\nChapter Six\t<br \/>\nProduct Pricing Checklist<br \/>\n<br \/>\nWrapping up, here&#39;s a checklist to help you decide your pricing:<br \/>\nWhat\u2019s your strategy?<br \/>\nAre you going to price low and sell lots, or price high and sell a few? How does this fit<br \/>\ninto your brand, the product you have and the image you want to project?<br \/>\nWhat\u2019s your product?<br \/>\nDon&#39;t forget that it&#39;s not just the software that you&#39;re selling. It&#39;s the entire package<br \/>\naround it.<br \/>\nHow will your customers judge the fairness of your pricing?<br \/>\nWhat reference points will they use? How will they determine what seems right? Will<br \/>\nthey baulk at the price you choose, or will they accept it?<br \/>\nWho are your customers?<br \/>\nHow does their business work, and how do they expect to be charged? How much<br \/>\nmoney do they have? Do they prefer a one-off fee, or a monthly subscription? Get<br \/>\nunder their skin.<br \/>\nWho are your competitors?<br \/>\nHow will they react to your pricing? How much more, or less, valuable is your product<br \/>\nthan theirs? What is their business model? What are their prices? If you undercut them,<br \/>\nwill you trigger a price war? If you do, are your pockets deep enough for you to win it?<br \/>\nDo you want to co-exist with your competitors, or destroy them?<br \/>\nHow are you going to sell your software?<br \/>\nDo you need to send out sales people to take customers golfing? Or are you planning<br \/>\nlow-touch sales over the internet? Will you require a telesales team? How much will\t<br \/>\n55<br \/>\n<br \/>\neach sale cost you? Do you need to sell via a channel or reseller? What cut will they<br \/>\ntake?<br \/>\nCan you segment your customers, and create versions?<br \/>\nIs your software worth different amounts to different people, and can you create pricing<br \/>\nthat reflects that? Students and business people for example, or normal and power<br \/>\nusers, or maybe you can split by geography or taste.<br \/>\nHow can you bundle your software?<br \/>\nCan you create a larger package that contains more than one software product?<br \/>\nMake an informed guess at your price<br \/>\nDespite all the psychology and economics, you ultimately just have to pick a price.<br \/>\nSome price \u2013 any price \u2013 is better than no price.<br \/>\nTry it out<br \/>\nPractice trumps theory. Try out your pricing and see what happens. If you&#39;ve got your<br \/>\npricing broadly right - and if you&#39;ve got this far you should do - then you can tweak it<br \/>\nlater.\t<br \/>\n56<br \/>\n<br \/>\nSummary &amp; Bibliography<br \/>\n<br \/>\nIn this usefully short guide to software pricing, we&#39;ve examined a bunch of useful<br \/>\ntechniques to better pricing your software. I hope you&#39;ve got something out of the past<br \/>\n60 pages. If you have, then please forward the eBook onto your friends and colleagues,<br \/>\ntweet about it and share it around your network.<br \/>\nIf you spot any errors in this book, it would be amazing if you could submit it to the<br \/>\nerrata page at https:\/\/efendibooks.com\/minibooks\/dont-just-roll-the-dice\/errata\/. If you<br \/>\nthink I&#39;ve left something out, or got something wrong, then drop me an e-mail at<br \/>\nneil.davidson@red-gate.com\t, or get in touch with me via Twitter,\t@neildavidson\t.\t<br \/>\nBibliography \/ Further Reading<br \/>\nAnderson, E . and Simester, D. (2003) &#39;Minding your pricing cues&#39;, Harvard Business<br \/>\nReview, September 2003<br \/>\nBreckon, N. (2009) &#39;Valve: Left 4 dead half-price sale saw 3000% increase, beat launch<br \/>\nnumbers&#39; http:\/\/www.shacknews.com\/onearticle.x\/57308<br \/>\nChapman, R. (2006) &#39;In Search of Stupidity&#39; ,2nd ed, Apress, Berkeley<br \/>\nCrampes, C . and Laffont, J-J. (2002) &#39;Copying and software pricing&#39;<br \/>\nCusumano, M. (2007) &#39;The changing labyrinth of software pricing&#39;, Communications of<br \/>\nthe ACM, Vol. 50, Issue 7, pp. 19-22<br \/>\nDavidow, W. (1986) &#39;Marketing high technology - an insider&#39;s view&#39;, The Free Press,<br \/>\nNew York<br \/>\nDistimo (2012) &#39;Distimo Report&#39;<br \/>\nGallaugher, J. and Wallace, E. (2002) &#39;Understanding network effects in software<br \/>\nmarkets: evidence from web server pricing&#39;, MIS Quarterly, Vol. 26, No. 4, pp. 303-327<br \/>\nGilbert, A. (2003) &#39;CRM software or CRM shelfware?&#39; http:\/\/news.cnet.com\/CRM-<br \/>\nsoftware-or-CRM-shelfware\/2100-1012_3-990880.html\t<br \/>\n58<br \/>\n<br \/>\nGourville, J. and Soman, D. (2002) &#39;Pricing and the psychology of consumption&#39;, Harvard<br \/>\nBusiness Review, September 2002<br \/>\nGourville, J. (2006) &#39;Eager sellers and stony buyers&#39;, Harvard Business Review, June<br \/>\n2006<br \/>\nGourville, J. and Soman, D. (2005) &#39;Overchoice and assortment type: when and why<br \/>\nvariety backfires&#39;, Marketing Science, Vol. 24, No. 3, pp. 382-395<br \/>\nGourville, J. and Soman, D. (2001) &#39;Transaction decoupling: how price bundling affects<br \/>\nthe decision to consume&#39;, Journal of Marketing Research, Vol. 38, pp. 30-44<br \/>\nGourville, J. and Soman, D. (2007) &#39;Extremeness seeking: when and why consumers<br \/>\nprefer the extremes&#39;, Harvard Business Review<br \/>\nHarford, T. (2008) &#39;Business life: Fair trade or foul&#39; http:\/\/timharford.com\/2008\/04\/<br \/>\nbusiness-life-fair-trade-or-foul\/<br \/>\nKnopf, J. (2000) &#39;The Origin of Shareware&#39; - http:\/\/www.asp-shareware.org\/users\/history-<br \/>\nof-shareware.asp<br \/>\nLevitt, T. (1980) &#39;Marketing success through differentiation - of anything&#39;, Harvard<br \/>\nBusiness Review, January - February 1980<br \/>\nMacrovision (2007) &#39;Key trends in software pricing and licensing&#39;<br \/>\nMason, M. (2008) &#39;The pirate&#39;s dilemma&#39;, The Free Press, New York<br \/>\nMiller, P. (2006) &#39;Sony losing mad loot on each PS3&#39; http:\/\/www.engadget.com\/2006\/11\/<br \/>\n16\/sony-losing-mad-loot-on-each-ps3\/<br \/>\nMurph, D. (2006) &#39;Wii Manufacturing Costs ring up to just $158?&#39;<br \/>\nhttp:\/\/www.engadget.com\/2006\/12\/15\/wii-manufacturing-costs-ring-up-to-just-158\/<br \/>\nPackard, D. (1996) &#39;The HP Way&#39;, HarperCollins, New York\t<br \/>\n59<br \/>\n<br \/>\nQuint, B. (2001) &#39;Dialog rolls out new connect-time pricing&#39; http:\/\/www.allbusiness.com\/<br \/>\nsales\/1012692-1.html<br \/>\nSink, E. (2004) &#39;Product Pricing Primer&#39;, http:\/\/www.ericsink.com\/bos\/<br \/>\nProduct_Pricing.html<br \/>\nSpolsky, J. (2004) &#39;Camels and Rubber Duckies&#39; http:\/\/joelonsoftware.com\/articles\/<br \/>\nCamelsandRubberDuckies.html<br \/>\nSpolsky, J. (2006) &#39;Simplicity&#39; - http:\/\/www.joelonsoftware.com\/items\/2006\/12\/09.html<br \/>\nStiff, D. (2007) &#39;How DeWALT turned customers into influencers&#39;<br \/>\nhttp:\/\/credibilitybranding.typepad.com\/blog\/2007\/03\/how_dewalt_turn.html<br \/>\nSutton, J. (2001) &#39;Technology and market structure&#39;, 2nd ed, The MIT Press<br \/>\nVarian, H. (2003) &#39;Intermediate Microeconomics&#39;, 6th ed, W.M. Norton, New York<br \/>\nWendt O., von Westarp, F. And K\u00c3\u00b6nig, W. (2000) &#39;Pricing in Network Effect Markets&#39;,<br \/>\nECIS Proceedings, Paper 82<br \/>\nWayne, B. (2009) &#39;YouTube is doomed (GOOG)&#39; http:\/\/www.businessinsider.com\/is-<br \/>\nyoutube-doomed-2009-4<br \/>\nWikipedia \u2013 &#39;3DO Interactive Multiplayer&#39; - http:\/\/en.wikipedia.org\/wiki\/<br \/>\n3DO_Interactive_Multiplayer\t<br \/>\n60<br \/>\n<br \/>\nefef\tendiendi\t<br \/>\nbookbook\tss\t<br \/>\nSmart, succinct books for web<br \/>\ndevelopers<br \/>\nhttps:\/\/efendibooks.com","created_at":"2020-01-12T21:49:06+0000","updated_at":"2020-01-12T21:49:06+0000","published_at":"2012-10-02T10:03:06+0000","published_by":["Neil Davidson"],"annotations":[],"mimetype":"application\/pdf","reading_time":91,"domain_name":"neildavidson.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1102,"title":"A Simple Guide to Writing & Publishing Emacs Packages","url":"https:\/\/spin.atomicobject.com\/2016\/05\/27\/write-emacs-package\/","content":"<p>I <a href=\"https:\/\/spin.atomicobject.com\/2016\/01\/26\/webstorm-nodejs-development\/\">wrote recently<\/a> about using <a href=\"https:\/\/www.gnu.org\/software\/emacs\">Emacs<\/a> as a JavaScript development environment. One of my chief complaints was the inability to easily run JavaScript tests from within Emacs. I practice TDD frequently, and having to context-switch out of the editor I\u2019m using to run tests is a big annoyance for me.<\/p>\n<p>I knew it was possible to do what I wanted from within Emacs, as evidenced by other test runner modes like <a href=\"https:\/\/github.com\/pezra\/rspec-mode\">RSpec-mode<\/a>. Armed with that knowledge, I decided to go through the process of learning enough <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\">Emacs Lisp<\/a> to make a <a href=\"https:\/\/mochajs.org\/\">Mocha<\/a> test runner. In the process, I learned a lot about developing Emacs packages and ended up with a really useful tool, so I thought I would share some of the things I learned.<br \/><\/p>\n<p>There is a lot of content here, and we are going to cover three main topics:\u00a0using Emacs as a Lisp IDE, writing a simple package, and publishing that package for others to use.<\/p>\n<h2>Emacs as an Emacs Lisp IDE<\/h2>\n<p>Unsuprisingly, Emacs itself is an excellent development environment for Emacs Lisp code. It can be easily cofigured to include IDE features for Lisp development, such as autocompletion, popup documentation, integrated debugging, and a REPL.<\/p>\n<h3>A few recommendations<\/h3>\n<p>Most of these features are built in, although I highly recommend installing the third-party packages <a href=\"http:\/\/company-mode.github.io\/\">company-mode<\/a> (autocompletion) and <a href=\"http:\/\/www.flycheck.org\/en\/latest\/\">Flycheck<\/a> (real-time syntax checking) if you\u2019re going to do Emacs Lisp development.<\/p>\n<p>I also recommend turning on the built-in <code>eldoc-mode,<\/code> which will pop up documentation and signatures for various functions and symbols as you write code.<\/p>\n<p>Lastly, I recommend familiarizing yourself with the built-in debugging and evaluation functions for Emacs Lisp. For evaluating code to test it, you can use the built-in <code>Lisp-interaction-mode<\/code>, which the <code>*scratch*<\/code> buffer usually has enabled by default. With the mode, you can paste Emacs Lisp code and then press <code>C-x C-e<\/code> to evaluate the code and see the results.<\/p>\n<p>Emacs also comes with <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Edebug.html\">Edebug<\/a>, a built-in stepping debugger for Emacs Lisp code. There are several ways to use it, but I most commonly use the interactive function <code>edebug-defun.<\/code>\u00a0When run inside the body of a function, it sets a breakpoint at the start of that function that will be hit the next time you run it.<\/p>\n<h2>Making a Custom Compilation Mode<\/h2>\n<p>Mocha is a CLI tool, and Emacs has a number of built-in utilities for running external CLI programs.<\/p>\n<h3>Compilation buffer<\/h3>\n<p>The most relevant one for something like a test runner is a compilation buffer. In Emacs, this\u00a0runs an external CLI process and displays the output in a buffer. This is useful for programs where you care about the output, like a compiler or test runner. It also includes some built-in niceties like the ability to highlight errors and jump to them.<\/p>\n<p>In fact, you don\u2019t even need to write any code to run an external command in a compilation buffer. You can just use the <code>M-x compile<\/code> command like so:<\/p>\n<p><a href=\"http:\/\/g.recordit.co\/wWAwkTuVH2.gif\"><img src=\"http:\/\/g.recordit.co\/wWAwkTuVH2.gif\" alt=\"Running a compilation command\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>This is a solid approach for a static compilation command like the default <code>make -k<\/code>. However, it doesn\u2019t scale well to something like a test runner, which needs to do the following:<\/p>\n<ol><li>Run a local script, requiring a consistent working directory or an absolute path (<code>M-x compile<\/code> will use the directory of the current file as the working directory).<\/li>\n<li>Pass dynamic configuration options like the file to test the runner.<\/li>\n<\/ol><h3>Custom compilation mode<\/h3>\n<p>The solution in Emacs is to programmatically create a custom compilation mode that can take these options and run using an interactive function. This is easy to do. In fact, the compilation mode for Mocha.el is only a couple of lines:<\/p>\n<pre><code class=\"language-Lisp\">\n(require 'compile)\n...\n(defvar node-error-regexp-alist\n  `((,node-error-regexp 1 2 3)))\n(defun mocha-compilation-filter ()\n  \"Filter function for compilation output.\"\n  (ansi-color-apply-on-region compilation-filter-start (point-max)))\n(define-compilation-mode mocha-compilation-mode \"Mocha\"\n  \"Mocha compilation mode.\"\n  (progn\n    (set (make-local-variable 'compilation-error-regexp-alist) node-error-regexp-alist)\n    (add-hook 'compilation-filter-hook 'mocha-compilation-filter nil t)\n  ))\n<\/code><\/pre>\n<p>While some of the syntax is a little cryptic (thanks, Lisp!), what it does\u00a0is very simple. We use the built-in <code>define-compilation-mode<\/code> macro to define a compilation mode named mocha-compilation-mode, and we do two things with it:<\/p>\n<ol><li>Pass it a regular expression that maps <a href=\"https:\/\/nodejs.org\/en\">Node.js<\/a> error output to filenames, line numbers, and column numbers.<\/li>\n<li>Add a processing hook which interprets ANSI escape codes and formats them properly.<\/li>\n<\/ol><p>The first enables us to quickly jump to the point of failure in a test. The second makes everything look nicer.<\/p>\n<h2>Running Test Commands<\/h2>\n<p>Now that we have a custom compilation mode that will nicely display our command output, we need to generate a test command and run it with the custom mode. Doing this will involve several simple steps.<\/p>\n<h3>Find project root<\/h3>\n<p>Many types of command line utilities need to be run from the project root. Fortunately, project roots are\u00a0generally easily identified by the presence of a particular file or directory (like a source control directory). Since this is such a common need, Emacs has a built-in function, <code>locate-dominating-file<\/code>, to recursively search up a directory tree for a particular file name. The Emacs documentation on this function explains how to use it better than I could:<\/p>\n<blockquote>\n<p><code>(locate-dominating-file FILE NAME)<\/code><br \/>\nLook up the directory hierarchy from FILE for a directory containing NAME. Stop at the first parent directory containing a file NAME, and return the directory. Return nil if not found. Instead of a string, NAME can also be a predicate taking one argument (a directory) and returning a non-nil value if that directory is the one for which we\u2019re looking.<\/p>\n<\/blockquote>\n<h3>Customize configuration<\/h3>\n<p>Unlike an actual compilation, which would involve rerunning a single static command, something like a test runner needs to be dynamically configurable. Fortunately, Emacs has Customize, an awesome built-in and extensible configuration interface for packages (and the core editor). Customize exposes several macros which can be used to define custom configuration parameters for a package and display them in an editable GUI.<\/p>\n<p>For example, here are the configurations we expose for our Mocha runner:<\/p>\n<pre><code class=\"language-Lisp\">\n(defgroup mocha nil\n  \"Tools for running mocha tests.\"\n  :group 'tools)\n(defcustom mocha-which-node \"node\"\n  \"The path to the node executable to run.\"\n  :type 'string\n  :group 'mocha)\n(defcustom mocha-command \"mocha\"\n  \"The path to the mocha command to run.\"\n  :type 'string\n  :group 'mocha)\n(defcustom mocha-environment-variables nil\n  \"Environment variables to run mocha with.\"\n  :type 'string\n  :group 'mocha)\n(defcustom mocha-options \"--recursive --reporter dot\"\n  \"Command line options to pass to mocha.\"\n  :type 'string\n  :group 'mocha)\n(defcustom mocha-debug-port \"5858\"\n  \"The port number to debug mocha tests at.\"\n  :type 'string\n  :group 'mocha)\n<\/code><\/pre>\n<p>And those show up in the customize GUI like so:<\/p>\n<p><a href=\"https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-1024x393.png\"><img data-attachment-id=\"125271\" data-permalink=\"https:\/\/spin.atomicobject.com\/2016\/05\/27\/write-emacs-package\/screen-shot-2016-05-14-at-7-04-02-pm\/\" data-orig-file=\"https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM.png\" data-orig-size=\"3834,1472\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screen Shot 2016-05-14 at 7.04.02 PM\" data-image-description=\"\" data-medium-file=\"https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-590x227.png\" data-large-file=\"https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-1024x393.png\" class=\"alignnone size-large wp-image-125271\" src=\"https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-1024x393.png\" alt=\"GUI interface for configuring our package\" width=\"640\" height=\"246\" srcset=\"https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-1024x393.png 1024w, https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-150x58.png 150w, https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-590x227.png 590w, https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-768x295.png 768w, https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-600x230.png 600w, https:\/\/spin.atomicobject.com\/wp-content\/uploads\/20160514211809\/Screen-Shot-2016-05-14-at-7.04.02-PM-1200x461.png 1200w\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>Since many of these options make sense to configure on a per-project rather than global basis, Emacs also supports a special file called <code>.dir-locals.el<\/code>, which can override these settings on a per-directory basis. A typical <code>.dir-locals.el<\/code> file might look like this:<\/p>\n<pre><code class=\"language-Lisp\">\n((nil . (\n            (mocha-which-node . \"\/Users\/ajs\/.nvm\/versions\/node\/v4.2.2\/bin\/node\")\n            (mocha-command . \"node_modules\/.bin\/mocha\")\n            (mocha-environment-variables . \"NODE_ENV=test\")\n            (mocha-options . \"--recursive --reporter dot -t 5000\")\n            (mocha-project-test-directory . \"test\")\n            )))\n<\/code><\/pre>\n<p>The syntax is a little cryptic, but if your Emacs working directory is in the same directory as this file or below it, it will respect these options in favor of any global configuration.<\/p>\n<p>Once we have these configuration options defined, it is easy to write a function that will concatenate all the strings together to create our test runner command!<\/p>\n<pre><code class=\"language-Lisp\">\n(defun mocha-generate-command (debug &amp;optional mocha-file test)\n  \"The test command to run.\nIf DEBUG is true, then make this a debug command.\nIf MOCHA-FILE is specified run just that file otherwise run\nMOCHA-PROJECT-TEST-DIRECTORY.\nIF TEST is specified run mocha with a grep for just that test.\"\n  (let ((path (or mocha-file mocha-project-test-directory))\n        (target (if test (concat \"--grep \\\"\" test \"\\\" \") \"\"))\n        (node-command (concat mocha-which-node (if debug (concat \" --debug=\" mocha-debug-port) \"\")))\n        (options (concat mocha-options (if debug \" -t 21600000\"))))\n    (concat mocha-environment-variables \" \"\n            node-command \" \"\n            mocha-command \" \"\n            options \" \"\n            target\n            path)))\n<\/code><\/pre>\n<h2>Generating and Running Compile Command<\/h2>\n<p>Now that we can configure our test command and find the root of our project, we are ready to run it with the custom compilation mode we made earlier. I\u2019m going to show you the most important code for doing that below, and then break it down and explain the different parts.<\/p>\n<pre><code class=\"language-Lisp\">\n(defun mocha-run (&amp;optional mocha-file test)\n  \"Run mocha in a compilation buffer.\nIf MOCHA-FILE is specified run just that file otherwise run\nMOCHA-PROJECT-TEST-DIRECTORY.\nIF TEST is specified run mocha with a grep for just that test.\"\n  (save-some-buffers (not compilation-ask-about-save)\n                     (when (boundp 'compilation-save-buffers-predicate)\n                       compilation-save-buffers-predicate))\n(when (get-buffer \"*mocha tests*\")\n    (kill-buffer \"*mocha tests*\"))\n  (let ((test-command-to-run (mocha-generate-command nil mocha-file test)) (root-dir (mocha-find-project-root)))\n    (with-current-buffer (get-buffer-create \"*mocha tests*\")\n      (setq default-directory root-dir)\n      (compilation-start test-command-to-run 'mocha-compilation-mode (lambda (m) (buffer-name))))))\n<\/code><\/pre>\n<p>Whew! That is some pretty dense code, so let\u2019s break it down bit by bit.<\/p>\n<h3>Check for unsaved buffers<\/h3>\n<p>The first thing this function does is check if there are any unsaved buffers open, and then prompt the user to save them. Sounds pretty complex, but since this is such a common operation, Emacs makes it possible with just a couple of lines.<\/p>\n<pre><code class=\"language-Lisp\">\n  (save-some-buffers (not compilation-ask-about-save)\n                     (when (boundp 'compilation-save-buffers-predicate)\n                       compilation-save-buffers-predicate))\n<\/code><\/pre>\n<h3>Clean up test buffer<\/h3>\n<p>Next, we search for the named buffer we use to run tests to see if it is still around from a previous test run. If it is, we kill it so we can get a fresh start.<\/p>\n<pre><code class=\"language-Lisp\">\n  (when (get-buffer \"*mocha tests*\")\n    (kill-buffer \"*mocha tests*\"))\n<\/code><\/pre>\n<h3>Bind values<\/h3>\n<p>After that, the real work begins. We start by binding two values: the actual test command we are going to run and the path to the project root directory. Both values are calculated using the techniques and code we defined above.<\/p>\n<pre><code class=\"language-Lisp\">\n  (let ((test-command-to-run (mocha-generate-command nil mocha-file test)) (root-dir (mocha-find-project-root)))\n<\/code><\/pre>\n<h3>Run test command<\/h3>\n<p>Finally, now that we have those two values, we actually run our test command. This is a three-step process of:<\/p>\n<ol><li>Creating and switching to the buffer our tests will run in.<\/li>\n<li>Changing the working directory to our project root.<\/li>\n<li>Running our test command in the buffer with our custom compilation mode.<\/li>\n<\/ol><p>All of this is done with the last three\u00a0lines of code:<\/p>\n<pre><code class=\"language-Lisp\">\n    (with-current-buffer (get-buffer-create \"*mocha tests*\")\n      (setq default-directory root-dir)\n      (compilation-start test-command-to-run 'mocha-compilation-mode (lambda (m) (buffer-name))))))\n<\/code><\/pre>\n<h3>Expose interface\u00a0to users<\/h3>\n<p>Now that we have the code to run our test commands, we need to expose it to users. For explicit actions like running commands, Emacs uses interactive functions, which can be called interactively by a user via either the <code>M-x<\/code> interface or a hotkey.<\/p>\n<p>To make a function interactive, you just include the <code>(interactive)<\/code> special form at the top of the function body like so:<\/p>\n<pre><code class=\"language-lisp\">\n;;;###autoload\n(defun mocha-test-file ()\n  \"Test the current file.\"\n  (interactive)\n  (mocha-run (buffer-file-name)))\n<\/code><\/pre>\n<p>If you are not exporting the function as part of a mode, it is also customary to add the <code>;;;###autoload<\/code> magic comment before the function, which helps other Emacs files referencing your package find the function so it\u00a0can be used (for example, to bind them to a hotkey).<\/p>\n<p>Once a function is defined as interactive, it will appear in the <code>M-x<\/code> interface and can be activated by a user.<\/p>\n<p><a href=\"https:\/\/raw.githubusercontent.com\/scottaj\/mocha.el\/master\/mocha.png\"><img src=\"https:\/\/raw.githubusercontent.com\/scottaj\/mocha.el\/master\/mocha.png\" alt=\"Interactive functions for our mode\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>And there you have it. With only a couple of functions and big dose of Emacs magic, we have created a highly configurable test runner that is integrated into our development environment.<\/p>\n<h2>Distributing on MELPA<\/h2>\n<p>Having done all the work to create a custom package, don\u2019t you just want to share it with the world? Fortunately for you, Emacs has a built-in package manager that makes this pretty easy. The package manager is backed by several different repositories, so making your package publicly available is just a matter of getting it\u00a0into one of these repositories.<\/p>\n<p>The three main package repositories are <a href=\"http:\/\/elpa.gnu.org\/\">ELPA<\/a>, <a href=\"https:\/\/marmalade-repo.org\/\">Marmalade<\/a>, and <a href=\"http:\/\/melpa.org\/\">MELPA<\/a>. ELPA is the offical GNU repository that comes with Emacs, while Marmalade and MELPA are third-party repositories. There are a number of differences between each of the repositories, the most significant being how they deal with licensing.<\/p>\n<p>ELPA and Marmalade both require that all packages are GPL- or GPL-compliant licensed. Additionally, ELPA requires you to complete an\u00a0<a href=\"http:\/\/www.gnu.org\/licenses\/why-assign.en.html\">FSF copyright assignment form<\/a>. MELPA, on the other hand, has no licensing requirements, although it does\u00a0have a code review process that all newly added packages must go through to ensure the code\u00a0is of suitable quality.<\/p>\n<p>Which package repositories you choose to put your code on is up to you, but I personally use MELPA and will talk about the process of getting a package into that repository.<\/p>\n<p>There are two basic steps to getting a project on to MELPA.<\/p>\n<h3>Format the package file<\/h3>\n<p>First, you need to follow standard Emacs Lisp conventions for formatting a package file, which includes adding a description header and several other sections to the file. The <a href=\"http:\/\/www.flycheck.org\/en\/latest\/\">Flycheck<\/a> package for Emacs is invaluable here, because it will mark all of the required sections that are missing as errors and guide you through adding them. Doing this correctly is important because the Emacs package manager actually parses these sections as metadata to use.<\/p>\n<h3>Add your recipe<\/h3>\n<p>Once your code is properly formatted, all you need to do is fork the <a href=\"https:\/\/github.com\/melpa\/melpa\">MELPA project on GitHub<\/a>\u00a0and add a recipe for your project. MELPA has docs for configuring more complex projects, but for a simple one-file package, the recipe is really easy.<\/p>\n<p>The recipe for the Mocha runner looks like this:<\/p>\n<pre><code class=\"language-Lisp\">\n(mocha\n :repo \"scottaj\/mocha.el\"\n :fetcher github)\n<\/code><\/pre>\n<p>That\u2019s it, just a path to the GitHub repository. Once the recipe is added, you can open a pull request against MELPA. Someone will review your package and may suggest code changes. Once those are done, your pull request will be merged and MELPA will start publishing your package in its regular builds. The best part is, since MELPA pulls your code straight from your source repository, you don\u2019t have to do anything to push updates to MELPA. It will just automatically pull down the latest version of your code.<\/p>\n<p>Well, that is my short guide to creating and publishing an Emacs package. You can find the Mocha.el package I used as an example <a href=\"https:\/\/github.com\/scottaj\/mocha.el\">here<\/a> and my Emacs config <a href=\"https:\/\/github.com\/scottaj\/new-emacs-config\">here<\/a>. Drop me a comment if you have any questions!<\/p>","created_at":"2020-01-12T02:45:55+0000","updated_at":"2020-01-12T02:45:55+0000","published_at":"2016-05-27T12:00:40+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":12,"domain_name":"spin.atomicobject.com","preview_picture":"http:\/\/g.recordit.co\/wWAwkTuVH2.gif","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","ring","compojure","swagger"],"is_public":false,"id":1101,"title":"metosin\/compojure-api","url":"https:\/\/github.com\/metosin\/compojure-api","content":"<p><strong>Psst!<\/strong> If you're starting a new project, why not try out <a href=\"https:\/\/github.com\/metosin\/reitit\">reitit<\/a>?<\/p><p>Stuff on top of <a href=\"https:\/\/github.com\/weavejester\/compojure\">Compojure<\/a> for making sweet web apis.<\/p><ul><li><a href=\"https:\/\/github.com\/Prismatic\/schema\">Schema<\/a> &amp; <a href=\"https:\/\/clojure.org\/about\/spec\" rel=\"nofollow\">clojure.spec<\/a> (2.0.0) for input &amp; output data coercion<\/li>\n<li><a href=\"http:\/\/swagger.io\/\" rel=\"nofollow\">Swagger<\/a> for api documentation, via <a href=\"https:\/\/github.com\/metosin\/ring-swagger\">ring-swagger<\/a> &amp; <a href=\"https:\/\/github.com\/metosin\/spec-tools\">spec-tools<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Async\">Async<\/a> with async-ring, <a href=\"https:\/\/github.com\/ztellman\/manifold\">manifold<\/a> and <a href=\"https:\/\/github.com\/clojure\/core.async\">core.async<\/a> (2.0.0)<\/li>\n<li>Client negotiable formats: <a href=\"http:\/\/www.json.org\/\" rel=\"nofollow\">JSON<\/a>, <a href=\"https:\/\/github.com\/edn-format\/edn\">EDN<\/a> &amp; <a href=\"https:\/\/github.com\/cognitect\/transit-format\">Transit<\/a>, optionally <a href=\"http:\/\/yaml.org\/\" rel=\"nofollow\">YAML<\/a> and <a href=\"http:\/\/msgpack.org\/\" rel=\"nofollow\">MessagePack<\/a><\/li>\n<li>Data-driven <a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Resources-and-Liberator\">resources<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Routing#bi-directional-routing\">Bi-directional<\/a> routing<\/li>\n<li>Bundled middleware for common api behavior (<a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Exception-handling\">exception handling<\/a>, parameters &amp; formats)<\/li>\n<li>Extendable route DSL via <a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Creating-your-own-metadata-handlers\">metadata handlers<\/a><\/li>\n<li>Route functions &amp; macros for putting things together, including the <a href=\"https:\/\/github.com\/wordnik\/swagger-ui\">Swagger-UI<\/a> via <a href=\"https:\/\/github.com\/metosin\/ring-swagger-ui\">ring-swagger-ui<\/a><\/li>\n<li>Requires Clojure 1.9.0 &amp; Java 1.8<\/li>\n<\/ul><p><a href=\"http:\/\/metosin.github.io\/compojure-api\/doc\/\" rel=\"nofollow\">API Docs<\/a> &amp; <a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\">Wiki<\/a><\/p><h2><a id=\"user-content-latest-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#latest-version\"><\/a>Latest version<\/h2><p><a href=\"http:\/\/clojars.org\/metosin\/compojure-api\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/36d2ddfe8a157ae84c2917e4d2c5c9be661d8ea2\/687474703a2f2f636c6f6a6172732e6f72672f6d65746f73696e2f636f6d706f6a7572652d6170692f6c61746573742d76657273696f6e2e737667\" alt=\"Clojars Project\" data-canonical-src=\"http:\/\/clojars.org\/metosin\/compojure-api\/latest-version.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>Latest non-alpha: <code>[metosin\/compojure-api \"1.1.13\"]<\/code>.<\/p><p>See <a href=\"https:\/\/github.com\/metosin\/compojure-api\/blob\/master\/CHANGELOG.md\">CHANGELOG<\/a> for details.<\/p><h2><a id=\"user-content-for-information-and-help\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-information-and-help\"><\/a>For information and help<\/h2><h3><a id=\"user-content-read-the-version-10-blog-post\" class=\"anchor\" aria-hidden=\"true\" href=\"#read-the-version-10-blog-post\"><\/a><a href=\"http:\/\/www.metosin.fi\/blog\/compojure-api-100\/\" rel=\"nofollow\">Read the Version 1.0 Blog Post<\/a><\/h3><h3><a id=\"user-content-schema--spec-coercion-with-200\" class=\"anchor\" aria-hidden=\"true\" href=\"#schema--spec-coercion-with-200\"><\/a><a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Coercion\">Schema &amp; Spec Coercion with 2.0.0<\/a><\/h3><h3><a id=\"user-content-check-wiki-for-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#check-wiki-for-documentation\"><\/a><a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\">Check wiki for documentation<\/a><\/h3><p><a href=\"https:\/\/clojurians.slack.com\/\" rel=\"nofollow\">Clojurians slack<\/a> (<a href=\"http:\/\/clojurians.net\/\" rel=\"nofollow\">join<\/a>) has a channel <a href=\"https:\/\/clojurians.slack.com\/messages\/ring-swagger\/\" rel=\"nofollow\">#ring-swagger<\/a> for talk about any libraries using Ring-swagger. You can also ask questions about Compojure-api and Ring-swagger on other channels at Clojurians Slack or at #clojure on Freenode IRC (mention <code>compojure-api<\/code> or <code>ring-swagger<\/code> to highlight us).<\/p><h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2><h3><a id=\"user-content-hello-world-api\" class=\"anchor\" aria-hidden=\"true\" href=\"#hello-world-api\"><\/a>Hello World Api<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(require '[compojure.api.sweet :refer :all])\n(require '[ring.util.http-response :refer :all])\n(def app\n  (api\n    (GET \"\/hello\" []\n      :query-params [name :- String]\n      (ok {:message (str \"Hello, \" name)}))))<\/pre><\/div><h3><a id=\"user-content-hello-world-async\" class=\"anchor\" aria-hidden=\"true\" href=\"#hello-world-async\"><\/a>Hello World, async<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(require '[compojure.api.sweet :refer :all])\n(require '[clojure.core.async :as a])\n(GET \"\/hello-async\" []\n  :query-params [name :- String]\n  (a\/go\n    (a\/&lt;! (a\/timeout 500))\n    (ok {:message (str \"Hello, \" name)})))<\/pre><\/div><p><sub>* requires server to be run in <a href=\"https:\/\/github.com\/metosin\/compojure-api\/wiki\/Async\">async mode<\/a><\/sub><\/p><h3><a id=\"user-content-hello-world-async--data-driven\" class=\"anchor\" aria-hidden=\"true\" href=\"#hello-world-async--data-driven\"><\/a>Hello World, async &amp; data-driven<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(require '[compojure.api.sweet :refer :all])\n(require '[clojure.core.async :as a])\n(require '[schema.core :as s])\n(context \"\/hello-async\" []\n  (resource\n    {:get\n     {:parameters {:query-params {:name String}}\n      :responses {200 {:schema {:message String}}\n                  404 {}\n                  500 {:schema s\/Any}}\n      :handler (fn [{{:keys [name]} :query-params}]\n                 (a\/go\n                   (a\/&lt;! (a\/timeout 500))\n                   (ok {:message (str \"Hello, \" name)})))}}))<\/pre><\/div><p><sub>* Note that empty body responses can be specified with <code>{}<\/code> or <code>{:schema s\/Any}<\/code><\/sub><\/p><h3><a id=\"user-content-hello-world-async-data-driven--clojurespec\" class=\"anchor\" aria-hidden=\"true\" href=\"#hello-world-async-data-driven--clojurespec\"><\/a>Hello World, async, data-driven &amp; clojure.spec<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(require '[compojure.api.sweet :refer :all])\n(require '[clojure.core.async :as a])\n(require '[clojure.spec.alpha :as s])\n(s\/def ::name string?)\n(s\/def ::message string?)\n(context \"\/hello-async\" []\n  (resource\n    {:coercion :spec\n     :get {:parameters {:query-params (s\/keys :req-un [::name])}\n           :responses {200 {:schema (s\/keys :req-un [::message])}}\n           :handler (fn [{{:keys [name]} :query-params}]\n                      (a\/go\n                        (a\/&lt;! (a\/timeout 500))\n                        (ok {:message (str \"Hello, \" name)})))}}))<\/pre><\/div><h3><a id=\"user-content-api-with-schema--swagger-docs\" class=\"anchor\" aria-hidden=\"true\" href=\"#api-with-schema--swagger-docs\"><\/a>Api with Schema &amp; Swagger-docs<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(require '[compojure.api.sweet :refer :all])\n(require '[schema.core :as s])\n(s\/defschema Pizza\n  {:name s\/Str\n   (s\/optional-key :description) s\/Str\n   :size (s\/enum :L :M :S)\n   :origin {:country (s\/enum :FI :PO)\n            :city s\/Str}})\n(def app\n  (api\n    {:swagger\n     {:ui \"\/api-docs\"\n      :spec \"\/swagger.json\"\n      :data {:info {:title \"Sample API\"\n                    :description \"Compojure Api example\"}\n             :tags [{:name \"api\", :description \"some apis\"}]\n             :consumes [\"application\/json\"]\n             :produces [\"application\/json\"]}}}\n    (context \"\/api\" []\n      :tags [\"api\"]\n      (GET \"\/plus\" []\n        :return {:result Long}\n        :query-params [x :- Long, y :- Long]\n        :summary \"adds two numbers together\"\n        (ok {:result (+ x y)}))\n      (POST \"\/echo\" []\n        :return Pizza\n        :body [pizza Pizza]\n        :summary \"echoes a Pizza\"\n        (ok pizza)))))<\/pre><\/div><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/wiki\/metosin\/compojure-api\/swagger-api.png\"><img src=\"https:\/\/raw.githubusercontent.com\/wiki\/metosin\/compojure-api\/swagger-api.png\" alt=\"swagger-api\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-more-samples\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-samples\"><\/a>More samples<\/h2><p>To try it yourself, clone this repository and do either:<\/p><ol><li><code>lein run<\/code><\/li>\n<li><code>lein repl<\/code> &amp; <code>(go)<\/code><\/li>\n<\/ol><h2><a id=\"user-content-quick-start-for-a-new-project\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-start-for-a-new-project\"><\/a>Quick start for a new project<\/h2><p>Use a Leiningen template, with or without tests:<\/p><pre>lein new compojure-api my-api\nlein new compojure-api my-api +midje\nlein new compojure-api my-api +clojure-test\n<\/pre><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright \u00a9 2014-2019 <a href=\"http:\/\/www.metosin.fi\" rel=\"nofollow\">Metosin Oy<\/a> and contributors.<\/p><p>Distributed under the Eclipse Public License, see <code>LICENSE<\/code>.<\/p>","created_at":"2020-01-12T02:04:41+0000","updated_at":"2020-01-12T02:04:48+0000","published_by":["metosin"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1569418?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1100,"title":"FastAPI","url":"https:\/\/fastapi.tiangolo.com\/","content":"<p style=\"text-align: center;\"><a href=\"https:\/\/fastapi.tiangolo.com\"><img src=\"https:\/\/fastapi.tiangolo.com\/img\/logo-margin\/logo-teal.png\" alt=\"FastAPI\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p style=\"text-align: center;\"><em>FastAPI framework, high performance, easy to learn, fast to code, ready for production<\/em><\/p><p style=\"text-align: center;\"><a href=\"https:\/\/travis-ci.com\/tiangolo\/fastapi\" target=\"_blank\"><img src=\"https:\/\/travis-ci.com\/tiangolo\/fastapi.svg?branch=master\" alt=\"Build Status\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/codecov.io\/gh\/tiangolo\/fastapi\" target=\"_blank\"><img src=\"https:\/\/codecov.io\/gh\/tiangolo\/fastapi\/branch\/master\/graph\/badge.svg\" alt=\"Coverage\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/pypi.org\/project\/fastapi\" target=\"_blank\"><img src=\"https:\/\/badge.fury.io\/py\/fastapi.svg\" alt=\"Package version\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/gitter.im\/tiangolo\/fastapi?utm_source=badge&amp;utm_medium=badge&amp;utm_campaign=pr-badge&amp;utm_content=badge\" target=\"_blank\"><img src=\"https:\/\/badges.gitter.im\/tiangolo\/fastapi.svg\" alt=\"Join the chat at https:\/\/gitter.im\/tiangolo\/fastapi\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><hr \/><p><strong>Documentation<\/strong>: <a href=\"https:\/\/fastapi.tiangolo.com\" target=\"_blank\">https:\/\/fastapi.tiangolo.com<\/a><\/p><p><strong>Source Code<\/strong>: <a href=\"https:\/\/github.com\/tiangolo\/fastapi\" target=\"_blank\">https:\/\/github.com\/tiangolo\/fastapi<\/a><\/p><hr \/><p>FastAPI is a modern, fast (high-performance), web framework for building APIs with Python 3.6+ based on standard Python type hints.<\/p><p>The key features are:<\/p><ul><li>\n<p><strong>Fast<\/strong>: Very high performance, on par with <strong>NodeJS<\/strong> and <strong>Go<\/strong> (thanks to Starlette and Pydantic). <a href=\"#performance\">One of the fastest Python frameworks available<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong>Fast to code<\/strong>: Increase the speed to develop features by about 200% to 300% *.<\/p>\n<\/li>\n<li><strong>Fewer bugs<\/strong>: Reduce about 40% of human (developer) induced errors. *<\/li>\n<li><strong>Intuitive<\/strong>: Great editor support. <abbr title=\"also known as auto-complete, autocompletion, IntelliSense\">Completion<\/abbr> everywhere. Less time debugging.<\/li>\n<li><strong>Easy<\/strong>: Designed to be easy to use and learn. Less time reading docs.<\/li>\n<li><strong>Short<\/strong>: Minimize code duplication. Multiple features from each parameter declaration. Fewer bugs.<\/li>\n<li><strong>Robust<\/strong>: Get production-ready code. With automatic interactive documentation.<\/li>\n<li><strong>Standards-based<\/strong>: Based on (and fully compatible with) the open standards for APIs: <a href=\"https:\/\/github.com\/OAI\/OpenAPI-Specification\" target=\"_blank\">OpenAPI<\/a> (previously known as Swagger) and <a href=\"http:\/\/json-schema.org\/\" target=\"_blank\">JSON Schema<\/a>.<\/li>\n<\/ul><p><small>* estimation based on tests on an internal development team, building production applications.<\/small><\/p><h2 id=\"opinions\">Opinions<\/h2><p>\"<em>[...] I'm using <strong>FastAPI<\/strong> a ton these days. [...] I'm actually planning to use it for all of my team's <strong>ML services at Microsoft<\/strong>. Some of them are getting integrated into the core <strong>Windows<\/strong> product and some <strong>Office<\/strong> products.<\/em>\"<\/p><p>Kabir Khan - <strong>Microsoft<\/strong> <a href=\"https:\/\/github.com\/tiangolo\/fastapi\/pull\/26\" target=\"_blank\"><small>(ref)<\/small><\/a><\/p><hr \/><p>\"<em>I\u2019m over the moon excited about <strong>FastAPI<\/strong>. It\u2019s so fun!<\/em>\"<\/p><p>Brian Okken - <strong><a href=\"https:\/\/pythonbytes.fm\/episodes\/show\/123\/time-to-right-the-py-wrongs?time_in_sec=855\" target=\"_blank\">Python Bytes<\/a> podcast host<\/strong> <a href=\"https:\/\/twitter.com\/brianokken\/status\/1112220079972728832\" target=\"_blank\"><small>(ref)<\/small><\/a><\/p><hr \/><p>\"<em>Honestly, what you've built looks super solid and polished. In many ways, it's what I wanted <strong>Hug<\/strong> to be - it's really inspiring to see someone build that.<\/em>\"<\/p><p>Timothy Crosley - <strong><a href=\"http:\/\/www.hug.rest\/\" target=\"_blank\">Hug<\/a> creator<\/strong> <a href=\"https:\/\/news.ycombinator.com\/item?id=19455465\" target=\"_blank\"><small>(ref)<\/small><\/a><\/p><hr \/><p>\"<em>If you're looking to learn one <strong>modern framework<\/strong> for building REST APIs, check out <strong>FastAPI<\/strong> [...] It's fast, easy to use and easy to learn [...]<\/em>\"<\/p><p>\"<em>We've switched over to <strong>FastAPI<\/strong> for our <strong>APIs<\/strong> [...] I think you'll like it [...]<\/em>\"<\/p><p>Ines Montani - Matthew Honnibal - <strong><a href=\"https:\/\/explosion.ai\" target=\"_blank\">Explosion AI<\/a> founders - <a href=\"https:\/\/spacy.io\" target=\"_blank\">spaCy<\/a> creators<\/strong> <a href=\"https:\/\/twitter.com\/_inesmontani\/status\/1144173225322143744\" target=\"_blank\"><small>(ref)<\/small><\/a> - <a href=\"https:\/\/twitter.com\/honnibal\/status\/1144031421859655680\" target=\"_blank\"><small>(ref)<\/small><\/a><\/p><hr \/><p>\"<em>We adopted the <strong>FastAPI<\/strong> library to spawn a <strong>REST<\/strong> server that can be queried to obtain <strong>predictions<\/strong>. [for Ludwig]<\/em>\"<\/p><p>Piero Molino, Yaroslav Dudin, and Sai Sumanth Miryala - <strong>Uber<\/strong> <a href=\"https:\/\/eng.uber.com\/ludwig-v0-2\/\" target=\"_blank\"><small>(ref)<\/small><\/a><\/p><hr \/><h2 id=\"requirements\">Requirements<\/h2><p>Python 3.6+<\/p><p>FastAPI stands on the shoulders of giants:<\/p><ul><li><a href=\"https:\/\/www.starlette.io\/\" target=\"_blank\">Starlette<\/a> for the web parts.<\/li>\n<li><a href=\"https:\/\/pydantic-docs.helpmanual.io\/\" target=\"_blank\">Pydantic<\/a> for the data parts.<\/li>\n<\/ul><h2 id=\"installation\">Installation<\/h2><div class=\"codehilite\"><pre>pip install fastapi\n<\/pre><\/div><p>You will also need an ASGI server, for production such as <a href=\"http:\/\/www.uvicorn.org\" target=\"_blank\">Uvicorn<\/a> or <a href=\"https:\/\/gitlab.com\/pgjones\/hypercorn\" target=\"_blank\">Hypercorn<\/a>.<\/p><div class=\"codehilite\"><pre>pip install uvicorn\n<\/pre><\/div><h2 id=\"example\">Example<\/h2><h3 id=\"create-it\">Create it<\/h3><ul><li>Create a file <code>main.py<\/code> with:<\/li>\n<\/ul><div class=\"codehilite\"><pre>from fastapi import FastAPI\napp = FastAPI()\n@app.get(\"\/\")\ndef read_root():\n    return {\"Hello\": \"World\"}\n@app.get(\"\/items\/{item_id}\")\ndef read_item(item_id: int, q: str = None):\n    return {\"item_id\": item_id, \"q\": q}\n<\/pre><\/div><details><summary>Or use <code>async def<\/code>...<\/summary><p>If your code uses <code>async<\/code> \/ <code>await<\/code>, use <code>async def<\/code>:<\/p>\n<div class=\"codehilite\"><pre>from fastapi import FastAPI\napp = FastAPI()\n@app.get(\"\/\")\nasync def read_root():\n    return {\"Hello\": \"World\"}\n@app.get(\"\/items\/{item_id}\")\nasync def read_item(item_id: int, q: str = None):\n    return {\"item_id\": item_id, \"q\": q}\n<\/pre><\/div>\n<p><strong>Note<\/strong>:<\/p>\n<p>If you don't know, check the <em>\"In a hurry?\"<\/em> section about <a href=\"https:\/\/fastapi.tiangolo.com\/async\/#in-a-hurry\" target=\"_blank\"><code>async<\/code> and <code>await<\/code> in the docs<\/a>.<\/p>\n<\/details><h3 id=\"run-it\">Run it<\/h3><p>Run the server with:<\/p><div class=\"codehilite\"><pre>uvicorn main:app --reload\n<\/pre><\/div><details><summary>About the command <code>uvicorn main:app --reload<\/code>...<\/summary><p>The command <code>uvicorn main:app<\/code> refers to:<\/p>\n<ul><li><code>main<\/code>: the file <code>main.py<\/code> (the Python \"module\").<\/li>\n<li><code>app<\/code>: the object created inside of <code>main.py<\/code> with the line <code>app = FastAPI()<\/code>.<\/li>\n<li><code>--reload<\/code>: make the server restart after code changes. Only do this for development.<\/li>\n<\/ul><\/details><h3 id=\"check-it\">Check it<\/h3><p>Open your browser at <a href=\"http:\/\/127.0.0.1:8000\/items\/5?q=somequery\" target=\"_blank\">http:\/\/127.0.0.1:8000\/items\/5?q=somequery<\/a>.<\/p><p>You will see the JSON response as:<\/p><div class=\"codehilite\"><pre>{\"item_id\": 5, \"q\": \"somequery\"}\n<\/pre><\/div><p>You already created an API that:<\/p><ul><li>Receives HTTP requests in the <em>paths<\/em> <code>\/<\/code> and <code>\/items\/{item_id}<\/code>.<\/li>\n<li>Both <em>paths<\/em> take <code>GET<\/code> <em>operations<\/em> (also known as HTTP <em>methods<\/em>).<\/li>\n<li>The <em>path<\/em> <code>\/items\/{item_id}<\/code> has a <em>path parameter<\/em> <code>item_id<\/code> that should be an <code>int<\/code>.<\/li>\n<li>The <em>path<\/em> <code>\/items\/{item_id}<\/code> has an optional <code>str<\/code> <em>query parameter<\/em> <code>q<\/code>.<\/li>\n<\/ul><h3 id=\"interactive-api-docs\">Interactive API docs<\/h3><p>Now go to <a href=\"http:\/\/127.0.0.1:8000\/docs\" target=\"_blank\">http:\/\/127.0.0.1:8000\/docs<\/a>.<\/p><p>You will see the automatic interactive API documentation (provided by <a href=\"https:\/\/github.com\/swagger-api\/swagger-ui\" target=\"_blank\">Swagger UI<\/a>):<\/p><p><img alt=\"Swagger UI\" src=\"https:\/\/fastapi.tiangolo.com\/img\/index\/index-01-swagger-ui-simple.png\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"alternative-api-docs\">Alternative API docs<\/h3><p>And now, go to <a href=\"http:\/\/127.0.0.1:8000\/redoc\" target=\"_blank\">http:\/\/127.0.0.1:8000\/redoc<\/a>.<\/p><p>You will see the alternative automatic documentation (provided by <a href=\"https:\/\/github.com\/Rebilly\/ReDoc\" target=\"_blank\">ReDoc<\/a>):<\/p><p><img alt=\"ReDoc\" src=\"https:\/\/fastapi.tiangolo.com\/img\/index\/index-02-redoc-simple.png\" referrerpolicy=\"no-referrer\" \/><\/p><h2 id=\"example-upgrade\">Example upgrade<\/h2><p>Now modify the file <code>main.py<\/code> to receive a body from a <code>PUT<\/code> request.<\/p><p>Declare the body using standard Python types, thanks to Pydantic.<\/p><div class=\"codehilite\"><pre>from fastapi import FastAPI\nfrom pydantic import BaseModel\napp = FastAPI()\nclass Item(BaseModel):\n    name: str\n    price: float\n    is_offer: bool = None\n@app.get(\"\/\")\ndef read_root():\n    return {\"Hello\": \"World\"}\n@app.get(\"\/items\/{item_id}\")\ndef read_item(item_id: int, q: str = None):\n    return {\"item_id\": item_id, \"q\": q}\n@app.put(\"\/items\/{item_id}\")\ndef update_item(item_id: int, item: Item):\n    return {\"item_name\": item.name, \"item_id\": item_id}\n<\/pre><\/div><p>The server should reload automatically (because you added <code>--reload<\/code> to the <code>uvicorn<\/code> command above).<\/p><h3 id=\"interactive-api-docs-upgrade\">Interactive API docs upgrade<\/h3><p>Now go to <a href=\"http:\/\/127.0.0.1:8000\/docs\" target=\"_blank\">http:\/\/127.0.0.1:8000\/docs<\/a>.<\/p><ul><li>The interactive API documentation will be automatically updated, including the new body:<\/li>\n<\/ul><p><img alt=\"Swagger UI\" src=\"https:\/\/fastapi.tiangolo.com\/img\/index\/index-03-swagger-02.png\" referrerpolicy=\"no-referrer\" \/><\/p><ul><li>Click on the button \"Try it out\", it allows you to fill the parameters and directly interact with the API:<\/li>\n<\/ul><p><img alt=\"Swagger UI interaction\" src=\"https:\/\/fastapi.tiangolo.com\/img\/index\/index-04-swagger-03.png\" referrerpolicy=\"no-referrer\" \/><\/p><ul><li>Then click on the \"Execute\" button, the user interface will communicate with your API, send the parameters, get the results and show them on the screen:<\/li>\n<\/ul><p><img alt=\"Swagger UI interaction\" src=\"https:\/\/fastapi.tiangolo.com\/img\/index\/index-05-swagger-04.png\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"alternative-api-docs-upgrade\">Alternative API docs upgrade<\/h3><p>And now, go to <a href=\"http:\/\/127.0.0.1:8000\/redoc\" target=\"_blank\">http:\/\/127.0.0.1:8000\/redoc<\/a>.<\/p><ul><li>The alternative documentation will also reflect the new query parameter and body:<\/li>\n<\/ul><p><img alt=\"ReDoc\" src=\"https:\/\/fastapi.tiangolo.com\/img\/index\/index-06-redoc-02.png\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"recap\">Recap<\/h3><p>In summary, you declare <strong>once<\/strong> the types of parameters, body, etc. as function parameters.<\/p><p>You do that with standard modern Python types.<\/p><p>You don't have to learn a new syntax, the methods or classes of a specific library, etc.<\/p><p>Just standard <strong>Python 3.6+<\/strong>.<\/p><p>For example, for an <code>int<\/code>:<\/p><div class=\"codehilite\"><pre>item_id: int\n<\/pre><\/div><p>or for a more complex <code>Item<\/code> model:<\/p><div class=\"codehilite\"><pre>item: Item\n<\/pre><\/div><p>...and with that single declaration you get:<\/p><ul><li>Editor support, including:\n<ul><li>Completion.<\/li>\n<li>Type checks.<\/li>\n<\/ul><\/li>\n<li>Validation of data:\n<ul><li>Automatic and clear errors when the data is invalid.<\/li>\n<li>Validation even for deeply nested JSON objects.<\/li>\n<\/ul><\/li>\n<li><abbr title=\"also known as: serialization, parsing, marshalling\">Conversion<\/abbr> of input data: coming from the network to Python data and types. Reading from:\n<ul><li>JSON.<\/li>\n<li>Path parameters.<\/li>\n<li>Query parameters.<\/li>\n<li>Cookies.<\/li>\n<li>Headers.<\/li>\n<li>Forms.<\/li>\n<li>Files.<\/li>\n<\/ul><\/li>\n<li><abbr title=\"also known as: serialization, parsing, marshalling\">Conversion<\/abbr> of output data: converting from Python data and types to network data (as JSON):\n<ul><li>Convert Python types (<code>str<\/code>, <code>int<\/code>, <code>float<\/code>, <code>bool<\/code>, <code>list<\/code>, etc).<\/li>\n<li><code>datetime<\/code> objects.<\/li>\n<li><code>UUID<\/code> objects.<\/li>\n<li>Database models.<\/li>\n<li>...and many more.<\/li>\n<\/ul><\/li>\n<li>Automatic interactive API documentation, including 2 alternative user interfaces:\n<ul><li>Swagger UI.<\/li>\n<li>ReDoc.<\/li>\n<\/ul><\/li>\n<\/ul><hr \/><p>Coming back to the previous code example, <strong>FastAPI<\/strong> will:<\/p><ul><li>Validate that there is an <code>item_id<\/code> in the path for <code>GET<\/code> and <code>PUT<\/code> requests.<\/li>\n<li>Validate that the <code>item_id<\/code> is of type <code>int<\/code> for <code>GET<\/code> and <code>PUT<\/code> requests.\n<ul><li>If it is not, the client will see a useful, clear error.<\/li>\n<\/ul><\/li>\n<li>Check if there is an optional query parameter named <code>q<\/code> (as in <code>http:\/\/127.0.0.1:8000\/items\/foo?q=somequery<\/code>) for <code>GET<\/code> requests.\n<ul><li>As the <code>q<\/code> parameter is declared with <code>= None<\/code>, it is optional.<\/li>\n<li>Without the <code>None<\/code> it would be required (as is the body in the case with <code>PUT<\/code>).<\/li>\n<\/ul><\/li>\n<li>For <code>PUT<\/code> requests to <code>\/items\/{item_id}<\/code>, Read the body as JSON:\n<ul><li>Check that it has a required attribute <code>name<\/code> that should be a <code>str<\/code>.<\/li>\n<li>Check that is has a required attribute <code>price<\/code> that has to be a <code>float<\/code>.<\/li>\n<li>Check that it has an optional attribute <code>is_offer<\/code>, that should be a <code>bool<\/code>, if present.<\/li>\n<li>All this would also work for deeply nested JSON objects.<\/li>\n<\/ul><\/li>\n<li>Convert from and to JSON automatically.<\/li>\n<li>Document everything with OpenAPI, that can be used by:\n<ul><li>Interactive documentation systems.<\/li>\n<li>Automatic client code generation systems, for many languages.<\/li>\n<\/ul><\/li>\n<li>Provide 2 interactive documentation web interfaces directly.<\/li>\n<\/ul><hr \/><p>We just scratched the surface, but you already get the idea of how it all works.<\/p><p>Try changing the line with:<\/p><div class=\"codehilite\"><pre>    return {\"item_name\": item.name, \"item_id\": item_id}\n<\/pre><\/div><p>...from:<\/p><div class=\"codehilite\"><pre>        ... \"item_name\": item.name ...\n<\/pre><\/div><p>...to:<\/p><div class=\"codehilite\"><pre>        ... \"item_price\": item.price ...\n<\/pre><\/div><p>...and see how your editor will auto-complete the attributes and know their types:<\/p><p><img alt=\"editor support\" src=\"https:\/\/fastapi.tiangolo.com\/img\/vscode-completion.png\" referrerpolicy=\"no-referrer\" \/><\/p><p>For a more complete example including more features, see the <a href=\"https:\/\/fastapi.tiangolo.com\/tutorial\/intro\/\">Tutorial - User Guide<\/a>.<\/p><p><strong>Spoiler alert<\/strong>: the tutorial - user guide includes:<\/p><ul><li>Declaration of <strong>parameters<\/strong> from other different places as: <strong>headers<\/strong>, <strong>cookies<\/strong>, <strong>form fields<\/strong> and <strong>files<\/strong>.<\/li>\n<li>How to set <strong>validation constraints<\/strong> as <code>maximum_length<\/code> or <code>regex<\/code>.<\/li>\n<li>A very powerful and easy to use <strong><abbr title=\"also known as components, resources, providers, services, injectables\">Dependency Injection<\/abbr><\/strong> system.<\/li>\n<li>Security and authentication, including support for <strong>OAuth2<\/strong> with <strong>JWT tokens<\/strong> and <strong>HTTP Basic<\/strong> auth.<\/li>\n<li>More advanced (but equally easy) techniques for declaring <strong>deeply nested JSON models<\/strong> (thanks to Pydantic).<\/li>\n<li>Many extra features (thanks to Starlette) as:\n<ul><li><strong>WebSockets<\/strong><\/li>\n<li><strong>GraphQL<\/strong><\/li>\n<li>extremely easy tests based on <code>requests<\/code> and <code>pytest<\/code><\/li>\n<li><strong>CORS<\/strong><\/li>\n<li><strong>Cookie Sessions<\/strong><\/li>\n<li>...and more.<\/li>\n<\/ul><\/li>\n<\/ul><h2 id=\"performance\">Performance<\/h2><p>Independent TechEmpower benchmarks show <strong>FastAPI<\/strong> applications running under Uvicorn as <a href=\"https:\/\/www.techempower.com\/benchmarks\/#section=test&amp;runid=7464e520-0dc2-473d-bd34-dbdfd7e85911&amp;hw=ph&amp;test=query&amp;l=zijzen-7\" target=\"_blank\">one of the fastest Python frameworks available<\/a>, only below Starlette and Uvicorn themselves (used internally by FastAPI). (*)<\/p><p>To understand more about it, see the section <a href=\"https:\/\/fastapi.tiangolo.com\/benchmarks\/\" target=\"_blank\">Benchmarks<\/a>.<\/p><h2 id=\"optional-dependencies\">Optional Dependencies<\/h2><p>Used by Pydantic:<\/p><ul><li><a href=\"https:\/\/github.com\/esnme\/ultrajson\" target=\"_blank\"><code>ujson<\/code><\/a> - for faster JSON <abbr title=\"converting the string that comes from an HTTP request into Python data\">\"parsing\"<\/abbr>.<\/li>\n<li><a href=\"https:\/\/github.com\/JoshData\/python-email-validator\" target=\"_blank\"><code>email_validator<\/code><\/a> - for email validation.<\/li>\n<\/ul><p>Used by Starlette:<\/p><ul><li><a href=\"http:\/\/docs.python-requests.org\" target=\"_blank\"><code>requests<\/code><\/a> - Required if you want to use the <code>TestClient<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/Tinche\/aiofiles\" target=\"_blank\"><code>aiofiles<\/code><\/a> - Required if you want to use <code>FileResponse<\/code> or <code>StaticFiles<\/code>.<\/li>\n<li><a href=\"http:\/\/jinja.pocoo.org\" target=\"_blank\"><code>jinja2<\/code><\/a> - Required if you want to use the default template configuration.<\/li>\n<li><a href=\"https:\/\/andrew-d.github.io\/python-multipart\/\" target=\"_blank\"><code>python-multipart<\/code><\/a> - Required if you want to support form <abbr title=\"converting the string that comes from an HTTP request into Python data\">\"parsing\"<\/abbr>, with <code>request.form()<\/code>.<\/li>\n<li><a href=\"https:\/\/pythonhosted.org\/itsdangerous\/\" target=\"_blank\"><code>itsdangerous<\/code><\/a> - Required for <code>SessionMiddleware<\/code> support.<\/li>\n<li><a href=\"https:\/\/pyyaml.org\/wiki\/PyYAMLDocumentation\" target=\"_blank\"><code>pyyaml<\/code><\/a> - Required for <code>SchemaGenerator<\/code> support.<\/li>\n<li><a href=\"https:\/\/graphene-python.org\/\" target=\"_blank\"><code>graphene<\/code><\/a> - Required for <code>GraphQLApp<\/code> support.<\/li>\n<li><a href=\"https:\/\/github.com\/esnme\/ultrajson\" target=\"_blank\"><code>ujson<\/code><\/a> - Required if you want to use <code>UJSONResponse<\/code>.<\/li>\n<\/ul><p>Used by FastAPI \/ Starlette:<\/p><ul><li><a href=\"http:\/\/www.uvicorn.org\" target=\"_blank\"><code>uvicorn<\/code><\/a> - for the server that loads and serves your application.<\/li>\n<\/ul><p>You can install all of these with <code>pip install fastapi[all]<\/code>.<\/p><h2 id=\"license\">License<\/h2><p>This project is licensed under the terms of the MIT license.<\/p>","created_at":"2020-01-11T15:17:48+0000","updated_at":"2020-01-11T15:17:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"fastapi.tiangolo.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","python"],"is_public":false,"id":1099,"title":"cnuernber\/libpython-clj","url":"https:\/\/github.com\/cnuernber\/libpython-clj","content":"<p>JNA libpython bindings to the tech ecosystem.<\/p><p><a href=\"https:\/\/clojars.org\/cnuernber\/libpython-clj\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/fc615c431473595e1521852726a47adf955025eb\/68747470733a2f2f696d672e736869656c64732e696f2f636c6f6a6172732f762f636e7565726e6265722f6c6962707974686f6e2d636c6a2e737667\" alt=\"Clojars Project\" data-canonical-src=\"https:\/\/img.shields.io\/clojars\/v\/cnuernber\/libpython-clj.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/travis-ci.com\/cnuernber\/libpython-clj\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/4a87e6df5edf37e5adad47087b317c07877fbff8\/68747470733a2f2f7472617669732d63692e636f6d2f636e7565726e6265722f6c6962707974686f6e2d636c6a2e7376673f6272616e63683d6d6173746572\" alt=\"travis integration\" data-canonical-src=\"https:\/\/travis-ci.com\/cnuernber\/libpython-clj.svg?branch=master\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><ul><li>Bridge between JVM objects and Python objects easily; use Python in your Java and use some Java in your Python.<\/li>\n<li>Python objects are linked to the JVM GC such that when they are no longer reachable from the JVM their references are released. Scope based resource contexts are <a href=\"https:\/\/github.com\/cnuernber\/libpython-clj\/blob\/master\/docs\/scopes-and-gc.md\">also available<\/a>.<\/li>\n<li>Finding the python libraries is done dynamically allowing one system to run on multiple versions of python.<\/li>\n<li>REPL oriented design means fast, smooth, iterative development.<\/li>\n<\/ul><h2><a id=\"user-content-vision\" class=\"anchor\" aria-hidden=\"true\" href=\"#vision\"><\/a>Vision<\/h2><p>We aim to integrate Python into Clojure at a deep level. This means that we want to be able to load\/use python modules almost as if they were Clojure namespaces. We also want to be able to use Clojure to extend Python objects. I gave a <a href=\"https:\/\/www.youtube.com\/watch?v=vQPW16_jixs\" rel=\"nofollow\">talk at Clojure Conj 2019<\/a> that outlines more of what is going on.<\/p><p>This code is a concrete example that generates an <a href=\"https:\/\/github.com\/cnuernber\/facial-rec\">embedding for faces<\/a>:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(ns facial-rec.face-feature\n  (:require [libpython-clj.require :refer [require-python]]\n            [libpython-clj.python :as py]\n            [tech.v2.datatype :as dtype]))\n(require-python '(mxnet mxnet.ndarray mxnet.module mxnet.io mxnet.model))\n(require-python 'cv2)\n(require-python '[numpy :as np])\n(defn load-model\n  [&amp; {:keys [model-path checkpoint]\n      :or {model-path \"models\/recognition\/model\"\n           checkpoint 0}}]\n  (let [[sym arg-params aux-params] (mxnet.model\/load_checkpoint model-path checkpoint)\n        all-layers (py\/$a sym get_internals)\n        target-layer (py\/get-item all-layers \"fc1_output\")\n        ;;TODO - We haven't overloaded enough of the IFn invoke methods for\n                ;;this to work without using call-kw\n        model (py\/call-kw mxnet.module\/Module [] {:symbol target-layer :context (mxnet\/cpu) :label_names nil})]\n    (py\/$a model bind :data_shapes [[\"data\" [1 3 112 112]]])\n    (py\/$a model set_params arg-params aux-params)\n    model))\n(defonce model (load-model))\n(defn face-&gt;feature\n  [img-path]\n  (py\/with-gil-stack-rc-context\n    (if-let [new-img (cv2\/imread img-path)]\n      (let [new-img (cv2\/cvtColor new-img cv2\/COLOR_BGR2RGB)\n            new-img (np\/transpose new-img [2 0 1])\n            input-blob (np\/expand_dims new-img :axis 0)\n            data (mxnet.ndarray\/array input-blob)\n            batch (mxnet.io\/DataBatch :data [data])]\n        (py\/$a model forward batch :is_train false)\n        (-&gt; (py\/$a model get_outputs)\n            first\n            (py\/$a asnumpy)\n            (#(dtype\/make-container :java-array :float32 %))))\n      (throw (Exception. (format \"Failed to load img: %s\" img-path))))))<\/pre><\/div><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (require '[libpython-clj.require :refer [require-python]])\n...logging info....\nnil\nuser&gt; (require-python '[numpy :as np])\nnil\nuser&gt; (def test-ary (np\/array [[1 2][3 4]]))\n#'user\/test-ary\nuser&gt; test-ary\n[[1 2]\n [3 4]]<\/pre><\/div><p>We have a <a href=\"https:\/\/github.com\/cnuernber\/libpython-clj\/blob\/master\/docs\/Usage.md\">document<\/a> on all the features but beginning usage is pretty simple. Import your modules, use the things from Clojure. We have put effort into making sure things like sequences and ranges transfer between the two languages.<\/p><h4><a id=\"user-content-environments\" class=\"anchor\" aria-hidden=\"true\" href=\"#environments\"><\/a>Environments<\/h4><p>One very complimentary aspect of Python with respect to Clojure is it's integration with cutting edge native libraries. Our support isn't perfect so some understanding of the mechanism is important to diagnose errors and issues.<\/p><p>Current, we launch the python3 executable and print out various different bits of configuration as json. We parse the json and use the output to attempt to find the <code>libpython3.Xm.so<\/code> shared library so for example if we are loading python 3.6 we look for <code>libpython3.6m.so<\/code> on Linux or <code>libpython3.6m.dylib<\/code> on the Mac.<\/p><p>This pathway has allowed us support Conda albeit with some work. For examples using Conda, check out the facial rec repository above or look into how we <a href=\"https:\/\/github.com\/cnuernber\/libpython-clj\/blob\/master\/scripts\/build-conda-docker\">build<\/a> our test <a href=\"https:\/\/github.com\/cnuernber\/libpython-clj\/blob\/master\/dockerfiles\/CondaDockerfile\">docker containers<\/a>.<\/p><h2><a id=\"user-content-further-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#further-information\"><\/a>Further Information<\/h2><h2><a id=\"user-content-new-to-clojure\" class=\"anchor\" aria-hidden=\"true\" href=\"#new-to-clojure\"><\/a>New To Clojure<\/h2><p>New to Clojure or the JVM? Try remixing the nextjournal entry and playing around there. For more resources on learning and getting more comfortable with Clojure, we have an <a href=\"https:\/\/github.com\/cnuernber\/libpython-clj\/blob\/master\/docs\/new-to-clojure.md\">introductory document<\/a>.<\/p><h2><a id=\"user-content-resources\" class=\"anchor\" aria-hidden=\"true\" href=\"#resources\"><\/a>Resources<\/h2><ul><li><a href=\"https:\/\/docs.python.org\/3.7\/c-api\/index.html#c-api-index\" rel=\"nofollow\">libpython C api<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/23930671\/how-to-create-n-dim-numpy-array-from-a-pointer\" rel=\"nofollow\">create numpy from C ptr<\/a><\/li>\n<li><a href=\"https:\/\/docs.scipy.org\/doc\/numpy\/reference\/generated\/numpy.ndarray.ctypes.html\" rel=\"nofollow\">create C ptr from numpy<\/a><\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright \u00a9 2019 Chris Nuernberger<\/p><p>This program and the accompanying materials are made available under the terms of the Eclipse Public License 2.0 which is available at <a href=\"http:\/\/www.eclipse.org\/legal\/epl-2.0\" rel=\"nofollow\">http:\/\/www.eclipse.org\/legal\/epl-2.0<\/a>.<\/p>","created_at":"2020-01-10T01:08:48+0000","updated_at":"2020-01-10T01:08:54+0000","published_by":["cnuernber"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/40426?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","clojurescript"],"is_public":false,"id":1098,"title":"redplanetlabs\/specter","url":"https:\/\/github.com\/redplanetlabs\/specter","content":"<p>Specter rejects Clojure's restrictive approach to immutable data structure manipulation, instead exposing an elegant API to allow any sort of manipulation imaginable. Specter especially excels at querying and transforming nested and recursive data, important use cases that are very complex to handle with vanilla Clojure.<\/p><p>Specter has an extremely simple core, just a single abstraction called \"navigator\". Queries and transforms are done by composing navigators into a \"path\" precisely targeting what you want to retrieve or change. Navigators can be composed with any other navigators, allowing sophisticated manipulations to be expressed very concisely.<\/p><p>In addition, Specter has performance rivaling hand-optimized code (see <a href=\"https:\/\/gist.github.com\/nathanmarz\/b7c612b417647db80b9eaab618ff8d83\">this benchmark<\/a>). Clojure's only comparable built-in operations are <code>get-in<\/code> and <code>update-in<\/code>, and the Specter equivalents are 30% and 85% faster respectively (while being just as concise). Under the hood, Specter uses <a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/Specter's-inline-caching-implementation\">advanced dynamic techniques<\/a> to strip away the overhead of composition.<\/p><p>There are some key differences between the Clojure approach to data manipulation and the Specter approach. Unlike Clojure, Specter always uses the most efficient method possible to implement an operation for a datatype (e.g. <code>last<\/code> vs. <code>LAST<\/code>). Clojure intentionally leaves out many operations, such as prepending to a vector or inserting into the middle of a sequence. Specter has navigators that cover these use cases (<code>BEFORE-ELEM<\/code> and <code>before-index<\/code>) and many more. Finally, Specter transforms always target precise parts of a data structure, leaving everything else the same. For instance, <code>ALL<\/code> targets every value within a sequence, and the resulting transform is always the same type as the input (e.g. a vector stays a vector, a sorted map stays a sorted map).<\/p><p>Consider these examples:<\/p><p><strong>Example 1: Increment every even number nested within map of vector of maps<\/strong><\/p><div class=\"highlight highlight-source-clojure\"><pre>(def data {:a [{:aa 1 :bb 2}\n               {:cc 3}]\n           :b [{:dd 4}]})\n;; Manual Clojure\n(defn map-vals [m afn]\n  (-&gt;&gt; m (map (fn [[k v]] [k (afn v)])) (into (empty m))))\n(map-vals data\n  (fn [v]\n    (mapv\n      (fn [m]\n        (map-vals\n          m\n          (fn [v] (if (even? v) (inc v) v))))\n      v)))\n;; Specter\n(transform [MAP-VALS ALL MAP-VALS even?] inc data)<\/pre><\/div><p><strong>Example 2: Append a sequence of elements to a nested vector<\/strong><\/p><div class=\"highlight highlight-source-clojure\"><pre>(def data {:a [1 2 3]})\n;; Manual Clojure\n(update data :a (fn [v] (into (if v v []) [4 5])))\n;; Specter\n(setval [:a END] [4 5] data)<\/pre><\/div><p><strong>Example 3: Increment the last odd number in a sequence<\/strong><\/p><div class=\"highlight highlight-source-clojure\"><pre>(def data [1 2 3 4 5 6 7 8])\n;; Manual Clojure\n(let [idx (reduce-kv (fn [res i v] (if (odd? v) i res)) nil data)]\n  (if idx (update data idx inc) data))\n;; Specter\n(transform [(filterer odd?) LAST] inc data)<\/pre><\/div><p><strong>Example 4: Map a function over a sequence without changing the type or order of the sequence<\/strong><\/p><div class=\"highlight highlight-source-clojure\"><pre>;; Manual Clojure\n(map inc data) ;; doesn't work, becomes a lazy sequence\n(into (empty data) (map inc data)) ;; doesn't work, reverses the order of lists\n;; Specter\n(transform ALL inc data) ;; works for all Clojure datatypes with near-optimal efficiency<\/pre><\/div><p>The latest release version of Specter is hosted on <a href=\"https:\/\/clojars.org\" rel=\"nofollow\">Clojars<\/a>:<\/p><p><a href=\"https:\/\/clojars.org\/com.rpl\/specter\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/67466df9e58ef88e2f05d5e9fa2b15c648f1fa4d\/68747470733a2f2f636c6f6a6172732e6f72672f636f6d2e72706c2f737065637465722f6c61746573742d76657273696f6e2e737667\" alt=\"Current Version\" data-canonical-src=\"https:\/\/clojars.org\/com.rpl\/specter\/latest-version.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><ul><li>Introductory blog post: <a href=\"http:\/\/nathanmarz.com\/blog\/clojures-missing-piece.html\" rel=\"nofollow\">Clojure's missing piece<\/a><\/li>\n<li>Presentation about Specter: <a href=\"https:\/\/www.youtube.com\/watch?v=VTCy_DkAJGk\" rel=\"nofollow\">Specter: Powerful and Simple Data Structure Manipulation<\/a>\n<ul><li>Note that this presentation was given before Specter's inline compilation\/caching system was developed. You no longer need to do anything special to get near-optimal performance.<\/li>\n<\/ul><\/li>\n<li>Screencast on Specter: <a href=\"https:\/\/www.youtube.com\/watch?v=rh5J4vacG98\" rel=\"nofollow\">Understanding Specter<\/a><\/li>\n<li>List of navigators with examples: <a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/List-of-Navigators\">This wiki page<\/a> provides a more comprehensive overview than the API docs about the behavior of specific navigators and includes many examples.<\/li>\n<li>Core operations and defining new navigators: <a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/List-of-Macros\">This wiki page<\/a> provides a more comprehensive overview than the API docs of the core select\/transform\/etc. operations and the operations for defining new navigators.<\/li>\n<li><a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/Using-Specter-Recursively\">This wiki page<\/a> explains how to do precise and efficient recursive navigation with Specter.<\/li>\n<li><a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/Using-Specter-With-Zippers\">This wiki page<\/a> provides a comprehensive overview of how to use Specter's zipper navigators. Zippers are a much slower navigation method but can perform certain tasks that are not possible with Specter's regular navigators. Note that zippers are rarely needed.<\/li>\n<li><a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/Cheat-Sheet\">Cheat Sheet<\/a><\/li>\n<li><a href=\"http:\/\/redplanetlabs.github.io\/specter\/\" rel=\"nofollow\">API docs<\/a><\/li>\n<li>Performance guide: <a href=\"https:\/\/github.com\/redplanetlabs\/specter\/wiki\/Specter's-inline-caching-implementation\">This post<\/a> provides an overview of how Specter achieves its performance.<\/li>\n<\/ul><p>Specter's API is contained in these files:<\/p><ul><li><a href=\"https:\/\/github.com\/redplanetlabs\/specter\/blob\/master\/src\/clj\/com\/rpl\/specter.cljc\">specter.cljc<\/a>: This contains the built-in navigators and the definition of the core operations.<\/li>\n<li><a href=\"https:\/\/github.com\/redplanetlabs\/specter\/blob\/master\/src\/clj\/com\/rpl\/specter\/transients.cljc\">transients.cljc<\/a>: This contains navigators for transient collections.<\/li>\n<li><a href=\"https:\/\/github.com\/redplanetlabs\/specter\/blob\/master\/src\/clj\/com\/rpl\/specter\/zipper.cljc\">zipper.cljc<\/a>: This integrates zipper-based navigation into Specter.<\/li>\n<\/ul><p>You can ask questions about Specter by <a href=\"https:\/\/github.com\/redplanetlabs\/specter\/issues?utf8=%E2%9C%93&amp;q=is%3Aissue+label%3Aquestion+\">opening an issue<\/a> on Github.<\/p><p>You can also find help in the #specter channel on <a href=\"http:\/\/clojurians.net\/\" rel=\"nofollow\">Clojurians<\/a>.<\/p><p>Increment all the values in maps of maps:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (use 'com.rpl.specter)\nuser&gt; (transform [MAP-VALS MAP-VALS]\n              inc\n              {:a {:aa 1} :b {:ba -1 :bb 2}})\n{:a {:aa 2}, :b {:ba 0, :bb 3}}<\/pre><\/div><p>Increment all the even values for :a keys in a sequence of maps:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [ALL :a even?]\n              inc\n              [{:a 1} {:a 2} {:a 4} {:a 3}])\n[{:a 1} {:a 3} {:a 5} {:a 3}]<\/pre><\/div><p>Retrieve every number divisible by 3 out of a sequence of sequences:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (select [ALL ALL #(= 0 (mod % 3))]\n              [[1 2 3 4] [] [5 3 2 18] [2 4 6] [12]])\n[3 3 18 6 12]<\/pre><\/div><p>Increment the last odd number in a sequence:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [(filterer odd?) LAST]\n              inc\n              [2 1 3 6 9 4 8])\n[2 1 3 6 10 4 8]<\/pre><\/div><p>Remove nils from a nested sequence:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (setval [:a ALL nil?] NONE {:a [1 2 nil 3 nil]})\n{:a [1 2 3]}<\/pre><\/div><p>Remove key\/value pair from nested map:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (setval [:a :b :c] NONE {:a {:b {:c 1}}})\n{:a {:b {}}}<\/pre><\/div><p>Remove key\/value pair from nested map, removing maps that become empty along the way:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (setval [:a (compact :b :c)] NONE {:a {:b {:c 1}}})\n{}<\/pre><\/div><p>Increment all the odd numbers between indices 1 (inclusive) and 4 (exclusive):<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [(srange 1 4) ALL odd?] inc [0 1 2 3 4 5 6 7])\n[0 2 2 4 4 5 6 7]<\/pre><\/div><p>Replace the subsequence from indices 2 to 4 with [:a :b :c :d :e]:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (setval (srange 2 4) [:a :b :c :d :e] [0 1 2 3 4 5 6 7 8 9])\n[0 1 :a :b :c :d :e 4 5 6 7 8 9]<\/pre><\/div><p>Concatenate the sequence [:a :b] to every nested sequence of a sequence:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (setval [ALL END] [:a :b] [[1] '(1 2) [:c]])\n[[1 :a :b] (1 2 :a :b) [:c :a :b]]<\/pre><\/div><p>Get all the numbers out of a data structure, no matter how they're nested:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (select (walker number?)\n              {2 [1 2 [6 7]] :a 4 :c {:a 1 :d [2 nil]}})\n[2 1 2 1 2 6 7 4]<\/pre><\/div><p>Navigate with string keys:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (select [\"a\" \"b\"]\n              {\"a\" {\"b\" 10}})\n[10]<\/pre><\/div><p>Reverse the positions of all even numbers between indices 4 and 11:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [(srange 4 11) (filterer even?)]\n              reverse\n              [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15])\n[0 1 2 3 10 5 8 7 6 9 4 11 12 13 14 15]<\/pre><\/div><p>Append [:c :d] to every subsequence that has at least two even numbers:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (setval [ALL\n               (selected? (filterer even?) (view count) (pred&gt;= 2))\n               END]\n              [:c :d]\n              [[1 2 3 4 5 6] [7 0 -1] [8 8] []])\n[[1 2 3 4 5 6 :c :d] [7 0 -1] [8 8 :c :d] []]<\/pre><\/div><p>When doing more involved transformations, you often find you lose context when navigating deep within a data structure and need information \"up\" the data structure to perform the transformation. Specter solves this problem by allowing you to collect values during navigation to use in the transform function. Here's an example which transforms a sequence of maps by adding the value of the :b key to the value of the :a key, but only if the :a key is even:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [ALL (collect-one :b) :a even?]\n              +\n              [{:a 1 :b 3} {:a 2 :b -10} {:a 4 :b 10} {:a 3}])\n[{:b 3, :a 1} {:b -10, :a -8} {:b 10, :a 14} {:a 3}]<\/pre><\/div><p>The transform function receives as arguments all the collected values followed by the navigated to value. So in this case <code>+<\/code> receives the value of the :b key followed by the value of the :a key, and the transform is performed to :a's value.<\/p><p>The four built-in ways for collecting values are <code>VAL<\/code>, <code>collect<\/code>, <code>collect-one<\/code>, and <code>putval<\/code>. <code>VAL<\/code> just adds whatever element it's currently on to the value list, while <code>collect<\/code> and <code>collect-one<\/code> take in a selector to navigate to the desired value. <code>collect<\/code> works just like <code>select<\/code> by finding a sequence of values, while <code>collect-one<\/code> expects to only navigate to a single value. Finally, <code>putval<\/code> adds an external value into the collected values list.<\/p><p>Increment the value for :a key by 10:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [:a (putval 10)]\n              +\n              {:a 1 :b 3})\n{:b 3 :a 11}<\/pre><\/div><p>For every map in a sequence, increment every number in :c's value if :a is even or increment :d if :a is odd:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (transform [ALL (if-path [:a even?] [:c ALL] :d)]\n              inc\n              [{:a 2 :c [1 2] :d 4} {:a 4 :c [0 10 -1]} {:a -1 :c [1 1 1] :d 1}])\n[{:c [2 3], :d 4, :a 2} {:c [1 11 0], :a 4} {:c [1 1 1], :d 2, :a -1}]<\/pre><\/div><p>\"Protocol paths\" can be used to navigate on polymorphic data. For example, if you have two ways of storing \"account\" information:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defrecord Account [funds])\n(defrecord User [account])\n(defrecord Family [accounts-list])<\/pre><\/div><p>You can make an \"AccountPath\" that dynamically chooses its path based on the type of element it is currently navigated to:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defprotocolpath AccountPath [])\n(extend-protocolpath AccountPath\n  User :account\n  Family [:accounts-list ALL])<\/pre><\/div><p>Then, here is how to select all the funds out of a list of <code>User<\/code> and <code>Family<\/code>:<\/p><div class=\"highlight highlight-source-clojure\"><pre>user&gt; (select [ALL AccountPath :funds]\n        [(-&gt;User (-&gt;Account 50))\n         (-&gt;User (-&gt;Account 51))\n         (-&gt;Family [(-&gt;Account 1)\n                    (-&gt;Account 2)])\n         ])\n[50 51 1 2]<\/pre><\/div><p>The next examples demonstrate recursive navigation. Here's one way to double all the even numbers in a tree:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defprotocolpath TreeWalker [])\n(extend-protocolpath TreeWalker\n  Object nil\n  clojure.lang.PersistentVector [ALL TreeWalker])\n(transform [TreeWalker number? even?] #(* 2 %) [:a 1 [2 [[[3]]] :e] [4 5 [6 7]]])\n;; =&gt; [:a 1 [4 [[[3]]] :e] [8 5 [12 7]]]<\/pre><\/div><p>Here's how to reverse the positions of all even numbers in a tree (with order based on a depth first search). This example uses conditional navigation instead of protocol paths to do the walk:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(def TreeValues\n  (recursive-path [] p\n    (if-path vector?\n      [ALL p]\n      STAY\n      )))\n(transform (subselect TreeValues even?)\n  reverse\n  [1 2 [3 [[4]] 5] [6 [7 8] 9 [[10]]]]\n  )\n;; =&gt; [1 10 [3 [[8]] 5] [6 [7 4] 9 [[2]]]]<\/pre><\/div><p>Specter supports ClojureScript! However, some of the differences between Clojure and ClojureScript affect how you use Specter in ClojureScript, in particular with the namespace declarations. In Clojure, you might <code>(use 'com.rpl.specter)<\/code> or say <code>(:require [com.rpl.specter :refer :all])<\/code> in your namespace declaration. But in ClojureScript, these options <a href=\"https:\/\/groups.google.com\/d\/msg\/clojurescript\/SzYK08Oduxo\/MxLUjg50gQwJ\" rel=\"nofollow\">aren't allowed<\/a>. Instead, consider using one of these options:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(:require [com.rpl.specter :as s])\n(:require [com.rpl.specter :as s :refer-macros [select transform]]) ;; add in the Specter macros that you need<\/pre><\/div><ul><li>Integrate Specter with other kinds of data structures, such as graphs. Desired navigations include: reduction in topological order, navigate to outgoing\/incoming nodes, to a subgraph (with metadata indicating how to attach external edges on transformation), to node attributes, to node values, to specific nodes.<\/li>\n<\/ul><p>Copyright 2015-2020 Red Planet Labs, Inc. Specter is licensed under Apache License v2.0.<\/p>","created_at":"2020-01-10T00:50:37+0000","updated_at":"2020-01-10T00:50:42+0000","published_by":["redplanetlabs"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/2371803?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","conj"],"is_public":false,"id":1097,"title":"Sherlock Holmes, Consulting Developer - Stuart Halloway","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=OUZZKtypink","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/OUZZKtypink?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2020-01-10T00:12:17+0000","updated_at":"2020-01-10T00:12:26+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/OUZZKtypink\/maxresdefault.jpg","http_status":"200","origin_url":"https:\/\/www.youtube.com\/watch?v=OUZZKtypink"},{"is_archived":0,"is_starred":0,"tags":["clojure","conj"],"is_public":false,"id":1096,"title":"The Hitchhikers Guide to Multiparadigm Programming - Ariel Ortiz","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=GvJm-eJ5o18","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/GvJm-eJ5o18?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2020-01-10T00:08:51+0000","updated_at":"2020-01-10T00:08:57+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/GvJm-eJ5o18\/maxresdefault.jpg","http_status":"200","origin_url":"https:\/\/www.youtube.com\/watch?v=GvJm-eJ5o18"},{"is_archived":0,"is_starred":0,"tags":["clojure","conj"],"is_public":false,"id":1095,"title":"Defeating the Four Horsemen of the Coding Apocalypse - Mark Bastian","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=jh4hMAvygjk","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/jh4hMAvygjk?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2020-01-10T00:08:25+0000","updated_at":"2020-01-10T00:08:29+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/jh4hMAvygjk\/maxresdefault.jpg","http_status":"200","origin_url":"https:\/\/www.youtube.com\/watch?v=jh4hMAvygjk"},{"is_archived":0,"is_starred":0,"tags":["linux","arch","configuration"],"is_public":false,"id":1094,"title":"CyberShadow\/aconfmgr","url":"https:\/\/github.com\/CyberShadow\/aconfmgr","content":"<p><code>aconfmgr<\/code> is a package to track, manage, and restore the configuration of an Arch Linux system. Its goals are:<\/p>\n<ul><li>Quickly configure a new system, or restore an existing system according to a saved configuration<\/li>\n<li>Track temporary\/undesired changes to the system's configuration<\/li>\n<li>Identify obsolete packages and maintain a lean system<\/li>\n<\/ul><p><code>aconfmgr<\/code> tracks the list of installed packages (both native and external), as well as changes to configuration files (<code>\/etc\/<\/code>). Since the system configuration is described as shell scripts, it is trivially extensible.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/fe5837cb8be337af8536cb4ad44b9c7880f8f448\/68747470733a2f2f64756d702e7468656379626572736861646f772e6e65742f36306366306233316434663939396666363539366165356361366138366436312f636f6d706f7365642e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/fe5837cb8be337af8536cb4ad44b9c7880f8f448\/68747470733a2f2f64756d702e7468656379626572736861646f772e6e65742f36306366306233316434663939396666363539366165356361366138366436312f636f6d706f7365642e706e67\" alt=\"screenshot\" data-canonical-src=\"https:\/\/dump.thecybershadow.net\/60cf0b31d4f999ff6596ae5ca6a86d61\/composed.png\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h2>Table of Contents<\/h2>\n<ul><li><a href=\"#usage\">Usage<\/a>\n<ul><li><a href=\"#installation\">Installation<\/a><\/li>\n<li><a href=\"#first-run\">First run<\/a><\/li>\n<li><a href=\"#maintenance\">Maintenance<\/a><\/li>\n<li><a href=\"#restoring\">Restoring<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#modus-operandi\">Modus operandi<\/a>\n<ul><li><a href=\"#packages\">Packages<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#advanced-usage\">Advanced Usage<\/a>\n<ul><li><a href=\"#configuration-syntax\">Configuration syntax<\/a><\/li>\n<li><a href=\"#ignoring-some-changes\">Ignoring some changes<\/a>\n<ul><li><a href=\"#ignoring-files\">Ignoring files<\/a><\/li>\n<li><a href=\"#ignoring-packages\">Ignoring packages<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#inlining-files\">Inlining files<\/a>\n<ul><li><a href=\"#inlining-file-content-entirely\">Inlining file content entirely<\/a><\/li>\n<li><a href=\"#inlining-file-edits\">Inlining file edits<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#managing-multiple-systems\">Managing multiple systems<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparisons\">Comparisons<\/a>\n<ul><li><a href=\"#aconfmgr-vs-puppetansible\">aconfmgr vs. Puppet\/Ansible<\/a><\/li>\n<li><a href=\"#aconfmgr-vs-nixos\">aconfmgr vs. NixOS<\/a><\/li>\n<li><a href=\"#aconfmgr-vs-lostfiles\">aconfmgr vs. lostfiles<\/a><\/li>\n<li><a href=\"#aconfmgr-vs-etckeeper\">aconfmgr vs. etckeeper<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#limitations\">Limitations<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2>Usage<\/h2>\n<h3>Installation<\/h3>\n<p>Simply clone (or <a href=\"https:\/\/github.com\/CyberShadow\/aconfmgr\/archive\/master.zip\">download<\/a>+unzip) the GitHub repository. <code>aconfmgr<\/code> will install dependencies as needed during execution. <a href=\"https:\/\/aur.archlinux.org\/packages\/aconfmgr-git\/\" rel=\"nofollow\">An AUR package is also available<\/a>.<\/p>\n<h3>First run<\/h3>\n<p>Run <code>aconfmgr save<\/code> to transcribe the system's configuration to the configuration directory. This will create the file <code>99-unsorted.sh<\/code> in the configuration directory, as well as other files describing the system configuration. (The configuration directory will usually be <code>~\/.config\/aconfmgr<\/code>, or <code>.\/config<\/code> if running directly from git, or it can be overridden with <code>-c<\/code>.)<\/p>\n<p>On the first run, <code>aconfmgr<\/code> will likely include some files which you may not want to include in your system configuration. These can be temporary or auto-generated files which are not directly owned by a package. To prevent <code>aconfmgr<\/code> from including these files in the configuration, create e.g. <code>10-ignores.sh<\/code> in the configuration directory, with the lines e.g. <code>IgnorePath '\/path\/to\/file.ext'<\/code> or <code>IgnorePath '\/path\/to\/dir\/*'<\/code>. (See <a href=\"#ignoring-files\">ignoring files<\/a> for details.) Delete everything from the configuration directory except that file and re-run <code>aconfmgr save<\/code> to regenerate a configuration minding these ignore rules.<\/p>\n<p>Once <code>aconfmgr save<\/code> finishes, you should review the contents of <code>99-unsorted.sh<\/code>, and sort it into one or more new files (e.g.: <code>10-base.sh<\/code>, <code>20-drivers.sh<\/code>, <code>30-gui.sh<\/code>, <code>50-misc.sh<\/code> ...). The files should have a <code>.sh<\/code> extension, and use <code>bash<\/code> syntax. I suggest adding a comment for each package describing why installing the package was needed, so it is clear when the package is no longer needed and can be removed.<\/p>\n<p>During this process, you may identify packages or system changes which are no longer needed. Do not sort them into your configuration files - instead, delete the file <code>99-unsorted.sh<\/code>, and run <code>aconfmgr apply<\/code>. This will synchronize the system state against your configuration, thus removing the omitted packages. (You will be given a chance to confirm all changes before they are applied.)<\/p>\n<p>Note: you don't need to run <code>aconfmgr<\/code> via <code>sudo<\/code>. It will elevate as necessary by invoking <code>sudo<\/code> itself.<\/p>\n<h3>Maintenance<\/h3>\n<p>The configuration directory should be versioned using a version control system (e.g. Git). Ideally, the file <code>99-unsorted.sh<\/code> should not be versioned - it will only be created when the current configuration does not reflect the current system state, therefore indicating that there are system changes that have not been accounted for.<\/p>\n<p>Periodic maintenance consists of running <code>aconfmgr save<\/code>; if this results in uncommitted changes to the configuration directory, then there are unaccounted system changes. The changes should be reviewed, sorted, documented, committed and pushed.<\/p>\n<h3>Restoring<\/h3>\n<p>To restore a system to its earlier state, or to set up a new system, simply make sure the correct configuration is in the configuration directory and run <code>aconfmgr apply<\/code>. You will be able to preview and confirm any actual system changes.<\/p>\n<h2>Modus operandi<\/h2>\n<p>The <code>aconfmgr<\/code> script has two subcommands:<\/p>\n<ul><li><code>aconfmgr save<\/code> calculates the difference between the current system's configuration and the configuration described by the configuration directory, and writes it back to the configuration directory.<\/li>\n<li><code>aconfmgr apply<\/code> applies the difference between the configuration described by the configuration directory and the current system's configuration, installing\/removing packages and creating\/editing configuration files.<\/li>\n<\/ul><p>The configuration directory contains shell scripts, initially generated by the <code>save<\/code> subcommand, and then usually edited by the user. Evaluating these scripts will <em>compile<\/em> a system configuration description in the <code>output<\/code> directory. The difference between that directory's contents, and the actual current system configuration, dictates the actions ultimately taken by <code>aconfmgr<\/code>.<\/p>\n<p><code>aconfmgr save<\/code> will write the difference to the file <code>99-unsorted.sh<\/code> (under the configuration directory) as a series of shell commands which attempt to bring the configuration up to date with the current system. When starting with an empty configuration, this difference will consist of the entire system description. Since the script only appends to that file, it may end up undoing configuration changes done earlier in the scripts (e.g. removing packages from the package list). It is up to the user to refactor the configuration to remove redundancies, document changes, and improve maintainability.<\/p>\n<p><code>aconfmgr apply<\/code> will apply the differences to the actual system.<\/p>\n<p>The contracts of both commands are that they are mutually idempotent: after a successful invocation of either, invoking either command immediately after will be a no-op.<\/p>\n<h3>Packages<\/h3>\n<p>Background: On Arch Linux, every installed package is installed either explicitly, or as a dependency for another package. Packages can also have mandatory (hard) or optional dependencies. You can view this information using <code>pacman -Qi &lt;package&gt;<\/code> (\"Install Reason\", \"Depends On\", \"Optional Deps\").<\/p>\n<p><code>aconfmgr<\/code> only tracks explicitly-installed packages, ignoring their hard dependencies. Therefore:<\/p>\n<ul><li><code>aconfmgr save<\/code> will only save installed packages that are marked as explicitly installed.<\/li>\n<li>Installed packages that are neither explicitly installed, nor are hard dependencies of other installed packages, are considered prunable orphans and will be removed.<\/li>\n<li>Packages that are only optional dependencies of other packages must be listed explicitly, otherwise they will be pruned.<\/li>\n<li><code>aconfmgr apply<\/code> removes unlisted packages by unpinning them (setting their install reason as \"installed as a dependency\"), after which it prunes all orphan packages. If the package is still required by another package, it will remain on the system (until it is no longer required); otherwise, it is removed.<\/li>\n<li>Packages that are installed and explicitly listed in the configuration will have their install reason set to \"explicitly installed\".<\/li>\n<\/ul><h2>Advanced Usage<\/h2>\n<h3>Configuration syntax<\/h3>\n<p>The configuration files use <code>bash<\/code> syntax. The easiest way to learn the syntax is to run <code>aconfmgr save<\/code> and examine its output (<code>99-unsorted.sh<\/code>).<\/p>\n<p>Some simple helper functions are defined in <code>src\/helpers.bash<\/code> (sourced automatically). You are encouraged to examine their implementation - their main goal is not so much to provide an API as simply to make the generated configuration terser and more readable. As such, their use is in no way required, and they can be substituted with their underlying implementations.<\/p>\n<p>The list of provided helper functions:<\/p>\n<ul><li><code>AddPackage [--foreign] PACKAGE...<\/code> - Adds a package to the list of packages to be installed.<\/li>\n<li><code>RemovePackage [--foreign] PACKAGE...<\/code> - Removes an earlier-added package to the list of packages to be installed.<\/li>\n<li><code>IgnorePackage [--foreign] PACKAGE...<\/code> - Adds a package to the list of <a href=\"#ignoring-packages\">packages to be ignored<\/a>.<\/li>\n<li><code>AddPackageGroup GROUP<\/code> - Adds all packages currently in the given group to the list of packages to be installed.<\/li>\n<li><code>CopyFile PATH [MODE [OWNER [GROUP]]]<\/code> - Copies a file from the <code>files<\/code> subdirectory to the output.<\/li>\n<li><code>CopyFileTo SRC-PATH DST-PATH [MODE [OWNER [GROUP]]]<\/code> - As above, but allows source and output paths to vary.<\/li>\n<li><code>CreateFile PATH [MODE [OWNER [GROUP]]]<\/code> - Creates an empty file, to be included in the output. Prints its absolute path to standard output.<\/li>\n<li><code>GetPackageOriginalFile PACKAGE PATH<\/code> - Extracts the original file from a package's archive for inclusion in the output. Prints its absolute path to standard output.<\/li>\n<li><code>CreateLink PATH TARGET [OWNER [GROUP]]<\/code> - Creates a symbolic link with the specified target.<\/li>\n<li><code>RemoveFile PATH<\/code> - Removes an earlier-added file.<\/li>\n<li><code>SetFileProperty PATH TYPE VALUE<\/code> - Sets a file property.<\/li>\n<li><code>IgnorePath PATTERN<\/code> - Adds the specified pattern to the list of <a href=\"#ignoring-files\">ignored paths<\/a>.<\/li>\n<\/ul><p>You are free to define your own custom helper functions as part of your configuration, e.g. by placing them in a <code>00-helpers.sh<\/code> file.<\/p>\n<p>Some examples can be found on <a href=\"https:\/\/github.com\/CyberShadow\/aconfmgr\/wiki\">the project wiki<\/a>.<\/p>\n<h3>Ignoring some changes<\/h3>\n<h4>Ignoring files<\/h4>\n<p>Some files will inevitably neither belong to or match any installed packages, nor can be considered part of the system configuration. This can include:<\/p>\n<ul><li>Temporary \/ cache \/ auto-generated \/ lock \/ pipe \/ pid \/ timestamp \/ database \/ backup \/ log files<\/li>\n<li>Files managed by third-party package managers, esp. programming languages' package managers (pip, gem, npm)<\/li>\n<li>Virtual machine disk images<\/li>\n<\/ul><p>Other files may not be desirable to include in the managed system configuration because they are security-sensitive (e.g. sshd private keys).<\/p>\n<p>To declare a group of files to be ignored by <code>aconfmgr<\/code>, you can use the provided <code>IgnorePath<\/code> function, e.g.:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>IgnorePath '\/var\/lib\/pacman\/local\/*' # package metadata\nIgnorePath '\/var\/lib\/pacman\/sync\/*.db' # repos\nIgnorePath '\/var\/lib\/pacman\/sync\/*.db.sig' # repo sigs<\/pre><\/div>\n<p>Make sure to quote the pattern argument to prevent globbing at configuration parse time.<\/p>\n<h4>Ignoring packages<\/h4>\n<p>To ignore the presence of some packages on the system, you can use the <code>IgnorePackage<\/code> function:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>IgnorePackage linux-git<\/pre><\/div>\n<p><code>aconfmgr save<\/code> will not update the configuration based on ignored packages' presence or absence, and <code>aconfmgr apply<\/code> will not install or uninstall them. The packages should also not be present in the configuration's package list, of course. To ignore a foreign package (e.g. a non-AUR foreign package), use the <code>--foreign<\/code> switch (e.g. <code>IgnorePackage --foreign my-flobulator<\/code>).<\/p>\n<h3>Inlining files<\/h3>\n<p>In the output generated by <code>aconfmgr save<\/code>, non-empty new or modified files are copied in their entirety to the configuration directory, under the <code>\/files\/<\/code> subdirectory, accompanied by a <code>Copyfile<\/code> invocation. However, this can sometimes be wasteful or unwieldy if the file consists of a single line, or has one line changed out of a thousand. This section describes some techniques to move the edits into the configuration files entirely.<\/p>\n<h4>Inlining file content entirely<\/h4>\n<p>Instead of copying, the file's can be inlined using a <code>bash<\/code> heredoc, or even an <code>echo<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre># Enable Magic SysRq\necho \"kernel.sysrq = 1\" &gt; \"$(CreateFile \/etc\/sysctl.d\/99-sysrq.conf)\"\n# https:\/\/wiki.archlinux.org\/index.php\/Getty#Have_boot_messages_stay_on_tty1\ncat &gt; \"$(CreateFile \/etc\/systemd\/system\/getty@tty1.service.d\/noclear.conf)\" &lt;&lt;EOF\n[Service]\nTTYVTDisallocate=no\nEOF<\/pre><\/div>\n<h4>Inlining file edits<\/h4>\n<p>It is also possible to generate an output file that is a modification of the original. For this purpose, the helper function <code>GetPackageOriginalFile<\/code> is provided. The function will extract the indicated file from the archive in pacman's package cache, downloading it first if necessary.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre># Append some options to systemd's system.conf\ncat &gt;&gt; \"$(GetPackageOriginalFile systemd \/etc\/systemd\/system.conf)\" &lt;&lt;EOF\nRuntimeWatchdogSec=10min\nShutdownWatchdogSec=10min\nDefaultTimeoutStartSec=30s\nDefaultTimeoutStopSec=30s\nEOF\n# Specify locales\nf=\"$(GetPackageOriginalFile glibc \/etc\/locale.gen)\"\nsed -i 's\/^#\\(en_US.UTF-8\\)\/\\1\/g' \"$f\"\nsed -i 's\/^#\\(en_DK.UTF-8\\)\/\\1\/g' \"$f\" # for ISO timestamps<\/pre><\/div>\n<p>The <a href=\"http:\/\/augeas.net\/\" rel=\"nofollow\">Augeas<\/a> package can assist in editing configuration files:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>AconfNeedProgram augtool augeas n # Install augeas now, if needed\naug() { augtool --root=\"$output_dir\/files\" \"$@\" ; }\n# DNS\nGetPackageOriginalFile filesystem \/etc\/resolv.conf &gt; \/dev\/null\naug set '\/files\/etc\/resolv.conf\/nameserver[101]' 127.0.0.1 # dnsmasq\naug set '\/files\/etc\/resolv.conf\/nameserver[102]' 8.8.8.8 # Google\naug set '\/files\/etc\/resolv.conf\/nameserver[103]' 8.8.4.4 # Google<\/pre><\/div>\n<h3>Managing multiple systems<\/h3>\n<p>You can use the same configuration repository to manage multiple sufficiently-similar systems. One way of doing so is e.g. Git branches (having one main branch plus one branch per machine, and periodically merge in changes from the main branch into the machine-specific branches); however, it is simpler to use shell scripting:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>AddPackage coreutils\n# ... more common packages ...\nif [[ \"$HOSTNAME\" == \"home.example.com\" ]]\nthen\n        AddPackage nvidia\n        AddPackage nvidia-utils\n        # ... more packages only for the home system ...\nfi<\/pre><\/div>\n<p>To manage different versions of a file for different systems, you can <a href=\"#inlining-files\">inline the file<\/a>, or copy a different version depending on the hostname.<\/p>\n<p>An example for inlining <code>\/etc\/hosts<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>f=\"$(GetPackageOriginalFile filesystem \/etc\/hosts)\"\necho '1.2.3.4 home.example.net' &gt;&gt; \"$f\"\nif [[ \"$HOSTNAME\" == laptop.example.net ]]\nthen\n        echo '127.0.1.1 laptop.example.net laptop' &gt;&gt; \"$f\"\nfi<\/pre><\/div>\n<p>If the file is binary, or too large to make inlining practical, <code>CopyFileTo<\/code> can be used:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>CopyFileTo \"\/etc\/hosts-$HOSTNAME\" \"\/etc\/hosts\"<\/pre><\/div>\n<p>This will use <code>files\/etc\/hosts-home.example.net<\/code> for the <code>home.example.net<\/code> machine, and so on.<\/p>\n<h2>Comparisons<\/h2>\n<h3>aconfmgr vs. Puppet\/Ansible<\/h3>\n<p>Although <code>aconfmgr<\/code> calls itself a configuration manager, it has a number of core distinctions from the more well-known ones. One big distinction is that <code>aconfmgr<\/code> manages only one system - the one it is running on. It does not depend on any background services or network agents to work.<\/p>\n<p>Another big distinction is the scope, and the direction of the flow of information. To clarify:<\/p>\n<ul><li>\n<p>Puppet and Ansible have limited discovery abilities, whereas <code>aconfmgr<\/code> attempts to discover and save the system configuration in its entirety. There is no Puppet\/Ansible command that can be run to save a running system's configuration to a file which, when executed, will produce a similarly-configured system; however, this is the goal of <code>aconfmgr save<\/code>.<\/p>\n<\/li>\n<li>\n<p>Puppet and Ansible manage the system's configuration insofar as it is defined by the configuration file, whereas <code>aconfmgr<\/code> manages the entire system; as such, the absence of an item in the configuration file indicates its absence on the system. This is not true for <a href=\"http:\/\/www.puppetcookbook.com\/posts\/remove-package.html\" rel=\"nofollow\">Puppet<\/a> or <a href=\"http:\/\/stackoverflow.com\/questions\/29914253\/remove-package-ansible-playbook\" rel=\"nofollow\">Ansible<\/a>, where, to remove a package, one must first push a configuration file that explicitly indicates that the package is to be removed, only after which can all mentions of the package be removed from the configuration file.<\/p>\n<\/li>\n<\/ul><h3>aconfmgr vs. NixOS<\/h3>\n<p>There are some similarities between the <a href=\"https:\/\/nixos.org\/nixos\/manual\/\" rel=\"nofollow\">NixOS configuration file<\/a> and <code>aconfmgr<\/code>: they both attempt to describe the entire system configuration from a text file, with any changes in the configuration reflecting on the system. However, while <code>NixOS<\/code> forbids directly editing files under its control, <code>aconfmgr<\/code> doesn't. As with Puppet\/Ansible, <code>aconfmgr<\/code> differs in that it provides a mechanism to transcribe changes in system state back to the configuration, making it idempotent.<\/p>\n<p>Another difference is that <code>NixOS<\/code> provides a specialized syntax for many common configuration settings of managed software (e.g. allowing syntax such as <code>boot.loader.grub.device = \"\/dev\/sda\";<\/code>). <code>aconfmgr<\/code> doesn't provide this directly, though this can be achieved to some extent using tools such as <a href=\"#inlining-file-edits\">Augeas<\/a>, without also sacrificing the extent of possible configuration to only <a href=\"https:\/\/nixos.org\/nixos\/manual\/options.html\" rel=\"nofollow\">the predefined set of available options<\/a>.<\/p>\n<h3>aconfmgr vs. lostfiles<\/h3>\n<p><a href=\"https:\/\/github.com\/graysky2\/lostfiles\">lostfiles<\/a> is a \"simple script that identifies files not owned and not created by any Arch Linux package\". <code>aconfmgr<\/code> provides a superset of its functionality, not least the ability to save exclusions to a configuration file.<\/p>\n<h3>aconfmgr vs. etckeeper<\/h3>\n<p><a href=\"https:\/\/joeyh.name\/code\/etckeeper\" rel=\"nofollow\">etckeeper<\/a> allows storing <code>\/etc<\/code> in a version control system. <code>aconfmgr<\/code> allows this as well, although it does not directly provide a way to automatically merge configuration files with upstream package versions. This can be done manually, by <a href=\"#inlining-files\">inlining file changes<\/a>.<\/p>\n<h2>Limitations<\/h2>\n<ul><li>Dependencies where more than one package provides something (e.g. <code>fcron<\/code> and <code>cronie<\/code> provide <code>cron<\/code>) are not tracked, and the desired dependency must be pinned or added to the configuration manually.<\/li>\n<li>Installing AUR packages that depend on virtual packages (such as <code>java-environment<\/code>) is not currently implemented. The desired dependency can be manually specified in the configuration, or a supported AUR helper can be used instead.<\/li>\n<\/ul><h2>License<\/h2>\n<p>Copyright (c) 2016-2018 aconfmgr authors.<\/p>\n<p>Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and\/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:<\/p>\n<p>The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.<\/p>\n<p>THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.<\/p>","created_at":"2020-01-09T14:03:49+0000","updated_at":"2020-02-27T13:06:43+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/160894?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["lisp","bash"],"is_public":false,"id":1093,"title":"chr15m\/flk","url":"https:\/\/github.com\/chr15m\/flk\/","content":"<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/chr15m\/flk\/blob\/master\/docs\/wordmark.svg?sanitize=true\"><img src=\"https:\/\/github.com\/chr15m\/flk\/raw\/master\/docs\/wordmark.svg?sanitize=true\" alt=\"Fleck wordmark\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><br \/>Fleck is a Clojure-like LISP that runs wherever Bash is.<\/p><p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/chr15m\/flk\/blob\/master\/docs\/screencast.svg?sanitize=true\"><img src=\"https:\/\/github.com\/chr15m\/flk\/raw\/master\/docs\/screencast.svg?sanitize=true\" alt=\"Fleck screencast\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><div class=\"highlight highlight-source-shell\"><pre>curl -s https:\/\/chr15m.github.io\/flk\/flk &gt; flk &amp;&amp; chmod 755 flk\n.\/flk<\/pre><\/div><h3><a id=\"user-content-examples--reference--faq--make-a-lisp\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples--reference--faq--make-a-lisp\"><\/a><a href=\"https:\/\/github.com\/chr15m\/flk\/blob\/master\/examples\">Examples<\/a> | <a href=\"#reference\">Reference<\/a> | <a href=\"#faq\">FAQ<\/a> | <a href=\"https:\/\/github.com\/kanaka\/mal\">make-a-lisp<\/a><\/h3><pre>$ echo '(println \"Hello world!\") (println \"Hostname:\" (sh* \"hostname\")))' &gt; example.clj\n$ .\/flk example.clj\nHello world!\nHostname: diziet\n<\/pre><p>Now you can use a humble LISP to do Bash things. Bash as a scripting language has many edges, but it is everywhere. Fleck attempts to round off the edges.<\/p><p>Fleck runs on Bash 4 and higher.<\/p><p>Almost all of this code is from the <a href=\"https:\/\/github.com\/kanaka\/mal\/\">make-a-LISP<\/a> project. All I've done is put together a simple Makefile to package it up into an easily deployable single-file bash script.<\/p><p>A list of variables, macros and functions that are present in Fleck.<\/p><h2><a id=\"user-content-built-ins\" class=\"anchor\" aria-hidden=\"true\" href=\"#built-ins\"><\/a>Built-ins<\/h2><p>This is the set of built-ins from the make-a-lisp project. These more or less work but are generally more limited in functionality than their Clojure equivalents. For example the addition function <code>(+)<\/code> can only add two integers at a time.<\/p><p><code>def!<\/code> | <code>defmacro!<\/code> | <code>if<\/code> | <code>do<\/code> | <code>fn*<\/code> | <code>try*<\/code> | <code>sh*<\/code> | <code>let*<\/code> | <code>quote<\/code> | <code>quasiquote<\/code> | <code>macroexpand<\/code> | <code>type<\/code> | <code>=<\/code> | <code>throw<\/code> | <code>nil?<\/code> | <code>true?<\/code> | <code>false?<\/code> | <code>string?<\/code> | <code>symbol<\/code> | <code>symbol?<\/code> | <code>keyword<\/code> | <code>keyword?<\/code> | <code>number?<\/code> | <code>fn?<\/code> | <code>macro?<\/code> | <code>pr-str<\/code> | <code>str<\/code> | <code>prn<\/code> | <code>println<\/code> | <code>readline<\/code> | <code>read-string<\/code> | <code>slurp<\/code> | <code>&lt;<\/code> | <code>&lt;=<\/code> | <code>&gt;<\/code> | <code>&gt;=<\/code> | <code>+<\/code> | <code>-<\/code> | <code>*<\/code> | <code>\/<\/code> | <code>time-ms<\/code> | <code>list<\/code> | <code>list?<\/code> | <code>vector<\/code> | <code>vector?<\/code> | <code>hash-map<\/code> | <code>map?<\/code> | <code>assoc<\/code> | <code>dissoc<\/code> | <code>get<\/code> | <code>contains?<\/code> | <code>keys<\/code> | <code>vals<\/code> | <code>sequential?<\/code> | <code>cons<\/code> | <code>concat<\/code> | <code>nth<\/code> | <code>first<\/code> | <code>rest<\/code> | <code>empty?<\/code> | <code>count<\/code> | <code>apply<\/code> | <code>map<\/code> | <code>conj<\/code> | <code>seq<\/code> | <code>with-meta<\/code> | <code>meta<\/code> | <code>atom<\/code> | <code>atom?<\/code> | <code>deref<\/code> | <code>reset!<\/code> | <code>swap!<\/code><\/p><ul><li><code>*ARGV*<\/code> - list of arguments passed on the command line.<\/li>\n<\/ul><h2><a id=\"user-content-aliases\" class=\"anchor\" aria-hidden=\"true\" href=\"#aliases\"><\/a>Aliases<\/h2><p>These are wrappers around the limited make-a-lisp versions and are much more limited than the Clojure equivalents.<\/p><p><code>let<\/code> | <code>when<\/code> | <code>def<\/code> | <code>fn<\/code> | <code>defn<\/code><\/p><h2><a id=\"user-content-mal-extras\" class=\"anchor\" aria-hidden=\"true\" href=\"#mal-extras\"><\/a>Mal extras<\/h2><p>These functions are pulled from a selection of <code>mal\/lib\/*.mal<\/code>.<\/p><p><code>partial<\/code> | <code>inc<\/code> | <code>dec<\/code> | <code>zero<\/code> | <code>identity<\/code> | <code>reduce<\/code> | <code>foldr<\/code><\/p><h2><a id=\"user-content-fleck-extras\" class=\"anchor\" aria-hidden=\"true\" href=\"#fleck-extras\"><\/a>Fleck extras<\/h2><p>These functions are hand crafted Fleck specials designed to make common shell scripting tasks easier.<\/p><ul><li><code>(str-replace STRING FIND REPLACE)<\/code> - Replace all occurrences of the string <code>FIND<\/code> in <code>STRING<\/code> with the string <code>REPLACE<\/code>.<\/li>\n<li><code>(str-split STRING SPLIT-CHARACTER)<\/code> - Split <code>STRING<\/code> into a list of strings on the single characters <code>SPLIT-CHARACTER<\/code>.<\/li>\n<li><code>(dc OPERATOR ARRAY-OF-NUMBERS)<\/code> - Wraps the <code>dc<\/code> command to do decimal math. E.g. <code>(dc '+ [1 2 3])<\/code> yeilds <code>6<\/code>.<\/li>\n<li><code>(env [KEY] [VALUE])<\/code> - Returns a <code>hash-map<\/code> of environment variables. Returns the value of <code>KEY<\/code> if present. Sets the value of <code>KEY<\/code> to <code>VAL<\/code> if the latter is present.<\/li>\n<\/ul><h2><a id=\"user-content-interop\" class=\"anchor\" aria-hidden=\"true\" href=\"#interop\"><\/a>Interop<\/h2><ul><li><code>(sh* COMMAND)<\/code> - Run arbitrary bash strings and return the stdout result.<\/li>\n<li><code>(env [KEY] [VALUE])<\/code> - See above section.<\/li>\n<\/ul><p>For examples of writing your own Fleck functions in Bash see <a href=\"https:\/\/github.com\/chr15m\/flk\/blob\/master\/src\/extras.sh\">src\/extras.sh<\/a>. Functions should set the special return value <code>r<\/code> and use Fleck type casting functions like <code>_string<\/code> to wrap the result in a reference. Internal Fleck functions such as <code>_string<\/code> automatically do this and can be used bare. Use <code>_fref<\/code> to make your function available to the Fleck namespace e.g. <code>_fref \"my-bash-function\" _my_bash_function<\/code>.<\/p><p>You can make a pure bash script from your Fleck script by bundling your script and Fleck together into a new script.<\/p><p>Say you have a Fleck script called <code>wow.clj<\/code>, you can bundle it as follows:<\/p><pre>make DEST=wow INSERT=.\/wow.clj NOREPL=1\n<\/pre><p>This will produce a new standalone script called <code>wow<\/code> with Fleck + <code>wow.clj<\/code> bundled together.<\/p><p>When you run <code>wow<\/code> the embedded <code>wow.clj<\/code> will be run by the embedded Fleck.<\/p><p>Think of this as homoiconic Bash rather than Clojure, and code as if you're in Bash.<\/p><h3><a id=\"user-content-will-my-favourite-piece-of-clojure-run-in-this\" class=\"anchor\" aria-hidden=\"true\" href=\"#will-my-favourite-piece-of-clojure-run-in-this\"><\/a>Will my favourite piece of Clojure run in this?<\/h3><p>No, it's bash.<\/p><p>Some subset of Clojure-like code will run. See the <a href=\"#reference\">documentation<\/a> and <a href=\"https:\/\/github.com\/chr15m\/flk\/blob\/master\/examples\">examples<\/a>.<\/p><h3><a id=\"user-content-how-do-i-access-command-line-arguments\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-access-command-line-arguments\"><\/a>How do I access command line arguments?<\/h3><p>Use the special global list <code>*ARGV*<\/code>.<\/p><h3><a id=\"user-content-how-do-i-access-and-modify-environment-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-access-and-modify-environment-variables\"><\/a>How do I access and modify environment variables?<\/h3><p>Check the <a href=\"#fleck-extras\"><code>(env)<\/code> function above<\/a>. See also <a href=\"https:\/\/github.com\/chr15m\/flk\/blob\/master\/examples\/environment-variables.clj\">examples\/environment-variables.clj<\/a>.<\/p><h3><a id=\"user-content-how-can-i-execute-a-one-liner-of-fleck-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-can-i-execute-a-one-liner-of-fleck-code\"><\/a>How can I execute a one-liner of Fleck code?<\/h3><p>Either of these methods will work:<\/p><div class=\"highlight highlight-source-shell\"><pre>flk &lt;&lt;&lt; '(println \"hi\")'\necho '(println \"hi\")' | flk<\/pre><\/div><h3><a id=\"user-content-why-cant-i-add-more-than-2-numbers-together\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-cant-i-add-more-than-2-numbers-together\"><\/a>Why can't I add more than 2 numbers together?<\/h3><p>It's bash. Try the <code>dc<\/code> function: <code>(dc '+ [1 2 3 4])<\/code><\/p><h3><a id=\"user-content-where-are-the-floating-point-numbers\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-are-the-floating-point-numbers\"><\/a>Where are the floating point numbers?<\/h3><p>It's bash. Try the <code>dc<\/code> function for decimals: <code>(dc '* [8.2 3.5])<\/code><\/p><p><code>dc<\/code> is set to keep four fractional digits in its results.<\/p><h3><a id=\"user-content-how-do-i-cast-a-string-to-a-number\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-cast-a-string-to-a-number\"><\/a>How do I cast a string to a number?<\/h3><p>Try <code>(read-str \"42\")<\/code> but also Bash doesn't care and <code>(+ \"1\" 1)<\/code> will yeild <code>2<\/code>.<\/p><h3><a id=\"user-content-why-cant-i-iterate-on-a-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-cant-i-iterate-on-a-string\"><\/a>Why can't I iterate on a string?<\/h3><p>Try <code>(seq \"somestring\")<\/code>.<\/p><h3><a id=\"user-content-how-do-i-do-destructuring\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-do-destructuring\"><\/a>How do I do destructuring?<\/h3><p>You can't.<\/p><h3><a id=\"user-content-how-do-i-use-a-keyhash-map-as-a-function-in-lookups\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-use-a-keyhash-map-as-a-function-in-lookups\"><\/a>How do I use a key\/<code>hash-map<\/code> as a function in lookups?<\/h3><p>You can't. You'll get an error with something like <code>(:a {:a 12})<\/code> or <code>({:a 12} :a)<\/code>.<\/p><p>Instead you must use <code>get<\/code> like this: <code>(get {:a 12} :a)<\/code>.<\/p><h3><a id=\"user-content-can-i-use-anything-as-a-hash-map-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#can-i-use-anything-as-a-hash-map-key\"><\/a>Can I use anything as a <code>hash-map<\/code> key?<\/h3><p>Seems unlikely. Better stick to strings.<\/p><h3><a id=\"user-content-this-is-even-slower-than-python\" class=\"anchor\" aria-hidden=\"true\" href=\"#this-is-even-slower-than-python\"><\/a>This is even slower than Python!<\/h3><p>Yes.<\/p><p>PS That is not actually a question.<\/p><h3><a id=\"user-content-havent-i-seen-this-before-somewhere\" class=\"anchor\" aria-hidden=\"true\" href=\"#havent-i-seen-this-before-somewhere\"><\/a>Haven't I seen this before somewhere?<\/h3><p>You're probably thinking of <a href=\"https:\/\/github.com\/alandipert\/gherkin\/\">Gherkin, the original Clojure-like LISP in Bash by Alan Dipert<\/a>. Gherkin helped kick off the <a href=\"https:\/\/github.com\/kanaka\/mal\/\">make-a-lisp<\/a> revolution.<\/p><h3><a id=\"user-content-why-is-it-called-fleck\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-is-it-called-fleck\"><\/a>Why is it called Fleck?<\/h3><p>At <code>36k<\/code> and running on any machine with Bash 4, the name seemed appropriate.<\/p><pre> fleck\n    n. A tiny mark or spot.\n    n. A small bit or flake.\n<\/pre>","created_at":"2020-01-08T23:11:39+0000","updated_at":"2020-01-09T14:04:19+0000","published_by":["chr15m"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/repository-images.githubusercontent.com\/224614194\/5d268580-120a-11ea-8db1-f3e7996e675c","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1092,"title":"pipx \u2014 Install and Run Python Applications in Isolated Environments","url":"https:\/\/pipxproject.github.io\/pipx\/","content":"<a href=\"https:\/\/github.com\/pipxproject\/pipx\/edit\/master\/docs\/index.md\" title=\"Edit this page\" class=\"md-icon md-content__icon\">\ue3c9<\/a><p style=\"text-align: center;\"><a href=\"https:\/\/pipxproject.github.io\/pipx\/\"><img src=\"https:\/\/github.com\/pipxproject\/pipx\/raw\/master\/logo.png\" referrerpolicy=\"no-referrer\" alt=\"image\" style=\"text-align: center;\" \/><\/a><\/p><p style=\"text-align: center;\"><a href=\"https:\/\/github.com\/pipxproject\/pipx\/raw\/master\/pipx_demo.gif\"><img src=\"https:\/\/github.com\/pipxproject\/pipx\/raw\/master\/pipx_demo.gif\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/p><p style=\"text-align: center;\"><a href=\"https:\/\/travis-ci.org\/pipxproject\/pipx\"><img src=\"https:\/\/travis-ci.org\/pipxproject\/pipx.svg?branch=master\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/pypi.python.org\/pypi\/pipx\/\"><img src=\"https:\/\/img.shields.io\/badge\/pypi-0.15.1.2-blue.svg\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a> <a href=\"https:\/\/github.com\/ambv\/black\"><img alt=\"Code style: black\" src=\"https:\/\/img.shields.io\/badge\/code%20style-black-000000.svg\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p><strong>Documentation<\/strong>: https:\/\/pipxproject.github.io\/pipx\/<\/p><p><strong>Source Code<\/strong>: https:\/\/github.com\/pipxproject\/pipx<\/p><p><em>For comparison to other tools including pipsi, see <a href=\"https:\/\/pipxproject.github.io\/pipx\/comparisons\/\">Comparison to Other Tools<\/a>.<\/em><\/p><h2 id=\"install-pipx\">Install pipx<\/h2><p>On macOS:<\/p><pre>brew install pipx\npipx ensurepath\n<\/pre><p>Upgrade pipx with <code>brew update &amp;&amp; brew upgrade pipx<\/code>.<\/p><p>Otherwise, install via pip:<\/p><pre>python3 -m pip install --user pipx\npython3 -m pipx ensurepath\n<\/pre><p>Upgrade pipx with <code>python3 -m pip install -U pipx<\/code>.<\/p><p>Shell completions are available by following the instructions printed with this command:<\/p><pre>pipx completions\n<\/pre><p>For more details, see the <a href=\"https:\/\/pipxproject.github.io\/pipx\/installation\/\">installation instructions<\/a>.<\/p><h2 id=\"overview-what-is-pipx\">Overview: What is <code>pipx<\/code>?<\/h2><p>pipx is a tool to help you install and run end-user applications written in Python. pipx <strong>is not<\/strong> a tool for development or publishing of your code -- it's only for consuming already published packages.<\/p><p>If you are familiar macOS's <code>brew<\/code>, it's kind of like that but for Python applications. If you are familiar with JavaScript's <a href=\"https:\/\/medium.com\/@maybekatz\/introducing-npx-an-npm-package-runner-55f7d4bd282b\">npx<\/a>, it's kind of like that but also allows you to install packages. If you are familiar with <code>pip<\/code>, it's kind of like that but is only focused on installing packages that have entry points.<\/p><p>pipx does not ship with pip, but installing it is often an important part of bootstrapping your system.<\/p><p>Python and PyPI allow developers to distribute code with \"console script entry points\". These scripts let users call into Python code from the command line, effectively acting like standalone applications. <code>pipx<\/code> is a tool to install and run any of these thousands of Python applications available on PyPI in a safe, convenient, and reliable way. <strong>In a way, it turns Python Package Index (PyPI) into a big app store for Python applications.<\/strong> Not all Python packages have entry points, but many do.<\/p><p><code>pipx<\/code> enables you to:<\/p><ul><li>Safely install packages to isolated environments, while globally exposing their CLI entry points so you can run them from anywhere (see the <code>install<\/code> command). This guarantees no dependency conflicts and clean uninstalls!<\/li>\n<li>Easily list, upgrade, and uninstall packages that were installed with pipx<\/li>\n<li>Run the latest version of a Python application in a temporary environment (see the <code>run<\/code> command)<\/li>\n<\/ul><p>Best of all, pipx runs with regular user permissions, never calling <code>sudo pip install<\/code> (you aren't doing that, are you? \ud83d\ude04).<\/p><h3 id=\"walkthrough-installing-a-package-and-its-applications-with-pipx\">Walkthrough: Installing a Package and its Applications With <code>pipx<\/code><\/h3><p>You can globally install an application by running<\/p><pre>pipx install PACKAGE\n<\/pre><p>This automatically creates a virtual environment, installs the package, and adds the package's associated applications (entry points) to a location on your <code>PATH<\/code>. For example, <code>pipx install pycowsay<\/code> makes the <code>pycowsay<\/code> command available globally, but sandboxes the pycowsay package in its own virtual environment. <strong>pipx never needs to run as sudo to do this.<\/strong><\/p><p>Example:<\/p><pre>&gt;&gt; pipx install pycowsay\n  installed package pycowsay 2.0.3, Python 3.7.3\n  These apps are now globally available\n    - pycowsay\ndone! \u2728 \ud83c\udf1f \u2728\n&gt;&gt; pipx list\nvenvs are in \/home\/user\/.local\/pipx\/venvs\napps are exposed on your $PATH at \/home\/user\/.local\/bin\n   package pycowsay 2.0.3, Python 3.7.3\n    - pycowsay\n# Now you can run pycowsay from anywhere\n&gt;&gt; pycowsay mooo\n  ____\n&lt; mooo &gt;\n  ====\n         \\\n          \\\n            ^__^\n            (oo)\\_______\n            (__)\\       )\\\/\\\n                ||----w |\n                ||     ||\n<\/pre><h3 id=\"walkthrough-running-an-application-in-a-temporary-sandboxed-environment\">Walkthrough: Running an Application in a Temporary, Sandboxed Environment<\/h3><p>I find this handy when I need to run the latest version of an app, but don't necessarily need it installed on my computer.<\/p><p>You may want to do this when you are initializing a new project and want to set up the right directory structure, when you want to view the help text of an application, or if you simply want to run an app in a one-off case and and leave your system untouched afterwards.<\/p><p>For example, the blog post <a href=\"https:\/\/sourcery.ai\/blog\/python-best-practices\/\">How to set up a perfect Python project<\/a> uses <code>pipx run<\/code> to kickstart a new project with <a href=\"https:\/\/github.com\/cookiecutter\/cookiecutter\">cookiecutter<\/a>, a tool that creates projects from project templates.<\/p><p>A nice side benefit is that you don't have to remember to upgrade the app since <code>pipx run<\/code> will automatically run a recent version for you.<\/p><p>Okay, let's see what this looks like in practice!<\/p><pre>pipx run APP [ARGS...]\n<\/pre><p>This will install the package in an isolated, temporary directory and invoke the app. Give it a try:<\/p><pre>&gt; pipx run pycowsay moo\n  ---\n&lt; moo &gt;\n  ---\n   \\   ^__^\n    \\  (oo)\\_______\n       (__)\\       )\\\/\\\n           ||----w |\n           ||     ||\n<\/pre><p>Notice that you <strong>don't need to execute any install commands to run the app<\/strong>.<\/p><p>Any arguments after the application name will be passed directly to the application:<\/p><pre>&gt; pipx run pycowsay these arguments are all passed to pycowsay!\n  -------------------------------------------\n&lt; these arguments are all passed to pycowsay! &gt;\n  -------------------------------------------\n   \\   ^__^\n    \\  (oo)\\_______\n       (__)\\       )\\\/\\\n           ||----w |\n           ||     ||\n<\/pre><p>Re-running the same app is quick because pipx caches Virtual Environments on a per-app basis. The caches only last a few days, and when they expire, pipx will again use the latest version of the package. This way you can be sure you're always running a new version of the package without having to manually upgrade.<\/p><p>You can run .py files directly, too.<\/p><pre>pipx run https:\/\/gist.githubusercontent.com\/cs01\/fa721a17a326e551ede048c5088f9e0f\/raw\/6bdfbb6e9c1132b1c38fdd2f195d4a24c540c324\/pipx-demo.py\npipx is working!\n<\/pre><h3 id=\"summary\">Summary<\/h3><p>That's it! Those are the most important commands <code>pipx<\/code> offers. To see all of pipx's documentation, run <code>pipx --help<\/code> or see the <a href=\"https:\/\/pipxproject.github.io\/pipx\/docs\/\">docs<\/a>.<\/p><h2 id=\"testimonials\">Testimonials<\/h2><div>\"Thanks for improving the workflow that pipsi has covered in the past. Nicely done!\"<p>\u2014<a href=\"https:\/\/twitter.com\/jezdez\">Jannis Leidel<\/a>, PSF fellow, former pip and Django core developer, and founder of the Python Packaging Authority (PyPA)<\/p><\/div><div>\"My setup pieces together pyenv, poetry, and pipx. [...] For the things I need, it\u2019s perfect.\"<p>\u2014<a href=\"https:\/\/twitter.com\/mkennedy\">Jacob Moss<\/a>, co-creator of Django in blog post <a href=\"https:\/\/jacobian.org\/2019\/nov\/11\/python-environment-2020\/\">My Python Development Environment, 2020 Edition<\/a><\/p><\/div><div>\"I'm a big fan of pipx. I think pipx is super cool.\"<p>\u2014<a href=\"https:\/\/twitter.com\/mkennedy\">Michael Kennedy<\/a>, co-host of PythonBytes podcast in <a href=\"https:\/\/pythonbytes.fm\/episodes\/transcript\/139\/f-yes-for-the-f-strings\">episode 139<\/a><\/p><\/div><h2 id=\"credits\">Credits<\/h2><p>pipx was inspired by <a href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> and <a href=\"https:\/\/github.com\/zkat\/npx\">npx<\/a>. It was created by <a href=\"https:\/\/github.com\/cs01\/\">Chad Smith<\/a> and has had lots of help from <a href=\"https:\/\/github.com\/pipxproject\/pipx\/graphs\/contributors\">contributors<\/a>. The logo was created by <a href=\"https:\/\/github.com\/IrishMorales\">@IrishMorales<\/a>.<\/p><p>pipx is maintained by a team of volunteers (in alphabetical order)<\/p><ul><li><a href=\"https:\/\/github.com\/gaborbernat\">Bern\u00e1t G\u00e1bor<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/cs01\">Chad Smith<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/aiguofer\">Diego Fernandez<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/itsayellow\">Matthew Clapp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/sahensley\">Shawn Hensley<\/a><\/li>\n<\/ul><h2 id=\"contributing\">Contributing<\/h2><p>Issues and Pull Requests are definitely welcome! Check out <a href=\"https:\/\/pipxproject.github.io\/pipx\/contributing\/\">Contributing<\/a> to get started.<\/p>","created_at":"2020-01-08T23:07:34+0000","updated_at":"2020-01-08T23:07:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"pipxproject.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1091,"title":"How to Build a Small Form Factor Gaming PC That Can Fit Anywhere","url":"https:\/\/theinventory-com.cdn.ampproject.org\/v\/s\/theinventory.com\/how-to-build-a-small-form-factor-gaming-pc-that-can-fit-1837032102\/amp?usqp=mq331AQCKAE=&amp_js_v=0.1","content":"<p>Commerce Content is independent of Editorial and Advertising, and if you buy something through our posts, we may get a small share of the sale. <a href=\"https:\/\/theinventory.com\/welcome-to-the-inventory-1825927863\" target=\"_top\" title=\"About The Inventory and Kinja Deals\">Click here<\/a> for more.<\/p><div class=\"post-content\"><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3RqW0Gs3--\/c_scale,f_auto,fl_progressive,q_80,w_800\/nmg9luobnwkgbybltgx4.jpg\" data-chomp-id=\"nmg9luobnwkgbybltgx4\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3RqW0Gs3--\/c_scale,f_auto,fl_progressive,q_80,w_800\/nmg9luobnwkgbybltgx4.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3RqW0Gs3--\/c_scale,f_auto,fl_progressive,q_80,w_800\/nmg9luobnwkgbybltgx4.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3RqW0Gs3--\/c_scale,f_auto,fl_progressive,q_80,w_800\/nmg9luobnwkgbybltgx4.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Small form factor cases come in all shapes and sizes, but take up significantly less space than a traditional mid tower.\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>When you think of a \u201cgaming PC,\u201d chances are you picture a tricked-out desktop tower the size of a large sock drawer. But why waste that much space? With a Mini-ITX motherboard and the right case, you can build a tiny PC that fits just about anywhere, without sacrificing on power.<br \/><\/p><p>\u201cSmall form factor PCs\u201d have grown in popularity the past few years, even spawning their own tight-knit communities rallied around building the smallest possible machines. These communities generally define small form factor as having a volume of 20 liters or less, and while you can\u2019t fit every configuration into that small a space, you can build most consumer-level gaming PCs in this form factor without too much trouble. I\u2019ve built quite a few over the years, and after building the most recent iteration of my living room gaming PC, I felt the undeniable urge to evangelize.<\/p><p>We\u2019re not going to detail the basics of building a PC in this guide\u2014I\u2019ve <a href=\"https:\/\/lifehacker.com\/how-to-build-a-computer-the-complete-guide-5828747\" rel=\"noopener noreferrer\" target=\"_blank\">done that before over at Lifehacker<\/a>, and while a bit old, most of those tips are still relevant today. Instead, we\u2019re going to focus on what makes small form factor PCs different, what to look for when you buy your parts, and a few tips for making things go smoothly.<\/p><aside class=\"referenced-wide referenced-fullwidth branded-item branded-item--lifehacker clearfix\"><div class=\"container referenced-item\"><div class=\"referenced-item-image referenced-item-processed\"><a class=\"inset--story__thumb\" href=\"https:\/\/lifehacker.com\/how-to-build-a-computer-the-complete-guide-5828747\" target=\"_blank\">\n<figure><img class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"77\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--UVrFzXQ1--\/c_fill,f_auto,fl_progressive,g_center,h_77,pg_1,q_80,w_137\/zdpg8ivxuaub1xsx4qkt.png\" width=\"137\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><\/a><\/div><div class=\"referenced-item-text\"><p class=\"text small-offset-4 medium-offset-3\">Building a computer from scratch gives you the perfect machine for your needs, but it can be\u2026<\/p><a class=\"readmore-referenced inset--story__readmore\" href=\"https:\/\/lifehacker.com\/how-to-build-a-computer-the-complete-guide-5828747\" target=\"_blank\">Read more Read more<\/a><\/div><\/div>\n<\/aside><aside class=\"pullquote align--center\">The Case<\/aside><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--X-5XeOPd--\/c_scale,f_auto,fl_progressive,q_80,w_800\/c7j9fcwyacu7uxm8v8fj.jpg\" data-chomp-id=\"c7j9fcwyacu7uxm8v8fj\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--X-5XeOPd--\/c_scale,f_auto,fl_progressive,q_80,w_800\/c7j9fcwyacu7uxm8v8fj.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--X-5XeOPd--\/c_scale,f_auto,fl_progressive,q_80,w_800\/c7j9fcwyacu7uxm8v8fj.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--X-5XeOPd--\/c_scale,f_auto,fl_progressive,q_80,w_800\/c7j9fcwyacu7uxm8v8fj.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Flat cases like the <a href=\"https:\/\/amazon.com\/Fractal-Design-FD-CA-NODE-202-BK-Case\/dp\/B015MKDOPU?tag=theinventory-20&amp;ascsubtag=0a378c4dfb3996c36c0974714256c0eb7ac0b3b3\" target=\"_top\">Node 202<\/a> are perfect for hiding in an entertainment unit...as long as your hardware doesn\u2019t get too hot!\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>Picking a case is, by far, the most difficult, important, and enjoyable part of building a pint-sized PC. You have a ton of variety to choose from, not just in style, but in size shape. There are \u201ccube\u201d cases like the <a href=\"https:\/\/amazon.com\/Thermaltake-Computer-Chassis-Interchangeable-CA-1B8-00S1WN-00\/dp\/B00M2UKGSM?tag=theinventory-20&amp;ascsubtag=e59494cc3457d1bfef86e8787910f630cf0bbd11\" target=\"_top\">Thermaltake Core V1<\/a> and <a href=\"https:\/\/amazon.com\/SilverStone-Technology-Mini-ITX-Computer-SST-SG13B-USA\/dp\/B07MNC3JCB?tag=theinventory-20&amp;ascsubtag=197524cb039aaa5b0af016d6c4e57ea8a00dc0af\" target=\"_top\">Silverstone SG13<\/a>, which are great for putting on a desk. There are \u201cflat\u201d cases that can lay in an entertainment center or stand up on end, like the <a href=\"https:\/\/amazon.com\/Fractal-Design-FD-CA-NODE-202-BK-Case\/dp\/B015MKDOPU?tag=theinventory-20&amp;ascsubtag=0a378c4dfb3996c36c0974714256c0eb7ac0b3b3\" target=\"_top\">Fractal Design Node 202<\/a>, <a href=\"https:\/\/amazon.com\/SilverStone-Technology-Evolution-Mini-ITX-Computer\/dp\/B01ECDWI3M?tag=theinventory-20&amp;ascsubtag=fbc26cc8517ac3e243e5cb642ed5e88f5ab234d5\" target=\"_top\">Silverstone RVZ01<\/a>, or <a href=\"https:\/\/amazon.com\/SilverStone-Technology-Mini-ITX-Computer-ML08B-H\/dp\/B0161UXWRA?tag=theinventory-20&amp;ascsubtag=494777a459d4ee415e878f64e51cb4d8bb56d2e1\" target=\"_top\">Silverstone ML08<\/a>. And there are mini desktop cases like the <a href=\"https:\/\/amazon.com\/Cougar-gaming-QBX-Case\/dp\/B00ZV5YKY6?tag=theinventory-20&amp;ascsubtag=4fe54cce787fdd6539b158e858689a2c0d2e0f1b\" target=\"_top\">Cougar QBX<\/a>. All of these are significantly smaller than a typical mid tower case, while still being spacious enough that building in them isn\u2019t <em>too<\/em> hard\u2014Fractal Design graciously lent me a Node 202 to use for this particular guide, and I was shocked at how easy it was to build in compared to other Mini-ITX cases I\u2019ve used in the past.<br \/><\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$52<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/SilverStone-Technology-Mini-ITX-Computer-SST-SG13B-USA\/dp\/B07MNC3JCB?tag=theinventory-20&amp;ascsubtag=fe159f26f19b1cae9fdf4f1628bc9b4ee55753a9\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--A2dRPisQ--\/c_fit,fl_progressive,h_166,q_80,w_295\/tducltz3g1nndnamfqwz.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>2Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/SilverStone-Technology-Mini-ITX-Computer-SST-SG13B-USA\/dp\/B07MNC3JCB?tag=theinventory-20&amp;ascsubtag=fe159f26f19b1cae9fdf4f1628bc9b4ee55753a9\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$80<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Fractal-Design-FD-CA-NODE-202-BK-Case\/dp\/B015MKDOPU?tag=theinventory-20&amp;ascsubtag=8f60201433fd278bff08d18a5579fda97d21caf1\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--Tjo54xOC--\/c_fit,fl_progressive,h_166,q_80,w_295\/ionq2ohkrcrbok4zyrin.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>17Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Fractal-Design-FD-CA-NODE-202-BK-Case\/dp\/B015MKDOPU?tag=theinventory-20&amp;ascsubtag=8f60201433fd278bff08d18a5579fda97d21caf1\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$117<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/SilverStone-Technology-Evolution-Mini-ITX-Computer\/dp\/B01ECDWI3M?tag=theinventory-20&amp;ascsubtag=6693d624cb7de52d8e33df13e256dfe0d319608e\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--PCGDQ7Iu--\/c_fit,fl_progressive,h_166,q_80,w_295\/cerckbqlpfpuxohczuku.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/SilverStone-Technology-Evolution-Mini-ITX-Computer\/dp\/B01ECDWI3M?tag=theinventory-20&amp;ascsubtag=6693d624cb7de52d8e33df13e256dfe0d319608e\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$61<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Cougar-gaming-QBX-Case\/dp\/B00ZV5YKY6?tag=theinventory-20&amp;ascsubtag=3c69950e464581cf455a9d7f9f8af0b25155c6bd\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3GOglZc9--\/c_fit,fl_progressive,h_166,q_80,w_295\/p01vxxs9poviyq1ijm3p.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>1Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Cougar-gaming-QBX-Case\/dp\/B00ZV5YKY6?tag=theinventory-20&amp;ascsubtag=3c69950e464581cf455a9d7f9f8af0b25155c6bd\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><p>If you really want to go tiny, though\u2014which requires a bit more perseverance and money\u2014boutique manufacturers offer some of the best sub-10L cases on the market, like the venerable <a href=\"http:\/\/www.ncases.com\/\" rel=\"noopener noreferrer\" target=\"_blank\">NCase M1<\/a>, <a href=\"https:\/\/amazon.com\/LOUQE-Ghost-S1-Ash\/dp\/B07FSKXG4F?tag=theinventory-20&amp;ascsubtag=42738d648f850bccca5028f3b5199a63a75e4b6f\" target=\"_top\">Ghost S1<\/a>, and <a href=\"https:\/\/www.dan-cases.com\/dana4.php\" rel=\"noopener noreferrer\" target=\"_blank\">Dan Case A4-SFX<\/a>\u2014not to mention the impossibly small <a href=\"https:\/\/nfc-systems.com\/skyreach-4-mini\" rel=\"noopener noreferrer\" target=\"_blank\">NFC SkyReach 4 Mini<\/a>, which is barely larger than a PS4. And frankly, this list just scratches the surface of what\u2019s available.<\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\"><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><p>From amazon<\/p><p>Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/LOUQE-Ghost-S1-Ash\/dp\/B07FSKXG4F?tag=theinventory-20&amp;ascsubtag=42e8b0773af8357ba4f42349f4151bd523713b8a\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><p>Each of these cases has their own unique benefits and challenges, especially when it comes to cooling (which we\u2019ll talk about in a moment). Pick a couple cases you like the look of, and poke around forums like <a href=\"https:\/\/www.reddit.com\/r\/sffpc\" rel=\"noopener noreferrer\" target=\"_blank\">\/r\/sffpc<\/a> and <a href=\"https:\/\/smallformfactor.net\/\" rel=\"noopener noreferrer\" target=\"_blank\">SFF.network<\/a> to see what you\u2019re in for. Ultimately, it\u2019s all about buying what you think looks good and what fits in your space, just know that the smaller you go, the more patient you\u2019ll have to be with the build. I\u2019ve been building PCs for almost a decade and these Mini-ITX projects still trip me up sometimes.<\/p><aside class=\"pullquote align--center\">The Motherboard, CPU, RAM, and Storage<\/aside><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--ceXPdAJf--\/c_scale,f_auto,fl_progressive,q_80,w_800\/x92ean1j9uqq39z0lolw.jpg\" data-chomp-id=\"x92ean1j9uqq39z0lolw\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--ceXPdAJf--\/c_scale,f_auto,fl_progressive,q_80,w_800\/x92ean1j9uqq39z0lolw.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--ceXPdAJf--\/c_scale,f_auto,fl_progressive,q_80,w_800\/x92ean1j9uqq39z0lolw.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--ceXPdAJf--\/c_scale,f_auto,fl_progressive,q_80,w_800\/x92ean1j9uqq39z0lolw.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Mini-ITX doesn\u2019t offer much in the way of expansion, but you get the essentials in a very tiny package.\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>Picking the brains of your PC is actually the easiest part of the process, since it\u2019s not <em>that<\/em> different from doing so on a larger PC. If you\u2019re going balls to the wall with your processor and graphics card, you may want to do some research about what kind of heat your particular case can handle, but if you\u2019re building a somewhat reasonable machine, you shouldn\u2019t need to worry too much\u2014buy the CPU that meets your needs.<\/p><p>I went with an <a href=\"https:\/\/amazon.com\/Intel-i5-9400F-Desktop-Processor-Graphics\/dp\/B07MRCGQQ4?tag=theinventory-20&amp;ascsubtag=0fc635746c402550c905d879030672de733b4580\" target=\"_top\">Intel i5-9400F<\/a> since I\u2019m mostly using this PC for casual gaming and VR, but you could go as high as an <a href=\"https:\/\/amazon.com\/Intel-i5-9600K-Desktop-Processor-Unlocked\/dp\/B07HHLX1R8?tag=theinventory-20&amp;ascsubtag=27c52edc0f581aeb887a3edd37e5a5eab615b93a\" target=\"_top\">i5-9600K<\/a> or <a href=\"https:\/\/amazon.com\/AMD-Ryzen-3600-12-Thread-Processor\/dp\/B07STGGQ18?tag=theinventory-20&amp;ascsubtag=7938f2665fd193312238a9a13fa2c90809057be9\" target=\"_top\">AMD Ryzen 5 3600<\/a> in most small form factor cases without issue. If you\u2019re building more of a home theater or emulation machine and don\u2019t need a dedicated GPU, an APU with integrated graphics like the <a href=\"https:\/\/amazon.com\/AMD-Ryzen-3200G-Unlocked-Processor\/dp\/B07STGHZK8?tag=theinventory-20&amp;ascsubtag=4cb4e7a885c6e848c6066a9cdb08a9e1b7e916bc\" target=\"_top\">Ryzen 3 3200G<\/a> would be perfect.<br \/><\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$159<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Intel-i5-9400F-Desktop-Processor-Graphics\/dp\/B07MRCGQQ4?tag=theinventory-20&amp;ascsubtag=7fca336ce64b552ac8b5369380829e206c4b4c5b\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--VRbLO5uz--\/c_fit,fl_progressive,h_166,q_80,w_295\/lmmz3pg6gi09z0wxrpuv.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>11Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Intel-i5-9400F-Desktop-Processor-Graphics\/dp\/B07MRCGQQ4?tag=theinventory-20&amp;ascsubtag=7fca336ce64b552ac8b5369380829e206c4b4c5b\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$207<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Intel-i5-9600K-Desktop-Processor-Unlocked\/dp\/B07HHLX1R8?tag=theinventory-20&amp;ascsubtag=345370eac229d31013472527eca5a5f89193a696\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--Vf6NWEwQ--\/c_fit,fl_progressive,h_166,q_80,w_295\/hvwjd11vy2pokg8l3ewx.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>31Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Intel-i5-9600K-Desktop-Processor-Unlocked\/dp\/B07HHLX1R8?tag=theinventory-20&amp;ascsubtag=345370eac229d31013472527eca5a5f89193a696\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$193<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/AMD-Ryzen-3600-12-Thread-Processor\/dp\/B07STGGQ18?tag=theinventory-20&amp;ascsubtag=6c43de2da9153b0d8e9bae9b9e8c05e3458000bf\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--na6RVAvI--\/c_fit,fl_progressive,h_166,q_80,w_295\/rg9drwhfzbnnxs9grh4j.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>20Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/AMD-Ryzen-3600-12-Thread-Processor\/dp\/B07STGGQ18?tag=theinventory-20&amp;ascsubtag=6c43de2da9153b0d8e9bae9b9e8c05e3458000bf\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$95<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/AMD-Ryzen-3200G-Unlocked-Processor\/dp\/B07STGHZK8?tag=theinventory-20&amp;ascsubtag=2b8462f6c34a3543aaa89f002caa60b0a8ab28c9\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--2NI5naEy--\/c_fit,fl_progressive,h_166,q_80,w_295\/vorhucl63vck3djbrs0e.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>3Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/AMD-Ryzen-3200G-Unlocked-Processor\/dp\/B07STGHZK8?tag=theinventory-20&amp;ascsubtag=2b8462f6c34a3543aaa89f002caa60b0a8ab28c9\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><p>Motherboard and RAM is similar\u2014buy what works for your needs and is compatible with your CPU. I went with an <a href=\"https:\/\/amazon.com\/ASRock-H370M-ITX-AC-Mini-ITX-Motherboard\/dp\/B07BMV3X6N?tag=theinventory-20&amp;ascsubtag=23d666ca8313326459ae853bc47626b10bb2d7ce\" target=\"_top\">ASRock H370M-ITX\/AC<\/a> to go with my i5, and a stick of <a href=\"https:\/\/amazon.com\/Crucial-Ballistix-Desktop-Gaming-BLS8G4D240FSBK\/dp\/B01M7STFAJ?tag=theinventory-20&amp;ascsubtag=ec2eee6c62c6315a8a19d875b8848d97d8809f7e\" target=\"_top\">Crucial Ballistix Sport RAM<\/a>. Just make sure you don\u2019t get blinged-out memory with huge heatsinks, since they might cause clearance problems with certain CPU coolers. For storage, I recommend going with <a href=\"https:\/\/theinventory.com\/m-2-sata-and-nvme-ssds-explained-which-should-you-bu-1836647003\" target=\"_top\">an M.2 drive<\/a> like the <a href=\"https:\/\/amazon.com\/Crucial-MX500-NAND-2280SS-Internal\/dp\/B0784SY515?tag=theinventory-20&amp;ascsubtag=a7e864b5abab56080edd6373b5eaba5576516bca\" target=\"_top\">Crucial MX500<\/a> or <a href=\"https:\/\/amazon.com\/Samsung-860-SATA-Internal-MZ-N6E1T0BW\/dp\/B07822Z77M?tag=theinventory-20&amp;ascsubtag=b7572e18ace26df38de9f329e95320233b937338\" target=\"_top\">Samsung 860 EVO<\/a>, since they attach directly to the motherboard and take up no extra space. You could go with a <a href=\"https:\/\/amazon.com\/Samsung-Inch-Internal-MZ-76E1T0B-AM\/dp\/B078DPCY3T?tag=theinventory-20&amp;ascsubtag=e979564c0fc5925c37df4b1d0d74d02c74b3472a\" target=\"_top\">regular 2.5-inch drive<\/a>, and many of the above cases do have mounting points for them, but that requires routing more cables. Plus, some cases may limit the GPU or other parts you can use if you use their SSD mounts. If you <em>have<\/em> to have a spinning hard drive for cost or space reasons, definitely go with a <a href=\"https:\/\/amazon.com\/Toshiba-HDWL110XZSTA-Laptop-Internal-Height\/dp\/B07CSDQ5RX?tag=theinventory-20&amp;ascsubtag=cb609b166ef556a91b4f8ef4c171e7a111e0d1e7\" target=\"_top\">2.5-inch laptop drive<\/a> rather than a 3.5-inch behemoth.<\/p><aside class=\"referenced-wide referenced-fullwidth branded-item branded-item--theinventory clearfix\"><div class=\"container referenced-item\"><div class=\"referenced-item-image referenced-item-processed\"><a class=\"inset--story__thumb\" href=\"https:\/\/theinventory.com\/m-2-sata-and-nvme-ssds-explained-which-should-you-bu-1836647003\" target=\"_blank\">\n<figure><img class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"77\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--5ep9wbwA--\/c_fill,f_auto,fl_progressive,g_center,h_77,pg_1,q_80,w_137\/i8jktmdmxfmgsii3rzct.jpg\" width=\"137\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><\/a><\/div><div class=\"referenced-item-text\"><p class=\"text small-offset-4 medium-offset-3\">SSDs are one of the best upgrades that mainstream computing has seen in quite some time\u2014they\u2019ll\u2026<\/p><a class=\"readmore-referenced inset--story__readmore\" href=\"https:\/\/theinventory.com\/m-2-sata-and-nvme-ssds-explained-which-should-you-bu-1836647003\" target=\"_blank\">Read more Read more<\/a><\/div><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$110<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Crucial-MX500-NAND-2280SS-Internal\/dp\/B0784SY515?tag=theinventory-20&amp;ascsubtag=7dc86a7a16b5713169a040554a034fd38707f76d\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--xJyYyBBN--\/c_fit,fl_progressive,h_166,q_80,w_295\/yd5z8dccyysfekixds1y.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>28Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Crucial-MX500-NAND-2280SS-Internal\/dp\/B0784SY515?tag=theinventory-20&amp;ascsubtag=7dc86a7a16b5713169a040554a034fd38707f76d\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$155<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Samsung-860-SATA-Internal-MZ-N6E1T0BW\/dp\/B07822Z77M?tag=theinventory-20&amp;ascsubtag=2a4c129ce54ea6bf9cf7cd7b5bce069fe5c76dc7\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--Prlm9CZg--\/c_fit,fl_progressive,h_166,q_80,w_295\/svwnyd8mypqfk3vqrzex.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>37Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Samsung-860-SATA-Internal-MZ-N6E1T0BW\/dp\/B07822Z77M?tag=theinventory-20&amp;ascsubtag=2a4c129ce54ea6bf9cf7cd7b5bce069fe5c76dc7\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"pullquote align--center\">The Graphics Card<\/aside><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lTF-SzLE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/znmvx6p19k0zpgbxnhrz.jpg\" data-chomp-id=\"znmvx6p19k0zpgbxnhrz\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lTF-SzLE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/znmvx6p19k0zpgbxnhrz.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lTF-SzLE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/znmvx6p19k0zpgbxnhrz.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lTF-SzLE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/znmvx6p19k0zpgbxnhrz.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Many small form factor builds use PCIe risers to put your graphics card in a separate chamber.\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>If you\u2019re aiming to do some real PC gaming, you\u2019ll probably want a dedicated graphics card in your mini machine\u2014but graphics cards are much bigger than they used to be. So check your case\u2019s spec sheet and see the maximum graphics card length it supports. Plenty of Mini-ITX cases can fit \u201cfull size\u201d dual-slot graphics cards, but some may require you to remove drive cages, fans, or other features in order to do so. In those situations, \u201cmini\u201d graphics cards are a great alternative. Honestly, even if you have the extra space, mini cards provide you a bit more room for building and cable management, which can be nice.<br \/><\/p><p>The bigger consideration is what style of cooling your card uses. GPUs tend to come in two main cooling flavors: <a href=\"https:\/\/amazon.com\/ASUS-GeForce-Turbo-GDDR6-RTX2060-6G\/dp\/B07MSK1H93?tag=theinventory-20&amp;ascsubtag=1eb63446846ac6d6010393b0b0c1ae5ae4696d0b\" target=\"_top\">\u201cblower\u201d style coolers<\/a>, which use a closed shroud to push air out the back of the case, and <a href=\"https:\/\/amazon.com\/MSI-GeForce-RTX-2060-VENTUS\/dp\/B07MC23VS4?tag=theinventory-20&amp;ascsubtag=bc5557b03b43f0dbf299341b64440b5c566fde91\" target=\"_top\">\u201copen\u201d coolers<\/a>, which tend to have less noise but also recirculate hot air within your case. Cooling is especially crucial when you\u2019re cramming a bunch of hardware into a tiny case, which is why many small form-factor enthusiasts recommend blower style cards\u2014they ensure hot air gets exhausted out the back instead of blown around inside.<\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$340<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/ASUS-GeForce-Turbo-GDDR6-RTX2060-6G\/dp\/B07MSK1H93?tag=theinventory-20&amp;ascsubtag=6c0361251121025775fddeb0175cfd75d08df67e\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--hyaVOUQR--\/c_fit,fl_progressive,h_166,q_80,w_295\/aqsjsif7v8sp7rfr1r1x.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>5Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/ASUS-GeForce-Turbo-GDDR6-RTX2060-6G\/dp\/B07MSK1H93?tag=theinventory-20&amp;ascsubtag=6c0361251121025775fddeb0175cfd75d08df67e\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$360<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/MSI-GeForce-RTX-2060-VENTUS\/dp\/B07MC23VS4?tag=theinventory-20&amp;ascsubtag=356044599414a0e17fa29e97fb692080024b72ff\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--LAzrLdjs--\/c_fit,fl_progressive,h_166,q_80,w_295\/qdjj8wia3bzlmxgg2bl0.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>10Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/MSI-GeForce-RTX-2060-VENTUS\/dp\/B07MC23VS4?tag=theinventory-20&amp;ascsubtag=356044599414a0e17fa29e97fb692080024b72ff\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\"><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><p>From amazon<\/p><p>Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/ZOTAC-GeForce-Graphics-IceStorm-ZT-T20600F-10M\/dp\/B07MBKKQPW?tag=theinventory-20&amp;ascsubtag=c762d0753ad90da98636a6b4ca4af631ad47e62e\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><p>You can, however, still go with an open air cooler if you\u2019d prefer to have less noise (or you can\u2019t get a blower-style cooler on the card you want). You just have to be a bit more careful about your cooling setup. Add some case fans near the card to suck cool air in from the outside, and choose a card with fins aligned <a href=\"https:\/\/c1.neweggimages.com\/NeweggImage\/ProductImageCompressAll1280\/14-500-458-V11.jpg\" rel=\"noopener noreferrer\" target=\"_blank\">horizontally<\/a> rather than <a href=\"https:\/\/c1.neweggimages.com\/NeweggImage\/ProductImageCompressAll1280\/14-126-284-V02.jpg\" rel=\"noopener noreferrer\" target=\"_blank\">vertically<\/a>, so some of that air gets directed out the back of the case instead of toward your other components. My Node 202 can fit a full-size card, but I chose a <a href=\"https:\/\/amazon.com\/ZOTAC-GeForce-192-bit-Graphics-Compact\/dp\/B07NMWQXLR?tag=theinventory-20&amp;ascsubtag=8966970b55000c929da709370eb356cab8e941e1\" target=\"_top\">Zotac GTX 1660 Ti<\/a> to give myself a bit more room for cable management and any <a href=\"https:\/\/www.reddit.com\/r\/Vive\/comments\/9iy0q7\/m2_to_pcie_adapter_works_with_the_vive_wireless\/\" rel=\"noopener noreferrer\" target=\"_blank\">extra peripherals<\/a> I may want to stuff in later on.<\/p><aside class=\"pullquote align--center\">The Power Supply<\/aside><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--sGkz58JA--\/c_scale,f_auto,fl_progressive,q_80,w_800\/eosw0z7wexnp539g4b2x.jpg\" data-chomp-id=\"eosw0z7wexnp539g4b2x\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--sGkz58JA--\/c_scale,f_auto,fl_progressive,q_80,w_800\/eosw0z7wexnp539g4b2x.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--sGkz58JA--\/c_scale,f_auto,fl_progressive,q_80,w_800\/eosw0z7wexnp539g4b2x.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--sGkz58JA--\/c_scale,f_auto,fl_progressive,q_80,w_800\/eosw0z7wexnp539g4b2x.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Some cases may route the power cord through the inside to save space on the rear panel.\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>Depending on the case you get, buying a power supply may be simple...or it may open up a whole new world of craziness.<br \/><\/p><p>Many cases, including the Node 202 I used in this build, support SFX power supplies, which are very similar to their ATX brethren\u2014just a bit more compact. You\u2019ll find some budget options out there, like the <a href=\"https:\/\/amazon.com\/SilverStone-Technology-Factor-BRONZE-ST30SF\/dp\/B00FA4KP8S?tag=theinventory-20&amp;ascsubtag=6703f4f8315751c9b18f128b3ec74301424590ae\" target=\"_top\">Silverstone ST30SF<\/a> and <a href=\"https:\/\/amazon.com\/SilverStone-Technology-Factor-Bronze-ST45SF-V3\/dp\/B01LZFMKCI?tag=theinventory-20&amp;ascsubtag=ec87ee408d9640045aa81fa85941779053b72a29\" target=\"_top\">ST45SF-V3<\/a>, but for a small form factor build, I can\u2019t recommend modular power supplies highly enough\u2014you\u2019ll be able to remove any cables you don\u2019t need, reducing clutter and making your build a heck of a lot easier. <a href=\"https:\/\/amazon.com\/Corsair-Platinum-Certified-Modular-Supply\/dp\/B01CGGOXWQ?tag=theinventory-20&amp;ascsubtag=42c176c3499c0934f7808b1f952078ddcdaf62c4\" target=\"_top\">Corsair\u2019s SF450<\/a> and <a href=\"https:\/\/amazon.com\/Corsair-Platinum-Certified-Modular-Supply\/dp\/B01CGI5M24?tag=theinventory-20&amp;ascsubtag=3a5bc660db4e4d9895dbbf41973dcbc8beff4298\" target=\"_top\">SF600<\/a> are good choices.<\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$73<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/SilverStone-Technology-Factor-Bronze-ST45SF-V3\/dp\/B01LZFMKCI?tag=theinventory-20&amp;ascsubtag=cabeb5f92a95a12d94b54e19ed589619e4cb5ec4\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--tgg-rjRR--\/c_fit,fl_progressive,h_166,q_80,w_295\/dwckquk7j7yxr0pspvqp.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>9Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/SilverStone-Technology-Factor-Bronze-ST45SF-V3\/dp\/B01LZFMKCI?tag=theinventory-20&amp;ascsubtag=cabeb5f92a95a12d94b54e19ed589619e4cb5ec4\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$108<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Corsair-Platinum-Certified-Modular-Supply\/dp\/B01CGGOXWQ?tag=theinventory-20&amp;ascsubtag=9363ba3cf36bceb407b10d734eb914d0d3dcc32d\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--v-eEkw6Y--\/c_fit,fl_progressive,h_166,q_80,w_295\/pgjlj9adsptzhjjdx9le.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>14Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Corsair-Platinum-Certified-Modular-Supply\/dp\/B01CGGOXWQ?tag=theinventory-20&amp;ascsubtag=9363ba3cf36bceb407b10d734eb914d0d3dcc32d\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><p>As cases get smaller, though, power supplies are one of the first things to shrink. Some may use a Flex-ATX power supply like the <a href=\"https:\/\/amazon.com\/FSP-Solution-Platinum-Efficiency-FSP500-50FSPT\/dp\/B01N4IGM0O?tag=theinventory-20&amp;ascsubtag=97afc306e4a5d3028056848eb1f4fb4c308c0074\" target=\"_top\">FSP500<\/a>, though these usually aren\u2019t modular and can be very loud, leading some folks to replace the 40mm fan inside with <a href=\"https:\/\/amazon.com\/Noctua-Cooling-Blades-Bearing-NF-A4x10\/dp\/B009NQLT0M?tag=theinventory-20&amp;ascsubtag=195c2a87e2d902d8cb71d2f6a948d23375da3fde\" target=\"_top\">something quieter<\/a>. Other folks eschew an internal power supply altogether, instead using an external laptop power brick with an <a href=\"https:\/\/www.hdplex.com\/hdplex-400w-hi-fi-dc-atx-power-supply-16v-24v-wide-range-voltage-input.html\" rel=\"noopener noreferrer\" target=\"_blank\">HDPlex<\/a> DC-to-ATX converter connected to the internal components. You can see how this gets a bit complicated\u2014if you opt for a case that requires something like this, just do your due diligence before you buy. Research what others recommend for your setup, <a href=\"https:\/\/outervision.com\/power-supply-calculator\" rel=\"noopener noreferrer\" target=\"_blank\">use a power supply calculator<\/a> to make sure you get enough wattage, and don\u2019t skimp on quality. Power supplies may be boring, but they\u2019re very important, and it\u2019s not a place to cheap out.<\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$140<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/FSP-Solution-Platinum-Efficiency-FSP500-50FSPT\/dp\/B01N4IGM0O?tag=theinventory-20&amp;ascsubtag=20bf2bf06e3f364f781fd32c5d94f7941d32d555\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--ABWitYUD--\/c_fit,fl_progressive,h_166,q_80,w_295\/qu8dpmrd6ykqwlti6ef0.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>1Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/FSP-Solution-Platinum-Efficiency-FSP500-50FSPT\/dp\/B01N4IGM0O?tag=theinventory-20&amp;ascsubtag=20bf2bf06e3f364f781fd32c5d94f7941d32d555\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"pullquote align--center\">The CPU Cooler and Case Fans<\/aside><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3kj6qlxi--\/c_scale,f_auto,fl_progressive,q_80,w_800\/du4obiwbjderyjclahyq.jpg\" data-chomp-id=\"du4obiwbjderyjclahyq\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3kj6qlxi--\/c_scale,f_auto,fl_progressive,q_80,w_800\/du4obiwbjderyjclahyq.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3kj6qlxi--\/c_scale,f_auto,fl_progressive,q_80,w_800\/du4obiwbjderyjclahyq.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3kj6qlxi--\/c_scale,f_auto,fl_progressive,q_80,w_800\/du4obiwbjderyjclahyq.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Some cases can fit larger coolers and hard drives, but it\u2019ll be a tight fit.\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>Finally, let\u2019s talk about cooling. You\u2019re stuffing an awful lot of hardware into a small space, which means you\u2019ll need to take a bit more care with your fans than you would in a normal-sized build.<br \/><\/p><p>For your CPU, the stock cooler may very well fit, but it\u2019ll be awfully loud. You\u2019re usually better off with a low-profile CPU cooler like the <a href=\"https:\/\/amazon.com\/quiet-BK002-Shadow-Rock-Cooler\/dp\/B00VGDN6F6?tag=theinventory-20&amp;ascsubtag=a92df88a5dd0acef8236d5f70aa78ff26fab4747\" target=\"_top\">Be Quiet Shadow Rock LP<\/a>, <a href=\"https:\/\/amazon.com\/Scythe-Shuriken-Profile-120mm-Cooler\/dp\/B07MDXVLCZ?tag=theinventory-20&amp;ascsubtag=854e59762b40a7c1ebd6ec832a2fda3180b8830c\" target=\"_top\">Scythe Big Shuriken 3<\/a>, or (for a super slim heatsink) the <a href=\"https:\/\/amazon.com\/Noctua-NH-L9i-Premium-Low-profile-LGA115x\/dp\/B009VCAJ7W?tag=theinventory-20&amp;ascsubtag=ce4dafab9d62c5ab6ea690405b61767ae74f5041\" target=\"_top\">Noctua NH-L9i<\/a>\/<a href=\"https:\/\/amazon.com\/Noctua-NH-L9a-Premium-Low-profile-Cooler\/dp\/B075SG1T3X?tag=theinventory-20&amp;ascsubtag=9555a2f8b32c48613c5dd0c9594eac1c4ef1be60\" target=\"_top\">L9a<\/a>, depending on the clearance you have in your specific case. There are a surprising amount of low-profile coolers on the market, so check the clearance spec for your case and shop around. Some cases might even be able to fit a small all-in-one liquid cooling loop like the <a href=\"https:\/\/amazon.com\/CORSAIR-Hydro-Liquid-Cooler-Radiator\/dp\/B079NXZQBC?tag=theinventory-20&amp;ascsubtag=30cc933de3fc09aba101738880d0e8e15f77603b\" target=\"_top\">Corsair H60<\/a>. Your CPU cooler will be very dependent on your case and your CPU\u2019s TDP, so again, it may help to see what others have done in their own builds.<\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$45<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/quiet-BK002-Shadow-Rock-Cooler\/dp\/B00VGDN6F6?tag=theinventory-20&amp;ascsubtag=958c3640e5fe4ed2d7e741cfa6d3169d76693aae\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--_7dAw_YT--\/c_fit,fl_progressive,h_166,q_80,w_295\/cyjcc9tq7rlzvdijmor4.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>8Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/quiet-BK002-Shadow-Rock-Cooler\/dp\/B00VGDN6F6?tag=theinventory-20&amp;ascsubtag=958c3640e5fe4ed2d7e741cfa6d3169d76693aae\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$47<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Scythe-Shuriken-Profile-120mm-Cooler\/dp\/B07MDXVLCZ?tag=theinventory-20&amp;ascsubtag=fbba643e16ee7ce1aa39d7c8271c90df6f1f6146\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--smU37r3S--\/c_fit,fl_progressive,h_166,q_80,w_295\/mqevz1zurhtqlbrhij0o.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>1Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Scythe-Shuriken-Profile-120mm-Cooler\/dp\/B07MDXVLCZ?tag=theinventory-20&amp;ascsubtag=fbba643e16ee7ce1aa39d7c8271c90df6f1f6146\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$40<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Noctua-NH-L9i-Premium-Low-profile-LGA115x\/dp\/B009VCAJ7W?tag=theinventory-20&amp;ascsubtag=5f03c0840f52ba6977d14954627b93896bfc9461\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--8tIN151o--\/c_fit,fl_progressive,h_166,q_80,w_295\/tvf9g5csu6bkqsnnsciu.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>44Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Noctua-NH-L9i-Premium-Low-profile-LGA115x\/dp\/B009VCAJ7W?tag=theinventory-20&amp;ascsubtag=5f03c0840f52ba6977d14954627b93896bfc9461\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$80<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/CORSAIR-Hydro-Liquid-Cooler-Radiator\/dp\/B079NXZQBC?tag=theinventory-20&amp;ascsubtag=f069070a306670a07416e99a812d04b4dc1c81bc\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lninuaMM--\/c_fit,fl_progressive,h_166,q_80,w_295\/gz1o22hyabyfegxcckhy.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>19Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/CORSAIR-Hydro-Liquid-Cooler-Radiator\/dp\/B079NXZQBC?tag=theinventory-20&amp;ascsubtag=f069070a306670a07416e99a812d04b4dc1c81bc\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><p>Finally, not all Mini-ITX cases come with fans built-in, so you may want to add a few case fans to your arsenal to aid in cooling. Standard 120mm fans like the <a href=\"https:\/\/amazon.com\/Noctua-NF-F12-chromax-Black-swap-Premium-Grade-Quiet\/dp\/B07654PNFQ?tag=theinventory-20&amp;ascsubtag=1981ceeb26467cae93cd2f598a70babbcb81bcdb\" target=\"_top\">Noctua NF-F12<\/a> may work fine in some cases, but chances are you\u2019ll want something a bit slimmer, like the <a href=\"https:\/\/amazon.com\/Noctua-NF-A12x15-PWM-Premium-Quiet\/dp\/B071W6HJP6?tag=theinventory-20&amp;ascsubtag=a086c46d9c5688336fecaf8c2fedd3a0239d72d5\" target=\"_top\">Noctua NF-A12x15<\/a> (for 120mm mounts) or <a href=\"https:\/\/amazon.com\/Noctua-NF-A9x14-Premium-Quiet-92x14mm\/dp\/B009NQM7V2?tag=theinventory-20&amp;ascsubtag=1a347753ca34624f5a9857732b91699edea891e1\" target=\"_top\">NF-A9x14<\/a> (for 92mm mounts). Look up the ideal configuration for your case and go to town.<\/p><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$23<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Noctua-NF-F12-chromax-Black-swap-Premium-Grade-Quiet\/dp\/B07654PNFQ?tag=theinventory-20&amp;ascsubtag=65025bc5de23f2764b787cdcb358322ca1c4b9b0\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--4p0c39EQ--\/c_fit,fl_progressive,h_166,q_80,w_295\/zpzc9jh2ura8moaygudh.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>33Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Noctua-NF-F12-chromax-Black-swap-Premium-Grade-Quiet\/dp\/B07654PNFQ?tag=theinventory-20&amp;ascsubtag=65025bc5de23f2764b787cdcb358322ca1c4b9b0\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$20<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/amazon.com\/Noctua-NF-A12x15-PWM-Premium-Quiet\/dp\/B071W6HJP6?tag=theinventory-20&amp;ascsubtag=9f81c8d149cf228212801c13d3825525adbb42c2\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--bh8AuV7d--\/c_fit,fl_progressive,h_166,q_80,w_295\/xdhhmnu8nz4y13ate49f.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>26Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/amazon.com\/Noctua-NF-A12x15-PWM-Premium-Quiet\/dp\/B071W6HJP6?tag=theinventory-20&amp;ascsubtag=9f81c8d149cf228212801c13d3825525adbb42c2\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><h3>Putting It All Together<\/h3><figure><img alt=\"null\" class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" data-asset-url=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--rCH00UUE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/u2ua6wurgom3ndowztup.jpg\" data-chomp-id=\"u2ua6wurgom3ndowztup\" data-format=\"jpg\" height=\"450\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--rCH00UUE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/u2ua6wurgom3ndowztup.jpg\" srcset=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w820\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--rCH00UUE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/u2ua6wurgom3ndowztup.jpg 820w, https:\/\/i-kinja--img-com.cdn.ampproject.org\/ii\/w1200\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--rCH00UUE--\/c_scale,f_auto,fl_progressive,q_80,w_800\/u2ua6wurgom3ndowztup.jpg 1200w\" width=\"800\" referrerpolicy=\"no-referrer\" \/><figcaption>Time to manage some cables.\nPhoto: Whitson Gordon\n<\/figcaption><\/figure><p>Again, we\u2019re not going to do a full-on PC building guide here (this has gotten long enough already), but after doing quite a few of these myself over the years, here are some tips I\u2019ve found helpful:<br \/><\/p><ul><li>Build and test your hardware outside the case first. Do you hear me? <em>Build outside the case first<\/em>. This has been <a href=\"https:\/\/lifehacker.com\/6-essential-tips-for-first-time-pc-builders-1458484291\" rel=\"noopener noreferrer\" target=\"_blank\">standard PC building advice for years<\/a>, but it\u2019s even more useful for small form factor builds\u2014they take longer, they\u2019re more frustrating, and if you spend all day building the thing only to find your motherboard was dead on arrival, you\u2019re going to curse yourself. Build outside your case first and make sure the machine POSTs.<\/li>\n<li>Many mini cases are split up into \u201cchambers,\u201d with one housing your motherboard, and another holding your graphics card (thanks to a PCIe riser that usually comes with the case). Take a step back and plan out your build before you start popping things in\u2014you may have to build in a different order than you\u2019re used to so everything fits properly (not to mention so you can manage your cables along the way).<\/li>\n<li>Give yourself an extra hour or two, if not more, to build. You\u2019re bound to run into some snags, and you may find yourself having to pull some stuff out just to re-build in the correct order, or to re-route your cables multiple times. In fact, you may even find you bought a combination of parts that doesn\u2019t fit in your case\u2014don\u2019t worry, it happens to the best of us\u2014and you have to wait a couple days for Amazon to ship you something else.<\/li>\n<li><a href=\"https:\/\/www.amazon.com\/TR-Industrial-TR88302-Multi-Purpose-Cable\/dp\/B01018DC96?tag=theinventory-20&amp;ascsubtag=3e96985128aa45c1dc042d8ead4ec82218c08615\" target=\"_top\">Zip ties<\/a> are your friend. If your case doesn\u2019t come with some, grab some separately to help manage your cables. Even if you don\u2019t care about how it looks, you may have to zip tie cables to certain parts of the case to keep them from interfering with spinning fans.<\/li>\n<\/ul><aside class=\"commerce-inset--v2 js_inset\"><div class=\"commerce-inset\">$7<div class=\"commerce-inset__element commerce-inset__element--image\"><a data-insetrender=\"0\" data-linktype=\"[t|mod-image\" href=\"https:\/\/www.amazon.com\/TR-Industrial-TR88302-Multi-Purpose-Cable\/dp\/B01018DC96?tag=theinventory-20&amp;ascsubtag=5fd4f8dbf00834324e2662a281f1b69860c3f42e\" target=\"_blank\"><img class=\"commerce-inset__product-image i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"166\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--3Sl7Yywp--\/c_fit,fl_progressive,h_166,q_80,w_295\/pp5qz5rkuc5vjtsp90pt.jpg\" width=\"295\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/div><div class=\"commerce-inset__content-wrapper commerce-inset__element commerce-inset__element--middle\"><header>\n<\/header><p>From amazon<\/p><p>372Gizmodo Media Group may get a commission<\/p><\/div><p><a class=\"button button--medium button--commerce\" data-insetrender=\"0\" data-linktype=\"[t|mod-button\" href=\"https:\/\/www.amazon.com\/TR-Industrial-TR88302-Multi-Purpose-Cable\/dp\/B01018DC96?tag=theinventory-20&amp;ascsubtag=5fd4f8dbf00834324e2662a281f1b69860c3f42e\" target=\"_blank\">Buy now<\/a><\/p><\/div>\n<\/aside><ul><li>Once you\u2019re finished installing Windows, be sure to monitor your temperatures closely for the first few days. Again, this is pretty good advice for building any PC, but it\u2019s especially important for these smaller builds. I like <a href=\"https:\/\/www.msi.com\/page\/afterburner\" rel=\"noopener noreferrer\" target=\"_blank\">MSI Afterburner<\/a> for monitoring CPU and GPU temperatures during games, though apps like <a href=\"https:\/\/www.hwinfo.com\/\" rel=\"noopener noreferrer\" target=\"_blank\">HWiNFO<\/a> are great on the desktop, too.<\/li>\n<\/ul><aside class=\"referenced-wide referenced-fullwidth branded-item branded-item--lifehacker clearfix\"><div class=\"container referenced-item\"><div class=\"referenced-item-image referenced-item-processed\"><a class=\"inset--story__thumb\" href=\"https:\/\/lifehacker.com\/6-essential-tips-for-first-time-pc-builders-1458484291\" target=\"_blank\">\n<figure><img class=\"i-amphtml-layout-responsive i-amphtml-layout-size-defined\" height=\"77\" src=\"https:\/\/i-kinja--img-com.cdn.ampproject.org\/i\/s\/i.kinja-img.com\/gawker-media\/image\/upload\/s--vnu3-KWM--\/c_fill,f_auto,fl_progressive,g_center,h_77,pg_1,q_80,w_137\/dbrbi5wvrlmslzghunbi.jpg\" width=\"137\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/figure><\/a><\/div><div class=\"referenced-item-text\"><p class=\"text small-offset-4 medium-offset-3\">Building a computer is the best way to get the perfect machine for your needs. We've shown you \u2026<\/p><a class=\"readmore-referenced inset--story__readmore\" href=\"https:\/\/lifehacker.com\/6-essential-tips-for-first-time-pc-builders-1458484291\" target=\"_blank\">Read more Read more<\/a><\/div><\/div>\n<\/aside><p>Believe it or not, this is the short version of this guide. Building a Mini-ITX rig is far more complex than building a normal-sized PC, but it\u2019s well worth the effort if you\u2019re short on space or just want the bragging rights of kicking the ass of computers twice its size. Do your research, choose your parts carefully, and don\u2019t get discouraged\u2014you\u2019ll be mighty proud of the final result.<\/p><hr \/><\/div>","created_at":"2020-01-08T16:50:35+0000","updated_at":"2020-01-08T16:50:35+0000","annotations":[],"mimetype":"text\/html","reading_time":13,"domain_name":"theinventory-com.cdn.ampproject.org","http_status":"200","origin_url":"https:\/\/theinventory-com.cdn.ampproject.org\/v\/s\/theinventory.com\/how-to-build-a-small-form-factor-gaming-pc-that-can-fit-1837032102\/amp?amp_js_v=a2&amp_gsa=1&usqp=mq331AQCKAE%3D#aoh=15785021532881&amp_ct=1578502175399&csi=1&referrer=https%3A%2F%2Fwww.google.com&amp_tf=From%20%251%24s&ampshare=https%3A%2F%2Ftheinventory.com%2Fhow-to-build-a-small-form-factor-gaming-pc-that-can-fit-1837032102"},{"is_archived":0,"is_starred":0,"tags":["stumpwm","configuration"],"is_public":false,"id":1090,"title":"tomoyuki28jp\/stumpwm","url":"https:\/\/github.com\/tomoyuki28jp\/stumpwm\/blob\/master\/.stumpwmrc","content":"; -*- Mode: Lisp; Mode: StumpWM; -*-\n; .stumpwmrc\n(in-package :stumpwm)\n\n(setq *startup-message* nil)\n(setq *startup-mode-line* t)\n\n(loop for file in '(\"battery\" \"notifications\")\ndo (load (make-pathname :name file :type \"lisp\" :directory\n'(:relative \".stumpwm.d\" \"contrib\"))))\n\n; --- process management ----------------------------------------\n(defun ps-exists (ps)\n(let ((f \"ps -ef | grep ~S | grep -v -e grep -e stumpish | wc -l\"))\n(&lt; 0 (parse-integer (run-shell-command (format nil f ps) t)))))\n\n(defun start-uniq-command-ps (command &amp;key options (background t))\n(unless (ps-exists command)\n(run-shell-command\n(concat command \" \" options \" \" (when background \"&amp;\")))))\n\n(defun kill-ps-command (command)\n(format nil \"kill -TERM `ps -ef | grep ~S | grep -v grep | awk '{print $2}'`\"\ncommand))\n\n(defun kill-ps (command)\n(run-shell-command (kill-ps-command command)))\n\n; --- stumpwm command definitions -------------------------------\n(defmacro def-run-or-raise-command (cmd prop)\n(let ((cmd-str (string-downcase (symbol-name cmd))))\n`(defcommand ,cmd () ()\n(run-or-raise ,cmd-str ,prop))))\n\n(def-run-or-raise-command gnome-terminal '(:class \"Gnome-terminal\"))\n;(def-run-or-raise-command mlterm '(:class \"mlterm\"))\n;(def-run-or-raise-command urxvt '(:class \"URxvt\"))\n(def-run-or-raise-command firefox '(:class \"Firefox\"))\n(def-run-or-raise-command soffice '(:class \"soffice\"))\n(def-run-or-raise-command virtualbox '(:class \"VirtualBox\"))\n\n(defcommand ps-kill (ps) ((:rest \"Process to kill: \"))\n(kill-ps ps))\n\n(defcommand emacs () ()\n(run-or-raise \"emacs\" '(:class \"Emacs\")))\n\n(defcommand nm-applet () ()\n(start-uniq-command-ps \"nm-applet\" :options \"--sm-disable\"))\n\n(defcommand show-battery () ()\n(echo-string (current-screen) (run-shell-command \"acpi\" t)))\n\n(defcommand uptime () ()\n(echo-string (current-screen) (run-shell-command \"uptime\" t)))\n\n(defcommand gnome-panel () ()\n(if (ps-exists \"gnome-panel\")\n(progn (kill-ps \"gnome-panel\")\n(sleep 0.3)\n(when *startup-mode-line*\n(run-shell-command \"stumpish mode-line\")))\n(progn (when (head-mode-line (current-head))\n(mode-line))\n(run-shell-command \"gnome-panel\"))))\n\n; --- run commands ---------------------------------------\n(run-shell-command \"xsetroot -solid black\")\n(run-shell-command \"\/opt\/atokx3\/sample\/iiimf_status_hide\")\n(nm-applet)\n(start-uniq-command-ps \"gnome-settings-daemon\")\n(start-uniq-command-ps \"gnome-power-manager\")\n(start-uniq-command-ps \"gnome-valume-manager\")\n(start-uniq-command-ps \"update-notifier\")\n(run-shell-command \"~\/src\/backup\/alert.sh\")\n(emacs)\n\n; --- sudo command definitions --------------------------------\n(define-stumpwm-type :password (input prompt)\n(let ((history *input-history*)\n(arg (argument-pop input))\n(fn (symbol-function 'draw-input-bucket)))\n(unless arg\n(unwind-protect\n(setf (symbol-function 'draw-input-bucket)\n(lambda (screen prompt input &amp;optional errorp)\n(let ((i (copy-structure input)))\n(setf (input-line-string i)\n(make-string (length (input-line-string i))\n:initial-element #\\*))\n(funcall fn screen prompt i)))\narg (read-one-line (current-screen) prompt))\n(setf (symbol-function 'draw-input-bucket) fn\n*input-history* history))\narg)))\n\n(defmacro define-sudo-command (name command &amp;key output)\n(let ((cmd (gensym)))\n`(defcommand ,name (password) ((:password \"sudo password: \"))\n(let ((,cmd (concat \"echo '\" password \"' | sudo -S \" ,command)))\n,(if output\n`(run-prog-collect-output *shell-program* \"-c\" ,cmd)\n`(run-prog *shell-program* :args (list \"-c\" ,cmd) :wait nil))))))\n\n(define-sudo-command eon\n(concat \"pon em\"\n(when (ps-exists \"nm-applet\")\n(concat \"&amp;&amp; sudo \" (kill-ps-command \"nm-applet\"))))\n:output t)\n\n(define-sudo-command eoff \"poff em\" :output t)\n(define-sudo-command reboot \"reboot\")\n(define-sudo-command shutdown \"shutdown -h now\")\n(define-sudo-command backup \"~\/src\/backup\/backup.sh\")\n(define-sudo-command apache-start \"\/etc\/init.d\/apache2 start\")\n(define-sudo-command apache-stop \"\/etc\/init.d\/apache2 stop\")\n\n; --- key bindings -----------------------------------------------\n(set-prefix-key (kbd \"C-,\"))\n(define-key *root-map* (kbd \"s\") \"colon exec firefox http:\/\/\")\n(define-key *root-map* (kbd \"*\") \"show-battery\")\n(define-key *root-map* (kbd \"w\") \"windowlist\")\n(define-key *root-map* (kbd \"c\") \"gnome-terminal\")\n(define-key *root-map* (kbd \"e\") \"emacs\")\n(define-key *root-map* (kbd \"b\") \"firefox\")\n(define-key *root-map* (kbd \"C-o\") \"soffice\")\n(define-key *root-map* (kbd \"v\") \"virtualbox\")\n(define-key *root-map* (kbd \"P\") \"gnome-panel\")\n(define-key *root-map* (kbd \"m\") \"mode-line\")\n(define-key *root-map* (kbd \"N\") '*notifications-map*)\n(define-key *root-map* (kbd \"u\") \"uptime\")\n; emacs style\n(define-key *root-map* (kbd \"0\") \"remove\")\n(define-key *root-map* (kbd \"1\") \"only\")\n(define-key *root-map* (kbd \"2\") \"vsplit\")\n(define-key *root-map* (kbd \"3\") \"hsplit\")\n(define-key *input-map* (kbd \"C-m\") 'input-submit)\n(define-key *input-map* (kbd \"C-h\") 'input-delete-backward-char)\n(define-key *input-map* (kbd \"M-p\") 'input-history-back)\n(define-key *input-map* (kbd \"M-n\") 'input-history-forward)\n(define-key *input-map* (kbd \"C-i\") 'input-complete-forward)\n(define-key *resize-map* (kbd \"C-m\") \"exit-iresize\")\n\n; --- mode line -------------------------------------------------\n(setf *mode-line-screen-position* :top)\n(setf *mode-line-frame-position* :top)\n(setf *mode-line-border-width* 1)\n(setf *mode-line-pad-x* 6)\n(setf *mode-line-pad-y* 0)\n(setf *mode-line-background-color* \"black\")\n(setf *mode-line-foreground-color* \"wheat\")\n(setf *mode-line-border-color* \"dark grey\")\n(setf *screen-mode-line-format*\n(list \"^03%N^n \"\n'(:eval (run-shell-command\n\"LANG=c; date '+%m\/%d(%a)_%H:%M'|tr -d [:cntrl:]\" t))\n\" | %b | %g\"))\n\n(when (and *startup-mode-line*\n(not (head-mode-line (current-head))))\n(mode-line))","created_at":"2020-01-07T23:53:12+0000","updated_at":"2020-01-07T23:53:17+0000","published_by":["tomoyuki28jp"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/42951?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","configuration","smartparens"],"is_public":false,"id":1089,"title":"a-few-paredit-keys-that-take-over-the-world.html","url":"https:\/\/endlessparentheses.com\/a-few-paredit-keys-that-take-over-the-world.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2020-01-07T14:52:25+0000","updated_at":"2020-01-07T14:52:33+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"endlessparentheses.com"},{"is_archived":0,"is_starred":0,"tags":["emacs","smartparens"],"is_public":false,"id":1088,"title":"Emacs and Pairs","url":"https:\/\/ebzzry.io\/en\/emacs-pairs\/","content":"<p><a href=\"https:\/\/ebzzry.io\/eo\/emakso-paroj\/\">Esperanto<\/a>\u00a0\u25aa\u00a0English<\/p><p>August 15, 2015<\/p><p>Last updated: March 9, 2019<\/p><blockquote>\n<p>The white noise that beats within the white darkness is the rhythm of life; it is that pulse which never truly left the stage.<br \/>\u2015Ergo Proxy, Ergo Proxy<\/p>\n<\/blockquote><p><img src=\"https:\/\/ebzzry.io\/bil\/wallhaven-578010-1008x250.jpg\" class=\"c1\" alt=\"wallhaven-578010\" title=\"wallhaven-578010\" referrerpolicy=\"no-referrer\" \/><\/p><h2>Table of contents<\/h2><ul><li><a href=\"#introduction\">Introduction<\/a><\/li>\n<li><a href=\"#installation\">Installation<\/a><\/li>\n<li><a href=\"#configuration\">Configuration<\/a><\/li>\n<li><a href=\"#usage\">Usage<\/a>\n<ul><li><a href=\"#basics\">Basics<\/a><\/li>\n<li><a href=\"#navigation\">Navigation<\/a>\n<ul><li><a href=\"#startsandends\">Starts and ends<\/a><\/li>\n<li><a href=\"#traversinglists\">Traversing lists<\/a><\/li>\n<li><a href=\"#blockmovements\">Block movements<\/a><\/li>\n<li><a href=\"#toplevel\">Top-level-ish traversal<\/a><\/li>\n<li><a href=\"#freeform\">Free-form movements<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#manipulation\">Manipulation<\/a>\n<ul><li><a href=\"#wrapping\">Wrapping<\/a><\/li>\n<li><a href=\"#unwrapping\">Unwrapping<\/a><\/li>\n<li><a href=\"#slurpandbarf\">Slurp and barf<\/a><\/li>\n<li><a href=\"#swapping\">Swapping<\/a><\/li>\n<li><a href=\"#killing\">Killing<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#keys\">Keys<\/a><\/li>\n<li><a href=\"#closing\">Closing remarks<\/a><\/li>\n<\/ul><h2>Introduction<\/h2><p>In this article, I\u2019ll exclusively talk about <em>smartparens<\/em>\u2014a package that you wish you should have used, earlier, presuming you don\u2019t use it yet. If you\u2019re new to it, read along; if not, this may be a good refresher.<\/p><p><em>smartparens<\/em> is one of those packages that drastically improves, and changes how one uses Emacs. It\u2019s like having cybernetic limbs\u2014it makes you jump higher and punch harder.<\/p><p>Take note, though, that the name is a misnomer, as it not only handles parentheses. It handles just about anything that pairs, and it performs those functions stellarly.<\/p><h2>Installation<\/h2><p>Installing smartparens is straightforward:<\/p><pre>M-x package-install RET smartparens RET<\/pre><h2>Configuration<\/h2><p>Let\u2019s enable smartparens on startup, and hook it with some major hooks:<\/p><pre class=\"lisp\">(use-package smartparens-config\n  :ensure smartparens\n  :config (progn (show-smartparens-global-mode t)))\n(add-hook 'prog-mode-hook 'turn-on-smartparens-strict-mode)\n(add-hook 'markdown-mode-hook 'turn-on-smartparens-strict-mode)\n<\/pre><h2>Usage<\/h2><p>Managing paired characters like parentheses, braces, brackets, quotation marks, angle brackets, and other conceivable pair-able characters has always been a pain. Other packages solve that problem partially. However, they still miss several points.<\/p><p>In the code snippets below, the <code>^<\/code> symbol will be used to represent <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Point.html\">point<\/a>:<\/p><h3>Basics<\/h3><p>In smartparens, when you input a pair-able character:<\/p><pre class=\"clojure\">\n(defn foo )\n          ^\n<\/pre><p>the matching pair gets inserted, too, and point is positioned inside the pair:<\/p><pre class=\"clojure\">\n(defn foo [])\n           ^\n<\/pre><h3>Navigation<\/h3><h4>Starts and ends<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n(let [x \"foo bar baz ... blah\"])\n                         ^\n<\/pre><p>and you want to move point to the start of the string:<\/p><pre class=\"clojure\">\n(let [x \"foo bar baz ... blah\"])\n         ^\n<\/pre><p>Execute <code>sp-beginning-of-sexp<\/code>. I bound it to <kbd>C-M-a<\/kbd>.<\/p><p>Conversely, to move point to the end of the expression:<\/p><pre class=\"clojure\">\n(let [x \"foo bar baz ... blah\"])\n                             ^\n<\/pre><p>Execute <code>sp-end-of-sexp<\/code>. I bound it to <kbd>C-M-e<\/kbd>.<\/p><h4>Traversing lists<\/h4><p>If you have the expression:<\/p><pre class=\"lisp\">\n(defun format-date (format)\n  \"Insert date with FORMAT specification using a specific locale.\"\n  (let ((system-time-locale \"en_US.UTF-8\"))\n    (insert (format-time-string format)))) ^\n<\/pre><p>and you want to move point to <code>insert<\/code>:<\/p><pre class=\"lisp\">\n(defun format-date (format)\n  \"Insert date with FORMAT specification using a specific locale.\"\n  (let ((system-time-locale \"en_US.UTF-8\"))\n    (insert (format-time-string format))))\n     ^\n<\/pre><p>Execute <code>sp-down-sexp<\/code>. I bound it to <kbd>C-down<\/kbd>.<\/p><p>If you have the expression:<\/p><pre class=\"clojure\">\n(str \"foo\" \"bar baz qux\")\n    ^\n<\/pre><p>and you want to move point after <code>)<\/code>:<\/p><pre class=\"clojure\">\n(str \"foo\" \"bar baz qux\")\n                         ^\n<\/pre><p>Execute <code>sp-up-sexp<\/code>. I bound it to <kbd>C-up<\/kbd>.<\/p><p>If you have the expression:<\/p><pre class=\"clojure\">\n(defn foo [bar] (let [x 0] x))\n                ^\n<\/pre><p>and you want to move point to the preceeding <code>]<\/code>:<\/p><pre class=\"clojure\">\n(defn foo [bar] (let [x 0] x))\n              ^\n<\/pre><p>Execute <code>sp-backward-down-sexp<\/code>. I bound it to <kbd>M-down<\/kbd>.<\/p><p>If you have the expression:<\/p><pre class=\"lisp\">\n(insert (format-time-string format))\n                           ^\n<\/pre><p>and you want to move point to <code>(format<\/code>:<\/p><pre class=\"lisp\">\n(insert (format-time-string format))\n        ^\n<\/pre><p>Execute <code>sp-backward-up-sexp<\/code>. I bound it to <kbd>M-up<\/kbd>.<\/p><h4>Block movements<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n(:require [clojure.string :as s])\n          ^\n<\/pre><p>and you want to move point after <code>]<\/code>:<\/p><pre class=\"clojure\">\n(:require [clojure.string :as s])\n                                ^\n<\/pre><p>Execute <code>sp-forward-sexp<\/code>. I bound it to <kbd>C-M-f<\/kbd>.<\/p><p>Conversely, to move it back to <code>[<\/code>:<\/p><pre class=\"clojure\">\n(:require [clojure.string :as s])\n          ^\n<\/pre><p>Execute <code>sp-backward-sexp<\/code>. I bound it to <kbd>C-M-b<\/kbd>.<\/p><h4>Top-level-ish traversal<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n(defn blah\n  \"Returns blah of foo.\"\n  [foo]                 ^\n  )\n<\/pre><p>and you want to move point to <code>[<\/code>:<\/p><pre class=\"clojure\">\n(defn blah\n  \"Returns blah of foo.\"\n  [foo]\n  ^)\n<\/pre><p>Execute <code>sp-next-sexp<\/code>. I bound it to <kbd>C-M-n<\/kbd>.<\/p><p>Conversely, to move it back:<\/p><pre class=\"clojure\">\n(defn blah\n  \"Returns blah of foo.\"\n  [foo]                 ^\n  )\n<\/pre><p>Execute <code>sp-previous-sexp<\/code>. I bound it to <kbd>C-M-p<\/kbd>.<\/p><h4>Free-form movements<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n(defn blah [] (let [x 0 y 1] (+ x 1)))\n               ^\n<\/pre><p>and you want to move point to <code>blah<\/code>:<\/p><pre class=\"clojure\">\n(defn blah [] (let [x 0 y 1] (+ x 1)))\n      ^\n<\/pre><p>Execute <code>sp-backward-symbol<\/code>. I bound it to <kbd>C-S-b<\/kbd>.<\/p><p>Conversely, if you have the expression:<\/p><pre class=\"clojure\">\n(defn blah [] (let [x 0 y 1] (+ x 1)))\n            ^\n<\/pre><p>and you want to move point just after <code>(let<\/code>:<\/p><pre class=\"clojure\">\n(defn blah [] (let [x 0 y 1] (+ x 1)))\n                  ^\n<\/pre><p>Execute <code>sp-forward-symbol<\/code>. I bound it to <kbd>C-S-f<\/kbd>.<\/p><p>What they do is that they navigate around expressions as if delimiters, like parens, brackets, and braces do not exist.<\/p><h3>Manipulation<\/h3><h4>Wrapping<\/h4><p>If you have the expression:<\/p><pre class=\"javascript\">\nvar mods = \"vars\";\n           ^\n<\/pre><p>and you want <code>\"vars\"<\/code> mto be surrounded by <code>[<\/code> and <code>]<\/code>:<\/p><pre class=\"javascript\">\nvar mods = [\"vars\"];\n            ^\n<\/pre><p>Pressing <kbd>C-M-Space<\/kbd> or <kbd>ESC C-Space<\/kbd> followed by <kbd>[<\/kbd> will make the whole region become surrounded by matching <code>[<\/code> and <code>]<\/code>. It also applies to keys like <code>(<\/code>, <code>{<\/code>, <code>\"<\/code>, <code>'<\/code>, <code>*<\/code>, <code>_<\/code>, etc, depending on the mode that you\u2019re using.<\/p><p>Alternatively, you may define wrapping functions:<\/p><pre class=\"lisp\">(defmacro def-pairs (pairs)\n  \"Define functions for pairing. PAIRS is an alist of (NAME . STRING)\nconses, where NAME is the function name that will be created and\nSTRING is a single-character string that marks the opening character.\n  (def-pairs ((paren . \\\"(\\\")\n              (bracket . \\\"[\\\"))\ndefines the functions WRAP-WITH-PAREN and WRAP-WITH-BRACKET,\nrespectively.\"\n  `(progn\n     ,@(loop for (key . val) in pairs\n             collect\n             `(defun ,(read (concat\n                             \"wrap-with-\"\n                             (prin1-to-string key)\n                             \"s\"))\n                  (&amp;optional arg)\n                (interactive \"p\")\n                (sp-wrap-with-pair ,val)))))\n(def-pairs ((paren . \"(\")\n            (bracket . \"[\")\n            (brace . \"{\")\n            (single-quote . \"'\")\n            (double-quote . \"\\\"\")\n            (back-quote . \"`\")))\n<\/pre><p>These have the advantage of not requiring a region to operate on. I bound the first three functions to <kbd>C-c (<\/kbd>, <kbd>C-c [<\/kbd>, and <kbd>C-c {<\/kbd>, respectively. So, if you have the expression:<\/p><pre class=\"clojure\">\n(defn foo args (let [x 0] (inc x)))\n          ^\n<\/pre><p>and you want to surround <code>args<\/code> with <code>[<\/code> and <code>]<\/code>:<\/p><pre class=\"clojure\">\n(defn foo [args] (let [x 0] (inc x)))\n           ^\n<\/pre><p>Press <kbd>C-c [<\/kbd>.<\/p><p>Sometimes, we inadvertently delete one of the pair characters\u2014this results in an unbalanced expression. smartparens prevents us from doing that. If you hit <kbd>Backspace<\/kbd> in this expression:<\/p><pre class=\"javascript\">\nvar mods = [\"vars\"];\n            ^\n<\/pre><p>Nothing will happen. smartparens saves us a lot of trouble, here.<\/p><h4>Unwrapping<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n(foo (bar x y z))\n     ^\n<\/pre><p>and you want to unwrap the <code>bar<\/code> expression, removing the parentheses around <code>foo<\/code>:<\/p><pre class=\"clojure\">\nfoo (bar x y z)\n    ^\n<\/pre><p>Execute <code>sp-backward-unwrap-sexp<\/code>. I bound it to <kbd>M-[<\/kbd><\/p><p>Conversely, if you want to unwrap the <code>bar<\/code> expression, removing the parentheses around <code>bar<\/code>:<\/p><pre class=\"clojure\">\n(foo bar x y z)\n     ^\n<\/pre><p>Execute <code>sp-unwrap-sexp<\/code>. I bound it to <kbd>M-]<\/kbd>.<\/p><h4>Slurp and barf<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n[foo bar] baz\n        ^\n<\/pre><p>and you want <code>baz<\/code> to become part of <code>foo<\/code> and <code>bar<\/code>:<\/p><pre class=\"clojure\">\n[foo bar baz]\n        ^\n<\/pre><p>Execute <code>sp-forward-slurp-sexp<\/code>. I bound it to <kbd>C-right<\/kbd>.<\/p><p>Conversely, if you want to remove <code>baz<\/code>:<\/p><pre class=\"clojure\">\n[foo bar] baz\n        ^\n<\/pre><p>Execute <code>sp-forward-barf-sexp<\/code>. I bound it to <kbd>M-right<\/kbd>.<\/p><p>If you have the expression:<\/p><pre class=\"clojure\">\nblah [foo bar]\n             ^\n<\/pre><p>and you want <code>blah<\/code> to become part of <code>foo<\/code> and <code>bar<\/code>:<\/p><pre class=\"clojure\">\n[blah foo bar]\n             ^\n<\/pre><p>Execute <code>sp-backward-slurp-sexp<\/code>. I bound it to <kbd>C-left<\/kbd>.<\/p><p>Conversely, if you want to remove <code>blah<\/code>:<\/p><pre class=\"clojure\">\nblah [foo bar]\n             ^\n<\/pre><p>Execute <code>sp-backward-barf-sexp<\/code>. I bound it to <kbd>M-left<\/kbd>.<\/p><h4>Swapping<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n\"foo\" \"bar\"\n      ^\n<\/pre><p>and you want <code>\"foo\"<\/code> and <code>\"bar\"<\/code> to trade places:<\/p><pre class=\"clojure\">\n\"bar\" \"foo\"\n      ^\n<\/pre><p>Execute <code>sp-transpose-sexp<\/code>. I bound it to <kbd>C-M-t<\/kbd>.<\/p><h4>Killing<\/h4><p>If you have the expression:<\/p><pre class=\"clojure\">\n(let [x \"xxx\" y \"y yy yyy\" z 0])\n               ^\n<\/pre><p>and you want to kill just <code>\"y yy yyy\"<\/code>:<\/p><pre class=\"clojure\">\n(let [x \"xxx\" y z 0])\n               ^\n<\/pre><p>Execute <code>sp-kill-sexp<\/code>. I bound it to <kbd>C-M-k<\/kbd>.<\/p><p>If you want to kill <code>\"y yy yyy\" z 0<\/code>:<\/p><pre class=\"clojure\">\n(let [x \"xxx\" y])\n               ^\n<\/pre><p>Execute <code>sp-kill-hybrid-sexp<\/code>. I bound it to <kbd>C-k<\/kbd>.<\/p><p>If you have the expression:<\/p><pre class=\"clojure\">\n(:require [clojure.string :as s])\n                                ^\n<\/pre><p>and you want to kill <code>[clojure.string :as s]<\/code>:<\/p><pre class=\"clojure\">\n(:require )\n          ^\n<\/pre><p>Execute <code>sp-backward-kill-sexp<\/code>. I bound it to <kbd>M-k<\/kbd>.<\/p><h2>Keys<\/h2><p>The following snippet summarizes the key bindings used in this article. I use <code>bind-keys<\/code> to conveniently map my keys. I discussed about it, in an <a href=\"https:\/\/ebzzry.io\/en\/emacs-tips-2\">earlier<\/a> article.<\/p><pre class=\"lisp\">(defmacro def-pairs (pairs)\n  \"Define functions for pairing. PAIRS is an alist of (NAME . STRING)\nconses, where NAME is the function name that will be created and\nSTRING is a single-character string that marks the opening character.\n  (def-pairs ((paren . \\\"(\\\")\n              (bracket . \\\"[\\\"))\ndefines the functions WRAP-WITH-PAREN and WRAP-WITH-BRACKET,\nrespectively.\"\n  `(progn\n     ,@(loop for (key . val) in pairs\n             collect\n             `(defun ,(read (concat\n                             \"wrap-with-\"\n                             (prin1-to-string key)\n                             \"s\"))\n                  (&amp;optional arg)\n                (interactive \"p\")\n                (sp-wrap-with-pair ,val)))))\n(def-pairs ((paren . \"(\")\n            (bracket . \"[\")\n            (brace . \"{\")\n            (single-quote . \"'\")\n            (double-quote . \"\\\"\")\n            (back-quote . \"`\")))\n(bind-keys\n :map smartparens-mode-map\n (\"C-M-a\" . sp-beginning-of-sexp)\n (\"C-M-e\" . sp-end-of-sexp)\n (\"C-&lt;down&gt;\" . sp-down-sexp)\n (\"C-&lt;up&gt;\"   . sp-up-sexp)\n (\"M-&lt;down&gt;\" . sp-backward-down-sexp)\n (\"M-&lt;up&gt;\"   . sp-backward-up-sexp)\n (\"C-M-f\" . sp-forward-sexp)\n (\"C-M-b\" . sp-backward-sexp)\n (\"C-M-n\" . sp-next-sexp)\n (\"C-M-p\" . sp-previous-sexp)\n (\"C-S-f\" . sp-forward-symbol)\n (\"C-S-b\" . sp-backward-symbol)\n (\"C-&lt;right&gt;\" . sp-forward-slurp-sexp)\n (\"M-&lt;right&gt;\" . sp-forward-barf-sexp)\n (\"C-&lt;left&gt;\"  . sp-backward-slurp-sexp)\n (\"M-&lt;left&gt;\"  . sp-backward-barf-sexp)\n (\"C-M-t\" . sp-transpose-sexp)\n (\"C-M-k\" . sp-kill-sexp)\n (\"C-k\"   . sp-kill-hybrid-sexp)\n (\"M-k\"   . sp-backward-kill-sexp)\n (\"C-M-w\" . sp-copy-sexp)\n (\"C-M-d\" . delete-sexp)\n (\"M-&lt;backspace&gt;\" . backward-kill-word)\n (\"C-&lt;backspace&gt;\" . sp-backward-kill-word)\n ([remap sp-backward-kill-word] . backward-kill-word)\n (\"M-[\" . sp-backward-unwrap-sexp)\n (\"M-]\" . sp-unwrap-sexp)\n (\"C-x C-t\" . sp-transpose-hybrid-sexp)\n (\"C-c (\"  . wrap-with-parens)\n (\"C-c [\"  . wrap-with-brackets)\n (\"C-c {\"  . wrap-with-braces)\n (\"C-c '\"  . wrap-with-single-quotes)\n (\"C-c \\\"\" . wrap-with-double-quotes)\n (\"C-c _\"  . wrap-with-underscores)\n (\"C-c `\"  . wrap-with-back-quotes))\n<\/pre><h2>Closing remarks<\/h2><p>The plethora of commands in smartparens may be daunting at first, but the investement in time in learning them, will be minimal compared to benefits that you will reap.<\/p><p>smartparens is the brainchild of <a href=\"https:\/\/github.com\/Fuco1\">Matus\u00a0Goljer<\/a>. For more information on smartparens, go <a href=\"https:\/\/github.com\/Fuco1\/smartparens\">here<\/a>. If you like this project, you may donate <a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=CEYP5YVHDRX8C\">here<\/a>.<\/p><hr \/>","created_at":"2020-01-07T14:32:08+0000","updated_at":"2020-01-07T14:32:14+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"ebzzry.io","preview_picture":"https:\/\/ebzzry.io\/images\/ico\/android-chrome-512x512.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","org-mode"],"is_public":false,"id":1087,"title":"novoid\/Memacs","url":"https:\/\/github.com\/novoid\/Memacs\/","content":"<p><strong>What were you doing<\/strong> on February 14th of 2007? On <strong>which tasks<\/strong> were you working on that very day you met your girl friend? When was the <strong>last appointments<\/strong> with your dentist? <strong>Who called<\/strong> you on telephone during that meeting with your customer last month?<\/p><p>Most people can not answer such questions. <strong>With Memacs you can!<\/strong><\/p><p>Memacs extracts metadata (subjects, timestamps, contact information, \u2026) from many different existing data sources (file names, emails, tweets, bookmarks, \u2026) on your computer and generates files which are readable by <a href=\"http:\/\/en.wikipedia.org\/wiki\/Emacs\" rel=\"nofollow\">GNU Emacs<\/a> with <a href=\"http:\/\/orgmode.org\/\" rel=\"nofollow\">Org mode<\/a>.<\/p><p>Example:<\/p><pre>   emails              -&gt; memacs-maildir.py  \\\n   firefox history     -&gt; memacs-firefox.py   |\n   SMS                 -&gt; memacs-sms.py       |\n   RSS-feeds           -&gt; memacs_rss.py       |\n   bank statements     -&gt; memacs-easybank.py  |&gt;  Memacs\n   postings            -&gt; memacs-slrn.org     |\n   git repository logs -&gt; memacs_git.py       |\n   svn repository logs -&gt; memacs_svn.py      \/\n   |_________________|   |_________________|     |______|\n   your personal data      Memacs modules        Org mode\n<\/pre><p>Memacs - as the central component of the system - is a hub for all the connectors that add data from individual data sources. Those connectors are called <strong>Memacs modules<\/strong> or short <em>module<\/em>.<\/p><ul><li><strong>target group<\/strong>: Users of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Emacs\" rel=\"nofollow\">Emacs<\/a> and <a href=\"http:\/\/orgmode.org\/\" rel=\"nofollow\">Org mode<\/a><\/li>\n<li><strong>skills necessary<\/strong>: creating\/modifying shell script code; creating scheduled tasks<\/li>\n<li><strong>project hosted on<\/strong> <a href=\"https:\/\/github.com\/novoid\/Memacs\">https:\/\/github.com\/novoid\/Memacs<\/a><\/li>\n<\/ul><h2><a id=\"user-content-python-2-to-3-migration\" class=\"anchor\" aria-hidden=\"true\" href=\"#python-2-to-3-migration\"><\/a>Python 2 to 3 Migration<\/h2><p>This repository was migrated from Python2 to Python3 (by <a href=\"https:\/\/github.com\/DerBeutlin\">DerBeutlin<\/a> - thanks!).<\/p><p>If you need to get the latest version that ran with Python2, do visit <a href=\"https:\/\/github.com\/novoid\/Memacs\/tree\/5d4b3f0802f43490af0d28f1eaf5fb07976cb0b3\">this version<\/a>. For the future, Python3 is the only way to go.<\/p><p>I\u2019ll write about migration impact (changing IDs, \u2026) when I have tested it myself.<\/p><h2><a id=\"user-content-table-of-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-contents\"><\/a>Table of Contents<\/h2><ul><li><a href=\"#installation\">Installation<\/a><\/li>\n<li><a href=\"#getting-started\">Getting started<\/a><\/li>\n<li><a href=\"#memacs-modules\">Memacs modules<\/a><\/li>\n<li><a href=\"#example-story\">Example story<\/a><\/li>\n<li><a href=\"#background\">Background<\/a><\/li>\n<li><a href=\"#contribute\">Contribute<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><ul><li>clone or download repository from github<\/li>\n<li>make sure to use <strong>python v3<\/strong><\/li>\n<li>create a virtualenv or<\/li>\n<li>export\/set <code>PYTHONPATH<\/code><\/li>\n<\/ul><pre>PYTHONPATH=\/path\/to\/memacs\n<\/pre><ul><li>install dependencies<\/li>\n<\/ul><pre>pip install -r requirements.txt\n<\/pre><h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting Started<\/h2><p>The basic concept of using a Memacs module is following:<\/p><ol><li>Choose a Memacs modules you want to use and set it up:\n<ul><li>Read the module descripion files in the <code>docs<\/code> folder.<\/li>\n<li>Each <a href=\"#memacs-modules\">module<\/a> shares global options but also comes with its very own set of custom arguments.<\/li>\n<li>Manually start the module of your choice from the <code>bin<\/code> folder of Memacs with the argument <code>--help<\/code>, e.g. <code>bin\/memacs_csv.py --help<\/code> in your command line in order to get an overview of the arguments of this module.<\/li>\n<li>You probably want to <strong>develop and test a script (=.bat= or =.sh=) for the module invocation<\/strong> in the command line and \u2026<\/li>\n<\/ul><\/li>\n<li>When this script works as expected, set up a periodical invocation so that your data gets converted to Org mode via Memacs regularily.\n<ul><li>This is usually done via operating system dependent methods:\n<ul><li>Windows: <a href=\"https:\/\/docs.microsoft.com\/en-us\/windows\/desktop\/taskschd\/task-scheduler-start-page\" rel=\"nofollow\">Task Scheduler<\/a><\/li>\n<li>GNU\/Linux: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cron\" rel=\"nofollow\">cron job<\/a> or via <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Systemd\/Timers\" rel=\"nofollow\">systemd<\/a><\/li>\n<li>macOS: <a href=\"http:\/\/www.launchd.info\/\" rel=\"nofollow\">launchd<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li>Think of another Memacs module you might want to try ;-)<\/li>\n<\/ol><p>Please make sure you also read the <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/FAQs_and_Best_Practices.org\">FAQ\u2019s and best practices<\/a>, as it contains many tips and tricks on how to meet your requirements and on how to choose a reasonable setup.<\/p><h2><a id=\"user-content-workflows\" class=\"anchor\" aria-hidden=\"true\" href=\"#workflows\"><\/a>Workflows<\/h2><p>Here are some workflows just to give you an initial impression how Memacs may give you much for digital fun.<\/p><h3><a id=\"user-content-the-basics\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-basics\"><\/a>The Basics<\/h3><p>When one or more Memacs modules are set-up and data starts flowing in to your Memacs Org mode files, you have many options to use Memacs for your personal workflows.<\/p><p>The most basic thing that changes with Memacs is that you might want to see Memacs-processed data in your agenda:<\/p><p>Either you are generating (1) Org mode files that are within your <a href=\"https:\/\/orgmode.org\/manual\/Agenda-files.html\" rel=\"nofollow\">agenda files<\/a> or you are generating (2) <a href=\"https:\/\/orgmode.org\/manual\/Archiving.html#Archiving\" rel=\"nofollow\">archive files<\/a> whose more or less empty Org mode file stub is within your agenda files (as described in the <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/FAQs_and_Best_Practices.org#performance-and-scalability\">FAQs<\/a>).<\/p><p>This way, the generated time-stamp information gets into your normal agenda (1) or in the <a href=\"https:\/\/orgmode.org\/manual\/Agenda-commands.html\" rel=\"nofollow\">extended in archives-mode<\/a> (by pressing <code>v A<\/code> in your agenda) that also shows content of the archive files (2).<\/p><h3><a id=\"user-content-filename-time-stamp-and-friends\" class=\"anchor\" aria-hidden=\"true\" href=\"#filename-time-stamp-and-friends\"><\/a>Filename Time-Stamp and Friends<\/h3><p>A Memacs feature I personally use all the time provides a somewhat more complex workflow whose basic data is provided by the <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_filenametimestamps.org\">filename time-stamp module<\/a>. The module indexes all my files that contain ISO-datestamps or ISO-timestamps in their file names just like <a href=\"http:\/\/linux-sxs.org\/utilities\/updatedb.html\" rel=\"nofollow\">updatedb<\/a> does for \u201clocate\u201d.<\/p><p>The fun starts, when this module is set-up and the file index is generated (nightly). With the elisp snippets mentioned in the module documentation, I am able to link to any of those files just by specifying their file name. This links never get broken since it does not matter in which folder files are located in any more.<\/p><p>Please refer to <a href=\"https:\/\/github.com\/novoid\/dot-emacs\/blob\/master\/config.org#handling-tsfile-links-memacs\">my Emacs configuration<\/a> and look out for all occurrences of <code>tsfile<\/code> (time-stamp file) which is my custom link for those files.<\/p><p>I even extended this workflow such that my <a href=\"https:\/\/github.com\/novoid\/lazyblorg\">blog system<\/a> is able to <a href=\"https:\/\/github.com\/novoid\/lazyblorg\/wiki\/Images#embedding-tsfile-image-files\">link to <code>tsfile:<\/code> images<\/a> independent of their location. Awesome stuff.<\/p><h2><a id=\"user-content-memacs-modules\" class=\"anchor\" aria-hidden=\"true\" href=\"#memacs-modules\"><\/a>Memacs Modules<\/h2><ul><li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_arbtt.org\">arbtt<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_battery.org\">battery<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_csv.org\">csv<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_filenametimestamps.org\">filenametimestamps<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_firefox_history.org\">firefox_history<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_chrome_history.org\">chrome_history<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_git.org\">git<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_gpx.org\">gpx<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_ical.org\">ical<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_imap.org\">imap<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_lastfm.org\">lastfm<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_mumail.org\">mu<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_phonecalls.org\">phonecalls<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_photos.org\">photos<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_rss.org\">rss<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_simplephonelogs.org\">simplephonelogs<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_sms.org\">sms<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_svn.org\">svn<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_twitter.org\">twitter<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_whatsapp.org\">whatsapp<\/a><\/li>\n<\/ul><p>This module is an example for developers: <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/memacs_csv.org\">docs\/memacs_csv.org<\/a><\/p><p>Those modules are <strong>not<\/strong> adopted to the new unified framework yet:<\/p><ul><li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/tmp\/emails\/maildir\/memacs_maildir.org\">maildir<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/tmp\/emails\/mbox\/memacs_mbox.org\">mbox<\/a> (also useful for newsgroups)\n<ul><li>see also <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/emails\/mbox\/works-for-me-hack\/memacs_mbox.org\">emails\/mbox\/works-for-me-hack\/memacs_mbox.org<\/a><\/li>\n<\/ul><\/li>\n<li><strong>bank account<\/strong>:\n<ul><li><a href=\"http:\/\/www.easybank.at\" rel=\"nofollow\">easybank.at<\/a>: see <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/tmp\/bank_statements\/easybank.at\/memacs_easybank.org\">tmp\/bank_statements\/easybank.at\/memacs_easybank.org<\/a><\/li>\n<\/ul><\/li>\n<li><strong>newsgroups<\/strong>\n<ul><li><strong><a href=\"http:\/\/en.wikipedia.org\/wiki\/Slrn\" rel=\"nofollow\">slrn<\/a><\/strong>: see <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/emails\/mbox\/works-for-me-hack\/memacs_mbox.org\">emails\/mbox\/works-for-me-hack\/memacs_mbox.org<\/a><\/li>\n<\/ul><\/li>\n<\/ul><p>Following modules exist as a rough idea only and might get implemented some day (by you?):<\/p><ul><li><strong><a href=\"http:\/\/karl-voit.at\/tagstore\/\" rel=\"nofollow\">tagstore<\/a><\/strong>: see <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/tmp\/tagstore\/memacs-tagstore.org\">tmp\/tagstore\/memacs-tagstore.org<\/a><\/li>\n<li><strong>calendar<\/strong>:\n<ul><li><a href=\"http:\/\/www.jpilot.org\/\" rel=\"nofollow\">JPilot<\/a>-datebook: see <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/calendar\/memacs_jpilot-datebook.org\">calendar\/memacs_jpilot-datebook.org<\/a><\/li>\n<\/ul><\/li>\n<li><strong>tasks<\/strong>:\n<ul><li><a href=\"http:\/\/www.jpilot.org\/\" rel=\"nofollow\">JPilot<\/a>-todos: see <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/tasklists\/jpilot-todos\/memacs_jpilot-todos.org\">tasklists\/jpilot-todos\/memacs_jpilot-todos.org<\/a><\/li>\n<\/ul><\/li>\n<li><strong>blog_systems<\/strong>:\n<ul><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Serendipity_(weblog_software)\" rel=\"nofollow\">Serendipity<\/a>: see <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/blog_systems\/serendipity\/memacs_serendipity.org\">blog_systems\/serendipity\/memacs_serendipity.org<\/a><\/li>\n<\/ul><\/li>\n<\/ul><p>Imagine you are already using Memacs.<\/p><p>When remembering that day, when you joined an interesting talk about \u00ab<a href=\"http:\/\/en.wikipedia.org\/wiki\/Getting_Things_Done\" rel=\"nofollow\">Getting Things Done<\/a>\u00bb (GTD), you start up your GNU Emacs with your main Org mode file. There you go to the Agenda-view and select this specific day a couple of months ago.<\/p><p>There it is, from 2pm to 3pm you scheduled this talk in your calendar. And then you realize that within this time frame, there appear some <a href=\"http:\/\/en.wikipedia.org\/wiki\/Jpeg\" rel=\"nofollow\">JPEG files<\/a> containing an <a href=\"http:\/\/www.cl.cam.ac.uk\/~mgk25\/iso-time.html\" rel=\"nofollow\">ISO 8601<\/a> time stamp[1] are indexed by one module. (filenametimestamp-module)<\/p><p>This image contains a slide you found interesting and which you photographed using your sleek smartphone. Who would remember having taken a picture during a talk?<\/p><p>Ten minutes after the talk you wrote a short message on <a href=\"http:\/\/Twitter.com\" rel=\"nofollow\">Twitter<\/a> where you mentioned useful URLs for your followers. This time those URLs are handy for yourself too! (Twitter-module)<\/p><p>On the evening of that day you see an entry of an incoming email from the author of the talk. Now you remember having had a cool conversation at the end of the talk where he promised you some additional information about that nice little GTD tool on his computer. Great that you got that link to that email too. Without Memacs you would probably never remembered that email again. (Maildir-module)<\/p><p>And then there were some bookmarks you saved this day, almost all related to great ideas you got from the GTD talk. (delicious-module)<\/p><p>This small story shows only a few use cases where different modules combine given data sources and their information to provide an overall view related to an event. Since Org mode has got links, no actual data has to be duplicated (except the meta data extracted by Memacs). Emails, files, bookmarks, and so forth are linked rather than copied.<\/p><p>[1] with periods instead of colons - just because the <a href=\"http:\/\/msdn.microsoft.com\/en-us\/library\/aa365247(v%3Dvs.85).aspx#naming_conventions\" rel=\"nofollow\">ancient limitations of Microsoft based file systems<\/a>; like \u00ab2011-02-14T14.35.42 ideas.jpg\u00bb<\/p><p>In 1945, <a href=\"http:\/\/en.wikipedia.org\/wiki\/Vannevar_Bush\" rel=\"nofollow\">Vannevar Bush<\/a> wrote a famous article \u00ab<a href=\"http:\/\/en.wikipedia.org\/wiki\/As_We_May_Think\" rel=\"nofollow\">As We May Think<\/a>\u00bb where he develops the idea of having a \u00abmemory extender\u00bb called <strong>Memex<\/strong>. The memex can store all letters, books, and other information which are related to a person.<\/p><p>Besides having foreseen several technologies like hypertext, he defined a device that holds all metadata and data and provides associative trails to access information.<\/p><p>In the last decade of the previous century Microsoft Research had a research program that resulted in <a href=\"http:\/\/en.wikipedia.org\/wiki\/MyLifeBits\" rel=\"nofollow\">MyLifeBits<\/a>. This software tried to store each information of the user like office documents, screenshots, name of active windows on the desktop computer, and even automatically took photographs (<a href=\"http:\/\/en.wikipedia.org\/wiki\/Sensecam\" rel=\"nofollow\">SenseCam<\/a>). This word did not result in any (open) software product. Bell and Gemmell wrote a book called \u00ab<a href=\"http:\/\/www.amazon.de\/gp\/product\/0525951342\/ref%3Das_li_ss_tl?ie%3DUTF8&amp;tag%3Dkarlssuder-21&amp;linkCode%3Das2&amp;camp%3D1638&amp;creative%3D19454&amp;creativeASIN%3D0525951342\" rel=\"nofollow\">Total Recall<\/a>\u00bb.<\/p><p>The Memacs project tries to implement the ideas of Vannevar Bush\u2019s vision with open source and open standards. Hence, it\u2019s name \u00abMemacs\u00bb is the obvious combination of \u00ab<a href=\"http:\/\/www.gnu.org\/software\/emacs\/\" rel=\"nofollow\">GNU Emacs<\/a>\u00bb and \u00abMemex\u00bb.<\/p><p>Memacs uses GNU Emacs Org mode to visualize and access information extracted by Memacs modules: using tags, time stamps, full text search, and so forth GNU Emacs is able to derive different views. The most important view probably is the <a href=\"http:\/\/orgmode.org\/org.html#Agenda-Views\" rel=\"nofollow\">Agenda-view<\/a> where you can see anything that happened during a specific day\/week\/month according to the time frame selected. But you can derive other views too. For example you can choose to generate a condensed search result using a <a href=\"http:\/\/en.wikipedia.org\/wiki\/Boolean_algebra_(logic)\" rel=\"nofollow\">boolean combination<\/a> of tags.<\/p><p>Related to Memacs, the project founder developed a research software called <strong><a href=\"http:\/\/karl-voit.at\/tagstore\/\" rel=\"nofollow\">tagstore<\/a><\/strong>. This system allows users to store (local) files using tags without a hierarchy of folders. As a natural extension, tagstore targets associative access for (local) files. You might want to check out tagstore too. Memacs and tagstore are a very useful combination.<\/p><p>If you do like to know how to efficiently organize digital files in a simple and operating system independent way, read <a href=\"http:\/\/karl-voit.at\/managing-digital-photographs\/\" rel=\"nofollow\">this blog post<\/a> from Karl. It might give you ideas for your workflows as well.<\/p><p>Karl also wrote <a href=\"http:\/\/arxiv.org\/abs\/1304.1332\" rel=\"nofollow\">a whitepaper on Memacs<\/a> which describes Memacs from a scientists point of view.<\/p><p>If you want to contribute to this cool project, please fork and contribute or write an additional module!<\/p><p>See <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/docs\/FAQs_and_Best_Practices.org\">docs\/FAQs_and_Best_Practices.org<\/a> for more developing information.<\/p><p>We are sure that there are a <strong>lot<\/strong> of cool ideas for other modules out there! This is just the beginning!<\/p><p>Memacs is designed with respect to minimal effort for new modules.<\/p><p>We are using <a href=\"http:\/\/www.python.org\/dev\/peps\/pep-0008\/\" rel=\"nofollow\">Python PEP8<\/a> and <a href=\"http:\/\/en.wikipedia.org\/wiki\/Test-driven_development\" rel=\"nofollow\">Test Driven Development (TDD)<\/a>.<\/p><p>Memacs is licensed under the GPLv3 <a href=\"https:\/\/github.com\/novoid\/Memacs\/blob\/master\/license.txt\">license<\/a>.<\/p>","created_at":"2020-01-07T03:42:51+0000","updated_at":"2020-01-07T03:42:57+0000","published_by":["novoid"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/727493?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","configuration"],"is_public":false,"id":1086,"title":"romatthe\/romacs","url":"https:\/\/github.com\/romatthe\/romacs","content":"<div class=\"application-main\" data-commit-hovercards-enabled=\"\"><main id=\"js-repo-pjax-container\" data-pjax-container=\"\"><div class=\"pagehead repohead readability-menu experiment-repo-nav\"><div class=\"repohead-details-container clearfix container-lg px-3\"><ul class=\"pagehead-actions\"><li><a class=\"tooltipped tooltipped-s btn btn-sm btn-with-count\" aria-label=\"You must be signed in to watch a repository\" rel=\"nofollow\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;notification subscription menu watch&quot;,&quot;repository_id&quot;:null,&quot;auth_type&quot;:&quot;LOG_IN&quot;,&quot;client_id&quot;:null,&quot;originating_request_id&quot;:&quot;CCBE:7459:66AD969:C46ADBA:5E13FB42&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/romatthe\/romacs&quot;,&quot;referrer&quot;:&quot;http:\/\/www.google.co.uk\/url?sa=t&amp;source=web&amp;cd=1&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"b6871c13d668641bbfc719e847584707b0d8e5714fb815743d55329617b5cbf3\" href=\"https:\/\/github.com\/login?return_to=%2Fromatthe%2Fromacs\"> Watch<\/a> <\/li>\n<li><a class=\"btn btn-sm btn-with-count tooltipped tooltipped-s\" aria-label=\"You must be signed in to star a repository\" rel=\"nofollow\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;star button&quot;,&quot;repository_id&quot;:220352370,&quot;auth_type&quot;:&quot;LOG_IN&quot;,&quot;client_id&quot;:null,&quot;originating_request_id&quot;:&quot;CCBE:7459:66AD969:C46ADBA:5E13FB42&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/romatthe\/romacs&quot;,&quot;referrer&quot;:&quot;http:\/\/www.google.co.uk\/url?sa=t&amp;source=web&amp;cd=1&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"b046a0367169b746dd5b1297da5799de7de315d214210f3e1d43d957a059dead\" href=\"https:\/\/github.com\/login?return_to=%2Fromatthe%2Fromacs\"> Star<\/a> <\/li>\n<li><a class=\"btn btn-sm btn-with-count tooltipped tooltipped-s\" aria-label=\"You must be signed in to fork a repository\" rel=\"nofollow\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;repo details fork button&quot;,&quot;repository_id&quot;:220352370,&quot;auth_type&quot;:&quot;LOG_IN&quot;,&quot;client_id&quot;:null,&quot;originating_request_id&quot;:&quot;CCBE:7459:66AD969:C46ADBA:5E13FB42&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/romatthe\/romacs&quot;,&quot;referrer&quot;:&quot;http:\/\/www.google.co.uk\/url?sa=t&amp;source=web&amp;cd=1&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"637b5d75a0c86dd08ea1afe84e2b4289c49bea7748978220afd4c6a15a378257\" href=\"https:\/\/github.com\/login?return_to=%2Fromatthe%2Fromacs\"> Fork<\/a> <\/li>\n<\/ul><\/div><\/div>\n<div class=\"container-lg clearfix new-discussion-timeline experiment-repo-nav px-3 repository-content\"><div class=\"signup-prompt-bg rounded-1 signup-prompt p-4 text-center mb-4 rounded-1 position-relative\"><h3 class=\"pt-2\">Join GitHub today<\/h3><p class=\"col-6 mx-auto\">GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.<\/p><a class=\"btn btn-primary\" data-hydro-click=\"{&quot;event_type&quot;:&quot;authentication.click&quot;,&quot;payload&quot;:{&quot;location_in_page&quot;:&quot;files signup prompt&quot;,&quot;repository_id&quot;:null,&quot;auth_type&quot;:&quot;SIGN_UP&quot;,&quot;client_id&quot;:null,&quot;originating_request_id&quot;:&quot;CCBE:7459:66AD969:C46ADBA:5E13FB42&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/romatthe\/romacs&quot;,&quot;referrer&quot;:&quot;http:\/\/www.google.co.uk\/url?sa=t&amp;source=web&amp;cd=1&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"5e97601b28166afe5f7f3710e5d8140c935c2cd0eb774bf185b80d133c36ea40\" data-ga-click=\"(Logged out) Sign up prompt, clicked Sign up, text:sign-up\" href=\"https:\/\/github.com\/join?source=prompt-code&amp;source_repo=romatthe%2Fromacs\">Sign up<\/a><\/div><p>\u270f\ufe0f A personal Emacs configuration<\/p><p><a class=\"topic-tag topic-tag-link\" href=\"https:\/\/github.com\/topics\/emacs\" title=\"Topic: emacs\" data-ga-click=\"Topic, repository page\" data-octo-click=\"topic_click\" data-octo-dimensions=\"topic:emacs\">emacs<\/a> <a class=\"topic-tag topic-tag-link\" href=\"https:\/\/github.com\/topics\/emacs-lisp\" title=\"Topic: emacs-lisp\" data-ga-click=\"Topic, repository page\" data-octo-click=\"topic_click\" data-octo-dimensions=\"topic:emacs-lisp\">emacs-lisp<\/a> <a class=\"topic-tag topic-tag-link\" href=\"https:\/\/github.com\/topics\/dotfiles\" title=\"Topic: dotfiles\" data-ga-click=\"Topic, repository page\" data-octo-click=\"topic_click\" data-octo-dimensions=\"topic:dotfiles\">dotfiles<\/a><\/p><details class=\"details-reset\"><summary title=\"Click for language details\" data-ga-click=\"Repository, language bar stats toggle, location:repo overview\">Emacs Lisp\n<\/summary><\/details><div class=\"file-navigation in-mid-page d-flex flex-items-start\"><details class=\"details-reset details-overlay select-menu branch-select-menu hx_rsm\" id=\"branch-select-menu\"><\/details><p><a class=\"btn btn-sm empty-icon float-right BtnGroup-item\" data-hydro-click=\"{&quot;event_type&quot;:&quot;repository.click&quot;,&quot;payload&quot;:{&quot;target&quot;:&quot;FIND_FILE_BUTTON&quot;,&quot;repository_id&quot;:220352370,&quot;client_id&quot;:null,&quot;originating_request_id&quot;:&quot;CCBE:7459:66AD969:C46ADBA:5E13FB42&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/romatthe\/romacs&quot;,&quot;referrer&quot;:&quot;http:\/\/www.google.co.uk\/url?sa=t&amp;source=web&amp;cd=1&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"c578d520d2373160144c7871297e761aa83999ef361926c50b4d0cc09d3a5b37\" data-ga-click=\"Repository, find file, location:repo overview\" data-hotkey=\"t\" data-pjax=\"true\" href=\"https:\/\/github.com\/romatthe\/romacs\/find\/master\">Find file<\/a><\/p><details class=\"get-repo-select-menu js-get-repo-select-menu js-anon-get-repo-select-menu position-relative details-overlay details-reset\"><summary class=\"btn btn-sm ml-2 btn-primary\" data-hydro-click=\"{&quot;event_type&quot;:&quot;repository.click&quot;,&quot;payload&quot;:{&quot;repository_id&quot;:220352370,&quot;target&quot;:&quot;CLONE_OR_DOWNLOAD_BUTTON&quot;,&quot;client_id&quot;:null,&quot;originating_request_id&quot;:&quot;CCBE:7459:66AD969:C46ADBA:5E13FB42&quot;,&quot;originating_url&quot;:&quot;https:\/\/github.com\/romatthe\/romacs&quot;,&quot;referrer&quot;:&quot;http:\/\/www.google.co.uk\/url?sa=t&amp;source=web&amp;cd=1&quot;,&quot;user_id&quot;:null}}\" data-hydro-click-hmac=\"1dd6ce33c8b4157e98cdefc1ff3c2a7883dd57f0e7bc70273d38ce003661223c\">Clone or download<\/summary>\n<\/details><\/div><div class=\"commit-tease js-details-container Details d-flex rounded-top-1\" data-issue-and-pr-hovercards-enabled=\"\"><div class=\"AvatarStack flex-self-start AvatarStack-body\"><a class=\"avatar\" data-skip-pjax=\"true\" data-hovercard-type=\"user\" data-hovercard-url=\"\/users\/romatthe\/hovercard\" data-octo-click=\"hovercard-link-click\" data-octo-dimensions=\"link_type:self\" href=\"https:\/\/github.com\/romatthe\"><img height=\"20\" width=\"20\" alt=\"@romatthe\" src=\"https:\/\/avatars0.githubusercontent.com\/u\/8373222?s=60&amp;v=4\" referrerpolicy=\"no-referrer\" \/><\/a><\/div><p> <a data-pjax=\"true\" title=\"Tweaked Rust configuration\" class=\"message text-inherit\" href=\"https:\/\/github.com\/romatthe\/romacs\/commit\/685ff5b2858eefc70e13549e3c687f77b05e38ef\">Tweaked Rust configuration<\/a><\/p><p>Latest commit <a class=\"commit-tease-sha mr-1\" href=\"https:\/\/github.com\/romatthe\/romacs\/commit\/685ff5b2858eefc70e13549e3c687f77b05e38ef\" data-pjax=\"\">685ff5b<\/a> Dec 14, 2019<\/p><\/div><div class=\"file-wrap\"><a class=\"d-none js-permalink-shortcut\" data-hotkey=\"y\" href=\"https:\/\/github.com\/romatthe\/romacs\/tree\/685ff5b2858eefc70e13549e3c687f77b05e38ef\">Permalink<\/a><table class=\"files js-navigation-container js-active-navigation-container\" data-pjax=\"\"><thead><tr><th>Type<\/th>\n<th>Name<\/th>\n<th>Latest commit message<\/th>\n<th>Commit time<\/th>\n<\/tr><\/thead><tbody><tr class=\"warning include-fragment-error\"><td class=\"icon\"><\/td>\n<td class=\"content\" colspan=\"3\">Failed to load latest commit information.<\/td>\n<\/tr><\/tbody><\/table><\/div><\/div>\n<\/main><\/div><div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n<details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark hx_rsm\" open=\"open\"><\/details>","created_at":"2020-01-07T03:30:10+0000","updated_at":"2020-01-07T03:30:16+0000","published_by":["romatthe"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/8373222?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["dominions"],"is_public":false,"id":1085,"title":"Steam Community :: Guide :: Rej's Guide to Winning Wars","url":"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=1224028251","content":"<p>3. The Best Defense is an All-Out, Screaming, Berserker Attack<\/p><div class=\"subSectionDesc\">When you own a province the enemy cannot move into it from anywhere but adjacent, and cannot move 'over' it as part of multiple-province movement unless they have flying or sailing (or teleporting). Let me reiterate that. THEIR MOVEMENT IS NOW BLOCKED FOR A TURN. Every time an enemy attacks you, there's a chance they'll slip up and give away their army strategy, which spells they plan to use, how they've ordered their troops. Every province you take is time bought for spells to come online, and it's a chance to gather intel on the enemy and what they plan to do in a real fight - a fight you plan to win.<p>Also, if the province is connected to your lands, you get income. Precious gold pounds and magical gems. If not, you deny it to the enemy and place economic and psychological pressure on them to hurry - and therefore, to make mistakes.<\/p><p>Detaching squads of soldiers with a commander, a mage with a few gems to spam elementals, or even a thug (very expensive for this kind of thing, but their mobility means you might use them to do this), a teleporting\/flying mage, or bloody anything you can get your hands on, mercs hired in a raided province, to take provinces reduces your opponent's ability to continue the war, their income, and their available forces, as well as the psychological toll. You will lose many of these guys! But you're paying for time with their lives - even if you don't earn back their cost, what they buy you is time, the use of the enemy's retaking forces, and the loss of enemy income as WELL as any income you gain.<\/p><p>You aren't 'raiding for gold', you're raiding for time, for position, and to draw the enemy into your trap. A trap that might have not been there the turn before, not some long laid plan, but simply an opportunity created by a relentless offensive. The more provinces, aka time, position, and intelligence, you have, the less the enemy does, and the more chance they'll blunder into something that'll rip them apart. <strong>Armies should shard raiding parties into each province they can reach<\/strong>, if you have the troops - or even split up into mini armies if that's not going to get them destroyed piecemeal. Forts should periodically spit out little tiny raider groups, or each mage picks up a few gems and goes off to do his bit for the glory of your god. You're feeding these into the teeth of enemy forces sometimes, but your willingness to do so buys you territory and time - and if the enemy is truly defending in depth, upscale, or abandon, these efforts as need be. Remember, a dead raiding force is a good investment if it takes a single territory - even if it doesn't pay back the gold cost, it pays in other ways.<\/p><\/div>","created_at":"2020-01-06T20:54:06+0000","updated_at":"2020-01-06T20:54:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"steamcommunity.com","preview_picture":"https:\/\/steamuserimages-a.akamaihd.net\/ugc\/879755598711080992\/7F8E4F44CDBE9379B94135B7973E5CF45899CE01\/?imw=512&imh=319&ima=fit&impolicy=Letterbox&imcolor=%23000000&letterbox=true","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["books","fantasy","fiction"],"is_public":false,"id":1084,"title":"15 new science fiction &amp; fantasy books to read in January 2020","url":"https:\/\/www.polygon.com\/2020\/1\/4\/21048332\/books-to-read-january-2020-science-fiction-fantasy-recommendations","content":"<p class=\"p--has-dropcap p-large-text\" id=\"MXeFSk\">The end of one year and the beginning of the next always has me working to read books that I missed or which I\u2019ve been really wanting to read for a while. Accordingly, I\u2019ve packed along three books while I\u2019ve been traveling to see family over the holidays, using the opportunity to catch up.<\/p><p id=\"Q2uD3t\">The first is William Gibson\u2019s 2014 science fiction novel <em>The Peripheral<\/em>, spurred on by news that Amazon is adapting the novel for an original streaming series, and the upcoming sequel called <em>Agency<\/em> (see more on that below). The novel is set between two time periods, the near future and slightly more distant future, with an interesting take on time travel. The second is Leigh Bardugo\u2019s debut adult novel, <em>Ninth House<\/em>, about Yale\u2019s mysterious, magical houses, and the efforts of a mysterious \u201cninth house\u201d that keep them all in check. It\u2019s fantastically detailed, and Bardugo has the perfect blend of characters and setting to keep me going. Finally, I\u2019ve been listening to Daniel Suarez\u2019s <em>Delta-V<\/em>, a techno thriller that takes place in the 2030s as the private space race kicks off, following cave diver James Tighe as he goes through an intense training regime as he prepares to become one of humanity\u2019s first asteroid miners.<\/p><p id=\"f1VkDb\">2020 brings a whole lot of new science fiction and fantasy novels to delve into. Here are 15 coming out in January to check out.<\/p><h2 id=\"IBBi5j\"><strong>Jan. 7<\/strong><\/h2><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/a0IQ5aOAMnKk8uRfLKyh5F4nNRg=\/0x0:463x700\/320x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/oqeiFkuDOIVawSwdZJgLfzQOq4M=\/0x0:463x700\/520x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/H5oU3C5HwE4C45T5kvj3Xe-AlT4=\/0x0:463x700\/720x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/Vv4zSkHaFgz7f30RLPib7-lesig=\/0x0:463x700\/920x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/LzvHbeHknDS9Ae-J2W9yYxzBogc=\/0x0:463x700\/1120x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/viyw8PePcmXbP-TJHBipRKdu01U=\/0x0:463x700\/1320x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/aYzMmqRcFzd4SE3K6h7pTCw2GLE=\/0x0:463x700\/1520x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/FzdMnbRTV2q70u7tymjBYHhfNXs=\/0x0:463x700\/1720x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/-_SVniPH-gSzXUKVNJ78fAz0uJM=\/0x0:463x700\/1920x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg 1920w\" alt=\"Resurgence by C.J. Cherryh cover men with guns stand in front of a snow covered train\" data-upload-width=\"463\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/vmP2P25gMXoqQDPSpLRhtIwWxRk=\/0x0:463x700\/1200x0\/filters:focal(0x0:463x700):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573009\/9780756414276.jpeg\" referrerpolicy=\"no-referrer\" \/><cite>Image: DAW Books<\/cite><\/figure><\/div><p id=\"VXNj16\"><em>Resurgence<\/em> is the 20th installment in Cherryh\u2019s popular, long-running Foreigner space-opera series, and the second in a trilogy of books that started with 2018\u2019s <em>Emergence<\/em>. That book followed a human diplomat named Bren Cameron, who was working to try and build a peace between the refugees of Alpha Station and the atevi, the natives of the world that the humans are orbiting. Bren has a new problem on his hands when Ilisidi, the widow of the atevi ruler, works to solve simmering tensions between the north and south atevi factions on her own, commandeering a train and is passengers to do so.<\/p><hr class=\"p-entry-hr\" id=\"wDjbL1\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/paZmqDf7GbYbTOzJ4HXkXpbqEmU=\/0x0:1200x1800\/320x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/zGvd9QOr-c0-W2kaSYdX5EESxPA=\/0x0:1200x1800\/520x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/1k07-DK6KhI5hd0dnSgutVVZf3w=\/0x0:1200x1800\/720x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/OFe5wrzlGjBwDQ8o8J1ENfPkUuM=\/0x0:1200x1800\/920x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/MCL6uc8fHyRN--AnV49zKUkMeKk=\/0x0:1200x1800\/1120x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/c-eDpEHNhduaYcuy3ir7EzPVWq0=\/0x0:1200x1800\/1320x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/OrnKrjxTWBch5uwRMvI-N8cg7To=\/0x0:1200x1800\/1520x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/cvBXzl104ZK9x9_lFXzqe5PEPA4=\/0x0:1200x1800\/1720x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/Wtn12i1z6jEtJHa0_XmwccWp--I=\/0x0:1200x1800\/1920x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg 1920w\" alt=\"the secret chapter cover with a shark shadow\" data-upload-width=\"1200\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/PgErup_7zWlLoH4Mrw0gJCMKQUw=\/0x0:1200x1800\/1200x0\/filters:focal(0x0:1200x1800):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573011\/the_secret_chapter_1.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Ace Books<\/cite><\/figure><\/div><p id=\"kt3HMH\">In the latest installment of her Invisible Library series, Genevieve Cogman follows librarian-spy Irene and dragon prince Kai as they jump through time and dimensions. Irene is summoned to the Library, where she\u2019s tasked with finding a book that will help save her world, which is currently on the verge of collapse. She\u2019s forced to track down a Fae trader named Mr. Nemo, and before he hands over the needed book, she and her companions have to steal a painting from another world. <a href=\"https:\/\/www.publishersweekly.com\/978-0-593-19784-4\"><em>Publishers Weekly<\/em> says<\/a> that \u201cCogman charts the heist with the fluid mix of humor and adventure series readers will expect, while adding new dimensions to fan favorite characters and introducing dangerous enemies.\u201d<\/p><hr class=\"p-entry-hr\" id=\"Pf1dj7\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/JnOrp0lvegMTEUlwwlPrPkk93ms=\/0x0:658x1000\/320x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/03t8DsI-Qsh1_mRWY2X53rLhffA=\/0x0:658x1000\/520x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/4n_oEczHCJ_1U5EWt4KYn_hJ9sI=\/0x0:658x1000\/720x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/-pD99YLrsnyT6Dmm6uEh58MvewA=\/0x0:658x1000\/920x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/AuI7FATIixfhoDJjrHoxjk7Ky4g=\/0x0:658x1000\/1120x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/cuRad3tiMpGRvENcN8dOGFYoXOs=\/0x0:658x1000\/1320x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/QnbGrefE7H4t4szFbEvYIRkSAT0=\/0x0:658x1000\/1520x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/rT8cFxTvPZw_uCkPKNsc8enmoYo=\/0x0:658x1000\/1720x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/QTxLdbzJ4G-dN6hoV7Z8uGKBsxc=\/0x0:658x1000\/1920x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg 1920w\" alt=\"a hand holds a crown in the air in the lady hotspur book cover\" data-upload-width=\"658\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/Zdrfz-5cdEKNdLvbFsa34qkmFiA=\/0x0:658x1000\/1200x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573012\/9780765392497.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Tor Books<\/cite><\/figure><\/div><p id=\"JRoBxN\">Tessa Gratton\u2019s latest is set in the same world as her last novel, <em>The Queens of Innis Lear<\/em>. The Lady Knights have a motto: Strike fast, love hard, live forever. They\u2019ve sworn to protect their kingdom and its heir apparent, Banna Mora. But when a revolution topples the throne, Mora faces a choice: throw away her duty to her nation, or retake the throne against her best friend, Hal Bolingbrooke, the head of the Lady Knights. Hal herself isn\u2019t comfortable with her position on the throne, and faces her own challenges, living up to the expectations that her fellow citizens have for her. Both seek the support of Lady Hotspur Persy, whose support could swing the tide of war in either direction.<\/p><p id=\"1imaUb\"><a href=\"https:\/\/www.kirkusreviews.com\/book-reviews\/tessa-gratton\/lady-hotspur\/\"><em>Kirkus Reviews<\/em> gave the book a starred review<\/a>, and characterizes the book as a retelling of William Shakespeare\u2019s <em>Henry IV, Part I<\/em>, saying that Gratton \u201cwrites in conversation with the bard instead of just copying him, using the play as a starting point for a tale about love, family, and creating space for your own story.\u201d<\/p><p id=\"Vizzaf\">Read an <a href=\"https:\/\/www.tor.com\/2019\/12\/02\/read-the-first-two-chapters-from-tessa-grattons-lady-hotspur\/\">excerpt<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"UZShvY\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/AvnM89yy93NzltLN3uEhupdt21I=\/0x0:1839x2775\/320x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/4VmJ5BPL66aga_alydO1bI0yZzY=\/0x0:1839x2775\/520x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/Ca1pjdfNCCx9_K2c1_UZx0y8Wn0=\/0x0:1839x2775\/720x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/MgxLEmC9po0ncYrkOxyBHYrJkK4=\/0x0:1839x2775\/920x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/5dn2eG6zic7vjSgB4jetoXJST9s=\/0x0:1839x2775\/1120x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/6A52yDlVrpDPMkZuLs_X39yGU-0=\/0x0:1839x2775\/1320x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/wN1Bq3TBhvQsxnDsbWcnyP-WXeM=\/0x0:1839x2775\/1520x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/YZON9-oGyncuJu2nhuz8UhDxVnk=\/0x0:1839x2775\/1720x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/Tf9pG5Vy0U9Rho0M5pbpU0a62mk=\/0x0:1839x2775\/1920x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg 1920w\" alt=\"qualityland book cover looks like an amazon package\" data-upload-width=\"1839\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/_xb5fEtgMc9pq4XCxqDsy5_PEq8=\/0x0:1839x2775\/1200x0\/filters:focal(0x0:1839x2775):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573013\/Kling_Qualityland_HC.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Grand Central Publishing<\/cite><\/figure><\/div><h2 id=\"M4RH6j\"><a href=\"https:\/\/www.grandcentralpublishing.com\/titles\/marc-uwe-kling\/qualityland\/9781538732960\/\"><strong>Qualityland<\/strong><\/a> <strong>by Marc-Uwe Kling (translated by Jamie Searle Romanelli)<\/strong><\/h2><p id=\"jHT3t9\">In the nation of QualityLand, citizens are ranked by an automated system that gives them social and employment opportunities, matches them up with the best partners, and connects them with the products that they need the most before they know they need them. Peter Jobless is a scrapper who can\u2019t bear to destroy imperfect machines, and accidentally becomes the leader of a band of robotic refugees and misfits. When he gets a product delivered to his workshop from TheShop, he knows its wrong, but returning it would mean that QualityLand\u2019s perfect algorithm is wrong, potentially shaking the nation to its foundations.<\/p><p id=\"3rUhCk\"><a href=\"https:\/\/www.publishersweekly.com\/978-1-5387-3296-0\"><em>Publishers Weekly<\/em> says that<\/a> Kling\u2019s book is \u201cSharp and biting, the most implausible aspect of Kling\u2019s novel is the relative note of optimism that ends it. This is spot-on satire.\u201d<\/p><p id=\"9SJvHo\">Listen to an excerpt <a href=\"https:\/\/go.redirectingat.com?id=66960X1516586&amp;xs=1&amp;url=https%3A%2F%2Fsoundcloud.com%2Fhachetteaudio%2Fqualityland-by-marc-uwe-kling&amp;referrer=polygon.com&amp;sref=https%3A%2F%2Fwww.polygon.com%2F2020%2F1%2F4%2F21048332%2Fbooks-to-read-january-2020-science-fiction-fantasy-recommendations\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"QuCEkJ\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/w5wZxB-OxGBFqF76ab42rSFvpCs=\/0x0:350x560\/320x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/R9EOSIghQbFD2xjXXs1IDFxnFxc=\/0x0:350x560\/520x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/iHZcS_-ryKMqE5RZyUsv04Eg_tM=\/0x0:350x560\/720x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/pkQJqGS4d2cqOhtIe_RxNKJ96Ac=\/0x0:350x560\/920x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/PTV0IXeNi52DGeABx5ZJIbIAEXM=\/0x0:350x560\/1120x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/45F55DidSU3Zrx1l_tYrWbvTM5g=\/0x0:350x560\/1320x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/llcOnMTSNu9ECnBWaH9TNhiwbOA=\/0x0:350x560\/1520x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/OSKiC_kevUtyGhMo4ZCogkfC588=\/0x0:350x560\/1720x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/-aSRlG-qRmI0LWGN0KMmDWmqkUw=\/0x0:350x560\/1920x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg 1920w\" alt=\"Come Tumbling Down book cover lightning strikes a door standing in the middle of a pasture\" data-upload-width=\"350\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/Q4D_WyK3DbwiQqM4PFb1ybB3UhY=\/0x0:350x560\/1200x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573014\/9780765399311.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Tor.com<\/cite><\/figure><\/div><p id=\"Axtves\">Now in its fifth installment, Seanan McGuire\u2019s Wayward Childrenseries has explored a number of fantasy tropes as her characters go through trans-dimensional doorways to find fantastical new worlds. <em>Come Tumbling Down<\/em> picks up some of the story from two prior installments, <em>Every Heart a Doorway<\/em> and <em>Down Among the Sticks and Bones.<\/em> In those installments, Jack Wolcott left Eleanor West\u2019s School for Wayward Children with her recently-murdered sister Jill in tow, but when she is brought back to the school, she realizes that a terrible mad scientist has switched their bodies.<\/p><p id=\"v0OPew\"><a href=\"https:\/\/www.kirkusreviews.com\/book-reviews\/seanan-mcguire\/come-tumbling-down\/\"><em>Kirkus Reviews<\/em> gave the book a starred review and says that McGuire<\/a> \u201cexamines the thin line separating heroes from monsters \u2014 and then blurs that line completely. As in the other series installments, she also argues that one\u2019s real or perceived flaws can prove to be a source of strength despite, or even because of, the pain they cause to oneself and others.\u201d<\/p><p id=\"cFAUma\">Read an <a href=\"https:\/\/us.macmillan.com\/excerpt?isbn=9780765399311\">excerpt<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"ACMSq4\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/0Y8P5W769pHxAr9sFGKspVczLO4=\/0x0:658x1000\/320x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/HRrXe0Xu1iWWaDyYiDzXd-rgevo=\/0x0:658x1000\/520x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/5k1HV1SE2IVP0H79EhGmyrgxNBc=\/0x0:658x1000\/720x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/QB2Wz34mQNUCouXSF8dLz3DWbO8=\/0x0:658x1000\/920x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/mMlqo0QVtUFuViZsuptb2jorNXA=\/0x0:658x1000\/1120x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/A6_NclLda-PeYY1U_RaI1IyUJgw=\/0x0:658x1000\/1320x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/mDViZo_V_v9ZEMvkvOa-W2ur-bU=\/0x0:658x1000\/1520x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/wYB_wSpB2P6Efihdpx5Khd-pxwE=\/0x0:658x1000\/1720x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/TjOyO32S66ga1rLUfYbfLAYqMUI=\/0x0:658x1000\/1920x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg 1920w\" alt=\"the god game book cover is glitchy\" data-upload-width=\"658\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/I1Qg1x9chSvpJSsPmnGI8vPm7xw=\/0x0:658x1000\/1200x0\/filters:focal(0x0:658x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573016\/9781250306142.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: St. Martin\u2019s Press<\/cite><\/figure><\/div><p id=\"1WMfpZ\">In the near future, Charlie and his friends are invited to a new video game, run by an artificial intelligence that believes that it\u2019s God. At first, it seems cool: they complete missions using their phones and AR glasses, earning expensive tech and money for rewards. But the missions become crueler, with deadly consequences as the AI demands that they worship and obey it. <a href=\"https:\/\/www.publishersweekly.com\/978-1-250-30614-2\"><em>Publishers Weekly<\/em> says that<\/a> \u201cthe heartlessness on display may put off some readers, but fans of AI run amok should relish this one.\u201d<\/p><p id=\"cAXNKx\">Read an <a href=\"https:\/\/us.macmillan.com\/excerpt?isbn=9781250306142\">excerpt<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"0fgE6O\" \/><h2 id=\"hwxQKe\"><strong>Jan. 14<\/strong><\/h2><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/Vu74BCgn3aJjXjV_YaVh_MzifRo=\/0x0:1701x2560\/320x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/-TJy6mene6sTORGtFm_44UHoyXg=\/0x0:1701x2560\/520x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/Iwmj43PZuJZxo79StFlJELDwSuQ=\/0x0:1701x2560\/720x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/A1G0aLytO3BDAAWC97JYpqt_IoM=\/0x0:1701x2560\/920x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/Y7Q9KBfqkILDdnMainSyAnx6zJc=\/0x0:1701x2560\/1120x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/PHRCW5xu2y9tnku-oKSv4O0PJDg=\/0x0:1701x2560\/1320x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/me6yeSgGUEKVyp_dH-vv1acv-uM=\/0x0:1701x2560\/1520x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/CtV7T2qxG2dFnJveCAuLKCdMUhA=\/0x0:1701x2560\/1720x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/vTjmAEaKZZi_5ZJRCET3EaZjl28=\/0x0:1701x2560\/1920x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg 1920w\" alt=\"a beginning at the end book cover has a collapsing golden gate bridge\" data-upload-width=\"1701\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/fMfRZo_4RybDt6gM2oDTOJonrww=\/0x0:1701x2560\/1200x0\/filters:focal(0x0:1701x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573017\/81RijIN96oL.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Mira Books<\/cite><\/figure><\/div><p id=\"GhDwNr\">In the aftermath of a global pandemic, the survivors of <em>A Beginning at the End<\/em> are in the midst of reconstruction. Mike Chen follows a series of characters working to rebuild their lives: former musician Moria, who is trying to restart her life, only to have those plans thwarted when her domineering father goes through great lengths to track her down. Krista runs errands for people too traumatized to go outside, while Rob faces problems when the government tries to separate him from his daughter. As a new outbreak looms, the new friends are forced to face their greatest fears.<\/p><p id=\"Cr5H68\"><a href=\"https:\/\/www.publishersweekly.com\/978-0-7783-0934-5\"><em>Publishers Weekly<\/em> says that<\/a> \u201cby foregrounding family, Chen manages to imbue his apocalypse with heart, hope, and humanity. Sci-fi fans will delight in this lovingly rendered tale.\u201d<\/p><p id=\"Xw9XbX\">Read an <a href=\"https:\/\/www.themarysue.com\/a-beginning-at-the-end-mike-chen-cover-excerpt\/\">excerpt<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"ECDM9S\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/1UbsezEqdddq927ffLPh_C9l0wM=\/0x0:596x900\/320x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/u-bh7jgHrBpmn30gSUXZeF47ar0=\/0x0:596x900\/520x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/VaX0Fdyg4jK1SwQDx81p91R43Pc=\/0x0:596x900\/720x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/2fCfcGid2snXfo3IJQeroyaLkz4=\/0x0:596x900\/920x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/Lqi01gROXj0XAeyPNtv68xdgbto=\/0x0:596x900\/1120x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/y-lAvOaFuc4nWAuWGFa5PXNWIaQ=\/0x0:596x900\/1320x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/NfBhor35AZf8278sR9LdrSLJdug=\/0x0:596x900\/1520x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/S1LPAmj5VspEQW_YI9-ZJzIOLN0=\/0x0:596x900\/1720x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/rnhpACBCyCP3GIV-vpMjsegymN0=\/0x0:596x900\/1920x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg 1920w\" alt=\"a mysterious girl looks at the reader in the book cover for a longer fall\" data-upload-width=\"596\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/4jUylEC-CM-114nVLuYyPbjLiTE=\/0x0:596x900\/1200x0\/filters:focal(0x0:596x900):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573019\/a_longer_fall_9781481494953_xlg.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Saga Press<\/cite><\/figure><\/div><p id=\"kM8dbi\">Charlaine Harris, best known for her Sookie Sackhouse and Midnight Texas novels, continues her latest series, about gunslinger Lizbeth Rose. Set in an alternate, broken United States, Rose faces off against Russian wizards and outlaws, and in the new book, she joins a new crew with a simple job: transport a crate into Dixie (the former American South). Elements from three territories want to get their hands on the cargo, and when it\u2019s stolen, Lizbeth has to go undercover to try and retrieve it. <a href=\"https:\/\/www.kirkusreviews.com\/book-reviews\/charlaine-harris\/a-longer-fall\/\"><em>Kirkus Reviews<\/em> says that<\/a> \u201cthe indomitable, quick-on-the-draw Lizbeth remains an irresistible heroine, and Harris proves she still has the magic touch.\u201d<\/p><p id=\"WoiUO4\">Read an <a href=\"https:\/\/www.simonandschuster.com\/books\/A-Longer-Fall\/Charlaine-Harris\/Gunnie-Rose\/9781481494953\">excerpt<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"zy7Gdw\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/-1En6wPVumwOOrngm_lyW-8y57U=\/0x0:652x1000\/320x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/de5bR05oTzZeyZuZMtlpb9mOXgg=\/0x0:652x1000\/520x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/mK-geJcxKTMFnXym1Nq7A-IQ31g=\/0x0:652x1000\/720x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/fVBXcNCfh7dxoqE-jMHK3S2VlQs=\/0x0:652x1000\/920x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/dfRbtSWE4RscFK_a2CVscKou89I=\/0x0:652x1000\/1120x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/wy0wMHIdyKZraOdetyhCRasLULg=\/0x0:652x1000\/1320x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/JSytzvbzH0n_Z7bT9wg0gLFmexM=\/0x0:652x1000\/1520x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/jpMM4xqWoomPECL1fxmW2j2YzE0=\/0x0:652x1000\/1720x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/yrPm9UCDuQcszeW-lRJgMZ-Jy_0=\/0x0:652x1000\/1920x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg 1920w\" alt=\"a boy with blue hair stands in front of a blood rose monster in the cover for burn the dark\" data-upload-width=\"652\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/iy7HPK1Bb-pmO_YSeyL7T_UXhz4=\/0x0:652x1000\/1200x0\/filters:focal(0x0:652x1000):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573022\/9781250306432.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Tor<\/cite><\/figure><\/div><p id=\"4FiWRz\">Robin Martine has gained an incredible following on YouTube for her channel, Malus Domestica, an incredibly realistic video series in which she hunts witches. Little do her followers know that the fights are actually real, and that she\u2019s on a mission to avenge the death of her mother. When she returns home to Blackfield, Georgia, she\u2019s haunted by a spectre called the Red Lord, which could derail her plans. <a href=\"https:\/\/www.publishersweekly.com\/978-1-250-30643-2\"><em>Publishers Weekly<\/em> says<\/a> that \u201cwith a detailed setting and an ear for dialogue, Hunt captures a distinct feeling of Southern Americana.\u201d<\/p><p id=\"exyLtH\">Read an excerpt <a href=\"https:\/\/us.macmillan.com\/excerpt?isbn=9781250306432\">here<\/a> and <a href=\"http:\/\/terribleminds.com\/ramble\/2019\/07\/04\/s-a-hunt-cover-reveal-and-excerpt-for-burn-the-dark\/\">here<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"DVQcVJ\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/Vt6bq9BuO1hF5QUf5qCnim6_gzs=\/0x0:1535x2327\/320x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/Q9BMTlgpbbg-fAikxwt1KJd9tnI=\/0x0:1535x2327\/520x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/AansVG-s0kaOvN5zLyrn6JqsI2M=\/0x0:1535x2327\/720x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/fwZjDA_L6UFIMG3Bj86mhKTcNhk=\/0x0:1535x2327\/920x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/yp4y13qBfkGhYS2oROUIQ0lcyxg=\/0x0:1535x2327\/1120x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/TINRnPvoLVkVIg8X4FpmoGQ4rso=\/0x0:1535x2327\/1320x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/rqSt-gj3p7wrLbsdMrz7uNf6tuA=\/0x0:1535x2327\/1520x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/7ZwEEwyziERV8F0R2O-mdTsUcPA=\/0x0:1535x2327\/1720x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/GmvXw72TKe2txbBV__HHAgDMsHE=\/0x0:1535x2327\/1920x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg 1920w\" alt=\"the broken heavens cover has planets swirling around a gyroscope\" data-upload-width=\"1535\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/Sqfxd2yjSHw7PCf8nm3dLaatcQc=\/0x0:1535x2327\/1200x0\/filters:focal(0x0:1535x2327):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573025\/810lxFBmbjL.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Angry Robot<\/cite><\/figure><\/div><p id=\"nQ08pW\">Kameron Hurley\u2019s conclusion to her Worldbreaker Saga has finally arrived, bringing to a close a fantasy epic that began with <em>The Mirror Empire.<\/em> The reappearance of the star Oma has brought the promise of destruction to the Dhai nation, which has broken apart under an onslaught from Tai Kao invaders from another world. As Oma begins to wane, only one world will survive, and as the Tai Kao work to establish a permanent presence, the surviving Dhai are forced to figure out how to move forward and build a new future for themselves.<\/p><hr class=\"p-entry-hr\" id=\"x6p3wz\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/uzw-KtiFHWU_favtlJ5UxC4jQoc=\/0x0:1875x2850\/320x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/2I9dgJgbSkFsvRifGXo1gSZf-7g=\/0x0:1875x2850\/520x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/ep2c2XckVeePg7SPMSq78GKAebU=\/0x0:1875x2850\/720x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/4zatEZHhT6pk5PW2Kw2oMQCUc1Q=\/0x0:1875x2850\/920x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/jcCie3k90YyVhv8lQ6oy3Hc2sbg=\/0x0:1875x2850\/1120x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/bbMU1BOivfARuMl6WqGGTjd7ZBI=\/0x0:1875x2850\/1320x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/WoGNnQZFRGu2qu72aVreHyzlaAM=\/0x0:1875x2850\/1520x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/-SbjhfA-CxoPMe9Kl3akiwglLto=\/0x0:1875x2850\/1720x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/aFYcqcjRpSDg3KQkeyTMBOv6yKs=\/0x0:1875x2850\/1920x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg 1920w\" alt=\"The Vanished Birds by Simon Jimenez cover has some multicolor stripes\" data-upload-width=\"1875\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/MhdmXcBqPH3N_3fD7LuT4Fqj0Nk=\/0x0:1875x2850\/1200x0\/filters:focal(0x0:1875x2850):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573027\/9780593128985.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Penguin Random House<\/cite><\/figure><\/div><p id=\"PDj8BY\">A hundred years from now, a starship engineer named Fumiko works to design starships that will help lift humanity away from a climate change-ravaged Earth, something he comes to regret as he leaves everyone behind. When a mute boy crash lands on a distant planet, starship captain Nia Imani agrees to take him to Pelican Station, forming an unusual bond with him as she carries him across space. Traveling via Pocket Space means that decades pass in just months, and when powerful individuals and corporations begin to realize the potential that the boy holds, they\u2019ll stop at nothing to get their hands on him. <em>Kirkus Reviews<\/em> gave the book a <a href=\"https:\/\/www.kirkusreviews.com\/book-reviews\/simon-jimenez\/the-vanished-birds\/\">starred review<\/a>, saying that it\u2019s \u201cthe best of what science fiction can be: a thought-provoking, heart-rending story about the choices that define our lives.\u201d<\/p><p id=\"LgsIHW\">Read an <a href=\"https:\/\/www.tor.com\/2019\/11\/25\/excerpts-the-vanished-birds-simon-jimenez\/\">excerpt<\/a>.<\/p><hr class=\"p-entry-hr\" id=\"L9ggel\" \/><h2 id=\"YkS03d\"><strong>Jan. 21<\/strong><\/h2><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/zksDXVyLmc1HWUNUWsPB8hdb2K8=\/0x0:1696x2560\/320x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/ced0yJLaVqy9IfFf3gtrPNjQBlg=\/0x0:1696x2560\/520x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/agSZXMjs-EMzxvhvDSfRkeDzRtM=\/0x0:1696x2560\/720x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/0IW3RE2jMSjNLWpVpqxbKKMCvvI=\/0x0:1696x2560\/920x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/rshii0mVst5TN2wA3L-sX5O36e8=\/0x0:1696x2560\/1120x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/h52Xewhq2R9r3lM8PQJiinYtsoQ=\/0x0:1696x2560\/1320x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/2icaq20woClp4xGMGP7a9QwT4Ps=\/0x0:1696x2560\/1520x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/k4eLtdIfLUPI-3M1afB1M-mY9D0=\/0x0:1696x2560\/1720x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/MpiwYt_6itzgD3olnMWvTCaQD9E=\/0x0:1696x2560\/1920x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg 1920w\" alt=\"agency by william gibson cover has a black woman out of focus\" data-upload-width=\"1696\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/qErQVlzOmmD5qhQO7Ot6i_o6sro=\/0x0:1696x2560\/1200x0\/filters:focal(0x0:1696x2560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573031\/71fj2SHMomL.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Putnam<\/cite><\/figure><\/div><h2 id=\"pRWVnW\"><a href=\"https:\/\/go.redirectingat.com?id=66960X1516586&amp;xs=1&amp;url=https%3A%2F%2Fwww.penguinrandomhouse.com%2Fbooks%2F530536%2Fagency-by-william-gibson%2F&amp;referrer=polygon.com&amp;sref=https%3A%2F%2Fwww.polygon.com%2F2020%2F1%2F4%2F21048332%2Fbooks-to-read-january-2020-science-fiction-fantasy-recommendations\" rel=\"nofollow noopener\" target=\"_blank\"><strong>Agency<\/strong><\/a> <strong>by William Gibson<\/strong><\/h2><p id=\"z5GDor\">As mentioned, William Gibson\u2019s 2014 novel <em>The Peripheral<\/em> is getting a sequel this month: <em>Agency<\/em>. In it, he returns to the same world and time periods, following Verity Jane, an app tester who is tasked with beta testing a new set of smart glasses, equipped with an artificial intelligence called Eunice. She realizes that Eunice is far more powerful than her employers realize, and sets out to prevent them from learning the truth. Meanwhile, a century ahead in a slightly different future, a man named Ainsley Lowbeer works to nudge those alternate pasts in certain directions.<\/p><p id=\"h7jrvH\"><a href=\"https:\/\/www.publishersweekly.com\/9781101986936\"><em>Publishers Weekly<\/em> says that the book<\/a> will have\u201cCyberpunk fans looking to dive into the \u2018what-if\u2019s\u2019 of an alternate timeline will be as enraptured as ever by Gibson\u2019s imagination.\u201d<\/p><hr class=\"p-entry-hr\" id=\"8qxMrO\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/IjamtNYbweHSQAzdi8gajPcoOLo=\/0x0:350x560\/320x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/Pz94YnBYP6PtEf1dIyNJ5fsxZKs=\/0x0:350x560\/520x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/LUqt6WLoZ-e7xZu5oC-M03VYIxI=\/0x0:350x560\/720x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/uN-skH17Qfrqrhooz_AEJxHXGqA=\/0x0:350x560\/920x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/A8_kw1NtAu7xFjWbSIql-jk_qBo=\/0x0:350x560\/1120x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/LqQH516Hx83vLvcPKngd0Yt5nZI=\/0x0:350x560\/1320x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/Q6YgmOTXf6ntTBrbyJ6IwWifs_c=\/0x0:350x560\/1520x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/_CBk4ACYpfSbHDlS2VE_1dGjEN4=\/0x0:350x560\/1720x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/3CXbVvW4hghFcm7oLdanqsmjT-4=\/0x0:350x560\/1920x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg 1920w\" alt=\"riot baby cover is a close up of a black woman\u2019s face\" data-upload-width=\"350\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/WzRsOSJd-oSFJ45huxzULNZyHEE=\/0x0:350x560\/1200x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573033\/9781250214751.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Tor.com<\/cite><\/figure><\/div><h2 id=\"N0AJY5\"><a href=\"https:\/\/publishing.tor.com\/riotbaby-tochionyebuchi\/9781250214751\/\"><strong>Riot Baby<\/strong><\/a> <strong>by Tochi Onyebuchi<\/strong><\/h2><p id=\"WxC8rf\">Ella has a bunch of special talents: the ability to see the fates and futures of the people around her, travel from place to place, move undetected, and destroy things with a thought. Her brother, Kev, wants to protect her, but when he\u2019s jailed, she works to keep him sane by showing him the future and the state of the world outside of his cell walls. <em>Publishers Weekly<\/em> gave the book a <a href=\"https:\/\/www.publishersweekly.com\/978-1-250-21475-1\">starred review<\/a>, saying that \u201cOnyebuchi\u2019s unexpectedly hopeful ending is just as powerful as his unflinching, heartbreaking depictions of racism and cruelty. This staggering story is political speculative fiction at its finest.\u201d<\/p><hr class=\"p-entry-hr\" id=\"Z8bvf7\" \/><h2 id=\"1S6rZX\">Jan. 28<\/h2><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/nf0hJ2rS-JAeMA_VCF-FzFuwylA=\/0x0:350x560\/320x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/jDQOu-zcZELObMYeh0IIi8IlIH0=\/0x0:350x560\/520x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/dc9vpvCkwJeowST-LpeYzc7elrk=\/0x0:350x560\/720x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/Oxu2aC90eCrI4vN5kxahE4A2c3o=\/0x0:350x560\/920x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/NGER9h5qBq_HNpxCX2C_yObR0Cw=\/0x0:350x560\/1120x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/imVnySBSlK_wltOorNmBar-T9AY=\/0x0:350x560\/1320x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/hPDABDjNLOdAwOEmQEUEzIRR8I0=\/0x0:350x560\/1520x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/ZCNrsaIrket_4eV720qZSwLBqSY=\/0x0:350x560\/1720x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/7UhktI0PWr-VoEYDQvVJ2HSfFpA=\/0x0:350x560\/1920x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg 1920w\" alt=\"a renaissance man\u2019s head turns to smoke in prosper\u2019s demon book cover\" data-upload-width=\"350\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/_xtfn8NuaialQcrne3akQH0hDDs=\/0x0:350x560\/1200x0\/filters:focal(0x0:350x560):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573034\/9781250260505.jpg\" referrerpolicy=\"no-referrer\" \/><cite>Image: Tor.com<\/cite><\/figure><\/div><p id=\"9frj2N\">In his latest novella, K.J. Parker takes readers to a world inspired by Renaissance Europe, following a narrator who can communicate with demons. He\u2019s been tasked with the Church to perform exorcisms. He doesn\u2019t particularly care what happens to the host, just so long as the demon\u2019s out. When a demon he knows is attracted to a princess\u2019s unborn child, he ventures to a palace, only to discover that one of the world\u2019s most brilliant individuals, Prosper of Schanz, is himself infected \u2014 and that the demon is probably responsible for his scientific and artistic genius, putting everyone into a serious moral quandary. <a href=\"https:\/\/www.publishersweekly.com\/978-1-250-26051-2\"><em>Publishers Weekly<\/em> says that<\/a> \u201cthe engrossing narrative voice and assured worldbuilding compensate for a slight plot and rushed conclusion.\u201d<\/p><hr class=\"p-entry-hr\" id=\"4x3m7I\" \/><div class=\"c-float-left\"><figure class=\"e-image\"><img srcset=\"https:\/\/cdn.vox-cdn.com\/thumbor\/suBPEhpA6c8C-APlOW-WacMHuK8=\/0x0:996x1500\/320x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 320w, https:\/\/cdn.vox-cdn.com\/thumbor\/rnUoC3qt8ZCGW9L80-V-_EtzdAE=\/0x0:996x1500\/520x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 520w, https:\/\/cdn.vox-cdn.com\/thumbor\/CrqV3gq_qgdRqOXnYrMGQm6dmNc=\/0x0:996x1500\/720x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 720w, https:\/\/cdn.vox-cdn.com\/thumbor\/B4sBuWah9yptM-cwp8uFPUAwc1Y=\/0x0:996x1500\/920x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 920w, https:\/\/cdn.vox-cdn.com\/thumbor\/RKAV2Ht7pr3zBnIvO0dYWO1xT7Q=\/0x0:996x1500\/1120x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 1120w, https:\/\/cdn.vox-cdn.com\/thumbor\/aNZ4FQftdRPfQoQ7i0lTOaZm9_o=\/0x0:996x1500\/1320x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 1320w, https:\/\/cdn.vox-cdn.com\/thumbor\/wjVGcUE1rFlXUM1wWyYOrsFAfuo=\/0x0:996x1500\/1520x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 1520w, https:\/\/cdn.vox-cdn.com\/thumbor\/ZtjOzr_awjzxbFxv3O938T6F40Y=\/0x0:996x1500\/1720x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 1720w, https:\/\/cdn.vox-cdn.com\/thumbor\/CIVi0UYSngxa9Hgw0fJCtNejeXo=\/0x0:996x1500\/1920x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image 1920w\" alt=\"interior chinatown cover has a chinese temple with exploding cartoons\" data-upload-width=\"996\" src=\"https:\/\/cdn.vox-cdn.com\/thumbor\/xeMI4m6FuYPBGKDdblBXDyY--7Y=\/0x0:996x1500\/1200x0\/filters:focal(0x0:996x1500):no_upscale()\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19573037\/image\" referrerpolicy=\"no-referrer\" \/><cite>Image: Pantheon<\/cite><\/figure><\/div><p id=\"q7IbWO\">Charles Yu has been one of my favorite novelists ever since I picked up his debut book <em>How to Live Safely in a Science Fictional Universe<\/em> a decade ago. His next, <em>Interior Chinatown<\/em>, looks to be just as wonderfully meta and strange as that first one. It follows as young actor named Willis Wu who aspires to play more than just token Asian roles in the backgrounds of films and TV productions. <a href=\"https:\/\/www.kirkusreviews.com\/book-reviews\/charles-yu\/interior-chinatown\/\"><em>Kirkus Reviews<\/em> gave the book a starred review<\/a>, saying that it\u2019s \u201can acid indictment of Asian stereotypes and a parable for outcasts feeling invisible in this fast-moving world.\u201d<\/p>","created_at":"2020-01-06T19:26:41+0000","updated_at":"2020-01-06T19:26:46+0000","published_at":"2020-01-04T11:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"www.polygon.com","preview_picture":"https:\/\/cdn.vox-cdn.com\/thumbor\/ulFQ04Gg11yxI5q_Ww7UIBOPzSM=\/0x146:2040x1214\/fit-in\/1200x630\/cdn.vox-cdn.com\/uploads\/chorus_asset\/file\/19571647\/aliptak_200103_ply0949_8912.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1083,"title":"Strategy Index - Guides for Dominions 5 :: Dominions 5 General Discussions","url":"https:\/\/steamcommunity.com\/app\/722060\/discussions\/0\/1636417554423298189\/","content":"\u201cKnowledge is knowing ... or where to find out.\u201d<br \/>- Alvin Toffler<p>Similar to the threads for Dominions 3 and 4, this is a quick reference list to Dominions guides.<\/p><p>Did I miss valuable references? Just post them here or PM me, so I can add them.<\/p><p>Table of Contents<\/p><ol><li>Strategy Indices &amp; Information Sources<br \/><\/li>\n<li>Dominions: Why play it? What is new in Dominions 5 and is it worth it?<br \/><\/li>\n<li>Tools of the Trade &amp; Hotkeys<br \/><\/li>\n<li>Beginner Guides<br \/><\/li>\n<li>Nation, Pretender and Scales Guides<br \/><\/li>\n<li>General Strategy &amp; Guides<br \/><\/li>\n<li>Magic Sites and Site Searching<br \/><\/li>\n<li>Magic, Spells and items<br \/><\/li>\n<li>Combat, Thugs and Super Combatants<br \/><\/li>\n<li>Mods &amp; Maps<br \/><\/li>\n<li>Multiplayer<br \/><\/li>\n<li>Informative YouTube Channels<\/li>\n<\/ol><p>1. Strategy Indices and other Information Sources<\/p><br \/>Some games have manuals you can ignore. Dominions isn't one of them.<p><strong>Manuals<\/strong><br \/><\/p><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/www.illwinter.com\/dom4\/\" target=\"_blank\" rel=\"noreferrer\">Dominions 4 Manual<\/a>[www.illwinter.com] (explains a lot of Dominion concepts, with in depth spell descriptions &amp; formulas, dated)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/www.illwinter.com\/dom5\/docs.html\" target=\"_blank\" rel=\"noreferrer\">Dominions 5 Manual<\/a>[www.illwinter.com] (updated for Dominons 5, but lacks detailed spell descriptions)<\/li>\n<\/ul><strong>Strategy Indices<\/strong><br \/><strong>Information Sources<\/strong><p>2. Dominions: Why play it? What is new in Dominions 5 and is it worth it?<\/p><br \/><strong>Why Dominions?<\/strong><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/www.pcgamer.com\/part-one-one-rocks-fight-to-become-top-deity\/\" target=\"_blank\" rel=\"noreferrer\">One rocks fight to become the top deity<\/a> (by Tom Senior)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.pcgamer.com\/dominions-5-gods-go-to-war-in-4x-showdown-for-the-ages\/\" target=\"_blank\" rel=\"noreferrer\">Gods go to war<\/a> (by Tom Senior)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.rockpapershotgun.com\/2013\/10\/18\/wot-i-think-dominions-4-single-player\/\" target=\"_blank\" rel=\"noreferrer\">Wot I Think: Dominions 4 Single Player<\/a> (by Adam Smith)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=https:\/\/devnada.wordpress.com\/2014\/10\/06\/drama-mama-introduction\/\" target=\"_blank\" rel=\"noreferrer\">Elvira: Drama Mama<\/a>[devnada.wordpress.com]<\/li>\n<\/ul><br \/><strong>What is new?<\/strong><br \/><strong>Is it worth it?<\/strong><p>Yes! If you are short on cash, wait for one of the <a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=https:\/\/steamdb.info\/app\/722060\/\" target=\"_blank\" rel=\"noreferrer\">Steam Sales<\/a>[steamdb.info], the best was 40% off.<br \/><\/p><p>3. Tools of the Trade &amp; Hotkeys<\/p><br \/>Software to help your daily Dominions needs:<br \/><strong>A Short Guide to Hotkeys<\/strong><p>Hotkeys can save you a lot of clicks and time! To view all available hotkeys of a screen press ?<br \/>The list below contains a few shortcuts, for the full list see <a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/www.illwinter.com\/dom4\/dom4_quickref.pdf\" target=\"_blank\" rel=\"noreferrer\">Illwinter's Quick Reference Guide pdf<\/a>[www.illwinter.com]<\/p><ul class=\"bb_ul\"><li><strong>Pretender Creation (under Main Menu &gt; Tools\/Manuals):<\/strong> b: view all blesses, s\/F5: view research \/ spells, n: see national units<br \/><\/li>\n<li><strong>Campaign Map - Menus<\/strong> F1: empire overview, F2: score graphs, F5: Research, F8: item treasury<br \/><\/li>\n<li><strong>Campaign Map - Provinces<\/strong> c: province combat reports, v: see province battles, r: recruitment, t: army setup, b: mercenaries<br \/><\/li>\n<li><strong>Campaign Map - Commanders<\/strong> n: select next idle commander, a: select all commanders, ctrl-click: add commander to selection<br \/><\/li>\n<li><strong>Campaign Map - Gem\/slave transfer, PD:<\/strong> +\/- change amount by 1, Shift +\/- change by 5<br \/><\/li>\n<li><strong>Campaign Map - Blood Hunting:<\/strong> V: transfers blood slave to selected commander, Z: places all slaves of commander into lab<br \/><\/li>\n<li><strong>Campaign Map - Unit Orders:<\/strong> R: Research, W: wait (wont show up as idle), S: site search, F: forge item, C: cast spells<br \/><\/li>\n<li><strong>Army Setup:<\/strong> Ctrl + 1-9: (saves orders under mouse pointer), 1-9 assign orders, g: guard commander, a: attack, h: hold &amp; attack, C: closest, r: rear<br \/><\/li>\n<li><strong>Battle orders:<\/strong> x: copy last order<br \/><\/li>\n<li><strong>Combat:<\/strong> c: colored squares, 1-3: detailed combat logs, v: see selected units combat log<\/li>\n<\/ul><p>4. Beginner Guides<\/p><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=https:\/\/dom4.fandom.com\/wiki\/Beginner%27s_Guide\" target=\"_blank\" rel=\"noreferrer\">Wiki: Beginner's Guide<\/a>[dom4.fandom.com]<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=599297365\" target=\"_blank\" rel=\"noreferrer\">Beginner-Intermediate Guide to Dom4<\/a> (by Reekris)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/watch?v=zGZxIKWarF0\" target=\"_blank\" rel=\"noreferrer\">\u201cA complete game in one sitting\u201d<\/a> (Youtube Intro to Dominions 5 by DasTactic)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/watch?v=vkQEoAnFqY4&amp;list=PLGB6RkFB7ZmNDvUGIMQZLrLs7NzTl0DzD\" target=\"_blank\" rel=\"noreferrer\">Dominions 5 Beginner's Guide<\/a> (Very detailed tutorial videos on many aspects of Dominions by DasTactic)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/old.reddit.com\/r\/Dominions5\/comments\/9lzrdf\/tips_for_expansion\/\" target=\"_blank\" rel=\"noreferrer\">Tips for expansion<\/a><\/li>\n<\/ul><p>5. Nation, Pretender and Scales Guides<\/p><br \/><strong>Nation Guide Collections<\/strong><br \/><strong>Pretender Design<\/strong><br \/><strong>Scales<\/strong><p>6. General Strategy &amp; Guides<\/p><br \/><strong>General Strategy<\/strong><br \/><strong>General Guides<\/strong><p>7. Magic Sites and Site Searching<\/p><p>8. Magic, Spells and Items<\/p><br \/><strong>Magic Primer<\/strong><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=692390381\" target=\"_blank\" rel=\"noreferrer\">Basic Magic, a Primer<\/a> (by Shinuyama)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=667432694\" target=\"_blank\" rel=\"noreferrer\">Blood Magic Primer<\/a> (by Shinuyama)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/sharedfiles\/filedetails\/?id=814311765\" target=\"_blank\" rel=\"noreferrer\">Communion Guide<\/a> (must read if you want to use communions \/ sabbath, <strong>Note for Dom5:<\/strong> slaves now always take 1 damage for master spells when over 200 fatigue and fatigue can kill mages when casting big spells!)<\/li>\n<\/ul><strong>Spells by Use<\/strong><br \/><strong>Spells by School<\/strong><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/watch?v=i7gYWqq-5TU&amp;list=PLunmZfRvOVBLIvyKhe1kvWJAPdJAXEqwk\" target=\"_blank\" rel=\"noreferrer\">Dominions 5: Spell Schools Videos<\/a> (Yakri evaluates spells, sorted by schools)<\/li>\n<\/ul><strong>Spells in Combat<\/strong><br \/><strong>Items<\/strong><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=https:\/\/docs.google.com\/document\/d\/10gLUCpnErweG6naoRvperf4hwGJrz_ckExot1h3bMt8\/edit\" target=\"_blank\" rel=\"noreferrer\">Dominions 5: Magic Booster &amp; Path Access<\/a>[docs.google.com] (by MelficeBelmont)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.moddb.com\/games\/dominions-4-thrones-of-ascensions\/forum\/thread\/dom4-magic-path-booster-guide-by-telos\" target=\"_blank\" rel=\"noreferrer\">Dominions 4: Magic Path Booster &amp; Mage Summons by telos<\/a> (Astral Rings now S6\/S7, Robe of Magi A5B5)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.reddit.com\/r\/Dominions4\/comments\/6dbpdx\/list_of_temporary_gem_items\/\" target=\"_blank\" rel=\"noreferrer\">Temporary Gem Items<\/a><\/li>\n<\/ul><p>9. Combat, Thugs and Super Combatants<\/p><br \/><strong>Combat<\/strong><br \/><strong>Thugs and Super Combatants<\/strong><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=https:\/\/dom4.wikia.com\/wiki\/Building_thugs_and_SCs\" target=\"_blank\" rel=\"noreferrer\">Wiki: Building Thugs &amp; SCs<\/a>[dom4.wikia.com] (Spells, items and example chassis)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=https:\/\/forum.shrapnelgames.com\/showthread.php?t=41970\" target=\"_blank\" rel=\"noreferrer\">Zen and the art of Thugging<\/a>[forum.shrapnelgames.com] (Building and countering Thugs &amp; SCs by baalz)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/watch?v=Y9qaBN1KaRU\" target=\"_blank\" rel=\"noreferrer\">Zen and the art of Thugging<\/a> (Youtube video by Lucid)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/watch?v=GTgWX8o5fzA\" target=\"_blank\" rel=\"noreferrer\">Killing Super Combatants - 101<\/a> (Youtube video by Lucid)<\/li>\n<\/ul><p>10. Mods &amp; Maps<\/p><br \/><strong>Catalog and Tools<\/strong><br \/><strong>Popular Mods<\/strong><br \/><ul class=\"bb_ul\"><li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/dominionsmods.com\/index.php?showtopic=3838\" target=\"_blank\" rel=\"noreferrer\">Worthy Heroes<\/a>[dominionsmods.com] (\u201cA popular mod buffing weaker heroes and adding many new heroes\")<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/dominionsmods.com\/index.php?showtopic=3225\" target=\"_blank\" rel=\"noreferrer\">Pretender Extended<\/a>[dominionsmods.com] (\u201dA pair of mods adding over 100 new Pretender chassis to the game, with Pretenders Enhanced also providing rebalancing of existing Pretenders.\")<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/dominionsmods.com\/index.php?showtopic=3175\" target=\"_blank\" rel=\"noreferrer\">Lucid's Thematic Gem Gen<\/a>[dominionsmods.com] (\"A mod adding various tempgem items to each path\" reducing gem micro)<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/dominionsmods.com\/index.php?showtopic=3945\" target=\"_blank\" rel=\"noreferrer\">AI No Recruit<\/a>[dominionsmods.com] (\u201dA mod preventing the AI from recruiting most independent and weak national units.\")<br \/><\/li>\n<li><a class=\"bb_link\" href=\"https:\/\/steamcommunity.com\/linkfilter\/?url=http:\/\/dominionsmods.com\/index.php?showtopic=3169\" target=\"_blank\" rel=\"noreferrer\">Debug Mod<\/a>[dominionsmods.com] (For testing purposes, all spells are researched and Wish \/ Divine Name are 1 gem)<\/li>\n<\/ul><br \/><strong>Maps<\/strong><p>11. Multiplayer<\/p><p>12. Informative YouTube Channels<\/p><p>Active<\/p><br \/><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/channel\/UCwURCFaZsdynkjjOuI_RV0A\" target=\"_blank\" rel=\"noreferrer\">Attica<\/a><br \/>Dom5 Basic Nation Analysis: EA Machaka, Mictlan, Ur, MA Marignon<p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/channel\/UCRyuNpdX1JvvbHmgmCeehFw\" target=\"_blank\" rel=\"noreferrer\">GeneralConfusion<\/a> (Dom5 SP and MP for beginners and intermediate players)<br \/>Dom5: EA Fomoria, Hinnom, Lanka, Machaka, Mekone, Rlyeh, Ulm; MA Ashdod, Machaka, Marignon, Phlegra, Tien Chi; LA Jomon, Phlegra<br \/>Basic Nation Analysis: many<\/p><p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/user\/lucidsentience\" target=\"_blank\" rel=\"noreferrer\">LucidTactics:<\/a> (Multiplayer Action for intermediate and advanced players, Nation Analysis &amp; MP strategy videos)<br \/>Dom4: EA Hinnom, Tien Chi; MA Agartha, Ermor<br \/>Dom5: EA Marverni, Yomi; MA Agartha, Bandar Log, Ctis, Machaka, Sceleria, Ulm; LA Lemuria, Mictlan, Ulm<\/p><p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/user\/darkchampi0n\/playlists\" target=\"_blank\" rel=\"noreferrer\">Nuclearmonkee<\/a> (Multiplayer action for intermediate and advanced players)<br \/>Dom4: MA Marignon, Xibalba; LA Patala, Pythium<br \/>Dom5: EA Sauromatia; MA Ctis, Ryleh, Phaecia; LA Utgard<\/p><p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/user\/darkchampi0n\/playlists\" target=\"_blank\" rel=\"noreferrer\">Perun<\/a> (Multiplayer action for intermediate players)<br \/>Dom5: MA Nazca, LA Atlantis<\/p><p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/user\/tokshen\/playlists\" target=\"_blank\" rel=\"noreferrer\">Tokshen<\/a> (many singleplayer roleplay series)<\/p><p>Inactive<\/p><br \/><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/user\/marcusaureliuslp\/playlists\" target=\"_blank\" rel=\"noreferrer\">Marcus Aurelius<\/a>(Dom4 &amp; Dom5 singleplayer roleplay series)<p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/user\/Alcoway\/\" target=\"_blank\" rel=\"noreferrer\">Sy<\/a> (Multiplayer Action for intermediate and advanced players)<br \/>Dom4: LA Agartha, Lemuria,<br \/>Dom5: EA Mictlan, MA Nazca<\/p><p><a class=\"bb_link\" href=\"https:\/\/www.youtube.com\/channel\/UC2g9ydQ3fIVvpWA9uLIpb-Q\" target=\"_blank\" rel=\"noreferrer\">Yakri:<\/a> (Bless and Spell tutorials)<br \/>Dom5 Nation Analysis: EA Ctis, Kailasa, Lanka, Yomi, MA Ermor, Ryleh, Shinuyama, LA Ulm<br \/><\/p>","created_at":"2020-01-06T15:27:45+0000","updated_at":"2020-01-06T15:27:45+0000","published_by":["30","50"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"steamcommunity.com","preview_picture":"https:\/\/steamcdn-a.akamaihd.net\/steamcommunity\/public\/images\/apps\/722060\/9fe595931d63cc7f45b2b6eae9d1d2d2da6f2eb7.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1082,"title":"Notes on Technical Writing","url":"https:\/\/mkaz.blog\/misc\/notes-on-technical-writing\/","content":"<p>Over the past year, I\u2019ve worked on and off documentation for WordPress. I started contributing during a freeze around launch to help developers transition to the new platform. I found writing documentation is something I enjoy, plus rewarding to help and educate people. Though it\u2019s not a primary part of my job, I\u2019ve continued to find time here and there to keep contributing.<\/p><p>In this time, I\u2019ve read various resources on technical writing and documentation. These are my notes, both to help me remember later, but also as a tool to help me think about writing now.<\/p>\n<h3>Principles<\/h3>\n<p>The following principles I try to keep in mind when writing documentation.<\/p>\n<ul><li>The purpose of technical writing is to help users accomplish tasks as quickly and effectively as possible.<\/li>\n<li>People learn by doing, prefer to be shown and not told.<\/li>\n<li>Get users to their first success quickly.<\/li>\n<li>There is more than one type of documentation.<\/li>\n<li>Keep it simple, write in plain language<\/li>\n<\/ul><h3>Tips<\/h3>\n<ul><li>Know your audience, know your purpose<\/li>\n<li>Put the most important information first<\/li>\n<li>Use bullet lists<\/li>\n<li>One idea per paragraph<\/li>\n<li>Edit, Edit, Edit<\/li>\n<\/ul><h4>Avoid Meta Writing<\/h4>\n<p>Avoid writing about the writing, text does not have a mind of its own. The active parties are the writer and the reader.<\/p>\n<p>You don\u2019t need to tell the reader what you\u2019re going to tell them. Just tell them.<\/p>\n<ul><li><strong>Bad<\/strong>: This chapter discusses the factors that cause names to rise and fall in popularity.<\/li>\n<li><strong>Good<\/strong>: What makes a name rise and fall in popularity?<\/li>\n<\/ul><p>Don\u2019t refer to sections or paragraphs in the writing.<\/p>\n<ul><li><strong>Bad<\/strong>: Summarizing the preceding section about\u2026<\/li>\n<li><strong>Good<\/strong>: In summary, \u2026<\/li>\n<\/ul><p>Don\u2019t tell the reader something is extremely complex; likewise don\u2019t tell the reader something is simple. The reader will make up their own mind about a concept being easy or hard.<\/p>\n<h2>Minimalist Instruction<\/h2>\n<p>John M. Carroll and his colleagues at IBM studied creating support documents analyzing the best way to get a user proficient. In the 1980s, they developed a minimalist approach and created a wonderful <a href=\"https:\/\/mkaz.blog\/wp-content\/uploads\/2020\/01\/IBM-Displaywriter-Manual.pdf\">Operator Training Manual for the IBM Displaywriter<\/a>, imagine having to introduce someone to using a word processor for the first time.<\/p>\n<p>The four principals of minimalist instruction:<\/p>\n<ul><li>Choose an action-oriented approach. Users typically want to do things. This principle reflects the user-centric nature of minimalism.<\/li>\n<li>Anchor the tool in the task domain. A tool is a means to an end. This principle asks designers to select training tasks that are meaningful for the user.<\/li>\n<li>Support error recognition and recovery. To err is human. There are several ways to increase user competence and comfort levels in handling mistakes.<\/li>\n<li>Support reading to do, study and locate. Designs must fit as much as possible the diverging needs and propensities of the intended audience. This principle reflects the user-centric nature of minimalism.<\/li>\n<\/ul><p><a href=\"https:\/\/www.utwente.nl\/en\/bms\/ist\/minimalism\/\">Read deeper about Minimalist Instruction<\/a> from the Department of Instructional Technology at University of Twente, Netherlands.<\/p>\n<h2>Constructivism<\/h2>\n<p>American psychologist Jerome Bruner studied human cognitive psychology. A major theme in Bruner\u2019s work is that learning is an active process in which learners construct new ideas based upon their own knowledge. Applying the principles of constructivism to documentation:<\/p>\n<ul><li>Try to encourage readers to discover principles by themselves<\/li>\n<li>Translate information into a format appropriate to the reader\u2019s current level<\/li>\n<li>Organize material in a spiral, so it continually builds upon what they have already learned.<\/li>\n<li>Engage in an active dialog, invite users to act \u201cPlease try\u201d, \u201cSee what happens\u201d, \u201cExplore yourself\u201d<\/li>\n<\/ul><p>Read more about <a href=\"https:\/\/www.instructionaldesign.org\/theories\/constructivist\/\">Bruner and Constructivist Theory<\/a><\/p>\n<h2>Types of Documentation<\/h2>\n<p>Daniele Procida wrote <a href=\"https:\/\/www.divio.com\/blog\/documentation\/\">What nobody tells you about documentation<\/a> that discusses there isn\u2019t just one thing called documentation, but there (at least) four types of documentation.<\/p>\n<ol><li><strong>Tutorials<\/strong> \u2013 A tutorial is learning-oriented for newcomers, a directed set of lessons to give the user basic confidence and skills. A tutorial should not assume knowledge or language, but is there to take a beginner who doesn\u2019t know what questions to ask through learning.<\/li>\n<li><strong>How-to Guides<\/strong> \u2013 A how-to guide, or FAQ is goal-oriented, directed by the user. A how-to guide can assume the user has basic knowledge and language, but needs to know how to solve a specific problem.<\/li>\n<li><strong>Explanation<\/strong> \u2013 An explanation or overview document is understanding-oriented providing background or additional detailed context.<\/li>\n<li><strong>Reference Guides<\/strong> \u2013 A reference guide is information-oriented, it is accurate and complete describing the machinery. API documentation is an example of a reference guide.<\/li>\n<\/ol><h2>Resources<\/h2>\n<ul><li><a href=\"https:\/\/amzn.to\/2ZNqUQl\">On Writing Well<\/a> (book) The classic guide to writing non-fiction by William Zinsser.<\/li>\n<li><a href=\"http:\/\/andyetter.com\/2018\/06\/14\/basic-writing\/\">Basic Writing Tips<\/a> an article by Andrew Etter<\/li>\n<li><a href=\"http:\/\/advice.writing.utoronto.ca\/wp-content\/uploads\/sites\/2\/wordiness.pdf\">Wordiness: Danger Signals<\/a> (2 page pdf) by Margaret Procter, University of Toronto<\/li>\n<li><a href=\"https:\/\/alistapart.com\/article\/writing-is-thinking\/\">Writing is Thinking<\/a> by Sally Kerrigan, A List Apart<\/li>\n<li><a href=\"https:\/\/blog.stephsmith.io\/learning-to-write-with-confidence\/\">Writing is Thinking<\/a>, by Steph Smith<\/li>\n<li><a href=\"https:\/\/www.nature.com\/articles\/d41586-019-02918-5\">Tips on Writing a Great Science Paper<\/a> from Cormac McCarthy<\/li>\n<li><a href=\"https:\/\/docs.microsoft.com\/en-us\/style-guide\/welcome\/\">Microsoft Writing Style Guide<\/a><\/li>\n<\/ul>","created_at":"2020-01-04T18:37:42+0000","updated_at":"2020-01-04T18:37:42+0000","published_at":"2020-01-04T15:04:12+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"mkaz.blog","preview_picture":"https:\/\/149354403.v2.pressablecdn.com\/wp-content\/uploads\/2018\/04\/cropped-mkaz-blog.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["systemd"],"is_public":false,"id":1081,"title":"Systemd: Service File Examples - ShellHacks","url":"https:\/\/www.shellhacks.com\/systemd-service-file-example\/","content":"<p>Most Linux distributions use systemd as a system and service manager.<\/p><p>The <code>systemctl<\/code> is the main command in systemd, used to control services.<\/p>\n<p>In this tutorial i will show how to create a systemd service file that will allow you to control your service using the <code>systemctl<\/code> command, how to restart systemd without reboot to reload unit files and how to enable your new service.<\/p>\n<p>I will also show and describe the most important systemd service file options with the live examples of the systemd service files.<\/p>\n<h2>Create Systemd Service File<\/h2>\n<p>Create a systemd service file <code>\/etc\/systemd\/system\/foo-daemon.service<\/code> (replace the <code>foo-daemon<\/code> with your service name):<\/p>\n<pre>$ sudo touch \/etc\/systemd\/system\/foo-daemon.service\n$ sudo chmod 664 \/etc\/systemd\/system\/foo-daemon.service<\/pre>\n<p>Open the <code>foo-daemon.service<\/code> file and add the minimal service configuration options that allow this service to be controlled via <code>systemctl<\/code>:<\/p>\n<pre>\n[Unit]\nDescription=Foo\n[Service]\nExecStart=\/usr\/sbin\/foo-daemon\n[Install]\nWantedBy=multi-user.target\n<\/pre>\n<p class=\"note-blue\"><strong>Path To Daemon<\/strong>: If you don\u2019t know the full path to a daemon, try <code>which foo-daemon<\/code>.<\/p>\n<p>Once the service file is changed, it needs to reload systemd configuration:<\/p>\n<pre>$ sudo systemctl daemon-reload<\/pre>\n<p>Now you should be able to <code>start<\/code>, <code>stop<\/code>, <code>restart<\/code> and check the service <code>status<\/code><\/p>\n<pre>$ sudo systemctl start foo-daemon\n$ sudo systemctl stop foo-daemon\n$ sudo systemctl restart foo-daemon\n$ systemctl status foo-daemon\n<\/pre>\n<p>To configure a service to start automatically on boot, you need to <code>enable<\/code> it:<\/p>\n<pre>$ sudo systemctl enable foo-daemon<\/pre>\n<p>To check the service logs, run:<\/p>\n<pre>$ journalctl -u foo-daemon<\/pre>\n<h2>Systemd Service File Options<\/h2>\n<p>Systemd service files typically consist of three sections.<\/p>\n<p>The common configuration items are configured in the generic <code>[Unit]<\/code> and <code>[Install]<\/code> sections.<\/p>\n<p>The service specific configuration options are configured in the <code>[Service]<\/code> section.<\/p>\n<h3>Important <code>[Unit]<\/code> Section Options<\/h3>\n<table class=\"ptable\"><tbody><tr><th>Option<\/th>\n<th>Description<\/th>\n<\/tr><tr><td><code>Description<\/code><\/td>\n<td>A short description of the unit.<\/td>\n<\/tr><tr><td><code>Documentation<\/code><\/td>\n<td>A list of URIs referencing documentation.<\/td>\n<\/tr><tr><td><code>Before<\/code>, <code>After<\/code><\/td>\n<td>The order in which units are started.<\/td>\n<\/tr><tr><td><code>Requires<\/code><\/td>\n<td>If this unit gets activated, the units listed here will be activated as well. If one of the other units gets deactivated or fails, this unit will be deactivated.<\/td>\n<\/tr><tr><td><code>Wants<\/code><\/td>\n<td>Configures weaker dependencies than <code>Requires<\/code>. If any of the listed units does not start successfully, it has no impact on the unit activation. This is the recommended way to establish custom unit dependencies.<\/td>\n<\/tr><tr><td><code>Conflicts<\/code><\/td>\n<td>If a unit has a <code>Conflicts<\/code> setting on another unit, starting the former will stop the latter and vice versa.<\/td>\n<\/tr><\/tbody><\/table><p>A complete list of <code>[Unit]<\/code> section options:<\/p>\n<pre>$ man systemd.unit<\/pre>\n<h3>Important <code>[Install]<\/code> Section Options<\/h3>\n<table class=\"ptable\"><tbody><tr><th>Option<\/th>\n<th>Description<\/th>\n<\/tr><tr><td><code>Alias<\/code><\/td>\n<td>A space-separated list of additional names for the unit. Most <code>systemctl<\/code> commands, excluding <code>systemctl enable<\/code>, can use aliases instead of the actual unit name.<\/td>\n<\/tr><tr><td><code>RequiredBy<\/code>, <code>WantedBy<\/code><\/td>\n<td>The current service will be started when the listed services are started. See the description of <code>Wants<\/code> and <code>Requires<\/code> in the <code>[Unit]<\/code> section for details.<\/td>\n<\/tr><tr><td><code>Also<\/code><\/td>\n<td>Specifies a list of units to be enabled or disabled along with this unit when a user runs <code>systemctl enable<\/code> or <code>systemctl disable<\/code>.<\/td>\n<\/tr><\/tbody><\/table><p>A complete list of <code>[Install]<\/code> section options:<\/p>\n<pre>$ man systemd.unit<\/pre>\n<h3>Important <code>[Service]<\/code> Section Options<\/h3>\n<table class=\"ptable\"><tbody><tr><th>Option<\/th>\n<th>Description<\/th>\n<\/tr><tr><td><code>Type<\/code><\/td>\n<td>Configures the process start-up type. One of:<br \/><code>simple<\/code> (default) \u2013 starts the service immediately. It is expected that the main process of the service is defined in <code>ExecStart<\/code>.<br \/><code>forking<\/code> \u2013 considers the service started up once the process forks and the parent has exited.<br \/><code>oneshot<\/code> \u2013 similar to <code>simple<\/code>, but it is expected that the process has to exit before systemd starts follow-up units (useful for scripts that do a single job and then exit). You may want to set <code>RemainAfterExit=yes<\/code> as well so that systemd still considers the service as active after the process has exited.<br \/><code>dbus<\/code> \u2013 similar to <code>simple<\/code>, but considers the service started up when the main process gains a D-Bus name.<br \/><code>notify<\/code> \u2013 similar to <code>simple<\/code>, but considers the service started up only after it sends a special signal to systemd.<br \/><code>idle<\/code> \u2013 similar to <code>simple<\/code>, but the actual execution of the service binary is delayed until all jobs are finished.<\/td>\n<\/tr><tr><td><code>ExecStart<\/code><\/td>\n<td>Commands with arguments to execute when the service is started. <code>Type=oneshot<\/code> enables specifying multiple custom commands that are then executed sequentially. <code>ExecStartPre<\/code> and <code>ExecStartPost<\/code> specify custom commands to be executed before and after <code>ExecStart<\/code>.<\/td>\n<\/tr><tr><td><code>ExecStop<\/code><\/td>\n<td>Commands to execute to stop the service started via <code>ExecStart<\/code>.<\/td>\n<\/tr><tr><td><code>ExecReload<\/code><\/td>\n<td>Commands to execute to trigger a configuration reload in the service.<\/td>\n<\/tr><tr><td><code>Restart<\/code><\/td>\n<td>With this option enabled, the service shall be restarted when the service process exits, is killed, or a timeout is reached with the exception of a normal stop by the <code>systemctl stop<\/code> command.<\/td>\n<\/tr><tr><td><code>RemainAfterExit<\/code><\/td>\n<td>If set to <code>True<\/code>, the service is considered active even when all its processes exited. Useful with <code>Type=oneshot<\/code>. Default value is <code>False<\/code>.<\/td>\n<\/tr><\/tbody><\/table><p>A complete list of <code>[Service]<\/code> section options:<\/p>\n<pre>$ man systemd.service<\/pre>\n<h2>Systemd Service File Examples<\/h2>\n<pre>\n[Unit]\nDescription=The NGINX HTTP and reverse proxy server\nAfter=syslog.target network.target remote-fs.target nss-lookup.target\n[Service]\nType=forking\nPIDFile=\/run\/nginx.pid\nExecStartPre=\/usr\/sbin\/nginx -t\nExecStart=\/usr\/sbin\/nginx\nExecReload=\/bin\/kill -s HUP $MAINPID\nExecStop=\/bin\/kill -s QUIT $MAINPID\nPrivateTmp=true\n[Install]\nWantedBy=multi-user.target\n<\/pre>\n<pre>\n[Unit]\nDescription=The Apache HTTP Server\nAfter=network.target remote-fs.target nss-lookup.target\n[Service]\nType=notify\nEnvironmentFile=\/etc\/sysconfig\/httpd\nExecStart=\/usr\/sbin\/httpd $OPTIONS -DFOREGROUND\nExecReload=\/usr\/sbin\/httpd $OPTIONS -k graceful\nExecStop=\/bin\/kill -WINCH ${MAINPID}\nKillSignal=SIGCONT\nPrivateTmp=true\n[Install]\nWantedBy=multi-user.target\n<\/pre>\n<pre>\n[Unit]\nDescription=Redis persistent key-value database\nAfter=network.target\n[Service]\nExecStart=\/usr\/bin\/redis-server \/etc\/redis.conf --daemonize no\nExecStop=\/usr\/bin\/redis-shutdown\nUser=redis\nGroup=redis\n[Install]\nWantedBy=multi-user.target\n<\/pre>\n<p>For more examples, check the <a href=\"https:\/\/www.freedesktop.org\/software\/systemd\/man\/systemd.service.html\" target=\"_blank\">systemd.service<\/a> and <a href=\"https:\/\/www.freedesktop.org\/software\/systemd\/man\/systemd.unit.html\" target=\"_blank\">systemd.unit<\/a> man pages.<\/p>","created_at":"2020-01-04T12:54:07+0000","updated_at":"2020-01-04T12:54:07+0000","published_at":"2018-03-20T13:02:32+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"www.shellhacks.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","configuration","init.el"],"is_public":false,"id":1080,"title":"Ladicle's Emacs Configuration","url":"https:\/\/ladicle.com\/post\/config\/#screenshot","content":"<h3 id=\"table-of-contents\">1.1 Table of Contents<\/h3><h3 id=\"screenshot\">1.2 ScreenShot<\/h3><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(setq user-emacs-directory \"~\/Developments\/src\/github.com\/Ladicle\/dotfiles\/common\/emacs.d\/elisp\")\n(eval-and-compile\n  (setq load-prefer-newer t\n        package-user-dir \"~\/Developments\/src\/github.com\/Ladicle\/dotfiles\/common\/emacs.d\/elpa\"\n        package--init-file-ensured t\n        package-enable-at-startup nil)\n  (unless (file-directory-p package-user-dir)\n    (make-directory package-user-dir t)))\n(eval-and-compile\n  (setq load-path (append load-path (directory-files package-user-dir t \"^[^.]\" t))))\n(eval-when-compile\n  (require 'package)\n  (package-initialize)\n  (add-to-list 'package-archives '(\"gnu\" . \"https:\/\/elpa.gnu.org\/packages\/\") t)\n  (add-to-list 'package-archives '(\"melpa\" . \"https:\/\/melpa.org\/packages\/\") t)\n  (add-to-list 'package-archives '(\"org\" . \"http:\/\/orgmode.org\/elpa\/\") t)\n  (unless (package-installed-p 'use-package)\n    (package-refresh-contents)\n    (package-install 'use-package)\n    (package-install 'diminish)\n    (package-install 'quelpa)\n    (package-install 'bind-key))\n  (setq use-package-always-ensure t)\n  (setq use-package-expand-minimally t)\n  (require 'use-package))\n(require 'diminish)\n(require 'bind-key)<\/pre><\/div><h3 id=\"generic\">1.4 Generic<\/h3><h4 id=\"server\">1.4.1 Server<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package server\n  :ensure nil\n  :hook (after-init . server-mode))<\/pre><\/div><h4 id=\"editing\">1.4.2 Editing<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">;; Ignore split window horizontally\n(setq split-width-threshold nil)\n(setq split-width-threshold 160)\n;; Default Encoding\n(prefer-coding-system 'utf-8-unix)\n(set-locale-environment \"en_US.UTF-8\")\n(set-default-coding-systems 'utf-8-unix)\n(set-selection-coding-system 'utf-8-unix)\n(set-buffer-file-coding-system 'utf-8-unix)\n(set-clipboard-coding-system 'utf-8) ; included by set-selection-coding-system\n(set-keyboard-coding-system 'utf-8) ; configured by prefer-coding-system\n(set-terminal-coding-system 'utf-8) ; configured by prefer-coding-system\n(setq buffer-file-coding-system 'utf-8) ; utf-8-unix\n(setq save-buffer-coding-system 'utf-8-unix) ; nil\n(setq process-coding-system-alist\n  (cons '(\"grep\" utf-8 . utf-8) process-coding-system-alist))\n;; Quiet Startup\n(setq inhibit-startup-screen t)\n(setq inhibit-startup-message t)\n(setq inhibit-startup-echo-area-message t)\n(setq initial-scratch-message nil)\n(defun display-startup-echo-area-message ()\n  (message \"\"))\n(setq frame-title-format nil)\n(setq ring-bell-function 'ignore)\n(setq uniquify-buffer-name-style 'post-forward-angle-brackets) ; Show path if names are same\n(setq adaptive-fill-regexp \"[ t]+|[ t]*([0-9]+.|*+)[ t]*\")\n(setq adaptive-fill-first-line-regexp \"^* *$\")\n(setq sentence-end \"\\\\([\u3002\u3001\uff01\uff1f]\\\\|\u2026\u2026\\\\|[,.?!][]\\\"')}]*\\\\($\\\\|[ \\t]\\\\)\\\\)[ \\t\\n]*\")\n(setq sentence-end-double-space nil)\n(setq delete-by-moving-to-trash t)    ; Deleting files go to OS's trash folder\n(setq make-backup-files nil)          ; Forbide to make backup files\n(setq auto-save-default nil)          ; Disable auto save\n(setq set-mark-command-repeat-pop t)  ; Repeating C-SPC after popping mark pops it again\n(setq track-eol t)                      ; Keep cursor at end of lines.\n(setq line-move-visual nil)             ; To be required by track-eol\n(setq-default kill-whole-line t)        ; Kill line including '\\n'\n(setq-default indent-tabs-mode nil)   ; use space\n(defalias 'yes-or-no-p #'y-or-n-p)\n(when (functionp 'mac-auto-ascii-mode)\n  (mac-auto-ascii-mode 1))\n;; Delete selection if insert someting\n(use-package delsel\n  :ensure nil\n  :hook (after-init . delete-selection-mode))\n;; Automatically reload files was modified by external program\n(use-package autorevert\n  :ensure nil\n  :diminish\n  :hook (after-init . global-auto-revert-mode))\n;; Hungry deletion\n(use-package hungry-delete\n  :diminish\n  :hook (after-init . global-hungry-delete-mode)\n  :config (setq-default hungry-delete-chars-to-skip \" \\t\\f\\v\"))<\/pre><\/div><h4 id=\"smartparens\">1.4.3 Smartparens<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package smartparens\n  :hook\n  (after-init . smartparens-global-mode)\n  :config\n  (require 'smartparens-config)\n  (sp-pair \"=\" \"=\" :actions '(wrap))\n  (sp-pair \"+\" \"+\" :actions '(wrap))\n  (sp-pair \"&lt;\" \"&gt;\" :actions '(wrap))\n  (sp-pair \"$\" \"$\" :actions '(wrap)))<\/pre><\/div><h4 id=\"history\">1.4.4 History<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">;; History\n(use-package saveplace\n  :ensure nil\n  :hook (after-init . save-place-mode))\n;; Recent files\n(use-package recentf\n  :ensure nil\n  :hook (after-init . recentf-mode)\n  :custom\n  (recentf-max-saved-items 20000000)\n  (recentf-auto-cleanup 'never)\n  (recentf-exclude '((expand-file-name package-user-dir)\n                     \".cache\"\n                     \"cache\"\n                     \"recentf\"\n                     \"COMMIT_EDITMSG\\\\'\"))\n  :preface\n  (defun ladicle\/recentf-save-list-silence ()\n    (interactive)\n    (let ((message-log-max nil))\n      (if (fboundp 'shut-up)\n          (shut-up (recentf-save-list))\n        (recentf-save-list)))\n    (message \"\"))\n  (defun ladicle\/recentf-cleanup-silence ()\n    (interactive)\n    (let ((message-log-max nil))\n      (if shutup-p\n          (shut-up (recentf-cleanup))\n        (recentf-cleanup)))\n    (message \"\"))\n  :hook\n  (focus-out-hook . (ladicle\/recentf-save-list-silence ladicle\/recentf-cleanup-silence)))<\/pre><\/div><h4 id=\"gui-font\">1.4.5 GUI\/Font<\/h4><p>font-size check<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">|abcdef ghijkl|\n|ABCDEF GHIJKL|\n|'\";:-+ =\/\\~`?|\n|\u221e\u2264\u2265\u220f\u2211\u222b \u00d7\u00b1\u2286\u2287|\n|\u03b1\u03b2\u03b3\u03b4\u03b5\u03b6 \u03b7\u03b8\u03b9\u03ba\u03bb\u03bc|\n|\u0391\u0392\u0393\u0394\u0395\u0396 \u0397\u0398\u0399\u039a\u039b\u039c|\n|\u65e5\u672c\u8a9e \u306e\u7f8e\u89b3|\n|\u3042\u3044\u3046 \u3048\u304a\u304b|\n|\u30a2\u30a4\u30a6 \u30a8\u30aa\u30ab|\n|\uff71\uff72\uff73\uff74\uff75\uff76 \uff77\uff78\uff79\uff7a\uff7b\uff7c|\n| hoge                 | hogeghoe | age               |\n|----------------------+----------+-------------------|\n| \u4eca\u65e5\u3082\u3044\u3044\u5929\u6c17\u3067\u3059\u306d | \u304a\u3001     | \u7b49\u5e45\u306b\u306a\u3063\u305f \ud83d\udc4d |<\/pre><\/div><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(if window-system\n    (progn\n      ;; UI parts\n      (toggle-scroll-bar 0)\n      (tool-bar-mode 0)\n      (menu-bar-mode 0)\n      ;; Japanese font settings\n      (defun set-japanese-font (family)\n        (set-fontset-font (frame-parameter nil 'font) 'japanese-jisx0208        (font-spec :family family))\n        (set-fontset-font (frame-parameter nil 'font) 'japanese-jisx0212        (font-spec :family family))\n        (set-fontset-font (frame-parameter nil 'font) 'katakana-jisx0201        (font-spec :family family)))\n      ;; Overwrite latin and greek char's font\n      (defun set-latin-and-greek-font (family)\n        (set-fontset-font (frame-parameter nil 'font) '(#x0250 . #x02AF) (font-spec :family family)) ; IPA extensions\n        (set-fontset-font (frame-parameter nil 'font) '(#x00A0 . #x00FF) (font-spec :family family)) ; latin-1\n        (set-fontset-font (frame-parameter nil 'font) '(#x0100 . #x017F) (font-spec :family family)) ; latin extended-A\n        (set-fontset-font (frame-parameter nil 'font) '(#x0180 . #x024F) (font-spec :family family)) ; latin extended-B\n        (set-fontset-font (frame-parameter nil 'font) '(#x2018 . #x2019) (font-spec :family family)) ; end quote\n        (set-fontset-font (frame-parameter nil 'font) '(#x2588 . #x2588) (font-spec :family family)) ; \u2588\n        (set-fontset-font (frame-parameter nil 'font) '(#x2500 . #x2500) (font-spec :family family)) ; \u2500\n        (set-fontset-font (frame-parameter nil 'font) '(#x2504 . #x257F) (font-spec :family family)) ; box character\n        (set-fontset-font (frame-parameter nil 'font) '(#x0370 . #x03FF) (font-spec :family family)))\n      (setq use-default-font-for-symbols nil)\n      (setq inhibit-compacting-font-caches t)\n      (setq jp-font-family \"SF Mono Square\")\n      (setq default-font-family \"FuraCode Nerd Font\")\n      ;; (set-face-attribute 'default nil :family default-font-family)\n      (when (eq system-type 'darwin)\n                (set-face-attribute 'default nil :family jp-font-family :height 140))\n      (when (eq system-type 'gnu\/linux)\n                (set-face-attribute 'default nil :family jp-font-family :height 150))\n      (set-japanese-font jp-font-family)\n      (set-latin-and-greek-font default-font-family)\n      (add-to-list 'face-font-rescale-alist (cons default-font-family 0.86))\n      (add-to-list 'face-font-rescale-alist (cons jp-font-family 1.0))))<\/pre><\/div><h4 id=\"os\">1.4.6 OS<\/h4><ol><li>\n<p>common<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package exec-path-from-shell\n  :custom\n  (exec-path-from-shell-check-startup-files nil)\n  (exec-path-from-shell-variables '(\"PATH\" \"GOPATH\"))\n  :config\n  (when (memq window-system '(mac ns x))\n    (exec-path-from-shell-initialize)))<\/pre><\/div>\n<\/li>\n<li>\n<p>Linux<\/p>\n<ol><li>\n<p>xclip<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package xclip\n  :if (eq system-type 'gnu\/linux)\n  :config\n  (xclip-mode 1))<\/pre><\/div>\n<\/li>\n<li>\n<p>Mozc<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package mozc\n  :if (eq system-type 'gnu\/linux)\n  :commands (mozc-mode mozc-handle-event)\n  :custom (default-input-method \"japanese-mozc\")\n  :bind ((\"M-l\" . mozc-start)\n         :map mozc-mode-map\n         (\",\" . (lambda () (interactive) (mozc-insert-str \"\u3001\")))\n         (\".\" . (lambda () (interactive) (mozc-insert-str \"\u3002\")))\n         (\"?\" . (lambda () (interactive) (mozc-insert-str \"\uff1f\")))\n         (\"!\" . (lambda () (interactive) (mozc-insert-str \"\uff01\")))\n         (\"l\" . mozc-end)\n         (\"C-m\" . (lambda () (interactive) (mozc-insert-str \"\")))\n         (\"&lt;enter&gt;\" . (lambda () (interactive) (mozc-insert-str \"\")))\n         (\"C-h\" . delete-backward-char))\n  :init\n  (defun mozc-start()\n    (interactive)\n    (set-cursor-color \"#ff79c6\")\n    (message \"Mozc start\")\n    (mozc-mode 1))\n  (defun mozc-end()\n    (interactive)\n    (mozc-handle-event 'enter)\n    (set-cursor-color \"cyan\")\n    (message \"Mozc end\")\n    (mozc-mode -1))\n  (defun mozc-insert-str (str)\n    (mozc-handle-event 'enter)\n    (insert str))\n  :config\n  (mozc-mode 1)\n  ;; Show candidates style\n  (use-package mozc-popup\n    :custom (mozc-candidates-style \"popup\")))<\/pre><\/div>\n<\/li>\n<\/ol><\/li>\n<li>\n<p>macOS<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(when (equal system-type 'darwin)\n  (setq mac-option-modifier 'super)\n  (setq mac-command-modifier 'meta)\n  (setq ns-auto-hide-menu-bar t)\n  (setq ns-use-proxy-icon nil)\n  (setq initial-frame-alist\n     (append\n      '((ns-transparent-titlebar . t)\n        (ns-appearance . dark)\n        (vertical-scroll-bars . nil)\n        (internal-border-width . 0)))))\n;; pbcopy\n(use-package pbcopy\n:if (eq system-type 'darwin)\n:hook (dashboard-mode . (turn-on-pbcopy)))<\/pre><\/div>\n<\/li>\n<\/ol><h4 id=\"icons\">1.4.7 Icons<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package all-the-icons\n  :defer t\n  :load-path \"~\/Developments\/src\/github.com\/Ladicle\/all-the-icons.el\/\")<\/pre><\/div><h4 id=\"posframe\">1.4.8 posframe<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package posframe)<\/pre><\/div><h4 id=\"point\">1.4.9 point<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package popwin)\n(use-package point-history\n  :load-path \"~\/Developments\/src\/github.com\/blue0513\/point-history\"\n  :config\n  (point-history-mode t))<\/pre><\/div><h3 id=\"key-bindings\">1.5 Key-bindings<\/h3><h4 id=\"mouse\">1.5.1 Mouse<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(xterm-mouse-mode t)\n(global-set-key   [mouse-4] '(lambda () (interactive) (scroll-down 1)))\n(global-set-key   [mouse-5] '(lambda () (interactive) (scroll-up   1)))<\/pre><\/div><h4 id=\"custom-maps\">1.5.2 Custom Maps<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(define-prefix-command 'ladicle-window-map)\n(define-key global-map (kbd \"M-o\") 'ladicle-window-map)\n(define-prefix-command 'ladicle-toggle-map)\n(define-key global-map (kbd \"M-t\") 'ladicle-toggle-map)\n(define-prefix-command 'ladicle-link-map)\n(define-key global-map (kbd \"M-o l\") 'ladicle-link-map)<\/pre><\/div><h4 id=\"global\">1.5.3 Global<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(global-unset-key \"\\C-z\")\n(global-set-key (kbd \"M-o m\")   'async-shell-command)\n(global-set-key (kbd \"M-o b\")   'switch-to-buffer)\n(global-set-key (kbd \"C-x C-b\") 'switch-to-buffer)\n(global-set-key (kbd \"C-M-g\")   'goto-line)\n(global-set-key (kbd \"C-h\")     'backward-delete-char)\n(global-set-key (kbd \"C-o\")     'other-window)\n(global-set-key (kbd \"M-:\")     'comment-dwim)\n(global-set-key (kbd \"C-m\")     'electric-newline-and-maybe-indent)\n(global-set-key (kbd \"C-j\")     'newline-and-indent)\n(global-set-key (kbd \"M-r\")     'rename-file)\n(global-set-key (kbd \"M-t l\")   'toggle-truncate-lines)<\/pre><\/div><h4 id=\"which-key\">1.5.4 which-key<\/h4><p>Guide Key-bindings <a href=\"https:\/\/github.com\/justbur\/emacs-which-key\">https:\/\/github.com\/justbur\/emacs-which-key<\/a><\/p><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package which-key\n  :diminish which-key-mode\n  :hook (after-init . which-key-mode))<\/pre><\/div><h4 id=\"hydra\">1.5.5 Hydra<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package hydra\n  :config\n  (use-package hydra-posframe\n    :load-path \"~\/Developments\/src\/github.com\/Ladicle\/hydra-posframe\"\n    :custom\n    (hydra-posframe-parameters\n      '((left-fringe . 5)\n        (right-fringe . 5)))\n    :custom-face\n    (hydra-posframe-border-face ((t (:background \"#6272a4\"))))\n    :hook (after-init . hydra-posframe-enable)))<\/pre><\/div><h3 id=\"undo-redo\">1.6 Undo\/Redo<\/h3><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/UndoTree\">https:\/\/www.emacswiki.org\/emacs\/UndoTree<\/a><\/p><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package undo-tree\n  :bind\n  (\"M-\/\" . undo-tree-redo)\n  :config\n  (global-undo-tree-mode))<\/pre><\/div><h4 id=\"projectile\">1.7.1 Projectile<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package projectile\n  :diminish\n  :bind\n  (\"M-o p\" . counsel-projectile-switch-project)\n  :config\n  (projectile-mode +1))<\/pre><\/div><h4 id=\"wgrep\">1.7.2 Wgrep<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package wgrep\n  :defer t\n  :custom\n  (wgrep-enable-key \"e\")\n  (wgrep-auto-save-buffer t)\n  (wgrep-change-readonly-file t))<\/pre><\/div><h4 id=\"ag\">1.7.3 Ag<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package ag\n  :custom\n  (ag-highligh-search t)\n  (ag-reuse-buffers t)\n  (ag-reuse-window t)\n  :bind\n  (\"M-s a\" . ag-project)\n  :config\n  (use-package wgrep-ag))<\/pre><\/div><h4 id=\"ivy\">1.7.4 Ivy<\/h4><p><a href=\"https:\/\/github.com\/abo-abo\/swiper\">https:\/\/github.com\/abo-abo\/swiper<\/a><\/p><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">  (use-package counsel\n    :diminish ivy-mode counsel-mode\n    :defines\n    (projectile-completion-system magit-completing-read-function)\n    :bind\n    ((\"C-s\" . swiper)\n    (\"M-s r\" . ivy-resume)\n    (\"C-c v p\" . ivy-push-view)\n    (\"C-c v o\" . ivy-pop-view)\n    (\"C-c v .\" . ivy-switch-view)\n    (\"M-s c\" . counsel-ag)\n    (\"M-o f\" . counsel-fzf)\n    (\"M-o r\" . counsel-recentf)\n    (\"M-y\" . counsel-yank-pop)\n    :map ivy-minibuffer-map\n    (\"C-w\" . ivy-backward-kill-word)\n    (\"C-k\" . ivy-kill-line)\n    (\"C-j\" . ivy-immediate-done)\n    (\"RET\" . ivy-alt-done)\n    (\"C-h\" . ivy-backward-delete-char))\n    :preface\n    (defun ivy-format-function-pretty (cands)\n      \"Transform CANDS into a string for minibuffer.\"\n      (ivy--format-function-generic\n       (lambda (str)\n         (concat\n             (all-the-icons-faicon \"hand-o-right\" :height .85 :v-adjust .05 :face 'font-lock-constant-face)\n             (ivy--add-face str 'ivy-current-match)))\n       (lambda (str)\n         (concat \"  \" str))\n       cands\n       \"\\n\"))\n    :hook\n    (after-init . ivy-mode)\n    (ivy-mode . counsel-mode)\n    :custom\n    (counsel-yank-pop-height 15)\n    (enable-recursive-minibuffers t)\n    (ivy-use-selectable-prompt t)\n    (ivy-use-virtual-buffers t)\n    (ivy-on-del-error-function nil)\n    (swiper-action-recenter t)\n    (counsel-grep-base-command \"ag -S --noheading --nocolor --nofilename --numbers '%s' %s\")\n    :config\n    ;; using ivy-format-fuction-arrow with counsel-yank-pop\n    (advice-add\n    'counsel--yank-pop-format-function\n    :override\n    (lambda (cand-pairs)\n      (ivy--format-function-generic\n       (lambda (str)\n         (mapconcat\n          (lambda (s)\n            (ivy--add-face (concat (propertize \"\u2503 \" 'face `(:foreground \"#61bfff\")) s) 'ivy-current-match))\n          (split-string\n           (counsel--yank-pop-truncate str) \"\\n\" t)\n          \"\\n\"))\n       (lambda (str)\n         (counsel--yank-pop-truncate str))\n       cand-pairs\n       counsel-yank-pop-separator)))\n    ;; NOTE: this variable do not work if defined in :custom\n    (setq ivy-format-function 'ivy-format-function-pretty)\n    (setq counsel-yank-pop-separator\n        (propertize \"\\n\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\\n\"\n               'face `(:foreground \"#6272a4\")))\n    ;; Integration with `projectile'\n    (with-eval-after-load 'projectile\n      (setq projectile-completion-system 'ivy))\n    ;; Integration with `magit'\n    (with-eval-after-load 'magit\n      (setq magit-completing-read-function 'ivy-completing-read))\n    ;; Enhance fuzzy matching\n    (use-package flx)\n    ;; Enhance M-x\n    (use-package amx)\n    ;; Ivy integration for Projectile\n    (use-package counsel-projectile\n      :config (counsel-projectile-mode 1))\n  ;; Show ivy frame using posframe\n  (use-package ivy-posframe\n    :custom\n    (ivy-display-function #'ivy-posframe-display-at-frame-center)\n    ;; (ivy-posframe-width 130)\n    ;; (ivy-posframe-height 11)\n    (ivy-posframe-parameters\n      '((left-fringe . 5)\n        (right-fringe . 5)))\n    :custom-face\n    (ivy-posframe ((t (:background \"#282a36\"))))\n    (ivy-posframe-border ((t (:background \"#6272a4\"))))\n    (ivy-posframe-cursor ((t (:background \"#61bfff\"))))\n    :hook\n    (ivy-mode . ivy-posframe-enable))\n  ;; ghq\n  (use-package ivy-ghq\n    :load-path \"~\/Developments\/src\/github.com\/analyticd\/ivy-ghq\"\n    :commands (ivy-ghq-open)\n    :bind\n    (\"M-o p\" . ivy-ghq-open-and-fzf)\n    :custom\n    (ivy-ghq-short-list t)\n    :preface\n    (defun ivy-ghq-open-and-fzf ()\n      (interactive)\n      (ivy-ghq-open)\n      (counsel-fzf)))\n  ;; More friendly display transformer for Ivy\n  (use-package ivy-rich\n    :defines (all-the-icons-dir-icon-alist bookmark-alist)\n    :functions (all-the-icons-icon-family\n                all-the-icons-match-to-alist\n                all-the-icons-auto-mode-match?\n                all-the-icons-octicon\n                all-the-icons-dir-is-submodule)\n    :preface\n    (defun ivy-rich-bookmark-name (candidate)\n      (car (assoc candidate bookmark-alist)))\n    (defun ivy-rich-repo-icon (candidate)\n      \"Display repo icons in `ivy-rich`.\"\n      (all-the-icons-octicon \"repo\" :height .9))\n    (defun ivy-rich-org-capture-icon (candidate)\n      \"Display repo icons in `ivy-rich`.\"\n      (pcase (car (last (split-string (car (split-string candidate)) \"-\")))\n         (\"emacs\" (all-the-icons-fileicon \"emacs\" :height .68 :v-adjust .001))\n         (\"schedule\" (all-the-icons-faicon \"calendar\" :height .68 :v-adjust .005))\n         (\"tweet\" (all-the-icons-faicon \"commenting\" :height .7 :v-adjust .01))\n         (\"link\" (all-the-icons-faicon \"link\" :height .68 :v-adjust .01))\n         (\"memo\" (all-the-icons-faicon \"pencil\" :height .7 :v-adjust .01))\n         (_       (all-the-icons-octicon \"inbox\" :height .68 :v-adjust .01))\n         ))\n    (defun ivy-rich-org-capture-title (candidate)\n      (let* ((octl (split-string candidate))\n             (title (pop octl))\n             (desc (mapconcat 'identity octl \" \")))\n        (format \"%-25s %s\"\n                 title\n                 (propertize desc 'face `(:inherit font-lock-doc-face)))))\n    (defun ivy-rich-buffer-icon (candidate)\n      \"Display buffer icons in `ivy-rich'.\"\n      (when (display-graphic-p)\n        (when-let* ((buffer (get-buffer candidate))\n                    (major-mode (buffer-local-value 'major-mode buffer))\n                    (icon (if (and (buffer-file-name buffer)\n                                   (all-the-icons-auto-mode-match? candidate))\n                              (all-the-icons-icon-for-file candidate)\n                            (all-the-icons-icon-for-mode major-mode))))\n          (if (symbolp icon)\n              (setq icon (all-the-icons-icon-for-mode 'fundamental-mode)))\n          (unless (symbolp icon)\n            (propertize icon\n                        'face `(\n                                :height 1.1\n                                :family ,(all-the-icons-icon-family icon)\n                                ))))))\n    (defun ivy-rich-file-icon (candidate)\n      \"Display file icons in `ivy-rich'.\"\n      (when (display-graphic-p)\n        (let ((icon (if (file-directory-p candidate)\n                        (cond\n                         ((and (fboundp 'tramp-tramp-file-p)\n                               (tramp-tramp-file-p default-directory))\n                          (all-the-icons-octicon \"file-directory\"))\n                         ((file-symlink-p candidate)\n                          (all-the-icons-octicon \"file-symlink-directory\"))\n                         ((all-the-icons-dir-is-submodule candidate)\n                          (all-the-icons-octicon \"file-submodule\"))\n                         ((file-exists-p (format \"%s\/.git\" candidate))\n                          (all-the-icons-octicon \"repo\"))\n                         (t (let ((matcher (all-the-icons-match-to-alist candidate all-the-icons-dir-icon-alist)))\n                              (apply (car matcher) (list (cadr matcher))))))\n                      (all-the-icons-icon-for-file candidate))))\n          (unless (symbolp icon)\n            (propertize icon\n                        'face `(\n                                :height 1.1\n                                :family ,(all-the-icons-icon-family icon)\n                                ))))))\n    :hook (ivy-rich-mode . (lambda ()\n                             (setq ivy-virtual-abbreviate\n                                   (or (and ivy-rich-mode 'abbreviate) 'name))))\n    :init\n    (setq ivy-rich-display-transformers-list\n          '(ivy-switch-buffer\n            (:columns\n             ((ivy-rich-buffer-icon)\n              (ivy-rich-candidate (:width 30))\n              (ivy-rich-switch-buffer-size (:width 7))\n              (ivy-rich-switch-buffer-indicators (:width 4 :face error :align right))\n              (ivy-rich-switch-buffer-major-mode (:width 12 :face warning))\n              (ivy-rich-switch-buffer-project (:width 15 :face success))\n              (ivy-rich-switch-buffer-path (:width (lambda (x) (ivy-rich-switch-buffer-shorten-path x (ivy-rich-minibuffer-width 0.3))))))\n             :predicate\n             (lambda (cand) (get-buffer cand)))\n            ivy-switch-buffer-other-window\n            (:columns\n             ((ivy-rich-buffer-icon)\n              (ivy-rich-candidate (:width 30))\n              (ivy-rich-switch-buffer-size (:width 7))\n              (ivy-rich-switch-buffer-indicators (:width 4 :face error :align right))\n              (ivy-rich-switch-buffer-major-mode (:width 12 :face warning))\n              (ivy-rich-switch-buffer-project (:width 15 :face success))\n              (ivy-rich-switch-buffer-path (:width (lambda (x) (ivy-rich-switch-buffer-shorten-path x (ivy-rich-minibuffer-width 0.3))))))\n             :predicate\n             (lambda (cand) (get-buffer cand)))\n            counsel-M-x\n            (:columns\n             ((counsel-M-x-transformer (:width 40))\n              (ivy-rich-counsel-function-docstring (:face font-lock-doc-face))))\n            counsel-describe-function\n            (:columns\n             ((counsel-describe-function-transformer (:width 45))\n              (ivy-rich-counsel-function-docstring (:face font-lock-doc-face))))\n            counsel-describe-variable\n            (:columns\n             ((counsel-describe-variable-transformer (:width 45))\n              (ivy-rich-counsel-variable-docstring (:face font-lock-doc-face))))\n            counsel-find-file\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            counsel-file-jump\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            counsel-dired-jump\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            counsel-git\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            counsel-recentf\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate (:width 110))))\n            counsel-bookmark\n            (:columns\n             ((ivy-rich-bookmark-type)\n              (ivy-rich-bookmark-name (:width 30))\n              (ivy-rich-bookmark-info (:width 80))))\n            counsel-projectile-switch-project\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            counsel-fzf\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            ivy-ghq-open\n            (:columns\n             ((ivy-rich-repo-icon)\n              (ivy-rich-candidate)))\n            ivy-ghq-open-and-fzf\n            (:columns\n             ((ivy-rich-repo-icon)\n              (ivy-rich-candidate)))\n            counsel-projectile-find-file\n            (:columns\n             ((ivy-rich-file-icon)\n              (ivy-rich-candidate)))\n            counsel-org-capture\n            (:columns\n             ((ivy-rich-org-capture-icon)\n              (ivy-rich-org-capture-title)\n              ))\n            counsel-projectile-find-dir\n            (:columns\n             ((ivy-rich-file-icon)\n              (counsel-projectile-find-dir-transformer)))))\n    (setq ivy-rich-parse-remote-buffer nil)\n    :config\n    (ivy-rich-mode 1))\n)<\/pre><\/div><h4 id=\"anzu\">1.7.5 Anzu<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package anzu\n  :diminish\n  :bind\n  (\"C-r\"   . anzu-query-replace-regexp)\n  (\"C-M-r\" . anzu-query-replace-at-cursor-thing)\n  :hook\n  (after-init . global-anzu-mode))<\/pre><\/div><h4 id=\"migemo\">1.7.6 Migemo<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package migemo\n  :custom\n  (migemo-command \"\/usr\/local\/bin\/cmigemo\")\n  (migemo-options '(\"-q\" \"--emacs\"))\n  (migemo-user-dictionary nil)\n  (migemo-regex-dictionary nil)\n  (migemo-coding-system 'utf-8-unix)\n  (migemo-dictionary \"\/usr\/share\/cmigemo\/utf-8\/migemo-dict\")\n  :config\n  (when (eq system-type 'darwin)\n    (setq migemo-dictionary \"\/usr\/local\/share\/migemo\/utf-8\/migemo-dict\"))\n  (migemo-init)\n  (use-package avy-migemo\n    :after swiper\n    :config\n    ;; (avy-migemo-mode 1)\n    (require 'avy-migemo-e.g.swiper))\n  )<\/pre><\/div><h3 id=\"cursor\">1.8 Cursor<\/h3><h4 id=\"multiple-cursor\">1.8.1 multiple-cursor<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package multiple-cursors\n  :functions hydra-multiple-cursors\n  :bind\n  (\"M-u\" . hydra-multiple-cursors\/body)\n  :preface\n  ;; insert specific serial number\n  (defvar ladicle\/mc\/insert-numbers-hist nil)\n  (defvar ladicle\/mc\/insert-numbers-inc 1)\n  (defvar ladicle\/mc\/insert-numbers-pad \"%01d\")\n  (defun ladicle\/mc\/insert-numbers (start inc pad)\n    \"Insert increasing numbers for each cursor specifically.\"\n    (interactive\n     (list (read-number \"Start from: \" 0)\n           (read-number \"Increment by: \" 1)\n           (read-string \"Padding (%01d): \" nil ladicle\/mc\/insert-numbers-hist \"%01d\")))\n    (setq mc--insert-numbers-number start)\n    (setq ladicle\/mc\/insert-numbers-inc inc)\n    (setq ladicle\/mc\/insert-numbers-pad pad)\n    (mc\/for-each-cursor-ordered\n     (mc\/execute-command-for-fake-cursor\n      'ladicle\/mc--insert-number-and-increase\n      cursor)))\n  (defun ladicle\/mc--insert-number-and-increase ()\n    (interactive)\n    (insert (format ladicle\/mc\/insert-numbers-pad mc--insert-numbers-number))\n    (setq mc--insert-numbers-number (+ mc--insert-numbers-number ladicle\/mc\/insert-numbers-inc)))\n  :config\n  (with-eval-after-load 'hydra\n    (defhydra hydra-multiple-cursors (:color pink :hint nil)\n\"\n                                                                        \u2554\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2557\n    Point^^^^^^             Misc^^            Insert                            \u2551 Cursor \u2551\n  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2568\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u255c\n     _k_    _K_    _M-k_    [_l_] edit lines  [_i_] 0...\n     ^\u2191^    ^\u2191^     ^\u2191^     [_m_] mark all    [_a_] letters\n    mark^^ skip^^^ un-mk^   [_s_] sort        [_n_] numbers\n     ^\u2193^    ^\u2193^     ^\u2193^\n     _j_    _J_    _M-j_\n  \u256d\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u256f\n                           [_q_]: quit, [Click]: point\n\"\n          (\"l\" mc\/edit-lines :exit t)\n          (\"m\" mc\/mark-all-like-this :exit t)\n          (\"j\" mc\/mark-next-like-this)\n          (\"J\" mc\/skip-to-next-like-this)\n          (\"M-j\" mc\/unmark-next-like-this)\n          (\"k\" mc\/mark-previous-like-this)\n          (\"K\" mc\/skip-to-previous-like-this)\n          (\"M-k\" mc\/unmark-previous-like-this)\n          (\"s\" mc\/mark-all-in-region-regexp :exit t)\n          (\"i\" mc\/insert-numbers :exit t)\n          (\"a\" mc\/insert-letters :exit t)\n          (\"n\" ladicle\/mc\/insert-numbers :exit t)\n          (\"&lt;mouse-1&gt;\" mc\/add-cursor-on-click)\n          ;; Help with click recognition in this hydra\n          (\"&lt;down-mouse-1&gt;\" ignore)\n          (\"&lt;drag-mouse-1&gt;\" ignore)\n          (\"q\" nil))))<\/pre><\/div><h4 id=\"avy-ace\">1.8.2 avy\/ace<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">  (use-package avy\n    :functions (hydra-avy hydra-viewer)\n    :bind\n    (\"C-'\"   . avy-resume)\n    (\"C-:\"   . avy-goto-char-2-below)\n    (\"C-;\"   . avy-goto-char)\n    (\"M-j\"   . hydra-avy\/body)\n    (\"C-M-v\" . hydra-viewer\/body)\n    :preface\n    ;; fixed cursor scroll-up\n    (defun scroll-up-in-place (n)\n      (interactive \"p\")\n      (forward-line (- n))\n      (scroll-down n))\n    ;; fixed cursor scroll-down\n    (defun scroll-down-in-place (n)\n      (interactive \"p\")\n      (forward-line n)\n      (scroll-up n))\n    ;; yank inner sexp\n    (defun yank-inner-sexp ()\n      (interactive)\n      (backward-list)\n      (mark-sexp)\n      (copy-region-as-kill (region-beginning) (region-end)))\n    :config\n    (when (eq system-type 'darwin)\n      (progn\n        (global-set-key (kbd \"C-:\") 'avy-goto-char)\n        (global-set-key (kbd \"C-;\") 'avy-goto-char-2-below)))\n    (use-package avy-zap\n      :bind\n      (\"M-z\" . avy-zap-to-char-dwim)\n      (\"M-z\" . avy-zap-up-to-char-dwim))\n    (with-eval-after-load 'hydra\n        (defhydra hydra-viewer (:color pink :hint nil)\n          \"\n                                                                        \u2554\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2557\n   Char\/Line^^^^^^  Word\/Page^^^^^^^^  Line\/Buff^^^^   Paren                              \u2551 Window \u2551\n  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2568\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u255c\n       ^^_k_^^          ^^_u_^^          ^^_g_^^       _(_ \u2190 _y_ \u2192 _)_\n       ^^^\u2191^^^          ^^^\u2191^^^          ^^^\u2191^^^       _,_ \u2190 _\/_ \u2192 _._\n   _h_ \u2190 _d_ \u2192 _l_  _H_ \u2190 _D_ \u2192 _L_  _a_ \u2190 _K_ \u2192 _e_\n       ^^^\u2193^^^          ^^^\u2193^^^          ^^^\u2193^\n       ^^_j_^^          ^^_n_^^          ^^_G_\n  \u256d\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u256f\n                           [_q_]: quit, [_&lt;SPC&gt;_]: center\n          \"\n          (\"j\" scroll-down-in-place)\n          (\"k\" scroll-up-in-place)\n          (\"l\" forward-char)\n          (\"d\" delete-char)\n          (\"h\" backward-char)\n          (\"L\" forward-word)\n          (\"H\" backward-word)\n          (\"u\" scroll-up-command)\n          (\"n\" scroll-down-command)\n          (\"D\" delete-word-at-point)\n          (\"a\" mwim-beginning-of-code-or-line)\n          (\"e\" mwim-end-of-code-or-line)\n          (\"g\" beginning-of-buffer)\n          (\"G\" end-of-buffer)\n          (\"K\" kill-whole-line)\n          (\"(\" backward-list)\n          (\")\" forward-list)\n          (\"y\" yank-inner-sexp)\n          (\".\" backward-forward-next-location)\n          (\",\" backward-forward-previous-location)\n          (\"\/\" avy-goto-char :exit t)\n          (\"&lt;SPC&gt;\" recenter-top-bottom)\n          (\"q\" nil))\n        (defhydra hydra-avy (:color pink :hint nil)\n          \"\n                                                                        \u2554\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2557\n        ^^Goto^^        Kill^^        Yank^^        Move^^        Misc            \u2551  Jump  \u2551\n  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2568\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u255c\n    _c_ \u2190 char^^        [_k_] region  [_y_] region  [_m_] region  [_n_] line number\n    _a_ \u2190 char2 \u2192 _b_   [_K_] line    [_Y_] line    [_M_] line    [_v_] Goto viewer\n    _w_ \u2190 word  \u2192 _W_   [_z_] zap^^^^                             [_o_] Goto clock\n    _l_ \u2190 line  \u2192 _e_   ^^^^^                                     _,_ \u2190 f!y \u2192 _._\n  \u256d\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u256f\n                      [_q_]: quit, [_i_]: imenu, [_&lt;SPC&gt;_]: resume\n\"\n          (\"c\" avy-goto-char :exit t)\n          (\"a\" avy-goto-char-2 :exit t)\n          (\"b\" avy-goto-char-below :exit t)\n          (\"w\" avy-goto-word-1 :exit t)\n          (\"W\" avy-goto-word-1-below :exit t)\n          (\"l\" avy-goto-line :exit t)\n          (\"e\" avy-goto-end-of-line :exit t)\n          (\"M\" avy-move-line)\n          (\"m\" avy-move-region)\n          (\"K\" avy-kill-whole-line)\n          (\"k\" avy-kill-region)\n          (\"Y\" avy-copy-line :exit t)\n          (\"y\" avy-copy-region :exit t)\n          (\"n\" goto-line :exit t)\n          (\"o\" org-clock-jump-to-current-clock :exit t)\n          (\"z\" avy-zap-to-char-dwim :exit t)\n          (\"v\" hydra-viewer\/body :exit t)\n          (\"&lt;SPC&gt;\" avy-resume :exit t)\n          (\"o\" org-clock-jump-to-current-clock :exit t)\n          (\"i\" counsel-imenu :exit t)\n          (\",\" flymake-goto-previous-error)\n          (\".\" flymake-goto-next-error)\n          (\"q\" nil))))\n  (use-package ace-window\n    :functions hydra-frame-window\/body\n    :bind\n    (\"C-M-o\" . hydra-frame-window\/body)\n    (\"M-t m\" . ladicle\/toggle-window-maximize)\n    :custom\n    (aw-keys '(?j ?k ?l ?i ?o ?h ?y ?u ?p))\n    :custom-face\n    (aw-leading-char-face ((t (:height 4.0 :foreground \"#f1fa8c\"))))\n    :preface\n    (defvar is-window-maximized nil)\n    (defun ladicle\/toggle-window-maximize ()\n        (interactive)\n        (progn\n          (if is-window-maximized\n              (balance-windows)\n            (maximize-window))\n          (setq is-window-maximized\n                (not is-window-maximized))))\n    (defun hydra-title(title) (propertize title 'face `(:inherit font-lock-warning-face :weight bold)))\n    (defun command-name(title) (propertize title 'face `(:foreground \"#f8f8f2\")))\n    (defun spacer() (propertize \".\" 'face `(:foreground \"#282a36\")))\n    :config\n    (use-package rotate\n        :load-path \"~\/Developments\/src\/github.com\/Ladicle\/dotfiles\/common\/emacs.d\/elisp\/emacs-rotate\"\n        :bind\n        (\"M-o SPC\" . rotate-layout))\n    (with-eval-after-load 'hydra\n        (defhydra hydra-frame-window (:color blue :hint nil)\n        (format\n         (format \"%s\" (propertize \"                                                                       \u2554\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2550\u2557\n    ((%s))^^^^^^^^   ((%s))^^^^  ((%s))^^  ((%s))^^  ((%s))^^^^^^  ((%s))^   \u2551 Window \u2551\n^^^^^^ \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2568\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u255c\n        ^_k_^        %s_+_         _-_       %s     _,_ \u2190 %s \u2192 _._^  %s\n        ^^\u2191^^          ^\u2191^         ^\u2191^       %s\n    _h_ \u2190   \u2192 _l_   ^^%s%s^^^^^    ^%s    ^^^%s^^^^     %s\n        ^^\u2193^^          ^\u2193^         ^\u2193^       %s^^       %s\n        ^_j_^        %s_=_         _\/_       %s\n^^^^^^ \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                           [_q_]: %s, [_&lt;SPC&gt;_]: %s\" 'face `(:inherit font-lock-doc-face)))\n                           (hydra-title \"Size\")\n                           (hydra-title \"Zoom\")\n                           (hydra-title \"Split\")\n                           (hydra-title \"Window\")\n                           (hydra-title \"Buffer\")\n                           (hydra-title \"Misc\")\n                           (all-the-icons-material \"zoom_in\" :height .85 :face 'font-lock-doc-face)\n                           (command-name \"_o_ther\")\n                           (command-name \"page\")\n                           (command-name \"_r_centf\")\n                           (command-name \"_s_wap\")\n                           (all-the-icons-faicon \"slideshare\" :height .85 :face 'font-lock-doc-face)\n                           (command-name \"_p_mode\")\n                           (command-name \"w_i_ndow\")\n                           (command-name \"_m_aximize\")\n                           (command-name \"_s_witch\")\n                           (command-name \"_d_elete\")\n                           (command-name \"_D_elete\")\n                           (all-the-icons-material \"zoom_out\" :height .85 :face 'font-lock-doc-face)\n                           (command-name \"del_O_thers\")\n                           (command-name \"quit\")\n                           (command-name \"rotate\")\n                           )\n          (\"K\" kill-current-buffer :exit t)\n          (\"D\" kill-buffer-and-window :exit t)\n          (\"O\" delete-other-windows  :exit t)\n          (\"F\" toggle-frame-fullscreen)\n          (\"i\" ace-window)\n          (\"s\" ace-swap-window :exit t)\n          (\"d\" ace-delete-window)\n          (\"m\" ladicle\/toggle-window-maximize :exit t)\n          (\"=\" text-scale-decrease)\n          (\"+\" text-scale-increase)\n          (\"-\" split-window-vertically)\n          (\"\/\" split-window-horizontally)\n          (\"h\" shrink-window-horizontally)\n          (\"k\" shrink-window)\n          (\"j\" enlarge-window)\n          (\"l\" enlarge-window-horizontally)\n          (\",\" previous-buffer)\n          (\".\" next-buffer)\n          (\"o\" other-window)\n          (\"p\" presentation-mode)\n          (\"r\" counsel-recentf :exit t)\n          (\"s\" switch-to-buffer :exit t)\n          (\"D\" kill-buffer-and-window)\n          (\"&lt;SPC&gt;\" rotate-layout)\n          (\"q\" nil)))\n          )<\/pre><\/div><h4 id=\"smart-move\">1.8.3 Smart move<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package mwim\n  :bind\n  (\"C-a\" . mwim-beginning-of-code-or-line)\n  (\"C-e\" . mwim-end-of-code-or-line))<\/pre><\/div><h3 id=\"check\">1.9 Check<\/h3><h4 id=\"flymake\">1.9.1 flymake<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package flymake-posframe\n  :load-path \"~\/Developments\/src\/github.com\/Ladicle\/flymake-posframe\"\n  :custom\n  (flymake-posframe-error-prefix \"\uf42d \")\n  :custom-face\n  (flymake-posframe-foreground-face ((t (:foreground \"white\"))))\n  :hook (flymake-mode . flymake-posframe-mode))\n(use-package flymake-diagnostic-at-point\n  :disabled\n  :after flymake\n  :custom\n  (flymake-diagnostic-at-point-timer-delay 0.1)\n  (flymake-diagnostic-at-point-error-prefix \"\uf42d \")\n  (flymake-diagnostic-at-point-display-diagnostic-function 'flymake-diagnostic-at-point-display-popup) ;; or flymake-diagnostic-at-point-display-minibuffer\n  :hook\n  (flymake-mode . flymake-diagnostic-at-point-mode))<\/pre><\/div><h4 id=\"flyspell\">1.9.2 flyspell<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package flyspell\n  :diminish\n  :if (executable-find \"aspell\")\n  :hook\n  ((org-mode yaml-mode markdown-mode git-commit-mode) . flyspell-mode)\n  (prog-mode . flyspell-prog-mode)\n  (before-save-hook . flyspell-buffer)\n  (flyspell-mode . (lambda ()\n                     (dolist (key '(\"C-;\" \"C-,\" \"C-.\"))\n                       (unbind-key key flyspell-mode-map))))\n  :custom\n  (flyspell-issue-message-flag nil)\n  (ispell-program-name \"aspell\")\n  (ispell-extra-args '(\"--sug-mode=ultra\" \"--lang=en_US\" \"--run-together\"))\n  :custom-face\n  (flyspell-incorrect ((t (:underline (:color \"#f1fa8c\" :style wave)))))\n  (flyspell-duplicate ((t (:underline (:color \"#50fa7b\" :style wave)))))\n  :preface\n  (defun message-off-advice (oldfun &amp;rest args)\n    \"Quiet down messages in adviced OLDFUN.\"\n    (let ((message-off (make-symbol \"message-off\")))\n      (unwind-protect\n          (progn\n            (advice-add #'message :around #'ignore (list 'name message-off))\n            (apply oldfun args))\n        (advice-remove #'message message-off))))\n  :config\n  (advice-add #'ispell-init-process :around #'message-off-advice)\n  (use-package flyspell-correct-ivy\n    :bind (\"C-M-:\" . flyspell-correct-at-point)\n    :config\n    (when (eq system-type 'darwin)\n      (progn\n        (global-set-key (kbd \"C-M-;\") 'flyspell-correct-at-point)))\n    (setq flyspell-correct-interface #'flyspell-correct-ivy)))<\/pre><\/div><h3 id=\"project\">1.10 Project<\/h3><h4 id=\"persp\">1.10.1 persp<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package persp-mode\n  :disabled\n  :diminish\n  :defines ivy-sort-functions-alist\n  :commands (get-current-persp persp-contain-buffer-p persp-add persp-by-name-and-exists)\n  :hook ((after-init . persp-mode)\n         (emacs-startup . toggle-frame-maximized))\n  :custom\n  (persp-keymap-prefix (kbd \"C-x p\"))\n  (persp-nil-name \"default\")\n  (persp-set-last-persp-for-new-frames nil)\n  (persp-auto-resume-time 0)\n  :config\n  ;; NOTE: Redefine `persp-add-new' to address.\n  ;; Issue: Unable to create\/handle persp-mode\n  ;; https:\/\/github.com\/Bad-ptr\/persp-mode.el\/issues\/96\n  ;; https:\/\/github.com\/Bad-ptr\/persp-mode-projectile-bridge.el\/issues\/4\n  ;; https:\/\/emacs-china.org\/t\/topic\/6416\/7\n  (defun* persp-add-new (name &amp;optional (phash *persp-hash*))\n    \"Create a new perspective with the given `NAME'. Add it to `PHASH'.\n  Return the created perspective.\"\n    (interactive \"sA name for the new perspective: \")\n    (if (and name (not (equal \"\" name)))\n        (destructuring-bind (e . p)\n            (persp-by-name-and-exists name phash)\n          (if e p\n            (setq p (if (equal persp-nil-name name)\n                        nil (make-persp :name name)))\n            (persp-add p phash)\n            (run-hook-with-args 'persp-created-functions p phash)\n            p))\n      (message \"[persp-mode] Error: Can't create a perspective with empty name.\")\n      nil))\n  ;; Ignore temporary buffers\n  (add-hook 'persp-common-buffer-filter-functions\n            (lambda (b) (or (string-prefix-p \"*\" (buffer-name b))\n                       (string-prefix-p \"magit\" (buffer-name b)))))\n  ;; Integrate IVY\n  (with-eval-after-load \"ivy\"\n    (add-hook 'ivy-ignore-buffers\n              #'(lambda (b)\n                  (when persp-mode\n                    (let ((persp (get-current-persp)))\n                      (if persp\n                          (not (persp-contain-buffer-p b persp))\n                        nil)))))\n    (setq ivy-sort-functions-alist\n          (append ivy-sort-functions-alist\n                  '((persp-kill-buffer   . nil)\n                    (persp-remove-buffer . nil)\n                    (persp-add-buffer    . nil)\n                    (persp-switch        . nil)\n                    (persp-window-switch . nil)\n                    (persp-frame-switch  . nil))))))<\/pre><\/div><h3 id=\"complition\">1.11 Complition<\/h3><h4 id=\"yasnippet\">1.11.1 Yasnippet<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package yasnippet\n  :diminish yas-minor-mode\n  :custom (yas-snippet-dirs '(\"~\/.emacs.d\/snippets\"))\n  :hook (after-init . yas-global-mode))<\/pre><\/div><h4 id=\"company\">1.11.2 Company<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package company\n  :diminish company-mode\n  :defines\n  (company-dabbrev-ignore-case company-dabbrev-downcase)\n  :bind\n  (:map company-active-map\n   (\"C-n\" . company-select-next)\n   (\"C-p\" . company-select-previous)\n   (\"&lt;tab&gt;\" . company-complete-common-or-cycle)\n   :map company-search-map\n   (\"C-p\" . company-select-previous)\n   (\"C-n\" . company-select-next))\n  :custom\n  (company-idle-delay 0)\n  (company-echo-delay 0)\n  (company-minimum-prefix-length 1)\n  :hook\n  (after-init . global-company-mode)\n  (plantuml-mode . (lambda () (set (make-local-variable 'company-backends)\n                            '((company-yasnippet\n                               ;; company-dabbrev\n                               )))))\n  ((go-mode\n    c++-mode\n    c-mode\n    objc-mode) . (lambda () (set (make-local-variable 'company-backends)\n                            '((company-yasnippet\n                               company-lsp\n                               company-files\n                               ;; company-dabbrev-code\n                               )))))\n  :config\n  ;; using child frame\n  (use-package company-posframe\n    :hook (company-mode . company-posframe-mode))\n  ;; Show pretty icons\n  (use-package company-box\n    :diminish\n    :hook (company-mode . company-box-mode)\n    :init (setq company-box-icons-alist 'company-box-icons-all-the-icons)\n    :config\n    (setq company-box-backends-colors nil)\n    (setq company-box-show-single-candidate t)\n    (setq company-box-max-candidates 50)\n    (defun company-box-icons--elisp (candidate)\n      (when (derived-mode-p 'emacs-lisp-mode)\n        (let ((sym (intern candidate)))\n          (cond ((fboundp sym) 'Function)\n                ((featurep sym) 'Module)\n                ((facep sym) 'Color)\n                ((boundp sym) 'Variable)\n                ((symbolp sym) 'Text)\n                (t . nil)))))\n    (with-eval-after-load 'all-the-icons\n      (declare-function all-the-icons-faicon 'all-the-icons)\n      (declare-function all-the-icons-fileicon 'all-the-icons)\n      (declare-function all-the-icons-material 'all-the-icons)\n      (declare-function all-the-icons-octicon 'all-the-icons)\n      (setq company-box-icons-all-the-icons\n            `((Unknown . ,(all-the-icons-material \"find_in_page\" :height 0.7 :v-adjust -0.15))\n              (Text . ,(all-the-icons-faicon \"book\" :height 0.68 :v-adjust -0.15))\n              (Method . ,(all-the-icons-faicon \"cube\" :height 0.7 :v-adjust -0.05 :face 'font-lock-constant-face))\n              (Function . ,(all-the-icons-faicon \"cube\" :height 0.7 :v-adjust -0.05 :face 'font-lock-constant-face))\n              (Constructor . ,(all-the-icons-faicon \"cube\" :height 0.7 :v-adjust -0.05 :face 'font-lock-constant-face))\n              (Field . ,(all-the-icons-faicon \"tags\" :height 0.65 :v-adjust -0.15 :face 'font-lock-warning-face))\n              (Variable . ,(all-the-icons-faicon \"tag\" :height 0.7 :v-adjust -0.05 :face 'font-lock-warning-face))\n              (Class . ,(all-the-icons-faicon \"clone\" :height 0.65 :v-adjust 0.01 :face 'font-lock-constant-face))\n              (Interface . ,(all-the-icons-faicon \"clone\" :height 0.65 :v-adjust 0.01))\n              (Module . ,(all-the-icons-octicon \"package\" :height 0.7 :v-adjust -0.15))\n              (Property . ,(all-the-icons-octicon \"package\" :height 0.7 :v-adjust -0.05 :face 'font-lock-warning-face)) ;; Golang module\n              (Unit . ,(all-the-icons-material \"settings_system_daydream\" :height 0.7 :v-adjust -0.15))\n              (Value . ,(all-the-icons-material \"format_align_right\" :height 0.7 :v-adjust -0.15 :face 'font-lock-constant-face))\n              (Enum . ,(all-the-icons-material \"storage\" :height 0.7 :v-adjust -0.15 :face 'all-the-icons-orange))\n              (Keyword . ,(all-the-icons-material \"filter_center_focus\" :height 0.7 :v-adjust -0.15))\n              (Snippet . ,(all-the-icons-faicon \"code\" :height 0.7 :v-adjust 0.02 :face 'font-lock-variable-name-face))\n              (Color . ,(all-the-icons-material \"palette\" :height 0.7 :v-adjust -0.15))\n              (File . ,(all-the-icons-faicon \"file-o\" :height 0.7 :v-adjust -0.05))\n              (Reference . ,(all-the-icons-material \"collections_bookmark\" :height 0.7 :v-adjust -0.15))\n              (Folder . ,(all-the-icons-octicon \"file-directory\" :height 0.7 :v-adjust -0.05))\n              (EnumMember . ,(all-the-icons-material \"format_align_right\" :height 0.7 :v-adjust -0.15 :face 'all-the-icons-blueb))\n              (Constant . ,(all-the-icons-faicon \"tag\" :height 0.7 :v-adjust -0.05))\n              (Struct . ,(all-the-icons-faicon \"clone\" :height 0.65 :v-adjust 0.01 :face 'font-lock-constant-face))\n              (Event . ,(all-the-icons-faicon \"bolt\" :height 0.7 :v-adjust -0.05 :face 'all-the-icons-orange))\n              (Operator . ,(all-the-icons-fileicon \"typedoc\" :height 0.65 :v-adjust 0.05))\n              (TypeParameter . ,(all-the-icons-faicon \"hashtag\" :height 0.65 :v-adjust 0.07 :face 'font-lock-const-face))\n              (Template . ,(all-the-icons-faicon \"code\" :height 0.7 :v-adjust 0.02 :face 'font-lock-variable-name-face))))))\n    ;; Show quick tooltip\n    (use-package company-quickhelp\n      :defines company-quickhelp-delay\n      :bind (:map company-active-map\n                  (\"M-h\" . company-quickhelp-manual-begin))\n      :hook (global-company-mode . company-quickhelp-mode)\n      :custom (company-quickhelp-delay 0.8)))<\/pre><\/div><h3 id=\"git\">1.12 Git<\/h3><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package git-timemachine\n  :bind (\"M-g t\" . git-timemachine-toggle))<\/pre><\/div><h4 id=\"diffview\">1.12.2 Diffview<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package diffview\n  :commands (diffview-region diffview-current)\n  :preface\n  (defun ladicle\/diffview-dwim ()\n    (interactive)\n    (if (region-active-p)\n        (diffview-region)\n      (diffview-current)))\n  :bind (\"M-g v\" . ladicle\/diffview-dwim))<\/pre><\/div><h4 id=\"magit\">1.12.3 Magit<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package magit\n  :custom\n  (magit-auto-revert-mode nil)\n  :bind\n  (\"M-g s\" . magit-status))<\/pre><\/div><h4 id=\"gitmodes\">1.12.4 GitModes<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package gitattributes-mode :defer t)\n(use-package gitconfig-mode :defer t)\n(use-package gitignore-mode :defer t)<\/pre><\/div><h4 id=\"gitgutter\">1.12.5 GitGutter<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package git-gutter\n    :custom\n    (git-gutter:modified-sign \"~\")              ; \uf46d\n    (git-gutter:added-sign    \"+\")              ; \uf46b\n    (git-gutter:deleted-sign  \"-\")              ; \uf46c\n    :custom-face\n    (git-gutter:modified ((t (:foreground \"#f1fa8c\" :background \"#f1fa8c\"))))\n    (git-gutter:added    ((t (:foreground \"#50fa7b\" :background \"#50fa7b\"))))\n    (git-gutter:deleted  ((t (:foreground \"#ff79c6\" :background \"#ff79c6\"))))\n    :config\n    (global-git-gutter-mode +1))<\/pre><\/div><h4 id=\"git-remote\">1.12.6 Git Remote<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package browse-at-remote\n  :bind (\"M-g r\" . browse-at-remote))<\/pre><\/div><h4 id=\"github\">1.12.7 GitHub<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package github-pullrequest)<\/pre><\/div><h4 id=\"smerge\">1.12.8 Smerge<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package smerge-mode\n  :diminish\n  :preface\n  (with-eval-after-load 'hydra\n    (defhydra smerge-hydra\n      (:color pink :hint nil :post (smerge-auto-leave))\n      \"\n^Move^       ^Keep^               ^Diff^                 ^Other^\n^^-----------^^-------------------^^---------------------^^-------\n_n_ext       _b_ase               _&lt;_: upper\/base        _C_ombine\n_p_rev       _u_pper              _=_: upper\/lower       _r_esolve\n^^           _l_ower              _&gt;_: base\/lower        _k_ill current\n^^           _a_ll                _R_efine\n^^           _RET_: current       _E_diff\n\"\n      (\"n\" smerge-next)\n      (\"p\" smerge-prev)\n      (\"b\" smerge-keep-base)\n      (\"u\" smerge-keep-upper)\n      (\"l\" smerge-keep-lower)\n      (\"a\" smerge-keep-all)\n      (\"RET\" smerge-keep-current)\n      (\"\\C-m\" smerge-keep-current)\n      (\"&lt;\" smerge-diff-base-upper)\n      (\"=\" smerge-diff-upper-lower)\n      (\"&gt;\" smerge-diff-base-lower)\n      (\"R\" smerge-refine)\n      (\"E\" smerge-ediff)\n      (\"C\" smerge-combine-with-next)\n      (\"r\" smerge-resolve)\n      (\"k\" smerge-kill-current)\n      (\"ZZ\" (lambda ()\n              (interactive)\n              (save-buffer)\n              (bury-buffer))\n       \"Save and bury buffer\" :color blue)\n      (\"q\" nil \"cancel\" :color blue)))\n  :hook ((find-file . (lambda ()\n                        (save-excursion\n                          (goto-char (point-min))\n                          (when (re-search-forward \"^&lt;&lt;&lt;&lt;&lt;&lt;&lt; \" nil t)\n                            (smerge-mode 1)))))\n         (magit-diff-visit-file . (lambda ()\n                                    (when smerge-mode\n                                      (smerge-hydra\/body))))))<\/pre><\/div><h3 id=\"tools\">1.13 Tools<\/h3><h4 id=\"google-translater\">1.13.1 Google Translater<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package google-translate\n  :bind\n  (\"M-o t\" . google-translate-at-point)\n  (\"M-o T\" . google-translate-at-point-reverse)\n  :custom\n  (google-translate-default-source-language \"en\")\n  (google-translate-default-target-language \"ja\"))<\/pre><\/div><h4 id=\"google-this\">1.13.2 google this!<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package google-this)<\/pre><\/div><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package wakatime-mode\n  :custom (wakatime-cli-path \"\/usr\/local\/bin\/wakatime\")\n  :config (global-wakatime-mode))<\/pre><\/div><h3 id=\"language\">1.14 Language<\/h3><h4 id=\"lsp\">1.14.1 LSP<\/h4><ul><li><a href=\"https:\/\/github.com\/saibing\/bingo\">bingo(LSP server)<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/go-delve\/delve\">delve(Go debugger)<\/a><\/li>\n<li><a href=\"https:\/\/marketplace.visualstudio.com\/items?itemName=ms-vscode.Go\">VSCode.Go(Dap delve server)<\/a><\/li>\n<li>\n<p><a href=\"https:\/\/marketplace.visualstudio.com\/items?itemName=webfreak.debug\">VSCode.NativeDebug(Dap GDB-LLDB server)<\/a><\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package lsp-mode\n:custom\n;; debug\n(lsp-print-io nil)\n(lsp-trace nil)\n(lsp-print-performance nil)\n;; general\n(lsp-auto-guess-root t)\n(lsp-document-sync-method 'incremental) ;; none, full, incremental, or nil\n(lsp-response-timeout 10)\n(lsp-prefer-flymake t) ;; t(flymake), nil(lsp-ui), or :none\n;; go-client\n(lsp-clients-go-server-args '(\"--cache-style=always\" \"--diagnostics-style=onsave\" \"--format-style=goimports\"))\n:hook\n((go-mode c-mode c++-mode) . lsp)\n:bind\n(:map lsp-mode-map\n(\"C-c r\"   . lsp-rename))\n:config\n(require 'lsp-clients)\n;; LSP UI tools\n(use-package lsp-ui\n:custom\n;; lsp-ui-doc\n(lsp-ui-doc-enable nil)\n(lsp-ui-doc-header t)\n(lsp-ui-doc-include-signature nil)\n(lsp-ui-doc-position 'at-point) ;; top, bottom, or at-point\n(lsp-ui-doc-max-width 120)\n(lsp-ui-doc-max-height 30)\n(lsp-ui-doc-use-childframe t)\n(lsp-ui-doc-use-webkit t)\n;; lsp-ui-flycheck\n(lsp-ui-flycheck-enable nil)\n;; lsp-ui-sideline\n(lsp-ui-sideline-enable nil)\n(lsp-ui-sideline-ignore-duplicate t)\n(lsp-ui-sideline-show-symbol t)\n(lsp-ui-sideline-show-hover t)\n(lsp-ui-sideline-show-diagnostics nil)\n(lsp-ui-sideline-show-code-actions t)\n(lsp-ui-sideline-code-actions-prefix \"\uf400\")\n;; lsp-ui-imenu\n(lsp-ui-imenu-enable t)\n(lsp-ui-imenu-kind-position 'top)\n;; lsp-ui-peek\n(lsp-ui-peek-enable t)\n(lsp-ui-peek-peek-height 20)\n(lsp-ui-peek-list-width 50)\n(lsp-ui-peek-fontify 'on-demand) ;; never, on-demand, or always\n:preface\n(defun ladicle\/toggle-lsp-ui-doc ()\n  (interactive)\n  (if lsp-ui-doc-mode\n    (progn\n      (lsp-ui-doc-mode -1)\n      (lsp-ui-doc--hide-frame))\n     (lsp-ui-doc-mode 1)))\n:bind\n(:map lsp-mode-map\n(\"C-c C-r\" . lsp-ui-peek-find-references)\n(\"C-c C-j\" . lsp-ui-peek-find-definitions)\n(\"C-c i\"   . lsp-ui-peek-find-implementation)\n(\"C-c m\"   . lsp-ui-imenu)\n(\"C-c s\"   . lsp-ui-sideline-mode)\n(\"C-c d\"   . ladicle\/toggle-lsp-ui-doc))\n:hook\n(lsp-mode . lsp-ui-mode))\n(lsp-register-client\n(make-lsp-client :new-connection (lsp-stdio-connection\n                             (lambda () (cons \"bingo\"\n                                              lsp-clients-go-server-args)))\n            :major-modes '(go-mode)\n            :priority 2\n            :initialization-options 'lsp-clients-go--make-init-options\n            :server-id 'go-bingo\n            :library-folders-fn (lambda (_workspace)\n                                  lsp-clients-go-library-directories)))\n;; DAP\n(use-package dap-mode\n:custom\n(dap-go-debug-program `(\"node\" \"~\/.extensions\/go\/out\/src\/debugAdapter\/goDebug.js\"))\n:config\n(dap-mode 1)\n(require 'dap-hydra)\n(require 'dap-gdb-lldb)     ; download and expand lldb-vscode to the =~\/.extensions\/webfreak.debug=\n(require 'dap-go)           ; download and expand vscode-go-extenstion to the =~\/.extensions\/go=\n(use-package dap-ui\n  :ensure nil\n  :config\n  (dap-ui-mode 1)))\n;; Lsp completion\n(use-package company-lsp\n:custom\n(company-lsp-cache-candidates t) ;; auto, t(always using a cache), or nil\n(company-lsp-async t)\n(company-lsp-enable-snippet t)\n(company-lsp-enable-recompletion t)))<\/pre><\/div>\n<\/li>\n<\/ul><h4 id=\"golang\">1.14.2 Golang<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package go-mode\n  :mode \"\\\\.go\\\\'\"\n  :custom (gofmt-command \"goimports\")\n  :bind (:map go-mode-map\n         (\"C-c C-n\" . go-run)\n         (\"C-c .\"   . go-test-current-test)\n         (\"C-c f\"   . go-test-current-file)\n         (\"C-c a\"   . go-test-current-project))\n  :config\n  (add-hook 'before-save-hook #'gofmt-before-save)\n  (use-package gotest)\n  (use-package go-tag\n    :config (setq go-tag-args (list \"-transform\" \"camelcase\"))))<\/pre><\/div><h4 id=\"c-c\">1.14.3 C\/C++<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package cc-mode\n  :bind (:map c-mode-base-map\n         (\"C-c c\" . compile))\n  :hook (c-mode-common . (lambda ()\n                            (c-set-style \"bsd\")\n                            (setq tab-width 4)\n                            (setq c-base-offset 4))))\n(use-package ccls\n  :custom\n  (ccls-executable \"\/usr\/local\/bin\/ccls\")\n  (ccls-sem-highlight-method 'font-lock)\n  :config\n  :hook ((c-mode c++-mode objc-mode) .\n         (lambda () (require 'ccls) (lsp))))<\/pre><\/div><h4 id=\"plantuml\">1.14.4 PlantUML<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package plantuml-mode\n  :custom\n  (plantuml-jar-path \"~\/.emacs.d\/plantuml.jar\")\n  :mode \"\\\\.uml\\\\'\")<\/pre><\/div><h4 id=\"org\">1.14.5 Org<\/h4><ol><li>\n<p>ob-async<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package ob-async\n  :load-path \"~\/Developments\/src\/github.com\/astahlman\/ob-async\"\n  :config\n  (add-hook 'org-babel-after-execute-hook 'org-redisplay-inline-images)\n  (add-hook 'ob-async-pre-execute-src-block-hook\n      '(lambda ()\n         (setq org-plantuml-jar-path \"~\/.emacs.d\/plantuml.jar\"))))<\/pre><\/div>\n<\/li>\n<li>\n<p>screenshot<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun my-org-screenshot ()\n  \"Take a screenshot into a time stamped unique-named file in the\nsame directory as the org-buffer and insert a link to this file.\"\n  (interactive)\n  (org-display-inline-images)\n  (setq filename\n        (concat\n         (make-temp-name\n          (concat (file-name-nondirectory (buffer-file-name))\n                  \"_imgs\/\"\n                  (format-time-string \"%Y%m%d_%H%M%S_\")) ) \".png\"))\n  (unless (file-exists-p (file-name-directory filename))\n    (make-directory (file-name-directory filename)))\n  ; take screenshot\n  (if (eq system-type 'darwin)\n      (call-process \"screencapture\" nil nil nil \"-i\" filename))\n  (if (eq system-type 'gnu\/linux)\n      (call-process \"import\" nil nil nil filename))\n  ; insert into file if correctly taken\n  (if (file-exists-p filename)\n    (insert (concat \"[[file:\" filename \"]]\"))))<\/pre><\/div>\n<\/li>\n<li>\n<p>org<\/p>\n<div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">  (use-package org\n    :custom\n    (org-src-fontify-natively t)\n    (private-directory \"~\/Private\/\")\n    (task-file (concat private-directory \"task.org\"))\n    (schedule-file (concat private-directory \"schedule.org\"))\n    (org-directory private-directory)\n    (org-ditaa-jar-path \"\/usr\/share\/ditaa\/ditaa.jar\")\n    (org-plantuml-jar-path \"~\/.emacs.d\/plantuml.jar\")\n    (org-confirm-babel-evaluate nil)\n    (org-clock-out-remove-zero-time-clocks t)\n    (org-startup-folded 'content)\n    (org-columns-default-format \"%50ITEM(Task) %5TODO(Todo) %10Effort(Effort){:} %10CLOCKSUM(Clock) %2PRIORITY %TAGS\")\n    (org-agenda-columns-add-appointments-to-effort-sum t)\n    (org-agenda-span 'day)\n    (org-agenda-log-mode-items (quote (closed clock)))\n    (org-agenda-clockreport-parameter-plist\n      '(:maxlevel 5 :block t :tstart t :tend t :emphasize t :link t :narrow 80 :indent t :formula nil :timestamp t :level 5 :tcolumns nil :formatter nil))\n    (org-global-properties (quote ((\n      \"Effort_ALL\" . \"00:05 00:10 00:15 00:30 01:00 01:30 02:00 02:30 03:00\"))))\n    (org-agenda-files (quote (\n      \"~\/Private\/task.org\"\n      \"~\/Private\/routine.org\"\n      \"~\/Private\/task.org_archive\"\n      \"~\/Private\/schedule.org\")))\n    :custom-face\n    (org-link ((t (:foreground \"#ebe087\" :underline t))))\n    (org-list-dt ((t (:foreground \"#bd93f9\"))))\n    (org-special-keyword ((t (:foreground \"#6272a4\"))))\n    (org-todo ((t (:background \"#272934\" :foreground \"#51fa7b\" :weight bold))))\n    (org-document-title ((t (:foreground \"#f1fa8c\" :weight bold))))\n    (org-done ((t (:background \"#373844\" :foreground \"#216933\" :strike-through nil :weight bold))))\n    (org-footnote ((t (:foreground \"#76e0f3\"))))\n    ;; do not scale outline header\n    ;; (org-level-1 ((t (:inherit outline-1 :height 1.0))))\n    ;; (org-level-2 ((t (:inherit outline-2 :height 1.0))))\n    ;; (org-level-3 ((t (:inherit outline-3 :height 1.0))))\n    ;; (org-level-4 ((t (:inherit outline-4 :height 1.0))))\n    ;; (org-level-5 ((t (:inherit outline-5 :height 1.0))))\n    :bind ((\"M-o c\" . counsel-org-capture)\n           (\"M-o a\" . org-agenda)\n           (\"C-x C-l\" . org-store-link)\n           :map org-mode-map\n           (\"C-c i\" . org-clock-in)\n           (\"C-c o\" . org-clock-out)\n           (\"C-c n\" . org-narrow-to-subtree)\n           (\"C-c b\" . org-narrow-to-block)\n           (\"C-c w\" . widen)\n           (\"C-c e\" . org-set-effort)\n           ;; custom functions\n           (\"C-c 1\" . (lambda () (interactive) (org-cycle-list-bullet 2)))\n           (\"M-o l i\" . (lambda () (interactive) (ladicle\/open-org-file task-file)))\n           (\"M-o l s\" . (lambda () (interactive) (ladicle\/open-org-file schedule-file)))\n           (\"M-o l y\" . (lambda () (interactive) (ladicle\/open-org-file (ladicle\/get-yesterday-diary))))\n           (\"M-o l p\" . (lambda () (interactive) (ladicle\/open-org-file (ladicle\/get-diary-from-cal))))\n           (\"M-o l t\" . (lambda () (interactive) (ladicle\/open-org-file (ladicle\/get-today-diary)))))\n    :hook\n    (kill-emacs . ladicle\/org-clock-out-and-save-when-exit)\n    (org-clock-in .\n                (lambda ()\n                  (setq org-mode-line-string (ladicle\/task-clocked-time))\n                  (run-at-time 0 60 '(lambda ()\n                                       (setq org-mode-line-string (ladicle\/task-clocked-time))\n                                       (force-mode-line-update)))\n                  (force-mode-line-update)))\n    (org-mode . (lambda ()\n                       (dolist (key '(\"C-'\" \"C-,\" \"C-.\"))\n                         (unbind-key key org-mode-map))))\n    :preface\n    (defun ladicle\/get-today-diary ()\n      (concat private-directory\n        (format-time-string \"diary\/%Y\/%m\/%Y-%m-%d.org\" (current-time))))\n    (defun ladicle\/get-yesterday-diary ()\n      (concat private-directory\n        (format-time-string \"diary\/%Y\/%m\/%Y-%m-%d.org\" (time-add (current-time) (* -24 3600)))))\n    (defun ladicle\/get-diary-from-cal ()\n      (concat private-directory\n        (format-time-string \"diary\/%Y\/%m\/%Y-%m-%d.org\"\n          (apply 'encode-time (parse-time-string (concat (org-read-date) \" 00:00\"))))))\n    (defun ladicle\/open-org-file (fname)\n      (switch-to-buffer (find-file-noselect fname)))\n    (defun ladicle\/org-get-time ()\n      (format-time-string \"&lt;%H:%M&gt;\" (current-time)))\n    (defun ladicle\/code-metadata ()\n      (concat \":\" (projectile-project-name) \":\"))\n    (defun ladicle\/org-clock-out-and-save-when-exit ()\n        \"Save buffers and stop clocking when kill emacs.\"\n          (ignore-errors (org-clock-out) t)\n          (save-some-buffers t))\n    (defun ladicle\/task-clocked-time ()\n        \"Return a string with the clocked time and effort, if any\"\n        (interactive)\n        (let* ((clocked-time (org-clock-get-clocked-time))\n               (h (truncate clocked-time 60))\n               (m (mod clocked-time 60))\n               (work-done-str (format \"%d:%02d\" h m)))\n          (if org-clock-effort\n              (let* ((effort-in-minutes\n                      (org-duration-to-minutes org-clock-effort))\n                     (effort-h (truncate effort-in-minutes 60))\n                     (effort-m (truncate (mod effort-in-minutes 60)))\n                     (effort-str (format \"%d:%02d\" effort-h effort-m)))\n                (format \"\uf446%s\/%s\" work-done-str effort-str))\n            (format \"\uf446%s\" work-done-str))))\n    :config\n    ;; Pomodoro\n    (use-package org-pomodoro\n    :after org-agenda\n    :custom\n    (org-pomodoro-ask-upon-killing t)\n    (org-pomodoro-format \"\ue001%s\") ;; \ue004 \ue001 \ue002 \ue003 \uf252\n    (org-pomodoro-short-break-format \"\uf0f4%s\")\n    (org-pomodoro-long-break-format  \"\ue006%s\")\n    :custom-face\n    (org-pomodoro-mode-line ((t (:foreground \"#ff5555\"))))\n    (org-pomodoro-mode-line-break   ((t (:foreground \"#50fa7b\"))))\n    :hook\n    (org-pomodoro-started . (lambda () (notifications-notify\n                                               :title \"org-pomodoro\"\n                                               :body \"Let's focus for 25 minutes!\"\n                                               :app-icon \"~\/.emacs.d\/img\/001-food-and-restaurant.png\")))\n    (org-pomodoro-finished . (lambda () (notifications-notify\n                                               :title \"org-pomodoro\"\n                                               :body \"Well done! Take a break.\"\n                                               :app-icon \"~\/.emacs.d\/img\/004-beer.png\")))\n    :config\n    (when (eq system-type 'darwin)\n      (setq alert-default-style 'osx-notifier))\n    (require 'alert)\n    :bind (:map org-agenda-mode-map\n                (\"p\" . org-pomodoro)))\n    (setq org-agenda-current-time-string \"\u2190 now\")\n    (setq org-agenda-time-grid ;; Format is changed from 9.1\n        '((daily today require-timed)\n          (0900 01000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400)\n          \"-\"\n          \"\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\"))\n    (setq org-todo-keyword-faces\n            '((\"WAIT\" . (:foreground \"#6272a4\":weight bold))\n              (\"NEXT\"   . (:foreground \"#f1fa8c\" :weight bold))\n              (\"CARRY\/O\" . (:foreground \"#6272a4\" :background \"#373844\" :weight bold))))\n    (setq org-capture-templates\n          '((\"tweet\" \"Write down the thoughts of this moment with a timestamp.\" item\n             (file+headline ladicle\/get-today-diary \"Log\")\n             \"%(ladicle\/org-get-time) %? \\n\")\n            ;; memo\n            (\"memo\" \"Memorize something in the memo section of today's diary.\" entry\n             (file+headline ladicle\/get-today-diary \"Memo\")\n             \"** %? \\n\"\n             :unnarrowed 1)\n            ;; tasks\n            (\"inbox\" \"Create a general task to the inbox and jump to the task file.\" entry\n             (file+headline \"~\/Private\/task.org\" \"Inbox\")\n             \"** TODO %?\"\n             :jump-to-captured 1)\n            (\"interrupt-task\" \"Create an interrupt task to the inbox and start clocking.\" entry\n             (file+headline \"~\/Private\/task.org\" \"Inbox\")\n             \"** TODO %?\"\n             :jump-to-captured 1 :clock-in 1 :clock-resume 1)\n            (\"hack-emacs\" \"Collect hacking Emacs ideas!\" item\n             (file+headline \"~\/Private\/task.org\" \"Hacking Emacs\")\n             \"- [ ] %?\"\n             :prepend t)\n            (\"private-schedule\" \"Add an event to the private calendar.\" entry\n             (file+olp schedule-file \"Calendar\" \"2019\" \"Private\")\n             \"** %?\\n   SCHEDULED: &lt;%(org-read-date)&gt;\\n\"\n             :prepend t)\n            (\"work-schedule\" \"Add an event to the work calendar.\" entry\n             (file+olp schedule-file \"Calendar\" \"2019\" \"Work\")\n             \"** %?\\n   SCHEDULED: &lt;%(org-read-date)&gt;\\n\")\n            (\"store-link\" \"Store the link of the current position in the clocking task.\" item\n             (clock)\n             \"- %A\\n\"\n             :immediate t :prepend t)\n            ;; code-reading\n            (\"code-link\" \"Store the code reading memo to today's diary with metadata.\" entry\n             (file+headline ladicle\/get-today-diary \"Code\")\n             ;;(file+headline ladicle\/get-today-diary \"Code\")\n             \"** %? %(ladicle\/code-metadata)\\n%A\\n\")))\n    ;; delete unused modules\n    (setq org-modules (delete '(org-gnus org-w3m org-bbdb org-bibtex org-docview org-info org-irc org-mhe org-rmail org-eww) org-modules))\n    ;; load org-protocol for capturing websites\n    (use-package org-protocol\n      :ensure nil)\n    ;; load babel languages\n    (org-babel-do-load-languages\n         'org-babel-load-languages\n         '((dot . t)\n           (plantuml . t)\n           (latex . t)\n           (R . t)))\n    ;; Pretty bullets\n    (use-package org-bullets\n      :custom (org-bullets-bullet-list '(\"\uf219\" \"\uf8a4\" \"\uf8a7\" \"\uf8aa\" \"\uf8ad\" \"\uf8b0\" \"\uf8b3\" \"\uf8b6\" \"\uf8b9\" \"\uf8bc\"))\n      :hook (org-mode . org-bullets-mode))\n  (use-package ox-hugo\n    :after ox\n    :custom\n    (org-blackfriday--org-element-string '((src-block . \"Code\")\n                                              (table . \"Table\")\n                                              (figure . \"Figure\"))))\n    ;; Download Drag&amp;Drop images\n    (use-package org-download)\n  (with-eval-after-load 'hydra\n    (eval-and-compile\n      (defun hot-expand (str &amp;optional mod)\n        \"Expand org template.\"\n        (let (text)\n          (when (region-active-p)\n            (setq text (buffer-substring (region-beginning) (region-end)))\n            (delete-region (region-beginning) (region-end)))\n          (insert str)\n          (org-try-structure-completion)\n          (when mod (insert mod) (forward-line))\n          (when text (insert text)))))\n    (defhydra hydra-org-template (:color blue :hint nil)\n      (format \"   %s^^     %s^^^^       %s^^^^^        %s\n%s\n  %s _e_lisp      %s _h_ugo        %s plant_u_ml      %s _s_ource\n  %s _r_uby       %s _c_aption     %s La_t_ex         %s _n_ote\n  %s _f_ish       %s _l_ink        %s i_P_ython       %s _i_nfo\n  %s _b_ash       %s %s^^^^^       %s %s^^^^          %s qu_o_te\n  %s _g_o\n  %s _y_aml\n%s\n                     %s quit%s insert\n\"\n(concat (propertize \"((\" 'face `(:foreground \"#6272a4\"))\n        (propertize \"CODE\" 'face `(:foreground \"#ff79c6\" :weight bold))\n        (propertize \"))\" 'face `(:foreground \"#6272a4\")))\n(concat (propertize \"((\" 'face `(:foreground \"#6272a4\"))\n        (propertize \"META\" 'face `(:foreground \"#ff79c6\" :weight bold))\n        (propertize \"))\" 'face `(:foreground \"#6272a4\")))\n(concat (propertize \"((\" 'face `(:foreground \"#6272a4\"))\n        (propertize \"DRAW\" 'face `(:foreground \"#ff79c6\" :weight bold))\n        (propertize \"))\" 'face `(:foreground \"#6272a4\")))\n(concat (propertize \"((\" 'face `(:foreground \"#6272a4\"))\n        (propertize \"BLOCK\" 'face `(:foreground \"#ff79c6\" :weight bold))\n        (propertize \"))\" 'face `(:foreground \"#6272a4\")))\n(propertize \" \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 \" 'face `(:foreground \"#6272a4\"))\n;; L1\n(all-the-icons-fileicon \"emacs\" :v-adjust .00001 :height .68 :face '(:foreground \"#6272a4\"))\n(all-the-icons-material \"web\" :v-adjust -.1 :height .7 :face '(:foreground \"#6272a4\"))\n(all-the-icons-material \"format_shapes\" :v-adjust -.15 :height .7 :face '(:foreground \"#6272a4\"))\n(all-the-icons-octicon \"code\" :v-adjust -.05 :height .75  :face '(:foreground \"#6272a4\"))\n;; L2\n(all-the-icons-alltheicon \"ruby-alt\" :v-adjust .0505 :height .7 :face '(:foreground \"#6272a4\"))\n(all-the-icons-faicon \"flag\" :v-adjust -.05 :height .69  :face '(:foreground \"#6272a4\"))\n(all-the-icons-faicon \"text-height\" :v-adjust -.05 :height .69 :face '(:foreground \"#6272a4\"))\n(all-the-icons-octicon \"light-bulb\" :v-adjust -.1 :height .78 :face '(:foreground \"#6272a4\"))\n;; L3\n(all-the-icons-alltheicon \"script\" :v-adjust .05 :height .7 :face '(:foreground \"#6272a4\"))\n(all-the-icons-faicon \"link\" :v-adjust -.05 :height .69 :face '(:foreground \"#6272a4\"))\n(all-the-icons-fileicon \"test-python\" :v-adjust -.1 :height .75 :face '(:foreground \"#6272a4\"))\n(all-the-icons-faicon \"info-circle\" :v-adjust -.1 :height .72 :face '(:foreground \"#6272a4\"))\n;; L4\n(all-the-icons-alltheicon \"script\" :v-adjust .05 :height .7 :face '(:foreground \"#6272a4\"))\n(all-the-icons-fileicon \"test-python\" :v-adjust -.1 :height .7 :face '(:foreground \"#282a36\")) ;; dummy\n(propertize \"link\" 'face `(:foreground \"#282a36\"))\n(all-the-icons-fileicon \"test-python\" :v-adjust -.1 :height .7 :face '(:foreground \"#282a36\")) ;; dummy\n(propertize \"latex\" 'face `(:foreground \"#282a36\"))\n(all-the-icons-faicon \"quote-right\" :v-adjust -.05 :height .65 :face '(:foreground \"#6272a4\"))\n;; L5\n(all-the-icons-fileicon \"go\" :v-adjust -.1 :height .75 :face '(:foreground \"#6272a4\"))\n;; L6\n(all-the-icons-octicon \"settings\" :v-adjust -.1 :height .75 :face '(:foreground \"#6272a4\"))\n;; Draw\n(propertize \" \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518 \" 'face `(:foreground \"#6272a4\"))\n(propertize \"[_q_]:\" 'face `(:foreground \"#6272a4\"))\n(propertize \", [_&lt;_]:\" 'face `(:foreground \"#6272a4\"))\n)\n      (\"s\" (hot-expand \"&lt;s\"))\n      (\"o\" (hot-expand \"&lt;q\"))\n      (\"c\" (hot-expand \"&lt;c\"))\n      (\"t\" (hot-expand \"&lt;L\"))\n      (\"c\" (insert \"#+CAPTION: \"))\n      (\"l\" (insert \"#+NAME: \"))\n      (\"n\" (insert \"#+BEGIN_NOTE\\n\\n#+END_NOTE\"))\n      (\"i\" (insert \"#+BEGIN_INFO\\n\\n#+END_INFO\"))\n      (\"h\" (insert \":PROPERTIES:\\n:EXPORT_FILE_NAME:\\n:EXPORT_HUGO_SECTION: pages\\n:EXPORT_HUGO_TAGS:\\n:EXPORT_HUGO_CATEGORIES:\\n:END:\"))\n      (\"e\" (hot-expand \"&lt;s\" \"emacs-lisp\"))\n      (\"f\" (hot-expand \"&lt;s\" \"fish\"))\n      (\"b\" (hot-expand \"&lt;s\" \"bash\"))\n      (\"y\" (hot-expand \"&lt;s\" \"yaml\"))\n      (\"P\" (hot-expand \"&lt;s\" \"ipython :session :exports both :async :cache yes :results raw drawer\\n$0\"))\n      (\"g\" (hot-expand \"&lt;s\" \"go\"))\n      (\"r\" (hot-expand \"&lt;s\" \"ruby\"))\n      (\"S\" (hot-expand \"&lt;s\" \"sh\"))\n      (\"u\" (hot-expand \"&lt;s\" \"plantuml :file overview.svg :cache yes :cmdline -config \\\"$HOME\/Private\/style.uml\\\" :async\"))\n      (\"&lt;\" self-insert-command)\n      (\"q\" nil))\n    (bind-key \"&lt;\"\n              (lambda () (interactive)\n                (if (or (region-active-p) (looking-back \"^\\s*\" 1))\n                    (hydra-org-template\/body)\n                  (self-insert-command 1)))\n              org-mode-map))\n)<\/pre><\/div>\n<\/li>\n<\/ol><h4 id=\"r\">1.14.6 R<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package ess\n  :disabled\n  :custom (ess-ask-for-ess-directory nil)\n  :config\n  (require 'ess-site))<\/pre><\/div><h4 id=\"html-css-js\">1.14.7 HTML\/CSS\/JS<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package js2-mode\n  :mode (\"\\\\.js\\\\'\" . js2-mode))<\/pre><\/div><h4 id=\"protobuf\">1.14.8 Protobuf<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package protobuf-mode\n  :mode \"\\\\.proto$\")<\/pre><\/div><h4 id=\"fish\">1.14.9 Fish<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package fish-mode\n  :mode \"\\\\.fish\\\\'\")<\/pre><\/div><h4 id=\"markdown\">1.14.10 Markdown<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package markdown-mode\n  :custom\n  (markdown-hide-markup nil)\n  (markdown-bold-underscore t)\n  (markdown-italic-underscore t)\n  (markdown-header-scaling t)\n  (markdown-indent-function t)\n  (markdown-enable-math t)\n  (markdown-hide-urls nil)\n  :custom-face\n  (markdown-header-delimiter-face ((t (:foreground \"mediumpurple\"))))\n  (markdown-header-face-1 ((t (:foreground \"violet\" :weight bold :height 1.0))))\n  (markdown-header-face-2 ((t (:foreground \"lightslateblue\" :weight bold :height 1.0))))\n  (markdown-header-face-3 ((t (:foreground \"mediumpurple1\" :weight bold :height 1.0))))\n  (markdown-link-face ((t (:background \"#0e1014\" :foreground \"#bd93f9\"))))\n  (markdown-list-face ((t (:foreground \"mediumpurple\"))))\n  (markdown-pre-face ((t (:foreground \"#bd98fe\"))))\n  :mode \"\\\\.md\\\\'\")\n(use-package markdown-toc)<\/pre><\/div><h4 id=\"log\">1.14.11 Log<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package logview :defer t)<\/pre><\/div><h4 id=\"systemd\">1.14.12 Systemd<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package systemd\n  :mode\n  (\"\\\\.service\\\\'\" \"\\\\.timer\\\\'\" \"\\\\.target\\\\'\" \"\\\\.mount\\\\'\"\n   \"\\\\.automount\\\\'\" \"\\\\.slice\\\\'\" \"\\\\.socket\\\\'\" \"\\\\.path\\\\'\"\n   \"\\\\.netdev\\\\'\" \"\\\\.network\\\\'\" \"\\\\.link\\\\'\"))<\/pre><\/div><h4 id=\"yaml\">1.14.13 YAML<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package yaml-mode\n  :mode (\"\\\\.yaml\\\\'\" \"\\\\.yml\\\\'\")\n  :custom-face\n  (font-lock-variable-name-face ((t (:foreground \"violet\")))))<\/pre><\/div><h4 id=\"bazel\">1.14.14 Bazel<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package bazel-mode :defer t)<\/pre><\/div><h4 id=\"dockerfile\">1.14.15 Dockerfile<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package dockerfile-mode\n  :mode \"\\\\Dockerfile\\\\'\")<\/pre><\/div><h4 id=\"change-letter-case\">1.15.1 Change letter case<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun upcase-backward-word (arg)\n  (interactive \"p\")\n  (upcase-word (- arg)))\n(defun downcase-backward-word (arg)\n  (interactive \"p\")\n  (downcase-word (- arg)))\n(defun capitalize-backward-word (arg)\n  (interactive \"p\")\n  (capitalize-word (- arg)))\n(global-set-key (kbd \"C-M-u\")    'upcase-backward-word)\n(global-set-key (kbd \"C-M-l\")    'downcase-backward-word)\n(global-set-key (kbd \"M-c\")      'capitalize-backward-word)<\/pre><\/div><h4 id=\"change-smart-kill-word\">1.15.2 Change smart Kill-word<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun kill-word-at-point ()\n  (interactive)\n  (let ((char (char-to-string (char-after (point)))))\n    (cond\n     ((string= \" \" char) (delete-horizontal-space))\n     ((string-match \"[\\t\\n -@\\[-`{-~],.\u3001\u3002\" char) (kill-word 1))\n     (t (forward-char) (backward-word) (kill-word 1)))))\n(global-set-key (kbd \"M-d\")  'kill-word-at-point)<\/pre><\/div><h4 id=\"kill-word-or-region\">1.15.3 Kill word or Region<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun backward-kill-word-or-region (&amp;optional arg)\n  (interactive \"p\")\n  (if (region-active-p)\n      (call-interactively #'kill-region)\n    (backward-kill-word arg)))\n(global-set-key (kbd \"C-w\")  'backward-kill-word-or-region)<\/pre><\/div><h4 id=\"add-execution-flag-to-a-file-with-shelbang\">1.15.4 Add execution flag to a file with shelbang<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun add-chmod()\n    (interactive)\n    (save-restriction\n      (widen)\n      (let ((name (buffer-file-name)))\n        (if (and (not (string-match \":\" name))\n                 (not (string-match \"\/\\\\.[^\/]+$\" name))\n                 (equal \"#!\" (buffer-substring 1 (min 3 (point-max)))))\n            (progn (set-file-modes name (logior (file-modes name) 73))\n                   (message \"Wrote %s (chmod +x)\" name))))))\n(add-hook 'after-save-hook\n          'executable-make-buffer-file-executable-if-script-p)<\/pre><\/div><h4 id=\"copy-file-path\">1.15.5 Copy file path<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun put-current-path-to-clipboard ()\n    (interactive)\n    (let ((file-path buffer-file-name)\n          (dir-path default-directory))\n      (cond (file-path\n             (kill-new (expand-file-name file-path))\n             (message \"This file path is on the clipboard!\"))\n            (dir-path\n             (kill-new (expand-file-name dir-path))\n             (message \"This directory path is on the clipboard!\"))\n            (t\n             (error-message-string \"Fail to get path name.\")))))\n  (defun put-current-filename-to-clipboard ()\n    (interactive)\n    (let ((file-path buffer-file-name)\n          (dir-path default-directory))\n      (cond (file-path\n             (kill-new (file-name-nondirectory file-path))\n             (message \"This file path is on the clipboard!\"))\n            (dir-path\n             (kill-new (file-name-nondirectory dir-path))\n             (message \"This directory path is on the clipboard!\"))\n            (t\n             (error-message-string \"Fail to get path name.\")))))\n  (defun put-current-filename-with-line-to-clipboard ()\n    (interactive)\n    (let ((file-path buffer-file-name)\n          (dir-path default-directory))\n      (cond (file-path\n             (kill-new (format \"%s:%s\"\n                               (file-name-nondirectory file-path)\n                               (count-lines (point-min) (point))))\n             (message \"This file path is on the clipboard!\"))\n            (dir-path\n             (kill-new (file-name-nondirectory dir-path))\n             (message \"This directory path is on the clipboard!\"))\n            (t\n             (error-message-string \"Fail to get path name.\")))))<\/pre><\/div><h3 id=\"ui\">1.16 UI<\/h3><h4 id=\"dashboard\">1.16.1 Dashboard<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package dashboard\n  :diminish\n  (dashboard-mode page-break-lines-mode)\n  :custom\n  (dashboard-center-content t)\n  (dashboard-startup-banner 4)\n  (dashboard-items '((recents . 15)\n                     (projects . 5)\n                     (bookmarks . 5)))\n  :custom-face\n  (dashboard-heading ((t (:foreground \"#f1fa8c\" :weight bold))))\n  :hook\n  (after-init . dashboard-setup-startup-hook))<\/pre><\/div><h4 id=\"general\">1.16.2 General<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">;; Change cursor style\n(add-to-list 'default-frame-alist '(cursor-type . bar))\n;; vertical border\n(let ((display-table (or standard-display-table (make-display-table))))\n  (set-display-table-slot display-table 'vertical-border (make-glyph-code ?\u2502)) ; or \u2503 \u2502\n  (setq standard-display-table display-table))\n(set-face-background 'vertical-border \"#0e0f1b\")\n(set-face-foreground 'vertical-border (face-background 'vertical-border))<\/pre><\/div><h4 id=\"persp-1\">1.16.3 persp<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package persp-mode\n  :disabled\n  :diminish\n  :commands (get-current-persp persp-contain-buffer-p)\n  :hook (after-init . persp-mode))<\/pre><\/div><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package imenu-list\n  :load-path \"~\/Developments\/src\/github.com\/Ladicle\/imenu-list\"\n  :bind\n  (\"&lt;f10&gt;\" . imenu-list-smart-toggle)\n  :custom-face\n  (imenu-list-entry-face-1 ((t (:foreground \"white\"))))\n  :custom\n  (imenu-list-focus-after-activation t)\n  (imenu-list-auto-resize t))<\/pre><\/div><h4 id=\"transparency\">1.16.5 Transparency<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(defun ladicle\/toggle-window-transparency ()\n  \"Cycle the frame transparency from default to transparent.\"\n  (interactive)\n  (let ((transparency 85)\n        (opacity 100))\n    (if (and (not (eq (frame-parameter nil 'alpha) nil))\n             (&lt; (frame-parameter nil 'alpha) opacity))\n        (set-frame-parameter nil 'alpha opacity)\n      (set-frame-parameter nil 'alpha transparency))))\n(global-set-key (kbd \"M-t t\") 'ladicle\/toggle-window-transparency)<\/pre><\/div><h4 id=\"minimap\">1.16.6 Minimap<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package minimap\n  :commands\n  (minimap-bufname minimap-create minimap-kill)\n  :custom\n  (minimap-major-modes '(prog-mode))\n  (minimap-window-location 'right)\n  (minimap-update-delay 0.2)\n  (minimap-minimum-width 20)\n  :bind\n  (\"M-t p\" . ladicle\/toggle-minimap)\n  :preface\n  (defun ladicle\/toggle-minimap ()\n    \"Toggle minimap for current buffer.\"\n    (interactive)\n    (if (null minimap-bufname)\n        (minimap-create)\n      (minimap-kill)))\n  :config\n  (custom-set-faces\n   '(minimap-active-region-background\n    ((((background dark)) (:background \"#555555555555\"))\n      (t (:background \"#C847D8FEFFFF\"))) :group 'minimap)))<\/pre><\/div><h4 id=\"neotree\">1.16.7 Neotree<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package neotree\n  :after\n  projectile\n  :commands\n  (neotree-show neotree-hide neotree-dir neotree-find)\n  :custom\n  (neo-theme 'nerd2)\n  :bind\n  (\"&lt;f8&gt;\" . neotree-current-dir-toggle)\n  (\"&lt;f9&gt;\" . neotree-projectile-toggle)\n  :preface\n  (defun neotree-projectile-toggle ()\n    (interactive)\n    (let ((project-dir\n           (ignore-errors\n           ;;; Pick one: projectile or find-file-in-project\n             (projectile-project-root)\n             ))\n          (file-name (buffer-file-name))\n          (neo-smart-open t))\n      (if (and (fboundp 'neo-global--window-exists-p)\n               (neo-global--window-exists-p))\n          (neotree-hide)\n        (progn\n          (neotree-show)\n          (if project-dir\n              (neotree-dir project-dir))\n          (if file-name\n              (neotree-find file-name))))))\n  (defun neotree-current-dir-toggle ()\n    (interactive)\n    (let ((project-dir\n           (ignore-errors\n             (ffip-project-root)\n             ))\n          (file-name (buffer-file-name))\n          (neo-smart-open t))\n      (if (and (fboundp 'neo-global--window-exists-p)\n               (neo-global--window-exists-p))\n          (neotree-hide)\n        (progn\n          (neotree-show)\n          (if project-dir\n              (neotree-dir project-dir))\n          (if file-name\n              (neotree-find file-name)))))))<\/pre><\/div><h4 id=\"doom-dracula-theme-modeline\">1.16.8 Doom dracula Theme\/Modeline<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package doom-themes\n  :custom\n  (doom-themes-enable-italic t)\n  (doom-themes-enable-bold t)\n  :custom-face\n  ;; (vertical-bar   (doom-darken base5 0.4))\n  ;; (doom-darken bg 0.4)\n  :config\n  (load-theme 'doom-dracula t)\n  (doom-themes-neotree-config)\n  (doom-themes-org-config)\n  ;; Modeline\n  (use-package doom-modeline\n    :custom\n    (doom-modeline-buffer-file-name-style 'truncate-with-project)\n    (doom-modeline-icon t)\n    (doom-modeline-major-mode-icon nil)\n    (doom-modeline-minor-modes nil)\n    :hook\n    (after-init . doom-modeline-mode)\n    :config\n    (set-cursor-color \"cyan\")\n    (line-number-mode 0)\n    (column-number-mode 0)\n    (doom-modeline-def-modeline 'main\n      '(bar workspace-number window-number evil-state god-state ryo-modal xah-fly-keys matches buffer-info remote-host buffer-position parrot selection-info)\n      '(misc-info persp-name lsp github debug minor-modes input-method major-mode process vcs checker))))<\/pre><\/div><h4 id=\"nyan-mode\">1.16.9 Nyan-mode<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package nyan-mode\n   :custom\n   (nyan-cat-face-number 4)\n   (nyan-animate-nyancat t)\n   :hook\n   (doom-modeline-mode . nyan-mode))<\/pre><\/div><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package hide-mode-line\n  :hook\n  ((neotree-mode imenu-list-minor-mode minimap-mode) . hide-mode-line-mode))<\/pre><\/div><h4 id=\"line-number\">1.16.11 Line number<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package display-line-numbers\n  :ensure nil\n  :hook\n  ((prog-mode yaml-mode systemd-mode) . display-line-numbers-mode))<\/pre><\/div><h4 id=\"dimmer\">1.16.12 Dimmer<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package dimmer\n  :disabled\n  :custom\n  (dimmer-fraction 0.5)\n  (dimmer-exclusion-regexp-list\n       '(\".*Minibuf.*\"\n         \".*which-key.*\"\n         \".*NeoTree.*\"\n         \".*Messages.*\"\n         \".*Async.*\"\n         \".*Warnings.*\"\n         \".*LV.*\"\n         \".*Ilist.*\"))\n  :config\n  (dimmer-mode t))<\/pre><\/div><h4 id=\"fill-column-indicator\">1.16.13 fill-column-indicator<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package fill-column-indicator\n  :hook\n  ((markdown-mode\n    git-commit-mode) . fci-mode))<\/pre><\/div><h4 id=\"presentation\">1.16.14 Presentation<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package presentation)<\/pre><\/div><h4 id=\"eshell\">1.16.15 Eshell<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(setq eshell-prompt-function\n      (lambda ()\n        (format \"%s %s\\n%s%s%s \"\n                (all-the-icons-octicon \"repo\")\n                (propertize (cdr (shrink-path-prompt default-directory)) 'face `(:foreground \"white\"))\n                (propertize \"\u276f\" 'face `(:foreground \"#ff79c6\"))\n                (propertize \"\u276f\" 'face `(:foreground \"#f1fa8c\"))\n                (propertize \"\u276f\" 'face `(:foreground \"#50fa7b\")))))\n(setq eshell-hist-ignoredups t)\n(setq eshell-cmpl-cycle-completions nil)\n(setq eshell-cmpl-ignore-case t)\n(setq eshell-ask-to-save-history (quote always))\n(setq eshell-prompt-regexp \"\u276f\u276f\u276f \")\n(add-hook 'eshell-mode-hook\n          '(lambda ()\n             (progn\n               (define-key eshell-mode-map \"\\C-a\" 'eshell-bol)\n               (define-key eshell-mode-map \"\\C-r\" 'counsel-esh-history)\n               (define-key eshell-mode-map [up] 'previous-line)\n               (define-key eshell-mode-map [down] 'next-line)\n               )))<\/pre><\/div><h3 id=\"high-lights\">1.17 High-Lights<\/h3><h4 id=\"highlight-line\">1.17.1 Highlight Line<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package hl-line\n  :ensure nil\n  :hook\n  (after-init . global-hl-line-mode))<\/pre><\/div><h4 id=\"paren\">1.17.2 Paren<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package paren\n  :ensure nil\n  :hook\n  (after-init . show-paren-mode)\n  :custom-face\n  (show-paren-match ((nil (:background \"#44475a\" :foreground \"#f1fa8c\")))) ;; :box t\n  :custom\n  (show-paren-style 'mixed)\n  (show-paren-when-point-inside-paren t)\n  (show-paren-when-point-in-periphery t))<\/pre><\/div><h4 id=\"highlight-symbol\">1.17.3 Highlight-symbol<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package highlight-symbol\n  :bind\n  (:map prog-mode-map\n  (\"M-o h\" . highlight-symbol)\n  (\"M-p\" . highlight-symbol-prev)\n  (\"M-n\" . highlight-symbol-next)))<\/pre><\/div><h4 id=\"beacon\">1.17.4 Beacon<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package beacon\n  :custom\n  (beacon-color \"#f1fa8c\")\n  :hook (after-init . beacon-mode))<\/pre><\/div><h4 id=\"rainbow-delimiters\">1.17.5 Rainbow delimiters<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package rainbow-delimiters\n  :hook\n  (prog-mode . rainbow-delimiters-mode))<\/pre><\/div><h4 id=\"rainbow-mode\">1.17.6 Rainbow Mode<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package rainbow-mode\n  :diminish\n  :hook (emacs-lisp-mode . rainbow-mode))<\/pre><\/div><h4 id=\"volatile\">1.17.7 Volatile<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package volatile-highlights\n  :diminish\n  :hook\n  (after-init . volatile-highlights-mode)\n  :custom-face\n  (vhl\/default-face ((nil (:foreground \"#FF3333\" :background \"#FFCDCD\")))))<\/pre><\/div><h4 id=\"indent\">1.17.8 Indent<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package highlight-indent-guides\n  :diminish\n  :hook\n  ((prog-mode yaml-mode) . highlight-indent-guides-mode)\n  :custom\n  (highlight-indent-guides-auto-enabled t)\n  (highlight-indent-guides-responsive t)\n  (highlight-indent-guides-method 'character)) ; column<\/pre><\/div><h3 id=\"mark\">1.18 Mark<\/h3><h4 id=\"backward-forward\">1.18.1 backward-forward<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package backward-forward\n  :disabled\n  :bind\n  (\"C-,\" . backward-forward-previous-location)\n  (\"C-.\" . backward-forward-next-location)\n  :custom\n  (mark-ring-max 60)\n  (set-mark-command-repeat-pop t)\n  :config\n  (backward-forward-mode t))<\/pre><\/div><h4 id=\"point-undo\">1.18.2 Point Undo<\/h4><div class=\"highlight\"><pre class=\"language-emacs-lisp\" data-lang=\"emacs-lisp\">(use-package point-undo\n  :load-path \"~\/Developments\/src\/github.com\/emacsmirror\/point-undo\"\n  :bind\n  (\"C-,\" . point-undo)\n  (\"C-.\" . point-redo))<\/pre><\/div>","created_at":"2020-01-02T21:57:44+0000","updated_at":"2020-01-02T21:57:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":53,"domain_name":"ladicle.com","preview_picture":"https:\/\/ladicle.com\/img\/avatar.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1079,"title":"MatthewZMD\/.emacs.d","url":"https:\/\/github.com\/MatthewZMD\/.emacs.d","content":"<ul><li><a href=\"#orgd7ad0a6\">M-EMACS<\/a><\/li>\n<li><a href=\"#orgb3d18a2\">About EMACS<\/a><\/li>\n<li><a href=\"#orge08bf91\">About M-EMACS<\/a>\n<ul><li><a href=\"#org651f9ba\">Screenshot<\/a><\/li>\n<li><a href=\"#orgb26ffff\">About README<\/a><\/li>\n<li><a href=\"#orgac6b52e\">Installation<\/a><\/li>\n<li><a href=\"#orgb143513\">Modification<\/a><\/li>\n<li><a href=\"#org63137da\">Contribution<\/a><\/li>\n<li><a href=\"#orgafbca06\">Special Thanks<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#org9a9485d\">Startup<\/a>\n<ul><li><a href=\"#orga6636da\">Lexical Binding<\/a><\/li>\n<li><a href=\"#org149ccc4\">Early Init<\/a><\/li>\n<li><a href=\"#orgdb5cf46\">Garbage Collection<\/a><\/li>\n<li><a href=\"#orgaa3c88c\">Load Path<\/a><\/li>\n<li><a href=\"#orgc7cbaaf\">Define Constants<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#org5fa6009\">Package Management<\/a>\n<ul><li><a href=\"#orgbc2ad2f\">Melpa Packages<\/a><\/li>\n<li><a href=\"#org7b158ab\">Non-Melpa Packages<\/a><\/li>\n<li><a href=\"#org549c0b6\">Configure Package Manager<\/a><\/li>\n<li><a href=\"#org2b67329\">Use Package<\/a><\/li>\n<li><a href=\"#org3c71e8f\">Auto Package Update<\/a><\/li>\n<li><a href=\"#orgbfe90da\">Diminish<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#orge495bba\">Global Functionalities<\/a>\n<ul><li><a href=\"#orga1ce3ce\">User Information<\/a><\/li>\n<li><a href=\"#org907487a\">Bindings<\/a><\/li>\n<li><a href=\"#org0de9036\">Avy<\/a><\/li>\n<li><a href=\"#org4d82128\">Crux<\/a><\/li>\n<li><a href=\"#org25b9061\">Ivy, Amx, Counsel, Swiper<\/a><\/li>\n<li><a href=\"#org8b1893b\">Color Ripgrep<\/a><\/li>\n<li><a href=\"#org430bf19\">Snails<\/a><\/li>\n<li><a href=\"#orge862725\">Files Directories<\/a><\/li>\n<li><a href=\"#orgd3584ea\">Winner<\/a><\/li>\n<li><a href=\"#org29a6dbc\">Which Key<\/a><\/li>\n<li><a href=\"#org41bb9dd\">Popup Kill Ring<\/a><\/li>\n<li><a href=\"#org6c1c57e\">Undo Tree<\/a><\/li>\n<li><a href=\"#org017082d\">Discover My Major<\/a><\/li>\n<li><a href=\"#orge2b713f\">Ace Window<\/a><\/li>\n<li><a href=\"#orgbfbccc2\">Shells<\/a><\/li>\n<li><a href=\"#org89b1d23\">Sudo Edit<\/a><\/li>\n<li><a href=\"#org0b2cbc0\">Configs<\/a><\/li>\n<li><a href=\"#org55ce802\">Functions<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#org0f2fa21\">User Interface Enhancements<\/a>\n<ul><li><a href=\"#org114fa31\">Doom Themes<\/a><\/li>\n<li><a href=\"#org147550c\">Doom Modeline<\/a><\/li>\n<li><a href=\"#org69ec7f6\">Dashboard<\/a><\/li>\n<li><a href=\"#orgde152e6\">Fonts and Icons<\/a><\/li>\n<li><a href=\"#orgaba331d\">Smooth Scrolling<\/a><\/li>\n<li><a href=\"#org6c9ea88\">Prettify Symbols<\/a><\/li>\n<li><a href=\"#org03b1615\">UI Configs<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#org6047b89\">General Programming<\/a>\n<ul><li><a href=\"#org4bd2328\">Magit<\/a><\/li>\n<li><a href=\"#org6ceaa37\">Projectile<\/a><\/li>\n<li><a href=\"#org721f82e\">Treemacs<\/a><\/li>\n<li><a href=\"#orga179a9a\">YASnippet<\/a><\/li>\n<li><a href=\"#org15721e1\">Flycheck<\/a><\/li>\n<li><a href=\"#orgf55c60a\">Dumb Jump<\/a><\/li>\n<li><a href=\"#org9609254\">Parenthesis<\/a><\/li>\n<li><a href=\"#orgbaa760b\">Indentation<\/a><\/li>\n<li><a href=\"#org9b42efe\">Quickrun<\/a><\/li>\n<li><a href=\"#org13b554c\">Format All<\/a><\/li>\n<li><a href=\"#org5284b3d\">Evil Nerd Commenter<\/a><\/li>\n<li><a href=\"#org0e18695\">Editing<\/a><\/li>\n<li><a href=\"#org76d35ce\">Headers<\/a><\/li>\n<li><a href=\"#org34717b6\">Jupyter Notebook<\/a><\/li>\n<li><a href=\"#org2da8617\">LSP<\/a><\/li>\n<li><a href=\"#org5dd5e1a\">DAP<\/a><\/li>\n<li><a href=\"#org5c23782\">Company<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#org01531f0\">Programming<\/a>\n<ul><li><a href=\"#orgb3bd2ea\">Java<\/a><\/li>\n<li><a href=\"#orgebddafb\">C\/C++\/Objective C<\/a><\/li>\n<li><a href=\"#org48d9c05\">Golang<\/a><\/li>\n<li><a href=\"#orgc6c7929\">Python<\/a><\/li>\n<li><a href=\"#org0e788da\">ESS<\/a><\/li>\n<li><a href=\"#orgcf6f6cc\">TeX<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#org0721002\">Web Development<\/a>\n<ul><li><a href=\"#org4890429\">Web Mode<\/a><\/li>\n<li><a href=\"#orge445e1f\">JavaScript\/TypeScript<\/a><\/li>\n<li><a href=\"#orgcc8e8ee\">Emmet<\/a><\/li>\n<li><a href=\"#org71dcd8e\">Instant Rename Tag<\/a><\/li>\n<li><a href=\"#org08be382\">JSON<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#orgd6db8ed\">Miscellaneous<\/a>\n<ul><li><a href=\"#org614bc8a\">Org<\/a><\/li>\n<li><a href=\"#orgf7fc98d\">EAF<\/a><\/li>\n<li><a href=\"#org61fa912\">ERC<\/a><\/li>\n<li><a href=\"#org0fdbc91\">EWW<\/a><\/li>\n<li><a href=\"#orgc1f893e\">MU4E<\/a><\/li>\n<li><a href=\"#orga4d9a9a\">Tramp<\/a><\/li>\n<li><a href=\"#org1bd0add\">PDF Tools<\/a><\/li>\n<li><a href=\"#org7cd3b5e\">LeetCode<\/a><\/li>\n<li><a href=\"#orge0a6564\">Pyim<\/a><\/li>\n<li><a href=\"#orgac58870\">EPaint<\/a><\/li>\n<li><a href=\"#orgc51f078\">Tetris<\/a><\/li>\n<li><a href=\"#org9caf302\">Speed Type<\/a><\/li>\n<li><a href=\"#orgf2169e0\">2048 Game<\/a><\/li>\n<li><a href=\"#org6e3bfd1\">Zone<\/a><\/li>\n<\/ul><\/li>\n<\/ul><p>Emacs changes how you <em>think<\/em> about programming.<\/p><p>Emacs is <strong>totally introspectable<\/strong>. You can always find out 'what code runs when I press this button?'.<\/p><p>Emacs is an <strong>incremental programming environment<\/strong>. There's no edit-compile-run cycle. There isn't even an edit-run cycle. You can execute snippets of code and gradually turn them into a finished project. There's no distinction between your editor and your interpreter.<\/p><p>Emacs is a <strong>mutable environment<\/strong>. You can set variables, tweak functions with advice, or redefine entire functions. Nothing is off-limits.<\/p><p>Emacs provides <strong>functionality without applications<\/strong>. Rather than separate applications, functionality is all integrated into your Emacs instance. Amazingly, this works. Ever wanted to use the same snippet tool for writing C++ classes as well as emails?<\/p><p>Emacs is full of <strong>incredible software concepts that haven't hit the mainstream yet<\/strong>. For example:<\/p><ul><li>Many platforms have a single item clipboard. Emacs has an <strong>infinite clipboard<\/strong>.<\/li>\n<li>If you undo a change, and then continue editing, you can't redo the original change. Emacs allows <strong>undoing to any historical state<\/strong>, even allowing tree-based exploration of history.<\/li>\n<li>Emacs supports a <strong>reverse variable search<\/strong>: you can find variables with a given value.<\/li>\n<li>You can perform <strong>structural editing<\/strong> of code, allowing you to make changes without breaking syntax. This works for lisps (paredit) and non-lisps (smartparens).<\/li>\n<li>Many applications use a modal GUI: for example, you can't do other edits during a find-and-replace operation. Emacs provides <strong>recursive editing<\/strong> that allow you to suspend what you're currently doing, perform other edits, then continue the original task.<\/li>\n<\/ul><p>Emacs has a <strong>documentation culture<\/strong>. Emacs includes a usage manual, a lisp programming manual, pervasive docstrings and even an interactive tutorial.<\/p><p>Emacs has a <strong>broad ecosystem<\/strong>. If you want to edit code in a niche language, there's probably an Emacs package for it.<\/p><p>Emacs doesn't have a monopoly on good ideas, and there are other great tools out there. Nonetheless, we believe the <a href=\"https:\/\/i.stack.imgur.com\/7Cu9Z.jpg\" rel=\"nofollow\">Emacs learning curve<\/a> pays off.<\/p><p><em>This beautifully written <strong>About EMACS<\/strong> section credits to <a href=\"https:\/\/github.com\/remacs\/remacs\">Remacs<\/a>.<\/em><\/p><p>M-EMACS is a custom GNU Emacs setup and configurations that aims not only to enhance the default Emacs experience, and hopefully be a sample that everyone can easily navigate and reference through a highly detailed README that contains 99% of the <strong>entire<\/strong> configuration code.<\/p><p>As a young EMACSer, I have experienced the struggle to find a detailed configuration that is loosely coupled and highly readable. This mostly due to the nature of source codes, sometimes comments are harder to notice or simply not enough. Therefore I decided to construct this README and present any human-readable explanation in a much more human-friendly way. Anyone, particularly Emacs beginners who have no idea where to start with their personal config, is more than welcome to read through this document and copy\/paste any part to use it on their own.<\/p><p>This configuration is designed and tested for <strong>GNU Emacs 26.1 and above only<\/strong>. However, it is suggested to use <strong>emacs27<\/strong>, the latest version currently available.<\/p><h2><a id=\"user-content-screenshot\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshot\"><\/a>Screenshot<\/h2><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/MatthewZMD\/.emacs.d\/blob\/master\/images\/Sample.png\"><img src=\"https:\/\/github.com\/MatthewZMD\/.emacs.d\/raw\/master\/images\/Sample.png\" alt=\"img\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><h2><a id=\"user-content-about-readme\" class=\"anchor\" aria-hidden=\"true\" href=\"#about-readme\"><\/a>About README<\/h2><p>This README is originated from <code>init.org<\/code> that is generated using <code>M-x org-gfm-export-to-markdown<\/code>. Every block of code is generated through this function - it exports sections of code from the <code>elisp\/<\/code> directory. You will not see their presence in <code>init.org<\/code>.<\/p><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><ol><li>\n<p>Install <a href=\"https:\/\/www.gnu.org\/software\/emacs\/\" rel=\"nofollow\">GNU Emacs<\/a>.<\/p>\n<ul><li>\n<p>(Optional) On Ubuntu, <code>emacs-snapshot<\/code> is a great way to get latest version of Emacs.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>sudo add-apt-repository -y ppa:ubuntu-elisp\nsudo apt-get update\nsudo apt-get install emacs-snapshot<\/pre><\/div>\n<\/li>\n<li>\n<p>(Optional) Build latest Emacs from source.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre># Install essential build tools\nsudo apt-get install build-essential texinfo libx11-dev libxpm-dev libjpeg-dev libpng-dev libgif-dev libtiff-dev libgtk2.0-dev libncurses-dev gnutls-dev libgtk-3-dev git autoconf\n# Clone source\ngit clone --depth=1 https:\/\/github.com\/emacs-mirror\/emacs.git\n# Go to source\ncd emacs\/\n# Build Emacs\n.\/autogen.sh\n.\/configure --with-mailutils\nmake\n# Install (optional)\nsudo make install<\/pre><\/div>\n<\/li>\n<\/ul><\/li>\n<li>\n<p>Clone this repo to <code>HOME<\/code> or <code>~\/<\/code> path using <a href=\"https:\/\/git-scm.com\/\" rel=\"nofollow\">git<\/a> and update all the submodules.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>cd ~\ngit clone --recurse-submodules -j8 https:\/\/github.com\/MatthewZMD\/.emacs.d.git\ncd .emacs.d<\/pre><\/div>\n<\/li>\n<li>\n<p>Ensure a stable connection to Melpa Packages, then open Emacs.<\/p>\n<\/li>\n<li>\n<p>Enter <code>y<\/code> when prompted with <code>Auto-update packages now?<\/code>, wait for all packages to install.<\/p>\n<\/li>\n<li>\n<p>In your favorite browser, <code>Ctrl-f Prerequisite<\/code> through this README and follow the <strong>Prerequisite<\/strong> instructions.<\/p>\n<\/li>\n<li>\n<p>Restart Emacs.<\/p>\n<\/li>\n<\/ol><h3><a id=\"user-content-further-updates\" class=\"anchor\" aria-hidden=\"true\" href=\"#further-updates\"><\/a>Further Updates<\/h3><p>I will be updating M-EMACS from time to time, it is best to <code>git pull<\/code> once a while to stay up to date.<\/p><p>Please also execute <code>git submodule update --recursive --remote<\/code> to sync with all the submodules.<\/p><h2><a id=\"user-content-modification\" class=\"anchor\" aria-hidden=\"true\" href=\"#modification\"><\/a>Modification<\/h2><p>You have the permission to use, modify, distribute in any way you want.<\/p><p>However, what is <em>free<\/em> stays <em>free<\/em>. After all, this is <a href=\"https:\/\/github.com\/MatthewZMD\/.emacs.d\/blob\/master\/LICENSE\">GPL<\/a>.<\/p><p><strong>Remember<\/strong> you must manually sync this README with all the new changes you made by:<\/p><ol><li>Please do <strong>NOT<\/strong> edit this <code>README.md<\/code> file, edit <code>init.org<\/code> instead!<\/li>\n<li>If you add a new mode, create a new <code>&lt;file-name&gt;.el<\/code> file in <code>elisp\/<\/code> directory.<\/li>\n<li>Put <code>(require '&lt;file-name&gt;)<\/code> in <a href=\"https:\/\/github.com\/MatthewZMD\/.emacs.d\/blob\/master\/init.el\">init.el<\/a> accordingly.<\/li>\n<li>Add <code>#+INCLUDE: \"~\/.emacs.d\/elisp\/&lt;place-holder&gt;.el\" src emacs-lisp :range-begin \"&lt;start-line-wrapper-exclusive&gt;\" :range-end \"&lt;end-line-wrapper-exclusive&gt;\"<\/code> in the appropriate section in <code>init.org<\/code>.<\/li>\n<li>Enter <code>C-x C-s<\/code> to save and update <code>:lines<\/code>. (if you don't see the updated effect, run <code>M-x save-and-update-includes<\/code> manually)<\/li>\n<li>Call <code>M-x org-gfm-export-to-markdown<\/code> to update <code>README.md<\/code> automatically.<\/li>\n<\/ol><h2><a id=\"user-content-contribution\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribution\"><\/a>Contribution<\/h2><p>If you spotted a bug or you have any suggestions, please fill in an issue. If you have something to fix, feel free to create a pull request.<\/p><h2><a id=\"user-content-special-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#special-thanks\"><\/a>Special Thanks<\/h2><p>Everyone starts somewhere, and I started here.<\/p><h2><a id=\"user-content-lexical-binding\" class=\"anchor\" aria-hidden=\"true\" href=\"#lexical-binding\"><\/a>Lexical Binding<\/h2><p>Use lexical-binding. <a href=\"https:\/\/nullprogram.com\/blog\/2016\/12\/22\/\" rel=\"nofollow\">Why?<\/a><\/p><blockquote>\n<p>Until Emacs 24.1 (June 2012), Elisp only had dynamically scoped variables, a feature, mostly by accident, common to old lisp dialects. While dynamic scope has some selective uses, it\u2019s widely regarded as a mistake for local variables, and virtually no other languages have adopted it.<\/p>\n<\/blockquote><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; init.el --- -*- lexical-binding: t -*-<\/pre><\/div><h2><a id=\"user-content-early-init\" class=\"anchor\" aria-hidden=\"true\" href=\"#early-init\"><\/a>Early Init<\/h2><p>Emacs27 introduces <code>early-init.el<\/code>, which is run before <code>init.el<\/code>, before package and UI initialization happens.<\/p><h3><a id=\"user-content-compatibility-with-26\" class=\"anchor\" aria-hidden=\"true\" href=\"#compatibility-with-26\"><\/a>Compatibility With 26<\/h3><p>Ensure <code>emacs-version&gt;=26<\/code>, manually require <code>early-init<\/code> configurations if <code>emacs-version&lt;27<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(cond ((version&lt; emacs-version \"26.1\")\n       (warn \"M-EMACS requires Emacs 26.1 and above!\"))\n      ((let* ((early-init-f (expand-file-name \"early-init.el\" user-emacs-directory))\n              (early-init-do-not-edit-d (expand-file-name \"early-init-do-not-edit\/\" user-emacs-directory))\n              (early-init-do-not-edit-f (expand-file-name \"early-init.el\" early-init-do-not-edit-d)))\n         (and (version&lt; emacs-version \"27\")\n              (or (not (file-exists-p early-init-do-not-edit-f))\n                  (file-newer-than-file-p early-init-f early-init-do-not-edit-f)))\n         (make-directory early-init-do-not-edit-d t)\n         (copy-file early-init-f early-init-do-not-edit-f t t t t)\n         (add-to-list 'load-path early-init-do-not-edit-d)\n         (require 'early-init))))<\/pre><\/div><h3><a id=\"user-content-defer-garbage-collection\" class=\"anchor\" aria-hidden=\"true\" href=\"#defer-garbage-collection\"><\/a>Defer Garbage Collection<\/h3><p>Defer garbage collection further back in the startup process, according to <a href=\"https:\/\/github.com\/hlissner\/doom-emacs\/blob\/develop\/docs\/faq.org#how-does-doom-start-up-so-quickly\">hlissner<\/a>.<\/p><blockquote>\n<p>The GC eats up quite a bit of time, easily doubling startup time. The trick is to turn up the memory threshold as early as possible.<\/p>\n<\/blockquote><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq gc-cons-threshold 100000000)<\/pre><\/div><h3><a id=\"user-content-disable-package-enable-at-startup\" class=\"anchor\" aria-hidden=\"true\" href=\"#disable-package-enable-at-startup\"><\/a>Disable <code>package-enable-at-startup<\/code><\/h3><p>Package initialize occurs automatically, before <code>user-init-file<\/code> is loaded, but after <code>early-init-file<\/code>. We handle package initialization, so we must prevent Emacs from doing it early!<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq package-enable-at-startup nil)<\/pre><\/div><h3><a id=\"user-content-unset-file-name-handler-alist\" class=\"anchor\" aria-hidden=\"true\" href=\"#unset-file-name-handler-alist\"><\/a>Unset <code>file-name-handler-alist<\/code><\/h3><p>Every file opened and loaded by Emacs will run through this list to check for a proper handler for the file, but during startup, it won\u2019t need any of them.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar file-name-handler-alist-original file-name-handler-alist)\n(setq file-name-handler-alist nil)<\/pre><\/div><h3><a id=\"user-content-disable-site-run-file\" class=\"anchor\" aria-hidden=\"true\" href=\"#disable-site-run-file\"><\/a>Disable <code>site-run-file<\/code><\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq site-run-file nil)<\/pre><\/div><h3><a id=\"user-content-disable-unnecessary-interface\" class=\"anchor\" aria-hidden=\"true\" href=\"#disable-unnecessary-interface\"><\/a>Disable Unnecessary Interface<\/h3><p>It will be faster to disable them here before they've been initialized.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(menu-bar-mode -1)\n(unless (and (display-graphic-p) (eq system-type 'darwin))\n  (push '(menu-bar-lines . 0) default-frame-alist))\n(push '(tool-bar-lines . 0) default-frame-alist)\n(push '(vertical-scroll-bars) default-frame-alist)<\/pre><\/div><h2><a id=\"user-content-garbage-collection\" class=\"anchor\" aria-hidden=\"true\" href=\"#garbage-collection\"><\/a>Garbage Collection<\/h2><h3><a id=\"user-content-set-gc-cons-threshold-smaller-for-interactive-use\" class=\"anchor\" aria-hidden=\"true\" href=\"#set-gc-cons-threshold-smaller-for-interactive-use\"><\/a>Set <code>gc-cons-threshold<\/code> Smaller for Interactive Use<\/h3><p>A large <code>gc-cons-threshold<\/code> may cause freezing and stuttering during long-term interactive use.<\/p><p>If you experience freezing, decrease this amount, if you increase stuttering, increase this amount.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar better-gc-cons-threshold 67108864 ; 64mb\n  \"The default value to use for `gc-cons-threshold'.\nIf you experience freezing, decrease this.  If you experience stuttering, increase this.\")\n(add-hook 'emacs-startup-hook\n          (lambda ()\n            (setq gc-cons-threshold better-gc-cons-threshold)\n            (setq file-name-handler-alist file-name-handler-alist-original)\n            (makunbound 'file-name-handler-alist-original)))<\/pre><\/div><p>Garbage Collect when Emacs is out of focus and avoid garbage collection when using minibuffer.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'emacs-startup-hook\n          (lambda ()\n            (if (boundp 'after-focus-change-function)\n                (add-function :after after-focus-change-function\n                              (lambda ()\n                                (unless (frame-focus-state)\n                                  (garbage-collect))))\n              (add-hook 'after-focus-change-function 'garbage-collect))\n            (defun gc-minibuffer-setup-hook ()\n              (setq gc-cons-threshold (* better-gc-cons-threshold 2)))\n            (defun gc-minibuffer-exit-hook ()\n              (garbage-collect)\n              (setq gc-cons-threshold better-gc-cons-threshold))\n            (add-hook 'minibuffer-setup-hook #'gc-minibuffer-setup-hook)\n            (add-hook 'minibuffer-exit-hook #'gc-minibuffer-exit-hook)))<\/pre><\/div><h2><a id=\"user-content-load-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#load-path\"><\/a>Load Path<\/h2><p>Since all the configuration files are stored in <code>elisp\/<\/code> folder, they need to be added to <code>load-path<\/code> now.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun update-to-load-path (folder)\n  \"Update FOLDER and its subdirectories to `load-path'.\"\n  (let ((base folder))\n    (unless (member base load-path)\n      (add-to-list 'load-path base))\n    (dolist (f (directory-files base))\n      (let ((name (concat base \"\/\" f)))\n        (when (and (file-directory-p name)\n                   (not (equal f \"..\"))\n                   (not (equal f \".\")))\n          (unless (member base load-path)\n            (add-to-list 'load-path name)))))))\n(update-to-load-path (expand-file-name \"elisp\" user-emacs-directory))<\/pre><\/div><h2><a id=\"user-content-define-constants\" class=\"anchor\" aria-hidden=\"true\" href=\"#define-constants\"><\/a>Define Constants<\/h2><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defconst *sys\/gui*\n  (display-graphic-p)\n  \"Are we running on a GUI Emacs?\")\n(defconst *sys\/win32*\n  (eq system-type 'windows-nt)\n  \"Are we running on a WinTel system?\")\n(defconst *sys\/linux*\n  (eq system-type 'gnu\/linux)\n  \"Are we running on a GNU\/Linux system?\")\n(defconst *sys\/mac*\n  (eq system-type 'darwin)\n  \"Are we running on a Mac system?\")\n(defconst *sys\/root*\n  (string-equal \"root\" (getenv \"USER\"))\n  \"Are you a ROOT user?\")\n(defconst *rg*\n  (executable-find \"rg\")\n  \"Do we have ripgrep?\")\n(defconst *python*\n  (executable-find \"python\")\n  \"Do we have python?\")\n(defconst *python3*\n  (executable-find \"python3\")\n  \"Do we have python3?\")\n(defconst *tr*\n  (executable-find \"tr\")\n  \"Do we have tr?\")\n(defconst *mvn*\n  (executable-find \"mvn\")\n  \"Do we have Maven?\")\n(defconst *clangd*\n  (or (executable-find \"clangd\")  ;; usually\n      (executable-find \"\/usr\/local\/opt\/llvm\/bin\/clangd\"))  ;; macOS\n  \"Do we have clangd?\")\n(defconst *gcc*\n  (executable-find \"gcc\")\n  \"Do we have gcc?\")\n(defconst *git*\n  (executable-find \"git\")\n  \"Do we have git?\")\n(defconst *pdflatex*\n  (executable-find \"pdflatex\")\n  \"Do we have pdflatex?\")\n(defconst *eaf-env*\n  (and *sys\/linux* *sys\/gui* *python3*\n       (executable-find \"pip\")\n       (not (equal (shell-command-to-string \"pip freeze | grep '^PyQt\\\\|PyQtWebEngine'\") \"\")))\n  \"Check basic requirements for EAF to run.\")<\/pre><\/div><p>Some packages are disabled with the <code>:disabled<\/code> tag, because I don't use them very often. They might not work.<\/p><h2><a id=\"user-content-melpa-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#melpa-packages\"><\/a>Melpa Packages<\/h2><p>Configure package archives, where to install online packages and add them to <code>load-path<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Select the folder to store packages\n;; Comment \/ Uncomment to use desired sites\n(setq package-user-dir (expand-file-name \"elpa\" user-emacs-directory)\n      package-archives\n      '((\"gnu\"   . \"https:\/\/elpa.gnu.org\/packages\/\")\n        (\"melpa\" . \"https:\/\/melpa.org\/packages\/\")\n        (\"cselpa\" . \"https:\/\/elpa.thecybershadow.net\/packages\/\")\n        ;; (\"melpa-cn\" . \"http:\/\/mirrors.cloud.tencent.com\/elpa\/melpa\/\")\n        ;; (\"gnu-cn\"   . \"http:\/\/mirrors.cloud.tencent.com\/elpa\/gnu\/\")\n        ))<\/pre><\/div><h2><a id=\"user-content-non-melpa-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#non-melpa-packages\"><\/a>Non-Melpa Packages<\/h2><p>Add packages contained in <code>site-elisp\/<\/code> to <code>load-path<\/code> too.<\/p><h3><a id=\"user-content-add-packages-manually-from-git\" class=\"anchor\" aria-hidden=\"true\" href=\"#add-packages-manually-from-git\"><\/a>Add Packages Manually from Git<\/h3><div class=\"highlight highlight-source-shell\"><pre>cd site-elisp\/\ngit submodule add https:\/\/github.com\/foo\/bar.git<\/pre><\/div><p>Verify <code>\/.gitmodules<\/code> file that the newly added package exist.<\/p><h3><a id=\"user-content-update-manually-added-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#update-manually-added-packages\"><\/a>Update Manually Added Packages<\/h3><div class=\"highlight highlight-source-shell\"><pre>git submodule init\ngit submodule update<\/pre><\/div><h2><a id=\"user-content-configure-package-manager\" class=\"anchor\" aria-hidden=\"true\" href=\"#configure-package-manager\"><\/a>Configure Package Manager<\/h2><div class=\"highlight highlight-source-emacs-lisp\"><pre>(unless (bound-and-true-p package--initialized)\n  (setq package-enable-at-startup nil)          ; To prevent initializing twice\n  (package-initialize))\n;; set use-package-verbose to t for interpreted .emacs,\n;; and to nil for byte-compiled .emacs.elc.\n(eval-and-compile\n  (setq use-package-verbose (not (bound-and-true-p byte-compile-current-file))))<\/pre><\/div><h2><a id=\"user-content-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-package\"><\/a>Use Package<\/h2><p>My Emacs configuration is almost entirely dependant on <a href=\"https:\/\/github.com\/jwiegley\/use-package\">use-package<\/a>.<\/p><blockquote>\n<p>The <code>use-package<\/code> macro allows you to isolate package configuration in your .emacs file in a way that is both performance-oriented and, well, tidy. I created it because I have over 80 packages that I use in Emacs, and things were getting difficult to manage. Yet with this utility my total load time is around 2 seconds, with no loss of functionality!<\/p>\n<\/blockquote><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Install use-package if not installed\n(unless (package-installed-p 'use-package)\n  (package-refresh-contents)\n  (package-install 'use-package))\n(eval-and-compile\n  (setq use-package-always-ensure t)\n  (setq use-package-expand-minimally t)\n  (setq use-package-compute-statistics t)\n  (setq use-package-enable-imenu-support t))\n(eval-when-compile\n  (require 'use-package)\n  (require 'bind-key))<\/pre><\/div><h2><a id=\"user-content-auto-package-update\" class=\"anchor\" aria-hidden=\"true\" href=\"#auto-package-update\"><\/a>Auto Package Update<\/h2><p><a href=\"https:\/\/github.com\/rranelli\/auto-package-update.el\">Auto package update<\/a> automatically updates installed packages if at least <code>auto-package-update-interval<\/code> days have passed since the last update.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package auto-package-update\n  :if (not (daemonp))\n  :custom\n  (auto-package-update-interval 7) ;; in days\n  (auto-package-update-prompt-before-update t)\n  (auto-package-update-delete-old-versions t)\n  (auto-package-update-hide-results t)\n  :config\n  (auto-package-update-maybe))<\/pre><\/div><h2><a id=\"user-content-diminish\" class=\"anchor\" aria-hidden=\"true\" href=\"#diminish\"><\/a>Diminish<\/h2><p><a href=\"https:\/\/github.com\/emacsmirror\/diminish\">Diminish<\/a>, a feature that removes certain minor modes from mode-line.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package diminish)<\/pre><\/div><h2><a id=\"user-content-user-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-information\"><\/a>User Information<\/h2><p><strong>Prerequisite<\/strong>: Please update this file your personal info.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq user-full-name \"Mingde (Matthew) Zeng\")\n(setq user-mail-address \"matthewzmd@gmail.com\")<\/pre><\/div><h2><a id=\"user-content-bindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#bindings\"><\/a>Bindings<\/h2><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Unbind unneeded keys\n(global-set-key (kbd \"C-z\") nil)\n(global-set-key (kbd \"M-z\") nil)\n(global-set-key (kbd \"C-x C-z\") nil)\n(global-set-key (kbd \"M-\/\") nil)\n;; Use iBuffer instead of Buffer List\n(global-set-key (kbd \"C-x C-b\") #'ibuffer)\n;; Truncate lines\n(global-set-key (kbd \"C-x C-l\") #'toggle-truncate-lines)\n;; Adjust font size like web browsers\n(global-set-key (kbd \"C-+\") #'text-scale-increase)\n(global-set-key (kbd \"C--\") #'text-scale-decrease)\n;; Move up\/down paragraph\n(global-set-key (kbd \"M-n\") #'forward-paragraph)\n(global-set-key (kbd \"M-p\") #'backward-paragraph)<\/pre><\/div><h2><a id=\"user-content-avy\" class=\"anchor\" aria-hidden=\"true\" href=\"#avy\"><\/a>Avy<\/h2><p><a href=\"https:\/\/github.com\/abo-abo\/avy\">Avy<\/a>, a nice way to move around text.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package avy\n  :defer t\n  :bind\n  ((\"C-z c\" . avy-goto-char-timer)\n   (\"C-z l\" . avy-goto-line))\n  :custom\n  (avy-timeout-seconds 0.3)\n  (avy-style 'pre)\n  :custom-face\n  (avy-lead-face ((t (:background \"#51afef\" :foreground \"#870000\" :weight bold)))));<\/pre><\/div><h2><a id=\"user-content-crux\" class=\"anchor\" aria-hidden=\"true\" href=\"#crux\"><\/a>Crux<\/h2><p><a href=\"https:\/\/github.com\/bbatsov\/crux\">Crux<\/a>, a Collection of Ridiculously Useful eXtensions for Emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package crux\n  :bind\n  ((\"C-a\" . crux-move-beginning-of-line)\n   (\"C-x 4 t\" . crux-transpose-windows)\n   (\"C-x K\" . crux-kill-other-buffers)\n   (\"C-k\" . crux-smart-kill-line))\n  :config\n  (crux-with-region-or-buffer indent-region)\n  (crux-with-region-or-buffer untabify)\n  (crux-with-region-or-point-to-eol kill-ring-save)\n  (defalias 'rename-file-and-buffer #'crux-rename-file-and-buffer))<\/pre><\/div><h2><a id=\"user-content-ivy-amx-counsel-swiper\" class=\"anchor\" aria-hidden=\"true\" href=\"#ivy-amx-counsel-swiper\"><\/a>Ivy, Amx, Counsel, Swiper<\/h2><p><a href=\"https:\/\/github.com\/abo-abo\/swiper\">Ivy<\/a>, a generic completion mechanism for Emacs. It utilizes <a href=\"https:\/\/github.com\/DarwinAwardWinner\/amx\">Amx<\/a>, <a href=\"https:\/\/github.com\/abo-abo\/swiper\">Counsel<\/a> and <a href=\"https:\/\/github.com\/abo-abo\/swiper\">Swiper<\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package ivy\n  :diminish\n  :init\n  (use-package amx :defer t)\n  (use-package counsel :diminish :config (counsel-mode 1))\n  (use-package swiper :defer t)\n  (ivy-mode 1)\n  :bind\n  ((\"C-s\" . swiper-isearch)\n   (\"C-z s\" . counsel-rg)\n   (\"C-z b\" . counsel-buffer-or-recentf)\n   (\"C-z C-b\" . counsel-ibuffer)\n   (:map ivy-minibuffer-map\n         (\"C-r\" . ivy-previous-line-or-history)\n         (\"M-RET\" . ivy-immediate-done))\n   (:map counsel-find-file-map\n         (\"C-~\" . counsel-goto-local-home)))\n  :custom\n  (ivy-use-virtual-buffers t)\n  (ivy-height 10)\n  (ivy-on-del-error-function nil)\n  (ivy-magic-slash-non-match-action 'ivy-magic-slash-non-match-create)\n  (ivy-count-format \"\u3010%d\/%d\u3011\")\n  (ivy-wrap t)\n  :config\n  (defun counsel-goto-local-home ()\n      \"Go to the $HOME of the local machine.\"\n      (interactive)\n    (ivy--cd \"~\/\")))<\/pre><\/div><h2><a id=\"user-content-color-ripgrep\" class=\"anchor\" aria-hidden=\"true\" href=\"#color-ripgrep\"><\/a>Color Ripgrep<\/h2><p><a href=\"https:\/\/github.com\/manateelazycat\/color-rg\">Color rg<\/a>, a search and refactoring tool based on <em>ripgrep<\/em> that is used to search text.<\/p><p><strong>Prerequisite<\/strong>: Please install <a href=\"https:\/\/github.com\/BurntSushi\/ripgrep#installation\">ripgrep<\/a> and ensure <code>rg<\/code> is in the <code>PATH<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package color-rg\n  :load-path (lambda () (expand-file-name \"site-elisp\/color-rg\" user-emacs-directory))\n  :if *rg*\n  :bind (\"C-M-s\" . color-rg-search-input))<\/pre><\/div><h2><a id=\"user-content-snails\" class=\"anchor\" aria-hidden=\"true\" href=\"#snails\"><\/a>Snails<\/h2><p><a href=\"https:\/\/github.com\/manateelazycat\/snails\">Snails<\/a>, a fuzzy search framework. It utilizes <a href=\"https:\/\/github.com\/purcell\/exec-path-from-shell\">exec-path-from-shell<\/a> if you are using Mac.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package snails\n  :load-path (lambda () (expand-file-name \"site-elisp\/snails\/\" user-emacs-directory))\n  :if *sys\/gui*\n  :custom-face\n  (snails-content-buffer-face ((t (:background \"#111\" :height 110))))\n  (snails-input-buffer-face ((t (:background \"#222\" :foreground \"gold\" :height 110))))\n  (snails-header-line-face ((t (:inherit font-lock-function-name-face :underline t :height 1.1))))\n  :config\n  (use-package exec-path-from-shell\n    :if (featurep 'cocoa) :defer t)\n  ;; Functions for specific backends\n  (defun snails-current-project ()\n    (interactive)\n    (snails '(snails-backend-projectile snails-backend-rg snails-backend-fd)))\n  (defun snails-active-recent-buffers ()\n    (interactive)\n    (snails '(snails-backend-buffer snails-backend-recentf)))\n  (defun snails-everywhere ()\n    (interactive)\n    (snails '(snails-backend-everything snails-backend-mdfind)))\n  :bind\n  ((\"M-s s\" . snails)\n   (\"M-s g\" . snails-current-project)\n   (\"M-s b\" . snails-active-recent-buffers)\n   (\"M-s e\" . snails-everywhere)))<\/pre><\/div><h2><a id=\"user-content-files-directories\" class=\"anchor\" aria-hidden=\"true\" href=\"#files-directories\"><\/a>Files Directories<\/h2><h3><a id=\"user-content-dired\" class=\"anchor\" aria-hidden=\"true\" href=\"#dired\"><\/a>Dired<\/h3><p>Dired, the directory editor.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package dired\n  :ensure nil\n  :bind\n  ((\"C-x C-j\" . dired-jump)\n   (\"C-x j\" . dired-jump-other-window))\n  :custom\n  ;; Always delete and copy recursively\n  (dired-recursive-deletes 'always)\n  (dired-recursive-copies 'always)\n  ;; Auto refresh Dired, but be quiet about it\n  (global-auto-revert-non-file-buffers t)\n  (auto-revert-verbose nil)\n  ;; Quickly copy\/move file in Dired\n  (dired-dwim-target t)\n  ;; Move files to trash when deleting\n  (delete-by-moving-to-trash t)\n  ;; Load the newest version of a file\n  (load-prefer-newer t)\n  ;; Detect external file changes and auto refresh file\n  (auto-revert-use-notify nil)\n  (auto-revert-interval 3) ; Auto revert every 3 sec\n  :config\n  ;; Enable global auto-revert\n  (global-auto-revert-mode t)\n  ;; Reuse same dired buffer, to prevent numerous buffers while navigating in dired\n  (put 'dired-find-alternate-file 'disabled nil)\n  :hook\n  (dired-mode . (lambda ()\n                  (local-set-key (kbd \"&lt;mouse-2&gt;\") #'dired-find-alternate-file)\n                  (local-set-key (kbd \"RET\") #'dired-find-alternate-file)\n                  (local-set-key (kbd \"^\")\n                                 (lambda () (interactive) (find-alternate-file \"..\"))))))<\/pre><\/div><h3><a id=\"user-content-disk-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#disk-usage\"><\/a>Disk Usage<\/h3><p><a href=\"https:\/\/gitlab.com\/ambrevar\/emacs-disk-usage\" rel=\"nofollow\">Disk Usage<\/a>, a file system analyzer that offers a tabulated view of file listings sorted by size.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package disk-usage\n  :commands (disk-usage))<\/pre><\/div><h3><a id=\"user-content-super-save\" class=\"anchor\" aria-hidden=\"true\" href=\"#super-save\"><\/a>Super Save<\/h3><p><a href=\"https:\/\/github.com\/bbatsov\/super-save\">Super Save<\/a>, enables save when switching between buffers, an Emacs frame losing focus, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package super-save\n  :diminish\n  :custom\n  (super-save-auto-save-when-idle t)\n  (auto-save-default nil)\n  (make-backup-files nil)\n  :config\n  (super-save-mode 1))<\/pre><\/div><h3><a id=\"user-content-save-all-buffers\" class=\"anchor\" aria-hidden=\"true\" href=\"#save-all-buffers\"><\/a>Save All Buffers<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun save-all-buffers ()\n  \"Instead of `save-buffer', save all opened buffers by calling `save-some-buffers' with ARG t.\"\n  (interactive)\n  (save-some-buffers t))\n(global-set-key (kbd \"C-x C-s\") nil)\n(global-set-key (kbd \"C-x C-s\") #'save-all-buffers)<\/pre><\/div><h2><a id=\"user-content-winner\" class=\"anchor\" aria-hidden=\"true\" href=\"#winner\"><\/a>Winner<\/h2><p>Winner, a mode to restore previous window layouts.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package winner\n  :ensure nil\n  :custom\n  (winner-boring-buffers\n   '(\"*Completions*\"\n     \"*Compile-Log*\"\n     \"*inferior-lisp*\"\n     \"*Fuzzy Completions*\"\n     \"*Apropos*\"\n     \"*Help*\"\n     \"*cvs*\"\n     \"*Buffer List*\"\n     \"*Ibuffer*\"\n     \"*esh command on file*\"))\n  :config\n  (winner-mode 1))<\/pre><\/div><h2><a id=\"user-content-which-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#which-key\"><\/a>Which Key<\/h2><p><a href=\"https:\/\/github.com\/justbur\/emacs-which-key\">Which Key<\/a>, a feature that displays the key bindings following the incomplete command.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package which-key\n  :diminish\n  :custom\n  (which-key-separator \" \")\n  (which-key-prefix-prefix \"+\")\n  :config\n  (which-key-mode))<\/pre><\/div><h2><a id=\"user-content-popup-kill-ring\" class=\"anchor\" aria-hidden=\"true\" href=\"#popup-kill-ring\"><\/a>Popup Kill Ring<\/h2><p><a href=\"https:\/\/github.com\/waymondo\/popup-kill-ring\">Popup Kill Ring<\/a>, a feature that provides the ability to browse Emacs kill ring in autocomplete style popup menu.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package popup-kill-ring\n  :bind (\"M-y\" . popup-kill-ring))<\/pre><\/div><h2><a id=\"user-content-undo-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#undo-tree\"><\/a>Undo Tree<\/h2><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/UndoTree\" rel=\"nofollow\">Undo tree<\/a>, a feature that provides a visualization of the undos in a file.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package undo-tree\n  :defer t\n  :diminish undo-tree-mode\n  :init (global-undo-tree-mode))<\/pre><\/div><h2><a id=\"user-content-discover-my-major\" class=\"anchor\" aria-hidden=\"true\" href=\"#discover-my-major\"><\/a>Discover My Major<\/h2><p><a href=\"https:\/\/github.com\/jguenther\/discover-my-major\">Discover my major<\/a>, a feature that discovers key bindings and their meaning for the current Emacs major mode.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package discover-my-major\n  :bind (\"C-h C-m\" . discover-my-major))<\/pre><\/div><h2><a id=\"user-content-ace-window\" class=\"anchor\" aria-hidden=\"true\" href=\"#ace-window\"><\/a>Ace Window<\/h2><p><a href=\"https:\/\/github.com\/abo-abo\/ace-window\">Ace Window<\/a>, a package for selecting windows to switch to.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package ace-window\n  :bind (\"C-x C-o\" . ace-window))<\/pre><\/div><h2><a id=\"user-content-shells\" class=\"anchor\" aria-hidden=\"true\" href=\"#shells\"><\/a>Shells<\/h2><h3><a id=\"user-content-aweshell\" class=\"anchor\" aria-hidden=\"true\" href=\"#aweshell\"><\/a>Aweshell<\/h3><p><a href=\"https:\/\/github.com\/manateelazycat\/aweshell\">Aweshell<\/a>, shell extension base on eshell with better features.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package aweshell\n  :load-path (lambda () (expand-file-name \"site-elisp\/aweshell\" user-emacs-directory))\n  :commands (aweshell-new aweshell-dedicated-open)\n  :bind\n  ((\"M-#\" . aweshell-dedicated-open)\n   (:map eshell-mode-map (\"M-#\" . aweshell-dedicated-close))))<\/pre><\/div><h3><a id=\"user-content-shell-here\" class=\"anchor\" aria-hidden=\"true\" href=\"#shell-here\"><\/a>Shell Here<\/h3><p><a href=\"https:\/\/github.com\/ieure\/shell-here\">Shell Here<\/a>, a tool that opens a shell buffer in (or relative to) <code>default-directory<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package shell-here\n  :bind (\"M-~\" . shell-here)\n  :config\n  (when *sys\/linux*\n    (setq explicit-shell-file-name \"\/bin\/bash\")))<\/pre><\/div><h3><a id=\"user-content-multiterm\" class=\"anchor\" aria-hidden=\"true\" href=\"#multiterm\"><\/a>MultiTerm<\/h3><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/MultiTerm\" rel=\"nofollow\">MultiTerm<\/a>, a mode based on term.el, for managing multiple terminal buffers in Emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package multi-term\n  :load-path (lambda () (expand-file-name \"site-elisp\/multi-term\" user-emacs-directory))\n  :commands (multi-term)\n  :bind\n  ((\"M-$\" . multi-term)\n   (:map dired-mode-map (\"M-$\" . multi-term)))\n  :custom\n  (multi-term-program (executable-find \"bash\")))<\/pre><\/div><h3><a id=\"user-content-term-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#term-keys\"><\/a>Term Keys<\/h3><p><a href=\"https:\/\/github.com\/CyberShadow\/term-keys\">Term Keys<\/a>, a lossless keyboard input for Emacs in terminal emulators.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package term-keys\n  :if (not *sys\/gui*)\n  :config (term-keys-mode t))<\/pre><\/div><h2><a id=\"user-content-sudo-edit\" class=\"anchor\" aria-hidden=\"true\" href=\"#sudo-edit\"><\/a>Sudo Edit<\/h2><p><a href=\"https:\/\/github.com\/nflath\/sudo-edit\">Sudo Edit<\/a>, an utility for opening files with <code>sudo<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package sudo-edit\n  :commands (sudo-edit))<\/pre><\/div><h2><a id=\"user-content-configs\" class=\"anchor\" aria-hidden=\"true\" href=\"#configs\"><\/a>Configs<\/h2><p>Some essential configs that make my life a lot easier.<\/p><h3><a id=\"user-content-utf-8-coding-system\" class=\"anchor\" aria-hidden=\"true\" href=\"#utf-8-coding-system\"><\/a>UTF-8 Coding System<\/h3><p>Use UTF-8 as much as possible with unix line endings.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(unless *sys\/win32*\n  (set-selection-coding-system 'utf-8)\n  (prefer-coding-system 'utf-8)\n  (set-language-environment \"UTF-8\")\n  (set-default-coding-systems 'utf-8)\n  (set-terminal-coding-system 'utf-8)\n  (set-keyboard-coding-system 'utf-8)\n  (setq locale-coding-system 'utf-8))\n;; Treat clipboard input as UTF-8 string first; compound text next, etc.\n(when *sys\/gui*\n  (setq x-select-request-type '(UTF8_STRING COMPOUND_TEXT TEXT STRING)))<\/pre><\/div><h3><a id=\"user-content-optimize-editing-experience\" class=\"anchor\" aria-hidden=\"true\" href=\"#optimize-editing-experience\"><\/a>Optimize Editing Experience<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Remove useless whitespace before saving a file\n(defun delete-trailing-whitespace-except-current-line ()\n  \"Sometimes `delete-trailing-whitespace' becomes very annoying.\nIt deletes trailing whitespace current line.  Therefore I use this alternative.\"\n  (interactive)\n  (let ((begin (line-beginning-position))\n        (end (line-end-position)))\n    (save-excursion\n      (when (&lt; (point-min) begin)\n        (save-restriction\n          (narrow-to-region (point-min) (1- begin))\n          (delete-trailing-whitespace)\n          (widen)))\n      (when (&gt; (point-max) end)\n        (save-restriction\n          (narrow-to-region (+ end 2) (point-max))\n          (delete-trailing-whitespace)\n          (widen))))))\n(add-hook 'before-save-hook #'delete-trailing-whitespace-except-current-line)\n;; Replace selection on insert\n(delete-selection-mode 1)\n;; Map Alt key to Meta\n(setq x-alt-keysym 'meta)<\/pre><\/div><h3><a id=\"user-content-history\" class=\"anchor\" aria-hidden=\"true\" href=\"#history\"><\/a>History<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package recentf\n  :ensure nil\n  :hook (after-init . recentf-mode)\n  :custom\n  (recentf-auto-cleanup \"05:00am\")\n  (recentf-max-saved-items 200)\n  (recentf-exclude '((expand-file-name package-user-dir)\n                     \".cache\"\n                     \".cask\"\n                     \".elfeed\"\n                     \"bookmarks\"\n                     \"cache\"\n                     \"ido.*\"\n                     \"persp-confs\"\n                     \"recentf\"\n                     \"undo-tree-hist\"\n                     \"url\"\n                     \"COMMIT_EDITMSG\\\\'\")))\n;; When buffer is closed, saves the cursor location\n(save-place-mode 1)\n;; Set history-length longer\n(setq-default history-length 500)<\/pre><\/div><h3><a id=\"user-content-small-configs\" class=\"anchor\" aria-hidden=\"true\" href=\"#small-configs\"><\/a>Small Configs<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Turn Off Cursor Alarms\n(setq ring-bell-function 'ignore)\n;; Show Keystrokes in Progress Instantly\n(setq echo-keystrokes 0.1)\n;; Don't Lock Files\n(setq-default create-lockfiles nil)\n;; Better Compilation\n(setq-default compilation-always-kill t) ; kill compilation process before starting another\n(setq-default compilation-ask-about-save nil) ; save all buffers on `compile'\n(setq-default compilation-scroll-output t)\n;; ad-handle-definition warnings are generated when functions are redefined with `defadvice',\n;; they are not helpful.\n(setq ad-redefinition-action 'accept)\n;; Move Custom-Set-Variables to Different File\n(setq custom-file (concat user-emacs-directory \"custom-set-variables.el\"))\n(load custom-file 'noerror)\n;; So Long mitigates slowness due to extremely long lines.\n;; Currently available in Emacs master branch *only*!\n(when (fboundp 'global-so-long-mode)\n  (global-so-long-mode))\n;; Add a newline automatically at the end of the file upon save.\n(setq require-final-newline t)\n;; Default .args, .in, .out files to text-mode\n(add-to-list 'auto-mode-alist '(\"\\\\.in\\\\'\" . text-mode))\n(add-to-list 'auto-mode-alist '(\"\\\\.out\\\\'\" . text-mode))\n(add-to-list 'auto-mode-alist '(\"\\\\.args\\\\'\" . text-mode))<\/pre><\/div><h2><a id=\"user-content-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#functions\"><\/a>Functions<\/h2><p>Important functions.<\/p><h3><a id=\"user-content-resize-window-width--height-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#resize-window-width--height-functions\"><\/a>Resize Window Width \/ Height Functions<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Resizes the window width based on the input\n(defun resize-window-width (w)\n  \"Resizes the window width based on W.\"\n  (interactive (list (if (&gt; (count-windows) 1)\n                         (read-number \"Set the current window width in [1~9]x10%: \")\n                       (error \"You need more than 1 window to execute this function!\"))))\n  (message \"%s\" w)\n  (window-resize nil (- (truncate (* (\/ w 10.0) (frame-width))) (window-total-width)) t))\n;; Resizes the window height based on the input\n(defun resize-window-height (h)\n  \"Resizes the window height based on H.\"\n  (interactive (list (if (&gt; (count-windows) 1)\n                         (read-number \"Set the current window height in [1~9]x10%: \")\n                       (error \"You need more than 1 window to execute this function!\"))))\n  (message \"%s\" h)\n  (window-resize nil (- (truncate (* (\/ h 10.0) (frame-height))) (window-total-height)) nil))\n;; Setup shorcuts for window resize width and height\n(global-set-key (kbd \"C-z w\") #'resize-window-width)\n(global-set-key (kbd \"C-z h\") #'resize-window-height)\n(defun resize-window (width delta)\n  \"Resize the current window's size.  If WIDTH is non-nil, resize width by some DELTA.\"\n  (if (&gt; (count-windows) 1)\n      (window-resize nil delta width)\n    (error \"You need more than 1 window to execute this function!\")))\n;; Setup shorcuts for window resize width and height\n(global-set-key (kbd \"M-W =\") (lambda () (interactive) (resize-window t 5)))\n(global-set-key (kbd \"M-W M-+\") (lambda () (interactive) (resize-window t 5)))\n(global-set-key (kbd \"M-W -\") (lambda () (interactive) (resize-window t -5)))\n(global-set-key (kbd \"M-W M-_\") (lambda () (interactive) (resize-window t -5)))\n(global-set-key (kbd \"M-H =\") (lambda () (interactive) (resize-window nil 5)))\n(global-set-key (kbd \"M-H M-+\") (lambda () (interactive) (resize-window nil 5)))\n(global-set-key (kbd \"M-H -\") (lambda () (interactive) (resize-window nil -5)))\n(global-set-key (kbd \"M-H M-_\") (lambda () (interactive) (resize-window nil -5)))<\/pre><\/div><h3><a id=\"user-content-edit-this-configuration-file-shortcut\" class=\"anchor\" aria-hidden=\"true\" href=\"#edit-this-configuration-file-shortcut\"><\/a>Edit This Configuration File Shortcut<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun edit-configs ()\n  \"Opens the README.org file.\"\n  (interactive)\n  (find-file \"~\/.emacs.d\/init.org\"))\n(global-set-key (kbd \"C-z e\") #'edit-configs)<\/pre><\/div><h3><a id=\"user-content-update-org-mode-include-automatically\" class=\"anchor\" aria-hidden=\"true\" href=\"#update-org-mode-include-automatically\"><\/a>Update Org Mode Include Automatically<\/h3><p>Update Org Mode INCLUDE Statements Automatically from <a href=\"http:\/\/endlessparentheses.com\/updating-org-mode-include-statements-on-the-fly.html\" rel=\"nofollow\">Artur Malabarba<\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun save-and-update-includes ()\n  \"Update the line numbers of #+INCLUDE:s in current buffer.\nOnly looks at INCLUDEs that have either :range-begin or :range-end.\nThis function does nothing if not in `org-mode', so you can safely\nadd it to `before-save-hook'.\"\n  (interactive)\n  (when (derived-mode-p 'org-mode)\n    (save-excursion\n      (goto-char (point-min))\n      (while (search-forward-regexp\n              \"^\\\\s-*#\\\\+INCLUDE: *\\\"\\\\([^\\\"]+\\\\)\\\".*:range-\\\\(begin\\\\|end\\\\)\"\n              nil 'noerror)\n        (let* ((file (expand-file-name (match-string-no-properties 1)))\n               lines begin end)\n          (forward-line 0)\n          (when (looking-at \"^.*:range-begin *\\\"\\\\([^\\\"]+\\\\)\\\"\")\n            (setq begin (match-string-no-properties 1)))\n          (when (looking-at \"^.*:range-end *\\\"\\\\([^\\\"]+\\\\)\\\"\")\n            (setq end (match-string-no-properties 1)))\n          (setq lines (decide-line-range file begin end))\n          (when lines\n            (if (looking-at \".*:lines *\\\"\\\\([-0-9]+\\\\)\\\"\")\n                (replace-match lines :fixedcase :literal nil 1)\n              (goto-char (line-end-position))\n              (insert \" :lines \\\"\" lines \"\\\"\"))))))))\n(add-hook 'before-save-hook #'save-and-update-includes)\n(defun decide-line-range (file begin end)\n  \"Visit FILE and decide which lines to include.\nBEGIN and END are regexps which define the line range to use.\"\n  (let (l r)\n    (save-match-data\n      (with-temp-buffer\n        (insert-file-contents file)\n        (goto-char (point-min))\n        (if (null begin)\n            (setq l \"\")\n          (search-forward-regexp begin)\n          (setq l (line-number-at-pos (match-beginning 0))))\n        (if (null end)\n            (setq r \"\")\n          (search-forward-regexp end)\n          (setq r (1+ (line-number-at-pos (match-end 0)))))\n        (format \"%s-%s\" (+ l 1) (- r 1)))))) ;; Exclude wrapper<\/pre><\/div><h3><a id=\"user-content-minibuffer-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#minibuffer-functions\"><\/a>MiniBuffer Functions<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun abort-minibuffer-using-mouse ()\n  \"Abort the minibuffer when using the mouse.\"\n  (when (and (&gt;= (recursion-depth) 1) (active-minibuffer-window))\n    (abort-recursive-edit)))\n(add-hook 'mouse-leave-buffer-hook 'abort-minibuffer-using-mouse)\n;; keep the point out of the minibuffer\n(setq-default minibuffer-prompt-properties '(read-only t point-entered minibuffer-avoid-prompt face minibuffer-prompt))<\/pre><\/div><h3><a id=\"user-content-display-line-overlay\" class=\"anchor\" aria-hidden=\"true\" href=\"#display-line-overlay\"><\/a>Display Line Overlay<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun display-line-overlay+ (pos str &amp;optional face)\n  \"Display line at POS as STR with FACE.\nFACE defaults to inheriting from default and highlight.\"\n  (let ((ol (save-excursion\n              (goto-char pos)\n              (make-overlay (line-beginning-position)\n                            (line-end-position)))))\n    (overlay-put ol 'display str)\n    (overlay-put ol 'face\n                 (or face '(:background null :inherit highlight)))\n    ol))<\/pre><\/div><h3><a id=\"user-content-read-lines-from-file\" class=\"anchor\" aria-hidden=\"true\" href=\"#read-lines-from-file\"><\/a>Read Lines From File<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun read-lines (filePath)\n  \"Return a list of lines of a file at FILEPATH.\"\n  (with-temp-buffer (insert-file-contents filePath)\n                    (split-string (buffer-string) \"\\n\" t)))<\/pre><\/div><h3><a id=\"user-content-where-am-i\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-am-i\"><\/a>Where Am I<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun where-am-i ()\n  \"An interactive function showing function `buffer-file-name' or `buffer-name'.\"\n  (interactive)\n  (message (kill-new (if (buffer-file-name) (buffer-file-name) (buffer-name)))))<\/pre><\/div><h2><a id=\"user-content-doom-themes\" class=\"anchor\" aria-hidden=\"true\" href=\"#doom-themes\"><\/a>Doom Themes<\/h2><p><a href=\"https:\/\/github.com\/hlissner\/emacs-doom-themes\">Doom Themes<\/a>, an UI plugin and pack of themes.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package doom-themes\n  :custom-face\n  (cursor ((t (:background \"BlanchedAlmond\"))))\n  :config\n  ;; flashing mode-line on errors\n  (doom-themes-visual-bell-config)\n  ;; Corrects (and improves) org-mode's native fontification.\n  (doom-themes-org-config)\n  (load-theme 'doom-one t))<\/pre><\/div><h2><a id=\"user-content-doom-modeline\" class=\"anchor\" aria-hidden=\"true\" href=\"#doom-modeline\"><\/a>Doom Modeline<\/h2><p><a href=\"https:\/\/github.com\/seagle0128\/doom-modeline\">Doom Modeline<\/a>, a modeline from DOOM Emacs, but more powerful and faster.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package doom-modeline\n  :hook (after-init . doom-modeline-mode)\n  :custom\n  ;; Don't compact font caches during GC. Windows Laggy Issue\n  (inhibit-compacting-font-caches t)\n  (doom-modeline-minor-modes t)\n  (doom-modeline-icon t)\n  (doom-modeline-major-mode-color-icon t)\n  (doom-modeline-height 15))<\/pre><\/div><h2><a id=\"user-content-dashboard\" class=\"anchor\" aria-hidden=\"true\" href=\"#dashboard\"><\/a>Dashboard<\/h2><h3><a id=\"user-content-dashboard-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#dashboard-1\"><\/a>Dashboard<\/h3><p><a href=\"https:\/\/github.com\/rakanalh\/emacs-dashboard\">Dashboard<\/a>, an extensible Emacs startup screen.<\/p><p>Use either <code>KEC_Dark_BK.png<\/code> or <code>KEC_Light_BK.png<\/code> depends on the backgrond theme.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package dashboard\n  :demand\n  :diminish (dashboard-mode page-break-lines-mode)\n  :bind (\"C-z d\" . open-dashboard)\n  :custom\n  (dashboard-banner-logo-title \"Close the world. Open the nExt.\")\n  (dashboard-startup-banner (expand-file-name \"images\/KEC_Dark_BK_Small.png\" user-emacs-directory))\n  (dashboard-items '((recents  . 7)\n                     (bookmarks . 7)\n                     (agenda . 5)))\n  (initial-buffer-choice (lambda () (get-buffer dashboard-buffer-name)))\n  (dashboard-set-heading-icons t)\n  (dashboard-set-navigator t)\n  (dashboard-navigator-buttons\n   (if (featurep 'all-the-icons)\n       `(((,(all-the-icons-octicon \"mark-github\" :height 1.1 :v-adjust -0.05)\n           \"M-EMACS\" \"Browse M-EMACS Homepage\"\n           (lambda (&amp;rest _) (browse-url \"https:\/\/github.com\/MatthewZMD\/.emacs.d\")))\n          (,(all-the-icons-fileicon \"elisp\" :height 1.0 :v-adjust -0.1)\n           \"Configuration\" \"\" (lambda (&amp;rest _) (edit-configs)))\n          (,(all-the-icons-faicon \"cogs\" :height 1.0 :v-adjust -0.1)\n           \"Update\" \"\" (lambda (&amp;rest _) (auto-package-update-now)))))\n     `(((\"\" \"M-EMACS\" \"Browse M-EMACS Homepage\"\n         (lambda (&amp;rest _) (browse-url \"https:\/\/github.com\/MatthewZMD\/.emacs.d\")))\n        (\"\" \"Configuration\" \"\" (lambda (&amp;rest _) (edit-configs)))\n        (\"\" \"Update\" \"\" (lambda (&amp;rest _) (auto-package-update-now)))))))\n  :custom-face\n  (dashboard-banner-logo-title ((t (:family \"Love LetterTW\" :height 123))))\n  :config\n  (dashboard-modify-heading-icons '((recents . \"file-text\")\n                                    (bookmarks . \"book\")))\n  (dashboard-setup-startup-hook)\n  ;; Open Dashboard function\n  (defun open-dashboard ()\n    \"Open the *dashboard* buffer and jump to the first widget.\"\n    (interactive)\n    (if (get-buffer dashboard-buffer-name)\n        (kill-buffer dashboard-buffer-name))\n    (dashboard-insert-startupify-lists)\n    (switch-to-buffer dashboard-buffer-name)\n    (goto-char (point-min))\n    (delete-other-windows)))<\/pre><\/div><h3><a id=\"user-content-page-break-lines\" class=\"anchor\" aria-hidden=\"true\" href=\"#page-break-lines\"><\/a>Page Break Lines<\/h3><p><a href=\"https:\/\/github.com\/purcell\/page-break-lines\">Page-break-lines<\/a>, a feature that displays ugly form feed characters as tidy horizontal rules.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package page-break-lines\n  :diminish\n  :init (global-page-break-lines-mode))<\/pre><\/div><h2><a id=\"user-content-fonts-and-icons\" class=\"anchor\" aria-hidden=\"true\" href=\"#fonts-and-icons\"><\/a>Fonts and Icons<\/h2><p><strong>Prerequisite<\/strong>: Install all the available fonts and icons from <code>fonts\/<\/code>.<\/p><h3><a id=\"user-content-fonts\" class=\"anchor\" aria-hidden=\"true\" href=\"#fonts\"><\/a>Fonts<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Input Mono, Monaco Style, Line Height 1.3 download from http:\/\/input.fontbureau.com\/\n(defvar font-list '((\"Input\" . 11) (\"SF Mono\" . 12) (\"Consolas\" . 12) (\"Love LetterTW\" . 12.5))\n  \"List of fonts and sizes.  The first one available will be used.\")<\/pre><\/div><p>Function to switch between fonts.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun change-font ()\n  \"Documentation.\"\n  (interactive)\n  (let* (available-fonts font-name font-size font-setting)\n    (dolist (font font-list (setq available-fonts (nreverse available-fonts)))\n      (when (member (car font) (font-family-list))\n        (push font available-fonts)))\n    (if (not available-fonts)\n        (message \"No fonts from the chosen set are available\")\n      (if (called-interactively-p 'interactive)\n          (let* ((chosen (assoc-string (completing-read \"What font to use? \" available-fonts nil t) available-fonts)))\n            (setq font-name (car chosen) font-size (read-number \"Font size: \" (cdr chosen))))\n        (setq font-name (caar available-fonts) font-size (cdar available-fonts)))\n      (setq font-setting (format \"%s-%d\" font-name font-size))\n      (set-frame-font font-setting nil t)\n      (add-to-list 'default-frame-alist (cons 'font font-setting)))))\n(when *sys\/gui*\n  (change-font))<\/pre><\/div><h3><a id=\"user-content-icons\" class=\"anchor\" aria-hidden=\"true\" href=\"#icons\"><\/a>Icons<\/h3><p><a href=\"https:\/\/github.com\/domtronn\/all-the-icons.el\">All The Icons<\/a>, a utility package to collect various Icon Fonts. Enable only in GUI Emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package all-the-icons :if *sys\/gui*)<\/pre><\/div><p><a href=\"https:\/\/github.com\/jtbm37\/all-the-icons-dired\">All The Icons Dired<\/a>, an icon set for Dired.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package all-the-icons-dired\n  :after all-the-icons\n  :if *sys\/gui*\n  :diminish\n  :custom-face\n  (all-the-icons-dired-dir-face ((t `(:foreground ,(face-background 'default)))))\n  :hook (dired-mode . all-the-icons-dired-mode)\n  :config\n  ;; Workaround for all-the-icons bug until PR merged https:\/\/github.com\/domtronn\/all-the-icons.el\/pull\/150\n  (when (require 'all-the-icons nil 'noerror)\n    (setq all-the-icons-mode-icon-alist\n          (delete '(erc-mode all-the-icons-faicon \"commenting-o\" :height 1.0 :v-adjust 0.0 :face all-the-icons-white) all-the-icons-mode-icon-alist))\n    (add-to-list 'all-the-icons-mode-icon-alist '(erc-mode all-the-icons-faicon \"commenting-o\" :height 1.0 :v-adjust 0.0))))<\/pre><\/div><h2><a id=\"user-content-smooth-scrolling\" class=\"anchor\" aria-hidden=\"true\" href=\"#smooth-scrolling\"><\/a>Smooth Scrolling<\/h2><p>Configurations to smooth scrolling.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Vertical Scroll\n(setq scroll-step 1)\n(setq scroll-margin 1)\n(setq scroll-conservatively 101)\n(setq scroll-up-aggressively 0.01)\n(setq scroll-down-aggressively 0.01)\n(setq auto-window-vscroll nil)\n(setq fast-but-imprecise-scrolling nil)\n(setq mouse-wheel-scroll-amount '(1 ((shift) . 1)))\n(setq mouse-wheel-progressive-speed nil)\n;; Horizontal Scroll\n(setq hscroll-step 1)\n(setq hscroll-margin 1)<\/pre><\/div><h2><a id=\"user-content-prettify-symbols\" class=\"anchor\" aria-hidden=\"true\" href=\"#prettify-symbols\"><\/a>Prettify Symbols<\/h2><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/PrettySymbol\" rel=\"nofollow\">Prettify symbols mode<\/a>, a built-in mode for displaying sequences of characters as fancy characters or symbols.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(global-prettify-symbols-mode 1)\n(defun add-pretty-lambda ()\n  \"Make some word or string show as pretty Unicode symbols.  See https:\/\/unicodelookup.com for more.\"\n  (setq prettify-symbols-alist\n        '(\n          (\"lambda\" . 955)\n          (\"delta\" . 120517)\n          (\"epsilon\" . 120518)\n          (\"-&gt;\" . 8594)\n          (\"&lt;=\" . 8804)\n          (\"&gt;=\" . 8805)\n          )))\n(add-hook 'prog-mode-hook 'add-pretty-lambda)\n(add-hook 'org-mode-hook 'add-pretty-lambda)<\/pre><\/div><h2><a id=\"user-content-ui-configs\" class=\"anchor\" aria-hidden=\"true\" href=\"#ui-configs\"><\/a>UI Configs<\/h2><h3><a id=\"user-content-title-bar\" class=\"anchor\" aria-hidden=\"true\" href=\"#title-bar\"><\/a>Title Bar<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq-default frame-title-format '(\"M-EMACS - \" user-login-name \"@\" system-name \" - %b\"))<\/pre><\/div><h3><a id=\"user-content-simplify-yesno-prompts\" class=\"anchor\" aria-hidden=\"true\" href=\"#simplify-yesno-prompts\"><\/a>Simplify Yes\/No Prompts<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(fset 'yes-or-no-p 'y-or-n-p)<\/pre><\/div><h3><a id=\"user-content-disable-splash-screen\" class=\"anchor\" aria-hidden=\"true\" href=\"#disable-splash-screen\"><\/a>Disable Splash Screen<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq inhibit-startup-screen t)\n(setq initial-major-mode 'text-mode)\n(setq initial-scratch-message \"Present Day, Present Time...\\n\")<\/pre><\/div><h3><a id=\"user-content-line-numbers\" class=\"anchor\" aria-hidden=\"true\" href=\"#line-numbers\"><\/a>Line Numbers<\/h3><p>Display line numbers, and column numbers in modeline.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Hook line numbers to only when files are opened, also use linum-mode for emacs-version&lt; 26\n(if (version&lt; emacs-version \"26\")\n    (global-linum-mode)\n  (add-hook 'text-mode-hook #'display-line-numbers-mode)\n  (add-hook 'prog-mode-hook #'display-line-numbers-mode))\n;; Display column numbers in modeline\n(column-number-mode 1)<\/pre><\/div><h3><a id=\"user-content-modeline-time-and-battery\" class=\"anchor\" aria-hidden=\"true\" href=\"#modeline-time-and-battery\"><\/a>Modeline Time and Battery<\/h3><p>Display time and battery information in modeline.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(display-time-mode 1)\n(display-battery-mode 1)<\/pre><\/div><h2><a id=\"user-content-magit\" class=\"anchor\" aria-hidden=\"true\" href=\"#magit\"><\/a>Magit<\/h2><p><a href=\"https:\/\/magit.vc\/\" rel=\"nofollow\">Magit<\/a>, an interface to the version control system Git.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package magit\n  :if *git*\n  :bind (\"C-x g\" . magit-status))<\/pre><\/div><h2><a id=\"user-content-projectile\" class=\"anchor\" aria-hidden=\"true\" href=\"#projectile\"><\/a>Projectile<\/h2><p><a href=\"https:\/\/github.com\/bbatsov\/projectile\">Projectile<\/a>, a Project Interaction Library for Emacs.<\/p><p><strong>Prerequisite<\/strong>: Windows OS: Install <a href=\"https:\/\/github.com\/bmatzelle\/gow\/releases\">Gow<\/a> and ensure it's in <code>PATH<\/code>.<\/p><p><a href=\"https:\/\/github.com\/bmatzelle\/gow\">Gow<\/a> is a lightweight installer that installs useful open source UNIX applications compiled as native win32 binaries. Specifically, <code>tr<\/code> is needed for Projectile alien indexing.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package projectile\n  :bind\n  (\"C-c p\" . projectile-command-map)\n  (\"C-z o\" . projectile-find-file)\n  (\"C-z p\" . projectile-add-known-project)\n  :custom\n  (projectile-completion-system 'ivy)\n  :config\n  (projectile-mode 1)\n  (when (and *sys\/win32* *tr*)\n    (setq projectile-indexing-method 'alien))\n  (add-to-list 'projectile-globally-ignored-directories \"node_modules\"))<\/pre><\/div><h2><a id=\"user-content-treemacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#treemacs\"><\/a>Treemacs<\/h2><p><a href=\"https:\/\/github.com\/Alexander-Miller\/treemacs\">Treemacs<\/a>, a tree layout file explorer for Emacs.<\/p><h3><a id=\"user-content-treemacs-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#treemacs-1\"><\/a>Treemacs<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package treemacs\n  :init\n  (with-eval-after-load 'winum\n    (define-key winum-keymap (kbd \"M-0\") #'treemacs-select-window))\n  :custom\n  (treemacs-collapse-dirs 3)\n  (treemacs-deferred-git-apply-delay 0.5)\n  (treemacs-display-in-side-window t)\n  (treemacs-file-event-delay 5000)\n  (treemacs-file-follow-delay 0.2)\n  (treemacs-follow-after-init t)\n  (treemacs-follow-recenter-distance 0.1)\n  (treemacs-git-command-pipe \"\")\n  (treemacs-goto-tag-strategy 'refetch-index)\n  (treemacs-indentation 2)\n  (treemacs-indentation-string \" \")\n  (treemacs-is-never-other-window nil)\n  (treemacs-max-git-entries 5000)\n  (treemacs-no-png-images nil)\n  (treemacs-no-delete-other-windows t)\n  (treemacs-project-follow-cleanup nil)\n  (treemacs-persist-file (expand-file-name \".cache\/treemacs-persist\" user-emacs-directory))\n  (treemacs-recenter-after-file-follow nil)\n  (treemacs-recenter-after-tag-follow nil)\n  (treemacs-show-cursor nil)\n  (treemacs-show-hidden-files t)\n  (treemacs-silent-filewatch nil)\n  (treemacs-silent-refresh nil)\n  (treemacs-sorting 'alphabetic-desc)\n  (treemacs-space-between-root-nodes t)\n  (treemacs-tag-follow-cleanup t)\n  (treemacs-tag-follow-delay 1.5)\n  (treemacs-width 35)\n  :config\n  ;; The default width and height of the icons is 22 pixels. If you are\n  ;; using a Hi-DPI display, uncomment this to double the icon size.\n  ;;(treemacs-resize-icons 44)\n  (treemacs-follow-mode t)\n  (treemacs-filewatch-mode t)\n  (treemacs-fringe-indicator-mode t)\n  :bind\n  ((\"M-0\"       . treemacs-select-window)\n   (\"C-x t 1\"   . treemacs-delete-other-windows)\n   (\"C-x t t\"   . treemacs)\n   (\"C-x t B\"   . treemacs-bookmark)\n   (\"C-x t C-t\" . treemacs-find-file)\n   (\"C-x t M-t\" . treemacs-find-tag))\n  (:map treemacs-mode-map (\"C-p\" . treemacs-previous-line)))<\/pre><\/div><h3><a id=\"user-content-treemacs-magit\" class=\"anchor\" aria-hidden=\"true\" href=\"#treemacs-magit\"><\/a>Treemacs Magit<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package treemacs-magit\n  :defer t\n  :after (treemacs magit))<\/pre><\/div><h3><a id=\"user-content-treemacs-projectile\" class=\"anchor\" aria-hidden=\"true\" href=\"#treemacs-projectile\"><\/a>Treemacs Projectile<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package treemacs-projectile\n  :defer t\n  :after (treemacs projectile))<\/pre><\/div><h2><a id=\"user-content-yasnippet\" class=\"anchor\" aria-hidden=\"true\" href=\"#yasnippet\"><\/a>YASnippet<\/h2><h3><a id=\"user-content-yasnippet-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#yasnippet-1\"><\/a>YASnippet<\/h3><p><a href=\"https:\/\/github.com\/joaotavora\/yasnippet\">YASnippet<\/a>, a programming template system for Emacs. It loads <a href=\"https:\/\/github.com\/AndreaCrotti\/yasnippet-snippets\">YASnippet Snippets<\/a>, a collection of yasnippet snippets for many languages.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package yasnippet\n  :diminish yas-minor-mode\n  :init\n  (use-package yasnippet-snippets :after yasnippet)\n  :hook ((prog-mode LaTeX-mode org-mode) . yas-minor-mode)\n  :bind\n  (:map yas-minor-mode-map (\"C-c C-n\" . yas-expand-from-trigger-key))\n  (:map yas-keymap\n        ((\"TAB\" . smarter-yas-expand-next-field)\n         ([(tab)] . smarter-yas-expand-next-field)))\n  :config\n  (yas-reload-all)\n  (defun smarter-yas-expand-next-field ()\n    \"Try to `yas-expand' then `yas-next-field' at current cursor position.\"\n    (interactive)\n    (let ((old-point (point))\n          (old-tick (buffer-chars-modified-tick)))\n      (yas-expand)\n      (when (and (eq old-point (point))\n                 (eq old-tick (buffer-chars-modified-tick)))\n        (ignore-errors (yas-next-field))))))<\/pre><\/div><h2><a id=\"user-content-flycheck\" class=\"anchor\" aria-hidden=\"true\" href=\"#flycheck\"><\/a>Flycheck<\/h2><p><a href=\"https:\/\/www.flycheck.org\/en\/latest\/\" rel=\"nofollow\">Flycheck<\/a>, a syntax checking extension.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package flycheck\n  :defer t\n  :hook (prog-mode . flycheck-mode)\n  :custom\n  (flycheck-emacs-lisp-load-path 'inherit)\n  :config\n  (flycheck-add-mode 'javascript-eslint 'js-mode)\n  (flycheck-add-mode 'typescript-tslint 'rjsx-mode))<\/pre><\/div><h2><a id=\"user-content-dumb-jump\" class=\"anchor\" aria-hidden=\"true\" href=\"#dumb-jump\"><\/a>Dumb Jump<\/h2><p><a href=\"https:\/\/github.com\/jacktasia\/dumb-jump\">Dumb jump<\/a>, an Emacs \"jump to definition\" package.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package dumb-jump\n  :bind\n  (:map prog-mode-map\n        ((\"C-c C-o\" . dumb-jump-go-other-window)\n         (\"C-c C-j\" . dumb-jump-go)\n         (\"C-c C-i\" . dumb-jump-go-prompt)))\n  :custom (dumb-jump-selector 'ivy))<\/pre><\/div><h2><a id=\"user-content-parenthesis\" class=\"anchor\" aria-hidden=\"true\" href=\"#parenthesis\"><\/a>Parenthesis<\/h2><h3><a id=\"user-content-smartparens\" class=\"anchor\" aria-hidden=\"true\" href=\"#smartparens\"><\/a>Smartparens<\/h3><p><a href=\"https:\/\/github.com\/Fuco1\/smartparens\">Smartparens<\/a>, a minor mode for dealing with pairs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package smartparens\n  :hook (prog-mode . smartparens-mode)\n  :diminish smartparens-mode\n  :bind\n  (:map smartparens-mode-map\n        (\"C-M-f\" . sp-forward-sexp)\n        (\"C-M-b\" . sp-backward-sexp)\n        (\"C-M-a\" . sp-backward-down-sexp)\n        (\"C-M-e\" . sp-up-sexp)\n        (\"C-M-w\" . sp-copy-sexp)\n        (\"C-M-k\" . sp-change-enclosing)\n        (\"M-k\" . sp-kill-sexp)\n        (\"C-M-&lt;backspace&gt;\" . sp-splice-sexp-killing-backward)\n        (\"C-S-&lt;backspace&gt;\" . sp-splice-sexp-killing-around)\n        (\"C-]\" . sp-select-next-thing-exchange))\n  :custom\n  (sp-escape-quotes-after-insert nil)\n  :config\n  ;; Stop pairing single quotes in elisp\n  (sp-local-pair 'emacs-lisp-mode \"'\" nil :actions nil)\n  (sp-local-pair 'org-mode \"[\" nil :actions nil)\n  ;; Smartparens is broken in `cc-mode' as of Emacs 27. See\n  ;; https:\/\/github.com\/Fuco1\/smartparens\/issues\/963\n  (unless (version&lt; emacs-version \"27\")\n    (dolist (fun '(c-electric-paren c-electric-brace))\n      (add-to-list 'sp--special-self-insert-commands fun))))<\/pre><\/div><h3><a id=\"user-content-match-parenthesis\" class=\"anchor\" aria-hidden=\"true\" href=\"#match-parenthesis\"><\/a>Match Parenthesis<\/h3><p>Match and automatically pair parenthesis, and show parenthesis even when it went offscreen from <a href=\"https:\/\/with-emacs.com\/posts\/editing\/show-matching-lines-when-parentheses-go-off-screen\/\" rel=\"nofollow\">Clemens Radermacher<\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Show matching parenthesis\n(show-paren-mode 1)\n;; we will call `blink-matching-open` ourselves...\n(remove-hook 'post-self-insert-hook\n             #'blink-paren-post-self-insert-function)\n;; this still needs to be set for `blink-matching-open` to work\n(setq blink-matching-paren 'show)\n(let ((ov nil)) ; keep track of the overlay\n  (advice-add\n   #'show-paren-function\n   :after\n    (defun show-paren--off-screen+ (&amp;rest _args)\n      \"Display matching line for off-screen paren.\"\n      (when (overlayp ov)\n        (delete-overlay ov))\n      ;; check if it's appropriate to show match info,\n      ;; see `blink-paren-post-self-insert-function'\n      (when (and (overlay-buffer show-paren--overlay)\n                 (not (or cursor-in-echo-area\n                          executing-kbd-macro\n                          noninteractive\n                          (minibufferp)\n                          this-command))\n                 (and (not (bobp))\n                      (memq (char-syntax (char-before)) '(?\\) ?\\$)))\n                 (= 1 (logand 1 (- (point)\n                                   (save-excursion\n                                     (forward-char -1)\n                                     (skip-syntax-backward \"\/\\\\\")\n                                     (point))))))\n        ;; rebind `minibuffer-message' called by\n        ;; `blink-matching-open' to handle the overlay display\n        (cl-letf (((symbol-function #'minibuffer-message)\n                   (lambda (msg &amp;rest args)\n                     (let ((msg (apply #'format-message msg args)))\n                       (setq ov (display-line-overlay+\n                                 (window-start) msg))))))\n          (blink-matching-open))))))<\/pre><\/div><h2><a id=\"user-content-indentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#indentation\"><\/a>Indentation<\/h2><p><a href=\"https:\/\/github.com\/DarthFennec\/highlight-indent-guides\">Highlight Indent Guides<\/a>, a feature that highlights indentation levels.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package highlight-indent-guides\n  :if *sys\/gui*\n  :diminish\n  :hook ((prog-mode web-mode nxml-mode) . highlight-indent-guides-mode)\n  :custom\n  (highlight-indent-guides-method 'character)\n  (highlight-indent-guides-responsive 'top)\n  (highlight-indent-guides-delay 0)\n  (highlight-indent-guides-auto-character-face-perc 7))<\/pre><\/div><p>Indentation Configuration<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq-default indent-tabs-mode nil)\n(setq-default indent-line-function 'insert-tab)\n(setq-default tab-width 4)\n(setq-default c-basic-offset 4)\n(setq-default js-switch-indent-offset 4)\n(c-set-offset 'comment-intro 0)\n(c-set-offset 'innamespace 0)\n(c-set-offset 'case-label '+)\n(c-set-offset 'access-label 0)\n(c-set-offset (quote cpp-macro) 0 nil)\n(add-hook 'after-change-major-mode-hook\n          (lambda () (if (equal electric-indent-mode 't)\n                    (when (derived-mode-p 'text-mode)\n                      (electric-indent-mode -1))\n                  (electric-indent-mode 1))))<\/pre><\/div><h2><a id=\"user-content-quickrun\" class=\"anchor\" aria-hidden=\"true\" href=\"#quickrun\"><\/a>Quickrun<\/h2><p><a href=\"https:\/\/github.com\/syohex\/emacs-quickrun\">Quickrun<\/a>, compile and run source code quickly.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package quickrun\n  :bind\n  ((\"&lt;f5&gt;\" . quickrun)\n   (\"M-&lt;f5&gt;\" . quickrun-shell)))<\/pre><\/div><h2><a id=\"user-content-format-all\" class=\"anchor\" aria-hidden=\"true\" href=\"#format-all\"><\/a>Format All<\/h2><p><a href=\"https:\/\/github.com\/lassik\/emacs-format-all-the-code\">Format all<\/a>, a feature that lets you auto-format source code.<\/p><p><strong>Prerequisite<\/strong>: Read <a href=\"https:\/\/github.com\/lassik\/emacs-format-all-the-code#supported-languages\">Supported Languages<\/a> to see which additional tool you need to install for the specific language.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package format-all\n  :bind (\"C-c C-f\" . format-all-buffer))<\/pre><\/div><h2><a id=\"user-content-evil-nerd-commenter\" class=\"anchor\" aria-hidden=\"true\" href=\"#evil-nerd-commenter\"><\/a>Evil Nerd Commenter<\/h2><p><a href=\"https:\/\/github.com\/redguardtoo\/evil-nerd-commenter\">Evil Nerd Commenter<\/a>, a tool that helps you comment code efficiently.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package evil-nerd-commenter\n  :bind\n  ((\"C-c M-;\" . c-toggle-comment-style)\n   (\"M-;\" . evilnc-comment-or-uncomment-lines)))<\/pre><\/div><h2><a id=\"user-content-editing\" class=\"anchor\" aria-hidden=\"true\" href=\"#editing\"><\/a>Editing<\/h2><h3><a id=\"user-content-iedit\" class=\"anchor\" aria-hidden=\"true\" href=\"#iedit\"><\/a>Iedit<\/h3><p><a href=\"https:\/\/github.com\/victorhge\/iedit\">Iedit<\/a>, a minor mode that allows editing multiple regions simultaneousy in a buffer or a region.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package iedit\n  :bind (\"C-z ,\" . iedit-mode)\n  :diminish)<\/pre><\/div><h3><a id=\"user-content-awesome-pair\" class=\"anchor\" aria-hidden=\"true\" href=\"#awesome-pair\"><\/a>Awesome Pair<\/h3><p><a href=\"https:\/\/github.com\/manateelazycat\/awesome-pair\">Awesome Pair<\/a>, a feature that provides grammatical parenthesis completion.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package awesome-pair\n  :load-path (lambda () (expand-file-name \"site-elisp\/awesome-pair\" user-emacs-directory))\n  :bind\n  (:map prog-mode-map\n        ((\"M-D\" . awesome-pair-kill)\n         (\"SPC\" . awesome-pair-space)\n         (\"=\" . awesome-pair-equal)\n         (\"M-F\" . awesome-pair-jump-right)\n         (\"M-B\" . awesome-pair-jump-left)))\n  :hook (prog-mode . awesome-pair-mode))<\/pre><\/div><h3><a id=\"user-content-delete-block\" class=\"anchor\" aria-hidden=\"true\" href=\"#delete-block\"><\/a>Delete Block<\/h3><p><a href=\"https:\/\/github.com\/manateelazycat\/delete-block\">Delete Block<\/a>, a feature that deletes block efficiently.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package delete-block\n  :load-path (lambda () (expand-file-name \"site-elisp\/delete-block\" user-emacs-directory))\n  :bind\n  ((\"M-d\" . delete-block-forward)\n   (\"C-&lt;backspace&gt;\" . delete-block-backward)\n   (\"M-&lt;backspace&gt;\" . delete-block-backward)\n   (\"M-DEL\" . delete-block-backward)))<\/pre><\/div><h2><a id=\"user-content-headers\" class=\"anchor\" aria-hidden=\"true\" href=\"#headers\"><\/a>Headers<\/h2><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/header2.el\" rel=\"nofollow\">Header2<\/a>, a support for creation and update of file headers.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package header2\n  :load-path (lambda () (expand-file-name \"site-elisp\/header2\" user-emacs-directory))\n  :custom\n  (header-copyright-notice (concat \"Copyright (C) 2019 \" (user-full-name) \"\\n\"))\n  :hook (emacs-lisp-mode . auto-make-header)\n  :config\n  (add-to-list 'write-file-functions 'auto-update-file-header)\n  (autoload 'auto-make-header \"header2\")\n  (autoload 'auto-update-file-header \"header2\"))<\/pre><\/div><h2><a id=\"user-content-jupyter-notebook\" class=\"anchor\" aria-hidden=\"true\" href=\"#jupyter-notebook\"><\/a>Jupyter Notebook<\/h2><p><a href=\"https:\/\/github.com\/millejoh\/emacs-ipython-notebook\">Emacs IPython Notebook<\/a>, a <a href=\"https:\/\/jupyter.org\/\" rel=\"nofollow\">Jupyter<\/a> (formerly IPython) client in Emacs.<\/p><h3><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h3><ol><li>Execute <code>M-x ein:run<\/code> to launch a local Jupyter session.<\/li>\n<li>Login with <code>M-x ein:login<\/code> to a local or remote session.<\/li>\n<li>Open <code>.ipynb<\/code> file and press <code>C-c C-o<\/code>.<\/li>\n<\/ol><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package ein\n  :disabled\n  :defer t)<\/pre><\/div><h2><a id=\"user-content-lsp\" class=\"anchor\" aria-hidden=\"true\" href=\"#lsp\"><\/a>LSP<\/h2><h3><a id=\"user-content-lsp-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#lsp-mode\"><\/a>LSP Mode<\/h3><p><a href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\">Language Server Protocol Mode<\/a>, a client\/library for the <a href=\"https:\/\/microsoft.github.io\/language-server-protocol\/\" rel=\"nofollow\">Language Server Protocol<\/a>. M-EMACS tries to use lsp-mode whenever possible.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lsp-mode\n  :defer t\n  :commands lsp\n  :custom\n  (lsp-auto-guess-root nil)\n  (lsp-prefer-flymake nil) ; Use flycheck instead of flymake\n  (lsp-file-watch-threshold 2000)\n  (read-process-output-max (* 1024 1024))\n  :bind (:map lsp-mode-map (\"C-c C-f\" . lsp-format-buffer))\n  :hook ((java-mode python-mode go-mode\n          js-mode js2-mode typescript-mode web-mode\n          c-mode c++-mode objc-mode) . lsp))<\/pre><\/div><h3><a id=\"user-content-lsp-ui\" class=\"anchor\" aria-hidden=\"true\" href=\"#lsp-ui\"><\/a>LSP UI<\/h3><p><a href=\"https:\/\/github.com\/emacs-lsp\/lsp-ui\">Language Server Protocol UI<\/a>, provides all the higher level UI modules of lsp-mode, like flycheck support and code lenses.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lsp-ui\n  :after lsp-mode\n  :diminish\n  :commands lsp-ui-mode\n  :custom-face\n  (lsp-ui-doc-background ((t (:background nil))))\n  (lsp-ui-doc-header ((t (:inherit (font-lock-string-face italic)))))\n  :bind (:map lsp-ui-mode-map\n              ([remap xref-find-definitions] . lsp-ui-peek-find-definitions)\n              ([remap xref-find-references] . lsp-ui-peek-find-references)\n              (\"C-c u\" . lsp-ui-imenu))\n  :custom\n  (lsp-ui-doc-enable t)\n  (lsp-ui-doc-header t)\n  (lsp-ui-doc-include-signature t)\n  (lsp-ui-doc-position 'top)\n  (lsp-ui-doc-border (face-foreground 'default))\n  (lsp-ui-sideline-enable nil)\n  (lsp-ui-sideline-ignore-duplicate t)\n  (lsp-ui-sideline-show-code-actions nil)\n  :config\n  ;; Use lsp-ui-doc-webkit only in GUI\n  (if *sys\/gui*\n      (setq lsp-ui-doc-use-webkit t))\n  ;; WORKAROUND Hide mode-line of the lsp-ui-imenu buffer\n  ;; https:\/\/github.com\/emacs-lsp\/lsp-ui\/issues\/243\n  (defadvice lsp-ui-imenu (after hide-lsp-ui-imenu-mode-line activate)\n    (setq mode-line-format nil)))<\/pre><\/div><h2><a id=\"user-content-dap\" class=\"anchor\" aria-hidden=\"true\" href=\"#dap\"><\/a>DAP<\/h2><p><a href=\"https:\/\/github.com\/emacs-lsp\/dap-mode\">Debug Adapter Protocol Mode<\/a>, a client\/library for the <a href=\"https:\/\/code.visualstudio.com\/api\/extension-guides\/debugger-extension\" rel=\"nofollow\">Debug Adapter Protocol<\/a>.<\/p><p><strong>Prerequisite<\/strong>: See <a href=\"https:\/\/github.com\/emacs-lsp\/dap-mode#configuration\">Configuration<\/a> to configure DAP appropriately.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package dap-mode\n  :diminish\n  :bind\n  (:map dap-mode-map\n        ((\"&lt;f12&gt;\" . dap-debug)\n         (\"&lt;f8&gt;\" . dap-continue)\n         (\"&lt;f9&gt;\" . dap-next)\n         (\"&lt;M-f11&gt;\" . dap-step-in)\n         (\"C-M-&lt;f11&gt;\" . dap-step-out)\n         (\"&lt;f7&gt;\" . dap-breakpoint-toggle)))\n  :hook ((after-init . dap-mode)\n         (dap-mode . dap-ui-mode)\n         (python-mode . (lambda () (require 'dap-python)))\n         (ruby-mode . (lambda () (require 'dap-ruby)))\n         (go-mode . (lambda () (require 'dap-go)))\n         (java-mode . (lambda () (require 'dap-java)))\n         ((c-mode c++-mode objc-mode swift) . (lambda () (require 'dap-lldb)))\n         (php-mode . (lambda () (require 'dap-php)))\n         (elixir-mode . (lambda () (require 'dap-elixir)))\n         ((js-mode js2-mode typescript-mode) . (lambda () (require 'dap-chrome)))))<\/pre><\/div><h2><a id=\"user-content-company\" class=\"anchor\" aria-hidden=\"true\" href=\"#company\"><\/a>Company<\/h2><h3><a id=\"user-content-company-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#company-mode\"><\/a>Company Mode<\/h3><p><a href=\"http:\/\/company-mode.github.io\/\" rel=\"nofollow\">Company<\/a>, a text completion framework for Emacs.<\/p><p>The function <code>smarter-yas-expand-next-field-complete<\/code> is to smartly resolve TAB conflicts in company and yasnippet packages.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package company\n  :diminish company-mode\n  :hook ((prog-mode LaTeX-mode latex-mode ess-r-mode) . company-mode)\n  :bind\n  (:map company-active-map\n        ([tab] . smarter-yas-expand-next-field-complete)\n        (\"TAB\" . smarter-yas-expand-next-field-complete))\n  :custom\n  (company-minimum-prefix-length 1)\n  (company-tooltip-align-annotations t)\n  (company-begin-commands '(self-insert-command))\n  (company-require-match 'never)\n  ;; Don't use company in the following modes\n  (company-global-modes '(not shell-mode eaf-mode))\n  ;; Trigger completion immediately.\n  (company-idle-delay 0.1)\n  ;; Number the candidates (use M-1, M-2 etc to select completions).\n  (company-show-numbers t)\n  :config\n  (unless *clangd* (delete 'company-clang company-backends))\n  (global-company-mode 1)\n  (defun smarter-yas-expand-next-field-complete ()\n    \"Try to `yas-expand' and `yas-next-field' at current cursor position.\nIf failed try to complete the common part with `company-complete-common'\"\n    (interactive)\n    (if yas-minor-mode\n        (let ((old-point (point))\n              (old-tick (buffer-chars-modified-tick)))\n          (yas-expand)\n          (when (and (eq old-point (point))\n                     (eq old-tick (buffer-chars-modified-tick)))\n            (ignore-errors (yas-next-field))\n            (when (and (eq old-point (point))\n                       (eq old-tick (buffer-chars-modified-tick)))\n              (company-complete-common))))\n      (company-complete-common))))<\/pre><\/div><h3><a id=\"user-content-company-lsp\" class=\"anchor\" aria-hidden=\"true\" href=\"#company-lsp\"><\/a>Company LSP<\/h3><p><a href=\"https:\/\/github.com\/tigersoldier\/company-lsp\">Company LSP<\/a>, a Company completion backend for lsp-mode.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package company-lsp\n  :defer t\n  :custom (company-lsp-cache-candidates 'auto))<\/pre><\/div><h3><a id=\"user-content-company-tabnine\" class=\"anchor\" aria-hidden=\"true\" href=\"#company-tabnine\"><\/a>Company TabNine<\/h3><p><a href=\"https:\/\/github.com\/TommyX12\/company-tabnine\">Company TabNine<\/a>, A company-mode backend for <a href=\"https:\/\/tabnine.com\/\" rel=\"nofollow\">TabNine<\/a>, the all-language autocompleter.<\/p><p>This is enabled by default, if ever you find it not good enough for a particular completion, simply use <code>M-q<\/code> to immediately switch to default backends.<\/p><p><strong>Prerequisite<\/strong>: Execute <code>M-x company-tabnine-install-binary<\/code> to install the TabNine binary for your system.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package company-tabnine\n  :defer 1\n  :custom\n  (company-tabnine-max-num-results 9)\n  :bind\n  ((\"M-q\" . company-other-backend)\n   (\"C-z t\" . company-tabnine))\n  :hook\n  (lsp-after-open . (lambda ()\n                      (setq company-tabnine-max-num-results 3)\n                      (add-to-list 'company-transformers 'company\/\/sort-by-tabnine t)\n                      (add-to-list 'company-backends '(company-lsp :with company-tabnine :separate))))\n  (kill-emacs . company-tabnine-kill-process)\n  :config\n  ;; Enable TabNine on default\n  (add-to-list 'company-backends #'company-tabnine)\n  ;; Integrate company-tabnine with lsp-mode\n  (defun company\/\/sort-by-tabnine (candidates)\n    (if (or (functionp company-backend)\n            (not (and (listp company-backend) (memq 'company-tabnine company-backend))))\n        candidates\n      (let ((candidates-table (make-hash-table :test #'equal))\n            candidates-lsp\n            candidates-tabnine)\n        (dolist (candidate candidates)\n          (if (eq (get-text-property 0 'company-backend candidate)\n                  'company-tabnine)\n              (unless (gethash candidate candidates-table)\n                (push candidate candidates-tabnine))\n            (push candidate candidates-lsp)\n            (puthash candidate t candidates-table)))\n        (setq candidates-lsp (nreverse candidates-lsp))\n        (setq candidates-tabnine (nreverse candidates-tabnine))\n        (nconc (seq-take candidates-tabnine 3)\n               (seq-take candidates-lsp 6))))))<\/pre><\/div><h3><a id=\"user-content-company-box\" class=\"anchor\" aria-hidden=\"true\" href=\"#company-box\"><\/a>Company Box<\/h3><p><a href=\"https:\/\/github.com\/sebastiencs\/company-box\">Company Box<\/a>, a company front-end with icons.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package company-box\n  :diminish\n  :functions (my-company-box--make-line\n              my-company-box-icons--elisp)\n  :commands (company-box--get-color\n             company-box--resolve-colors\n             company-box--add-icon\n             company-box--apply-color\n             company-box--make-line\n             company-box-icons--elisp)\n  :hook (company-mode . company-box-mode)\n  :custom\n  (company-box-backends-colors nil)\n  (company-box-show-single-candidate t)\n  (company-box-max-candidates 50)\n  (company-box-doc-delay 0.3)\n  :config\n  ;; Support `company-common'\n  (defun my-company-box--make-line (candidate)\n    (-let* (((candidate annotation len-c len-a backend) candidate)\n            (color (company-box--get-color backend))\n            ((c-color a-color i-color s-color) (company-box--resolve-colors color))\n            (icon-string (and company-box--with-icons-p (company-box--add-icon candidate)))\n            (candidate-string (concat (propertize (or company-common \"\") 'face 'company-tooltip-common)\n                                      (substring (propertize candidate 'face 'company-box-candidate) (length company-common) nil)))\n            (align-string (when annotation\n                            (concat \" \" (and company-tooltip-align-annotations\n                                             (propertize \" \" 'display `(space :align-to (- right-fringe ,(or len-a 0) 1)))))))\n            (space company-box--space)\n            (icon-p company-box-enable-icon)\n            (annotation-string (and annotation (propertize annotation 'face 'company-box-annotation)))\n            (line (concat (unless (or (and (= space 2) icon-p) (= space 0))\n                            (propertize \" \" 'display `(space :width ,(if (or (= space 1) (not icon-p)) 1 0.75))))\n                          (company-box--apply-color icon-string i-color)\n                          (company-box--apply-color candidate-string c-color)\n                          align-string\n                          (company-box--apply-color annotation-string a-color)))\n            (len (length line)))\n      (add-text-properties 0 len (list 'company-box--len (+ len-c len-a)\n                                       'company-box--color s-color)\n                           line)\n      line))\n  (advice-add #'company-box--make-line :override #'my-company-box--make-line)\n  ;; Prettify icons\n  (defun my-company-box-icons--elisp (candidate)\n    (when (derived-mode-p 'emacs-lisp-mode)\n      (let ((sym (intern candidate)))\n        (cond ((fboundp sym) 'Function)\n              ((featurep sym) 'Module)\n              ((facep sym) 'Color)\n              ((boundp sym) 'Variable)\n              ((symbolp sym) 'Text)\n              (t . nil)))))\n  (advice-add #'company-box-icons--elisp :override #'my-company-box-icons--elisp)\n  (when (and *sys\/gui*\n             (require 'all-the-icons nil t))\n    (declare-function all-the-icons-faicon 'all-the-icons)\n    (declare-function all-the-icons-material 'all-the-icons)\n    (declare-function all-the-icons-octicon 'all-the-icons)\n    (setq company-box-icons-all-the-icons\n          `((Unknown . ,(all-the-icons-material \"find_in_page\" :height 0.85 :v-adjust -0.2))\n            (Text . ,(all-the-icons-faicon \"text-width\" :height 0.8 :v-adjust -0.05))\n            (Method . ,(all-the-icons-faicon \"cube\" :height 0.8 :v-adjust -0.05 :face 'all-the-icons-purple))\n            (Function . ,(all-the-icons-faicon \"cube\" :height 0.8 :v-adjust -0.05 :face 'all-the-icons-purple))\n            (Constructor . ,(all-the-icons-faicon \"cube\" :height 0.8 :v-adjust -0.05 :face 'all-the-icons-purple))\n            (Field . ,(all-the-icons-octicon \"tag\" :height 0.8 :v-adjust 0 :face 'all-the-icons-lblue))\n            (Variable . ,(all-the-icons-octicon \"tag\" :height 0.8 :v-adjust 0 :face 'all-the-icons-lblue))\n            (Class . ,(all-the-icons-material \"settings_input_component\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-orange))\n            (Interface . ,(all-the-icons-material \"share\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-lblue))\n            (Module . ,(all-the-icons-material \"view_module\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-lblue))\n            (Property . ,(all-the-icons-faicon \"wrench\" :height 0.8 :v-adjust -0.05))\n            (Unit . ,(all-the-icons-material \"settings_system_daydream\" :height 0.85 :v-adjust -0.2))\n            (Value . ,(all-the-icons-material \"format_align_right\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-lblue))\n            (Enum . ,(all-the-icons-material \"storage\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-orange))\n            (Keyword . ,(all-the-icons-material \"filter_center_focus\" :height 0.85 :v-adjust -0.2))\n            (Snippet . ,(all-the-icons-material \"format_align_center\" :height 0.85 :v-adjust -0.2))\n            (Color . ,(all-the-icons-material \"palette\" :height 0.85 :v-adjust -0.2))\n            (File . ,(all-the-icons-faicon \"file-o\" :height 0.85 :v-adjust -0.05))\n            (Reference . ,(all-the-icons-material \"collections_bookmark\" :height 0.85 :v-adjust -0.2))\n            (Folder . ,(all-the-icons-faicon \"folder-open\" :height 0.85 :v-adjust -0.05))\n            (EnumMember . ,(all-the-icons-material \"format_align_right\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-lblue))\n            (Constant . ,(all-the-icons-faicon \"square-o\" :height 0.85 :v-adjust -0.05))\n            (Struct . ,(all-the-icons-material \"settings_input_component\" :height 0.85 :v-adjust -0.2 :face 'all-the-icons-orange))\n            (Event . ,(all-the-icons-faicon \"bolt\" :height 0.8 :v-adjust -0.05 :face 'all-the-icons-orange))\n            (Operator . ,(all-the-icons-material \"control_point\" :height 0.85 :v-adjust -0.2))\n            (TypeParameter . ,(all-the-icons-faicon \"arrows\" :height 0.8 :v-adjust -0.05))\n            (Template . ,(all-the-icons-material \"format_align_center\" :height 0.85 :v-adjust -0.2)))\n          company-box-icons-alist 'company-box-icons-all-the-icons)))<\/pre><\/div><h2><a id=\"user-content-java\" class=\"anchor\" aria-hidden=\"true\" href=\"#java\"><\/a>Java<\/h2><h3><a id=\"user-content-lsp-java\" class=\"anchor\" aria-hidden=\"true\" href=\"#lsp-java\"><\/a>LSP Java<\/h3><p><a href=\"https:\/\/github.com\/emacs-lsp\/lsp-java\">LSP Java<\/a>, Emacs Java IDE using <a href=\"https:\/\/projects.eclipse.org\/projects\/eclipse.jdt.ls\" rel=\"nofollow\">Eclipse JDT Language Server<\/a>. Note that this package is dependant on <a href=\"https:\/\/github.com\/tkf\/emacs-request\">Request<\/a>.<\/p><p><strong>Prerequisite<\/strong>: Install <a href=\"https:\/\/maven.apache.org\/download.cgi\" rel=\"nofollow\">Maven<\/a> and ensure it's in <code>PATH<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lsp-java\n  :after lsp-mode\n  :if *mvn*\n  :config\n  (use-package request :defer t)\n  :custom\n  (lsp-java-server-install-dir (expand-file-name \"~\/.emacs.d\/eclipse.jdt.ls\/server\/\"))\n  (lsp-java-workspace-dir (expand-file-name \"~\/.emacs.d\/eclipse.jdt.ls\/workspace\/\")))<\/pre><\/div><h2><a id=\"user-content-ccobjective-c\" class=\"anchor\" aria-hidden=\"true\" href=\"#ccobjective-c\"><\/a>C\/C++\/Objective C<\/h2><p><strong>Prerequisite<\/strong>: Since all completion features are provided by <a href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\">LSP Mode<\/a>, it needs to setup.<\/p><ul><li>Install <a href=\"https:\/\/cmake.org\/download\/\" rel=\"nofollow\">CMake<\/a> &gt;= 3.8 for all OS.<\/li>\n<li>*nix OS:\n<ul><li>It is suggested to use <a href=\"https:\/\/github.com\/MaskRay\/ccls\">CCLS<\/a> as LSP server. Now <a href=\"https:\/\/github.com\/MaskRay\/ccls\/wiki\/Build\">build<\/a> it.<\/li>\n<li>Set <code>ccls-executable<\/code> to the directory where your ccls is built.<\/li>\n<\/ul><\/li>\n<li>Windows OS:\n<ul><li>Install <a href=\"http:\/\/www.mingw.org\/wiki\/Install_MinGW\" rel=\"nofollow\">MinGW<\/a> for Compilation.<\/li>\n<li>It is a pain to build CCLS on Windows, install <a href=\"https:\/\/clang.llvm.org\/extra\/clangd\/Installation.html\" rel=\"nofollow\">Clangd<\/a> and ensure it's in <code>PATH<\/code> instead.<\/li>\n<\/ul><\/li>\n<\/ul><h3><a id=\"user-content-ccls\" class=\"anchor\" aria-hidden=\"true\" href=\"#ccls\"><\/a>CCLS<\/h3><p><a href=\"https:\/\/github.com\/MaskRay\/emacs-ccls\">Emacs CCLS<\/a>, a client for <a href=\"https:\/\/github.com\/MaskRay\/ccls\">CCLS<\/a>, a C\/C++\/Objective-C language server supporting multi-million line C++ code-bases, powered by libclang.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package ccls\n  :defer t\n  :if (not *sys\/win32*)\n  :hook ((c-mode c++-mode objc-mode) .\n         (lambda () (require 'ccls) (lsp)))\n  :custom\n  (ccls-executable (executable-find \"ccls\")) ; Add ccls to path if you haven't done so\n  (ccls-sem-highlight-method 'font-lock)\n  (ccls-enable-skipped-ranges nil)\n  :config\n  (lsp-register-client\n   (make-lsp-client\n    :new-connection (lsp-tramp-connection (cons ccls-executable ccls-args))\n    :major-modes '(c-mode c++-mode cuda-mode objc-mode)\n    :server-id 'ccls-remote\n    :multi-root nil\n    :remote? t\n    :notification-handlers\n    (lsp-ht (\"$ccls\/publishSkippedRanges\" #'ccls--publish-skipped-ranges)\n            (\"$ccls\/publishSemanticHighlight\" #'ccls--publish-semantic-highlight))\n    :initialization-options (lambda () ccls-initialization-options)\n    :library-folders-fn nil)))<\/pre><\/div><h3><a id=\"user-content-modern-c-font-lock\" class=\"anchor\" aria-hidden=\"true\" href=\"#modern-c-font-lock\"><\/a>Modern C++ Font Lock<\/h3><p><a href=\"https:\/\/github.com\/ludwigpacifici\/modern-cpp-font-lock\">Modern CPP Font Lock<\/a>, font-locking for \"Modern C++\".<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package modern-cpp-font-lock\n  :diminish t\n  :init (modern-c++-font-lock-global-mode t))<\/pre><\/div><h2><a id=\"user-content-golang\" class=\"anchor\" aria-hidden=\"true\" href=\"#golang\"><\/a>Golang<\/h2><p><a href=\"https:\/\/github.com\/dominikh\/go-mode.el\">Go Mode<\/a>, an Emacs mode for editing Golang code.<\/p><p><strong>Prerequisite<\/strong>: <a href=\"https:\/\/github.com\/golang\/go\/wiki\/gopls\">gopls<\/a> is suggested for Golang's LSP support.<\/p><div class=\"highlight highlight-source-shell\"><pre>go get golang.org\/x\/tools\/gopls@latest<\/pre><\/div><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package go-mode\n  :mode \"\\\\.go\\\\'\"\n  :hook (before-save . gofmt-before-save))<\/pre><\/div><h2><a id=\"user-content-python\" class=\"anchor\" aria-hidden=\"true\" href=\"#python\"><\/a>Python<\/h2><h3><a id=\"user-content-python-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#python-configuration\"><\/a>Python Configuration<\/h3><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package python-mode\n  :ensure nil\n  :after flycheck\n  :mode \"\\\\.py\\\\'\"\n  :custom\n  (python-indent-offset 4)\n  (flycheck-python-pycompile-executable \"python3\")\n  (python-shell-interpreter \"python3\"))<\/pre><\/div><h3><a id=\"user-content-lsp-python-ms\" class=\"anchor\" aria-hidden=\"true\" href=\"#lsp-python-ms\"><\/a>LSP Python MS<\/h3><p><a href=\"https:\/\/github.com\/andrew-christianson\/lsp-python-ms\">LSP Python MS<\/a>, a lsp-mode client leveraging <a href=\"https:\/\/github.com\/Microsoft\/python-language-server\">Microsoft's Python Language Server<\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lsp-python-ms\n  :after lsp-mode python\n  :if (or *python3* *python*)\n  :custom\n  (lsp-python-executable-cmd \"python3\"))<\/pre><\/div><h2><a id=\"user-content-ess\" class=\"anchor\" aria-hidden=\"true\" href=\"#ess\"><\/a>ESS<\/h2><p><a href=\"https:\/\/ess.r-project.org\/\" rel=\"nofollow\">Emacs Speaks Statistics<\/a>, short for ESS, it's designed to support editing of scripts and interaction with various statistical analysis programs such as R, S-Plus, SAS, Stata and OpenBUGS\/JAGS.<\/p><p><strong>Prerequisite<\/strong>: Install <a href=\"https:\/\/cran.r-project.org\/mirrors.html\" rel=\"nofollow\">R<\/a> to start using ESS with R.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package ess\n  :defer t\n  :commands R\n  :config\n  (load \"ess-autoloads\"))<\/pre><\/div><h2><a id=\"user-content-tex\" class=\"anchor\" aria-hidden=\"true\" href=\"#tex\"><\/a>TeX<\/h2><p><strong>Prerequisite<\/strong>: Please install <a href=\"https:\/\/www.tug.org\/texlive\/quickinstall.html\" rel=\"nofollow\">TeX Live<\/a>.<\/p><h3><a id=\"user-content-auctex\" class=\"anchor\" aria-hidden=\"true\" href=\"#auctex\"><\/a>AUCTeX<\/h3><p><a href=\"https:\/\/www.gnu.org\/software\/auctex\/\" rel=\"nofollow\">AUCTeX<\/a>, an extensible package for writing and formatting TeX files. It supports many different TeX macro packages, including AMS-TEX, LaTeX, Texinfo, ConTEXt, and docTEX (dtx files).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package tex\n  :ensure auctex\n  :defer t\n  :custom\n  (TeX-auto-save t)\n  (TeX-parse-self t)\n  (TeX-master nil)\n  ;; to use pdfview with auctex\n  (TeX-view-program-selection '((output-pdf \"pdf-tools\"))\n                              TeX-source-correlate-start-server t)\n  (TeX-view-program-list '((\"pdf-tools\" \"TeX-pdf-tools-sync-view\")))\n  (TeX-after-compilation-finished-functions #'TeX-revert-document-buffer)\n  :hook\n  (LaTeX-mode . (lambda ()\n                  (turn-on-reftex)\n                  (setq reftex-plug-into-AUCTeX t)\n                  (reftex-isearch-minor-mode)\n                  (setq TeX-PDF-mode t)\n                  (setq TeX-source-correlate-method 'synctex)\n                  (setq TeX-source-correlate-start-server t)))\n  :config\n  (when (version&lt; emacs-version \"26\")\n    (add-hook LaTeX-mode-hook #'display-line-numbers-mode)))<\/pre><\/div><h3><a id=\"user-content-org-edit-latex\" class=\"anchor\" aria-hidden=\"true\" href=\"#org-edit-latex\"><\/a>Org Edit LaTeX<\/h3><p><a href=\"https:\/\/github.com\/et2010\/org-edit-latex\">Org Edit LaTeX<\/a>, an extension to edit LaTeX fragment\/environment in an edit buffer, even to complete and preview LaTeX in the edit buffer.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package org-edit-latex\n  :defer t\n  :after org)<\/pre><\/div><p><strong>Prerequisite<\/strong>: Install <a href=\"https:\/\/nodejs.org\/en\/download\/\" rel=\"nofollow\">NodeJS<\/a> and ensure it's in <code>PATH<\/code>. Execute following commands to enable LSP for JavaScript\/TypeScript\/HTML:<\/p><div class=\"highlight highlight-source-shell\"><pre>npm i -g typescript\nnpm i -g typescript-language-server<\/pre><\/div><h2><a id=\"user-content-web-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#web-mode\"><\/a>Web Mode<\/h2><p><a href=\"https:\/\/github.com\/fxbois\/web-mode\">Web mode<\/a>, a major mode for editing web templates.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package web-mode\n  :custom-face\n  (css-selector ((t (:inherit default :foreground \"#66CCFF\"))))\n  (font-lock-comment-face ((t (:foreground \"#828282\"))))\n  :mode\n  (\"\\\\.phtml\\\\'\" \"\\\\.tpl\\\\.php\\\\'\" \"\\\\.[agj]sp\\\\'\" \"\\\\.as[cp]x\\\\'\"\n   \"\\\\.erb\\\\'\" \"\\\\.mustache\\\\'\" \"\\\\.djhtml\\\\'\" \"\\\\.[t]?html?\\\\'\"))<\/pre><\/div><h2><a id=\"user-content-javascripttypescript\" class=\"anchor\" aria-hidden=\"true\" href=\"#javascripttypescript\"><\/a>JavaScript\/TypeScript<\/h2><h3><a id=\"user-content-javascript2-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#javascript2-mode\"><\/a>JavaScript2 Mode<\/h3><p><a href=\"https:\/\/github.com\/mooz\/js2-mode\">JS2 mode<\/a>, a feature that offers improved JavsScript editing mode.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package js2-mode\n  :mode \"\\\\.js\\\\'\"\n  :interpreter \"node\")<\/pre><\/div><h3><a id=\"user-content-typescript-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#typescript-mode\"><\/a>TypeScript Mode<\/h3><p><a href=\"https:\/\/github.com\/emacs-typescript\/typescript.el\">TypeScript mode<\/a>, a feature that offers TypeScript support for Emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package typescript-mode\n  :mode \"\\\\.ts\\\\'\"\n  :commands (typescript-mode))<\/pre><\/div><h2><a id=\"user-content-emmet\" class=\"anchor\" aria-hidden=\"true\" href=\"#emmet\"><\/a>Emmet<\/h2><p><a href=\"https:\/\/github.com\/smihica\/emmet-mode\">Emmet<\/a>, a feature that allows writing HTML using CSS selectors along with <code>C-j<\/code>. See <a href=\"https:\/\/github.com\/smihica\/emmet-mode#usage\">usage<\/a> for more information.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package emmet-mode\n  :hook ((web-mode . emmet-mode)\n         (css-mode . emmet-mode)))<\/pre><\/div><h2><a id=\"user-content-instant-rename-tag\" class=\"anchor\" aria-hidden=\"true\" href=\"#instant-rename-tag\"><\/a>Instant Rename Tag<\/h2><p><a href=\"https:\/\/github.com\/manateelazycat\/instant-rename-tag\">Instant Rename Tag<\/a>, a plugin that provides ability to rename html tag pairs instantly.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package instant-rename-tag\n  :load-path (lambda () (expand-file-name \"site-elisp\/instant-rename-tag\" user-emacs-directory))\n  :bind (\"C-z &lt;\" . instant-rename-tag))<\/pre><\/div><h2><a id=\"user-content-json\" class=\"anchor\" aria-hidden=\"true\" href=\"#json\"><\/a>JSON<\/h2><p><a href=\"https:\/\/github.com\/joshwnj\/json-mode\">JSON Mode<\/a>, a major mode for editing JSON files.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package json-mode\n  :mode \"\\\\.json\\\\'\")<\/pre><\/div><h2><a id=\"user-content-org\" class=\"anchor\" aria-hidden=\"true\" href=\"#org\"><\/a>Org<\/h2><p><a href=\"https:\/\/orgmode.org\/\" rel=\"nofollow\">Org<\/a>, a Emacs built-in tool for keeping notes, maintaining TODO lists, planning projects, and authoring documents with a fast and effective plain-text system.<\/p><p><strong>Prerequisite<\/strong>: Configure <code>(org-agenda-files (list \"~\/org\/agenda\/\"))<\/code> to your agenda folder to use org-agenda.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package org\n  :ensure nil\n  :defer t\n  :bind\n  (\"C-c l\" . org-store-link)\n  (\"C-c a\" . org-agenda)\n  (\"C-c c\" . org-capture)\n  (\"C-c b\" . org-switch)\n  (:map org-mode-map (\"C-c C-p\" . org-export-as-pdf-and-open))\n  :custom\n  (org-log-done 'time)\n  (org-export-backends (quote (ascii html icalendar latex md odt)))\n  (org-use-speed-commands t)\n  (org-confirm-babel-evaluate 'nil)\n  (org-todo-keywords\n   '((sequence \"TODO\" \"IN-PROGRESS\" \"REVIEW\" \"|\" \"DONE\")))\n  (org-agenda-window-setup 'other-window)\n  :config\n  (unless (version&lt; org-version \"9.2\")\n    (require 'org-tempo))\n  (when (file-directory-p \"~\/org\/agenda\/\")\n    (setq org-agenda-files (list \"~\/org\/agenda\/\")))\n  (defun org-export-turn-on-syntax-highlight ()\n    \"Setup variables to turn on syntax highlighting when calling `org-latex-export-to-pdf'.\"\n    (interactive)\n    (setq org-latex-listings 'minted\n          org-latex-packages-alist '((\"\" \"minted\"))\n          org-latex-pdf-process\n          '(\"pdflatex -shelnl-escape -interaction nonstopmode -output-directory %o %f\"\n            \"pdflatex -shell-escape -interaction nonstopmode -output-directory %o %f\")))\n  (defun org-export-as-pdf-and-open ()\n    \"Run `org-latex-export-to-pdf', delete the tex file and open pdf in a new buffer.\"\n    (interactive)\n    (save-buffer)\n    (let* ((pdf-path (org-latex-export-to-pdf))\n           (pdf-name (file-name-nondirectory pdf-path)))\n      (if (try-completion pdf-name (mapcar #'buffer-name (buffer-list)))\n          (progn\n            (kill-matching-buffers (concat \"^\" pdf-name) t t)\n            (find-file-other-window pdf-name))\n        (find-file-other-window pdf-name))\n      (delete-file (concat (substring pdf-path 0 (string-match \"[^\\.]*\\\/?$\" pdf-path)) \"tex\")))))<\/pre><\/div><h3><a id=\"user-content-toc-org\" class=\"anchor\" aria-hidden=\"true\" href=\"#toc-org\"><\/a>TOC Org<\/h3><p><a href=\"https:\/\/github.com\/snosov1\/toc-org\">TOC Org<\/a> generates table of contents for <code>.org<\/code> files<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package toc-org\n  :hook (org-mode . toc-org-mode))<\/pre><\/div><h3><a id=\"user-content-htmlize\" class=\"anchor\" aria-hidden=\"true\" href=\"#htmlize\"><\/a>HTMLize<\/h3><p><a href=\"https:\/\/github.com\/hniksic\/emacs-htmlize\">HTMLize<\/a>, a tool that converts buffer text and decorations to HTML.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package htmlize :defer t)<\/pre><\/div><h3><a id=\"user-content-gfm-exporter\" class=\"anchor\" aria-hidden=\"true\" href=\"#gfm-exporter\"><\/a>GFM Exporter<\/h3><p><a href=\"https:\/\/github.com\/larstvei\/ox-gfm\">OX-GFM<\/a>, a Github Flavored Markdown exporter for Org Mode.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package ox-gfm :defer t)<\/pre><\/div><h3><a id=\"user-content-plantuml-and-graphviz\" class=\"anchor\" aria-hidden=\"true\" href=\"#plantuml-and-graphviz\"><\/a>PlantUML and Graphviz<\/h3><p><a href=\"https:\/\/github.com\/skuro\/plantuml-mode\">PlantUML Mode<\/a>, a major mode for editing PlantUML sources.<\/p><p><strong>Prerequisite<\/strong>:<\/p><ol><li>Install <a href=\"http:\/\/plantuml.com\/download\" rel=\"nofollow\">plantuml<\/a> and configure <code>(org-plantuml-jar-path (expand-file-name \"path\/to\/plantuml.jar\"))<\/code>.<\/li>\n<li>Install <a href=\"https:\/\/graphviz.gitlab.io\/download\/\" rel=\"nofollow\">Graphviz<\/a> on your system to support graph visualization. Execute <code>sudo apt install graphviz<\/code> in Ubuntu.<\/li>\n<\/ol><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package plantuml-mode\n  :defer t\n  :custom\n  (org-plantuml-jar-path (expand-file-name \"~\/tools\/plantuml\/plantuml.jar\"))\n  :config\n  (org-babel-do-load-languages\n   'org-babel-load-languages\n   '(;; other Babel languages\n     (plantuml . t))))<\/pre><\/div><h2><a id=\"user-content-eaf\" class=\"anchor\" aria-hidden=\"true\" href=\"#eaf\"><\/a>EAF<\/h2><p><a href=\"https:\/\/github.com\/manateelazycat\/emacs-application-framework\">Emacs Application Framework<\/a>, a development framework that integrates any PyQt program into Emacs.<\/p><p><strong>Prerequisite<\/strong>: Please install <code>python3<\/code> and <code>pip3<\/code>. Then execute below command:<\/p><div class=\"highlight highlight-source-shell\"><pre>sudo pip3 install dbus-python PyMuPDF grip qrcode pyqt5 python-xlib PyQtWebEngine<\/pre><\/div><p>Note that If you are using Debian\/Ubuntu, it is possible that <code>QtWebEngine<\/code> is <a href=\"https:\/\/marc.info\/?l=kde-core-devel&amp;m=142954900813235&amp;w=2\" rel=\"nofollow\">not working<\/a>. Install the following:<\/p><div class=\"highlight highlight-source-shell\"><pre>sudo apt-get install python3-pyqt5.qtwebengine python3-pyqt5.qtmultimedia<\/pre><\/div><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package eaf\n  :load-path (lambda () (expand-file-name \"site-elisp\/emacs-application-framework\" user-emacs-directory))\n  :if *eaf-env*\n  :custom\n  (eaf-find-alternate-file-in-dired t)\n  (browse-url-browser-function 'eaf-open-browser) ;; Make EAF Browser my default browser\n  :config\n  (defalias 'browse-web #'eaf-open-browser)\n  ;; I already bind \"RET\", \"&lt;mouse-2&gt;\", \"^\" to `dired-find-alternate-file' in `init-dired.el'.\n  ;; Comment this line out of you don't want to use EAF to open available files in dired.\n  ;; (global-set-key [remap dired-find-alternate-file] #'eaf-file-open-in-dired)\n  (eaf-bind-key scroll_up \"RET\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key scroll_down_page \"DEL\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key scroll_down_page \"u\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key scroll_up_page \"d\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key scroll_to_end \"M-&gt;\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key scroll_to_home \"M-&lt;\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key quit-window \"q\" eaf-pdf-viewer-keybinding)\n  (eaf-bind-key take_photo \"p\" eaf-camera-keybinding))<\/pre><\/div><h2><a id=\"user-content-erc\" class=\"anchor\" aria-hidden=\"true\" href=\"#erc\"><\/a>ERC<\/h2><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/ERC\" rel=\"nofollow\">Emacs Relay Chat<\/a>, a powerful, modular, and extensible <a href=\"http:\/\/www.irc.org\/\" rel=\"nofollow\">IRC<\/a> client for Emacs. It utilizes <a href=\"https:\/\/github.com\/leathekd\/erc-hl-nicks\">erc-hl-nicks<\/a> for nickname highlighting and <a href=\"https:\/\/github.com\/kidd\/erc-image.el\">erc-image<\/a> to fetch and show received images in ERC.<\/p><p><strong>Prerequisite<\/strong>: Put IRC credentials in the file <code>~\/.authinfo<\/code> and configure <code>erc-nick<\/code> to your IRC nickname.<\/p><pre>machine irc.freenode.net login &lt;nickname&gt; password &lt;password&gt; port 6697\n<\/pre><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package erc\n  :ensure nil\n  :init\n  (use-package erc-hl-nicks :defer t)\n  (use-package erc-image :defer t)\n  :custom-face\n  (erc-notice-face ((t (:foreground \"#ababab\"))))\n  :custom\n  (erc-autojoin-channels-alist '((\"freenode.net\" \"#emacs\")))\n  (erc-track-exclude-types '(\"NICK\" \"PART\" \"MODE\" \"324\" \"329\" \"332\" \"333\" \"353\" \"477\"))\n  (erc-hide-list '(\"JOIN\" \"PART\" \"QUIT\"))\n  (erc-lurker-hide-list '(\"JOIN\" \"PART\" \"QUIT\"))\n  (erc-server-coding-system '(utf-8 . utf-8))\n  (erc-interpret-mirc-color t)\n  (erc-kill-buffer-on-part t)\n  (erc-kill-queries-on-quit t)\n  (erc-kill-server-buffer-on-quit t)\n  (erc-autojoin-timing 'ident)\n  (erc-fill-function 'erc-fill-static)\n  (erc-fill-static-center 15)\n  (erc-lurker-threshold-time 43200)\n  (erc-server-reconnect-attempts 5)\n  (erc-server-reconnect-timeout 3)\n  (erc-prompt-for-password nil)\n  (erc-prompt-for-nickserv-password nil)\n  :config\n  ;; Prerequisite: Configure this to your IRC nickname\n  (defvar erc-nick \"\"\n    \"The nickname used to login into ERC\")\n  (add-to-list 'erc-modules 'notifications)\n  (erc-track-mode t)\n  (erc-services-mode 1)\n  (defun erc-start-or-switch ()\n    \"Start ERC or switch to ERC buffer if it has started already.\"\n    (interactive)\n    (if (get-buffer \"irc.freenode.net:6697\")\n        (erc-track-switch-buffer 1)\n      (erc-tls :server \"irc.freenode.net\" :port 6697 :nick erc-nick)))\n  (defun erc-count-users ()\n    \"Displays the number of users and ops connected on the current channel.\"\n    (interactive)\n    (if (get-buffer \"irc.freenode.net:6697\")\n        (let ((channel (erc-default-target)))\n          (if (and channel (erc-channel-p channel))\n              (let ((hash-table (with-current-buffer (erc-server-buffer)\n                                  erc-server-users))\n                    (users 0)\n                    (ops 0))\n                (maphash (lambda (k v)\n                           (when (member (current-buffer)\n                                         (erc-server-user-buffers v))\n                             (cl-incf users))\n                           (when (erc-channel-user-op-p k)\n                             (cl-incf ops)))\n                         hash-table)\n                (message \"%d users (%s ops) are online on %s\" users ops channel))\n            (user-error \"The current buffer is not a channel\")))\n      (user-error \"You must first be connected on IRC\")))\n  (defun erc-get-ops ()\n    \"Displays the names of ops users on the current channel.\"\n    (interactive)\n    (if (get-buffer \"irc.freenode.net:6697\")\n        (let ((channel (erc-default-target)))\n          (if (and channel (erc-channel-p channel))\n              (let (ops)\n                (maphash (lambda (nick cdata)\n                           (if (and (cdr cdata)\n                                    (erc-channel-user-op (cdr cdata)))\n                               (setq ops (cons nick ops))))\n                         erc-channel-users)\n                (if ops\n                    (message \"The online ops users are: %s\"  (mapconcat 'identity ops \" \"))\n                  (message \"There are no ops users online on %s\" channel)))\n            (user-error \"The current buffer is not a channel\")))\n      (user-error \"You must first be connected on IRC\")))\n  (defun erc-notify (nickname message)\n    \"Displays a notification message for ERC.\"\n    (let* ((channel (buffer-name))\n           (nick (erc-hl-nicks-trim-irc-nick nickname))\n           (title (if (string-match-p (concat \"^\" nickname) channel)\n                      nick\n                    (concat nick \" (\" channel \")\")))\n           (msg (s-trim (s-collapse-whitespace message))))\n      (alert (concat nick \": \" msg) :title title)))\n  :bind\n  (\"M-z i\" . erc-start-or-switch)\n  :hook\n  (ercn-notify . erc-notify))<\/pre><\/div><h2><a id=\"user-content-eww\" class=\"anchor\" aria-hidden=\"true\" href=\"#eww\"><\/a>EWW<\/h2><p>Emacs Web Wowser, the HTML-based Emacs Web Browser.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package eww\n  :ensure nil\n  :commands (eww)\n  :hook (eww-mode . (lambda ()\n                      \"Rename EWW's buffer so sites open in new page.\"\n                      (rename-buffer \"eww\" t)))\n  :config\n  ;; I am using EAF-Browser instead of EWW\n  (unless *eaf-env*\n    (setq browse-url-browser-function 'eww-browse-url))) ; Hit &amp; to browse url with system browser<\/pre><\/div><h2><a id=\"user-content-mu4e\" class=\"anchor\" aria-hidden=\"true\" href=\"#mu4e\"><\/a>MU4E<\/h2><p><a href=\"https:\/\/www.djcbsoftware.nl\/code\/mu\/mu4e.html\" rel=\"nofollow\">Mu4e<\/a>, a package that provides an emacs-based e-mail client which uses <a href=\"https:\/\/www.djcbsoftware.nl\/code\/mu\/\" rel=\"nofollow\">mu<\/a> as its backend. This mu4e configuration is tailored for Gmail.<\/p><p><strong>Prerequisite<\/strong>:<\/p><ol><li>\n<p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Isync\" rel=\"nofollow\">Configure<\/a> IMAP using <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Isync\" rel=\"nofollow\">mbsync<\/a>, put your <code>.mbsyncrc<\/code> config file in <code>~\/.emacs.d\/mu4e\/<\/code>. A <a href=\"https:\/\/gist.github.com\/MatthewZMD\/39cc00260486d17450f7228a4f36891f\">sample<\/a> is provided.<\/p>\n<\/li>\n<li>\n<p>Install <a href=\"https:\/\/www.djcbsoftware.nl\/code\/mu\/\" rel=\"nofollow\">mu<\/a>.<\/p>\n<\/li>\n<li>\n<p>Execute the follwing commands<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>mbsync -c ~\/.emacs.d\/mu4e\/.mbsyncrc -Dmn gmail\nmbsync -c ~\/.emacs.d\/mu4e\/.mbsyncrc -a\nmu index --maildir=~\/Maildir\/<\/pre><\/div>\n<ul><li>If you are getting <code>Invalid Credentials<\/code> error and you are sure the password is correct, check <a href=\"https:\/\/appuals.com\/fix-your-imap-server-wants-to-alert-you-invalid-credentials\/\" rel=\"nofollow\">this<\/a> link.<\/li>\n<\/ul><\/li>\n<\/ol><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package mu4e\n  :ensure nil\n  :commands (mu4e)\n  :init\n  (use-package mu4e-alert\n    :defer t\n    :config\n    (when (executable-find \"notify-send\")\n      (mu4e-alert-set-default-style 'libnotify))\n    :hook\n    ((after-init . mu4e-alert-enable-notifications)\n     (after-init . mu4e-alert-enable-mode-line-display)))\n  (use-package mu4e-overview :defer t)\n  :bind (\"M-z m\" . mu4e)\n  :custom\n  (mu4e-maildir (expand-file-name \"~\/Maildir\"))\n  (mu4e-get-mail-command \"mbsync -c ~\/.emacs.d\/mu4e\/.mbsyncrc -a\")\n  (mu4e-view-prefer-html t)\n  (mu4e-update-interval 180)\n  (mu4e-headers-auto-update t)\n  (mu4e-compose-signature-auto-include nil)\n  (mu4e-compose-format-flowed t)\n  (mu4e-view-show-images t)\n  (mu4e-sent-messages-behavior 'delete)\n  (mu4e-change-filenames-when-moving t) ; work better for mbsync\n  (mu4e-attachment-dir \"~\/Downloads\")\n  (message-kill-buffer-on-exit t)\n  (mu4e-compose-dont-reply-to-self t)\n  (mu4e-view-show-addresses t)\n  (mu4e-confirm-quit nil)\n  (mu4e-use-fancy-chars t)\n  :hook\n  ((mu4e-view-mode . visual-line-mode)\n   (mu4e-compose-mode . (lambda ()\n                          (visual-line-mode)\n                          (use-hard-newlines -1)\n                          (flyspell-mode)))\n   (mu4e-view-mode . (lambda() ;; try to emulate some of the eww key-bindings\n                       (local-set-key (kbd \"&lt;tab&gt;\") 'shr-next-link)\n                       (local-set-key (kbd \"&lt;backtab&gt;\") 'shr-previous-link)))\n   (mu4e-headers-mode . (lambda ()\n                          (interactive)\n                          (setq mu4e-headers-fields\n                                `((:human-date . 25) ;; alternatively, use :date\n                                  (:flags . 6)\n                                  (:from . 22)\n                                  (:thread-subject . ,(- (window-body-width) 70)) ;; alternatively, use :subject\n                                  (:size . 7))))))\n  :config\n  (add-to-list 'mu4e-view-actions\n               '(\"ViewInBrowser\" . mu4e-action-view-in-browser) t)\n  (setq mu4e-contexts\n        (list\n         (make-mu4e-context\n          :name \"gmail\"\n          :enter-func (lambda () (mu4e-message \"Entering context gmail\"))\n          :leave-func (lambda () (mu4e-message \"Leaving context gmail\"))\n          :match-func\n          (lambda (msg)\n            (when msg\n              (mu4e-message-contact-field-matches\n               msg '(:from :to :cc :bcc) user-mail-address))) ; Set to your email address\n          :vars '((mu4e-refile-folder \"\/gmail\/Archive\")\n                  (mu4e-sent-folder . \"\/gmail\/[email].Sent Mail\")\n                  (mu4e-drafts-folder . \"\/gmail\/[email].Drafts\")\n                  (mu4e-trash-folder . \"\/gmail\/[email].Trash\")\n                  (mu4e-compose-signature . user-full-name)\n                  (mu4e-compose-format-flowed . t)\n                  (smtpmail-queue-dir . \"~\/Maildir\/gmail\/queue\/cur\")\n                  (message-send-mail-function . smtpmail-send-it)\n                  (smtpmail-smtp-user . \"matthewzmd\") ; Set to your username\n                  (smtpmail-starttls-credentials . ((\"smtp.gmail.com\" 587 nil nil)))\n                  (smtpmail-auth-credentials . (expand-file-name \"~\/.authinfo.gpg\"))\n                  (smtpmail-default-smtp-server . \"smtp.gmail.com\")\n                  (smtpmail-smtp-server . \"smtp.gmail.com\")\n                  (smtpmail-smtp-service . 587)\n                  (smtpmail-debug-info . t)\n                  (smtpmail-debug-verbose . t)\n                  (mu4e-maildir-shortcuts . ( (\"\/gmail\/INBOX\"            . ?i)\n                                              (\"\/gmail\/[email].Sent Mail\" . ?s)\n                                              (\"\/gmail\/[email].Trash\"       . ?t)\n                                              (\"\/gmail\/[email].All Mail\"  . ?a)\n                                              (\"\/gmail\/[email].Starred\"   . ?r)\n                                              (\"\/gmail\/[email].Drafts\"    . ?d))))))))<\/pre><\/div><h2><a id=\"user-content-tramp\" class=\"anchor\" aria-hidden=\"true\" href=\"#tramp\"><\/a>Tramp<\/h2><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/TrampMode\" rel=\"nofollow\">Tramp<\/a>, short for Transparent Remote Access, Multiple Protocols is a package for editing remote files using a remote shell connection (rlogin, telnet, ssh).<\/p><h3><a id=\"user-content-google-cloud-platform\" class=\"anchor\" aria-hidden=\"true\" href=\"#google-cloud-platform\"><\/a>Google Cloud Platform<\/h3><p>Connect to Google Cloud Platform using the following:<\/p><pre>\/gssh:some-instance:\/path\/to\/file\n<\/pre><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package tramp\n  :ensure nil\n  :defer 1\n  :config\n  (add-to-list 'tramp-remote-path 'tramp-own-remote-path)\n  ;; TRAMP gcloud ssh\n  (add-to-list 'tramp-methods\n               '(\"gssh\"\n                 (tramp-login-program        \"gcloud compute ssh\")\n                 (tramp-login-args           ((\"%h\")))\n                 (tramp-async-args           ((\"-q\")))\n                 (tramp-remote-shell         \"\/bin\/bash\")\n                 (tramp-remote-shell-args    (\"-c\"))\n                 (tramp-gw-args              ((\"-o\" \"GlobalKnownHostsFile=\/dev\/null\")\n                                              (\"-o\" \"UserKnownHostsFile=\/dev\/null\")\n                                              (\"-o\" \"StrictHostKeyChecking=no\")))\n                 (tramp-default-port         22))))<\/pre><\/div><h2><a id=\"user-content-pdf-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#pdf-tools\"><\/a>PDF Tools<\/h2><p><a href=\"https:\/\/github.com\/politza\/pdf-tools\">PDF Tools<\/a>, an Emacs support library for PDF files. It works best on non-Windows OS.<\/p><p><strong>Note<\/strong>: You need <a href=\"https:\/\/linux.die.net\/man\/1\/convert\" rel=\"nofollow\">convert<\/a> provided from imagemagick to <em>Pick a Link and Jump<\/em> with F.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package pdf-tools-install\n  :ensure pdf-tools\n  :if (and *sys\/gui* (not *sys\/win32*))\n  :mode \"\\\\.pdf\\\\'\"\n  :commands (pdf-loader-install)\n  :custom\n  (TeX-view-program-selection '((output-pdf \"pdf-tools\")))\n  (TeX-view-program-list '((\"pdf-tools\" \"TeX-pdf-tools-sync-view\")))\n  :hook\n  (pdf-view-mode . (lambda () (display-line-numbers-mode -1)))\n  :config\n  (pdf-loader-install))<\/pre><\/div><h2><a id=\"user-content-leetcode\" class=\"anchor\" aria-hidden=\"true\" href=\"#leetcode\"><\/a>LeetCode<\/h2><p><a href=\"https:\/\/github.com\/kaiwk\/leetcode.el\">LeetCode<\/a>, an Emacs LeetCode client. Note that this package is dependant on <a href=\"https:\/\/github.com\/skeeto\/emacs-aio\">aio<\/a> and <a href=\"https:\/\/github.com\/davazp\/graphql-mode\">GraphQL<\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package leetcode\n  :load-path (lambda () (expand-file-name \"site-elisp\/leetcode.el\" user-emacs-directory))\n  :commands (leetcode)\n  :init\n  (use-package graphql :defer t)\n  (use-package aio :defer t)\n  :custom\n  (url-debug t)\n  (leetcode-prefer-language \"python3\"))<\/pre><\/div><h2><a id=\"user-content-pyim\" class=\"anchor\" aria-hidden=\"true\" href=\"#pyim\"><\/a>Pyim<\/h2><p><a href=\"https:\/\/github.com\/tumashu\/pyim\">Pyim<\/a>, an Emacs Chinese Pinyin Input. It uses <a href=\"https:\/\/github.com\/tumashu\/posframe\">posframe<\/a> package to display candidates.<\/p><p>\u6211\u5df2\u7ecf\u505c\u6b62\u4f7f\u7528\u4f5c\u8005\u63a8\u8350\u7684\u65e0\u75db\u4e2d\u82f1\u5207\u6362\uff0c\u5b83\u5bf9\u9700\u8981\u540c\u65f6\u6253\u82f1\u6587\u548c\u4e2d\u6587\u7684\u60c5\u51b5\u4e0d\u662f\u5f88\u53cb\u597d\u3002\u5982\u9700\u5207\u6362\u8f93\u5165\u6cd5\uff0c\u8bf7\u5584\u7528 <code>C-\\<\/code> \u3002<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package pyim\n  :init\n  (use-package posframe :defer t)\n  :custom\n  (default-input-method \"pyim\")\n  (pyim-default-scheme 'quanpin)\n  (pyim-page-tooltip 'posframe)\n  (pyim-page-length 9)\n  :config\n  (pyim-isearch-mode 1)\n  (setq-default pyim-english-input-switch-functions\n                '(pyim-probe-isearch-mode\n                  pyim-probe-org-structure-template))\n  (setq-default pyim-punctuation-half-width-functions\n                '(pyim-probe-punctuation-line-beginning\n                  pyim-probe-punctuation-after-punctuation))\n  :bind\n  (\"M-j\" . pyim-convert-string-at-point)) ; M-j \u5f3a\u5236\u5c06\u5149\u6807\u524d\u7684\u62fc\u97f3\u5b57\u7b26\u4e32\u8f6c\u6362\u4e3a\u4e2d\u6587\u3002<\/pre><\/div><h3><a id=\"user-content-pyim-basedict\" class=\"anchor\" aria-hidden=\"true\" href=\"#pyim-basedict\"><\/a>Pyim BaseDict<\/h3><p><a href=\"https:\/\/github.com\/tumashu\/pyim-basedict\">Pyim BaseDict<\/a>, the default Chinese-Pyim dictionary.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package pyim-basedict\n  :after pyim\n  :config (pyim-basedict-enable))<\/pre><\/div><h2><a id=\"user-content-epaint\" class=\"anchor\" aria-hidden=\"true\" href=\"#epaint\"><\/a>EPaint<\/h2><p><a href=\"https:\/\/github.com\/chuntaro\/epaint\">EPaint<\/a>, a simple paint tool for emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package epaint\n  :if *sys\/gui*\n  :load-path (lambda () (expand-file-name \"site-elisp\/epaint\" user-emacs-directory))\n  :commands (epaint)\n  :init\n  (with-eval-after-load (quote epaint-context)\n    (unless (boundp (quote cl-struct-epaint-drawable))\n      (defvar cl-struct-epaint-drawable (quote epaint-drawable)))\n    (unless (boundp (quote cl-struct-epaint-gc))\n      (defvar cl-struct-epaint-gc (quote epaint-gc)))))<\/pre><\/div><h2><a id=\"user-content-tetris\" class=\"anchor\" aria-hidden=\"true\" href=\"#tetris\"><\/a>Tetris<\/h2><p>Although <a href=\"https:\/\/www.emacswiki.org\/emacs\/TetrisMode\" rel=\"nofollow\">Tetris<\/a> is part of Emacs, but there still could be some configurations.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package tetris\n  :ensure nil\n  :commands (tetris)\n  :bind\n  (:map tetris-mode-map\n        (\"C-p\" . tetris-rotate-prev)\n        (\"C-n\" . tetris-rotate-down)\n        (\"C-b\" . tetris-move-left)\n        (\"C-f\" . tetris-move-right)\n        (\"C-SPC\" . tetris-move-bottom))\n  :config\n  (defadvice tetris-end-game (around zap-scores activate)\n    (save-window-excursion ad-do-it)))<\/pre><\/div><h2><a id=\"user-content-speed-type\" class=\"anchor\" aria-hidden=\"true\" href=\"#speed-type\"><\/a>Speed Type<\/h2><p><a href=\"https:\/\/github.com\/hagleitn\/speed-type\">Speed type<\/a>, a game to practice touch\/speed typing in Emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package speed-type\n  :commands (speed-type-text))<\/pre><\/div><h2><a id=\"user-content-2048-game\" class=\"anchor\" aria-hidden=\"true\" href=\"#2048-game\"><\/a>2048 Game<\/h2><p><a href=\"https:\/\/bitbucket.org\/zck\/2048.el\" rel=\"nofollow\">2048 Game<\/a>, an implementation of 2048 in Emacs.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package 2048-game\n  :commands (2048-game))<\/pre><\/div><h2><a id=\"user-content-zone\" class=\"anchor\" aria-hidden=\"true\" href=\"#zone\"><\/a>Zone<\/h2><p><a href=\"https:\/\/www.emacswiki.org\/emacs\/ZoneMode\" rel=\"nofollow\">Zone<\/a>, a minor-mode 'zones' Emacs out, choosing one of its random modes to obfuscate the current buffer.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package zone\n  :ensure nil\n  :defer 5\n  :config\n  ;; (zone-when-idle 600) ; in seconds\n  (defun zone-choose (pgm)\n    \"Choose a PGM to run for `zone'.\"\n    (interactive\n     (list\n      (completing-read\n       \"Program: \"\n       (mapcar 'symbol-name zone-programs))))\n    (let ((zone-programs (list (intern pgm))))\n      (zone))))<\/pre><\/div>","created_at":"2020-01-01T17:39:12+0000","updated_at":"2020-01-01T17:39:12+0000","published_by":["MatthewZMD"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":61,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/12422335?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","clojurescript","reagent","server-side-rendering"],"is_public":false,"id":1078,"title":"Rendering Reagent on the Server using Hiccup","url":"https:\/\/yogthos.net\/posts\/2015-11-24-Serverside-Reagent.html","content":"<p>I recently watched a great talk titled <a href=\"https:\/\/www.youtube.com\/watch?v=fICC26GGBpg\">Optimizing ClojureScript Apps For Speed<\/a>, where Allen Rohner discusses the benefits and hurdles of server-side rendering.<\/p><p>React supports hooking in to server generated HTML at runtime. However, since React is a JavaScript library it becomes problematic to leverage this functionality from Clojure. While the JVM provides a Js runtime with Nashorn, it's extremely slow and requires a lot of twiddling to work for even basic examples.<\/p><p>Another approach is to run an instance of Node.js and farm out React rendering to it. This avoids the limitations of Nashorn, but introduces a host of new problems described in the talk.<\/p><p>Allen then proposes an alternative approach where he implements parts of the Om API and cross-compiles the components that way. You can see how this works in his <a href=\"https:\/\/github.com\/arohner\/foam\">Foam<\/a> library.<\/p><p>The main difficulty identified in the talk is in implementing a sufficient amount of Om API in order to generate HTML on the server.<\/p><p>This got me thinking about what it would take to leverage this approach using Reagent. Unlike Om, Reagent has a tiny API and the only part of it used to create components is the Reagent atom implementation. The components themselves are written using plain Hiccup syntax.<\/p><p>Let's see how this could work. We'll start by creating a new Reagent project:<\/p><pre>lein new reagent reagent-serverside\n<\/pre><p>Next, we'll add a new namespace in called <code>reagent-serverside.home<\/code> <code>src\/cljc\/reagent_serverside\/home.cljc<\/code>. This namespace will house the home page component that we'll pre-render on the server.<\/p><p>All we have to do now is to use a reader conditional to only require the Reagent atom during ClojureScript compilation:<\/p><pre class=\"clojure\">(ns reagent-serverside.home\n  #?(:cljs\n     (:require [reagent.core :as reagent :refer [atom]])))\n<\/pre><p>We can now write our components as we would normally:<\/p><pre class=\"clojure\">(ns reagent-serverside.home\n  #?(:cljs\n     (:require [reagent.core :as reagent :refer [atom]])))\n(def items (atom nil))\n(defn item-list [items]\n  [:ul\n   (for [item items]\n     ^{:key item}\n     [:li item])])\n(defn add-item-button [items]\n  [:button\n   {:on-click #(swap! items conj (count @items))}\n   \"add item\"])\n(defn home-page []\n  [:div [:h2 \"Welcome to reagent-serverside\"]\n   [add-item-button items]\n   [item-list @items]])\n<\/pre><p>We'll have the <code>items<\/code> atom to house a collection of items, an <code>item-list<\/code> function to render it, and the <code>home-page<\/code> function that will use the <code>item-list<\/code> component. We also have a button that lets the user add new items with an <code>:on-click<\/code> event. This is all standard Reagent code.<\/p><h3 id=\"rendering_on_the_server\">Rendering on the Server<\/h3><p>Now, let's navigate to the <code>reagent-serverside.handler<\/code> namespace and reference the <code>reagent-serverside.home<\/code> we just created.<\/p><pre class=\"clojure\">(ns reagent-serverside.handler\n  (:require ...\n            [reagent-serverside.home :refer [items home-page]]))\n<\/pre><p>We'll now have to write the functions that will traverse the components and render them as appropriate. We'll attach a <code>:data-reactid<\/code> key to each one to give it an identifier that React looks for, and inject the result into our Hiccup markup.<\/p><pre class=\"clojure\">(defn react-id-str [react-id]\n  (assert (vector? react-id))\n  (str \".\" (clojure.string\/join \".\" react-id)))\n(defn set-react-id [react-id element]\n  (update element 1 merge {:data-reactid (react-id-str react-id)}))\n(defn normalize [component]\n  (if (map? (second component))\n    component\n    (into [(first component) {}] (rest component))))\n(defn render\n  ([component] (render [0] component))  \n  ([id component]\n   (cond\n    (fn? component)\n    (render (component))    \n    (not (coll? component))\n    component\n    (coll? (first component))\n    (map-indexed #(render (conj id %1) %2) component)\n    (keyword? (first component))\n    (let [[tag opts &amp; body] (normalize component)]\n      (-&gt;&gt; body\n           (map-indexed #(render (conj id %1) %2))\n           (into [tag opts])\n           (set-react-id id)))\n    (fn? (first component))\n    (render id (apply (first component) (rest component))))))\n(reset! items (range 10))\n(def mount-target\n  [:div#app (render home-page)])\n<\/pre><p>The <code>render<\/code> function will recursively walk the components evaluating any functions it finds and assigning the React id to each element.<\/p><p>Next, we'll set the <code>items<\/code> atom to a range of numbers, and then call <code>render<\/code> inside the <code>mount-target<\/code> to generate the markup.<\/p><h3 id=\"rendering_on_the_client\">Rendering on the Client<\/h3><p>Finally, let's navigate to the <code>reagent-serverside.core<\/code> namespace in the <code>src\/cljs<\/code> source path. We'll update it to reference the <code>home<\/code> namespace we created and render the <code>home-page<\/code> component on load.<\/p><pre class=\"clojure\">(ns reagent-serverside.core\n    (:require [reagent.core :as reagent :refer [atom]]\n              [reagent-serverside.home :refer [items home-page]]))\n(defn mount-root []\n  (reagent\/render [home-page] (.getElementById js\/document \"app\")))\n(defn init! []\n  (reset! items (range 20))\n  (mount-root))\n<\/pre><p>When we load the page we'll immediately see the server generated markup and then it will be updated by Reagent when ClojureScript is loaded. There are a few caveats here that you should be aware of.<\/p><p>Any components you wish to render on the server have to be written in <code>cljc<\/code>, so you may end up having to add some shims for things like Ajax calls.<\/p><p>The component syntax has to work with both Reagent and Hiccup, so you have to be mindful to use the common subset.<\/p><p>React is fairly picky about the structure and the <code>data-reactid<\/code> tags. So, it can be tricky to generate a DOM tree that it likes. The example in the post will give a React warning about the DOM being different. Some more work is needed around this.<\/p><p>However, even in the case that React doesn't reuse the DOM, the user will see the page immediately and you'll get the benefits of SEO for your site.<\/p><p>Full source is available on <a href=\"https:\/\/github.com\/yogthos\/reagent-serverside\">GitHub<\/a>.<\/p><h3 id=\"conclusions\">Conclusions<\/h3><p>Overall, I'm very happy with the results and it looks like it would be fairly easy to wrap this up into a library. The data focused approach is a huge win for Reagent here in my opinion. Since the components are laid out using regular Clojure data structures there's no need to implement any special API and things just work out of the box.<\/p>","created_at":"2019-12-31T18:58:53+0000","updated_at":"2019-12-31T18:59:05+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"yogthos.net","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","clojurescript","reagent","server-side-rendering"],"is_public":false,"id":1077,"title":"reagent-project\/reagent-cookbook","url":"https:\/\/github.com\/reagent-project\/reagent-cookbook\/tree\/master\/recipes\/reagent-server-rendering","content":"<h2><a id=\"user-content-problem\" class=\"anchor\" aria-hidden=\"true\" href=\"#problem\"><\/a>Problem<\/h2><p>You want to Reagent to generate HTML on the server and send a rendered page to the browser.<\/p><h2><a id=\"user-content-solution\" class=\"anchor\" aria-hidden=\"true\" href=\"#solution\"><\/a>Solution<\/h2><p>We're going to follow <a href=\"https:\/\/carouselapps.com\/2015\/09\/11\/isomorphic-clojurescriptjavascript-for-pre-rendering-single-page-applications-part-1\/\" rel=\"nofollow\">this example<\/a> using Nashorn to render Reagent on the server.<\/p><p><em>Steps<\/em><\/p><ol><li>Create a new project<\/li>\n<li>Add the necessary dependencies and plugins to <code>project.clj<\/code><\/li>\n<li>Add a Clojure namespace for the server functions<\/li>\n<li>Add the functions to create the Nashorn JavaScript engine and render Reagent components<\/li>\n<li>Create a function to render the HTML page using Hiccup<\/li>\n<li>Define server routes for the application<\/li>\n<li>Modify the ClojureScript namespace to provide functions to render the pages on the server and the client<\/li>\n<\/ol><h3><a id=\"user-content-step-1-create-a-new-project\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-1-create-a-new-project\"><\/a>Step 1: Create a new project<\/h3><pre>lein new rc reagent-server-rendering\n<\/pre><h3><a id=\"user-content-step-2-add-the-necessary-dependencies-and-plugins-to-projectclj\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-2-add-the-necessary-dependencies-and-plugins-to-projectclj\"><\/a>Step 2: Add the necessary dependencies and plugins to <code>project.clj<\/code><\/h3><div class=\"highlight highlight-source-clojure\"><pre>:dependencies [[org.clojure\/clojure \"1.7.0\"]\n                 [org.clojure\/clojurescript \"1.7.122\"]\n                 [reagent \"0.5.1\"]\n                 ;;Ring is needed to start an HTTP server\n                 [ring \"1.4.0\"]\n                 ;;provides default middleware for Ring\n                 [ring\/ring-defaults \"0.1.5\"]\n                 ;;server-side HTML templating\n                 [hiccup \"1.0.5\"]\n                 ;;thread pool management\n                 [aleph \"0.4.0\"]]\n  :plugins [[lein-cljsbuild \"1.0.6\"]\n            ;;plugin for starting the HTTP server\n            [lein-ring \"0.9.6\"]]\n  ;;specifies the HTTP handler\n  :ring {:handler reagent-server-rendering.handler\/app}<\/pre><\/div><h3><a id=\"user-content-step-3-add-a-clojure-namespace-for-the-server-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-3-add-a-clojure-namespace-for-the-server-functions\"><\/a>Step 3: Add a Clojure namespace for the server functions<\/h3><p>We need to create a Clojure namespace that will house the functions needed to start the JavaScript engine and render the HTML. We'll create a new file <code>src\/clj\/reagent_server_rendering\/handler.clj<\/code>. The namespace will have the following references in its declaration.<\/p><div class=\"highlight highlight-source-clojure\"><pre>(ns reagent-server-rendering.handler\n  (:require [aleph.flow :as flow]\n            [clojure.java.io :as io]\n            [hiccup.core :refer [html]]\n            [hiccup.page :refer [include-js include-css]]\n            [compojure.core :refer [GET defroutes]]\n            [compojure.route :refer [not-found resources]]\n            [ring.middleware.defaults :refer [site-defaults wrap-defaults]])\n  (:import [io.aleph.dirigiste Pools]\n           [javax.script ScriptEngineManager Invocable]))<\/pre><\/div><h3><a id=\"user-content-step-4-add-the-functions-to-create-the-nashorn-javascript-engine-and-render-reagent-components\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-4-add-the-functions-to-create-the-nashorn-javascript-engine-and-render-reagent-components\"><\/a>Step 4: Add the functions to create the Nashorn JavaScript engine and render Reagent components<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(defn- create-js-engine []\n  (doto (.getEngineByName (ScriptEngineManager.) \"nashorn\")\n    ; React requires either \"window\" or \"global\" to be defined.\n    (.eval \"var global = this\")\n    (.eval (-&gt; \"public\/js\/compiled\/app.js\"\n               io\/resource\n               io\/reader))))\n; We have one and only one key in the pool, it's a constant.\n(def ^:private js-engine-key \"js-engine\")\n(def ^:private js-engine-pool\n  (flow\/instrumented-pool\n    {:generate   (fn [_] (create-js-engine))\n     :controller (Pools\/utilizationController 0.9 10000 10000)}))\n(defn- render-page [page-id]\n  (let [js-engine @(flow\/acquire js-engine-pool js-engine-key)]\n    (try (.invokeMethod\n           ^Invocable js-engine\n           (.eval js-engine \"reagent_server_rendering.core\")\n           \"render_page\" (object-array [page-id]))\n         (finally (flow\/release js-engine-pool js-engine-key js-engine)))))<\/pre><\/div><p>The <code>create-js-engine<\/code> function will setup a new instance of the engine that will run the compiled JavaScript produced by compiling ClojureScript.<\/p><p>The <code>js-engine-pool<\/code> keeps a pool of running engines.<\/p><p>The <code>render-page<\/code> function will acquire an engine from the pool and call the <code>render-page<\/code> function from the <code>reagent-server-rendering.core<\/code> ClojureScript namespace. It will pass this function the page id as the argument. Once the page is rendered the JavaScript engine is released back into the pool.<\/p><h3><a id=\"user-content-step-5-create-a-function-to-render-the-html-page-using-hiccup\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-5-create-a-function-to-render-the-html-page-using-hiccup\"><\/a>Step 5: Create a function to render the HTML page using Hiccup<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(defn page [page-id]\n  (html\n   [:html\n    [:head\n     [:meta {:charset \"utf-8\"}]\n     [:meta {:name \"viewport\"\n             :content \"width=device-width, initial-scale=1\"}]\n     (include-css \"css\/site.css\")]\n    [:body\n     [:div#app\n      (render-page page-id)]\n     (include-js \"js\/compiled\/app.js\")\n     [:script {:type \"text\/javascript\"}\n      (str \"reagent_server_rendering.core.main('\" page-id \"');\")]]]))<\/pre><\/div><p>The <code>page<\/code> function uses Hiccup to generate a static page and inject the generated HTML produced by the <code>render-page<\/code> function into it.<\/p><h3><a id=\"user-content-step-6-define-server-routes-for-the-application\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-6-define-server-routes-for-the-application\"><\/a>Step 6: Define server routes for the application<\/h3><p>We'll now define routes for <code>\/<\/code> and <code>\/about<\/code> pages. Each route will pass a unique page id to the <code>page<\/code> function:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(defroutes app-routes\n  (GET \"\/\" [] (page \"home\"))\n  (GET \"\/about\" [] (page \"about\"))\n  (resources \"\/\")\n  (not-found \"Not Found\"))\n(def app (wrap-defaults app-routes site-defaults))<\/pre><\/div><h3><a id=\"user-content-step-7-modify-the-clojurescript-namespace-to-provide-functions-to-render-the-pages-on-the-server-and-the-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-7-modify-the-clojurescript-namespace-to-provide-functions-to-render-the-pages-on-the-server-and-the-client\"><\/a>Step 7: Modify the ClojureScript namespace to provide functions to render the pages on the server and the client<\/h3><div class=\"highlight highlight-source-clojure\"><pre>(ns reagent-server-rendering.core\n    (:require [reagent.core :as reagent]))\n(defn home-page []\n  [:div [:h2 \"Welcome to reagent-server-rendering\"]\n   [:div [:a {:href \"\/about\"} \"go to about page\"]]])\n(defn about-page []\n  [:div [:h2 \"About reagent-server-rendering\"]\n   [:div [:a {:href \"\/\"} \"go to the home page\"]]])\n(def pages\n  {\"home\"  home-page\n   \"about\" about-page})\n(defn ^:export render-page [page-id]\n  (reagent\/render-to-string [(get pages page-id)]))\n(defn ^:export main [page-id]\n  (reagent\/render [(get pages page-id)] (.getElementById js\/document \"app\")))<\/pre><\/div><p>The <code>render-page<\/code> function will render the HTML string given the page id that will be served from the server. The <code>main<\/code> function will render Reagent components at runtime when ClojureScript is loaded in the browser.<\/p><div class=\"highlight highlight-source-shell\"><pre># watch files for changes and rebuild...\nlein cljsbuild auto\n# ...or only build once\nlein cljsbuild once\n# start the server (in another terminal if you watch)\nlein ring server-headless<\/pre><\/div><p>Open <a href=\"http:\/\/localhost:3000\" rel=\"nofollow\">localhost:3000<\/a> to see the page rendered.<\/p><p>If you're seeing an error like this: <code>java.lang.IllegalArgumentException: No implementation of method: :make-reader of protocol: #'clojure.java.io\/IOFactory found for class: nil<\/code>, you probably forgot to build cljs, or there was an error building it.<\/p>","created_at":"2019-12-31T18:37:49+0000","updated_at":"2019-12-31T18:38:00+0000","published_by":["reagent-project"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/9254615?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["infrastructure","load-balancing","optimization"],"is_public":false,"id":1076,"title":"Arkwrite","url":"https:\/\/arkwright.github.io\/scaling-react-server-side-rendering.html","content":"<p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/scaling-react.svg\" width=\"250\" alt=\"Giant stack of React service instances in the shape of a shoddy pyramid, with the top one falling off, possibly to crush a tiny developer below. 8\/10 pretty good metaphor.\" referrerpolicy=\"no-referrer\" \/><\/p><p>I had the opportunity to work on scaling a React rendering service, adapting a fixed hardware provision to deal with increasing load. Over the course of many months, incremental improvements were made to the system to enable it to cope with demand. I thought it might be useful to share the more interesting insights that I gained during this process.<\/p><p>Some of the insights here are React-specific, but many are simply generic scalability challenges, or simple mistakes that were made. React server-side performance optimization has been covered elsewhere, so I\u2019m not going to provide an overview of React performance, generally. I\u2019m going to focus on the \u201cbig wins\u201d that we enjoyed, along with the subtle, fascinating <a href=\"https:\/\/en.wiktionary.org\/wiki\/footgun\">footguns<\/a>. My hope is that I can give you something interesting to think about, beyond the standard advice of setting <code class=\"highlighter-rouge\">NODE_ENV=production<\/code>. Something based on the real, honest-to-goodness challenges we had to overcome.<\/p><p>What I found so interesting about this project was where the investigative trail led. I assumed that improving React server-side performance would boil down to correctly implementing a number of React-specific best practices. Only later did I realize that I was looking for performance in the wrong places. With any luck, these stories will enable you to diagnose or avoid your own performance pitfalls!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/pitfall.svg\" width=\"280\" alt=\"Stick figure swinging on a rope over a bottomless pit, towards a shimmering React logo. Remember Pitfall?\" referrerpolicy=\"no-referrer\" \/><\/p><ul><li><a href=\"#scaling-react-server-side-rendering\">Introduction<\/a><\/li>\n<li><a href=\"#the-situation\">The Situation<\/a><\/li>\n<li><a href=\"#load-balancing\">Load Balancing<\/a>\n<ul><li><a href=\"#i-got-99-percentiles\">I Got 99 Percentiles<\/a><\/li>\n<li><a href=\"#seasonality\">Seasonality<\/a><\/li>\n<li><a href=\"#randomness\">Randomness<\/a><\/li>\n<li><a href=\"#load-balancing-strategies\">Load Balancing Strategies<\/a><\/li>\n<li><a href=\"#load-shedding-with-random-retries\">Load Shedding With Random Retries<\/a><\/li>\n<li><a href=\"#round-robin\">Round-Robin<\/a><\/li>\n<li><a href=\"#join-shortest-queue\">Join-Shortest-Queue<\/a><\/li>\n<li><a href=\"#fabio\">Fabio<\/a><\/li>\n<li><a href=\"#great-success\">Great Success<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#client-side-rendering-fallback\">Client-Side Rendering Fallback<\/a>\n<ul><li><a href=\"#elastic-inelasticity\">Elastic Inelasticity<\/a><\/li>\n<li><a href=\"#how-it-works\">How It Works<\/a><\/li>\n<li><a href=\"#the-results\">The Results<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#load-shedding\">Load Shedding<\/a>\n<ul><li><a href=\"#why-you-need-load-shedding\">Why You Need Load Shedding<\/a><\/li>\n<li><a href=\"#not-so-fast\">Not So Fast<\/a><\/li>\n<li><a href=\"#interleaved-shedding\">Interleaved Shedding<\/a><\/li>\n<li><a href=\"#io-and-worker-processes\">I\/O And Worker Processes<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#component-caching\">Component Caching<\/a>\n<ul><li><a href=\"#the-idea-of-caching\">The Idea Of Caching<\/a><\/li>\n<li><a href=\"#two-hard-things-in-computer-science\">Two Hard Things In Computer Science<\/a><\/li>\n<li><a href=\"#caching-and-interpolation\">Caching And Interpolation<\/a><\/li>\n<li><a href=\"#murphys-law\">Murphy\u2019s Law<\/a><\/li>\n<li><a href=\"#oh-fouc\">Oh FOUC!<\/a><\/li>\n<li><a href=\"#exploding-cache\">Exploding Cache<\/a><\/li>\n<li><a href=\"#making-the-opposite-mistake\">Making The Opposite Mistake<\/a><\/li>\n<li><a href=\"#cache-rules-everything-around-me\">Cache Rules Everything Around Me<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#dependencies\">Dependencies<\/a>\n<ul><li><a href=\"#dont-get-hacked\">Don\u2019t Get Hacked<\/a><\/li>\n<li><a href=\"#do-you-like-free-things\">Do You Like Free Things?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#isomorphic-rendering\">Isomorphic Rendering<\/a>\n<ul><li><a href=\"#the-browser-as-your-server\">The Browser As Your Server<\/a><\/li>\n<li><a href=\"#pairs-of-pages\">Pairs Of Pages<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#the-aggregation-of-marginal-gains\">The Aggregation Of Marginal Gains<\/a><\/li>\n<li><a href=\"#all-your-servers-are-belong-to-redux\">All Your Servers Are Belong To Redux<\/a><\/li>\n<\/ul><h2 id=\"the-situation\">The Situation<\/h2><p>Our team was looking to revitalize the front-end architecture for our product. As tends to be the case with a many years-old monolith, the technical debt had piled up, and front-end modifications were becoming difficult. Increasingly, we were telling product managers that their requested changes were infeasible. It was time to get serious about sustainability.<\/p><p>Within the front-end team, a consensus was quickly reached that a component-oriented architecture built on React and Redux was the best bet for a sustainable future. Our collective experience and intuition favored separating concerns at the component level, extracting reusable components wherever possible, and embracing functional programming.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/redux-logo.svg\" width=\"250\" alt=\"It is surprisingly hard to draw the Redux logo! (Along with three very poorly drawn Redux logos.)\" referrerpolicy=\"no-referrer\" \/><\/p><p>We were beginning with the fairly modest, spaghetti front-end that most monolithic applications seem to evolve into. Browser requests would hit a load balancer, which would forward requests to one of several instances of a Java\/Spring monolith. JSP-generated HTML templates were returned, styled with CSS (LESS), and dynamic client functionality was bolted on with a gratuitous amount of jQuery.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/starting-architecture.svg\" width=\"500\" alt=\"Diagram of monocled, top-hatted user connecting to a Load Balancer, which forwards the request to a Monolith. Monolith responds by returning a rendered JSP document to the load balancer, which sends it to the user. Pretty boring stuff.\" referrerpolicy=\"no-referrer\" \/><\/p><p>The question was how to integrate our desire for a React front-end with a Java monolith. SEO was a very important consideration \u2014 we had full-time SEO consultants on staff \u2014 and we wanted to provide the best possible page load speed, so server-side rendering quickly became a requirement. We knew that React was capable of isomorphic (client- and server-side) rendering. The back-end team was already on their journey towards breaking up the monolith into a microservice architecture. It therefore seemed only natural to extract our React server-side rendering into its own Node.js service.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/final-architecture.svg\" width=\"650\" alt=\"Diagram of our MacGuffin, the monocled, top-hatted user, connecting once again to the Monolith via the Load Balancer. This time, Monolith requests some React component renders from the React service, which sends a response containing the rendered components, a serialized Redux store, and mounting instructions. The Monolith takes these pieces and merges them into a JSP, sending the final output through the Load Balancer and back to the user. Slightly less boring stuff.\" referrerpolicy=\"no-referrer\" \/><\/p><p>The idea was that the monolith would continue to render JSP templates, but would delegate some parts of the page to the React service. The monolith would send rendering requests to the React service, including the names of components to render, and any data that the component would require. The React service would render the requested components, returning embeddable HTML, React mounting instructions, and the serialized Redux store to the monolith. Finally, the monolith would insert these assets into the final, rendered template. In the browser, React would handle any dynamic re-rendering. The result was a single codebase which renders on both the client and server \u2014\u00a0a huge improvement upon the status quo.<\/p><p>As we gained confidence with this new approach, we would build more and more of our features using React, eventually culminating with the entire page render being delegated to the React service. This approach allowed us to migrate safely and incrementally, avoiding a big-bang rewrite.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/starting-transition-goal.svg\" width=\"280\" alt=\"The same web page, composed three different ways, demonstrating an incremental migration path. Starting form is entirely JSP-rendered. Transition form contains a mixture of JSP and React-rendered elements. Goal form is one huge React-rendered component.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Our service would be deployed as a Docker container within a Mesos\/Marathon infrastructure. Due to extremely complex and boring internal dynamics, we did not have much horizontal scaling capacity. We weren\u2019t in a position to be able to provision additional machines for the cluster. We were limited to approximately 100 instances of our React service. It wouldn\u2019t always be this way, but during the period of transition to isomorphic rendering, we would have to find a way to work within these constraints.<\/p><h3 id=\"i-got-99-percentiles\">I Got 99 Percentiles<\/h3><p>The initial stages of this transition weren\u2019t without their hiccups, but our React service rendering performance was reasonable.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/response-latency-5-50ms.svg\" width=\"280\" alt=\"Graph of Response Latency (ms) over time. p50 response time is plotted as being fairly consistent at around 5ms. p99 response time is more erratic but generally around 50ms.\" referrerpolicy=\"no-referrer\" \/><\/p><p>As we ported more and more portions of the site to React, we noticed that our render times were increasing \u2014\u00a0which was expected \u2014 but our 99th percentile was particularly egregious.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/response-latency-30-250ms.svg\" width=\"300\" alt=\"Graph of Response Latency (ms) over time. p50 response time is plotted as being fairly consistent around 30ms. p99 response time is fairly erratic but generally around 250ms. Yeah, I know, graphs are boring. Drawing them was surprisingly fun, though. At least one of us is having fun.\" referrerpolicy=\"no-referrer\" \/><\/p><p>To make matters worse, when our traffic peaked in the evening, we would see large spikes in 99th percentile response time.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/response-latency-evening.svg\" width=\"320\" alt=\"Graph of Response Latency (ms) over the course of about 12 hours, from 12pm to 12am. p50 response time is plotted as being fairly consistent at around 30ms, and increases to about 50ms in the evening, starting at about 6pm. p99 response time is more erratic, hovering around 250ms, and increases to around 350ms in the evening, also starting around 6pm. By 12am response times are clearly decreasing again. If the previous graphs were boring, this graph is like the beginning of an episode of your favourite TV crime drama.\" referrerpolicy=\"no-referrer\" \/><\/p><p>We knew from our benchmarks that it simply does not take 400ms to render even a fairly complex page in React. We profiled and made lots of improvements to the service\u2019s rendering efficiency, including streaming responses, refactoring React component elements to DOM node elements, various Webpack shenanigans, and introducing cached renders for some components. These measures mitigated the problem, and for a while we were hovering right on the edge of acceptable performance.<\/p><h3 id=\"seasonality\">Seasonality<\/h3><p>One day I was looking at our response latency graph, and I noticed that the problem had returned. Unusually high traffic during the previous evening had pushed our 99th percentile response times past the acceptable threshold. I shrugged it off as an outlier \u2014 we were incredibly busy, and I didn\u2019t have time to investigate.<\/p><p>This trend continued for a few days. Every evening when traffic peaked, we would set a new record. Zooming out to show the last few days, there was a clear trend of increasing response time.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/response-latency-week.svg\" width=\"400\" alt=\"Graph of Response Latency (ms) from Monday to Friday. p50 response time is plotted as being fairly consistent at around 50ms, though daily peaks and troughs are clear. p99 response time is more erratic at around 250ms, also with daily peaks and troughs. Importantly, the peak p99 response time is increasing with each passing day, and approaches 400ms on Friday. Seeing something like this in real, production graphs makes you feel like a cast member in Armageddon. Now you can't get that Aerosmith song out of your head. Look, I didn't say you would *like* the alt text.\" referrerpolicy=\"no-referrer\" \/><\/p><p>There was a clear correlation in the graphs between traffic volume and response time. We could attempt to duct tape the problem, but if traffic were to increase, we would be in bad shape. We needed to scale horizontally, but we couldn\u2019t. So how close were we to a calamity? I pulled up an annual traffic graph, and promptly spit out my tea.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/requests-per-minute-annual.svg\" width=\"650\" alt=\"Graph of Requests Per Minute over the course of one year, from January to the end of December. Y-axis goes from zero to 'lots' of requests. A single line plots requests per minute, which vary a bit through the day, so the line is slightly erratic. Still, a general trend is clear, with requests hitting a low point in December, and increasing until a high point in July. A 'You are here.' arrow points to early March, which is where I was situated when I first looked at this graph (and promptly spit out my tea).\" referrerpolicy=\"no-referrer\" \/><\/p><p>Without a doubt our response times would dramatically increase with traffic. It was currently spring \u2014\u00a0roughly the annual midpoint for traffic \u2014 and by summer we would be drowning in requests. This was Very Bad.<\/p><p>But how could we have missed this? We thought we had solved this problem already. What gives?<\/p><p>I\u2019m pretty sure we were caught off guard due to the seasonality of our traffic. Starting the previous summer \u2014 when traffic was at its peak \u2014 we began moving more and more functionality to React. If traffic had remained constant, the increased component rendering load would have caused our response times to increase. Instead, as the year progressed, traffic was decreasing. Requests were going down, but the per-request workload was going up! The result was a roughly flat response time during the fall and winter seasons. As traffic picked up again in the spring, our response times rapidly increased, and this time the effect was magnified by the increased per-request workload.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/response-latency-vs-components.svg\" width=\"650\" alt=\"Graph of Response Latency (ms) vs # React components. X-axis spans one year, from July to July. Y-axis ranges from 0 to 400 ms response latency. Two lines are plotted: p99 response latency and number of components. Both lines suddenly begin in mid-July. Response latency starts around 100ms (hey, this graph isn't exactly to scale, I'm drawing it from memory), and increases to around 300ms in December, then decreases during the winter, to about 250ms in February, finally increasing roughly linearly to a peak of 400ms in July. The number of components line increases purely linearly from mid-July to the following July, though I didn't provide a value for its Y-axis, simply showing that it starts from a small number of components, and increases to 'lots'. The general idea of this graph is that the number of components can increase while response times decrease, because demand for rendering can fluctuate seasonally. I think this is the biggest alt text I've ever written. In a pinch, you could use this alt text to filibuster.\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"randomness\">Randomness<\/h3><p>Out of ideas for squeezing easy performance wins out of the system, I started asking some of my colleagues for suggestions. During one one of these conversations, somebody mentioned the fact that our service discovery mechanism, Consul, returns three random service instances for every service discovery request.<\/p><p>I remembered reading a <a href=\"https:\/\/genius.com\/James-somers-herokus-ugly-secret-annotated\">fantastic Genius article<\/a> several years ago, which told the story of the performance regressions that they experienced when Heroku silently switched to a randomized load balancing strategy, causing a 50x decrease in scaling efficiency. If we were using a similar load balancing strategy, then we were likely to be suffering the same fate. I did a bit of spelunking and confirmed that this was indeed the case.<\/p><p>Basically, when the monolith needs to make a request to the React service, it needs to know the IP address and port where it can locate an instance of that service. To get this information, a DNS request is sent to Consul, which keeps track of every active service instance. In our configuration, for each service discovery request, Consul returns three <em>random<\/em> instances from the pool. This was the only load balancing mechanism within the system. Yikes!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/service-discovery.svg\" width=\"450\" alt=\"Diagram of how service discovery works. Monolith makes a call to Consul requesting the location of the React service. Consul responds with three different IP addresses and port number combinations, which are the locations of three random React service instances. The Monolith then sends a request to one of these React service instances, asking for a Header component to be rendered. The React service renders the component, and returns it to the Monolith.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Before I continue, I should explain why random load balancing is inefficient.<\/p><p>Let\u2019s say you have a load balancer and three service instances. If the load balancer routes requests <em>randomly<\/em> to those instances, the distribution of requests will always be severely uneven.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/random-load-balancing.svg\" width=\"500\" alt=\"Diagram showing Load Balancer routing requests randomly to three service instances. Instances receive 7, 2, and 4 requests, respectively. Random load balancing always produces a distribution like this. Yes, it does. YES IT DOES!!!1\" referrerpolicy=\"no-referrer\" \/><\/p><p>I have explained this problem to many people, and it confuses a huge number of them. It reminds me of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Monty_Hall_problem\">Monty Hall problem<\/a> \u2014\u00a0even though it\u2019s true, people find it hard to believe.<\/p><p>But yes, it\u2019s true: random load balancing does not balance load at all! This can be easier to understand if you flip a coin, counting the number of heads and tails. The balance is almost always uneven.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/coin-tosses.svg\" class=\"breakout\" alt=\"Diagram of coin tosses over time, titled, 'Coin tosses are random yet uneven!' Describing this is going to be tricky, since I had to invent this visualization to make the point, but here we go. The x-axis represents time, and the y-axis represents bias in coin tosses towards heads or tails. With each coin toss, the plotted line moves incrementally upwards (towards heads), or downwards (towards tails). Each point on the line is marked with an 'H' or a 'T', respectively. When coin tosses are evenly distributed, the line remains roughly horizontal, moving up, down, up, down, etc. When coin tosses are uneven, such as when we get four heads in a row towards the end of the graph, the line moves up, up, up, up. This diagram shows that, at any given point in time, coin toss results can be unbalanced in favor of heads or tails, despite the long-term trend being towards evenness. These were actual coin tosses that I performed at like 1am, when I was suddenly inspired to draw this diagram. That's how committed I am to you. I'm here for you.\" referrerpolicy=\"no-referrer\" \/><\/p><p>A common response is that the load may not be balanced at the beginning, but over time the load will \u201caverage out\u201d so that each instance will handle the same number of requests. This is correct, but unfortunately it misses the point: at almost every <em>moment<\/em>, the load will be unevenly distributed across instances. Virtually all of the time, some servers will be concurrently handling more requests than the others. The problem arises when a server decides what to do with those extra requests.<\/p><p>When a server is under too much load, it has a couple of options. One option is to drop the excess requests, such that some clients will not receive a response, a strategy known as <em>load shedding<\/em>. Another option is to queue the requests, such that every client will receive a response, but that response might take a long time, since it must wait its turn in the queue. To be honest, both options are unacceptable.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/load-shedding.svg\" width=\"500\" alt=\"Diagram of how load shedding works. Load Balancer sends 3 requests ('too many requests!') to an instance of a service. The instance sends a response for request #1, and and discards requests #2 and #3 into a bottomless pit, which is actually how servers work and not at all a metaphor.\" referrerpolicy=\"no-referrer\" \/><\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/queueing.svg\" width=\"500\" alt=\"Diagram of how queuing works. Load Balancer sends 3 requests ('too many requests!') to an instance of a service. The instance sends a response for request #1, and enqueues the remaining two requests for later processing. I drew tiny front and back doors on the service, which the queued requests can use to enter and exit. One cannot live on boxes and arrows alone.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Our Node servers were queueing excess requests. If we have at least one service instance per concurrent request, the queue length for each instance will always be zero, and response times will be normal, provided that we are balancing the load evenly. But when we are using a random load balancing strategy, some instances will <em>always<\/em> receive an unfair share of requests, forcing them to queue the excess ones. The requests at the back of a queue must wait for the <em>entire<\/em> queue to be processed, dramatically increasing their response time.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/queues-increase-response-time.svg\" width=\"500\" alt=\"Diagram entitled 'Queues increase response time!' Load Balancer sends 3 requests ('too many requests!') to an instance of a service. The instance sends a response for request #1, and enqueues the remaining two requests for later processing. Each requests requires 10ms to process, so request #1 experiences 10ms latency. Request #2 must wait for request #1's 10ms processing, and then also its own, suffering 20ms latency in total. Request #3 accordingly must wait for 30ms. This cycle continues until either all requests are handled or your server explodes, all of your users defect, and your team disbands.\" referrerpolicy=\"no-referrer\" \/><\/p><p>To make matters worse, it doesn\u2019t matter how many service instances we have. The random allocation of requests guarantees that some instances will always be sitting idle, while other instances are being crushed by too much traffic. Adding more instances will reduce the probability that multiple requests will be routed to the same instance, but it doesn\u2019t eliminate it. To really fix this problem, you need load balancing.<\/p><p>I installed metrics to graph request queue length per service instance, and it was clear that some services were queueing more requests than others. The distribution would change over time, as the random load balancing just happened to select different instances.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/instance-queue-length.svg\" width=\"450\" alt=\"Graph of Request Queue Length (Per Instance). X-axis represents time, and y-axis spans from 0 to 5 requests enqueued. Three hypothetical service instances are plotted, with queue lengths that vary from 0 to 3 requests. Crucially: the queue lengths for each service are different, and constantly fluctuating, a symptom of uneven load balancing.\" referrerpolicy=\"no-referrer\" \/><\/p><p>So we need to ensure that the load is evenly distributed across instances. Not wishing to repeat past mistakes, I began researching load balancing strategies. This is a really fascinating topic, and if you\u2019re interested in learning more, I highly recommend Tyler McMullen\u2019s presentation, <em><a href=\"https:\/\/www.infoq.com\/presentations\/load-balancing\">Load Balancing is Impossible<\/a><\/em>.<\/p><p>Unfortunately, there are so many permutations of load balancing strategies that it would be impossible to test them all in a production environment. The iteration cost for each strategy would be too great. So I followed Genius\u2019 lead and wrote a simple in-memory load balancing simulator which enabled me to experiment with dozens of strategies over the course of a few hours. This gave me much greater confidence in the shortlist of solutions that would be tested in production.<\/p><p>One clever solution involves configuring our React service to shed load, returning a <code class=\"highlighter-rouge\">503 Service Unavailable<\/code> instead of queueing excess requests. The monolith would receive the <code class=\"highlighter-rouge\">503<\/code> more or less immediately, and would then retry its request on a different, randomly selected node. Each retry has an exponentially decreasing probability of reaching another overloaded instance.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/load-shedding-503.svg\" width=\"500\" alt=\"Diagram of how load shedding can be used in conjunction with random retries to achieve a kind of load balancing. Monolith sends a request to an instance of a service, which responds with a 503 Service Unavailable, because it has too many requests in its queue. Monolith then retries its requests on a different, randomly selected instance. The second instance does not have a queue and so it happily responds with a cool, sunglasses emoji, or whatever other metaphor you want to use for a successful response. If you had requested a poop emoji, the response would be a poop emoji, but then it seem look like a failed response, so I went with sunglasses emoji instead.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Unfortunately, when I simulated this approach I discovered that it was not the most efficient. It was certainly better than a single, random attempt, but it does not perform as well as a round-robin algorithm, for example.<\/p><p>There are a few reasons for this. First, each retry adds additional network latency to the ultimate response time. All other things being equal, an algorithm which does not issue redundant requests will not suffer this overhead.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/latency-adds-up.svg\" width=\"500\" alt=\"Diagram of how load shedding with random retries adds latency with each unsuccessful retry. Assumem 5ms latency for every request or response sent. Monolith sends a request to an instance of a service, which responds with a 503 Service Unavailable. Monolith then retries on a different instance, which responds successfully. The total network cost is 20ms, instead of the 10ms you would have suffered if there had been no retry. Neutral face emoji demonstrates the unsatisfactory delay. Really, you only need a couple of emojis to handle most situations. Smiley, frowny, neutral, poop, sunglasses. I guess poop emoji is usually smiling, but that's a technicality. Definitely the person looking back at poop is not smiling. Well, I guess that also depends. It's complicated. Let's move on.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Second, as the cluster of service instances becomes saturated with traffic, the probability that a retry will reach a healthy instance decreases! Think about a 5 instance cluster, with 4 instances at capacity, unable to handle additional requests \u2014\u00a0the odds that a retry will reach the 1 available instance are only 20%! This means that some requests will suffer many retries in order to receive a response.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/probability.svg\" width=\"650\" alt=\"Diagram showing Monolith choosing randomly among 5 service instances for its next request. Instances have 3, 2, 1, 0, and 4 requests in queue, respectively. Random selection has a 20% chance of picking the instance with no requests in its queue, i.e. the 'best' instance. Of course, I love all my instances equally. None of them are truly the 'best'. Some are just more enqueued than others.\" referrerpolicy=\"no-referrer\" \/><\/p><p>This problem is less pronounced when you can scale horizontally, but hopefully the inefficiency of this solution is clear. I wanted to do better, if possible.<\/p><h3 id=\"round-robin\">Round-Robin<\/h3><p>A much better approach is to route each request, in turn, to the next instance in the cluster, commonly known as a <em>round-robin<\/em> algorithm.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/round-robin.svg\" width=\"500\" alt=\"Diagram of how round-robin load balancing works. Load balancer sends 9 requests to 3 service instances. Instance #1 receives requests 1, 4, and 7. Instance #2 receives requests 2, 5, and 8. Instance #3 receives requests 3, 6, and 9. Round-robin simply picks the next instance, in order, and sends a request to it. The result is an even distribution of requests across instances. I don't know why it's called round-robin, though. It doesn't have anything to do with robins. Or even Batman &amp; Robin, which was mediocre at best. George Clooney is too pretty to be Batman.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Round-robin guarantees that each service instance will receive exactly its fair share of requests. This is the simplest load balancing algorithm that we can honestly say is <em>balancing<\/em> load in a meaningful way. Accordingly, it vastly outperforms random, and load shedding with random retries.<\/p><p>Deceptively, round-robin is not the absolute most efficient approach, because requests can vary in the amount of work that they require the server to perform. One request might require 5ms to render a single React component, while another may require 50ms to render a page filled with hundreds of components. This natural variance in per-request workload means that round-robin can send requests to instances which are still processing a previous request, while other instances remain idle. This is because round-robin does not take an instance\u2019s workload into account. It <em>strictly<\/em> allocates requests as a blackjack dealer would deal cards: everybody gets the same number of cards, but some cards are better than others!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/blackjack.svg\" width=\"300\" alt=\"Comic of a blackjack dealer talking to a single player. 'Dealer has 20. You have... 96'. Caption underneath reads, 'Distributed systems blackjack.' I really messed up the bow tie on the dealer, but I'm too lazy to draw it again.\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"join-shortest-queue\">Join-Shortest-Queue<\/h3><p>Obviously we can\u2019t speak of the \u201cbest\u201d load balancing algorithm, because the \u201cbest\u201d choice depends on your particular circumstances. But I would be remiss not to describe what is probably the most widely useful approach, which is a <em>join-shortest-queue<\/em> strategy.<\/p><p>I\u2019m going to lump a few variations of this strategy together. Sometimes we might use a <em>least-connections<\/em>, or a <em>join-idle-queue<\/em> approach, but the unifying principle is the same: try to send requests to the instance which is least overloaded. We can use different heuristics to approximate \u201cload\u201d, including the number of requests in the instance\u2019s queue, or the number of outstanding connections, or having each instance self-report when they are ready to handle another request.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/join-shortest-queue.svg\" width=\"550\" alt=\"Diagram of how join-shortest-queue load balancing works. Three service instances have 2, 3, and 1 requests enqueued, respectively. Load Balancer sends the next request to the instance with only 1 request in queue. It is observed that round-robin might pick the other two instances, because round-robin is about as good at load balancing as George Clooney is at playing Batman. Great actor, just not for Batman. Great actor, though.\" referrerpolicy=\"no-referrer\" \/><\/p><p>The join-shortest-queue approach outperforms round-robin because it attempts to take the per-request workload into account. It does this by keeping track of the number of responses it is waiting for from each instance. If one instance is struggling to process a gigantic request, its queue length will be 1. Meanwhile, another instance might complete all of its requests, reducing its queue length to 0, at which point the load balancer will prefer to send requests to it.<\/p><h3 id=\"fabio\">Fabio<\/h3><p>So how did we resolve our load balancing woes? We ended up implementing a round-robin load balancer, <a href=\"https:\/\/github.com\/fabiolb\/fabio\">Fabio<\/a>, as a compromise solution, trading off performance for convenience.<\/p><p>While Fabio does not support a join-shortest-queue load balancing strategy, it integrates seamlessly with Consul, giving us <a href=\"http:\/\/microservices.io\/patterns\/server-side-discovery.html\">server-side service discovery<\/a>. This means that our monolith can simply send requests to Fabio, and Fabio figures out both how to get them to the React service, and also how to balance the load in a reasonable way.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/fabio.svg\" width=\"550\" alt=\"Diagram of how Fabio acts as a load balancer within the architecture. Monolith sends requests to Fabio, which then contacts Consul to get the IP addresses and port numbers of the destination service instances. Fabio then forwards the request to a service instance, using a round-robin algorithm. Service instance sends a response to Fabio, which forwards it back to the Monolith. In reality the Consul service discovery lookups are cached, otherwise too much latency would be introduced. I reserve the right to simplify things for pedagogical purposes. If you don't like it, draw your own diagrams.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Of course, in this configuration our load balancer becomes a single point of failure \u2014 if it dies, we can\u2019t render any web pages!<\/p><p>To provide an availability strategy, we implemented our Fabio load balancer as just another containerized service \u2014\u00a0load balancing as a service. The monolith would use Consul to discover a <em>random<\/em> Fabio instance, and send requests to that instance. If a Fabio instance dies, Consul would automatically detect this and stop offering that instance as one of the random options. We tested failover in production by sending a small amount of traffic through Fabio, and then manually killing a Fabio instance. Consul would reliably recover from this failure within a couple of seconds. Not bad!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/fabio-failover.svg\" width=\"600\" alt=\"Diagram of how multiple Fabio instances provide failover capability. Monolith requests the IP and port of a Fabio instance from Consul. Meanwhile, a Fabio instance has exploded. Consul quickly detects that Fabio has exploded, and does not return that Fabio instance as an option for the Monolith to connect to. The Monolith is therefore unaware of the failure, and blissfully sends its traffic to one of the remaining Fabio instances. Fabio then forwards requests to the React service, which sends responses back to Fabio, which forwards them back to the Monolith.\" referrerpolicy=\"no-referrer\" \/><\/p><p>We might be tempted to assume that randomly selecting a load balancer would preserve the performance issue we are trying to solve, but in practice this is not a problem. Each instance of Fabio can easily accommodate all of the traffic destined for our React service cluster. If our load balancers are sufficiently fast, it doesn\u2019t matter if the load is evenly balanced across the load balancers themselves. We have multiple load balancers purely to provide failover capability.<\/p><h3 id=\"great-success\">Great Success<\/h3><p>When the new round-robin load balancing strategy was productionized and ramped up to 100% of traffic, our React service instance queue lengths were a sight to behold. All of the queues converged around the same length. The system works!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/queue-length-with-load-balancing.svg\" width=\"450\" alt=\"Graph of Request Queue Length (Per Instance). This time, with load balancing installed! X-axis is time, and y-axis is request queue length, from 0 to 5. Three service instance request queues are plotted, and you know what? They're all sitting steady at 2 requests in queue! Ideally we would have less than 1 request in queue, but the point is that with load balancing finally installed, no instance can become overloaded. Make sure to balance your loads, people.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Even better, our original problem was solved: peak traffic response latency spikes smoothed out, and our 99th percentile latency dropped. Everything \u201cjust worked\u201d, as we had originally hoped.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/latency-with-load-balancing.svg\" width=\"350\" alt=\"Graph of Response Latency (ms) during the activation of load balancing. p50 response time is about 50ms, until load balancing is activated, at which point it drops to about 40ms. p99 response time is always erratic, and hovers around 350ms until load balancing is activated, after which it drops to around 200ms. A terrific win!\" referrerpolicy=\"no-referrer\" \/><\/p><h2 id=\"client-side-rendering-fallback\">Client-Side Rendering Fallback<\/h2><h3 id=\"elastic-inelasticity\">Elastic Inelasticity<\/h3><p>The addition of load balancing to our system effectively solved our high latency issues, and the efficiency gains provided a modest amount of additional capacity. But we were still concerned about extraordinary scenarios. Bots would scrape our website, triggering a huge surge in requests. Seasonality, including holidays, could also trigger unexpected increases in traffic. We had enough server capacity to keep up with normal traffic growth, but we could only sleep easily with the knowledge that our system would be resilient under significantly higher load.<\/p><p>Ideally we would build an auto-scaling system which could detect surges in traffic, and scale horizontally to accommodate them. Of course, this was not an option for us. We also couldn\u2019t simply provision 10x more capacity than required. Was there <em>any<\/em> way we could add some kind of margin of safety? As it turns out, there was.<\/p><p>We couldn\u2019t shed load by dropping requests, but I started thinking about load shedding more generally, and I began to wonder if some kind of load <em>throttling<\/em> would be possible. Late one evening, a solution popped into my head. We were using Redux, and one of the nice things about Redux is that it makes serialization of state very easy, enabling isomorphic rendering. We were rendering requests on the server, and then handling re-renders on the client, yet isomorphic rendering allows us to render on <em>either<\/em> the server <em>or<\/em> client. We don\u2019t always have to do both.<\/p><p>So the way to throttle load was profound in its simplicity: when the server is under high load, skip the server-side render, and force the browser to perform the initial render. In times of great need, our rendering capacity would automatically expand to include every single user\u2019s computer. We would trade a bit of page load speed for the ability to elastically scale on a fixed amount of hardware. Redux is the gift that just keeps on giving!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/server-side-rendering.svg\" width=\"550\" alt=\"Diagram of how React server-side rendering works. Browser sends a request to the Monolith, which requests some React component renders from the React service. The React service responds with the rendered components, serialized Redux store, and mounting instructions. These pieces are merged into a JSP template by the Monolith, and sent back to the browser. Pretty straightforward.\" referrerpolicy=\"no-referrer\" \/><\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/client-side-rendering-fallback.svg\" width=\"550\" alt=\"Diagram of how client-side rendering fallback works. Browser sends a request to the Monolith, which requests some React component renders from the React service. This time, the React service is under heavy load and skips the React component rendering, responding only with a serialized Redux store, and mounting instructions. These pieces are merged into a JSP template by the Monolith, and sent back to the browser. When React mounts in the browser, it performs an initial render, and the user can finally see a picture of a cat, or whatever they were looking for. That's really none of our business. We are in the React and Redux businesses, respectively.\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"how-it-works\">How It Works<\/h3><p>Building a client-side rendering fallback system is remarkably straightforward.<\/p><p>The Node server simply maintains a request queue length counter. For every request received, increment the counter, and for every error or response sent, decrement the counter. When the queue length is less than or equal to <code class=\"highlighter-rouge\">n<\/code>, perform regular data fetching, Redux store hydration, and a server-side React render. When the queue length is greater than <code class=\"highlighter-rouge\">n<\/code>, skip the server-side React rendering part \u2014 the browser will handle that, using the data from the Redux store.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/client-side-rendering-fallback-queue.svg\" width=\"600\" alt=\"Diagram of how client-side rendering fallback acts as a kind of load throttling. Monolith sends 7 requests to the React service. The first request is server-side rendered, because at that point in time, the service has 0 requests in queue. The next 6 requests are triaged and queued. The first 2 requests are queued for server-side rendering, because we have arbitrarily chosen a queue length of &lt; 3 as our light load cutoff for server-side rendering. The next 3 requests are queued for client-side rendering, because we have arbitrarily chosen a queue length of &lt; 6 as our heavy load cutoff for client-side rendering. The final request exceeds our maximum queue length, and is dropped in order to shed load.\" referrerpolicy=\"no-referrer\" \/><\/p><p>The exact value of <code class=\"highlighter-rouge\">n<\/code> will need to be tuned to match the characteristics of your application. Generally speaking, <code class=\"highlighter-rouge\">n<\/code> should be slightly larger than the typical queue length during peak expected load.<\/p><p>Of course, if SEO is a requirement, this approach contains a slight problem: if a search engine crawls the site during a traffic surge, it may not receive a server-side rendered response, and therefore it may not index your pages! Fortunately this is an easy problem to solve: provide an exception for known search engine user agent strings.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/club-react.svg\" width=\"500\" alt=\"Comic of people lined up to get into Club React. Google made it past the bouncer, who is speaking with the next person in line, who is arguing, 'I'm a friend of Google.'\" referrerpolicy=\"no-referrer\" \/><\/p><p>There is a possibility that the search engine will punish our rankings for treating it differently than other clients. However, it is important to remember that the client-side rendering fallback exists to prevent us from dropping requests during traffic surges, or server failures. It is a safety net for rare, exceptional circumstances. The alternative is to risk sending <em>nothing<\/em> to the crawler, which could also result in punishment. In addition, we aren\u2019t serving <em>different<\/em> content to the search engine, we are merely providing it with priority rendering. Plenty of users will receive server-side rendered responses, but search engines will always receive one. And of course, it is easy to remove this priority if it is considered counter-productive.<\/p><h3 id=\"the-results\">The Results<\/h3><p>The day after we deployed client-side rendering fallback to production, a traffic spike occurred and the results were outstanding. The system performed exactly as we had hoped. Our React service instances automatically began delegating rendering to the browser. Client-side renders increased, while server-side request latency held roughly constant.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/client-side-render-requests.svg\" width=\"350\" alt=\"Graph of Requests Per Minute with client-side rendering fallback installed. X-axis is time, and y-axis ranges from 0 to 'lots' of requests. Two lines are plotted: total requests, and client-side rendered requests. Total requests is roughly constant until a big spike occurs, which lasts for a while, and then returns to normal levels. During the spike, client-side renders increase from 0 to a level which mirrors the increase in total requests. When the spike ends, client-side renders return to zero. The system works!\" referrerpolicy=\"no-referrer\" \/><\/p><p>We benchmarked the efficiency gained through this approach, and found that it provides a roughly 8x increase in capacity. This system went on to save us multiple times over the next several months, including during a deployment error which significantly reduced the number of React service instances. I\u2019m extremely pleased with the results, and I do recommend that you experiment with this approach in your own isomorphic rendering setup.<\/p><p>Previously I mentioned that load shedding could be used in conjunction with random retries to provide an improvement over purely random load balancing. But even if a different load balancing strategy is used, it is still important to ensure that the React service can shed load by dropping excess requests.<\/p><p>We discovered this the hard way during a freak operations accident. A Puppet misconfiguration accidentally restarted Docker on every machine in the cluster, <em>simultaneously<\/em>. When Marathon attempted to restart the React service instances, the first ones to register with Consul would have 100% of the normal request load routed to them. A single instance could be swamped with 100x its normal request load. This is very bad, because the instance may then exceed the Docker container\u2019s memory limit, triggering the container\u2019s death. With one less active instance, the other instances are now forced to shoulder the additional load. If we aren\u2019t lucky, a cascade failure can occur, and the entire cluster can fail to start!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/cascade-failure.svg\" width=\"550\" alt=\"Diagram of how services which lack load shedding can experience a cascade failure when starting. Load Balancer sends a firehose of requests to one instance of React service, which enqueues as many as it can until its memory is exhausted and it explodes. Another instance of React service comes online, and the Load Balancer, with nowhere else to send its requests, directs the firehose toward this new instance, which promptly screams 'Crap!' The cycle continues until either you go out of business or somebody installs load shedding.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Checking our graphs during this incident, I saw request queue lengths spike into the <em>thousands<\/em> for some service instances. We were lucky the service recovered, and we immediately installed a load shedding mechanism to cap the request queue length at a reasonable number.<\/p><h3 id=\"not-so-fast\">Not So Fast<\/h3><p>Unfortunately the Node event loop makes load shedding tricky. When we shed a request, we want to return a <code class=\"highlighter-rouge\">503 Service Unavailable<\/code> response so that the client can implement its fallback plan. But we can\u2019t return a response until all earlier requests in the queue have been processed. This means that the <code class=\"highlighter-rouge\">503<\/code> response will not be sent immediately, and could be waiting a long time in the queue. This in turn will keep the client waiting for a response, which could ruin its fallback plan, especially if that plan was to retry the request on a different instance.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/shed-queue.svg\" width=\"400\" alt=\"Diagram of an instance of React service which implements load shedding in conjunction with request queues. A request at the back of the queue, which is due to be shed, must wait for all of the other requests before actually being shed. Meanwhile, the developer also sheds... tears.\" referrerpolicy=\"no-referrer\" \/><\/p><p>If we want load shedding to be useful, we need to send the <code class=\"highlighter-rouge\">503<\/code> response almost immediately after the doomed request is received.<\/p><h3 id=\"interleaved-shedding\">Interleaved Shedding<\/h3><p>After a bit of brainstorming, I realized that we could provide fast shedding by interleaving request rendering and shedding.<\/p><p>I built a proof of concept by pushing all requests to be rendered into a rendering queue, implemented with a simple array. When a new request arrived, if the queue was smaller than <code class=\"highlighter-rouge\">m<\/code> \u2014 where <code class=\"highlighter-rouge\">m<\/code> is the maximum number of concurrent requests to accept \u2014 I would push the request object into the array. If the queue has grown too large, a <code class=\"highlighter-rouge\">503<\/code> response is immediately sent.<\/p><p>When the server starts, I call a function which pulls a single request from the head of the rendering queue, and renders it. When the request has finished rendering, the response is sent, and the function is recursively called with <code class=\"highlighter-rouge\">setImmediate()<\/code>. This schedules the next single request render <em>after<\/em> the Node event loop processes accumulated I\/O events, giving us a chance to shed the excess requests.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/interleaved-shedding.svg\" class=\"breakout\" alt=\"Diagram of how interleaved shedding works. 12 requests arrive as events in the Node event loop over time. If our max queue length is 3, the first 3 requests are placed into a rendering queue, which is implemented as an array. The next 3 requests are shed, with 503 responses sent for each. A setImmediate() callback, enqueued at application start, is then called, pulling request #1 from the render queue, rendering it, and sending a response to the client. The setImmediate() render callback recursively calls itself, enqueuing another render event behind requests 7 through 9, which arrived in the meantime. Request #7 is added to the render queue, bringing it back up to a maximum of 3 requests in queue. Requests #8 and #9 are shed, with 503 responses sent for each. The setImmediate() render callback is the next event in the Node event loop, and pulls request #2 from the render queue, rendering it, and sending a response. This cycle continues forever: render one request, then shed everything that doesn't fit in the render queue. This approach limits shed response latency to be approximately the processing time of the render request which lies ahead of it. This keeps our shed latency to reasonable, but not excellent, levels.\" referrerpolicy=\"no-referrer\" \/><\/p><p>The effect is that a single request is rendered, then <em>all<\/em> excess requests are shed, then another single request is rendered, and so on. This approach limits the shed response latency to approximately the length of the request that was rendered before it.<\/p><p>Of course, it is possible to provide even faster shedding.<\/p><h3 id=\"io-and-worker-processes\">I\/O And Worker Processes<\/h3><p>To achieve almost instantaneous load shedding, we refactored our application to spawn a <a href=\"https:\/\/nodejs.org\/api\/cluster.html\">cluster<\/a> of Node processes.<\/p><p>The idea was simple: dedicate one process exclusively to load shedding. When the service starts, the cluster master process forks a number of worker processes. The master process handles I\/O, receiving incoming requests and immediately returning a <code class=\"highlighter-rouge\">503<\/code> if the worker processes are too busy. If a worker is idle, the master process sends requests to it. The worker performs all of the heavy lifting, including React component rendering, and returns a response to the master. The master process finally sends the HTTP response to the client.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/io-worker.svg\" width=\"500\" alt=\"Diagram of how I\/O and worker process architecture works, with 1 I\/O process and 2 worker processes. Load Balancer sends 3 requests to the React service, which are received by its I\/O process. The I\/O process forwards requests #1 and #2 to each of the worker processes, but immediately sends a 503 response to shed request #3, since all of the workers are now considered busy. When the workers are finished processing requests #1 and #2, those responses are forwarded to the I\/O process which sends them back to the Load Balancer for final delivery. This is a bit of a simplification: in reality, the I\/O process can implement a queue so that you have a bit of elastic capacity in the event of fluctuating demand, or to implement a client-side rendering fallback threshold, etc.\" referrerpolicy=\"no-referrer\" \/><\/p><p>This is the approach we shipped to production. Although it is a bit more complicated, it gives us the flexibility to experiment with various numbers of worker processes. It is also important, when evolving towards a microservice architecture, to take the easy latency wins where we can have them.<\/p><h2 id=\"component-caching\">Component Caching<\/h2><h3 id=\"the-idea-of-caching\">The Idea Of Caching<\/h3><p>Whenever we\u2019re attempting to improve performance, the topic of caching is going to come up. Out of the box, React server-side rendering performance is not nearly as fast as, say, a JSP template, and so there has been considerable interest in implementing caching strategies for React.<\/p><p>Walmart Labs has produced a very fancy <a href=\"https:\/\/github.com\/electrode-io\/electrode-react-ssr-caching\">caching library<\/a>, <code class=\"highlighter-rouge\">electrode-react-ssr-caching<\/code>, which provides caching of HTML output on a per-component basis. For dynamic rendering, prop values can either be cached or interpolated. It\u2019s a very impressive system.<\/p><p>And whoa, it\u2019s fast! Liberal use of caching can reduce render times to sub-millisecond levels. This is clearly the approach which offers the greatest performance gains.<\/p><h3 id=\"two-hard-things-in-computer-science\">Two Hard Things In Computer Science<\/h3><p>Unfortunately, this approach is not without its cost. To implement caching, <code class=\"highlighter-rouge\">electrode-react-ssr-caching<\/code> relies on React private APIs, and mutates some of them. This effectively ties the library to React 15, since a complete rewrite of React\u2019s core algorithm shipped with React 16.<\/p><p>Even more pernicious, there is that old saw looming in the background:<\/p><blockquote>\n<p>There are only two hard things in Computer Science: cache invalidation and naming things. \u2014 Phil Karlton<\/p>\n<\/blockquote><p>At it turns out, implementing caching on a per-component basis produces a lot of subtle problems.<\/p><h3 id=\"caching-and-interpolation\">Caching And Interpolation<\/h3><p>In order to cache a rendered React component, <code class=\"highlighter-rouge\">electrode-react-ssr-caching<\/code> needs to know what to do with the component\u2019s props. Two strategies are available, \u201csimple\u201d and \u201ctemplate\u201d, but I will use the more descriptive terms, \u201cmemoization\u201d and \u201cinterpolation\u201d.<\/p><p>Imagine a <code class=\"highlighter-rouge\">&lt;Greeting&gt;<\/code> component, which renders a greeting for the user. To keep things simple, let\u2019s assume we only support English and French greetings. The component accepts a <code class=\"highlighter-rouge\">language<\/code> prop, which could be either <code class=\"highlighter-rouge\">en<\/code> or <code class=\"highlighter-rouge\">fr<\/code>. Eventually, two versions of the component would be cached in memory.<\/p><p>When using the memoization strategy, the component is rendered normally, and one or more of its props are used to generate a cache key. Every time a relevant prop value changes, a different, rendered copy of the component is stored in the cache.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/memoization.svg\" width=\"350\" alt=\"Table illustrating that the 'Greeting_en' cache key corresponds with the '&lt;p&gt;Hello!&lt;\/p&gt;' rendered component HTML, and the 'Greeting_fr' cache key corresponds with the '&lt;p&gt;Bonjour!&lt;\/p&gt;' rendered component HTML.\" referrerpolicy=\"no-referrer\" \/><\/p><p>By contrast, the interpolation strategy treats the component as a template generation function. It renders the component once, stores the output in cache, and for subsequent renders it merges the props into the cached output.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/interpolation.svg\" width=\"550\" alt=\"'Greeting' cache key corresponds with the '&lt;p&gt;@1@&lt;\/p&gt;' rendered component HTML template. When rendering a Greeting component with 'language' prop 'fr', the resulting HTML is '&lt;p&gt;fr&lt;\/p&gt;', which is obviously not what we want. When rendering a Greeting component with 'language' prop 'Bonjour!', the resulting HTML is '&lt;p&gt;Bonjour!&lt;\/p&gt;', which is the original intention.\" referrerpolicy=\"no-referrer\" \/><\/p><p>It is important to note that we can\u2019t simply pass a language code to the <code class=\"highlighter-rouge\">&lt;Greeting&gt;<\/code> component when we are using interpolation. The <em>exact<\/em> prop values are merged into the cached component template. In order to render English and French messages, we have to pass those exact messages into the component as props \u2014\u00a0conditional logic is not usable inside interpolated component <code class=\"highlighter-rouge\">render()<\/code> methods.<\/p><h3 id=\"murphys-law\">Murphy\u2019s Law<\/h3><p>How do we choose between prop memoization and interpolation strategies for our cached components? A global configuration object stores the choice of strategy for each component. Developers must manually register components and their strategies with the caching config. This means that if, as a component evolves, its prop strategy needs to change, the developer must remember to update the strategy in the caching config. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Murphy%27s_law\">Murphy\u2019s Law<\/a> tells us that sometimes we will forget to do so. The consequences of this dependence on human infallibility can be startling.<\/p><p>Let\u2019s say our <code class=\"highlighter-rouge\">&lt;Greeting&gt;<\/code> component is using a memoization strategy for its props, and the <code class=\"highlighter-rouge\">language<\/code> prop value is still being used to generate the cache key. We decide that we would like to display a more personalized greeting, so we add a second prop to the component, <code class=\"highlighter-rouge\">name<\/code>.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/memoize-name.svg\" width=\"480\" alt=\"Rendering a memoized Greeting component which receives a 'language' prop of 'en', and a 'name' prop of 'Brutus', will result in '&lt;p&gt;Hello, Brutus!&lt;\/p&gt;'.\" referrerpolicy=\"no-referrer\" \/><\/p><p>In order to accomplish this, we must update the component\u2019s entry in the caching config so that it uses the interpolation strategy instead.<\/p><p>But if we forget to update the strategy, <em>both prop values<\/em> will be memoized. The first two user names to be rendered within the <code class=\"highlighter-rouge\">&lt;Greeting&gt;<\/code> component will be cached, one per language, and will accidentally appear for all users!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/memoize-gone-wrong.svg\" width=\"550\" alt=\"Rendering a Greeting component which we intended to interpolate but accidentally memoized produces unexpected results. If the Greeting component receives a 'language' prop of 'en', and a 'name' prop of 'Brutus', and the cache key only takes the 'language' prop into account, it will result in '&lt;p&gt;Hello, Brutus!&lt;\/p&gt;'. If the Greeting component is rendered a second time with 'name' prop set to 'Not Brutus', the same HTML output is produced.\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"oh-fouc\">Oh FOUC!<\/h3><p>It gets worse. Since component caching is only used for server-side renders, and since all of our state is stored in Redux, when React mounts in the browser its virtual DOM will <em>not<\/em> match the server-side rendered DOM! React will correct the situation by reconciling in favor of the virtual DOM. The user will experience something like a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Flash_of_unstyled_content\">flash of unstyled content (FOUC)<\/a>. The wrong name will appear for a split-second, and then the correct one will suddenly render!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/fouc.svg\" width=\"650\" alt=\"Diagram of how a flash-of-unstyled-content (FOUC) and SEO problems can manifest when using per-component caching. React Service renders a memoized component with the value 'Brutus', although 'Not Brutus' is the dynamic value which should have been rendered. This mistake is sent to the Monolith, which dutifully includes it in the response to the browser. The browser initially displays 'Brutus', but React detects a discrepancy between the virtual and real DOMs, and re-renders the component, finally and correctly displaying 'Not Brutus', albeit after a FOUC. Meanwhile, Google receives the cached, server-side rendered, mistaken 'Brutus' response and indexes it forever and ever.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Now imagine that this content is being served to a search engine crawler. When a human looks at the page, they are unlikely to notice the error, because the client-side re-render fixes the issue in the blink of an eye. But search engines will index the incorrect content. We are in danger of shipping serious SEO defects, potentially for long periods of time, with no obvious symptoms.<\/p><h3 id=\"exploding-cache\">Exploding Cache<\/h3><p>It gets even worse. Let\u2019s assume our application has one million users, and that we generate cache keys for the <code class=\"highlighter-rouge\">&lt;Greeting&gt;<\/code> component using both <code class=\"highlighter-rouge\">language<\/code> and <code class=\"highlighter-rouge\">name<\/code> prop values. Accidentally forgetting to switch from memoization to interpolation means that the new <code class=\"highlighter-rouge\">name<\/code> prop, which will be rendered with one million unique values, will generate one million cache entries. The cache has exploded in size!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/cache-explosion.svg\" width=\"500\" alt=\"Illustration huge explosion occurring on Earth due to passing millions of user names to an accidentally memoized component, which inflates memory usage until probably something explodes. Probably.\" referrerpolicy=\"no-referrer\" \/><\/p><p>If this accident exhausts available memory, the service will terminate. This failure will probably sneak up on us, as cache misses don\u2019t all occur at once.<\/p><p>Even if we set a maximum cache size and employ a cache replacement policy \u2014 such as <em>least recently used<\/em> (LRU) \u2014 the cache explosion runs a serious risk of exhausting cache storage. Components that would have been cached are now competing for cache space with all of the other debris. Cache misses will increase, and rendering performance could severely degrade.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/warp-core.svg\" width=\"300\" alt=\"Comic of Star Trek bridge, with someone saying, 'Captain, our cache size has reached critical limits.' The Captain responds, 'Eject the warp core!' The caption reads, 'Don't let this happen to you.'\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"making-the-opposite-mistake\">Making The Opposite Mistake<\/h3><p>Now let\u2019s imagine that we <em>do<\/em> remember to update the caching config, changing the prop strategy to from memoization to interpolation for our <code class=\"highlighter-rouge\">&lt;Greeting&gt;<\/code> component. If we do this, but forget to update the component\u2019s prop usage, we will ship a broken component to production.<\/p><p>Recall that interpolated prop values are merged as-is into the rendered component template. Conditional logic inside a component\u2019s <code class=\"highlighter-rouge\">render()<\/code> method \u2014 such as the selection of a greeting based on the value of the <code class=\"highlighter-rouge\">language<\/code> prop \u2014 will only ever execute <em>once<\/em>. If the first render happens to produce an English greeting, the template will be cached with the English greeting baked-in. For all subsequent renders, the user\u2019s name will be successfully interpolated, but the rest of the greeting will only ever render in English.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/broken-interpolation.svg\" width=\"650\" alt=\"Diagram of interpolated Greeting component with 'language' and 'name' props being rendered for the first time, with values 'en' and 'Brutus', respectively. The 'language' prop value does not appear in the rendered output, but is instead used in a conditional to select either a 'Hello' or 'Bonjour' greeting. The resulting template is '&lt;p&gt;Hello, @2@!&lt;\/p&gt;'. The first interpolation of this template, using values 'en' and 'Brutus', produces the output '&lt;p&gt;Hello, Brutus!&lt;\/p&gt;'. The second interpolation of this template, using values 'fr' and 'Brutus', produces the output '&lt;p&gt;Hello, Brutus!&lt;\/p&gt;' again! This demonstrates how easy it is to introduce subtle bugs when using interpolation.\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"cache-rules-everything-around-me\">Cache Rules Everything Around Me<\/h3><p>No matter which way we look at it, modifying the props of a cached component becomes fraught with danger. The developer must take special care to ensure that caching is correctly implemented for each component. React components experience a lot of churn as new features are added, so there are constant opportunities to make an innocuous change which destroys SEO performance, or destroys rendering performance, or renders incorrect data, or renders private user data for every user, or brings the UI down entirely.<\/p><p>Due to these problems, I\u2019m not comfortable recommending per-component caching as a primary scaling strategy. The speed gains are incredible, and you should consider implementing this style of caching when you have run out of other options. But in my view, the biggest advantage of isomorphic rendering is that it unifies your codebase. Developers no longer need to cope with both client- and server-side logic, and the duplication that arrangement entails. The potential for subtle, pernicious bugs creates the need to think very carefully about both client- and server-side rendering, which is precisely the wasteful paradigm we were trying to get away from.<\/p><h2 id=\"dependencies\">Dependencies<\/h2><h3 id=\"dont-get-hacked\">Don\u2019t Get Hacked<\/h3><p>I would be remiss not to mention the disgustingly cheap performance wins we were able to achieve by keeping our dependencies up to date. Dependencies such as Node.js and React.<\/p><p>It is important to keep your dependencies up to date so that you don\u2019t get hacked. If you\u2019re on the fence about this, just ask Equifax <a href=\"https:\/\/www.nytimes.com\/2017\/09\/14\/business\/equifax-hack-what-we-know.html\">how well that worked out for them<\/a>.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/equifax.svg\" width=\"250\" alt=\"A newscaster sitting at their newscaster's desk reads the following cast of news. 'Equifax revealed that a cyberattack potentially compromised confidential information of 143 million Americans. The breach was open from mid-May to July 29. That was when Equifax first detected it. This security weakness was publicly identified in March and a patch to fix it had been available since then.\" referrerpolicy=\"no-referrer\" \/><\/p><h3 id=\"do-you-like-free-things\">Do You Like Free Things?<\/h3><p><em>But that\u2019s not all!<\/em> If you act now, your dependency upgrades will come with a free <em>performance boost!<\/em><\/p><p>Because we were seeking to improve performance, we became interested in benchmarking upgrades to major dependencies. While your mileage may vary, upgrading from Node 4 to Node 6 decreased our response times by about 20%. Upgrading from Node 6 to Node 8 brought a 30% improvement. Finally, upgrading from React 15 to 16 yielded a 25% improvement. The cumulative effect of these upgrades is to more than <em>double<\/em> our performance, and therefore our service capacity.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/dependencies.svg\" width=\"350\" alt=\"Bar graph entitled 'Response Latency', showing relative improvements in response latency as dependencies are upgraded to newer major versions over time. Upgrading Node 4 + React 15 to Node 6 produced a 20% improvement in response times. Upgrading to Node 8 produced a further 30% improvement. Upgrading Node 8 + React 15 to React 16 produced another 25% improvement. That's a lot of free performance! Thanks, open source buddies!\" referrerpolicy=\"no-referrer\" \/><\/p><p>Profiling your code can be important, as well. But the open source community is a <em>vast<\/em> ocean of talent. Very smart people are working incredibly hard, often for free, to speed up your application for you. They\u2019re standing on the corner of a busy intersection, handing out free performance chocolate bars. Take one, and thank them!<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/free-performance.svg\" width=\"350\" alt=\"Comic of person handing out free performance chocolate bars to a turtle, saying, 'Free performance?' An arrow points out that the turtle is an 'Insecure turtle.' Slow, and without the latest dependencies, also insecure. Get it? GET IT? OH COME ON!!!1\" referrerpolicy=\"no-referrer\" \/><\/p><h2 id=\"isomorphic-rendering\">Isomorphic Rendering<\/h2><h3 id=\"the-browser-as-your-server\">The Browser As Your Server<\/h3><p>Isomorphic rendering is a huge simplicity booster for developers, who for too long have been forced to maintain split templates and logic for both client- and server-side rendering contexts. It also enables a dramatic reduction in server resource consumption, by offloading re-renders onto the web browser. The first page of a user\u2019s browsing session can be rendered server-side, providing a first-render performance boost along with basic SEO. All subsequent page views may then fetch their data from JSON endpoints, rendering exclusively within the browser, and managing browser history via the history API.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/ssr-csr.svg\" width=\"550\" alt=\"Diagram illustrating how isomorphic rendering works. For the first page viewed, the browser sends a request to the Monolith, in this case for the \/home page. The Monolith then requests that a Home component be rendered by the React service, which performs a server-side render and returns the result. The Monolith integrates this result into its response to the browser, and the user can now see the first page of their session. When the user navigates to their next page, the \/search page, our React client-side app notices this, fetches data from the Monolith, and renders the page entirely on the client. Client-side renders are ideally used for every page in the session, except for the first one.\" referrerpolicy=\"no-referrer\" \/><\/p><p>If a typical user session consists of 5 page views, rendering only the first page server-side will reduce your server resource consumption by 80%. Another way to think of this is that it would achieve a 5x increase in server-side rendering capacity. This is a huge win!<\/p><h3 id=\"pairs-of-pages\">Pairs Of Pages<\/h3><p>Evolving toward this capability in a legacy application requires patience. A big-bang rewrite of the front-end, in addition to being incredibly risky, is usually off the table because it is a very expensive prospect. A long-term, incremental strategy is therefore required.<\/p><p>I think it makes sense to conceive of this problem in terms of <em>pairs<\/em> of pages. Imagine a simple, e-commerce website, with home, search results, and individual product pages.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/page-flow.svg\" width=\"330\" alt=\"Diagram of a common web page architecture, with many users starting on the Home page, proceeding to the Search page, which displays search results, and then finally ending up on a Product page.\" referrerpolicy=\"no-referrer\" \/><\/p><p>If you upgrade both the home and search results pages to take advantage of isomorphic rendering, most users will hit the homepage first and can therefore render the search results page entirely within the browser. The same is true for the search results and product page combination.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/strategic-page-pairings.svg\" width=\"380\" alt=\"Diagram of the Home, Search, Product page user flow, with the Home-Search and Search-Product page pairings highlighted. Ideally pages will transition to an isomorphic rendering strategy in adjacent pairs, such as those identified here.\" referrerpolicy=\"no-referrer\" \/><\/p><p>But it\u2019s easy to miss out on these strategic pairings. Let\u2019s say your search results page is where all of the money is made, and so the product team is hesitant to modify it. If we invest our time into improving the home and product pages, making them isomorphic in the process, we won\u2019t see much uptake in client-side rendering. This is because in order to get from the homepage to a product page, most users will navigate <em>through<\/em> a search results page. Because the search results page is not isomorphic, a server-side render will be required. If we\u2019re not careful, it\u2019s easy to perform a kind of inverse <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pareto_principle\">Pareto optimization<\/a>, investing 80% of the resources to achieve only 20% of the gains.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/inefficient-pairing.svg\" width=\"300\" alt=\"Diagram of the Home, Search, Product page user flow, with the Home and Product pages having transitioned to an isomorphic rendering strategy, while the Search page remains server-side only. Since few users jump from the Home page to a Product page, client-side rendering cannot do much to reduce our server-side rendering load. Since many users progress from the Home page through Search to a Product page, these users are forced to experience a server-side render and full page refresh when transitioning into and out of the Search page.\" referrerpolicy=\"no-referrer\" \/><\/p><h2 id=\"the-aggregation-of-marginal-gains\">The Aggregation Of Marginal Gains<\/h2><p>It is astonishing how a large number of small improvements, when compounded, can add up to produce one enormous performance boost. I recently learned that the term <em>aggregation of marginal gains<\/em> describes this phenomenon. It is famously associated with Dave Brailsford, head of British Cycling, who <a href=\"https:\/\/hbr.org\/2015\/10\/how-1-performance-improvements-led-to-olympic-gold\">used this philosophy<\/a> to turn the British Cycling team into a dominant force.<\/p><p>It is important to emphasize the <em>compounding<\/em> effect of these gains. If we implement two improvements which, in isolation, double performance, combining them will <em>quadruple<\/em> performance. Various fixed costs and overhead will affect the final result, but in general this principle applies.<\/p><p>Human psychology seems at odds with this approach. We tend to prefer quick wins, and short-term improvements. We tend not to consider a long-term roadmap of improvements in aggregate, and certainly not their compounding effects. These tendencies discourage us from exploring viable strategies. Comparing React server-side rendering to traditional server-rendered templating, React at first seems like it \u201cdoesn\u2019t scale\u201d. But as we layer performance improvement techniques, we can see that we have enormous performance headroom.<\/p><p>How much performance can we gain? And in which order should we pursue these techniques? Ultimately, the exact techniques and their order of implementation will depend on your specific situation. Your mileage may vary. But as a generic starting point from which to plan your journey, I recommend the following approach.<\/p><ol><li>First, upgrade your Node and React dependencies. This is likely the easiest performance win you will achieve. In my experience, upgrading from Node 4 and React 15, to Node 8 and React 16, increased performance by approximately 2.3x.<\/li>\n<li>Double-check your load balancing strategy, and fix it if necessary. This is probably the next-easiest win. While it doesn\u2019t improve average render times, we must always provision for the worst-case scenario, and so reducing 99th percentile response latency counts as a capacity increase in my book. I would conservatively estimate that switching from random to round-robin load balancing bought us a 1.4x improvement in headroom.<\/li>\n<li>Implement a client-side rendering fallback strategy. This is fairly easy if you are already server-side rendering a serialized Redux store. In my experience, this provides a roughly 8x improvement in emergency, elastic capacity. This capability can give you a lot of flexibility to defer other performance upgrades. And even if your performance is fine, it\u2019s always nice to have a safety net.<\/li>\n<li>Implement isomorphic rendering for entire pages, in conjunction with client-side routing. The goal here is to server-side render only the first page in a user\u2019s browsing session. Upgrading a legacy application to use this approach will probably take a while, but it can be done incrementally, and it can be Pareto-optimized by upgrading strategic pairs of pages. All applications are different, but if we assume an average of 5 pages visited per user session, we can increase capacity by 5x with this strategy.<\/li>\n<li>Install per-component caching in low-risk areas. I have already outlined the pitfalls of this caching strategy, but certain rarely modified components, such as the page header, navigation, and footer, provide a better risk-to-reward ratio. I saw a roughly 1.4x increase in capacity when a handful of rarely modified components were cached.<\/li>\n<li>Finally, for situations requiring both maximum risk and maximum reward, cache as many components as possible. A 10x or greater improvement in capacity is easily achievable with this approach. It does, however, require very careful attention to detail.<\/li>\n<\/ol><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/1288x.svg\" width=\"450\" alt=\"Bar graph of relative capacity increase provided by compounding each successive technique. Baseline is 1x. Upgrading Dependencies produces another 2.3x improvement for 2.3x total capacity. Fixing Load Balancing produces another 1.4x improvement for 3.2x total capacity. Client-Side Fallback produces another 8x improvement for 25x total capacity. Isomorphic rendering produces another 5x improvement for 128x total capacity. Some Caching produces another 1.4x improvement for 180x total capacity. Maximum Caching produces a 10x improvement on top of all other techniques, for 1288x total capacity. All bars are rendered as precariously stacked React service instances. Maximum Caching bar contains a stick figure holding up some of the instances, in a nod to the ongoing maintenance that technique requires.\" referrerpolicy=\"no-referrer\" \/><\/p><p>Given reasonable estimates, when we compound these improvements, we can achieve an astounding 1288x improvement in total capacity! Your mileage will of course vary, but a three orders of magnitude improvement can easily change your technology strategy.<\/p><h2 id=\"all-your-servers-are-belong-to-redux\">All Your Servers Are Belong To Redux<\/h2><p>I feel a lot better about the viability of React server-side rendering, now that I have waded through the fires and come out with only minor burns. As with virtually everything in the world of technology, exploring an approach for the first time carries the bulk of the cost. But even if you leave it to somebody else to blaze the trails, there will still be a first time for <em>you<\/em>. You can\u2019t escape that. Waiting for other people to perfect the backstroke is a very slow way to learn how to swim.<\/p><p><img src=\"https:\/\/arkwright.github.io\/images\/scaling-react-server-side-rendering\/swimming.svg\" width=\"550\" alt=\"Comic of a beach, with a body of water in the background. Lots of people are swimming and playing in the water, while a single person stands on the beach and addresses the reader: 'I'm not sure that swimming is production-ready yet.' The sun shines in the sky, wearing sunglasses, as suns do.\" referrerpolicy=\"no-referrer\" \/><\/p><p>I know so much more about this topic than I did when I first started. This isn\u2019t to say that my next attempt will be devoid of problems, but knowing exactly where many trap doors and power-ups lie could easily make the next project an order of magnitude cheaper. I\u2019m looking forward to a world where, rather than something to aspire towards, component-oriented, isomorphic architecture is the standard approach. We\u2019re getting there!<\/p>","created_at":"2019-12-31T13:55:22+0000","updated_at":"2019-12-31T13:55:36+0000","annotations":[],"mimetype":"text\/html","reading_time":36,"domain_name":"arkwright.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":1075,"title":"raxod502\/straight.el","url":"https:\/\/github.com\/raxod502\/straight.el","content":"<p><strong>straight.el<\/strong>: next-generation, purely functional package manager for the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/\" rel=\"nofollow\">Emacs<\/a> hacker.<\/p><p><strong>Please check out the <a href=\"#faq\">FAQ<\/a> and <a href=\"#news\">news<\/a> :)<\/strong><\/p><p><a href=\"https:\/\/travis-ci.org\/raxod502\/straight.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e1b9504ca74a584011596e46a78873de1fa1784a\/68747470733a2f2f7472617669732d63692e6f72672f7261786f643530322f73747261696768742e656c2e7376673f6272616e63683d646576656c6f70\" alt=\"Build status\" data-canonical-src=\"https:\/\/travis-ci.org\/raxod502\/straight.el.svg?branch=develop\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/gitter.im\/raxod502\/straight.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/82630e77f45b173fed5fc2dbb99461abb2c402e7\/68747470733a2f2f6261646765732e6769747465722e696d2f7261786f643530322f73747261696768742e656c2e737667\" alt=\"Gitter chat\" data-canonical-src=\"https:\/\/badges.gitter.im\/raxod502\/straight.el.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><ul><li><a href=\"#features\">Features<\/a><\/li>\n<li><a href=\"#guiding-principles\">Guiding principles<\/a><\/li>\n<li><a href=\"#getting-started\">Getting started<\/a>\n<ul><li><a href=\"#debugging\">Debugging<\/a><\/li>\n<\/ul><ul><li><a href=\"#install-packages\">Install packages<\/a><\/li>\n<li><a href=\"#but-what-about-my-fork-of-obscure-el-package\">But what about my fork of (obscure .el package)?<\/a><\/li>\n<li><a href=\"#integration-with-use-package\">Integration with <code>use-package<\/code><\/a><\/li>\n<li><a href=\"#edit-packages-locally\">Edit packages locally<\/a><\/li>\n<li><a href=\"#automatic-repository-management\">Automatic repository management<\/a><\/li>\n<li><a href=\"#configuration-reproducibility\">Configuration reproducibility<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#conceptual-overview\">Conceptual overview<\/a>\n<ul><li><a href=\"#tldr\">TL;DR<\/a><\/li>\n<li><a href=\"#what-is-a-package\">What is a package?<\/a><\/li>\n<li><a href=\"#where-do-packages-come-from\">Where do packages come from?<\/a><\/li>\n<li><a href=\"#what-does-this-look-like-on-disk\">What does this look like on disk?<\/a><\/li>\n<li><a href=\"#where-do-repositories-come-from\">Where do repositories come from?<\/a><\/li>\n<li><a href=\"#what-does-it-mean-to-load-a-package\">What does it mean to load a package?<\/a><\/li>\n<li><a href=\"#where-do-recipes-come-from\">Where do recipes come from?<\/a><\/li>\n<li><a href=\"#what-happens-when-i-call-straight-use-package\">What happens when I call <code>straight-use-package<\/code>?<\/a><\/li>\n<li><a href=\"#what-does-it-mean-to-register-a-package\">What does it mean to register a package?<\/a><\/li>\n<li><a href=\"#how-does-straightel-know-when-to-rebuild-packages\">How does <code>straight.el<\/code> know when to rebuild packages?<\/a><\/li>\n<li><a href=\"#how-does-straightel-know-what-packages-are-installed\">How does <code>straight.el<\/code> know what packages are installed?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparison-to-other-package-managers\">Comparison to other package managers<\/a>\n<ul><li><a href=\"#tldr-1\">TL;DR<\/a><\/li>\n<li><a href=\"#comparison-to-packageel\">Comparison to <code>package.el<\/code><\/a>\n<ul><li><a href=\"#advantages-of-straightel\">Advantages of <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#advantages-of-packageel\">Advantages of <code>package.el<\/code><\/a><\/li>\n<li><a href=\"#additional-notes\">Additional notes<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparison-to-quelpa\">Comparison to Quelpa<\/a>\n<ul><li><a href=\"#advantages-of-straightel-1\">Advantages of <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#advantages-of-quelpa\">Advantages of Quelpa<\/a><\/li>\n<li><a href=\"#additional-notes-1\">Additional notes<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparison-to-cask\">Comparison to Cask<\/a>\n<ul><li><a href=\"#advantages-of-straightel-2\">Advantages of <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#advantages-of-cask\">Advantages of Cask<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparison-to-el-get\">Comparison to el-get<\/a>\n<ul><li><a href=\"#advantages-of-straightel-3\">Advantages of <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#advantages-of-el-get\">Advantages of el-get<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparison-to-borg\">Comparison to Borg<\/a>\n<ul><li><a href=\"#advantages-of-straightel-4\">Advantages of <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#advantages-of-borg\">Advantages of Borg<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#comparison-to-the-manual-approach\">Comparison to the manual approach<\/a>\n<ul><li><a href=\"#advantages-of-straightel-5\">Advantages of <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#advantages-of-the-manual-approach\">Advantages of the manual approach<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-manual\">User manual<\/a>\n<ul><li><a href=\"#bootstrapping-straightel\">Bootstrapping <code>straight.el<\/code><\/a><\/li>\n<li><a href=\"#installing-packages-programmatically\">Installing packages programmatically<\/a>\n<ul><li><a href=\"#installing-with-a-custom-recipe\">Installing with a custom recipe<\/a><\/li>\n<li><a href=\"#additional-arguments-to-straight-use-package\">Additional arguments to <code>straight-use-package<\/code><\/a><\/li>\n<li><a href=\"#variants-of-straight-use-package\">Variants of <code>straight-use-package<\/code><\/a><\/li>\n<li><a href=\"#customizing-when-packages-are-built\">Customizing when packages are built<\/a>\n<ul><li><a href=\"#custom-or-manual-modification-detection\">Custom or manual modification detection<\/a><\/li>\n<li><a href=\"#summary-of-options-for-package-modification-detection\">Summary of options for package modification detection<\/a>\n<ul><li><a href=\"#find-at-startup\"><code>find-at-startup<\/code><\/a><\/li>\n<li><a href=\"#check-on-save\"><code>check-on-save<\/code><\/a><\/li>\n<li><a href=\"#watch-files\"><code>watch-files<\/code><\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#customizing-how-packages-are-built\">Customizing how packages are built<\/a><\/li>\n<li><a href=\"#customizing-how-packages-are-made-available\">Customizing how packages are made available<\/a><\/li>\n<li><a href=\"#hooks-run-by-straight-use-package\">Hooks run by <code>straight-use-package<\/code><\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#the-recipe-format\">The recipe format<\/a>\n<ul><li><a href=\"#version-control-backends\">Version-control backends<\/a><\/li>\n<li><a href=\"#git-backend\">Git backend<\/a>\n<ul><li><a href=\"#deprecated-upstream-keyword\">Deprecated <code>:upstream<\/code> keyword<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#recipe-lookup\">Recipe lookup<\/a>\n<ul><li><a href=\"#updating-recipe-repositories\">Updating recipe repositories<\/a><\/li>\n<li><a href=\"#customizing-recipe-repositories\">Customizing recipe repositories<\/a>\n<ul><li><a href=\"#gnu-elpa\">GNU ELPA<\/a><\/li>\n<li><a href=\"#emacsmirror\">Emacsmirror<\/a><\/li>\n<li><a href=\"#defining-new-recipe-repositories\">Defining new recipe repositories<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#overriding-recipes\">Overriding recipes<\/a>\n<ul><li><a href=\"#overriding-the-recipe-for-straightel\">Overriding the recipe for <code>straight.el<\/code><\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#interactive-usage\">Interactive usage<\/a>\n<ul><li><a href=\"#version-control-operations\">Version control operations<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#lockfile-management\">Lockfile management<\/a>\n<ul><li><a href=\"#the-profile-system\">The profile system<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#packages-and-the-init-file\">Packages and the init-file<\/a><\/li>\n<li><a href=\"#using-straightel-to-reproduce-bugs\">Using <code>straight.el<\/code> to reproduce bugs<\/a><\/li>\n<li><a href=\"#integration-with-other-packages\">Integration with other packages<\/a>\n<ul><li><a href=\"#integration-with-use-package-1\">Integration with <code>use-package<\/code><\/a><\/li>\n<li><a href=\"#integration-with-packageel\">\"Integration\" with <code>package.el<\/code><\/a><\/li>\n<li><a href=\"#integration-with-org\">Integration with Org<\/a><\/li>\n<li><a href=\"#integration-with-hydra\">Integration with Hydra<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#miscellaneous\">Miscellaneous<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#developer-manual\">Developer manual<\/a><\/li>\n<li><a href=\"#trivia\">Trivia<\/a>\n<ul><li><a href=\"#comments-and-docstrings\">Comments and docstrings<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#contributing\">Contributing<\/a><\/li>\n<li><a href=\"#faq\">FAQ<\/a>\n<ul><li><a href=\"#my-init-time-got-slower\">My init time got slower<\/a><\/li>\n<li><a href=\"#how-do-i-update-melpa-et-al\">How do I update MELPA et al.?<\/a><\/li>\n<li><a href=\"#the-wrong-version-of-my-package-was-loaded\">The wrong version of my package was loaded<\/a><\/li>\n<li><a href=\"#i-get-could-not-read-usernamepassword-errors\">I get \"could not read username\/password\" errors<\/a><\/li>\n<li><a href=\"#the-interactive-version-control-operations-are-confusing\">The interactive version-control operations are confusing<\/a><\/li>\n<li><a href=\"#how-do-i-pin-package-versions-or-use-only-tagged-releases\">How do I pin package versions or use only tagged releases?<\/a><\/li>\n<li><a href=\"#how-can-i-use-the-built-in-version-of-a-package\">How can I use the built-in version of a package?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#news\">News<\/a>\n<ul><li><a href=\"#july-6-2019\">July 6, 2019<\/a><\/li>\n<li><a href=\"#may-24-2019\">May 24, 2019<\/a><\/li>\n<li><a href=\"#may-22-2019\">May 22, 2019<\/a><\/li>\n<li><a href=\"#may-1-2019\">May 1, 2019<\/a><\/li>\n<li><a href=\"#march-15-2019\">March 15, 2019<\/a><\/li>\n<li><a href=\"#december-22-2018\">December 22, 2018<\/a><\/li>\n<li><a href=\"#september-12-2018\">September 12, 2018<\/a><\/li>\n<li><a href=\"#july-19-2018\">July 19, 2018<\/a><\/li>\n<li><a href=\"#july-12-2018\">July 12, 2018<\/a><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2><ul><li>Install Emacs packages listed on <a href=\"http:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a>, <a href=\"https:\/\/elpa.gnu.org\/\" rel=\"nofollow\">GNU ELPA<\/a>, or <a href=\"https:\/\/emacsmirror.net\/\" rel=\"nofollow\">Emacsmirror<\/a>, or provide your own recipes.<\/li>\n<li>Packages are cloned as Git (or other) repositories, not as opaque tarballs.<\/li>\n<li>Make changes to a package simply by editing its source code, no additional steps required. Contribute upstream just by pushing your changes.<\/li>\n<li>Powerful interactive workflows (with popups \u00e0 la Magit) for performing bulk maintenance on the Git repositories for your packages.<\/li>\n<li>Save and load version lockfiles that ensure 100% reproducibility for your Emacs configuration. Package state is defined entirely by your init-file and (optional) lockfile, with no extra persistent data floating around.<\/li>\n<li>Specify package descriptions using a powerful recipe format that supports everything from <a href=\"https:\/\/github.com\/melpa\/melpa#recipe-format\">MELPA recipes<\/a> and more.<\/li>\n<li><a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a> integration.<\/li>\n<li>Modular: you can install your packages manually and straight.el will load them for you. Or you can also have straight.el install your packages, while you provide the recipes explicitly. Or straight.el can also fetch recipes, if you want. Bulk repository management and package updates are also optional.<\/li>\n<li>Extensible APIs to add new recipe sources and version-control backends.<\/li>\n<li>The cleanest source code you've ever seen. <a href=\"#comments-and-docstrings\">45%<\/a> of <code>straight.el<\/code> is comments and docstrings.<\/li>\n<\/ul><h2><a id=\"user-content-guiding-principles\" class=\"anchor\" aria-hidden=\"true\" href=\"#guiding-principles\"><\/a>Guiding principles<\/h2><ul><li>Init-file and version lockfiles as the sole source of truth. No persistent state kept elsewhere.<\/li>\n<li>100% reproducible package management, accounting for changes in packages, recipe repositories, configuration, and the package manager itself.<\/li>\n<li>No support whatsoever for <code>package.el<\/code>.<\/li>\n<li>Edit packages by editing their code, no extra steps required. Allow for manual version control operations.<\/li>\n<li>Compatibility with MELPA, GNU ELPA, and Emacsmirror.<\/li>\n<li>Trivial to quickly try out a package without permanently installing it.<\/li>\n<li>Good for reproducing an issue with <code>emacs -Q<\/code>.<\/li>\n<\/ul><h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting started<\/h2><blockquote>\n<p><strong>Note: <code>straight.el<\/code> supports a minimum version of Emacs 24.5, and works on macOS, Windows, and most flavors of Linux. You must install <a href=\"https:\/\/git-scm.com\/\" rel=\"nofollow\">Git<\/a> in order to use <code>straight.el<\/code>.<\/strong><\/p>\n<\/blockquote><p>First, place the following bootstrap code in your init-file:<\/p><pre>(defvar bootstrap-version)\n(let ((bootstrap-file\n       (expand-file-name \"straight\/repos\/straight.el\/bootstrap.el\" user-emacs-directory))\n      (bootstrap-version 5))\n  (unless (file-exists-p bootstrap-file)\n    (with-current-buffer\n        (url-retrieve-synchronously\n         \"https:\/\/raw.githubusercontent.com\/raxod502\/straight.el\/develop\/install.el\"\n         'silent 'inhibit-cookies)\n      (goto-char (point-max))\n      (eval-print-last-sexp)))\n  (load bootstrap-file nil 'nomessage))\n<\/pre><p>Here are some variables you may be interested in (some of them must be set <strong>before<\/strong> the bootstrap code runs, if they might affect how <code>straight.el<\/code> itself is loaded):<\/p><ul><li><a href=\"#overriding-the-recipe-for-straightel\"><code>straight-repository-branch<\/code><\/a> -- to get the latest version of <code>straight.el<\/code> from the <code>develop<\/code> branch, rather than the default <code>master<\/code> which is updated less frequently but which is ostensibly more stable.<\/li>\n<li><a href=\"#customizing-when-packages-are-built\"><code>straight-check-for-modifications<\/code><\/a> -- to configure an alternate way for <code>straight.el<\/code> to check for modifications made to package source code, rather than the default (which is 100% reliable, but has a minor cost to startup time).<\/li>\n<li><a href=\"#integration-with-use-package-1\"><code>straight-use-package-by-default<\/code><\/a> -- if you use <a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a>, then this makes each <code>use-package<\/code> form also invoke <code>straight.el<\/code> to install the package, unless otherwise specified.<\/li>\n<li><a href=\"#git-backend\"><code>straight-vc-git-default-protocol<\/code><\/a> -- by default, <code>straight.el<\/code> clones over HTTPS. If you need packages from private Git repositories in your configuration, then you might want to use SSH instead.<\/li>\n<\/ul><p>You should remove any code that relates to <code>package.el<\/code>; for example, references to <code>package-initialize<\/code>, <code>package-archives<\/code>, and (if you're using <a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a>) <code>:ensure<\/code> or <code>use-package-always-ensure<\/code>. While it is technically possible to use both <code>package.el<\/code> and <code>straight.el<\/code> at the same time, there is no real reason to, and it might result in oddities like packages getting loaded more than once.<\/p><h4><a id=\"user-content-debugging\" class=\"anchor\" aria-hidden=\"true\" href=\"#debugging\"><\/a>Debugging<\/h4><p>On macOS, you may receive an error:<\/p><pre>Could not create connection to raw.githubusercontent.com:443\n<\/pre><p>There are two ways to solve this problem. One way is to install a version of Emacs that is linked with GnuTLS. The <a href=\"https:\/\/brew.sh\/\" rel=\"nofollow\">Homebrew<\/a> formula for Emacs was recently updated to link with GnuTLS by default, so you need only do this:<\/p><pre>$ brew upgrade emacs\n<\/pre><p>The other way is to let Emacs use certificates provided by LibreSSL, which you can do by running this command:<\/p><pre>$ brew install gnutls libressl\n<\/pre><p>And adding this to your init-file, <em>before<\/em> the bootstrap snippet:<\/p><pre>(with-eval-after-load 'gnutls\n  (add-to-list 'gnutls-trustfiles \"\/usr\/local\/etc\/libressl\/cert.pem\"))\n<\/pre><h3><a id=\"user-content-install-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-packages\"><\/a>Install packages<\/h3><p>Out of the box, you can install any package listed on <a href=\"http:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a>, <a href=\"https:\/\/elpa.gnu.org\/\" rel=\"nofollow\">GNU ELPA<\/a>, or <a href=\"https:\/\/emacsmirror.net\/\" rel=\"nofollow\">Emacsmirror<\/a>, which is to say any package in existence. (Although MELPA is used as a package listing, packages are installed by cloning their Git repositories rather than by downloading tarballs like <code>package.el<\/code> does.) To install a package temporarily (until you restart Emacs), run <code>M-x straight-use-package<\/code> and select the package you want. To install a package permanently, place a call to <code>straight-use-package<\/code> in your init-file, like:<\/p><pre>(straight-use-package 'el-patch)\n<\/pre><p>Note that installing a package will activate all of its autoloads, but it will not actually <code>require<\/code> the features provided by the package. This means that you might need to use <code>require<\/code> or <code>autoload<\/code> for some antiquated packages that do not properly declare their autoloads.<\/p><p>To learn more, see the documentation on <a href=\"#installing-packages-programmatically\">the package lifecycle<\/a>.<\/p><h3><a id=\"user-content-but-what-about-my-fork-of-obscure-el-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#but-what-about-my-fork-of-obscure-el-package\"><\/a>But what about my fork of (obscure .el package)?<\/h3><p>Instead of passing just a package name to <code>straight-use-package<\/code>, you can pass a list (\"recipe\"). You can see the default recipe for any given package by running <code>M-x straight-get-recipe<\/code>. For example, the recipe for <code>el-patch<\/code> is:<\/p><pre>(el-patch :type git :host github :repo \"raxod502\/el-patch\")\n<\/pre><p>So, if you have forked <code>el-patch<\/code> and you want to use your fork instead of the upstream, do:<\/p><pre>(straight-use-package\n '(el-patch :type git :host github :repo \"your-name\/el-patch\"))\n<\/pre><p>In fact, <code>straight.el<\/code> has explicit support for using a forked package, since this is so common:<\/p><pre>(straight-use-package\n '(el-patch :type git :host github :repo \"raxod502\/el-patch\"\n            :fork (:host github\n                   :repo \"your-name\/el-patch\")))\n<\/pre><p>You may also omit the <code>:type git<\/code> if you leave <code>straight-default-vc<\/code> at its default value of <code>git<\/code>.<\/p><p>To learn more, see the documentation on <a href=\"#the-recipe-format\">the recipe format<\/a>.<\/p><h3><a id=\"user-content-integration-with-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-use-package\"><\/a>Integration with <code>use-package<\/code><\/h3><p><a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a> is a macro that provides convenient syntactic sugar for many common tasks related to installing and configuring Emacs packages. Of course, it does not actually install the packages, but instead defers to a package manager, like <code>straight.el<\/code> (which comes with <code>use-package<\/code> integration by default).<\/p><p>To use <code>use-package<\/code>, first install it with <code>straight.el<\/code>:<\/p><pre>(straight-use-package 'use-package)\n<\/pre><p>Now <code>use-package<\/code> will use <code>straight.el<\/code> to automatically install missing packages if you provide <code>:straight t<\/code>:<\/p><pre>(use-package el-patch\n  :straight t)\n<\/pre><p>You can still provide a custom recipe for the package:<\/p><pre>(use-package el-patch\n  :straight (el-patch :type git :host github :repo \"raxod502\/el-patch\"\n                      :fork (:host github\n                             :repo \"your-name\/el-patch\")))\n<\/pre><p>Specifying <code>:straight t<\/code> is unnecessary if you set <code>straight-use-package-by-default<\/code> to a non-nil value. (Note that the variable <code>use-package-always-ensure<\/code> is associated with <code>package.el<\/code>, and you should not use it with <code>straight.el<\/code>.)<\/p><p>To learn more, see the documentation on <a href=\"#integration-with-use-package-1\"><code>straight.el<\/code>'s <code>use-package<\/code> integration<\/a>.<\/p><h3><a id=\"user-content-edit-packages-locally\" class=\"anchor\" aria-hidden=\"true\" href=\"#edit-packages-locally\"><\/a>Edit packages locally<\/h3><p>One of the biggest strengths of <code>straight.el<\/code> is that editing packages locally is trivial. You literally just edit the files (<code>find-function<\/code> and friends all work as you would expect). Packages will be automatically rebuilt if necessary when Emacs next starts up.<\/p><p>You can even commit your changes and push or pull to various remotes using Git. You have complete control over your packages' Git repositories.<\/p><p>To learn more, see the documentation on <a href=\"#installing-packages-programmatically\">the package lifecycle<\/a>.<\/p><h3><a id=\"user-content-automatic-repository-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#automatic-repository-management\"><\/a>Automatic repository management<\/h3><p>While being able to make arbitrary changes to your packages is very powerful, it can also get tiring to keep track of the all those changes. For this reason, <code>straight.el<\/code> provides a suite of powerful interactive workflows to perform bulk operations on your packages.<\/p><ul><li>\n<p>To restore each package to its canonical state (a clean working directory with the main branch checked out, and the remotes set correctly), run <code>M-x straight-normalize-package<\/code> or <code>M-x straight-normalize-all<\/code>.<\/p>\n<\/li>\n<li>\n<p>To fetch from each package's configured remote, run <code>M-x straight-fetch-package-and-deps<\/code> or <code>M-x straight-fetch-all<\/code>; to also fetch from the upstream for forked packages, supply a prefix argument.<\/p>\n<\/li>\n<li>\n<p>To merge changes from each package's configured remote, run <code>M-x straight-merge-package-and-deps<\/code> or <code>M-x straight-merge-all<\/code>; to also merge from the upstream for forked packages, supply a prefix argument.<\/p>\n<\/li>\n<li>\n<p>To push all local changes to each package's configured remote, run <code>M-x straight-push-package<\/code> or <code>M-x straight-push-all<\/code>.<\/p>\n<\/li>\n<\/ul><p>All of these commands are highly interactive and ask you before making any changes. At any point, you can stop and perform manual operations with Magit or other tools in a recursive edit.<\/p><p>To learn more, see the documentation on <a href=\"#version-control-operations\">bulk repository management<\/a>.<\/p><h3><a id=\"user-content-configuration-reproducibility\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration-reproducibility\"><\/a>Configuration reproducibility<\/h3><p>To save the currently checked out revisions of all of your packages, run <code>M-x straight-freeze-versions<\/code>. The resulting file (<code>~\/.emacs.d\/straight\/versions\/default.el<\/code>), together with your init-file, perfectly define your package configuration. Keep your version lockfile checked into version control; when you install your Emacs configuration on another machine, the versions of packages specified in your lockfile will automatically be checked out after the packages are installed. You can manually revert all packages to the revisions specified in the lockfile by running <code>M-x straight-thaw-versions<\/code>.<\/p><p>To learn more, see the documentation on <a href=\"#lockfile-management\">version lockfiles<\/a>.<\/p><h2><a id=\"user-content-conceptual-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#conceptual-overview\"><\/a>Conceptual overview<\/h2><p>This section describes, at a high level, how the different mechanisms in <code>straight.el<\/code> play together. This illustrates how <code>straight.el<\/code> manages to accomplish all of its <a href=\"#guiding-principles\">guiding principles<\/a>.<\/p><h3><a id=\"user-content-tldr\" class=\"anchor\" aria-hidden=\"true\" href=\"#tldr\"><\/a>TL;DR<\/h3><p><code>straight.el<\/code> operates by cloning Git repositories and then symlinking files into Emacs' load path. The collection of symlinked files constitutes the package, which is defined by its recipe. The recipe also describes which local repository to link the files from, and how to clone that repository, if it is absent.<\/p><p>When you call <code>straight-use-package<\/code>, the recipe you provide is registered with <code>straight.el<\/code> for future reference. Then the package's repository is cloned if it is absent, the package is rebuilt if its files have changed since the last build (as determined by <code>find(1)<\/code>), and its autoloads are evaluated.<\/p><p>You can also provide only a package name, in which case the recipe will be looked up in one of several configurable recipe repositories, which are just packages themselves (albeit with the build step disabled).<\/p><p><code>straight.el<\/code> determines which packages are installed solely by how and when <code>straight-use-package<\/code> is invoked in your init-file.<\/p><h3><a id=\"user-content-what-is-a-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-is-a-package\"><\/a>What is a package?<\/h3><p>A <em>package<\/em> is a collection of Emacs Lisp (and possibly other) files. The most common case is just a single <code>.el<\/code> file, but some packages have many <code>.el<\/code> files, and some even have a directory structure.<\/p><p>Note that a package is defined only as a collection of files. It doesn't necessarily correspond to a Git repository, or an entry on MELPA, or anything like that. Frequently there is a relationship between all of these concepts, but that relationship does not always have to be direct or one-to-one.<\/p><p>A package also has a name, which must be unique. This is the name that is used for the folder holding the package's files. It is frequently the same as the name of a Git repository, or an entry on MELPA, but again this does not have to be the case.<\/p><h3><a id=\"user-content-where-do-packages-come-from\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-do-packages-come-from\"><\/a>Where do packages come from?<\/h3><p>If you really wanted all of your packages to be unambiguously defined, you could just copy and paste all of their files into version control. But that would defeat the purpose of using a package manager like <code>straight.el<\/code>. In <code>straight.el<\/code>, packages are defined by two sources of information:<\/p><ul><li>a <em>local repository<\/em><\/li>\n<li>a <em>build recipe<\/em><\/li>\n<\/ul><p>The local repository is just a directory containing some files. Of course, it also has a name, which may or may not be the same as the package's name. Frequently, the local repository is also a Git repository, but this is not necessary.<\/p><p>The build recipe is not a literal data structure. It is a concept that represents a certain subset of the package's recipe. Specifically, the <code>:files<\/code>, <code>:local-repo<\/code>, and <code>:no-build<\/code> keywords.<\/p><p>To transform this <em>information<\/em> into an actual package that Emacs can load, <code>straight.el<\/code> <em>builds<\/em> the package. This means that some symbolic links are created in the package's directory that point back into the local repository's directory. Exactly how these symlinks are created is determined by the <code>:files<\/code> directive, and which local repository the symlinks point to is determined by the <code>:local-repo<\/code> directive.<\/p><p>After the symlinks are created, the resulting files are byte-compiled, and their autoloads are generated and written into a file in the package's directory.<\/p><p>If <code>:no-build<\/code> is specified, however, this entire process is skipped. This mechanism is used for recipe repositories.<\/p><h3><a id=\"user-content-what-does-this-look-like-on-disk\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-does-this-look-like-on-disk\"><\/a>What does this look like on disk?<\/h3><p>The local repositories are kept in <code>~\/.emacs.d\/straight\/repos<\/code>, and the built packages are kept in <code>~\/.emacs.d\/straight\/build<\/code>. If you have initialized <code>straight.el<\/code> and loaded package <code>el-patch<\/code>, then your <code>~\/.emacs.d\/straight<\/code> directory will look roughly like this (some irrelevant details have been omitted for pedagogical purposes):<\/p><pre>straight\n\u251c\u2500\u2500 build\n\u2502   \u251c\u2500\u2500 el-patch\n\u2502   \u2502   \u251c\u2500\u2500 el-patch-autoloads.el\n\u2502   \u2502   \u251c\u2500\u2500 el-patch.el -&gt; ~\/.emacs.d\/straight\/repos\/el-patch\/el-patch.el\n\u2502   \u2502   \u2514\u2500\u2500 el-patch.elc\n\u2502   \u2514\u2500\u2500 straight\n\u2502       \u251c\u2500\u2500 straight-autoloads.el\n\u2502       \u251c\u2500\u2500 straight.el -&gt; ~\/.emacs.d\/straight\/repos\/straight.el\/straight.el\n\u2502       \u2514\u2500\u2500 straight.elc\n\u2514\u2500\u2500 repos\n    \u251c\u2500\u2500 el-patch\n    \u2502   \u251c\u2500\u2500 CHANGELOG.md\n    \u2502   \u251c\u2500\u2500 LICENSE.md\n    \u2502   \u251c\u2500\u2500 README.md\n    \u2502   \u2514\u2500\u2500 el-patch.el\n    \u2514\u2500\u2500 straight.el\n        \u251c\u2500\u2500 LICENSE.md\n        \u251c\u2500\u2500 Makefile\n        \u251c\u2500\u2500 README.md\n        \u251c\u2500\u2500 bootstrap.el\n        \u251c\u2500\u2500 install.el\n        \u2514\u2500\u2500 straight.el\n<\/pre><p>As you can see, the package names are <code>el-patch<\/code> and <code>straight<\/code>. While <code>el-patch<\/code> is built from a local repository of the same name, <code>straight<\/code> is built from a local repository by the name <code>straight.el<\/code>. Also note that only <code>.el<\/code> files are symlinked, since only they are relevant to Emacs.<\/p><h3><a id=\"user-content-where-do-repositories-come-from\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-do-repositories-come-from\"><\/a>Where do repositories come from?<\/h3><p>Local repositories provide a way to define packages without specifying the contents of all of their files explicitly. But that's not helpful without a higher-level way to define local repositories without specifying the contents of all of <em>their<\/em> files. In <code>straight.el<\/code>, local repositories are defined by two sources of information:<\/p><ul><li>a <em>fetch recipe<\/em><\/li>\n<li>the <em>version lockfiles<\/em><\/li>\n<\/ul><p>The fetch recipe is, like the build recipe, a concept representing a certain subset of the package's overall recipe. The situation is more interesting here because <code>straight.el<\/code> supports multiple version-control backends. The version-control backend specified by the fetch recipe is determined by the <code>:type<\/code> directive (which defaults to <code>straight-default-vc<\/code>). Each version-control backend then accepts some set of additional directives. For example, the <code>git<\/code> backend accepts:<\/p><ul><li><code>:repo<\/code><\/li>\n<li><code>:host<\/code><\/li>\n<li><code>:branch<\/code><\/li>\n<li><code>:nonrecursive<\/code><\/li>\n<li><code>:fork<\/code><\/li>\n<\/ul><p>If a local repository is not present, then its fetch recipe describes how to obtain it. This is done using the <code>straight-vc-clone<\/code> function, which delegates to one of the backend implementations of the <code>clone<\/code> operation, according to <code>:type<\/code>. (The option <code>:type built-in<\/code> is a special case that results in all version-control operations for the package being ignored.)<\/p><p>However, even with a particular repository source specified, there is still the question of which version of the repository to use. This is where the version lockfiles come in. When a local repository is cloned, the version lockfiles are searched to see if there is a particular commit specified for that local repository's name. If so, that commit is checked out. (For the <code>git<\/code> backend, commits are 40-character strings representing SHA-1 hashes, but the representation of a commit identifier could be different across different backends.)<\/p><p>The <code>straight-freeze-versions<\/code> and <code>straight-thaw-versions<\/code> methods also use backend-delegating methods; in this case, they are <code>straight-vc-get-commit<\/code> and <code>straight-vc-check-out-commit<\/code>.<\/p><p>The fetch recipe and version lockfiles, together with the configuration options for <code>straight.el<\/code>, precisely define the state of a local repository. Of course, you may make any changes you want to the local repository. But this information defines a \"canonical\" state that you may revert to at any time.<\/p><p>When this information is combined with the build recipe, <code>straight.el<\/code> is able to construct canonical, universal versions of your Emacs packages that will be the same everywhere and forever.<\/p><p>Note that you do not have to provide fetch recipes or version lockfiles. You may manage your local repositories manually, if you wish, although this has obvious disadvantages in terms of repeatability and maintainability.<\/p><h3><a id=\"user-content-what-does-it-mean-to-load-a-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-does-it-mean-to-load-a-package\"><\/a>What does it mean to load a package?<\/h3><p>A prerequisite to loading a package is making sure the package has been built. After that is done, loading the package means adding its directory to the load path and evaluating its autoloads file.<\/p><p>Adding the directory to the load path means that you can use <code>require<\/code> to load the package's files. Note that <code>straight.el<\/code> does not do this for you, since loading packages immediately is usually not necessary and it immensely slows down Emacs startup.<\/p><p>Evaluating the autoloads file means that calling the functions that are defined in the autoloads file will automatically <code>require<\/code> the files that define those functions. All modern packages define their functions in autoloads and are designed to be loaded on-demand when those functions are called. Antiquated packages may need you to explicitly define autoloads, or to just <code>require<\/code> the package right away.<\/p><h3><a id=\"user-content-where-do-recipes-come-from\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-do-recipes-come-from\"><\/a>Where do recipes come from?<\/h3><p><code>straight-use-package<\/code> does not require an actual recipe. You can just give it a package name, and it will look up the recipe. This is done using <em>recipe repositories<\/em>. Recipe repositories are set up as a swappable backend system, much like the version-control backend system.<\/p><p>A recipe repository consists of four parts:<\/p><ul><li>a fetch recipe for the local repository (this will typically include the <code>:no-build<\/code> directive, since recipe repositories usually do not need to be built)<\/li>\n<li>a function that, provided the local repository is already available, returns a list of all packages that have recipes in the recipe repository<\/li>\n<li>a function that, given a package name, returns the recipe for that package, or nil if the recipe repository does not provide a recipe for the package<\/li>\n<li>an entry in <code>straight-recipe-repositories<\/code> indicating that the recipe provided actually corresponds to a recipe repository (otherwise it would just be a regular package)<\/li>\n<\/ul><p>Note that recipe repositories are implemented as regular packages! This means that all the usual package management operations work on them as well. It also means that you use <code>straight-use-package<\/code> to register them (although typically you will provide arguments to <code>straight-use-package<\/code> so that the recipe repository is only registered, and not cloned until it is needed; see <a href=\"#what-happens-when-i-call-straight-use-package\">the section on <code>straight-use-package<\/code><\/a>).<\/p><p>If you give <code>straight-use-package<\/code> just a package name, then each recipe repository in <code>straight-recipe-repositories<\/code> is checked for a recipe for that package. Once one is found, it is used. Otherwise, an error is signaled (unless the package is built-in to Emacs, according to <code>package.el<\/code>).<\/p><p>Note that <code>straight.el<\/code> uses its own recipe format which is similar, but not identical, to the one used by MELPA. The recipe repository backends abstract over the formatting differences in different recipe sources to translate recipes into the uniform format used by <code>straight.el<\/code>. When you run <code>M-x straight-get-recipe<\/code>, the translated recipe is what is returned.<\/p><h3><a id=\"user-content-what-happens-when-i-call-straight-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-happens-when-i-call-straight-use-package\"><\/a>What happens when I call <code>straight-use-package<\/code>?<\/h3><p>There are three actions that <code>straight-use-package<\/code> can take:<\/p><ul><li>Register a package's recipe with <code>straight.el<\/code>.<\/li>\n<li>Clone a package's local repository, if it is missing.<\/li>\n<li>Build a package, if it has been changed since the last time it was built, and load it.<\/li>\n<\/ul><p>These actions must be performed in order. Depending on the arguments you pass to <code>straight-use-package<\/code>, one, two, or all three may be performed.<\/p><p>The normal case is to do all three. The fetch recipe is only required if the local repository is actually missing, but the build recipe is always required.<\/p><p>Deferred installation can be accomplished by telling <code>straight-use-package<\/code> to stop if the local repository is not already available. The deferred installation can be triggered by invoking <code>straight-use-package<\/code> again, but telling it to go ahead and clone the repository (this is the default behavior). Because <code>straight-use-package<\/code> already registered the package's recipe the first time, you don't have to provide it again.<\/p><p>In some extraordinary circumstances (such as when <code>straight.el<\/code> is bootstrapping its own installation), it may be desirable to clone a package's local repository if it is missing, but to stop before building and loading the package. This can also be done by <code>straight-use-package<\/code>.<\/p><h3><a id=\"user-content-what-does-it-mean-to-register-a-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-does-it-mean-to-register-a-package\"><\/a>What does it mean to register a package?<\/h3><p>Package registration is the first action taken by <code>straight-use-package<\/code>, before building and cloning. First, if only a package name was provided to <code>straight-use-package<\/code>, a recipe is obtained from the configured recipe repositories. Next, the resulting recipe is recorded in various caches.<\/p><p>This is important, since it allows for several things to happen:<\/p><ul><li>if you later want to perform another operation on the package using <code>straight.el<\/code>, you do not need to provide the recipe again<\/li>\n<li>if you use a custom recipe for Package A, and Package B requires Package A as a dependency, your custom recipe is remembered and re-used when Package A is used as a dependency, to avoid conflicts.<\/li>\n<li>when multiple packages are built from the same local repository, and you have specified a custom fetch recipe for one of those packages, <code>straight.el<\/code> can intelligently merge that fetch recipe into the automatically retrieved recipes of dependencies, in order to avoid conflicts.<\/li>\n<li><code>straight.el<\/code> knows which packages you have installed, if you want to perform interactive operations on them.<\/li>\n<li>if you accidentally provide two different recipes for the same package, <code>straight.el<\/code> can issue a helpful warning, since this may lead to surprising behavior.<\/li>\n<\/ul><h3><a id=\"user-content-how-does-straightel-know-when-to-rebuild-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-does-straightel-know-when-to-rebuild-packages\"><\/a>How does <code>straight.el<\/code> know when to rebuild packages?<\/h3><p>When you request for <code>straight.el<\/code> to load a package (using <code>straight-use-package<\/code>), it first checks if the package needs to be rebuilt. This means that some of the files in its local repository have been modified since the last time the package was built. <code>straight.el<\/code> uses an optimized <code>find(1)<\/code> command to check for package modifications, and it uses some caching mechanisms to perform bulk <code>find(1)<\/code> operations on multiple packages, to speed up these checks (although it never performs optimizations that may result in erroneous behavior).<\/p><p>This check occurs during Emacs init, when your init-file makes calls to <code>straight-use-package<\/code>. You may notice a significant delay on the first <code>straight-use-package<\/code> call, because this is when <code>straight.el<\/code> performs a bulk <code>find(1)<\/code> call and caches the results for later usage (this speeds up init considerably). The total delay is likely to be on the order of 100ms for a double-digit number of packages.<\/p><p>The rebuild detection system is what allows for you to make changes to packages whenever you would like, without performing any additional operations.<\/p><p>(Packages are also rebuilt when their recipes change, of course.)<\/p><h3><a id=\"user-content-how-does-straightel-know-what-packages-are-installed\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-does-straightel-know-what-packages-are-installed\"><\/a>How does <code>straight.el<\/code> know what packages are installed?<\/h3><p><code>straight.el<\/code> does not require you to declare a central list of packages anywhere, like Cask does. Instead, it determines what packages are to be loaded implicitly, by your invocations of <code>straight-use-package<\/code> during Emacs initialization. Furthermore, <code>straight.el<\/code> allows you to install packages after initialization using <code>M-x straight-install-package<\/code> (or even by evaluating <code>straight-use-package<\/code> forms). However, <code>straight.el<\/code> still provides advanced features such as bulk package management and version locking. This creates some interesting challenges which other package managers do not have to deal with.<\/p><p><code>straight.el<\/code> solves these problems using a concept called <em>transactions<\/em>. The operation of the transaction system is mostly transparent to the user, at least in recent versions of <code>straight.el<\/code>. Basically, it provides a way for <code>straight.el<\/code> to keep track of what happens within a single user operation (e.g. evaluate a buffer of <code>straight-use-package<\/code> calls, or load the init-file).<\/p><p><code>straight.el<\/code> uses the transaction system to keep track of what packages you request in your init-file. If you invoke <code>straight-use-package<\/code> interactively, then this invalidates that information, since you have now requested a package that is not in your init-file. For this reason, if you have invoked <code>straight-use-package<\/code> interactively, running <code>M-x straight-freeze-versions<\/code> will prompt you to first reload your init-file.<\/p><h2><a id=\"user-content-comparison-to-other-package-managers\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-other-package-managers\"><\/a>Comparison to other package managers<\/h2><p>(Disclaimer: while I try to be as objective and comprehensive as possible here, I'm obviously biased. Please submit corrections if I have unfairly disparaged your favorite package manager!)<\/p><p>There are many package managers for Emacs, ranging from simple scripts to download files from EmacsWiki to full-featured package management solutions like <code>straight.el<\/code>. Here are the most feature-rich alternatives to <code>straight.el<\/code>:<\/p><ul><li><a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Packages.html\" rel=\"nofollow\"><code>package.el<\/code><\/a>: de facto standard, bundled with Emacs.<\/li>\n<li><a href=\"https:\/\/github.com\/quelpa\/quelpa\">Quelpa<\/a>: allows you to use external sources like GitHub with <code>package.el<\/code>. Essentially a local <a href=\"http:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a>: another <code>package.el<\/code> wrapper. Specify your dependencies in a <code>Cask<\/code> file; can be used for project management or an Emacs configuration.<\/li>\n<li><a href=\"https:\/\/github.com\/dimitri\/el-get\">el-get<\/a>: ridiculously OP in terms of how many different sources you can pull packages from (<code>package.el<\/code>, every known VCS, distro package managers, <code>go get<\/code>(!!)).<\/li>\n<li><a href=\"https:\/\/github.com\/emacscollective\/borg\">Borg<\/a>: assimilates packages as Git submodules into <code>.emacs.d<\/code>, relying on <a href=\"https:\/\/emacsmirror.net\/\" rel=\"nofollow\">Emacsmirror<\/a>.<\/li>\n<li>\"Screw package managers! I'll just handle it all myself!\"<\/li>\n<\/ul><h3><a id=\"user-content-tldr-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#tldr-1\"><\/a>TL;DR<\/h3><p>Here is a summary of the main differences in philosophy between the package managers:<\/p><ul><li>Use <code>package.el<\/code> if you want package management to be as easy as possible, and do not much care for installing packages from specific sources, keeping track of their versions, or doing local development on them.<\/li>\n<li>Use Quelpa if you like <code>package.el<\/code> but really wish you could specify the sources of your packages.<\/li>\n<li>Use Cask if you like <code>package.el<\/code> but wish it came with some project management tools, as well.<\/li>\n<li>Use el-get if you want to easily install packages from as many different sources as possible.<\/li>\n<li>Use Borg if you like a more lightweight approach to package management that leverages existing solutions, if contributing changes to packages upstream is important to you, and if using Git submodules isn't a deal-breaker.<\/li>\n<li>Use the manual approach if you need to contribute changes to a package that is versioned in something other than Git.<\/li>\n<li>Use <code>straight.el<\/code> if you like reproducibility in your Emacs configuration, you regularly contribute changes to packages upstream, you think deferred installation is a really great idea, or you are writing an Emacs configuration to be used by others.<\/li>\n<\/ul><p>And here is a brief list of the main reasons you might not want to use <code>straight.el<\/code>:<\/p><ul><li><code>straight.el<\/code> is largely unusable if you do not have Git installed, although it is still possible to use the package-building features if you manage your repositories manually (you also cannot use the magic bootstrap snippet, in that case). If you don't want to install Git, you'll have to use <code>package.el<\/code> or take the manual approach.<\/li>\n<li><code>straight.el<\/code> is not built in to Emacs. If you want something that will work right out of the box, you're stuck with <code>package.el<\/code> or the manual approach.<\/li>\n<li><code>straight.el<\/code> takes a minute or two to update all your packages, since it does not rely on a centralized server. If you want quick update checking, you'll have to use <code>package.el<\/code>.<\/li>\n<li><code>straight.el<\/code> does not provide any user interface for package management. For that, you'll have to use <code>package.el<\/code>, el-get, Cask, or Borg (which expects you to use <a href=\"https:\/\/github.com\/emacscollective\/epkg\"><code>epkg<\/code><\/a> for browsing packages).<\/li>\n<li><code>straight.el<\/code> does not currently support using only stable versions of packages (i.e. tagged revisions), although this is a <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/31\">planned feature<\/a>. If this is important to you, you probably want to go with <code>package.el<\/code> (with GNU ELPA and MELPA Stable), Cask, or Quelpa.<\/li>\n<li><code>straight.el<\/code> does not currently support arbitrary build commands like <code>make<\/code>, although this is a <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/72\">planned feature<\/a>. This feature is supported by el-get and Borg.<\/li>\n<li>If you don't like having to modify your init-file to do package management, then <code>straight.el<\/code> is absolutely not for you. You want <code>package.el<\/code>, Quelpa, el-get, or Borg.<\/li>\n<li>If you really want to contribute changes to packages that are not versioned in Git, then <code>straight.el<\/code> will not help you. You'll have to manage the package's repository manually. Unfortunately, there is no existing package manager that supports both non-Git version-control systems and contributing changes upstream. You'll have to go with the manual approach.<\/li>\n<li><code>straight.el<\/code> does not provide project management tools. It is a package manager. If you want project management tools, check out Cask.<\/li>\n<li><code>straight.el<\/code> is quite new and moving fast. Things might break. The other package managers can generally be ranked as follows, from most active to least active: el-get, Quelpa, Borg, Cask, <code>package.el<\/code> (glacial).<\/li>\n<\/ul><h3><a id=\"user-content-comparison-to-packageel\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-packageel\"><\/a>Comparison to <code>package.el<\/code><\/h3><ul><li><code>package.el<\/code> downloads pre-built packages from central servers using a special (undocumented?) HTTP protocol, while <code>straight.el<\/code> clones Git (or other) repositories and builds packages locally.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-straightel\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-straightel\"><\/a>Advantages of <code>straight.el<\/code><\/h4><ul><li><code>straight.el<\/code> allows you to install a package from any branch of any Git repository. <code>package.el<\/code> only allows you to install a package from a <code>package.el<\/code>-compliant central server.<\/li>\n<li><code>straight.el<\/code> allows you to check out any Git revision of any package. <code>package.el<\/code> only allows you to install the latest version, and there is no way to downgrade.<\/li>\n<li><code>straight.el<\/code> supports Emacsmirror, while <code>package.el<\/code> does not.<\/li>\n<li><code>straight.el<\/code> uses your init-file as the sole source of truth for package operations. <code>package.el<\/code> loads every package you ever installed at startup, even if some of those packages are no longer referenced by your init-file.<\/li>\n<li><code>straight.el<\/code> supports 100% reproducibility for your Emacs packages with version lockfiles. <code>package.el<\/code> cannot provide reproducibility for the set of packages installed, the central servers they were installed from, or the versions in use.<\/li>\n<li><code>straight.el<\/code> allows you to make arbitrary changes to your packages locally. While it is possible to make local changes to <code>package.el<\/code> packages, these changes cannot be version-controlled and they will be silently overwritten whenever <code>package.el<\/code> performs an update.<\/li>\n<li><code>straight.el<\/code> allows you to perform arbitrary version-control operations on your package's Git repositories, including contributing changes upstream. <code>straight.el<\/code> has explicit support for specifying both an upstream and a fork for a package. Contributing changes upstream with <code>package.el<\/code> is impossible.<\/li>\n<li><code>straight.el<\/code> is designed with <code>emacs -Q<\/code> bug reports in mind. <code>package.el<\/code> is unsuitable for minimal bug reproductions, since it automatically loads all of your packages on any package operation, even in <code>emacs -Q<\/code>.<\/li>\n<li><code>straight.el<\/code> operates quietly when all is going well. <code>package.el<\/code> displays all messages, errors, and warnings that come from byte-compilation and autoload generation.<\/li>\n<li><code>straight.el<\/code> considers modifying the user's init-file extremely uncouth. <code>package.el<\/code> aggressively inserts a call to <code>package-initialize<\/code> into the init-file if it is not already present, whenever any package management operation is performed.<\/li>\n<li><code>straight.el<\/code> has a profile system that allows users of someone else's Emacs configuration to manage an additional subset of packages, or to override upstream package configuration, without forking the upstream. <code>package.el<\/code> has no such concept.<\/li>\n<li><code>straight.el<\/code> is developed openly on GitHub, using a modern <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\">issue tracker<\/a> and continuous integration from <a href=\"https:\/\/travis-ci.org\/raxod502\/straight.el\" rel=\"nofollow\">Travis CI<\/a>. It welcomes contributions of any type. <code>straight.el<\/code> is licensed under the permissive MIT license and does not require a copyright assignment. <code>straight.el<\/code> is developed actively and has explicit support for installing development versions of itself, as well as for contributing upstream changes. <code>package.el<\/code> is maintained as a part of Emacs core, meaning that the contribution process is poorly documented and discouraging. Releases of <code>package.el<\/code> coincide with releases of Emacs, which are infrequent and inflexible. There is no issue tracker specifically for <code>package.el<\/code>, only the Emacs bug tracker and the emacs-devel mailing list. Contributing to <code>package.el<\/code> requires a poorly-documented, cumbersome copyright assignment process and is done by submitting patches to an antiquated mailing list, unsupported by modern code review tooling or continuous integration.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-packageel\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-packageel\"><\/a>Advantages of <code>package.el<\/code><\/h4><ul><li><code>package.el<\/code> does not require that you have Git installed, since the central server deals with where the packages originally came from. <code>straight.el<\/code> cannot be used at all without Git.<\/li>\n<li><code>package.el<\/code> is built in to Emacs and does not require additional configuration to get started with. <code>straight.el<\/code> requires the use of a 10-line bootstrap snippet in your init-file.<\/li>\n<li><code>package.el<\/code> can perform bulk package updates more quickly since it relies on central servers.<\/li>\n<li><code>package.el<\/code> has a user interface for package management that also displays package metadata. <code>straight.el<\/code> has no user interface for package management; any UI is provided by the user's <code>completing-read<\/code> framework.<\/li>\n<li><code>package.el<\/code> does not require you to touch your init-file to install packages, while <code>straight.el<\/code> absolutely refuses to permanently install a package without an explicit reference to it in your init-file (although this may be considered an advantage, depending on your perspective).<\/li>\n<li>Using MELPA Stable, <code>package.el<\/code> can install only stable versions of packages. By default, <code>package.el<\/code> also installs only stable versions of packages from GNU ELPA. These modes of operation are unsupported by <code>straight.el<\/code> at this time, although this is a <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/31\">planned feature<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-additional-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#additional-notes\"><\/a>Additional notes<\/h4><ul><li><code>package.el<\/code> and <code>straight.el<\/code> usually take approximately the same time to install packages, despite the fact that <code>straight.el<\/code> is cloning entire Git repositories. This is because network latency and byte-compilation are the dominant factors.<\/li>\n<li>Some <code>package.el<\/code> servers build packages from non-Git upstreams. <code>package.el<\/code> can install these packages, while <code>straight.el<\/code> cannot. However, since <code>package.el<\/code> has no version-control support, this is more or less equivalent to installing those packages from the <a href=\"https:\/\/emacsmirror.net\/\" rel=\"nofollow\">Emacsmirror<\/a>, which <code>straight.el<\/code> can do by default.<\/li>\n<\/ul><h3><a id=\"user-content-comparison-to-quelpa\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-quelpa\"><\/a>Comparison to Quelpa<\/h3><ul><li>Quelpa allows for fetching packages from arbitrary sources and building them into a format that can be installed by <code>package.el<\/code>. <code>straight.el<\/code> has a philosophy which is fundamentally incompatible with <code>package.el<\/code>, and non-compatibility with <code>package.el<\/code> is one of its design goals.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-straightel-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-straightel-1\"><\/a>Advantages of <code>straight.el<\/code><\/h4><ul><li><code>straight.el<\/code> has out-of-the-box compatibility with MELPA, GNU ELPA, and Emacsmirror, while Quelpa only has support for MELPA. To use GNU ELPA, you must drop down to <code>package.el<\/code>. <a href=\"https:\/\/emacsmirror.net\/\" rel=\"nofollow\">Emacsmirror<\/a> is not supported by default, although it is easy to specify an Emacsmirror repository in a recipe. While Quelpa allows you to specify custom recipe folders, it does not have support for cloning these folders automatically from version control, nor for generating the recipes in any way other than copying them literally from files. <code>straight.el<\/code> allows you full flexibility in this regard.<\/li>\n<li><code>straight.el<\/code> has integrated support for selecting particular Git revisions of packages. This process is more manual in Quelpa, as it requires placing the commit hash into the recipe, which disables updates.<\/li>\n<li><code>straight.el<\/code> uses your init-file as the sole source of truth for package operations. Since Quelpa is based on <code>package.el<\/code>, it also loads every package you ever installed at startup, even if those packages are no longer referenced by your init-file. Furthermore, there is an additional caching layer, so that deleting a package from the <code>package.el<\/code> interface and removing it from your init-file still does not actually delete it.<\/li>\n<li><code>straight.el<\/code> supports 100% reproducibility for your Emacs packages with version lockfiles. Quelpa can theoretically provide some measure of reproducibility, but this requires significant manual effort since all packages are not associated with specific revisions by default, nor is the revision of MELPA saved anywhere.<\/li>\n<li><code>straight.el<\/code> allows you to make arbitrary changes to your packages locally. While it is possible to make local changes to Quelpa packages, there are two places to do so: the built package, which is the default destination of <code>find-function<\/code>, and the original repository. Changes to the former are not version-controlled and will be silently overwritten by <code>package.el<\/code> operations, while changes to the latter will be silently overwritten by Quelpa operations.<\/li>\n<li><code>straight.el<\/code> has explicit support for configuring both an upstream repository and a fork for the same package. Quelpa does not have such a concept.<\/li>\n<li><code>straight.el<\/code> allows you complete control over how your repositories are managed, and the default behavior is to draw all packages versioned in a single repository from a single copy of that repository. Quelpa is hardcoded to require a separate repository for each package, so that installing Magit requires three copies of the Magit repository.<\/li>\n<li><code>straight.el<\/code> builds packages using symlinks, meaning that <code>find-function<\/code> works as expected. Quelpa builds packages by copying, a feature inherited from MELPA. This means that <code>find-function<\/code> brings you to the built package, instead of the actual repository, which is not version-controlled and will be overwritten whenever <code>package.el<\/code> performs an update.<\/li>\n<li><code>straight.el<\/code> allows you to perform arbitrary version-control operations on your package's Git repositories. Quelpa allows this, but all local changes will be silently overridden whenever Quelpa performs an update.<\/li>\n<li><code>straight.el<\/code> is designed with <code>emacs -Q<\/code> bug reports in mind. Since Quelpa is based on <code>package.el<\/code>, it is also unsuitable for minimal bug reproductions, since it automatically loads all of your packages on any package operation, even in <code>emacs -Q<\/code>.<\/li>\n<li><code>straight.el<\/code> operates quietly when all is going well. Since Quelpa is based on <code>package.el<\/code>, it displays all messages, errors, and warnings that come from byte-compilation and autoload generation. It also displays additional messages while cloning Git repositories, downloading files, and building packages from their repositories into <code>package.el<\/code> format.<\/li>\n<li><code>straight.el<\/code> does not modify your init-file. Since Quelpa is based on <code>package.el<\/code>, it inherits the behavior of aggressively inserting a call to <code>package-initialize<\/code> into your init-file on any package management operation.<\/li>\n<li><code>straight.el<\/code> has a profile system that allows users of someone else's Emacs configuration to manage an additional subset of packages, or to override upstream package configuration, without forking the upstream. Quelpa has no such concept.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-quelpa\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-quelpa\"><\/a>Advantages of Quelpa<\/h4><ul><li>Quelpa supports all the version-control systems supported by MELPA, which is to say almost every commonly and uncommonly used VCS. <code>straight.el<\/code> only supports Git, although it is designed to support other version-control backends.<\/li>\n<li>Quelpa allows for installing only stable versions of packages, from any source. This mode of operation is unsupported by <code>straight.el<\/code>, although it is a <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/31\">planned feature<\/a>.<\/li>\n<li>Since Quelpa is based on <code>package.el<\/code>, it inherits a user interface for package management that also displays package metadata. <code>straight.el<\/code> has no such interface.<\/li>\n<\/ul><h4><a id=\"user-content-additional-notes-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#additional-notes-1\"><\/a>Additional notes<\/h4><ul><li><code>straight.el<\/code> and Quelpa both allow you to manage your package's local repositories manually, if you wish.<\/li>\n<li>In principle, <code>straight.el<\/code> and Quelpa have identical package installation times, since they are performing the same operations. In practice, Quelpa is slightly slower since it builds packages by copying rather than symlinking, and it clones multiple copies of the same Git repository when multiple packages are built from it.<\/li>\n<li><code>straight.el<\/code> encourages you to keep a tight handle on your package versions by default. Quelpa encourages you to stick to the latest versions of your packages, and to upgrade them automatically.<\/li>\n<\/ul><h3><a id=\"user-content-comparison-to-cask\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-cask\"><\/a>Comparison to Cask<\/h3><p>I have not used Cask extensively, so please feel especially free to offer corrections for this section.<\/p><ul><li>Cask installs packages using the <code>package.el<\/code> protocol, based on a <code>Cask<\/code> file written in the Cask DSL. <code>straight.el<\/code> eschews <code>package.el<\/code> entirely, and clones packages from source based on how you invoke <code>straight-use-package<\/code> in your init-file.<\/li>\n<li>Cask focuses more on being a build manager, like Maven or Leiningen, while <code>straight.el<\/code> focuses <em>exclusively<\/em> on being a package manager.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-straightel-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-straightel-2\"><\/a>Advantages of <code>straight.el<\/code><\/h4><ul><li><code>straight.el<\/code> has out-of-the-box compatibility with Emacsmirror, while Cask only supports <code>package.el<\/code>-compliant repositories. However, it is easy to specify an Emacsmirror repository in a recipe. Cask does not support custom package sources. <code>straight.el<\/code> supports MELPA, GNU ELPA, and Emacsmirror, and allows you to add any other sources you would like.<\/li>\n<li><code>straight.el<\/code> has integrated support for selecting particular Git revisions of packages. This process is more manual in Cask, as it requires placing the commit hash into the recipe, which disables updates.<\/li>\n<li><code>straight.el<\/code> uses your init-file as the sole source of truth for package operations. Since Cask is based on <code>package.el<\/code>, it loads every package you ever installed at startup, even if some of those packages are no longer referenced by your <code>Cask<\/code> file.<\/li>\n<li><code>straight.el<\/code> determines your package management configuration implicitly by detecting how you call <code>straight-use-package<\/code> in your init-file and making the appropriate changes immediately. Cask requires manual intervention (for example, issuing a <code>cask install<\/code> command when you have updated your <code>Cask<\/code> file). However, both <code>straight.el<\/code> and Cask can be considered declarative package managers.<\/li>\n<li><code>straight.el<\/code> supports 100% reproducibility for your Emacs packages with version lockfiles. Cask can theoretically provide some measure of reproducibility, but this requires significant manual effort since all packages are not associated with specific revisions by default, nor is the revision of Cask saved anywhere.<\/li>\n<li><code>straight.el<\/code> allows you to make arbitrary changes to your packages locally. While it is possible to make local changes to Cask packages, these will not be version-controlled and they will be silently overwritten or shadowed when Cask performs an update.<\/li>\n<li><code>straight.el<\/code> allows you to perform arbitrary version-control operations on your package's Git repositories, including contributing changes upstream. <code>straight.el<\/code> has explicit support for specifying both an upstream and a fork for a package. Contributing changes upstream with Cask is impossible.<\/li>\n<li><code>straight.el<\/code> is designed with <code>emacs -Q<\/code> bug reports in mind. Cask appears to be unsuitable for minimal bug reproductions, since there does not appear to be a straightforward way to load a single package, without loading all other packages configured in your <code>Cask<\/code> file.<\/li>\n<li><code>straight.el<\/code> operates quietly when all is going well. Since Cask is based on <code>package.el<\/code>, it displays all messages, errors, and warnings that come from byte-compilation and autoload generation.<\/li>\n<li><code>straight.el<\/code> has a profile system that allows users of someone else's Emacs configuration to manage an additional subset of packages, or to override upstream package configuration, without forking the upstream. Cask has no such concept.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-cask\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-cask\"><\/a>Advantages of Cask<\/h4><ul><li>Cask provides a useful toolbox of operations for project management, which are completely absent from <code>straight.el<\/code>.<\/li>\n<li>Since Cask is based on <code>package.el<\/code>, it does not require that you have Git installed. (It does require Python, however.) <code>straight.el<\/code> is mostly unusable without Git.<\/li>\n<li>Since Cask is based on <code>package.el<\/code>, it can perform bulk package updates more quickly than <code>straight.el<\/code>.<\/li>\n<li>Since Cask is based on <code>package.el<\/code>, it inherits a user interface for package management that also displays package metadata.<\/li>\n<li>Since Cask is based on <code>package.el<\/code>, you can install packages without editing a file manually, although this rather defeats the entire purpose of using Cask instead of <code>package.el<\/code>. <code>straight.el<\/code> absolutely refuses to permanently install a package without an explicit reference to it in your init-file (although this may be considered an advantage, depending on your perspective).<\/li>\n<li>Using MELPA Stable, Cask can install only stable versions of packages. By default, Cask also installs only stable versions of packages from GNU ELPA. These modes of operation are unsupported by <code>straight.el<\/code> at this time, although this is a <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/31\">planned feature<\/a>.<\/li>\n<li>Cask supports more version-control systems than <code>straight.el<\/code> (which only supports Git).<\/li>\n<\/ul><h3><a id=\"user-content-comparison-to-el-get\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-el-get\"><\/a>Comparison to el-get<\/h3><p>I have not used el-get extensively, so please feel especially free to offer corrections for this section.<\/p><ul><li>Both el-get and <code>straight.el<\/code> implement their own package management abstractions instead of delegating to <code>package.el<\/code>. However:\n<ul><li>el-get goes the route of adding as many package sources as possible (e.g. <code>package.el<\/code>, many different version-control systems, various specific websites, and even system package managers) so that packages can be used very easily.<\/li>\n<li><code>straight.el<\/code> only supports Git and in doing so is able to provide more advanced package management features.<\/li>\n<\/ul><\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-straightel-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-straightel-3\"><\/a>Advantages of <code>straight.el<\/code><\/h4><ul><li><code>straight.el<\/code> uses your init-file as the sole source of truth for package operations. el-get has additional metadata stored outside the init-file, although specifying all packages in your init-file is a supported mode of operation.<\/li>\n<li><code>straight.el<\/code> supports 100% reproducibility for your Emacs packages with version lockfiles. el-get can theoretically provide some measure of reproducibility, but this requires significant manual effort since all packages are not associated with specific revisions by default, nor is the revision of el-get saved anywhere.<\/li>\n<li><code>straight.el<\/code> allows you to make arbitrary changes to your packages locally, and conflicts during updates are presented to the user and resolved interactively. While it is possible to make local changes to el-get packages, the el-get manual warns that such changes may break the update mechanism.<\/li>\n<li><code>straight.el<\/code> has explicit support for configuring both an upstream repository and a fork for the same package. el-get does not have such a concept.<\/li>\n<li><code>straight.el<\/code> allows you to perform arbitrary version-control operartions on your package's Git repositories. el-get allows this, but local changes will be overwritten when el-get performs an update.<\/li>\n<li><code>straight.el<\/code> provides a suite of powerful interactive workflows for performing bulk operations on your package's Git repositories. el-get only allows you to install, uninstall, and update packages.<\/li>\n<li><code>straight.el<\/code> operates quietly when all is going well. el-get reports its progress verbosely.<\/li>\n<li><code>straight.el<\/code> has a profile system that allows users of someone else's Emacs configuration to manage an additional subset of packages, or to override upstream package configuration, without forking the upstream. el-get has no such concept.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-el-get\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-el-get\"><\/a>Advantages of el-get<\/h4><ul><li>el-get supports virtually all known version-control systems, as well as system package managers, EmacsWiki, arbitrary HTTP, and even <code>go get<\/code>. <code>straight.el<\/code> supports only Git, although it does allow you to manage your local repositories manually if you would like.<\/li>\n<li>el-get has been around since 2010 and is on its fifth major version, whereas <code>straight.el<\/code> was created in January 2017 and is only now approaching a 1.0 release. Clearly, el-get is more stable, although despite its recency <code>straight.el<\/code> is already almost 50% of the size of el-get, by the line count. Both package managers are actively maintained.<\/li>\n<li>el-get has a recipe format which is several orders of magnitude more powerful than that of <code>straight.el<\/code>, since it supports many more package sources that can be configured and since it allows for a more complex build process.<\/li>\n<li>el-get provides a number of features for running per-package initialization and setup code, including pulling that code from arbitrary sources. <code>straight.el<\/code> does not support this and expects you to use a dedicated tool like <a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a> (with which integration is built in) for that purpose.<\/li>\n<li>el-get has a user interface for package management that also displays package metadata, similarly to <code>package.el<\/code>. <code>straight.el<\/code> has no user interface for package management; any UI is provided by the user's <code>completing-read<\/code> framework.<\/li>\n<\/ul><h3><a id=\"user-content-comparison-to-borg\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-borg\"><\/a>Comparison to Borg<\/h3><ul><li>Borg and <code>straight.el<\/code> are perhaps the two most similar package managers on this list. The difference is that Borg is very minimal and expects you to complement it with other tools such as <a href=\"https:\/\/magit.vc\/\" rel=\"nofollow\">Magit<\/a>, <a href=\"https:\/\/github.com\/emacscollective\/epkg\">epkg<\/a>, <a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a>, and <a href=\"https:\/\/github.com\/tarsius\/auto-compile\">auto-compile<\/a>. On the other hand, <code>straight.el<\/code> aspires to be a one-stop package management solution, although it does not try to replace dedicated version-control packages (Magit) or dedicated package <em>configuration<\/em> packages (<code>use-package<\/code>).<\/li>\n<li>Borg uses Git submodules, while <code>straight.el<\/code> uses independently managed Git repositories.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-straightel-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-straightel-4\"><\/a>Advantages of <code>straight.el<\/code><\/h4><ul><li><code>straight.el<\/code> supports MELPA, GNU ELPA, Emacsmirror, and custom recipe sources. Borg only supports Emacsmirror and custom recipe sources. However, as the Emacsmirror is a near-complete superset of both GNU ELPA and MELPA, this does not necessarily mean you have access to more packages: it just means you benefit from the recipe maintenance efforts of the MELPA team and the Emacsmirror team, rather than only the latter.<\/li>\n<li>Borg, even when combined with related tools, do not allow for the kind of massive interactive repository management provided by <code>straight.el<\/code>.<\/li>\n<li><code>straight.el<\/code> supports deferred and conditional installation. This is not supported by Borg, although it could in principle be implemented via lazy cloning of submodules.<\/li>\n<li><code>straight.el<\/code> provides an API designed for other version-control backends to be added in future. Borg is inextricably tied to Git.<\/li>\n<li>The interface for Git submodules has a number of sharp edges.<\/li>\n<li><code>straight.el<\/code> provides dependency management. This is a manual process in Borg.<\/li>\n<li><code>straight.el<\/code> provides mechanisms for updating your packages. This is a manual process in Borg.<\/li>\n<li><code>straight.el<\/code> is configured solely by how you use in your init-file. Configuring Borg requires customizing <code>~\/.emacs.d\/.gitmodules<\/code>, which means (for example) that you cannot generate recipes dynamically. (However, the handling of configuration is <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/95#issuecomment-316379495\">planned<\/a> to be improved in a future release.)<\/li>\n<li><code>straight.el<\/code> has a profile system that allows users of someone else's Emacs configuration to manage an additional subset of packages, or to override upstream package configuration, without forking the upstream. Borg has no such concept.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-borg\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-borg\"><\/a>Advantages of Borg<\/h4><ul><li>Borg does a heck of a lot less magic, so if you want a solution with simple implementation details, <code>straight.el<\/code> may not be for you. (But see the developer manual and docstrings, first.)<\/li>\n<li>Borg supports arbitrary build commands; <code>straight.el<\/code> does not (although this is a <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/72\">planned feature<\/a>).<\/li>\n<\/ul><h3><a id=\"user-content-comparison-to-the-manual-approach\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-the-manual-approach\"><\/a>Comparison to the manual approach<\/h3><ul><li>The manual approach is to download packages yourself and put them on your <code>load-path<\/code>. <code>straight.el<\/code> is more or less what you get when you take the manual approach, think very hard about the best way to do everything, and then automate all of it.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-straightel-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-straightel-5\"><\/a>Advantages of <code>straight.el<\/code><\/h4><ul><li><code>straight.el<\/code> figures out where to clone your packages from for you.<\/li>\n<li><code>straight.el<\/code> byte-compiles your packages for you and generates their autoloads automatically.<\/li>\n<li><code>straight.el<\/code> frees you from needing to manually recompile and regenerate autoloads.<\/li>\n<li><code>straight.el<\/code> keeps track of dependencies for you.<\/li>\n<li><code>straight.el<\/code> provides tools to manage all your packages in bulk, which would otherwise be a long, tedious process.<\/li>\n<li><code>straight.el<\/code> allows you to get reproducibility for your configuration without needing to keep all of your packages under version control.<\/li>\n<li><code>straight.el<\/code> (when used with <a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a>) automates the complex process of deferred installation.<\/li>\n<li><code>straight.el<\/code> links packages into a separate build directories. Running packages directly from their repositories has a number of problems, including:\n<ul><li>making it impossible to run only one package, if others are provided in the same repository.<\/li>\n<li>making your working directory dirty when the package author forgot to add their build artifacts like <code>*.elc<\/code> and autoload files to the <code>.gitignore<\/code>.<\/li>\n<li>instantly losing compatibility with MELPA recipes.<\/li>\n<\/ul><\/li>\n<li><code>straight.el<\/code> offers you a single entry point to install only a single package in isolation, for a minimal bug reproduction. With the manual approach this would be more complicated, especially if the package has dependencies.<\/li>\n<li><code>straight.el<\/code> frees you from needing to think about package management, since I already did all the thinking to figure how best to design everything.<\/li>\n<\/ul><h4><a id=\"user-content-advantages-of-the-manual-approach\" class=\"anchor\" aria-hidden=\"true\" href=\"#advantages-of-the-manual-approach\"><\/a>Advantages of the manual approach<\/h4><ul><li>No dependencies.<\/li>\n<li>You learn a lot, if you don't give up first.<\/li>\n<li>You might end up writing a package manager (case in point).<\/li>\n<li>This is the only way to deal with packages that have non-Git upstreams which you need to contribute changes to. (However, you can always use the manual approach for one package and <code>straight.el<\/code> for the rest. Or you can just eschew <code>straight.el<\/code>'s version-control support for that package, and use it only for building the package.)<\/li>\n<\/ul><h2><a id=\"user-content-user-manual\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-manual\"><\/a>User manual<\/h2><p>This section tells you everything you need to know about the user-facing features of <code>straight.el<\/code>. For implementation details, see the <a href=\"#developer-manual\">developer manual<\/a>. It may also be helpful to get some perspective on the overarching concepts of <code>straight.el<\/code> from the <a href=\"#conceptual-overview\">conceptual overview<\/a>.<\/p><h3><a id=\"user-content-bootstrapping-straightel\" class=\"anchor\" aria-hidden=\"true\" href=\"#bootstrapping-straightel\"><\/a>Bootstrapping <code>straight.el<\/code><\/h3><p>In order to use <code>straight.el<\/code>, you will need to somehow get it loaded into Emacs. (This is easy for <code>package.el<\/code>, since <code>package.el<\/code> is built in to Emacs. <code>straight.el<\/code> must work a little harder.)<\/p><p><code>straight.el<\/code> comes with a file to do just this, <code>bootstrap.el<\/code>. All you need to do is load that file. You can do this with <code>M-x load-file<\/code> or by a call to <code>load<\/code> in your init-file. However, there is an obvious shortcoming: <code>bootstrap.el<\/code> will only be available once <code>straight.el<\/code> is already installed.<\/p><p>You could just invoke <code>git clone<\/code> from your init-file, if <code>straight.el<\/code> is not installed, but then you would have to manually take care of selecting the correct branch, parsing your version lockfile to check out the right revision, and so on. Instead, you can just use this snippet, which uses a copious amount of magic to take care of all these details for you:<\/p><pre>(defvar bootstrap-version)\n(let ((bootstrap-file\n       (expand-file-name \"straight\/repos\/straight.el\/bootstrap.el\" user-emacs-directory))\n      (bootstrap-version 5))\n  (unless (file-exists-p bootstrap-file)\n    (with-current-buffer\n        (url-retrieve-synchronously\n         \"https:\/\/raw.githubusercontent.com\/raxod502\/straight.el\/develop\/install.el\"\n         'silent 'inhibit-cookies)\n      (goto-char (point-max))\n      (eval-print-last-sexp)))\n  (load bootstrap-file nil 'nomessage))\n<\/pre><p>Despite the reference to <code>develop<\/code>, this snippet actually installs from the <code>master<\/code> branch by default, just like every other package. Furthermore, the correct revision of <code>straight.el<\/code> is checked out, if you have one specified in your lockfile. Even better, you can <a href=\"#overriding-the-recipe-for-straightel\">override the recipe for <code>straight.el<\/code><\/a>, just like for any other package.<\/p><h3><a id=\"user-content-installing-packages-programmatically\" class=\"anchor\" aria-hidden=\"true\" href=\"#installing-packages-programmatically\"><\/a>Installing packages programmatically<\/h3><p>The primary entry point to <code>straight.el<\/code> is the <code>straight-use-package<\/code> function. It can be invoked interactively (for installing a package temporarily) or programmatically (for installing a package permanently). This section covers the programmatic usage; see <a href=\"#interactive-usage\">later<\/a> for interactive usage.<\/p><p>Here is the basic usage of <code>straight-use-package<\/code>:<\/p><pre>(straight-use-package 'el-patch)\n<\/pre><p>This will ensure that the package <code>el-patch<\/code> is installed and loaded. (Note that <code>straight-use-package<\/code> takes a symbol, not a string, for the name of the package.) Precisely, this is what happens:<\/p><ul><li>If the local Git repository for <code>el-patch<\/code> is not available, it is cloned, and the appropriate revision is checked out (if one is specified in your version lockfiles).<\/li>\n<li>If the local Git repository has been modified since the last time the package was built, it is rebuilt. This means:\n<ul><li>The <code>.el<\/code> files are symlinked into a separate directory to isolate them from other, irrelevant files.<\/li>\n<li>The main package file is checked for dependencies, which are installed recursively if necessary using <code>straight-use-package<\/code>.<\/li>\n<li>The <code>.el<\/code> files are byte-compiled.<\/li>\n<li>Autoloads are extracted from the <code>.el<\/code> files and saved into a separate file.<\/li>\n<\/ul><\/li>\n<li>The package's directory is added to Emacs' <code>load-path<\/code>.<\/li>\n<li>The package's autoloads are evaluated.<\/li>\n<\/ul><p>Package authors should note that <code>straight.el<\/code> checks for dependencies that are specified in the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Packaging-Basics.html\" rel=\"nofollow\"><code>package.el<\/code> format<\/a>. To spare you reading that documentation, this is either a <code>Package-Requires<\/code> header in <code>PACKAGENAME.el<\/code>, or an argument to a <code>define-package<\/code> invocation in <code>PACKAGENAME-pkg.el<\/code>. Despite the <a href=\"#advantages-of-straightel\">many shortcomings<\/a> of <code>package.el<\/code>, it has done a good job of creating a standardized format for dependency declarations.<\/p><p>There is one exception to the above statement: not all entries specified in the <code>Package-Requires<\/code> header necessarily correspond to packages. For example, specifying a minimum Emacs version for a package is done by depending on the <code>emacs<\/code> pseudo-package. Such packages are simply ignored by <code>straight.el<\/code>, using the variable <code>straight-built-in-pseudo-packages<\/code>.<\/p><p>Note that loading a package does not entail invoking <code>require<\/code> on any of its features. If you wish to actually load the files of the package, you need to do this separately. This is because most packages do not need to be loaded immediately, and are better served by the autoload system.<\/p><h4><a id=\"user-content-installing-with-a-custom-recipe\" class=\"anchor\" aria-hidden=\"true\" href=\"#installing-with-a-custom-recipe\"><\/a>Installing with a custom recipe<\/h4><p><code>straight-use-package<\/code> can also take a list instead of a symbol. In that case, the first member of the list is a symbol giving the package name, and the remainder of the list is a <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Property-Lists.html\" rel=\"nofollow\">property list<\/a> providing information about how to install and build the package. Here is an example:<\/p><pre>(straight-use-package\n '(el-patch :type git :host github :repo \"raxod502\/el-patch\"\n            :fork (:host github\n                   :repo \"your-name\/el-patch\")))\n<\/pre><p>If you give <code>straight-use-package<\/code> just a package name, then a recipe will be looked up by default (see the section on <a href=\"#recipe-lookup\">recipe lookup<\/a>). You can see the default recipe for a package by invoking <a href=\"#interactive-usage\"><code>M-x straight-get-recipe<\/code><\/a>.<\/p><p>If <code>straight-allow-recipe-inheritance<\/code> is non-nil, then you only need to specify the components of the recpie that you want to override. All other components will still be looked up in the default recipe. In the example above, we are only interested in changing the <code>:fork<\/code> component. Therefore if <code>straight-allow-recipe-inheritance<\/code> is set, the recipe could be simplifed as follows:<\/p><pre>(straight-use-package\n '(el-patch :fork (:repo \"your-name\/el-patch\")))\n<\/pre><p>or even simpler:<\/p><pre>(straight-use-package\n '(el-patch :fork \"your-name\/el-patch\"))\n<\/pre><p>The <code>:files<\/code> keyword and all version control keywords support inheritance.<\/p><p>To learn more, see the section on <a href=\"#the-recipe-format\">the recipe format<\/a>.<\/p><h4><a id=\"user-content-additional-arguments-to-straight-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#additional-arguments-to-straight-use-package\"><\/a>Additional arguments to <code>straight-use-package<\/code><\/h4><p>The full user-facing signature of <code>straight-use-package<\/code> is:<\/p><pre>(straight-use-package PACKAGE-OR-RECIPE &amp;optional NO-CLONE NO-BUILD)\n<\/pre><p>As discussed <a href=\"#installing-packages-programmatically\">previously<\/a>, by default <code>straight-use-package<\/code> will do three things:<\/p><ul><li>Register the recipe provided with <code>straight.el<\/code>.<\/li>\n<li>Clone the package's local repository, if it is absent.<\/li>\n<li>Rebuild the package if necessary, and load it.<\/li>\n<\/ul><p>By providing the optional arguments, you may cause processing to halt before all three of these tasks are completed. Specifically, providing <code>NO-CLONE<\/code> causes processing to halt after registration but before cloning, and providing <code>NO-BUILD<\/code> causes processing to halt after cloning (if necessary) but before building and loading.<\/p><p><code>straight.el<\/code> supports lazy-loading by means of a special value for <code>NO-CLONE<\/code>, the symbol <code>lazy<\/code>. If this symbol is passed, then processing will halt at the clone step, unless the package is already cloned. This means that the package is built and loaded if it is already installed, but otherwise installation is deferred until later. When you want to trigger the lazy installation, simply call <code>straight-use-package<\/code> again, but without <code>NO-CLONE<\/code>. (There is no need to pass the recipe again; see <a href=\"#recipe-lookup\">recipe lookup<\/a>.)<\/p><p>You can also pass functions for <code>NO-CLONE<\/code> or <code>NO-BUILD<\/code>, which will be called with the package name as a string; their return values will then be used instead.<\/p><p>Note that if it makes no sense to build a package, then you should put <code>:no-build t<\/code> in its <a href=\"#the-recipe-format\">recipe<\/a>, rather than specifying <code>NO-BUILD<\/code> every time you register it with <code>straight.el<\/code>. (This is especially relevant when writing recipes for <a href=\"#customizing-recipe-repositories\">recipe repositories<\/a>.)<\/p><h4><a id=\"user-content-variants-of-straight-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#variants-of-straight-use-package\"><\/a>Variants of <code>straight-use-package<\/code><\/h4><p>For convenience, <code>straight.el<\/code> provides some functions that wrap <code>straight-use-package<\/code> with particular arguments, to cover all of the common cases. Each of these functions takes only a package name or recipe, and no additional arguments.<\/p><ul><li><code>straight-register-package<\/code>: always stop after the registration step. This may be useful for specifying the recipe for an optional dependency (see <a href=\"#recipe-lookup\">recipe lookup<\/a>, but see also <a href=\"#overriding-recipes\">recipe overrides<\/a>).<\/li>\n<li><code>straight-use-package-no-build<\/code>: always stop before the build step. This is used by <a href=\"#lockfile-management\"><code>straight-freeze-versions<\/code><\/a> to make sure packages are cloned, since building them is unnecessary for writing the lockfiles.<\/li>\n<li><code>straight-use-package-lazy<\/code>: stop at the clone step if the package's local repository is not already cloned. This is used for lazy-loading.<\/li>\n<\/ul><h4><a id=\"user-content-customizing-when-packages-are-built\" class=\"anchor\" aria-hidden=\"true\" href=\"#customizing-when-packages-are-built\"><\/a>Customizing when packages are built<\/h4><p>By default, when <code>straight.el<\/code> is bootstrapped during Emacs init, it uses a bulk <code>find(1)<\/code> command to identify files that were changed since the last time a package depending on them was built. These packages are then rebuilt when they are requested via <code>straight-use-package<\/code>. Normally, <code>straight.el<\/code> will try to detect what sort of <code>find(1)<\/code> program is installed, and issue the appropriate command. If it makes a mistake, then you can manually customize <code>straight-find-flavor<\/code>. Alternately, you can install GNU find and customize the variable <code>straight-find-executable<\/code> to point to it.<\/p><p>For about 100 packages on an SSD, calling <code>find(1)<\/code> to detect modifications takes about 500ms. You can save this time by customizing <code>straight-check-for-modifications<\/code>. This is a list of symbols which determines how <code>straight.el<\/code> detects package modifications. The default value is <code>(find-at-startup find-when-checking)<\/code>, which means that <code>find(1)<\/code> is used to detect modifications at startup, and also when you invoke <code>M-x straight-check-package<\/code> or <code>M-x straight-check-all<\/code>. If you prefer to avoid this performance hit, or do not have <code>find(1)<\/code> installed, then you can remove these symbols from the list. In that case, you will probably want to add either <code>check-on-save<\/code> or <code>watch-files<\/code> to the list.<\/p><p><code>check-on-save<\/code> causes <code>straight.el<\/code> to use <code>before-save-hook<\/code> to detect package modifications as you make them (modifications made by the <code>straight.el<\/code> repository management commands are also detected). This reduces init time, but modifications made outside of Emacs (or modifications that bypass <code>before-save-hook<\/code>) are not detected. Pull requests extending the number of cases in which <code>straight.el<\/code> is able to detect live modifications are welcome. Also, for the sake of efficiency, this form of modification checking is restricted to subdirectories of <code>~\/.emacs.d\/straight\/repos<\/code>, so you must put your local repositories into that directory for it to work. (Pull requests to change this would be welcome.)<\/p><p><code>watch-files<\/code> causes <code>straight.el<\/code> to automatically invoke a filesystem watcher to detect modifications as they are made, inside or outside of Emacs. For this setting to work, you must have <a href=\"https:\/\/www.python.org\/\" rel=\"nofollow\"><code>python3<\/code><\/a> and <a href=\"https:\/\/github.com\/mattgreen\/watchexec\"><code>watchexec<\/code><\/a> installed on your <code>PATH<\/code>. By default, the watcher persists after Emacs is closed. You can stop it manually by running <code>M-x straight-watcher-stop<\/code>, and start it again by running <code>M-x straight-watcher-start<\/code>. The watcher script is designed so that when one instance is started, all the others gracefully shut down, so you don't have to worry about accidentally ending up with more than one. There is nothing exciting in the process buffer for the watcher, but if you are interested in it then its name is given by <code>straight-watcher-process-buffer<\/code>. (By default, the name has a leading space so that the buffer does not appear in the buffer list.)<\/p><p>There is probably no good reason to use both <code>check-on-save<\/code> and <code>watch-files<\/code> at the same time. Your configuration can dynamically switch between which one is used depending on <code>(executable-find \"watchexec\")<\/code> or similar.<\/p><p>If you prefer to eschew automatic package rebuilding entirely, you can just set <code>straight-check-for-modifications<\/code> to <code>nil<\/code>. In that case, packages will only be rebuilt when metadata (e.g. the recipe or the Emacs version) changes, or when you manually invoke <code>M-x straight-rebuild-package<\/code> or <code>M-x straight-rebuild-all<\/code>.<\/p><p>Regardless of your preferred setting for <code>straight-check-for-modifications<\/code>, you should set it before the <code>straight.el<\/code> bootstrap snippet is run, since hooks relating to this variable are set during bootstrap.<\/p><p>On Microsoft Windows, <code>find(1)<\/code> is generally not available, so the default value of <code>straight-check-for-modifications<\/code> is instead <code>(check-on-save)<\/code>.<\/p><h5><a id=\"user-content-custom-or-manual-modification-detection\" class=\"anchor\" aria-hidden=\"true\" href=\"#custom-or-manual-modification-detection\"><\/a>Custom or manual modification detection<\/h5><p>You can also use the low-level functions for modification detection directly.<\/p><p>The function <code>straight-register-repo-modification<\/code> takes a string (e.g. <code>\"straight.el\"<\/code>) corresponding to the name of a local repository, and marks all packages from that local repository to be rebuilt at next Emacs startup. This function silently ignores local repositories which contain slashes, a limitation which might be removed in future.<\/p><p>The function <code>straight-register-file-modification<\/code> takes no arguments and checks if the file visited by the current buffer (if any) is contained by any local repository. If so, it delegates to <code>straight-register-repo-modification<\/code>. The <code>check-on-save<\/code> value for <code>straight-check-for-modifications<\/code> just adds <code>straight-register-file-modification<\/code> to <code>before-save-hook<\/code>.<\/p><h5><a id=\"user-content-summary-of-options-for-package-modification-detection\" class=\"anchor\" aria-hidden=\"true\" href=\"#summary-of-options-for-package-modification-detection\"><\/a>Summary of options for package modification detection<\/h5><h6><a id=\"user-content-find-at-startup\" class=\"anchor\" aria-hidden=\"true\" href=\"#find-at-startup\"><\/a><code>find-at-startup<\/code><\/h6><p>Save build timestamps and run <code>find(1)<\/code> at startup to detect changes<\/p><ul><li>Most reliable, never misses changes<\/li>\n<li>Requires <code>find(1)<\/code><\/li>\n<li>Slows down startup<\/li>\n<\/ul><h6><a id=\"user-content-check-on-save\" class=\"anchor\" aria-hidden=\"true\" href=\"#check-on-save\"><\/a><code>check-on-save<\/code><\/h6><p>Use <code>before-save-hook<\/code> to detect changes<\/p><ul><li>No external dependencies<\/li>\n<li>No startup delay<\/li>\n<li>No additional CPU or memory impact<\/li>\n<li>Doesn't catch changes made except via <code>save-file<\/code> inside Emacs<\/li>\n<\/ul><h6><a id=\"user-content-watch-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#watch-files\"><\/a><code>watch-files<\/code><\/h6><p>Run filesystem watcher to detect changes<\/p><ul><li>Requires Python 3 and Watchexec<\/li>\n<li>No startup delay<\/li>\n<li>Takes a few seconds to build virtualenv the first time<\/li>\n<li>Memory and CPU impact of running filesystem watcher<\/li>\n<li>Only misses changes if you make them after booting the system but before starting Emacs<\/li>\n<\/ul><hr \/><h4><a id=\"user-content-customizing-how-packages-are-built\" class=\"anchor\" aria-hidden=\"true\" href=\"#customizing-how-packages-are-built\"><\/a>Customizing how packages are built<\/h4><p>By specifying a non-nil value for the <code>:no-build<\/code> attribute in a package's <a href=\"#the-recipe-format\">recipe<\/a>, you may prevent the package from being built at all. This is usually useful for recipe repositories which do not bundle executable Lisp code. (Make sure to use <a href=\"#customizing-recipe-repositories\"><code>straight-use-recipes<\/code><\/a> for registering recipe repositories.)<\/p><p>By specifying a non-nil value for the <code>:no-autoloads<\/code> attribute in a package's recipe, you may prevent any autoloads provided by the package from being generated and loaded into Emacs. This is mostly useful if the package provides a large number of autoloads, you know you need only a few of them, and you wish to optimize your startup time (although this is almost certainly premature optimization unless you <em>really<\/em> know what you're doing). You can also customize the variable <code>straight-disable-autoloads<\/code> to effect this change on all recipes which do not explicitly specify a <code>:no-autoloads<\/code> attribute.<\/p><p>By specifying a non-nil value for the <code>:no-byte-compile<\/code> attribute in a package's recipe, you may inhibit byte-compilation. See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/357\">this issue<\/a> for discussion of why this might be useful. You can also customize the variable <code>straight-disable-byte-compilation<\/code> to effect this change on all recipes which do not explicitly specify a <code>:no-byte-compile<\/code> attribute.<\/p><p>Usually, <code>straight.el<\/code> uses symbolic links (\"symlinks\") to make package files available from the build directory. This happens when <code>straight-use-symlinks<\/code> is non-nil, the default. On Microsoft Windows, however, support for symlinks is not always available, so the default value of <code>straight-use-symlinks<\/code> is nil on that platform. That causes copying to be used instead, and an advice is placed on <code>find-file<\/code> to cause the copied files to act as symlinks if you try to edit them.<\/p><p>If you want to activate symlink-support on MS Windows 7, 8, or 10, you should ensure the following requirements:<\/p><ul><li>\n<p><code>straight-use-symlinks<\/code> has to be set to non-nil manually.<\/p>\n<\/li>\n<li>\n<p>Your user-account needs to be assigned the right to create symbolic links. To do so, run \"secpol.msc\" and in \"Local Policies \u2192 User Rights Assignment\" assign the right to \"Create symbolic links\" to your user-account.<\/p>\n<\/li>\n<li>\n<p>If you have User Account Control (UAC) enabled and your user-account belongs to the the <em>Administrators<\/em> group you'll need to run Emacs in elevated mode to be able to create symlinks (see <a href=\"https:\/\/community.perforce.com\/s\/article\/3472\" rel=\"nofollow\">here<\/a> and <a href=\"https:\/\/stackoverflow.com\/a\/29065060\/3538165\" rel=\"nofollow\">here<\/a> and, for an official reference, section Access Token Changes <a href=\"https:\/\/msdn.microsoft.com\/en-us\/library\/bb530410.aspx#vistauac_topic8\" rel=\"nofollow\">in this document<\/a>.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/blogs.windows.com\/buildingapps\/2016\/12\/02\/symlinks-windows-10\/\" rel=\"nofollow\">Windows Creators Update<\/a> supports symlink-creation without any special permission setup.<\/p>\n<\/li>\n<\/ul><h4><a id=\"user-content-customizing-how-packages-are-made-available\" class=\"anchor\" aria-hidden=\"true\" href=\"#customizing-how-packages-are-made-available\"><\/a>Customizing how packages are made available<\/h4><p>By setting the variable <code>straight-cache-autoloads<\/code> to a non-nil value, you can cause <code>straight.el<\/code> to cache the autoloads of all used packages in a single file on disk, and load them from there instead of from the individual package files if they are still up to date. This reduces the number of disk IO operations during startup from O(number of packages) to O(1), so it should improve performance. No other configuration should be necessary to make this work; however, you may wish to call <a href=\"#interactive-usage\"><code>straight-prune-build<\/code><\/a> occasionally, since otherwise this cache file may grow quite large over time.<\/p><h4><a id=\"user-content-hooks-run-by-straight-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#hooks-run-by-straight-use-package\"><\/a>Hooks run by <code>straight-use-package<\/code><\/h4><p>Currently, <code>straight-use-package<\/code> supports two hooks:<\/p><ul><li><code>straight-use-package-prepare-functions<\/code>: The functions in this hook are run just before a package would be built, even if the package does not actually need to be rebuilt. They are passed the name of the package being (maybe) built as a string, and should take and ignore any additional arguments.<\/li>\n<li><code>straight-use-package-pre-build-functions<\/code>: The functions in this hook are run just before building a package (and only if the package needs to be built). They are passed the name of the package being built as a string, and should take and ignore any additional arguments.<\/li>\n<\/ul><h3><a id=\"user-content-the-recipe-format\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-recipe-format\"><\/a>The recipe format<\/h3><p>The general format for a <code>straight.el<\/code> recipe is:<\/p><pre>(package-name :keyword value :keyword value ...)\n<\/pre><p>Note that if you wish to pass a recipe to <code>straight-use-package<\/code>, you will need to quote it. If you need to compute part of the recipe dynamically, use backquoting:<\/p><pre>(straight-use-package\n `(el-patch :type git :repo ,(alist-get 'el-patch my-package-urls)))\n<\/pre><p>Here is a comprehensive list of all keywords which have special meaning in a recipe (unknown keywords are ignored but preserved):<\/p><ul><li>\n<p><code>:local-repo<\/code><\/p>\n<p>This is the name of the local repository that is used for the package. If a local repository by that name does not exist when you invoke <code>straight-use-package<\/code>, one will be cloned according to the package's <a href=\"#version-control-backends\">version-control settings<\/a>.<\/p>\n<p>Multiple packages can use the same local repository. If so, then a change to the local repository will cause both packages to be rebuilt. Typically, if multiple packages are drawn from the same repository, both should specify a <code>:files<\/code> directive.<\/p>\n<p>If you do not provide <code>:local-repo<\/code>, then it defaults to a value derived from the <a href=\"#version-control-backends\">version-control settings<\/a>, or as a last resort the package name.<\/p>\n<\/li>\n<li>\n<p><code>:files<\/code><\/p>\n<p>This is a list specifying which files in a package's local repository need to be symlinked into its build directory, and how to arrange the symlinks. For most packages, the default value (<code>straight-default-files-directive<\/code>) will suffice, and you do not need to specify anything.<\/p>\n<p>If you do need to override the <code>:files<\/code> directive (this happens most commonly when you are taking a single package from a repository that holds multiple packages), it is almost always sufficient to just specify a list of globs or filenames. All matching files will be linked into the top level of the package's build directory.<\/p>\n<p>In spite of this, the <code>:files<\/code> directive supports an almost comically powerful DSL (with nested excludes and everything!) that allows you full flexibility on how the links are made; see the docstring of <code>straight-expand-files-directive<\/code> for the full details.<\/p>\n<\/li>\n<li>\n<p><code>:flavor<\/code><\/p>\n<p>If this is non-nil, then it changes the interpretation of the <code>:files<\/code> directive for compatibility with an existing upstream recipe source. See the docstring of <code>straight-expand-files-directive<\/code> for details.<\/p>\n<\/li>\n<li>\n<p><code>:no-build<\/code><\/p>\n<p>If this is non-nil, then it causes the build step to be skipped entirely and unconditionally. You should specify this for <a href=\"#customizing-recipe-repositories\">recipe repository recipes<\/a>.<\/p>\n<\/li>\n<li>\n<p><code>:type<\/code><\/p>\n<p>This specifies the version-control backend to use for cloning and managing the package's local repository. It defaults to the value of <code>straight-default-vc<\/code>, which defaults to <code>git<\/code>.<\/p>\n<p>The only traditional version-control backend currently supported is <code>git<\/code>, although more backends may be added.<\/p>\n<p>As a special case, however, you may specify the value <code>built-in<\/code>, which results in all version-control operations on the package being ignored. This allows you to tell <code>straight.el<\/code> that a package has already been provided (for example, because a version of it is shipped by Emacs) and does not have a local repository which needs to be cloned, updated, and so on. Here is how you can tell <code>straight.el<\/code> that you would like to use the Emacs-provided copy of Org, rather than cloning it from the upstream repository if another package declares it as a dependency:<\/p>\n<pre>(straight-use-package '(org :type built-in))\n<\/pre><\/li>\n<li>\n<p>backend-specific keywords<\/p>\n<p>Depending on the value of <code>:type<\/code>, additional keywords (relevant to how the package's repository is cloned and managed) will be meaningful. See the next section.<\/p>\n<p>The <code>built-in<\/code> pseudo-backend does not take any other keywords.<\/p>\n<\/li>\n<\/ul><h4><a id=\"user-content-version-control-backends\" class=\"anchor\" aria-hidden=\"true\" href=\"#version-control-backends\"><\/a>Version-control backends<\/h4><p>Defining a version-control backend consists of declaring a number of functions named as <code>straight-vc-BACKEND-METHOD<\/code>, where <code>BACKEND<\/code> is the name of the version-control backend being defined and <code>METHOD<\/code> is a backend API method. The relevant methods are:<\/p><ul><li><code>clone<\/code>: given a recipe and a commit object, clone the repository and attempt to check out the given commit.<\/li>\n<li><code>commit-present-p<\/code>: given a recipe and a commit object, return whether the commit can be checked out offline, i.e., without fetching from the remote.<\/li>\n<li><code>normalize<\/code>: given a recipe, \"normalize\" the repository (this generally means reverting it to a standard state, such as a clean working directory, but does not entail checking out any particular commit).<\/li>\n<li><code>fetch-from-remote<\/code>: given a recipe, fetch the latest version from its configured remote, if one is specified.<\/li>\n<li><code>fetch-from-upstream<\/code>: given a recipe, fetch the latest version from its configured upstream, if the package is forked.<\/li>\n<li><code>merge-from-remote<\/code>: given a recipe, merge the latest version fetched from the configured remote, if any, to the local copy.<\/li>\n<li><code>merge-from-upstream<\/code>: given a recipe, merge the latest version fetched from the configured upstream, if the package is forked, to the local copy.<\/li>\n<li><code>push-to-remote<\/code>: given a recipe, push the current version of the repository to its configured remote, if one is specified.<\/li>\n<li><code>check-out-commit<\/code>: given a recipe and a commit object, attempt to check out that commit in the repository for that recipe.<\/li>\n<li><code>get-commit<\/code>: given a local repository name, return the commit object that is currently checked out.<\/li>\n<li><code>local-repo-name<\/code>: given a recipe, return a good name for the local repository, or nil.<\/li>\n<li><code>keywords<\/code>: return a list of keywords which are meaningful for this version-control backend.<\/li>\n<\/ul><p>Most of these methods are highly interactive: they don't actually do anything without prompting you to confirm it, and very often they will offer you a number of different options to proceed (including starting a recursive edit and allowing you to do whatever you would like).<\/p><p>Also, all of the methods in this section take <a href=\"#developer-manual\"><code>straight.el<\/code>-style recipes<\/a>; see the section on <a href=\"#developer-manual\">defining VC backends<\/a> in the developer manual for more details.<\/p><h4><a id=\"user-content-git-backend\" class=\"anchor\" aria-hidden=\"true\" href=\"#git-backend\"><\/a>Git backend<\/h4><p>These are the keywords meaningful for the <code>git<\/code> backend:<\/p><ul><li><code>:repo<\/code>: the clone URL for the Git repository.<\/li>\n<li><code>:host<\/code>: either nil or one of the symbols <code>github<\/code>, <code>gitlab<\/code>, <code>bitbucket<\/code>. If non-nil, then <code>:repo<\/code> should just be a string \"username\/repo\", and the URL is constructed automatically.<\/li>\n<li><code>:branch<\/code>: the name of the branch used for primary development, as a string. If your version lockfiles do not specify a commit to check out when the repository is cloned, then this branch is checked out, if possible. This branch is also viewed as the \"primary\" branch for the purpose of normalization and interaction with the remote.<\/li>\n<li><code>:remote<\/code>: the name to use for the Git remote. If the package is forked, this name is used for the upstream remote.<\/li>\n<li><code>:nonrecursive<\/code>: if non-nil, then submodules are not cloned. This is particularly important for the Emacsmirror recipe repository, which contains every known Emacs package in existence as submodules.<\/li>\n<li><code>:fork<\/code>: a plist which specifies settings for a fork, if desired. This causes the <code>fetch-from-remote<\/code> method to operate on the fork; you can use the <code>fetch-from-upstream<\/code> method to operate on the upstream instead. The allowed keywords are <code>:repo<\/code>, <code>:host<\/code>, <code>:branch<\/code>, and <code>:remote<\/code>.<\/li>\n<li><code>:depth<\/code>: either the symbol <code>full<\/code> or an integer. If <code>full<\/code>, then the repository is cloned with its whole history. If an integer <code>N<\/code>, then the repository is cloned with the option <code>--depth N<\/code>, unless a commit is specified (e.g. by version lockfiles). The default value is <code>full<\/code>.<\/li>\n<\/ul><p>This section tells you how the <code>git<\/code> backend, specifically, implements the version-control backend API:<\/p><ul><li><code>clone<\/code>: clones the repository, including submodules if <code>:nonrecursive<\/code> is not provided. Checks out the commit specified in your revision lockfile, or the <code>:branch<\/code> (from the <code>:fork<\/code> configuration, if given), or <code>origin\/HEAD<\/code>. If a <code>:fork<\/code> is specified, also fetches from the upstream.<\/li>\n<li><code>commit-present-p<\/code>: checks if the commit SHA is among the revisions that are present locally.<\/li>\n<li><code>normalize<\/code>: verifies that remote URLs are set correctly, that no merge is in progress, that the worktree is clean, and that the primary <code>:branch<\/code> (from the <code>:fork<\/code> configuration, if given) is checked out.<\/li>\n<li><code>fetch-from-remote<\/code>: checks that remote URLs are set correctly, then fetches from the primary remote (the fork, if the package is forked).<\/li>\n<li><code>fetch-from-upstream<\/code>: checks that remote URLs are set correctly, then fetches from the upstream remote. If the package is not a fork, does nothing.<\/li>\n<li><code>merge-from-remote<\/code>: performs normalization, then merges from the primary remote (the fork, if the package is forked) into the primary local <code>:branch<\/code>.<\/li>\n<li><code>merge-from-upstream<\/code>: performs normalization, then merges from the upstream remote into the primary local <code>:branch<\/code>. If the package is not a fork, does not attempt to merge.<\/li>\n<li><code>push-to-remote<\/code>: performs normalization, pulls from the primary remote if necessary, and then pushes if necessary. This operation acts on the fork, if the package is forked.<\/li>\n<li><code>check-out-commit<\/code>: verifies that no merge is in progress and that the worktree is clean, then resets the worktree to the specified commit.<\/li>\n<li><code>get-commit<\/code>: returns HEAD as a 40-character string.<\/li>\n<li><code>local-repo-name<\/code>: if <code>:host<\/code> is non-nil, then <code>:repo<\/code> will be of the form \"username\/repository\", and \"repository\" is used. Otherwise, if the URL is of the form <code>...\/&lt;something&gt;.git<\/code>, then <code>&lt;something&gt;<\/code> is used. Otherwise, nil is returned.<\/li>\n<li><code>keywords<\/code>: see the list of keywords above.<\/li>\n<\/ul><p>You can customize the following user options:<\/p><ul><li><code>straight-vc-git-default-branch<\/code>: if <code>:branch<\/code> is unspecified, then this is used instead. Defaults to \"master\".<\/li>\n<li><code>straight-vc-git-default-remote-name<\/code>: the name to use for the primary remote, or the upstream remote if the package is forked. Defaults to \"origin\". The <code>:remote<\/code> keyword may be used to override the value of this variable on a per-repository basis.<\/li>\n<li><code>straight-vc-git-default-fork-name<\/code>: the name to use for the fork remote, if the package is forked. Defaults to \"fork\". The <code>:remote<\/code> keyword may be used to override the value of this variable on a per-repository basis.<\/li>\n<li><code>straight-vc-git-default-protocol<\/code>: the default protocol to use for automatically generated URLs when <code>:host<\/code> is non-nil. It can be either <code>https<\/code> or <code>ssh<\/code>, and defaults to <code>https<\/code> because this requires less work to set up.<\/li>\n<li><code>straight-vc-git-force-protocol<\/code>: if this is non-nil, then HTTPS and SSH URLs are not treated as equivalent, so that bulk version-control operations will offer to re-set your remote URLs from HTTPS to SSH or vice versa, depending on the value of <code>straight-vc-git-default-protocol<\/code>. This is nil by default.<\/li>\n<li><code>straight-vc-git-auto-fast-forward<\/code>: if this is non-nil, pulling will quietly do fast-forward, to suppress asking for instructions on each package with updates, unless they're not trivial. Set to nil if you'd prefer to inspect all changes.<\/li>\n<li><code>straight-vc-git-default-clone-depth<\/code>: the default value for the <code>:depth<\/code> keyword. It can be either the symbol <code>full<\/code> or an integer, and defaults to <code>full<\/code>. Setting this variable to a small integer will reduce the size of repositories. Note that this variable does <em>not<\/em> affect packages whose versions are locked.<\/li>\n<\/ul><h5><a id=\"user-content-deprecated-upstream-keyword\" class=\"anchor\" aria-hidden=\"true\" href=\"#deprecated-upstream-keyword\"><\/a>Deprecated <code>:upstream<\/code> keyword<\/h5><p><code>straight.el<\/code> previously supported fork configuration in recipes using an <code>:upstream<\/code> keyword rather than a <code>:fork<\/code> keyword. For various reasons, this was more complex to handle, which is why the change was made. For backwards compatibility, the <code>:upstream<\/code> keyword is still accepted, with the following behavior.<\/p><p>When <code>straight.el<\/code> processes a recipe which uses the <code>:upstream<\/code> keyword, it moves the <code>:repo<\/code>, <code>:host<\/code>, and <code>:branch<\/code> keywords from that sub-plist to the top level, and moves those top-level keywords to a new <code>:fork<\/code> sub-plist. Then it sets the top-level and <code>:fork<\/code> sub-plist values of <code>:remote<\/code> to the values of the deprecated variables <code>straight-vc-git-upstream-remote<\/code> (defaults to \"upstream\") and <code>straight-vc-git-primary-remote<\/code> (defaults to \"origin\"), respectively.<\/p><p>For backwards compatibility, if <code>straight-vc-git-primary-remote<\/code> differs from its default value of \"origin\", then its value is used in place of <code>straight-vc-git-default-remote-name<\/code>.<\/p><h3><a id=\"user-content-recipe-lookup\" class=\"anchor\" aria-hidden=\"true\" href=\"#recipe-lookup\"><\/a>Recipe lookup<\/h3><p>If you only provide a symbol (package name) to <code>straight-use-package<\/code>, then the recipe is looked up automatically. By default, <a href=\"http:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a>, <a href=\"https:\/\/elpa.gnu.org\/\" rel=\"nofollow\">GNU ELPA<\/a>, and <a href=\"https:\/\/emacsmirror.net\/\" rel=\"nofollow\">Emacsmirror<\/a> are searched for recipes, in that order. This means that one or more of them may need to be cloned. Recipe repositories are actually just the same as ordinary packages, except that their recipes specify <code>:no-build<\/code>, so they are not symlinked or added to the <code>load-path<\/code> or anything.<\/p><p>Note that dependencies always use the default recipes, since the only information <code>straight.el<\/code> gets about a package's dependencies are their names.<\/p><p>This leads to a few interesting questions regarding requesting a package multiple times. For example, you might need to load two features using <a href=\"https:\/\/github.com\/jwiegley\/use-package\"><code>use-package<\/code><\/a> that are provided from the same package, or one of the packages you have installed is also requested as a dependency by another package. <code>straight.el<\/code> uses a number of heuristics to try to make these interactions as intuitive and painless as possible:<\/p><ul><li>\n<p>The first time a package is registered with <code>straight.el<\/code>, its recipe (either the recipe that you provided, or the one that was looked up from a recipe repository) is recorded. In future registration, if you just provide the package name to <code>straight-use-package<\/code>, the existing recipe is reused.<\/p>\n<p>Note, however: <em>if<\/em> you want to use a custom recipe for a given package, you must load it <em>before<\/em> all of its dependencies. Otherwise, the package will first be registered as a dependency, using the default recipe.<\/p>\n<\/li>\n<li>\n<p>If a package has already been registered with <code>straight.el<\/code>, and you attempt to load it again with an explicit recipe which is different from the one previously registered, the new recipe is used but a warning is signalled.<\/p>\n<\/li>\n<li>\n<p>If you attempt to register a package which shares a <code>:local-repo<\/code> (either by default, or due to explicit specification) with a previously registered package, and the two packages specify different values for their version-control keywords (see <a href=\"#version-control-backends\">version-control backends<\/a>), then the new recipe is used but a warning is signalled. If the repository was already cloned, this means the second recipe will have no effect.<\/p>\n<p>But if the second recipe was fetched automatically from a recipe repository, all of its version-control keywords will be silently overwritten with the ones from the first recipe, to avoid conflicts (although if there are conflicts in other parts of the recipe, a warning will still be displayed).<\/p>\n<\/li>\n<\/ul><h4><a id=\"user-content-updating-recipe-repositories\" class=\"anchor\" aria-hidden=\"true\" href=\"#updating-recipe-repositories\"><\/a>Updating recipe repositories<\/h4><p>As mentioned in the <a href=\"#where-do-recipes-come-from\">conceptual overview<\/a>, recipe repositories are just regular packages, with some extra code to look up recipes in the relevant local repository.<\/p><p>This means that updating a recipe repository is done the same way as updating a regular package, i.e. with <a href=\"#version-control-operations\"><code>M-x straight-pull-package<\/code><\/a>. You should do this if you find that a package isn't listed by <code>M-x straight-use-package<\/code>\u2014perhaps it was added recently.<\/p><p>Note that there is currently some potentially surprising behavior if you update all packages at once using <code>M-x straight-pull-all<\/code> or <code>M-x straight-merge-all<\/code>, and this bulk update includes recipe repository updates: see <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/323\">#323<\/a>.<\/p><h4><a id=\"user-content-customizing-recipe-repositories\" class=\"anchor\" aria-hidden=\"true\" href=\"#customizing-recipe-repositories\"><\/a>Customizing recipe repositories<\/h4><p>The recipe repository system is designed to be extended. Firstly, you can control which recipe repositories are searched, and in what order of precedence, by customizing <code>straight-recipe-repositories<\/code>. The default value is:<\/p><pre>(org-elpa melpa gnu-elpa-mirror emacsmirror)\n<\/pre><h5><a id=\"user-content-gnu-elpa\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnu-elpa\"><\/a>GNU ELPA<\/h5><p>You can customize the following user options:<\/p><ul><li>\n<p><code>straight-recipes-gnu-elpa-url<\/code>: The Git URL to use for the <code>gnu-elpa<\/code> recipe repository.<\/p>\n<\/li>\n<li>\n<p><code>straight-recipes-gnu-elpa-use-mirror<\/code>: GNU ELPA uses a stupidly complex build system for no good reason, and it's unlikely to change any time soon. What this means for you is that you have to run the Elisp-powered Makefile of the GNU ELPA repository (which has a fatal bug last I checked, so you'd have to patch it locally) <em>and<\/em> have a full checkout of the Emacs source repository (more than 1GB) if you want all the packages to work correctly. To work around this problem, I maintain a <a href=\"https:\/\/github.com\/emacs-straight\">full mirror of GNU ELPA on GitHub<\/a>. (The tooling used to maintain this mirror is located <a href=\"https:\/\/github.com\/raxod502\/gnu-elpa-mirror\">here<\/a>.) By default, <code>straight.el<\/code> retrieves packages from this mirror instead of the source repository; this behavior is controlled by the value of <code>straight-recipes-gnu-elpa-use-mirror<\/code>. You must do any customization of this variable <em>before<\/em> the <code>straight.el<\/code> <a href=\"#getting-started\">bootstrap<\/a>. Note that setting the value of this user option to nil causes the default value of <code>straight-recipe-repositories<\/code> to shift to:<\/p>\n<pre>(org-elpa melpa gnu-elpa emacsmirror)\n<\/pre><\/li>\n<\/ul><h5><a id=\"user-content-emacsmirror\" class=\"anchor\" aria-hidden=\"true\" href=\"#emacsmirror\"><\/a>Emacsmirror<\/h5><p>You can customize the following user option:<\/p><ul><li>\n<p><code>straight-recipes-emacsmirror-use-mirror<\/code>: Yes, there is also a mirror for Emacsmirror. This is because the <a href=\"https:\/\/github.com\/emacsmirror\/epkgs\">epkgs<\/a> repository contains a (frequently updated) SQLite database in it, which means the Git repository takes <em>forever<\/em> to clone (see <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/356\">#356<\/a>). My solution to this problem is to generate a new repository which contains the information that <code>straight.el<\/code> needs but which is much smaller. By default, <code>straight.el<\/code> uses the official <a href=\"https:\/\/github.com\/emacsmirror\/epkgs\">epkgs<\/a> repository to find packages on Emacsmirror, but you can tell it to use my mirror by configuring the value of this variable to non-nil. You must do any customization of this variable <em>before<\/em> the <code>straight.el<\/code> <a href=\"#getting-started\">bootstrap<\/a>. Note that setting the value of this user option to non-nil causes the default value of <code>straight-recipe-repositories<\/code> to shift to:<\/p>\n<pre>(org-elpa melpa gnu-elpa-mirror emacsmirror-mirror)\n<\/pre><\/li>\n<\/ul><h5><a id=\"user-content-defining-new-recipe-repositories\" class=\"anchor\" aria-hidden=\"true\" href=\"#defining-new-recipe-repositories\"><\/a>Defining new recipe repositories<\/h5><p>To define a new recipe repository called <code>NAME<\/code>, you should do the following things:<\/p><ul><li>Define a function <code>straight-recipes-NAME-retrieve<\/code>, which takes a package name as a symbol and returns a recipe for that package if it is available, else nil. This is used for recipe lookup. This function may assume that the local repository for the recipe repository has already been cloned, and that <code>default-directory<\/code> has been set to that local repository. This is used for recipe lookup during the course of <code>straight-use-package<\/code>.<\/li>\n<li>Define a function <code>straight-recipes-NAME-list<\/code>, which takes no arguments and returns a list of strings representing packages for which recipes are available. It is permissible to return some strings for which recipes are actually not available, for performance reasons. However, this is discouraged. (The <a href=\"http:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a> backend uses this functionality, since all files in the <code>recipes<\/code> directory are potentially recipes, but only the Git-based ones can actually be used.)<\/li>\n<li>(Optional) Define a function <code>straight-recipes-NAME-version<\/code> which returns a non-nil value indicating the current version of the logic in your <code>straight-recipes-NAME-retrieve<\/code> function. Each time you change the logic, this version value must be changed. If this function is defined, then <code>straight.el<\/code> automatically and transparently caches calls to <code>straight-recipes-NAME-retrieve<\/code> persistently, using your version value (and its detection of modifications to the recipe repository) to decide when to invalidate the cache.<\/li>\n<li>Call <code>straight-use-recipes<\/code> with the recipe for your recipe repository. Make sure to include <code>:no-build<\/code> in the recipe, unless you also want to use the recipe repository as an executable Emacs Lisp package. Alternatively, you can take the manual approach:\n<ul><li>Call <code>straight-use-package-lazy<\/code> with the recipe for your recipe repository.<\/li>\n<li>Add the symbol for your recipe repository's name (the car of the recipe you provided, that is) to <code>straight-recipe-repositories<\/code>, at the appropriate place.<\/li>\n<\/ul><\/li>\n<\/ul><h3><a id=\"user-content-overriding-recipes\" class=\"anchor\" aria-hidden=\"true\" href=\"#overriding-recipes\"><\/a>Overriding recipes<\/h3><p>You can always use <code>straight-register-package<\/code> to specify a specific recipe for a package without cloning or building it, so that just in case that package is requested later (possibly as a dependency, or in somebody else's code) your recipe will be used instead of the default one. However, this does not help in the case that a specific recipe is passed to <code>straight-use-package<\/code>.<\/p><p>Also, it is obviously impossible to call <code>straight-register-package<\/code> before <code>straight.el<\/code> has been loaded, so you can't use it to specify a custom recipe for <code>straight.el<\/code> itself.<\/p><p>To remedy these difficulties, <code>straight.el<\/code> provides a mechanism for specifically overriding the recipe for a particular package. You can use it by customizing <code>straight-recipe-overrides<\/code>, or by calling <code>straight-override-recipe<\/code>.<\/p><p><code>straight-recipe-overrides<\/code> is an association list from <a href=\"#the-profile-system\">profile names<\/a> to <em>override alists<\/em>. If you don't care about the profile system, you can just use a single override specification, with the profile name nil. Each override alist is just a list of recipes. Because the car of a recipe is just the package name as a symbol, this list of recipes is also an alist whose keys are recipe names and whose values are the plists for those recipes.<\/p><p>Even if an explicit recipe is supplied to <code>straight-use-package<\/code>, the one given in <code>straight-recipe-overrides<\/code> will be used instead, if such a recipe is specified there.<\/p><p>For convenience, you may add to <code>straight-recipe-overrides<\/code> by passing a recipe to <code>straight-override-recipe<\/code>. This will register it in the override alist for the current profile. Note that if you do this, you will probably want to explicitly set <code>straight-recipe-overrides<\/code> to nil before bootstrapping <code>straight.el<\/code>. This will make it so that if you remove a call to <code>straight-override-recipe<\/code> from your init-file and then reload it, the entry will actually be removed from <code>straight-recipe-overrides<\/code>.<\/p><h4><a id=\"user-content-overriding-the-recipe-for-straightel\" class=\"anchor\" aria-hidden=\"true\" href=\"#overriding-the-recipe-for-straightel\"><\/a>Overriding the recipe for <code>straight.el<\/code><\/h4><p>As was briefly mentioned earlier, you can actually override the recipe of <code>straight.el<\/code> itself using <code>straight-recipe-overrides<\/code>! How does this work? Well, it's basically black magic. If you want the details, go read the <a href=\"#developer-manual\">developer manual<\/a>. All you need to know is that you can set <code>straight-recipe-overrides<\/code>, and it will magically work. The only caveat is that if you change the <code>:local-repo<\/code> for <code>straight.el<\/code>, then you will also need to adjust the value of <code>bootstrap-file<\/code> in the <a href=\"#getting-started\">bootstrap snippet<\/a> accordingly, since otherwise your init-file will not know where to find <code>straight.el<\/code>. (You must use <code>straight-recipe-overrides<\/code> instead of <code>straight-override-recipe<\/code>, since the latter function definition hasn't been loaded yet before <code>straight.el<\/code> is installed and bootstrapped.)<\/p><p>Here is the default recipe used for <code>straight.el<\/code>, if you don't override it:<\/p><pre>(straight :type git :host github\n          :repo ,(format \"%s\/straight.el\" straight-repository-user)\n          :files (\"straight*.el\")\n          :branch ,straight-repository-branch)\n<\/pre><p>Note that even though the bootstrap snippet references the <code>develop<\/code> branch of <code>straight.el<\/code>, the default recipe installs from <code>master<\/code>.<\/p><p>If all you want to do is change which branch you are installing <code>straight.el<\/code> from, simply customize the variable <code>straight-repository-branch<\/code>, which is provided for this purpose. (Although using <code>straight-recipe-overrides<\/code> will work just as well, at least until the recipe happens to be changed upstream and your init-file isn't updated.)<\/p><p>Similarly, if all you want to do is switch to your own fork of <code>straight.el<\/code> on GitHub, simply customize the variable <code>straight-repository-user<\/code> to your GitHub username.<\/p><p>There is one minor caveat to the above discussion. If your fork makes changes to the way in which recipes are interpreted, then those changes will not be effective during the interpretation of your own recipe for <code>straight.el<\/code>. If you wish for them to be, then you will have to follow the same procedure that is followed in <code>straight.el<\/code> itself for making changes to recipe interpretation. These details are outlined in the <a href=\"#developer-manual\">developer manual<\/a>; see also <code>install.el<\/code> for an explanation of this aspect of the bootstrap mechanism.<\/p><h3><a id=\"user-content-interactive-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#interactive-usage\"><\/a>Interactive usage<\/h3><p>The primary usage of <code>straight.el<\/code> is expected to be in your init-file. For example, this is where you will need to put the bootstrap code as well as any packages that you always want to be installed. However, there are three important interactive uses of <code>straight.el<\/code>: temporary installation of packages, various helpful utility functions, and <a href=\"#version-control-operations\">version control operations<\/a>.<\/p><p>To install a package temporarily, run <code>M-x straight-use-package<\/code>. All registered recipe repositories will be cloned, and you will be presented with a combined list of all recipes available from them. Simply select a package and it will be cloned, built, and loaded automatically. This does not affect future Emacs sessions.<\/p><p>If you provide a prefix argument to <code>M-x straight-use-package<\/code>, then you are presented with a list of registered recipe repositories. After you select one, you are shown a list of recipes specifically from that recipe repository. This is helpful if you do not want to clone all registered recipe repositories, or you have a particular recipe repository in mind.<\/p><p>You can also call <code>M-x straight-get-recipe<\/code>, which has the same interface as <code>M-x straight-use-package<\/code>, except that instead of the package being cloned, built, and loaded, its recipe is copied to the kill ring. If you are writing a custom recipe, this may be helpful, because you may be able to reuse parts of the existing recipe, particularly the <code>:files<\/code> directive.<\/p><p>Normally, packages are rebuilt automatically if needed, when Emacs restarts. If you for some reason want them to be rebuilt at another time, you can call <code>M-x straight-check-all<\/code> to rebuild all packages that have been modified since their last build. Alternatively, use <code>M-x straight-rebuild-all<\/code> to unconditionally rebuild all packages. Note that this will probably take a while. There are also <code>M-x straight-check-package<\/code> and <code>M-x straight-rebuild-package<\/code>, which allow you to select a particular package to check or rebuild.<\/p><p>Finally, you may use <code>M-x straight-prune-build<\/code> in order to tell <code>straight.el<\/code> to forget about any packages which were not registered since the last time you loaded your init-file. This may improve performance, although only slightly, and will clean up stale entries in the <code>build<\/code> directory. You can call this function in your init-file if you really wish your filesystem to be as clean as possible, although it's not particularly recommended as the performance implications are uninvestigated. If you do call it in your init-file, be sure to only call it on a fully successful init; otherwise, an error during init will result in some packages' build information being discarded, and they will need to be rebuilt next time.<\/p><p>If you have enabled <a href=\"#customizing-how-packages-are-made-available\">autoloads caching<\/a>, it is advisable to call <code>straight-prune-build<\/code> occasionally, since otherwise the build cache may grow quite large over time.<\/p><h4><a id=\"user-content-version-control-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#version-control-operations\"><\/a>Version control operations<\/h4><p><code>straight.el<\/code> provides a number of highly interactive workflows for managing your package's local repositories, using the configured <a href=\"#version-control-backends\">version-control backends<\/a>. They are as follows:<\/p><ul><li><code>M-x straight-normalize-package<\/code>: normalize a package<\/li>\n<li><code>M-x straight-normalize-all<\/code>: normalize all packages<\/li>\n<li><code>M-x straight-fetch-package<\/code>: fetch from a package's configured remote; with prefix argument, then for forks also fetch from the upstream<\/li>\n<li><code>M-x straight-fetch-package-and-deps<\/code>: fetch from the configured remotes of a package and all of its dependencies (including the dependencies of its dependencies); with prefix argment, then for forks also fetch from the upstream<\/li>\n<li><code>M-x straight-fetch-all<\/code>: fetch from all packages' configured remotes; with prefix argument, then for forks also fetch from the upstreams<\/li>\n<li><code>M-x straight-merge-package<\/code>: merge the latest version fetched from a package's configured remote into the local copy; with prefix argument, then for forks also merge from the upstream<\/li>\n<li><code>M-x straight-merge-package-and-deps<\/code>: merge the latest versions fetched from the configured remotes of a package and all of its dependencies (including the dependencies of its dependencies); with prefix argment, then for forks also merge from the upstreams<\/li>\n<li><code>M-x straight-merge-all<\/code>: merge the latest versions fetched from each package's configured remote into its local copy; with prefix argument, then for forks also merge from the upstreams<\/li>\n<li><code>M-x straight-pull-package<\/code>: combination of <code>M-x straight-fetch-package<\/code> and <code>M-x straight-merge-package<\/code><\/li>\n<li><code>M-x straight-pull-package-and-deps<\/code>: combination of <code>M-x straight-fetch-package-and-deps<\/code> and <code>M-x straight-merge-package-and-deps<\/code><\/li>\n<li><code>M-x straight-pull-all<\/code>: combination of <code>M-x straight-fetch-all<\/code> and <code>M-x straight-merge-all<\/code><\/li>\n<li><code>M-x straight-push-package<\/code>: push a package to its remote, if necessary<\/li>\n<li><code>M-x straight-push-all<\/code>: push all packages to their remotes, if necessary<\/li>\n<\/ul><p>See the sections on <a href=\"#version-control-backends\">version-control backends<\/a> and the <a href=\"#git-backend\">Git backend<\/a> in particular for more information about the meanings of these operations.<\/p><h3><a id=\"user-content-lockfile-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#lockfile-management\"><\/a>Lockfile management<\/h3><p><code>straight.el<\/code> determines your package management configuration from two, and only two, sources: the contents of your init-file, and your version lockfiles (which are optional). Your init-file specifies the configuration of <code>straight.el<\/code> (for example, the values of <code>straight-recipe-overrides<\/code> and <code>straight-default-vc<\/code>), the packages you want to use, and their recipes. Your version lockfiles specify the exact revisions of each package, recipe repository, and even <code>straight.el<\/code> itself. Together, they lock down your Emacs configuration to a state of no uncertainty: perfect reproducibility.<\/p><p>To write the current revisions of all your packages into version lockfiles, run <code>M-x straight-freeze-versions<\/code>. This will first check that <code>straight.el<\/code> has an up-to-date account of what packages are installed by your init-file, then ensure that all your local changes are pushed (remember, we are aiming for perfect reproducibility!). If you wish to bypass these checks, provide a prefix argument.<\/p><p>Version lockfiles are written into <code>~\/.emacs.d\/straight\/versions<\/code>. By default, there will be one, called <code>default.el<\/code>. It is recommended that you keep your version lockfiles under version control with the rest of your Emacs configuration. If you symlink your init-file into <code>~\/.emacs.d<\/code> from somewhere else, you should also make sure to symlink your version lockfiles into <code>~\/.emacs.d\/straight\/versions<\/code>. On a new machine, do this <em>before<\/em> launching Emacs: that way, <code>straight.el<\/code> can make sure to check out the specified revisions of each package when cloning them for the first time.<\/p><p>To install the versions of the packages specified in your version lockfiles, run <code>M-x straight-thaw-versions<\/code>. Thawing will interactively check for local changes before checking out the relevant revisions, so don't worry about things getting overwritten.<\/p><h4><a id=\"user-content-the-profile-system\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-profile-system\"><\/a>The profile system<\/h4><p><code>straight.el<\/code> has support for writing multiple version lockfiles, instead of just one. Why? Consider a large Emacs configuration such as <a href=\"https:\/\/github.com\/raxod502\/radian\">Radian<\/a>, <a href=\"http:\/\/spacemacs.org\/\" rel=\"nofollow\">Spacemacs<\/a>, or <a href=\"https:\/\/github.com\/bbatsov\/prelude\">Prelude<\/a>, which is used by many different people. There are two parts to the configuration that is actually loaded: the \"default\" part, and the local customizations that each user has added. Generally, these configurations have a mechanism for making local customizations without forking the entire project.<\/p><p>So Radian will have some set of packages that it requires, and my local customizations of Radian have some other set of packages that they require. In order for me to maintain Radian, I need to be able to separate Radian's packages (which go into a versions lockfile in the Radian repository) from my own local packages (which go into a versions lockfile in my own private local dotfiles repository). <code>straight.el<\/code> provides this ability through the <em>profile system<\/em>.<\/p><p>The idea is that whenever a package is registered, either directly or as a dependency, it is associated with a given profile. Any given package can be associated with multiple profiles.<\/p><p>When you call <code>straight-use-package<\/code>, which profile the registered packages are associated with is determined by the value of <code>straight-current-profile<\/code>, which defaults to nil. In Radian, for example, <code>straight-current-profile<\/code> is bound to <code>radian<\/code> while the Radian libraries are being loaded, and it is bound to <code>radian-local<\/code> while the user's local customizations are being loaded. This results in Radian packages being associated with the <code>radian<\/code> profile, and the user's local packages being associated with the <code>radian-local<\/code> profile.<\/p><p>When you call <code>M-x straight-freeze-versions<\/code>, one or more version lockfiles are written, according to the value of <code>straight-profiles<\/code>. This variable is an association list whose keys are symbols naming profiles and whose values are filenames for the corresponding version lockfiles to be written into <code>~\/.emacs.d\/straight\/versions<\/code>. You should make sure that each potential value of <code>straight-current-profile<\/code> has a corresponding entry in <code>straight-profiles<\/code>, since otherwise some packages might not be written into your lockfiles.<\/p><p>When customizing <a href=\"#overriding-recipes\"><code>straight-recipe-overrides<\/code><\/a>, note that if multiple profiles are set to override the same recipe, then the last one listed in <code>straight-profiles<\/code> will take precedence. Similarly, when using <code>M-x straight-thaw-versions<\/code>, if different lockfiles specify revisions for the same local repository, the last one in <code>straight-profiles<\/code> will take precedence.<\/p><h3><a id=\"user-content-packages-and-the-init-file\" class=\"anchor\" aria-hidden=\"true\" href=\"#packages-and-the-init-file\"><\/a>Packages and the init-file<\/h3><p>Package managers like <code>package.el<\/code> store mutable state outside your init-file, including the set of packages that are installed. <code>straight.el<\/code> does not do this, so it has a rather different way of determining what packages are installed. To <code>straight.el<\/code>, a package is part of your Emacs configuration if it is passed to <code>straight-use-package<\/code> when your init-file is loaded.<\/p><p>Note that this means packages installed interactively (using <code>M-x straight-use-package<\/code>) are not considered part of your Emacs configuration, since the invocation of <code>straight-use-package<\/code> does not happen in your init-file.<\/p><p>This raises an interesting question: if you <em>add<\/em> a package to your init-file, how can you convince <code>straight.el<\/code> that it really is part of your init-file, and not just part of a temporary <code>straight-use-package<\/code> form that you evaluated ad-hoc? The answer is simple: <em>reload your entire init-file<\/em>. That way, <code>straight.el<\/code> will see whether or not that package is registered during your init-file.<\/p><p>One might ask how <code>straight.el<\/code> determines that you have finished loading your init-file. The answer is simple: <code>run-with-idle-timer<\/code> is used to execute code only after the current interactive operation has finished. The implementation of this concept is part of the <em>transaction system<\/em> of <code>straight.el<\/code>, and it is also used to amortize certain performance costs when many calls to <code>straight-use-package<\/code> are made sequentially. However, since the transaction system (at least in recent versions of <code>straight.el<\/code>) operates transparently, its details are relegated to the <a href=\"#developer-manual\">developer manual<\/a>.<\/p><h3><a id=\"user-content-using-straightel-to-reproduce-bugs\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-straightel-to-reproduce-bugs\"><\/a>Using <code>straight.el<\/code> to reproduce bugs<\/h3><p>One of the major reasons I wanted to write <code>straight.el<\/code> was that existing package managers were not good for reproducing bugs. For instance, some of them would load all installed packages when the package manager was initialized! Obviously that is not acceptable for a \"minimal test case\".<\/p><p>On the contrary, bootstrapping <code>straight.el<\/code> does not load anything except for <code>straight.el<\/code> itself (the default recipe repositories are registered, but not cloned until needed). You should normally be loading <code>straight.el<\/code> by means of the <a href=\"#getting-started\">bootstrap snippet<\/a>, but when you are in <code>emacs -Q<\/code>, here is how you can initialize <code>straight.el<\/code>:<\/p><pre>M-x load-file RET ~\/.emacs.d\/straight\/repos\/straight.el\/bootstrap.el RET\n<\/pre><p>You can also do this from the command line, perhaps by creating an alias for it:<\/p><pre>$ emacs -Q -l ~\/.emacs.d\/straight\/repos\/straight.el\/bootstrap.el\n<\/pre><p>Let's say you are making a bug report for Projectile. To load just Projectile and all of its dependencies, run:<\/p><pre>M-x straight-use-package RET projectile RET\n<\/pre><p>Note that this will use the currently checked-out revisions of Projectile and all of its dependencies, so you should take note of those in order to make your bug report.<\/p><h3><a id=\"user-content-integration-with-other-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-other-packages\"><\/a>Integration with other packages<\/h3><h4><a id=\"user-content-integration-with-use-package-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-use-package-1\"><\/a>Integration with <code>use-package<\/code><\/h4><p>By default, <code>straight.el<\/code> installs a new keyword <code>:straight<\/code> for <code>use-package<\/code> which may be used to install packages via <code>straight.el<\/code>. The algorithm is extremely simple. This:<\/p><pre>(use-package el-patch\n  :straight t)\n<\/pre><p>macroexpands (essentially) to:<\/p><pre>(straight-use-package 'el-patch)\n<\/pre><p>And this:<\/p><pre>(use-package el-patch\n  :straight (:host github :repo \"raxod502\/el-patch\"\n             :branch \"develop\"))\n<\/pre><p>becomes:<\/p><pre>(straight-use-package\n '(el-patch :host github :repo \"raxod502\/el-patch\"\n            :branch \"develop\"))\n<\/pre><p>If the feature you are requiring with <code>use-package<\/code> is different from the package name, you can provide a full recipe:<\/p><pre>(use-package tex-site\n  :straight (auctex :host github\n                    :repo \"emacsmirror\/auctex\"\n                    :files (:defaults (:exclude \"*.el.in\"))))\n<\/pre><p>And you may also provide just the package name:<\/p><pre>(use-package tex-site\n  :straight auctex)\n<\/pre><p>If you don't provide <code>:straight<\/code>, then by default nothing happens. You may customize <code>straight-use-package-by-default<\/code> to make it so that <code>:straight t<\/code> is assumed unless you explicitly override it with <code>:straight nil<\/code>.<\/p><p>Previously, <code>straight.el<\/code> used a different syntax for its <code>use-package<\/code> integration. For backwards compatibility, you can use this syntax instead by customizing <code>straight-use-package-version<\/code>.<\/p><p>You can disable <code>use-package<\/code> integration entirely by customizing <code>straight-enable-use-package-integration<\/code>.<\/p><h4><a id=\"user-content-integration-with-packageel\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-packageel\"><\/a>\"Integration\" with <code>package.el<\/code><\/h4><p>By default, <code>package.el<\/code> will automatically insert a call to <code>package-initialize<\/code> into your init-file as soon as Emacs starts, which is ridiculous. It will also do this when you perform any package management operation. A separate system inserts some <code>custom<\/code> forms into your init-file when you install a package. <code>straight.el<\/code> disables all of these \"features\" by setting <code>package-enable-at-startup<\/code> to nil and enabling some advices. You can override this behavior by customizing <code>straight-enable-package-integration<\/code>, however.<\/p><p>To help avoid you shooting yourself in the foot by using both <code>:ensure<\/code> and <code>:straight<\/code> at the same time in a <code>use-package<\/code> form (which would cause the same package to be installed twice using two different package managers), <code>straight.el<\/code> will helpfully disable <code>:ensure<\/code> whenever you include <code>:straight<\/code> in a <code>use-package<\/code> form. See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/425\">#425<\/a>.<\/p><h4><a id=\"user-content-integration-with-org\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-org\"><\/a>Integration with Org<\/h4><p>Org expects you to run <code>make<\/code> in its source repository before you run it, but <code>straight.el<\/code> does not yet support running such build systems automatically (see <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/72\">#72<\/a>). This presents two problems:<\/p><ul><li>Byte-compiling Org without running <code>make<\/code> first produces some annoying warnings.<\/li>\n<li>Running <code>make<\/code> generates a file <code>org-version.el<\/code> which provides the functions <code>org-git-version<\/code> and <code>org-release<\/code>. Thus the version of Org provided by <code>straight.el<\/code> does not include these functions, but the obsolete version of Org provided by Emacs (see <a href=\"#the-wrong-version-of-my-package-was-loaded\">the FAQ<\/a>) does. This can result in the obsolete version getting partially loaded, which is confusing.<\/li>\n<\/ul><p>See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/211\">#211<\/a> for discussion.<\/p><p>By default, <code>straight.el<\/code> installs a hack (namely, defining the functions <code>org-git-version<\/code> and <code>org-release<\/code> itself) whenever you ask it to install Org. This functionality is implemented using <a href=\"#hooks-run-by-straight-use-package\"><code>straight-use-package-prepare-functions<\/code><\/a>. You can disable it by setting the value of the variable <code>straight-fix-org<\/code> to nil.<\/p><p>Please be careful with setting <code>straight-vc-git-default-clone-depth<\/code>, which may break some packages' installing processes such as <code>elfeed<\/code> that depend on <code>org<\/code>.<\/p><h4><a id=\"user-content-integration-with-hydra\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-hydra\"><\/a>Integration with Hydra<\/h4><p>See <a href=\"https:\/\/github.com\/abo-abo\/hydra\/wiki\/straight.el\">the Hydra wiki<\/a>.<\/p><h3><a id=\"user-content-miscellaneous\" class=\"anchor\" aria-hidden=\"true\" href=\"#miscellaneous\"><\/a>Miscellaneous<\/h3><ul><li>\n<p>By default, <code>straight.el<\/code> explains what it is doing in the echo area, like this:<\/p>\n<pre>Looking for cider recipe \u2192 Cloning melpa...\n<\/pre>\n<p>If your terminal does not support Unicode characters nicely, you can customize <code>straight-arrow<\/code> to display something else for the arrow.<\/p>\n<\/li>\n<li>\n<p>By default, <code>straight.el<\/code> reports process output the <code>*straight-process*<\/code> buffer. You can customize the name of this buffer via the <code>straight-process-buffer<\/code> user option. If you want to hide this buffer by default, consider adding a leading space to the name.<\/p>\n<\/li>\n<li>\n<p>You can prevent <code>straight.el<\/code> from making any modifications to the filesystem (though it may still read) by customizing the user option <code>straight-safe-mode<\/code> to non-nil. This may be useful for running tasks automatically in batch mode, to avoid multiple concurrent Emacs processes all making changes to the filesystem. For an example of how this feature may be used to safely implement asynchronous byte-compilation of the init-file on successful startup, see <a href=\"https:\/\/github.com\/raxod502\/radian\">Radian<\/a>.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-developer-manual\" class=\"anchor\" aria-hidden=\"true\" href=\"#developer-manual\"><\/a>Developer manual<\/h2><p>This section tells you about all the interesting implementation details and design decisions that power <code>straight.el<\/code> behind the scenes. It assumes you have already read the <a href=\"#user-manual\">user manual<\/a> and the <a href=\"#conceptual-overview\">conceptual overview<\/a>.<\/p><p>Unfortunately, I haven't yet had time to write it. See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/51\">#51<\/a>.<\/p><h2><a id=\"user-content-trivia\" class=\"anchor\" aria-hidden=\"true\" href=\"#trivia\"><\/a>Trivia<\/h2><p>This section has random, (possibly) interesting tidbits about <code>straight.el<\/code> that don't fit in the other sections.<\/p><h3><a id=\"user-content-comments-and-docstrings\" class=\"anchor\" aria-hidden=\"true\" href=\"#comments-and-docstrings\"><\/a>Comments and docstrings<\/h3><p>How did I get that statistic about the percentage of <code>straight.el<\/code> that is comments and docstrings? Simple: by abusing the syntax highlighting.<\/p><pre>(let ((lines (make-hash-table :test #'equal)))\n  (goto-char (point-min))\n  (while (&lt; (point) (point-max))\n    (when (memq (face-at-point)\n                '(font-lock-comment-face\n                  font-lock-doc-face))\n      (puthash (line-number-at-pos) t lines))\n    (forward-char))\n  (* (\/ (float (length (hash-table-keys lines)))\n        (line-number-at-pos))\n     100))\n<\/pre><p>Note that you will have to scroll through the entire buffer first, since <code>font-lock-mode<\/code> computes syntax highlighting lazily.<\/p><h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2><p>Please do! Development takes place on the <code>develop<\/code> branch. You can switch to that branch with<\/p><pre>(setq straight-repository-branch \"develop\")\n<\/pre><p>and base your pull requests from it. If you have an outstanding pull request whose features you would like to use in your configuration, there is full support for defining <code>straight.el<\/code> as coming from any branch in any fork:<\/p><pre>(setq straight-repository-user \"my-github-username\")\n(setq straight-repository-branch \"feat\/my-cool-feature\")\n<\/pre><p>Please try to follow the style of the surrounding code and documentation, but anything is welcome. All text, including comments and docstrings, should be formatted using Emacs' <code>M-x fill-paragraph<\/code> command (bound to <code>M-q<\/code> by default).<\/p><p>We require that the linting pass on all new commits. You can check this easily by installing <a href=\"https:\/\/www.docker.com\/\" rel=\"nofollow\">Docker<\/a> and running<\/p><pre>$ make docker\n<\/pre><p>Then you will be in a shell with the <code>straight.el<\/code> source code, and to test your commits you'll be able to run<\/p><pre>$ make lint\n<\/pre><p>or to see what other targets are available<\/p><pre>$ make help\n<\/pre><p>If you don't want to install Docker, you can simply skip the <code>make docker<\/code> step, and everything will work the same! However, you'll then have to make sure you have the relevant dependencies installed locally.<\/p><p>When you create a pull request, it will be <a href=\"https:\/\/circleci.com\/gh\/raxod502\/straight.el\" rel=\"nofollow\">tested automatically<\/a> on <a href=\"https:\/\/circleci.com\/\" rel=\"nofollow\">CircleCI<\/a> and the status will be reported. Please make sure the CI build is passing before asking for review.<\/p><h2><a id=\"user-content-faq\" class=\"anchor\" aria-hidden=\"true\" href=\"#faq\"><\/a>FAQ<\/h2><h3><a id=\"user-content-my-init-time-got-slower\" class=\"anchor\" aria-hidden=\"true\" href=\"#my-init-time-got-slower\"><\/a>My init time got slower<\/h3><p>Your first step should be to customize the value of <code>straight-check-for-modifications<\/code>. The best setting is <code>(watch-files find-when-checking)<\/code>; this is not enabled by default because it is impolite to spawn persistent background processes without asking, and because you must install <a href=\"https:\/\/www.python.org\/\" rel=\"nofollow\">Python 3<\/a> and <a href=\"https:\/\/github.com\/mattgreen\/watchexec\"><code>watchexec<\/code><\/a> for it to work. If you can't stand the extra dependencies and background process, consider the setting <code>(check-on-save find-when-checking)<\/code> instead, which is just as fast but won't catch package modifications unless they are made within Emacs via the <code>save-buffer<\/code> command.<\/p><p>Even with lazy modification detection, as described above, <code>straight.el<\/code> is not quite as fast as <code>package.el<\/code> (by a few percentage points). There are some planned changes which will make <code>straight.el<\/code> just as fast as <code>package.el<\/code>, if not faster. See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/9\">#9<\/a>.<\/p><h3><a id=\"user-content-how-do-i-update-melpa-et-al\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-update-melpa-et-al\"><\/a>How do I update MELPA et al.?<\/h3><p>Using <a href=\"#version-control-operations\"><code>M-x straight-pull-package<\/code><\/a>, like for any other package. <a href=\"#updating-recipe-repositories\">Read more.<\/a><\/p><h3><a id=\"user-content-the-wrong-version-of-my-package-was-loaded\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-wrong-version-of-my-package-was-loaded\"><\/a>The wrong version of my package was loaded<\/h3><p>To explain this problem, let us consider a concrete example. In <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/355\">this issue<\/a>, a user found that the code<\/p><pre>(straight-use-package 'company-lsp)\n(straight-use-package 'eglot)\n<\/pre><p>sometimes resulted in runtime errors because an old version of Flymake was being used.<\/p><p>The root problem here is that you want the most recent version of Flymake to be installed by <code>straight.el<\/code>, but Emacs also ships an older version, and that older version is getting loaded instead.<\/p><p>The older version will be loaded if <code>(require 'flymake)<\/code> (or similar) is invoked before <code>straight.el<\/code> has made Flymake available (by means of <code>(straight-use-package 'flymake)<\/code> or similar). But why would <code>straight.el<\/code> not make Flymake available?<\/p><p>The only way that <code>straight.el<\/code> knows to make Flymake available is if either you manually invoke <code>straight-use-package<\/code> in your init-file, or if one of the packages that you request in your init-file declares Flymake as a dependency. Now, any package that uses Flymake ought to declare it as a dependency. Thus, there should be no way for a package to load the Emacs-provided version of Flymake. However, sometimes package authors overlook this problem (it does not always cause an error, and sometimes package authors do not test exhaustively enough).<\/p><p>In this case, the problem was that <code>company-lsp<\/code> declared a dependency on <code>lsp-mode<\/code>, and <code>lsp-mode<\/code> used Flymake without declaring a dependency on <code>flymake<\/code>. There are two ways to work around the problem:<\/p><ul><li>(Preferable) Fix <code>lsp-mode<\/code> to declare a dependency on <code>flymake<\/code>.<\/li>\n<li>(Workaround) Manually invoke <code>(straight-use-package 'flymake)<\/code> before <code>(straight-use-package 'company-lsp)<\/code>.<\/li>\n<\/ul><p>If you test this yourself, you might find it difficult to reproduce the problem. That is because there is only an issue when Flymake is actually loaded, and this doesn't necessarily happen when invoking <code>(straight-use-package 'company-lsp)<\/code> <em>unless<\/em> <code>straight.el<\/code> needs to rebuild the relevant packages (which includes byte-compilation, which sometimes means actually loading dependencies). Keep this in mind when testing.<\/p><p>This problem commonly occurs with Org, since (1) Org is popular, (2) Emacs ships an obsolete version of Org, (3) many users want to use the up-to-date version, and (4) Org breaks backwards compatibility frequently. To solve it, simply make sure that you invoke <code>(straight-use-package 'org)<\/code> or <code>(straight-use-package 'org-plus-contrib)<\/code> before running any code that could load Org, including installing any package that lists it as a dependency. See also the <a href=\"#integration-with-org\">integration with Org<\/a> section for more fun problems you can encounter with Org.<\/p><p>See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/236\">this issue<\/a> for discussion about ways of mitigating the bad UX of this situation.<\/p><h3><a id=\"user-content-i-get-could-not-read-usernamepassword-errors\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-get-could-not-read-usernamepassword-errors\"><\/a>I get \"could not read username\/password\" errors<\/h3><p>This is because <code>straight.el<\/code> is not currently able to detect when SSH or Git asks for your username and\/or password\/passphrase and then pipe that prompt through to the minibuffer (<a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/334\">#334<\/a>).<\/p><p>To work around the problem, set up <a href=\"https:\/\/git-scm.com\/docs\/git-credential-cache\" rel=\"nofollow\">git-credential-cache<\/a> if you use HTTPS, and <a href=\"https:\/\/www.ssh.com\/ssh\/agent\" rel=\"nofollow\">ssh-agent<\/a> if you use SSH. That way, you won't be prompted for your username\/password. When setting up ssh-agent, be careful to make sure that the relevant environment variables get set in Emacs. This might be tricky since starting Emacs from the desktop (rather than from the command line) sometimes results in it not inheriting any environment variables from your shell.<\/p><h3><a id=\"user-content-the-interactive-version-control-operations-are-confusing\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-interactive-version-control-operations-are-confusing\"><\/a>The interactive version-control operations are confusing<\/h3><p>This part of <code>straight.el<\/code> still needs some work; see <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/54\">#54<\/a> about the UX of pushing and pulling.<\/p><h3><a id=\"user-content-how-do-i-pin-package-versions-or-use-only-tagged-releases\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-pin-package-versions-or-use-only-tagged-releases\"><\/a>How do I pin package versions or use only tagged releases?<\/h3><p>This is a planned feature. In the meantime, contributors have proposed various workarounds. See <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/246\">#246<\/a> and <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/31\">#31<\/a>.<\/p><p><code>straight-x.el<\/code> now contains an experimental solution. In order to use it you will need to add similar snippets to your Emacs configuration.<\/p><p>First you need to add a new profile to <code>straight-profiles<\/code> which also needs to be the last profile in the list. This should be done before you bootstrap <code>straight.el<\/code>.<\/p><pre>;; Tell straight.el about the profiles we are going to be using.\n(setq straight-profiles\n      '((nil . \"default.el\")\n        ;; Packages which are pinned to a specific commit.\n        (pinned . \"pinned.el\")))\n<\/pre><p>After straight's install procedure you will need to add <code>straight-x.el<\/code> and load the required commands.<\/p><pre>(autoload #'straight-x-pull-all \"straight-x\")\n(autoload #'straight-x-freeze-versions \"straight-x\")\n<\/pre><p>A variable called <code>straight-x-pinned-packages<\/code> has been defined in <code>straight-x.el<\/code> and will contain your list of pinned packages.<\/p><p>From now on, you can pin a package to a specific commit like in the following example which will pin <code>org-mode<\/code> to the 9.2.3 release version:<\/p><pre>(let ((straight-current-profile 'pinned))\n  (straight-use-package 'org-plus-contrib)\n  (straight-use-package 'org)\n  ;; Pin org-mode version.\n  (add-to-list 'straight-x-pinned-packages\n               '(\"org\" . \"924308a150ab82014b69c46c04d1ab71e874a2e6\")))\n<\/pre><p>If you invoke <code>straight-x-freeze-versions<\/code> it will first write the default lockfile and then pinned lockfile which takes precedence over the default one if packages are thawed. <code>straight-x-pull-all<\/code> will first invoke <code>straight-pull-all<\/code> and then restore all pinned packages.<\/p><p>You might want to assign the following aliases for more convenience:<\/p><pre>(defalias 'straight-pull-all #'straight-x-pull-all)\n(defalias 'straight-freeze-versions #'straight-x-freeze-versions)\n<\/pre><p>Please keep in mind that this is only a temporary solution and experimental!<\/p><h3><a id=\"user-content-how-can-i-use-the-built-in-version-of-a-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-can-i-use-the-built-in-version-of-a-package\"><\/a>How can I use the built-in version of a package?<\/h3><p>To tell <code>straight.el<\/code> that you want to use the version of Org shipped with Emacs, rather than cloning the upstream repository:<\/p><pre>(straight-use-package '(org :type built-in))\n<\/pre><p><a href=\"#the-recipe-format\">Read more.<\/a><\/p><h2><a id=\"user-content-news\" class=\"anchor\" aria-hidden=\"true\" href=\"#news\"><\/a>News<\/h2><h3><a id=\"user-content-july-6-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#july-6-2019\"><\/a>July 6, 2019<\/h3><p>The default value of the user option <code>straight-emacsmirror-use-mirror<\/code> is now non-nil.<\/p><h3><a id=\"user-content-may-24-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#may-24-2019\"><\/a>May 24, 2019<\/h3><p>I have completely rewritten the transaction system. The practical impact of this is that you no longer have to care about it. Simply remove all references to the following functions and variables from your configuration, and everything should \"just work\":<\/p><ul><li><code>straight-transaction<\/code><\/li>\n<li><code>straight-begin-transaction<\/code><\/li>\n<li><code>straight-finalize-transaction<\/code><\/li>\n<li><code>straight-mark-transaction-as-init<\/code><\/li>\n<li><code>straight-treat-as-init<\/code><\/li>\n<\/ul><h3><a id=\"user-content-may-22-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#may-22-2019\"><\/a>May 22, 2019<\/h3><p>I now maintain a mirror of Emacsmirror. (Bear with me here.) The advantage of using my mirror is that cloning it is several orders of magnitude faster than cloning the official Emacsmirror. You can tell <code>straight.el<\/code> to do so by customizing the user option <code>straight-emacsmirror-use-mirror<\/code> to non-nil.<\/p><h3><a id=\"user-content-may-1-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#may-1-2019\"><\/a>May 1, 2019<\/h3><p><code>straight-thaw-versions<\/code> now fetches in a repository if a commit in your lockfile can not be found and normalizes the repository to the recipe's default branch. This should ensure that versions you have frozen can be quickly restored and that they can also be saved back to the version lock file. This addresses issues <a href=\"https:\/\/github.com\/raxod502\/straight.el\/issues\/58\">#58<\/a>, [#66], and [#294].<\/p><h3><a id=\"user-content-march-15-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#march-15-2019\"><\/a>March 15, 2019<\/h3><p><code>straight.el<\/code> <a href=\"#integration-with-org\">now installs a hack for Org by default<\/a>, so Org should work out of the box with no compile warnings. If you have code to work around the problem in your init-file, you can safely remove it. To avoid installing the hack, customize the variable <code>straight-fix-org<\/code>.<\/p><h3><a id=\"user-content-december-22-2018\" class=\"anchor\" aria-hidden=\"true\" href=\"#december-22-2018\"><\/a>December 22, 2018<\/h3><p>The default value of the user option <code>straight-recipes-gnu-elpa-use-mirror<\/code> is now non-nil.<\/p><h3><a id=\"user-content-september-12-2018\" class=\"anchor\" aria-hidden=\"true\" href=\"#september-12-2018\"><\/a>September 12, 2018<\/h3><p><code>straight.el<\/code> now supports specifying configuration for your fork of a package via the new <code>:fork<\/code> keyword. The previously supported <code>:upstream<\/code> keyword is now deprecated, but still works for backwards compatibility. To support this change, the user options <code>straight-vc-git-primary-remote<\/code> and <code>straight-vc-git-upstream-remote<\/code> are deprecated (but still work for backwards compatibility), as they have been superseded by the new user options <code>straight-vc-git-default-remote-name<\/code> and <code>straight-vc-git-default-fork-name<\/code>. Your usage should be updated.<\/p><h3><a id=\"user-content-july-19-2018\" class=\"anchor\" aria-hidden=\"true\" href=\"#july-19-2018\"><\/a>July 19, 2018<\/h3><p><code>straight.el<\/code> now automatically caches the recipes it looks up in recipe repositories. This should lead to a reduction in <code>straight.el<\/code>-related startup time of as much as 50% if you also use live modification detection, as disk IO and usage of external processes are reduced significantly.<\/p><p>No changes to user configuration are necessary; however, if you define a custom recipe repository (call it <code>NAME<\/code>) then caching is not enabled by default. To enable caching, define a <code>straight-recipes-NAME-version<\/code> function which returns a non-nil value indicating the current version of the logic in <code>straight-recipes-NAME-retrieve<\/code>. This version value needs to be changed each time you change the logic, so that the recipe lookup cache for that recipe repository may automatically be invalidated.<\/p><h3><a id=\"user-content-july-12-2018\" class=\"anchor\" aria-hidden=\"true\" href=\"#july-12-2018\"><\/a>July 12, 2018<\/h3><p>I now maintain a <a href=\"https:\/\/github.com\/emacs-straight\">full mirror of GNU ELPA on GitHub<\/a>. You can tell <code>straight.el<\/code> to use it by customizing the user option <code>straight-recipes-gnu-elpa-use-mirror<\/code>, and this will allow you to use packages such as AUCTeX correctly, which was previously impossible. Note that the user option must be customized <em>before<\/em> the <code>straight.el<\/code> <a href=\"#getting-started\">bootstrap<\/a>.<\/p>","created_at":"2019-12-27T15:16:29+0000","updated_at":"2019-12-27T15:16:31+0000","published_by":["raxod502"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":102,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/6559064?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1074,"title":"rejeep\/f.el","url":"https:\/\/github.com\/rejeep\/f.el","content":"<p>Much inspired by <a href=\"https:\/\/github.com\/magnars\">@magnars<\/a>'s excellent <a href=\"https:\/\/github.com\/magnars\/s.el\">s.el<\/a> and <a href=\"https:\/\/github.com\/magnars\/dash.el\">dash.el<\/a>, <a href=\"https:\/\/github.com\/rejeep\/f.el\">f.el<\/a> is a modern API for working with files and directories in Emacs.<\/p><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>It's available on <a href=\"https:\/\/melpa.org\/\" rel=\"nofollow\">Melpa<\/a>:<\/p><pre>M-x package-install f\n<\/pre><p>Or you can just dump <code>f.el<\/code> in your load path somewhere.<\/p><h2><a id=\"user-content-api\" class=\"anchor\" aria-hidden=\"true\" href=\"#api\"><\/a>API<\/h2><h3><a id=\"user-content-paths\" class=\"anchor\" aria-hidden=\"true\" href=\"#paths\"><\/a>Paths<\/h3><ul><li><a href=\"#f-join-rest-args\">f-join<\/a> <code>(&amp;rest args)<\/code><\/li>\n<li><a href=\"#f-split-path\">f-split<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-expand-path-optional-dir\">f-expand<\/a> <code>(path &amp;optional dir)<\/code><\/li>\n<li><a href=\"#f-filename-path\">f-filename<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-dirname-path\">f-dirname<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-common-parent-paths\">f-common-parent<\/a> <code>(paths)<\/code><\/li>\n<li><a href=\"#f-ext-path\">f-ext<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-no-ext-path\">f-no-ext<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-swap-ext-path-ext\">f-swap-ext<\/a> <code>(path ext)<\/code><\/li>\n<li><a href=\"#f-base-path\">f-base<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-relative-path-optional-dir\">f-relative<\/a> <code>(path &amp;optional dir)<\/code><\/li>\n<li><a href=\"#f-short-path\">f-short<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-long-path\">f-long<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-canonical-path\">f-canonical<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-slash-path\">f-slash<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-full-path\">f-full<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-uniquify-paths\">f-uniquify<\/a> <code>(paths)<\/code><\/li>\n<li><a href=\"#f-uniquify-alist-paths\">f-uniquify-alist<\/a> <code>(paths)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-io\" class=\"anchor\" aria-hidden=\"true\" href=\"#io\"><\/a>I\/O<\/h3><ul><li><a href=\"#f-read-bytes-path\">f-read-bytes<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-write-bytes-data-path\">f-write-bytes<\/a> <code>(data path)<\/code><\/li>\n<li><a href=\"#f-read-text-path-optional-coding\">f-read-text<\/a> <code>(path &amp;optional coding)<\/code><\/li>\n<li><a href=\"#f-write-text-text-coding-path\">f-write-text<\/a><code>(text coding path)<\/code><\/li>\n<li><a href=\"#f-append-text-text-coding-path\">f-append-text<\/a><code>(text coding path)<\/code><\/li>\n<li><a href=\"#f-append-data-path\">f-append-bytes<\/a><code>(text coding path)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-destructive\" class=\"anchor\" aria-hidden=\"true\" href=\"#destructive\"><\/a>Destructive<\/h3><ul><li><a href=\"#f-mkdir-rest-dirs\">f-mkdir<\/a> <code>(&amp;rest dirs)<\/code><\/li>\n<li><a href=\"#f-delete-path-optional-force\">f-delete<\/a> <code>(path &amp;optional force)<\/code><\/li>\n<li><a href=\"#f-symlink-source-path\">f-symlink<\/a> <code>(source path)<\/code><\/li>\n<li><a href=\"#f-move-from-to\">f-move<\/a> <code>(from to)<\/code><\/li>\n<li><a href=\"#f-copy-from-to\">f-copy<\/a> <code>(from to)<\/code><\/li>\n<li><a href=\"#f-copy-contents-from-to\">f-copy-contenst<\/a> <code>(from to)<\/code><\/li>\n<li><a href=\"#f-touch-path\">f-touch<\/a> <code>(path)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-predicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#predicates\"><\/a>Predicates<\/h3><ul><li><a href=\"#f-exists-path\">f-exists?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-directory-path\">f-directory?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-file-path\">f-file?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-symlink-path\">f-symlink?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-readable-path\">f-readable?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-writable-path\">f-writable?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-executable-path\">f-executable?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-absolute-path\">f-absolute?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-relative-path\">f-relative?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-root-path\">f-root?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-ext-path-ext\">f-ext?<\/a> <code>(path ext)<\/code><\/li>\n<li><a href=\"#f-same-path-a-path-b\">f-same?<\/a> <code>(path-a path-b)<\/code><\/li>\n<li><a href=\"#f-parent-of-path-a-path-b\">f-parent-of?<\/a> <code>(path-a path-b)<\/code><\/li>\n<li><a href=\"#f-child-of-path-a-path-b\">f-child-of?<\/a> <code>(path-a path-b)<\/code><\/li>\n<li><a href=\"#f-ancestor-of-path-a-path-b\">f-ancestor-of?<\/a> <code>(path-a path-b)<\/code><\/li>\n<li><a href=\"#f-descendant-of-path-a-path-b\">f-descendant-of?<\/a> <code>(path-a path-b)<\/code><\/li>\n<li><a href=\"#f-hidden-path\">f-hidden?<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-empty-path\">f-empty?<\/a> <code>(path)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-stats\" class=\"anchor\" aria-hidden=\"true\" href=\"#stats\"><\/a>Stats<\/h3><ul><li><a href=\"#f-size-path\">f-size<\/a> <code>(path)<\/code><\/li>\n<li><a href=\"#f-depth-path\">f-depth<\/a> <code>(path)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-misc\" class=\"anchor\" aria-hidden=\"true\" href=\"#misc\"><\/a>Misc<\/h3><ul><li><a href=\"#f-this-file-\">f-this-file<\/a> <code>()<\/code><\/li>\n<li><a href=\"#f-path-separator-\">f-path-separator<\/a> <code>()<\/code><\/li>\n<li><a href=\"#f-glob-pattern-optional-path\">f-glob<\/a> <code>(pattern &amp;optional path)<\/code><\/li>\n<li><a href=\"#f-entries-path-optional-fn-recursive\">f-entries<\/a> <code>(path &amp;optional fn recursive)<\/code><\/li>\n<li><a href=\"#f-directories-path-optional-fn-recursive\">f-directories<\/a> <code>(path &amp;optional fn recursive)<\/code><\/li>\n<li><a href=\"#f-files-path-optional-fn-recursive\">f-files<\/a> <code>(path &amp;optional fn recursive)<\/code><\/li>\n<li><a href=\"#f-root-\">f-root<\/a> <code>()<\/code><\/li>\n<li><a href=\"#f-traverse-upwards-fn-optional-path\">f-traverse-upwards<\/a> <code>(fn &amp;optional path)<\/code><\/li>\n<li><a href=\"#f-with-sandbox-path-or-paths-rest-body\">f-with-sandbox<\/a> <code>(path-or-paths &amp;rest body)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-documentation-and-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation-and-examples\"><\/a>Documentation and examples<\/h2><h3><a id=\"user-content-f-join-rest-args\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-join-rest-args\"><\/a>f-join <code>(&amp;rest args)<\/code><\/h3><p>Join ARGS to a single path.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-join \"path\") ;; =&gt; \"path\"\n(f-join \"path\" \"to\") ;; =&gt; \"path\/to\"\n(f-join \"\/\" \"path\" \"to\" \"heaven\") ;; =&gt; \"\/path\/to\/heaven\"<\/pre><\/div><h3><a id=\"user-content-f-split-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-split-path\"><\/a>f-split <code>(path)<\/code><\/h3><p>Split PATH and return list containing parts.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-split \"path\") ;; =&gt; '(\"path\")\n(f-split \"path\/to\") ;; =&gt; '(\"path\" \"to\")\n(f-split \"\/path\/to\/heaven\") ;; =&gt; '(\"\/\" \"path\" \"to\" \"heaven\")<\/pre><\/div><h3><a id=\"user-content-f-expand-path-optional-dir\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-expand-path-optional-dir\"><\/a>f-expand <code>(path &amp;optional dir)<\/code><\/h3><p>Expand PATH relative to DIR (or \u2018default-directory\u2019).<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-expand \"name\") ;; =&gt; \"\/default\/directory\/name\"\n(f-expand \"name\" \"other\/directory\") ;; =&gt; \"other\/directory\/name\"<\/pre><\/div><h3><a id=\"user-content-f-filename-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-filename-path\"><\/a>f-filename <code>(path)<\/code><\/h3><p>Return the name of PATH.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-filename \"path\/to\/file.ext\") ;; =&gt; \"file.ext\"\n(f-filename \"path\/to\/directory\") ;; =&gt; \"directory\"<\/pre><\/div><h3><a id=\"user-content-f-dirname-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-dirname-path\"><\/a>f-dirname <code>(path)<\/code><\/h3><p>Return the parent directory to PATH.<\/p><p>Alias: <code>f-parent<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-dirname \"path\/to\/file.ext\") ;; =&gt; \"path\/to\"\n(f-dirname \"path\/to\/directory\") ;; =&gt; \"path\/to\"\n(f-dirname \"\/\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-common-parent-paths\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-common-parent-paths\"><\/a>f-common-parent <code>(paths)<\/code><\/h3><p>Return the deepest common parent directory of PATHS.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-common-parent '(\"foo\/bar\/baz\" \"foo\/bar\/qux\" \"foo\/bar\/mux\")) ;; =&gt; \"foo\/bar\/\"\n(f-common-parent '(\"\/foo\/bar\/baz\" \"\/foo\/bar\/qux\" \"\/foo\/bax\/mux\")) ;; =&gt; \"\/foo\/\"\n(f-common-parent '(\"foo\/bar\/baz\" \"quack\/bar\/qux\" \"lack\/bar\/mux\")) ;; =&gt; \"\"<\/pre><\/div><h3><a id=\"user-content-f-ext-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-ext-path\"><\/a>f-ext <code>(path)<\/code><\/h3><p>Return the file extension of PATH.<\/p><p>The extension, in a file name, is the part that follows the last \u2019.\u2019, excluding version numbers and backup suffixes.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-ext \"path\/to\/file.ext\") ;; =&gt; \"ext\"\n(f-ext \"path\/to\/directory\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-no-ext-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-no-ext-path\"><\/a>f-no-ext <code>(path)<\/code><\/h3><p>Return everything but the file extension of PATH.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-no-ext \"path\/to\/file.ext\") ;; =&gt; \"path\/to\/file\"\n(f-no-ext \"path\/to\/directory\") ;; =&gt; \"path\/to\/directory\"<\/pre><\/div><h3><a id=\"user-content-f-swap-ext-path-ext\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-swap-ext-path-ext\"><\/a>f-swap-ext <code>(path ext)<\/code><\/h3><p>Return PATH but with EXT as the new extension. EXT must not be nil or empty.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-swap-ext \"path\/to\/file.ext\" \"org\") ;; =&gt; \"path\/to\/file.org\"\n(f-swap-ext \"path\/to\/file.ext\" \"\") ;; =&gt; error<\/pre><\/div><h3><a id=\"user-content-f-base-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-base-path\"><\/a>f-base <code>(path)<\/code><\/h3><p>Return the name of PATH, excluding the extension of file.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-base \"path\/to\/file.ext\") ;; =&gt; \"file\"\n(f-base \"path\/to\/directory\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-relative-path-optional-dir\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-relative-path-optional-dir\"><\/a>f-relative <code>(path &amp;optional dir)<\/code><\/h3><p>Return PATH relative to DIR.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-relative \"\/some\/path\/relative\/to\/my\/file.txt\" \"\/some\/path\/\") ;; =&gt; relative\/to\/my\/file.txt\n(f-relative \"\/default\/directory\/my\/file.txt\") ;; =&gt; my\/file.txt<\/pre><\/div><h3><a id=\"user-content-f-short-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-short-path\"><\/a>f-short <code>(path)<\/code><\/h3><p>Return abbrev of PATH. See \u2018abbreviate-file-name\u2019.<\/p><p>Alias: <code>f-abbrev<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-short \"\/Users\/foo\/Code\/bar\") ;; =&gt; ~\/Code\/bar\n(f-short \"\/path\/to\/Code\/bar\") ;; =&gt; \/path\/to\/Code\/bar<\/pre><\/div><h3><a id=\"user-content-f-long-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-long-path\"><\/a>f-long <code>(path)<\/code><\/h3><p>Return long version of PATH.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-long \"~\/Code\/bar\") ;; =&gt; \/Users\/foo\/Code\/bar\n(f-long \"\/path\/to\/Code\/bar\") ;; =&gt; \/path\/to\/Code\/bar<\/pre><\/div><h3><a id=\"user-content-f-canonical-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-canonical-path\"><\/a>f-canonical <code>(path)<\/code><\/h3><p>Return the canonical name of PATH.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-canonical \"\/path\/to\/real\/file\") ;; =&gt; \/path\/to\/real\/file\n(f-canonical \"\/link\/to\/file\") ;; =&gt; \/path\/to\/real\/file<\/pre><\/div><h3><a id=\"user-content-f-slash-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-slash-path\"><\/a>f-slash <code>(path)<\/code><\/h3><p>Append slash to PATH unless one already.<\/p><p>Some functions, such as \u2018call-process\u2019 requires there to be an ending slash.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-slash \"\/path\/to\/file\") ;; =&gt; \/path\/to\/file\n(f-slash \"\/path\/to\/dir\") ;; =&gt; \/path\/to\/dir\/\n(f-slash \"\/path\/to\/dir\/\") ;; =&gt; \/path\/to\/dir\/<\/pre><\/div><h3><a id=\"user-content-f-full-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-full-path\"><\/a>f-full <code>(path)<\/code><\/h3><p>Return absolute path to PATH, with ending slash.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-full \"~\/path\/to\/file\") ;; =&gt; \/home\/path\/to\/file\n(f-full \"~\/path\/to\/dir\") ;; =&gt; \/home\/path\/to\/dir\/\n(f-full \"~\/path\/to\/dir\/\") ;; =&gt; \/home\/path\/to\/dir\/<\/pre><\/div><h3><a id=\"user-content-f-uniquify-paths\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-uniquify-paths\"><\/a>f-uniquify <code>(paths)<\/code><\/h3><p>Return unique suffixes of FILES.<\/p><p>This function expects no duplicate paths.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-uniquify '(\"\/foo\/bar\" \"\/foo\/baz\" \"\/foo\/quux\")) ;; =&gt; '(\"bar\" \"baz\" \"quux\")\n(f-uniquify '(\"\/foo\/bar\" \"\/www\/bar\" \"\/foo\/quux\")) ;; =&gt; '(\"foo\/bar\" \"www\/bar\" \"quux\")\n(f-uniquify '(\"\/foo\/bar\" \"\/www\/bar\" \"\/www\/bar\/quux\")) ;; =&gt; '(\"foo\/bar\" \"www\/bar\" \"quux\")\n(f-uniquify '(\"\/foo\/bar\" \"\/foo\/baz\" \"\/home\/www\/bar\" \"\/home\/www\/baz\" \"\/var\/foo\" \"\/opt\/foo\/www\/baz\")) ;; =&gt; '(\"foo\/bar\" \"www\/bar\" \"foo\/baz\" \"home\/www\/baz\" \"foo\/www\/baz\" \"foo\")<\/pre><\/div><h3><a id=\"user-content-f-uniquify-alist-paths\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-uniquify-alist-paths\"><\/a>f-uniquify-alist <code>(paths)<\/code><\/h3><p>Return alist mapping FILES to unique suffixes of FILES.<\/p><p>This function expects no duplicate paths.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-uniquify-alist '(\"\/foo\/bar\" \"\/foo\/baz\" \"\/foo\/quux\")) ;; =&gt; '((\"\/foo\/bar\" . \"bar\") (\"\/foo\/baz\" . \"baz\") (\"\/foo\/quux\" . \"quux\"))\n(f-uniquify-alist '(\"\/foo\/bar\" \"\/www\/bar\" \"\/foo\/quux\")) ;; =&gt; '((\"\/foo\/bar\" . \"foo\/bar\") (\"\/www\/bar\" . \"www\/bar\") (\"\/foo\/quux\" . \"quux\"))\n(f-uniquify-alist '(\"\/foo\/bar\" \"\/www\/bar\" \"\/www\/bar\/quux\")) ;; =&gt; '((\"\/foo\/bar\" . \"foo\/bar\") (\"\/www\/bar\" . \"www\/bar\") (\"\/www\/bar\/quux\" . \"quux\"))\n(f-uniquify-alist '(\"\/foo\/bar\" \"\/foo\/baz\" \"\/home\/www\/bar\" \"\/home\/www\/baz\" \"\/var\/foo\" \"\/opt\/foo\/www\/baz\")) ;; =&gt; '((\"\/foo\/bar\" . \"foo\/bar\") (\"\/home\/www\/bar\" . \"www\/bar\") (\"\/foo\/baz\" . \"foo\/baz\") (\"\/home\/www\/baz\" . \"home\/www\/baz\") (\"\/opt\/foo\/www\/baz\" . \"foo\/www\/baz\") (\"\/var\/foo\" . \"foo\"))<\/pre><\/div><h3><a id=\"user-content-f-read-bytes-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-read-bytes-path\"><\/a>f-read-bytes <code>(path)<\/code><\/h3><p>Read binary data from PATH.<\/p><p>Return the binary data as unibyte string.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-read-bytes \"path\/to\/binary\/data\")<\/pre><\/div><h3><a id=\"user-content-f-write-bytes-data-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-write-bytes-data-path\"><\/a>f-write-bytes <code>(data path)<\/code><\/h3><p>Write binary DATA to PATH.<\/p><p>DATA is a unibyte string. PATH is a file name to write to.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-write-bytes (unibyte-string 72 101 108 108 111 32 119 111 114 108 100) \"path\/to\/binary\/data\")<\/pre><\/div><h3><a id=\"user-content-f-read-text-path-optional-coding\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-read-text-path-optional-coding\"><\/a>f-read-text <code>(path &amp;optional coding)<\/code><\/h3><p>Read text with PATH, using CODING.<\/p><p>CODING defaults to \u2018utf-8\u2019.<\/p><p>Return the decoded text as multibyte string.<\/p><p>Alias: <code>f-read<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-read-text \"path\/to\/file.txt\" 'utf-8)\n(f-read \"path\/to\/file.txt\" 'utf-8)<\/pre><\/div><h3><a id=\"user-content-f-write-text-text-coding-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-write-text-text-coding-path\"><\/a>f-write-text <code>(text coding path)<\/code><\/h3><p>Write TEXT with CODING to PATH.<\/p><p>TEXT is a multibyte string. CODING is a coding system to encode TEXT with. PATH is a file name to write to.<\/p><p>Alias: <code>f-write<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-write-text \"Hello world\" 'utf-8 \"path\/to\/file.txt\")\n(f-write \"Hello world\" 'utf-8 \"path\/to\/file.txt\")<\/pre><\/div><h3><a id=\"user-content-f-append-text-text-coding-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-append-text-text-coding-path\"><\/a>f-append-text <code>(text coding path)<\/code><\/h3><p>Append TEXT with CODING to PATH.<\/p><p>If PATH does not exist, it is created.<\/p><p>Alias: <code>f-append<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-append-text \"Hello world\" 'utf-8 \"path\/to\/file.txt\")\n(f-append \"Hello world\" 'utf-8 \"path\/to\/file.txt\")<\/pre><\/div><h3><a id=\"user-content-f-append-bytes-data-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-append-bytes-data-path\"><\/a>f-append-bytes <code>(data path)<\/code><\/h3><p>Append binary DATA to PATH.<\/p><p>If PATH does not exist, it is created.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-append-bytes \"path\/to\/file\" (unibyte-string 72 101 108 108 111 32 119 111 114 108 100))<\/pre><\/div><h3><a id=\"user-content-f-mkdir-rest-dirs\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-mkdir-rest-dirs\"><\/a>f-mkdir <code>(&amp;rest dirs)<\/code><\/h3><p>Create directories DIRS.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-mkdir \"dir\") ;; =&gt; \/default\/directory\/dir\n(f-mkdir \"other\" \"dir\") ;; =&gt; \/default\/directory\/other\/dir<\/pre><\/div><h3><a id=\"user-content-f-delete-path-optional-force\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-delete-path-optional-force\"><\/a>f-delete <code>(path &amp;optional force)<\/code><\/h3><p>Delete PATH, which can be file or directory.<\/p><p>If FORCE is t, a directory will be deleted recursively.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-delete \"dir\")\n(f-delete \"other\/dir\" t)\n(f-delete \"path\/to\/file.txt\")<\/pre><\/div><h3><a id=\"user-content-f-symlink-source-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-symlink-source-path\"><\/a>f-symlink <code>(source path)<\/code><\/h3><p>Create a symlink to SOURCE from PATH.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-symlink \"path\/to\/source\" \"path\/to\/link\")<\/pre><\/div><h3><a id=\"user-content-f-move-from-to\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-move-from-to\"><\/a>f-move <code>(from to)<\/code><\/h3><p>Move or rename FROM to TO.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-move \"path\/to\/file.txt\" \"new-file.txt\")\n(f-move \"path\/to\/file.txt\" \"other\/path\")<\/pre><\/div><h3><a id=\"user-content-f-copy-from-to\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-copy-from-to\"><\/a>f-copy <code>(from to)<\/code><\/h3><p>Copy file or directory FROM to TO.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-copy \"path\/to\/file.txt\" \"new-file.txt\")\n(f-copy \"path\/to\/dir\" \"other\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-copy-contents-from-to\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-copy-contents-from-to\"><\/a>f-copy-contents <code>(from to)<\/code><\/h3><p>Copy contents in directory FROM, to directory TO.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-copy-contents \"path\/to\/dir\" \"path\/to\/other\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-touch-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-touch-path\"><\/a>f-touch <code>(path)<\/code><\/h3><p>Update PATH last modification date or create if it does not exist.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-touch \"path\/to\/existing\/file.txt\")\n(f-touch \"path\/to\/non\/existing\/file.txt\")<\/pre><\/div><h3><a id=\"user-content-f-exists-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-exists-path\"><\/a>f-exists? <code>(path)<\/code><\/h3><p>Return t if PATH exists, false otherwise.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-exists? \"path\/to\/file.txt\")\n(f-exists? \"path\/to\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-directory-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-directory-path\"><\/a>f-directory? <code>(path)<\/code><\/h3><p>Return t if PATH is directory, false otherwise.<\/p><p>Aliases: <code>f-directory-p f-dir? f-dir-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-directory? \"path\/to\/file.txt\") ;; =&gt; nil\n(f-directory? \"path\/to\/dir\") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-f-file-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-file-path\"><\/a>f-file? <code>(path)<\/code><\/h3><p>Return t if PATH is file, false otherwise.<\/p><p>Alias: <code>f-file-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-file? \"path\/to\/file.txt\") ;; =&gt; t\n(f-file? \"path\/to\/dir\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-symlink-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-symlink-path\"><\/a>f-symlink? <code>(path)<\/code><\/h3><p>Return t if PATH is symlink, false otherwise.<\/p><p>Alias: <code>f-symlink-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-symlink? \"path\/to\/file.txt\") ;; =&gt; nil\n(f-symlink? \"path\/to\/dir\") ;; =&gt; nil\n(f-symlink? \"path\/to\/link\") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-f-readable-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-readable-path\"><\/a>f-readable? <code>(path)<\/code><\/h3><p>Return t if PATH is readable, false otherwise.<\/p><p>Alias: <code>f-readable-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-readable? \"path\/to\/file.txt\")\n(f-readable? \"path\/to\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-writable-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-writable-path\"><\/a>f-writable? <code>(path)<\/code><\/h3><p>Return t if PATH is writable, false otherwise.<\/p><p>Alias: <code>f-writable-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-writable? \"path\/to\/file.txt\")\n(f-writable? \"path\/to\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-executable-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-executable-path\"><\/a>f-executable? <code>(path)<\/code><\/h3><p>Return t if PATH is executable, false otherwise.<\/p><p>Alias: <code>f-executable-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-executable? \"path\/to\/file.txt\")\n(f-executable? \"path\/to\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-absolute-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-absolute-path\"><\/a>f-absolute? <code>(path)<\/code><\/h3><p>Return t if PATH is absolute, false otherwise.<\/p><p>Alias: <code>f-absolute-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-absolute? \"path\/to\/dir\") ;; =&gt; nil\n(f-absolute? \"\/full\/path\/to\/dir\") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-f-relative-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-relative-path\"><\/a>f-relative? <code>(path)<\/code><\/h3><p>Return t if PATH is relative, false otherwise.<\/p><p>Alias: <code>f-relative-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-relative? \"path\/to\/dir\") ;; =&gt; t\n(f-relative? \"\/full\/path\/to\/dir\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-root-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-root-path\"><\/a>f-root? <code>(path)<\/code><\/h3><p>Return t if PATH is root directory, false otherwise.<\/p><p>Alias: <code>f-root-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-root? \"\/\") ;; =&gt; t\n(f-root? \"\/not\/root\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-ext-path-ext\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-ext-path-ext\"><\/a>f-ext? <code>(path ext)<\/code><\/h3><p>Return t if extension of PATH is EXT, false otherwise.<\/p><p>If EXT is nil or omitted, return t if PATH has any extension, false otherwise.<\/p><p>The extension, in a file name, is the part that follows the last \u2019.\u2019, excluding version numbers and backup suffixes.<\/p><p>Alias: <code>f-ext-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-ext? \"path\/to\/file.el\" \"el\") ;; =&gt; t\n(f-ext? \"path\/to\/file.el\" \"txt\") ;; =&gt; nil\n(f-ext? \"path\/to\/file.el\") ;; =&gt; t\n(f-ext? \"path\/to\/file\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-same-path-a-path-b\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-same-path-a-path-b\"><\/a>f-same? <code>(path-a path-b)<\/code><\/h3><p>Return t if PATH-A and PATH-B are references to same file.<\/p><p>Aliases: <code>f-same-p f-equal? f-equal-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-same? \"foo.txt\" \"foo.txt\") ;; =&gt; t\n(f-same? \"\/path\/to\/foo.txt\" \"\/path\/to\/bar.txt\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-parent-of-path-a-path-b\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-parent-of-path-a-path-b\"><\/a>f-parent-of? <code>(path-a path-b)<\/code><\/h3><p>Return t if PATH-A is parent of PATH-B.<\/p><p>Alias: <code>f-parent-of-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-parent-of? \"\/path\/to\" \"\/path\/to\/dir\") ;; =&gt; t\n(f-parent-of? \"\/path\/to\/dir\" \"\/path\/to\") ;; =&gt; nil\n(f-parent-of? \"\/path\/to\" \"\/path\/to\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-child-of-path-a-path-b\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-child-of-path-a-path-b\"><\/a>f-child-of? <code>(path-a path-b)<\/code><\/h3><p>Return t if PATH-A is child of PATH-B.<\/p><p>Alias: <code>f-child-of-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-child-of? \"\/path\/to\" \"\/path\/to\/dir\") ;; =&gt; nil\n(f-child-of? \"\/path\/to\/dir\" \"\/path\/to\") ;; =&gt; t\n(f-child-of? \"\/path\/to\" \"\/path\/to\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-ancestor-of-path-a-path-b\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-ancestor-of-path-a-path-b\"><\/a>f-ancestor-of? <code>(path-a path-b)<\/code><\/h3><p>Return t if PATH-A is ancestor of PATH-B.<\/p><p>Alias: <code>f-ancestor-of-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-ancestor-of? \"\/path\/to\" \"\/path\/to\/dir\") ;; =&gt; t\n(f-ancestor-of? \"\/path\" \"\/path\/to\/dir\") ;; =&gt; t\n(f-ancestor-of? \"\/path\/to\/dir\" \"\/path\/to\") ;; =&gt; nil\n(f-ancestor-of? \"\/path\/to\" \"\/path\/to\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-descendant-of-path-a-path-b\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-descendant-of-path-a-path-b\"><\/a>f-descendant-of? <code>(path-a path-b)<\/code><\/h3><p>Return t if PATH-A is desendant of PATH-B.<\/p><p>Alias: <code>f-descendant-of-p<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-descendant-of? \"\/path\/to\/dir\" \"\/path\/to\") ;; =&gt; t\n(f-descendant-of? \"\/path\/to\/dir\" \"\/path\") ;; =&gt; t\n(f-descendant-of? \"\/path\/to\" \"\/path\/to\/dir\") ;; =&gt; nil\n(f-descendant-of? \"\/path\/to\" \"\/path\/to\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-f-hidden-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-hidden-path\"><\/a>f-hidden? <code>(path)<\/code><\/h3><p>Return t if PATH is hidden, nil otherwise.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-hidden? \"\/path\/to\/foo\") ;; =&gt; nil\n(f-hidden? \"\/path\/to\/.foo\") ;; =&gt; t<\/pre><\/div><p>Alias: <code>f-hidden-p<\/code><\/p><h3><a id=\"user-content-f-empty-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-empty-path\"><\/a>f-empty? <code>(path)<\/code><\/h3><p>If PATH is a file, return t if the file in PATH is empty, nil otherwise. If PATH is directory, return t if directory has no files, nil otherwise.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-empty? \"\/path\/to\/empty-file\") ;; =&gt; t\n(f-empty? \"\/path\/to\/file-with-contents\") ;; =&gt; nil\n(f-empty? \"\/path\/to\/empty-dir\/\") ;; =&gt; t\n(f-empty? \"\/path\/to\/dir-with-contents\/\") ;; =&gt; nil<\/pre><\/div><p>Alias: <code>f-empty-p<\/code><\/p><h3><a id=\"user-content-f-size-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-size-path\"><\/a>f-size <code>(path)<\/code><\/h3><p>Return size of PATH.<\/p><p>If PATH is a file, return size of that file. If PATH is directory, return sum of all files in PATH.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-size \"path\/to\/file.txt\")\n(f-size \"path\/to\/dir\")<\/pre><\/div><h3><a id=\"user-content-f-depth-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-depth-path\"><\/a>f-depth <code>(path)<\/code><\/h3><p>Return the depth of PATH.<\/p><p>At first, PATH is expanded with \u2018f-expand\u2019. Then the full path is used to detect the depth. \u2019\/\u2019 will be zero depth, \u2019\/usr\u2019 will be one depth. And so on.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-depth \"\/\") ;; 0\n(f-depth \"\/var\/\") ;; 1\n(f-depth \"\/usr\/local\/bin\") ;; 3<\/pre><\/div><h3><a id=\"user-content-f-this-file-\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-this-file-\"><\/a>f-this-file <code>()<\/code><\/h3><p>Return path to this file.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-this-file) ;; =&gt; \/path\/to\/this\/file<\/pre><\/div><h3><a id=\"user-content-f-path-separator-\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-path-separator-\"><\/a>f-path-separator <code>()<\/code><\/h3><p>Return path separator.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-path-separator) ;; =&gt; \/<\/pre><\/div><h3><a id=\"user-content-f-glob-pattern-optional-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-glob-pattern-optional-path\"><\/a>f-glob <code>(pattern &amp;optional path)<\/code><\/h3><p>Find PATTERN in PATH.<\/p><p>See: <code>file-expand-wildcards<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-glob \"path\/to\/*.el\")\n(f-glob \"*.el\" \"path\/to\")<\/pre><\/div><h3><a id=\"user-content-f-entries-path-optional-fn-recursive\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-entries-path-optional-fn-recursive\"><\/a>f-entries <code>(path &amp;optional fn recursive)<\/code><\/h3><p>Find all files and directories in PATH.<\/p><p>FN - called for each found file and directory. If FN returns a thruthy value, file or directory will be included. RECURSIVE - Search for files and directories recursive.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-entries \"path\/to\/dir\")\n(f-entries \"path\/to\/dir\" (lambda (file) (s-matches? \"test\" file)))\n(f-entries \"path\/to\/dir\" nil t)\n(f--entries \"path\/to\/dir\" (s-matches? \"test\" it))<\/pre><\/div><h3><a id=\"user-content-f-directories-path-optional-fn-recursive\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-directories-path-optional-fn-recursive\"><\/a>f-directories <code>(path &amp;optional fn recursive)<\/code><\/h3><p>Find all directories in PATH. See \u2018f-entries\u2018.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-directories \"path\/to\/dir\")\n(f-directories \"path\/to\/dir\" (lambda (dir) (equal (f-filename dir) \"test\")))\n(f-directories \"path\/to\/dir\" nil t)\n(f--directories \"path\/to\/dir\" (equal (f-filename it) \"test\"))<\/pre><\/div><h3><a id=\"user-content-f-files-path-optional-fn-recursive\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-files-path-optional-fn-recursive\"><\/a>f-files <code>(path &amp;optional fn recursive)<\/code><\/h3><p>Find all files in PATH. See \u2018f-entries\u2018.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-files \"path\/to\/dir\")\n(f-files \"path\/to\/dir\" (lambda (file) (equal (f-ext file) \"el\")))\n(f-files \"path\/to\/dir\" nil t)\n(f--files \"path\/to\/dir\" (equal (f-ext it) \"el\"))<\/pre><\/div><h3><a id=\"user-content-f-root-\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-root-\"><\/a>f-root <code>()<\/code><\/h3><p>Return absolute root.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-root) ;; =&gt; \"\/\"<\/pre><\/div><h3><a id=\"user-content-f-traverse-upwards-fn-optional-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-traverse-upwards-fn-optional-path\"><\/a>f-traverse-upwards <code>(fn &amp;optional path)<\/code><\/h3><p>Traverse up as long as FN return nil, starting at PATH.<\/p><p>If FN returns a non-nil value, the path sent as argument to FN is returned. If no function callback return a non-nil value, nil is returned.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-traverse-upwards\n (lambda (path)\n   (f-exists? (f-expand \".git\" path)))\n start-path)\n(f--traverse-upwards (f-exists? (f-expand \".git\" it)) start-path) ;; same as above<\/pre><\/div><h3><a id=\"user-content-f-with-sandbox-path-or-paths-rest-body\" class=\"anchor\" aria-hidden=\"true\" href=\"#f-with-sandbox-path-or-paths-rest-body\"><\/a>f-with-sandbox <code>(path-or-paths &amp;rest body)<\/code><\/h3><p>Only allow PATH-OR-PATHS and descendants to be modified in BODY.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(f-with-sandbox foo-path\n  (f-touch (f-expand \"foo\" foo-path)))\n(f-with-sandbox (list foo-path bar-path)\n  (f-touch (f-expand \"foo\" foo-path))\n  (f-touch (f-expand \"bar\" bar-path)))\n(f-with-sandbox foo-path\n  (f-touch (f-expand \"bar\" bar-path))) ;; \"Destructive operation outside sandbox\"<\/pre><\/div><h2><a id=\"user-content-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#example\"><\/a>Example<\/h2><p>Here's an example of a function that finds the Git project root.<\/p><h3><a id=\"user-content-using-standard-emacs-builtin-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-standard-emacs-builtin-functions\"><\/a>Using standard Emacs builtin functions<\/h3><div class=\"highlight highlight-source-lisp\"><pre>(defun find-git-root (&amp;optional dir)\n  (unless dir (setq dir (expand-file-name (file-name-directory (buffer-file-name)))))\n  (let ((parent (expand-file-name \"..\" dir)))\n    (unless (equal parent dir)\n      (if (file-exists-p (expand-file-name \".git\" dir))\n          dir\n        (find-git-root parent)))))<\/pre><\/div><h3><a id=\"user-content-using-fel\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-fel\"><\/a>Using <code>f.el<\/code><\/h3><div class=\"highlight highlight-source-lisp\"><pre>(defun find-git-root (&amp;optional dir)\n  (interactive)\n  (unless dir (setq dir (f-dirname (buffer-file-name))))\n  (let ((parent (f-parent dir)))\n    (unless (f-root? parent)\n      (if (f-exists? (f-expand \".git\" dir))\n          dir\n        (find-git-root parent)))))<\/pre><\/div><p>Now, try writing it even simpler yourself. Hint, check out <code>f-traverse-upwards<\/code>.<\/p><h2><a id=\"user-content-contribution\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribution\"><\/a>Contribution<\/h2><p>Be sure to!<\/p><p>Install <a href=\"https:\/\/github.com\/rejeep\/cask.el\">Cask<\/a> if you haven't already.<\/p><p>Run the unit tests with:<\/p><pre>$ make test\n<\/pre><p>Do not change <code>README.md<\/code> directly. If you want to change the README or if you change any function comments, update the README with:<\/p><pre>$ make docs\n<\/pre>","created_at":"2019-12-25T03:46:38+0000","updated_at":"2019-12-25T03:46:38+0000","published_by":["rejeep"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/20277?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1072,"title":"magnars\/dash.el","url":"https:\/\/github.com\/magnars\/dash.el","content":"<p>A modern list api for Emacs. No 'cl required.<\/p><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>It's available on <a href=\"https:\/\/melpa.org\/\" rel=\"nofollow\">Melpa<\/a>:<\/p><pre>M-x package-install dash\n<\/pre><p>Or you can just dump <code>dash.el<\/code> in your load path somewhere.<\/p><p>If you want the function combinators, then also:<\/p><pre>M-x package-install dash-functional\n<\/pre><h2><a id=\"user-content-using-in-a-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-in-a-package\"><\/a>Using in a package<\/h2><p>Add this to the big comment block at the top:<\/p><pre>;; Package-Requires: ((dash \"2.16.0\"))\n<\/pre><p>To get function combinators:<\/p><pre>;; Package-Requires: ((dash \"2.16.0\") (dash-functional \"1.2.0\") (emacs \"24\"))\n<\/pre><h2><a id=\"user-content-upcoming-breaking-change\" class=\"anchor\" aria-hidden=\"true\" href=\"#upcoming-breaking-change\"><\/a>Upcoming breaking change!<\/h2><ul><li>For backward compatibility reasons <code>-zip<\/code> return a cons-cell instead of a list with two elements when called on two lists. This is a clunky API, and in an upcoming 3.0 release of Dash it will always return a list. If you rely on the cons-cell return value, use <code>-zip-pair<\/code> instead.<\/li>\n<\/ul><h2><a id=\"user-content-syntax-highlighting-of-dash-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#syntax-highlighting-of-dash-functions\"><\/a>Syntax highlighting of dash functions<\/h2><p>Font lock of dash functions in emacs lisp buffers is now optional. Include this in your emacs settings to get syntax highlighting:<\/p><pre>(eval-after-load 'dash '(dash-enable-font-lock))\n<\/pre><h2><a id=\"user-content-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#functions\"><\/a>Functions<\/h2><p>All functions and constructs in the library are prefixed with a dash (-).<\/p><p>There are also anaphoric versions of functions where that makes sense, prefixed with two dashes instead of one.<\/p><p>While <code>-map<\/code> takes a function to map over the list, you can also use the anaphoric form with double dashes - which will then be executed with <code>it<\/code> exposed as the list item. Here's an example:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map (lambda (n) (* n n)) '(1 2 3 4)) ;; normal version\n(--map (* it it) '(1 2 3 4)) ;; anaphoric version<\/pre><\/div><p>of course the original can also be written like<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun square (n) (* n n))\n(-map 'square '(1 2 3 4))<\/pre><\/div><p>which demonstrates the usefulness of both versions.<\/p><h3><a id=\"user-content-maps\" class=\"anchor\" aria-hidden=\"true\" href=\"#maps\"><\/a>Maps<\/h3><p>Functions in this category take a transforming function, which is then applied sequentially to each or selected elements of the input list. The results are collected in order and returned as new list.<\/p><ul><li><a href=\"#-map-fn-list\">-map<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-map-when-pred-rep-list\">-map-when<\/a> <code>(pred rep list)<\/code><\/li>\n<li><a href=\"#-map-first-pred-rep-list\">-map-first<\/a> <code>(pred rep list)<\/code><\/li>\n<li><a href=\"#-map-last-pred-rep-list\">-map-last<\/a> <code>(pred rep list)<\/code><\/li>\n<li><a href=\"#-map-indexed-fn-list\">-map-indexed<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-annotate-fn-list\">-annotate<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-splice-pred-fun-list\">-splice<\/a> <code>(pred fun list)<\/code><\/li>\n<li><a href=\"#-splice-list-pred-new-list-list\">-splice-list<\/a> <code>(pred new-list list)<\/code><\/li>\n<li><a href=\"#-mapcat-fn-list\">-mapcat<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-copy-arg\">-copy<\/a> <code>(arg)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-sublist-selection\" class=\"anchor\" aria-hidden=\"true\" href=\"#sublist-selection\"><\/a>Sublist selection<\/h3><p>Functions returning a sublist of the original list.<\/p><ul><li><a href=\"#-filter-pred-list\">-filter<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-remove-pred-list\">-remove<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-remove-first-pred-list\">-remove-first<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-remove-last-pred-list\">-remove-last<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-remove-item-item-list\">-remove-item<\/a> <code>(item list)<\/code><\/li>\n<li><a href=\"#-non-nil-list\">-non-nil<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-slice-list-from-optional-to-step\">-slice<\/a> <code>(list from &amp;optional to step)<\/code><\/li>\n<li><a href=\"#-take-n-list\">-take<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-take-last-n-list\">-take-last<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-drop-n-list\">-drop<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-drop-last-n-list\">-drop-last<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-take-while-pred-list\">-take-while<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-drop-while-pred-list\">-drop-while<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-select-by-indices-indices-list\">-select-by-indices<\/a> <code>(indices list)<\/code><\/li>\n<li><a href=\"#-select-columns-columns-table\">-select-columns<\/a> <code>(columns table)<\/code><\/li>\n<li><a href=\"#-select-column-column-table\">-select-column<\/a> <code>(column table)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-list-to-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#list-to-list\"><\/a>List to list<\/h3><p>Functions returning a modified copy of the input list.<\/p><ul><li><a href=\"#-keep-fn-list\">-keep<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-concat-rest-lists\">-concat<\/a> <code>(&amp;rest lists)<\/code><\/li>\n<li><a href=\"#-flatten-l\">-flatten<\/a> <code>(l)<\/code><\/li>\n<li><a href=\"#-flatten-n-num-list\">-flatten-n<\/a> <code>(num list)<\/code><\/li>\n<li><a href=\"#-replace-old-new-list\">-replace<\/a> <code>(old new list)<\/code><\/li>\n<li><a href=\"#-replace-first-old-new-list\">-replace-first<\/a> <code>(old new list)<\/code><\/li>\n<li><a href=\"#-replace-last-old-new-list\">-replace-last<\/a> <code>(old new list)<\/code><\/li>\n<li><a href=\"#-insert-at-n-x-list\">-insert-at<\/a> <code>(n x list)<\/code><\/li>\n<li><a href=\"#-replace-at-n-x-list\">-replace-at<\/a> <code>(n x list)<\/code><\/li>\n<li><a href=\"#-update-at-n-func-list\">-update-at<\/a> <code>(n func list)<\/code><\/li>\n<li><a href=\"#-remove-at-n-list\">-remove-at<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-remove-at-indices-indices-list\">-remove-at-indices<\/a> <code>(indices list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-reductions\" class=\"anchor\" aria-hidden=\"true\" href=\"#reductions\"><\/a>Reductions<\/h3><p>Functions reducing lists into single value.<\/p><ul><li><a href=\"#-reduce-from-fn-initial-value-list\">-reduce-from<\/a> <code>(fn initial-value list)<\/code><\/li>\n<li><a href=\"#-reduce-r-from-fn-initial-value-list\">-reduce-r-from<\/a> <code>(fn initial-value list)<\/code><\/li>\n<li><a href=\"#-reduce-fn-list\">-reduce<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-reduce-r-fn-list\">-reduce-r<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-reductions-from-fn-init-list\">-reductions-from<\/a> <code>(fn init list)<\/code><\/li>\n<li><a href=\"#-reductions-r-from-fn-init-list\">-reductions-r-from<\/a> <code>(fn init list)<\/code><\/li>\n<li><a href=\"#-reductions-fn-list\">-reductions<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-reductions-r-fn-list\">-reductions-r<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-count-pred-list\">-count<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-sum-list\">-sum<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-running-sum-list\">-running-sum<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-product-list\">-product<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-running-product-list\">-running-product<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-inits-list\">-inits<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-tails-list\">-tails<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-common-prefix-rest-lists\">-common-prefix<\/a> <code>(&amp;rest lists)<\/code><\/li>\n<li><a href=\"#-common-suffix-rest-lists\">-common-suffix<\/a> <code>(&amp;rest lists)<\/code><\/li>\n<li><a href=\"#-min-list\">-min<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-min-by-comparator-list\">-min-by<\/a> <code>(comparator list)<\/code><\/li>\n<li><a href=\"#-max-list\">-max<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-max-by-comparator-list\">-max-by<\/a> <code>(comparator list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-unfolding\" class=\"anchor\" aria-hidden=\"true\" href=\"#unfolding\"><\/a>Unfolding<\/h3><p>Operations dual to reductions, building lists from seed value rather than consuming a list to produce a single value.<\/p><ul><li><a href=\"#-iterate-fun-init-n\">-iterate<\/a> <code>(fun init n)<\/code><\/li>\n<li><a href=\"#-unfold-fun-seed\">-unfold<\/a> <code>(fun seed)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-predicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#predicates\"><\/a>Predicates<\/h3><ul><li><a href=\"#-any-pred-list\">-any?<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-all-pred-list\">-all?<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-none-pred-list\">-none?<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-only-some-pred-list\">-only-some?<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-contains-list-element\">-contains?<\/a> <code>(list element)<\/code><\/li>\n<li><a href=\"#-same-items-list-list2\">-same-items?<\/a> <code>(list list2)<\/code><\/li>\n<li><a href=\"#-is-prefix-prefix-list\">-is-prefix?<\/a> <code>(prefix list)<\/code><\/li>\n<li><a href=\"#-is-suffix-suffix-list\">-is-suffix?<\/a> <code>(suffix list)<\/code><\/li>\n<li><a href=\"#-is-infix-infix-list\">-is-infix?<\/a> <code>(infix list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-partitioning\" class=\"anchor\" aria-hidden=\"true\" href=\"#partitioning\"><\/a>Partitioning<\/h3><p>Functions partitioning the input list into a list of lists.<\/p><ul><li><a href=\"#-split-at-n-list\">-split-at<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-split-with-pred-list\">-split-with<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-split-on-item-list\">-split-on<\/a> <code>(item list)<\/code><\/li>\n<li><a href=\"#-split-when-fn-list\">-split-when<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-separate-pred-list\">-separate<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-partition-n-list\">-partition<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-partition-all-n-list\">-partition-all<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-partition-in-steps-n-step-list\">-partition-in-steps<\/a> <code>(n step list)<\/code><\/li>\n<li><a href=\"#-partition-all-in-steps-n-step-list\">-partition-all-in-steps<\/a> <code>(n step list)<\/code><\/li>\n<li><a href=\"#-partition-by-fn-list\">-partition-by<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-partition-by-header-fn-list\">-partition-by-header<\/a> <code>(fn list)<\/code><\/li>\n<li><a href=\"#-partition-after-pred-pred-list\">-partition-after-pred<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-partition-before-pred-pred-list\">-partition-before-pred<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-partition-before-item-item-list\">-partition-before-item<\/a> <code>(item list)<\/code><\/li>\n<li><a href=\"#-partition-after-item-item-list\">-partition-after-item<\/a> <code>(item list)<\/code><\/li>\n<li><a href=\"#-group-by-fn-list\">-group-by<\/a> <code>(fn list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-indexing\" class=\"anchor\" aria-hidden=\"true\" href=\"#indexing\"><\/a>Indexing<\/h3><p>Return indices of elements based on predicates, sort elements by indices etc.<\/p><ul><li><a href=\"#-elem-index-elem-list\">-elem-index<\/a> <code>(elem list)<\/code><\/li>\n<li><a href=\"#-elem-indices-elem-list\">-elem-indices<\/a> <code>(elem list)<\/code><\/li>\n<li><a href=\"#-find-index-pred-list\">-find-index<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-find-last-index-pred-list\">-find-last-index<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-find-indices-pred-list\">-find-indices<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-grade-up-comparator-list\">-grade-up<\/a> <code>(comparator list)<\/code><\/li>\n<li><a href=\"#-grade-down-comparator-list\">-grade-down<\/a> <code>(comparator list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-set-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#set-operations\"><\/a>Set operations<\/h3><p>Operations pretending lists are sets.<\/p><ul><li><a href=\"#-union-list-list2\">-union<\/a> <code>(list list2)<\/code><\/li>\n<li><a href=\"#-difference-list-list2\">-difference<\/a> <code>(list list2)<\/code><\/li>\n<li><a href=\"#-intersection-list-list2\">-intersection<\/a> <code>(list list2)<\/code><\/li>\n<li><a href=\"#-powerset-list\">-powerset<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-permutations-list\">-permutations<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-distinct-list\">-distinct<\/a> <code>(list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-other-list-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-list-operations\"><\/a>Other list operations<\/h3><p>Other list functions not fit to be classified elsewhere.<\/p><ul><li><a href=\"#-rotate-n-list\">-rotate<\/a> <code>(n list)<\/code><\/li>\n<li><a href=\"#-repeat-n-x\">-repeat<\/a> <code>(n x)<\/code><\/li>\n<li><a href=\"#-cons-rest-args\">-cons*<\/a> <code>(&amp;rest args)<\/code><\/li>\n<li><a href=\"#-snoc-list-elem-rest-elements\">-snoc<\/a> <code>(list elem &amp;rest elements)<\/code><\/li>\n<li><a href=\"#-interpose-sep-list\">-interpose<\/a> <code>(sep list)<\/code><\/li>\n<li><a href=\"#-interleave-rest-lists\">-interleave<\/a> <code>(&amp;rest lists)<\/code><\/li>\n<li><a href=\"#-zip-with-fn-list1-list2\">-zip-with<\/a> <code>(fn list1 list2)<\/code><\/li>\n<li><a href=\"#-zip-rest-lists\">-zip<\/a> <code>(&amp;rest lists)<\/code><\/li>\n<li><a href=\"#-zip-fill-fill-value-rest-lists\">-zip-fill<\/a> <code>(fill-value &amp;rest lists)<\/code><\/li>\n<li><a href=\"#-unzip-lists\">-unzip<\/a> <code>(lists)<\/code><\/li>\n<li><a href=\"#-cycle-list\">-cycle<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-pad-fill-value-rest-lists\">-pad<\/a> <code>(fill-value &amp;rest lists)<\/code><\/li>\n<li><a href=\"#-table-fn-rest-lists\">-table<\/a> <code>(fn &amp;rest lists)<\/code><\/li>\n<li><a href=\"#-table-flat-fn-rest-lists\">-table-flat<\/a> <code>(fn &amp;rest lists)<\/code><\/li>\n<li><a href=\"#-first-pred-list\">-first<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-some-pred-list\">-some<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-last-pred-list\">-last<\/a> <code>(pred list)<\/code><\/li>\n<li><a href=\"#-first-item-list\">-first-item<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-second-item-arg1\">-second-item<\/a> <code>(arg1)<\/code><\/li>\n<li><a href=\"#-third-item-arg1\">-third-item<\/a> <code>(arg1)<\/code><\/li>\n<li><a href=\"#-fourth-item-list\">-fourth-item<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-fifth-item-list\">-fifth-item<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-last-item-list\">-last-item<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-butlast-list\">-butlast<\/a> <code>(list)<\/code><\/li>\n<li><a href=\"#-sort-comparator-list\">-sort<\/a> <code>(comparator list)<\/code><\/li>\n<li><a href=\"#-list-rest-args\">-list<\/a> <code>(&amp;rest args)<\/code><\/li>\n<li><a href=\"#-fix-fn-list\">-fix<\/a> <code>(fn list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-tree-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#tree-operations\"><\/a>Tree operations<\/h3><p>Functions pretending lists are trees.<\/p><ul><li><a href=\"#-tree-seq-branch-children-tree\">-tree-seq<\/a> <code>(branch children tree)<\/code><\/li>\n<li><a href=\"#-tree-map-fn-tree\">-tree-map<\/a> <code>(fn tree)<\/code><\/li>\n<li><a href=\"#-tree-map-nodes-pred-fun-tree\">-tree-map-nodes<\/a> <code>(pred fun tree)<\/code><\/li>\n<li><a href=\"#-tree-reduce-fn-tree\">-tree-reduce<\/a> <code>(fn tree)<\/code><\/li>\n<li><a href=\"#-tree-reduce-from-fn-init-value-tree\">-tree-reduce-from<\/a> <code>(fn init-value tree)<\/code><\/li>\n<li><a href=\"#-tree-mapreduce-fn-folder-tree\">-tree-mapreduce<\/a> <code>(fn folder tree)<\/code><\/li>\n<li><a href=\"#-tree-mapreduce-from-fn-folder-init-value-tree\">-tree-mapreduce-from<\/a> <code>(fn folder init-value tree)<\/code><\/li>\n<li><a href=\"#-clone-list\">-clone<\/a> <code>(list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-threading-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#threading-macros\"><\/a>Threading macros<\/h3><ul><li><a href=\"#--x-optional-form-rest-more\">-&gt;<\/a> <code>(x &amp;optional form &amp;rest more)<\/code><\/li>\n<li><a href=\"#--x-optional-form-rest-more\">-&gt;&gt;<\/a> <code>(x &amp;optional form &amp;rest more)<\/code><\/li>\n<li><a href=\"#---x-rest-forms\">--&gt;<\/a> <code>(x &amp;rest forms)<\/code><\/li>\n<li><a href=\"#-as--value-variable-rest-forms\">-as-&gt;<\/a> <code>(value variable &amp;rest forms)<\/code><\/li>\n<li><a href=\"#-some--x-optional-form-rest-more\">-some-&gt;<\/a> <code>(x &amp;optional form &amp;rest more)<\/code><\/li>\n<li><a href=\"#-some--x-optional-form-rest-more\">-some-&gt;&gt;<\/a> <code>(x &amp;optional form &amp;rest more)<\/code><\/li>\n<li><a href=\"#-some---x-optional-form-rest-more\">-some--&gt;<\/a> <code>(x &amp;optional form &amp;rest more)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-binding\" class=\"anchor\" aria-hidden=\"true\" href=\"#binding\"><\/a>Binding<\/h3><p>Convenient versions of <code>let<\/code> and <code>let*<\/code> constructs combined with flow control.<\/p><ul><li><a href=\"#-when-let-var-val-rest-body\">-when-let<\/a> <code>(var-val &amp;rest body)<\/code><\/li>\n<li><a href=\"#-when-let-vars-vals-rest-body\">-when-let*<\/a> <code>(vars-vals &amp;rest body)<\/code><\/li>\n<li><a href=\"#-if-let-var-val-then-rest-else\">-if-let<\/a> <code>(var-val then &amp;rest else)<\/code><\/li>\n<li><a href=\"#-if-let-vars-vals-then-rest-else\">-if-let*<\/a> <code>(vars-vals then &amp;rest else)<\/code><\/li>\n<li><a href=\"#-let-varlist-rest-body\">-let<\/a> <code>(varlist &amp;rest body)<\/code><\/li>\n<li><a href=\"#-let-varlist-rest-body\">-let*<\/a> <code>(varlist &amp;rest body)<\/code><\/li>\n<li><a href=\"#-lambda-match-form-rest-body\">-lambda<\/a> <code>(match-form &amp;rest body)<\/code><\/li>\n<li><a href=\"#-setq-rest-forms\">-setq<\/a> <code>(&amp;rest forms)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-side-effects\" class=\"anchor\" aria-hidden=\"true\" href=\"#side-effects\"><\/a>Side-effects<\/h3><p>Functions iterating over lists for side-effect only.<\/p><ul><li><a href=\"#-each-list-fn\">-each<\/a> <code>(list fn)<\/code><\/li>\n<li><a href=\"#-each-while-list-pred-fn\">-each-while<\/a> <code>(list pred fn)<\/code><\/li>\n<li><a href=\"#-each-indexed-list-fn\">-each-indexed<\/a> <code>(list fn)<\/code><\/li>\n<li><a href=\"#-each-r-list-fn\">-each-r<\/a> <code>(list fn)<\/code><\/li>\n<li><a href=\"#-each-r-while-list-pred-fn\">-each-r-while<\/a> <code>(list pred fn)<\/code><\/li>\n<li><a href=\"#-dotimes-num-fn\">-dotimes<\/a> <code>(num fn)<\/code><\/li>\n<li><a href=\"#-doto-eval-initial-value-rest-forms\">-doto<\/a> <code>(eval-initial-value &amp;rest forms)<\/code><\/li>\n<li><a href=\"#--doto-eval-initial-value-rest-forms\">--doto<\/a> <code>(eval-initial-value &amp;rest forms)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-destructive-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#destructive-operations\"><\/a>Destructive operations<\/h3><ul><li><a href=\"#cons-car-cdr\">!cons<\/a> <code>(car cdr)<\/code><\/li>\n<li><a href=\"#cdr-list\">!cdr<\/a> <code>(list)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-function-combinators\" class=\"anchor\" aria-hidden=\"true\" href=\"#function-combinators\"><\/a>Function combinators<\/h3><p>These combinators require Emacs 24 for its lexical scope. So they are offered in a separate package: <code>dash-functional<\/code>.<\/p><ul><li><a href=\"#-partial-fn-rest-args\">-partial<\/a> <code>(fn &amp;rest args)<\/code><\/li>\n<li><a href=\"#-rpartial-fn-rest-args\">-rpartial<\/a> <code>(fn &amp;rest args)<\/code><\/li>\n<li><a href=\"#-juxt-rest-fns\">-juxt<\/a> <code>(&amp;rest fns)<\/code><\/li>\n<li><a href=\"#-compose-rest-fns\">-compose<\/a> <code>(&amp;rest fns)<\/code><\/li>\n<li><a href=\"#-applify-fn\">-applify<\/a> <code>(fn)<\/code><\/li>\n<li><a href=\"#-on-operator-transformer\">-on<\/a> <code>(operator transformer)<\/code><\/li>\n<li><a href=\"#-flip-func\">-flip<\/a> <code>(func)<\/code><\/li>\n<li><a href=\"#-const-c\">-const<\/a> <code>(c)<\/code><\/li>\n<li><a href=\"#-cut-rest-params\">-cut<\/a> <code>(&amp;rest params)<\/code><\/li>\n<li><a href=\"#-not-pred\">-not<\/a> <code>(pred)<\/code><\/li>\n<li><a href=\"#-orfn-rest-preds\">-orfn<\/a> <code>(&amp;rest preds)<\/code><\/li>\n<li><a href=\"#-andfn-rest-preds\">-andfn<\/a> <code>(&amp;rest preds)<\/code><\/li>\n<li><a href=\"#-iteratefn-fn-n\">-iteratefn<\/a> <code>(fn n)<\/code><\/li>\n<li><a href=\"#-fixfn-fn-optional-equal-test-halt-test\">-fixfn<\/a> <code>(fn &amp;optional equal-test halt-test)<\/code><\/li>\n<li><a href=\"#-prodfn-rest-fns\">-prodfn<\/a> <code>(&amp;rest fns)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-maps-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#maps-1\"><\/a>Maps<\/h2><p>Functions in this category take a transforming function, which is then applied sequentially to each or selected elements of the input list. The results are collected in order and returned as new list.<\/p><h4><a id=\"user-content--map-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-map-fn-list\"><\/a>-map <code>(fn list)<\/code><\/h4><p>Return a new list consisting of the result of applying <code>fn<\/code> to the items in <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map (lambda (num) (* num num)) '(1 2 3 4)) ;; =&gt; '(1 4 9 16)\n(-map 'square '(1 2 3 4)) ;; =&gt; '(1 4 9 16)\n(--map (* it it) '(1 2 3 4)) ;; =&gt; '(1 4 9 16)<\/pre><\/div><h4><a id=\"user-content--map-when-pred-rep-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-map-when-pred-rep-list\"><\/a>-map-when <code>(pred rep list)<\/code><\/h4><p>Return a new list where the elements in <code>list<\/code> that do not match the <code>pred<\/code> function are unchanged, and where the elements in <code>list<\/code> that do match the <code>pred<\/code> function are mapped through the <code>rep<\/code> function.<\/p><p>Alias: <code>-replace-where<\/code><\/p><p>See also: <a href=\"#-update-at-n-func-list\"><code>-update-at<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map-when 'even? 'square '(1 2 3 4)) ;; =&gt; '(1 4 3 16)\n(--map-when (&gt; it 2) (* it it) '(1 2 3 4)) ;; =&gt; '(1 2 9 16)\n(--map-when (= it 2) 17 '(1 2 3 4)) ;; =&gt; '(1 17 3 4)<\/pre><\/div><h4><a id=\"user-content--map-first-pred-rep-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-map-first-pred-rep-list\"><\/a>-map-first <code>(pred rep list)<\/code><\/h4><p>Replace first item in <code>list<\/code> satisfying <code>pred<\/code> with result of <code>rep<\/code> called on this item.<\/p><p>See also: <a href=\"#-map-when-pred-rep-list\"><code>-map-when<\/code><\/a>, <a href=\"#-replace-first-old-new-list\"><code>-replace-first<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map-first 'even? 'square '(1 2 3 4)) ;; =&gt; '(1 4 3 4)\n(--map-first (&gt; it 2) (* it it) '(1 2 3 4)) ;; =&gt; '(1 2 9 4)\n(--map-first (= it 2) 17 '(1 2 3 2)) ;; =&gt; '(1 17 3 2)<\/pre><\/div><h4><a id=\"user-content--map-last-pred-rep-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-map-last-pred-rep-list\"><\/a>-map-last <code>(pred rep list)<\/code><\/h4><p>Replace last item in <code>list<\/code> satisfying <code>pred<\/code> with result of <code>rep<\/code> called on this item.<\/p><p>See also: <a href=\"#-map-when-pred-rep-list\"><code>-map-when<\/code><\/a>, <a href=\"#-replace-last-old-new-list\"><code>-replace-last<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map-last 'even? 'square '(1 2 3 4)) ;; =&gt; '(1 2 3 16)\n(--map-last (&gt; it 2) (* it it) '(1 2 3 4)) ;; =&gt; '(1 2 3 16)\n(--map-last (= it 2) 17 '(1 2 3 2)) ;; =&gt; '(1 2 3 17)<\/pre><\/div><h4><a id=\"user-content--map-indexed-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-map-indexed-fn-list\"><\/a>-map-indexed <code>(fn list)<\/code><\/h4><p>Return a new list consisting of the result of (<code>fn<\/code> index item) for each item in <code>list<\/code>.<\/p><p>In the anaphoric form <code>--map-indexed<\/code>, the index is exposed as symbol <code>it-index<\/code>.<\/p><p>See also: <a href=\"#-each-indexed-list-fn\"><code>-each-indexed<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map-indexed (lambda (index item) (- item index)) '(1 2 3 4)) ;; =&gt; '(1 1 1 1)\n(--map-indexed (- it it-index) '(1 2 3 4)) ;; =&gt; '(1 1 1 1)<\/pre><\/div><h4><a id=\"user-content--annotate-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-annotate-fn-list\"><\/a>-annotate <code>(fn list)<\/code><\/h4><p>Return a list of cons cells where each cell is <code>fn<\/code> applied to each element of <code>list<\/code> paired with the unmodified element of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-annotate '1+ '(1 2 3)) ;; =&gt; '((2 . 1) (3 . 2) (4 . 3))\n(-annotate 'length '((\"h\" \"e\" \"l\" \"l\" \"o\") (\"hello\" \"world\"))) ;; =&gt; '((5 \"h\" \"e\" \"l\" \"l\" \"o\") (2 \"hello\" \"world\"))\n(--annotate (&lt; 1 it) '(0 1 2 3)) ;; =&gt; '((nil . 0) (nil . 1) (t . 2) (t . 3))<\/pre><\/div><h4><a id=\"user-content--splice-pred-fun-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-splice-pred-fun-list\"><\/a>-splice <code>(pred fun list)<\/code><\/h4><p>Splice lists generated by <code>fun<\/code> in place of elements matching <code>pred<\/code> in <code>list<\/code>.<\/p><p><code>fun<\/code> takes the element matching <code>pred<\/code> as input.<\/p><p>This function can be used as replacement for <code>,@<\/code> in case you need to splice several lists at marked positions (for example with keywords).<\/p><p>See also: <a href=\"#-splice-list-pred-new-list-list\"><code>-splice-list<\/code><\/a>, <a href=\"#-insert-at-n-x-list\"><code>-insert-at<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-splice 'even? (lambda (x) (list x x)) '(1 2 3 4)) ;; =&gt; '(1 2 2 3 4 4)\n(--splice 't (list it it) '(1 2 3 4)) ;; =&gt; '(1 1 2 2 3 3 4 4)\n(--splice (equal it :magic) '((list of) (magical) (code)) '((foo) (bar) :magic (baz))) ;; =&gt; '((foo) (bar) (list of) (magical) (code) (baz))<\/pre><\/div><h4><a id=\"user-content--splice-list-pred-new-list-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-splice-list-pred-new-list-list\"><\/a>-splice-list <code>(pred new-list list)<\/code><\/h4><p>Splice <code>new-list<\/code> in place of elements matching <code>pred<\/code> in <code>list<\/code>.<\/p><p>See also: <a href=\"#-splice-pred-fun-list\"><code>-splice<\/code><\/a>, <a href=\"#-insert-at-n-x-list\"><code>-insert-at<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-splice-list 'keywordp '(a b c) '(1 :foo 2)) ;; =&gt; '(1 a b c 2)\n(-splice-list 'keywordp nil '(1 :foo 2)) ;; =&gt; '(1 2)\n(--splice-list (keywordp it) '(a b c) '(1 :foo 2)) ;; =&gt; '(1 a b c 2)<\/pre><\/div><h4><a id=\"user-content--mapcat-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-mapcat-fn-list\"><\/a>-mapcat <code>(fn list)<\/code><\/h4><p>Return the concatenation of the result of mapping <code>fn<\/code> over <code>list<\/code>. Thus function <code>fn<\/code> should return a list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-mapcat 'list '(1 2 3)) ;; =&gt; '(1 2 3)\n(-mapcat (lambda (item) (list 0 item)) '(1 2 3)) ;; =&gt; '(0 1 0 2 0 3)\n(--mapcat (list 0 it) '(1 2 3)) ;; =&gt; '(0 1 0 2 0 3)<\/pre><\/div><h4><a id=\"user-content--copy-arg\" class=\"anchor\" aria-hidden=\"true\" href=\"#-copy-arg\"><\/a>-copy <code>(arg)<\/code><\/h4><p>Create a shallow copy of <code>list<\/code>.<\/p><p>(fn <code>list<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-copy '(1 2 3)) ;; =&gt; '(1 2 3)\n(let ((a '(1 2 3))) (eq a (-copy a))) ;; =&gt; nil<\/pre><\/div><h2><a id=\"user-content-sublist-selection-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#sublist-selection-1\"><\/a>Sublist selection<\/h2><p>Functions returning a sublist of the original list.<\/p><h4><a id=\"user-content--filter-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-filter-pred-list\"><\/a>-filter <code>(pred list)<\/code><\/h4><p>Return a new list of the items in <code>list<\/code> for which <code>pred<\/code> returns a non-nil value.<\/p><p>Alias: <code>-select<\/code><\/p><p>See also: <a href=\"#-keep-fn-list\"><code>-keep<\/code><\/a>, <a href=\"#-remove-pred-list\"><code>-remove<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-filter (lambda (num) (= 0 (% num 2))) '(1 2 3 4)) ;; =&gt; '(2 4)\n(-filter 'even? '(1 2 3 4)) ;; =&gt; '(2 4)\n(--filter (= 0 (% it 2)) '(1 2 3 4)) ;; =&gt; '(2 4)<\/pre><\/div><h4><a id=\"user-content--remove-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-remove-pred-list\"><\/a>-remove <code>(pred list)<\/code><\/h4><p>Return a new list of the items in <code>list<\/code> for which <code>pred<\/code> returns nil.<\/p><p>Alias: <code>-reject<\/code><\/p><p>See also: <a href=\"#-filter-pred-list\"><code>-filter<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-remove (lambda (num) (= 0 (% num 2))) '(1 2 3 4)) ;; =&gt; '(1 3)\n(-remove 'even? '(1 2 3 4)) ;; =&gt; '(1 3)\n(--remove (= 0 (% it 2)) '(1 2 3 4)) ;; =&gt; '(1 3)<\/pre><\/div><h4><a id=\"user-content--remove-first-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-remove-first-pred-list\"><\/a>-remove-first <code>(pred list)<\/code><\/h4><p>Return a new list with the first item matching <code>pred<\/code> removed.<\/p><p>Alias: <code>-reject-first<\/code><\/p><p>See also: <a href=\"#-remove-pred-list\"><code>-remove<\/code><\/a>, <a href=\"#-map-first-pred-rep-list\"><code>-map-first<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-remove-first 'even? '(1 3 5 4 7 8 10)) ;; =&gt; '(1 3 5 7 8 10)\n(-remove-first 'stringp '(1 2 \"first\" \"second\" \"third\")) ;; =&gt; '(1 2 \"second\" \"third\")\n(--remove-first (&gt; it 3) '(1 2 3 4 5 6 7 8 9 10)) ;; =&gt; '(1 2 3 5 6 7 8 9 10)<\/pre><\/div><h4><a id=\"user-content--remove-last-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-remove-last-pred-list\"><\/a>-remove-last <code>(pred list)<\/code><\/h4><p>Return a new list with the last item matching <code>pred<\/code> removed.<\/p><p>Alias: <code>-reject-last<\/code><\/p><p>See also: <a href=\"#-remove-pred-list\"><code>-remove<\/code><\/a>, <a href=\"#-map-last-pred-rep-list\"><code>-map-last<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-remove-last 'even? '(1 3 5 4 7 8 10 11)) ;; =&gt; '(1 3 5 4 7 8 11)\n(-remove-last 'stringp '(1 2 \"last\" \"second\" \"third\")) ;; =&gt; '(1 2 \"last\" \"second\")\n(--remove-last (&gt; it 3) '(1 2 3 4 5 6 7 8 9 10)) ;; =&gt; '(1 2 3 4 5 6 7 8 9)<\/pre><\/div><h4><a id=\"user-content--remove-item-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-remove-item-item-list\"><\/a>-remove-item <code>(item list)<\/code><\/h4><p>Remove all occurrences of <code>item<\/code> from <code>list<\/code>.<\/p><p>Comparison is done with <code>equal<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-remove-item 3 '(1 2 3 2 3 4 5 3)) ;; =&gt; '(1 2 2 4 5)\n(-remove-item 'foo '(foo bar baz foo)) ;; =&gt; '(bar baz)\n(-remove-item \"bob\" '(\"alice\" \"bob\" \"eve\" \"bob\" \"dave\")) ;; =&gt; '(\"alice\" \"eve\" \"dave\")<\/pre><\/div><h4><a id=\"user-content--non-nil-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-non-nil-list\"><\/a>-non-nil <code>(list)<\/code><\/h4><p>Return all non-nil elements of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-non-nil '(1 nil 2 nil nil 3 4 nil 5 nil)) ;; =&gt; '(1 2 3 4 5)<\/pre><\/div><h4><a id=\"user-content--slice-list-from-optional-to-step\" class=\"anchor\" aria-hidden=\"true\" href=\"#-slice-list-from-optional-to-step\"><\/a>-slice <code>(list from &amp;optional to step)<\/code><\/h4><p>Return copy of <code>list<\/code>, starting from index <code>from<\/code> to index <code>to<\/code>.<\/p><p><code>from<\/code> or <code>to<\/code> may be negative. These values are then interpreted modulo the length of the list.<\/p><p>If <code>step<\/code> is a number, only each STEPth item in the resulting section is returned. Defaults to 1.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-slice '(1 2 3 4 5) 1) ;; =&gt; '(2 3 4 5)\n(-slice '(1 2 3 4 5) 0 3) ;; =&gt; '(1 2 3)\n(-slice '(1 2 3 4 5 6 7 8 9) 1 -1 2) ;; =&gt; '(2 4 6 8)<\/pre><\/div><h4><a id=\"user-content--take-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-take-n-list\"><\/a>-take <code>(n list)<\/code><\/h4><p>Return a new list of the first <code>n<\/code> items in <code>list<\/code>, or all items if there are fewer than <code>n<\/code>.<\/p><p>See also: <a href=\"#-take-last-n-list\"><code>-take-last<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-take 3 '(1 2 3 4 5)) ;; =&gt; '(1 2 3)\n(-take 17 '(1 2 3 4 5)) ;; =&gt; '(1 2 3 4 5)<\/pre><\/div><h4><a id=\"user-content--take-last-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-take-last-n-list\"><\/a>-take-last <code>(n list)<\/code><\/h4><p>Return the last <code>n<\/code> items of <code>list<\/code> in order.<\/p><p>See also: <a href=\"#-take-n-list\"><code>-take<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-take-last 3 '(1 2 3 4 5)) ;; =&gt; '(3 4 5)\n(-take-last 17 '(1 2 3 4 5)) ;; =&gt; '(1 2 3 4 5)\n(-take-last 1 '(1 2 3 4 5)) ;; =&gt; '(5)<\/pre><\/div><h4><a id=\"user-content--drop-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-drop-n-list\"><\/a>-drop <code>(n list)<\/code><\/h4><p>Return the tail of <code>list<\/code> without the first <code>n<\/code> items.<\/p><p>See also: <a href=\"#-drop-last-n-list\"><code>-drop-last<\/code><\/a><\/p><p>(fn <code>n<\/code> <code>list<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-drop 3 '(1 2 3 4 5)) ;; =&gt; '(4 5)\n(-drop 17 '(1 2 3 4 5)) ;; =&gt; '()<\/pre><\/div><h4><a id=\"user-content--drop-last-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-drop-last-n-list\"><\/a>-drop-last <code>(n list)<\/code><\/h4><p>Remove the last <code>n<\/code> items of <code>list<\/code> and return a copy.<\/p><p>See also: <a href=\"#-drop-n-list\"><code>-drop<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-drop-last 3 '(1 2 3 4 5)) ;; =&gt; '(1 2)\n(-drop-last 17 '(1 2 3 4 5)) ;; =&gt; '()<\/pre><\/div><h4><a id=\"user-content--take-while-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-take-while-pred-list\"><\/a>-take-while <code>(pred list)<\/code><\/h4><p>Return a new list of successive items from <code>list<\/code> while (<code>pred<\/code> item) returns a non-nil value.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-take-while 'even? '(1 2 3 4)) ;; =&gt; '()\n(-take-while 'even? '(2 4 5 6)) ;; =&gt; '(2 4)\n(--take-while (&lt; it 4) '(1 2 3 4 3 2 1)) ;; =&gt; '(1 2 3)<\/pre><\/div><h4><a id=\"user-content--drop-while-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-drop-while-pred-list\"><\/a>-drop-while <code>(pred list)<\/code><\/h4><p>Return the tail of <code>list<\/code> starting from the first item for which (<code>pred<\/code> item) returns nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-drop-while 'even? '(1 2 3 4)) ;; =&gt; '(1 2 3 4)\n(-drop-while 'even? '(2 4 5 6)) ;; =&gt; '(5 6)\n(--drop-while (&lt; it 4) '(1 2 3 4 3 2 1)) ;; =&gt; '(4 3 2 1)<\/pre><\/div><h4><a id=\"user-content--select-by-indices-indices-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-select-by-indices-indices-list\"><\/a>-select-by-indices <code>(indices list)<\/code><\/h4><p>Return a list whose elements are elements from <code>list<\/code> selected as <code>(nth i list)<\/code> for all i from <code>indices<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-select-by-indices '(4 10 2 3 6) '(\"v\" \"e\" \"l\" \"o\" \"c\" \"i\" \"r\" \"a\" \"p\" \"t\" \"o\" \"r\")) ;; =&gt; '(\"c\" \"o\" \"l\" \"o\" \"r\")\n(-select-by-indices '(2 1 0) '(\"a\" \"b\" \"c\")) ;; =&gt; '(\"c\" \"b\" \"a\")\n(-select-by-indices '(0 1 2 0 1 3 3 1) '(\"f\" \"a\" \"r\" \"l\")) ;; =&gt; '(\"f\" \"a\" \"r\" \"f\" \"a\" \"l\" \"l\" \"a\")<\/pre><\/div><h4><a id=\"user-content--select-columns-columns-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#-select-columns-columns-table\"><\/a>-select-columns <code>(columns table)<\/code><\/h4><p>Select <code>columns<\/code> from <code>table<\/code>.<\/p><p><code>table<\/code> is a list of lists where each element represents one row. It is assumed each row has the same length.<\/p><p>Each row is transformed such that only the specified <code>columns<\/code> are selected.<\/p><p>See also: <a href=\"#-select-column-column-table\"><code>-select-column<\/code><\/a>, <a href=\"#-select-by-indices-indices-list\"><code>-select-by-indices<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-select-columns '(0 2) '((1 2 3) (a b c) (:a :b :c))) ;; =&gt; '((1 3) (a c) (:a :c))\n(-select-columns '(1) '((1 2 3) (a b c) (:a :b :c))) ;; =&gt; '((2) (b) (:b))\n(-select-columns nil '((1 2 3) (a b c) (:a :b :c))) ;; =&gt; '(nil nil nil)<\/pre><\/div><h4><a id=\"user-content--select-column-column-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#-select-column-column-table\"><\/a>-select-column <code>(column table)<\/code><\/h4><p>Select <code>column<\/code> from <code>table<\/code>.<\/p><p><code>table<\/code> is a list of lists where each element represents one row. It is assumed each row has the same length.<\/p><p>The single selected column is returned as a list.<\/p><p>See also: <a href=\"#-select-columns-columns-table\"><code>-select-columns<\/code><\/a>, <a href=\"#-select-by-indices-indices-list\"><code>-select-by-indices<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-select-column 1 '((1 2 3) (a b c) (:a :b :c))) ;; =&gt; '(2 b :b)<\/pre><\/div><h2><a id=\"user-content-list-to-list-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#list-to-list-1\"><\/a>List to list<\/h2><p>Functions returning a modified copy of the input list.<\/p><h4><a id=\"user-content--keep-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-keep-fn-list\"><\/a>-keep <code>(fn list)<\/code><\/h4><p>Return a new list of the non-nil results of applying <code>fn<\/code> to the items in <code>list<\/code>.<\/p><p>If you want to select the original items satisfying a predicate use <a href=\"#-filter-pred-list\"><code>-filter<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-keep 'cdr '((1 2 3) (4 5) (6))) ;; =&gt; '((2 3) (5))\n(-keep (lambda (num) (when (&gt; num 3) (* 10 num))) '(1 2 3 4 5 6)) ;; =&gt; '(40 50 60)\n(--keep (when (&gt; it 3) (* 10 it)) '(1 2 3 4 5 6)) ;; =&gt; '(40 50 60)<\/pre><\/div><h4><a id=\"user-content--concat-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-concat-rest-lists\"><\/a>-concat <code>(&amp;rest lists)<\/code><\/h4><p>Return a new list with the concatenation of the elements in the supplied <code>lists<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-concat '(1)) ;; =&gt; '(1)\n(-concat '(1) '(2)) ;; =&gt; '(1 2)\n(-concat '(1) '(2 3) '(4)) ;; =&gt; '(1 2 3 4)<\/pre><\/div><h4><a id=\"user-content--flatten-l\" class=\"anchor\" aria-hidden=\"true\" href=\"#-flatten-l\"><\/a>-flatten <code>(l)<\/code><\/h4><p>Take a nested list <code>l<\/code> and return its contents as a single, flat list.<\/p><p>Note that because <code>nil<\/code> represents a list of zero elements (an empty list), any mention of nil in <code>l<\/code> will disappear after flattening. If you need to preserve nils, consider <a href=\"#-flatten-n-num-list\"><code>-flatten-n<\/code><\/a> or map them to some unique symbol and then map them back.<\/p><p>Conses of two atoms are considered \"terminals\", that is, they aren't flattened further.<\/p><p>See also: <a href=\"#-flatten-n-num-list\"><code>-flatten-n<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-flatten '((1))) ;; =&gt; '(1)\n(-flatten '((1 (2 3) (((4 (5))))))) ;; =&gt; '(1 2 3 4 5)\n(-flatten '(1 2 (3 . 4))) ;; =&gt; '(1 2 (3 . 4))<\/pre><\/div><h4><a id=\"user-content--flatten-n-num-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-flatten-n-num-list\"><\/a>-flatten-n <code>(num list)<\/code><\/h4><p>Flatten <code>num<\/code> levels of a nested <code>list<\/code>.<\/p><p>See also: <a href=\"#-flatten-l\"><code>-flatten<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-flatten-n 1 '((1 2) ((3 4) ((5 6))))) ;; =&gt; '(1 2 (3 4) ((5 6)))\n(-flatten-n 2 '((1 2) ((3 4) ((5 6))))) ;; =&gt; '(1 2 3 4 (5 6))\n(-flatten-n 3 '((1 2) ((3 4) ((5 6))))) ;; =&gt; '(1 2 3 4 5 6)<\/pre><\/div><h4><a id=\"user-content--replace-old-new-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-replace-old-new-list\"><\/a>-replace <code>(old new list)<\/code><\/h4><p>Replace all <code>old<\/code> items in <code>list<\/code> with <code>new<\/code>.<\/p><p>Elements are compared using <code>equal<\/code>.<\/p><p>See also: <a href=\"#-replace-at-n-x-list\"><code>-replace-at<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-replace 1 \"1\" '(1 2 3 4 3 2 1)) ;; =&gt; '(\"1\" 2 3 4 3 2 \"1\")\n(-replace \"foo\" \"bar\" '(\"a\" \"nice\" \"foo\" \"sentence\" \"about\" \"foo\")) ;; =&gt; '(\"a\" \"nice\" \"bar\" \"sentence\" \"about\" \"bar\")\n(-replace 1 2 nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--replace-first-old-new-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-replace-first-old-new-list\"><\/a>-replace-first <code>(old new list)<\/code><\/h4><p>Replace the first occurrence of <code>old<\/code> with <code>new<\/code> in <code>list<\/code>.<\/p><p>Elements are compared using <code>equal<\/code>.<\/p><p>See also: <a href=\"#-map-first-pred-rep-list\"><code>-map-first<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-replace-first 1 \"1\" '(1 2 3 4 3 2 1)) ;; =&gt; '(\"1\" 2 3 4 3 2 1)\n(-replace-first \"foo\" \"bar\" '(\"a\" \"nice\" \"foo\" \"sentence\" \"about\" \"foo\")) ;; =&gt; '(\"a\" \"nice\" \"bar\" \"sentence\" \"about\" \"foo\")\n(-replace-first 1 2 nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--replace-last-old-new-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-replace-last-old-new-list\"><\/a>-replace-last <code>(old new list)<\/code><\/h4><p>Replace the last occurrence of <code>old<\/code> with <code>new<\/code> in <code>list<\/code>.<\/p><p>Elements are compared using <code>equal<\/code>.<\/p><p>See also: <a href=\"#-map-last-pred-rep-list\"><code>-map-last<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-replace-last 1 \"1\" '(1 2 3 4 3 2 1)) ;; =&gt; '(1 2 3 4 3 2 \"1\")\n(-replace-last \"foo\" \"bar\" '(\"a\" \"nice\" \"foo\" \"sentence\" \"about\" \"foo\")) ;; =&gt; '(\"a\" \"nice\" \"foo\" \"sentence\" \"about\" \"bar\")\n(-replace-last 1 2 nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--insert-at-n-x-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-insert-at-n-x-list\"><\/a>-insert-at <code>(n x list)<\/code><\/h4><p>Return a list with <code>x<\/code> inserted into <code>list<\/code> at position <code>n<\/code>.<\/p><p>See also: <a href=\"#-splice-pred-fun-list\"><code>-splice<\/code><\/a>, <a href=\"#-splice-list-pred-new-list-list\"><code>-splice-list<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-insert-at 1 'x '(a b c)) ;; =&gt; '(a x b c)\n(-insert-at 12 'x '(a b c)) ;; =&gt; '(a b c x)<\/pre><\/div><h4><a id=\"user-content--replace-at-n-x-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-replace-at-n-x-list\"><\/a>-replace-at <code>(n x list)<\/code><\/h4><p>Return a list with element at Nth position in <code>list<\/code> replaced with <code>x<\/code>.<\/p><p>See also: <a href=\"#-replace-old-new-list\"><code>-replace<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-replace-at 0 9 '(0 1 2 3 4 5)) ;; =&gt; '(9 1 2 3 4 5)\n(-replace-at 1 9 '(0 1 2 3 4 5)) ;; =&gt; '(0 9 2 3 4 5)\n(-replace-at 4 9 '(0 1 2 3 4 5)) ;; =&gt; '(0 1 2 3 9 5)<\/pre><\/div><h4><a id=\"user-content--update-at-n-func-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-update-at-n-func-list\"><\/a>-update-at <code>(n func list)<\/code><\/h4><p>Return a list with element at Nth position in <code>list<\/code> replaced with <code>(func (nth n list))<\/code>.<\/p><p>See also: <a href=\"#-map-when-pred-rep-list\"><code>-map-when<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-update-at 0 (lambda (x) (+ x 9)) '(0 1 2 3 4 5)) ;; =&gt; '(9 1 2 3 4 5)\n(-update-at 1 (lambda (x) (+ x 8)) '(0 1 2 3 4 5)) ;; =&gt; '(0 9 2 3 4 5)\n(--update-at 2 (length it) '(\"foo\" \"bar\" \"baz\" \"quux\")) ;; =&gt; '(\"foo\" \"bar\" 3 \"quux\")<\/pre><\/div><h4><a id=\"user-content--remove-at-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-remove-at-n-list\"><\/a>-remove-at <code>(n list)<\/code><\/h4><p>Return a list with element at Nth position in <code>list<\/code> removed.<\/p><p>See also: <a href=\"#-remove-at-indices-indices-list\"><code>-remove-at-indices<\/code><\/a>, <a href=\"#-remove-pred-list\"><code>-remove<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-remove-at 0 '(\"0\" \"1\" \"2\" \"3\" \"4\" \"5\")) ;; =&gt; '(\"1\" \"2\" \"3\" \"4\" \"5\")\n(-remove-at 1 '(\"0\" \"1\" \"2\" \"3\" \"4\" \"5\")) ;; =&gt; '(\"0\" \"2\" \"3\" \"4\" \"5\")\n(-remove-at 2 '(\"0\" \"1\" \"2\" \"3\" \"4\" \"5\")) ;; =&gt; '(\"0\" \"1\" \"3\" \"4\" \"5\")<\/pre><\/div><h4><a id=\"user-content--remove-at-indices-indices-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-remove-at-indices-indices-list\"><\/a>-remove-at-indices <code>(indices list)<\/code><\/h4><p>Return a list whose elements are elements from <code>list<\/code> without elements selected as <code>(nth i list)<\/code> for all i from <code>indices<\/code>.<\/p><p>See also: <a href=\"#-remove-at-n-list\"><code>-remove-at<\/code><\/a>, <a href=\"#-remove-pred-list\"><code>-remove<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-remove-at-indices '(0) '(\"0\" \"1\" \"2\" \"3\" \"4\" \"5\")) ;; =&gt; '(\"1\" \"2\" \"3\" \"4\" \"5\")\n(-remove-at-indices '(0 2 4) '(\"0\" \"1\" \"2\" \"3\" \"4\" \"5\")) ;; =&gt; '(\"1\" \"3\" \"5\")\n(-remove-at-indices '(0 5) '(\"0\" \"1\" \"2\" \"3\" \"4\" \"5\")) ;; =&gt; '(\"1\" \"2\" \"3\" \"4\")<\/pre><\/div><h2><a id=\"user-content-reductions-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#reductions-1\"><\/a>Reductions<\/h2><p>Functions reducing lists into single value.<\/p><h4><a id=\"user-content--reduce-from-fn-initial-value-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reduce-from-fn-initial-value-list\"><\/a>-reduce-from <code>(fn initial-value list)<\/code><\/h4><p>Return the result of applying <code>fn<\/code> to <code>initial-value<\/code> and the first item in <code>list<\/code>, then applying <code>fn<\/code> to that result and the 2nd item, etc. If <code>list<\/code> contains no items, return <code>initial-value<\/code> and do not call <code>fn<\/code>.<\/p><p>In the anaphoric form <code>--reduce-from<\/code>, the accumulated value is exposed as symbol <code>acc<\/code>.<\/p><p>See also: <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a>, <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reduce-from '- 10 '(1 2 3)) ;; =&gt; 4\n(-reduce-from (lambda (memo item) (format \"(%s - %d)\" memo item)) \"10\" '(1 2 3)) ;; =&gt; \"(((10 - 1) - 2) - 3)\"\n(--reduce-from (concat acc \" \" it) \"START\" '(\"a\" \"b\" \"c\")) ;; =&gt; \"START a b c\"<\/pre><\/div><h4><a id=\"user-content--reduce-r-from-fn-initial-value-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reduce-r-from-fn-initial-value-list\"><\/a>-reduce-r-from <code>(fn initial-value list)<\/code><\/h4><p>Replace conses with <code>fn<\/code>, nil with <code>initial-value<\/code> and evaluate the resulting expression. If <code>list<\/code> is empty, <code>initial-value<\/code> is returned and <code>fn<\/code> is not called.<\/p><p>Note: this function works the same as <a href=\"#-reduce-from-fn-initial-value-list\"><code>-reduce-from<\/code><\/a> but the operation associates from right instead of from left.<\/p><p>See also: <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a>, <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reduce-r-from '- 10 '(1 2 3)) ;; =&gt; -8\n(-reduce-r-from (lambda (item memo) (format \"(%d - %s)\" item memo)) \"10\" '(1 2 3)) ;; =&gt; \"(1 - (2 - (3 - 10)))\"\n(--reduce-r-from (concat it \" \" acc) \"END\" '(\"a\" \"b\" \"c\")) ;; =&gt; \"a b c END\"<\/pre><\/div><h4><a id=\"user-content--reduce-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reduce-fn-list\"><\/a>-reduce <code>(fn list)<\/code><\/h4><p>Return the result of applying <code>fn<\/code> to the first 2 items in <code>list<\/code>, then applying <code>fn<\/code> to that result and the 3rd item, etc. If <code>list<\/code> contains no items, return the result of calling <code>fn<\/code> with no arguments. If <code>list<\/code> contains a single item, return that item and do not call <code>fn<\/code>.<\/p><p>In the anaphoric form <code>--reduce<\/code>, the accumulated value is exposed as symbol <code>acc<\/code>.<\/p><p>See also: <a href=\"#-reduce-from-fn-initial-value-list\"><code>-reduce-from<\/code><\/a>, <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reduce '- '(1 2 3 4)) ;; =&gt; -8\n(-reduce 'list '(1 2 3 4)) ;; =&gt; '(((1 2) 3) 4)\n(--reduce (format \"%s-%d\" acc it) '(1 2 3)) ;; =&gt; \"1-2-3\"<\/pre><\/div><h4><a id=\"user-content--reduce-r-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reduce-r-fn-list\"><\/a>-reduce-r <code>(fn list)<\/code><\/h4><p>Replace conses with <code>fn<\/code> and evaluate the resulting expression. The final nil is ignored. If <code>list<\/code> contains no items, return the result of calling <code>fn<\/code> with no arguments. If <code>list<\/code> contains a single item, return that item and do not call <code>fn<\/code>.<\/p><p>The first argument of <code>fn<\/code> is the new item, the second is the accumulated value.<\/p><p>Note: this function works the same as <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a> but the operation associates from right instead of from left.<\/p><p>See also: <a href=\"#-reduce-r-from-fn-initial-value-list\"><code>-reduce-r-from<\/code><\/a>, <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reduce-r '- '(1 2 3 4)) ;; =&gt; -2\n(-reduce-r (lambda (item memo) (format \"%s-%d\" memo item)) '(1 2 3)) ;; =&gt; \"3-2-1\"\n(--reduce-r (format \"%s-%d\" acc it) '(1 2 3)) ;; =&gt; \"3-2-1\"<\/pre><\/div><h4><a id=\"user-content--reductions-from-fn-init-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reductions-from-fn-init-list\"><\/a>-reductions-from <code>(fn init list)<\/code><\/h4><p>Return a list of the intermediate values of the reduction.<\/p><p>See <a href=\"#-reduce-from-fn-initial-value-list\"><code>-reduce-from<\/code><\/a> for explanation of the arguments.<\/p><p>See also: <a href=\"#-reductions-fn-list\"><code>-reductions<\/code><\/a>, <a href=\"#-reductions-r-fn-list\"><code>-reductions-r<\/code><\/a>, <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reductions-from (lambda (a i) (format \"(%s FN %d)\" a i)) \"INIT\" '(1 2 3 4)) ;; =&gt; '(\"INIT\" \"(INIT FN 1)\" \"((INIT FN 1) FN 2)\" \"(((INIT FN 1) FN 2) FN 3)\" \"((((INIT FN 1) FN 2) FN 3) FN 4)\")\n(-reductions-from 'max 0 '(2 1 4 3)) ;; =&gt; '(0 2 2 4 4)\n(-reductions-from '* 1 '(1 2 3 4)) ;; =&gt; '(1 1 2 6 24)<\/pre><\/div><h4><a id=\"user-content--reductions-r-from-fn-init-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reductions-r-from-fn-init-list\"><\/a>-reductions-r-from <code>(fn init list)<\/code><\/h4><p>Return a list of the intermediate values of the reduction.<\/p><p>See <a href=\"#-reduce-r-from-fn-initial-value-list\"><code>-reduce-r-from<\/code><\/a> for explanation of the arguments.<\/p><p>See also: <a href=\"#-reductions-r-fn-list\"><code>-reductions-r<\/code><\/a>, <a href=\"#-reductions-fn-list\"><code>-reductions<\/code><\/a>, <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reductions-r-from (lambda (i a) (format \"(%d FN %s)\" i a)) \"INIT\" '(1 2 3 4)) ;; =&gt; '(\"(1 FN (2 FN (3 FN (4 FN INIT))))\" \"(2 FN (3 FN (4 FN INIT)))\" \"(3 FN (4 FN INIT))\" \"(4 FN INIT)\" \"INIT\")\n(-reductions-r-from 'max 0 '(2 1 4 3)) ;; =&gt; '(4 4 4 3 0)\n(-reductions-r-from '* 1 '(1 2 3 4)) ;; =&gt; '(24 24 12 4 1)<\/pre><\/div><h4><a id=\"user-content--reductions-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reductions-fn-list\"><\/a>-reductions <code>(fn list)<\/code><\/h4><p>Return a list of the intermediate values of the reduction.<\/p><p>See <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a> for explanation of the arguments.<\/p><p>See also: <a href=\"#-reductions-from-fn-init-list\"><code>-reductions-from<\/code><\/a>, <a href=\"#-reductions-r-fn-list\"><code>-reductions-r<\/code><\/a>, <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reductions (lambda (a i) (format \"(%s FN %d)\" a i)) '(1 2 3 4)) ;; =&gt; '(1 \"(1 FN 2)\" \"((1 FN 2) FN 3)\" \"(((1 FN 2) FN 3) FN 4)\")\n(-reductions '+ '(1 2 3 4)) ;; =&gt; '(1 3 6 10)\n(-reductions '* '(1 2 3 4)) ;; =&gt; '(1 2 6 24)<\/pre><\/div><h4><a id=\"user-content--reductions-r-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-reductions-r-fn-list\"><\/a>-reductions-r <code>(fn list)<\/code><\/h4><p>Return a list of the intermediate values of the reduction.<\/p><p>See <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a> for explanation of the arguments.<\/p><p>See also: <a href=\"#-reductions-r-from-fn-init-list\"><code>-reductions-r-from<\/code><\/a>, <a href=\"#-reductions-fn-list\"><code>-reductions<\/code><\/a>, <a href=\"#-reduce-fn-list\"><code>-reduce<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-reductions-r (lambda (i a) (format \"(%d FN %s)\" i a)) '(1 2 3 4)) ;; =&gt; '(\"(1 FN (2 FN (3 FN 4)))\" \"(2 FN (3 FN 4))\" \"(3 FN 4)\" 4)\n(-reductions-r '+ '(1 2 3 4)) ;; =&gt; '(10 9 7 4)\n(-reductions-r '* '(1 2 3 4)) ;; =&gt; '(24 24 12 4)<\/pre><\/div><h4><a id=\"user-content--count-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-count-pred-list\"><\/a>-count <code>(pred list)<\/code><\/h4><p>Counts the number of items in <code>list<\/code> where (<code>pred<\/code> item) is non-nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-count 'even? '(1 2 3 4 5)) ;; =&gt; 2\n(--count (&lt; it 4) '(1 2 3 4)) ;; =&gt; 3<\/pre><\/div><h4><a id=\"user-content--sum-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-sum-list\"><\/a>-sum <code>(list)<\/code><\/h4><p>Return the sum of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-sum '()) ;; =&gt; 0\n(-sum '(1)) ;; =&gt; 1\n(-sum '(1 2 3 4)) ;; =&gt; 10<\/pre><\/div><h4><a id=\"user-content--running-sum-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-running-sum-list\"><\/a>-running-sum <code>(list)<\/code><\/h4><p>Return a list with running sums of items in <code>list<\/code>.<\/p><p><code>list<\/code> must be non-empty.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-running-sum '(1 2 3 4)) ;; =&gt; '(1 3 6 10)\n(-running-sum '(1)) ;; =&gt; '(1)\n(-running-sum '()) ;; Error<\/pre><\/div><h4><a id=\"user-content--product-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-product-list\"><\/a>-product <code>(list)<\/code><\/h4><p>Return the product of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-product '()) ;; =&gt; 1\n(-product '(1)) ;; =&gt; 1\n(-product '(1 2 3 4)) ;; =&gt; 24<\/pre><\/div><h4><a id=\"user-content--running-product-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-running-product-list\"><\/a>-running-product <code>(list)<\/code><\/h4><p>Return a list with running products of items in <code>list<\/code>.<\/p><p><code>list<\/code> must be non-empty.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-running-product '(1 2 3 4)) ;; =&gt; '(1 2 6 24)\n(-running-product '(1)) ;; =&gt; '(1)\n(-running-product '()) ;; Error<\/pre><\/div><h4><a id=\"user-content--inits-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-inits-list\"><\/a>-inits <code>(list)<\/code><\/h4><p>Return all prefixes of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-inits '(1 2 3 4)) ;; =&gt; '(nil (1) (1 2) (1 2 3) (1 2 3 4))\n(-inits nil) ;; =&gt; '(nil)\n(-inits '(1)) ;; =&gt; '(nil (1))<\/pre><\/div><h4><a id=\"user-content--tails-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tails-list\"><\/a>-tails <code>(list)<\/code><\/h4><p>Return all suffixes of <code>list<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tails '(1 2 3 4)) ;; =&gt; '((1 2 3 4) (2 3 4) (3 4) (4) nil)\n(-tails nil) ;; =&gt; '(nil)\n(-tails '(1)) ;; =&gt; '((1) nil)<\/pre><\/div><h4><a id=\"user-content--common-prefix-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-common-prefix-rest-lists\"><\/a>-common-prefix <code>(&amp;rest lists)<\/code><\/h4><p>Return the longest common prefix of <code>lists<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-common-prefix '(1)) ;; =&gt; '(1)\n(-common-prefix '(1 2) '(3 4) '(1 2)) ;; =&gt; nil\n(-common-prefix '(1 2) '(1 2 3) '(1 2 3 4)) ;; =&gt; '(1 2)<\/pre><\/div><h4><a id=\"user-content--common-suffix-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-common-suffix-rest-lists\"><\/a>-common-suffix <code>(&amp;rest lists)<\/code><\/h4><p>Return the longest common suffix of <code>lists<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-common-suffix '(1)) ;; =&gt; '(1)\n(-common-suffix '(1 2) '(3 4) '(1 2)) ;; =&gt; nil\n(-common-suffix '(1 2 3 4) '(2 3 4) '(3 4)) ;; =&gt; '(3 4)<\/pre><\/div><h4><a id=\"user-content--min-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-min-list\"><\/a>-min <code>(list)<\/code><\/h4><p>Return the smallest value from <code>list<\/code> of numbers or markers.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-min '(0)) ;; =&gt; 0\n(-min '(3 2 1)) ;; =&gt; 1\n(-min '(1 2 3)) ;; =&gt; 1<\/pre><\/div><h4><a id=\"user-content--min-by-comparator-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-min-by-comparator-list\"><\/a>-min-by <code>(comparator list)<\/code><\/h4><p>Take a comparison function <code>comparator<\/code> and a <code>list<\/code> and return the least element of the list by the comparison function.<\/p><p>See also combinator <a href=\"#-on-operator-transformer\"><code>-on<\/code><\/a> which can transform the values before comparing them.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-min-by '&gt; '(4 3 6 1)) ;; =&gt; 1\n(--min-by (&gt; (car it) (car other)) '((1 2 3) (2) (3 2))) ;; =&gt; '(1 2 3)\n(--min-by (&gt; (length it) (length other)) '((1 2 3) (2) (3 2))) ;; =&gt; '(2)<\/pre><\/div><h4><a id=\"user-content--max-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-max-list\"><\/a>-max <code>(list)<\/code><\/h4><p>Return the largest value from <code>list<\/code> of numbers or markers.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-max '(0)) ;; =&gt; 0\n(-max '(3 2 1)) ;; =&gt; 3\n(-max '(1 2 3)) ;; =&gt; 3<\/pre><\/div><h4><a id=\"user-content--max-by-comparator-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-max-by-comparator-list\"><\/a>-max-by <code>(comparator list)<\/code><\/h4><p>Take a comparison function <code>comparator<\/code> and a <code>list<\/code> and return the greatest element of the list by the comparison function.<\/p><p>See also combinator <a href=\"#-on-operator-transformer\"><code>-on<\/code><\/a> which can transform the values before comparing them.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-max-by '&gt; '(4 3 6 1)) ;; =&gt; 6\n(--max-by (&gt; (car it) (car other)) '((1 2 3) (2) (3 2))) ;; =&gt; '(3 2)\n(--max-by (&gt; (length it) (length other)) '((1 2 3) (2) (3 2))) ;; =&gt; '(1 2 3)<\/pre><\/div><h2><a id=\"user-content-unfolding-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#unfolding-1\"><\/a>Unfolding<\/h2><p>Operations dual to reductions, building lists from seed value rather than consuming a list to produce a single value.<\/p><h4><a id=\"user-content--iterate-fun-init-n\" class=\"anchor\" aria-hidden=\"true\" href=\"#-iterate-fun-init-n\"><\/a>-iterate <code>(fun init n)<\/code><\/h4><p>Return a list of iterated applications of <code>fun<\/code> to <code>init<\/code>.<\/p><p>This means a list of form:<\/p><pre>(init (fun init) (fun (fun init)) ...)\n<\/pre><p><code>n<\/code> is the length of the returned list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-iterate '1+ 1 10) ;; =&gt; '(1 2 3 4 5 6 7 8 9 10)\n(-iterate (lambda (x) (+ x x)) 2 5) ;; =&gt; '(2 4 8 16 32)\n(--iterate (* it it) 2 5) ;; =&gt; '(2 4 16 256 65536)<\/pre><\/div><h4><a id=\"user-content--unfold-fun-seed\" class=\"anchor\" aria-hidden=\"true\" href=\"#-unfold-fun-seed\"><\/a>-unfold <code>(fun seed)<\/code><\/h4><p>Build a list from <code>seed<\/code> using <code>fun<\/code>.<\/p><p>This is \"dual\" operation to <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a>: while -reduce-r consumes a list to produce a single value, <a href=\"#-unfold-fun-seed\"><code>-unfold<\/code><\/a> takes a seed value and builds a (potentially infinite!) list.<\/p><p><code>fun<\/code> should return <code>nil<\/code> to stop the generating process, or a cons (<code>a<\/code> . <code>b<\/code>), where <code>a<\/code> will be prepended to the result and <code>b<\/code> is the new seed.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-unfold (lambda (x) (unless (= x 0) (cons x (1- x)))) 10) ;; =&gt; '(10 9 8 7 6 5 4 3 2 1)\n(--unfold (when it (cons it (cdr it))) '(1 2 3 4)) ;; =&gt; '((1 2 3 4) (2 3 4) (3 4) (4))\n(--unfold (when it (cons it (butlast it))) '(1 2 3 4)) ;; =&gt; '((1 2 3 4) (1 2 3) (1 2) (1))<\/pre><\/div><h2><a id=\"user-content-predicates-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#predicates-1\"><\/a>Predicates<\/h2><h4><a id=\"user-content--any-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-any-pred-list\"><\/a>-any? <code>(pred list)<\/code><\/h4><p>Return t if (<code>pred<\/code> x) is non-nil for any x in <code>list<\/code>, else nil.<\/p><p>Alias: <code>-any-p<\/code>, <code>-some?<\/code>, <code>-some-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-any? 'even? '(1 2 3)) ;; =&gt; t\n(-any? 'even? '(1 3 5)) ;; =&gt; nil\n(-any? 'null '(1 3 5)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--all-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-all-pred-list\"><\/a>-all? <code>(pred list)<\/code><\/h4><p>Return t if (<code>pred<\/code> x) is non-nil for all x in <code>list<\/code>, else nil.<\/p><p>Alias: <code>-all-p<\/code>, <code>-every?<\/code>, <code>-every-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-all? 'even? '(1 2 3)) ;; =&gt; nil\n(-all? 'even? '(2 4 6)) ;; =&gt; t\n(--all? (= 0 (% it 2)) '(2 4 6)) ;; =&gt; t<\/pre><\/div><h4><a id=\"user-content--none-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-none-pred-list\"><\/a>-none? <code>(pred list)<\/code><\/h4><p>Return t if (<code>pred<\/code> x) is nil for all x in <code>list<\/code>, else nil.<\/p><p>Alias: <code>-none-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-none? 'even? '(1 2 3)) ;; =&gt; nil\n(-none? 'even? '(1 3 5)) ;; =&gt; t\n(--none? (= 0 (% it 2)) '(1 2 3)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--only-some-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-only-some-pred-list\"><\/a>-only-some? <code>(pred list)<\/code><\/h4><p>Return <code>t<\/code> if at least one item of <code>list<\/code> matches <code>pred<\/code> and at least one item of <code>list<\/code> does not match <code>pred<\/code>. Return <code>nil<\/code> both if all items match the predicate or if none of the items match the predicate.<\/p><p>Alias: <code>-only-some-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-only-some? 'even? '(1 2 3)) ;; =&gt; t\n(-only-some? 'even? '(1 3 5)) ;; =&gt; nil\n(-only-some? 'even? '(2 4 6)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--contains-list-element\" class=\"anchor\" aria-hidden=\"true\" href=\"#-contains-list-element\"><\/a>-contains? <code>(list element)<\/code><\/h4><p>Return non-nil if <code>list<\/code> contains <code>element<\/code>.<\/p><p>The test for equality is done with <code>equal<\/code>, or with <code>-compare-fn<\/code> if that's non-nil.<\/p><p>Alias: <code>-contains-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-contains? '(1 2 3) 1) ;; =&gt; t\n(-contains? '(1 2 3) 2) ;; =&gt; t\n(-contains? '(1 2 3) 4) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--same-items-list-list2\" class=\"anchor\" aria-hidden=\"true\" href=\"#-same-items-list-list2\"><\/a>-same-items? <code>(list list2)<\/code><\/h4><p>Return true if <code>list<\/code> and <code>list2<\/code> has the same items.<\/p><p>The order of the elements in the lists does not matter.<\/p><p>Alias: <code>-same-items-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-same-items? '(1 2 3) '(1 2 3)) ;; =&gt; t\n(-same-items? '(1 2 3) '(3 2 1)) ;; =&gt; t\n(-same-items? '(1 2 3) '(1 2 3 4)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--is-prefix-prefix-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-is-prefix-prefix-list\"><\/a>-is-prefix? <code>(prefix list)<\/code><\/h4><p>Return non-nil if <code>prefix<\/code> is prefix of <code>list<\/code>.<\/p><p>Alias: <code>-is-prefix-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-is-prefix? '(1 2 3) '(1 2 3 4 5)) ;; =&gt; t\n(-is-prefix? '(1 2 3 4 5) '(1 2 3)) ;; =&gt; nil\n(-is-prefix? '(1 3) '(1 2 3 4 5)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--is-suffix-suffix-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-is-suffix-suffix-list\"><\/a>-is-suffix? <code>(suffix list)<\/code><\/h4><p>Return non-nil if <code>suffix<\/code> is suffix of <code>list<\/code>.<\/p><p>Alias: <code>-is-suffix-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-is-suffix? '(3 4 5) '(1 2 3 4 5)) ;; =&gt; t\n(-is-suffix? '(1 2 3 4 5) '(3 4 5)) ;; =&gt; nil\n(-is-suffix? '(3 5) '(1 2 3 4 5)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--is-infix-infix-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-is-infix-infix-list\"><\/a>-is-infix? <code>(infix list)<\/code><\/h4><p>Return non-nil if <code>infix<\/code> is infix of <code>list<\/code>.<\/p><p>This operation runs in <code>o<\/code>(n^2) time<\/p><p>Alias: <code>-is-infix-p<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-is-infix? '(1 2 3) '(1 2 3 4 5)) ;; =&gt; t\n(-is-infix? '(2 3 4) '(1 2 3 4 5)) ;; =&gt; t\n(-is-infix? '(3 4 5) '(1 2 3 4 5)) ;; =&gt; t<\/pre><\/div><h2><a id=\"user-content-partitioning-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#partitioning-1\"><\/a>Partitioning<\/h2><p>Functions partitioning the input list into a list of lists.<\/p><h4><a id=\"user-content--split-at-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-split-at-n-list\"><\/a>-split-at <code>(n list)<\/code><\/h4><p>Return a list of ((-take <code>n<\/code> <code>list<\/code>) (-drop <code>n<\/code> <code>list<\/code>)), in no more than one pass through the list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-split-at 3 '(1 2 3 4 5)) ;; =&gt; '((1 2 3) (4 5))\n(-split-at 17 '(1 2 3 4 5)) ;; =&gt; '((1 2 3 4 5) nil)<\/pre><\/div><h4><a id=\"user-content--split-with-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-split-with-pred-list\"><\/a>-split-with <code>(pred list)<\/code><\/h4><p>Return a list of ((-take-while <code>pred<\/code> <code>list<\/code>) (-drop-while <code>pred<\/code> <code>list<\/code>)), in no more than one pass through the list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-split-with 'even? '(1 2 3 4)) ;; =&gt; '(nil (1 2 3 4))\n(-split-with 'even? '(2 4 5 6)) ;; =&gt; '((2 4) (5 6))\n(--split-with (&lt; it 4) '(1 2 3 4 3 2 1)) ;; =&gt; '((1 2 3) (4 3 2 1))<\/pre><\/div><h4><a id=\"user-content--split-on-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-split-on-item-list\"><\/a>-split-on <code>(item list)<\/code><\/h4><p>Split the <code>list<\/code> each time <code>item<\/code> is found.<\/p><p>Unlike <a href=\"#-partition-by-fn-list\"><code>-partition-by<\/code><\/a>, the <code>item<\/code> is discarded from the results. Empty lists are also removed from the result.<\/p><p>Comparison is done by <code>equal<\/code>.<\/p><p>See also <a href=\"#-split-when-fn-list\"><code>-split-when<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-split-on '| '(Nil | Leaf a | Node [Tree a])) ;; =&gt; '((Nil) (Leaf a) (Node [Tree a]))\n(-split-on ':endgroup '(\"a\" \"b\" :endgroup \"c\" :endgroup \"d\" \"e\")) ;; =&gt; '((\"a\" \"b\") (\"c\") (\"d\" \"e\"))\n(-split-on ':endgroup '(\"a\" \"b\" :endgroup :endgroup \"d\" \"e\")) ;; =&gt; '((\"a\" \"b\") (\"d\" \"e\"))<\/pre><\/div><h4><a id=\"user-content--split-when-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-split-when-fn-list\"><\/a>-split-when <code>(fn list)<\/code><\/h4><p>Split the <code>list<\/code> on each element where <code>fn<\/code> returns non-nil.<\/p><p>Unlike <a href=\"#-partition-by-fn-list\"><code>-partition-by<\/code><\/a>, the \"matched\" element is discarded from the results. Empty lists are also removed from the result.<\/p><p>This function can be thought of as a generalization of <code>split-string<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-split-when 'even? '(1 2 3 4 5 6)) ;; =&gt; '((1) (3) (5))\n(-split-when 'even? '(1 2 3 4 6 8 9)) ;; =&gt; '((1) (3) (9))\n(--split-when (memq it '(&amp;optional &amp;rest)) '(a b &amp;optional c d &amp;rest args)) ;; =&gt; '((a b) (c d) (args))<\/pre><\/div><h4><a id=\"user-content--separate-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-separate-pred-list\"><\/a>-separate <code>(pred list)<\/code><\/h4><p>Return a list of ((-filter <code>pred<\/code> <code>list<\/code>) (-remove <code>pred<\/code> <code>list<\/code>)), in one pass through the list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-separate (lambda (num) (= 0 (% num 2))) '(1 2 3 4 5 6 7)) ;; =&gt; '((2 4 6) (1 3 5 7))\n(--separate (&lt; it 5) '(3 7 5 9 3 2 1 4 6)) ;; =&gt; '((3 3 2 1 4) (7 5 9 6))\n(-separate 'cdr '((1 2) (1) (1 2 3) (4))) ;; =&gt; '(((1 2) (1 2 3)) ((1) (4)))<\/pre><\/div><h4><a id=\"user-content--partition-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-n-list\"><\/a>-partition <code>(n list)<\/code><\/h4><p>Return a new list with the items in <code>list<\/code> grouped into <code>n-<\/code>sized sublists. If there are not enough items to make the last group <code>n-<\/code>sized, those items are discarded.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition 2 '(1 2 3 4 5 6)) ;; =&gt; '((1 2) (3 4) (5 6))\n(-partition 2 '(1 2 3 4 5 6 7)) ;; =&gt; '((1 2) (3 4) (5 6))\n(-partition 3 '(1 2 3 4 5 6 7)) ;; =&gt; '((1 2 3) (4 5 6))<\/pre><\/div><h4><a id=\"user-content--partition-all-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-all-n-list\"><\/a>-partition-all <code>(n list)<\/code><\/h4><p>Return a new list with the items in <code>list<\/code> grouped into <code>n-<\/code>sized sublists. The last group may contain less than <code>n<\/code> items.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-all 2 '(1 2 3 4 5 6)) ;; =&gt; '((1 2) (3 4) (5 6))\n(-partition-all 2 '(1 2 3 4 5 6 7)) ;; =&gt; '((1 2) (3 4) (5 6) (7))\n(-partition-all 3 '(1 2 3 4 5 6 7)) ;; =&gt; '((1 2 3) (4 5 6) (7))<\/pre><\/div><h4><a id=\"user-content--partition-in-steps-n-step-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-in-steps-n-step-list\"><\/a>-partition-in-steps <code>(n step list)<\/code><\/h4><p>Return a new list with the items in <code>list<\/code> grouped into <code>n-<\/code>sized sublists at offsets <code>step<\/code> apart. If there are not enough items to make the last group <code>n-<\/code>sized, those items are discarded.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-in-steps 2 1 '(1 2 3 4)) ;; =&gt; '((1 2) (2 3) (3 4))\n(-partition-in-steps 3 2 '(1 2 3 4)) ;; =&gt; '((1 2 3))\n(-partition-in-steps 3 2 '(1 2 3 4 5)) ;; =&gt; '((1 2 3) (3 4 5))<\/pre><\/div><h4><a id=\"user-content--partition-all-in-steps-n-step-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-all-in-steps-n-step-list\"><\/a>-partition-all-in-steps <code>(n step list)<\/code><\/h4><p>Return a new list with the items in <code>list<\/code> grouped into <code>n-<\/code>sized sublists at offsets <code>step<\/code> apart. The last groups may contain less than <code>n<\/code> items.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-all-in-steps 2 1 '(1 2 3 4)) ;; =&gt; '((1 2) (2 3) (3 4) (4))\n(-partition-all-in-steps 3 2 '(1 2 3 4)) ;; =&gt; '((1 2 3) (3 4))\n(-partition-all-in-steps 3 2 '(1 2 3 4 5)) ;; =&gt; '((1 2 3) (3 4 5) (5))<\/pre><\/div><h4><a id=\"user-content--partition-by-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-by-fn-list\"><\/a>-partition-by <code>(fn list)<\/code><\/h4><p>Apply <code>fn<\/code> to each item in <code>list<\/code>, splitting it each time <code>fn<\/code> returns a new value.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-by 'even? '()) ;; =&gt; '()\n(-partition-by 'even? '(1 1 2 2 2 3 4 6 8)) ;; =&gt; '((1 1) (2 2 2) (3) (4 6 8))\n(--partition-by (&lt; it 3) '(1 2 3 4 3 2 1)) ;; =&gt; '((1 2) (3 4 3) (2 1))<\/pre><\/div><h4><a id=\"user-content--partition-by-header-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-by-header-fn-list\"><\/a>-partition-by-header <code>(fn list)<\/code><\/h4><p>Apply <code>fn<\/code> to the first item in <code>list<\/code>. That is the header value. Apply <code>fn<\/code> to each item in <code>list<\/code>, splitting it each time <code>fn<\/code> returns the header value, but only after seeing at least one other value (the body).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(--partition-by-header (= it 1) '(1 2 3 1 2 1 2 3 4)) ;; =&gt; '((1 2 3) (1 2) (1 2 3 4))\n(--partition-by-header (&gt; it 0) '(1 2 0 1 0 1 2 3 0)) ;; =&gt; '((1 2 0) (1 0) (1 2 3 0))\n(-partition-by-header 'even? '(2 1 1 1 4 1 3 5 6 6 1)) ;; =&gt; '((2 1 1 1) (4 1 3 5) (6 6 1))<\/pre><\/div><h4><a id=\"user-content--partition-after-pred-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-after-pred-pred-list\"><\/a>-partition-after-pred <code>(pred list)<\/code><\/h4><p>Partition directly after each time <code>pred<\/code> is true on an element of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-after-pred #'odd? '()) ;; =&gt; '()\n(-partition-after-pred #'odd? '(1)) ;; =&gt; '((1))\n(-partition-after-pred #'odd? '(0 1)) ;; =&gt; '((0 1))<\/pre><\/div><h4><a id=\"user-content--partition-before-pred-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-before-pred-pred-list\"><\/a>-partition-before-pred <code>(pred list)<\/code><\/h4><p>Partition directly before each time <code>pred<\/code> is true on an element of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-before-pred #'odd? '()) ;; =&gt; '()\n(-partition-before-pred #'odd? '(1)) ;; =&gt; '((1))\n(-partition-before-pred #'odd? '(0 1)) ;; =&gt; '((0) (1))<\/pre><\/div><h4><a id=\"user-content--partition-before-item-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-before-item-item-list\"><\/a>-partition-before-item <code>(item list)<\/code><\/h4><p>Partition directly before each time <code>item<\/code> appears in <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-before-item 3 '()) ;; =&gt; '()\n(-partition-before-item 3 '(1)) ;; =&gt; '((1))\n(-partition-before-item 3 '(3)) ;; =&gt; '((3))<\/pre><\/div><h4><a id=\"user-content--partition-after-item-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partition-after-item-item-list\"><\/a>-partition-after-item <code>(item list)<\/code><\/h4><p>Partition directly after each time <code>item<\/code> appears in <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-partition-after-item 3 '()) ;; =&gt; '()\n(-partition-after-item 3 '(1)) ;; =&gt; '((1))\n(-partition-after-item 3 '(3)) ;; =&gt; '((3))<\/pre><\/div><h4><a id=\"user-content--group-by-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-group-by-fn-list\"><\/a>-group-by <code>(fn list)<\/code><\/h4><p>Separate <code>list<\/code> into an alist whose keys are <code>fn<\/code> applied to the elements of <code>list<\/code>. Keys are compared by <code>equal<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-group-by 'even? '()) ;; =&gt; '()\n(-group-by 'even? '(1 1 2 2 2 3 4 6 8)) ;; =&gt; '((nil 1 1 3) (t 2 2 2 4 6 8))\n(--group-by (car (split-string it \"\/\")) '(\"a\/b\" \"c\/d\" \"a\/e\")) ;; =&gt; '((\"a\" \"a\/b\" \"a\/e\") (\"c\" \"c\/d\"))<\/pre><\/div><h2><a id=\"user-content-indexing-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#indexing-1\"><\/a>Indexing<\/h2><p>Return indices of elements based on predicates, sort elements by indices etc.<\/p><h4><a id=\"user-content--elem-index-elem-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-elem-index-elem-list\"><\/a>-elem-index <code>(elem list)<\/code><\/h4><p>Return the index of the first element in the given <code>list<\/code> which is equal to the query element <code>elem<\/code>, or nil if there is no such element.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-elem-index 2 '(6 7 8 2 3 4)) ;; =&gt; 3\n(-elem-index \"bar\" '(\"foo\" \"bar\" \"baz\")) ;; =&gt; 1\n(-elem-index '(1 2) '((3) (5 6) (1 2) nil)) ;; =&gt; 2<\/pre><\/div><h4><a id=\"user-content--elem-indices-elem-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-elem-indices-elem-list\"><\/a>-elem-indices <code>(elem list)<\/code><\/h4><p>Return the indices of all elements in <code>list<\/code> equal to the query element <code>elem<\/code>, in ascending order.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-elem-indices 2 '(6 7 8 2 3 4 2 1)) ;; =&gt; '(3 6)\n(-elem-indices \"bar\" '(\"foo\" \"bar\" \"baz\")) ;; =&gt; '(1)\n(-elem-indices '(1 2) '((3) (1 2) (5 6) (1 2) nil)) ;; =&gt; '(1 3)<\/pre><\/div><h4><a id=\"user-content--find-index-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-find-index-pred-list\"><\/a>-find-index <code>(pred list)<\/code><\/h4><p>Take a predicate <code>pred<\/code> and a <code>list<\/code> and return the index of the first element in the list satisfying the predicate, or nil if there is no such element.<\/p><p>See also <a href=\"#-first-pred-list\"><code>-first<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-find-index 'even? '(2 4 1 6 3 3 5 8)) ;; =&gt; 0\n(--find-index (&lt; 5 it) '(2 4 1 6 3 3 5 8)) ;; =&gt; 3\n(-find-index (-partial 'string-lessp \"baz\") '(\"bar\" \"foo\" \"baz\")) ;; =&gt; 1<\/pre><\/div><h4><a id=\"user-content--find-last-index-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-find-last-index-pred-list\"><\/a>-find-last-index <code>(pred list)<\/code><\/h4><p>Take a predicate <code>pred<\/code> and a <code>list<\/code> and return the index of the last element in the list satisfying the predicate, or nil if there is no such element.<\/p><p>See also <a href=\"#-last-pred-list\"><code>-last<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-find-last-index 'even? '(2 4 1 6 3 3 5 8)) ;; =&gt; 7\n(--find-last-index (&lt; 5 it) '(2 7 1 6 3 8 5 2)) ;; =&gt; 5\n(-find-last-index (-partial 'string-lessp \"baz\") '(\"q\" \"foo\" \"baz\")) ;; =&gt; 1<\/pre><\/div><h4><a id=\"user-content--find-indices-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-find-indices-pred-list\"><\/a>-find-indices <code>(pred list)<\/code><\/h4><p>Return the indices of all elements in <code>list<\/code> satisfying the predicate <code>pred<\/code>, in ascending order.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-find-indices 'even? '(2 4 1 6 3 3 5 8)) ;; =&gt; '(0 1 3 7)\n(--find-indices (&lt; 5 it) '(2 4 1 6 3 3 5 8)) ;; =&gt; '(3 7)\n(-find-indices (-partial 'string-lessp \"baz\") '(\"bar\" \"foo\" \"baz\")) ;; =&gt; '(1)<\/pre><\/div><h4><a id=\"user-content--grade-up-comparator-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-grade-up-comparator-list\"><\/a>-grade-up <code>(comparator list)<\/code><\/h4><p>Grade elements of <code>list<\/code> using <code>comparator<\/code> relation, yielding a permutation vector such that applying this permutation to <code>list<\/code> sorts it in ascending order.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-grade-up '&lt; '(3 1 4 2 1 3 3)) ;; =&gt; '(1 4 3 0 5 6 2)\n(let ((l '(3 1 4 2 1 3 3))) (-select-by-indices (-grade-up '&lt; l) l)) ;; =&gt; '(1 1 2 3 3 3 4)<\/pre><\/div><h4><a id=\"user-content--grade-down-comparator-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-grade-down-comparator-list\"><\/a>-grade-down <code>(comparator list)<\/code><\/h4><p>Grade elements of <code>list<\/code> using <code>comparator<\/code> relation, yielding a permutation vector such that applying this permutation to <code>list<\/code> sorts it in descending order.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-grade-down '&lt; '(3 1 4 2 1 3 3)) ;; =&gt; '(2 0 5 6 3 1 4)\n(let ((l '(3 1 4 2 1 3 3))) (-select-by-indices (-grade-down '&lt; l) l)) ;; =&gt; '(4 3 3 3 2 1 1)<\/pre><\/div><h2><a id=\"user-content-set-operations-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#set-operations-1\"><\/a>Set operations<\/h2><p>Operations pretending lists are sets.<\/p><h4><a id=\"user-content--union-list-list2\" class=\"anchor\" aria-hidden=\"true\" href=\"#-union-list-list2\"><\/a>-union <code>(list list2)<\/code><\/h4><p>Return a new list containing the elements of <code>list<\/code> and elements of <code>list2<\/code> that are not in <code>list<\/code>. The test for equality is done with <code>equal<\/code>, or with <code>-compare-fn<\/code> if that's non-nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-union '(1 2 3) '(3 4 5)) ;; =&gt; '(1 2 3 4 5)\n(-union '(1 2 3 4) '()) ;; =&gt; '(1 2 3 4)\n(-union '(1 1 2 2) '(3 2 1)) ;; =&gt; '(1 1 2 2 3)<\/pre><\/div><h4><a id=\"user-content--difference-list-list2\" class=\"anchor\" aria-hidden=\"true\" href=\"#-difference-list-list2\"><\/a>-difference <code>(list list2)<\/code><\/h4><p>Return a new list with only the members of <code>list<\/code> that are not in <code>list2<\/code>. The test for equality is done with <code>equal<\/code>, or with <code>-compare-fn<\/code> if that's non-nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-difference '() '()) ;; =&gt; '()\n(-difference '(1 2 3) '(4 5 6)) ;; =&gt; '(1 2 3)\n(-difference '(1 2 3 4) '(3 4 5 6)) ;; =&gt; '(1 2)<\/pre><\/div><h4><a id=\"user-content--intersection-list-list2\" class=\"anchor\" aria-hidden=\"true\" href=\"#-intersection-list-list2\"><\/a>-intersection <code>(list list2)<\/code><\/h4><p>Return a new list containing only the elements that are members of both <code>list<\/code> and <code>list2<\/code>. The test for equality is done with <code>equal<\/code>, or with <code>-compare-fn<\/code> if that's non-nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-intersection '() '()) ;; =&gt; '()\n(-intersection '(1 2 3) '(4 5 6)) ;; =&gt; '()\n(-intersection '(1 2 3 4) '(3 4 5 6)) ;; =&gt; '(3 4)<\/pre><\/div><h4><a id=\"user-content--powerset-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-powerset-list\"><\/a>-powerset <code>(list)<\/code><\/h4><p>Return the power set of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-powerset '()) ;; =&gt; '(nil)\n(-powerset '(x y z)) ;; =&gt; '((x y z) (x y) (x z) (x) (y z) (y) (z) nil)<\/pre><\/div><h4><a id=\"user-content--permutations-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-permutations-list\"><\/a>-permutations <code>(list)<\/code><\/h4><p>Return the permutations of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-permutations '()) ;; =&gt; '(nil)\n(-permutations '(1 2)) ;; =&gt; '((1 2) (2 1))\n(-permutations '(a b c)) ;; =&gt; '((a b c) (a c b) (b a c) (b c a) (c a b) (c b a))<\/pre><\/div><h4><a id=\"user-content--distinct-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-distinct-list\"><\/a>-distinct <code>(list)<\/code><\/h4><p>Return a new list with all duplicates removed. The test for equality is done with <code>equal<\/code>, or with <code>-compare-fn<\/code> if that's non-nil.<\/p><p>Alias: <code>-uniq<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-distinct '()) ;; =&gt; '()\n(-distinct '(1 2 2 4)) ;; =&gt; '(1 2 4)\n(-distinct '(t t t)) ;; =&gt; '(t)<\/pre><\/div><h2><a id=\"user-content-other-list-operations-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-list-operations-1\"><\/a>Other list operations<\/h2><p>Other list functions not fit to be classified elsewhere.<\/p><h4><a id=\"user-content--rotate-n-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-rotate-n-list\"><\/a>-rotate <code>(n list)<\/code><\/h4><p>Rotate <code>list<\/code> <code>n<\/code> places to the right. With <code>n<\/code> negative, rotate to the left. The time complexity is <code>o<\/code>(n).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-rotate 3 '(1 2 3 4 5 6 7)) ;; =&gt; '(5 6 7 1 2 3 4)\n(-rotate -3 '(1 2 3 4 5 6 7)) ;; =&gt; '(4 5 6 7 1 2 3)\n(-rotate 16 '(1 2 3 4 5 6 7)) ;; =&gt; '(6 7 1 2 3 4 5)<\/pre><\/div><h4><a id=\"user-content--repeat-n-x\" class=\"anchor\" aria-hidden=\"true\" href=\"#-repeat-n-x\"><\/a>-repeat <code>(n x)<\/code><\/h4><p>Return a list with <code>x<\/code> repeated <code>n<\/code> times. Return nil if <code>n<\/code> is less than 1.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-repeat 3 :a) ;; =&gt; '(:a :a :a)\n(-repeat 1 :a) ;; =&gt; '(:a)\n(-repeat 0 :a) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--cons-rest-args\" class=\"anchor\" aria-hidden=\"true\" href=\"#-cons-rest-args\"><\/a>-cons* <code>(&amp;rest args)<\/code><\/h4><p>Make a new list from the elements of <code>args<\/code>.<\/p><p>The last 2 members of <code>args<\/code> are used as the final cons of the result so if the final member of <code>args<\/code> is not a list the result is a dotted list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-cons* 1 2) ;; =&gt; '(1 . 2)\n(-cons* 1 2 3) ;; =&gt; '(1 2 . 3)\n(-cons* 1) ;; =&gt; 1<\/pre><\/div><h4><a id=\"user-content--snoc-list-elem-rest-elements\" class=\"anchor\" aria-hidden=\"true\" href=\"#-snoc-list-elem-rest-elements\"><\/a>-snoc <code>(list elem &amp;rest elements)<\/code><\/h4><p>Append <code>elem<\/code> to the end of the list.<\/p><p>This is like <code>cons<\/code>, but operates on the end of list.<\/p><p>If <code>elements<\/code> is non nil, append these to the list as well.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-snoc '(1 2 3) 4) ;; =&gt; '(1 2 3 4)\n(-snoc '(1 2 3) 4 5 6) ;; =&gt; '(1 2 3 4 5 6)\n(-snoc '(1 2 3) '(4 5 6)) ;; =&gt; '(1 2 3 (4 5 6))<\/pre><\/div><h4><a id=\"user-content--interpose-sep-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-interpose-sep-list\"><\/a>-interpose <code>(sep list)<\/code><\/h4><p>Return a new list of all elements in <code>list<\/code> separated by <code>sep<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-interpose \"-\" '()) ;; =&gt; '()\n(-interpose \"-\" '(\"a\")) ;; =&gt; '(\"a\")\n(-interpose \"-\" '(\"a\" \"b\" \"c\")) ;; =&gt; '(\"a\" \"-\" \"b\" \"-\" \"c\")<\/pre><\/div><h4><a id=\"user-content--interleave-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-interleave-rest-lists\"><\/a>-interleave <code>(&amp;rest lists)<\/code><\/h4><p>Return a new list of the first item in each list, then the second etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-interleave '(1 2) '(\"a\" \"b\")) ;; =&gt; '(1 \"a\" 2 \"b\")\n(-interleave '(1 2) '(\"a\" \"b\") '(\"A\" \"B\")) ;; =&gt; '(1 \"a\" \"A\" 2 \"b\" \"B\")\n(-interleave '(1 2 3) '(\"a\" \"b\")) ;; =&gt; '(1 \"a\" 2 \"b\")<\/pre><\/div><h4><a id=\"user-content--zip-with-fn-list1-list2\" class=\"anchor\" aria-hidden=\"true\" href=\"#-zip-with-fn-list1-list2\"><\/a>-zip-with <code>(fn list1 list2)<\/code><\/h4><p>Zip the two lists <code>list1<\/code> and <code>list2<\/code> using a function <code>fn<\/code>. This function is applied pairwise taking as first argument element of <code>list1<\/code> and as second argument element of <code>list2<\/code> at corresponding position.<\/p><p>The anaphoric form <code>--zip-with<\/code> binds the elements from <code>list1<\/code> as symbol <code>it<\/code>, and the elements from <code>list2<\/code> as symbol <code>other<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-zip-with '+ '(1 2 3) '(4 5 6)) ;; =&gt; '(5 7 9)\n(-zip-with 'cons '(1 2 3) '(4 5 6)) ;; =&gt; '((1 . 4) (2 . 5) (3 . 6))\n(--zip-with (concat it \" and \" other) '(\"Batman\" \"Jekyll\") '(\"Robin\" \"Hyde\")) ;; =&gt; '(\"Batman and Robin\" \"Jekyll and Hyde\")<\/pre><\/div><h4><a id=\"user-content--zip-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-zip-rest-lists\"><\/a>-zip <code>(&amp;rest lists)<\/code><\/h4><p>Zip <code>lists<\/code> together. Group the head of each list, followed by the second elements of each list, and so on. The lengths of the returned groupings are equal to the length of the shortest input list.<\/p><p>If two lists are provided as arguments, return the groupings as a list of cons cells. Otherwise, return the groupings as a list of lists.<\/p><p>Please note! This distinction is being removed in an upcoming 3.0 release of Dash. If you rely on this behavior, use -zip-pair instead.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-zip '(1 2 3) '(4 5 6)) ;; =&gt; '((1 . 4) (2 . 5) (3 . 6))\n(-zip '(1 2 3) '(4 5 6 7)) ;; =&gt; '((1 . 4) (2 . 5) (3 . 6))\n(-zip '(1 2 3 4) '(4 5 6)) ;; =&gt; '((1 . 4) (2 . 5) (3 . 6))<\/pre><\/div><h4><a id=\"user-content--zip-fill-fill-value-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-zip-fill-fill-value-rest-lists\"><\/a>-zip-fill <code>(fill-value &amp;rest lists)<\/code><\/h4><p>Zip <code>lists<\/code>, with <code>fill-value<\/code> padded onto the shorter lists. The lengths of the returned groupings are equal to the length of the longest input list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-zip-fill 0 '(1 2 3 4 5) '(6 7 8 9)) ;; =&gt; '((1 . 6) (2 . 7) (3 . 8) (4 . 9) (5 . 0))<\/pre><\/div><h4><a id=\"user-content--unzip-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-unzip-lists\"><\/a>-unzip <code>(lists)<\/code><\/h4><p>Unzip <code>lists<\/code>.<\/p><p>This works just like <a href=\"#-zip-rest-lists\"><code>-zip<\/code><\/a> but takes a list of lists instead of a variable number of arguments, such that<\/p><pre>(-unzip (-zip `l1` `l2` `l3` ...))\n<\/pre><p>is identity (given that the lists are the same length).<\/p><p>See also: <a href=\"#-zip-rest-lists\"><code>-zip<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-unzip (-zip '(1 2 3) '(a b c) '(\"e\" \"f\" \"g\"))) ;; =&gt; '((1 2 3) (a b c) (\"e\" \"f\" \"g\"))\n(-unzip '((1 2) (3 4) (5 6) (7 8) (9 10))) ;; =&gt; '((1 3 5 7 9) (2 4 6 8 10))<\/pre><\/div><h4><a id=\"user-content--cycle-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-cycle-list\"><\/a>-cycle <code>(list)<\/code><\/h4><p>Return an infinite copy of <code>list<\/code> that will cycle through the elements and repeat from the beginning.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-take 5 (-cycle '(1 2 3))) ;; =&gt; '(1 2 3 1 2)\n(-take 7 (-cycle '(1 \"and\" 3))) ;; =&gt; '(1 \"and\" 3 1 \"and\" 3 1)\n(-zip (-cycle '(1 2 3)) '(1 2)) ;; =&gt; '((1 . 1) (2 . 2))<\/pre><\/div><h4><a id=\"user-content--pad-fill-value-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-pad-fill-value-rest-lists\"><\/a>-pad <code>(fill-value &amp;rest lists)<\/code><\/h4><p>Appends <code>fill-value<\/code> to the end of each list in <code>lists<\/code> such that they will all have the same length.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-pad 0 '()) ;; =&gt; '(nil)\n(-pad 0 '(1)) ;; =&gt; '((1))\n(-pad 0 '(1 2 3) '(4 5)) ;; =&gt; '((1 2 3) (4 5 0))<\/pre><\/div><h4><a id=\"user-content--table-fn-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-table-fn-rest-lists\"><\/a>-table <code>(fn &amp;rest lists)<\/code><\/h4><p>Compute outer product of <code>lists<\/code> using function <code>fn<\/code>.<\/p><p>The function <code>fn<\/code> should have the same arity as the number of supplied lists.<\/p><p>The outer product is computed by applying fn to all possible combinations created by taking one element from each list in order. The dimension of the result is (length lists).<\/p><p>See also: <a href=\"#-table-flat-fn-rest-lists\"><code>-table-flat<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-table '* '(1 2 3) '(1 2 3)) ;; =&gt; '((1 2 3) (2 4 6) (3 6 9))\n(-table (lambda (a b) (-sum (-zip-with '* a b))) '((1 2) (3 4)) '((1 3) (2 4))) ;; =&gt; '((7 15) (10 22))\n(apply '-table 'list (-repeat 3 '(1 2))) ;; =&gt; '((((1 1 1) (2 1 1)) ((1 2 1) (2 2 1))) (((1 1 2) (2 1 2)) ((1 2 2) (2 2 2))))<\/pre><\/div><h4><a id=\"user-content--table-flat-fn-rest-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#-table-flat-fn-rest-lists\"><\/a>-table-flat <code>(fn &amp;rest lists)<\/code><\/h4><p>Compute flat outer product of <code>lists<\/code> using function <code>fn<\/code>.<\/p><p>The function <code>fn<\/code> should have the same arity as the number of supplied lists.<\/p><p>The outer product is computed by applying fn to all possible combinations created by taking one element from each list in order. The results are flattened, ignoring the tensor structure of the result. This is equivalent to calling:<\/p><pre>(-flatten-n (1- (length lists)) (apply '-table fn lists))\n<\/pre><p>but the implementation here is much more efficient.<\/p><p>See also: <a href=\"#-flatten-n-num-list\"><code>-flatten-n<\/code><\/a>, <a href=\"#-table-fn-rest-lists\"><code>-table<\/code><\/a><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-table-flat 'list '(1 2 3) '(a b c)) ;; =&gt; '((1 a) (2 a) (3 a) (1 b) (2 b) (3 b) (1 c) (2 c) (3 c))\n(-table-flat '* '(1 2 3) '(1 2 3)) ;; =&gt; '(1 2 3 2 4 6 3 6 9)\n(apply '-table-flat 'list (-repeat 3 '(1 2))) ;; =&gt; '((1 1 1) (2 1 1) (1 2 1) (2 2 1) (1 1 2) (2 1 2) (1 2 2) (2 2 2))<\/pre><\/div><h4><a id=\"user-content--first-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-first-pred-list\"><\/a>-first <code>(pred list)<\/code><\/h4><p>Return the first x in <code>list<\/code> where (<code>pred<\/code> x) is non-nil, else nil.<\/p><p>To get the first item in the list no questions asked, use <code>car<\/code>.<\/p><p>Alias: <code>-find<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-first 'even? '(1 2 3)) ;; =&gt; 2\n(-first 'even? '(1 3 5)) ;; =&gt; nil\n(-first 'null '(1 3 5)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--some-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-some-pred-list\"><\/a>-some <code>(pred list)<\/code><\/h4><p>Return (<code>pred<\/code> x) for the first <code>list<\/code> item where (<code>pred<\/code> x) is non-nil, else nil.<\/p><p>Alias: <code>-any<\/code><\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-some 'even? '(1 2 3)) ;; =&gt; t\n(-some 'null '(1 2 3)) ;; =&gt; nil\n(-some 'null '(1 2 nil)) ;; =&gt; t<\/pre><\/div><h4><a id=\"user-content--last-pred-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-last-pred-list\"><\/a>-last <code>(pred list)<\/code><\/h4><p>Return the last x in <code>list<\/code> where (<code>pred<\/code> x) is non-nil, else nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-last 'even? '(1 2 3 4 5 6 3 3 3)) ;; =&gt; 6\n(-last 'even? '(1 3 7 5 9)) ;; =&gt; nil\n(--last (&gt; (length it) 3) '(\"a\" \"looong\" \"word\" \"and\" \"short\" \"one\")) ;; =&gt; \"short\"<\/pre><\/div><h4><a id=\"user-content--first-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-first-item-list\"><\/a>-first-item <code>(list)<\/code><\/h4><p>Return the first item of <code>list<\/code>, or nil on an empty list.<\/p><p>See also: <a href=\"#-second-item-arg1\"><code>-second-item<\/code><\/a>, <a href=\"#-last-item-list\"><code>-last-item<\/code><\/a>.<\/p><p>(fn <code>list<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-first-item '(1 2 3)) ;; =&gt; 1\n(-first-item nil) ;; =&gt; nil\n(let ((list (list 1 2 3))) (setf (-first-item list) 5) list) ;; =&gt; '(5 2 3)<\/pre><\/div><h4><a id=\"user-content--second-item-arg1\" class=\"anchor\" aria-hidden=\"true\" href=\"#-second-item-arg1\"><\/a>-second-item <code>(arg1)<\/code><\/h4><p>Return the second item of <code>list<\/code>, or nil if <code>list<\/code> is too short.<\/p><p>See also: <a href=\"#-third-item-arg1\"><code>-third-item<\/code><\/a>.<\/p><p>(fn <code>list<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-second-item '(1 2 3)) ;; =&gt; 2\n(-second-item nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--third-item-arg1\" class=\"anchor\" aria-hidden=\"true\" href=\"#-third-item-arg1\"><\/a>-third-item <code>(arg1)<\/code><\/h4><p>Return the third item of <code>list<\/code>, or nil if <code>list<\/code> is too short.<\/p><p>See also: <a href=\"#-fourth-item-list\"><code>-fourth-item<\/code><\/a>.<\/p><p>(fn <code>list<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-third-item '(1 2 3)) ;; =&gt; 3\n(-third-item nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--fourth-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-fourth-item-list\"><\/a>-fourth-item <code>(list)<\/code><\/h4><p>Return the fourth item of <code>list<\/code>, or nil if <code>list<\/code> is too short.<\/p><p>See also: <a href=\"#-fifth-item-list\"><code>-fifth-item<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-fourth-item '(1 2 3 4)) ;; =&gt; 4\n(-fourth-item nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--fifth-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-fifth-item-list\"><\/a>-fifth-item <code>(list)<\/code><\/h4><p>Return the fifth item of <code>list<\/code>, or nil if <code>list<\/code> is too short.<\/p><p>See also: <a href=\"#-last-item-list\"><code>-last-item<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-fifth-item '(1 2 3 4 5)) ;; =&gt; 5\n(-fifth-item nil) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--last-item-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-last-item-list\"><\/a>-last-item <code>(list)<\/code><\/h4><p>Return the last item of <code>list<\/code>, or nil on an empty list.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-last-item '(1 2 3)) ;; =&gt; 3\n(-last-item nil) ;; =&gt; nil\n(let ((list (list 1 2 3))) (setf (-last-item list) 5) list) ;; =&gt; '(1 2 5)<\/pre><\/div><h4><a id=\"user-content--butlast-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-butlast-list\"><\/a>-butlast <code>(list)<\/code><\/h4><p>Return a list of all items in list except for the last.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-butlast '(1 2 3)) ;; =&gt; '(1 2)\n(-butlast '(1 2)) ;; =&gt; '(1)\n(-butlast '(1)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--sort-comparator-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-sort-comparator-list\"><\/a>-sort <code>(comparator list)<\/code><\/h4><p>Sort <code>list<\/code>, stably, comparing elements using <code>comparator<\/code>. Return the sorted list. <code>list<\/code> is <code>not<\/code> modified by side effects. <code>comparator<\/code> is called with two elements of <code>list<\/code>, and should return non-nil if the first element should sort before the second.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-sort '&lt; '(3 1 2)) ;; =&gt; '(1 2 3)\n(-sort '&gt; '(3 1 2)) ;; =&gt; '(3 2 1)\n(--sort (&lt; it other) '(3 1 2)) ;; =&gt; '(1 2 3)<\/pre><\/div><h4><a id=\"user-content--list-rest-args\" class=\"anchor\" aria-hidden=\"true\" href=\"#-list-rest-args\"><\/a>-list <code>(&amp;rest args)<\/code><\/h4><p>Return a list with <code>args<\/code>.<\/p><p>If first item of <code>args<\/code> is already a list, simply return <code>args<\/code>. If not, return a list with <code>args<\/code> as elements.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-list 1) ;; =&gt; '(1)\n(-list 1 2 3) ;; =&gt; '(1 2 3)\n(-list '(1 2 3)) ;; =&gt; '(1 2 3)<\/pre><\/div><h4><a id=\"user-content--fix-fn-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-fix-fn-list\"><\/a>-fix <code>(fn list)<\/code><\/h4><p>Compute the (least) fixpoint of <code>fn<\/code> with initial input <code>list<\/code>.<\/p><p><code>fn<\/code> is called at least once, results are compared with <code>equal<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-fix (lambda (l) (-non-nil (--mapcat (-split-at (\/ (length it) 2) it) l))) '((1 2 3 4 5 6))) ;; =&gt; '((1) (2) (3) (4) (5) (6))\n(let ((data '((\"starwars\" \"scifi\") (\"jedi\" \"starwars\" \"warrior\")))) (--fix (-uniq (--mapcat (cons it (cdr (assoc it data))) it)) '(\"jedi\" \"book\"))) ;; =&gt; '(\"jedi\" \"starwars\" \"warrior\" \"scifi\" \"book\")<\/pre><\/div><h2><a id=\"user-content-tree-operations-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#tree-operations-1\"><\/a>Tree operations<\/h2><p>Functions pretending lists are trees.<\/p><h4><a id=\"user-content--tree-seq-branch-children-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-seq-branch-children-tree\"><\/a>-tree-seq <code>(branch children tree)<\/code><\/h4><p>Return a sequence of the nodes in <code>tree<\/code>, in depth-first search order.<\/p><p><code>branch<\/code> is a predicate of one argument that returns non-nil if the passed argument is a branch, that is, a node that can have children.<\/p><p><code>children<\/code> is a function of one argument that returns the children of the passed branch node.<\/p><p>Non-branch nodes are simply copied.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-seq 'listp 'identity '(1 (2 3) 4 (5 (6 7)))) ;; =&gt; '((1 (2 3) 4 (5 (6 7))) 1 (2 3) 2 3 4 (5 (6 7)) 5 (6 7) 6 7)\n(-tree-seq 'listp 'reverse '(1 (2 3) 4 (5 (6 7)))) ;; =&gt; '((1 (2 3) 4 (5 (6 7))) (5 (6 7)) (6 7) 7 6 5 4 (2 3) 3 2 1)\n(--tree-seq (vectorp it) (append it nil) [1 [2 3] 4 [5 [6 7]]]) ;; =&gt; '([1 [2 3] 4 [5 [6 7]]] 1 [2 3] 2 3 4 [5 [6 7]] 5 [6 7] 6 7)<\/pre><\/div><h4><a id=\"user-content--tree-map-fn-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-map-fn-tree\"><\/a>-tree-map <code>(fn tree)<\/code><\/h4><p>Apply <code>fn<\/code> to each element of <code>tree<\/code> while preserving the tree structure.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-map '1+ '(1 (2 3) (4 (5 6) 7))) ;; =&gt; '(2 (3 4) (5 (6 7) 8))\n(-tree-map '(lambda (x) (cons x (expt 2 x))) '(1 (2 3) 4)) ;; =&gt; '((1 . 2) ((2 . 4) (3 . 8)) (4 . 16))\n(--tree-map (length it) '(\"&lt;body&gt;\" (\"&lt;p&gt;\" \"text\" \"&lt;\/p&gt;\") \"&lt;\/body&gt;\")) ;; =&gt; '(6 (3 4 4) 7)<\/pre><\/div><h4><a id=\"user-content--tree-map-nodes-pred-fun-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-map-nodes-pred-fun-tree\"><\/a>-tree-map-nodes <code>(pred fun tree)<\/code><\/h4><p>Call <code>fun<\/code> on each node of <code>tree<\/code> that satisfies <code>pred<\/code>.<\/p><p>If <code>pred<\/code> returns nil, continue descending down this node. If <code>pred<\/code> returns non-nil, apply <code>fun<\/code> to this node and do not descend further.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-map-nodes 'vectorp (lambda (x) (-sum (append x nil))) '(1 [2 3] 4 (5 [6 7] 8))) ;; =&gt; '(1 5 4 (5 13 8))\n(-tree-map-nodes 'keywordp (lambda (x) (symbol-name x)) '(1 :foo 4 ((5 6 :bar) :baz 8))) ;; =&gt; '(1 \":foo\" 4 ((5 6 \":bar\") \":baz\" 8))\n(--tree-map-nodes (eq (car-safe it) 'add-mode) (-concat it (list :mode 'emacs-lisp-mode)) '(with-mode emacs-lisp-mode (foo bar) (add-mode a b) (baz (add-mode c d)))) ;; =&gt; '(with-mode emacs-lisp-mode (foo bar) (add-mode a b :mode emacs-lisp-mode) (baz (add-mode c d :mode emacs-lisp-mode)))<\/pre><\/div><h4><a id=\"user-content--tree-reduce-fn-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-reduce-fn-tree\"><\/a>-tree-reduce <code>(fn tree)<\/code><\/h4><p>Use <code>fn<\/code> to reduce elements of list <code>tree<\/code>. If elements of <code>tree<\/code> are lists themselves, apply the reduction recursively.<\/p><p><code>fn<\/code> is first applied to first element of the list and second element, then on this result and third element from the list etc.<\/p><p>See <a href=\"#-reduce-r-fn-list\"><code>-reduce-r<\/code><\/a> for how exactly are lists of zero or one element handled.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-reduce '+ '(1 (2 3) (4 5))) ;; =&gt; 15\n(-tree-reduce 'concat '(\"strings\" (\" on\" \" various\") ((\" levels\")))) ;; =&gt; \"strings on various levels\"\n(--tree-reduce (cond ((stringp it) (concat it \" \" acc)) (t (let ((sn (symbol-name it))) (concat \"&lt;\" sn \"&gt;\" acc \"&lt;\/\" sn \"&gt;\")))) '(body (p \"some words\") (div \"more\" (b \"bold\") \"words\"))) ;; =&gt; \"&lt;body&gt;&lt;p&gt;some words&lt;\/p&gt; &lt;div&gt;more &lt;b&gt;bold&lt;\/b&gt; words&lt;\/div&gt;&lt;\/body&gt;\"<\/pre><\/div><h4><a id=\"user-content--tree-reduce-from-fn-init-value-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-reduce-from-fn-init-value-tree\"><\/a>-tree-reduce-from <code>(fn init-value tree)<\/code><\/h4><p>Use <code>fn<\/code> to reduce elements of list <code>tree<\/code>. If elements of <code>tree<\/code> are lists themselves, apply the reduction recursively.<\/p><p><code>fn<\/code> is first applied to <code>init-value<\/code> and first element of the list, then on this result and second element from the list etc.<\/p><p>The initial value is ignored on cons pairs as they always contain two elements.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-reduce-from '+ 1 '(1 (1 1) ((1)))) ;; =&gt; 8\n(--tree-reduce-from (-concat acc (list it)) nil '(1 (2 3 (4 5)) (6 7))) ;; =&gt; '((7 6) ((5 4) 3 2) 1)<\/pre><\/div><h4><a id=\"user-content--tree-mapreduce-fn-folder-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-mapreduce-fn-folder-tree\"><\/a>-tree-mapreduce <code>(fn folder tree)<\/code><\/h4><p>Apply <code>fn<\/code> to each element of <code>tree<\/code>, and make a list of the results. If elements of <code>tree<\/code> are lists themselves, apply <code>fn<\/code> recursively to elements of these nested lists.<\/p><p>Then reduce the resulting lists using <code>folder<\/code> and initial value <code>init-value<\/code>. See <a href=\"#-reduce-r-from-fn-initial-value-list\"><code>-reduce-r-from<\/code><\/a>.<\/p><p>This is the same as calling <a href=\"#-tree-reduce-fn-tree\"><code>-tree-reduce<\/code><\/a> after <a href=\"#-tree-map-fn-tree\"><code>-tree-map<\/code><\/a> but is twice as fast as it only traverse the structure once.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-mapreduce 'list 'append '(1 (2 (3 4) (5 6)) (7 (8 9)))) ;; =&gt; '(1 2 3 4 5 6 7 8 9)\n(--tree-mapreduce 1 (+ it acc) '(1 (2 (4 9) (2 1)) (7 (4 3)))) ;; =&gt; 9\n(--tree-mapreduce 0 (max acc (1+ it)) '(1 (2 (4 9) (2 1)) (7 (4 3)))) ;; =&gt; 3<\/pre><\/div><h4><a id=\"user-content--tree-mapreduce-from-fn-folder-init-value-tree\" class=\"anchor\" aria-hidden=\"true\" href=\"#-tree-mapreduce-from-fn-folder-init-value-tree\"><\/a>-tree-mapreduce-from <code>(fn folder init-value tree)<\/code><\/h4><p>Apply <code>fn<\/code> to each element of <code>tree<\/code>, and make a list of the results. If elements of <code>tree<\/code> are lists themselves, apply <code>fn<\/code> recursively to elements of these nested lists.<\/p><p>Then reduce the resulting lists using <code>folder<\/code> and initial value <code>init-value<\/code>. See <a href=\"#-reduce-r-from-fn-initial-value-list\"><code>-reduce-r-from<\/code><\/a>.<\/p><p>This is the same as calling <a href=\"#-tree-reduce-from-fn-init-value-tree\"><code>-tree-reduce-from<\/code><\/a> after <a href=\"#-tree-map-fn-tree\"><code>-tree-map<\/code><\/a> but is twice as fast as it only traverse the structure once.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-tree-mapreduce-from 'identity '* 1 '(1 (2 (3 4) (5 6)) (7 (8 9)))) ;; =&gt; 362880\n(--tree-mapreduce-from (+ it it) (cons it acc) nil '(1 (2 (4 9) (2 1)) (7 (4 3)))) ;; =&gt; '(2 (4 (8 18) (4 2)) (14 (8 6)))\n(concat \"{\" (--tree-mapreduce-from (cond ((-cons-pair? it) (concat (symbol-name (car it)) \" -&gt; \" (symbol-name (cdr it)))) (t (concat (symbol-name it) \" : {\"))) (concat it (unless (or (equal acc \"}\") (equal (substring it (1- (length it))) \"{\")) \", \") acc) \"}\" '((elips-mode (foo (bar . booze)) (baz . qux)) (c-mode (foo . bla) (bum . bam))))) ;; =&gt; \"{elips-mode : {foo : {bar -&gt; booze}, baz -&gt; qux}, c-mode : {foo -&gt; bla, bum -&gt; bam}}\"<\/pre><\/div><h4><a id=\"user-content--clone-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#-clone-list\"><\/a>-clone <code>(list)<\/code><\/h4><p>Create a deep copy of <code>list<\/code>. The new list has the same elements and structure but all cons are replaced with new ones. This is useful when you need to clone a structure such as plist or alist.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let* ((a '(1 2 3)) (b (-clone a))) (nreverse a) b) ;; =&gt; '(1 2 3)<\/pre><\/div><h2><a id=\"user-content-threading-macros-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#threading-macros-1\"><\/a>Threading macros<\/h2><h4><a id=\"user-content---x-optional-form-rest-more\" class=\"anchor\" aria-hidden=\"true\" href=\"#--x-optional-form-rest-more\"><\/a>-&gt; <code>(x &amp;optional form &amp;rest more)<\/code><\/h4><p>Thread the expr through the forms. Insert <code>x<\/code> as the second item in the first form, making a list of it if it is not a list already. If there are more forms, insert the first form as the second item in second form, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt; '(2 3 5)) ;; =&gt; '(2 3 5)\n(-&gt; '(2 3 5) (append '(8 13))) ;; =&gt; '(2 3 5 8 13)\n(-&gt; '(2 3 5) (append '(8 13)) (-slice 1 -1)) ;; =&gt; '(3 5 8)<\/pre><\/div><h4><a id=\"user-content---x-optional-form-rest-more-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#--x-optional-form-rest-more-1\"><\/a>-&gt;&gt; <code>(x &amp;optional form &amp;rest more)<\/code><\/h4><p>Thread the expr through the forms. Insert <code>x<\/code> as the last item in the first form, making a list of it if it is not a list already. If there are more forms, insert the first form as the last item in second form, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-&gt;&gt; '(1 2 3) (-map 'square)) ;; =&gt; '(1 4 9)\n(-&gt;&gt; '(1 2 3) (-map 'square) (-remove 'even?)) ;; =&gt; '(1 9)\n(-&gt;&gt; '(1 2 3) (-map 'square) (-reduce '+)) ;; =&gt; 14<\/pre><\/div><h4><a id=\"user-content----x-rest-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#---x-rest-forms\"><\/a>--&gt; <code>(x &amp;rest forms)<\/code><\/h4><p>Starting with the value of <code>x<\/code>, thread each expression through <code>forms<\/code>.<\/p><p>Insert <code>x<\/code> at the position signified by the symbol <code>it<\/code> in the first form. If there are more forms, insert the first form at the position signified by <code>it<\/code> in in second form, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(--&gt; \"def\" (concat \"abc\" it \"ghi\")) ;; =&gt; \"abcdefghi\"\n(--&gt; \"def\" (concat \"abc\" it \"ghi\") (upcase it)) ;; =&gt; \"ABCDEFGHI\"\n(--&gt; \"def\" (concat \"abc\" it \"ghi\") upcase) ;; =&gt; \"ABCDEFGHI\"<\/pre><\/div><h4><a id=\"user-content--as--value-variable-rest-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#-as--value-variable-rest-forms\"><\/a>-as-&gt; <code>(value variable &amp;rest forms)<\/code><\/h4><p>Starting with <code>value<\/code>, thread <code>variable<\/code> through <code>forms<\/code>.<\/p><p>In the first form, bind <code>variable<\/code> to <code>value<\/code>. In the second form, bind <code>variable<\/code> to the result of the first form, and so forth.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-as-&gt; 3 my-var (1+ my-var) (list my-var) (mapcar (lambda (ele) (* 2 ele)) my-var)) ;; =&gt; '(8)\n(-as-&gt; 3 my-var 1+) ;; =&gt; 4\n(-as-&gt; 3 my-var) ;; =&gt; 3<\/pre><\/div><h4><a id=\"user-content--some--x-optional-form-rest-more\" class=\"anchor\" aria-hidden=\"true\" href=\"#-some--x-optional-form-rest-more\"><\/a>-some-&gt; <code>(x &amp;optional form &amp;rest more)<\/code><\/h4><p>When expr is non-nil, thread it through the first form (via <a href=\"#--x-optional-form-rest-more\"><code>-&gt;<\/code><\/a>), and when that result is non-nil, through the next form, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-some-&gt; '(2 3 5)) ;; =&gt; '(2 3 5)\n(-some-&gt; 5 square) ;; =&gt; 25\n(-some-&gt; 5 even? square) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--some--x-optional-form-rest-more-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#-some--x-optional-form-rest-more-1\"><\/a>-some-&gt;&gt; <code>(x &amp;optional form &amp;rest more)<\/code><\/h4><p>When expr is non-nil, thread it through the first form (via <a href=\"#--x-optional-form-rest-more\"><code>-&gt;&gt;<\/code><\/a>), and when that result is non-nil, through the next form, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-some-&gt;&gt; '(1 2 3) (-map 'square)) ;; =&gt; '(1 4 9)\n(-some-&gt;&gt; '(1 3 5) (-last 'even?) (+ 100)) ;; =&gt; nil\n(-some-&gt;&gt; '(2 4 6) (-last 'even?) (+ 100)) ;; =&gt; 106<\/pre><\/div><h4><a id=\"user-content--some---x-optional-form-rest-more\" class=\"anchor\" aria-hidden=\"true\" href=\"#-some---x-optional-form-rest-more\"><\/a>-some--&gt; <code>(x &amp;optional form &amp;rest more)<\/code><\/h4><p>When expr in non-nil, thread it through the first form (via <a href=\"#---x-rest-forms\"><code>--&gt;<\/code><\/a>), and when that result is non-nil, through the next form, etc.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-some--&gt; \"def\" (concat \"abc\" it \"ghi\")) ;; =&gt; \"abcdefghi\"\n(-some--&gt; nil (concat \"abc\" it \"ghi\")) ;; =&gt; nil\n(-some--&gt; '(1 3 5) (-filter 'even? it) (append it it) (-map 'square it)) ;; =&gt; nil<\/pre><\/div><h2><a id=\"user-content-binding-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#binding-1\"><\/a>Binding<\/h2><p>Convenient versions of <code>let<\/code> and <code>let*<\/code> constructs combined with flow control.<\/p><h4><a id=\"user-content--when-let-var-val-rest-body\" class=\"anchor\" aria-hidden=\"true\" href=\"#-when-let-var-val-rest-body\"><\/a>-when-let <code>(var-val &amp;rest body)<\/code><\/h4><p>If <code>val<\/code> evaluates to non-nil, bind it to <code>var<\/code> and execute body.<\/p><p>Note: binding is done according to <a href=\"#-let-varlist-rest-body\"><code>-let<\/code><\/a>.<\/p><p>(fn (<code>var<\/code> <code>val<\/code>) &amp;rest <code>body<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-when-let (match-index (string-match \"d\" \"abcd\")) (+ match-index 2)) ;; =&gt; 5\n(-when-let ((&amp;plist :foo foo) (list :foo \"foo\")) foo) ;; =&gt; \"foo\"\n(-when-let ((&amp;plist :foo foo) (list :bar \"bar\")) foo) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--when-let-vars-vals-rest-body\" class=\"anchor\" aria-hidden=\"true\" href=\"#-when-let-vars-vals-rest-body\"><\/a>-when-let* <code>(vars-vals &amp;rest body)<\/code><\/h4><p>If all <code>vals<\/code> evaluate to true, bind them to their corresponding <code>vars<\/code> and execute body. <code>vars-vals<\/code> should be a list of (<code>var<\/code> <code>val<\/code>) pairs.<\/p><p>Note: binding is done according to <a href=\"#-let-varlist-rest-body\"><code>-let*<\/code><\/a>. <code>vals<\/code> are evaluated sequentially, and evaluation stops after the first nil <code>val<\/code> is encountered.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-when-let* ((x 5) (y 3) (z (+ y 4))) (+ x y z)) ;; =&gt; 15\n(-when-let* ((x 5) (y nil) (z 7)) (+ x y z)) ;; =&gt; nil<\/pre><\/div><h4><a id=\"user-content--if-let-var-val-then-rest-else\" class=\"anchor\" aria-hidden=\"true\" href=\"#-if-let-var-val-then-rest-else\"><\/a>-if-let <code>(var-val then &amp;rest else)<\/code><\/h4><p>If <code>val<\/code> evaluates to non-nil, bind it to <code>var<\/code> and do <code>then<\/code>, otherwise do <code>else<\/code>.<\/p><p>Note: binding is done according to <a href=\"#-let-varlist-rest-body\"><code>-let<\/code><\/a>.<\/p><p>(fn (<code>var<\/code> <code>val<\/code>) <code>then<\/code> &amp;rest <code>else<\/code>)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-if-let (match-index (string-match \"d\" \"abc\")) (+ match-index 3) 7) ;; =&gt; 7\n(--if-let (even? 4) it nil) ;; =&gt; t<\/pre><\/div><h4><a id=\"user-content--if-let-vars-vals-then-rest-else\" class=\"anchor\" aria-hidden=\"true\" href=\"#-if-let-vars-vals-then-rest-else\"><\/a>-if-let* <code>(vars-vals then &amp;rest else)<\/code><\/h4><p>If all <code>vals<\/code> evaluate to true, bind them to their corresponding <code>vars<\/code> and do <code>then<\/code>, otherwise do <code>else<\/code>. <code>vars-vals<\/code> should be a list of (<code>var<\/code> <code>val<\/code>) pairs.<\/p><p>Note: binding is done according to <a href=\"#-let-varlist-rest-body\"><code>-let*<\/code><\/a>. <code>vals<\/code> are evaluated sequentially, and evaluation stops after the first nil <code>val<\/code> is encountered.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-if-let* ((x 5) (y 3) (z 7)) (+ x y z) \"foo\") ;; =&gt; 15\n(-if-let* ((x 5) (y nil) (z 7)) (+ x y z) \"foo\") ;; =&gt; \"foo\"\n(-if-let* (((_ _ x) '(nil nil 7))) x) ;; =&gt; 7<\/pre><\/div><h4><a id=\"user-content--let-varlist-rest-body\" class=\"anchor\" aria-hidden=\"true\" href=\"#-let-varlist-rest-body\"><\/a>-let <code>(varlist &amp;rest body)<\/code><\/h4><p>Bind variables according to <code>varlist<\/code> then eval <code>body<\/code>.<\/p><p><code>varlist<\/code> is a list of lists of the form (<code>pattern<\/code> <code>source<\/code>). Each <code>pattern<\/code> is matched against the <code>source<\/code> \"structurally\". <code>source<\/code> is only evaluated once for each <code>pattern<\/code>. Each <code>pattern<\/code> is matched recursively, and can therefore contain sub-patterns which are matched against corresponding sub-expressions of <code>source<\/code>.<\/p><p>All the SOURCEs are evalled before any symbols are bound (i.e. \"in parallel\").<\/p><p>If <code>varlist<\/code> only contains one (<code>pattern<\/code> <code>source<\/code>) element, you can optionally specify it using a vector and discarding the outer-most parens. Thus<\/p><pre>(-let ((`pattern` `source`)) ..)\n<\/pre><p>becomes<\/p><pre>(-let [`pattern` `source`] ..).\n<\/pre><p><a href=\"#-let-varlist-rest-body\"><code>-let<\/code><\/a> uses a convention of not binding places (symbols) starting with _ whenever it's possible. You can use this to skip over entries you don't care about. However, this is not <em>always<\/em> possible (as a result of implementation) and these symbols might get bound to undefined values.<\/p><p>Following is the overview of supported patterns. Remember that patterns can be matched recursively, so every a, b, aK in the following can be a matching construct and not necessarily a symbol\/variable.<\/p><p>Symbol:<\/p><pre>a - bind the `source` to `a`.  This is just like regular `let`.\n<\/pre><p>Conses and lists:<\/p><pre>(a) - bind `car` of cons\/list to `a`\n(a . b) - bind car of cons to `a` and `cdr` to `b`\n(a b) - bind car of list to `a` and `cadr` to `b`\n(a1 a2 a3  ...) - bind 0th car of list to `a1`, 1st to `a2`, 2nd to `a3` ...\n(a1 a2 a3 ... aN . rest) - as above, but bind the Nth cdr to `rest`.\n<\/pre><p>Vectors:<\/p><pre>[a] - bind 0th element of a non-list sequence to `a` (works with\n      vectors, strings, bit arrays...)\n[a1 a2 a3 ...] - bind 0th element of non-list sequence to `a0`, 1st to\n                 `a1`, 2nd to `a2`, ...\n                 If the `pattern` is shorter than `source`, the values at\n                 places not in `pattern` are ignored.\n                 If the `pattern` is longer than `source`, an `error` is\n                 thrown.\n[a1 a2 a3 ... &amp;rest rest] - as above, but bind the rest of\n                            the sequence to `rest`.  This is\n                            conceptually the same as improper list\n                            matching (a1 a2 ... aN . rest)\n<\/pre><p>Key\/value stores:<\/p><pre>(&amp;plist key0 a0 ... keyN aN) - bind value mapped by keyK in the\n                               `source` plist to aK.  If the\n                               value is not found, aK is nil.\n                               Uses `plist-get` to fetch values.\n(&amp;alist key0 a0 ... keyN aN) - bind value mapped by keyK in the\n                               `source` alist to aK.  If the\n                               value is not found, aK is nil.\n                               Uses `assoc` to fetch values.\n(&amp;hash key0 a0 ... keyN aN) - bind value mapped by keyK in the\n                              `source` hash table to aK.  If the\n                              value is not found, aK is nil.\n                              Uses `gethash` to fetch values.\n<\/pre><p>Further, special keyword &amp;keys supports \"inline\" matching of plist-like key-value pairs, similarly to &amp;keys keyword of <code>cl-defun<\/code>.<\/p><pre>(a1 a2 ... aN &amp;keys key1 b1 ... keyN bK)\n<\/pre><p>This binds <code>n<\/code> values from the list to a1 ... aN, then interprets the cdr as a plist (see key\/value matching above).<\/p><p><code>a<\/code> shorthand notation for kv-destructuring exists which allows the patterns be optionally left out and derived from the key name in the following fashion:<\/p><ul><li>a key :foo is converted into <code>foo<\/code> pattern,<\/li>\n<li>a key 'bar is converted into <code>bar<\/code> pattern,<\/li>\n<li>a key \"baz\" is converted into <code>baz<\/code> pattern.<\/li>\n<\/ul><p>That is, the entire value under the key is bound to the derived variable without any further destructuring.<\/p><p>This is possible only when the form following the key is not a valid pattern (i.e. not a symbol, a cons cell or a vector). Otherwise the matching proceeds as usual and in case of an invalid spec fails with an error.<\/p><p>Thus the patterns are normalized as follows:<\/p><pre> ;; derive all the missing patterns\n (&amp;plist :foo 'bar \"baz\") =&gt; (&amp;plist :foo foo 'bar bar \"baz\" baz)\n ;; we can specify some but not others\n (&amp;plist :foo 'bar explicit-bar) =&gt; (&amp;plist :foo foo 'bar explicit-bar)\n ;; nothing happens, we store :foo in x\n (&amp;plist :foo x) =&gt; (&amp;plist :foo x)\n ;; nothing happens, we match recursively\n (&amp;plist :foo (a b c)) =&gt; (&amp;plist :foo (a b c))\n<\/pre><p>You can name the source using the syntax <code>symbol<\/code> &amp;as <code>pattern<\/code>. This syntax works with lists (proper or improper), vectors and all types of maps.<\/p><pre>(list &amp;as a b c) (list 1 2 3)\n<\/pre><p>binds <code>a<\/code> to 1, <code>b<\/code> to 2, <code>c<\/code> to 3 and <code>list<\/code> to (1 2 3).<\/p><p>Similarly:<\/p><pre>(bounds &amp;as beg . end) (cons 1 2)\n<\/pre><p>binds <code>beg<\/code> to 1, <code>end<\/code> to 2 and <code>bounds<\/code> to (1 . 2).<\/p><pre>(items &amp;as first . rest) (list 1 2 3)\n<\/pre><p>binds <code>first<\/code> to 1, <code>rest<\/code> to (2 3) and <code>items<\/code> to (1 2 3)<\/p><pre>[vect &amp;as _ b c] [1 2 3]\n<\/pre><p>binds <code>b<\/code> to 2, <code>c<\/code> to 3 and <code>vect<\/code> to [1 2 3] (_ avoids binding as usual).<\/p><pre>(plist &amp;as &amp;plist :b b) (list :a 1 :b 2 :c 3)\n<\/pre><p>binds <code>b<\/code> to 2 and <code>plist<\/code> to (:a 1 :b 2 :c 3). Same for &amp;alist and &amp;hash.<\/p><p>This is especially useful when we want to capture the result of a computation and destructure at the same time. Consider the form (function-returning-complex-structure) returning a list of two vectors with two items each. We want to capture this entire result and pass it to another computation, but at the same time we want to get the second item from each vector. We can achieve it with pattern<\/p><pre>(result &amp;as [_ a] [_ b]) (function-returning-complex-structure)\n<\/pre><p>Note: Clojure programmers may know this feature as the \":as binding\". The difference is that we put the &amp;as at the front because we need to support improper list binding.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-let (([a (b c) d] [1 (2 3) 4])) (list a b c d)) ;; =&gt; '(1 2 3 4)\n(-let [(a b c . d) (list 1 2 3 4 5 6)] (list a b c d)) ;; =&gt; '(1 2 3 (4 5 6))\n(-let [(&amp;plist :foo foo :bar bar) (list :baz 3 :foo 1 :qux 4 :bar 2)] (list foo bar)) ;; =&gt; '(1 2)<\/pre><\/div><h4><a id=\"user-content--let-varlist-rest-body-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#-let-varlist-rest-body-1\"><\/a>-let* <code>(varlist &amp;rest body)<\/code><\/h4><p>Bind variables according to <code>varlist<\/code> then eval <code>body<\/code>.<\/p><p><code>varlist<\/code> is a list of lists of the form (<code>pattern<\/code> <code>source<\/code>). Each <code>pattern<\/code> is matched against the <code>source<\/code> structurally. <code>source<\/code> is only evaluated once for each <code>pattern<\/code>.<\/p><p>Each <code>source<\/code> can refer to the symbols already bound by this <code>varlist<\/code>. This is useful if you want to destructure <code>source<\/code> recursively but also want to name the intermediate structures.<\/p><p>See <a href=\"#-let-varlist-rest-body\"><code>-let<\/code><\/a> for the list of all possible patterns.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-let* (((a . b) (cons 1 2)) ((c . d) (cons 3 4))) (list a b c d)) ;; =&gt; '(1 2 3 4)\n(-let* (((a . b) (cons 1 (cons 2 3))) ((c . d) b)) (list a b c d)) ;; =&gt; '(1 (2 . 3) 2 3)\n(-let* (((&amp;alist \"foo\" foo \"bar\" bar) (list (cons \"foo\" 1) (cons \"bar\" (list 'a 'b 'c)))) ((a b c) bar)) (list foo a b c bar)) ;; =&gt; '(1 a b c (a b c))<\/pre><\/div><h4><a id=\"user-content--lambda-match-form-rest-body\" class=\"anchor\" aria-hidden=\"true\" href=\"#-lambda-match-form-rest-body\"><\/a>-lambda <code>(match-form &amp;rest body)<\/code><\/h4><p>Return a lambda which destructures its input as <code>match-form<\/code> and executes <code>body<\/code>.<\/p><p>Note that you have to enclose the <code>match-form<\/code> in a pair of parens, such that:<\/p><pre>(-lambda (x) body)\n(-lambda (x y ...) body)\n<\/pre><p>has the usual semantics of <code>lambda<\/code>. Furthermore, these get translated into normal lambda, so there is no performance penalty.<\/p><p>See <a href=\"#-let-varlist-rest-body\"><code>-let<\/code><\/a> for the description of destructuring mechanism.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map (-lambda ((x y)) (+ x y)) '((1 2) (3 4) (5 6))) ;; =&gt; '(3 7 11)\n(-map (-lambda ([x y]) (+ x y)) '([1 2] [3 4] [5 6])) ;; =&gt; '(3 7 11)\n(funcall (-lambda ((_ . a) (_ . b)) (-concat a b)) '(1 2 3) '(4 5 6)) ;; =&gt; '(2 3 5 6)<\/pre><\/div><h4><a id=\"user-content--setq-rest-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#-setq-rest-forms\"><\/a>-setq <code>(&amp;rest forms)<\/code><\/h4><p>Bind each <code>match-form<\/code> to the value of its <code>val<\/code>.<\/p><p><code>match-form<\/code> destructuring is done according to the rules of <a href=\"#-let-varlist-rest-body\"><code>-let<\/code><\/a>.<\/p><p>This macro allows you to bind multiple variables by destructuring the value, so for example:<\/p><pre>(-setq (a b) x\n       (&amp;plist :c c) plist)\n<\/pre><p>expands roughly speaking to the following code<\/p><pre>(setq a (car x)\n      b (cadr x)\n      c (plist-get plist :c))\n<\/pre><p>Care is taken to only evaluate each <code>val<\/code> once so that in case of multiple assignments it does not cause unexpected side effects.<\/p><p>(fn [<code>match-form<\/code> <code>val<\/code>]...)<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(progn (-setq a 1) a) ;; =&gt; 1\n(progn (-setq (a b) (list 1 2)) (list a b)) ;; =&gt; '(1 2)\n(progn (-setq (&amp;plist :c c) (list :c \"c\")) c) ;; =&gt; \"c\"<\/pre><\/div><h2><a id=\"user-content-side-effects-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#side-effects-1\"><\/a>Side-effects<\/h2><p>Functions iterating over lists for side-effect only.<\/p><h4><a id=\"user-content--each-list-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-each-list-fn\"><\/a>-each <code>(list fn)<\/code><\/h4><p>Call <code>fn<\/code> with every item in <code>list<\/code>. Return nil, used for side-effects only.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (s) (-each '(1 2 3) (lambda (item) (setq s (cons item s))))) ;; =&gt; nil\n(let (s) (-each '(1 2 3) (lambda (item) (setq s (cons item s)))) s) ;; =&gt; '(3 2 1)\n(let (s) (--each '(1 2 3) (setq s (cons it s))) s) ;; =&gt; '(3 2 1)<\/pre><\/div><h4><a id=\"user-content--each-while-list-pred-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-each-while-list-pred-fn\"><\/a>-each-while <code>(list pred fn)<\/code><\/h4><p>Call <code>fn<\/code> with every item in <code>list<\/code> while (<code>pred<\/code> item) is non-nil. Return nil, used for side-effects only.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (s) (-each-while '(2 4 5 6) 'even? (lambda (item) (!cons item s))) s) ;; =&gt; '(4 2)\n(let (s) (--each-while '(1 2 3 4) (&lt; it 3) (!cons it s)) s) ;; =&gt; '(2 1)<\/pre><\/div><h4><a id=\"user-content--each-indexed-list-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-each-indexed-list-fn\"><\/a>-each-indexed <code>(list fn)<\/code><\/h4><p>Call (<code>fn<\/code> index item) for each item in <code>list<\/code>.<\/p><p>In the anaphoric form <code>--each-indexed<\/code>, the index is exposed as symbol <code>it-index<\/code>.<\/p><p>See also: <a href=\"#-map-indexed-fn-list\"><code>-map-indexed<\/code><\/a>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (s) (-each-indexed '(a b c) (lambda (index item) (setq s (cons (list item index) s)))) s) ;; =&gt; '((c 2) (b 1) (a 0))\n(let (s) (--each-indexed '(a b c) (setq s (cons (list it it-index) s))) s) ;; =&gt; '((c 2) (b 1) (a 0))<\/pre><\/div><h4><a id=\"user-content--each-r-list-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-each-r-list-fn\"><\/a>-each-r <code>(list fn)<\/code><\/h4><p>Call <code>fn<\/code> with every item in <code>list<\/code> in reversed order. Return nil, used for side-effects only.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (s) (-each-r '(1 2 3) (lambda (item) (setq s (cons item s))))) ;; =&gt; nil\n(let (s) (-each-r '(1 2 3) (lambda (item) (setq s (cons item s)))) s) ;; =&gt; '(1 2 3)\n(let (s) (--each-r '(1 2 3) (setq s (cons it s))) s) ;; =&gt; '(1 2 3)<\/pre><\/div><h4><a id=\"user-content--each-r-while-list-pred-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-each-r-while-list-pred-fn\"><\/a>-each-r-while <code>(list pred fn)<\/code><\/h4><p>Call <code>fn<\/code> with every item in reversed <code>list<\/code> while (<code>pred<\/code> item) is non-nil. Return nil, used for side-effects only.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (s) (-each-r-while '(2 4 5 6) 'even? (lambda (item) (!cons item s))) s) ;; =&gt; '(6)\n(let (s) (--each-r-while '(1 2 3 4) (&gt;= it 3) (!cons it s)) s) ;; =&gt; '(3 4)<\/pre><\/div><h4><a id=\"user-content--dotimes-num-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-dotimes-num-fn\"><\/a>-dotimes <code>(num fn)<\/code><\/h4><p>Repeatedly calls <code>fn<\/code> (presumably for side-effects) passing in integers from 0 through <code>num-1<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (s) (-dotimes 3 (lambda (n) (!cons n s))) s) ;; =&gt; '(2 1 0)\n(let (s) (--dotimes 5 (!cons it s)) s) ;; =&gt; '(4 3 2 1 0)<\/pre><\/div><h4><a id=\"user-content--doto-eval-initial-value-rest-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#-doto-eval-initial-value-rest-forms\"><\/a>-doto <code>(eval-initial-value &amp;rest forms)<\/code><\/h4><p>Eval a form, then insert that form as the 2nd argument to other forms. The <code>eval-initial-value<\/code> form is evaluated once. Its result is passed to <code>forms<\/code>, which are then evaluated sequentially. Returns the target form.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-doto '(1 2 3) (!cdr) (!cdr)) ;; =&gt; '(3)\n(-doto '(1 . 2) (setcar 3) (setcdr 4)) ;; =&gt; '(3 . 4)<\/pre><\/div><h4><a id=\"user-content---doto-eval-initial-value-rest-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#--doto-eval-initial-value-rest-forms\"><\/a>--doto <code>(eval-initial-value &amp;rest forms)<\/code><\/h4><p>Anaphoric form of <a href=\"#-doto-eval-initial-value-rest-forms\"><code>-doto<\/code><\/a>. Note: <code>it<\/code> is not required in each form.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(gethash \"key\" (--doto (make-hash-table :test 'equal) (puthash \"key\" \"value\" it))) ;; =&gt; \"value\"<\/pre><\/div><h2><a id=\"user-content-destructive-operations-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#destructive-operations-1\"><\/a>Destructive operations<\/h2><h4><a id=\"user-content-cons-car-cdr\" class=\"anchor\" aria-hidden=\"true\" href=\"#cons-car-cdr\"><\/a>!cons <code>(car cdr)<\/code><\/h4><p>Destructive: Set <code>cdr<\/code> to the cons of <code>car<\/code> and <code>cdr<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let (l) (!cons 5 l) l) ;; =&gt; '(5)\n(let ((l '(3))) (!cons 5 l) l) ;; =&gt; '(5 3)<\/pre><\/div><h4><a id=\"user-content-cdr-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#cdr-list\"><\/a>!cdr <code>(list)<\/code><\/h4><p>Destructive: Set <code>list<\/code> to the cdr of <code>list<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(let ((l '(3))) (!cdr l) l) ;; =&gt; '()\n(let ((l '(3 5))) (!cdr l) l) ;; =&gt; '(5)<\/pre><\/div><h2><a id=\"user-content-function-combinators-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#function-combinators-1\"><\/a>Function combinators<\/h2><p>These combinators require Emacs 24 for its lexical scope. So they are offered in a separate package: <code>dash-functional<\/code>.<\/p><h4><a id=\"user-content--partial-fn-rest-args\" class=\"anchor\" aria-hidden=\"true\" href=\"#-partial-fn-rest-args\"><\/a>-partial <code>(fn &amp;rest args)<\/code><\/h4><p>Takes a function <code>fn<\/code> and fewer than the normal arguments to <code>fn<\/code>, and returns a fn that takes a variable number of additional <code>args<\/code>. When called, the returned function calls <code>fn<\/code> with <code>args<\/code> first and then additional args.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-partial '- 5) 3) ;; =&gt; 2\n(funcall (-partial '+ 5 2) 3) ;; =&gt; 10<\/pre><\/div><h4><a id=\"user-content--rpartial-fn-rest-args\" class=\"anchor\" aria-hidden=\"true\" href=\"#-rpartial-fn-rest-args\"><\/a>-rpartial <code>(fn &amp;rest args)<\/code><\/h4><p>Takes a function <code>fn<\/code> and fewer than the normal arguments to <code>fn<\/code>, and returns a fn that takes a variable number of additional <code>args<\/code>. When called, the returned function calls <code>fn<\/code> with the additional args first and then <code>args<\/code>.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-rpartial '- 5) 8) ;; =&gt; 3\n(funcall (-rpartial '- 5 2) 10) ;; =&gt; 3<\/pre><\/div><h4><a id=\"user-content--juxt-rest-fns\" class=\"anchor\" aria-hidden=\"true\" href=\"#-juxt-rest-fns\"><\/a>-juxt <code>(&amp;rest fns)<\/code><\/h4><p>Takes a list of functions and returns a fn that is the juxtaposition of those fns. The returned fn takes a variable number of args, and returns a list containing the result of applying each fn to the args (left-to-right).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-juxt '+ '-) 3 5) ;; =&gt; '(8 -2)\n(-map (-juxt 'identity 'square) '(1 2 3)) ;; =&gt; '((1 1) (2 4) (3 9))<\/pre><\/div><h4><a id=\"user-content--compose-rest-fns\" class=\"anchor\" aria-hidden=\"true\" href=\"#-compose-rest-fns\"><\/a>-compose <code>(&amp;rest fns)<\/code><\/h4><p>Takes a list of functions and returns a fn that is the composition of those fns. The returned fn takes a variable number of arguments, and returns the result of applying each fn to the result of applying the previous fn to the arguments (right-to-left).<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-compose 'square '+) 2 3) ;; =&gt; (square (+ 2 3))\n(funcall (-compose 'identity 'square) 3) ;; =&gt; (square 3)\n(funcall (-compose 'square 'identity) 3) ;; =&gt; (square 3)<\/pre><\/div><h4><a id=\"user-content--applify-fn\" class=\"anchor\" aria-hidden=\"true\" href=\"#-applify-fn\"><\/a>-applify <code>(fn)<\/code><\/h4><p>Changes an n-arity function <code>fn<\/code> to a 1-arity function that expects a list with n items as arguments<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-map (-applify '+) '((1 1 1) (1 2 3) (5 5 5))) ;; =&gt; '(3 6 15)\n(-map (-applify (lambda (a b c) `(,a (,b (,c))))) '((1 1 1) (1 2 3) (5 5 5))) ;; =&gt; '((1 (1 (1))) (1 (2 (3))) (5 (5 (5))))\n(funcall (-applify '&lt;) '(3 6)) ;; =&gt; t<\/pre><\/div><h4><a id=\"user-content--on-operator-transformer\" class=\"anchor\" aria-hidden=\"true\" href=\"#-on-operator-transformer\"><\/a>-on <code>(operator transformer)<\/code><\/h4><p>Return a function of two arguments that first applies <code>transformer<\/code> to each of them and then applies <code>operator<\/code> on the results (in the same order).<\/p><p>In types: (b -&gt; b -&gt; c) -&gt; (a -&gt; b) -&gt; a -&gt; a -&gt; c<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-sort (-on '&lt; 'length) '((1 2 3) (1) (1 2))) ;; =&gt; '((1) (1 2) (1 2 3))\n(-min-by (-on '&gt; 'length) '((1 2 3) (4) (1 2))) ;; =&gt; '(4)\n(-min-by (-on 'string-lessp 'number-to-string) '(2 100 22)) ;; =&gt; 22<\/pre><\/div><h4><a id=\"user-content--flip-func\" class=\"anchor\" aria-hidden=\"true\" href=\"#-flip-func\"><\/a>-flip <code>(func)<\/code><\/h4><p>Swap the order of arguments for binary function <code>func<\/code>.<\/p><p>In types: (a -&gt; b -&gt; c) -&gt; b -&gt; a -&gt; c<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-flip '&lt;) 2 1) ;; =&gt; t\n(funcall (-flip '-) 3 8) ;; =&gt; 5\n(-sort (-flip '&lt;) '(4 3 6 1)) ;; =&gt; '(6 4 3 1)<\/pre><\/div><h4><a id=\"user-content--const-c\" class=\"anchor\" aria-hidden=\"true\" href=\"#-const-c\"><\/a>-const <code>(c)<\/code><\/h4><p>Return a function that returns <code>c<\/code> ignoring any additional arguments.<\/p><p>In types: a -&gt; b -&gt; a<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-const 2) 1 3 \"foo\") ;; =&gt; 2\n(-map (-const 1) '(\"a\" \"b\" \"c\" \"d\")) ;; =&gt; '(1 1 1 1)\n(-sum (-map (-const 1) '(\"a\" \"b\" \"c\" \"d\"))) ;; =&gt; 4<\/pre><\/div><h4><a id=\"user-content--cut-rest-params\" class=\"anchor\" aria-hidden=\"true\" href=\"#-cut-rest-params\"><\/a>-cut <code>(&amp;rest params)<\/code><\/h4><p>Take n-ary function and n arguments and specialize some of them. Arguments denoted by &lt;&gt; will be left unspecialized.<\/p><p>See <code>srfi-26<\/code> for detailed description.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-cut list 1 &lt;&gt; 3 &lt;&gt; 5) 2 4) ;; =&gt; '(1 2 3 4 5)\n(-map (-cut funcall &lt;&gt; 5) '(1+ 1- (lambda (x) (\/ 1.0 x)))) ;; =&gt; '(6 4 0.2)\n(-map (-cut &lt;&gt; 1 2 3) (list 'list 'vector 'string)) ;; =&gt; '((1 2 3) [1 2 3] \"\ufffd\ufffd\ufffd\")<\/pre><\/div><h4><a id=\"user-content--not-pred\" class=\"anchor\" aria-hidden=\"true\" href=\"#-not-pred\"><\/a>-not <code>(pred)<\/code><\/h4><p>Take a unary predicate <code>pred<\/code> and return a unary predicate that returns t if <code>pred<\/code> returns nil and nil if <code>pred<\/code> returns non-nil.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-not 'even?) 5) ;; =&gt; t\n(-filter (-not (-partial '&lt; 4)) '(1 2 3 4 5 6 7 8)) ;; =&gt; '(1 2 3 4)<\/pre><\/div><h4><a id=\"user-content--orfn-rest-preds\" class=\"anchor\" aria-hidden=\"true\" href=\"#-orfn-rest-preds\"><\/a>-orfn <code>(&amp;rest preds)<\/code><\/h4><p>Take list of unary predicates <code>preds<\/code> and return a unary predicate with argument x that returns non-nil if at least one of the <code>preds<\/code> returns non-nil on x.<\/p><p>In types: [a -&gt; Bool] -&gt; a -&gt; Bool<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(-filter (-orfn 'even? (-partial (-flip '&lt;) 5)) '(1 2 3 4 5 6 7 8 9 10)) ;; =&gt; '(1 2 3 4 6 8 10)\n(funcall (-orfn 'stringp 'even?) \"foo\") ;; =&gt; t<\/pre><\/div><h4><a id=\"user-content--andfn-rest-preds\" class=\"anchor\" aria-hidden=\"true\" href=\"#-andfn-rest-preds\"><\/a>-andfn <code>(&amp;rest preds)<\/code><\/h4><p>Take list of unary predicates <code>preds<\/code> and return a unary predicate with argument x that returns non-nil if all of the <code>preds<\/code> returns non-nil on x.<\/p><p>In types: [a -&gt; Bool] -&gt; a -&gt; Bool<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-andfn (-cut &lt; &lt;&gt; 10) 'even?) 6) ;; =&gt; t\n(funcall (-andfn (-cut &lt; &lt;&gt; 10) 'even?) 12) ;; =&gt; nil\n(-filter (-andfn (-not 'even?) (-cut &gt;= 5 &lt;&gt;)) '(1 2 3 4 5 6 7 8 9 10)) ;; =&gt; '(1 3 5)<\/pre><\/div><h4><a id=\"user-content--iteratefn-fn-n\" class=\"anchor\" aria-hidden=\"true\" href=\"#-iteratefn-fn-n\"><\/a>-iteratefn <code>(fn n)<\/code><\/h4><p>Return a function <code>fn<\/code> composed <code>n<\/code> times with itself.<\/p><p><code>fn<\/code> is a unary function. If you need to use a function of higher arity, use <a href=\"#-applify-fn\"><code>-applify<\/code><\/a> first to turn it into a unary function.<\/p><p>With n = 0, this acts as identity function.<\/p><p>In types: (a -&gt; a) -&gt; Int -&gt; a -&gt; a.<\/p><p>This function satisfies the following law:<\/p><pre>(funcall (-iteratefn fn n) init) = (-last-item (-iterate fn init (1+ n))).\n<\/pre><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-iteratefn (lambda (x) (* x x)) 3) 2) ;; =&gt; 256\n(funcall (-iteratefn '1+ 3) 1) ;; =&gt; 4\n(funcall (-iteratefn 'cdr 3) '(1 2 3 4 5)) ;; =&gt; '(4 5)<\/pre><\/div><h4><a id=\"user-content--fixfn-fn-optional-equal-test-halt-test\" class=\"anchor\" aria-hidden=\"true\" href=\"#-fixfn-fn-optional-equal-test-halt-test\"><\/a>-fixfn <code>(fn &amp;optional equal-test halt-test)<\/code><\/h4><p>Return a function that computes the (least) fixpoint of <code>fn<\/code>.<\/p><p><code>fn<\/code> must be a unary function. The returned lambda takes a single argument, <code>x<\/code>, the initial value for the fixpoint iteration. The iteration halts when either of the following conditions is satisfied:<\/p><ol><li>\n<p>Iteration converges to the fixpoint, with equality being tested using <code>equal-test<\/code>. If <code>equal-test<\/code> is not specified, <code>equal<\/code> is used. For functions over the floating point numbers, it may be necessary to provide an appropriate appoximate comparison test.<\/p>\n<\/li>\n<li>\n<p><code>halt-test<\/code> returns a non-nil value. <code>halt-test<\/code> defaults to a simple counter that returns t after <code>-fixfn-max-iterations<\/code>, to guard against infinite iteration. Otherwise, <code>halt-test<\/code> must be a function that accepts a single argument, the current value of <code>x<\/code>, and returns non-nil as long as iteration should continue. In this way, a more sophisticated convergence test may be supplied by the caller.<\/p>\n<\/li>\n<\/ol><p>The return value of the lambda is either the fixpoint or, if iteration halted before converging, a cons with car <code>halted<\/code> and cdr the final output from <code>halt-test<\/code>.<\/p><p>In types: (a -&gt; a) -&gt; a -&gt; a.<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-fixfn 'cos 'approx-equal) 0.7) ;; ~&gt; 0.7390851332151607\n(funcall (-fixfn (lambda (x) (expt (+ x 10) 0.25))) 2.0) ;; =&gt; 1.8555845286409378\n(funcall (-fixfn 'sin 'approx-equal) 0.1) ;; =&gt; '(halted . t)<\/pre><\/div><h4><a id=\"user-content--prodfn-rest-fns\" class=\"anchor\" aria-hidden=\"true\" href=\"#-prodfn-rest-fns\"><\/a>-prodfn <code>(&amp;rest fns)<\/code><\/h4><p>Take a list of n functions and return a function that takes a list of length n, applying i-th function to i-th element of the input list. Returns a list of length n.<\/p><p>In types (for n=2): ((a -&gt; b), (c -&gt; d)) -&gt; (a, c) -&gt; (b, d)<\/p><p>This function satisfies the following laws:<\/p><pre>(-compose (-prodfn f g ...) (-prodfn f' g' ...)) = (-prodfn (-compose f f') (-compose g g') ...)\n(-prodfn f g ...) = (-juxt (-compose f (-partial 'nth 0)) (-compose g (-partial 'nth 1)) ...)\n(-compose (-prodfn f g ...) (-juxt f' g' ...)) = (-juxt (-compose f f') (-compose g g') ...)\n(-compose (-partial 'nth n) (-prod f1 f2 ...)) = (-compose fn (-partial 'nth n))\n<\/pre><div class=\"highlight highlight-source-emacs-lisp\"><pre>(funcall (-prodfn '1+ '1- 'number-to-string) '(1 2 3)) ;; =&gt; '(2 1 \"3\")\n(-map (-prodfn '1+ '1-) '((1 2) (3 4) (5 6) (7 8))) ;; =&gt; '((2 1) (4 3) (6 5) (8 7))\n(apply '+ (funcall (-prodfn 'length 'string-to-number) '((1 2 3) \"15\"))) ;; =&gt; 18<\/pre><\/div><h2><a id=\"user-content-contribute\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribute\"><\/a>Contribute<\/h2><p>Yes, please do. Pure functions in the list manipulation realm only, please. There's a suite of tests in <code>dev\/examples.el<\/code>, so remember to add tests for your function, or I might break it later.<\/p><p>You'll find the repo at:<\/p><pre>https:\/\/github.com\/magnars\/dash.el\n<\/pre><p>Run the tests with<\/p><pre>.\/run-tests.sh\n<\/pre><p>Create the docs with<\/p><pre>.\/create-docs.sh\n<\/pre><p>I highly recommend that you install these as a pre-commit hook, so that the tests are always running and the docs are always in sync:<\/p><pre>cp pre-commit.sh .git\/hooks\/pre-commit\n<\/pre><p>Oh, and don't edit <code>README.md<\/code> directly, it is auto-generated. Change <code>readme-template.md<\/code> or <code>examples-to-docs.el<\/code> instead.<\/p><h2><a id=\"user-content-changelist\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelist\"><\/a>Changelist<\/h2><h3><a id=\"user-content-from-215-to-216\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-215-to-216\"><\/a>From 2.15 to 2.16<\/h3><ul><li>Added <code>--doto<\/code>, anaphoric version of <code>-doto<\/code> (#282)<\/li>\n<li>Aliased <code>-cons-pair-p<\/code> to <code>-cons-pair?<\/code>(#288)<\/li>\n<li>Generalized <code>-rotate<\/code> for |n| greater than the length of the list (@leungbk, #290)<\/li>\n<li>Added a mechanism to extend destructuring with custom matchers (@yyoncho, #277)<\/li>\n<\/ul><h3><a id=\"user-content-from-214-to-215\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-214-to-215\"><\/a>From 2.14 to 2.15<\/h3><p>This release brings new destructuring features, some new control flow functions and performance optimizations.<\/p><ul><li>Added <code>-setq<\/code> with destructuring binding support similar to <code>-let<\/code> family (<a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/116\">#116<\/a>)<\/li>\n<li>Added smarter key destructuring in <code>-let<\/code> and friends where variables are auto-derived from keys (<a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/111\">#111<\/a>)<\/li>\n<li>Allow <code>-let<\/code> bindings with place only (<a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/256\">#256<\/a>)<\/li>\n<li>Added <code>-each-r<\/code> and <code>-each-r-while<\/code> (@doublep, <a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/159\">#159<\/a>)<\/li>\n<li>Added <code>-common-suffix<\/code> (@basil-conto, <a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/263\">#263<\/a>)<\/li>\n<li>Improved performance of folds (<code>-reduce<\/code> and friends) (@basil-conto, <a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/264\">#264<\/a>)<\/li>\n<\/ul><h3><a id=\"user-content-from-213-to-214\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-213-to-214\"><\/a>From 2.13 to 2.14<\/h3><p>This release retires Emacs 23 support. We will still try to keep things compatible but no future guarantees are made.<\/p><ul><li>Added edebug support for threading macros (@Wilfred)<\/li>\n<li>Added <code>-unzip<\/code><\/li>\n<li>Added gv setters for <code>-first-item<\/code> and <code>-last-item<\/code><\/li>\n<li>Added <code>-powerset<\/code> and <code>-permutations<\/code> (@holomorph)<\/li>\n<li>Added <code>-as-&gt;<\/code> for threading a named variable (@zck)<\/li>\n<li>Added <code>-partition-after-pred<\/code>, <code>-partition-before-pred<\/code>, <code>-partition-after-item<\/code>, <code>-partition-before-item<\/code> (@zck)<\/li>\n<li>Fixed a bug in <code>-any-p<\/code> and friends testing for <code>null<\/code> on lists containing <code>nil<\/code> (#239)<\/li>\n<li>Fixed infinite loop bug in <code>-zip<\/code> and <code>-interleave<\/code> when called with empty input.<\/li>\n<li>Added <code>-second-item<\/code> through to <code>-fifth-item<\/code> as an alternative to <code>nth<\/code> (@Wilfred)<\/li>\n<li>Added <code>-tails<\/code> and <code>-inits<\/code><\/li>\n<li>Added <code>-running-sum<\/code> and <code>-running-product<\/code><\/li>\n<li>Added <code>-reductions[-r][-from]<\/code> family of functions (like <code>-reduce<\/code> but collecting intermediate results)<\/li>\n<li>Added <code>-common-prefix<\/code> (@basil-conto)<\/li>\n<\/ul><h3><a id=\"user-content-from-212-to-213\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-212-to-213\"><\/a>From 2.12 to 2.13<\/h3><ul><li><code>-let<\/code> now supports <code>&amp;alist<\/code> in destructuring.<\/li>\n<li>Various performance improvements.<\/li>\n<li><code>-zip<\/code> will change in future so it always returns lists. Added <code>-zip-pair<\/code> for users who explicitly want the old behavior.<\/li>\n<li>Added lexical binding pragma to dash.el, fixes <a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/130\">#130<\/a> in Emacs 24+.<\/li>\n<li>Added <code>-select-column<\/code> and <code>-select-columns<\/code>.<\/li>\n<li>Fixed an issue with <code>-map-last<\/code> and <code>--remove-last<\/code> where they modified their inputs (<a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/158\">#158<\/a>).<\/li>\n<li>Added <code>-each-indexed<\/code> and <code>--each-indexed<\/code>.<\/li>\n<li>Added <code>-take-last<\/code> and <code>-drop-last<\/code>.<\/li>\n<li>Added <code>-doto<\/code> macro.<\/li>\n<li><code>-cut &lt;&gt;<\/code> is now treated as a function, consistent with SRFI 26 (<a href=\"https:\/\/github.com\/magnars\/dash.el\/issues\/185\">#185<\/a>)<\/li>\n<\/ul><h3><a id=\"user-content-from-211-to-212\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-211-to-212\"><\/a>From 2.11 to 2.12<\/h3><ul><li>Add GNU ELPA support. (Phillip Lord)<\/li>\n<li>Add <code>-some-&gt;<\/code>, <code>-some-&gt;&gt;<\/code>, and <code>-some--&gt;<\/code> macros. (Cam Saul)<\/li>\n<li><code>-is-suffix?<\/code> no longer destroys input list.<\/li>\n<li>Faster hashtable implementation for <code>-union<\/code>.<\/li>\n<li>Improvements to docstrings and examples<\/li>\n<\/ul><h3><a id=\"user-content-from-210-to-211\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-210-to-211\"><\/a>From 2.10 to 2.11<\/h3><ul><li>Lots of clean up wrt byte compilation, debug macros and tests<\/li>\n<\/ul><h3><a id=\"user-content-from-29-to-210\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-29-to-210\"><\/a>From 2.9 to 2.10<\/h3><ul><li>Add <code>-let<\/code> destructuring to <code>-if-let<\/code> and <code>-when-let<\/code> (Fredrik Bergroth)<\/li>\n<\/ul><h3><a id=\"user-content-from-28-to-29\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-28-to-29\"><\/a>From 2.8 to 2.9<\/h3><ul><li>Add <code>-let<\/code>, <code>-let*<\/code> and <code>-lambda<\/code> with destructuring<\/li>\n<li>Add <code>-tree-seq<\/code> and <code>-tree-map-nodes<\/code><\/li>\n<li>Add <code>-non-nil<\/code><\/li>\n<li>Add <code>-fix<\/code><\/li>\n<li>Add <code>-fixfn<\/code> (dash-functional 1.2)<\/li>\n<li>Add <code>-copy<\/code> (Wilfred Hughes)<\/li>\n<\/ul><h3><a id=\"user-content-from-27-to-28\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-27-to-28\"><\/a>From 2.7 to 2.8<\/h3><ul><li>Add <code>-butlast<\/code><\/li>\n<\/ul><h3><a id=\"user-content-from-26-to-27\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-26-to-27\"><\/a>From 2.6 to 2.7<\/h3><ul><li><code>-zip<\/code> now supports more than two lists (Steve Lamb)<\/li>\n<li>Add <code>-cycle<\/code> , <code>-pad<\/code> , <code>-annotate<\/code> , <code>-zip-fill<\/code> (Steve Lamb)<\/li>\n<li>Add <code>-table<\/code>, <code>-table-flat<\/code> (finite cartesian product)<\/li>\n<li>Add <code>-flatten-n<\/code><\/li>\n<li><code>-slice<\/code> now supports \"step\" argument<\/li>\n<li>Add functional combinators <code>-iteratefn<\/code>, <code>-prodfn<\/code><\/li>\n<li>Add <code>-replace<\/code>, <code>-splice<\/code>, <code>-splice-list<\/code> which generalize <code>-replace-at<\/code> and <code>-insert-at<\/code><\/li>\n<li>Add <code>-compose<\/code>, <code>-iteratefn<\/code> and <code>-prodfn<\/code> (dash-functional 1.1)<\/li>\n<\/ul><h3><a id=\"user-content-from-25-to-26\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-25-to-26\"><\/a>From 2.5 to 2.6<\/h3><ul><li>Add <code>-is-prefix-p<\/code>, <code>-is-suffix-p<\/code>, <code>-is-infix-p<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-iterate<\/code>, <code>-unfold<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-split-on<\/code>, <code>-split-when<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-find-last-index<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-list<\/code> (Johan Andersson)<\/li>\n<\/ul><h3><a id=\"user-content-from-24-to-25\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-24-to-25\"><\/a>From 2.4 to 2.5<\/h3><ul><li>Add <code>-same-items?<\/code> (Johan Andersson)<\/li>\n<li>A few bugfixes<\/li>\n<\/ul><h3><a id=\"user-content-from-23-to-24\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-23-to-24\"><\/a>From 2.3 to 2.4<\/h3><ul><li>Add <code>-snoc<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-replace-at<\/code>, <code>-update-at<\/code>, <code>-remove-at<\/code>, and <code>-remove-at-indices<\/code> (Matus Goljer)<\/li>\n<\/ul><h3><a id=\"user-content-from-22-to-23\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-22-to-23\"><\/a>From 2.2 to 2.3<\/h3><ul><li>Add tree operations (Matus Goljer)<\/li>\n<li>Make font-lock optional<\/li>\n<\/ul><h3><a id=\"user-content-from-21-to-22\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-21-to-22\"><\/a>From 2.1 to 2.2<\/h3><ul><li>Add <code>-compose<\/code> (Christina Whyte)<\/li>\n<\/ul><h3><a id=\"user-content-from-20-to-21\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-20-to-21\"><\/a>From 2.0 to 2.1<\/h3><ul><li>Add indexing operations (Matus Goljer)<\/li>\n<\/ul><h3><a id=\"user-content-from-18-to-20\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-18-to-20\"><\/a>From 1.8 to 2.0<\/h3><ul><li>Split out <code>dash-functional.el<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-andfn<\/code>, <code>-orfn<\/code>, <code>-not<\/code>, <code>-cut<\/code>, <code>-const<\/code>, <code>-flip<\/code> and <code>-on<\/code>. (Matus Goljer)<\/li>\n<li>Fix <code>-min<\/code>, <code>-max<\/code>, <code>-min-by<\/code> and <code>-max-by<\/code> (Matus Goljer)<\/li>\n<\/ul><h3><a id=\"user-content-from-17-to-18\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-17-to-18\"><\/a>From 1.7 to 1.8<\/h3><ul><li>Add <code>-first-item<\/code> and <code>-last-item<\/code> (Wilfred Hughes)<\/li>\n<\/ul><h3><a id=\"user-content-from-16-to-17\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-16-to-17\"><\/a>From 1.6 to 1.7<\/h3><ul><li>Add <code>-rotate<\/code> (Matus Goljer)<\/li>\n<\/ul><h3><a id=\"user-content-from-15-to-16\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-15-to-16\"><\/a>From 1.5 to 1.6<\/h3><ul><li>Add <code>-min<\/code>, <code>-max<\/code>, <code>-min-by<\/code> and <code>-max-by<\/code> (Johan Andersson)<\/li>\n<\/ul><h3><a id=\"user-content-from-14-to-15\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-14-to-15\"><\/a>From 1.4 to 1.5<\/h3><ul><li>Add <code>-sum<\/code> and <code>-product<\/code> (Johan Andersson)<\/li>\n<\/ul><h3><a id=\"user-content-from-13-to-14\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-13-to-14\"><\/a>From 1.3 to 1.4<\/h3><ul><li>Add <code>-sort<\/code><\/li>\n<li>Add <code>-reduce-r<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-reduce-r-from<\/code> (Matus Goljer)<\/li>\n<\/ul><h3><a id=\"user-content-from-12-to-13\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-12-to-13\"><\/a>From 1.2 to 1.3<\/h3><ul><li>Add <code>-partition-in-steps<\/code><\/li>\n<li>Add <code>-partition-all-in-steps<\/code><\/li>\n<\/ul><h3><a id=\"user-content-from-11-to-12\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-11-to-12\"><\/a>From 1.1 to 1.2<\/h3><ul><li>Add <code>-last<\/code> (Matus Goljer)<\/li>\n<li>Add <code>-insert-at<\/code> (Emanuel Evans)<\/li>\n<li>Add <code>-when-let<\/code> and <code>-if-let<\/code> (Emanuel Evans)<\/li>\n<li>Add <code>-when-let*<\/code> and <code>-if-let*<\/code> (Emanuel Evans)<\/li>\n<li>Some bugfixes<\/li>\n<\/ul><h2><a id=\"user-content-contributors\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributors\"><\/a>Contributors<\/h2><ul><li><a href=\"https:\/\/github.com\/Fuco1\">Matus Goljer<\/a> contributed lots of features and functions.<\/li>\n<li><a href=\"https:\/\/github.com\/tkf\">Takafumi Arakaki<\/a> contributed <code>-group-by<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/tali713\">tali713<\/a> is the author of <code>-applify<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/vemv\">V\u00edctor M. Valenzuela<\/a> contributed <code>-repeat<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/nicferrier\">Nic Ferrier<\/a> contributed <code>-cons*<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/Wilfred\">Wilfred Hughes<\/a> contributed <code>-slice<\/code>, <code>-first-item<\/code> and <code>-last-item<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/shosti\">Emanuel Evans<\/a> contributed <code>-if-let<\/code>, <code>-when-let<\/code> and <code>-insert-at<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/rejeep\">Johan Andersson<\/a> contributed <code>-sum<\/code>, <code>-product<\/code> and <code>-same-items?<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/kurisuwhyte\">Christina Whyte<\/a> contributed <code>-compose<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/steventlamb\">Steve Lamb<\/a> contributed <code>-cycle<\/code>, <code>-pad<\/code>, <code>-annotate<\/code>, <code>-zip-fill<\/code> and an n-ary version of <code>-zip<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/fbergroth\">Fredrik Bergroth<\/a> made the <code>-if-let<\/code> family use <code>-let<\/code> destructuring and improved script for generating documentation.<\/li>\n<li><a href=\"https:\/\/github.com\/holomorph\">Mark Oteiza<\/a> contributed the script to create an info manual.<\/li>\n<li><a href=\"https:\/\/github.com\/wasamasa\">Vasilij Schneidermann<\/a> contributed <code>-some<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/occidens\">William West<\/a> made <code>-fixfn<\/code> more robust at handling floats.<\/li>\n<li><a href=\"https:\/\/github.com\/camsaul\">Cam Sa\u00fcl<\/a> contributed <code>-some-&gt;<\/code>, <code>-some-&gt;&gt;<\/code>, and <code>-some--&gt;<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/basil-conto\">Basil L. Contovounesios<\/a> contributed <code>-common-prefix<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/doublep\">Paul Pogonyshev<\/a> contributed <code>-each-r<\/code> and <code>-each-r-while<\/code>.<\/li>\n<\/ul><p>Thanks!<\/p><p>New contributors are welcome. To ensure that dash.el can be distributed with ELPA or Emacs, we would request that all contributors assign copyright to the Free Software Foundation.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright (C) 2012-2016 Free Software Foundation, Inc.<\/p><p>Authors: Magnar Sveen <a href=\"mailto:magnars@gmail.com\">magnars@gmail.com<\/a> Keywords: lists<\/p><p>This program is free software; you can redistribute it and\/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.<\/p><p>This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.<\/p><p>You should have received a copy of the GNU General Public License along with this program. If not, see <a href=\"http:\/\/www.gnu.org\/licenses\/\" rel=\"nofollow\">http:\/\/www.gnu.org\/licenses\/<\/a>.<\/p>","created_at":"2019-12-25T03:44:05+0000","updated_at":"2019-12-25T03:44:05+0000","published_by":["magnars"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":83,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/265785?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1073,"title":"magnars\/s.el","url":"https:\/\/github.com\/magnars\/s.el","content":"<p>The long lost Emacs string manipulation library.<\/p><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2><p>It's available on <a href=\"http:\/\/marmalade-repo.org\/\" rel=\"nofollow\">marmalade<\/a> and <a href=\"https:\/\/melpa.org\/\" rel=\"nofollow\">Melpa<\/a>:<\/p><pre>M-x package-install s\n<\/pre><p>Or you can just dump <code>s.el<\/code> in your load path somewhere.<\/p><h2><a id=\"user-content-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#functions\"><\/a>Functions<\/h2><h3><a id=\"user-content-tweak-whitespace\" class=\"anchor\" aria-hidden=\"true\" href=\"#tweak-whitespace\"><\/a>Tweak whitespace<\/h3><ul><li><a href=\"#s-trim-s\">s-trim<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-trim-left-s\">s-trim-left<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-trim-right-s\">s-trim-right<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-chomp-s\">s-chomp<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-collapse-whitespace-s\">s-collapse-whitespace<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-word-wrap-len-s\">s-word-wrap<\/a> <code>(len s)<\/code><\/li>\n<li><a href=\"#s-center-len-s\">s-center<\/a> <code>(len s)<\/code><\/li>\n<li><a href=\"#s-pad-left-len-padding-s\">s-pad-left<\/a> <code>(len padding s)<\/code><\/li>\n<li><a href=\"#s-pad-right-len-padding-s\">s-pad-right<\/a> <code>(len padding s)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-to-shorter-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#to-shorter-string\"><\/a>To shorter string<\/h3><ul><li><a href=\"#s-truncate-len-s\">s-truncate<\/a> <code>(len s)<\/code><\/li>\n<li><a href=\"#s-left-len-s\">s-left<\/a> <code>(len s)<\/code><\/li>\n<li><a href=\"#s-right-len-s\">s-right<\/a> <code>(len s)<\/code><\/li>\n<li><a href=\"#s-chop-suffix-suffix-s\">s-chop-suffix<\/a> <code>(suffix s)<\/code><\/li>\n<li><a href=\"#s-chop-suffixes-suffixes-s\">s-chop-suffixes<\/a> <code>(suffixes s)<\/code><\/li>\n<li><a href=\"#s-chop-prefix-prefix-s\">s-chop-prefix<\/a> <code>(prefix s)<\/code><\/li>\n<li><a href=\"#s-chop-prefixes-prefixes-s\">s-chop-prefixes<\/a> <code>(prefixes s)<\/code><\/li>\n<li><a href=\"#s-shared-start-s1-s2\">s-shared-start<\/a> <code>(s1 s2)<\/code><\/li>\n<li><a href=\"#s-shared-end-s1-s2\">s-shared-end<\/a> <code>(s1 s2)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-to-longer-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#to-longer-string\"><\/a>To longer string<\/h3><ul><li><a href=\"#s-repeat-num-s\">s-repeat<\/a> <code>(num s)<\/code><\/li>\n<li><a href=\"#s-concat-rest-strings\">s-concat<\/a> <code>(&amp;rest strings)<\/code><\/li>\n<li><a href=\"#s-prepend-prefix-s\">s-prepend<\/a> <code>(prefix s)<\/code><\/li>\n<li><a href=\"#s-append-suffix-s\">s-append<\/a> <code>(suffix s)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-to-and-from-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#to-and-from-lists\"><\/a>To and from lists<\/h3><ul><li><a href=\"#s-lines-s\">s-lines<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-match-regexp-s-optional-start\">s-match<\/a> <code>(regexp s &amp;optional start)<\/code><\/li>\n<li><a href=\"#s-match-strings-all-regex-string\">s-match-strings-all<\/a> <code>(regex string)<\/code><\/li>\n<li><a href=\"#s-matched-positions-all-regexp-string-optional-subexp-depth\">s-matched-positions-all<\/a> <code>(regexp string &amp;optional subexp-depth)<\/code><\/li>\n<li><a href=\"#s-slice-at-regexp-s\">s-slice-at<\/a> <code>(regexp s)<\/code><\/li>\n<li><a href=\"#s-split-separator-s-optional-omit-nulls\">s-split<\/a> <code>(separator s &amp;optional omit-nulls)<\/code><\/li>\n<li><a href=\"#s-split-up-to-separator-s-n-optional-omit-nulls\">s-split-up-to<\/a> <code>(separator s n &amp;optional omit-nulls)<\/code><\/li>\n<li><a href=\"#s-join-separator-strings\">s-join<\/a> <code>(separator strings)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-predicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#predicates\"><\/a>Predicates<\/h3><ul><li><a href=\"#s-equals-s1-s2\">s-equals?<\/a> <code>(s1 s2)<\/code><\/li>\n<li><a href=\"#s-less-s1-s2\">s-less?<\/a> <code>(s1 s2)<\/code><\/li>\n<li><a href=\"#s-matches-regexp-s-optional-start\">s-matches?<\/a> <code>(regexp s &amp;optional start)<\/code><\/li>\n<li><a href=\"#s-blank-s\">s-blank?<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-present-s\">s-present?<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-ends-with-suffix-s-optional-ignore-case\">s-ends-with?<\/a> <code>(suffix s &amp;optional ignore-case)<\/code><\/li>\n<li><a href=\"#s-starts-with-prefix-s-optional-ignore-case\">s-starts-with?<\/a> <code>(prefix s &amp;optional ignore-case)<\/code><\/li>\n<li><a href=\"#s-contains-needle-s-optional-ignore-case\">s-contains?<\/a> <code>(needle s &amp;optional ignore-case)<\/code><\/li>\n<li><a href=\"#s-lowercase-s\">s-lowercase?<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-uppercase-s\">s-uppercase?<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-mixedcase-s\">s-mixedcase?<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-capitalized-s\">s-capitalized?<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-numeric-s\">s-numeric?<\/a> <code>(s)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-the-misc-bucket\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-misc-bucket\"><\/a>The misc bucket<\/h3><ul><li><a href=\"#s-replace-old-new-s\">s-replace<\/a> <code>(old new s)<\/code><\/li>\n<li><a href=\"#s-replace-all-replacements-s\">s-replace-all<\/a> <code>(replacements s)<\/code><\/li>\n<li><a href=\"#s-downcase-s\">s-downcase<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-upcase-s\">s-upcase<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-capitalize-s\">s-capitalize<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-titleize-s\">s-titleize<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-with-s-form-rest-more\">s-with<\/a> <code>(s form &amp;rest more)<\/code><\/li>\n<li><a href=\"#s-index-of-needle-s-optional-ignore-case\">s-index-of<\/a> <code>(needle s &amp;optional ignore-case)<\/code><\/li>\n<li><a href=\"#s-reverse-s\">s-reverse<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-presence-s\">s-presence<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-format-template-replacer-optional-extra\">s-format<\/a> <code>(template replacer &amp;optional extra)<\/code><\/li>\n<li><a href=\"#s-lex-format-format-str\">s-lex-format<\/a> <code>(format-str)<\/code><\/li>\n<li><a href=\"#s-count-matches-regexp-s-optional-start-end\">s-count-matches<\/a> <code>(regexp s &amp;optional start end)<\/code><\/li>\n<li><a href=\"#s-wrap-s-prefix-optional-suffix\">s-wrap<\/a> <code>(s prefix &amp;optional suffix)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-pertaining-to-words\" class=\"anchor\" aria-hidden=\"true\" href=\"#pertaining-to-words\"><\/a>Pertaining to words<\/h3><ul><li><a href=\"#s-split-words-s\">s-split-words<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-lower-camel-case-s\">s-lower-camel-case<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-upper-camel-case-s\">s-upper-camel-case<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-snake-case-s\">s-snake-case<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-dashed-words-s\">s-dashed-words<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-capitalized-words-s\">s-capitalized-words<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-titleized-words-s\">s-titleized-words<\/a> <code>(s)<\/code><\/li>\n<li><a href=\"#s-word-initials-s\">s-word-initials<\/a> <code>(s)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-documentation-and-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation-and-examples\"><\/a>Documentation and examples<\/h2><h3><a id=\"user-content-s-trim-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-trim-s\"><\/a>s-trim <code>(s)<\/code><\/h3><p>Remove whitespace at the beginning and end of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-trim \"trim \") ;; =&gt; \"trim\"\n(s-trim \" this\") ;; =&gt; \"this\"\n(s-trim \" only  trims beg and end  \") ;; =&gt; \"only  trims beg and end\"<\/pre><\/div><h3><a id=\"user-content-s-trim-left-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-trim-left-s\"><\/a>s-trim-left <code>(s)<\/code><\/h3><p>Remove whitespace at the beginning of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-trim-left \"trim \") ;; =&gt; \"trim \"\n(s-trim-left \" this\") ;; =&gt; \"this\"<\/pre><\/div><h3><a id=\"user-content-s-trim-right-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-trim-right-s\"><\/a>s-trim-right <code>(s)<\/code><\/h3><p>Remove whitespace at the end of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-trim-right \"trim \") ;; =&gt; \"trim\"\n(s-trim-right \" this\") ;; =&gt; \" this\"<\/pre><\/div><h3><a id=\"user-content-s-chomp-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-chomp-s\"><\/a>s-chomp <code>(s)<\/code><\/h3><p>Remove one trailing <code>\\n<\/code>, <code>\\r<\/code> or <code>\\r\\n<\/code> from <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-chomp \"no newlines\\n\") ;; =&gt; \"no newlines\"\n(s-chomp \"no newlines\\r\\n\") ;; =&gt; \"no newlines\"\n(s-chomp \"some newlines\\n\\n\") ;; =&gt; \"some newlines\\n\"<\/pre><\/div><h3><a id=\"user-content-s-collapse-whitespace-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-collapse-whitespace-s\"><\/a>s-collapse-whitespace <code>(s)<\/code><\/h3><p>Convert all adjacent whitespace characters to a single space.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-collapse-whitespace \"only   one space   please\") ;; =&gt; \"only one space please\"\n(s-collapse-whitespace \"collapse \\n all \\t sorts of \\r whitespace\") ;; =&gt; \"collapse all sorts of whitespace\"<\/pre><\/div><h3><a id=\"user-content-s-word-wrap-len-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-word-wrap-len-s\"><\/a>s-word-wrap <code>(len s)<\/code><\/h3><p>If <code>s<\/code> is longer than <code>len<\/code>, wrap the words with newlines.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-word-wrap 10 \"This is too long\") ;; =&gt; \"This is\\ntoo long\"\n(s-word-wrap 10 \"This is way way too long\") ;; =&gt; \"This is\\nway way\\ntoo long\"\n(s-word-wrap 10 \"It-wraps-words-but-does-not-break-them\") ;; =&gt; \"It-wraps-words-but-does-not-break-them\"<\/pre><\/div><h3><a id=\"user-content-s-center-len-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-center-len-s\"><\/a>s-center <code>(len s)<\/code><\/h3><p>If <code>s<\/code> is shorter than <code>len<\/code>, pad it with spaces so it is centered.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-center 5 \"a\") ;; =&gt; \"  a  \"\n(s-center 5 \"ab\") ;; =&gt; \"  ab \"\n(s-center 1 \"abc\") ;; =&gt; \"abc\"<\/pre><\/div><h3><a id=\"user-content-s-pad-left-len-padding-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-pad-left-len-padding-s\"><\/a>s-pad-left <code>(len padding s)<\/code><\/h3><p>If <code>s<\/code> is shorter than <code>len<\/code>, pad it with <code>padding<\/code> on the left.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-pad-left 3 \"0\" \"3\") ;; =&gt; \"003\"\n(s-pad-left 3 \"0\" \"23\") ;; =&gt; \"023\"\n(s-pad-left 3 \"0\" \"1234\") ;; =&gt; \"1234\"<\/pre><\/div><h3><a id=\"user-content-s-pad-right-len-padding-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-pad-right-len-padding-s\"><\/a>s-pad-right <code>(len padding s)<\/code><\/h3><p>If <code>s<\/code> is shorter than <code>len<\/code>, pad it with <code>padding<\/code> on the right.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-pad-right 3 \".\" \"3\") ;; =&gt; \"3..\"\n(s-pad-right 3 \".\" \"23\") ;; =&gt; \"23.\"\n(s-pad-right 3 \".\" \"1234\") ;; =&gt; \"1234\"<\/pre><\/div><h3><a id=\"user-content-s-truncate-len-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-truncate-len-s\"><\/a>s-truncate <code>(len s)<\/code><\/h3><p>If <code>s<\/code> is longer than <code>len<\/code>, cut it down to <code>len<\/code> - 3 and add ... at the end.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-truncate 6 \"This is too long\") ;; =&gt; \"Thi...\"\n(s-truncate 16 \"This is also too long\") ;; =&gt; \"This is also ...\"\n(s-truncate 16 \"But this is not!\") ;; =&gt; \"But this is not!\"<\/pre><\/div><h3><a id=\"user-content-s-left-len-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-left-len-s\"><\/a>s-left <code>(len s)<\/code><\/h3><p>Returns up to the <code>len<\/code> first chars of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-left 3 \"lib\/file.js\") ;; =&gt; \"lib\"\n(s-left 3 \"li\") ;; =&gt; \"li\"<\/pre><\/div><h3><a id=\"user-content-s-right-len-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-right-len-s\"><\/a>s-right <code>(len s)<\/code><\/h3><p>Returns up to the <code>len<\/code> last chars of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-right 3 \"lib\/file.js\") ;; =&gt; \".js\"\n(s-right 3 \"li\") ;; =&gt; \"li\"<\/pre><\/div><h3><a id=\"user-content-s-chop-suffix-suffix-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-chop-suffix-suffix-s\"><\/a>s-chop-suffix <code>(suffix s)<\/code><\/h3><p>Remove <code>suffix<\/code> if it is at end of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-chop-suffix \"-test.js\" \"penguin-test.js\") ;; =&gt; \"penguin\"\n(s-chop-suffix \"\\n\" \"no newlines\\n\") ;; =&gt; \"no newlines\"\n(s-chop-suffix \"\\n\" \"some newlines\\n\\n\") ;; =&gt; \"some newlines\\n\"<\/pre><\/div><h3><a id=\"user-content-s-chop-suffixes-suffixes-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-chop-suffixes-suffixes-s\"><\/a>s-chop-suffixes <code>(suffixes s)<\/code><\/h3><p>Remove <code>suffixes<\/code> one by one in order, if they are at the end of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-chop-suffixes '(\"_test.js\" \"-test.js\" \"Test.js\") \"penguin-test.js\") ;; =&gt; \"penguin\"\n(s-chop-suffixes '(\"\\r\" \"\\n\") \"penguin\\r\\n\") ;; =&gt; \"penguin\\r\"\n(s-chop-suffixes '(\"\\n\" \"\\r\") \"penguin\\r\\n\") ;; =&gt; \"penguin\"<\/pre><\/div><h3><a id=\"user-content-s-chop-prefix-prefix-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-chop-prefix-prefix-s\"><\/a>s-chop-prefix <code>(prefix s)<\/code><\/h3><p>Remove <code>prefix<\/code> if it is at the start of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-chop-prefix \"\/tmp\" \"\/tmp\/file.js\") ;; =&gt; \"\/file.js\"\n(s-chop-prefix \"\/tmp\" \"\/tmp\/tmp\/file.js\") ;; =&gt; \"\/tmp\/file.js\"<\/pre><\/div><h3><a id=\"user-content-s-chop-prefixes-prefixes-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-chop-prefixes-prefixes-s\"><\/a>s-chop-prefixes <code>(prefixes s)<\/code><\/h3><p>Remove <code>prefixes<\/code> one by one in order, if they are at the start of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-chop-prefixes '(\"\/tmp\" \"\/my\") \"\/tmp\/my\/file.js\") ;; =&gt; \"\/file.js\"\n(s-chop-prefixes '(\"\/my\" \"\/tmp\") \"\/tmp\/my\/file.js\") ;; =&gt; \"\/my\/file.js\"<\/pre><\/div><h3><a id=\"user-content-s-shared-start-s1-s2\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-shared-start-s1-s2\"><\/a>s-shared-start <code>(s1 s2)<\/code><\/h3><p>Returns the longest prefix <code>s1<\/code> and <code>s2<\/code> have in common.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-shared-start \"bar\" \"baz\") ;; =&gt; \"ba\"\n(s-shared-start \"foobar\" \"foo\") ;; =&gt; \"foo\"\n(s-shared-start \"bar\" \"foo\") ;; =&gt; \"\"<\/pre><\/div><h3><a id=\"user-content-s-shared-end-s1-s2\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-shared-end-s1-s2\"><\/a>s-shared-end <code>(s1 s2)<\/code><\/h3><p>Returns the longest suffix <code>s1<\/code> and <code>s2<\/code> have in common.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-shared-end \"bar\" \"var\") ;; =&gt; \"ar\"\n(s-shared-end \"foo\" \"foo\") ;; =&gt; \"foo\"\n(s-shared-end \"bar\" \"foo\") ;; =&gt; \"\"<\/pre><\/div><h3><a id=\"user-content-s-repeat-num-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-repeat-num-s\"><\/a>s-repeat <code>(num s)<\/code><\/h3><p>Make a string of <code>s<\/code> repeated <code>num<\/code> times.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-repeat 10 \" \") ;; =&gt; \"          \"\n(s-concat (s-repeat 8 \"Na\") \" Batman!\") ;; =&gt; \"NaNaNaNaNaNaNaNa Batman!\"<\/pre><\/div><h3><a id=\"user-content-s-concat-rest-strings\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-concat-rest-strings\"><\/a>s-concat <code>(&amp;rest strings)<\/code><\/h3><p>Join all the string arguments into one string.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-concat \"abc\" \"def\" \"ghi\") ;; =&gt; \"abcdefghi\"<\/pre><\/div><h3><a id=\"user-content-s-prepend-prefix-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-prepend-prefix-s\"><\/a>s-prepend <code>(prefix s)<\/code><\/h3><p>Concatenate <code>prefix<\/code> and <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-prepend \"abc\" \"def\") ;; =&gt; \"abcdef\"<\/pre><\/div><h3><a id=\"user-content-s-append-suffix-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-append-suffix-s\"><\/a>s-append <code>(suffix s)<\/code><\/h3><p>Concatenate <code>s<\/code> and <code>suffix<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-append \"abc\" \"def\") ;; =&gt; \"defabc\"<\/pre><\/div><h3><a id=\"user-content-s-lines-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-lines-s\"><\/a>s-lines <code>(s)<\/code><\/h3><p>Splits <code>s<\/code> into a list of strings on newline characters.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-lines \"abc\\ndef\\nghi\") ;; =&gt; '(\"abc\" \"def\" \"ghi\")\n(s-lines \"abc\\rdef\\rghi\") ;; =&gt; '(\"abc\" \"def\" \"ghi\")\n(s-lines \"abc\\r\\ndef\\r\\nghi\") ;; =&gt; '(\"abc\" \"def\" \"ghi\")<\/pre><\/div><h3><a id=\"user-content-s-match-regexp-s-optional-start\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-match-regexp-s-optional-start\"><\/a>s-match <code>(regexp s &amp;optional start)<\/code><\/h3><p>When the given expression matches the string, this function returns a list of the whole matching string and a string for each matched subexpressions. If it did not match the returned value is an empty list (nil).<\/p><p>When <code>start<\/code> is non-nil the search will start at that index.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-match \"^def\" \"abcdefg\") ;; =&gt; nil\n(s-match \"^abc\" \"abcdefg\") ;; =&gt; '(\"abc\")\n(s-match \"^\/.*\/\\\\([a-z]+\\\\)\\\\.\\\\([a-z]+\\\\)\" \"\/some\/weird\/file.html\") ;; =&gt; '(\"\/some\/weird\/file.html\" \"file\" \"html\")<\/pre><\/div><h3><a id=\"user-content-s-match-strings-all-regex-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-match-strings-all-regex-string\"><\/a>s-match-strings-all <code>(regex string)<\/code><\/h3><p>Return a list of matches for <code>regex<\/code> in <code>string<\/code>.<\/p><p>Each element itself is a list of matches, as per <code>match-string<\/code>. Multiple matches at the same position will be ignored after the first.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-match-strings-all \"{\\\\([^}]+\\\\)}\" \"x is {x} and y is {y}\") ;; =&gt; '((\"{x}\" \"x\") (\"{y}\" \"y\"))\n(s-match-strings-all \"ab.\" \"abXabY\") ;; =&gt; '((\"abX\") (\"abY\"))\n(s-match-strings-all \"\\\\&lt;\" \"foo bar baz\") ;; =&gt; '((\"\") (\"\") (\"\"))<\/pre><\/div><h3><a id=\"user-content-s-matched-positions-all-regexp-string-optional-subexp-depth\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-matched-positions-all-regexp-string-optional-subexp-depth\"><\/a>s-matched-positions-all <code>(regexp string &amp;optional subexp-depth)<\/code><\/h3><p>Return a list of matched positions for <code>regexp<\/code> in <code>string<\/code>. <code>subexp-depth<\/code> is 0 by default.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-matched-positions-all \"l+\" \"{{Hello}} World, {{Emacs}}!\" 0) ;; =&gt; '((4 . 6) (13 . 14))\n(s-matched-positions-all \"{{\\\\(.+?\\\\)}}\" \"{{Hello}} World, {{Emacs}}!\" 0) ;; =&gt; '((0 . 9) (17 . 26))\n(s-matched-positions-all \"{{\\\\(.+?\\\\)}}\" \"{{Hello}} World, {{Emacs}}!\" 1) ;; =&gt; '((2 . 7) (19 . 24))<\/pre><\/div><h3><a id=\"user-content-s-slice-at-regexp-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-slice-at-regexp-s\"><\/a>s-slice-at <code>(regexp s)<\/code><\/h3><p>Slices <code>s<\/code> up at every index matching <code>regexp<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-slice-at \"-\" \"abc\") ;; =&gt; '(\"abc\")\n(s-slice-at \"-\" \"abc-def\") ;; =&gt; '(\"abc\" \"-def\")\n(s-slice-at \"[.#]\" \"abc.def.ghi#id\") ;; =&gt; '(\"abc\" \".def\" \".ghi\" \"#id\")<\/pre><\/div><h3><a id=\"user-content-s-split-separator-s-optional-omit-nulls\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-split-separator-s-optional-omit-nulls\"><\/a>s-split <code>(separator s &amp;optional omit-nulls)<\/code><\/h3><p>Split <code>s<\/code> into substrings bounded by matches for regexp <code>separator<\/code>. If <code>omit-nulls<\/code> is non-nil, zero-length substrings are omitted.<\/p><p>This is a simple wrapper around the built-in <code>split-string<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-split \"|\" \"a|bc|12|3\") ;; =&gt; '(\"a\" \"bc\" \"12\" \"3\")\n(s-split \":\" \"a,c,d\") ;; =&gt; '(\"a,c,d\")\n(s-split \"\\n\" \"z\\nefg\\n\") ;; =&gt; '(\"z\" \"efg\" \"\")<\/pre><\/div><h3><a id=\"user-content-s-split-up-to-separator-s-n-optional-omit-nulls\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-split-up-to-separator-s-n-optional-omit-nulls\"><\/a>s-split-up-to <code>(separator s n &amp;optional omit-nulls)<\/code><\/h3><p>Split <code>s<\/code> up to <code>n<\/code> times into substrings bounded by matches for regexp <code>separator<\/code>.<\/p><p>If <code>omit-nulls<\/code> is non-nil, zero-length substrings are omitted.<\/p><p>See also <code>s-split<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-split-up-to \"\\\\s-*-\\\\s-*\" \"Author - Track-number-one\" 1) ;; =&gt; '(\"Author\" \"Track-number-one\")\n(s-split-up-to \"\\\\s-*-\\\\s-*\" \"Author - Track-number-one\" 2) ;; =&gt; '(\"Author\" \"Track\" \"number-one\")\n(s-split-up-to \"|\" \"foo||bar|baz|qux\" 3 t) ;; =&gt; '(\"foo\" \"bar\" \"baz|qux\")<\/pre><\/div><h3><a id=\"user-content-s-join-separator-strings\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-join-separator-strings\"><\/a>s-join <code>(separator strings)<\/code><\/h3><p>Join all the strings in <code>strings<\/code> with <code>separator<\/code> in between.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-join \"+\" '(\"abc\" \"def\" \"ghi\")) ;; =&gt; \"abc+def+ghi\"\n(s-join \"\\n\" '(\"abc\" \"def\" \"ghi\")) ;; =&gt; \"abc\\ndef\\nghi\"<\/pre><\/div><h3><a id=\"user-content-s-equals-s1-s2\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-equals-s1-s2\"><\/a>s-equals? <code>(s1 s2)<\/code><\/h3><p>Is <code>s1<\/code> equal to <code>s2<\/code>?<\/p><p>This is a simple wrapper around the built-in <code>string-equal<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-equals? \"abc\" \"ABC\") ;; =&gt; nil\n(s-equals? \"abc\" \"abc\") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-s-less-s1-s2\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-less-s1-s2\"><\/a>s-less? <code>(s1 s2)<\/code><\/h3><p>Is <code>s1<\/code> less than <code>s2<\/code>?<\/p><p>This is a simple wrapper around the built-in <code>string-lessp<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-less? \"abc\" \"abd\") ;; =&gt; t\n(s-less? \"abd\" \"abc\") ;; =&gt; nil\n(s-less? \"abc\" \"abc\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-matches-regexp-s-optional-start\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-matches-regexp-s-optional-start\"><\/a>s-matches? <code>(regexp s &amp;optional start)<\/code><\/h3><p>Does <code>regexp<\/code> match <code>s<\/code>? If <code>start<\/code> is non-nil the search starts at that index.<\/p><p>This is a simple wrapper around the built-in <code>string-match-p<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-matches? \"^[0-9]+$\" \"123\") ;; =&gt; t\n(s-matches? \"^[0-9]+$\" \"a123\") ;; =&gt; nil\n(s-matches? \"1\" \"1a\" 1) ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-blank-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-blank-s\"><\/a>s-blank? <code>(s)<\/code><\/h3><p>Is <code>s<\/code> nil or the empty string?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-blank? \"\") ;; =&gt; t\n(s-blank? nil) ;; =&gt; t\n(s-blank? \" \") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-present-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-present-s\"><\/a>s-present? <code>(s)<\/code><\/h3><p>Is <code>s<\/code> anything but nil or the empty string?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-present? \"\") ;; =&gt; nil\n(s-present? nil) ;; =&gt; nil\n(s-present? \" \") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-s-ends-with-suffix-s-optional-ignore-case\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-ends-with-suffix-s-optional-ignore-case\"><\/a>s-ends-with? <code>(suffix s &amp;optional ignore-case)<\/code><\/h3><p>Does <code>s<\/code> end with <code>suffix<\/code>?<\/p><p>If <code>ignore-case<\/code> is non-nil, the comparison is done without paying attention to case differences.<\/p><p>Alias: <code>s-suffix?<\/code><\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-ends-with? \".md\" \"readme.md\") ;; =&gt; t\n(s-ends-with? \".MD\" \"readme.md\") ;; =&gt; nil\n(s-ends-with? \".MD\" \"readme.md\" t) ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-s-starts-with-prefix-s-optional-ignore-case\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-starts-with-prefix-s-optional-ignore-case\"><\/a>s-starts-with? <code>(prefix s &amp;optional ignore-case)<\/code><\/h3><p>Does <code>s<\/code> start with <code>prefix<\/code>?<\/p><p>If <code>ignore-case<\/code> is non-nil, the comparison is done without paying attention to case differences.<\/p><p>Alias: <code>s-prefix?<\/code>. This is a simple wrapper around the built-in <code>string-prefix-p<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-starts-with? \"lib\/\" \"lib\/file.js\") ;; =&gt; t\n(s-starts-with? \"LIB\/\" \"lib\/file.js\") ;; =&gt; nil\n(s-starts-with? \"LIB\/\" \"lib\/file.js\" t) ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-s-contains-needle-s-optional-ignore-case\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-contains-needle-s-optional-ignore-case\"><\/a>s-contains? <code>(needle s &amp;optional ignore-case)<\/code><\/h3><p>Does <code>s<\/code> contain <code>needle<\/code>?<\/p><p>If <code>ignore-case<\/code> is non-nil, the comparison is done without paying attention to case differences.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-contains? \"file\" \"lib\/file.js\") ;; =&gt; t\n(s-contains? \"nope\" \"lib\/file.js\") ;; =&gt; nil\n(s-contains? \"^a\" \"it's not ^a regexp\") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-s-lowercase-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-lowercase-s\"><\/a>s-lowercase? <code>(s)<\/code><\/h3><p>Are all the letters in <code>s<\/code> in lower case?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-lowercase? \"file\") ;; =&gt; t\n(s-lowercase? \"File\") ;; =&gt; nil\n(s-lowercase? \"fil\u00e4\") ;; =&gt; t<\/pre><\/div><h3><a id=\"user-content-s-uppercase-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-uppercase-s\"><\/a>s-uppercase? <code>(s)<\/code><\/h3><p>Are all the letters in <code>s<\/code> in upper case?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-uppercase? \"HULK SMASH\") ;; =&gt; t\n(s-uppercase? \"Bruce no smash\") ;; =&gt; nil\n(s-uppercase? \"F\u00f6B\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-mixedcase-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-mixedcase-s\"><\/a>s-mixedcase? <code>(s)<\/code><\/h3><p>Are there both lower case and upper case letters in <code>s<\/code>?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-mixedcase? \"HULK SMASH\") ;; =&gt; nil\n(s-mixedcase? \"Bruce no smash\") ;; =&gt; t\n(s-mixedcase? \"BR\u00dcCE\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-capitalized-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-capitalized-s\"><\/a>s-capitalized? <code>(s)<\/code><\/h3><p>In <code>s<\/code>, is the first letter upper case, and all other letters lower case?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-capitalized? \"Capitalized\") ;; =&gt; t\n(s-capitalized? \"I am capitalized\") ;; =&gt; t\n(s-capitalized? \"I Am Titleized\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-numeric-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-numeric-s\"><\/a>s-numeric? <code>(s)<\/code><\/h3><p>Is <code>s<\/code> a number?<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-numeric? \"123\") ;; =&gt; t\n(s-numeric? \"onetwothree\") ;; =&gt; nil\n(s-numeric? \"7a\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-replace-old-new-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-replace-old-new-s\"><\/a>s-replace <code>(old new s)<\/code><\/h3><p>Replaces <code>old<\/code> with <code>new<\/code> in <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-replace \"file\" \"nope\" \"lib\/file.js\") ;; =&gt; \"lib\/nope.js\"\n(s-replace \"^a\" \"\\\\1\" \"it's not ^a regexp\") ;; =&gt; \"it's not \\\\1 regexp\"<\/pre><\/div><h3><a id=\"user-content-s-replace-all-replacements-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-replace-all-replacements-s\"><\/a>s-replace-all <code>(replacements s)<\/code><\/h3><p><code>replacements<\/code> is a list of cons-cells. Each <code>car<\/code> is replaced with <code>cdr<\/code> in <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-replace-all '((\"lib\" . \"test\") (\"file\" . \"file_test\")) \"lib\/file.js\") ;; =&gt; \"test\/file_test.js\"\n(s-replace-all '((\"lib\" . \"test\") (\"test\" . \"lib\")) \"lib\/test.js\") ;; =&gt; \"test\/lib.js\"<\/pre><\/div><h3><a id=\"user-content-s-downcase-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-downcase-s\"><\/a>s-downcase <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to lower case.<\/p><p>This is a simple wrapper around the built-in <code>downcase<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-downcase \"ABC\") ;; =&gt; \"abc\"<\/pre><\/div><h3><a id=\"user-content-s-upcase-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-upcase-s\"><\/a>s-upcase <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to upper case.<\/p><p>This is a simple wrapper around the built-in <code>upcase<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-upcase \"abc\") ;; =&gt; \"ABC\"<\/pre><\/div><h3><a id=\"user-content-s-capitalize-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-capitalize-s\"><\/a>s-capitalize <code>(s)<\/code><\/h3><p>Convert the first word's first character to upper case and the rest to lower case in <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-capitalize \"abc DEF\") ;; =&gt; \"Abc def\"\n(s-capitalize \"abc.DEF\") ;; =&gt; \"Abc.def\"<\/pre><\/div><h3><a id=\"user-content-s-titleize-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-titleize-s\"><\/a>s-titleize <code>(s)<\/code><\/h3><p>Convert each word's first character to upper case and the rest to lower case in <code>s<\/code>.<\/p><p>This is a simple wrapper around the built-in <code>capitalize<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-titleize \"abc DEF\") ;; =&gt; \"Abc Def\"\n(s-titleize \"abc.DEF\") ;; =&gt; \"Abc.Def\"<\/pre><\/div><h3><a id=\"user-content-s-with-s-form-rest-more\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-with-s-form-rest-more\"><\/a>s-with <code>(s form &amp;rest more)<\/code><\/h3><p>Threads <code>s<\/code> through the forms. Inserts <code>s<\/code> as the last item in the first form, making a list of it if it is not a list already. If there are more forms, inserts the first form as the last item in second form, etc.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-with \"   hulk smash   \" s-trim s-upcase) ;; =&gt; \"HULK SMASH\"\n(s-with \"My car is a Toyota\" (s-replace \"car\" \"name\") (s-replace \"a Toyota\" \"Bond\") (s-append \", James Bond\")) ;; =&gt; \"My name is Bond, James Bond\"\n(s-with \"abc \\ndef  \\nghi\" s-lines (mapcar 's-trim) (s-join \"-\") s-reverse) ;; =&gt; \"ihg-fed-cba\"<\/pre><\/div><h3><a id=\"user-content-s-index-of-needle-s-optional-ignore-case\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-index-of-needle-s-optional-ignore-case\"><\/a>s-index-of <code>(needle s &amp;optional ignore-case)<\/code><\/h3><p>Returns first index of <code>needle<\/code> in <code>s<\/code>, or nil.<\/p><p>If <code>ignore-case<\/code> is non-nil, the comparison is done without paying attention to case differences.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-index-of \"abc\" \"abcdef\") ;; =&gt; 0\n(s-index-of \"CDE\" \"abcdef\" t) ;; =&gt; 2\n(s-index-of \"n.t\" \"not a regexp\") ;; =&gt; nil<\/pre><\/div><h3><a id=\"user-content-s-reverse-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-reverse-s\"><\/a>s-reverse <code>(s)<\/code><\/h3><p>Return the reverse of <code>s<\/code>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-reverse \"abc\") ;; =&gt; \"cba\"\n(s-reverse \"ab xyz\") ;; =&gt; \"zyx ba\"\n(s-reverse \"\") ;; =&gt; \"\"<\/pre><\/div><h3><a id=\"user-content-s-presence-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-presence-s\"><\/a>s-presence <code>(s)<\/code><\/h3><p>Return <code>s<\/code> if it's <code>s-present?<\/code>, otherwise return nil.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-presence nil) ;; =&gt; nil\n(s-presence \"\") ;; =&gt; nil\n(s-presence \"foo\") ;; =&gt; \"foo\"<\/pre><\/div><h3><a id=\"user-content-s-format-template-replacer-optional-extra\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-format-template-replacer-optional-extra\"><\/a>s-format <code>(template replacer &amp;optional extra)<\/code><\/h3><p>Format <code>template<\/code> with the function <code>replacer<\/code>.<\/p><p><code>replacer<\/code> takes an argument of the format variable and optionally an extra argument which is the <code>extra<\/code> value from the call to <code>s-format<\/code>.<\/p><p>Several standard <code>s-format<\/code> helper functions are recognized and adapted for this:<\/p><pre>(s-format \"${name}\" 'gethash hash-table)\n(s-format \"${name}\" 'aget alist)\n(s-format \"$0\" 'elt sequence)\n<\/pre><p>The <code>replacer<\/code> function may be used to do any other kind of transformation.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-format \"help ${name}! I'm ${malady}\" 'aget '((\"name\" . \"nic\") (\"malady\" . \"on fire\"))) ;; =&gt; \"help nic! I'm on fire\"\n(s-format \"hello ${name}, nice day\" (lambda (var-name) \"nic\")) ;; =&gt; \"hello nic, nice day\"\n(s-format \"hello $0, nice $1\" 'elt '(\"nic\" \"day\")) ;; =&gt; \"hello nic, nice day\"<\/pre><\/div><h3><a id=\"user-content-s-lex-format-format-str\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-lex-format-format-str\"><\/a>s-lex-format <code>(format-str)<\/code><\/h3><p><code>s-format<\/code> with the current environment.<\/p><p><code>format-str<\/code> may use the <code>s-format<\/code> variable reference to refer to any variable:<\/p><p>(let ((x 1)) (s-lex-format \"x is: ${x}\"))<\/p><p>The values of the variables are interpolated with \"%s\" unless the variable <code>s-lex-value-as-lisp<\/code> is <code>t<\/code> and then they are interpolated with \"%S\".<\/p><div class=\"highlight highlight-source-lisp\"><pre>(let ((x 1)) (s-lex-format \"x is ${x}\")) ;; =&gt; \"x is 1\"\n(let ((str1 \"this\") (str2 \"that\")) (s-lex-format \"${str1} and ${str2}\")) ;; =&gt; \"this and that\"\n(let ((foo \"Hello\\\\nWorld\")) (s-lex-format \"${foo}\")) ;; =&gt; \"Hello\\\\nWorld\"<\/pre><\/div><h3><a id=\"user-content-s-count-matches-regexp-s-optional-start-end\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-count-matches-regexp-s-optional-start-end\"><\/a>s-count-matches <code>(regexp s &amp;optional start end)<\/code><\/h3><p>Count occurrences of <code>regexp<\/code> in `s'.<\/p><p><code>start<\/code>, inclusive, and <code>end<\/code>, exclusive, delimit the part of <code>s<\/code> to match.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-count-matches \"a\" \"aba\") ;; =&gt; 2\n(s-count-matches \"a\" \"aba\" 0 2) ;; =&gt; 1\n(s-count-matches \"\\\\w\\\\{2\\\\}[0-9]+\" \"ab1bab2frobinator\") ;; =&gt; 2<\/pre><\/div><h3><a id=\"user-content-s-wrap-s-prefix-optional-suffix\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-wrap-s-prefix-optional-suffix\"><\/a>s-wrap <code>(s prefix &amp;optional suffix)<\/code><\/h3><p>Wrap string <code>s<\/code> with <code>prefix<\/code> and optionally <code>suffix<\/code>.<\/p><p>Return string <code>s<\/code> with <code>prefix<\/code> prepended. If <code>suffix<\/code> is present, it is appended, otherwise <code>prefix<\/code> is used as both prefix and suffix.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-wrap \"foo\" \"\\\"\") ;; =&gt; \"\\\"foo\\\"\"\n(s-wrap \"foo\" \"(\" \")\") ;; =&gt; \"(foo)\"\n(s-wrap \"foo\" \"bar\") ;; =&gt; \"barfoobar\"<\/pre><\/div><h3><a id=\"user-content-s-split-words-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-split-words-s\"><\/a>s-split-words <code>(s)<\/code><\/h3><p>Split <code>s<\/code> into list of words.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-split-words \"under_score\") ;; =&gt; '(\"under\" \"score\")\n(s-split-words \"some-dashed-words\") ;; =&gt; '(\"some\" \"dashed\" \"words\")\n(s-split-words \"evenCamelCase\") ;; =&gt; '(\"even\" \"Camel\" \"Case\")<\/pre><\/div><h3><a id=\"user-content-s-lower-camel-case-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-lower-camel-case-s\"><\/a>s-lower-camel-case <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to lowerCamelCase.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-lower-camel-case \"some words\") ;; =&gt; \"someWords\"\n(s-lower-camel-case \"dashed-words\") ;; =&gt; \"dashedWords\"\n(s-lower-camel-case \"under_scored_words\") ;; =&gt; \"underScoredWords\"<\/pre><\/div><h3><a id=\"user-content-s-upper-camel-case-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-upper-camel-case-s\"><\/a>s-upper-camel-case <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to UpperCamelCase.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-upper-camel-case \"some words\") ;; =&gt; \"SomeWords\"\n(s-upper-camel-case \"dashed-words\") ;; =&gt; \"DashedWords\"\n(s-upper-camel-case \"under_scored_words\") ;; =&gt; \"UnderScoredWords\"<\/pre><\/div><h3><a id=\"user-content-s-snake-case-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-snake-case-s\"><\/a>s-snake-case <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to snake_case.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-snake-case \"some words\") ;; =&gt; \"some_words\"\n(s-snake-case \"dashed-words\") ;; =&gt; \"dashed_words\"\n(s-snake-case \"camelCasedWords\") ;; =&gt; \"camel_cased_words\"<\/pre><\/div><h3><a id=\"user-content-s-dashed-words-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-dashed-words-s\"><\/a>s-dashed-words <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to dashed-words.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-dashed-words \"some words\") ;; =&gt; \"some-words\"\n(s-dashed-words \"under_scored_words\") ;; =&gt; \"under-scored-words\"\n(s-dashed-words \"camelCasedWords\") ;; =&gt; \"camel-cased-words\"<\/pre><\/div><h3><a id=\"user-content-s-capitalized-words-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-capitalized-words-s\"><\/a>s-capitalized-words <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to Capitalized words.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-capitalized-words \"some words\") ;; =&gt; \"Some words\"\n(s-capitalized-words \"under_scored_words\") ;; =&gt; \"Under scored words\"\n(s-capitalized-words \"camelCasedWords\") ;; =&gt; \"Camel cased words\"<\/pre><\/div><h3><a id=\"user-content-s-titleized-words-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-titleized-words-s\"><\/a>s-titleized-words <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to Titleized Words.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-titleized-words \"some words\") ;; =&gt; \"Some Words\"\n(s-titleized-words \"under_scored_words\") ;; =&gt; \"Under Scored Words\"\n(s-titleized-words \"camelCasedWords\") ;; =&gt; \"Camel Cased Words\"<\/pre><\/div><h3><a id=\"user-content-s-word-initials-s\" class=\"anchor\" aria-hidden=\"true\" href=\"#s-word-initials-s\"><\/a>s-word-initials <code>(s)<\/code><\/h3><p>Convert <code>s<\/code> to its initials.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(s-word-initials \"some words\") ;; =&gt; \"sw\"\n(s-word-initials \"under_scored_words\") ;; =&gt; \"usw\"\n(s-word-initials \"camelCasedWords\") ;; =&gt; \"cCW\"<\/pre><\/div><h2><a id=\"user-content-whats-with-the-built-in-wrappers\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-with-the-built-in-wrappers\"><\/a>What's with the built-in wrappers?<\/h2><p>Imagine looking through the function list and seeing <code>s-ends-with?<\/code>, but <code>s-starts-with?<\/code> is nowhere to be found. Why? Well, because Emacs already has <code>string-prefix-p<\/code>. Now you're starting out slightly confused, then have to go somewhere else to dig for the command you were looking for.<\/p><p>The wrapping functions serve as both documentation for existing functions and makes for a consistent API.<\/p><h2><a id=\"user-content-other-string-related-libraries\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-string-related-libraries\"><\/a>Other string related libraries<\/h2><ul><li>\n<p><a href=\"https:\/\/github.com\/eschulte\/jump.el\/blob\/master\/inflections.el\">inflections<\/a> package provides functions for strings pluralization and singularization.<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/emacswiki.org\/emacs\/levenshtein.el\" rel=\"nofollow\">levenshtein<\/a> package provides a function to calculate the Levenshtein distance between two strings.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/rolandwalker\/string-utils\">string-utils<\/a> is another general string manipulation library.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-changelist\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelist\"><\/a>Changelist<\/h2><h3><a id=\"user-content-from-1110-to-1120\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-1110-to-1120\"><\/a>From 1.11.0 to 1.12.0<\/h3><ul><li>Alias all functions ending in <code>?<\/code> (Tianxiang Xiong)<\/li>\n<li>Add <code>s-blank-str?<\/code> (Aborn Jiang)<\/li>\n<li>Several bugfixes<\/li>\n<\/ul><h3><a id=\"user-content-from-1100-to-1110\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-1100-to-1110\"><\/a>From 1.10.0 to 1.11.0<\/h3><ul><li>Add <code>s-matched-positions-all<\/code> (ono hiroko)<\/li>\n<\/ul><h3><a id=\"user-content-from-190-to-1100\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-190-to-1100\"><\/a>From 1.9.0 to 1.10.0<\/h3><ul><li>Add <code>s-wrap<\/code> (Johan Andersson)<\/li>\n<li>Add <code>s-split-up-to<\/code> (Matus Goljer)<\/li>\n<li>Fix <code>s-reverse<\/code> for Unicode combining characters. (Christopher Wellons)<\/li>\n<\/ul><h3><a id=\"user-content-from-180-to-190\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-180-to-190\"><\/a>From 1.8.0 to 1.9.0<\/h3><ul><li>Add <code>s-count-matches<\/code> (Lars Andersen)<\/li>\n<\/ul><h3><a id=\"user-content-from-170-to-180\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-170-to-180\"><\/a>From 1.7.0 to 1.8.0<\/h3><ul><li>Add <code>s-present?<\/code> and <code>s-present?<\/code> (Johan Andersson)<\/li>\n<li>Better handling of international characters<\/li>\n<\/ul><h3><a id=\"user-content-from-160-to-170\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-160-to-170\"><\/a>From 1.6.0 to 1.7.0<\/h3><ul><li>Add <code>s-word-initials<\/code> (Sylvain Rousseau)<\/li>\n<li>Better handling of camel cased strings (@Bruce-Connor)<\/li>\n<\/ul><h3><a id=\"user-content-from-150-to-160\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-150-to-160\"><\/a>From 1.5.0 to 1.6.0<\/h3><ul><li>Add <code>s-pad-left<\/code> and <code>s-pad-right<\/code><\/li>\n<li>Bugfixes for <code>s-format<\/code> (Nic Ferrier)<\/li>\n<\/ul><h3><a id=\"user-content-from-140-to-150\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-140-to-150\"><\/a>From 1.4.0 to 1.5.0<\/h3><ul><li>Add <code>s-all-match-strings<\/code> (Geoff Gole)<\/li>\n<li>Add <code>s-lex-format<\/code> (Nic Ferrier)<\/li>\n<\/ul><h3><a id=\"user-content-from-131-to-140\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-131-to-140\"><\/a>From 1.3.1 to 1.4.0<\/h3><ul><li>Add <code>s-capitalized?<\/code><\/li>\n<li>Add <code>s-replace-all<\/code><\/li>\n<li>Add <code>s-slice-at<\/code><\/li>\n<li>Add <code>s-split<\/code> alias for <code>split-string<\/code> (R\u00fcdiger Sonderfeld)<\/li>\n<li>Add <code>s-less?<\/code> predicate (R\u00fcdiger Sonderfeld)<\/li>\n<li>Add START parameter to <code>s-matches?<\/code> (R\u00fcdiger Sonderfeld)<\/li>\n<li>Bugfixes<\/li>\n<\/ul><h3><a id=\"user-content-from-130-to-131\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-130-to-131\"><\/a>From 1.3.0 to 1.3.1<\/h3><ul><li>Add <code>s-numeric?<\/code><\/li>\n<li>Add <code>s-match<\/code> (Arthur Andersen)<\/li>\n<li>Add <code>s-format<\/code> (Nic Ferrier)<\/li>\n<li>Move .el files out of root to avoid problems with require.<\/li>\n<\/ul><h3><a id=\"user-content-from-121-to-130\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-121-to-130\"><\/a>From 1.2.1 to 1.3.0<\/h3><ul><li>\n<p><strong>Breaking change:<\/strong> <code>s-capitalize<\/code> now converts the first word's first character to upper case and the rest to lower case. <code>s-titleize<\/code> works like the old <code>s-capitalize<\/code> and capitalizes each word. (Johan Andersson)<\/p>\n<\/li>\n<li>\n<p><code>s-capitalized-words<\/code> and <code>s-titleized-words<\/code> mirror this change.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-contributors\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributors\"><\/a>Contributors<\/h2><p>Thanks!<\/p><h2><a id=\"user-content-contribute\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribute\"><\/a>Contribute<\/h2><p>Yes, please do. Pure functions in the string manipulation realm only, please. There's a suite of tests in <code>dev\/examples.el<\/code>, so remember to add tests for your function, or I might break it later.<\/p><p>You'll find the repo at:<\/p><pre>https:\/\/github.com\/magnars\/s.el\n<\/pre><p>Run the tests with<\/p><pre>.\/run-tests.sh\n<\/pre><p>Create the docs with<\/p><pre>.\/create-docs.sh\n<\/pre><p>I highly recommend that you install these as a pre-commit hook, so that the tests are always running and the docs are always in sync:<\/p><pre>cp pre-commit.sh .git\/hooks\/pre-commit\n<\/pre><p>Oh, and don't edit <code>README.md<\/code> directly, it is auto-generated. Change <code>readme-template.md<\/code> or <code>examples-to-docs.el<\/code> instead.<\/p><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2><p>Copyright (C) 2012-2015 Magnar Sveen<\/p><p>Authors: Magnar Sveen <a href=\"mailto:magnars@gmail.com\">magnars@gmail.com<\/a> Keywords: strings<\/p><p>This program is free software; you can redistribute it and\/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.<\/p><p>This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.<\/p><p>You should have received a copy of the GNU General Public License along with this program. If not, see <a href=\"http:\/\/www.gnu.org\/licenses\/\" rel=\"nofollow\">http:\/\/www.gnu.org\/licenses\/<\/a>.<\/p>","created_at":"2019-12-25T03:44:05+0000","updated_at":"2019-12-25T03:44:05+0000","published_by":["magnars"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":19,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/265785?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1071,"title":"Death to the Shell","url":"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-transcript.html","content":"<p>Remember how I mention that we use shell in two distinct ways, interactively and programmatically. I\u2019ve shown my idea for the first, let\u2019s chat about the second.<\/p><p>While I\u2019ve been thinking about this for many years, not long ago, I came across a Reddit Question:<\/p><figure><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/piper-presentation-a.png\" alt=\"Lisp people: What's your go-to language for when you want to write a quick script. Python seems to be the default for me, but I do genuinely find Lisp joyful to write in.\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>To which I responded that as a Lisp, Emacs Lisp is great as integrating any good idea that comes along, and stealing good ideas has made this system from the twentieth century still useful and almost modern.<\/p><figure><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/piper-presentation-b.png\" alt=\"After many years of shell and Perl scripts that grew in complexity, I switch to writing scripts where it is easiest to call, Emacs Lisp. With modern libraries and completing libraries like Helm and Ivy, I just found a nice go to language.\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><figure id=\"orgd91c85c\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-12.png\" alt=\"I skipped showing a couple of slides that I just described. This slide is titled, Emacs as a Go To Scripting Language\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>The old adage that Emacs is an operating system is pretty true, it is just an operating system of functions, and as such, it really good for the hacker. You want to either overwrite or just advice every function. Global variables, normally a bad thing when you are sharing code with others, turns out to be really easy to prototype ideas when it is just you.<\/p><p>But think about all the really great libraries and support that we have that are new. I didn\u2019t have these in the 1980s:<\/p><ul class=\"org-ul\"><li>Common Lisp library<\/li>\n<li><a href=\"https:\/\/github.com\/magnars\/dash.el\">Dash library<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/magnars\/s.el\">S library<\/a> for strings<\/li>\n<li><a href=\"https:\/\/github.com\/rejeep\/f.el\">F Library<\/a> for files<\/li>\n<li>Web-focused Libraries for JSON, Yaml, XML<\/li>\n<\/ul><p>But I think the most compelling argument for using Emacs functions instead of shell scripts is the improved user interface. I\u2019m serious. Think about all those command line options to each command you use, and how no amount of tab completion helps. In emacs, we can easily bind functions to keys, use two great completing systems (<a href=\"https:\/\/github.com\/abo-abo\/swiper\">Ivy<\/a> or <a href=\"https:\/\/github.com\/emacs-helm\/helm\">Helm<\/a>) that can give you history of the arguments, just not the command.<\/p><p>Finally, think of using Hydra and Magit\u2019s interface, <a href=\"https:\/\/github.com\/magit\/transient\">Transient<\/a>, to really have a flexible approach to running commands.<\/p><figure id=\"org5b922cc\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-13.png\" alt=\"Shell scripts are both good and bad\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>Shell scripts, especially small ones can be really readable, and they are good as running commands and connecting them with streams, however, any script that grows to any size seems to become completely unreadable. The other issue is that since everything is a string, we can have conversion issues into numbers or needing to escape white space. Sure, modern versions of the shell have arrays, but they\u2019re pretty kludgy at best.<\/p><p>So, can we take the best of both Lisp and Shell scripts? If we could, this could be a win for us using Emacs, as calling a <em>function<\/em> is better than a <em>script<\/em>.<\/p><figure id=\"org55b36c0\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-14.png\" alt=\"Maybe we could bring the good parts of shell scripts into Emacs.\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>What should we <em>bring over<\/em> from shells? The <code>format<\/code> function is fairly good, but the shell, along with most modern languages, have variable substitution within strings. This seems more readable. Also, since the shell has embedded itself in my DNA, I am comfortable with <code>mkdir<\/code> and <code>cp<\/code>. Besides, since Emacs doesn\u2019t have namespaces, our functions tend to be pretty lengthy, which often makes our code is little more difficult to read.<\/p><p>But when we do, we can get a slew of Lispy goodness, like data structures, better iterations, logical variable scoping, and more functions than you can shake a stick.<\/p><figure id=\"org08050f7\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-15.png\" alt=\"Desired Features\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>The magic starts with a couple of macros that will convert the <em>forms within<\/em> to be more script-like. I have five specific goals for how this macro should behave. Let me explain each of these in details.<\/p><figure id=\"orgfb3be37\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-16.png\" alt=\"Shell Executables should just Work\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>Should be easy to call a shell command, using the <code>shell<\/code> form:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(piper-script\n (shell \"ls \/tmp\"))\n<\/pre><\/div><p>This should also work well with files that have spaces:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(piper-script\n (shell \"ls\" \"~\/Google Drive File Stream\/My Drive\/\"))\n<\/pre><\/div><p>Give the <code>shell<\/code> form function multiple strings, and it will assume that all strings are parameters, and not do any parsing, but this function will split single strings on spaces.<\/p><p>How about making an alias for the <code>shell<\/code> function name?<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(piper-script\n ($ \"ls \/tmp\"))\n<\/pre><\/div><p>Looks like a prompt character.<\/p><figure id=\"orgef93771\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-17.png\" alt=\"Pipes and Data Transformations\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>I use the same technique described earlier of using a temporary buffer, and running shell commands such that the contents of the buffer become <em>standard in<\/em> to a shell command, and the results of the command replace the contents of the buffer for the next command. This allows an obvious pipe-like approach:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(piper-script\n ($ \"ls\" \"-1\" \"\/tmp\")\n (| \"grep Temp\")\n (| \"tr [a-z] [A-Z]\"))\n<\/pre><\/div><p>Keep in mind you can use any Emacs function that works with buffers.<\/p><figure id=\"org1476641\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-18.png\" alt=\"Wildcard Expansion\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>File expansion should work as expected:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(piper-script\n ($ \"ls\" \"~\/.emacs.d\/*.el\"))\n<\/pre><\/div><p>Keep in mind that the <code>piper-script<\/code> looks at <em>every string<\/em>, so to make this work, if an expansion fails to find matching files, I just return the original string.<\/p><figure id=\"org0331f2e\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-19.png\" alt=\"Embedded Strings\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>While <code>format<\/code> is fine, <em>string-variable substitution<\/em> is more readable, plus I want to be able to insert both Emacs variables as well as environment variables:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(let ((some-var \"fooey\"))\n  (piper-script\n   (let ((nudder-var \"chop\"))\n     (echo \"$HOME\/projects\/${nudder-var}\/${some-var}\"))))\n<\/pre><\/div><p>Should return:<\/p><pre class=\"example\">\n\"\/home\/howard\/projects\/chop\/fooey\"\n<\/pre><figure id=\"orga006232\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-20.png\" alt=\"Shell-Like Commands?\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>I\u2019m not convinced, but perhaps when converted shell script should have shorter, more shell-like function names. In others words, I could use the function <code>grep<\/code> instead of <code>keep-lines<\/code>. In this case, if <code>grep<\/code> were in quotes and passed to the <code>shell<\/code> function it would call out to the executable, and if not, to the Emacs function, so maybe this would be more confusing.<\/p><p>However, I do need to write <em>some functions<\/em>, like <code>sudo<\/code> is a <code>let<\/code> wrapper around a modification to convert the <code>default-directory<\/code> variable to reference Tramp.<\/p><p>Let me run through a couple of examples, and for each, I want to show my original shell script, and how I converted it.<\/p><figure id=\"org4d28751\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-21.png\" alt=\"First Example\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>First, let\u2019s see if I can get the resident memory size of two running browsers.<\/p><figure id=\"org28b9eda\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-22.png\" alt=\"First example's Original Script\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>While the <code>ps<\/code> has a lot of command line options, I\u2019m going to use grep to look for the top-level process, that is, the ones where the parent process is <code>1<\/code>, and I want to ignore all the extra service frameworks that Chrome starts:<\/p><div class=\"org-src-container\"><pre class=\"src src-shell\">ps -u $USER -o ppid,rss,command | egrep 'Chrome|Firefox' | grep '^ *1' | grep -v Frameworks\n<\/pre><\/div><figure id=\"orgea00a67\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-23.png\" alt=\"New Piper Script for First example\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>The most prominent advantage is the use of Emacs\u2019 <code>rx<\/code> macro for the regular expression (which I pass to the <code>keep-lines<\/code> alias):<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(let ((browsers      (rx (or \"Chrome\" \"Firefox\")))\n      (session-leads (rx line-start (one-or-more blank) \"1\")))\n  (piper-script\n   ($ \"ps\" \"-u $USER\" \"-o ppid,rss,command\")\n   (grep browsers)\n   (grep session-leads)\n   (grep-v \"Frameworks\"))))\n<\/pre><\/div><figure id=\"org49dfd21\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-24.png\" alt=\"Second Example\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>I\u2019ve installed Ubuntu on an old Macbook laptop, and I noticed that after closing the lid, the computer would repeated turn itself on and off. The reason is that some devices are set to wake up the system, and magic process entry can be used to both read the state of all these devices, but if you write the name of the device back to this process entry, it will turn it off.<\/p><p>Why yes, this does seem like a job for a script, eh?<\/p><figure id=\"org517f76f\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-25.png\" alt=\"Second example's original script\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>My original script uses <code>cat<\/code> to expose all the entries, which text I can massage with a little pipe sequence. Eventually storing each enabled device in a variable, and writing the value of that variable back into the process entry.<\/p><figure id=\"orgee9db29\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-26.png\" alt=\"New Piper Script for Second example\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>The <code>piper-script<\/code> version starts by looking a lot like the original, but I use a Lisp-looking looping macro I call <code>for<\/code> and a function that converts all the lines in this temporary buffer into a list of strings for it to consume:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(piper-script\n (sudo\n  (cat \"\/proc\/acpi\/wakeup\")\n  (grep \"enabled\")\n  (replace-regexp \" .*\" \"\")\n  (for (device (read-all-lines))\n    (write-into device \"\/proc\/acpi\/wakeup\"))))\n<\/pre><\/div><figure id=\"org7fd16be\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-27.png\" alt=\"Show us the Magic\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>The magic is obviously a nifty macro. I simply use a function from the dash library to analyze and possibly change every element I find in the forms given to it, without changing the structure of those forms:<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(defmacro piper-script (&amp;rest forms)\n  \"Runs the FORMS in a shell-like DSL.\"\n  (cons 'with-temp-buffer\n        (-tree-map #'piper--script-transform forms)))\n<\/pre><\/div><p>Since all command works with a buffer, we create one, run through the converted forms, and not shown here, but we return the contents of the buffer as a string.<\/p><figure id=\"org55639ce\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/piper-presentation-slides\/slide-28.png\" alt=\"Piper Script Transform\" width=\"1000\" referrerpolicy=\"no-referrer\" \/><\/figure><p>The followup magic happens with a function that returns a potentially changed version of the form. It is just a lengthy <code>cond<\/code> statement, but you can see that at first, I change strings, so that I can expand any wildcards or variable substitution. Then I just swap out various function names for their alias. The way it is currently written, I can\u2019t use any variables with matching names as I don\u2019t currently distinguish their position. Did I mention that this was a proof of concept?<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(defun piper--script-transform (element)\n  \"Helper for `piper-script' to convert forms and strings. \"\n  (cond\n   ((stringp element) `(piper-script-get-files ,element))\n   ((eq element '$) 'piper-script-shell)\n   ((eq element '|) 'piper-script-shell)\n   ((eq element 'echo) 'piper-script-echo)\n   ((eq element 'ifsh) 'piper-script-shell-if)\n   ((eq element 'for) 'piper-script-loop)\n   ((eq element 'cat) 'insert-file-contents)\n   ((eq element 'touch) 'f-touch)\n   ((eq element 'ln-s) 'f-symlink)   ; Perhaps make-symbolic-link\n   ((eq element 'mkdir) 'f-mkdir)    ; ...\n   ((eq element 'write-into) 'piper-script-write-into)\n   ((eq element 'to-clipboard) 'piper-script-to-clipboard)\n   ((eq element 'read-all-lines) 'piper-script-read-all-lines)\n   ;; ...\n   (t element)))\n<\/pre><\/div>","created_at":"2019-12-25T03:19:16+0000","updated_at":"2019-12-25T03:19:16+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"howardism.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1070,"title":"Open-Source Spotify App | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=21867728","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bransonf\" class=\"hnuser\">bransonf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868136\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=trashcat\" class=\"hnuser\">trashcat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869245\">57 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Macha\" class=\"hnuser\">Macha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869383\">26 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=snek\" class=\"hnuser\">snek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868991\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haolez\" class=\"hnuser\">haolez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868870\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868930\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jedimastert\" class=\"hnuser\">jedimastert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869002\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mendelmaleh\" class=\"hnuser\">mendelmaleh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869033\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869413\">20 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anamexis\" class=\"hnuser\">anamexis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869450\">11 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=piracy1\" class=\"hnuser\">piracy1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868358\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868486\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=fao_\" class=\"hnuser\">fao_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868765\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bathtub365\" class=\"hnuser\">bathtub365<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869020\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868918\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haecceity\" class=\"hnuser\">haecceity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869146\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=vinkelhake\" class=\"hnuser\">vinkelhake<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869435\">15 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=snissn\" class=\"hnuser\">snissn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869165\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bransonf\" class=\"hnuser\">bransonf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869214\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Macha\" class=\"hnuser\">Macha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869422\">19 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868265\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868556\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868578\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=setr\" class=\"hnuser\">setr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868934\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868959\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869358\">32 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=1wheel\" class=\"hnuser\">1wheel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868653\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dehrmann\" class=\"hnuser\">dehrmann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868681\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868431\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868476\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868562\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868676\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Bootwizard\" class=\"hnuser\">Bootwizard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869003\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869014\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868592\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868951\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869062\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bilekas\" class=\"hnuser\">bilekas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868531\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dickeytk\" class=\"hnuser\">dickeytk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868704\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=uncletaco\" class=\"hnuser\">uncletaco<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868802\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpdt\" class=\"hnuser\">cpdt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868925\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=saagarjha\" class=\"hnuser\">saagarjha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868311\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sod\" class=\"hnuser\">sod<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868391\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericlewis\" class=\"hnuser\">ericlewis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868977\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thasian\" class=\"hnuser\">thasian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21868496\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=verisimilitudes\" class=\"hnuser\">verisimilitudes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869169\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869455\">10 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=djsumdog\" class=\"hnuser\">djsumdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869444\">12 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Wowfunhappy\" class=\"hnuser\">Wowfunhappy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21869447\">12 minutes ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>","created_at":"2019-12-24T02:59:37+0000","updated_at":"2019-12-24T02:59:37+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["design","undraw","free","pictures"],"is_public":false,"id":1069,"title":"UnDraw: Open-source illustrations for every project you can imagine and create","url":"https:\/\/news.ycombinator.com\/item?id=21858822","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ignoramous\" class=\"hnuser\">ignoramous<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21859930\">20 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Vincenius\" class=\"hnuser\">Vincenius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863146\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hmhrex\" class=\"hnuser\">hmhrex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863616\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bjelkeman-again\" class=\"hnuser\">bjelkeman-again<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861971\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=netcyrax\" class=\"hnuser\">netcyrax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862708\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonquark\" class=\"hnuser\">jonquark<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21864770\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonquark\" class=\"hnuser\">jonquark<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21864831\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=vinrob92\" class=\"hnuser\">vinrob92<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861496\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thojest\" class=\"hnuser\">thojest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861590\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=EGreg\" class=\"hnuser\">EGreg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861879\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=xeeeeeeeeeeenu\" class=\"hnuser\">xeeeeeeeeeeenu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860434\">18 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=koboll\" class=\"hnuser\">koboll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860455\">18 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hinkley\" class=\"hnuser\">hinkley<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861455\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=usrusr\" class=\"hnuser\">usrusr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862652\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=amelius\" class=\"hnuser\">amelius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862961\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=prepend\" class=\"hnuser\">prepend<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863170\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bryanrasmussen\" class=\"hnuser\">bryanrasmussen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863567\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=usrusr\" class=\"hnuser\">usrusr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863790\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=resoluteteeth\" class=\"hnuser\">resoluteteeth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21864733\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gvjddbnvdrbv\" class=\"hnuser\">gvjddbnvdrbv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863622\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=guelo\" class=\"hnuser\">guelo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860692\">17 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=eitland\" class=\"hnuser\">eitland<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861926\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tyingq\" class=\"hnuser\">tyingq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863194\">5 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cmurf\" class=\"hnuser\">cmurf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860678\">17 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860846\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mc3\" class=\"hnuser\">mc3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860900\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jordigh\" class=\"hnuser\">jordigh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861508\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860950\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mc3\" class=\"hnuser\">mc3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860957\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Eikon\" class=\"hnuser\">Eikon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861367\">14 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jordigh\" class=\"hnuser\">jordigh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861515\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rambojazz\" class=\"hnuser\">rambojazz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861806\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Eikon\" class=\"hnuser\">Eikon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861884\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lopmotr\" class=\"hnuser\">lopmotr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861970\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=drusepth\" class=\"hnuser\">drusepth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21859834\">20 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860217\">19 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dorkandstormy\" class=\"hnuser\">dorkandstormy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863556\">4 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863690\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=DeadBabyOrgasm\" class=\"hnuser\">DeadBabyOrgasm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21864928\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ItsMattyG\" class=\"hnuser\">ItsMattyG<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21864574\">1 hour ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=VvR-Ox\" class=\"hnuser\">VvR-Ox<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862629\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863698\">3 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=artyomavanesov\" class=\"hnuser\">artyomavanesov<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860914\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=TeMPOraL\" class=\"hnuser\">TeMPOraL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861959\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zo1\" class=\"hnuser\">zo1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862100\">10 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=eitland\" class=\"hnuser\">eitland<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861964\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cyberferret\" class=\"hnuser\">cyberferret<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862677\">7 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=amelius\" class=\"hnuser\">amelius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862969\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=brabel\" class=\"hnuser\">brabel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21863119\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thorentis\" class=\"hnuser\">Thorentis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861293\">14 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=system2\" class=\"hnuser\">system2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861836\">11 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=elamje\" class=\"hnuser\">elamje<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861365\">14 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=masukomi\" class=\"hnuser\">masukomi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21864048\">2 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ourmandave\" class=\"hnuser\">ourmandave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860857\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mc3\" class=\"hnuser\">mc3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860871\">16 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=userbinator\" class=\"hnuser\">userbinator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861342\">14 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=flamey\" class=\"hnuser\">flamey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860108\">19 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860224\">19 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pesterazor\" class=\"hnuser\">pesterazor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862921\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Havoc\" class=\"hnuser\">Havoc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861434\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=solarkraft\" class=\"hnuser\">solarkraft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860169\">19 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anges244\" class=\"hnuser\">anges244<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860232\">19 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=blondin\" class=\"hnuser\">blondin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861246\">14 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tuukkah\" class=\"hnuser\">tuukkah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862544\">8 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=aehtyb\" class=\"hnuser\">aehtyb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21861581\">13 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=chris_engel\" class=\"hnuser\">chris_engel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21860514\">18 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cerberusss\" class=\"hnuser\">cerberusss<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21862911\">6 hours ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>","created_at":"2019-12-23T18:08:26+0000","updated_at":"2019-12-23T18:08:35+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1068,"title":"unDraw | Colorful illustrations","url":"https:\/\/undraw.co\/","content":"<div class=\"point container point__texts\"><h5 class=\"point__title\">Infinite scalability<\/h5><p class=\"point__text\">Scale images without quality change. Not just Retina but future ready.<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Tiny file size<\/h5><p class=\"point__text\">Use tens of svg images without worrying about the page load speed.<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Your own color<\/h5><p class=\"point__text\">Use cutting-edge tech to change colors for images that match your brand.<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Embeddable format<\/h5><p class=\"point__text\">Embed the code directly in your html and save your server requests.<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Naturally material<\/h5><p class=\"point__text\">Subtle shadows &amp; multi-level objects for perfect modern web designs.<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Do whatever you want<\/h5><p class=\"point__text\">Use it on commercial or personal projects without worries. Open <a href=\"https:\/\/undraw.co\/license\" class=\"link link--landing\">license<\/a>, babe!<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Editable code<\/h5><p class=\"point__text\">Mess around with colors, animate with plain css and other crazy possibilities!<\/p><\/div><div class=\"point container point__texts\"><h5 class=\"point__title\">Perfect for everything<\/h5><p class=\"point__text\">Cool designs, landing pages, mobile apps, sites or whatever you can create!<\/p><\/div>","created_at":"2019-12-23T14:13:52+0000","updated_at":"2019-12-23T14:13:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"undraw.co","preview_picture":"https:\/\/42f2671d685f51e10fc6-b9fcecea3e50b3b59bdc28dead054ebc.ssl.cf5.rackcdn.com\/assets\/share_new.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1067,"title":"Show HN: Angeldust \u2013 a fast and efficient video game","url":"https:\/\/news.ycombinator.com\/item?id=21858226","content":"<a href=\"https:\/\/news.ycombinator.com\/item?id=21858226\" class=\"storylink\">Show HN: Angeldust \u2013 a fast and efficient video game<\/a>\n149 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=ItsFirefly\" class=\"hnuser\">ItsFirefly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21858226\">3 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=21858226&amp;goto=item%3Fid%3D21858226\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Show%20HN%3A%20Angeldust%20%E2%80%93%20a%20fast%20and%20efficient%20video%20game&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Show%20HN%3A%20Angeldust%20%E2%80%93%20a%20fast%20and%20efficient%20video%20game\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=21858226&amp;auth=33097867fd9c7cd5e7c73d01e3b1ada7d6ff4485\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=21858226\">34\u00a0comments<\/a>\nAfter years of solo development I'm presenting: Angeldust, an <em>omni-compatible<\/em> action RPG and building game!\n<p>Website: <a href=\"http:\/\/angeldu.st\" rel=\"nofollow\">http:\/\/angeldu.st<\/a><\/p>\n<p>My single server hosts 250+K active players in one giant, dynamic fantasy world. Angeldust's game server is pervasively multithreaded, implementing John Carmack's dream of processing an immutable world state each tick.<\/p>\n<p>Both world handling and network traffic routing uses any number of threads. For networking, a typical game session only uses ~3Kbps of bandwidth, enabling decent play even over 2G\/EDGE networks.<\/p>\n<p>The client is programmed in C++ with bits of C, Objective-C, Java and PHP. It runs really well on very low-end and obsolete hardware, all the way up to modern systems with fast GPUs.<\/p>\n<p>Angeldust works perfectly on Windows XP+, even without hardware OpenGL; Mac OS X 10.6+, even on the first 2006 32-bit Intel GMA945 MacBook1,1; Linux 64-bit glibc 2.17+; Android v2.2.3+, supporting Bluetooth game controllers; and iOS 6+, though iTunes App Store forces iOS 8+ (any way around this?).<\/p>\n<p>I live stream weekly on YouTube and Twitch, even right now! Come ask me your development questions! I often see developers playing together with their kids.<\/p>\n<p>YouTube channel: <a href=\"https:\/\/www.youtube.com\/AngeldustLive\" rel=\"nofollow\">https:\/\/www.youtube.com\/AngeldustLive<\/a><\/p>\n<p>One more thing: the Angeldust website ( <a href=\"https:\/\/angeldu.st\" rel=\"nofollow\">https:\/\/angeldu.st<\/a> ) uses plain HTML and CSS to offer interactive, game-related actions. It's fully functional even without JavaScript enabled.<\/p>","created_at":"2019-12-22T20:00:55+0000","updated_at":"2019-12-22T20:00:55+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1066,"title":"Dungeon Generation in Enter The Gungeon","url":"https:\/\/www.boristhebrave.com\/2019\/07\/28\/dungeon-generation-in-enter-the-gungeon\/","content":"<p>I\u2019ve been playing a lot of <a href=\"https:\/\/store.steampowered.com\/app\/311690\/Enter_the_Gungeon\/\">Enter The Gungeon<\/a> recently. It\u2019s a great, brutally hard, twin stick bullet hell that reminded me a lot of <a href=\"https:\/\/store.steampowered.com\/app\/113200\/The_Binding_of_Isaac\/\">Binding of Isaac<\/a>. But as I\u2019ve been playing it more and more, I realized that the dungeon design actually shows some subtle genius.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_screenshot.jpeg\" alt=\"\" class=\"wp-image-408\" srcset=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_screenshot.jpeg 658w, https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_screenshot-300x169.jpeg 300w\" referrerpolicy=\"no-referrer\" \/><\/figure><\/div>\n<p>There are many procedural generators out there that produce sensible level layouts that manage pacing and rewards correctly, and there are other generators out there that provide levels that include loops and compact layouts. But it\u2019s hard to find both in a single game. Really, the only other game I\u2019ve heard attempting this is <a href=\"https:\/\/store.steampowered.com\/app\/506870\/Unexplored\/\">Unexplored<\/a>.<\/p>\n<p>So, naturally, I broke out the decompiler to reveal all of <em>Gungeon<\/em>\u2018s secrets to me. I\u2019ll share with you what I found.<\/p>\n<hr class=\"wp-block-separator\" \/><p>On the face of it, <em>Gungeon<\/em>\u2018s levels look rather straightforward. Here\u2019s a typical map.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle.png\" alt=\"\" class=\"wp-image-390\" srcset=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle.png 660w, https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle-253x300.png 253w\" referrerpolicy=\"no-referrer\" \/><figcaption>The map from a typical level 1<\/figcaption><\/figure><\/div>\n<p>The player starts at the entrance <img class=\"wp-image-392 c1\" src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_entrance.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/>, swings by the chests <img class=\"wp-image-393 c2\" src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_chest_blue.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/><img class=\"wp-image-394 c3\" src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_chest_black.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/> and the shopkeeper <img class=\"wp-image-395 c4\" src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_shopkeeper.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/> to get some loot, before finally defeating the boss<img class=\"wp-image-396 c5\" src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_boss.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/>. Teleporters can be used to revisit rooms quickly, if you\u2019ve gone down a long dead end. The individual rooms themselves are clearly pre-authored, and populated with a wide variety of enemies to gun down. As you move onto later stages in the game, you will see this pattern over and over again, just with larger stages.<\/p>\n<p>So far this seems unremarkable. You can find a lot of level generators online that jam together a bunch of random rooms with connections between them.<\/p>\n<p>What\u2019s special becomes apparent as you play. The levels\u2026 just feel a bit more planned than mere chance would indicate. The boss room is always a reasonable distance from the start. Rooms with enemies are always reasonably spaced with breather rooms, shops and hubs. And most of all, many of the chests are hidden behind one way loops.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle_detail.png\" alt=\"\" class=\"wp-image-398\" referrerpolicy=\"no-referrer\" \/><\/figure><\/div>\n<p>That red line is a one way corridor. If you want to get to the room with a chest in, you have to go the long way around. The majority of chests in the game either are at the end of a one way loop, or deep enough into the level to force you to fight many rooms just for that chest. No risk \u2013 no reward.<\/p>\n<p>The secret to this pacing is that general layouts have been carefully pre-authored. Here\u2019s the layout that was used to generate the above.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/boristhebrave.com\/content\/2019\/07\/F1_Castle_Flow_01.svg\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/figure><\/div>\n<p>Normal rooms are randomly picked rooms with enemies, hub rooms are larger rooms with more exits. Reward and boss should be self-explanatory. Not shown are \u201cconnector\u201d rooms, which are enemyless rooms, often with a environmental hazard. The other rooms are either prespecified, or drawn from a specific table of rooms.<\/p>\n<p>There\u2019s a handful of these layouts, called flows, per stage in <em>Gungeon<\/em>. The Hollow has the fewest with 4, while the Gungeon Proper has the most at 8. They\u2019re not generic \u2013 each is designed around a specific feature that you can notice with experienced play. This might be a gigantic loop, or an important multi way fork, or having to travel through the shop to get to the end. They\u2019re so noticeable, that speedrunners eventually spotted the differences and <a href=\"https:\/\/www.speedrun.com\/Enter_the_Gungeon\/guide\/1ijr7\">created charts<\/a> indicating how to find the boss as fast as possible. I\u2019ve made a full list of layouts to download <a href=\"http:\/\/boristhebrave.com\/content\/2019\/07\/gungeon_flows.zip\">here<\/a>.<\/p>\n<p>You may have spotted that the flow diagram and the map don\u2019t actually correspond perfectly. There\u2019s an extra room directly below the shopkeeper room that doesn\u2019t correspond, and there\u2019s these weird corridor rooms <img class=\"wp-image-401 c6\" src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_corridor.png\" alt=\"\" referrerpolicy=\"no-referrer\" \/>. Let\u2019s go over the full process, it contains a lot of clever ideas.<\/p>\n<p>The process starts with a randomly picked flow file, like the one drawn above. It\u2019s a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Graph_(discrete_mathematics)\">graph data structure<\/a>, meaning it stores the relationships between rooms, but nothing about their positioning. Each room includes metadata about what type of room it is, and what connections it should have. The connections are actually directed \u2013 each flow chart starts from a root node, and then forms a tree of children nodes. Then extra connections break the tree structure to add loops. I think this is mainly a quirk of how the game was developed, but it is handy for some of the map analysis routines that all loops have a well defined beginning and end.<\/p>\n<h4>Flow Transformation<\/h4>\n<p>The flow file is transformed in a few different ways. Firstly, some specified rooms are replaced with lines of rooms of random length. This feature is only used on the later, larger, stages. Similarly, some sections of the flow file have alternate paths, and one of them is selected at random. This feature is only used twice.<\/p>\n<p>Then, some extra nodes are \u201cinjected\u201d. This feature is quite flexible and is is used for all sorts of purposes.<\/p>\n<p>Each injection contains data specifying what sort of thing should be inserted, where it should be inserted, probability of spawning, and any pre-conditions that must be met (such as having a <a href=\"https:\/\/enterthegungeon.gamepedia.com\/Master_Round\">master round<\/a>, high <a href=\"https:\/\/enterthegungeon.gamepedia.com\/Curse\">curse<\/a>, or having not yet rescued a character). For example, secret rooms usually spawn at dead ends, but have a 1\/5 chance of being attached to any room. They have a 90% chance of appearing, and have no pre-requisites.<\/p>\n<p>Nearly every special room in the game is specified with node injection, including vendors (other than main shop), jails, fireplace and elevator rooms.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_cadence.png\" alt=\"\" class=\"wp-image-416\" srcset=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_cadence.png 416w, https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_cadence-234x300.png 234w\" referrerpolicy=\"no-referrer\" \/><figcaption>One of the jail cells that can be injected into a level<\/figcaption><\/figure><\/div>\n<p>It\u2019s also at this point that the generator picks a specific room for each node. This is mostly based on what stage you are on and what type of room is wanted. There\u2019s a huge list of rooms \u2013 nearly 300 for the first stage, but it\u2019ll avoid picking the same room twice.<\/p>\n<p>Nodes of type \u201cconnector\u201d work differently. They have their rooms picked later, while layout occurs. These are often long, thin rooms, so picking one with a sensible orientation is vital.<\/p>\n<h4>Composites<\/h4>\n<p>Once we\u2019ve got the completed flow, it is broken down into \u201ccomposites\u201d. Each composite is either a single loop of rooms, or a set of connected, loopless rooms (i.e. a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Tree_(graph_theory)\">tree<\/a>). This is done by finding the smallest loop in the map, and cutting it out as a composite. This is repeated until no loops are left. The rest of the map is a bunch of disconnected trees, and connections between individual composites.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/boristhebrave.com\/content\/2019\/07\/F1_Castle_Flow_01_by_composite.svg\" alt=\"\" referrerpolicy=\"no-referrer\" \/><figcaption>The same flow, after injection and separating into composites<\/figcaption><\/figure><\/div>\n<h4>Composite Layout<\/h4>\n<p>Each composite is then laid out separately, in a separate map. They\u2019ll get composed together later.<\/p>\n<p>To lay out a composite, the first room is put an arbitrary spot. Then rooms are added to the layout one by one, by picking a pair of exits, one in the new room, and one in the layout. Exits are are pre-defined locations in the metadata for each room. The new room is then aligned so that its exit connects directly to the exit of the previous room. Then we repeat.<\/p>\n<p>To be more specific, tree composites are laid out by walking the tree <a href=\"https:\/\/en.wikipedia.org\/wiki\/Depth-first_search\">depth first<\/a>. The algorithm only picks pairs of exits that locate the new room without overlapping with what has been placed so far. Generally, the algorithm prefers to select exits far from already used exits. If layout isn\u2019t possible, it will <a href=\"https:\/\/en.wikipedia.org\/wiki\/Backtracking\">backtrack<\/a> and regen the choice of room, up to 3 times.<\/p>\n<p>Meanwhile, loop composites are laid out by adding entries from the loop, alternately at either end of a line. To begin with, exit pairs are picked randomly (with preference for opposing walls, west-east or north-south). Once the loop is half constructed, it starts to prefer exit pairs that bring the two unclosed ends of the loop closer together. When it has constructed all the rooms, it still needs to add one more connection between the last two rooms. It picks another pair of exits. If possible, it constructs a small rectangular room between those exits. Otherwise it pathfinds between the exits, and makes a \u201croom\u201d that is just a narrow corridor. The corridor must be between 4 and 30 units long (50 in the mines).<\/p>\n<h4>Final Assembly<\/h4>\n<p>At this point small individual pieces of the dungeon have been connected to each other, but the composites themselves still needed to be put together into the full map.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle_by_composite.png\" alt=\"\" class=\"wp-image-406\" srcset=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle_by_composite.png 945w, https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle_by_composite-300x203.png 300w, https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_castle_by_composite-768x521.png 768w\" referrerpolicy=\"no-referrer\" \/><figcaption>The same map as composites before final assembly<\/figcaption><\/figure><\/div>\n<p>As you can see, the remaining connections don\u2019t really leave much choice in this case. But there can be harder cases than before. The map is walked over, starting with the room with the most connections. As before, for each connection to make, a pair of exits are chosen. If the two rooms are in disconnected parts of the map, then the two map pieces are aligned next to each other to give a short path. Otherwise, pathfinding is used to construct a route.<\/p>\n<p>And that\u2019s it for level layout. The rest is picking enemies and decorations for the rooms, which is a different matter altogether.<\/p>\n<h2>Final thoughts<\/h2>\n<p>It seems designing a procedural generator that gave a satisfactory play experience was a key goal of the developers. They obviously went through a lot of iterations to get it right \u2013 there is loads of generation code that appears to be unused as they tweaked the formula to get it perfect.<\/p>\n<p>One particular trick is that they generate the parts of the map that are hardest \/ most important first. The generator emphasizes making tight loops and short corridors for the most central parts of the level, and then tries to fit everything else around that.<\/p>\n<p>As with my <a href=\"https:\/\/www.boristhebrave.com\/2019\/07\/14\/dungeon-generation-in-diablo-1\/\">Diablo 1 writeup<\/a>, I\u2019m struck by how effective it is to generate part of the dungeon in a more abstract form, in this case, a graph with no position information, and only later make everything more concrete. The \u201cinjection\u201d customization would just not be possible if we went straight to dealing with a tile based map, and it allows the control of pacing and scope to be easily separated from details of positioning and details.<\/p>\n<p>I also admire how extensible the whole system is. Unity encourages a very <a href=\"https:\/\/en.wikipedia.org\/wiki\/Data-driven_programming\">data-driven<\/a> approach. Adding a new room, layout, or even special behaviour are all doable by adding additional objects into the appropriate tables. This must have been a huge boon as Dodge Roll cranked out several free DLCs, and no doubt encourages <a href=\"https:\/\/modthegungeon.eu\/\">mods<\/a>.<\/p>\n<div class=\"wp-block-image\">\n<figure class=\"aligncenter\"><img src=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_resourceful_rat.png\" alt=\"\" class=\"wp-image-413\" srcset=\"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_resourceful_rat.png 640w, https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_resourceful_rat-300x169.png 300w\" referrerpolicy=\"no-referrer\" \/><\/figure><\/div>\n<h2>Bonus stuff<\/h2>\n<p>Studying this generation was my first chance to observe how a professional unity game is constructed. The devs at Dodge Roll did a great job writing this code. It\u2019s very readable, and in some places, quite funny \u2013 it seems their fondness for punny names extends to the code itself. Some of my favourites include:<\/p>\n<ul><li>The fluid engine in the game called <code>DeadlyDeadlyGoopManager<\/code><\/li>\n<li>The dungeon generation code is called <code>Dungeonator<\/code><\/li>\n<li>The various stages are called <code>CASTLEGEON\/GUNGEON\/MINEGEON\/CATACOMBGEON<\/code> etc. I have to wonder if the devs were inspired by Diablo 1, which has a very similar layout.<\/li>\n<li>Literally every room is named, usually something punny (or after Joe, the presumably egotistic artist on team)<\/li>\n<\/ul><p>I also noticed originally there were plans for space, jungle, wild west and \u201cbelly\u201d themed stages. Alas \u2019twas not to be. Dodge Roll have decided their work on <em>Gungeon<\/em> <a href=\"https:\/\/www.pcgamesinsider.biz\/indie-interview\/68862\/why-dodge-roll-has-decided-to-part-ways-with-enter-the-gungeon\/\">is done<\/a>. I look forward to playing their next game, and hope they put just as much love and thought into it.<\/p>","created_at":"2019-12-21T17:01:46+0000","updated_at":"2019-12-21T17:01:46+0000","published_at":"2019-07-28T16:55:57+0000","annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":9,"domain_name":"www.boristhebrave.com","preview_picture":"https:\/\/www.boristhebrave.com\/wp-content\/uploads\/2019\/07\/gungeon_resourceful_rat.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1065,"title":"Thread by @gravislizard: \"almost everything on computers is perceptually slower than it was in 1983 amber-screen library computer in 1998: type in two words and hit F [\u2026]\"","url":"https:\/\/threadreaderapp.com\/thread\/927593460642615296.html","content":"<p>almost everything on computers is perceptually slower than it was in 1983<\/p><p>amber-screen library computer in 1998: type in two words and hit F3. search results appear instantly.<\/p><p>now: type in two words, wait for an AJAX popup. get a throbber for five seconds. oops you pressed a key, your results are erased<\/p><p>one of the things that makes me steaming mad is how the entire field of web apps ignores 100% of learned lessons from desktop apps<\/p><p>data in webpages in 2017 is distressingly fragile. go to google maps and try and find an action that *doesn't* erase what you're doing<\/p><p>drag the map even a pixel? it erases all your results and closes the infobox you were looking at.<\/p><p>you have a list of interesting locations on the screen but you want to figure out how far they are from the center of town? you can't.<\/p><p>you can open a new tab, do the search there, then flip back and forth manually in the browser. there's no other way.<\/p><p>that is to say, once the data's up on the screen, you *can't add to it*. which is one of the core functions of computers, generally.<\/p><p>one of the primary reasons computers were *created* was to cross reference data. that is nearly impossible in most software now.<\/p><p>maps are a particularly hot item for this. christ, what about looking at a map ISN'T about cross ref'ing data? it's the WHOLE POINT<\/p><p>you have a start and a finish and need to integrate that with geography and roads. and gmaps, bing, etc. are all the worst choice for this.<\/p><p>you are, literally, better off taking a screenshot of the map, dropping it in ms paint and manually plotting there.<\/p><p>gmaps wildly thrashes the map around every time you do anything. Any time you search, almost any time you click on anything<\/p><p>it's a bewildering whirl of colors and shapes that has gotten worse every six months for 15 years<\/p><p>and in doing so it has made humans worse and worse and worse at doing things that computers were created to replace and improve<\/p><p>in 1998 if you were planning a trip you might have gotten out a paper road map and put marks on it for interesting locations along the way<\/p><p>with online maps you CAN do that, but the entire process is built assuming you already know everywhere you're going<\/p><p>It APPEARS to be what you want - you can keep putting in locations and it'll keep plotting them - but in truth it's not at all<\/p><p>The process you WANT: pick your start and end. now start searching for places in between. Your start and end are saved.<\/p><p>When you find someplace interesting, add it to your list. Keep doing that, keep searching and adding.<\/p><p>Search far and wide. Search for cities and then click around inside them. Read reviews. Do street view.<\/p><p>When you're all done, you go back to your plotted trip and start laying out the chosen locations and optimizing your path.<\/p><p>You can do this with a paper map. You can't do this with gmaps. So you just don't do it.<\/p><p>You do something halfass and unsatisfying instead, using multiple tabs or a text file you save addresses in or some shit<\/p><p>You don't even realize why the process is frustrating because it's just The Way It Is.<\/p><p>And everything on computers is like this. It's just How It Is now. You can't fail quickly and iterate.<\/p><p>On the library computer in 1998 I could retry searches over and over and over until I found what I was looking for because it was quick<\/p><p>Now I have to wait for a huge page to load, wait while the page elements shift all over, GOD FORBID i click on anything while its loading<\/p><p>how many times have i typed in a search box, seen what i wanted pop up as i was typing, go to click on it, then have it disappear<\/p><p>I make no secret of hating the mouse. I think it's a crime. I think it's stifling humanitys progress, a gimmick we can't get over.<\/p><p>The mouse is the CueCat except it didn't get ridiculed and reviled as it should have been. It's inappropriate for almost everything we do.<\/p><p>There's no reason for Twitter to use a mouse. There's nothing mousey about this website, not a damn thing<\/p><p>Mice are for rapidly navigating through a complex and unstructured set of objects, like an app with dozens of options and input types<\/p><div class=\"row t-tweet col-12 content-tweet allow-preview\">twitter: i need to navigate through a linear list and perform one of four actions on discrete items, almost all text-based<br \/>...<img alt=\"\" src=\"https:\/\/pbs.twimg.com\/media\/DN9_tLVVwAAh_iS.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><p>When computers used interfaces like these they were lightning fast, universally. I've used a lot of them.<\/p><p>The library computers, the homebrew POS at my first job, even the machines at Fry's Electronics use an interface like this.<\/p><p>If you ever go to Fry's, watch the employees use this style of UI. Blinding speed. They ricochet around like pros, even the new people.<\/p><p>And I have to say that it's solely because the mouse is not being used. Mice are bad. Mice are absolutely terrible.<\/p><p>The reason mice are terrible is a matter of basic facts about human brains, hands, eyes and muscles. Hell, I think Jef Raskin covered it.<\/p><p>Keyboards present fewer possible discrete options. Mice present a continuum. One can be operated blind; the other requires feedback.<\/p><p>You cannot use a mouse without using your eyes to confirm everything. At my first job I rang people up without looking at the screen<\/p><p>POS software is designed VERY pragmatically. Typically there is no concept of \"input focus\"; all text goes to one field.<\/p><div class=\"row t-tweet col-12 content-tweet allow-preview\">In this UI, most likely at any given moment keyboard entry will ONLY go to the \"Stock Number\" field.<img alt=\"\" src=\"https:\/\/pbs.twimg.com\/media\/DN-BTjQV4AIt-hr.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><p>When you walk up to this computer, you can rest assured of the following:<br \/>a) your input will go there<br \/>b) if not, pressing escape will fix it<\/p><p>If you don't know the stock number, press F1 to be transported AWAY FROM THIS INTERFACE to look it up. this is critical.<\/p><p>i posit that nobody wants autocomplete-style live DB lookups. They don't fit the mold that autocomplete fits in.<\/p><p>When you're unsure of even what search term you want, you are already context switching. That means changing interfaces is okay.<\/p><p>If you DON'T do that, you're trying to cram a side-mission - figure out what search term I'm looking for - into a UI not designed for it.<\/p><p>Anyway. Once you find the stock number you want, the price shows up. Hit F3 to edit it. * to change quantity. Enter to add to invoice.<\/p><p>In practice this is incredibly fast. You can't even see it happening, because there's no reason to except to verify after the fact.<\/p><p>The operator doesn't have to stop for every action and figure out what they're doing time and time again. It's all muscle memory.<\/p><p>The cashiers hands are just a blur on the keys. They're looking at the products, not the screen. And they get it right.<\/p><p>Part of the reason for this is that the entire keyboard gets used, including the function keys. INCREDIBLY powerful, those.<\/p><p>God. God! What a tragedy, that we left them in the past. What a heinous crime that we forgot their value.<\/p><p>Look at that UI up there. Look how many functions you can access AT ANY TIME. You can RELY on them. They will ALWAYS be there.<\/p><p>In these systems, the screen was always dominated by a specific form that you could identify from across the room.<\/p><div class=\"row t-tweet col-12 content-tweet allow-preview\">Invoice form<br \/>Stock lookup form<br \/>Customer lookup form<br \/>Payment form<p>And the F-keys were *always* bound to that form.<\/p><\/div><p>This allowed your \/human brain\/, a flawed but powerful tool, to use the computer INTUITIVELY<\/p><p>As you approach e.g. the cash register, you take one look at the screen and instantly your synapses wire themselves up for the F-key layout<\/p><p>If the register is on the stock lookup screen, your hands fall on the keyboard and automatically do what's needed<\/p><p>Escape-escape to get to the login screen, fingers on the numpad hammer in your employee ID, F2 to cancel retail customer lookup<\/p><p>And now you're ready to start keying in items. No \"getting your bearings.\" You don't even stop talking to the customer. 1\/10th of a second.<\/p><p>To be clear, the web as I envision it does not look like what we have. Not one bit. It's completely different.<\/p><p>And it's worth noting that HYPERTEXT, specifically, is best with a mouse in a lot of cases. Wikipedia would suck on keyboard.<\/p><div class=\"row t-tweet col-12 content-tweet allow-preview\">This is a FANTASTIC example of a mouse-optimal document. Any keyboard approach would be mediocre at best.,<img alt=\"\" src=\"https:\/\/pbs.twimg.com\/media\/DN-GjbPVAAAd9lD.jpg\" referrerpolicy=\"no-referrer\" \/><\/div><p>Your brain is GREAT at identifying points of interest here. From this array of 20+ unstructured links I can grab the ones I want.<\/p><p>Like a grocery store shelf, you can double-fist it if things are organized appropriately. I've grabbed beans and rice simultaneously before<\/p><p>Your brain highlights the points of interest like a T-1000 and bam bam bam you grab everything you need. New tab new tab new tab.<\/p><p>Twitter, however, is deeply linear, as is Google, eBay, Dropbox, bug trackers, IM clients<\/p><p>I could continue to talk for hours about this but I think I covered the bases. RT as desired.<\/p><p>A comment - although I appreciate all of you that are recommending apps better than gmaps and keyboard browser plugins and the like...<\/p><p>I want to clarify that I am literally talking about the future of the human race and I am deadly serious about this. It's not about me.<\/p><p>I do OK. I'm an extremely experienced user, I can overcome everything this garbage industry throws at me, albeit bitterly.<\/p><p>I am mad that commodity software that people don't even RECOGNIZE as software, like gmaps, is so far through the looking glass.<\/p><p>It makes me mad that Joe Q Middleage, Safeway manager in New Jersey, can't use the computer better bc it sucks so bad.<\/p><p>Joe is using the computer to emulate labor-intensive manual work on paper and he doesn't even know he's being shafted<\/p><p>Joe should be offered incredible power by software that encourages him to use it with purpose-built interfaces optimized for him.<\/p><p>I am upset by the way that computers disenfranchise non-nerds. I wish it was better for me; I wish it WORKED AT ALL for everyone else.<\/p><p>another thing you should know about me - and this is where i think a lot of people will start disagreeing - I don't believe in \"intuitive\"<\/p><p>I think it's a crock of shit. Hogwash. Absolutely false. The belief that GUIs are more intuitive is marketing for GUIs.<\/p><p>GUIs are in no way more intuitive than keyboard interfaces using function keys such as the POS I posted earlier. Nor do they need to be.<\/p><p>GUIs require you to learn how to use a mouse, how input focus works, how multiple windows work, how modal dialogs work.<\/p><p>I believe well designed keyboard interfaces and well designed GUI interfaces have exactly the same learning curve.<\/p><p>The problem is you can't prove that now because we spent the last 25 years teaching people ONLY GUI and mouse.<\/p><p>Nobody will agree with me, citing anecdotes and examples that are meaningless in the current zeitgeist.<\/p><p>All UIs require a common language. The common language of quality keyboard interfaces was no different than GUIs.<\/p><p>People get upset by OPAQUE keyboard interfaces, which I don't disagree with. The POS software I posted earlier is not opaque.<\/p><p>Once you understand the basic language of keyb UI, those textmode interfaces are completely intuitive by the same standards.<\/p><p>The other thing I think maybe some people don't get - I'm not advocating textmode. Why would I do that? Graphics and keyboard work fine.<\/p><p>Character cell graphics offer some benefits that aren't worth discussing; they are not an option for plenty of reasons now.<\/p><p>The cash register software I used at my last job was graphical AND keyboard based. Worked fine. Used function keys.<\/p><p>If I could do this thread over again, I'd just alter 1983 to 1987, I think. '83 might have been too aggressive. But it wasn't literal anyway<\/p><p>If this thread positively affected you, can I ask you to throw me a couple bucks? <a class=\"entity-url\" data-preview=\"true\" href=\"http:\/\/ko-fi.com\/gravis\">ko-fi.com\/gravis<\/a> I recently had a water main burst and the plumbing cost a fortune.<\/p><p>Check out my YouTube channel where I look at old software, and should be posting some videos on 80s utilities soon <a class=\"entity-url\" data-preview=\"true\" href=\"https:\/\/www.youtube.com\/channel\/UCXnNibvR_YIdyPs8PZIBoEw\">youtube.com\/channel\/UCXnNi\u2026<\/a><\/p>","created_at":"2019-12-19T15:28:34+0000","updated_at":"2019-12-19T15:28:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"threadreaderapp.com","preview_picture":"https:\/\/threadreaderapp.com\/images\/screenshots\/thread\/927593460642615296.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["gaming","headset","headphones"],"is_public":false,"id":1064,"title":"B07YQFJ2DR","url":"https:\/\/www.amazon.com\/Sennheiser-GSP-370-Over-Ear-Wireless\/dp\/B07YQFJ2DR?tag=rps-m-20&ascsubtag=2019-12-18-sennheiser-gsp-370-review-","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2019-12-18T19:54:36+0000","updated_at":"2019-12-18T19:54:46+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.amazon.com"},{"is_archived":0,"is_starred":0,"tags":["procgen","interactive-fiction","inkle"],"is_public":false,"id":1063,"title":"Writing web-based interactive fiction with ink","url":"https:\/\/www.inklestudios.com\/ink\/web-tutorial\/","content":"<h2><em>For complete beginners!<\/em><\/h2><p>Writing interactive fiction has never been easier - there are so many authoring systems available. Twine has become especially popular for web-based IF because of its user friendly approach, using visual tools to allow you plan your story with \u201csticky notes and string\u201d.<\/p><p>Ink by comparison, is a system that wasn\u2019t originally designed for the web. It was designed as a pluggable component that integrates into a traditional game engine. At <a href=\"https:\/\/www.inklestudios.com\/\">inkle<\/a>, we\u2019ve used it for almost everything we\u2019ve created, from <a href=\"https:\/\/www.inklestudios.com\/80days\">80 Days<\/a> (initially for mobile) to the upcoming <a href=\"https:\/\/www.inklestudios.com\/heavensvault\">Heaven\u2019s Vault<\/a> for PS4 and PC. Since we open sourced it, it\u2019s also been used in a host of highly respected indie games including <a href=\"http:\/\/burymemylove.arte.tv\/\">Bury Me My Love<\/a> and <a href=\"https:\/\/www.wherethewatertasteslikewine.com\/\">Where the Water Tastes Like Wine<\/a>, both nominated for IGFs in 2018.<\/p><p>But it's easy to write stories for the web in ink! What\u2019s been lacking until now is a straightforward tutorial that demonstrates the basics of writing a story in ink, exporting it for web, and making it available for others to play.<\/p><p><strong>This guide will assume zero prior knowledge of ink or web technology<\/strong>, so that less technical writers have everything they need to get started, but we\u2019ll provide jumping off points throughout if you want to learn more about a topic in depth.<\/p><p>Here\u2019s the rough structure that we'll follow. Feel free to skip or skim over any topics that you are already familiar with.<\/p><ul><li><strong>Getting started with ink<\/strong>: downloading the tools and writing a simple branching narrative.<\/li>\n<li><strong>Exporting for web<\/strong> - how to create a shareable web page of your story<\/li>\n<li>Optional: <strong>Customising the look and feel<\/strong> of the page, including an explanation of the basics of CSS.<\/li>\n<li><strong>Uploading to itch.io<\/strong>: the perfect place to upload your work for others to play.<\/li>\n<\/ul><h2>Getting started with ink<\/h2><p>There's an <a href=\"https:\/\/github.com\/inkle\/ink\/blob\/master\/Documentation\/WritingWithInk.md\">official manual<\/a> for writing in ink that tells you everything you would need to know, from the basics to the complex and powerful structures.<\/p><p>But this guide will jump through all that, pointing out only the most easily-understood features that will allow you to write a simple <em>Choose Your Own Adventure<\/em> style structure. At any point though, feel free to jump into the official guide if you want to try something more advanced.<\/p><p>We'll be writing in <strong>Inky<\/strong> - the official editor for authoring ink scripts. So, go ahead and <a href=\"https:\/\/github.com\/inkle\/inky\/releases\">click here to download it<\/a>!<\/p><p>When you launch it, you'll see a window with two columns. On the left hand side is where you write your script, and on the right hand side is where you'll be able to play a preview of it.<\/p><p>The first thing to know about ink is that fundamentally, it's just pure written text, with special annotations (\"markup\") that make it interactive. So go ahead - write a sentence on the left hand side, and you'll see it appear on the play area on the right. Adding multiple paragraphs works just fine too.<\/p><p><img src=\"https:\/\/www.inklestudios.com\/ink\/web-tutorial\/basic-text.jpg\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><h3>The anatomy of an ink story<\/h3><p>Now, let's jump straight in and take a look at a simple example of a branching story.<\/p><p>The blue markup in the screenshot below might look initially intimidating if you've never done any programming before (lots of symbols!), but we'll go through it bit by bit so you can understand how it fits together.<\/p><p><img src=\"https:\/\/www.inklestudios.com\/ink\/web-tutorial\/full-story.jpg\" alt=\"\" referrerpolicy=\"no-referrer\" \/><\/p><p>For the full text of the story, <a href=\"https:\/\/www.inklestudios.com\/ink\/web-tutorial\/80-days-example.txt\" target=\"_blank\">click here<\/a>.<\/p><p>Feel free to copy and paste this directly into the left hand pane of Inky, and then play through it in the right-hand pane. To reset the story back to the beginning, press the double arrow at the top right (the other arrow rewinds a single choice).<\/p><p>While you're playing, you can simply tweak the text on the left, and it'll update live in the preview, even if you're in the middle of the story. You can even alt-click a word on the right hand pane and it'll select the source text on the left hand pane (this is extremely useful when a story gets quite complex!)<\/p><p>There are just three anatomical parts you have to understand in this simple story: knots, diverts and choices.<\/p><h4>1. KNOTS<\/h4><p>The story is comprised of multiple linked sections which we call \"knots\" in ink terminology. The start of a knot is indicated in ink using at least two equals signs the left hand side knot's name, and optionally on the right too (so for example <code>=== london ===<\/code>). The blue highlighting in Inky indicates ink markup that has recognised, confirming to us that it's been written correctly, and making it stand out as being separate to our content.<\/p><p>All the content below a knot header then belongs to that knot.<\/p><p>The knot's name (which can't have any spaces in it) won't appear in the story itself; instead it's used when we want to link sections of our interactive narrative together.<\/p><h4>2. DIVERTS<\/h4><p>Now, talking of linking, this bit is called a <strong><em>divert<\/em><\/strong>:<\/p><p><code>-&gt; london<\/code><\/p><p>Using the divert arrow (entered as \"minus\" then \"arrow bracket\"), we tell the story to go to a different knot. When the story is played, the join will be automatic and invisible to the player.<\/p><p>So, at the very beginning of our story, after showing the introductory paragraph, we have the <code>-&gt; london<\/code> divert which takes the player directly into the knot with that name.<\/p><p>There's also a special kind of divert at the end of this story. If you divert to \"END\", that simply tells ink <em>we're done; this ending is intentional<\/em>. Try removing the line with <code>-&gt; END<\/code> on it: Inky will show a warning icon in the margin. If you mouse over the warning, it'll tell you that you have a \"loose end\".<\/p><p>This is an important feature of ink, and one which will help you as you story becomes increasingly complex. As you diverge your story in different directions, you'll realise that as you write, you have to leave divergent threads of your story unfinished. These warnings remain as reminders of which parts you still have left to write.<\/p><h4>3. CHOICES<\/h4><p>Talking of divergence, the final crucial feature that we need in order to become dangerous with interactive writing is of course, choices:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ \"Around the world, Monsieur?\"\n<\/pre><\/div><p>Choices in ink are designed to look like a series of bullet points. Frequently, they're directly combined with diverts in order to create a choice that directly takes you elsewhere in the story:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ [Nod curtly.] -&gt; nod\n<\/pre><\/div><p>There are a few things to note about the way we've authored these ones.<\/p><p>Normally, the text you put in your choice becomes both the text of the clickable choice, as well as text that appears in the main body after you click it. However, if you want a \"silent\" choice, with text that only appears on the choice itself, you can put the text in <code>[<\/code>square brackets <code>]<\/code>, as you can see with the second example.<\/p><p>This can be used stylistically, and is often used for the difference between a command-like choice and a choice that sounds narrated, as part of the story. For example:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ [Nod curtly.]     -&gt; nod\n+ I nodded curtly.  -&gt; nod\n== nod ==\nIt seemed unbelievable.\n<\/pre><\/div><p>The first choice would only produce the <code>It seemed unbelievable.<\/code> text after picking the choice, whereas the second would produce <code>I nodded curtly. It seemed unbelievable.<\/code>.<\/p><p>You can also add extra lines of text under the choice. This will be used as content to be shown after the choice has been taken, and not before. For example:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ \"Around the world, Monsieur?\"\n    I was utterly astonished.\n    \"You are in jest!\" I told him in dignified affront.\n    -&gt; astonished\n=== astonished ===\n...etc...\n<\/pre><\/div><p>... is equivalent to:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ \"Around the world, Monsieur?\" -&gt; astonished\n=== astonished ===\nI was utterly astonished. \n\"You are in jest!\" I told him in dignified affront.\n...etc...\n<\/pre><\/div><p>ink is very flexible - it allows you lots of different ways to structure your narrative.<\/p><p>And a final note about the overall structure of an ink story: the ink will always begin at the top of the file, and work its way down. But if <em>all<\/em> of your content is split into knots, it's important to make sure that at the very least you have an initial divert at the top of the file (such as <code>-&gt; intro<\/code>) to tell it which knot to begin with.<\/p><p>Now, why not experiment with this story? Try inserting extra text, extra choices, extra diverts, extra endings, and see what happens! As an exercise to get started - perhaps try and turn the introductory paragraph into its own knot?<\/p><h3>If you like: Conditional content<\/h3><p>So far, everything we've written could've been authored as a printed <em>Choose Your Own Adventure<\/em> type novel. But what if we want to show different text and choices depending on what the player did in the past?<\/p><p>The ink engine takes note of every section of content that the player reads, so that you can query this later on. For example, to determine whether the player saw a particular knot called <code>catacombs<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">{ catacombs:\n    It was darker here than the Paris catacombs.\n}\n<\/pre><\/div><p>Within one of these curly-brace sections you can include multiple lines of content. You can also include diverts and choices in them.<\/p><p>However, a more succinct way of writing conditional choices is the following, which is the neatest way to provide a specific choice only if a particular knot has been seen:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ {catacombs} [Tell her what you found] -&gt; tell_her\n<\/pre><\/div><p>If you want to invert the condition - in other words if you want to check if they player <em>didn't<\/em> visit the catacombs, you can simply prepend it with <code>not<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">+ {not catacombs} [Visit the catacombs] -&gt; catacombs\n<\/pre><\/div><p>Finally, you can use <code>and<\/code> and <code>or<\/code> to form more complex conditions, and you can use parentheses to clarify the specifics of the logic. A couple of examples:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">{ catacombs and not pick_up_ring:\n    \"So you didn't find it then?\" she enquired.\n    + [Apologise.] -&gt; apologise\n}\n{ (catacombs or cross_river or sing_in_rain) and not buy_new_shoes:\n    My shoes were sodden from earlier in the day.\n}\n<\/pre><\/div><h3>Advanced ink<\/h3><p>We've only looked at a fraction of the powerful features that are available in ink. If you want to read more about it or just dive more deeply into something we've already seen, check out the <a href=\"https:\/\/github.com\/inkle\/ink\/blob\/master\/Documentation\/WritingWithInk.md\">official documentation<\/a>. To give you a taste:<\/p><ul><li>As with any programming language, you can create custom variables, and perform mathematical calculations.<\/li>\n<li>We've been using <code>+<\/code> as choice bullets, whereas normally we would recommend using <code>*<\/code>. The difference is that a <code>*<\/code> choice will never reappear once it's been chosen - great for repeated hubs of content where you don't want the player to read a specific section repeatedly.<\/li>\n<li>There's a harder-to-learn but easier-to-write system for writing intricate branches called \"weave\" that doesn't require you to name every section with its own header. Most of our game <em>80 Days<\/em> was written this way.<\/li>\n<li>Within knots you can also have sub-sections called \"stitches\".<\/li>\n<li>Writing content that varies and branches mid-sentence is both possible and straightforward.<\/li>\n<li>You can split up your ink file into smaller ones that are linked together.<\/li>\n<li>And much much more!<\/li>\n<\/ul><h2>Exporting for web<\/h2><p>The next step is to save our story out to a set of files that can be uploaded somewhere as a webpage. To do this from Inky, choose File, then \"Export for web...\", and choose the name of your story. The name you choose will be used both for the visible title on the webpage (so use spaces and proper capitalisation!), as well as the name of the folder that it gets saved into.<\/p><p>As you make changes to the story, and <em>especially<\/em> if you decide to make changes to the look and feel (see below), then in future you should choose the \"Export story.js only...\" option, saving it into the same folder and overwriting \"yourStoryName.js\". This is the only file that needs to be updated as you change your ink content - the rest is purely part of the web template. So, if you change the look and feel of the template, you don't want to accidentally overwrite it.<\/p><p>At this point, you're already done! You can skip to the <strong>Upload to itch.io<\/strong> step, if you like.<\/p><p>However, if you'd like to add images, set a byline, use a dark theme, or have a play with the look and feel of your story, read on!<\/p><h2>Special features for the web<\/h2><p>ink has been designed to be as focussed as possible on one task: writing interactive text. We've resisted overloading the language with extra features in order to keep a good balance of power, flexibility and simplicity.<\/p><p>So, ink-the-language doesn't have any <em>built-in<\/em> tools for things such as inserting images, since it's totally dependent on the type of story you're making. If you were writing dialogue for a 3D game, specifying an image to be displayed wouldn't necessarily make sense.<\/p><p>However, <em>Inky<\/em> does provide some extra features that can be used specifically when saving for the web (since <a href=\"https:\/\/github.com\/inkle\/inky\/releases\">version 0.10.0<\/a>). These extra features all use ink's tag system, which allows you to provide special text annotations to each line, either before it or above. The tags are just pure text that's invisible to the reader, but can be read by the game system, or in our case, the web template.<\/p><h3>Images<\/h3><p>Here's how you insert an image:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># IMAGE: imageName.jpg\n<\/pre><\/div><p>The image file should be in the same folder as your other files, or if you want you can use a path such as:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># IMAGE: myImages\/imageName.jpg\n<\/pre><\/div><p>You can also add an image tag on the end of the line. The image will always be shown above the text in the game.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">The above picture is a dog.     # IMAGE: dog.jpg\n<\/pre><\/div><p>(This is actually because all tags are associated with particular lines of text, and when you have a tag on a line on its own, it'll be associated with the line of text below it.)<\/p><h3>Clearing the screen<\/h3><p>Another feature that's specific to Inky's web template is the <em>clear<\/em> tag:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># CLEAR\n<\/pre><\/div><p>This will remove all the text in the current flow, and start again from the top of the page. Make sure you only insert this directly after a choice! If you insert it in the middle of a section of text, the top part will be cleared before it's seen.<\/p><h3>Global tags: dark theme and author<\/h3><p>To use the dark theme, simply put this at the very top of your ink file:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># theme: dark\n<\/pre><\/div><p>The web template is able to read this in as a <em>global tag<\/em>, and has built in styling to change theme. (It uses \"dark\" as a class for the overall page. When you see uses of the <code>.dark<\/code> selector in the CSS file, these are the dark theme specific overrides.)<\/p><p>To set the author name, put this at the top of your ink file:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># author: Your Name\n<\/pre><\/div><p>This will be found by the inky web template and used to insert a byline directly below the main title.<\/p><h3>Custom CSS classes<\/h3><p>You can also fully customise the appearance of individual lines of text using <em>CSS classes<\/em>. In your ink, add a tag to a line like this:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">You make your way into the forest.\nThere's a pool of blood on the ground.    # CLASS: danger\n<\/pre><\/div><p>You can then tell the browser what the \"danger\" class looks like by adding the following to your <strong>style.css<\/strong>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">.danger {\n    color: red;\n}\n<\/pre><\/div><p>The dot followed by a word is how we define a style class in CSS. Most of our previous examples didn't have dots since they were built in elements such as \"p\" for \"paragraph\". These can also be combined, so <code>p.danger<\/code> would mean \"a paragraph with the <em>danger<\/em> class\", and would work just as well in our example.<\/p><h3>Custom CSS Example: <em>The End<\/em><\/h3><p>The Inky template contains built in styling for a class called <code>end<\/code>, if you want to show the words <em>The End<\/em> in a centred\/bold style. So, to use it, put the following at the end of your story:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># CLASS: end\nThe End\n-&gt; END\n<\/pre><\/div><p>To recap:<\/p><ul><li>The middle line is the text itself.<\/li>\n<li>The first line applies the CSS class \"end\" to the text.<\/li>\n<li>The last line tells the ink engine that we're intentionally finished, it's not a loose end.<\/li>\n<\/ul><p>And to tweak the styling of the <em>end<\/em> CSS class, look for <code>.end<\/code> at the bottom of the CSS file.<\/p><h3>Restarting the story<\/h3><p>The Inky web template has one more built in feature. If you insert a <code># RESTART<\/code> tag, it will immediately clear the story's progress (including all history of the player's choices and all variables) and start again from the beginning. Since it's immediate, you'll probably want to put it behind a choice, like this:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">You died!\n+ Try again from the beginning.\n    # RESTART\n    -&gt; END\n<\/pre><\/div><p>The <code>-&gt; END<\/code> divert will never actually be reached, but it's good practice: you will still want to have it there to prevent Inky from complaining that you have a loose end.<\/p><h2>Customising the look and feel<\/h2><p>This guide is targetted at complete beginners - it assumes no knowledge of web technology.<\/p><p>The above export step should've saved out a folder with the following files:<\/p><ul><li><strong>index.html<\/strong> - the HTML file is the one that binds the others together, and is the main one that you double click if you want to view your story. It tells the browser the rough structure of the page, and tells it to load the other files.<\/li>\n<li><strong>main.js<\/strong> - this is the main JavaScript file where you can customise the <em>behaviour<\/em> of the way the story is presented.<\/li>\n<li><strong>style.css<\/strong> - this is the main CSS file, which determines things like colours, fonts as well as sizing and spacing information such as the widths of paragraphs and the margins. We can also define transitions, such as the way the paragraphs fade in.<\/li>\n<li><strong>yourStoryName.js<\/strong> - this is your actual ink content, as exported by Inky. This is the only file that's fully unique to your story, as opposed to part of a generic template saved out by Inky. You should never edit this file directly.<\/li>\n<li><strong>ink.js<\/strong> - This is the <a href=\"https:\/\/github.com\/y-lohse\/inkjs\">JavaScript port of the ink engine<\/a> itself.<\/li>\n<\/ul><p>To change any of these files, you should open them up in a pure text editor. We recommend <a href=\"https:\/\/www.sublimetext.com\/\">Sublime Text<\/a>, but you could also use Notepad++ on Windows, TextEdit on Mac, or any other editor you're familiar with. However, do not use Microsoft Word!<\/p><h3>How to change fonts and colours<\/h3><p>Fonts and colours are all controlled from the CSS file - <strong>style.css<\/strong>. If you open up the file, you'll see a series of blocks like this one:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">body {\n    font-family: 'Open Sans', sans-serif;\n    font-weight: lighter;\n    background: white;\n}\n<\/pre><\/div><p>Each one starts with a name like \"body\" or \"h1\". In CSS this is called a <em>selector<\/em> since it selects a particular part of the webpage for styling. In the above case it selects the body, which is effectively the entire page.<\/p><p>The list of fonts following <code>font-family<\/code> is in priority order; it depends on what's available on the user's computer. In the above example <em>Open Sans<\/em> is a web font that's imported at the top of the file - just visit <a href=\"https:\/\/fonts.google.com\/\">fonts.google.com<\/a> for a selection to choose from, and instructions on how to use them (it gives you one of those <code>@import<\/code> lines you can paste in yourself). If for some reason this font can't be downloaded, it'll fall back to the <code>sans-serif<\/code> option.<\/p><p>You can also use a font that's built into the user's system, but the problem is that there's a <a href=\"https:\/\/www.w3schools.com\/cssref\/css_websafe_fonts.asp\">very limited selection<\/a> that are guaranteed to exist. The very most basic fallbacks can be specified using simply <code>serif<\/code> and <code>sans-serif<\/code>.<\/p><p>A few of the CSS selectors and rules in the file include:<\/p><ul><li><code>h1<\/code> - this stands for \"header level 1\" - i.e. the largest header, which is used for the main title of the story at the top of the page. If you wanted change the size of the title, you could change the <code>font-size<\/code> property in here. Similarly the <code>h1, h2<\/code> is a shortcut for setting rules on two selectors at once - header levels 1 and 2. The later is used for the byline if you have an <code># author: Your Name<\/code> tag at the top of your ink.<\/li>\n<li><code>.header<\/code> - A name preceeded by a dot is a <em>class<\/em> name. This particular example selects an element that's simply a container for the title and <a href=\"#globalTags\">byline<\/a>. Often in HTML\/CSS, elements are given containers to help with structuring or to give fine control over spacing and layout.<\/li>\n<li><code>p<\/code> - this stands for \"paragraph\". It's the selector for the main body text on the page. So this is where you should look if you want to change the font size, text colour, etc.<\/li>\n<li><code>a<\/code> - This stands for <em>anchor tag<\/em>, which is HTML terminology for clickable link - these are mostly just used on our page for clickable choices, but could be anything.<\/li>\n<li><code>p.choice<\/code> - This translates to \"a paragraph with the 'choice' class\". If it was just <code>.choice<\/code>, it would mean \"anything with a choice class\". Choices on our page are actually structured as specially styled paragraphs that contain clickable links (anchor tags).<\/li>\n<li><code>a:hover<\/code> - Here, <code>:hover<\/code> is called a pseudo-class, and it's used to select an element when the user has their mouse pointer over it.<\/li>\n<\/ul><p>CSS applies styles according to priorities that are based on how specific the selector is (more specific is higher priority) and position in the file (lower down is higher priority). This allows you to define styles initially with broad strokes and then progressively override them. For example, we define <code>h2<\/code> first, then lower down <code>.dark h2<\/code> overrides part of h2's style: we change the colouring <em>when<\/em> the dark theme is active (when the <code>.dark<\/code> class is given to the whole page - ).<\/p><p>For more on all the different ways you can change the styling using CSS alone, try googling \"CSS tutorial\" - there are loads of great resources out there!<\/p><p>Here's a couple of tutorials to get started:<\/p><ul><li><a href=\"https:\/\/www.w3schools.com\/css\/\">CSS tutorial on w3schools.com<\/a><\/li>\n<li><a href=\"https:\/\/www.w3schools.com\/html\/\">HTML tutorial on w3schools.com<\/a> - the HTML file describes the structure of the page, so if you want to change the overall layout and customise it further, you should learn both HTML and CSS together - they go hand in hand.<\/li>\n<\/ul><h3>Example: Big fancy title<\/h3><p>Select a font on Google Fonts, and insert the <code>@import<\/code> line at the top of your CSS file, such as:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">@import url('https:\/\/fonts.googleapis.com\/css?family=Lobster');\n<\/pre><\/div><p>Then change your header rule to something like:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">h1 {\n    font-family: 'Lobster', cursive;\n    font-size: 80pt;\n    text-align: center;\n}\n<\/pre><\/div><h2>Upload to itch.io<\/h2><p>You've finished your masterpiece! Time to share it with the world.<\/p><p>You could choose to upload your files anywhere that allows you to host a webpage, but we're suggesting <strong>itch.io<\/strong> because it has a great community of indie game creators, it's one of the few places that allow you to upload pure web content, and best of all, it's free! You can even choose to sell your content, if you like.<\/p><p>itch.io also lets you upload a private draft of your game content so you can test out the process before you release it.<\/p><p>Finally, <a href=\"https:\/\/itch.io\/jam\/inkjam\">ink jam<\/a> is hosted from itch.io, making it trivial to enter if you've already got your game uploaded on the platform.<\/p><p>First of all, <a href=\"https:\/\/itch.io\/register\">register for an account here<\/a> if you haven't already. Then:<\/p><ul><li>From <a href=\"https:\/\/itch.io\/dashboard\">dashboard<\/a> choose <em>Create New Project<\/em>.<\/li>\n<li>Enter the details of your game\/story, making sure to choose \"HTML\" for type of project.<\/li>\n<li>Zip up all the files exported from inky and upload them.<\/li>\n<li>Choose \"This file will be played in the browser\".<\/li>\n<li>Embed: We suggest you experiment with this setting to see what you prefer.<\/li>\n<li>Size: We suggest 800 x 600 is better for reading text than default size. It'll still downsize okay for mobile.<\/li>\n<li>Mobile friendly: yes<\/li>\n<li>Automatically start: We suggest you play with it and decide what you prefer.<\/li>\n<li>Full screen button: We suggest yes, but it's up to you.<\/li>\n<li>Enable scrollbars: This isn't necessary since the template has scrolling built in.<\/li>\n<\/ul><p>You can view a preview of your game on itch.io all the while you have <em>Draft<\/em> selected under <em>Visibility &amp; access<\/em>. As soon as you're ready to put the game live, simply select <em>public<\/em>.<\/p>","created_at":"2019-12-17T15:27:39+0000","updated_at":"2019-12-17T15:27:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":20,"domain_name":"www.inklestudios.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["gnu","make","makefile"],"is_public":false,"id":1062,"title":"Your Makefiles are wrong","url":"https:\/\/tech.davis-hansson.com\/p\/make\/","content":"<p>Your Makefiles are full of tabs and errors. An opinionated approach to writing (GNU) Makefiles that I learned from <a href=\"https:\/\/github.com\/benbc\">Ben<\/a> may still be able to salvage them.<\/p><h2 id=\"an-opinionated-approach-to-gnu-make\">An opinionated approach to (GNU) Make<\/h2><p>This is my second hand account of the approach to Make that I learned from <a href=\"https:\/\/github.com\/benbc\">Ben<\/a>. If something is wrong, assume it was lost in translation.<\/p><p>The big things I hope you take away are:<\/p><ul><li>The file system is a fundamental part of Make, don't fight it<\/li>\n<li>Use sentinel files for targets that do not yield exactly one file<\/li>\n<li>Don't use tabs, set -ea -o pipefail, and a few other sensible defaults<\/li>\n<li>Use the above as guidelines, not dogma<\/li>\n<\/ul><p>Because it'll make the examples throughout easier to read, lets start with setting sensible defaults.<\/p><h2 id=\"sensible-defaults\">Sensible defaults<\/h2><h3 id=\"dont-use-tabs\">Don't use tabs<\/h3><p>Make leans heavily on the shell, and in the shell spaces matter. Hence, the default behavior in Make of using tabs forces you to mix tabs and spaces, and that leads to readability issues.<\/p><p>Instead, ask make to use <code>&gt;<\/code> as the block character, by adding this at the top of your makefile:<\/p><pre>ifeq ($(origin .RECIPEPREFIX), undefined)\n  $(error This Make does not support .RECIPEPREFIX. Please use GNU Make 4.0 or later)\nendif\n.RECIPEPREFIX = &gt;\n<\/pre><p>You really just need the <code>.RECIPEPREFIX = &gt;<\/code> part, but the version check is good since there are a lot of old make binaries floating around.<\/p><p>With this change, a makefile that previously looked like this:<\/p><pre>hello:\n    echo \"Hello\"\n    echo \"world\"\n.PHONY: hello\n<\/pre><p>Would look like:<\/p><pre>hello:\n&gt; echo \"Hello\"\n&gt; echo \"world\"\n.PHONY: hello\n<\/pre><p>And you will never again pull your hair out because some editor swapped a tab for four spaces and made Make do insane things.<\/p><h3 id=\"always-use-a-recent-bash\">Always use (a recent) bash<\/h3><p>Make uses <code>\/bin\/sh<\/code> as the default shell. People have <code>\/bin\/sh<\/code> set up to point to different shell implementations. \u201cPortable shell scripts\u201d is a pipe dream (see what I did there?).<\/p><p>Tell Make \u201cthis Makefile is written with Bash as the shell\u201d by adding this to the top of the Makefile:<\/p><pre>SHELL := bash\n<\/pre><p>With a well-known shell targeted, you can confidently skip workarounds for cross-shell compatibility, and use modern bash features.<\/p><p>The key message here, of course, is to choose a specific shell. If you'd rather use ZSH, or Python or Node for that matter, set it to that. Pick a specific language so you can stop targeting a lowest common denominator\/<\/p><h3 id=\"use-bash-strict-mode\">Use bash strict mode<\/h3><p>This holds true any time you use bash, but matters particularly here because without this your build may keep executing even if there was a failure in one of the targets.<\/p><p>There's a good writeup of what these flags do <a href=\"http:\/\/redsymbol.net\/articles\/unofficial-bash-strict-mode\/\">here<\/a>.<\/p><p>Add this to your Makefile:<\/p><pre>.SHELLFLAGS := -eu -o pipefail -c  \n<\/pre><h3 id=\"change-some-make-defaults\">Change some Make defaults<\/h3><p><code>.ONESHELL<\/code> ensures each Make recipe is ran as one single shell session, rather than one new shell per line. This both - in my opinion - is more intuitive, and it lets you do things like loops, variable assignments and so on in bash.<\/p><p><code>.DELETE_ON_ERROR<\/code> does what it says on the box - if a Make rule fails, it's target file is deleted. This ensures the next time you run Make, it'll properly re-run the failed rule, and guards against broken files.<\/p><pre>MAKEFLAGS += --warn-undefined-variables\n<\/pre><p>This does what it says on the tin - if you are referring to Make variables that don't exist, that's probably wrong and it's good to get a warning.<\/p><pre>MAKEFLAGS += --no-builtin-rules\n<\/pre><p>This disables the <a href=\"https:\/\/www.gnu.org\/software\/make\/manual\/html_node\/Catalogue-of-Rules.html\">bewildering array<\/a> of built in rules to automatically build Yacc grammars out of your data if you accidentally add the wrong file suffix. Remove magic, don't do things unless I tell you to, Make.<\/p><h3 id=\"tldr\">TL;DR<\/h3><p>Use this preamble in your Makefiles:<\/p><pre>SHELL := bash\n.ONESHELL:\n.SHELLFLAGS := -eu -o pipefail -c\n.DELETE_ON_ERROR:\nMAKEFLAGS += --warn-undefined-variables\nMAKEFLAGS += --no-builtin-rules\nifeq ($(origin .RECIPEPREFIX), undefined)\n  $(error This Make does not support .RECIPEPREFIX. Please use GNU Make 4.0 or later)\nendif\n.RECIPEPREFIX = &gt;\n<\/pre><h2 id=\"the-file-system\">The file system<\/h2><p>The trigger for this post was a Makefile that was sent to me as the official \u201cthis is how you do it\u201d example from a large software company.<\/p><p>Like all the Makefiles I wrote before Ben showed me better, it used Make without regard for its connection to the file system. In this rendition, Make is a tool for writing tab-indented named shell snippets that depend on each other.<\/p><p>These uses would look something like this:<\/p><pre>test:\n&gt; npm run test\ndeploy: test\n&gt; gsutil .. \n<\/pre><p>This is bad. Make sees each target (\u201ctest\u201d and \u201cdeploy\u201d) as the output file name of the associated block of code. If the output file exists, Make may decide that it doesn't need to run the code, since the file is already there.<\/p><p>Imagine you stashed some data in a file named <code>test<\/code>. The above Makefile will see that <code>test<\/code> is a file that exists, and that there are no dependencies for the <code>test<\/code> rule. Hence, no need to \u201crebuild\u201d the test file, and voila you've deployed to prod without running tests.<\/p><p>If you strictly want a target that generates no files, tell Make by saying it is \u201c.PHONY\u201d:<\/p><pre>test:\n&gt; npm run test\n.PHONY: test  # Make will not look for a file named `test` on the file system\n<\/pre><h3 id=\"generating-output-files\">Generating output files<\/h3><p>Briefly, then, each Make rule follows a convention like<\/p><pre>&lt;output-file&gt;: &lt;input file1&gt; &lt;input file 2&gt; &lt;input file n&gt;\n&gt; &lt;script to create output-file from input files&gt;\n<\/pre><p>Make calls the output the \u201ctarget\u201d, the inputs the \u201cprerequisites\u201d and the script the \u201crecipe\u201d. The entire thing is called a \u201crule\u201d.<\/p><p>Say you want to build a docker image for your little Node app. The output file might be a small file that records the image id:<\/p><pre>out\/image-id: \n&gt; image_id=\"example.com\/my-app:$$(pwgen -1)\"\n&gt; docker build --tag=\"$${image_id}\n&gt; echo \"$${image_id}\" &gt; out\/image-id\n<\/pre><p>Minor asides:<\/p><ul><li>Note the use of $$ instead of $ for bash variables and subshells; Make uses $ for it's own templating variables, $$ escapes it so bash sees a single $.<\/li>\n<li>Please ignore the pwgen bit, just need something that generates a random id for the image; much better option here is to hash the image inputs<\/li>\n<\/ul><p>This Make rule now will generate an output file that contains our image id. Other rules can depend on it, and it's easy for us to use the image id in deploy scripts and so on.<\/p><p>However, <em>there are problems here<\/em>. Our rule specifies no prerequisites, so once the file exists, Make will never rebuild it.<\/p><h3 id=\"specifying-inputs\">Specifying inputs<\/h3><p>To have make understand when to rebuild our docker image and when not to, we need to tell it what our docker image is built out of. When those prerequisites change, it needs to rebuild the image.<\/p><pre>out\/image-id: src\/main.js src\/views.js\n&gt; image_id=\"example.com\/my-app:$$(pwgen -1)\"\n&gt; docker build --tag=\"$${image_id}\n&gt; echo \"$${image_id}\" &gt; out\/image-id\n<\/pre><p>Now, if you made a change to <code>main.js<\/code> or <code>views.js<\/code>, Make will rebuild the image. However, maintaining a list of all the source files is a massive pain - and of course unnecessary, the file system already knows this information, so lets just ask it.<\/p><p>Like mentioned earlier, Make recognizes $-prefixed variables and commands. One such command is <code>shell<\/code>, which lets us run a shell command and have its output substituted into our Makefile.<\/p><p>To find all the source files, you might do something like: <code>find src\/ -type f<\/code>. Simply replace the manual listing of prerequisite files with this.<\/p><pre>out\/image-id: $(shell find src -type f)\n&gt; image_id=\"example.com\/my-app:$$(pwgen -1)\"\n&gt; docker build --tag=\"$${image_id}\n&gt; echo \"$${image_id}\" &gt; out\/image-id\n<\/pre><h2 id=\"sentinel-files\">Sentinel files<\/h2><p>It is possible to have one rule generate multiple files. For instance, the Webpack webapp bundler might yield lots of output files.<\/p><pre>index.html: $(shell find src -type f)\n&gt; node run webpack ..\n<\/pre><p>Here, Make will only rebuild the target if <code>index.html<\/code> is older than either of the dependencies. What if some other file outputted by webpack is much older? You can write a target with a templated output to have Make handle this. If you hate yourself.<\/p><p>This becomes particularly insane once you start running Make in parallel.<\/p><p>A better option is to have rules that yield lots of files have a \u201csentinel\u201d output. A single file that you know Make will look at to tell if a rebuild is needed:<\/p><pre>out\/.packed.sentinel: $(shell find src -type f)\n&gt; node run webpack ..\n&gt; touch out\/.packed.sentinel\n<\/pre><h2 id=\"on-make-magic-variables\">On Make magic variables<\/h2><p>Make has a bunch of cryptic magic variables that refer to things like the targets and prerequisites of rules. I mostly think these should be avoided, because they are hard to read.<\/p><p>However, for the sentinel file pattern, the magic variable <code>$(@D)<\/code>, which refers to the directory the target should go in, and <code>$@<\/code>, which refers to the target, are common enough that you quickly learn to recognize what they mean:<\/p><pre>out\/.packed.sentinel: $(shell find src -type f)\n&gt; mkdir -p $(@D) # expands to `mkdir -p out`\n&gt; node run webpack ..\n&gt; touch $@  # expands to `touch out\/.packed.sentinel`\n<\/pre><h2 id=\"putting-it-all-together\">Putting it all together<\/h2><p>In the end, for a Makefile to build a Node app and package it into a docker image, you might do something like:<\/p><pre>SHELL := bash\n.ONESHELL:\n.SHELLFLAGS := -eu -o pipefail -c\n.DELETE_ON_ERROR:\nMAKEFLAGS += --warn-undefined-variables\nMAKEFLAGS += --no-builtin-rules\nifeq ($(origin .RECIPEPREFIX), undefined)\n  $(error This Make does not support .RECIPEPREFIX. Please use GNU Make 4.0 or later)\nendif\n.RECIPEPREFIX = &gt;\n# Default - top level rule is what gets ran when you run just `make`\nbuild: out\/image-id\n.PHONY: build\n# Clean up the output directories; since all the sentinel files go under tmp, this will cause everything to get rebuilt\nclean:\n&gt; rm -rf tmp\n&gt; rm -rf out\n.PHONY: clean\n# Tests - re-ran if any file under src has been changed since tmp\/.tests-passed.sentinel was last touched\ntmp\/.tests-passed.sentinel: $(shell find src -type f)\n&gt; mkdir -p $(@D)\n&gt; node run test\n&gt; touch $@ \n# Webpack - re-built if the tests have been rebuilt (and so, by proxy, whenever the source files have changed)\ntmp\/.packed.sentinel: tmp\/.tests-passed.sentinel\n&gt; mkdir -p $(@D)\n&gt; webpack ..\n&gt; touch $@\n# Docker image - re-built if the webpack output has been rebuilt\nout\/image-id: tmp\/.packed.sentinel\n&gt; mkdir -p $(@D)\n&gt; image_id=\"example.com\/my-app:$$(pwgen -1)\"\n&gt; docker build --tag=\"$${image_id}\n&gt; echo \"$${image_id}\" &gt; out\/image-id\n<\/pre><h2 id=\"this-is-not-dogma\">This is not dogma<\/h2><p>One of the virtues of Make, to me, is that it is pragmatic. Counterintuitively, I think part of the reason Make has managed to maintain simplicity is precisely because it has allowed some hacks to accumulate in the tool itself.<\/p><p>The alternative is to expand the model to be more generic, to incorporate new concepts, making the core model more complex.<\/p><p>In the end, the hacks have allowed the exceptional beauty of the core model - a way to describe file construction pipelines tightly connected with the file system tree - to remain simple.<\/p><hr \/><h3>See also<\/h3><ul><li><a href=\"https:\/\/tech.davis-hansson.com\/p\/congress-is-a-vm\/\">The United States Congress is an elegant stack machine<\/a><\/li>\n<li><a href=\"https:\/\/tech.davis-hansson.com\/p\/price-wars.html\">How we use Dynamic Programming to find the best price for our customers<\/a><\/li>\n<li><a href=\"https:\/\/tech.davis-hansson.com\/p\/clickbait\/\">I gave commit rights to someone I didn't know, I could never have guessed what happened next!<\/a><\/li>\n<li><a href=\"https:\/\/tech.davis-hansson.com\/p\/tx-patterns\/\">Asynchronous transactional patterns<\/a><\/li>\n<li><a href=\"https:\/\/tech.davis-hansson.com\/p\/professionalism\/\">On professionalism in software<\/a><\/li>\n<\/ul>","created_at":"2019-12-17T15:07:10+0000","updated_at":"2019-12-17T15:07:16+0000","published_at":"2019-12-15T00:00:00+0000","published_by":["Jacob Davis-Hansson"],"annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"tech.davis-hansson.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1061,"title":"Krabby","url":"https:\/\/krabby.netlify.com\/","content":"<p>Krabby is a web extension (for <a href=\"https:\/\/google.com\/chrome\/\">Chrome<\/a>, <a href=\"https:\/\/mozilla.org\/firefox\/\">Firefox<\/a> and <a href=\"https:\/\/surf.suckless.org\">surf<\/a>) for keyboard-based navigation, inspired by <a href=\"https:\/\/kakoune.org\">Kakoune<\/a>.<\/p><p>The main differences with existing extensions are: multiple selections, keyboard layout agnostic, SOV (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Subject%E2%80%93object%E2%80%93verb\">subject\u2013object\u2013verb<\/a>) constructs and simple interaction with external programs. It is also quite usable with the mouse.<\/p>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Subject%E2%80%93object%E2%80%93verb\">Subject\u2013object\u2013verb<\/a> order. Learn more at <a href=\"https:\/\/github.com\/alexherbo2\/krabby\/blob\/master\/doc\/differences-with-vim.md#a-linguistic-twist\" title=\"A linguistic twist\">Differences with Vi(m)<\/a>.<\/li>\n<li>Keyboard layout agnostic. Maps to physical keys, but displays with the <a href=\"https:\/\/en.wikipedia.org\/wiki\/QWERTY\">US layout<\/a>.<\/li>\n<li>Multiple selections as a central way of interacting<\/li>\n<li>Powerful selection manipulation primitives<\/li>\n<li>Link hints<\/li>\n<li>External editor support. Open your favorite editor on text inputs.<\/li>\n<li>Tab search<\/li>\n<li>Contextual commands<\/li>\n<li>Contextual help<\/li>\n<li>Simple interaction with external programs<\/li>\n<li>Usable with the mouse. You can for example create a selection by grabbing a link with the mouse and  to create the selection. Or create multiple selections by clicking on hints with . If you hold <kbd>Control<\/kbd> or <kbd>Command<\/kbd>, you can even make selections with the mouse selection area.<\/li>\n<li>Limited scope. Krabby is not <a href=\"https:\/\/vim.org\">Vim<\/a>, nor <a href=\"https:\/\/kakoune.org\">Kakoune<\/a>. Krabby is trying to be a navigation tool and some concepts from code editors don\u2019t translate well. For example, there is no concept of mode in Krabby, but a system of contexts.<\/li>\n<li>Composability. Composability is better than extensibility. Krabby tries hard to provide features that interact well with each other. For example,  and  are two building blocks to focus a link and yank its URL. Depending on the context,  will yank the page or a link URL.<\/li>\n<li>Modular (think <a href=\"https:\/\/gnu.org\/software\/emacs\/\">Emacs<\/a>). Krabby is divided into multiple small projects. At its core Krabby just holds the configuration and grabs its functionalities from other projects.<\/li>\n<li><a href=\"https:\/\/google.com\/chrome\/\">Chrome<\/a>, <a href=\"https:\/\/mozilla.org\/firefox\/\">Firefox<\/a> and <a href=\"https:\/\/surf.suckless.org\">surf<\/a> support. It has also been tested on <a href=\"https:\/\/opera.com\">Opera<\/a>, <a href=\"https:\/\/vivaldi.com\">Vivaldi<\/a> and <a href=\"https:\/\/brave.com\">Brave<\/a>.<\/li>\n<\/ul>\n<div class=\"carousel slides\"><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/aXaFt75lIqo\">Download 3-gatsu no Lion episodes from HorribleSubs<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/PJXCnRBkHDY\">Extract from Wikipedia the list of cities and towns in Russia<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/gp4_6VGXkOk\">Quickly move around a document with no table of contents<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/tgrmss3u2aE\">Tab search with dmenu<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/zgSx1AE6pig\">Playing videos in picture-in-picture mode<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/v2Jvk1rhIlc\">Opening links in a paragraph<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/gYTi-eXuWdI\">Play videos with mpv<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/gmx3eEm6L3M\">Copy selection with Pandoc<\/a><\/figcaption><\/figure><p><\/p><figure><figcaption><a href=\"https:\/\/youtu.be\/rRGmuJhopJg\">Copy multiple links<\/a><\/figcaption><\/figure><\/div>\n<ul><li>Make sure to deactivate your extension and browser bindings.<\/li>\n<li>Press  to activate Krabby.<\/li>\n<li>Press  for help.<\/li>\n<\/ul><h3>Notes<\/h3>\n<ul><li>Commands are bound to <strong>physical keys<\/strong> and <strong>displayed<\/strong> with the <a href=\"https:\/\/en.wikipedia.org\/wiki\/QWERTY\">US layout<\/a>.<\/li>\n<li>You can click on the keys themselves to run the commands.<\/li>\n<\/ul>\n<table class=\"tutorial\"><thead><tr><th>Keys<\/th>\n<th>Example<\/th>\n<\/tr><\/thead><tbody><tr id=\"1\"><td class=\"keys\">\n<\/td><td class=\"example\">Scroll down.<\/td>\n<\/tr><tr><td class=\"keys\">\n<\/td><td class=\"example\">Scroll up.<\/td>\n<\/tr><tr><td class=\"keys\">\n<\/td><td class=\"example\">Scroll to the bottom of the page.<\/td>\n<\/tr><tr><td class=\"keys\">\n<\/td><td class=\"example\">Scroll to the top of the page.<\/td>\n<\/tr><tr><td class=\"keys\">\n<\/td><td class=\"example\">Scroll down.<\/td>\n<\/tr><tr><td class=\"keys\">\n<\/td><td class=\"example\">Copy page address.<\/td>\n<\/tr><tr><td class=\"keys\">\n<\/td><td class=\"example\">Scroll down.<\/td>\n<\/tr><tr id=\"2\"><td class=\"keys\">\n<ul><li> <code>{hint}<\/code><\/li>\n<li>\n<\/li><\/ul><\/td>\n<td class=\"example\">\n<figure><a href=\"https:\/\/www.pokemon.com\" target=\"_blank\"><img src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/9\/98\/International_Pok%C3%A9mon_logo.svg\" height=\"128\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a>\n<figcaption>\n<\/figcaption><\/figure><\/td>\n<\/tr><tr id=\"3\"><td class=\"keys\">\n<ul><li> <code>{hint}<\/code><\/li>\n<li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li> or  or <\/li>\n<li>\n<\/li><\/ul><\/td>\n<td class=\"example\">In <a href=\"https:\/\/www.pokemon.com\/us\/pokemon-video-games\/pokemon-red-version-and-pokemon-blue-version\/\">Pok\u00e9mon Red and Blue<\/a> and <a href=\"https:\/\/www.pokemon.com\/us\/pokemon-video-games\/pokemon-firered-version-and-pokemon-leafgreen-version\/\">Pok\u00e9mon FireRed and LeafGreen<\/a>, the player receives one of the Kanto starters as his first Pok\u00e9mon from <a href=\"https:\/\/www.pokemon.com\/us\/pokemon-news\/trainer-spotlight-professor-oak\/\">Professor Oak<\/a>. The player\u2019s rival receives the starter Pok\u00e9mon whose type is stronger against the player\u2019s chosen starter.<\/td>\n<\/tr><tr id=\"4\"><td class=\"keys\">\n<ul><li> <code>{hint}<\/code><\/li>\n<li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><\/ul><\/td>\n<td class=\"example\">\n<table><tr><th><a href=\"https:\/\/www.pokemon.com\/us\/pokedex\/bulbasaur\">Bulbasaur<\/a><\/th>\n<th><a href=\"https:\/\/www.pokemon.com\/us\/pokedex\/charmander\">Charmander<\/a><\/th>\n<th><a href=\"https:\/\/www.pokemon.com\/us\/pokedex\/squirtle\">Squirtle<\/a><\/th>\n<\/tr><tr><td><a href=\"https:\/\/www.pokemon.com\/us\/pokedex\/bulbasaur\"><img src=\"https:\/\/assets.pokemon.com\/assets\/cms2\/img\/pokedex\/full\/001.png\" height=\"128\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td><a href=\"https:\/\/www.pokemon.com\/us\/pokedex\/charmander\"><img src=\"https:\/\/assets.pokemon.com\/assets\/cms2\/img\/pokedex\/full\/004.png\" height=\"128\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td><a href=\"https:\/\/www.pokemon.com\/us\/pokedex\/squirtle\"><img src=\"https:\/\/assets.pokemon.com\/assets\/cms2\/img\/pokedex\/full\/007.png\" height=\"128\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<\/tr><tr><td>\n<ul><li class=\"pokemon type grass\"><a href=\"https:\/\/www.pokemon.com\/us\/pokedex?type=grass\">Grass<\/a><\/li>\n<li class=\"pokemon type poison\"><a href=\"https:\/\/www.pokemon.com\/us\/pokedex?type=poison\">Poison<\/a><\/li>\n<\/ul><\/td>\n<td>\n<\/td>\n<td>\n<\/td>\n<\/tr><\/table><\/td>\n<\/tr><tr id=\"5\"><td class=\"keys\">\n<ul><li> <code>{hint}<\/code><\/li>\n<li>\n<\/li><\/ul><\/td>\n<td class=\"example\"><label for=\"pokemon-nickname\">Pok\u00e9mon\u2019s nickname:<\/label> <\/td>\n<\/tr><tr id=\"6\"><td class=\"keys top\">\n<ul id=\"6.1\"><li> <code>{hint}<\/code><\/li>\n<li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><li>\n<\/li><\/ul><\/td>\n<td class=\"example\">\n<\/td>\n<\/tr><\/tbody><\/table>","created_at":"2019-12-15T13:05:28+0000","updated_at":"2019-12-15T13:05:28+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"krabby.netlify.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["design","consultant"],"is_public":false,"id":1060,"title":"About \u2014 Cuberto","url":"https:\/\/cuberto.com\/about\/","content":"<div class=\"cb-view\" data-controller=\"aboutController\" id=\"view-main\"><section class=\"cb-description\"><div class=\"cb-description-wrap cb-container -lg cb-description-grid -middle\"><h2><strong>Who<\/strong> we are<\/h2><div class=\"cb-description-grid-col cb-description-text -full\" data-aos=\"fade-up\"><p>Cuberto is a passionate, tight-knit team of premier specialists who can take on any challenge in the sphere of web and mobile app development. We love what we do, but what really makes us shine is years of experience and a deep knowledge of internal business processes that empower us to develop products which are exactly as they should be.<\/p><p>We never take the easy way and keep up with the times, always sticking to the newest trends in design and development.<\/p><\/div><\/div><\/section><section class=\"cb-award\"><div class=\"cb-award-wrap cb-container -lg\"><h2><strong>Achievements<\/strong><\/h2><div class=\"cb-award-ach\"><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/awwwards-2.png\" srcset=\"\/assets\/img\/awards\/awwwards-2@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x5<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/fwa-2.png\" srcset=\"\/assets\/img\/awards\/fwa-2@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x1<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/css-design.png\" srcset=\"\/assets\/img\/awards\/css-design@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x7<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/tagline.png\" srcset=\"\/assets\/img\/awards\/tagline@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x1<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/tappawards.png\" srcset=\"\/assets\/img\/awards\/tappawards@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x3<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/bma.png\" srcset=\"\/assets\/img\/awards\/bma@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x2<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/gs_greyen.png\" srcset=\"\/assets\/img\/awards\/gs_greyen@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x3<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/ratingruneta.png\" srcset=\"\/assets\/img\/awards\/ratingruneta@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x2<\/p><\/div><div class=\"cb-award-ach-item\" data-aos=\"fade-up\"><div class=\"cb-award-ach-item-img\"><img src=\"https:\/\/cuberto.com\/assets\/img\/awards\/dxd-masthead.png\" srcset=\"\/assets\/img\/awards\/dxd-masthead@2x.png 2x\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>x2<\/p><\/div><\/div><\/div><\/section><p><\/p><h2>Clients<\/h2>\n<em>Most of our clients are startups for whom we develop gorgeous, memorable branding.<\/em><section class=\"cb-request -half\" data-aos=\"fade-up\"><p><a href=\"https:\/\/cuberto.com\/contacts\/\"><strong>Send<\/strong> a Request<\/a><\/p><\/section><\/div><p><\/p><h2>Thank you!<br \/><strong>We will contact you soon.<\/strong><\/h2>\n<h2>An error occurred...<br \/><strong>Check form data and try again or email us.<\/strong><\/h2>\n<h4><strong>What are you looking for?<\/strong><em>Drop us a line<\/em><\/h4>","created_at":"2019-12-11T14:48:03+0000","updated_at":"2019-12-11T14:50:29+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"cuberto.com","preview_picture":"https:\/\/cuberto.com\/assets\/img\/about\/bg@2x.png)","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["python","mypy"],"is_public":false,"id":1059,"title":"Python Type Checking (Guide)","url":"https:\/\/realpython.com\/python-type-checking\/","content":"<div class=\"border rounded p-3 card mb-2\"><p class=\"mb-0\">Watch Now This tutorial has a related video course created by the Real Python team. Watch it together with the written tutorial to deepen your understanding: <a class=\"stretched-link text-success\" href=\"https:\/\/realpython.com\/courses\/python-type-checking\/\"><strong>Python Type Checking<\/strong><\/a><\/p><\/div><p>In this guide, you will get a look into Python type checking. Traditionally, types have been handled by the Python interpreter in a flexible but implicit way. Recent versions of Python allow you to specify explicit type hints that can be used by different tools to help you develop your code more efficiently.<\/p><p><strong>In this tutorial, you\u2019ll learn about the following:<\/strong><\/p><ul><li>Type annotations and type hints<\/li>\n<li>Adding static types to code, both your code and the code of others<\/li>\n<li>Running a static type checker<\/li>\n<li>Enforcing types at runtime<\/li>\n<\/ul><p>This is a comprehensive guide that will cover a lot of ground. If you want to just get a quick glimpse of how type hints work in Python, and see whether type checking is something you would include in your code, you don\u2019t need to read all of it. The two sections <a href=\"#hello-types\">Hello Types<\/a> and <a href=\"#pros-and-cons\">Pros and Cons<\/a> will give you a taste of how type checking works and recommendations about when it\u2019ll be useful.<\/p><div class=\"alert alert-warning\" role=\"alert\"><p><strong>Free Bonus:<\/strong> <a href=\"https:\/\/realpython.com\/python-type-checking\/\" class=\"alert-link\" data-toggle=\"modal\" data-target=\"#modal-python-mastery-course\" data-focus=\"false\">5 Thoughts On Python Mastery<\/a>, a free course for Python developers that shows you the roadmap and the mindset you'll need to take your Python skills to the next level.<\/p><\/div><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"type-systems\">Type Systems<\/h2><p>All programming languages include some kind of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Type_system\">type system<\/a> that formalizes which categories of objects it can work with and how those categories are treated. For instance, a type system can define a numerical type, with <code>42<\/code> as one example of an object of numerical type.<\/p><h3 id=\"dynamic-typing\">Dynamic Typing<\/h3><p>Python is a dynamically typed language. This means that the Python interpreter does type checking only as code runs, and that the type of a variable is allowed to change over its lifetime. The following dummy examples demonstrate that Python has dynamic typing:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; if False:\n...     1 + \"two\"  # This line never runs, so no TypeError is raised\n... else:\n...     1 + 2\n...\n3\n&gt;&gt;&gt; 1 + \"two\"  # Now this is type checked, and a TypeError is raised\nTypeError: unsupported operand type(s) for +: 'int' and 'str'\n<\/pre><\/div><p>In the first example, the branch <code>1 + \"two\"<\/code> never runs so it\u2019s never type checked. The second example shows that when <code>1 + \"two\"<\/code> is evaluated it raises a <code>TypeError<\/code> since you can\u2019t add an integer and a string in Python.<\/p><p>Next, let\u2019s see if variables can change type:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; thing = \"Hello\"\n&gt;&gt;&gt; type(thing)\n&lt;class 'str'&gt;\n&gt;&gt;&gt; thing = 28.1\n&gt;&gt;&gt; type(thing)\n&lt;class 'float'&gt;\n<\/pre><\/div><p><code>type()<\/code> returns the type of an object. These examples confirm that the type of <code>thing<\/code> is allowed to change, and Python correctly infers the type as it changes.<\/p><h3 id=\"static-typing\">Static Typing<\/h3><p>The opposite of dynamic typing is static typing. Static type checks are performed without running the program. In most statically typed languages, for instance C and Java, this is done as your program is compiled.<\/p><p>With static typing, variables generally are not allowed to change types, although mechanisms for casting a variable to a different type may exist.<\/p><p>Let\u2019s look at a quick example from a statically typed language. Consider the following Java snippet:<\/p><div class=\"highlight java\"><pre>String thing;\nthing = \"Hello\";\n<\/pre><\/div><p>The first line declares that the variable name <code>thing<\/code> is bound to the <code>String<\/code> type at compile time. The name can never be rebound to another type. In the second line, <code>thing<\/code> is assigned a value. It can never be assigned a value that is not a <code>String<\/code> object. For instance, if you were to later say <code>thing = 28.1f<\/code> the compiler would raise an error because of incompatible types.<\/p><p>Python will always <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0484\/#non-goals\">remain a dynamically typed language<\/a>. However, <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0484\/\">PEP 484<\/a> introduced type hints, which make it possible to also do static type checking of Python code.<\/p><p>Unlike how types work in most other statically typed languages, type hints by themselves don\u2019t cause Python to enforce types. As the name says, type hints just suggest types. There are other tools, which <a href=\"#static-type-checking\">you\u2019ll see later<\/a>, that perform static type checking using type hints.<\/p><h3 id=\"duck-typing\">Duck Typing<\/h3><p>Another term that is often used when talking about Python is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Duck_typing\">duck typing<\/a>. This moniker comes from the phrase \u201cif it walks like a duck and it quacks like a duck, then it must be a duck\u201d (or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Duck_test#History\">any of its variations<\/a>).<\/p><p>Duck typing is a concept related to dynamic typing, where the type or the class of an object is less important than the methods it defines. Using duck typing you do not check types at all. Instead you check for the presence of a given method or attribute.<\/p><p>As an example, you can call <code>len()<\/code> on any Python object that defines a <code>.__len__()<\/code> method:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; class TheHobbit:\n...     def __len__(self):\n...         return 95022\n...\n&gt;&gt;&gt; the_hobbit = TheHobbit()\n&gt;&gt;&gt; len(the_hobbit)\n95022\n<\/pre><\/div><p>Note that the call to <code>len()<\/code> gives the return value of the <code>.__len__()<\/code> method. In fact, the implementation of <code>len()<\/code> is essentially equivalent to the following:<\/p><div class=\"highlight python\"><pre>def len(obj):\n    return obj.__len__()\n<\/pre><\/div><p>In order to call <code>len(obj)<\/code>, the only real constraint on <code>obj<\/code> is that it must define a <code>.__len__()<\/code> method. Otherwise, the object can be of types as different as <code>str<\/code>, <code>list<\/code>, <code>dict<\/code>, or <code>TheHobbit<\/code>.<\/p><p>Duck typing is somewhat supported when doing static type checking of Python code, using <a href=\"https:\/\/en.wikipedia.org\/wiki\/Structural_type_system\">structural subtyping<\/a>. You\u2019ll learn <a href=\"#duck-types-and-protocols\">more about duck typing<\/a> later.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"hello-types\">Hello Types<\/h2><p>In this section you\u2019ll see how to add type hints to a function. The following function turns a text string into a headline by adding proper capitalization and a decorative line:<\/p><div class=\"highlight python\"><pre>def headline(text, align=True):\n    if align:\n        return f\"{text.title()}\\n{'-' * len(text)}\"\n    else:\n        return f\" {text.title()} \".center(50, \"o\")\n<\/pre><\/div><p>By default the function returns the headline left aligned with an underline. By setting the <code>align<\/code> flag to <code>False<\/code> you can alternatively have the headline be centered with a surrounding line of <code>o<\/code>:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; print(headline(\"python type checking\"))\nPython Type Checking\n--------------------\n&gt;&gt;&gt; print(headline(\"python type checking\", align=False))\noooooooooooooo Python Type Checking oooooooooooooo\n<\/pre><\/div><p>It\u2019s time for our first type hints! To add information about types to the function, you simply annotate its arguments and return value as follows:<\/p><div class=\"highlight python\"><pre>def headline(text: str, align: bool = True) -&gt; str:\n    ...\n<\/pre><\/div><p>The <code>text: str<\/code> syntax says that the <code>text<\/code> argument should be of type <code>str<\/code>. Similarly, the optional <code>align<\/code> argument should have type <code>bool<\/code> with the default value <code>True<\/code>. Finally, the <code>-&gt; str<\/code> notation specifies that <code>headline()<\/code> will return a string.<\/p><p>In <a href=\"https:\/\/realpython.com\/python-pep8\/\">terms of style<\/a>, <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0008\/#other-recommendations\">PEP 8<\/a> recommends the following:<\/p><ul><li>Use normal rules for colons, that is, no space before and one space after a colon: <code>text: str<\/code>.<\/li>\n<li>Use spaces around the <code>=<\/code> sign when combining an argument annotation with a default value: <code>align: bool = True<\/code>.<\/li>\n<li>Use spaces around the <code>-&gt;<\/code> arrow: <code>def headline(...) -&gt; str<\/code>.<\/li>\n<\/ul><p>Adding type hints like this has no runtime effect: they are only hints and are not enforced on their own. For instance, if we use a wrong type for the (admittedly badly named) <code>align<\/code> argument, the code still runs without any problems or warnings:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; print(headline(\"python type checking\", align=\"left\"))\nPython Type Checking\n--------------------\n<\/pre><\/div><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note:<\/strong> The reason this seemingly works is that the string <code>\"left\"<\/code> <a href=\"https:\/\/realpython.com\/python-operators-expressions\/#evaluation-of-non-boolean-values-in-boolean-context\">compares as truthy<\/a>. Using <code>align=\"center\"<\/code> would not have the desired effect as <code>\"center\"<\/code> is also truthy.<\/p><\/div><p>To catch this kind of error you can use a static type checker. That is, a tool that checks the types of your code without actually running it in the traditional sense.<\/p><p>You might already have such a type checker built into your editor. For instance <a href=\"https:\/\/www.jetbrains.com\/pycharm\/\">PyCharm<\/a> immediately gives you a warning:<\/p><p><a href=\"https:\/\/files.realpython.com\/media\/pycharm_type_error.76a49b9d4ff1.png\" target=\"_blank\"><img class=\"img-fluid mx-auto d-block border w-66\" src=\"https:\/\/files.realpython.com\/media\/pycharm_type_error.76a49b9d4ff1.png\" width=\"748\" height=\"297\" srcset=\"https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/pycharm_type_error.76a49b9d4ff1.png&amp;w=187&amp;sig=a258fbc862d03ec121a257be645b15168d1fb727 187w, https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/pycharm_type_error.76a49b9d4ff1.png&amp;w=374&amp;sig=9510f9ce9f753dc9f55c8f9244dde031b4fb660f 374w, https:\/\/files.realpython.com\/media\/pycharm_type_error.76a49b9d4ff1.png 748w\" alt=\"PyCharm flagging a type error\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>The most common tool for doing type checking is <a href=\"http:\/\/mypy-lang.org\/\">Mypy<\/a> though. You\u2019ll get a short introduction to Mypy in a moment, while you can learn much more about how it works <a href=\"#the-mypy-project\">later<\/a>.<\/p><p>If you don\u2019t already have Mypy on your system, you can install it using <code>pip<\/code>:<\/p><div class=\"highlight sh\"><pre>$ pip install mypy\n<\/pre><\/div><p>Put the following code in a file called <code>headlines.py<\/code>:<\/p><div class=\"highlight python\"><pre> 1 # headlines.py\n 2 \n 3 def headline(text: str, align: bool = True) -&gt; str:\n 4     if align:\n 5         return f\"{text.title()}\\n{'-' * len(text)}\"\n 6     else:\n 7         return f\" {text.title()} \".center(50, \"o\")\n 8 \n 9 print(headline(\"python type checking\"))\n10 print(headline(\"use mypy\", align=\"center\"))\n<\/pre><\/div><p>This is essentially the same code you saw earlier: the definition of <code>headline()<\/code> and two examples that are using it.<\/p><p>Now run Mypy on this code:<\/p><div class=\"highlight sh\"><pre>$ mypy headlines.py\nheadlines.py:10: error: Argument \"align\" to \"headline\" has incompatible\n                        type \"str\"; expected \"bool\"\n<\/pre><\/div><p>Based on the type hints, Mypy is able to tell us that we are using the wrong type on line 10.<\/p><p>To fix the issue in the code you should change the value of the <code>align<\/code> argument you are passing in. You might also rename the <code>align<\/code> flag to something less confusing:<\/p><div class=\"highlight python\"><pre> 1 # headlines.py\n 2 \n 3 def headline(text: str, centered: bool = False) -&gt; str:\n 4     if not centered:\n 5         return f\"{text.title()}\\n{'-' * len(text)}\"\n 6     else:\n 7         return f\" {text.title()} \".center(50, \"o\")\n 8 \n 9 print(headline(\"python type checking\"))\n10 print(headline(\"use mypy\", centered=True))\n<\/pre><\/div><p>Here you\u2019ve changed <code>align<\/code> to <code>centered<\/code>, and correctly used a boolean value for <code>centered<\/code> when calling <code>headline()<\/code>. The code now passes Mypy:<\/p><div class=\"highlight sh\"><pre>$ mypy headlines.py\n$ \n<\/pre><\/div><p>No output from Mypy means that no type errors were detected. Furthermore, when you run the code you see the expected output:<\/p><div class=\"highlight sh\"><pre>$ python headlines.py\nPython Type Checking\n--------------------\noooooooooooooooooooo Use Mypy oooooooooooooooooooo\n<\/pre><\/div><p>The first headline is aligned to the left, while the second one is centered.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"pros-and-cons\">Pros and Cons<\/h2><p>The previous section gave you a little taste of what type checking in Python looks like. You also saw an example of one of the advantages of adding types to your code: type hints help <strong>catch certain errors<\/strong>. Other advantages include:<\/p><ul><li>\n<p>Type hints help <strong>document your code<\/strong>. Traditionally, you would use <a href=\"https:\/\/realpython.com\/documenting-python-code\/\">docstrings<\/a> if you wanted to document the expected types of a function\u2019s arguments. This works, but as there is no standard for docstrings (despite <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0257\/\">PEP 257<\/a> they can\u2019t be easily used for automatic checks.<\/p>\n<\/li>\n<li>\n<p>Type hints <strong>improve IDEs and linters<\/strong>. They make it much easier to statically reason about your code. This in turn allows IDEs to offer better code completion and similar features. With the type annotation, PyCharm knows that <code>text<\/code> is a string, and can give specific suggestions based on this:<\/p>\n<p><a href=\"https:\/\/files.realpython.com\/media\/pycharm_code_completion.82857c2750f6.png\" target=\"_blank\"><img class=\"img-fluid mx-auto d-block border w-66\" src=\"https:\/\/files.realpython.com\/media\/pycharm_code_completion.82857c2750f6.png\" width=\"1122\" height=\"515\" srcset=\"https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/pycharm_code_completion.82857c2750f6.png&amp;w=280&amp;sig=1003ed3b6f9346a44ac4690308feb53cec74b210 280w, https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/pycharm_code_completion.82857c2750f6.png&amp;w=561&amp;sig=a0c5aba588155ab7faa3574a64aacbe2206a97c4 561w, https:\/\/files.realpython.com\/media\/pycharm_code_completion.82857c2750f6.png 1122w\" alt=\"Code completion in PyCharm on a typed variable\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<\/li>\n<li>\n<p>Type hints help you <strong>build and maintain a cleaner architecture<\/strong>. The act of writing type hints forces you to think about the types in your program. While the dynamic nature of Python is one of its great assets, being conscious about relying on duck typing, overloaded methods, or multiple return types is a good thing.<\/p>\n<\/li>\n<\/ul><p>Of course, static type checking is not all peaches and cream. There are also some downsides you should consider:<\/p><ul><li>\n<p>Type hints <strong>take developer time and effort to add<\/strong>. Even though it probably pays off in spending less time debugging, you will spend more time entering code.<\/p>\n<\/li>\n<li>\n<p>Type hints <strong>work best in modern Pythons<\/strong>. Annotations were introduced in Python 3.0, and it\u2019s possible to use <a href=\"#type-comments\">type comments<\/a> in Python 2.7. Still, improvements like <a href=\"#variable-annotations\">variable annotations<\/a> and <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0563\/\">postponed evaluation of type hints<\/a> mean that you\u2019ll have a better experience doing type checks using Python 3.6 or even <a href=\"https:\/\/realpython.com\/python37-new-features\/\">Python 3.7<\/a>.<\/p>\n<\/li>\n<li>\n<p>Type hints <strong>introduce a slight penalty in start-up time<\/strong>. If you need to use the <a href=\"#sequences-and-mappings\"><code>typing<\/code> module<\/a> the import time may be significant, especially in short scripts.<\/p>\n<\/li>\n<\/ul><div class=\"card mb-3\" id=\"collapse_cardca6bfe\"><div id=\"collapseca6bfe\" class=\"collapse card-body\" data-parent=\"#collapse_cardca6bfe\"><p>You\u2019ll <a href=\"#sequences-and-mappings\">later<\/a> learn about the <code>typing<\/code> module, and how it\u2019s necessary in most cases when you add type hints. Importing modules necessarily take some time, but how much?<\/p><p>To get some idea about this, create two files: <code>empty_file.py<\/code> should be an empty file, while <code>import_typing.py<\/code> should contain the following line:<\/p><div class=\"highlight python\"><pre>import typing\n<\/pre><\/div><p>On Linux it\u2019s quite easy to check how much time the <code>typing<\/code> import takes using the <a href=\"https:\/\/perf.wiki.kernel.org\/\"><code>perf<\/code> utility<\/a>:<\/p><div class=\"highlight sh\"><pre>$ perf stat -r 1000 python3.6 import_typing.py\n Performance counter stats for 'python3.6 import_typing.py' (1000 runs):\n [ ... extra information hidden for brevity ... ]\n       0.045161650 seconds time elapsed    ( +-  0.77% )\n<\/pre><\/div><p>So running the <code>import typing.py<\/code> script takes about 45 milliseconds. Of course this is not all time spent on importing <code>typing<\/code>. Some of this is overhead in starting the Python interpreter, so let\u2019s compare to running Python on an empty file:<\/p><div class=\"highlight sh\"><pre>$ perf stat -r 1000 python3.6 empty_file.py\n Performance counter stats for 'python3.6 empty_file.py' (1000 runs):\n [ ... extra information hidden for brevity ... ]\n       0.028077845 seconds time elapsed    ( +-  0.49% )\n<\/pre><\/div><p>Based on this test, the import of the <code>typing<\/code> module takes around 17 milliseconds on Python 3.6.<\/p><p>One of the advertised improvements in Python 3.7 is <a href=\"https:\/\/docs.python.org\/3\/whatsnew\/3.7.html#optimizations\">faster startup<\/a>. Let\u2019s see if the results are different:<\/p><div class=\"highlight sh\"><pre>$ perf stat -r 1000 python3.7 import_typing.py\n [...]\n       0.025979806 seconds time elapsed    ( +-  0.31% )\n$ perf stat -r 1000 python3.7 empty_file.py\n [...]\n       0.020002505 seconds time elapsed    ( +-  0.30% )\n<\/pre><\/div><p>Indeed, the general startup time is reduced by about 8 milliseconds, and the time to import <code>typing<\/code> is down from 17 to around 6 milliseconds\u2014almost 3 times faster.<\/p><p><strong>Using <code>timeit<\/code><\/strong><\/p><p>There are similar tools on other platforms. Python itself comes with the <code>timeit<\/code> module in the standard library. Typically, we would directly use <code>timeit<\/code> for the timings above. However, <code>timeit<\/code> struggles to time imports reliably because Python is clever about importing modules only once. Consider the following example:<\/p><div class=\"highlight sh\"><pre>$ python3.6 -m timeit \"import typing\"\n10000000 loops, best of 3: 0.134 usec per loop\n<\/pre><\/div><p>While you get a result, you should be suspicious about the result: 0.1 microsecond is more than 100000 times faster than what <code>perf<\/code> measured! What <code>timeit<\/code> has actually done is to run the <code>import typing<\/code> statement 30 million times, with Python actually only importing <code>typing<\/code> once.<\/p><p>To get reasonable results you can tell <code>timeit<\/code> to only run once:<\/p><div class=\"highlight sh\"><pre>$ python3.6 -m timeit -n 1 -r 1 \"import typing\"\n1 loops, best of 1: 9.77 msec per loop\n$ python3.7 -m timeit -n 1 -r 1 \"import typing\"\n1 loop, best of 1: 1.97 msec per loop\n<\/pre><\/div><p>These results are on the same scale as the results from <code>perf<\/code> above. However, since these are based on only one execution of the code, they are not as reliable as those based on multiple runs.<\/p><p>The conclusion in both these cases is that importing <code>typing<\/code> takes a few milliseconds. For the majority of programs and scripts you write this will probably not be an issue.<\/p><p><strong>The New <code>importtime<\/code> Option<\/strong><\/p><p>In Python 3.7 there is also a new command line option that can be used to figure out how much time imports take. Using <code>-X importtime<\/code> you\u2019ll get a report about all imports that are made:<\/p><div class=\"highlight sh\"><pre>$ python3.7 -X importtime import_typing.py\nimport time: self [us] | cumulative | imported package\n[ ... some information hidden for brevity ... ]\nimport time:       358 |        358 | zipimport\nimport time:      2107 |      14610 | site\nimport time:       272 |        272 |   collections.abc\nimport time:       664 |       3058 |   re\nimport time:      3044 |       6373 | typing\n<\/pre><\/div><p>This shows a similar result. Importing <code>typing<\/code> takes about 6 milliseconds. If you\u2019ll read the report closely you can notice that around half of this time is spent on importing the <code>collections.abc<\/code> and <code>re<\/code> modules which <code>typing<\/code> depends on.<\/p><\/div><\/div><p>So, should you use static type checking in your own code? Well, it\u2019s not an all-or-nothing question. Luckily, Python supports the concept of <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0483\/\">gradual typing<\/a>. This means that you can gradually introduce types into your code. Code without type hints will be ignored by the static type checker. Therefore, you can start adding types to critical components, and continue as long as it adds value to you.<\/p><p>Looking at the lists above of pros and cons you\u2019ll notice that adding types will have no effect on your running program or the users of your program. Type checking is meant to make your life as a developer better and more convenient.<\/p><p>A few rules of thumb on whether to add types to your project are:<\/p><ul><li>\n<p>If you are just beginning to learn Python, you can safely wait with type hints until you have more experience.<\/p>\n<\/li>\n<li>\n<p>Type hints add little value in <a href=\"https:\/\/www.youtube.com\/watch?v=Jd8ulMb6_ls\">short throw-away scripts<\/a>.<\/p>\n<\/li>\n<li>\n<p>In libraries that will be used by others, especially ones <a href=\"https:\/\/realpython.com\/pypi-publish-python-package\/\">published on PyPI<\/a>, type hints add a lot of value. Other code using your libraries need these type hints to be properly type checked itself. For examples of projects using type hints see <a href=\"https:\/\/github.com\/Bogdanp\/cursive_re\/blob\/master\/cursive_re\/exprs.py\"><code>cursive_re<\/code><\/a>, <a href=\"https:\/\/github.com\/ambv\/black\/blob\/master\/black.py\"><code>black<\/code><\/a>, our own <a href=\"https:\/\/github.com\/realpython\/reader\/blob\/master\/reader\/feed.py\">Real Python Reader<\/a>, and <a href=\"https:\/\/github.com\/python\/mypy\/blob\/master\/mypy\/build.py\">Mypy<\/a> itself.<\/p>\n<\/li>\n<li>\n<p>In bigger projects, type hints help you understand how types flow through your code, and are highly recommended. Even more so in projects where you cooperate with others.<\/p>\n<\/li>\n<\/ul><p>In his excellent article <a href=\"https:\/\/www.bernat.tech\/the-state-of-type-hints-in-python\/\">The State of Type Hints in Python<\/a> Bern\u00e1t G\u00e1bor recommends that \u201c<strong>type hints should be used whenever unit tests are worth writing<\/strong>.\u201d Indeed, type hints play a similar role as <a href=\"https:\/\/realpython.com\/python-testing\/\">tests<\/a> in your code: they help you as a developer write better code.<\/p><p>Hopefully you now have an idea about how type checking works in Python and whether it\u2019s something you would like to employ in your own projects.<\/p><p>In the rest of this guide, we\u2019ll go into more detail about the Python type system, including how you run static type checkers (with particular focus on Mypy), how you type check code that uses libraries without type hints, and how you use annotations at runtime.<\/p><h2 id=\"annotations\">Annotations<\/h2><p>Annotations were <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-3107\/\">introduced in Python 3.0<\/a>, originally without any specific purpose. They were simply a way to associate arbitrary expressions to function arguments and return values.<\/p><p>Years later, <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0484\/\">PEP 484<\/a> defined how to add type hints to your Python code, based off work that Jukka Lehtosalo had done on his Ph.D. project\u2014Mypy. The main way to add type hints is using annotations. As type checking is becoming more and more common, this also means that annotations should mainly be reserved for type hints.<\/p><p>The next sections explain how annotations work in the context of type hints.<\/p><p>For functions, you can annotate arguments and the return value. This is done as follows:<\/p><div class=\"highlight python\"><pre>def func(arg: arg_type, optarg: arg_type = default) -&gt; return_type:\n    ...\n<\/pre><\/div><p>For arguments the syntax is <code>argument: annotation<\/code>, while the return type is annotated using <code>-&gt; annotation<\/code>. Note that the annotation must be a valid Python expression.<\/p><p>The following simple example adds annotations to a function that calculates the circumference of a circle:<\/p><div class=\"highlight python\"><pre>import math\ndef circumference(radius: float) -&gt; float:\n    return 2 * math.pi * radius\n<\/pre><\/div><p>When running the code, you can also inspect the annotations. They are stored in a special <code>.__annotations__<\/code> attribute on the function:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; circumference(1.23)\n7.728317927830891\n&gt;&gt;&gt; circumference.__annotations__\n{'radius': &lt;class 'float'&gt;, 'return': &lt;class 'float'&gt;}\n<\/pre><\/div><p>Sometimes you might be confused by how Mypy is interpreting your type hints. For those cases there are special Mypy expressions: <code>reveal_type()<\/code> and <code>reveal_locals()<\/code>. You can add these to your code before running Mypy, and Mypy will dutifully report which types it has inferred. As an example, save the following code to <code>reveal.py<\/code>:<\/p><div class=\"highlight python\"><pre> 1 # reveal.py\n 2 \n 3 import math\n 4 reveal_type(math.pi)\n 5 \n 6 radius = 1\n 7 circumference = 2 * math.pi * radius\n 8 reveal_locals()\n<\/pre><\/div><p>Next, run this code through Mypy:<\/p><div class=\"highlight sh\"><pre>$ mypy reveal.py\nreveal.py:4: error: Revealed type is 'builtins.float'\nreveal.py:8: error: Revealed local types are:\nreveal.py:8: error: circumference: builtins.float\nreveal.py:8: error: radius: builtins.int\n<\/pre><\/div><p>Even without any annotations Mypy has correctly inferred the types of the built-in <code>math.pi<\/code>, as well as our local variables <code>radius<\/code> and <code>circumference<\/code>.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note:<\/strong> The reveal expressions are only meant as a tool helping you add types and debug your type hints. If you try to run the <code>reveal.py<\/code> file as a Python script it will crash with a <code>NameError<\/code> since <code>reveal_type()<\/code> is not a function known to the Python interpreter.<\/p><\/div><p>If Mypy says that \u201cName \u2018<code>reveal_locals<\/code>\u2018 is not defined\u201d you might need to update your Mypy installation. The <code>reveal_locals()<\/code> expression is available in <a href=\"http:\/\/mypy-lang.blogspot.com\/2018\/06\/mypy-0610-released.html\">Mypy version 0.610<\/a> and later.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"variable-annotations\">Variable Annotations<\/h3><p>In the definition of <code>circumference()<\/code> in the previous section, you only annotated the arguments and the return value. You did not add any annotations inside the function body. More often than not, this is enough.<\/p><p>However, sometimes the type checker needs help in figuring out the types of variables as well. Variable annotations were defined in <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0526\/\">PEP 526<\/a> and introduced in Python 3.6. The syntax is the same as for function argument annotations:<\/p><div class=\"highlight python\"><pre>pi: float = 3.142\ndef circumference(radius: float) -&gt; float:\n    return 2 * pi * radius\n<\/pre><\/div><p>The variable <code>pi<\/code> has been annotated with the <code>float<\/code> type hint.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note:<\/strong> Static type checkers are more than able to figure out that <code>3.142<\/code> is a float, so in this example the annotation of <code>pi<\/code> is not necessary. As you learn more about the Python type system, you\u2019ll see more relevant examples of variable annotations.<\/p><\/div><p>Annotations of variables are stored in the module level <code>__annotations__<\/code> dictionary:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; circumference(1)\n6.284\n&gt;&gt;&gt; __annotations__\n{'pi': &lt;class 'float'&gt;}\n<\/pre><\/div><p>You\u2019re allowed to annotate a variable without giving it a value. This adds the annotation to the <code>__annotations__<\/code> dictionary, while the variable remains undefined:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; nothing: str\n&gt;&gt;&gt; nothing\nNameError: name 'nothing' is not defined\n&gt;&gt;&gt; __annotations__\n{'nothing': &lt;class 'str'&gt;}\n<\/pre><\/div><p>Since no value was assigned to <code>nothing<\/code>, the name <code>nothing<\/code> is not yet defined.<\/p><p>As mentioned, annotations were introduced in Python 3, and they\u2019ve not been backported to Python 2. This means that if you\u2019re writing code that needs to support <a href=\"https:\/\/pythonclock.org\/\">legacy Python<\/a>, you can\u2019t use annotations.<\/p><p>Instead, you can use type comments. These are specially formatted comments that can be used to add type hints compatible with older code. To add type comments to a function you do something like this:<\/p><div class=\"highlight python\"><pre>import math\ndef circumference(radius):\n    # type: (float) -&gt; float\n    return 2 * math.pi * radius\n<\/pre><\/div><p>The type comments are just comments, so they can be used in any version of Python.<\/p><p>Type comments are handled directly by the type checker, so these types are not available in the <code>__annotations__<\/code> dictionary:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; circumference.__annotations__\n{}\n<\/pre><\/div><p>A type comment must start with the <code>type:<\/code> literal, and be on the same or the following line as the function definition. If you want to annotate a function with several arguments, you write each type separated by comma:<\/p><div class=\"highlight python\"><pre>def headline(text, width=80, fill_char=\"-\"):\n    # type: (str, int, str) -&gt; str\n    return f\" {text.title()} \".center(width, fill_char)\nprint(headline(\"type comments work\", width=40))\n<\/pre><\/div><p>You are also allowed to write each argument on a separate line with its own annotation:<\/p><div class=\"highlight python\"><pre> 1 # headlines.py\n 2 \n 3 def headline(\n 4     text,           # type: str\n 5     width=80,       # type: int\n 6     fill_char=\"-\",  # type: str\n 7 ):                  # type: (...) -&gt; str\n 8     return f\" {text.title()} \".center(width, fill_char)\n 9 \n10 print(headline(\"type comments work\", width=40))\n<\/pre><\/div><p>Run the example through Python and Mypy:<\/p><div class=\"highlight sh\"><pre>$  python headlines.py\n---------- Type Comments Work ----------\n$ mypy headline.py\n$\n<\/pre><\/div><p>If you have errors, for instance if you happened to call <code>headline()<\/code> with <code>width=\"full\"<\/code> on line 10, Mypy will tell you:<\/p><div class=\"highlight sh\"><pre>$ mypy headline.py\nheadline.py:10: error: Argument \"width\" to \"headline\" has incompatible\n                       type \"str\"; expected \"int\"\n<\/pre><\/div><p>You can also add type comments to variables. This is done similarly to how you add type comments to arguments:<\/p><div class=\"highlight python\"><pre>pi = 3.142  # type: float\n<\/pre><\/div><p>In this example, <code>pi<\/code> will be type checked as a float variable.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><p>Should you use annotations or type comments when adding type hints to your own code? In short: <strong>Use annotations if you can, use type comments if you must.<\/strong><\/p><p>Annotations provide a cleaner syntax keeping type information closer to your code. They are also the <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0484\/\">officially recommended way<\/a> of writing type hints, and will be further developed and properly maintained in the future.<\/p><p>Type comments are more verbose and might conflict with other kinds of comments in your code like <a href=\"https:\/\/realpython.com\/python-code-quality\/\">linter directives<\/a>. However, they can be used in code bases that don\u2019t support annotations.<\/p><p>There is also hidden option number three: <a href=\"https:\/\/github.com\/python\/mypy\/wiki\/Creating-Stubs-For-Python-Modules\">stub files<\/a>. You will learn about these later, when we discuss <a href=\"#adding-stubs\">adding types to third party libraries<\/a>.<\/p><p>Stub files will work in any version of Python, at the expense of having to maintain a second set of files. In general, you only want to use stub files if you can\u2019t change the original source code.<\/p><h2 id=\"playing-with-python-types-part-1\">Playing With Python Types, Part 1<\/h2><p>Up until now you\u2019ve only used basic types like <code>str<\/code>, <code>float<\/code>, and <code>bool<\/code> in your type hints. The Python type system is quite powerful, and supports many kinds of more complex types. This is necessary as it needs to be able to reasonably model Python\u2019s dynamic duck typing nature.<\/p><p>In this section you will learn more about this type system, while implementing a simple card game. You will see how to specify:<\/p><ul><li>The type of <a href=\"#sequences-and-mappings\">sequences and mappings<\/a> like tuples, lists and dictionaries<\/li>\n<li><a href=\"#type-aliases\">Type aliases<\/a> that make code easier to read<\/li>\n<li>That functions and methods <a href=\"#functions-without-return-values\">do not return anything<\/a><\/li>\n<li>Objects that may be of <a href=\"#the-any-type\">any type<\/a><\/li>\n<\/ul><p>After a short detour into some <a href=\"#type-theory\">type theory<\/a> you will then see <a href=\"#playing-with-python-types-part-2\">even more ways to specify types in Python<\/a>. You can find the code examples from this section <a href=\"https:\/\/github.com\/realpython\/materials\/tree\/master\/python-type-checking\">here<\/a>.<\/p><h3 id=\"example-a-deck-of-cards\">Example: A Deck of Cards<\/h3><p>The following example shows an implementation of a <a href=\"https:\/\/en.wikipedia.org\/wiki\/French_playing_cards\">regular (French) deck of cards<\/a>:<\/p><div class=\"highlight python\"><pre> 1 # game.py\n 2 \n 3 import random\n 4 \n 5 SUITS = \"\u2660 \u2661 \u2662 \u2663\".split()\n 6 RANKS = \"2 3 4 5 6 7 8 9 10 J Q K A\".split()\n 7 \n 8 def create_deck(shuffle=False):\n 9     \"\"\"Create a new deck of 52 cards\"\"\"\n10     deck = [(s, r) for r in RANKS for s in SUITS]\n11     if shuffle:\n12         random.shuffle(deck)\n13     return deck\n14 \n15 def deal_hands(deck):\n16     \"\"\"Deal the cards in the deck into four hands\"\"\"\n17     return (deck[0::4], deck[1::4], deck[2::4], deck[3::4])\n18 \n19 def play():\n20     \"\"\"Play a 4-player card game\"\"\"\n21     deck = create_deck(shuffle=True)\n22     names = \"P1 P2 P3 P4\".split()\n23     hands = {n: h for n, h in zip(names, deal_hands(deck))}\n24 \n25     for name, cards in hands.items():\n26         card_str = \" \".join(f\"{s}{r}\" for (s, r) in cards)\n27         print(f\"{name}: {card_str}\")\n28 \n29 if __name__ == \"__main__\":\n30     play()\n<\/pre><\/div><p>Each card is represented as a tuple of strings denoting the suit and rank. The deck is represented as a list of cards. <code>create_deck()<\/code> creates a regular deck of 52 playing cards, and optionally shuffles the cards. <code>deal_hands()<\/code> deals the deck of cards to four players.<\/p><p>Finally, <code>play()<\/code> plays the game. As of now, it only prepares for a card game by constructing a shuffled deck and dealing cards to each player. The following is a typical output:<\/p><div class=\"highlight sh\"><pre>$ python game.py\nP4: \u26639 \u26629 \u26612 \u26627 \u26617 \u2663A \u26606 \u2661K \u26615 \u26626 \u26623 \u26633 \u2663Q\nP1: \u2661A \u26602 \u266010 \u2662J \u266310 \u26634 \u26605 \u2661Q \u26625 \u26636 \u2660A \u26635 \u26624\nP2: \u26622 \u26607 \u26618 \u2662K \u26603 \u26613 \u2663K \u2660J \u2662A \u26637 \u26616 \u266110 \u2660K\nP3: \u26632 \u26638 \u26608 \u2663J \u2662Q \u26619 \u2661J \u26604 \u26628 \u266210 \u26609 \u26614 \u2660Q\n<\/pre><\/div><p>You will see how to extend this example into a more interesting game as we move along.<\/p><h3 id=\"sequences-and-mappings\">Sequences and Mappings<\/h3><p>Let\u2019s add type hints to our card game. In other words, let\u2019s annotate the functions <code>create_deck()<\/code>, <code>deal_hands()<\/code>, and <code>play()<\/code>. The first challenge is that you need to annotate composite types like the list used to represent the deck of cards and the tuples used to represent the cards themselves.<\/p><p>With simple types like <code>str<\/code>, <code>float<\/code>, and <code>bool<\/code>, adding type hints is as easy as using the type itself:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; name: str = \"Guido\"\n&gt;&gt;&gt; pi: float = 3.142\n&gt;&gt;&gt; centered: bool = False\n<\/pre><\/div><p>With composite types, you are allowed to do the same:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; names: list = [\"Guido\", \"Jukka\", \"Ivan\"]\n&gt;&gt;&gt; version: tuple = (3, 7, 1)\n&gt;&gt;&gt; options: dict = {\"centered\": False, \"capitalize\": True}\n<\/pre><\/div><p>However, this does not really tell the full story. What will be the types of <code>names[2]<\/code>, <code>version[0]<\/code>, and <code>options[\"centered\"]<\/code>? In this concrete case you can see that they are <code>str<\/code>, <code>int<\/code>, and <code>bool<\/code>, respectively. However, the type hints themselves give no information about this.<\/p><p>Instead, you should use the special types defined in the <a href=\"https:\/\/docs.python.org\/library\/typing.html\"><code>typing<\/code> module<\/a>. These types add syntax for specifying the types of elements of composite types. You can write the following:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; from typing import Dict, List, Tuple\n&gt;&gt;&gt; names: List[str] = [\"Guido\", \"Jukka\", \"Ivan\"]\n&gt;&gt;&gt; version: Tuple[int, int, int] = (3, 7, 1)\n&gt;&gt;&gt; options: Dict[str, bool] = {\"centered\": False, \"capitalize\": True}\n<\/pre><\/div><p>Note that each of these types start with a capital letter and that they all use square brackets to define item types:<\/p><ul><li><strong><code>names<\/code><\/strong> is a list of strings<\/li>\n<li><strong><code>version<\/code><\/strong> is a 3-tuple consisting of three integers<\/li>\n<li><strong><code>options<\/code><\/strong> is a dictionary mapping strings to boolean values<\/li>\n<\/ul><p>The <code>typing<\/code> module contains many more composite types, including <code>Counter<\/code>, <code>Deque<\/code>, <code>FrozenSet<\/code>, <code>NamedTuple<\/code>, and <code>Set<\/code>. In addition, the module includes other kinds of types that you\u2019ll see in later sections.<\/p><p>Let\u2019s return to the card game. A card is represented by a tuple of two strings. You can write this as <code>Tuple[str, str]<\/code>, so the type of the deck of cards becomes <code>List[Tuple[str, str]]<\/code>. Therefore you can annotate <code>create_deck()<\/code> as follows:<\/p><div class=\"highlight python\"><pre> 8 def create_deck(shuffle: bool = False) -&gt; List[Tuple[str, str]]:\n 9     \"\"\"Create a new deck of 52 cards\"\"\"\n10     deck = [(s, r) for r in RANKS for s in SUITS]\n11     if shuffle:\n12         random.shuffle(deck)\n13     return deck\n<\/pre><\/div><p>In addition to the return value, you\u2019ve also added the <code>bool<\/code> type to the optional <code>shuffle<\/code> argument.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note:<\/strong> Tuples and lists are annotated differently.<\/p><p>A tuple is an immutable sequence, and typically consists of a fixed number of possibly differently typed elements. For example, we represent a card as a tuple of suit and rank. In general, you write <code>Tuple[t_1, t_2, ..., t_n]<\/code> for an n-tuple.<\/p><p>A list is a mutable sequence and usually consists of an unknown number of elements of the same type, for instance a list of cards. No matter how many elements are in the list there is only one type in the annotation: <code>List[t]<\/code>.<\/p><\/div><p>In many cases your functions will expect some kind of <a href=\"https:\/\/docs.python.org\/glossary.html#term-sequence\">sequence<\/a>, and not really care whether it is a list or a tuple. In these cases you should use <code>typing.Sequence<\/code> when annotating the function argument:<\/p><div class=\"highlight python\"><pre>from typing import List, Sequence\ndef square(elems: Sequence[float]) -&gt; List[float]:\n    return [x**2 for x in elems]\n<\/pre><\/div><p>Using <code>Sequence<\/code> is an example of using duck typing. A <code>Sequence<\/code> is anything that supports <code>len()<\/code> and <code>.__getitem__()<\/code>, independent of its actual type.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"type-aliases\">Type Aliases<\/h3><p>The type hints might become quite oblique when working with nested types like the deck of cards. You may need to stare at <code>List[Tuple[str, str]]<\/code> a bit before figuring out that it matches our representation of a deck of cards.<\/p><p>Now consider how you would annotate <code>deal_hands()<\/code>:<\/p><div class=\"highlight python\"><pre>15 def deal_hands(\n16     deck: List[Tuple[str, str]]\n17 ) -&gt; Tuple[\n18     List[Tuple[str, str]],\n19     List[Tuple[str, str]],\n20     List[Tuple[str, str]],\n21     List[Tuple[str, str]],\n22 ]:\n23     \"\"\"Deal the cards in the deck into four hands\"\"\"\n24     return (deck[0::4], deck[1::4], deck[2::4], deck[3::4])\n<\/pre><\/div><p>That\u2019s just terrible!<\/p><p>Recall that type annotations are regular Python expressions. That means that you can define your own type aliases by assigning them to new variables. You can for instance create <code>Card<\/code> and <code>Deck<\/code> type aliases:<\/p><div class=\"highlight python\"><pre>from typing import List, Tuple\nCard = Tuple[str, str]\nDeck = List[Card]\n<\/pre><\/div><p><code>Card<\/code> can now be used in type hints or in the definition of new type aliases, like <code>Deck<\/code> in the example above.<\/p><p>Using these aliases, the annotations of <code>deal_hands()<\/code> become much more readable:<\/p><div class=\"highlight python\"><pre>15 def deal_hands(deck: Deck) -&gt; Tuple[Deck, Deck, Deck, Deck]:\n16     \"\"\"Deal the cards in the deck into four hands\"\"\"\n17     return (deck[0::4], deck[1::4], deck[2::4], deck[3::4])\n<\/pre><\/div><p>Type aliases are great for making your code and its intent clearer. At the same time, these aliases can be inspected to see what they represent:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; from typing import List, Tuple\n&gt;&gt;&gt; Card = Tuple[str, str]\n&gt;&gt;&gt; Deck = List[Card]\n&gt;&gt;&gt; Deck\ntyping.List[typing.Tuple[str, str]]\n<\/pre><\/div><p>Note that when printing <code>Deck<\/code>, it shows that it\u2019s an alias for a list of 2-tuples of strings.<\/p><p>You may know that functions without an explicit return still return <code>None<\/code>:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; def play(player_name):\n...     print(f\"{player_name} plays\")\n...\n&gt;&gt;&gt; ret_val = play(\"Jacob\")\nJacob plays\n&gt;&gt;&gt; print(ret_val)\nNone\n<\/pre><\/div><p>While such functions technically return something, that return value is not useful. You should add type hints saying as much by using <code>None<\/code> also as the return type:<\/p><div class=\"highlight python\"><pre> 1 # play.py\n 2 \n 3 def play(player_name: str) -&gt; None:\n 4     print(f\"{player_name} plays\")\n 5 \n 6 ret_val = play(\"Filip\")\n<\/pre><\/div><p>The annotations help catch the kinds of subtle bugs where you are trying to use a meaningless return value. Mypy will give you a helpful warning:<\/p><div class=\"highlight sh\"><pre>$ mypy play.py\nplay.py:6: error: \"play\" does not return a value\n<\/pre><\/div><p>Note that being explicit about a function not returning anything is different from not adding a type hint about the return value:<\/p><div class=\"highlight python\"><pre># play.py\ndef play(player_name: str):\n    print(f\"{player_name} plays\")\nret_val = play(\"Henrik\")\n<\/pre><\/div><p>In this latter case Mypy has no information about the return value so it will not generate any warning:<\/p><div class=\"highlight sh\"><pre>$ mypy play.py\n$\n<\/pre><\/div><p>As a more exotic case, note that you can also annotate functions that are never expected to return normally. This is done using <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/more_types.html#the-noreturn-type\"><code>NoReturn<\/code><\/a>:<\/p><div class=\"highlight python\"><pre>from typing import NoReturn\ndef black_hole() -&gt; NoReturn:\n    raise Exception(\"There is no going back ...\")\n<\/pre><\/div><p>Since <code>black_hole()<\/code> always raises an exception, it will never return properly.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"example-play-some-cards\">Example: Play Some Cards<\/h3><p>Let\u2019s return to our <a href=\"#example-a-deck-of-cards\">card game example<\/a>. In this second version of the game, we deal a hand of cards to each player as before. Then a start player is chosen and the players take turns playing their cards. There are not really any rules in the game though, so the players will just play random cards:<\/p><div class=\"highlight python\"><pre> 1 # game.py\n 2 \n 3 import random\n 4 from typing import List, Tuple\n 5 \n 6 SUITS = \"\u2660 \u2661 \u2662 \u2663\".split()\n 7 RANKS = \"2 3 4 5 6 7 8 9 10 J Q K A\".split()\n 8 \n 9 Card = Tuple[str, str]\n10 Deck = List[Card]\n11 \n12 def create_deck(shuffle: bool = False) -&gt; Deck:\n13     \"\"\"Create a new deck of 52 cards\"\"\"\n14     deck = [(s, r) for r in RANKS for s in SUITS]\n15     if shuffle:\n16         random.shuffle(deck)\n17     return deck\n18 \n19 def deal_hands(deck: Deck) -&gt; Tuple[Deck, Deck, Deck, Deck]:\n20     \"\"\"Deal the cards in the deck into four hands\"\"\"\n21     return (deck[0::4], deck[1::4], deck[2::4], deck[3::4])\n22 \n23 def choose(items):\n24     \"\"\"Choose and return a random item\"\"\"\n25     return random.choice(items)\n26 \n27 def player_order(names, start=None):\n28     \"\"\"Rotate player order so that start goes first\"\"\"\n29     if start is None:\n30         start = choose(names)\n31     start_idx = names.index(start)\n32     return names[start_idx:] + names[:start_idx]\n33 \n34 def play() -&gt; None:\n35     \"\"\"Play a 4-player card game\"\"\"\n36     deck = create_deck(shuffle=True)\n37     names = \"P1 P2 P3 P4\".split()\n38     hands = {n: h for n, h in zip(names, deal_hands(deck))}\n39     start_player = choose(names)\n40     turn_order = player_order(names, start=start_player)\n41 \n42     # Randomly play cards from each player's hand until empty\n43     while hands[start_player]:\n44         for name in turn_order:\n45             card = choose(hands[name])\n46             hands[name].remove(card)\n47             print(f\"{name}: {card[0] + card[1]:&lt;3}  \", end=\"\")\n48         print()\n49 \n50 if __name__ == \"__main__\":\n51     play()\n<\/pre><\/div><p>Note that in addition to changing <code>play()<\/code>, we have added two new functions that need type hints: <code>choose()<\/code> and <code>player_order()<\/code>. Before discussing how we\u2019ll add type hints to them, here is an example output from running the game:<\/p><div class=\"highlight sh\"><pre>$ python game.py\nP3: \u266210  P4: \u26634   P1: \u26618   P2: \u2661Q\nP3: \u26638   P4: \u26606   P1: \u26605   P2: \u2661K\nP3: \u26629   P4: \u2661J   P1: \u2663A   P2: \u2661A\nP3: \u2660Q   P4: \u26603   P1: \u26607   P2: \u2660A\nP3: \u26614   P4: \u26616   P1: \u26632   P2: \u2660K\nP3: \u2663K   P4: \u26637   P1: \u26617   P2: \u26602\nP3: \u266310  P4: \u26604   P1: \u26625   P2: \u26613\nP3: \u2663Q   P4: \u2662K   P1: \u2663J   P2: \u26619\nP3: \u26622   P4: \u26624   P1: \u26609   P2: \u266010\nP3: \u2662A   P4: \u26615   P1: \u2660J   P2: \u2662Q\nP3: \u26608   P4: \u26627   P1: \u26623   P2: \u2662J\nP3: \u26633   P4: \u266110  P1: \u26639   P2: \u26612\nP3: \u26626   P4: \u26636   P1: \u26635   P2: \u26628\n<\/pre><\/div><p>In this example, player <code>P3<\/code> was randomly chosen as the starting player. In turn, each player plays a card: first <code>P3<\/code>, then <code>P4<\/code>, then <code>P1<\/code>, and finally <code>P2<\/code>. The players keep playing cards as long as they have any left in their hand.<\/p><h3 id=\"the-any-type\">The <code>Any<\/code> Type<\/h3><p><code>choose()<\/code> works for both lists of names and lists of cards (and any other sequence for that matter). One way to add type hints for this would be the following:<\/p><div class=\"highlight python\"><pre>import random\nfrom typing import Any, Sequence\ndef choose(items: Sequence[Any]) -&gt; Any:\n    return random.choice(items)\n<\/pre><\/div><p>This means more or less what it says: <code>items<\/code> is a sequence that can contain items of any type and <code>choose()<\/code> will return one such item of any type. Unfortunately, this is not that useful. Consider the following example:<\/p><div class=\"highlight python\"><pre> 1 # choose.py\n 2 \n 3 import random\n 4 from typing import Any, Sequence\n 5 \n 6 def choose(items: Sequence[Any]) -&gt; Any:\n 7     return random.choice(items)\n 8 \n 9 names = [\"Guido\", \"Jukka\", \"Ivan\"]\n10 reveal_type(names)\n11 \n12 name = choose(names)\n13 reveal_type(name)\n<\/pre><\/div><p>While Mypy will correctly infer that <code>names<\/code> is a list of strings, that information is lost after the call to <code>choose()<\/code> because of the use of the <code>Any<\/code> type:<\/p><div class=\"highlight sh\"><pre>$ mypy choose.py\nchoose.py:10: error: Revealed type is 'builtins.list[builtins.str*]'\nchoose.py:13: error: Revealed type is 'Any'\n<\/pre><\/div><p>You\u2019ll see a better way shortly. First though, let\u2019s have a more theoretical look at the Python type system, and the special role <code>Any<\/code> plays.<\/p><h2 id=\"type-theory\">Type Theory<\/h2><p>This tutorial is mainly a practical guide and we will only scratch the surface of the theory underpinning Python type hints. For more details <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0483\/\">PEP 483<\/a> is a good starting point. If you want to get back to the practical examples, feel free to <a href=\"#playing-with-python-types-part-2\">skip to the next section<\/a>.<\/p><h3 id=\"subtypes\">Subtypes<\/h3><p>One important concept is that of <strong>subtypes<\/strong>. Formally, we say that a type <code>T<\/code> is a subtype of <code>U<\/code> if the following two conditions hold:<\/p><ul><li>Every value from <code>T<\/code> is also in the set of values of <code>U<\/code> type.<\/li>\n<li>Every function from <code>U<\/code> type is also in the set of functions of <code>T<\/code> type.<\/li>\n<\/ul><p>These two conditions guarantees that even if type <code>T<\/code> is different from <code>U<\/code>, variables of type <code>T<\/code> can always pretend to be <code>U<\/code>.<\/p><p>For a concrete example, consider <code>T = bool<\/code> and <code>U = int<\/code>. The <code>bool<\/code> type takes only two values. Usually these are denoted <code>True<\/code> and <code>False<\/code>, but these names are just aliases for the integer values <code>1<\/code> and <code>0<\/code>, respectively:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; int(False)\n0\n&gt;&gt;&gt; int(True)\n1\n&gt;&gt;&gt; True + True\n2\n&gt;&gt;&gt; issubclass(bool, int)\nTrue\n<\/pre><\/div><p>Since 0 and 1 are both integers, the first condition holds. Above you can see that booleans can be added together, but they can also do anything else integers can. This is the second condition above. In other words, <code>bool<\/code> is a subtype of <code>int<\/code>.<\/p><p>The importance of subtypes is that a subtype can always pretend to be its supertype. For instance, the following code type checks as correct:<\/p><div class=\"highlight python\"><pre>def double(number: int) -&gt; int:\n    return number * 2\nprint(double(True))  # Passing in bool instead of int\n<\/pre><\/div><p>Subtypes are somewhat related to subclasses. In fact all subclasses corresponds to subtypes, and <code>bool<\/code> is a subtype of <code>int<\/code> because <code>bool<\/code> is a subclass of <code>int<\/code>. However, there are also subtypes that do not correspond to subclasses. For instance <code>int<\/code> is a subtype of <code>float<\/code>, but <code>int<\/code> is not a subclass of <code>float<\/code>.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"covariant-contravariant-and-invariant\">Covariant, Contravariant, and Invariant<\/h3><p>What happens when you use subtypes inside composite types? For instance, is <code>Tuple[bool]<\/code> a subtype of <code>Tuple[int]<\/code>? The answer depends on the composite type, and whether that type is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Covariance_and_contravariance_%28computer_science%29\">covariant, contravariant, or invariant<\/a>. This gets technical fast, so let\u2019s just give a few examples:<\/p><ul><li>\n<p><code>Tuple<\/code> is covariant. This means that it preserves the type hierarchy of its item types: <code>Tuple[bool]<\/code> is a subtype of <code>Tuple[int]<\/code> because <code>bool<\/code> is a subtype of <code>int<\/code>.<\/p>\n<\/li>\n<li>\n<p><code>List<\/code> is invariant. Invariant types give no guarantee about subtypes. While all values of <code>List[bool]<\/code> are values of <code>List[int]<\/code>, you can append an <code>int<\/code> to <code>List[int]<\/code> and not to <code>List[bool]<\/code>. In other words, the second condition for subtypes does not hold, and <code>List[bool]<\/code> is not a subtype of <code>List[int]<\/code>.<\/p>\n<\/li>\n<li>\n<p><code>Callable<\/code> is contravariant in its arguments. This means that it reverses the type hierarchy. You will see how <code>Callable<\/code> works <a href=\"#callables\">later<\/a>, but for now think of <code>Callable[[T], ...]<\/code> as a function with its only argument being of type <code>T<\/code>. An example of a <code>Callable[[int], ...]<\/code> is the <code>double()<\/code> function defined above. Being contravariant means that if a function operating on a <code>bool<\/code> is expected, then a function operating on an <code>int<\/code> would be acceptable.<\/p>\n<\/li>\n<\/ul><p>In general, you don\u2019t need to keep these expression straight. However, you should be aware that subtypes and composite types may not be simple and intuitive.<\/p><h3 id=\"gradual-typing-and-consistent-types\">Gradual Typing and Consistent Types<\/h3><p>Earlier we mentioned that Python supports <a href=\"http:\/\/wphomes.soic.indiana.edu\/jsiek\/what-is-gradual-typing\/\">gradual typing<\/a>, where you can gradually add type hints to your Python code. Gradual typing is essentially made possible by the <code>Any<\/code> type.<\/p><p>Somehow <code>Any<\/code> sits both at the top and at the bottom of the type hierarchy of subtypes. Any type behaves as if it is a subtype of <code>Any<\/code>, and <code>Any<\/code> behaves as if it is a subtype of any other type. Looking at the definition of subtypes above this is not really possible. Instead we talk about <strong>consistent types<\/strong>.<\/p><p>The type <code>T<\/code> is consistent with the type <code>U<\/code> if <code>T<\/code> is a subtype of <code>U<\/code> or either <code>T<\/code> or <code>U<\/code> is <code>Any<\/code>.<\/p><p>The type checker only complains about inconsistent types. The takeaway is therefore that you will never see type errors arising from the <code>Any<\/code> type.<\/p><p>This means that you can use <code>Any<\/code> to explicitly fall back to dynamic typing, describe types that are too complex to describe in the Python type system, or describe items in composite types. For instance, a dictionary with string keys that can take any type as its values can be annotated <code>Dict[str, Any]<\/code>.<\/p><p>Do remember, though, if you use <code>Any<\/code> the static type checker will effectively not do any type any checking.<\/p><h2 id=\"playing-with-python-types-part-2\">Playing With Python Types, Part 2<\/h2><p>Let\u2019s return to our practical examples. Recall that you were trying to annotate the general <code>choose()<\/code> function:<\/p><div class=\"highlight python\"><pre>import random\nfrom typing import Any, Sequence\ndef choose(items: Sequence[Any]) -&gt; Any:\n    return random.choice(items)\n<\/pre><\/div><p>The problem with using <code>Any<\/code> is that you are needlessly losing type information. You know that if you pass a list of strings to <code>choose()<\/code>, it will return a string. Below you\u2019ll see how to express this using type variables, as well as how to work with:<\/p><ul><li><a href=\"#duck-types-and-protocols\">Duck types and protocols<\/a><\/li>\n<li>Arguments with <a href=\"#the-optional-type\"><code>None<\/code> as default value<\/a><\/li>\n<li><a href=\"#type-hints-for-methods\">Class methods<\/a><\/li>\n<li><a href=\"#classes-as-types\">The type of your own classes<\/a><\/li>\n<li><a href=\"#annotating-args-and-kwargs\">Variable number of arguments<\/a><\/li>\n<\/ul><h3 id=\"type-variables\">Type Variables<\/h3><p>A type variable is a special variable that can take on any type, depending on the situation.<\/p><p>Let\u2019s create a type variable that will effectively encapsulate the behavior of <code>choose()<\/code>:<\/p><div class=\"highlight python\"><pre> 1 # choose.py\n 2 \n 3 import random\n 4 from typing import Sequence, TypeVar\n 5 \n 6 Choosable = TypeVar(\"Chooseable\")\n 7 \n 8 def choose(items: Sequence[Choosable]) -&gt; Choosable:\n 9     return random.choice(items)\n10 \n11 names = [\"Guido\", \"Jukka\", \"Ivan\"]\n12 reveal_type(names)\n13 \n14 name = choose(names)\n15 reveal_type(name)\n<\/pre><\/div><p>A type variable must be defined using <code>TypeVar<\/code> from the <code>typing<\/code> module. When used, a type variable ranges over all possible types and takes the most specific type possible. In the example, <code>name<\/code> is now a <code>str<\/code>:<\/p><div class=\"highlight sh\"><pre>$ mypy choose.py\nchoose.py:12: error: Revealed type is 'builtins.list[builtins.str*]'\nchoose.py:15: error: Revealed type is 'builtins.str*'\n<\/pre><\/div><p>Consider a few other examples:<\/p><div class=\"highlight python\"><pre> 1 # choose_examples.py\n 2 \n 3 from choose import choose\n 4 \n 5 reveal_type(choose([\"Guido\", \"Jukka\", \"Ivan\"]))\n 6 reveal_type(choose([1, 2, 3]))\n 7 reveal_type(choose([True, 42, 3.14]))\n 8 reveal_type(choose([\"Python\", 3, 7])\n<\/pre><\/div><p>The first two examples should have type <code>str<\/code> and <code>int<\/code>, but what about the last two? The individual list items have different types, and in that case the <code>Choosable<\/code> type variable does its best to accommodate:<\/p><div class=\"highlight sh\"><pre>$ mypy choose_examples.py\nchoose_examples.py:5: error: Revealed type is 'builtins.str*'\nchoose_examples.py:6: error: Revealed type is 'builtins.int*'\nchoose_examples.py:7: error: Revealed type is 'builtins.float*'\nchoose_examples.py:8: error: Revealed type is 'builtins.object*'\n<\/pre><\/div><p>As you\u2019ve already seen <code>bool<\/code> is a subtype of <code>int<\/code>, which again is a subtype of <code>float<\/code>. So in the third example the return value of <code>choose()<\/code> is guaranteed to be something that can be thought of as a <code>float<\/code>. In the last example, there is no subtype relationship between <code>str<\/code> and <code>int<\/code>, so the best that can be said about the return value is that it is an object.<\/p><p>Note that none of these examples raised a type error. Is there a way to tell the type checker that <code>choose()<\/code> should accept both strings and numbers, but not both at the same time?<\/p><p>You can constrain type variables by listing the acceptable types:<\/p><div class=\"highlight python\"><pre> 1 # choose.py\n 2 \n 3 import random\n 4 from typing import Sequence, TypeVar\n 5 \n 6 Choosable = TypeVar(\"Choosable\", str, float)\n 7 \n 8 def choose(items: Sequence[Choosable]) -&gt; Choosable:\n 9     return random.choice(items)\n10 \n11 reveal_type(choose([\"Guido\", \"Jukka\", \"Ivan\"]))\n12 reveal_type(choose([1, 2, 3]))\n13 reveal_type(choose([True, 42, 3.14]))\n14 reveal_type(choose([\"Python\", 3, 7]))\n<\/pre><\/div><p>Now <code>Choosable<\/code> can only be either <code>str<\/code> or <code>float<\/code>, and Mypy will note that the last example is an error:<\/p><div class=\"highlight sh\"><pre>$ mypy choose.py\nchoose.py:11: error: Revealed type is 'builtins.str*'\nchoose.py:12: error: Revealed type is 'builtins.float*'\nchoose.py:13: error: Revealed type is 'builtins.float*'\nchoose.py:14: error: Revealed type is 'builtins.object*'\nchoose.py:14: error: Value of type variable \"Choosable\" of \"choose\"\n                     cannot be \"object\"\n<\/pre><\/div><p>Also note that in the second example the type is considered <code>float<\/code> even though the input list only contains <code>int<\/code> objects. This is because <code>Choosable<\/code> was restricted to strings and floats and <code>int<\/code> is a subtype of <code>float<\/code>.<\/p><p>In our card game we want to restrict <code>choose()<\/code> to be used for <code>str<\/code> and <code>Card<\/code>:<\/p><div class=\"highlight python\"><pre>Choosable = TypeVar(\"Choosable\", str, Card)\ndef choose(items: Sequence[Choosable]) -&gt; Choosable:\n    ...\n<\/pre><\/div><p>We briefly mentioned that <code>Sequence<\/code> represents both lists and tuples. As we noted, a <code>Sequence<\/code> can be thought of as a duck type, since it can be any object with <code>.__len__()<\/code> and <code>.__getitem__()<\/code> implemented.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"duck-types-and-protocols\">Duck Types and Protocols<\/h3><p>Recall the following example from <a href=\"#duck-typing\">the introduction<\/a>:<\/p><div class=\"highlight python\"><pre>def len(obj):\n    return obj.__len__()\n<\/pre><\/div><p><code>len()<\/code> can return the length of any object that has implemented the <code>.__len__()<\/code> method. How can we add type hints to <code>len()<\/code>, and in particular the <code>obj<\/code> argument?<\/p><p>The answer hides behind the academic sounding term <a href=\"https:\/\/en.wikipedia.org\/wiki\/Structural_type_system\">structural subtyping<\/a>. One way to categorize type systems is by whether they are <strong>nominal<\/strong> or <strong>structural<\/strong>:<\/p><ul><li>\n<p>In a <strong>nominal<\/strong> system, comparisons between types are based on names and declarations. The Python type system is mostly nominal, where an <code>int<\/code> can be used in place of a <code>float<\/code> because of their subtype relationship.<\/p>\n<\/li>\n<li>\n<p>In a <strong>structural<\/strong> system, comparisons between types are based on structure. You could define a structural type <code>Sized<\/code> that includes all instances that define <code>.__len__()<\/code>, irrespective of their nominal type.<\/p>\n<\/li>\n<\/ul><p>There is ongoing work to bring a full-fledged structural type system to Python through <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0544\/\">PEP 544<\/a> which aims at adding a concept called protocols. Most of PEP 544 is already <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/protocols.html\">implemented in Mypy<\/a> though.<\/p><p>A protocol specifies one or more methods that must be implemented. For example, all classes defining <code>.__len__()<\/code> fulfill the <code>typing.Sized<\/code> protocol. We can therefore annotate <code>len()<\/code> as follows:<\/p><div class=\"highlight python\"><pre>from typing import Sized\ndef len(obj: Sized) -&gt; int:\n    return obj.__len__()\n<\/pre><\/div><p>Other <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/protocols.html#predefined-protocols\">examples of protocols<\/a> defined in the <code>typing<\/code> module include <code>Container<\/code>, <code>Iterable<\/code>, <code>Awaitable<\/code>, and <code>ContextManager<\/code>.<\/p><p>You can also define your own protocols. This is done by inheriting from <code>Protocol<\/code> and defining the function signatures (with empty function bodies) that the protocol expects. The following example shows how <code>len()<\/code> and <code>Sized<\/code> could have been implemented:<\/p><div class=\"highlight python\"><pre>from typing_extensions import Protocol\nclass Sized(Protocol):\n    def __len__(self) -&gt; int: ...\ndef len(obj: Sized) -&gt; int:\n    return obj.__len__()\n<\/pre><\/div><p>At the time of writing the support for self-defined protocols is still experimental and only available through the <code>typing_extensions<\/code> module. This module must be explicitly installed from <a href=\"https:\/\/pypi.org\/project\/typing-extensions\/\">PyPI<\/a> by doing <code>pip install typing-extensions<\/code>.<\/p><h3 id=\"the-optional-type\">The <code>Optional<\/code> Type<\/h3><p>A common pattern in Python is to use <code>None<\/code> as a default value for an argument. This is usually done either to avoid problems with <a href=\"https:\/\/docs.quantifiedcode.com\/python-anti-patterns\/correctness\/mutable_default_value_as_argument.html\">mutable default values<\/a> or to have a sentinel value flagging special behavior.<\/p><p>In the card example, the <code>player_order()<\/code> function uses <code>None<\/code> as a sentinel value for <code>start<\/code> saying that if no start player is given it should be chosen randomly:<\/p><div class=\"highlight python\"><pre>27 def player_order(names, start=None):\n28     \"\"\"Rotate player order so that start goes first\"\"\"\n29     if start is None:\n30         start = choose(names)\n31     start_idx = names.index(start)\n32     return names[start_idx:] + names[:start_idx]\n<\/pre><\/div><p>The challenge this creates for type hinting is that in general <code>start<\/code> should be a string. However, it may also take the special non-string value <code>None<\/code>.<\/p><p>In order to annotate such arguments you can use the <code>Optional<\/code> type:<\/p><div class=\"highlight python\"><pre>from typing import Sequence, Optional\ndef player_order(\n    names: Sequence[str], start: Optional[str] = None\n) -&gt; Sequence[str]:\n    ...\n<\/pre><\/div><p>The <code>Optional<\/code> type simply says that a variable either has the type specified or is <code>None<\/code>. An equivalent way of specifying the same would be using the <code>Union<\/code> type: <code>Union[None, str]<\/code><\/p><p>Note that when using either <code>Optional<\/code> or <code>Union<\/code> you must take care that the variable has the correct type when you operate on it. This is done in the example by testing whether <code>start is None<\/code>. Not doing so would cause both static type errors as well as possible runtime errors:<\/p><div class=\"highlight python\"><pre> 1 # player_order.py\n 2 \n 3 from typing import Sequence, Optional\n 4 \n 5 def player_order(\n 6     names: Sequence[str], start: Optional[str] = None\n 7 ) -&gt; Sequence[str]:\n 8     start_idx = names.index(start)\n 9     return names[start_idx:] + names[:start_idx]\n<\/pre><\/div><p>Mypy tells you that you have not taken care of the case where <code>start<\/code> is <code>None<\/code>:<\/p><div class=\"highlight sh\"><pre>$ mypy player_order.py\nplayer_order.py:8: error: Argument 1 to \"index\" of \"list\" has incompatible\n                          type \"Optional[str]\"; expected \"str\"\n<\/pre><\/div><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note:<\/strong> The use of <code>None<\/code> for optional arguments is so common that Mypy handles it automatically. Mypy assumes that a default argument of <code>None<\/code> indicates an optional argument even if the type hint does not explicitly say so. You could have used the following:<\/p><div class=\"highlight python\"><pre>def player_order(names: Sequence[str], start: str = None) -&gt; Sequence[str]:\n    ...\n<\/pre><\/div><p>If you don\u2019t want Mypy to make this assumption you can turn it off with the <code>--no-implicit-optional<\/code> command line option.<\/p><\/div><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"example-the-objective-of-the-game\">Example: The Object(ive) of the Game<\/h3><p>Let\u2019s rewrite the card game to be more <a href=\"https:\/\/realpython.com\/python3-object-oriented-programming\/\">object-oriented<\/a>. This will allow us to discuss how to properly annotate classes and methods.<\/p><p>A more or less direct translation of our card game into code that uses classes for <code>Card<\/code>, <code>Deck<\/code>, <code>Player<\/code>, and <code>Game<\/code> looks something like the following:<\/p><div class=\"highlight python\"><pre> 1 # game.py\n 2 \n 3 import random\n 4 import sys\n 5 \n 6 class Card:\n 7     SUITS = \"\u2660 \u2661 \u2662 \u2663\".split()\n 8     RANKS = \"2 3 4 5 6 7 8 9 10 J Q K A\".split()\n 9 \n10     def __init__(self, suit, rank):\n11         self.suit = suit\n12         self.rank = rank\n13 \n14     def __repr__(self):\n15         return f\"{self.suit}{self.rank}\"\n16 \n17 class Deck:\n18     def __init__(self, cards):\n19         self.cards = cards\n20 \n21     @classmethod\n22     def create(cls, shuffle=False):\n23         \"\"\"Create a new deck of 52 cards\"\"\"\n24         cards = [Card(s, r) for r in Card.RANKS for s in Card.SUITS]\n25         if shuffle:\n26             random.shuffle(cards)\n27         return cls(cards)\n28 \n29     def deal(self, num_hands):\n30         \"\"\"Deal the cards in the deck into a number of hands\"\"\"\n31         cls = self.__class__\n32         return tuple(cls(self.cards[i::num_hands]) for i in range(num_hands))\n33 \n34 class Player:\n35     def __init__(self, name, hand):\n36         self.name = name\n37         self.hand = hand\n38 \n39     def play_card(self):\n40         \"\"\"Play a card from the player's hand\"\"\"\n41         card = random.choice(self.hand.cards)\n42         self.hand.cards.remove(card)\n43         print(f\"{self.name}: {card!r:&lt;3}  \", end=\"\")\n44         return card\n45 \n46 class Game:\n47     def __init__(self, *names):\n48         \"\"\"Set up the deck and deal cards to 4 players\"\"\"\n49         deck = Deck.create(shuffle=True)\n50         self.names = (list(names) + \"P1 P2 P3 P4\".split())[:4]\n51         self.hands = {\n52             n: Player(n, h) for n, h in zip(self.names, deck.deal(4))\n53         }\n54 \n55     def play(self):\n56         \"\"\"Play a card game\"\"\"\n57         start_player = random.choice(self.names)\n58         turn_order = self.player_order(start=start_player)\n59 \n60         # Play cards from each player's hand until empty\n61         while self.hands[start_player].hand.cards:\n62             for name in turn_order:\n63                 self.hands[name].play_card()\n64             print()\n65 \n66     def player_order(self, start=None):\n67         \"\"\"Rotate player order so that start goes first\"\"\"\n68         if start is None:\n69             start = random.choice(self.names)\n70         start_idx = self.names.index(start)\n71         return self.names[start_idx:] + self.names[:start_idx]\n72 \n73 if __name__ == \"__main__\":\n74     # Read player names from command line\n75     player_names = sys.argv[1:]\n76     game = Game(*player_names)\n77     game.play()\n<\/pre><\/div><p>Now let\u2019s add types to this code.<\/p><h3 id=\"type-hints-for-methods\">Type Hints for Methods<\/h3><p>First of all type hints for methods work much the same as type hints for functions. The only difference is that the <code>self<\/code> argument need not be annotated, as it always will be a class instance. The types of the <code>Card<\/code> class are easy to add:<\/p><div class=\"highlight python\"><pre> 6 class Card:\n 7     SUITS = \"\u2660 \u2661 \u2662 \u2663\".split()\n 8     RANKS = \"2 3 4 5 6 7 8 9 10 J Q K A\".split()\n 9 \n10     def __init__(self, suit: str, rank: str) -&gt; None:\n11         self.suit = suit\n12         self.rank = rank\n13 \n14     def __repr__(self) -&gt; str:\n15         return f\"{self.suit}{self.rank}\"\n<\/pre><\/div><p>Note that the <code>.__init__()<\/code> method always should have <code>None<\/code> as its return type.<\/p><h3 id=\"classes-as-types\">Classes as Types<\/h3><p>There is a correspondence between classes and types. For example, all instances of the <code>Card<\/code> class together form the <code>Card<\/code> type. To use classes as types you simply use the name of the class.<\/p><p>For example, a <code>Deck<\/code> essentially consists of a list of <code>Card<\/code> objects. You can annotate this as follows:<\/p><div class=\"highlight python\"><pre>17 class Deck:\n18     def __init__(self, cards: List[Card]) -&gt; None:\n19         self.cards = cards\n<\/pre><\/div><p>Mypy is able to connect your use of <code>Card<\/code> in the annotation with the definition of the <code>Card<\/code> class.<\/p><p>This doesn\u2019t work as cleanly though when you need to refer to the class currently being defined. For example, the <code>Deck.create()<\/code> class method returns an object with type <code>Deck<\/code>. However, you can\u2019t simply add <code>-&gt; Deck<\/code> as the <code>Deck<\/code> class is not yet fully defined.<\/p><p>Instead, you are allowed to use string literals in annotations. These strings will only be evaluated by the type checker later, and can therefore contain self and forward references. The <code>.create()<\/code> method should use such string literals for its types:<\/p><div class=\"highlight python\"><pre>20 class Deck:\n21     @classmethod\n22     def create(cls, shuffle: bool = False) -&gt; \"Deck\":\n23         \"\"\"Create a new deck of 52 cards\"\"\"\n24         cards = [Card(s, r) for r in Card.RANKS for s in Card.SUITS]\n25         if shuffle:\n26             random.shuffle(cards)\n27         return cls(cards)\n<\/pre><\/div><p>Note that the <code>Player<\/code> class also will reference the <code>Deck<\/code> class. This is however no problem, since <code>Deck<\/code> is defined before <code>Player<\/code>:<\/p><div class=\"highlight python\"><pre>34 class Player:\n35     def __init__(self, name: str, hand: Deck) -&gt; None:\n36         self.name = name\n37         self.hand = hand\n<\/pre><\/div><p>Usually annotations are not used at runtime. This has given wings to the idea of <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0563\/\">postponing the evaluation of annotations<\/a>. Instead of evaluating annotations as Python expressions and storing their value, the proposal is to store the string representation of the annotation and only evaluate it when needed.<\/p><p>Such functionality is planned to become standard in the still mythical <a href=\"http:\/\/www.curiousefficiency.org\/posts\/2014\/08\/python-4000.html\">Python 4.0<\/a>. However, in <a href=\"https:\/\/realpython.com\/python37-new-features\/\">Python 3.7<\/a> and later, forward references are available through a <code>__future__<\/code> import:<\/p><div class=\"highlight python\"><pre>from __future__ import annotations\nclass Deck:\n    @classmethod\n    def create(cls, shuffle: bool = False) -&gt; Deck:\n        ...\n<\/pre><\/div><p>With the <code>__future__<\/code> import you can use <code>Deck<\/code> instead of <code>\"Deck\"<\/code> even before <code>Deck<\/code> is defined.<\/p><h3 id=\"returning-self-or-cls\">Returning <code>self<\/code> or <code>cls<\/code><\/h3><p>As noted, you should typically not annotate the <code>self<\/code> or <code>cls<\/code> arguments. Partly, this is not necessary as <code>self<\/code> points to an instance of the class, so it will have the type of the class. In the <code>Card<\/code> example, <code>self<\/code> has the implicit type <code>Card<\/code>. Also, adding this type explicitly would be cumbersome since the class is not defined yet. You would have to use the string literal syntax, <code>self: \"Card\"<\/code>.<\/p><p>There is one case where you might want to annotate <code>self<\/code> or <code>cls<\/code>, though. Consider what happens if you have a superclass that other classes inherit from, and which has methods that return <code>self<\/code> or <code>cls<\/code>:<\/p><div class=\"highlight python\"><pre> 1 # dogs.py\n 2 \n 3 from datetime import date\n 4 \n 5 class Animal:\n 6     def __init__(self, name: str, birthday: date) -&gt; None:\n 7         self.name = name\n 8         self.birthday = birthday\n 9 \n10     @classmethod\n11     def newborn(cls, name: str) -&gt; \"Animal\":\n12         return cls(name, date.today())\n13 \n14     def twin(self, name: str) -&gt; \"Animal\":\n15         cls = self.__class__\n16         return cls(name, self.birthday)\n17 \n18 class Dog(Animal):\n19     def bark(self) -&gt; None:\n20         print(f\"{self.name} says woof!\")\n21 \n22 fido = Dog.newborn(\"Fido\")\n23 pluto = fido.twin(\"Pluto\")\n24 fido.bark()\n25 pluto.bark()\n<\/pre><\/div><p>While the code runs without problems, Mypy will flag a problem:<\/p><div class=\"highlight sh\"><pre>$ mypy dogs.py\ndogs.py:24: error: \"Animal\" has no attribute \"bark\"\ndogs.py:25: error: \"Animal\" has no attribute \"bark\"\n<\/pre><\/div><p>The issue is that even though the inherited <code>Dog.newborn()<\/code> and <code>Dog.twin()<\/code> methods will return a <code>Dog<\/code> the annotation says that they return an <code>Animal<\/code>.<\/p><p>In cases like this you want to be more careful to make sure the annotation is correct. The return type should match the type of <code>self<\/code> or the instance type of <code>cls<\/code>. This can be done using type variables that keep track of what is actually passed to <code>self<\/code> and <code>cls<\/code>:<\/p><div class=\"highlight python\"><pre># dogs.py\nfrom datetime import date\nfrom typing import Type, TypeVar\nTAnimal = TypeVar(\"TAnimal\", bound=\"Animal\")\nclass Animal:\n    def __init__(self, name: str, birthday: date) -&gt; None:\n        self.name = name\n        self.birthday = birthday\n    @classmethod\n    def newborn(cls: Type[TAnimal], name: str) -&gt; TAnimal:\n        return cls(name, date.today())\n    def twin(self: TAnimal, name: str) -&gt; TAnimal:\n        cls = self.__class__\n        return cls(name, self.birthday)\nclass Dog(Animal):\n    def bark(self) -&gt; None:\n        print(f\"{self.name} says woof!\")\nfido = Dog.newborn(\"Fido\")\npluto = fido.twin(\"Pluto\")\nfido.bark()\npluto.bark()\n<\/pre><\/div><p>There are a few things to note in this example:<\/p><ul><li>\n<p>The type variable <code>TAnimal<\/code> is used to denote that return values might be instances of subclasses of <code>Animal<\/code>.<\/p>\n<\/li>\n<li>\n<p>We specify that <code>Animal<\/code> is an upper bound for <code>TAnimal<\/code>. Specifying <code>bound<\/code> means that <code>TAnimal<\/code> will only be <code>Animal<\/code> or one of its subclasses. This is needed to properly restrict the types that are allowed.<\/p>\n<\/li>\n<li>\n<p>The <code>typing.Type[]<\/code> construct is the typing equivalent of <code>type()<\/code>. You need it to note that the class method expects a class and returns an instance of that class.<\/p>\n<\/li>\n<\/ul><h3 id=\"annotating-args-and-kwargs\">Annotating <code>*args<\/code> and <code>**kwargs<\/code><\/h3><p>In the <a href=\"#example-the-objective-of-the-game\">object oriented version<\/a> of the game, we added the option to name the players on the command line. This is done by listing player names after the name of the program:<\/p><div class=\"highlight sh\"><pre>$ python game.py GeirArne Dan Joanna\nDan: \u2662A   Joanna: \u26619   P1: \u2663A   GeirArne: \u26632\nDan: \u2661A   Joanna: \u26616   P1: \u26604   GeirArne: \u26628\nDan: \u2662K   Joanna: \u2662Q   P1: \u2663K   GeirArne: \u26605\nDan: \u26612   Joanna: \u2661J   P1: \u26607   GeirArne: \u2661K\nDan: \u266210  Joanna: \u26633   P1: \u26624   GeirArne: \u26608\nDan: \u26636   Joanna: \u2661Q   P1: \u2663Q   GeirArne: \u2662J\nDan: \u26622   Joanna: \u26614   P1: \u26638   GeirArne: \u26617\nDan: \u266110  Joanna: \u26623   P1: \u26613   GeirArne: \u26602\nDan: \u2660K   Joanna: \u26635   P1: \u26637   GeirArne: \u2660J\nDan: \u26606   Joanna: \u26629   P1: \u2663J   GeirArne: \u266310\nDan: \u26603   Joanna: \u26615   P1: \u26639   GeirArne: \u2660Q\nDan: \u2660A   Joanna: \u26609   P1: \u266010  GeirArne: \u26618\nDan: \u26626   Joanna: \u26625   P1: \u26627   GeirArne: \u26634\n<\/pre><\/div><p>This is implemented by unpacking and passing in <code>sys.argv<\/code> to <code>Game()<\/code> when it\u2019s instantiated. The <code>.__init__()<\/code> method uses <code>*names<\/code> to pack the given names into a tuple.<\/p><p>Regarding type annotations: even though <code>names<\/code> will be a tuple of strings, you should only annotate the type of each name. In other words, you should use <code>str<\/code> and not <code>Tuple[str]<\/code>:<\/p><div class=\"highlight python\"><pre>46 class Game:\n47     def __init__(self, *names: str) -&gt; None:\n48         \"\"\"Set up the deck and deal cards to 4 players\"\"\"\n49         deck = Deck.create(shuffle=True)\n50         self.names = (list(names) + \"P1 P2 P3 P4\".split())[:4]\n51         self.hands = {\n52             n: Player(n, h) for n, h in zip(self.names, deck.deal(4))\n53         }\n<\/pre><\/div><p>Similarly, if you have a function or method accepting <code>**kwargs<\/code> you should only annotate the type of each possible keyword argument.<\/p><h3 id=\"callables\">Callables<\/h3><p>Functions are <a href=\"https:\/\/dbader.org\/blog\/python-first-class-functions\">first-class objects<\/a> in Python. This means that you can use functions as arguments to other functions. That also means that you need to be able to add type hints representing functions.<\/p><p>Functions, as well as lambdas, methods and classes, are <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/kinds_of_types.html#callable-types-and-lambdas\">represented by <code>typing.Callable<\/code><\/a>. The types of the arguments and the return value are usually also represented. For instance, <code>Callable[[A1, A2, A3], Rt]<\/code> represents a function with three arguments with types <code>A1<\/code>, <code>A2<\/code>, and <code>A3<\/code>, respectively. The return type of the function is <code>Rt<\/code>.<\/p><p>In the following example, the function <code>do_twice()<\/code> calls a given function twice and prints the return values:<\/p><div class=\"highlight python\"><pre> 1 # do_twice.py\n 2 \n 3 from typing import Callable\n 4 \n 5 def do_twice(func: Callable[[str], str], argument: str) -&gt; None:\n 6     print(func(argument))\n 7     print(func(argument))\n 8 \n 9 def create_greeting(name: str) -&gt; str:\n10     return f\"Hello {name}\"\n11 \n12 do_twice(create_greeting, \"Jekyll\")\n<\/pre><\/div><p>Note the annotation of the <code>func<\/code> argument to <code>do_twice()<\/code> on line 5. It says that <code>func<\/code> should be a callable with one string argument, that also returns a string. One example of such a callable is <code>create_greeting()<\/code> defined on line 9.<\/p><p>Most callable types can be annotated in a similar manner. However, if you need more flexibility, check out <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/protocols.html#callback-protocols\">callback protocols<\/a> and <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/additional_features.html#extended-callable-types\">extended callable types<\/a>.<\/p><h3 id=\"example-hearts\">Example: Hearts<\/h3><p>Let\u2019s end with a full example of the game of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hearts\">Hearts<\/a>. You might already know this game from other computer simulations. Here is a quick recap of the rules:<\/p><ul><li>\n<p>Four players play with a hand of 13 cards each.<\/p>\n<\/li>\n<li>\n<p>The player holding the \u26632 starts the first round, and must play \u26632.<\/p>\n<\/li>\n<li>\n<p>Players take turns playing cards, following the leading suit if possible.<\/p>\n<\/li>\n<li>\n<p>The player playing the highest card in the leading suit wins the trick, and becomes start player in the next turn.<\/p>\n<\/li>\n<li>\n<p>A player can not lead with a \u2661 until a \u2661 has already been played in an earlier trick.<\/p>\n<\/li>\n<li>\n<p>After all cards are played, players get points if they take certain cards:<\/p>\n<ul><li>13 points for the \u2660Q<\/li>\n<li>1 point for each \u2661<\/li>\n<\/ul><\/li>\n<li>\n<p>A game lasts several rounds, until one player has 100 points or more. The player with the least points wins.<\/p>\n<\/li>\n<\/ul><p>More details can be found <a href=\"https:\/\/www.bicyclecards.com\/how-to-play\/hearts\">found online<\/a>.<\/p><p>There are not many new typing concepts in this example that you have not already seen. We\u2019ll therefore not go through this code in detail, but leave it as an example of annotated code.<\/p><div class=\"card mb-3\" id=\"collapse_card311f38\"><div id=\"collapse311f38\" class=\"collapse card-body\" data-parent=\"#collapse_card311f38\"><p>You can download this code and other examples from <a href=\"https:\/\/github.com\/realpython\/materials\/tree\/master\/python-type-checking\">GitHub<\/a>:<\/p><div class=\"highlight python\"><pre># hearts.py\nfrom collections import Counter\nimport random\nimport sys\nfrom typing import Any, Dict, List, Optional, Sequence, Tuple, Union\nfrom typing import overload\nclass Card:\n    SUITS = \"\u2660 \u2661 \u2662 \u2663\".split()\n    RANKS = \"2 3 4 5 6 7 8 9 10 J Q K A\".split()\n    def __init__(self, suit: str, rank: str) -&gt; None:\n        self.suit = suit\n        self.rank = rank\n    @property\n    def value(self) -&gt; int:\n        \"\"\"The value of a card is rank as a number\"\"\"\n        return self.RANKS.index(self.rank)\n    @property\n    def points(self) -&gt; int:\n        \"\"\"Points this card is worth\"\"\"\n        if self.suit == \"\u2660\" and self.rank == \"Q\":\n            return 13\n        if self.suit == \"\u2661\":\n            return 1\n        return 0\n    def __eq__(self, other: Any) -&gt; Any:\n        return self.suit == other.suit and self.rank == other.rank\n    def __lt__(self, other: Any) -&gt; Any:\n        return self.value &lt; other.value\n    def __repr__(self) -&gt; str:\n        return f\"{self.suit}{self.rank}\"\nclass Deck(Sequence[Card]):\n    def __init__(self, cards: List[Card]) -&gt; None:\n        self.cards = cards\n    @classmethod\n    def create(cls, shuffle: bool = False) -&gt; \"Deck\":\n        \"\"\"Create a new deck of 52 cards\"\"\"\n        cards = [Card(s, r) for r in Card.RANKS for s in Card.SUITS]\n        if shuffle:\n            random.shuffle(cards)\n        return cls(cards)\n    def play(self, card: Card) -&gt; None:\n        \"\"\"Play one card by removing it from the deck\"\"\"\n        self.cards.remove(card)\n    def deal(self, num_hands: int) -&gt; Tuple[\"Deck\", ...]:\n        \"\"\"Deal the cards in the deck into a number of hands\"\"\"\n        return tuple(self[i::num_hands] for i in range(num_hands))\n    def add_cards(self, cards: List[Card]) -&gt; None:\n        \"\"\"Add a list of cards to the deck\"\"\"\n        self.cards += cards\n    def __len__(self) -&gt; int:\n        return len(self.cards)\n    @overload\n    def __getitem__(self, key: int) -&gt; Card: ...\n    @overload\n    def __getitem__(self, key: slice) -&gt; \"Deck\": ...\n    def __getitem__(self, key: Union[int, slice]) -&gt; Union[Card, \"Deck\"]:\n        if isinstance(key, int):\n            return self.cards[key]\n        elif isinstance(key, slice):\n            cls = self.__class__\n            return cls(self.cards[key])\n        else:\n            raise TypeError(\"Indices must be integers or slices\")\n    def __repr__(self) -&gt; str:\n        return \" \".join(repr(c) for c in self.cards)\nclass Player:\n    def __init__(self, name: str, hand: Optional[Deck] = None) -&gt; None:\n        self.name = name\n        self.hand = Deck([]) if hand is None else hand\n    def playable_cards(self, played: List[Card], hearts_broken: bool) -&gt; Deck:\n        \"\"\"List which cards in hand are playable this round\"\"\"\n        if Card(\"\u2663\", \"2\") in self.hand:\n            return Deck([Card(\"\u2663\", \"2\")])\n        lead = played[0].suit if played else None\n        playable = Deck([c for c in self.hand if c.suit == lead]) or self.hand\n        if lead is None and not hearts_broken:\n            playable = Deck([c for c in playable if c.suit != \"\u2661\"])\n        return playable or Deck(self.hand.cards)\n    def non_winning_cards(self, played: List[Card], playable: Deck) -&gt; Deck:\n        \"\"\"List playable cards that are guaranteed to not win the trick\"\"\"\n        if not played:\n            return Deck([])\n        lead = played[0].suit\n        best_card = max(c for c in played if c.suit == lead)\n        return Deck([c for c in playable if c &lt; best_card or c.suit != lead])\n    def play_card(self, played: List[Card], hearts_broken: bool) -&gt; Card:\n        \"\"\"Play a card from a cpu player's hand\"\"\"\n        playable = self.playable_cards(played, hearts_broken)\n        non_winning = self.non_winning_cards(played, playable)\n        # Strategy\n        if non_winning:\n            # Highest card not winning the trick, prefer points\n            card = max(non_winning, key=lambda c: (c.points, c.value))\n        elif len(played) &lt; 3:\n            # Lowest card maybe winning, avoid points\n            card = min(playable, key=lambda c: (c.points, c.value))\n        else:\n            # Highest card guaranteed winning, avoid points\n            card = max(playable, key=lambda c: (-c.points, c.value))\n        self.hand.cards.remove(card)\n        print(f\"{self.name} -&gt; {card}\")\n        return card\n    def has_card(self, card: Card) -&gt; bool:\n        return card in self.hand\n    def __repr__(self) -&gt; str:\n        return f\"{self.__class__.__name__}({self.name!r}, {self.hand})\"\nclass HumanPlayer(Player):\n    def play_card(self, played: List[Card], hearts_broken: bool) -&gt; Card:\n        \"\"\"Play a card from a human player's hand\"\"\"\n        playable = sorted(self.playable_cards(played, hearts_broken))\n        p_str = \"  \".join(f\"{n}: {c}\" for n, c in enumerate(playable))\n        np_str = \" \".join(repr(c) for c in self.hand if c not in playable)\n        print(f\"  {p_str}  (Rest: {np_str})\")\n        while True:\n            try:\n                card_num = int(input(f\"  {self.name}, choose card: \"))\n                card = playable[card_num]\n            except (ValueError, IndexError):\n                pass\n            else:\n                break\n        self.hand.play(card)\n        print(f\"{self.name} =&gt; {card}\")\n        return card\nclass HeartsGame:\n    def __init__(self, *names: str) -&gt; None:\n        self.names = (list(names) + \"P1 P2 P3 P4\".split())[:4]\n        self.players = [Player(n) for n in self.names[1:]]\n        self.players.append(HumanPlayer(self.names[0]))\n    def play(self) -&gt; None:\n        \"\"\"Play a game of Hearts until one player go bust\"\"\"\n        score = Counter({n: 0 for n in self.names})\n        while all(s &lt; 100 for s in score.values()):\n            print(\"\\nStarting new round:\")\n            round_score = self.play_round()\n            score.update(Counter(round_score))\n            print(\"Scores:\")\n            for name, total_score in score.most_common(4):\n                print(f\"{name:&lt;15}{round_score[name]:&gt;3}{total_score:&gt;3}\")\n        winners = [n for n in self.names if score[n] == min(score.values())]\n        print(f\"\\n{' and '.join(winners)} won the game\")\n    def play_round(self) -&gt; Dict[str, int]:\n        \"\"\"Play a round of the Hearts card game\"\"\"\n        deck = Deck.create(shuffle=True)\n        for player, hand in zip(self.players, deck.deal(4)):\n            player.hand.add_cards(hand.cards)\n        start_player = next(\n            p for p in self.players if p.has_card(Card(\"\u2663\", \"2\"))\n        )\n        tricks = {p.name: Deck([]) for p in self.players}\n        hearts = False\n        # Play cards from each player's hand until empty\n        while start_player.hand:\n            played: List[Card] = []\n            turn_order = self.player_order(start=start_player)\n            for player in turn_order:\n                card = player.play_card(played, hearts_broken=hearts)\n                played.append(card)\n            start_player = self.trick_winner(played, turn_order)\n            tricks[start_player.name].add_cards(played)\n            print(f\"{start_player.name} wins the trick\\n\")\n            hearts = hearts or any(c.suit == \"\u2661\" for c in played)\n        return self.count_points(tricks)\n    def player_order(self, start: Optional[Player] = None) -&gt; List[Player]:\n        \"\"\"Rotate player order so that start goes first\"\"\"\n        if start is None:\n            start = random.choice(self.players)\n        start_idx = self.players.index(start)\n        return self.players[start_idx:] + self.players[:start_idx]\n    @staticmethod\n    def trick_winner(trick: List[Card], players: List[Player]) -&gt; Player:\n        lead = trick[0].suit\n        valid = [\n            (c.value, p) for c, p in zip(trick, players) if c.suit == lead\n        ]\n        return max(valid)[1]\n    @staticmethod\n    def count_points(tricks: Dict[str, Deck]) -&gt; Dict[str, int]:\n        return {n: sum(c.points for c in cards) for n, cards in tricks.items()}\nif __name__ == \"__main__\":\n    # Read player names from the command line\n    player_names = sys.argv[1:]\n    game = HeartsGame(*player_names)\n    game.play()\n<\/pre><\/div><\/div><\/div><p>Here are a few points to note in the code:<\/p><ul><li>\n<p>For type relationships that are hard to express using <code>Union<\/code> or type variables, you can use the <code>@overload<\/code> decorator. See <code>Deck.__getitem__()<\/code> for an example and <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/more_types.html#function-overloading\">the documentation<\/a> for more information.<\/p>\n<\/li>\n<li>\n<p>Subclasses correspond to subtypes, so that a <code>HumanPlayer<\/code> can be used wherever a <code>Player<\/code> is expected.<\/p>\n<\/li>\n<li>\n<p>When a subclass reimplements a method from a superclass, the type annotations must match. See <code>HumanPlayer.play_card()<\/code> for an example.<\/p>\n<\/li>\n<\/ul><p>When starting the game, you control the first player. Enter numbers to choose which cards to play. The following is an example of game play, with the highlighted lines showing where the player made a choice:<\/p><div class=\"highlight sh\"><pre>$ python hearts.py GeirArne Aldren Joanna Brad\nStarting new round:\nBrad -&gt; \u26632\n  0: \u26635  1: \u2663Q  2: \u2663K  (Rest: \u26626 \u266110 \u26616 \u2660J \u26613 \u26619 \u266210 \u26607 \u2660K \u26604)\n  GeirArne, choose card: 2\nGeirArne =&gt; \u2663K\nAldren -&gt; \u266310\nJoanna -&gt; \u26639\nGeirArne wins the trick\n  0: \u26604  1: \u26635  2: \u26626  3: \u26607  4: \u266210  5: \u2660J  6: \u2663Q  7: \u2660K  (Rest: \u266110 \u26616 \u26613 \u26619)\n  GeirArne, choose card: 0\nGeirArne =&gt; \u26604\nAldren -&gt; \u26605\nJoanna -&gt; \u26603\nBrad -&gt; \u26602\nAldren wins the trick\n...\nJoanna -&gt; \u2661J\nBrad -&gt; \u26612\n  0: \u26616  1: \u26619  (Rest: )\n  GeirArne, choose card: 1\nGeirArne =&gt; \u26619\nAldren -&gt; \u2661A\nAldren wins the trick\nAldren -&gt; \u2663A\nJoanna -&gt; \u2661Q\nBrad -&gt; \u2663J\n  0: \u26616  (Rest: )\n  GeirArne, choose card: 0\nGeirArne =&gt; \u26616\nAldren wins the trick\nScores:\nBrad             14  14\nAldren           10  10\nGeirArne          1   1\nJoanna            1   1\n<\/pre><\/div><p>So far you have seen how to add type hints to your code. In this section you\u2019ll learn more about how to actually perform static type checking of Python code.<\/p><h3 id=\"the-mypy-project\">The Mypy Project<\/h3><p>Mypy was started by Jukka Lehtosalo during his Ph.D. studies at Cambridge around 2012. Mypy was originally envisioned as a Python variant with seamless dynamic and static typing. See <a href=\"https:\/\/www.slideshare.net\/jukkaleh\/mypy-pyconfi2012\">Jukka\u2019s slides from PyCon Finland 2012<\/a> for examples of the original vision of Mypy.<\/p><p>Most of those original ideas still play a big part in the Mypy project. In fact, the slogan \u201cSeamless dynamic and static typing\u201d is still <a href=\"http:\/\/mypy-lang.org\">prominently visible on Mypy\u2019s home page<\/a> and describes the motivation for using type hints in Python well.<\/p><p>The biggest change since 2012 is that Mypy is no longer a <em>variant<\/em> of Python. In its first versions Mypy was a stand-alone language that was compatible with Python except for its type declarations. Following a <a href=\"https:\/\/twitter.com\/gvanrossum\/status\/700741601966985216\">suggestion by Guido van Rossum<\/a>, Mypy was rewritten to use annotations instead. Today Mypy is a static type checker for <em>regular<\/em> Python code.<\/p><h3 id=\"running-mypy\">Running Mypy<\/h3><p>Before running Mypy for the first time, you must install the program. This is most easily done using <code>pip<\/code>:<\/p><div class=\"highlight sh\"><pre>$ pip install mypy\n<\/pre><\/div><p>With Mypy installed, you can run it as a regular command line program:<\/p><div class=\"highlight sh\"><pre>$ mypy my_program.py\n<\/pre><\/div><p>Running Mypy on your <code>my_program.py<\/code> Python file will check it for type errors without actually executing the code.<\/p><p>There are many available options when type checking your code. As Mypy is still under very active development, command line options are liable to change between versions. You should refer to Mypy\u2019s help to see which settings are default on your version:<\/p><div class=\"highlight sh\"><pre>$ mypy --help\nusage: mypy [-h] [-v] [-V] [more options; see below]\n            [-m MODULE] [-p PACKAGE] [-c PROGRAM_TEXT] [files ...]\nMypy is a program that will type check your Python code.\n[... The rest of the help hidden for brevity ...]\n<\/pre><\/div><p>Additionally, the <a href=\"https:\/\/mypy.readthedocs.io\/en\/stable\/command_line.html#command-line\">Mypy command line documentation online<\/a> has a lot of information.<\/p><p>Let\u2019s look at some of the most common options. First of all, if you are using third-party packages without type hints, you may want to silence Mypy\u2019s warnings about these. This can be done with the <code>--ignore-missing-imports<\/code> option.<\/p><p>The following example uses <a href=\"https:\/\/realpython.com\/numpy-array-programming\/\">Numpy<\/a> to calculate and print the cosine of several numbers:<\/p><div class=\"highlight python\"><pre> 1 # cosine.py\n 2 \n 3 import numpy as np\n 4 \n 5 def print_cosine(x: np.ndarray) -&gt; None:\n 6     with np.printoptions(precision=3, suppress=True):\n 7         print(np.cos(x))\n 8 \n 9 x = np.linspace(0, 2 * np.pi, 9)\n10 print_cosine(x)\n<\/pre><\/div><p>Note that <a href=\"https:\/\/docs.scipy.org\/doc\/numpy\/reference\/generated\/numpy.printoptions.html\"><code>np.printoptions()<\/code><\/a> is only available in version 1.15 and later of Numpy. Running this example prints some numbers to the console:<\/p><div class=\"highlight sh\"><pre>$ python cosine.py\n[ 1.     0.707  0.    -0.707 -1.    -0.707 -0.     0.707  1.   ]\n<\/pre><\/div><p>The actual output of this example is not important. However, you should note that the argument <code>x<\/code> is annotated with <code>np.ndarray<\/code> on line 5, as we want to print the cosine of a full array of numbers.<\/p><p>You can run Mypy on this file as usual:<\/p><div class=\"highlight sh\"><pre>$ mypy cosine.py \ncosine.py:3: error: No library stub file for module 'numpy'\ncosine.py:3: note: (Stub files are from https:\/\/github.com\/python\/typeshed)\n<\/pre><\/div><p>These warnings may not immediately make much sense to you, but you\u2019ll learn about <a href=\"#adding-stubs\">stubs<\/a> and <a href=\"#typeshed\">typeshed<\/a> soon. You can essentially read the warnings as Mypy saying that the Numpy package does not contain type hints.<\/p><p>In most cases, missing type hints in third-party packages is not something you want to be bothered with so you can silence these messages:<\/p><div class=\"highlight sh\"><pre>$ mypy --ignore-missing-imports cosine.py \n$\n<\/pre><\/div><p>If you use the <code>--ignore-missing-import<\/code> command line option, Mypy will <a href=\"https:\/\/mypy.readthedocs.io\/en\/stable\/running_mypy.html#how-mypy-handles-imports\">not try to follow or warn about any missing imports<\/a>. This might be a bit heavy-handed though, as it also ignores actual mistakes, like misspelling the name of a package.<\/p><p>Two less intrusive ways of handling third-party packages are using type comments or configuration files.<\/p><p>In a simple example as the one above, you can silence the <code>numpy<\/code> warning by adding a type comment to the line containing the import:<\/p><div class=\"highlight python\"><pre> 3 import numpy as np  # type: ignore\n<\/pre><\/div><p>The literal <code># type: ignore<\/code> tells Mypy to ignore the import of Numpy.<\/p><p>If you have several files, it might be easier to keep track of which imports to ignore in a configuration file. Mypy reads a file called <code>mypy.ini<\/code> in the current directory if it is present. This configuration file must contain a section called <code>[mypy]<\/code> and may contain module specific sections of the form <code>[mypy-module]<\/code>.<\/p><p>The following configuration file will ignore that Numpy is missing type hints:<\/p><div class=\"highlight ini\"><pre># mypy.ini\n[mypy]\n[mypy-numpy]\nignore_missing_imports = True\n<\/pre><\/div><p>There are many options that can be specified in the configuration file. It is also possible to specify a global configuration file. See the <a href=\"https:\/\/mypy.readthedocs.io\/en\/stable\/config_file.html\">documentation<\/a> for more information.<\/p><h3 id=\"adding-stubs\">Adding Stubs<\/h3><p>Type hints are available for all the packages in the Python standard library. However, if you are using third-party packages you\u2019ve already seen that the situation can be different.<\/p><p>The following example uses the <a href=\"https:\/\/pypi.org\/project\/parse\/\">Parse package<\/a> to do simple text parsing. To follow along you should first install Parse:<\/p><div class=\"highlight sh\"><pre>$ pip install parse\n<\/pre><\/div><p>Parse can be used to recognize simple patterns. Here is a small program that tries its best to figure out your name:<\/p><div class=\"highlight python\"><pre> 1 # parse_name.py\n 2 \n 3 import parse\n 4 \n 5 def parse_name(text: str) -&gt; str:\n 6     patterns = (\n 7         \"my name is {name}\",\n 8         \"i'm {name}\",\n 9         \"i am {name}\",\n10         \"call me {name}\",\n11         \"{name}\",\n12     )\n13     for pattern in patterns:\n14         result = parse.parse(pattern, text)\n15         if result:\n16             return result[\"name\"]\n17     return \"\"\n18 \n19 answer = input(\"What is your name? \")\n20 name = parse_name(answer)\n21 print(f\"Hi {name}, nice to meet you!\")\n<\/pre><\/div><p>The main flow is defined in the last three lines: ask for your name, parse the answer, and print a greeting. The <code>parse<\/code> package is called on line 14 in order to try to find a name based on one of the patterns listed on lines 7-11.<\/p><p>The program can be used as follows:<\/p><div class=\"highlight sh\"><pre>$ python parse_name.py\nWhat is your name? I am Geir Arne\nHi Geir Arne, nice to meet you!\n<\/pre><\/div><p>Note that even though I answer <code>I am Geir Arne<\/code>, the program figures out that <code>I am<\/code> is not part of my name.<\/p><p>Let\u2019s add a small bug to the program, and see if Mypy is able to help us detect it. Change line 16 from <code>return result[\"name\"]<\/code> to <code>return result<\/code>. This will return a <code>parse.Result<\/code> object instead of the string containing the name.<\/p><p>Next run Mypy on the program:<\/p><div class=\"highlight sh\"><pre>$ mypy parse_name.py \nparse_name.py:3: error: Cannot find module named 'parse'\nparse_name.py:3: note: (Perhaps setting MYPYPATH or using the\n                       \"--ignore-missing-imports\" flag would help)\n<\/pre><\/div><p>Mypy prints a similar error to the one you saw in the previous section: It doesn\u2019t know about the <code>parse<\/code> package. You could try to ignore the import:<\/p><div class=\"highlight sh\"><pre>$ mypy parse_name.py --ignore-missing-imports\n$\n<\/pre><\/div><p>Unfortunately, ignoring the import means that Mypy has no way of discovering the bug in our program. A better solution would be to add type hints to the Parse package itself. As <a href=\"https:\/\/github.com\/r1chardj0n3s\/parse\">Parse is open source<\/a> you can actually add types to the source code and send a pull request.<\/p><p>Alternatively, you can add the types in a <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/stubs.html\">stub file<\/a>. A stub file is a text file that contains the signatures of methods and functions, but not their implementations. Their main function is to add type hints to code that you for some reason can\u2019t change. To show how this works, we will add some stubs for the Parse package.<\/p><p>First of all, you should put all your stub files inside one common directory, and set the <code>MYPYPATH<\/code> environment variable to point to this directory. On Mac and Linux you can set <code>MYPYPATH<\/code> as follows:<\/p><div class=\"highlight sh\"><pre>$ export MYPYPATH=\/home\/gahjelle\/python\/stubs\n<\/pre><\/div><p>You can set the variable permanently by adding the line to your <code>.bashrc<\/code> file. On Windows you can click the start menu and search for <em>environment variables<\/em> to set <code>MYPYPATH<\/code>.<\/p><p>Next, create a file inside your stubs directory that you call <code>parse.pyi<\/code>. It must be named for the package that you are adding type hints for, with a <code>.pyi<\/code> suffix. Leave this file empty for now. Then run Mypy again:<\/p><div class=\"highlight sh\"><pre>$ mypy parse_name.py\nparse_name.py:14: error: Module has no attribute \"parse\"\n<\/pre><\/div><p>If you have set everything up correctly, you should see this new error message. Mypy uses the new <code>parse.pyi<\/code> file to figure out which functions are available in the <code>parse<\/code> package. Since the stub file is empty, Mypy assumes that <code>parse.parse()<\/code> does not exist, and then gives the error you see above.<\/p><p>The following example does not add types for the whole <code>parse<\/code> package. Instead it shows the type hints you need to add in order for Mypy to type check your use of <code>parse.parse()<\/code>:<\/p><div class=\"highlight python\"><pre># parse.pyi\nfrom typing import Any, Mapping, Optional, Sequence, Tuple, Union\nclass Result:\n    def __init__(\n        self,\n        fixed: Sequence[str],\n        named: Mapping[str, str],\n        spans: Mapping[int, Tuple[int, int]],\n    ) -&gt; None: ...\n    def __getitem__(self, item: Union[int, str]) -&gt; str: ...\n    def __repr__(self) -&gt; str: ...\ndef parse(\n    format: str,\n    string: str,\n    evaluate_result: bool = ...,\n    case_sensitive: bool = ...,\n) -&gt; Optional[Result]: ...\n<\/pre><\/div><p>The ellipsis <code>...<\/code> are part of the file, and should be written exactly as above. The stub file should only contain type hints for variables, attributes, functions, and methods, so the implementations should be left out and replaced by <code>...<\/code> markers.<\/p><p>Finally Mypy is able to spot the bug we introduced:<\/p><div class=\"highlight sh\"><pre>$ mypy parse_name.py\nparse_name.py:16: error: Incompatible return value type (got\n                         \"Result\", expected \"str\")\n<\/pre><\/div><p>This points straight to line 16 and the fact that we return a <code>Result<\/code> object and not the name string. Change <code>return result<\/code> back to <code>return result[\"name\"]<\/code>, and run Mypy again to see that it\u2019s happy.<\/p><h3 id=\"typeshed\">Typeshed<\/h3><p>You\u2019ve seen how to use stubs to add type hints without changing the source code itself. In the previous section we added some type hints to the third-party Parse package. Now, it wouldn\u2019t be very effective if everybody needs to create their own stubs files for all third-party packages they are using.<\/p><p><a href=\"https:\/\/github.com\/python\/typeshed\">Typeshed<\/a> is a Github repository that contains type hints for the Python standard library, as well as many third-party packages. Typeshed comes included with Mypy so if you are using a package that already has type hints defined in Typeshed, the type checking will just work.<\/p><p>You can also <a href=\"https:\/\/github.com\/python\/typeshed\/blob\/master\/CONTRIBUTING.md\">contribute type hints to Typeshed<\/a>. Make sure to get the permission of the owner of the package first though, especially because they might be working on adding type hints into the source code itself\u2014which is the <a href=\"https:\/\/github.com\/python\/typeshed\/blob\/master\/CONTRIBUTING.md#adding-a-new-library\">preferred approach<\/a>.<\/p><h3 id=\"other-static-type-checkers\">Other Static Type Checkers<\/h3><p>In this tutorial, we have mainly focused on type checking using Mypy. However, there are other static type checkers in the Python ecosystem.<\/p><p>The <a href=\"https:\/\/realpython.com\/python-ides-code-editors-guide\/#pycharm\"><strong>PyCharm<\/strong><\/a> editor comes with its own type checker included. If you are using PyCharm to write your Python code, it will be automatically type checked.<\/p><p>Facebook has developed <a href=\"https:\/\/pyre-check.org\/\"><strong>Pyre<\/strong><\/a>. One of its stated goals is to be fast and performant. While there are some differences, Pyre functions mostly similar to Mypy. See the <a href=\"https:\/\/pyre-check.org\/docs\/overview.html\">documentation<\/a> if you\u2019re interested in trying out Pyre.<\/p><p>Furthermore, Google has created <a href=\"https:\/\/github.com\/google\/pytype\"><strong>Pytype<\/strong><\/a>. This type checker also works mostly the same as Mypy. In addition to checking annotated code, Pytype has some support for running type checks on unannotated code and even adding annotations to code automatically. See the <a href=\"https:\/\/github.com\/google\/pytype\/blob\/master\/docs\/quickstart.md\">quickstart<\/a> document for more information.<\/p><p>As a final note, it\u2019s possible to use type hints also at runtime during execution of your Python program. Runtime type checking will probably never be natively supported in Python.<\/p><p>However, the type hints are available at runtime in the <code>__annotations__<\/code> dictionary, and you can use those to do type checks if you desire. Before you run off and write your own package for enforcing types, you should know that there are already several packages doing this for you. Have a look at <a href=\"https:\/\/pypi.org\/project\/enforce\/\"><strong>Enforce<\/strong><\/a>, <a href=\"https:\/\/pypi.org\/project\/pydantic\/\"><strong>Pydantic<\/strong><\/a>, or <a href=\"https:\/\/pypi.org\/project\/pytypes\/\"><strong>Pytypes<\/strong><\/a> for some examples.<\/p><p>Another use of type hints is for translating your Python code to C and compiling it for optimization. The popular <a href=\"https:\/\/cython.org\/\"><strong>Cython<\/strong> project<\/a> uses a hybrid C\/Python language to write statically typed Python code. However, since version 0.27 Cython has also supported type annotations. Recently, the <a href=\"https:\/\/github.com\/mypyc\/mypyc\"><strong>Mypyc<\/strong> project<\/a> has become available. While not yet ready for general use, it can compile some type annotated Python code to C extensions.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>Type hinting in Python is a very useful feature that you can happily live without. Type hints don\u2019t make you capable of writing any code you can\u2019t write without using type hints. Instead, using type hints makes it easier for you to reason about code, find subtle bugs, and maintain a clean architecture.<\/p><p>In this tutorial you have learned how type hinting works in Python, and how gradual typing makes type checks in Python more flexible than in many other languages. You\u2019ve seen some of the pros and cons of using type hints, and how they can be added to code using annotations or type comments. Finally you saw many of the different types that Python supports, as well as how to perform static type checking.<\/p><p>There are many resources to learn more about static type checking in Python. <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0483\/\">PEP 483<\/a> and <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0484\/\">PEP 484<\/a> give a lot of background about how type checking is implemented in Python. The <a href=\"https:\/\/mypy.readthedocs.io\/\">Mypy documentation<\/a> has a great <a href=\"https:\/\/mypy.readthedocs.io\/en\/stable\/builtin_types.html\">reference section<\/a> detailing all the different types available.<\/p><div class=\"border rounded p-3 card mb-2\"><p class=\"mb-0\">Watch Now This tutorial has a related video course created by the Real Python team. Watch it together with the written tutorial to deepen your understanding: <a class=\"stretched-link text-success\" href=\"https:\/\/realpython.com\/courses\/python-type-checking\/\"><strong>Python Type Checking<\/strong><\/a><\/p><\/div>","created_at":"2019-12-10T20:34:37+0000","updated_at":"2019-12-10T20:34:43+0000","published_at":"2019-01-07T14:00:00+0000","published_by":["Real Python"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":73,"domain_name":"realpython.com","preview_picture":"https:\/\/files.realpython.com\/media\/The-Ultimate-Guide-to-Python-Type-Checking_Watermarked.6fc1c7e40ada.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1058,"title":"Python Tips and Trick, You Haven\u2019t Already Seen","url":"https:\/\/towardsdatascience.com\/python-tips-and-trick-you-havent-already-seen-37825547544f?gi=516fc773b4f3","content":"<div><div class=\"ep n eq er es et\"><div class=\"o n\"><div><a href=\"https:\/\/towardsdatascience.com\/@martin.heinz?source=post_page-----37825547544f----------------------\" rel=\"noopener\"><img alt=\"Martin Heinz\" class=\"r eu ev ew\" src=\"https:\/\/miro.medium.com\/fit\/c\/96\/96\/2*p3sIQ1Ga2bfAbZYzoCcACw.jpeg\" width=\"48\" height=\"48\" referrerpolicy=\"no-referrer\" \/><\/a><\/div><div class=\"ex ak r\"><p><a class=\"dc dd bb bc bd be bf bg bh bi fg bl bm fh fi\" rel=\"noopener\" href=\"https:\/\/towardsdatascience.com\/python-tips-and-trick-you-havent-already-seen-37825547544f?source=post_page-----37825547544f----------------------\">Sep 29<\/a> \u00b7 5 min read<\/p><\/div><\/div><\/div><\/div><p id=\"edb4\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\"><em class=\"gw\">Note: This was originally posted at<\/em> <a href=\"https:\/\/martinheinz.dev\/blog\/1\" class=\"dc by gx gy gz ha\" target=\"_blank\" rel=\"noopener nofollow\"><em class=\"gw\">martinheinz.dev<\/em><\/a><\/p><p id=\"2bb0\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">There are plenty of articles written about lots of cool features in Python such as variable unpacking, partial functions, enumerating iterables, but there is much more to talk about when it comes to Python, so here I will try to show some of the features I know and use, that I haven\u2019t yet seen mentioned elsewhere. So here we go.<\/p><figure class=\"hc hd he hf hg hh do dp paragraph-image\">\n<\/figure><h2 id=\"0b90\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Sanitizing String Input<\/h2><p id=\"5149\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">Problem of sanitizing user input applies to almost every program you might write. Often it\u2019s enough to convert characters to lower or upper-case, sometimes you can use Regex to do the work, but for complex cases, there might be a better way:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"4728\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">In this example you can see that whitespace characters \u201c\\n\u201d and \u201c\\t\u201d have been replaced by single space and \u201c\\r\u201d has been removed completely. This is a simple example, but we could take it much further and generate big remapping tables using <code class=\"hs is it iu iv b\">unicodedata<\/code> package and its <code class=\"hs is it iu iv b\">combining()<\/code> function to generate and map which we could use to remove all accents from string.<\/p><h2 id=\"b3d8\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Taking Slice of an Iterator<\/h2><p id=\"99b4\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">If you try to take slice of an Iterator, you will get a <code class=\"hs is it iu iv b\">TypeError<\/code>, stating that generator object is not subscriptable, but there is a easy solution to that:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"e763\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">Using <code class=\"hs is it iu iv b\">itertools.islice<\/code> we can create a <code class=\"hs is it iu iv b\">islice<\/code> object which is an iterator that produces desired items. It's important to note though, that this consumes all generator items up until the start of slice and also all the items in our <code class=\"hs is it iu iv b\">islice<\/code> object.<\/p><h2 id=\"977f\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Skipping Begining of Iterable<\/h2><p id=\"96eb\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">Sometimes you have to work with files which you know that start with variable number of unwanted lines such as comments. <code class=\"hs is it iu iv b\">itertools<\/code> again provides easy solution to that:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"661b\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">This code snippet produces only lines after initial comment section. This approach can be useful in case we only want to discard items (lines in this instance) at the beginning of the iterable and don\u2019t know how many of them there are.<\/p><h2 id=\"b983\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Functions with only Keyword Arguments (kwargs)<\/h2><p id=\"1ec1\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">It can be helpful to create function that only takes keyword arguments to provide (force) more clarity when using such function:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"0eba\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">As you can see this can be easily solved by placing single <code class=\"hs is it iu iv b\">*<\/code> argument before keyword arguments. There can obviously be positional arguments if we place them before the <code class=\"hs is it iu iv b\">*<\/code> argument.<\/p><h2 id=\"5269\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Creating Object That Supports <code class=\"hs is it iu iv b\">with<\/code> Statements<\/h2><p id=\"c697\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">We all know how to, for example open file or maybe acquire locks using <code class=\"hs is it iu iv b\">with<\/code> statement, but can we actually implement our own? Yes, we can implement context-manager protocol using <code class=\"hs is it iu iv b\">__enter__<\/code> and <code class=\"hs is it iu iv b\">__exit__<\/code> methods:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"9e0f\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">This is the most common way to implement context management in Python, but there is easier way to do it:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"18a0\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">The snippet above implements the content management protocol using <code class=\"hs is it iu iv b\">contextmanager<\/code> manager decorator. The first part of the <code class=\"hs is it iu iv b\">tag<\/code> function (before <code class=\"hs is it iu iv b\">yield<\/code>) is executed when entering the <code class=\"hs is it iu iv b\">with<\/code> block, then the block is executed and finally rest of the <code class=\"hs is it iu iv b\">tag<\/code> function is executed.<\/p><h2 id=\"1921\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Saving Memory with <code class=\"hs is it iu iv b\">__slots__<\/code><\/h2><p id=\"07ce\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">If you ever wrote a program that was creating really big number of instances of some class, you might have noticed that your program suddenly needed a lot of memory. That is because Python uses dictionaries to represent attributes of instances of classes, which makes it fast but not very memory efficient, which is usually not a problem. However, if it becomes a problem for your program you might try using <code class=\"hs is it iu iv b\">__slots__<\/code>:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"c5a0\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">What happens here is that when we define <code class=\"hs is it iu iv b\">__slots__<\/code> attribute, Python uses small fixed-size array for the attributes instead of dictionary, which greatly reduces memory needed for each instance. There are also some downsides to using <code class=\"hs is it iu iv b\">__slots__<\/code> - we can't declare any new attributes and we are restricted to using ones on <code class=\"hs is it iu iv b\">__slots__<\/code>. Also classes with <code class=\"hs is it iu iv b\">__slots__<\/code> can't use multiple inheritance.<\/p><h2 id=\"5237\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Limiting CPU and Memory Usage<\/h2><p id=\"ddab\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">If instead of optimizing your program memory or CPU usage, you want to just straight up limit it to some hard number, then Python has a library for that too:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"67d7\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">Here we can see both options to set maximum CPU runtime as well as maximum memory used limit. For CPU limit we first get soft and hard limit for that specific resource ( <code class=\"hs is it iu iv b\">RLIMIT_CPU<\/code>) and then set it using number of seconds specified by argument and previously retrieved hard limit. Finally, we register signal that causes system exit if CPU time is exceeded. As for the memory, we again retrieve soft and hard limit and set it using <code class=\"hs is it iu iv b\">setrlimit<\/code> with size argument and retrieved hard limit.<\/p><h2 id=\"4d31\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Controlling What Can Be Imported and What Not<\/h2><p id=\"f0a0\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">Some languages have very obvious mechanism for exporting members (variables, methods, interfaces) such as Golang, where only members starting with upper-case letter are exported. In Python on the other hand, everything is exported, unless we use <code class=\"hs is it iu iv b\">__all__<\/code>:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"f749\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">Based on code snippet above, we know that only <code class=\"hs is it iu iv b\">bar<\/code> function will be exported. Also, we can leave <code class=\"hs is it iu iv b\">__all__<\/code> empty and nothing will be exported casing <code class=\"hs is it iu iv b\">AttributeError<\/code> when importing from this module.<\/p><h2 id=\"5796\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Comparison Operators the Easy Way<\/h2><p id=\"583a\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">It can be pretty annoying to implement all the comparison operators for one class, considering there are quite a few of them \u2014 <code class=\"hs is it iu iv b\">__lt__ , __le__ , __gt__ ,<\/code> or <code class=\"hs is it iu iv b\">__ge__<\/code>. But what if there was an easier way to do it? <code class=\"hs is it iu iv b\">functools.total_ordering<\/code> to the rescue:<\/p><figure class=\"hc hd he hf hg hh\">\n<\/figure><p id=\"8571\" class=\"gi gj ef at gk b gl gm gn go gp gq gr gs gt gu gv\">How does this actually work? <code class=\"hs is it iu iv b\">total_ordering<\/code> decorator is used to simplify the process of implementing ordering of instances for our class. It's only needed to define <code class=\"hs is it iu iv b\">__lt__<\/code> and <code class=\"hs is it iu iv b\">__eq__<\/code>, which is the minimum needed for mapping of remaining operations and that's the job of decorator - it fills the gaps for us.<\/p><h2 id=\"22c4\" class=\"hy hz ef at as ia ib ic id ie if ig ih ii ij ik il\">Conclusion<\/h2><p id=\"fe67\" class=\"gi gj ef at gk b gl im gn in gp io gr ip gt iq gv\">Not all these features are essential and useful in day-to-day Python programming, but some of them might come in handy from time to time and they also might simplify task that would be otherwise quite lengthy and annoying to implement. Also I want to note that all those features are part of Python standard library, while some of them seem to me like pretty non-standard things to have in standard library, so whenever you decide to implement something in Python first go looking for it in standard library and if you can\u2019t find it, then you are probably not looking hard enough (if it\u2019s really not there, then it\u2019s surely in some third party library). \ud83d\ude42<\/p>","created_at":"2019-12-09T18:39:42+0000","updated_at":"2019-12-09T18:39:42+0000","published_at":"2019-09-29T14:41:45+0000","published_by":["Martin Heinz"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"towardsdatascience.com","preview_picture":"https:\/\/miro.medium.com\/max\/1200\/1*qplDceAY84tdlnllh7wlEg.jpeg","http_status":"200","origin_url":"https:\/\/towardsdatascience.com\/python-tips-and-trick-you-havent-already-seen-37825547544f"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1057,"title":"Orger: plaintext reflection of your digital self","url":"https:\/\/beepb00p.xyz\/orger.html","content":"<section class=\"post-title\"><p>Mirror your personal data as org-mode for instant access and search<\/p><\/section><section class=\"content\"><div id=\"table-of-contents\"><h2>Table of Contents<\/h2><div id=\"text-table-of-contents\"><ul><li><a href=\"#intro\">1. Intro<\/a><\/li>\n<li><a href=\"#what\">2. What Orger does<\/a><\/li>\n<li><a href=\"#demo\">3. Demo: displaying Pocket data via Orger<\/a><\/li>\n<li><a href=\"#examples\">4. More examples<\/a>\n<ul><li><a href=\"#org0000005\">instapaper<\/a><\/li>\n<li><a href=\"#org0000006\">hypothesis<\/a><\/li>\n<li><a href=\"#org0000007\">kobo<\/a><\/li>\n<li><a href=\"#org0000008\">pinboard<\/a><\/li>\n<li><a href=\"#org0000009\">pdfs<\/a><\/li>\n<li><a href=\"#org000000a\">twitter<\/a><\/li>\n<li><a href=\"#org000000b\">rtm2org<\/a><\/li>\n<li><a href=\"#org000000c\">telegram2org<\/a><\/li>\n<li><a href=\"#org000000d\">reddit<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#why_hard\">5. It does sound very simple. Does that really deserve a post?<\/a><\/li>\n<li><a href=\"#org0000010\">6. What makes Orger good?<\/a><\/li>\n<li><a href=\"#using\">7. Using Orger views<\/a>\n<ul><li><a href=\"#usecases\">Typical use patterns<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#improvements\">8. Potential improvements<\/a>\n<ul><li><a href=\"#org0000013\">TODO more frequent, ideally realtime updates to views<\/a><\/li>\n<li><a href=\"#org0000014\">TODO alternative export formats<\/a><\/li>\n<li><a href=\"#org0000015\">TODO two-way data flow<\/a><\/li>\n<li><a href=\"#org0000016\">TODO potential for race condition<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#similar\">9. Similar projects<\/a><\/li>\n<li><a href=\"#org0000019\">10. ----<\/a><\/li>\n<\/ul><\/div><\/div><p>TLDR: I'll write about <a href=\"https:\/\/github.com\/karlicoss\/orger\">orger<\/a>, a tool I'm using to convert my personal data into easily readable and searchable org-mode views. I'll present some examples and use cases, that will hopefully be helpful to you as well even if you are not sold by using my tool.<\/p><p>There is also <a href=\"https:\/\/beepb00p.xyz\/orger-todos.html\">second part<\/a> where I'm explaining how it can be used to process Reddit, create quick tasks from Telegram messages and help with spaced repetition. If you're impatient, you can jump straight to a <a href=\"#demo\">demo<\/a>.<\/p><div class=\"outline-2\" id=\"outline-container-org0000001\"><div class=\"outline-text-2\" id=\"text-intro\"><p>I consume lots of digital content (books, articles, reddit, youtube, reddit, etc.) and most of it I find somewhat useful and insightful. I want to use that knowledge later, act and build on it. But there's an obstacle: human brain.<\/p><p>It would be cool to be capable of always remembering and instantly recalling information you've interacted with, metadata and your thoughts on it. Until we get augmented though, there are two options: first is just to accept it and live with it. You might have guessed this is not the option I'm taking.<\/p><div>Second option is compensating for your sloppy meaty memory and having information you've read at hand and a quick way of searching over it.<aside class=\"sidenote\">see <a class=\"post-tag\" href=\"https:\/\/beepb00p.xyz\/tags.html#extendedmind\">#extendedmind<\/a><\/aside><\/div><p>That sounds simple enough but as with many simple things, on practice you run into obstacles. I'll give some I've personally been overcoming as examples:<\/p><ul class=\"org-ul\"><li>convenience of access, e.g.:\n<ul class=\"org-ul\"><li>to access highlights and notes on my Kobo ebook I need to actually reach my reader and tap through e-ink touch screen. Not much fun!<\/li>\n<li>\n<div>if you want to search over annotations in your PDF collections\u2026 well good luck, I'm just not aware of such a tool. It's actually way worse: many PDF viewers wouldn't even let you search through highlights within the file you're currently viewing.<aside class=\"sidenote\">see PDF readers <a href=\"https:\/\/beepb00p.xyz\/annotating.html#pdfs\">overview<\/a><\/aside><\/div>\n<\/li>\n<li>there is no easy way to access all of your twitter favorites, people suggest using hacks like <a href=\"https:\/\/webapps.stackexchange.com\/a\/131374\/68104\">autoscroll extension<\/a>.<\/li>\n<\/ul><\/li>\n<li>searching data, e.g.:\n<ul class=\"org-ul\"><li>search function often just isn't available at all or search is not incremental, e.g. on Instapaper, you can't restrict search to highlights<\/li>\n<li>builtin browser search (<code class=\"inline\">Ctrl-F<\/code>) sucks for the most part: it's not very easy to navigate as you don't get previews and you have to look through every match<\/li>\n<li>sometimes you vaguely recall reading about something but don't remember where exactly. Was it on stackoverflow? Or in some github issue? Or in a conversation with friend?<\/li>\n<\/ul><\/li>\n<li>data ownership and liberation, e.g.\n<ul class=\"org-ul\"><li>what happens if data disappears or service is down (temporary\/permanently) or banned by your government?<\/li>\n<li>\n<div>99% of services don't have support for offline mode. This may be just a small inconvenience if you're on a train or something, but there is more to it. What if some sort of apocalypse happens and you lose all access to data? That depends on your paranoia level of course, and apocalypse is bad enough as it is, but my take on it is that at least I'd have my data :)<aside class=\"sidenote\">see <a class=\"post-tag\" href=\"https:\/\/beepb00p.xyz\/tags.html#offline\">#offline<\/a><\/aside><\/div>\n<\/li>\n<li>if you delete a book on Kobo, not only you can't access its annotations, but they seem to get <strong>wiped<\/strong> from the database.<\/li>\n<li>in 2018, instapaper was unavailable in Europe for several months (!) due to missing the GDPR deadline<\/li>\n<\/ul><\/li>\n<\/ul><p>Thinking about that and tinkering helped me understand what I want: some sort of <strong>search engine<\/strong>, over my <strong>personal data<\/strong>, with <strong>uniform<\/strong> and <strong>always available<\/strong> way of accessing it.<\/p><p>So, I present you a system that I've developed and that solves all my problems\u2122: <a href=\"https:\/\/github.com\/karlicoss\/orger\">orger<\/a>.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000003\"><h2 id=\"what\">2 What Orger does<\/h2><div class=\"outline-text-2\" id=\"text-what\"><p>It's really so quite trivial that it's almost stupid. Orger provides a simple python API to render any data as an Org-mode file. It's easier to give an example:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">from orger import StaticView\nfrom orger.inorganic import node, link\nfrom orger.common import dt_heading\nimport my.github_data\nclass Github(StaticView):\n  def get_items(self):\n    for event in my.github_data.get_events():\n      yield node(dt_heading(event.dt, event.summary))\nGithub.main()\n<\/pre><\/div><p>That ten line program results in a file <samp class=\"inline\">Github.org<\/samp>:<\/p><div class=\"org-src-container\"><pre class=\"src src-org\"># AUTOGENERATED BY \/code\/orger\/github.py\n* [2016-10-30 Sun 10:29] opened PR Add __enter__ and __exit__ to Pool stub\n* [2016-11-10 Thu 09:29] opened PR Update gradle to 2.14.1 and gradle plugin to 2.1.1\n* [2016-11-16 Wed 20:20] commented on issue Linker error makes it impossible to use a stack-provided ghc\n* [2016-12-30 Fri 11:57] commented on issue Fix performance in the rare case of hashCode evaluating to zero \n* [2019-09-21 Sat 16:51] commented on issue Tags containing letters outside of a-zA-Z\n....\n<\/pre><\/div><p>Even with event summaries only it can already be very useful to search over. What you can potentially do really depends on your imagination and needs! You can also add:<\/p><ul class=\"org-ul\"><li>links<\/li>\n<li>tags<\/li>\n<li>timestamps<\/li>\n<li>properties<\/li>\n<li>child nodes<\/li>\n<\/ul><p>See <a href=\"#examples\">'Examples'<\/a> section for more.<\/p><p>So as you can see orger itself is a really not sophisticated tool, at least until you spend time trying to reimplement the same. As always the devil is in the details (look at that cheeky <code class=\"inline\">my.github_data<\/code> import), which I'll explain <a href=\"#org0000002\">further<\/a>.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000004\"><h2 id=\"demo\">3 Demo: displaying Pocket data via Orger<\/h2><div class=\"outline-text-2\" id=\"text-demo\"><p>I've documented one of modules, <a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/pocket_demo.py\"><samp class=\"inline\">pocket_demo<\/samp><\/a> so you could get the sense of using Orger.<\/p><details><summary>Click to view the code<\/summary><div class=\"org-src-container\"><pre class=\"src src-python\">#!\/usr\/bin\/env python3\n\"\"\"\nDemo Orger adapter for Pocket data. For documentation purposes, so please modify pocket.py if you want to contribute.\n\"\"\"\n\"\"\"\nFirst we define some abstractions for Pocket entities (articles and highlights).\nWhile it's not that necessary and for one script you can get away with using json directly,\n it does help to separate parsing and rendering, allows you to reuse parsing for other projects\n and generally makes everything clean.\nAlso see https:\/\/github.com\/karlicoss\/my package for some inspiration.\n\"\"\"\nfrom datetime import datetime\nfrom pathlib import Path\nfrom typing import NamedTuple, Sequence, Any\nclass Highlight(NamedTuple):\n    \"\"\"\n    Abstract representation of Pocket highlight\n    \"\"\"\n    json: Any\n    @property\n    def text(self) -&gt; str:\n        return self.json['quote']\n    @property\n    def created(self) -&gt; datetime:\n        return datetime.strptime(self.json['created_at'], '%Y-%m-%d %H:%M:%S')\nclass Article(NamedTuple):\n    \"\"\"\n    Abstract representation of Pocket saved page\n    \"\"\"\n    json: Any\n    @property\n    def url(self) -&gt; str:\n        return self.json['given_url']\n    @property\n    def title(self) -&gt; str:\n        return self.json['given_title']\n    @property\n    def pocket_link(self) -&gt; str:\n        return 'https:\/\/app.getpocket.com\/read\/' + self.json['item_id']\n    @property\n    def added(self) -&gt; datetime:\n        return datetime.fromtimestamp(int(self.json['time_added']))\n    @property\n    def highlights(self) -&gt; Sequence[Highlight]:\n        raw = self.json.get('annotations', [])\n        return list(map(Highlight, raw))\n    # TODO add tags?\ndef get_articles(json_path: Path) -&gt; Sequence[Article]:\n    \"\"\"\n    Parses Pocket export produced by https:\/\/github.com\/karlicoss\/pockexport\n    \"\"\"\n    import json\n    raw = json.loads(json_path.read_text())['list']\n    return list(map(Article, raw.values()))\n\"\"\"\nOk, now we can get to implementing the adapter.\n\"\"\"\nfrom orger import StaticView\n\"\"\"\nStaticView means it's meant to be read-only view onto data (as opposed to InteractiveView).\n\"\"\"\nfrom orger.inorganic import node, link\nfrom orger.common import dt_heading\nclass PocketView(StaticView):\n    def get_items(self):\n        \"\"\"\n        get_items returns a sequence\/iterator of nodes\n        see orger.inorganic.OrgNode to find out about attributes you can use\n        \"\"\"\n        export_file = self.cmdline_args.file # see setup_parser\n        for a in get_articles(export_file):\n            yield node(\n                heading=dt_heading(\n                    a.added,\n                    link(title=a.title, url=a.url)\n                ),\n                body=link(title='Pocket link', url=a.pocket_link), # permalink is pretty convenient to jump straight into Pocket app\n                children=[node( # comments are displayed as org-mode child entries\n                    heading=dt_heading(hl.created, hl.text)\n                ) for hl in a.highlights]\n            )\ndef setup_parser(p):\n    \"\"\"\n    Optional hooks for extra arguments if you need them in your adapter\n    \"\"\"\n    p.add_argument('--file', type=Path, help='JSON file from API export', required=True)\nif __name__ == '__main__':\n    \"\"\"\n    Usage example: .\/pocket.py --file \/backups\/pocket\/last-backup.json --to \/data\/orger\/pocket.org\n    \"\"\"\n    PocketView.main(setup_parser=setup_parser)\n\"\"\"\nExample pocket.org output:\n# AUTOGENERATED BY \/L\/zzz_syncthing\/coding\/orger\/pocket.py\n* [2018-07-09 Mon 10:56] [[https:\/\/www.gwern.net\/Complexity-vs-AI][Complexity no Bar to AI - Gwern.net]]\n [[https:\/\/app.getpocket.com\/read\/1949330650][Pocket link]]\n* [2016-10-21 Fri 14:42] [[https:\/\/johncarlosbaez.wordpress.com\/2016\/09\/09\/struggles-with-the-continuum-part-2\/][Struggles with the Continuum (Part 2) | Azimuth]]\n [[https:\/\/app.getpocket.com\/read\/1407671000][Pocket link]]\n* [2016-05-31 Tue 18:25] [[http:\/\/www.scottaaronson.com\/blog\/?p=2464][Bell inequality violation finally done right]]\n [[https:\/\/app.getpocket.com\/read\/1042711293][Pocket link]]\n* [2016-05-31 Tue 18:24] [[https:\/\/packetzoom.com\/blog\/how-to-test-your-app-in-different-network-conditions.html][How to test your app in different network conditions -]]\n [[https:\/\/app.getpocket.com\/read\/1188624587][Pocket link]]\n* [2016-05-31 Tue 18:24] [[http:\/\/www.schibsted.pl\/2016\/02\/hood-okhttps-cache\/][What's under the hood of the OkHttp's cache?]]\n [[https:\/\/app.getpocket.com\/read\/1191143185][Pocket link]]\n* [2016-03-15 Tue 17:27] [[http:\/\/joeduffyblog.com\/2016\/02\/07\/the-error-model\/][Joe Duffy - The Error Model]]\n [[https:\/\/app.getpocket.com\/read\/1187239791][Pocket link]]\n** [2019-09-25 Wed 18:20] A bug is a kind of error the programmer didn\u2019t expect. Inputs weren\u2019t validated correctly, logic was written wrong, or any host of problems have arisen.\n** [2019-09-25 Wed 18:19] First, throwing an exception is usually ridiculously expensive. This is almost always due to the gathering of a stack trace.\n** [2019-09-25 Wed 18:20] In other words, an exception, as with error codes, is just a different kind of return value!\n\"\"\"\n<\/pre><\/div>\n<\/details><details><summary>Click to view the output<\/summary><div id=\"content\"><div class=\"outline-2\" id=\"outline-container-org6385060\"><div class=\"outline-text-2\" id=\"text-1\"><p><a href=\"https:\/\/app.getpocket.com\/read\/1949330650\">Pocket link<\/a><\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-orgf66ad5f\"><div class=\"outline-text-2\" id=\"text-2\"><p><a href=\"https:\/\/app.getpocket.com\/read\/1407671000\">Pocket link<\/a><\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-orga907d2c\"><div class=\"outline-text-2\" id=\"text-3\"><p><a href=\"https:\/\/app.getpocket.com\/read\/1042711293\">Pocket link<\/a><\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-orgaa7b53a\"><div class=\"outline-text-2\" id=\"text-4\"><p><a href=\"https:\/\/app.getpocket.com\/read\/1188624587\">Pocket link<\/a><\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org01f1d77\"><div class=\"outline-text-2\" id=\"text-5\"><p><a href=\"https:\/\/app.getpocket.com\/read\/1191143185\">Pocket link<\/a><\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-orgdd27fdd\"><div class=\"outline-text-2\" id=\"text-6\"><p><a href=\"https:\/\/app.getpocket.com\/read\/1187239791\">Pocket link<\/a><\/p><\/div><p><\/p><h3 id=\"org92bca32\">[2019-09-25 Wed 18:20] A bug is a kind of error the programmer didn\u2019t expect. Inputs weren\u2019t validated correctly, logic was written wrong, or any host of problems have arisen.<\/h3><p><\/p><h3 id=\"org41ce646\">[2019-09-25 Wed 18:19] First, throwing an exception is usually ridiculously expensive. This is almost always due to the gathering of a stack trace.<\/h3><p><\/p><h3 id=\"orge27b04d\">[2019-09-25 Wed 18:20] In other words, an exception, as with error codes, is just a different kind of return value!<\/h3><\/div><\/div>\n<\/details><p>As you can see it's quite easy to search in your highlights and jump straight in the pocket app to the article you were reading.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org000000e\"><h2 id=\"examples\">4 More examples<\/h2><div class=\"outline-text-2\" id=\"text-examples\"><p>I'm using more than ten different Orger modules, most of which I've moved into the <a href=\"https:\/\/github.com\/karlicoss\/orger\/tree\/master\/modules\">repository<\/a>. Here I'll describe some featured views I'm generating.<\/p><p>To give you a heads up, if you read the code, you'll see bunch of imports like <code class=\"inline\">from my.hypothesis import ...<\/code>. I find it easier to move all data parsing in a separate <a href=\"https:\/\/github.com\/karlicoss\/my\"><samp class=\"inline\">my<\/samp><\/a> package, that deals with parsing and converting input data (typically, some JSON). That makes everything less messy, separates data and rendering and lets me reuse abstract models in other tools. Also that lets me access my data from any python code, which makes it way easier to use and interact with data.<\/p><p>Some of these are still private so if you're interested in something not present in the github repo, please don't be shy and open an issue so I can prioritize.<\/p><p>Hopefully the code is readable enough and will give you some inspiration. If you find something confusing or you write your own module and want to contribute, please feel free to open issue\/PR!<\/p><\/div><div class=\"outline-3\" id=\"outline-container-org0000005\"><h3 id=\"org0000005\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/instapaper.py\">instapaper<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org0000005\"><p>Instapaper <a href=\"https:\/\/beepb00p.xyz\/annotating.html#instapaper\">doesn't have search<\/a> over annotations, so I implemented my own!<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000006\"><h3 id=\"org0000006\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/hypothesis.py\">hypothesis<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org0000006\"><p><a href=\"https:\/\/hypothes.is\">Hypothesis<\/a> does have search, but it's still way quicker for me to invoke search in Emacs (takes literally less than a second) than do that in web browser.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000007\"><h3 id=\"org0000007\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/kobo.py\">kobo<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org0000007\"><p>Generates views for all highlights and comments along with book titles from my Kobo database export.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000008\"><h3 id=\"org0000008\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/pinboard.py\">pinboard<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org0000008\"><p>Searches over my Pinboard bookmarks.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000009\"><h3 id=\"org0000009\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/pdfs.py\">pdfs<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org0000009\"><p>Crawls my filesystem for PDF files and collects all highlights and comments in a single view.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000000a\"><h3 id=\"org000000a\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/twitter.py\">twitter<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org000000a\"><p>It's got two modes<\/p><ul class=\"org-ul\"><li>First mode generates a view of everything I've ever tweeted, so I can search over it.<\/li>\n<li>Second mode generates a view of all older tweets from the previous years posted on the same day. I find it quite fascinating to read through it and observe how I've been changing over years.<\/li>\n<\/ul><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000000b\"><h3 id=\"org000000b\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/rtm2org.py\">rtm2org<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org000000b\"><p>I stopped using Remember The Milk a while ago, but there are still some tasks and notes I've left behind, which I'm slowly moving to org-mode or canceling over time.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000000c\"><h3 id=\"org000000c\"><a href=\"https:\/\/github.com\/karlicoss\/telegram2org\">telegram2org<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org000000c\"><p>Lets me create todo tasks from Telegram messages in couple of taps (you can't use share function on them in Android).<\/p><p>I write about it <a href=\"https:\/\/beepb00p.xyz\/orger-todos.html#telegram2org\">in the second part<\/a>.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000000d\"><h3 id=\"org000000d\"><a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/modules\/reddit.py\">reddit<\/a><\/h3><div class=\"outline-text-3\" id=\"text-org000000d\"><p>Displays and lets me search my Reddit saved posts\/comments.<\/p><p>I write about it <a href=\"https:\/\/beepb00p.xyz\/orger-todos.html#reddit\">in the second part<\/a>.<\/p><\/div><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org000000f\"><h2 id=\"why_hard\">5 It does sound very simple. Does that really deserve a post?<\/h2><div class=\"outline-text-2\" id=\"text-why_hard\"><p>Well yeah it really does seem simple\u2026 until you try to do it.<\/p><ul class=\"org-ul\"><li>\n<p>emitting Org-mode<\/p>\n<p>While it's plaintext, and generating simple outlines is trivial, with more sophisticated inputs, there is some nasty business of escaping and sanitizing that has to be dealt with. I didn't manage to find any Python libraries capable of emitting Org-mode. Only project I knew of was <a href=\"https:\/\/github.com\/bjonnh\/PyOrgMode\">PyOrgMode<\/a> but the author abandoned it.<\/p>\n<p>When it comes to generating 10+ views from different data sources, you <strong>really<\/strong> want to make sure it's as little effort and minimal boilerplate as it can possibly be.<\/p>\n<p>That's how <a href=\"https:\/\/github.com\/karlicoss\/inorganic\">inorganic<\/a> library was born.<\/p>\n<\/li>\n<li>\n<p>accessing data sources and exposing it through Python interfaces<\/p>\n<p>This is probably where most of effort was spent. All sorts of stupid APIs, tedious parsing, you can imagine.<\/p>\n<p>I'll write separately about it sometime, for now you can see some of the code I prettified and shared in my github <a href=\"https:\/\/github.com\/search?q=topic:export+user:karlicoss&amp;type=Repositories\">'export'<\/a> and <a href=\"https:\/\/github.com\/karlicoss\/my\">my<\/a> packages. I tried to make sure they are easy to use for other people and not specific to my use cases.<\/p>\n<\/li>\n<li>\n<p>keeping track of already processed items for <samp class=\"inline\">Interactive<\/samp> views<\/p>\n<p>Because there is no feedback from org-mode files back to data sources, you want to keep track of items already added in the file, otherwise you're gonna have duplicates.<\/p>\n<p>It's not rocket science of course, but it is quite tedious. There is some additional logic that checks for lock files, makes sure writes are atomic, etc. You really don't want to implement it more than once. I figured it was worth extracting this 'pattern' in a separate <a href=\"https:\/\/github.com\/karlicoss\/orger\/blob\/master\/src\/orger\/state.py\">python module<\/a>.<\/p>\n<\/li>\n<\/ul><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000010\"><h2 id=\"org0000010\">6 What makes Orger good?<\/h2><div class=\"outline-text-2\"><ul class=\"org-ul\"><li>\n<p>it solves <a href=\"#org0000000\">my problems<\/a>!<\/p>\n<p>I won't go long into Org-mode propaganda, there are people that do it better than me out there, but for me it's good because it's a decent balance between ease of use and ease of augmenting.<\/p>\n<ul class=\"org-ul\"><li>it's easy to do unstructured (i.e. grep) or structured (i.e. tag search in emacs) search on any of your devices be it desktop or phone<\/li>\n<li>you can open it anywhere you can open a text file<\/li>\n<li>tasks as easy to create as any other Org outline so it can integrate with your todo list and agenda (see more in <a href=\"https:\/\/beepb00p.xyz\/orger-todos.html\">the second part<\/a>).<\/li>\n<\/ul><\/li>\n<li>\n<p>it doesn't require Emacs<\/p>\n<p>If you're not willing to go full on Emacs, you can still benefit from this setup by using plaintext viewer and search tool of your choice.<\/p>\n<\/li>\n<li>\n<p>written in Python. I don't claim at all that Python is the best programming language, but that's the one I'm most productive on as well as many other people.<\/p>\n<p>Also the fact that it's a real programming language rather than some YAML config makes sure you can do anything and not restricted by stupid DSL.<\/p>\n<\/li>\n<li>it's extremely easy to add new views \u2014 a matter of 10-20 lines of code.<\/li>\n<li>agnostic to what you feed in it \u2013 it could be offline data from your regular backups, or it could be fresh API data. Again, it's a real programming language, you can do literally anything.<\/li>\n<\/ul><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000012\"><h2 id=\"using\">7 Using Orger views<\/h2><div class=\"outline-text-2\" id=\"text-using\"><p>Apart from, obviously, opening org mode file in your favorite text editor, one major strength of this system is being able to effortlessly search over them.<\/p><p>I'm writing extensively about my information search setup <a href=\"https:\/\/beepb00p.xyz\/pkm-search.html#personal_information\">here<\/a>. In summary:<\/p><ul class=\"org-ul\"><li>on my desktop I'm just using spacemacs or <a href=\"https:\/\/github.com\/karlicoss\/cloudmacs\">cloudmacs<\/a> from web browser<\/li>\n<li>in Emacs, I'm usually just using <a href=\"https:\/\/github.com\/syohex\/emacs-helm-ag\">helm-ag<\/a> with <a href=\"https:\/\/github.com\/BurntSushi\/ripgrep\">ripgrep<\/a><\/li>\n<li>sometimes <a href=\"https:\/\/github.com\/emacsorphanage\/helm-swoop\">helm-swoop<\/a> is very convenient<\/li>\n<li><a href=\"https:\/\/orgmode.org\/manual\/Matching-tags-and-properties.html\"><samp class=\"inline\">org-tags-view<\/samp><\/a> or <a href=\"https:\/\/github.com\/alphapapa\/org-ql#helm-org-ql\"><samp class=\"inline\">helm-org-ql<\/samp><\/a> for structured Org-mode search<\/li>\n<li>I've got hotkeys set up that invoke Emacs window with search prompt in a blink<\/li>\n<\/ul><p>On my phone I'm using:<\/p><ul class=\"org-ul\"><li><a href=\"https:\/\/github.com\/orgzly\/orgzly-android\">orglzy<\/a> for structured search\/viewing Org-mode files<\/li>\n<li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.docsearch.pro\">Docsearch +<\/a> for<\/li>\n<\/ul><p>You can also set up some proper indexing daemon like <a href=\"https:\/\/www.lesbonscomptes.com\/recoll\">recoll<\/a>.<\/p><\/div><div class=\"outline-3\" id=\"outline-container-org0000011\"><h3 id=\"usecases\">Typical use patterns<\/h3><div class=\"outline-text-3\" id=\"text-usecases\"><p>I'll just give some of my use cases:<\/p><ul class=\"org-ul\"><li>\n<p>While running tests for <a href=\"https:\/\/github.com\/karlicoss\/orgparse\">orgparse<\/a> I started randomly getting <samp class=\"inline\">AssertionError: Cannot find component 'A@3' for 'orgparse.A@3<\/samp>.<\/p>\n<p>I recall that I had same issue few month ago but don't quite remember what was the fix. I press <samp class=\"inline\">F1<\/samp> which invokes <code class=\"inline\">helm-ag<\/code> for me and type 'cannot find component'. I instantly find a github <a href=\"https:\/\/github.com\/python\/mypy\/issues\/7281\">issue<\/a> I opened in <samp class=\"inline\">github.org<\/samp> and figure out what I need to do to work around the problem.<\/p>\n<\/li>\n<li>\n<p>While discussing special relativity with a friend, I recall watching some intuitive rationale for Maxwell's equations, but don't quite recall what was the video.<\/p>\n<p>I press <samp class=\"inline\">F1<\/samp>, type 'Special relativity' and instantly get few results, in particular <a href=\"https:\/\/www.youtube.com\/watch?v=1TKSfAkWWN0\">this<\/a> awesome Veritasium video in <samp class=\"inline\">youtube.org<\/samp>, which I was looking for.<\/p>\n<\/li>\n<li>\n<p>Recommending books<\/p>\n<p>I often struggle to recall the details why I liked a particular book, especially fiction. Having all annotations in my <samp class=\"inline\">kobo.org<\/samp> file lets me quickly look up and skim through highlighted bits so I can freshen up my memory.<\/p>\n<\/li>\n<\/ul><\/div><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000017\"><h2 id=\"improvements\">8 Potential improvements<\/h2><div class=\"outline-3\" id=\"outline-container-org0000013\"><h3 id=\"org0000013\">TODO more frequent, ideally realtime updates to views<\/h3><div class=\"outline-text-3\" id=\"text-org0000013\"><p>If the API doesn't provide push-based interface (as most of them), ultimately it's a question of polling them carefully to avoid rate limiting penalties.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000014\"><h3 id=\"org0000014\">TODO alternative export formats<\/h3><div class=\"outline-text-3\" id=\"text-org0000014\"><p>There is nothing really about Org-mode that's specific to this system. For instance, there are markdown-based organizers out there and people could benefit from using Orger for them.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000015\"><h3 id=\"org0000015\">TODO two-way data flow<\/h3><div class=\"outline-text-3\" id=\"text-org0000015\"><p>It would be cool to implement feedback from emacs, e.g. editing Github comment when you edit the corresponding Orger item. But it requires considerably more effort and would only work within emacs.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000016\"><h3 id=\"org0000016\">TODO potential for race condition<\/h3><div class=\"outline-text-3\" id=\"text-org0000016\"><p>Unfortunately there is a little space for race condition if Orger appends something while you're editing file. Orger tries to detect emacs and vim swap\/lock files, but it's if you're very unlucky or using different setup it's still possible. Hopefully your text editor warns you when the file had been overwritten while you were editing it (e.g. as emacs does).<\/p><p>Also I run Orger jobs at night (via cron) so it's quite unlikely to overlap with editing anything.<\/p><\/div><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000018\"><h2 id=\"similar\">9 Similar projects<\/h2><div class=\"outline-text-2\" id=\"text-similar\"><ul class=\"org-ul\"><li>\n<p><a href=\"https:\/\/github.com\/novoid\/Memacs\">Memacs<\/a> by Karl Voit.<\/p>\n<p>I only discovered it after I released Orger, so frankly haven't got to try it yet! It looks very similar in terms of goals and seems we can cooperate on rendering parts at least.<\/p>\n<p>One (as I see it) big advantage of my setup is that data providers are abstracted away in <code class=\"inline\">my.<\/code> package, which makes everything more modular and resilient. However Memacs seems to be flexible as well, so it can be used with e.g. <code class=\"inline\">my.<\/code> package as well.<\/p>\n<p>If someone compares Memacs and Orger please let me know, I'd be happy to link it! It would also be more objective than comparison by myself!<\/p>\n<\/li>\n<\/ul><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000019\"><h2 id=\"org0000019\">10 ----<\/h2><div class=\"outline-text-2\" id=\"text-10\"><p>I'd be interesting in hearing your thoughts or feature requests.<\/p><p>This post ended up longer that I expected so in the next part I will tell about more use cases, in particular how I'm using Orger to process Reddit.<\/p><\/div><\/div><\/section><p>Discussion:<\/p><ul><li><a href=\"https:\/\/www.reddit.com\/r\/orgmode\/comments\/da1sf0\/orger_plaintext_reflection_of_your_digital_self\">\/r\/orgmode<\/a><\/li>\n<li><a href=\"https:\/\/lobste.rs\/s\/rr9pxd\/orger_plaintext_reflection_your_digital\">lobste.rs<\/a><\/li>\n<\/ul>","created_at":"2019-12-08T18:07:17+0000","updated_at":"2019-12-08T18:07:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"beepb00p.xyz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1056,"title":"Python: better typed than you think","url":"https:\/\/beepb00p.xyz\/mypy-error-handling.html","content":"<section class=\"post-title\"><p>mypy assisted error handling, exception mechanisms in other languages, fun with pattern matching and type variance<\/p><\/section><section class=\"content\"><div id=\"table-of-contents\"><h2>Table of Contents<\/h2><div id=\"text-table-of-contents\"><ul><li><a href=\"#intro\">1. Intro aka computers are hard<\/a><\/li>\n<li><a href=\"#problem\">2. The problem: parsing Kindle highlights<\/a><\/li>\n<li><a href=\"#logging\">3. A non-solution #1: logging<\/a><\/li>\n<li><a href=\"#error_object\">4. A non-solution #2: special error value<\/a><\/li>\n<li><a href=\"#container\">5. Almost solution #1: Result container<\/a>\n<ul><li><a href=\"#iterator\">By the way, what's up with <code class=\"inline\">Iterator<\/code> everywhere?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#combinators\">6. Almost solution #2: use error combinators<\/a><\/li>\n<li><a href=\"#pair\">7. Still-not-quite-a-solution #3: (Value, Error) pairs<\/a><\/li>\n<li><a href=\"#kiss\">8. Solution: keep it simple<\/a><\/li>\n<li><a href=\"#tips_tricks\">9. Tips &amp; tricks<\/a>\n<ul><li><a href=\"#api\">Custom error type<\/a><\/li>\n<li><a href=\"#unwrap\">unwrap<\/a><\/li>\n<li><a href=\"#global_policy\">Global error policy<\/a><\/li>\n<li><a href=\"#error_context\">Improving error context<\/a><\/li>\n<li><a href=\"#warnings\">Fine grained defensiveness<\/a><\/li>\n<li><a href=\"#dataframe\">Error values, revisited<\/a><\/li>\n<li><a href=\"#cursed_pattern_matching\">Cursed pattern matching mechanism<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#fin\">10. Closing points<\/a><\/li>\n<li><a href=\"#links\">11. Other links<\/a><\/li>\n<li><a href=\"#org0000020\">12. --<\/a><\/li>\n<\/ul><\/div><\/div><p>TLDR: I overview few error handling techniques (with the emphasis on Python, although I mention few other programming languages), some existing Python libraries and suggesting a simple and clean <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/introduction.html\">mypy<\/a>-based approach.<\/p><p>You might learn few things about error handling in different languages, pattern matching, type variance, mypy's capabilities in general and clues for making your code and interfaces more mypy-friendly (and IDE friendly if you're using LSP\/Intellij).<\/p><div class=\"outline-2\" id=\"outline-container-org0000000\"><div class=\"outline-text-2\" id=\"text-intro\"><p>I am somewhat obsessed with personal data and information, analyzing data for quantified self, lifelogging etc. I am trying to integrate all my information sources and make it easy to access and search. You can see some examples in <a href=\"https:\/\/github.com\/karlicoss\/my\">my<\/a> package and Orger: <a href=\"https:\/\/beepb00p.xyz\/orger.html\">part I<\/a>, <a href=\"https:\/\/beepb00p.xyz\/orger-todos.html\">part II<\/a>.<\/p><p>To get this data, manipulate with it and interact with, of course, you need to extract it first (e.g. from json\/csv), parse it (e.g. from plaintext), or even worse, reverse engineer it from vendor locked formats (e.g. in my <a href=\"https:\/\/github.com\/karlicoss\/kobuddy\">kobo parsing library<\/a>).<\/p><p>If you ever worked with data and had to parse some semi-structured data (let alone natural language), or scraped web pages, you might start getting flashbacks now. Undocumented APIs, bad characters, cryptic regexes, corrupt fields, unexpected nulls, logical inconsistencies, all sorts of things. You will almost never get it right from the first few attempts, and then when it finally does what your want\u2026 it breaks after couple of days because of course you missed some edge cases or data provider just gives you utter garbage for no reason. And thing you've spent so much effort on stops working, spams your mailbox and requires attention.<\/p><p>Ew. <strong>Data is messy<\/strong>.<\/p><p>Most modern programming languages are fairly unforgiving to unexpected, and would crash at the slightest opportunity. Some languages do have quirks (e.g. 'undefined' in JS), but generally well written software aborts very soon after something unexpected starts happening. And for good reasons:<\/p><ul class=\"org-ul\"><li>\n<p>if it didn't, your program's state would lose the properties the author intended it to have.<\/p>\n<p>Ignoring the errors will almost surely prevent the program from getting to desired result anyway and end up with even more severe, or potentially catastrophic inconsistencies. How about formatting your disk if you're <em>really<\/em> unlucky?<\/p>\n<\/li>\n<li>another good reason to fail fast is that it makes the programmer more likely to notice and then fix the bug<\/li>\n<\/ul><p>So in most cases, as long as you can get away with it, it's <strong>good to throw exception or abort<\/strong> the program immediately in some way. You might not be able to do that if you're literally doing rocket science or <a href=\"https:\/\/isocpp.org\/wiki\/faq\/exceptions#why-exceptions\">flight control software<\/a>, but most of us aren't. For typical software engineering problems, some errors are less crucial and more manageable than other errors. So we try to be pragmatic when we program, evaluate failure risks and use try\/catch mechanisms where appropriate.<\/p><p>Now, I'm sure we as an engineers we could handwave about about that stuff forever, so let me be more specific straightaway and introduce a motivating <strong>real life problem<\/strong> that I actually had to solve.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000003\"><h2 id=\"problem\">2 The problem: parsing Kindle highlights<\/h2><div class=\"outline-text-2\" id=\"text-problem\"><p>Say, you own a Kindle book. Electronic books are great. Yeah okay they don't smell like the real thing, but the possibility of highlighting bits of text and typing your comment without distracting on external means of annotation is incredibly helpful. However, then when you want to go through your highlights after reading to refresh your memory or perhaps to share with a friend, you find out it's <a href=\"https:\/\/beepb00p.xyz\/annotating.html#kindle\">not so convenient<\/a> to actually quickly access them.<\/p><p>So you decide to write a script that would process the highlights, perhaps group them by book, displays timestamps and render a nice HTML page so you could easily open it from phone and recall latest books you read to discuss with friends.<\/p><p>On device, Kindle keeps bookmarks and highlights are stored \u2026 in <samp class=\"inline\">My Clippings.txt<\/samp> file.<\/p><details><summary>Click to view 'clippings.txt'<\/summary><pre class=\"example\">\nPHYS771 Lecture 12: Proof (scottaaronson.com)\n- Your Highlight on Page 2 | Added on Sunday, July 21, 2013 10:06:53 AM\nRoger Penrose likes to talk about making direct contact with Platonic reality, but it's a bit embarrassing when you think you've made such contact and it turns out the next morning that you were wrong!\n==========\n[Tong][2013] Dynamics and Relativity  \n- Your Highlight on Page 120 | Added on Sunday, August 4, 2013 6:17:21 PM\nIt is worth mentioning that although the two people disagree on whether the light hits the walls at the same time, this does not mean that they can't be friends.\n==========\nPHYS771 Lecture 12: Proof (scottaaronson.com)\n- Your Highlight on Page 14 | Added on Sunday, August 4, 2013 8:41:53 PM\nNo hidden-variable theory can be local (I think some guy named Bell proved that).\n<\/pre><\/details><p>Yes, it's a messy format and not very machine friendly. But oh well it's a file, you're a programmer. You know the drill.<\/p><div class=\"org-src-container\"><pre class=\"src src-python\"> 1: from datetime import datetime\n 2: from typing import NamedTuple, Sequence\n 3: import re\n 4: from pathlib import Path\n 5: from itertools import groupby\n 6: from textwrap import wrap\n 7: \n 8: class Highlight(NamedTuple):\n 9:     dt: datetime # date when highlight was made\n10:     title: str   # book title\n11:     page: str    # highlight location\n12:     text: str    # highlighted text\n13: \n14: class Book(NamedTuple):\n15:     \"Represents book along with its highlights\"\n16:     title: str\n17:     highlights: Sequence[Highlight]\n18: \n19: def parse_entry(entry: str) -&gt; Highlight:\n20:     groups = re.search(\n21:         r'(?P&lt;title&gt;.*)$\\n.*Highlight on Page (?P&lt;page&gt;\\d+).*Added on (?P&lt;dts&gt;.*)$\\n\\n(?P&lt;text&gt;.*)$', \n22:         entry, \n23:         re.MULTILINE,\n24:     )\n25:     assert groups is not None, \"Couldn't match regex!\"\n26:     dt = datetime.strptime(groups['dts'], '%A, %B %d, %Y %I:%M:%S %p')\n27:     return Highlight(\n28:         dt=dt,\n29:         title=groups['title'],\n30:         page=groups['page'],\n31:         text=groups['text'],\n32:     )\n33: \n34: def iter_highlights():\n35:     data = Path(clippings_file).read_text()\n36:     for entry in data.split('=========='):\n37:         yield parse_entry(entry.strip())\n38: \n39: def iter_books():\n40:     key = lambda e: e.title\n41:     for book, hls in groupby(sorted(iter_highlights(), key=key), key=key):\n42:         highlights = list(sorted(hls, key=lambda hl: hl.dt))\n43:         yield Book(title=book, highlights=highlights)\n44: \n45: def print_books():\n46:     for r in iter_books():\n47:         print(f'* {r.title}')\n48:         for h in r.highlights:\n49:             text = \"\\n      \".join(wrap(h.text))\n50:             print(f'  - {h.dt:%d %b %Y %H:%M}  {text} [Page {h.page}]')\n51:         print()\n52: print_books()\n<\/pre><\/div><pre class=\"example\">\n* PHYS771 Lecture 12: Proof (scottaaronson.com)\n  - 21 Jul 2013 10:06  Roger Penrose likes to talk about making direct contact with Platonic\n      reality, but it's a bit embarrassing when you think you've made such\n      contact and it turns out the next morning that you were wrong! [Page 2]\n  - 04 Aug 2013 20:41  No hidden-variable theory can be local (I think some guy named Bell\n      proved that). [Page 14]\n* [Tong][2013] Dynamics and Relativity  \n  - 04 Aug 2013 18:17  It is worth mentioning that although the two people disagree on\n      whether the light hits the walls at the same time, this does not mean\n      that they can't be friends. [Page 120]\n<\/pre><div>For the purposes of this post, to keep the example output clean, I am just using plain text. Even though it's not quite HTML with CSS, it still looks kinda nice, doesn't it?<aside class=\"sidenote\">if you're wondering why <code class=\"inline\">yield<\/code> I'll explain it <a href=\"#iterator\">further down<\/a><\/aside><\/div><p>Now:<\/p><ul class=\"org-ul\"><li>\n<p>imagine you've set this script to run in cron, and it's been fine for a while. You left for a three week holiday to finally get some rest from programming; started reading this new book about quant finance (yeah, you've always had interesting ways of getting a rest from computer) and\u2026 your script stopped working.<\/p>\n<pre class=\"example\">\nTraceback (most recent call last):\n  File \"&lt;stdin&gt;\", line 55, in &lt;module&gt;\n  File \"&lt;stdin&gt;\", line 49, in print_books\n  File \"&lt;stdin&gt;\", line 44, in iter_books\n  File \"&lt;stdin&gt;\", line 34, in iter_highlights\n  File \"&lt;stdin&gt;\", line 21, in parse_entry\nAssertionError: Couldn't match regex!\n<\/pre>\n<p>You swear out loud, reach for the laptop you promised to distance yourself from and turns our your parser chokes over <samp class=\"inline\">page<\/samp> instead of <samp class=\"inline\">Page<\/samp> in one of new entries. (and yes, this was actually the case in my Kindle export)<\/p>\n<details><summary>Click to view updated 'clippings.txt'<\/summary><pre class=\"example\">\nPHYS771 Lecture 12: Proof (scottaaronson.com)\n- Your Highlight on Page 2 | Added on Sunday, July 21, 2013 10:06:53 AM\nRoger Penrose likes to talk about making direct contact with Platonic reality, but it's a bit embarrassing when you think you've made such contact and it turns out the next morning that you were wrong!\n==========\n[Tong][2013] Dynamics and Relativity  \n- Your Highlight on Page 120 | Added on Sunday, August 4, 2013 6:17:21 PM\nIt is worth mentioning that although the two people disagree on whether the light hits the walls at the same time, this does not mean that they can't be friends.\n==========\nPHYS771 Lecture 12: Proof (scottaaronson.com)\n- Your Highlight on Page 14 | Added on Sunday, August 4, 2013 8:41:53 PM\nNo hidden-variable theory can be local (I think some guy named Bell proved that).\n==========\nMy Life as a Quant: Reflections on Physics and Finance (Emanuel Derman)\n- Your Highlight on page 54 | Added on Tuesday, October 4, 2013 12:11:16 PM\nThe Black-Scholes model allows us to determine the fair value of a stock option.\n<\/pre><\/details><p>You could argue that you should have made the regex in <code class=\"inline\">parse_entry<\/code> case independent in the first place, but it's not something you would normally expect. Kindle specifically got all sorts of nasty things: roman numerals for page numbers, locale dependent dates, inconsistent separators, and so on.<\/p>\n<p>Perhaps you even fix this particular problem, but it's a <strong>matter of short time till next parsing issue<\/strong>. It's quite sad if you have to constantly tend for things that are meant to simplify and enhance your life.<\/p>\n<\/li>\n<\/ul><p>Or,<\/p><ul class=\"org-ul\"><li>\n<p>you wrote this parser and decided that it could be useful for other people.<\/p>\n<p>So for a small fee, you are providing a service that fetches highlights from their Kindles, displays on profile pages and lets their friends comment.<\/p>\n<p>Imagine user's highlights result in the same error described above. It would be pretty sad if parsing a single entry took down the whole user's page or prevented updates. No matter how fast you'd be willing to fix these things, <strong>users would leave discouraged<\/strong>.<\/p>\n<\/li>\n<\/ul><p>With the way code is written at the moment, any exception would take the whole program down. So, we need some way of getting around these errors and carrying on.<\/p><p>What do we do?<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000004\"><div class=\"outline-text-2\" id=\"text-logging\"><p>One simple strategy would be to make parsing fully defensive, wrap the whole <code class=\"inline\">parse_entry<\/code> call in <code class=\"inline\">try\/except<\/code> and log:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">33: import logging\n34: def iter_highlights():\n35:     data = Path(clippings_file).read_text()\n36:     for entry in data.split('=========='):\n37:         try:\n38:             yield parse_entry(entry.strip())\n39:         except Exception as e:\n40:             logging.exception(e)\n<\/pre><\/div><p>Logging typically works well for minor things not worthy a proper error (i.e. warnings) and as a means of retrospective error analysis and debugging. In our case logging wouldn't do the job:<\/p><ul class=\"org-ul\"><li>you're not aware that error is happening at all. If it's your personal tool, chances are you don't have time to go through all the logs and inspect them regularly.<\/li>\n<li>user expects to see their data, but can't find it. It's pretty frustrating.<\/li>\n<\/ul><p>What do we want?<\/p><ul class=\"org-ul\"><li>keep track of errors, render as much as we can, but terminate with non-zero exit code<\/li>\n<li>potentially present errors in the interface so you or your users wouldn't worry about lost data<\/li>\n<\/ul><p>So we need some way of propagating the errors up the call hierarchy instead of throwing immediately or suppressing.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000005\"><h2 id=\"error_object\">4 A non-solution #2: special error value<\/h2><div class=\"outline-text-2\" id=\"text-error_object\"><p>Often it's tempting to fallback to some sort of special 'default' or 'error' value. I bet you've seen this before: <code class=\"inline\">0<\/code> or <code class=\"inline\">INT_MAX<\/code> meaning error for integer type, or <code class=\"inline\">\"\"<\/code> for string types. We could try something similar and squeeze exception into the <code class=\"inline\">Highlight<\/code> object itself.<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">33: def iter_highlights():\n34:     data = Path(clippings_file).read_text()\n35:     for entry in data.split('=========='):\n36:         try:\n37:             yield parse_entry(entry.strip())\n38:         except Exception as e:\n39:             yield Highlight(dt=datetime.now(), page='', book=\"ERROR\", text=str(e))\n<\/pre><\/div><p>One obvious problem is that it's very nontransparent and <strong>relies on implicit convention<\/strong>: there is no way of telling that this function might return some special <code class=\"inline\">Highlight<\/code> which should be treated as error. That not only complicates code, but might also introduce logical inconsistencies.<\/p><p>E.g. if your <code class=\"inline\">Highlight<\/code> object also returned book's ISBN and you filled it with some arbitrary text, it would almost surely not be a valid ISBN, that might cause failures down the pipeline.<\/p><p>Sometimes it's inevitable though, e.g. I'm giving an example <a href=\"#dataframe\">later<\/a>.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org000000c\"><h2 id=\"container\">5 Almost solution #1: Result container<\/h2><div class=\"outline-text-2\" id=\"text-container\"><p>An abstraction that stood the test of time well is a container that holds a <strong>result<\/strong> representing one of two:<\/p><ul class=\"org-ul\"><li><strong>success value<\/strong>, representing the <strong>desired outcome<\/strong> of type <code class=\"inline\">T<\/code><\/li>\n<li>or <strong>'error value'<\/strong>, holding <strong>error description<\/strong> of type <code class=\"inline\">E<\/code>.<\/li>\n<\/ul><p>I will try to stick to the same semantics further down, 'result' typically meaning that it could be either desired value or error.<\/p><p>You can vaguely think of it as an interface <code class=\"inline\">Result<\/code>, and two implementations: <code class=\"inline\">Ok<\/code> and <code class=\"inline\">Error<\/code>. In runtime, you can ask the instance behind <code class=\"inline\">Result<\/code>, which of these alternative it holds and act accordingly.<\/p><p>It has manifested as:<\/p><ul class=\"org-ul\"><li>\n<p>in Rust: <a href=\"https:\/\/doc.rust-lang.org\/std\/result\/enum.Result.html\">std::result::Result<\/a>. Example borrowed from <a href=\"https:\/\/doc.rust-lang.org\/1.30.0\/book\/second-edition\/ch09-02-recoverable-errors-with-result.html\">here<\/a>:<\/p>\n<div class=\"org-src-container\"><pre class=\"src src-rust\">let f: Result&lt;File, io::Error&gt; = File::open(\"hello.txt\");\nlet f = match f {\n    Ok(file) =&gt; file,\n    Err(error) =&gt; {\n        panic!(\"There was a problem opening the file: {:?}\", error)\n    },\n};\n<\/pre><\/div>\n<\/li>\n<li>\n<p>in Haskell: <a href=\"https:\/\/wiki.haskell.org\/Handling_errors_in_Haskell#Error_using_the_Either_type\"><code class=\"inline\">Either E T<\/code><\/a><\/p>\n<div class=\"org-src-container\"><pre class=\"src src-haskell\">main = do\n  line &lt;- getLine\n  case runParser emailParser line of\n    Right (user, domain) -&gt; print (\"The email is OK.\", user, domain)\n    Left  (pos, err)     -&gt; putStrLn (\"Parse error on \" &lt;&gt; pos &lt;&gt; \": \" &lt;&gt; err)\n<\/pre><\/div>\n<p>Yes, <code class=\"inline\">Left<\/code> meaning error and <code class=\"inline\">Right<\/code> meaning success are not necessarily obvious. It's kinda a pun: \"right\" also means \"correct\". Also notice that error is not just a string, but also contains the position where parsing failed.<\/p>\n<\/li>\n<li>in C++: there is a proposal for <a href=\"https:\/\/issues.isocpp.org\/show_bug.cgi?id=29\"><code class=\"inline\">std::expected&lt;E, T&gt;<\/code><\/a><\/li>\n<\/ul><p>So, Rust and Haskell programmers seem to be quite happy with it? <strong>Why can't we have same in Python?<\/strong> Well, some people tried! So I'll review a python library that does that: <a href=\"https:\/\/github.com\/dbrgn\/result#result\">result.Result<\/a><\/p><p>Let's try it on our program and see how it works. To make it easier to compare to <a href=\"#org0000002\">the original code<\/a> I suggest duplicating the tab in a separate window and tiling them side by side.<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">33: from result import Ok, Err\n34: def iter_highlights():\n35:     data = Path(clippings_file).read_text()\n36:     for entry in data.split('=========='):\n37:         try:\n38:             yield Ok(parse_entry(entry.strip()))\n39:         except Exception as e:\n40:             yield Err(str(e))\n<\/pre><\/div><p>We've had to wrap success and error values in <code class=\"inline\">Ok<\/code> and <code class=\"inline\">Err<\/code>, but so far it's not too bad.<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">41: from itertools import tee\n42: def iter_books():\n43:     vit, eit = tee(iter_highlights())\n44:     values = (r.value for r in vit if r.is_ok())\n45:     errors = (r.err() for r in eit if r.is_err())\n46:     key = lambda e: e.title\n47:     for book, hls in groupby(sorted(values, key=key), key=key):\n48:         highlights = list(sorted(hls, key=lambda hl: hl.dt))\n49:         yield Ok(Book(title=book, highlights=highlights))\n50:     yield from map(Err, errors)\n<\/pre><\/div><p>We use <a href=\"https:\/\/docs.python.org\/3\/library\/itertools.html#itertools.tee\"><code class=\"inline\">itertools.tee<\/code><\/a> here so we don't have to pollute our code with temporary lists.<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">51: def print_books():\n52:     for r in iter_books():\n53:         if r.is_ok():\n54:             v = r.value\n55:             print(f'* {v.title}')\n56:             for h in v.highlights:\n57:                 text = \"\\n      \".join(wrap(h.text))\n58:                 print(f'  - {h.dt:%d %b %Y %H:%M}  {text} [Page {h.page}]')\n59:             print()\n60:         else:\n61:             e = r.err()\n62:             print(f\"* ERROR: {e}\")\n63: print_books()\n<\/pre><\/div><pre class=\"example\">\n* PHYS771 Lecture 12: Proof (scottaaronson.com)\n  - 21 Jul 2013 10:06  Roger Penrose likes to talk about making direct contact with Platonic\n      reality, but it's a bit embarrassing when you think you've made such\n      contact and it turns out the next morning that you were wrong! [Page 2]\n  - 04 Aug 2013 20:41  No hidden-variable theory can be local (I think some guy named Bell\n      proved that). [Page 14]\n* [Tong][2013] Dynamics and Relativity  \n  - 04 Aug 2013 18:17  It is worth mentioning that although the two people disagree on\n      whether the light hits the walls at the same time, this does not mean\n      that they can't be friends. [Page 120]\n* ERROR: Couldn't match regex!\n<\/pre><p>Cool, we rendered as much as we can, and we get the error displayed as well, so nothing crashes and the users are not as unhappy. The error looks a bit out of nowhere, but at least it's there. We will address how we can improve it <a href=\"#error_context\">later<\/a>.<\/p><p>Sadly, for someone else who looks at <code class=\"inline\">iter_highlights<\/code> or <code class=\"inline\">iter_books<\/code> signatures, it's not obvious that it yields <code class=\"inline\">Result<\/code> objects, not <code class=\"inline\">Book\/Highlight<\/code> objects without reading the code. It's a thankless job for a human to keep track of, and <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/introduction.html\"><strong>mypy<\/strong><\/a> is a perfect fit for this task. Gladly, <code class=\"inline\">result<\/code> library already <a href=\"https:\/\/github.com\/dbrgn\/result\/blob\/master\/result\/result.py\">comes with type annotations<\/a>.<\/p><p>So, let's try to use mypy to aid us at writing correct code.<\/p><p>Let's focus just on <code class=\"inline\">iter_highlights<\/code> and <code class=\"inline\">iter_books<\/code> and use the <code class=\"inline\">Result<\/code> type.<\/p><div class=\"org-src-container\"><pre class=\"src src-python\" id=\"org0000007\">34: from result import Ok, Err, Result\n35: from typing import Iterator\n36: Error = str\n37: \n38: def iter_highlights() -&gt; Iterator[Result[Error, Highlight]]:\n39:     data = Path(clippings_file).read_text()\n40:     for entry in data.split('=========='):\n41:         try:\n42:             yield Ok(parse_entry(entry.strip()))\n43:         except Exception as e:\n44:             yield Err(str(e))\n<\/pre><\/div><div class=\"org-src-container\"><pre class=\"src src-mypy\">45: from itertools import tee\n46: def iter_books() -&gt; Iterator[Result[Error, Book]]:\n47:     vit, eit = tee(iter_highlights())\n48:     values = (r.ok() for r in vit if r.is_ok())\n49:     errors = (r      for r in eit if r.is_err())\n50:     key = lambda e: e.title\n51:     for book, hls in groupby(sorted(values, key=key), key=key):\n52:         highlights = list(sorted(hls, key=lambda hl: hl.dt))\n53:         yield Ok(Book(title=book, highlights=highlights))\n54:     yield from errors\n<\/pre><\/div><pre class=\"example\">\nMypy output [exit code 1]:\ninput.py: note: In function \"iter_books\":\ninput.py:52: error: Item \"None\" of \"Optional[Highlight]\" has no\nattribute \"dt\"  [union-attr]\n            highlights = list(sorted(hls, key=lambda hl: hl.dt))\n                                                         ^\ninput.py:53: error: Argument \"highlights\" to \"Book\" has incompatible\ntype \"List[Optional[Highlight]]\"; expected \"Sequence[Highlight]\"  [arg-type]\n            yield Ok(Book(title=book, highlights=highlights))\n                                                 ^\ninput.py:54: error: Incompatible types in \"yield from\" (actual type\n\"Result[str, Highlight]\", expected type \"Result[str, Book]\")  [misc]\n        yield from errors\n        ^\nFound 3 errors in 1 file (checked 1 source file)\n<\/pre><p>Umm. Let's go through the errors:<\/p><ul class=\"org-ul\"><li>errors 1 and 2 are due to <code class=\"inline\">ok()<\/code> method being too defensive and <a href=\"https:\/\/github.com\/dbrgn\/result\/blob\/0778597ddb737754780b3aca956ad944282ee870\/result\/result.py#L75-L81\">returning <code class=\"inline\">None<\/code><\/a> if <code class=\"inline\">is_ok<\/code> is <code class=\"inline\">False<\/code>. Ideally, you'd throw exception here, because such a situation is a <strong>programming bug<\/strong>. We can just enforce non-optional type here via <code class=\"inline\">unopt<\/code> helper.<\/li>\n<li>error 3 happens because even though we filtered error values, mypy has no idea about that, so it still assumes that <code class=\"inline\">errors<\/code> might hold <code class=\"inline\">Highlight<\/code> objects. You could blame mypy of not being smart enough, but it would be a very <strong>hard if not impossible analysis<\/strong> in general case. We can get around this by unpacking error and wrapping back in <code class=\"inline\">Err<\/code>.<\/li>\n<\/ul><p>Let's apply these insights and try again:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">45: from typing import Optional, TypeVar\n46: X = TypeVar('X')\n47: def unopt(x: Optional[X]) -&gt; X:\n48:     # similar to https:\/\/doc.rust-lang.org\/std\/option\/enum.Option.html#method.unwrap\n49:     assert x is not None\n50:     return x\n51: \n52: from itertools import tee\n53: def iter_books() -&gt; Iterator[Result[Error, Book]]:\n54:     vit, eit = tee(iter_highlights())\n55:     values = (unopt(r.ok())  for r in vit if r.is_ok())\n56:     errors = (unopt(r.err()) for r in eit if r.is_err())\n57:     key = lambda e: e.title\n58:     for book, hls in groupby(sorted(values, key=key), key=key):\n59:         highlights = list(sorted(hls, key=lambda hl: hl.dt))\n60:         yield Ok(Book(title=book, highlights=highlights))\n61:     for err in errors:\n62:         yield Err(err)\n<\/pre><\/div><pre class=\"example\">\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>Phew! With some minor changes and restructuring we've convinced mypy.<\/p><p>It does come with some downsides:<\/p><ul class=\"org-ul\"><li><strong>readability<\/strong>: there is a bit of visual noise since you need to add <code class=\"inline\">Ok\/Err<\/code> wrappers and access the success value via <code class=\"inline\">.value<\/code> property<\/li>\n<li>\n<p><strong>safety<\/strong>: you could forget to call <code class=\"inline\">is_ok\/is_err<\/code> before calling <code class=\"inline\">ok\/err<\/code>, and mypy won't even blink.<\/p>\n<p>The contract <samp class=\"inline\">if .is_ok() is True, then it's safe to call .ok()<\/samp> is too complicated to be encoded as a type that mypy can handle. You'll get <code class=\"inline\">None<\/code> or exception thrown in runtime. The author of the library admits it by the way, so it's not a criticism, just highlighting limitations of mypy here!<\/p>\n<\/li>\n<\/ul><p>Ok, we've learned something, let's try again.<\/p><\/div><div class=\"outline-3\" id=\"outline-container-org000000b\"><h3 id=\"iterator\">By the way, what's up with <a href=\"https:\/\/docs.python.org\/3\/library\/typing.html#typing.Iterator\"><code class=\"inline\">Iterator<\/code><\/a> everywhere?<\/h3><div class=\"outline-text-3\" id=\"text-iterator\"><p>Glad you asked! Several reasons I'm using generators here:<\/p><ul class=\"org-ul\"><li>it makes code cleaner because there is <strong>no need for temporary lists<\/strong>, calling <code class=\"inline\">.append<\/code> and then returning them in the end.<\/li>\n<li>it makes code faster (again, no temporary lists), and also it <strong>feels faster<\/strong> because you print items as soon as you process<\/li>\n<li><code class=\"inline\">Iterator<\/code> type is <strong>covariant<\/strong>, whereas <code class=\"inline\">List<\/code> is not. I'm elaborating on it <a href=\"#org000000a\">later<\/a>. I'm also using <a href=\"https:\/\/docs.python.org\/3\/library\/typing.html#typing.Sequence\"><code class=\"inline\">Sequence<\/code><\/a> for the same reason.<\/li>\n<\/ul><\/div><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000010\"><h2 id=\"combinators\">6 Almost solution #2: use error combinators<\/h2><div class=\"outline-text-2\" id=\"text-combinators\"><p>Now, let's try out <a href=\"https:\/\/github.com\/dry-python\/returns#result-container\">returns.result<\/a> library, clearly inspired by Haskell's <code class=\"inline\">Either<\/code> monad and <code class=\"inline\">do<\/code> notation. I'm quite glad someone already implemented it and I didn't have to reinvent the wheel here.<\/p><p>So, let's try and rewrite the code using <code class=\"inline\">returns.result.Result<\/code>:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\" id=\"org000000d\">19: from returns.result import safe\n20: \n21: @safe\n22: def parse_entry(entry: str) -&gt; Highlight:\n23:     groups = re.search(\n24:         r'(?P&lt;title&gt;.*)$\\n.*Highlight on Page (?P&lt;page&gt;\\d+).*Added on (?P&lt;dts&gt;.*)$\\n\\n(?P&lt;text&gt;.*)$', \n25:         entry, \n26:         re.MULTILINE,\n27:     )\n28:     assert groups is not None, \"Couldn't match regex!\"\n29:     dt = datetime.strptime(groups['dts'], '%A, %B %d, %Y %I:%M:%S %p')\n30:     return Highlight(\n31:         dt=dt,\n32:         title=groups['title'],\n33:         page=groups['page'],\n34:         text=groups['text'],\n35:     )\n36: \n37: from returns.result import Result\n38: from typing import Iterator\n39: def iter_highlights() -&gt; Iterator[Result[Highlight, Exception]]:\n40:     data = Path(clippings_file).read_text()\n41:     for entry in data.split('=========='):\n42:         yield parse_entry(entry.strip())\n<\/pre><\/div><p>So far the only difference from <a href=\"#org0000002\">the original code<\/a> is <a href=\"https:\/\/returns.readthedocs.io\/en\/latest\/pages\/result.html#safe\"><code class=\"inline\">@safe<\/code><\/a> decorator on <code class=\"inline\">parse_entry<\/code>, which basically deals with catching all exceptions and wrapping into <code class=\"inline\">Result<\/code>.<\/p><p>As a consequence, <code class=\"inline\">iter_highlights<\/code> required no changes in its body. (which may not be a desirable thing as we'll see <a href=\"#error_context\">later<\/a>)<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\" id=\"org000000e\">43: from typing import cast\n44: from returns.result import Success, Failure\n45: from itertools import tee\n46: def iter_books() -&gt; Iterator[Result[Book, Exception]]:\n47:     vit, eit = tee(iter_highlights())\n48:     sentinel = cast(Highlight, object())\n49:     values = (r.unwrap()  for r in vit if r.value_or(sentinel) is not sentinel)\n50:     errors = (r.failure() for r in eit if r.value_or(sentinel) is     sentinel)\n51:     key = lambda e: e.title\n52:     for book, hls in groupby(sorted(values, key=key), key=key):\n53:         highlights = list(sorted(hls, key=lambda hl: hl.dt))\n54:         yield Success(Book(title=book, highlights=highlights))\n55:     for e in errors:\n56:         yield Failure(e)\n<\/pre><\/div><p>Ok, that definitely requires some explanation\u2026<\/p><p><code class=\"inline\">returns<\/code> library public API doesn't provide any way to tell between success and failure (<a href=\"https:\/\/returns.readthedocs.io\/en\/latest\/pages\/result.html#what-is-the-difference-between-success-and-success\">kind of deliberately<\/a>). The types <code class=\"inline\">_Success<\/code> and <code class=\"inline\">_Failure<\/code> are private, and the only method that we can use seems to be <a href=\"https:\/\/returns.readthedocs.io\/en\/latest\/pages\/result.html#returns.result.Result.value_or\"><code class=\"inline\">result.value_or(default)<\/code><\/a>. This method returns the success value if <code class=\"inline\">result<\/code> is <code class=\"inline\">Success<\/code> and falls back to <code class=\"inline\">default<\/code> if <code class=\"inline\">result<\/code> is a <code class=\"inline\">Failure<\/code>. So we use a <strong>sentinel object to distinguish<\/strong> between actual success values and <code class=\"inline\">default<\/code> ones, and also have to trick mypy with a <code class=\"inline\">cast<\/code>.<\/p><p>Apart from this obscurity, the function suffers from exactly the same issues as the <code class=\"inline\">iter_books<\/code> implementation from <a href=\"#org0000008\">the previous section<\/a>, and for the same reason: contract is too complicated to be expressed in mypy.<\/p><p>One could argue that this function is going to look awkward anyway since we need to separate list of results into successes and errors. Let's see the function that should be more straightforward:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\" id=\"org000000f\">57: from typing import Callable\n58: def print_books() -&gt; None:\n59:     for r in iter_books():\n60:         def print_ok(r: Book) -&gt; None:\n61:             print(f'* {r.title}')\n62:             for h in r.highlights:\n63:                 text = \"\\n      \".join(wrap(h.text))\n64:                 print(f'  - {h.dt:%d %b %Y %H:%M}  {text} [Page {h.page}]')\n65:         print_error = lambda e: print(f\"* ERROR: {e}\")\n66:         r.map(print_ok).fix(print_error)\n<\/pre><\/div><p>The idea here is that we can use <code class=\"inline\">map<\/code> method (that works like <code class=\"inline\">fmap<\/code> in Haskell) and use it to print successful results, and chain it with <code class=\"inline\">fix<\/code> that works like like <code class=\"inline\">fmap<\/code>, but for errors. In a sense, these methods encapsulate pattern matching (which Python lacks syntactically) so as long the implementor did the dirty business of correctly doing it dynamically, you're safe. However I feel that this particular library overdid this encapsulation a bit, hence very hacky implementation of <code class=\"inline\">iter_books<\/code>.<\/p><p>Lambdas <a href=\"https:\/\/stackoverflow.com\/a\/1233520\/706389\"><strong>can't be multiline<\/strong><\/a>, so we have to define a local function for <code class=\"inline\">print_ok<\/code>.<\/p><p>There is a <a href=\"https:\/\/github.com\/python\/mypy\/issues\/4226\">bug in mypy<\/a> that sometimes prevents you from inlining the lambda and struggles with type inference. Here I'm hitting this bug with <code class=\"inline\">print_error<\/code>, that's why it's not <code class=\"inline\">.fix(lambda e: print(f\"* ERROR: {e}\"))<\/code>.<\/p><p>Another potential problem is one could forget to implement one of <code class=\"inline\">map\/fix<\/code> clauses, since <strong>nothing enforces calling them<\/strong>. Even if you're detecting unused variables, missing <code class=\"inline\">.fix<\/code> clause could stay unnoticed forever. It's very similar to forgetting <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Global_Objects\/Promise\/catch\"><code class=\"inline\">catch<\/code><\/a> when using Javascript Promises.<\/p><p>It might be possible to enforce with some static analysis though, e.g. via mypy plugin by flagging dangling\/temporary <code class=\"inline\">Result<\/code> values (e.g. similarly to <a href=\"https:\/\/doc.rust-lang.org\/std\/result\/#results-must-be-used\"><code class=\"inline\">must_use<\/code><\/a> attribute in Rust), but it's a project on its own.<\/p><p>Well at the very least it works and type checks!<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">67: print_books()\n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 0]:\n* PHYS771 Lecture 12: Proof (scottaaronson.com)\n  - 21 Jul 2013 10:06  Roger Penrose likes to talk about making direct contact with Platonic\n      reality, but it's a bit embarrassing when you think you've made such\n      contact and it turns out the next morning that you were wrong! [Page 2]\n  - 04 Aug 2013 20:41  No hidden-variable theory can be local (I think some guy named Bell\n      proved that). [Page 14]\n* [Tong][2013] Dynamics and Relativity  \n  - 04 Aug 2013 18:17  It is worth mentioning that although the two people disagree on\n      whether the light hits the walls at the same time, this does not mean\n      that they can't be friends. [Page 120]\n* ERROR: Couldn't match regex!\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>Overall I'm not sold, Python simply <strong>lacks syntax<\/strong> that lets you unpack and compose <code class=\"inline\">Result<\/code> objects in a clean way and you end up with boilerplate. <a href=\"https:\/\/returns.readthedocs.io\/en\/latest\/pages\/io.html#returns.io.IO.lift\"><code class=\"inline\">lifts<\/code><\/a> are not very readable in Haskell, let alone in Python.<\/p><p>I think authors did a great experiment though, the more people have fun with types, the more good abstractions we'll find.<\/p><p>I don't want to discourage people from using their library, so if it's your personal project and it makes your code more manageable or it just feels fun then by all means go for it!<\/p><p>But as much as I like ideas from functional programming, I'm almost certain that it's gonna look confusing to an average Python programmer, and won't be welcome warmly in your team.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000011\"><h2 id=\"pair\">7 Still-not-quite-a-solution #3: (Value, Error) pairs<\/h2><div class=\"outline-text-2\" id=\"text-pair\"><p>Before we go on to the solution I propose let me mention another notable pattern of error handling.<\/p><p>It's commonly used in <a href=\"https:\/\/blog.golang.org\/error-handling-and-go\">Go<\/a>.<\/p><div class=\"org-src-container\"><pre class=\"src src-go\">f, err := os.Open(\"filename.ext\")\nif err != nil {\n    log.Fatal(err)\n}\n\/\/ do something with the open *File f\n<\/pre><\/div><p>However, it's not limited only by Go, e.g. you'd often encounter it implicitly in C (which had no exceptions) or C++ code. For instance, <a href=\"https:\/\/en.cppreference.com\/w\/cpp\/filesystem\/is_symlink\"><code class=\"inline\">std::filesystem::is_symlink<\/code><\/a> comes in two flavours:<\/p><ul class=\"org-ul\"><li><code class=\"inline\">bool is_symlink( const std::filesystem::path&amp; p )<\/code>, which throws exceptions on errors.<\/li>\n<li>\n<p><code class=\"inline\">bool is_symlink( const std::filesystem::path&amp; p, std::error_code&amp; ec ) noexcept<\/code>, which sets <code class=\"inline\">ec<\/code> on errors.<\/p>\n<p>You can think of it as if it returned <code class=\"inline\">std::tuple&lt;bool, std::error_code&gt;<\/code>. I assume it's not that way because the compiler wouldn't be able to distinguish between signatures.<\/p>\n<\/li>\n<\/ul><p>Personally I as well as many other people find it pretty ugly. No judgment here though as I have no idea behind the design requirements and rationale for such a model in Go. Pretty sure one can get used to it after a while and that there are some static flow analyzers that help to ensure correct error handling.<\/p><p>Main issue with this approach regarding Python is that it's not mypy friendly as return type of <code class=\"inline\">Open<\/code> would have to be <code class=\"inline\">Tuple[Optional[Success], Optional[Error]]<\/code>. In the type theory language, it is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Product_type\">product type<\/a>, so in addition to all members of <code class=\"inline\">Success<\/code> type and all members of <code class=\"inline\">Error<\/code> type, it also got inhabitants that <strong>don't make sense for our program<\/strong>, such as <code class=\"inline\">(None, None)<\/code> and also all of <code class=\"inline\">Tuple[Success, Error]<\/code>.<\/p><p>In other words, nothing on type level prevents the callee (<code class=\"inline\">os.Open<\/code>) from returning something like <code class=\"inline\">(file_descriptor, \"whoops\")<\/code>, which has ambiguous meaning. If we use it we would have to pay with sacrificing type safety or <strong>extra code on caller site<\/strong> to eliminate these impossible program states:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">f, err = open('filename.ext')\nif err is None:\n    assert f is not None\n    # ok, now we have both mypy and runtime safety: open returned error\nelif:\n    assert f is None\n    # ok, now we have both mypy and runtime safety: open returned value\n<\/pre><\/div><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000012\"><h2 id=\"kiss\">8 Solution: keep it simple<\/h2><div class=\"outline-text-2\" id=\"text-kiss\"><p>It seems that we were on the right track with the <a href=\"#container\">container type<\/a> and <a href=\"#combinators\">combinators<\/a>, but never completely satisfied. Let's recall the problems we had again:<\/p><ul class=\"org-ul\"><li>\n<p><strong>readability<\/strong>: extra wrappers and accessor methods like <code class=\"inline\">Ok\/Success\/Error\/.is_ok()\/.unwrap()<\/code>.<\/p>\n<p>It's <strong>visual noise<\/strong> and also they creep throughout the code, so if you decide you won't need them later, you might have to refactor a lot of code.<\/p>\n<\/li>\n<\/ul><ul class=\"org-ul\"><li><strong>safety<\/strong>: it's still possible to write logically inconsistent code like <code class=\"inline\">if res.is_error(): return res.value * 10<\/code>.<\/li>\n<li><strong>composability<\/strong>: <code class=\"inline\">fmap<\/code>-style combinators are not really going to look good because Python lacks multiline lambdas.<\/li>\n<li><strong>performance and memory use<\/strong>: not going to make claims here as I haven't benchmarked, but there is a potential for overhead caused by extra wrapper objects.<\/li>\n<\/ul><p>First, we'll attack readability and safety. Yes, at the same time!<\/p><p>In part it's solved with syntactic sugar in other languages like <code class=\"inline\">do<\/code> syntax in <code class=\"inline\">Haskell<\/code>, or <code class=\"inline\">try!<\/code> macro and <code class=\"inline\">?<\/code> operator in <a href=\"https:\/\/doc.rust-lang.org\/edition-guide\/rust-2018\/error-handling-and-panics\/the-question-mark-operator-for-easier-error-handling.html\">Rust<\/a>. Sometimes it's inevitable and you have to inject values into rust's <code class=\"inline\">Result<\/code> explicitly via <code class=\"inline\">Ok\/Err<\/code> constructors. However checking for <code class=\"inline\">.is_ok()<\/code> or <code class=\"inline\">isRight<\/code> is really not that common in idiomatic Rust and Haskell. Reason is <strong>pattern matching<\/strong>! E.g. if we had pattern matching in Python we could write something like:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">def print_books():\n    for r in iter_books():\n        match r:\n            Book b:\n                print(f'* {b.title}')\n                for h in b.highlights:\n                    text = \"\\n      \".join(wrap(h.text))\n                    print(f'  - {h.dt:%d %b %Y %H:%M}  {text} [Page {h.page}]')\n                print()\n            Error e:\n                print(f\"* ERROR: {e}\")\n<\/pre><\/div><p>That's cleaner than checking for <code class=\"inline\">is_ok\/is_err<\/code> and unpacking; and also makes it type safe because <code class=\"inline\">b<\/code> and <code class=\"inline\">e<\/code> already have the appropriate types. In our imaginary world where python had this syntax, surely mypy would have supported it too, right?<\/p><p>Oh wait. It kind of <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/kinds_of_types.html#union-types\">supports it already<\/a>!<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">from typing import Union\ndef f(x: Union[int, str]) -&gt; None:\n    x + 1     # Error: str + int is not valid\n    if isinstance(x, int):\n        # Here type of x is int.\n        x + 1      # OK\n    else:\n        # Here type of x is str.\n        x + 'a'    # OK\n<\/pre><\/div><p>So, mypy keeps track of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Typing_environment\">typing context<\/a> and <strong>narrows it down<\/strong> after certain operations, in particular, <code class=\"inline\">isinstance<\/code> checks and <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/kinds_of_types.html#optional-types-and-the-none-type\"><code class=\"inline\">is None\/is not None<\/code> checks<\/a>.<\/p><p>That looks <strong>very similar to pattern matching<\/strong> both in terms of syntax and typing rules.<\/p><p>So, it seems that <code class=\"inline\">Union<\/code> would represent our result type. Do we still need to come up with some special wrapper for errors? Not really, Python already has a fairly convenient candidate for it: <code class=\"inline\">Exception<\/code>! Most often you have it anyway in <code class=\"inline\">except<\/code> clause, if it's not enough, you can inherit it, add extra fields and treat as any other type.<\/p><p>On the other hand, Exceptions almost never end up as function return values (and when they do, it's normally some fairly unambiguous code dealing specifically with error handling). Hmm, how convenient \ud83e\udd14.<\/p><p>So even though we don't have explicit tagged unions in Python, if we agree that error values are represented as Exceptions, then we do get a disjoint type (i.e. <code class=\"inline\">Ok<\/code> and <code class=\"inline\">Error<\/code> are mutually exclusive) at runtime.<\/p><p>So, rules of thumb:<\/p><ul class=\"org-ul\"><li>use <code class=\"inline\">Union[T, Exception]<\/code> to represent type for results that hold <code class=\"inline\">T<\/code> but can also end up with an error<\/li>\n<li><code class=\"inline\">return<\/code> or <code class=\"inline\">yield<\/code> exceptions and success values without using any extra wrappers<\/li>\n<li>'pattern match' through <code class=\"inline\">isinstance<\/code><\/li>\n<\/ul><p>Let's see how we can rewrite our program by employing these principles:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">33: from typing import TypeVar, Union\n34: T = TypeVar('T')\n35: Res = Union[T, Exception]\n36: \n37: from typing import Iterator\n38: \n39: def iter_highlights() -&gt; Iterator[Res[Highlight]]:\n40:     data = Path(clippings_file).read_text()\n41:     for entry in data.split('=========='):\n42:         try:\n43:             yield parse_entry(entry.strip())\n44:         except Exception as e:\n45:             yield e\n46: \n47: from itertools import tee\n48: \n49: def iter_books() -&gt; Iterator[Res[Book]]:\n50:     vit, eit = tee(iter_highlights())\n51:     values = (r for r in vit if not isinstance(r, Exception))\n52:     errors = (r for r in eit if     isinstance(r, Exception))\n53:     key = lambda e: e.title\n54:     for book, hls in groupby(sorted(values, key=key), key=key):\n55:         highlights = list(sorted(hls, key=lambda hl: hl.dt))\n56:         yield Book(title=book, highlights=highlights)\n57:     yield from errors\n58: \n59: def print_books() -&gt; None:\n60:     for r in iter_books():\n61:         if not isinstance(r, Exception):\n62:             print(f'* {r.title}')\n63:             for h in r.highlights:\n64:                 text = \"\\n      \".join(wrap(h.text))\n65:                 print(f'  - {h.dt:%d %b %Y %H:%M}  {text} [Page {h.page}]')\n66:             print()\n67:         else:\n68:             print(f\"* ERROR: {r}\")\n69: print_books()\n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 0]:\n* PHYS771 Lecture 12: Proof (scottaaronson.com)\n  - 21 Jul 2013 10:06  Roger Penrose likes to talk about making direct contact with Platonic\n      reality, but it's a bit embarrassing when you think you've made such\n      contact and it turns out the next morning that you were wrong! [Page 2]\n  - 04 Aug 2013 20:41  No hidden-variable theory can be local (I think some guy named Bell\n      proved that). [Page 14]\n* [Tong][2013] Dynamics and Relativity  \n  - 04 Aug 2013 18:17  It is worth mentioning that although the two people disagree on\n      whether the light hits the walls at the same time, this does not mean\n      that they can't be friends. [Page 120]\n* ERROR: Couldn't match regex!\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>Yay, it works and typechecks. Now you can decide for yourself how clean it is by comparing it side by side with <a href=\"#org0000002\">the original code<\/a> without error handling. You'd see that the only differences (apart from indentation) is code for error handling.<\/p><p>Here's what I like about this approach:<\/p><ul class=\"org-ul\"><li>\n<p>no extra wrapper classes, code is <strong>clean and readable<\/strong><\/p>\n<p>Also note that surprisingly, Python's dynamic nature actually helps here. E.g. if you rewrote <code class=\"inline\">iter_books<\/code> in Rust, you'd have to use <code class=\"inline\">Ok<\/code> and <code class=\"inline\">Err<\/code> to wrap the return values into <code class=\"inline\">Res<\/code> object. I can imagine you might be able get away with explicit wrapping if you use language with conversions like <code class=\"inline\">Scala<\/code> or <code class=\"inline\">C++<\/code>.<\/p>\n<\/li>\n<li>\n<p>because of no runtime wrappers, <strong>on the 'successful' code path, the callee doesn't need extra code<\/strong> to wrap\/unwrap anything.<\/p>\n<p>You can prototype and mess with your program in the interpreter without having to think about errors. If you do get an error, it would just most likely crash the whole program with <code class=\"inline\">AttributeError<\/code>, which is essentially the desired non-defensive behaviour during prototyping.<\/p>\n<p>You can completely ignore mypy and error handling, until you're happy, then you harden your program by making sure it complies to mypy.<\/p>\n<\/li>\n<li>\n<p><strong>no memory overhead<\/strong> caused by constant wrapping and unwrapping.<\/p>\n<p>I don't really want to make claims about CPU here. I tried isolated micro benchmarking; using <code class=\"inline\">isinstance(r, Exception)<\/code> runs in 50ns, using <code class=\"inline\">is_err()<\/code> call and then unpacking <code class=\"inline\">err()<\/code> runs is 60ns. But these numbers might not make sense under a realistic data flow.<\/p>\n<\/li>\n<li>\n<p><strong>easy to operate and transform<\/strong> values, you just write regular Python code without extra lambdas or kludgy local functions.<\/p>\n<p>If you don't need to handle the error, you can just yield it up the call stack as we do in <code class=\"inline\">iter_books<\/code>.<\/p>\n<\/li>\n<li><strong>doesn't require modifying existing types<\/strong>, and introducing invalid states that signal errors (mentioned <a href=\"#error_object\">here<\/a>)<\/li>\n<li>\n<p><strong>correct variance<\/strong> for free<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Covariance_and_contravariance_(computer_science)\">Variance<\/a> reflects how compound types (e.g. containers\/functions) behave with respect to inheritance of their arguments and return types. You might have also heard of this as Liskov substitution principle. I wouldn't try to explain it here, as it's a topic that deserves a whole post and something you need to experiment with and get comfortable. You can also find some explanations and examples <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/generics.html#variance-of-generic-types\">here<\/a>.<\/p>\n<p>It short, we can let <code class=\"inline\">Res[T]<\/code> to be covariant with respect to <code class=\"inline\">T<\/code>, because it's a simple immutable wrapper around <code class=\"inline\">T<\/code>.<\/p>\n<p>If you were defining your own generic class, you'd have to declare <code class=\"inline\">T = TypeVar('T', covariant=True)<\/code>. It's somewhat misleading, because <strong>variance is a property of a generic container<\/strong>, however for some historic reasons in mypy, you specify variance in the definition of type variable. However, because <code class=\"inline\">Res<\/code> is merely an alias to <code class=\"inline\">Union<\/code>, you <a href=\"https:\/\/mypy.readthedocs.io\/en\/latest\/generics.html#generic-type-aliases\">don't have to remember to do it<\/a>, because <code class=\"inline\">Union<\/code> is already defined as covariant in both its type arguments.<\/p>\n<\/li>\n<\/ul><p>Downsides:<\/p><ul class=\"org-ul\"><li>\n<p><code class=\"inline\">isinstance<\/code> looks a bit verbose and might be frowned upon as it's <strong>often considered as code smell<\/strong><\/p>\n<p>We can't get around this and hide in a helper function for the same reason mentioned <a href=\"#org0000009\">above<\/a>, but it might be solved in mypy in <a href=\"https:\/\/github.com\/python\/mypy\/issues\/5206\">some near future<\/a>, though.<\/p>\n<\/li>\n<\/ul><p>That's basically what I wanted to show! I've been using this pattern for a while now and I think it could work well. Remember about typing contexts and how <code class=\"inline\">isinstance<\/code> \/ <code class=\"inline\">is None<\/code> checks impact it, and you can keep your code clean and safe.<\/p><p>Not suggesting you to go and rewrite all your code from using <code class=\"inline\">try\/catch<\/code> now though. Every error handling style has its place, and hopefully you'll figure out parts of your projects where it's applicable.<\/p><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org000001d\"><h2 id=\"tips_tricks\">9 Tips &amp; tricks<\/h2><div class=\"outline-3\" id=\"outline-container-org0000014\"><h3 id=\"api\">Custom error type<\/h3><div class=\"outline-text-3\" id=\"text-api\"><p>While <a class=\"coderef\" href=\"#coderef-simple_error_api\">the three line API<\/a> is enough in most cases, you might want something more fancy.<\/p><p>One improvement is allowing <strong>arbitrary error type<\/strong>.<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\" id=\"org0000013\">1: from typing import TypeVar, Union\n2: T = TypeVar('T')\n3: E = TypeVar('E')\n4: ResT = Union[T, E]\n<\/pre><\/div><div class=\"org-src-container\"><pre class=\"src src-mypy\"> 5: from typing import NamedTuple, Iterator\n 6: class Error(NamedTuple):\n 7:     text: str\n 8: \n 9: Res = ResT[T, Error]\n10: def iter_numbers() -&gt; Iterator[Res[int]]:\n11:     for s in ['1', 'two', '3', '4']:\n12:         try:\n13:             yield int(s)\n14:         except Exception as e:\n15:             yield Error(str(e))\n16: \n17: def print_negated() -&gt; None:\n18:     for n in iter_numbers():\n19:         if not isinstance(n, Error):\n20:             print(-n)\n21:         else:\n22:             print('ERROR! ' + n.text)\n23: \n24: print_negated()\n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 0]:\n-1\nERROR! invalid literal for int() with base 10: 'two'\n-3\n-4\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>The downside now is that you do need to wrap your exception (i.e. presumably you still want to keep the message and stacktrace) in <code class=\"inline\">Error<\/code> container.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000015\"><h3 id=\"unwrap\">unwrap<\/h3><div class=\"outline-text-3\" id=\"text-unwrap\"><p>Sometimes it's desirable to quickly switch result back to non-defensive version. You can do it by using a simple helper function <code class=\"inline\">unwrap<\/code> (naming inspired by <a href=\"https:\/\/doc.rust-lang.org\/std\/result\/enum.Result.html#method.unwrap\">rust<\/a>):<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">from typing import Union, TypeVar\nT = TypeVar('T', covariant=True)\nRes = Union[T, Exception]\ndef unwrap(res: Res[T]) -&gt; T:\n    if isinstance(res, Exception):\n        raise res\n    else:\n        return res\ngood: Res[int] = 123\nbad: Res[int] = RuntimeError('bad')\nprint(unwrap(good))\nprint(unwrap(bad))\n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 1]:\n123\nTraceback (most recent call last):\n  File \"input.py\", line 13, in &lt;module&gt;\n    print(unwrap(bad))\n  File \"input.py\", line 6, in unwrap\n    raise res\nRuntimeError: bad\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000018\"><h3 id=\"global_policy\">Global error policy<\/h3><div class=\"outline-text-3\" id=\"text-global_policy\"><p>When you're actively working on your code and running tests, you want to make sure that there are no errors and be as non-defensive as possible. However, in the field, you want to keep the code more defensive. To switch behaviours quickly, you can use the following trick:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\" id=\"org0000016\"> 5: from typing import Generic\n 6: X = TypeVar('X', bound=Exception, covariant=True)\n 7: \n 8: class Error(Generic[X]):\n 9:     defensive_policy = True\n10: \n11:     def __init__(self, exc: X) -&gt; None:\n12:         self.exc = exc\n13:         if not Error.defensive_policy:\n14:             raise exc\n15: \n16: Res = ResT[T, Error[Exception]]\n<\/pre><\/div><p>The idea here is <code class=\"inline\">Error.defensive_policy<\/code> determines if exception will be handled defensively or thrown straightaway. <strong>This is enforced on type level<\/strong>, because in order to get <code class=\"inline\">Error<\/code> you need to call its constructor at some point.<\/p><p>Also note the use of <code class=\"inline\">bound=Exception<\/code> on the type variable, this is because we can only <code class=\"inline\">raise<\/code> something that inherits <code class=\"inline\">Exception<\/code>.<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\" id=\"org0000017\">17: \n18: from typing import Iterator\n19: def iter_numbers() -&gt; Iterator[Res[int]]:\n20:     for s in ['1', 'two', '3', '4']:\n21:         try:\n22:             yield int(s)\n23:         except Exception as e:\n24:             yield Error(e)\n25: \n26: def print_negated() -&gt; None:\n27:     for n in iter_numbers():\n28:         if not isinstance(n, Error):\n29:             print(-n)\n30:         else:\n31:             print('ERROR! ' + str(n.exc))\n<\/pre><\/div><p>Now, the default behavior is defensive:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">32: print_negated() \n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 0]:\n-1\nERROR! invalid literal for int() with base 10: 'two'\n-3\n-4\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>And if we set the error policy to non-defensive, we get exception as soon as we get parsing error:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">33: Error.defensive_policy = False\n34: print_negated()\n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 1]:\n-1\nTraceback (most recent call last):\n  File \"input.py\", line 33, in &lt;module&gt;\n    print_negated()\n  File \"input.py\", line 27, in print_negated\n    for n in iter_numbers():\n  File \"input.py\", line 24, in iter_numbers\n    yield Error(e)\n  File \"input.py\", line 14, in __init__\n    raise exc\n  File \"input.py\", line 22, in iter_numbers\n    yield int(s)\nValueError: invalid literal for int() with base 10: 'two'\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>Even though you never actually return <code class=\"inline\">Error<\/code> under the non-defensive policy, you <strong>don't have to change any of the type signatures<\/strong>: <code class=\"inline\">Iterator[int]<\/code> is still a perfectly good <code class=\"inline\">Iterator[Res[int]]<\/code>. Thanks, <a href=\"#org000000a\">covariance<\/a>!<\/p><p>I'm using this technique in my <a href=\"https:\/\/github.com\/karlicoss\/kobuddy#as-a-standalone-app\">Kobo parser<\/a> and control it via <samp class=\"inline\">--errors<\/samp> argument. On CI, it runs in non-defensive mode of course. However when other people use the library for the first time they, something is likely to fail. It deals with decoding binary blobs in unspecified format after all! So one can run it in defensive mode, get most of their data and just ignore (hopefully few) errors till they are fixed.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org0000019\"><h3 id=\"error_context\">Improving error context<\/h3><div class=\"outline-text-3\" id=\"text-error_context\"><p>If you remember <a href=\"#org0000006\">the output<\/a>, we got a rather cryptic <samp class=\"inline\">ERROR: Couldn't match regex!<\/samp>. That's of course not desirable because you can't easily tell what exactly is causing the error.<\/p><p>Normally, you'd use <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-3134\/#explicit-exception-chaining\">exception chaining<\/a>, i.e. <code class=\"inline\">raise EXCEPTION from CAUSE<\/code> syntax <a class=\"coderef\" href=\"#coderef-throw_exc\">here<\/a>.<\/p><div>However the problem is that<code class=\"inline\">raise ... from ...<\/code>is a compound statement, so you can't write<code class=\"inline\">yield RuntimeError(entry) from e<\/code>.<aside class=\"sidenote\">see my investigation attempt <a href=\"https:\/\/stackoverflow.com\/questions\/58355639\/why-raise-exception-from-cause-not-raise-exception-from-cause\">here<\/a><\/aside><\/div><p>I find it handy to have a helper function here:<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">from typing import TypeVar\nE = TypeVar('E', bound=Exception)\ndef echain(e: E, from_: Exception) -&gt; E:\n    e.__cause__ = from_\n    return e\n<\/pre><\/div><p>, then you can write <code class=\"inline\">yield echain(RuntimeError(entry), from_=e)<\/code>, and use <code class=\"inline\">traceback.format_exception<\/code> to unroll it and get the stacktrace. The result looks like this:<\/p><pre class=\"example\">\n* ERROR: Traceback (most recent call last):\n        File \"\/tmp\/tmp.afhyiITIK2\", line 45, in iter_highlights\n          yield parse_entry(entry.strip())\n        File \"\/tmp\/tmp.afhyiITIK2\", line 26, in parse_entry\n          assert groups is not None, \"Couldn't match regex!\"\n      AssertionError: Couldn't match regex!\n      The above exception was the direct cause of the following exception:\n      RuntimeError: \n      My Life as a Quant: Reflections on Physics and Finance (Emanuel Derman)\n      - Your Highlight on page 54 | Added on Tuesday, October 4, 2013 12:11:16 PM\n      The Black-Scholes model allows us to determine the fair value of a stock option.\n<\/pre><p>Now that's better!<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000001a\"><h3 id=\"warnings\">Fine grained defensiveness<\/h3><div class=\"outline-text-3\" id=\"text-warnings\"><p>Remember <a href=\"#org0000001\"><code class=\"inline\">parse_entry<\/code><\/a>? Its return type is <code class=\"inline\">Highlight<\/code>, so it can return a single highlight or throw a single error, that will be handled by <code class=\"inline\">iter_highlights<\/code>.<\/p><p>If we change return type to <code class=\"inline\">Iterator[Res[Highlight]]<\/code>, we can be more defensive and do some <strong>neat fallbacks<\/strong>:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">def parse_entry(entry: str) -&gt; Iterator[Res[Highlight]]:\n    groups = re.search(\n        r'(?P&lt;title&gt;.*)$\\n.*Highlight on Page (?P&lt;page&gt;\\d+).*Added on (?P&lt;dts&gt;.*)$\\n\\n(?P&lt;text&gt;.*)$', \n        entry, \n        re.MULTILINE,\n    )\n    assert groups is not None, \"Couldn't match regex!\"\n    dts   = groups['dts']\n    title = groups['title']\n    page  = groups['page']\n    text  = groups['text']\n    if len(dts) == 0:\n        yield Exception(\"Bad timestamp!\")\n        dt = datetime.now() # might be better than no highlight at all\n    else:\n        dt = datetime.strptime(dts, '%A, %B %d, %Y %I:%M:%S %p')\n    if len(text) == 0:\n        yield Exception(\"Empty highlight, something might be wrong\")\n    yield Highlight(\n        dt=dt,\n        title=title,\n        page=page,\n        text=text,\n    )\n<\/pre><\/div><p>You can think of <code class=\"inline\">Exceptions<\/code> coming from <code class=\"inline\">parse_entry<\/code> as sort of warnings and you can handle them accordingly in <code class=\"inline\">iter_highlights<\/code>, e.g. attach extra context.<\/p><p>Of course, this complicates code, and you can't predict all possible errors anyway, so there is always some balance of how defensive you can be.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000001b\"><h3 id=\"dataframe\">Error values, revisited<\/h3><div class=\"outline-text-3\" id=\"text-dataframe\"><p>One case where I find <a href=\"#error_object\">'special error value'<\/a> more or less appropriate is when your function returns a pandas <code class=\"inline\">DataFrame<\/code>.<\/p><p>When manipulating dataframes, you typically don't iterate explicitly, but apply more idiomatic (and often efficient!) combinators like <code class=\"inline\">merge<\/code>, <code class=\"inline\">join<\/code>, <code class=\"inline\">concat<\/code> etc, so it makes sense to try and keep errors inside the dataframe. For me, it looks somewhat like this:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">def iter_workout_data() -&gt; Iterable[ResT[Exercise, ParsingException]]:\n    ...\ndef rows() -&gt; Iterable[Dict]:\n    for r in iter_workout_data():\n        if isinstance(r, ParsingException):\n            yield {\n                'timestamp': r.timestamp,\n                'error'    : 'parsing failed',\n            }\n        else: # otherwise it's an instance of Exercise\n            yield {\n                'timestamp': r.timestamp,\n                'exercise' : r.exercise_name,\n                'volume'   : r.exercise_volume,\n            }\ndef make_dataframe() -&gt; pandas.DataFrame:\n    return pandas.DataFrame(rows())\n<\/pre><\/div><p>It looks pretty clean since <code class=\"inline\">DataFrame<\/code> constructor automatically creates the necessary columns and fills missing values with <code class=\"inline\">None<\/code>. (you can see some frame examples <a href=\"https:\/\/beepb00p.xyz\/heartbeats_vs_kcals.html\">here<\/a>).<\/p><p>Then in the dataframe processing code I would typically check for presence of non-nil value in 'error' column and act accordingly. E.g. <a href=\"https:\/\/gateway.serph.network\/ipfs\/Qmc5hBuomCLzb25kg14dAesAcnRSJZHPvFSUifZUJexKRx\">here<\/a> I'm using the timestamp attached to the parsing errors to plot them neatly close to the rest of data.<\/p><\/div><\/div><div class=\"outline-3\" id=\"outline-container-org000001c\"><h3 id=\"cursed_pattern_matching\">Cursed pattern matching mechanism<\/h3><div class=\"outline-text-3\" id=\"text-cursed_pattern_matching\"><p>This is forbidden knowledge liberated during the latest Area 51 raid. Tsss\u2026 don't tell the government.<\/p><p>Have to admit, this is a pretty weird idea that I haven't got practical use for, but still.<\/p><p>What's a construction in Python language that's dispatching objects according to their type? <code class=\"inline\">try\/catch<\/code>!<\/p><div class=\"org-src-container\"><pre class=\"src src-mypy\">class A(Exception): pass\nclass B(Exception): pass\nclass C(Exception): pass\nfrom typing import Any\ndef dispatch(x: Any) -&gt; None:\n    try:\n        raise x\n    except A as e:\n        print(\"Matched A!\")\n    except B as e:\n        print(\"Matched B!\")\n    except Exception as e:\n        print(f\"Unhandled object: {type(e)} {e}\")\ndispatch(B())\ndispatch(C())\ndispatch(A())\n<\/pre><\/div><pre class=\"example\">\nPython output [exit code 0]:\nMatched B!\nUnhandled object: &lt;class '__main__.C'&gt; \nMatched A!\nMypy output [exit code 0]:\nSuccess: no issues found in 1 source file\n<\/pre><p>It certainly looks unconventional, and you can only use that as long as your object inherits from <code class=\"inline\">Exception<\/code>.<\/p><p>We can exploit this for our specific case on <code class=\"inline\">Union[T, Exception]<\/code> by using <a href=\"#unwrap\"><code class=\"inline\">unwrap<\/code><\/a>:<\/p><div class=\"org-src-container\"><pre class=\"src src-python\">def print_books():\n    for r in iter_books():\n        try:\n            b = unwrap(r)\n        except Exception as e:\n            # e has type Exception (duh!)\n            print(f\"* ERROR: {e}\")\n        else:\n            # b has type Highlight!\n            print(f'* {b.title}')\n            for h in b.highlights:\n                text = \"\\n      \".join(wrap(h.text))\n                print(f'  - {h.dt:%d %b %Y %H:%M}  {text} [Page {h.page}]')\n            print()\n<\/pre><\/div><p>This looks a bit odd. We still have to type <code class=\"inline\">Exception<\/code>, you can't just write <code class=\"inline\">except e<\/code>, which hardly makes it different from <code class=\"inline\">isinstance<\/code>. Note that we have to use <code class=\"inline\">else<\/code> block: if you put code in it under <code class=\"inline\">try<\/code>, you'll start catching exceptions coming from the printing code, which is unintended.<\/p><p>And the obvious downside is that there is a potential to forget to handle exception signaled by <code class=\"inline\">unwrap<\/code> and mypy can't help you here.<\/p><\/div><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org000001e\"><h2 id=\"fin\">10 Closing points<\/h2><div class=\"outline-text-2\" id=\"text-fin\"><ul class=\"org-ul\"><li>mypy is your best friend<\/li>\n<li>\n<p>sometimes existing and simple things work better and cleaner<\/p>\n<p>Not trying to advocate avoiding syntactic sugar, decorators and libraries at any cost, however you might experience friction while trying to introduce them in more conservative teams.<\/p>\n<\/li>\n<li>\n<p>it's kind of ironic that you can't achieve similar level of safety and cleanliness in many statically typed programming languages<\/p>\n<p>Python is often hated by static typing advocates (I suppose as any other dynamically typed language). Have to admit, I was one of these haters few years ago. But in this case Python nails it.<\/p>\n<\/li>\n<li>\n<p>writing is damn hard<\/p>\n<p>Literate programming is even harder, however I'm glad I've started doing this in Emacs and Org mode. That saved me from otherwise massive amounts of code duplication and reference rot.<\/p>\n<\/li>\n<\/ul><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org000001f\"><div class=\"outline-text-2\" id=\"text-links\"><ul class=\"org-ul\"><li>A good overview of different approaches to error handling: <a href=\"http:\/\/joeduffyblog.com\/2016\/02\/07\/the-error-model\">Joe Duffy - The Error Model<\/a><\/li>\n<li><a href=\"http:\/\/www.open-std.org\/jtc1\/sc22\/wg21\/docs\/papers\/2019\/p0709r4.pdf\">Zero-overhead deterministic exceptions: Throwing values<\/a> by Herb Sutter<\/li>\n<li><a href=\"https:\/\/barahilia.github.io\/blog\/computers\/2017\/01\/04\/to-yield-or-not-to-yield.html\">To yield or not to yield<\/a>: good summary of Python's generator's strengths<\/li>\n<\/ul><\/div><\/div><div class=\"outline-2\" id=\"outline-container-org0000020\"><h2 id=\"org0000020\">12 --<\/h2><div class=\"outline-text-2\" id=\"text-12\"><p>Let me know what you think! I'm open to all feedback.<\/p><\/div><\/div><\/section><p>Discussion:<\/p><ul><li><a href=\"https:\/\/lobste.rs\/s\/xxzwij\/python_is_better_typed_than_you_think\">lobste.rs<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/Python\/comments\/dplkz6\/python_safer_typed_than_you_think\">reddit<\/a><\/li>\n<li><a href=\"https:\/\/news.ycombinator.com\/item?id=21397150\">hackernews<\/a><\/li>\n<li><a href=\"https:\/\/news.ycombinator.com\/item?id=21456706\">hackernews<\/a><\/li>\n<\/ul>","created_at":"2019-12-08T17:58:35+0000","updated_at":"2019-12-08T17:58:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":43,"domain_name":"beepb00p.xyz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1055,"title":"Async IO in Python: A Complete Walkthrough","url":"https:\/\/realpython.com\/async-io-python\/","content":"<div class=\"border rounded p-3 card mb-2\"><p class=\"mb-0\">Watch Now This tutorial has a related video course created by the Real Python team. Watch it together with the written tutorial to deepen your understanding: <a class=\"stretched-link text-success\" href=\"https:\/\/realpython.com\/courses\/python-3-concurrency-asyncio-module\/\"><strong>Hands-on Python 3 Concurrency With the asyncio Module<\/strong><\/a><\/p><\/div><p>Async IO is a concurrent programming design that has received dedicated support in Python, evolving rapidly from Python 3.4 through 3.7, and <a href=\"https:\/\/twitter.com\/1st1\/status\/1041855365745455104\">probably beyond<\/a>.<\/p><p>You may be thinking with dread, \u201cConcurrency, parallelism, threading, multiprocessing. That\u2019s a lot to grasp already. Where does async IO fit in?\u201d<\/p><p>This tutorial is built to help you answer that question, giving you a firmer grasp of Python\u2019s approach to async IO.<\/p><p><strong>Here\u2019s what you\u2019ll cover:<\/strong><\/p><ul><li>\n<p><strong>Asynchronous IO (async IO)<\/strong>: a language-agnostic paradigm (model) that has implementations across a host of programming languages<\/p>\n<\/li>\n<li>\n<p><strong><code>async<\/code>\/<code>await<\/code><\/strong>: two new Python keywords that are used to define coroutines<\/p>\n<\/li>\n<li>\n<p><strong><code>asyncio<\/code><\/strong>: the Python package that provides a foundation and API for running and managing coroutines<\/p>\n<\/li>\n<\/ul><p>Coroutines (specialized generator functions) are the heart of async IO in Python, and we\u2019ll dive into them later on.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note<\/strong>: In this article, I use the term <strong>async IO<\/strong> to denote the language-agnostic design of asynchronous IO, while <code>asyncio<\/code> refers to the Python package.<\/p><\/div><p>Before you get started, you\u2019ll need to make sure you\u2019re set up to use <code>asyncio<\/code> and other libraries found in this tutorial.<\/p><div class=\"alert alert-warning\" role=\"alert\"><p><strong>Free Bonus:<\/strong> <a href=\"https:\/\/realpython.com\/async-io-python\/\" class=\"alert-link\" data-toggle=\"modal\" data-target=\"#modal-python-mastery-course\" data-focus=\"false\">5 Thoughts On Python Mastery<\/a>, a free course for Python developers that shows you the roadmap and the mindset you'll need to take your Python skills to the next level.<\/p><\/div><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"setting-up-your-environment\">Setting Up Your Environment<\/h2><p>You\u2019ll need Python 3.7 or above to follow this article in its entirety, as well as the <code>aiohttp<\/code> and <code>aiofiles<\/code> packages:<\/p><div class=\"highlight sh\"><pre>$ python3.7 -m venv .\/py37async\n$ source .\/py37async\/bin\/activate  # Windows: .\\py37async\\Scripts\\activate.bat\n$ pip install --upgrade pip aiohttp aiofiles  # Optional: aiodns\n<\/pre><\/div><p>For help with installing Python 3.7 and setting up a virtual environment, check out <a href=\"https:\/\/realpython.com\/installing-python\/\">Python 3 Installation &amp; Setup Guide<\/a> or <a href=\"https:\/\/realpython.com\/python-virtual-environments-a-primer\/\">Virtual Environments Primer<\/a>.<\/p><p>With that, let\u2019s jump in.<\/p><h2 id=\"the-10000-foot-view-of-async-io\">The 10,000-Foot View of Async IO<\/h2><p>Async IO is a bit lesser known than its tried-and-true cousins, multiprocessing and threading. This section will give you a fuller picture of what async IO is and how it fits into its surrounding landscape.<\/p><h3 id=\"where-does-async-io-fit-in\">Where Does Async IO Fit In?<\/h3><p>Concurrency and parallelism are expansive subjects that are not easy to wade into. While this article focuses on async IO and its implementation in Python, it\u2019s worth taking a minute to compare async IO to its counterparts in order to have context about how async IO fits into the larger, sometimes dizzying puzzle.<\/p><p><strong>Parallelism<\/strong> consists of performing multiple operations at the same time. <strong>Multiprocessing<\/strong> is a means to effect parallelism, and it entails spreading tasks over a computer\u2019s central processing units (CPUs, or cores). Multiprocessing is well-suited for CPU-bound tasks: tightly bound <code>for<\/code> loops and mathematical computations usually fall into this category.<\/p><p><strong>Concurrency<\/strong> is a slightly broader term than parallelism. It suggests that multiple tasks have the ability to run in an overlapping manner. (There\u2019s a saying that concurrency does not imply parallelism.)<\/p><p><strong>Threading<\/strong> is a concurrent execution model whereby multiple <a href=\"https:\/\/en.wikipedia.org\/wiki\/Thread_(computing)\">threads<\/a> take turns executing tasks. One process can contain multiple threads. Python has a complicated relationship with threading thanks to its <a href=\"https:\/\/realpython.com\/python-gil\/\">GIL<\/a>, but that\u2019s beyond the scope of this article.<\/p><p>What\u2019s important to know about threading is that it\u2019s better for IO-bound tasks. While a CPU-bound task is characterized by the computer\u2019s cores continually working hard from start to finish, an IO-bound job is dominated by a lot of waiting on input\/output to complete.<\/p><p>To recap the above, concurrency encompasses both multiprocessing (ideal for CPU-bound tasks) and threading (suited for IO-bound tasks). Multiprocessing is a form of parallelism, with parallelism being a specific type (subset) of concurrency. The Python standard library has offered longstanding <a href=\"https:\/\/docs.python.org\/3\/library\/concurrency.html\">support for both of these<\/a> through its <code>multiprocessing<\/code>, <code>threading<\/code>, and <code>concurrent.futures<\/code> packages.<\/p><p>Now it\u2019s time to bring a new member to the mix. Over the last few years, a separate design has been more comprehensively built into CPython: asynchronous IO, enabled through the standard library\u2019s <code>asyncio<\/code> package and the new <code>async<\/code> and <code>await<\/code> language keywords. To be clear, async IO is not a newly invented concept, and it has existed or is being built into other languages and runtime environments, such as <a href=\"https:\/\/gobyexample.com\/goroutines\">Go<\/a>, <a href=\"https:\/\/docs.microsoft.com\/en-us\/dotnet\/csharp\/async\">C#<\/a>, or <a href=\"https:\/\/docs.scala-lang.org\/sips\/async.html\">Scala<\/a>.<\/p><p>The <code>asyncio<\/code> package is billed by the Python documentation as <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio.html\">a library to write concurrent code<\/a>. However, async IO is not threading, nor is it multiprocessing. It is not built on top of either of these.<\/p><p>In fact, async IO is a single-threaded, single-process design: it uses <strong>cooperative multitasking<\/strong>, a term that you\u2019ll flesh out by the end of this tutorial. It has been said in other words that async IO gives a feeling of concurrency despite using a single thread in a single process. Coroutines (a central feature of async IO) can be scheduled concurrently, but they are not inherently concurrent.<\/p><p>To reiterate, async IO is a style of concurrent programming, but it is not parallelism. It\u2019s more closely aligned with threading than with multiprocessing but is very much distinct from both of these and is a standalone member in concurrency\u2019s bag of tricks.<\/p><p>That leaves one more term. What does it mean for something to be <strong>asynchronous<\/strong>? This isn\u2019t a rigorous definition, but for our purposes here, I can think of two properties:<\/p><ul><li>Asynchronous routines are able to \u201cpause\u201d while waiting on their ultimate result and let other routines run in the meantime.<\/li>\n<li>Asynchronous code, through the mechanism above, facilitates concurrent execution. To put it differently, asynchronous code gives the look and feel of concurrency.<\/li>\n<\/ul><p>Here\u2019s a diagram to put it all together. The white terms represent concepts, and the green terms represent ways in which they are implemented or effected:<\/p><p><a href=\"https:\/\/files.realpython.com\/media\/Screen_Shot_2018-10-17_at_3.18.44_PM.c02792872031.jpg\" target=\"_blank\"><img class=\"img-fluid mx-auto d-block w-70\" src=\"https:\/\/files.realpython.com\/media\/Screen_Shot_2018-10-17_at_3.18.44_PM.c02792872031.jpg\" width=\"504\" height=\"411\" srcset=\"https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/Screen_Shot_2018-10-17_at_3.18.44_PM.c02792872031.jpg&amp;w=126&amp;sig=49a0ee12fcb52037d02ba7dba0959616a45e72f1 126w, https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/Screen_Shot_2018-10-17_at_3.18.44_PM.c02792872031.jpg&amp;w=252&amp;sig=de481cd508527d65e641c75e8d98cde592b82d0f 252w, https:\/\/files.realpython.com\/media\/Screen_Shot_2018-10-17_at_3.18.44_PM.c02792872031.jpg 504w\" alt=\"Concurrency versus parallelism\" referrerpolicy=\"no-referrer\" \/><\/a><\/p><p>I\u2019ll stop there on the comparisons between concurrent programming models. This tutorial is focused on the subcomponent that is async IO, how to use it, and the APIs that have sprung up around it. For a thorough exploration of threading versus multiprocessing versus async IO, pause here and check out Jim Anderson\u2019s <a href=\"https:\/\/realpython.com\/python-concurrency\/\">overview of concurrency in Python<\/a>. Jim is way funnier than me and has sat in more meetings than me, to boot.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"async-io-explained\">Async IO Explained<\/h3><p>Async IO may at first seem counterintuitive and paradoxical. How does something that facilitates concurrent code use a single thread and a single CPU core? I\u2019ve never been very good at conjuring up examples, so I\u2019d like to paraphrase one from Miguel Grinberg\u2019s 2017 PyCon talk, which explains everything quite beautifully:<\/p><blockquote>\n<p>Chess master Judit Polg\u00e1r hosts a chess exhibition in which she plays multiple amateur players. She has two ways of conducting the exhibition: synchronously and asynchronously.<\/p>\n<p>Assumptions:<\/p>\n<ul><li>24 opponents<\/li>\n<li>Judit makes each chess move in 5 seconds<\/li>\n<li>Opponents each take 55 seconds to make a move<\/li>\n<li>Games average 30 pair-moves (60 moves total)<\/li>\n<\/ul><p><strong>Synchronous version<\/strong>: Judit plays one game at a time, never two at the same time, until the game is complete. Each game takes <em>(55 + 5) * 30 == 1800<\/em> seconds, or 30 minutes. The entire exhibition takes <em>24 * 30 == 720<\/em> minutes, or <strong>12 hours<\/strong>.<\/p>\n<p><strong>Asynchronous version<\/strong>: Judit moves from table to table, making one move at each table. She leaves the table and lets the opponent make their next move during the wait time. One move on all 24 games takes Judit <em>24 * 5 == 120<\/em> seconds, or 2 minutes. The entire exhibition is now cut down to <em>120 * 30 == 3600<\/em> seconds, or just <strong>1 hour<\/strong>. <a href=\"https:\/\/youtu.be\/iG6fr81xHKA?t=4m29s\">(Source)<\/a><\/p>\n<\/blockquote><p>There is only one Judit Polg\u00e1r, who has only two hands and makes only one move at a time by herself. But playing asynchronously cuts the exhibition time down from 12 hours to one. So, cooperative multitasking is a fancy way of saying that a program\u2019s event loop (more on that later) communicates with multiple tasks to let each take turns running at the optimal time.<\/p><p>Async IO takes long waiting periods in which functions would otherwise be blocking and allows other functions to run during that downtime. (A function that blocks effectively forbids others from running from the time that it starts until the time that it returns.)<\/p><h3 id=\"async-io-is-not-easy\">Async IO Is Not Easy<\/h3><p>I\u2019ve heard it said, \u201cUse async IO when you can; use threading when you must.\u201d The truth is that building durable multithreaded code can be hard and error-prone. Async IO avoids some of the potential speedbumps that you might otherwise encounter with a threaded design.<\/p><p>But that\u2019s not to say that async IO in Python is easy. Be warned: when you venture a bit below the surface level, async programming can be difficult too! Python\u2019s async model is built around concepts such as callbacks, events, transports, protocols, and futures\u2014just the terminology can be intimidating. The fact that its API has been changing continually makes it no easier.<\/p><p>Luckily, <code>asyncio<\/code> has matured to a point where most of its features are no longer provisional, while its documentation has received a huge overhaul and some quality resources on the subject are starting to emerge as well.<\/p><h2 id=\"the-asyncio-package-and-asyncawait\">The <code>asyncio<\/code> Package and <code>async<\/code>\/<code>await<\/code><\/h2><p>Now that you have some background on async IO as a design, let\u2019s explore Python\u2019s implementation. Python\u2019s <code>asyncio<\/code> package (introduced in Python 3.4) and its two keywords, <code>async<\/code> and <code>await<\/code>, serve different purposes but come together to help you declare, build, execute, and manage asynchronous code.<\/p><h3 id=\"the-asyncawait-syntax-and-native-coroutines\">The <code>async<\/code>\/<code>await<\/code> Syntax and Native Coroutines<\/h3><div class=\"alert alert-primary\" role=\"alert\"><p><strong>A Word of Caution<\/strong>: Be careful what you read out there on the Internet. Python\u2019s async IO API has evolved rapidly from Python 3.4 to Python 3.7. Some old patterns are no longer used, and some things that were at first disallowed are now allowed through new introductions. For all I know, this tutorial will join the club of the outdated soon too.<\/p><\/div><p>At the heart of async IO are coroutines. A coroutine is a specialized version of a Python generator function. Let\u2019s start with a baseline definition and then build off of it as you progress here: a coroutine is a function that can suspend its execution before reaching <code>return<\/code>, and it can indirectly pass control to another coroutine for some time.<\/p><p>Later, you\u2019ll dive a lot deeper into how exactly the traditional generator is repurposed into a coroutine. For now, the easiest way to pick up how coroutines work is to start making some.<\/p><p>Let\u2019s take the immersive approach and write some async IO code. This short program is the <code>Hello World<\/code> of async IO but goes a long way towards illustrating its core functionality:<\/p><div class=\"highlight python\"><pre>#!\/usr\/bin\/env python3\n# countasync.py\nimport asyncio\nasync def count():\n    print(\"One\")\n    await asyncio.sleep(1)\n    print(\"Two\")\nasync def main():\n    await asyncio.gather(count(), count(), count())\nif __name__ == \"__main__\":\n    import time\n    s = time.perf_counter()\n    asyncio.run(main())\n    elapsed = time.perf_counter() - s\n    print(f\"{__file__} executed in {elapsed:0.2f} seconds.\")\n<\/pre><\/div><p>When you execute this file, take note of what looks different than if you were to define the functions with just <code>def<\/code> and <code>time.sleep()<\/code>:<\/p><div class=\"highlight sh\"><pre>$ python3 countasync.py\nOne\nOne\nOne\nTwo\nTwo\nTwo\ncountasync.py executed in 1.01 seconds.\n<\/pre><\/div><p>The order of this output is the heart of async IO. Talking to each of the calls to <code>count()<\/code> is a single event loop, or coordinator. When each task reaches <code>await asyncio.sleep(1)<\/code>, the function yells up to the event loop and gives control back to it, saying, \u201cI\u2019m going to be sleeping for 1 second. Go ahead and let something else meaningful be done in the meantime.\u201d<\/p><p>Contrast this to the synchronous version:<\/p><div class=\"highlight python\"><pre>#!\/usr\/bin\/env python3\n# countsync.py\nimport time\ndef count():\n    print(\"One\")\n    time.sleep(1)\n    print(\"Two\")\ndef main():\n    for _ in range(3):\n        count()\nif __name__ == \"__main__\":\n    s = time.perf_counter()\n    main()\n    elapsed = time.perf_counter() - s\n    print(f\"{__file__} executed in {elapsed:0.2f} seconds.\")\n<\/pre><\/div><p>When executed, there is a slight but critical change in order and execution time:<\/p><div class=\"highlight sh\"><pre>$ python3 countsync.py\nOne\nTwo\nOne\nTwo\nOne\nTwo\ncountsync.py executed in 3.01 seconds.\n<\/pre><\/div><p>While using <code>time.sleep()<\/code> and <code>asyncio.sleep()<\/code> may seem banal, they are used as stand-ins for any time-intensive processes that involve wait time. (The most mundane thing you can wait on is a <code>sleep()<\/code> call that does basically nothing.) That is, <code>time.sleep()<\/code> can represent any time-consuming blocking function call, while <code>asyncio.sleep()<\/code> is used to stand in for a non-blocking call (but one that also takes some time to complete).<\/p><p>As you\u2019ll see in the next section, the benefit of awaiting something, including <code>asyncio.sleep()<\/code>, is that the surrounding function can temporarily cede control to another function that\u2019s more readily able to do something immediately. In contrast, <code>time.sleep()<\/code> or any other blocking call is incompatible with asynchronous Python code, because it will stop everything in its tracks for the duration of the sleep time.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"the-rules-of-async-io\">The Rules of Async IO<\/h3><p>At this point, a more formal definition of <code>async<\/code>, <code>await<\/code>, and the coroutine functions that they create are in order. This section is a little dense, but getting a hold of <code>async<\/code>\/<code>await<\/code> is instrumental, so come back to this if you need to:<\/p><ul><li>\n<p>The syntax <code>async def<\/code> introduces either a <strong>native coroutine<\/strong> or an <strong>asynchronous generator<\/strong>. The expressions <code>async with<\/code> and <code>async for<\/code> are also valid, and you\u2019ll see them later on.<\/p>\n<\/li>\n<li>\n<p>The keyword <code>await<\/code> passes function control back to the event loop. (It suspends the execution of the surrounding coroutine.) If Python encounters an <code>await f()<\/code> expression in the scope of <code>g()<\/code>, this is how <code>await<\/code> tells the event loop, \u201cSuspend execution of <code>g()<\/code> until whatever I\u2019m waiting on\u2014the result of <code>f()<\/code>\u2014is returned. In the meantime, go let something else run.\u201d<\/p>\n<\/li>\n<\/ul><p>In code, that second bullet point looks roughly like this:<\/p><div class=\"highlight python\"><pre>async def g():\n    # Pause here and come back to g() when f() is ready\n    r = await f()\n    return r\n<\/pre><\/div><p>There\u2019s also a strict set of rules around when and how you can and cannot use <code>async<\/code>\/<code>await<\/code>. These can be handy whether you are still picking up the syntax or already have exposure to using <code>async<\/code>\/<code>await<\/code>:<\/p><ul><li>\n<p>A function that you introduce with <code>async def<\/code> is a coroutine. It may use <code>await<\/code>, <code>return<\/code>, or <code>yield<\/code>, but all of these are optional. Declaring <code>async def noop(): pass<\/code> is valid:<\/p>\n<ul><li>\n<p>Using <code>await<\/code> and\/or <code>return<\/code> creates a coroutine function. To call a coroutine function, you must <code>await<\/code> it to get its results.<\/p>\n<\/li>\n<li>\n<p>It is less common (and only recently legal in Python) to use <code>yield<\/code> in an <code>async def<\/code> block. This creates an <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0525\/\">asynchronous generator<\/a>, which you iterate over with <code>async for<\/code>. Forget about async generators for the time being and focus on getting down the syntax for coroutine functions, which use <code>await<\/code> and\/or <code>return<\/code>.<\/p>\n<\/li>\n<li>\n<p>Anything defined with <code>async def<\/code> may not use <code>yield from<\/code>, which will raise a <code>SyntaxError<\/code>.<\/p>\n<\/li>\n<\/ul><\/li>\n<li>\n<p>Just like it\u2019s a <code>SyntaxError<\/code> to use <code>yield<\/code> outside of a <code>def<\/code> function, it is a <code>SyntaxError<\/code> to use <code>await<\/code> outside of an <code>async def<\/code> coroutine. You can only use <code>await<\/code> in the body of coroutines.<\/p>\n<\/li>\n<\/ul><p>Here are some terse examples meant to summarize the above few rules:<\/p><div class=\"highlight python\"><pre>async def f(x):\n    y = await z(x)  # OK - `await` and `return` allowed in coroutines\n    return y\nasync def g(x):\n    yield x  # OK - this is an async generator\nasync def m(x):\n    yield from gen(x)  # No - SyntaxError\ndef m(x):\n    y = await z(x)  # Still no - SyntaxError (no `async def` here)\n    return y\n<\/pre><\/div><p>Finally, when you use <code>await f()<\/code>, it\u2019s required that <code>f()<\/code> be an object that is <a href=\"https:\/\/docs.python.org\/3\/reference\/datamodel.html#awaitable-objects\">awaitable<\/a>. Well, that\u2019s not very helpful, is it? For now, just know that an awaitable object is either (1) another coroutine or (2) an object defining an <code>.__await__()<\/code> dunder method that returns an iterator. If you\u2019re writing a program, for the large majority of purposes, you should only need to worry about case #1.<\/p><p>That brings us to one more technical distinction that you may see pop up: an older way of marking a function as a coroutine is to decorate a normal <code>def<\/code> function with <code>@asyncio.coroutine<\/code>. The result is a <strong>generator-based coroutine<\/strong>. This construction has been outdated since the <code>async<\/code>\/<code>await<\/code> syntax was put in place in Python 3.5.<\/p><p>These two coroutines are essentially equivalent (both are awaitable), but the first is <strong>generator-based<\/strong>, while the second is a <strong>native coroutine<\/strong>:<\/p><div class=\"highlight python\"><pre>import asyncio\n@asyncio.coroutine\ndef py34_coro():\n    \"\"\"Generator-based coroutine, older syntax\"\"\"\n    yield from stuff()\nasync def py35_coro():\n    \"\"\"Native coroutine, modern syntax\"\"\"\n    await stuff()\n<\/pre><\/div><p>If you\u2019re writing any code yourself, prefer native coroutines for the sake of being explicit rather than implicit. Generator-based coroutines will be <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio-task.html#generator-based-coroutines\">removed<\/a> in Python 3.10.<\/p><p>Towards the latter half of this tutorial, we\u2019ll touch on generator-based coroutines for explanation\u2019s sake only. The reason that <code>async<\/code>\/<code>await<\/code> were introduced is to make coroutines a standalone feature of Python that can be easily differentiated from a normal generator function, thus reducing ambiguity.<\/p><p>Don\u2019t get bogged down in generator-based coroutines, which have been <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0492\/#rationale-and-goals\">deliberately outdated<\/a> by <code>async<\/code>\/<code>await<\/code>. They have their own small set of rules (for instance, <code>await<\/code> cannot be used in a generator-based coroutine) that are largely irrelevant if you stick to the <code>async<\/code>\/<code>await<\/code> syntax.<\/p><p>Without further ado, let\u2019s take on a few more involved examples.<\/p><p>Here\u2019s one example of how async IO cuts down on wait time: given a coroutine <code>makerandom()<\/code> that keeps producing random integers in the range [0, 10], until one of them exceeds a threshold, you want to let multiple calls of this coroutine not need to wait for each other to complete in succession. You can largely follow the patterns from the two scripts above, with slight changes:<\/p><div class=\"highlight python\"><pre>#!\/usr\/bin\/env python3\n# rand.py\nimport asyncio\nimport random\n# ANSI colors\nc = (\n    \"\\033[0m\",   # End of color\n    \"\\033[36m\",  # Cyan\n    \"\\033[91m\",  # Red\n    \"\\033[35m\",  # Magenta\n)\nasync def makerandom(idx: int, threshold: int = 6) -&gt; int:\n    print(c[idx + 1] + f\"Initiated makerandom({idx}).\")\n    i = random.randint(0, 10)\n    while i &lt;= threshold:\n        print(c[idx + 1] + f\"makerandom({idx}) == {i} too low; retrying.\")\n        await asyncio.sleep(idx + 1)\n        i = random.randint(0, 10)\n    print(c[idx + 1] + f\"---&gt; Finished: makerandom({idx}) == {i}\" + c[0])\n    return i\nasync def main():\n    res = await asyncio.gather(*(makerandom(i, 10 - i - 1) for i in range(3)))\n    return res\nif __name__ == \"__main__\":\n    random.seed(444)\n    r1, r2, r3 = asyncio.run(main())\n    print()\n    print(f\"r1: {r1}, r2: {r2}, r3: {r3}\")\n<\/pre><\/div><p>The colorized output says a lot more than I can and gives you a sense for how this script is carried out:<\/p><figure class=\"figure mx-auto d-block\"><a href=\"https:\/\/files.realpython.com\/media\/asyncio-rand.dffdd83b4256.gif\" target=\"_blank\"><img class=\"img-fluid mx-auto d-block w-70\" src=\"https:\/\/files.realpython.com\/media\/asyncio-rand.dffdd83b4256.gif\" width=\"300\" height=\"400\" srcset=\"https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/asyncio-rand.dffdd83b4256.gif&amp;w=75&amp;sig=1730fc7b8bca67960626d95ccf73c50dfe4c66dd 75w, https:\/\/robocrop.realpython.net\/?url=https%3A\/\/files.realpython.com\/media\/asyncio-rand.dffdd83b4256.gif&amp;w=150&amp;sig=35259ed0d3ddd6b7d3165c31498b67d657715b21 150w, https:\/\/files.realpython.com\/media\/asyncio-rand.dffdd83b4256.gif 300w\" alt=\"rand.py program execution\" referrerpolicy=\"no-referrer\" \/><\/a>\n<figcaption class=\"figure-caption text-center\">rand.py execution<\/figcaption><\/figure><p>This program uses one main coroutine, <code>makerandom()<\/code>, and runs it concurrently across 3 different inputs. Most programs will contain small, modular coroutines and one wrapper function that serves to chain each of the smaller coroutines together. <code>main()<\/code> is then used to gather tasks (futures) by mapping the central coroutine across some iterable or pool.<\/p><p>In this miniature example, the pool is <code>range(3)<\/code>. In a fuller example presented later, it is a set of URLs that need to be requested, parsed, and processed concurrently, and <code>main()<\/code> encapsulates that entire routine for each URL.<\/p><p>While \u201cmaking random integers\u201d (which is CPU-bound more than anything) is maybe not the greatest choice as a candidate for <code>asyncio<\/code>, it\u2019s the presence of <code>asyncio.sleep()<\/code> in the example that is designed to mimic an IO-bound process where there is uncertain wait time involved. For example, the <code>asyncio.sleep()<\/code> call might represent sending and receiving not-so-random integers between two clients in a message application.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><p>Async IO comes with its own set of possible script designs, which you\u2019ll get introduced to in this section.<\/p><h3 id=\"chaining-coroutines\">Chaining Coroutines<\/h3><p>A key feature of coroutines is that they can be chained together. (Remember, a coroutine object is awaitable, so another coroutine can <code>await<\/code> it.) This allows you to break programs into smaller, manageable, recyclable coroutines:<\/p><div class=\"highlight python\"><pre>#!\/usr\/bin\/env python3\n# chained.py\nimport asyncio\nimport random\nimport time\nasync def part1(n: int) -&gt; str:\n    i = random.randint(0, 10)\n    print(f\"part1({n}) sleeping for {i} seconds.\")\n    await asyncio.sleep(i)\n    result = f\"result{n}-1\"\n    print(f\"Returning part1({n}) == {result}.\")\n    return result\nasync def part2(n: int, arg: str) -&gt; str:\n    i = random.randint(0, 10)\n    print(f\"part2{n, arg} sleeping for {i} seconds.\")\n    await asyncio.sleep(i)\n    result = f\"result{n}-2 derived from {arg}\"\n    print(f\"Returning part2{n, arg} == {result}.\")\n    return result\nasync def chain(n: int) -&gt; None:\n    start = time.perf_counter()\n    p1 = await part1(n)\n    p2 = await part2(n, p1)\n    end = time.perf_counter() - start\n    print(f\"--&gt;Chained result{n} =&gt; {p2} (took {end:0.2f} seconds).\")\nasync def main(*args):\n    await asyncio.gather(*(chain(n) for n in args))\nif __name__ == \"__main__\":\n    import sys\n    random.seed(444)\n    args = [1, 2, 3] if len(sys.argv) == 1 else map(int, sys.argv[1:])\n    start = time.perf_counter()\n    asyncio.run(main(*args))\n    end = time.perf_counter() - start\n    print(f\"Program finished in {end:0.2f} seconds.\")\n<\/pre><\/div><p>Pay careful attention to the output, where <code>part1()<\/code> sleeps for a variable amount of time, and <code>part2()<\/code> begins working with the results as they become available:<\/p><div class=\"highlight sh\"><pre>$ python3 chained.py 9 6 3\npart1(9) sleeping for 4 seconds.\npart1(6) sleeping for 4 seconds.\npart1(3) sleeping for 0 seconds.\nReturning part1(3) == result3-1.\npart2(3, 'result3-1') sleeping for 4 seconds.\nReturning part1(9) == result9-1.\npart2(9, 'result9-1') sleeping for 7 seconds.\nReturning part1(6) == result6-1.\npart2(6, 'result6-1') sleeping for 4 seconds.\nReturning part2(3, 'result3-1') == result3-2 derived from result3-1.\n--&gt;Chained result3 =&gt; result3-2 derived from result3-1 (took 4.00 seconds).\nReturning part2(6, 'result6-1') == result6-2 derived from result6-1.\n--&gt;Chained result6 =&gt; result6-2 derived from result6-1 (took 8.01 seconds).\nReturning part2(9, 'result9-1') == result9-2 derived from result9-1.\n--&gt;Chained result9 =&gt; result9-2 derived from result9-1 (took 11.01 seconds).\nProgram finished in 11.01 seconds.\n<\/pre><\/div><p>In this setup, the runtime of <code>main()<\/code> will be equal to the maximum runtime of the tasks that it gathers together and schedules.<\/p><h3 id=\"using-a-queue\">Using a Queue<\/h3><p>The <code>asyncio<\/code> package provides <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio-queue.html\">queue classes<\/a> that are designed to be similar to classes of the <a href=\"https:\/\/docs.python.org\/3\/library\/queue.html#module-queue\"><code>queue<\/code><\/a> module. In our examples so far, we haven\u2019t really had a need for a queue structure. In <code>chained.py<\/code>, each task (future) is composed of a set of coroutines that explicitly await each other and pass through a single input per chain.<\/p><p>There is an alternative structure that can also work with async IO: a number of producers, which are not associated with each other, add items to a queue. Each producer may add multiple items to the queue at staggered, random, unannounced times. A group of consumers pull items from the queue as they show up, greedily and without waiting for any other signal.<\/p><p>In this design, there is no chaining of any individual consumer to a producer. The consumers don\u2019t know the number of producers, or even the cumulative number of items that will be added to the queue, in advance.<\/p><p>It takes an individual producer or consumer a variable amount of time to put and extract items from the queue, respectively. The queue serves as a throughput that can communicate with the producers and consumers without them talking to each other directly.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note<\/strong>: While queues are often used in threaded programs because of the thread-safety of <code>queue.Queue()<\/code>, you shouldn\u2019t need to concern yourself with thread safety when it comes to async IO. (The exception is when you\u2019re combining the two, but that isn\u2019t done in this tutorial.)<\/p><p>One use-case for queues (as is the case here) is for the queue to act as a transmitter for producers and consumers that aren\u2019t otherwise directly chained or associated with each other.<\/p><\/div><p>The synchronous version of this program would look pretty dismal: a group of blocking producers serially add items to the queue, one producer at a time. Only after all producers are done can the queue be processed, by one consumer at a time processing item-by-item. There is a ton of latency in this design. Items may sit idly in the queue rather than be picked up and processed immediately.<\/p><p>An asynchronous version, <code>asyncq.py<\/code>, is below. The challenging part of this workflow is that there needs to be a signal to the consumers that production is done. Otherwise, <code>await q.get()<\/code> will hang indefinitely, because the queue will have been fully processed, but consumers won\u2019t have any idea that production is complete.<\/p><p>(Big thanks for some help from a StackOverflow <a href=\"https:\/\/stackoverflow.com\/a\/52615705\/7954504\">user<\/a> for helping to straighten out <code>main()<\/code>: the key is to <code>await q.join()<\/code>, which blocks until all items in the queue have been received and processed, and then to cancel the consumer tasks, which would otherwise hang up and wait endlessly for additional queue items to appear.)<\/p><p>Here is the full script:<\/p><div class=\"highlight python\"><pre>#!\/usr\/bin\/env python3\n# asyncq.py\nimport asyncio\nimport itertools as it\nimport os\nimport random\nimport time\nasync def makeitem(size: int = 5) -&gt; str:\n    return os.urandom(size).hex()\nasync def randsleep(a: int = 1, b: int = 5, caller=None) -&gt; None:\n    i = random.randint(0, 10)\n    if caller:\n        print(f\"{caller} sleeping for {i} seconds.\")\n    await asyncio.sleep(i)\nasync def produce(name: int, q: asyncio.Queue) -&gt; None:\n    n = random.randint(0, 10)\n    for _ in it.repeat(None, n):  # Synchronous loop for each single producer\n        await randsleep(caller=f\"Producer {name}\")\n        i = await makeitem()\n        t = time.perf_counter()\n        await q.put((i, t))\n        print(f\"Producer {name} added &lt;{i}&gt; to queue.\")\nasync def consume(name: int, q: asyncio.Queue) -&gt; None:\n    while True:\n        await randsleep(caller=f\"Consumer {name}\")\n        i, t = await q.get()\n        now = time.perf_counter()\n        print(f\"Consumer {name} got element &lt;{i}&gt;\"\n              f\" in {now-t:0.5f} seconds.\")\n        q.task_done()\nasync def main(nprod: int, ncon: int):\n    q = asyncio.Queue()\n    producers = [asyncio.create_task(produce(n, q)) for n in range(nprod)]\n    consumers = [asyncio.create_task(consume(n, q)) for n in range(ncon)]\n    await asyncio.gather(*producers)\n    await q.join()  # Implicitly awaits consumers, too\n    for c in consumers:\n        c.cancel()\nif __name__ == \"__main__\":\n    import argparse\n    random.seed(444)\n    parser = argparse.ArgumentParser()\n    parser.add_argument(\"-p\", \"--nprod\", type=int, default=5)\n    parser.add_argument(\"-c\", \"--ncon\", type=int, default=10)\n    ns = parser.parse_args()\n    start = time.perf_counter()\n    asyncio.run(main(**ns.__dict__))\n    elapsed = time.perf_counter() - start\n    print(f\"Program completed in {elapsed:0.5f} seconds.\")\n<\/pre><\/div><p>The first few coroutines are helper functions that return a random string, a fractional-second performance counter, and a random integer. A producer puts anywhere from 1 to 5 items into the queue. Each item is a tuple of <code>(i, t)<\/code> where <code>i<\/code> is a random string and <code>t<\/code> is the time at which the producer attempts to put the tuple into the queue.<\/p><p>When a consumer pulls an item out, it simply calculates the elapsed time that the item sat in the queue using the timestamp that the item was put in with.<\/p><p>Keep in mind that <code>asyncio.sleep()<\/code> is used to mimic some other, more complex coroutine that would eat up time and block all other execution if it were a regular blocking function.<\/p><p>Here is a test run with two producers and five consumers:<\/p><div class=\"highlight sh\"><pre>$ python3 asyncq.py -p 2 -c 5\nProducer 0 sleeping for 3 seconds.\nProducer 1 sleeping for 3 seconds.\nConsumer 0 sleeping for 4 seconds.\nConsumer 1 sleeping for 3 seconds.\nConsumer 2 sleeping for 3 seconds.\nConsumer 3 sleeping for 5 seconds.\nConsumer 4 sleeping for 4 seconds.\nProducer 0 added &lt;377b1e8f82&gt; to queue.\nProducer 0 sleeping for 5 seconds.\nProducer 1 added &lt;413b8802f8&gt; to queue.\nConsumer 1 got element &lt;377b1e8f82&gt; in 0.00013 seconds.\nConsumer 1 sleeping for 3 seconds.\nConsumer 2 got element &lt;413b8802f8&gt; in 0.00009 seconds.\nConsumer 2 sleeping for 4 seconds.\nProducer 0 added &lt;06c055b3ab&gt; to queue.\nProducer 0 sleeping for 1 seconds.\nConsumer 0 got element &lt;06c055b3ab&gt; in 0.00021 seconds.\nConsumer 0 sleeping for 4 seconds.\nProducer 0 added &lt;17a8613276&gt; to queue.\nConsumer 4 got element &lt;17a8613276&gt; in 0.00022 seconds.\nConsumer 4 sleeping for 5 seconds.\nProgram completed in 9.00954 seconds.\n<\/pre><\/div><p>In this case, the items process in fractions of a second. A delay can be due to two reasons:<\/p><ul><li>Standard, largely unavoidable overhead<\/li>\n<li>Situations where all consumers are sleeping when an item appears in the queue<\/li>\n<\/ul><p>With regards to the second reason, luckily, it is perfectly normal to scale to hundreds or thousands of consumers. You should have no problem with <code>python3 asyncq.py -p 5 -c 100<\/code>. The point here is that, theoretically, you could have different users on different systems controlling the management of producers and consumers, with the queue serving as the central throughput.<\/p><p>So far, you\u2019ve been thrown right into the fire and seen three related examples of <code>asyncio<\/code> calling coroutines defined with <code>async<\/code> and <code>await<\/code>. If you\u2019re not completely following or just want to get deeper into the mechanics of how modern coroutines came to be in Python, you\u2019ll start from square one with the next section.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"async-ios-roots-in-generators\">Async IO\u2019s Roots in Generators<\/h2><p>Earlier, you saw an example of the old-style generator-based coroutines, which have been outdated by more explicit native coroutines. The example is worth re-showing with a small tweak:<\/p><div class=\"highlight python\"><pre>import asyncio\n@asyncio.coroutine\ndef py34_coro():\n    \"\"\"Generator-based coroutine\"\"\"\n    # No need to build these yourself, but be aware of what they are\n    s = yield from stuff()\n    return s\nasync def py35_coro():\n    \"\"\"Native coroutine, modern syntax\"\"\"\n    s = await stuff()\n    return s\nasync def stuff():\n    return 0x10, 0x20, 0x30\n<\/pre><\/div><p>As an experiment, what happens if you call <code>py34_coro()<\/code> or <code>py35_coro()<\/code> on its own, without <code>await<\/code>, or without any calls to <code>asyncio.run()<\/code> or other <code>asyncio<\/code> \u201cporcelain\u201d functions? Calling a coroutine in isolation returns a coroutine object:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; py35_coro()\n&lt;coroutine object py35_coro at 0x10126dcc8&gt;\n<\/pre><\/div><p>This isn\u2019t very interesting on its surface. The result of calling a coroutine on its own is an awaitable <strong>coroutine object<\/strong>.<\/p><p>Time for a quiz: what other feature of Python looks like this? (What feature of Python doesn\u2019t actually \u201cdo much\u201d when it\u2019s called on its own?)<\/p><p>Hopefully you\u2019re thinking of <strong>generators<\/strong> as an answer to this question, because coroutines are enhanced generators under the hood. The behavior is similar in this regard:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; def gen():\n...     yield 0x10, 0x20, 0x30\n...\n&gt;&gt;&gt; g = gen()\n&gt;&gt;&gt; g  # Nothing much happens - need to iterate with `.__next__()`\n&lt;generator object gen at 0x1012705e8&gt;\n&gt;&gt;&gt; next(g)\n(16, 32, 48)\n<\/pre><\/div><p>Generator functions are, as it so happens, the foundation of async IO (regardless of whether you declare coroutines with <code>async def<\/code> rather than the older <code>@asyncio.coroutine<\/code> wrapper). Technically, <code>await<\/code> is more closely analogous to <code>yield from<\/code> than it is to <code>yield<\/code>. (But remember that <code>yield from x()<\/code> is just syntactic sugar to replace <code>for i in x(): yield i<\/code>.)<\/p><p>One critical feature of generators as it pertains to async IO is that they can effectively be stopped and restarted at will. For example, you can <code>break<\/code> out of iterating over a generator object and then resume iteration on the remaining values later. When a generator function reaches <code>yield<\/code>, it yields that value, but then it sits idle until it is told to yield its subsequent value.<\/p><p>This can be fleshed out through an example:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; from itertools import cycle\n&gt;&gt;&gt; def endless():\n...     \"\"\"Yields 9, 8, 7, 6, 9, 8, 7, 6, ... forever\"\"\"\n...     yield from cycle((9, 8, 7, 6))\n&gt;&gt;&gt; e = endless()\n&gt;&gt;&gt; total = 0\n&gt;&gt;&gt; for i in e:\n...     if total &lt; 30:\n...         print(i, end=\" \")\n...         total += i\n...     else:\n...         print()\n...         # Pause execution. We can resume later.\n...         break\n9 8 7 6 9 8 7 6 9 8 7 6 9 8\n&gt;&gt;&gt; # Resume\n&gt;&gt;&gt; next(e), next(e), next(e)\n(6, 9, 8)\n<\/pre><\/div><p>The <code>await<\/code> keyword behaves similarly, marking a break point at which the coroutine suspends itself and lets other coroutines work. \u201cSuspended,\u201d in this case, means a coroutine that has temporarily ceded control but not totally exited or finished. Keep in mind that <code>yield<\/code>, and by extension <code>yield from<\/code> and <code>await<\/code>, mark a break point in a generator\u2019s execution.<\/p><p>This is the fundamental difference between functions and generators. A function is all-or-nothing. Once it starts, it won\u2019t stop until it hits a <code>return<\/code>, then pushes that value to the caller (the function that calls it). A generator, on the other hand, pauses each time it hits a <code>yield<\/code> and goes no further. Not only can it push this value to calling stack, but it can keep a hold of its local variables when you resume it by calling <code>next()<\/code> on it.<\/p><p>There\u2019s a second and lesser-known feature of generators that also matters. You can send a value into a generator as well through its <code>.send()<\/code> method. This allows generators (and coroutines) to call (<code>await<\/code>) each other without blocking. I won\u2019t get any further into the nuts and bolts of this feature, because it matters mainly for the implementation of coroutines behind the scenes, but you shouldn\u2019t ever really need to use it directly yourself.<\/p><p>If you\u2019re interested in exploring more, you can start at <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0342\/\">PEP 342<\/a>, where coroutines were formally introduced. Brett Cannon\u2019s <a href=\"https:\/\/snarky.ca\/how-the-heck-does-async-await-work-in-python-3-5\/\">How the Heck Does Async-Await Work in Python<\/a> is also a good read, as is the <a href=\"https:\/\/pymotw.com\/3\/asyncio\/coroutines.html\">PYMOTW writeup on <code>asyncio<\/code><\/a>. Lastly, there\u2019s David Beazley\u2019s <a href=\"http:\/\/www.dabeaz.com\/coroutines\/\">Curious Course on Coroutines and Concurrency<\/a>, which dives deep into the mechanism by which coroutines run.<\/p><p>Let\u2019s try to condense all of the above articles into a few sentences: there is a particularly unconventional mechanism by which these coroutines actually get run. Their result is an attribute of the exception object that gets thrown when their <code>.send()<\/code> method is called. There\u2019s some more wonky detail to all of this, but it probably won\u2019t help you use this part of the language in practice, so let\u2019s move on for now.<\/p><p>To tie things together, here are some key points on the topic of coroutines as generators:<\/p><ul><li>\n<p>Coroutines are <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0492\/#differences-from-generators\">repurposed generators<\/a> that take advantage of the peculiarities of generator methods.<\/p>\n<\/li>\n<li>\n<p>Old generator-based coroutines use <code>yield from<\/code> to wait for a coroutine result. Modern Python syntax in native coroutines simply replaces <code>yield from<\/code> with <code>await<\/code> as the means of waiting on a coroutine result. The <code>await<\/code> is analogous to <code>yield from<\/code>, and it often helps to think of it as such.<\/p>\n<\/li>\n<li>\n<p>The use of <code>await<\/code> is a signal that marks a break point. It lets a coroutine temporarily suspend execution and permits the program to come back to it later.<\/p>\n<\/li>\n<\/ul><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"other-features-async-for-and-async-generators-comprehensions\">Other Features: <code>async for<\/code> and Async Generators + Comprehensions<\/h3><p>Along with plain <code>async<\/code>\/<code>await<\/code>, Python also enables <code>async for<\/code> to iterate over an <strong>asynchronous iterator<\/strong>. The purpose of an asynchronous iterator is for it to be able to call asynchronous code at each stage when it is iterated over.<\/p><p>A natural extension of this concept is an <strong>asynchronous generator<\/strong>. Recall that you can use <code>await<\/code>, <code>return<\/code>, or <code>yield<\/code> in a native coroutine. Using <code>yield<\/code> within a coroutine became possible in Python 3.6 (via PEP 525), which introduced asynchronous generators with the purpose of allowing <code>await<\/code> and <code>yield<\/code> to be used in the same coroutine function body:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; async def mygen(u: int = 10):\n...     \"\"\"Yield powers of 2.\"\"\"\n...     i = 0\n...     while i &lt; u:\n...         yield 2 ** i\n...         i += 1\n...         await asyncio.sleep(0.1)\n<\/pre><\/div><p>Last but not least, Python enables <strong>asynchronous comprehension<\/strong> with <code>async for<\/code>. Like its synchronous cousin, this is largely syntactic sugar:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; async def main():\n...     # This does *not* introduce concurrent execution\n...     # It is meant to show syntax only\n...     g = [i async for i in mygen()]\n...     f = [j async for j in mygen() if not (j \/\/ 3 % 5)]\n...     return g, f\n...\n&gt;&gt;&gt; g, f = asyncio.run(main())\n&gt;&gt;&gt; g\n[1, 2, 4, 8, 16, 32, 64, 128, 256, 512]\n&gt;&gt;&gt; f\n[1, 2, 16, 32, 256, 512]\n<\/pre><\/div><p>This is a crucial distinction: <strong>neither asynchronous generators nor comprehensions make the iteration concurrent<\/strong>. All that they do is provide the look-and-feel of their synchronous counterparts, but with the ability for the loop in question to give up control to the event loop for some other coroutine to run.<\/p><p>In other words, asynchronous iterators and asynchronous generators are not designed to concurrently map some function over a sequence or iterator. They\u2019re merely designed to let the enclosing coroutine allow other tasks to take their turn. The <code>async for<\/code> and <code>async with<\/code> statements are only needed to the extent that using plain <code>for<\/code> or <code>with<\/code> would \u201cbreak\u201d the nature of <code>await<\/code> in the coroutine. This distinction between asynchronicity and concurrency is a key one to grasp.<\/p><h3 id=\"the-event-loop-and-asynciorun\">The Event Loop and <code>asyncio.run()<\/code><\/h3><p>You can think of an event loop as something like a <code>while True<\/code> loop that monitors coroutines, taking feedback on what\u2019s idle, and looking around for things that can be executed in the meantime. It is able to wake up an idle coroutine when whatever that coroutine is waiting on becomes available.<\/p><p>Thus far, the entire management of the event loop has been implicitly handled by one function call:<\/p><div class=\"highlight python\"><pre>asyncio.run(main())  # Python 3.7+\n<\/pre><\/div><p><a href=\"https:\/\/github.com\/python\/cpython\/blob\/d4c76d960b8b286b75c933780416ace9cda682fd\/Lib\/asyncio\/runners.py#L8\"><code>asyncio.run()<\/code><\/a>, introduced in Python 3.7, is responsible for getting the event loop, running tasks until they are marked as complete, and then closing the event loop.<\/p><p>There\u2019s a more long-winded way of managing the <code>asyncio<\/code> event loop, with <code>get_event_loop()<\/code>. The typical pattern looks like this:<\/p><div class=\"highlight python\"><pre>loop = asyncio.get_event_loop()\ntry:\n    loop.run_until_complete(main())\nfinally:\n    loop.close()\n<\/pre><\/div><p>You\u2019ll probably see <code>loop.get_event_loop()<\/code> floating around in older examples, but unless you have a specific need to fine-tune control over the event loop management, <code>asyncio.run()<\/code> should be sufficient for most programs.<\/p><p>If you do need to interact with the event loop within a Python program, <code>loop<\/code> is a good-old-fashioned Python object that supports introspection with <code>loop.is_running()<\/code> and <code>loop.is_closed()<\/code>. You can manipulate it if you need to get more fine-tuned control, such as in <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#asyncio-example-lowlevel-helloworld\">scheduling a callback<\/a> by passing the loop as an argument.<\/p><p>What is more crucial is understanding a bit beneath the surface about the mechanics of the event loop. Here are a few points worth stressing about the event loop.<\/p><p><strong>#1:<\/strong> Coroutines don\u2019t do much on their own until they are tied to the event loop.<\/p><p>You saw this point before in the explanation on generators, but it\u2019s worth restating. If you have a main coroutine that awaits others, simply calling it in isolation has little effect:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; import asyncio\n&gt;&gt;&gt; async def main():\n...     print(\"Hello ...\")\n...     await asyncio.sleep(1)\n...     print(\"World!\")\n&gt;&gt;&gt; routine = main()\n&gt;&gt;&gt; routine\n&lt;coroutine object main at 0x1027a6150&gt;\n<\/pre><\/div><p>Remember to use <code>asyncio.run()<\/code> to actually force execution by scheduling the <code>main()<\/code> coroutine (future object) for execution on the event loop:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; asyncio.run(routine)\nHello ...\nWorld!\n<\/pre><\/div><p>(Other coroutines can be executed with <code>await<\/code>. It is typical to wrap just <code>main()<\/code> in <code>asyncio.run()<\/code>, and chained coroutines with <code>await<\/code> will be called from there.)<\/p><p><strong>#2:<\/strong> By default, an async IO event loop runs in a single thread and on a single CPU core. Usually, running one single-threaded event loop in one CPU core is more than sufficient. It is also possible to run event loops across multiple cores. Check out this <a href=\"https:\/\/youtu.be\/0kXaLh8Fz3k?t=10m30s\">talk by John Reese<\/a> for more, and be warned that your laptop may spontaneously combust.<\/p><p><strong>#3.<\/strong> Event loops are pluggable. That is, you could, if you really wanted, write your own event loop implementation and have it run tasks just the same. This is wonderfully demonstrated in the <a href=\"https:\/\/github.com\/MagicStack\/uvloop\"><code>uvloop<\/code><\/a> package, which is an implementation of the event loop in Cython.<\/p><p>That is what is meant by the term \u201cpluggable event loop\u201d: you can use any working implementation of an event loop, unrelated to the structure of the coroutines themselves. The <code>asyncio<\/code> package itself ships with <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#event-loop-implementations\">two different event loop implementations<\/a>, with the default being based on the <a href=\"https:\/\/docs.python.org\/3\/library\/selectors.html#module-selectors\"><code>selectors<\/code><\/a> module. (The second implementation is built for Windows only.)<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"a-full-program-asynchronous-requests\">A Full Program: Asynchronous Requests<\/h2><p>You\u2019ve made it this far, and now it\u2019s time for the fun and painless part. In this section, you\u2019ll build a web-scraping URL collector, <code>areq.py<\/code>, using <code>aiohttp<\/code>, a blazingly fast async HTTP client\/server framework. (We just need the client part.) Such a tool could be used to map connections between a cluster of sites, with the links forming a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Directed_graph\">directed graph<\/a>.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note<\/strong>: You may be wondering why Python\u2019s <code>requests<\/code> package isn\u2019t compatible with async IO. <code>requests<\/code> is built on top of <code>urllib3<\/code>, which in turn uses Python\u2019s <code>http<\/code> and <code>socket<\/code> modules.<\/p><p>By default, socket operations are blocking. This means that Python won\u2019t like <code>await requests.get(url)<\/code> because <code>.get()<\/code> is not awaitable. In contrast, almost everything in <code>aiohttp<\/code> is an awaitable coroutine, such as <a href=\"https:\/\/github.com\/aio-libs\/aiohttp\/blob\/508adbb656da2e9ae660da5e98e1e5fa6669a3f4\/aiohttp\/client.py#L225\"><code>session.request()<\/code><\/a> and <a href=\"https:\/\/github.com\/aio-libs\/aiohttp\/blob\/da75122f6089a250128d2736f2bd88d10e97ca17\/aiohttp\/client_reqrep.py#L913\"><code>response.text()<\/code><\/a>. It\u2019s a great package otherwise, but you\u2019re doing yourself a disservice by using <code>requests<\/code> in asynchronous code.<\/p><\/div><p>The high-level program structure will look like this:<\/p><ol><li>\n<p>Read a sequence of URLs from a local file, <code>urls.txt<\/code>.<\/p>\n<\/li>\n<li>\n<p>Send GET requests for the URLs and decode the resulting content. If this fails, stop there for a URL.<\/p>\n<\/li>\n<li>\n<p>Search for the URLs within <code>href<\/code> tags in the HTML of the responses.<\/p>\n<\/li>\n<li>\n<p>Write the results to <code>foundurls.txt<\/code>.<\/p>\n<\/li>\n<li>\n<p>Do all of the above as asynchronously and concurrently as possible. (Use <code>aiohttp<\/code> for the requests, and <code>aiofiles<\/code> for the file-appends. These are two primary examples of IO that are well-suited for the async IO model.)<\/p>\n<\/li>\n<\/ol><p>Here are the contents of <code>urls.txt<\/code>. It\u2019s not huge, and contains mostly highly trafficked sites:<\/p><div class=\"highlight sh\"><pre>$ cat urls.txt\nhttps:\/\/regex101.com\/\nhttps:\/\/docs.python.org\/3\/this-url-will-404.html\nhttps:\/\/www.nytimes.com\/guides\/\nhttps:\/\/www.mediamatters.org\/\nhttps:\/\/1.1.1.1\/\nhttps:\/\/www.politico.com\/tipsheets\/morning-money\nhttps:\/\/www.bloomberg.com\/markets\/economics\nhttps:\/\/www.ietf.org\/rfc\/rfc2616.txt\n<\/pre><\/div><p>The second URL in the list should return a 404 response, which you\u2019ll need to handle gracefully. If you\u2019re running an expanded version of this program, you\u2019ll probably need to deal with much hairier problems than this, such a server disconnections and endless redirects.<\/p><p>The requests themselves should be made using a single session, to take advantage of reusage of the session\u2019s internal connection pool.<\/p><p>Let\u2019s take a look at the full program. We\u2019ll walk through things step-by-step after:<\/p><div class=\"highlight python\"><pre>#!\/usr\/bin\/env python3\n# areq.py\n\"\"\"Asynchronously get links embedded in multiple pages' HMTL.\"\"\"\nimport asyncio\nimport logging\nimport re\nimport sys\nfrom typing import IO\nimport urllib.error\nimport urllib.parse\nimport aiofiles\nimport aiohttp\nfrom aiohttp import ClientSession\nlogging.basicConfig(\n    format=\"%(asctime)s%(levelname)s:%(name)s: %(message)s\",\n    level=logging.DEBUG,\n    datefmt=\"%H:%M:%S\",\n    stream=sys.stderr,\n)\nlogger = logging.getLogger(\"areq\")\nlogging.getLogger(\"chardet.charsetprober\").disabled = True\nHREF_RE = re.compile(r'href=\"(.*?)\"')\nasync def fetch_html(url: str, session: ClientSession, **kwargs) -&gt; str:\n    \"\"\"GET request wrapper to fetch page HTML.\n    kwargs are passed to `session.request()`.\n    \"\"\"\n    resp = await session.request(method=\"GET\", url=url, **kwargs)\n    resp.raise_for_status()\n    logger.info(\"Got response [%s] for URL: %s\", resp.status, url)\n    html = await resp.text()\n    return html\nasync def parse(url: str, session: ClientSession, **kwargs) -&gt; set:\n    \"\"\"Find HREFs in the HTML of `url`.\"\"\"\n    found = set()\n    try:\n        html = await fetch_html(url=url, session=session, **kwargs)\n    except (\n        aiohttp.ClientError,\n        aiohttp.http_exceptions.HttpProcessingError,\n    ) as e:\n        logger.error(\n            \"aiohttp exception for %s [%s]: %s\",\n            url,\n            getattr(e, \"status\", None),\n            getattr(e, \"message\", None),\n        )\n        return found\n    except Exception as e:\n        logger.exception(\n            \"Non-aiohttp exception occured:  %s\", getattr(e, \"__dict__\", {})\n        )\n        return found\n    else:\n        for link in HREF_RE.findall(html):\n            try:\n                abslink = urllib.parse.urljoin(url, link)\n            except (urllib.error.URLError, ValueError):\n                logger.exception(\"Error parsing URL: %s\", link)\n                pass\n            else:\n                found.add(abslink)\n        logger.info(\"Found %d links for %s\", len(found), url)\n        return found\nasync def write_one(file: IO, url: str, **kwargs) -&gt; None:\n    \"\"\"Write the found HREFs from `url` to `file`.\"\"\"\n    res = await parse(url=url, **kwargs)\n    if not res:\n        return None\n    async with aiofiles.open(file, \"a\") as f:\n        for p in res:\n            await f.write(f\"{url}\\t{p}\\n\")\n        logger.info(\"Wrote results for source URL: %s\", url)\nasync def bulk_crawl_and_write(file: IO, urls: set, **kwargs) -&gt; None:\n    \"\"\"Crawl &amp; write concurrently to `file` for multiple `urls`.\"\"\"\n    async with ClientSession() as session:\n        tasks = []\n        for url in urls:\n            tasks.append(\n                write_one(file=file, url=url, session=session, **kwargs)\n            )\n        await asyncio.gather(*tasks)\nif __name__ == \"__main__\":\n    import pathlib\n    import sys\n    assert sys.version_info &gt;= (3, 7), \"Script requires Python 3.7+.\"\n    here = pathlib.Path(__file__).parent\n    with open(here.joinpath(\"urls.txt\")) as infile:\n        urls = set(map(str.strip, infile))\n    outpath = here.joinpath(\"foundurls.txt\")\n    with open(outpath, \"w\") as outfile:\n        outfile.write(\"source_url\\tparsed_url\\n\")\n    asyncio.run(bulk_crawl_and_write(file=outpath, urls=urls))\n<\/pre><\/div><p>This script is longer than our initial toy programs, so let\u2019s break it down.<\/p><p>The constant <code>HREF_RE<\/code> is a regular expression to extract what we\u2019re ultimately searching for, <code>href<\/code> tags within HTML:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; HREF_RE.search('Go to &lt;a href=\"https:\/\/realpython.com\/\"&gt;Real Python&lt;\/a&gt;')\n&lt;re.Match object; span=(15, 45), match='href=\"https:\/\/realpython.com\/\"'&gt;\n<\/pre><\/div><p>The coroutine <code>fetch_html()<\/code> is a wrapper around a GET request to make the request and decode the resulting page HTML. It makes the request, awaits the response, and raises right away in the case of a non-200 status:<\/p><div class=\"highlight python\"><pre>resp = await session.request(method=\"GET\", url=url, **kwargs)\nresp.raise_for_status()\n<\/pre><\/div><p>If the status is okay, <code>fetch_html()<\/code> returns the page HTML (a <code>str<\/code>). Notably, there is no exception handling done in this function. The logic is to propagate that exception to the caller and let it be handled there:<\/p><div class=\"highlight python\"><pre>html = await resp.text()\n<\/pre><\/div><p>We <code>await<\/code> <code>session.request()<\/code> and <code>resp.text()<\/code> because they\u2019re awaitable coroutines. The request\/response cycle would otherwise be the long-tailed, time-hogging portion of the application, but with async IO, <code>fetch_html()<\/code> lets the event loop work on other readily available jobs such as parsing and writing URLs that have already been fetched.<\/p><p>Next in the chain of coroutines comes <code>parse()<\/code>, which waits on <code>fetch_html()<\/code> for a given URL, and then extracts all of the <code>href<\/code> tags from that page\u2019s HTML, making sure that each is valid and formatting it as an absolute path.<\/p><p>Admittedly, the second portion of <code>parse()<\/code> is blocking, but it consists of a quick regex match and ensuring that the links discovered are made into absolute paths.<\/p><p>In this specific case, this synchronous code should be quick and inconspicuous. But just remember that any line within a given coroutine will block other coroutines unless that line uses <code>yield<\/code>, <code>await<\/code>, or <code>return<\/code>. If the parsing was a more intensive process, you might want to consider running this portion in its own process with <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#executing-code-in-thread-or-process-pools\"><code>loop.run_in_executor()<\/code><\/a>.<\/p><p>Next, the coroutine <code>write()<\/code> takes a file object and a single URL, and waits on <code>parse()<\/code> to return a <code>set<\/code> of the parsed URLs, writing each to the file asynchronously along with its source URL through use of <code>aiofiles<\/code>, a package for async file IO.<\/p><p>Lastly, <code>bulk_crawl_and_write()<\/code> serves as the main entry point into the script\u2019s chain of coroutines. It uses a single session, and a task is created for each URL that is ultimately read from <code>urls.txt<\/code>.<\/p><p>Here are a few additional points that deserve mention:<\/p><ul><li>\n<p>The default <code>ClientSession<\/code> has an <a href=\"https:\/\/aiohttp.readthedocs.io\/en\/stable\/client_reference.html#connectors\">adapter<\/a> with a maximum of 100 open connections. To change that, pass an instance of <code>asyncio.connector.TCPConnector<\/code> to <code>ClientSession<\/code>. You can also specify limits on a per-host basis.<\/p>\n<\/li>\n<li>\n<p>You can specify max <a href=\"https:\/\/aiohttp.readthedocs.io\/en\/stable\/client_quickstart.html#timeouts\">timeouts<\/a> for both the session as a whole and for individual requests.<\/p>\n<\/li>\n<li>\n<p>This script also uses <code>async with<\/code>, which works with an <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0492\/#asynchronous-context-managers-and-async-with\">asynchronous context manager<\/a>. I haven\u2019t devoted a whole section to this concept because the transition from synchronous to asynchronous context managers is fairly straightforward. The latter has to define <code>.__aenter__()<\/code> and <code>.__aexit__()<\/code> rather than <code>.__exit__()<\/code> and <code>.__enter__()<\/code>. As you might expect, <code>async with<\/code> can only be used inside a coroutine function declared with <code>async def<\/code>.<\/p>\n<\/li>\n<\/ul><p>If you\u2019d like to explore a bit more, the <a href=\"https:\/\/github.com\/realpython\/materials\/tree\/master\/asyncio-walkthrough\">companion files<\/a> for this tutorial up at GitHub have comments and docstrings attached as well.<\/p><p>Here\u2019s the execution in all of its glory, as <code>areq.py<\/code> gets, parses, and saves results for 9 URLs in under a second:<\/p><div class=\"highlight sh\"><pre>$ python3 areq.py\n21:33:22 DEBUG:asyncio: Using selector: KqueueSelector\n21:33:22 INFO:areq: Got response [200] for URL: https:\/\/www.mediamatters.org\/\n21:33:22 INFO:areq: Found 115 links for https:\/\/www.mediamatters.org\/\n21:33:22 INFO:areq: Got response [200] for URL: https:\/\/www.nytimes.com\/guides\/\n21:33:22 INFO:areq: Got response [200] for URL: https:\/\/www.politico.com\/tipsheets\/morning-money\n21:33:22 INFO:areq: Got response [200] for URL: https:\/\/www.ietf.org\/rfc\/rfc2616.txt\n21:33:22 ERROR:areq: aiohttp exception for https:\/\/docs.python.org\/3\/this-url-will-404.html [404]: Not Found\n21:33:22 INFO:areq: Found 120 links for https:\/\/www.nytimes.com\/guides\/\n21:33:22 INFO:areq: Found 143 links for https:\/\/www.politico.com\/tipsheets\/morning-money\n21:33:22 INFO:areq: Wrote results for source URL: https:\/\/www.mediamatters.org\/\n21:33:22 INFO:areq: Found 0 links for https:\/\/www.ietf.org\/rfc\/rfc2616.txt\n21:33:22 INFO:areq: Got response [200] for URL: https:\/\/1.1.1.1\/\n21:33:22 INFO:areq: Wrote results for source URL: https:\/\/www.nytimes.com\/guides\/\n21:33:22 INFO:areq: Wrote results for source URL: https:\/\/www.politico.com\/tipsheets\/morning-money\n21:33:22 INFO:areq: Got response [200] for URL: https:\/\/www.bloomberg.com\/markets\/economics\n21:33:22 INFO:areq: Found 3 links for https:\/\/www.bloomberg.com\/markets\/economics\n21:33:22 INFO:areq: Wrote results for source URL: https:\/\/www.bloomberg.com\/markets\/economics\n21:33:23 INFO:areq: Found 36 links for https:\/\/1.1.1.1\/\n21:33:23 INFO:areq: Got response [200] for URL: https:\/\/regex101.com\/\n21:33:23 INFO:areq: Found 23 links for https:\/\/regex101.com\/\n21:33:23 INFO:areq: Wrote results for source URL: https:\/\/regex101.com\/\n21:33:23 INFO:areq: Wrote results for source URL: https:\/\/1.1.1.1\/\n<\/pre><\/div><p>That\u2019s not too shabby! As a sanity check, you can check the line-count on the output. In my case, it\u2019s 626, though keep in mind this may fluctuate:<\/p><div class=\"highlight sh\"><pre>$ wc -l foundurls.txt\n     626 foundurls.txt\n$ head -n 3 foundurls.txt\nsource_url  parsed_url\nhttps:\/\/www.bloomberg.com\/markets\/economics https:\/\/www.bloomberg.com\/feedback\nhttps:\/\/www.bloomberg.com\/markets\/economics https:\/\/www.bloomberg.com\/notices\/tos\n<\/pre><\/div><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Next Steps<\/strong>: If you\u2019d like to up the ante, make this webcrawler recursive. You can use <a href=\"https:\/\/github.com\/aio-libs\/aioredis\"><code>aio-redis<\/code><\/a> to keep track of which URLs have been crawled within the tree to avoid requesting them twice, and connect links with Python\u2019s <code>networkx<\/code> library.<\/p><p>Remember to be nice. Sending 1000 concurrent requests to a small, unsuspecting website is bad, bad, bad. There are ways to limit how many concurrent requests you\u2019re making in one batch, such as in using the <a href=\"https:\/\/stackoverflow.com\/q\/40836800\/7954504\">sempahore<\/a> objects of <code>asyncio<\/code> or using a pattern <a href=\"https:\/\/www.artificialworlds.net\/blog\/2017\/05\/31\/python-3-large-numbers-of-tasks-with-limited-concurrency\/\">like this one<\/a>. If you don\u2019t heed this warning, you may get a massive batch of <code>TimeoutError<\/code> exceptions and only end up hurting your own program.<\/p><\/div><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h2 id=\"async-io-in-context\">Async IO in Context<\/h2><p>Now that you\u2019ve seen a healthy dose of code, let\u2019s step back for a minute and consider when async IO is an ideal option and how you can make the comparison to arrive at that conclusion or otherwise choose a different model of concurrency.<\/p><h3 id=\"when-and-why-is-async-io-the-right-choice\">When and Why Is Async IO the Right Choice?<\/h3><p>This tutorial is no place for an extended treatise on async IO versus threading versus multiprocessing. However, it\u2019s useful to have an idea of when async IO is probably the best candidate of the three.<\/p><p>The battle over async IO versus multiprocessing is not really a battle at all. In fact, they can be <a href=\"https:\/\/youtu.be\/0kXaLh8Fz3k?t=10m30s\">used in concert<\/a>. If you have multiple, fairly uniform CPU-bound tasks (a great example is a <a href=\"http:\/\/scikit-learn.org\/stable\/modules\/grid_search.html#parallelism\">grid search<\/a> in libraries such as <code>scikit-learn<\/code> or <code>keras<\/code>), multiprocessing should be an obvious choice.<\/p><p>Simply putting <code>async<\/code> before every function is a bad idea if all of the functions use blocking calls. (This can actually slow down your code.) But as mentioned previously, there are places where async IO and multiprocessing can <a href=\"https:\/\/youtu.be\/0kXaLh8Fz3k?t=10m30s\">live in harmony<\/a>.<\/p><p>The contest between async IO and threading is a little bit more direct. I mentioned in the introduction that \u201cthreading is hard.\u201d The full story is that, even in cases where threading seems easy to implement, it can still lead to infamous impossible-to-trace bugs due to race conditions and memory usage, among other things.<\/p><p>Threading also tends to scale less elegantly than async IO, because threads are a system resource with a finite availability. Creating thousands of threads will fail on many machines, and I don\u2019t recommend trying it in the first place. Creating thousands of async IO tasks is completely feasible.<\/p><p>Async IO shines when you have multiple IO-bound tasks where the tasks would otherwise be dominated by blocking IO-bound wait time, such as:<\/p><ul><li>\n<p>Network IO, whether your program is the server or the client side<\/p>\n<\/li>\n<li>\n<p>Serverless designs, such as a peer-to-peer, multi-user network like a group chatroom<\/p>\n<\/li>\n<li>\n<p>Read\/write operations where you want to mimic a \u201cfire-and-forget\u201d style but worry less about holding a lock on whatever you\u2019re reading and writing to<\/p>\n<\/li>\n<\/ul><p>The biggest reason not to use it is that <code>await<\/code> only supports a specific set of objects that define a specific set of methods. If you want to do async read operations with a certain DBMS, you\u2019ll need to find not just a Python wrapper for that DBMS, but one that supports the <code>async<\/code>\/<code>await<\/code> syntax. Coroutines that contain synchronous calls block other coroutines and tasks from running.<\/p><p>For a shortlist of libraries that work with <code>async<\/code>\/<code>await<\/code>, see the <a href=\"#libraries-that-work-with-asyncawait\">list<\/a> at the end of this tutorial.<\/p><h3 id=\"async-io-it-is-but-which-one\">Async IO It Is, but Which One?<\/h3><p>This tutorial focuses on async IO, the <code>async<\/code>\/<code>await<\/code> syntax, and using <code>asyncio<\/code> for event-loop management and specifying tasks. <code>asyncio<\/code> certainly isn\u2019t the only async IO library out there. This observation from Nathaniel J. Smith says a lot:<\/p><blockquote>\n<p>[In] a few years, <code>asyncio<\/code> might find itself relegated to becoming one of those stdlib libraries that savvy developers avoid, like <code>urllib2<\/code>.<\/p>\n<p>\u2026<\/p>\n<p>What I\u2019m arguing, in effect, is that <code>asyncio<\/code> is a victim of its own success: when it was designed, it used the best approach possible; but since then, work inspired by <code>asyncio<\/code> \u2013 like the addition of <code>async<\/code>\/<code>await<\/code> \u2013 has shifted the landscape so that we can do even better, and now <code>asyncio<\/code> is hamstrung by its earlier commitments. <a href=\"https:\/\/vorpus.org\/blog\/some-thoughts-on-asynchronous-api-design-in-a-post-asyncawait-world\/\">(Source)<\/a><\/p>\n<\/blockquote><p>To that end, a few big-name alternatives that do what <code>asyncio<\/code> does, albeit with different APIs and different approaches, are <a href=\"https:\/\/github.com\/dabeaz\/curio\"><code>curio<\/code><\/a> and <a href=\"https:\/\/github.com\/python-trio\/trio\"><code>trio<\/code><\/a>. Personally, I think that if you\u2019re building a moderately sized, straightforward program, just using <code>asyncio<\/code> is plenty sufficient and understandable, and lets you avoid adding yet another large dependency outside of Python\u2019s standard library.<\/p><p>But by all means, check out <code>curio<\/code> and <code>trio<\/code>, and you might find that they get the same thing done in a way that\u2019s more intuitive for you as the user. Many of the package-agnostic concepts presented here should permeate to alternative async IO packages as well.<\/p><h2 id=\"odds-and-ends\">Odds and Ends<\/h2><p>In these next few sections, you\u2019ll cover some miscellaneous parts of <code>asyncio<\/code> and <code>async<\/code>\/<code>await<\/code> that haven\u2019t fit neatly into the tutorial thus far, but are still important for building and understanding a full program.<\/p><p>In addition to <code>asyncio.run()<\/code>, you\u2019ve seen a few other package-level functions such as <code>asyncio.create_task()<\/code> and <code>asyncio.gather()<\/code>.<\/p><p>You can use <code>create_task()<\/code> to schedule the execution of a coroutine object, followed by <code>asyncio.run()<\/code>:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; import asyncio\n&gt;&gt;&gt; async def coro(seq) -&gt; list:\n...     \"\"\"'IO' wait time is proportional to the max element.\"\"\"\n...     await asyncio.sleep(max(seq))\n...     return list(reversed(seq))\n...\n&gt;&gt;&gt; async def main():\n...     # This is a bit redundant in the case of one task\n...     # We could use `await coro([3, 2, 1])` on its own\n...     t = asyncio.create_task(coro([3, 2, 1]))  # Python 3.7+\n...     await t\n...     print(f't: type {type(t)}')\n...     print(f't done: {t.done()}')\n...\n&gt;&gt;&gt; t = asyncio.run(main())\nt: type &lt;class '_asyncio.Task'&gt;\nt done: True\n<\/pre><\/div><p>There\u2019s a subtlety to this pattern: if you don\u2019t <code>await t<\/code> within <code>main()<\/code>, it may finish before <code>main()<\/code> itself signals that it is complete. Because <code>asyncio.run(main())<\/code> <a href=\"https:\/\/github.com\/python\/cpython\/blob\/7e18deef652a9d413d5dbd19d61073ba7eb5460e\/Lib\/asyncio\/runners.py#L43\">calls <code>loop.run_until_complete(main())<\/code><\/a>, the event loop is only concerned (without <code>await t<\/code> present) that <code>main()<\/code> is done, not that the tasks that get created within <code>main()<\/code> are done. Without <code>await t<\/code>, the loop\u2019s other tasks <a href=\"https:\/\/github.com\/python\/cpython\/blob\/7e18deef652a9d413d5dbd19d61073ba7eb5460e\/Lib\/asyncio\/runners.py#L46\">will be cancelled<\/a>, possibly before they are completed. If you need to get a list of currently pending tasks, you can use <code>asyncio.Task.all_tasks()<\/code>.<\/p><div class=\"alert alert-primary\" role=\"alert\"><p><strong>Note<\/strong>: <code>asyncio.create_task()<\/code> was introduced in Python 3.7. In Python 3.6 or lower, use <code>asyncio.ensure_future()<\/code> in place of <code>create_task()<\/code>.<\/p><\/div><p>Separately, there\u2019s <code>asyncio.gather()<\/code>. While it doesn\u2019t do anything tremendously special, <code>gather()<\/code> is meant to neatly put a collection of coroutines (futures) into a single future. As a result, it returns a single future object, and, if you <code>await asyncio.gather()<\/code> and specify multiple tasks or coroutines, you\u2019re waiting for all of them to be completed. (This somewhat parallels <code>queue.join()<\/code> from our earlier example.) The result of <code>gather()<\/code> will be a list of the results across the inputs:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; import time\n&gt;&gt;&gt; async def main():\n...     t = asyncio.create_task(coro([3, 2, 1]))\n...     t2 = asyncio.create_task(coro([10, 5, 0]))  # Python 3.7+\n...     print('Start:', time.strftime('%X'))\n...     a = await asyncio.gather(t, t2)\n...     print('End:', time.strftime('%X'))  # Should be 10 seconds\n...     print(f'Both tasks done: {all((t.done(), t2.done()))}')\n...     return a\n...\n&gt;&gt;&gt; a = asyncio.run(main())\nStart: 16:20:11\nEnd: 16:20:21\nBoth tasks done: True\n&gt;&gt;&gt; a\n[[1, 2, 3], [0, 5, 10]]\n<\/pre><\/div><p>You probably noticed that <code>gather()<\/code> waits on the entire result set of the Futures or coroutines that you pass it. Alternatively, you can loop over <code>asyncio.as_completed()<\/code> to get tasks as they are completed, in the order of completion. The function returns an iterator that yields tasks as they finish. Below, the result of <code>coro([3, 2, 1])<\/code> will be available before <code>coro([10, 5, 0])<\/code> is complete, which is not the case with <code>gather()<\/code>:<\/p><div class=\"highlight python repl\">&gt;&gt;&gt;<pre>&gt;&gt;&gt; async def main():\n...     t = asyncio.create_task(coro([3, 2, 1]))\n...     t2 = asyncio.create_task(coro([10, 5, 0]))\n...     print('Start:', time.strftime('%X'))\n...     for res in asyncio.as_completed((t, t2)):\n...         compl = await res\n...         print(f'res: {compl} completed at {time.strftime(\"%X\")}')\n...     print('End:', time.strftime('%X'))\n...     print(f'Both tasks done: {all((t.done(), t2.done()))}')\n...\n&gt;&gt;&gt; a = asyncio.run(main())\nStart: 09:49:07\nres: [1, 2, 3] completed at 09:49:10\nres: [0, 5, 10] completed at 09:49:17\nEnd: 09:49:17\nBoth tasks done: True\n<\/pre><\/div><p>Lastly, you may also see <code>asyncio.ensure_future()<\/code>. You should rarely need it, because it\u2019s a lower-level plumbing API and largely replaced by <code>create_task()<\/code>, which was introduced later.<\/p><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"the-precedence-of-await\">The Precedence of <code>await<\/code><\/h3><p>While they behave somewhat similarly, the <code>await<\/code> keyword has significantly higher precedence than <code>yield<\/code>. This means that, because it is more tightly bound, there are a number of instances where you\u2019d need parentheses in a <code>yield from<\/code> statement that are not required in an analogous <code>await<\/code> statement. For more information, see <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0492\/#examples-of-await-expressions\">examples of <code>await<\/code> expressions<\/a> from PEP 492.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>You\u2019re now equipped to use <code>async<\/code>\/<code>await<\/code> and the libraries built off of it. Here\u2019s a recap of what you\u2019ve covered:<\/p><ul><li>\n<p>Asynchronous IO as a language-agnostic model and a way to effect concurrency by letting coroutines indirectly communicate with each other<\/p>\n<\/li>\n<li>\n<p>The specifics of Python\u2019s new <code>async<\/code> and <code>await<\/code> keywords, used to mark and define coroutines<\/p>\n<\/li>\n<li>\n<p><code>asyncio<\/code>, the Python package that provides the API to run and manage coroutines<\/p>\n<\/li>\n<\/ul><h2 id=\"resources\">Resources<\/h2><h3 id=\"python-version-specifics\">Python Version Specifics<\/h3><p>Async IO in Python has evolved swiftly, and it can be hard to keep track of what came when. Here\u2019s a list of Python minor-version changes and introductions related to <code>asyncio<\/code>:<\/p><ul><li>\n<p>3.3: The <code>yield from<\/code> expression allows for generator delegation.<\/p>\n<\/li>\n<li>\n<p>3.4: <code>asyncio<\/code> was introduced in the Python standard library with provisional API status.<\/p>\n<\/li>\n<li>\n<p>3.5: <code>async<\/code> and <code>await<\/code> became a part of the Python grammar, used to signify and wait on coroutines. They were not yet reserved keywords. (You could still define functions or variables named <code>async<\/code> and <code>await<\/code>.)<\/p>\n<\/li>\n<li>\n<p>3.6: Asynchronous generators and asynchronous comprehensions were introduced. The API of <code>asyncio<\/code> was declared stable rather than provisional.<\/p>\n<\/li>\n<li>\n<p>3.7: <code>async<\/code> and <code>await<\/code> became reserved keywords. (They cannot be used as identifiers.) They are intended to replace the <code>asyncio.coroutine()<\/code> decorator. <code>asyncio.run()<\/code> was introduced to the <code>asyncio<\/code> package, among <a href=\"https:\/\/docs.python.org\/3\/whatsnew\/3.7.html#whatsnew37-asyncio\">a bunch of other features<\/a>.<\/p>\n<\/li>\n<\/ul><p>If you want to be safe (and be able to use <code>asyncio.run()<\/code>), go with Python 3.7 or above to get the full set of features.<\/p><h3 id=\"articles\">Articles<\/h3><p>Here\u2019s a curated list of additional resources:<\/p><ul><li>Real Python: <a href=\"https:\/\/realpython.com\/python-concurrency\/\">Speed up your Python Program with Concurrency<\/a><\/li>\n<li>Real Python: <a href=\"https:\/\/realpython.com\/python-gil\/\">What is the Python Global Interpreter Lock?<\/a><\/li>\n<li>CPython: The <code>asyncio<\/code> package <a href=\"https:\/\/github.com\/python\/cpython\/tree\/master\/Lib\/asyncio\">source<\/a><\/li>\n<li>Python docs: <a href=\"https:\/\/docs.python.org\/3\/reference\/datamodel.html#coroutines\">Data model &gt; Coroutines<\/a><\/li>\n<li>TalkPython: <a href=\"https:\/\/training.talkpython.fm\/courses\/details\/async-in-python-with-threading-and-multiprocessing\">Async Techniques and Examples in Python<\/a><\/li>\n<li>Brett Cannon: <a href=\"https:\/\/snarky.ca\/how-the-heck-does-async-await-work-in-python-3-5\/\">How the Heck Does Async-Await Work in Python 3.5?<\/a><\/li>\n<li>PYMOTW: <a href=\"https:\/\/pymotw.com\/3\/asyncio\/\"><code>asyncio<\/code><\/a><\/li>\n<li>A. Jesse Jiryu Davis and Guido van Rossum: <a href=\"http:\/\/aosabook.org\/en\/500L\/a-web-crawler-with-asyncio-coroutines.html\">A Web Crawler With asyncio Coroutines<\/a><\/li>\n<li>Andy Pearce: <a href=\"http:\/\/www.andy-pearce.com\/blog\/posts\/2016\/Jun\/the-state-of-python-coroutines-yield-from\/\">The State of Python Coroutines: <code>yield from<\/code><\/a><\/li>\n<li>Nathaniel J. Smith: <a href=\"https:\/\/vorpus.org\/blog\/some-thoughts-on-asynchronous-api-design-in-a-post-asyncawait-world\/\">Some Thoughts on Asynchronous API Design in a Post-<code>async<\/code>\/<code>await<\/code> World<\/a><\/li>\n<li>Armin Ronacher: <a href=\"http:\/\/lucumr.pocoo.org\/2016\/10\/30\/i-dont-understand-asyncio\/\">I don\u2019t understand Python\u2019s Asyncio<\/a><\/li>\n<li>Andy Balaam: <a href=\"http:\/\/www.artificialworlds.net\/blog\/2017\/05\/31\/basic-ideas-of-python-3-asyncio-concurrency\/\">series on <code>asyncio<\/code><\/a> (4 posts)<\/li>\n<li>Stack Overflow: <a href=\"https:\/\/stackoverflow.com\/q\/40836800\/7954504\">Python <code>asyncio.semaphore<\/code> in <code>async<\/code>-<code>await<\/code> function<\/a><\/li>\n<li>Yeray Diaz:\n<ul><li><a href=\"https:\/\/hackernoon.com\/asyncio-for-the-working-python-developer-5c468e6e2e8e\">AsyncIO for the Working Python Developer<\/a><\/li>\n<li><a href=\"https:\/\/medium.com\/python-pandemonium\/asyncio-coroutine-patterns-beyond-await-a6121486656f\">Asyncio Coroutine Patterns: Beyond <code>await<\/code><\/a><\/li>\n<\/ul><\/li>\n<\/ul><p>A few Python <em>What\u2019s New<\/em> sections explain the motivation behind language changes in more detail:<\/p><ul><li><a href=\"https:\/\/docs.python.org\/3\/whatsnew\/3.3.html#pep-380\">What\u2019s New in Python 3.3<\/a> (<code>yield from<\/code> and PEP 380)<\/li>\n<li><a href=\"https:\/\/docs.python.org\/3\/whatsnew\/3.6.html#whatsnew36-pep525\">What\u2019s New in Python 3.6<\/a> (PEP 525 &amp; 530)<\/li>\n<\/ul><p>From David Beazley:<\/p><ul><li><a href=\"http:\/\/www.dabeaz.com\/generators\/\">Generator: Tricks for Systems Programmers<\/a><\/li>\n<li><a href=\"http:\/\/www.dabeaz.com\/coroutines\/\">A Curious Course on Coroutines and Concurrency<\/a><\/li>\n<li><a href=\"http:\/\/dabeaz.com\/finalgenerator\/index.html\">Generators: The Final Frontier<\/a><\/li>\n<\/ul><p>YouTube talks:<\/p><ul><li><a href=\"https:\/\/youtu.be\/0kXaLh8Fz3k\">John Reese - Thinking Outside the GIL with AsyncIO and Multiprocessing - PyCon 2018<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/ZzfHjytDceU\">Keynote David Beazley - Topics of Interest (Python Asyncio)<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/MCs5OvhV9S4\">David Beazley - Python Concurrency From the Ground Up: LIVE! - PyCon 2015<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/9zinZmE3Ogk\">Raymond Hettinger, Keynote on Concurrency, PyBay 2017<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/Bv25Dwe84g0\">Thinking about Concurrency, Raymond Hettinger, Python core developer<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/iG6fr81xHKA\">Miguel Grinberg Asynchronous Python for the Complete Beginner PyCon 2017<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/2ZFFv-wZ8_g\">Yury Selivanov asyncawait and asyncio in Python 3 6 and beyond PyCon 2017<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/E-1Y4kSsAFc\">Fear and Awaiting in Async: A Savage Journey to the Heart of the Coroutine Dream<\/a><\/li>\n<li><a href=\"https:\/\/youtu.be\/kdzL3r-yJZY\">What Is Async, How Does It Work, and When Should I Use It? (PyCon APAC 2014)<\/a><\/li>\n<\/ul><h3 id=\"related-peps\">Related PEPs<\/h3><p><\/p><table class=\"table table-hover\"><thead><tr><th>PEP<\/th>\n<th>Date Created<\/th>\n<\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0342\/\">PEP 342 \u2013 Coroutines via Enhanced Generators<\/a><\/td>\n<td>2005-05<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0380\/\">PEP 380 \u2013 Syntax for Delegating to a Subgenerator<\/a><\/td>\n<td>2009-02<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-3153\/\">PEP 3153 \u2013 Asynchronous IO support<\/a><\/td>\n<td>2011-05<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-3156\/\">PEP 3156 \u2013 Asynchronous IO Support Rebooted: the \u201casyncio\u201d Module<\/a><\/td>\n<td>2012-12<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0492\/\">PEP 492 \u2013 Coroutines with async and await syntax<\/a><\/td>\n<td>2015-04<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0525\/\">PEP 525 \u2013 Asynchronous Generators<\/a><\/td>\n<td>2016-07<\/td>\n<\/tr><tr><td><a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0530\/\">PEP 530 \u2013 Asynchronous Comprehensions<\/a><\/td>\n<td>2016-09<\/td>\n<\/tr><\/tbody><\/table><p><a class=\"small text-muted js-disclosure c3\" href=\"https:\/\/realpython.com\/account\/join\/\" rel=\"nofollow\">Remove ads<\/a><\/p><h3 id=\"libraries-that-work-with-asyncawait\">Libraries That Work With <code>async<\/code>\/<code>await<\/code><\/h3><p>From <a href=\"https:\/\/github.com\/aio-libs\">aio-libs<\/a>:<\/p><ul><li><a href=\"https:\/\/github.com\/aio-libs\/aiohttp\"><code>aiohttp<\/code><\/a>: Asynchronous HTTP client\/server framework<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/aioredis\"><code>aioredis<\/code><\/a>: Async IO Redis support<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/aiopg\"><code>aiopg<\/code><\/a>: Async IO PostgreSQL support<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/aiomcache\"><code>aiomcache<\/code><\/a>: Async IO memcached client<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/aiokafka\"><code>aiokafka<\/code><\/a>: Async IO Kafka client<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/aiozmq\"><code>aiozmq<\/code><\/a>: Async IO ZeroMQ support<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/aiojobs\"><code>aiojobs<\/code><\/a>: Jobs scheduler for managing background tasks<\/li>\n<li><a href=\"https:\/\/github.com\/aio-libs\/async_lru\"><code>async_lru<\/code><\/a>: Simple LRU cache for async IO<\/li>\n<\/ul><p>From <a href=\"https:\/\/magic.io\/\">magicstack<\/a>:<\/p><ul><li><a href=\"https:\/\/github.com\/MagicStack\/uvloop\"><code>uvloop<\/code><\/a>: Ultra fast async IO event loop<\/li>\n<li><a href=\"https:\/\/github.com\/MagicStack\/asyncpg\"><code>asyncpg<\/code><\/a>: (Also very fast) async IO PostgreSQL support<\/li>\n<\/ul><p>From other hosts:<\/p><ul><li><a href=\"https:\/\/github.com\/python-trio\/trio\"><code>trio<\/code><\/a>: Friendlier <code>asyncio<\/code> intended to showcase a radically simpler design<\/li>\n<li><a href=\"https:\/\/github.com\/Tinche\/aiofiles\"><code>aiofiles<\/code><\/a>: Async file IO<\/li>\n<li><a href=\"https:\/\/github.com\/theelous3\/asks\"><code>asks<\/code><\/a>: Async requests-like http library<\/li>\n<li><a href=\"https:\/\/github.com\/jonathanslenders\/asyncio-redis\"><code>asyncio-redis<\/code><\/a>: Async IO Redis support<\/li>\n<li><a href=\"https:\/\/github.com\/dano\/aioprocessing\"><code>aioprocessing<\/code><\/a>: Integrates <code>multiprocessing<\/code> module with <code>asyncio<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Scille\/umongo\"><code>umongo<\/code><\/a>: Async IO MongoDB client<\/li>\n<li><a href=\"https:\/\/github.com\/alex-sherman\/unsync\"><code>unsync<\/code><\/a>: Unsynchronize <code>asyncio<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/vxgmichel\/aiostream\"><code>aiostream<\/code><\/a>: Like <code>itertools<\/code>, but async<\/li>\n<\/ul><div class=\"border rounded p-3 card mb-2\"><p class=\"mb-0\">Watch Now This tutorial has a related video course created by the Real Python team. Watch it together with the written tutorial to deepen your understanding: <a class=\"stretched-link text-success\" href=\"https:\/\/realpython.com\/courses\/python-3-concurrency-asyncio-module\/\"><strong>Hands-on Python 3 Concurrency With the asyncio Module<\/strong><\/a><\/p><\/div>","created_at":"2019-12-07T19:04:33+0000","updated_at":"2019-12-07T19:04:33+0000","published_at":"2019-01-16T14:00:00+0000","published_by":["Real Python"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":55,"domain_name":"realpython.com","preview_picture":"https:\/\/files.realpython.com\/media\/A-Complete-Walkthrough-of-Pythons-Asyncio_Watermarked.5b6b9a01fdc9.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1054,"title":"Coroutines and Tasks \u2014 Python 3.8.0 documentation","url":"https:\/\/docs.python.org\/3\/library\/asyncio-task.html","content":"<p>This section outlines high-level asyncio APIs to work with coroutines and Tasks.<\/p><div class=\"contents local topic\" id=\"contents\"><ul class=\"simple\"><li>\n<p><a class=\"reference internal\" href=\"#coroutines\" id=\"id1\">Coroutines<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#awaitables\" id=\"id2\">Awaitables<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#running-an-asyncio-program\" id=\"id3\">Running an asyncio Program<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#creating-tasks\" id=\"id4\">Creating Tasks<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#sleeping\" id=\"id5\">Sleeping<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#running-tasks-concurrently\" id=\"id6\">Running Tasks Concurrently<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#shielding-from-cancellation\" id=\"id7\">Shielding From Cancellation<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#timeouts\" id=\"id8\">Timeouts<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#waiting-primitives\" id=\"id9\">Waiting Primitives<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#scheduling-from-other-threads\" id=\"id10\">Scheduling From Other Threads<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#introspection\" id=\"id11\">Introspection<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#task-object\" id=\"id12\">Task Object<\/a><\/p>\n<\/li>\n<li>\n<p><a class=\"reference internal\" href=\"#generator-based-coroutines\" id=\"id13\">Generator-based Coroutines<\/a><\/p>\n<\/li>\n<\/ul><\/div><div class=\"section\" id=\"coroutines\"><p><a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/glossary.html#term-coroutine\">Coroutines<\/a> declared with the async\/await syntax is the preferred way of writing asyncio applications. For example, the following snippet of code (requires Python 3.7+) prints \u201chello\u201d, waits 1 second, and then prints \u201cworld\u201d:<\/p><div class=\"highlight-python3 notranslate highlight\"><pre>&gt;&gt;&gt; import asyncio\n&gt;&gt;&gt; async def main():\n...     print('hello')\n...     await asyncio.sleep(1)\n...     print('world')\n&gt;&gt;&gt; asyncio.run(main())\nhello\nworld\n<\/pre><\/div><p>Note that simply calling a coroutine will not schedule it to be executed:<\/p><div class=\"highlight-python3 notranslate highlight\"><pre>&gt;&gt;&gt; main()\n&lt;coroutine object main at 0x1053bb7c8&gt;\n<\/pre><\/div><p>To actually run a coroutine, asyncio provides three main mechanisms:<\/p><ul><li>\n<p>The <a class=\"reference internal\" href=\"#asyncio.run\" title=\"asyncio.run\"><code class=\"xref py py-func docutils literal notranslate\">asyncio.run()<\/code><\/a> function to run the top-level entry point \u201cmain()\u201d function (see the above example.)<\/p>\n<\/li>\n<li>\n<p>Awaiting on a coroutine. The following snippet of code will print \u201chello\u201d after waiting for 1 second, and then print \u201cworld\u201d after waiting for <em>another<\/em> 2 seconds:<\/p>\n<div class=\"highlight-python3 notranslate highlight\"><pre>import asyncio\nimport time\nasync def say_after(delay, what):\n    await asyncio.sleep(delay)\n    print(what)\nasync def main():\n    print(f\"started at {time.strftime('%X')}\")\n    await say_after(1, 'hello')\n    await say_after(2, 'world')\n    print(f\"finished at {time.strftime('%X')}\")\nasyncio.run(main())\n<\/pre><\/div>\n<p>Expected output:<\/p>\n<div class=\"highlight-python3 notranslate highlight\"><pre>started at 17:13:52\nhello\nworld\nfinished at 17:13:55\n<\/pre><\/div>\n<\/li>\n<li>\n<p>The <a class=\"reference internal\" href=\"#asyncio.create_task\" title=\"asyncio.create_task\"><code class=\"xref py py-func docutils literal notranslate\">asyncio.create_task()<\/code><\/a> function to run coroutines concurrently as asyncio <a class=\"reference internal\" href=\"#asyncio.Task\" title=\"asyncio.Task\"><code class=\"xref py py-class docutils literal notranslate\">Tasks<\/code><\/a>.<\/p>\n<p>Let\u2019s modify the above example and run two <code class=\"docutils literal notranslate\">say_after<\/code> coroutines <em>concurrently<\/em>:<\/p>\n<div class=\"highlight-python3 notranslate highlight\"><pre>async def main():\n    task1 = asyncio.create_task(\n        say_after(1, 'hello'))\n    task2 = asyncio.create_task(\n        say_after(2, 'world'))\n    print(f\"started at {time.strftime('%X')}\")\n    # Wait until both tasks are completed (should take\n    # around 2 seconds.)\n    await task1\n    await task2\n    print(f\"finished at {time.strftime('%X')}\")\n<\/pre><\/div>\n<p>Note that expected output now shows that the snippet runs 1 second faster than before:<\/p>\n<div class=\"highlight-python3 notranslate highlight\"><pre>started at 17:14:32\nhello\nworld\nfinished at 17:14:34\n<\/pre><\/div>\n<\/li>\n<\/ul><\/div><div class=\"section\" id=\"awaitables\"><p>We say that an object is an <strong>awaitable<\/strong> object if it can be used in an <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/reference\/expressions.html#await\"><code class=\"xref std std-keyword docutils literal notranslate\">await<\/code><\/a> expression. Many asyncio APIs are designed to accept awaitables.<\/p><p>There are three main types of <em>awaitable<\/em> objects: <strong>coroutines<\/strong>, <strong>Tasks<\/strong>, and <strong>Futures<\/strong>.<\/p><p class=\"rubric\">Coroutines<\/p><p>Python coroutines are <em>awaitables<\/em> and therefore can be awaited from other coroutines:<\/p><div class=\"highlight-python3 notranslate highlight\"><pre>import asyncio\nasync def nested():\n    return 42\nasync def main():\n    # Nothing happens if we just call \"nested()\".\n    # A coroutine object is created but not awaited,\n    # so it *won't run at all*.\n    nested()\n    # Let's do it differently now and await it:\n    print(await nested())  # will print \"42\".\nasyncio.run(main())\n<\/pre><\/div><div class=\"admonition important\"><p class=\"admonition-title\">Important<\/p><p>In this documentation the term \u201ccoroutine\u201d can be used for two closely related concepts:<\/p><ul class=\"simple\"><li>\n<p>a <em>coroutine function<\/em>: an <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/reference\/compound_stmts.html#async-def\"><code class=\"xref std std-keyword docutils literal notranslate\">async def<\/code><\/a> function;<\/p>\n<\/li>\n<li>\n<p>a <em>coroutine object<\/em>: an object returned by calling a <em>coroutine function<\/em>.<\/p>\n<\/li>\n<\/ul><\/div><p>asyncio also supports legacy <a class=\"reference internal\" href=\"#asyncio-generator-based-coro\">generator-based<\/a> coroutines.<\/p><p class=\"rubric\">Tasks<\/p><p><em>Tasks<\/em> are used to schedule coroutines <em>concurrently<\/em>.<\/p><p>When a coroutine is wrapped into a <em>Task<\/em> with functions like <a class=\"reference internal\" href=\"#asyncio.create_task\" title=\"asyncio.create_task\"><code class=\"xref py py-func docutils literal notranslate\">asyncio.create_task()<\/code><\/a> the coroutine is automatically scheduled to run soon:<\/p><div class=\"highlight-python3 notranslate highlight\"><pre>import asyncio\nasync def nested():\n    return 42\nasync def main():\n    # Schedule nested() to run soon concurrently\n    # with \"main()\".\n    task = asyncio.create_task(nested())\n    # \"task\" can now be used to cancel \"nested()\", or\n    # can simply be awaited to wait until it is complete:\n    await task\nasyncio.run(main())\n<\/pre><\/div><p class=\"rubric\">Futures<\/p><p>A <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future\" title=\"asyncio.Future\"><code class=\"xref py py-class docutils literal notranslate\">Future<\/code><\/a> is a special <strong>low-level<\/strong> awaitable object that represents an <strong>eventual result<\/strong> of an asynchronous operation.<\/p><p>When a Future object is <em>awaited<\/em> it means that the coroutine will wait until the Future is resolved in some other place.<\/p><p>Future objects in asyncio are needed to allow callback-based code to be used with async\/await.<\/p><p>Normally <strong>there is no need<\/strong> to create Future objects at the application level code.<\/p><p>Future objects, sometimes exposed by libraries and some asyncio APIs, can be awaited:<\/p><div class=\"highlight-python3 notranslate highlight\"><pre>async def main():\n    await function_that_returns_a_future_object()\n    # this is also valid:\n    await asyncio.gather(\n        function_that_returns_a_future_object(),\n        some_python_coroutine()\n    )\n<\/pre><\/div><p>A good example of a low-level function that returns a Future object is <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#asyncio.loop.run_in_executor\" title=\"asyncio.loop.run_in_executor\"><code class=\"xref py py-meth docutils literal notranslate\">loop.run_in_executor()<\/code><\/a>.<\/p><\/div><div class=\"section\" id=\"task-object\"><dl class=\"class\"><dt id=\"asyncio.Task\"><em class=\"property\">class<\/em> <code class=\"descclassname\">asyncio.<\/code><code class=\"descname\">Task<\/code>(<em>coro<\/em>, <em>*<\/em>, <em>loop=None<\/em>, <em>name=None<\/em>)<\/dt>\n<dd>\n<p>A <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future\" title=\"asyncio.Future\"><code class=\"xref py py-class docutils literal notranslate\">Future-like<\/code><\/a> object that runs a Python <a class=\"reference internal\" href=\"#coroutine\">coroutine<\/a>. Not thread-safe.<\/p>\n<p>Tasks are used to run coroutines in event loops. If a coroutine awaits on a Future, the Task suspends the execution of the coroutine and waits for the completion of the Future. When the Future is <em>done<\/em>, the execution of the wrapped coroutine resumes.<\/p>\n<p>Event loops use cooperative scheduling: an event loop runs one Task at a time. While a Task awaits for the completion of a Future, the event loop runs other Tasks, callbacks, or performs IO operations.<\/p>\n<p>Use the high-level <a class=\"reference internal\" href=\"#asyncio.create_task\" title=\"asyncio.create_task\"><code class=\"xref py py-func docutils literal notranslate\">asyncio.create_task()<\/code><\/a> function to create Tasks, or the low-level <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#asyncio.loop.create_task\" title=\"asyncio.loop.create_task\"><code class=\"xref py py-meth docutils literal notranslate\">loop.create_task()<\/code><\/a> or <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.ensure_future\" title=\"asyncio.ensure_future\"><code class=\"xref py py-func docutils literal notranslate\">ensure_future()<\/code><\/a> functions. Manual instantiation of Tasks is discouraged.<\/p>\n<p>To cancel a running Task use the <a class=\"reference internal\" href=\"#asyncio.Task.cancel\" title=\"asyncio.Task.cancel\"><code class=\"xref py py-meth docutils literal notranslate\">cancel()<\/code><\/a> method. Calling it will cause the Task to throw a <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.CancelledError\" title=\"asyncio.CancelledError\"><code class=\"xref py py-exc docutils literal notranslate\">CancelledError<\/code><\/a> exception into the wrapped coroutine. If a coroutine is awaiting on a Future object during cancellation, the Future object will be cancelled.<\/p>\n<p><a class=\"reference internal\" href=\"#asyncio.Task.cancelled\" title=\"asyncio.Task.cancelled\"><code class=\"xref py py-meth docutils literal notranslate\">cancelled()<\/code><\/a> can be used to check if the Task was cancelled. The method returns <code class=\"docutils literal notranslate\">True<\/code> if the wrapped coroutine did not suppress the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.CancelledError\" title=\"asyncio.CancelledError\"><code class=\"xref py py-exc docutils literal notranslate\">CancelledError<\/code><\/a> exception and was actually cancelled.<\/p>\n<p><a class=\"reference internal\" href=\"#asyncio.Task\" title=\"asyncio.Task\"><code class=\"xref py py-class docutils literal notranslate\">asyncio.Task<\/code><\/a> inherits from <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future\" title=\"asyncio.Future\"><code class=\"xref py py-class docutils literal notranslate\">Future<\/code><\/a> all of its APIs except <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future.set_result\" title=\"asyncio.Future.set_result\"><code class=\"xref py py-meth docutils literal notranslate\">Future.set_result()<\/code><\/a> and <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future.set_exception\" title=\"asyncio.Future.set_exception\"><code class=\"xref py py-meth docutils literal notranslate\">Future.set_exception()<\/code><\/a>.<\/p>\n<p>Tasks support the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/contextvars.html#module-contextvars\" title=\"contextvars: Context Variables\"><code class=\"xref py py-mod docutils literal notranslate\">contextvars<\/code><\/a> module. When a Task is created it copies the current context and later runs its coroutine in the copied context.<\/p>\n<div class=\"versionchanged\"><p>Changed in version 3.7: Added support for the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/contextvars.html#module-contextvars\" title=\"contextvars: Context Variables\"><code class=\"xref py py-mod docutils literal notranslate\">contextvars<\/code><\/a> module.<\/p><\/div>\n<div class=\"versionchanged\"><p>Changed in version 3.8: Added the <code class=\"docutils literal notranslate\">name<\/code> parameter.<\/p><\/div>\n<div class=\"deprecated-removed\"><p>Deprecated since version 3.8, will be removed in version 3.10: The <em>loop<\/em> parameter.<\/p><\/div>\n<dl class=\"method\"><dt id=\"asyncio.Task.cancel\"><code class=\"descname\">cancel<\/code>()<\/dt>\n<dd>\n<p>Request the Task to be cancelled.<\/p>\n<p>This arranges for a <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.CancelledError\" title=\"asyncio.CancelledError\"><code class=\"xref py py-exc docutils literal notranslate\">CancelledError<\/code><\/a> exception to be thrown into the wrapped coroutine on the next cycle of the event loop.<\/p>\n<p>The coroutine then has a chance to clean up or even deny the request by suppressing the exception with a <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/reference\/compound_stmts.html#try\"><code class=\"xref std std-keyword docutils literal notranslate\">try<\/code><\/a> \u2026 \u2026 <code class=\"docutils literal notranslate\">except CancelledError<\/code> \u2026 <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/reference\/compound_stmts.html#finally\"><code class=\"xref std std-keyword docutils literal notranslate\">finally<\/code><\/a> block. Therefore, unlike <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future.cancel\" title=\"asyncio.Future.cancel\"><code class=\"xref py py-meth docutils literal notranslate\">Future.cancel()<\/code><\/a>, <a class=\"reference internal\" href=\"#asyncio.Task.cancel\" title=\"asyncio.Task.cancel\"><code class=\"xref py py-meth docutils literal notranslate\">Task.cancel()<\/code><\/a> does not guarantee that the Task will be cancelled, although suppressing cancellation completely is not common and is actively discouraged.<\/p>\n<p id=\"asyncio-example-task-cancel\">The following example illustrates how coroutines can intercept the cancellation request:<\/p>\n<div class=\"highlight-python3 notranslate highlight\"><pre>async def cancel_me():\n    print('cancel_me(): before sleep')\n    try:\n        # Wait for 1 hour\n        await asyncio.sleep(3600)\n    except asyncio.CancelledError:\n        print('cancel_me(): cancel sleep')\n        raise\n    finally:\n        print('cancel_me(): after sleep')\nasync def main():\n    # Create a \"cancel_me\" Task\n    task = asyncio.create_task(cancel_me())\n    # Wait for 1 second\n    await asyncio.sleep(1)\n    task.cancel()\n    try:\n        await task\n    except asyncio.CancelledError:\n        print(\"main(): cancel_me is cancelled now\")\nasyncio.run(main())\n# Expected output:\n#\n#     cancel_me(): before sleep\n#     cancel_me(): cancel sleep\n#     cancel_me(): after sleep\n#     main(): cancel_me is cancelled now\n<\/pre><\/div>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.cancelled\"><code class=\"descname\">cancelled<\/code>()<\/dt>\n<dd>\n<p>Return <code class=\"docutils literal notranslate\">True<\/code> if the Task is <em>cancelled<\/em>.<\/p>\n<p>The Task is <em>cancelled<\/em> when the cancellation was requested with <a class=\"reference internal\" href=\"#asyncio.Task.cancel\" title=\"asyncio.Task.cancel\"><code class=\"xref py py-meth docutils literal notranslate\">cancel()<\/code><\/a> and the wrapped coroutine propagated the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.CancelledError\" title=\"asyncio.CancelledError\"><code class=\"xref py py-exc docutils literal notranslate\">CancelledError<\/code><\/a> exception thrown into it.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.done\"><code class=\"descname\">done<\/code>()<\/dt>\n<dd>\n<p>Return <code class=\"docutils literal notranslate\">True<\/code> if the Task is <em>done<\/em>.<\/p>\n<p>A Task is <em>done<\/em> when the wrapped coroutine either returned a value, raised an exception, or the Task was cancelled.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.result\"><code class=\"descname\">result<\/code>()<\/dt>\n<dd>\n<p>Return the result of the Task.<\/p>\n<p>If the Task is <em>done<\/em>, the result of the wrapped coroutine is returned (or if the coroutine raised an exception, that exception is re-raised.)<\/p>\n<p>If the Task has been <em>cancelled<\/em>, this method raises a <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.CancelledError\" title=\"asyncio.CancelledError\"><code class=\"xref py py-exc docutils literal notranslate\">CancelledError<\/code><\/a> exception.<\/p>\n<p>If the Task\u2019s result isn\u2019t yet available, this method raises a <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.InvalidStateError\" title=\"asyncio.InvalidStateError\"><code class=\"xref py py-exc docutils literal notranslate\">InvalidStateError<\/code><\/a> exception.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.exception\"><code class=\"descname\">exception<\/code>()<\/dt>\n<dd>\n<p>Return the exception of the Task.<\/p>\n<p>If the wrapped coroutine raised an exception that exception is returned. If the wrapped coroutine returned normally this method returns <code class=\"docutils literal notranslate\">None<\/code>.<\/p>\n<p>If the Task has been <em>cancelled<\/em>, this method raises a <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.CancelledError\" title=\"asyncio.CancelledError\"><code class=\"xref py py-exc docutils literal notranslate\">CancelledError<\/code><\/a> exception.<\/p>\n<p>If the Task isn\u2019t <em>done<\/em> yet, this method raises an <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-exceptions.html#asyncio.InvalidStateError\" title=\"asyncio.InvalidStateError\"><code class=\"xref py py-exc docutils literal notranslate\">InvalidStateError<\/code><\/a> exception.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.add_done_callback\"><code class=\"descname\">add_done_callback<\/code>(<em>callback<\/em>, <em>*<\/em>, <em>context=None<\/em>)<\/dt>\n<dd>\n<p>Add a callback to be run when the Task is <em>done<\/em>.<\/p>\n<p>This method should only be used in low-level callback-based code.<\/p>\n<p>See the documentation of <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future.add_done_callback\" title=\"asyncio.Future.add_done_callback\"><code class=\"xref py py-meth docutils literal notranslate\">Future.add_done_callback()<\/code><\/a> for more details.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.remove_done_callback\"><code class=\"descname\">remove_done_callback<\/code>(<em>callback<\/em>)<\/dt>\n<dd>\n<p>Remove <em>callback<\/em> from the callbacks list.<\/p>\n<p>This method should only be used in low-level callback-based code.<\/p>\n<p>See the documentation of <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-future.html#asyncio.Future.remove_done_callback\" title=\"asyncio.Future.remove_done_callback\"><code class=\"xref py py-meth docutils literal notranslate\">Future.remove_done_callback()<\/code><\/a> for more details.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.get_stack\"><code class=\"descname\">get_stack<\/code>(<em>*<\/em>, <em>limit=None<\/em>)<\/dt>\n<dd>\n<p>Return the list of stack frames for this Task.<\/p>\n<p>If the wrapped coroutine is not done, this returns the stack where it is suspended. If the coroutine has completed successfully or was cancelled, this returns an empty list. If the coroutine was terminated by an exception, this returns the list of traceback frames.<\/p>\n<p>The frames are always ordered from oldest to newest.<\/p>\n<p>Only one stack frame is returned for a suspended coroutine.<\/p>\n<p>The optional <em>limit<\/em> argument sets the maximum number of frames to return; by default all available frames are returned. The ordering of the returned list differs depending on whether a stack or a traceback is returned: the newest frames of a stack are returned, but the oldest frames of a traceback are returned. (This matches the behavior of the traceback module.)<\/p>\n<\/dd>\n<\/dl><dl class=\"method\">\n<dd>\n<p>Print the stack or traceback for this Task.<\/p>\n<p>This produces output similar to that of the traceback module for the frames retrieved by <a class=\"reference internal\" href=\"#asyncio.Task.get_stack\" title=\"asyncio.Task.get_stack\"><code class=\"xref py py-meth docutils literal notranslate\">get_stack()<\/code><\/a>.<\/p>\n<p>The <em>limit<\/em> argument is passed to <a class=\"reference internal\" href=\"#asyncio.Task.get_stack\" title=\"asyncio.Task.get_stack\"><code class=\"xref py py-meth docutils literal notranslate\">get_stack()<\/code><\/a> directly.<\/p>\n<p>The <em>file<\/em> argument is an I\/O stream to which the output is written; by default output is written to <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/sys.html#sys.stderr\" title=\"sys.stderr\"><code class=\"xref py py-data docutils literal notranslate\">sys.stderr<\/code><\/a>.<\/p>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.get_coro\"><code class=\"descname\">get_coro<\/code>()<\/dt>\n<dd>\n<p>Return the coroutine object wrapped by the <a class=\"reference internal\" href=\"#asyncio.Task\" title=\"asyncio.Task\"><code class=\"xref py py-class docutils literal notranslate\">Task<\/code><\/a>.<\/p>\n<div class=\"versionadded\"><p>New in version 3.8.<\/p><\/div>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.get_name\"><code class=\"descname\">get_name<\/code>()<\/dt>\n<dd>\n<p>Return the name of the Task.<\/p>\n<p>If no name has been explicitly assigned to the Task, the default asyncio Task implementation generates a default name during instantiation.<\/p>\n<div class=\"versionadded\"><p>New in version 3.8.<\/p><\/div>\n<\/dd>\n<\/dl><dl class=\"method\"><dt id=\"asyncio.Task.set_name\"><code class=\"descname\">set_name<\/code>(<em>value<\/em>)<\/dt>\n<dd>\n<p>Set the name of the Task.<\/p>\n<p>The <em>value<\/em> argument can be any object, which is then converted to a string.<\/p>\n<p>In the default Task implementation, the name will be visible in the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/functions.html#repr\" title=\"repr\"><code class=\"xref py py-func docutils literal notranslate\">repr()<\/code><\/a> output of a task object.<\/p>\n<div class=\"versionadded\"><p>New in version 3.8.<\/p><\/div>\n<\/dd>\n<\/dl><dl class=\"classmethod\"><dt id=\"asyncio.Task.all_tasks\"><em class=\"property\">classmethod<\/em> <code class=\"descname\">all_tasks<\/code>(<em>loop=None<\/em>)<\/dt>\n<dd>\n<p>Return a set of all tasks for an event loop.<\/p>\n<p>By default all tasks for the current event loop are returned. If <em>loop<\/em> is <code class=\"docutils literal notranslate\">None<\/code>, the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#asyncio.get_event_loop\" title=\"asyncio.get_event_loop\"><code class=\"xref py py-func docutils literal notranslate\">get_event_loop()<\/code><\/a> function is used to get the current loop.<\/p>\n<div class=\"deprecated-removed\"><p>Deprecated since version 3.7, will be removed in version 3.9: Do not call this as a task method. Use the <a class=\"reference internal\" href=\"#asyncio.all_tasks\" title=\"asyncio.all_tasks\"><code class=\"xref py py-func docutils literal notranslate\">asyncio.all_tasks()<\/code><\/a> function instead.<\/p><\/div>\n<\/dd>\n<\/dl><dl class=\"classmethod\"><dt id=\"asyncio.Task.current_task\"><em class=\"property\">classmethod<\/em> <code class=\"descname\">current_task<\/code>(<em>loop=None<\/em>)<\/dt>\n<dd>\n<p>Return the currently running task or <code class=\"docutils literal notranslate\">None<\/code>.<\/p>\n<p>If <em>loop<\/em> is <code class=\"docutils literal notranslate\">None<\/code>, the <a class=\"reference internal\" href=\"https:\/\/docs.python.org\/3\/library\/asyncio-eventloop.html#asyncio.get_event_loop\" title=\"asyncio.get_event_loop\"><code class=\"xref py py-func docutils literal notranslate\">get_event_loop()<\/code><\/a> function is used to get the current loop.<\/p>\n<div class=\"deprecated-removed\"><p>Deprecated since version 3.7, will be removed in version 3.9: Do not call this as a task method. Use the <a class=\"reference internal\" href=\"#asyncio.current_task\" title=\"asyncio.current_task\"><code class=\"xref py py-func docutils literal notranslate\">asyncio.current_task()<\/code><\/a> function instead.<\/p><\/div>\n<\/dd>\n<\/dl><\/dd>\n<\/dl><\/div><div class=\"section\" id=\"generator-based-coroutines\"><div class=\"admonition note\"><p class=\"admonition-title\">Note<\/p><p>Support for generator-based coroutines is <strong>deprecated<\/strong> and is scheduled for removal in Python 3.10.<\/p><\/div><p>Generator-based coroutines predate async\/await syntax. They are Python generators that use <code class=\"docutils literal notranslate\">yield from<\/code> expressions to await on Futures and other coroutines.<\/p><p>Generator-based coroutines should be decorated with <a class=\"reference internal\" href=\"#asyncio.coroutine\" title=\"asyncio.coroutine\"><code class=\"xref py py-func docutils literal notranslate\">@asyncio.coroutine<\/code><\/a>, although this is not enforced.<\/p><\/div>","created_at":"2019-12-07T17:23:08+0000","updated_at":"2019-12-07T17:23:08+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"docs.python.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1053,"title":"encode\/starlette","url":"https:\/\/github.com\/encode\/starlette","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p style=\"text-align: center;\"><a href=\"https:\/\/www.starlette.io\/\" rel=\"nofollow\"><img width=\"420px\" src=\"https:\/\/raw.githubusercontent.com\/encode\/starlette\/master\/docs\/img\/starlette.png\" alt=\"starlette\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p style=\"text-align: center;\"><em>\u2728 The little ASGI framework that shines. \u2728<\/em><\/p>\n<p style=\"text-align: center;\"><a href=\"https:\/\/travis-ci.org\/encode\/starlette\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/55249beea9d5375ce4551686496f8bc2ec41c83f\/68747470733a2f2f7472617669732d63692e6f72672f656e636f64652f737461726c657474652e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/encode\/starlette.svg?branch=master\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/codecov.io\/gh\/encode\/starlette\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/cbd567ea07608d19226b0cb37e400af4a2f066e3\/68747470733a2f2f636f6465636f762e696f2f67682f656e636f64652f737461726c657474652f6272616e63682f6d61737465722f67726170682f62616467652e737667\" alt=\"Coverage\" data-canonical-src=\"https:\/\/codecov.io\/gh\/encode\/starlette\/branch\/master\/graph\/badge.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/pypi.org\/project\/starlette\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/3b392e5c7245158cb4369bbe50b2a9bcc7538157\/68747470733a2f2f62616467652e667572792e696f2f70792f737461726c657474652e737667\" alt=\"Package version\" data-canonical-src=\"https:\/\/badge.fury.io\/py\/starlette.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<hr \/><p><strong>Documentation<\/strong>: <a href=\"https:\/\/www.starlette.io\/\" rel=\"nofollow\">https:\/\/www.starlette.io\/<\/a><\/p>\n<p><strong>Community<\/strong>: <a href=\"https:\/\/discuss.encode.io\/c\/starlette\" rel=\"nofollow\">https:\/\/discuss.encode.io\/c\/starlette<\/a><\/p>\n<hr \/>\n<p>Starlette is a lightweight <a href=\"https:\/\/asgi.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">ASGI<\/a> framework\/toolkit, which is ideal for building high performance asyncio services.<\/p>\n<p>It is production-ready, and gives you the following:<\/p>\n<ul><li>Seriously impressive performance.<\/li>\n<li>WebSocket support.<\/li>\n<li>GraphQL support.<\/li>\n<li>In-process background tasks.<\/li>\n<li>Startup and shutdown events.<\/li>\n<li>Test client built on <code>requests<\/code>.<\/li>\n<li>CORS, GZip, Static Files, Streaming responses.<\/li>\n<li>Session and Cookie support.<\/li>\n<li>100% test coverage.<\/li>\n<li>100% type annotated codebase.<\/li>\n<li>Zero hard dependencies.<\/li>\n<\/ul><h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<p>Python 3.6+<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<div class=\"highlight highlight-source-shell\">\n<pre>$ pip3 install starlette<\/pre><\/div>\n<p>You'll also want to install an ASGI server, such as <a href=\"http:\/\/www.uvicorn.org\/\" rel=\"nofollow\">uvicorn<\/a>, <a href=\"https:\/\/github.com\/django\/daphne\/\">daphne<\/a>, or <a href=\"https:\/\/pgjones.gitlab.io\/hypercorn\/\" rel=\"nofollow\">hypercorn<\/a>.<\/p>\n<div class=\"highlight highlight-source-shell\">\n<pre>$ pip3 install uvicorn<\/pre><\/div>\n<h2><a id=\"user-content-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#example\"><\/a>Example<\/h2>\n<p><strong>example.py<\/strong>:<\/p>\n<div class=\"highlight highlight-source-python\">\n<pre>from starlette.applications import Starlette\nfrom starlette.responses import JSONResponse\nfrom starlette.routing import Route\nasync def homepage(request):\n    return JSONResponse({'hello': 'world'})\nroutes = [\n    Route(\"\/\", endpoint=homepage)\n]\napp = Starlette(debug=True, routes=routes)<\/pre><\/div>\n<p>Then run the application using Uvicorn:<\/p>\n<div class=\"highlight highlight-source-shell\">\n<pre>$ uvicorn example:app<\/pre><\/div>\n<p>For a more complete example, see <a href=\"https:\/\/github.com\/encode\/starlette-example\">encode\/starlette-example<\/a>.<\/p>\n<h2><a id=\"user-content-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>Dependencies<\/h2>\n<p>Starlette does not have any hard dependencies, but the following are optional:<\/p>\n<ul><li><a href=\"http:\/\/docs.python-requests.org\/en\/master\/\" rel=\"nofollow\"><code>requests<\/code><\/a> - Required if you want to use the <code>TestClient<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/Tinche\/aiofiles\"><code>aiofiles<\/code><\/a> - Required if you want to use <code>FileResponse<\/code> or <code>StaticFiles<\/code>.<\/li>\n<li><a href=\"http:\/\/jinja.pocoo.org\/\" rel=\"nofollow\"><code>jinja2<\/code><\/a> - Required if you want to use <code>Jinja2Templates<\/code>.<\/li>\n<li><a href=\"https:\/\/andrew-d.github.io\/python-multipart\/\" rel=\"nofollow\"><code>python-multipart<\/code><\/a> - Required if you want to support form parsing, with <code>request.form()<\/code>.<\/li>\n<li><a href=\"https:\/\/pythonhosted.org\/itsdangerous\/\" rel=\"nofollow\"><code>itsdangerous<\/code><\/a> - Required for <code>SessionMiddleware<\/code> support.<\/li>\n<li><a href=\"https:\/\/pyyaml.org\/wiki\/PyYAMLDocumentation\" rel=\"nofollow\"><code>pyyaml<\/code><\/a> - Required for <code>SchemaGenerator<\/code> support.<\/li>\n<li><a href=\"https:\/\/graphene-python.org\/\" rel=\"nofollow\"><code>graphene<\/code><\/a> - Required for <code>GraphQLApp<\/code> support.<\/li>\n<li><a href=\"https:\/\/github.com\/esnme\/ultrajson\"><code>ujson<\/code><\/a> - Required if you want to use <code>UJSONResponse<\/code>.<\/li>\n<\/ul><p>You can install all of these with <code>pip3 install starlette[full]<\/code>.<\/p>\n<h2><a id=\"user-content-framework-or-toolkit\" class=\"anchor\" aria-hidden=\"true\" href=\"#framework-or-toolkit\"><\/a>Framework or Toolkit<\/h2>\n<p>Starlette is designed to be used either as a complete framework, or as an ASGI toolkit. You can use any of its components independently.<\/p>\n<div class=\"highlight highlight-source-python\">\n<pre>from starlette.responses import PlainTextResponse\nasync def app(scope, receive, send):\n    assert scope['type'] == 'http'\n    response = PlainTextResponse('Hello, world!')\n    await response(scope, receive, send)<\/pre><\/div>\n<p>Run the <code>app<\/code> application in <code>example.py<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\">\n<pre>$ uvicorn example:app\nINFO: Started server process [11509]\nINFO: Uvicorn running on http:\/\/127.0.0.1:8000 (Press CTRL+C to quit)<\/pre><\/div>\n<p>Run uvicorn with <code>--reload<\/code> to enable auto-reloading on code changes.<\/p>\n<h2><a id=\"user-content-modularity\" class=\"anchor\" aria-hidden=\"true\" href=\"#modularity\"><\/a>Modularity<\/h2>\n<p>The modularity that Starlette is designed on promotes building re-usable components that can be shared between any ASGI framework. This should enable an ecosystem of shared middleware and mountable applications.<\/p>\n<p>The clean API separation also means it's easier to understand each component in isolation.<\/p>\n<h2><a id=\"user-content-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance\"><\/a>Performance<\/h2>\n<p>Independent TechEmpower benchmarks show Starlette applications running under Uvicorn as <a href=\"https:\/\/www.techempower.com\/benchmarks\/#section=data-r17&amp;hw=ph&amp;test=fortune&amp;l=zijzen-1\" rel=\"nofollow\">one of the fastest Python frameworks available<\/a>. <em>(*)<\/em><\/p>\n<p>For high throughput loads you should:<\/p>\n<ul><li>Make sure to install <code>ujson<\/code> and use <code>UJSONResponse<\/code>.<\/li>\n<li>Run using gunicorn using the <code>uvicorn<\/code> worker class.<\/li>\n<li>Use one or two workers per-CPU core. (You might need to experiment with this.)<\/li>\n<li>Disable access logging.<\/li>\n<\/ul><p>Eg.<\/p>\n<div class=\"highlight highlight-source-shell\">\n<pre>gunicorn -w 4 -k uvicorn.workers.UvicornWorker --log-level warning example:app<\/pre><\/div>\n<p>Several of the ASGI servers also have pure Python implementations available, so you can also run under <code>PyPy<\/code> if your application code has parts that are CPU constrained.<\/p>\n<p>Either programatically:<\/p>\n<div class=\"highlight highlight-source-python\">\n<pre>uvicorn.run(..., http='h11', loop='asyncio')<\/pre><\/div>\n<p>Or using Gunicorn:<\/p>\n<div class=\"highlight highlight-source-shell\">\n<pre>gunicorn -k uvicorn.workers.UvicornH11Worker ...<\/pre><\/div>\n<p style=\"text-align: center;\">\u2014 \u2b50\ufe0f \u2014<\/p>\n<p style=\"text-align: center;\"><em>Starlette is <a href=\"https:\/\/github.com\/encode\/starlette\/blob\/master\/LICENSE.md\">BSD licensed<\/a> code. Designed &amp; built in Brighton, England.<\/em><\/p>\n<\/article>","created_at":"2019-12-07T17:18:21+0000","updated_at":"2019-12-07T17:18:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/19159390?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1051,"title":"Hexing the Technical Interview (2017)","url":"https:\/\/news.ycombinator.com\/item?id=21717039","content":"<table border=\"0\" cellpadding=\"0\" class=\"c5\" style=\"border-spacing: 0px; width: 100%;\"><tr><td class=\"c2\"><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" class=\"c1\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/a><\/td>\n<td class=\"c3\"><strong class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/strong> <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a><\/td>\n<td class=\"c4\"><a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D21717039\">login<\/a><\/td>\n<\/tr><\/table>\n\n<table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"21717039\"><td align=\"right\" valign=\"top\" class=\"title\">\n<\/td><td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"title\"><a href=\"https:\/\/aphyr.com\/posts\/341-hexing-the-technical-interview\" class=\"storylink\">Hexing the Technical Interview (2017)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=aphyr.com\">aphyr.com<\/a>)<\/td>\n<\/tr><tr><td class=\"subtext\">247 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=kick\" class=\"hnuser\">kick<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21717039\">5 hours ago<\/a> | <a href=\"https:\/\/news.ycombinator.com\/hide?id=21717039&amp;goto=item%3Fid%3D21717039\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Hexing%20the%20Technical%20Interview&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Hexing%20the%20Technical%20Interview\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=21717039&amp;auth=658ee545e3ed1ab143e5acf9f9e6ffdab3af7a2d\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=21717039\">57\u00a0comments<\/a><\/td>\n<\/tr><tr><td>\n<\/td>\n<\/tr><\/table>\n<img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><br \/><div class=\"c7\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>|<a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>|<a href=\"mailto:hn@ycombinator.com\">Support<\/a>|<a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>|<a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>|<a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>|<a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>|<a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>|<a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>|<a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\">\n<p>Search: <\/p>\n<\/form><\/div>","created_at":"2019-12-06T03:54:16+0000","updated_at":"2019-12-06T03:54:16+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1050,"title":"Clojure from the ground up","url":"https:\/\/aphyr.com\/tags\/Clojure-from-the-ground-up","content":"<p>This guide aims to introduce newcomers and experienced programmers alike to the beauty of functional programming, starting with the simplest building blocks of software. You\u2019ll need a computer, basic proficiency in the command line, a text editor, and an internet connection. By the end of this series, you\u2019ll have a thorough command of the Clojure programming language.<\/p>\n<p>Science, technology, engineering, and mathematics are deeply rewarding fields, yet few women enter STEM as a career path. Still more are discouraged by a culture which repeatedly asserts that women lack the analytic aptitude for writing software, that they are not driven enough to be successful scientists, that it\u2019s not cool to pursue a passion for structural engineering. Those few with the talent, encouragement, and persistence to break in to science and tech are discouraged by persistent sexism in practice: the old boy\u2019s club of tenure, being passed over for promotions, isolation from peers, and flat-out assault. This landscape sucks. I want to help change it.<\/p>\n<p><a href=\"https:\/\/twitter.com\/WomenWhoCode\">Women Who Code<\/a>, <a href=\"http:\/\/www.pyladies.com\/\">PyLadies<\/a>, <a href=\"http:\/\/www.blackgirlscode.com\/\">Black Girls Code<\/a>, <a href=\"http:\/\/railsbridge.org\/\">RailsBridge<\/a>, <a href=\"http:\/\/www.girlswhocode.com\/about-us\/\">Girls Who Code<\/a>, <a href=\"http:\/\/www.girldevelopit.com\/\">Girl Develop It<\/a>, and <a href=\"http:\/\/www.lambdaladies.com\/\">Lambda Ladies<\/a> are just a few of the fantastic groups helping women enter and thrive in software. I wholeheartedly support these efforts.<\/p>","created_at":"2019-12-06T03:27:44+0000","updated_at":"2019-12-06T03:27:48+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"aphyr.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1049,"title":"1u7flclharvMchwWHY7Ya41NKjX3dkslu","url":"https:\/\/colab.research.google.com\/drive\/1u7flclharvMchwWHY7Ya41NKjX3dkslu#forceEdit=true&sandboxMode=true&scrollTo=YjArwbWh6XwN","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.","created_at":"2019-12-05T23:14:26+0000","updated_at":"2019-12-05T23:14:27+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"colab.research.google.com"},{"is_archived":0,"is_starred":0,"tags":["selfhosted","analytics"],"is_public":false,"id":1048,"title":"matomo-org\/matomo","url":"https:\/\/github.com\/matomo-org\/matomo","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p><a href=\"https:\/\/matomo.org\/download\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d8d88d71a5c724e2759dc39f2cdbae66e8beeb7d\/68747470733a2f2f706f7365722e707567782e6f72672f706977696b2f706977696b2f762f737461626c65\" alt=\"Latest Stable Version\" data-canonical-src=\"https:\/\/poser.pugx.org\/piwik\/piwik\/v\/stable\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/packagist.org\/packages\/piwik\/piwik\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/c05066bbf6331f903ec65158f56e3285bcfc446b\/68747470733a2f2f706f7365722e707567782e6f72672f706977696b2f706977696b2f762f756e737461626c65\" alt=\"Latest Unstable Version\" data-canonical-src=\"https:\/\/poser.pugx.org\/piwik\/piwik\/v\/unstable\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/matomo.org\/free-software\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5b83ce52f6b9b45c75d8a38b0a74da9d22f3a839\/68747470733a2f2f706f7365722e707567782e6f72672f706977696b2f706977696b2f6c6963656e7365\" alt=\"License\" data-canonical-src=\"https:\/\/poser.pugx.org\/piwik\/piwik\/license\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h2><a id=\"user-content-code-status\" class=\"anchor\" aria-hidden=\"true\" href=\"#code-status\"><\/a>Code Status<\/h2>\n<p><a href=\"https:\/\/travis-ci.org\/matomo-org\/matomo\/branches\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/416a364bf0caf53884078da4b84559f578125d43\/68747470733a2f2f7472617669732d63692e6f72672f6d61746f6d6f2d6f72672f6d61746f6d6f2e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/matomo-org\/matomo.svg?branch=master\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/isitmaintained.com\/project\/matomo-org\/matomo\" title=\"Percentage of issues still open\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f9997bca64263e7a149cdaf7eab5a764391743c4\/687474703a2f2f697369746d61696e7461696e65642e636f6d2f62616467652f6f70656e2f6d61746f6d6f2d6f72672f6d61746f6d6f2e737667\" alt=\"Percentage of issues still open\" data-canonical-src=\"http:\/\/isitmaintained.com\/badge\/open\/matomo-org\/matomo.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h2><a id=\"user-content-description\" class=\"anchor\" aria-hidden=\"true\" href=\"#description\"><\/a>Description<\/h2>\n<p>Matomo is the leading Free\/Libre open analytics platform.<\/p>\n<p>Matomo is a full-featured PHP MySQL software program that you download and install on your own webserver. At the end of the five-minute installation process, you will be given a JavaScript code. Simply copy and paste this tag on websites you wish to track and access your analytics reports in real-time.<\/p>\n<p>Matomo aims to be a Free software alternative to Google Analytics and is already used on more than 1,400,000 websites. Privacy is built-in!<\/p>\n<h2><a id=\"user-content-mission-statement\" class=\"anchor\" aria-hidden=\"true\" href=\"#mission-statement\"><\/a>Mission Statement<\/h2>\n<blockquote>\n<p>\u00ab To create, as a community, the leading international open source digital analytics platform, that gives every user full control of their data. \u00bb<\/p>\n<\/blockquote>\n<p>Or in short:<\/p>\n<blockquote>\n<p>\u00ab Liberate Web Analytics \u00bb<\/p>\n<\/blockquote>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Matomo is released under the GPL v3 (or later) license, see <a href=\"https:\/\/github.com\/matomo-org\/matomo\/blob\/3.x-dev\/LICENSE\">LICENSE<\/a><\/p>\n<h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<ul><li>PHP 5.5.9 or greater<\/li>\n<li>MySQL version 5.5 or greater, or MariaDB<\/li>\n<li>PHP extension pdo and pdo_mysql, or the MySQLi extension.<\/li>\n<li>Matomo is OS \/ server independent<\/li>\n<\/ul><p>See <a href=\"https:\/\/matomo.org\/docs\/requirements\/\" rel=\"nofollow\">https:\/\/matomo.org\/docs\/requirements\/<\/a><\/p>\n<h2><a id=\"user-content-install-matomo\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-matomo\"><\/a>Install Matomo<\/h2>\n<ul><li><a href=\"https:\/\/matomo.org\/download\/\" rel=\"nofollow\">Download Matomo<\/a><\/li>\n<li>Upload matomo to your webserver<\/li>\n<li>Point your browser to the directory<\/li>\n<li>Follow the steps<\/li>\n<li>Add the given javascript code to your pages<\/li>\n<li>(You may also generate fake data to experiment, by enabling the plugin VisitorGenerator)<\/li>\n<\/ul><p>See <a href=\"https:\/\/matomo.org\/docs\/installation\/\" rel=\"nofollow\">https:\/\/matomo.org\/docs\/installation\/<\/a><\/p>\n<p>(When using Matomo for development you need to <a href=\"https:\/\/matomo.org\/faq\/how-to-install\/faq_18271\/\" rel=\"nofollow\">install Matomo from the Git repository.<\/a>)<\/p>\n<h2><a id=\"user-content-free-trial\" class=\"anchor\" aria-hidden=\"true\" href=\"#free-trial\"><\/a>Free trial<\/h2>\n<p>If you do not have a server or don't want to host yourself you can use our Matomo Cloud partner service (30 day free trial): <a href=\"https:\/\/matomo.org\/start-30-day-free-analytics-trial\/\" rel=\"nofollow\">https:\/\/matomo.org\/start-30-day-free-analytics-trial\/<\/a><\/p>\n<h2><a id=\"user-content-online-demo\" class=\"anchor\" aria-hidden=\"true\" href=\"#online-demo\"><\/a>Online Demo<\/h2>\n<p>Check out the online demo for Matomo at <a href=\"https:\/\/demo.matomo.org\/\" rel=\"nofollow\">demo.matomo.org<\/a><\/p>\n<h2><a id=\"user-content-changelog\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelog\"><\/a>Changelog<\/h2>\n<p>For the list of all tickets closed in the current and past releases, see <a href=\"https:\/\/matomo.org\/changelog\/\" rel=\"nofollow\">matomo.org\/changelog\/<\/a>. For the list of technical changes in the Matomo platform, see <a href=\"https:\/\/developer.matomo.org\/changelog\" rel=\"nofollow\">developer.matomo.org\/changelog<\/a>.<\/p>\n<h2><a id=\"user-content-get-involved\" class=\"anchor\" aria-hidden=\"true\" href=\"#get-involved\"><\/a>Get involved!<\/h2>\n<p>We believe in liberating Web Analytics, providing a free platform for simple and advanced analytics. Matomo was built by dozens of people like you, and we need your help to make Matomo better\u2026 Why not participate in a useful project today? <a href=\"https:\/\/matomo.org\/get-involved\" rel=\"nofollow\">Learn how you can contribute to Matomo.<\/a><\/p>\n<h2><a id=\"user-content-quality-assurance\" class=\"anchor\" aria-hidden=\"true\" href=\"#quality-assurance\"><\/a>Quality Assurance<\/h2>\n<p>The Matomo project uses an ever-expanding comprehensive set of thousands of unit tests and hundreds of automated integration tests, system tests, JavaScript tests, and screenshot UI tests, running on a continuous integration server as part of its software quality assurance. <a href=\"https:\/\/developer.matomo.org\/guides\/tests\" rel=\"nofollow\">Learn more<\/a><\/p>\n<p>We use <a href=\"https:\/\/www.browserstack.com\/\" rel=\"nofollow\">BrowserStack.com<\/a> testing tool to help check the Matomo user interface is compatible with many browsers.<\/p>\n<h2><a id=\"user-content-security\" class=\"anchor\" aria-hidden=\"true\" href=\"#security\"><\/a>Security<\/h2>\n<p>Security is a top priority at Matomo. As potential issues are discovered, we validate, patch and release fixes as quickly as we can. We have a security bug bounty program in place that rewards researchers for finding security issues and disclosing them to us.<\/p>\n<p><a href=\"https:\/\/matomo.org\/security\/\" rel=\"nofollow\">Learn more<\/a> or check out our <a href=\"https:\/\/hackerone.com\/matomo\" rel=\"nofollow\">HackerOne program<\/a>.<\/p>\n<h2><a id=\"user-content-support-for-matomo\" class=\"anchor\" aria-hidden=\"true\" href=\"#support-for-matomo\"><\/a>Support for Matomo<\/h2>\n<p>For <strong>Free support<\/strong>, post a message in our community forums: <a href=\"https:\/\/forum.matomo.org\/\" rel=\"nofollow\">forum.matomo.org<\/a><\/p>\n<p>For <strong>Professional paid support<\/strong>, purchase a Matomo On-Premises Support Plan: <a href=\"https:\/\/matomo.org\/support-plans\/\" rel=\"nofollow\">matomo.org\/support-plans<\/a><\/p>\n<h2><a id=\"user-content-contact\" class=\"anchor\" aria-hidden=\"true\" href=\"#contact\"><\/a>Contact<\/h2>\n<p>Website: <a href=\"https:\/\/matomo.org\" rel=\"nofollow\">matomo.org<\/a><\/p>\n<p>About us: <a href=\"https:\/\/matomo.org\/team\/\" rel=\"nofollow\">matomo.org\/team\/<\/a><\/p>\n<p>Contact us: <a href=\"https:\/\/matomo.org\/contact\/\" rel=\"nofollow\">matomo.org\/contact\/<\/a><\/p>\n<h2><a id=\"user-content-more-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-information\"><\/a>More information<\/h2>\n<p>What makes Matomo unique from the competition:<\/p>\n<ul><li>\n<p>You own your web analytics data: since Matomo is installed on your server, the data is stored in your own database and you can get all the statistics using the powerful Matomo Analytics API.<\/p>\n<\/li>\n<li>\n<p>Matomo is a Free Software which can easily be configured to respect your visitors' privacy.<\/p>\n<\/li>\n<li>\n<p>Modern, easy to use User Interface: you can fully customize your dashboard, drag and drop widgets and more.<\/p>\n<\/li>\n<li>\n<p>Matomo features are built inside plugins: you can add new features and remove the ones you don\u2019t need. You can build your own web analytics plugins or hire a consultant to have your custom feature built-in Matomo.<\/p>\n<\/li>\n<li>\n<p>A vibrant international Open community of more than 200,000 active users (tracking even more websites!)<\/p>\n<\/li>\n<li>\n<p>Advanced Web Analytics capabilities such as E-commerce Tracking, Goal tracking, Campaign tracking, Custom Variables, Email Reports, Custom Segment Editor, Geo Location, Real-time visits and maps, <a href=\"https:\/\/matomo.org\/feature-overview\/\" rel=\"nofollow\">and a lot more!<\/a><\/p>\n<\/li>\n<\/ul><p>Documentation and more info on <a href=\"https:\/\/matomo.org\" rel=\"nofollow\">https:\/\/matomo.org<\/a><\/p>\n<p>We are together creating the best open analytics platform in the world!<\/p>\n<\/article>","created_at":"2019-12-05T22:08:20+0000","updated_at":"2019-12-05T22:08:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/698038?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["selfhosted","analytics"],"is_public":false,"id":1047,"title":"usefathom\/fathom","url":"https:\/\/github.com\/usefathom\/fathom","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p><a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=LJ5WZVA9ER9GJ\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d5d24e33e2f4b6fe53987419a21b203c03789a8f\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f446f6e6174652d50617950616c2d677265656e2e737667\" alt=\"Donate\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/Donate-PayPal-green.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/goreportcard.com\/report\/github.com\/usefathom\/fathom\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/9c8d41c246f17d94e7f4ef9de3f4ee3cb2656187\/68747470733a2f2f676f7265706f7274636172642e636f6d2f62616467652f6769746875622e636f6d2f757365666174686f6d2f666174686f6d\" alt=\"Go Report Card\" data-canonical-src=\"https:\/\/goreportcard.com\/badge\/github.com\/usefathom\/fathom\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/raw.githubusercontent.com\/usefathom\/fathom\/master\/LICENSE\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/890acbdcb87868b382af9a4b1fac507b9659d9bf\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d626c75652e737667\" alt=\"MIT licensed\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-MIT-blue.svg\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p><a href=\"https:\/\/usefathom.com\/\" rel=\"nofollow\">Fathom Analytics<\/a> is a simpler and more privacy-focused alternative to Google Analytics.<\/p>\n<p>Collecting information on the internet is important, but it\u2019s broken. We\u2019ve become complacent in trading information for free access to web services, and then complaining when those web services do crappy things with that data.<\/p>\n<p>The problem is this: <em>if we aren\u2019t paying for the product, we are the product<\/em>.<\/p>\n<p>Google Analytics may give you free access to their services but in turn, they\u2019re assembling data profiles on your website visitors, which they can then use for better targeting of advertisements across their network.<\/p>\n<p>We need to stop giving away our data and our users' privacy for free access to a tool.<\/p>\n<p>Fathom respects the privacy of your users and does not collect any personally identifiable information. All while giving you the information you need about your site, so you can make smarter decisions about your design and content.<\/p>\n<p>At present, Fathom Analytics Lite is not PECR compliant due to the fact that it uses an anonymous cookie. Our <a href=\"https:\/\/usefathom.com\" rel=\"nofollow\">PRO version<\/a> is PECR compliant, and we'll be making changes to this codebase some time in the future to make it compliant.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/usefathom\/fathom\/raw\/master\/assets\/src\/img\/fathom.jpg?v=7\"><img src=\"https:\/\/github.com\/usefathom\/fathom\/raw\/master\/assets\/src\/img\/fathom.jpg?v=7\" alt=\"Screenshot of the Fathom dashboard\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h2><a id=\"user-content-lite-vs-pro\" class=\"anchor\" aria-hidden=\"true\" href=\"#lite-vs-pro\"><\/a>Lite vs PRO<\/h2>\n<p>We offer a <a href=\"https:\/\/usefathom.com\/#pricing\" rel=\"nofollow\">PRO version of Fathom Analytics<\/a> that starts at $14 \/ month. If you\u2019d like to become a customer, we\u2019d love to have you on board. You can signup for a <a href=\"https:\/\/usefathom.com\/ref\/GITHUB\" rel=\"nofollow\">free 7-day trial of Fathom Analytics here<\/a> <strong>(this link gives you $10 off your first invoice!)<\/strong>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/ff38065df00f650166cadb747cbc0bc345bd1627\/68747470733a2f2f757365666174686f6d2e636f6d2f6173736574732f666174686f6d2d616e616c79746963732e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/ff38065df00f650166cadb747cbc0bc345bd1627\/68747470733a2f2f757365666174686f6d2e636f6d2f6173736574732f666174686f6d2d616e616c79746963732e706e67\" alt=\"Screenshot of the PRO Dashboard\" data-canonical-src=\"https:\/\/usefathom.com\/assets\/fathom-analytics.png\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<table><thead><tr><th>Lite<\/th>\n<th>Pro<\/th>\n<\/tr><\/thead><tbody><tr><td>Uses Cookies<\/td>\n<td>Cookie-free<\/td>\n<\/tr><tr><td>-<\/td>\n<td>Track goal completions<\/td>\n<\/tr><tr><td>-<\/td>\n<td>Share your site publicly or privately<\/td>\n<\/tr><tr><td>-<\/td>\n<td>Device, Browser &amp; Country Data<\/td>\n<\/tr><tr><td>No support (you can post an issue to our repo)<\/td>\n<td>Fast and responsive support from the Fathom founders<\/td>\n<\/tr><tr><td>No guaranteed uptime<\/td>\n<td>Fully redundant cloud-based architecture that keeps your stats online<\/td>\n<\/tr><tr><td>Scaling requires you to power down and then upgrade your server<\/td>\n<td>On-demand, automatic scaling, so even if your site goes viral, your stats won\u2019t stop or slow down - we can handle billions of page views each month<\/td>\n<\/tr><tr><td>Manual backups, if you know how to set them up on your server<\/td>\n<td>Real time backups included in the cost<\/td>\n<\/tr><tr><td>No data protection<\/td>\n<td>Continuous data protection (in the event of a database hardware failure, we have a live standby database ready to go in another availability zone)<\/td>\n<\/tr><tr><td>Manual updates, via the command line<\/td>\n<td>Automatic updates, patches and new versions at no extra cost, no coding required<\/td>\n<\/tr><tr><td>Manual server hardening for security<\/td>\n<td>Totally secure server, monitored and maintained by us, included in the price<\/td>\n<\/tr><tr><td>You pay for hosting, you have to do all the work to maintain the server, the code and backups<\/td>\n<td>You pay us, we take care of everything for you<\/td>\n<\/tr><tr><td>Tracker file served via single server, from a single location<\/td>\n<td>Tracker file served via our super-fast CDN, with endpoints located around the world to ensure fast page loads<\/td>\n<\/tr><tr><td>Data aggregation performed on a single server<\/td>\n<td>Super fast data-aggregation spread across our cloud architecture<\/td>\n<\/tr><tr><td>Contribute to our repo<\/td>\n<td>Supporting a privacy-focused, indie software company<\/td>\n<\/tr><tr><td><a href=\"https:\/\/app.usefathom.com\/register\" rel=\"nofollow\">Get started for free<\/a><\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<h3><a id=\"user-content-production\" class=\"anchor\" aria-hidden=\"true\" href=\"#production\"><\/a>Production<\/h3>\n<p>You can install Fathom on your server by following <a href=\"https:\/\/github.com\/usefathom\/fathom\/blob\/master\/docs\/Installation%20instructions.md\">our simple instructions<\/a>.<\/p>\n<h3><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h3>\n<p>For getting a development version of Fathom up &amp; running, go through the following steps.<\/p>\n<ol><li>Ensure you have <a href=\"https:\/\/golang.org\/doc\/install#install\" rel=\"nofollow\">Go<\/a> and <a href=\"https:\/\/www.npmjs.com\" rel=\"nofollow\">NPM<\/a> installed<\/li>\n<li>Download the code: <code>git clone https:\/\/github.com\/usefathom\/fathom.git $GOPATH\/src\/github.com\/usefathom\/fathom<\/code><\/li>\n<li>Compile the project into an executable: <code>make build<\/code><\/li>\n<li>(Optional) Set <a href=\"https:\/\/github.com\/usefathom\/fathom\/blob\/master\/docs\/Configuration.md\">custom configuration values<\/a><\/li>\n<li>(Required) Register a user account: <code>.\/fathom user add --email=&lt;email&gt; --password=&lt;password&gt;<\/code><\/li>\n<li>Start the webserver: <code>.\/fathom server<\/code> and then visit <strong><a href=\"http:\/\/localhost:8080\" rel=\"nofollow\">http:\/\/localhost:8080<\/a><\/strong> to access your analytics dashboard<\/li>\n<\/ol><h2><a id=\"user-content-docker\" class=\"anchor\" aria-hidden=\"true\" href=\"#docker\"><\/a>Docker<\/h2>\n<h3><a id=\"user-content-building\" class=\"anchor\" aria-hidden=\"true\" href=\"#building\"><\/a>Building<\/h3>\n<p>Ensure you have Docker installed and run <code>docker build -t fathom .<\/code>. Run the container with <code>docker run -d -p 8080:8080 fathom<\/code>.<\/p>\n<h3><a id=\"user-content-running\" class=\"anchor\" aria-hidden=\"true\" href=\"#running\"><\/a>Running<\/h3>\n<p>To run <a href=\"https:\/\/hub.docker.com\/r\/usefathom\/fathom\/\" rel=\"nofollow\">our pre-built Docker image<\/a>, run <code>docker run -d -p 8080:8080 usefathom\/fathom:latest<\/code><\/p>\n<h2><a id=\"user-content-tracking-snippet\" class=\"anchor\" aria-hidden=\"true\" href=\"#tracking-snippet\"><\/a>Tracking snippet<\/h2>\n<p>To start tracking, create a site in your Fathom dashboard and copy the tracking snippet to the website(s) you want to track.<\/p>\n<h3><a id=\"user-content-content-security-policy\" class=\"anchor\" aria-hidden=\"true\" href=\"#content-security-policy\"><\/a>Content Security Policy<\/h3>\n<p>If you use a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/HTTP\/CSP\" rel=\"nofollow\">Content Security Policy (CSP)<\/a> to specify security policies for your website, Fathom requires the following CSP directives (replace <code>yourfathom.com<\/code> with the URL to your Fathom instance):<\/p>\n<pre>script-src: yourfathom.com;\nimg-src: yourfathom.com;\n<\/pre>\n<h2><a id=\"user-content-donation\" class=\"anchor\" aria-hidden=\"true\" href=\"#donation\"><\/a>Donation<\/h2>\n<p>If you're self-hosting Fathom Analytics Lite and want to support it's development, you can:<\/p>\n<p><a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=LJ5WZVA9ER9GJ\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f896f7d176663a1559376bb56aac4bdbbbe85ed1\/68747470733a2f2f7777772e70617970616c6f626a656374732e636f6d2f656e5f55532f692f62746e2f62746e5f646f6e61746543435f4c472e676966\" alt=\"paypal\" data-canonical-src=\"https:\/\/www.paypalobjects.com\/en_US\/i\/btn\/btn_donateCC_LG.gif\" class=\"c2\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<h2><a id=\"user-content-copyright-and-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#copyright-and-license\"><\/a>Copyright and license<\/h2>\n<p>MIT licensed. Fathom and Fathom logo are trademarks of Fathom Analytics.<\/p>\n<\/article>","created_at":"2019-12-05T22:08:08+0000","updated_at":"2019-12-05T22:08:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/38684088?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["revealjs"],"is_public":false,"id":1046,"title":"Creating Presentation Slides in Plain Text","url":"https:\/\/plaintextproject.online\/2019\/09\/04\/slides.html","content":"<p>\n04 Sep 2019 | by: Scott Nesbitt<br \/>You'll need to set aside about 3 minute(s) to read this article<\/p><div><p>Most people turn to desktop software like Powerpoint or LibreOffice Impress, or to web-based tools like Google Slides, when they need to cobble together slides for their presentations. There\u2019s nothing wrong with any of them, but I think that for most people those tools are overkill.<\/p><p>Why not turn to plain text instead? Yes, you <em>can<\/em> create slides in your favourite text editor. And, no, those slides aren\u2019t merely Courier text on a white background. I\u2019m talking about slides that present the information and ideas that you want to present. I\u2019m talking about slides that are more than simply workable.<\/p><p>Let\u2019s take a quick look at three ways in which you can craft slides for your presentations using plain text.<\/p><h2 id=\"revealjs\">Reveal.js<\/h2><p>On those rare occasions I give a talk, I turn to <a href=\"https:\/\/revealjs.com\/#\/\">Reveal.js<\/a> to create my slides. All you need to do is write and format your slides using HTML in a text editor. When you\u2019re ready to present, you open your slides file in a web browser and Reveal.js works all the magic of:<\/p><ul><li>Breaking your file into individual slides<\/li>\n<li>Adding navigation and effects<\/li>\n<li>Applying a theme to the slides<\/li>\n<\/ul><p>You can do just about anything with your slides: add tables, include lists, and insert images. How you structure your slides is only limited by your knowledge of HTML.<\/p><p>To learn more, <a href=\"https:\/\/github.com\/hakimel\/reveal.js#instructions\">browse the documentation<\/a>. I\u2019ve also created a template for some <a href=\"https:\/\/plaintextproject.online\/templates\/slides-simple.html\">simple slides<\/a> that you can use and modify.<\/p><h2 id=\"latex-beamer\">LaTeX Beamer<\/h2><p>If you want to embrace your inner plain text geek a bit more, you\u2019ll want to check out the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Beamer_(LaTeX)\">LaTeX Beamer class<\/a>. LaTeX, as you might remember from <a href=\"https:\/\/plaintextproject.online\/2018\/12\/19\/latex.html\">this article<\/a>, is a powerful typesetting system. The <a href=\"https:\/\/github.com\/josephwright\/beamer\">Beamer class<\/a> helps you create <em>(on-screen) presentations, along with support material such as handouts and speaker notes<\/em>.<\/p><p>A slide show is contained in one file. Each slide is created in a <em>frame<\/em> in the file. The frame denotes a single page in the file that you output. When you\u2019re ready to publish, you can create a PDF file of your slides and another of your handouts.<\/p><p>You can also include images, apply <a href=\"https:\/\/en.wikibooks.org\/wiki\/LaTeX\/Presentations#Themes\">themes<\/a> and colour, as well as change the font and split a slide into two columns.<\/p><p>Actually, there\u2019s a lot more you can do with Beamer. It\u2019s a very powerful and flexible way of creating slides for anyone who uses LaTeX or who is interested in learning it.<\/p><h2 id=\"pandoc\">Pandoc<\/h2><p>One tool that indispensible here at The Plain Text Project is <a href=\"https:\/\/pandoc.org\">Pandoc<\/a>. In case you don\u2019t know what Pandoc is, it\u2019s a program that converts between markup languages and formats.<\/p><p>So what does that have to do with creating presentation slides? Using Pandoc, you can format your slides using <a href=\"https:\/\/gumroad.com\/l\/learnmarkdown\">Markdown<\/a> and output them as HTML that works with these presentation frameworks:<\/p><ul><li><a href=\"https:\/\/revealjs.com\/#\/\">Slidy<\/a><\/li>\n<li><a href=\"https:\/\/goessner.net\/articles\/slideous\/\">Slideous<\/a><\/li>\n<li><a href=\"http:\/\/paulrouget.com\/dzslides\/\">DZSlides<\/a><\/li>\n<li><a href=\"https:\/\/meyerweb.com\/eric\/tools\/s5\/\">S5<\/a><\/li>\n<li><a href=\"https:\/\/revealjs.com\/#\/\">Reveal.js<\/a><\/li>\n<\/ul><p>You can also output LaTeX Beamer slides. Yeah, it\u2019s pretty flexible.<\/p><p>The source file for your slides needs to <a href=\"https:\/\/pandoc.org\/MANUAL.html#producing-slide-shows-with-pandoc\">follow some formatting rules<\/a>, but you can add some variables for <a href=\"https:\/\/pandoc.org\/MANUAL.html#variables-for-html-slides\">HTML slides<\/a> and <a href=\"https:\/\/pandoc.org\/MANUAL.html#variables-for-beamer-slides\">Beamer slides<\/a> to change how they look.<\/p><p>If you\u2019re creating HTML slides, you need to have the supporting files for your preferred presentation framework installed on your computer. Pandoc only produced the HTML slide file.<\/p><h2 id=\"a-few-articles-on-this-subject\">A Few Articles on This Subject<\/h2><p>Crafting slides in plain text is a viable alternative to using big, bulky presentation tools. But don\u2019t take my word for it. Here are a few articles that you might find useful:<\/p><ul><li><a href=\"https:\/\/opensource.com\/article\/18\/2\/how-create-slides-emacs-org-mode-and-revealjs\">How to create slides with Emacs Org mode and Reveal.js<\/a><\/li>\n<li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/how-to-create-presentation-slides-with-html-and-css--net-19870\">How to Create Presentation Slides with HTML and CSS<\/a><\/li>\n<li><a href=\"https:\/\/tech.graze.com\/2015\/07\/31\/easily-create-slideshow-presentations-from-markdown-with-remark-js.html\">Easily create slideshow presentations from markdown with remark.js<\/a><\/li>\n<li><a href=\"https:\/\/nval.andreasherten.de\/2016\/01\/26\/latex-beamer-with-markdown.html\">Talks with LaTeX Beamer, written in Markdown<\/a><\/li>\n<li><a href=\"https:\/\/computers.tutsplus.com\/tutorials\/how-to-create-a-slideshow-presentation-from-markdown-notes--cms-23062\">How to Create a Slideshow Presentation From Markdown Notes<\/a><\/li>\n<\/ul><\/div>","created_at":"2019-12-03T05:33:26+0000","updated_at":"2019-12-03T05:33:55+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"plaintextproject.online","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1045,"title":"Oleg's gists - Do you have a problem? Write a compiler!","url":"http:\/\/oleg.fi\/gists\/posts\/2019-09-26-write-a-compiler.html","content":"<div class=\"row columns\"><p>Posted on 2019-09-26 by Oleg Grenrus<\/p><p>These are notes of the talk I gave at <a href=\"https:\/\/clojutre.org\/2019\/\">ClojuTre 2019<\/a>. There is <a href=\"https:\/\/www.youtube.com\/watch?v=kOXfdZRD0wM\">a video recording of the talk<\/a>, and <a href=\"https:\/\/oleg.fi\/pdfs\/write-a-compiler.pdf\">the slide deck as a PDF<\/a>.<\/p><p><a href=\"https:\/\/creativecommons.org\/licenses\/by-sa\/2.0\/\"><img title=\"cc-by-sa\" src=\"http:\/\/oleg.fi\/gists\/images\/by-sa.svg\" referrerpolicy=\"no-referrer\" alt=\"image\" \/> This work is licensed under a \u201cCC BY SA 4.0\u201d license.<\/a><\/p><div class=\"c1\"><img title=\"slide-00\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-00-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Hello ClojuTre. I'm Oleg from Helsinki.<\/p><hr \/><div class=\"c1\"><img title=\"slide-01\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-01-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Imagine you are writing a cool new rogue-like game. So cool many have no idea what's going on. A definitive character of rogue-likes is procedural generation of content. You'll need a random number generator for that.<\/p><hr \/><div class=\"c1\"><img title=\"slide-02\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-02-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>SplitMix is a fast, splittable pseudorandom number generator. Being able to split a generator into two independent generators is a property you'll want if you use the generator in a functional programming language. Look it up. We'll concentrate on the being fast part. Obviously you want things to be fast.<\/p><p>SplitMix is fast. It does only 9 operations per generated number, one addition to advance the random seed, and <code>xor<\/code>, <code>shift<\/code>, <code>multiply<\/code> <code>xor<\/code>, <code>shift<\/code>, <code>multiply<\/code>, <code>xor<\/code> and <code>shift<\/code>. 9 in total.<\/p><p>However, for the maximum reach, we want our game to run in the web browsers.<\/p><hr \/><div class=\"c1\"><img title=\"slide-03\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-03-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><ul><li>JavaScript is a great platform<\/li>\n<li>But a terrible programming language<\/li>\n<\/ul><hr \/><div class=\"c1\"><img title=\"slide-04\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-04-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Look carefully. When we multiply two \"big\" odd numbers, we do get even result. We shouldn't: product of two odd numbers is odd.<\/p><p>JavaScript numbers are a mess. Next a bit of arithmetics.<\/p><hr \/><div class=\"c1\"><img title=\"slide-05\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-05-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Instead of multiplying two big numbers, we split them in high and low parts and multiply many small numbers. Like in an elementary school. Then we'll have enough precision.<\/p><hr \/><div class=\"c1\"><img title=\"slide-06\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-06-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>So as we are multiplying 32bit numbers, and are interested only in lower 32 bits of 64 bit result. Then we need to do only three multiplications.<\/p><p>At the end, we can get correct results, i.e. good random numbers even in JavaScript.<\/p><hr \/><div class=\"c1\"><img title=\"slide-07\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-07-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Next if we replace the multiplication in <code>xor-shift-multiply<\/code> with a macro doing the right thing, and actually change everything to be a macro, and expand we'll get...<\/p><hr \/><div class=\"c1\"><img title=\"slide-08\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-08-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>a lot of code which barely fits on the slide.<\/p><p>I had to shorted <code>bit-shift-left<\/code> to <code>bsl<\/code> and <code>unsigned-bit-shift-right<\/code> to <code>ubsr<\/code>.<\/p><hr \/><div class=\"c1\"><img title=\"slide-09\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-09-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Look closely. Would you write this kind of code.<\/p><hr \/><div class=\"c1\"><img title=\"slide-10\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-10-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>We bind (\"assign\") a constant value to <code>uv<\/code> variable.<\/p><hr \/><div class=\"c1\"><img title=\"slide-11\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-11-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>And the calculate the high and low 16 bit parts of it. Something we could write directly: <code>let [u 0x85eb v ca6b]<\/code>, i.e. <em>optimize by hand<\/em>. Should we do so?<\/p><hr \/><div class=\"c1\"><img title=\"slide-12\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-12-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>No. Let's rather write an (optimizing) compiler. We don't have time to optimize by hand.<\/p><hr \/><div class=\"c1\"><img title=\"slide-13\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-13-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Recall, we have a very specific problem. Working with a very very tiny subset of Clojure. Some simple arithmetic and bit mangling of 32bit numbers.<\/p><hr \/><div class=\"c1\"><img title=\"slide-14\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-14-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>We'd like to add a little of <code>magic<\/code>, which would make the program magically run faster. In my toy-micro benchmarks the speed-up is 10 percent. It's definitely worth it.<\/p><hr \/><div class=\"c1\"><img title=\"slide-15\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-15-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>What is this magic?<\/p><hr \/><div class=\"c1\"><img title=\"slide-16\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-16-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>A little cute macro. Of course.<\/p><hr \/><div class=\"c1\"><img title=\"slide-17\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-17-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>We get a <code>form<\/code> and convert it into internal representation on a way in, and back to Clojure on the way back.<\/p><p>Working the whole Clojure syntax directly is insane task. We want only deal with a small sane subset of it. Also in a easier to manipulate format. Clojure as a pragmatic language is still optimized for writing and reading, not machine manipulation so much.<\/p><hr \/><div class=\"c1\"><img title=\"slide-18\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-18-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>The next step is to expand a multiplication using a trick we have seen previously. We could have done it already in <code>from-clojure<\/code> step, but it's good engineering to have only one thing per step.<\/p><hr \/><div class=\"c1\"><img title=\"slide-19\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-19-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>And the important part is the <code>optimise<\/code> function.<\/p><hr \/><div class=\"c1\"><img title=\"slide-20\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-20-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>The internal representation I used is nested vectors with a node type as a first element. An uniform representation made it way easier to do everything else. Note how literals 1 and 3 are wrapped into vector too. We can simply look at the head of a vector to know what we are dealing with.<\/p><p>Code is Data.<\/p><hr \/><div class=\"c1\"><img title=\"slide-21\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-21-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>A difficult part in <em>Code is Data<\/em> are local variables. I chose to use <em>de Bruijn<\/em> indices, so instead of names: <code>x<\/code> and <code>y<\/code> or <code>a<\/code> and <code>b<\/code> there are numbers counting towards the corresponding <code>let<\/code> (which binds only one variable at the time by the way).<\/p><p>de Bruijn indices are tricky to grok. Look at the colors, they are there to help. The blue <code>:var 1<\/code> references \"one away\" <code>let<\/code>.<\/p><p>I don't expect you to understand them. It's one way to represent bindings. Perfectly there should be a library, so you don't need to think about low-level details.<\/p><hr \/><div class=\"c1\"><img title=\"slide-22\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-22-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Once we got rid out of names, we can still keep them around using metadata. That's a really cool feature in Clojure, I have to admit. The metadata is there, but not in your way. And having names around is actually useful when you try to debug things.<\/p><hr \/><div class=\"c1\"><img title=\"slide-23\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-23-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Now we have a setup done. Let's jump into optimizations.<\/p><hr \/><div class=\"c1\"><img title=\"slide-24\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-24-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Recall our code snippet. There's <code>uv<\/code> which is bound to a constant. And then it's used an expression which could simplify if we do this and that...<\/p><hr \/><div class=\"c1\"><img title=\"slide-25\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-25-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Let's keep it super simple.<\/p><ul><li>We can have a small set of simple local rewrite rules. Local meaning, we don't need to look around, only at one subexpression at the time.<\/li>\n<li>Then we try to match the rule everywhere, and if it match, perform the rewrite.<\/li>\n<li>And loop until there's nothing to do.<\/li>\n<\/ul><hr \/><div class=\"c1\"><img title=\"slide-26\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-26-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>The first optimization is inlining. In a sense it's most powerful one, as it makes opportunities for other optimizations to fire, even that on itself it doesn't do much.<\/p><p>So if we have a let-binding, then in some cases we perform a substitution. Replace all <code>x<\/code>s with an <code>expression<\/code> inside a body.<\/p><p><code>let x 1 y 1 (+ x y)<\/code> to a lot simpler <code>(+ 1 2)<\/code>. But nothing more, just that.<\/p><hr \/><div class=\"c1\"><img title=\"slide-27\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-27-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>I need to point out, that optimizing is somewhat of an art. Sometimes it work, sometimes it don't.<\/p><p>For example, we don't want to duplicate an expensive <code>(fibonacci 100)<\/code> expression. We want to evaluate it once and share the result.<\/p><p>On the other hand, if someone already went and computed the value, then we can push it to the leaves of an expression tree.<\/p><p>Heuristics are tricky.<\/p><p>Luckily for our needs simple heuristics work well.<\/p><hr \/><div class=\"c1\"><img title=\"slide-28\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-28-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>When <em>inlining<\/em> is a valid rewrite?<\/p><hr \/><div class=\"c1\"><img title=\"slide-29\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-29-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Not in every language. Consider this not-so-functional example.<\/p><p>If we substitute everything, the <code>(do-foo)<\/code> and <code>(do-bar)<\/code> would be in different order. And <code>(do-quux)<\/code> will be gone completely, hopefully it didn't anything important!<\/p><p>We need a language where there are no side-effects, nor there are so much difference in the execution order. Whole Clojure isn't such language. Our tiny subset is.<\/p><p>I have heard there are programming languages which behave like our small one, but are more general purpose!<\/p><p>OK. Let's move to the next optimization.<\/p><hr \/><div class=\"c1\"><img title=\"slide-30\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-30-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>When we have something simple as <code>(+ 1 2)<\/code>, let us evaluate it already at compile time.<\/p><p>For every primitive operation, if the arguments are known constants, just do it.<\/p><hr \/><div class=\"c1\"><img title=\"slide-31\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-31-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Now, I ask you when <em>constant folding<\/em> is a valid rewrite.<\/p><hr \/><div class=\"c1\"><img title=\"slide-32\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-32-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Well, that was a trick question. It really depends on primitives, whether it make sense to perform them at compile time. (Even pure languages have primitives to print stuff on a screen).<\/p><p>But you could think about that <code>precalculate<\/code> example. When does it makes to perform the calculation (assuming we somehow know it terminates):<\/p><ul><li>At compile time?<\/li>\n<li>At start up?<\/li>\n<li>At first access?<\/li>\n<\/ul><p>It really depends, and there are no single simple answer.<\/p><p>Again, optimizations is an art.<\/p><hr \/><div class=\"c1\"><img title=\"slide-33\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-33-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Because we work with nice internal representation, writing individual optimizations is so nice.<\/p><ul><li>if a node is not one of special nodes<\/li>\n<li>and all node arguments are constants<\/li>\n<li>evaluate it.<\/li>\n<\/ul><p>The code is shorter than my explanation. And still understandable.<\/p><hr \/><div class=\"c1\"><img title=\"slide-34\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-34-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>With these two optimizations, inlining and constant folding, we get from this big (and repetitive) code blob to...<\/p><hr \/><div class=\"c1\"><img title=\"slide-35\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-35-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>Something which actually fits on the slide without font size scaling. It's not super-pretty, but it's hard to spot if something can be done there.<\/p><hr \/><div class=\"c1\"><img title=\"slide-36\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-36-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>We can make it look nice with one more optimization. When you bind a let-expression to a variable in outer let-expression, we can float out the inner one.<\/p><p>A very old idea it is.<\/p><hr \/><div class=\"c1\"><img title=\"slide-37\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-37-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>And then we get (in my opinion) a very nice direct code. Six intermediate results to get final one.<\/p><hr \/><div class=\"c1\"><img title=\"slide-38\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-38-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><p>That's what I wanted to tell you.<\/p><hr \/><div class=\"c1\"><img title=\"slide-39\" src=\"http:\/\/oleg.fi\/gists\/images\/write-a-compiler-39-fs8.png\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div><ul><li>Implementing small (domain specific) languages is fun.<\/li>\n<li>If you approach problems with \"let's write a programming language to describe them\" -attitude there a lot of big hammers in your disposal. A lot of wheels is already invented.<\/li>\n<li>Languages don't need only be about numerics, it could be HTTP routing, authorisation rules, UI-workflows, CI-scripts (I could bash about bash), data descriptions, you name it.<\/li>\n<li>But make your languages <em>typed, lazy, pure, total or and even dependent<\/em> for extra fun and interesting new problems. ;)<\/li>\n<\/ul><p>Thank you.<\/p><\/div><div class=\"row columns\"><p><a href=\"http:\/\/www.twitter.com\/phadej\">twitter<\/a> <a href=\"https:\/\/github.com\/phadej\">github<\/a> <a href=\"mailto:oleg.grenrus@iki.fi\">email<\/a> <a href=\"https:\/\/www.linkedin.com\/in\/oleggrenrus\">linkedin<\/a> <a href=\"http:\/\/stackoverflow.com\/users\/1308058\/oleg-grenrus\">stackoverflow<\/a> <a href=\"http:\/\/www.strava.com\/athletes\/498127\">strava<\/a><\/p><\/div>","created_at":"2019-12-03T05:32:02+0000","updated_at":"2019-12-03T05:32:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"oleg.fi","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1044,"title":"Ask HN: What are weird and\/or novel ways to do web UIs?","url":"https:\/\/news.ycombinator.com\/item?id=21144228","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zxcvbn4038\" class=\"hnuser\">zxcvbn4038<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145334\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145549\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=marapuru\" class=\"hnuser\">marapuru<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146606\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=klenwell\" class=\"hnuser\">klenwell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147564\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gdhbcc\" class=\"hnuser\">gdhbcc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21164703\">58 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=wtvanhest\" class=\"hnuser\">wtvanhest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148139\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yesbabyyes\" class=\"hnuser\">yesbabyyes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155652\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=marapuru\" class=\"hnuser\">marapuru<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155694\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yesbabyyes\" class=\"hnuser\">yesbabyyes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21156584\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=riazrizvi\" class=\"hnuser\">riazrizvi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147139\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexis_fr\" class=\"hnuser\">alexis_fr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148342\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevesimmons\" class=\"hnuser\">stevesimmons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21161991\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jedieaston\" class=\"hnuser\">jedieaston<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146364\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=auxym\" class=\"hnuser\">auxym<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21162320\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacobush\" class=\"hnuser\">jacobush<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152799\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=AlphaWeaver\" class=\"hnuser\">AlphaWeaver<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146029\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=AndrewStephens\" class=\"hnuser\">AndrewStephens<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148100\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=smacktoward\" class=\"hnuser\">smacktoward<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148135\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=moonlighter\" class=\"hnuser\">moonlighter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149938\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=AHThomas\" class=\"hnuser\">AHThomas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150245\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=godson_drafty\" class=\"hnuser\">godson_drafty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151413\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmarreck\" class=\"hnuser\">pmarreck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148502\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tripzilch\" class=\"hnuser\">tripzilch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151490\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmarreck\" class=\"hnuser\">pmarreck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21172222\">57 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=shanemhansen\" class=\"hnuser\">shanemhansen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151925\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=simonebrunozzi\" class=\"hnuser\">simonebrunozzi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21153977\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=choiway\" class=\"hnuser\">choiway<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148021\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=slothtrop\" class=\"hnuser\">slothtrop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147367\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=xRahul\" class=\"hnuser\">xRahul<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145552\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=davinic\" class=\"hnuser\">davinic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148862\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mikepurvis\" class=\"hnuser\">mikepurvis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147117\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=htfu\" class=\"hnuser\">htfu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148120\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zxcvbn4038\" class=\"hnuser\">zxcvbn4038<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21157176\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=htfu\" class=\"hnuser\">htfu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21159237\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=TekMol\" class=\"hnuser\">TekMol<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145233\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=oftenwrong\" class=\"hnuser\">oftenwrong<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145936\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=heffer\" class=\"hnuser\">heffer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147483\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thirdsun\" class=\"hnuser\">thirdsun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21223458\">52 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=synack\" class=\"hnuser\">synack<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21157205\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fnoord\" class=\"hnuser\">Fnoord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147424\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=htfu\" class=\"hnuser\">htfu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148381\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fnoord\" class=\"hnuser\">Fnoord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21157213\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=philipodonnell\" class=\"hnuser\">philipodonnell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149387\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=porker\" class=\"hnuser\">porker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145361\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mirimir\" class=\"hnuser\">mirimir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145323\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fnoord\" class=\"hnuser\">Fnoord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147324\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=justaguyhere\" class=\"hnuser\">justaguyhere<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146372\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=davidivadavid\" class=\"hnuser\">davidivadavid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150358\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=toddmorey\" class=\"hnuser\">toddmorey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146841\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<div class=\"c7\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/div>\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/item?id=21147407\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Vosporos\" class=\"hnuser\">Vosporos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147781\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yakshaving_jgt\" class=\"hnuser\">yakshaving_jgt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145750\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=el_dev_hell\" class=\"hnuser\">el_dev_hell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21153812\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yakshaving_jgt\" class=\"hnuser\">yakshaving_jgt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155953\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Namrog84\" class=\"hnuser\">Namrog84<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144875\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacobush\" class=\"hnuser\">jacobush<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145161\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Roritharr\" class=\"hnuser\">Roritharr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145356\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cr0sh\" class=\"hnuser\">cr0sh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149384\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Namrog84\" class=\"hnuser\">Namrog84<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150954\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kqr\" class=\"hnuser\">kqr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155323\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomcatfish\" class=\"hnuser\">tomcatfish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21162549\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=giancarlostoro\" class=\"hnuser\">giancarlostoro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21159781\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kqr\" class=\"hnuser\">kqr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145389\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sbayeta\" class=\"hnuser\">sbayeta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146142\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Namrog84\" class=\"hnuser\">Namrog84<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150966\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jedimastert\" class=\"hnuser\">jedimastert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148443\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=melvinroest\" class=\"hnuser\">melvinroest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145187\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=j-me\" class=\"hnuser\">j-me<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145972\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=melvinroest\" class=\"hnuser\">melvinroest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152465\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nestorD\" class=\"hnuser\">nestorD<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145497\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pjc50\" class=\"hnuser\">pjc50<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145218\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=melvinroest\" class=\"hnuser\">melvinroest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145255\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=neurostimulant\" class=\"hnuser\">neurostimulant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149899\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kall\" class=\"hnuser\">kall<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21156093\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ulucs\" class=\"hnuser\">ulucs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146637\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ilovetux\" class=\"hnuser\">ilovetux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148470\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=melvinroest\" class=\"hnuser\">melvinroest<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152483\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jetti\" class=\"hnuser\">jetti<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149077\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mpfundstein\" class=\"hnuser\">mpfundstein<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21165129\">58 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jetti\" class=\"hnuser\">jetti<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21202749\">54 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ggerganov\" class=\"hnuser\">ggerganov<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145314\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=helltone\" class=\"hnuser\">helltone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145577\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sbayeta\" class=\"hnuser\">sbayeta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146114\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ag8\" class=\"hnuser\">ag8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148019\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=some1else\" class=\"hnuser\">some1else<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149643\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ssully\" class=\"hnuser\">ssully<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152099\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rwmj\" class=\"hnuser\">rwmj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145328\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=regularfry\" class=\"hnuser\">regularfry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145833\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zxcmx\" class=\"hnuser\">zxcmx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21154096\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thrower123\" class=\"hnuser\">thrower123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21154154\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Doxin\" class=\"hnuser\">Doxin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147421\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=severine\" class=\"hnuser\">severine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21153704\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Doxin\" class=\"hnuser\">Doxin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21156132\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tiborsaas\" class=\"hnuser\">tiborsaas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144911\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomglynch\" class=\"hnuser\">tomglynch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144664\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nothis\" class=\"hnuser\">nothis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147318\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dandare\" class=\"hnuser\">dandare<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144996\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=htfu\" class=\"hnuser\">htfu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148557\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=porker\" class=\"hnuser\">porker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145364\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=brianpgordon\" class=\"hnuser\">brianpgordon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151058\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yesbabyyes\" class=\"hnuser\">yesbabyyes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155828\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=amoerie\" class=\"hnuser\">amoerie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151552\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sesutton\" class=\"hnuser\">sesutton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151797\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacobush\" class=\"hnuser\">jacobush<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152885\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=superkuh\" class=\"hnuser\">superkuh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146433\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Bilters\" class=\"hnuser\">Bilters<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145291\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Ruphin\" class=\"hnuser\">Ruphin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146154\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hazzamanic\" class=\"hnuser\">hazzamanic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21158476\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=justusthane\" class=\"hnuser\">justusthane<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151492\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=orthecreedence\" class=\"hnuser\">orthecreedence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152066\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tontonius\" class=\"hnuser\">tontonius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145310\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=iLemming\" class=\"hnuser\">iLemming<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21161585\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ssijak\" class=\"hnuser\">ssijak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148063\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ilovetux\" class=\"hnuser\">ilovetux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148539\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sp332\" class=\"hnuser\">sp332<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147911\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sksksk\" class=\"hnuser\">sksksk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145240\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=daliusd\" class=\"hnuser\">daliusd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145321\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147302\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=TeMPOraL\" class=\"hnuser\">TeMPOraL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145168\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=daosyn\" class=\"hnuser\">daosyn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152083\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=atrilumen\" class=\"hnuser\">atrilumen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149180\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=license2e\" class=\"hnuser\">license2e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148427\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dec0dedab0de\" class=\"hnuser\">dec0dedab0de<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146869\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nothis\" class=\"hnuser\">nothis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147242\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=miki123211\" class=\"hnuser\">miki123211<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151768\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mwcampbell\" class=\"hnuser\">mwcampbell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21167265\">58 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=miki123211\" class=\"hnuser\">miki123211<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21177711\">57 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haolez\" class=\"hnuser\">haolez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146971\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=grenoire\" class=\"hnuser\">grenoire<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147405\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=NoGravitas\" class=\"hnuser\">NoGravitas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147715\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haolez\" class=\"hnuser\">haolez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149264\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=NoGravitas\" class=\"hnuser\">NoGravitas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149805\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=radimm\" class=\"hnuser\">radimm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150006\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haolez\" class=\"hnuser\">haolez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21154079\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=8jef\" class=\"hnuser\">8jef<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145304\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ryanianian\" class=\"hnuser\">ryanianian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145561\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=laszlokorte\" class=\"hnuser\">laszlokorte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146615\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mzzter\" class=\"hnuser\">mzzter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149097\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=khuknows\" class=\"hnuser\">khuknows<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145391\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=juancampa\" class=\"hnuser\">juancampa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148550\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xelbair\" class=\"hnuser\">Xelbair<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146400\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tim1994\" class=\"hnuser\">tim1994<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145826\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jho406\" class=\"hnuser\">jho406<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146526\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=faizshah\" class=\"hnuser\">faizshah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148432\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jho406\" class=\"hnuser\">jho406<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21153416\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=carapace\" class=\"hnuser\">carapace<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149153\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=fragmede\" class=\"hnuser\">fragmede<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149578\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=carapace\" class=\"hnuser\">carapace<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150795\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=WillPontificate\" class=\"hnuser\">WillPontificate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145851\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xelbair\" class=\"hnuser\">Xelbair<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146240\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=WillPontificate\" class=\"hnuser\">WillPontificate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147023\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=onemoresoop\" class=\"hnuser\">onemoresoop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148747\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ahartmetz\" class=\"hnuser\">ahartmetz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155518\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pssflops\" class=\"hnuser\">pssflops<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147076\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ehnto\" class=\"hnuser\">ehnto<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148367\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=htfu\" class=\"hnuser\">htfu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148488\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=p_l\" class=\"hnuser\">p_l<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21171163\">57 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=neurostimulant\" class=\"hnuser\">neurostimulant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147910\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lykr0n\" class=\"hnuser\">lykr0n<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149137\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=exikyut\" class=\"hnuser\">exikyut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21153997\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway_bad\" class=\"hnuser\">throwaway_bad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145352\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nathell\" class=\"hnuser\">nathell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146679\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=specialist\" class=\"hnuser\">specialist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147186\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=harryf\" class=\"hnuser\">harryf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145322\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pjungwir\" class=\"hnuser\">pjungwir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155246\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zimpenfish\" class=\"hnuser\">zimpenfish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21155330\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=laumars\" class=\"hnuser\">laumars<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145400\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway_bad\" class=\"hnuser\">throwaway_bad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145188\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=boldslogan\" class=\"hnuser\">boldslogan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145272\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=subpixel\" class=\"hnuser\">subpixel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151785\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mattasher\" class=\"hnuser\">Mattasher<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145726\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bullen\" class=\"hnuser\">bullen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146716\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=spiralganglion\" class=\"hnuser\">spiralganglion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21157767\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=theossuary\" class=\"hnuser\">theossuary<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21157986\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=spiralganglion\" class=\"hnuser\">spiralganglion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21163216\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mo1ok\" class=\"hnuser\">mo1ok<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151393\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=clemensnk\" class=\"hnuser\">clemensnk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146178\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sktrdie\" class=\"hnuser\">sktrdie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145942\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jaimex2\" class=\"hnuser\">jaimex2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146065\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=alluro2\" class=\"hnuser\">alluro2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148511\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jaimex2\" class=\"hnuser\">jaimex2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21169812\">58 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bradenb\" class=\"hnuser\">bradenb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147913\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=avaer\" class=\"hnuser\">avaer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145450\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Zarel\" class=\"hnuser\">Zarel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146992\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=l00sed\" class=\"hnuser\">l00sed<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148751\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=teknopaul\" class=\"hnuser\">teknopaul<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152423\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=xylophone\" class=\"hnuser\">xylophone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146611\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rescbr\" class=\"hnuser\">rescbr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146711\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nick_vdb31\" class=\"hnuser\">nick_vdb31<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147154\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tsak\" class=\"hnuser\">tsak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146166\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zitterbewegung\" class=\"hnuser\">zitterbewegung<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147115\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147234\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=neurostimulant\" class=\"hnuser\">neurostimulant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148052\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zitterbewegung\" class=\"hnuser\">zitterbewegung<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150855\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dugluak\" class=\"hnuser\">dugluak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148047\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dynamite-ready\" class=\"hnuser\">dynamite-ready<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147021\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gameswithgo\" class=\"hnuser\">gameswithgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146559\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=macca321\" class=\"hnuser\">macca321<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145071\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=JDiculous\" class=\"hnuser\">JDiculous<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144668\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sowrabh\" class=\"hnuser\">sowrabh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145069\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=PavlovsCat\" class=\"hnuser\">PavlovsCat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146619\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=neurostimulant\" class=\"hnuser\">neurostimulant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150017\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dkersten\" class=\"hnuser\">dkersten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144823\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=philtar\" class=\"hnuser\">philtar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147945\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=therealmarv\" class=\"hnuser\">therealmarv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144870\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=meremortals\" class=\"hnuser\">meremortals<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147290\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=atannen\" class=\"hnuser\">atannen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152761\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=christophilus\" class=\"hnuser\">christophilus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146020\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=non-entity\" class=\"hnuser\">non-entity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150755\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=intrasight\" class=\"hnuser\">intrasight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146915\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=RocketSyntax\" class=\"hnuser\">RocketSyntax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146507\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jawns\" class=\"hnuser\">jawns<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146633\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=guhcampos\" class=\"hnuser\">guhcampos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145992\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=AkshatM\" class=\"hnuser\">AkshatM<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145748\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=itronitron\" class=\"hnuser\">itronitron<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145951\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=breakerbox\" class=\"hnuser\">breakerbox<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146422\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=exikyut\" class=\"hnuser\">exikyut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149573\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mtrycz2\" class=\"hnuser\">mtrycz2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146102\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=vollmond\" class=\"hnuser\">vollmond<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21160575\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jedimastert\" class=\"hnuser\">jedimastert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148390\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thewway481\" class=\"hnuser\">thewway481<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21144777\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=michannne\" class=\"hnuser\">michannne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146179\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bkanber\" class=\"hnuser\">bkanber<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151983\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yummypaint\" class=\"hnuser\">yummypaint<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21156982\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=hmhrex\" class=\"hnuser\">hmhrex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149249\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ransom1538\" class=\"hnuser\">ransom1538<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146293\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=AnIdiotOnTheNet\" class=\"hnuser\">AnIdiotOnTheNet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146658\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147025\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrighele\" class=\"hnuser\">mrighele<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148566\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Epskampie\" class=\"hnuser\">Epskampie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150119\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21174535\">57 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=grenoire\" class=\"hnuser\">grenoire<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147340\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mthoms\" class=\"hnuser\">mthoms<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148060\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yazan94\" class=\"hnuser\">yazan94<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150096\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=edoceo\" class=\"hnuser\">edoceo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147857\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nvarsj\" class=\"hnuser\">nvarsj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146418\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=joshmn\" class=\"hnuser\">joshmn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146650\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cies\" class=\"hnuser\">cies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146676\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jamesgeck0\" class=\"hnuser\">jamesgeck0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148212\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cabernal\" class=\"hnuser\">cabernal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146666\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mister_hn\" class=\"hnuser\">mister_hn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145784\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=madprops\" class=\"hnuser\">madprops<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146137\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tekkk\" class=\"hnuser\">tekkk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148066\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrisweekly\" class=\"hnuser\">chrisweekly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146038\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=benbristow\" class=\"hnuser\">benbristow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146873\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147174\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rmdashrfstar\" class=\"hnuser\">rmdashrfstar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147997\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jng\" class=\"hnuser\">jng<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149419\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kls\" class=\"hnuser\">kls<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150204\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Cyberdog\" class=\"hnuser\">Cyberdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149760\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149874\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=diminoten\" class=\"hnuser\">diminoten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150471\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21151980\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=diminoten\" class=\"hnuser\">diminoten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21153769\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21159811\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=diminoten\" class=\"hnuser\">diminoten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21161383\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"400\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21161672\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"440\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=diminoten\" class=\"hnuser\">diminoten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21164031\">58 days ago<\/a> [flagged]<\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"480\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21169916\">58 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"520\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=diminoten\" class=\"hnuser\">diminoten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21174990\">57 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21158342\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=benbristow\" class=\"hnuser\">benbristow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147423\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=collyw\" class=\"hnuser\">collyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21207278\">54 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickopotomus\" class=\"hnuser\">kickopotomus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150595\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=CameronNemo\" class=\"hnuser\">CameronNemo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148345\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149776\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=benbristow\" class=\"hnuser\">benbristow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150945\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=debaserab2\" class=\"hnuser\">debaserab2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21152176\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=louismerlin\" class=\"hnuser\">louismerlin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21145829\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=FlorianRappl\" class=\"hnuser\">FlorianRappl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146598\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=flanbiscuit\" class=\"hnuser\">flanbiscuit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146820\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=petilon\" class=\"hnuser\">petilon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21147105\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=schwartzworld\" class=\"hnuser\">schwartzworld<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150811\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kls\" class=\"hnuser\">kls<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150284\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kls\" class=\"hnuser\">kls<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150021\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=y-c-o-m-b\" class=\"hnuser\">y-c-o-m-b<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21148568\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=FlorianRappl\" class=\"hnuser\">FlorianRappl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146582\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway_bad\" class=\"hnuser\">throwaway_bad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21146717\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=codecurve\" class=\"hnuser\">codecurve<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149031\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mister_hn\" class=\"hnuser\">mister_hn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150319\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tqkxzugoaupvwqr\" class=\"hnuser\">tqkxzugoaupvwqr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150512\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=helb\" class=\"hnuser\">helb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21157576\">59 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=schwartzworld\" class=\"hnuser\">schwartzworld<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150823\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=pier25\" class=\"hnuser\">pier25<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21149797\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=yazan94\" class=\"hnuser\">yazan94<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21150152\">60 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=21144228&amp;p=2\" class=\"morelink\" rel=\"next\">More<\/a><\/td>\n<\/tr><\/table>","created_at":"2019-12-03T05:32:01+0000","updated_at":"2019-12-03T05:32:01+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1043,"title":"Hard comp-fi reading list | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=21115324","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=manifestsilence\" class=\"hnuser\">manifestsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117281\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=technofiend\" class=\"hnuser\">technofiend<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117779\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=irrational\" class=\"hnuser\">irrational<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119744\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=technofiend\" class=\"hnuser\">technofiend<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119852\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tsmarsh\" class=\"hnuser\">tsmarsh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117849\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=wil421\" class=\"hnuser\">wil421<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120004\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=otikik\" class=\"hnuser\">otikik<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117771\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=manifestsilence\" class=\"hnuser\">manifestsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118407\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=anderspitman\" class=\"hnuser\">anderspitman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117435\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=beezle\" class=\"hnuser\">beezle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120330\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=manifestsilence\" class=\"hnuser\">manifestsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118770\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nurettin\" class=\"hnuser\">nurettin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119629\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=twic\" class=\"hnuser\">twic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117963\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=joshmarlow\" class=\"hnuser\">joshmarlow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118531\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mellosouls\" class=\"hnuser\">mellosouls<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118563\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=zem\" class=\"hnuser\">zem<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120123\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=haasted\" class=\"hnuser\">haasted<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117176\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=eigenhombre\" class=\"hnuser\">eigenhombre<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119599\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ThinkBeat\" class=\"hnuser\">ThinkBeat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120680\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=darkfire613\" class=\"hnuser\">darkfire613<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117743\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=NCO2\" class=\"hnuser\">NCO2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119808\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dcra\" class=\"hnuser\">dcra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118682\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sramsay\" class=\"hnuser\">sramsay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117603\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=entropicdrifter\" class=\"hnuser\">entropicdrifter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118050\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Alex63\" class=\"hnuser\">Alex63<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118467\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MarcScott\" class=\"hnuser\">MarcScott<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118372\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bart_spoon\" class=\"hnuser\">bart_spoon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117688\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=waterfowl\" class=\"hnuser\">waterfowl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21116770\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=stuxnet79\" class=\"hnuser\">stuxnet79<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117421\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=teraflop\" class=\"hnuser\">teraflop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118410\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=DEADBEEFC0FFEE\" class=\"hnuser\">DEADBEEFC0FFEE<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119392\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gjm11\" class=\"hnuser\">gjm11<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21123584\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=monday_\" class=\"hnuser\">monday_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117999\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ZhuanXia\" class=\"hnuser\">ZhuanXia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120157\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=monday_\" class=\"hnuser\">monday_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21139657\">61 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=scrumper\" class=\"hnuser\">scrumper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117555\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=twirlip\" class=\"hnuser\">twirlip<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117774\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dcra\" class=\"hnuser\">dcra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118691\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=emikulic\" class=\"hnuser\">emikulic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21123217\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=defertoreptar\" class=\"hnuser\">defertoreptar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117247\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=adaml_623\" class=\"hnuser\">adaml_623<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21116946\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sundarurfriend\" class=\"hnuser\">sundarurfriend<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21123610\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=frikk\" class=\"hnuser\">frikk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117240\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=squeaky-clean\" class=\"hnuser\">squeaky-clean<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119502\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=bryanlarsen\" class=\"hnuser\">bryanlarsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117241\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=guskel\" class=\"hnuser\">guskel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21116833\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=codeulike\" class=\"hnuser\">codeulike<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118724\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevekemp\" class=\"hnuser\">stevekemp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118843\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21126354\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21124662\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=codeulike\" class=\"hnuser\">codeulike<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21124796\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=qubex\" class=\"hnuser\">qubex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118680\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=philsnow\" class=\"hnuser\">philsnow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119340\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=qubex\" class=\"hnuser\">qubex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21122827\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sddfd\" class=\"hnuser\">sddfd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118921\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gattr\" class=\"hnuser\">gattr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117962\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ak217\" class=\"hnuser\">ak217<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120804\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=scrumper\" class=\"hnuser\">scrumper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119669\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=gattr\" class=\"hnuser\">gattr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21122362\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Zanni\" class=\"hnuser\">Zanni<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118097\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Baeocystin\" class=\"hnuser\">Baeocystin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118758\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=farnsworth\" class=\"hnuser\">farnsworth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21121067\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Zanni\" class=\"hnuser\">Zanni<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21122976\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jorjordandan\" class=\"hnuser\">jorjordandan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118716\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thrower123\" class=\"hnuser\">thrower123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21116937\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=snitch182\" class=\"hnuser\">snitch182<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117046\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=rgacote\" class=\"hnuser\">rgacote<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21115385\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=korse\" class=\"hnuser\">korse<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118523\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=blacksmith_tb\" class=\"hnuser\">blacksmith_tb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119615\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=drak0n1c\" class=\"hnuser\">drak0n1c<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117714\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=berbec\" class=\"hnuser\">berbec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118359\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mcguire\" class=\"hnuser\">mcguire<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118895\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=CobrastanJorji\" class=\"hnuser\">CobrastanJorji<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120849\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=EvanAnderson\" class=\"hnuser\">EvanAnderson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117508\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MarcScott\" class=\"hnuser\">MarcScott<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118334\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ArtWomb\" class=\"hnuser\">ArtWomb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118304\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=blacksmith_tb\" class=\"hnuser\">blacksmith_tb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118678\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=nils-m-holm\" class=\"hnuser\">nils-m-holm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118818\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=blacksmith_tb\" class=\"hnuser\">blacksmith_tb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119567\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=simmons\" class=\"hnuser\">simmons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117518\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=stefs\" class=\"hnuser\">stefs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118245\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Apocryphon\" class=\"hnuser\">Apocryphon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119335\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=ChuckMcM\" class=\"hnuser\">ChuckMcM<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117325\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mellosouls\" class=\"hnuser\">mellosouls<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117678\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lachlan-sneff\" class=\"hnuser\">lachlan-sneff<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117799\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=twic\" class=\"hnuser\">twic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118001\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=WalterBright\" class=\"hnuser\">WalterBright<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120058\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=omnibrain\" class=\"hnuser\">omnibrain<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117360\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=e12e\" class=\"hnuser\">e12e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119333\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=lonalzarus\" class=\"hnuser\">lonalzarus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117203\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=RmDen\" class=\"hnuser\">RmDen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117320\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MrFantastic\" class=\"hnuser\">MrFantastic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120246\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jlangemeier\" class=\"hnuser\">jlangemeier<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117987\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dsr_\" class=\"hnuser\">dsr_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119408\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=cylinder714\" class=\"hnuser\">cylinder714<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117964\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tombert\" class=\"hnuser\">tombert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117459\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=planck01\" class=\"hnuser\">planck01<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117665\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MichaelMoser123\" class=\"hnuser\">MichaelMoser123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118152\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=traverseda\" class=\"hnuser\">traverseda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119014\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=carapace\" class=\"hnuser\">carapace<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118598\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=f00zz\" class=\"hnuser\">f00zz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118014\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=swayvil\" class=\"hnuser\">swayvil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119682\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=philhd\" class=\"hnuser\">philhd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117344\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=inetsee\" class=\"hnuser\">inetsee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117507\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=manifestsilence\" class=\"hnuser\">manifestsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117389\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=krapp\" class=\"hnuser\">krapp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117736\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=manifestsilence\" class=\"hnuser\">manifestsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118677\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117674\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=TickleSteve\" class=\"hnuser\">TickleSteve<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118842\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=kenforthewin\" class=\"hnuser\">kenforthewin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117482\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=skosch\" class=\"hnuser\">skosch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21121513\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhbadger\" class=\"hnuser\">jhbadger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118834\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=davedx\" class=\"hnuser\">davedx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119624\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=OkGoDoIt\" class=\"hnuser\">OkGoDoIt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21127898\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=demygale\" class=\"hnuser\">demygale<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117493\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=krapp\" class=\"hnuser\">krapp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118606\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrewf\" class=\"hnuser\">andrewf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21121351\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=qznc\" class=\"hnuser\">qznc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119307\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=virtualritz\" class=\"hnuser\">virtualritz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117648\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=orjan\" class=\"hnuser\">orjan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117825\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=atrn\" class=\"hnuser\">atrn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119985\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dstroot\" class=\"hnuser\">dstroot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119001\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=josephwegner\" class=\"hnuser\">josephwegner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119049\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=geocrasher\" class=\"hnuser\">geocrasher<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21124902\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=agentwiggles\" class=\"hnuser\">agentwiggles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21126059\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mcguire\" class=\"hnuser\">mcguire<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119030\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mslev\" class=\"hnuser\">mslev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119151\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=tcpekin\" class=\"hnuser\">tcpekin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21130007\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mirimir\" class=\"hnuser\">mirimir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119925\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117664\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Blakestr\" class=\"hnuser\">Blakestr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21131921\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=dllthomas\" class=\"hnuser\">dllthomas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119626\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21121134\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=MaysonL\" class=\"hnuser\">MaysonL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119756\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21126563\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jasoneckert\" class=\"hnuser\">jasoneckert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120954\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=thecleaner\" class=\"hnuser\">thecleaner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119591\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sundarurfriend\" class=\"hnuser\">sundarurfriend<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21123555\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=samirm\" class=\"hnuser\">samirm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119110\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=onemoresoop\" class=\"hnuser\">onemoresoop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117778\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mannykannot\" class=\"hnuser\">mannykannot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118177\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118158\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=onemoresoop\" class=\"hnuser\">onemoresoop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120470\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=iamaelephant\" class=\"hnuser\">iamaelephant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119782\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=jlangemeier\" class=\"hnuser\">jlangemeier<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21128307\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=DennisP\" class=\"hnuser\">DennisP<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21121097\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21126690\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=DennisP\" class=\"hnuser\">DennisP<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21128144\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=mordant\" class=\"hnuser\">mordant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21123456\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=aledalgrande\" class=\"hnuser\">aledalgrande<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120805\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=justaj\" class=\"hnuser\">justaj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21122428\">62 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=wintorez\" class=\"hnuser\">wintorez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21119699\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=Bootwizard\" class=\"hnuser\">Bootwizard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21117949\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=sideshowb\" class=\"hnuser\">sideshowb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21118142\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>\n\n<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" referrerpolicy=\"no-referrer\" alt=\"image\" \/><\/td>\n<td valign=\"top\" class=\"votelinks\">\n<\/td>\n<td class=\"default\">\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=52-hertz_whale\" class=\"hnuser\">52-hertz_whale<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=21120690\">63 days ago<\/a><\/p>\n<br \/><\/td>\n<\/tr><\/table>","created_at":"2019-12-03T05:31:59+0000","updated_at":"2019-12-03T05:31:59+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["procgen","gamedev","noita"],"is_public":false,"id":1042,"title":"Exploring the Tech and Design of 'Noita'","url":"https:\/\/www.gdcvault.com\/play\/1025695\/Exploring-the-Tech-and-Design","content":"GDC Vault - Exploring the Tech and Design of 'Noita'\n       \n         <div id=\"iribbon-container\"><div id=\"iribbon-detail\" class=\"ribbon-hide\"><div id=\"iribbon-left\"><p>GDC Vault is part of the Informa Tech Division of Informa PLC<\/p><ul><li class=\"iribbon-mainlink\"><a href=\"http:\/\/informa.com\/\" tabindex=\"-1\" class=\"iribbon-opacity-out\">Informa PLC<\/a><\/li>\n                   <li><a href=\"http:\/\/informa.com\/About-Us\/\" tabindex=\"-1\" class=\"iribbon-opacity-out\">About us<\/a><\/li>\n                   <li><a href=\"http:\/\/informa.com\/Investors\/\" tabindex=\"-1\" class=\"iribbon-opacity-out\">Investor relations<\/a><\/li>\n                   <li><a href=\"http:\/\/informa.com\/Talent\/\" tabindex=\"-1\" class=\"iribbon-opacity-out\">Talent<\/a><\/li>\n               <\/ul><\/div><div id=\"iribbon-right\"><p>This is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.<\/p><\/div><\/div><\/div>\n         <div id=\"bumpedWrapper\"><div id=\"bumped\">You've been logged out of GDC Vault since the maximum users allowed for this account has been reached.\n                        To access Members Only content on GDC Vault, please log out of GDC Vault from the computer which last accessed this account.<p>Click <strong><a href=\"https:\/\/www.gdcvault.com\/demo\">here<\/a><\/strong> to find out about GDC Vault Membership options for more users.\n                        <\/p><p><a id=\"bumpClose\">close<\/a>\n                <\/p><\/div><\/div>\n        <header>\n        <\/header><div class=\"wrapper\"><div id=\"player\"><table><tr><td><strong>Session Name:<\/strong><\/td>\n        <td>Exploring the Tech and Design of 'Noita'<\/td>\n    <\/tr><tr><td><strong>Speaker(s):<\/strong><\/td>\n        <td>\n                                Petri Purho                        <\/td>\n    <\/tr><tr><td><strong>Company Name(s):<\/strong><\/td>\n        <td>\n                                Nolla Games                        <\/td>\n    <\/tr><tr><td><strong>Track \/ Format:<\/strong> <\/td>\n        <td>Independent Games Summit<\/td>\n    <\/tr><tr><td><strong>Overview:<\/strong><\/td>\n        <td>\n            'Noita' is a rogue-lite set in a procedurally generated physics based destructible world. In this talk, co-creator Petri Purho talks about the technical details of the physics engine, including scaling up the falling sand simulation to support large continuous worlds and integrating destructible rigid body physics, while demonstrating the emergent physics based gameplay.\n        <\/td>\n    <\/tr><\/table><br \/><section id=\"recommended\"><p><\/p><h3>Recommended<\/h3><img src=\"https:\/\/www.gdcvault.com\/img\/icon_members_only_text.png\" alt=\"image\" \/><\/section><\/div><\/div>\n       \n        \n                &#13;\n&#13;\n&#13;\n&#13;\n<noscript><img src=\"https:\/\/cmp.d3.sc.omtrdc.net\/b\/ss\/cmpglobalvista\/1\/H.27--NS\/0\" height=\"1\" width=\"1\" alt=\"\" \/><\/noscript>&#13;\n&#13;","created_at":"2019-10-01T20:31:41+0000","updated_at":"2019-10-01T20:31:46+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.gdcvault.com","preview_picture":"https:\/\/twvideo01.ubm-us.net\/o1\/vault\/gdc2019\/images\/gdc19-thumbnail.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1040,"title":"Event Modeling: What is it?","url":"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/","content":"<h2 id=\"motivation\">Motivation<\/h2><h3 id=\"moore-s-law\">Moore\u2019s Law<\/h3><p>Digitized Information Systems are a relatively new concept. Humans have been working with information systems for thousands of years. Over centuries banks, insurance companies and many other large scale organizations have managed to succeed.<\/p><p>With the advent of the transistor, the speed and accuracy of processing information increased by orders of magnitude. What did not gain the same quantum leap is digital storage. This imbalance caused information systems to be optimized for a very small amount of online information. You can see this in the advent of RDBMS technology. What it mean is that the compromise was to throw information away.<\/p><h3 id=\"human-memory\">Human Memory<\/h3><p>Story telling is something that enables humans to pass knowledge on to subsequent generations and relies heavily on how we store memories - whether logical, visual, auditory or other. This is important because there is a parallel with how information systems were constructed. There is a \u201cmemory\u201d of all your visits to the doctor. It\u2019s the ledger of the forms that are filled in with each visit.<\/p><p>Specifications by example are a way to show how something is supposed to work. This can be seen in successful practices in software such as Behaviour Driven Development. This works well because we communicate by stories more effectively. It ties back to story telling as a way to keep information in society. Our brains are built for it more than they are built for flow-charts and other formats.<\/p><h2 id=\"life-after-the-dawn-of-the-computer-age\">Life After the Dawn of the Computer Age<\/h2><p>In recent decades, Moore\u2019s Law from the side of online storage has caught up. This means that after the initial few decades of living with computer systems, our information systems that are now digitized can use the mechanics that made them effective throughout history.<\/p><p>This means we have enough storage to not throw away information. The ability to be able to keep a history of all that has happened allows systems to be more reliable by means of audit and specification by example that literally translates to how the system is implemented.<\/p><p>We also have enough storage to have a cache of different views into what has happened in the system. This is important as we now have made the task of trying to fit all our concerns into one model an unnecessary constraint. In 1956, an IBM harddrive that stored 10MB cost $1M and required $30K monthly budget.<\/p><h3 id=\"reality-of-current-tooling\">Reality of Current Tooling<\/h3><p>So we are now at a cross-roads where we have very mature tooling, but that tooling is made for solving a problem we no longer have - being efficient with storage constraints. The new tooling that we see on the rise is what information systems always had: a ledger of what happened - storage is not a major issue anymore. There are many benefits to keeping ledger. They represent the natural way we think about systems - digital or not.<\/p><h2 id=\"the-model-that-works\">The Model That Works<\/h2><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/blueprint.jpg\" alt=\"blueprint\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/blueprint_large.jpg\">high res version<\/a><\/sub><\/p><p>Time is a concept that is now a core piece of describing a system. The components and classes that we saw in computing are not as important. We can show, by example, what a system is supposed to do from start to finish, on a time line and with no branching - again to make use of that memory aspect of our brains. This is the Event Model. It is used to follow all field values in the UI to the storage of those value to where they finally end up on a report or a screen. It\u2019s generally done with sticky notes on a wall or whiteboard - or an online version of a whiteboard. We\u2019ll see that simplicity is at the heart of the approach as we will only use 3 types of building blocks as well as traditional wireframes or mockups. Further to keep things simple, we will rely on only 4 patterns of how we structure the diagram.<\/p><h3 id=\"simplicity\">Simplicity<\/h3><p>When we want to adopt certain practices or processes to help one another understand and communicate, it is inversely proportional to the amount of learning individuals must do to be proficient in those methods. Put in another way, if an organization chooses to adopt a process called \u201cX\u201d, and X requires one book and a workshop that takes a week to go through, it nullifies the effectiveness of X, and here\u2019s the worst part, <strong>no matter how good X is<\/strong>.<\/p><p>When the book is a required reading by the people in an organization, everyone will say they have read it; only half will have actually read it; half of those will claim they understood it; and only half of those will have understood it; and half of those will be able to apply it.<\/p><p>This is why Event Modeling only uses 3 moving pieces and 4 patterns based on 2 ideas. It takes a few minutes to explain and the rest of the learning is done in practice, transparently where any deficiencies in the understanding of even those few core ideas are quickly corrected.<\/p><p>This is how you get to an understanding in an organization.<\/p><h3 id=\"events\">Events<\/h3><p>Let\u2019s say we want to design a hotel website for a hotel chain for allowing our customers to book rooms online and for us to schedule cleaning and any other hotel concerns. We can show what events, or facts, are stored on a timeline of the year in that business. We can pretend we have the system already and ask ourselves what facts were stored as we move forward through time.<\/p><h3 id=\"wireframes\">Wireframes<\/h3><p>To bring in the visual part of story-telling we show wireframes or web page mockups across the top. These can be organized in swim-lanes to show different people (or sometimes systems) interacting with our system. We also show any automation here with a symbol like gears to illustrate that the system is doing something. This has an easy to understand set of mechanics of a todo list that a process goes and does and marks items as done. In our hotel example, this could be a payment system or notification system.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/innovate.jpg\" alt=\"innovation\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/innovate_large.jpg\">high res version<\/a><\/sub><\/p><p>At this point we have enough to be able to design some systems with some UX\/UI people. But there are 2 very fundamental pieces that must be added to the blueprint which show 2 core features of any information system: Empowering the user and informing the user.<\/p><h3 id=\"commands\">Commands<\/h3><p>Most information systems must give an ability for a user to affect state of the system. In our example, we must allow the booking of a room to change the system so that we don\u2019t over-book and when that person arrives at that future date, they have a room ready for them.<\/p><p>Intentions to change the system are encapsulated in a command. As opposed to simply saving form data to a table in a database, this allows us to have a non-technical way to show the intentions while allowing any implementation - although certain ones have advantages as we will see.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/empower.jpg\" alt=\"empower\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/empower_large.jpg\">high res version<\/a><\/sub><\/p><p>From the UI and UX perspective this drives a \u201ccommand based UI\u201d which goes a long way into helping make composable UIs. With this pattern, it\u2019s a lot clearer what the transactional boundaries are both from the technical and business perspectives. The hotel guest either registered successfully or not.<\/p><p>When there are nuances to what the prerequisites are for having a command succeed, they are elaborated on \u201cGiven-When-Then\u201d style specifications. This is, again, a way to tell a story of what success looks like. There may be a few of these stories to show how a command can and cannot succeed.<\/p><p>An example might be \u201c<strong>Given<\/strong>: We have registered, and added a payment method, <strong>When<\/strong>: We try to book a room, <strong>Then<\/strong>: a room is booked.\u201d This form of specification is also referred to as \u201cArrange, Act, Assert\u201d and in the UX\/UI world \u201cSituation, Motivation, Value\u201d.<\/p><p>The second part of any information system is the ability to inform the user about the state of the system. Our hotel guest should know about what days are available for certain types of rooms they are interested in staying in. There are usually many of these and support the multi-model aspect of information systems.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/inform.jpg\" alt=\"inform\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/inform_large.jpg\">high res version<\/a><\/sub><\/p><p>A view into the facts already in the system has been changing as these new events were being stored. In our hotel system, this calendar view was being updated as new events that affected inventory were happening. Other views may be for the cleaning staff to see which rooms are ready to be cleaned as events about guests checking out are being stored.<\/p><p>Specifying how a view behaves is very similar to the way we specify how we accept commands with one difference. The views are passive and cannot reject an event after it\u2019s been stored in the system. We have \u201c<strong>Given<\/strong>: hotel is set up with 12 ocean view rooms, ocean view room was booked from April 4th - 12th X 12, <strong>Then<\/strong>: the calendar should show all dates except April 4th - 12th for ocean view availability\u201d.<\/p><h3 id=\"integration\">Integration<\/h3><p>We just covered the first 2 patterns of the 4 that are needed to describe most systems. Systems can get information from other systems and send information to other systems. It would be tempting to force these 2 patterns to be an extension of the first 2 and share the same space. However, these interactions are harder to communicate as they don\u2019t have that human-visible aspect to them and require some higher level patterns.<\/p><h3 id=\"translation\">Translation<\/h3><p>When we have an external system that\u2019s providing us with information, it\u2019s helpful to translate that information into a form that is more familiar in our own system. In our hotel system, we may get events from guests\u2019 GPS coordinates if they opted in to our highly reactive cleaning crew. We would not want to use longitude and latitude pairs as events to specify preconditions in our system. We would rather have events that mean something to us like \u201cGuest left hotel\u201d, \u201cGuest returned to hotel room\u201d.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/understand.jpg\" alt=\"understand\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/understand_large.jpg\">high res version<\/a><\/sub><\/p><p>Often, translations are simple enough to represent as views that get their information from external events. If we don\u2019t use them as any \u201cGiven\u201d parts of tests, the values they store in that view model are simply represented in the command parameters in our state change tests.<\/p><h3 id=\"automation\">Automation<\/h3><p>Our system is going to need to communicate with external services. When the guests in our hotel are paying for their stay when they check out, our system makes a call to a payment processor. We can make the concept of how this occurs with the idea of a \u201ctodo list\u201d for some processor in our system. This todo list shows tasks we need to complete. Our processor goes through that list from time to time (could be milliseconds or days) and sends out a command to the external system to process the payment, as an example. The reply from the external system is then translated into an event that we store back in our system. This way we keep the building blocks that we use in our system as something that\u2019s meaningful to us.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/automate.jpg\" alt=\"automate\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/automate_large.jpg\">high res version<\/a><\/sub><\/p><p>We show this by putting a processor in the top of our blueprint which has the wireframes. This shows that there are things not evident on the screen but are happening behind the scenes. A user may expect a spinning icon to indicate a delay due to background tasks needing to finish. The specification for this has the form of \u201c<strong>Given<\/strong>: A view of the tasks to do, <strong>When<\/strong> This command is launched for each item, <strong>Then<\/strong> These events are expected back.\u201d<\/p><p>In reality, these may be implemented in many different ways such as queues, reactive or real-time constructs. They may even actually be manual todo lists that our employees use. The goal here is to communicate how our system communicates with the outside world when it needs to affect it.<\/p><h2 id=\"workshop-format-the-7-steps\">Workshop Format - The 7 Steps<\/h2><p>Event Modeling is done in 7 steps. We explained the end-goal already. So let\u2019s rewind to the beginning and show how to build up to the blueprint:<\/p><h3 id=\"1-brain-storming\">1. Brain Storming<\/h3><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-1.jpg\" alt=\"Step 1\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-1_large.jpg\">high res version<\/a><\/sub><\/p><p>We have someone explain the goals of the project and other information. The participants then envision what system would look and behave like. They put down all the events that they can conceive of having happened. Here we gently introduce the concept that only state-changing events are to be specified. Often, people will name \u201cguest viewed calendar for room availability\u201d. We put those aside for now - they are not events.<\/p><h3 id=\"2-the-plot\">2. The Plot<\/h3><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-2.jpg\" alt=\"Step 2\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-2_large.jpg\">high res version<\/a><\/sub><\/p><p>Now the task is to create a plausible story made of these events. So they are arranged in a line and everyone reviews this time line to understand that this makes sense as events that happen in order.<\/p><h3 id=\"3-the-story-board\">3. The Story Board<\/h3><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-3.jpg\" alt=\"Step 3\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-3_large.jpg\">high res version<\/a><\/sub><\/p><p>Next, the wireframes or mockups of the story are needed to address those that are visual learners. More importantly, each field must be represented so that the blueprint for the system has the source of and destination of the information represented from the user\u2019s perspective.<\/p><h4 id=\"3-1-ux-concurrency\">3.1 UX Concurrency<\/h4><p>The wireframes are generally put at the top of the blueprint. They can be divided into separate swimlanes to show what each user sees if there is more than one. There are no screens that appear above one another as we need to capture each change in the system state as a separate vertical slice of the blueprint. The different ordering can be shown in the various specifications. If it is core to the system or very important to communicate, alternate workflows will need to be added to the blueprint. This is part of the last step that shows organization but can be done earlier if helpful.<\/p><h3 id=\"4-identify-inputs\">4. Identify Inputs<\/h3><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-4.jpg\" alt=\"Step 4\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-4_large.jpg\">high res version<\/a><\/sub><\/p><p>From the earlier section we saw that we need to show how we enable the user to change the state of the system. This is usually the step in which we do this introduction of these blue boxes. Each time an event is stored due to a users action, we link that to the UI by a command that shows what we are getting from the screen or implicitly from client state if it\u2019s a web application.<\/p><h3 id=\"5-identify-outputs\">5. Identify Outputs<\/h3><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-5.jpg\" alt=\"Step 5\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-5_large.jpg\">high res version<\/a><\/sub><\/p><p>Again looking back at our goals for the blueprint, we now have to link information accumulated by storing events back into the UI via views (aka read-models). These may be things like the calendar view in our hotel system that will show the availability of rooms when a user is looking to book a room.<\/p><h3 id=\"6-apply-conway-s-law\">6. Apply Conway\u2019s Law<\/h3><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-6.jpg\" alt=\"Step 6\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/Step-6_large.jpg\">high res version<\/a><\/sub><\/p><p>Now that we know how information gets in and out of our system, we can start to look at organizing the events themselves into swimlanes. We need to do this to allow the system to exist as a set of autonomous parts that separate teams can own. This allows specialization to happen to a level that we control instead of falling out of the composition of teams. See <a href=\"http:\/\/melconway.com\/Home\/Conways_Law.html\">Conway\u2019s Law<\/a> by Mel Conway.<\/p><h3 id=\"7-elaborate-scenarios\">7. Elaborate Scenarios<\/h3><p>Each workflow step is tied to either a command or a view\/read-model. The specifications were explained earlier on. How we make them is still collaboratively with all participants in the same space. A Give-When-Then or Given-Then can be constructed one after the other very rapidly while being reviewed by multiple role representatives. This allows what is traditionally done as user story writing by a dedicated product owner in isolation in a text format, to be done visually in a very small amount of time collaboratively. What\u2019s very critical here, is that each specification is tied to exactly one command or view.<\/p><h3 id=\"completeness-check\">Completeness Check<\/h3><p>At this time the event model should have every field accounted for. All information has to have an origin and a destination. Events must facilitate this transition and hold the necessary fields to do so. This rigor is what is required to get the most benefits of the technique.<\/p><p>A variation of this is where we don\u2019t do this final check and rely on absorbing the rework costs. There are scenarios where this is desired.<\/p><h2 id=\"project-management\">Project Management<\/h2><p>The final output of the exercise if done to completion is a set of very small projects defined by all the scenarios for each workflow step. They are in a format that allows them to be directly translated to what developers will use to make their unit tests. They are also coupled to the adjacent workflow steps by only the contract.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/parallel.jpg\" alt=\"parallel effort\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/parallel_large.jpg\">high res version<\/a><\/sub><\/p><h3 id=\"strong-contracts\">Strong Contracts<\/h3><p>Many project management, business and coordination issues are mitigated by the fact that we have made explicit contracts as to the shape of the information of when we start a particular step of the workflow and what is the shape of the data when it\u2019s finished. These pre- and post-conditions are what allows the work to be completed in relative isolation and later snap together with the adjoining steps as designed.<\/p><h4 id=\"flat-cost-curve\">Flat Cost Curve<\/h4><p>The biggest impact of using Event Modeling is the flat cost curve of the average feature cost. This is due to the fact that the effort of building each workflow step is not impacted by the development of other workflows. One important thing to understand, is that a workflow step is considered to be repeated on the event model if it uses the same command or view.<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/flat-cost-curve.jpg\" alt=\"flat cost curve\" \/><\/p><p>The impact of this is very far reaching because it is what changes software development back into an engineering practice. It\u2019s what makes creating an information system work like the construction of a house. Features can be created in any order. Traditional development cannot rely on estimates because whether the feature gets developed early on versus later in the project impacts the amount of work required. Reprioritizing work makes any previous estimates unreliable.<\/p><h4 id=\"done-is-done-done-right\">Done is Done Done Right<\/h4><p>When a workflow step is implemented, the act of implementing any other workflow step does not cause the need to revisit this already complete workflow step. It\u2019s the reason that the constant feature cost curve can be realized.<\/p><h4 id=\"estimates-without-estimating\">Estimates without Estimating<\/h4><p>With a constant cost curve, the effort for an organization to implement can simply be measured over many features over time. This is an impartial way to empirically determine the velocity of teams. These numbers are then used to scope, schedule and cost out future projects.<\/p><p>This is the impact of the adoption of Agile practices in the industry to put band-aids over the core issue of lack of design. Because the scope of each set of requirements is now per workflow step, the refactoring step of TDD does not impact other workflow steps in the event model. When we don\u2019t have an event model, refactoring goes unrestricted and previously completed pieces of work have to be adjusted. The more work is already completed, the more that has to be reviewed and adjusted with each new addition as we build the solution.<\/p><h4 id=\"subcontracting\">Subcontracting<\/h4><p>The constant cost curve gives the opportunity to do fixed-cost projects. Once there is a velocity established for a team, you have the cost of the software for your organization. With this number, you now can price out what you are willing to give contractors in pay for each workflow step they complete.<\/p><h5 id=\"guarantees\">Guarantees<\/h5><p>Since each workflow step is protected from being affected by other workflow steps, any deficiencies are to be guaranteed by who is delivering them with non-billable work. So in the case of a subcontractor doing a bad job just to get more billable items done quickly, they will have to have the next hours of work dedicated to fixing deficiencies of work already done before. This evens out their effective rate of pay because they are not working on new delivarables.<\/p><p>This can be carried out over longer periods within an employee engagement by making these metrics available through different checkpoints for performance.<\/p><p>Due to the effective pay self-adjusting to the capability of the individual, it is also a way to on-board new employees and pay them fairly while they are in the probation stage of the engagement. This contract-to-hire process removes the subjective and largely ineffective interview process for technical positions.<\/p><h4 id=\"prioritization\">Prioritization<\/h4><p>Moving work on a schedule as to what steps are going to be implemented first is done without changing the estimated costs of each item. This ensures that prioritization of work has no impact in the total cost also. The constant cost curve is required to allow this \u201cagility\u201d of reprioritizing features.<\/p><h4 id=\"change-management\">Change Management<\/h4><p>When the plans change, we simply adjust the event model. This is usually done by just copying the current one and adjusting. Now we can see where the differences are. If a new piece of information is added to one event, that constitutes a new version of the workflow that creates it. Same with the views. If these have not been implemented yet, they don\u2019t change our estimate. If they are already implemented, they add another unit of work to our plan because it\u2019s considered a replacement. There are a few more rules around this. The end result is a definitive guide for change management.<\/p><h2 id=\"security\">Security<\/h2><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/arrows.jpg\" alt=\"security arrows\" \/><sub><a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/arrows_large.jpg\">high res version<\/a><\/sub><\/p><p>With an event model, the solution shows exactly where, and equally importantly, when sensitive data crosses boundaries. With traditional audits, the number of interviews with staff was time consuming and at risk of missing important areas. Security concerns are addressed most responsibly when the applications have an event model to reference.<\/p><h2 id=\"legacy-systems\">Legacy Systems<\/h2><p>Most of the scenarios that real organizations face is where a system is already in place. The main way to deal with a system that is hard to manage because of complexity and lack of understanding is to either rewrite it or to refactor it while it runs. Both of these are very costly.<\/p><p>A third, less risky option exists: Freeze the old system. With proper buy-in, the organization can agree to not alter the existing system. Instead, dealing with bugs and adding new functionality is done on the side as a side-car solution.<\/p><p>Events can be gathered from the database of the old system and make views of that state - employing the <a href=\"#translation\">translate<\/a> pattern described previously. Y-valve redirection of user action can add new functionality in the side solution. An example which fixes a bug (notice that we use the <a href=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/#automation\">external integration pattern<\/a> and extends the old system to add profile pictures is shown here:<\/p><p><img src=\"https:\/\/eventmodeling.org\/posts\/what-is-event-modeling\/event-modeling-legacy-side-car.jpg\" alt=\"legacy side car event model\" \/><\/p><p>This pattern allows an organization to stop putting energy into the sub-optimal existing system and get unblocked from delivering value via the patterns that enable the benefits of the Event Model.<\/p><h2 id=\"conclusion-for-now\">Conclusion for Now<\/h2><p>Event Modeling is changing how information systems are built. With simple repeatable patterns, information systems are as predicable as engineering efforts should be.<\/p><p>(to be continued)<\/p><p>** This is a periodically updated article that will migrate to a page on the site as a resource.<\/p>","created_at":"2019-09-30T12:16:46+0000","updated_at":"2019-09-30T12:16:46+0000","published_at":"2019-06-23T19:01:58+0000","annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"eventmodeling.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1041,"title":"No title found","url":"http:\/\/mesh-spreadsheet.com\/","content":"&#13;\n&#13;\n    &#13;\n&#13;\n    <div id=\"detail\"><h2>What is Mesh?<\/h2><p>Mesh is a JavaScript code editor that feels like a spreadsheet.&#13;\n    &#13;\n    <\/p><p>It runs in a browser (but doesn't require internet) and can be carried with you on a USB. It:<\/p><ul><li>uses JavaScript as its formula language, and stores sheet files as JavaScript source code&#13;\n    <\/li><li>is lightweight: ~2000 lines of source code (excluding libraries)&#13;\n    <\/li><li>has no install or privacy headaches&#13;\n    <\/li><li>strives for maximum compatibility: runs in Internet Explorer 11+ and is based on ECMAScript 5&#13;\n    <\/li><li>is free (as in beer), open-source and permissively licensed (Apache 2.0).&#13;\n    <\/li><\/ul><p>Mesh is really a 'user experience' layer on top of the JavaScript that's built into browsers.<\/p><h2>How do I use Mesh?<\/h2><p>Download it <a href=\"https:\/\/github.com\/chrispsn\/mesh\/archive\/master.zip\">here<\/a> or try it out <a href=\"http:\/\/mesh-spreadsheet.com\/try-mesh.html\">here<\/a> (takes a couple of seconds to load). Here's the <a href=\"https:\/\/github.com\/chrispsn\/mesh\/blob\/master\/README.md#quick-user-guide\">user guide<\/a>.<\/p><p>Mesh emphasises structured data (arrays, tables, etc) more than Excel and Google Sheets, which can take getting used to.<\/p><h2>Why may I want to use Mesh instead of...<\/h2><h3>Excel or Google Docs?<\/h3><p>Aside from being free, with Mesh you can:<\/p><ul><li>work with data of arbitrary length (including changing the row count of tables at calculation time)&#13;\n    <\/li><li>compare spreadsheets using standard tools such as Word's 'compare docs', or 'diff' them at the command prompt&#13;\n    <\/li><li>integrate sheets within websites, as calculation engines&#13;\n    <\/li><li>manage sheets with the rest of your code base using your existing version control tools (they're just text files)&#13;\n    <\/li><li>use and write custom themes (using CSS).&#13;\n    <\/li><\/ul><h3>Libre Office?<\/h3><p>Mesh fits into &lt;1MB, so it has a much smaller download size. Also, it has a different feature mix.<\/p><p>I'd like to think that Mesh's code base is small enough that anyone can read it and be ready to contribute within an hour or two.<\/p><h3>A 'real' programming language?<\/h3><p>You can write real programs in a spreadsheet! For example, <a href=\"https:\/\/twitter.com\/MeshSpreadsheet\/status\/1019555310363504640\">here's what Mesh's own code looks like<\/a>.<\/p><p>Writing code in a spreadsheet format provides you with immediate feedback and lets you lay out your program on a 2D canvas. I highly recommend trying it.<\/p><hr \/><h2>Roadmap<\/h2><h3>Alpha<\/h3><ul><li>Allow loading and saving&#13;\n    <\/li><li>Anonymous-first. Make it so you don't need to name data to enter it. Right now empty cells generate names, will that change?<\/li>\n    <li>Any code layout changes. &#13;\n        <ul><li>Remove <code>_OUTPUT<\/code> and just add props via code to cells?<\/li>\n            <li>Figure out whether managing locations and other props in a separate object<\/li>\n        <\/ul><\/li><li>Docs:&#13;\n            <ul><li>Create short tutorial of how to use and when better than Excel, Google Sheets, etc. [shortcut list]<\/li>\n                <li>Update README, including better-sized GIFs that show off the features, and interop (messaging)<\/li>\n                <li>Libraries (and how to \"import\" and use without a build process)<\/li>\n            <\/ul><\/li><li>Flush out any remaining cache logic issues, eg:&#13;\n            <ul><li>Don't recalculate if the cell's a primitive value and it didn't change&#13;\n                <\/li><li>Let's say we override b, then we send in new data in a subsequent recalc. Should b now return to normal?&#13;\n                <\/li><li>What if a cell happens to recalc to the same value as before? Should we not invalidate its calc chain from that point?&#13;\n                <\/li><li>Deleting names has a bug&#13;\n            <\/li><\/ul><\/li><li>Add Mesh to Product Hunt and <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/README.md#office-suites\">this free office suite list<\/a> and other 'self-hosted' app lists<\/li>\n    <li>Cleanup:&#13;\n        <ul><li>Figure out whether need to add Apache license to each source file? (+ remove NOTICE?)<\/li>\n            <li>Update boilerplate with latest from app.js<\/li>\n            <li>Scan code for TODOs and comments<\/li>\n            <li>Fix themes being gross in some browsers<\/li>\n            <li>Finish light theme<\/li>\n        <\/ul><\/li><li>Get 'new' working again (as in <code>=new Object<\/code>)<\/li>\n    <li>When creating name, check for valid symbols and existing names<\/li>\n    <li>Make sure edit without replacement works on strings, dates etc<\/li>\n    <li>Add back array and object literal support? If so, stop extra brackets being added when add new key to object literal?<\/li>\n\n    <\/ul><h3>Beta<\/h3><ul><li>Undo\/redo<\/li>\n    <li>RO table support<\/li>\n    <li>IO and sharing (use <a href=\"https:\/\/github.com\/chrispsn\/monotreme\">Monotreme<\/a>?):&#13;\n        <ul><li>Share sheets via url hash<\/li>\n            <li>Override cells via url querystring<\/li>\n            <li>WebDAV and remote servers<\/li>\n        <\/ul><\/li><li>Add highlighting behind occupied cells to show they're taken (eg if empty string or renders to eqivalent)<\/li>\n    <li>Style regex differently<\/li>\n    <li>Make a Mesh electron package<\/li>\n    <li>Click on cells to reference in formula (or move cursor with arrow keys)<\/li>\n    <li><a href=\"https:\/\/twitter.com\/MeshSpreadsheet\/status\/1010857552551206912\">Aggregations - SUM, AVERAGE, etc.<\/a> Predefine these functions? Have the UI insert them automatically?<\/li>\n    <li>Make it easier to manage dates<\/li>\n    <li><a href=\"https:\/\/news.ycombinator.com\/item?id=11874367\">Why lookup tables are awesome (incl use of 'find')<\/a><\/li> \n    <li>Copy and paste data in\/out<\/li>\n    <li>Guide on setting up a WebDAV server<\/li>\n    <li>Move around cell locations via dragging mouse<\/li>\n    <li>Split reducer table so that each 'state component' has its own column (separate fns?), eg formula bar, cells, etc<\/li>\n    <li>Edit table header to rename<\/li>\n\n    <\/ul><h3>Miscellaneous<\/h3><ul><li>Empty string entry<\/li>\n    <li>Change add row\/col labels so take up less height\/width?<\/li>\n    <li>Figure out how to manage row\/col widths&#13;\n    <\/li><li>Add row\/column indices (useful for counting \/ IDing). Similarly, way to see the coords of the selected cell?&#13;\n    <\/li><li>Excel -&gt; JS function equivalent table&#13;\n    <\/li><li>APL\/K support&#13;\n    <\/li><li>Version that supports TypeScript or all the latest JS features (bigints, realms, etc)&#13;\n    <\/li><li>Figure out how to manage Recast being pasted into the spreadsheet app (how to speed up parsing on change?)&#13;\n    <\/li><li>What to display for cells that have an invisible string? What about objects that have no string rep (eg Array.prototype)?&#13;\n    <\/li><li>Center function symbol in cell?&#13;\n    <\/li><li>Overflow - make expand to literal size?<\/li>\n    <li>Boolean cells - hardcode vs softcode (take into account colour vs background)<\/li>\n    <li>Should there be a function that lets you specify which cells you want the values of? IE don't get it all, all the time (feels wasteful)<\/li>\n    <li><a href=\"http:\/\/archive.vector.org.uk\/art10004590\">APLers don't use libraries<\/a><\/li>\n\n    <\/ul><\/div>&#13;\n&#13;","created_at":"2019-09-30T12:16:46+0000","updated_at":"2019-09-30T12:16:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"mesh-spreadsheet.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1039,"title":"VCV Rack - Open-source virtual modular synthesizer","url":"https:\/\/vcvrack.com","content":"<div class=\"modal\" id=\"login-register-modal\"><div class=\"modal-container\"><form class=\"form\" id=\"register-form\" action=\"action\">\n          <h4>Register<\/h4><p>Registering allows you to add plugins (free, commercial, and third-party) and install them automatically within Rack.<\/p>\n          <p>\n            <label class=\"label\" for=\"register-email\">Email:<\/label>\n          <\/p>\n          \n          <p>\n            <label class=\"label\" for=\"register-password\">Password:<\/label>\n          <\/p>\n          \n          <p>\n            <label class=\"label\" for=\"register-password-again\">Password (again):<\/label>\n          <\/p>\n          \n          <p>\n            <label class=\"label-inline\" for=\"newsletter\">Receive infrequent emails about VCV updates<\/label>\n          <\/p>\n          \n          \n        <\/form><form class=\"form\" id=\"login-form\" action=\"action\">\n          <h4>Log in<\/h4>\n          <p>\n            <label class=\"label\" for=\"login-email\">Email:<\/label>\n          <\/p>\n          \n          <p>\n            <label class=\"label\" for=\"login-password\">Password:<\/label>\n          <\/p>\n          \n          \n          \n        <\/form><p><a id=\"login-forgot\" href=\"#\">I forgot my password<\/a><\/p><p>Email <a href=\"mailto:contact@vcvrack.com\">contact@vcvrack.com<\/a> for account assistance.<\/p><p><a href=\"https:\/\/vcvrack.com\/privacy.html\">Privacy Policy<\/a><\/p><\/div><\/div><div class=\"modal\" id=\"settings-modal\"><div class=\"modal-container\"><form class=\"form\" id=\"settings-form\" action=\"action\">\n          <h4>Settings<\/h4>\n          <p>\n            <label class=\"label\" for=\"settings-email\">New email (leave blank if no change):<\/label>\n          <\/p>\n          \n          <p>\n            <label class=\"label\" for=\"settings-password\">New password (leave blank if no change):<\/label>\n          <\/p>\n          \n          <p>\n            <label class=\"label\" for=\"settings-password-again\">New password (again):<\/label>\n          <\/p>\n          \n          \n          <p>\n            <label class=\"label-inline\" for=\"settings-newsletter\">Receive infrequent emails about VCV updates<\/label>\n          <\/p><p>Email <a href=\"mailto:contact@vcvrack.com\">contact@vcvrack.com<\/a> for account assistance.<\/p>\n          \n        <\/form><\/div><\/div>","created_at":"2019-09-28T15:33:46+0000","updated_at":"2019-09-28T15:33:46+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"vcvrack.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","elisp","comint"],"is_public":false,"id":1037,"title":"Get tab completion in custom comint mode","url":"https:\/\/stackoverflow.com\/questions\/20396698\/get-tab-completion-in-custom-comint-mode","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I usually try to start CLI's (might not be defining it properly) using commands like:<\/p>\n<pre>(make-comint-in-buffer \"PUTTY\" nil \".\/plink.exe\" nil \"dbserver\")\n(make-comint-in-buffer \"Python\" nil \"c:\/Python26\/python.exe\" nil \"-i\")\n(make-comint-in-buffer \"Python\" nil \"c:\/Python27\/python.exe\" nil \"-i\")\n(make-comint-in-buffer \"git\" nil \"C:\/Program Files (x86)\/Git\/bin\/sh.exe\" nil \"--login\" \"-i\")\n<\/pre>\n<p>Starting the above specially the first and the last in command prompt provides the facility of <kbd>tab<\/kbd> completion. But in <code>comint-mode<\/code> I don't seem to get this feature.<\/p>\n<p>Entering <kbd>tab<\/kbd> adds a simple tab character under the point instead of passing it to the process and getting back the completion candidates. <code>comint-dynamic-complete<\/code> and <kbd>C-q<\/kbd><kbd>tab<\/kbd> doesn't produce the desired result.<\/p>\n<p>How can I utilize in above said feature in <code>comint-mode<\/code> running different processes?<\/p><\/div><div class=\"user-action-time\">&#13;\n        asked Dec 5 '13 at 9:54&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1309352\/bleeding-fingers\">Bleeding Fingers<\/a>Bleeding Fingers&#13;\n        <div class=\"-flair\">&#13;\n            4,62777 gold badges3535 silver badges6060 bronze badges&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"1\">&#13;\n                                    1 Answer&#13;\n                                1&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<h2>Short answer<\/h2>\n<p>You need to write your own completion functions inside Emacs. These may pull info in from the external program, but you can't have direct access to the tab-completion provided by python or git etc.<\/p>\n<h2>Long answer:<\/h2>\n<p>When you run <code>make-comint-in-buffer<\/code>, you are setting up the most basic setup for interactively sending and receiving data from an external process. Whenever you press <code>enter<\/code>, Emacs sends the current line to the underlying process, and then prints back the output from the process. The external process doesn't know what you're doing until Emacs sends a full line to it - no single characters are passed directly.<\/p>\n<p>This is different from running the external program directly in a terminal. When you run python in a terminal, it interprets every character as you type it (or at least it can). So when you hit tab it knows to try to complete something. When you run python inside Emacs, Emacs is processing all of your input until you hit enter. Which means when you hit tab, Emacs just inserts a tab character, and python doesn't know to try and complete something.<\/p>\n<p>Emacs can provide tab-completion for comint modes, but you need to build the completion targets on the Emacs side, you can't rely on the native completion of the external program. To get an idea of how this works, take a look at the completion functions in shell.el, which you can open with <code>M-x find-library shell<\/code>. <\/p>\n<p>This isn't something you can do with a line or two of elisp. It's an important part of major-modes written to interact with external processes - things like ESS for R <a href=\"http:\/\/ess.r-project.org\/\" rel=\"nofollow\">http:\/\/ess.r-project.org\/<\/a><\/p>\n<p>That said, if you dig into the guts of shell.el you may find enough code you can modify to get you started.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Dec 6 '13 at 21:42&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/523044\/tyler\">Tyler<\/a>Tyler&#13;\n        <div class=\"-flair\">&#13;\n            8,70711 gold badge2626 silver badges5050 bronze badges&#13;\n        <\/div>&#13;<\/div>","created_at":"2019-09-25T21:20:16+0000","updated_at":"2019-09-25T21:20:25+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1036,"title":"PDF","url":"https:\/\/www.seas.upenn.edu\/~zives\/03f\/cis550\/codd.pdf","content":"Information Retrieval \tP. BAXENDALE, Editor \t<br \/>\nA Relational Model of Data for <br \/>\nLarge Shared Data Banks <br \/>\nE. F. CODD <br \/>\nIBM Research Laboratory, San Jose, California <br \/>\nFuture users of large data banks must be protected from <br \/>\nhaving to know how the data is organized in the machine (the <br \/>\ninternal representation). A prompting service which supplies <br \/>\nsuch information is not a satisfactory solution. Activities of users <br \/>\nat terminals and most application programs should remain <br \/>\nunaffected when the internal representation of data is changed <br \/>\nand even when some aspects of the external representation <br \/>\nare changed. Changes in data representation will often be <br \/>\nneeded as a result of changes in query, update, and report <br \/>\ntraffic and natural growth in the types of stored information. <br \/>\nExisting noninferential, formatted data systems provide users <br \/>\nwith tree-structured files or slightly more general network <br \/>\nmodels of the data. In Section 1, inadequacies of these models <br \/>\nare discussed. A model based on n-ary relations, a normal <br \/>\nform for data base relations, and the concept of a universal <br \/>\ndata sublanguage are introduced. In Section 2, certain opera- <br \/>\ntions on relations (other than logical inference) are discussed <br \/>\nand applied to the problems of redundancy and consistency <br \/>\nin the user\u2019s model. <br \/>\nKEY WORDS AND PHRASES: data bank, data base, \t<br \/>\ndata \tstructure, data <br \/>\norganization, hierarchies of data, networks of data, relations, derivability, <br \/>\nredundancy, consistency, composition, join, retrieval language, predicate <br \/>\ncalculus, security, data integrity <br \/>\nCR CATEGORIES: 3.70, 3.73, 3.75, 4.20, 4.22, 4.29 \t<br \/>\n1. Relational Model and Normal Form 1 .I. INTR~xJ~TI~N <br \/>\nThis paper is concerned with the application of ele- <br \/>\nmentary relation theory to systems which provide shared <br \/>\naccess to large banks of formatted data. Except for a paper <br \/>\nby Childs [l], the principal application of relations to data <br \/>\nsystems has been to deductive question-answering systems. <br \/>\nLevein and Maron [2] provide numerous references to work <br \/>\nin this area. <br \/>\nIn contrast, the problems treated here are those of data <br \/>\nindependence-the independence of application programs <br \/>\nand terminal activities from growth in data types and <br \/>\nchanges in data representation-and certain kinds of data <br \/>\ninconsistency which are expected to become troublesome <br \/>\neven in nondeductive systems. <br \/>\nVolume 13 \/ Number 6 \/ June, 1970 The relational view (or model) of data described in <br \/>\nSection 1 appears to be superior in several respects to the <br \/>\ngraph or network model [3,4] presently in vogue for non- <br \/>\ninferential systems. It provides a means of describing data <br \/>\nwith its natural structure only-that is, without superim- <br \/>\nposing any additional structure for machine representation <br \/>\npurposes. Accordingly, it provides a basis for a high level <br \/>\ndata language which will yield maximal independence be- <br \/>\ntween programs on the one hand and machine representa- <br \/>\ntion and organization of data on the other. <br \/>\nA further advantage of the relational view is that it <br \/>\nforms a sound basis for treating derivability, redundancy, <br \/>\nand consistency of relations-these are discussed in Section <br \/>\n2. The network model, on the other hand, has spawned a <br \/>\nnumber of confusions, not the least of which is mistaking <br \/>\nthe derivation of connections for the derivation of rela- <br \/>\ntions (see remarks in Section 2 on the \u201cconnection trap\u201d). <br \/>\nFinally, the relational view permits a clearer evaluation <br \/>\nof the scope and logical limitations of present formatted <br \/>\ndata systems, and also the relative merits (from a logical <br \/>\nstandpoint) of competing representations of data within a <br \/>\nsingle system. Examples of this clearer perspective are <br \/>\ncited in various parts of this paper. Implementations of <br \/>\nsystems to support the relational model are not discussed. <br \/>\n1.2. DATA DEPENDENCIES IN PRESENT SYSTEMS <br \/>\nThe provision of data description tables in recently de- <br \/>\nveloped information systems represents a major advance <br \/>\ntoward the goal of data independence [5,6,7]. Such tables <br \/>\nfacilitate changing certain characteristics of the data repre- <br \/>\nsentation stored in a data bank. However, the variety of <br \/>\ndata representation characteristics which can be changed <br \/>\nwithout logically impairing some application programs is <br \/>\nstill quite limited. Further, the model of data with which <br \/>\nusers interact is still cluttered with representational prop- <br \/>\nerties, particularly in regard to the representation of col- <br \/>\nlections of data (as opposed to individual items). Three of <br \/>\nthe principal kinds of data dependencies which still need <br \/>\nto be removed are: ordering dependence, indexing depend- <br \/>\nence, and access path dependence. In some systems these <br \/>\ndependencies are not clearly separable from one another. <br \/>\n1.2.1. Ordering Dependence. Elements of data in a <br \/>\ndata bank may be stored in a variety of ways, some involv- <br \/>\ning no concern for ordering, some permitting each element <br \/>\nto participate in one ordering only, others permitting each <br \/>\nelement to participate in several orderings. Let us consider <br \/>\nthose existing systems which either require or permit data <br \/>\nelements to be stored in at least one total ordering which is <br \/>\nclosely associated with the hardware-determined ordering <br \/>\nof addresses. For example, the records of a file concerning <br \/>\nparts might be stored in ascending order by part serial <br \/>\nnumber. Such systems normally permit application pro- <br \/>\ngrams to assume that the order of presentation of records <br \/>\nfrom such a file is identical to (or is a subordering of) the \tCommunications of the ACM 377<br \/>\n<br \/>\nstored ordering. Those application programs which take <br \/>\nadvantage of the stored ordering of a file are likely to fail <br \/>\nto operate correctly if for some reason it becomes necessary <br \/>\nto replace that ordering by a different one. Similar remarks <br \/>\nhold for a stored ordering implemented by means of <br \/>\npointers. <br \/>\nIt is unnecessary to single out any system as an example, <br \/>\nbecause all the well-known information systems that are <br \/>\nmarketed today fail to make a clear distinction between <br \/>\norder of presentation on the one hand and stored ordering <br \/>\non the other. Significant implementation problems must be <br \/>\nsolved to provide this kind of independence. <br \/>\n1.2.2. Indexing Dependence. In the context of for- <br \/>\nmatted data, an index is usually thought of as a purely <br \/>\nperformance-oriented component of the data representa- <br \/>\ntion. It tends to improve response to queries and updates <br \/>\nand, at the same time, slow down response to insertions <br \/>\nand deletions. From an informational standpoint, an index <br \/>\nis a redundant component of the data representation. If a <br \/>\nsystem uses indices at all and if it is to perform well in an <br \/>\nenvironment with changing patterns of activity on the data <br \/>\nbank, an ability to create and destroy indices from time to <br \/>\ntime will probably be necessary. The question then arises: <br \/>\nCan application programs and terminal activities remain <br \/>\ninvariant as indices come and go? <br \/>\nPresent formatted data systems take widely different <br \/>\napproaches to indexing. TDMS [7] unconditionally pro- <br \/>\nvides indexing on all attributes. The presently released <br \/>\nversion of IMS [5] provides the user with a choice for each <br \/>\nfile: a choice between no indexing at all (the hierarchic se- <br \/>\nquential organization) or indexing on the primary key <br \/>\nonly (the hierarchic indexed sequent,ial organization). In <br \/>\nneither case is the user\u2019s application logic dependent on the <br \/>\nexistence of the unconditionally provided indices. IDS <br \/>\n[8], however, permits the fle designers to select attributes <br \/>\nto be indexed and to incorporate indices into the file struc- <br \/>\nture by means of additional chains. Application programs <br \/>\ntaking advantage of the performance benefit of these in- <br \/>\ndexing chains must refer to those chains by name. Such pro- <br \/>\ngrams do not operate correctly if these chains are later <br \/>\nremoved. <br \/>\n1.2.3. Access Path Dependence. Many of the existing <br \/>\nformatted data systems provide users with tree-structured <br \/>\nfiles or slightly more general network models of the data. <br \/>\nApplication programs developed to work with these sys- <br \/>\ntems tend to be logically impaired if the trees or networks <br \/>\nare changed in structure. A simple example follows. <br \/>\nSuppose the data bank contains information about parts <br \/>\nand projects. For each part, the part number, part name, <br \/>\npart description, quantity-on-hand, and quantity-on-order <br \/>\nare recorded. For each project, the project number, project <br \/>\nname, project description are recorded. Whenever a project <br \/>\nmakes use of a certain part, the quantity of that part com- <br \/>\nmitted to the given project is also recorded. Suppose that <br \/>\nthe system requires the user or file designer to declare or <br \/>\ndefine the data in terms of tree structures. Then, any one <br \/>\nof the hierarchical structures may be adopted for the infor- <br \/>\nmation mentioned above (see Structures l-5). <br \/>\n378 \t<br \/>\nCommunications of the ACM \t<br \/>\nStructure 1. Projects \tSubordinate to Parts <br \/>\nFile <br \/>\nSegment Fields <br \/>\nF PART <br \/>\npart # <br \/>\npart name <br \/>\npart description <br \/>\nquantity-on-hand <br \/>\nquantity-on-order PROJECT <br \/>\nproject # <br \/>\nproject name <br \/>\nproject description <br \/>\nquantity committed <br \/>\nStructure 2. Parts Subordinate to Projects <br \/>\nFile <br \/>\nSqmeut Fields <br \/>\nF PROJECT <br \/>\nproject # <br \/>\nproject name <br \/>\nproject description <br \/>\nPART <br \/>\npart # <br \/>\npart name <br \/>\npart description <br \/>\nquantity-on-hand <br \/>\nquantity-on-order <br \/>\nquantity committed <br \/>\nStructure 3. Parts and Projects as Peers <br \/>\nCommitment Relationship Subordinate to Projects <br \/>\nFile Segment Fields <br \/>\nF PART <br \/>\npart # <br \/>\npart name <br \/>\npart description <br \/>\nquantity-on-hand <br \/>\nquantity-on-order <br \/>\nG PROJECT project # <br \/>\nproject name <br \/>\nproject description <br \/>\nPART <br \/>\npart # <br \/>\nquantity committed <br \/>\nStructure 4. Parts and Projects as Peers <br \/>\nCommitment Relationship Subordinate to Parts <br \/>\nFile \t<br \/>\nSegnren1 Fields F <br \/>\nPART <br \/>\npart # <br \/>\npart description <br \/>\nquantity-on-hand <br \/>\nquantity-on-order <br \/>\nPROJECT project # <br \/>\nquantity committed \t<br \/>\nG \tPROJECT project # <br \/>\nproject name <br \/>\nproject description <br \/>\nStructure 5. Parts, Projects, and <br \/>\nCommitment Relationship as Peers \t<br \/>\nFCZC .%&amp;-,,ZC,,t Ficlds F \t<br \/>\nPART <br \/>\npart # <br \/>\npart name <br \/>\npart description <br \/>\nquantity-on-hand <br \/>\nquantity-on-order <br \/>\nG PROJECT project # <br \/>\nproject name <br \/>\nproject description <br \/>\nH COMMIT <br \/>\npart # \t<br \/>\nproject # <br \/>\nquantity committed \t<br \/>\nVolume 13 \/ Number 6 \/ June, 1970<br \/>\n<br \/>\nNow, consider the problem of printing out the part <br \/>\nnumber, part name, and quantity committed for every part <br \/>\nused in the project whose project name is \u201calpha.\u201d The <br \/>\nfollowing observations may be made regardless of which <br \/>\navailable tree-oriented information system is selected to <br \/>\ntackle this problem. If a program \t<br \/>\nP \tis developed for this <br \/>\nproblem assuming one of the five structures above-that <br \/>\nis, \t<br \/>\nP \tmakes no test to determine which structure is in ef- <br \/>\nfect-then \t<br \/>\nP \twill fail on at least three of the remaining <br \/>\nstructures. More specifically, if \t<br \/>\nP succeeds with structure 5, <br \/>\nit will fail with all the others; if <br \/>\nP \tsucceeds with structure 3 <br \/>\nor 4, it will fail with at least 1,2, and 5; if \t<br \/>\nP \tsucceeds with <br \/>\n1 or 2, it will fail with at least 3, 4, and 5. The reason is <br \/>\nsimple in each case. In the absence of a test to determine <br \/>\nwhich structure is in effect, \t<br \/>\nP \tfails because an attempt is <br \/>\nmade to exceute a reference to a nonexistent file (available <br \/>\nsystems treat this as an error) or no attempt is made to <br \/>\nexecute a reference to a file containing needed information. <br \/>\nThe reader who is not convinced should develop sample <br \/>\nprograms for this simple problem. <br \/>\nSince, in general, it is not practical to develop applica- <br \/>\ntion programs which test for all tree structurings permitted <br \/>\nby the system, these programs fail when a change in <br \/>\n&amp;ructure becomes necessary. <br \/>\nSystems which provide users with a network model of <br \/>\nthe data run into similar difficulties. In both the tree and <br \/>\nnetwork cases, the user (or his program) is required to <br \/>\nexploit a collection of user access paths to the data. It does <br \/>\nnot matter whether these paths are in close correspondence <br \/>\nwith pointer-defined paths in the stored representation-in <br \/>\nIDS the correspondence is extremely simple, in TDMS it is <br \/>\njust the opposite. The consequence, regardless of the stored <br \/>\nrepresentation, is that terminal activities and programs be- <br \/>\ncome dependent on the continued existence of the user <br \/>\naccess paths. <br \/>\nOne solution to this is to adopt the policy that once a <br \/>\nuser access path is defined it will not be made obsolete un- <br \/>\ntil all application programs using that path have become <br \/>\nobsolete. Such a policy is not practical, because the number <br \/>\nof access paths in the total model for the community of <br \/>\nusers of a data bank would eventually become excessively <br \/>\nlarge. <br \/>\n1.3. A \t<br \/>\nRELATIONAL VIEW OF DATA The term relation is used here in its accepted mathe- <br \/>\nmatical sense. Given sets X1 , S, , . . . , S, (not necessarily <br \/>\ndistinct), <br \/>\nR is a relation on these n sets if it is a set of n- <br \/>\ntuples each of which has its first element from S1, its <br \/>\nsecond element from Sz , and so on.\u2019 We shall refer to Si as <br \/>\nthe jth domain of \t<br \/>\nR. As defined above, \tR is said to have <br \/>\ndegree n. Relations of degree 1 are often called unary, de- <br \/>\ngree 2 binary, degree 3 ternary, and degree n n-ary. <br \/>\nFor expository reasons, we shall frequently make use of <br \/>\nan array representation of relations, but it must be re- <br \/>\nmembered that this particular representation is not an es- <br \/>\nsential part of the relational view being expounded. An ar- \t<br \/>\n1 More concisely, R \tis a subset of the Cartesian product 81 X <br \/>\nsz x *.* x 87%. ray which represents an n-ary relation \t<br \/>\nR has the following <br \/>\nproperties : <br \/>\n(1) Each row represents an n-tuple of \t<br \/>\nR. (2) The ordering of rows is immaterial. <br \/>\n(3) All rows are distinct. <br \/>\n(4) The ordering of columns is significant-it corre- <br \/>\nsponds to the ordering S1, Sz , . . . , S, of the do- <br \/>\nmains on which \t<br \/>\nR is defined (see, however, remarks <br \/>\nbelow on domain-ordered and domain-unordered <br \/>\nrelations ) . <br \/>\n(5) The significance of each column is partially con- <br \/>\nveyed by labeling it with the name of the corre- <br \/>\nsponding domain. <br \/>\nThe example in Figure 1 illustrates a relation of degree <br \/>\n4, called supply, which reflects the shipments-in-progress <br \/>\nof parts from specified suppliers to specified projects in <br \/>\nspecified quantities. \t<br \/>\nsupply (supplier part project quantity) \t<br \/>\n1 \t2 <br \/>\n5 17 \t<br \/>\n1 \t3 5 23 <br \/>\n2 3 7 9 <br \/>\n2 7 5 4 <br \/>\n4 <br \/>\n1 1 12 <br \/>\nFIG. \t<br \/>\n1. \tA relation of degree 4 One might ask: If the columns are labeled by the name <br \/>\nof corresponding domains, why should the ordering of col- <br \/>\numns matter? As the example in Figure 2 shows, two col- <br \/>\numns may have identical headings (indicating identical <br \/>\ndomains) but possess distinct meanings with respect to the <br \/>\nrelation. The relation depicted is called component. It is a <br \/>\nternary relation, whose first two domains are called part <br \/>\nand third domain is called quantity. The meaning of com- <br \/>\nponent (2, y, z) is that part x is an immediate component <br \/>\n(or subassembly) of part y, and z units of part 5 are needed <br \/>\nto assemble one unit of part y. It is a relation which plays <br \/>\na critical role in the parts explosion problem. \t<br \/>\ncomponent <br \/>\n(part part quantity) \t<br \/>\n1 \t5 9 <br \/>\n2 <br \/>\n5 7 <br \/>\n3 5 2 <br \/>\n2 6 12 <br \/>\n3 6 3 <br \/>\n4 <br \/>\n7 \t<br \/>\n1 \t<br \/>\n6 7 \t1 <br \/>\nFIG. \t<br \/>\n2. \tA relation with-two identical domains It is a remarkable fact that several existing information <br \/>\nsystems (chiefly those based on tree-structured files) fail <br \/>\nto provide data representations for relations which have <br \/>\ntwo or more identical domains. The present version of <br \/>\nIMS\/360 [5] is an example of such a system. <br \/>\nThe totality of data in a data bank may be viewed as a <br \/>\ncollection of time-varying relations. These relations are of <br \/>\nassorted degrees. As time progresses, each n-ary relation <br \/>\nmay be subject to insertion of additional n-tuples, deletion <br \/>\nof existing ones, and alteration of components of any of its <br \/>\nexisting n-tuples. \t<br \/>\nVolume 13 \/ Number 6 \/ June, \t1970 \tCommunications of the ACM \t379<br \/>\n<br \/>\nIn many commercial, governmental, and scientific data <br \/>\nbanks, however, some of the relations are of quite high de- <br \/>\ngree (a degree of 30 is not at all uncommon). Users should <br \/>\nnot normally be burdened with remembering the domain <br \/>\nordering of any relation (for example, the ordering supplier, <br \/>\nthen part, then project, then quantity in the relation supply). <br \/>\nAccordingly, we propose that users deal, not with relations <br \/>\nwhich are domain-ordered, but with relationships which are <br \/>\ntheir domain-unordered counterparts.2 To accomplish this, <br \/>\ndomains must be uniquely identifiable at least within any <br \/>\ngiven relation, without using position. Thus, where there <br \/>\nare two or more identical domains, we require in each case <br \/>\nthat the domain name be qualified by a distinctive role <br \/>\nname, which serves to identify the role played by that <br \/>\ndomain in the given relation. For example, in the relation <br \/>\ncomponent of Figure 2, the first domain part might be <br \/>\nqualified by the role name sub, and the second by super, so <br \/>\nthat users could deal with the relationship component and <br \/>\nits domains-sub.part \t<br \/>\nsuper.part, quantity-without regard <br \/>\nto any ordering between these domains. <br \/>\nTo sum up, it is proposed that most users should interact <br \/>\nwith a relational model of the data consisting of a collection <br \/>\nof time-varying relationships (rather than relations). Each <br \/>\nuser need not know more about any relationship than its <br \/>\nname together with the names of its domains (role quali- <br \/>\nfied whenever necessary): Even this information might be <br \/>\noffered in menu style by the system (subject to security <br \/>\nand privacy constraints) upon request by the user. <br \/>\nThere are usually many alternative ways in which a re- <br \/>\nlational model may be established for a data bank. In <br \/>\norder to discuss a preferred way (or normal form), we <br \/>\nmust first introduce a few additional concepts (active <br \/>\ndomain, primary key, foreign key, nonsimple domain) <br \/>\nand establish some links with terminology currently in use <br \/>\nin information systems programming. In the remainder of <br \/>\nthis paper, we shall not bother to distinguish between re- <br \/>\nlations and relationships except where it appears advan- <br \/>\ntageous to be explicit. <br \/>\nConsider an example of a data bank which includes rela- <br \/>\ntions concerning \t<br \/>\nparts, projects, and suppliers. One rela- <br \/>\ntion called part is defined on the following domains: <br \/>\n(1) part number <br \/>\n(2) part name <br \/>\n(3) part color <br \/>\n(4) part weight <br \/>\n(5) quantity on hand <br \/>\n(6) quantity on order <br \/>\nand possibly other domains as well. Each of these domains <br \/>\nis, in effect, a pool of values, some or all of which may be <br \/>\nrepresented in the data bank at any instant. While it is <br \/>\nconceivable that, at some instant, all part colors are pres- <br \/>\nent, it is unlikely that all possible part weights, part \t<br \/>\n2 In mathematical terms, a relationship is an equivalence class of <br \/>\nthose relations that are equivalent under permutation of domains <br \/>\n(see Section 2.1.1). <br \/>\n* Naturally, as with any data put into and retrieved from a com- <br \/>\nputer system, the user will normally make far more effective use <br \/>\nof the data if he is aware of its meaning. names, and part numbers are. We shall call the set of <br \/>\nvalues represented at some instant the active domain at that <br \/>\ninstant. <br \/>\nNormally, one domain (or combination of domains) of a <br \/>\ngiven relation has values which uniquely identify each ele- <br \/>\nment (n-tuple) of that relation. Such a domain (or com- <br \/>\nbination) is called a primary key. In the example above, <br \/>\npart number would be a primary key, while part color <br \/>\nwould not be. A primary key is nonredundant if it is either <br \/>\na simple domain (not a combination) or a combination <br \/>\nsuch that none of the participating simple domains is <br \/>\nsuperfluous in uniquely identifying each element. A rela- <br \/>\ntion may possess more than one nonredundant primary <br \/>\nkey. This would be the case in the example if different parts <br \/>\nwere always given distinct names. Whenever a relation <br \/>\nhas two or more nonredundant primary keys, one of them <br \/>\nis arbitrarily selected and called the primary key of that re- <br \/>\nlation. <br \/>\nA common requirement is for elements of a relation to <br \/>\ncross-reference other elements of the same relation or ele- <br \/>\nments of a different relation. Keys provide a user-oriented <br \/>\nmeans (but not the only means) of expressing such cross- <br \/>\nreferences. We shall call a domain (or domain combma- <br \/>\ntion) of relation R a foreign key if it is not the primary key <br \/>\nof R but its elements are values of the primary key of some <br \/>\nrelation S (the possibility that S and R are identical is not <br \/>\nexcluded). In the relation supply of Figure 1, the combina- <br \/>\ntion of supplier, part, project is the primary key, while each <br \/>\nof these three domains taken separately is a foreign key. <br \/>\nIn previous work there has been a strong tendency to <br \/>\ntreat the data in a data bank as consisting of two parts, one <br \/>\npart consisting of entity descriptions (for example, descrip- <br \/>\ntions of suppliers) and the other part consisting of rela- <br \/>\ntions between the various entities or types of entities (for <br \/>\nexample, the supply relation). This distinction is difficult <br \/>\nto maintain when one may have foreign keys in any rela- <br \/>\ntion whatsoever. In the user\u2019s relational model there ap- <br \/>\npears to be no advantage to making such a distinction <br \/>\n(there may be some advantage, however, when one applies <br \/>\nrelational concepts to machine representations of the user\u2019s <br \/>\nset of relationships). <br \/>\nSo far, we have discussed examples of relations which are <br \/>\ndefined on simple domains-domains whose elements are <br \/>\natomic (nondecomposable) values. Nonatomic values can <br \/>\nbe discussed within the relational framework. Thus, some <br \/>\ndomains may have relations as elements. These relations <br \/>\nmay, in turn, be defined on nonsimple domains, and so on. <br \/>\nFor example, one of the domains on which the relation em- <br \/>\nployee is defined might be salary history. An element of the <br \/>\nsalary history domain is a binary relation defined on the do- <br \/>\nmain date and the domain salary. The salary history domain <br \/>\nis the set of all such binary relations. At any instant of time <br \/>\nthere are as many instances of the salary history relation <br \/>\nin the data bank as there are employees. In contrast, there <br \/>\nis only one instance of the employee relation. <br \/>\nThe terms attribute and repeating group in present data <br \/>\nbase terminology are roughly analogous to simple domain <br \/>\n380 \t<br \/>\nCommunications of the ACM <br \/>\nVolume 13 \/ Number 6 \/ June, 1970<br \/>\n<br \/>\nand nonsimple domain, respectively. Much of the confusion <br \/>\nin present terminology is due to failure to distinguish be- <br \/>\ntween type and instance (as in \u201crecord\u201d) and between <br \/>\ncomponents of a user model of the data on the one hand <br \/>\nand their machine representation counterparts on the <br \/>\nother hand (again, we cite \u201crecord\u201d as an example). <br \/>\n1.4. \t<br \/>\nNORMAL FORM A relation whose domains are all simple can be repre- <br \/>\nsented in storage by a two-dimensional column-homo- <br \/>\ngeneous array of the kind discussed above. Some more <br \/>\ncomplicated data structure is necessary for a relation with <br \/>\none or more nonsimple domains. For this reason (and others <br \/>\nto be cited below) the possibility of eliminating nonsimple <br \/>\ndomains appears worth investigating! There is, in fact, a <br \/>\nvery simple elimination procedure, which we shall call <br \/>\nnormalization. <br \/>\nConsider, for example, the collection of relations ex- <br \/>\nhibited in Figure 3 (a). Job history and children are non- <br \/>\nsimple domains of the relation employee. Salary history is a <br \/>\nnonsimple domain of the relation job history. The tree in <br \/>\nFigure 3 (a) shows just these interrelationships of the non- <br \/>\nsimple domains. \t<br \/>\nemployee <br \/>\nI <br \/>\njobhistory <br \/>\nI <br \/>\nsalaryhistory children <br \/>\nemployee (man#, name, birthdate, jobhistory, children) <br \/>\njobhistory (jobdate, title, salaryhistory) <br \/>\nsalaryhistory (salarydate, salary) <br \/>\nchildren (childname, birthyear) \t<br \/>\nFIG. \t3(a). Unnormalized set <br \/>\nemployee\u2019 (man#, name, birthdate) <br \/>\njobhistory\u2019 (man#, jobdate, title) <br \/>\nsalaryhistory\u2019 (man#, jobdate, salarydate, salary) <br \/>\nchildren\u2019 (man#, childname, birthyear) \t<br \/>\nFIG. \t3(b). Normalized set Normalization proceeds as follows. Starting with the re- <br \/>\nlation at the top of the tree, take its primary key and ex- <br \/>\npand each of the immediately subordinate relations by <br \/>\ninserting this primary key domain or domain combination. <br \/>\nThe primary key of each expanded relation consists of the <br \/>\nprimary key before expansion augmented by the primary <br \/>\nkey copied down from the parent relation. Now, strike out <br \/>\nfrom the parent relation all nonsimple domains, remove the <br \/>\ntop node of the tree, and repeat the same sequence of <br \/>\noperations on each remaining subtree. <br \/>\nThe result of normalizing the collection of relations in <br \/>\nFigure 3 (a) is the collection in Figure 3 (b). The primary <br \/>\nkey of each relation is italicized to show how such keys <br \/>\nare expanded by the normalization. \t<br \/>\n4 M. E. Sanko of IBM, San Jose, independently recognized the <br \/>\ndesirability of eliminating nonsimple domains. If normalization as described above is to be applicable, <br \/>\nthe unnormalized collection of relations must satisfy the <br \/>\nfollowing conditions : <br \/>\n(1) The graph of interrelationships of the nonsimple <br \/>\ndomains is a collection of trees. <br \/>\n(2) No primary key has a component domain which is <br \/>\nnonsimple. <br \/>\nThe writer knows of no application which would require <br \/>\nany relaxation of these conditions. Further operations of a <br \/>\nnormalizing kind are possible. These are not discussed in <br \/>\nthis paper. <br \/>\nThe simplicity of the array representation which becomes <br \/>\nfeasible when all relations are cast in normal form is not <br \/>\nonly an advantage for storage purposes but also for com- <br \/>\nmunication of bulk data between systems which use widely <br \/>\ndifferent representations of the data. The communication <br \/>\nform would be a suitably compressed version of the array <br \/>\nrepresentation and would have the following advantages: <br \/>\n(1) It would be devoid of pointers (address-valued or <br \/>\ndisplacement-valued ) . <br \/>\n(2) It would avoid all dependence on hash addressing <br \/>\nschemes. <br \/>\n(3) It would contain no indices or ordering lists. <br \/>\nIf the user\u2019s relational model is set up in normal form, <br \/>\nnames of items of data in the data bank can take a simpler <br \/>\nform than would otherwise be the case. A general name <br \/>\nwould take a form such as <br \/>\nR (g).r.d <br \/>\nwhere R is a relational name; g is a generation identifier <br \/>\n(optional); r is a role name (optional); d is a domain name. <br \/>\nSince g is needed only when several generations of a given <br \/>\nrelation exist, or are anticipated to exist, and r is needed <br \/>\nonly when the relation R has two or more domains named <br \/>\nd, the simple form R.d will often be adequate. <br \/>\n1.5. \t<br \/>\nSOME LINGUISTIC ASPECTS The adoption of a relational model of data, as described <br \/>\nabove, permits the development of a universal data sub- <br \/>\nlanguage based on an applied predicate calculus. A first- <br \/>\norder predicate calculus s&amp;ices if the collection of relations <br \/>\nis in normal form. Such a language would provide a yard- <br \/>\nstick of linguistic power for all other proposed data Ian- <br \/>\nguages, and would itself be a strong candidate for embed- <br \/>\nding (with appropriate syntactic modification) in a variety <br \/>\nof host Ianguages (programming, command- or problem- <br \/>\noriented). While it is not the purpose of this paper to <br \/>\ndescribe such a language in detail, its salient features <br \/>\nwould be as follows. <br \/>\nLet us denote the data sublanguage by R and the host <br \/>\nlanguage by H. R permits the declaration of relations and <br \/>\ntheir domains. Each declaration of a relation identifies the <br \/>\nprimary key for that relation. Declared relations are added <br \/>\nto the system catalog for use by any members of the user <br \/>\ncommunity who have appropriate authorization. H per- <br \/>\nmits supporting declarations which indicate, perhaps less <br \/>\npermanently, how these relations are represented in stor- \t<br \/>\nVolume 13 \/ Number 6 \/ June, 1970 <br \/>\nCommunications of the ACM 381<br \/>\n<br \/>\nage. R permits the specification for retrieval of any subset <br \/>\nof data from the data bank. Action on such a retrieval re- <br \/>\nquest is subject to security constraints. <br \/>\nThe universality of the data sublanguage lies in its <br \/>\ndescriptive ability (not its computing ability). In a large <br \/>\ndata bank each subset of the data has a very large number <br \/>\nof possible (and sensible) descriptions, even when we as- <br \/>\nsume (as we do) that there is only a finite set of function <br \/>\nsubroutines to which the system has access for use in <br \/>\nqualifying data for retrieval. Thus, the class of qualification <br \/>\nexpressions which can be used in a set specification must <br \/>\nhave the descriptive power of the class of well-formed <br \/>\nformulas of an applied predicate calculus. It is well known <br \/>\nthat to preserve this descriptive power it is unnecessary to <br \/>\nexpress (in whatever syntax is chosen) every formula of <br \/>\nthe selected predicate calculus. For example, just those in <br \/>\nprenex normal form are adequate [9]. <br \/>\nArithmetic functions may be needed in the qualification <br \/>\nor other parts of retrieval statements. Such functions can <br \/>\nbe defined in H and invoked in R. <br \/>\nA set so specified may be fetched for query purposes <br \/>\nonly, or it may be held for possible changes. Insertions t,ake <br \/>\nthe form of adding new elements to declared relations with- <br \/>\nout regard to any ordering that may be present in their <br \/>\nmachine representation. Deletions which are effective for <br \/>\nthe community (as opposed to the individual user or sub- <br \/>\ncommunities) take the form of removing elements from de- <br \/>\nclared relations. Some deletions and updates may be trig- <br \/>\ngered by others, if deletion and update dependencies be- <br \/>\ntween specified relations are declared in R. <br \/>\nOne important effect that the view adopted toward data <br \/>\nhas on the language used to retrieve it is in the naming of <br \/>\ndata elements and sets. Some aspects of this have been dis- <br \/>\ncussed in the previous section. With the usual network <br \/>\nview, users will often be burdened with coining and using <br \/>\nmore relat,ion names than are absolutely necessary, since <br \/>\nnames are associated with paths (or path types) rather <br \/>\nthan with relations. <br \/>\nOnce a user is aware that a certain relation is stored, he <br \/>\nwill expect to be able to exploit5 it using any combination <br \/>\nof its arguments as \u201cknowns\u201d and the remaining argu- <br \/>\nments as \u201cunknowns,\u201d because the information (like <br \/>\nEverest) is there. This is a system feature (missing from <br \/>\nmany current informat.ion systems) which we shall call <br \/>\n(logically) symmetric expZoitation of relations. Naturally, <br \/>\nsymmetry in performance is not to be expected. <br \/>\nTo support symmetric exploitation of a single binary re- <br \/>\nlation, two directed paths are needed. For a relation of de- <br \/>\ngree n, the number of paths to be named and controlled is <br \/>\nn factorial. <br \/>\nAgain, if a relational view is adopted in which every n- <br \/>\nary relation (n &gt; 2) has to be expressed by the user as a <br \/>\nnested expression involving only binary relations (see <br \/>\nFeldman\u2019s LEAP System [lo], for example) then 2n - 1 <br \/>\nnames have to be coined instead of only n + 1 with direct <br \/>\nn-ary notation as described in Section 1.2. For example, the <br \/>\n6 Exploiting a relation includes query, update, and delete. 4-ary relation supply of Figure 1, which entails 5 names in <br \/>\nn-ary notation, would be represented in the form <br \/>\nP (supplier, &amp; (part, R (project, quantity))) <br \/>\nin nested binary notation and, thus, employ 7 names. <br \/>\n-4 further disadvantage of this kind of expression is its <br \/>\nasymmetry. Although this asymmetry does not prohibit <br \/>\nsymmetric exploitation, it certainly makes some bases of <br \/>\ninterrogation very awkward for the user to express (con- <br \/>\nsider, for example, a query for those parts and quantities <br \/>\nrelated to certain given projects via &amp; and R). <br \/>\n1.6. \t<br \/>\nEXPRESSIBLE, NAMED, AND STORED RELATIONS Associated with a data bank are two collections of rela- <br \/>\ntions: the <br \/>\nnamed set and the expressible set. The named set <br \/>\nis the collection of all those relations that the community of <br \/>\nusers can identify by means of a simple name (or identifier). <br \/>\nA relation R acquires membership in the named set when a <br \/>\nsuitably authorized user declares R; it loses membership <br \/>\nwhen a suitably authorized user cancels the declaration of <br \/>\nR. <br \/>\nThe expressible set is the total collection of relations that <br \/>\ncan be designated by expressions in the data language. Such <br \/>\nexpressions are constructed from simple names of relations <br \/>\nin the named set; names of generations, roles and domains; <br \/>\nlogical connectives; the quantifiers of the predicate calcu- <br \/>\n1~s;~ and certain constant relation symbols such as = , &gt; . <br \/>\nThe named set is a subset of the expressible set-usually a <br \/>\nvery small subset. <br \/>\nSince some relations in the named set may be time-inde- <br \/>\npendent combinations of others in that set, it is useful to <br \/>\nconsider associating with the named set a collection of <br \/>\nstatements that define these time-independent constraints. <br \/>\nWe shall postpone further discussion of this until we have <br \/>\nintroduced several operations on relations (see Section 2). <br \/>\nOne of the major problems confronting the designer of a <br \/>\ndata system which is to support a relational model for its <br \/>\nusers is that of determining the class of stored representa- <br \/>\ntions to be supported. Ideally, the variety of permitted <br \/>\ndata representations should be just adequate to cover the <br \/>\nspectrum of performance requirements of the total col- <br \/>\nlection of installations. Too great a variety leads to un- <br \/>\nnecessary overhead in storage and continual reinterpreta- <br \/>\ntion of descriptions for the structures currently in effect. <br \/>\nFor any selected class of stored representations the data <br \/>\nsystem must provide a means of translating user requests <br \/>\nexpressed in the data language of the relational model into <br \/>\ncorresponding-and elhcient-actions on the current <br \/>\nstored representation. For a high level data language this <br \/>\npresents a challenging design problem. Nevertheless, it is a <br \/>\nproblem which must be solved-as more users obtain con- <br \/>\ncurrent access to a large data bank, responsibility for pro- <br \/>\nviding efficient response and throughput shifts from the <br \/>\nindividual user to the data system. \t<br \/>\n6 Because each relation in a practical data bank is a finite set at <br \/>\nevery instant of time, the existential and universal quantifiers <br \/>\ncan be expressed in terms of a function that counts the number of <br \/>\nelements in any finite set. 382 \t<br \/>\nCommunications of the ACM Volume 13 \/ Number 6 \/ June, 1970<br \/>\n<br \/>\n2. Redundancy and Consistency 2.1. \t<br \/>\nOPERATIONS ON RELATIONS Since relations are sets, all of the usual set operations are <br \/>\napplicable to them. Nevertheless, the result may not be a <br \/>\nrelation; for example, the union of a binary relation and a <br \/>\nternary relation is not a relation. <br \/>\nThe operations discussed below are specifically for rela- <br \/>\ntions. These operations are introduced because of their key <br \/>\nrole in deriving relations from other relations. Their <br \/>\nprincipal application is in noninferential information sys- <br \/>\ntems-systems which do not provide logical inference <br \/>\nservices-although their applicability is not necessarily <br \/>\ndestroyed when such services are added. <br \/>\nMost users would not be directly concerned with these <br \/>\noperations. Information systems designers and people con- <br \/>\ncerned with data bank control should, however, be thor- <br \/>\noughly familiar with them. <br \/>\n2.1.1. <br \/>\nPermutation. A binary relation has an array <br \/>\nrepresentation with two columns. Interchanging these col- <br \/>\numns yields the converse relation. More generally, if a <br \/>\npermutation is applied to the columns of an n-ary relation, <br \/>\nthe resulting relation is said to be a \t<br \/>\npermutation of the <br \/>\ngiven relation. There are, for example, 4! = 24 permuta- <br \/>\ntions of the relation \t<br \/>\nsupply in Figure 1, if we include the <br \/>\nidentity permutation which leaves the ordering of columns <br \/>\nunchanged. <br \/>\nSince the user\u2019s relational model consists of a collection <br \/>\nof relationships (domain-unordered relations), permuta- <br \/>\ntion is not relevant to such a model considered in isolation. <br \/>\nIt is, however, relevant to the consideration of stored <br \/>\nrepresentations of the model. In a system which provides <br \/>\nsymmetric exploitation of relations, the set of queries <br \/>\nanswerable by a stored relation is identical to the set <br \/>\nanswerable by any permutation of that relation. Although <br \/>\nit is logically unnecessary to store both a relation and some <br \/>\npermutation of it, performance considerations could make <br \/>\nit advisable. <br \/>\n2.1.2. \t<br \/>\nProjection. Suppose now we select certain col- <br \/>\numns of a relation (striking out the others) and then re- <br \/>\nmove from the resulting array any duplication in the rows. <br \/>\nThe final array represents a relation which is said to be a \t<br \/>\nprojection of the given relation. <br \/>\nA selection operator ?r is used to obtain any desired <br \/>\npermutation, projection, or combination of the two opera- <br \/>\ntions. Thus, if \t<br \/>\nL is a list of \tlc indices7 \tL = i1, ii, - . - , ik and \t<br \/>\nR is an n-ary relation (n 2 k ), then rrL \t(R ) is the k-ary <br \/>\nrelation whose jth column is column ii of \t<br \/>\nR (j = 1,2, * * . ,k) <br \/>\nexcept that duplication in resulting rows is removed. Con- <br \/>\nsider the relation \t<br \/>\nsupply of Figure 1. A permuted projection <br \/>\nof this relation is exhibited in Figure 4. Note that, in this <br \/>\nparticular case, the projection has fewer n-tuples than the <br \/>\nrelation from which it is derived. <br \/>\n2.1.3. \t<br \/>\nJoin. Suppose we are given two binary rela- <br \/>\ntions, which have some domain in common. Under what <br \/>\ncircumstances can we combine these relations to form a \t<br \/>\n7 When dealing with relationships, we use domain names (role- <br \/>\nqualified whenever necessary) instead of domain positions. ternary relation which preserves all of the information in <br \/>\nthe given relations? <br \/>\nThe example in Figure 5 shows two relations \t<br \/>\nR, S, which <br \/>\nare joinable without loss of information, while Figure 6 <br \/>\nshows a join of \t<br \/>\nR with S. A binary relation \tR is joinable <br \/>\nwith a binary relation S if there exists a ternary relation U <br \/>\nsuch that 7r12 (U) = \t<br \/>\nR and \t\u20181~23 (U) = S. Any such ternary <br \/>\nrelation is called \t<br \/>\na join of \tR with S. If \tR, S are binary rela- <br \/>\ntions such that ~2 \t<br \/>\n(R) = ~1 (S), then \tR is joinable with S. <br \/>\nOne join that always exists in such <br \/>\na case is the \tnatural join of \t<br \/>\nR with S defined by \t<br \/>\nR*S = {(a, b, c):R(a, b) A S(b, c)) where \t<br \/>\nR (a, b) has the value \ttrue if (a, b) is a member of \tR and similarly for S(b, c). It is immediate that <br \/>\nand \t<br \/>\nTB(R*S) = R <br \/>\nT33(R*S) = S. Note that the join shown in Figure 6 is the natural join <br \/>\nof \t<br \/>\nR with S from Figure 5. Another join is shown in Figure <br \/>\n7. \t<br \/>\nII31 hPPb\/) \t(project supplier) <br \/>\n5 1 <br \/>\n5 2 <br \/>\n1 4 <br \/>\n7 2 <br \/>\nFIG. 4. A permuted projection of the relation in Figure 1 <br \/>\nR (supplier <br \/>\nPart) S (part project) <br \/>\n1 1 <br \/>\n1 1 <br \/>\n2 1 <br \/>\n1 2 <br \/>\n2 2 2 1 \t<br \/>\nFIG. \t5. Two joinable relations <br \/>\nR*S (supplier part project) <br \/>\n1 1 <br \/>\n1 <br \/>\n1 1 <br \/>\n2 <br \/>\n2 <br \/>\n1 1 <br \/>\n2 <br \/>\n1 2 <br \/>\n2 2 1 \t<br \/>\nFIG. \t6. The natural join of R with S (from Figure 5) <br \/>\nU (supplier part project) <br \/>\n1 1 2 <br \/>\n2 1 <br \/>\n1 <br \/>\n2 <br \/>\n2 1 <br \/>\nFIG. \t7. Another join of R with S (from Figure 5) Inspection of these relations reveals an element (ele- <br \/>\nment 1) of the domain \t<br \/>\npart (the domain on which the join <br \/>\nis to be made) with the property that it possesses more <br \/>\nthan one relative under <br \/>\nR and also under S. It is this ele- \t<br \/>\nVolume 13 \/ Number 6 \/ June, 1970 Communications of the ACM \t383<br \/>\n<br \/>\nment which gives rise to the plurality of joins. Such an ele- <br \/>\nment in the joining domain is called a point of ambiguity <br \/>\nwith respect to the joining of R with S. <br \/>\nIf either ~1 (R) or S is a function: no point of ambiguity <br \/>\ncan occur in joining R with S. In such a case, the natural <br \/>\njoin of R with S is the only join of R with S. Note that the <br \/>\nreiterated qualification \u201cof R with S\u201d is necessary, because <br \/>\nS might be joinable with R (as well as R with S), and this <br \/>\njoin would be an entirely separate consideration. In Figure <br \/>\n5, none of the relations R, 7r21(R), S, ?rzl(S) is a function. <br \/>\nAmbiguity in the joining of R with S can sometimes be <br \/>\nresolved by means of other relations. Suppose we are given, <br \/>\nor can derive from sources independent of R and S, a rela- <br \/>\ntion T on the domains project and supplier with the follow- <br \/>\ning properties : <br \/>\n(1) m(T) = m(S), <br \/>\n(2) m(T) = al(R), <br \/>\n(3) T(i s&gt; +~P(R(&amp; P&gt; A S(P,~))~ <br \/>\n(4) \t<br \/>\nR(s, P&gt; -+ 3j(Soj\\,.i) * T(.A s)), (5) <br \/>\nS@,.i) + 3s(T(.?, \ts&gt; A R(s, P&gt;&gt;, then we may form a three-way join of R, S, T; that is, a <br \/>\nternary relation such that <br \/>\nmz(U) = R, \t<br \/>\n7r23(U) \t= s, ml(U) = T. <br \/>\nSuch a join will be called a cyclic 3-join to distinguish it <br \/>\nfrom a linear S-join which would be a quaternary relation <br \/>\nV such that \t<br \/>\nm(V) = R, \tlr23W) \t= s, lr34(V) = T. <br \/>\nWhile it is possible for more than one cyclic 3-join to exist <br \/>\n(see Figures 8,9, for an example), the circumstances under <br \/>\nwhich this can occur entail much more severe constraints <br \/>\nR (s \t<br \/>\nP) s (P 23 T 0\u2019 s) <br \/>\n1 a a d d 1 <br \/>\n2 a <br \/>\n2 b b&amp;Ii d 2 <br \/>\ne 2 <br \/>\nb e e 2 <br \/>\nFIG. 8. Binary relations with a plurality of cyclic 3-joins u b \t<br \/>\nP 8 TJ\u2019 (s P i) <br \/>\n1 a d 1 a d <br \/>\n2 a e 2 a d <br \/>\n2 b d 2 \t<br \/>\na e <br \/>\n2 b e 2 <br \/>\nb d <br \/>\n2 b e <br \/>\nFIG. 9. Two cyclic 3-joins of the relations in Figure 8 than those for a plurality of 2-joins. To be specific, the re- <br \/>\nlations R, S, T must possess points of ambiguity with <br \/>\nrespect to joining R with S (say point x), S with T (say \t<br \/>\n8 A function is a binary relation, which is one-one or many-one, <br \/>\nbut not one-many. y), and T with R (say a), and, furthermore, y must be a <br \/>\nrelative of x under S, z a relative of y under T, and x a <br \/>\nrelative of z under R. Note that in Figure 8 the points <br \/>\n2 = a; y = d; x = 2 have this property. <br \/>\nThe natural linear 3-join of three binary relations R, S, <br \/>\nT is given by <br \/>\nR*S*T = { (a, b, c, d):R (a, b) A S (b, c) A T (c, d)} <br \/>\nwhere parentheses are not needed on the left-hand side be- <br \/>\ncause the natural 2-join (*) is associative. To obtain the <br \/>\ncyclic counterpart, we introduce the operator y which pro- <br \/>\nduces a relation of degree n - 1 from a relation of degree n <br \/>\nby tying its ends together. Thus, if R is an n-ary relation <br \/>\n(n 2 2), the tie of R is defined by the equation <br \/>\nr(R) = {(a~, a2, .-. , a,-l):R(ul, az, ... , a,-~, a,) <br \/>\nA al = a,). <br \/>\nWe may now represent the natural cyclic S-join of R, S, T <br \/>\nby the expression <br \/>\ny (R*S*T). <br \/>\nExtension of the notions of linear and cyclic S-join and <br \/>\ntheir natural counterparts to the joining of n binary rela- <br \/>\ntions (where n 2 3) is obvious. A few words may be ap- <br \/>\npropriate, however, regarding the joining of relations which <br \/>\nare not necessarily binary. Consider the case of two rela- <br \/>\ntions R (degree r ), S (degree s) which are to be joined on <br \/>\np of their domains (p &lt; T, p &lt; s). For simplicity, sup- <br \/>\npose these p domains are the last p of the r domains of R, <br \/>\nand the first p of the s domains of S. If this were not so, we <br \/>\ncould always apply appropriate permutations to make it <br \/>\nso. Now, take the Cartesian product of the first r-p do- <br \/>\nmains of R, and call this new domain A. Take the Car- <br \/>\ntesian product of the last p domains of R, and call this B. <br \/>\nTake the Cartesian product of the last s-p domains of S <br \/>\nand call this C. <br \/>\nWe can treat R as if it were a binary relation on the <br \/>\ndomains A, B. Similarly, we can treat S as if it were a bi- <br \/>\nnary relation on the domains B, C. The notions of linear <br \/>\nand cyclic S-join are now directly applicable. A similar ap- <br \/>\nproach can be taken with the linear and cyclic n-joins of n <br \/>\nrelations of assorted degrees. <br \/>\n2.1.4. Composition. The reader is probably familiar <br \/>\nwith the notion of composition applied to functions. We <br \/>\nshall discuss a generalization of that concept and apply it <br \/>\nfirst to binary relations. Our definitions of composition <br \/>\nand composability are based very directly on the definitions <br \/>\nof join and joinability given above. <br \/>\nSuppose we are given two relations R, S. T is a cam- <br \/>\nposition of R with S if there exists a. join U of R with S such <br \/>\nthat T = aI3 (U) . Thus, two relations are composable if <br \/>\nand only if they are joinable. However, the existence of <br \/>\nmore than one join of R with S does not imply the existence <br \/>\nof more than one composition of R with S. <br \/>\nCorresponding to the natural join of R with S is the <br \/>\n384 \t<br \/>\nCommunications of the ACM <br \/>\nVolume 13 \/ Number 6 \/ June, 1970<br \/>\n<br \/>\nndural composition9 of R with S defined by <br \/>\nR.S = TH(R*S). <br \/>\nTaking the relations R, S from Figure 5, their natural com- <br \/>\nposition is exhibited in Figure 10 and another composition <br \/>\nis exhibited in Figure 11 (derived from the join exhibited <br \/>\nin Figure 7). \t<br \/>\nR. S (project \tsupplier) \t<br \/>\n1 \t1 \t1 \t2 <br \/>\n2 1 \t<br \/>\nFIG. 10. The natural composition of R with S (from Figure 5) <br \/>\nT \t<br \/>\n(project supplier) \t1 \t2 <br \/>\n2 <br \/>\n1 <br \/>\nFIQ. 11. Another composition of R with S (from Figure 5) When two or more joins exist, the number of distinct <br \/>\ncompositions may be as few as one or as many as the num- <br \/>\nber of distinct joins. Figure 12 shows an example of two <br \/>\nrelations which \t<br \/>\nhave several joins but only one composition. <br \/>\nNote that the ambiguity of point c is lost in composing R <br \/>\nwith S, because of unambiguous associations made via the <br \/>\npoints a, b, d, e. \t<br \/>\nR (supplier \tpart) \tS (part \tproject) \t<br \/>\n1 <br \/>\n1 z ; ; <br \/>\n1 c c f <br \/>\n2 <br \/>\n2 8 : is <br \/>\n2 e e f \t<br \/>\nFICA 12. Many joins, only one composition Extension of composition to pairs of relations which are <br \/>\nnot necessarily binary (and which may be of different de- <br \/>\ngrees) follows the same pattern as extension of pairwise <br \/>\njoining to such relations. <br \/>\nA lack of understanding of relational composition has led <br \/>\nseveral systems designers into what may be called the <br \/>\nconnection trap. This \t<br \/>\ntrap may be described in terms of the <br \/>\nfollowing example. Suppose each supplier description is <br \/>\nlinked by pointers to the descriptions of each part supplied <br \/>\nby that supplier, and each part description is similarly <br \/>\nlinked to the descriptions of each project which uses that <br \/>\npart. A conclusion is now drawn which is, in general, er- <br \/>\nroneous: namely that, if all possible paths are followed from <br \/>\na given supplier via the parts he supplies to the projects <br \/>\nusing those parts, one will obtain a valid set of all projects <br \/>\nsupplied by that supplier. Such a conclusion is correct <br \/>\nonly in the very special case that the target relation be- <br \/>\ntween projects and suppliers is, in fact, the natural com- <br \/>\nposition of the other two relations-and we must normally <br \/>\nadd the phrase \u201cfor all time,\u201d because this is usually im- <br \/>\nplied in claims concerning path-following techniques. \t<br \/>\n0 Other writers tend to ignore compositions other than the na- <br \/>\ntural one, and accordingly refer to this particular composition as <br \/>\nthe \tcomposition-see, for example, Kelley\u2019s \u201cGeneral Topology.\u201d 2.1.5. Restriction. A subset of a relation is a relation. <br \/>\nOne way in which a relation S \t<br \/>\nmay act on a relation R to <br \/>\ngenerate a subset of R is through the operation restriction <br \/>\nof R by S. This operation is a generalization of the restric- <br \/>\ntion of a function to a subset of its domain, and is defined <br \/>\nas follows. <br \/>\nLet L, M be equal-length lists of indices such that <br \/>\nL = iI,&amp;., *** <br \/>\n,ik,M = \t<br \/>\njI,j2, e-s , jk where k 5 degree <br \/>\nof R and k 6 degree of S. Then the L, M restriction of R by <br \/>\nS denoted RLIMS is the maximal subset R\u2019 of R such that <br \/>\n?rL(R\u2019) = TM(S). <br \/>\nThe operation is defined only if equality is applicable be- <br \/>\ntween elements of ?T+, (R) on the one hand and rjh (S) on <br \/>\nthe other for all h = 1, 2, . . . , k. <br \/>\nThe three relations R, S, R\u2019 of Figure 13 satisfy the equa- <br \/>\ntion \t<br \/>\nR' = Rw~wsS. <br \/>\nR (8 \t<br \/>\nP \t~3 s \t(P \t8 R' (8 \tP \t3 \t<br \/>\n1aA a A 1aA 2 a \tA c B \t2 \ta A \t2 \ta B b B \t2 \tb B <br \/>\n2bA \t<br \/>\n2 \tb B <br \/>\nFIG. 13. Example of restriction We are now in a position to consider various applications <br \/>\nof these operations on relations. <br \/>\n2.2. \t<br \/>\nREDUNDANCY Redundancy in the named set of relations must be dis- <br \/>\ntinguished from redundancy in the stored set of representa- <br \/>\ntions. We are primarily concerned here with the former. <br \/>\nTo begin with, we need a precise notion of derivability for <br \/>\nrelations. <br \/>\nSuppose 0 is a collection of operations on relations and <br \/>\neach operation has the property that from its operands it <br \/>\nyields a unique relation (thus natural join is eligible, but <br \/>\njoin is not ). A relation R is O-derivable from a set S of rela- <br \/>\ntions if there exists a sequence of operations from the col- <br \/>\nlection 0 which, for all time, yields R from members of S. <br \/>\nThe phrase \u201cfor all time\u201d is present, because we are dealing <br \/>\nwith time-varying relations, and our interest is in derivabil- <br \/>\nity which holds over a significant period of time. For the <br \/>\nnamed set of relationships in noninferential systems, it ap- <br \/>\npears that an adequate collection &amp; contains the following <br \/>\noperations: projection, natural join, tie, and restriction. <br \/>\nPermutation is irrelevant and natural composition need <br \/>\nnot be included, because it is obtainable by taking a natural <br \/>\njoin and then a projection. For the stored set of representa- <br \/>\ntions, an adequate collection e2 of operations would include <br \/>\npermutation and additional operations concerned with sub- <br \/>\nsetting and merging relations, and ordering and connecting <br \/>\ntheir elements. <br \/>\n2.2.1. Strong Redundancy. A set of relations is strongly <br \/>\nredundant if it contains at least one relation that possesses <br \/>\na projection which is derivable from other projections of <br \/>\nrelations in the set. The following two examples are in- <br \/>\ntended to explain why strong redundancy is defined this <br \/>\nway, and to demonstrate its practical use. In the first ex- \t<br \/>\nVolume 13 \/ Number 6 \/ June, 1970 Communications of the ACM 385<br \/>\n<br \/>\nample the collection of relations consists of just the follow- <br \/>\ning relation : <br \/>\nemployee (serial #, name, manager#, managername) <br \/>\nwith serial# as the primary key and manager# as a foreign <br \/>\nkey. Let us denote the active domain by \t<br \/>\nA,, and suppose <br \/>\nthat <br \/>\nA, (munuger#) c A, (serial#) <br \/>\nand <br \/>\nAt (managername) C At (name) <br \/>\nfor all time t. In this case the redundancy is obvious: the <br \/>\ndomain managername is unnecessary. To see that it is a <br \/>\nstrong redundancy as defined above, we observe that <br \/>\nm4 (employee) = ~12 (empZoyee)&amp;r3 (employee). <br \/>\nIn the second example the collection of relations includes a <br \/>\nrelation S describing suppliers with primary key s#, a re- <br \/>\nlation D describing departments with primary key d#, a <br \/>\nrelation J describing projects with primary key j#, and the <br \/>\nfollowing relations : <br \/>\nlJ (s#, d#, - * * &gt;t \t<br \/>\nQ(s#,j#, .--&gt;, R(d#,j#, **-&gt;, <br \/>\nwhere in each case - - - denotes domains other than s#, d#, <br \/>\nj#. Let us suppose the following condition C is known to <br \/>\nhold independent of time: supplier s supplies department <br \/>\nd (relation P ) if and only if supplier s supplies some project <br \/>\nj (relation Q) to which d is assigned (relation R). Then, we <br \/>\ncan write the equation \t<br \/>\nm(P) = m(Q)-w(R) and thereby exhibit a strong redundancy. <br \/>\nAn important reason for the existence of strong re- <br \/>\ndundancies in the named set of relationships is user con- <br \/>\nvenience. A particular case of this is the retention of semi- <br \/>\nobsolete relationships in the named set so that old pro- <br \/>\ngrams that refer to them by name can continue to run cor- <br \/>\nrectly. Knowledge of the existence of strong redundancies <br \/>\nin the named set enables a system or data base adminis- <br \/>\ntrator greater freedom in the selection of stored representa- <br \/>\ntions to cope more efficiently with current traffic. If the <br \/>\nstrong redundancies in the named set are directly reflected <br \/>\nin strong redundancies in the stored set (or if other strong <br \/>\nredundancies are introduced into the stored set), then, gen- <br \/>\nerally speaking, extra storage space and update time are <br \/>\nconsumed with a potential drop in query time for some <br \/>\nqueries and in load on the central processing units. <br \/>\n2.2.2. Weak Redundancy. A second type of redun- <br \/>\ndancy may exist. In contrast to strong redundancy it is not <br \/>\ncharacterized by an equation. A colIection of relations is <br \/>\nweakly redundant if it contains a relation that has a projec- <br \/>\ntion which is not derivable from other members but is at <br \/>\nall times a projection of some join of other projections of <br \/>\nrelations in the collection. <br \/>\nWe can exhibit a weak redundancy by taking the second <br \/>\nexample (cited above) for a strong redundancy, and as- <br \/>\nsuming now that condition C does not hold at all times. The relations al2 (P), 7r12 (Q), 7r12 (R ) are complexlo relations <br \/>\nwith the possibility of points of ambiguity occurring from <br \/>\ntime to time in the potential joining of any two. Under <br \/>\nthese circumstances, none of them is derivable from the <br \/>\nother two. However, constraints do exist between them, <br \/>\nsince each is a projection of some cyclic join of the three of <br \/>\nthem. One of the weak redundancies can be characterized <br \/>\nby the statement: for all time, 1r12 (P) is some composition <br \/>\nof ~12 (Q) with \u2018~~1 (R). The composition in question might <br \/>\nbe the natural one at some instant and a nonnatural one at <br \/>\nanother instant. <br \/>\nGenerally speaking, weak redundancies are inherent in <br \/>\nthe logical needs of the community of users. They are not <br \/>\nremovable by the system or data base administrator. If <br \/>\nthey appear at all, they appear in both the named set and <br \/>\nthe stored set of representations. <br \/>\n2.3. \t<br \/>\nCONSISTENCY Whenever the named set of relations is redundant in <br \/>\neither sense, we shall associate with that set a collection of <br \/>\nstatements which define all of the redundancies which hold <br \/>\nindependent of time between the member relations. If the <br \/>\ninformation system lacks-and it most probably will-de- <br \/>\ntailed semantic information about each named relation, it <br \/>\ncannot deduce the redundancies applicable to the named <br \/>\nset. It might, over a period of time, make attempts to <br \/>\ninduce the redundancies, but such attempts would be fal- <br \/>\nlible. <br \/>\nGiven a collection C of time-varying relations, an as- <br \/>\nsociated set Z of constraint statements and an instantaneous <br \/>\nvalue V for C, we shall call the state (C, 2, V) consistent <br \/>\nor inconsistent according as V does or does not satisfy 2. <br \/>\nFor example, given stored relations R, S, T together with <br \/>\nthe constraint statement \u201cnlz(T) is a composition of <br \/>\n5~12 (R) with ~12 (X)\u201c, we may check from time to time that <br \/>\nthe values stored for R, S, T satisfy this constraint. An al- <br \/>\ngorithm for making this check would examine the first two <br \/>\ncolumns of each of R, S, T (in whatever way they are repre- <br \/>\nsented in the system) and determine whether <br \/>\n(1) ?rl(T) = rl(R), \t<br \/>\n(2) ?rz(T) = \t7r2@), (3) for every element pair (a, c) in the relation al2 (T) <br \/>\nthere is an element b such that (a, b) is in a12(R) <br \/>\nand (b, c) is in 7r12(S). <br \/>\nThere are practica1 problems (which we shall not discuss <br \/>\nhere) in taking an instantaneous snapshot of a collection <br \/>\nof relations, some of which may be very large and highly <br \/>\nvariable. <br \/>\nIt is important to note that consistency as defined above <br \/>\nis a property of the instantaneous state of a data bank, and <br \/>\nis independent of how that state came about. Thus, in <br \/>\nparticular, there is no distinction made on the basis of <br \/>\nwhether a user generated an inconsistency due to an act of <br \/>\nomission or an act of commission. Examination of a simple \t<br \/>\nIO A binary relation is complex if neither it nor its converse is a <br \/>\nfunction. 386 \t<br \/>\nCommunications of the AMC Volume 13 \/ Number 6 \/ June, 1970<br \/>\n<br \/>\nexample will show the reasonableness of this (possibly un- <br \/>\nconventional) approach to consistency. <br \/>\nSuppose the named set C includes the relations S, J, D, <br \/>\nP, Q, R of the example in Section 2.2 and that P, Q, R <br \/>\npossess either the strong or weak redundancies described <br \/>\ntherein (in the particular case now under consideration, it <br \/>\ndoes not matter which kind of redundancy occurs). Further, <br \/>\nsuppose that at some time t the data bank state is consistent <br \/>\nand contains no project j such that supplier 2 supplies <br \/>\nproject j and j is assigned to department 5. Accordingly, <br \/>\nthere is no element (2,5) in ~12 (P). Now, a user introduces <br \/>\nthe element (2, 5) into 7~12 (P) by inserting some appropri- <br \/>\nate element into P. The data bank state is now inconsistent. <br \/>\nThe inconsistency could have arisen from an act of omis- <br \/>\nsion, if the input (2, 5) is correct, and there does exist a <br \/>\nproject j such that supplier 2 supplies j and j is assigned to <br \/>\ndepartment 5. In this case, it is very likely that the user <br \/>\nintends in the near future to insert elements into Q and R <br \/>\nwhich will have the effect of introducing (2, j) into al2 (Q) <br \/>\nand (5, j) in W(R). On the other hand, the input (2, 5) <br \/>\nmight have been faulty. It could be the case that the user <br \/>\nintended to insert some other element into P-an element <br \/>\nwhose insertion would transform a consistent state into <br \/>\na consistent state. The point is that the system will <br \/>\nnormally have no way of resolving this question without <br \/>\ninterrogating its environment (perhaps the user who cre- <br \/>\nated the inconsistency ). <br \/>\nThere are, of course, several possible ways in which a <br \/>\nsystem can detect inconsistencies and respond to them. <br \/>\nIn one approach the system checks for possible inconsist- <br \/>\nency whenever an insertion, deletion, or key update occurs. <br \/>\nNaturally, such checking will slow these operations down. <br \/>\nIf an inconsistency has been generated, details are logged <br \/>\ninternally, and if it is not remedied within some reasonable <br \/>\ntime interval, either the user or someone responsible for <br \/>\nthe security and integrity of the data is notified. Another <br \/>\napproach is to conduct consistency checking as a batch <br \/>\noperation once a day or less frequently. Inputs causing the <br \/>\ninconsistencies which remain in the data bank state at <br \/>\nchecking time can be tracked down if the system main- <br \/>\ntains a journal of all state-changing transactions. This <br \/>\nlatter approach would certainly be superior if few non- <br \/>\ntransitory inconsistencies occurred. <br \/>\n2.4. SUMMARY In Section 1 a relational model of data is proposed as a <br \/>\nbasis for protecting users of formatted data systems from <br \/>\nthe potentially disruptive changes in data representation <br \/>\ncaused by growth in the data bank and changes in traffic. <br \/>\nA normal form for the time-varying collection of relation- <br \/>\nships is introduced. In Section 2 operations on relations and two types of <br \/>\nredundancy are defined and applied to the problem of <br \/>\nmaintaining the data in a consistent state. This is bound to <br \/>\nbecome a serious practical problem as more and more dif- <br \/>\nferent types of data are integrated together into common <br \/>\ndata banks. <br \/>\nMany questions are raised and left unanswered. For <br \/>\nexample, only a few of the more important properties of <br \/>\nthe data sublanguage in Section 1.4 are mentioned. Neither <br \/>\nthe purely linguistic details of such a language nor the <br \/>\nimplementation problems are discussed. Nevertheless, the <br \/>\nmaterial presented should be adequate for experienced <br \/>\nsystems programmers to visualize several approaches. It <br \/>\nis also hoped that this paper can contribute to greater pre- <br \/>\ncision in work on formatted data systems. <br \/>\nAcknowledgment. It was C. T. Davies of IBM Pough- <br \/>\nkeepsie who convinced the author of the need for data <br \/>\nindependence in future information systems. The author <br \/>\nwishes to thank him and also F. P. Palermo, C. P. Wang, <br \/>\nE. B. Altman, and M. E. Senko of the IBM San Jose Re- <br \/>\nsearch Laboratory for helpful discussions. \t<br \/>\nRECEIVED SEPTEMBER, \t1969; \tREVISED FEBRUARY, 1970 \t<br \/>\nREFERENCES <br \/>\n1. \t<br \/>\nCHILDS, \tD. L. Feasibility of a set-theoretical data structure <br \/>\n-a general structure based on a reconstituted definition of <br \/>\nrelation. Proc. IFIP Cong., 1968, North Holland Pub. Co., <br \/>\nAmsterdam, p. 162-172. \t<br \/>\n2. LEVEIN, \tR. \tE., AND MARON, \tM. E. A computer system for <br \/>\ninference execution and data retrieval. Comm. \tACM 10, \t<br \/>\n11 (Nov. 1967), 715-721. \t3. BACHMAN, \tC. W. Software for random access processing. <br \/>\nDatumation (Apr. 1965), 3641. \t<br \/>\n4. MCGEE, \tW. C. Generalized file processing. In \tAnnual Re- <br \/>\nview in Automatic \tProgramming 6, 13, Pergamon Press, <br \/>\nNew York, 1969, pp. 77-149. <br \/>\n5. Information Management System\/360, Application Descrip- <br \/>\ntion Manual H20-0524-1. IBM Corp., White Plains, N. Y., <br \/>\nJuly 1968. <br \/>\n6. GIS (Generalized Information System), Application Descrip- <br \/>\ntion Manual H20-0574. IBM Corp., White Plains, N. Y., <br \/>\n1965. \t<br \/>\n7. BLEIER, \tR. E. Treating hierarchical data structures in the <br \/>\nSDC time-shared data management system (TDMS). <br \/>\nProc. ACM 22nd Nat. Conf., 1967, MD1 Publications, <br \/>\nWayne, Pa., pp. 41-49. <br \/>\n8. IDS Reference Manual GE 625\/635, GE Inform. Sys. Div., <br \/>\nPheonix, Ariz., CPB 1093B, Feb. 1968. \t<br \/>\n9. CHURCH, A. An Introduction to Mathematical Logic I. \tPrince- <br \/>\nton U. Press, Princeton, N.J., 1956. <br \/>\n10. \t<br \/>\nFELDMAN, J. \tA., \tAND ROVNER, \tP. D. An Algol-based associ- <br \/>\native language. Stanford Artificial Intelligence Rep. AI-66, <br \/>\nAug. 1, 1968. \t<br \/>\nVolume 13 \/ Number 6 \/ June, 1970 <br \/>\nCommunications of the ACM 38\u20197","created_at":"2019-09-25T02:08:10+0000","updated_at":"2019-09-25T02:08:10+0000","published_at":"2000-02-11T10:47:51+0000","annotations":[],"mimetype":"application\/pdf","reading_time":50,"domain_name":"www.seas.upenn.edu","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1035,"title":"google\/python-fire","url":"https:\/\/github.com\/google\/python-fire","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p><em>Python Fire is a library for automatically generating command line interfaces\n(CLIs) from absolutely any Python object.<\/em><\/p>\n<ul><li>Python Fire is a simple way to create a CLI in Python. <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/benefits.md#simple-cli\">[1]<\/a><\/li>\n<li>Python Fire is a helpful tool for developing and debugging Python code. <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/benefits.md#debugging\">[2]<\/a><\/li>\n<li>Python Fire helps with exploring existing code or turning other people's code\ninto a CLI. <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/benefits.md#exploring\">[3]<\/a><\/li>\n<li>Python Fire makes transitioning between Bash and Python easier. <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/benefits.md#bash\">[4]<\/a><\/li>\n<li>Python Fire makes using a Python REPL easier by setting up the REPL with the\nmodules and variables you'll need already imported and created. <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/benefits.md#repl\">[5]<\/a><\/li>\n<\/ul><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>To install Python Fire with pip, run: <code>pip install fire<\/code><\/p>\n<p>To install Python Fire with conda, run: <code>conda install fire -c conda-forge<\/code><\/p>\n<p>To install Python Fire from source, first clone the repository and then run:\n<code>python setup.py install<\/code><\/p>\n<h2><a id=\"user-content-basic-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-usage\"><\/a>Basic Usage<\/h2>\n<p>You can call <code>Fire<\/code> on any Python object:<br \/>functions, classes, modules, objects, dictionaries, lists, tuples, etc.\nThey all work!<\/p>\n<p>Here's an example of calling Fire on a function.<\/p>\n<div class=\"highlight highlight-source-python\"><pre>import fire\ndef hello(name=\"World\"):\n  return \"Hello %s!\" % name\nif __name__ == '__main__':\n  fire.Fire(hello)<\/pre><\/div>\n<p>Then, from the command line, you can run:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>python hello.py  # Hello World!\npython hello.py --name=David  # Hello David!\npython hello.py --help  # Shows usage information.<\/pre><\/div>\n<p>Here's an example of calling Fire on a class.<\/p>\n<div class=\"highlight highlight-source-python\"><pre>import fire\nclass Calculator(object):\n  \"\"\"A simple calculator class.\"\"\"\n  def double(self, number):\n    return 2 * number\nif __name__ == '__main__':\n  fire.Fire(Calculator)<\/pre><\/div>\n<p>Then, from the command line, you can run:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>python calculator.py double 10  # 20\npython calculator.py double --number=15  # 30<\/pre><\/div>\n<p>To learn how Fire behaves on functions, objects, dicts, lists, etc, and to learn\nabout Fire's other features, see the <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md\">Using a Fire CLI page<\/a>.<\/p>\n<p>For additional examples, see <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/guide.md\">The Python Fire Guide<\/a>.<\/p>\n<h2><a id=\"user-content-why-is-it-called-fire\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-is-it-called-fire\"><\/a>Why is it called Fire?<\/h2>\n<p>When you call <code>Fire<\/code>, it fires off (executes) your command.<\/p>\n<h2><a id=\"user-content-where-can-i-learn-more\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-can-i-learn-more\"><\/a>Where can I learn more?<\/h2>\n<p>Please see <a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/guide.md\">The Python Fire Guide<\/a>.<\/p>\n<h2><a id=\"user-content-reference\" class=\"anchor\" aria-hidden=\"true\" href=\"#reference\"><\/a>Reference<\/h2>\n<table><thead><tr><th align=\"left\">Setup<\/th>\n<th align=\"left\">Command<\/th>\n<th align=\"left\">Notes<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\">install<\/td>\n<td align=\"left\"><code>pip install fire<\/code><\/td>\n<td align=\"left\">\n<\/td><\/tr><\/tbody><\/table><table><thead><tr><th align=\"left\">Creating a CLI<\/th>\n<th align=\"left\">Command<\/th>\n<th align=\"left\">Notes<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\">import<\/td>\n<td align=\"left\"><code>import fire<\/code><\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\">Call<\/td>\n<td align=\"left\"><code>fire.Fire()<\/code><\/td>\n<td align=\"left\">Turns the current module into a Fire CLI.<\/td>\n<\/tr><tr><td align=\"left\">Call<\/td>\n<td align=\"left\"><code>fire.Fire(component)<\/code><\/td>\n<td align=\"left\">Turns <code>component<\/code> into a Fire CLI.<\/td>\n<\/tr><\/tbody><\/table><table><thead><tr><th align=\"left\">Using a CLI<\/th>\n<th align=\"left\">Command<\/th>\n<th align=\"left\">Notes<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md#help-flag\">Help<\/a><\/td>\n<td align=\"left\"><code>command --help<\/code> or <code>command -- --help<\/code><\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md#interactive-flag\">REPL<\/a><\/td>\n<td align=\"left\"><code>command -- --interactive<\/code><\/td>\n<td align=\"left\">Enters interactive mode.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md#separator-flag\">Separator<\/a><\/td>\n<td align=\"left\"><code>command -- --separator=X<\/code><\/td>\n<td align=\"left\">Sets the separator to <code>X<\/code>. The default separator is <code>-<\/code>.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md#completion-flag\">Completion<\/a><\/td>\n<td align=\"left\"><code>command -- --completion [shell]<\/code><\/td>\n<td align=\"left\">Generates a completion script for the CLI.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md#trace-flag\">Trace<\/a><\/td>\n<td align=\"left\"><code>command -- --trace<\/code><\/td>\n<td align=\"left\">Gets a Fire trace for the command.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/docs\/using-cli.md#verbose-flag\">Verbose<\/a><\/td>\n<td align=\"left\"><code>command -- --verbose<\/code><\/td>\n<td align=\"left\">\n<\/td><\/tr><\/tbody><\/table><p><em>Note that these flags are separated from the Fire command by an isolated <code>--<\/code>.<\/em><\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Licensed under the\n<a href=\"https:\/\/github.com\/google\/python-fire\/blob\/master\/LICENSE\">Apache 2.0<\/a> License.<\/p>\n<h2><a id=\"user-content-disclaimer\" class=\"anchor\" aria-hidden=\"true\" href=\"#disclaimer\"><\/a>Disclaimer<\/h2>\n<p>This is not an official Google product.<\/p>\n<\/article>","created_at":"2019-09-24T14:08:03+0000","updated_at":"2019-09-24T14:08:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/1342004?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["keyboard","laptop","repair","x1","lenovo"],"is_public":false,"id":1034,"title":"Lenovo Thinkpad X1 Gen 3 change \/replace keyboard","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=BnQiL4cRL3U","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/BnQiL4cRL3U?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-09-24T11:33:49+0000","updated_at":"2019-09-24T11:33:58+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/BnQiL4cRL3U\/maxresdefault.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["keyboard","repair","x1","lenovo"],"is_public":false,"id":1033,"title":"Once You Know, You Newegg","url":"https:\/\/www.newegg.com\/p\/0GA-001W-003K6?Description=lenovo%20x1%20carbon%20keyboard&cm_re=lenovo_x1_carbon_keyboard-_-9SIAH8W9PY2509-_-Product","content":"&#13;\n                    \t<div><a class=\"noline\" href=\"http:\/\/www.new%65%67%67.com\"><img id=\"logo\" alt=\"Newegg.com - Computers, Computer Parts and Electronics\" border=\"0\" src=\"http:\/\/images10.newegg.com\/WebResource\/Themes\/2005\/Nest\/logo.gif\" title=\"Newegg.com - Computers, Computer Parts and Electronics\" \/><\/a><\/div>&#13;\n&#13;\n                   &#13;\n                 &#13;\n\t\t\t\t\t<p>We're sorry, but our systems have detected the possible use of an automated program to visit Newegg.com. For more information on appropriate use of the Newegg website, please see the excerpt from Newegg.com's Policy &amp; Agreement below. If you feel you've reached this page in error, please send an email to <a href=\"mailto:webmaster@newegg.com\">webmaster@newegg.com<\/a> with your IP address and the error code listed below.<\/p>&#13;\n\t\t\t\t\t<p>Error code: 2<\/p>&#13;\n\t\t\t\t\t<p>Your IP:104.236.87.208<\/p>&#13;\n\t\t\t\t\t&#13;\n<p>You agree to use Newegg.com website (the \"Website\") only for purposes that are permitted by this Policy &amp; Agreement and any applicable law, regulation or generally accepted practices or guidelines in the relevant jurisdictions.  You agree that you will not access or attempt to access the Website other than through the interface that is provided by Newegg.com, unless you have specifically been permitted to do so in a separate written agreement signed by an authorized representative of Newegg.com.  You specifically agree not to access or attempt to access the Website, or any portion thereof, through any automated means, including but not limited to the use of scripts or web crawlers.  You agree that you will not engage in any activity that disrupts or otherwise interferes with the Website (or the servers and networks which are connected to the Website).  You agree that you will not duplicate, reproduce, copy, sell, trade or resell the Website for any purpose.  You agree that you are solely responsible for any breach of your obligations under this Policy &amp; Agreement and for the consequences (including any loss or damage which Newegg.com may suffer) of any such breach.<\/p>&#13;\n\t\t     &#13;\n                 &#13;\n&#13;\n                    &#13;\n                 \t\t  \t&#13;\n            \t\t\t    &#13;\n                        \t\t     All rights reserved.&#13;\n                   \t\t\t&#13;","created_at":"2019-09-24T11:28:35+0000","updated_at":"2019-09-24T11:28:43+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.newegg.com","http_status":"403"},{"is_archived":0,"is_starred":0,"tags":["laptop","repair","x1","lenovo"],"is_public":false,"id":1032,"title":"Lenovo X1 Carbon Keyboard Repair and Replacement","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=kmzYTvRKxj8","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/kmzYTvRKxj8?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-09-24T11:27:25+0000","updated_at":"2019-09-24T11:27:40+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/kmzYTvRKxj8\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["orm","sql","typescript"],"is_public":false,"id":1031,"title":"jawj\/mostly-ormless","url":"https:\/\/github.com\/jawj\/mostly-ormless\/blob\/master\/README.md","content":"<p><em>I gave <a href=\"https:\/\/news.ycombinator.com\/item?id=19853066\" rel=\"nofollow\">a very brief outline of this on HN<\/a> and a few people sounded interested, so here goes. It's extracted from my work as co-founder\/CTO at <a href=\"https:\/\/www.psyt.co.uk\" rel=\"nofollow\">PSYT<\/a>.<\/em><\/p>\n<p>TypeScript and VS Code are <a href=\"https:\/\/stackoverflow.com\/questions\/12694530\/what-is-typescript-and-why-would-i-use-it-in-place-of-javascript\/35048303#35048303\" rel=\"nofollow\">wonderful<\/a>. ORMs are <a href=\"https:\/\/en.wikipedia.org\/wiki\/Object-relational_impedance_mismatch\" rel=\"nofollow\">problematic<\/a>.<\/p>\n<p>Using an ORM within TypeScript can be particularly aggravating, since you can find yourself specifying lots of type information twice over (once to define the database schema, and once to define the associated TypeScript types).<\/p>\n<p>That's why, in the course of converting a project to TypeScript recently, I decided to switch it from <a href=\"http:\/\/docs.sequelizejs.com\/\" rel=\"nofollow\">Sequelize<\/a> to <a href=\"https:\/\/typeorm.io\" rel=\"nofollow\">TypeORM<\/a>. The first day was wonderful: shiny TypeScript toys! The rest of the week was less so: perhaps even more than other ORMs, TypeORM just doesn't seem to <em>get<\/em> SQL.<\/p>\n<p>So, let's back up a bit. What am I really after?<\/p>\n<p>Well, I like and understand Postgres. I like and understand SQL. When I'm working with data, I'd like a good mental model of what SQL is going to be issued. I don't mind writing the SQL myself, but in that case let me <em>write<\/em> the SQL, rather than deal with the endlessly chained method calls of a 'query builder'.<\/p>\n<p>Also, I'm greedy. I want all that, but I want to keep all the type-checking goodness of TypeScript \u2014 speeding development, preventing bugs, simplifying refactoring \u2014\u00a0too. And I'd <em>strongly<\/em> prefer not to spend hours laboriously adding and later maintaining lots of type information to facilitate it.<\/p>\n<p>In short, I want <a href=\"https:\/\/www.theguardian.com\/books\/2018\/jul\/05\/word-of-the-week-cake\" rel=\"nofollow\">cake<\/a>. So I set out to see if I could rustle some up. As it turns out, this was about a week-long project in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Three-act_structure\" rel=\"nofollow\">three acts<\/a>:<\/p>\n<ul><li><em><a href=\"#act1\">Act 1<\/a>: In which we translate a Postgres schema into TypeScript types.<\/em> Spoiler: we'll use a project called <a href=\"https:\/\/github.com\/SweetIQ\/schemats\">schemats<\/a> here.<\/li>\n<li><em><a href=\"#act2\">Act 2<\/a>: In which we significantly expand that type information and use it to improve the ergonomics of writing raw SQL.<\/em> Spoiler: ES2015 <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Template_literals#Tagged_templates\" rel=\"nofollow\">tagged templates<\/a> will play a starring role.<\/li>\n<li><em><a href=\"#act3\">Act 3<\/a>: In which we create a set of simple shortcut functions on top, which looks a little bit like an ORM but really isn't.<\/em> Spoiler: Typescript <a href=\"https:\/\/www.typescriptlang.org\/docs\/handbook\/functions.html\" rel=\"nofollow\">function signature overloads<\/a> on concrete string values will be these functions' secret sauce.<\/li>\n<\/ul><p>(There is a also a brief <a href=\"#encore\">encore<\/a> in which I discuss transactions).<\/p>\n<h2><a id=\"user-content-act-1-in-which-we-translate-a-postgres-schema-into-typescript-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#act-1-in-which-we-translate-a-postgres-schema-into-typescript-types\"><\/a>Act 1: In which we translate a Postgres schema into TypeScript types<\/h2>\n<p>If you want type information that matches a database schema, you have two ways to maintain a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Single_source_of_truth\" rel=\"nofollow\">single source of truth<\/a>.<\/p>\n<p>On the one hand, you can define a schema in TypeScript (or JavaScript), then automagically issue database commands to sync up the tables and columns. That's the approach taken by both Sequelize and TypeScript. This approach helps manage schema updates across multiple databases (e.g. test, staging, production). But I've always been a bit leery of it: I'm not sure I want any code, let alone a third party npm library, issuing <code>CREATE<\/code>\/<code>ALTER<\/code>\/<code>DROP TABLE<\/code> statements at will where production data may be involved.<\/p>\n<p>So, on the other hand, you can take the database schema as your source of truth, then interrogate it to discover the tables and column types. This is what <a href=\"https:\/\/guides.rubyonrails.org\/active_record_basics.html\" rel=\"nofollow\">Active Record<\/a> does (or at least, what it did when I last used it, about 10 years ago), and what I'm going to do here. Managing schema updates across multiple databases is therefore something I'm not going to address.<\/p>\n<p>Active Record interrogates the database at runtime, but I'm going to do it prior to runtime: discovering types only at runtime would be largely pointless, since what we want is type information to ease development.<\/p>\n<p>As it happens, there's already a library that can do most of the heavy lifting we need, called <a href=\"https:\/\/github.com\/SweetIQ\/schemats\">schemats<\/a>.<\/p>\n<p>How does it work? Well, take this simple database schema:<\/p>\n<div class=\"highlight highlight-source-sql\"><pre>CREATE TABLE authors \n( \"id\" SERIAL PRIMARY KEY\n, \"name\" TEXT NOT NULL\n, \"isLiving\" BOOLEAN\n);\nCREATE TABLE books \n( \"id\" SERIAL PRIMARY KEY\n, \"authorId\" INTEGER NOT NULL REFERENCES \"authors\"(\"id\")\n, \"title\" TEXT\n, \"createdAt\" TIMESTAMPTZ NOT NULL DEFAULT now()\n, \"updatedAt\" TIMESTAMPTZ NOT NULL DEFAULT now()\n);<\/pre><\/div>\n<p>If we run vanilla schemats against this, we get the following TypeScript definitions for the <code>books<\/code> table:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>export namespace booksFields {\n    export type id = number;\n    export type authorId = number;\n    export type title = string | null;\n    export type createdAt = Date;\n    export type updatedAt = Date;\n}\nexport interface books {\n    id: booksFields.id;\n    authorId: booksFields.authorId;\n    title: booksFields.title;\n    createdAt: booksFields.createdAt;\n    updatedAt: booksFields.updatedAt;\n}<\/pre><\/div>\n<p>This is a helpful description of what you might get back from a <code>SELECT<\/code> query, and we can use it as follows<\/p>\n<p>First, some housekeeping \u2014\u00a0we set up a connection pool:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>import * as pg from \"pg\";\nconst pool = new pg.Pool({ \/* connection options *\/ });<\/pre><\/div>\n<p>Then:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const\n  authorId = 123,  \/\/ in production this could be untrusted input\n  result = await pool.query({\n    text: 'SELECT * FROM \"books\" WHERE \"authorId\" = $1',\n    values: [authorId],\n  }),\n  existingBooks: books[] = result.rows;<\/pre><\/div>\n<p>Our results now pop out fully typed. But there are still some real annoyances here:<\/p>\n<ul><li>There's nothing to stop me absent-mindedly or fat-fingeredly trying to query from the non-existent tables <code>\"novels\"<\/code> or <code>\"boks\"<\/code>, or according to the non-existent columns <code>\"writerId\"<\/code> or <code>\"authorid\"<\/code>, or indeed trying to equate my <code>\"authorId\"<\/code> with a string or date. There's no auto-complete to help me out, and any errors (whether introduced now or by a future schema update) will only become apparent later on, at runtime.<\/li>\n<li>I have to keep track of the order of interpolated parameters, and provide them separately from the query text. That's only <code>authorId<\/code> in this instance, but there could be dozens.<\/li>\n<li>Things are even worse for <code>INSERT<\/code>s and <code>UPDATE<\/code>s, where I need to generate separate, matching-ordered lists of column names and their corresponding values.<\/li>\n<\/ul><p>So, here's the plan. I'm going to come up with some ES2015 <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Template_literals#Tagged_templates\" rel=\"nofollow\">tagged templates<\/a> to improve these SQL-writing ergonomics in TypeScript. But first, to support that, I'm going to need some more complete type information about my tables.<\/p>\n<h2><a id=\"user-content-act-2-in-which-we-expand-that-type-information-and-use-it-to-improve-the-ergonomics-of-writing-raw-sql\" class=\"anchor\" aria-hidden=\"true\" href=\"#act-2-in-which-we-expand-that-type-information-and-use-it-to-improve-the-ergonomics-of-writing-raw-sql\"><\/a>Act 2: In which we expand that type information and use it to improve the ergonomics of writing raw SQL<\/h2>\n<p>I said I wanted more complete type information about my tables. Specifically, I want the types I generate to specify not just what I'll get back from a <code>SELECT<\/code> query, but also what I'm allowed to use in a <code>WHERE<\/code> condition, and what I can <code>INSERT<\/code> and <code>UPDATE<\/code> too.<\/p>\n<p>So these are the four main interfaces I'm going to define:<\/p>\n<ul><li><strong><code>Selectable<\/code><\/strong>: what I'll get back from a <code>SELECT<\/code> query. This is what schemats was already giving us.<\/li>\n<li><strong><code>Whereable<\/code><\/strong>: what I can use in a <code>WHERE<\/code> condition. This is approximately the same as <code>Selectable<\/code>, but all columns are optional. It is (subject to some later tweaks) a <code>Partial&lt;Selectable&gt;<\/code>.<\/li>\n<li><strong><code>Insertable<\/code><\/strong>: what I can <code>INSERT<\/code> into a table. This is also similar to <code>Selectable<\/code>, but any fields that are <code>NULL<\/code>able and\/or have <code>DEFAULT<\/code> values are allowed to be missing, <code>NULL<\/code> or <code>DEFAULT<\/code>.<\/li>\n<li><strong><code>Updatable<\/code><\/strong>: what I can <code>UPDATE<\/code> a row with. This is similar to what I can <code>INSERT<\/code>, but all columns are optional: it is a simple <code>Partial&lt;Insertable&gt;<\/code>.<\/li>\n<\/ul><p>I <a href=\"https:\/\/github.com\/PSYT\/schemats\">forked schemats<\/a> to generate these interfaces. While I was at it I also got rid of the verbose, original, two-stage approach, where a type alias is created for every column.<\/p>\n<p>Ignoring a few additional bells and whistles, my schemats fork therefore now generates something like the following when run against the <code>books<\/code> table:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>export namespace books {\n  export interface Selectable {\n      id: number;\n      authorId: number;\n      title: string | null;\n      createdAt: Date;\n      updatedAt: Date;\n  }\n  export type Whereable = Partial&lt;Selectable&gt;;\n  export interface Insertable {\n      id?: number | DefaultType;\n      authorId: number;\n      title?: string | null | DefaultType;\n      createdAt?: Date | DefaultType;\n      updatedAt?: Date | DefaultType;\n  }\n  export type Updatable = Partial&lt;Insertable&gt;;\n}\nexport const Default = Symbol('DEFAULT');\nexport type DefaultType = typeof Default;<\/pre><\/div>\n<p>If I'm querying that table, I now type the returned rows as <code>books.Selectable[]<\/code> instead of just <code>books[]<\/code>:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const existingBooks: books.Selectable[] = \/* database query goes here *\/;<\/pre><\/div>\n<p>And if I have an object with data that I plan to insert into the <code>books<\/code> table, or that I plan to use to look up a particular book, I can now also have that type-checked and auto-completed in VS Code, like so:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const newBook: books.Insertable = { \n  authorId: 123,\n  title: \"One Hundred Years of Solitude\",\n};\nconst bookConditions: books.Whereable = {\n  authorId: 456,\n};<\/pre><\/div>\n<p>But how am I going to use these extra types? Enter the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Template_literals#Tagged_templates\" rel=\"nofollow\">tagged templates<\/a> I mentioned earlier.<\/p>\n<p>Tagged templates are pleasingly flexible. Critically, neither a template's interpolated values nor its return value have to be strings. And TypeScript supports tagged templates fully, applying the tag function's signature to the interpolated values.<\/p>\n<p>This is useful because:<\/p>\n<ul><li>I'm going to define a typed tagged template function, <code>sql<\/code>, which we'll use to write queries.<\/li>\n<li>It's going to be generic, so I can tell it which table(s) it targets: e.g. <code>sql&lt;books.SQL&gt;<\/code> <code>`-- database query goes here`<\/code><\/li>\n<li>It's not going to return a string (yet), but an instance of a class I'm going to call <code>SQLFragment<\/code>.<\/li>\n<li>Different kinds of interpolated values will be processed in different ways, and they'll be type-checked and auto-completed appropriately.<\/li>\n<li>An important kind of interpolated values is other <code>SQLFragment<\/code>s.<\/li>\n<li>When it comes to making the query, the root <code>SQLFragment<\/code> will be compiled recursively to produce the appropriate <code>{ text: '', values: [] }<\/code> object that can be passed on to <code>pg<\/code>.<\/li>\n<\/ul><p>It's still going to be possible to write invalid SQL, of course, but a bunch of simple errors (including most of the errors that could be introduced by future refactorings) will be caught at compile time or earlier.<\/p>\n<p><strong><code>SELECT<\/code><\/strong><\/p>\n<p>How does this look in practice? Well, the simple query above ends up looking like this:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const \n  authorId = 123,\n  query = sql&lt;books.SQL&gt;`SELECT * FROM ${\"books\"} WHERE ${{ authorId }}`,\n  existingBooks: books.Selectable[] = await query.run(pool);<\/pre><\/div>\n<p>What's happening here?<\/p>\n<p>Well, first, the table name is now being interpolated, which means it will be type checked and auto-completed. That means: I can pick it from a list; if I get it wrong, I'll be told so immediately; and if I later change its name in the database but forget to change it here, I'll be told so immediately too.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/jawj\/mostly-ormless\/blob\/master\/README-resources\/table-type-check.png\"><img src=\"https:\/\/github.com\/jawj\/mostly-ormless\/raw\/master\/README-resources\/table-type-check.png\" alt=\"Screenshot: table type-checking\" \/><\/a><\/p>\n<p>So the first kind of interpolated value my <code>sql<\/code> template tag function supports is a plain string. This is type-checked\/auto-completed to correspond to a valid table or column name. It will be double-quoted in the generated raw SQL, to protect any idiomatic JavaScript\/TypeScript camelCased column names.<\/p>\n<p>And, second, the object I've passed as my <code>WHERE<\/code> clause is being type-checked\/auto-completed as the appropriate <code>books.Whereable<\/code> type. It will be compiled to <code>(\"authorId\" = $1)<\/code>, and its value will be inserted at the appropriate index of the pg query <code>values<\/code> array.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/jawj\/mostly-ormless\/blob\/master\/README-resources\/column-auto-complete.png\"><img src=\"https:\/\/github.com\/jawj\/mostly-ormless\/raw\/master\/README-resources\/column-auto-complete.png\" alt=\"Screenshot: column auto-completion\" \/><\/a><\/p>\n<p>So the second kind of interpolated value I can use in a <code>sql<\/code> template is a plain object. This is type-checked\/auto-completed as an appropriate <code>Whereable<\/code>, and compiled to a set of <code>WHERE<\/code> conditions connected with <code>AND<\/code>s.<\/p>\n<p>These <code>Whereable<\/code> objects in fact have more flexibility than I let on before, since they also accept <code>SQLFragment<\/code> values. For example, what if I want to see only books for a particular author that were added during the last <em>N<\/em> days?<\/p>\n<p>Then I can write the following:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const\n  authorId = 123,\n  days = 7,\n  query = sql&lt;books.SQL&gt;`\n    SELECT * FROM ${\"books\"} \n    WHERE ${{\n      authorId,\n      createdAt: sql&lt;books.SQL&gt;`\n        ${self} &gt; now() - ${param(period)} * INTERVAL '1 DAY'`,\n    }}`;<\/pre><\/div>\n<p>And to confirm how this is compiled:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>&gt; console.log(query.compile());\n{ text: 'SELECT * FROM \"books\" WHERE (\"authorId\" = $1 AND (\"createdAt\" &gt; now() - $2 * INTERVAL \\'1 DAY\\'))',\n  values: [ 123, 7 ] }<\/pre><\/div>\n<p>You can see the <code>SQLFragment<\/code> nested inside the <code>Whereable<\/code>. You can also see a symbol I've called <code>self<\/code>, which is compiled to its parent column name. Finally, I'm using my <code>param<\/code> wrapper function, which lets me manually insert parameters to be represented as <code>$1<\/code>, <code>$2<\/code> (etc.) and added to the pg query's <code>values<\/code> array.<\/p>\n<p><strong><code>INSERT<\/code> and <code>UPDATE<\/code><\/strong><\/p>\n<p>There are some additional supported interpolation types that can help with <code>INSERT<\/code> and <code>UPDATE<\/code> queries. For example, what if I wanted to save the <code>Insertable<\/code> object I showed you before?<\/p>\n<p>I can write:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const \n  newBook: books.Insertable = { \n    authorId: 123,\n    title: \"One Hundred Years of Solitude\",\n  },\n  query = sql&lt;books.SQL&gt;`\n    INSERT INTO ${\"books\"} (${cols(newBook)})\n    VALUES (${vals(newBook)})`;<\/pre><\/div>\n<p>And this gives the expected:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>&gt; console.log(query.compile());\n{ text: 'INSERT INTO \"books\" (\"authorId\", \"title\") VALUES ($1, $2)',\n  values: [ 123, 'One Hundred Years of Solitude' ] }<\/pre><\/div>\n<p>The <code>cols<\/code> and <code>vals<\/code> wrapper functions ultimately produce identically-ordered sequences of the object's keys (quoted) and values (safely passed via pg's <code>values<\/code> array) respectively.<\/p>\n<p>And, of course, this is all being type-checked and auto-completed as I type it:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/jawj\/mostly-ormless\/blob\/master\/README-resources\/column-title-auto-complete.png\"><img src=\"https:\/\/github.com\/jawj\/mostly-ormless\/raw\/master\/README-resources\/column-title-auto-complete.png\" alt=\"Screenshot: column name auto-completion\" \/><\/a><\/p>\n<p>And if, for example, someone renames the <code>title<\/code> column to <code>name<\/code> in future, TypeScript will complain before I even try to run the program.<\/p>\n<p>For further flexibility, there are three further kinds of interpolated value that are supported in a <code>sql<\/code> template. They are:<\/p>\n<ul><li>The <code>Default<\/code> symbol I mentioned earlier, compiling to the <code>DEFAULT<\/code> keyword.<\/li>\n<li><code>raw()<\/code> \u2014 as in <code>raw(name)<\/code>, where <code>name<\/code> is <code>\"Robert'); -- DROP TABLE Students; --\"<\/code> (<a href=\"https:\/\/www.xkcd.com\/327\/\" rel=\"nofollow\">XKCD<\/a>) \u2014 for times when you really know what you're doing.<\/li>\n<li>Arrays containing any of the other supported interpolation types, whose compiled results are simply concatenated together.<\/li>\n<\/ul><p>Some of this relies on a few further types that I tweaked schemats to generate too:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>export namespace books {\n  \/* ... Selectable, Whereable, Insertable, Updatable, mostly as before,\n    but all except Selectable now also allow SQLFragment values ... *\/\n  export type Table = \"books\";\n  export type Column = keyof Selectable;\n  export type SQLExpression = GenericSQLExpression | Table | Whereable | Column | ColumnNames&lt;Updatable&gt; | ColumnValues&lt;Updatable&gt;;\n  \/\/ all building up to the books.SQL type we saw above:\n  export type SQL = SQLExpression | SQLExpression[];\n}\n\/\/ these are mostly the types of symbols or wrapper classes (not shown)\nexport type GenericSQLExpression = SQLFragment | Parameter | DefaultType | DangerousRawString | SelfType;  <\/pre><\/div>\n<p><strong><code>JOIN<\/code>s<\/strong><\/p>\n<p>We can of course use all of these types to help us in more complicated queries too. Take this <code>JOIN<\/code>, for example, retrieving each book with its (single) author. We can use Postgres's excellent JSON support to produce a sensibly-structured return value and avoid column name clashes:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>type bookAuthorSQL = books.SQL | authors.SQL | \"author\";\ntype bookAuthorSelectable = books.Selectable &amp; { author: authors.Selectable };\nconst query = sql&lt;bookAuthorSQL&gt;`\n  SELECT ${\"books\"}.*, to_jsonb(${\"authors\"}.*) as ${\"author\"}\n  FROM ${\"books\"} JOIN ${\"authors\"} \n    ON ${\"books\"}.${\"authorId\"} = ${\"authors\"}.${\"id\"}`;\nconst bookAuthors: bookAuthorSelectable[] = await query.run(pool);<\/pre><\/div>\n<p>Producing:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>&gt; console.log(bookAuthors[0]);\n{\n  \"id\": 14,\n  \"authorId\": 123,\n  \"title\": \"One Hundred Years of Solitude\",\n  \"createdAt\": \"2019-06-11T14:52:30.291Z\",\n  \"updatedAt\": \"2019-06-11T14:52:30.291Z\",\n  \"author\": {\n    \"id\": 123,\n    \"name\": \"Gabriel Garcia Marquez\",\n    \"isLiving\": false\n  }\n}<\/pre><\/div>\n<p>In which everything is appropriately typed:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/jawj\/mostly-ormless\/blob\/master\/README-resources\/selectable-auto-complete.png\"><img src=\"https:\/\/github.com\/jawj\/mostly-ormless\/raw\/master\/README-resources\/selectable-auto-complete.png\" alt=\"Screenshot: column name auto-completion\" \/><\/a><\/p>\n<p>Of course, we might also want the converse query, retrieving each author with their (many) books. This is also easy enough to arrange:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>type authorBooksSQL = authors.SQL | books.SQL;\ntype authorBooksSelectable = authors.Selectable &amp; { books: books.Selectable[] };\nconst\n  query = sql&lt;authorBooksSQL&gt;`\n    SELECT ${\"authors\"}.*, jsonb_agg(${\"books\"}.*) AS ${\"books\"}\n    FROM ${\"authors\"} JOIN ${\"books\"} \n      ON ${\"authors\"}.${\"id\"} = ${\"books\"}.${\"authorId\"}\n    GROUP BY ${\"authors\"}.${\"id\"}`,\n  authorBooks: authorBooksSelectable[] = await query.run(pool);<\/pre><\/div>\n<p>Which gives:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>&gt; console.dir(authorBooks[0]);\n[ { id: 123,\n    name: 'Gabriel Garcia Marquez',\n    isLiving: false,\n    books:\n     [ { id: 139,\n         title: 'Love in the Time of Cholera',\n         authorId: 123,\n         createdAt: '2019-09-22T19:49:32.373132+01:00',\n         updatedAt: '2019-09-22T19:49:32.373132+01:00' },\n       { id: 140,\n         title: 'One Hundred Years of Solitude',\n         authorId: 123,\n         createdAt: '2019-09-22T19:49:32.380854+01:00',\n         updatedAt: '2019-09-22T19:49:32.380854+01:00' } ] } ]<\/pre><\/div>\n<p>(It also demonstrates a useful SQL fact: selecting all fields in a <code>GROUP BY<\/code> query is <a href=\"https:\/\/dba.stackexchange.com\/questions\/158015\/why-can-i-sel\" rel=\"nofollow\">permitted as long as you're grouping by primary key<\/a>).<\/p>\n<p>Note that if you want to include authors with no books, you need a <code>LEFT JOIN<\/code> in this query, and then you'll want to fix the annoying <a href=\"https:\/\/stackoverflow.com\/questions\/24155190\/postgresql-left-join-json-agg-ignore-remove-null\" rel=\"nofollow\"><code>[null]<\/code> array results <code>jsonb_agg<\/code> will return for those authors<\/a>.<\/p>\n<p>One way to do that is to define a simple function to replace <code>[null]<\/code> with the empty array <code>[]<\/code>:<\/p>\n<div class=\"highlight highlight-source-sql\"><pre>CREATE OR REPLACE FUNCTION empty_nulls(jsonb) RETURNS jsonb AS $$\n  SELECT CASE $1 WHEN '[null]'::jsonb THEN '[]'::jsonb ELSE $1 END\n$$ LANGUAGE SQL IMMUTABLE;<\/pre><\/div>\n<p>And wrap it around the <code>jsonb_agg<\/code> call:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const query = sql&lt;authorBooksSQL&gt;`\n  SELECT ${\"authors\"}.*, empty_nulls(jsonb_agg(${\"books\"}.*)) AS ${\"books\"}\n  FROM ${\"authors\"} LEFT JOIN ${\"books\"} \n    ON ${\"authors\"}.${\"id\"} = ${\"books\"}.${\"authorId\"}\n  GROUP BY ${\"authors\"}.${\"id\"}`;<\/pre><\/div>\n<p><strong>Field subsets<\/strong><\/p>\n<p>Unless you have very wide tables and\/or very large values, it might be a <a href=\"https:\/\/softwareengineering.stackexchange.com\/questions\/80084\/is-premature-optimization-really-the-root-of-all-evil\" rel=\"nofollow\">premature optimization<\/a> to query only for a subset of fields. If you do need to limit your queries to particular fields, you might find TypeScript's <a href=\"https:\/\/www.typescriptlang.org\/docs\/handbook\/advanced-types.html\" rel=\"nofollow\">mapped and conditional types<\/a> useful.<\/p>\n<p>In this example we create a custom type for the selected data, and also for the SQL query. That latter type (<code>bookDatumSQL<\/code>) is kind of fiddly, and quite frankly it's probably better to stick to <code>books.SQL<\/code>, since you might anyway want to use other column names from the table \u2014\u00a0in a <code>WHERE<\/code> clause, for example.<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>type bookDatum = Pick&lt;books.Selectable, 'id' | 'title'&gt;;\ntype bookDatumSQL = Exclude&lt;books.SQL, Exclude&lt;keyof books.Selectable, keyof bookDatum&gt;&gt;;  \/\/ but maybe don't bother with this one\nconst\n  query = sql&lt;bookDatumSQL&gt;`SELECT ${\"id\"}, ${\"title\"} FROM ${\"books\"}`,\n  bookData: bookDatum[] = await query.run(pool);<\/pre><\/div>\n<p>Giving:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>&gt; console.log(bookData);\n[ { id: 153, title: 'Pride and Prejudice' },\n  { id: 154, title: 'Love in the Time of Cholera' },\n  { id: 155, title: 'One Hundred Years of Solitude' } ]<\/pre><\/div>\n<p>And that marks the end of Act 2. SQL queries are now being auto-completed and type-checked for me, which is excellent. But a lot of the simple stuff still feels a bit boiler-platey and verbose.<\/p>\n<h2><a id=\"user-content-act-3-in-which-we-create-a-set-of-simple-shortcut-functions-on-top-which-looks-a-little-bit-like-an-orm-but-really-isnt\" class=\"anchor\" aria-hidden=\"true\" href=\"#act-3-in-which-we-create-a-set-of-simple-shortcut-functions-on-top-which-looks-a-little-bit-like-an-orm-but-really-isnt\"><\/a>Act 3: In which we create a set of simple shortcut functions on top, which looks a little bit like an ORM but really isn't<\/h2>\n<p>In the final Act, I aim to make the simple one-table stuff fall-off-a-log easy with some straightforward helper functions. For these helper functions I generate a set of overloaded signatures per table, which means the return type and all other argument types can be inferred automatically from the table name argument.<\/p>\n<p><strong><code>SELECT<\/code><\/strong><\/p>\n<p>First, <code>SELECT<\/code>s. I define a <code>select<\/code> function, and my schemats fork generates appropriate signatures for it for every table in the database.<\/p>\n<p>Using that function turns this query:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const \n  query = sql&lt;books.SQL&gt;`SELECT * FROM ${\"books\"} WHERE ${{ authorId }}`,\n  existingBooks: books.Selectable[] = await query.run(pool);<\/pre><\/div>\n<p>Into this:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const existingBooks = await select(pool, \"books\", { authorId });<\/pre><\/div>\n<p>What's really nice here is that, thanks to the schemats-generated function signatures, once I've typed in <code>\"books\"<\/code> as the second argument to the function, TypeScript and VS Code know how to type-check and auto-complete both the third argument (which, if present, must be a <code>books.Whereable<\/code>) and the return value (which must be a <code>Promise&lt;books.Selectable[]&gt;<\/code>).<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/jawj\/mostly-ormless\/blob\/master\/README-resources\/return-type.png\"><img src=\"https:\/\/github.com\/jawj\/mostly-ormless\/raw\/master\/README-resources\/return-type.png\" alt=\"Screenshot: inferred return type\" \/><\/a><\/p>\n<p>The generated function signatures that make this happen look approximately like so:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>interface SelectSignatures {\n  (client: Queryable, table: authors.Table, where?: authors.Whereable, options?: authors.SelectOptions): \n    Promise&lt;authors.Selectable[]&gt;;\n  (client: Queryable, table: books.Table, where?: books.Whereable, options?: books.SelectOptions): \n    Promise&lt;books.Selectable[]&gt;;\n}<\/pre><\/div>\n<p>The <code>options<\/code> keys, as you might have spotted, are <code>order<\/code>, <code>limit<\/code> and <code>offset<\/code>, so I can do this kind of thing:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const [lastButOneBook] = await select(pool, \"books\", { authorId }, \n  { order: [{ by: \"createdAt\", direction: \"DESC\" }], limit: 1, offset: 1 });<\/pre><\/div>\n<p>The <code>by<\/code> option is being type-checked and auto-completed, of course.<\/p>\n<p>I used destructuring assignment here (<code>[lastButOneBook] = \/* ... *\/<\/code>) to account for the fact that I know this query is only going to return one response. Unfortunately, destructuring is just syntactic sugar for indexing, and indexing in TypeScript <a href=\"https:\/\/github.com\/Microsoft\/TypeScript\/issues\/13778\">doesn't reflect that the result may be undefined<\/a>. That means that <code>lastButOneBook<\/code> is now typed as a <code>books.Selectable<\/code>, but it could actually be <code>undefined<\/code>, and that could lead to errors down the line.<\/p>\n<p>To work around this, I also define a <code>selectOne<\/code> helper function. This turns the query above into the following:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const lastButOneBook = await selectOne(pool, \"books\", { authorId }, {\n  order: [{ by: \"createdAt\", direction: 'DESC' }], offset: 1\n});<\/pre><\/div>\n<p>The <code>limit<\/code> option is applied automatically in this case, and the return type following <code>await<\/code> is now, correctly, <code>books.Selectable | undefined<\/code>.<\/p>\n<p>\"Oh, great\", I hear you say, unenthusiastically. \"We've come all this way, and you've finished up reinventing the ORMs and query builders you promised we were going to escape\".<\/p>\n<p>Well, I don't think so. This is not an ORM. There's no weird magic going on behind the scenes \u2014 not even the merest hint that any semantics are being obscured. There's just a mechanical transformation of some function parameters into a single SQL query, which makes my life easier. One call to <code>select<\/code> means one unsurprising <code>SELECT<\/code> query is issued, and happily both the inputs and outputs of that query are now fully type-checked.<\/p>\n<p>It's also not a query builder.\u00a0There's no interminable method chaining. There's just type-checking and some additional conveniences applied to writing raw SQL queries, and some shortcut functions that can create basic SQL queries on your behalf.<\/p>\n<p>With that out of the way, onwards and upwards: <code>INSERT<\/code>s, <code>UPDATE<\/code>s, and (because, hey, now we can do anything Postgres can do) <a href=\"https:\/\/wiki.postgresql.org\/wiki\/UPSERT\" rel=\"nofollow\">UPSERT<\/a>s!<\/p>\n<p><strong><code>INSERT<\/code><\/strong><\/p>\n<p>The <code>insert<\/code> helper has the same sort of table-specific signatures as select, and it turns the <code>INSERT<\/code> query we saw earlier into this:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const savedBook = await insert(pool, \"books\", { \n  authorId: 123,\n  title: \"One Hundred Years of Solitude\",\n});<\/pre><\/div>\n<p>This produces the same query we wrote previously, but now with the addition of a <code>RETURNING *<\/code> clause at the end, meaning that we get back a <code>books.Selectable<\/code> that includes values for all the columns with defaults, such as the <code>id<\/code> serial.<\/p>\n<p>In addition, we can insert multiple rows in one query: the function is written (and has appropriate type signatures) such that, if I instead give it an array-valued <code>books.Insertable[]<\/code>, it gives me an array-valued <code>books.Selectable[]<\/code> back:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const savedBooks = await insert(pool, \"books\", [{ \n  authorId: 123,\n  title: \"One Hundred Years of Solitude\",\n}, {\n  authorId: 456,\n  title: \"Cheerio, and Thanks for All the Fish\",\n}]);<\/pre><\/div>\n<p><strong><code>UPDATE<\/code><\/strong><\/p>\n<p>Updates are likewise straightforward, fully typed, and also return the affected rows as a <code>Selectable[]<\/code>:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const \n  fishBookId = savedBooks[1].id,\n  properTitle = \"So Long, and Thanks for All the Fish\",\n  [updatedBook] = await update(pool, \"books\", \n    { title: properTitle },\n    { id: fishBookId }\n  );<\/pre><\/div>\n<p>We can again use a <code>SQLFragment<\/code> in an <code>update<\/code> for further flexibility \u2014 for example, if we needed to atomically increment a count.<\/p>\n<p>In a different context \u2014  for a schema that's not shown \u2014 I do something like the following:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>await update(pool, \"emailAuthentication\", { \n    consecutiveFailedLogins: sql`${self} + 1`,\n    lastFailedLogin: sql`now()`,\n  },\n  { email }\n);<\/pre><\/div>\n<p><strong>'UPSERT'<\/strong><\/p>\n<p>This being the thinnest possible wrapper around Postgres, we can of course do native 'UPSERT's too:<\/p>\n<p>I'll illustrate this one with a simplified example culled from our actual codebase, which deals with In-App Purchase updates for our iOS apps.<\/p>\n<div class=\"highlight highlight-source-sql\"><pre>CREATE TABLE \"appleTransactions\" \n( \"environment\" \"appleEnvironment\" NOT NULL  -- ENUM of 'PROD' or 'Sandbox'\n, \"originalTransactionId\" TEXT NOT NULL\n, \"accountId\" INTEGER REFERENCES \"accounts\"(\"id\") NOT NULL\n, \"latestReceiptData\" TEXT\n-- ... lots more fields ...\n);\nALTER TABLE \"appleTransactions\" \n  ADD CONSTRAINT \"appleTransPKey\" \n  PRIMARY KEY (\"environment\", \"originalTransactionId\");<\/pre><\/div>\n<p>When we receive a purchase receipt, either via our iOS app or direct from Apple's servers, we need to either store a new record or update an existing record for each 'original transaction ID' it contains.<\/p>\n<p>We therefore <code>map<\/code> the transaction data in the receipt into an <code>appleTransactions.Insertable[]<\/code>, and do what's needed with a single <code>upsert<\/code> call. Hard-coding the <code>Insertable[]<\/code> for ease of exposition:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const \n  newTransactions: s.appleTransactions.Insertable[] = [{\n    environment: 'PROD',\n    originalTransactionId: '123456',\n    accountId: 123,\n    latestReceiptData: \"TWFuIGlzIGRpc3Rp\",\n  }, {\n    environment: 'PROD',\n    originalTransactionId: '234567',\n    accountId: 234,\n    latestReceiptData: \"bmd1aXNoZWQsIG5v\",\n  }],\n  result = await upsert(pool, \"appleTransactions\", newTransactions, \n    [\"environment\", \"originalTransactionId\"]);<\/pre><\/div>\n<p>The last argument to <code>upsert<\/code> is the key or array of keys on which there could be a conflict.<\/p>\n<p>In this case, the following query is issued:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>{ text: `INSERT INTO \"appleTransactions\"\n    (\"accountId\", \"environment\", \"latestReceiptData\", \"originalTransactionId\") \n    VALUES ($1, $2, $3, $4), ($5, $6, $7, $8) \n    ON CONFLICT (\"environment\", \"originalTransactionId\") DO UPDATE \n    SET (\"accountId\", \"latestReceiptData\") = \n      ROW(EXCLUDED.\"accountId\", EXCLUDED.\"latestReceiptData\") \n    RETURNING *, \n      CASE xmax WHEN 0 THEN 'INSERT' ELSE 'UPDATE' END AS \"$action\"`,\n  values: [ \n    123, 'PROD', 'TWFuIGlzIGRpc3Rp', '123456',\n    234, 'PROD', 'bmd1aXNoZWQsIG5v', '234567' \n  ] }<\/pre><\/div>\n<p>The (awaited) return value is an <code>appleTransactions.UpsertReturnable[]<\/code>. An <code>UpsertReturnable<\/code> is just like a <code>Selectable<\/code> but with one extra property, <code>$action: \"INSERT\" | \"UPDATE\"<\/code>, valued according to what in fact happened. This is modelled on <a href=\"https:\/\/www.mssqltips.com\/sqlservertip\/1704\/using-merge-in-sql-server-to-insert-update-and-delete-at-the-same-time\/\" rel=\"nofollow\">a similar approach in MS SQL Server<\/a>.<\/p>\n<p>The <code>result<\/code> of the above query might therefore be:<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>&gt; console.log(result);\n[{\n  \"environment\": \"PROD\",\n  \"originalTransactionId\": \"123456\",\n  \"accountId\": 123,\n  \"latestReceiptData\": \"TWFuIGlzIGRpc3Rp\",\n  \"$action\": \"UPDATE\"\n}, {\n  \"environment\": \"PROD\",\n  \"originalTransactionId\": \"234567\",\n  \"accountId\": 234,\n  \"latestReceiptData\": \"bmd1aXNoZWQsIG5v\",\n  \"$action\": \"INSERT\"\n}]<\/pre><\/div>\n<h2><a id=\"user-content-encore-transactions\" class=\"anchor\" aria-hidden=\"true\" href=\"#encore-transactions\"><\/a>Encore: Transactions<\/h2>\n<p>The 'transactions' we just saw were data about people giving us money, but \u2014 for added confusion \u2014 let's talk for a moment about <em>database<\/em> transactions (as in: <code>BEGIN TRANSACTION<\/code> and <code>COMMIT TRANSACTION<\/code>).<\/p>\n<p>Transactions are an area where I've found TypeORM especially clumsy, and both Sequelize and TypeORM very footgun-prone. Both these ORMs encourage you not to think very often about which DB client will be running your query, and also make it very easy to issue some commands unintentionally outside of an open transaction (in Sequelize, you have to add a <code>transaction<\/code> option on the end of every query; in TypeORM, you have to remember to only use some decorator-injected transaction-aware contraption or other).<\/p>\n<p>That's why everything I've shown here requires a client argument to be specified explicitly, and all the above helper functions take that client as their first argument.<\/p>\n<p>To make life easier, I've written a <code>transaction<\/code> helper function that handles issuing a <code>ROLLBACK<\/code> on error, releasing the database client in a <code>finally<\/code> clause (i.e. whether or not an error was thrown), and automatically retrying queries in case of serialization failures.<\/p>\n<p>The helper simply takes the isolation level you're looking for and calls the function you provide, handing over an appropriate database client.<\/p>\n<div class=\"highlight highlight-source-ts\"><pre>const result = transaction(Isolation.Serializable, async txnClient =&gt; {\n  \/* transaction queries go here, using txnClient instead of pool *\/\n});<\/pre><\/div>\n<p>In addition, it provides a set of hierarchical isolation types, so that you can declare a function as requiring a particular isolation level <em>or above<\/em>. Type-checking will then allow you to pass a client associated with any appropriate level. For example, if you type the <code>txnClient<\/code> argument to your function as <code>TxnSatisfying.RepeatableRead<\/code>, you can call it with <code>Isolation.Serializable<\/code> or <code>Isolation.RepeatableRead<\/code>, but not <code>Isolation.ReadCommitted<\/code>.<\/p>\n<h2><a id=\"user-content-try-it-out\" class=\"anchor\" aria-hidden=\"true\" href=\"#try-it-out\"><\/a>Try it out<\/h2>\n<p>Everything you need to give this approach a spin is included in this repo. Make sure you have node, npm and Postgres installed, then:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>cd \/some\/appropriate\/path\ngit clone https:\/\/github.com\/jawj\/mostly-ormless.git\ncd mostly-ormless\nnpm install\ncreatedb mostly_ormless\npsql -d mostly_ormless &lt; src\/schema.sql<\/pre><\/div>\n<p>Then:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>npx ts-node src\/demo.ts  # runs all the examples in this README<\/pre><\/div>\n<p>Once you've got <code>demo.ts<\/code> running successfully, you can play around with your own queries in there.<\/p>\n<p>You can also modify the database schema, of course. For this, you'll need my schemats fork, so in the root of this project run:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>npm submodule init\nnpm submodule update\ncd schemats\nnpm install\ncd ..<\/pre><\/div>\n<p>Then, whenever you change the schema in Postgres, regenerate <code>src\/schema.ts<\/code> by running (as one long line):<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>npx ts-node schemats\/bin\/schemats.ts generate -c postgres:\/\/localhost\/mostly_ormless -o src\/schema.ts <\/pre><\/div>\n<h2><a id=\"user-content-where-next\" class=\"anchor\" aria-hidden=\"true\" href=\"#where-next\"><\/a>Where next?<\/h2>\n<p>If you think this approach could be useful in your own projects, feel free to adopt\/adapt it. I'd be interested to hear how you get on. I could of course create an npm library for all this, but I think it's much more truly ORMless if you just take the code (<code>core.ts<\/code> is less than 400 lines), understand it, and make it your own.<\/p>","created_at":"2019-09-23T16:52:47+0000","updated_at":"2019-12-07T17:10:18+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":23,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/141620?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1030,"title":"Tweet from \n\n    \n      adamgryu","url":"https:\/\/mobile.twitter.com\/adamgryu\/status\/1156201582641537025?s=09","content":"<div class=\"dir-ltr\" dir=\"ltr\">  \ud83d\udc26\n\ud83e\udd7e\ud83e\udd7e\n\n\ud83c\udf32\ud83c\udf32\ud83c\udf32 A Short Hike \ud83c\udf32\ud83c\udf32\ud83c\udf32\nis OUT NOW on PC, Mac, and Linux!\ud83c\udfde\ufe0f\n\nExplore the peaceful and quiet island of Hawk Peak Provincial Park as you search for... cellphone reception\ud83d\udcf1\n\n\u27a1\ufe0fSteam: <a href=\"https:\/\/t.co\/77nCsfbbQW\" rel=\"nofollow noopener\" dir=\"ltr\" data-expanded-url=\"https:\/\/store.steampowered.com\/app\/1055540\/A_Short_Hike\/\" data-url=\"https:\/\/store.steampowered.com\/app\/1055540\/A_Short_Hike\/\" class=\"twitter_external_link dir-ltr tco-link\" target=\"_blank\" title=\"https:\/\/store.steampowered.com\/app\/1055540\/A_Short_Hike\/\">store.steampowered.com\/app\/1055540\/A_\u2026<\/a>\n\u27a1\ufe0f Itch: <a href=\"https:\/\/t.co\/v57thXdShD\" rel=\"nofollow noopener\" dir=\"ltr\" data-expanded-url=\"https:\/\/adamgryu.itch.io\/a-short-hike\" data-url=\"https:\/\/adamgryu.itch.io\/a-short-hike\" class=\"twitter_external_link dir-ltr tco-link\" target=\"_blank\" title=\"https:\/\/adamgryu.itch.io\/a-short-hike\">adamgryu.itch.io\/a-short-hike<\/a> <a href=\"https:\/\/t.co\/FGhxohTqo8\" data-pre-embedded=\"true\" rel=\"nofollow\" data-entity-id=\"1156200349730050048\" dir=\"ltr\" data-url=\"https:\/\/twitter.com\/adamgryu\/status\/1156201582641537025\/video\/1\" data-tco-id=\"FGhxohTqo8\" class=\"twitter_external_link dir-ltr tco-link has-expanded-path\" target=\"_top\" data-expanded-path=\"\/adamgryu\/status\/1156201582641537025\/video\/1\">pic.twitter.com\/FGhxohTqo8<\/a>\n<\/div>","created_at":"2019-09-23T02:16:06+0000","updated_at":"2019-09-23T02:16:06+0000","published_by":["adamgryu"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"mobile.twitter.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1029,"title":"Using Gooey as a Universal frontend for any Language or CLI Application","url":"https:\/\/chriskiehl.com\/article\/gooey-as-a-universal-frontend","content":"<p><img alt=\"image\" src=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-1000px.jpeg\" srcset=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-1000px.jpeg 1000w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-833px.jpeg 833w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-666px.jpeg 666w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-499px.jpeg 499w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-332px.jpeg 332w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/40c812f6-image-165px.jpeg 165w\" \/><\/p><p><sup>Published: 2019-09-22<\/sup><\/p><p><a href=\"https:\/\/github.com\/chriskiehl\/Gooey\">Gooey<\/a> is a tool for transforming command line interfaces into beautiful desktop applications. It can be used as the frontend client for <i>any<\/i> language or program. Whether you've built your application in Java, Node, or Haskell, or you just want to put a pretty interface on an existing tool like FFMPEG, Gooey can be used to create a fast, practically free UI with just a little bit of Python (about 20 lines!).\u00a0<\/p><blockquote><p> Don't want to write no stinkin' Python? Good news! Gooey is actually controlled entirely by a plain ol' JSON file. We're only using Python here because it has convenient bindings for generating the JSON. <a href=\"https:\/\/github.com\/chriskiehl\/Gooey\">Contribute bindings for your language<\/a>! <\/p><\/blockquote><p>To show how this all fits together, and that it really works for <i>anything<\/i>, we're going to walk through building a graphical interface to one of my favorite tools of all time: <a href=\"https:\/\/ffmpeg.org\/\">FFMPEG<\/a>. These steps apply to anything, though! You could swap out FFMPEG for a <code>.jar<\/code> you've written, or an arbitrary windows <code>.exe<\/code>, an OSX <code>.app<\/code> bundle, or anything on linux that's executable!\u00a0<\/p><p>Why wrap up another program at all? While FFMPEG is amazing, it is anything but friendly. Borderline hostile, in fact. The learning curve is basically just a brick wall. To make it friendly for our theoretical end user, we're going to expose a subset of it as a simple, familiar GUI in which they can point-and-click their way to success. Additionally, since FFMPEG has a staggeringly deep suite of video processing tools, to keep things simple we're going to expose just a very small subset of its functionality. We're going to build a UI in which users can specify 1. an input video file, 2. a timestamp where they'd like to extract a screenshot, and finally 3. the output filename where they want the screenshot saved.<\/p><p>In short, it will transform a \"scary\" terminal command line this:<\/p><pre>ffmpeg -ss 300 -i \/home\/user\/path\/to\/movie.mkv -frames:v 1 outfile.png\n<\/pre><p>into an easy to use desktop application that you could hand over to users.<\/p><p><img alt=\"image\" src=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-647px.jpeg\" srcset=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-647px.jpeg 647w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-539px.jpeg 539w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-431px.jpeg 431w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-323px.jpeg 323w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-215px.jpeg 215w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/2c86b50c-image-107px.jpeg 107w\" \/><\/p><p><strong>Getting Started<\/strong><\/p><p>To play along at home, grab the appropriate FFMPEG binary for your OS <a href=\"https:\/\/johnvansickle.com\/ffmpeg\/\">here<\/a> and install it.<\/p><p>If all has gone well, you should be able to invoke it from the command line and get back some useful version info<\/p><pre>ffmpeg -version\n<\/pre><p>With that working, let's get onto building the UI side of things.<\/p><p><strong>Step 1. Installing Python<\/strong><\/p><p>If you're coming from other languages \/ backgrounds, step one will be installing Python. You can find an appropriate guide for your OS <a href=\"https:\/\/wiki.python.org\/moin\/BeginnersGuide\/Download\">here<\/a>. If you're on Windows, the shortest path to done boils down to visiting <a href=\"https:\/\/www.python.org\/\">python.org<\/a>, clicking the big \"Download\" button, and then double clicking the downloaded file.<\/p><p><strong>Step 2. Create a directory and source file for the project<\/strong><\/p><p>We'll need a single python file to describe the GUI we want to build. To stay organized, we'll put any files we create in a directory called <code>myproject<\/code>.<\/p><pre>mkdir myproject\u00a0\ncd myproject\n<\/pre><p>Inside here we'll create the source file we need.<\/p><pre>touch main.py\n<\/pre><p><strong>Step 3. Installing Gooey<\/strong><\/p><p>Before we start writing the little bits of code required, we have to install Gooey. In Python land, this is done via a tool called <code>pip<\/code>. This should have been installed automatically along side Python in step 1.<\/p><pre>pip install Gooey\n<\/pre><blockquote><p> Note: you would normally install Python dependencies inside of a <a href=\"https:\/\/virtualenv.pypa.io\/en\/stable\/\">virtual environment<\/a>, but we're ignoring that detail for ease of example. <\/p><\/blockquote><p><strong>Step 4. Let's make a UI!<\/strong><\/p><p>Open the <code>main.py<\/code> file we created in step 2 and paste in the following.<\/p><pre class=\"python\">from gooey import Gooey, GooeyParser\n@Gooey(target=\"ffmpeg\", program_name='Frame Extraction v1.0', suppress_gooey_flag=True)\u00a0\ndef main():\n\u00a0\u00a0\u00a0 parser = GooeyParser(description=\"Extracting frames from a movie using FFMPEG\")\n\u00a0\u00a0\u00a0 ffmpeg = parser.add_argument_group('Frame Extraction Util')\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('-i',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Input Movie',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='The movie for which you want to extract frames',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 widget='FileChooser')\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('output',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Output Image',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='Where to save the extracted frame',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 widget='FileSaver',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 )\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('-ss',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Timestamp',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='Timestamp of snapshot (in seconds)')\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('-frames:v',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Timestamp',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 default=1,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 gooey_options={'visible': False})\n\u00a0\u00a0\u00a0 parser.parse_args()\nif __name__ == '__main__':\n\u00a0\u00a0\u00a0 main()\n<\/pre><blockquote><p>note: if you're familiar with Python, this code will look suspiciously similar to the <a href=\"https:\/\/docs.python.org\/3\/library\/argparse.html\">Argparse<\/a> library, and that's because it <i>is<\/i> Argparse code! Gooey is a drop-in replacement for all projects backed by Argparse! <\/p><\/blockquote><p>That's all it takes to get a UI bolted on in front of an existing CLI application! Before we walk through the details of the code, let's give it a run and enjoy the spoils of our work!<\/p><pre>python main.py\n<\/pre><p><img alt=\"image\" src=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-736px.jpeg\" srcset=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-736px.jpeg 736w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-613px.jpeg 613w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-490px.jpeg 490w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-367px.jpeg 367w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-244px.jpeg 244w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/0143068f-image-121px.jpeg 121w\" \/><\/p><p>Pretty awesome!<\/p><p><strong>How it works:<\/strong><\/p><p>A fair bit of the code is just boilerplate: function definitions, imports, bookkeeping. You can learn about them in detail via the <a href=\"https:\/\/github.com\/chriskiehl\/Gooey\">docs<\/a>. We'll focus just on the interesting bits. Going from top to bottom, the first notable line in our code is the one that begins with <code>@Gooey<\/code><\/p><pre>@Gooey(target=\"ffmpeg\", program_name='Frame Extraction v1.0', suppress_gooey_flag=True)\n<\/pre><p><code>target<\/code> is the most important argument on this line. This is how we tell Gooey that we're trying to run an <code>ffmpeg<\/code> command. If you were using Gooey with, say, a <code>.jar<\/code>, you'd similarly provide that info here (e.g. <code>target='java -jar '). At run time, the input provided by the user in the UI gets stitched together and handed over to whatever is specified in <\/code>target`. This is what enables Gooey to work with any CLI app!\u00a0<\/p><p>The next notable lines are the ones beginning with <code>ffmpeg.add_argument<\/code>. These are how we specify what shows up as a form field in the UI. Each field we want to present to the user will have an associated call to <code>add_argument<\/code>. They all have a few arguments in common:<\/p><ul><li>first positional argument - this is where we specify the CLI variable for which we're receiving input. In our case, we've set it to <code>-i<\/code>, which is how you declare the input file to FFMPEG<\/li><li><code>metavar<\/code> - This is the human readable name we want to appear over the input field.<\/li><li><code>help<\/code> - This is the descriptive text that appears next to the input field<\/li><li><code>widget<\/code> - Gooey has a bunch of different input widgets. Since we're dealing with file IO, we set the relevant fields to <code>FileChooser<\/code>\/<code>FileSaver<\/code> which spawns the native file dialog used by your host OS.<\/li><\/ul><p>There are a few slightly advanced pieces of Gooey in here as well.<\/p><pre class=\"python\">\u00a0\u00a0\u00a0 ffmpeg.add_argument('-frames:v',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Timestamp',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 default=1,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 required=True,\u00a0 # Look at me!\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 gooey_options={'visible': False})\n<\/pre><p>This one uses the <code>gooey_options<\/code> API to hide this specific component from the interface. The <code>frames:v<\/code> argument is a minor detail of FFMPEG that end users need not worry about, so we set it up with the values ffmpeg needs, but don't actually expose it.<\/p><p>And that's all there is to it! In just a few lines of Python, we've build a super polished UI that's easy to use and turns the complex task of frame extraction via FFMPEG into something that can be used by anyone.\u00a0<\/p><p>We can take this a little further, though..!<\/p><p>Our UI is pretty good, but it has a few problems. For instance, users could click <code>start<\/code> without filling out all the fields, or they could fill out the fields with junk data! Let's be helpful citizens and give them some guidance.<\/p><p><strong>Make required fields required:<\/strong><\/p><p>We can add <code>required=True<\/code> to any fields we want to force the user to manage.<\/p><pre>@Gooey(target=\"ffmpeg\", program_name='Frame Extraction v1.0', suppress_gooey_flag=True)\ndef main():\n\u00a0\u00a0\u00a0 parser = GooeyParser(description=\"Extracting frames from a movie using FFMPEG\")\n\u00a0\u00a0\u00a0 ffmpeg = parser.add_argument_group('Frame Extraction Util')\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('-i',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Input Movie',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='The movie for which you want to extract frames',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 required=True,\u00a0 # Look at me!\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 widget='FileChooser')\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('output',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Output Image',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='Where to save the extracted frame',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 widget='FileSaver',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 )\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('-ss',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Timestamp',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 required=True,\u00a0 # Look at me!\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='Timestamp of snapshot (in seconds)')\n\u00a0\u00a0\u00a0 ffmpeg.add_argument('-frames:v',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Timestamp',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 default=1,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 required=True,\u00a0 # Look at me!\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 gooey_options={'visible': False})\n\u00a0\u00a0\u00a0 parser.parse_args()\n<\/pre><p>Now if we try to advance without filling out any of the forms we get a helpful notification!<\/p><p><img alt=\"image\" src=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-641px.jpeg\" srcset=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-641px.jpeg 641w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-534px.jpeg 534w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-427px.jpeg 427w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-320px.jpeg 320w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-213px.jpeg 213w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/092c4d1d-image-106px.jpeg 106w\" \/><\/p><p><strong>Input Validation:<\/strong><\/p><p>Now let's make sure that the user enters appropriate data types. Update the <code>-ss<\/code> argument to look like the following.<\/p><pre class=\"python\">\u00a0\u00a0\u00a0 ffmpeg.add_argument('-ss',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 metavar='Timestamp',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 required=True,\u00a0 # Look at me!\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 help='Timestamp of snapshot (in seconds)',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 gooey_options={\u00a0\u00a0\u00a0 # NEW!\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 'validator': {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 'test': 'user_input.isdigit()',\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 'message': 'Please enter a number'\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 })\n<\/pre><p>You can read all about validation in the docs, but in short, this tells Gooey to verify that the user's input is actually a number before passing things of to ffmpeg. With this in place, now we've gone really far towards making something that's truly user friendly!<\/p><p><img alt=\"image\" src=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-649px.jpeg\" srcset=\"https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-649px.jpeg 649w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-540px.jpeg 540w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-431px.jpeg 431w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-322px.jpeg 322w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-213px.jpeg 213w, https:\/\/s3.amazonaws.com\/awsblogstore\/articles\/5c0aa089\/75fa3e0e-image-104px.jpeg 104w\" \/><\/p><h3>Wrapping Up<\/h3><p>Where should we go from here? There are lots of options! You could <a href=\"https:\/\/github.com\/chriskiehl\/Gooey#packaging\">package your new Gooey powered app as a stand alone executable<\/a>, customize the <a href=\"https:\/\/github.com\/chriskiehl\/Gooey#customizing-icons\">look and feel<\/a> of the UI, <a href=\"https:\/\/github.com\/chriskiehl\/Gooey#internationalization\">internationalize<\/a> your text, and a whole lot more!<\/p>","created_at":"2019-09-23T01:35:20+0000","updated_at":"2019-09-23T01:35:20+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"chriskiehl.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["datalog","datomic","unifydb","datahike"],"is_public":false,"id":1028,"title":"Datahike Release 0.2.0","url":"https:\/\/lambdaforge.io\/2019\/09\/20\/datahike-release-0.2.0.html","content":"<p>Last week we released version <code class=\"highlighter-rouge\">0.2.0<\/code> of <strong>datahike<\/strong> which brings new features\nand marks a step away from its datascript origin. The major new features are\nschema flexibility and time travel. After integrating the\nlatest datascript code we extended it through protocols on storage and index level.<\/p><p>While <strong>datahike<\/strong> began as a fork of\n<a href=\"https:\/\/github.com\/tonsky\/datascript\">datascript<\/a> and integration of the\nhitchhiker tree, we are now adding novel features and moving away from the\noriginal project. For this release we integrated the latest datascript code and\nrefactored it in preparation for the next features.<\/p><p>Additionally protocols were created for both index and storage layer. <a href=\"https:\/\/github.com\/purrgrammer\">Alejandro\nG\u00f3mez<\/a> provided protocols for the storage\nbackends that support different solutions like <code class=\"highlighter-rouge\">in-memory<\/code>, <code class=\"highlighter-rouge\">file-based<\/code>,\n<code class=\"highlighter-rouge\">LevelDB<\/code>, or <code class=\"highlighter-rouge\">PostgreSQL<\/code>. New stores can be added by implementing four\nfunctions on top of <a href=\"https:\/\/github.com\/replikativ\/konserve\">konserve<\/a> bindings.\nA good example are the <a href=\"https:\/\/github.com\/replikativ\/datahike\/blob\/master\/src\/datahike\/store.cljc#L80\">PostgreSQL\ncalls<\/a>.<\/p><p>The underlying index is now also configurable supporting both the persistent\nsorted set used in datascript and the <a href=\"https:\/\/blog.datopia.io\/2018\/11\/03\/hitchhiker-tree\/\">hitchhiker\ntree<\/a>. Through protocols\nother interesting index data structures can be added easily.<\/p><p>Since our partners from the <a href=\"https:\/\/datopia.io\/\">datopia project<\/a> needed the\ndatalog parsing functionality we moved all functions and tests around datalog\nquery parsing into a <a href=\"https:\/\/github.com\/lambdaforge\/datalog-parser\">separate\nproject<\/a> that will also support\nfunctions not used in <strong>datahike<\/strong>, like creating datalog queries from its syntax\ntree, so you can parse a datalog query, add addtionaly data, and create an\noptimized query for a database to use.<\/p><p>Up until now <strong>datahike<\/strong> did not enforce any schema on its data, so any badly shaped\ndata could be added. Starting with <strong>datahike<\/strong> <code class=\"highlighter-rouge\">0.2.0<\/code> the schema-on-write\napproach is supported where the schema has to be defined explicitly and\nthe transactor ensures all transacted data conforms to it. Through configuration\nthe previous schema-on-read is still supported.<\/p><p><strong>datahike<\/strong> supports large parts of <a href=\"https:\/\/docs.datomic.com\/on-prem\/schema.html\">Datomic\u2019s schema\ndefinition<\/a> except tuples, full\ntext search and byte types. The data validation happens on transaction level using <code class=\"highlighter-rouge\">clojure.spec<\/code>.<\/p><p>In contrast to datascript\u2019s schema in the database record, <strong>datahike<\/strong> stores the\nschema additionally as transactions in\nthe index, which allows queries and time travel\nthat can help auditing schema changes.<\/p><p>Modern database solutions require auditable data history without integrating time stamp\nattributes to all data models. Similar to <a href=\"https:\/\/docs.datomic.com\/on-prem\/getting-started\/see-historic-data.html\">Datomic\u2019s historic data<\/a>\n<strong>datahike<\/strong> supports time travel capabilities that allow datalog queries against\nthe whole history, against any point in the past, or against the\ndifference since a certain point in time.<\/p><p>In accordance to the <a href=\"https:\/\/gdpr-info.eu\/\">GDPR regulations<\/a> data\npurging capabilities were added that completely remove data either from\nall indices or only from the historical indices.<\/p><p>In order to keep the current data view fast and clean, a separate set of\nindices were created that track all past data. This separation makes it very easy\nto clean up data after certain retention periods without interfering with the\ncurrent view of data.<\/p><p>With the addition of the latest features, <strong>datahike<\/strong> required better configuration.\nTherefore a clean concept was implemented to allow picking only those\nfunctions from datahike that you need the most at database creation. For example,\nboth time travel and schema-on-write capabilities can be deactivated if you have\nno use for it.<\/p><p>Since it is always helpful to have examples, we provide <a href=\"https:\/\/github.com\/replikativ\/datahike\/tree\/master\/examples\">projects<\/a> inside\nthe code base that show different features as well as documentation about\nnew features like <a href=\"https:\/\/github.com\/replikativ\/datahike\/blob\/master\/doc\/time_variance.md\">time\ntravel<\/a>,\n<a href=\"https:\/\/github.com\/replikativ\/datahike\/blob\/master\/doc\/schema.md\">schema<\/a>, and\n<a href=\"https:\/\/github.com\/replikativ\/datahike\/blob\/master\/doc\/config.md\">configuration<\/a>.<\/p><p>These projects just need a REPL and can be worked through at your own pace.\nWhere it was possible we added explanations and different ways to use a certain\nfeature. Have a look at the <a href=\"https:\/\/github.com\/replikativ\/datahike\/tree\/master\/examples\/basic\">basic\nexample<\/a>\nwhich introduces topics like stores, schema or time travel.\nOver the next weeks we will add more examples about queries and transactions.<\/p><p>For the next release we are planning to make datalog available to a wider audience by\nreleasing API bindings for Java\/Scala of <strong>datahike<\/strong>.<\/p><p>In collaboration with our partners from the <a href=\"https:\/\/github.com\/sixthnormal\/clj-3df\">3DF\nproject<\/a> we aim to support remote\ncapabilities with a HTTP layer, having only a thin client on the application\nside instead of the full <strong>datahike<\/strong> stack.<\/p><p>With support of a datalog query planner we will to be able to predict the\ncosts of each query and recommend possible solutions to improve your overall queries.<\/p><p>For security reasons it would be useful to have an <a href=\"https:\/\/en.wikipedia.org\/wiki\/Identity_management\">identity and access\nmanagement<\/a> concept ready to\nbe used. Therefore user and role definitions will be added to the system entities.<\/p><p>Since we have experience with probabilistic programming in\n<a href=\"http:\/\/anglican.ml\">anglican<\/a> we\nwill try to integrate probabilistic reasoning in datalog and <em>datahike<\/em> itself.<\/p><p>From a development side we will try to release improvements on a\nsmaller scale with more minor and patch releases and a more transparent\ndevelopment process around GitHub issues.<\/p><p>Many thanks to all the people from the <a href=\"https:\/\/jobtechdev.se\/\">Job Tech\nproject<\/a> supporting us towards the implementation of the\nschema and time travel functionalities. Also thanks to <a href=\"https:\/\/github.com\/grischoun\">Chrislain\nRazafimahefa<\/a> for his code reviews, <a href=\"https:\/\/github.com\/purrgrammer\">Alejandro\nG\u00f3mez<\/a> for his storage contributions, and all\nyou guys either on slack, GitHub or personally for fruitful discussions and\nfeedback.<\/p><p>Have fun with the latest version of <strong>datahike<\/strong> and let us know what you think\nabout it.<\/p>","created_at":"2019-09-21T18:34:51+0000","updated_at":"2019-09-21T18:34:58+0000","annotations":[],"mimetype":"text\/html","language":"de","reading_time":4,"domain_name":"lambdaforge.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["graphs","dgraph","auth","unifydb"],"is_public":false,"id":1027,"title":"How Access Control Works in Dgraph - Dgraph Blog","url":"https:\/\/blog.dgraph.io\/post\/access-control-in-dgraph\/","content":"<p>With <a href=\"https:\/\/en.wikipedia.org\/wiki\/General_Data_Protection_Regulation\">GDPR<\/a> passing in the EU on April 14th, 2016, users on the Internet are demanding\nprotection of their privacy and data from Internet companies. The regulation has profound\nimplications. Many companies take the law and users\u2019 demand seriously: they start encrypting data\nfor traffic <em>within<\/em> their data centers, and employing strict control on who can access users\u2019 data.\nThe Dgraph team takes full notice of that. Access control on the data stored in Dgraph is very\nimportant to our enterprise customers. <strong>This feature has been in-demand and we plan to release it\nin the upcoming Dgraph v1.1 as a proprietary feature.<\/strong> In this article, you will find some details\nabout how Access Control List (ACL) works in Dgraph.<\/p><h3 id=\"the-requirements\">The requirements<\/h3><p>We started the project by asking what are the features that we must support for\naccess control to work both securely and seamlessly for our customers. After a\nfew rounds of brainstorming and feedback, we settled on the following list:<\/p><ul><li>We should be able to create user accounts, e.g. Alice and Bob, and groups, e.g. devs and SREs.<\/li>\n<li>Each user can be assigned to multiple groups.<\/li>\n<li>Access control should be enforced on predicates, and each predicate should\nhave 3 types of access:\n<ol><li>Read: being able to query data under the predicate<\/li>\n<li>Write: being able to mutate data under the predicate<\/li>\n<li>Modify: being able to alter the schema of the predicate.<\/li>\n<\/ol><\/li>\n<li>Access control should be granted on the granularity of groups. For example, we can grant Read\naccess of the predicate <code>friend<\/code> to the group <code>dev<\/code>, and Read+Write access of the predicate <code>friend<\/code> to\nthe group <code>sre<\/code>.<\/li>\n<li>There should be a root user account that can do any type of access to any predicate in the db. The\nroot account is needed to bootstrap access control by creating other users\u2019 accounts and setting\nup the initial ACL rules. Also the root account is necessary for mediation if ACL rules are\nmisconfigured.<\/li>\n<li>Multiple predicates matching a regular expression, e.g. <code>^user(.\\*)name$<\/code>, can be granted to a group at once. When\nsuch a rule exists, predicates created later matching the regex should have access control\nenforced automatically.<\/li>\n<li>There should be a way to query:\n<ol><li>All the groups of a given user<\/li>\n<li>All the users belonging to a given group<\/li>\n<li>All the predicates and the granted permissions associated with a given group.<\/li>\n<\/ol><\/li>\n<li>There should be logging of all the access attempts, including who is trying to perform what\noperation on which predicate, and whether the access is accepted or denied.<\/li>\n<\/ul><h3 id=\"how-acl-data-is-stored\">How ACL data is stored<\/h3><p>With the requirements settled, now let\u2019s dive into some details. The very first detailed\ndesign question is how to store the ACL data. We discussed several options each\nhaving their pros and cons. In the end we decided to pick the following approach:\n<img src=\"https:\/\/blog.dgraph.io\/images\/groups.png\" alt=\"Predicates assigned to groups\" \/><\/p><p>To make it easier for operators to turn on the ACL feature, we are storing the\nACL rules inside Dgraph instead of relying upon an external system. The example\ndemonstrates a cluster with a Zero group and a number of Alpha groups. The\nZero group maintains metadata in the cluster, including which group a server\nbelongs to and which group a predicate is assigned to. The Alpha groups\nhost the actual user data, with each one hosting a portion of the whole data\nset. The whole dataset is divided into shards and in Dgraph sharding is based on\npredicates.<\/p><p>To store the ACL data, we need a total of four predicates. And these four predicates\nare treated specially by group zero in the sense that they are always assigned\nto group 1. Other groups maintain a local in-memory cache of all the ACL\nrules and refresh it periodically by querying group 1<sup class=\"footnote-ref\" id=\"fnref:1\"><a href=\"#fn:1\">1<\/a><\/sup>. Thus, every Alpha\nserver knows about the ACL rules, so each Alpha is capable to applying those\nrules to all the incoming requests, without having to make network calls.<\/p><p>Now let\u2019s take a closer look at the four predicates:<\/p><ul><li><code>dgraph.xid<\/code> is for storing the name of a user or group<\/li>\n<li><code>dgraph.password<\/code> is for storing a user\u2019s hashed password<\/li>\n<li><code>dgraph.user.group<\/code> is for storing user \u2192 group mapping and uses a reverse index to retrieve group \u2192 users mapping.<\/li>\n<li><code>dgraph.group.acl<\/code> is for storing all predicates assigned to a group (or regexps against which predicates will be matched), and the types of access allowed for\nthe group, e.g. <code>(friend, Read), (name, Read+Write), (^user.*name$, Read+Write+Modify)<\/code>.<\/li>\n<\/ul><figure><img src=\"https:\/\/blog.dgraph.io\/images\/groot.jpg\" alt=\"image\" \/><figcaption>I AM GROOT\n        <\/figcaption><\/figure><p>These predicates are proposed by an Alpha server when it starts up if the ACL feature is turned on\nthrough command line options and created when they do not already exist in the cluster.\nUnlike a regular data predicate, they are not removed by the DropAll request.<\/p><p>Also being created during an Alpha server start-up time is the root user account if it does not\nalready exist, and we give it the nickname <code>groot<\/code> (graph root) along with a default password which\ncan be changed using our command line tools explained below.<\/p><h3 id=\"acl-clients\">ACL clients<\/h3><p>Now that the server side storage is ready, let\u2019s switch our attention to the client side.\nFrom the ACL perspective, there are two types of clients:<\/p><ul><li>The command line tools for changing and querying ACL data<\/li>\n<li>The clients used to access data protected by ACL rules, e.g. the\n<a href=\"https:\/\/github.com\/dgraph-io\/dgo\">dgo<\/a> client written in Go, the <a href=\"https:\/\/github.com\/dgraph-io\/dgraph4j\">dgraph4j<\/a> client in Java, the\n<a href=\"https:\/\/play.dgraph.io\/\">Ratel<\/a> web portal etc.<\/li>\n<\/ul><h4 id=\"the-command-line-tools\">The command line tools<\/h4><p>The command line tools are available under a new <code>acl<\/code> subcommand within the\nDgraph binary.<\/p><ul><li><code>dgraph acl add<\/code> is for creating a new user account or a new group<\/li>\n<li><code>dgraph acl del<\/code> is for deleting a user account or a group<\/li>\n<li><code>dgraph acl mod<\/code> is for changing a user\u2019s password, or a group\u2019s permissions on predicates<\/li>\n<li><code>dgraph acl info<\/code> is for querying a user\u2019s groups, or a group\u2019s users, or predicate permissions granted\nto a group<\/li>\n<\/ul><p>Concrete examples for turning on the ACL feature in a cluster and setting up ACL rules can be found\nat <a href=\"https:\/\/docs.dgraph.io\/enterprise-features\/#access-control-lists\">our documentation page<\/a>.<\/p><h4 id=\"client-libraries\">Client Libraries<\/h4><p>The second type of clients, e.g. dgo or dgraph4j, accesses data in Dgraph by\nexecuting transactions.  Each transaction can have a mix of <code>query<\/code> and <code>mutate<\/code>\nrequests. Also a client can be used to <code>alter<\/code> the schema, e.g. changing a\npredicate type from an int to a string. When a request is received by an Alpha\nserver, it is evaluated using rules in the ACL cache and is only accepted if the\noperations for all predicates in the request are allowed. For example, for the\nquery below to work, the current user must belong to one or more groups that\ncollectively have been granted Read access to both the <code>name<\/code> and <code>friend<\/code>\npredicate.<\/p><pre>{\n  friends_of_alice(func: eq(name, \"Alice\")) {\n    name\n    friend {\n      name\n    }\n  }\n}\n<\/pre><h3 id=\"the-jwts\">The JWTs<\/h3><p><em>But wait a second!<\/em> How does an Alpha server know the\nquery\u2019s current user or its groups?  The trick is that each request carries some\nmetadata encoded as <a href=\"https:\/\/jwt.io\/\">JSON Web Token<\/a> that tells the Alpha server about the\ncurrent user and the groups they belong to.<\/p><figure><img src=\"https:\/\/blog.dgraph.io\/images\/accessJWT.png\" alt=\"image\" \/><figcaption>Access Token\n        <\/figcaption><\/figure><h4 id=\"access-token\">Access Token<\/h4><p><strong>Here is how JWT works:<\/strong> Before running any transactions, the client, e.g. dgo\nneeds to make a <code>Login<\/code> call to an Alpha server using a user name and password.\nWhen processing the <code>Login<\/code> request, the Alpha server 1) queries all of the\nuser\u2019s groups, 2) puts an expiration timestamp depending on a server-configured\nTTL, and 3) signs all that data with a secret key. The group list, expiration\ntimestamp, and signature together form an access JWT which is sent back to the\nclient. The client then remembers this JWT and attaches it to every subsequent\nrequest it sends to the cluster, e.g. the query request example we saw earlier.<\/p><h4 id=\"refresh-token\">Refresh Token<\/h4><p>As explained above, JWTs contain the groups a logged-in user belongs to. This\ncan cause a problem if this user is assigned a new group. Until the Access Token\nexpires, a new JWT would not be issued to reflect that change.<\/p><figure><img src=\"https:\/\/blog.dgraph.io\/images\/refreshJWT.png\" alt=\"image\" \/><figcaption>Refresh Token\n        <\/figcaption><\/figure><p>Therefore, Access Token expiration is typically set to a short duration, like a\nfew hours. But, this causes an issue with long running services, which would\nsuddenly get their access denied after this token expires.<\/p><p>When a <code>Login<\/code> happens, a refresh token is issued as well. This token\nis typically valid for a much longer duration, like a month. When access token\nexpires and an unauthorized response is received from the server, Dgraph clients\nwould automatically retry a login using the refresh token (if available). This would\ncause a new JWT to be issued with the updated groups.<\/p><p>Notice that when an access JWT expiration causes an automatic login using the refresh JWT, not only\ndoes the access JWT get replaced, but the refresh JWT itself also gets renewed with a later\nexpiration timestamp. Thus the same client can be used continually without an explicit call to\nLogin again after it\u2019s initialized, thanks to the rolling JWTs.<\/p><h3 id=\"logging\">Logging<\/h3><p>We log all of the <em>failed<\/em> access attempts in the main log files of Alpha servers, together with\nother types of events. All the log entries for ACL have a prefix <code>ACL-LOG<\/code>, so operators can run\na simple grep and filter them out. Here is an example log entry:<\/p><pre class=\"language-bash\">    I0320 18:36:44.386417       1 access_ee.go:650] ACL-LOG Authorizing user \"alice\" with groups \"\" on predicates \"name,friend\" for \"Read\", allowed:false\n<\/pre><p>To reduce verbosity of logging, we send the <em>authorized<\/em> access attempts as <a href=\"https:\/\/opencensus.io\/\">OpenCensus<\/a> traces,\nwhich can be further analyzed with <a href=\"https:\/\/www.jaegertracing.io\/\">Jaeger<\/a>. Below is a screenshot of the ACL log from the\nJaeger UI:\n<img src=\"https:\/\/blog.dgraph.io\/images\/jaeger-acl-log.png\" alt=\"ACL Log in Jaeger\" \/><\/p><h3 id=\"show-case\">Show case<\/h3><p>Below is a quick demo of the ACL feature in action using the command line tool.\n<img src=\"https:\/\/blog.dgraph.io\/images\/acl-demo.gif\" alt=\"ACL demo\" \/><\/p><h3 id=\"got-feedback\">Got feedback?<\/h3><p>That covers all the aspects of our first ACL release. We cannot wait to hear your feedback at\n<a href=\"https:\/\/discuss.dgraph.io\/\">our discussion forum.<\/a><\/p><hr \/><div class=\"engage\"><p>\n\t\t<strong>\n\t\t\tWe are building a fast, transactional and distributed graph database.\n\t\t\tIf you like what you read above,\n\t\t\t<a href=\"https:\/\/dgraph.io\/careers\">come join<\/a> the team!\n\t\t<\/strong>\n\t<\/p><table><thead><tr><\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/docs.dgraph.io\" target=\"_blank\">Get started with Dgraph<\/a><\/td>\n      <td><a href=\"https:\/\/docs.dgraph.io\" target=\"_blank\">https:\/\/docs.dgraph.io<\/a><\/td>\n    <\/tr><tr><td><a href=\"https:\/\/github.com\/dgraph-io\/dgraph\" target=\"_blank\">Star us on GitHub<\/a><\/td>\n      <td><a href=\"https:\/\/github.com\/dgraph-io\/dgraph\" target=\"_blank\">https:\/\/github.com\/dgraph-io\/dgraph<\/a><\/td>\n    <\/tr><tr><td><a href=\"https:\/\/discuss.dgraph.io\" target=\"_blank\">Ask us questions<\/a><\/td>\n      <td><a href=\"https:\/\/discuss.dgraph.io\" target=\"_blank\">https:\/\/discuss.dgraph.io<\/a><\/td>\n    <\/tr><\/tbody><\/table><strong>Fortune 500 companies are using Dgraph in production. If you need production support,\n    <a href=\"https:\/\/blog.dgraph.io\/cdn-cgi\/l\/email-protection#3f525e51564c577f5b584d5e4f57115650\">talk to us.<\/a><\/strong><\/div><br \/><em>Top image: <a href=\"https:\/\/en.wikipedia.org\/wiki\/STARSHINE_(satellite)\">On December 5th, 2001, STARSHINE-2 was launched from the Space Shuttle Endeavour on STS-108<\/a><\/em><div class=\"footnotes\"><hr \/><ol><li id=\"fn:1\">For now, each refresh retrieves the full set of ACL data. In the future, we\u2019ll switch to using a live stream of updates which is in the works.\n <a class=\"footnote-return\" href=\"#fnref:1\"><sup>[return]<\/sup><\/a><\/li>\n<\/ol><\/div>","created_at":"2019-09-21T16:14:49+0000","updated_at":"2019-09-21T16:15:00+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"blog.dgraph.io","preview_picture":"https:\/\/blog.dgraph.io\/images\/starshine-2.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","clojure"],"is_public":false,"id":1026,"title":"clj-refactor-unleash-your-clojure-wizard.html","url":"https:\/\/endlessparentheses.com\/clj-refactor-unleash-your-clojure-wizard.html","created_at":"2019-09-20T21:30:55+0000","updated_at":"2019-09-20T21:30:58+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"endlessparentheses.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1025,"title":"How to Generate Random Colors Programmatically","url":"https:\/\/martin.ankerl.com\/2009\/12\/09\/how-to-create-random-colors-programmatically\/","content":"<p>Creating random colors is actually more difficult than it seems. The randomness itself is easy, but aesthetically pleasing randomness is more difficult. For a little project at work I needed to automatically generate multiple background colors with the following properties:<\/p><ul><li>Text over the colored background should be easily readable<\/li>\n<li>Colors should be very distinct<\/li>\n<li>The number of required colors is not initially known<\/li>\n<\/ul><h2 id=\"na\u00efve-approach\">Na\u00efve Approach<\/h2><p>The first and simplest approach is to create random colors by simply using a random number between <code class=\"highlighter-rouge\">[0, 256[<\/code> for the R, G, B values. I have created a little Ruby script to generate sample HTML code:<\/p><div class=\"language-ruby highlighter-rouge\"><div class=\"highlight\"><pre># generates HTML code for 26 background colors given R, G, B values.\ndef gen_html\n  ('A'..'Z').each do |c|\n    r, g, b = yield\n    printf \"&lt;span style=\\\"background-color:#%02x%02x%02x; padding:5px; -moz-border-radius:3px; -webkit-border-radius:3px;\\\"&gt;#{c}&lt;\/span&gt; \", r, g, b\n  end\nend\n<\/pre><\/div><\/div><h2 id=\"naive-approach-generate-purely-random-colors\">naive approach: generate purely random colors<\/h2><div class=\"language-ruby highlighter-rouge\"><div class=\"highlight\"><pre>gen_html { [rand(256), rand(256), rand(256)] }\n<\/pre><\/div><\/div><p>The generated output looks like this:<\/p><p>A B C D E F G H I J K L M N O P Q R S T U V W X Y Z\n<\/p><p>As you can see this is quite suboptimal. Some letters are hard to read because the background is too dark (B, Q, S), other colors look very similar (F, R).<\/p><h2 id=\"using-hsv-color-space\">Using HSV Color Space<\/h2><p>Let\u2019s fix the too dark \/ too bright problem first. A convenient way to do this is to not use the RGB color space, but <a href=\"https:\/\/en.wikipedia.org\/wiki\/HSL_and_HSV\">HSV<\/a> (Hue, Saturation, Value). Here you get equally bright and colorful colors by using a fixed value for saturation and value, and just modifying the hue:<\/p><p><img src=\"https:\/\/martin.ankerl.com\/img\/2009\/12\/HSV_cylinder_small.png\" alt=\"HSV Cylinder\" style=\"text-align: middle;\" \/><\/p><p>Based on the description provided by the wikipedia article on <a href=\"https:\/\/en.wikipedia.org\/wiki\/HSL_and_HSV#Converting_to_RGB\">conversion from HSV to RGB<\/a> I have implemented a converter:<\/p><div class=\"language-ruby highlighter-rouge\"><div class=\"highlight\"><pre># HSV values in [0..1[\n# returns [r, g, b] values from 0 to 255\ndef hsv_to_rgb(h, s, v)\n  h_i = (h*6).to_i\n  f = h*6 - h_i\n  p = v * (1 - s)\n  q = v * (1 - f*s)\n  t = v * (1 - (1 - f) * s)\n  r, g, b = v, t, p if h_i==0\n  r, g, b = q, v, p if h_i==1\n  r, g, b = p, v, t if h_i==2\n  r, g, b = p, q, v if h_i==3\n  r, g, b = t, p, v if h_i==4\n  r, g, b = v, p, q if h_i==5\n  [(r*256).to_i, (g*256).to_i, (b*256).to_i]\nend\n<\/pre><\/div><\/div><p>Using the generator and fixed values for saturation and value:<\/p><div class=\"language-ruby highlighter-rouge\"><div class=\"highlight\"><pre># using HSV with variable hue\ngen_html { hsv_to_rgb(rand, 0.5, 0.95) }\n<\/pre><\/div><\/div><p>returns something like this:<\/p><div style=\"text-align: center;\">\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z\n<\/div><p>Much better. The text is easily readable, and all colors have a similar brightness. Unfortunately, since we have limited us to less colors now, the difference between the randomly generated colors is even less than in the first approach.<\/p><h2 id=\"golden-ratio\">Golden Ratio<\/h2><p>Using just <code class=\"highlighter-rouge\">rand()<\/code> to choose different values for hue does not lead to a good use of the whole color spectrum, it simply is too random.<\/p><p>Here I have generated 2, 4, 8, 16, and 32 random values and printed them all on a scale. Its easy to see that some values are very tightly packed together, which we do not want.<\/p><p><img src=\"https:\/\/martin.ankerl.com\/img\/2009\/12\/distribution-random.png\" alt=\"random distribution\" \/><\/p><p>Lo and behold, some mathematician has discovered the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Golden_ratio\">Golden Ratio<\/a> more than 2400 years ago. It has lots of interesting properties, but for us only one is interesting:<\/p><blockquote>\n<p>[\u2026] Furthermore, it is a property of the golden ratio, <em>\u03a6<\/em>, that each subsequent hash value divides the interval into which it falls according to the golden ratio!<\/p>\n<p>\u2013 <a href=\"http:\/\/brpreiss.com\/books\/opus4\/html\/page214.html\">Bruno R. Preiss, P.Eng.<\/a><\/p>\n<\/blockquote><p>Using the golden ratio as the spacing, the generated values look like this:<\/p><p><img src=\"https:\/\/martin.ankerl.com\/img\/2009\/12\/distribution-goldenratio.png\" alt=\"golden ratio distribution\" \/><\/p><p>Much better! The values are very evenly distributed, regardless how many values are used. Also, the algorithm for this is extremely simple. Just add 1\/\u03a6 and modulo 1 for each subsequent color.<\/p><div class=\"language-ruby highlighter-rouge\"><div class=\"highlight\"><pre># use golden ratio\ngolden_ratio_conjugate = 0.618033988749895\nh = rand # use random start value\ngen_html {\n  h += golden_ratio_conjugate\n  h %= 1\n  hsv_to_rgb(h, 0.5, 0.95)\n}\n<\/pre><\/div><\/div><p>The final result:<\/p><div style=\"text-align: center;\">\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z\n<\/div><p>You can see that the first few values are very different, and the difference decreases as more colors are added (Z and E are already quite similar). Anyways, this is good enough for me.<\/p><p>And because it is so beautiful, here are some more colors <img class=\"emoji\" title=\":smiley:\" alt=\":smiley:\" src=\"https:\/\/github.githubassets.com\/images\/icons\/emoji\/unicode\/1f603.png\" height=\"20\" width=\"20\" \/><\/p><h3 id=\"s099-v099\">s=0.99, v=0.99<\/h3><div style=\"text-align: center;\">\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z\n<\/div><h3 id=\"s025-h08\">s=0.25, h=0.8<\/h3><div style=\"text-align: center;\">\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z\n<\/div><h3 id=\"s03-v099\">s=0.3, v=0.99<\/h3><div style=\"text-align: center;\">\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z\n<\/div><p>Have fun!\nMartin<\/p>","created_at":"2019-09-20T20:10:18+0000","updated_at":"2019-09-20T20:10:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"martin.ankerl.com","preview_picture":"https:\/\/martin.ankerl.com\/img\/me.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["java","jar","shell"],"is_public":false,"id":1024,"title":"Making Really Executable Jars","url":"https:\/\/skife.org\/java\/unix\/2011\/06\/20\/really_executable_jars.html","content":"<p>One of the more annoying things about writing command line applications in Java is that the Java model of an executable is the so called executable jar, which is executed via an incantation like<\/p><div class=\"highlight\"><pre class=\"sh\">$ java -jar .\/waffles-1.2.3.jar --some-flag=blue hello \n<\/pre><\/div><p>There has long been a hack known in some circles, but not widely known, to make jars <em>really<\/em> executable, in the <code>chmod +x<\/code> sense. The hack takes advantage of the fact that jar files are zip files, and zip files allow arbitrary cruft to be prepended to the zip file itself (this is how self-extracting zip files work).<\/p><p>To do this for jar files, on unix-like operating systems, create a little shell script which looks like:<\/p><div class=\"highlight\"><pre class=\"sh\">#!\/bin\/sh\nexec java -jar $0 \"$@\"\n<\/pre><\/div><p>You can make it fancier, doing things like looking for JAVA_HOME and so on, but the above is enough to get started. Make sure to add a few newlines at the end, they are very important. If you leave them out it will not work.<\/p><p>Now that you have your little shell script, cat the executable jar you want onto the end of it, set the script +x, and go to town. If you script is named <code>waffles<\/code>, then you would do that like:<\/p><div class=\"highlight\"><pre class=\"sh\">$ cat .\/waffles-1.2.3.jar &gt;&gt; .\/waffles\n$ chmod +x .\/waffles\n$ .\/waffles --some-flag=blue hello\n<\/pre><\/div><p>and there you go! I have a little maven plugin that will do this for you automagically, but haven\u2019t had a chance to get it into central yet. I guess I should probably stop writing and go do so\u2026<\/p><ul><li>\n<p><a href=\"http:\/\/twitter.com\/electrum32\">David Phillips<\/a> suggested putting the <code>$@<\/code> in parens as it can contain spaces. I have updated the post to do so.<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/twitter.com\/ssch0ber\">Sven Schober<\/a> pointed out a bug in the original form of the shell script I posted. I forgot the extremely important <code>$0<\/code>. That is what I get for writing from memory and not unit testing my blog posts! The post has been fixed.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/twitter.com\/jeffreymcmanus\">Jeffrey McManus<\/a> found a typo, I had <code>chomd<\/code> instead of <code>chmod<\/code>. Fixed, thank you! I really need to find a way to unit test blog posts!<\/p>\n<\/li>\n<\/ul><noscript><div>Please enable JavaScript to view the <a href=\"http:\/\/disqus.com\/?ref_noscript\">comments powered by Disqus.<\/a><\/div><\/noscript><p><a href=\"https:\/\/skife.org\/src\/libraries\/2011\/07\/15\/long_tail_treasure_trove_2.html\" class=\"forward\">Yet More of the Long Tail Treasure Trove \u276f\u276f<\/a>\n        \n        \n        <a href=\"https:\/\/skife.org\/\">Home<\/a> | <a href=\"https:\/\/skife.org\/archive.html\">Archives<\/a>\n        \n        \n            <a href=\"https:\/\/skife.org\/emacs\/mac\/osx\/2011\/03\/17\/emacs-client-mac.html\" class=\"back\">\u276e\u276e Emacs Client on Mac OS X<\/a><\/p>","created_at":"2019-09-18T19:21:18+0000","updated_at":"2019-09-18T19:21:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"skife.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["datalog","logic-programming"],"is_public":false,"id":1023,"title":"Inference in Datalog","url":"https:\/\/cs.nyu.edu\/faculty\/davise\/ai\/datalog.html","content":"Inference in Datalog \nDatalog: The subset of the predicate calculus consisting of sentences with\n<ul><li> No function symbols.\n<\/li><li> No existential quantifiers. \n<\/li><li> All sentences have the form \n<blockquote><div>\n\"{ forall(VARS) } ATOMIC-FORM\"    or <br \/>\"{ forall(VARS) } [ATOMIC-FORM ^ ATOMIC-FORM ^ ... ^ ATOMIC-FORM] =&amp;gt \nATOMIC-FORM\"\n<\/div><\/blockquote>\n<\/li><\/ul>\nSince all variables are universally quantified, with a scope of the entire\nsentence, there is no need to actually write down the quantifiers,\nas long as there is some convention about which symbols are variables and\nwhich are constants.  Here, I will use upper case for variables and lower\ncase for constants.\n<p>\nThus, all sentences have either the form.\n<\/p><blockquote><div>\n1. PRED(TERM ... TERM);  or <br \/>2. [PRED(TERM ...  TERM) ^ PRED(TERM ...  TERM) ^ ... ^ \n    PRED(TERM ...  TERM)] =&amp;gt <br \/>PRED(TERM ...  TERM)\n<\/div><\/blockquote>\nwhere all the TERM's are either variables or constants.\n<p>\nSentences of form (1) are called \"facts\"; sentences of form (2) are\ncalled \"rules\".  The formula on the right hand of the implication of a rule\nis called the \"head\" of the rule; the formulas on the left hand are the\n\"tail\" of the rule.\n<\/p><h4> Example <\/h4>\nThe following is a Datalog knowledge base describing a family tree and rules\nfor relations:\n<blockquote>\n<div style=\"text-align: center;\">\n<b> Rules <\/b>\n<\/div><div>\n1. parent(X,Y) ^ male(X) =&amp;gt father(X,Y). <\/div><div><br \/>2. father(X,Y) =&amp;gt parent(X,Y). <br \/>3. father(X,Y) =&amp;gt male(X). <br \/>4. parent(X,Y) ^ female(X) =&amp;gt mother(X,Y). <br \/>5. mother(X,Y) =&amp;gt parent(X,Y). <br \/>6. mother(X,Y) =&amp;gt female(X). <br \/>7. child(X,Y,Z) =&amp;gt parent(Y,X). <br \/>8. child(X,Y,Z) =&amp;gt parent(Z,X). <br \/>9. son(X,Y,Z) =&amp;gt child(X,Y,Z). <br \/>10. son(X,Y,Z) =&amp;gt male(X). <br \/>11. male(X) ^ child(X,Y,Z) =&amp;gt son(X,Y,Z). <br \/>12. daughter(X,Y,Z) =&amp;gt child(X,Y,Z). <br \/>13. daughter(X,Y,Z) =&amp;gt female(X). <br \/>14. female(X) ^ child(X,Y,Z) =&amp;gt daughter(X,Y,Z). \n<div style=\"text-align: center;\">\n<b> Facts<\/b>\n<\/div>\n15. male(philip). <br \/>16. female(elizabeth). <br \/>17. son(charles,philip,elizabeth). <br \/>18. daughter(anne,philip,elizabeth). <br \/>19. son(andrew,philip,elizabeth). <br \/>20. son(edward,philip,elizabeth). <br \/>21. son(william,charles,diana). <br \/>22. son(harry,charles,diana). <br \/><\/div><\/blockquote>\nThe following sentences in predicate calculus <em> cannot <\/em> be expressed\nin Datalog:\n<ul><li> forall(X) exists(Y,Z) child(X,Y,Z).  (Datalog doesn't deal with exists)\n<\/li><li> forall(X) male(X) V female(X).  (nor with disjunction)\n<\/li><li> forall(X) not(male(X) ^ female(X)) (nor with negation).\n<\/li><\/ul><h3> Inference <\/h3>\nThe method of inference we will discuss is a mixture of two ideas:\n<p>\nI. <b>Modus Ponens.<\/b> From the sentences\n<\/p><blockquote><div>\n[P^Q^R^S] =&amp;gt W. <br \/>P. <br \/>Q. <br \/>R. <br \/>S. <br \/><\/div><\/blockquote>\nyou can infer W.\n<p>\nII. (Universal instantiation)\nFrom the sentence \"forall(X) P(X)\"  you may infer P(T) for any term T.\n<\/p><p>\nNow, suppose we have the sentences\n<\/p><blockquote><div>\n1. forall(X) man(X) =&amp;gt mortal(X).  (All men are mortal) <br \/>2. man(socrates).\n<\/div><\/blockquote>\nThen we can proceed as follows:\n<ul><li> First, use universal instantiation to replace \"X\" in (1) by \"socrates\"\ngiving \n<blockquote><div>\n3. \"man(socrates) =&amp;gt mortal(socrates)\"\n<\/div><\/blockquote>\n<\/li><li> Second, apply modus ponens to (2) and (3), and infer \"mortal(socrates)\".\n<\/li><\/ul>\nThe question is, how did we know to replace X in (1) by \"socrates\" rather\nthan, say \"eiffel_tower\" or anything else?  The answer is that we chose\nto replace X by socrates in order to make possible the use of modus ponens.\nThis is done by <em> unifying <\/em> the atomic formula \"man(X)\" in (1)\nwith the formula \"man(socrates)\" in (2).\n<h3> Substitution and Unification in Datalog <\/h3>\nA <em> substitution <\/em> in Datalog is a mapping of variable symbols to\nvariable or constant symbols.  That is, it is a set of pairs of the form\nV -&gt; W where:\n<ul><li> V is a variable. \n<\/li><li> W is either a variable or a constant.\n<\/li><li> No variable appears on the left hand side of two different pairs.\n<\/li><li> No variable appears on the left hand side of one pair and the right\nhand side of a different pair.\n<\/li><\/ul><p>\n<b> Example: <\/b> The set { X -&gt; a, Y -&gt; a, Z -&gt; b, P -&gt; Q, M -&gt; Q}\nis a substitution.\n(Again, variables are upper-case; constants are lower case.)\n<\/p><p>\n<b> Non-examples.<\/b>  The set {X -&gt; a, X -&gt; b, Y -&gt; z} is not valid because\nX is on the left hand of two different pairs. The set\n{X -&gt; a, b -&gt; a, Y -&gt; b} is not valid because the constant b appears on\nthe left hand  of a pair. The set { X -&gt; a, Y -&gt; X, Z -&gt; b } is not valid\nbecause X appears on both the left and right side of a pair.\nappears in two sets). The collection {{X,a,b}, {Y}} is not a valid\n<\/p><h4> Applying substitution U to sentence S = PRED(T1 ... Tk) <\/h4>\nIf V -&gt; W occurs in U, then replace every occurrence of V in S by W.\n<p>\n<b> Example <\/b>\nLet U be the substitution \n{ X -&gt; a, Y -&gt; a, Z -&gt; b, P -&gt; Q, M -&gt; Q}. <br \/>Let S be the sentence\nf(X,Y,c) ^ g(Y,Z,W) ^ h(P,c,X) =&gt; h(X,c,Q)\n<\/p><p>\n The result of applying U to S is then <br \/>f(a,a,c) ^ g(a,b,W) ^ h(Q,c,a) =&gt; h(a,c,Q)\n<\/p><h4> Unifying two atomic formulas <\/h4>\nLet F1 and F2 be two atomic formulas with no variables in common.\nA substitution B <em> unifies <\/em> formulas F1 and F2 if the result of\napply(B,F1) = apply(B,F2).\n<p>\nLet F1 be the formula \"P(T1 ... Tk)\" and let F2 be the formula \"Q(W1 ... Wm)\"\n<\/p><pre>\nunify(F1,F2) : returns either a unifying substitution, or NIL if there is no\n               unifying substitution\n{ if (P != Q or k != m) then return(NIL);\n  C = collection containing each of the symbols in different sets;\n  for i = 1 to k do { \n    let E(Ti) and E(Vi) be the sets in C containing Ti and Vi;\n    if E(Ti) and E(Vi) each contain a constant (necessarily different)\n       then return(NIL);\n    if (Ti != Vi) then merge Ti with Vi in C;\n    } \n  B := emptyset;\n  for (S in C) {\n    if (S has more than one element) {\n      if (S contains constant symbol X) \/* S contains at most one constant *\/\n        for (each variable Q in S) add Q -&gt; X to B;\n      else { X = some variable in S;\n             for (Q != X in S) add Q -&gt; X to B;\n    } }\n}\n<\/pre>\nThis can be made very efficient using the UNION-FIND or MERGE-FIND algorithm.\n<p>\nExample: To unify the formula F1=\"p(a,X,b,X,Y,W)\" with F2=\"p(U,U,T,V,c,T)\":\n<\/p><pre>\nSince both formulas have predicate \"p\" and 6 arguments,\nInitialize C to be {{a}, {b}, {c}, {T}, {U}, {V}, {W}, {X}, {Y}}\nFirst loop:\nStep 1: T1=\"a\", V1=\"U\". Merge {a} with {X}.\nC  = {{a,U}, {b}, {c}, {T}, {V}, {W}, {X}, {Y}}\nStep 2: T2=\"X\", V2=\"U\". Merge {X} with {a,U} \nC  = {{a,U,X}, {b}, {c}, {T}, {V}, {W}, {Y}}\nStep 3: T3=\"b\", V3=\"T\". Merge {b} with {T} \nC  = {{a,U,X}, {b,T}, {c}, {V}, {W}, {Y}}\nStep 4: T4=\"X\", V4=\"V\". Merge {a,U,X} with {V} \nC  = {{a,U,X,V}, {b,T}, {c}, {W}, {Y}}\nStep 5: T5=\"Y\", V5=\"c\". Merge {Y} with {c} \nC  = {{a,U,X,V}, {b,T}, {c,Y}, {W}} \nStep 6: T6=\"W\", V6=\"T\". Merge {W} with {b,T}\nC = {{a,U,X,V}, {b,T,W}, {c,Y} }\nSecond loop:\nB = { U -&gt; a, X -&gt; a, V -&gt; a, T -&gt; b, W -&gt; b, Y -&gt; c}.\n<\/pre>\nNote that apply(C,F1) = apply(C,F2) = \"p(a,a,b,a,c,b)\"\n<h3> Forward Chaining <\/h3>\nThe basic idea in forward chaining is this: You look for unifications that\nwill match each of the formulas in the left hand side of a known rule\nto known facts. You can then infer the right hand side of the rule \nunder the specified substitution. This is a fact that you can add to the\nknowledge base. You keep doing this until there is nothing new to be derived, \nthen stop.\n<p>\nExample: Suppose you start with the following knowledge base (a subset of\nthe knowledge base at the start of these notes: hence the gappy numbering).\n<\/p><blockquote>\n<div style=\"text-align: center;\">\n<b> Rules <\/b>\n<\/div><div>\n1. parent(X,Y) ^ male(X) =&amp;gt father(X,Y). <\/div><div><br \/>2. father(X,Y) =&amp;gt parent(X,Y). <br \/>3. father(X,Y) =&amp;gt male(X). <br \/>4. parent(X,Y) ^ female(X) =&amp;gt mother(X,Y). <br \/>5. mother(X,Y) =&amp;gt parent(X,Y). <br \/>6. mother(X,Y) =&amp;gt female(X). <br \/>7. child(X,Y,Z) =&amp;gt parent(Y,X). <br \/>8. child(X,Y,Z) =&amp;gt parent(Z,X). <br \/>9. son(X,Y,Z) =&amp;gt child(X,Y,Z). <br \/>10. son(X,Y,Z) =&amp;gt male(X). <br \/>11. male(X) ^ child(X,Y,Z) =&amp;gt son(X,Y,Z). <br \/><div style=\"text-align: center;\">\n<b> Facts<\/b>\n<\/div>\n15. male(philip). <br \/>16. female(elizabeth). <br \/>17. son(charles,philip,elizabeth). <br \/><\/div><\/blockquote>\nWe can proceed as follows:\n<blockquote><div>\nCombining rule (9) with fact (17) \nwith substitution {X -&gt; charles, Y -&gt; philip, Z -&gt; elizabeth}, \ninfer (23) child(charles,philip,elizabeth)\n<p>\nCombining rule (10) with fact (17) \nwith substitution {X -&gt; charles, Y -&gt; philip, Z -&gt; elizabeth}, \ninfer (24) male(charles).\n<\/p><p>\nCombining rule (7) with fact (23) \nwith substitution {X -&gt; charles, Y -&gt; philip, Z -&gt; elizabeth}, \ninfer (25) parent(philip,charles)\n<\/p><p>\nCombining rule (8) with fact (23) \nwith substitution {X -&gt; charles, Y -&gt; philip, Z -&gt; elizabeth}, \ninfer (26) parent(elizabeth,charles)\n<\/p><p>\nCombining rule (1) with facts (15) and (25) \nwith substitution {X -&gt; philip, Y -&gt; charles}\ninfer (27) father(philip,charles).\n<\/p><p>\nCombining rule (4) with facts (16) and (26) with \nwith substitution {X -&gt; elizabeth, Y -&gt; charles}\ninfer (28) mother(elizabeth,charles).\n<\/p><p>\nNo further facts can be derived.\n<\/p><\/div><\/blockquote>\n<p>\nIMPORTANT NOTE: Each time you are trying to unify a fact with a set of rules,\nyou must rename the variables so that no two of the sentences use\nthe same variable names.\n<\/p><p>\nHere is some pseudo-code to do forward chaining exhaustively.\nI am\nnot going to go through this in detail.\n<\/p><pre>\nfc(RR : set of rules; FF: set of facts) return set of facts. {\nrepeat { NEW_FACT := false;\n         for each rule R = \"T1 ^ T2 ^ ... ^ Tk =&amp;gt H\" do {\n           for each fact F1 in consequences(R,FF) do {\n             if (F1 is not in FF)  then\n               { NEW_FACT := true;\n                 FF := add F1 to FF;\n                 }\n            }\n         }\n        }\n  until not(NEW_FACT);\n}\n\/* \"consequences\" recurs down the formulas in the tail of rule R *\/\nconsequences(R : rule or fact; FF : set of facts) return set of facts;\n{ if (R is a fact) then return(R); \/* Base case of the recursion *\/\n  T1 := the first literal in the tail of R;\n  NEW_FACTS := empty set;\n  for each fact (F in FF) {\n     rename the variables in F so as not to conflict with the variables in R;\n     U = unify(T1,F);\n     if (U != NIL) then {\n       R1 := drop T1 from R;\n       R1 := apply(U,R1);\n       NEW_FACTS := NEW_FACTS union consequences(R1,FF)\n      }\n    }\n  return(NEW_FACTS);\n<\/pre>  \n    \nA trace of the subroutine \"consequences\" for a reasonably complicated\nexample is given <a href=\"https:\/\/cs.nyu.edu\/faculty\/davise\/ai\/fc.html\"> here <\/a>\n    \n<h3> Backward Chaining <\/h3>\nIn backward chaining you start with a <em> goal <\/em>, a statement that\nyou want to prove, in addition to the knowledge base, and you look\nfor derivations that allow you to prove the goal. You look for a rule R or\na fact F.\nwhose head unifies with the goal, and then recursively try to prove\nall the formulas in the tail of R.\n<p>\nFor instance, suppose\nthat you start with the following knowledge base, \n<\/p><blockquote>\n<div style=\"text-align: center;\">\n<b> Rules <\/b>\n<\/div><div>\n1. parent(X,Y) ^ male(X) =&amp;gt father(X,Y). <\/div><div><br \/>2. parent(X,Y) ^ female(X) =&amp;gt mother(X,Y). <br \/>3. child(X,Y,Z) =&amp;gt parent(Y,X). <br \/>4. child(X,Y,Z) =&amp;gt parent(Z,X). <br \/>5. male(X) ^ child(X,Y,Z) =&amp;gt son(X,Y,Z). <br \/>6. female(X) ^ child(X,Y,Z) =&amp;gt daughter(X,Y,Z).\n<div style=\"text-align: center;\">\n<b> Facts<\/b>\n<\/div>\n7. male(philip). <br \/>8. female(elizabeth). <br \/>9. child(charles,philip,elizabeth). <br \/>10. child(anne,philip,elizabeth). <br \/>11. male(charles). <br \/>12. female(anne).\n<\/div><\/blockquote>\nand the goal is\n\"mother(elizabeth,charles)\".  Backward chaining proceeds as follows:\n<pre>\nGoal G0: mother(elizabeth,charles).\n  Unify G0 with the head of rule (1) with substitution {X1-&gt;elizabeth, Y1-&gt;charles}}\n  Subgoals: G1 = \"parent(elizabeth,charles)\", G2=\"female(elizabeth)\".\n    Goal G1: parent(elizabeth,charles)\n      Unify G1 with the head of rule (3) with substitution X2 -&gt; charles, Y2 -&gt; elizabeth\n      Subgoal G3 = \"child(charles,elizabeth,Z2\")\n         No fact or rule unifies with G3.\n         Fail G3\n      Try again with G1.\n      Unify G1 with the head of rule (4) with substitution X2 -&gt; charles, Y2 -&gt; elizabeth\n      Subgoal G3 = \"child(charles,Y3,elizabeth)\"\n         G3 unifies with fact (9) with substitution Y3=elizabeth.\n         G3 succeeds.\n      G1 succeeds.\n    Goal G2: female(elizabeth)\n       G2 unifies with fact 8.\n       G2 succeeds\nG0 succeeds.\n<\/pre>\nIf you have a variable in the goal in backward chaining, then the\ngoal is to find _some_ value of the variable that satisfies the formula.\nFor example, if the goal is \"child(charles,J,K)\", that means \"find J and\nK such that charles is the child of J and K.  Backward chaining in that case\nproceeds as follows.\n<pre>\nG0: child(charles,J,K)\n  G0 unifies with fact 9, under substitution {J-&gt;philip, K-&gt;elizabeth}.\nG0 succeeds with J=philip and K=elizabeth.\n<\/pre>\nAnother example: if \nthe goal is \"daughter(J,K,L)\" (find J,K,L such that J is the daughter of K and L)\nthen backward chaining proceeds as follows:\n<pre>\nGoal G0: daughter(J,K,L)\n   G0 unifies with rule 6 under the substitution {X1-&gt;J, Y1-&gt;K, Z1-&gt;L}\n   Subgoals: G1=\"female(J)\", G2=\"child(J,K,L)\"\n   Goal G1: female(J).\n     Unify G1 with fact 8, with binding J-&gt;elizabeth.\n     G1 succeeds with binding J-&gt;elizabeth\n   Goal G2: child(elizabeth,K,L)\n     G2 does not unify with any fact or rule.\n     G2 fails\n   Return to G1 (last point where there was a choice).\n     Try unifying G1 with fact 12, with binding J=anne.\n     G1 succeeds with binding J=anne.\n   Goal G2: child(anne,K,L)\n      G2 unifies with fact 10 with binding K-&gt;philip, L-&gt;elizabeth.\n   \n   G0 succeeds with J=anne, K=elizabeth, L=philip.\n<\/pre>\nUnlike forward chaining, backward chaining can potentially go into an\ninfinite loop.  There are two ways to avoid this:\n<ul><li> Responsibility of the knowledge base designer to set up the rules so\nthat this doesn't happen. (Prolog)\n \n<\/li><li> Repeated goal detection: If goal Gi is a subgoal of Gj (direct or\nindirect) and Gi is identical to Gj, then fail Gi.\n<\/li><\/ul>\nIn Datalog, repeated goal detection is sufficient to remove all infinite \nloops.\n<h3> Reformulating Some Predicate Calculus Sentences into Datalog <\/h3>\nIn some cases a predicate calculus sentence which does not immediately\nappear to satisfy the conditions for Datalog can be reformulated\ninto a logically equivalent form that does satisfy the conditions. The\nmost important case involves existentially quantified variables on the \nleft-hand side of an implication.  Suppose you have the rule, ``X is\na grandparent of Z if there exists a Y such that X is a parent of Y\nand Y is a parent of Z.'' The natural formulation of this in predicate \ncalculus is, \n<blockquote><div>\n(A) forall(X,Z) [exists(Y) parent(X,Y) ^ parent(Y,Z)] =&gt; grandparent(X,Z).\n<\/div><\/blockquote>\nIn that form, this does not satisfy the conditions on Datalog because of\nthe existential quantifier.  However, it is logically equivalent to\n<blockquote><div>\n(B) forall(X,Y,Z) parent(X,Y) ^ parent(Y,Z) =&gt; grandparent(X,Z)\n<\/div><\/blockquote>\nwhich is valid Datalog.\n<p>\nLet me give two arguments for this equivalence: an intuitive argument and\na formal argument.\n<\/p><p>\n<b> Intuitive argument:<\/b> <br \/>Sentence (A) means \"For any X and Z that we choose, if we can find a Y\nsuch that parent(X,Y) and parent(Y,Z) then grandparent(X,Z).\" Now suppose\nthat Elizabeth is a parent of Charles and Charles is a parent of William,\nthus satisfying the condition of (B).\nThen the condition of (A) is satisfied for X=Elizabeth and Z=William \nwe can choose a Y such that parent(elizabeth,Y) and \nparent(Y,william); namely, Charles. \nConversely, suppose that Jesse and Solomon satisfy\nthe condition for (B).  That is, there exists some particular Y1 such that\nJesse is a parent of Y1 and Y1 is a parent of Solomon.  Therefore, the\ncondition in (A) is satisfied with X=Jesse, Y=Y1, and Z=Solomon. Thus, (A) and\n(B) are equivalent.\n<\/p><p>\n<b> Formal argument:<\/b> <br \/>(A) forall(X,Z) [exists(Y) parent(X,Y) ^ parent(Y,Z)] =&gt; grandparent(X,Z) <br \/>is equivalent to  <br \/>(C) forall(X,Z) ~[exists(Y) parent(X,Y) ^ parent(Y,Z)] V grandparent(X,Z) <br \/>which is equivalent to  <br \/>(D) forall(X,Z) [forall(Y) ~[parent(X,Y) ^ parent(Y,Z)]] V grandparent(X,Z) <br \/>which is equivalent to  <br \/>(E) forall(X,Z) [forall(Y) ~[parent(X,Y) ^ parent(Y,Z)] V grandparent(X,Z)] <br \/>which is equivalent to  <br \/>(B) forall(X,Z) [forall(Y) [parent(X,Y) ^ parent(Y,Z)] =&gt; grandparent(X,Z) \n<\/p><p>\n(A) is equivalent to (C) and (E) is equivalent to (B) \nbecause P=&gt;Q is always equivalent to ~P V Q. <br \/>(C) is equivalent to (D) because ~exists(X) P(X) is always equivalent to \nforall(X) ~P(X). (If there does not exists anything satisfying P, then nothing\nsatisfies P.) <br \/>(D) is equivalent to (C) because [forall(X) P(X)] V Q is equivalent to\n[forall(X) P(X) V Q] if variable X does not occur in Q.\n<\/p>","created_at":"2019-09-12T06:18:41+0000","updated_at":"2019-09-12T06:18:48+0000","annotations":[],"mimetype":"text\/html","reading_time":14,"domain_name":"cs.nyu.edu","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["databases","datalog","datomic"],"is_public":false,"id":1022,"title":"The Architecture of Datomic","url":"https:\/\/www.infoq.com\/articles\/Architecture-Datomic\/","content":"<p><a href=\"http:\/\/datomic.com\/\" target=\"_blank\">Datomic<\/a> is a new database designed as a composition of simple services. It strives to strike a balance between the capabilities of the traditional RDBMS and the elastic scalability of the new generation of redundant distributed storage systems.<\/p><h2>Motivations<\/h2><p>Datomic seeks to accomplish the following goals:<\/p><ul><li>Provide for a sound information model, eschewing update-in-place<\/li>\n    <li>Leverage redundant, scalable storage systems<\/li>\n    <li>Provide ACID transactions and consistency<\/li>\n    <li>Enable declarative data programming in applications<\/li>\n<\/ul><p>Any database system needs to have a point of view about the data model it supports. Traditional RDBMS support the relational model combined with some sort of world-update semantic. On the opposite end of the spectrum, some of the new NoSQL systems know little or nothing of the information they contain, simply storing blobs at keys, with eventual consistency. Datomic considers a database to be an <em>information system<\/em>, where information is a set of facts, and facts are things that have happened. Since one can't change the past, this implies that the database <em>accumulates<\/em> facts, rather than updates places, and that while the past may be forgotten, it is immutable. Thus if someone 'changes' their address, Datomic stores the fact that they have a new address without replacing the old fact (it has simply been retracted as of this point in time). This immutability leads to many important architectural benefits and opportunities.<\/p><p>Many great strides have been made in defining highly available, redundant, distributed and scalable storage systems, such as Amazon's <a href=\"http:\/\/www.allthingsdistributed.com\/2007\/10\/amazons_dynamo.html\" target=\"_blank\">Dynamo<\/a>. In moving away from a monolithic design, Datomic seeks to support these systems directly, offering users their choice of storage system, location, cost and scalability characteristics.<\/p><p>ACID transactions are quite important to many business processes. While much has been made of the difficulty of scaling ACID, as we'll see, Datomic separates transactions from read\/query, and in doing so maximizes the throughput of the transactional component. That said, there is a definite tradeoff here, one that Datomic makes in favor of ACID transactions instead of unlimited write scalability, and targets domains where that tradeoff is appropriate. Consistency is a greatly simplifying characteristic of systems, leading to more robust solutions, and shouldn't be discarded without good reason. (There are, sometimes, good reasons, but make sure they apply to you!).<\/p><p>Traditional RDBMS provide powerful logical, declarative and set-oriented language for manipulating data. Unfortunately, that power is trapped in servers, and once the data reaches the application, too often vanishes in a sea of nested <em>for<\/em> loops and imperative manipulation. Datomic provides what is fundamentally a distributed index, and allows the declarative query component to reside in the application server tier. This query engine can manipulate both Datomic-hosted data and in-memory data sources, providing for a consistent, high-level approach.<\/p><h2>Overarching Concerns<\/h2><p>There are many concepts that pervade the Datomic architecture, among them:<\/p><ul><li>&#13;\n    <p>Simplicity and composition. The system should comprise a set of services each of which does one thing well, and knows little or nothing about the others. Where well-regarded, simple components already exist in a particular domain, such as storage or caching, they should be integrable.<\/p>&#13;\n    <\/li>\n    <li>&#13;\n    <p>Scalability and elasticity of storage and query. If you need more storage or query capability, you should be able to dynamically add machines, and if less, remove them, without preconfiguration.<\/p>&#13;\n    <\/li>\n    <li>&#13;\n    <p>Location flexibility. The components should not care where the other components reside.<\/p>&#13;\n    <\/li>\n    <li>&#13;\n    <p>Pervasive immutability. Programming using update-in-place is an idea whose time has passed. Most programming should be done with immutable values. In Datomic, even the database itself can be treated as a value.<\/p>&#13;\n    <\/li>\n    <li>&#13;\n    <p>Data as interface. The system should be programmable. In order to be so, the primary interfaces should be data-driven, not syntax-laden strings. In Datomic, schema, transactions, queries and query results are all defined in terms of ordinary data structures such as lists and maps.<\/p>&#13;\n    <\/li>\n<\/ul><h2>Logical Model<\/h2><p>Datomic comprises the following services:<br class=\"image-wide\" \/><img alt=\"\" src=\"https:\/\/www.infoq.com\/articles\/Architecture-Datomic\/articles\/Architecture-Datomic\/en\/resources\/DatomicLogicalModel.jpg\" style=\"margin-top: 2px; margin-bottom: 2px; margin-left: 2px; margin-right: 2px; float: right;\" \/><\/p><ul><li>Storage<\/li>\n    <li>Transaction coordination and notification<\/li>\n    <li>Indexing<\/li>\n    <li>Application data model and query<\/li>\n    <li>Caching<\/li>\n<\/ul><p>The capabilities required of storage are minimal - primarily storage of small (&lt; 64k) blobs. Storage is used by the system much in the same way a file system might be used by a traditional database - for log and index segments. These segments are immutable. In addition, a small set of keys are used as references to the log and index roots. These are the only mutable use of storage, and require the storage system to provide consistent write\/read, and, at present, conditional or versioned put.<\/p><p>All additions of information to the database take the form of submissions of novelty to the transaction coordinator. The coordinator serializes all transactions, and ensures ACID semantics, placing the new information in log segments in storage. Upon durable storage, the coordinator notifies the transaction submitter and any connected parties of the new information resulting from the transaction.<\/p><p>The second use of storage, the index, is actually several sets of the database data, sorted in various orders, represented as trees of segments. Periodically, the indexing service incorporates new information from the logs in storage into these sorted sets. Since the index segments are immutable, the indexing process always creates a new index tree, that tree sharing structure with the prior index tree as much as possible.<\/p><p>Datomic presents to the application server tier a view of the database as if it was a value in memory. This database value provides a view of the database as a graph of entities and their attributes. In addition, the database object provides direct iterable access to the indexes themselves. Finally, the database value can serve as an argument to a declarative query. In order to present what appears to be a 'current' view of the database, the database object must merge the most recent index with an in-memory index of the novelty that has occurred since the last durable index was produced.<\/p><p>Datomic also provides to the application server an embedded query engine which can operate upon the database values, as well as in-memory collections.<\/p><p>Caching is fully integrated into the Datomic model. Unlike the typical scenario, where caching is an application-level problem, used to cache query results in an effort to ease the load on servers, Datomic manages caches itself, and caches the index segments themselves, in order to bring them into memory from storage, or into the application process. This ability to cache 'under' the queries derives from the fact that index segments are immutable, and query is distributed.<\/p><h2>Physical Model<\/h2><p>The logical model maps in a straightforward way to the physical model, with independent subsystems for most services.<\/p><p class=\"image-wide\"><img alt=\"\" src=\"https:\/\/www.infoq.com\/articles\/Architecture-Datomic\/articles\/Architecture-Datomic\/en\/resources\/large-DatomicPhysicalModel1.jpg\" \/><\/p><ul><li>Storage services<\/li>\n<\/ul><p>Perhaps surprisingly, Datomic is not in the business of disks. Instead, it integrates with a wide variety of storage servers and systems, as long as they satisfy the minimal requirements mentioned previously. Examples of supported storages are:<\/p><ul><li>In-process memory (no storage). Datomic can run completely in memory, in-process, which is great for development and testing.<\/li>\n    <li>SQL. Any JDBC-compliant SQL database can serve as storage.<\/li>\n    <li>Key\/value stores. Next-generation distributed, redundant key\/value stores such as Amazon's <a href=\"http:\/\/aws.amazon.com\/dynamodb\/\" target=\"_blank\">DynamoDB<\/a> are a great fit for Datomic.<\/li>\n    <li>Memory grids. Distributed data grids such as <a href=\"http:\/\/www.jboss.org\/infinispan\/\" target=\"_blank\">Infinispan<\/a> can be used when redundant memory is sufficiently durable.<\/li>\n<\/ul><p>The result of this 'servicification' of storage is a tremendous amount of flexibility. Note also that in many cases (e.g. SQL and Infinispan), Datomic need not have a dedicated storage resource. You can have Datomic use an existing SQL server by simply adding a single table. When used with a highly-scalable storage such as DynamoDB, Datomic acquires the same elastic read scalability. This is another benefit of the component-based architecture - read and write scaling become orthogonal decisions, with different tradeoffs.<\/p><p>Datomic encapsulates the differences between storages, and operates identically on all of them. You can move a system from one storage to another by a simple export\/import to\/from an S3 bucket or local directory, and change your app from one storage to another with a few property file changes and a different URI.<\/p><ul><li>the Transactor<\/li>\n<\/ul><p>The transactor is a process dedicated to transaction coordination. It accepts persistent connections from application server peers. Over these connections it accepts transactions, serializes them, commits log segments to storage, and broadcasts novelty to all connected peers. Currently, indexing also runs as a background process on the transactor. The transactor can be made highly available by running a second transactor as a standby. Since storage is shared truth, no dedicated connections are required between the active transactor and the standby. The transactor does not have any storage of its own, and does not service any read or query load. Thus, there is far less contention for it vs a monolithic model. That said, Datomic is not the right choice when you require unlimited write scalability.<\/p><ul><li>the Peer library<\/li>\n<\/ul><p>The database value code, connectivity code, memory index and query engine are provided in a JVM library called the 'peer' library, since application servers running this library are fully empowered equals. This library presents a Java API suitable for use by Java, Clojure, Scala, JRuby, Groovy and all other JVM languages. Each peer has direct (read-only) access to storage. The peer also contains the top-level cache of index segments, in which the segments have been fully expanded into JVM objects.<\/p><p>Programming at the peer level feels categorically different from client-server database programming. While peers are not full replicas, they quickly cache their working sets, and the database feels like an object in memory. This is not a veneer, as is provided by an ORM. Furthermore, query is local. Thus, rather than queries running within the context of a database, databases become arguments to queries. Queries can take more than one database, and\/or a combination of database and ordinary collections.<\/p><p>Datomic comes with a Datalog query engine. Datalog is an easy-to-learn, pattern oriented query language well suited to Datomic's data model and programmatic collections. It is set-oriented and relational, like SQL. There is a sense in which even the query engine is orthogonal though, as the direct index access provided supports the use of other query languages. For instance, Clojure users are able to use Clojure's Prolog-like core.logic library to query Datomic as well.<\/p><p>All of this application interaction happens through a data-driven API with an extremely small surface area.<\/p><ul><li>memcache<\/li>\n<\/ul><p>Datomic can automatically utilize a memcache cluster as second-tier cache between the in-process peer cache and storage. As described above, this is a cache 'under' query, of index segments. Because they are immutable, and are the source of answers, rather than specific answers, there is tremendous leverage in caching index segments, with none of the normal complexity of caching, i.e. cache coherence, invalidation etc. The transactor can be configured to prefill memcache as it writes to storage, so even the first reads will result in cache hits. This can add a level of scalability and performance not otherwise present in the base storage (e.g. if it is a single-instance SQL server). As with storage, Datomic doesn't need a dedicated memcache cluster. It uses UUID-based keys that will not conflict with other use of the same cache.<\/p><h2>The REST API and clients<\/h2><p>Datomic comes with a REST component that turns a process running the peer library into an HTTP service supporting the Datomic API. This can be used for non-JVM languages, or for lighter-weight clients that don't need the full in-process capabilities of the peer. The REST service is completely discoverable and self-describing via the text\/html media type. Applications will generally request the application\/edn media type, which supports data transmission in the <a href=\"http:\/\/edn-format.org\/\" target=\"_blank\">edn format<\/a>.<\/p><p>Clients of the REST API lose very few of the benefits of sharing the process with the database cache, and otherwise retain the important characteristics of the model - programming with data, queries can cross data sources, or incorporate program data, and independent queries can act upon the same basis, without any transaction, due to the fact that particular database values can have permalinks.<\/p><p>Clients pass and receive the exact same data structures as do peer API callers, with an isomorphic interface. This says less about the REST API than it does about the peer API - programming with data is a good idea, on the wire or off.<\/p><h2>Summary<\/h2><p>I hope this has given you a feel for the nature of Datomic's architecture and some of its details. If nothing else, it provides many examples of the architectural value of immutability. Databases as compositions of services, such as Datomic, are likely to become more common as we move away from the limitations of monolithic architectures while trying to retain their capabilities, and reach for new ones.<\/p><p>There are many more aspects of Datomic, including the information model, the reification of process, the database as a value, the query model and time model, all of which are synergistic with the architecture. You can learn more from the <a href=\"http:\/\/docs.datomic.com\/\" target=\"_blank\">Datomic documentation<\/a>.<\/p><h2>About the Author<\/h2><p><strong><img alt=\"\" src=\"https:\/\/www.infoq.com\/articles\/Architecture-Datomic\/articles\/Architecture-Datomic\/en\/resources\/richHickey.jpg\" style=\"margin-top: 4px; margin-bottom: 4px; margin-left: 4px; margin-right: 4px; float: left;\" \/>Rich Hickey<\/strong>, the author of Clojure and designer of Datomic, is a software developer with over 25 years of experience in various domains. Rich has worked on scheduling systems, broadcast automation, audio analysis and fingerprinting, database design, yield management, exit poll systems, and machine listening, in a variety of languages.<\/p>","created_at":"2019-09-12T04:31:05+0000","updated_at":"2019-09-12T04:31:13+0000","published_at":"2012-11-02T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"www.infoq.com","preview_picture":"https:\/\/res.infoq.com\/articles\/Architecture-Datomic\/en\/smallimage\/article-logo.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["web","webdev","rust","wasm"],"is_public":false,"id":1021,"title":"yewstack\/yew","url":"https:\/\/github.com\/yewstack\/yew","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2>\n<p><strong>Yew<\/strong> is a modern Rust framework inspired by Elm and React for\ncreating multi-threaded frontend apps with WebAssembly.<\/p>\n<p>The framework supports <em><strong>multi-threading &amp; concurrency<\/strong><\/em> out of the box.\nIt uses <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Web_Workers_API\" rel=\"nofollow\">Web Workers API<\/a> to spawn actors (agents) in separate threads\nand uses a local scheduler attached to a thread for concurrent tasks.<\/p>\n<p><a href=\"https:\/\/www.patreon.com\/deniskolodin\" rel=\"nofollow\">Become a sponsor on Patreon<\/a><\/p>\n<p><a href=\"https:\/\/yew-todomvc.netlify.com\/\" rel=\"nofollow\">Check out a live demo<\/a> powered by <a href=\"https:\/\/github.com\/yewstack\/yew-wasm-pack-template\"><code>yew-wasm-pack-template<\/code><\/a><\/p>\n<h2><a id=\"user-content-cutting-edge-technologies\" class=\"anchor\" aria-hidden=\"true\" href=\"#cutting-edge-technologies\"><\/a>Cutting Edge technologies<\/h2>\n<h3><a id=\"user-content-rust-to-wasm-compilation\" class=\"anchor\" aria-hidden=\"true\" href=\"#rust-to-wasm-compilation\"><\/a>Rust to WASM compilation<\/h3>\n<p>This framework is designed to be compiled into modern browsers' runtimes: wasm, asm.js, emscripten.<\/p>\n<p>To prepare the development environment use the installation instruction here: <a href=\"https:\/\/github.com\/raphamorim\/wasm-and-rust\">wasm-and-rust<\/a>.<\/p>\n<h3><a id=\"user-content-clean-mvc-approach-inspired-by-elm-and-redux\" class=\"anchor\" aria-hidden=\"true\" href=\"#clean-mvc-approach-inspired-by-elm-and-redux\"><\/a>Clean MVC approach inspired by Elm and Redux<\/h3>\n<p>Yew implements strict application state management based on message passing and updates:<\/p>\n<p><code>src\/main.rs<\/code><\/p>\n<div class=\"highlight highlight-source-rust\"><pre>use yew::{html, Component, ComponentLink, Html, Renderable, ShouldRender};\nstruct Model { }\nenum Msg {\n    DoIt,\n}\nimpl Component for Model {\n    \/\/ Some details omitted. Explore the examples to see more.\n    type Message = Msg;\n    type Properties = ();\n    fn create(_: Self::Properties, _: ComponentLink&lt;Self&gt;) -&gt; Self {\n        Model { }\n    }\n    fn update(&amp;mut self, msg: Self::Message) -&gt; ShouldRender {\n        match msg {\n            Msg::DoIt =&gt; {\n                \/\/ Update your model on events\n                true\n            }\n        }\n    }\n}\nimpl Renderable&lt;Model&gt; for Model {\n    fn view(&amp;self) -&gt; Html&lt;Self&gt; {\n        html! {\n            \/\/ Render your model here\n            &lt;button onclick=|_| Msg::DoIt&gt;{ \"Click me!\" }&lt;\/button&gt;\n        }\n    }\n}\nfn main() {\n    yew::start_app::&lt;Model&gt;();\n}<\/pre><\/div>\n<p>Predictable mutability and lifetimes (thanks Rust!) make it possible to reuse a single instance of the model\nwithout a need to create a fresh one on every update. It also helps to reduce memory allocations.<\/p>\n<h3><a id=\"user-content-jsx-like-templates-with-html-macro\" class=\"anchor\" aria-hidden=\"true\" href=\"#jsx-like-templates-with-html-macro\"><\/a>JSX-like templates with <code>html!<\/code> macro<\/h3>\n<p>Feel free to put pure Rust code into HTML tags with all the compiler and borrow checker's benefits.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>html! {\n    &lt;section class=\"todoapp\"&gt;\n        &lt;header class=\"header\"&gt;\n            &lt;h1&gt;{ \"todos\" }&lt;\/h1&gt;\n            { view_input(&amp;model) }\n        &lt;\/header&gt;\n        &lt;section class=\"main\"&gt;\n            &lt;input class=\"toggle-all\"\n                   type=\"checkbox\"\n                   checked=model.is_all_completed()\n                   onclick=|_| Msg::ToggleAll \/&gt;\n            { view_entries(&amp;model) }\n        &lt;\/section&gt;\n    &lt;\/section&gt;\n}<\/pre><\/div>\n<h3><a id=\"user-content-agents---actors-model-inspired-by-erlang-and-actix\" class=\"anchor\" aria-hidden=\"true\" href=\"#agents---actors-model-inspired-by-erlang-and-actix\"><\/a>Agents - actors model inspired by Erlang and Actix<\/h3>\n<p>Every <code>Component<\/code> can spawn an agent and attach to it.\nAgents are separate tasks that work concurrently.<\/p>\n<p>Create your worker\/agent (in <code>context.rs<\/code> for example):<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>use yew::worker::*;\nstruct Worker {\n    link: AgentLink&lt;Worker&gt;,\n}\n#[derive(Serialize, Deserialize, Debug)]\npub enum Request {\n    Question(String),\n}\n#[derive(Serialize, Deserialize, Debug)]\npub enum Response {\n    Answer(String),\n}\nimpl Agent for Worker {\n    \/\/ Available:\n    \/\/ - `Job` (one per bridge)\n    \/\/ - `Context` (shared in the same thread)\n    \/\/ - `Public` (separate thread).\n    type Reach = Context; \/\/ Spawn only one instance per thread (all components could reach this)\n    type Message = Msg;\n    type Input = Request;\n    type Output = Response;\n    \/\/ Create an instance with a link to agent's environment.\n    fn create(link: AgentLink&lt;Self&gt;) -&gt; Self {\n        Worker { link }\n    }\n    \/\/ Handle inner messages (of services of `send_back` callbacks)\n    fn update(&amp;mut self, msg: Self::Message) { \/* ... *\/ }\n    \/\/ Handle incoming messages from components of other agents.\n    fn handle(&amp;mut self, msg: Self::Input, who: HandlerId) {\n        match msg {\n            Request::Question(_) =&gt; {\n                self.link.response(who, Response::Answer(\"That's cool!\".into()));\n            },\n        }\n    }\n}<\/pre><\/div>\n<p>Build the bridge to an instance of this agent.\nIt spawns a worker automatically or reuses an existing one, depending on the type of the agent:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>struct Model {\n    context: Box&lt;Bridge&lt;context::Worker&gt;&gt;,\n}\nenum Msg {\n    ContextMsg(context::Response),\n}\nimpl Component for Model {\n    type Message = Msg;\n    type Properties = ();\n    fn create(_: Self::Properties, link: ComponentLink&lt;Self&gt;) -&gt; Self {\n        let callback = link.send_back(|_| Msg::ContextMsg);\n        \/\/ `Worker::bridge` spawns an instance if no one is available\n        let context = context::Worker::bridge(callback); \/\/ Connected! :tada:\n        Model { context }\n    }\n}<\/pre><\/div>\n<p>You can use as many agents as you want. For example you could separate all interactions\nwith a server to a separate thread (a real OS thread because Web Workers map to the native threads).<\/p>\n<blockquote>\n<p><strong>REMEMBER!<\/strong> Not every API is available for every environment. For example you can't use\n<code>StorageService<\/code> from a separate thread. It won't work with <code>Public<\/code> agents,\nonly with <code>Job<\/code> and <code>Context<\/code> ones.<\/p>\n<\/blockquote>\n<h3><a id=\"user-content-components\" class=\"anchor\" aria-hidden=\"true\" href=\"#components\"><\/a>Components<\/h3>\n<p>Yew supports components! You could create a new one by implementing a <code>Component<\/code> trait\nand including it directly into the <code>html!<\/code> template:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>html! {\n    &lt;nav class=\"menu\"&gt;\n        &lt;MyButton title=\"First Button\" \/&gt;\n        &lt;MyButton title=\"Second Button \"\/&gt;\n        &lt;MyList name=\"Grocery List\"&gt;\n          &lt;MyListItem text=\"Apples\" \/&gt;\n        &lt;\/MyList&gt;\n    &lt;\/nav&gt;\n}<\/pre><\/div>\n<h3><a id=\"user-content-scopes\" class=\"anchor\" aria-hidden=\"true\" href=\"#scopes\"><\/a>Scopes<\/h3>\n<p>Components live in an Angular-like scopes with <strong>parent-to-child<\/strong> <em>(properties)<\/em> and\n<strong>child-to-parent<\/strong> <em>(events)<\/em> interaction.<\/p>\n<p>Properties are also pure Rust types with strict type-checking during the compilation.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>\/\/ my_button.rs\n#[derive(Properties, PartialEq)]\npub struct Properties {\n    pub hidden: bool,\n    #[props(required)]\n    pub color: Color,\n    #[props(required)]\n    pub onclick: Callback&lt;()&gt;,\n}\n<\/pre><\/div>\n<div class=\"highlight highlight-source-rust\"><pre>\/\/ confirm_dialog.rs\nhtml! {\n    &lt;div class=\"confirm-dialog\"&gt;\n        &lt;MyButton onclick=|_| DialogMsg::Cancel color=Color::Red hidden=true \/&gt;\n        &lt;MyButton onclick=|_| DialogMsg::Submit color=Color::Blue \/&gt;\n    &lt;\/div&gt;\n}<\/pre><\/div>\n<h3><a id=\"user-content-fragments\" class=\"anchor\" aria-hidden=\"true\" href=\"#fragments\"><\/a>Fragments<\/h3>\n<p>Yew supports fragments: elements without a parent which could be attached somewhere later.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>html! {\n    &lt;&gt;\n        &lt;tr&gt;&lt;td&gt;{ \"Row\" }&lt;\/td&gt;&lt;\/tr&gt;\n        &lt;tr&gt;&lt;td&gt;{ \"Row\" }&lt;\/td&gt;&lt;\/tr&gt;\n        &lt;tr&gt;&lt;td&gt;{ \"Row\" }&lt;\/td&gt;&lt;\/tr&gt;\n    &lt;\/&gt;\n}<\/pre><\/div>\n<h3><a id=\"user-content-virtual-dom-independent-loops-fine-updates\" class=\"anchor\" aria-hidden=\"true\" href=\"#virtual-dom-independent-loops-fine-updates\"><\/a>Virtual DOM, independent loops, fine updates<\/h3>\n<p>Yew uses its own <strong>virtual-dom<\/strong> implementation. It updates the browser's DOM\nwith tiny patches when properties of elements have changed. Every component lives\nin its own independent loop interacting with the environment (<code>Scope<\/code>) through message passing\nand supports a fine control of rendering.<\/p>\n<p>The <code>ShouldRender<\/code> returns the value which informs the loop when the component should be re-rendered:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>fn update(&amp;mut self, msg: Self::Message) -&gt; ShouldRender {\n    match msg {\n        Msg::UpdateValue(value) =&gt; {\n            self.value = value;\n            true\n        }\n        Msg::Ignore =&gt; {\n            false\n        }\n    }\n}<\/pre><\/div>\n<p>Using <code>ShouldRender<\/code> is more effective than comparing the model after every update because not every model\nchange leads to a view update. It allows the framework to skip the model comparison checks entirely.\nThis also allows you to control updates as precisely as possible.<\/p>\n<h3><a id=\"user-content-rustjsc-style-comments-in-templates\" class=\"anchor\" aria-hidden=\"true\" href=\"#rustjsc-style-comments-in-templates\"><\/a>Rust\/JS\/C-style comments in templates<\/h3>\n<p>Use single-line or multi-line Rust comments inside html-templates.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>html! {\n    &lt;section&gt;\n   \/* Write some ideas\n    * in multiline comments\n    *\/\n    &lt;p&gt;{ \"and tags can be placed between comments!\" }&lt;\/p&gt;\n    \/\/ &lt;li&gt;{ \"or single-line comments\" }&lt;\/li&gt;\n    &lt;\/section&gt;\n}<\/pre><\/div>\n<h3><a id=\"user-content-third-party-crates-and-pure-rust-expressions-inside\" class=\"anchor\" aria-hidden=\"true\" href=\"#third-party-crates-and-pure-rust-expressions-inside\"><\/a>Third-party crates and pure Rust expressions inside<\/h3>\n<p>Use external crates and put values from them into the template:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>extern crate chrono;\nuse chrono::prelude::*;\nimpl Renderable&lt;Model&gt; for Model {\n    fn view(&amp;self) -&gt; Html&lt;Self&gt; {\n        html! {\n            &lt;p&gt;{ Local::now() }&lt;\/p&gt;\n        }\n    }\n}<\/pre><\/div>\n<blockquote>\n<p>Some crates don't support the true wasm target (<code>wasm32-unknown-unknown<\/code>) yet.<\/p>\n<\/blockquote>\n<h3><a id=\"user-content-services\" class=\"anchor\" aria-hidden=\"true\" href=\"#services\"><\/a>Services<\/h3>\n<p>Yew has implemented pluggable services that allow you to call external APIs, such as:\nJavaScript alerts, timeout, storage, fetches and websockets.\nIt's a handy alternative to subscriptions.<\/p>\n<p>Implemented:<\/p>\n<ul><li><code>IntervalService<\/code><\/li>\n<li><code>RenderService<\/code><\/li>\n<li><code>TimeoutService<\/code><\/li>\n<li><code>StorageService<\/code><\/li>\n<li><code>DialogService<\/code><\/li>\n<li><code>FetchService<\/code><\/li>\n<li><code>WebSocketService<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-rust\"><pre>use yew::services::{ConsoleService, TimeoutService};\nstruct Model {\n    link: ComponentLink&lt;Model&gt;,\n    console: ConsoleService,\n    timeout: TimeoutService,\n}\nimpl Component for Model {\n    fn update(&amp;mut self, msg: Self::Message) -&gt; ShouldRender {\n        match msg {\n            Msg::Fire =&gt; {\n                let send_msg = self.link.send_back(|_| Msg::Timeout);\n                self.timeout.spawn(Duration::from_secs(5), send_msg);\n            }\n            Msg::Timeout =&gt; {\n                self.console.log(\"Timeout!\");\n            }\n        }\n    }\n}<\/pre><\/div>\n<p>Can't find an essential service? Want to use a library from <code>npm<\/code>?\nYou can reuse <code>JavaScript<\/code> libraries with <code>stdweb<\/code> capabilities and create\nyour own service implementation. Here's an example below of how to wrap the\n<a href=\"https:\/\/www.npmjs.com\/package\/ccxt\" rel=\"nofollow\">ccxt<\/a> library:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>pub struct CcxtService(Option&lt;Value&gt;);\nimpl CcxtService {\n    pub fn new() -&gt; Self {\n        let lib = js! {\n            return ccxt;\n        };\n        CcxtService(Some(lib))\n    }\n    pub fn exchanges(&amp;mut self) -&gt; Vec&lt;String&gt; {\n        let lib = self.0.as_ref().expect(\"ccxt library object lost\");\n        let v: Value = js! {\n            var ccxt = @{lib};\n            console.log(ccxt.exchanges);\n            return ccxt.exchanges;\n        };\n        let v: Vec&lt;String&gt; = v.try_into().expect(\"can't extract exchanges\");\n        v\n    }\n    \/\/ Wrap more methods here!\n}<\/pre><\/div>\n<h3><a id=\"user-content-easy-to-use-data-conversion-and-destructuring\" class=\"anchor\" aria-hidden=\"true\" href=\"#easy-to-use-data-conversion-and-destructuring\"><\/a>Easy-to-use data conversion and destructuring<\/h3>\n<p>Yew allows for serialization (store\/send and restore\/receive) formats.<\/p>\n<p>Implemented: <code>JSON<\/code>, <code>TOML<\/code>, <code>YAML<\/code>, <code>MSGPACK<\/code>, <code>CBOR<\/code>.<\/p>\n<p>In development: <code>BSON<\/code>, <code>XML<\/code>.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>use yew::format::Json;\n#[derive(Serialize, Deserialize)]\nstruct Client {\n    first_name: String,\n    last_name: String,\n}\nstruct Model {\n    local_storage: StorageService,\n    clients: Vec&lt;Client&gt;,\n}\nimpl Component for Model {\n    fn update(&amp;mut self, msg: Self::Message) -&gt; ShouldRender {\n        Msg::Store =&gt; {\n            \/\/ Stores it, but in JSON format\/layout\n            self.local_storage.store(KEY, Json(&amp;model.clients));\n        }\n        Msg::Restore =&gt; {\n            \/\/ Tries to read and destructure it as JSON formatted data\n            if let Json(Ok(clients)) = self.local_storage.restore(KEY) {\n                model.clients = clients;\n            }\n        }\n    }\n}<\/pre><\/div>\n<p>Only <code>JSON<\/code> is available by default but you can activate the rest through features in\nyour project's <code>Cargo.toml<\/code>:<\/p>\n<div class=\"highlight highlight-source-toml\"><pre>[dependencies]\nyew = { git = \"https:\/\/github.com\/yewstack\/yew\", features = [\"toml\", \"yaml\", \"msgpack\", \"cbor\"] }<\/pre><\/div>\n<h2><a id=\"user-content-development-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#development-setup\"><\/a>Development setup<\/h2>\n<p>Clone or download this repository.<\/p>\n<h3><a id=\"user-content-install-cargo-web\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-cargo-web\"><\/a>Install <a href=\"https:\/\/github.com\/koute\/cargo-web\">cargo-web<\/a><\/h3>\n<p>This is an optional tool that simplifies deploying web applications:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>cargo install cargo-web<\/pre><\/div>\n<blockquote>\n<p>Add <code>--force<\/code> option to ensure you install the latest version.<\/p>\n<\/blockquote>\n<h3><a id=\"user-content-build\" class=\"anchor\" aria-hidden=\"true\" href=\"#build\"><\/a>Build<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>cargo web build\n# without cargo-web, only the wasm32-unknown-unknown target is supported\ncargo build --target wasm32-unknown-unknown<\/pre><\/div>\n<h3><a id=\"user-content-running-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-tests\"><\/a>Running Tests<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>.\/ci\/run_tests.sh<\/pre><\/div>\n<h3><a id=\"user-content-running-the-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-the-examples\"><\/a>Running the examples<\/h3>\n<p>There are many examples that show how the framework works:\n<a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/counter\">counter<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/crm\">crm<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/custom_components\">custom_components<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/dashboard\">dashboard<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/fragments\">fragments<\/a>,\n<a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/game_of_life\">game_of_life<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/mount_point\">mount_point<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/npm_and_rest\">npm_and_rest<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/timer\">timer<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/todomvc\">todomvc<\/a>, <a href=\"https:\/\/github.com\/yewstack\/yew\/blob\/master\/examples\/two_apps\">two_apps<\/a>.<\/p>\n<p>To start an example enter its directory and start it with <a href=\"https:\/\/github.com\/koute\/cargo-web\">cargo-web<\/a>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>cargo web start<\/pre><\/div>\n<p>To run an optimised build instead of a debug build use:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>cargo web start --release<\/pre><\/div>\n<p>This will use the <code>wasm32-unknown-unknown<\/code> target by default, which is Rust's native WebAssembly target.\nThe Emscripten-based <code>wasm32-unknown-emscripten<\/code> and <code>asmjs-unknown-emscripten<\/code> targets are also supported\nif you tell the <code>cargo-web<\/code> to build for them using the <code>--target<\/code> parameter.<\/p>\n<h2><a id=\"user-content-project-templates\" class=\"anchor\" aria-hidden=\"true\" href=\"#project-templates\"><\/a>Project templates<\/h2>\n<\/article>","created_at":"2019-09-11T18:01:34+0000","updated_at":"2019-09-11T18:01:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/49116234?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["python","logic-programming","minikanren"],"is_public":false,"id":1020,"title":"A Primer\u00b6","url":"https:\/\/microkanrenpy.readthedocs.io\/en\/latest\/primer.html","content":"<p>This short document aims to show the basic concepts underlying <em>\u03bckanren<\/em>,\nnamely relations as goals, and how to use primitive goals as building blocks to\nbuild and define relations. We start from two primitive relations toward\nrecursive relations, through constructors, fresh logic variables, unification\nand \u03b7-inversion.<\/p>\n<p>First of all we start importing objects and definitions of the logic system:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; from muk.core import *\n&gt;&gt;&gt; from muk.ext import *\n&gt;&gt;&gt; from muk.utils import *\n<\/pre><\/div>\n<\/div>\n<p>Moreover, our definitions lies on the following imports:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; from collections import defaultdict, Counter\n&gt;&gt;&gt; from functools import lru_cache\n<\/pre><\/div>\n<\/div>\n<p>We use the interface function <a class=\"reference internal\" href=\"https:\/\/microkanrenpy.readthedocs.io\/en\/latest\/under_the_hood.html#muk.core.run\" title=\"muk.core.run\"><code class=\"xref py py-func docutils literal\">run<\/code><\/a> to ask for\nsubstitutions that satisfy the relation under study; moreover, nouns <em>relation<\/em>,\n<em>goal<\/em> and <em>predicate<\/em> are equivalent notions and denotes the same concept in\nthe rest of this presentation.<\/p>\n<div class=\"section\" id=\"primitive-goals\">\n<h2>Primitive goals<a class=\"headerlink\" href=\"#primitive-goals\" title=\"Permalink to this headline\">\u00b6<\/a><\/h2>\n<p>The two primitive goals <a class=\"reference internal\" href=\"https:\/\/microkanrenpy.readthedocs.io\/en\/latest\/under_the_hood.html#muk.core.succeed\" title=\"muk.core.succeed\"><code class=\"xref py py-func docutils literal\">succeed<\/code><\/a> and\n<a class=\"reference internal\" href=\"https:\/\/microkanrenpy.readthedocs.io\/en\/latest\/under_the_hood.html#muk.core.fail\" title=\"muk.core.fail\"><code class=\"xref py py-func docutils literal\">fail<\/code><\/a> denote <em>truth<\/em> and <em>falsehood<\/em> in classic\nlogic, respectively.<\/p>\n<div class=\"section\" id=\"succeed\">\n<h3><code class=\"docutils literal\">succeed<\/code><a class=\"headerlink\" href=\"#succeed\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>Asking for all substitutions that satisfy the <code class=\"docutils literal\">succeed<\/code>\nrelation yields a tautology, namely any substitution works:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(succeed)\n[Tautology]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"fail\">\n<h3><code class=\"docutils literal\">fail<\/code><a class=\"headerlink\" href=\"#fail\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>On the other hand, there is no substitution that satisfies the <code class=\"docutils literal\">fail<\/code> relation:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fail)\n[]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"goal-ctors\">\n<h2>Goal ctors<a class=\"headerlink\" href=\"#goal-ctors\" title=\"Permalink to this headline\">\u00b6<\/a><\/h2>\n<p>Before introducing goal constructors we show how a logic variable is\nrepresented when it doesn\u2019t get any association:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; rvar(0)\n\u25a2\u2080\n<\/pre><\/div>\n<\/div>\n<p>where class <code class=\"docutils literal\">rvar<\/code> builds a <em>reified variable<\/em> providing an\ninteger, namely <code class=\"docutils literal\">0<\/code>, to track the variable\u2019s birthdate.<\/p>\n<div class=\"section\" id=\"fresh\">\n<h3><code class=\"docutils literal\">fresh<\/code><a class=\"headerlink\" href=\"#fresh\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>It is a goal ctor that introduces <em>new<\/em> logic variables, namely variables without\nassociations, according to the plugged-in lambda expression, which receives\nvars and returns a goal. Consider the following query where the variable <code class=\"docutils literal\">q<\/code>\nis introduced fresh:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: succeed))\n[\u25a2\u2080]\n<\/pre><\/div>\n<\/div>\n<p>since the returned goal is <code class=\"docutils literal\">succeed<\/code> and var <code class=\"docutils literal\">q<\/code> doesn\u2019t get any\nassociation, it remains fresh in the list of values, result of the whole\n<code class=\"docutils literal\">run<\/code> invocation, that satisfy the goal <code class=\"docutils literal\">fresh(lambda q: succeed)<\/code>.<\/p>\n<p>As particular case, it is possible to plugin a <em>thunk<\/em>, namely a lambda\nexpression without argument:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda: succeed))\n[Tautology]\n<\/pre><\/div>\n<\/div>\n<p>at a first this could look useless but it is of great help for the definition\nof <em>recursive relations<\/em> as we will see in later examples (it is an instance of\n\u03b7-inversion, formally).<\/p>\n<\/div>\n<div class=\"section\" id=\"unify\">\n<h3><code class=\"docutils literal\">unify<\/code><a class=\"headerlink\" href=\"#unify\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>It is a goal ctor that attempts to make two arbitrary objects equal, recording\nassociations when fresh variables appears in the nested structures under\nunification. Here we show two simple examples of unification, the first succeeds\nwhile the second doesn\u2019t:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(unify(3, 3))\n[Tautology]\n&gt;&gt;&gt; run(unify([1, 2, 3], [[1]]))\n[]\n<\/pre><\/div>\n<\/div>\n<p>On the other hand, things get interesting when fresh variables are mixed in:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: unify(3, q)))\n[3]\n&gt;&gt;&gt; run(fresh(lambda q: unify([1, 2, 3], [1] + q)))\n[[2, 3]]\n&gt;&gt;&gt; run(fresh(lambda q: unify([[2, 3], 1, 2, 3], [q, 1] + q)))\n[[2, 3]]\n<\/pre><\/div>\n<\/div>\n<p>When two fresh vars are unified it is said that they <em>share<\/em> or <em>co-refer<\/em>:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q, z: unify(q, z)))\n[\u25a2\u2080]\n&gt;&gt;&gt; run(fresh(lambda q, z: unify(q, z) &amp; unify(z, 3)),\n...     var_selector=lambda q, z: q)\n[3]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"disj\">\n<h3><code class=\"docutils literal\">disj<\/code><a class=\"headerlink\" href=\"#disj\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>It is a goal ctor that consumes two goals and returns a new goal that can be\nsatisfied when <em>either<\/em> the former <em>or<\/em> the latter goal can be satisfied:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(succeed | fail)\n[Tautology]\n&gt;&gt;&gt; run(fail | fresh(lambda q: unify(q, True)))\n[Tautology]\n&gt;&gt;&gt; run(fresh(lambda q: fail | fail))\n[]\n&gt;&gt;&gt; run(fresh(lambda q: unify(q, False) | unify(q, True)))\n[False, True]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"conj\">\n<h3><code class=\"docutils literal\">conj<\/code><a class=\"headerlink\" href=\"#conj\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>It is a goal ctor that consumes two goals and returns a new goal that can be\nsatisfied when <em>both<\/em> the former <em>and<\/em> the latter goal can be satisfied:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(succeed &amp; fail)\n[]\n&gt;&gt;&gt; run(fail &amp; fresh(lambda q: unify(q, True)))\n[]\n&gt;&gt;&gt; run(fresh(lambda q: unify(q, 3) &amp; succeed))\n[3]\n&gt;&gt;&gt; run(fresh(lambda q: unify(q, False) &amp; unify(q, True)))\n[]\n&gt;&gt;&gt; run(fresh(lambda q: fresh(lambda q: unify(q, False)) &amp;\n...                     unify(q, True)))\n[True]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"facts-and-recursive-relations\">\n<h2>Facts and recursive relations<a class=\"headerlink\" href=\"#facts-and-recursive-relations\" title=\"Permalink to this headline\">\u00b6<\/a><\/h2>\n<p>In order to represent <em>facts<\/em> we introduce the <code class=\"xref py py-obj docutils literal\">conde<\/code> goal ctor,\nwhich is defined as a combination of conjs and disjs and we show how to write\nrecursive relation, possibly satisfied by a countably infinite number of values.<\/p>\n<div class=\"section\" id=\"conde\">\n<h3><code class=\"docutils literal\">conde<\/code><a class=\"headerlink\" href=\"#conde\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>The following simple example resembles facts declaration in Prolog about the father and\ngrandfather toy example:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def father(p, s):\n...     return conde([unify(p, 'paul'), unify(s, 'jason')],\n...                  [unify(p, 'john'), unify(s, 'henry')],\n...                  [unify(p, 'jason'), unify(s, 'tom')],\n...                  [unify(p, 'peter'), unify(s, 'brian')],\n...                  [unify(p, 'tom'), unify(s, 'peter')])\n...\n&gt;&gt;&gt; def grand_father(g, s):\n...     return fresh(lambda p: father(g, p) &amp; father(p, s))\n...\n&gt;&gt;&gt; run(fresh(lambda rel, p, s: grand_father(p, s) &amp; unify([p, s], rel)))\n[['paul', 'tom'], ['jason', 'peter'], ['tom', 'brian']]\n<\/pre><\/div>\n<\/div>\n<p>Introducing guided search:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def recuro(e):\n...     return fresh(lambda a, b: conde([unify(e, ['x']), succeed],\n...                                     [unify(e, ['b']+a), recuro(a)],\n...                                     [unify(e, ['a']+b), recuro(b)]))\n...\n&gt;&gt;&gt; run(fresh(recuro), n=5)\n[['x'], ['b', 'x'], ['b', 'b', 'x'], ['b', 'b', 'b', 'x'], ['b', 'b', 'b', 'b', 'x']]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"inversion\">\n<h3>\u03b7-inversion<a class=\"headerlink\" href=\"#inversion\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>Let us define a relation that yields countably many 5 objects; in order to do\nthat, the usual solution is to write a recursive definition. However, we\nproceed step by step, adjusting and learning from the Python semantic of\nargument evaluation at <em>function-call time<\/em>.  Consider the following as initial\ndefinition:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def fives(x):\n...     return unify(5, x) | fives(x)\n...\n&gt;&gt;&gt; run(fresh(lambda x: fives(x)))\nTraceback (most recent call last):\n...\nRecursionError: maximum recursion depth exceeded while calling a Python object\n<\/pre><\/div>\n<\/div>\n<p>Exception <code class=\"docutils literal\">RecursionError<\/code> is raised because in the body of function <code class=\"docutils literal\">fives<\/code> it\nis required to evaluate <code class=\"docutils literal\">fives(x)<\/code> in order to return a <code class=\"docutils literal\">disj<\/code> object, but this is\nthe point from were we started, hence no progress for recursion.<\/p>\n<p>Keeping in mind the previous argument, why not wrapping the recursion on\n<code class=\"docutils literal\">fives(x)<\/code> inside a <code class=\"docutils literal\">fresh<\/code> ctor in order to refresh the var <code class=\"docutils literal\">x<\/code> at\neach invocation?<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def fives(x):\n...     return unify(5, x) | fresh(lambda x: fives(x))\n...\n&gt;&gt;&gt; run(fresh(lambda x: fives(x)))\nTraceback (most recent call last):\n...\nRecursionError: maximum recursion depth exceeded while calling a Python object\n<\/pre><\/div>\n<\/div>\n<p>Again the same exception as before, this time for a different reason, however:\nsince we ask for all associations that satisfy the <em>countably infinite<\/em>\nrelation <code class=\"docutils literal\">fives<\/code>, function <code class=\"docutils literal\">run<\/code> continue to look for such values which are\ninfinite, of course. So, select only the first 10 objects:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def fives(x):\n...     return unify(5, x) | fresh(lambda x: fives(x))\n...\n&gt;&gt;&gt; run(fresh(lambda x: fives(x)), n=10)\n[5, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080, \u25a2\u2080]\n<\/pre><\/div>\n<\/div>\n<p>Although not a list of 10 fives objects, it makes sense: the very first 5 gets\nassociated to the var <code class=\"docutils literal\">x<\/code> introduced by the goal provided to <code class=\"docutils literal\">run<\/code> by the\n<code class=\"docutils literal\">fresh<\/code> ctor, and this association is only one way to satisfy the <code class=\"docutils literal\">disj<\/code> in\nthe definition of relation <code class=\"docutils literal\">fives<\/code>. Looking for other associations that work,\nwe attempt to satisfy the second goal in the <code class=\"docutils literal\">disj<\/code>, namely <code class=\"docutils literal\">fresh(lambda x:\nfives(x))<\/code>: it introduces a new var <code class=\"docutils literal\">x<\/code>, different from the previous one,\nand then recurs, leaving the original var without association. Since associations shown\nin the output list refer to the very first var <code class=\"docutils literal\">x<\/code>, we get many <code class=\"docutils literal\">\u25a2\u2080<\/code> symbols\nwhich represent the absence of association, therefore <code class=\"docutils literal\">x<\/code> remains fresh.<\/p>\n<div class=\"section\" id=\"fives\">\n<h4><code class=\"docutils literal\">fives<\/code><a class=\"headerlink\" href=\"#fives\" title=\"Permalink to this headline\">\u00b6<\/a><\/h4>\n<p>One way to actually get a list of fives is to unify inside the inner <code class=\"docutils literal\">fresh<\/code>, as follows:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def fives(x):\n...     return unify(5, x) | fresh(lambda y: fives(y) &amp; unify(y, x))\n...\n&gt;&gt;&gt; run(fresh(lambda x: fives(x)), n=10)\n[5, 5, 5, 5, 5, 5, 5, 5, 5, 5]\n<\/pre><\/div>\n<\/div>\n<p>or to use <code class=\"docutils literal\">fresh<\/code> as <em>\u03b7-inversion<\/em> rule, as follows:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def fives(x):\n...     return unify(5, x) | fresh(lambda: fives(x))\n...\n&gt;&gt;&gt; run(fresh(lambda x: fives(x)), n=10)\n[5, 5, 5, 5, 5, 5, 5, 5, 5, 5]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"nats\">\n<h4><code class=\"docutils literal\">nats<\/code><a class=\"headerlink\" href=\"#nats\" title=\"Permalink to this headline\">\u00b6<\/a><\/h4>\n<p>Just for fun, using the previous trick and abstracting out the 5s,\nwe can generate the naturals, taking only the first 10 as follows:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def nats(x, n=0):\n...     return unify(n, x) | fresh(lambda: nats(x, n+1))\n...\n&gt;&gt;&gt; run(fresh(lambda x: nats(x)), n=10)\n[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"append\">\n<h4><code class=\"docutils literal\">append<\/code><a class=\"headerlink\" href=\"#append\" title=\"Permalink to this headline\">\u00b6<\/a><\/h4>\n<p>Here we want to define the relation <code class=\"docutils literal\">append(r, s, o)<\/code> which holds if\n<code class=\"docutils literal\">r + s == o<\/code> where <code class=\"docutils literal\">r, s, o<\/code> are both lists. First of all we need an\nhelper relation <code class=\"docutils literal\">nullo<\/code> which holds if <code class=\"docutils literal\">l == []<\/code>:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def nullo(l):\n...     return unify([], l)\n<\/pre><\/div>\n<\/div>\n<p>so it follows the recursive definition, as usual:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def append(r, s, out):\n...     def A(r, out):\n...         return conde([nullo(r), unify(s, out)],\n...                      else_clause=fresh(lambda a, d, res:\n...                                            unify([a]+d, r) &amp;\n...                                            unify([a]+res, out) &amp;\n...                                            fresh(lambda: A(d, res))))\n...     return A(r, out)\n<\/pre><\/div>\n<\/div>\n<p>Some examples follow:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: append([1,2,3], [4,5,6], q)))\n[[1, 2, 3, 4, 5, 6]]\n&gt;&gt;&gt; run(fresh(lambda l, q: append([1,2,3]+q, [4,5,6], l)), n=4) \n[[1, 2, 3, 4, 5, 6],\n [1, 2, 3, \u25a2\u2080, 4, 5, 6],\n [1, 2, 3, \u25a2\u2080, \u25a2\u2081, 4, 5, 6],\n [1, 2, 3, \u25a2\u2080, \u25a2\u2081, \u25a2\u2082, 4, 5, 6]]\n&gt;&gt;&gt; run(fresh(lambda r, x, y:\n...             append(x, y, ['cake', 'with', 'ice', 'd', 't']) &amp;\n...             unify([x, y], r))) \n[[[], ['cake', 'with', 'ice', 'd', 't']],\n [['cake'], ['with', 'ice', 'd', 't']],\n [['cake', 'with'], ['ice', 'd', 't']],\n [['cake', 'with', 'ice'], ['d', 't']],\n [['cake', 'with', 'ice', 'd'], ['t']],\n [['cake', 'with', 'ice', 'd', 't'], []]]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"combinatorics\">\n<h3>Combinatorics<a class=\"headerlink\" href=\"#combinatorics\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>It is possible to generate the set of <em>Dyck paths<\/em> as it follows:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def dycko(\u03b1):\n...     return conde([nullo(\u03b1), succeed],\n...                  else_clause=fresh(lambda \u03b2, \u03b3:\n...                                     append(['(']+\u03b2, [')']+\u03b3, \u03b1) @ (dycko(\u03b2) @ dycko(\u03b3))))\n...\n&gt;&gt;&gt; paths = run(fresh(lambda \u03b1: dycko(\u03b1)), n=80)\n&gt;&gt;&gt; D = defaultdict(list)\n&gt;&gt;&gt; for \u03b1 in map(lambda \u03b1: ''.join(\u03b1), paths):\n...     D[len(\u03b1)\/\/2].append(\u03b1)\n...\n&gt;&gt;&gt; dict(D) \n{0: [''],\n 1: ['()'],\n 2: ['()()', '(())'],\n 3: ['()()()', '(())()', '()(())', '(()())', '((()))'],\n 4: ['(())()()', '()()()()', '(())(())', '()(())()', '(()()())', '(()())()', '()()(())', '(()(()))', '((()))()', '()(()())', '((())())', '()((()))', '((()()))', '(((())))'],\n 5: ['(())()()()', '(())(())()', '(()()()())', '(()()())()', '(()())()()', '(())()(())', '()(())()()', '(()()(()))', '((()))()()', '(())(()())', '()()()()()', '(()(()))()', '(()())(())', '((())()())', '((()))(())', '()(())(())', '(())((()))', '()()(())()', '((())())()', '()(()()())', '(()(())())', '()(()())()'],\n 6: ['(()()()()())', '(()()()())()', '(())(())()()', '(()()()(()))', '(()()())()()', '(())()()()()', '(()()(()))()', '(()())()()()', '((())()()())', '((()))()()()', '(())(())(())', '(()(()))()()', '(()())(())()', '(())()(())()', '((())()())()', '(()()())(())', '((()))(())()', '(())(()()())', '()(())()()()', '(()(())()())', '(()())()(())'],\n 7: ['(()()()()()())', '(()()()()())()', '(()()()()(()))', '(()()()())()()', '(()()()(()))()', '((())()()()())', '(()()(()))()()'],\n 8: ['(()()()()()()())', '(()()()()()())()', '(()()()()()(()))', '(()()()()())()()', '(()()()()(()))()'],\n 9: ['(()()()()()()()())', '(()()()()()()())()']}\n<\/pre><\/div>\n<\/div>\n<p>Fibonacci numbers as tailing problem:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def fibo(\u03b1):\n...     return conde([nullo(\u03b1), succeed],\n...                  else_clause=fresh(lambda \u03b2: condi([unify(['v']+\u03b2, \u03b1), fibo(\u03b2)],\n...                                                    [unify(['hh']+\u03b2, \u03b1), fibo(\u03b2)])))\n...\n&gt;&gt;&gt; strips = run(fresh(lambda \u03b1: fibo(\u03b1)), n=80)\n&gt;&gt;&gt; D = defaultdict(list)\n&gt;&gt;&gt; for \u03b1 in map(lambda \u03b1: ''.join(\u03b1), strips):\n...     D[len(\u03b1)].append(\u03b1)\n...\n&gt;&gt;&gt; dict(D) \n{0: [''],\n 1: ['v'],\n 2: ['hh', 'vv'],\n 3: ['hhv', 'vhh', 'vvv'],\n 4: ['hhhh', 'hhvv', 'vhhv', 'vvhh', 'vvvv'],\n 5: ['hhhhv', 'hhvhh', 'vhhhh', 'hhvvv', 'vhhvv', 'vvhhv', 'vvvhh', 'vvvvv'],\n 6: ['hhhhhh', 'hhhhvv', 'hhvhhv', 'vhhhhv', 'hhvvhh', 'vhhvhh', 'vvhhhh', 'hhvvvv', 'vhhvvv', 'vvhhvv', 'vvvhhv', 'vvvvhh', 'vvvvvv'],\n 7: ['hhhhhhv', 'hhhhvhh', 'hhvhhhh', 'vhhhhhh', 'hhhhvvv', 'hhvhhvv', 'vhhhhvv', 'hhvvhhv', 'vhhvhhv', 'vvhhhhv', 'hhvvvhh', 'vhhvvhh', 'vvhhvhh', 'vvvhhhh', 'hhvvvvv', 'vhhvvvv', 'vvhhvvv', 'vvvhhvv', 'vvvvhhv'],\n 8: ['hhhhhhhh', 'hhhhhhvv', 'hhhhvhhv', 'hhvhhhhv', 'vhhhhhhv', 'hhhhvvhh', 'hhvhhvhh', 'vhhhhvhh', 'hhvvhhhh', 'vhhvhhhh', 'vvhhhhhh', 'hhhhvvvv', 'hhvhhvvv', 'vhhhhvvv', 'hhvvhhvv', 'vhhvhhvv', 'vvhhhhvv'],\n 9: ['hhhhhhhhv', 'hhhhhhvhh', 'hhhhvhhhh', 'hhvhhhhhh', 'vhhhhhhhh', 'hhhhhhvvv', 'hhhhvhhvv', 'hhvhhhhvv', 'vhhhhhhvv'],\n 10: ['hhhhhhhhhh', 'hhhhhhhhvv']}\n&gt;&gt;&gt; run(fibo(['v', 'h', 'v']))\n[]\n<\/pre><\/div>\n<\/div>\n<p>Unfoldings within the Pascal triangle:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; from sympy import IndexedBase, symbols, latex\n&gt;&gt;&gt; P = IndexedBase('P')\n&gt;&gt;&gt; n, m = symbols('n m')\n&gt;&gt;&gt; def pascalo(depth, r, c, \u03b1):\n...     if not depth: return unify([P[r,c]], \u03b1)\n...     return fresh(lambda \u03b2, \u03b3: (pascalo(depth-1, r-1, c-1, \u03b2) @\n...                                pascalo(depth-1, r-1, c, \u03b3) @\n...                                append(\u03b2, \u03b3, \u03b1)))\n...\n&gt;&gt;&gt; unfoldings = [run(fresh(lambda \u03b1: pascalo(d, n, m, \u03b1))) for d in range(1, 6)]\n&gt;&gt;&gt; list(map(lambda unfold: sum(unfold[0]), unfoldings)) \n[P[n - 1, m - 1] + P[n - 1, m],\n 2*P[n - 2, m - 1] + P[n - 2, m - 2] + P[n - 2, m],\n 3*P[n - 3, m - 1] + 3*P[n - 3, m - 2] + P[n - 3, m - 3] + P[n - 3, m],\n 4*P[n - 4, m - 1] + 6*P[n - 4, m - 2] + 4*P[n - 4, m - 3] + P[n - 4, m - 4] + P[n - 4, m],\n 5*P[n - 5, m - 1] + 10*P[n - 5, m - 2] + 10*P[n - 5, m - 3] + 5*P[n - 5, m - 4] + P[n - 5, m - 5] + P[n - 5, m]]\n<\/pre><\/div>\n<\/div>\n<p>Fibonacci rabbits, again:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; f = IndexedBase('f')\n&gt;&gt;&gt; def rabbitso(depth, r, \u03b1):\n...     if not depth: return unify([f[r]], \u03b1)\n...     return fresh(lambda \u03b2, \u03b3: (rabbitso(depth-1, r-1, \u03b2) @\n...                                rabbitso(depth-1, r-2, \u03b3) @\n...                                append(\u03b2, \u03b3, \u03b1)))\n...\n&gt;&gt;&gt; unfoldings = [run(fresh(lambda \u03b1: rabbitso(d, n, \u03b1))) for d in range(1, 6)]\n&gt;&gt;&gt; list(map(lambda unfold: sum(unfold[0]), unfoldings)) \n[f[n - 1] + f[n - 2],\n f[n - 2] + 2*f[n - 3] + f[n - 4],\n f[n - 3] + 3*f[n - 4] + 3*f[n - 5] + f[n - 6],\n f[n - 4] + 4*f[n - 5] + 6*f[n - 6] + 4*f[n - 7] + f[n - 8],\n f[n - 10] + f[n - 5] + 5*f[n - 6] + 10*f[n - 7] + 10*f[n - 8] + 5*f[n - 9]]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"interleaving\">\n<h3>Interleaving<a class=\"headerlink\" href=\"#interleaving\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>Let us now define an operator <code class=\"docutils literal\">repeat<\/code> which makes relations that succeed by\nunifying the curried variable with some plugged in object, infinitely many\ntimes:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def repeat(r):\n...     def R(x):\n...         return unify(x, r) | fresh(lambda: R(x))\n...     return R\n...\n<\/pre><\/div>\n<\/div>\n<p>consequently, use it to define four streams of numbers:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; ones = repeat(1)\n&gt;&gt;&gt; twos = repeat(2)\n&gt;&gt;&gt; threes = repeat(3)\n&gt;&gt;&gt; fours = repeat(4)\n<\/pre><\/div>\n<\/div>\n<p>In the following example we use <code class=\"docutils literal\">conde<\/code>, which yields solutions of a cond\nline staying there when that line supplies multiple satisfying states, here\ninfinite when it meets <code class=\"docutils literal\">fives<\/code>, by the way:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def facts(x):\n...     return conde([unify(x, 'hello'), succeed],\n...                  [unify(x, 'pluto'), succeed],\n...                  [fives(x), succeed],\n...                  [unify(x, 'topolino'), succeed],\n...                  else_clause=unify(x, 'paperone'))\n...\n&gt;&gt;&gt; run(fresh(lambda x: facts(x)), n=10)\n['hello', 'pluto', 5, 5, 5, 5, 5, 5, 5, 5]\n&gt;&gt;&gt; groups_with_positions_notation(_)\nhello\u2080\npluto\u2081\n5\u2082, 5\u2083, 5\u2084, 5\u2085, 5\u2086, 5\u2087, 5\u2088, 5\u2089\n<\/pre><\/div>\n<\/div>\n<p>On the other hand, <code class=\"docutils literal\">condi<\/code> allows us to explore the solutions space by interleaving:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def facts(x):\n...     return condi([unify(x, 'hello'), succeed],\n...                  [unify(x, 'pluto'), succeed],\n...                  [fives(x), succeed],\n...                  [ones(x), succeed],\n...                  else_clause=unify(x, 'paperone'))\n...\n&gt;&gt;&gt; run(fresh(lambda x: facts(x)), n=15)\n['hello', 'pluto', 5, 1, 5, 'paperone', 1, 5, 1, 5, 1, 5, 1, 5, 1]\n&gt;&gt;&gt; groups_with_positions_notation(_)\nhello\u2080\npluto\u2081\n5\u2082, 5\u2084, 5\u2087, 5\u2089, 5\u2081\u2081, 5\u2081\u2083\n1\u2083, 1\u2086, 1\u2088, 1\u2081\u2080, 1\u2081\u2082, 1\u2081\u2084\npaperone\u2085\n<\/pre><\/div>\n<\/div>\n<p>Now, use only streams of numbers defined before and ask for the first 20\nassociations that satisfy their <code class=\"docutils literal\">condi<\/code> combination, providing the keyword argument <code class=\"docutils literal\">dovetail=False<\/code>\nwhich <em>associates streams on the right<\/em>:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: condi([succeed, ones(q)],\n...                           [succeed, twos(q)],\n...                           [succeed, threes(q)],\n...                           [succeed, fours(q)],\n...                           dovetail=False)), n=20)\n...\n[1, 2, 1, 3, 1, 2, 1, 4, 1, 2, 1, 3, 1, 2, 1, 4, 1, 2, 1, 3]\n&gt;&gt;&gt; groups_with_positions_notation(_)\n1\u2080, 1\u2082, 1\u2084, 1\u2086, 1\u2088, 1\u2081\u2080, 1\u2081\u2082, 1\u2081\u2084, 1\u2081\u2086, 1\u2081\u2088\n2\u2081, 2\u2085, 2\u2089, 2\u2081\u2083, 2\u2081\u2087\n3\u2083, 3\u2081\u2081, 3\u2081\u2089\n4\u2087, 4\u2081\u2085\n<\/pre><\/div>\n<\/div>\n<p>on the contrary, using <code class=\"docutils literal\">dovetail=True<\/code> allows us to get a more balanced\ninterleaving (this is the default behaviour):<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: condi([succeed, ones(q)],\n...                           [succeed, twos(q)],\n...                           [succeed, threes(q)],\n...                           [succeed, fours(q)])), n=20)\n...\n[1, 2, 1, 3, 2, 1, 4, 3, 2, 1, 4, 3, 2, 1, 4, 3, 2, 1, 4, 3]\n&gt;&gt;&gt; groups_with_positions_notation(_)\n1\u2080, 1\u2082, 1\u2085, 1\u2089, 1\u2081\u2083, 1\u2081\u2087\n2\u2081, 2\u2084, 2\u2088, 2\u2081\u2082, 2\u2081\u2086\n3\u2083, 3\u2087, 3\u2081\u2081, 3\u2081\u2085, 3\u2081\u2089\n4\u2086, 4\u2081\u2080, 4\u2081\u2084, 4\u2081\u2088\n<\/pre><\/div>\n<\/div>\n<p>Previous expression is equivalent to combining streams with <code class=\"docutils literal\">disj<\/code> getting\nthe same behavior of the last but one using the same keyword argument:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: disj(ones(q), twos(q), threes(q), fours(q),\n...                          dovetail=False)), n=20)\n[1, 2, 1, 3, 1, 2, 1, 4, 1, 2, 1, 3, 1, 2, 1, 4, 1, 2, 1, 3]\n&gt;&gt;&gt; groups_with_positions_notation(_)\n1\u2080, 1\u2082, 1\u2084, 1\u2086, 1\u2088, 1\u2081\u2080, 1\u2081\u2082, 1\u2081\u2084, 1\u2081\u2086, 1\u2081\u2088\n2\u2081, 2\u2085, 2\u2089, 2\u2081\u2083, 2\u2081\u2087\n3\u2083, 3\u2081\u2081, 3\u2081\u2089\n4\u2087, 4\u2081\u2085\n<\/pre><\/div>\n<\/div>\n<p>Finally, it is possible to request a <em>fair enumeration<\/em> by <em>dovetail<\/em> strategy,\nsetting the keyword argument <code class=\"docutils literal\">dovetail=True<\/code>, which is the default behavior:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: disj(ones(q), twos(q), threes(q), fours(q))), n=20)\n[1, 2, 1, 3, 2, 1, 4, 3, 2, 1, 4, 3, 2, 1, 4, 3, 2, 1, 4, 3]\n&gt;&gt;&gt; groups_with_positions_notation(_)\n1\u2080, 1\u2082, 1\u2085, 1\u2089, 1\u2081\u2083, 1\u2081\u2087\n2\u2081, 2\u2084, 2\u2088, 2\u2081\u2082, 2\u2081\u2086\n3\u2083, 3\u2087, 3\u2081\u2081, 3\u2081\u2085, 3\u2081\u2089\n4\u2086, 4\u2081\u2080, 4\u2081\u2084, 4\u2081\u2088\n<\/pre><\/div>\n<\/div>\n<p>As the last but one example, combining streams with the binary operator <code class=\"docutils literal\">|<\/code>\nyields a yet different result because the <em>disjunction binary operator\nassociates on the left<\/em>:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda q: ones(q) | twos(q) | threes(q) | fours(q)), n=20)\n[1, 4, 3, 4, 2, 4, 3, 4, 1, 4, 3, 4, 2, 4, 3, 4, 1, 4, 3, 4]\n&gt;&gt;&gt; groups_with_positions_notation(_)\n1\u2080, 1\u2088, 1\u2081\u2086\n4\u2081, 4\u2083, 4\u2085, 4\u2087, 4\u2089, 4\u2081\u2081, 4\u2081\u2083, 4\u2081\u2085, 4\u2081\u2087, 4\u2081\u2089\n3\u2082, 3\u2086, 3\u2081\u2080, 3\u2081\u2084, 3\u2081\u2088\n2\u2084, 2\u2081\u2082\n<\/pre><\/div>\n<\/div>\n<\/div>\n<div class=\"section\" id=\"difference-structures\">\n<h3>Difference structures<a class=\"headerlink\" href=\"#difference-structures\" title=\"Permalink to this headline\">\u00b6<\/a><\/h3>\n<p>The following implementation of <code class=\"docutils literal\">appendo<\/code> uses <em>difference lists<\/em>:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; def appendo(X, Y, XY):\n...     return fresh(lambda \u03b1, \u03b2, \u03b3: unify(X, \u03b1-\u03b2) &amp;\n...                                  unify(Y, \u03b2-\u03b3) &amp;\n...                                  unify(XY, \u03b1-\u03b3))\n...\n<\/pre><\/div>\n<\/div>\n<p>and can be used as follows:<\/p>\n<div class=\"highlight-default\"><div class=\"highlight\"><pre>&gt;&gt;&gt; run(fresh(lambda \u03b1\u03b2, \u03b1, \u03b2: appendo(([1,2,3]+\u03b1)-\u03b1, ([4,5,6]+\u03b2)-\u03b2, \u03b1\u03b2)))\n[([1, 2, 3, 4, 5, 6] + \u25a2\u2080) - \u25a2\u2080]\n&gt;&gt;&gt; run(fresh(lambda out, X, Y, \u03b1, \u03b2:\n...         appendo(([1,2,3]+\u03b1)-\u03b1, ([4,5,6]+\u03b2)-\u03b2, X-Y) &amp;\n...         unify([X, Y], out)))\n[[[1, 2, 3, 4, 5, 6] + \u25a2\u2080, \u25a2\u2080]]\n<\/pre><\/div>\n<\/div>\n<\/div>\n<\/div>","created_at":"2019-09-10T22:43:19+0000","updated_at":"2019-09-10T22:43:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"microkanrenpy.readthedocs.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1019,"title":"ledger, a powerful command-line accounting system","url":"https:\/\/www.ledger-cli.org","content":"<p><b>Tweets (#ledgercli OR #plaintextaccounting):<\/b><\/p><p>Ledger is a powerful, double-entry accounting system that is accessed from\nthe UNIX command-line. Ledger, begun in 2003, is written by John Wiegley\nand released under the BSD license.  It has also inspired several\n<a href=\"https:\/\/github.com\/ledger\/ledger\/wiki\/Ports\">ports<\/a> to other languages.<\/p><p>To get started with Ledger, add transactions to a text file in Ledger\u2019s\nown textual format. Here\u2019s what a single transaction might look like:<\/p><div class=\"highlight\"><pre>2015\/10\/12 Exxon\n    Expenses:Auto:Gas         $10.00\n    Liabilities:MasterCard   $-10.00\n<\/pre><\/div><p>You use the <code>ledger<\/code> command line program to see the balance of your accounts:<\/p><div class=\"highlight\"><pre>$ ledger -f your-file.dat balance\n              $10.00  Expenses:Auto:Gas\n             $-10.00  Liabilities:MasterCard\n--------------------\n                   0\n<\/pre><\/div><p>or the history of a particular account:<\/p><div class=\"highlight\"><pre>$ ledger -f your-file.dat register Expenses\n15-Oct-12 Exxon               Expenses:Auto:Gas          $10.00      $10.00\n<\/pre><\/div><p>Read more about Ledger\u2019s <a href=\"https:\/\/www.ledger-cli.org\/features.html\">features<\/a>.<\/p>","created_at":"2019-09-10T11:21:19+0000","updated_at":"2019-09-10T11:21:19+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.ledger-cli.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["prolog","logic-programming","minikanren"],"is_public":false,"id":1018,"title":"What are the main technical differences between Prolog and miniKanren, with respect to logic programming?","url":"https:\/\/stackoverflow.com\/questions\/28467011\/what-are-the-main-technical-differences-between-prolog-and-minikanren-with-resp","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>When I want to read up on logic programming I always stumble over two \"main\" ways to do it nowadays:<\/p>\n<ul><li><strong>miniKanren<\/strong>, a minilanguage introduced in <a href=\"https:\/\/mitpress.mit.edu\/books\/reasoned-schemer\" rel=\"noreferrer\">The Reasoned Schemer<\/a> and popular at the moment due to <a href=\"https:\/\/github.com\/clojure\/core.logic\" rel=\"noreferrer\">core.logic<\/a>.<\/li>\n<li><strong>Prolog<\/strong>, the first \"big\" logic programming language.<\/li>\n<\/ul><p>What I'm interested in now: What are the principal technical differences between the two? Are they very similar in approach and implementation, or do they take completely different approaches to logic programming? Which branches of mathematics do they come from, and what are the theoretical foundations?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/28467011\/revisions\" title=\"show all edits to this post\">edited Feb 18 '15 at 4:09<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/415518\/lindsey-kuper\">Lindsey Kuper<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            68144 silver badges2020 bronze badges&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Feb 12 '15 at 0:04&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1382925\/profpatsch\">Profpatsch<\/a>Profpatsch&#13;\n        <div class=\"-flair\">&#13;\n            2,18355 gold badges2121 silver badges3030 bronze badges&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"2\">&#13;\n                                    2 Answers&#13;\n                                2&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>First, allow me to compliment you on your fine pw0n1e icon.<\/p>\n<p>This is a tricky question to answer, largely because there are so many variants of both miniKanren and Prolog.  miniKanren and Prolog are really families of languages, which makes it difficult to compare their features, or even how they are used in practice.  Because of this, please take everything I'm about to say with caution: if I say that Prolog uses depth-first search, be aware that many Prolog implementations support other search strategies, and that alternate search strategies can also be encoded at the meta-interpreter level.  Still, miniKanren and Prolog have different design philosophies, and make different trade-offs.<\/p>\n<p>Prolog is one of the two classic languages for symbolic artificial intelligence programming (the other classic language being Lisp).  Prolog excels at implementing symbolic rule-based systems in which declarative knowledge is encoded in first-order logic.  The language is optimized for expressiveness and efficiency for these types of applications, sometimes at the expense of logical purity.  For example, by default Prolog does not use the \"occur check\" in unification.  From a math\/logic standpoint, this version of unification is incorrect.  However, the occur check is expensive, and in most cases the lack of the occur check is not a problem.  This is a very pragmatic design decision, as is Prolog's use of depth-first search, and use of cut (<code>!<\/code>) to control backtracking.  I'm sure these decisions were absolutely necessary when running on the hardware of the 1970s, and today are very useful when working on large problems, and when dealing with huge (often infinite!) search spaces.<\/p>\n<p>Prolog supports many \"extra-logical\" or \"non-logical\" features, including cut, <code>assert<\/code> and <code>retract<\/code>, projection of variables for arithmetic using <code>is<\/code>, and so forth.  Many of these features make it easier to express complex control flow, and to manipulate Prolog's global database of facts.  One very interesting feature of Prolog is that Prolog code is itself stored in the global database of facts, and can be queried against at run time.  This makes it trivial to write meta-interpreters that modify the behavior of Prolog code under interpretation.  For example, it is possible to encode breadth-first search in Prolog using a meta-interpreter that changes the search order.  This is an extremely powerful technique that is not well known outside of the Prolog world.  'The Art of Prolog' describes this technique in detail.<\/p>\n<p>Tremendous effort has gone into improving Prolog implementations, most of which are based on the Warren Abstract Machine (WAM).  The WAM uses a side-effecting model in which values are destructively assigned to logic variables, with these side-effects being undone upon backtracking.  Many features can be added to Prolog by extending the instructions of the WAM.  One disadvantage of this approach is that Prolog implementation papers can be difficult to read without a solid understanding of the WAM.  On the other hand, Prolog implementer have a common model for discussing implementation issues.  There has been a great deal of research in parallel Prolog, culminating in Andorra Prolog in the 1990s.  At least some of these ideas live on in Ciao Prolog.  (Ciao Prolog is full of interesting ideas, many of which go far beyond the Prolog standard.)<\/p>\n<p>Prolog has a beautiful unification-based \"pattern-matching\"-style syntax that results in very succinct programs.  Prologers love their syntax, just like Lispers love their s-expressions.  Prolog also has a large library of standard predicates.  Due to all of the engineering that has gone into making the WAM fast, there are very capable and mature Prolog implementations.  As a result, many large knowledge-based systems have been written entirely in Prolog.<\/p>\n<p>miniKanren was designed as a minimal logic programming language, with a small, easily understandable, and easily hackable implementation.  miniKanren was originally embedded in Scheme, and has been ported to dozens of other host languages over the past decade.  The most popular miniKanren implementation is 'core.logic' in Clojure, which now has many Prolog-like extensions and a number of optimizations.  Recently the core of the miniKanren implementation has been simplified even further, resulting in a tiny \"micro kernel\" called \"microKanren.\"  miniKanren can then be implemented on top of this microKanren core.  Porting microKanren or miniKanren to a new host language has become a standard exercise for programmers learning miniKanren.  As a result, most popular high-level languages have at least one miniKanren or microKanren implementation.<\/p>\n<p>The standard implementations of miniKanren and microKanren contain no mutation or other side-effects, with a single exception:  some versions of miniKanren use pointer equality for comparison of logic variables.  I consider this a \"benign effect,\" although many implementations avoid even this effect by passing a counter through the implementation.  There is also no global fact database.  miniKanren's implementation philosophy is inspired by functional programming:  mutation and effects should be avoided, and all language constructs should respect lexical scope.  If you look carefully at the implementation you might even spot a couple of monads.  The search implementation is based on combining and manipulating lazy streams, once again without using mutation.  These implementation choices lead to very different trade-offs than in Prolog.  In Prolog, variable lookup is constant time, but backtracking requires undoing side-effects.  In miniKanren variable lookup is more expensive, but backtracking is \"free.\"  In fact, there is no backtracking in miniKanren, due to how the streams are handled.<\/p>\n<p>One interesting aspect of the miniKanren implementation is that the code is inherently thread-safe and---at least in theory---trivially parallelizable.  Of course, parallelizing the code without making it <em>slower<\/em> is not trivial, given that each thread or process must be given enough work to make up for the overhead of parallelization.  Still, this is an area of miniKanren implementation that I hope will receive more attention and experimentation.<\/p>\n<p>miniKanren uses the occur check for unification, and uses a complete interleaving search instead of depth-first search.  Interleaving search uses more memory than depth-first search, but can find answers in some cases in which depth-first search will diverge\/loop forever.  miniKanren <em>does<\/em> support a few extra-logical operators---<code>conda<\/code>, <code>condu<\/code>, and <code>project<\/code>, for example.  <code>conda<\/code> and <code>condu<\/code> can be used to simulate Prolog's cut, and <code>project<\/code> can be used to get the value associated with a logic variable.<\/p>\n<p>The presence of <code>conda<\/code>, <code>condu<\/code>, and <code>project<\/code>---and the ability to easily modify the search strategy---allows programmers to use miniKanren as an embedded Prolog-like language.  This is especially true for users of Clojure's 'core.logic', which includes many Prolog-like extensions.  This \"pragmatic\" use of miniKanren seems to account for the majority of miniKanren's use in industry.  Programmers who want to add a knowledge-based reasoning system to an existing application written in Clojure or Python or JavaScript are generally not interested in rewriting their entire application in Prolog.  Embedding a small logic programming language in Clojure or Python is much more appealing.  An embedded Prolog implementation would work just as well for this purpose, presumably.  I suspect miniKanren has become popular as an embedded logic language because of the tiny and pure core implementation, along with the talks, blog posts, tutorials, and other educational materials that have come out since 'The Reasoned Schemer' was published.<\/p>\n<p>In addition to the use of miniKanren as a pragmatic embedded logic programming language similar in spirit to Prolog, miniKanren is being used for research in \"relational\" programming.  That is, in writing programs that behave as mathematical relations rather than mathematical functions.  For example, in Scheme the <code>append<\/code> function can append two lists, returning a new list: the function call <code>(append '(a b c) '(d e))<\/code> returns the list <code>(a b c d e)<\/code>.  We can, however, also treat <code>append<\/code> as a three-place relation rather than as a two-argument function.  The call <code>(appendo '(a b c) '(d e) Z)<\/code> would then associate the logic variable <code>Z<\/code> with the list <code>(a b c d e)<\/code>.  Of course things get more interesting when we place logic variables in other positions.  The call <code>(appendo X '(d e) '(a b c d e))<\/code> associates <code>X<\/code> with <code>(a b c)<\/code>, while the call <code>(appendo X Y '(a b c d e))<\/code> associates <code>X<\/code> and <code>Y<\/code> with pairs of lists that, when appended, are equal to <code>(a b c d e)<\/code>.  For example <code>X<\/code> = <code>(a b)<\/code> and <code>Y<\/code> = <code>(c d e)<\/code> are one such pair of values.  We can also write <code>(appendo X Y Z)<\/code>, which will produce infinitely many triples of lists <code>X<\/code>, <code>Y<\/code>, and <code>Z<\/code> such that appending <code>X<\/code> to <code>Y<\/code> produces <code>Z<\/code>.<\/p>\n<p>This relational version of <code>append<\/code> can be easily expressed in Prolog, and indeed is shown in many Prolog tutorials.  In practice, more complex Prolog programs tend to use at least a few extra-logical features, such as cut, which inhibit the ability to treat the resulting program as a relation.  In contrast, miniKanren is explicitly designed to support this style of relational programming.  More recent versions of miniKanren have support for symbolic constraint solving (<code>symbolo<\/code>, <code>numbero<\/code>, <code>absento<\/code>, disequality constraints, nominal logic programming) to make it easier to write non-trivial programs as relations.  In practice I never use any of the extra-logical features of miniKanren, and I write all of my miniKanren programs as relations.  The most interesting relational programs are the relational interpreters for a subset of Scheme.  These interpreters have many interesting abilities, such as generating a million Scheme programs that evaluate to the list <code>(I love you)<\/code>, or trivially generating quines (programs that evaluate to themselves).<\/p>\n<p>miniKanren makes a number of trade-offs to enable this relational style of programming, which are very different from the trade-offs Prolog makes.  Over time miniKanren has added more symbolic constraints, really becoming a symbolically-oriented Constraint Logic Programming language.  In many cases these symbolic constraints make it practical to avoid using extra-logical operators like <code>condu<\/code> and <code>project<\/code>.  In other cases, these symbolic constraints are not sufficient.  Better support for symbolic constraints is one active area of miniKanren research, along with the broader question of how to write larger and more complex programs as relations.<\/p>\n<p>In short, both miniKanren and Prolog have interesting features, implementations, and uses, and I think it is worth learning the ideas from both languages.  There are other very interesting logic programming languages as well, such as Mercury, Curry, and G\u00f6del, each of which has its own take on logic programming.<\/p>\n<p>I'll end with a few miniKanren resources: <\/p>\n<p>The main miniKanren website:\n<a href=\"http:\/\/minikanren.org\/\">http:\/\/minikanren.org\/<\/a><\/p>\n<p>An interview I gave on relational programming and miniKanren, including a comparison with Prolog: \n<a href=\"http:\/\/www.infoq.com\/interviews\/byrd-relational-programming-minikanren\">http:\/\/www.infoq.com\/interviews\/byrd-relational-programming-minikanren<\/a><\/p>\n<p>Cheers,<\/p>\n<p>--Will<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/28556223\/revisions\" title=\"show all edits to this post\">edited Feb 18 '15 at 10:20<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2561452\/robert-oschler\">Robert Oschler<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            7,1801717 gold badges5555 silver badges176176 bronze badges&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 17 '15 at 6:49&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4570634\/william-e-byrd\">William E. Byrd<\/a>William E. Byrd&#13;\n        <div class=\"-flair\">&#13;\n            2,45711 gold badge66 silver badges55 bronze badges&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Tentative answer: <\/p>\n<p>AFAIK, \"The Reasoned Schemer\" introduced basic logic programming in a Scheme-y syntax and functional programming style, adding in particular the constant goals \"#u\" (fail) and \"#s\" (suceeed) to the boolean values \"#t\" and \"#f\". It used the same approach to logic programming as Prolog: Unification and backtracking search. I will see whether I have some time to retrieve that book from my shelf over the weekend. The branch of mathematics is a restricted form first-order logic, in this case Horn clauses, and the Resolution Unfication. See: <a href=\"http:\/\/www.computational-logic.org\/iccl\/downloads\/Robinson-CL2000.pdf\" rel=\"nofollow noreferrer\">Computational Logic: Memories of the Past\nand Challenges for the Future<\/a> by John Alan Robinson and <a href=\"http:\/\/www.doc.ic.ac.uk\/~rak\/papers\/the%20early%20years.pdf\" rel=\"nofollow noreferrer\">The early years of logic programming<\/a> by Robert Kowalski for a cold start.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Feb 13 '15 at 21:21&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/884463\/david-tonhofer\">David Tonhofer<\/a>David Tonhofer&#13;\n        <div class=\"-flair\">&#13;\n            5,97011 gold badge3737 silver badges3333 bronze badges&#13;\n        <\/div>&#13;<\/div>","created_at":"2019-09-10T02:23:56+0000","updated_at":"2019-09-10T02:24:06+0000","annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1017,"title":"Devilspie2","url":"http:\/\/www.nongnu.org\/devilspie2\/","content":"<p>\n\t    Devilspie2 is a window matching utility, allowing the user to perform\n\t\tscripted actions on windows as they are created. For example you can\n\t\tscript a terminal program to always be positioned at a specific screen\n\t\tposition, or position a window on a specific workspace.\n\t\tThe code is free software, released under the\n\t\t<a href=\"http:\/\/www.gnu.org\/copyleft\/gpl.html\" class=\"extern\">GPL3<\/a>\n\t\tlicense and is availible in a\n\t\t<a href=\"http:\/\/git.savannah.gnu.org\/cgit\/devilspie2.git\" class=\"extern\">git\n\t\trepository on savannah<\/a>.\n\t<\/p><p>\n\t\tDevilspie2 is based on the program <a href=\"http:\/\/www.burtonini.com\/wordpress\/tag\/devilspie\/\" class=\"extern\">Devilspie<\/a> by Ross Burton. The difference is that\n\t\tDevilspie2 has replaced the symbolic expressions of the original with a\n\t\t<a href=\"http:\/\/www.lua.org\" class=\"extern\">LUA<\/a> interpreter, making it\n\t\tmuch easier to maintain and extend.\n\t<\/p>\n\t<p>\n\t\tPlease note that the rules of Devilspie2 is <b><i>not<\/i><\/b> compatible with the\n\t\tsymbolic expressions of the original Devilspie.\n\t<\/p>\n\t<p>\n\t\tAlso see the latest version of the complete <a href=\"http:\/\/git.savannah.gnu.org\/cgit\/devilspie2.git\/plain\/README\" class=\"extern\">README<\/a> and the <a href=\"http:\/\/www.nongnu.org\/devilspie2\/manual.php\">manual<\/a> (WIP).\n\t<\/p>\n\t\n\t<p>\n\t\tThis program is free software: you can redistribute it and\/or modify it\n\t\tunder the terms of the GNU General Public License as published by the Free\n\t\tSoftware Foundation, either version 3 of the License, or (at your option)\n\t\tany later version. This program is distributed in the hope that it will be\n\t\tuseful, but <b>without any warranty<\/b>. See the <a class=\"extern\" href=\"http:\/\/www.gnu.org\/copyleft\/gpl.html\">GNU General Public License<\/a>\n\t\tfor more details.\n\t<\/p>\n\t\n\t<p>\n\t\tThe following downloads are available:\n\t<\/p>\n\t\n\t<p>\n\t\tCurrent version: \n\t\t<img src=\"http:\/\/www.nongnu.org\/devilspie2\/images\/empty.png\" alt=\"\" \/><a href=\"http:\/\/download.savannah.nongnu.org\/releases\/devilspie2\/devilspie2_0.43-src.tar.gz\">\n\t\tdevilspie2_0.43-src.tar.gz<\/a> \n\t\t<img src=\"http:\/\/www.nongnu.org\/devilspie2\/images\/empty.png\" alt=\"\" \/>\n\t\t[<a href=\"http:\/\/download.savannah.nongnu.org\/releases\/devilspie2\/devilspie2_0.43-src.tar.gz.md5.txt\">md5sum<\/a>]\n\t\t[<a href=\"http:\/\/download.savannah.nongnu.org\/releases\/devilspie2\/devilspie2_0.43-src.tar.gz.asc\">gpgsig<\/a>]\n\t<\/p>\n\t<p>\n\t\tOlder versions available <a href=\"http:\/\/download.savannah.nongnu.org\/releases\/devilspie2\/\">here<\/a>.\n\t<\/p>","created_at":"2019-09-07T11:44:02+0000","updated_at":"2019-09-07T11:44:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.nongnu.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1015,"title":"Introduction to Stock & Options","url":"https:\/\/blog.dweek.ly\/introduction-to-stock-options-startup-founder-entrepreneur-employee\/","content":"<p><a href=\"http:\/\/www.amazon.com\/dp\/B0055PQ4H8\">Buy this book on Amazon\u2019s Kindle Store!<\/a><\/p><div><a href=\"http:\/\/www.scribd.com\/doc\/55945011\" target=\"_blank\">View this document on Scribd<\/a><\/div>\n<div><a href=\"http:\/\/www.scribd.com\/doc\/93815774\" target=\"_blank\">View this document on Scribd<\/a><\/div>\n<div class=\"jetpack-video-wrapper\"><iframe class=\"youtube-player\" width=\"840\" height=\"503\" src=\"https:\/\/www.youtube.com\/embed\/NLM89kLn8Sk?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/div>","created_at":"2019-09-06T03:35:04+0000","updated_at":"2019-09-06T03:35:04+0000","published_at":"2012-02-02T23:50:28+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"blog.dweek.ly","preview_picture":"http:\/\/img.youtube.com\/vi\/NLM89kLn8Sk\/0.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1016,"title":"How to do a code review","url":"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/","content":"<p>The pages in this section contain recommendations on the best way to do code\nreviews, based on long experience. All together they represent one complete\ndocument, broken up into many separate sections. You don\u2019t have to read them\nall, but many people have found it very helpful to themselves and their team to\nread the entire set.<\/p><ul><li><a href=\"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/standard.html\">The Standard of Code Review<\/a><\/li>\n  <li><a href=\"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/looking-for.html\">What to Look For In a Code Review<\/a><\/li>\n  <li><a href=\"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/navigate.html\">Navigating a CL in Review<\/a><\/li>\n  <li><a href=\"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/speed.html\">Speed of Code Reviews<\/a><\/li>\n  <li><a href=\"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/comments.html\">How to Write Code Review Comments<\/a><\/li>\n  <li><a href=\"https:\/\/google.github.io\/eng-practices\/review\/reviewer\/pushback.html\">Handling Pushback in Code Reviews<\/a><\/li>\n<\/ul><p>See also the <a href=\"https:\/\/google.github.io\/eng-practices\/review\/developer\/\">CL Author\u2019s Guide<\/a>, which gives detailed guidance\nto developers whose CLs are undergoing review.<\/p>","created_at":"2019-09-06T03:35:04+0000","updated_at":"2019-09-06T03:35:04+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"google.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1014,"title":"jamiebuilds\/babel-handbook","url":"https:\/\/github.com\/jamiebuilds\/babel-handbook\/blob\/master\/translations\/en\/README.md","content":"<p>This handbook is divided into two parts:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/jamiebuilds\/babel-handbook\/blob\/master\/translations\/en\/user-handbook.md\">User Handbook<\/a> - How to setup\/configure Babel and more.<\/li>\n<li><a href=\"https:\/\/github.com\/jamiebuilds\/babel-handbook\/blob\/master\/translations\/en\/plugin-handbook.md\">Plugin Handbook<\/a> - How to create plugins for Babel.<\/li>\n<\/ul><blockquote>\n<p>For future updates, follow <a href=\"https:\/\/twitter.com\/thejameskyle\" rel=\"nofollow\">@thejameskyle<\/a>\non Twitter.<\/p>\n<\/blockquote>\n<p>If you are reading a non-english translation of this handbook you may still find\nenglish sections that have not yet been translated. If you would like to\ncontribute to one of the translations you must do so through Crowdin. Please\nread the <a href=\"https:\/\/github.com\/jamiebuilds\/babel-handbook\/blob\/master\/CONTRIBUTING.md\">contributing guidelines<\/a> for more information. You\nwill find a number of english words that are programming concepts. If these were\ntranslated to other languages there would be a lack of consistency and fluency\nwhen reading about them. In many cases you will find the literal translation\nfollowed by the english term in parenthesis <code>()<\/code>. For example: Abstract Syntax\nTrees (ASTs).<\/p>","created_at":"2019-09-06T03:35:03+0000","updated_at":"2019-09-06T03:35:03+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/952783?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1013,"title":"walkable.gitlab.io","url":"https:\/\/walkable.gitlab.io\/","created_at":"2019-09-02T13:59:08+0000","updated_at":"2019-09-02T13:59:08+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"walkable.gitlab.io"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1012,"title":"The \u201cmail is hard\u201d myth","url":"https:\/\/news.ycombinator.com\/item?id=20851880","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D20851880\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"20851880\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/poolp.org\/posts\/2019-08-30\/you-should-not-run-your-mail-server-because-mail-is-hard\/\" class=\"storylink\">The \u201cmail is hard\u201d myth<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=poolp.org\">poolp.org<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        189 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=tambourine_man\" class=\"hnuser\">tambourine_man<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=20851880\">3 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=20851880&amp;goto=item%3Fid%3D20851880\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=The%20%E2%80%9Cmail%20is%20hard%E2%80%9D%20myth&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=The%20%E2%80%9Cmail%20is%20hard%E2%80%9D%20myth\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=20851880&amp;auth=cce25917b20f514cc55540218aac5adcee27b513\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=20851880\">106\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-09-01T16:09:46+0000","updated_at":"2019-09-01T16:09:46+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1011,"title":"rss.xml","url":"rss.xml","created_at":"2019-08-28T15:17:01+0000","updated_at":"2019-08-28T15:17:01+0000","published_by":[""],"annotations":[],"reading_time":0},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1010,"title":"Datomic - Our Customers","url":"https:\/\/www.datomic.com\/customers.html","content":"<p>Our customers are using Datomic in a variety of industries, and a variety of ways.  Read some of their stories.<\/p><div class=\"w-row d-customers-row\"><div class=\"w-col w-col-4 w-col-small-small-stack w-col-stack\"><div class=\"d-customers-customer\"><img class=\"d-customers-logo\" src=\"https:\/\/www.datomic.com\/images\/customers-room-key.png\" alt=\"image\" \/><p>Hotel search personalization - Room Key leverages the scalability and simplicity of Datomic to serve millions of customer visits.<\/p><p><a href=\"https:\/\/www.datomic.com\/room-keys-story.html\" class=\"d-learn-more-button\">Read More<\/a><\/p><\/div><\/div><div class=\"w-col w-col-4 w-col-small-small-stack w-col-stack\"><div class=\"d-customers-customer\"><img class=\"d-customers-logo\" src=\"https:\/\/www.datomic.com\/images\/customers-vital-reactor.png\" alt=\"image\" \/><p>Switched from schema-less NoSQL store to Datomic; leveraging rich, flexible information model to simplify complex development problems, and Datomic's built-in notion of time for auditing and data provenance.<\/p><p><a href=\"https:\/\/www.datomic.com\/vitalreactors-story.html\" class=\"d-learn-more-button\">Read More<\/a><\/p><\/div><\/div><div class=\"w-col w-col-4 w-col-small-small-stack w-col-stack\"><div class=\"d-customers-customer\"><img class=\"d-customers-logo\" src=\"https:\/\/www.datomic.com\/images\/customers-pellucid.png\" alt=\"image\" \/><p>Data analysis for investment bankers - Pellucid expands the speed, power and scalability of its data system through Datomic.<\/p><p><a href=\"https:\/\/www.datomic.com\/pellucids-story.html\" class=\"d-learn-more-button\">Read More<\/a><\/p><\/div><\/div><\/div><div class=\"w-row d-customers-row\"><div class=\"w-col w-col-4 w-col-small-small-stack w-col-stack\"><div class=\"d-customers-customer\"><img class=\"d-customers-logo\" src=\"https:\/\/www.datomic.com\/images\/customers-rally.png\" alt=\"image\" \/><p>Enterprise software for agile project management - Rally uses Datomic to add real-time collaboration features to its backend datastore.<\/p><p><a href=\"https:\/\/www.datomic.com\/rallys-story.html\" class=\"d-learn-more-button\">Read More<\/a><\/p><\/div><\/div><div class=\"w-col w-col-4 w-col-small-small-stack w-col-stack\"><div class=\"d-customers-customer\"><img class=\"d-customers-logo\" src=\"https:\/\/www.datomic.com\/images\/customers-brigade.png\" alt=\"image\" \/><p>Creating interactive digital experiences - The Brigade quickly created a social music application for Spotify using Datomic.<\/p><p><a href=\"https:\/\/www.datomic.com\/the-brigades-story.html\" class=\"d-learn-more-button\">Read More<\/a><\/p><\/div><\/div><div class=\"w-col w-col-4 w-col-small-small-stack w-col-stack\"><div class=\"d-customers-customer\"><img class=\"d-customers-logo\" src=\"https:\/\/www.datomic.com\/images\/customers-nubank.png\" alt=\"image\" \/><p>Disrupting the financial services industry with superior technology - Nubank uses Datomic to power their nimble, elastic microservices platform.<\/p><p><a href=\"https:\/\/www.datomic.com\/nubanks-story.html\" class=\"d-learn-more-button\">Read More<\/a><\/p><\/div><\/div><\/div>","created_at":"2019-08-27T18:28:00+0000","updated_at":"2019-08-27T18:28:00+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.datomic.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1009,"title":"i18next\/react-i18next","url":"https:\/\/github.com\/i18next\/react-i18next","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p><a href=\"https:\/\/circleci.com\/gh\/i18next\/react-i18next\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/bef46a56c6b9b151b1f68eb06bce8f737ff09dff\/68747470733a2f2f636972636c6563692e636f6d2f67682f6931386e6578742f72656163742d6931386e6578742e7376673f7374796c653d737667\" alt=\"CircleCI\" data-canonical-src=\"https:\/\/circleci.com\/gh\/i18next\/react-i18next.svg?style=svg\" \/><\/a>\n<a href=\"https:\/\/codeclimate.com\/github\/i18next\/react-i18next\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7df6907249ab542adcc9bb9be2c947b9bb2f5a9e\/68747470733a2f2f636f6465636c696d6174652e636f6d2f6769746875622f636f6465636c696d6174652f636f6465636c696d6174652f6261646765732f6770612e737667\" alt=\"Code Climate\" data-canonical-src=\"https:\/\/codeclimate.com\/github\/codeclimate\/codeclimate\/badges\/gpa.svg\" \/><\/a>\n<a href=\"https:\/\/coveralls.io\/github\/i18next\/react-i18next\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7a6d5d077c0feed62ed4513e04717dba835676bb\/68747470733a2f2f636f766572616c6c732e696f2f7265706f732f6769746875622f6931386e6578742f72656163742d6931386e6578742f62616467652e737667\" alt=\"Coverage Status\" data-canonical-src=\"https:\/\/coveralls.io\/repos\/github\/i18next\/react-i18next\/badge.svg\" \/><\/a>\n<a href=\"http:\/\/packagequality.com\/#?package=react-i18next\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/93dd733ff314c1a0023e05bd11d832c2c78337bb\/687474703a2f2f6e706d2e7061636b6167657175616c6974792e636f6d2f736869656c642f72656163742d6931386e6578742e737667\" alt=\"Quality\" data-canonical-src=\"http:\/\/npm.packagequality.com\/shield\/react-i18next.svg\" \/><\/a>\n<a href=\"https:\/\/david-dm.org\/i18next\/react-i18next\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f5e01b6da75f01df5cfb890bd75ab1bb4ad6afdb\/68747470733a2f2f64617669642d646d2e6f72672f6931386e6578742f72656163742d6931386e6578742e706e67\" alt=\"dependencies\" data-canonical-src=\"https:\/\/david-dm.org\/i18next\/react-i18next.png\" \/><\/a>\n<a href=\"https:\/\/david-dm.org\/i18next\/react-i18next#info=devDependencies\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/ac0e364698d5290aeade8ed2a3a71d4ebd3c23da\/68747470733a2f2f64617669642d646d2e6f72672f6931386e6578742f72656163742d6931386e6578742f6465762d7374617475732e706e67\" alt=\"devdependencies\" data-canonical-src=\"https:\/\/david-dm.org\/i18next\/react-i18next\/dev-status.png\" \/><\/a><\/p>\n<h3><a id=\"user-content-important\" class=\"anchor\" aria-hidden=\"true\" href=\"#important\"><\/a>IMPORTANT:<\/h3>\n<p>Master Branch is the new v10 using hooks.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ v10.0.0\nnpm i react-i18next<\/pre><\/div>\n<p><strong>react-native: not yet supports hooks (hooks are part of react-native v0.59.0)!!!<\/strong><\/p>\n<p>For the legacy version please use the <a href=\"https:\/\/github.com\/i18next\/react-i18next\/tree\/v9.x.x\">v9.x.x Branch<\/a><\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ v9.0.10 (legacy)\nnpm i react-i18next@legacy<\/pre><\/div>\n<h3><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h3>\n<p>The documentation is published on <a href=\"https:\/\/react.i18next.com\" rel=\"nofollow\">react.i18next.com<\/a><\/p>\n<h3><a id=\"user-content-what-will-my-code-look-like\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-will-my-code-look-like\"><\/a>What will my code look like?<\/h3>\n<p><strong>Before:<\/strong> Your react code would have looked something like:<\/p>\n<div class=\"highlight highlight-source-js-jsx\"><pre>...\n&lt;div&gt;Just simple content&lt;\/div&gt;\n&lt;div&gt;\n  Hello &lt;strong title=\"this is your name\"&gt;{name}&lt;\/strong&gt;, you have {count} unread message(s). &lt;Link to=\"\/msgs\"&gt;Go to messages&lt;\/Link&gt;.\n&lt;\/div&gt;\n...<\/pre><\/div>\n<p><strong>After:<\/strong> With the trans component just change it to:<\/p>\n<div class=\"highlight highlight-source-js-jsx\"><pre>...\n&lt;div&gt;{t('simpleContent')}&lt;\/div&gt;\n&lt;Trans i18nKey=\"userMessagesUnread\" count={count}&gt;\n  Hello &lt;strong title={t('nameTitle')}&gt;{{name}}&lt;\/strong&gt;, you have {{count}} unread message. &lt;Link to=\"\/msgs\"&gt;Go to messages&lt;\/Link&gt;.\n&lt;\/Trans&gt;\n...<\/pre><\/div>\n<p>Head over to the <strong>interactive playground<\/strong> at <a href=\"https:\/\/codesandbox.io\/s\/1zxox032q\" rel=\"nofollow\">codesandbox<\/a>.<\/p>\n<h3><a id=\"user-content--what-others-say\" class=\"anchor\" aria-hidden=\"true\" href=\"#-what-others-say\"><\/a>\ud83d\udcd6 What others say<\/h3>\n<ul><li><a href=\"https:\/\/alligator.io\/react\/i18n-with-react-and-i18next\" rel=\"nofollow\">I18n with React and i18next<\/a> via Alligator.io by Danny Hurlburt<\/li>\n<li><a href=\"https:\/\/itnext.io\/ultimate-localization-of-react-mobx-app-with-i18next-efab77712149\" rel=\"nofollow\">Ultimate Localization of React (Mobx) App with i18next<\/a> via itnext.io by Viktor Shevchenko<\/li>\n<li><a href=\"https:\/\/reactjsexample.com\/internationalization-for-react-done-right-using-the-i18next-i18n-ecosystem\/\" rel=\"nofollow\">Internationalization for react done right Using the i18next i18n ecosystem<\/a> via reactjsexample.com<\/li>\n<li><a href=\"https:\/\/codetain.co\/2018\/05\/using-i18next-to-translate-reactjs-application\/\" rel=\"nofollow\">Using i18next to translate React.js application<\/a> via codetain.co by Kasia Dadek<\/li>\n<li><a href=\"https:\/\/www.gatsbyjs.org\/blog\/2017-10-17-building-i18n-with-gatsby\/\" rel=\"nofollow\">Building i18n with Gatsby<\/a> via gatsbyjs.org by Samuel Goudie<\/li>\n<li><a href=\"https:\/\/medium.com\/@jamuhl\/get-your-react-js-application-translated-with-style-4ad090aefc2c\" rel=\"nofollow\">Get your react.js application translated with style<\/a> by Jan M\u00fchlemann<\/li>\n<li><a href=\"https:\/\/medium.com\/@jamuhl\/translate-your-expo-io-react-native-mobile-application-aa220b2362d2\" rel=\"nofollow\">Translate your expo.io \/ react-native mobile application<\/a> by Jan M\u00fchlemann<\/li>\n<li>you're welcome to share your story...<\/li>\n<\/ul><h3><a id=\"user-content-why-i18next\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-i18next\"><\/a>Why i18next?<\/h3>\n<ul><li><strong>Simplicity:<\/strong> no need to change your webpack configuration or adding additional babel transpilers, just use create-react-app and go<\/li>\n<li><strong>Production ready<\/strong> we know there are more needs for production than just doing i18n on the clientside. So we offer wider support on <a href=\"https:\/\/www.i18next.com\/supported-frameworks.html\" rel=\"nofollow\">serverside<\/a> too (nodejs, php, ruby, .net, ...). <strong>Learn once - translate everywhere<\/strong>.<\/li>\n<li><strong>Beyond i18n<\/strong> comes with <a href=\"https:\/\/locize.com\" rel=\"nofollow\">locize<\/a> bridging the gap between developement and translations - covering the whole translation process.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/9d49328bd3dd451dcec282529179102be66a3d12\/68747470733a2f2f626c6f627363646e2e676974626f6f6b2e636f6d2f76302f622f676974626f6f6b2d32383432372e61707073706f742e636f6d2f6f2f6173736574732532462d4c39695336576d3268796e53354839476a376a2532462d4c396953374c6c543257377746744a482d326e2532462d4c3969534250395536352d62484a42525344762532466931386e6578742d65636f73797374656d2e6a70673f67656e65726174696f6e3d3135323333343533313831323239313326616c743d6d65646961\"><img src=\"https:\/\/camo.githubusercontent.com\/9d49328bd3dd451dcec282529179102be66a3d12\/68747470733a2f2f626c6f627363646e2e676974626f6f6b2e636f6d2f76302f622f676974626f6f6b2d32383432372e61707073706f742e636f6d2f6f2f6173736574732532462d4c39695336576d3268796e53354839476a376a2532462d4c396953374c6c543257377746744a482d326e2532462d4c3969534250395536352d62484a42525344762532466931386e6578742d65636f73797374656d2e6a70673f67656e65726174696f6e3d3135323333343533313831323239313326616c743d6d65646961\" alt=\"i18next ecossystem\" width=\"400\" data-canonical-src=\"https:\/\/blobscdn.gitbook.com\/v0\/b\/gitbook-28427.appspot.com\/o\/assets%2F-L9iS6Wm2hynS5H9Gj7j%2F-L9iS7LlT2W7wFtJH-2n%2F-L9iSBP9U65-bHJBRSDv%2Fi18next-ecosystem.jpg?generation=1523345318122913&amp;alt=media\" \/><\/a><\/p>\n<h3><a id=\"user-content-localization-workflow\" class=\"anchor\" aria-hidden=\"true\" href=\"#localization-workflow\"><\/a>Localization workflow<\/h3>\n<p>Want to learn more about how seamless your internationalization and translation process can be?<\/p>\n<p><a href=\"https:\/\/www.youtube.com\/watch?v=9NOzJhgmyQE\" rel=\"nofollow\"><img src=\"https:\/\/github.com\/i18next\/react-i18next\/raw\/master\/example\/v9.x.x\/locize\/video_sample.png\" alt=\"video\" \/><\/a><\/p>\n<p><a href=\"https:\/\/www.youtube.com\/watch?v=9NOzJhgmyQE\" rel=\"nofollow\">watch the video<\/a><\/p>\n<h3><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h3>\n<p>Source can be loaded via <a href=\"https:\/\/www.npmjs.com\/package\/react-i18next\" rel=\"nofollow\">npm<\/a> or <a href=\"https:\/\/github.com\/i18next\/react-i18next\/blob\/master\/react-i18next.min.js\">downloaded<\/a> from this repo.<\/p>\n<pre># npm package\n$ npm install react-i18next\n<\/pre>\n<ul><li>If you don't use a module loader it will be added to <code>window.reactI18next<\/code><\/li>\n<\/ul><h3><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h3>\n<h4><a id=\"user-content-v9-samples\" class=\"anchor\" aria-hidden=\"true\" href=\"#v9-samples\"><\/a>v9 samples<\/h4>\n<h3><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h3>\n<ul><li>react &gt;= <strong>16.8.0<\/strong><\/li>\n<li>react-dom &gt;= <strong>16.8.0<\/strong><\/li>\n<li>react-native &gt;= <strong>0.59.0<\/strong><\/li>\n<li>i18next &gt;= <strong>10.0.0<\/strong> (typescript users: &gt;=17.0.9)<\/li>\n<\/ul><h4><a id=\"user-content-v9\" class=\"anchor\" aria-hidden=\"true\" href=\"#v9\"><\/a>v9<\/h4>\n<ul><li>react &gt;= <strong>0.14.0<\/strong> (in case of &lt; v16 or preact you will need to define parent in <a href=\"https:\/\/react.i18next.com\/legacy-v9\/trans-component#trans-props\" rel=\"nofollow\">Trans component<\/a> or globally in <a href=\"https:\/\/react.i18next.com\/legacy-v9\/trans-component#additional-options-on-i-18-next-init\" rel=\"nofollow\">i18next.react options<\/a>)<\/li>\n<li>i18next &gt;= <strong>2.0.0<\/strong><\/li>\n<\/ul><p>Thanks goes to these wonderful people (<a href=\"https:\/\/github.com\/kentcdodds\/all-contributors#emoji-key\">emoji key<\/a>):<\/p>\n<p>This project follows the <a href=\"https:\/\/github.com\/kentcdodds\/all-contributors\">all-contributors<\/a> specification. Contributions of any kind welcome!<\/p>\n<hr \/><h3 style=\"text-align: center;\"><a id=\"user-content-gold-sponsors\" class=\"anchor\" aria-hidden=\"true\" href=\"#gold-sponsors\"><\/a>Gold Sponsors<\/h3>\n<p style=\"text-align: center;\">\n  <a href=\"https:\/\/locize.com\/\" rel=\"nofollow\">\n    <img src=\"https:\/\/raw.githubusercontent.com\/i18next\/i18next\/master\/assets\/locize_sponsor_240.gif\" width=\"240px\" alt=\"image\" \/><\/a>\n<\/p>\n<hr \/><p><strong>localization as a service - locize.com<\/strong><\/p>\n<p>Needing a translation management? Want to edit your translations with an InContext Editor? Use the orginal provided to you by the maintainers of i18next!<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/d76639e73ab0f69c7310bb482bb4b7cd75b24c1e\/68747470733a2f2f6c6f63697a652e636f6d2f696d672f6164732f6769746875625f6c6f63697a652e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/d76639e73ab0f69c7310bb482bb4b7cd75b24c1e\/68747470733a2f2f6c6f63697a652e636f6d2f696d672f6164732f6769746875625f6c6f63697a652e706e67\" alt=\"locize\" data-canonical-src=\"https:\/\/locize.com\/img\/ads\/github_locize.png\" \/><\/a><\/p>\n<p>With using <a href=\"http:\/\/locize.com\/?utm_source=react_i18next_readme&amp;utm_medium=github\" rel=\"nofollow\">locize<\/a> you directly support the future of i18next and react-i18next.<\/p>\n<hr \/><\/article>","created_at":"2019-08-26T15:23:29+0000","updated_at":"2019-08-26T15:23:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/8546082?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["ux","design"],"is_public":false,"id":1008,"title":"4 Rules for Intuitive UX \u2013 Learn UI Design","url":"https:\/\/learnui.design\/blog\/4-rules-intuitive-ux.html","content":"<p>This is my advice on improving the UX of your designs WITHOUT hours of user research sessions, paper prototyping playtime, or any other trendy UX buzzwords.<\/p><p>(Seriously, search \u201cdesign thinking\u201d. 0 results. <em>Nailed it<\/em>!)<\/p>\n<p>Who\u2019s this article for? I\u2019m looking at you:<\/p>\n<ul><li><strong>Developers<\/strong>. You created your own app, but every time someone downloads it, they struggle to use it. And you know if they\u2019re telling you this, then it\u2019s <em>really<\/em> bad.<\/li>\n  <li><strong>Graphic designers<\/strong>. Looking to make the transition into digital, but trying to learn UX by reading articles online is\u2026 a very painful way to die \ud83d\ude2c<\/li>\n  <li><strong>PMs<\/strong>. Your job is already like 25% UX designer. Would be nice to level up those skills.<\/li>\n  <li><strong>And the hustlers<\/strong>. Anyone working on digital side projects nights\/weekends. This one\u2019s for you too \ud83c\udf7b<\/li>\n<\/ul><p>If you\u2019re already a UX designer, I don\u2019t expect this article to go over super well with you. I\u2019m basically skipping over entire chunks of our field in favor of focusing entirely on <strong>the single most lacking skill in aspiring UX designers<\/strong> (or UX-adjacent folks who find themselves designing screens).<\/p>\n<p>I call it \u201c<strong>speaking interface<\/strong>\u201d.<\/p>\n<p>When I started as a professional UX designer, I was <em>shocked<\/em> how many times my clients would hand me the initial wireframes (or the living, breathing, in-browser MVP) and there\u2019d be <strong><em>completely obvious UX mistakes<\/em><\/strong> all over them. I\u2019m not talking about things you need hours of research and A\/B testing to discover. I\u2019m talking, like, <em>dead simple<\/em> mistakes.<\/p>\n<p>For lack of a better example:<\/p>\n<blockquote class=\"twitter-tweet\"><p lang=\"en\" dir=\"ltr\">Somewhere out there, there's a team that knows HTML, but doesn't know the difference between a radio button and a checkbox. <a href=\"https:\/\/t.co\/VBwk8Jxekd\">pic.twitter.com\/VBwk8Jxekd<\/a><\/p><div>\u2014 Erik D. Kennedy (@erikdkennedy) <\/div><a href=\"https:\/\/twitter.com\/erikdkennedy\/status\/867412999115464705?ref_src=twsrc%5Etfw\">May 24, 2017<\/a><\/blockquote>\n<p>Now my clients weren\u2019t <em>this<\/em> bad, but <em>look\u200a<\/em>-\u200ayou don\u2019t need to be Bret Victor to understand that if you can only select ONE thing from a list, you need RADIO BUTTONS, not checkboxes. To understand <em>that<\/em>, you just need to be able to <em><strong>speak interface<\/strong><\/em>. And that\u2019s the craziest thing to me. Interface fluency is something <em>anyone<\/em> can achieve. You don\u2019t need college, you don\u2019t need Lambda school, yadda yadda.<\/p>\n<p>Frankly, you just need the <strong>presence of mind<\/strong> to (A) pause every single time you\u2019re confused or frustrated by some app, (B) verbalize <em>what<\/em> about the interface makes you confused\/frustrated\/etc., and then (C) figure out how you could avoid that specific snafu that in your own designs.<\/p>\n<p>Rinse and repeat that non-stop and you\u2019ll be a pro in no time.<\/p>\n<p>What I want to talk about today is <strong>four little rules<\/strong> that will help eliminate these pain points in your own designs. They\u2019re the heuristics that are a level or two deeper than \u201cuse radio buttons if the user can only select one thing\u201d. But, if you can remember to obey the things in this checklist, you\u2019ll be that much closer to creating designs that your users can use easily right off the bat, freeing up your time for other, more important things.<\/p>\n<p>(That\u2019s when the other UX designers can lecture you on the newest academic user research methodologies!)<\/p>\n<p>Here\u2019s what we\u2019ll be covering:<\/p>\n<ol><li><a href=\"#1-obey-the-law-of-locality\">Obey the Law of Locality<\/a><\/li>\n  <li><a href=\"#2-abd-anything-but-dropdowns\">ABD: Anything But Dropdowns<\/a><\/li>\n  <li><a href=\"#3-pass-the-squint-test\">Pass the Squint Test<\/a><\/li>\n  <li><a href=\"#4-teach-by-example\">Teach by example<\/a><\/li>\n<\/ol><p>Any questions? Let\u2019s dive right in.<\/p>\n<h2 id=\"1-obey-the-law-of-locality\">1. Obey the Law of Locality<\/h2>\n<p><em>Put interface elements where they affect change.<\/em><\/p>\n<p>All else being equal, you should put the elements in your interface <em>near where they affect change<\/em>. This is because, when a user wants to make a change to the system, <strong>they will unwittingly <em>glance at where that change will happen<\/em><\/strong>.<\/p>\n<p>So, let\u2019s say you have a list of <em>things<\/em>. Where do you put the \u201cADD A NEW THING\u201d button?<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-where.png\" alt=\"law of locality in playlist illustration\" \/><\/figure><p>Q: Well, where does the <em>change<\/em> happen?<\/p>\n<p>A: At the end of the list.<\/p>\n<p>Great, put the button at the end of the list.<\/p>\n<p>WAIT. You\u2019d think this would be pretty simple. But there\u2019s a <em>temptation<\/em>.<\/p>\n<p><strong>The temptation is to just put it <em>where we have space for it<\/em>.<\/strong><\/p>\n<p>For instance, if you have a menu, maybe you\u2019d think \u201cWe have a menu! Why not just put it in the menu!?\u201d<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-menu-no.png\" alt=\"the law of locality violated in a music UI\" \/><\/figure><p>The answer is, of course, because <strong>users won\u2019t look for it there<\/strong>.<\/p>\n<p>(And the ultimate answer is that having a place where \u201cwe just put things\u201d will ultimately render your app an unusable mess that people will abandon the first chance they see a half-viable alternative)<\/p>\n<p>Don\u2019t think I\u2019m joking. Have you ever noted <em>this<\/em> interface?<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-evernote.png\" alt=\"the law of locality violated in evernote's interface\" \/><\/figure><p>An equally-bad\/common alternative is to just take a solution that you\u2019ve seen applied by A Respected Tech Company without any thought as to if it makes sense for <em>you<\/em>. \u201cWe need an \u2018Add\u2019 button? I\u2019ve seen one of those. Hold my beer!\u201d<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-fab-no.png\" alt=\"the law of locality violated with a floating action button\" \/><\/figure><p>Look. Another button in a place users <strong><em>will never look for it<\/em><\/strong>. To compound things, users will suspect <em>this<\/em> button actually adds a new whatever-is-currently-displayed-on-the-big-blank-white-space. Because that\u2019s where the control <em>is<\/em>.<\/p>\n<p>Your users <em>want<\/em> you to follow the Law of Locality.<\/p>\n<p>So, now that we know it, let\u2019s use it.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-bottom-of-list.png\" alt=\"the law of locality in a list of music playlists\" \/><\/figure><p>Bam.<\/p>\n<p>But maybe you\u2019re a born UX designer and <strong>you always visualize what happens when there\u2019s 1000 items instead of 5<\/strong> and you realize: <em>there\u2019s still an issue here<\/em>. If the user creates a TON of playlists, this button will disappear hundreds of pixels offscreen!<\/p>\n<p>So maybe you could anchor the button <em>near the bottom of the list<\/em>, but have it <em>always be visible<\/em>, no matter how many hundreds of playlists the user has created.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-spotify-new-bottom.png\" alt=\"the law of locality in Spotify's UI\" \/><figcaption class=\"post__figcaption\">For bonus points, (1) use the inline button UNTIL it's about to go offscreen, and at that point switch to the anchored solution and (2) make it more visible than Spotify's button, which took me months to notice while I haplessly right-clicked individual songs to add them to my playlists!<\/figcaption><\/figure><p>Brilliant! And this is what Spotify has done.<\/p>\n<p>Another possibility is to say \u201cHey, we can\u2019t <em>reliably and consistently<\/em> show the button at the bottom of the list. Where\u2019s the <em>nearest logical place<\/em> to put it?\u201d<\/p>\n<p>And the answer is, (I think pretty obviously) <em>the top of the list<\/em>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-spotify-new-top.png\" alt=\"the law of locality obeyed in Spotify's UI\" \/><figcaption class=\"post__figcaption\">I wish.<\/figcaption><\/figure><p>Sacrebleu! This is actually just what Spotify-competitor Rdio did, before they were acqui-shut-down by Pandora.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/lol-rdio.png\" alt=\"law of locality obeyed in rdio's UI\" \/><figcaption class=\"post__figcaption\">Reconstructed from memory (like all reality, if you think about\u00a0it)<\/figcaption><\/figure><p>The lesson here is clear. Never sell your company, and always always obey the Law of Locality.<\/p>\n<p>(There are actually 3 laws of locality, and \u201cPut UI elements where they affect change\u201d is only the first. If you\u2019re interested, read more <a href=\"https:\/\/learnui.design\/blog\/the-3-laws-of-locality.html\">here<\/a>)<\/p>\n<p>Next!<\/p>\n<h2 id=\"2-abd-anything-but-dropdowns\">2. ABD: Anything but Dropdowns<\/h2>\n<p><em>Any time you feel tempted to use a dropdown, ask yourself if one of these 12 controls is better instead.<\/em><\/p>\n<p>One <strong>non-obvious lesson of UX design<\/strong> is that <strong>dropdowns<\/strong> are pretty much the <strong>worst control<\/strong>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-hell.png\" alt=\"3 dropdowns to specify one simple date\" \/><figcaption class=\"post__figcaption\">Welcome to\u00a0hell!<\/figcaption><\/figure><p>They\u2019re not <em>always<\/em> bad, but you\u2019re working against the following:<\/p>\n<ul><li>Dropdowns take <strong>too many clicks\/taps<\/strong>. One to open, a few more to scroll around to the right option (on mobile), another to select the right option, and (on mobile) another to close. (Compare to the <em>single click use-cases<\/em> of many of the options listed below)<\/li>\n  <li>Dropdowns <strong>don\u2019t show you the options<\/strong>! You have to click into them to see the possible values, and on mobile, you can often only see a couple at a time.<\/li>\n  <li>Long dropdowns are <strong>ridiculous to navigate<\/strong>. A country dropdown for an app used worldwide could have 195+ countries. At some point, almost <em>any other method<\/em> of asking a user their country would be quicker than having them scroll through a dropdown (\u201cSmoke signals?\u201d AGCKKHKGH).<\/li>\n<\/ul><p>This is pretty straightforward, so let\u2019s just cover some examples for the various major cases of dropdown replacement.<\/p>\n<h3 id=\"if-youre-choosing-between-2options\">If you\u2019re choosing between 2\u00a0options\u2026<\/h3>\n<p>We already have some fantastic options for allowing users to choose 1 of 2 things, all of which (A) show the options right away and (B) require fewer taps\/clicks.<\/p>\n<p>For questions to which there is no \u201cdefault\u201d answer, and either might be picked with roughly equal frequency, try a <strong>segmented button<\/strong>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-two-segment.png\" alt=\"segmented button instead of dropdown control\" \/><\/figure><p>If there is a \u201cdefault state\u201d that corresponds to \u201cOff\u201d, try a <strong>checkbox<\/strong>. A checkbox is also good for settings that don\u2019t affect change <em>until the user presses Save or Submit<\/em>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-two-checkbox.png\" alt=\"checkbox instead of dropdown control\" \/><\/figure><p>Similar to the checkbox is the <strong>switch<\/strong>, which is good for changes that should apply <em>immediately<\/em>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-two-switch.png\" alt=\"switch instead of dropdown control\" \/><\/figure><p>Checkboxes and switches <em>only<\/em> make sense when there are two options. However, the following controls make sense for 2 to roughly 5 options, so you might try some of the following instead.<\/p>\n<h3 id=\"if-youre-choosing-between-25options\">If you\u2019re choosing between 2\u20135\u00a0options\u2026<\/h3>\n<p>We covered <strong>segmented buttons<\/strong> above (and they apply here too) but it\u2019s worth mentioning that when there are more options, <strong>vertical segmented buttons<\/strong> allow even more flexibility of answer length.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-few-segment.png\" alt=\"vertical segmented button instead of dropdown control\" \/><\/figure><p><strong>Radio buttons<\/strong> are similar, but particularly useful if you need to display a couple sub-elements for each choice.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-few-radio.png\" alt=\"radio button instead of dropdown control\" \/><\/figure><p>For detailed displays of just a few choices, <strong>cards<\/strong> are where it\u2019s at.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-few-cards.png\" alt=\"cards instead of dropdown control\" \/><\/figure><p>One trick I like is displaying <strong>visual options<\/strong> literally.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-tesla-visual.png\" alt=\"visual options instead of dropdown control\" \/><figcaption class=\"post__figcaption\">Tesla likes it too, apparently.<\/figcaption><\/figure><h3 id=\"if-youre-choosing-between-manyoptions\">If you\u2019re choosing between many\u00a0options\u2026<\/h3>\n<p>When there are enough options that scrolling through them is annoying, consider a <strong>typeahead<\/strong> control. It\u2019s like a search bar that shows top matching results as you type.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-many-typeahead.png\" alt=\"typeahead control instead of dropdown control\" \/><\/figure><h3 id=\"if-youre-choosing-adate\">If you\u2019re choosing a\u00a0date\u2026<\/h3>\n<p>Picking a date from dropdowns is the <em>worst<\/em>. If I ever do this, then I\u2019ve <em>really<\/em> failed as a UX designer.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-dates-dont.png\" alt=\"don't use dropdown controls for choosing dates\" \/><\/figure><p>But what do you use instead? Well, it depends. First question: what <em>type of date are you picking<\/em>?<\/p>\n<ol><li><strong>Poisson dates<\/strong>. Dates <em>most likely to be in the near future<\/em>, tapering off as you go farther into the future (or nearer to the present), e.g. date of an appointment you\u2019re scheduling, date of a flight you\u2019re purchasing<\/li>\n  <li><strong>High-variability dates<\/strong>. Dates that have a <em>similar probability of being anywhere in a wide range of time<\/em>, e.g. date of birth, day-and-month of your birthday<\/li>\n<\/ol><p>(Yes, I named \u201cPoisson dates\u201d after the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Poisson_distribution\" target=\"_blank\">mathematical distribution<\/a> \ud83e\udd13)<\/p>\n<p>For different types of date-picking, you should use different controls.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-dates-chart.png\" alt=\"chart of poisson vs. wide-range dates\" \/><\/figure><p>For Poisson dates, you want to make it DEAD SIMPLE to pick dates in the most common range (e.g. for scheduling an appointment, it might be the next, say, 14 days). It\u2019s perfectly OK if picking dates outside of that range is a little tougher.<\/p>\n<p>A <strong>calendar control<\/strong> fits the bill rather well for Poisson dates. If you know the date to-be-picked is most likely in the next 2\u20134 weeks, you\u2019re golden.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-date-calendar.png\" alt=\"calendar control instead of dropdown control\" \/><\/figure><p>Rather creatively, <a href=\"https:\/\/www.google.com\/flights\" target=\"_blank\">Google Flights<\/a> <em>defaults<\/em> to you selecting a flight roughly 2 weeks in the future, which is perhaps an opportunity for confusion (\u201cI didn\u2019t choose this!\u201d), but probably a better date to default to, and closer to the hump in the Poisson curve.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-google-flights.png\" alt=\"Google Flights defaults to hump in Poisson distribution fo flight dates\" \/><\/figure><p><strong>Date text inputs<\/strong> are probably the best option for high-variability dates, where (A) there\u2019s no reason to favor any date over another, meaning (B) all options will be equally difficult to select.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-date-text.png\" alt=\"date text input instead of dropdown control\" \/><figcaption class=\"post__figcaption\">Remember, input[type=date] is your friend\u2026 on desktop, at\u00a0least<\/figcaption><\/figure><h3 id=\"if-youre-choosing-anumber\">If you\u2019re choosing a\u00a0number\u2026<\/h3>\n<p>Numbers come in all kinds of flavors, but you\u2019re most likely to be tempted to use dropdowns when you\u2019re dealing with <strong><em>counts\u200a<\/em><\/strong>-\u200ae.g. the number of <em>tickets<\/em>, the number of <em>people<\/em>, the number of <em>rooms<\/em>, etc.<\/p>\n<p>How often do you need 1 ticket? <em>Plenty<\/em>.<\/p>\n<p>How often do you need 10 tickets? <em>Not so much<\/em>.<\/p>\n<p>How often do you need 10,000 tickets? <em>Is this some kind of cruel joke?<\/em><\/p>\n<p>For counts of things, you\u2019re also dealing with Poisson distributions, and should use a control that biases towards lower numbers\u200a-\u200alike a <strong>stepper<\/strong>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-number-stepper.png\" alt=\"stepper control instead of dropdown control\" \/><\/figure><p>For wide-range numbers (like, say, SSNs), you weren\u2019t going to use a dropdown anyways\u2026 <em>I hope<\/em>.<\/p>\n<h3 id=\"so-can-i-ever-use-a-dropdown\">So can I ever use a dropdown?<\/h3>\n<p>Sure.<\/p>\n<p>Remember, they work OK when\u2026<\/p>\n<ul><li>Users <strong>rarely need to change the default<\/strong> value<\/li>\n  <li>There are <strong>very few options<\/strong>\u200a-\u200ae.g. only 3 will be visible on the default iOS control<\/li>\n  <li>The user is <strong>not on mobile<\/strong> (whereby many of these problems are mitigated)<\/li>\n<\/ul><p>The particularly observant among you may have noticed that the Google Flights interface I lauded above actually has <strong>three prominent dropdowns<\/strong>!<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/dropdown-google-flights.gif\" alt=\"dropdown controls on google flights\" \/><figcaption class=\"post__figcaption\">Brilliant detail: on mobile, the 'Economy' dropdown is\u00a0removed.<\/figcaption><\/figure><p>They actually do a great job with this. The potential usability issues are swiftly mitigated with:<\/p>\n<ul><li><strong>Custom controls<\/strong> that show all options on tap (including on mobile) \u2013 and replace 4 dropdowns (for Adults, Children, and Seated Infants and Lap Infants) with <strong>4 steppers in a single dropdown<\/strong>.<\/li>\n  <li><strong>Removing<\/strong> the \u201cEconomy\u201d dropdown on mobile<\/li>\n  <li><strong>Few options<\/strong> and <strong>smart defaults<\/strong> for each control<\/li>\n<\/ul><p>If you want to print this section out and stick it on your wall, I\u2019ve created a <a href=\"https:\/\/learnui.design\/extras\/selection-controls-ux-checklist.html\" target=\"_blank\">printable cheatsheet<\/a> of dropdown replacements.<\/p>\n<p>Anyhow. Let\u2019s move on.<\/p>\n<h2 id=\"3-pass-the-squint-test\">3. Pass the Squint Test<\/h2>\n<p><em>If you squint your eyes, the Most Important Thing should catch your eye first\u200a-\u200aand the least important elements should catch your eye last.<\/em><\/p>\n<p>Pop quiz: what does a user need to do to <em>use this page<\/em>?<\/p>\n<p>(NB: I\u2019ve blurred it out so you have to go by gut instinct, but it\u2019s a data entry form, to give you a hint)<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-ak1-blur.png\" alt=\"blurred out version of a train ticketing ui\" \/><\/figure><p>My best guess is <strong>two things<\/strong>:<\/p>\n<ol><li>Check any <strong>applicable checkboxes<\/strong> (??) in the yellow area<\/li>\n  <li>Press the <strong>blue \u201cSubmit\u201d button<\/strong><\/li>\n<\/ol><p>Did you guess the same?<\/p>\n<p>Wrong and wrong.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-ak1.png\" alt=\"train ticketing ui violates squint test\" \/><\/figure><ol><li>The \u201ccheckboxes\u201d are actually very small numerical text inputs. (If you already read Anything But Dropdowns, you know Poisson numbers should be steppers)<\/li>\n  <li>The <strong>Most Important Thing<\/strong> (\u201cFind Options\u201d \u2013 which is a very confusing way to say \u201cSubmit\u201d, by the way) is <strong>gray and unnoticeable<\/strong>. A much <em>less<\/em> important thing (\u201cHelp\u201d) is immediately next it, but bigger and <em>more<\/em> visible.<\/li>\n<\/ol><p>The Squint Test says the <em>Most Important Thing<\/em> must be the <em>most visible thing<\/em>. What\u2019s the MIT? The ticket textbox (or stepper \ud83d\ude09) controls and \u201cSubmit\u201d button.<\/p>\n<p>If you make it past this page, the next page is even worse.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-ak2-blur.png\" alt=\"blurred out version of a train ticketing ui\" \/><\/figure><p>What will you click: gray button the <em>left<\/em>, or identical gray button on the <em>right<\/em>?<\/p>\n<p>Hope you chose left!<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-ak2.png\" alt=\"train ticketing ui violating the squint test\" \/><figcaption class=\"post__figcaption\">In rushing through this form, I actually clicked 'help' first. Oops. My second time on this page, I clicked 'Go Back', having processed there was an 'Add' and 'Go Back' button, and in the other 99.999% of (left-to-right language) websites, 'Go Back' is always on the\u00a0left.<\/figcaption><\/figure><p>Again, <em>when I squint my eyes and look at the design, I can\u2019t tell what\u2019s important<\/em>.<\/p>\n<p>Like the Law of Locality and Anything But Dropdowns, the Squint Test is a fairly simple law to enforce. Here\u2019s like a 30-second wireframey redesign.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-ak2-redesign.png\" alt=\"wireframe redesign of a train ticketing ui to pass the squint test\" \/><figcaption class=\"post__figcaption\">In a real redesign, I'd also want to consider allowing the user to specify number of tickets ON THIS PAGE. But that's another law for another\u00a0time.<\/figcaption><\/figure><p>Does it work?<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-ak2-redesign-blur.png\" alt=\"blurred out version of a train ticketing ui wireframe reddesign to pass the squint test\" \/><\/figure><p>You tell me. Four radios and a button. And a tiny little link below it.<\/p>\n<p>I\u2019m not trying to pick on AlaskaTrain.com. You see this kind of stuff all over.<\/p>\n<p>Here\u2019s the signup screen for my beloved recommendation-based social network, Foursquare (blurred, of course).<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/foursquare-blur.png\" alt=\"blurred out foursquare ui\" \/><\/figure><p>How do you actually submit the required data? (i.e. the Most Important Thing)<\/p>\n<p>Hint: it\u2019s hidden in plain text in the upper-right corner.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-foursquare.png\" alt=\"foursquare ui redesigned to pass squint test\" \/><\/figure><p>But Foursquare is just following Apple\u2019s design standards here. Unfortunately, violating the Squint Test is a tradition even among industry leaders.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-calendar.png\" alt=\"ios calendar app failing the squint test\" \/><\/figure><p>One way to find the Most Important Thing is to consider what percentage of pageviews will involve a certain action. Here\u2019s flashcard\/memorization software Anki analyzed in this way.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-anki.png\" alt=\"action frequency analysis of Anki UI\" \/><\/figure><p>For every 100 flashcards I view, I will <em>then go on to\u2026<\/em><\/p>\n<ul><li>Show the answer (approx. 95 times)<\/li>\n  <li>Navigate back to the list of decks (twice)<\/li>\n  <li>Start adding cards (twice)<\/li>\n  <li>Use some other feature (very rarely)<\/li>\n<\/ul><p>This sort of analysis really hints at what kind of interface would work better here.<\/p>\n<ul><li><strong>Emphasize<\/strong> the most-commonly used functionality (at first approximation, \u201cmost used\u201d equals \u201cmost important\u201d)<\/li>\n  <li><strong>Deemphasize, hide, or remove<\/strong> the less commonly used functionality<\/li>\n<\/ul><figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/squint-anki-redesign.png\" alt=\"wireframe redesign of Anki UI to pass the squint test\" \/><\/figure><p>Now this is just a start (I\u2019d want to see if users understood that the unlabelled plus button <em>added cards<\/em>, for instance). But with just a couple simple heuristics, we\u2019ve reduced a cluttered, confusing interface of 10 UI elements down to just 5. A reduction of\u2026 check my math here\u2026 50%.<\/p>\n<p>For more on the Squint Test, check out my YouTube <a href=\"https:\/\/youtu.be\/B4XHaboNMOI\" target=\"_blank\">video redesign<\/a> of the Timezon.es web app. Or, if you don\u2019t have 10 minutes, here\u2019s a <a href=\"https:\/\/learnui.design\/blog\/squint-test-ui-design-case-study.html\">scannable, illustrated blog post<\/a> with the same step-by-step redesign.<\/p>\n<iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/B4XHaboNMOI\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>\n<h2 id=\"4-teach-by-example\">4. Teach by example<\/h2>\n<p><em>If you\u2019re introducing users to new concepts, a few examples can be worth 1000 words\u200a-\u200awhich your users wouldn\u2019t read, anyways.<\/em><\/p>\n<p>We have a weird tendency to try and explain things <em>in words<\/em> when <em>examples<\/em> would be much clearer.<\/p>\n<p>Consider the new startup Teeming.ai, who recently reached out to me to ask about their homepage design. Headlines on the page read:<\/p>\n<ul><li>\u201cTeeming takes the <strong>isolation out of remote work<\/strong>\u201d<\/li>\n  <li>\u201cTeeming helps with <strong>remote team building<\/strong>\u201d as well as \u201c<strong>learning, problem solving, having fun, and motivating each other<\/strong>\u201d<\/li>\n  <li>\u201cTeeming and video for <strong>synchronous [communication]<\/strong>\u201d<\/li>\n  <li>\u201cWorks with all your <strong>favorite video platforms<\/strong>\u201d<\/li>\n<\/ul><p>But here\u2019s my question for you. <strong><em>What does Teeming actually do?<\/em><\/strong><\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-teeming-1.png\" alt=\"teeming.ai UI doesn't teach by example\" \/><\/figure><p>It\u2019s tough to tell. I know it has something to do with\u2026 <em>good vibes for remote workers<\/em>? But I have no concrete idea how it would help me, so I wouldn\u2019t otherwise try it, recommend it, etc.<\/p>\n<p>(Sorry Teeming, you know I \u2764\ufe0f you)<\/p>\n<p>Next, let\u2019s look at <a href=\"https:\/\/ifttt.com\/\" target=\"_blank\">IFTTT<\/a>. Maybe you already know what they do\u200a-\u200ain which case, <em>pretend you don\u2019t<\/em>, and try to figure it out from these headlines on their homepage:<\/p>\n<ul><li>Automatically light the way for the pizza delivery guy (Dominoes+Hue)<\/li>\n  <li>Post your photo anywhere and see it everywhere (Instagram+twitter)<\/li>\n  <li>Make your voice assistant more personal (Google Assistant+iOS Calendar)<\/li>\n<\/ul><figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-ifttt-1.png\" alt=\"IFTTT UI teaches by example\" \/><\/figure><figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-ifttt-2.png\" alt=\"IFTTT UI teaches by example\" \/><\/figure><figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-ifttt-3.png\" alt=\"IFTTT UI teaches by example\" \/><\/figure><p>You don\u2019t have to list too many examples to paint a decently clear picture: <em>IFTTT hooks apps together to do things they couldn\u2019t do alone<\/em>.<\/p>\n<p>The crazy part is, if you visit their homepage, they first explain it in text:<\/p>\n<p><em>IFTTT helps your apps and devices work together in new ways. IFTTT is the free way to get all your apps and devices talking to each other. Not everything on the internet plays nice, so we\u2019re on a mission to build a more connected world.<\/em><\/p>\n<p>YAAAAWN.<\/p>\n<p>My question: <strong>which gives you a better idea of the app?<\/strong> The examples, or the description? \ud83e\udd14<\/p>\n<p>I think it\u2019s the <em>examples<\/em>. The description only resonates once I see a few examples of how it can help me.<\/p>\n<blockquote>\n  <p>The description of your complex new app\/feature only resonates once I see a few examples of how it can help\u00a0me.<\/p>\n<\/blockquote>\n<p>But examples aren\u2019t just for landing pages. Here\u2019s what you see when you first sign into project management tool <a href=\"https:\/\/basecamp.com\/\" target=\"_blank\">Basecamp<\/a>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-basecamp-1.png\" alt=\"Basecamp UI teaches by example\" \/><\/figure><p>Rather than seeing a totally blank page, you see <em>two obviously pre-fabricated example projects<\/em> that <strong>teach you, by example, how the whole app works<\/strong> (and also gives you an idea of what the tool will look and feel like when you\u2019ve been using it a while).<\/p>\n<p>Seriously, I can browse through fake <em>chat logs<\/em> by fake <em>users<\/em> discussing fake <em>file uploads<\/em> and fake <em>to-do items<\/em>.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-basecamp-2.png\" alt=\"Basecamp UI teaches by example\" \/><\/figure><p>There\u2019s even a friendly\u2026 <em>mountain?<\/em>\u2026 telling me I can watch a 2-minute explanatory video about this sample project.<\/p>\n<p>And thank you, Mr. Mountain, for the lead-in: <em>providing videos showing usage is another way of teaching by example<\/em>! Not only does the sample project model teach by example what my projects will look\/feel like, but the video teaches by example what it looks like to <em>use<\/em> the software.<\/p>\n<p>Brilliant.<\/p>\n<p>If your app allows users to <em>create<\/em> something, a <strong>showcase<\/strong> is a great way to teach by example just what\u2019s possible.<\/p>\n<p>The beloved painting app <a href=\"https:\/\/procreate.art\" target=\"_blank\">Procreate<\/a> won an Apple Design Award, the App Store Editor\u2019s Choice, the App Store Essential awards, and John Gruber called it \u201cgroundbreaking\u201d, etc. \u2013 and yet none of this is as <em>viscerally<\/em> exciting as seeing what you can create with it.<\/p>\n<figure class=\"post__figure\"><img src=\"https:\/\/learnui.design\/blog\/img\/4-rules\/tbe-procreate.png\" alt=\"Procreate gallery UI teaches by example\" \/><figcaption class=\"post__figcaption\">This ain't no ordinary painting\u00a0app.<\/figcaption><\/figure><p>Whoa.<\/p>\n<p>That\u2019s no MS Paint.<\/p>\n<p>The showcase is a <strong>powerful tool<\/strong> for making it clear just what\u2019s possible with your app.<\/p>\n<p>So: if your app does something new and unfamiliar \u2013 or relies on new and unfamiliar concepts \u2013 you should get acquainted with <strong>the ways of teaching by example<\/strong>. The <em>moment you realize<\/em> that you\u2019re introducing users won\u2019t have seen before, you should start thinking: <em>how can I give an example to make this clearer<\/em>?<\/p>\n<blockquote>\n  <p>The moment you realize that you\u2019re introducing users won\u2019t have seen before, you should start thinking: how can I give an example to make this\u00a0clearer?<\/p>\n<\/blockquote>\n<p>In review, my favorite ways of doing this:<\/p>\n<ol><li>On any page that tries to get the user to use a feature\/app\/etc., show examples of what they can do with your tool<\/li>\n  <li>Use the \u201cfirst load\u201d experience to provide sample data, showing by example what the properly-working app will look like<\/li>\n  <li>Strategically inject help content (like articles, videos, or tooltips) inline with the feature that show how to use it<\/li>\n  <li>Does your app allow users to create something? Include a user-submitted gallery of examples to spur imaginations<\/li>\n<\/ol><p>Make sense? Let\u2019s call it a day.<\/p>\n<hr \/><p>Alright, that wraps things up.<\/p>\n<p>There are plenty more rules for \u201cspeaking interface\u201d that I cover in my video course <a href=\"https:\/\/learnui.design\/courses\/learn-ux-design.html\">Learn UX Design<\/a>, but these are some of the ones that I\u2019ve used the most over the years. If you like these, check out more of my design writing on the Design Newsletter, where I send occasional, original design writing\u00a0\u2013 as well as updates when <a href=\"https:\/\/learnui.design\/courses\/learn-ux-design.html\">Learn UX Design<\/a> is open for enrollment.<\/p>\n<div class=\"magnet\">\n\t\n\t<div class=\"magnet__content\">\n\t    \n\t    \n\t    \n\t    \n\t    <p class=\"secondary\">Over 30,000 subscribed.<br \/>No spam. Unsubscribe anytime.<\/p>\n\t<\/div>\n<\/div>","created_at":"2019-08-26T14:27:17+0000","updated_at":"2019-08-26T14:27:20+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"learnui.design","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["career"],"is_public":false,"id":1007,"title":"Get your work recognized: write a brag document","url":"https:\/\/jvns.ca\/blog\/brag-documents\/","content":"<p>There\u2019s this idea that, if you do great work at your job, people will (or should!) automatically\nrecognize that work and reward you for it with promotions \/ increased pay. In practice, it\u2019s often\nmore complicated than that \u2013 some kinds of important work are more visible\/memorable than others.\nIt\u2019s frustrating to have done something really important and later realize that you didn\u2019t get\nrewarded for it just because the people making the decision didn\u2019t understand or remember what you\ndid. So I want to talk about a tactic that I and lots of people I work with have used!<\/p><p>This blog post isn\u2019t just about being promoted or getting raises though. The ideas here have\nactually been more useful to me to help me reflect on themes in my work, what\u2019s important to me,\nwhat I\u2019m learning, and what I\u2019d like to be doing differently. But they\u2019ve definitely helped with\npromotions!<\/p>\n<p>You can also <a href=\"#template\">skip to the brag document template at the end<\/a>.<\/p>\n<h3 id=\"you-don-t-remember-everything-you-did\">you don\u2019t remember everything you did<\/h3>\n<p>One thing I\u2019m always struck by when it comes to performance review time is a feeling of \u201cwait, what\n<em>did<\/em> I do in the last 6 months?\u201c. This is a kind of demoralizing feeling and it\u2019s usually not based\nin reality, more in \u201cI forgot what cool stuff I actually did\u201d.<\/p>\n<p>I invariably end up having to spend a bunch of time looking through my pull requests, tickets,\nlaunch emails, design documents, and more. I always end up finding small (and sometimes\nnot-so-small) things that I completely forgot I did, like:<\/p>\n<ul><li>mentored an intern 5 months ago<\/li>\n<li>did a small-but-important security project<\/li>\n<li>spent a few weeks helping get an important migration over the line<\/li>\n<li>helped X put together this design doc<\/li>\n<li>etcetera!<\/li>\n<\/ul><h3 id=\"your-manager-doesn-t-remember-everything-you-did\">your manager doesn\u2019t remember everything you did<\/h3>\n<p>And if you don\u2019t remember everything important you did, your manager (no matter how great they are!)\nprobably doesn\u2019t either. And they need to explain to other people why you should be promoted or\ngiven an evaluation like \u201cexceeds expectations\u201d (\u201cX\u2019s work is so awesome!!!!\u201d doesn\u2019t fly).<\/p>\n<p>So if your manager is going to effectively advocate for you, they need help.<\/p>\n<h3 id=\"here-s-the-tactic-write-a-document-listing-your-accomplishments\">here\u2019s the tactic: write a document listing your accomplishments<\/h3>\n<p>The tactic is pretty simple! Instead of trying to remember everything you did with your brain,\nmaintain a \u201cbrag document\u201d that lists everything so you can refer to it when you get to performance\nreview season! This is a pretty common tactic \u2013 when I started doing this I mentioned it to more\nexperienced people and they were like \u201coh yeah, I\u2019ve been doing that for a long time, it really\nhelps\u201d.<\/p>\n<p>Where I work we call this a \u201cbrag document\u201d but I\u2019ve heard other names for the same concept like\n\u201chype document\u201d or \u201clist of stuff I did\u201d :).<\/p>\n<p>There\u2019s a basic template for a brag document at the end of this post.<\/p>\n<h3 id=\"share-your-brag-document-with-your-manager\">share your brag document with your manager<\/h3>\n<p>When I first wrote a brag document I was kind of nervous about sharing it with my manager. It felt\nweird to be like \u201chey, uh, look at all the awesome stuff I did this year, I wrote a long document\nlisting everything\u201d. But my manager was really thankful for it \u2013 I think his perspective was \u201cthis\nmakes my job way easier, now I can look at the document when writing your perf review instead of\ntrying to remember what happened\u201d.<\/p>\n<p>Giving them a document that explains your accomplishments will really help your manager advocate for\nyou in discussions about your performance and come to any meetings they need to have prepared.<\/p>\n<p>Brag documents also <strong>really<\/strong> help with manager transitions \u2013 if you get a new manager 3 months\nbefore an important performance review that you want to do well on, giving them a brag document\noutlining your most important work &amp; its impact will help them understand what you\u2019ve been doing\neven though they may not have been aware of any of your work before.<\/p>\n<h3 id=\"share-it-with-your-peer-reviewers\">share it with your peer reviewers<\/h3>\n<p>Similarly, if your company does peer feedback as part of the promotion\/perf process \u2013 share your\nbrag document with your peer reviewers!! Every time someone shares their doc with me I find it SO\nHELPFUL with writing their review for much the same reasons it\u2019s helpful to share it with your\nmanager \u2013 it reminds me of all the amazing things they did, and when they list their goals in their\nbrag document it also helps me see what areas they might be most interested in feedback on.<\/p>\n<p>On some teams at work it\u2019s a team norm to share a brag document with peer reviewers to make it\neasier for them.<\/p>\n<h3 id=\"explain-the-big-picture\">explain the big picture<\/h3>\n<p>In addition to just listing accomplishments, in your brag document you can write the narrative explaining the big picture of your work. Have you been really focused on security? On building your product skills &amp; having really good relationships with your users? On building a strong culture of code review on the team?<\/p>\n<p>In my brag document, I like to do this by making a section for areas that I\u2019ve been focused on (like \u201csecurity\u201d) and listing all the work I\u2019ve done in that area there. This is especially good if you\u2019re working on something fuzzy like \u201cbuilding a stronger culture of code review\u201d where all the individual actions you do towards that might be relatively small and there isn\u2019t a big shiny ship.<\/p>\n<h3 id=\"use-your-brag-document-to-notice-patterns\">use your brag document to notice patterns<\/h3>\n<p>In the past I\u2019ve found the brag document useful not just to hype my accomplishments, but also to reflect on the work I\u2019ve done. Some questions it\u2019s helped me with:<\/p>\n<ul><li>What work do I feel most proud of?<\/li>\n<li>Are there themes in these projects I should be thinking about? What\u2019s the big picture of what I\u2019m working on? (am I working a lot on security? localization?).<\/li>\n<li>What do I wish I was doing more \/ less of?<\/li>\n<li>Which of my projects had the effect I wanted, and which didn\u2019t? Why might that have been?<\/li>\n<li>What could have gone better with project X? What might I want to do differently next time?<\/li>\n<\/ul><h3 id=\"you-can-write-it-all-at-once-or-update-it-every-2-weeks\">you can write it all at once or update it every 2 weeks<\/h3>\n<p>Many people have told me that it works best for them if they take a few minutes to update their brag document every 2 weeks ago. For me it actually works better to do a single marathon session every 6 months or every year where I look through everything I did and reflect on it all at once. Try out different approaches and see what works for you!<\/p>\n<h3 id=\"don-t-forget-to-include-the-fuzzy-work\">don\u2019t forget to include the fuzzy work<\/h3>\n<p>A lot of us work on fuzzy projects that can feel hard to quantify, like:<\/p>\n<ul><li>improving code quality on the team \/ making code reviews a little more in depth<\/li>\n<li>making on call easier<\/li>\n<li>building a more fair interview process \/ performance review system<\/li>\n<li>refactoring \/ driving down technical debt<\/li>\n<\/ul><p>A lot of people will leave this kind of work out because they don\u2019t know how to explain why it\u2019s important. But I think this kind of work is especially important to put into your brag document because it\u2019s the most likely to fall under the radar! One way to approach this is to, for each goal:<\/p>\n<ol><li>explain your goal for the work (why do you think it\u2019s important to refactor X piece of code?)<\/li>\n<li>list some things you\u2019ve done towards that goal<\/li>\n<li>list any effects you\u2019ve seen of the work, even if they\u2019re a little indirect<\/li>\n<\/ol><p>If you tell your coworkers this kind of work is important to you and tell them what you\u2019ve been\ndoing, maybe they can also give you ideas about how to do it more effectively or make the effects of\nthat work more obvious!<\/p>\n<h3 id=\"encourage-each-other-to-celebrate-accomplishments\">encourage each other to celebrate accomplishments<\/h3>\n<p>One nice side effect of having a shared idea that it\u2019s normal\/good to maintain a brag document at\nwork is that I sometimes see people encouraging each other to record &amp; celebrate their\naccomplishments (\u201chey, you should put that in your brag doc, that was really good!\u201d). It can be hard\nto see the value of your work sometimes, especially when you\u2019re working on something hard, and an\noutside perspective from a friend or colleague can really help you see why what you\u2019re doing is\nimportant.<\/p>\n<p>Brag documents are good when you use them on your own to advocate for yourself, but I think they\u2019re\nbetter as a collaborative effort to recognize where people are excelling.<\/p>\n<p>Next, I want to talk about a couple of structures that we\u2019ve used to help people recognize their accomplishments.<\/p>\n<h3 id=\"the-brag-workshop-help-people-list-their-accomplishments\">the brag workshop: help people list their accomplishments<\/h3>\n<p>The way this \u201cbrag document\u201d practice started in the first place is that my coworker <a href=\"https:\/\/karla.io\/\">Karla<\/a> and I wanted to help other women in engineering advocate for themselves more in the performance review process. The idea is that some people undersell their accomplishments more than they should, so we wanted to encourage those people to \u201cbrag\u201d a little bit and write down what they did that was important.<\/p>\n<p>We did this by running a \u201cbrag workshop\u201d just before performance review season. The format of the workshop is like this:<\/p>\n<p><strong>Part 1: write the document: 1-2 hours<\/strong>. Everybody sits down with their laptop, starts looking\nthrough their pull requests, tickets they resolved, design docs, etc, and puts together a list of\nimportant things they did in the last 6 months.<\/p>\n<p><strong>Part 2: pair up and make the impact of your work clearer: 1 hour<\/strong>. The goal of this part is to\npair up, review each other\u2019s documents, and identify places where people haven\u2019t bragged \u201cenough\u201d \u2013\nmaybe they worked on an extremely critical project to the company but didn\u2019t highlight how important\nit was, maybe they improved test performance but didn\u2019t say that they made the tests 3 times faster\nand that it improved everyone\u2019s developer experience. It\u2019s easy to accidentally write \u201cI shipped\n$feature\u201d and miss the follow up (\u201c\u2026 which caused $thing to happen\u201d). Another person reading\nthrough your document can help you catch the places where you need to clarify the impact.<\/p>\n<h3 id=\"biweekly-brag-document-writing-session\">biweekly brag document writing session<\/h3>\n<p>Another approach to helping people remember their accomplishments: my friend Dave gets some friends\ntogether every couple of weeks or so for everyone to update their brag documents. It\u2019s a nice way\nfor people to talk about work that they\u2019re happy about &amp; celebrate it a little bit, and updating\nyour brag document as you go can be easier than trying to remember everything you did all at once at\nthe end of the year.<\/p>\n<p>These don\u2019t have to be people in the same company or even in the same city \u2013 that group meets over\nvideo chat and has people from many different companies doing this together from Portland, Toronto,\nNew York, and Montreal.<\/p>\n<p>In general, especially if you\u2019re someone who really cares about your work, I think it\u2019s really\npositive to share your goals &amp; accomplishments (and the things that haven\u2019t gone so well too!) with\nyour friends and coworkers. It makes it feel less like you\u2019re working alone and more like\neveryone is supporting each other in helping them accomplish what they want.<\/p>\n<h3 id=\"thanks\">thanks<\/h3>\n<p>Thanks to Karla Burnett who I worked with on spreading this idea at work, to Dave Vasilevsky for\nrunning brag doc writing sessions, to Will Larson who encouraged me to start one <a href=\"https:\/\/lethain.com\/career-narratives\/\">of these<\/a> in the\nfirst place, to my manager Jay Shirley for always being encouraging &amp; showing me that this is a\nuseful way to work with a manager, and to Allie, Dan, Laura, Julian, Kamal, Stanley, and Vaibhav for\nreading a draft of this.<\/p>\n<p>I\u2019d also recommend the blog post <a href=\"http:\/\/blog.aashni.me\/2019\/01\/hype-yourself-youre-worth-it\/\">Hype Yourself! You\u2019re Worth It!<\/a> by Aashni Shah which talks about a similar approach.<\/p>\n<p>Here\u2019s a template for a brag document! Usually I make one brag document per year. (\u201cJulia\u2019s\n2017 brag document\u201d). I think it\u2019s okay to make it quite long \/ comprehensive \u2013 5-10 pages or more\nfor a year of work doesn\u2019t seem like too much to me, especially if you\u2019re including some\ngraphs\/charts \/ screenshots to show the effects of what you did.<\/p>\n<p>One thing I want to emphasize, for people who don\u2019t like to brag, is \u2013 <strong>you don\u2019t have to try to\nmake your work sound better than it is<\/strong>. Just make it sound <strong>exactly as good as it is<\/strong>! For\nexample \u201cwas the primary contributor to X new feature that\u2019s now used by 60% of our customers and\nhas gotten Y positive feedback\u201d.<\/p>\n<h3 id=\"goals-for-this-year\">Goals for this year:<\/h3>\n<ul><li>List your major goals here! Sharing your goals with your manager &amp; coworkers is really nice because it helps them see how they can support you in accomplishing those goals!<\/li>\n<\/ul><h3 id=\"goals-for-next-year\">Goals for next year<\/h3>\n<ul><li>If it\u2019s getting towards the end of the year, maybe start writing down what you think your goals for next year might be.<\/li>\n<\/ul><h3 id=\"projects\">Projects<\/h3>\n<p>For each one, go through:<\/p>\n<ul><li>What your contributions were (did you come up with the design? Which components did you build? Was there some useful insight like \u201cwait, we can cut scope and do what we want by doing way less work\u201d that you came up with?)<\/li>\n<li>The impact of the project \u2013 who was it for? Are there numbers you can attach to it? (saved X dollars? shipped new feature that has helped sell Y big deals? Improved performance by X%? Used by X internal users every day?). Did it support some important non-numeric company goal (required to pass an audit? helped retain an important user?)<\/li>\n<\/ul><p>Remember: don\u2019t forget to explain what the results of you work actually were! It\u2019s often important to go back a few months later and fill in what actually happened after you launched the project.<\/p>\n<h3 id=\"collaboration-mentorship\">Collaboration &amp; mentorship<\/h3>\n<p>Examples of things in this category:<\/p>\n<ul><li>Helping others in an area you\u2019re an expert in (like \u201cother engineers regularly ask me for one-off help\nsolving weird bugs in their CSS\u201d or \u201cquoting from the C standard at just the right moment\u201d)<\/li>\n<li>Mentoring interns \/ helping new team members get started<\/li>\n<li>Writing really clear emails\/meeting notes<\/li>\n<li>Foundational code that other people built on top of<\/li>\n<li>Improving monitoring \/ dashboards \/ on call<\/li>\n<li>Any code review that you spent a particularly long time on \/ that you think was especially important<\/li>\n<li>Important questions you answered (\u201chelped Risha from OTHER_TEAM with a lot of questions related to Y\u201d)<\/li>\n<li>Mentoring someone on a project (\u201cgave Ben advice from time to time on leading his first big project\u201d)<\/li>\n<li>Giving an internal talk or workshop<\/li>\n<\/ul><h3 id=\"design-documentation\">Design &amp; documentation<\/h3>\n<p>List design docs &amp; documentation that you worked on<\/p>\n<ul><li>Design docs: I usually just say \u201cwrote design for X\u201d or \u201creviewed design for X\u201d<\/li>\n<li>Documentation: maybe briefly explain the goal behind this documentation (for example \u201cwe were getting a lot of questions about X, so I documented it and now we can answer the questions more quickly\u201d)<\/li>\n<\/ul><h3 id=\"company-building\">Company building<\/h3>\n<p>This is a category we have at work \u2013 it basically means \u201cthings you did to help the company overall, not just your project \/ team\u201d. Some things that go in here:<\/p>\n<ul><li>Going above &amp; beyond with interviewing or recruiting (doing campus recruiting, etc)<\/li>\n<li>Improving important processes, like the interview process or writing better onboarding materials<\/li>\n<\/ul><h3 id=\"what-you-learned\">What you learned<\/h3>\n<p>My friend Julian suggested this section and I think it\u2019s a great idea \u2013 try listing important\nthings you learned or skills you\u2019ve acquired recently! Some examples of skills you might be\nlearning or improving:<\/p>\n<ul><li>how to do performance analysis &amp; make code run faster<\/li>\n<li>internals of an important piece of software (like the JVM or Postgres or Linux)<\/li>\n<li>how to use a library (like React)<\/li>\n<li>how to use an important tool (like the command line or Firefox dev tools)<\/li>\n<li>about a specific area of programming (like localization or timezones)<\/li>\n<li>an area like product management \/ UX design<\/li>\n<li>how to write a clear design doc<\/li>\n<li>a new programming language<\/li>\n<\/ul><p>It\u2019s really easy to lose track of what skills you\u2019re learning, and usually when I reflect on this I\nrealize I learned a lot more than I thought and also notice things that I\u2019m <em>not<\/em> learning that I\nwish I was.<\/p>\n<h3 id=\"outside-of-work\">Outside of work<\/h3>\n<p>It\u2019s also often useful to track accomplishments outside of work, like:<\/p>\n<ul><li>blog posts<\/li>\n<li>talks\/panels<\/li>\n<li>open source work<\/li>\n<li>Industry recognition<\/li>\n<\/ul><p>I think this can be a nice way to highlight how you\u2019re thinking about your career outside of strictly what you\u2019re doing at work.<\/p>\n<p>This can also include other non-career-related things you\u2019re proud of, if that feels good to you! Some people like to keep a combined personal + work brag document.<\/p>\n<h3 id=\"general-prompts\">General prompts<\/h3>\n<p>If you\u2019re feeling stuck for things to mention, try:<\/p>\n<ul><li>If you were trying to convince a friend to come join your company\/team, what would you tell them about your work?<\/li>\n<li>Did anybody tell you you did something well recently?<\/li>\n<\/ul>","created_at":"2019-08-26T14:22:57+0000","updated_at":"2019-08-26T14:22:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"jvns.ca","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["career","networking","leadership"],"is_public":false,"id":1006,"title":"How to manage up effectively - Know Your Team | Blog","url":"https:\/\/knowyourteam.com\/blog\/2019\/08\/15\/how-to-manage-up-effectively\/","content":"<h4>5 not-so-often-shared ways to manage up and have a better relationship with your boss.<\/h4>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2019\/08\/managing-up-1024x429.png\" alt=\"Managing up\" class=\"wp-image-4578\" srcset=\"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2019\/08\/managing-up-1024x429.png 1024w, https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2019\/08\/managing-up-300x126.png 300w, https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2019\/08\/managing-up-768x321.png 768w\" \/><\/figure><p>You want to manage up \u2013 but what you really mean is that you simply want to work well with your boss. <\/p>\n\n\n\n<p>Who doesn\u2019t? Especially when your boss is pestering you with questions via Slack after work-hours, or failing to give you enough time to complete projects\u2026 <\/p>\n\n\n\n<p>You sigh and think yourself:  \u201cHow do I manage up effectively?\u201d<\/p>\n\n\n\n<p>This is one of the most common questions I\u2019m asked \u2013 and, unfortunately, one of the most common situations that you might face, whether you\u2019re a manager, executive or individual contributor. <\/p>\n\n\n\n<p>Based on research we\u2019ve done over the past five years with hundreds of managers and employees, and the insights shared in our online leadership community, <a href=\"http:\/\/thewatercooler.io\/\">The Watercooler<\/a>, here are the 5 distinct ways you can manage up to have a better relationship with your boss:<\/p>\n\n\n\n<h3 id=\"share-progress-more-than-you-usually-would.\">Share progress more often than you typically would.<\/h3>\n\n\n\n<p>Your boss is craving to know what you\u2019re working on. We recently conducted a survey of 355 people and learned that the #1 piece of information that managers want to know is the progress that\u2019s being made on a project. As a result, you\u2019ll want to ask yourself:  <em>Am I sharing the progress I\u2019m making day-to-day or week-to-week? <\/em>You can also ask your boss directly:  \u201cHow can I give you more visibility into my work?\u201d or \u201cAre there any decisions or projects you wish I were more transparent about?\u201d Additionally, in <a href=\"https:\/\/knowyourteam.com\">Know Your Team<\/a> in fact, we have a tool called <a href=\"https:\/\/knowyourteam.com\/m\/features\">\u201cThe Heartbeat\u201d<\/a> that helps you do this with little to zero effort \u2013 it\u2019s an easy way to automate daily or weekly check-ins with your team. I highly recommend taking a peek if you feel like your team doesn\u2019t have an effective process for sharing progress.<\/p>\n\n\n\n<h3 id=\"uncover-your-boss'-work-preferences,-even-if-they-don't-do-the-same-for-you.\">Uncover your boss\u2019 work preferences (even if they don\u2019t do the same for you).<\/h3>\n\n\n\n<p>We all have preferences around the way we like to work\u00a0\u2013 including our boss. Make sure you spend time asking questions to your boss to figure out what theirs are. The more you know about how your boss likes to work (e.g., how they prefer to receive feedback, when they\u2019re most productive during the day), the more you can adjust your own behavior to increase the likelihood of a positive relationship with them.  It may seem frustrating (rightfully!) that you have to be the one to instigate this conversation, if they haven\u2019t already (they are <em>your<\/em> manager, after all) \u2013 but some people need the nudge. You can provide it \ud83d\ude42<\/p>\n\n\n\n<p>Here are some of my favorite questions to ask around uncovering someone\u2019s work preferences:<\/p>\n\n\n\n<ul><li>Where on the spectrum of extrovert to introvert would you place yourself?<\/li><li>What\u2019s your preferred way to receive feedback, in terms of format?<\/li><li>What\u2019s your preferred way to receive feedback, in terms of speed?<\/li><li>What\u2019s your orientation toward conflict?<\/li><li>What time of day are you most productive?<\/li><li>How would you describe your communication style?<\/li><li>What motivates you the most? <\/li><li>Who is your hero? Why?<\/li><li>What do you consider your \u201csuper power\u201d?<\/li><li>Who\u2019s been the best coworker \/ team you\u2019ve worked with? Why? <\/li><li>Who\u2019s been the best boss \/ mentor you\u2019ve ever had? Why?<\/li><li>When have you worked with someone and noticed it not going well? <\/li><li>What do you think you\u2019re more sensitive about, compared to others?<\/li><li>What\u2019s your biggest work-related pet peeves (i.e., that thing other people do that totally annoys you when you work with them)? <\/li><li>What would others who\u2019ve worked with you say are your greatest strengths? <\/li><li>What would others who\u2019ve worked with you say are your greatest weaknesses?<\/li><\/ul><p> (Note:  These are also fantastic questions to ask when <a href=\"https:\/\/knowyourteam.com\/blog\/2019\/06\/20\/how-to-onboard-new-hire\/\">onboarding a new hire<\/a>, in general, and not just to your boss. )<\/p>\n\n\n\n<h3 id=\"show-how-absurd-it'd-be-not-to-trust-you.\">Build trust through your actions, not just your words.<\/h3>\n\n\n\n<p>Let\u2019s get real:  Your boss might not <a href=\"https:\/\/knowyourteam.com\/blog\/2019\/01\/15\/building-trust-in-teams-what-and-why\/\">trust<\/a> you. Oftentimes, a boss\u2019 <a href=\"https:\/\/knowyourteam.com\/blog\/2018\/11\/29\/you-have-a-micromanaging-boss-what-can-you-do\/\">micromanaging tendencies<\/a> are derived from their lack of trust in your ability to follow through on something. Now, this may not be your fault by any means \u2013 your manager could be projecting their own past experiences onto you. But it does reveal that if you want a strong relationship with your boss, you must enable them to feel they can trust in your capabilities. <\/p>\n\n\n\n<p>To be clear: You\u2019re not trying to blow smoke or curry favor with your boss. Trust does not mean likability. Rather, you should be focused making sure your behavior matches your intention \u2013 that\u2019s <a href=\"https:\/\/knowyourteam.com\/blog\/2019\/02\/12\/the-3-most-effective-ways-to-build-trust-as-a-leader\/\">the best way to build trust<\/a>. Your boss is going to give you more space the more you show them that they can give you that space. This means taking a close look and asking yourself:  <em>How well do I follow through on my word? <\/em>The more closely that your actions mirror your words, the more likely your manager will have a greater degree of trust in you.<\/p>\n\n\n\n<h3 id=\"foster-real-rapport-\u2013-not-the-superficial-stuff.\">Foster real rapport.<\/h3>\n\n\n\n<p>Intuitively, we tend to have more positive relationships with people we feel interpersonally connected with. So, looking to form some sort of rapport with your boss is helpful to manage up, as well. This doesn\u2019t mean you need to become best friends with them, do a whole in-depth reveal on your personal life, or ask them a bunch of corny icebreaker questions. Instead, it means you can share about what makes you tick, what you\u2019re passionate about, what you like to do outside of work, etc\u2026 And, you can ask questions to uncover similar aspects of who your boss is, too. Here are some of my favorite questions to ask that you can consider asking over a lunch or coffee chat with your boss, beyond the typical icebreakers:<\/p>\n\n\n\n<ul><li>What was your first job?<\/li><li>What are you reading right now?<\/li><li>Who\u2019s someone you really admire?<\/li><li>Seen any good movies lately you\u2019d recommend?<\/li><li>Who had the most influence on you growing up?<\/li><li>What\u2019s the best advice you\u2019ve ever heard?<\/li><li>Been pleasantly surprised by anything lately?<\/li><\/ul><h3 id=\"get-crystal-clear-on-expectations.\">Get crystal clear on expectations.<\/h3>\n\n\n\n<p>Possibly the most constructive way to manage up is to make sure you\u2019re clear on what your manager expects of you. Without clear expectations, naturally, it\u2019s easy to step on each other\u2019s toes, misinterpret a comment, or be offended by a request. Instead, make it your mission to get clear on two things in particular:  (1) What \u201csuccess\u201d looks like (2) How to communicate well to achieve \u201csuccess.\u201d  Here are some questions you can ask to do this\u2026<\/p>\n\n\n\n<ul><li>What things will I need have accomplished this year for you to view me in this role as \u201csuccessful\u201d?<\/li><li>What does quality work look like to you?<\/li><li>If I need your attention on something urgent, what form of communication do you prefer? (E.g., Slack, an email, a phone call)<\/li><li>What do you consider \u201curgent\u201d versus \u201cnot urgent\u201d?<\/li><li>If I have any feedback, how would they prefer you receive it: In person, in writing, over the phone, or during a video call?<\/li><li>If I have a suggestion about something, how would you prefer you receive it: In writing, in person, over the phone, or during a video call?<\/li><li>If I have a question, what form of communication should I pose the questions to you? (E.g., In writing, in person, over the phone, during a video call)<\/li><\/ul><h3 id=\"dissent-is-your-obligation\">Remind yourself:  Dissent is an obligation.<\/h3>\n\n\n\n<p>Managing up also means speaking up. Your manager is never going to know where you\u2019re coming from and what you think  if you don\u2019t vocalize it. This can be daunting, not to mention implausible, if your manager isn\u2019t receptive to feedback or your workplace environment is toxic. However, if you do perceive even a slight chance that your manager might be receptive, I suggest you take it \u2013\u00a0feedback can be delivered both honestly and kindly. To <a href=\"https:\/\/knowyourteam.com\/blog\/2018\/10\/15\/how-to-deliver-negative-feedback-well-the-4-things-that-good-managers-do-that-bad-managers\/\">deliver difficult feedback well<\/a>, here are a few phrases you can use\u2026<\/p>\n\n\n\n<ul><li><em>\u201cWould you be open to hearing some quick feedback around a few things I noticed?\u201d<\/em><\/li><li><em>\u201cI heard a few things on a call the other day that I thought we could talk through together\u200a\u2014\u200awould you be open to that?\u201d<\/em><\/li><li><em>\u201cHappen to have time later today to chat around a few things I saw?\u201d<\/em><\/li><li><em>\u201cWould you want to sit down and talk about different ways we can both improve?\u201d<\/em><\/li><li><em>\u201cI\u2019m sharing this because I want to team to succeed\u2026\u201d<\/em><\/li><li><em>\u201cThis is important to me because I care about the company\u2019s direction as a whole\u2026\u201d<\/em><\/li><\/ul><p>Take note:  This is a general map, not a step-by-step guide, to manage up. You\u2019ll have to assess, for yourself, your own manager\u2019s propensity to be open to some of these questions and conversations. <\/p>\n\n\n\n<p>However, most broadly, thinking about these 5 aspects \u2013\u00a0sharing progress, uncovering work preferences, building both trust and rapport, clarifying expectations, and sharing feedback\u00a0\u2013 are things you can experiment with as you aspire to have a strong working relationship with your boss. <\/p>\n\n\n\n<p>The ball is in your court, more than you might think.<\/p>\n\n\n\n<div class=\"newsletter-post-form-container\">\n<h3>Enjoy this? Get our latest leadership insights, as they\u2019re released.<\/h3>\n\n<form class=\"newsletter-form newsletter-post-form\" action=\"https:\/\/www.getdrip.com\/forms\/855299083\/submissions\" method=\"post\" data-drip-embedded-form=\"855299083\">\n  <div><input type=\"email\" id=\"drip-email\" name=\"fields[email]\" value=\"\" placeholder=\"Type your email address...\" \/><input type=\"submit\" value=\"Subscribe now\" data-drip-attribute=\"sign-up-button\" \/><\/div><\/form>\n<\/div>\n\n\n<h3 class=\"crp-list-title\">You might also enjoy reading\u2026<\/h3><ul class=\"crp-list\"><li class=\"crp-list-item crp-list-item-image-above crp-list-item-has-image\"><div class=\"crp-list-item-image\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/10\/12\/new-managers-heres-how-to-run-your-first-team-meeting\/\"><img width=\"50\" height=\"50\" src=\"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2018\/12\/143df-1-lfopueazajaomtpwiknug-150x150.jpeg\" class=\"attachment-50x50 size-50x50 wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"crp-list-item-title\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/10\/12\/new-managers-heres-how-to-run-your-first-team-meeting\/\">New managers, here\u2019s how to run your first team meeting<\/a><\/div><\/li><li class=\"crp-list-item crp-list-item-image-above crp-list-item-has-image\"><div class=\"crp-list-item-image\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/02\/01\/the-5-ways-to-improve-your-self-awareness-as-a-leader\/\"><img width=\"50\" height=\"50\" src=\"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2018\/02\/self-awareness-150x150.jpg\" class=\"attachment-50x50 size-50x50 wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"crp-list-item-title\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/02\/01\/the-5-ways-to-improve-your-self-awareness-as-a-leader\/\">The 5 ways to improve your self-awareness as a leader<\/a><\/div><\/li><li class=\"crp-list-item crp-list-item-image-above crp-list-item-has-image\"><div class=\"crp-list-item-image\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/01\/11\/the-8-best-questions-to-ask-during-a-one-on-one-meeting\/\"><img width=\"50\" height=\"50\" src=\"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2018\/12\/22e24-1fhcjchvmp0qczswldyvycw-150x150.jpeg\" class=\"attachment-50x50 size-50x50 wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"crp-list-item-title\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/01\/11\/the-8-best-questions-to-ask-during-a-one-on-one-meeting\/\">The 8 best questions to ask during a one-on-one meeting<\/a><\/div><\/li><li class=\"crp-list-item crp-list-item-image-above crp-list-item-has-image\"><div class=\"crp-list-item-image\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/11\/29\/you-have-a-micromanaging-boss-what-can-you-do\/\"><img width=\"50\" height=\"50\" src=\"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2018\/12\/d09e6-1jchzvkpvy6581ls1nq0d-g-150x150.jpeg\" class=\"attachment-50x50 size-50x50 wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"crp-list-item-title\"><a href=\"https:\/\/knowyourteam.com\/blog\/2018\/11\/29\/you-have-a-micromanaging-boss-what-can-you-do\/\">You have a micromanaging boss. What can you do?<\/a><\/div><\/li><\/ul><div class=\"author-profile\">\n\n\t\t\t<div class=\"author-profile-avatar\">\n\t\t\t\t<a href=\"https:\/\/knowyourteam.com\/blog\/author\/claire\/\" title=\"Posts by Claire Lew\">\n\t\t\t\t\t<img alt=\"\" src=\"https:\/\/secure.gravatar.com\/avatar\/c35e01544a0dd94a2cf1619ee8a42ebb?s=96&amp;r=g\" srcset=\"https:\/\/secure.gravatar.com\/avatar\/c35e01544a0dd94a2cf1619ee8a42ebb?s=192&amp;r=g 2x\" class=\"avatar avatar-96 photo\" height=\"96\" width=\"96\" \/><\/a>\n\t\t\t<\/div>\n\n\t\t\t<div class=\"author-profile-info\">\n\t\t\t\t<h3 class=\"author-profile-title\">\n\t\t\t\t\t\t\t\t\t\t\tWritten by Claire Lew\t\t\t\t\t\t\t\t\t<\/h3>\n\n\t\t\t\t<div class=\"author-description\">\n\t\t\t\t\t<p>CEO of Know Your Team. My mission in life is to help people become happier at work. Say hi to me on Twitter at @clairejlew.<\/p>\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>","created_at":"2019-08-26T14:05:53+0000","updated_at":"2019-08-26T14:06:06+0000","published_at":"2019-08-15T17:12:48+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"knowyourteam.com","preview_picture":"https:\/\/knowyourteam.com\/blog\/wp-content\/uploads\/2019\/08\/managing-up-1024x429.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["k8s","tutorial"],"is_public":false,"id":1005,"title":"An Introduction to Kubernetes","url":"https:\/\/www.digitalocean.com\/community\/tutorials\/an-introduction-to-kubernetes","content":"<p>Kubernetes is a powerful open-source system, initially developed by Google, for managing containerized applications in a clustered environment.  It aims to provide better ways of managing related, distributed components and services across varied infrastructure.<\/p><p>In this guide, we\u2019ll discuss some of Kubernetes\u2019 basic concepts.  We will talk about the architecture of the system, the problems it solves, and the model that it uses to handle containerized deployments and scaling.<\/p><h2 id=\"what-is-kubernetes\">What is Kubernetes?<\/h2><p><strong>Kubernetes<\/strong>, at its basic level, is a system for running and coordinating containerized applications across a cluster of machines.  It is a platform designed to completely manage the life cycle of containerized applications and services using methods that provide predictability, scalability, and high availability.<\/p><p>As a Kubernetes user, you can define how your applications should run and the ways they should be able to interact with other applications or the outside world.  You can scale your services up or down, perform graceful rolling updates, and switch traffic between different versions of your applications to test features or rollback problematic deployments.  Kubernetes provides interfaces and composable platform primitives that allow you to define and manage your applications with high degrees of flexibility, power, and reliability.<\/p><h2 id=\"kubernetes-architecture\">Kubernetes Architecture<\/h2><p>To understand how Kubernetes is able to provide these capabilities, it is helpful to get a sense of how it is designed and organized at a high level.  Kubernetes can be visualized as a system built in layers, with each higher layer abstracting the complexity found in the lower levels.<\/p><p>At its base, Kubernetes brings together individual physical or virtual machines into a cluster using a shared network to communicate between each server.  This cluster is the physical platform where all Kubernetes components, capabilities, and workloads are configured.<\/p><p>The machines in the cluster are each given a role within the Kubernetes ecosystem.  One server (or a small group in highly available deployments) functions as the <strong>master<\/strong> server.  This server acts as a gateway and brain for the cluster by exposing an API for users and clients, health checking other servers, deciding how best to split up and assign work (known as \u201cscheduling\u201d), and orchestrating communication between other components.  The master server acts as the primary point of contact with the cluster and is responsible for most of the centralized logic Kubernetes provides.<\/p><p>The other machines in the cluster are designated as <strong>nodes<\/strong>: servers responsible for accepting and running workloads using local and external resources.  To help with isolation, management, and flexibility, Kubernetes runs applications and services in <strong>containers<\/strong>, so each node needs to be equipped with a container runtime (like Docker or rkt).  The node receives work instructions from the master server and creates or destroys containers accordingly, adjusting networking rules to route and forward traffic appropriately.<\/p><p>As mentioned above, the applications and services themselves are run on the cluster within containers.  The underlying components make sure that the desired state of the applications matches the actual state of the cluster.  Users interact with the cluster by communicating with the main API server either directly or with clients and libraries.  To start up an application or service, a declarative plan is submitted in JSON or YAML defining what to create and how it should be managed.  The master server then takes the plan and figures out how to run it on the infrastructure by examining the requirements and the current state of the system.  This group of user-defined applications running according to a specified plan represents Kubernetes\u2019 final layer.<\/p><h2 id=\"master-server-components\">Master Server Components<\/h2><p>As we described above, the master server acts as the primary control plane for Kubernetes clusters.  It serves as the main contact point for administrators and users, and also provides many cluster-wide systems for the relatively unsophisticated worker nodes.  Overall, the components on the master server work together to accept user requests, determine the best ways to schedule workload containers, authenticate clients and nodes, adjust cluster-wide networking, and manage scaling and health checking responsibilities.<\/p><p>These components can be installed on a single machine or distributed across multiple servers.  We will take a look at each of the individual components associated with master servers in this section.<\/p><h3 id=\"etcd\">etcd<\/h3><p>One of the fundamental components that Kubernetes needs to function is a globally available configuration store.  The <a href=\"https:\/\/coreos.com\/etcd\/docs\/latest\/\"><strong>etcd<\/strong> project<\/a>, developed by the team at CoreOS, is a lightweight, distributed key-value store that can be configured to span across multiple nodes.<\/p><p>Kubernetes uses <code>etcd<\/code> to store configuration data that can be accessed by each of the nodes in the cluster.  This can be used for service discovery and can help components configure or reconfigure themselves according to up-to-date information.  It also helps maintain cluster state with features like leader election and distributed locking.  By providing a simple HTTP\/JSON API, the interface for setting or retrieving values is very straight forward.<\/p><p>Like most other components in the control plane, <code>etcd<\/code> can be configured on a single master server or, in production scenarios, distributed among a number of machines.  The only requirement is that it be network accessible to each of the Kubernetes machines.<\/p><h3 id=\"kube-apiserver\">kube-apiserver<\/h3><p>One of the most important master services is an API server.  This is the main management point of the entire cluster as it allows a user to configure Kubernetes\u2019 workloads and organizational units.  It is also responsible for making sure that the <code>etcd<\/code> store and the service details of deployed containers are in agreement.  It acts as the bridge between various components to maintain cluster health and disseminate information and commands.<\/p><p>The API server implements a RESTful interface, which means that many different tools and libraries can readily communicate with it.  A client called <strong>kubectl<\/strong> is available as a default method of interacting with the Kubernetes cluster from a local computer.<\/p><h3 id=\"kube-controller-manager\">kube-controller-manager<\/h3><p>The controller manager is a general service that has many responsibilities.  Primarily, it manages different controllers that regulate the state of the cluster, manage workload life cycles, and perform routine tasks.  For instance, a replication controller ensures that the number of replicas (identical copies) defined for a pod matches the number currently deployed on the cluster. The details of these operations are written to <code>etcd<\/code>, where the controller manager watches for changes through the API server.<\/p><p>When a change is seen, the controller reads the new information and implements the procedure that fulfills the desired state.  This can involve scaling an application up or down, adjusting endpoints, etc.<\/p><h3 id=\"kube-scheduler\">kube-scheduler<\/h3><p>The process that actually assigns workloads to specific nodes in the cluster is the scheduler.  This service reads in a workload\u2019s operating requirements, analyzes the current infrastructure environment, and places the work on an acceptable node or nodes.<\/p><p>The scheduler is responsible for tracking available capacity on each host to make sure that workloads are not scheduled in excess of the available resources.  The scheduler must know the total capacity as well as the resources already allocated to existing workloads on each server.<\/p><h3 id=\"cloud-controller-manager\">cloud-controller-manager<\/h3><p>Kubernetes can be deployed in many different environments and can interact with various infrastructure providers to understand and manage the state of resources in the cluster.  While Kubernetes works with generic representations of resources like attachable storage and load balancers, it needs a way to map these to the actual resources provided by non-homogeneous cloud providers.<\/p><p>Cloud controller managers act as the glue that allows Kubernetes to interact providers with different capabilities, features, and APIs while maintaining relatively generic constructs internally.  This allows Kubernetes to update its state information according to information gathered from the cloud provider, adjust cloud resources as changes are needed in the system, and create and use additional cloud services to satisfy the work requirements submitted to the cluster.<\/p><h2 id=\"node-server-components\">Node Server Components<\/h2><p>In Kubernetes, servers that perform work by running containers are known as <strong>nodes<\/strong>.  Node servers have a few requirements that are necessary for communicating with master components, configuring the container networking, and running the actual workloads assigned to them.<\/p><h3 id=\"a-container-runtime\">A Container Runtime<\/h3><p>The first component that each node must have is a container runtime.  Typically, this requirement is satisfied by installing and running <a href=\"https:\/\/www.docker.com\/\">Docker<\/a>, but alternatives like <a href=\"https:\/\/coreos.com\/rkt\/\">rkt<\/a> and <a href=\"https:\/\/github.com\/opencontainers\/runc\">runc<\/a> are also available.<\/p><p>The container runtime is responsible for starting and managing containers, applications encapsulated in a relatively isolated but lightweight operating environment.  Each unit of work on the cluster is, at its basic level, implemented as one or more containers that must be deployed.  The container runtime on each node is the component that finally runs the containers defined in the workloads submitted to the cluster.<\/p><h3 id=\"kubelet\">kubelet<\/h3><p>The main contact point for each node with the cluster group is a small service called <strong>kubelet<\/strong>.  This service is responsible for relaying information to and from the control plane services, as well as interacting with the <code>etcd<\/code> store to read configuration details or write new values.<\/p><p>The <code>kubelet<\/code> service communicates with the master components to authenticate to the cluster and receive commands and work.  Work is received in the form of a <strong>manifest<\/strong> which defines the workload and the operating parameters.  The <code>kubelet<\/code> process then assumes responsibility for maintaining the state of the work on the node server.  It controls the container runtime to launch or destroy containers as needed.<\/p><h3 id=\"kube-proxy\">kube-proxy<\/h3><p>To manage individual host subnetting and make services available to other components, a small proxy service called <strong>kube-proxy<\/strong> is run on each node server.  This process forwards requests to the correct containers, can do primitive load balancing, and is generally responsible for making sure the networking environment is predictable and accessible, but isolated where appropriate.<\/p><h2 id=\"kubernetes-objects-and-workloads\">Kubernetes Objects and Workloads<\/h2><p>While containers are the underlying mechanism used to deploy applications, Kubernetes uses additional layers of abstraction over the container interface to provide scaling, resiliency, and life cycle management features.  Instead of managing containers directly, users define and interact with instances composed of various primitives provided by the Kubernetes object model. We will go over the different types of objects that can be used to define these workloads below.<\/p><h3 id=\"pods\">Pods<\/h3><p>A <strong>pod<\/strong> is the most basic unit that Kubernetes deals with.  Containers themselves are not assigned to hosts.  Instead, one or more tightly coupled containers are encapsulated in an object called a pod.<\/p><p>A pod generally represents one or more containers that should be controlled as a single application.  Pods consist of containers that operate closely together, share a life cycle, and should always be scheduled on the same node.  They are managed entirely as a unit and share their environment, volumes, and IP space.  In spite of their containerized implementation, you should generally think of pods as a single, monolithic application to best conceptualize how the cluster will manage the pod\u2019s resources and scheduling.<\/p><p>Usually, pods consist of a main container that satisfies the general purpose of the workload and optionally some helper containers that facilitate closely related tasks.  These are programs that benefit from being run and managed in their own containers, but are tightly tied to the main application.  For example, a pod may have one container running the primary application server and a helper container pulling down files to the shared filesystem when changes are detected in an external repository.  Horizontal scaling is generally discouraged on the pod level because there are other higher level objects more suited for the task.<\/p><p>Generally, users should not manage pods themselves, because they do not provide some of the features typically needed in applications (like sophisticated life cycle management and scaling).  Instead, users are encouraged to work with higher level objects that use pods or pod templates as base components but implement additional functionality.<\/p><h3 id=\"replication-controllers-and-replication-sets\">Replication Controllers and Replication Sets<\/h3><p>Often, when working with Kubernetes, rather than working with single pods, you will instead be managing groups of identical, replicated pods.  These are created from pod templates and can be horizontally scaled by controllers known as replication controllers and replication sets.<\/p><p>A <strong>replication controller<\/strong> is an object that defines a pod template and control parameters to scale identical replicas of a pod horizontally by increasing or decreasing the number of running copies.  This is an easy way to distribute load and increase availability natively within Kubernetes.  The replication controller knows how to create new pods as needed because a template that closely resembles a pod definition is embedded within the replication controller configuration.<\/p><p>The replication controller is responsible for ensuring that the number of pods deployed in the cluster matches the number of pods in its configuration.  If a pod or underlying host fails, the controller will start new pods to compensate.  If the number of replicas in a controller\u2019s configuration changes, the controller either starts up or kills containers to match the desired number.  Replication controllers can also perform rolling updates to roll over a set of pods to a new version one by one, minimizing the impact on application availability.<\/p><p><strong>Replication sets<\/strong> are an iteration on the replication controller design with greater flexibility in how the controller identifies the pods it is meant to manage.  Replication sets are beginning to replace replication controllers because of their greater replica selection capabilities, but they are not able to do rolling updates to cycle backends to a new version like replication controllers can.  Instead, replication sets are meant to be used inside of additional, higher level units that provide that functionality.<\/p><p>Like pods, both replication controllers and replication sets are rarely the units you will work with directly.  While they build on the pod design to add horizontal scaling and reliability guarantees, they lack some of the fine grained life cycle management capabilities found in more complex objects.<\/p><h3 id=\"deployments\">Deployments<\/h3><p><strong>Deployments<\/strong> are one of the most common workloads to directly create and manage.  Deployments use replication sets as a building block, adding flexible life cycle management functionality to the mix.<\/p><p>While deployments built with replications sets may appear to duplicate the functionality offered by replication controllers, deployments solve many of the pain points that existed in the implementation of rolling updates.  When updating applications with replication controllers, users are required to submit a plan for a new replication controller that would replace the current controller.  When using replication controllers, tasks like tracking history, recovering from network failures during the update, and rolling back bad changes are either difficult or left as the user\u2019s responsibility.<\/p><p>Deployments are a high level object designed to ease the life cycle management of replicated pods.  Deployments can be modified easily by changing the configuration and Kubernetes will adjust the replica sets, manage transitions between different application versions, and optionally maintain event history and undo capabilities automatically.  Because of these features, deployments will likely be the type of Kubernetes object you work with most frequently.<\/p><h3 id=\"stateful-sets\">Stateful Sets<\/h3><p><strong>Stateful sets<\/strong> are specialized pod controllers that offer ordering and uniqueness guarantees.  Primarily, these are used to have more fine-grained control when you have special requirements related to deployment ordering, persistent data, or stable networking.  For instance, stateful sets are often associated with data-oriented applications, like databases, which need access to the same volumes even if rescheduled to a new node.<\/p><p>Stateful sets provide a stable networking identifier by creating a unique, number-based name for each pod that will persist even if the pod needs to be moved to another node.  Likewise, persistent storage volumes can be transferred with a pod when rescheduling is necessary.  The volumes persist even after the pod has been deleted to prevent accidental data loss.<\/p><p>When deploying or adjusting scale, stateful sets perform operations according to the numbered identifier in their name.  This gives greater predictability and control over the order of execution, which can be useful in some cases.<\/p><h3 id=\"daemon-sets\">Daemon Sets<\/h3><p><strong>Daemon sets<\/strong> are another specialized form of pod controller that run a copy of a pod on each node in the cluster (or a subset, if specified).  This is most often useful when deploying pods that help perform maintenance and provide services for the nodes themselves.<\/p><p>For instance, collecting and forwarding logs, aggregating metrics, and running services that increase the capabilities of the node itself are popular candidates for daemon sets.  Because daemon sets often provide fundamental services and are needed throughout the fleet, they can bypass pod scheduling restrictions that prevent other controllers from assigning pods to certain hosts.  As an example, because of its unique responsibilities, the master server is frequently configured to be unavailable for normal pod scheduling, but daemon sets have the ability to override the restriction on a pod-by-pod basis to make sure essential services are running.<\/p><h3 id=\"jobs-and-cron-jobs\">Jobs and Cron Jobs<\/h3><p>The workloads we\u2019ve described so far have all assumed a long-running, service-like life cycle.  Kubernetes uses a workload called <strong>jobs<\/strong> to provide a more task-based workflow where the running containers are expected to exit successfully after some time once they have completed their work.  Jobs are useful if you need to perform one-off or batch processing instead of running a continuous service.<\/p><p>Building on jobs are <strong>cron jobs<\/strong>.  Like the conventional <code>cron<\/code> daemons on Linux and Unix-like systems that execute scripts on a schedule, cron jobs in Kubernetes provide an interface to run jobs with a scheduling component.  Cron jobs can be used to schedule a job to execute in the future or on a regular, reoccurring basis.  Kubernetes cron jobs are basically a reimplementation of the classic cron behavior, using the cluster as a platform instead of a single operating system.<\/p><h2 id=\"other-kubernetes-components\">Other Kubernetes Components<\/h2><p>Beyond the workloads you can run on a cluster, Kubernetes provides a number of other abstractions that help you manage your applications, control networking, and enable persistence.  We will discuss a few of the more common examples here.<\/p><h3 id=\"services\">Services<\/h3><p>So far, we have been using the term \u201cservice\u201d in the conventional, Unix-like sense: to denote long-running processes, often network connected, capable of responding to requests.  However, in Kubernetes, a <strong>service<\/strong> is a component that acts as a basic internal load balancer and ambassador for pods.  A service groups together logical collections of pods that perform the same function to present them as a single entity.<\/p><p>This allows you to deploy a service that can keep track of and route to all of the backend containers of a particular type.  Internal consumers only need to know about the stable endpoint provided by the service.  Meanwhile, the service abstraction allows you to scale out or replace the backend work units as necessary.  A service\u2019s IP address remains stable regardless of changes to the pods it routes to.  By deploying a service, you easily gain discoverability and can simplify your container designs.<\/p><p>Any time you need to provide access to one or more pods to another application or to external consumers, you should configure a service.  For instance, if you have a set of pods running web servers that should be accessible from the internet, a service will provide the necessary abstraction.  Likewise, if your web servers need to store and retrieve data, you would want to configure an internal service to give them access to your database pods.<\/p><p>Although services, by default, are only available using an internally routable IP address, they can be made available outside of the cluster by choosing one of several strategies.  The <strong>NodePort<\/strong> configuration works by opening a static port on each node\u2019s external networking interface.  Traffic to the external port will be routed automatically to the appropriate pods using an internal cluster IP service.<\/p><p>Alternatively, the <strong>LoadBalancer<\/strong> service type creates an external load balancer to route to the service using a cloud provider\u2019s Kubernetes load balancer integration.  The cloud controller manager will create the appropriate resource and configure it using the internal service service addresses.<\/p><h3 id=\"volumes-and-persistent-volumes\">Volumes and Persistent Volumes<\/h3><p>Reliably sharing data and guaranteeing its availability between container restarts is a challenge in many containerized environments.  Container runtimes often provide some mechanism to attach storage to a container that persists beyond the lifetime of the container, but implementations typically lack flexibility.<\/p><p>To address this, Kubernetes uses its own <strong>volumes<\/strong> abstraction that allows data to be shared by all containers within a pod and remain available until the pod is terminated.  This means that tightly coupled pods can easily share files without complex external mechanisms.  Container failures within the pod will not affect access to the shared files.  Once the pod is terminated, the shared volume is destroyed, so it is not a good solution for truly persistent data.<\/p><p><strong>Persistent volumes<\/strong> are a mechanism for abstracting more robust storage that is not tied to the pod life cycle.  Instead, they allow administrators to configure storage resources for the cluster that users can request and claim for the pods they are running.  Once a pod is done with a persistent volume, the volume\u2019s reclamation policy determines whether the volume is kept around until manually deleted or removed along with the data immediately.  Persistent data can be used to guard against node-based failures and to allocate greater amounts of storage than is available locally.<\/p><h3 id=\"labels-and-annotations\">Labels and Annotations<\/h3><p>A Kubernetes organizational abstraction related to, but outside of the other concepts, is labeling.  A <strong>label<\/strong> in Kubernetes is a semantic tag that can be attached to Kubernetes objects to mark them as a part of a group.  These can then be selected for when targeting different instances for management or routing.  For instance, each of the controller-based objects use labels to identify the pods that they should operate on.  Services use labels to understand the backend pods they should route requests to.<\/p><p>Labels are given as simple key-value pairs.  Each unit can have more than one label, but each unit can only have one entry for each key.  Usually, a \u201cname\u201d key is used as a general purpose identifier, but you can additionally classify objects by other criteria like development stage, public accessibility, application version, etc.<\/p><p><strong>Annotations<\/strong> are a similar mechanism that allows you to attach arbitrary key-value information to an object.  While labels should be used for semantic information useful to match a pod with selection criteria, annotations are more free-form and can contain less structured data.  In general, annotations are a way of adding rich metadata to an object that is not helpful for selection purposes.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>Kubernetes is an exciting project that allows users to run scalable, highly available containerized workloads on a highly abstracted platform.  While Kubernetes\u2019 architecture and set of internal components can at first seem daunting, their power, flexibility, and robust feature set are unparalleled in the open-source world.  By understanding how the basic building blocks fit together, you can begin to design systems that fully leverage the capabilities of the platform to run and manage your workloads at scale.<\/p>","created_at":"2019-08-26T13:45:50+0000","updated_at":"2019-08-26T13:45:55+0000","published_at":"2014-10-01T21:58:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"www.digitalocean.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["books","fantasy"],"is_public":false,"id":1004,"title":"Books like Game of Thrones: What to read after George R.R. Martin's fantasy series","url":"https:\/\/qz.com\/quartzy\/1690491\/books-like-game-of-thrones-what-to-read-after-george-r-r-martins-fantasy-series\/","content":"<p class=\"a1dbe d0e40\">In an <a href=\"https:\/\/www.theguardian.com\/books\/2019\/aug\/18\/george-rr-martin-interview-game-of-thrones-at-own-pace-now\">interview with the Guardian<\/a> this week, <em>A Song of Ice and Fire\u00a0<\/em>author George R.R. Martin admitted what his fans have long suspected: It\u2019s going to be awhile before the wildly popular fantasy book series is complete.<\/p><p class=\"a1dbe d0e40\">Now that HBO\u2019s <em>Game of Thrones <\/em>is over, fans of both the show and the books that inspired it are eagerly waiting to see how the writer wraps up his version of the story\u2014and if it will differ from the TV show\u2019s <a href=\"https:\/\/qz.com\/quartzy\/1623459\/game-of-thrones-finale-was-a-lesser-lord-of-the-rings-ending\/\">divisive ending<\/a>. Martin reportedly told the show\u2019s producers how he intends to finish his books, but the plot of the prestige drama had already passed that of Martin\u2019s novels. He is currently in the middle of writing the series\u2019 sixth book, <em>The Winds of Winter<\/em>, and has not yet started on its final installment, <em>A Dream of Spring<\/em>.<\/p><p class=\"a1dbe d0e40\">Martin is 70, and it\u2019s fair to wonder if at his current pace he will actually <em>ever<\/em> finish the series that made him famous. At the earliest, it will be several years before the books are done, giving fans of Martin\u2019s work ample time to read other stuff that might fill the <em>Game of Thrones<\/em>-sized holes in their libraries.<\/p><p class=\"a1dbe d0e40\">The Quartz newsroom came up with 7 of its favorite book series\u2014some obvious, some less so\u2014to read if you\u2019re a fan of the <em>Game of Thrones<\/em> show, Martin\u2019s <em>A Song of Ice and Fire<\/em> book series, or both:<\/p><h2 class=\"_7d6a7 d0e40\">The Lord of the Rings<\/h2><p class=\"a1dbe d0e40\">Duh. The gold standard for fantasy writing, J.R.R. Tolkien\u2019s\u00a0<em>The Lord of the Rings\u00a0<\/em>trilogy has inspired every fantasy story that\u2019s come since\u2014including and especially <em>A Song of Ice and Fire<\/em>. Martin has repeatedly called Tolkien one of his favorite writers and\u00a0<em>The Lord of the Rings\u00a0<\/em>one of the biggest inspirations of his own mythologies. He even credited his knack for killing off popular characters to the Tolkien character <a href=\"https:\/\/qz.com\/quartzy\/1355619\/game-of-thrones-character-deaths-blame-lord-of-the-rings-and-gandalf\/\">Gandalf<\/a>, the friendly wizard who (spoiler alert) shockingly \u201cdies\u201d in the first book of the trilogy. (He comes back, even more powerful, in the next installment.)<\/p><p class=\"a1dbe d0e40\">Tolkien\u2019s writing is at times almost impossibly beautiful, and his vision unparalleled. An academic and a philologist in addition to being a writer, Tolkien created multiple languages from scratch and populated his fictional universe with thousands of years worth of stories, settings, and characters. Without <em>The Lord of the Rings<\/em>, <em>Game of Thrones<\/em> very likely would not exist. It\u2019s the best and most important work of fantasy in modern history.<\/p><h2 class=\"_7d6a7 d0e40\">Dune<\/h2><p class=\"a1dbe d0e40\">Frank Herbert\u2019s 1965 novel (and the five subsequent books in the series) is more science fiction than fantasy, but it will still feel familiar to <em>Thrones\u00a0<\/em>fans. That\u2019s because the story revolves around a far-future feudalistic society in which warring families and factions struggle over the universe\u2019s labor and resources\u2014namely the desert planet Arrakis and its magical substance called \u201cthe spice.\u201d It will surely remind <em>A Song of Ice and Fire\u00a0<\/em>readers of its many Great Houses of the Seven Kingdoms: The Lannisters, the Starks, the Baratheons, etc.<\/p><p class=\"a1dbe d0e40\">If that sounds interesting, you might want to prioritize reading <em>Dune<\/em>. A film adaptation, directed by <em>Arrival<\/em> and <em>Blade Runner 2049<\/em> filmmaker Denis Villeneuve, will be released in Dec. 2020.<\/p><h2 class=\"_7d6a7 d0e40\">The Vorrh<\/h2><p class=\"a1dbe d0e40\">Brian Catling, a sculptor and performance artist before he became a published author, has written something that is totally unique. <em>The Vorrh<\/em> trilogy mixes elements of fantasy, sci-fi, steampunk, historical fiction, surrealism, horror, and just about every other genre into one expansive story. While there\u2019s too much going on to explain in a short summary, a very basic premise of the first book would go something like this: Somewhere in Africa lies an immense, mysterious forest, in which the properties of time and space do not appear to operate as humans understand them. A man is sent on a quest to traverse this forest, while others are tasked with stopping him.<\/p><p class=\"a1dbe d0e40\">That doesn\u2019t remotely do Catling\u2019s narrative justice, but if something like\u00a0<em>Heart of Darkness\u00a0<\/em>meets <em>Mad Max<\/em> meets <em>Game of Thrones<\/em> piques your interest, then give <em>The Vorrh<\/em> a try. It\u2019s obviously quite different from <em>Thrones<\/em> in genre and subject matter, but it shares quite a bit with Martin\u2019s story in tone and scope. It\u2019s a dark, challenging, but always fascinating read.<\/p><h2 class=\"_7d6a7 d0e40\">The Kingkiller Chronicle<\/h2><p class=\"a1dbe d0e40\">You\u2019re going to be hearing this title a lot in the near future. The fantasy series by Patrick Rothfuss, which has resulted in two novels so far, is in the process of being adapted into a TV series, a feature film, and even a video game. Lin-Manuel Miranda will co-produce and write the music for the TV series for Showtime (an HBO rival clearly hoping to conjure some <em>Thrones<\/em>-level magic), while Sam Raimi is set to direct the movie for Lionsgate.<\/p><p class=\"a1dbe d0e40\">There\u2019s a lot in Rothfuss\u2019 work that\u2019s reminiscent of Martin\u2019s: a medieval-esque setting, class struggles, magic, civil war, stories within stories, characters with the title \u201cMasters\u201d (though Martin spells it \u201cMaester\u201d). In 2008, Martin called the first book, <em>The Name of the Wind<\/em>, \u201cthe best epic fantasy I read last year.\u201d Martin added, \u201cHe\u2019s bloody good, this Rothfuss guy.\u201d<\/p><h2 class=\"_7d6a7 d0e40\">Black Leopard, Red Wolf<\/h2><p class=\"a1dbe d0e40\">Marlon James, the winner of the 2015 Man Booker Prize for <em>A Brief History of Seven Killings<\/em>, was joking when he dubbed his planned fantasy trilogy \u201cthe African <em>Game of Thrones<\/em>,\u201d but that hasn\u2019t stopped everyone from <a href=\"https:\/\/www.newyorker.com\/magazine\/2019\/01\/28\/why-marlon-james-decided-to-write-an-african-game-of-thrones\">describing it that way<\/a>, or from comparing it to the best of Martin\u2019s work. The first book in the trilogy, <em>Black Leopard, Red Wolf<\/em>, came out earlier this year to rave reviews.<\/p><p class=\"a1dbe d0e40\">Drawing from African mythology and history, James\u2019 story follows a Tracker searching for a missing boy across several landscapes, surrounded by a cast of characters that includes shape-shifters, mercenaries, giants, and zombie-like creatures. Swirling around Tracker\u2019s quest are political tensions between rival tribes\u2014again, not unlike the tensions at the heart of Martin\u2019s Westeros narrative.<\/p><h2 class=\"_7d6a7 d0e40\">Broken Earth series<\/h2><p class=\"a1dbe d0e40\">If your favorite part of <em>Game of Thrones<\/em> is its weird weather patterns (winters that last a lifetime?), then N.J. Jemisin\u2019s Broken Earth trilogy is for you. The stories take place on a planet, where, every few centuries, a \u201cFifth Season\u201d of cataclysms devastates the world, essentially bringing about the apocalypse over and over. Within this geological disarray is a caste-based society, of which the protagonist, Essun\u2014an \u201corogene\u201d who can manipulate energy\u2014is a part. Jemisin\u2019s sprawling narrative is a clever synthesis of <em>Thrones<\/em>-like fantasy and drama with science and our very real anxieties about climate change and social justice.<\/p><h2 class=\"_7d6a7 d0e40\">Thomas Cromwell trilogy<\/h2><p class=\"a1dbe d0e40\">If your favorite part of <em>Game of Thrones<\/em> is the royal political machinations and you don\u2019t much care for dragons or magic, then you\u2019ll probably take a liking to Hilary Mantel\u2019s trilogy chronicling the rise and fall of Thomas Crowell, King Henry VIII\u2019s chief minister. It\u2019s like a real-life version of the Lannisters.<\/p><p class=\"a1dbe d0e40\">Mantel\u2019s first book, <em>Wolf Hall<\/em>, won the Man Booker Prize and the National Book Critics Circle award and was adapted into a successful BBC miniseries starring Oscar-winner Mark Rylance as Cromwell. Another book soon followed: <em>Bring Up the Bodies<\/em>, while a third and final installment, <em>The Mirror and the Light<\/em>, will be released in March of next year.<\/p><h3 class=\"a83d8 d0e40\">Even more options:<\/h3><ul class=\"_487dc d0e40\"><li><em>His Dark Materials<\/em><\/li><li><em>Ancillary Justice<\/em><\/li><li><em>The Dark Tower<\/em><\/li><li><em>The Wheel of Time<\/em><\/li><li><em>The Stormlight Archive<\/em><\/li><li><em>Hainish Cycle<\/em><\/li><li><a href=\"https:\/\/qz.com\/quartzy\/1315808\/george-rr-martin-recommends-nine-books-for-game-of-thrones-fans\/\">Books Martin himself recommends<\/a><\/li><\/ul>","created_at":"2019-08-26T13:25:10+0000","updated_at":"2019-08-26T13:25:15+0000","published_at":"2019-08-24T10:00:01+0000","published_by":["Adam Epstein"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"qz.com","preview_picture":"https:\/\/cms.qz.com\/wp-content\/uploads\/2019\/08\/game-of-thrones-brienne-book-e1566320418255.jpg?quality=75&strip=all&w=1400","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1003,"title":"The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery","url":"https:\/\/gizmodo.com\/the-gorgeous-typeface-that-drove-men-mad-and-sparked-a-1686081182","content":"<div class=\"js_share_tools_side share-toolbar-sticky\"><figure class=\"js_marquee-assetfigure align--bleed\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--7pX1eL8W--\/c_scale,f_auto,fl_progressive,q_80,w_800\/k6a5pggjjvwabsfy11nc.png\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--7pX1eL8W--\/c_scale,f_auto,fl_progressive,q_80,w_800\/k6a5pggjjvwabsfy11nc.png 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--7pX1eL8W--\/c_scale,f_auto,fl_progressive,q_80,w_800\/k6a5pggjjvwabsfy11nc.png 3x\" data-width=\"920\" data-chomp-id=\"k6a5pggjjvwabsfy11nc\" data-format=\"png\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><p>\n\tNo one seemed to notice him: A dark figure who often came to stand at the edge of London's Hammersmith Bridge on nights in 1916. No one seemed to notice, either, that during his visits he was dropping something into the River Thames. Something heavy.<\/p><div class=\"ad-unit js_ad-unit\"><div class=\"ad-inner google-adx hide\"><div class=\"ad-container dfp dfp-slot-ADX_PERMALINK_TEXT ad-adx-ad\" data-ad-unit=\"ADX_PERMALINK_TEXT\" data-targeting=\"{&quot;pos&quot;:&quot;adx_native&quot;}\"><div id=\"ad-container-35067803\" class=\"ad-container js_ad-container js_ad-container-top\" data-zone-type=\"ADX_PERMALINK_TEXT\"><\/div><\/div><p>\n\tOver the course of more than a hundred illicit nightly trips, this man was committing a crime\u2014against his partner, a man who owned half of what was being heaved into the Thames, and against himself, the force that had spurred its creation. This venerable figure, founder of the legendary <a href=\"http:\/\/www.artsandcraftsmuseum.org.uk\/Arts_and_Crafts_Movement\/Emery_Walker_Library\/The_Private_Press_Movement\/Doves_Press.aspx\" target=\"_blank\" rel=\"noopener\">Doves Press<\/a> and the mastermind of its typeface, was a man named T.J. Cobden Sanderson. And he was taking the metal type that he had painstakingly overseen and dumping thousands of pounds of it into the river.<\/p><div class=\"ad-container dfp dfp-slot-SPLASHYMID_MOBILE ad-splashy-mid\" data-ad-unit=\"SPLASHYMID_MOBILE\" data-targeting=\"{&quot;pos&quot;:&quot;splashymid&quot;}\"><div id=\"swappable-mobile-ad-container\" class=\"js_ad-mobile-dynamic swappable-mobile-ad-container js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container js_swappable-mobile-ad dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><div class=\"instream-native-video instream-native-video--mobile instream-permalink\"><p>As a driving force in the Arts &amp; Crafts movement in England, Cobden Sanderson championed traditional craftsmanship against the rising tides of industrialization. He was brilliant and creative, and in some ways, a luddite\u2014because he was concerned that the typeface he had designed would be sold to a mechanized printing press after his death by his business partner, with whom he was feuding.<\/p><figure class=\"js_marquee-assetfigure align--center\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--arq7psNt--\/c_fit,f_auto,fl_progressive,q_80,w_320\/x8dny20pubhh2kk6ny6z.png 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--arq7psNt--\/c_fit,f_auto,fl_progressive,q_80,w_320\/x8dny20pubhh2kk6ny6z.png 3x\" data-width=\"720\" data-chomp-id=\"x8dny20pubhh2kk6ny6z\" data-format=\"png\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><small><em>Alfred, Lord Tennyson, Seven Poems and Two Translations, Doves Press 1902.<\/em><\/small><\/p><p>So, night after night, he was making it his business to \"bequeath\" it to the river, in his words, screwing his partner out of his half of their work and destroying a legendarily beautiful typeface forever. Or so it seemed.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><hr class=\"storybreak-stars\" \/><p>\n\tAlmost exactly a century later, this November, a cadre of ex-military divers who work for the Port of London Authority were gearing up to descend into the Thames to look for the small metal bits\u2014perhaps hundreds of thousands of them\u2014that Cobden Sanderson had thrown overboard so many years ago.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tThey were doing this at the behest and personal expense of <a href=\"https:\/\/twitter.com\/thedovestype\" target=\"_blank\" rel=\"noopener\">Robert Green<\/a>, a designer who has spent years <a href=\"http:\/\/www.typespec.co.uk\/doves-type\/\" target=\"_blank\" rel=\"noopener\">researching and recreating the lost typeface<\/a>, which is available on <a href=\"http:\/\/www.typespec.co.uk\/doves-type\/\" target=\"_blank\" rel=\"noopener\">Typespec<\/a>. As Green told me over the phone recently, the Port of London Authority had been hesitant about letting him pay its diving team to search for the lost type. \"They were actually concerned that I was some crazy bloke looking for a needle in a haystack and throwing a couple grand away,\" he laughs.<\/p><figure class=\"js_marquee-assetfigure align--center\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--Te872M9r--\/c_fit,f_auto,fl_progressive,q_80,w_320\/ucqqa8v0ycgnv4qpmqnu.png 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--Te872M9r--\/c_fit,f_auto,fl_progressive,q_80,w_320\/ucqqa8v0ycgnv4qpmqnu.png 3x\" data-width=\"737\" data-chomp-id=\"ucqqa8v0ycgnv4qpmqnu\" data-format=\"png\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><small><em>Courtesy of Robert Green.<br \/><\/em><\/small><\/p><p>\n\tIt's not hard to imagine how crazy he must have seemed. A civilian offering to pay the city's salvage divers to troll the depths of the muddy Thames, possibly for weeks, looking for tiny chunks of metal that were thrown there by a deranged designer more than a century ago? Yeah, that's pretty crazy.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tIn the end, it only took them 20 minutes to find some.<\/p><p>\n\tGreen has spent years researching Cobden Sanderson's story, using what amounts to forensic psychology to understand the actions of a man who lived 100 years ago, studying how and where he would have dumped his illicit cargo. Green had narrowed it down to a small dip in the river, and it was there that the city's divers uncovered most of their haul. \"They were really into it,\" Green remembers. \"They wanted to find something, which they did.\"<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><figure class=\"js_marquee-assetfigure align--center\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lK_TTSp_--\/c_fit,f_auto,fl_progressive,q_80,w_320\/meimxwemcklbuxhm3qc3.png 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--lK_TTSp_--\/c_fit,f_auto,fl_progressive,q_80,w_320\/meimxwemcklbuxhm3qc3.png 3x\" data-width=\"700\" data-chomp-id=\"meimxwemcklbuxhm3qc3\" data-format=\"png\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><p><small><em>Courtesy of Robert Green.<br \/><\/em><\/small><\/p><p>\n\tWhat they ended up uncovering over their two day dive was several hundred pieces of type, as documented by <a href=\"http:\/\/www.thesundaytimes.co.uk\/sto\/Magazine\/article1502054.ece\" target=\"_blank\" rel=\"noopener\">The Sunday Times' Justin Quirk<\/a>, who attended the dive. It was far from the full haul, either: Green points out the Hammersmith Bridge has been the target of two IRA bombings, one of which blew water 60 feet into the air after a suitcase packed with explosives was heaved into the river around the spot the type was dumped.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><figure class=\"js_marquee-assetfigure align--bleed\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--V_NMAfv4--\/c_scale,f_auto,fl_progressive,q_80,w_800\/aspdoeanfwob9vdfwawk.jpg 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--V_NMAfv4--\/c_scale,f_auto,fl_progressive,q_80,w_800\/aspdoeanfwob9vdfwawk.jpg 3x\" data-width=\"1000\" data-chomp-id=\"aspdoeanfwob9vdfwawk\" data-format=\"jpg\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><p>As a result, it's possible some of the metal punches were blown to other locations\u2014it's also possible that they were embedded in concrete poured around the bridge as part of repairs.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><hr class=\"storybreak-stars\" \/><p>Now would probably be a good time to explain exactly what \"it\" is. Today, typefaces\u2014or fonts, as we usually call them these digital days\u2014are essentially just little bits of binary on our computers. But the age of digital type is young, at only a few decades.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tCobden Sanderson and his partner, Emery Walker, founded the Doves Press in 1900. Walker was a businessman, with plenty of other concerns in the world, but Cobden Sanderson was a creative perfectionist\u2014a man obsessed with authenticity and craft. Together, they commissioned a typeface for their press, to be based on a 15th century Venetian type. That meant paying a \"punchcutter\" to create steel \"punches\" for each letter in the type\u2014from which a matrix would be made by pressing a piece of copper into the metal punch. Then, the type itself could be cast from the matrix.<\/p><figure class=\"js_marquee-assetfigure align--bleed\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--EF5b4p7k--\/c_scale,f_auto,fl_progressive,q_80,w_800\/dx3edfgtas6hx1c7kemn.jpg 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--EF5b4p7k--\/c_scale,f_auto,fl_progressive,q_80,w_800\/dx3edfgtas6hx1c7kemn.jpg 3x\" data-width=\"910\" data-chomp-id=\"dx3edfgtas6hx1c7kemn\" data-format=\"jpg\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><em>Photograph by <a href=\"http:\/\/www.samarmstrong.co.uk\/\" target=\"_blank\" rel=\"noopener\">Sam Armstrong<\/a>, courtesy of <a href=\"http:\/\/www.thesundaytimes.co.uk\/sto\/Magazine\/article1502054.ece\" target=\"_blank\" rel=\"noopener\">The Sunday Times<\/a>.<\/em><\/p><p>\n\tTheir type was created in 1899, and the duo would use it to print indescribably beautiful books, bound by hand and designed with the perfect balance of craftsmanship and modern utility. Cobden Sanderson was a bit of a snob in the sense that he only wanted to commit his designs to the finest literature, the \"most beautiful words.\" They printed <a href=\"https:\/\/www.flickr.com\/photos\/artsandcraftsmuseumcheltenham\/5409736693\/in\/set-72157625834581201\" target=\"_blank\" rel=\"noopener\">Paradise Lost<\/a>. They printed the<a href=\"https:\/\/www.flickr.com\/photos\/artsandcraftsmuseumcheltenham\/5410347798\/in\/set-72157625834581201\" target=\"_blank\" rel=\"noopener\"> English Bible<\/a>. Today, copies of these books are extremely rare, and they command thousands of dollars at auction.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tBut soon, the Doves Press was in trouble. According to <a href=\"http:\/\/www.typespec.co.uk\/doves-type-revival\/\" target=\"_blank\" rel=\"noopener\">TypeSpec's own account of the partnership<\/a>, Walker wanted to shut it down and divide the metal evidence\u2014thousands of pounds of it\u2014of the type between himself and Cobden Sanderson, and for them to go their separate ways. As Quirk explains in <a href=\"http:\/\/www.thesundaytimes.co.uk\/sto\/Magazine\/article1502054.ece\" target=\"_blank\" rel=\"noopener\">The Sunday Times<\/a>, they landed on an agreement that Cobden Sanderson would keep the type until his death, at which point Walker would own it. But Cobden Sanderson was horrified by the idea of letting what he thought of as his own work go to Walker\u2014and so slowly, over the course of the next few years, he decided upon a plan of action that would deprive Walker of his end of the bargain. <\/p><figure class=\"js_marquee-assetfigure align--bleed\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--wN2IZ0l8--\/c_scale,f_auto,fl_progressive,q_80,w_800\/vdrctk4xup018iku5msy.jpg 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--wN2IZ0l8--\/c_scale,f_auto,fl_progressive,q_80,w_800\/vdrctk4xup018iku5msy.jpg 3x\" data-width=\"3133\" data-chomp-id=\"vdrctk4xup018iku5msy\" data-format=\"jpg\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><small><em>Courtesy of Robert Green.<\/em><\/small><\/p><p>\n\t\"It took him a few years to actually decide to throw away the type\u2014he ruminated for years about whether or not he should do it,\" Green says. He wrote about the process in his lengthy journals (\"he would have been a bit of an over-sharer\" today, Green adds), leaving behind detailed accounts of his inner turmoil. Eventually, he decided he's rather destroy the type than see it made into a mechanical version of its former glory. \"He reveled in it,\" Green says. Cobden Sanderson said it himself: \"If Emery Walker wants to find it, he'll have to dive for it.\"<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><hr class=\"storybreak-stars\" \/><p>\n\tGreen has spent years researching the Doves Press type\u2014he even redrew it, after thousands of hours of painstaking research work, and published his revival in 2013 as a digital typeface<a href=\"http:\/\/www.typespec.co.uk\/doves-type-revival\/\" target=\"_blank\" rel=\"noopener\"> called the Doves Type that anyone can buy<\/a>.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tBut about a year ago, he says, he started wondering if there wasn't something to be salvaged from the river. \"People kept saying nobody's ever found it,\" he says. \"But nowhere could I find an account of anybody searching for it.\"<\/p><div class=\"ad-container dfp dfp-slot-SPLASHYMID ad-splashy-mid\" data-ad-unit=\"SPLASHYMID\" data-targeting=\"{&quot;pos&quot;:&quot;splashymid&quot;}\"><div class=\"instream-native-video instream-native-video--desktop instream-permalink\"><figure class=\"js_marquee-assetfigure align--bleed\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--PuokT_t0--\/c_scale,f_auto,fl_progressive,q_80,w_800\/agnvypct4vpkbmvslyuj.jpg 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--PuokT_t0--\/c_scale,f_auto,fl_progressive,q_80,w_800\/agnvypct4vpkbmvslyuj.jpg 3x\" data-width=\"905\" data-chomp-id=\"agnvypct4vpkbmvslyuj\" data-format=\"jpg\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><small><em>The updated type, courtesy of Robert Green.<\/em><\/small><\/p><p>\n\tWhich brings us to a very good question: Why would anyone search for it? What made it so special, so worth saving?<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tThe Doves Press was a unique entity, but in some ways it mirrored what's happening in today's design world. At the cusp of the modern age, Doves was founded to preserve a craft that went back centuries. It was also destined to fail, to end up as a historical eccentricity that died out just as the mechanized printing press sprang onto the scene. It valued one thing above all others: Doing things by hand, and doing them with utter devotion.<\/p><figure class=\"js_marquee-assetfigure align--center\"><div class=\"js_img-wrapper img-wrapper lazy-image\"><div class=\"img-permalink-sub-wrapper\"><img src=\"denied:data:image\/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==\" class=\"lazyload ls-lazy-image-tag\" data-sizes=\"auto\" data-srcset=\"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--S_pTqG9A--\/c_fit,f_auto,fl_progressive,q_80,w_320\/uhwwjehqlobgiki2si9k.png 2x, https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--S_pTqG9A--\/c_fit,f_auto,fl_progressive,q_80,w_320\/uhwwjehqlobgiki2si9k.png 3x\" data-width=\"720\" data-chomp-id=\"uhwwjehqlobgiki2si9k\" data-format=\"png\" alt=\"Illustration for article titled The Gorgeous Typeface That Drove Men Mad and Sparked a 100-Year Mystery\" \/><\/div><\/div><\/figure><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><small><em>'Oenone', Alfred, Lord Tennyson, Seven Poems and Two Translations, Doves Press 1902.<\/em><\/small><\/p><p>\n\tFor Green, who has worked in the design world since he was young, the Arts &amp; Crafts' glorification of handicraft resonates even today. \"The Industrial Revolution scared the crap out of them, and quite rightly,\" he says of those turn-of-the-century designers, pointing out how digitization has further devalued the skills of graphic designers today. \"A whole swathe of the middle class is being knocked out,\" he adds.  \"You look at what's happened to graphic design now, it's been completely devalued and demonetized. It's very hard to earn a living as a graphic designer.\"<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tNow, traditional methods\u2014like letterpress printing\u2014are wildly popular once more among contemporary designers. \"People are turning back to craft to earn a living,\" says Green, a bit like Cobden Sanderson and his contemporaries did. Not only because it gives their work authenticity, he says, but also: \"because it's fun.\"<\/p><hr class=\"storybreak-stars\" \/><p>\n\tIt's strange to imagine that a designer who was born more than a century after Cobden Sanderson has been the one to rebuild his ruined life's work.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tIn an odd way, he's also healing the rift between Cobden Sanderson and his partner, Emery Walker. Rather than sell the metal punches he lifted from the Thames, Green is keeping half and giving half to the Emery Walker Trust, which maintains his <a href=\"http:\/\/emerywalker.org.uk\/\" target=\"_blank\" rel=\"noopener\">former home as a museum to his work<\/a>. 100 years ago, Cobden Sanderson had said that Walker would \"have to dive for it\" if he wanted his half of the business. Oddly enough, he's getting half after all\u2014with Green acting as his generous, devoted proxy diver.<\/p><p class=\"has-video media-large video-embed align--bleed\"><iframe allowfullscreen=\"allowfullscreen\" data-chomp-id=\"LsQftfMOdC4\" scrolling=\"no\" data-src=\"\/ajax\/inset\/iframe?id=youtube-video-LsQftfMOdC4&amp;start=63\" data-recommend-id=\"youtube:\/\/LsQftfMOdC4\" height=\"450\" id=\"youtube-LsQftfMOdC4\" class=\"core-inset lazyload\" width=\"800\" frameborder=\"0\">[embedded content]<\/iframe><\/p><p>\n\tToday, anyone can <a href=\"http:\/\/www.typespec.co.uk\/doves-type-revival\/\" target=\"_blank\" rel=\"noopener\">download and buy Green's revival<\/a> of Cobden Sanderson's type online. \"He probably would've been horrified,\" Green laughs. But then again, he doesn't see his Doves Type as an exact recreation of the original. It's more like an echo or a simulacrum\u2014it has a life of its own.<\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p>\n\tIt's a story that ties together the most important and controversial ideas in the last century of design. Cobden Sanderson was reacting\u2014criminally!\u2014to the threat of his profession being made irrelevant by the machine age. Today, designers are still struggling to find meaning and reconcile their work with a kind of machine logic born by technologies that Cobden Sanderson couldn't have even imagined.<\/p><p>\n\t100 years later, the concerns of a man obsessed with craft still resonate with us. But then again, without computers\u2014a product of the machines that Arts &amp; Crafts strove against\u2014Doves Type wouldn't exist. Today, it's a living, breathing thing, an amalgam of the technologies and machines that were just being born when it was dropped into the river. <\/p><div class=\"js_ad-mobile-dynamic  js_ad-dynamic ad-mobile-dynamic movable-ad\"><div class=\"ad-unit ad-mobile\"><div class=\"ad-mobile-inner\"><p class=\"ad-label proxima\"><small class=\"proxima\">Advertisement<\/small><\/p><div class=\"ad-container dfp dfp-slot-MOBILE_IN_POST ad-mobile\" data-ad-unit=\"MOBILE_IN_POST\" data-targeting=\"{&quot;pos&quot;:&quot;mobile_inpost&quot;}\"><div class=\"bt-wrapper\"><p class=\"ad-label-bottom\"><\/p><\/div><\/div><\/div><p><small><em>This post has been updated to reflect more details of the bargain between the pair, as well as the fact that Doves Type is really the name of Green's revival. <\/em><\/small><em>You can read more about the revival and the story on <a href=\"http:\/\/www.typespec.co.uk\/doves-type\/\" target=\"_blank\" rel=\"noopener\">TypeSpec<\/a> or in <a href=\"http:\/\/www.thesundaytimes.co.uk\/sto\/Magazine\/article1502054.ece\" target=\"_blank\" rel=\"noopener\">The Sunday Times<\/a>. Lead image: <a href=\"https:\/\/www.flickr.com\/photos\/psg3\/8246586343\" target=\"_blank\" rel=\"noopener\">Peter Glenday<\/a>\/CC. Second image: <a href=\"http:\/\/www.shutterstock.com\/gallery-697858p1.html\" target=\"_blank\" rel=\"noopener\"><small>Dafinka<\/small><\/a><small>\/Shutterstock. All images courtesy of Robert Green unless otherwise noted. <\/small><\/em><\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2019-08-25T17:34:16+0000","updated_at":"2019-08-25T17:34:16+0000","published_at":"2015-02-16T13:00:00+0000","published_by":["Kelsey Campbell-Dollaghan"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":9,"domain_name":"gizmodo.com","preview_picture":"https:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--4OyG85j_--\/c_fill,fl_progressive,g_center,h_900,q_80,w_1600\/k6a5pggjjvwabsfy11nc.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1002,"title":"A quick beginner\u2019s guide to drawing","url":"https:\/\/medium.com\/personal-growth\/a-quick-beginners-guide-to-drawing-58213877715e?_branch_match_id=694009154483098151","content":"<p><\/p><h2 class=\"ao cy kj kk at\">6 drawing exercises to get you started right now!<\/h2><div class=\"kl\"><div class=\"ag af\"><div><a href=\"https:\/\/medium.com\/@ralphammer?source=post_page-----58213877715e----------------------\"><div class=\"ek km kn\"><img alt=\"Ralph Ammer\" src=\"https:\/\/miro.medium.com\/fit\/c\/96\/96\/1*dpnpENXIDu8KMxIaTjbcZA.png\" class=\"l fq kn km\" width=\"48\" height=\"48\" \/><\/div><\/a><\/div><div class=\"ks ab l\"><div class=\"af\"><div><div class=\"kt af ag ku\"><a class=\"dd de ax ay az ba bb bc bd be jg bh bi dh di\" href=\"https:\/\/medium.com\/@ralphammer?source=post_page-----58213877715e----------------------\">Ralph Ammer<\/a><\/div><\/div><\/div><p><a class=\"dd de ax ay az ba bb bc bd be jg bh bi dh di\" href=\"https:\/\/medium.com\/personal-growth\/a-quick-beginners-guide-to-drawing-58213877715e?source=post_page-----58213877715e----------------------\">Aug 22, 2017<\/a> \u00b7 7 min read<\/p><\/div><\/div><\/div><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"9f2b\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">The <strong class=\"ma mm\">basic craft of drawing<\/strong> is about two things: you learn <strong class=\"ma mm\">to control your hand<\/strong> and <strong class=\"ma mm\">to see<\/strong>.<\/p><p id=\"617b\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">Tip: For the following 6 exercises I suggest you <\/em><strong class=\"ma mm\"><em class=\"mn\">stick with one pen and one particular type of paper<\/em><\/strong><em class=\"mn\"> (for instance A5).<\/em><\/p><p id=\"8908\" class=\"ly lz ch ap ma b mb nb md nc mf nd mh ne mj nf ml\">The first two exercises are about controlling your hand. We want to <strong class=\"ma mm\">build muscles<\/strong> and train our <strong class=\"ma mm\">hand-eye-coordination<\/strong>. Mechanical exercises like these are great for beginners. Later on you might use them to <strong class=\"ma mm\">explore new pens<\/strong> or get started <strong class=\"ma mm\">when you don\u2019t know what to draw <\/strong>yet.<\/p><p id=\"76a6\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">They are also terrific ways to relax your mind.<\/p><h2 id=\"8ce4\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Exercise 1: Circles \u2014 more is more!<\/h2><blockquote class=\"ns nt nu\"><p id=\"0af6\" class=\"ly lz ch mn ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Distribute circles of various sizes on a piece of paper until the paper is filled. Make sure the circles do not overlap.<\/strong><\/p><\/blockquote><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"244e\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Drawing circles is <a class=\"dd bt nx ny nz oa\" target=\"_blank\" href=\"https:\/\/medium.com\/personal-growth\/why-perfection-is-boring-1079cb3bf5d1\">not as easy as you might think<\/a>. Notice how the circles become harder, the bigger you make them? Try them in both directions \u2014 and make lots of them.<\/p><p id=\"4a07\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">Tip: <\/em><strong class=\"ma mm\"><em class=\"mn\">Shake your hand<\/em><\/strong><em class=\"mn\"> when it starts cramping! This is a workout for our hands after all.<\/em><\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><h2 id=\"7bf2\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Exercise 2: Hatching \u2014 the joy of structure<\/h2><blockquote class=\"ns nt nu\"><p id=\"95cf\" class=\"ly lz ch mn ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Fill a piece of paper with parallel lines.<\/strong><\/p><\/blockquote><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"9063\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Diagonal lines come easiest for us because<strong class=\"ma mm\"> they comply to the motion of our wrist<\/strong>. Did you notice how left-handed people prefer the opposite direction than right-handed people? Have a look at drawings by your favourite draughtswoman or draughtsman (in my case: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Leonardo_da_Vinci\" class=\"dd bt nx ny nz oa\" target=\"_blank\">Leonardo<\/a>), and guess which hand they used!<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"826c\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Now make sure to try other directions as well<\/strong>. Have fun! Combine various hatchings and enjoy watching the darkness spread on your paper.<\/p><p id=\"a677\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">Tip: <\/em><strong class=\"ma mm\"><em class=\"mn\">Don\u2019t rotate the paper<\/em><\/strong><em class=\"mn\">. The whole point here is to train your hand to get comfortable with <\/em><strong class=\"ma mm\"><em class=\"mn\">all<\/em><\/strong><em class=\"mn\"> directions.<\/em><\/p><p id=\"99d3\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">So now that we have worked our hands a little, let\u2019s train our eyes!<\/strong><\/p><p id=\"cc35\" class=\"ly lz ch ap ma b mb nb md nc mf nd mh ne mj nf ml\">Drawing is mainly about <a class=\"dd bt nx ny nz oa\" target=\"_blank\" href=\"https:\/\/medium.com\/personal-growth\/stop-taking-pictures-and-start-drawing-b1642aded2b6\"><strong class=\"ma mm\">seeing and understanding what you see<\/strong><\/a>. People often assume that everybody sees the same, but actually seeing is a skill which you can improve. <strong class=\"ma mm\">The more you draw, the more you see.<\/strong> The next four exercises <strong class=\"ma mm\">will make you see more<\/strong>.<\/p><h2 id=\"6d89\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Exercise 3: Contour \u2014 Show me your hands!<\/h2><blockquote class=\"ns nt nu\"><p id=\"6565\" class=\"ly lz ch mn ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">You see all those fascinating contours of your hand? Collect them on a piece of paper! Don\u2019t try to draw the whole hand yet, just pick some of those lovely lines.<\/strong><\/p><\/blockquote><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"0822\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Whether you draw a person, a plant, or your favourite animal, it is usually the <strong class=\"ma mm\">contours<\/strong> that define a body or object and make others <strong class=\"ma mm\">recognise<\/strong> them. The challenge is less to draw those distinctive lines but rather <strong class=\"ma mm\">to see them<\/strong> in the first place!<\/p><p id=\"3649\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Even if you think you already know the shape of an object, it is always worth to take a closer look and rediscover it<\/strong>.<\/p><h2 id=\"56b1\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Exercise 4: Chiaroscuro \u2014 folding light and darkness<\/h2><blockquote class=\"ns nt nu\"><p id=\"5f5b\" class=\"ly lz ch mn ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Arrange and draw a piece of cloth. Start with the contours and then \u2014 using your hatching skills \u2014 create an interplay of light und darkness.<\/strong><\/p><\/blockquote><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"8c83\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">This exercise gives you a sense for light and darkness. I have to admit it is not the easiest and might as well be part of an advanced tutorial. Keep in mind: <strong class=\"ma mm\">This one is not about getting it perfectly \u201cright\u201d<\/strong>. The cloth is a <strong class=\"ma mm\">playground<\/strong> to try out those various hatchings you have practised before and feel how you can create light and shadows with your \u2014 almost \u2014 bare hands!<\/p><p id=\"17d9\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">Tip: You can use <\/em><strong class=\"ma mm\"><em class=\"mn\">curved hatching<\/em><\/strong><em class=\"mn\"> to modulate the shapes, and <\/em><strong class=\"ma mm\"><em class=\"mn\">crosshatching<\/em><\/strong><em class=\"mn\"> to achieve darker areas that resemble woven structures.<\/em><\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"f23b\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">Tip: Close your eyes just a little bit when you look at the cloth. You will see everything blurred, but you will also <\/em><strong class=\"ma mm\"><em class=\"mn\">see an enhanced contrast between light and dark<\/em><\/strong><em class=\"mn\">.<\/em><\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"2e6d\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">The arrangement of light is a great way to <strong class=\"ma mm\">show what is important<\/strong> in a picture. Just have a look at paintings by <a href=\"https:\/\/de.wikipedia.org\/wiki\/Rembrandt_van_Rijn\" class=\"dd bt nx ny nz oa\" target=\"_blank\">Rembrandt<\/a> or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Georges_de_La_Tour\" class=\"dd bt nx ny nz oa\" target=\"_blank\">Georges de la Tour<\/a>. And the next time you watch a movie look out for dramatic effects achieved with light and shadow.<\/p><h2 id=\"8248\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Exercise 5: Perspective \u2014 lost in space<\/h2><blockquote class=\"ns nt nu\"><p id=\"759c\" class=\"ly lz ch mn ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Let\u2019s draw some cubes! Just follow the simple steps below.<\/strong><\/p><\/blockquote><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"342b\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Perspective drawing is basically a <strong class=\"ma mm\">projection of a 3D environment on a 2D surface<\/strong> (your piece of paper).<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"c297\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Constructing perspective drawings is a bit of a science and cannot be covered in detail within the attention span of an online article. Nevertheless we can have some fun with <strong class=\"ma mm\">a simple technique<\/strong> that gives us <strong class=\"ma mm\">an intuitive feel<\/strong> for the magic of perspective drawing:<\/p><p id=\"4b2d\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 1:<\/em><\/strong> Draw a horizontal line. This is the <strong class=\"ma mm\">horizon<\/strong> of your picture.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"71b6\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 2:<\/em><\/strong> Define two points on the horizon near the edge of the paper. These are your two <strong class=\"ma mm\">vanishing points<\/strong>.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"cbc0\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 3:<\/em><\/strong> Draw a vertical line somewhere.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"e2c9\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 4:<\/em><\/strong> Connect the endpoints of the vertical line with the vanishing points.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"2b18\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 5:<\/em><\/strong> Add two vertical lines like this:<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"8fc0\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 6:<\/em><\/strong> Connect them with the vanishing points.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"2300\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\"><em class=\"mn\">Step 7:<\/em><\/strong> Now use a dark pencil or pen to emphasise the cube. Voil\u00e0!<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"0ce7\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Repeat steps 3\u20137 as often as you like.<\/strong> Have fun! If you feel adventurous you might even hatch the sides of the cubes.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"4557\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">Tip: When you draw lines that meet, it is usually a good idea to be confident and <\/em><strong class=\"ma mm\"><em class=\"mn\">let them overlap a little bit<\/em><\/strong><em class=\"mn\">. The shapes will look better defined.<\/em><\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"0ddb\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Mastering perspective drawings will give you the power to create illusions of depth. But most importantly,<strong class=\"ma mm\"> you teach your brain how to think in three dimensions<\/strong>. So even if you choose to draw \u201cflat\u201d drawings or mess with the \u201crules\u201d of perspective \u2014 which I like to do \u2014<strong class=\"ma mm\"> to understand perspective drawing is still one of the most precious drawing skills you can learn<\/strong>.<\/p><h2 id=\"4cb0\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Exercise 6: Composition \u2014 Why is this there?<\/h2><blockquote class=\"ns nt nu\"><p id=\"196c\" class=\"ly lz ch mn ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Make 5 different drawings of an object. Arrange the object differently on the paper each time!<\/strong><\/p><\/blockquote><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"6378\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">Composition is a great tool to <strong class=\"ma mm\">\u201csay\u201d something with a drawing<\/strong>, to <strong class=\"ma mm\">shape its meaning or message<\/strong>.<\/p><p id=\"1424\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">To understand how it works we have to bear in mind that our perception has been <strong class=\"ma mm\">shaped by everyday experiences<\/strong>. For instance horizontal and vertical lines seem more \u201cstable\u201d to us than diagonals, which might \u201cfall over\u201d any second. And when we see a big dark shape at the bottom we somehow assume it must be \u201cheavy\u201d.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><p id=\"a48f\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\">As you try different arrangements of your subject on a piece of paper, notice how this changes their connotations \u2014 their meaning.<\/p><figure class=\"lk ll lm ln lo ew x y paragraph-image\"><h2 id=\"2189\" class=\"ng mp ch ap ao bz nh ni nj nk nl nm nn no np nq nr\">Let me know what you think!<\/h2><p id=\"0cef\" class=\"ly lz ch ap ma b mb nb md nc mf nd mh ne mj nf ml\">Since this is my first drawing tutorial, I am curious about which parts you enjoyed. What else would you like to learn about drawing? <strong class=\"ma mm\">Please leave ideas and suggestions in the comment section below!<\/strong><\/p><p id=\"592e\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><strong class=\"ma mm\">Oh, and here is part II: \u201c<\/strong><a class=\"dd bt nx ny nz oa\" target=\"_blank\" href=\"https:\/\/medium.com\/@ralphammer\/5-more-drawing-exercises-9c0df4645387\"><strong class=\"ma mm\">5 more drawing exercises<\/strong><\/a><strong class=\"ma mm\">\u201d!<\/strong><\/p><p id=\"0b15\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><a href=\"https:\/\/www.youtube.com\/watch?v=ZqlTSCvP-Z0&amp;fbclid=IwAR2-WMNWlyJHJYTg4vgfG5Y8UrOgCxJ1PaxR1EYF2zyC1Nf1B15PXL2W3Qs\" class=\"dd bt nx ny nz oa\" target=\"_blank\"><strong class=\"ma mm\">Here<\/strong><\/a><strong class=\"ma mm\"> is a video of me <\/strong><a href=\"https:\/\/www.youtube.com\/watch?v=ZqlTSCvP-Z0&amp;fbclid=IwAR2-WMNWlyJHJYTg4vgfG5Y8UrOgCxJ1PaxR1EYF2zyC1Nf1B15PXL2W3Qs\" class=\"dd bt nx ny nz oa\" target=\"_blank\"><strong class=\"ma mm\">talking about drawing<\/strong><\/a><strong class=\"ma mm\"> (with animations of course)!<\/strong><\/p><p id=\"75ed\" class=\"ly lz ch ap ma b mb mc md me mf mg mh mi mj mk ml\"><em class=\"mn\">If you enjoyed this article, please hit those little <\/em>\ud83d\udc4f<em class=\"mn\"> below (you can \u201cclap\u201d multiple times!), share the story with your friends, and <\/em><a href=\"http:\/\/eepurl.com\/cJJLR1\" class=\"dd bt nx ny nz oa\" target=\"_blank\"><strong class=\"ma mm\"><em class=\"mn\">subscribe to my mailing list<\/em><\/strong><\/a><em class=\"mn\">. You can find an overview of my articles <\/em><a href=\"http:\/\/ralphammer.com\/writing\" class=\"dd bt nx ny nz oa\" target=\"_blank\"><strong class=\"ma mm\"><em class=\"mn\">here<\/em><\/strong><\/a><em class=\"mn\">.<\/em><\/p><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure><\/figure>","created_at":"2019-08-25T03:15:03+0000","updated_at":"2019-08-25T03:15:03+0000","published_at":"2019-01-18T19:51:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"medium.com","preview_picture":"https:\/\/miro.medium.com\/freeze\/max\/350\/1*rUnTj2M6B-pWZAGEJzUXIQ.gif","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":1001,"title":"acko.net","url":"https:\/\/acko.net\/","created_at":"2019-08-25T03:15:01+0000","updated_at":"2019-08-25T03:15:01+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"acko.net"},{"is_archived":0,"is_starred":0,"tags":["python","unicode","charsets"],"is_public":false,"id":1000,"title":"Ned Batchelder: Pragmatic Unicode","url":"https:\/\/nedbatchelder.com\/text\/unipain.html","content":"<p class=\"first\">This is a presentation I gave at PyCon 2012.&#13;\nYou can read the slides and text on this page, or open the <a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html\">actual presentation<\/a> in your browser,&#13;\nor watch the video:<\/p><iframe src=\"https:\/\/www.youtube.com\/embed\/sgHbC6udIqc\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>Also, clicking the slide images will jump into the full presentation at that point.&#13;\nThe Symbola font is included, but will have to be downloaded before some of the&#13;\nspecial symbols will appear.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#1\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/000.png\" alt=\"Pragmatic Unicode~ or ~How Do I Stop the Pain?\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Hi, I\u2019m Ned Batchelder.  I\u2019ve been writing in Python for over ten years,&#13;\n    which means at least a half-dozen times, I\u2019ve made the same Unicode&#13;\n    mistakes that everyone else has.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#2\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/001.png\" alt=\"The past\" width=\"512\" height=\"384\" \/><\/a><\/p><p>If you\u2019re like most Python programmers, you\u2019ve done it too: you\u2019ve built&#13;\n    a nice application, and everything seemed to be going fine.  Then one day&#13;\n    an accented character appeared out of nowhere, and your program started&#13;\n    belching UnicodeErrors.<\/p><p>You kind of knew what to do with those, so you added an encode or a&#13;\n    decode where the error was raised, but the UnicodeError happened somewhere&#13;\n    else.  You went to the new place, and added a decode, maybe an encode.&#13;\n    After playing whack-a-mole like this for a while, the problem seemed to be&#13;\n    fixed.<\/p><p>Then a few days later, another accent appeared in another place, and you&#13;\n    had to play a little bit more whack-a-mole until the problem finally&#13;\n    stopped.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#3\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/002.png\" alt=\"You\" width=\"512\" height=\"384\" \/><\/a><\/p><p>So now you have a program that works, but you\u2019re annoyed and&#13;\n    uncomfortable, it took too long, you know it isn\u2019t \u201cright,\u201d and you hate&#13;\n    yourself.  And the main thing you know about Unicode is that you don\u2019t like&#13;\n    Unicode.<\/p><p>You don\u2019t want to know about weirdo character sets, you just want &#13;\n    to be able to write a program that doesn\u2019t make you feel bad.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#4\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/003.png\" alt=\"This talk\" width=\"512\" height=\"384\" \/><\/a><\/p><p>You don\u2019t have to play whack-a-mole.  Unicode isn\u2019t simple, but it isn\u2019t&#13;\n    difficult either. With knowledge and discipline, you can deal with Unicode&#13;\n    easily and with grace.<\/p><p>I\u2019ll teach you five Facts of Life, and give you three Pro Tips that will&#13;\n    solve your Unicode problems. We\u2019re going to cover the basics of Unicode,&#13;\n    and how both Python 2 and Python 3 work.  They are different, but the&#13;\n    strategies you\u2019ll use are basically the same.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#5\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/004.png\" alt=\"The World &amp; Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>We\u2019ll start with the basics of Unicode.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#6\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/005.png\" alt=\"Bytes\" width=\"512\" height=\"384\" \/><\/a><\/p><p>The first Fact of Life: everything in a computer is bytes.  Files on&#13;\n    disk are a series of bytes, and network connections only transmit bytes.&#13;\n    Almost without exception, all the data going into or out of any program you&#13;\n    write, is bytes.<\/p><p>The problem with bytes is that by themselves they are meaningless, we&#13;\n    need conventions to give them meaning.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#7\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/006.png\" alt=\"ASCII\" width=\"512\" height=\"384\" \/><\/a><\/p><p>To represent text, we\u2019ve been using the ASCII code for nearly 50 years.&#13;\n    Every byte is assigned one of 95 symbols. When I send you a byte 65, you&#13;\n    know that I mean an upper-case A, but only because we\u2019ve agreed beforehand&#13;\n    on what each byte represents.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#8\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/007.png\" alt=\"ISO 8859-1\" width=\"512\" height=\"384\" \/><\/a><\/p><p>ISO Latin 1, or 8859-1, is ASCII extended with 96 more symbols.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#9\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/008.png\" alt=\"Windows-1252\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Windows added 27 more symbols to produce CP1252.  This is pretty much&#13;\n    the best you can do to represent text as single bytes, because there\u2019s not&#13;\n    much room left to add more symbols.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#10\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/009.png\" alt=\"Tower of Babel\" width=\"512\" height=\"384\" \/><\/a><\/p><p>With character sets like these, we can represent at most 256 characters.&#13;\n    But Fact of Life #2 is that there are way more than 256 symbols in the&#13;\n    world\u2019s text.  A single byte simply can\u2019t represent text world-wide.&#13;\n    During your darkest whack-a-mole moments, you may have wished that everyone&#13;\n    spoke English, but it simply isn\u2019t so. People need lots of symbols to&#13;\n    communicate.<\/p><p>Fact of Life #1 and Fact of Life #2 together create a fundamental&#13;\n    conflict between the structure of our computing devices, and the needs of&#13;\n    the world\u2019s people.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#11\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/010.png\" alt=\"Character codes\" width=\"512\" height=\"384\" \/><\/a><\/p><p>There have been a number of attempts to resolve this conflict.&#13;\n    Single-byte character codes like ASCII map bytes to characters.&#13;\n    Each one pretends that Fact of Life #2 doesn\u2019t exist.<\/p><p>There are many single-byte codes, and they don\u2019t solve the problem.&#13;\n    Each is only good for representing one small slice of human language.  They&#13;\n    can\u2019t solve the global text problem.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#12\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/011.png\" alt=\"Character codes\" width=\"512\" height=\"384\" \/><\/a><\/p><p>People tried creating double-byte character sets, but they were still&#13;\n    fragmented, serving different subsets of people.   There were multiple&#13;\n    standards in place, and they still weren\u2019t large enough to deal&#13;\n    with all the symbols needed.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#13\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/012.png\" alt=\"Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Unicode was designed to deal decisively with the issues with&#13;\n    older character codes.  Unicode assigns integers, known as code points, to&#13;\n    characters.  It has room for 1.1 million code points, and only 110,000 are&#13;\n    already assigned, so there\u2019s plenty of room for future growth.<\/p><p>Unicode\u2019s goal is to have everything.  It starts with ASCII, and&#13;\n    includes thousands of symbols, including the famous Snowman, covers all the&#13;\n    writing systems of the world, and is constantly being expanded.  For&#13;\n    example, the latest update gave us the symbol PILE OF POO.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#14\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/013.png\" alt=\"Sample Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Here is a string of six exotic Unicode characters.  Unicode code points&#13;\n    are written as 4-, 5-, or 6-digits of hex with a U+ prefix.  Every&#13;\n    character has an unambiguous full name which is always in uppercase ASCII.&#13;\n    <\/p><p>This string is designed to look like the word \u201cPython\u201d, but doesn\u2019t use&#13;\n    any ASCII characters at all.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#15\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/014.png\" alt=\"Encodings\" width=\"512\" height=\"384\" \/><\/a><\/p><p>So Unicode makes room for all of the characters we could ever need,&#13;\n    but we still have Fact of Life #1 to deal with: computers need bytes.&#13;\n    We need a way to represent Unicode code points as bytes in order to&#13;\n    store or transmit them.<\/p><p>The Unicode standard defines a number of ways to represent code points&#13;\n    as bytes.  These are called encodings.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#16\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/015.png\" alt=\"UTF-8\" width=\"512\" height=\"384\" \/><\/a><\/p><p>UTF-8 is easily the most popular encoding for storage and transmission&#13;\n    of Unicode.  It uses a variable number of bytes for each code point. The&#13;\n    higher the code point value, the more bytes it needs in UTF-8. ASCII&#13;\n    characters are one byte each, using the same values as ASCII, so ASCII is a&#13;\n    subset of UTF-8.<\/p><p>Here we show our exotic string as UTF-8.  The ASCII characters H and i are&#13;\n    single bytes, and other characters use two or three bytes depending on their&#13;\n    code point value. Some Unicode code points require four bytes, but we aren\u2019t &#13;\n    using any of those here.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#17\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/016.png\" alt=\"Python 2\" width=\"512\" height=\"384\" \/><\/a><\/p><p>OK, enough theory, let\u2019s talk about Python 2. In the slides, Python 2&#13;\n    samples have a big 2 in the upper-right corner, Python 3 samples will have&#13;\n    a big 3.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#18\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/017.png\" alt=\"Str vs Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>In Python 2, there are two different string data types.  A plain-old&#13;\n    string literal gives you a \u201cstr\u201d object, which stores bytes.  If you use a&#13;\n    \u201cu\u201d prefix, you get a \u201cunicode\u201d object, which stores code points.  In a&#13;\n    unicode string literal, you can use backslash-u to insert any Unicode code&#13;\n    point.<\/p><p>Notice that the word \u201cstring\u201d is problematic.  Both \u201cstr\u201d and \u201cunicode\u201d&#13;\n    are kinds of strings, and it\u2019s tempting to call either or both of them&#13;\n    \u201cstring,\u201d but it\u2019s better to use more specific terms to keep things&#13;\n    straight.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#19\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/018.png\" alt=\".encode() and .decode()\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Byte strings and unicode strings each have a method to convert it to&#13;\n    the other type of string.  Unicode&#13;\n    strings have a .encode() method that produces bytes, and byte strings have&#13;\n    a .decode() method that produces unicode. Each takes an argument, which&#13;\n    is the name of the encoding to use for the operation.<\/p><p>We can define a Unicode string named my_unicode, and see that it has&#13;\n    9 characters.  We can encode it to UTF-8 to create the my_utf8 byte string,&#13;\n    which has 19 bytes.  As you\u2019d expect, reversing the operation by decoding &#13;\n    the UTF-8 string produces the original Unicode string.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#20\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/019.png\" alt=\"Encoding errors\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Unfortunately, encoding and decoding can produce errors if the data&#13;\n    isn\u2019t appropriate for the specified encoding.  Here we try to encode our&#13;\n    exotic Unicode string to ASCII.  It fails because ASCII can only represent&#13;\n    charaters in the range 0 to 127, and our Unicode string has code points&#13;\n    outside that range.<\/p><p>The UnicodeEncodeError that\u2019s raised indicates the encoding being used,&#13;\n    in the form of the \u201ccodec\u201d (short for coder\/decoder), and the actual&#13;\n    position of the character that caused the problem.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#21\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/020.png\" alt=\"Decoding errors\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Decoding can also produce errors.  Here we try to decode our UTF-8&#13;\n    string as ASCII and get a UnicodeDecodeError because again, ASCII can only&#13;\n    accept values up to 127, and our UTF-8 string has bytes outside that&#13;\n    range.<\/p><p>Even UTF-8 can\u2019t decode any sequence of bytes.  Next we try to decode&#13;\n    some random junk, and it also produces a UnicodeDecodeError.   Actually,&#13;\n    one of UTF-8\u2019s advantages is that there are invalid sequences of bytes,&#13;\n    which helps to build robust systems:  mistakes in data won\u2019t be accepted as&#13;\n    if they were valid.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#22\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/021.png\" alt=\"Error handling\" width=\"512\" height=\"384\" \/><\/a><\/p><p>When encoding or decoding, you can specify what should happen when&#13;\n    the codec can\u2019t handle the data.  An optional second argument to encode or&#13;\n    decode specifies the policy.  The default value is \u201cstrict\u201d, which means&#13;\n    raise an error, as we\u2019ve seen.<\/p><p>A value of \u201creplace\u201d means, give me a standard replacement character.&#13;\n    When encoding, the replacement character is a question mark, so any code&#13;\n    point that can\u2019t be encoded using the specified encoding will simply&#13;\n    produce a \u201c?\u201d.<\/p><p>Other error handlers are more useful. \u201cxmlcharrefreplace\u201d produces&#13;\n    an HTML\/XML character entity reference, so that \\u01B4 becomes \u201c&amp;#436;\u201d&#13;\n    (hex 01B4 is decimal 436.)  This is very useful if you need to output&#13;\n    unicode for an HTML file.<\/p><p>Notice that different error policies are used for different reasons.&#13;\n    \u201cReplace\u201d is a defensive mechanism against data that cannot be interpreted,&#13;\n    and loses information. \u201cXmlcharrefreplace\u201d preserves all the original&#13;\n    information, and is used when outputting data where XML escapes are&#13;\n    acceptable.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#23\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/022.png\" alt=\"Error handling\" width=\"512\" height=\"384\" \/><\/a><\/p><p>You can also specify error handling when decoding.  \u201cIgnore\u201d will drop&#13;\n    bytes that can\u2019t decode properly.  \u201cReplace\u201d will insert a Unicode U+FFFD,&#13;\n    \u201cREPLACEMENT CHARACTER\u201d for problem bytes.  Notice that since the decoder&#13;\n    can\u2019t decode the data, it doesn\u2019t know how many Unicode characters were&#13;\n    intended.  Decoding our UTF-8 bytes as ASCII produces 16 replacement&#13;\n    characters, one for each byte that couldn\u2019t be decoded, while those bytes&#13;\n    were meant to only produce 6 Unicode characters.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#24\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/023.png\" alt=\"Implicit conversion\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Python 2 tries to be helpful when working with unicode and byte strings.&#13;\n    If you try to perform a string operation that combines a unicode string&#13;\n    with a byte string, Python 2 will automatically decode the byte string to&#13;\n    produce a second unicode string, then will complete the operation with the&#13;\n    two unicode strings.<\/p><p>For example, we try to concatenate a unicode \u201cHello \u201d with a byte string&#13;\n    \u201cworld\u201d.  The result is a unicode \u201cHello world\u201d.  On our behalf, Python 2&#13;\n    is decoding the byte string \u201cworld\u201d using the ASCII codec.  The encoding&#13;\n    used for these implicit decodings is the value of&#13;\n    sys.getdefaultencoding().<\/p><p>The implicit encoding is ASCII because it\u2019s the only safe guess: ASCII&#13;\n    is so widely accepted, and is a subset of so many encodings, that it\u2019s &#13;\n    unlikely to produce false positives.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#25\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/024.png\" alt=\"Implicit decoding errors\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Of course, these implicit decodings are not immune to decoding errors.  If you&#13;\n    try to combine a byte string with a unicode string and the byte string can\u2019t be&#13;\n    decoded as ASCII, then the operation will raise a UnicodeDecodeError.<\/p><p>This is the source of those painful UnicodeErrors.  Your code&#13;\n    inadvertently mixes unicode strings and byte strings, and as long as the&#13;\n    data is all ASCII, the implicit conversions silently succeed.  Once a&#13;\n    non-ASCII character finds its way into your program, an implicit decode&#13;\n    will fail, causing a UnicodeDecodeError.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#26\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/025.png\" alt=\"Python 2 is \u201chelpful\u201d\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Python 2\u2019s philosophy was that unicode strings and byte strings are &#13;\n    confusing, and it tried to ease your burden by automatically converting&#13;\n    between them, just as it does for ints and floats.  But the conversion&#13;\n    from int to float can\u2019t fail, while byte string to unicode string can.<\/p><p>Python 2 silently glosses over byte to unicode conversions, making it&#13;\n    much easier to write code that deals with ASCII.  The price you pay is&#13;\n    that it will fail with non-ASCII data.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#27\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/026.png\" alt=\"Other implicit conversions\" width=\"512\" height=\"384\" \/><\/a><\/p><p>There are lots of ways to combine two strings, and all of them will&#13;\n    decode bytes to unicode, so you have to watch out for them.<\/p><p>Here we use an ASCII format string, with unicode data.  The format&#13;\n    string will be decoded to unicode, then the formatting performed, resulting&#13;\n    in a unicode string.<\/p><p>Next we switch the two:  A unicode format string and a byte string again&#13;\n    combine to produce a unicode string, because the byte string data is&#13;\n    decoded as ASCII.<\/p><p>Even just attempting to print a unicode string will cause an implicit&#13;\n    encoding: output is always bytes, so the unicode string has to be encoded&#13;\n    into bytes before it can be printed.<\/p><p>The next one is truly confusing: we ask to encode a byte string to UTF-8,&#13;\n    and get an error about not being about to decode as ASCII!  The problem here&#13;\n    is that byte strings can\u2019t be encoded: remember encode is how you turn unicode&#13;\n    into bytes.  So to perform the encoding you want, Python 2 needs a unicode&#13;\n    string, which it tries to get by implicitly decoding your bytes as ASCII.<\/p><p>So you asked to encode to UTF-8, and you get an error about decoding ASCII.&#13;\n    It pays to look carefully at the error, it has clues about what operation is&#13;\n    being attempted, and how it failed.<\/p><p>Lastly, we encode an ASCII string to UTF-8, which is silly, encode&#13;\n    should be used on unicode string.  To make it work, Python performs the&#13;\n    same implicit decode to get a unicode string we can encode, but since the&#13;\n    string is ASCII, it succeeds, and then goes on to encode it as UTF-8,&#13;\n    producing the original byte string, since ASCII is a subset of UTF-8.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#28\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/027.png\" alt=\"Bytes and Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>This is the most important Fact of Life: bytes and unicode are both&#13;\n    important, and you need to deal with both of them.  You can\u2019t pretend&#13;\n    that everything is bytes, or everything is unicode.  You need to use &#13;\n    each for their purpose, and explicitly convert between them as needed.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#29\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/028.png\" alt=\"Python 3\" width=\"512\" height=\"384\" \/><\/a><\/p><p>We\u2019ve seen the source of Unicode pain in Python 2, now let\u2019s take&#13;\n    a look at Python 3.  The biggest change from Python 2 to Python 3 is &#13;\n    their treatment of Unicode.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#30\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/029.png\" alt=\"Str vs bytes\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Just as in Python 2, Python 3 has two string types, one for unicode and&#13;\n    one for bytes, but they are named differently.<\/p><p>Now the \u201cstr\u201d type that you get from a plain string literal stores unicode,&#13;\n    and the \u201cbytes\u201d types stores bytes.  You can create a bytes literal with a b&#13;\n    prefix.<\/p><p>So \u201cstr\u201d in Python 2 is now called \u201cbytes,\u201d and \u201cunicode\u201d in Python 2 is&#13;\n    now called \u201cstr\u201d.  This makes more sense than the Python 2 names, since &#13;\n    Unicode is how you want all text stored, and byte strings are only for &#13;\n    when you are dealing with bytes.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#31\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/030.png\" alt=\"No coercion!\" width=\"512\" height=\"384\" \/><\/a><\/p><p>The biggest change in the Unicode support in Python 3 is that there is&#13;\n    no automatic decoding of byte strings.  If you try to combine a byte string&#13;\n    with a unicode string, you will get an error all the time, regardless of&#13;\n    the data involved!<\/p><p>All of those operations I showed where Python 2 silently converted byte&#13;\n    strings to unicode strings to complete an operation, every one of them is&#13;\n    an error in Python 3.<\/p><p>In addition, Python 2 considers a unicode string and a byte string&#13;\n    equal if they contain the same ASCII bytes, and Python 3 won\u2019t.  A&#13;\n    consequence of this is that unicode dictionary keys can\u2019t be found with&#13;\n    byte strings, and vice-versa, as they can be in Python 2.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#32\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/031.png\" alt=\"Python 3 pain\" width=\"512\" height=\"384\" \/><\/a><\/p><p>This drastically changes the nature of Unicode pain in Python 3.  In&#13;\n    Python 2, mixing unicode and bytes succeeds so long as you only use ASCII&#13;\n    data.  In Python 3, it fails immediately regardless of the data.<\/p><p>So Python 2\u2019s pain is deferred: you think your program is correct, and&#13;\n    find out later that it fails with exotic characters.<\/p><p>With Python 3, your code fails immediately, so even if you are&#13;\n    only handling ASCII, you have to explicitly deal with the difference&#13;\n    between bytes and unicode.<\/p><p>Python 3 is strict about the difference between bytes and unicode.&#13;\n    You are forced to be clear in your code which you are dealing with.&#13;\n    This has been controversial, and can cause you pain.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#33\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/032.png\" alt=\"Reading files\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Because of this new strictness, Python 3 has changed how you read files.&#13;\n    Python has always had two modes for reading files: binary and text.  In Python 2, it &#13;\n    only affected the line endings, and on Unix platforms, even that was a no-op.<\/p><p>In Python 3, the two modes produce different results.  When you open a file in text&#13;\n    mode, either with \u201cr\u201d, or by defaulting the mode entirely, the data read from the &#13;\n    file is implicitly decoded into Unicode, and you get str objects.<\/p><p>If you open a file in binary mode, by supplying \u201crb\u201d as the mode, then the data&#13;\n    read from the file is bytes, with no processing done on them.<\/p><p>The implicit conversion from bytes to unicode uses the encoding returned&#13;\n    from locale.getpreferredencoding(), and it may not give you the results you&#13;\n    expect.  For example, when we read hi_utf8.txt, it\u2019s being decoded using&#13;\n    the locale\u2019s preferred encoding, which since I created these samples on&#13;\n    Windows, is \u201ccp1252\u201d.  Like ISO 8859-1, CP-1252 is a one-byte character&#13;\n    code that will accept any byte value, so it will never raise a&#13;\n    UnicodeDecodeError.  That also means that it will happily decode data that&#13;\n    isn\u2019t actually CP-1252, and produce garbage.<\/p><p>To get the file read properly, you should specify an encoding to use.&#13;\n    The open() function now has an optional encoding parameter.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#34\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/033.png\" alt=\"Pain relief\" width=\"512\" height=\"384\" \/><\/a><\/p><p>OK, so how do we deal with all this pain? The good news is that the rules&#13;\n    to remember are simple, and they\u2019re the same for Python 2 and Python 3.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#35\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/034.png\" alt=\"Pro tip #1: Unicode sandwich\" width=\"512\" height=\"384\" \/><\/a><\/p><p>As we saw with Fact of Life #1, the data coming into and going out of&#13;\n    your program must be bytes.  But you don\u2019t need to deal with bytes on the&#13;\n    inside of your program.  The best strategy is to decode incoming bytes as&#13;\n    soon as possible, producing unicode.  You use unicode throughout your&#13;\n    program, and then when outputting data, encode it to bytes as late as&#13;\n    possible.<\/p><p>This creates a Unicode sandwich: bytes on the outside, Unicode on the&#13;\n    inside.<\/p><p>Keep in mind that sometimes, a library you\u2019re using may do some of these&#13;\n    conversions for you. The library may present you with Unicode input, or&#13;\n    will accept Unicode for output, and the library will take care of the edge&#13;\n    conversion to and from bytes.  For example, Django provides Unicode, as&#13;\n    does the json module.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#36\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/035.png\" alt=\"Pro tip #2: Know what you have\" width=\"512\" height=\"384\" \/><\/a><\/p><p>The second rule is, you have to know what kind of data you are dealing with.&#13;\n    At any point in your program, you need to know whether you have a byte string&#13;\n    or a unicode string.  This shouldn\u2019t be a matter of guessing, it should be&#13;\n    by design.<\/p><p>In addition, if you have a byte string, you should know what encoding it&#13;\n    is if you ever intend to deal with it as text.<\/p><p>When debugging your code, you can\u2019t simply print a value to see what it&#13;\n    is.  You need to look at the type, and you may need to look at the repr of&#13;\n    the value in order to get to the bottom of what data you have.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#37\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/036.png\" alt=\"Encoding is out-of-band\" width=\"512\" height=\"384\" \/><\/a><\/p><p>I said you have to understand what encoding your byte strings are.&#13;\n    Here\u2019s Fact of Life #4: You can\u2019t determine the encoding of a byte string&#13;\n    by examining it.  You need to know through other means. For example,&#13;\n    many protocols include ways to specify the encoding.  Here we have&#13;\n    examples from HTTP, HTML, XML, and Python source files.  You may&#13;\n    also know the encoding by prior arrangement, for example, the spec&#13;\n    for a data source may specify the encoding.<\/p><p>There are ways to guess at the encoding of the bytes, but they are&#13;\n    just guesses. The only way to be sure of the encoding is to find it &#13;\n    out some other way.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#38\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/037.png\" alt=\"Poo happens\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Here\u2019s an example of our exotic Unicode string, encoded as UTF-8, and&#13;\n    then mistakenly decoded in a variety of encodings.  As you can see, decoding&#13;\n    with an incorrect encoding might succeed, but produce the wrong characters.&#13;\n    Your program can\u2019t tell it\u2019s decoding wrong, only when people try to read&#13;\n    the text will you know something has gone wrong.<\/p><p>This is a good demonstration of Fact of Life #4: the same stream of bytes&#13;\n    is decodable using a number of different encodings.  The bytes themselves&#13;\n    don\u2019t indicate what encoding they use.<\/p><p>BTW, there\u2019s a term for this garbage display, from the Japanese who have&#13;\n    been dealing with this for years and years: Mojibake.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#39\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/038.png\" alt=\"Data is dirty\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Unfortunately, because the encoding for bytes has to be communicated&#13;\n    separately from the bytes themselves, sometimes the specified encoding is&#13;\n    wrong.  For example, you may pull an HTML page from a web server, and the&#13;\n    HTTP header claims the page is 8859-1, but in fact, it is encoded with&#13;\n    UTF-8.<\/p><p>In some cases, the encoding mismatch will succeed and cause mojibake.&#13;\n    Other times, the encoding is invalid for the bytes, and will cause a&#13;\n    UnicodeError of some sort.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#40\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/039.png\" alt=\"Pro tip #3: Test Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>It should go without saying: you should explicitly test your Unicode&#13;\n    support.  To do this, you need challenging Unicode data to pump through&#13;\n    your code.  If you are an English-only speaker, you may have a problem&#13;\n    doing this, because lots of non-ASCII data is hard to read.  Luckily, the&#13;\n    variety of Unicode code points mean you can construct complex Unicode&#13;\n    strings that are still readable by English speakers.<\/p><p>Here\u2019s an example of overly-accented text, readable pseudo-ASCII&#13;\n    text, and upside-down text.  One good source of these sorts of strings&#13;\n    are various web sites that offer strings like this for teenagers to&#13;\n    paste into social networking sites.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#41\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/040.png\" alt=\"More Unicode\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Depending on your application, you may need to dig deeper into&#13;\n    the other complexities in the Unicode world.  There are many&#13;\n    details that I haven\u2019t covered here, and they can be very involved.&#13;\n    I call this Fact of Life #5\u00bd because you may not have to&#13;\n    deal with any of this.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#42\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/041.png\" alt=\"Facts of Life\" width=\"512\" height=\"384\" \/><\/a><\/p><p>To review, these are the five unavoidable Facts of Life:<\/p><ol><li>All input and output of your program is bytes.<\/li><li>The world needs more than 256 symbols to communicate text.<\/li><li>Your program has to deal with both bytes and Unicode.<\/li><li>A stream of bytes can\u2019t tell you its encoding.<\/li><li>Encoding specifications can be wrong.<\/li><\/ol><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#43\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/042.png\" alt=\"Pro tips\" width=\"512\" height=\"384\" \/><\/a><\/p><p>These are the three Pro Tips to keep in mind as you build your software&#13;\n    to keep your code Unicode-clean:<\/p><ol><li>Unicode sandwich: keep all text in your program as Unicode, and convert&#13;\n    as close to the edges as possible.<\/li><li>Know what your strings are: you should be able to explain which of your&#13;\n    strings are Unicode, which are bytes, and for your byte strings, what &#13;\n    encoding they use.<\/li><li>Test your Unicode support.  Use exotic strings throughout your test&#13;\n    suites to be sure you\u2019re covering all the cases.<\/li><\/ol><p>If you follow these tips, you\u2019ll write good solid code that deals&#13;\n    well with Unicode, and won\u2019t fall over no matter how wild the Unicode&#13;\n    it encounters.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#44\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/043.png\" alt=\"See also\" width=\"512\" height=\"384\" \/><\/a><\/p><p>Other resources you might find helpful:<\/p><p>Joel Spolsky wrote <a href=\"http:\/\/www.joelonsoftware.com\/articles\/Unicode.html\" rel=\"external\">The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)<\/a>,&#13;\n    which covers how Unicode works and why.  It has no Python-specific information,&#13;\n    but is better written than this talk!<\/p><p>If you need to deal with the semantics of arbitrary Unicode characters, the <a href=\"http:\/\/docs.python.org\/library\/unicodedata.html\" rel=\"external\">unicodedata module<\/a>&#13;\n    in the Python standard library has functions that can help.<\/p><p>For testing Unicode, the various <a href=\"http:\/\/fsymbols.com\/generators\/encool\" rel=\"external\">\u201cfancy\u201d text generators<\/a>&#13;\n    for use on social networks work great.<\/p><p class=\"figure\"><a href=\"https:\/\/nedbatchelder.com\/text\/unipain\/unipain.html#45\"><img src=\"https:\/\/nedbatchelder.com\/text\/unipain_pix\/044.png\" alt=\"Thank You\" width=\"512\" height=\"384\" \/><\/a><\/p>","created_at":"2019-08-22T20:54:01+0000","updated_at":"2019-08-22T20:54:08+0000","published_by":["about"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":21,"domain_name":"nedbatchelder.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["vpn","dns","networking","ssh"],"is_public":false,"id":999,"title":"sshuttle\/sshuttle","url":"https:\/\/github.com\/sshuttle\/sshuttle","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p>As far as I know, sshuttle is the only program that solves the following\ncommon case:<\/p>\n<ul><li>Your client machine (or router) is Linux, FreeBSD, or MacOS.<\/li>\n<li>You have access to a remote network via ssh.<\/li>\n<li>You don't necessarily have admin access on the remote network.<\/li>\n<li>The remote network has no VPN, or only stupid\/complex VPN\nprotocols (IPsec, PPTP, etc). Or maybe you <em>are<\/em> the\nadmin and you just got frustrated with the awful state of\nVPN tools.<\/li>\n<li>You don't want to create an ssh port forward for every\nsingle host\/port on the remote network.<\/li>\n<li>You hate openssh's port forwarding because it's randomly\nslow and\/or stupid.<\/li>\n<li>You can't use openssh's PermitTunnel feature because\nit's disabled by default on openssh servers; plus it does\nTCP-over-TCP, which has terrible performance (see below).<\/li>\n<\/ul>\n<h2><a id=\"user-content-obtaining-sshuttle\" class=\"anchor\" aria-hidden=\"true\" href=\"#obtaining-sshuttle\"><\/a>Obtaining sshuttle<\/h2>\n<ul><li><p>Debian stretch or later:<\/p>\n<pre>apt-get install sshuttle\n<\/pre>\n<\/li>\n<li><p>Arch Linux:<\/p>\n<pre>pacman -S sshuttle\n<\/pre>\n<\/li>\n<li><p>Fedora:<\/p>\n<pre>dnf install sshuttle\n<\/pre>\n<\/li>\n<li><p>From PyPI:<\/p>\n<pre>sudo pip install sshuttle\n<\/pre>\n<\/li>\n<li><p>Clone:<\/p>\n<pre>git clone https:\/\/github.com\/sshuttle\/sshuttle.git\ncd sshuttle\nsudo .\/setup.py install\n<\/pre>\n<\/li>\n<\/ul><p>It is also possible to install into a virtualenv as a non-root user.<\/p>\n<ul><li><p>From PyPI:<\/p>\n<pre>virtualenv -p python3 \/tmp\/sshuttle\n. \/tmp\/sshuttle\/bin\/activate\npip install sshuttle\n<\/pre>\n<\/li>\n<li><p>Clone:<\/p>\n<pre>virtualenv -p python3 \/tmp\/sshuttle\n. \/tmp\/sshuttle\/bin\/activate\ngit clone https:\/\/github.com\/sshuttle\/sshuttle.git\ncd sshuttle\n.\/setup.py install\n<\/pre>\n<\/li>\n<li><p>Homebrew:<\/p>\n<pre>brew install sshuttle\n<\/pre>\n<\/li>\n<\/ul>\n<h2><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h2>\n<p>The documentation for the stable version is available at:\n<a href=\"https:\/\/sshuttle.readthedocs.org\/\" rel=\"nofollow\">https:\/\/sshuttle.readthedocs.org\/<\/a><\/p>\n<p>The documentation for the latest development version is available at:\n<a href=\"https:\/\/sshuttle.readthedocs.org\/en\/latest\/\" rel=\"nofollow\">https:\/\/sshuttle.readthedocs.org\/en\/latest\/<\/a><\/p>\n<\/article>","created_at":"2019-08-21T17:57:18+0000","updated_at":"2019-08-21T17:57:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/8774006?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":998,"title":"SB-489","url":"https:\/\/lola.atlassian.net\/browse\/SB-489","created_at":"2019-08-20T15:48:34+0000","updated_at":"2019-08-20T15:48:34+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"lola.atlassian.net"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":997,"title":"The Architecture of Open Source Applications","url":"http:\/\/www.aosabook.org\/en\/index.html","content":"<p>\n            Architects look at thousands of buildings during their\n            training, and study critiques of those buildings written\n            by masters.  In contrast, most software developers only\n            ever get to know a handful of large programs\n            well\u2014usually programs they wrote\n            themselves\u2014and never study the great programs of\n            history.  As a result, they repeat one another's mistakes\n            rather than building on one another's successes.\n          <\/p><p>\n            Our goal is to change that.  In these two books, the authors of\n            four dozen open source applications explain how their software\n            is structured, and why.  What are each program's major\n            components?  How do they interact?  And what did their builders\n            learn during their development?  In answering these questions,\n            the contributors to these books provide unique insights into how\n            they think.\n          <\/p><p>\n            If you are a junior developer, and want to learn how your\n            more experienced colleagues think, these books are the place\n            to start.  If you are an intermediate or senior developer,\n            and want to see how your peers have solved hard design\n            problems, these books can help you too.\n          <\/p><p>\n            Follow us on our blog at <a href=\"http:\/\/aosabook.org\/blog\/\">http:\/\/aosabook.org\/blog\/<\/a>\n            or on Twitter at <a href=\"http:\/\/twitter.com\/aosabook\">@aosabook<\/a>\n            and using the <a href=\"http:\/\/twitter.com\/#!\/search\/%23aosa\">#aosa<\/a> hashtag.\n          <\/p>","created_at":"2019-08-18T18:05:49+0000","updated_at":"2019-08-18T18:05:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.aosabook.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":996,"title":"Paged Out!","url":"https:\/\/pagedout.institute\/","content":"<div class=\"row\"><div class=\"col\"><h3>What is Paged Out!?<\/h3><p><b>Paged Out! is a new experimental (one article == one page) free magazine<\/b> about programming (especially programming tricks!), <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hacker\">hacking<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Security_hacker\">security hacking<\/a>, retro computers, modern computers, electronics, demoscene, and other similar topics.<\/p><p>It's <b>made by the community for the community<\/b> - the project is led by <a href=\"https:\/\/gynvael.coldwind.pl\/?id=50\">Gynvael Coldwind<\/a> with <a href=\"https:\/\/pagedout.institute\/?page=about.php\">multiple folks helping<\/a>. And it's not-for-profit (though in time we hope it will be self-sustained) - this means that the issues will always be free to download, share and print.<\/p><p>If you're interested in more details, check our our <a href=\"https:\/\/pagedout.institute\/?page=faq.php\">FAQ<\/a> and <a href=\"https:\/\/pagedout.institute\/?page=about.php\">About<\/a> pages!<\/p><h4>Zine Issues<\/h4><div><div><a href=\"https:\/\/pagedout.institute\/download\/PagedOut_001_beta1.pdf\"><img src=\"https:\/\/pagedout.institute\/img\/issue_1_cover_small.png\" alt=\"image\" \/><\/a><br \/>Cover art by<a href=\"https:\/\/www.deviantart.com\/refiend\">ReFiend<\/a>.<br \/><\/div><div><p>The first Paged Out! issue has arrived!\n        <\/p><ul><li><a href=\"https:\/\/pagedout.institute\/download\/PagedOut_001_beta1.pdf\">Paged Out! #1 (web PDF)<\/a> (12MB)<\/li>\n          \n        <\/ul><p><b>Note<\/b>: This is a \"beta build\" of the PDF, i.e. we will be re-publishing it with compatibility\/size\/menu\/layout improvements multiple times in the next days. We'll also publish soon:\n      <\/p><ul><li>PDFs for printing (A4+bleed, US Letter+bleed) - our scripts don't build them yet.<\/li>\n        <li>High-res wallpaper based on Issue #1 cover art.<\/li>\n      <\/ul><\/div><\/div><h4>Next issue<\/h4><p>If you like our work, <b><a href=\"https:\/\/pagedout.institute\/?page=cfp.php\">how about writing an article for Paged Out!<\/a>?<\/b> It's only one page after all - easy. ;)<\/p><p>\n      <b>Next issue progress tracker<\/b> (unit of measurement: article count):<br \/><small>(article submission deadline for Issue #2 is 20 Oct 2019)<\/small>\n      <\/p><br \/><div><a href=\"https:\/\/pagedout.institute\/img\/po2_cover_design_rc1.jpg\"><img src=\"https:\/\/pagedout.institute\/img\/t_po2_cover_design_rc1.jpg\" alt=\"image\" \/><\/a><br \/><a href=\"https:\/\/pagedout.institute\/?page=cfp.php\">Call For Articles<\/a>for Issue #2!<br \/>Cover art by Vlad Gradobyk (<a href=\"https:\/\/instagram.com\/vladgradobyk\">Insta<\/a>,<a href=\"https:\/\/facebook.com\/gradobyk.graphic\">FB<\/a>).<br \/><\/div><h4>Notify me when the new issue is out!<\/h4><p>Sure! Just subscribe to this newsletter group: <a href=\"https:\/\/groups.google.com\/forum\/#!forum\/pagedout-notifications\">pagedout-notifications (googlegroups.com)<\/a> (be sure to select you want e-mail notifications about every message when subscribing).<\/p><p>We will only send e-mails to this group about new Paged Out! issues (both the free electronic ones and special issues if we ever get to that). No spam will be sent there and your e-mail will be visible only to group owners.<\/p><\/div><\/div>","created_at":"2019-08-12T03:25:23+0000","updated_at":"2019-08-12T03:25:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"pagedout.institute","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":995,"title":"Alan Kay's answer to Experienced programmers and computer scientists, what are some really old (or even nearly forgotten) books you think every new programmer should read?","url":"https:\/\/www.quora.com\/Experienced-programmers-and-computer-scientists-what-are-some-really-old-or-even-nearly-forgotten-books-you-think-every-new-programmer-should-read\/answer\/Alan-Kay-11?share=1","content":"<div id=\"wseHt23H49\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para u-ltr u-text-align--start\">I love that \u201c2006\u201d and \u201c2008\u201d (in another answer) must be considered <i>\u201creally old\u201d<\/i> (which is what the question requests) \u2026<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">I\u2019m still a big fan of the <b>\u201cLisp 1.5 Programmers Manual\u201d<\/b> (MIT Press \u2014 still in print). This version of the language is no longer with us, but the book \u2014 first written ca 1962 \u2014 by John McCarthy, who invented, and his colleagues, who implemented, is a perfect classic.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">It starts with a version of John\u2019s first papers about Lisp, and develops the ideas in a few pages of examples to culminate on page 13 with Lisp <i>eval<\/i> and <i>apply<\/i> defined in itself. There are many other thought provoking ideas and examples throughout the rest of the book.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">The way to grow from this book is to deeply learn what they did and how they did it, and then try to rewrite page 13 in a number of ways. How nicely can this be written in \u201ca lisp\u201d using recursion. How nicely can this be written without recursion? (In both cases, look ahead in the book to see that Lisp 1.5 had gotten to the idea of EXPRs and FEXPRs (functions which don\u2019t <i>eval<\/i> their arguments before the call \u2014 thus they can be used to replace all the \u201cspecial forms\u201d \u2014 do a Lisp made from FEXPRs and get the rest by definition, etc.).<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">What is a neat bootstrapping path? How could you combine this with Val Shorre\u2019s \u201cMeta II\u201d programmatic parser to make a really extensible language? What does it take to get to \u201cobjects\u201d? What are three or four really interesting (and different ways) to think about objects here? (Hints: how many different ways can you define \u201cclosures\u201d in a language that executes? What about using Lisp atoms as a model for objects? Etc.)<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">The idea is that Lisp is not just a language but a really deep \u201cbuilding material\u201d that is tidy enough to \u201cthink with\u201d not just make things (it\u2019s a \u201cbuilding material\u201d for thoughts as well as computer processes).<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">Dani Richard reminded me to mention: <b>\u201cComputation: Finite and Infinite Machines\u201d <\/b>by Marvin Minsky (Prentice-Hall, 1967), which \u2014 since it is one of my favorite books of all time \u2014 I\u2019m surprised I didn\u2019t include in the original list. Marvin could really write, and in this book he is at his best. It is actually a \u201cmath book\u201d \u2014 with lots of ideas, theorems, proofs, etc., \u2014 but presented in the friendliest way imaginable by a great mind who treated everyone \u2014 including children \u2014 as equal to him, and as fellow appreciators of great ideas. There are lots of interesting things to ponder in this book, but perhaps it is the <i>approach<\/i> that beckons to the reader to start thinking \u201clike this\u201d that is the most rewarding.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\"><b>\u201cAdvances in Programming and Non-Numerical Computation\u201d<\/b> (Ed. L. Fox) mid-60s. The papers presented at a 1963 summer workshop in the UK. The most provocative ones were by Christopher Strachey and several by Peter Landin. This was one of the books that Bob Barton had us read in his famous advanced systems design class in 1967.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">Try <b>\u201cThe Mythical Man-Month\u201d<\/b> by Fred Brooks, for an early look and experience with timeless truths (and gotchas) from systems building with teams \u2026<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">Try <b>\u201cThe Sciences of the Artificial\u201d<\/b> by Herb Simon. A much stronger way to think about computing \u2014 and what \u201cComputer Science\u201d might mean \u2014 by a much stronger thinker than most today.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\"><b>\u201cA Programming Language\u201d<\/b> by Ken Iverson (ca 1962). This has the same thought expanding properties of Lisp. And, like Lisp, the way to learn from these really old ideas is to concentrate on what is unique and powerful in the approach (we know how to better improve both Lisp and APL today, but the deep essence is perhaps easier to grasp in the original manifestations of the ideas). Another book that Barton had us read.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">I like Dave Fisher\u2019s 1970 CMU Thesis \u2014 <b>\u201cControl Structures for Programming Languages\u201d<\/b> \u2014 most especially the first 100 pages. Still a real gem for helping to think about design and implementations.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">More recent: (80s) <b>\u201cThe Meta-Object Protocol\u201d<\/b> by Kiczales, et al. The first section and example is a must to read and understand.<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\"><b>Joe Armstrong\u2019s PhD thesis <\/b>\u2014 after many years of valuable experience with Erlang \u2014 was published as a book ca 2003 \u2026<\/p><p class=\"ui_qtext_para u-ltr u-text-align--start\">Lots more out there for curious minds \u2026.<\/p><\/div><\/div><div id=\"wseHt23H54\"><div id=\"__w2_wseHt23H48_survey_wrapper\"><\/div><\/div><\/div>","created_at":"2019-08-09T13:21:28+0000","updated_at":"2019-08-09T13:21:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"www.quora.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","elisp","testing"],"is_public":false,"id":994,"title":"Testing Emacs programs with Buttercup","url":"http:\/\/www.modernemacs.com\/post\/testing-emacs\/","content":"<p><a href=\"https:\/\/github.com\/jorgenschaefer\/emacs-buttercup\" title=\"Buttercup\">Buttercup<\/a> is a testing framework for emacs-lisp. It is used by large projects\nlike <a href=\"https:\/\/github.com\/clojure-emacs\/cider\" title=\"Clojure's CIDER\">Clojure's CIDER<\/a> to write clean, concise, and descriptive tests.<\/p><p>I introduce Buttercup and build up to advanced usages with the faint, unlikely\ndream that some Emacs programmer decides to add tests to their library...<\/p>\n<h3 id=\"introduction\">Introduction<\/h3>\n<p>Buttercup's entry points are: <code>describe<\/code>, <code>it<\/code>, and <code>expect<\/code>.<\/p>\n<p>We <code>describe<\/code> a test suite with a name. Test cases within the possibly nested\nsuites are done with <code>it<\/code> and assertions as <code>expect<\/code> blocks within.<\/p>\n<pre class=\"language-lisp\">(describe \"Four\"\n  (describe \"comparisons\"\n    (it \"is greater than one\"\n      (expect (&gt; 4 1)))\n    (it \"and less than five\"\n      (expect 4 :to-be-less-than 5)))\n  (it \"is a number\"\n    (expect (numberp 4))))\n<\/pre>\n<p>passes with testing output:<\/p>\n<pre>Four\n  comparisons\n    is greater than one (0.24ms)\n    and less than five (0.12ms)\n  is a number (0.09ms)\n<\/pre>\n<h3 id=\"setup-and-teardown\">Setup and Teardown<\/h3>\n<p>Buttercup provides <code>before-each<\/code>, <code>after-each<\/code>, <code>before-all<\/code>, and <code>after-all<\/code> to\nreduce boilerplate with setting up and tearing down test suites.<\/p>\n<pre class=\"language-lisp\">(describe \"Lisp mode syntax\"\n  (before-all (set-syntax-table lisp-mode-syntax-table))\n  (after-each (delete-region (point-min) (point-max)))\n  (it \"sets comments\"\n    (insert \";; foo\")\n    (expect (nth 4 (syntax-ppss))))\n  (it \"sets strings\"\n    (insert \"\\\"foo\\\"\")\n    (backward-char)\n    (expect (nth 3 (syntax-ppss)))))\n<\/pre>\n<h3 id=\"matchers\">Matchers<\/h3>\n<p>The <code>expect<\/code> has more utility than simple tests of truth. <strong>Matchers<\/strong> are\nkeywords that tailor the expectation.<\/p>\n<p>Some example matcher expansions:<\/p>\n<dl><dt>:to-be<\/dt>\n<dd><code>(eq foo bar)<\/code><\/dd>\n<dt>:to-equal<\/dt>\n<dd><code>(equal foo bar)<\/code><\/dd>\n<dt>:to-be-in<\/dt>\n<dd><code>(member foo bar)<\/code><\/dd>\n<dt>:to-be-close-to<\/dt>\n<dd><code>(foo bar precisision)<\/code><\/dd>\n<dt>:to-throw<\/dt>\n<dd><code>(expr &amp;optional signal signal-args)<\/code><\/dd>\n<\/dl><p>Some other more advanced matchers include: <code>:to-have-same-items-as<\/code>, <code>:to-match<\/code>, and <code>:to-have-been-called<\/code>.<\/p>\n<p>These matchers may be combined too: eg. <code>(expect 4 :not :to-be-greater-than 5)<\/code>.<\/p>\n<p>Matchers are more than just transforms+comparisons. They give information about\nthe failure.<\/p>\n<pre class=\"language-lisp\">(describe \"Example Matchers\"\n  (it \"regexes\"\n    (expect (s-concat \"foo\" \"bar\")\n            :to-match (rx word-start \"foo\" word-end))))\n<\/pre>\n<p><code>Expected `(s-concat \"foo\" \"bar\")' with value \"foobar\" to match the regexp \"\\\\&lt;foo\\\\&gt;\", but instead it was \"foobar\".<\/code><\/p>\n<h3 id=\"running-it\">Running It<\/h3>\n<p>I recommend using <code>Cask<\/code> and executing tests with <code>cask exec buttercup -L .<\/code> in the project root.<\/p>\n<p>For example, have a file named <code>Cask<\/code> in the project root with:<\/p>\n<pre class=\"language-lisp\">(source gnu)\n(source melpa)\n(package-file \"test-stuff-i-beg-you-mode.el\")\n;; Project Dependencies\n(depends-on \"dash\")\n;; Additional Testing Dependencies\n(development\n (depends-on \"buttercup\")\n (depends-on \"faceup\"))\n<\/pre>\n<p>A folder named <code>test\/<\/code> should be present and contain <code>test-stuff-i-beg-you-mode-test.el<\/code>.<\/p>\n<p>This file should have your tests, set up the load path if needed, and require\neverything you need.<\/p>\n<p>Lastly I will mention some other useful features before diving in to Buttercup:<\/p>\n<ul><li>Variables can be defined with let syntax with <code>:var<\/code> in <code>describe<\/code> blocks.<\/li>\n<li>Buttercup has good support for spying on function calls.<\/li>\n<li>Adding an <code>x<\/code>, so it's <code>xit<\/code> and <code>xdescribe<\/code>, mark the test as pending so it<\/li>\n    \n<\/ul><p>You have written <code>yet-another-lisp-like-mode<\/code> you affectionately call <code>yall-mode<\/code> and want to test its indentation.<\/p>\n<p>Lets write a skeleton to test the simplest cases:<\/p>\n<pre class=\"language-lisp\">;; Want to test these two cases:\n;; (foo\n;;  bar)\n;; (foo bar\n;;      baz)\n(describe \"Indentation\"\n  (before-all (setq indent-line-function #'yall-indent-line))\n  (describe \"standard cases\"\n    (it \"opening line has one sexp - so indentation doesn't carry\"\n      (expect ???))\n    (it \"opening line has two+ sexps - so indentation carries\"\n      (expect ???))))\n<\/pre>\n<p>To test indentation - all we need is the text we expect, as the text alone\ndetermines the indent.<\/p>\n<p>Buttercup allows us to achieve this via <strong>custom matchers<\/strong>. We can bypass <em>all\nboilerplate<\/em> and write our expectations as simply as:<\/p>\n<pre class=\"language-lisp\">(expect \"\n(foo\n bar)\n\" :indented)\n<\/pre>\n<p>The macro <code>buttercup-define-matcher<\/code> allows defining our own matcher, that will\nperform transforms, assertions, and <em>give descriptive failures<\/em>.<\/p>\n<p>Lets implement our <code>:indented<\/code> matcher:<\/p>\n<pre class=\"language-lisp\">(defun yall-trim-indent (text)\n  \"Remove indentation from TEXT.\"\n  (-&gt;&gt; text s-lines (-map #'s-trim-left) (s-join \"\\n\")))\n(defun yall-buffer-string ()\n  \"Return buffer as text with beginning and ending empty space trimmed.\"\n  (s-trim (buffer-substring-no-properties (point-min) (point-max))))\n(buttercup-define-matcher :indented (text)\n  (let* ((text (s-trim (funcall text)))\n         (text-no-indent (yall-trim-indent text)))\n    (insert text-no-indent)\n    (indent-region-line-by-line (point-min) (point-max))\n    (let ((text-with-indent (yall-buffer-string)))\n      (delete-region (point-min) (point-max))\n      (if (string= text text-with-indent)\n          t\n        `(nil . ,(format \"\\nGiven indented text \\n%s\\nwas instead indented to \\n%s\\n\"\n                         text text-with-indent))))))\n<\/pre>\n<p>Now we can see the power of buttercup when we accidentally write:<\/p>\n<pre class=\"language-lisp\">(describe \"Indentation\"\n  (before-all (setq indent-line-function #'yall-indent-line))\n  (describe \"standard cases\"\n    (it \"opening line has two+ sexps - so indentation carries\"\n      (expect \"\n(foo bar\n      baz)\n\" :indented))))\n<\/pre>\n<p>and are given the failure:<\/p>\n<pre>FAILED:\nGiven indented text\n(foo bar\n      baz)\nwas instead indented to\n(foo bar\n     baz)\n<\/pre>\n<p>We know exactly what went wrong, with nearly all the implementation details\nseparated from the testcase with boilerplate just <code>:indented<\/code>.<\/p>\n<p>Testing Emacs programs doesn't have to be painful - buttercup is a great and\nbattle-tested library for writing quality Emacs programs.<\/p>","created_at":"2019-08-09T12:23:58+0000","updated_at":"2019-08-09T12:24:09+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"www.modernemacs.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":993,"title":"Managing code with Outlines | Modern Emacs","url":"http:\/\/www.modernemacs.com\/post\/outline-ivy\/","content":"<p>I consider outlines an under-utilized yet killer feature of Emacs.<\/p><p>This post is split into two parts:<\/p>\n<ol><li>Introducing and customizing <code>outline-mode<\/code> and <code>outshine-mode<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/ekaschalk\/.spacemacs.d\/blob\/master\/layers\/personal\/local\/outline-ivy\/outline-ivy.el\" title=\"My package\">My package<\/a> <code>outline-ivy<\/code> for jumping to outlines.<\/li>\n<\/ol><p><img src=\"http:\/\/www.modernemacs.com\/img\/outline-ivy.gif\" alt=\"\/img\/outline-ivy.gif\" title=\"\/img\/outline-ivy.gif\" \/><\/p>\n<h2 id=\"background\">Background<\/h2>\n<p>Outline-mode is a general framework for headers. Org-mode itself uses\noutline-mode.<\/p>\n<p>Headers are demarcated by the current major-mode's comment syntax, typically\nwith levels determined by the proceeding number of <code>*<\/code>.<\/p>\n<ul><li>Python: <code># *<\/code> is a level 1 header, <code># **<\/code> a level 2 header...<\/li>\n<li>Haskell: <code>--*<\/code> <code>--**<\/code><\/li>\n<li>Clojure: <code>;; *<\/code> <code>;; **<\/code><\/li>\n<\/ul><p>Emacs lisp uses <code>;;;<\/code>, <code>;;;;<\/code>, ... for compatibility with the many packages that\nuse the original, not asterisk-based, outline format.<\/p>\n<p>The package <code>outshine<\/code> gives utility like narrowing and cycling to these\noutlines.<\/p>\n<h2 id=\"setup\">Setup<\/h2>\n<p>Enable outlines and outshine with:<\/p>\n<pre class=\"language-lisp\">;; Require packages for following code\n(require 'dash)\n(require 'outshine)\n;; Required for outshine\n(add-hook 'outline-minor-mode-hook 'outshine-hook-function)\n;; Enables outline-minor-mode for *ALL* programming buffers\n(add-hook 'prog-mode-hook 'outline-minor-mode)\n<\/pre>\n<h2 id=\"keybindings\">Keybindings<\/h2>\n<p>I remap <code>outline-minor-mode-map<\/code> to mirror org-mode. Provided are evil and\nleader-key based bindings. Reference org-mode for developing your own\nemacs-style bindings.<\/p>\n<pre class=\"language-lisp\">;; Narrowing now works within the headline rather than requiring to be on it\n(advice-add 'outshine-narrow-to-subtree :before\n            (lambda (&amp;rest args) (unless (outline-on-heading-p t)\n                                   (outline-previous-visible-heading 1))))\n(spacemacs\/set-leader-keys\n  ;; Narrowing\n  \"nn\" 'outshine-narrow-to-subtree\n  \"nw\" 'widen\n  ;; Structural edits\n  \"nj\" 'outline-move-subtree-down\n  \"nk\" 'outline-move-subtree-up\n  \"nh\" 'outline-promote\n  \"nl\" 'outline-demote)\n(let ((kmap outline-minor-mode-map))\n  (define-key kmap (kbd \"M-RET\") 'outshine-insert-heading)\n  (define-key kmap (kbd \"&lt;backtab&gt;\") 'outshine-cycle-buffer)\n  ;; Evil outline navigation keybindings\n  (evil-define-key '(normal visual motion) kmap\n    \"gh\" 'outline-up-heading\n    \"gj\" 'outline-forward-same-level\n    \"gk\" 'outline-backward-same-level\n    \"gl\" 'outline-next-visible-heading\n    \"gu\" 'outline-previous-visible-heading))\n<\/pre>\n<h2 id=\"styling\">Styling<\/h2>\n<h3 id=\"faces\">Faces<\/h3>\n<p>Outline headers fulfill a different goal than org headers. Outlines are for\nstructuring semantically similar blocks of code. Org headers additionally\nincorporate todos, priorities, tags, and so on.<\/p>\n<p>Outlines are necessarily further apart than org headers, which often have no\nsubtext at all.<\/p>\n<p>So I recommend setting both the <code>:background<\/code> and <code>:height<\/code> face attributes to\nhighlight and make obvious the different sections.<\/p>\n<p>This can be done by updating the <code>outline-1\/2\/3...<\/code> faces.\nOrg mode header faces can set separately using the <code>org-level-1\/2\/3...<\/code> faces.<\/p>\n<pre class=\"language-lisp\">(setq my-black \"#1b1b1e\")\n(custom-theme-set-faces\n 'solarized-dark\n `(outline-1 ((t (:height 1.25 :background \"#268bd2\"\n                          :foreground ,my-black :weight bold))))\n `(outline-2 ((t (:height 1.15 :background \"#2aa198\"\n                          :foreground ,my-black :weight bold))))\n `(outline-3 ((t (:height 1.05 :background \"#b58900\"\n                          :foreground ,my-black :weight bold)))))\n<\/pre>\n<h3 id=\"bullets\">Bullets<\/h3>\n<p>Now these outlines will be displayed with the comment syntax, there is no mirror\nof <code>org-bullets-bullet-list<\/code> for setting icons to replace the eg. ';;;'.<\/p>\n<p>We need to use <code>compose-region<\/code> to manually replace the headers with our custom\nbullets. Here is an image of my org-bullets replicated for outlines.\n<img src=\"http:\/\/www.modernemacs.com\/img\/outline-bullets.png\" alt=\"\/img\/outline-bullets.png\" title=\"\/img\/outline-bullets.png\" \/><\/p>\n<p>The following code is generalized to replace symbols for possibly many modes and\nis a bit complex. Here I implement the outline bullets for lisp-like modes and\npython.<\/p>\n<p>UPDATE: Check <a href=\"http:\/\/www.modernemacs.com\/post\/outline-bullets\/\">this post for full implementation.<\/a><\/p>\n<pre class=\"language-lisp\">(defun -add-font-lock-kwds (FONT-LOCK-ALIST)\n  (font-lock-add-keywords\n   nil (--map (-let (((rgx uni-point) it))\n                `(,rgx (0 (progn\n                            (compose-region (match-beginning 1) (match-end 1)\n                                            ,(concat \"\\t\" (list uni-point)))\n                            nil))))\n              FONT-LOCK-ALIST)))\n(defmacro add-font-locks (FONT-LOCK-HOOKS-ALIST)\n  `(--each ,FONT-LOCK-HOOKS-ALIST\n     (-let (((font-locks . mode-hooks) it))\n       (--each mode-hooks\n         (add-hook it (-partial '-add-font-lock-kwds\n                                (symbol-value font-locks)))))))\n(defconst emacs-outlines-font-lock-alist\n  ;; Outlines\n  '((\"\\\\(^;;;\\\\) \"          ?\u25a0)\n    (\"\\\\(^;;;;\\\\) \"         ?\u25cb)\n    (\"\\\\(^;;;;;\\\\) \"        ?\u2738)\n    (\"\\\\(^;;;;;;\\\\) \"       ?\u273f)))\n(defconst lisp-outlines-font-lock-alist\n  ;; Outlines\n  '((\"\\\\(^;; \\\\*\\\\) \"          ?\u25a0)\n    (\"\\\\(^;; \\\\*\\\\*\\\\) \"       ?\u25cb)\n    (\"\\\\(^;; \\\\*\\\\*\\\\*\\\\) \"    ?\u2738)\n    (\"\\\\(^;; \\\\*\\\\*\\\\*\\\\*\\\\) \" ?\u273f)))\n(defconst python-outlines-font-lock-alist\n  '((\"\\\\(^# \\\\*\\\\) \"          ?\u25a0)\n    (\"\\\\(^# \\\\*\\\\*\\\\) \"       ?\u25cb)\n    (\"\\\\(^# \\\\*\\\\*\\\\*\\\\) \"    ?\u2738)\n    (\"\\\\(^# \\\\*\\\\*\\\\*\\\\*\\\\) \" ?\u273f)))\n(add-font-locks\n '((emacs-outlines-font-lock-alist emacs-lisp-mode-hook)\n   (lisp-outlines-font-lock-alist clojure-mode-hook hy-mode-hook)\n   (python-outlines-font-lock-alist python-mode-hook)))\n<\/pre>\n<p>This could eventually be improved to work for any number of levels and\nauto-determine the outline regex base on the comment syntax.<\/p>\n<h3 id=\"ellipsis\">Ellipsis<\/h3>\n<p>Org-mode has the variable <code>org-ellipsis<\/code> for setting the trailing chars for\ncollapsed headers.<\/p>\n<p>We can set our own outline ellipsis icon as follows:<\/p>\n<pre class=\"language-lisp\">(defvar outline-display-table (make-display-table))\n(set-display-table-slot outline-display-table 'selective-display\n                        (vector (make-glyph-code ?\u25bc 'escape-glyph)))\n(defun set-outline-display-table ()\n  (setf buffer-display-table outline-display-table))\n(add-hook 'outline-mode-hook 'set-outline-display-table)\n(add-hook 'outline-minor-mode-hook 'set-outline-display-table)\n<\/pre>\n<table><thead><tr><th>All the outlines<\/th>\n<th>Searching catches children<\/th>\n<th>Restricting Levels<\/th>\n<\/tr><\/thead><tbody><tr><td><img src=\"http:\/\/www.modernemacs.com\/img\/outline-ivy-raw.png\" alt=\"\/img\/outline-ivy-raw.png\" title=\"\/img\/outline-ivy-raw.png\" \/><\/td>\n<td><img src=\"http:\/\/www.modernemacs.com\/img\/outline-ivy-many.png\" alt=\"\/img\/outline-ivy-many.png\" title=\"\/img\/outline-ivy-many.png\" \/><\/td>\n<td><img src=\"http:\/\/www.modernemacs.com\/img\/outline-ivy-level.png\" alt=\"\/img\/outline-ivy-level.png\" title=\"\/img\/outline-ivy-level.png\" \/><\/td>\n<\/tr><\/tbody><\/table><p>Current methods for jumping to outlines have significant limitations.<\/p>\n<p><a href=\"https:\/\/github.com\/ekaschalk\/.spacemacs.d\/blob\/master\/layers\/personal\/local\/outline-ivy\/outline-ivy.el\" title=\"Outline-ivy\">Outline-ivy<\/a> makes outlines a proper navigational, not just organizational, tool\nwith <code>oi-jump<\/code>.<\/p>\n<p>All outlines are collected, stylized, and associated with their parents.<\/p>\n<p>Parents are inserted as invisible text for child outlines. This way, searching\nfor eg. \"Display\" catches all its children.<\/p>\n<p>The level is also inserted and hidden enabling the search \"1 spacemacs\" to catch\nonly top-level headings matching spacemacs. It also overrides <code>ivy-match-face<\/code> and\n <code>ivy-height<\/code> to play nice with the propertized prompt strings.<\/p>\n<p>The default configuration:<\/p>\n<pre class=\"language-lisp\">(defvar oi-height 20\n  \"Number of outlines to display, overrides ivy-height.\")\n(defface oi-match-face\n  '((t :height 1.10 :foreground \"light gray\"))\n  \"Match face for ivy outline prompt.\")\n(defface oi-face-1\n  '((t :foreground \"#268bd2\" :height 1.25 :underline t :weight ultra-bold))\n  \"Ivy outline face for level 1\")\n(defface oi-face-2\n  '((t :foreground \"#2aa198\" :height 1.1 :weight semi-bold))\n  \"Ivy outline face for level 2\")\n(defface oi-face-3\n  '((t :foreground \"steel blue\"))\n  \"Ivy outline face for level 3\")\n;; My keybinding for oi-jump, unbound by default.\n(global-set-key (kbd \"C-j\") 'oi-jump)\n<\/pre>\n<p>Many emacs lisp packages both intentionally and unintentionally use the outline\nsyntax for organizing their source. These changes can make understanding and\nnavigating the source significantly easier.<\/p>\n<p>Below is a screenshot jumping to an outline in <code>org.el<\/code>, the core org source\nfile.<\/p>\n<p><img src=\"http:\/\/www.modernemacs.com\/img\/outlines-org.png\" alt=\"\/img\/outlines-org.png\" title=\"\/img\/outlines-org.png\" \/><\/p>\n<p>I'm not sure how difficult it would be to port this package to <code>helm<\/code>. This is\none of many updates I'll look to add eventually, such as adding more quick\nactions to the prompt like jump-and-narrow or a projective jump version.<\/p>\n<p>Perhaps the best judge of the impact of some configuration is how often you\nfind yourself reaching for it. For buffer-wide navigation to a specific area\nrather than some specific symbol or function, I now almost exclusively use\n <code>oi-jump<\/code>.<\/p>\n<p>I consider outlines to be one of the most practical features of Emacs. All\nsource code, in any language, I organize with outlines. Given how unobtrusive\nthe syntax is, it shouldn't be difficult to implement in a collaborative\nproject.<\/p>","created_at":"2019-08-09T12:19:56+0000","updated_at":"2019-08-09T12:19:59+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"www.modernemacs.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["databases","log","data"],"is_public":false,"id":992,"title":"The Log: What every software engineer should know about real-time data's unifying abstraction","url":"https:\/\/engineering.linkedin.com\/distributed-systems\/log-what-every-software-engineer-should-know-about-real-time-datas-unifying","content":"<p> I joined LinkedIn about six years ago at a particularly interesting time. We were just beginning to run up against the limits of our monolithic, centralized database and needed to start the transition to a portfolio of specialized distributed systems. This has been an interesting experience: we built, deployed, and run to this day a distributed graph database, a distributed search backend, a Hadoop installation, and a first and second generation key-value store. <\/p><p> One of the most useful things I learned in all this was that many of the things we were building had a very simple concept at their heart: the log. Sometimes called write-ahead logs or commit logs or transaction logs, logs have been around almost as long as computers and are at the heart of many distributed data systems and real-time application architectures. <\/p><p> You can't fully understand databases, NoSQL stores, key value stores, replication, paxos, hadoop, version control, or almost any software system without understanding logs; and yet, most software engineers are not familiar with them. I'd like to change that. In this post, I'll walk you through everything you need to know about logs, including what is log and how to use logs for data integration, real time processing, and system building. <\/p><h2>Part One: What Is a Log?<\/h2>A log is perhaps the simplest possible storage abstraction. It is an append-only, totally-ordered sequence of records ordered by time. It looks like this:<div><img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/log.png\" alt=\"image\" \/><\/div><p> Records are appended to the end of the log, and reads proceed left-to-right. Each entry is assigned a unique sequential log entry number. <\/p><p> The ordering of records defines a notion of \"time\" since entries to the left are defined to be older then entries to the right. The log entry number can be thought of as the \"timestamp\" of the entry. Describing this ordering as a notion of time seems a bit odd at first, but it has the convenient property that it is decoupled from any particular physical clock. This property will turn out to be essential as we get to distributed systems. <\/p><p> The contents and format of the records aren't important for the purposes of this discussion. Also, we can't just keep adding records to the log as we'll eventually run out of space. I'll come back to this in a bit. <\/p><p> So, a log is not all that different from a file or a table. A file is an array of bytes, a table is an array of records, and a log is really just a kind of table or file where the records are sorted by time. <\/p><p> At this point you might be wondering why it is worth talking about something so simple? How is a append-only sequence of records in any way related to data systems? The answer is that logs have a specific purpose: they record what happened and when. For distributed data systems this is, in many ways, the very heart of the problem. <\/p><p> But before we get too far let me clarify something that is a bit confusing. Every programmer is familiar with another definition of logging\u2014the unstructured error messages or trace info an application might write out to a local file using syslog or log4j. For clarity I will call this \"application logging\". The application log is a degenerative form of the log concept I am describing. The biggest difference is that text logs are meant to be primarily for humans to read and the \"journal\" or \"data logs\" I'm describing are built for programmatic access. <\/p><p> (Actually, if you think about it, the idea of humans reading through logs on individual machines is something of an anachronism. This approach quickly becomes an unmanageable strategy when many services and servers are involved and the purpose of logs quickly becomes as an input to queries and graphs to understand behavior across many machines\u2014something for which english text in files is not nearly as appropriate as the kind structured log described here.) <\/p><h3>Logs in databases<\/h3><p> I don't know where the log concept originated\u2014probably it is one of those things like binary search that is too simple for the inventor to realize it was an invention. It is present as early as IBM's <a href=\"http:\/\/www.cs.berkeley.edu\/~brewer\/cs262\/SystemR.pdf\" target=\"_blank\">System R<\/a>. The usage in databases has to do with keeping in sync the variety of data structures and indexes in the presence of crashes. To make this atomic and durable, a database uses a log to write out information about the records they will be modifying, before applying the changes to all the various data structures it maintains. The log is the record of what happened, and each table or index is a projection of this history into some useful data structure or index. Since the log is immediately persisted it is used as the authoritative source in restoring all other persistent structures in the event of a crash. <\/p><p> Over-time the usage of the log grew from an implementation detail of ACID to a method for replicating data between databases. It turns out that the sequence of changes that happened on the database is exactly what is needed to keep a remote replica database in sync. Oracle, MySQL, and PostgreSQL include log shipping protocols to transmit portions of log to replica databases which act as slaves. Oracle has productized the log as a general data subscription mechanism for non-oracle data subscribers with their <a href=\"http:\/\/docs.oracle.com\/cd\/E11882_01\/server.112\/e16545\/xstrm_intro.htm\" target=\"_blank\">XStreams<\/a> and <a href=\"http:\/\/www.oracle.com\/technetwork\/middleware\/goldengate\/overview\/index.html\" target=\"_blank\">GoldenGate<\/a> and similar facilities in MySQL and PostgreSQL are key components of many data architectures. <\/p><p> Because of this origin, the concept of a machine readable log has largely been confined to database internals. The use of logs as a mechanism for data subscription seems to have arisen almost by chance. But this very abstraction is ideal for supporting all kinds of messaging, data flow, and real-time data processing. <\/p><h3>Logs in distributed systems<\/h3><p> The two problems a log solves\u2014ordering changes and distributing data\u2014are even more important in distributed data systems. Agreeing upon an ordering for updates (or agreeing to disagree and coping with the side-effects) are among the core design problems for these systems. <\/p><p> The log-centric approach to distributed systems arises from a simple observation that I will call the State Machine Replication Principle: <\/p><p>If two identical, deterministic processes begin in the same state and get the same inputs in the same order, they will produce the same output and end in the same state.<\/p><p> This may seem a bit obtuse, so let's dive in and understand what it means. <\/p><p> <a href=\"http:\/\/en.wikipedia.org\/wiki\/Deterministic_algorithm\" target=\"_blank\">Deterministic<\/a> means that the processing isn't timing dependent and doesn't let any other \"out of band\" input influence its results. For example a program whose output is influenced by the particular order of execution of threads or by a call to <code>gettimeofday<\/code> or some other non-repeatable thing is generally best considered as non-deterministic. <\/p><p> The <i>state<\/i> of the process is whatever data remains on the machine, either in memory or on disk, at the end of the processing. <\/p><p> The bit about getting the same input in the same order should ring a bell\u2014that is where the log comes in. This is a very intuitive notion: if you feed two deterministic pieces of code the same input log, they will produce the same output. <\/p><p> The application to distributed computing is pretty obvious. You can reduce the problem of making multiple machines all do the same thing to the problem of implementing a distributed consistent log to feed these processes input. The purpose of the log here is to squeeze all the non-determinism out of the input stream to ensure that each replica processing this input stays in sync. <\/p><p> When you understand it, there is nothing complicated or deep about this principle: it more or less amounts to saying \"deterministic processing is deterministic\". Nonetheless, I think it is one of the more general tools for distributed systems design. <\/p><p> One of the beautiful things about this approach is that the time stamps that index the log now act as the clock for the state of the replicas\u2014you can describe each replica by a single number, the timestamp for the maximum log entry it has processed. This timestamp combined with the log uniquely captures the entire state of the replica. <\/p><p> There are a multitude of ways of applying this principle in systems depending on what is put in the log. For example, we can log the incoming requests to a service, or the state changes the service undergoes in response to request, or the transformation commands it executes. Theoretically, we could even log a series of machine instructions for each replica to execute or the method name and arguments to invoke on each replica. As long as two processes process these inputs in the same way, the processes will remaining consistent across replicas. <\/p><p> Different groups of people seem to describe the uses of logs differently. Database people generally differentiate between <i>physical<\/i> and <i>logical<\/i> logging. Physical logging means logging the contents of each row that is changed. Logical logging means logging not the changed rows but the SQL commands that lead to the row changes (the insert, update, and delete statements). <\/p><p> The distributed systems literature commonly distinguishes two broad approaches to processing and replication. The \"state machine model\" usually refers to an active-active model where we keep a log of the incoming requests and each replica processes each request. A slight modification of this, called the \"primary-backup model\", is to elect one replica as the leader and allow this leader to process requests in the order they arrive and log out the changes to its state from processing the requests. The other replicas apply in order the state changes the leader makes so that they will be in sync and ready to take over as leader should the leader fail. <\/p><div><img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/active_and_passive_arch.png\" alt=\"image\" \/><\/div><p> To understand the difference between these two approaches, let's look at a toy problem. Consider a replicated \"arithmetic service\" which maintains a single number as its state (initialized to zero) and applies additions and multiplications to this value. The active-active approach might log out the transformations to apply, say \"+1\", \"*2\", etc. Each replica would apply these transformations and hence go through the same set of values. The \"active-passive\" approach would have a single master execute the transformations and log out the <i>result<\/i>, say \"1\", \"3\", \"6\", etc. This example also makes it clear why ordering is key for ensuring consistency between replicas: reordering an addition and multiplication will yield a different result. <\/p><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/paxos_postcard.jpg\" alt=\"image\" \/> The distributed log can be seen as the data structure which models the problem of consensus. A log, after all, represents a series of decisions on the \"next\" value to append. You have to squint a little to see a log in the Paxos family of algorithms, though log-building is their most common practical application. With Paxos, this is usually done using an extension of the protocol called \"multi-paxos\", which models the log as a series of consensus problems, one for each slot in the log. The log is much more prominent in other protocols such as <a href=\"http:\/\/www.stanford.edu\/class\/cs347\/reading\/zab.pdf\" target=\"_blank\">ZAB<\/a><a>, <\/a><a href=\"https:\/\/ramcloud.stanford.edu\/wiki\/download\/attachments\/11370504\/raft.pdf\">RAFT<\/a>, and <a href=\"http:\/\/pmg.csail.mit.edu\/papers\/vr-revisited.pdf\" target=\"_blank\">Viewstamped Replication<\/a>, which directly model the problem of maintaining a distributed, consistent log. <\/p><p> My suspicion is that our view of this is a little bit biased by the path of history, perhaps due to the few decades in which the theory of distributed computing outpaced its practical application. In reality, the consensus problem is a bit too simple. Computer systems rarely need to decide a single value, they almost always handle a sequence of requests. So a log, rather than a simple single-value register, is the more natural abstraction. <\/p><p> Furthermore, the focus on the algorithms obscures the underlying log abstraction systems need. I suspect we will end up focusing more on the log as a commoditized building block irrespective of its implementation in the same way we often talk about a hash table without bothering to get in the details of whether we mean the murmur hash with linear probing or some other variant. The log will become something of a commoditized interface, with many algorithms and implementations competing to provide the best guarantees and optimal performance. <\/p><h3>Changelog 101: Tables and Events are Dual<\/h3><p> Let's come back to databases for a bit. There is a facinating duality between a log of changes and a table. The log is similar to the list of all credits and debits and bank processes; a table is all the current account balances. If you have a log of changes, you can apply these changes in order to create the table capturing the current state. This table will record the latest state for each key (as of a particular log time). There is a sense in which the log is the more fundamental data structure: in addition to creating the original table you can also transform it to create all kinds of derived tables. (And yes, table can mean keyed data store for the non-relational folks.) <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/yin-yang.jpg\" alt=\"image\" \/><\/p><p> This process works in reverse too: if you have a table taking updates, you can record these changes and publish a \"changelog\" of all the updates to the state of the table. This changelog is exactly what you need to support near-real-time replicas. So in this sense you can see tables and events as dual: tables support data at rest and logs capture change. The magic of the log is that if it is a <i>complete<\/i> log of changes, it holds not only the contents of the final version of the table, but also allows recreating all other versions that might have existed. It is, effectively, a sort of backup of <i>every<\/i> previous state of the table. <\/p><p> This might remind you of source code version control. There is a close relationship between source control and databases. Version control solves a very similar problem to what distributed data systems have to solve\u2014managing distributed, concurrent changes in state. A version control system usually models the sequence of patches, which is in effect a log. You interact directly with a checked out \"snapshot\" of the current code which is analogous to the table. You will note that in version control systems, as in other distributed stateful systems, replication happens via the log: when you update, you pull down just the patches and apply them to your current snapshot. <\/p><p> Some people have seen some of these ideas recently from <a href=\"http:\/\/www.datomic.com\/\" target=\"_blank\">Datomic<\/a>, a company selling a log-centric database. This <a href=\"https:\/\/www.youtube.com\/watch?v=Cym4TZwTCNU\">presentation<\/a> gives a great overview of how they have applied the idea in their system. These ideas are not unique to this system, of course, as they have been a part of the distributed systems and database literature for well over a decade. <\/p><p> This may all seem a little theoretical. Do not despair! We'll get to practical stuff pretty quickly. <\/p><h3>What's next<\/h3><p> In the remainder of this article I will try to give a flavor of what a log is good for that goes beyond the internals of distributed computing or abstract distributed computing models. This includes: <\/p><ol><li><i>Data Integration<\/i>\u2014Making all of an organization's data easily available in all its storage and processing systems. <\/li><li><i>Real-time data processing<\/i>\u2014Computing derived data streams. <\/li><li><i>Distributed system design<\/i>\u2014How practical systems can by simplified with a log-centric design. <\/li><\/ol><p> These uses all resolve around the idea of a log as a stand-alone service. <\/p><p> In each case, the usefulness of the log comes from simple function that the log provides: producing a persistent, re-playable record of history. Surprisingly, at the core of these problems is the ability to have many machines playback history at their own rate in a deterministic manner. <\/p><h2>Part Two: Data Integration<\/h2><p> Let me first say what I mean by \"data integration\" and why I think it's important, then we'll see how it relates back to logs. <\/p><p>Data integration is making all the data an organization has available in all its services and systems.<\/p><p> This phrase \"data integration\" isn't all that common, but I don't know a better one. The more recognizable term <a href=\"http:\/\/en.wikipedia.org\/wiki\/Extract,_transform,_load\" target=\"_blank\">ETL<\/a> usually covers only a limited part of data integration\u2014populating a relational data warehouse. But much of what I am describing can be thought of as ETL generalized to cover real-time systems and processing flows. <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/cabling.jpg\" alt=\"image\" \/><\/p><p> You don't hear much about data integration in all the breathless interest and hype around the idea of <i>big data<\/i>, but nonetheless, I believe this mundane problem of \"making the data available\" is one of the more valuable things an organization can focus on. <\/p><p> Effective use of data follows a kind of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Maslow's_hierarchy_of_needs\" target=\"_blank\">Maslow's hierarchy of needs<\/a>. The base of the pyramid involves capturing all the relevant data, being able to put it together in an applicable processing environment (be that a fancy real-time query system or just text files and python scripts). This data needs to be modeled in a uniform way to make it easy to read and process. Once these basic needs of capturing data in a uniform way are taken care of it is reasonable to work on infrastructure to process this data in various ways\u2014MapReduce, real-time query systems, etc. <\/p><p> It's worth noting the obvious: without a reliable and complete data flow, a Hadoop cluster is little more than a very expensive and difficult to assemble space heater. Once data and processing are available, one can move concern on to more refined problems of good data models and consistent well understood semantics. Finally, concentration can shift to more sophisticated processing\u2014better visualization, reporting, and algorithmic processing and prediction. <\/p><p> In my experience, most organizations have huge holes in the base of this pyramid\u2014they lack reliable complete data flow\u2014but want to jump directly to advanced data modeling techniques. This is completely backwards. <\/p><p> So the question is, how can we build reliable data flow throughout all the data systems in an organization? <\/p><h3>Data Integration: Two complications<\/h3><p> Two trends make data integration harder. <\/p><p> <b>The event data firehose<\/b> <\/p><p> The first trend is the rise of event data. Event data records things that happen rather than things that are. In web systems, this means user activity logging, but also the machine-level events and statistics required to reliably operate and monitor a data center's worth of machines. People tend to call this \"log data\" since it is often written to application logs, but that confuses form with function. This data is at the heart of the modern web: Google's fortune, after all, is generated by a relevance pipeline built on clicks and impressions\u2014that is, events. <\/p><p> And this stuff isn't limited to web companies, it's just that web companies are already fully digital, so they are easier to instrument. Financial data has long been event-centric. <a href=\"http:\/\/en.wikipedia.org\/wiki\/RFID\" target=\"_blank\">RFID<\/a> adds this kind of tracking to physical objects. I think this trend will continue with the <a href=\"http:\/\/online.wsj.com\/article\/SB10001424053111903480904576512250915629460.html\" target=\"_blank\">digitization<\/a> of traditional businesses and activities. <\/p><p> This type of event data records what happened, and tends to be several orders of magnitude larger than traditional database uses. This presents significant challenges for processing. <\/p><p> <b>The explosion of specialized data systems<\/b> <\/p><p> The second trend comes from the <a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/summary?doi=10.1.1.68.9136\" target=\"_blank\">explosion<\/a> of specialized data systems that have become popular and often freely available in the last five years. Specialized systems exist for <a href=\"https:\/\/github.com\/metamx\/druid\/wiki\">OLAP<\/a>, <a href=\"http:\/\/www.elasticsearch.org\/\" target=\"_blank\">search<\/a>, <a href=\"http:\/\/www.rethinkdb.com\/\" target=\"_blank\">simple<\/a> <a href=\"http:\/\/www.slideshare.net\/amywtang\/espresso-20952131\" target=\"_blank\">online<\/a> <a href=\"http:\/\/cassandra.apache.org\/\" target=\"_blank\">storage<\/a>, <a href=\"http:\/\/hadoop.apache.org\/\" target=\"_blank\">batch processing<\/a>, <a href=\"http:\/\/graphlab.org\/\" target=\"_blank\">graph analysis<\/a>, and <a href=\"http:\/\/redis.io\/\" target=\"_blank\">so<\/a> <a href=\"http:\/\/spark.incubator.apache.org\/\" target=\"_blank\">on<\/a>. <\/p><p> The combination of more data of more varieties and a desire to get this data into more systems leads to a huge data integration problem. <\/p><h3>Log-structured data flow<\/h3>The log is the natural data structure for handling data flow between systems. The recipe is very simple:<p>Take all the organization's data and put it into a central log for real-time subscription.<\/p><p> Each logical data source can be modeled as its own log. A data source could be an application that logs out events (say clicks or page views), or a database table that accepts modifications. Each subscribing system reads from this log as quickly as it can, applies each new record to its own store, and advances its position in the log. Subscribers could be any kind of data system\u2014a cache, Hadoop, another database in another site, a search system, etc. <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/log_subscription.png\" alt=\"image\" \/><\/p><p> For example, the log concept gives a logical clock for each change against which all subscribers can be measured. This makes reasoning about the state of the different subscriber systems with respect to one another far simpler, as each has a \"point in time\" they have read up to. <\/p><p> To make this more concrete, consider a simple case where there is a database and a collection of caching servers. The log provides a way to synchronize the updates to all these systems and reason about the point of time of each of these systems. Let's say we write a record with log entry X and then need to do a read from the cache. If we want to guarantee we don't see stale data, we just need to ensure we don't read from any cache which has not replicated up to X. <\/p><p> The log also acts as a buffer that makes data production asynchronous from data consumption. This is important for a lot of reasons, but particularly when there are multiple subscribers that may consume at different rates. This means a subscribing system can crash or go down for maintenance and catch up when it comes back: the subscriber consumes at a pace it controls. A batch system such as Hadoop or a data warehouse may consume only hourly or daily, whereas a real-time query system may need to be up-to-the-second. Neither the originating data source nor the log has knowledge of the various data destination systems, so consumer systems can be added and removed with no change in the pipeline. <\/p><div><a href=\"http:\/\/en.wikipedia.org\/wiki\/Anna_Karenina_principle\" target=\"_blank\"> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/tolstoy.jpg\" alt=\"image\" \/><\/a><p><a href=\"http:\/\/en.wikipedia.org\/wiki\/Anna_Karenina_principle\" target=\"_blank\">\"Each working data pipeline is designed like a log; each broken data pipeline is broken in its own way.\"\u2014Count Leo Tolstoy (translation by the author)<\/a><\/p><\/div><p> Of particular importance: the destination system only knows about the log and not any details of the system of origin. The consumer system need not concern itself with whether the data came from an RDBMS, a new-fangled key-value store, or was generated without a real-time query system of any kind. This seems like a minor point, but is in fact critical. <\/p><p> I use the term \"log\" here instead of \"messaging system\" or \"pub sub\" because it is a lot more specific about semantics and a much closer description of what you need in a practical implementation to support data replication. I have found that \"publish subscribe\" doesn't imply much more than indirect addressing of messages\u2014if you compare any two messaging systems promising publish-subscribe, you find that they guarantee very different things, and most models are not useful in this domain. You can think of the log as acting as a kind of messaging system with durability guarantees and strong ordering semantics. In distributed systems, this model of communication sometimes goes by the (somewhat terrible) name of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Atomic_broadcast\" target=\"_blank\">atomic broadcast<\/a>. <\/p><p> It's worth emphasizing that the log is still just the infrastructure. That isn't the end of the story of mastering data flow: the rest of the story is around metadata, schemas, compatibility, and all the details of handling data structure and evolution. But until there is a reliable, general way of handling the mechanics of data flow, the semantic details are secondary. <\/p><h3>At LinkedIn<\/h3>I got to watch this data integration problem emerge in fast-forward as LinkedIn moved from a centralized relational database to a collection of distributed systems.<p> These days our major data systems include: <\/p><ul><li> <a href=\"http:\/\/data.linkedin.com\/projects\/search\" target=\"_blank\">Search<\/a> <\/li><li> <a href=\"http:\/\/engineering.linkedin.com\/real-time-distributed-graph\/using-set-cover-algorithm-optimize-query-latency-large-scale-distributed\">Social Graph<\/a> <\/li><li> <a href=\"http:\/\/project-voldemort.com\/\" target=\"_blank\">Voldemort<\/a> (key-value store) <\/li><li> <a href=\"http:\/\/data.linkedin.com\/projects\/espresso\" target=\"_blank\">Espresso<\/a> (document store) <\/li><li> <a href=\"http:\/\/www.quora.com\/LinkedIn-Recommendations\/How-does-LinkedIns-recommendation-system-work\" target=\"_blank\">Recommendation engine<\/a> <\/li><li> OLAP query engine <\/li><li> <a href=\"http:\/\/hadoop.apache.org\/\" target=\"_blank\">Hadoop<\/a> <\/li><li> <a href=\"http:\/\/www.teradata.com\/\" target=\"_blank\">Terradata<\/a> <\/li><li> <a href=\"http:\/\/engineering.linkedin.com\/52\/autometrics-self-service-metrics-collection\">Ingraphs<\/a> (monitoring graphs and metrics services) <\/li><\/ul><p> Each of these is a specialized distributed system that provides advanced functionality in its area of specialty. <\/p><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/linkedin.png\" alt=\"image\" \/><\/p><p> This idea of using logs for data flow has been floating around LinkedIn since even before I got here. One of the earliest pieces of infrastructure we developed was a service called <a href=\"https:\/\/github.com\/linkedin\/databus\">databus<\/a> that provided a log caching abstraction on top of our early Oracle tables to scale subscription to database changes so we could feed our social graph and search indexes. <\/p><p> I'll give a little bit of the history to provide context. My own involvement in this started around 2008 after we had shipped our key-value store. My next project was to try to get a working Hadoop setup going, and move some of our recommendation processes there. Having little experience in this area, we naturally budgeted a few weeks for getting data in and out, and the rest of our time for implementing fancy prediction algorithms. So began a long slog. <\/p><p> We originally planned to just scrape the data out of our existing Oracle data warehouse. The first discovery was that getting data out of Oracle quickly is something of a dark art. Worse, the data warehouse processing was not appropriate for the production batch processing we planned for Hadoop\u2014much of the processing was non-reversable and specific to the reporting being done. We ended up avoiding the data warehouse and going directly to source databases and log files. Finally, we implemented another pipeline to <a href=\"http:\/\/data.linkedin.com\/blog\/2009\/06\/building-a-terabyte-scale-data-cycle-at-linkedin-with-hadoop-and-project-voldemort\" target=\"_blank\">load data into our key-value store<\/a> for serving results. <\/p><p> This mundane data copying ended up being one of the dominate items for the original development. Worse, any time there was a problem in any of the pipelines, the Hadoop system was largely useless\u2014running fancy algorithms on bad data just produces more bad data. <\/p><p> Although we had built things in a fairly generic way, each new data source required custom configuration to set up. It also proved to be the source of a huge number of errors and failures. The site features we had implemented on Hadoop became popular and we found ourselves with a long list of interested engineers. Each user had a list of systems they wanted integration with and a long list of new data feeds they wanted. <\/p><div><a href=\"http:\/\/en.wikipedia.org\/wiki\/Sisyphus\" target=\"_blank\"> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/sisyphus.jpg\" alt=\"image\" \/><\/a><p>ETL in Ancient Greece. Not much has changed.<\/p><\/div><p> A few things slowly became clear to me. <\/p><p> First, the pipelines we had built, though a bit of a mess, were actually extremely valuable. Just the process of making data available in a new processing system (Hadoop) unlocked a lot of possibilities. New computation was possible on the data that would have been hard to do before. Many new products and analysis just came from putting together multiple pieces of data that had previously been locked up in specialized systems. <\/p><p> Second, it was clear that reliable data loads would require deep support from the data pipeline. If we captured all the structure we needed, we could make Hadoop data loads fully automatic, so that no manual effort was expanded adding new data sources or handling schema changes\u2014data would just magically appear in HDFS and Hive tables would automatically be generated for new data sources with the appropriate columns. <\/p><p> Third, we still had very low data coverage. That is, if you looked at the overall percentage of the data LinkedIn had that was available in Hadoop, it was still very incomplete. And getting to completion was not going to be easy given the amount of effort required to operationalize each new data source. <\/p><p> The way we had been proceeding, building out custom data loads for each data source and destination, was clearly infeasible. We had dozens of data systems and data repositories. Connecting all of these would have lead to building custom piping between each pair of systems something like this: <\/p><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/datapipeline_complex.png\" alt=\"image\" \/><\/p><p> Note that data often flows in both directions, as many systems (databases, Hadoop) are both sources and destinations for data transfer. This meant we would end up building two pipelines per system: one to get data in and one to get data out. <\/p><p> This clearly would take an army of people to build and would never be operable. As we approached fully connectivity we would end up with something like O(N<sup>2<\/sup>) pipelines. <\/p><p> Instead, we needed something generic like this: <\/p><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/datapipeline_simple.png\" alt=\"image\" \/><\/p><p> As much as possible, we needed to isolate each consumer from the source of the data. They should ideally integrate with just a single data repository that would give them access to everything. <\/p><p> The idea is that adding a new data system\u2014be it a data source or a data destination\u2014should create integration work only to connect it to a single pipeline instead of each consumer of data. <\/p><p> This experience lead me to focus on building <a href=\"http:\/\/kafka.apache.org\/\" target=\"_blank\">Kafka<\/a> to combine what we had seen in messaging systems with the log concept popular in databases and distributed system internals. We wanted something to act as a central pipeline first for all activity data, and eventually for many other uses, including data deployment out of Hadoop, monitoring data, etc. <\/p><p> For a long time, Kafka was a little unique (some would say odd) as an infrastructure product\u2014neither a database nor a log file collection system nor a traditional messaging system. But recently Amazon has offered a service that is very very similar to Kafka called <a href=\"http:\/\/aws.amazon.com\/kinesis\" target=\"_blank\">Kinesis<\/a>. The similarity goes right down to the way partitioning is handled, data is retained, and the fairly odd split in the Kafka API between high- and low-level consumers. I was pretty happy about this. A sign you've created a good infrastructure abstraction is that AWS offers it as a service! Their vision for this seems to be exactly similar to what I am describing: it is the piping that connects all their distributed systems\u2014DynamoDB, RedShift, S3, etc.\u2014as well as the basis for distributed stream processing using EC2. <\/p><h3>Relationship to ETL and the Data Warehouse<\/h3><p> Let's talk data warehousing for a bit. The data warehouse is meant to be a repository of the clean, integrated data structured to support analysis. This is a great idea. For those not in the know, the data warehousing methodology involves periodically extracting data from source databases, munging it into some kind of understandable form, and loading it into a central data warehouse. Having this central location that contains a clean copy of all your data is a hugely valuable asset for data-intensive analysis and processing. At a high level, this methodology doesn't change too much whether you use a traditional data warehouse like Oracle or Teradata or Hadoop, though you might <a href=\"http:\/\/searchdatamanagement.techtarget.com\/definition\/Extract-Load-Transform-ELT\" target=\"_blank\">switch up<\/a> the order of loading and munging. <\/p><p> A data warehouse containing clean, integrated data is a phenomenal asset, but the mechanics of getting this are a bit out of date. <\/p><img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/oracle.jpg\" alt=\"image\" \/><p> The key problem for a data-centric organization is coupling the clean integrated data to the data warehouse. A data warehouse is a piece of batch query infrastructure which is well suited to many kinds of reporting and ad hoc analysis, particularly when the queries involve simple counting, aggregation, and filtering. But having a batch system be the only repository of clean complete data means the data is unavailable for systems requiring a real-time feed\u2014real-time processing, search indexing, monitoring systems, etc. <\/p><p> In my view, ETL is really two things. First, it is an extraction and data cleanup process\u2014essentially liberating data locked up in a variety of systems in the organization and removing an system-specific non-sense. Secondly, that data is restructured for data warehousing queries (i.e. made to fit the type system of a relational DB, forced into a star or snowflake schema, perhaps broken up into a high performance <a href=\"http:\/\/parquet.io\/\" target=\"_blank\">column<\/a> <a href=\"http:\/\/docs.hortonworks.com\/HDPDocuments\/HDP2\/HDP-2.0.0.2\/ds_Hive\/orcfile.html\" target=\"_blank\">format<\/a>, etc). Conflating these two things is a problem. The clean, integrated repository of data should be available in real-time as well for low-latency processing as well as indexing in other real-time storage systems. <\/p><p> I think this has the added benefit of making data warehousing ETL much more <i>organizationally<\/i> scalable. The classic problem of the data warehouse team is that they are responsible for collecting and cleaning all the data generated by every other team in the organization. The incentives are not aligned: data producers are often not very aware of the use of the data in the data warehouse and end up creating data that is hard to extract or requires heavy, hard to scale transformation to get into usable form. Of course, the central team never quite manages to scale to match the pace of the rest of the organization, so data coverage is always spotty, data flow is fragile, and changes are slow. <\/p><p> A better approach is to have a central pipeline, the log, with a well defined API for adding data. The responsibility of integrating with this pipeline and providing a clean, well-structured data feed lies with the producer of this data feed. This means that as part of their system design and implementation they must consider the problem of getting data out and into a well structured form for delivery to the central pipeline. The addition of new storage systems is of no consequence to the data warehouse team as they have a central point of integration. The data warehouse team handles only the simpler problem of loading structured feeds of data from the central log and carrying out transformation specific to their system. <\/p><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/pipeline_ownership.png\" alt=\"image\" \/><\/p><p> This point about organizational scalability becomes particularly important when one considers adopting additional data systems beyond a traditional data warehouse. Say, for example, that one wishes to provide search capabilities over the complete data set of the organization. Or, say that one wants to provide sub-second monitoring of data streams with real-time trend graphs and alerting. In either of these cases, the infrastructure of the traditional data warehouse or even a Hadoop cluster is going to be inappropriate. Worse, the ETL processing pipeline built to support database loads is likely of no use for feeding these other systems, making bootstrapping these pieces of infrastructure as large an undertaking as adopting a data warehouse. This likely isn't feasible and probably helps explain why most organizations do not have these capabilities easily available for all their data. By contrast, if the organization had built out feeds of uniform, well-structured data, getting any new system full access to all data requires only a single bit of integration plumbing to attach to the pipeline. <\/p><p> This architecture also raises a set of different options for where a particular cleanup or transformation can reside: <\/p><ol><li>It can be done by the data producer prior to adding the data to the company wide log. <\/li><li>It can be done as a real-time transformation on the log (which in turn produces a new, transformed log) <\/li><li>It can be done as part of the load process into some destination data system <\/li><\/ol><p> The best model is to have <i>cleanup<\/i> done prior to publishing the data to the log by the publisher of the data. This means ensuring the data is in a canonical form and doesn't retain any hold-overs from the particular code that produced it or the storage system in which it may have been maintained. These details are best handled by the team that creates the data since they know the most about their own data. Any logic applied in this stage should be lossless and reversible. <\/p><p> Any kind of value-added transformation that can be done in real-time should be done as post-processing on the raw log feed produced. This would include things like sessionization of event data, or the addition of other derived fields that are of general interest. The original log is still available, but this real-time processing produces a derived log containing augmented data. <\/p><p> Finally, only aggregation that is specific to the destination system should be performed as part of the loading process. This might include transforming data into a particular star or snowflake schema for analysis and reporting in a data warehouse. Because this stage, which most naturally maps to the traditional ETL process, is now done on a far cleaner and more uniform set of streams, it should be much simplified. <\/p><h3>Log Files and Events<\/h3><p> Let's talk a little bit about a side benefit of this architecture: it enables decoupled, event-driven systems. <\/p><p> The typical approach to activity data in the web industry is to log it out to text files where it can be scrapped into a data warehouse or into Hadoop for aggregation and querying. The problem with this is the same as the problem with all batch ETL: it couples the data flow to the data warehouse's capabilities and processing schedule. <\/p><p> At LinkedIn, we have built our event data handling in a log-centric fashion. We are using Kafka as the central, multi-subscriber event log. We have defined several hundred event types, each capturing the unique attributes about a particular type of action. This covers everything from page views, ad impressions, and searches, to service invocations and application exceptions. <\/p><p> To understand the advantages of this, imagine a simple event\u2014showing a job posting on the job page. The job page should contain only the logic required to display the job. However, in a fairly dynamic site, this could easily become larded up with additional logic unrelated to showing the job. For example let's say we need to integrate the following systems: <\/p><ol><li> We need to send this data to Hadoop and data warehouse for offline processing purposes <\/li><li> We need to count the view to ensure that the viewer is not attempting some kind of content scraping <\/li><li> We need to aggregate this view for display in the Job poster's analytics page <\/li><li> We need to record the view to ensure we properly impression cap any job recommendations for that user (we don't want to show the same thing over and over) <\/li><li> Our recommendation system may need to record the view to correctly track the popularity of that job <\/li><li> Etc <\/li><\/ol><p> Pretty soon, the simple act of displaying a job has become quite complex. And as we add other places where jobs are displayed\u2014mobile applications, and so on\u2014this logic must be carried over and the complexity increases. Worse, the systems that we need to interface with are now somewhat intertwined\u2014the person working on displaying jobs needs to know about many other systems and features and make sure they are integrated properly. This is just a toy version of the problem, any real application would be more, not less, complex. <\/p><p> The \"event-driven\" style provides an approach to simplifying this. The job display page now just shows a job and records the fact that a job was shown along with the relevant attributes of the job, the viewer, and any other useful facts about the display of the job. Each of the other interested systems\u2014the recommendation system, the security system, the job poster analytics system, and the data warehouse\u2014all just subscribe to the feed and do their processing. The display code need not be aware of these other systems, and needn't be changed if a new data consumer is added. <\/p><h3>Building a Scalable Log<\/h3><p> Of course, separating publishers from subscribers is nothing new. But if you want to keep a commit log that acts as a multi-subscriber real-time journal of everything happening on a consumer-scale website, scalability will be a primary challenge. Using a log as a universal integration mechanism is never going to be more than an elegant fantasy if we can't build a log that is fast, cheap, and scalable enough to make this practical at scale. <\/p><p> Systems people typically think of a distributed log as a slow, heavy-weight abstraction (and usually associate it only with the kind of \"metadata\" uses for which Zookeeper might be appropriate). But with a thoughtful implementation focused on journaling large data streams, this need not be true. At LinkedIn we are currently running over 60 billion unique message writes through Kafka per day (several hundred billion if you count the writes from <a href=\"http:\/\/kafka.apache.org\/documentation.html#datacenters\" target=\"_blank\">mirroring between datacenters<\/a>). <\/p><p> We used a few tricks in Kafka to support this kind of scale: <\/p><ol><li>Partitioning the log <\/li><li>Optimizing throughput by batching reads and writes <\/li><li>Avoiding needless data copies <\/li><\/ol><p> In order to allow horizontal scaling we chop up our log into partitions: <\/p><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/partitioned_log.png\" alt=\"image\" \/> Each partition is a totally ordered log, but there is no global ordering between partitions (other than perhaps some wall-clock time you might include in your messages). The assignment of the messages to a particular partition is controllable by the writer, with most users choosing to partition by some kind of key (e.g. user id). Partitioning allows log appends to occur without co-ordination between shards and allows the throughput of the system to scale linearly with the Kafka cluster size. <\/p><p> Each partition is replicated across a configurable number of replicas, each of which has an identical copy of the partition's log. At any time, a single one of them will act as the leader; if the leader fails, one of the replicas will take over as leader. <\/p><p> Lack of a global order across partitions is a limitation, but we have not found it to be a major one. Indeed, interaction with the log typically comes from hundreds or thousands of distinct processes so it is not meaningful to talk about a total order over their behavior. Instead, the guarantees that we provide are that each partition is order preserving, and Kafka guarantees that appends to a particular partition from a single sender will be delivered in the order they are sent. <\/p><p> A log, like a filesystem, is easy to optimize for linear read and write patterns. The log can group small reads and writes together into larger, high-throughput operations. Kafka pursues this optimization aggressively. Batching occurs from client to server when sending data, in writes to disk, in replication between servers, in data transfer to consumers, and in acknowledging committed data. <\/p><p> Finally, Kafka uses a simple binary format that is maintained between in-memory log, on-disk log, and in network data transfers. This allows us to make use of numerous optimizations including <a href=\"https:\/\/www.ibm.com\/developerworks\/library\/j-zerocopy\">zero-copy data transfer<\/a>. <\/p><p> The cumulative effect of these optimizations is that you can usually write and read data at the rate supported by the disk or network, even while maintaining data sets that vastly exceed memory. <\/p><p> This write-up isn't meant to be primarily about Kafka so I won't go into further details. You can read a more detailed overview of LinkedIn's approach <a href=\"http:\/\/sites.computer.org\/debull\/A12june\/pipeline.pdf\" target=\"_blank\">here<\/a> and a thorough overview of Kafka's design <a href=\"http:\/\/kafka.apache.org\/documentation.html#design\" target=\"_blank\">here<\/a>. <\/p><h2>Part Three: Logs &amp; Real-time Stream Processing<\/h2><p> So far, I have only described what amounts to a fancy method of copying data from place-to-place. But shlepping bytes between storage systems is not the end of the story. It turns out that \"log\" is another word for \"stream\" and logs are at the heart of <a href=\"http:\/\/highlyscalable.wordpress.com\/2013\/08\/20\/in-stream-big-data-processing\/\" target=\"_blank\">stream processing<\/a>. <\/p><p> But, wait, what exactly is stream processing? <\/p><p> If you are a fan of late 90s and early 2000s <a href=\"http:\/\/cs.brown.edu\/research\/aurora\/vldb03_journal.pdf\" target=\"_blank\">database<\/a> <a href=\"http:\/\/db.cs.berkeley.edu\/papers\/cidr03-tcq.pdf\" target=\"_blank\">literature<\/a> or semi-successful <a href=\"http:\/\/www-03.ibm.com\/software\/products\/us\/en\/infosphere-streams\" target=\"_blank\">data<\/a> <a href=\"http:\/\/en.wikipedia.org\/wiki\/StreamBase_Systems\" target=\"_blank\">infrastructure<\/a> <a href=\"http:\/\/en.wikipedia.org\/wiki\/Truviso\" target=\"_blank\">products<\/a>, you likely associate stream processing with efforts to build a SQL engine or \"boxes and arrows\" interface for event driven processing. <\/p><p> If you follow the explosion of open source data systems, you likely associate stream processing with some of the systems in this space\u2014for example, <a href=\"http:\/\/storm-project.net\/\" target=\"_blank\">Storm<\/a>, <a href=\"http:\/\/akka.io\/\" target=\"_blank\">Akka<\/a>, <a href=\"http:\/\/incubator.apache.org\/s4\" target=\"_blank\">S4<\/a>, and Samza. But most people see these as a kind of asynchronous message processing system not that different from a cluster-aware RPC layer (and in fact some things in this space are exactly that). <\/p><p> Both these views are a little limited. Stream processing has nothing to do with SQL. Nor is it limited to real-time processing. There is no inherent reason you can't process the stream of data from yesterday or a month ago using a variety of different languages to express the computation. <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/census.jpg\" alt=\"image\" \/><\/p><p> I see stream processing as something much broader: infrastructure for continuous data processing. I think the computational model can be as general as MapReduce or other distributed processing frameworks, but with the ability to produce low-latency results. <\/p><p> The real driver for the processing model is the method of data collection. Data which is collected in batch is naturally processed in batch. When data is collected continuously, it is naturally processed continuously. <\/p><p> The US census provides a good example of batch data collection. The census periodically kicks off and does a brute force discovery and enumeration of US citizens by having people walking around door-to-door. This made a lot of sense in 1790 when the census was first begun. Data collection at the time was inherently batch oriented, it involved riding around on horseback and writing down records on paper, then transporting this batch of records to a central location where humans added up all the counts. These days, when you describe the census process one immediately wonders why we don't keep a journal of births and deaths and produce population counts either continuously or with whatever granularity is needed. <\/p><p> This is an extreme example, but many data transfer processes still depend on taking periodic dumps and bulk transfer and integration. The only natural way to process a bulk dump is with a batch process. But as these processes are replaced with continuous feeds, one naturally starts to move towards continuous processing to smooth out the processing resources needed and reduce latency. <\/p><p> LinkedIn, for example, has almost no batch data collection at all. The majority of our data is either activity data or database changes, both of which occur continuously. In fact, when you think about any business, the underlying mechanics are almost always a continuous process\u2014events happen in real-time, as Jack Bauer would tell us. When data is collected in batches, it is almost always due to some manual step or lack of digitization or is a historical relic left over from the automation of some non-digital process. Transmitting and reacting to data used to be very slow when the mechanics were mail and humans did the processing. A first pass at automation always retains the form of the original process, so this often lingers for a long time. <\/p><p> Production \"batch\" processing jobs that run daily are often effectively mimicking a kind of continuous computation with a window size of one day. The underlying data is, of course, always changing. These were actually so common at LinkedIn (and the mechanics of making them work in Hadoop so tricky) that we implemented a whole <a href=\"http:\/\/engineering.linkedin.com\/datafu\/datafus-hourglass-incremental-data-processing-hadoop\">framework<\/a> for managing incremental Hadoop workflows. <\/p><p> Seen in this light, it is easy to have a different view of stream processing: it is just processing which includes a notion of time in the underlying data being processed and does not require a static snapshot of the data so it can produce output at a user-controlled frequency instead of waiting for the \"end\" of the data set to be reached. In this sense, stream processing is a generalization of batch processing, and, given the prevalence of real-time data, a very important generalization. <\/p><p> So why has the traditional view of stream processing been as a niche application? I think the biggest reason is that a lack of real-time data collection made continuous processing something of an academic concern. <\/p><p> I think the lack of real-time data collection is likely what doomed the commercial stream-processing systems. Their customers were still doing file-oriented, daily batch processing for ETL and data integration. Companies building stream processing systems focused on providing processing engines to attach to real-time data streams, but it turned out that at the time very few people actually had real-time data streams. Actually, very early at my career at LinkedIn, a company tried to sell us a very cool stream processing system, but since all our data was collected in hourly files at that time, the best application we could come up with was to pipe the hourly files into the stream system at the end of the hour! They noted that this was a fairly common problem. The exception actually proves the rule here: finance, the one domain where stream processing has met with some success, was exactly the area where real-time data streams were already the norm and processing had become the bottleneck. <\/p><p> Even in the presence of a healthy batch processing ecosystem, I think the actual applicability of stream processing as an infrastructure style is quite broad. I think it covers the gap in infrastructure between real-time request\/response services and offline batch processing. For modern internet companies, I think around 25% of their code falls into this category. <\/p><p> It turns out that the log solves some of the most critical technical problems in stream processing, which I'll describe, but the biggest problem that it solves is just making data available in real-time multi-subscriber data feeds. For those interested in more details, we have open sourced <a href=\"http:\/\/samza.incubator.apache.org\/\" target=\"_blank\">Samza<\/a>, a stream processing system explicitly built on many of these ideas. We describe a lot of these applications in more detail in the documentation <a href=\"http:\/\/samza.incubator.apache.org\/learn\/documentation\/0.7.0\" target=\"_blank\">here<\/a>. <\/p><h3>Data flow graphs<\/h3><img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/dag.png\" alt=\"image\" \/><p> The most interesting aspect of stream processing has nothing to do with the internals of a stream processing system, but instead has to do with how it extends our idea of what a data feed is from the earlier data integration discussion. We discussed primarily feeds or logs of primary data\u2014the events and rows of data produced in the execution of various applications. But stream processing allows us to also include feeds computed off other feeds. These derived feeds look no different to consumers then the feeds of primary data from which they are computed. These derived feeds can encapsulate arbitrary complexity. <\/p><p> Let's dive into this a bit. A stream processing job, for our purposes, will be anything that reads from logs and writes output to logs or other systems. The logs they use for input and output join these processes into a graph of processing stages. Indeed, using a centralized log in this fashion, you can view all the organization's data capture, transformation, and flow as just a series of logs and processes that write to them. <\/p><p> A stream processor need not have a fancy framework at all: it can be any process or set of processes that read and write from logs, but additional infrastructure and support can be provided for helping manage processing code. <\/p><p> The purpose of the log in the integration is two-fold. <\/p><p> First, it makes each dataset multi-subscriber and ordered. Recall our \"state replication\" principle to remember the importance of order. To make this more concrete, consider a stream of updates from a database\u2014if we re-order two updates to the same record in our processing we may produce the wrong final output. This order is more permanent than what is provided by something like TCP as it is not limited to a single point-to-point link and survives beyond process failures and reconnections. <\/p><p> Second, the log provides buffering to the processes. This is very fundamental. If processing proceeds in an unsynchronized fashion it is likely to happen that an upstream data producing job will produce data more quickly than another downstream job can consume it. When this occurs processing must block, buffer or drop data. Dropping data is likely not an option; blocking may cause the entire processing graph to grind to a halt. The log acts as a very, very large buffer that allows process to be restarted or fail without slowing down other parts of the processing graph. This isolation is particularly important when extending this data flow to a larger organization, where processing is happening by jobs made by many different teams. We cannot have one faulty job cause back-pressure that stops the entire processing flow. <\/p><p> Both <a href=\"http:\/\/storm-project.net\/\" target=\"_blank\">Storm<\/a> and <a href=\"http:\/\/samza.incubator.apache.org\/\" target=\"_blank\">Samza<\/a> are built in this fashion and can use Kafka or other similar systems as their log. <\/p><h3>Stateful Real-Time Processing<\/h3><p> Some real-time stream processing is just stateless record-at-a-time transformation, but many of the uses are more sophisticated counts, aggregations, or joins over windows in the stream. One might, for example, want to enrich an event stream (say a stream of clicks) with information about the user doing the click\u2014in effect joining the click stream to the user account database. Invariably, this kind of processing ends up requiring some kind of state to be maintained by the processor: for example, when computing a count, you have the count so far to maintain. How can this kind of state be maintained correctly if the processors themselves can fail? <\/p><p> The simplest alternative would be to keep state in memory. However if the process crashed it would lose its intermediate state. If state is only maintained over a window, the process could just fall back to the point in the log where the window began. However, if one is doing a count over an hour, this may not be feasible. <\/p><p> An alternative is to simply store all state in a remote storage system and join over the network to that store. The problem with this is that there is no locality of data and lots of network round-trips. <\/p><p> How can we support something like a \"table\" that is partitioned up with our processing? <\/p><p> Well recall the discussion of the duality of tables and logs. This gives us exactly the tool to be able to convert streams to tables co-located with our processing, as well as a mechanism for handling fault tolerance for these tables. <\/p><p> A stream processor can keep it's state in a local \"table\" or \"index\"\u2014a <a href=\"http:\/\/www.oracle.com\/technetwork\/products\/berkeleydb\" target=\"_blank\">bdb<\/a>, <a href=\"https:\/\/code.google.com\/p\/leveldb\">leveldb<\/a>, or even something more unusual such as a <a href=\"http:\/\/lucene.apache.org\/\" target=\"_blank\">Lucene<\/a> or <a href=\"https:\/\/sdm.lbl.gov\/fastbit\">fastbit<\/a> index. The contents of this this store is fed from its input streams (after first perhaps applying arbitrary transformation). It can journal out a changelog for this local index it keeps to allow it to restore its state in the event of a crash and restart. This mechanism allows a generic mechanism for keeping co-partitioned state in arbitrary index types local with the incoming stream data. <\/p><p> When the process fails, it restores its index from the changelog. The log is the transformation of the local state into a sort of incremental record at a time backup. <\/p><p> This approach to state management has the elegant property that the state of the processors is also maintained as a log. We can think of this log just like we would the log of changes to a database table. In fact, the processors have something very like a co-partitioned table maintained along with them. Since this state is itself a log, other processors can subscribe to it. This can actually be quite useful in cases when the goal of the processing is to update a final state and this state is the natural output of the processing. <\/p><p> When combined with the logs coming out of databases for data integration purposes, the power of the log\/table duality becomes clear. A change log may be extracted from a database and indexed in different forms by various stream processors to join against event streams. <\/p><p> We give more detail on this style of managing stateful processing in Samza and a lot more practical examples <a href=\"http:\/\/samza.incubator.apache.org\/learn\/documentation\/0.7.0\/container\/state-management.html\" target=\"_blank\">here<\/a>. <\/p><h3>Log Compaction<\/h3><p> <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/log_compaction_0.png\" alt=\"image\" \/> Of course, we can't hope to keep a complete log for all state changes for all time. Unless one wants to use infinite space, somehow the log must be cleaned up. I'll talk a little about the implementation of this in Kafka to make it more concrete. In Kafka, cleanup has two options depending on whether the data contains keyed updates or event data. For event data, Kafka supports just retaining a window of data. Usually, this is configured to a few days, but the window can be defined in terms of time or space. For keyed data, though, a nice property of the complete log is that you can replay it to recreate the state of the source system (potentially recreating it in another system). <\/p><p> However, retaining the complete log will use more and more space as time goes by, and the replay will take longer and longer. Hence, in Kafka, we support a different type of retention. Instead of simply throwing away the old log, we remove obsolete records\u2014i.e. records whose primary key has a more recent update. By doing this, we still guarantee that the log contains a complete backup of the source system, but now we can no longer recreate <i>all<\/i> previous states of the source system, only the more recent ones. We call this feature <a href=\"https:\/\/cwiki.apache.org\/confluence\/display\/KAFKA\/Log+Compaction\">log compaction<\/a>. <\/p><h2>Part Four: System Building<\/h2><p> The final topic I want to discuss is the role of the log in data system design for online data systems. <\/p><p> There is an analogy here between the role a log serves for data flow inside a distributed database and the role it serves for data integration in a larger organization. In both cases, it is responsible for data flow, consistency, and recovery. What, after all, is an organization, if not a very complicated distributed data system? <\/p><h3>Unbundling?<\/h3><p> So maybe if you squint a bit, you can see the whole of your organization's systems and data flows as a single distributed database. You can view all the individual query-oriented systems (Redis, SOLR, Hive tables, and so on) as just particular indexes on your data. You can view the stream processing systems like Storm or Samza as just a very well-developed trigger and view materialization mechanism. Classical database people, I have noticed, like this view very much because it finally explains to them what on earth people are doing with all these different data systems\u2014they are just different index types! <\/p><p> There is undeniably now an explosion of types of data systems, but in reality, this complexity has always existed. Even in the heyday of the relational database, organizations had lots and lots of relational databases! So perhaps real integration hasn't existed since the mainframe when all the data really was in one place. There are many motivations for segregating data into multiple systems: scale, geography, security, and performance isolation are the most common. But these issues can be addressed by a good system: it is possible for an organization to have a single Hadoop cluster, for example, that contains all the data and serves a large and diverse constituency. <\/p><p> So there is already one possible simplification in the handling of data that has become possible in the move to distributed systems: coalescing lots of little instances of each system into a few big clusters. Many systems aren't good enough to allow this yet: they don't have security, or can't guarantee performance isolation, or just don't scale well enough. But each of these problems is solvable. <\/p><p> My take is that the explosion of different systems is caused by the difficulty of building distributed data systems. By cutting back to a single query type or use case each system is able to bring its scope down into the set of things that are feasible to build. But running all these systems yields too much complexity. <\/p><p> I see three possible directions this could follow in the future. <\/p><p> The first possibility is a continuation of the status quo: the separation of systems remains more or less as it is for a good deal longer. This could happen either because the difficulty of distribution is too hard to overcome or because this specialization allows new levels of convenience and power for each system. As long as this remains true, the data integration problem will remain one of the most centrally important things for the successful use of data. In this case, an external log that integrates data will be very important. <\/p><p> The second possibility is that there could be a re-consolidation in which a single system with enough generality starts to merge back in all the different functions into a single uber-system. This uber-system could be like the relational database superficially, but it's use in an organization would be far different as you would need only one big one instead of umpteen little ones. In this world, there is no real data integration problem except what is solved inside this system. I think the practical difficulties of building such a system make this unlikely. <\/p><p> There is another possible outcome, though, which I actually find appealing as an engineer. One interesting facet of the new generation of data systems is that they are virtually all open source. Open source allows another possibility: data infrastructure could be unbundled into a collection of services and application-facing system apis. You already see this happening to a certain extent in the Java stack: <\/p><ul><li><a href=\"http:\/\/zookeeper.apache.org\/\" target=\"_blank\">Zookeeper<\/a> handles much of the system co-ordination (perhaps with a bit of help from higher-level abstractions like <a href=\"http:\/\/helix.incubator.apache.org\/\" target=\"_blank\">Helix<\/a> or <a href=\"http:\/\/curator.incubator.apache.org\/\" target=\"_blank\">Curator<\/a>). <\/li><li><a href=\"http:\/\/mesos.apache.org\/\" target=\"_blank\">Mesos<\/a> and <a href=\"http:\/\/hadoop.apache.org\/docs\/current\/hadoop-yarn\/hadoop-yarn-site\/YARN.html\" target=\"_blank\">YARN<\/a> do process virtualization and resource management <\/li><li>Embedded libraries like <a href=\"http:\/\/lucene.apache.org\/\" target=\"_blank\">Lucene<\/a> and <a href=\"https:\/\/code.google.com\/p\/leveldb\">LevelDB<\/a> do indexing <\/li><li><a href=\"http:\/\/netty.io\/\" target=\"_blank\">Netty<\/a>, <a href=\"http:\/\/www.eclipse.org\/jetty\" target=\"_blank\">Jetty<\/a> and higher-level wrappers like <a href=\"http:\/\/twitter.github.io\/finagle\" target=\"_blank\">Finagle<\/a> and <a href=\"http:\/\/rest.li\/\" target=\"_blank\">rest.li<\/a> handle remote communication <\/li><li><a href=\"http:\/\/avro.apache.org\/\" target=\"_blank\">Avro<\/a>, <a href=\"https:\/\/code.google.com\/p\/protobuf\">Protocol Buffers<\/a>, <a href=\"http:\/\/thrift.apache.org\/\" target=\"_blank\">Thrift<\/a>, and <a href=\"https:\/\/github.com\/eishay\/jvm-serializers\/wiki\">umpteen zillion<\/a> other libraries handle serialization <\/li><li><a href=\"http:\/\/kafka.apache.org\/\" target=\"_blank\">Kafka<\/a> and <a href=\"http:\/\/zookeeper.apache.org\/bookkeeper\" target=\"_blank\">Bookeeper<\/a> provide a backing log. <\/li><\/ul><p> If you stack these things in a pile and squint a bit, it starts to look a bit like a lego version of distributed data system engineering. You can piece these ingredients together to create a vast array of possible systems. This is clearly not a story relevant to end-users who presumably care primarily more about the API then how it is implemented, but it might be a path towards getting the simplicity of the single system in a more diverse and modular world that continues to evolve. If the implementation time for a distributed system goes from years to weeks because reliable, flexible building blocks emerge, then the pressure to coalesce into a single monolithic system disappears. <\/p><h3>The place of the log in system architecture<\/h3><p> A system that assumes an external log is present allows the individual systems to relinquish a lot of their own complexity and rely on the shared log. Here are the things I think a log can do: <\/p><ul><li>Handle data consistency (whether eventual or immediate) by sequencing concurrent updates to nodes <\/li><li>Provide data replication between nodes <\/li><li>Provide \"commit\" semantics to the writer (i.e. acknowledging only when your write guaranteed not to be lost) <\/li><li>Provide the external data subscription feed from the system <\/li><li>Provide the capability to restore failed replicas that lost their data or bootstrap new replicas <\/li><li>Handle rebalancing of data between nodes. <\/li><\/ul><p> This is actually a substantial portion of what a distributed data system does. In fact, the majority of what is left over is related to the final client-facing query API and indexing strategy. This is exactly the part that should vary from system to system: for example, a full-text search query may need to query all partitions whereas a query by primary key may only need to query a single node responsible for that key's data. <\/p><p> Here is how this works. The system is divided into two logical pieces: the log and the serving layer. The log captures the state changes in sequential order. The serving nodes store whatever index is required to serve queries (for example a key-value store might have something like a btree or sstable, a search system would have an inverted index). Writes may either go directly to the log, though they may be proxied by the serving layer. Writing to the log yields a logical timestamp (say the index in the log). If the system is partitioned, and I assume it is, then the log and the serving nodes will have the same number of partitions, though they may have very different numbers of machines. <img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/system.png\" alt=\"image\" \/><\/p><p> The serving nodes subscribe to the log and apply writes as quickly as possible to its local index in the order the log has stored them. <\/p><p> The client can get read-your-write semantics from any node by providing the timestamp of a write as part of its query\u2014a serving node receiving such a query will compare the desired timestamp to its own index point and if necessary delay the request until it has indexed up to at least that time to avoid serving stale data. <\/p><p> The serving nodes may or may not need to have any notion of \"mastership\" or \"leader election\". For many simple use cases, the serving nodes can be completely without leaders, since the log is the source of truth. <\/p><p> One of the trickier things a distributed system must do is handle restoring failed nodes or moving partitions from node to node. A typical approach would have the log retain only a fixed window of data and combine this with a snapshot of the data stored in the partition. It is equally possible for the log to retain a complete copy of data and <a href=\"https:\/\/cwiki.apache.org\/confluence\/display\/KAFKA\/Log+Compaction\">garbage collect the log itself<\/a>. This moves a significant amount of complexity out of the serving layer, which is system-specific, and into the log, which can be general purpose. <\/p><p> By having this log system, you get a fully developed subscription API for the contents of the data store which feeds ETL into other systems. In fact, many systems can share the same the log while providing different indexes, like this: <\/p><div><img src=\"https:\/\/content.linkedin.com\/content\/dam\/engineering\/en-us\/blog\/migrated\/full-stack.png\" alt=\"image\" \/><\/div><p> Note how such a log-centric system is itself immediately a provider of data streams for processing and loading in other systems. Likewise, a stream processor can consume multiple input streams and then serve them via another system that indexes that output. <\/p><p> I find this view of systems as factored into a log and query api to very revealing, as it lets you separate the query characteristics from the availability and consistency aspects of the system. I actually think this is even a useful way to mentally factor a system that isn't built this way to better understand it. <\/p><p> It's worth noting that although Kafka and Bookeeper are consistent logs, this is not a requirement. You could just as easily factor a <a href=\"http:\/\/www.allthingsdistributed.com\/2007\/10\/amazons_dynamo.html\" target=\"_blank\">Dynamo<\/a>-like database into an eventually consistent <a href=\"http:\/\/en.wikipedia.org\/wiki\/CAP_theorem\" target=\"_blank\">AP<\/a> log and a key-value serving layer. Such a log is a bit tricky to work with, as it will redeliver old messages and depends on the subscriber to handle this (much like Dynamo itself). <\/p><p> The idea of having a separate copy of data in the log (especially if it is a complete copy) strikes many people as wasteful. In reality, though there are a few factors that make this less of an issue. First, the log can be a particularly efficient storage mechanism. We store over 75TB per datacenter on our production Kafka servers. Meanwhile many serving systems require much more memory to serve data efficiently (text search, for example, is often all in memory). The serving system may also use optimized hardware. For example, most our live data systems either serve out of memory or else use SSDs. In contrast, the log system does only linear reads and writes, so it is quite happy using large multi-TB hard drives. Finally, as in the picture above, in the case where the data is served by multiple systems, the cost of the log is amortized over multiple indexes. This combination makes the expense of an external log pretty minimal. <\/p><p> This is exactly the pattern that LinkedIn has used to build out many of its own real-time query systems. These systems feed off a database (using Databus as a log abstraction or off a dedicated log from Kafka) and provide a particular partitioning, indexing, and query capability on top of that data stream. This is the way we have implemented our search, social graph, and OLAP query systems. In fact, it is quite common to have a single data feed (whether a live feed or a derived feed coming from Hadoop) replicated into multiple serving systems for live serving. This has proven to be an enormous simplifying assumption. None of these systems need to have an externally accessible write api at all, Kafka and databases are used as the system of record and changes flow to the appropriate query systems through that log. Writes are handled locally by the nodes hosting a particular partition. These nodes blindly transcribe the feed provided by the log to their own store. A failed node can be restored by replaying the upstream log. <\/p><p> The degree to which these systems rely on the log varies. A fully reliant system could make use of the log for data partitioning, node restore, rebalancing, and all aspects of consistency and data propagation. In this setup, the actual serving tier is actually nothing less than a sort of \"cache\" structured to enable a particular type of processing with writes going directly to the log. <\/p><h2>The End<\/h2><p> If you made it this far you know most of what I know about logs. <\/p><p> Here are a few interesting references you may want to check out. <\/p><p> Everyone seems to uses different terms for the same things so it is a bit of a puzzle to connect the database literature to the distributed systems stuff to the various enterprise software camps to the open source world. Nonetheless, here are a few pointers in the general direction. <\/p><p> Academic papers, systems, talks, and blogs: <\/p><ul><li>A good overview of <a href=\"http:\/\/www.cs.cornell.edu\/fbs\/publications\/smsurvey.pdf\" target=\"_blank\">state machine<\/a> and <a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/summary?doi=10.1.1.20.5896\" target=\"_blank\">primary-backup<\/a> replication <\/li><li><a href=\"http:\/\/research.microsoft.com\/apps\/pubs\/default.aspx?id=66814\" target=\"_blank\">PacificA<\/a> is a generic framework for implementing log-based distributed storage systems at Microsoft. <\/li><li><a href=\"http:\/\/static.googleusercontent.com\/external_content\/untrusted_dlcp\/research.google.com\/en\/us\/archive\/spanner-osdi2012.pdf\" target=\"_blank\">Spanner<\/a>\u2014Not everyone loves logical time for their logs. Google's new database tries to use physical time and models the uncertainty of clock drift directly by treating the timestamp as a range. <\/li><li><a href=\"http:\/\/www.datomic.com\/\" target=\"_blank\">Datanomic<\/a>: <a href=\"https:\/\/www.youtube.com\/watch?v=Cym4TZwTCNU\">Deconstructing the database<\/a> is a great presentation by Rich Hickey, the creator of Clojure, on his startup's database product. <\/li><li><a href=\"http:\/\/www.cs.utexas.edu\/~lorenzo\/papers\/SurveyFinal.pdf\" target=\"_blank\">A Survey of Rollback-Recovery Protocols in Message-Passing Systems<\/a>. I found this to be a very helpful introduction to fault-tolerance and the practical application of logs to recovery outside databases. <\/li><li><a href=\"http:\/\/www.reactivemanifesto.org\/\" target=\"_blank\">Reactive Manifesto<\/a>\u2014I'm actually not quite sure what is meant by reactive programming, but I think it means the same thing as \"event driven\". This link doesn't have much info, but <a href=\"https:\/\/www.coursera.org\/course\/reactive\">this class<\/a> by Martin Odersky (of Scala fame) looks facinating. <\/li><li>Paxos! <ul><li>Original paper is <a href=\"http:\/\/research.microsoft.com\/en-us\/um\/people\/lamport\/pubs\/lamport-paxos.pdf\" target=\"_blank\">here<\/a>. Leslie Lamport has an interesting <a href=\"http:\/\/research.microsoft.com\/en-us\/um\/people\/lamport\/pubs\/pubs.html#lamport-paxos\" target=\"_blank\">history<\/a> of how the algorithm was created in the 1980s but not published until 1998 because the reviewers didn't like the Greek parable in the paper and he didn't want to change it. <\/li><li>Even once the original paper was published it wasn't well understood. Lamport <a href=\"http:\/\/research.microsoft.com\/en-us\/um\/people\/lamport\/pubs\/paxos-simple.pdf\" target=\"_blank\">tries again<\/a> and this time even includes a few of the \"uninteresting details\" of how to put it to use using these new-fangled automatic computers. It is still not widely understood. <\/li><li><a href=\"http:\/\/www.cs.cornell.edu\/fbs\/publications\/SMSurvey.pdf\" target=\"_blank\">Fred Schneider<\/a> and <a href=\"http:\/\/research.microsoft.com\/en-us\/um\/people\/blampson\/58-consensus\/Abstract.html\">Butler Lampson<\/a> each give more detailed overview of applying Paxos in real systems. <\/li><li>A few Google engineers summarize <a href=\"http:\/\/www.cs.utexas.edu\/users\/lorenzo\/corsi\/cs380d\/papers\/paper2-1.pdf\" target=\"_blank\">their experience<\/a> implementing Paxos in Chubby. <\/li><li>I actually found all the Paxos papers pretty painful to understand but dutifully struggled through. But you don't need to because <a href=\"https:\/\/www.youtube.com\/watch?v=JEpsBg0AO6o\">this video<\/a> by <a href=\"http:\/\/www.stanford.edu\/~ouster\/cgi-bin\/papers\/lfs.pdf\" target=\"_blank\">John Ousterhout<\/a> (of log-structured filesystem fame!) will make it all very simple. Somehow these consensus algorithms are much better presented by drawing them as the communication rounds unfold, rather than in a static presentation in a paper. Ironically, this video was created in an attempt to show that Paxos was hard to understand. <\/li><li><a href=\"http:\/\/arxiv.org\/pdf\/1103.2408.pdf\" target=\"_blank\">Using Paxos to Build a Scalable Consistent Data Store<\/a>: This is a cool paper on using a log to build a data store, by Jun, one of the co-authors is also one of the earliest engineers on Kafka. <\/li><\/ul><\/li><li>Paxos has competitors! Actually each of these map a lot more closely to the implementation of a log and are probably more suitable for practical implementation: <ul><li><a href=\"http:\/\/pmg.csail.mit.edu\/papers\/vr-revisited.pdf\" target=\"_blank\">Viewstamped Replication<\/a> by Barbara Liskov is an early algorithm to directly model log replication. <\/li><li><a href=\"http:\/\/www.stanford.edu\/class\/cs347\/reading\/zab.pdf\" target=\"_blank\">Zab<\/a> is the algorithm used by Zookeeper. <\/li><li><a href=\"https:\/\/ramcloud.stanford.edu\/wiki\/download\/attachments\/11370504\/raft.pdf\">RAFT<\/a> is an attempt at a more understandable consensus algorithm. The <a href=\"https:\/\/www.youtube.com\/watch?v=YbZ3zDzDnrw\">video presentation<\/a>, also by John Ousterhout, is great too. <\/li><\/ul><\/li><li>You can see the role of the log in action in different real distributed databases. <ul><li><a href=\"http:\/\/www.mpi-sws.org\/~druschel\/courses\/ds\/papers\/cooper-pnuts.pdf\" target=\"_blank\">PNUTS<\/a> is a system which attempts to apply to log-centric design of traditional distributed databases at large scale. <\/li><li><a href=\"http:\/\/hbase.apache.org\/\" target=\"_blank\">HBase<\/a> and <a href=\"http:\/\/research.google.com\/archive\/bigtable.html\">Bigtable<\/a> both give another example of logs in modern databases. <\/li><li>LinkedIn's own distributed database <a href=\"http:\/\/www.slideshare.net\/amywtang\/espresso-20952131\" target=\"_blank\">Espresso<\/a>, like PNUTs, uses a log for replication, but takes a slightly different approach using the underlying table itself as the source of the log. <\/li><\/ul><\/li><li>If you find yourself comparison shopping for a replication algorithm, <a href=\"http:\/\/arxiv.org\/abs\/1309.5671\" target=\"_blank\">this paper<\/a> may help you out. <\/li><li><a href=\"http:\/\/www.amazon.com\/Replication-Practice-Lecture-Computer-Theoretical\/dp\/3642112935\" target=\"_blank\">Replication: Theory and Practice<\/a> is a great book that collects a bunch of summary papers on replication in distributed systems. Many of the chapters are online (e.g. <a href=\"http:\/\/disi.unitn.it\/~montreso\/ds\/papers\/replication.pdf\" target=\"_blank\">1<\/a>, <a href=\"http:\/\/research.microsoft.com\/en-us\/people\/aguilera\/stumbling-chapter.pdf\" target=\"_blank\">4<\/a>, <a href=\"http:\/\/www.distributed-systems.net\/papers\/2010.verita.pdf\" target=\"_blank\">5<\/a>, <a href=\"http:\/\/www.cs.cornell.edu\/ken\/history.pdf\" target=\"_blank\">6<\/a>, <a href=\"http:\/\/www.pmg.csail.mit.edu\/papers\/vr-to-bft.pdf\" target=\"_blank\">7<\/a>, <a href=\"http:\/\/engineering.linkedin.com\/distributed-systems\/www.cs.cornell.edu\/fbs\/publications\/TrustSurveyTR.pdf\">8<\/a>). <\/li><li>Stream processing. This is a bit too broad to summarize, but here are a few things I liked. <ul><li><a href=\"http:\/\/infolab.usc.edu\/csci599\/Fall2002\/paper\/DML2_streams-issues.pdf\" target=\"_blank\">Models and Issues in Data Stream Systems<\/a>: probably the best overview of the early research in this area. <\/li><li><a href=\"http:\/\/cs.brown.edu\/research\/aurora\/hwang.icde05.ha.pdf\" target=\"_blank\">High-Availability Algorithms for Distributed Stream Processing<\/a> <\/li><li>A couple of random systems papers: <ul><li><a href=\"http:\/\/db.cs.berkeley.edu\/papers\/cidr03-tcq.pdf\" target=\"_blank\">TelegraphCQ<\/a> <\/li><li><a href=\"http:\/\/cs.brown.edu\/research\/aurora\/vldb03_journal.pdf\" target=\"_blank\">Aurora<\/a> <\/li><li><a href=\"http:\/\/research.cs.wisc.edu\/niagara\/papers\/NiagaraCQ.pdf\" target=\"_blank\">NiagaraCQ<\/a> <\/li><li><a href=\"http:\/\/www.cs.berkeley.edu\/~matei\/papers\/2012\/hotcloud_spark_streaming.pdf\" target=\"_blank\">Discretized Streams<\/a>: This paper discusses Spark's streaming system. <\/li><li><a href=\"http:\/\/research.google.com\/pubs\/pub41378.html\" target=\"_blank\">MillWheel<\/a> is one of Google's stream processing systems. <\/li><li><a href=\"http:\/\/research.microsoft.com\/apps\/pubs\/?id=201100\" target=\"_blank\">Naiad: A Timely Dataflow System<\/a> <\/li><\/ul><\/li><\/ul><\/li><\/ul>Enterprise software has all the same problems but with different names, a smaller scale, and XML. Ha ha, just kidding. Kind of.<ul><li><a href=\"http:\/\/martinfowler.com\/eaaDev\/EventSourcing.html\" target=\"_blank\">Event Sourcing<\/a>\u2014As far as I can tell this is basically the enterprise software engineer's way of saying \"state machine replication\". It's interesting that the same idea would be invented again in such a different context. Event sourcing seems to focus on smaller, in-memory use cases. This approach to application development seems to combine the \"stream processing\" that occurs on the log of events with the application. Since this becomes pretty non-trivial when the processing is large enough to require data partitioning for scale I focus on stream processing as a separate infrastructure primitive. <\/li><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Change_data_capture\" target=\"_blank\">Change Data Capture<\/a>\u2014There is a small industry around getting data out of databases, and this is the most log-friendly style of data extraction. <\/li><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Enterprise_application_integration\" target=\"_blank\">Enterprise Application Integration<\/a> seems to be about solving the data integration problem when what you have is a collection of off-the-shelf enterprise software like CRM or supply-chain management software. <\/li><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Complex_event_processing\" target=\"_blank\">Complex Event Processing (CEP)<\/a>: Fairly certain nobody knows what this means or how it actually differs from stream processing. The difference seems to be that the focus is on unordered streams and on event filtering and detection rather than aggregation, but this, in my opinion is a distinction without a difference. I think any system that is good at one should be good at another. <\/li><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Enterprise_service_bus\" target=\"_blank\">Enterprise Service Bus<\/a>\u2014I think the enterprise service bus concept is very similar to some of the ideas I have described around data integration. This idea seems to have been moderately successful in enterprise software communities and is mostly unknown among web folks or the distributed data infrastructure crowd. <\/li><\/ul>Interesting open source stuff:<ul><li><a href=\"http:\/\/kafka.apache.org\/\" target=\"_blank\">Kafka<\/a> Is the \"log as a service\" project that is the basis for much of this post. <\/li><li><a href=\"http:\/\/zookeeper.apache.org\/bookkeeper\/\" target=\"_blank\">Bookeeper<\/a> and <a href=\"https:\/\/cwiki.apache.org\/confluence\/display\/BOOKKEEPER\/HedWig\">Hedwig<\/a> comprise another open source \"log as a service\". They seem to be more targeted at data system internals then at event data. <\/li><li><a href=\"https:\/\/github.com\/linkedin\/databus\">Databus<\/a> is a system that provides a log-like overlay for database tables. <\/li><li><a href=\"http:\/\/akka.io\/\" target=\"_blank\">Akka<\/a> is an actor framework for Scala. It has an add on, <a href=\"https:\/\/github.com\/eligosource\/eventsourced\">eventsourced<\/a>, that provides persistence and journaling. <\/li><li><a href=\"http:\/\/samza.incubator.apache.org\/\" target=\"_blank\">Samza<\/a> is a stream processing framework we are working on at LinkedIn. It uses a lot of the ideas in this article as well as integrating with Kafka as the underlying log. <\/li><li><a href=\"http:\/\/storm-project.net\/\" target=\"_blank\">Storm<\/a> is popular stream processing framework that integrates well with Kafka. <\/li><li><a href=\"http:\/\/spark.incubator.apache.org\/docs\/0.7.3\/streaming-programming-guide.html\" target=\"_blank\">Spark Streaming<\/a> is a stream processing framework that is part of <a href=\"http:\/\/spark.incubator.apache.org\/\" target=\"_blank\">Spark<\/a>. <\/li><li><a href=\"https:\/\/blog.twitter.com\/2013\/streaming-mapreduce-with-summingbird\">Summingbird<\/a> is a layer on top of Storm or Hadoop that provides a convenient computing abstraction. <\/li><\/ul><p> I try to keep up on this area so if you know of some things I've left out, let me know. <\/p><p> I leave you with this message: <\/p><iframe frameborder=\"0\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/2C7mNr5WMjA\" width=\"420\"> <\/iframe>","created_at":"2019-08-08T13:04:42+0000","updated_at":"2019-08-08T13:04:53+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":63,"domain_name":"engineering.linkedin.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","datalog","datomic"],"is_public":false,"id":991,"title":"Rich Hickey: Deconstructing the Database","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=Cym4TZwTCNU","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/Cym4TZwTCNU?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-08-07T22:30:27+0000","updated_at":"2019-08-07T22:30:38+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/Cym4TZwTCNU\/maxresdefault.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["career","engineering","advice"],"is_public":false,"id":990,"title":"All the best engineering advice I stole from non-technical people","url":"https:\/\/medium.com\/@bellmar\/all-the-best-engineering-advice-i-stole-from-non-technical-people-eb7f90ca2f5f","content":"<div><div class=\"jk\"><div class=\"as ar\"><div><a href=\"https:\/\/medium.com\/@bellmar?source=post_page---------------------------\"><img alt=\"Marianne Bellotti\" src=\"https:\/\/miro.medium.com\/fit\/c\/96\/96\/0*rw2MhHHnUwwnb-Op.\" class=\"l ey jl jm\" width=\"48\" height=\"48\" \/><\/a><\/div><\/div><\/div><\/div><figure class=\"jv jw jx jy jz ei ka v w paragraph-image\"><figcaption class=\"ck eg ki kj hg dg v w kk cf ef\">Vector created by <a href=\"https:\/\/www.freepik.com\/free-photos-vectors\/business\" class=\"bp db kl km kn ko\">rawpixel<\/a><\/figcaption><\/figure><p id=\"1eff\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">As I focus on becoming a better manager of engineers, I have been reflecting more and more on the advice that produced a 10X boost in my abilities at that same stage. More often than not the best advice, the things that stuck with me, came from people who had no background at all in software.<\/p><p id=\"4730\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">I have been blessed (or cursed depending on your viewpoint I guess) with the privilege of high impact work on critical systems. This is work that naturally brings out a lot of self-doubt and anxiety. Over time I have found myself facing the same challenges, writing down the same thoughts and retelling the same stories in response. It\u2019s intriguing that the stuff that really seems to make a difference in the quality of software never seems to be about software.<\/p><p id=\"56fb\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">These are five of my favorites.<\/p><p id=\"b423\" class=\"kp kq be cg kr b ks lq ku lr kw ls ky lt la lu lc\"><strong class=\"kr lv\">Who said it:<\/strong> a Senior Official at the National Security Agency (NSA)<\/p><p id=\"dd14\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">Context:<\/strong> There\u2019s a long story about how I ended up in a State Department conference room with the NSA, but suffice to say it was a pretty crazy meeting in which a lot of very important people with impressive titles said a lot of stupid things about computer security. This would be the first of many times in my career in the federal government where we came in expecting to have a productive meeting with senior level leadership and were treated instead to a completely pointless dick measuring contest between three letter agencies. I remember the woman \u2014 and yes it was a woman \u2014 who made this comment was redirecting the conversation after her colleague had said something both wrong and insulting in response to a question from DHS. She was trying to defuse a situation, I doubt she thought much about her choice of words, but her insight was sharp and to the point. Security and reliability are more likely to go wrong in the seams between components. That means literal integrations, but it also means organization seams. Places where no one is sure who owns what, or who is responsible for what are unlikely to have proper monitoring and much more likely to be two or three upgrades behind. The seams are where things get lost, sometimes for years. So if your mandate is security or availability the seams are your best bet of finding a big pay off.<\/p><p id=\"d94f\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">How I changed my approach: <\/strong>Shortly there after I started applying a creative writing technique called 100:10:1 to evaluating and rescuing software systems. The idea behind 100:10:1 is simple:<\/p><ul class=\"\"><li id=\"a7d3\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc lw lx ly\">Brainstorm <strong class=\"kr lv\">100<\/strong> things that could go wrong<\/li><li id=\"11e3\" class=\"kp kq be cg kr b ks lz ku ma kw mb ky mc la md lc lw lx ly\">Pick <strong class=\"kr lv\">10<\/strong> on that list that feel like the most likely and investigate them<\/li><li id=\"34df\" class=\"kp kq be cg kr b ks lz ku ma kw mb ky mc la md lc lw lx ly\">Find the <strong class=\"kr lv\">1 <\/strong>critical problem you\u2019re going to focus on.<\/li><\/ul><p id=\"5495\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">The numbers themselves are not significant. The idea behind 100:10:1 is that you generate so many potential ideas you scrape the absolute bottom of the barrel and bring non-obvious things to the forefront. As you dig into the ones that feel the most likely, over time the list whittles itself down. Some ideas just end up flat out wrong, some merge with other ideas, some split off, some get ruled out because there\u2019s already someone looking at them\u2026 eventually you hit on something in the seams that everyone has missed.<\/p><p id=\"447c\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Approaching problems this way also changed the way my team behaved for the better. Often the hardest part of working as a team is giving people permission to bring up ideas that are ultimately ruled out. People tend to self-censor to avoid looking or sounding wrong, because they assume everyone else is keepIng score. When we allow this on a team it creates blindspots. Fortunately, humans are very bad with probabilities. Very bad. I became comfortable with being wrong 95% of the time because I understood that when I found that one critical problem all the ideas I had before that had been ruled out would be forgotten. I found a few problems much more experienced engineers had missed and it gave me super powers in the eyes of my colleagues.<\/p><p id=\"db2c\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">When I shifted to a leader of teams the benefits became more pronounced. We talk a lot about psychological safety and giving people permission to speak their mind, but there\u2019s not a whole lot of guidance around how you do that. Being a person who is not afraid to ask the stupid questions and who lets ideas get thrown out easily without a big argument has created an environment where correcting each other and debate doesn\u2019t feel like that big a deal. It normalizes the process and people stop keeping score because it\u2019s clear that their boss isn\u2019t.<\/p><p id=\"543b\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Often I find it useful to just acknowledge the situation up front: I tend to think in edge cases, but edge cases by definition are unlikely. As a result I tend to bring up a lot of irrelevant things. When you\u2019re discussing a problem or situation there are basically three buckets of information: things that are true, things that are false, and things that are true but irrelevant. Having a culture where people are not optimizing to avoid mentioning something true but irrelevant means we make better technical decisions overall because we get the benefit of every teammate\u2019s complete perspective.<\/p><p id=\"9a57\" class=\"kp kq be cg kr b ks lq ku lr kw ls ky lt la lu lc\"><strong class=\"kr lv\">Who said it:<\/strong> Leslie Ryan, my TOEFL instructor<\/p><p id=\"514d\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">Context:<\/strong> Leslie certified me to teach english as a foreign language, a skill that allowed me to travel the world and work for small NGOs that could not otherwise pay me a living wage or sponsor a visa. He told a group of us this on the eve of our first solo experience in front of a classroom. I got the impression that he gave this speech to every group he certified at precisely that moment in the program. Knowing that made it no less impactful.<\/p><p id=\"3a7b\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Leslie pointed out to us that most of our future students would be professional people. They didn\u2019t need us to teach them about business, law, medicine, or economics. They were already experts in those things. What they needed from us was to learn how to do those things in English. They wanted to access our native fluency in order to supplement their existing skills. We should allow them to be experts in what they were experts in and remember that what they had hired us to be experts in.<\/p><p id=\"cda5\" class=\"kp kq be cg kr b ks lq ku lr kw ls ky lt la lu lc\"><strong class=\"kr lv\">Who said it:<\/strong> Mikey Dickerson, Administrator of USDS<\/p><p id=\"f68f\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">Context:<\/strong> One of a few meetings I had with Mikey in the basement of USDS\u2019s headquarters on Jackson Place just in front of the White House. I don\u2019t remember exactly what we were talking about, but I remember how this comment felt like a shot to the heart. Mikey had a habit of cutting to the quick like that. For many people it was unnerving, a point of contention and an endless source of misunderstandings and hurt feelings, but you got used to it after a while.<\/p><p id=\"1dab\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">I hadn\u2019t thought I was making decisions based on imposture syndrome, but I did feel completely out of place working alongside some of the people who worked for USDS at the time. Engineers who had gotten in at the ground floor of some of the largest and most prestigious tech companies today \u2014 Google, Facebook, Twitter, Netflix, Amazon. People who had grown those companies to their current size. How could I NOT be overwhelmed?<\/p><p id=\"5a94\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">I don\u2019t think Mikey meant to be profound. He was just stating the truth as he saw it: USDS understood that I was not the greatest software engineer in the world. They knew where I was in my career, what my skill set was and where I was weak. They had hired me because I seemed to know how to get technical things done in a sustainable fashion in government. That was the skill set they were looking for me to bring to the table.<\/p><p id=\"e0fa\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">In the same conversation Mikey went on to tell me that \u201cI would have paid you your whole year\u2019s salary just based on your first three months of work. Even if you had done nothing else, I would have happily just signed the checks.\u201d<\/p><p id=\"f06f\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">After that I stopped getting so concerned about all these great engineers realizing I didn\u2019t belong there and started to focus on learning as much as I could from them.<\/p><p id=\"35a0\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">How I changed my approach:<\/strong> I think about Leslie\u2019s comment a lot when I\u2019m working with engineering teams. Too often engineers get so wrapped up in wanting to prove they\u2019re good enough they railroad colleagues with complimentary skillsets without realizing it. Software engineers suddenly act like they can be their own product managers, designers, copy editors, and sales people. It isn\u2019t always clear to professionals in other disciplines that this need to be a one person team is how engineers look for acceptance and respect.<\/p><p id=\"2153\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">When I find myself itching to interrupt someone with my thoughts about a topic I try to ask myself \u201cwhat am I being asked to be an expert in here?\u201d Often I realize that in my enthusiasm to show my casual knowledge I\u2019m about to correct someone that who has devoted a considerable amount of time and effort into developing their expertise in the topic. I never regret keeping my mouth shut and letting them speak.<\/p><p id=\"bb63\" class=\"kp kq be cg kr b ks lq ku lr kw ls ky lt la lu lc\"><strong class=\"kr lv\">Who said it:<\/strong> A group therapy facilitator at NAMI<\/p><p id=\"88d4\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">Context<\/strong>: The entire series of events that lead me to NAMI\u2019s friends and family group session in the summer of 2014 it too long and way too personal for this blog post, but I did end up there \u2014 $14,000 in debt, utterly heartbroken. I thought about this statement a lot back then. Later on when life returned to normal I realized that those same troublesome patterns of behavior I had had to confront lurk in much more banal situations too. The extremes only made them obvious, not unique.<\/p><p id=\"77d8\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">How I changed my approach:<\/strong> A big part of what I do as an engineering manager is stopping truly brilliant people from executing on plans that begin with the words \u201cI can just do this myself in a weekend.\u201d<\/p><p id=\"fc3e\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">If you\u2019re doing things for colleagues and denying them the ability to either do it themselves or participate in the process, you\u2019re not helping, you\u2019re making them dependent on you. Even if the end result is much faster doing it yourself.<\/p><p id=\"8a71\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Not long ago we had a Staff Engineer on another team reach out to my team and offer to migrate us to a new system for deploys. He had just done it for one of his team\u2019s services and was confident he could finish it up without any help from our team. He highlighted this as an advantage. We didn\u2019t need to reprioritize. He wouldn\u2019t need any effort from us at all! When I said no I got push back from my team. It sounded like a great deal. The Staff Engineer was so talented, why would I turn down the help?<\/p><p id=\"f3a5\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">I pointed out that if someone did it for us, then we wouldn\u2019t know how the new system worked and that might cause problems long term that would invalidate the time saved by accepting the offer. Still my team was not convinced, so we compromised and requested a meeting to go through the steps to move to the new system. If it was as easy as it sounded we would accept his offer.<\/p><p id=\"8619\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Turns out there were several requirements specific to our service that weren\u2019t available yet in the new system. Even worse, some of these missing pieces had no clear resolution. The Staff Engineer would have gotten stuck and either had to pull my team into a migration we weren\u2019t ready to do yet, abandon it halfway through, or make critical architecture decisions for us.<\/p><p id=\"86b8\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Doing things for people is rarely as helpful as it seems like it should be. If they don\u2019t understand what you built, you\u2019ve made things worse. If they don\u2019t know how to maintain it, you\u2019ve made things worse. If you didn\u2019t know enough about their requirements to get implementation correct, you\u2019ve made things worse. If they don\u2019t care about using or maintaining the thing you\u2019ve built because they didn\u2019t build it and have no sense of ownership or obligation to it, you\u2019ve made things worse.<\/p><p id=\"5194\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">It\u2019s almost always better to help by supplementing existing efforts rather than taking some work away and bringing back a solution.<\/p><p id=\"3603\" class=\"kp kq be cg kr b ks lq ku lr kw ls ky lt la lu lc\"><strong class=\"kr lv\">Who said it: <\/strong>Sifu Jesse Teasley<\/p><p id=\"7ba2\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">Context<\/strong>: I met Sifu when I was about 16 years old. He had dreadlocks and taught me how to fight with a sword. He also introduced me to Chinese philosophy, specifically: Taoism.<\/p><p id=\"cbc7\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">How I changed my approach:<\/strong> One of the core concepts of Taoism is the harmony of opposites. Basically Taoism acknowledges that the human mind loves binaries: white -vs- black, right -vs- left, soft -vs- hard, good -vs- evil. This is the way we are wired to think, but it\u2019s also inaccurate. Nature does not honor such strict distinctions. To the point that their usefulness at all is debatable. They make complex thought easier by eliminating some aspects of variation, but they also make us vulnerable to big mistakes when the pattern doesn\u2019t hold.<\/p><p id=\"0b17\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Taoists understand that we\u2019re inclined to over-optimize based on these false dichotomies. If softness is a desirable feature in given context we try to eliminate all hardness. If white is preferable, we scrub away the black. These states of purity are not just unattainable, they are built on a dichotomy that is not real. This insight went on to influence Buddhism, particularly what would ultimately become Zen Buddhism.<\/p><p id=\"9e85\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Taoists believe that the path to the truth is in harmonizing the dichotomy. Recognizing that humans are unlikely to be able to abandon binaries altogether and simply putting a bit of black in the white, and a bit of white in the black. That\u2019s why the symbol of Taoism is the yin-yang.<\/p><p id=\"27bd\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">I liked listening to Sifu because he framed this stuff in a very pragmatic way which made it less spiritual and more applicable to me. Another thing he used to say a lot was \u201cwhen you punch someone you need to pull your arm back, before you launch it forward. If you don\u2019t your hit will be weak.\u201d It was a variation on the same thing: we are stronger by considering the opposite first.<\/p><p id=\"25d2\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">We all know we should challenge assumptions, but that calls on us to realize we are making an assumption in the first place, which is not as easy. The Taoist advice of not moving in any direction before looking for and considering the opposite path bypasses this problem. Everything is an assumption when you are doomed to see in binaries.<\/p><p id=\"bbb0\" class=\"kp kq be cg kr b ks lq ku lr kw ls ky lt la lu lc\"><strong class=\"kr lv\">Who said it: <\/strong>Unknown<\/p><p id=\"3b5b\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">Context<\/strong>: I always associate this statement with an older woman whose charming pre-war Bronx apartment I ended up at for a Winter Solstice party as the college student, but that person was a stranger and it seems unlikely that we talked that much. I have tried to match up this advice against other possible sources from my past lives to no avail. It\u2019s a false memory that has stuck with me throughout my entire adult life \ud83e\udd37\u200d\u2640\ufe0f<\/p><p id=\"9960\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\"><strong class=\"kr lv\">How I changed my approach: <\/strong>On a personal level it gave me permission to take time when I needed time. Why should I feel guilty about leaving the office to go on a walk? Thinking is also work.<\/p><p id=\"589b\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">But it also influenced how I ran engineering teams. When I was in a traditional office environment I used to tell my people: If it\u2019s 2pm and you\u2019ve finished your work for the day and you have no meetings, just go home. You\u2019re not cheating the organization, you\u2019re putting that energy in the bank. You\u2019re going to have some on call rotation where you get paged at 3am. Or a hard week where we have to work late to get something out the door. These things happen and are impossible to predict exactly when. If you\u2019re done for the day go home, relax, spend some time with family. Put that time in the bank because we will certainly spent it later.<\/p><p id=\"47c8\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">It seems silly that people need to be told that, but many of us have been trained to believe that if we\u2019re not seen to be working, then management will assume we\u2019re slacking off.<\/p><p id=\"3085\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">I think of my role as a manager as part ad exec. My job is to sell the engineers that report up to me back to the organization that hired them in the first place. An organization\u2019s nature is to drift toward over optimization. This means it isn\u2019t enough for all our employees to be super productive, there has to be a clearly defined method and process that creates that productivity so that it can be reproduced to <em class=\"mq\">guarantee<\/em> productivity.<\/p><p id=\"a617\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Yet you could fill a whole library with research that shows this is not how effective teams are built or run. Effective teams need trust. That\u2019s not to say that frameworks for decision making or metrics tracking are not useful, they are <strong class=\"kr lv\">critical <\/strong>\u2014 but replacing trust with process is called bureaucracy. The trust must exist first in order for KPIs, OKRs, SLOs (or whatever acronym is selling books now) to work.<\/p><p id=\"d299\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">But trust also degrades naturally over time. Italian researchers Cristiano Castelfranchi and Rino Falcone have a model of trust in which it\u2019s <strong class=\"kr lv\"><em class=\"mq\">observability<\/em><\/strong> not success that is the key factor. Under their theory an entity that is silently successful can end up seen as less trustworthy than an entity that visibly fails. If we recover from failure quickly and efficiently, trust increases. Whereas when we succeed and no one notices we become more and more unknown and uncertain. This explains what is known as the <em class=\"mq\">service recovery paradox<\/em>,<em class=\"mq\"> <\/em>when consumers trust a service provider more after a failure than they did before the failure.<\/p><p id=\"d2b5\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Organizations have to constantly be reminded that they hired great people who know what they are doing. They know that in the beginning, but over time if the value of their employees is not observable then the trust degrades and bureaucracy becomes more and more attractive to leadership. The reason why people have to be told to take time off when they have unlimited vacation time, or go home early if they\u2019re finished at 2pm, or not to answer emails after midnight is because they have innate understanding that being <em class=\"mq\">observed<\/em> working is more valuable than <strong class=\"kr lv\"><em class=\"mq\">the results of their work<\/em><\/strong>.<\/p><p id=\"eac6\" class=\"kp kq be cg kr b ks kt ku kv kw kx ky kz la lb lc\">Obviously this creates a culture that leaves everyone worse off. The employees burn out. The organization\u2019s efficiency degrades. Poor performance decreases trust which increases bureaucracy\u2026 The turning point in my life was the day I realized to run great engineering teams I didn\u2019t need to be the best engineer in the world, I needed to get good at advertising my people and their stories up the chain of command. I needed to improve their observability so that we can keep bureaucracy at bay by maintaining a high level of trust.<\/p>","created_at":"2019-08-05T18:19:26+0000","updated_at":"2019-08-05T18:19:37+0000","published_at":"2019-08-05T02:58:49+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"medium.com","preview_picture":"https:\/\/miro.medium.com\/max\/1200\/1*0ESRdJgzIKVIKsvmzPEWnQ.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":989,"title":"No title found","url":"https:\/\/standardebooks.org\/","content":"<img alt=\"Ereaders with a Standard Ebook open.\" src=\"https:\/\/standardebooks.org\/images\/devices.png\" \/><p>Standard Ebooks is a volunteer driven, not-for-profit project that produces new editions of public domain ebooks that are lovingly formatted, open source, and free.<\/p><p>Ebook projects like <a href=\"https:\/\/www.gutenberg.org\">Project Gutenberg<\/a> transcribe ebooks and make them available for the widest number of reading devices. Standard Ebooks takes ebooks from sources like Project Gutenberg, formats and typesets them using a carefully designed and professional-grade style manual, fully proofreads and corrects them, and then builds them to create a new edition that takes advantage of state-of-the-art ereader and browser technology.<\/p>\n\t<p>Standard Ebooks aren\u2019t just a beautiful addition to your digital library\u2014they\u2019re a high quality standard to build your own ebooks on.<\/p><a class=\"button next\" href=\"https:\/\/standardebooks.org\/ebooks\/\">Browse our library of free ebooks<\/a>\n\t<section><section><h3>Modern &amp; consistent typography<\/h3><div><div><p>Other free ebooks don\u2019t put much effort into professional-quality typography: they use \"straight\" quotes instead of \u201ccurly\u201d quotes, they ignore details like em- and en-dashes, and they look more like early-90\u2019s web pages instead of actual books.<\/p><p>The Standard Ebooks project applies a rigorous and modern <a href=\"https:\/\/standardebooks.org\/contribute\/typography\">typography manual<\/a> when developing each and every ebook to ensure they meet a professional-grade and consistent typographical standard. Our ebooks look <em>good<\/em>.<\/p><\/div><figure class=\"stacked\"><img alt=\"An example of bad typography.\" class=\"bottom\" src=\"https:\/\/standardebooks.org\/images\/typography-bad.png\" \/><img alt=\"An example of Standard Ebooks typography.\" class=\"top\" src=\"https:\/\/standardebooks.org\/images\/typography-good.png\" \/><img alt=\"An arrow pointing from bad typography to good typography.\" class=\"arrow\" src=\"https:\/\/standardebooks.org\/images\/arrow-down.png\" \/><\/figure><\/div><\/section><section><h3>Full proofing with careful corrections<\/h3><div><div><p>Transcriptions from other sources are often filled with typos or suffer from issues like inconsistent spelling, missing accent marks, or missing punctuation. Submitting corrections to such sources can be difficult or impossible, so errors are rarely fixed.<\/p><p>At Standard Ebooks, we do a careful and complete readthrough of each ebook before releasing it, checking it against a scan of the original pages to fix as many typos as possible. Even if we <em>do<\/em> miss something, our ebooks are stored in the hugely popular Git source control system, allowing anyone to easily submit a correction.<\/p><\/div><figure><img alt=\"A text with proofreader's marks.\" src=\"https:\/\/standardebooks.org\/images\/proofreading.svg\" \/><\/figure><\/div><\/section><section><h3>Rich &amp; detailed metadata<\/h3><div><p>Our ebooks include complete, well-researched, and consistent metadata, including original, detailed book blurbs and links to encyclopedia sources. Perfect for machine processing or for extra-curious, technically-minded readers.<\/p><figure><img alt=\"An application dialog displaying ebook metadata.\" class=\"top\" src=\"https:\/\/standardebooks.org\/images\/metadata.png\" \/><\/figure><\/div><\/section><section><h3>State-of-the-art technology<\/h3><div><div><p>Each Standard Ebook takes full advantage of the latest ereader technology, including:<\/p><ul><li>\n\t\t\t\t\t\t\t<p>Hyphenation support,<\/p>\n\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t<li>\n\t\t\t\t\t\t\t<p>Popup footnotes,<\/p>\n\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t<li>\n\t\t\t\t\t\t\t<p>High-resolution and scalable vector graphics,<\/p>\n\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t<li>\n\t\t\t\t\t\t\t<p>Ereader-compatible tables of contents,<\/p>\n\t\t\t\t\t\t<\/li>\n\t\t\t\t\t<\/ul><p>and more. One of our goals is to ensure our ebooks stay up-to-date with the best reading experience technology can provide. Just because it\u2019s a classic doesn\u2019t mean it has to use old technology.<\/p><\/div><figure class=\"stacked\"><img alt=\"A screenshot of a popup endnote.\" class=\"bottom\" src=\"https:\/\/standardebooks.org\/images\/endnote.png\" \/><img alt=\"A screenshot of an ebook's table of contents.\" class=\"top\" src=\"https:\/\/standardebooks.org\/images\/toc.png\" \/><\/figure><\/div><\/section><section><h3>Quality covers<\/h3><div><div><p>Everyone knows a book is judged by its cover, but most free ebooks leave it to your ereader software to generate a drab default cover.<\/p><p>Standard Ebooks draws from a vast collection of public domain fine art to create attractive, unique, appropriate, and consistent covers for each of our ebooks.<\/p><\/div><figure><img alt=\"An ebookshelf featuring Standard Ebooks covers.\" src=\"https:\/\/standardebooks.org\/images\/covers.png\" \/><\/figure><\/div><\/section><section><h3>Clean code &amp; semantic markup<\/h3><div><div><p>Our strict coding standards allow technologists and ebook producers to use Standard Ebooks files as reliable, easy to read, and robust bases for their own work\u2014not to mention as models of what well-crafted ebook files look like. Common code patterns are repeated through different ebooks, so the code never surprises you.<\/p><p>Each ebook is also enhanced with careful standards-based semantic markup that opens the gateway for exciting new kinds of machine processing.<\/p><\/div><figure><img alt=\"A screenshot of the source code of a Standard Ebook file.\" src=\"https:\/\/standardebooks.org\/images\/code.png\" \/><\/figure><\/div><\/section><section><h3>Free, open-source, &amp; public domain<\/h3><div><div><p>We use the popular Git source control system to track each and every change made to our ebooks. Anyone can easily see a history of changes, or contribute their own changes with the click of a mouse.<\/p><p>And while all of the ebooks we feature and the cover art we draw from are <em>already<\/em> believed to be in the public domain in the US, Standard Ebooks releases all of the work we put in to each ebook into the public domain too. That makes each and every one of our ebook files not just free, but <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gratis_versus_libre\">libre<\/a> too\u2014because the world deserves more unrestricted culture.<\/p><\/div><figure class=\"oss\"><img alt=\"The Git SCM logo.\" src=\"https:\/\/standardebooks.org\/images\/git.svg\" \/><img alt=\"The no-copyright symbol.\" src=\"https:\/\/standardebooks.org\/images\/no-copyright.svg\" \/><img alt=\"The anti-DRM symbol.\" src=\"https:\/\/standardebooks.org\/images\/no-drm.svg\" \/><\/figure><\/div><\/section><\/section>","created_at":"2019-08-03T02:23:47+0000","updated_at":"2019-08-03T02:23:47+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"standardebooks.org","preview_picture":"https:\/\/standardebooks.org\/images\/logo.svg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":988,"title":"Sharon Broder, Clinical Social Work\/Therapist, Boston, MA, 02116 | Psychology Today","url":"https:\/\/www.psychologytoday.com\/us\/therapists\/bluecross-and-blueshield\/ma\/boston\/305339?sid=5d42385eb6b20&ref=1&tr=ResultsName","content":"<p>I work with professionals who are seeking support in navigating complicated relationships, pursuing that ever-elusive idea of work\/life balance, and managing challenging life situations at home and in the office. Because psychotherapy is my second career, I combine my clinical training as a therapist with my own work experience in both profit and non-profit organizations. I welcome the opportunity to understand your story in the service of shoring up your ability to cope, to change, and to have new experiences.<\/p><p>Getting ourselves into good working order starts with learning how to manage difficult feelings and understanding our authentic selves more deeply. We'll go beneath your surface presenting problems in order to discover and treat what's really at stake. We'll identify your values, your goals, and explore ways you might become the person you want to be.<\/p><p>Over time you'll develop strategies to feel less overwhelmed by painful feelings, modify behaviors that might be getting in your own way, and do things in life that bring you closer to your best self. The point will be for you to take our work with you after we say goodbye and to continue to reap the benefits through life's ups and downs.<\/p>","created_at":"2019-08-03T02:23:46+0000","updated_at":"2019-08-03T02:23:46+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.psychologytoday.com","preview_picture":"http:\/\/cdn4.sussexdirectories.com\/rms\/rms_photos\/sized\/39\/53\/305339-839338-3_320x400.jpg?pu=1485554617","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","org-mode"],"is_public":false,"id":987,"title":"DanielDe\/org-web","url":"https:\/\/github.com\/DanielDe\/org-web","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p>org-web lets you directly edit your org files online. It syncs with Dropbox and Google Drive and is optimized for mobile use. You can check out a <a href=\"https:\/\/www.youtube.com\/watch?v=ijNqZQxNgiA\" rel=\"nofollow\">video demo here<\/a>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/DanielDe\/org-web\/blob\/master\/screenshots\/overview.png\"><img src=\"https:\/\/github.com\/DanielDe\/org-web\/raw\/master\/screenshots\/overview.png\" alt=\".\/screenshots\/overview.png\" \/><\/a><\/p>\n<p>It was born out of my desire to be able to easily view and edit my org files from my phone. It\u2019s available for use <a href=\"https:\/\/org-web.org\" rel=\"nofollow\">here<\/a>.<\/p>\n<p>If you don\u2019t already keep your org files in Dropbox or Google Drive, I recommend symlinking them in:<\/p>\n<pre>ln -s ~\/Documents\/todo.org ~\/Dropbox\/todo.org\n<\/pre>\n<h2><a id=\"user-content-gitter-chat\" class=\"anchor\" aria-hidden=\"true\" href=\"#gitter-chat\"><\/a>Gitter chat<\/h2>\n<p>If you\u2019re interested in learning more about using or developing org-web, feel free to ask for help in <a href=\"https:\/\/gitter.im\/org-web\/Lobby?source=orgpage\" rel=\"nofollow\">this Gitter room<\/a>. It\u2019s quiet right now, but I still monitor it closely :)<\/p>\n<h2><a id=\"user-content-capture-params-and-siri-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#capture-params-and-siri-support\"><\/a>Capture params and Siri support<\/h2>\n<p>org-web supports a flexible mechanism for capturing using URL parameters. This mechanism integrates very nicely with the new <a href=\"https:\/\/support.apple.com\/guide\/shortcuts\/welcome\/ios\" rel=\"nofollow\">Siri Shortcuts<\/a> feature in iOS 12, allowing you to use Siri to execute capture templates.<\/p>\n<p>You can use <a href=\"https:\/\/www.icloud.com\/shortcuts\/4e51e8b748d14a50aa70ac6fb963f775\" rel=\"nofollow\">this sample Shortcut<\/a> to get started with this right away in iOS 12. Open the link on your iOS device and click \u201cGet Shortcut\u201d. Then open up the Shortcuts app and edit the template by following the directions in the comments. Then record a Siri trigger and you\u2019re good to go!<\/p>\n<p>Alternatively, you can take advantage of the URL parameters yourself to build your own custom capture mechanism. All three of the following parameters are required and must be URL encoded:<\/p>\n<ul><li><code>captureTemplateName<\/code>: the name of the capture template to use. This capture template must already exist in Settings &gt; Capture templates.<\/li>\n  <li><code>captureFile<\/code>: the path (for Dropbox) or id (for Google Drive) of the file in which to execute the capture template.<\/li>\n  <li><code>captureContent<\/code>: the content you\u2019d like to capture. This content will be placed at the cursor position if specified in the capture template (with <code>%?<\/code>), or at the end of the template if its not specified.<\/li>\n<\/ul><p>You can also specify additional custom variables for use in your templates. They should be in the format <code>captureVariable_&lt;your custom variable&gt;<\/code>, and should also be URL encoded. In your capture template they\u2019d show up as <code>%&lt;your custom variable&gt;<\/code>.<\/p>\n<p>Here\u2019s an example URL:<\/p>\n<p><a href=\"https:\/\/org-web.org?captureTemplateName=Grocery%20item&amp;captureContent=Lemon&amp;captureFile=\/org\/todo.org&amp;captureVariable_location=Home\" rel=\"nofollow\">https:\/\/org-web.org?captureTemplateName=Grocery%20item&amp;captureContent=Lemon&amp;captureFile=\/org\/todo.org&amp;captureVariable_location=Home<\/a><\/p>\n<h2><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h2>\n<p>org-web is built with React and Redux. It was bootstrapped with <a href=\"https:\/\/github.com\/facebook\/create-react-app\">Create React App<\/a>.<\/p>\n<p>To install the necessary packages, run:<\/p>\n<pre>yarn install\n<\/pre>\n<p>To run the app, run:<\/p>\n<pre>yarn start\n<\/pre>\n<p>To test against your actual Dropbox, you\u2019ll need to create a <code>.env<\/code> file by copying <code>.env.sample<\/code> to just <code>.env<\/code>. To test against your Google Drive you\u2019ll need to generate an API key as described on <a href=\"https:\/\/developers.google.com\/drive\/api\/v3\/quickstart\/js\" rel=\"nofollow\">this page<\/a>.<\/p>\n<p>Note that this will only work if you\u2019re running the app on <code>http:\/\/localhost:3000<\/code> because all redirect URIs must be specified ahead of time on the Dropbox and Google developer consoles.<\/p>\n<p>To make contributions to this repository, please install <a href=\"https:\/\/prettier.io\/\" rel=\"nofollow\">Prettier<\/a> and use it to format your code. This repository includes a <code>.prettierrc.json<\/code> with some configuration options that Prettier will use automatically. I also highly recommend installing <a href=\"https:\/\/github.com\/prettier\/prettier-emacs\">prettier-emacs<\/a> which will format your files on save, saving you the trouble of manually running Prettier.<\/p>\n<h2><a id=\"user-content-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#thanks\"><\/a>Thanks<\/h2>\n<p>org-web is built on the shoulders of giants. These are some packages I rely on that I think are absolutely fantastic:<\/p>\n<\/article>","created_at":"2019-07-29T13:35:44+0000","updated_at":"2019-07-29T13:35:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/595798?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["rpg","dnd"],"is_public":false,"id":986,"title":"Fronts","url":"https:\/\/book.dwgazetteer.com\/fronts.html","content":"<p>Fronts are secret tomes of GM knowledge. Each is a collection of linked dangers\u2014threats to the characters specifically and to the people, places, and things the characters care about. It also includes one or more impending dooms, the horrible things that will happen without the characters\u2019 intervention. \u201cFronts\u201d comes, of course, from \u201cfighting on two fronts\u201d which is just where you want the characters to be\u2014surrounded by threats, danger and adventure.<\/p><p>Fronts are built outside of active play. They\u2019re the solo fun that you get to have between games\u2014rubbing your hands and cackling evilly to yourself as you craft the foes with which to challenge your PCs. You may tweak or adjust your fronts during play (who knows when inspiration will strike?) but the meat of them comes from preparation between sessions.<\/p>\n<p>Fronts are designed to help you organize your thoughts on what opposes the players. They\u2019re here to contain your notes, ideas, and plans for these opposing forces. When you\u2019re in a bind your fronts are where you\u2019re going to turn and say, \u201cOh, so <em>that\u2019s<\/em> what I should do.\u201d Consider them an organizational tool, as inspiration for present and future mayhem.<\/p>\n<p>When you\u2019re building fronts, think about all the creepy dungeon denizens, the rampaging hordes and ancient cults that you\u2019d like to see in your game. Think in broad strokes at first and then, as you build dangers into your fronts, you\u2019ll be able to narrow those ideas down. When you write your campaign front, think about session-to-session trends. When you write your adventure fronts, think about what\u2019s important right here and right now. When you\u2019re done writing a few fronts you\u2019ll be equipped with all the tools you\u2019ll need to challenge your players and ready to run Dungeon World.<\/p>\n<h3>When to Make Fronts<\/h3>\n<p>You\u2019ll make your campaign front and first adventure fronts <em>after<\/em> your first session. Your campaign front may not be complete when you first make it\u2014that\u2019s great! Just like blanks on a map, unknown parts of your campaign front are opportunities for future creativity.<\/p>\n<p>After that first session you\u2019ll also make some adventure fronts. One or two is usually a good number. If you find yourself with more adventure fronts consider leaving some possible fronts as just notes for now.<\/p>\n<h3>Campaign and Adventure Fronts<\/h3>\n<p>At their core, all fronts contain the same components. They sort and gather your dangers into easy-to-use clusters. There are, however, two different <em>kinds<\/em> of fronts available to you. On the session-to-session level there are  your adventure fronts. These fronts will see use for a few sessions each. They\u2019re tied to one problem and will be dealt with or cast aside as the characters wander the dungeon or uncover the plot at hand. Think of them as episodic content: \u201cToday, on Dungeon World\u2026\u201d<\/p>\n<p>Tying your adventure fronts together is your campaign front. While the adventure fronts will contain immediate dangers\u2014the orcs in Hargrosh Pass, say\u2014the campaign front contains the Dark God Grishkar who drives the orcs to their pillaging. The campaign front is the unifying element that spans all the sessions of your Dungeon World game. It will have slower-burning portents but they\u2019ll be bigger in scope and have a deeper impact on the world. Most importantly they\u2019ll be scarier if they\u2019re allowed to resolve.<\/p>\n<p>When a danger from an adventure front goes without resolution you\u2019ll have to make a decision. If the danger is something you like and feel has a place in the larger world of your game don\u2019t hesitate to move it to the campaign front. You\u2019re able to make smaller dangers that went unresolved into bigger dangers some day later on. You can move dangers from the campaign fronts to an adventure front if they\u2019ve come to bear, too.<\/p>\n<h2>Creating Fronts<\/h2>\n<p>Here\u2019s how a front comes together:<\/p>\n<ul><li>Choose campaign front or adventure front<\/li>\n<li>Create 2-3 dangers<\/li>\n<li>Choose an impending doom for each danger<\/li>\n<li>Add grim portents (1-3 for an adventure front, 3-5 for the campaign front)<\/li>\n<li>Write 1\u20133 stakes questions<\/li>\n<li>List the general cast of the front<\/li><\/ul><h2>Creating Dangers<\/h2>\n<p>Not every element of your game will warrant a danger\u2014traps, some roving monsters, and other bits of ephemera may just be there to add context but aren\u2019t important enough to warrant inclusion. That\u2019s okay. Fronts are here to keep you apprised of the bigger picture. Dangers are divided into a handful of categories, each with its own name and <strong>impulse<\/strong>.<\/p>\n<p>Every danger has a crucial motivation that drives it called its impulse. The impulse exists to help you understand that danger. What pushes it to fulfill its impending doom? Impulses can help you translate the danger into action.<\/p>\n<p>When creating dangers for your front, think about how each one interacts as a facet of the front as a whole. Keep in mind the people, places, and things that might be a part of the threat to the world that the front represents. How does each danger contribute to the front?<\/p>\n<p class=\"Example\">Let\u2019s say we have an idea for a front\u2014an ancient portal has been discovered in the icy north. We\u2019ll call our front \u201cThe Opening of the White Gate.\u201d<\/p>\n<p>The easiest place to start is with people and monsters. Cultists, ogre chieftains, demonic overlords, and the like are all excellent dangers. These are the creatures that have risen above mere monster status to become serious threats on their own. Groups of monsters can be dangers too\u2014goblin tribes or a rampaging centaur khanate, for example.<\/p>\n<p class=\"Example\">For the front we\u2019re creating, we can pick a few different groups or people who might be interested in the gate. The College of Arcanists, perhaps. There\u2019s a golem, too, we\u2019ve decided, that protects the forgotten portal. The golem is just an obstacle, so we won\u2019t make him a danger.<\/p>\n<p>Thinking more broadly, less obvious elements of the world can be dangers. Blasted landscapes, intelligent magical items, ancient spells woven into the fabric of time. These things fulfill the same purposes as a mad necromancer\u2014they\u2019re part of the front, a danger to the world.<\/p>\n<p class=\"Example\">For our front, we\u2019ll add the gate itself as a danger.<\/p>\n<p>Lastly, if we think ahead, we can include some overarching dangers. The sorts of things that are in play outside the realm of the obvious\u2014godly patrons, hidden conspiracies and cursed prophecies waiting to be fulfilled.<\/p>\n<p class=\"Example\">Perhaps the White Gate was carved in the ancient past, hidden by a race of angels until the Day of Judgement. We\u2019ll add the Argent Seraphim to our front as a new danger.<\/p>\n<p>There\u2019s always more dangers you could add to a front, but limit yourself to 3 at most and leave room for discovery. Like a map, blank spaces can always be filled in later. Leaving room for player contribution and future inspiration means you\u2019ll have freedom to alter the front and make it fit the game. Not every bad thing that could happen deserves to be made into a danger. If you\u2019re uncertain, think about it this way: dangers can always get worse.<\/p>\n<p class=\"Example\">A barbarian tribe near the gate, the frozen tundra itself, a band of rival adventurers; all these things could be dangerous elements of the game but they\u2019re not important enough just yet to deserve to be dangers.<\/p>\n<p>Creating dangers is a way to slice up your overall front concept into smaller, easier to manage pieces. Dangers are tools for adding detail to the right parts of the front and for making the front easier to manage in the long run.<\/p>\n<p>Once you\u2019ve named and added a danger to the front you need to choose a type for that danger from the list below. Alternately, you can use the list of types to inspire dangers: with your front in mind, peruse the list and pick one or two that fit.<\/p>\n<p class=\"Example\">For our three dangers (The College of Arcanists, The White Gate and the Argent Seraphim) we\u2019ve selected Cabal, Dark Portal and Choir of Angels, respectively.<\/p>\n<h2>Types of Dangers<\/h2>\n<ul><li>Ambitious Organizations<\/li>\n<li>Planar Forces<\/li>\n<li>Arcane Enemies<\/li>\n<li>Hordes<\/li>\n<li>Cursed Places<\/li><\/ul><h3>Ambitious Organizations<\/h3>\n<ul><li>Misguided Good (<em>impulse: to do what is \u201cright\u201d no matter the cost<\/em>)<\/li>\n<li>Thieves Guild (<em>impulse: to take by subterfuge<\/em>)<\/li>\n<li>Cult (<em>impulse: to infest from within<\/em>)<\/li>\n<li>Religious Organization (<em>impulse: to establish and follow doctrine<\/em>)<\/li>\n<li>Corrupt Government (<em>impulse: to maintain the status quo<\/em>)<\/li>\n<li>Cabal (<em>impulse: to absorb those in power, to grow<\/em>)<\/li><\/ul><h4>GM Moves for Ambitious Organizations<\/h4>\n<ul><li>Attack someone by stealthy means (kidnapping, etc.)<\/li>\n<li>Attack someone directly (with a gang or single assailant)<\/li>\n<li>Absorb or buy out someone important (an ally, perhaps)<\/li>\n<li>Influence a powerful institution (change a law, manipulate doctrine)<\/li>\n<li>Establish a new rule (within the organization)<\/li>\n<li>Claim territory or resources<\/li>\n<li>Negotiate a deal<\/li>\n<li>Observe a potential foe in great detail<\/li><\/ul><h3>Planar Forces<\/h3>\n<ul><li>God (<em>impulse: to gather worshippers<\/em>)<\/li>\n<li>Demon Prince (<em>impulse: to open the gates of Hell<\/em>)<\/li>\n<li>Elemental Lord (<em>impulse: to tear down creation to its component parts<\/em>)<\/li>\n<li>Force of Chaos (<em>impulse: to destroy all semblance of order<\/em>)<\/li>\n<li>Choir of Angels (<em>impulse: to pass judgement<\/em>)<\/li>\n<li>Construct of Law (<em>impulse: to eliminate perceived disorder<\/em>)<\/li><\/ul><h4>GM Moves for Planar Forces<\/h4>\n<ul><li>Turn an organization (corrupt or infiltrate with influence)<\/li>\n<li>Give dreams of prophecy<\/li>\n<li>Lay a Curse on a foe<\/li>\n<li>Extract a promise in exchange for a boon<\/li>\n<li>Attack indirectly, through intermediaries<\/li>\n<li>Rarely, when the stars are right, attack directly<\/li>\n<li>Foster rivalries with other, similar powers<\/li>\n<li>Expose someone to a Truth, wanted or otherwise<\/li><\/ul><h3>Arcane Enemies<\/h3>\n<ul><li>Lord of the Undead (<em>impulse: to seek true immortality<\/em>)<\/li>\n<li>Power-mad Wizard (<em>impulse: to seek magical power<\/em>)<\/li>\n<li>Sentient Artifact (<em>impulse: to find a worthy wielder<\/em>)<\/li>\n<li>Ancient Curse (<em>impulse: to ensnare<\/em>)<\/li>\n<li>Chosen One (<em>impulse: to fulfill or resent their destiny<\/em>)<\/li>\n<li>Dragon (<em>impulse: to hoard gold and jewels, to protect the clutch<\/em>)<\/li><\/ul><h4>GM Moves for Arcane Enemies<\/h4>\n<ul><li>Learn forbidden knowledge<\/li>\n<li>Cast a spell over time and space<\/li>\n<li>Attack a foe with magic, directly or otherwise<\/li>\n<li>Spy on someone with a scrying spell<\/li>\n<li>Recruit a follower or toady<\/li>\n<li>Tempt someone with promises<\/li>\n<li>Demand a sacrifice<\/li><\/ul><h3>Hordes<\/h3>\n<ul><li>Wandering Barbarians (<em>impulse: to grow strong, to drive their enemies before them<\/em>)<\/li>\n<li>Humanoid Vermin (<em>impulse: to breed, to multiply and consume<\/em>)<\/li>\n<li>Underground Dwellers (<em>impulse: to defend the complex from outsiders<\/em>)<\/li>\n<li>Plague of the Undead (<em>impulse: to spread<\/em>)<\/li><\/ul><h4>GM Moves for Hordes<\/h4>\n<ul><li>Assault a bastion of civilization<\/li>\n<li>Embrace internal chaos<\/li>\n<li>Change direction suddenly<\/li>\n<li>Overwhelm a weaker force<\/li>\n<li>Perform a show of dominance<\/li>\n<li>Abandon an old home, find a new one<\/li>\n<li>Grow in size by breeding or conquest<\/li>\n<li>Appoint a champion<\/li>\n<li>Declare war and act upon that declaration without hesitation or deliberation<\/li><\/ul><h3>Cursed Places<\/h3>\n<ul><li>Abandoned Tower (<em>impulse: to draw in the weak-willed<\/em>)<\/li>\n<li>Unholy Ground (<em>impulse: to spawn evil<\/em>)<\/li>\n<li>Elemental Vortex (<em>impulse: to grow, to tear apart reality<\/em>)<\/li>\n<li>Dark Portal (<em>impulse: to disgorge demons<\/em>)<\/li>\n<li>Shadowland (<em>impulse: to corrupt or consume the living<\/em>)<\/li>\n<li>Place of Power (<em>impulse: to be controlled or tamed<\/em>)<\/li><\/ul><h4>GM Moves for Cursed Places<\/h4>\n<ul><li>Vomit forth a lesser monster<\/li>\n<li>Spread to an adjacent place<\/li>\n<li>Lure someone in<\/li>\n<li>Grow in intensity or depth<\/li>\n<li>Leave a lingering effect on an inhabitant or visitor<\/li>\n<li>Hide something from sight<\/li>\n<li>Offer power<\/li>\n<li>Dampen magic or increase its effects<\/li>\n<li>Confuse or obfuscate truth or direction<\/li>\n<li>Corrupt a natural law<\/li><\/ul><h3>Description and Cast<\/h3>\n<p>Write up something short to remind you just what this danger is about, something to describe it in a nutshell. Don\u2019t worry about where it\u2019s going or what could happen\u2014grim portents and the impending doom will handle that for you; you\u2019ll get to those in a bit. If there are multiple people involved in the danger (an orc warlord and his clansmen, a hateful god and his servants) go ahead and give them names and a detail or two now. Leave yourself some space as you\u2019ll be adding to this section as you play.<\/p>\n<h3>Custom Moves<\/h3>\n<p>Sometimes a danger will suggest a move that isn\u2019t covered by any existing ones. You can write custom moves to fill the gaps or to add the right effects for the danger. They can be player moves or GM moves, as you see fit. Of course, if you\u2019re writing a player move, keep your hands off the dice and mind the basic structure of a move. A 10+ is a complete success, while a 7\u20139 is a partial success. On a miss, maybe the custom move does something specific, or maybe not\u2014maybe you just get to make a move or work towards fulfilling a grim portent. The formatting of these moves varies from move to move.<\/p>\n<p class=\"Example\">For the Opening of the White Gate, I just know some fool PC is going to end up in the light that spills from the gate, so I\u2019m writing a move to show what might occur.<\/p>\n<p class=\"Example\">When you <strong>stand in the presence of the Light From Beyond<\/strong>, roll+WIS: \u2022 On a 10+ you are judged worthy, the Argent Seraphim will grant you a vision or boon. \u2022 On a 7-9 you are under suspicion and see a vision of what dark fate might befall you if you do not correct your ways. \u2022 On a miss, thou art weighed in the balance and art found wanting.<\/p>\n<h3>Grim Portents<\/h3>\n<p>Grim portents are dark designs for what could happen if a danger goes unchecked. Think about what would happen if the danger existed in the world but the PCs didn\u2019t\u2014if all these awful things you\u2019ve conjured up had their run of the world. Scary, huh? The grim portents are your way to codify the plans and machinations of your dangers. A grim portent can be a single interesting event or a chain of steps. When you\u2019re not sure what to do next, push your danger towards resolving a grim portent.<\/p>\n<p>More often than not grim portents have a logical order. The orcs tear down the city only after the peace talks fail, for example. A simple front will progress from bad to worse to much worse in a clear path forward. Sometimes, grim portents are unconnected pathways to the impending doom. The early manifestations of danger might not all be related. It\u2019s up to you to decide how complex your front will be. Whenever a danger comes to pass, check the other dangers in the front. In a complex front, you may need to cross off or alter the grim portents. That\u2019s fine, you\u2019re allowed. Keep scale in mind, too. Grim portents don\u2019t all have to be world-shaking. They can simply represent a change in direction for a danger. Some new way for it to cause trouble in the world.<\/p>\n<p>Think of your grim portents as possible moves waiting in the wings. When the time is right, unleash them on the world.<\/p>\n<p class=\"Example\">I\u2019ve chosen a few grim portents for my new front.<\/p>\n<ul><li class=\"Exampleli\">The College sends an expedition to the Gate<\/li>\n<li class=\"Exampleli\">The Key is discovered<\/li>\n<li class=\"Exampleli\">The First Trumpet sounds<\/li>\n<li class=\"Exampleli\">A Champion is chosen<\/li>\n<li class=\"Exampleli\">The Second Trumpet sounds<\/li>\n<li class=\"Exampleli\">The Herald appears<\/li>\n<li class=\"Exampleli\">The Gate is Opened<\/li><\/ul><p>When a grim portent comes to pass, check it off\u2014the prophecy has come true! A grim portent that has come to pass might have ramifications for your other fronts, too. Have a quick look when your players aren\u2019t demanding your attention and feel free to make changes. One small grim portent may resound across the whole campaign in subtle ways.<\/p>\n<p>You can advance a grim portent descriptively or prescriptively. Descriptively means that you\u2019ve seen the change happen during play, so you mark it off. Maybe the players sided with the goblin tribes against their lizardman enemies\u2014now the goblins control the tunnels. Lo and behold, this was the next step in a grim portent. Prescriptive is when, due to a failed player move or a golden opportunity, you advance the grim portent as your hard move. That step comes to pass, show its effects and keep on asking, \u201cWhat do you do, now?\u201d<\/p>\n<h3>Impending Doom<\/h3>\n<p>At the end of every danger\u2019s path is an impending doom. This is the final toll of the bell that signals the danger\u2019s triumphant resolution. When a grim portent comes to pass the impending doom grows stronger, more apparent and present in the world. These are the very bad things that every danger, in some way, seeks to bring into effect. Choose one of the types of impending dooms and give it a concrete form in your front. These often change in play, as the characters meddle in the affairs of the world. Don\u2019t fret, you can change them later.<\/p>\n<ul><li>Tyranny (of the strong over the weak or the few over the many)<\/li>\n<li>Pestilence (the spread of sickness and disease, the end of wellness)<\/li>\n<li>Destruction (apocalypse, ruin and woe)<\/li>\n<li>Usurpation (the chain of order comes apart, someone rightful is displaced)<\/li>\n<li>Impoverishment (enslavement, the abandonment of goodness and right)<\/li>\n<li>Rampant Chaos (laws of reality, of society, or any order is dissolved)<\/li><\/ul><p>When all of the grim portents of a danger come to pass, the impending doom sets in. The danger is then resolved but the setting has changed in some meaningful way. This will almost certainly change the front at large as well. Making sure that these effects reverberate throughout the world is a big part of making them feel real.<\/p>\n<h3>Stakes<\/h3>\n<p>Your stakes questions are 1-3 questions about people, places, or groups that you\u2019re interested in. People include PCs and NPCs, your choice. Remember that your agenda includes \u201cPlay to find out what happens?\u201d Stakes are a way of reminding yourself what you want to find out.<\/p>\n<p>Stakes are concrete and clear. Don\u2019t write stakes about vague feelings or incremental changes. Stakes are about important changes that affect the PCs and the world. A good stakes question is one that, when it\u2019s resolved, means that things will never be the same again.<\/p>\n<p>The most important thing about stakes is that you find them interesting. Your stakes should be things that you genuinely want to know, but that you\u2019re also willing to leave to be resolved through play. Once you\u2019ve written it as a stake, it\u2019s out of your hands, you don\u2019t get to just make it up anymore. Now you have to play to find out.<\/p>\n<p>Playing to find out is one of the biggest rewards of playing Dungeon World. You\u2019ve written down something tied to events happening in the world that you want to find out about\u2014now you get to do just that.<\/p>\n<p>Once you have your stakes your front is ready to play.<\/p>\n<p class=\"Example\">My stakes questions include, as tailored to my group:<\/p>\n<ul><li class=\"Exampleli\">Who will be the champion?<\/li>\n<li class=\"Exampleli\">How will Lux respond to the Light From Beyond?<\/li>\n<li class=\"Exampleli\">Will the College be able to recruit Avon?<\/li><\/ul><h2>Resolving a Front<\/h2>\n<p>Often a front will be resolved in a simple and straightforward manner. A front representing a single dungeon may have its dangers killed, turned to good, or overcome by some act of heroism. In this case the front is dissolved and set aside. Maybe there are elements of the front\u2014dangers that go unresolved or leftover members of a danger that\u2019s been cleared\u2014that live on. Maybe they move to the campaign front as brand new dangers?<\/p>\n<p>The campaign front will need a bit more effort to resolve. It\u2019ll be working slowly and subtly as the course of the campaign rolls along. You won\u2019t introduce or resolve it all at once, but in pieces. The characters work towards defeating the various minions of the big bad that lives in your campaign front. In the end, though, you\u2019ll know that the campaign front is resolved when the Dark God is confronted or the undead plague claims the world and the heroes emerge bloodied but victorious or defeated and despairing. Campaign fronts take longer to deal with but in the end they\u2019re the most satisfying to resolve.<\/p>\n<p>When a front is resolved take some extra time to sit down and look at the aftermath. Did any grim portents come to pass? Even if a danger is stopped, if any grim portents are fulfilled, the world is changed, if only in subtle ways. Keep this in mind when you write your future fronts. Is there anyone who could be moved from the now-defeated front to somewhere else? Anyone get promoted or reduced in stature? The resolution of a front is an important event!<\/p>\n<p>When you resolve an adventure front usually that means the adventure itself has been resolved. This is a great time to take a break and look at your campaign front. Let it inspire your next adventure front. Write up a new adventure front or polish off one you\u2019ve been working on, draw a few maps to go with it and get ready for the next big thing.<\/p>\n<h2>Multiple Adventure Fronts<\/h2>\n<p>As you start your campaign you\u2019re likely to have a lightly detailed campaign front and one or two detailed adventure fronts. Characters may choose, part-way through an adventure, to pursue some other course. You might end up with a handful of partly-resolved adventure fronts. Not only is this okay, it\u2019s a great way to explore a world that feels alive and organic. Always remember, fronts continue along apace no matter whether the characters are there to see them or not. Think offscreen, especially where fronts are concerned.<\/p>\n<p>When running two adventure fronts at the same time they can be intertwined or independent. The anarchists corrupting the city from the inside are a different front from the orcs massing outside the walls, but they\u2019d both be in play at once. On the other hand one dungeon could have multiple fronts at play within its walls: the powers and effects of the cursed place itself and the warring humanoid tribes that inhabit it.<\/p>\n<p>A situation warrants multiple adventure fronts when there are multiple impending dooms, all equally potent but not necessarily related. The impending doom of the anarchists is chaos in the city, the impending doom of the orcs is its utter ruination. They are two separate fronts with their own dangers. They\u2019ll deal with each other, as well, so there\u2019s some room for the players choosing sides or attempting to turn the dangers of one front against the other.<\/p>\n<p>When dealing with multiple adventure fronts the players are likely to prioritize. The cult needs attention now, the orcs can wait, or vice versa. These decisions lead to the slow advancement of the neglected front, eventually causing more problems for the characters and leading to new adventures. This can get complex once you\u2019ve got three or four fronts in play. Take care not to get overwhelmed.<\/p>\n<h2>An Example Front: The Opening of the White Gate<\/h2>\n<h2>Dangers<\/h2>\n<h3>The College of Arcanists (Cabal)<\/h3>\n<p><strong>Impulse:<\/strong> to absorb those in power, to grow<\/p>\n<h4>Grim Portents<\/h4>\n<ul><li>The College sends an expedition to the Gate<\/li>\n<li>The Key is discovered<\/li>\n<li>The Gate\u2019s Power is harnessed<\/li>\n<li>The College seizes control<\/li><\/ul><p><strong>Impending Doom:<\/strong> Usurpation<\/p>\n<h3>The White Gate (Dark Portal)<\/h3>\n<p><strong>Impulse:<\/strong> to disgorge demons<\/p>\n<h4>Grim Portents<\/h4>\n<ul><li>The First Trumpet sounds<\/li>\n<li>The Second Trumpet sounds<\/li>\n<li>The Gate is opened<\/li><\/ul><p><strong>Impending Doom:<\/strong> Destruction<\/p>\n<h3>The Argent Seraphim (Choir of Angels)<\/h3>\n<p><strong>Impulse:<\/strong> to pass judgement<\/p>\n<h4>Grim Portents<\/h4>\n<ul><li>A Champion is chosen<\/li>\n<li>An organization of power is formed or co-opted<\/li>\n<li>The Herald appears<\/li>\n<li>Judgement is passed<\/li><\/ul><p><strong>Impending Doom:<\/strong> Tyranny<\/p>\n<h2>Description and Cast<\/h2>\n<p>An ancient gate, buried for aeons in the icy north. It opens into a realm of pure light, guarded by the Argent Seraphim. It was crafted only to be opened at Judgement Day, so that the Seraphim could come forth and purge the realm of men. It was recently uncovered by the College of Arcanists, who do not yet understand its terrible power.<\/p>\n<ul><li>Oren Balserus, Arcanist Supreme<\/li>\n<li>Hali\u2019el, voice of the Seraphim<\/li>\n<li>Drudge, a manservant<\/li><\/ul><h2>Custom Moves<\/h2>\n<p>When you <strong>stand in the presence of the Light From Beyond<\/strong>, roll+WIS. \u2022 On a 10+ you are judged worthy, the Argent Seraphim will grant you a vision or boon. \u2022 On a 7-9 you are under suspicion and see a vision of what dark fate might befall you if you do not correct your ways. \u2022 On a miss, thou art weighed in the balance and art found wanting.<\/p>\n<h2>Stakes<\/h2>\n<ul><li>Who will be the Champion?<\/li>\n<li>How will Lux respond to the holy light?<\/li>\n<li>Will the College be able to recruit Avon?<\/li><\/ul>","created_at":"2019-07-25T20:18:54+0000","updated_at":"2019-07-25T20:18:59+0000","annotations":[],"mimetype":"text\/html","reading_time":20,"domain_name":"book.dwgazetteer.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":985,"title":"Alpaca - Commission-Free API First Stock Brokerage","url":"https:\/\/alpaca.markets\/","content":"<section class=\"section api\"><div class=\"container\"><div class=\"row justify-content-center align-items-center\"><div class=\"col-12 col-lg-7\"><p class=\"section__desc\">With Alpaca Securities LLC\u2019s commission-free API, you can run algorithmic trading &amp; portfolio management bots and build your own mobile app with ease.<\/p><a href=\"https:\/\/docs.alpaca.markets\/api-documentation\" target=\"_blank\" class=\"section__link\">Learn More<\/a><p class=\"api__note\">\n                            Commission-Free trading means that there are no commission charges for Alpaca self-directed\n                            individual cash brokerage accounts that trade U.S. listed securities through an API. <a href=\"https:\/\/support.alpaca.markets\/hc\/en-us\/articles\/360010814292-Regulatory-Fees-\" target=\"_blank\">Relevant SEC and FINRA fees may apply.<\/a>\n                        <\/p><\/div><div class=\"col-12 col-lg-4\" data-aos=\"fade-left\" data-once=\"true\"><img class=\"api__img\" src=\"https:\/\/alpaca.markets\/img\/currency.png\" alt=\"Commission-Free Stock Trading\" \/><\/div><\/div><\/div><\/section><section class=\"section broker\"><div class=\"container\"><div class=\"row justify-content-center align-items-center\"><div data-aos=\"fade-right\" data-once=\"true\" class=\"col-12 col-lg-4 order-2 order-lg-1\"><img class=\"broker__img\" src=\"https:\/\/alpaca.markets\/img\/broker-img.svg\" alt=\"Alpaca is a Registered Securities Broker\" \/><\/div><div class=\"col-12 col-lg-7 order-1 order-lg-2\"><h3 class=\"title broker__title\">Registered Securities Broker<\/h3><p class=\"section__desc broker__desc\">Alpaca Securities LLC is a member of FInancial Industry Regulatory Authority, Inc. (<a href=\"http:\/\/www.finra.org\/\" target=\"_blank\">FINRA<\/a>). You can check the background of Alpaca Securities on <a href=\"https:\/\/brokercheck.finra.org\/firm\/summary\/288202\" target=\"_blank\">FINRA\u2019s BrokerCheck<\/a>. Alpaca Securities is also a member of SIPC - securities in your account are protected up to $500,000. For details, please visit <a href=\"https:\/\/www.sipc.org\/\" target=\"_blank\">www.sipc.org<\/a>.<\/p><\/div><\/div><\/div><\/section><section class=\"section try\"><div class=\"container\"><div class=\"row\"><div class=\"col-12\"><div class=\"try__grid\"><h3 class=\"title try__title\">Built For Makers<br \/>In Open source &amp; Community<\/h3><p class=\"try__desc\">We truly believe open source changes the financial systems and how we interact with them. Alpaca grows with our community and knowledges we build together.<\/p><p class=\"try__note\">Technology and services are offered by AlpacaDB, Inc.<br \/>Securities are offered through Alpaca Securities LLC - <a href=\"https:\/\/alpaca.markets\/disclosures\" target=\"_blank\">Important disclosures<\/a><\/p><\/div><\/div><\/div><\/div><\/section><section class=\"section testimonials\"><div class=\"container\"><div class=\"row\"><div class=\"col-12\"><h3 class=\"title testimonials__title\">What our Users are Saying<\/h3><div class=\"row justify-content-center\"><div class=\"col-10\"><div class=\"testimonials__grid\"><div class=\"testimonials__item\"><img class=\"testimonials__quote\" src=\"https:\/\/alpaca.markets\/img\/icons\/icon-quote.svg\" alt=\"quote\" \/><p class=\"testimonials__text\">Finally, a professional and easy to use API to\n                                            connect to the markets courtesy of Alpaca. Their team is the real deal. I\n                                            can't wait to see what's next.<\/p><img class=\"testimonials__quote\" src=\"https:\/\/alpaca.markets\/img\/icons\/icon-quote.svg\" alt=\"quote\" \/><p class=\"testimonials__name\">John S.<br \/>from Stamford, CT<\/p><\/div><div class=\"testimonials__item\"><img class=\"testimonials__quote\" src=\"https:\/\/alpaca.markets\/img\/icons\/icon-quote.svg\" alt=\"quote\" \/><p class=\"testimonials__text\">I love Alpaca and their API first approach to trading,\n                                             I now can add smarter rules for buying and selling stocks for free.<\/p><img class=\"testimonials__quote\" src=\"https:\/\/alpaca.markets\/img\/icons\/icon-quote.svg\" alt=\"quote\" \/><p class=\"testimonials__name\">Kiran H.<br \/>Software Engineer, San Francisco<\/p><\/div><div class=\"testimonials__item\"><img class=\"testimonials__quote\" src=\"https:\/\/alpaca.markets\/img\/icons\/icon-quote.svg\" alt=\"quote\" \/><p class=\"testimonials__text\">The Alpaca API is elegant, fast, and easy to understand and work with.<\/p><img class=\"testimonials__quote\" src=\"https:\/\/alpaca.markets\/img\/icons\/icon-quote.svg\" alt=\"quote\" \/><p class=\"testimonials__name\">Jesse C.<br \/>Software Engineer, Seattle<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"testimonials__note\"><div class=\"container\"><div class=\"justify-content-center row\"><div class=\"col-12 col-lg-10\"><p>These testimonials may not be representative of the experience of other customers and are not a guarantee of future performance or success.\n            <br \/>Technology and services are offered by AlpacaDB, Inc.<\/p><\/div><section class=\"section features\"><div class=\"container\"><div class=\"row\"><div class=\"col-12\"><div class=\"features__icon\"><p class=\"features__icon-name\">Free features. Forever.<\/p><\/div><div class=\"row\"><div class=\"col-12 col-sm-6 col-lg-3 features__col\"><div class=\"features__item\"><p class=\"features__item-name\">Real-timeMarket Data<\/p><\/div><\/div><div class=\"col-12 col-sm-6 col-lg-3 features__col\"><div class=\"features__item\"><p class=\"features__item-name\">REST API + WebSocket stream <\/p><\/div><\/div><div class=\"col-12 col-sm-6 col-lg-3 features__col\"><div class=\"features__item\"><p class=\"features__item-name\">Agile Management Dashboard<\/p><\/div><\/div><div class=\"col-12 col-sm-6 col-lg-3 features__col\"><div class=\"features__item\"><p class=\"features__item-name\">Unlimited Live Paper Trading<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/section><\/div><\/div><\/div><\/section>","created_at":"2019-07-25T11:13:29+0000","updated_at":"2019-07-25T11:13:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"alpaca.markets","preview_picture":"https:\/\/s3.amazonaws.com\/marketing.alpaca.markets\/social\/banners\/201811_paper-trading.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":984,"title":"Space Engine \u2013 the universe simulator","url":"http:\/\/spaceengine.org","content":"<div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-6\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><div class=\"wpb_text_column wpb_content_element\"><div class=\"wpb_wrapper\"><p>SpaceEngine is a realistic virtual Universe you can explore on your computer. You can travel from star to star, from galaxy to galaxy, landing on any planet, moon, or asteroid with the ability to explore its alien landscape. You can alter the speed of time and observe any celestial phenomena you please. All transitions are completely seamless, and this virtual universe has a size of billions of light-years across and contains trillions upon trillions of planetary systems. The procedural generation is based on real scientific knowledge, so SpaceEngine depicts the universe the way it is thought to be by modern science. Real celestial objects are also present if you want to visit them, including the planets and moons of our Solar system, thousands of nearby stars with newly discovered exoplanets, and thousands of galaxies that are currently known.<\/p><\/div><\/div><div class=\"vc_row wpb_row vc_inner section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><a href=\"#arrowdownload\" target=\"_self\" class=\"qbutton  big_large transparent qbutton_with_icon icon_right rotate_icon animate_button\"> Download SpaceEngineDownload SpaceEngine <\/a><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\">News<div class=\"latest_post_holder  boxes three_columns\"><ul class=\"post_list\"><li class=\"clearfix\"><div class=\"box_padding_border\"><div class=\"boxes_image\"><a href=\"http:\/\/spaceengine.org\/news\/blog190720\/\"><img width=\"800\" height=\"600\" src=\"http:\/\/spaceengine.org\/media\/2019\/07\/scr00000-800x600.png\" class=\"attachment-eltd-se-portfolio-landscape size-eltd-se-portfolio-landscape wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"latest_post\"><div class=\"latest_post_text\"><p><\/p><h5 class=\"latest_post_title\"><a href=\"http:\/\/spaceengine.org\/news\/blog190720\/\">50 years since Moon landing<\/a><\/h5><p class=\"excerpt\">Today is the 50th anniversary of humanity\u2019s greatest achievement \u2013 landing a crewed spaceship on another celestial body for the first time...<\/p><\/div><\/div><\/div><\/li><li class=\"clearfix\"><div class=\"box_padding_border\"><div class=\"boxes_image\"><a href=\"http:\/\/spaceengine.org\/news\/blog190703\/\"><img width=\"800\" height=\"600\" src=\"http:\/\/spaceengine.org\/media\/2019\/07\/scr01114-800x600.jpg\" class=\"attachment-eltd-se-portfolio-landscape size-eltd-se-portfolio-landscape wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"latest_post\"><div class=\"latest_post_text\"><p><\/p><h5 class=\"latest_post_title\"><a href=\"http:\/\/spaceengine.org\/news\/blog190703\/\">Update 0.990.36.1665 (beta)<\/a><\/h5><p class=\"excerpt\">Todays build 0.990.36.1665 is released in the beta branch, so if you want to test this, feel free to join the beta. Changes: Multithreaded saving of s...<\/p><\/div><\/div><\/div><\/li><li class=\"clearfix\"><div class=\"box_padding_border\"><div class=\"boxes_image\"><a href=\"http:\/\/spaceengine.org\/news\/blog190629\/\"><img width=\"800\" height=\"600\" src=\"http:\/\/spaceengine.org\/media\/2019\/06\/scr01109-800x600.jpg\" class=\"attachment-eltd-se-portfolio-landscape size-eltd-se-portfolio-landscape wp-post-image\" alt=\"\" \/><\/a><\/div><div class=\"latest_post\"><div class=\"latest_post_text\"><p><\/p><h5 class=\"latest_post_title\"><a href=\"http:\/\/spaceengine.org\/news\/blog190629\/\">Update 0.990.35.1660<\/a><\/h5><p class=\"excerpt\">Todays update \u2013 build 1660, it is merged with previous beta build 1655. Build 0.990.35.1660 \u2013 Automatic saturation of stellar illumination...<\/p><\/div><\/div><\/div><\/li><\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\">Capabilities<\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><p>EXPLORE THE ENTIRE UNIVERSE FORGET ABOUT SPACE AND TIME RESTRICTIONS BECOME A PIONEER WITH SPACEENGINE<\/p><div class=\"wpb_text_column wpb_content_element\"><p><\/p><h3>And much more<\/h3><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-6\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Movement made possible with free, spacecraft or aircraft mode<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">\"Select and fly\" autopilot to automatically go directly to the object<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Automatic binding of the observer to moving objects<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Automatic selection of optimum flight speed<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Built-in wiki system with descriptions and ability to extend<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Ability to import user addons: models, textures, catalogs<\/p><\/div><\/div><\/div><\/div><div class=\"wpb_column vc_column_container vc_col-sm-6\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">3D models of galaxies and nebulae with interstellar dust clouds<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Accurate planetary atmosphere models<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Controllable space ships<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Original music with context-dependent track switching<\/p><\/div><div class=\"eltd_icon_list left\"><p class=\"eltd_icon_list_text\">Localization in many languages, with the ability to add new ones<\/p><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"arrowdownload\" class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\">Download<\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-6\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><a href=\"https:\/\/store.steampowered.com\/app\/314650\/SpaceEngine\/\" target=\"_self\" class=\"qbutton  big_large transparent qbutton_with_icon icon_right rotate_icon animate_button\"> SpaceEngine 0.990 on SteamSpaceEngine 0.990 on Steam <\/a><\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\">Support the project<\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-6\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\"><div class=\"wpb_text_column wpb_content_element\"><div class=\"wpb_wrapper\"><p>If you like SpaceEngine, buy it on <a href=\"https:\/\/store.steampowered.com\/app\/314650\/SpaceEngine\/\" rel=\"noopener\" target=\"_blank\">Steam<\/a> or provide additional support via Paypal. This greatly assists SpaceEngine\u2019s development!<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"vc_row wpb_row section\"><div class=\"full_section_inner clearfix\"><div class=\"wpb_column vc_column_container vc_col-sm-12\"><div class=\"vc_column-inner\"><div class=\"wpb_wrapper\">System requirements<\/div><\/div><\/div><\/div><\/div>","created_at":"2019-07-25T01:15:46+0000","updated_at":"2019-07-25T01:15:46+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"spaceengine.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":983,"title":"Operating a Large, Distributed System in a Reliable Way: Practices I Learned","url":"https:\/\/blog.pragmaticengineer.com\/operating-a-high-scale-distributed-system\/","content":"<p>For the past few years, I've been building and operating a large distributed system: the <a href=\"https:\/\/eng.uber.com\/payments-platform\/\">payments system at Uber<\/a>. I've learned a lot about <a href=\"https:\/\/blog.pragmaticengineer.com\/distributed-architecture-concepts-i-have-learned-while-building-payments-systems\/\">distributed architecture concepts <\/a>during this time and seen first-hand how high-load and high-availability systems are challenging not just to build, but to operate as well. Building the system itself is a fun job. Planning how the system will handle 10x\/100x traffic increase, ensuring data is durable, regardless of hardware failures is intellectually rewarding. However, <strong>operating a large, distributed system has been an eye-opening experience <\/strong>for myself.<\/p><p>The larger a system, the more Murphy's law of \"what can go wrong, will go wrong\" applies. This is especially true with frequent deploys, lots of developers deploying code, multiple data centers involved, and the system being used globally by a large number of users. The past few years, I've experienced a variety of system failures, many of which surprised me. These went from predictable things like hardware faults or innocent bugs making it into production, all the way to network cables connecting data centers being cut or multiple cascading failures occurring at the same time. I've gone through dozens of outages, where parts of the system were not working correctly, resulting in large business impact.<\/p><p><strong>This post is the collection of the practices I've found useful to reliably operate a large system <\/strong>at Uber, while working here. My experience is not unique - people working on similar sized systems go through a similar journey. I've talked with engineers at Google, Facebook, and Netflix, who shared similar experiences and solutions. Many of the ideas and processes listed here should apply to systems of similar scale, regardless of running on own data centers (like Uber mostly does) or on the cloud (where Uber <a href=\"https:\/\/aws.amazon.com\/blogs\/startups\/how-uber-survives-its-busiest-nights-of-the-year\/\">sometimes scales to<\/a>). However, the practices might be an overkill for smaller or less mission-critical systems.<\/p><p>There's much ground to cover - I'll go through the following topics:<\/p><ul><li><a href=\"#monitoring\">Monitoring<\/a><\/li><li><a href=\"#oncall-anomaly-detection-and-alerting\">Oncall, Anomaly Detection &amp; Alerting<\/a><\/li><li><a href=\"#outages-incident-management-processes\">Outages &amp; Incident Management Processes<\/a><\/li><li><a href=\"#postmortems-incident-reviews-a-culture-of-ongoing-improvements\">Postmortems, Incident Reviews &amp; a Culture of Ongoing Improvements<\/a><\/li><li><a href=\"#failover-drills-planned-downtime-capacity-planning-blackbox-testing\">Failover Drills, Capacity Planning &amp; Blackbox Testing<\/a><\/li><li><a href=\"#slos-slas-reporting-on-them\">SLOs, SLAs &amp; Reporting on Them<\/a><\/li><li><a href=\"#sre-as-an-independent-team\">SRE as an Independent Team<\/a><\/li><li><a href=\"#reliability-as-an-ongoing-investment\">Reliability as an Ongoing Investment<\/a><\/li><li><a href=\"#further-recommended-reading\">Further Recommended Reading<\/a><\/li><\/ul><h2 id=\"monitoring\">Monitoring<\/h2><p>To know if a system is healthy, we need to answer the question \"Is my system working correctly\"? To do so, it is vital to collect data on critical parts of the system. With distributed systems that run multiple services, on multiple machines and data centers, it can be difficult to decide what key things <em>really<\/em> need to be monitored.<\/p><p><strong>Infrastructure health monitoring.<\/strong> If one or more machines\/virtual machines are overloaded, parts of the distributed system can degrade. The health stats for machines a service operates on - their CPU utilization, memory usage - are basics that are worth monitoring. There are platforms that take care of this monitoring and auto-scaling instances out of the box. At Uber, \u00a0we have a great <a href=\"https:\/\/eng.uber.com\/tag\/core-infrastructure\/\">core infrastructure team<\/a> providing infra level monitoring and alerting out of the box. Regardless of how this implemented, being aware when things are in the red for instances or the infrastructure of the service is must-have information. <\/p><p><strong>Service health monitoring: traffic, errors, latency<\/strong>. Answering the question \"is this backend service healthy?\" is a pretty common one. Observing things like how much traffic is hitting the endpoint, the error rate, and endpoint latency all provide valuable information on service health. I prefer to have dashboards on all of these. When building a new service, by using the right HTTP response mappings and monitoring the relevant codes can tell a lot about the system. So by ensuring 4XX mappings are returned on client errors, and 5xx codes for server errors, this kind of monitoring is easy to build and easy to interpret.<\/p><p>Monitoring latency deserves one more thought. For production services, the goal is for most end users to have a good experience. Turns out, measuring the average latency is not a great metric to do so because this average can hide a small percentage of high latency requests. Measuring the p95, the p99 or p999 - latency experienced by the 95th, 99th or 99.9th percentile of requests - is a lot better metric. These numbers help answer questions like \"How fast will the request be for 99% of people?\" (p99) or \"At least how slow is the request for one in 1,000 people?\" (p999) . For those more interested in this topic, this <a href=\"https:\/\/igor.io\/latency\/\">latencies primer article<\/a> is good further read.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/blog.pragmaticengineer.com\/content\/images\/2019\/06\/Screenshot-2019-05-29-at-11.47.04.png\" class=\"kg-image\" alt=\"image\" \/><figcaption>Average, p95 and p99 latency visualized. Note how though the average latency for this endpoint is under 1s, 1% of requests take 2 seconds or more to complete - something the average measurement would have masked.<\/figcaption><\/figure><p>There is a lot more depth around monitoring and observability. Two resources worth reading more are Google's <a href=\"https:\/\/landing.google.com\/sre\/sre-book\/\">SRE book<\/a> and the section on the <a href=\"https:\/\/landing.google.com\/sre\/sre-book\/chapters\/monitoring-distributed-systems\/#xref_monitoring_golden-signals\">four golden signals<\/a> of distributed systems monitoring. They suggest if you can only measure four metrics of your user-facing system, focus on traffic, errors, latency, and saturation. The short read is the <a href=\"https:\/\/www.oreilly.com\/library\/view\/distributed-systems-observability\/9781492033431\/\">Distributed Systems Observability<\/a> e-book from <a href=\"https:\/\/twitter.com\/copyconstruct\">Cindy Sridharan<\/a>, which touches on other useful tools like event logs, metrics and tracing best practices.<\/p><p><strong>Business metrics monitoring.<\/strong> While monitoring services tells us a lot about how correctly the service seems to work, it tells nothing about whether it works as intended, allowing for \"business as usual.\" In the case of the payments system, a key question is, \"can people take trips using a specific payment method?\". Identifying business events that the service enables and monitoring these business events is one of the most important monitoring steps.<\/p><p>Business metrics monitoring was something my team built after having been bruised by outages that we had no way to detect otherwise. All our services looked like they were operating normally, yet key functionality was down across services. This kind of monitoring was quite bespoke to our organization and domain. As such, we've had to put a lot of thought and effort into customizing this type of monitoring for ourselves, building on the <a href=\"https:\/\/eng.uber.com\/observability-at-scale\/\">observability tech stack at Uber<\/a>.<\/p><h2 id=\"oncall-anomaly-detection-and-alerting\">Oncall, Anomaly Detection, and Alerting<\/h2><p>Monitoring is a great tool for people to inspect the current state of the system. But it's really just a stepping stone to automatically detect when things go wrong and raise alerts that people can take action on.<\/p><p>Oncall itself is a broad topic - Increment magazine did a great job covering many aspects in its <a href=\"https:\/\/increment.com\/on-call\/\">On-Call issue<\/a>. I strongly prefer thinking of oncall as a follow-up to the \"you build it, you own it\" mindset. Teams who build services own them, owning the oncall as well. My team owned oncall for the payments services that we built. So whenever an alert fires, the oncall engineer would respond and look into the details. But how do we get from monitoring to alerts?<\/p><p><strong>Detecting anomalies from the monitoring data is a tough challenge<\/strong> and an area where machine learning can shine. There are plenty of third-party services that offer anomaly detection. Again lucky for my team, we had an in-house machine learning team to work with, who tailored their solutions to Uber use-cases. The \u00a0NYC-based Observability team wrote a helpful article on how <a href=\"https:\/\/eng.uber.com\/anomaly-detection\/\">anomaly detection works at Uber<\/a>. From my team's point of view, we push our monitoring data to this team's pipeline and get alerts with the respective confidence levels. Then we decide if we should page an engineer or not.<\/p><p>When to fire an alert is an interesting question. Too few alerts can mean missing impactful outages. Too many can cause sleepless nights and burn people out. <strong>Tracking and categorizing alerts and measuring signal-to-noise is essential in tweaking the alerting system<\/strong>. Going through alerts and tagging whether they were actionable or not, then taking steps to reduce non-actionable alerts is a good step towards having <a href=\"https:\/\/increment.com\/on-call\/crafting-sustainable-on-call-rotations\/\">sustainable on-call rotations<\/a> in place.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/blog.pragmaticengineer.com\/content\/images\/2019\/06\/oncall.png\" class=\"kg-image\" alt=\"image\" \/><figcaption>Example of the <a href=\"https:\/\/eng.uber.com\/on-call-dashboard\/\">in-house oncall dashboard<\/a> used at Uber, built by the Uber Developer Experience team in Vilnius<\/figcaption><\/figure><p>The Uber developer tooling team based in Vilnius <a href=\"https:\/\/eng.uber.com\/on-call-dashboard\/\">builds neat on-call tooling<\/a> that we use to annotate alerts and visualize the oncall shifts. Our the team holds a weekly review of the last oncall shift, analyzing pain points and spending time to improve the oncall experience, week after week.<\/p><p>Imagine this: <em>you're the engineer oncall for the week. A pager alert wakes you up the middle of the night. You investigate if there's a production outage happening. Uh oh. Seems like part of the system is down. Now what?<\/em> Monitoring and alerting just got very real.<\/p><p>Outages might not be that big of a deal for small systems, where the oncall engineer can understand what is happening and why. They are usually quick to understand and easy to mitigate. For complex systems with multiple (micro)services, lots of engineers pushing code to production, it is challenging enough just to pinpoint where the underlying problem is happening. Having a few standard processes to help with this makes a huge difference.<\/p><p><strong>Runbooks<\/strong> that are attached to alerts, describing simple mitigation steps are the first line of defense. For teams that have good runbooks, it's rarely a problem if the engineer oncall does not know the system in-depth. Runbooks need to be kept up to date, updated, and re-worked with new type of mitigations, as they happen.<\/p><p><strong>Communicating outages across the organization<\/strong> become essential as soon as there are more than a few teams that deploy services. In the environment I work at, thousands of engineers deploy services they work on into production when they see fit, resulting in hundreds of deploys per hour. A seemingly unrelated deployment on one service might impact another one. Here, standardized outage broadcasting and communication channels make a big difference. I've had multiple cases when an alert was unlike anything I've seen before - and realizing that other people in other teams are seeing similarly odd things. Being in a central chat group for outages, we pinpointed the service causing the outage and mitigated the issue quickly. We did this much faster, as a group than any one of us would have done it by ourselves.<\/p><p><strong>Mitigation now, investigation tomorrow.<\/strong> In the middle of an outage, I often got the \"adrenalin rush\" of wanting to fix what went wrong. Often the root cause was a bad code deploy, with a glaring bug being present in the code change. In the past, I would have jumped to fix the bug, push the fix and close the outage over just rolling back the code change. However, fixing the root cause in the middle of an outage is a <em>terrible<\/em> idea. <strong>There is little to gain and a lot to lose with a forward fix<\/strong>. Because the new fix has to be done quickly, testing has to be done in production. Which is the recipe to introduce a second bug - or a second outage - on top of the existing one. I have seen outages blow up, just like this. Just focus on mitigating first, resisting the urge to fix or investigate the root cause. A proper investigation can wait until the next business day.<\/p><figure class=\"kg-card kg-image-card\"><img src=\"https:\/\/blog.pragmaticengineer.com\/content\/images\/2019\/07\/Screenshot-2019-07-03-at-17.14.53.png\" class=\"kg-image\" alt=\"image\" \/><\/figure><h2 id=\"postmortems-incident-reviews-a-culture-of-ongoing-improvements\">Postmortems, Incident Reviews &amp; a Culture of Ongoing Improvements<\/h2><p>It is telling of a team how they handle the aftermath of an outage. Do they carry on? Do they do a small investigation? Do they perhaps spend a surprisingly <em>large<\/em> effort on the follow-up, stopping product work to make system-level fixes?<\/p><p><strong>Postmortems<\/strong> done right are a building block for robust systems. A good postmortem is blameless yet thorough. Templates for postmortems at Uber evolved over time across engineering, having sections for the summary of the incident, impact overview, the timeline of events unfolding, a root cause analysis, lessons learned and a list of detailed follow-up items.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/blog.pragmaticengineer.com\/content\/images\/2019\/07\/Screenshot-2019-07-15-at-15.20.31.png\" class=\"kg-image\" alt=\"image\" \/><figcaption>A postmortem template similar to what I've used at Uber<\/figcaption><\/figure><p>Good postmortems dig deep down into the root cause and come up with improvements to make preventing, detecting or mitigating all similar outages a lot faster. When I say dig deep, I mean they won't stop at the root cause being a bad code change with a bug that the code reviewer did not spot. They dig deeper with the <a href=\"https:\/\/en.wikipedia.org\/wiki\/5_Whys\">\"5 Whys\"<\/a> exploration technique, to dig deeper, arriving at a more meaningful conclusion. Take this example:<\/p><ul><li><em>Why did the issue happen? --&gt;<\/em> A bug was committed as part of the code.<\/li><li><em>Why did the bug not get caught by someone else? --&gt;<\/em> The code reviewer did not notice that the code change could cause such an issue.<\/li><li><em>Why did we depend on only a code reviewer catching this bug?<\/em> ---&gt; Because we don't have an automated test for this use case.<\/li><li><em>\"Why don't we have an automated test for this use case?\"<\/em> --&gt; Because it is difficult to test without test accounts.<\/li><li><em>Why don't we have test accounts?<\/em> --&gt; Because this system does not yet support them<\/li><li>Conclusion: this issue points to the systemic issue of not having test accounts. Suggesting to add test accounts support to the system. Following this, write automated tests for all future, similar code changes.<\/li><\/ul><p><strong>Incident reviews<\/strong> are an important accompanying tool for postmortems. While many teams do a thorough job with postmortems, others can benefit from additional input and being challenged on preventative improvements. It's also essential that teams feel accountable and empowered to carry out the system level improvements they propose.<\/p><p>For organizations that are serious about reliability, the most severe incidents get reviewed and challenged by the experienced engineers. Organizational level engineering management should also be present to provide empowerment to go through with the fixes - <em>especially<\/em> when those are time-consuming and block other work. <strong>Robust systems are not built overnight: they are built through continuous iterations<\/strong>. Iterations that come from an organization-wide culture of continuous improvements, following learning from incidents.<\/p><p>There are a couple of regular activities that require significant investments but are critical in keeping a large distributed system up and running. These are concepts I came across at Uber for the first time - at previous companies, we did not need to use these as our scale and infrastructure did not prompt us to do so.<\/p><p><strong>A data center failover drill<\/strong> is something I assumed was dull until I observed a a few of them. I was initially thinking, designing robust distributed systems is precisely about being resilient to data centers going down. Why test it regularly, if it should <em>theoretically<\/em> just work? The answer has to do with scale and testing whether services can efficiently handle a sudden increase in traffic in a new data center.<\/p><p>The most common failure scenario I have observed is services not having enough resources in a new data center to handle global traffic, in case of a failover. Imagine ServiceA and ServiceB running from two data centers each. Let's assume that utilization of the resources is at 60% - with tens or hundreds of VMs running in each DC - and alerts are set to trigger at 70%. Now let's do a failover, directing all traffic from DataCenterA to DataCenterB. DataCenterB is suddenly unable to cope with the load, without provisioning new machines. Provisioning could take long enough time that requests get piled up and start dropping. This blocking could start to impact other services, causing a cascading failure on other systems, that are not even part of this failover.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/blog.pragmaticengineer.com\/content\/images\/2019\/07\/Failover-drill-drawing.png\" class=\"kg-image\" alt=\"image\" \/><figcaption>Possible ways a data center failover could go wrong<\/figcaption><\/figure><p>Other common failure scenarios involve routing level issues, network capacity problems, or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Backpressure_routing\">back pressure<\/a> pain points. Data center failovers are drills that any reliable distributed system should be able to perform without any user impact. I am emphasizing <em>should<\/em> - this drill is one of the most useful exercises to test the reliability of a web of distributed systems.<\/p><p><strong>Planned service downtime exercises <\/strong>are an excellent way to test the resiliency of the overall system. These are also a great way to discover hidden dependencies or inappropriate\/unexpected usages of a specific system. While this exercise can be done relatively easily with services that are client facing and have few known dependencies, it is less straightforward to do with critical systems that require high uptime and or with ones that many other systems depend on. But what will happen when this critical system will be unavailable, someday? It is better to validate the answer with a controlled drill, all teams being aware and ready, over an unexpected outage.<\/p><p><strong>Black-box testing <\/strong>is a way to measure the correctness of a system as close to the conditions as an end user would see it. This type of testing is similar to end-to-end testing. Except that for most products, having proper black-box tests warrant their own investment. Key user flows, and the most common, user-facing test scenarios are good ones to make \"black box testable\": doing this in a way that they can be triggered any time, to check if the systems work.<\/p><p>Taking Uber as an example, an obvious black-box test is checking if the rider-driver flow works at a city level. That is, can a rider in a specific city request an Uber, being matched with driver partners and get a ride? Once this scenario is automated, this test can be regularly run, simulating different cities. Having robust black-box testing systems makes it easier to verify that the system - or parts of the system - work correctly. It also helps a lot with failover drills: the quickest way to get feedback on the failover is to run the black-box tests.<\/p><figure class=\"kg-card kg-image-card kg-card-hascaption\"><img src=\"https:\/\/blog.pragmaticengineer.com\/content\/images\/2019\/07\/Screenshot-2019-07-03-at-22.07.07.png\" class=\"kg-image\" alt=\"image\" \/><figcaption>Example for using black-box tests during a failed failover drill, rolling back manually minutes into the drill.<\/figcaption><\/figure><p><strong>Capacity planning<\/strong> becomes equally important for large distributed systems. By large, I mean the cost of compute and storage being in the tens- or hundreds of thousands dollars per month. At this scale, having a fixed number of deployments might be cheaper over using self-scaling cloud solutions. At the very least, fixed deployments should handle the \"business as usual\" traffic, with automatic scaling happening for peak loads. But what is the minimum number of instances that should be running for the next month? The next three months? The next year?<\/p><p>Forecasting the future traffic pattern for systems that are mature and have good historical data is not difficult. Yet it is important both for budgeting, for choosing vendors or for locking in discounts with cloud providers. If your service has a large bill and you've not thought of capacity planning, you are missing out on an easy way to reduce and control cost.<\/p><h2 id=\"slos-slas-reporting-on-them\">SLOs, SLAs &amp; Reporting on Them<\/h2><p>SL0 stands for <a href=\"https:\/\/cloud.google.com\/blog\/products\/gcp\/sre-fundamentals-slis-slas-and-slos\"><strong>Service Level Objective<\/strong><\/a> - a numerical target for system availability. For each independent service, defining <strong>service-level SLOs<\/strong> like targets for capacity, latency, accuracy, and availability are good practices. These SLOs can then serve as triggers for alerting. An example service-level SLO could look like this:<\/p><table><thead><tr><th>SLO Metric<\/th>\n    <th>Subcategory<\/th>\n    <th>Value for Service<\/th>\n    <\/tr><\/thead><tr><td><strong>Capacity<\/strong><\/td>\n    <td>Minumum throughput<\/td>\n    <td>500 req\/sec<\/td>\n<\/tr><tr><td>Maximum expected throughput<\/td>\n    <td>2,500 req\/sec<\/td>\n<\/tr><tr><td><strong>Latency<\/strong><\/td>\n    <td>Expected median response time<\/td>\n    <td>50-90ms<\/td>\n<\/tr><tr><td>Expected p99 response time<\/td>\n    <td>500-800ms<\/td>\n<\/tr><tr><td><strong>Accuracy<\/strong><\/td>\n    <td>Maximum error rate<\/td>\n    <td>0.5%<\/td>\n<\/tr><tr><td><strong>Availability<\/strong><\/td>\n    <td>Guaranteed uptime<\/td>\n    <td>99.9%<\/td>\n<\/tr><\/table><p><strong>Business-level SLOs<\/strong> or functional SLOs are an abstraction above the services. They'll cover user or business-facing metrics. For example, a business-level SLO could be this: expecting 99.99% of email receipts to be sent within 1 minute of the trip completed. This SLO might map to service-level SLOs (e.g. latencies of payment and email receipt systems), or they might need to be measured differently.<\/p><p><strong>SLA - Service Level Agreement <\/strong>- is a broader agreement between a service provider and a service consumer. Usually, multiple SLOs make up an SLA. For example, the payments system being 99.99% available could be an SLA, that breaks down to specific SLOs for each supporting system.<\/p><p>After defining the SLOs, the next step is to measure these and report on them. <strong>Automating monitoring and reporting on SLAs and SLOs is often a complex project<\/strong>, one that is tempting to de-prioritize for both the engineering team and the business. The engineering team won't be as interested, they already have various levels of monitoring that detect outages realtime. The business, on the other hand, would rather prioritize delivering functionality over investing in a complex project that doesn't have immediate business impact.<\/p><p>Which brings us to the next topic: organizations that are operating large distributed systems sooner or later need to invest dedicated staffing for the reliability of these systems. Let's talk about the Site Reliability Engineering team.<\/p><h2 id=\"sre-as-an-independent-team\">SRE as an Independent Team<\/h2><p>Site Reliability Engineering <a href=\"https:\/\/en.wikipedia.org\/wiki\/Site_Reliability_Engineering\">originated<\/a> from Google from around 2003 - to more than 1,500 SRE engineers today. As operating a production environment becomes a more complex of a task, with more and more automation required, soon enough, this work becomes a full-time job. It varies when companies recognize that they have engineers working close to full time on production automation: the more critical these systems are and the more failures they have, the earlier this happens.<\/p><p>Fast growing tech companies often put an SRE team in-place early on, who build their own roadmap. At Uber, the <a href=\"https:\/\/eng.uber.com\/sre-talks-feb-2016\/\">SRE team was founded in 2015<\/a> with the mission of managing system complexity over time. Other companies might couple starting such a team to when a dedicated infrastructure team is created. <strong>When a company grows to reliability work across teams takes up more than a few engineers' time, it's time to put a dedicated team for this in place.<\/strong><\/p><p>With an SRE team in-place, this team makes the operational aspects of keeping large, distributed systems a lot easier for all engineers. The SRE team likely owns standard monitoring and alerting tools. They probably buy or build oncall tooling and are the goto team for oncall best practices. They might facilitate incident reviews and build systems to make detecting, mitigating, and preventing outages easier. They certainly facilitate failover drills, often drive black-box testing, and are involved with capacity planning. They drive choosing, customizing, or building standard tools to define and measure SLOs and report on them.<\/p><p>Given that companies have different pain areas they look for SRE to solve, SRE organizations are different across companies. The name is often something else as well: it could be called operations, platform engineering, or infrastructure. Google published two <a href=\"https:\/\/landing.google.com\/sre\/books\/\">must-read books on site reliability that are freely available<\/a> and is an excellent read to go deeper into SRE.<\/p><h2 id=\"reliability-as-an-ongoing-investment\">Reliability as an Ongoing Investment<\/h2><p>When building any product, building the first version is just the start. After the v1, new features get added in the iterations to come. If the product is successful and brings in business results, the work just keeps on coming.<\/p><p>Distributed systems have a similar lifecycle, except they need more investment, not just for new features, but to keep up with scaling up. As a system starts to take more load, stores more data, has more engineers work on it, it needs continuous care to keep running smoothly. Many people building distributed systems the first time assume this system to be like a car: once built, it only needing essential maintenance every few months. This comparison could not be further from how systems like these operate.<\/p><p><strong>I like to think of the effort to operate a distributed system being similar to operating a large organization, like a hospital<\/strong>. To make sure a hospital operates well, continuous validation and checks are needed (monitoring, alerting, black-box testing). New staff and equipment (new engineers and new services) need to be onboarded all the time. As the number of people and the number of services grows, new the old ways of doing things become inefficient: the same way a tiny clinic on the countryside works different to a large hospital in a metropolis. It becomes a full-time job to come up with more efficient ways of doing things, and measuring and reporting on efficiency becomes important. Just like large hospitals have more supporting officer staff, like finance, HR or security; operating larger distributed systems similarly rely on supporting teams like infrastructure and SRE.<\/p><p>For a team to run a reliable distributed system, the organization needs to be investing continuously in the operations of these systems, as well as the platforms that they are built on.<\/p><h2 id=\"further-recommended-reading\">Further Recommended Reading<\/h2><p>Though this post is lengthy in its content, it still only touches the surface. To dive deeper into operating distributed systems, I recommend the following resources:<\/p><p><strong>Books<\/strong><\/p><ul><li><a href=\"https:\/\/landing.google.com\/sre\/\">Google SRE Book<\/a> - a great and free book from Google. The <a href=\"https:\/\/landing.google.com\/sre\/sre-book\/chapters\/monitoring-distributed-systems\/\">Monitoring Distributed Systems charter<\/a> is especially relevant for this post.<\/li><li><a href=\"https:\/\/www.oreilly.com\/library\/view\/distributed-systems-observability\/9781492033431\/\">Distributed Systems Observability<\/a> by <a href=\"https:\/\/twitter.com\/copyconstruct\">Cindy Sridharan<\/a>, another excellent and free book with great points on monitoring distributed systems.<\/li><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1449373321\/ref=as_li_tl?ie=UTF8&amp;tag=gregdoesit-20&amp;camp=1789&amp;creative=9325&amp;linkCode=as2&amp;creativeASIN=1449373321&amp;linkId=adc1dc62fd3463b173cfd92dbe4ed821\">Designing Data-Intensive Applications<\/a> by <a href=\"https:\/\/twitter.com\/martinkl\">Dr Martin Kleppmann<\/a> - the most practical book I have found so far on distributed systems concepts. However, this book does not touch much on the operational aspects discussed in the post.<\/li><\/ul><p><strong>Online resources<\/strong><\/p><ul><li><a href=\"https:\/\/increment.com\/on-call\/\">On-Call issue of the Increment magazine<\/a>: a series of articles on incident response processes of companies like Amazon, Dropbox, Facebook, Google, and Netflix.<\/li><li><a href=\"http:\/\/brooker.co.za\/blog\/2019\/04\/03\/learning.html\">Learning to Build Distributed Systems<\/a> - a post by AWS engineer <a href=\"https:\/\/twitter.com\/MarcJBrooker\">Marc Brooker<\/a>, answering the question \"how do I learn to build big distributed systems\"?<\/li><\/ul><p>See <a href=\"https:\/\/news.ycombinator.com\/item?id=20462349\">comments for this article on Hacker News<\/a>.<\/p>","created_at":"2019-07-17T21:02:53+0000","updated_at":"2019-07-17T21:02:53+0000","published_at":"2019-07-15T15:47:04+0000","published_by":["Gergely Orosz"],"annotations":[],"mimetype":"text\/html","reading_time":21,"domain_name":"blog.pragmaticengineer.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":982,"title":"emacs-elsa\/Elsa","url":"https:\/\/github.com\/emacs-elsa\/Elsa","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p style=\"text-align: center;\">(Your favourite princess now in Emacs!)<\/p>\n<p><a href=\"https:\/\/coveralls.io\/github\/Fuco1\/Elsa?branch=master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/dd0bcbf7c160108694a238ec123c61a5e366719b\/68747470733a2f2f636f766572616c6c732e696f2f7265706f732f6769746875622f4675636f312f456c73612f62616467652e7376673f6272616e63683d6d6173746572\" alt=\"Coverage Status\" data-canonical-src=\"https:\/\/coveralls.io\/repos\/github\/Fuco1\/Elsa\/badge.svg?branch=master\" \/><\/a> <a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=A5PMGVKCQBT88\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e79b5862b1027f34216f36685d178b856fa72a72\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f50617950616c2d446f6e6174652d6f72616e67652e7376673f6c6f676f3d70617970616c\" alt=\"Paypal logo\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/PayPal-Donate-orange.svg?logo=paypal\" \/><\/a> <a href=\"https:\/\/www.patreon.com\/user?u=3282358\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/11ef9285067312fe2fb18b032b77ef8a29851a9c\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f50617472656f6e2d4265636f6d6525323061253230706174726f6e2d6f72616e67652e7376673f6c6f676f3d70617472656f6e\" alt=\"Patreon\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/Patreon-Become%20a%20patron-orange.svg?logo=patreon\" \/><\/a><\/p>\n<p>Elsa is a tool that analyses your code without loading or running it.\nIt can track types and provide helpful hints when things don't match\nup before you even try to run the code.<\/p>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#state-of-the-project\">State of the project<\/a><\/li>\n<li><a href=\"#non-exhaustive-list-of-features\">Non-exhaustive list of features<\/a>\n<ul><li><a href=\"#detect-dead-code\">Detect dead code<\/a><\/li>\n<li><a href=\"#enforce-style-rules\">Enforce style rules<\/a><\/li>\n<li><a href=\"#look-for-suspicious-code\">Look for suspicious code<\/a><\/li>\n<li><a href=\"#track-types-of-expressions\">Track types of expressions<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#how-do-i-run-it\">How do I run it<\/a>\n<ul><li><a href=\"#flycheck-integration\">Flycheck integration<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#configuration\">Configuration<\/a>\n<ul><li><a href=\"#analysis-extension\">Analysis extension<\/a><\/li>\n<li><a href=\"#rulesets\">Rulesets<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#type-annotations\">Type annotations<\/a><\/li>\n<li><a href=\"#how-can-i-contribute-to-this-project\">How can I contribute to this project<\/a><\/li>\n<li><a href=\"#faq\">F.A.Q.<\/a>\n<ul><li><a href=\"#whats-up-with-the-logo\">What's up with the logo?<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#for-developers\">For developers<\/a>\n<ul><li><a href=\"#how-to-write-an-extension-for-your-favourite-package\">How to write an extension for your-favourite-package<\/a><\/li>\n<li><a href=\"#how-to-write-a-ruleset\">How to write a ruleset<\/a><\/li>\n<\/ul><\/li>\n<\/ul>\n<p>We are currently in a very early <em>ALPHA<\/em> phase.  API is somewhat\nstable but the type system and annotations are under constant\ndevelopment.  Things might break at any point.<\/p>\n<p>Here comes a non-exhaustive list of some more interesting features.<\/p>\n<p>The error highlightings in the screenshots are provided by <a href=\"https:\/\/github.com\/emacs-elsa\/flycheck-elsa\">Elsa\nFlycheck extension<\/a>.<\/p>\n<p>Everything you see here actually works, this is not just for show!<\/p>\n<h2><a id=\"user-content-detect-dead-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#detect-dead-code\"><\/a>Detect dead code<\/h2>\n<h3><a id=\"user-content-detect-suspicious-branching-logic\" class=\"anchor\" aria-hidden=\"true\" href=\"#detect-suspicious-branching-logic\"><\/a>Detect suspicious branching logic<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/dead-code-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/dead-code-1.png\" alt=\"\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/dead-code-2.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/dead-code-2.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-find-unreachable-code-in-short-circuiting-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#find-unreachable-code-in-short-circuiting-forms\"><\/a>Find unreachable code in short-circuiting forms<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/unreachable-code-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/unreachable-code-1.png\" alt=\"\" \/><\/a><\/p>\n<h2><a id=\"user-content-enforce-style-rules\" class=\"anchor\" aria-hidden=\"true\" href=\"#enforce-style-rules\"><\/a>Enforce style rules<\/h2>\n<h3><a id=\"user-content-provide-helpful-tips-for-making-code-cleaner\" class=\"anchor\" aria-hidden=\"true\" href=\"#provide-helpful-tips-for-making-code-cleaner\"><\/a>Provide helpful tips for making code cleaner<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/useless-code-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/useless-code-1.png\" alt=\"\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/useless-code-2.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/useless-code-2.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-add-custom-rules-for-your-own-project-with-rulesets\" class=\"anchor\" aria-hidden=\"true\" href=\"#add-custom-rules-for-your-own-project-with-rulesets\"><\/a>Add custom rules for your own project with rulesets<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/custom-ruleset-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/custom-ruleset-1.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-make-formatting-consistent\" class=\"anchor\" aria-hidden=\"true\" href=\"#make-formatting-consistent\"><\/a>Make formatting consistent<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/formatting-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/formatting-1.png\" alt=\"\" \/><\/a><\/p>\n<h2><a id=\"user-content-look-for-suspicious-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#look-for-suspicious-code\"><\/a>Look for suspicious code<\/h2>\n<h3><a id=\"user-content-find-references-to-freeunbound-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#find-references-to-freeunbound-variables\"><\/a>Find references to free\/unbound variables<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/unbound-variable-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/unbound-variable-1.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-dont-assign-to-free-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#dont-assign-to-free-variables\"><\/a>Don't assign to free variables<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/unbound-variable-2.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/unbound-variable-2.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-detect-conditions-which-are-always-true-or-false\" class=\"anchor\" aria-hidden=\"true\" href=\"#detect-conditions-which-are-always-true-or-false\"><\/a>Detect conditions which are always true or false<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/always-nil-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/always-nil-1.png\" alt=\"\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/always-non-nil-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/always-non-nil-1.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-make-sure-functions-are-passed-enough-arguments\" class=\"anchor\" aria-hidden=\"true\" href=\"#make-sure-functions-are-passed-enough-arguments\"><\/a>Make sure functions are passed enough arguments<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/number-of-args-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/number-of-args-1.png\" alt=\"\" \/><\/a><\/p>\n<h3><a id=\"user-content-make-sure-functions-are-not-passed-too-many-arguments\" class=\"anchor\" aria-hidden=\"true\" href=\"#make-sure-functions-are-not-passed-too-many-arguments\"><\/a>Make sure functions are not passed too many arguments<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/number-of-args-2.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/number-of-args-2.png\" alt=\"\" \/><\/a><\/p>\n<h2><a id=\"user-content-track-types-of-expressions\" class=\"anchor\" aria-hidden=\"true\" href=\"#track-types-of-expressions\"><\/a>Track types of expressions<\/h2>\n<h3><a id=\"user-content-check-types-of-arguments-passed-to-functions-for-compatibility\" class=\"anchor\" aria-hidden=\"true\" href=\"#check-types-of-arguments-passed-to-functions-for-compatibility\"><\/a>Check types of arguments passed to functions for compatibility<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/type-inference-1.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/type-inference-1.png\" alt=\"\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/type-inference-2.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/type-inference-2.png\" alt=\"\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/blob\/master\/images\/type-inference-3.png\"><img src=\"https:\/\/github.com\/emacs-elsa\/Elsa\/raw\/master\/images\/type-inference-3.png\" alt=\"\" \/><\/a><\/p>\n<p>Currently we only support running Elsa with <a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a>.<\/p>\n<h3><a id=\"user-content-recommended-using-packaged-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#recommended-using-packaged-version\"><\/a>[RECOMMENDED] Using packaged version<\/h3>\n<p>The easiest and fastest way to install Elsa is through\n<a href=\"http:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a> and <a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a>.<\/p>\n<ol><li>Add <code>(depends-on \"elsa\")<\/code> to <code>Cask<\/code> file of your project.<\/li>\n<li>Run <code>cask install<\/code>.<\/li>\n<li><code>cask exec elsa FILE-TO-ANALYSE [ANOTHER-FILE...]<\/code> to analyse the file.<\/li>\n<\/ol><h3><a id=\"user-content-using-development-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-development-version\"><\/a>Using development version<\/h3>\n<p>To use the development version you can clone the repository and use\n<code>cask link<\/code> feature to use the code from the clone.<\/p>\n<ol><li><code>git clone https:\/\/github.com\/emacs-elsa\/Elsa.git<\/code> somewhere to your computer.<\/li>\n<li>Add <code>(depends-on \"elsa\")<\/code> to <code>Cask<\/code> file of your project.<\/li>\n<li>Run <code>cask link elsa &lt;path-to-elsa-repo&gt;<\/code>.<\/li>\n<li><code>cask exec elsa FILE-TO-ANALYSE [ANOTHER-FILE...]<\/code> to analyse the file.<\/li>\n<\/ol><h2><a id=\"user-content-flycheck-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#flycheck-integration\"><\/a>Flycheck integration<\/h2>\n<p>If you use <a href=\"https:\/\/github.com\/flycheck\/flycheck\">flycheck<\/a> you can\nuse the <a href=\"https:\/\/github.com\/emacs-elsa\/flycheck-elsa\">flycheck-elsa<\/a>\npackage which integrates Elsa with Flycheck.<\/p>\n<p>By default Elsa core comes with very little built-in logic, only\nunderstanding the elisp <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Special-Forms.html\" rel=\"nofollow\">special\nforms<\/a>.<\/p>\n<p>However, we ship a large number of extensions for popular packages\nsuch as <code>eieio<\/code>, <code>cl<\/code>, <code>dash<\/code> or even <code>elsa<\/code> itself.<\/p>\n<p>You can configure Elsa by adding an <code>Elsafile.el<\/code> to your project.\nThe <code>Elsafile.el<\/code> should be located next to the <code>Cask<\/code> file.<\/p>\n<p>There are multiple ways to extend the capabilities of Elsa.<\/p>\n<h2><a id=\"user-content-analysis-extension\" class=\"anchor\" aria-hidden=\"true\" href=\"#analysis-extension\"><\/a>Analysis extension<\/h2>\n<p>One is by providing special analysis rules for more forms and\nfunctions where we can exploit the knowledge of how the function\nbehaves to narrow the analysis down more.<\/p>\n<p>For example, we can say that if the input of <code>not<\/code> is <code>t<\/code>, the return\nvalue is always <code>nil<\/code>.  This encodes our domain knowledge in form of\nan analysis rule.<\/p>\n<p>All the rules are added in form of extensions.  Elsa has few core\nextensions for most common built-in functions such as list\nmanupulation (<code>car<\/code>, <code>nth<\/code>...), predicates (<code>stringp<\/code>, <code>atomp<\/code>...),\nlogical functions (<code>not<\/code>, ...) and so on.  These are automatically\nloaded because the functions are so common virtually every project is\ngoing to use them.<\/p>\n<p>Additional extensions are provided for popular external packages such\nas <a href=\"https:\/\/github.com\/magnars\/dash.el\">dash.el<\/a>.  To use them, add to\nyour <code>Elsafile.el<\/code> the <code>register-extensions<\/code> form, like so<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(register-extensions\n dash\n ;; more extensions here\n )<\/pre><\/div>\n<h2><a id=\"user-content-rulesets\" class=\"anchor\" aria-hidden=\"true\" href=\"#rulesets\"><\/a>Rulesets<\/h2>\n<p>After analysis of the forms is done we have all the type information\nand the AST ready to be further processed by various checks and rules.<\/p>\n<p>These can be (non-exhaustive list):<\/p>\n<ul><li>Stylistic, such as checking that a variable uses <code>lisp-case<\/code> for\nnaming instead of <code>snake_case<\/code>.<\/li>\n<li>Syntactic, such as checking we are not wrapping the else branch of\n<code>if<\/code> with a useless <code>progn<\/code>.<\/li>\n<li>Semantic, such as checking that the condition of <code>if<\/code> does not\nalways evaluate to <code>non-nil<\/code> (in which case the <code>if<\/code> form is\nuseless).<\/li>\n<\/ul><p>Elsa provides some built-in rulesets and more can also be used by loading extensions.<\/p>\n<p>To register a ruleset, add the following form to <code>Elsafile.el<\/code><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(register-ruleset\n dead-code\n style\n ;; more rulesets here\n )<\/pre><\/div>\n<p>In Elisp users are not required to provide type annotations to their\ncode.  While at many places the types can be inferred there are\nplaces, especially in user-defined functions, where we can not guess\nthe correct type (we can only infer what we see during runtime).<\/p>\n<p>Users can annotate their <code>defun<\/code> definitions like this:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; (elsa-pluralize :: String -&gt; Int -&gt; String)\n(defun elsa-pluralize (word n)\n  \"Return singular or plural of WORD based on N.\"\n  (if (= n 1)\n      word\n    (concat word \"s\")))<\/pre><\/div>\n<p>The <code>(elsa-pluralise :: ...)<\/code> inside a comment form provides\nadditional information to the Elsa analysis.  Here we say that the\nfunction following such a comment takes two arguments, string and int,\nand returns a string.<\/p>\n<p>The syntax of the type annotation is somewhat modeled after Haskell\nbut there are some special constructs available to Elsa<\/p>\n<p>Here are general guidelines on how the types are constructed.<\/p>\n<ul><li>For built-in types with test predicates, drop the <code>p<\/code> or <code>-p<\/code> suffix and PascalCase to get the type:\n<ul><li><code>stringp<\/code> \u2192 <code>String<\/code><\/li>\n<li><code>integerp<\/code> \u2192 <code>Integer<\/code> (<code>Int<\/code> is also accepted)<\/li>\n<li><code>markerp<\/code> \u2192 <code>Marker<\/code><\/li>\n<li><code>hash-table-p<\/code> \u2192 <code>HashTable<\/code><\/li>\n<\/ul><\/li>\n<li>A type for everything is called <code>Mixed<\/code>.  It accepts anything and is\nalways nullable.  This is the default type for when we lack type\ninformation.<\/li>\n<li>Sum types can be specified with <code>|<\/code> syntax, so <code>String | Integer<\/code> is\na type accepting both strings or integers.<\/li>\n<li>Cons types are specified by prefixing wrapping the <code>car<\/code> and <code>cdr<\/code>\ntypes with a <code>Cons<\/code> constructor, so <code>Cons Int Int<\/code> is a type where\nthe <code>car<\/code> is an int and <code>cdr<\/code> is also an int, for example <code>(1 . 3)<\/code>.<\/li>\n<li>List types are specified by wrapping a type in a vector <code>[]<\/code>\nconstructor, so <code>[Int]<\/code> is a list of integers and <code>[String | Int]<\/code>\nis a list of items where each item is either a string or an integer.\nA type constructor <code>List<\/code> is also supported.<\/li>\n<li>Function types are created by separating argument types and the\nreturn type with <code>-&gt;<\/code> token.<\/li>\n<li>To make variadic types (for the <code>&amp;rest<\/code> keyword) add three dots\n<code>...<\/code> after the type, so <code>String... -&gt; String<\/code> is a function taking\nany number of strings and returning a string, such as <code>concat<\/code>.\nNote: a variadic type is internally just a list of the same base\ntype but it has a flag that allows the function be of variable\narity.  A <code>Variadic<\/code> type constructor is also available to construct\ncomplex types.<\/li>\n<li>To mark type as nullable you can attach <code>?<\/code> to the end of it, so\nthat <code>Int?<\/code> accepts any integer and also a <code>nil<\/code>.  A <code>Maybe<\/code> type\nconstructor is also available to construct complex types.<\/li>\n<\/ul><p>Some type constructors have optional arguments, for example writing\njust <code>Cons<\/code> will assume the <code>car<\/code> and <code>cdr<\/code> are of type <code>Mixed<\/code>.<\/p>\n<p>Open an issue if you want to work on something (not necessarily listed\nbelow in the roadmap) so we won't duplicate work.  Or just give us\nfeedback or helpful tips.<\/p>\n<p>You can provide type definitions for built-in functions by extending\n<code>elsa-typed-builtin.el<\/code>.  There is plenty to go.  Some of the types\nnecessary to express what we want might not exist or be supported yet,\nopen an issue so we can discuss how to model things.<\/p>\n<h2><a id=\"user-content-whats-up-with-the-logo\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-up-with-the-logo\"><\/a>What's up with the logo?<\/h2>\n<p>See the <a href=\"https:\/\/github.com\/emacs-elsa\/Elsa\/issues\/80\">discussion<\/a>.<\/p>\n<p>After calling <code>(require 'elsa-font-lock)<\/code> there is a function\n<code>elsa-setup-font-lock<\/code> which can be called from <code>emacs-lisp-mode-hook<\/code>\nto set up some additional font-locking for Elsa types.<\/p>\n<h2><a id=\"user-content-how-to-write-an-extension-for-your-favourite-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-write-an-extension-for-your-favourite-package\"><\/a>How to write an extension for your-favourite-package<\/h2>\n<h2><a id=\"user-content-how-to-write-a-ruleset\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-write-a-ruleset\"><\/a>How to write a ruleset<\/h2>\n<\/article>","created_at":"2019-07-16T12:08:09+0000","updated_at":"2019-07-16T12:08:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/42645208?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":981,"title":"Passing web request context transparently to a celery task","url":"https:\/\/stackoverflow.com\/questions\/21756797\/passing-web-request-context-transparently-to-a-celery-task","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I've a multi-tenant setup where I'd like to pass certain customer specific information, specifically <em>request.host<\/em> to the celery task, where ideally it should be available in a global variable. Is there a way to set this up, in a manner transparent to the application?<\/p>\n<p>the task would be called the same way:<\/p>\n<pre>my_background_func.delay(foo, bar)\n<\/pre>\n<p>the task is defined the same way, except that it has access to a global variable called 'request' having an attribute 'host':<\/p>\n<pre>@celery_app.task\ndef my_background_func(foo, bar):\n    print \"running the task for host:\" + request.host\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/21756797\/revisions\" title=\"show all edits to this post\">edited Jul 12 '18 at 13:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4720935\/mel\">Mel<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            3,95699 gold badges2626 silver badges3232 bronze badges&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Feb 13 '14 at 14:19&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1893487\/vrtx54234\">vrtx54234<\/a>vrtx54234&#13;\n        <div class=\"-flair\">&#13;\n            75011 gold badge1313 silver badges3131 bronze badges&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"1\">&#13;\n                                1 Answer&#13;\n                            1&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>here's how I solved it ... <\/p>\n<pre>class MyTask(Task):\n    abstract = True\n    def delay(self, *args, **kwargs):\n        return self.apply_async(args, kwargs, headers={'host': request.host})\n<\/pre>\n<p>on the client:<\/p>\n<pre> @celery_app.task(base=MyTask, bind=True)\n def hellohost(task):\n     return \"hello \" +  task.request.headers['host']\n<\/pre>\n<p>it works, but strangely hellohost.delay().get() hangs on the client<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/21778918\/revisions\" title=\"show all edits to this post\">edited Jul 12 '18 at 13:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4720935\/mel\">Mel<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            3,95699 gold badges2626 silver badges3232 bronze badges&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 14 '14 at 12:03&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1893487\/vrtx54234\">vrtx54234<\/a>vrtx54234&#13;\n        <div class=\"-flair\">&#13;\n            75011 gold badge1313 silver badges3131 bronze badges&#13;\n        <\/div>&#13;<\/div>","created_at":"2019-07-16T03:27:01+0000","updated_at":"2019-07-16T03:27:01+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":980,"title":"Thread Local Context \u2014 structlog documentation","url":"https:\/\/www.structlog.org\/en\/stable\/thread-local.html","content":"<div class=\"section\" id=\"immutability\"><h2>Immutability<\/h2><blockquote>\n<div><p>You should call some functions with some arguments.<\/p><p class=\"attribution\">\u2014David Reid<\/p><\/div><\/blockquote><p>The behavior of copying itself, adding new values, and returning the result is useful for applications that keep somehow their own context using classes or closures.\nTwisted is a <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/examples.html#twisted-example\">fine example<\/a> for that.\nAnother possible approach is passing wrapped loggers around or log only within your view where you gather errors and events using return codes and exceptions.\nIf you are willing to do that, you should stick to it because <a class=\"reference external\" href=\"https:\/\/en.wikipedia.org\/wiki\/Immutable_object\">immutable state<\/a> is a very good thing<a class=\"footnote-reference\" href=\"#id3\" id=\"id1\">[*]<\/a>.\nSooner or later, global state and mutable data lead to unpleasant surprises.<\/p><p>However, in the case of conventional web development, we realize that passing loggers around seems rather cumbersome, intrusive, and generally against the mainstream culture.\nAnd since it\u2019s more important that people actually <em>use<\/em> <code class=\"docutils literal notranslate\">structlog<\/code> than to be pure and snobby, <code class=\"docutils literal notranslate\">structlog<\/code> contains a dirty but convenient trick: thread local context storage which you may already know from <a class=\"reference external\" href=\"http:\/\/flask.pocoo.org\/docs\/design\/#thread-locals\">Flask<\/a>:<\/p><p>Thread local storage makes your logger\u2019s context global but <em>only within the current thread<\/em><a class=\"footnote-reference\" href=\"#id4\" id=\"id2\">[\u2020]<\/a>.\nIn the case of web frameworks this usually means that your context becomes global to the current request.<\/p><p>The following explanations may sound a bit confusing at first but the <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/examples.html#flask-example\">Flask example<\/a> illustrates how simple and elegant this works in practice.<\/p><\/div><div class=\"section\" id=\"wrapped-dicts\"><h2>Wrapped Dicts<\/h2><p>In order to make your context thread local, <code class=\"docutils literal notranslate\">structlog<\/code> ships with a function that can wrap any dict-like class to make it usable for thread local storage: <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#structlog.threadlocal.wrap_dict\" title=\"structlog.threadlocal.wrap_dict\"><code class=\"xref py py-func docutils literal notranslate\">structlog.threadlocal.wrap_dict()<\/code><\/a>.<\/p><p>Within one thread, every instance of the returned class will have a <em>common<\/em> instance of the wrapped dict-like class:<\/p><div class=\"highlight-pycon notranslate\"><div class=\"highlight\"><pre>&gt;&gt;&gt; from structlog.threadlocal import wrap_dict\n&gt;&gt;&gt; WrappedDictClass = wrap_dict(dict)\n&gt;&gt;&gt; d1 = WrappedDictClass({\"a\": 1})\n&gt;&gt;&gt; d2 = WrappedDictClass({\"b\": 2})\n&gt;&gt;&gt; d3 = WrappedDictClass()\n&gt;&gt;&gt; d3[\"c\"] = 3\n&gt;&gt;&gt; d1 is d3\nFalse\n&gt;&gt;&gt; d1 == d2 == d3 == WrappedDictClass()\nTrue\n&gt;&gt;&gt; d3  \n&lt;WrappedDict-...({'a': 1, 'b': 2, 'c': 3})&gt;\n<\/pre><\/div><\/div><p>To enable thread local context use the generated class as the context class:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>configure(context_class=WrappedDictClass)\n<\/pre><\/div><\/div><div class=\"admonition note\"><p class=\"first admonition-title\">Note<\/p><p class=\"last\">Creation of a new <code class=\"docutils literal notranslate\">BoundLogger<\/code> initializes the logger\u2019s context as <code class=\"docutils literal notranslate\">context_class(initial_values)<\/code>, and then adds any values passed via <code class=\"docutils literal notranslate\">.bind()<\/code>.\nAs all instances of a wrapped dict-like class share the same data, in the case above, the new logger\u2019s context will contain all previously bound values in addition to the new ones.<\/p><\/div><p><a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#structlog.threadlocal.wrap_dict\" title=\"structlog.threadlocal.wrap_dict\"><code class=\"xref py py-func docutils literal notranslate\">structlog.threadlocal.wrap_dict()<\/code><\/a> returns always a completely <em>new<\/em> wrapped class:<\/p><div class=\"highlight-pycon notranslate\"><div class=\"highlight\"><pre>&gt;&gt;&gt; from structlog.threadlocal import wrap_dict\n&gt;&gt;&gt; WrappedDictClass = wrap_dict(dict)\n&gt;&gt;&gt; AnotherWrappedDictClass = wrap_dict(dict)\n&gt;&gt;&gt; WrappedDictClass() != AnotherWrappedDictClass()\nTrue\n&gt;&gt;&gt; WrappedDictClass.__name__  \nWrappedDict-41e8382d-bee5-430e-ad7d-133c844695cc\n&gt;&gt;&gt; AnotherWrappedDictClass.__name__   \nWrappedDict-e0fc330e-e5eb-42ee-bcec-ffd7bd09ad09\n<\/pre><\/div><\/div><p>In order to be able to bind values temporarily to a logger, <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#module-structlog.threadlocal\" title=\"structlog.threadlocal\"><code class=\"xref py py-mod docutils literal notranslate\">structlog.threadlocal<\/code><\/a> comes with a <a class=\"reference external\" href=\"https:\/\/docs.python.org\/2\/library\/stdtypes.html#context-manager-types\">context manager<\/a>: <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#structlog.threadlocal.tmp_bind\" title=\"structlog.threadlocal.tmp_bind\"><code class=\"xref py py-func docutils literal notranslate\">tmp_bind()<\/code><\/a>:<\/p><div class=\"highlight-pycon notranslate\"><div class=\"highlight\"><pre>&gt;&gt;&gt; log.bind(x=42)  \n&lt;BoundLogger(context=&lt;WrappedDict-...({'x': 42})&gt;, ...)&gt;\n&gt;&gt;&gt; log.msg(\"event!\")\nx=42 event='event!'\n&gt;&gt;&gt; with tmp_bind(log, x=23, y=\"foo\") as tmp_log:\n...     tmp_log.msg(\"another event!\")\nx=23 y='foo' event='another event!'\n&gt;&gt;&gt; log.msg(\"one last event!\")\nx=42 event='one last event!'\n<\/pre><\/div><\/div><p>The state before the <code class=\"docutils literal notranslate\">with<\/code> statement is saved and restored once it\u2019s left.<\/p><p>If you want to detach a logger from thread local data, there\u2019s <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#structlog.threadlocal.as_immutable\" title=\"structlog.threadlocal.as_immutable\"><code class=\"xref py py-func docutils literal notranslate\">structlog.threadlocal.as_immutable()<\/code><\/a>.<\/p><\/div><div class=\"section\" id=\"downsides-caveats\"><h2>Downsides &amp; Caveats<\/h2><p>The convenience of having a thread local context comes at a price though:<\/p><div class=\"admonition warning\"><p class=\"first admonition-title\">Warning<\/p><ul class=\"last\"><li><p class=\"first\">If you can\u2019t rule out that your application re-uses threads, you <em>must<\/em> remember to <strong>initialize your thread local context<\/strong> at the start of each request using <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#structlog.BoundLogger.new\" title=\"structlog.BoundLogger.new\"><code class=\"xref py py-func docutils literal notranslate\">new()<\/code><\/a> (instead of <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/api.html#structlog.BoundLogger.bind\" title=\"structlog.BoundLogger.bind\"><code class=\"xref py py-func docutils literal notranslate\">bind()<\/code><\/a>).\nOtherwise you may start a new request with the context still filled with data from the request before.<\/p>\n<\/li>\n<li><p class=\"first\"><strong>Don\u2019t<\/strong> stop assigning the results of your <code class=\"docutils literal notranslate\">bind()<\/code>s and <code class=\"docutils literal notranslate\">new()<\/code>s!<\/p>\n<p><strong>Do<\/strong>:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>log = log.new(y=23)\nlog = log.bind(x=42)\n<\/pre><\/div><\/div>\n<p><strong>Don\u2019t<\/strong>:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>log.new(y=23)\nlog.bind(x=42)\n<\/pre><\/div><\/div>\n<p>Although the state is saved in a global data structure, you still need the global wrapped logger produce a real bound logger.\nOtherwise each log call will result in an instantiation of a temporary BoundLogger.\nSee <a class=\"reference internal\" href=\"https:\/\/www.structlog.org\/en\/stable\/configuration.html#configuration\">Configuration<\/a> for more details.<\/p>\n<\/li>\n<\/ul><\/div><p>The general sentiment against thread locals is that they\u2019re hard to test.\nIn this case we feel like this is an acceptable trade-off.\nYou can easily write deterministic tests using a call-capturing processor if you use the API properly (cf. warning above).<\/p><p>This big red box is also what separates immutable local from mutable global data.<\/p><table class=\"docutils footnote\" frame=\"void\" id=\"id3\" rules=\"none\"><colgroup><col class=\"label\" \/><col \/><\/colgroup><tbody valign=\"top\"><tr><td class=\"label\"><a class=\"fn-backref\" href=\"#id1\">[*]<\/a><\/td><td>In the spirit of Python\u2019s \u2018consenting adults\u2019, <code class=\"docutils literal notranslate\">structlog<\/code> doesn\u2019t enforce the immutability with technical means.\nHowever, if you don\u2019t meddle with undocumented data, the objects can be safely considered immutable.<\/td><\/tr><\/tbody><\/table><table class=\"docutils footnote\" frame=\"void\" id=\"id4\" rules=\"none\"><colgroup><col class=\"label\" \/><col \/><\/colgroup><tbody valign=\"top\"><tr><td class=\"label\"><a class=\"fn-backref\" href=\"#id2\">[\u2020]<\/a><\/td><td>Special care has been taken to detect and support greenlets properly.<\/td><\/tr><\/tbody><\/table><\/div>","created_at":"2019-07-15T15:15:48+0000","updated_at":"2019-07-15T15:15:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.structlog.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["indiehackers","bootstrapping"],"is_public":false,"id":979,"title":"lightningminers\/awesome-indie","url":"https:\/\/github.com\/lightningminers\/awesome-indie","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p>Resources to help independent developers make money.<\/p>\n<p>Making money from your own code has never been easier, but it's still hard. Here, I try to curate a list of resources to help everyone interested in making money from their digital products get inspired and give it a try. Be it monetizing side-projects, bootstrapping business ideas, etc.<\/p>\n<p>Projects don't need to be big, actually, it's better if they are not, they just need to be profitable. They might provide a passive income after the initial bulk of work, and who knows, maybe even help you reach financial independence.<\/p>\n<p>Going indie does have lots of lifestyle perks :)<\/p>\n<p>All these resources might help inspire you and put you on the right track but in the end what matters is to <strong>MAKE and TRY stuff<\/strong>. So devour this content responsibly :)<\/p>\n<p>Contributions are more than welcome.<\/p>\n<p>by <strong><a href=\"https:\/\/twitter.com\/mezood\" rel=\"nofollow\">@mezod<\/a><\/strong> (Doing things every day with <strong><a href=\"https:\/\/everyday.app\" rel=\"nofollow\">everyday.app<\/a><\/strong>)<\/p>\n<p><strong>Notes:<\/strong><\/p>\n<ul><li>If [$], it means it's not free and probably means I haven't tried it but still decided to add it because of the several references.<\/li>\n<li>This repo was once <a href=\"https:\/\/medium.com\/@mezood\/making-money-from-your-own-code-has-never-been-easier-but-its-still-hard-2242ddaae6a1\" rel=\"nofollow\">#1 on Github Trending<\/a>! :_)<\/li>\n<\/ul><h2><a id=\"user-content-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#contents\"><\/a>Contents<\/h2>\n<ul><li><a href=\"#communities\">Communities<\/a><\/li>\n<li><a href=\"#newsletters\">Newsletters<\/a><\/li>\n<li><a href=\"#podcasts\">Podcasts<\/a><\/li>\n<li><a href=\"#talks\">Talks<\/a><\/li>\n<li><a href=\"#posts\">Posts<\/a><\/li>\n<li><a href=\"#blogs\">Blogs<\/a><\/li>\n<li><a href=\"#case-studies\">Case Studies<\/a><\/li>\n<li><a href=\"#events\">Events<\/a><\/li>\n<li><a href=\"#books\">Books<\/a><\/li>\n<li><a href=\"#tools\">Tools<\/a><\/li>\n<li><a href=\"#courses\">Courses<\/a><\/li>\n<li><a href=\"#specific-topics\">Specific Topics<\/a><\/li>\n<\/ul><h3><a id=\"user-content-communities\" class=\"anchor\" aria-hidden=\"true\" href=\"#communities\"><\/a>Communities<\/h3>\n<ul><li>\n<p><strong><a href=\"https:\/\/www.indiehackers.com\/businesses\" rel=\"nofollow\">IndieHackers<\/a><\/strong><\/p>\n<p>Learn from profitable businesses and side projects. IndieHackers is probably the catalyst for this new wave of interest on making an independent income. It is a collection of interviews to successful developers about their products. IndieHackers has been especially interesting to follow since it's been an indie project itself. <a href=\"https:\/\/twitter.com\/csallen?lang=ca\" rel=\"nofollow\">Courtland Allen<\/a> has been very transparent about <a href=\"https:\/\/www.indiehackers.com\/blog\" rel=\"nofollow\">how he built it<\/a>. The last news make one think that the forum\/community will grow.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/barnacl.es\/\" rel=\"nofollow\">Barnacl.es<\/a><\/strong><\/p>\n<p>News for bootstrappers, not VC hackers. Hacker-News like forum. <a href=\"https:\/\/twitter.com\/pushcx\" rel=\"nofollow\">Peter Bhat Harkins<\/a> started it in early 2016 and I've found it of great value ever since. Also, it runs on <a href=\"https:\/\/github.com\/jcs\/lobsters\">great software<\/a>, which is always a thing to appreciate!<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/discuss.bootstrapped.fm\/\" rel=\"nofollow\">bootstrapped.fm Forum<\/a><\/strong><\/p>\n<p>A forum for bootstrappers. Started in 2013 by <a href=\"https:\/\/twitter.com\/ianlandsman\" rel=\"nofollow\">Ian Landsman<\/a> and <a href=\"https:\/\/twitter.com\/andrey_butov\" rel=\"nofollow\">Andrey Butov<\/a> as a continuation to the now defunct <a href=\"http:\/\/discuss.joelonsoftware.com\/default.asp?biz\" rel=\"nofollow\">The Business of Software<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/webchat.freenode.net\/?channels=%23%23passiveincome&amp;uio=d4\" rel=\"nofollow\">##passiveincome<\/a><\/strong> on freenode<\/p>\n<p><a href=\"https:\/\/news.ycombinator.com\/user?id=matbram\" rel=\"nofollow\">Matthew Mallard<\/a> created this channel on the freenode IRC network in early 2017 and it has been pretty active and growing ever since.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/news.ycombinator.com\/\" rel=\"nofollow\">HackerNews<\/a><\/strong><\/p>\n<p>News for hackers, mostly technology and entrepreneurship. Not specific to indie devs but a must for everyone interested in computer science, creating products and making money. A lot of indie devs lurking around. Started by <a href=\"https:\/\/twitter.com\/paulg\" rel=\"nofollow\">Paul Graham<\/a> and maintained by the seed accelerator YCombinator.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/productpeople.club\/\" rel=\"nofollow\">ProductPeople.club<\/a><\/strong> [$]<\/p>\n<p>Exclusive community for indie makers run by <a href=\"https:\/\/twitter.com\/mijustin\" rel=\"nofollow\">Justin Jackson<\/a>. Currently there's a waitlist. Cost is $249 for lifetime access. I'm not a member so I can't personally endorse it. Generally, I'm skeptic on products targeting indie makers but I still found it relevant to add this specific product because of cross references. Some people might find it interesting. Feel free to reference to reviews!<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.foundercafe.com\/\" rel=\"nofollow\">FounderCafe<\/a><\/strong> [$]<\/p>\n<p>FounderCafe is an invite-only, online community of entrepreneurs helping each other succeed run by <a href=\"https:\/\/twitter.com\/robwalling\" rel=\"nofollow\">Rob Walling<\/a> and <a href=\"https:\/\/twitter.com\/SingleFounder\" rel=\"nofollow\">Mike Taber<\/a>. $99 every 3 months. I'm not a member so I can't personally endorse it.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/nugget.one\/\" rel=\"nofollow\">Nugget.one<\/a><\/strong> [$]<\/p>\n<p>Nugget.one focuses on helping software developers launch successful side projects. Being part of the community gives access to case studios, bi-weekly webinars and live mentoring. Run by <a href=\"https:\/\/twitter.com\/justinvincent\" rel=\"nofollow\">Justin Vincent<\/a>. $19.99 per month.<\/p>\n<\/li>\n<\/ul><h5><a id=\"user-content-other\" class=\"anchor\" aria-hidden=\"true\" href=\"#other\"><\/a>Other<\/h5>\n<ul><li><strong><a href=\"https:\/\/www.reddit.com\/r\/Entrepreneur\/\" rel=\"nofollow\">r\/Entrepreneur<\/a><\/strong> - Not exclusive to indies\/bootstrappers but still with relevant content for us.<\/li>\n<li><strong><a href=\"https:\/\/bootstrapped.slack.com\" rel=\"nofollow\">Bootstrapped Slack<\/a><\/strong> (<a href=\"http:\/\/www.bootstrappedchat.com\/\" rel=\"nofollow\">invite here<\/a> invite link currently dead, working on it!) - Slack that probably originated from the bootstrapped.fm forum. Not really moderated.<\/li>\n<li><strong><a href=\"https:\/\/www.reddit.com\/r\/SideProject\/\" rel=\"nofollow\">r\/SideProject<\/a><\/strong> - Quite active<\/li>\n<li><strong><a href=\"https:\/\/sideprojectxyz.slack.com\" rel=\"nofollow\">Sideprojectxyz<\/a><\/strong> (<a href=\"http:\/\/sideproject.xyz\/\" rel=\"nofollow\">invite here<\/a>) - Slack community for people working on their sideprojects.<\/li>\n<li><strong><a href=\"https:\/\/www.reddit.com\/r\/indiebiz\/\" rel=\"nofollow\">r\/IndieBiz<\/a><\/strong> - A subreddit for small and independent businesses.<\/li>\n<li><strong><a href=\"https:\/\/www.reddit.com\/r\/EntrepreneurRideAlong\/\" rel=\"nofollow\">\/r\/EntrepreneurRideAlong<\/a><\/strong><\/li>\n<li><strong><a href=\"https:\/\/wip.chat\" rel=\"nofollow\">Work In Progress<\/a><\/strong> [$] - A public todo list for accountability, and a <a href=\"https:\/\/t.me\/wipchat\" rel=\"nofollow\">Telegram chat<\/a> of 800+ makers. Freemium.<\/li>\n<li><strong><a href=\"https:\/\/www.reddit.com\/r\/startups\/\" rel=\"nofollow\">r\/Startups<\/a><\/strong> - Not exclusive to indies\/bootstrappers but still with relevant content for us.<\/li>\n<\/ul><h3><a id=\"user-content-newsletters\" class=\"anchor\" aria-hidden=\"true\" href=\"#newsletters\"><\/a>Newsletters<\/h3>\n<ul><li>\n<p><strong><a href=\"https:\/\/bootstrappers.io\/\" rel=\"nofollow\">Bootstrappers.io<\/a><\/strong><\/p>\n<p>Weekly digest of the best user-submitted articles featured on bootstrapped.io. Curated by <a href=\"https:\/\/twitter.com\/ninjamonk\" rel=\"nofollow\">Darren Stuart<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.indiehackers.com\/businesses\" rel=\"nofollow\">IndieHackers<\/a><\/strong><\/p>\n<p>Newsletter containing the newest interviews of the site and also the best threads from the forum. Normally ships on Thursdays. By <a href=\"https:\/\/twitter.com\/csallen?lang=ca\" rel=\"nofollow\">Courtland Allen<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/justinjackson.ca\/indie\/\" rel=\"nofollow\">Indie<\/a><\/strong><\/p>\n<p>Stories, lessons, and inspiration for independent makers. By <a href=\"https:\/\/twitter.com\/mijustin\" rel=\"nofollow\">Justin Jackson<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/thesaasbootstrapper.com\/resources\/\" rel=\"nofollow\">The SaaS Bootstrapper<\/a><\/strong><\/p>\n<p>Content about bootstrapping, SaaS, growth, marketing, and other startup topics curated by <a href=\"http:\/\/macmartine.com\/\" rel=\"nofollow\">Mac Martine<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/hackerpreneur.co\/\" rel=\"nofollow\">Hackerpreneur<\/a><\/strong><\/p>\n<p>For indie entrepreneurs &amp; makers who live and work on their own terms. Every Friday, you\u2019ll get hand-picked startup resources &amp; essays exploring the future of work in the exponential age. By <a href=\"https:\/\/twitter.com\/MarcEglon\" rel=\"nofollow\">Marc Eglon<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/bootstrap.money\" rel=\"nofollow\">Bootstrap Money<\/a><\/strong><\/p>\n<p>Weekly roundups of stories, guides and resources to help you grow your internet business. Every Friday. Free.\nCurated by <a href=\"https:\/\/twitter.com\/ndethore\" rel=\"nofollow\">Nicolas de Thor\u00e9<\/a><\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-podcasts\" class=\"anchor\" aria-hidden=\"true\" href=\"#podcasts\"><\/a>Podcasts<\/h3>\n<ul><li>\n<p><strong><a href=\"http:\/\/www.startupsfortherestofus.com\/\" rel=\"nofollow\">Startups for the rest of us<\/a><\/strong><\/p>\n<p>Started in 2010 by <a href=\"https:\/\/twitter.com\/robwalling\" rel=\"nofollow\">Rob Walling<\/a> and <a href=\"https:\/\/twitter.com\/SingleFounder\" rel=\"nofollow\">Mike Taber<\/a> have covered almost everything there is to cover in micropreneurship. <a href=\"http:\/\/www.startupsfortherestofus.com\/greatest-hits\" rel=\"nofollow\">Here<\/a> you can find a 'best of' selection of their podcast.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/www.productpeople.tv\/\" rel=\"nofollow\">Product People<\/a><\/strong><\/p>\n<p>A podcast focused on great products and the people who make them. Since 2012. By <a href=\"https:\/\/twitter.com\/mijustin\" rel=\"nofollow\">Justin Jackson<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/bootstrapped.fm\/\" rel=\"nofollow\">bootstrapped.fm<\/a><\/strong><\/p>\n<p>Several different topics around bootstrapping online businesses. Started in 2013 by <a href=\"https:\/\/twitter.com\/ianlandsman\" rel=\"nofollow\">Ian Landsman<\/a> and <a href=\"https:\/\/twitter.com\/andrey_butov\" rel=\"nofollow\">Andrey Butov<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/bootstrappedweb.com\" rel=\"nofollow\">Bootstrapped Web<\/a><\/strong><\/p>\n<p>\"To share the lessons we learn by doing, and learn from the things that others are doing, to help us all get ahead as we bootstrap our businesses online.\" Since 2013 by <a href=\"https:\/\/twitter.com\/CasJam\" rel=\"nofollow\">Brian Casel<\/a> and <a href=\"https:\/\/twitter.com\/JordanGal\" rel=\"nofollow\">Jordan Gal<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/entreprogrammers.com\/\" rel=\"nofollow\">Entreprogrammers<\/a><\/strong><\/p>\n<p>A podcast for developreneurs. Run by <a href=\"https:\/\/twitter.com\/josh_earl\" rel=\"nofollow\">Josh Earl<\/a>, <a href=\"https:\/\/twitter.com\/jsonmez\" rel=\"nofollow\">John Sonmez<\/a>, <a href=\"https:\/\/twitter.com\/derickbailey\" rel=\"nofollow\">Derick Bailey<\/a> and <a href=\"https:\/\/twitter.com\/cmaxw\" rel=\"nofollow\">Charles Max Wood<\/a> since 2014.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.indiehackers.com\/businesses\" rel=\"nofollow\">IndieHackers<\/a><\/strong><\/p>\n<p>Podcast where <a href=\"https:\/\/twitter.com\/csallen?lang=ca\" rel=\"nofollow\">Courtland Allen<\/a> interviews some of the most successful indiehackers featured on the site. Started in early 2017 already has 12 episodes with great content and very carefully crafted related notes and links.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/thesaasbootstrapper.com\/\" rel=\"nofollow\">The SaaS Bootstrapper<\/a><\/strong><\/p>\n<p>Interviews with entrepreneurs and bootstrappers about bootstrapping, SaaS, growth, marketing, and other startup topics. By <a href=\"http:\/\/macmartine.com\/\" rel=\"nofollow\">Mac Martine<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.smartpassiveincome.com\/podcasts\/\" rel=\"nofollow\">Smart Passive Income<\/a><\/strong><\/p>\n<p>Not really specific to developers but with great content on the several types of products to make a living online. By <a href=\"https:\/\/twitter.com\/patflynn\" rel=\"nofollow\">Pat Flynn<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/giantrobots.fm\/\" rel=\"nofollow\">Giant Robots<\/a><\/strong><\/p>\n<p><a href=\"http:\/\/www.benorenstein.com\/\" rel=\"nofollow\">Ben Orenstein<\/a> (former growth\/product lead at Thoughtbot) and <a href=\"https:\/\/twitter.com\/derrickreimer\" rel=\"nofollow\">Derrick Reimer<\/a> (co-founder of Drip) talk weekly about growing their SaaS businesses.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/itunes.apple.com\/us\/podcast\/mixergy-startup-stories-1000-entrepreneurs-businesses\/id348690336?mt=2\" rel=\"nofollow\">Mixergy<\/a><\/strong><\/p>\n<p>Startup Interview podcast hosted by <a href=\"https:\/\/twitter.com\/andrewwarner\" rel=\"nofollow\">Andrew Warner<\/a>, who made $1 million a month from an online greeting cards business.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.npr.org\/podcasts\/510313\/how-i-built-this\" rel=\"nofollow\">How I Built This<\/a><\/strong><\/p>\n<p>A podcast hosted by <a href=\"https:\/\/twitter.com\/guyraz\" rel=\"nofollow\">Guy Raz<\/a> where startup founders tell their story.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-talks\" class=\"anchor\" aria-hidden=\"true\" href=\"#talks\"><\/a>Talks<\/h3>\n<ul><li>\n<p><strong><a href=\"https:\/\/www.youtube.com\/watch?v=0CDXJ6bMkMY\" rel=\"nofollow\">David Heinemeier Hansson at Startup School 08<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/dhh\" rel=\"nofollow\">David Heinemeier Hansson<\/a><\/p>\n<p>Bootstrapping as the better alternative to VC funded startups.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.youtube.com\/watch?v=i3d1asTrWUQ\" rel=\"nofollow\">Tyler Tringas on Storemapper and MicroSaaS<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/tylertringas\" rel=\"nofollow\">Tyler Tringas<\/a><\/p>\n<p>15 minutes filled with a lot of valuable information and insights where Tyler presents the concept of Micro-SaaS and tells us his story.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.youtube.com\/watch?v=m3nb_Qj3mRA\" rel=\"nofollow\">Bootstrapping Side Projects To Profit by Pieter Levels<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/levelsio\" rel=\"nofollow\">Pieter Levels<\/a><\/p>\n<p>Probably one of the most known indie hackers of the last couple of years talks about how to turn our sideprojects into profitable businesses. Very straight to the point, Pieter always provides a lot of insightful thoughts. By the way, don't get fooled by his lifestyle, his success is 500% work.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/businessofsoftware.org\/2011\/03\/hello-ladies-patrick-mckenzie-at-business-of-software-2010-marketing-to-minorities\/\" rel=\"nofollow\">Marketing for Minorities<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/patio11?lang=ca\" rel=\"nofollow\">Patrick McKenzie<\/a><\/p>\n<p>Lightning 7 minute talk by one of the most respected indie developers. The point Patrick is trying to make, is that solo developers can target way smaller and niche markets which might have low competition and be very profitable from a solo business point of view.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.youtube.com\/watch?v=J8UwcyYT3z0\" rel=\"nofollow\">Talking about CoderPad and business at Dropbox<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/fulligin?lang=ca\" rel=\"nofollow\">Vincent Woo<\/a><\/p>\n<p>Talk on building a successful small solo business.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.youtube.com\/watch?v=EZxP0i9ah8E\" rel=\"nofollow\">Paul Buchheit at Startup School 08<\/a><\/strong><\/p>\n<p>In this classic talk, AdSense\/GMail\/like button inventor explains how to listen to your users and why you should be critical of startup advice.<\/p>\n<\/li>\n<\/ul><h5><a id=\"user-content-other-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-1\"><\/a>Other<\/h5>\n<ul><li><strong><a href=\"https:\/\/vimeo.com\/user12790628\/videos\" rel=\"nofollow\">Talks from the MicroConf<\/a><\/strong><\/li>\n<li><strong><a href=\"http:\/\/baconbiz.com\/\" rel=\"nofollow\">Talks from the BaconBizConf<\/a><\/strong><\/li>\n<li><strong><a href=\"http:\/\/businessofsoftware.org\/2016\/07\/all-talks-from-business-of-software-conferences-in-one-place-saas-software-talks\/\" rel=\"nofollow\">Talks from Business of Software<\/a><\/strong><\/li>\n<\/ul><h3><a id=\"user-content-posts\" class=\"anchor\" aria-hidden=\"true\" href=\"#posts\"><\/a>Posts<\/h3>\n<ul><li>\n<p><strong><a href=\"https:\/\/levels.io\/12-startups-12-months\/\" rel=\"nofollow\">I\u2019m Launching 12 Startups in 12 Months<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/levelsio\" rel=\"nofollow\">Pieter Levels<\/a><\/p>\n<p>The challenge that started it all for Pieter. It inspired me to do something similar. What matters is to create and ship a lot, learning in the process, until something is worth pushing further.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/www.kalzumeus.com\/2011\/10\/28\/dont-call-yourself-a-programmer\/\" rel=\"nofollow\">Don't Call Yourself A Programmer, And Other Career Advice<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/patio11?lang=ca\" rel=\"nofollow\">Patrick McKenzie<\/a><\/p>\n<p>Patrick has written over 500 posts and I have obviously not read all of them, but of the many I've read, I chose this one...<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/www.kalzumeus.com\/2010\/03\/20\/running-a-software-business-on-5-hours-a-week\/\" rel=\"nofollow\">Running A Software Business On 5 Hours A Week<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/patio11?lang=ca\" rel=\"nofollow\">Patrick McKenzie<\/a><\/p>\n<p>...and this one too :D<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/tylertringas.com\/storemapper-bootstrapped-to-50000year-in-2-years-with-live-metrics\/\" rel=\"nofollow\">Storemapper: Bootstrapped to $50,000\/year in 2 years (with live metrics)<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/tylertringas\" rel=\"nofollow\">Tyler Tringas<\/a><\/p>\n<p>This is just an example of a business more than one would be happy with :-) Besides liking how Tyler shares his story building Storemapper, I think he always maximizes for valuable content.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/unicornfree.com\/2013\/why-you-should-do-a-tiny-product-first\" rel=\"nofollow\">Why You Should Do A Tiny Product First<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/amyhoy\" rel=\"nofollow\">Amy Hoy<\/a><\/p>\n<p>This is the first of a series by Amy Hoy. I'll be honest, I don't even remember what this was about, just that I found it worth my time. Oh wait, yeah, Amy told me not to code, at least at first. I haven't still gotten there...<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/medium.com\/@cliffordoravec\/the-epic-guide-to-bootstrapping-a-saas-startup-from-scratch-by-yourself-part-1-4d834e1df8c1\" rel=\"nofollow\">The Epic Guide to Bootstrapping a SaaS Startup from Scratch\u200a\u2014\u200aBy Yourself<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/cliffordoravec\" rel=\"nofollow\">Clifford Oravec<\/a><\/p>\n<p>Another series on what solo bootstrapping is about, with an offensive twist.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/medium.com\/product-people\/how-to-get-more-customers-54f64220e430\" rel=\"nofollow\">How to get more customers<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/mijustin\" rel=\"nofollow\">Justin Jackson<\/a>.<\/p>\n<p>This is a bit of a too specific topic for this general selection of posts, but the hell, you need to read it before it's too late. Ok, I'm starting to sound marketingy :D<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/hackernoon.com\/indie-startups-the-ingredients-of-success-74531fe3a019\" rel=\"nofollow\">Indie Startups: The Ingredients of Success<\/a><\/strong> by <a href=\"https:\/\/hackernoon.com\/@fpgaminer\" rel=\"nofollow\">fpgaminer<\/a><\/p>\n<p>Numbers with meaning. fpgaminer analyzed all the products featured on IndieHackers, extracted some statistics and got to some conclusions. Also worth reading for the approach.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/qotoqot.com\/blog\/founder-skills\/\" rel=\"nofollow\">What you should know as a founder of a software company<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/ivmirx\" rel=\"nofollow\">Ivan Mir<\/a><\/p>\n<p>A guide to the complex world of modern software business: everything from UX design to marketing and writing, with a list of resources to start with each topic.<\/p>\n<\/li>\n<\/ul><h5><a id=\"user-content-others\" class=\"anchor\" aria-hidden=\"true\" href=\"#others\"><\/a>Others<\/h5>\n<p>There's obviously a million posts on the topic. I'm sure this is not a good selection but it'll get you started. Here you can find some more:<\/p>\n<ul><li><strong><a href=\"https:\/\/bootstrappers.io\/\" rel=\"nofollow\">Bootstrappers.io<\/a><\/strong><\/li>\n<li><strong><a href=\"https:\/\/hackernoon.com\/\" rel=\"nofollow\">HackerNoon<\/a><\/strong><\/li>\n<li><strong><a href=\"https:\/\/news.ycombinator.com\/\" rel=\"nofollow\">HackerNews<\/a><\/strong><\/li>\n<li><strong><a href=\"https:\/\/hackernoon.com\/7-articles-for-indie-devs-to-start-shipping-12b22f5e5660\" rel=\"nofollow\">7 Articles for Indie Devs to Start Shipping<\/a><\/strong> - great selection by <a href=\"https:\/\/twitter.com\/mskvsk\" rel=\"nofollow\">Alex Moskovski<\/a><\/li>\n<\/ul><h3><a id=\"user-content-blogs\" class=\"anchor\" aria-hidden=\"true\" href=\"#blogs\"><\/a>Blogs<\/h3>\n<ul><li>\n<p><strong><a href=\"http:\/\/www.kalzumeus.com\/\" rel=\"nofollow\">Patrick Mackenzie<\/a><\/strong><\/p>\n<p>Over 500 posts that show how Patrick went from being a software engineer to running successful software businesses.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/www.softwarebyrob.com\/\" rel=\"nofollow\">Rob Walling<\/a><\/strong><\/p>\n<p>Lessons Learned by a Solo Entrepreneur.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/levels.io\/\" rel=\"nofollow\">Pieter Levels<\/a><\/strong><\/p>\n<p>Lots of advice on bootstrapping businesses if you dig between Pieter's multiple interests.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/tylertringas.com\/\" rel=\"nofollow\">Tyler Tringas<\/a><\/strong><\/p>\n<p>On SaaS and his journey with Storemapper.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.joelonsoftware.com\/\" rel=\"nofollow\">Joel Spolsky<\/a><\/strong><\/p>\n<p>Software and business have always been topics Joel has covered. StackOverflow and Trello prove he knows what he's talking about.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/m.signalvnoise.com\/\" rel=\"nofollow\">Signal v. Noise<\/a><\/strong><\/p>\n<p>The blog of the basecamp people, mostly David Heinemeier Hansson and Jason Fried, on business and how to run bootstrapped companies.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/blog.pinboard.in\/\" rel=\"nofollow\">Pinboard<\/a><\/strong><\/p>\n<p>Maciej Ceg\u0142owski's blog for his bookmarking tool, pinboard, useful in showing the growth of a successful solo business.<\/p>\n<\/li>\n<\/ul><h5><a id=\"user-content-others-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#others-1\"><\/a>Others<\/h5>\n<ul><li><strong><a href=\"https:\/\/thefoundation.com\/blog\" rel=\"nofollow\">The Foundation<\/a><\/strong> - Turning the process of creating a B2B SaaS business into a formula.<\/li>\n<li><strong><a href=\"http:\/\/www.priceintelligently.com\/blog\" rel=\"nofollow\">Price Intelligently<\/a><\/strong> - Blog focused on the economics of the SaaS business model and SaaS pricing.<\/li>\n<li><strong><a href=\"https:\/\/moz.com\/blog\" rel=\"nofollow\">Moz Blog<\/a><\/strong> - For anything related to SEO.<\/li>\n<li><strong><a href=\"http:\/\/www.paulgraham.com\/articles.html\" rel=\"nofollow\">Paul Graham<\/a><\/strong> - A lot of content for big funded startups, but still relevant to bootstrappers.<\/li>\n<li><strong><a href=\"https:\/\/john.onolan.org\/\" rel=\"nofollow\">John O'Nolan<\/a><\/strong> - Especially the older posts on how he started ghost.<\/li>\n<\/ul><h3><a id=\"user-content-case-studies\" class=\"anchor\" aria-hidden=\"true\" href=\"#case-studies\"><\/a>Case Studies<\/h3>\n<ul><li>\n<p><strong><a href=\"https:\/\/www.indiehackers.com\/blog\" rel=\"nofollow\">IndieHackers<\/a><\/strong><\/p>\n<p>From capturing a trend to getting acquired in less than one year.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/levels.io\/product-hunt-hacker-news-number-one\/\" rel=\"nofollow\">NomadList<\/a><\/strong><\/p>\n<p>Solving your own pain, representing a movement, moving fast and working hard by a serial bootstrapper.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.bugmuncher.com\/blog\/\" rel=\"nofollow\">Bugmuncher<\/a><\/strong><\/p>\n<p>Transparent growth and financial reports by a solo saas founder.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/37signals.com\/bootstrapped\" rel=\"nofollow\">37signals<\/a><\/strong><\/p>\n<p>A collection of bootstrapped companies by the people at 37 signals.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/baremetrics.com\/blog\" rel=\"nofollow\">Baremetrics<\/a><\/strong><\/p>\n<p>An example of riding the wave.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/themakingof.carrd.co\/#\" rel=\"nofollow\">Carrd.co<\/a><\/strong><\/p>\n<p><a href=\"https:\/\/twitter.com\/ajlkn\" rel=\"nofollow\">AJ<\/a> writes about the making of carrd.co and the different steps that he took on the way to a successful product.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.candyjapan.com\/behind-the-scenes\" rel=\"nofollow\">CandyJapan<\/a><\/strong><\/p>\n<p>Revenue numbers and learnings from running a Japanese candy subscription box.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-events\" class=\"anchor\" aria-hidden=\"true\" href=\"#events\"><\/a>Events<\/h3>\n<ul><li>\n<p><strong><a href=\"http:\/\/www.microconf.com\/\" rel=\"nofollow\">MicroConf<\/a><\/strong> [$]<\/p>\n<p>\"The worlds biggest conference for the worlds smallest self-funded software companies\". It has two editions, one more focused to starters, and another one focused on growth. Both last two days. Las Vegas. Run by <a href=\"https:\/\/twitter.com\/robwalling\" rel=\"nofollow\">Rob Walling<\/a> and <a href=\"https:\/\/twitter.com\/SingleFounder\" rel=\"nofollow\">Mike Taber<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/www.microconfeurope.com\/\" rel=\"nofollow\">MicroConf Europe<\/a><\/strong> [$]<\/p>\n<p>MicroConf's European edition. Barcelona. Run by <a href=\"https:\/\/twitter.com\/robwalling\" rel=\"nofollow\">Rob Walling<\/a> and <a href=\"https:\/\/twitter.com\/SingleFounder\" rel=\"nofollow\">Mike Taber<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/baconbiz.com\/\" rel=\"nofollow\">BaconBizConf<\/a><\/strong> [$]<\/p>\n<p>\"An alliance for B2B bootstrappers\". Philadelphia. By <a href=\"https:\/\/twitter.com\/amyhoy\" rel=\"nofollow\">Amy Hoy<\/a> and <a href=\"https:\/\/twitter.com\/alexhillman\" rel=\"nofollow\">Alex Hillman<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/www.startupschool.org\/\" rel=\"nofollow\">StartupSchool<\/a><\/strong><\/p>\n<p>Free event organized by seed accelerator <a href=\"http:\/\/www.ycombinator.com\/\" rel=\"nofollow\">YCombinator<\/a>. You have to be a bit lucky to get in. Also, this suggestion goes a little bit against the idea of \"bootstrapped\" but YCombinator has a lot of good resources for us indie hackers!<\/p>\n<\/li>\n<\/ul><h5><a id=\"user-content-not-specific-to-indie-founders\" class=\"anchor\" aria-hidden=\"true\" href=\"#not-specific-to-indie-founders\"><\/a>Not specific to indie founders:<\/h5>\n<ul><li>\n<p><strong><a href=\"http:\/\/businessofsoftware.org\/\" rel=\"nofollow\">Business of Software<\/a><\/strong> [$]<\/p>\n<p>\"Business of Software is not based in Silicon Valley for good reason. It is about building long term, sustainable software businesses. \u2026a thoughtful conference, where people come to listen and learn, not self promote and shout\". Boston. <a href=\"http:\/\/thebln.com\/team\/\" rel=\"nofollow\">Team behind BoS<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/businessofsoftware.eu\/\" rel=\"nofollow\">Business of Software Europe<\/a><\/strong> [$]<\/p>\n<p>BoS European edition. London. <a href=\"http:\/\/thebln.com\/team\/\" rel=\"nofollow\">Team behind BoS<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/b2brocks.co\/\" rel=\"nofollow\">B2B Rocks<\/a><\/strong> [$]<\/p>\n<p>\"Leading conferences for B2B and SaaS startups\". Popular business model for indie devs. Paris. Sydney.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"http:\/\/peersconf.com\/\" rel=\"nofollow\">PeersConf<\/a><\/strong> [$]<\/p>\n<p>\"A conference for those who make the web, and make it their business\". Seattle (changes every year). By <a href=\"https:\/\/twitter.com\/justjessdc\" rel=\"nofollow\">Jessica D'Amico<\/a>.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-books\" class=\"anchor\" aria-hidden=\"true\" href=\"#books\"><\/a>Books<\/h3>\n<ul><li><strong><a href=\"https:\/\/www.amazon.com\/Start-Small-Stay-Developers-Launching-ebook\/dp\/B003YH9MMI\/ref=sr_1_1?s=digital-text&amp;ie=UTF8&amp;qid=1493983910&amp;sr=1-1&amp;keywords=start+small+stay+small\" rel=\"nofollow\">Start Small, Stay Small<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/robwalling\" rel=\"nofollow\">Rob Walling<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/tylertringas.com\/micro-saas-ebook\/\" rel=\"nofollow\">Building Micro-SaaS Businesses<\/a><\/strong> by <a href=\"https:\/\/twitter.com\/tylertringas\" rel=\"nofollow\">Tyler Tringas<\/a>. - A collection of very well structured posts that cover the most initial struggles of an indie dev on the way to build a MicroSaaS. I can personally endorse it and I'm eager to read the yet to be written chapters :)<\/li>\n<li><strong><a href=\"http:\/\/www.singlefounderhandbook.com\/\" rel=\"nofollow\">The Single Founder Handbook<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/SingleFounder\" rel=\"nofollow\">Mike Taber<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/gettingreal.37signals.com\/\" rel=\"nofollow\">Getting Real<\/a><\/strong> by <a href=\"https:\/\/37signals.com\" rel=\"nofollow\">37signals<\/a>. - More focused on building web applications, I really recommend it. Short, straight to the point, very easy to read and full of valuable content.<\/li>\n<li><strong><a href=\"https:\/\/makebook.io\/\" rel=\"nofollow\">MAKE<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/levelsio\" rel=\"nofollow\">Pieter Levels<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/37signals.com\/rework\" rel=\"nofollow\">Rework<\/a><\/strong> [$] by <a href=\"https:\/\/37signals.com\" rel=\"nofollow\">37signals<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/unicornfree.com\/just-fucking-ship\/\" rel=\"nofollow\">Just Fucking Ship<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/amyhoy\" rel=\"nofollow\">Amy Hoy<\/a> and <a href=\"https:\/\/twitter.com\/alexhillman\" rel=\"nofollow\">Alex Hillman<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/justinjackson.withcoach.com\/marketing-for-developers-the-book\" rel=\"nofollow\">Marketing for Developers<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/mijustin\" rel=\"nofollow\">Justin Jackson<\/a>.<\/li>\n<li><strong><a href=\"http:\/\/www.paulgraham.com\/hackpaint.html\" rel=\"nofollow\">Hackers&amp;Painters<\/a><\/strong> [$] by <a href=\"http:\/\/www.paulgraham.com\/\" rel=\"nofollow\">Paul Graham<\/a>.<\/li>\n<\/ul><h5><a id=\"user-content-other-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-2\"><\/a>Other<\/h5>\n<ul><li><strong><a href=\"https:\/\/groundera.com\/\" rel=\"nofollow\">Groundera.com<\/a><\/strong> - A collection of indie books, most of them free, for entrepreneurs. By <a href=\"https:\/\/groundera.com\/about\" rel=\"nofollow\">Philipp<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/books.makesideproject.com\/\" rel=\"nofollow\">SideProject Books<\/a><\/strong> - Side Project Books is a curated collection of books for anyone interested in building a business on the side. By <a href=\"https:\/\/twitter.com\/mubashariqbal\" rel=\"nofollow\">Mubashar Iqbal<\/a><\/li>\n<li><strong><a href=\"http:\/\/thesaasbootstrapper.com\/book-list\/\" rel=\"nofollow\">The SaaS Bootstrapper book list<\/a><\/strong> - Collection curated by <a href=\"http:\/\/macmartine.com\/\" rel=\"nofollow\">Mac Martine<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/www.amazon.com\/Badass-Making-Awesome-Kathy-Sierra-ebook\/dp\/B00VAUIM18\/ref=sr_1_1?s=books&amp;ie=UTF8&amp;qid=1493985078&amp;sr=1-1&amp;keywords=badass+kathy+sierra\" rel=\"nofollow\">Badass: Making Users Awesome<\/a><\/strong> [$] by <a href=\"http:\/\/seriouspony.com\/\" rel=\"nofollow\">Kathy Sierra<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/www.amazon.com\/dp\/B00SI0B5FS\/ref=dp-kindle-redirect?_encoding=UTF8&amp;btkr=1\" rel=\"nofollow\">Anything You Want<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/sivers\" rel=\"nofollow\">Derek Sivers<\/a>.<\/li>\n<li><strong><a href=\"https:\/\/www.amazon.com\/Founders-Work-Stories-Startups-Early-ebook\/dp\/B009IXMK4O\/ref=sr_1_1?s=digital-text&amp;ie=UTF8&amp;qid=1493985002&amp;sr=1-1&amp;keywords=founders+at+work\" rel=\"nofollow\">Founders at Work<\/a><\/strong> [$] by <a href=\"https:\/\/twitter.com\/jesslivingston\" rel=\"nofollow\">Jessica Livingston<\/a>.<\/li>\n<li><strong><a href=\"http:\/\/www.sideprojectbook.com\/\" rel=\"nofollow\">SideProject Book<\/a><\/strong> - [$49].<\/li>\n<\/ul><h3><a id=\"user-content-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#tools\"><\/a>Tools<\/h3>\n<ul><li>\n<p><strong><a href=\"http:\/\/thesaasbootstrapper.com\/saas-startup-resources\/\" rel=\"nofollow\">The SaaS Bootstrapper Startup Resources<\/a><\/strong><\/p>\n<p>A collection of resources for bootstrappers put together by <a href=\"http:\/\/macmartine.com\/\" rel=\"nofollow\">Mac Martine<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/ripienaar\/free-for-dev\">free-for-dev<\/a><\/strong><\/p>\n<p>A list of SaaS, PaaS and IaaS offerings that have free tiers of interest to devops and infradev. Very useful for bootstrappers to keep initial costs low ;) By <a href=\"https:\/\/github.com\/ripienaar\/\">R. I. Pienaar<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/255kb\/stack-on-a-budget\">Stack-on-a-budget<\/a><\/strong><\/p>\n<p>Another list of SaaS, PaaS, and Iaas offerings that have free tiers for developers on a budget. Very useful for exploring ideas and building out proof of concept. By <a href=\"https:\/\/github.com\/255kb\/\">Guillaume (256kb)<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade\">Tools Of The Trade<\/a><\/strong><\/p>\n<p>Spectacular list of categorized tools for developers resulting from HackerNews threads. By <a href=\"https:\/\/twitter.com\/chrisbarber\" rel=\"nofollow\">Chris Barber<\/a> and <a href=\"https:\/\/twitter.com\/davisonio\" rel=\"nofollow\">Craig Davison<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\">awesome-selfhosted<\/a><\/strong><\/p>\n<p>\"This is a list of Free Software network services and web applications which can be hosted locally. Selfhosting is the process of locally hosting and managing applications instead of renting from SaaS providers.\" Lots of salvageable software :P by <a href=\"https:\/\/github.com\/Kickball\">Edward D<\/a>.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-courses\" class=\"anchor\" aria-hidden=\"true\" href=\"#courses\"><\/a>Courses<\/h3>\n<p>Again, I'm not quite fond of listing [$] products made by indie developers targeting indie developers, especially because I haven't tried them myself, but due to several references I assume they are probably worth it and some people might find value in learning about them.<\/p>\n<ul><li>\n<p><strong><a href=\"https:\/\/devmarketing.xyz\/\" rel=\"nofollow\">Marketing for Developers<\/a><\/strong> [$]<\/p>\n<p>\"An online marketing course for software developers &amp; product people\". Read some reviews <a href=\"https:\/\/www.producthunt.com\/posts\/marketing-for-developers\" rel=\"nofollow\">here<\/a>. By <a href=\"https:\/\/twitter.com\/mijustin\" rel=\"nofollow\">Justin Jackson<\/a>.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/30x500.com\/academy\/\" rel=\"nofollow\">30x500 Academy<\/a><\/strong> [$]<\/p>\n<p>\"Pick your customer. Learn how they tick. Learn what they need, want, and are ready to buy.\" Find reviews <a href=\"http:\/\/codefol.io\/posts\/30x500-why\" rel=\"nofollow\">here<\/a>(good), <a href=\"https:\/\/news.ycombinator.com\/item?id=8119354\" rel=\"nofollow\">here<\/a>(bad) and <a href=\"https:\/\/www.quora.com\/How-do-30x500-and-How-To-Make-Your-First-Dollar-compare\" rel=\"nofollow\">here<\/a>. By <a href=\"https:\/\/twitter.com\/amyhoy\" rel=\"nofollow\">Amy Hoy<\/a> and <a href=\"https:\/\/twitter.com\/alexhillman\" rel=\"nofollow\">Alex Hillman<\/a>.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-specific-topics\" class=\"anchor\" aria-hidden=\"true\" href=\"#specific-topics\"><\/a>Specific topics<\/h2>\n<p>In the future, I'd like to have a list of resources for more specific topics or FAQs such as:<\/p>\n<ul><li>'how to get ideas'<\/li>\n<\/ul><ul><li>'how to find betatesters'<\/li>\n<li>'how to find my first 100 users'<\/li>\n<\/ul><ul><li>'how to validate my idea'<\/li>\n<\/ul><ul><li>'where to advertise my project'<\/li>\n<li>'what paid ads work better'<\/li>\n<li>'i've got 100 users, how to grow'<\/li>\n<li>'how do i deal with support'<\/li>\n<li>'types of digital products'<\/li>\n<li>'legal aspects'<\/li>\n<\/ul><ul><li>'typical mistakes'<\/li>\n<\/ul><p>Help me make a proper list of specific topics!<\/p>\n<\/article>","created_at":"2019-07-11T17:08:54+0000","updated_at":"2019-07-11T17:09:08+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/40381643?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":978,"title":"tonsky\/uberdeps","url":"https:\/\/github.com\/tonsky\/uberdeps\/blob\/master\/README.md","content":"<p>Takes deps.edn and packs an uberjar out of it.<\/p>\n<ul><li>Fast: Does not unpack intermediate jars on disk.<\/li>\n<li>Explicit: Prints dependency tree. Realize how much crap you\u2019re packing.<\/li>\n<li>Standalone: does not depend on current classpath, does not need live Clojure environment.<\/li>\n<li>Embeddable and configurable: fine-tune your build by combining config options and calling specific steps from your code.<\/li>\n<\/ul><h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>Add to your bash aliases:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>clj -Sdeps '{:deps {uberdeps {:mvn\/version \"0.1.4\"}}}' -m uberdeps.uberjar<\/pre><\/div>\n<p>Or add to your <code>deps.edn<\/code> or <code>~\/.clojure\/deps.edn<\/code>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:aliases {\n  :uberjar {:extra-deps {uberdeps {:mvn\/version \"0.1.4\"}}\n            :main-opts [\"-m\" \"uberdeps.uberjar\"]}\n}<\/pre><\/div>\n<p>Supported command-line options are:<\/p>\n<pre>--deps-file &lt;file&gt;                Which deps.edn file to use to build classpath. Defaults to 'deps.edn'\n--aliases &lt;alias:alias:...&gt;       Colon-separated list of alias names to include from deps file. Defaults to nothing\n--target &lt;file&gt;                   Jar file to ouput to. Defaults to 'target\/&lt;directory-name&gt;.jar'\n--level (debug|info|warn|error)   Verbose level. Defaults to debug\n<\/pre>\n<h2><a id=\"user-content-programmatic-api\" class=\"anchor\" aria-hidden=\"true\" href=\"#programmatic-api\"><\/a>Programmatic API<\/h2>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[uberdeps.api :as uberdeps])\n(let [exclusions (into uberdeps\/exclusions [#\"\\.DS_Store\" #\".*\\.cljs\" #\"cljsjs\/.*\"])\n      deps       (clojure.edn\/read-string (slurp \"deps.edn\"))]\n  (binding [uberdeps\/exclusions exclusions\n            uberdeps\/level      :warn]\n    (uberdeps\/package deps \"target\/uber.jar\" {:aliases #{:uberjar}})))<\/pre><\/div>\n<h2><a id=\"user-content-using-the-generated-uberjar\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-the-generated-uberjar\"><\/a>Using the generated uberjar<\/h2>\n<p>If your project has a <code>-main<\/code> function, you can run it from within the generated uberjar:<\/p>\n<pre>java -cp target\/&lt;your project&gt;.jar clojure.main -m &lt;your namespace with main&gt;\n<\/pre>\n<h2><a id=\"user-content-changelog\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelog\"><\/a>Changelog<\/h2>\n<h3><a id=\"user-content-014---june-5-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#014---june-5-2019\"><\/a>0.1.4 - June 5, 2019<\/h3>\n<ul><li>Package paths before jars so that local files take priority over deps #9<\/li>\n<\/ul><h3><a id=\"user-content-013---may-30-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#013---may-30-2019\"><\/a>0.1.3 - May 30, 2019<\/h3>\n<ul><li>Fixed NPE when target is located in current dir #7<\/li>\n<\/ul><h3><a id=\"user-content-012---may-27-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#012---may-27-2019\"><\/a>0.1.2 - May 27, 2019<\/h3>\n<ul><li>Make target dirs if don\u2019t exist #4<\/li>\n<\/ul><h3><a id=\"user-content-011---may-3-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#011---may-3-2019\"><\/a>0.1.1 - May 3, 2019<\/h3>\n<ul><li>Normalize dependencies without namespaces #3<\/li>\n<\/ul><h3><a id=\"user-content-010---april-29-2019\" class=\"anchor\" aria-hidden=\"true\" href=\"#010---april-29-2019\"><\/a>0.1.0 - April 29, 2019<\/h3>\n<ul><li>Initial version<\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2019 Nikita Prokopov<\/p>\n<p>Licensed under MIT (see <a href=\"https:\/\/github.com\/tonsky\/uberdeps\/blob\/master\/LICENSE\">LICENSE<\/a>).<\/p>","created_at":"2019-07-10T11:19:25+0000","updated_at":"2019-07-10T11:19:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/285292?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":977,"title":"mcohen01\/amazonica","url":"https:\/\/github.com\/mcohen01\/amazonica\/blob\/master\/README.md","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/mcohen01\/amazonica\/blob\/master\/claws.png\"><img src=\"https:\/\/github.com\/mcohen01\/amazonica\/raw\/master\/claws.png\" alt=\"AWS logo\" \/><\/a><\/p>\n<p>A comprehensive Clojure client for the entire <a href=\"http:\/\/aws.amazon.com\/documentation\/\" rel=\"nofollow\">Amazon AWS api<\/a>.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>Leiningen coordinates:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>[amazonica \"0.3.143\"]<\/pre><\/div>\n<p>For Maven users:<\/p>\n<p>add the following repository definition to your <code>pom.xml<\/code>:<\/p>\n<div class=\"highlight highlight-text-xml\"><pre>&lt;repository&gt;\n  &lt;id&gt;clojars.org&lt;\/id&gt;\n  &lt;url&gt;http:\/\/clojars.org\/repo&lt;\/url&gt;\n&lt;\/repository&gt;<\/pre><\/div>\n<p>and the following dependency:<\/p>\n<div class=\"highlight highlight-text-xml\"><pre>&lt;dependency&gt;\n  &lt;groupId&gt;amazonica&lt;\/groupId&gt;\n  &lt;artifactId&gt;amazonica&lt;\/artifactId&gt;\n  &lt;version&gt;0.3.143&lt;\/version&gt;\n&lt;\/dependency&gt;<\/pre><\/div>\n<h2><a id=\"user-content-supported-services\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-services\"><\/a>Supported Services<\/h2>\n<ul><li>Api Gateway<\/li>\n<li><a href=\"#autoscaling\">Autoscaling<\/a><\/li>\n<li><a href=\"#batch\">Batch<\/a><\/li>\n<li>Budgets<\/li>\n<li>CloudDirectory<\/li>\n<li><a href=\"#cloudformation\">CloudFormation<\/a><\/li>\n<li><a href=\"#cloudfront\">CloudFront<\/a><\/li>\n<li><a href=\"#cloudsearch\">CloudSearch<\/a><\/li>\n<li><a href=\"#cloudsearchv2\">CloudSearchV2<\/a><\/li>\n<li><a href=\"#cloudsearchdomain\">CloudSearchDomain<\/a><\/li>\n<li><a href=\"#cloudwatch\">CloudWatch<\/a><\/li>\n<li><a href=\"#cloudwatchevents\">CloudWatchEvents<\/a><\/li>\n<li>CodeCommit<\/li>\n<li><a href=\"#codedeploy\">CodeDeploy<\/a><\/li>\n<li>CodePipeline<\/li>\n<li>CodeStar<\/li>\n<li><a href=\"#cognitoidentityproviders\">CognitoIdentityProviders<\/a><\/li>\n<li><a href=\"#comprehend\">Comprehend<\/a><\/li>\n<li>Config<\/li>\n<li>CostAndUsageReport<\/li>\n<li>CostExplorer<\/li>\n<li>DatabaseMigrationService<\/li>\n<li><a href=\"#datapipeline\">DataPipeline<\/a><\/li>\n<li>Dax<\/li>\n<li>DeviceFarm<\/li>\n<li>DirectConnect<\/li>\n<li>Directory<\/li>\n<li><a href=\"#dynamodbv2\">DynamoDBV2<\/a><\/li>\n<li><a href=\"#ec2\">EC2<\/a><\/li>\n<li><a href=\"#ecr\">ECR<\/a><\/li>\n<li><a href=\"#ecs\">ECS<\/a><\/li>\n<li><a href=\"#elasticache\">ElastiCache<\/a><\/li>\n<li><a href=\"#elasticbeanstalk\">ElasticBeanstalk<\/a><\/li>\n<li>ElasticFileSystem<\/li>\n<li><a href=\"#elasticloadbalancing\">ElasticLoadBalancing<\/a><\/li>\n<li><a href=\"#elasticmapreduce\">ElasticMapReduce<\/a><\/li>\n<li><a href=\"#elasticsearch\">Elasticsearch<\/a><\/li>\n<li><a href=\"#elastictranscoder\">ElasticTranscoder<\/a><\/li>\n<li>GameLift<\/li>\n<li><a href=\"#glacier\">Glacier<\/a><\/li>\n<li>Glue<\/li>\n<li>GreenGrass<\/li>\n<li>GuardDuty<\/li>\n<li><a href=\"#identitymanagement\">IdentityManagement<\/a><\/li>\n<li>ImportExport<\/li>\n<li><a href=\"#iot\">IoT<\/a><\/li>\n<li><a href=\"#kinesis\">Kinesis<\/a><\/li>\n<li><a href=\"#kinesisanalytics\">Kinesis Analytics<\/a><\/li>\n<li><a href=\"#kinesisfirehose\">KinesisFirehose<\/a><\/li>\n<li><a href=\"#kms\">KMS<\/a><\/li>\n<li><a href=\"#logs\">Logs<\/a><\/li>\n<li><a href=\"#lambda\">Lambda<\/a><\/li>\n<li>MachineLearning<\/li>\n<li>MediaConvert<\/li>\n<li>MediaLive<\/li>\n<li>MediaPackage<\/li>\n<li>MediaStore<\/li>\n<li>MigrationHub<\/li>\n<li>Mobile<\/li>\n<li>MQ<\/li>\n<li>MechanicalTurk<\/li>\n<li><a href=\"#opsworks\">OpsWorks<\/a><\/li>\n<li><a href=\"#pinpoint\">Pinpoint<\/a><\/li>\n<li>Pricing<\/li>\n<li>Polly<\/li>\n<li>RDS<\/li>\n<li><a href=\"#redshift\">Redshift<\/a><\/li>\n<li><a href=\"#route53\">Route53<\/a><\/li>\n<li><a href=\"#route53domains\">Route53Domains<\/a><\/li>\n<li><a href=\"#s3\">S3<\/a><\/li>\n<li>ServerMigration<\/li>\n<li>ServiceCatalog<\/li>\n<li><a href=\"#simpledb\">SimpleDB<\/a><\/li>\n<li><a href=\"#simpleemail\">SimpleEmail<\/a><\/li>\n<li><a href=\"#SimpleSystemsManager\">SimpleSystemsManager<\/a><\/li>\n<li><a href=\"#simpleworkflow\">SimpleWorkflow<\/a><\/li>\n<li><a href=\"#sns\">SNS<\/a><\/li>\n<li><a href=\"#sqs\">SQS<\/a><\/li>\n<li><a href=\"#stepfunctions\">StepFunctions<\/a><\/li>\n<li>StorageGateway<\/li>\n<li>WAF<\/li>\n<li>Workspaces<\/li>\n<\/ul><h2><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h2>\n<p><a href=\"http:\/\/blog.fogus.me\/2012\/08\/23\/minimum-viable-snippet\/\" rel=\"nofollow\">Minimum Viable Snippet<\/a>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.ec2]))\n(describe-instances)\n(create-snapshot :volume-id   \"vol-8a4857fa\"\n                 :description \"my_new_snapshot\")<\/pre><\/div>\n<p>Amazonica reflectively delegates to the Java client library, as such it supports the complete set of remote service calls implemented by each of the service-specific AWS client classes (e.g. AmazonEC2Client, AmazonS3Client, etc.), the documentation for which can be found in the <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/index.html\" rel=\"nofollow\">AWS Javadocs<\/a>. <a href=\"https:\/\/cljdoc.org\/d\/amazonica\/amazonica\/CURRENT\/api\/amazonica\" rel=\"nofollow\">cljdoc function references<\/a> are also available.<\/p>\n<p>Reflection is used to create idiomatically named Clojure Vars in the library namespaces corresponding to the AWS service. camelCase Java methods become lower-case, hyphenated Clojure functions. So for example, if you want to create a snapshot of a running EC2 instance, you'd simply<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(create-snapshot :volume-id \"vol-8a4857fa\"\n                 :description \"my_new_snapshot\")<\/pre><\/div>\n<p>which delegates to the <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/AmazonEC2Client.html#createSnapshot(com.amazonaws.services.ec2.model.CreateSnapshotRequest)\" rel=\"nofollow\">createSnapshot()<\/a> method of AmazonEC2Client. If the Java method on the Amazon*Client takes a parameter, such as <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/model\/CreateSnapshotRequest.html\" rel=\"nofollow\">CreateSnapshotRequest<\/a> in this case, the bean properties exposed via mutators of the form set* can be supplied as key-value pairs passed as arguments to the Clojure function.<\/p>\n<p>All of the AWS Java apis (except S3) follow this pattern, either having a single implementation method which takes an AWS Java bean as its only argument, or being overloaded and having a no-arg implementation. The corresponding Clojure function will either require key-value pairs as arguments, or be variadic and allow a no-arg invocation.<\/p>\n<p>For example, AmazonEC2Client's <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/AmazonEC2Client.html#describeImages()\" rel=\"nofollow\">describeImages()<\/a> method is overloaded, and can be invoked either with no args, or with a <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/model\/DescribeImagesRequest.html\" rel=\"nofollow\">DescribeImagesRequest<\/a>. So the Clojure invocation would look like<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(describe-images)<\/pre><\/div>\n<p>or<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(describe-images :owners [\"self\"]\n                 :image-ids [\"ami-f00f9699\" \"ami-e0d30c89\"])<\/pre><\/div>\n<h3><a id=\"user-content-conversion-of-returned-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#conversion-of-returned-types\"><\/a>Conversion of Returned Types<\/h3>\n<p><code>java.util.Collections<\/code> are converted to the corresponding Clojure collection type. <code>java.util.Maps<\/code> are converted to <code>clojure.lang.IPersistentMaps<\/code>, <code>java.util.Lists<\/code> are converted to <code>clojure.lang.IPersistentVectors<\/code>, etc.<\/p>\n<p><code>java.util.Dates<\/code> are automatically converted to Joda Time <code>DateTime<\/code> instances.<\/p>\n<p>Amazon AWS object types are returned as Clojure maps, with conversion taking place recursively, so, \"Clojure data all the way down.\"<\/p>\n<p>For example, a call to<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(describe-instances)<\/pre><\/div>\n<p>invokes a Java method on AmazonEC2Client which returns a <code>com.amazonaws.services.ec2.model.DescribeInstancesResult<\/code>. However, this is recursively converted to Clojure data, yielding a map of <code>Reservations<\/code>, like so:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:owner-id \"676820690883\",\n   :group-names [\"cx\"],\n   :groups [{:group-name \"cx\", :group-id \"sg-38f45150\"}],\n   :instances\n   [{:instance-type \"m1.large\",\n     :kernel-id \"aki-825ea7eb\",\n     :hypervisor \"xen\",\n     :state {:name \"running\", :code 16},\n     :ebs-optimized false,\n     :public-dns-name \"ec2-154-73-176-213.compute-1.amazonaws.com\",\n     :root-device-name \"\/dev\/sda1\",\n     :virtualization-type \"paravirtual\",\n     :root-device-type \"ebs\",\n     :block-device-mappings\n     [{:device-name \"\/dev\/sda1\",\n       :ebs\n       {:status \"attached\",\n        :volume-id \"vol-b0e519c3\",\n        :attach-time #&lt;DateTime 2013-03-21T22:00:56.000-07:00&gt;,\n        :delete-on-termination true}}],\n     :network-interfaces [],\n     :public-ip-address \"164.73.176.213\",\n     :placement\n     {:availability-zone \"us-east-1a\",\n      :group-name \"\",\n      :tenancy \"default\"},\n     :private-ip-address \"10.116.187.19\",\n     :security-groups [{:group-name \"cx\", :group-id \"sg-38f45150\"}],\n     :state-transition-reason \"\",\n     :private-dns-name \"ip-10-116-187-19.ec2.internal\",\n     :instance-id \"i-cefbe7a2\",\n     :key-name \"cxci\",\n     :architecture \"x86_64\",\n     :client-token \"\",\n     :image-id \"ami-baba68d3\",\n     :ami-launch-index 0,\n     :monitoring {:state \"disabled\"},\n     :product-codes [],\n     :launch-time #&lt;DateTime 2013-03-21T22:00:52.000-07:00&gt;,\n     :tags [{:value \"CXCI_nightly\", :key \"Name\"}]}],\n   :reservation-id \"r-8a23d6f7\"}<\/pre><\/div>\n<p>If you look at the <code>Reservation<\/code> <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/model\/Reservation.html\" rel=\"nofollow\">Javadoc<\/a> you'll see that <code>getGroups()<\/code> returns a <code>java.util.List<\/code> of <code>GroupIdentifiers<\/code>, which is converted to a vector of maps containing keys <code>:group-name<\/code> and <code>:group-id<\/code>, under the <code>:groups<\/code> key. Ditto for :block-device-mappings and :tags, and so and so on...<\/p>\n<p>Similar in concept to JSON unwrapping in Jackson, Amazonica supports root unwrapping of the returned data. So calling<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>; dynamodb\n(list-tables)<\/pre><\/div>\n<p>by default would return<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:table-names [\"TableOne\" \"TableTwo\" \"TableThree\"]}<\/pre><\/div>\n<p>However, if you call<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(set-root-unwrapping! true)<\/pre><\/div>\n<p>then single keyed top level maps will be \"unwrapped\" like so:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(list-tables)\n=&gt; [\"TableOne\" \"TableTwo\" \"TableThree\"]<\/pre><\/div>\n<p>The returned data can be \"round tripped\" as well. So the returned Clojure data structures can be supplied as arguments to function calls which delegate to Java methods taking the same object type as an argument. See the section below for more on this.<\/p>\n<h3><a id=\"user-content-argument-coercion\" class=\"anchor\" aria-hidden=\"true\" href=\"#argument-coercion\"><\/a>Argument Coercion<\/h3>\n<p>Coercion of any types that are part of the java.lang wrapper classes happens transparently. So for example, Clojure's preferred longs are automatically converted to ints where required.<\/p>\n<p>Clojure data structures automatically participate in the Java Collections abstractions, and so no explicit coercion is necessary. Typically when service calls take collections as parameter arguments, as in the case above, the values in the collections are most often instances of the Java wrapper classes.<\/p>\n<p>When complex objects consisting of types outside of those in the <code>java.lang<\/code> package are required as argument parameters, smart conversions are attempted based on the argument types of the underlying Java method signature. Methods requiring a <code>java.util.Date<\/code> argument can take Joda Time <code>org.joda.time.base.AbstractInstants<\/code>, longs, or Strings (default pattern is \"yyyy-MM-dd\"), with conversion happening automatically.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(set-date-format! \"MM-dd-yyyy\")<\/pre><\/div>\n<p>can be used to set the pattern supplied to the underlying <code>java.text.SimpleDateFormat<\/code>.<\/p>\n<p>In cases where collection arguments contain instances of AWS \"model\" classes, Clojure maps will be converted to the appropriate AWS Java bean instance. So for example, <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/AmazonEC2Client.html#describeAvailabilityZones(com.amazonaws.services.ec2.model.DescribeAvailabilityZonesRequest)\" rel=\"nofollow\">describeAvailabilityZones()<\/a> can take a <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/ec2\/model\/DescribeAvailabilityZonesRequest.html\" rel=\"nofollow\">DescribeAvailabilityZonesRequest<\/a> which itself has a <code>filters<\/code> property, which is a <code>java.util.List<\/code> of <code>com.amazonaws.services.ec2.model.Filters<\/code>. Passing the filters argument would look like:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(describe-availability-zones :filters [{:name   \"environment\"\n                                        :values [\"dev\" \"qa\" \"staging\"]}])<\/pre><\/div>\n<p>and return the following Clojure collection:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:availability-zones\n [{:state \"available\",\n   :region-name \"us-east-1\",\n   :zone-name \"us-east-1a\",\n   :messages []}\n  {:state \"available\",\n   :region-name \"us-east-1\",\n   :zone-name \"us-east-1b\",\n   :messages []}\n  {:state \"available\",\n   :region-name \"us-east-1\",\n   :zone-name \"us-east-1c\",\n   :messages []}\n  {:state \"available\",\n   :region-name \"us-east-1\",\n   :zone-name \"us-east-1d\",\n   :messages []}\n  {:state \"available\",\n   :region-name \"us-east-1\",\n   :zone-name \"us-east-1e\",\n   :messages []}]}<\/pre><\/div>\n<h3><a id=\"user-content-extension-points\" class=\"anchor\" aria-hidden=\"true\" href=\"#extension-points\"><\/a>Extension points<\/h3>\n<p>Clojure apis built specifically to wrap a Java client, such as this one, often provide \"conveniences\" for the user of the api, to remove boilerplate. In Amazonica this is accomplished via the IMarshall protocol, which defines the contract for converting the returned Java result from the AWS service call to Clojure data, and the<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(amazonica.core\/register-coercions)<\/pre><\/div>\n<p>function, which takes a map of class\/function pairs defining how a value should be coerced to a specific AWS Java bean. You can find a good example of this in the <code>amazonica.aws.dynamodb<\/code> namespace. Consider the following DynamoDB service call:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(get-item :table-name \"MyTable\"\n          :key \"foo\")<\/pre><\/div>\n<p>The <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/dynamodb\/model\/GetItemRequest.html\" rel=\"nofollow\">GetItemRequest<\/a> takes a <code>com.amazonaws.services.dynamodb.model.Key<\/code> which is composed of a hash key of type <code>com.amazonaws.services.dynamodb.model.AttributeValue<\/code> and optional range key also of type <code>AttributeValue<\/code>. Without the coercions registered for <code>Key<\/code> and <code>AttributeValue<\/code> in <code>amazonica.aws.dynamodb<\/code> we would need to write:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(get-item :table-name \"TestTable\"\n          :key {:hash-key-element {:s \"foo\"}})<\/pre><\/div>\n<p>Note that either form will work. This allows contributors to the library to incrementally evolve the api independently from the core of the library, as well as maintain backward compatibility of existing code written against prior versions of the library which didn't contain the conveniences.<\/p>\n<h3><a id=\"user-content-authentication\" class=\"anchor\" aria-hidden=\"true\" href=\"#authentication\"><\/a>Authentication<\/h3>\n<p>The default authentication scheme is to use the <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/DefaultAWSCredentialsProviderChain.html\" rel=\"nofollow\">chained Provider class<\/a> from the AWS SDK, whereby authentication is attempted in the following order:<\/p>\n<ul><li>Environment Variables - AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY<\/li>\n<li>Java System Properties - aws.accessKeyId and aws.secretKey<\/li>\n<li>Credential profiles file at the default location (~\/.aws\/credentials) shared by all AWS SDKs and the AWS CLI<\/li>\n<li>Instance profile credentials delivered through the Amazon EC2 metadata service<\/li>\n<\/ul><p>Note that in order for the Instance Profile Metadata to be found, you must have launched the instance with a provided IAM role, and the same permissions as the IAM Role the instance was launched with will apply.<\/p>\n<p>See the <a href=\"http:\/\/docs.aws.amazon.com\/AWSSdkDocsJava\/latest\/DeveloperGuide\/java-dg-roles.html\" rel=\"nofollow\">AWS docs<\/a> for reference.<\/p>\n<p>Additionally, all of the functions may take as their first argument an optional map of credentials:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def cred {:access-key \"aws-access-key\"\n           :secret-key \"aws-secret-key\"\n           :endpoint   \"us-west-1\"})\n(describe-instances cred)<\/pre><\/div>\n<p>The credentials map may contain zero or one of the following:<\/p>\n<ul><li><code>:access-key<\/code> and <code>:secret-key<\/code>, in which case an instance of <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/BasicAWSCredentials.html\" rel=\"nofollow\"><code>BasicAWSCredentials<\/code><\/a> will be created.<\/li>\n<li><code>:session-token<\/code>, in which case an instance of <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/BasicSessionCredentials.html\" rel=\"nofollow\"><code>BasicSessionCredentials<\/code><\/a> will be created.<\/li>\n<li><code>:profile<\/code>, in which case an instance of <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/profile\/ProfileCredentialsProvider.html\" rel=\"nofollow\"><code>ProfileCredentialsProvider<\/code><\/a> will be created.<\/li>\n<li><code>:cred<\/code>, in which case the <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/AWSCredentialsProvider.html\" rel=\"nofollow\"><code>AWSCredentialsProvider<\/code><\/a> instance provided will be used.<\/li>\n<li>Or rather than a Clojure map, the argument may be an actual instance or subclass of either <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/AWSCredentialsProvider.html\" rel=\"nofollow\"><code>AWSCredentialsProvider<\/code><\/a> or <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/auth\/AWSCredentials.html\" rel=\"nofollow\"><code>AWSCredentials<\/code><\/a>.<\/li>\n<\/ul><p>In addition, the credentials map may contain an <code>:endpoint<\/code> entry. If the value of the <code>:endpoint<\/code> key is a lower case, hyphenated translation of one of the <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/regions\/Regions.html\" rel=\"nofollow\">Regions enums<\/a>, <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/AmazonWebServiceClient.html#setRegion(com.amazonaws.regions.Region)\" rel=\"nofollow\">.setRegion<\/a> will be called on the Client, otherwise <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/AmazonWebServiceClient.html#setEndpoint(java.lang.String)\" rel=\"nofollow\">.setEndpoint<\/a> will be called.<\/p>\n<p><strong>Note:<\/strong> The first function called (for each distinct AWS service namespace, e.g. amazonica.aws.ec2) creates an Amazon*Client, which is effectively cached via memoization.  Therefore, if you explicitly pass different credentials maps to different functions, you will effectively have different Clients.<\/p>\n<p>For example, to work with ec2 instances in different regions you might do something like:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(ec2\/create-image {:endpoint \"us-east-1\"} :instance-id \"i-1b9a9f71\")\n(ec2\/create-image {:endpoint \"us-west-2\"} :instance-id \"i-kj239d7d\")<\/pre><\/div>\n<p>You will have created two AmazonEC2Clients, pointing to the two different regions. Likewise, if you omit the explicit credentials map then the DefaultAWSCredentialsProviderChain will be used. So in the following scenario you will again have two different Amazon*Clients:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(set-s3client-options :path-style-access true)\n(create-bucket credentials \"foo\")<\/pre><\/div>\n<p>The call to <code>set-s3client-options<\/code> will use a DefaultAWSCredentialsProviderChain, while the <code>create-bucket<\/code> call will create a separate AmazonS3Client with BasicAWSCredentials.<\/p>\n<p>As a convenience, users may call <code>(defcredential)<\/code> before invoking any service functions and passing in their AWS key pair and an optional endpoint:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defcredential \"aws-access-key\" \"aws-secret-key\" \"us-west-1\")<\/pre><\/div>\n<p>All subsequent API calls will use the specified credential. If you need to execute a service call with alternate credentials, or against a different region than the one passed to <code>(defcredential)<\/code>, you can wrap these ad-hoc calls in the <code>(with-credential)<\/code> macro, which takes a vector of key pair credentials and an optional endpoint, like so:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defcredential \"account-1-aws-access-key\" \"aws-secret-key\" \"us-west-1\")\n(describe-instances)\n; returns instances visible to account-1\n(with-credential [\"account-2-aws-access-key\" \"secret\" \"us-east-1\"]\n  (describe-instances))\n; returns EC2 instances visible to account-2 running in US-East region\n(describe-images :owners [\"self\"])\n; returns images belonging to account-1<\/pre><\/div>\n<h3><a id=\"user-content-client-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#client-configuration\"><\/a>Client configuration<\/h3>\n<p>You can supply a <code>:client-config<\/code> entry in the credentials map to configure the <a href=\"http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/ClientConfiguration.html\" rel=\"nofollow\">ClientConfiguration<\/a> that the Amazon client uses. This is useful if you need to use a proxy.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(describe-images {:client-config {:proxy-host \"proxy.address.com\" :proxy-port 8080}})<\/pre><\/div>\n<h3><a id=\"user-content-exception-handling\" class=\"anchor\" aria-hidden=\"true\" href=\"#exception-handling\"><\/a>Exception Handling<\/h3>\n<p>All functions throw <code>com.amazonaws.AmazonServiceExceptions<\/code>. If you wish to catch exceptions you can convert the AWS object to a Clojure map like so:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(try\n  (create-snapshot :volume-id \"vol-ahsg23h\"\n                   :description \"daily backup\")\n  (catch Exception e\n    (log (ex-&gt;map e))))\n; {:error-code \"InvalidParameterValue\",\n;  :error-type \"Unknown\",\n;  :status-code 400,\n;  :request-id \"9ba69e16-ed63-41d4-ac02-1f6032cb64de\",\n;  :service-name \"AmazonEC2\",\n;  :message\n;  \"Value (vol-ahsg23h) for parameter volumeId is invalid. Expected: 'vol-...'.\",\n;  :stack-trace \"Status Code: 400, AWS Service: AmazonEC2, AWS Request ID: a5b0340a-8f37-4122-941c-ed8d5472b11d, AWS Error Code: InvalidParameterValue, AWS Error Message: Value (vol-ahsg23h) for parameter volumeId is invalid. Expected: 'vol-...'.\n;  at com.amazonaws.http.AmazonHttpClient.handleErrorResponse(AmazonHttpClient.java:644)\n;   at com.amazonaws.http.AmazonHttpClient.executeHelper(AmazonHttpClient.java:338)\n;   at com.amazonaws.http.AmazonHttpClient.execute(AmazonHttpClient.java:190)\n;   at com.amazonaws.services.ec2.AmazonEC2Client.invoke(AmazonEC2Client.java:6199)\n;   at com.amazonaws.services.ec2.AmazonEC2Client.createSnapshot(AmazonEC2Client.java:1531)\n;   .....<\/pre><\/div>\n<h4><a id=\"user-content-for-the-memory-constrained\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-the-memory-constrained\"><\/a>For the memory constrained<\/h4>\n<p>If you're especially concerned about the size of your uberjar, you can limit the transitive dependencies pulled in by the AWS Java SDK, which currently total 35mb. You'll need to exclude the entire AWS Java SDK and the Amazon Kinesis Client, and then add back only those services you'll be using (although core is always required). So for example, if you were only using S3, you could restrict the dependencies to only include the required jars like so:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:dependencies [[org.clojure\/clojure \"1.7.0\"]\n               [amazonica \"0.3.48\" :exclusions [com.amazonaws\/aws-java-sdk\n                                                com.amazonaws\/amazon-kinesis-client]]\n               [com.amazonaws\/aws-java-sdk-core \"1.10.49\"]\n               [com.amazonaws\/aws-java-sdk-s3 \"1.10.49\"]]<\/pre><\/div>\n<h3><a id=\"user-content-running-the-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-the-tests\"><\/a>Running the tests<\/h3>\n<p>As always, <code>lein test<\/code> will run all the tests. Note that some of the namespaces require the file <code>~\/.aws\/credentials<\/code> to be present and be of the same form as required by the official AWS tools:<\/p>\n<pre>[default]\naws_access_key_id = AKIAABCDEFGHIEJK\naws_secret_access_key = 6rqzvpAbcd1234++zyx987WUV654sRq\n<\/pre>\n<h3><a id=\"user-content-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance\"><\/a>Performance<\/h3>\n<p>Amazonica uses reflection extensively, to generate the public Vars, to set the bean properties passed as arguments to those functions, and to invoke the actual service method calls on the underlying AWS Client class. As such, one may wonder if such pervasive use of reflection will result in unacceptable performance. In general, this shouldn't be an issue, as the cost of reflection should be relatively minimal compared to the latency incurred by making a remote call across the network. Furthermore, typical AWS usage is not going to be terribly concerned with performance, except with specific services such as DynamoDB, RDS, SimpleDB, or SQS. But we have done some basic benchmarking against the excellent DynamoDB <a href=\"https:\/\/github.com\/weavejester\/rotary\">rotary<\/a> library, which uses no explicit reflection. Results are shown below. Benchmarking code is available at <a href=\"https:\/\/github.com\/mcohen01\/amazonica-benchmark\">https:\/\/github.com\/mcohen01\/amazonica-benchmark<\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/70acfef82942652955aa1e795858b314cb4c0751\/68747470733a2f2f7261772e6769746875622e636f6d2f6d636f68656e30312f616d617a6f6e6963612d62656e63686d61726b2f6d61737465722f7265666c656374696f6e2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/70acfef82942652955aa1e795858b314cb4c0751\/68747470733a2f2f7261772e6769746875622e636f6d2f6d636f68656e30312f616d617a6f6e6963612d62656e63686d61726b2f6d61737465722f7265666c656374696f6e2e706e67\" alt=\"Benchmark results\" data-canonical-src=\"https:\/\/raw.github.com\/mcohen01\/amazonica-benchmark\/master\/reflection.png\" \/><\/a><\/p>\n<h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<h3><a id=\"user-content-autoscaling\" class=\"anchor\" aria-hidden=\"true\" href=\"#autoscaling\"><\/a>Autoscaling<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.autoscaling]))\n(create-launch-configuration :launch-configuration-name \"aws_launch_cfg\"\n                             :block-device-mappings [\n                              {:device-name \"\/dev\/sda1\"\n                               :virtual-name \"vol-b0e519c3\"\n                               :ebs {:snapshot-id \"snap-36295e51\"\n                                     :volume-size 32}}]\n                             :ebs-optimized true\n                             :image-id \"ami-6fde0d06\"\n                             :instance-type \"m1.large\"\n                             :spot-price \".10\")\n(create-auto-scaling-group :auto-scaling-group-name \"aws_autoscale_grp\"\n                           :availability-zones [\"us-east-1a\" \"us-east-1b\"]\n                           :desired-capacity 3\n                           :health-check-grace-period 300\n                           :health-check-type \"EC2\"\n                           :launch-configuration-name \"aws_launch_cfg\"\n                           :min-size 3\n                           :max-size 3)\n(describe-auto-scaling-instances)\n<\/pre><\/div>\n<h3><a id=\"user-content-batch\" class=\"anchor\" aria-hidden=\"true\" href=\"#batch\"><\/a>Batch<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.batch]))\n(submit-job :job-name \"my-job\"\n            :job-definition \"my-job-definition\"\n            :job-queue \"my-job-queue\"\n            :parameters {\"example-url\" \"example.com\"})<\/pre><\/div>\n<h3><a id=\"user-content-cloudformation\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudformation\"><\/a>CloudFormation<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.cloudformation]))\n(create-stack :stack-name \"my-stack\"\n              :template-url \"abcd1234.s3.amazonaws.com\")\n(describe-stack-resources :stack-name \"my_cloud_stack\")\n<\/pre><\/div>\n<h3><a id=\"user-content-cloudfront\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudfront\"><\/a>CloudFront<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.cloudfront]))\n(create-distribution  :distribution-config {\n                      :enabled true\n                      :default-root-object \"index.html\"\n                      :origins\n                       {:quantity 0\n                        :items []}\n                      :logging\n                       {:enabled false\n                        :include-cookies false\n                        :bucket \"abcd1234.s3.amazonaws.com\"\n                        :prefix \"cflog_\"}\n                      :caller-reference 12345\n                      :aliases\n                       {:items [\"m.example.com\" \"www.example.com\"]\n                        :quantity 2}\n                      :cache-behaviors\n                       {:quantity 0\n                        :items []}\n                      :comment \"example\"\n                      :default-cache-behavior\n                       {:target-origin-id \"MyOrigin\"\n                        :forwarded-values\n                          {:query-string false\n                           :cookies\n                             {:forward \"none\"}}}\n                       :trusted-signers\n                         {:enabled false\n                          :quantity 0}\n                       :viewer-protocol-policy \"allow-all\"\n                       :min-ttl 3600}\n                      :price-class \"PriceClass_All\"})\n(list-distributions :max-items 10)\n<\/pre><\/div>\n<h3><a id=\"user-content-cloudsearch\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudsearch\"><\/a>CloudSearch<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.cloudsearch]))\n(create-domain :domain-name \"my-index\")\n(index-documents :domain-name \"my-index\")\n<\/pre><\/div>\n<h3><a id=\"user-content-cloudsearchv2\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudsearchv2\"><\/a>CloudSearchV2<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.cloudsearchv2]))\n(create-domain :domain-name \"my-index\")\n(index-documents :domain-name \"my-index\")\n(build-suggesters :domain-name \"my-index\")\n(list-domains)\n<\/pre><\/div>\n<h3><a id=\"user-content-cloudsearchdomain\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudsearchdomain\"><\/a>CloudSearchDomain<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>;; get the document and search service endpoints\n(clojure.pprint\/pprint\n  (amazonica.aws.cloudsearchv2\/describe-domains))\n(csd\/set-endpoint \"doc-domain-name-6fihexkq1234567895wm.us-east-1.cloudsearch.amazonaws.com\")\n(csd\/upload-documents\n  :content-type \"application\/json\"\n  :documents (io\/input-stream json-documents))\n(csd\/set-endpoint \"search-domain-name-6fihexkq1234567895wm.us-east-1.cloudsearch.amazonaws.com\")\n;; http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/cloudsearchdomain\/model\/SearchRequest.html\n(csd\/search :query \"drumpf\")\n(csd\/suggest :query \"{\\\"query\\\": \\\"make donald drumpf\\\"\"\n             :suggester \"url_suggester\")\n<\/pre><\/div>\n<h3><a id=\"user-content-cloudwatch\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudwatch\"><\/a>CloudWatch<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.cloudwatch]))\n(put-metric-alarm :alarm-name \"my-alarm\"\n                  :actions-enabled true\n                  :evaluation-periods 5\n                  :period 60\n                  :metric-name \"CPU\"\n                  :threshold \"50%\")\n<\/pre><\/div>\n<p>To put metric data.   <a href=\"http:\/\/docs.aws.amazon.com\/AmazonCloudWatch\/latest\/APIReference\/API_MetricDatum.html\" rel=\"nofollow\">UnitTypes<\/a><\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(put-metric-data\n    {:endpoint \"us-west-1\"} ;; Defaults to us-east-1\n    :namespace \"test_namespace\"\n    :metric-data [{:metric-name \"test_metric\"\n                   :unit \"Count\"\n                   :value 1.0\n                   :dimensions [{:name \"test_name\" :value \"test_value\"}]}])<\/pre><\/div>\n<p>To batch get metric data.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(get-metric-data\n :start-time \"2018-06-14T00:00:00Z\"\n :end-time   \"2018-06-15T00:00:00Z\"\n :metric-data-queries [{:id \"test\"\n                        :metricStat {:metric {:namespace \"AWS\/DynamoDB\"\n                                              :metricName \"ProvisionedReadCapacityUnits\"\n                                              :dimensions [{:name \"TableName\"\n                                                            :value \"MyTableName\"}]}\n                                     :period 86400\n                                     :stat \"Sum\"\n                                     :unit \"Count\"}}])<\/pre><\/div>\n<h3><a id=\"user-content-cloudwatchevents\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloudwatchevents\"><\/a>CloudWatchEvents<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.cloudwatchevents]))\n(put-rule\n    :name \"nightly-backup\"\n    :description \"Backup DB nightly at 10:00 UTC (2 AM or 3 AM Pacific)\"\n    :schedule-expression \"cron(0 10 * * ? *)\")\n(put-targets\n    :rule \"nightly-backup\"\n    :targets [{:id    \"backup-lambda\"\n               :arn   \"arn:aws:lambda:us-east-1:123456789012:function:backup-lambda\"\n               :input (json\/write-str {\"whatever\" \"arguments\"})}])<\/pre><\/div>\n<h3><a id=\"user-content-codedeploy\" class=\"anchor\" aria-hidden=\"true\" href=\"#codedeploy\"><\/a>CodeDeploy<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.codedeploy]))\n(list-applications)\n<\/pre><\/div>\n<h3><a id=\"user-content-cognitoidentityproviders\" class=\"anchor\" aria-hidden=\"true\" href=\"#cognitoidentityproviders\"><\/a>CognitoIdentityProviders<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.cognitoidp :refer :all]))\n(list-user-pools {:max-results 2})\n=&gt; {:user-pools [{:lambda-config {}, :last-modified-date \"2017-06-16T14:16:28.950-03:00\"], :creation-date \"2017-06-15T16:23:04.555-03:00\"], :name \"Amazonica\", :id \"us-west-1_example\"}]}\n<\/pre><\/div>\n<h3><a id=\"user-content-comprehend\" class=\"anchor\" aria-hidden=\"true\" href=\"#comprehend\"><\/a>Comprehend<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n(:require [amazonica.aws.comprehend :refer :all]))\n(amazonica.aws.comprehend\/detect-entities {:language-code \"en\" :text \"Hi my name is Joe Bloggs and I live in Glasgow, Scotland\"})\n=&gt; {:entities [{:type \"PERSON\", :text \"Joe Bloggs\", :score 0.99758613, :begin-offset 14, :end-offset 24} {:type \"LOCATION\", :text \"Glasgow, Scotland\", :score 0.93267196, :begin-offset 39, :end-offset 56}]}<\/pre><\/div>\n<h3><a id=\"user-content-datapipeline\" class=\"anchor\" aria-hidden=\"true\" href=\"#datapipeline\"><\/a>DataPipeline<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.datapipeline]))\n(create-pipeline :name \"my-pipeline\"\n                 :unique-id \"mp\")\n(put-pipeline-definition  :pipeline-id \"df-07746012XJFK4DK1D4QW\"\n                          :pipeline-objects [{:name \"my-pipeline-object\"\n                                              :id \"my-pl-object-id\"\n                                              :fields [{:key \"some-key\"\n                                                        :string-value \"foobar\"}]}])\n(list-pipelines)\n(delete-pipeline :pipeline-id pid)\n<\/pre><\/div>\n<h3><a id=\"user-content-dynamodbv2\" class=\"anchor\" aria-hidden=\"true\" href=\"#dynamodbv2\"><\/a>DynamoDBV2<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.dynamodbv2]))\n(def cred {:access-key \"aws-access-key\"\n           :secret-key \"aws-secret-key\"\n           :endpoint   \"http:\/\/localhost:8000\"})\n(create-table cred\n              :table-name \"TestTable\"\n              :key-schema\n                [{:attribute-name \"id\"   :key-type \"HASH\"}\n                 {:attribute-name \"date\" :key-type \"RANGE\"}]\n              :attribute-definitions\n                [{:attribute-name \"id\"      :attribute-type \"S\"}\n                 {:attribute-name \"date\"    :attribute-type \"N\"}\n                 {:attribute-name \"column1\" :attribute-type \"S\"}\n                 {:attribute-name \"column2\" :attribute-type \"S\"}]\n              :local-secondary-indexes\n                [{:index-name \"column1_idx\"\n                  :key-schema\n                   [{:attribute-name \"id\"   :key-type \"HASH\"}\n                    {:attribute-name \"column1\" :key-type \"RANGE\"}]\n                 :projection\n                   {:projection-type \"INCLUDE\"\n                    :non-key-attributes [\"id\" \"date\" \"column1\"]}}\n                 {:index-name \"column2_idx\"\n                  :key-schema\n                   [{:attribute-name \"id\"   :key-type \"HASH\"}\n                    {:attribute-name \"column2\" :key-type \"RANGE\"}]\n                 :projection {:projection-type \"ALL\"}}]\n              :provisioned-throughput\n                {:read-capacity-units 1\n                 :write-capacity-units 1})\n(put-item cred\n          :table-name \"TestTable\"\n          :return-consumed-capacity \"TOTAL\"\n          :return-item-collection-metrics \"SIZE\"\n          :item {\n            :id \"foo\"\n            :date 123456\n            :text \"barbaz\"\n            :column1 \"first name\"\n            :column2 \"last name\"\n            :numberSet #{1 2 3}\n            :stringSet #{\"foo\" \"bar\"}\n            :mixedList [1 \"foo\"]\n            :mixedMap {:name \"baz\" :secret 42}})\n(get-item cred\n          :table-name \"TestTable\"\n          :key {:id {:s \"foo\"}\n                :date {:n 123456}})\n(query cred\n       :table-name \"TestTable\"\n       :limit 1\n       :index-name \"column1_idx\"\n       :select \"ALL_ATTRIBUTES\"\n       :scan-index-forward true\n       :key-conditions\n        {:id      {:attribute-value-list [\"foo\"]      :comparison-operator \"EQ\"}\n         :column1 {:attribute-value-list [\"first na\"] :comparison-operator \"BEGINS_WITH\"}})\n(batch-write-item\n  cred\n  :return-consumed-capacity \"TOTAL\"\n  :return-item-collection-metrics \"SIZE\"\n  :request-items\n    {\"TestTable\"\n      [{:delete-request\n         {:key {:id \"foo\"\n                :date 123456}}}\n       {:put-request\n         {:item {:id \"foobar\"\n                 :date 3172671\n                 :text \"bunny\"\n                 :column1 \"funky\"}}}]})\n;; dynamodb-expressions https:\/\/github.com\/brabster\/dynamodb-expressions\n;; exists to make update expressions easier to write for Amazonica.\n(update-item\n  cred\n  :table-name \"TestTable\"\n  :key {:id \"foo\"}\n  :update-expression \"ADD #my_foo :x SET bar.baz = :y\"\n  :expression-attribute-names {\"#my_foo\" \"my-foo\"}\n  :expression-attribute-values {\":x\" 1\n                                \":y\" \"barbaz\"})\n(batch-get-item\n  cred\n  :return-consumed-capacity \"TOTAL\"\n  :request-items {\n  \"TestTable\" {:keys [{\"id\"   {:s \"foobar\"}\n                       \"date\" {:n 3172671}}\n                      {\"id\"   {:s \"foo\"}\n                       \"date\" {:n 123456}}]\n               :consistent-read true\n               :attributes-to-get [\"id\" \"text\" \"column1\"]}})\n(scan cred :table-name \"TestTable\")\n(describe-table cred :table-name \"TestTable\")\n(delete-table cred :table-name \"TestTable\")\n;; Amazonica depends on `[com.amazonaws\/amazon-kinesis-client]`\n;; which has a dependency on `[com.amazonaw\/aws-java-sdk-dynamodb]`.\n;; The version of this dependency is too old to support TTL,\n;; so you'll need to exclude it and explicitly depend on a recent version\n;; of `com.amazonaw\/aws-java-sdk-dynamodb` like `1.0.9` to use this feature for now.\n(update-time-to-live\n  cred\n  :table-name \"TestTable\"\n  :time-to-live-specification {:attribute-name \"foo\" :enabled true}\n<\/pre><\/div>\n<h3><a id=\"user-content-ec2\" class=\"anchor\" aria-hidden=\"true\" href=\"#ec2\"><\/a>EC2<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.ec2]))\n(-&gt; (run-instances :image-id \"ami-54f71039\"\n                   :instance-type \"c3.large\"\n                   :min-count 1\n                   :max-count 1)\n    (get-in [:reservation :instances 0 :instance-id]))\n(describe-images :owners [\"self\"])\n(describe-instances :filters [{:name \"tag:env\" :values [\"production\"]}])\n(create-image :name \"my_test_image\"\n              :instance-id \"i-1b9a9f71\"\n              :description \"test image - safe to delete\"\n              :block-device-mappings [\n                {:device-name  \"\/dev\/sda1\"\n                 :virtual-name \"myvirtual\"\n                 :ebs {\n                   :volume-size 8\n                   :volume-type \"standard\"\n                   :delete-on-termination true}}])\n(create-snapshot :volume-id   \"vol-8a4857fa\"\n                 :description \"my_new_snapshot\")\n<\/pre><\/div>\n<h3><a id=\"user-content-ecs\" class=\"anchor\" aria-hidden=\"true\" href=\"#ecs\"><\/a>ECS<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.ecs :refer :all]))\n(register-task-definition\n {:family \"grafana2\",\n  :container-definitions [{:name \"grafana2\"\n                           :image \"bbinet\/grafana2\",\n                           :port-mappings [{:container-port 3000, :host-port 3000}]\n                           :memory 300\n                           :cpu 300\n                           }]})\n(describe-task-definition :task-definition \"grafana2\")\n(list-task-definitions :family-prefix \"grafana2\")\n;; create cluster\n(create-cluster :cluster-name \"Amazonica\")\n(list-clusters)\n(describe-clusters)\n(create-service :cluster \"Amazonica\"\n                :service-name \"grafana2\"\n                :task-definition \"grafana2\" :desired-count 1\n                ;;:role \"ecsServiceRole\"\n                ;;:load-balancers [...]\n                )\n(list-services :cluster \"Amazonica\")\n(describe-services :cluster \"Amazonica\" :services [\"grafana2\"])\n;; add ec2 instances to your cluster\n(update-service :cluster \"Amazonica\" :service \"grafana2\" :desired-count 0)\n(delete-service :cluster \"Amazonica\" :service \"grafana2\")\n(delete-cluster :cluster \"Amazonica\")\n;; run task\n(run-task\n  :cluster \"Amazonica\"\n  :launch-type LaunchType\/FARGATE\n  :task-definition \"task-def-name\"\n  :overrides {:container-overrides [{:name    \"container-name\"\n                                     :command [\"java\" \"-jar\" \"artifact.jar\" \"arg1\" \"arg2\"]}]}\n  :network-configuration {:aws-vpc-configuration {:assign-public-ip AssignPublicIp\/ENABLED\n                                                  :subnets          [\"subnet-XXXXXXXX\"]\n                                                  :security-groups  [\"sg-XXXXXXXXXXXXXXXX\"]}})<\/pre><\/div>\n<h3><a id=\"user-content-ecr\" class=\"anchor\" aria-hidden=\"true\" href=\"#ecr\"><\/a>ECR<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[amazonica.aws.ecr :as ecr])\n(ecr\/describe-repositories {})\n(ecr\/create-repository :repository-name \"amazonica\")\n(ecr\/get-authorization-token {})\n(ecr\/list-images :repository-name \"amazonica\")\n(ecr\/delete-repository :repository-name \"amazonica\")<\/pre><\/div>\n<h3><a id=\"user-content-elasticache\" class=\"anchor\" aria-hidden=\"true\" href=\"#elasticache\"><\/a>Elasticache<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.elasticache]))\n(describe-cache-engine-versions)\n(create-cache-cluster :engine \"memcached\"\n                      :engine-version \"1.4.14\"\n                      :num-cache-nodes 1\n                      :cache-node-type \"cache.t1.micro\"\n                      :cache-cluster-id \"memcached-cluster\")\n(describe-cache-clusters)\n(describe-events)\n(delete-cache-cluster :cache-cluster-id \"memcached-cluster\")\n<\/pre><\/div>\n<h3><a id=\"user-content-elasticbeanstalk\" class=\"anchor\" aria-hidden=\"true\" href=\"#elasticbeanstalk\"><\/a>ElasticBeanstalk<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.elasticbeanstalk]))\n(describe-applications)\n(describe-environments)\n(create-environment creds\n                    {:application-name \"app\"\n                     :environment-name \"env\"\n                     :version-label \"1.0\"\n                     :solution-stack-name \"64bit Amazon Linux 2014.09 v1.0.9 running Docker 1.2.0\"\n                     :option-settings [{:namespace \"aws:elb:loadbalancer\"\n                                        :option-name \"LoadBalancerHTTPSPort\"\n                                        :value \"443\"}\n                                       {:namespace \"aws:elb:loadbalancer\"\n                                        :option-name \"LoadBalancerHTTPPort\"\n                                        :value \"OFF\"}]}))\n(describe-configuration-settings {:application-name \"app\" :environment-name \"env\"})<\/pre><\/div>\n<h3><a id=\"user-content-elasticloadbalancing\" class=\"anchor\" aria-hidden=\"true\" href=\"#elasticloadbalancing\"><\/a>ElasticLoadBalancing<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.elasticloadbalancing]))\n(deregister-instances-from-load-balancer :load-balancer-name \"my-ELB\"\n                                         :instances [{:instance-id \"i-1ed40bad\"}])\n(register-instances-with-load-balancer :load-balancer-name \"my-ELB\"\n                                       :instances [{:instance-id \"i-1fa370ea\"}])\n<\/pre><\/div>\n<h3><a id=\"user-content-elasticmapreduce\" class=\"anchor\" aria-hidden=\"true\" href=\"#elasticmapreduce\"><\/a>ElasticMapReduce<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws\n          elasticmapreduce\n          s3]))\n(create-bucket :bucket-name \"emr-logs\"\n               :access-control-list {:grant-permission [\"LogDelivery\" \"Write\"]})\n(set-bucket-logging-configuration :bucket-name \"emr-logs\"\n                                  :logging-configuration\n                                    {:log-file-prefix \"hadoop-job_\"\n                                     :destination-bucket-name \"emr-logs\"})\n(run-job-flow :name \"my-job-flow\"\n              :log-uri \"s3n:\/\/emr-logs\/logs\"\n              :instances\n                {:instance-groups [\n                   {:instance-type \"m1.large\"\n                    :instance-role \"MASTER\"\n                    :instance-count 1\n                    :market \"SPOT\"\n                    :bid-price \"0.10\"}]}\n              :steps [\n                {:name \"my-step\"\n                 :hadoop-jar-step\n                   {:jar \"s3n:\/\/beee0534-ad04-4143-9894-8ddb0e4ebd31\/hadoop-jobs\/bigml\"\n                    :main-class \"bigml.core\"\n                    :args [\"s3n:\/\/beee0534-ad04-4143-9894-8ddb0e4ebd31\/data\" \"output\"]}}])\n(list-clusters)\n(describe-cluster :cluster-id \"j-38BW9W0NN8YGV\")\n(list-steps :cluster-id \"j-38BW9W0NN8YGV\")\n(list-bootstrap-actions :cluster-id \"j-38BW9W0NN8YGV\")\n<\/pre><\/div>\n<h3><a id=\"user-content-elasticsearchservice\" class=\"anchor\" aria-hidden=\"true\" href=\"#elasticsearchservice\"><\/a>ElasticsearchService<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.awselasticsarch]))\n(list-domain-names {})<\/pre><\/div>\n<h3><a id=\"user-content-elastictranscoder\" class=\"anchor\" aria-hidden=\"true\" href=\"#elastictranscoder\"><\/a>ElasticTranscoder<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n(:use [amazonica.aws.elastictranscoder))\n(list-pipelines)\n;; -&gt; {:pipelines []}\n(list-presets)\n;; -&gt; {:presets [{:description \"System preset generic 1080p\", ....}]}\n;; status can be :Submitted :Progressing :Complete :Canceled :Error\n(list-jobs-by-status :status :Complete)\n;; -&gt; ...\n(def new-pipeline-id (-&gt; (create-pipeline\n                           :role \"arn:aws:iam::289431957111:role\/Elastic_Transcoder_Default_Role\",\n                           :name \"avi-to-mp4\",\n                           :input-bucket \"avi-to-convert\",\n                           :output-bucket \"converted-mp4\")\n                       :pipeline\n                       :id))\n;; -&gt; \"1111111111111-11aa11\"\n(create-job :pipeline-id \"1111111111111-11aa11\"\n            :input {:key \"my\/s3\/input\/obj\/key.avi\"}\n            :outputs [{:key \"my\/s3\/output\/obj\/key.avi\"\n                       :preset-id \"1351620000001-000030\"}])<\/pre><\/div>\n<h3><a id=\"user-content-glacier\" class=\"anchor\" aria-hidden=\"true\" href=\"#glacier\"><\/a>Glacier<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.glacier]))\n(create-vault :vault-name \"my-vault\")\n(describe-vault :vault-name \"my-vault\")\n(list-vaults :limit 10)\n(upload-archive :vault-name \"my-vault\"\n                :body \"upload.txt\")\n(delete-archive :account-id \"-\"\n                :vault-name \"my-vault\"\n                :archive-id \"pgy30P2FTNu_d7buSVrGawDsfKczlrCG7Hy6MQg53ibeIGXNFZjElYMYFm90mHEUgEbqjwHqPLVko24HWy7DU9roCnZ1djEmT-1REvnHKHGPgkuzVlMIYk3bn3XhqxLJ2qS22EYgzg\", :checksum \"83a05fd1ce759e401b44fff8f34d40e17236bbdd24d771ec2ca4886b875430f9\", :location \"\/676820690883\/vaults\/my-vault\/archives\/pgy30P2FTNu_d7buSVrGawDsfKczlrCG7Hy6MQg53ibeIGXNFZjElYMYFm90mHEUgEbqjwHqPLVko24HWy7DU9roCnZ1djEmT-1REvnHKHGPgkuzVlMIYk3bn3XhqxLJ2qS22EYgzg\")\n(delete-vault :vault-name \"my-vault\")\n<\/pre><\/div>\n<h3><a id=\"user-content-identitymanagement\" class=\"anchor\" aria-hidden=\"true\" href=\"#identitymanagement\"><\/a>IdentityManagement<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.identitymanagement]))\n(def policy \"{\\\"Version\\\": \\\"2012-10-17\\\", \\\"Statement\\\": [{\\\"Action\\\": [\\\"s3:*\\\"], \\\"Effect\\\": \\\"Allow\\\", \\\"Resource\\\": [\\\"arn:aws:s3:::bucket-name\/*\\\"]}]}\")\n(create-user :user-name \"amazonica\")\n(create-access-key :user-name \"amazonica\")\n(put-user-policy\n  :user-name \"amazonica\"\n  :policy-name \"s3policy\"\n  :policy-document policy)\n<\/pre><\/div>\n<h3><a id=\"user-content-iot\" class=\"anchor\" aria-hidden=\"true\" href=\"#iot\"><\/a>IoT<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.iot :refer :all]))\n(list-things {})\n;; =&gt; {:things [{:thing-name \"YourThing\"}]}\n(create-thing :thing-name \"MyThing\")\n;; =&gt; {:thing-name \"MyThing\" :thing-arn \"arn:aws:iot:...thing\/MyThing\"}<\/pre><\/div>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.iotdata :refer :all]))\n(get-thing-shadow :thing-name \"MyThing\")<\/pre><\/div>\n<h3><a id=\"user-content-kinesis\" class=\"anchor\" aria-hidden=\"true\" href=\"#kinesis\"><\/a>Kinesis<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.kinesis]))\n(create-stream \"my-stream\" 1)\n(list-streams)\n(describe-stream \"my-stream\")\n(merge-shards \"my-stream\" \"shardId-000000000000\" \"shardId-000000000001\")\n(split-shard \"my-stream\" \"shard-id\" \"new-starting-hash-key\")\n;; write to the stream\n;; #'put-record takes the name of the stream, any value as data, and the partition key\n(let [data {:name \"any data\"\n            :col  #{\"anything\" \"at\" \"all\"}\n            :date now}]\n  (put-record \"my-stream\"\n              data\n              (str (UUID\/randomUUID))))\n;; if anything BUT a java.nio.ByteBuffer is supplied as the second\n;; argument, then the data will be transparently serialized and compressed\n;; using Nippy, and deserialized on calls to (get-records), or via a worker\n;; (see below), provided that a deserializer function is NOT supplied. If\n;; you do pass a ByteBuffer instance as the data argument, then you'll need\n;; to also provide a deserializer function when fetching records.\n;; For bulk uploading, we provide a `put-records` function which takes in a sequence of maps\n;; that contain the partition-key and data.  As with `put-record` the data will be handled via\n;; Nippy if it is not of a `java.nio.ByteBuffer`.\n(put-records \"my-stream\"\n             [{:partition-key \"x5h2ch\" :data [\"foo\" \"bar\" \"baz\"]}\n              {:partition-key \"x5j3ak\" :data [\"quux\"]}])\n;; optional :deserializer function which will be passed the raw\n;; java.nio.ByteBuffer representing the data blob of each record\n(defn- get-raw-bytes [byte-buffer]\n  (let [b (byte-array (.remaining byte-buffer))]\n    (.get byte-buffer b)\n    b))\n;; manually read from a specific shard\n;; this is not the preferred way to consume a shard\n(get-records :deserializer get-raw-bytes\n             :shard-iterator (get-shard-iterator \"my-stream\"\n                                                 shard-id\n                                                 \"TRIM_HORIZON\"))\n;; if no :deserializer function is supplied then it will be assumed\n;; that the records were put into Kinesis by Amazonica, and hence,\n;; the data was serialized and compressed by Nippy (e.g. Snappy)\n;; better way to consume a shard....create and run a worker\n;; :app :stream and :processor keys are required\n;; :credentials, :checkpoint and :dynamodb-adaptor-client? keys are optional\n;; if no :checkpoint is provided the worker will automatically checkpoint every 60 seconds\n;; alternatively, supply a numeric value for duration in seconds between checkpoints\n;; for full checkpoint control, set :checkpoint to false and return true from the\n;; :processor function only when you want checkpoint to be called\n;; if no :credentials key is provided the default authentication scheme is used (preferable),\n;; see the [Authentication] #(authentication) section above\n;; if no :dynamodb-adaptor-client? is provided, then it defaults to not using the\n;; DynamoDB Streams Kinesis Adaptor. Set this flag to true when consuming streams\n;; from DynamoDB\n;; returns the UUID assigned to this worker\n(worker! :app \"app-name\"\n         :stream \"my-stream\"\n         :checkpoint false ;; default to disabled checkpointing, can still force\n                           ;; a checkpoint by returning true from the processor function\n         :processor (fn [records]\n                      (doseq [row records]\n                        (println (:data row)\n                                 (:sequence-number row)\n                                 (:partition-key row)))))\n(delete-stream \"my-stream\")\n<\/pre><\/div>\n<h3><a id=\"user-content-kinesis-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#kinesis-analytics\"><\/a>Kinesis Analytics<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.kinesisanalytics :as ka]))\n(ka\/create-application\n  :application-name \"my-ka-app\"\n  :inputs [\n    {:name-prefix \"prefix_\"\n     :input-schema {:record-format {:record-format-type \"json\"}}\n     :kinesis-treams-input {:resource-ARN \"fobar\"}}\n  ]\n  :outputs [...]})\n<\/pre><\/div>\n<h3><a id=\"user-content-kinesisfirehose\" class=\"anchor\" aria-hidden=\"true\" href=\"#kinesisfirehose\"><\/a>KinesisFirehose<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.kinesisfirehose :as fh])\n  (:import [java.nio ByteBuffer]))\n;; List delivery streams\n(fh\/list-delivery-streams)\n;; =&gt; {:delivery-stream-names (\"test-firehose\" \"test-firehose-2\"), :has-more-delivery-streams false}\n(fh\/describe-delivery-stream :delivery-stream-name \"my-test-firehose\")\n;; =&gt; {:delivery-stream-description\n;;       {:version-id \"2\", ....}}\n(fh\/create-delivery-stream :delivery-stream-name \"my-test-firehose-2\"\n                           :s3DestinationConfiguration {:role-arn  \"arn:aws:iam:xxxx:role\/firehose_delivery_role\",\n                                                        :bucket-arn \"arn:aws:s3:::my-test-bucket\"})\n;; =&gt; {:delivery-stream-arn \"arn:aws:firehose:us-west-2:xxxxx:deliverystream\/my-test-firehose-2\"}\n;; Describe delivery stream\n(fh\/describe-delivery-stream cred :delivery-stream-name stream-name)\n;; Update destination\n(fh\/update-destination cred {:current-delivery-stream-version-id version-id\n                             :delivery-stream-name stream-name\n                             :destination-id destination-id\n                             :s3-destination-update {:BucketARN (str \"arn:aws:s3:::\" new-bucket-name)\n                                                     :BufferingHints {:IntervalInSeconds 300\n                                                                      :SizeInMBs 5}\n                                                     :CompressionFormat \"UNCOMPRESSED\"\n                                                     :EncryptionConfiguration {:NoEncryptionConfig \"NoEncryption\"}\n                                                     :Prefix \"string\"\n                                                     :RoleARN \"arn:aws:iam::123456789012:role\/firehose_delivery_role\"}})\n;; Put batch of records to stream. Records are converted to instances of ByteBuffer if possible. Sequences are converted to CSV formatted strings for injestion into RedShift.\n(fh\/put-record-batch cred stream-name [[1 2 3 4][\"test\" 2 3 4] \"\\\"test\\\",2,3,4\" (ByteBuffer. (.getBytes \"test,2,3,4\"))])\n;; Put individual record to stream.\n(fh\/put-record stream-name \"test\")\n;; Delete delivery stream\n(fh\/delete-delivery-stream \"stream-name\")\n<\/pre><\/div>\n<h3><a id=\"user-content-kms\" class=\"anchor\" aria-hidden=\"true\" href=\"#kms\"><\/a>KMS<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.kms]))\n(create-key)\n(list-keys)\n(disable-key \"xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx\")<\/pre><\/div>\n<h3><a id=\"user-content-logs\" class=\"anchor\" aria-hidden=\"true\" href=\"#logs\"><\/a>Logs<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.logs]))\n(describe-log-streams :log-group-name \"my-log-group\"\n                      :order-by \"LastEventTime\"\n                      :descending true)<\/pre><\/div>\n<h3><a id=\"user-content-lambda\" class=\"anchor\" aria-hidden=\"true\" href=\"#lambda\"><\/a>Lambda<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.lambda]))\n(let [role \"arn:aws:iam::123456789012:role\/some-lambda-role\"\n      handler \"exports.helloWorld = function(event, context) {\n                  console.log('value1 = ' + event.key1)\n                  console.log('value2 = ' + event.key2)\n                  console.log('value3 = ' + event.key3)\n                  context.done(null, 'Hello World')\n                }\"]\n  (create-function :role role :function handler))\n(invoke :function-name \"helloWorld\"\n        :payload \"{\\\"key1\\\": 1, \\\"key2\\\": 2, \\\"key3\\\": 3}\")\n<\/pre><\/div>\n<h3><a id=\"user-content-opsworks\" class=\"anchor\" aria-hidden=\"true\" href=\"#opsworks\"><\/a>OpsWorks<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.opsworks]))\n (create-stack :name \"my-stack\"\n               :region \"us-east-1\"\n               :default-os \"Ubuntu 12.04 LTS\"\n               :service-role-arn \"arn:aws:iam::676820690883:role\/aws-opsworks-service-role\")\n(create-layer :name \"webapp-layer\"\n              :stack-id \"dafa328e-c529-41af-89d3-12840a31abad\"\n              :enable-auto-healing true\n              :auto-assign-elastic-ips true\n              :volume-configurations [\n                {:mount-point \"\/data\"\n                 :number-of-disks 1\n                 :size 50}])\n(create-instance :hostname \"node-app-1\"\n                 :instance-type \"m1.large\"\n                 :stack-id \"dafa328e-c529-41af-89d3-12840a31abad\"\n                 :layer-ids [\"660d00da-c533-43d4-8c7f-2df240fd563f\"]\n                 :availability-zone \"us-east-1a\"\n                 :autoscaling-type \"LoadBasedAutoScaling\"\n                 :os \"Ubuntu 12.04 LTS\"\n                 :ssh-key-name \"admin\")\n(describe-stacks :stack-ids [\"dafa328e-c529-41af-89d3-12840a31abad\"])\n(describe-layers :stack-id \"dafa328e-c529-41af-89d3-12840a31abad\")\n(describe-instances :stack-id \"dafa328e-c529-41af-89d3-12840a31abad\"\n                    :layer-id \"660d00da-c533-43d4-8c7f-2df240fd563f\"\n                    :instance-id \"93bc5049-1bd4-49c8-a6ef-e84145807f71\")\n(start-stack :stack-id \"660d00da-c533-43d4-8c7f-2df240fd563f\")\n(start-instance :instance-id \"93bc5049-1bd4-49c8-a6ef-e84145807f71\")\n<\/pre><\/div>\n<h3><a id=\"user-content-pinpoint\" class=\"anchor\" aria-hidden=\"true\" href=\"#pinpoint\"><\/a>Pinpoint<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.pinpoint :as pp]))\n(defn app-id []\n  (-&gt; (pp\/get-apps {})\n      :applications-response\n      :item\n      first\n      :id))\n(pp\/create-segment {:application-id (app-id)})\n(pp\/create-campaign\n  {:application-id (app-id)\n   :write-campaign-request\n    {:segment-id \"a668b484bec94cb1252772032ecdf540\"\n     :name \"my-campaign\"\n     :schedule\n       {:frequency \"ONCE\"\n        :start-time \"2017-09-27T20:36:11+00:00\"}\n     :message-configuration\n       {:default-message\n         {:body \"hello world\"}}}})\n(pp\/send-messages\n  {:application-id (app-id)\n   :message-request\n     {:addresses {\"+18132401139\" {:channel-type \"SMS\"}}\n      :message-configuration {:default-message {:body \"hello world\"}}}})<\/pre><\/div>\n<h3><a id=\"user-content-redshift\" class=\"anchor\" aria-hidden=\"true\" href=\"#redshift\"><\/a>Redshift<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.redshift]))\n(create-cluster :availability-zone \"us-east-1a\"\n                :cluster-type \"multi-node\"\n                :db-name \"dw\"\n                :master-username \"scott\"\n                :master-user-password \"tiger\"\n                :number-of-nodes 3)\n<\/pre><\/div>\n<h3><a id=\"user-content-route53\" class=\"anchor\" aria-hidden=\"true\" href=\"#route53\"><\/a>Route53<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.route53]))\n(create-health-check :health-check-config {:port 80,\n                                           :type \"HTTP\",\n                                           :ipaddress \"127.0.0.1\",\n                                           :fully-qualified-domain-name \"example.com\"})\n(get-health-check :health-check-id \"ce6a4aeb-acf1-4923-a116-cd9ae2c30ee3\")\n(create-hosted-zone :name \"example69.com\"\n                    :caller-reference (str (java.util.UUID\/randomUUID)))\n(get-hosted-zone :id \"Z3TKY0VR5CH45U\")\n(list-hosted-zones)\n(list-health-checks)\n(list-resource-record-sets :hosted-zone-id \"ZN8D0HXQLVRRL\")\n(delete-health-check :health-check-id \"99999999-1234-4923-a116-cd9ae2c30ee3\")\n(delete-hosted-zone :id \"my-bogus-hosted-zone\")\n<\/pre><\/div>\n<h3><a id=\"user-content-route53domains\" class=\"anchor\" aria-hidden=\"true\" href=\"#route53domains\"><\/a>Route53Domains<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.route53domains]))\n(list-domains)\n(check-domain-availability :domain-name \"amazon.com\")\n(check-domain-transferability :domain-name \"amazon.com\")\n(get-domain-detail :domain-name \"amazon.com\")\n(let [contact {:first-name \"Michael\"\n               :last-name \"Cohen\"\n               :organization-name \"amazonica\"\n               :address-line1 \"375 11th St\"\n               :city \"San Francisco\"\n               :state \"CA\"\n               :zip-code \"94103-2097\"\n               :country-code \"US\"\n               :email \"\"\n               :phone-number \"+1.4158675309\"\n               :contact-type \"PERSON\"}]\n  (register-domain :domain-name \"amazon.com\"\n                   :duration-in-years 10\n                   :auto-renew true\n                   :tech-contact contact\n                   :admin-contact contact\n                   :registrant-contact contact))\n<\/pre><\/div>\n<h3><a id=\"user-content-s3\" class=\"anchor\" aria-hidden=\"true\" href=\"#s3\"><\/a>S3<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.s3]\n        [amazonica.aws.s3transfer]))\n(create-bucket \"two-peas\")\n;; put object with server side encryption\n(put-object :bucket-name \"two-peas\"\n            :key \"foo\"\n            :metadata {:server-side-encryption \"AES256\"}\n            :file upload-file)\n(copy-object bucket1 \"key-1\" bucket2 \"key-2\")\n(-&gt; (get-object bucket2 \"key-2\")\n    :input-stream\n    slurp)\n;; (note that the InputStream returned by GetObject should be closed,\n;; e.g. via slurp here, or the HTTP connection pool will be exhausted\n;; after several objects are retrieved)\n(delete-object :bucket-name \"two-peas\" :key \"foo\")\n(generate-presigned-url bucket1 \"key-1\" (-&gt; 6 hours from-now))\n(def file \"big-file.jar\")\n(def down-dir (java.io.File. (str \"\/tmp\/\" file)))\n(def bucket \"my-bucket\")\n;; set S3 Client Options\n(s3\/list-buckets\n  {:client-config {\n    :path-style-access-enabled false\n    :chunked-encoding-disabled false\n    :accelerate-mode-enabled false\n    :payload-signing-enabled true\n    :dualstack-enabled true\n    :force-global-bucket-access-enabled true}})\n;; list objects in bucket\n(list-objects-v2\n  {:bucket-name bucket\n   :prefix \"keys\/start\/with\/this\"  ; optional\n   :continuation-token (:next-continuation-token prev-response)})  ; when paging through results\n(def key-pair\n    (let [kg (KeyPairGenerator\/getInstance \"RSA\")]\n      (.initialize kg 1024 (SecureRandom.))\n      (.generateKeyPair kg)))\n;; put object with client side encryption\n(put-object :bucket-name bucket1\n            :key \"foo\"\n            :encryption {:key-pair key-pair}\n            :file upload-file)\n;; get object and decrypt\n(get-object :bucket-name bucket1\n            :encryption {:key-pair key-pair}\n            :key \"foo\")))))\n;; get tags for the bucket\n(get-bucket-tagging-configuration {:bucket-name bucket})\n;; get just object metadata, e.g. content-length without fetching content:\n(get-object-metadata :bucket-name bucket1\n                     :key \"foo\")\n;; put object from stream\n(def some-bytes (.getBytes \"Amazonica\" \"UTF-8\"))\n(def input-stream (java.io.ByteArrayInputStream. some-bytes))\n(put-object :bucket-name bucket1\n            :key \"stream\"\n            :input-stream input-stream\n            :metadata {:content-length (count some-bytes)}\n            :return-values \"ALL_OLD\")\n(let [upl (upload bucket\n                  file\n                  down-dir)]\n  ((:add-progress-listener upl) #(println %)))\n(let [dl  (download bucket\n                    file\n                    down-dir)\n      listener #(if (= :completed (:event %))\n                    (println ((:object-metadata dl)))\n                    (println %))]\n  ((:add-progress-listener dl) listener))\n;; setup S3 bucket for static website hosting\n(create-bucket bucket-name)\n(put-object bucket-name\n            \"index.html\"\n            (java.io.File. \"index.html\"))\n(let [policy {:Version \"2012-10-17\"\n              :Statement [{\n                :Sid \"PublicReadGetObject\"\n                :Effect \"Allow\"\n                :Principal \"*\"\n                  :Action [\"s3:GetObject\"]\n                  :Resource [(str \"arn:aws:s3:::\" bucket-name \"\/*\")]}]}\n      json (cheshire.core\/generate-string policy true)]\n  (set-bucket-policy bucket-name json))\n(set-bucket-website-configuration\n  :bucket-name bucket-name\n  :configuration {\n    :index-document-suffix \"index.html\"})\n(s3\/set-bucket-notification-configuration\n  :bucket-name \"my.bucket.name\"\n  :notification-configuration\n    {:configurations\n      {:some-config-name\n        {:queue \"arn:aws:sqs:eu-west-1:123456789012:my-sqs-queue-name\"\n         :events #{\"s3:ObjectCreated:*\"}\n         ;; list of key value pairs as maps or nexted 2 element list\n         :filter [{\"foo\" \"bar\"}\n                  {:baz \"quux\"}\n                  [\"key\" \"value\"]]}}})\n(s3\/set-bucket-tagging-configuration\n   :bucket-name \"my.bucket.name\"\n   :tagging-configuration\n     {:tag-sets [{:Formation \"notlive\" :foo \"bar\" :baz \"quux\"}]})\n<\/pre><\/div>\n<h3><a id=\"user-content-simpledb\" class=\"anchor\" aria-hidden=\"true\" href=\"#simpledb\"><\/a>SimpleDB<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.simpledb :as sdb]))\n(sdb\/create-domain :domain-name \"domain\")\n(sdb\/list-domains)\n(sdb\/put-attributes :domain-name \"domain\"\n                    :item-name \"my-item\"\n                    :attributes [{:name \"foo\"\n                                  :value \"bar\"}\n                                 {:name \"baz\"\n                                  :value 42}])\n(sdb\/select :select-expression\n            \"select * from `test.domain` where baz = '42' \")\n(sdb\/delete-domain :domain-name \"domain\")\n<\/pre><\/div>\n<h3><a id=\"user-content-simpleemail\" class=\"anchor\" aria-hidden=\"true\" href=\"#simpleemail\"><\/a>SimpleEmail<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.simpleemail :as ses]))\n(ses\/send-email :destination {:to-addresses [\"example@example.com\"]}\n                :source \"no-reply@example.com\"\n                :message {:subject \"Test Subject\"\n                          :body {:html \"testing 1-2-3-4\"\n                                 :text \"testing 1-2-3-4\"}})<\/pre><\/div>\n<h3><a id=\"user-content-simplesystemsmanager\" class=\"anchor\" aria-hidden=\"true\" href=\"#simplesystemsmanager\"><\/a>SimpleSystemsManager<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:require [amazonica.aws.simplesystemsmanagement :as ssm]))\n(ssm\/get-parameter :name \"my-param-name\")<\/pre><\/div>\n<h3><a id=\"user-content-simpleworkflow\" class=\"anchor\" aria-hidden=\"true\" href=\"#simpleworkflow\"><\/a>SimpleWorkflow<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.simpleworkflow]))\n(def domain \"my-wkfl\")\n(def version \"1.0\")\n(register-domain :name domain\n                 :workflow-execution-retention-period-in-days \"30\")\n(register-activity-type :domain domain\n                        :name \"my-worflow\"\n                        :version version)\n(register-workflow-type :domain domain\n                        :name \"my-worflow\"\n                        :version version)\n(deprecate-activity-type :domain domain\n                         :activity-type {:name \"my-worflow\"\n                                         :version version})\n(deprecate-workflow-type :domain domain\n                         :workflowType {:name \"my-worflow\"\n                                        :version version})\n(deprecate-domain :name domain)\n<\/pre><\/div>\n<h3><a id=\"user-content-sns\" class=\"anchor\" aria-hidden=\"true\" href=\"#sns\"><\/a>SNS<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.sns]))\n(create-topic :name \"my-topic\")\n(list-topics)\n(subscribe :protocol \"email\"\n           :topic-arn \"arn:aws:sns:us-east-1:676820690883:my-topic\"\n           :endpoint \"mcohen01@gmail.com\")\n(subscribe :protocol \"lambda\"\n           :topic-arn \"arn:aws:sns:us-east-1:676820690883:my-topic\"\n           :endpoint \"arn:aws:lambda:us-east-1:676820690883:function:my-function\")\n;; provide endpoint in creds for topics in non-default region\n(subscribe {:endpoint \"eu-west-1\"}\n       :protocol \"lambda\"\n           :topic-arn \"arn:aws:sns:eu-west-1:676820690883:my-topic\"\n           :endpoint \"arn:aws:lambda:us-east-1:676820690883:function:my-function\")\n(clojure.pprint\/pprint\n  (list-subscriptions))\n(publish :topic-arn \"arn:aws:sns:us-east-1:676820690883:my-topic\"\n         :subject \"test\"\n         :message (str \"Todays is \" (java.util.Date.))\n         :message-attributes {\"attr\" \"value\"})\n(unsubscribe :subscription-arn \"arn:aws:sns:us-east-1:676820690883:my-topic:33fb2721-b639-419f-9cc3-b4adec0f4eda\")\n<\/pre><\/div>\n<h3><a id=\"user-content-sqs\" class=\"anchor\" aria-hidden=\"true\" href=\"#sqs\"><\/a>SQS<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.sqs]))\n(create-queue :queue-name \"my-queue\"\n              :attributes\n                {:VisibilityTimeout 30 ; sec\n                 :MaximumMessageSize 65536 ; bytes\n                 :MessageRetentionPeriod 1209600 ; sec\n                 :ReceiveMessageWaitTimeSeconds 10}) ; sec\n;; full list of attributes at\n;; http:\/\/docs.aws.amazon.com\/AWSJavaSDK\/latest\/javadoc\/com\/amazonaws\/services\/sqs\/model\/GetQueueAttributesRequest.html\n(create-queue \"DLQ\")\n(list-queues)\n(def queue (find-queue \"my-queue\"))\n(assign-dead-letter-queue\n  queue\n  (find-queue \"DLQ\")\n  10)\n(send-message queue \"hello world\")\n(def msgs (receive-message queue))\n(delete-message (-&gt; msgs\n                    :messages\n                    first\n                    (assoc :queue-url queue)))\n(receive-message :queue-url queue\n                 :wait-time-seconds 6\n                 :max-number-of-messages 10\n                 :delete true ;; deletes any received messages after receipt\n                 :attribute-names [\"All\"])\n(-&gt; \"my-queue\" find-queue delete-queue)\n(-&gt; \"DLQ\" find-queue delete-queue)\n<\/pre><\/div>\n<h3><a id=\"user-content-stepfunctions\" class=\"anchor\" aria-hidden=\"true\" href=\"#stepfunctions\"><\/a>StepFunctions<\/h3>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns com.example\n  (:use [amazonica.aws.stepfunctions]))\n;this is to start the execution, then you need to run get-activity-task-result ultimately to monitor for pending requests from the state machine components\n;to execute a worker task.\n(start-state-machine \"{\\\"test\\\":\\\"test\\\"}\" \"arn:aws:states:us-east-1:xxxxxxxxxx:stateMachine:test-sf\")\n;this will block until it returns a task in the queue from a state machine execution,\n;so you need to run it in a while loop on the worker side of your app.\n(let [tr (get-activity-task-result \"arn:aws:states:us-east-1:xxxxxxxxx:activity:test-sf-activity\")\n      input (:input tr)\n      token (:task-token tr)]\n      (if (&lt;validate input here....&gt;)\n        (mark-task-success \"&lt;json stuff to pipe back into the state machine....&gt;\" token)\n        (mark-task-failure token))\n      )\n<\/pre><\/div>\n<h3><a id=\"user-content-acknowledgements\" class=\"anchor\" aria-hidden=\"true\" href=\"#acknowledgements\"><\/a>Acknowledgements<\/h3>\n<p>YourKit is kindly supporting the Amazonica open source project with its full-featured Java Profiler. YourKit, LLC is the creator of innovative and intelligent tools for profiling Java and .NET applications. Take a look at YourKit's leading software products: <a href=\"http:\/\/www.yourkit.com\/java\/profiler\/index.jsp\" rel=\"nofollow\">YourKit Java Profiler<\/a> and <a href=\"http:\/\/www.yourkit.com\/.net\/profiler\/index.jsp\" rel=\"nofollow\">YourKit .NET Profiler<\/a>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/0427980e3e2c00e6047f3f6e5a35f51d85872dc5\/68747470733a2f2f656e637279707465642d74626e322e677374617469632e636f6d2f696d616765733f713d74626e3a414e64394763525f4351775a496f7376444a55304e746d745f37304873627763724a70524f4b344a626633356d5163736a584e304d587a55\"><img src=\"https:\/\/camo.githubusercontent.com\/0427980e3e2c00e6047f3f6e5a35f51d85872dc5\/68747470733a2f2f656e637279707465642d74626e322e677374617469632e636f6d2f696d616765733f713d74626e3a414e64394763525f4351775a496f7376444a55304e746d745f37304873627763724a70524f4b344a626633356d5163736a584e304d587a55\" alt=\"YourKit logo\" data-canonical-src=\"https:\/\/encrypted-tbn2.gstatic.com\/images?q=tbn:ANd9GcR_CQwZIosvDJU0Ntmt_70HsbwcrJpROK4Jbf35mQcsjXN0MXzU\" \/><\/a><\/p>\n<p><a href=\"https:\/\/github.com\/mcohen01\/amazonica\"><img src=\"https:\/\/camo.githubusercontent.com\/53188ba7367c9e5657f2e39cbc208d2b25860b44\/68747470733a2f2f67612d626561636f6e2e61707073706f742e636f6d2f55412d353932353337342d31312f616d617a6f6e6963612f726561646d653f706978656c\" alt=\"Analytics\" data-canonical-src=\"https:\/\/ga-beacon.appspot.com\/UA-5925374-11\/amazonica\/readme?pixel\" \/><\/a><\/p>\n<h3><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h3>\n<p>Copyright (C) 2013 Michael Cohen<\/p>\n<p>Distributed under the Eclipse Public License, the same as Clojure.<\/p>","created_at":"2019-07-10T03:55:17+0000","updated_at":"2019-07-10T03:55:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":34,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/1295642?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":976,"title":"Smuggler: A Fifth Edition Background","url":"https:\/\/koboldpress.com\/smuggler-a-fifth-edition-background\/","content":"<p><a href=\"https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg.jpg\" rel=\"attachment wp-att-21007\"><img src=\"https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-300x204.jpg\" alt=\"Philippe Jacques de Loutherbourg (French, 1740\u20131812),  The Smugglers Return , 1801\" width=\"300\" height=\"204\" class=\"alignright size-medium wp-image-21007\" srcset=\"https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-300x204.jpg 300w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-150x102.jpg 150w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-250x170.jpg 250w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-220x150.jpg 220w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-187x127.jpg 187w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-453x309.jpg 453w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg-140x94.jpg 140w, https:\/\/koboldpress.com\/wp-content\/uploads\/2016\/03\/Philippe-Jacques-de-Loutherbourg.jpg 950w\" \/><\/a>Knife blades and harness buckles smeared with black soot; muffled oar-locks and a moonless, midnight-rendezvous, torchlights on the moors and signal lights beyond the breakers\u00a0. . .<\/p><p>Throughout the ages and across civilizations, there has and will always arise the inevitable desire, demand, even desperate need, for those things regulated or forbidden. These desires and demands, in turn, frequently spell opportunity to an entrepreneurial few. Individuals like you, who\u2014despite oppressive laws, exorbitant taxes, or religious or political creeds\u2014have ventured to make available to many what some would see reserved for the few.<\/p>\n<p>Perhaps you bootlegged monastic whiskey down secret mountain paths, spiriting them away to richer markets abroad. Or did you hone your craft duping city watchmen and tariff inspectors, moving shipments of stolen art, antiquities, or even people under their very noses and into their beloved city? Possibly your motivations rose higher and you rowed ashore through storm and darkness to bring weapons and supplies to desperate revolutionaries? Or evaded murderous bounty hunters, while escorting escaped slaves along forgotten caravan tracks in a last-ditch bid for freedom? In any case, you have left in your wake a colorful trail of bribed customs officials, befuddled border patrols, and \u201cmissing\u201d tax collectors. All in all, not bad seeing you have only just begun to (re-) make your name in the world.<\/p>\n<p><strong>Skill Proficiencies: <\/strong>\u00a0Deception, Persuasion<\/p>\n<p><strong>Tool Proficiencies: <\/strong>One vehicle proficiency of your choice (land or water), and choose one from navigator\u2019s tools, disguise kit, or forgery kit<\/p>\n<p><strong>Equipment: \u00a0<\/strong>A set of tools (matching your tool proficiency choice), a hooded signal lantern, dark traveler\u2019s clothes, and a leather belt-pouch containing 5 gp<\/p>\n<p><strong>Feature: \u201cEven <em>I<\/em> get boarded <em>sometimes<\/em>.\u201d <\/strong>As a master of contraband logistics and camouflage, all Intelligence (Investigation) and Wisdom (Perception) checks made to discover items that you have hidden on a person, vehicle, or pack animal are made at disadvantage. If the characters conducting the search can hear you, you can oppose their checks with a Charisma (Deception) check.<\/p>\n<p>Additionally, you have advantage on Intelligence (Investigation) and Wisdom (Perception) checks made to <em>discover<\/em> the concealment of goods or items on a person, vehicle, or pack animal.<\/p>\n<p><strong>Suggested Characteristics: <\/strong>Often well traveled and worldly, those who take up the smuggler\u2019s life are usually acute observers and mental note-takers, with a knack for numbers and a gift for smooth (fast) talk. Their adrenaline-fueled lifestyle of danger, intrigue, and constant vigilance has left more than a few with some \u201ccolorful\u201d and noteworthy personality traits.<\/p>\n<p><strong>d6\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Personality Trait<\/strong><\/p>\n<ol><li>I am a natural gambler; prone to risking it all if the pay-off seems exciting, amusing, or lucrative.<\/li>\n<li>Among the first things I do when arriving at a new place is begin to collect information and establish contacts to \u201cfeel out\u201d the import-export potential of a place.<\/li>\n<li>I tend to be vigilant about covering my tracks, checking for tails, knocking on walls, and so on.<\/li>\n<li>My long association with persons of dubious character has made me slow to trust or make friends.<\/li>\n<li>I hide my origins and history under a shroud of absolute secrecy. So much so that I haven\u2019t used my real name in years.<\/li>\n<li>I have an appraiser\u2019s eye; I am frequently found assessing the wholesale and logistics values of almost any type of potential \u201cswag\u201d I come across.<\/li>\n<li>In-depth planning and organization, security and reconnaissance details, time tables and signal codes\u2014I am a preparation perfectionist. I leave little to chance.<\/li>\n<li>Some might consider me cold or aloof. I am quiet, but supremely confident in my own abilities; I rarely rely on others for anything.<\/li>\n<\/ol><p><strong>d6\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Ideal<\/strong><\/p>\n<ol><li><strong>Self.<\/strong> It\u2019s nothing personal, truly. Just don\u2019t ever get between me and anything I need. (Neutral)<\/li>\n<li><strong>Benevolence.<\/strong> Sometimes, bending a few rules and breaking a few laws is necessary to help those in need. (Good)<\/li>\n<li><strong>Efficiency.<\/strong> Proactive thinking and organization, inventory control, balanced books and timetables\u2014these are the tools of a successful life. (Lawful)<\/li>\n<li><strong>Cupidity<\/strong>\u201cGreed is right, greed works, greed clarifies.\u2026\u201d (Evil)<\/li>\n<li><strong>Easy-Rider. <\/strong>My lifestyle will never be dictated by the borders, rules, or the laws of others. (Chaotic)<\/li>\n<li><strong>Aspiration.<\/strong> Someday I will be a ship or caravan master and blaze my own trails and trade routes to wealth and adventures as yet unheard of. (Any)<\/li>\n<\/ol><p><strong>d6\u00a0 Bond<\/strong><\/p>\n<ol><li>My word is my bond; I have never welched on a deal, dishonored a handshake, or double-crossed a partner.<\/li>\n<li>I am hastening to pay off an old debt, which could have assassins put on my trail any day now.<\/li>\n<li>A corrupt magistrate ruined my good name and drove me underground; soon I will have my revenge.<\/li>\n<li>Ill-gotten or not, a goodly portion of my gains goes to the humble, little monastery and the brothers there, who took me in that terrifying night so long ago.<\/li>\n<li>I am known to have been a witness to a horrible crime. I became a smuggler as a means to survive while remaining in hiding from the crime\u2019s perpetrator.<\/li>\n<li>I will do whatever I must to prove my superiority and to discredit or destroy a past associate, who has turned into fierce rival.<\/li>\n<\/ol><p><strong>d6\u00a0 Flaw.<\/strong><\/p>\n<ol><li>I am an obsessive haggler; from coppers for the shoeshine boy, to the uncut emeralds for the latest harbormaster\u2019s bribe, I never settle for first price.<\/li>\n<li>I always take a cut, skimming a small percentage off the top of almost anything I handle.<\/li>\n<li>Some think me paranoid because of my hyper-acute awareness and sense for spies, eaves-droppers, and imposters, magical or otherwise<\/li>\n<li>I\u2019ll go a long way to win the favor of the local beauties and fine-featured folk.<\/li>\n<li>My loyalties go to the highest bidder or most lucrative contract, regardless of personal feelings or desires.<\/li>\n<li>I mentally figure out the net worth of people in terms of liquid assets.<\/li>\n<\/ol>","created_at":"2019-07-06T04:02:13+0000","updated_at":"2019-07-06T04:02:13+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"koboldpress.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":975,"title":"The Ultimate Synthesizer Tutorial","url":"https:\/\/www.syntorial.com\/","content":"<div class=\"e12561-1 x-section\">\n      \n<div class=\"e12561-2 x-container max width\">\n    \n<div class=\"e12561-3 x-column x-sm x-1-2\" data-fade=\"1\" data-x-element=\"column\" data-x-params=\"{&quot;fade&quot;:true}\">\n    \n<div class=\"e12561-4 x-text\">\n  <p class=\"bklt-txt\">\u201cSyntorial is the most direct route between hearing a sound in your head and knowing how to bring it to life.\u201d<\/p><p class=\"bklt-ftr\">\u2013 Electronic Musician Magazine<\/p><\/div>\n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2015\/03\/editors-choice.png\" width=\"245\" height=\"90\" \/><\/div>\n<div class=\"e12561-6 x-column x-sm x-1-2\">\n    \n<div class=\"e12561-7 x-frame ho-rw-one-vid x-frame-video-embed\">\n  <div class=\"x-frame-inner\">\n    \n<div class=\"e12561-7 x-video x-video-embed\">\n  <iframe width=\"450\" height=\"260\" src=\"https:\/\/www.youtube.com\/embed\/a6ZK4X7vryQ\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/div>\n  <\/div>\n<\/div>\n<\/div>\n<\/div>\n  <\/div>\n<div class=\"e12561-8 x-section\">\n      \n<div class=\"e12561-9 x-container max width\">\n    \n<div class=\"e12561-10 x-column x-sm x-1-1\">\n    \n<div class=\"e12561-12 x-text\">\n  <p>Syntorial is video game-like training software, that will teach you how to program synth patches by ear. With almost 200 lessons, combining video demonstrations with interactive challenges, you\u2019ll get hands on experience programming patches on a built-in soft synth, and learn everything you need to know to start making your own sounds with ease.<\/p>\n<p>This isn\u2019t your ordinary synth tutorial. <strong>This is Syntorial.<\/strong><\/p><\/div>\n<div class=\"x-text cs-ta-center\"><p><a class=\"x-btn x-btn-x-large\" href=\"https:\/\/www.syntorial.com\/try-for-free\/\" data-options=\"thumbnail: ''\">TRY FOR FREE<\/a><\/p>\n<\/div><\/div>\n<\/div>\n  <\/div>\n<div class=\"e12561-14 x-section indus-imag\">\n      \n<div class=\"e12561-18 x-container max width\">\n    \n<div class=\"e12561-19 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2015\/03\/editors-choice.png\" width=\"245\" height=\"90\" \/><\/div>\n<div class=\"e12561-21 x-column x-sm x-3-4\">\n    \n<div class=\"e12561-22 x-text\">\n  <p>\u201cSyntorial is the most direct route between hearing a sound in your head and knowing how to bring it to life. Lesson-by-lesson, module-by-module, Syntorial shows you almost everything that synths can do without getting bogged down in theory. Learn to aurally recognize VCO waveforms, filter bandwidths, LFO rates, ADSR parameters, effects types, and more. Get the edge on sound design by educating yourself with this new approach to tutorial software.\u201d\u00a0<em class=\"hmbdy-row-clr\">\u2013<a href=\"https:\/\/www.emusician.com\/gear\/rewind-the-2014-editors-choice-awards\" target=\"_blank\" rel=\"noopener\">Electronic Musician Magazine<\/a><\/em><\/p><\/div>\n<\/div>\n<\/div>\n<div class=\"e12561-23 x-container max width\">\n    \n<div class=\"e12561-24 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2015\/02\/keyboard_magazine_logo.png\" width=\"880\" height=\"301\" \/><\/div>\n<div class=\"e12561-26 x-column x-sm x-3-4\">\n    \n<div class=\"e12561-27 x-text\">\n  <p>\u201cThe most comprehensive and fun tool for learning synthesizer programming, hands down.\u201d\u00a0<em class=\"hmbdy-row-clr\">\u2013<a href=\"https:\/\/www.keyboardmag.com\/gear\/syntorial-the-most-fun-way-to-learn-synth-programming\" target=\"_blank\">Keyboard Magazine<\/a><\/em><\/p><\/div>\n<\/div>\n<\/div>\n<div class=\"e12561-28 x-container max width\">\n    \n<div class=\"e12561-29 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2015\/02\/musictech_award_logos.png\" width=\"880\" height=\"301\" \/><\/div>\n<div class=\"e12561-31 x-column x-sm x-3-4\">\n    \n<div class=\"e12561-32 x-text\">\n  <p>\u201cThe best training in synth sound-design we\u2019ve come across. A unique and fun way to train your ears in re-creating synth sounds.\u201d\u00a0<em class=\"hmbdy-row-clr\">\u2013<a href=\"http:\/\/www.musictech.net\/2014\/02\/audible-genius-syntorial-review\/\" target=\"_blank\">MusicTech Magazine<\/a><\/em><\/p><\/div>\n<\/div>\n<\/div>\n  <\/div>\n<div class=\"e12561-33 x-section custo-imag\">\n      \n<div class=\"e12561-37 x-container max width\">\n    \n<div class=\"e12561-38 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2016\/04\/morganpage.png\" width=\"173\" height=\"148\" \/><\/div>\n<div class=\"e12561-40 x-column x-sm x-3-4\">\n    \n<div class=\"e12561-41 x-text\">\n  <p>\u201cSyntorial is an excellent resource for learning and mastering synthesis. It\u2019s really important to understand the fundamentals that are often lost in today\u2019s YouTube tutorials. I\u2019m using it to re-learn synthesis and get a better grasp of concepts I missed along the way.\u201d\u00a0<em class=\"hmbdy-row-clr\">Morgan Page, Two-Time Grammy Nominee<\/em><\/p><\/div>\n<\/div>\n<\/div>\n<div class=\"e12561-42 x-container max width\">\n    \n<div class=\"e12561-43 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2015\/02\/mike-huckaby.png\" width=\"172\" height=\"148\" \/><\/div>\n<div class=\"e12561-45 x-column x-sm x-3-4\">\n    \n<div class=\"e12561-46 x-text\">\n  <p>\u201cWhere was this program when i was trying to learn synthesis ? This program is the most cutting edge method to learn synthesis. Highly recommended.\u201d\u00a0<em class=\"hmbdy-row-clr\">\u2013 Mike Huckaby<\/em><\/p><\/div>\n<\/div>\n<\/div>\n<div class=\"e12561-47 x-container max width\">\n    \n<div class=\"e12561-48 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2017\/04\/Keith_Hancock.png\" width=\"173\" height=\"148\" \/><\/div>\n<div class=\"e12561-50 x-column x-sm x-3-4\">\n    \n<div class=\"e12561-51 x-text\">\n  <p>\u201cIn all of my research, I haven\u2019t found another piece of software that teaches synthesis as well as Syntorial.  It moves at a great pace, and the logical sequence of the lessons truly trains the ears to hear the slight differences in tone when adjusting the various parameters.  It is the perfect tool for anyone looking to increase their understanding of the complex world of sound design.\u201d\u00a0<em class=\"hmbdy-row-clr\">\u2013 Keith Hancock, 2017 Grammy Music Educator of the Year<\/em><\/p><\/div>\n<\/div>\n<\/div>\n  <\/div>\n<div class=\"e12561-52 x-section no-btm-padng\">\n      \n<div class=\"e12561-80 x-container max width no-top-padng\">\n    \n<div class=\"e12561-82 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2017\/03\/classroom_1.jpg\" width=\"300\" height=\"225\" \/><\/div>\n<div class=\"e12561-84 x-column x-sm x-1-4\">\n    \n  <img alt=\"image\" src=\"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2017\/04\/classroom_2.jpg\" width=\"300\" height=\"225\" \/><\/div>\n<\/div>\n  <\/div>\n<div class=\"e12561-87 x-section\">\n      \n<div class=\"e12561-91 x-container max width\">\n    \n<div class=\"e12561-92 x-column x-sm x-1-1\">\n    \n<div class=\"e12561-93 x-text\">\n  Test drive Syntorial today. Download a free demo which contains the first 22 (out of 199) lessons, and there\u2019s no time limit.<\/div>\n<div class=\"e12561-94 x-text\">\n  <a class=\"x-btn x-btn-x-large\" href=\"https:\/\/www.syntorial.com\/try-for-free\/\" data-options=\"thumbnail: ''\">TRY FOR FREE<\/a><\/div>\n<\/div>\n<\/div>\n  <\/div>","created_at":"2019-06-26T20:13:36+0000","updated_at":"2019-06-26T20:13:36+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.syntorial.com","preview_picture":"https:\/\/mk0syntorial49drlf7i.kinstacdn.com\/wp-content\/uploads\/2015\/03\/LogoBlackNoBGMedium.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":974,"title":"Learning Synths","url":"https:\/\/learningsynths.ableton.com\/","content":"Learning Synths\n        \n        <noscript>\n            \n        <\/noscript>\n        \n        \n        \n        <p>Your browser does not support this site. Please try using a recent version of Chrome, Firefox, Safari, Edge, or Opera.<\/p>\n        <noscript>\n            <p>Please enable JavaScript in order to use this site.<\/p>\n        <\/noscript>","created_at":"2019-06-26T14:16:59+0000","updated_at":"2019-06-26T14:16:59+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"learningsynths.ableton.com","preview_picture":"https:\/\/learningsynths.ableton.com\/meta\/og-learningsynths-1200x630.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":973,"title":"Startup idea checklist","url":"https:\/\/www.defmacro.org\/2019\/03\/26\/startup-checklist.html","content":"<p>I\u2019ve been tinkering with different startup ideas and needed a good\nchecklist to think through them. There are great templates for this\nalready: <a href=\"https:\/\/apply.ycombinator.com\/\">The YC application<\/a>, Amazon\u2019s <a href=\"https:\/\/www.quora.com\/What-is-Amazons-approach-to-product-development-and-product-management\/answer\/Ian-McAllister\">internal press release<\/a>,\nand Sequoia\u2019s <a href=\"https:\/\/www.sequoiacap.com\/article\/writing-a-business-plan\/\">Writing a Business Plan<\/a>. I found myself mixing and\ntweaking these templates because they don\u2019t exactly match my model of\nthe world, so I wrote up my own list.<\/p><p>I use this list both to develop ideas and filter them. If you adopt\nit, be careful about using it as a filter. Remember that in the early\nstages, good ideas are very easy to kill.<\/p>\n<ol><li>Who are the users?\n<br \/>&lt;= 70 chars<\/li>\n  <li>What is the essence of their dissatisfaction? If they read this\nanswer, would they say \u201cthanks, I wish I\u2019d thought of putting it\nthat way\u201d?\n<br \/>&lt;= 240 chars\n<br \/><a href=\"https:\/\/twitter.com\/benedictevans\/status\/1110538673873805314\">@benedictevans<\/a>: \u201cThe iTunes Store\nsolved a user problem. So did the App Store. And so did Spotify and\nApple Music, and indeed Apple News. But what user problem is solved\nby Apple\u2019s commissioning TV shows?\u201d<\/li>\n  <li>What are you building for them?\n<br \/>&lt;= 70 chars\n<br \/><a href=\"https:\/\/www.amazon.com\/Effective-Executive-Definitive-Harperbusiness-Essentials\/dp\/0060833459\/\">Peter Drucker<\/a>: Is the product\nbeing designed <strong>for<\/strong> the customer, or <strong>at<\/strong> the customer?<\/li>\n  <li>Write a tweet from a hypothetical customer explaining the product\nand how it eliminates their dissatisfaction.\n<br \/><a href=\"https:\/\/twitter.com\/BrianNorgard\/status\/1110915013085028353\">@BrianNorgard<\/a>: No one cares\nabout your product. Who built it, its features, the origin story \u2014\nit\u2019s all superfluous. People only find value in what your product\ncan do for them right now. Save people time. Save people money.\nGive people an escape. The selfish hand will always govern.\n<br \/><a href=\"https:\/\/www.amazon.com\/Effective-Executive-Definitive-Harperbusiness-Essentials\/dp\/0060833459\/\">Peter Drucker<\/a>: Are you <strong>really<\/strong>\ndoing the best you can to help the customer?<\/li>\n  <li>Write a blog post title for your product launch. Is it\nsurprising? Is it new? Will your target customers want to click on\nit? Will they want to share the link? Will they still share it the\nnext day?\n<br \/>&lt;= 70 chars<\/li>\n  <li>Write the first paragraph of your product announcement blog post.\nInclude the product name, an explanation of what the product is,\nthe target market, the main benefit, and the call to action.\n<br \/>&lt;= 240 chars<\/li>\n  <li>What \u201cmetrics of goodness\u201d do your target customers care about?\nDoes your product dominate every available alternative on these\nmetrics?\n<br \/>&lt;= 240 chars\n<br \/><strong>See also:<\/strong> <a href=\"https:\/\/www.jwz.org\/doc\/worse-is-better.html\">The Rise of Worse is Better<\/a>,\n<a href=\"https:\/\/www.artima.com\/weblogs\/viewpost.jsp?thread=24807\">Worse is worse<\/a><\/li>\n<\/ol>\n<ol start=\"8\"><li>Fill in the bottom-up market size equation: <code class=\"highlighter-rouge\">NUM_USERS * ACV =\nMARKET_SIZE<\/code>. Are your numbers credible? Find a good reference\nclass if you\u2019re building something completely new.\n<br \/><strong>See also:<\/strong> <a href=\"https:\/\/wiki.lesswrong.com\/wiki\/Shut_up_and_multiply\">Shut up and multiply<\/a><\/li>\n  <li>Which subset of your target customers are so constrained by the\nstatus quo, they\u2019ll welcome a buggy product?\n<br \/>&lt;= 140 chars<\/li>\n  <li>List your first ten customers.\n<br \/>&lt;= 240 chars\n<br \/><strong>See also:<\/strong> <a href=\"http:\/\/paulgraham.com\/ds.html\">Do Things that Don\u2019t Scale<\/a><\/li>\n  <li>Which playbook will you use to get customers after the first ten?\n<br \/>&lt;= 240 chars\n<br \/><strong>See also:<\/strong> <a href=\"http:\/\/christophjanz.blogspot.com\/2014\/10\/five-ways-to-build-100-million-business.html\">Five ways to build a $100 million business<\/a><\/li>\n  <li>What would need to be true in 18 months for you to get essentially\nunlimited cheap capital? How will you achieve that?\n<br \/>&lt;= 240 chars<\/li>\n<\/ol>\n<ol start=\"13\"><li>Why now? What\u2019s true about the world that nobody else figured out\nyet?\n<br \/>&lt;= 240 chars<\/li>\n  <li>What is the most ambitious achievable milestone for your company\nwithin a 25 year time horizon?\n<br \/>&lt;= 70 chars<\/li>\n  <li>Is your product a credible advance toward this milestone?\n<br \/>Yes\/no<\/li>\n  <li>What\u2019s the <em>next<\/em> credible advance toward this milestone? The one\nafter that? The one after that?\n<br \/>&lt;= 240 chars\n<br \/><strong>See also:<\/strong> <a href=\"https:\/\/twitter.com\/spakhm\/status\/1111411471869595648\">Tesla master plan<\/a>,\n<a href=\"https:\/\/twitter.com\/stevesi\/status\/1111092932252041216\">iPhone runs OSX<\/a><\/li>\n  <li>How will you build a moat?\n<br \/>&lt;= 240 chars\n<br \/><strong>See also:<\/strong> <a href=\"https:\/\/hbr.org\/1979\/03\/how-competitive-forces-shape-strategy\">How Competitive Forces Shape Strategy<\/a><\/li>\n<\/ol>\n<ol start=\"18\"><li>What would reaching your 25 year milestone mean for the world? Is\nthis future <em>really<\/em> exciting? How many years of your life would you\ngive up to teleport there? If you found yourself in this\ncounterfactual world, would you want to go back?\n<br \/>&lt;= 140 chars<\/li>\n  <li>If another company was working on this idea and not you, what would\nyou think about it? Would you join them?\n<br \/>Yes\/no<\/li>\n  <li>Imagine yourself standing in front of your team, investors, family,\nand friends. You\u2019ve failed, and they\u2019re waiting for you to speak.\nWhat will you say? Are you willing to work on this problem given\nthat failure is the default?\n<br \/>&lt;= 480 chars\n<br \/><strong>See also:<\/strong> <a href=\"https:\/\/twitter.com\/statsepi\/status\/1021334815822548992\">Your intervention won\u2019t work<\/a><\/li>\n<\/ol>\n<ol start=\"21\"><li>\n    <p>What\u2019s your company\u2019s stock ticker symbol?\n<br \/><a href=\"https:\/\/twitter.com\/sama\/status\/571733273996488704\">@sama<\/a>: \u201cit\u2019s easy\/fun to say every\nnew startup you hear about is bad. you will usually be right. you will\nnever be successful.\u201d<\/p>\n  <\/li>\n  <li>\n    <p>Is it likely to be the most important company started this year?\n<br \/><strong>George Orwell<\/strong>: \u201cWhoever is winning\nat the moment will always seem to be invincible.\u201d<\/p>\n  <\/li>\n<\/ol><hr \/><p><em>Thanks to <a href=\"https:\/\/twitter.com\/darryl_ramm\/status\/1111829084202397696\">Darryl Ramm<\/a> for feedback on this post.<\/em><\/p>","created_at":"2019-06-24T07:01:04+0000","updated_at":"2019-06-24T07:01:04+0000","published_at":"2019-03-26T00:00:00+0000","published_by":["Slava Akhmechet"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"www.defmacro.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":972,"title":"RethinkDB: why we failed","url":"https:\/\/www.defmacro.org\/2017\/01\/18\/why-rethinkdb-failed.html","content":"<p>When we <a href=\"https:\/\/rethinkdb.com\/blog\/rethinkdb-shutdown\/\">announced<\/a> that RethinkDB is shutting\ndown, I promised to write a post-mortem. I took some time to process\nthe experience, and I can now write about it clearly.<\/p><p>In the <a href=\"https:\/\/news.ycombinator.com\/item?id=12649414\">HN discussion thread<\/a> people proposed many\nreasons for why RethinkDB failed, from inexplicable perversity of\nhuman nature and clever machinations of MongoDB\u2019s marketing people, to\nfailure to build an experienced go-to-market team, to lack of numeric\ntype support beyond 64-bit <code class=\"highlighter-rouge\">float<\/code>. I aggregated the comments into a\nlist of proposed failure reasons <a href=\"https:\/\/gist.github.com\/coffeemug\/af8dcb6f653a7f9c31daedbbdaa3402c\">here<\/a>.<\/p>\n<p>Some of these reasons have a ring of truth to them, but they\u2019re\nsymptoms rather than causes. For example, saying that we failed to\nmonetize is tautological. It doesn\u2019t illuminate the reasons for <em>why<\/em>\nwe failed.<\/p>\n<p>In hindsight, two things went wrong \u2013 we picked a terrible market and\noptimized the product for the wrong metrics of goodness. Each mistake\nlikely cut RethinkDB\u2019s valuation by one to two orders of magnitude. So\nif we got either of these right, RethinkDB would have been the size of\nMongoDB, and if we got both of them right, we eventually could have\nbeen the size of Red Hat[1].<\/p>\n<p>Our thinking went something like this. New companies aren\u2019t getting\nbuilt on top of Oracle, so there is a window of opportunity to build a\nnew infrastructure company. The database market is huge. If we build a\nproduct that captures some of that market, we\u2019ll end up building a\nvery successful company.<\/p>\n<p>Unfortunately you\u2019re not in the market you think you\u2019re in \u2013 you\u2019re\nin the market <em>your users<\/em> think you\u2019re in. And our users clearly\nthought of us as an open-source developer tools company, because\nthat\u2019s what we really were. Which turned out to be very unfortunate,\nbecause the open-source developer tools market is one of the <a href=\"https:\/\/techcrunch.com\/2014\/02\/13\/please-dont-tell-me-you-want-to-be-the-next-red-hat\/\">worst\nmarkets<\/a> one could possibly <a href=\"https:\/\/techcrunch.com\/2014\/08\/22\/will-developer-tools-startups-ever-find-investors\/\">end up\nin<\/a>. Thousands of people used RethinkDB, often in\nbusiness contexts, but most were willing to pay less for the lifetime\nof usage than the price of a single Starbucks coffee (which is to say,\nthey weren\u2019t willing to pay anything at all).<\/p>\n<p>This wasn\u2019t because the product was so good people didn\u2019t need to pay\nfor support, or because developers don\u2019t control budgets, or because\nof failure of capitalism. The answer is basic\nmicroeconomics. Developers love building developer tools, often for\nfree. So while there is massive demand, the supply vastly outstrips\nit. This <a href=\"https:\/\/en.wikipedia.org\/wiki\/Porter's_five_forces_analysis\">drives<\/a> the number of alternatives up, and the\nprices down to zero.<\/p>\n<p>To see how this plays out for other companies consider MongoDB (valued\nat roughly $1.6B with ~700 employees), and Docker (valued at roughly\n$1B with ~300 employees). Both companies completely dominate in their\nrespective markets. Two very rough rules of thumb for private growth\nstage technology companies is that valuations are a 10x multiple of\nannual revenue, and that <a href=\"https:\/\/www.saastr.com\/how-to-figure-out-your-competitors-revenues-in-about-70-seconds\/\">revenue per employee<\/a> is\naround $200K\/year. Which means that MongoDB\u2019s annual revenue is around\n$140-$160M, and Docker\u2019s annual revenue is around $60-$100M.<\/p>\n<p>That looks pretty good, until you look at dominant B2B technology\ncompanies in markets that <em>aren\u2019t<\/em> developer tools. Companies like\nSalesForce, or Palantir, or Box (which faces stiff competition). All\nof a sudden MongoDB and Docker start looking tiny.<\/p>\n<p>And these are massive successes. If relatively established companies\nwith existing partnerships, distribution infrastructure, and access to\nlarge accounts are having trouble growing, what does it mean for a\nstartup in its germination stage?<\/p>\n<p>For us, it meant an intractable customer acquisition funnel. If a\nstartup in a fertile B2B market has to process a hundred leads to get\nto ten opportunities to get to a single sale, for a developer tools\nstartup that number goes up 10x. You have access to plenty of high\nquality prospects \u2013 lots of people are downloading your product and\nengaging with you, but you have to burn through a ridiculous number of\nleads to converge to a single sale.<\/p>\n<p>This has disastrous domino effects. It demoralizes the team, and\nmakes it very challenging to attract investment and hire top\ntalent. In turn, that constrains your resources so you can\u2019t make\nsufficient investment in product and distribution. Startups live and\ndie by momentum, and early distribution challenges almost always doom\nyou to eventual death.<\/p>\n<p>Ok, so the market is bad, but other developer tools companies are\nstill selling a lot of product. Why not RethinkDB?<\/p>\n<p>While we couldn\u2019t do anything about the dynamics of the market (other\nthan building something else), the product decisions were entirely\nwithin our control. We wanted to build an elegant, robust, and\nbeautiful product, so we optimized for the following metrics:<\/p>\n<ul><li><strong>Correctness.<\/strong> We made very strict guarantees, and\n<a href=\"https:\/\/aphyr.com\/posts\/329-jepsen-rethinkdb-2-1-5\">fulfilled<\/a> them <a href=\"https:\/\/aphyr.com\/posts\/330-jepsen-rethinkdb-2-2-3-reconfiguration\">religiously<\/a>.<\/li>\n  <li><strong>Simplicity of the interface.<\/strong> We took on most of the complexity\nin the implementation, so application developers wouldn\u2019t have to.<\/li>\n  <li><strong>Consistency.<\/strong> We made everything from the query language, to the\nclient drivers, to cluster configuration, to documentation, to the\nmarketing copy on the front page as consistent as possible.<\/li>\n<\/ul><p>If these trade-offs seem familiar, they\u2019re straight from the <a href=\"http:\/\/dreamsongs.com\/RiseOfWorseIsBetter.html\">worse is\nbetter<\/a> essay. It turned out that correctness,\nsimplicity of the interface, and consistency are the wrong <a href=\"http:\/\/www.artima.com\/weblogs\/viewpost.jsp?thread=24807\">metrics of\ngoodness<\/a> for most users. The majority of users wanted\nthese three trade-offs instead:<\/p>\n<ul><li><strong>Timely arrival.<\/strong> They wanted the product to actually exist when\nthey needed it, not three years later.<\/li>\n  <li><strong>Palpable speed.<\/strong> People wanted RethinkDB to be fast on workloads\nthey <em>actually tried<\/em>, rather than \u201creal world\u201d workloads we\nsuggested. For example, they\u2019d write quick scripts to measure how\nlong it takes to insert ten thousand documents without ever reading\nthem back. MongoDB mastered these workloads brilliantly, while we\nfought the losing battle of educating the market.<\/li>\n  <li><strong>A use case.<\/strong> We set out to build <em>a good database system<\/em>, but\nusers wanted <em>a good way to do X<\/em> (e.g. a good way to store JSON\ndocuments from hapi, a good way to store and analyze logs, a good\nway to create reports, etc.)<\/li>\n<\/ul><p>It\u2019s not that we didn\u2019t try to ship quickly, make RethinkDB fast, and\nbuild the ecosystem around it to make doing useful work easy. We\ndid. But correct, simple, and consistent software takes a very long\ntime to build. That put us three years behind the market.<\/p>\n<p>By the time we felt RethinkDB satisfied our design goals and we were\nconfident enough to recommend it to be used in production, almost\neveryone was asking \u201chow is RethinkDB different from MongoDB?\u201d We\nworked hard to explain why correctness, simplicity, and consistency\nare important, but ultimately these weren\u2019t the metrics of goodness\nthat mattered to most users.<\/p>\n<p>To be honest, it hurt. It hurt a lot. It was unfathomable to us why\npeople would choose a system that barely does the thing it\u2019s supposed\nto do (store data), has a big kernel lock, throws away errors at\nrandom, implements single node features that stop working when you\nshard, has a barely working sharding system despite it being one of\nthe core features of the product, provides essentially no correctness\nguarantees, and exposes a hodge-podge of interfaces that have no\ndiscernible consistency or unity of vision.<\/p>\n<p>Every time MongoDB shipped a new release and people congratulated them\non making improvements, I felt pangs of resentment. They\u2019d announce\nthey fixed the BKL, but really they\u2019d get the granularity level down\nfrom a database to a collection. They\u2019d add more operations, but\ninstead of a composable interface that fits with the rest of the\nsystem, they\u2019d simply bolt on one-off commands. They\u2019d make sharding\nimprovements, but it was obvious they were unwilling or unable to make\neven rudimentary data consistency guarantees.<\/p>\n<p>But over time I learned to appreciate the wisdom of the\ncrowds. MongoDB turned regular developers into heroes <em>when people\nneeded it<\/em>, not years after the fact. It made data storage fast, and\nlet people ship products quickly. And over time, MongoDB grew up. One\nby one, they fixed the issues with the architecture, and now it is an\nexcellent product. It may not be as beautiful as we would have wanted,\nbut it does the job, and it does it well.<\/p>\n<p>When it became clear in mid-2014 that we couldn\u2019t compete, we worked\nhard to differentiate from MongoDB. We found a very elegant way to add\n<a href=\"https:\/\/rethinkdb.com\/blog\/1.16-release\/\">realtime push<\/a>, hoping to enable developers to build a\ngeneration of apps they couldn\u2019t build before. But that wasn\u2019t\nenough. Suddenly we found ourselves competing with Meteor and\nFirebase, companies that were dedicated to solving the realtime\nproblem for years before we even thought of it. Again we were three\nyears behind the market, and again we found ourselves unable to\ncompete.<\/p>\n<p>A few people suggested that we should have built a cloud offering. We\nactually did have one in the works, so it\u2019s an interesting topic I\u2019d\nlike to cover.<\/p>\n<p>The obvious problem with a small database company building a cloud\nservice is that it pattern matches to a common startup failure mode \u2013\nsplitting focus. Building, shipping, and operating reliable\nmulti-tenant cloud services is hard. It requires non-trivial expertise\nand resources, so if you go down that path you find yourself running\ntwo startups at once. But we were facing an existential threat and\nwere rapidly running out of options, so we gave it a shot\nanyway. Let\u2019s suppose for the moment we could have pulled it off.<\/p>\n<p>Our reasoning went like this. A database cloud offering could mean one\nof three things: managed hosting, database as a service (DBaaS), or\nvalue-added platform as a service (PaaS). Let\u2019s do a quick back of the\nnapkin market analysis using a $200K\/employee in annual revenue <a href=\"https:\/\/www.saastr.com\/how-to-figure-out-your-competitors-revenues-in-about-70-seconds\/\">rule\nof thumb<\/a> we used above:<\/p>\n<table><thead><tr><th>\u00a0<\/th>\n      <th>Managed Hosting<\/th>\n      <th>DBaaS<\/th>\n      <th>PaaS<\/th>\n    <\/tr><\/thead><tbody><tr><td><strong>Company<\/strong><\/td>\n      <td>Compose.io, mLab<\/td>\n      <td>FaunaDB<\/td>\n      <td>Parse, Firebase, Meteor<\/td>\n    <\/tr><tr><td><strong>Employees<\/strong><\/td>\n      <td>~30<\/td>\n      <td>~30<\/td>\n      <td>~30<\/td>\n    <\/tr><tr><td><strong>Revenue<\/strong><\/td>\n      <td>&lt; $10M<\/td>\n      <td>&lt; $10M<\/td>\n      <td>&lt; $10M<\/td>\n    <\/tr><\/tbody><\/table><p>So these markets are small, even smaller than the database market\nitself. But could one of them be a better bet than others?<\/p>\n<p>Managed hosting is essentially running the database for people on AWS\nso they don\u2019t have to. The alternative to using these services is\nsetting up the database on AWS yourself. That\u2019s a pain, but it isn\u2019t\nactually <em>that<\/em> hard. So there is a very hard cap on how much managed\ndatabase hosting services can charge. Considering that Compose.io and\nmLab are offering MongoDB which has one to two orders of magnitude\nmore users than RethinkDB, we reasoned that offering managed hosting\nwouldn\u2019t make a dent.<\/p>\n<p>Database as a service is a more complex version of managed hosting \u2013\nDBaaS offerings abstract node management from the user entirely. You\nsimply run your queries and the system handles them. You don\u2019t know\nanything about how many nodes are run under the hood. This business is\nvery challenging \u2013 partly because DBaaS companies have to compete with\nthe giants (e.g. DynamoDB and DocumentDB), and partly because\ncustomers are very resistant to completely hand off data management to\na startup when there are so many other substitutes and alternatives\n(do <em>you<\/em> know anyone who uses a DBaaS offering from a startup?) So a\nDBaaS offering was out.<\/p>\n<p>The last option was to build a value-added platform as a service. We\nthought this was a promising direction because here we had a massive\ntechnical advantage. Firebase and Meteor had to build\napplication-level realtime logic on top of MongoDB, which\nfundamentally limits the realtime querying capabilities and\nperformance at scale. On the other hand, we controlled the stack all\nthe way down, so we could offer significant advantages Firebase and\nMeteor couldn\u2019t build.<\/p>\n<p>So we built <a href=\"http:\/\/horizon.io\/\">Horizon<\/a> and started working on Horizon Cloud \u2013\na way for users to deploy and scale RethinkDB\/Horizon apps. The\nchallenges of building three large projects (RethinkDB, Horizon, and\nHorizon Cloud) with a very small team eventually caught up with us,\nand we never managed to ship the cloud offering before we ran out of\nmoney. Kudos to the engineering team, though. They came very, very\nclose.<\/p>\n<p>There is one more level of root cause analysis that we can do. Why did\nwe pick a bad market and optimize the product for the wrong metrics?<\/p>\n<p>When I was a little kid I wanted to build my own radio. I made a box\nout of plywood, threw some metal junk inside, and connected the box to\na power cord. I had books on electronics at home, but didn\u2019t think I\nneeded them \u2013 I had unwavering faith that I could do it on my\nown. Eventually I did build a working receiver, but it took me years\nto finally realize I needed to learn basic electronics.<\/p>\n<p>Early RethinkDB was quite a bit like that. We had no intuition for\nproducts or markets, so we\u2019d go through the motions of building a\ncompany without actually understanding what we were doing. What\u2019s\nmore, we had enormous <a href=\"https:\/\/doc.research-and-analytics.csfb.com\/docView?document_id=1048541371&amp;serialid=mofPYk1Y4WanTeErbeMtPx6ur0SCIcSlaZ7sKGPdQQU%3D\">optimism bias<\/a>. Just like\nphysicians know that gifts from pharmaceutical companies have biasing\neffects for other physicians but believe they are immune from the\neffect, we believed we were immune from the laws of economics and the\nmath of running a business. The math, of course, eventually caught up\nwith us.<\/p>\n<p>Could we have done anything to avoid these mistakes? Not any more than\nI could have built a working radio as a little kid. We were\nunconsciously incompetent, and it took years for that incompetence to\nbecome conscious.<\/p>\n<p>A few people pointed out that we would have done better if we had\nbuilt an experienced go-to-market team. That\u2019s 100% true, but the\ntiming of our personal development didn\u2019t line up with the needs of\nthe company. Initially we didn\u2019t know we needed go-to-market\nexpertise, so we didn\u2019t seek to include it on the founding team[2]. By\nthe time we built up a mental model that maps well to reality, we\nfound ourselves short on cash, in a difficult market filled with\ncapable competitors, and a product that\u2019s three years behind. By then,\nthe best go-to-market team in the world couldn\u2019t have saved us.<\/p>\n<p>Many people have very strong feelings about the developer tools\nmarket. Engineers love building developer tools, so they badly want\ndeveloper tools companies to thrive.<\/p>\n<p>I am hesitant to dismiss the market entirely \u2013 partly because I don\u2019t\nwant to generalize from a single experience, partly because I don\u2019t\nlike saying \u201cit cannot be done\u201d, and partly because there are quite a\nfew exceptions. GitHub, MongoDB, and Docker have built formidable\ncompanies. GitLab and Unity seem to be doing well.<\/p>\n<p>If you do set out to build a developer tools company, tread\ncarefully. The market is filled with good alternatives. User\nexpectations are high and prices are low. Think deeply about the value\nyou\u2019re offering to the customer. Remember \u2013 wanting the world to be a\ncertain way doesn\u2019t make it so.<\/p>\n<p>In 2009, we were pitching the early idea for RethinkDB (we had no\nsoftware yet) to an audience of investors at the YCombinator demo\nday. We ended the pitch with a slide of three key points to\nremember. \u201cIf you only remember three things about RethinkDB,\u201d we\nsaid, \u201cremember these.\u201d It worked. People didn\u2019t remember anything\nelse about the pitch, but they did remember the three points at the\nend.<\/p>\n<p>I\u2019ll now leave you with three key points to remember. If you remember\nanything about this post, remember these:<\/p>\n<ul><li>Pick a large market but build for specific users.<\/li>\n  <li>Learn to recognize the talents you\u2019re missing, then work like hell\nto get them on your team.<\/li>\n  <li>Read <a href=\"http:\/\/www.economist.com\/\">The Economist<\/a> religiously. It will make you better faster.<\/li>\n<\/ul><hr \/><p><em>[1] Don\u2019t read into these numbers too closely. I\u2019m ball-parking it,\nbut it should give you a general idea of the cost of these mistakes.<\/em><\/p>\n<p><em>[2] Incidentally, recognizing good business people without having\nstrong business intuition is about as hard as recognizing good\nengineers without having a strong intuition for engineering.<\/em><\/p>","created_at":"2019-06-23T16:38:15+0000","updated_at":"2019-06-23T16:38:15+0000","published_at":"2017-01-18T00:00:00+0000","published_by":["Slava Akhmechet"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"www.defmacro.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":971,"title":"atomontage\/xterm-color","url":"https:\/\/github.com\/atomontage\/xterm-color","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p><a href=\"https:\/\/opensource.org\/licenses\/BSD-2-Clause\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/6def34e1aa4e2e9e81448c8a57cf3e09d8af28cf\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4253442d626c75652e737667\" alt=\"https:\/\/img.shields.io\/badge\/license-BSD-blue.svg\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-BSD-blue.svg\" \/><\/a>\n  <a href=\"http:\/\/melpa.org\/#\/xterm-color\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/bf30294fffecd40f494edbec1010c4b4bada485b\/68747470733a2f2f6d656c70612e6f72672f7061636b616765732f787465726d2d636f6c6f722d62616467652e737667\" alt=\"https:\/\/melpa.org\/packages\/xterm-color-badge.svg\" data-canonical-src=\"https:\/\/melpa.org\/packages\/xterm-color-badge.svg\" \/><\/a>\n  <a href=\"http:\/\/stable.melpa.org\/#\/xterm-color\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/dc663957809e2d96bce9c0c259c3e12088a8dfa1\/68747470733a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f787465726d2d636f6c6f722d62616467652e737667\" alt=\"https:\/\/stable.melpa.org\/packages\/xterm-color-badge.svg\" data-canonical-src=\"https:\/\/stable.melpa.org\/packages\/xterm-color-badge.svg\" \/><\/a><\/p>\n<p>xterm-color.el is an ANSI control sequence to text-property translator<\/p>\n<ul><li>Regular ANSI color support<\/li>\n  <li>XTERM 256 color support<\/li>\n  <li>AIXTERM bright foreground color<\/li>\n  <li>AIXTERM bright background color (since 1.8)<\/li>\n  <li>Use bold instead of bright (since 1.8)<\/li>\n  <li>Works with compilation-mode<\/li>\n  <li>Works with eshell<\/li>\n  <li>Bonus: Faster than ansi-color.el<\/li>\n<\/ul>\n<p>Interactively or from Emacs Lisp, call xterm-color-colorize-buffer to colorize an entire buffer.<\/p>\n<p>In Emacs Lisp code call xterm-color-filter to propertize strings that you can then insert into\n  a buffer. All state is kept in buffer-local variables which means that\n  control sequences can span <i>xterm-color-filter<\/i> call boundaries.<\/p>\n<p>You may customize <i>xterm-color-debug<\/i> (default NIL, if T you will get warnings in <b>Messages<\/b> when\n  unsupported escape sequences are encountered), <i>xterm-color-use-bold-for-bright<\/i> (default NIL),\n  <i>xterm-color-names<\/i>,  <i>xterm-color-names-bright<\/i>. Additionally, you may set\n  <i>xterm-color-preserve-properties<\/i> to T (default NIL, should be set to T if using xterm-color with\n  eshell, see below). A buffer-local face attribute cache is used since 1.8 to improve performance.\n  This means that if you change <i>xterm-color-names<\/i> or <i>xterm-color-names-bright<\/i> or\n  <i>xterm-color-use-bold-for-bright<\/i> at runtime and want the changes to take effect in a pre-existing\n  buffer with activated xterm-color, you will need to run <i>xterm-color-clear-cache<\/i> in relevant buffer.<\/p>\n<p>Example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(let ((buffer (generate-new-buffer \"*xterm-color-test*\")))\n  (with-current-buffer buffer\n    (insert (xterm-color-filter \"\\x1b[0;1;3;4\"))\n    (insert (xterm-color-filter \";35\"))\n    (insert (xterm-color-filter \";51mThis is only a test\"))\n    (insert (xterm-color-filter \"\\x1b[0m\")))\n  (switch-to-buffer buffer))<\/pre><\/div>\n<p>If you are inserting into a buffer that has activated font locking, you\n  <b>NO LONGER<\/b> need to set <i>font-lock-unfontify-region-function<\/i> to\n  <i>xterm-color-unfontify-region<\/i> as the latter has been removed.<\/p>\n<p>You can replace ansi-color.el with xterm-color for all comint buffers,\n  but this may create problems with modes that propertize strings and\n  feed them through comint-preoutput-filter-functions since xterm-color-filter\n  will strip all text properties.<\/p>\n<p>The recommended configuration is to remove ansi-color-process-output from\n  comint-output-filter-functions and add xterm-color-filter as the <b>first<\/b>\n  hook in the buffer-local comint-preoutput-filter-functions for any comint-based\n  mode that you would like it to affect (e.g. shell-mode). Additionally, it is\n  recommended to disable font-locking for shell-mode buffers\n  <a href=\"https:\/\/github.com\/atomontage\/xterm-color\/issues\/28\">since it seems to interact badly with comint and drastically affect performance<\/a>.\n  Font locking in shell-mode buffers is superfluous since xterm-color.el will\n  handle faces fine by itself.<\/p>\n<p>An example configuration for shell-mode (M-x shell) is shown below:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq comint-output-filter-functions\n      (remove 'ansi-color-process-output comint-output-filter-functions))\n(add-hook 'shell-mode-hook\n          (lambda ()\n            ;; Disable font-locking in this buffer to improve performance\n            (font-lock-mode -1)\n            ;; Prevent font-locking from being re-enabled in this buffer\n            (make-local-variable 'font-lock-function)\n            (setq font-lock-function (lambda (_) nil))\n            (add-hook 'comint-preoutput-filter-functions 'xterm-color-filter nil t)))\n;; Also set TERM accordingly (xterm-256color)\n;; You can also use it with eshell (and thus get color output from system ls):\n(require 'eshell)\n(add-hook 'eshell-before-prompt-hook\n          (lambda ()\n            (setq xterm-color-preserve-properties t)))\n(add-to-list 'eshell-preoutput-filter-functions 'xterm-color-filter)\n(setq eshell-output-filter-functions (remove 'eshell-handle-ansi-color eshell-output-filter-functions))\n;;  Don't forget to setenv TERM xterm-256color\n<\/pre><\/div>\n<p>You may use `compilation-shell-minor-mode\u2019 with shell-mode buffers\n  and the configuration previously described.<\/p>\n<p>For standalone compilation buffers use the following configuration:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq compilation-environment '(\"TERM=xterm-256color\"))\n(add-hook 'compilation-start-hook\n          (lambda (proc)\n            ;; We need to differentiate between compilation-mode buffers\n            ;; and running as part of comint (which at this point we assume\n            ;; has been configured separately for xterm-color)\n            (when (eq (process-filter proc) 'compilation-filter)\n              ;; This is a process associated with a compilation-mode buffer.\n              ;; We may call `xterm-color-filter' before its own filter function.\n              (set-process-filter\n               proc\n               (lambda (proc string)\n                 (funcall 'compilation-filter proc\n                          (xterm-color-filter string)))))))<\/pre><\/div>\n<p>Standalone compilation-mode buffer support is experimental, report any\n  issues that may arise.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>M-x xterm-color-test\nFor comint or eshell:\nM-x shell || M-x eshell\nperl tests\/xterm-colortest &amp;&amp; perl tests\/256colors2.pl\n<\/pre><\/div>\n<p><a href=\"https:\/\/github.com\/atomontage\/xterm-color\/blob\/master\/img\/xterm-color.png\"><img src=\"https:\/\/github.com\/atomontage\/xterm-color\/raw\/master\/img\/xterm-thumb.png\" alt=\"img\/xterm-thumb.png\" \/><\/a>  <a href=\"https:\/\/github.com\/atomontage\/xterm-color\/blob\/master\/img\/godwars2.png\"><img src=\"https:\/\/github.com\/atomontage\/xterm-color\/raw\/master\/img\/godwars2-thumb.png\" alt=\"img\/godwars2-thumb.png\" \/><\/a><\/p>\n<p>The code comes with a BSD-style license so you can basically do with it\n  whatever you want.<\/p>\n<p>xristos (AT) sdf (DOT) lonestar (DOT) org<\/p>\n<\/article>","created_at":"2019-06-18T13:42:00+0000","updated_at":"2019-06-18T13:42:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/453886?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":970,"title":"Presenting the Eshell","url":"http:\/\/www.howardism.org\/Technical\/Emacs\/eshell-present.html","content":"<p>\nDemonstrated the iPython approach by entering the following into an\n<code>ipython<\/code> REPL:\n<\/p><pre class=\"example\">\nPython 2.7.12 (default, Nov 19 2016, 06:48:10)\nType \"copyright\", \"credits\" or \"license\" for more information.\nIPython 2.4.1 -- An enhanced Interactive Python.\n?         -&gt; Introduction and overview of IPython's features.\n%quickref -&gt; Quick reference.\nhelp      -&gt; Python's own help system.\nobject?   -&gt; Details about 'object', use 'object??' for extra details.\nIn [1]: 2 ** 60\nOut[1]: 1152921504606846976\nIn [2]: cd \/tmp\/testing\n\/tmp\/testing\nIn [3]: ls\npi.py  pi.rb  README  src\/  tests\/\nIn [4]: cat pi.py\nimport math\nprint(math.pi)\nprint(math.cos(math.pi))\nIn [5]: chmod a+x pi.py\n  File \"&lt;ipython-input-5-b9e30b9b8e31&gt;\", line 1\n    chmod a+x pi.py\n          ^\nSyntaxError: invalid syntax\nIn [6]: run pi\n3.14159265359\n-1.0\nIn [7]: system ruby pi.rb\nOut[7]: ['-0.9999999999964793', '-3.141592653589793']\nIn [8]: cat README\nCiao, all you cool cats.\nOh, and hey to all my dawgs.\nIn [9]: def cat(arg=None):\n   ...:     return 'Meow!'\n   ...:\nIn [10]: cat\nOut[10]: &lt;function __main__.cat&gt;\nIn [11]: cat()\nOut[11]: 'Meow!'\nIn [12]: cat README\n  File \"&lt;ipython-input-12-01ccd53eccb4&gt;\", line 1\n    cat README\n             ^\nSyntaxError: invalid syntax\n<\/pre><p>\nIn answer to your question, I haven\u2019t looked to see why we have an\narray when calling the <code>system<\/code> function.\n<\/p><p>\nIf you want to see what <i>shell-like<\/i> functions iPython has, type <code>%<\/code>\nand hit the Tab key, which shows something like:\n<\/p><pre class=\"example\">\nIn [13]: %\nDisplay all 122 possibilities? (y or n)\n%%!                      %doctest_mode            %pfile\n%%HTML                   %ed                      %pinfo\n%%SVG                    %edit                    %pinfo2\n%%bash                   %env                     %popd\n%%capture                %gui                     %pprint\n%%debug                  %hist                    %precision\n%%file                   %history                 %profile\n%%html                   %install_default_config  %prun\n%%javascript             %install_ext             %psearch\n%%latex                  %install_profiles        %psource\n%%perl                   %killbgscripts           %pushd\n%%prun                   %ldir                    %pwd\n%%pypy                   %less                    %pycat\n%%python                 %lf                      %pylab\n%%python2                %lk                      %quickref\n%%python3                %ll                      %recall\n%%ruby                   %load                    %rehashx\n%%script                 %load_ext                %reload_ext\n%%sh                     %loadpy                  %rep\n%%svg                    %logoff                  %rerun\n%%sx                     %logon                   %reset\n%%system                 %logstart                %reset_selective\n%%time                   %logstate                %rm\n%%timeit                 %logstop                 %rmdir\n%%writefile              %ls                      %run\n%alias                   %lsmagic                 %save\n%alias_magic             %lx                      %sc\n%autocall                %macro                   %store\n%autoindent              %magic                   %sx\n%automagic               %man                     %system\n%bookmark                %matplotlib              %tb\n%cat                     %mkdir                   %time\n%cd                      %more                    %timeit\n%clear                   %mv                      %unalias\n%colors                  %notebook                %unload_ext\n%config                  %page                    %who\n%cp                      %paste                   %who_ls\n%cpaste                  %pastebin                %whos\n%debug                   %pdb                     %xdel\n%dhist                   %pdef                    %xmode\n%dirs                    %pdoc\n<\/pre>","created_at":"2019-06-18T13:37:49+0000","updated_at":"2019-06-18T13:37:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.howardism.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":969,"title":"Procedural content generation: creating a universe","url":"https:\/\/news.ycombinator.com\/item?id=20145984","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D20145984\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"20145984\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/blog.rabidgremlin.com\/2015\/01\/14\/procedural-content-generation-creating-a-universe\/\" class=\"storylink\">Procedural content generation: creating a universe<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=rabidgremlin.com\">rabidgremlin.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        52 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=20145984\">4 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=20145984&amp;goto=item%3Fid%3D20145984\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Procedural%20content%20generation%3A%20creating%20a%20universe&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Procedural%20content%20generation%3A%20creating%20a%20universe\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=20145984&amp;auth=af5ca2c7e39ae0b2bcc75f33434eed5689388cb9\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=20145984\">9\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-06-10T22:03:36+0000","updated_at":"2019-06-10T22:03:36+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":968,"title":"johannesgerer\/buchhaltung","url":"https:\/\/github.com\/johannesgerer\/buchhaltung","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><blockquote>\n<p>What advantages does he derive from the system of book-keeping by double entry! It is among the finest inventions of the human mind; every prudent master of a house should introduce it into his economy.\n-- Johann Wolfgang von Goethe<\/p>\n<\/blockquote>\n<p><em>Buchhaltung<\/em> (['bu\u02d0\u03c7\u02cchalt\u028a\u014b], German <em>book keeping<\/em>), written in Haskell, helps you keep track of your finances on the command line with minimal effort. It provides tools that help you in creating a complete ledger of all your bank and savings accounts', credit cards', and other transactions, in a text-based ledger format, that is readable by the <a href=\"http:\/\/www.ledger-cli.org\/\" rel=\"nofollow\">ledger CLI tool<\/a> and its many <a href=\"http:\/\/plaintextaccounting.org\/\" rel=\"nofollow\">derivatives<\/a>.<\/p>\n<ul><li>Fetch your bank transaction directly via FinTS\/HBCI\/OFXDirectConnect<\/li>\n<li>Import transactions from\n<ul><li>PayPal<\/li>\n<li>Barclaycard Visa<\/li>\n<li>comdirect Visa<\/li>\n<li>Revolut<\/li>\n<li>Barclays UK CSV<\/li>\n<li>NatWest International CSV<\/li>\n<li>and Monefy expense tracker app<\/li>\n<li>... or define your own formats<\/li>\n<\/ul><\/li>\n<li>Semi-automatically match transactions to accounts using Bayesian classification<\/li>\n<li>Semi-automatic transaction entry with meaningful suggestions in keyboard-based speed mode<\/li>\n<li>It is couples\/room-mates aware: Create several transaction simultaneously (see <a href=\"#multi-user-add\">Multi-user add<\/a>)<\/li>\n<\/ul><h2><a id=\"user-content-status--aim\" class=\"anchor\" aria-hidden=\"true\" href=\"#status--aim\"><\/a>Status &amp; aim<\/h2>\n<p>I am actively and successfully using this software since 2010 and my ledger now contains more than 12,000 transactions accurately and continuously tracking the finances of my spouse and me including four checking and two savings accounts, one credit card, two PayPal accounts, two cash wallets in EUR, bitcoin trading (both physical and on exchanges) and other currencies like USD, GPB used on trips.<\/p>\n<p>The software is in alpha phase and I am looking for early adopters and their use cases. The aim of this stage is to agree about the functionality and customizability and produce a first shippable version, that can be used without tinkering with the source.<\/p>\n<p>Right now, I am using it on Linux but it should also run wherever GHC runs.<\/p>\n<h2><a id=\"user-content-prerequisites\" class=\"anchor\" aria-hidden=\"true\" href=\"#prerequisites\"><\/a>Prerequisites<\/h2>\n<ul><li>\n<p><a href=\"https:\/\/haskell-lang.org\/get-started\" rel=\"nofollow\">Haskell Stack<\/a>, more specifically the <a href=\"https:\/\/www.haskell.org\/\" rel=\"nofollow\">Glasgow Haskell Compiler<\/a> and <a href=\"https:\/\/docs.haskellstack.org\/en\/stable\/README\/\" rel=\"nofollow\">Stack<\/a><\/p>\n<p>Required to <strong>compile<\/strong> the software.<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/www2.aquamaniac.de\/sites\/aqbanking\/index.php\" rel=\"nofollow\">AqBanking Command Line Tool<\/a> (optional)<\/p>\n<p>This is required for <strong>direct retrieval of bank transactions<\/strong> via FinTS\/HBCI\/EBICS (Germany) or OFXDirectConnect (USA, Canada, UK). Packages available e.g. on Ubuntu (aqbanking-tools) and ArchLinux (aqbanking). (AqBanking is also the used by <a href=\"http:\/\/wiki.gnucash.org\/wiki\/AqBanking\" rel=\"nofollow\">GnuCash<\/a> for this purpose.)<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/dbacl.sourceforge.net\/\" rel=\"nofollow\">dbacl<\/a> (optional, needed to run <a href=\"#match-accounts\"><code>match<\/code><\/a>)<\/p>\n<p>Bayesian classifier used to <strong>match transaction to accounts<\/strong>. Packages available e.g. on Ubuntu and ArchLinux (<a href=\"https:\/\/aur.archlinux.org\/packages\/dbacl\/\" rel=\"nofollow\">AUR<\/a>).<\/p>\n<\/li>\n<li>\n<p><a href=\"http:\/\/www.ledger-cli.org\/\" rel=\"nofollow\">ledger CLI tool<\/a> or a compatible <a href=\"http:\/\/plaintextaccounting.org\/\" rel=\"nofollow\">derivative<\/a> (optional)<\/p>\n<p>... to <strong>query the ledger, create balance and report statements<\/strong>, <a href=\"http:\/\/hledger.org\/hledger-web.html\" rel=\"nofollow\">web interface<\/a>, etc.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-under-the-hood\" class=\"anchor\" aria-hidden=\"true\" href=\"#under-the-hood\"><\/a>Under the hood<\/h3>\n<p>AqBanking or any other supported CSV source provides the raw transaction data. The <code>buchhaltung<\/code> binary converts this data into the ledger format and puts the original source in the comments; thereby producing selfcontained transactions without the need for an external database or log. These comments  are used later for duplicates detection and account classification.<\/p>\n<p>The <code>buchhaltung<\/code> binary internally uses the great <code>hledger<\/code> library. The resulting ledger files can be processed\/queried with one of the <code>hledger<\/code> tools or any other compatible tool.<\/p>\n<h2><a id=\"user-content-download-compile--install\" class=\"anchor\" aria-hidden=\"true\" href=\"#download-compile--install\"><\/a>Download, compile &amp; install<\/h2>\n<div class=\"highlight highlight-source-shell\"><pre># download\ngit clone https:\/\/github.com\/johannesgerer\/buchhaltung.git\ncd buchhaltung\n# compile and install (usually in ~\/.local\/bin)\nstack install\n<\/pre><\/div>\n<h2><a id=\"user-content-configure\" class=\"anchor\" aria-hidden=\"true\" href=\"#configure\"><\/a>Configure<\/h2>\n<ol><li>\n<p>Create a folder that will hold all your config and possibly ledger files:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>mkdir ~\/.buchhaltung\ncp \/path\/to\/buchhaltung\/config.yml ~\/.buchhaltung\/config.yml<\/pre><\/div>\n<p>If you want a folder under a different location, either create a symlink or set the <code>BUCHHALTUNG<\/code> environment variable to that location.<\/p>\n<\/li>\n<li>\n<p>Edit the <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>config.yml<\/code><\/a>.<\/p>\n<\/li>\n<li>\n<p>Make sure the configured ledger files exist.<\/p>\n<\/li>\n<\/ol><ul><li>The <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>config.yml<\/code><\/a> file provides excessive comments.<\/li>\n<li>This readme documents most functionality.<\/li>\n<li>Every command and subcommand shows a help message when invoked with <code>-h<\/code>.<\/li>\n<li>Read the haddock documentation and source code on <a href=\"https:\/\/hackage.haskell.org\/package\/buchhaltung\" rel=\"nofollow\">Hackage<\/a>.<\/li>\n<li>Open an issue.<\/li>\n<li>Write an email.<\/li>\n<\/ul><h2><a id=\"user-content-first-usage--clean\" class=\"anchor\" aria-hidden=\"true\" href=\"#first-usage--clean\"><\/a>First usage \/ clean<\/h2>\n<p>To initialize AqBanking after you edited the config file, you need to run:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung setup<\/pre><\/div>\n<p>To clean everything aqbanking related remove the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>aqBanking.configDir<\/code><\/a>  and rerun the <code>setup<\/code> command.<\/p>\n<h3><a id=\"user-content-manual-aqbanking-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#manual-aqbanking-setup\"><\/a>Manual AqBanking setup<\/h3>\n<p>Currently only the HBCI <code>PinTan<\/code> method is supported by the <code>setup<\/code> command (pull requests welcome). For other methods or if the AqBanking setup fails due to other reasons, you can configure AqBanking manually into the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>aqBanking.configDir<\/code><\/a> (see for help <a href=\"https:\/\/www.aquamaniac.de\/sites\/download\/download.php?package=09&amp;release=09&amp;file=01&amp;dummy=aqbanking4-handbook-20091231.pdf\" rel=\"nofollow\">here<\/a> or <a href=\"https:\/\/wiki.gnucash.org\/wiki\/AqBanking\" rel=\"nofollow\">here<\/a>, usually via <code>aqhbci-tool4 -C &lt;aqBanking.configDir&gt;<\/code>).<\/p>\n<p>So far, I have not tested the OFXDirectConnect capabilities of AqBanking (please share your experiences), but <code>buchhaltung<\/code> should transparently support any method offered by AqBanking.<\/p>\n<h2><a id=\"user-content-importing-transactions\" class=\"anchor\" aria-hidden=\"true\" href=\"#importing-transactions\"><\/a>Importing transactions<\/h2>\n<p>There various ways (including from PayPal CSV files) to import transactions into your configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>ledgers.imported<\/code><\/a> file. They are presented in the following, but consult<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung import -h<\/pre><\/div>\n<p>and the <code>-h<\/code> calls to its subcommands to see the currently available functionality.<\/p>\n<p>The accounts of the imported transactions will be taken from the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>bankAccounts<\/code><\/a> and the offsetting balance will be posted to an account named <code>TODO<\/code>, and will be replaced by <a href=\"#match-accounts\"><code>match<\/code><\/a>.<\/p>\n<p>The original source information will be included in the second posting's comment and used for learning the account mappings and to find and handle duplicates.<\/p>\n<h3><a id=\"user-content-aqbanking\" class=\"anchor\" aria-hidden=\"true\" href=\"#aqbanking\"><\/a>AqBanking<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung update<\/pre><\/div>\n<p>This command fetches and imports all available transactions from all configured AqBanking connections.<\/p>\n<p>In order to import an existing file in <a href=\"https:\/\/hackage.haskell.org\/package\/buchhaltung-0.0.3\/docs\/buchhaltung\/src\/Buchhaltung-Importers.html#aqbankingImport\" rel=\"nofollow\">AqBanking format<\/a>, use:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung import &lt;file.csv&gt; aqbanking<\/pre><\/div>\n<h3><a id=\"user-content-other-formats\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-formats\"><\/a>Other formats<\/h3>\n<p>For other CSV formats you have two options:<\/p>\n<ol><li>\n<p>Use the importers for PayPal, Barclaycard Visa, comdirect Visa, Revolut, and Monefy expense tracker app.<\/p>\n<\/li>\n<li>\n<p>Bring your CSV into a currently supported format. The AqBanking format works especially well<\/p>\n<\/li>\n<li>\n<p>Hack the source, which is fun and should be easy since you only need to\ncopy and adapt from the other importers.<\/p>\n<\/li>\n<\/ol><p>You can look at <code>aqbankingImporter<\/code>, <code>comdirectVisaImporter<\/code> and\n<code>paypalImporter<\/code>\nin\n<a href=\"https:\/\/hackage.haskell.org\/package\/buchhaltung-0.0.3\/docs\/buchhaltung\/Buchhaltung-Importers.html\" rel=\"nofollow\">Buchhaltung\/Importers.hs<\/a> (or on <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/src\/Buchhaltung\/Importers.hs\">GitHub<\/a>).<\/p>\n<h3><a id=\"user-content-resolve-duplicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#resolve-duplicates\"><\/a>Resolve duplicates<\/h3>\n<p>Banks often minimally change the way they report transactions which leads to unwanted duplicates.<\/p>\n<p>When importing, <em>Buchhaltung<\/em> will identify duplicates based on <code>([(Commodity,Quantity)], AccountName, Day)<\/code> and interactively resolve them by showing the user what fields have changed. If there are several candidates, it sorts the candidates according to similarity using <code>levenshteinDistance<\/code> from <a href=\"https:\/\/hackage.haskell.org\/package\/edit-distance-0.2.1.3\" rel=\"nofollow\">edit-distance<\/a>. (See <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/src\/Buchhaltung\/Uniques.hs#L27\"><code>Buchhaltung.Uniques.addNew<\/code><\/a>)<\/p>\n<h2><a id=\"user-content-match-accounts\" class=\"anchor\" aria-hidden=\"true\" href=\"#match-accounts\"><\/a>Match accounts<\/h2>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung match<\/pre><\/div>\n<p>This command asks the user for the offsetting accounts of imported transactions, or more specifically, transaction whose second posting's account begins with <code>TODO<\/code> and does not match any regex in <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>ignoredAccountsOnMatch<\/code><\/a>.<\/p>\n<p>Have a look at the example output <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/example_output\/match.md\">here<\/a>.<\/p>\n<p>The significantly speed up this process, it learns the account mapping from existing transactions in the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>ledgers.imported<\/code><\/a> file using the original source of the imported transaction.<\/p>\n<p>Please note that you will have to enter account information in <strong>reverse order<\/strong>: for example <code>Expenses:Food<\/code> has to be entered as <code>Food:Expenses<\/code>.\nSee <a href=\"#input-and-tab-completion\">this<\/a> information about the account input field.<\/p>\n<h3><a id=\"user-content-best-practices\" class=\"anchor\" aria-hidden=\"true\" href=\"#best-practices\"><\/a>Best practices<\/h3>\n<p>The Bayesian classifier can only work if similar transactions are always matched with the same account.<\/p>\n<p>Consider frequent credit card payments to Starbucks:<\/p>\n<ul><li>\n<p>Match them with <code>Expenses:Food:Starbucks<\/code> if you know that these should always be booked to that account.<\/p>\n<\/li>\n<li>\n<p>Match them with <code>Accounts receivable:Starbucks<\/code> and <a href=\"#enter-transactions\">manually enter<\/a> your paper receipts if<\/p>\n<ul><li>\n<p>you want to make sure they charge you the correct amounts.<\/p>\n<\/li>\n<li>\n<p>you sometimes order for friends and get reimbursed later.<\/p>\n<\/li>\n<li>\n<p>...<\/p>\n<\/li>\n<\/ul><\/li>\n<\/ul><p>See also <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/issues\/16\">this issues<\/a> on how you could use this.<\/p>\n<h2><a id=\"user-content-enter-transactions\" class=\"anchor\" aria-hidden=\"true\" href=\"#enter-transactions\"><\/a>Enter transactions<\/h2>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung add<\/pre><\/div>\n<p>This command opens a transaction editor. <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/example_output\/add.md\">Here<\/a> is an example of the output of this command.<\/p>\n<p>The amount of manual typing is kept to a minimum by two clever suggestion mechanisms and TAB completion.<\/p>\n<h3><a id=\"user-content-input-and-tab-completion\" class=\"anchor\" aria-hidden=\"true\" href=\"#input-and-tab-completion\"><\/a>Input and TAB completion<\/h3>\n<p>All input fields save their history in the current directory. It can be browsed using up and down arrow keys.<\/p>\n<p>The account input fields support TAB completion. To make this even more useful, the account hierarchy is read in reverse order. For example <code>Expenses:Food<\/code> has to be entered as <code>Food:Expenses<\/code>. This behaviour can be deactivated by setting <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>reverseAccountInput<\/code><\/a> to <code>False<\/code>.<\/p>\n<h3><a id=\"user-content-suggested-transactions\" class=\"anchor\" aria-hidden=\"true\" href=\"#suggested-transactions\"><\/a>Suggested transactions<\/h3>\n<p>After the amount is entered, the user can select a transaction whose title, date, amount and second posting's account will used to pre fill an offsetting transaction. Suggestions will consist of all transactions<\/p>\n<ul><li>whose second posting has not been cleared (i.e. marked with an asterisk in front of the account name, and<\/li>\n<li>whose first posting's amount has the absolute value as the entered amount<\/li>\n<li>whose first posting's account is contained in the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>bankAccounts<\/code><\/a> and does not match any of regexes in <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>ignoredAccountsOnAdd<\/code><\/a>.<\/li>\n<\/ul><h3><a id=\"user-content-suggested-accounts\" class=\"anchor\" aria-hidden=\"true\" href=\"#suggested-accounts\"><\/a>Suggested accounts<\/h3>\n<p>Once the first posting's account has been entered, the editor suggests accounts for the second posting based on the frequency of the resulting transaction's accounts in the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>ledgers.addedByThisUser<\/code><\/a> file.<\/p>\n<h3><a id=\"user-content-assertions--assignments\" class=\"anchor\" aria-hidden=\"true\" href=\"#assertions--assignments\"><\/a>Assertions &amp; assignments<\/h3>\n<p>Amounts can be entered with <a href=\"http:\/\/hledger.org\/manual.html#balance-assertions\" rel=\"nofollow\">assertions<\/a> or can be <a href=\"http:\/\/hledger.org\/manual.html#balance-assignments\" rel=\"nofollow\">assigned<\/a>.<\/p>\n<h3><a id=\"user-content-default-currency\" class=\"anchor\" aria-hidden=\"true\" href=\"#default-currency\"><\/a>Default currency<\/h3>\n<p>In order to be able to enter naked amounts and have the currency added automatically, add a <a href=\"http:\/\/hledger.org\/manual.html#default-commodity\" rel=\"nofollow\">default currency<\/a> to the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>addedByThisUser<\/code><\/a> ledger file. Example:<\/p>\n<pre>D 1,000.000 EUR\n<\/pre>\n<h3><a id=\"user-content-multi-user-add\" class=\"anchor\" aria-hidden=\"true\" href=\"#multi-user-add\"><\/a>Multi-user add<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung add -w alice<\/pre><\/div>\n<p>If there is more than one user configured \u2014 possibly each with their own ledger, they can be included\/activated via the command-line argument <code>-w<\/code>. This enables you to enter a transaction where postings belong to different users. When done, a transaction for each user will be generated containing their respective postings and a balancing posting to an account prefixed with the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>accountPrefixOthers<\/code><\/a>.<\/p>\n<p>Example taken from the <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/example_output\/add_multi_user.md\">output<\/a> of the above command:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n2016\/12\/19 Dinner\n        Account          |  Amount  | Assertion\n-------------------------+----------+-----------+----------\n0,jo    Wallet:Assets    | $ -100.0 |           |\n1,jo    Food:Expenses    | $ 50     |           |\n2,alice Food:Expenses    | $ 50     |           |\n-------------------------+----------+-----------+----------\nOpen Balance             | 0        |<\/pre><\/div>\n<p>generates the following transactions<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>#######  jo:  Balanced Transaction   #######\n2016\/12\/19 Dinner    ; Entered on \"2016-12-19T19:01:00Z\" by 'buchhaltung' user jo\n    Wallet:Assets                             $ -100.0\n    Food:Expenses                                 $ 50\n    Accounts receivable:Friends:alice:jo          $ 50\n#######  alice:  Balanced Transaction   #######\n2016\/12\/19 Dinner    ; Entered on \"2016-12-19T19:01:00Z\" by 'buchhaltung' user jo\n    Accounts receivable:Friends:jo:jo         $ -50\n    Food:Expenses                              $ 50<\/pre><\/div>\n<h2><a id=\"user-content-getting-results\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-results\"><\/a>Getting results<\/h2>\n<h3><a id=\"user-content-get-current-aqbanking-account-balances\" class=\"anchor\" aria-hidden=\"true\" href=\"#get-current-aqbanking-account-balances\"><\/a>Get current AqBanking account balances<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung lb<\/pre><\/div>\n<h3><a id=\"user-content-call-ledger-or-hledger\" class=\"anchor\" aria-hidden=\"true\" href=\"#call-ledger-or-hledger\"><\/a>Call <code>ledger<\/code> or <code>hledger<\/code><\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung ledger\nbuchhaltung hledger<\/pre><\/div>\n<p>This calls the respective program with the <code>LEDGER<\/code> environment variable set to the configured <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>mainLedger<\/code><\/a> or <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>mainHledger<\/code><\/a>.<\/p>\n<h3><a id=\"user-content-commit-the-changes\" class=\"anchor\" aria-hidden=\"true\" href=\"#commit-the-changes\"><\/a>Commit the changes<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung commit -a -m'checking account ok'<\/pre><\/div>\n<p>or<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>buchhaltung commitHledger -a -m'checking account ok'<\/pre><\/div>\n<p>this commits all changes to the git repository that contains the <a href=\"https:\/\/github.com\/johannesgerer\/buchhaltung\/blob\/master\/config.yml\"><code>mainLedger<\/code><\/a> file. The commit message will also contain the output of <code>buchhaltung lb<\/code> and <code>buchhaltung ledger balance --end tomorrow<\/code>.<\/p>\n<p>Try to get Barclays UK direct feed working?\n<a href=\"https:\/\/stackoverflow.com\/questions\/3469628\/banking-api-protocol\" rel=\"nofollow\">https:\/\/stackoverflow.com\/questions\/3469628\/banking-api-protocol<\/a><\/p>\n<\/article>","created_at":"2019-06-07T13:13:28+0000","updated_at":"2019-06-07T13:13:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/253448?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":967,"title":"Tangentially related: there are quite a lot of good quality free software fonts ...","url":"https:\/\/news.ycombinator.com\/item?id=20097427","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D20097427\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"20097427\"><td class=\"ind\"><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Freak_NL\" class=\"hnuser\">Freak_NL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=20097427\">17 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/item?id=20096120\">parent<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=20097427&amp;auth=1fbebf659ccfe432f0ffb177a9266b85557cffce\">favorite<\/a>           | on: <a href=\"https:\/\/news.ycombinator.com\/item?id=20096120\">National Park Typeface<\/a><\/p><br \/><\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-06-05T11:02:55+0000","updated_at":"2019-06-05T11:02:55+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":966,"title":"National Park Typeface","url":"https:\/\/nationalparktypeface.com\/","content":"<h2>I saw those familiar words. Set \u201cNational Park Service, United States Department of the Interior\u201d \u2014 style. I wondered if it actually was a typeface or \u201cfont\u201d that anyone could download and use?\u00a0<i>Do park rangers have this as a \ntypeface on their computers to set in their word docs, pdfs and power point slides?<\/i><\/h2><p><\/p><h2>I had a sketchbook with me and took some rubbings of the letterforms and asked my friend Miles Barger, the Visual Information Specialist for Rocky, if he had the typeface. He asked the sign shop. No one has it? Turns out it isn\u2019t a typeface at all but a system of paths, points and curves that a router follows.<\/h2>\n<br \/><h2> <\/h2><div class=\"image-gallery\" data-gallery=\"%7B%22mode_id%22%3A3%2C%22gallery_instance_id%22%3A25%2C%22name%22%3A%22Justify%22%2C%22path%22%3A%22justify%22%2C%22data%22%3A%7B%22image_padding%22%3A2%2C%22row_height%22%3A20%2C%22variation_index%22%3A0%2C%22variation%22%3A0%2C%22variation_mode%22%3A0%2C%22variation_seed%22%3A0%2C%22responsive%22%3Atrue%2C%22meta_data%22%3A%7B%7D%2C%22mobile_data%22%3A%7B%22image_padding%22%3A1%2C%22row_height%22%3A60%2C%22separate_mobile_view%22%3Afalse%7D%7D%7D\"><img width=\"2000\" height=\"917\" data-mid=\"22520305\" border=\"0\" src=\"https:\/\/freight.cargo.site\/t\/original\/i\/0ceb661469dbd1b1b4e9bda2e0dc6c4a801008b0b8cc931b9d64ed5b01833851\/3.png\" alt=\"image\" \/><img width=\"1001\" height=\"1001\" data-mid=\"22517195\" border=\"0\" src=\"https:\/\/freight.cargo.site\/t\/original\/i\/ab1793016ab5e8afa1ea008282a694ef401579d63c366a89dd93e0b8cfae1c35\/All_Outline-01.png\" alt=\"image\" \/><\/div><br \/><h2>The router\u2019s \"bit\" follows the path<br \/>and gives the letters its stroke<br \/>weight or thickness only when engraving a sign.It doesn't really exist as a typeface unless a sign is made.<\/h2>","created_at":"2019-06-05T01:09:09+0000","updated_at":"2019-06-05T01:09:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"nationalparktypeface.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":965,"title":"How to Configure NGINX","url":"https:\/\/www.linode.com\/docs\/web-servers\/nginx\/how-to-configure-nginx\/","content":"<p><img src=\"https:\/\/www.linode.com\/docs\/web-servers\/nginx\/how-to-configure-nginx\/how_to_configure_nginx.png\" alt=\"Introduction to NGINX\" title=\"Introduction to NGINX\" \/><\/p><p><a href=\"https:\/\/www.nginx.com\/\">NGINX<\/a> is a lightweight, high-performance web server designed for high-traffic use cases.<\/p><p>One of NGINX\u2019s strongest features is the ability to efficiently serve static content such as HTML and media files. NGINX uses an asynchronous event-driven model which provides predictable performance under load.<\/p><p>NGINX hands off dynamic content to CGI, FastCGI, or other web servers such as Apache. This content is then passed back to NGINX for delivery to the client. This document will familiarize you with basic NGINX parameters and conventions.<\/p><h2 id=\"directives-blocks-and-contexts\">Directives, Blocks, and Contexts<\/h2><p>All NGINX configuration files are located in the <code>\/etc\/nginx\/<\/code> directory. The primary configuration file is <code>\/etc\/nginx\/nginx.conf<\/code>.<\/p><p>Configuration options in NGINX are called <a href=\"http:\/\/nginx.org\/en\/docs\/dirindex.html\">directives<\/a>. Directives are organized into groups known as <strong>blocks<\/strong> or <strong>contexts<\/strong>. The two terms are synonymous.<\/p><p>Lines preceded by a <code>#<\/code> character are comments and not interpreted by NGINX. Lines containing directives must end with a <code>;<\/code> or NGINX will fail to load the configuration and report an error.<\/p><p>Below is a condensed copy of the <code>\/etc\/nginx\/nginx.conf<\/code> file that is included with installations from the NGINX repositories. The file starts with 4 directives: <code>user<\/code>, <code>worker_processes<\/code>, <code>error_log<\/code>, and <code>pid<\/code>. These are outside any specific block or context, so they\u2019re said to exist in the <code>main<\/code> context. The <code>events<\/code> and <code>http<\/code> blocks are areas for additional directives, and they also exist in the <code>main<\/code> context.<\/p><p>See <a href=\"https:\/\/nginx.org\/en\/docs\/ngx_core_module.html\">the NGINX docs<\/a> for explanations of these directives and others available in the <code>main<\/code> context.<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/nginx.conf \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-bash\" data-lang=\"bash\"> 1\n 2\n 3\n 4\n 5\n 6\n 7\n 8\n 9\n10\n11\n12\n13\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-bash\" data-lang=\"bash\">user  nginx;\nworker_processes  1;\nerror_log  \/var\/log\/nginx\/error.log warn;\npid        \/var\/run\/nginx.pid;\nevents {\n       . . .\n}\nhttp {\n       . . .\n}<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><h2 id=\"the-http-block\">The http Block<\/h2><p>The <code>http<\/code> block contains directives for handling web traffic. These directives are often referred to as <em>universal<\/em> because they are passed on to all website configurations NGINX serves. See <a href=\"https:\/\/nginx.org\/en\/docs\/http\/ngx_http_core_module.html\">the NGINX docs<\/a> for a list of available directives for the <code>http<\/code> block.<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/nginx.conf \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\"> 1\n 2\n 3\n 4\n 5\n 6\n 7\n 8\n 9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">http {\n    include       \/etc\/nginx\/mime.types;\n    default_type  application\/octet-stream;\n    log_format  main  '$remote_addr - $remote_user [$time_local] \"$request\" '\n                      '$status $body_bytes_sent \"$http_referer\" '\n                      '\"$http_user_agent\" \"$http_x_forwarded_for\"';\n    access_log  \/var\/log\/nginx\/access.log  main;\n    sendfile        on;\n    #tcp_nopush     on;\n    keepalive_timeout  65;\n    #gzip  on;\n    include \/etc\/nginx\/conf.d\/*.conf;\n}<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><h2 id=\"server-blocks\">Server Blocks<\/h2><p>The <code>http<\/code> block above contains an <code>include<\/code> directive which tells NGINX where website configuration files are located.<\/p><ul><li><p>If you installed from the official NGINX repository, this line will say <code>include \/etc\/nginx\/conf.d\/*.conf;<\/code> as it does in the <code>http<\/code> block above. Each website you host with NGINX should have its own configuration file in <code>\/etc\/nginx\/conf.d\/<\/code>, with the name formatted as <code>example.com.conf<\/code>. Sites which are disabled (not being served by NGINX) should be named <code>example.com.conf.disabled<\/code>.<\/p><\/li>\n<li><p>If you installed NGINX from the Debian or Ubuntu repositories, this line will say <code>include \/etc\/nginx\/sites-enabled\/*;<\/code>. The <code>..\/sites-enabled\/<\/code> folder contains symlinks to the site configuration files stored in <code>\/etc\/nginx\/sites-available\/<\/code>. Sites in <code>sites-available<\/code> can be disabled by removing the symlink to <code>sites-enabled<\/code>.<\/p><\/li>\n<li><p>Depending on your installation source, you\u2019ll find an example configuration file at <code>\/etc\/nginx\/conf.d\/default.conf<\/code> or <code>etc\/nginx\/sites-enabled\/default<\/code>.<\/p><\/li>\n<\/ul><p>Regardless of the installation source, server configuration files will contain a <code>server<\/code> block (or blocks) for a website. For example:<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/conf.d\/example.com.conf \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n3\n4\n5\n6\n7\n8\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">server {\n    listen         80 default_server;\n    listen         [::]:80 default_server;\n    server_name    example.com www.example.com;\n    root           \/var\/www\/example.com;\n    index          index.html;\n    try_files $uri \/index.html;\n}<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><h3 id=\"listening-ports\">Listening Ports<\/h3><p>The <code>listen<\/code> directive tells NGINX the hostname\/IP and the TCP port where it should listen for HTTP connections. The argument <code>default_server<\/code> means this virtual host will answer requests on port 80 that don\u2019t specifically match another virtual host\u2019s listen statement. The second statement listens over IPv6 and behaves similarly.<\/p><h3 id=\"name-based-virtual-hosting\">Name-Based Virtual Hosting<\/h3><p>The <code>server_name<\/code> directive allows multiple domains to be served from a single IP address. The server decides which domain to serve based on the request header it receives.<\/p><p>You typically should create one file per domain or site you want to host on your server. Here are some examples:<\/p><ol><li><p>Process requests for both <code>example.com<\/code> and <code>www.example.com<\/code>:<\/p>\n<dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/conf.d\/example.com.conf \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">server_name   example.com www.example.com;<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><\/li>\n<li><p>The <code>server_name<\/code> directive can also use wildcards. <code>*.example.com<\/code> and <code>.example.com<\/code> both instruct the server to process requests for all subdomains of <code>example.com<\/code>:<\/p>\n<dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/conf.d\/example.com.conf \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">server_name   *.example.com;\nserver_name   .example.com;<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><\/li>\n<li><p>Process requests for all domain names beginning with <code>example.<\/code>:<\/p>\n<dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/conf.d\/example.com.conf \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">server_name   example.*;<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><\/li>\n<\/ol><p>NGINX allows you to specify server names that are not valid domain names. NGINX uses the name from the HTTP header to answer requests, regardless of whether the domain name is valid or not.<\/p><p>Using non-domain hostnames is useful if your server is on a LAN, or if you already know all of the clients that will be making requests of the server. This includes front-end proxy servers with <code>\/etc\/hosts<\/code> entries configured for the IP address on which NGINX is listening.<\/p><h2 id=\"location-blocks\">Location Blocks<\/h2><p>The <code>location<\/code> setting lets you configure how NGINX will respond to requests for resources within the server. Just like the <code>server_name<\/code> directive tells NGINX how to process requests for the domain, <code>location<\/code> directives cover requests for specific files and folders, such as <code>http:\/\/example.com\/blog\/<\/code>. Here are some examples:<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n3\n4\n5\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location \/ { }\nlocation \/images\/ { }\nlocation \/blog\/ { }\nlocation \/planet\/ { }\nlocation \/planet\/blog\/ { }<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>The locations above are <em>literal string<\/em> matches, which match any part of an HTTP request that comes after the host segment:<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/<\/code><\/p><p><strong>Returns:<\/strong> Assuming that there is a <code>server_name<\/code> entry for <code>example.com<\/code>, the <code>location \/<\/code> directive will determine what happens with this request.<\/p><p>NGINX always fulfills requests using the most specific match:<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/planet\/blog\/<\/code> or <code>http:\/\/example.com\/planet\/blog\/about\/<\/code><\/p><p><strong>Returns:<\/strong> This is fulfilled by the <code>location \/planet\/blog\/<\/code> directive because it is more specific, even though <code>location \/planet\/<\/code> also matches this request.<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location ~ IndexPage\\.php$ { }\nlocation ~ ^\/BlogPlanet(\/|\/index\\.php)$ { }<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>When a <code>location<\/code> directive is followed by a tilde (<strong>~<\/strong>), NGINX performs a <em>regular expression<\/em> match. These matches are always case-sensitive. So, <code>IndexPage.php<\/code> would match the first example above, but <code>indexpage.php<\/code> would not. In the second example, the regular expression <code>^\/BlogPlanet(\/|index\\.php)$<\/code> will match requests for <code>\/BlogPlanet\/<\/code> and <code>\/BlogPlanet\/index.php<\/code>, but <strong>not<\/strong> <code>\/BlogPlanet<\/code>, <code>\/blogplanet\/<\/code>, or <code>\/blogplanet\/index.php<\/code>. NGINX uses <a href=\"http:\/\/perldoc.perl.org\/perlre.html\">Perl Compatible Regular Expressions<\/a> (PCRE).<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location ~* \\.(pl|cgi|perl|prl)$ { }\nlocation ~* \\.(md|mdwn|txt|mkdn)$ { }<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>If you want matches to be case-<em>insensitive<\/em>, use a tilde with an asterisk (<strong>~<\/strong>*). The examples above all specify how nginx should process requests that end in a particular file extension. In the first example, any file ending in: <code>.pl<\/code>, <code>.PL<\/code>, <code>.cgi<\/code>, <code>.CGI<\/code>, <code>.perl<\/code>, <code>.Perl<\/code>, <code>.prl<\/code>, and <code>.PrL<\/code> (among others) will match the request.<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location ^~ \/images\/IndexPage\/ { }\nlocation ^~ \/blog\/BlogPlanet\/ { }<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>Adding a caret and tilde (<strong>^~<\/strong>) to your <code>location<\/code> directives tells NGINX, if it matches a particular string, to stop searching for more specific matches and use the directives here instead. Other than that, these directives work like the literal string matches in the first group. Even if there\u2019s a more specific match later, if a request matches one of these directives, the settings here will be used. See below for more information about the order and priority of <code>location<\/code> directive processing.<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location = \/ { }<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>Finally, if you add an equals sign (<strong>=<\/strong>) to the <code>location<\/code> setting, this forces an exact match with the path requested and then stops searching for more specific matches. For instance, the final example will match only <code>http:\/\/example.com\/<\/code>, not <code>http:\/\/example.com\/index.html<\/code>. Using exact matches can speed up request times slightly, which can be useful if you have some requests that are particularly popular.<\/p><p>Directives are processed in the following order:<\/p><ol><li>Exact string matches are processed first. If a match is found, NGINX stops searching and fulfills the request.<\/li>\n<li>Remaining literal string directives are processed next. If NGINX encounters a match where the <strong>^~<\/strong> argument is used, it stops here and fulfills the request. Otherwise, NGINX continues to process location directives.<\/li>\n<li>All location directives with regular expressions (<strong>~<\/strong> and <strong>~<\/strong>*) are processed. If a regular expression matches the request, nginx stops searching and fulfills the request.<\/li>\n<li>If no regular expressions match, the most specific literal string match is used.<\/li>\n<\/ol><p>Make sure each file and folder under a domain will match at least one <code>location<\/code> directive.<\/p><blockquote class=\"note\">\n  <div><strong class=\"callout-title\">Note<\/strong>\n  <p>Nested location blocks are not recommended or supported.<\/p>\n<\/div><\/blockquote><h3 id=\"location-root-and-index\">Location Root and Index<\/h3><p>The <code>location<\/code> setting is another variable that has its own block of arguments.<\/p><p>Once NGINX has determined which <code>location<\/code> directive best matches a given request, the response to this request is determined by the contents of the associated <code>location<\/code> directive block. Here\u2019s an example:<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">1\n2\n3\n4\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location \/ {\n    root html;\n    index index.html index.htm;\n}<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>In this example, the document root is located in the <code>html\/<\/code> directory. Under the default installation prefix for NGINX, the full path to this location is <code>\/etc\/nginx\/html\/<\/code>.<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/blog\/includes\/style.css<\/code><\/p><p><strong>Returns:<\/strong> NGINX will attempt to serve the file located at <code>\/etc\/nginx\/html\/blog\/includes\/style.css<\/code><\/p><blockquote class=\"note\">\n  <div><strong class=\"callout-title\">Note<\/strong>\n  <div>You can use absolute paths for the<code>root<\/code>directive if desired.<\/div>\n<\/div><\/blockquote><p>The <code>index<\/code> variable tells NGINX which file to serve if none is specified. For example:<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com<\/code><\/p><p><strong>Returns:<\/strong> NGINX will attempt to serve the file located at <code>\/etc\/nginx\/html\/index.html<\/code>.<\/p><p>If multiple files are specified for the <code>index<\/code> directive, NGINX will process the list in order and fulfill the request with the first file that exists. If <code>index.html<\/code> doesn\u2019t exist in the relevant directory, then <code>index.htm<\/code> will be used. If neither exists, a 404 message will be sent.<\/p><p>Here\u2019s a more complex example, showcasing a set of <code>location<\/code> directives for a server responding to the domain <code>example.com<\/code>:<\/p><dl class=\"file\"><dt>\n\t\t\/etc\/nginx\/sites-available\/example.com location directive \n<\/dt>\n<dd>\n<div class=\"highlight\"><div class=\"chroma\"><table class=\"lntable\"><tr><td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\"> 1\n 2\n 3\n 4\n 5\n 6\n 7\n 8\n 9\n10\n11\n12\n<\/pre><\/td>\n<td class=\"lntd\">\n<pre class=\"language-nginx\" data-lang=\"nginx\">location \/ {\n    root   \/srv\/www\/example.com\/public_html;\n    index  index.html index.htm;\n}\nlocation ~ \\.pl$ {\n    gzip off;\n    include \/etc\/nginx\/fastcgi_params;\n    fastcgi_pass unix:\/var\/run\/fcgiwrap.socket;\n    fastcgi_index index.pl;\n    fastcgi_param SCRIPT_FILENAME \/srv\/www\/example.com\/public_html$fastcgi_script_name;\n}<\/pre><\/td><\/tr><\/table><\/div><\/div>\n<\/dd>\n<\/dl><p>In this example, all requests for resources that end in a <code>.pl<\/code> extension are handled by the second location block, which specifies a <code>fastcgi<\/code> handler for these requests. Otherwise, NGINX uses the first location directive. Resources are located on the file system at <code>\/srv\/www\/example.com\/public_html\/<\/code>. If no file name is specified in the request, NGINX will look for and provide the <code>index.html<\/code> or <code>index.htm<\/code> file. If no <code>index<\/code> files are found, the server will return a 404 error.<\/p><p>Let\u2019s analyze what happens during a few requests:<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/<\/code><\/p><p><strong>Returns:<\/strong> <code>\/srv\/www\/example.com\/public_html\/index.html<\/code> if it exists. If that file doesn\u2019t exist, it will serve <code>\/srv\/www\/example.com\/public_html\/index.htm<\/code>. If neither exists, NGINX returns a 404 error.<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/blog\/<\/code><\/p><p><strong>Returns:<\/strong> <code>\/srv\/www\/example.com\/public_html\/blog\/index.html<\/code> if it exists. If that file doesn\u2019t exist, it will serve <code>\/srv\/www\/example.com\/public_html\/blog\/index.htm<\/code>. If neither exists, NGINX returns a 404 error.<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/tasks.pl<\/code><\/p><p><strong>Returns:<\/strong> NGINX will use the FastCGI handler to execute the file located at <code>\/srv\/www\/example.com\/public_html\/tasks.pl<\/code> and return the result.<\/p><p><strong>Request:<\/strong> <code>http:\/\/example.com\/username\/roster.pl<\/code><\/p><p><strong>Returns:<\/strong> NGINX will use the FastCGI handler to execute the file located at <code>\/srv\/www\/example.com\/public_html\/username\/roster.pl<\/code> and return the result.<\/p><p><a href=\"https:\/\/www.linode.com\/community\/questions\/\" target=\"_blank\">Find answers, ask questions, and help others.<\/a><\/p><p class=\"doc-license\">This guide is published under a <a href=\"https:\/\/creativecommons.org\/licenses\/by-nd\/4.0\">CC BY-ND 4.0<\/a> license.<\/p>","created_at":"2019-06-02T14:35:09+0000","updated_at":"2019-06-02T14:35:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"www.linode.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":964,"title":"Next","url":"https:\/\/next.atlas.engineer\/","content":"<header><main>\n\n<div class=\"pure-g feature\"><div class=\"pure-u-1 pure-u-md-8-24\"><p class=\"feature-description\">Jump exactly where you want to. Move through long documents quickly to find what you need. No more wasted searching.\n    <\/p><\/div><p><\/p><\/div>\n<div class=\"pure-g feature\"><div class=\"pure-u-1 pure-u-md-8-24\"><p class=\"feature-description\">Don't lose track of where you've been. With next every page you've visited is always accessible.\n    <\/p><\/div><p><\/p><\/div>\n<div class=\"pure-g feature\"><div class=\"pure-u-1 pure-u-md-8-24\"><p class=\"feature-description\">Do everything with the keyboard. Don't let your browser get in the way of your workflow\n    <\/p><\/div><p><\/p><\/div>\n<\/main><\/header><footer><\/footer>","created_at":"2019-06-02T11:54:08+0000","updated_at":"2019-06-02T11:54:08+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"next.atlas.engineer","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":963,"title":"Wilfred\/helpful","url":"https:\/\/github.com\/Wilfred\/helpful","content":"<p>Helpful is an alternative to the built-in Emacs help that provides\nmuch more contextual information.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful.png\" alt=\"screenshot\" \/><\/a><\/p><h2>Usage<\/h2><p>Install from MELPA, then call one of the following commands:<\/p><ul><li><code>helpful-callable<\/code><\/li>\n<li><code>helpful-function<\/code><\/li>\n<li><code>helpful-macro<\/code><\/li>\n<li><code>helpful-command<\/code><\/li>\n<li><code>helpful-key<\/code><\/li>\n<li><code>helpful-variable<\/code><\/li>\n<li><code>helpful-at-point<\/code><\/li>\n<\/ul><p>If you want to replace the default Emacs help keybindings, you can do\nso:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Note that the built-in `describe-function' includes both functions\n;; and macros. `helpful-function' is functions only, so we provide\n;; `helpful-callable' as a drop-in replacement.\n(global-set-key (kbd \"C-h f\") #'helpful-callable)\n(global-set-key (kbd \"C-h v\") #'helpful-variable)\n(global-set-key (kbd \"C-h k\") #'helpful-key)<\/pre><\/div><p>I also recommend the following keybindings to get the most out of\nhelpful:<\/p><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Lookup the current symbol at point. C-c C-d is a common keybinding\n;; for this in lisp modes.\n(global-set-key (kbd \"C-c C-d\") #'helpful-at-point)\n;; Look up *F*unctions (excludes macros).\n;;\n;; By default, C-h F is bound to `Info-goto-emacs-command-node'. Helpful\n;; already links to the manual, if a function is referenced there.\n(global-set-key (kbd \"C-h F\") #'helpful-function)\n;; Look up *C*ommands.\n;;\n;; By default, C-h C is bound to describe `describe-coding-system'. I\n;; don't find this very useful, but it's frequently useful to only\n;; look at interactive functions.\n(global-set-key (kbd \"C-h C\") #'helpful-command)<\/pre><\/div><h2>Features<\/h2><h3>Source code<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_source.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_source.png\" alt=\"screenshot\" \/><\/a><\/p><p>Helpful will try really hard to show the source code. It shows the\nsource code for interactively defined functions (unlike the built-in\nHelp) and falls back to the raw sexp if no source is available.<\/p><h3>View Callers<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_refs.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_refs.png\" alt=\"screenshot\" \/><\/a><\/p><p>Helpful will show you where a function is being called!<\/p><h3>Prettier Docstrings<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_docstring.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_docstring.png\" alt=\"screenshot\" \/><\/a><\/p><p>Docstrings in helpful:<\/p><ul><li>Highlight the summary (the first sentence)<\/li>\n<li>Include cross-references to other functions\/variables<\/li>\n<li>Linkify references to Info nodes<\/li>\n<li>Hide superfluous punctuation<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_view_in_manual.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_view_in_manual.png\" alt=\"screenshot\" \/><\/a><\/p><p>If a symbol is also documented in the Info manual, helpful will\nprovide a link to the relevant section too.<\/p><h3>Symbol Properties<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_props.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_props.png\" alt=\"screenshot\" \/><\/a><\/p><p>Helpful will show you the properties that have been applied to the\ncurrent symbol. This provides visibility of features\nlike edebug or byte-code optimisation.<\/p><p>Helpful will also highlight any symbol aliases.<\/p><h3>Describe Commands<\/h3><p>Helpful provides a separate <code>helpful-command<\/code> function, for when you\njust want to view interactive functions.<\/p><h3>View Keymaps<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_bindings.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_bindings.png\" alt=\"screenshot\" \/><\/a><\/p><p>Helpful displays any keybindings that apply to interactive functions.<\/p><h3>Integrated Tooling<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_tools.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_tools.png\" alt=\"screenshot\" \/><\/a><\/p><p>You can trace, debug or disassemble functions from inside\nHelpful. This is discoverable and doesn't require memorisation of\ncommands.<\/p><h3>Aliases<\/h3><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/helpful\/blob\/master\/screenshots\/helpful_aliases.png\"><img src=\"https:\/\/github.com\/Wilfred\/helpful\/raw\/master\/screenshots\/helpful_aliases.png\" alt=\"screenshot\" \/><\/a><\/p><p>If a function has multiple aliases in Emacs, Helpful will show all of\nthe aliases defined.<\/p><h2>Inspirations<\/h2><p>This project has been heavily influenced by:<\/p><ul><li><a href=\"https:\/\/www.emacswiki.org\/emacs\/help+.el\" rel=\"nofollow\">help+.el<\/a>, help-fns+.el, help-mode+.el<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=pQQTScuApWk\" rel=\"nofollow\">Dave Williams' demo of Lucid's Energize<\/a><\/li>\n<\/ul><h2>License<\/h2><p>GPLv3+.<\/p><p>I am providing code in the repository to you under an open source\nlicense. Because this is my personal repository, the license you\nreceive to my code is from me and not my employer.<\/p>","created_at":"2019-05-30T21:02:12+0000","updated_at":"2019-05-30T21:02:18+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/70800?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":962,"title":"Deft for Emacs","url":"https:\/\/jblevins.org\/projects\/deft\/","content":"<header><p>Deft is an Emacs mode for quickly browsing, filtering, and editing\ndirectories of plain text notes, inspired by Notational Velocity.\nIt was designed for increased productivity when writing and taking\nnotes by making it fast and simple to find the right file at the\nright time and by automating many of the usual tasks such as\ncreating new files and saving files.<\/p><p><a href=\"https:\/\/jblevins.org\/projects\/deft\/deft-v0.6.gif\">Deft Screencast<\/a><\/p><h2>Obtaining Deft<\/h2><p>Deft is open source software and may be freely distributed and\nmodified under the <abbr class=\"abbr\" title=\"Berkeley System Distribution\">BSD<\/abbr> license. The latest stable release is\nversion 0.8, released on January 12, 2018.<\/p><p><strong>Installation via MELPA Stable<\/strong><\/p><p>The recommended way to install Deft is to obtain the stable version\nfrom <a href=\"https:\/\/stable.melpa.org\/#\/deft\">MELPA Stable<\/a> using\n<code>package.el<\/code>. First, configure <code>package.el<\/code> and the MELPA Stable\nrepository by adding the following to your <code>.emacs<\/code>, <code>init.el<\/code>, or\nequivalent startup file:<\/p><pre class=\"lisp\">(require 'package)\n(add-to-list 'package-archives\n             '(\"melpa-stable\" . \"https:\/\/stable.melpa.org\/packages\/\"))\n(package-initialize)\n<\/pre><p>Then, after restarting Emacs or evaluating the above statements, issue\nthe following command: <kbd>M-x package-install RET deft RET<\/kbd>.<\/p><p><strong>Direct Download<\/strong><\/p><p>Alternatively you can manually download and install Deft.\nFirst, download the latest stable version of and save the file\nwhere Emacs can find it\u2014a directory in your <code>load-path<\/code>:<\/p><p>Then, add the following line to your startup file:<\/p><pre class=\"lisp\">(require 'deft)\n<\/pre><p><strong>Development Version<\/strong><\/p><p>To follow or contribute to Deft development, you can browse or\nclone the Git repository <a href=\"https:\/\/github.com\/jrblevin\/deft\">on GitHub<\/a>:<\/p><pre>git clone https:\/\/github.com\/jrblevin\/deft.git\n<\/pre><p>If you prefer to install and use the development version, which may\nbecome unstable at some times, you can either clone the Git\nrepository as above or install Deft from\n<a href=\"https:\/\/melpa.org\/#\/deft\">MELPA<\/a>.<\/p><p>If you clone the repository directly, then make sure that Emacs can\nfind it by adding the following line to your startup file:<\/p><pre class=\"lisp\">(add-to-list 'load-path \"\/path\/to\/deft\/repository\")\n<\/pre><h2>Overview<\/h2><p>The Deft buffer is simply a file browser which lists the titles of\nall text files in the Deft directory followed by short summaries\nand last modified times. The title is taken to be the first line\nof the file and the summary is extracted from the text that\nfollows. Files are, by default, sorted in terms of the last\nmodified date, from newest to oldest.<\/p><p>All Deft files or notes are simple plain text files where the first\nline contains a title. As an example, the following directory\nstructure generated the screenshot above.<\/p><pre>% ls ~\/.deft\nabout.txt    browser.txt     directory.txt   operations.txt\nack.txt      completion.txt  extensions.org\nbinding.txt  creation.txt    filtering.txt\n% cat ~\/.deft\/about.txt\n# About\nAn Emacs mode for slicing and dicing plain text files.\n<\/pre><p>Deft\u2019s primary operation is searching and filtering. The list of\nfiles can be limited or filtered using a search string, which will\nmatch both the title and the body text. To initiate a filter,\nsimply start typing. Filtering happens on the fly. As you type,\nthe file browser is updated to include only files that match the\ncurrent string.<\/p><p>To open the first matching file, simply press <kbd>RET<\/kbd>. If no files\nmatch your search string, pressing <kbd>RET<\/kbd> will create a new file\nusing the string as the title. This is a very fast way to start\nwriting new notes. The filename will be generated automatically.\nIf you prefer to provide a specific filename, use <kbd>C-RET<\/kbd> instead.<\/p><p>To open files other than the first match, navigate up and down\nusing <kbd>C-p<\/kbd> and <kbd>C-n<\/kbd> and press <kbd>RET<\/kbd> on the file you want to open.\nWhen opening a file, Deft searches forward and leaves the point\nat the end of the first match of the filter string.<\/p><p>You can also press <kbd>C-o<\/kbd> to open a file in another window, without\nswitching to the other window. Issue the same command with a prefix\nargument, <kbd>C-u C-o<\/kbd>, to open the file in another window and switch\nto that window.<\/p><p>To edit the filter string, press <code>DEL<\/code> (backspace) to remove the\nlast character or <kbd>M-DEL<\/kbd> to remove the last \u201cword\u201d. To yank\n(paste) the most recently killed (cut or copied) text into the\nfilter string, press <kbd>C-y<\/kbd>. Press <kbd>C-c C-c<\/kbd> to clear the filter\nstring and display all files and <kbd>C-c C-g<\/kbd> to refresh the file\nbrowser using the current filter string.<\/p><p>For more advanced editing operations, you can also edit the filter\nstring in the minibuffer by pressing <kbd>C-c C-l<\/kbd>. While in the\nminibuffer, the history of previous edits can be cycled through by\npressing <kbd>M-p<\/kbd> and <kbd>M-n<\/kbd>. This form of static, one-time filtering\n(as opposed to incremental, on-the-fly filtering) may be preferable\nin some situations, such as over slow connections or on systems\nwhere interactive filtering performance is poor.<\/p><p>By default, Deft filters files in incremental string search mode,\nwhere \u201csearch string\u201d will match all files containing both \u201csearch\u201d\nand \u201cstring\u201d in any order. Alternatively, Deft supports direct\nregexp filtering, where the filter string is interpreted as a\nformal regular expression. For example, <code>^\\(foo\\|bar\\)<\/code> matches\nfoo or bar at the beginning of a line. Pressing <kbd>C-c C-t<\/kbd> will\ntoggle between incremental and regexp search modes. Regexp\nsearch mode is indicated by an \u201cR\u201d in the mode line.<\/p><p>Common file operations can also be carried out from within Deft.\nFiles can be renamed using <kbd>C-c C-r<\/kbd> or deleted using <kbd>C-c C-d<\/kbd>.\nNew files can also be created using <kbd>C-c C-n<\/kbd> for quick creation or\n<kbd>C-c C-m<\/kbd> for a filename prompt. You can leave Deft at any time\nwith <kbd>C-c C-q<\/kbd>.<\/p><p>Unused files can be archived by pressing <kbd>C-c C-a<\/kbd>. Files will be\nmoved to <code>deft-archive-directory<\/code>, which is a directory named\n<code>archive<\/code> within your <code>deft-directory<\/code> by default.<\/p><p>Files opened with deft are automatically saved after Emacs has been\nidle for a customizable number of seconds. This value is a floating\npoint number given by <code>deft-auto-save-interval<\/code> (default: 1.0).<\/p><h2>Getting Started<\/h2><p>Once you have installed Deft following one of the above methods,\nyou can simply run <kbd>M-x deft<\/kbd> to start Deft. It is useful\nto create a global keybinding for the <code>deft<\/code> function (e.g., a\nfunction key) to start it quickly (see below for details).<\/p><p>When you first run Deft, it will complain that it cannot find the\n<code>~\/.deft<\/code> directory. You can either create a symbolic link to\nanother directory where you keep your notes or run <kbd>M-x deft-setup<\/kbd>\nto create the <code>~\/.deft<\/code> directory automatically.<\/p><p>One useful way to use Deft is to keep a directory of notes in a\nDropbox folder. This can be used with other applications and\nmobile devices, for example, <a href=\"http:\/\/brettterpstra.com\/projects\/nvalt\/\">nvALT<\/a>, <a href=\"http:\/\/notational.net\/\">Notational Velocity<\/a>, or\n<a href=\"http:\/\/simplenote.com\/\">Simplenote<\/a> on <abbr class=\"abbr\" title=\"Operating System\">OS<\/abbr> X or <a href=\"https:\/\/geo.itunes.apple.com\/us\/app\/editorial\/id673907758?mt=8&amp;uo=6&amp;at=11l5Vs&amp;ct=deft\">Editorial<\/a>, <a href=\"https:\/\/geo.itunes.apple.com\/us\/app\/byword\/id482063361?mt=8&amp;uo=6&amp;at=11l5Vs&amp;ct=deft\">Byword<\/a>, or <a href=\"https:\/\/geo.itunes.apple.com\/us\/app\/1writer-note-taking-writing\/id680469088?mt=8&amp;uo=6&amp;at=11l5Vs&amp;ct=deft\">1Writer<\/a>\non iOS.<\/p><h2>Basic Customization<\/h2><p>You can customize items in the <code>deft<\/code> group to change the default\nfunctionality.<\/p><p>By default, Deft looks for notes by searching for files with the\nextensions <code>.txt<\/code>, <code>.text<\/code>, <code>.md<\/code>, <code>.markdown<\/code>, or <code>.org<\/code> in the\n<code>~\/.deft<\/code> directory. You can customize both the file extension and\nthe Deft directory by running <kbd>M-x customize-group<\/kbd> and typing\n<code>deft<\/code>. Alternatively, you can configure them in your <code>.emacs<\/code>\nfile:<\/p><pre class=\"lisp\">(setq deft-extensions '(\"txt\" \"tex\" \"org\"))\n(setq deft-directory \"~\/Dropbox\/notes\")\n<\/pre><p>The first element of <code>deft-extensions<\/code> (or in Lisp parlance, the\ncar) is the default extension used to create new files.<\/p><p>By default, Deft only searches for files in <code>deft-directory<\/code> but\nnot in any subdirectories. All files in <code>deft-directory<\/code> with one\nof the specified extensions will be included except for those\nmatching <code>deft-ignore-file-regexp<\/code>. Set <code>deft-recursive<\/code> to a\nnon-nil value to enable searching for files in subdirectories\n(those not matching <code>deft-recursive-ignore-dir-regexp<\/code>):<\/p><pre class=\"lisp\">(setq deft-recursive t)\n<\/pre><p>You can easily set up a global keyboard binding for Deft. For\nexample, to bind it to F8, add the following code to your <code>.emacs<\/code>\nfile:<\/p><pre class=\"lisp\">(global-set-key [f8] 'deft)\n<\/pre><p>If you manage loading packages with <a href=\"https:\/\/github.com\/jwiegley\/use-package\">use-package<\/a>, then you can\nconfigure by adding a declaration such as this one to your init\nfile:<\/p><pre class=\"lisp\">(use-package deft\n  :bind (\"&lt;f8&gt;\" . deft)\n  :commands (deft)\n  :config (setq deft-directory \"~\/Dropbox\/notes\"\n                deft-extensions '(\"md\" \"org\")))\n<\/pre><h2>Reading Files<\/h2><p>The displayed title of each file is taken to be the first line of\nthe file, with certain characters removed from the beginning. Hash\ncharacters, as used in Markdown headers, and asterisks, as in Org\nMode headers, are removed. Additionally, Org mode <code>#+TITLE:<\/code> tags,\nMultiMarkdown <code>Title:<\/code> tags, LaTeX comment markers, and\nEmacs mode-line declarations (e.g., <code>-*-mode-*-<\/code>) are stripped from\ndisplayed titles. This can be customized by changing\n<code>deft-strip-title-regexp<\/code>.<\/p><p>More generally, the title post-processing function itself can be\ncustomized by setting <code>deft-parse-title-function<\/code>, which accepts\nthe first line of the file as an argument and returns the parsed\ntitle to display in the file browser. The default function is\n<code>deft-strip-title<\/code>, which removes all occurrences of\n<code>deft-strip-title-regexp<\/code> as described above.<\/p><p>For compatibility with other applications which use the filename as\nthe title of a note (rather than the first line of the file), set the\n<code>deft-use-filename-as-title<\/code> flag to a non-<code>nil<\/code> value. Deft will then\nuse note filenames to generate the displayed titles in the Deft\nfile browser. To enable this, add the following to your <code>.emacs<\/code> file:<\/p><pre class=\"lisp\">(setq deft-use-filename-as-title t)\n<\/pre><p>Finally, the short summary that is displayed following the file\ntitle can be customized by changing <code>deft-strip-summary-regexp<\/code>. By\ndefault, this is set to remove certain org-mode metadata statements\nsuch as <code>#+OPTIONS:<\/code> and <code>#+AUTHOR:<\/code>.<\/p><h2>Creating Files<\/h2><p>Filenames for newly created files are generated by Deft automatically.\nThe process for doing so is determined by the variables\n<code>deft-use-filename-as-title<\/code> and <code>deft-use-filter-string-for-filename<\/code>\nas well as the rules in the <code>deft-file-naming-rules<\/code> alist.\nThe possible cases are as follows:<\/p><ol><li><p><strong>Default<\/strong> (<code>deft-use-filename-as-title<\/code> and\n<code>deft-use-filter-string-for-filename<\/code> are both <code>nil<\/code>):<\/p>\n<p>The filename will be automatically generated using an short,\n<abbr class=\"abbr\" title=\"International Organization for Standardization\">ISO<\/abbr>-like timestamp as in <code>2016-05-12T09:00.txt<\/code>. The format\ncan be customized by setting the variable\n<code>deft-new-file-format<\/code>. The filter string will be inserted as\nthe first line of the file (which is also used as the display\ntitle). In case of file name conflicts, an underscore and a\nnumerical suffix (e.g., <code>_2<\/code>) will be appended before the\nextension.<\/p><\/li>\n<li><p><strong>Filenames as titles<\/strong> (<code>deft-use-filename-as-title<\/code> is non-<code>nil<\/code>):<\/p>\n<p>When <code>deft-use-filename-as-title<\/code> is non-<code>nil<\/code>, the filter string\nwill be used as the filename for new files (with the appropriate\nfile extension appended to the end). An example of new file creation\nin this case:<\/p>\n<ul><li>Filter string: \u201cMy New Project\u201d<\/li>\n<li>File name: \u201cMy New Project.txt\u201d<\/li>\n<li>File contents: [empty]<\/li>\n<\/ul><\/li>\n<li><p><strong>Readable filenames<\/strong> (<code>deft-use-filename-as-title<\/code> is\n<code>nil<\/code> but <code>deft-use-filter-string-for-filename<\/code> is non-<code>nil<\/code>):<\/p>\n<p>In this case you can choose to display the title as parsed from\nthe first line of the file while also generating readable\nfilenames for new files based on the filter string. The\nvariable <code>deft-use-filter-string-for-filename<\/code> controls this\nbehavior and decouples the title display\n(<code>deft-use-filename-as-title<\/code>) from the actual filename. New\nfilenames will be generated from the filter string and\nprocessed according to the rules defined in the\n<code>deft-file-naming-rules<\/code> alist. By default, slashes are removed\nand replaced by hyphens, but many other options are possible\n(camel case, replacing spaces by hyphens, and so on). See the\ndocumentation for <code>deft-file-naming-rules<\/code> for additional\ndetails.<\/p>\n<p>As an example, with the following value for\n<code>deft-file-naming-rules<\/code>, Deft will replace all slashes and\nspaces with hyphens and will convert the file name to\nlowercase:<\/p>\n<pre>``` lisp\n(setq deft-file-naming-rules\n      '((noslash . \"-\")\n        (nospace . \"-\")\n        (case-fn . downcase)))\n```\n<\/pre>\n<p>Below is an example in this case, with the above file naming\nrules. Notice that the filter string is inserted as the first\nline of the file but it is also used to generate a \u201creadable\u201d\nfile name.<\/p>\n<ul><li>Filter string: \u201cMy New Project\u201d<\/li>\n<li>File name: \u201cmy-new-project.txt\u201d<\/li>\n<li>File contents: \u201cMy New Project\u201d<\/li>\n<\/ul><\/li>\n<\/ol><p>Titles inserted into files from the filter string can also be\ncustomized for two common modes, <code>markdown-mode<\/code> and <code>org-mode<\/code>, by\nsetting the following variables:<\/p><ul><li><p><code>deft-markdown-mode-title-level<\/code> - When set to a positive\n integer, determines how many hash marks will be added to titles\n in new Markdown files. In other words, setting\n <code>deft-markdown-mode-title-level<\/code> to <kbd>2<\/kbd> will result in new files\n being created with level\u20132 headings of the form <code>## Title<\/code>.<\/p><\/li>\n<li><p><code>deft-org-mode-title-prefix<\/code> - When non-nil, automatically\n generated titles in new <code>org-mode<\/code> files will be prefixed with\n <code>#+TITLE:<\/code>.<\/p><\/li>\n<\/ul><h2>Other Customizations<\/h2><p>Deft, by default, lists files from newest to oldest. You can set\n<code>deft-current-sort-method<\/code> to \u2019title to sort by file titles, case\nignored. Or, you can toggle sorting method using\n<code>deft-toggle-sort-method<\/code>.<\/p><p>Incremental string search is the default method of filtering on\nstartup, but you can set <code>deft-incremental-search<\/code> to nil to make\nregexp search the default.<\/p><p>Deft also provides a function for opening files without using the\nDeft buffer directly. Calling <code>deft-find-file<\/code> will prompt for a\nfile to open, much like <code>find-file<\/code>, but limits consideration to\nfiles in <code>deft-directory<\/code> that are known to Deft (i.e., those files\nmatching <code>deft-extensions<\/code>). Unlike <code>find-file<\/code>, a list of all\nsuch files is provided and the desired file name can be completed\nusing <code>completing-read<\/code> (and, as a result, <code>deft-find-file<\/code> will\nread\/complete filenames using ido, helm, etc. when enabled). If\nthe selected file is in <code>deft-directory<\/code>, it is opened with the\nusual Deft features (automatic saving, automatic updating of the\nDeft buffer, etc.). Otherwise, the file will be opened by\n<code>find-file<\/code> as usual. Therefore, you can set up a global\nkeybinding for this function to open Deft files anywhere. For\nexample, to use <kbd>C-x C-g<\/kbd>, a neighbor of <kbd>C-x C-f<\/kbd>, use the\nfollowing:<\/p><pre class=\"lisp\">(global-set-key (kbd \"C-x C-g\") 'deft-find-file)\n<\/pre><p>The faces used for highlighting various parts of the screen can\nalso be customized. By default, these faces inherit their\nproperties from the standard font-lock faces defined by your current\ncolor theme.<\/p><p>Deft also provides several hooks: <code>deft-mode-hook<\/code>,\n<code>deft-filter-hook<\/code>, and <code>deft-open-file-hook<\/code>. See the\ndocumentation for these variables for further details.<\/p><h2>Acknowledgments<\/h2><p>Thanks to Konstantinos Efstathiou for writing simplenote.el, from\nwhich I borrowed liberally, and to Zachary Schneirov for writing\nNotational Velocity, whose functionality and spirit I wanted to\nbring to Emacs.<\/p><h2>History<\/h2><p>Version 0.8 (2018\u201301\u201312):<\/p><ul><li>Limit <code>deft-find-file<\/code> to files known to Deft and support\n completing-read.<\/li>\n<li>Keep subdirectory portion when displaying filenames.<\/li>\n<li>New variable <code>deft-width-offset<\/code> for custom summary line width\n offset.<\/li>\n<li>Attempt to restore point after refreshing browser and preserve\n position while filtering.<\/li>\n<li>Add hooks: <code>deft-filter-hook<\/code> for filter string changes and\n <code>deft-open-file-hook<\/code> which runs after opening a file.<\/li>\n<li>Prevent spurious Deft browser refreshes, which fixes an issue\n with <code>sublimity-mode<\/code>.<\/li>\n<li>More reliable browser updates when window size changes.<\/li>\n<li>Only update width when buffer is visible.<\/li>\n<li>Lazily update the Deft buffer after saving files.<\/li>\n<li>Close open buffer when deleting a file.<\/li>\n<li>Initialize width even when started in background.<\/li>\n<li>Omit files generated from org or markdown.<\/li>\n<li>Custom format string <code>deft-new-file-format<\/code> for new file names.<\/li>\n<li>Reduce summary line width when there is no fringe.<\/li>\n<li>Support Org links.<\/li>\n<li>Option <code>deft-filter-only-filenames<\/code> to filter only on file names.<\/li>\n<\/ul><p>Version 0.7 (2015\u201312\u201321):<\/p><ul><li>Add custom regular expression <code>deft-strip-summary-regexp<\/code> for\n stripping extraneous text for generating the summary line. Strip\n all <code>org-mode<\/code> metadata by default.<\/li>\n<li>New customizable regular expressions for ignoring files and\n directories. See <code>deft-recursive-ignore-dir-regexp<\/code> and\n <code>deft-ignore-file-regexp<\/code>.<\/li>\n<li>Bug fix: Prevent lines from wrapping in console mode.<\/li>\n<li>Bug fix: Setup <code>deft-extensions<\/code> and <code>deft-default-extension<\/code> at\n load time.<\/li>\n<li>Bug fix: Try to prevent false title matches in org-mode notes\n where the string <code>#+TITLE:<\/code> might also appear in the body.<\/li>\n<li>Bug fix: Use <code>with-current-buffer<\/code> instead of <code>save-excursion<\/code>\n while auto-saving files since we do not want to save the point.<\/li>\n<li>Bug fix: Don\u2019t escape quotes in <code>deft-file-naming-rules<\/code>.<\/li>\n<\/ul><p>Version 0.6 (2015\u201306\u201326):<\/p><ul><li>Recursive search in subdirectories (optional). Set\n <code>deft-recursive<\/code> to a non-nil value to enable.<\/li>\n<li>Support for multiple extensions via the <code>deft-extensions<\/code> list.\n As such, <code>deft-extension<\/code> is now deprecated.<\/li>\n<li>New variable <code>deft-create-file-from-filter-string<\/code> can enable\n generation of new filenames based on the filter string. This decouples\n the title display (<code>deft-use-filename-as-title<\/code>) from the actual filename\n generation.<\/li>\n<li>New variable <code>deft-file-naming-rules<\/code> allows customizing generation\n of filenames with regard to letter case and handling of spaces.<\/li>\n<li>New variables <code>deft-markdown-mode-title-level<\/code> and\n <code>deft-org-mode-title-prefix<\/code> for automatic insertion of title markup.<\/li>\n<li>Archiving of files in <code>deft-archive-directory<\/code>.<\/li>\n<li>Ability to sort by either title or modification time via\n <code>deft-current-sort-method<\/code>.<\/li>\n<li>Update default <code>deft-strip-title-regexp<\/code> to remove the following:\n<ul><li>org-mode <code>#+TITLE:<\/code> tags<\/li>\n<li>MultiMarkdown <code>Title:<\/code> tags<\/li>\n<li>LaTeX comment markers<\/li>\n<li>Emacs mode-line declarations (e.g., <code>-*-mode-*-<\/code>)<\/li>\n<\/ul><\/li>\n<li>Remove leading and trailing whitespace from titles.<\/li>\n<li>Disable visual line mode to prevent lines from wrapping.<\/li>\n<li>Enable line truncation to avoid displaying truncation characters.<\/li>\n<li>Show the old filename as the default prompt when renaming a file.<\/li>\n<li>Call <code>hack-local-variables<\/code> to read file-local variables when\n opening files.<\/li>\n<li>Fixed several byte-compilation warnings.<\/li>\n<li>Bug fix: more robust handling of relative and absolute filenames.<\/li>\n<li>Bug fix: use width instead of length of strings for calculations.<\/li>\n<li>Bug fix: fix <code>string-width<\/code> error with empty file.<\/li>\n<\/ul><p>Version 0.5.1 (2013\u201301\u201328):<\/p><ul><li>Bug fix: creating files with <kbd>C-c C-n<\/kbd> when both the filter string and\n <code>deft-use-filename-as-title<\/code> are non-nil resulted in an invalid path.<\/li>\n<li>Bug fix: killed buffers would persist in <code>deft-auto-save-buffers<\/code>.<\/li>\n<\/ul><p>Version 0.5 (2013\u201301\u201325):<\/p><ul><li>Implement incremental string search (default) and regexp search.\n These search modes can be toggled by pressing <kbd>C-c C-t<\/kbd>.<\/li>\n<li>Default search method can be changed by setting <code>deft-incremental-search<\/code>.<\/li>\n<li>Support custom <code>deft-parse-title-function<\/code> for post-processing titles.<\/li>\n<li>The default <code>deft-parse-title-function<\/code> simply strips occurrences of\n <code>deft-strip-title-regexp<\/code>, which removes Markdown and Org headings.<\/li>\n<li>Open files in another window with <kbd>C-o<\/kbd>. Prefix it with <kbd>C-u<\/kbd> to\n switch to the other window.<\/li>\n<li>For symbolic links, use modification time of taget for sorting.<\/li>\n<li>When opening files, move point to the end of the first match of\n the filter string.<\/li>\n<li>Improved filter editing: delete (<code>DEL<\/code>), delete word (<kbd>M-DEL<\/kbd>),\n and yank (<kbd>C-y<\/kbd>).<\/li>\n<li>Advanced filter editing in minibuffer (<kbd>C-c C-l<\/kbd>).<\/li>\n<\/ul><p>Version 0.4 (2011\u201312\u201311):<\/p><ul><li>Improved filtering performance.<\/li>\n<li>Optionally take title from filename instead of first line of the\n contents (see <code>deft-use-filename-as-title<\/code>).<\/li>\n<li>Dynamically resize width to fit the entire window.<\/li>\n<li>Customizable time format (see <code>deft-time-format<\/code>).<\/li>\n<li>Handle <code>deft-directory<\/code> properly with or without a trailing slash.<\/li>\n<\/ul><p>Version 0.3 (2011\u201309\u201311):<\/p><ul><li>Internationalization: support filtering with multibyte characters.<\/li>\n<\/ul><p>Version 0.2 (2011\u201308\u201322):<\/p><ul><li>Match filenames when filtering.<\/li>\n<li>Automatically save opened files (optional).<\/li>\n<li>Address some byte-compilation warnings.<\/li>\n<\/ul><p>Deft was originally written by <a href=\"https:\/\/jblevins.org\/\">Jason Blevins<\/a>.\nThe initial version, 0.1, was released on August 6, 2011.<\/p><\/header>","created_at":"2019-05-30T20:15:56+0000","updated_at":"2019-05-30T20:16:01+0000","published_by":["Jason Blevins"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"jblevins.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["datalog"],"is_public":false,"id":961,"title":"What you always wanted to know about Datalog (and never dared to ask) - Knowledge and Data Engineering, IEEE Transactions on ","url":"http:\/\/www.lifl.fr\/~kuttler\/elfe\/biblio\/datalog-overview-gottlob.pdf","content":"146 \tIEEE  TRANSACTIONS \tON \tKNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL \tI, NO \t1, MARCH \t1989 \t<br \/>\nWhat \tYou \tAlways Wanted \tto \tKnow \tAbout \tDatalog \t<br \/>\n(And \tNever  Dared \tto \tAsk) \t<br \/>\nSTEFAN0 \tCERI, \tGEORG \tGOTTLOB, \tAND \tLETIZIA \tTANCA \t<br \/>\nAbstract-Datalog \tis  a  database query  language based \ton \tthe  logic \tprogramming paradigm; \tit has  been designed and intensively studied \tover the last \tfive \tyears. \tWe \tpresent the syntax and semantics \tof Datalog \tand its \tuse \tfor querying a relational database. Then, \twe \tclassify opti- \tmization methods  for  achieving  efficient evaluations  of  Datalog  que- ries, and  present the most  relevant methods.  Finally, \twe \tdiscuss var- \tious \texhancements of Datalog, currently under study, and indicate \twhat \tis still needed \tin order to extend Datalog\u2019s applicability to the solution \tof  real-life problems.  The aim \tof this  paper \tis to \tprovide a  survey \tof \tresearch performed on  Datalog,  also  addressed \tto those  members of \tthe database community who are not too familiar with logic program- ming concepts. <br \/>\nZndex \tTerms-Deductive \tdatabases,  logic  programming,  recursive \tqueries, relational databases,  query optimization. \t<br \/>\nI. \tINTRODUCTION \t<br \/>\necent years \thave \tseen  substantial  efforts \tin the direc- \t<br \/>\nR \ttion \tof \tmerging \tartificial  intelligence \tand \tdatabase \t<br \/>\ntechnologies \tfor \tthe \tdevelopment \tof \tlarge \tand \tpersistent \t<br \/>\nknowledge \tbases. \tAn \timportant  contribution  towards  this \t<br \/>\ngoal \tcomes from \tthe integration \tof \tlogic \tprogramming \tand \t<br \/>\ndatabases. \tThe \tfocus has \tbeen \tmostly  concentrated \tby \tthe \t<br \/>\ndatabase  theory \tcommunity  on \twell-formalized \tissues, \t<br \/>\nlike  the  definition \tof \ta new \trule-based \tlanguage, \tcalled \t<br \/>\nDatalog,  specifically  designed \tfor \tinteracting  with \tlarge \t<br \/>\ndatabases; \tand \tthe  definition \tof \toptimization \tmethods for \t<br \/>\nvarious types \tof \tDatalog  rules,  together \twith \tthe  study \tof \t<br \/>\ntheir \tefficiency \t[ 1201, \t[ 151, \t[ 161. \tIn  parallel,  various  ex- \t<br \/>\nperimental  projects \thave shown \tthe  feasibility  of Datalog \t<br \/>\nprogramming \tenvironments \t[ 1191, \t[24], \t[88]. \t<br \/>\nPresent efforts in the integration \tof \tartificial intelligence \t<br \/>\nand \tdatabases  take \ta much \tmore \tbasic \tand \tpragmatic  ap- \t<br \/>\nproach; \tin particular,  several attempts  fall \tin the category \t<br \/>\nof \t\u201cloose  coupling,\u201d \twhere \texisting \tAI \tand \tDB environ- \t<br \/>\nments \tare  interconnected  through  ad-hoc  interfaces.  In \t<br \/>\nother \tcases, \tAI \tsystems \thave \tsolved persistency  issues \tby \t<br \/>\ndeveloping intemal  databases \tfor \ttheir \ttools; \tbut \tthese  in- \t<br \/>\nManuscript  received March \t1, 1989. This \tpaper \tis based on Parts \tI1 and \tI11 of the book by \tS. Ceri, \tG. Gottlob, \tand \tL. Tanca, \tLogic Programming \tand \tDatabases \t(New \tYork: \tSpringer-Verlag. to \tbe published). \tS. Ceri is \twith  the \tDipartimento \tdi Matematica, \tUniversita \tdi Modena, \tModena, \tItaly. \tG. Gottlob \tis with the Institut \tfur \tAngenwandte \tInformatik, \tTU, \tWien, \tAustria. L. Tanca \tis with the \tDipartimento \tdi Elettronica, Politecnico di \tMilano, \tMilano, \tItaly. \tIEEE \tLog \tNumber 8928 \t155. \t\u2018The \tterm \t\u201cDatalog\u201d  was \talso \tused  by \tMaier and  Warren \tin 1821 \tto \tdenote \ta subset \tof Prolog. \t<br \/>\ntemal  databases  typically \tdo \tnot \tallow  data  sharing \tand \t<br \/>\nrecovery,  thus \tdo \tnot \tproperly  belong \tto \tcurrent database \t<br \/>\ntechnology \t[26]. \tThe \tspread \tand \tsuccess \tof \tsuch  en- \t<br \/>\nhanced \tAI \tsystems, \thowever, \tindicate that there is \ta great \t<br \/>\nneed \tfor \tthem. \t<br \/>\nLoose \tcoupling has \tbeen \tattempted \tin the area \tof \tLogic \t<br \/>\nProgramming \tand \tdatabases \tas  well, \tby \tinterconnecting \t<br \/>\nProlog  systems \tto \trelational  databases \t1421, \tt2.51, \t1461, \t<br \/>\n[MI, \t[331, \tP71. \tAlthough \tinteresting  results \thave  been \tachieved, \tmost \t<br \/>\nstudies  indicate  that  simple  interfaces \tare \ttoo  inefficient; \t<br \/>\nan \tenhancement \tin \tefficiency \tis  achieved \tby \tintelligent \t<br \/>\ninterfaces \t[64], \t[33].  This  indicates  that  loose  coupling \t<br \/>\nmight \tin \tfact  solve  today\u2019s  problems, \tbut  on \tthe  long \t<br \/>\nrange,  strong  integration  is required. \tMore \tgenerally, \twe \t<br \/>\nexpect  that \tknowledge \tbase \tmanagement \tsystems  will \t<br \/>\nprovide  direct  access  to  data \tand \twill  support  rule-based \t<br \/>\ninteraction \tas  one \tof \tthe \tprogramming \tparadigms. \tDa- \t<br \/>\ntalog \tis a first \tstep \tin this \tdirection. \t<br \/>\nThe \treaction \tof \tthe database \tcommunity \tto Datalog  has \t<br \/>\noften \tbeen  marked \tby \tskepticism;  in  particular,  the  im- \t<br \/>\nmediate \tor \teven \tfuture  practical  use \tof \tresearch \ton \tso- \t<br \/>\nphisticated  rule-based  interactions  has \tbeen \tquestioned \t<br \/>\n[94].  Nevertheless, \twe  do \texpect  that  Datalog\u2019s  experi- \t<br \/>\nence, \tproperly  filtered, will teach  important lessons \tto re- \t<br \/>\nsearchers involved \tin the \tdevelopment \tof \tknowledge \tbase \t<br \/>\nsystems. \tThe \tpurpose \tof \tthis  paper \tis to  give \ta self-con- \t<br \/>\ntained  survey \tof \tthe  research  that  has \tbeen performed \tre- \t<br \/>\ncently \ton \tDatalog. \tMore \texhaustive  treatments  can \tbe \t<br \/>\nfound \tin the \tbooks [34], \t[ 1221, and \t[92]. \tOther interesting \t<br \/>\nsurvey papers,  approaching the subject \tfrom \tdifferent per- \t<br \/>\nspectives,  are  [53] \tand \t[loo]. \t<br \/>\nThis  paper  is  organized \tas \tfollows. \tIn \tSection \tI1 \twe \t<br \/>\npresent the foundations \tof \tDatalog:  the syntactic structure \t<br \/>\nand \tthe  semantics of  Datalog  programs. \tIn \tSection \tI11 \twe \t<br \/>\nexplain  how  Datalog  is \tused as a \tquery  language  over \tre- \t<br \/>\nlational databases; \tin particular,  we indicate \thow \tDatalog \t<br \/>\ncan  be \timmediately  translated \tto \tequations \tof \trelational \t<br \/>\nalgebra.  In Section \tIV \twe  present \ta taxonomy \tof \tthe var- \t<br \/>\nious \toptimization  methods, \temphasizing \tthe  distinction \t<br \/>\nbetween program \ttransformation \tand \tevaluation  methods. \t<br \/>\nIn \tSection \tV, \twe  present \ta survey \tof \tevaluation \tmethods \t<br \/>\nand program \ttransformation techniques,  selecting the \tmost \t<br \/>\nrepresentative  ones  within  the  classes  outlined \tin \tSection \t<br \/>\nIV. \tIn \tSection \tVI, \twe  present  several  formal  extensions \t<br \/>\ngiven  to  the  Datalog  language  to \tenhance \tits  expressive- \t<br \/>\nness,  and \twe \tsurvey \tsome \tof \tthe current research projects \t<br \/>\n1041-4347\/89\/0300-0146$01 \t.OO \t0 \t1989 \tIEEE \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERI \tet al.: \tDATALOG \t147 \t<br \/>\nin this \tarea. \tFinally, \tin \tSection \tVI1 \twe  attempt \tan \tevalu- \t<br \/>\nation  of  what  will \tbe \trequired  in \tDatalog \tin \torder \tto \tbe- \t<br \/>\ncome more attractive and usable. \t<br \/>\n11. \tDATALOG: \tSEMANTICS \tAND \tEVALUATION \t<br \/>\nPARADIGMS \t<br \/>\nIn this section we \tdefine \tthe \tsyntax  of \tthe Datalog \tquery \t<br \/>\nlanguage  and  explain \tits \tlogical  semantics. \tWe \tgive \ta \t<br \/>\nmodel-theoretic characterization \tof \tDatalog programs and \t<br \/>\nshow  how  they \tcan be evaluated \tin \ta  bottom-up \tfashion. \t<br \/>\nThis evaluation corresponds \tto computing \ta least \tfixpoint. \t<br \/>\nFinally,  we \tbriefly \tmention \tanother evaluation paradigm \t<br \/>\ncalled \t\u201ctop-down  evaluation, \t\u201d and \tcompare  Datalog \tto \t<br \/>\nthe \twell-known \tlogic programming  language \tProlog. \t<br \/>\nA. \tThe Syntax \tof \tDatalog  Programs \t<br \/>\nDatalog is \tin many  respects  a simplified  version  of \tgen- \t<br \/>\neral \tLogic  Programming \t1781. \tA  logic \tprogram  consists \t<br \/>\nof \ta finite \tset \toffacts \tand \trules. \tFacts are assertions about \t<br \/>\na relevant  piece  of \tthe world, \tsuch \tas: \t\u201cJohn \tis the father \t<br \/>\nof \tHarry\u201d. \tRules \tare sentences \twhich \tallow \tus \tto \tdeduce \t<br \/>\nfacts  from other  facts. \tAn \texample \tof  a  rule \tis: \t\u201cIf \tX is \t<br \/>\na  parent  of \tY and \tif Y is  a  parent \tof \t2, \tthen \tX is a  grand- \t<br \/>\nparent  of \tZ\u201d. \tNote \tthat \trules, \tin order \tto be \tgeneral, \tusu- \t<br \/>\nally \tcontain \tvariables \t(in \tour \tcase, \tX, \tY, \tand \t2). \tBoth \t<br \/>\nfacts and \trules \tare \tparticular forms \tof \tknowledge. \t<br \/>\nIn \tthe  formalism \tof \tDatalog \tboth  facts \tand \trules \tare \t<br \/>\nrepresented \tas \tHorn  clauses \tof \tthe  general shape \t<br \/>\nLo \t:- \tL1, \t. . . \t, L, \t<br \/>\nwhere each \tL, \tis a \tliteral \tof \tthe form \tpI ( tl, \t. . . \t, tk,) \tsuch \t<br \/>\nthat \tpI \tis a predicate \tsymbol \tand the \tt, are \tterms. \tA term \t<br \/>\nis  either \ta constant \tor \ta variable.* \tThe \tleft-hand \tside \t<br \/>\n(LHS)  of  a \tDatalog clause \tis called \tits \thead \tand \tthe  right- \t<br \/>\nhand  side \t(RHS) \tis  called  its \tbody. \tThe \tbody  of  a \tclause \t<br \/>\nmay \tbe \tempty.  Clauses \twith \tan \tempty \tbody  represent \t<br \/>\nfacts; clauses \twith \tat least \tone \tliteral \tin the \tbody  represent \t<br \/>\nrules. <br \/>\nThe \tfact \t\u201cJohn \tis the father \tof \tBob\u201d, \tfor example, \tcan \t<br \/>\nbe \trepresented \tas father(bob, \tjohn). \tThe \trule  \u201cIf \tX is a \t<br \/>\nparent \tof \tY and \tif \tY is a  parent  of \tZ, \tthen \tX is  a \tgrand- \t<br \/>\nparent \tof \tZ\u201d \tcan be \trepresented \tas \t<br \/>\ngrandpar(Z, \tX) \t: - par( \tY, \tX), \tpar(Z, \tY). \t<br \/>\nHere  the  symbols \tpar \tand \tgrandpar \tare \tpredicate \tsym- \t<br \/>\nbols, \tthe \tsymbols \tjohn \tand \tbob \tare \tconstants,  and \tthe \t<br \/>\nsymbols \tX, \tY, \tand \tZ are \tvariables. \tWe \twill \tuse  the  fol- \t<br \/>\nlowing \tnotational \tconvention:  constants  and  predicate \t<br \/>\nsymbols  are  strings beginning \twith  a \tlower-case  charac- \t<br \/>\nter;  variables  are  strings \tbeginning  with \tan \tupper-case \t<br \/>\ncharacter.  Note that \tfor \ta given Datalog program \tit  is \tal- \t<br \/>\nways \tclear from the context whether \ta particular nonvari- \t<br \/>\nable  symbol \tis  a \tconstant \tor \ta  predicate \tsymbol. \tWe \tre- \t<br \/>\nquire \tthat  all \tliterals \twith \tthe  same predicate  symbol  are \t<br \/>\nof \tthe \tsame \tarity, \ti.e., \tthat \tthey \thave \tthe \tsame number \tof \t<br \/>\narguments. \tA literal, \tfact, \trule, \tor \tclause \twhich \tdoes \tnot \t<br \/>\ncontain  any \tvariables \tis called \tground. \t<br \/>\nAny \tDatalog  program \tP \tmust  satisfy \tthe \tfollowing \t<br \/>\nsafety \tconditions. <br \/>\nEach \tfact \tof \tP is ground. \tEach  variable \twhich \toccurs \tin  the  head  of  a  rule \tof \t<br \/>\nThese conditions guarantee \tthat \tthe \tset \tof \tall \tfacts \tthat \t<br \/>\nP must also occur \tin the \tbody  of \tthe \tsame \trule. \t<br \/>\ncan \tbe \tderived from \ta Datalog program is finite. \t<br \/>\nB. \tDatalog  and  Relational  Databases <br \/>\nIn \tthe \tcontext \tof \tgeneral Logic Programming \tit is usu- \t<br \/>\nally \tassumed \tthat  all \tthe knowledge \t(facts \tand \trules)  rel- \t<br \/>\nevant \tto \ta particular application \tis contained \twithin  a  sin- \t<br \/>\ngle logic program \tP. \tDatalog,  on the \tother hand, \thas \tbeen \t<br \/>\ndeveloped \tfor \tapplications \twhich \tuse \ta large  number \tof \t<br \/>\nfacts \tstored \tin  a  relational \tdatabase. \tTherefore, \twe \twill \t<br \/>\nalways consider \ttwo sets \tof \tclauses: \ta set of \tground \tfacts, \t<br \/>\ncalled \tthe \tExtensional  Database \t(EDB), \tphysically  stored \t<br \/>\nin \ta  relational \tdatabase,  and \ta Datalog  program \tP called \t<br \/>\nthe \tInzpional \tDatabase  (ZDB) \t.3 The \tpredicates  occur- \t<br \/>\nring  in \tthe \tEDB \tand \tin P are \tdivided into \ttwo \tdisjoint \tsets: \t<br \/>\nthe \tEDB-predicates, \twhich \tare \tall  those \toccurring \tin \tthe \t<br \/>\nextensional \tdatabase and \tthe \tIDB-predicates, \twhich \toccur \t<br \/>\nin P but  not \tin \tthe \tEDB. \tWe \trequire that the \thead  predi- \t<br \/>\ncate \tof \teach \tclause in \tP be \tan \tIDB-predicate. \tEDB-pred- \t<br \/>\nicates \tmay \toccur in \tP, \tbut \tonly \tin \tclause bodies. \t<br \/>\nGround facts are stored \tin  a  relational \tdatabase; \twe \tas- \t<br \/>\nsume \tthat \teach \tEDB-predicate \tr corresponds \tto \texactly \t<br \/>\none \trelation \tR of \tour \tdatabase  such  that \teach \tfact \tr( cI, \t<br \/>\n. . . , ck) \tof \tthe EDB \tis stored \tas \ta tuple \t&lt; cl, \t. . . \t, e, &gt; \t<br \/>\nof \tR. <br \/>\nAlso \tthe IDB-predicates \tof \tP can \tbe \tidentified \twith \tre- \t<br \/>\nlations, \tcalled \tIDB-relations, \tor \talso \tderived \trelations, \t<br \/>\ndefined \tthrough  the  program \tP and  the \tEDB. \tIDB \trela- \t<br \/>\ntions \tare \tnot  stored \texplicitly,  and  correspond \tto \trela- \t<br \/>\ntional \tviews. \tThe \tmaterialization  of these \tviews, \ti.e., \ttheir \t<br \/>\neffective \t(and \tefficient) \tcomputation, \tis the \tmain  task \tof \t<br \/>\na Datalog compiler \tor \tinterpreter. \t<br \/>\nAs \tan \texample \tof a relational \tEDB, \tconsider \ta database \t<br \/>\nEl \tconsisting \tof \ttwo \trelations  with  respective \tschemes \t<br \/>\nPERSON \t(NAME) \tand \tPAR \t(CHILD, \tPARENT). \tThe \tfirst \t<br \/>\ncontains the names \tof \tpersons and the \tsecond \texpresses \ta \t<br \/>\nparent \trelationship  between  persons.  Let  the  actual \tin- \t<br \/>\nstances \tof \tthese \trelations \thave \tthe \tfollowing \tvalues: \t<br \/>\nPERSON \t= \t{ &lt; ann \t&gt; , &lt; \tbertrand \t&gt; , &lt; Charles \t&gt; , \t<br \/>\n&lt; dorothy \t&gt; , &lt; evelyn \t&gt; , &lt; ped \t&gt; , \t<br \/>\n&lt; george \t&gt; , &lt; hilary \t&gt; } \t<br \/>\nPAR \t= \t{ &lt; dorothy, \tgeorge \t&gt; , &lt; evelyn, \tgeorge \t&gt; , \t<br \/>\n&lt; bertrand, \tdorothy \t&gt; , &lt; ann, \tdorothy \t&gt; , \t<br \/>\n&lt; ann, \thilaly \t&gt; , &lt; Charles, evelyn \t&gt; 1. \t<br \/>\n\u2019Note  that \tin general \tLogic Programming, \ta term \tcan \talso \tbe a complex \tnested  structure  made \tof function \tsymbols, \tconstants, \tand \tvariables. \tEx- \ttensions \tof Datalog \tto cover \tsuch \tcomplex \tterms \tare  outlined \tin Section \tVI. \tis harmless. <br \/>\n\u2018In \tthe \tliterature \tan IDB \tis usually  defined \tas a set \tof rules. \tHere \tthe \tDatalog  program \tP may \talso \tcontain \tfacts. \tThis \tterminological  extension \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n148 \tIEEE \tTRANSACTIONS \tON  KNOWLEDGE  AND \tDATA \tENGINEERING, \tVOL. \t1. NO. \t1, MARCH \t1989 \t<br \/>\nThese \trelations \texpress \tthe \tset \tof  ground  facts \tE = \t{per- \t<br \/>\nson \t(ann), \tperson \t(bertrand), \t. . . \t, par \t(dorothy, george), \t<br \/>\n. . . \t, par(charles, \tevelyn)}. \t<br \/>\nLet \tPI \tbe \ta Datalog \tprogram consisting \tof \tthe \tfollowing \t<br \/>\nclauses: <br \/>\nrl: \tsgc(X, \tX) \t:- \tperson(X). \t<br \/>\nr2: \tsgc(X, \tY) \t:- \tpar(X, \tXI), \tsgc(X1, \tYI),par(Y, \t~1). \t<br \/>\nDue \tto  rule \trl, \tthe  derived \trelation \tSGC \twill  contain  a \t<br \/>\ntuple \t&lt; p, \tp &gt; for \teach \tpersonp. \tRule \tr2 is recursive and \t<br \/>\nstates that \ttwo \tpersons \tare same \tgeneration cousins \twhen- \t<br \/>\never \tthey \thave \tparents  which \tare \tin turn  same generation \t<br \/>\ncousins. \tFurther examples \tof tuples belonging \tto SGC \tare: \t<br \/>\n&lt; dorothy, \tevelyn \t&gt; , &lt; Charles, \tann \t&gt; , and \t&lt; ann, \t<br \/>\nCharles \t&gt; . \t<br \/>\nNote  that  program \tPI \tcan \tbe \tconsidered \tas \ta query \t<br \/>\nagainst \tthe  EDB \tEl, \tproducing  as  answer  the \trelation \t<br \/>\nSGC. \tIn \tthis  setting, the \tdistinction \tbetween the two \tsets \t<br \/>\nof \tclauses, \tE and \tP, \tmakes \tyet \tmore  sense. \tUsually  a \t<br \/>\ndatabase \t(in \tour \tcase  the  EDB)  is  considered \tas \ta  time- \t<br \/>\nvarying \tcollection \tof \tinformation. \tA query \t(in \tour case, \t<br \/>\na program \tP), \ton the other \thand, \tis a time-invariant  map- \t<br \/>\nping \twhich \tassociates \ta  result \tto \teach \tpossible \tdatabase \t<br \/>\nstate. \tFor \tthis  reason  we  will \tformally \tdefine \tthe  seman- \t<br \/>\ntics  of  a  Datalog  program \tP as \ta mapping  from database \t<br \/>\nstates  to  result \tstates. \tThe \tdatabase states are collections \t<br \/>\nof \tEDB-facts \tand  the  result  states \tare \tIDB-facts. \t<br \/>\nUsually \tDatalog  programs \tdefine \tlarge \tIDB-relations. \t<br \/>\nIt often \thappens \tthat  a  user \tis \tinterested \tin \ta subset \tof \t<br \/>\nthese relations. \tFor \tinstance, he \tor \tshe \tmight want \tto know \t<br \/>\nthe \tsame  generation  cousins \tof \tAnn \tonly \trather \tthan  all \t<br \/>\nsame generation cousins \tof \tall \tpersons \tin the database. \tTo \t<br \/>\nexpress  such  an \tadditional \tconstraint, \tone \tcan \tspecify  a \t<br \/>\ngoal \tto \ta Datalog program. \tA  goal \tis a single \tliteral  pre- \t<br \/>\nceded by  a \tquestion \tmark  and  a \tdash, \tfor \texample, \tin our \t<br \/>\ncase, \t?-sgc(ann, \tX). \tGoals \tusually \tserve \tto formulate ad \t<br \/>\nhoc  queries \tagainst \ta view  defined \tby \ta Datalog \tprogram. \t<br \/>\nC. \tThe \tLogical  Semantics \tof Datalog \t<br \/>\nEach \tDatalog \tfact \tF can \tbe \tidentified  with \tan  atomic \t<br \/>\nformula \tF* \tof \tFirst-Order Logic.  Each Datalog rule \tR of \t<br \/>\nthe  form \tLo \t: - \tL1, \t. . . \t, L, \trepresents \ta  first-order \tfor- \t<br \/>\nmula \tR* \tof \tthe form \tVX,, \t. . . VX,(L1 \tA . . . \tA L, \t* \tLo), \t<br \/>\nwhere \tXI, \t. . . \t, X, \tare \tall \tthe variables occurring in \tR. \tA \t<br \/>\nset \tS of \tDatalog \tclauses \tcorresponds \tto \tthe \tconjunction \t<br \/>\nS* \tof \tall \tformulas \tC* \tsuch  that \tC E S. \t<br \/>\nThe \tHerbrand \tBase \tHB \tis  the \tset  of  all \tfacts \tthat \twe \t<br \/>\ncan  express \tin \tthe  language \tof \tDatalog, \ti.e., \tall  literals \t<br \/>\nof \tthe \tform \tP( \tc1, \t. . \t. , ck) \tsuch \tthat  all \tci are \tconstants. \t<br \/>\nFurthermore, \tlet \tEHB \tdenote  the  extensional \tpart  of \tthe \t<br \/>\nHerbrand \tbase, \ti.e., \tall \tliterals \tof \tHB \twhose predicate \tis \t<br \/>\nan \tEDB-predicate \tand, \taccordingly, let \tZHB \tdenote the \tset \t<br \/>\nof \tall  literals  of \tHB \twhose \tpredicate is an IDB-predicate. \t<br \/>\nIf \tS is \ta finite \tset \tof \tDatalog \tclauses, \twe \tdenote \tby \t<br \/>\ncons \t(S \t) the \tset  of  all  facts \tthat are logical consequences \t<br \/>\nof \ts*. <br \/>\nThe \tsemantics \tof  a \tDatalog \tprogram \tP can \tnow \tbe de- \t<br \/>\nscribed \tas \ta mapping \t3np \tfrom \tEHB \tto IHB \twhich  to each \t<br \/>\npossible \textensional database \tE C \tEHB \tassociates the \tset \t<br \/>\n3np \t( E ) of \tintensional \t\u2018 \u2018result facts\u201d \tdefined \tby \t3np \t(E \t) \t<br \/>\n= cons \t(P \tU \tE) \tfl \tZHB. \t<br \/>\nLet \tK \tand \tL be \ttwo \tliterals  (not \tnecessarily \tground). \t<br \/>\nWe \tsay \tthat \tK subsumes \tL, \tdenoted \tby \tK \tD \tL, \tif there \t<br \/>\nexists \ta  substitution \t0 of \tvariables such \tthat \tKO \t= L, \ti.e., \t<br \/>\nif 0 applied \tto \tK yields \tL. \tIf K \tD \tL we \talso \tsay \tthat \tL is \t<br \/>\nan \tinstance \tof \tK. \tFor \texample, \tq(a, \tb, \tb) \tand \tq(c, \tc, c) \t<br \/>\nare both  instances \tof \tq(X, \tY, \tY), \tbut \tq(b, \tb, \ta) \tis not. \t<br \/>\nWhen \ta  goal \t\u201c?- \tG\u201d \tis given, \tthen  the  semantics \tof \t<br \/>\nthe \tprogram \tP with  respect \tto this \tgoal \tis a mapping \t3np, \t<br \/>\nfrom \tEHB \tto \tZHB \tdefined \tas \tfollows \t<br \/>\nVEG \tEHB3np,G(E)= \t{HIHE~~~(E)AG&gt;H}. \t<br \/>\nD. \tThe Model  Theory \tof \tDatalog \t<br \/>\nModel \ttheory \tis a branch \tof  mathematical \tlogic \twhich \t<br \/>\ndefines \tthe  semantics \tof \tformal  systems, \tby \tconsidering \t<br \/>\ntheir possible  interpretations, \ti.e., \tthe \tdifferent  intended \t<br \/>\nmeanings  that \tthe \tused \tsymbols  and  formulas \tmay \tas- \t<br \/>\nsume. \tEarly \tdevelopments \tof \tgeneral \tmodel  theory \tdate \t<br \/>\nback \tto works \tof \tLowenheim [79], \tSkolem \t[ 1131, \tand \tGO- \t<br \/>\ndel \t[56]. \tA \tcomprehensive  exposition \tis \tgiven \tin \t[39]. \t<br \/>\nThe \tmodel  theoretic \tcharacterization \tof \tgeneral logic \tthe- \t<br \/>\nories often \trequires \tthe use \tof \tquite \tsophisticated  tools  of \t<br \/>\nmodern \talgebra \t[83]. \tThe \tmodel  theory  of \tHorn  clause \t<br \/>\nsystems  and \tlogic programs,  however,  is \tless  difficult.  It \t<br \/>\nhas  been \tdeveloped  in  [125], \t[ll] \tand \tis  also \texplained \t<br \/>\nin \t[78]. \tDatalog, being \ta  simplified \tversion \tof \tLogic \tPro- \t<br \/>\ngramming, \tcan \tbe \tdescribed \tvery  easily \tin terms \tof model \t<br \/>\ntheory. <br \/>\nAn \tinterpretation \t(in \tthe  context \tof \tDatalog) \tconsists \t<br \/>\nof \tan  assignment \tof  a \tconcrete  meaning \tto \tconstant \tand \t<br \/>\npredicate symbols. \tA Datalog clause can be \tinterpreted \tin \t<br \/>\nseveral \tdifferent \tways. \tA clause \tmay \tbe \ttrue under \ta  cer- \t<br \/>\ntain  interpretation  and \tfalse under another \tone. \tIf a clause \t<br \/>\nC is  true  under \ta given  interpretation, \twe \tsay  that  this \t<br \/>\ninterpretation \tsatisfies \tC. \t<br \/>\nThe \tconcept \tof \tlogical  consequence, \tin \tthe \tcontext  of \t<br \/>\nDatalog,  can \tbe \tdefined \tas follows: \ta fact \tF follows log- \t<br \/>\nically \tfrom \ta  set \tS of \tclauses, \tiff \teach \tinterpretation \tsat- \t<br \/>\nisfying every \tclause \tof \tS also \tsatisfies \tF. \tIf F follows from \t<br \/>\nS, \twe  write \tS E \tF. \t<br \/>\nNote \tthat \tthis \tdefinition \tcaptures quite \twell \tour \tintuitive \t<br \/>\nunderstanding  of \tlogical  consequence.  However,  since \t<br \/>\ngeneral \tinterpretations \tare \tquite \tunhandy \tobjects, \twe \twill \t<br \/>\nlimit \tourselves \tto \tconsider interpretations \tof  a  particular \t<br \/>\ntype, \tcalled \tHerbrand  Interpretations. \tThe \trecognition \t<br \/>\nthat \twe \tmay  forget \tabout \tall \tother \tinterpretations \tis due \t<br \/>\nto \tthe  famous  logicians  Lowenheim,  Skolem, \tand \tHer- \t<br \/>\nbrand.4 <br \/>\nA \tHerbrand \tinterpretation \tassigns \tto \teach  constant \t<br \/>\nsymbol \t\u201citself\u2018 \t\u2019 , i.e., \ta lexicographic  entity.  Predicate \t<br \/>\nsymbols \tare \tassigned \tpredicates \tranging \tover  constant \t<br \/>\nsymbols. \tThus, \ttwo \tnonidentical \tHerbrand \tinterpretations \t<br \/>\ndiffer \tonly \tin \tthe \trespective \tinterpretations \tof \tthe \tpredi- \t<br \/>\ncate  symbols. \tFor \tinstance,  one  Herbrand \tinterpretation \t<br \/>\nmay  satisfy \tthe  fact \t1 (t, \tq) \tand \tanother one \tmay  not \tsat- \t<br \/>\n4Actually, \tHerbrand  interpretations \tcan  be \tdefined  in \tthe \tmuch \tmore \tgeneral \tcontext \tof Clausal \tLogic \t[39], \t(801 \tand \tLogic \tProgramming \t[125], \t1781. \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERI \tet al.: \tDATALOG \t149 \t<br \/>\nisfy this \tfact. \tFor \tthis \treason, any \tHerbrand  interpretation \t<br \/>\ncan \tbe \tidentified with  a \tsubset \t5 of \tthe Herbrand base \tHB. \t<br \/>\nThis \tsubset \tcontains  all \tthe \tground \tfacts  which \tare  true \t<br \/>\nunder \tthe  interpretation. \tThus, \ta ground fact \tp ( c,, \t. . . \t, \t<br \/>\nc,) \tis true under the \tinterpretation \t5 iff p ( c,, \t. . . , c,) \tE \t<br \/>\n5. \tA Datalog \trule  of \tthe  form \t&amp;: \t-L,, \t. . . \t, L, \tis true \t<br \/>\nunder \t5 iff \tfor each \tsubstitution \t8 which  replaces \tvariables \t<br \/>\nby \tconstants, whenever \tLl0 \tE 5 A . . . \tA L,8 \tE 5, \tthen \tit \t<br \/>\nalso \tholds  that \tLOB \tE 9. \t<br \/>\nA Herbrand \tinterpretation  which  satisfies  a \tclause \tC or \t<br \/>\na  set \tof \tclauses \tS is \tcalled \ta Herbrand  model \tfor \tC or, \t<br \/>\nrespectively, \tfor \tS. \t<br \/>\nConsider, \tfor \texample,  the  Herbrand \tinterpretations \t<br \/>\n4 , \t= \t{person \t(john), \tperson \t(jack), \tperson \t( jim), \t<br \/>\nsgc(john,  john), \tpar(jim, \tjohn),  par( jack, \tjohn), \t<br \/>\nsgc \t(john, \tjohn), \tsgc \t(jack, \tjack), \tsgc \t( jim, \tjim)} \t. Ob- \t<br \/>\nviously, \t5, \tis not  a \tHerbrand \tmodel  of \tthe \tprogram \tP, \t<br \/>\ngiven \tin \tSection \t11-B. \tLet \t5, \t= \t5, \tU \t{sgc(jim, \tjack), \t<br \/>\nsgc( \tjack, \tjim)} \t. It is \teasy \tto  see \tthat \t9, \tis \ta Herbrand \t<br \/>\nmodel \tof \tP,. \t<br \/>\nThe \tset \tcons \t(S \t) of  all \tconsequence facts \tof \ta  set \tS of \t<br \/>\nDatalog \tclauses \tcan \tthus \tbe \tcharacterized  as  follows: \t<br \/>\ncons \t(S \t) is the  set  of  all  ground \tfacts which \tare \tsatisfied \t<br \/>\nby \teach  Herbrand \tmodel  of \tS. \tSince \ta  ground  fact \tF is \t<br \/>\nsatisfied \tby \ta Herbrand \tinterpretation \t9 iff F E 9, \tcons \t( S ) \t<br \/>\nis equal \tto \tthe \tintersection \tof \tall  Herbrand  models \tof \tS. \t<br \/>\nSummarized \t<br \/>\ncons(S) \t= \t{FE \tHB(S \tE \tF} \t<br \/>\n= \tfl \t{ 5 15 \tis a Herbrand  model  of \tS}. \t<br \/>\nLet \t9, \tand \t5, \tbe \ttwo Herbrand  models \tof \tS. \tIt is easy  to \t<br \/>\nsee  that  their \tintersection \t9* \tfl \t$J2 \tis \talso \ta Herbrand \t<br \/>\nmodel \tof \tS. \tMore  generally, \tit can be shown \tthat \tthe  in- \t<br \/>\ntersection  of \tan \tarbitrary \t(possibly  infinite) \tnumber \tof \t<br \/>\nHerbrand models \tof \tS always \tyields  a \tHerbrand \tmodel \tof \t<br \/>\nS. \tThis \tproperty \tis called the \tmodel \tintersection \tproperty. \t<br \/>\nNote \tthat  the \tmodel \tintersection \tproperty \tnot \tonly \tholds \t<br \/>\nfor Datalog clauses, \tbut \tfor \ta more \tgeneral \ttype \tof \tclause \t<br \/>\ncalled \tdefinite \tHorn \tclauses \t[ 1251, \t[78]. \t<br \/>\nFrom \tthe \tmodel \tintersection \tproperty  it \tfollows, \tin par- \t<br \/>\nticular, that \tfor \teach \tset \tS of \tDatalog clauses, \tcons \t(S \t) is \t<br \/>\na Herbrand \tmodel  of \tS. \tSince \tcons \t(S \t) is a  subset \tof \tany \t<br \/>\nother  Herbrand \tmodel  of \tS, \twe \tcall \tcons(S) \tthe \tleast \t<br \/>\nHerbrand  model \tof \tS. \t<br \/>\nUp \tto \there \tnothing \thas \tbeen  said \tabout  how \tcons(S) \t<br \/>\ncan \tbe \tcomputed. \tThe \tfollowing  subsections  as \twell \tas \t<br \/>\nSections \t111-V \tdeal \twith \tthis problem. \tIn the \trest  of \tthis \t<br \/>\nsection \twe \twill \tremain \tat \ta quite  abstract  level  and \twe \t<br \/>\nber \ttheory) \tis faced \twith \tconsiderable  problems \tsuch \tas \t<br \/>\nconsistency  and \tincompleteness \tissues \t[SI, \t[57], \tthe \t<br \/>\nproof  theoretic \tanalysis \tof \tsubformalisms \tof  first-order \t<br \/>\nlogic,  such  as  Logic  Programming \tor \tDatalog, \tis easier \t<br \/>\nand  much \tless \tambitious,  and \tis \tmore  oriented  towards \t<br \/>\nalgorithmic issues. <br \/>\nIn \tthis \tsection \twe  show  how \tDatalog \trules  can \tbe \tused \t<br \/>\nto produce \tnew  facts \tfrom given facts. \tWe \tdefine  the  no- \t<br \/>\ntion  of \t\u201cfact \tinference\u201d \tand \tintroduce \ta  proof-theoretic \t<br \/>\nframework \twhich \tallows  one \tto \tinfer \tall  ground  facts \t<br \/>\nwhich \tare consequences \tof  a finite  set \tof \tDatalog clauses. \t<br \/>\nConsider \ta Datalog \trule \tR of \tthe  form \tLo \t: - LI, \t. . . , \t<br \/>\nL, \tand  a list \tof \tground facts \tF,, \t. . \t. , F,. \tIf a  substitution \t<br \/>\n8 exists such \tthat \tfor \teach \t1 5 \ti I \tn Lie \t= Fi then, \tfrom \t<br \/>\nrule \tR and from \tthe \tfacts \tF,, \t. . . \t, F,, \twe \tcan \tinfer \tin one \t<br \/>\nstep \tthe \tfact \tLOO. \tThe \tinferred fact \tmay \tbe \teither \ta new \t<br \/>\nfact \tor \tit may \tbe \talready \tknown. \t<br \/>\nWhat \twe \thave \tjust \tdescribed is \ta general \tinference \trule, \t<br \/>\nwhich  produces \tnew \tDatalog \tfacts \tfrom \tgiven  Datalog \t<br \/>\nrules  and \tfacts. \tWe \trefer \tto \tthis  rule \tas \tthe \tElementary \t<br \/>\nProduction \tPrinciple \t(EPP). \tIn \tsome  sense, \tEPP \tcan \tbe \t<br \/>\nconsidered \tas \tbeing \ta meta-rule, \tsince \tit is independent \t<br \/>\nof \tany \tparticular  Datalog  rules, \tand  treats  them \tjust  as \t<br \/>\nsyntactic \tentities. \t<br \/>\nFor \texample,  consider  the  Datalog \trule \trl \tof  program \t<br \/>\nP, : \t<br \/>\nr,: \tsgc(X, \tX) \t:- \tperson(X). \t<br \/>\nFrom this rule \tand \tfrom the  fact \tperson \t(george) \twe  can \t<br \/>\ninfer \tin \tone \tstep \tsgc(george, \tgeorge). \tThe \tsubstitution \t<br \/>\nused \there \twas \t8 = \t{X \t+ \tgeorge} \t. Now  recall \tthe \tsecond \t<br \/>\nrule  of \tP, \t<br \/>\nr,: \tsgc(X, \tY) \t:- \tpar(&amp; \tXI), \tsgc(X1, \tYI), \tpar( \tY, \tYI) \t<br \/>\nand \tconsider the facts \tpar(dorothy, \tgeorge), \tsgc \t(george, \t<br \/>\ngeorge), \tand \tpar \t(evelyn, george). \tBy \tapplying \tEPP \tand \t<br \/>\nusing \tthe  substitution \t8 = \t{X \t+ \tdoruthy, \tY + \tevelyn, \t<br \/>\nXI \t+ \tgeorge, \tYl \t+ \tgeorge}, \twe \tinfer \tin \tone  step \t<br \/>\nsgc \t(dorothy, \tevelyn). \t<br \/>\nLet \tS be \ta  set of \tDatalog clauses.  Informally, \ta ground \t<br \/>\nfact \tF can \tbe \tinferred \tfrom \tS, \tdenoted \tby \tS I- \tF iff \teither \t<br \/>\nF E S or \tF can \tbe \tobtained \tby \tapplying \tthe  inference  rule \t<br \/>\nEPP \ta  finite \tnumber \tof \ttimes. \tThe \trelationship \t\u201c t- \u201d is \t<br \/>\nmore \tprecisely  defined \tby \tthe following \trecursive \trules: \t<br \/>\nSt- \tFifFES. \t<br \/>\nS t- \tF if a  rule \tR E S and \tground  facts \tF,, \t. . \t. , F, \t<br \/>\nexist \tsuch \tthat \tVl \t5 \ti 5 \tn S \tFi \tand \tF can \tbe \tinferred \tin \tone  step \tby \tthe application \tof \tEPP \tto \tR and \tF,, \t. . \t. , \tr- \twill \tnot \tcare about implementation \tand \tstorage issues. \tIn \t<br \/>\nparticular, \twe \twill \tignore \tthat  a \tset \tS of \tDatalog  clauses \t<br \/>\nmay \tconsist \tof \tprogram clauses and \tof \tEDB-clauses. \tThe \t<br \/>\nproblem  of  retrieving \tfacts  from  an  EDB  is  deferred \tto \t<br \/>\nSections \t111-V. \t<br \/>\nrn\u2019 The \tsequence \tof \tapplications \tof \tEPP \twhich \tis used  to \t<br \/>\ninfer  a \tground  fact \tF from \tS is called \ta proof \tof \tF from \t<br \/>\nS. \tAny \tproof  can \tbe \trepresented \tas \ta proof \ttree \twith \tdif- \t<br \/>\nferent levels \tand  with \tthe derived \tfact \tF at the \ttop. \t<br \/>\nLet \tSI \tdenote \tthe \tset \tof \tall \tclauses \tpresent \tin \tour \tex- \t<br \/>\nample-program \tP1 \tand \tin \tour \texample-database \tE,, \ti.e., \t<br \/>\nSI \t= \tPI \tU \tE,. \tIt \tis  easy \tto \tsee \tthat \tSI \tF- \tsgc(ann, \t<br \/>\nCharles). \tThe \tcorresponding \tproof-tree \tis \tdepicted \tin \t<br \/>\nFig. \t1. \t<br \/>\nNow \tthat  we \thave \ta proof-theoretic  framework \twhich \t<br \/>\nallows \tus \tto \tinfer  new \tground  facts  from  an \toriginal \tset \t<br \/>\nE. \tThe Proof \tTheory \tof \tDatalog \t<br \/>\nProof  theory \tis  concerned \twith \tthe  analysis \tof \tlogical \t<br \/>\ninference. \tAs \ta branch \tof \tmathematical \tlogic, \tthis  disci- \t<br \/>\npline  was  founded \tby \tHilbert \t[60]. \tWhile  the  investiga- \t<br \/>\ntion  of relevant \textensions \tof  first-order \tlogic \t(e.g., \tnum- \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n150 \tIEEE \tTRANSACTIONS \t<br \/>\nI \tr\u2019 \t<br \/>\nFig. \t1. Proof \ttree \tof sgc \t(ann, \tcharles). \t<br \/>\nof  Datalog \tclauses \tS, \tlet \tus \tcompare \tthis \tapproach \tto \tthe \t<br \/>\nmodel-theoretic \tapproach \tpresented \tin the \tlast \tsubsection. \t<br \/>\nThe \tfollowing  important  theorem \tholds. \t<br \/>\nSoundness  and \tCompleteness  Theorem: \tLet \tS be  a  set \t<br \/>\nof \tDatalog  clauses and  let \tF be \ta ground  fact. \tS I- \tF iff \tS E \tF. \t<br \/>\nA proof \tof \tthis theorem can \tbe \tfound \tin \t[34]. \t<br \/>\nIn \torder \tto check \twhether \tEPP \tapplies \tto a rule \tR of the \t<br \/>\nform \tLo \t: - LI, \t. . . \t, L, \tand \tto \ta (ordered) \tlist  of \tground \t<br \/>\nfacts \tF1, \t. . . , F,,, \tone \thas \tto \tfind \tan  appropriate \tsubsti- \t<br \/>\ntution \t8 such  that \tfor each \t1 I \ti I \tn Lie \t= Fi. \tSuch  a \t<br \/>\nsubstitution \tcan \tbe  found \tby \tmatching \teach \tL, \tin \tthe \trule \t<br \/>\nbody \tagainst \tthe corresponding fact \tFi. \tSuch  a  matching \t<br \/>\neither \tfails \tor \tprovides \ta \u2018\u2018local\u2019\u2019 \tsubstitution \tBi for \tall \t<br \/>\nvariables \toccurring \tin Li. \tIf at least \tone \tmatching  fails \tor \t<br \/>\nif the \tlocal  substitutions \tare \tnot \tmutually \tcompatible \tbe- \t<br \/>\ncause \tthey \tassign  different \tconstants \tto the same variable, \t<br \/>\nthen \tEPP \tdoes \tnot \tapply. \tIf all \tmatchings \tare \tsuccessful \t<br \/>\nand \tif the \tsubstitutions \tare \tall compatible, then \twe \tobtain \t<br \/>\nthe \tglobal  substitution \t8 by \tforming \tthe \tunion  of  all  local \t<br \/>\nsubstitutions \t8 = \tU \t- \tU \t8,. \t(Note \tthat \tmatching \tis \t<br \/>\na particular \tform \tof \tunification \t[34] \t.) \t<br \/>\nThere \texists  a  very \tsimple \tmethod  of \tcomputing \t<br \/>\ncons \t(S \t) for each \tfinite  set  of \tDatalog \tclauses: \t<br \/>\nFUNCTION \tINFER(S) \t<br \/>\nINPUT: \ta  finite  set \tS of \tDatalog clauses \t<br \/>\nOUTPUT: \tcons(S) \t<br \/>\nBEGIN <br \/>\nw \t:= \ts; \t<br \/>\nWHILE \tEPP \tapplies \tto \tsome \trule \tand facts \tof \tW \t<br \/>\nproducing  a \tnew \tground fact \tF $ W \t<br \/>\nDO \tW:= \tW \tU \t{F}; \t<br \/>\nRETURN( \tW \tfl \tHB) \t\/ * all facts \tof \tW, \tbut \tnot \tthe \trules \t<br \/>\n*\/ <br \/>\nEND. <br \/>\nThe INFER \talgorithm \talways \tterminates  and  produces \tas \t<br \/>\noutput \tajinite \tset of \tfacts \tcons \t(S \t). If we \tassume \tthat \tthe \t<br \/>\narities  of  all  predicates  that \tmay \toccur  in \ta Datalog \tpro- \t<br \/>\ngram \tare \tbound \tby \ta constant, then the output \tof \tthe \tIN- \t<br \/>\nFER \talgorithm \tand its runtime \tare \tboth \tpolynomial \tin the \t<br \/>\nsize  of  its  input \tS. \t<br \/>\nThe order in \twhich \tINFER \tgenerates \tnew \tfacts \tcorre- \t<br \/>\nsponds  to \tthe bottom-up order \tof proof \ttrees. For \tthis \trea- \tson, \tthe  principle  underlying \tINFER \tis called \tbottom-up \t<br \/>\nevaluation. \tIn the \tterminology  of \tartificial \tintelligence this \t<br \/>\nON  KNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL. \tI, NO. \tI. MARCH \t1989 \t<br \/>\nprinciple \tis also \treferred \tto \tas forward  chaining, \tbecause \t<br \/>\nthe \tDatalog  rules \tare \tprocessed \tforward, \ti.e., \tin the \tsense \t<br \/>\nof \tthe \tlogical  implication \tsign, \tfrom \tpremises \tto \tconclu- \t<br \/>\nsions. <br \/>\nF. \tFixpoint \tCharacterization \tof \tCons \t(S \t) \t<br \/>\nLet \tus \tshow \tthat \tthe \tset \tcons \t(S \t) can \tbe  characterized \t<br \/>\nas  the  least \tfixpoint  of  a  mapping \tTs \tfrom \t6(HB) \tto \t<br \/>\n6 (HB), \twhere \t6 (HB) \tdenotes the \tpowerset  of \tHB. \t<br \/>\nIf S is a  set  of \tDatalog clauses, \tthen \tlet \tinfer1 \t(S \t) de- \t<br \/>\nnote \tthe \tset \tof \tall \tground  facts which  can \tbe \tinferred \tin \t<br \/>\none step from the \trules  and \tfacts \tof \tS through \tEPP. \tFur- \t<br \/>\nthermore,  let \tFACTS( \tS ) denote \tthe \tset \tof \tall \tfacts \tof \tS \t<br \/>\nand \tlet \tRULES(S) \tdenote  the  set \tof \tall \trules \tof \tS. \tOb- \t<br \/>\nviously, \twe  have \tS = FACTS(S) \tU \tRULES(S). \t<br \/>\nThe \ttransformation \tTs associated \tto S is a mapping from \t<br \/>\n6 (HB) \tto \t6 (HB) \tdefined \tas  follows: \t<br \/>\nVW \tE \tHB:Ts(W) \t= \tW \tU \tFACTS(S) \t<br \/>\nU \tinferl(RULES(S) \tU \tW). \t<br \/>\nIt is easy \tto \tsee \tthat \tany Herbrand \tinterpretation \t9 G \tHB \t<br \/>\nis a Herbrand \tmodel  of \tS iff \t9 is \tajixpoint \tof \tT,, \ti.e., \t<br \/>\nTs( \t9) \t= \t9. \tIn \tparticular,  the  least  Herbrand \tmodel \t<br \/>\ncons \t( S ) is the \tleast jixpoint \tof \tTs. \t<br \/>\ncons \t(S \t) can \tbe \tcomputed \tby \tleast jixpoint \titeration, \t<br \/>\ni.e., \tby \tcomputing \tin \torder, \tTs(0), \tTs(Ts(0)), \t<br \/>\nTs(Ts(Ts(0&gt;&gt;&gt;, \t*  * \t, until \tone \tterm \tis equal  to  its pre- \t<br \/>\ndecessor. \tThis \tfinal \tterm is \tcons \t(S \t), the \tleast  fixpoint of \t<br \/>\nS. \tThis  computation \tis quite  similar \tto \tthe \tINFER \talgo- \t<br \/>\nrithm,  however,  instead \tof \tadding \tone \tnew \tfact \tat \teach \t<br \/>\niteration \tstep, \ta set \tof \tnew \tfacts \tare \tadded: \tall  those \tnew \t<br \/>\nfacts  which \tcan \tbe  inferred \tin one step from the \told \tones. \t<br \/>\nThere \texists \teven \ta transformation \tTh \twhich \tis  some- \t<br \/>\nwhat  less \tcomplicated \tthan \tTs \tand \thas \tthe  same \tleast \tfix- \t<br \/>\npoint \t<br \/>\nVW \tc \tHB:Tk(W) \t= FACTS(S) \t<br \/>\nU \tinferZ(RULES(S) \tU \tw). \t<br \/>\nThus, \tTi \tdiffers \tfrom \tTs \tby \tthe omission \tof \tthe \tterm \tWin \t<br \/>\nthe \tunion. <br \/>\nAlthough  both \tTs \tand \tTk \thave \tthe  same \tleast \tfixpoint \t<br \/>\ncons \t(S \t), they \tdo \tnot \thave  the  same \tset \tof \tfixpoints.  In \t<br \/>\nparticular, \tnot \tall  Herbrand \tmodels  of \tS are \tfixpoints \tof \t<br \/>\nTk. <br \/>\nG. \tTop-Down \tEvaluation \tof \tDatalog  Goals \t<br \/>\nThe \ttop-down \tmethod \tis \ta  radically  different \tway \tof \t<br \/>\nevaluating  Datalog \tprograms. \tProof \ttrees \tare \tconstructed \t<br \/>\nfrom \tthe \ttop \tto \tthe  bottom. \tThis \tmethod \tis particularly \t<br \/>\nappropriate \twhen  a \tgoal  is \tspecified \ttogether \twith \ta Da- \t<br \/>\ntalog \tprogram. \t<br \/>\nConsider the \tprogram \tPI \tand \tthe \tEDB \tEl \tof \tour \t\u201csame- \t<br \/>\ngeneration\u201d  example.  Assume \tthat \tthe \tgoal \t?-sgc \t(ann, \t<br \/>\nX) \tis \tspecified. \tOne \tpossibility  of \tfinding \tthe \trequired \t<br \/>\nanswers is \tto \tcompute \tfirst \tthe \tentire set \tcons \t(PI \tU \tEl \t) \t<br \/>\nby \tbottom-up  evaluation  and \tthen \tthrow  away  all  facts \t<br \/>\nwhich \tare \tnot \tsubsumed \tby \tour \tgoal.  This \twould  be  a \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERl \ter U\/.: \tDATALOG \t151 \t<br \/>\nnoticeable  waste \tof \tenergy, \tsince \twe \twould \tderive  much \t<br \/>\nmore \tfacts than  necessary. \t<br \/>\nThe \tother  possibility  is  to  start \twith \tthe  goal \tand \tcon- \t<br \/>\nstruct \tproof \ttrees \tfrom \tthe \ttop  to  the \tbottom \tby \tapplying \t<br \/>\nEPP \u201cbackwards\u201d. \tIn the context \tof \tartificial intelligence \t<br \/>\nsuch \tmethods \tare  also  referred  to  as \tbackward \tchaining. \t<br \/>\nThe \tgeneral  principle \tof \tbackward \tchaining  is  described \t<br \/>\nin \t[ 1141 \tand \tin \t[34]. \t<br \/>\nIn \tSection \tV of \tthis  paper  we  will  present \ta top-down \t<br \/>\nmethod \tfor evaluating  Datalog \tprograms \tagainst \tan EDB. \t<br \/>\nThis \tmethod, \tcalled  the \tQuery-Subquery  approach \t<br \/>\n(QSQ), \timplicitly  constructs  all  proof  trees \tfor \ta given \t<br \/>\ngoal \tin a recursive \tfashion. \t<br \/>\nAlso \tthe \twell-known  programming \tlanguage  Prolog \t<br \/>\n[45] \tis based \ton \tthe principle \tof backward \tchaining. \tHow- \t<br \/>\never,  as \twe  outline  in \tthe \tnext \tsubsection, \tthe \tsemantics \t<br \/>\nof \tProlog differs \tfrom \tthat \tof \tDatalog. \t<br \/>\nH. \tDatalog \tand \tProlog \t<br \/>\nFrom \tthe  syntactical  point \tof \tview, \tDatalog  is \ta subset \t<br \/>\nof \tProlog; \thence \teach  set \tof \tDatalog  clauses  could \tbe \t<br \/>\nparsed  and \texecuted \tby \ta Prolog  interpreter. \tHowever, \t<br \/>\nDatalog \tand \tProlog  differ \tin their \tsemantics. \t<br \/>\nWhile \tDatalog, \tas \ta simplified version of general \tLogic \t<br \/>\nProgramming \t[78], \thas \ta purely \tdeclarative semantics,  the \t<br \/>\nmeaning \tof \tProlog \tprograms \tis  defined \tby \tan  operational \t<br \/>\nsemantics, \ti.e., \tby \tthe  specification \tof  how \tProlog  pro- \t<br \/>\ngrams \tmust \tbe \texecuted. \tA Prolog \tprogram \tis processed \t<br \/>\naccording to \ta resolution  strategy \twhich \tuses \ta depth-first \t<br \/>\nsearch \tmethod \twith  backtracking \tfor \tconstructing \tproof \t<br \/>\ntrees \tand \trespects  the \torder \tof \tthe \tclauses \tand \tliterals  as \t<br \/>\nthey \tappear  in \tthe \tprogram \t[45]. \tThis \tstrategy  does \tnot \t<br \/>\nguarantee  termination. \tThe \ttermination  of \ta recursive \t<br \/>\nProlog \tprogram depends \tstrongly \ton \tthe order \tof \tthe rules \t<br \/>\nin the program, \tand \ton \tthe order \tof \tthe literals in the rules. \t<br \/>\nConsider,  for \texample, \tthe \tprogram \tPi \tconsisting \tof \t<br \/>\nthe  following  clauses: <br \/>\nr;: \tsgc(X, \tY) \t: - sgc(X1, \tYl), \tpar(X, \tXl), \tpar( \tY, \tY1). \t<br \/>\nri: \tsgc(X, \tX) \t:- \tperson(X). \t<br \/>\nNote \tthat this \tprogram \tdiffers \tfrom program \tP1 \tof \tSection \t<br \/>\n11-B \tonly \tby \tthe \torder of \tthe  rules, \tand \tof  literals \tin \tthe \t<br \/>\nrule  bodies. \tFrom  a \tDatalog  viewpoint, \tthe  order \tof \t<br \/>\nclauses \tand \tliterals \tis \ttotally \tirrelevant,  hence \tP;, \tas  a \t<br \/>\nDatalog \tprogram, \tis equivalent  to \tPI. \tOn \tthe \tother hand, \t<br \/>\nif we  submit \tP,\u2019 \tand \tthe  EDB \tEl \tto \ta Prolog  interpreter \t<br \/>\nand \tactivate \tit, \tsay, \twith \tthe goal \t\u201c?-sgc(ann, \tX)\u201d, \tthen \t<br \/>\nwe \twould  run \tinto  infinite  recursion  without  getting \tany \t<br \/>\nresult. <br \/>\nProlog  has  several \tsystem \tpredicates,  such  as \tthe \tcut, \t<br \/>\nwhich \trender  the  language \teven  more \tprocedural.  It \tis, \t<br \/>\nhowever, \ta very  rich \tand \tflexible \tprogramming \tlanguage \t<br \/>\nwhich \thas gained \tenormous \tpopularity \tover the \tlast years. \t<br \/>\nIt is possible  to  couple  Prolog  to  an  external  database. \t<br \/>\nA Prolog interpreter can then distinguish \tbetween IDB \tand \t<br \/>\nto \tretrieve \ta matching \ttuple \tfrom  mass  memory.  Due \tto \t<br \/>\nthe  procedural  semantics \tof \tProlog, \twhich \tprescribes \ta \t<br \/>\nparticular \torder \tof  visiting  goals \tand \tsubgoals, \tthe \tre- \t<br \/>\nquired interaction \tbetween \tthe \tinterpreter \tand \tthe external \t<br \/>\ndatabase  is  of  the  type \tone-tuple-at-a-time. \tThis \tmethod \t<br \/>\nof \taccessing \tmass \tstorage  data  is  quite \tinefficient \tcom- \t<br \/>\npared to \tthe \tset-oriented \tmethods \tused \tby \thigh-level query \t<br \/>\nlanguages.  Several \tenhanced \tcoupling \tmechanisms  have \t<br \/>\nbeen proposed \t[a], \t[33], \tbut \tno \tone \ttakes  full advantage \t<br \/>\nof \tset-oriented  techniques. \tThis  is \tprobably  not  possible \t<br \/>\nwithout \tcompromising \tProlog\u2019s  semantics. \t<br \/>\nIt is the \taim \tof  the \tnext \tsection \tto \tshow \tthat  Datalog \tis \t<br \/>\nwell \tsuited \tfor \tset-oriented  techniques. \t<br \/>\n111. \tDATALOG \tIs \tREALLY \tA DATABASE \tLANGUAGE \t<br \/>\nAlthough \texpressing  queries \tand \tviews \tin \tDatalog \tis \t<br \/>\nquite  intuitive \tand \tfascinating \tfrom \tthe user\u2019s  viewpoint, \t<br \/>\nwe \tshould  not  forget  that  the \taim \tof \tdatabase  query  lan- \t<br \/>\nguages \tlike Datalog \tis providing  access to large quantities \t<br \/>\nof \tdata  stored  in \tmass memory. \tThus, \tin \torder \tto  enable \t<br \/>\nan  easy  integration \tbetween \tDatalog \tand \tdatabase \tman- \t<br \/>\nagement \tsystems, \twe \tneed \tto  relate  the  logic \tprogram- \t<br \/>\nming \tformalism \tto the \tmost \tcommon \tdatabase  languages. \t<br \/>\nWe \thave chosen \trelational  algebra \tas \tsuch \ta data  retrieval \t<br \/>\nlanguage. \tThis \tsection  provides  an  informal  description \t<br \/>\nof \tthe \ttranslation \tof \tDatalog  programs \tand \tgoals  into \tre- \t<br \/>\nlational \talgebra. \t<br \/>\nA. \tTranslation \tof \tDatalog  Queries \tinto \tRelational \t<br \/>\nAlgebra <br \/>\nEach \tclause \tof \ta Datalog \tprogram \tis  translated, \tby \ta \t<br \/>\nsyntax-directed  translation  algorithm,  into  an  inclusion <br \/>\nrelationship \tof \trelational \talgebra. \tThe \tset \tof \tinclusion  re- \t<br \/>\nlationships  that  refer  to \tthe  same \tpredicate  is  then  inter- \t<br \/>\npreted \tas \tan \tequation \tof \trelational  algebra. \tThus, \twe \tsay \t<br \/>\nthat \ta Datalog \tprogram \tgives  rise to \ta system \tof \talgebraic \t<br \/>\nequations. \tEach \tIDB-predicate \tof \tthe \tDatalog \tprogram \t<br \/>\ncorresponds to \ta variable relation; \teach EDB-predicate \tof \t<br \/>\nthe  Datalog \tprogram \tcorresponds  to \ta constant  relation. \t<br \/>\nDetermining \ta solution \tof \tthe  system \tcorresponds  to  de- \t<br \/>\ntermining \tthe \tvalue \tof \tthe \tvariable  relations \twhich \tsatisfy \t<br \/>\nthe \tsystem \tof \tequations \t[61], \t[63]. \tThe \ttranslation \tfrom \t<br \/>\nDatalog \tto \trelational  algebra \tis \tdescribed \tin \t[28],  [34], \t<br \/>\nand \t[122]. <br \/>\nLet \tus \tconsider \ta Datalog  clause \tC:p \t( al, \t. . . , a,) \t: - \t<br \/>\n41 ( PI, \t. . . , Pk), \t. . . qm( \tP,, \t. . . , Oh). \tThe \ttranslation \t<br \/>\nassociates  to \tC an  inclusion  relationship \tExpr( \tQ,, \t. . . , \t<br \/>\nQ,) \tE \tP, \tamong \tthe  relations \tP, \tQl, \t. . . , Q, \tthat  cor- \t<br \/>\nrespond to predicates \tp, \tql, \t. . \t. , q,,5 \twith the convention \t<br \/>\nthat \trelation  attributes \tare \tnamed \tby \tthe \tnumber \tof \tthe \t<br \/>\ncorresponding \targument \tin \tthe \trelated  predicate. \tFor \tex- \t<br \/>\nample, \tthe  Datalog  rules  of \tprogram \tP1 \tfrom \tSection \t11: \t<br \/>\nr,: \tsgc(X, \tX) \t:- \tperson(X). \t<br \/>\nr2: \tsgc(X, \tY) \t: - par(X, \tXI), \tsgc(X1, \tYl), \tpar( \tY, \tY1) \t<br \/>\nEDB-predicates. \tWhen \tan \tEDB \tgoal  is  encountered  dur- \t<br \/>\ning the execution \tOf a Prolog \tprogram, \tthe \tinterpreter tries \t5Note \tthat \tsome \tof the \tq, might \tbe p itself, \tyielding \ta  recursive \trule \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n- <br \/>\n152 \tIEEE \tTRANSACTIONS \tON \tKNOWLEDGE \tAND \tDATA \tENCilNEtRING. \tVOL \tI, NO \tI, MARCH \t1989 \t<br \/>\nare \ttranslated  into \tthe \tinclusion \trelationships: \t<br \/>\n((PAR \tw \tSGC) \tw \tPAR) \tE \tSGC \t2=1 \t4=2 \t<br \/>\nHI, \tI PERSON \tC \tSGC \t<br \/>\nThe \trationale  of \tthe \ttranslation \tis that  literals  with \tcom- \t<br \/>\nmon \tvariables  give \trise  to \tjoins, \twhile \tthe \thead  literal \t<br \/>\ndetermines \tthe \tprojection. \tDetails  of  translation  rules \tcan \t<br \/>\nbe found \tin \t1281. \tNote \tthat, \tin \torder \tto \tobtain \ta two-col- \t<br \/>\numn \trelation \tSGC \tin the \tsecond \tinclusion \trelationship, \twe \t<br \/>\nhave  performed \ta double \tprojection \tof \tthe unique column \t<br \/>\nof \trelation \tPERSON. \t<br \/>\nFor \teach \tIDB \tpredicate \tp, \twe \tnow \tcollect  all  the  inclu- \t<br \/>\nsion  relationships  of  the \ttype \tExpr, \t(Q,, \t. . . , Qm) \tC \tP, \t<br \/>\nand \tgenerate \tan \talgebraic equation \thaving \tP as \tLHS, \tand \t<br \/>\nthe \tunion \tof  all  the \tleft-hand  sides \tof \tthe \tinclusion  rela- \t<br \/>\ntionships \tas RHS: \t<br \/>\nP \tExP~I(QI, \t. . . \t&gt; Qm) \tU \tExP~,(QI, \t. . \t. 9 em). \t. . \t<br \/>\nU \tExPrm,,(Ql, \t. . . 2 em). \t<br \/>\nFor \tinstance,  from  the  above \tinclusion  relationships  we \t<br \/>\nobtain \tthe \tfollowing equation: <br \/>\nHI,,((PAR \tW \tSGC) \tW \tPAR) \t2= \tI \t4=2 \t<br \/>\nU \tII,.l \tPERSON \t= \tSGC \t<br \/>\nNote  that \tthe \ttransformation  of  several  disequations  into \t<br \/>\none  equation \treally \tcaptures \tthe  minimality  requirement \t<br \/>\ncontained \tin the  least \tHerbrand \tmodel \tsemantics \tof \ta Da- \t<br \/>\ntalog \tprogram. \tIn \tfact, \tit expresses \tthe  fact  that  we \tare \t<br \/>\nonly \tinterested \tin \tthose  ground  facts \tthat \tare \tconse- \t<br \/>\nquences \tof \tour \tprogram. \t<br \/>\nWe \talso  translate \tlogic  goals \tinto \talgebraic  queries. \t<br \/>\nInput  Datalog \tgoals \tare \ttranslated  into  projections  and \tse- \t<br \/>\nlections \tover one \tvariable  relation  of \tthe  system \tof \talge- \t<br \/>\nbraic \tequations. \tFor \texample,  the logic \tgoal \t\u201c?-p \t(X \t). \u201d \t<br \/>\nis equivalent  to \tthe  algebraic \tquery \t\u201cP\u201d, \tand \t\u201c?-q(a, \t<br \/>\nX)\u201d. \tis equivalent \tto \t\u201ca, \t=.e\u201d. \t<br \/>\nB. \tThe \tExpressive \tPower \tof \tDatalog \t<br \/>\nThe \tsystem \tproduced \tby \tthe \tabove \ttranslation  includes \t<br \/>\nall \tthe \tclassical  relational \toperations, \twith \tthe  exception \t<br \/>\nof \tdifference;  we  say  that  it \tis \twritten \tin \tpositive \trela- \t<br \/>\ntional \talgebra, \tRA\u2019. \tIt can \tbe \teasily \tshown \tthat \teach de- \t<br \/>\nfining \texpression \tof \tRAt \tcan \talso \tbe \ttranslated  into \ta Da- \t<br \/>\ntalog  program \t[34]. \tThis means \tthat  Datalog \tis at  least  as \t<br \/>\nexpressive \tas \tRAt; \tin \tfact,  Datalog  is \tstrictly  more \tex- \t<br \/>\npressive  than \tRA+ \tbecause in \tDatalog \tit is possible  to \tex- \t<br \/>\npress  recursive \tqueries, \twhich \tare \tnot \texpressible \tin RA\u2019. \t<br \/>\nHowever,  there  are expressions \tin \tfull  relational \talge- \t<br \/>\nbra \tthat \tcannot \tbe \texpressed \tby \tDatalog programs.  These \t<br \/>\nare \tthe \tqueries \tthat \tmake \tuse \tof \tthe \tdiference \toperator. \t<br \/>\nFor example, \tgiven \ttwo \tbinary  relations \tR and \tS, \tthere \tis \t<br \/>\nno \tDatalog  rule  defining \tR \t- \tS. \tFig. \t2 graphically  rep- \t<br \/>\nresents \tthe \tsituation. \tWe \twill \tsee \tin Section \tVI \tthat \tthese \t<br \/>\nexpressions  can \tbe \tcaptured \tby \tenriching \tpure  Datalog \t<br \/>\nwith \tthe use \tof \tlogical \tnegation \t( 1 ). \t<br \/>\nNote \talso  that,  even \tthough \tDatalog \tis syntactically \ta \t<br \/>\nsubset \tof \tfirst-order \tlogic, \tstrictly \tspeaking \tthey \tare \tnot \t<br \/>\nFig. \t2. The \texpressive \tpower \tof Datalog \t<br \/>\ncomparable. \tIndeed, \tthe semantics \tof  Datalog \tis based \ton \t<br \/>\nthe choice \tof \ta specific model \t(the least Herbrand \tmodel), \t<br \/>\nwhile  first-order  logic \tdoes \tnot \ta priori \trequire \ta partic- \t<br \/>\nular  choice \tof \tthe \tmodel.  Interesting \ttractations \tof \tthe \t<br \/>\nproblem  of \tthe  expressive power \tof  relational  query  lan- \t<br \/>\nguages can \tbe \tfound \tin \t101 \tand \t[2]. \t<br \/>\nIV. \tTHE \tOPTIMIZATION \tPROBLEM: \tAN \tOVERVIEW \t<br \/>\nIn \tthis  section  we \tprovide \ta classification \tof \tefficient \t<br \/>\nmethods \tfor evaluating Datalog \tgoals. \tA systematic over- \t<br \/>\nview \tof \tmethods \tis \trequired,  since  optimization  can  be \t<br \/>\nachieved  using \ta variety  of \ttechniques, \tand understanding \t<br \/>\ntheir \trelationships \tis  not \tobvious. \t<br \/>\nWe \tclassify \toptimization methods according \tto the \tfor- \t<br \/>\nmalism, \tto  the \tsearch  strategy, \tto  the \tobjective \tof \tthe \top- \t<br \/>\ntimization, \tand \tto \tthe \ttype \tof \tconsidered \tinformation. \t<br \/>\nA. \tLogic \tand \tAlgebraic \tFormalism \t<br \/>\nIt comes \tout  from  Section \tI11 \tthat  programs \tin \tDatalog \t<br \/>\ncan \tequivalently \tbe  expressed \tas \tsystems \tof \tequations \tof \t<br \/>\nRA\u2019. \tSo, \twe \tconsider \ttwo \talternative formalisms, \tthat  we \t<br \/>\nregard,  respectively, \tas \tlogic \tand \talgebraic, \tand  we  dis- \t<br \/>\ncuss \toptimization \tmethods  that  belong  to  both \tworlds. \t<br \/>\nWe \temphasize \tthe \tconvenience \tof \tmapping \tto \tthe  al- \t<br \/>\ngebraic \tformalism, \tas \tthis \tallows \treusing  classical  results, \t<br \/>\nsuch \tas \tconjunctive \tquery \toptimization,  common  sub- \t<br \/>\nexpression \tanalysis, \tand \tthe quantitative \tdetermination \tof \t<br \/>\ncosts \tassociated \tto  each \toperation \tof \trelational \talgebra \t<br \/>\n~91, \t~341, \t[w. \t<br \/>\nB. \tSearch  Strategy <br \/>\nWe \trecall  from \tSection \tI1 that \tthe  evaluation \tof \ta Da- \t<br \/>\ntalog  goal \tcan  be  performed \tin \ttwo \tdifferent \tways: \tbot- \t<br \/>\ntom-up, \tstarting \tfrom the existing \tfacts  and  inferring \tnew \t<br \/>\nfacts, \tor \trather \ttop-down, \ttrying  to  verify  the  premises \t<br \/>\nwhich \tare \tneeded \tin \torder for the conclusion \tto \thold. \t<br \/>\nIn \tfact, \tthese two evaluation \tstrategies  represent \tdiffer- \t<br \/>\nent  interpretations  of \ta rule. \tBottom-up \tevaluations \tcon- \t<br \/>\nsider \trules \tas productions, \tthat  apply  the  initial  program \t<br \/>\nto  the \tEDB, \tand \tproduce \tall the  possible \tconsequences \tof \t<br \/>\nthe \tprogram, \tuntil \tno  new \tfact \tcan  be  deduced. \tBottom- \t<br \/>\nup \tmethods  can naturally \tbe \tapplied \tin a set-oriented  fash- \t<br \/>\nion, \ti.e., \ttaking \tas \tinput \tthe entire \trelations \tof \tthe \tEDB. \t<br \/>\nThis \tis a desirable feature \tin \tthe \tDatalog \tcontext, \twhere \t<br \/>\nlarge quantities \tof \tdata must \tbe \tretrieved  from  mass mem- \t<br \/>\nory. \tOn \tthe  other  hand,  bottom-up \tmethods \tdo \tnot \ttake \t<br \/>\nimmediate \tadvantage \tof \tthe  selectivity \tdue \tto \tthe \texis- \t<br \/>\ntence  of \targuments \tbound \tto \tconstants \tin \tthe \tgoal \tpredi- \t<br \/>\ncate. <br \/>\nIn top-down evaluation, instead, \trules \tare \tseen \tas prob- \t<br \/>\nlem \tgenerators. \tEach \tgoal \tis considered \tas \ta problem  that \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERl \tet al.: \tDATALOG \t<br \/>\nFormdtsm \t<br \/>\n153 \t<br \/>\nlogic \tvs. relational  algebra \t<br \/>\nmust \tbe \tsolved. \tThe \tinitial \tgoal is \tmatched  with \tthe \tleft- \t<br \/>\nhand \tside \tof \tsome rule, and \tgenerates \tother problems cor- \t<br \/>\nresponding \tto the \tright-hand \tside \tpredicates \tof  that \trule; \t<br \/>\nthis \tprocess \tis continued \tuntil \tno \tnew \tproblems \tare \tgen- \t<br \/>\nerated. \tIn  this \tcase, \tif the  goal  contains \tsome \tbound  ar- \t<br \/>\ngument,  then  only  facts  that \tare \tsomehow \trelated \tto  the \t<br \/>\ngoal \tconstants \tare \tinvolved \tin the computation. \tThus, \tthis \t<br \/>\nevaluation mode already performs \ta  relevant \toptimization \t<br \/>\nbecause \tthe \tcomputation  automatically  disregards \tmany \t<br \/>\nof \tthe \tfacts \twhich \tare \tnot \tuseful \tfor \tproducing \tthe \tresult. \t<br \/>\nOn \tthe \tother \thand, \tin top-down methods \tit is more \tnatural \t<br \/>\nto \tproduce \tthe  answer \tone-tuple-at-a-time, \tand this \tis an \t<br \/>\nundesirable feature in Datalog. <br \/>\nIf we \trestrict \tour \tattention \tto \tthe \ttop-down \tapproach, \t<br \/>\nwe \tcan  further distinguish \ttwo \tsearch methods: \tbreadth- \t<br \/>\nfirst \tor \tdepth-jirst. \tWith \tthe \tdepth-first \tapproach,  we \tface \t<br \/>\nthe  disadvantage that  the \torder \tof \tliterals  in  rule  bodies \t<br \/>\nstrongly \taffects \tthe \tperformance \tof \tmethods.  This  hap- \t<br \/>\npens \tin \tProlog,  where \tnot \tonly \tefficiency,  but \teven \tter- \t<br \/>\nmination \tof \tprograms is \taffected \tby \tthe left-to-right \torder \t<br \/>\nof \tsubgoals in the rule \tbodies. Instead, Datalog goals seem \t<br \/>\nmore \tnaturally \texecuted through \tbreadth-first \ttechniques, \t<br \/>\nas \tthe \tresult \tof \tthe computation \tis neither \taffected \tby \tthe \t<br \/>\norder \tof \tpredicates \twithin \tthe \tright-hand \tsides \t(RHS) \tof \t<br \/>\nrules,  nor \tby \tthe \torder \tof \trules \twithin \tthe program. \t<br \/>\nC. \tObjectives \tof \tOptimization Methods \t<br \/>\nOur \tthird \tclassification \tcriterion is \tbased \ton the \tdiffer- \t<br \/>\nent \tobjectives \tof \toptimization \tmethods: \tsome  methods \t<br \/>\nperform \tprogram transformation, \tnamely, transforming \ta \t<br \/>\nprogram  into \tanother \tprogram  which  is written  in \tthe \tsame \t<br \/>\nformalism,  but  yields \ta more \tefficient \tcomputation \twhen \t<br \/>\none \tapplies \tan \tevaluation \tmethod \tto  it; we \trefer \tto \tthese \t<br \/>\nas \trewriting \tmethods. \tGiven \ta goal \tG and \ta  program \tP, \t<br \/>\nthe \trewritten  program \tP\u2019 \tis equivalent \tto \tP with  respect \t<br \/>\nto \tG, \tas \tit produces  the \tsame \tresult.  Formally, \trecalling \t<br \/>\nthe \tdefinition  of \tXp, \tof \tSection \t11, \ttwo \tprograms \tP and \t<br \/>\nP\u2018 \tare \tequivalent \twith  respect  to  a \tgoal \tG \tiff \t3np,G \t= \t<br \/>\nThese \tmethods \tcontrast with the \tpure \tevaluation  meth- \t<br \/>\nods, \twhich \tpropose \teffective \tevaluation strategies, where \t<br \/>\nthe optimization is performed during the evaluation itself. <br \/>\nD. \tType \tof \tConsidered \tInformation \t<br \/>\nOptimization methods \tdiffer \tin the \ttype \tof \tinformation \t<br \/>\nused \tin the optimization. \t<br \/>\nSyntactic optimization \tis the \tmost  widely \tused; \tit deals \t<br \/>\nwith \tthose  transformations \tto  a \tprogram  which  descend \t<br \/>\nfrom \tthe  program\u2019s  syntactic  features. \tIn \tparticular, \twe \t<br \/>\ndistinguish \ttwo \tkinds  of \tstructural properties. \tOne \tis the \t<br \/>\nanalysis \tof \tthe \tprogram  structure, \tand  in  particular  the \t<br \/>\ntype \tof  rules  which \tconstitute \tthe \tprogram. \tFor \texample, \t<br \/>\nsome \tmethods \texploit the \tlinearity \t(see \tSection \tV) \tof \tthe \t<br \/>\nrules \tto \tproduce optimized  forms \tof \tevaluation. \tThe \tsec- \t<br \/>\nond \tone \tis the \tstructure \tof \tthe \tgoal, \tand \tin particular the \t<br \/>\nselectivity \tthat comes from goal constants. \tThese two \tap- \t<br \/>\nproaches \tare \tnot mutually \texclusive: \tit is possible \tto build \t<br \/>\nsyntactic \tmethods  which \tcombine \tboth \tcases. \t<br \/>\nSemantic \toptimization, \tinstead, concerns \tthe \tuse \tof \tad- \t<br \/>\nditional \tsemantic knowledge  about the  database \tin \torder \t<br \/>\nXP\u2019,G. \t<br \/>\n7h\u201dsal \torder \t<br \/>\nObjective Approach <br \/>\nStructure <br \/>\nCRITERION \t1 \tALTERNATIVES \tI \t<br \/>\ndepth-first \tvs. breadth-first \t<br \/>\nrewriting \tvs. pure evaluation \tsyntactic \tvs. semantic \t<br \/>\nrule structure \tvs. coal  structure \t<br \/>\nBOTTOM-UP <br \/>\n1611. \t1151 \tSemi-naive \t[IS] \tHenschen~Naqvi \t1591 \t<br \/>\nLOGIC \tALGEBRAIC <br \/>\nConstant  reduction \tMagic \tCounling \t1106) \tStahc \tfiltering \t168) <br \/>\n(b) \tFig. \t3. Classification \tof evaluation \tand \toptimization \tmethods. \t<br \/>\nto produce \tan \tefficient \tanswer to \ta query; the combination \t<br \/>\nof \tthe  query \twith \tadditional semantic information is per- \t<br \/>\nformed automatically.  Semantic methods are often \tbased \t<br \/>\non \tintegrity \tconstraints, \twhich \texpress \tproperties \tof \tvalid \t<br \/>\ndatabases. \tFor \tinstance, \ta constraint might state \tthat \t\u201call \t<br \/>\nintercontinental \tjights \tdirected \tto \tMilan \tland \tin \tMal- \t<br \/>\npensa  airport\u201d. \tThis \tconstraint  can  be  used \tto \tproduce \t<br \/>\nthe answer \tof a \tgoal asking \tfor \t\u201cthe arrival airport \tof \tthe \t<br \/>\nintercontinental \tjight \tA2747 \tfrom \tNew \tYork \tto \tMilan \t\u201d \t<br \/>\nwithout \taccessing  the \tEDB. \tAlthough \twe \tthink \tthat \tse- \t<br \/>\nmantic  optimization  has  the \tpotential \tfor \tsignificant \tim- \t<br \/>\nprovements \tof  query \tprocessing strategies, \twe \tdo \tnot  fur- \t<br \/>\nther \tconsider semantic \toptimization  methods  in this paper: \t<br \/>\nvarious \tapproaches \tto semantic optimization \tare \tproposed \t<br \/>\nin the literature; among others; \tsee \t[70], \t[36]. \t<br \/>\nE. \tClassijication \tof \tEvaluation \tand \tOptimization \t<br \/>\nMethods <br \/>\nFig. \t3(a) summarizes \tclassification \tcriteria \tfor \tDatalog \t<br \/>\noptimization  methods; \tthey \tconcern \tthe \tsearch  strategy \t<br \/>\n(bottom-up \tor \ttop-down), \tthe \tobjective \t(rewriting \tor \tpure \t<br \/>\nevaluation),  and  the \tformalism \t(logic \tor \talgebraic). \tBy \t<br \/>\ncombining  approaches  and  excluding  some  alternatives <br \/>\nthat \tdo \tnot \tcorrespond \tto \trelevant \tclasses \tof \tmethods,  we \t<br \/>\nobtain \tfour \tclasses,  each  including  rather  homogeneous \t<br \/>\nmethods: <br \/>\n1) \ttop-down evaluation methods \t<br \/>\n2) \tbottom-up evaluation methods \t<br \/>\n3) \tlogic \trewriting \tmethods \t<br \/>\n4) \talgebraic \trewriting \tmethods. \t<br \/>\nFig. \t3(b)  shows  some \tof \tthe  most  well-known \tevalua- \t<br \/>\ntion  and \toptimization  methods  present \tin \tthe  literature. \t<br \/>\nThis \tlist \tis not \texhaustive for obvious \tbrevity \treasons; we \t<br \/>\napologize \tto \tauthors  whose \tmethods \tare \tomitted. \tIn  the \t<br \/>\nnext  section \twe \tsurvey \tinformally \tsome \tof \tthese methods. \t<br \/>\nFinally, \twe \texpect  optimization  methods \tto \tsatisfy \tthree \t<br \/>\nimportant \tproperties. \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nON \tKNOWLEDGE \tANI) \tDAIA \tENGINEERING. \tVOI.. \tI. VO. \tI. hlAKCH \t1089 \t<br \/>\nMethods \tmust \tbe \tsound: \tthey \tshould \tnot \tinclude \tin \t<br \/>\nthe  result \ttuples \twhich \tdo \tnot \tbelong \tto \tit. \t<br \/>\nMethods \tmust \tbe \tcomplete: \tthey  must \tproduce \tall  the \t<br \/>\ntuples \tof \tthe  result. \t<br \/>\nMethods \tmust \tterminate: \tthe \tcomputation \tshould  be \t<br \/>\nperformed \tin \tfinite \ttime. \t<br \/>\nAlthough  we \tomit  to  present  formal \tproofs, \tall \tthe \t<br \/>\nmethods  surveyed \tin  the \tnext  section  satisfy \tall \tthese \t<br \/>\nproperties. \t<br \/>\nv. \tSURVEY \tOF \tEXECUTION \tMETHODS \tAND \t<br \/>\nOPI.IMI%ATION \tTECHNIQIJES \t<br \/>\nIn \tthis \tsection \twe \tpresent  methods  for \toptimizing \ta Da- \t<br \/>\ntalog \tprogram, \ti.e., \tfor \tgenerating \tefficiently  the \tactual \t<br \/>\nset \tof \ttuples  which  satisfy \ta given  goal \tfor \ta given  set \tof \t<br \/>\nDatalog  rules. \tIn \torder \tto \tdo this, \twe \tfirst \tpresent \ta bot- \t<br \/>\ntom-up \tnaive \tevaluation  method  (called \tthe \tGauss-Seidel \t<br \/>\nmethod), \tand  discuss \tbriefly  its \tsources \tof \tinefficiency. \t<br \/>\nThen, \twe \texamine \thow \tto  reduce \tthese \tsources  of \tineffi- \t<br \/>\nciency: \twe \tintroduce \tthe \tSemi-Naive \tapproach \twhich \tim- \t<br \/>\nproves \ta bottom-up \tcomputation \tof \tlinear \trules, \tand \twe \t<br \/>\npresent \ta \ttop-down \tefficient \tstrategy, \tcalled \tQuery- \t<br \/>\nSubquery. \tThen, \twe \tpresent  two  of  the \tmost  known  re- \t<br \/>\nwriting \tmethods, \tMagic \tSets \tand \tCounting, \twhich \tare \t<br \/>\nused \tto optimize \tthe \tbehavior  of \tbottom-up computations. \t<br \/>\nFinally, \twe  briefly \toverview \tother \toptimization \tmethods. \t<br \/>\nA. \tBottom- \tUp \tEiuluation \t<br \/>\nThe \tGauss-Seidel  method \tis an \talgebraic \tversion  of \tthe \t<br \/>\nnni~be \triwfuLition \tparadigm, \twhich \tis present \tin  many \tdif- \t<br \/>\nferent  forms \tin \tthe \tliterature \t1131, \t1851, \t[41], \t[86], \t1871. \t<br \/>\nThe \tmethod \tis \talso \twell \tknown \tin \tnumerical \tanalysis, \t<br \/>\nwhere \tit is used \tfor determining \tthe  iterative  solution \t(fix- \t<br \/>\npoint) \tof \tsystems \tof \tequations. Assume \tthe following  sys- \t<br \/>\ntem \tC of \trelational \tequations: \t<br \/>\nR, \t= E,(R,, \t. . \t. , R,), \t(i \t= \t1, \t. . . \t, n). \t<br \/>\nThe  Gauss-Seidel  method  proceeds  as \tfollows: \tinitially, \t<br \/>\nthe  variable  relations \tRi \tare  set  equal \tto  the \tempty \tset. \t<br \/>\nThen, \tthe  computation \tRi \t:= \t,?;(RI, \t. . . \t, R,,), \t(i \t= \t1, \t<br \/>\n. .  . \t, n) \tis iterated \tuntil  all  the \tRi \tdo \tnot \tchange  between \t<br \/>\ntwo \tconsecutive  iterations \t(namely, \tuntil \tthe \tRi \thave \t<br \/>\nreached \ta jxpoint). \tAt \tthe  end \tof \tthe \tcomputation, \tthe \t<br \/>\nvalue \tassumed \tby \tthe  variable  relations \tR, \tis the  solution \t<br \/>\nof \tthe \tsystem \tE. \t<br \/>\nGAUSS-SEIDEL \tMETHOD \t<br \/>\nINPUT: <br \/>\nExtensional \tDatabase \tEDB. \t<br \/>\nOUTPUT: <br \/>\n... \t,R,,. \t<br \/>\nMETHOD: <br \/>\nFORI':== \tlTOnDORi:= \tD; \t<br \/>\nREPEAT \t<br \/>\nA \tsystem \tof \talgebraic  equations \tE, \tand  an \t<br \/>\nThe \tvalues  of  the  variable  relations \tRI, \t<br \/>\ncond \t: = true; \t<br \/>\nFori \t:= \t1 TO \tn DO \t<br \/>\nBEGIN <br \/>\nS := \tRI; \t<br \/>\nRi \t:= \tEi(R,, \t. . \t. , Rn); \t<br \/>\nIF \tR, \t# \tS THEN \tcond \t: = false; \t<br \/>\nEND; \t<br \/>\nUNTIL \tcond; \t<br \/>\nFOR \ti : = \t1 TO \tn DO \tOUTPUT(R,). \t<br \/>\nENDMETHOD <br \/>\nNote \tthat \tstep \t\"Ri \t: = E;( \tRI, \t. . \t. , R,)\" \tof \tthe \tGauss- \t<br \/>\nSeidel  algorithm  has \tthe same \teffect \tas \tthe  application  of \t<br \/>\nrule \tEPP \tof \tSection \t11. \tHowever, \tinstead \tof \tacting \ton \tsin- \t<br \/>\ngle  tuples, \there  we  apply \talgebraic \toperations \tsimultu- \t<br \/>\nneously \tto \tentire \trelations. \t<br \/>\nVariants \tof \tthe  Gauss-Seidel  method \tare \tthe \tJcicobi \t<br \/>\nmethod \tand \tthe \tChaotic \tmethod \t1281. \tThe \tlatter \tis ob- \t<br \/>\ntained \tby \tcomputing \tthe  various \talgebraic \texpressions \tnot \t<br \/>\nin \ta strict  sequential \torder. \tDifferent  versions  of \tthe cha- \t<br \/>\notic  method  yield  the  so-called \tlazy \tand \tdata$ow \tevalu- \t<br \/>\nations, that \tcorrespond \tto \tstarting \tthe \tevaluation \tof \tcom- \t<br \/>\nputable \trelations,  respectively, \tat \tthe  latest \tor \tat  the \t<br \/>\nearliest \tconvenience. \t<br \/>\nThe \tGauss-Seidel  method  has  two \tsources \tof  ineffi- \t<br \/>\nciency. \t<br \/>\na) \tSeveral  tuples \tare \tcomputed  multiple  times  during \t<br \/>\nthe \titeration \tprocess. \tIn \tparticular, \tduring  the  iterative \t<br \/>\nevaluation \tof \ta relation \tR, \ttuples  belonging  to  relations \t<br \/>\nR\") \twill \talso \tbelong  to  all \tsubsequent \trelations \tR\" \t), j 2 \t<br \/>\ni, until \tthe  fixpoint \tis reached. \t<br \/>\nb)  The \tmethod  produces  the \tentire \tresult  relations. \tIf \t<br \/>\nthe  goal  contains  constant \targuments, \tthey \tare \tselected \t<br \/>\nonly  at \tthe \tend. \tIn \tthis \tway, \tseveral  tuples \tare \tcomputed \t<br \/>\nwithout  being  really \trequired, \tand \teliminated \tby \tthe \tfinal \t<br \/>\nselection. <br \/>\nInefficiencies \tdue \tto \tobservation \ta) \tabove are \tpartially \t<br \/>\neliminated  through \tthe \tsemi-naive \tmethods  discussed \tin \t<br \/>\nthe  remainder \tof \tthis subsection; \tinefficiencies \tdue \tto ob- \t<br \/>\nservation \tb) \tabove  are \tdealt \twith \tby \tevaluation  methods \t<br \/>\n(like \tquerysubquery) \tand  rewriting  methods \t(like \tmagic \t<br \/>\nsets, \tcounting, \tetc.), \tdiscussed \tin \tthe \tnext \tsubsections. \tSemi-naive  evaluation \tis \ta bottom-up \ttechnique \tde- \t<br \/>\nsigned \tfor \teliminating  redundancy \tin the \tevaluation \tof \ttu- \t<br \/>\nples \tat different \titerations. \tSeveral versions \tof this \tmethod \t<br \/>\ncan \talso \tbe  found \tin \tthe \tliterature [13], \t[28], \t1201. \t<br \/>\nConsider \tthe \tGauss-Seidel \talgorithm, \tapplied \tto \tsolve \t<br \/>\na single \tequation  on  variable \tR. \tLet \tR'k' \tbe \tthe temporary \t<br \/>\nvalue \tof \trelation \tR at  iteration \tstep \tk. \tThe \tdifferential \tof \t<br \/>\nR at step \tk of \tthe  iteration \tis defined \tas \t<br \/>\nD'k) \t= R(k) \t- R(k-1). \t<br \/>\nThe \tdifferential \tterm,  expressing \tthe \tnew \ttuples  of \tR(kJ \tat \t<br \/>\neach  iteration \tk, \tis exactly \twhat \twe  would  like \tto  use \tat \t<br \/>\neach \titeration, \tinstead  of  the \tentire \trelation \tR'\";; \tthis \tis \t<br \/>\nlegal \twith \tlinear  eqiiations. \tAn \tequation \tof \trelational \tal- \t<br \/>\ngebra \t<br \/>\nR \t= E(R) \t<br \/>\nis linear \twith \trespect \tto R if it verifies  the \tfollowing  prop- \t<br \/>\nerty \t: \t<br \/>\nE(R' \tU \tR\") \t= E(R') \tU \tE(R\") \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERI \tet al.: \tDATALOG \t155 \t<br \/>\nfor  any \ttwo \trelations \tR\u2018 \tand \tRff \thaving \tthe same \tarity \tas \t<br \/>\nR. \tNote \tthat  linearity  is \tensured \tif only \tone \toccurrence \tof \t<br \/>\nR appears \tin \tthe  expression \tE(R). \tThe \tdifferential \tof \tR \t<br \/>\nat step \tk of \tthe \titeration  is  simply \tE(D( \tR)): \t<br \/>\nE(D(~)) \t= E(R\u2018~) \t- ~(k-1)) \t= E(R(~)) \t- \t~(~(k-1) \t) \t<br \/>\nso \tthat \tone \tcan compute \tR(k+l) \t= \tE(R(k\u2019) \tas \tthe \tunion \t<br \/>\nE(R\u2018k-I)) \tU \tE(D\u2018k\u2019). \t<br \/>\nThe \tadvantage \tof \tthis formula \tis that, \tat each \titeration \tk, \t<br \/>\nwe \tneed \tto \tcompute \tE(D\u2018k\u2019) \trather \tthan \tE(R(k\u2019). \tThis \t<br \/>\ncan  be  generalized \tto \tsystems \tof \tequations  (see  [29], \t<br \/>\nExtensions  of \tthis \tmethod, \tthe \tgeneral \tsemi-naive, \tand \t<br \/>\nthe \tpseudo-rewriting  semi-naive \t(discussed,  among  oth- \t<br \/>\ners, \tin \t[34], [135], \tand  [63]) enable \ta less \tefficient \tappli- \t<br \/>\ncation  of  a \tsimilar approach \tto \tnonlinear equations. \t<br \/>\n~31). <br \/>\nB. \tTop-Down  Evaluation <br \/>\nThe \tquery-subquery \t(QSQ) \talgorithm  [129] \tis an \teffi- \t<br \/>\ncient  top-down \tevaluation  algorithm,  optimizing \tthe \tbe- \t<br \/>\nhavior \tof \tbackward-chaining methods \tas described \tin Sec- \t<br \/>\ntion \t11. <br \/>\nThe \tobjective \tof \tthe \tQSQ \tmethod  is \tto \taccess the \tmin- \t<br \/>\nimum \tnumber \tof  facts  needed  in \torder \tto \tdetermine  the \t<br \/>\nanswer. \tIn \torder \tto \tdo \tthis, \tthe  fundamental \tnotion \tof \t<br \/>\nsubquery \tis introduced. \tA goal, \ttogether \twith  a \tprogram, \t<br \/>\ndetermines \ta query. \tLiterals \tin the \tbody  of  any \tone \tof \tthe \t<br \/>\nrules  defining \tthe goal predicate \tare \tsubgoals \tof \tthe given \t<br \/>\ngoal. \tThus, \ta subgoal,  together \twith \tthe  program, \tyields \t<br \/>\na subquery; \tthis \tdefinition \tapplies \trecursively \tto \tsubgoals \t<br \/>\nof  rules  which \tare \tsubsequently activated. \tIn \torder to \tan- \t<br \/>\nswer the query, each goal \tis expanded \tin a list of \tsubgoals, \t<br \/>\nwhich \tare \trecursively \texpanded \tin \ttheir turn. \t<br \/>\nThe \tmethod \tmaintains \ttwo \tsets: \ta  set \tP of \tanswer  tu- \t<br \/>\nples, \tcontaining \tanswers \tto \tthe \tmain \tgoal and answers \tto \t<br \/>\nintermediate subqueries, \twhich  is  represented \tby \ta  set \tof \t<br \/>\ntemporary  relations \t(one \trelation \tfor \teach  IDB-predi- \t<br \/>\ncate);  and \ta  set \tQ of \tcurrent  subqueries \t(or \tsubquery \tin- \t<br \/>\nstances), \twhich \tcontains \tall \tthe  subgoals \tthat \tare \tcur- \t<br \/>\nrently \tunder consideration. \t<br \/>\nThus, \tthe \tfunction \tof \tthe  QSQ  algorithm  is \ttwofold: \t<br \/>\ngenerating  new \tanswers \tand \tgenerating \tnew \tsubqueries \t<br \/>\nthat must \tbe answered. \tThere \tare \ttwo \tversions \tof the \tQSQ \t<br \/>\nalgorithm, \tan \titerative \tone \t(QSQI)  and \ta recursive \tone \t<br \/>\n(QSQR). \tThe \tdifference \tbetween the \ttwo \tconcerns \twhich \t<br \/>\nof \tthese \ttwo \tfunctions has \tpriority \tover \tthe  other:  QSQI \t<br \/>\nprivileges \tthe \tproduction \tof \tanswers,  thus, \twhen  a  new \t<br \/>\nsubquery  is \tencountered, \tit  is \tsuspended \tuntil \tthe  end \tof \t<br \/>\nthe \tproduction \tof \tall \tthe  possible  answers that \tdo \tnot \tre- \t<br \/>\nquire \tusing \tthe \tnew \tsubquery. \tQSQR \tbehaves \tin \tthe \top- \t<br \/>\nposite  way: \twhenever \ta  new  subquery \tis found, \tit  is \tre- \t<br \/>\ncursively \texpanded  and  the  answering \tto \tthe  current \t<br \/>\nsubquery \tis postponed \tto when the \tnew  subquery \thas been \t<br \/>\ncompletely \tsolved. \t<br \/>\nAt \tthe  end \tof \tthe  computation, \tP includes  the  answer \t<br \/>\nto \tthe  goal;  hence, \tas \tin  the \tGauss-Seidel \tmethod, \tit  is \t<br \/>\nrequired \tto perform (una tantum) the \tfinal \tselection. How- \t<br \/>\never, \tthe \treader \tshould  notice \tthat \tthe \tQSQ \tmethod  uses \t<br \/>\nthe  information \tabout  constants  in  the \tgoal, \thence, \tthe \t<br \/>\nsize \tof \tP and \tof \tall  the  relations involved \tin \tthe  compu- \t<br \/>\ntations  is \tcomparatively \tmuch \tsmaller than \tthe \tsize \tof  all \t<br \/>\nthe \trelations  involved  in \tthe \tGauss-Seidel \tcomputation. \t<br \/>\nThis algorithm \tcan be \tcompared \tto \tthe  Prolog inferen- \t<br \/>\ntial \tmachine, \twhich \tis also top-down. \tThe \tcomparison \tis \t<br \/>\npurely \tindicative, based \ton \tthe  idea \tof  using  a \tProlog \tin- \t<br \/>\nferential \tmachine \tto \texecute \ta Datalog program. \tWe \tnote \t<br \/>\nthat \tProlog \tacts \tone-tuple-at-a-time, while \tQSQ \tis set-ori- \t<br \/>\nented, \tas \tit processes whole relations. \tIn  this \tsense, QSQ \t<br \/>\nis \tappropriate \tfor \tdatabase  processing. \tAlso, \tQSQ \tis \t<br \/>\nbreadth-first \tand always terminates, while \tProlog \tis depth- \t<br \/>\nfirst \tand \tmay  instead  not \tterminate \tin \tsome cases. \t<br \/>\nThe \tquery-subquery \talgorithm \twas  introduced \tby \t<br \/>\nVieille  in \t[ 1281. \tThe \tversion \tintroduced \tin that \tpaper \twas \t<br \/>\nfound \tto be \tincomplete. \tThe \tauthor \t[129], \t[130] and oth- \t<br \/>\ners \t[93], \t[ 1001 \thave subsequently \tprovided \tcorrections \tor \t<br \/>\ncomplete versions \tof \tthe \talgorithm. \tA detailed \tdescription \t<br \/>\nof \tthe \tmethod \tcan \tbe \tfound \tin \t[34]. \t<br \/>\nC. \tMagic  Sets <br \/>\nThe \tmethod  of  the \tmagic \tsets \tis \ta logical \trewriting \t<br \/>\nmethod  which \ttransforms \ta program \tinto  a \tlarger \tone, \t<br \/>\ncontaining  some  more  rules  that \tdefine  new \tIDB \tpredi- \t<br \/>\ncates.  These \tIDB \tpredicates  serve \tas \tconstraints, \twhich \t<br \/>\nforce  the  program  variables \tto \tsatisfy \tsome \tadditional \t<br \/>\nconditions. \tThus, \tduring bottom-up computation, \tthe vari- \t<br \/>\nables \tof \tthe \tmodified \trules \tmay \tassume only \tsome \tof \tthe \t<br \/>\nvalues \tthat were \tinstead \tallowed \tfor \tvariables \tof \tthe \torig- \t<br \/>\ninal \trules. \tIn  most \tcases, \tthis  makes  the \tnew  program \t<br \/>\nmore  efficient. <br \/>\nDetails about \tthis \talgorithm can \tbe \tfound in \t[ 141, \t[ 171, \t<br \/>\n[22], [34]. \tWe \tshow \there \ta significant \texample, \tand \tmake \t<br \/>\nsome comments on \tthe \trewritten program.  Consider again \t<br \/>\nthe  program \tPI \twith \tthe \tgoal \t?-sgc(ann, \tX). \tAfter \tthe \t<br \/>\nmagic \tsets transformation,  the program becomes: <br \/>\nr;: \tsgc(X, \tX) \t: - person(X). \t<br \/>\nr;: \tsgc(X, \tY) \t:- \tmugic(Xl), par(X, \tXl), \t<br \/>\nsgc(X1, \tYl), \tpar(Y, \tYl). \t<br \/>\nr; : magic \t(ann \t). \t<br \/>\nr4: \tmagic(X1) \t: - magic(X), \tpar(X, \tXl). \t<br \/>\nLet \tus \tfirst \tnotice \twhat  has  happened \tto  the \tinitial  pro- \t<br \/>\ngram:  two \tnew \trules \thave  been \tadded, \tand \tone \trule \thas \t<br \/>\nbeen  modified. \tLet \tus \tconsider this \trule \t(r;). \tA  new \tlit- \t<br \/>\neral \thas been added \tto its \tbody: \tmagic \t(XI \t). The \tpresence \t<br \/>\nof \tthis \tnew \tliteral  forces  the  argument \tX1 \tof \tthe \tEDB \t<br \/>\nrelation \tPAR \tto \tassume  only  some \tspecific \tvalues, \ti.e., \t<br \/>\nvalues \twhich \talso belong \tto the IDB \trelation \tMAGIC. \t<br \/>\nLet \tus \tnow \tsee \thow  this  new  relation  is \tdefined. \tThe \t<br \/>\nfirst \trule \t(ri \t) simply  says \tthat \tthe  constant \tann \tbelongs \t<br \/>\nto \tit.  Rule \tri \tsays  that, \tif X1 \tis  in \tMAGIC, \tand \tX is  a \t<br \/>\nparent  of \tX1, \tthen \talso \tX is in MAGIC. \tThe \tresult \tof \tthe \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n156 \tIEEE \tTRANSACTIONS \tON \tKNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL. \tI, NO. \tI. MARCH \t1989 \t<br \/>\ncomputation \tof \trelation \tMAGIC \tis \t<br \/>\nMAGIC \t= \t{ &lt; ann \t&gt; , &lt; dorothy \t&gt; , \t<br \/>\n&lt; hilary \t&gt; , &lt; george \t&gt; }. \t<br \/>\nThus, \tthe \tIDB \trelation \tMAGIC \tcontains \tall  the ancestors \t<br \/>\nofann. \tThe \ttuples \tof \tthe \trelation \tMAGIC \tdefined \tby \tthe \t<br \/>\nmagic  rules  form \tthe \tmagic \tset. \tIt is often \tcalled \tthe \tcone \t<br \/>\nof \tann, \treferring \tto  the \tfact \tthat, \tstarting  from \tann, \tthe \t<br \/>\nancestors \tgrow \t\u201cfanning \tout\u201d \tlike \ta cone. \tBy \timposing \t<br \/>\nthat, \tin \tr;, \tXI \tmust  belong \tto \trelation \tMAGIC, \twe  are \t<br \/>\nimposing \tthat, \tin the \tcomputation \tof \tAnn\u2019s cousins \tat the \t<br \/>\nsame \tgeneration, \twe \tonly \thave \tto \tconsider  those  other \t<br \/>\npairs \tof \tsame \tgeneration \tcousins \twhose \tfirst \telement \tis an \t<br \/>\nancestor \tof \tAnn. \t<br \/>\nThis  example \tcan \tbe \tused \tto \tintroduce  the \tidea  of \t<br \/>\nsideways \tinformation  passing \t(SIP) \t[22]. \tIntuitively, \t<br \/>\ngiven  a \tcertain \trule  and  a  literal \tin \tthe \trule \tbody \twith \t<br \/>\nsome \tbound \targument(s), \tone \tcan  use  the  knowledge \t<br \/>\nabout  the  relation \tcorresponding \tto  this  literal  to \tobtain \t<br \/>\nbindings \tfor \tuninstantiated \tvariables \tin \tother  argument \t<br \/>\npositions. \tThis \tprocess \tcan  be \titerated \tfor \teach \tliteral \tin \t<br \/>\nthe \trule \tbody, \tand  recursively \ton  other  IDB-predicates. \t<br \/>\nThus, \tknown information  (bindings to constants) \tis passed \t<br \/>\nsideways \twithin \tthe \trule \tbody. \tAs we \thave \tseen \tin Section \t<br \/>\nV-B,  this \tis the  normal \tbehavior \tof  top-down \tevaluation \t<br \/>\nmethods,  for \tinstance \tQSQ. \t<br \/>\nAfter \tthe  magic \tsets transformation,  the \tresulting  pro- \t<br \/>\ngram  can \tbe \tevaluated \tby  a \tsimple algorithm like \tGauss- \t<br \/>\nSeidel \tor \tsemi-naive, \tstill taking \tadvantage \tof \tthe \tbinding \t<br \/>\npassing \tstrategy.  However,  the  application \tof  a  bottom- \t<br \/>\nup \tcomputation \tmethod \tto the \trewritten  program  may pro- \t<br \/>\nduce  more  tuples \tthan  exactly \tthose \tof \tthe  goal  answer, \t<br \/>\nas \tit includes  all  the \tsame \tgeneration \tpairs  of  all \tpeople \t<br \/>\nbelonging \tto \tthe  cone \tof \tann. \tThus, \tas \tin \tQSQ, \tthe \tre- \t<br \/>\nsulting  relation  must \tbe \tfinally \tselected \tto \tobtain the \tan- \t<br \/>\nswer. <br \/>\nThe \tmagic \tsets \tmethod \thas \tbeen \textended \tby \tSacca\u2019 \t<br \/>\nand \tZaniolo \tto \ta class \tof \tqueries \tto \tlogic \tprograms  that \t<br \/>\ncontain  function \tsymbols \t[ 1051. \tA \tmore \tsophisticated \t<br \/>\ntechnique  called \t\u201cSupplementary  Magic \tSets\u201d, \thas \tbeen \t<br \/>\nintroduced \tby \tBeeri  and \tothers \t[22], \twhose \tvirtue  is \tto \t<br \/>\neliminate some \trepeated \tcomputations. Another \timprove- \t<br \/>\nment  of \tthe  magic \tset \tmethod \tis proposed \tin \t[99], \twhere \t<br \/>\ntop-down \tevaluation \tis completely \tmimicked \tby \tbottom- \t<br \/>\nup, \tusing  a \tsemi-naive evaluation \tof  rewritten \trules;  the \t<br \/>\nrule  rewriting  uses  initial \tgoal \tbindings \tin \ta  very  sophis- \t<br \/>\nticated \tway; \tfor \tinstance, goals \tlike \tp (X, \tX), \tgive \trise \tto \t<br \/>\nthe  binding \tof \tthe \ttwo \toccurrences \tof \tvariables \tX to \teach \t<br \/>\nother.  This \twas \tnot \tcovered \tby \tthe \toriginal  magic \tsets \t<br \/>\nmethod. <br \/>\nD. \tCounting <br \/>\nThe \tCounting \tmethod \tis a  rewriting  method  based \ton \t<br \/>\nthe  knowledge  of  the  goal \tbindings;  the \tmethod  includes \t<br \/>\nthe \tcomputation \tof \tthe \tmagic \tset, \tbut \teach element \tof  the \t<br \/>\nmagic  set \tis complemented \tby \tadditional \tinformation \tex- \t<br \/>\npressing  its \t\u201cdistance\u201d  from  the \tgoal \tconstant. \tThe \t<br \/>\ncounting \tmethod  was \talso \tfirst \tpresented \tin \t[14]; \tan \tim- \t<br \/>\nproved \tversion was \tintroduced \tin \t[22]. \t<br \/>\nConsider again  the \tprogram \tPI \tused \tas  an  example \tof \t<br \/>\nthe \tmagic  set \tmethod \tin \tthe \tprevious \tsubsection. \tThe \t<br \/>\nmagic \tset \tmethod  restricts \tthe  computation \tto \tthe  ances- \t<br \/>\ntors  of \tann; \tfor \teach \tof \tthese  elements,  the  counting \t<br \/>\nmethod  maintains \tthe \tinformation \twhether \tit is \tone \tof \t<br \/>\nann\u2019s \tparents \t(distance \tl), \tann\u2019s \tgrandparents  (distance \t<br \/>\n2), \tann\u2019s \tgrand-grandparents  (distance \t3), \tetc.  The \tre- \t<br \/>\nwritten \tprogram  contains  the  computation \tof \tthese \tdis- \t<br \/>\ntances. \tAt \tthis  point,  computation \tmay \tbe  restricted, \tre- \t<br \/>\nspectively, \tto \tthe  children \tof \tann\u2019s \tparents, \tto \tthe \t<br \/>\ngrandchildren  of \tann\u2019s \tgrandparents, \tto \tthe grand-grand- \t<br \/>\nchildren \tof \tann\u2019s \tgrand-grandparents, \tetc. \t<br \/>\nThe \tfollowing \tis  the \tresult  of \tapplying \tthe \tcounting \t<br \/>\ntransformation \tto \tthe output \tof \tthe \tmagic  sets \tmethod: \t<br \/>\nsgc\u2019(X, \tX, \tI) \t:- \tperson(X), \tinteger(1). \t<br \/>\nsgc\u2018(X, \tY, \tI) \t:- \tcounting \t<br \/>\nsgc \t\u2019(Xl \t, Y1 \t<br \/>\nI=J- \t1. \t<br \/>\ncounting(ann, \t0). \t<br \/>\ncounting \t(XI, \tI) \t: - \tcounting \t<br \/>\nI=J+ \t1. \t<br \/>\nWith \tthe goal: \t<br \/>\n? - sgc\u2019(ann, \tY, \t0). \t<br \/>\nWith \tsome  liberality, \twe \tuse \tbuilt-in  predicates \tfor \tad- \t<br \/>\ndition  and \tsubtraction,  that \twill \tbe \tdiscussed \tin \tSection \t<br \/>\nVI-A. \tThe \treader \tcan \tobserve \tthat \tthe counting predicate \t<br \/>\nincrements \tgeneration levels from \tann \tupwards, \tmarking \t<br \/>\nthe  level \tof \teach \telement \tof \tthe  magic \tset, \twhereas  the \t<br \/>\nsgc\u2019 \tpredicate decrements generation levels. \tThe \tgoal  se- \t<br \/>\nlects  only \tthe \tsgc \tpairs  of \tlevel \t0, \ti.e., \tthose \tat the \tsame \t<br \/>\nlevel \tas \tann. \tAt \teach step, among \tthe \ttuples \tof \tthe \tmagic \t<br \/>\nset, the program \tonly \tuses \tthe \ttuples that have the \tappro- \t<br \/>\npriate distance from \tann. \t<br \/>\nThe \tapplication  of \tthe \tcounting \tmethod  has \tthe \tpoten- \t<br \/>\ntial \tfor \timproving \tthe \tefficiency \tof \tthe  computation, \tbut \t<br \/>\nclearly \tthe \tmethod \tdoes \tnot \tterminate \twhen  the \tdatabase \t<br \/>\nis cyclic \t(as \tthe increment \tof \tthe counting variable is \tnot \t<br \/>\narrested). \tThus,  the \tcounting \tmethod \tapplies \tto \tacyclic \t<br \/>\ndatabases only. \tA \tfurther hypothesis  is \tthat \tthe \tprogram \t<br \/>\nbe \tlinear, \twith \tat most \tone \trecursive rule \tfor \teach \tpredi- \t<br \/>\ncate. <br \/>\nIn \torder \tto \timprove  the \tapplicability  of \tthe  counting \t<br \/>\nmethod \twhenever \tit is \tnot \tknown \ta priori \twhether  the \t<br \/>\ndatabase  is  cyclic,  Sacca\u2019  and  Zaniolo  have \tintroduced \t<br \/>\n[ 1061 \tthe \t\u201cmagic \tcounting \tmethod\u201d, \twhich  constantly \t<br \/>\nmonitors \tthe counting computation \tin \torder \tto \tdetermine \t<br \/>\nwhether \tthe \tunderlying \tdatabase is cyclic; \tif so, \tit switches \t<br \/>\nto the magic set computation. <br \/>\nE. \tThe Method \tof \tHenschen \tand \tNaqvi \t<br \/>\nOne \tof \tthe  earliest  pure  evaluation \tmethods  proposed \t<br \/>\nin \tthe  literature \twas \tdeveloped \tby \tHenschen  and \tNaqvi \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERl \trf a1 : DATALOG \t157 \t<br \/>\n[59]; \tthe  method \tapplies \tto \tlinear Datalog programs \twith \t<br \/>\ngoals \tthat \tcontain \tbound \targuments. \t<br \/>\nThis \tmethod \tproduces \tan \titerative \tprogram  that \tevalu- \t<br \/>\nates \tthe  goal  through  several \tsteps. \tEach  step  produces \t<br \/>\nsome \tof \tthe \tanswer \ttuples, \tand,  at \tthe \tsame time,  com- \t<br \/>\nputes  symbolically \ta new \texpression \tthat \thas \tto \tbe  eval- \t<br \/>\nuated \tat \tthe  following \tstep. \tThe \tmethod  is  based \ton \ta \t<br \/>\n\u2018\u2018functional\u2019\u2019 \tinterpretation \tof \tpredicates:  we  can \tview \t<br \/>\nany \tpredicate \tp having \tat \tleast \ttwo \targuments \tas \ta set \t<br \/>\nfunction \tfrom \tsome \tof \tits \targuments \tto \tthe \tremaining \t<br \/>\nones,  associating \tto \teach \tset \tS of \tvalues \tof \tits \tfirst \targu- \t<br \/>\nments \ta set \tS\u2019 \tof \tvalues \tof  its \tsecond arguments; \tfor \tin- \t<br \/>\nstance, \tifp \tis a binary \tpredicate, \twe \tdenote \tby \tfp \tthe func- \t<br \/>\ntional  mapping \t<br \/>\nS\u2018 \t= fp(S) \t= \t(ylx~S \tand \tp(x,y)}. \t<br \/>\nWith  this \tnotation, \tone \tcan also \tperform \tthe composition \t<br \/>\nof \tpredicate  functions, \tin \tthe \tusual  mathematical \tsense. \t<br \/>\nThis \tcan \tbe \tapplied \tto  the \tsolution \tof \tlinear Datalog \tpro- \t<br \/>\ngrams, \tby \tconsidering the  bound  arguments \tas \tthe  func- \t<br \/>\ntion\u2019s \tdomains. \tWe \texemplify \tthe  method  of \tHenschen- \t<br \/>\nNaqvi \ton \tprogram \tPI, \twith \tthe goal: \t?-sgc(a, \tX). \t<br \/>\nfperson \t( a ) U frap \t( fpmm \t( hur \t(a ) 1) \t<br \/>\nU frap \t(frap(rb=~s~~.(fpar(fpar(a))))) \tU \t*  * \t* \t<br \/>\nwhere \tfperJon \tdenotes \ta unary \tfunction \treturning  all \tper- \t<br \/>\nsons, \tfPur \tis the \tset \tfunction  from \tthe \tfirst \tto \tthe \tsecond \t<br \/>\nargument \tof \tthe \trelation \tPAR, \tandf,,, \tis the \tset \tfunction \t<br \/>\nfrom the second \tto the \tfirst \targument \tof \tthe \trelation \tPAR. \t<br \/>\nThe \tmost \tinteresting feature \tof \tthe \tmethod  of \tHenschen \t<br \/>\nand \tNaqvi \tis that  it \tintegrates \ttwo \tkinds  of \tcomputation: \t<br \/>\nat a certain \tstep, \tsome tuples \tof \tthe answer \tare \tcomputed, \t<br \/>\nbut \talso  some  symbolic manipulation  is  performed. \tThe \t<br \/>\nfirst \tkind \tof \tcomputation  is \ttypical  of \tpure  evaluation \t<br \/>\nmethods, \tthe \tsecond \tis characteristic \tof  rewriting  meth- \t<br \/>\nods. <br \/>\nSubstantially  the \tsame \tfunctional \tinterpretation \tof pred- \t<br \/>\nicates is \tgiven in \ta more general form \tby \tGardarin and \tDe \t<br \/>\nMaidreville \tin \t[52], \twho propose \ta method \tfor \tevaluating \t<br \/>\nqueries \tas \tfunction \tseries, \twhere \tthe \tfunctions \tinvolved \t<br \/>\ncorrespond  to \tthe  functional \tinterpretation  of \tthe  predi- \t<br \/>\ncates.  Follow-up work can be found \tin \t[54]. \t<br \/>\nF. \tOther \tEficient \tEvaluation  Methods \t<br \/>\nSeveral \tother \tmethods \tcan \tbe \tfound \tin \tthe  literature; \t<br \/>\nmost \tof \tthem \tshare some characteristics with the methods \t<br \/>\npresented \tabove. \t<br \/>\nStatic  Filtering \tis a rewriting \tmethod \tintroduced \tby \tKi- \t<br \/>\nfer \tand  Lozinskii \tin \t[68].  In \tthis  method, \ta bottom-up \t<br \/>\nevaluation  is  viewed \tas \taflow \tof \ttuples  through \ta graph \t<br \/>\nderived \tfrom  the  program,  called \trelation-axiom \tgraph, \t<br \/>\nwith \ttwo \ttypes \tof \tnodes: \trelation \tnodes, \tassociated \tto \t<br \/>\npredicates, \tand \taxiom \tnodes, \tassociated \tto \trules. \tCom- \t<br \/>\nputation  is  ideally \tpreformed  inside axiom nodes. \t<br \/>\nWhen \tthe \tgoal has some bound arguments, those tuples \t<br \/>\nproduced \tduring \tthe \tgraph traversal \twhich \tdo \tnot  satisfy \t<br \/>\nthe  bindings  can  be  eliminated \tat the  end \tof \tthe \tcompu- \t<br \/>\ntation; the \tidea of \tthe \tmethod  is that \tof \t\u201ccutting\u201d \toff \tuse- \t<br \/>\nless tuples \tfrom the \tcomputation  at an \tearlier stage \tof \ttheir \t<br \/>\nflow \ttowards the goal node. \tThis \tis achieved \tby \timposing \t<br \/>\nconditions \ton  predicate  arguments, \tcalled \tjilrers, \tto  the \t<br \/>\noutput \tedges \tof \teach \trelation \tnode; the expressions \tof \tfil- \t<br \/>\nters \tare  derived  from  the  goal  bindings, \tand  propagated \t<br \/>\nalong \tthe graph \tby \ta push \toperation. \t<br \/>\nA \tsimilar  approach \twas \tpresented \tby \tDevanbu \tand \t<br \/>\nAgrawal  in \t[47], \tbut  restricting \tthe \tapplication  to \tlinear \t<br \/>\nrules \twith \tonly \tone \toccurrence \tof \tthe  recursive predicate \t<br \/>\nin \tthe \tRHS. \tAnother \tinteresting \t(pure  evaluation)  algo- \t<br \/>\nrithm  (the  Apex \tmethod) \twas  previously  introduced \tby \t<br \/>\nLozinskii  in \t[SI]. \tThe \tfirst \tmethod \tfor \tpushing selections \t<br \/>\ninto \trecursive expressions \twas  proposed \tby \tAho and \tU11- \t<br \/>\nman  in \ta seminal \tpaper \t[lo]; \tthe static \tfiltering  technique \t<br \/>\ncan \tbe \tconsidered \tas \ta generalization \tof  that \tconcept. \t<br \/>\nAnother generalization \tof \t[ 101 \tis presented \tby \tCeri and \t<br \/>\nTanca \tin \t[29], \tby \tintroducing the  pair \tof \tmethods \tVuri- \t<br \/>\nable  Reduction \tand \tConstant  Reduction \twhich  apply  to \t<br \/>\ngeneric \tsystems \tof \talgebraic  equations. \tBoth  methods \t<br \/>\npush \tthe \tinitial \tselections \tso \tas \tto  use \ttheir  \u201cfiltering\u201d \t<br \/>\nability \tas \tsoon \tas \tpossible; \tthe \ttwo \tmethods  act \teither \tby \t<br \/>\nrewriting \tequations  into  equivalent  ones \twith \tdifferent \t<br \/>\n(smaller) \tvariables \tor \tby \treducing  the \tsize \tof \tinvolved \tre- \t<br \/>\nlations \tprior \tto \tthe equations\u2019 evaluation. \tThese \tmethods \t<br \/>\nare \tpart \tof \ta structured  approach \tto \tthe \toptimization \tof \t<br \/>\nsystems \tof \talgebraic  equations  derived \tfrom \tlogic \tpro- \t<br \/>\ngrams, \tpresented  in \t[29]. \t<br \/>\nAn \tidea \tsimilar \tto  magic  sets  is \texhibited \tin \tthe \tAlex- \t<br \/>\nander \tMethod, \tpublished \tin \t[loll; \tthe  method \tconsists \tof \t<br \/>\nrewriting  the rules \tfor each \trecursive  predicate \tso \tthat \tthey \t<br \/>\nrepresent,  respectively,  problems  and  solutions  for \tthe \t<br \/>\npredicate; all \tnew rules \tobtained \tare \tevaluated bottom-up, \t<br \/>\nbut  the \tevaluation \tof \tthe \tproblem  rules \tsimulates \tin  fact \t<br \/>\ntop-down  evaluation \tand  allows \tbinding \tpassing  among \t<br \/>\nsubgoals. <br \/>\nThere  are \talso \tother \tpossible \tapproaches \tto \toptimiza- \t<br \/>\ntion. \tOne \tof \tthem  is  the  method \tfor \tredundancy \telimina- \t<br \/>\ntion \tof \tSagiv: \tin \t[lo71 \tan \talgorithm \tis  presented \tfor \tmin- \t<br \/>\nimizing \tthe \tsize \t(in  terms \tof \tnumber \tof \trules  in  the \t<br \/>\nprogram \tand \tof \tnumber \tof \tatoms \tin \ta rule) \tof \ta Datalog \t<br \/>\nprogram \tunder \ta \tdecidable  condition \tcalled \tuniform \t<br \/>\nequivalence. \tAnother \tkind  of \toptimization \tis achieved \tby \t<br \/>\nRamakrishnan and others, \tin \t[98], \twhere \tthe \tobjective \tof \t<br \/>\noptimization \tis \tpushing \tprojections, \trather \tthan  selec- \t<br \/>\ntions, \tin the \tbody  of \ta Datalog rule.  This means deleting \t<br \/>\nsome argument positions \tin \tthe  literals \tof \tthe  rule \tbody, \t<br \/>\nwhich \talso  has,  sometimes,  the \teffect \tof  making \tsome \t<br \/>\nrules  redundant \tfor \tthe \tcomputation. \t<br \/>\nA comparative evaluation \tof \tthe  performance \tof \tmeth- \t<br \/>\nods \tpresented \tprior \tto \t1986 \tis presented \tby \tBancilhon \tand \t<br \/>\nRamakrishnan \t[15]. \tThe \tstudy \tuses \ta benchmark \twhich \t<br \/>\nincludes  only \tfew, \tconventional  programs; \tit excludes \t<br \/>\ncyclic databases, \twhich \tare quite \tcommon. However, \tthis \t<br \/>\npaper \tremains \tone \tof \tthe  few  quantitative approaches \tto \t<br \/>\nthe \tcomparison \tof \toptimization  methods;  we \tthink \tthat \t<br \/>\nthis  topic \tdeserves \ta much \tmore  thorough  treatment,  al- \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n158 \tIEEE \tTRANSACTIONS \tON \tKNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL. \tI, NO. \tI. MARCH \t1989 \t<br \/>\nthough \tachieving \ta comparative \tevaluation \tof \tmethods\u2019 \t<br \/>\nperformance \tis extremely \tdifficult \tdue \tto \tthe \tvariety  and \t<br \/>\ninherent  complexity  of \tmethods and \tto the \tcontinuous ev- \t<br \/>\nolution  of \tthe \tfield. \t<br \/>\nG. \tComputing  Transitive \tClosures \tEficiently \t<br \/>\nComputing \ttransitive \tclosures \tefficiently \thas  been  rec- \t<br \/>\nognized \tas \ta  significant \tsubproblem \tof \tthe \tefficient \tcom- \t<br \/>\nputation \tof \tgeneral \trecursive \tqueries \t[621, \t[7]-[9], \t11231. \t<br \/>\nIn \tfact, \tmany  proposals  of \tother  extensions \tto \tthe \trela- \t<br \/>\ntional  query \tlanguages \tin \t\u2018\u2018nontraditional\u201d \tdirections  in- \t<br \/>\nclude \tthis \toperation \t[8], \t[102], \t[58], \t[136]. \tThus, \tan  in- \t<br \/>\ndependent \tarea \tof \tresearch  has  been \tdeveloped \twhich \t<br \/>\nstudies \tthe \tefficient \timplementation of \ttransitive  closures \t<br \/>\nimproving \tthe \tnaive \tand  semi-naive  methods.  The \teffi- \t<br \/>\ncient  methods \tare \tmainly \tconcerned \twith  transitive  clo- \t<br \/>\nsures  of  binary  relations  representing  graphs  and \ttrees, \t<br \/>\nvery \toften \tmaking  use \tof \tsparse  matrices \tto  represent \t<br \/>\nthem. <br \/>\nIn \t[62], \ta logarithmic technique  is \tapplied \tin  order \tto \t<br \/>\nbuild \tmore \tefficient \titerations. \tThe \tpaper \tprovides \tan \tin- \t<br \/>\nteresting  study \tof \tperformance \ton \ttree-shaped \tdatabases, \t<br \/>\nshowing  that \tlogarithmic techniques \tperform  much  better \t<br \/>\nthan \tthe \ttraditional \tones (as \tit  could \tbe \texpected). \tIn \t[7], \t<br \/>\ndirect \talgorithms \tare \tproposed, \twhich \tare \tbased  on \tstud- \t<br \/>\nies  performed \tin \tdifferent \tcontexts \t[108],  [109],  [131], \t<br \/>\n[132]. \tThe \tname  \u201cdirect\u201d \talgorithm \tdescends  from  the \t<br \/>\nfact \tthat \tthe \tlength  of \tthe \tcomputation  does \tnot \tdepend \t<br \/>\non \tthe \tlength  of  paths \tin  the \tunderlying \tgraph. \tSeveral \t<br \/>\nresearch  efforts \thave  also \tbeen  directed  towards \tthe \tpar- \t<br \/>\nallel \tcomputation of transitive closures. \tIn \tthese \tpropos- \t<br \/>\nals,  the  computation \tof  transitive  closures \tis performed \t<br \/>\non  several  processors \tat \tthe  same  time;  the \tsignificant \t<br \/>\nproblem \tis avoiding  maintaining \ttoo \tmany \tduplicates, and \t<br \/>\nto \tperform \ttoo \tmany \tduplicate computations \ton \tdifferent \t<br \/>\nprocessors. \tTwo \tsuch algorithms were \tproposed \tin \t[ 1241. \t<br \/>\nOther \tsignificant  papers \tare \t[9] \tand \t1671. \t<br \/>\nVI. \tEXTENSIONS \tOF \tPURE \tDATALOG \t<br \/>\nThe \tDatalog \tsyntax \twe \thave \tbeen  considering \tso \tfar \t<br \/>\ncorresponds  to  a  very  restricted \tsubset \tof \tfirst-order  logic \t<br \/>\nand  is \toften  referred \tto \tas \tpure \tDatalog. \tSeveral \texten- \t<br \/>\nsions \tof \tpure \tDatalog \thave \tbeen  proposed \tin the literature \t<br \/>\nor \tare \tcurrently \tunder  investigation. \tThe \tmost \timportant \t<br \/>\nof \tthese extensions \tare \tbuilt-in \tpredicates,  negation, \tand \t<br \/>\ncomplex \tobjects. \t<br \/>\nA. \tBuilt-In Predicates <br \/>\nBuilt-in  predicates \t(or  \u201cbuilt-ins\u201d) \tare \texpressed \tby \t<br \/>\nspecial predicate symbols such as \t&gt; , &lt; , 1, \t5, \t=, \t# \t<br \/>\nwith  a  predefined \tmeaning.  These  symbols  can  occur \tin \t<br \/>\nthe  right-hand \tside \tof  a \tDatalog  rule; \tthey \tare \tusually \t<br \/>\nwritten \tin \tinfix \tnotation. \t<br \/>\nConsider, \tfor \texample,  the \tfollowing  program \tP2 \tcon- \t<br \/>\nsisting  of  a \tsingle \trule \twhere \tpar \tis an EDB predicate: \t<br \/>\nP2: \tsibling(X, \tY) \t:- \tpar(Z, \tX), \tpar(Z, \tY), \tX \t+ \tY. \t<br \/>\nThe \tmeaning  of  this  program \tis obvious. \tBy \tuse  of  the \t<br \/>\ninequality  built-in \tpredicate \twe avoid \tthat \ta person \tis con- \t<br \/>\nsidered \tas \this \town sibling. \t<br \/>\nFrom \ta formal \tpoint  of \tview, \tbuilt-ins \tcan \tbe \tconsid- \t<br \/>\nered \tas EDB-predicates \twith  a  different  physical  realiza- \t<br \/>\ntion  than  ordinary \tEDB-predicates: \tthey \tare \tnot \texplicitly \t<br \/>\nstored \tin \tthe \tEDB \tbut \tare \timplemented \tas \tprocedures \t<br \/>\nwhich \tare \tevaluated \tduring \tthe \texecution \tof \ta Datalog \t<br \/>\nprogram. However, \tbuilt-ins  correspond \tin most  cases \tto \t<br \/>\ninJnite \trelations, and \tthis \tmay \tendanger the \tsafety \tof \tDa- \t<br \/>\ntalog \tprograms. \t<br \/>\nSafety  means \tthat \ta Datalog \tprogram  should \talways \t<br \/>\nhave  a finite \toutput, \ti.e., \tthe \tintensional  relations \tdefined \t<br \/>\nby \ta Datalog \tprogram \tmust \tbe \tfinite.  It \tis easy \tto see \tthat \t<br \/>\nsafety \tcan \tbe \tguaranteed \tby \trequiring  that  each  variable \t<br \/>\noccurring \tas argument \tof a built-in  predicate \tin a rule body \t<br \/>\nmust also occur in an \tordinary \tpredicate \tof \tthe \tsame \trule \t<br \/>\nbody, \tor \tmust \tbe \tbound \tby \tan \tequality \t(or \ta sequence \tof \t<br \/>\nequalities)  to  a  variable  of \tsuch an \tordinary  predicate \tor \t<br \/>\nto  a \tconstant. \tHere, \tby \t\u201cordinary  predicate\u201d, \twe \tmean  a \t<br \/>\nnonbuilt-in \tpredicate. \t<br \/>\nDuring \tthe \tevaluation \tof \ta Datalog \trule  with  built-in \t<br \/>\npredicates, the \tfollowing \tprinciple \thas \tto be \tobserved: de- \t<br \/>\nfer  the  evaluation  of  a  built-in  predicate  until \tall \targu- \t<br \/>\nments  of \tthis \tpredicate \tare \tbound \tto \tconstants. \tAn \texcep- \t<br \/>\ntion \tto \tthis \tprinciple \tcan (sometimes \tshould)  be \tmade for \t<br \/>\nthe equality \tpredicate. \tAn \tequality  should \tbe \tevaluated \tas \t<br \/>\nsoon \tas \tone \tof  its \ttwo arguments \tis a constant \tor \tis bound \t<br \/>\nto  a \tconstant. <br \/>\nIn \ta similar \tway, \tarithmetical  built-in predicates \tcan \t<br \/>\nbe used.  For \tinstance, \ta predicate \tplus \t(X, \tY, \tZ) \tmay \tbe \t<br \/>\nused \tfor expressing \tX \t+ \tY = \tZ, \twhere \tthe \tvariables \tX, \t<br \/>\nY, \tand \tZ are \tsupposed \tto \trange \tover \ta numeric  domain. \t<br \/>\nDuring the \tevaluation \tof \ta rule \tbody, \tsuch \ta predicate \tcan \t<br \/>\nbe evaluated \tas \tsoon as bindings \tfor \tits \u201cinput  variables\u201d \t<br \/>\n(here \tX and \tY) \tare \tprovided. \t<br \/>\nFinally, \tlet \tus \tremark \tthat \twhen \tDatalog \trules \tare \ttrans- \t<br \/>\nformed \tinto \talgebraic equations, \tthen \tmany \tbuilt-in  pred- \t<br \/>\nicates can \tbe \texpressed through \tjoin \tconditions. \tThe \tabove \t<br \/>\ndefined \tprogram \tP,, \tfor \tinstance,  is \ttranslated \tinto \tthe \t<br \/>\nfollowing  equation  of  relational \talgebra: \t<br \/>\nSIB \t= IT \t2,4(PAR \tw1=1,-,2+2 \tPAR) \t<br \/>\nwhere \tSIB \tand \tPAR \tdenote the \trelations  corresponding  to \t<br \/>\nthe \tpredicates \tsibling \tand \tparent \t, respectively. \t<br \/>\nB. \tIncorporating \tNegation \tinto \tDatalog: \tThe Problem \t<br \/>\nIn \tpure \tDatalog,  the \tnegation  sign \t\u2018\u2018 1 \u201d is not \tallowed \t<br \/>\nto \tappear.  However, \tby \tadopting \tthe \tClosed  World \tAs- \t<br \/>\nsumption \t(CWA), \twe  may \tinfer \tnegative \tfacts from \ta  set \t<br \/>\nof  pure  Datalog \tclauses. \t<br \/>\nNote \tthat \tthe \tCWA \tis  not \ta  universally  valid  logical \t<br \/>\nrule, \tbut \tjust \ta  principle  that \tone \tmay \tor \tmay  not \tadopt, \t<br \/>\ndepending \ton \tthe  semantics \tgiven  to  a \tlanguage. \tIn \tthe \t<br \/>\ncontext  of \tDatalog,  the \tCWA \tcan \tbe \tformulated \tas  fol- \t<br \/>\nlows: <br \/>\nCWA: \tIf a fact does \tnot \tlogically  follow  from  a  set \tof \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n159 \tCERl \tet al.: \tDATALOG \t<br \/>\nDatalog \tclauses, \tthen \twe \tconclude \tthat  the  negation \tof \t<br \/>\nthis  fact \tis true. \t<br \/>\nNegative Datalog facts \tare \tpositive ground \tliterals  pre- \t<br \/>\nceded \tby \tthe negation \tsign, for \tinstance, \t1 sgc(bertrand, \t<br \/>\nhilary). \tNote \tthat  this \tnegative fact follows \tby \tthe \tCWA \t<br \/>\nfrom \tP, \tU \tE,, \tsince \tsgc(bertrand, \thilary) \tdoes \tnot \tfol- \t<br \/>\nlow \tfrom \tPi \tU \tE,. \tIf F denotes \ta negative \tground \tfact, \t<br \/>\nthen \t1 F I denotes  its  positive  counterpart. \tFor \texample: \t<br \/>\nI 1 sgc(bertrand, \thilary) \tI = sgc(bertrand, \thilary). \t<br \/>\nThe \tCWA \tapplied  to \tpure Datalog allows \tus \tto deduce \t<br \/>\nnegative  facts \tfrom \ta set \tS of \tDatalog clauses. \tIt does not, \t<br \/>\nhowever,  allow \tus \tto \tuse \tthese negative facts \tin order \tto \t<br \/>\ndeduce  some  further  facts. \tIn \treal \tlife, \tit is often \tneces- \t<br \/>\nsary \tto express \trules \twhose premises contain negative in- \t<br \/>\nformation, \tfor \tinstance: \t\u201cifX \tis a  student \tand \tX is not \ta \t<br \/>\ngraduate student, \tthen \tX is an undergraduate \tstudent\u201d.  In \t<br \/>\npure \tDatalog,  there \tis no \tway \tto \trepresent \tsuch \ta rule. \t<br \/>\nNote \tthat \tin \trelational \talgebra  an  expression  corre- \t<br \/>\nsponding \tto \tthe  above \trule \tcan \tbe \tformulated \twith \tease \t<br \/>\nby \tuse  of \tthe \tset-difference \toperator \t\u201c-\u201d. \tAssume  that \t<br \/>\na one-column  relation \tSTUD \tcontains the \tnames  of all \tstu- \t<br \/>\ndents \tand another one-column \trelation \tGRAD \tcontains \tthe \t<br \/>\nnames  of  all \tgraduate students. \tThen we \tobtain \tthe \trela- \t<br \/>\ntion \tUND \tof \tall  undergraduate  students \tby \tsimply  sub- \t<br \/>\ntracting \tGRAD \tfrom \tSTUD, \tthus \t<br \/>\nUND \t= \tSTUD \t- \tGRAD. \t<br \/>\nOur \tintention \tis now \tto \textend  pure Datalog \tby \tallowing \t<br \/>\nnegated \tliterals  in  rule  bodies.  Assume \tthat  the \tunary \t<br \/>\npredicate \tsymbols \tstud, \tund, \tand \tgrad \trepresent \tthe \tprop- \t<br \/>\nerties \tof \tbeing  a \tstudent,  an undergraduate, \tand  a \tgradu- \t<br \/>\nate, \trespectively. \tOur \trule  could  then \tbe \tformulated \tas \t<br \/>\nfollows: \t<br \/>\nund(X) \t:- \tstud(X), \t1 grad(X). \t<br \/>\nMore  formally,  let \tus \tdefine  Datalog\u2019 \tas the  language \t<br \/>\nwhose  syntax \tis that \tof \tDatalog  except \tthat  negated  lit- \t<br \/>\nerals are allowed in \trule \tbodies.  Accordingly, \ta Datalog\u2019 \t<br \/>\nclause  is  either \ta positive  (ground)  fact \tor \ta  rule \twhere \t<br \/>\nnegative \tliterals \tare \tallowed \tto \tappear  in  the  body. \tFor \t<br \/>\nsafety  reasons \twe \talso require \tthat \teach \tvariable occurring \t<br \/>\nin a negative \tliteral \tof a \trule \tbody \talso occurs \tin a positive \t<br \/>\nliteral  of \tthe same \trule body. \t<br \/>\nIn \torder \tto \tdiscuss \tthe \tsemantics \tof \tDatalog\u2019  pro- \t<br \/>\ngrams, \twe \tfirst \tgeneralize  the \tnotion  of \tthe \tHerbrand \t<br \/>\nModel \t(see Section \t11-D) \tto \tcover \tnegation \tin rule \tbodies. \t<br \/>\nLet \t9 be \ta Herbrand interpretation, \ti.e., \ta subset \tof \tthe \t<br \/>\nHerbrand \tbase \tHB. \tLet \tF denote \ta positive \tor \tnegative \t<br \/>\nDatalog \tfact. \t<br \/>\n(\u201c \tis a positive \tfact  and \tF E 9, \tor \tF is  satisfied \tin \t9 iff \tLF \tis a negative \tfact  and \tIF \t1 g! 9. \t<br \/>\nNow, \tlet \tR \tbe \ta  Datalog\u2019 \trule \tof \tthe  form \t: - \tLi, \t. . . \t, L, \tand let \t9 be \ta Herbrand  interpretation. \tR is sat- \t<br \/>\nisfied  in \t9 iff \tfor \teach \tground \tsubstitution \tO for \tR, \twhen- \t<br \/>\never \tit  holds  that \tfor \tall \t1 5 \ti I \tn, \tLie \tis satisfied \tin \t9, \t<br \/>\nthen  it \talso \tholds  that \tLOO \tis satisfied \tin \t9. \t(Note \tthat \tLOO \t<br \/>\nis satisfied \tin \t9 iff \tLOO \tE 4, \tsince \tLOO \tis positive.) \t<br \/>\nLet \tS be \ta  set  of  Datalog\u2019 \tclauses. \tA Herbrand \tinter- \t<br \/>\npretation \t9 is a  Herbrand  model  of \tS iff \tall \tfacts \tand  rules \t<br \/>\nof \tS are \tsatisfied \tin \t9. \t<br \/>\nIn  analogy  to  pure \tDatalog, \twe  require  that \tthe \tset  of \t<br \/>\nall  positive \tfacts  derivable  from \ta set \tS of \tDatalog\u2019 \t<br \/>\nclauses \tbe \ta  minimal  model  of \tS. \tHowever, \ta  set \tS of \t<br \/>\nDatalog: \tclauses \tmay \thave more \tthan \tone \tminimal \tHer- \t<br \/>\nbrand \tmodel. \tFor \tinstance, \tif S, \t= \t{boring \t(chess) \t: - \t<br \/>\n1 interesting \t(chess)}, \tthen \tS has \ttwo  minimal \tHerbrand \t<br \/>\nmodels: \tHa \t= \t{interesting(chess)) \tand \tHb \t= \t{bor- \t<br \/>\ning \t(chess) \t} . \t<br \/>\nThe \texistence \tof \tseveral \tminimal \tHerbrand \tmodels \tfor \t<br \/>\na  set  of \tDatalog\u2019 \tclauses  entails \tdifficulties \tin \tdefining \t<br \/>\nthe semantics \tof  Datalog\u2019 \tprograms: \twhich  of \tthe \tdiffer- \t<br \/>\nent \tminimal Herbrand  models  should \tbe \tchosen?  Note \talso \t<br \/>\nthat \tthe \tmodel  minimality  requirement \tis inconsistent \twith \t<br \/>\nthe \tCWA. \tBy \tthe  CWA \tboth \tfacts \t1 boring(chess) \tand \t<br \/>\n1 interesting(chess) \tcan  be \tdeduced  from  the  above \tset \t<br \/>\nS,. \tThus, \tneither \tof \tthe \tmodels \tHa \tand \tHb \tare \tconsistent \t<br \/>\nwith \tthe \tCWA. \t<br \/>\nIn \tthe \tfollowing \twe \tdescribe \ta  policy  which \tis  com- \t<br \/>\nmonly  referred  to \tas \tstratijied \tevaluation \tof \tDatalog\u2019 \t<br \/>\nprograms, \tor \tsimply \tas \tstratijied  Datalog\u2019. \tThis \tpolicy \t<br \/>\npermits \tus \tto \tselect  a \tdistinguished  minimal \tHerbrand \t<br \/>\nmodel \tin \ta very \tnatural \tand intuitive \tway \tby \tapproximat- \t<br \/>\ning \tthe \tCWA. \tHowever, \tas \twe will \tsee later, \tthis  method \t<br \/>\ndoes \tnot \tapply \tto all  Datalog\u2019 \tprograms, \tbut  only \tto par- \t<br \/>\nticular \tsubclass;  the \tso-called \tstratijied \tprograms \t[ 371, \t<br \/>\n[ 121. \tNote also that this technique can be \tused \tin the \tmore \t<br \/>\ngeneral  context  of \tLogic Programming \twith  negation \t[ 121 \t<br \/>\nas well. <br \/>\nC. \tStratijied \tDatalog\u2019 \t<br \/>\nThis \tpolicy \tof \tchoosing \ta particular  Herbrand  model, \t<br \/>\nand  thus  of \tdetermining the semantics \tof \ta  Datalog\u2019  pro- \t<br \/>\ngram \tis guided \tby \tthe \tfollowing \tintuition: \twhen \tevaluat- \t<br \/>\ning  a  rule \twith \tone \tor \tmore negative literals \tin \tthe body, \t<br \/>\nfirst \tevaluate the \tpredicates \tcorresponding \tto \tthese \tnega- \t<br \/>\ntive \tliterals. \tThen the CWA \tis \u201clocally\u201d \tapplied \tto these \t<br \/>\npredicates. <br \/>\nFor \tinstance, \tthe \tclause set \tS,, \tas \tdefined \tabove, \twould \t<br \/>\nbe  evaluated \tas \tfollows: \tbefore \ttrying \tto \tevaluate  the \t<br \/>\npredicate \tboring, \twe \tevaluate  the  predicate \tinteresting \t<br \/>\nwhich \toccurs \tnegatively \tin the \trule \tbody.  Since \tthere \tare \t<br \/>\nno \trules \tand facts \tin \tS, \tallowing \tus \tto deduce \tany \tfact \tof \t<br \/>\nthe  form \tinteresting(a), \tthe  set \tof \tpositive  answers \tto \t<br \/>\nthis \tpredicate \tis \tempty. \tIn \tparticular, \tinteresting \t(chess) \t<br \/>\ncannot \tbe \tderived. \tHence, \tby \tapplying \tthe \tCWA \t\u201clo- \t<br \/>\ncally\u201d \tto \tthe \tinteresting \tpredicate, \twe \tderive \t1 interest- \t<br \/>\ning \t(chess). \tNow \twe \tevaluate the unique \trule \tof \ts, and get \t<br \/>\nboring \t(chess). \tThus, the \tcomputed Herbrand \tmodel \tis Hb \t<br \/>\n= \t(boring \t(chess)}. \t<br \/>\nWhen several \trules \toccur \tin \ta  Datalog\u2019 \tprogram, \tthen \t<br \/>\nthe \tevaluation \tof a rule \tbody \tmay \tengender the \tevaluation \t<br \/>\nof \tsubsequent \trules. \tThese \trules  may \tcontain \tin turn  neg- \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n160 \tIEEE \tTRANSACTIONS  ON \tKNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL. \t1, NO. \t1, MARCH \t1989 \t<br \/>\native literals \tin their \tbodies  and \tso \ton. Thus, \tit is required \t<br \/>\nthat \tbefore evaluating \ta predicate \tin \ta rule head, \tit is al- \t<br \/>\nways \tpossible \tto  completely \tevaluate \tall \tthe \tpredicates \t<br \/>\nwhich \toccur \tnegatively \tin the  rule \tbody \tor \tin \tthe \tbodies \t<br \/>\nof \tsome \tsubsequent  rules \tand  all  those \tpredicates  which \t<br \/>\nare \tneeded \tin order \tto \tevaluate these negative predicates. \t<br \/>\nIf a program \tfulfills \tthis condition it \tis called \tstratzjied. \t<br \/>\nAny \tstratified  program \tP can be \tpartitioned \tinto \tdisjoint \t<br \/>\nsets \tof \tclauses \tP \t= P\u2019 \tU \t. . . \tU \tPi \tU \t. . . \tU \tP \tcalled \t<br \/>\nstrata, \tsuch \tthat \teach IDB-predicate \tof \tP has \tits \tdefining \t<br \/>\nclauses  within \tone \tstratum \tand \tP\u2019 \tcontains \tonly \tclauses \t<br \/>\nwith \teither \tno \tnegative  literals \tor \twith  negative \tliterals \t<br \/>\ncorresponding \tto \tEDB-predicates  and  each \tstratum \tP\u2019 \t<br \/>\ncontains  only \tclauses whose  negative literals correspond \t<br \/>\nto \tpredicates  defined \tin \tlower \tstrata. \tThe \tpartition  of \tP \t<br \/>\ninto \tP\u2019 \t. . \t. P\u201c \tis called  a \tstratijication \tof \tP. \t<br \/>\nAssume  a  stratified  program \tP with  given  stratification \t<br \/>\nP\u2019 \t. . . P\u201c \thas \tto  be \tevaluated  against  an  EDB \tE. \tThe \t<br \/>\nevaluation \tis done \tstratum-by-stratum  as  follows.  First, \t<br \/>\nP\u2019 \tis evaluated \tby \tapplying the \tCWA \tlocally \tto \tthe \tEDB, \t<br \/>\ni.e., \tby \tassuming \t1 p (c,, \t. . . \t, ck) \tfor \teach \tk-ary \tEDB- \t<br \/>\npredicate \tp and \tconstants \tc1, \t. . . \t, ck \twhere \tp( \tc1, \t. . . \t, \t<br \/>\nck) \tE. \tThen  the \tother \tstrata \tare \tevaluated \tin \tascending \torder. \tDuring the evaluation \tof \teach \tstratum \tPi, \tthe \tresult \t<br \/>\nof the previous  computations \tis used and the \tCWA \tis made \t<br \/>\n\u201clocally\u201d  for \tall \tEDB-predicates \tand \tfor \tall  predicates \t<br \/>\ndefined  by \tlower strata. \t<br \/>\nConsider,  for \texample, \tthe \tfollowing  program \tP,, \t<br \/>\nwhere \td is the \tonly \tEDB-predicate: \t<br \/>\nr4: \tr(X, \tY): \t- d( \tY, \tX). \t<br \/>\nr5: \ts(X, \tY) \t:- \tq(X, \tZ), \tq(Y, \tT), \tX \t+ \tY \t<br \/>\nA \tstratification \tof \tP,~ \tis: \tP: \t= \t{ r4}, \t~,2 \t= \t{ r2, \tr3, \tr5}, \t<br \/>\ne \t= \t{ rl }. \tAssume \tP, \tis evaluated \tover \tthe \tEDB \tE, \t= \t<br \/>\n{d(a, \tb), \td(b, \tc), \td(e, \te)}. \tThe \tevaluation \tof \tthe \tfirst \t<br \/>\nstratum  produces  the \tnew \tfacts: \tr(b, \ta), \tr(c, \tb), \tr(e, \t<br \/>\ne). \tThe \tcomputation \tof \tthe \tsecond stratum \tyields \tthe \tfol- \t<br \/>\nlowing  additional \tfacts: \tq(a, \tb), \tq(b, \tc), \tq(a, \tc), \ts(a, \t<br \/>\nb), \ts(b, \ta). \tFinally, \tby \tevaluating the \tthird \tstratum \twe \t<br \/>\nNote \tthat  a  stratified \tprogram \thas, \tin \tgeneral,  several \t<br \/>\ndifferent stratifications. \tThe \tprogram \tP,, \tfor \texample, has \t<br \/>\nthe \tfollowing \talternative \tstratification: \tP: \t= \t{ r4 } , P: \t= \t<br \/>\n{ r2, \tr3 }, Pa \t= \t{ rl, \tr5 } . However, \tit can \tbe \tshown \t[12] \t<br \/>\nthat all  stratifications \tare equivalent, \ti.e., \tthe \tresult  of \tthe \t<br \/>\nevaluation  of  a  stratified  Datalog\u2019  program \tP is indepen- \t<br \/>\ndent \tof \tthe \tstratification \tused. \t<br \/>\nIt  is \teasy \tto \tdecide whether \ta given \tDatalog\u2019  program \t<br \/>\nis stratified \tby \tanalyzing \tthe \tExtended \tDependency  Graph \t<br \/>\nEDG \t(P) \tof \tP. \tThe \tnodes \tof \tEDG \t(P) \tconsist \tof \tthe IDB- \t<br \/>\npredicate \tsymbols \toccurring \tin \tP. \tThere \tis \ta  (directed) \t<br \/>\nedge \t&lt; p, \tq &gt; \tin \tEDG(P) \tiff \tthe predicate symbol \tq oc- \t<br \/>\ncurs  positively \tor \tnegatively \tin \ta  body  of  a  rule  whose \t<br \/>\ngetp(b, \ta). \t<br \/>\nhead \tpredicate \tis p. \tFurthermore,  the \tedge \t&lt; p, \tq &gt; \tis \t<br \/>\nmarked \twith \ta \u201c 1 \u201d sign \tiff \tthere exists \tat least \tone \trule \t<br \/>\nR with  head \tpredicate \tp such that \tq occurs \tnegatively \tin \t<br \/>\nthe \tbody  of \tR. \tThe \textended dependency graph \tEDG (P,) \t<br \/>\nof \tthe \tprogram \tP, \tis depicted \tin \tFig. \t4. \t<br \/>\nA \tDatalog\u2019 \tprogram \tP is \tstratified \tiff \tEDG( \tP) \tdoes \t<br \/>\nnot  contain  any \tcycle  involving  an \tedge \tlabeled \twith \t<br \/>\n\u201c 1 \u201d. \tIf P is stratified, then \tit is quite easy \tto \tconstruct \t<br \/>\na particular \tstratification \tfor \tP from \tEDG(P) \t[34]. \tAn- \t<br \/>\nother \tmethod \tfor \tconstructing \ta  stratification \tis  given \tin \t<br \/>\n[122]. <br \/>\nIt can \tbe \tshown that the strata-by-strata evaluation \tof  a \t<br \/>\nstratified  program \tP on the base \tof \tan \tunderlying \tEDB \tE \t<br \/>\nalways \tproduces  a \tminimal Herbrand model \tof \tP U \tEDB. \t<br \/>\nThis \tmodel \tis also \tcalled \tthe \tPerjiect \tModel \tand \tcan \tbe \t<br \/>\ncharacterized \tin \ta  purely \tnonprocedural \tway \t[12], \t[95]. \t<br \/>\nLocal \tstratijication, \ta  refinement  of  stratification, \tis pro- \t<br \/>\nposed \tin  [95]. \t<br \/>\nD. \tInJationary  Evaluation \tand \tExpressive Power \tof \t<br \/>\nDatalog \t\u2019 Programs \t<br \/>\nAnother  evaluation  paradigm \tfor \tDatalog \t\u2019 programs \t<br \/>\nhas \tbeen  proposed  recently \tin \t[4],  [72].  This \tmethod, \t<br \/>\ncalled \tinjationary  evaluation, \thas  the  advantage \tof \tap- \t<br \/>\nplying  to all Datalog\u2019 \tprograms \tand \tnot \tjust \tstratified \tDa- \t<br \/>\ntalog \t\u2019 programs. \t<br \/>\nLet \tP be \ta  Datalog\u2019 \tprogram \tand \tE an \tEDB. \tThe \tin- \t<br \/>\nflationary \tevaluation \tof \tP on \tE is performed \titeratively \tso \t<br \/>\nthat \tall \trules  of \tP are \tprocessed \tin \tparallel \tat \teach \tstep. \t<br \/>\nFrom  the \tEDB \tand \tthe \tfacts  already  derived, \tnew \tfacts \t<br \/>\nare  derived \tby \tapplying  the \trules  of \tP. \tThese \tnew  facts \t<br \/>\nare  added \tto \tthe \tresult \tat the \tend \tof \teach \tstep. \tAt \teach \t<br \/>\nstep, the CWA is made \ttemporarily \tduring the evaluation \t<br \/>\nof \tthe \trule  bodies: \tit is  assumed \tthat \tthe \tnegation  of  all \t<br \/>\nfacts \tnot \tyet  derived  is  valid. \tThe \tprocedure  terminates \t<br \/>\nwhen \tno \tmore additional facts \tcan \tbe derived. \t<br \/>\nConsider, \tfor \texample, \tthe \tfollowing \tnonstratified  pro- \t<br \/>\ngram \tPi, \twhere \td is the \tonly \tEDB-predicate: \t<br \/>\nrl: \ts(X) \t:- \tp(X), \tq(X), \tir(X). \t<br \/>\nr2: \tp(X) \t:- \td(X), \t1 q(X). \t<br \/>\nr3: \tq(X) \t:- \td(X), \t1p(X). \t<br \/>\nr4: \tr(X) \t:- \td(X), \td( \tb). \t<br \/>\nAssume  that  this  program  is  evaluated \tinflationarily \t<br \/>\nagainst  an \tEDB \tEi \t= \t{ d(a)}. \tDuring  the \tfirst \titeration \t<br \/>\nstep,  rule \tr2 produces the \tnew \tfact \tp (a), \tand \trule \tr3 pro- \t<br \/>\nduces \tthe \tnew \tfact \tq( \ta). \tDuring the second iteration step, \t<br \/>\nrule \tr1 produces the \tnew \tfact \ts(a). \tSince no further \tfacts \t<br \/>\nare \tderivable, the procedure \tstops  with \tthe \tresult \t{ p(a), \t<br \/>\nIt is  easy \tto \tsee \tthe \tinflationary \tevaluation \tof  a \tDa- \t<br \/>\ntalog\u2019 \tprogram \tP corresponds \tto \tthe  computation \tof  a \t<br \/>\nleast \tfixpoint \t[72],  [34]. Furthermore,  the \tresult  united \tto \t<br \/>\nE is a Herbrand \tmodel  of \tP \tU \tE. \tHowever, \tin \tgeneral, \t<br \/>\nit  is \tnot \ta minimal \tHerbrand  model. \tFor \tinstance, \tin \tthe \t<br \/>\nabove example the  computed  Herbrand \tmodel \tis \t{ d (a), \t<br \/>\np(a), \tq(a), \ts(a)}, \tbut \tthis  model \tis not  minimal \tsince \t<br \/>\n4(a), s(a)&gt;. \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERI \tet al.: \tDATALOG \t161 \t<br \/>\ni\/ \t<br \/>\nFig. \t4. Extended dependency \tgraph \tEDG(P,) \t<br \/>\n{d(a), \tq(a)} \tand \t{d(a), \tp(a)} \tare \tsmaller  Herbrand \t<br \/>\nmodels \tof \tPi \tU \tEi. \t<br \/>\nJust \tas  for \tpure  Datalog  programs, \tone \tcan \tspecify  a \t<br \/>\ngoal \ttogether  with \ta  Datalog\u2019 \tprogram. \tIn  that \tcase, \t<br \/>\nagain, \tthe  output  consists \tof  all \tthose  derived  IDB-facts \t<br \/>\nwhich \tare \tsubsumed \tby \tthe \tgiven  goal.  It \tcan be \tshown \t<br \/>\n[4], [7 \t11, \t[72], \tthat \tfor \teach \tstratified \tDatalog\u2019 \tprogram \t<br \/>\nP with \tgoal \tG there  exists \ta  Datalog\u2019 \tprogram \tP\u2018 \tsuch \t<br \/>\nthat  the \toutput \tof \tthe \tstratified \tcomputation \tof \tP on \tany \t<br \/>\nEDB \tE w.r.t. \tG is equal \tto \tthe output \tof \tthe \tinflationary \t<br \/>\ncomputation \tof \tP\u2018 \ton \tE w.r.t. \tG. \tThis \tmeans \tthat \tinfla- \t<br \/>\ntionary \tDatalog\u2019 \tis at least \tas \texpressive \tas \tstratified \tDa- \t<br \/>\ntalog\u2019 \t. Moreover,  there  exist programs  whose \tinflation- \t<br \/>\nary \tevaluation \t(w.r.t. \ta given  goal)  cannot \tbe \tsimulated \t<br \/>\nby \tany \tstrata-by-strata  computation \tof  a  stratified \tDa- \t<br \/>\ntalog\u2019 \tprogram. \tThus, \tinflationary  Datalog\u2019  is \tcompu- \t<br \/>\ntationally  strictly \tmore  expressive \tthan  stratified \tDa- \t<br \/>\ntalog\u2019 \t. Furthermore,  it has  been  shown \tthat  inflationary \t<br \/>\nDatalog\u2019 \thas the \tsame expressive power \tas Fixpoint \tLogic \t<br \/>\non \tFinite \tStructures, \ta well-known formalism obtained \tby \t<br \/>\nextending  first-order \tlogic \twith  a \tleast \tfixpoint \toperator \t<br \/>\nfor positive \tfirst-order \tformulas \t[6]. \t<br \/>\nIt is also easy \tto  see \tthat  stratified  Datalog\u2019  is  in  turn \t<br \/>\nstrictly \tmore  expressive  than  pure  Datalog  [4]. \tFig. \t5 \t<br \/>\nshows the \thierarchy  of \texpressiveness \tof \tthe \tdifferent \tlan- \t<br \/>\nguages and  formalisms presented \tin  this \tpaper.  More  in- \t<br \/>\nformation on \tthis \tsubject can \tbe \tfound \tin \t[38]. \t<br \/>\nWe \tfinish \tour \tdiscussion \ton \tincorporating \tnegation \tinto \t<br \/>\nDatalog \tby \tgiving \ta  brief \tsurvey  on \tother \trelevant  work \t<br \/>\non \tthis \ttopic. \t<br \/>\nA different \tapproach \tfor \tdefining \tthe \tmeaning  of \tlogic \t<br \/>\nprograms \twith  negation \tis based  on \ta proposal \tof \tClark \t<br \/>\n[44]. \tHe \tessentially \tconsiders \tthe \tcompletion \tof  a \tlogic \t<br \/>\nprogram  by \tviewing the \tdefinitions  of \tderived predicates \t<br \/>\nas \tlogical \tequivalences  rather  than  logical  implications. \t<br \/>\nThe \tsemantics \tof \ta logic program with \tnegation \tcan then \t<br \/>\nbe \tdefined \tas \ta minimal  model \tof  its \tcompletion.  This \t<br \/>\napproach \twas \tdiscussed \tby \tSheperdson \t[ 1 lo], \t[ 11 11 \tand \t<br \/>\nLloyd \t[78].  Fitting \t[49] \tand  Kunen  [74] \trefined \tthis \tap- \t<br \/>\nproach \tby \tusing \tthree-valued logic: \tin their setting, \ta fact \t<br \/>\ncan be \ttrue, \tfalse, \tor \tundefined.  In \tthe context \tof \tDatalog \t<br \/>\nthese approaches \tare \tnot \tcompletely satisfactory. \t<br \/>\nA \trecently \tintroduced  and \tvery \tpromising  approach, \t<br \/>\nalso \tbased \ton three-valued \tlogic, \tis the \twell-founded \tse- \t<br \/>\nmantics \tby \tVan \tGelder, \tRoss, \tand \tSchlipf  [126]. \tTheir \t<br \/>\nmethod  nicely \textends the \tstratified \tapproach \tto \tarbitrary \t<br \/>\n(= \tnonrecursive \tRecursive \tDatalog) \t( \t(, \tqueries \t<br \/>\nd\u2019 \tStratified \tDatalogy \t<br \/>\nFig. \t5. Hierarchy \tof expressiveness \tof different \tversions \tof Datalog \t<br \/>\nlogic \tprograms \twith \tnegation. \tIn \tparticular,  every \tstrati- \t<br \/>\nfied \tprogram \tis \tsemantically  characterized \tby \ta total \t<br \/>\nmodel, \ti.e., \ta  model \tsuch \tthat \teach fact \tof \tthe  Herbrand \t<br \/>\nbase \thas \teither \ttruth \tvalue \t\u201ctrue\u201d \tor \t\u201cfalse\u201d. \tThis \tmodel \t<br \/>\ncoincides \twith \tthe  perfect model \tmentioned \tabove. \tNon- \t<br \/>\nstratified \tprograms, \ton \tthe \tother \thand, \tcan \tbe  character- \t<br \/>\nized \tby \tpartial \tmodels,  where  single  facts \tmay \tassume \t<br \/>\ntruth \tvalue \t\u201cundejined.\u201d \tA fixpoint  method \tfor comput- \t<br \/>\ning \tthe well-founded  partial  model  is given \tin [ 1271, \twhile \t<br \/>\nresolution-based \tprocedural  semantics \tfor \twell-founded \t<br \/>\nnegation  is  provided  in \t[103].  Further \timportant  papers \t<br \/>\nrelated  to \twell-founded semantics \tare [27], \twhere \tthe \tre- \t<br \/>\nlationship \tto  logical \tconstructivism \tis \tinvestigated,  and \t<br \/>\n[961. <br \/>\nE. \tComplex \tObjects \t<br \/>\nThe \t\u201cobjects\u201d \thandled \tby \tpure \tDatalog programs cor- \t<br \/>\nrespond  to \tthe tuples \tof  relations  which \tin \tturn \tare \tmade \t<br \/>\nof \tattribute  values.  Each  attribute  value \tis atomic, \ti.e., \t<br \/>\nnot \tcomposed \tof \tsub-objects;  thus \tthe  underlying \tdata \t<br \/>\nmodel \tconsists \tof \trelations  in \tfirst \tnormal \tform. \tThis \t<br \/>\nmodel  has \tthe  advantage \tof  being  both  mathematically \t<br \/>\nsimple and easy \tto implement. \tOn \tthe other \thand, \tseveral \t<br \/>\nnew \tapplication areas (such \tas \tcomputer-aided design, of- \t<br \/>\nfice \tautomation,  and  knowledge \trepresentation)  require \t<br \/>\nthe \tstorage and \tmanipulation  of \t(deeply \tnested)  structured \t<br \/>\nobjects \tof \thigh complexity.  Such complex objects cannot \t<br \/>\nbe \trepresented \tas \tatomic  entities \tin \tthe  normalized \trela- \t<br \/>\ntional \tmodel \tbut \tare \tbroken  into  several \tautonomous \tob- \t<br \/>\njects.  This \timplies \ta number \tof \tsevere problems \tof \tcon- \t<br \/>\nceptual \tand \ttechnical \tnature. \t<br \/>\nFor \tthis reason, the \trelational  model \thas been \textended \t<br \/>\nin \tseveral \tways \tto \tallow  the  compact \trepresentation  of \t<br \/>\ncomplex  objects.  Datalog  can  be  extended  accordingly. <br \/>\nThe \tmain \tfeatures that  are \tadded  to \tDatalog \tin \torder \tto \t<br \/>\nrepresent \tand  manipulate  complex  objects \tare \tfinction \t<br \/>\nsymbols \tas \ta glue \tfor \tcomposing objects from \tsub-objects \t<br \/>\nand \tset \tconstructors \tfor \tbeing able \tto \tbuild \tobjects \twhich \t<br \/>\nare  collections \tof \tother \tobjects.  Function  symbols  are \t<br \/>\n\u201cuniterpreted\u201d, \ti.e., \tthey \tdo \tnot \thave \tany  predefined \t<br \/>\nmeaning. \tUsually \tone \talso  adds \ta number \tof \tpredefined \t<br \/>\nfunctions \tfor \tmanipulating \tsets \tand elements \tof \tsets to the \t<br \/>\nstandard \tvocabulary  of \tDatalog. \t<br \/>\nThere \texist several \tdifferent \tapproaches \tfor \tincorporat- \t<br \/>\ning \tthe  concept \tof \ta complex  structured  object \tinto  the \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n162 \tIEEE \tTRANSACTIONS \tON \tKNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL. \tI, NO. \tI, MARCH \t1989 \t<br \/>\nformalism  of \tDatalog \t[ 1331, \t[ 1191, \t[75]. \tOne \tof \tthe \tmost \t<br \/>\nwell-known \tapproaches \thas  been \tdeveloped \twithin  the \t<br \/>\nLDL \t(Logic Data Language) Project, \tcarried  out at \tMCC, \t<br \/>\nAustin, \tTX. \tThe \tnotation \tfor \trepresenting \tsets \tand \trelated \t<br \/>\nissues \tused \tin \tthis \tsubsection \tis the \tone \tof \tLDL. \t<br \/>\nExamples of complex facts \tinvolving  function  symbols \t<br \/>\nand \tsets \tare: \t<br \/>\nperson \t(name( \tjoe, \tberger), \tbirthdate \t( 1956, \tjune, \t30), \t<br \/>\nchildren \t( {mar, \tsarah, \tjim} \t)). \t<br \/>\nperson \t(name( \tjoe, \tcoker), \tbirthdate \t( 1956,,june, \t30), \t<br \/>\nchildren \t( {bill, \tsarah \t} )). \t<br \/>\nperson \t(name( \tbebe, \tsuong), \tbirthdate \t( 1958, \tmay, \t5), \t<br \/>\nchildren \t( { jim, \tmar, \tsarah \t} )) . \t<br \/>\nHere \tname, \tbirthdate, \tand \tchildren \tare \tfunction \tsymbols. \t<br \/>\nVariables \tmay \trepresent \tatomic objects \t(i.e., \tconstants) \t<br \/>\nor \tcompound \tobjects. \tThe \tfollowing rule \trelates \tthe \tlast \t<br \/>\nnames  of \tall \tpersons  having \tthe  same  birthdate \tand  the \t<br \/>\nsame \tfirst \tname: \t<br \/>\nsimilar(X, \tY) \t:- \tperson(name(2, \tX), \tB, \tC), \t<br \/>\nperson(name(2, \tY), \tB, \t0). \t<br \/>\nBy \tthis \trule, \twe \tcan \tderive, \tfor \tinstance, \tthe \tnew \tfact \t<br \/>\nsimilar \t(berger, \tcoker \t). \t<br \/>\nTwo \tsets are \tconsidered  equal \tiff \tthey  contain \tthe \tsame \t<br \/>\nelements, \tindependently  of \tthe  order  in \twhich  these  ele- \t<br \/>\nments \tappear. \tThe \tfollowing \trule  defines  a  predicate \t<br \/>\neqchilds \t(X, \tY) \tstating  that \tX and \tY are \tthe \tnames  of \tper- \t<br \/>\nsons \twhose \tchildren have \texactly \tthe  same \tfirst \tnames: \t<br \/>\neqchilds \t(X, \tY) \t: - person \t(X, \tB, \tC ), person \t( Y, \tD, \tC ). \t<br \/>\nBy \tthis \trule \twe \tcan  derive, \tfor \tinstance, \teqchilds \t<br \/>\n( name \t( joe, \tberger \t) , name \t( bebe, \tsuong \t) ) . \t<br \/>\nLDL \toffers \tseveral  built-in  predicates  and  functions \tfor \t<br \/>\nhandling \tsets.  The \tmost \timportant \tare: \t<br \/>\nmember( \tt, S ), a built-in  predicate \tfor \texpressing that \t<br \/>\nt is an element \tof \tthe set \tS. \tNotice that \tt can \tbe \ta complex \t<br \/>\nterm \tand \tmay \tcontain \tsets \tas \tcomponents. \t<br \/>\nunion(S, \tA, \tB), \ta  built-in \tpredicate \tfor \texpressing \t<br \/>\nthe \tS = A \tU \tB. \t<br \/>\nSets can \tbe \tintroduced \tnot \tonly \tby \tenumeration \tbut \talso \t<br \/>\nby \tgrouping. \tGrouping allows \tus \tto \tdefine  a \tset \tin \ta rule \t<br \/>\nhead \tby \tindicating \tthe \tproperties  of \tits  elements \tin \tthe \t<br \/>\ncorresponding  rule \tbody. \tThe \tfollowing \trule, \tfor \tin- \t<br \/>\nstance, \tdefines \tthe  set  of \tall  persons  (identified \tby \ttheir \t<br \/>\nlast  name) \twho have \ta  child \tcalled Sarah. \t<br \/>\nsarahpar \t( &lt; X &gt; ) : - \t<br \/>\nperson(nume(A, \tX), \tB, \tC), \tmember(saruh, \tC). \t<br \/>\nThis \trule \tgenerates \tthe \tnew \tfact \tsarahpar( \t{ berger, \t<br \/>\ncoker, \tsuong \t} ). \tUsing \tcomplex  objects \tin \tDatalog  is \tnot \tas \teasy \tas \tit \t<br \/>\nmight \tappear. Several problems have \tto be \ttaken \tinto \tcon- \t<br \/>\nsideration. First \tof \tall, \tthe  use \tof  function \tsymbols \tmay \t<br \/>\nendanger \tthe \tsafety of programs.  It \tis undecidable  whether \t<br \/>\na  Datalog  program  with  function \tsymbols \thas  a \tfinite \tor \t<br \/>\nan  infinite  result. \tThe \tsimplest \tsolution \tis to leave the \tre- \t<br \/>\nsponsibility  to \tthe programmer. \tA similar \tproblem \tis the \t<br \/>\nfiniteness \tof  sets.  Furthermore, \tnot \tall \tDatalog \tprograms \t<br \/>\nwith \tsets have \ta well-defined \tsemantics. \tIn \tparticular, one \t<br \/>\nshould  avoid \tself-referential \tset  definitions  such \tas \t<br \/>\np ( &lt; X &gt; ) : - p (X \t). Such \tdefinitions \tcome \tclose \tto Rus- \t<br \/>\nsell\u2019s \tparadox. \tA large class \tof \tprograms free \tof  self-ref- \t<br \/>\nerence,  called \tadmissible \tprograms, \tis \tdefined \tin \t[21]. \t<br \/>\nNote also that \tthe \ttest \twhether two \tterms \t(or \tliterals)  in- \t<br \/>\nvolving \tsets \tmatch \tis \ta  computationally \thard \tproblem. \t<br \/>\nThis \tis a  particular \tcase \tof \ttheory \tunification \t[116]. \t<br \/>\nAnother \tinteresting  problem \tis the \tconsistency  problem \t<br \/>\nfor \tmonovalued \tdata functions. \tA  monovalued \tdata \tfunc- \t<br \/>\ntion \tf is  an  evaluable  function  symbol, \tinterpreted \tas \ta \t<br \/>\nmathematical \tfunction.  Such functions \tcan be \tdefined \tby \t<br \/>\nthe \trules  of  a \tlogic program.  However,  the \tunicity  of \tthe \t<br \/>\nfunction \tvalue \tmust \tbe \tensured. References \t[3], \t[76], \tand \t<br \/>\n[77] \tdeal \twith  this \tproblem. \t<br \/>\nA Logic Programming  language \tfor \tdata \tmanipulation \t<br \/>\nsuch as \tLDL \tshould \tbe \tconceived in  accordance \twith \tan \t<br \/>\nappropriate data \tmodel \twhich \tformalizes \tthe  storage \tand \t<br \/>\nretrieval  principles  and \tthe manipulation \tprimitives  that  a \t<br \/>\nDBMS  offers \tfor \tthe objects \treferenced  by \tthe language. \t<br \/>\nPure Datalog, \tfor \tinstance, can \tbe \tbased \ton \tthe \trelational \t<br \/>\nmodel \tin  first \tnormal form (nowadays often called \tthejat \t<br \/>\nrelational  model)  because \tthe  concept \tof  a \tliteral \tnicely \t<br \/>\nmatches \tthe \tone \tof  a \ttuple \tin \ta  relation  and  because \tthe \t<br \/>\nsingle evaluation steps \tof a \tDatalog \tprogram \tcan \tbe \ttrans- \t<br \/>\nlated into appropriate sequences \tof  relational \toperations. \t<br \/>\nOn the other hand, Logic Programming \tlanguages \tdealing \t<br \/>\nwith  structured \tobjects  such  as \tLDL \trequire  more  com- \t<br \/>\nplex \tdata models. <br \/>\nQuite  a \tnumber \tof \textensions \tof \tthe \trelational  model \t<br \/>\nhave been  developed \tin the \tlast \tyears \tin order \tto allow \tthe \t<br \/>\nstorage  and  manipulation  of \tcomplex  objects. \tThe \tmost \t<br \/>\nfamous  ones  are  the \tNF2 \tmodel \tby \tJaeschke \tand \tSchek \t<br \/>\n[66], \tthe model \tof \tnested  relations \tby \tFisher \tand \tThomas \t<br \/>\n[48], \tthe \tmodel \tof \tAbiteboul  and \tBeen \t[l], \t[2] \t(which \tis \t<br \/>\nmore \tgeneral than the former two models),  the  \u201cFranco- \t<br \/>\nArmenian \tData  model\u201d \tFAD \tby \tBancilhon,  Briggs, \t<br \/>\nKhoshafian, \tand \tValduriez \t[19] \t(based  on  a  calculus \tfor \t<br \/>\ncomplex \tobjects \tby \tBancilhon \tand \tKhoshafian \t[18]). \tALGRES, \ta quite \tpowerful  data-model \tfor \tcomplex  ob- \t<br \/>\njects, \tsupports \tan extended \trelational  algebra  augmented \t<br \/>\nwith \ta  fixpoint \toperator,  and \tthus \tis an \tideal \tbase \tfor \tim- \t<br \/>\nplementing  interpreters \tor \tcompilers \tfor \tlogic \tdata \tlan- \t<br \/>\nguages  with \tcomplex objects \t[30],  [31]. \t<br \/>\nAn \texcellent overview and  comparison \tof \tdata \tmodels \t<br \/>\nfor complex \tobjects \tis given \tin \t[ 11. \t<br \/>\nF. \tAn \tOverview \tof \tResearch \tPrototypes \t<br \/>\nWe \tmention  here \tsome \tof \tthe \tresearch  prototypes \tcur- \t<br \/>\nrently \tunder development \tbased  on \tthe language \tDatalog \t<br \/>\n(or \tits \tvariations). \t<br \/>\nThe \tLDL \tproject \tis  under  development \tat  Microelec- \t<br \/>\ntronics  and \tComputer  Technology  Corporation \t(MCC), \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERI \ter al.: \tDATALOG \tI63 \t<br \/>\nAustin, \tTX.  The \tproject\u2019s \tgoal \tis  to \timplement \tan \tinte- \t<br \/>\ngrated \tsystem  for processing  queries \tin \tLogic  Data  Lan- \t<br \/>\nguage \t(LDL), \ta language \twhich \textends  Datalog. \tWe \t<br \/>\ngave some examples \tof \tthe \tmost  significant \tconstructs \tof \t<br \/>\nLDL \tin \tSection \tVI. \tA general overview \tof \tthe \tLDL \tproj- \t<br \/>\nect can \tbe \tfound \tin \t[43]. \tFeatures \tfor \tdealing \twith \tcom- \t<br \/>\nplex \tterms \tin \tLDL \tare \tpresented \tby \tZaniolo [133]; \ta lan- \t<br \/>\nguage  overview \tis given \tby \tTsur \tand  Zaniolo  [119]; the \t<br \/>\ntreatment  of \tsets  and  negation \tis \tformally \tpresented \tby \t<br \/>\nBeeri \tet al. \t[21]; \tother papers \ton \tthe \tsubject \tare \t[19] and \t<br \/>\n[73]. \tAn \texcellent  description \tof \tthe \tLDL \tlanguage  and \t<br \/>\nof  related \tfeatures \tis given \tin \t[92]. \t<br \/>\nThe \tNAIL!  project \t(Not  Another  Implementation \tof \t<br \/>\nLogic!) \tis under development \tat  Stanford  University  with \t<br \/>\nthe  support  of \tNSF \tand \tIBM. \tNAIL! processes queries \tin \t<br \/>\nDatalog, \tbut \tinterfaces \ta conventional \tSQL \tdatabase sys- \t<br \/>\ntem (running on \tIBM \tPC\/RT). \tAn \toverview \tof \tthe \tNAIL! \t<br \/>\nproject  is \tpresented \tby \tMorris,  Ullman,  and  Van  Gelder \t<br \/>\nin \t[88]; \tan \tupdate can be found \tin \t[89]. \tVarious \tkinds \tof \t<br \/>\nalgorithms applied \tin \tthe  prototype \tare \tdiscussed \tin \t[90] \t<br \/>\nand \t[121]. <br \/>\nThe \tKIWI  Esprit  project,  sponsored \tby \tthe \tEEC, \tis  a \t<br \/>\njoint \teffort \tfor the development \tof \tknowledge \tbases, \tpro- \t<br \/>\ngrammed \tthrough \tan \tobject-oriented  language  (OOPS) \t, \t<br \/>\nand \tinterfaced  to \tan \texisting \trelational \tdatabase. \tThe \tAd- \t<br \/>\nvanced \tDatabase  Interface \t(ADE) \tof \tKIWI \tis developed \t<br \/>\njointly \tby \tthe \tUniversity  of \tCalabria, \tCRAI, \tand \tENI- \t<br \/>\nDATA (Italy). \tA general overview \tof \tKIWI and \tADE \tcan \t<br \/>\nbe \tfound \tin [ 106al. \tThe \tmini-magic \tvariation \tto the magic \t<br \/>\nset \tapproach, \tused  in \tKIWI, \tis  described \tby \tSacca\u2019  and \t<br \/>\nZaniolo \t[ 1041. \t<br \/>\nThe \tALGRES Project,  under development in the \tframe \t<br \/>\nof \tthe \tMETEOR \tEsprit project, \tis also  sponsored \tby \tthe \t<br \/>\nEEC. The \tALGRES \tproject \textends the \trelational \tmodel \t<br \/>\nto \tsupport \tnonnormalized  relations and \ta  fixpoint \topera- \t<br \/>\ntor,  and  supports  Datalog \tas \tprogramming  language. \t<br \/>\nALGRES is \ta joint \teffort \tof the \tPolitecnico \tdi \tMilano and \t<br \/>\nof \tTXT-Techint  (Italy). \tA \tgeneral  overview \tof \tthe \t<br \/>\nALGRES Project \tcan be \tfound in \ta paper \tby \tCeri, \tCrespi- \t<br \/>\nReghizzi \tet al. \t[32]. \tOther \tpapers \tare \t[30] and  [31]. \t<br \/>\nOther \trelevant \tresearch  projects \twhich \tdeal \twith  rule- \t<br \/>\nbased computations \tare POSTGRES, \tunder development \t<br \/>\nat Berkeley \tUniversity \t[ 1171, \tand \tthe \t5th \tGENERATION \t<br \/>\nProject \t[50], \t[91], \tunder development \tat the Institute \tfor \t<br \/>\nNew \tGeneration  Computer  Technology \t(ICOT),  Tokio, \t<br \/>\nJapan. \tMore \tdetails  about \tthe \tabove-mentioned  projects \t<br \/>\ncan \tbe found \tin \t[34]. \tOther \trecent \toverviews \tof \tthe \tproj- \t<br \/>\nects  on \tdatabases \tand \tlogic \twere  presented \tby \tZaniolo \t<br \/>\n[ 1341 \ton \ta dedicated issue \tof \tIEEE-Data \tEngineering and \t<br \/>\nby \tGardarin and Simon  [53] \ton \tTSI. \t<br \/>\nVII.  CONCLUSIONS \t<br \/>\nThere are \tno \tdoubts \tthat \tDatalog \ttheory  has \tbeen \tnicely \t<br \/>\ndeveloped in the last \tfive \tyears, \tthrough \tthe \tflourishing  of \t<br \/>\nmany \telegant  contributions. \tThe \tmain \tattraction \tof \tDa- \t<br \/>\ntalog  is \tthe \tpossibility  of \tdealing, \twithin  a \tunique  for- \t<br \/>\nmalism,  with nonrecursive expressions (or \tviews) \tas \twell \t<br \/>\nas \twith \trecursive \tones. \tAlthough \tthis  area \tis  still  very \t<br \/>\nactive,  we \tfeel  that \tsome \tbasic  understanding  has \tbeen \t<br \/>\nestablished,  thus allowing \tfor \tsystematic treatment. \t<br \/>\nOne \tof \tthe  major  challenges that  Datalog \tresearch \thas \t<br \/>\nstill to \tmeet is \tto convince the knowledge \tbase  community \t<br \/>\nof \tthe practical \tmerits  of  this \ttheory. \tThe \tweaknesses \tof \t<br \/>\nDatalog \twork \thave been  indicated \tas \tfollows. \t<br \/>\na) \tVery \tfew  applications  have  been  shown \twhich  can \t<br \/>\ntake  full \tadvantage \tof \tDatalog\u2019s expressive power. \tIn par- \t<br \/>\nticular,  no \tuseful \tapplications  have  been \treported \tso \tfar \t<br \/>\nfor \tnonlinear \tor \tmutually \trecursive rules. \t<br \/>\nb) \tDatalog  is \tnot \tconsidered \tas \ta programming \tlan- \t<br \/>\nguage,  but  rather \tas \ta \u201cpure\u201d \tcomputational  paradigm. \t<br \/>\nFor \tinstance,  Datalog  does \tnot \tprovide  support \tfor \twrit- \t<br \/>\ning \tuser\u2019s  interfaces,  and  does \tnot \tsupport  quite \tuseful \t<br \/>\nprogramming  tools,  such \tas \tmodularization \tand \tstruc- \t<br \/>\ntured types. <br \/>\nc) \tDatalog  does \tnot \tcompromise \tits \tclean  declarative \t<br \/>\nstyle \tin  any \tway;  while  sometimes \tit is  required  that \tthe \t<br \/>\nprogrammer \tmay \ttake control \ton \tinference processing, \tby \t<br \/>\nstating \tthe \torder \tand \tmethod \tof \texecution \tof \trules.  This \t<br \/>\nis typical, \tfor \tinstance, \tof \tmany \texpert systems shells. \t<br \/>\nd) \tDatalog  systems  have  been  considered, \tuntil \tnow, \t<br \/>\nas \tclosed \tworlds, \tthat \tdo \tnot \ttalk \tto \tother \tsystems; \twhile \t<br \/>\nthe \tcurrent trend is towards supporting heterogeneous sys- \t<br \/>\ntems. <br \/>\nSome \tof \tthe above  criticisms \tare \tin  fact \twell  founded, \t<br \/>\nand  provide \tan \tindication \tof \tthe  directions \tin \twhich \twe \t<br \/>\nexpect  Datalog \tto  move \tin \torder \tto \tbecome \tfully \tappli- \t<br \/>\ncable.  Datalog  research \twill \thave \tto \tconsider \twith  great \t<br \/>\ncare  the  advances \tin \tother  research \tareas; \tin \tparticular, \t<br \/>\nwe \thave  indicated \tin \tSection \tVI  that \tDatalog  can  be  ex- \t<br \/>\ntended \tto \tsupport  complex  terms; \tthis \tis \ta first \tstep \tto- \t<br \/>\nwards \tthe development \tof  new \tlanguage paradigms \twhich \t<br \/>\nuse some \tof \tthe concepts from object-oriented databases. \t<br \/>\nThe \tfoundation \tof \tsuch \tan \tevolution  have \talready  been \t<br \/>\nplaced \tby \tBeeri \t[23]  and \tAbiteboul \tand  Kanellakis  [5]; \t<br \/>\nthis  work \thas  shown  that \trule-based \tand  object-oriented \t<br \/>\napproaches \tare \tnot  in \topposition, \tbut \trather \tthey \tare \tca- \t<br \/>\npable \tof  providing \tuseful  programming  concepts \tto \teach \t<br \/>\nother. <br \/>\nIn \tsummary, \twe \texpect  that  supporting \trule \tcomputa- \t<br \/>\ntion  will \tbe \tone \tof \tthe \tingredients \tof \tfuture  knowledge \t<br \/>\nbase systems;  Datalog research has \tprovided \texact \tmeth- \t<br \/>\nods  and \ta  fairly \tgood understanding \tfor \tapproaching \tthis \t<br \/>\nissue. \t<br \/>\nREFERENCES \t<br \/>\nS. Abiteboul  and \tS. Gmmbach, \t\u201cBases \tde donnees \tet objets \tcom- \tplexes,\u201d \tTech. \tSci. \tInform., \tvol. \t6, no. \t5, 1987. \tS. Abiteboul  and \tC. \tBeeri, \t\u201cOn \tthe \tpower \tof languages  for \tthe \tmanipulation \tof complex \tobjects,\u201d \tin Proc. \tInr. \tWorkshop  Theory \tAppl. \tNested  Relations  Complex \tObjects, \tDarmstadt, \tWest \tGer- \tmany, \tabstract,  1987. \tS. Abiteboul and \tR. Hull, \t\u201cData \tfunctions, datalog \tand \tnegation,\u201d \tin Proc. \tACM-SIGMOD \tConf., \t1988. \tS. Abiteboul  and \tV. Vianu, \t\u201cProcedural \tand \tdeclarative database \tupdate \tlanguages,\u201d \tin Proc. \tACM  SIGMOD-SIGACT \tSymp. \tPrin- \tciples \tDatabase \tSysr., \t1988, pp. \t240-250. \tS. Abiteboul  and \tP. C. \tKanellakis, \t\u201cObject \tidentity \tas a query \tlanguage \tprimitive, \t\u201d manuscript,  1989. \tP. Aczel, \t\u201cAn \tintroduction \tto inductive definitions,\u201d \tin The \tHand- \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nI64 \tIEEE \tTRANSACTIONS \tON \tKNOWLEDGE  AND \tDATA \tENGINEERING, \tVOL. \tI, NO. \tI. MARCH \t1989 \t<br \/>\nbook \tof Mathematical \tLogic, \tJ. Banvise, \tEd. \tAmsterdam, \tThe \tNetherlands: \tNorth-Holland,  1977, pp.  739-782. \t[7] \tR. Agrawal \tand \tH. V. Jagadish,  \u201cDirect  algorithms for computing \tthe  transitive  closure  of  database  relations,\u201d \tin Proc. \t13th  Int. \tConf. \tVery \tLarge \tDatabases, \tBrighton, \tU.K., \t1987. \t[8] \t-, \t\u201cAlpha: \tAn \textension of relational algebra to express \ta class \tof recursive \tqueries,\u201d \tin Proc.  IEEE  3rd  Int. \tConf. \tData \tEng., \tLos \tAngeles, \tCA, \tFeb. \t1987. \t[9] \t-, \t\u201cMultiprocessor  transitive  closure \talgorithms,\u2019\u2019 \tData \tEng. \t(Special Issue \ton Databases \tfor \tParallel and Distributed Systems), \tvol.  12. Mar.  1989. \t[IO] \tA. \tV. \tAho  and \tJ. D. \tUllman, \t\u201cUniversity  of  data \tretrieval \tlan- \tguages,\u201d \tpresented \tat the \t6th \tACM  Symp. \tPrinciples  Program- \tming \tLanguages, \tSan \tAntonio, \tTX, \tJan. \t1979. \t[ 111 \tK. R. Apt and M. \tH. VanEniden,  \u201cContributions  to the \ttheory \tof \tlogic \tprogramming,\u201d \tJ. ACM, \tvol. 29, no. \t3, 1982. \t[ 121 \tC. Apt, \tH. Blair, and \tA. Walker, \t\u201cTowards \ta theory \tof declarative \tknowledge,\u201d \tIBMRes. \tRep. \tRC \t11681, \tApr. \t1986. \t[I31 \tF. Bancilhon, \t\u201cNaive \tevaluation \tof recursively  defined \trelations,\u201d \tin On \tKnowledge  Based  Management  Systems-Integrating  Data- \tbase \tand \tAI Systems, \tBrodie and  Mylopoulos,  Eds. \tNew \tYork: \tSpringer-Verlag,  1985. \t[I41 \tF. Bancilhon, \tD. Maier, \tY. Sagiv, \tand \tJ. D. Ullman,  \u201cMagic  sets \tand \tother  strange  ways  to  implement  logic \tprograms,\u201d \tin Proc. \tACM \tSIGMOD-SIGAT  Symp.  Principles  Database \tSyst., \tCam- \tbridge, \tMA, \tMar.  1986. \t[I51 \tF. Bancilhon \tand \tR. \tRamakrishnan, \t\u201cAn \tamateur\u2019s  introduction \tto  recursive \tquery \tprocessing,\u201d \tin Proc.  ACM-SIGMOD \tConf., \tMay \t1986. \t1161 \t-, \t\u201cPerformance  evaluation of data intensive logic \tprograms,\u201d \tin Foundations \tof Deductive \tDatabases  and \tLogic Programming, \tJ. Minker, \tEd. \t[17] \tF. Bancilhon, \tD. Maier, \tY. Sagiv, and \tJ. D. Ullman,  \u201cMagic  sets: \tAlgorithms \tand \texamples,\u201d \tmanuscript,  1986. \t[I81 \tF. Bancilhon \tand \tS. Khoshafian, \t\u201cA \tcalculus  for  complex  ob- \tjects,\u201d \tin Proc.  SIGMOD \t86, \t1986. \t[I91 \tF. Bancilhon, \tT. Briggs, \tS. Khoshafian, and \tP. Valduriez, \t\u201cFAD, \tA powerful and \tsimple database language,\u201d \tin Proc. \t13th \tInt.  Conf. \tVery \tLarge  Data \tBases, \tBrighton, \tU.K., \t1987. \t[20] \tR. Bayer,  \u201cQuery  evaluation \tand \trecursion in deductive database \tsystems,\u201d \tmanuscript, \t1985. \t[21] \tC. Beeri, \tet al., \t\u201cSets  and negation \tin  a \tlogical database language \t(LDLl),\u201d \tin Proc. \tACM \tSIGMOD-SIGACT \tSymp.  Principles \tDa- \ttabase \tSyst., \tSan Diego, \tCA, \tMar.  1987. \t[22] \tC. Beeri and \tR. Ramakrishnan, \t\u201cOn \tthe power \tof magic,\u201d  in \tProc. \tACM  SIGMOD-SIGACT  Symp.  Principles  Database \tSyst., \tSan \tDiego, \tCA, \tMar.  1987. \t[23] \tC. \tBeeri, \t\u201cData \tmodels \tand \tlanguages \tfor \tdatabases,\u201d  in \tProc. \t2nd \tInt. \tConf. \tDatabase \tTheory, \tBruges,  Belgium,  1988; \tand \tin \tLNCS \t326. \tNew \tYork:  Springer-Verlag.  1988. \t[24] \tJ. Bocca, \tH. \tDecker, \tJ.-M. \tNicolas, \tL. Vielle,  and \tM. \tWallace, \t\u201cSome \tsteps toward \ta DBMS-based \tKBMS,\u201d \tin Proc. \tIFIP \tWorld \tConf., \tDublin,  1986. \t[25] \tJ. Bocca, \t\u201cOn \tthe evaluation strategy of \tEDUCE,\u201d \tin Proc. \tACM- \tSIGMOD \tConf., \tWashington, \tDC, \tMay \t1986. \t[26] \tM. L. \tBrodie, \t\u201cFuture \tintelligent information systems: AI and da- \ttabase  technologies  working \ttogether,\u201d \tin Readings \tin Artijicial \tIntelligence \tand \tDatabases. \tSan \tMateo, CA: \tMorgan Kaufman, \t1988. \t[27] \tF. Bry,  \u201cLogic  programming as  constructivism: \tA formalization \tand \tits application to \tdatabases,\u201d \tin 8th  ACM  Symp.  Principles \tDatabase  Syst. \t(PODS), \tMar. \t1989, pp.  34-50. \t[28] \tS. Ceri, \tG. Gottlob,  and \tL. Lavazza,  \u201cTranslation \tand \toptimiza- \ttion \tof logic \tqueries: the  algebraic \tapproach,\u201d \tin Proc. \t12th  Int. \tConf. \tVery \tLarge  Data \tBases, \tKyoto,  Aug.  1986. \t[29] \tS. Ceri and \tL. Tanca, \t\u201cOptimization \tof systems \tof algebraic equa- \ttions \tfor \tevaluating \tDatalog \tqueries,\u201d \tin Proc.  13th Int. \tCon$ \tVery \tLarge  Data \tBases, \tBrighton, \tU.K., \tSept.  1987. \t[30] \tS. Ceri, \tS. Crespi Reghizzi, \tG. Gottlob, \tF. Lamperti, \tL. Lavazza, \tL. Tanca, \tand \tR. Zicari, \t\u201cThe  ALGRES \tproject,\u201d \tin Proc. \tInt. \tCon$ \tExtending \tDatabase \tTechnol. \t(EDBT88), \tVenice,  1988. \t[31] \tS. \tCeri, \tS. \tCrespi-Reghizzi, \tL. \tLavazza,  and \tR. \tZicari, \t\u201cALGRES: \tA system for the \tspecification \tand prototyping of \tcom- \tplex \tdatabases,\u201d Dip. \tElettronica, Politecnico \tdi Milano, \tInt. Rep. \t<br \/>\n[32] \tS. Ceri, \tS. Crespi-Reghizzi, \tG. Lamperti, \tL. Lavazza, and \tR. Zi- \tcari,  \u201cALGRES: \tAn  advanced  database  system \tfor \tcomplex  ap- \tplications,\u201d \tIEEE \tSoftware, \tto be published. \t<br \/>\nWashington, \tDC, \t1986. \t<br \/>\n87-018,  1987. \t<br \/>\n[33] \tS. Ceri, \tG. \tGottlob, \tand \tG. \tWiederhold, \t\u201cEfficient \tdatabase ac- \tcess \tthrough \tProlog,\u201d \tIEEE \tTrans.  Software \tEng., \tFeb. \t1989. \t[34] \tS. Ceri, \tG. \tGottlob,  and \tL. Tanca, \tLogic \tProgramming  and \tDa- \ttabases. \tNew \tYork:  Springer-Verlag, to \tbe published. \t[35]  U. \tS. Chakravarthy, \tJ. Minker,  and \tJ. Grant,  \u201cSemantic \tquery \toptimization:  Additional  constraints  and  control  strategies,\u201d \tin \tProc. \t1st \tInt. \tConf. \tExpert  Database \tSyst., \tL. Kerschberg, \tEd., \tCharleston,  1986; and \tin Expert \tDatabase Systems. \tMenlo Park, \tCA: \tBenjamin-Cummings,  1987. \t[36] \tU. S. Chakravarthy, \tJ. Grant, \tand \tJ. Minker,  \u201cFoundations \tof se- \tmantic  query \toptimization \tfor \tdeductive \tdatabases,\u201d \tin Proc.  Int. \tWorkshop Foundations  Deductive  Databases \tLogic Programming, \tJ. Minker, \tEd., \tAug. \t1986. \t[37]  A. \tChandra \tand \tD. \tHarel, \t\u201cHorn \tclause  queries and  generaliza- \ttions,\u201d \tJ. Logic Programming, \tvol. \t1, pp. \t1-15, \t1985. \t[38] \tA. Chandra, \t\u201cA \ttheory \tof database \tqueries,\u201d \tin Proc. \tACM \tSIG- \tMOD-SIGACT Symp. \tPrinciples  Database \tSyst., \tMar.  1988. \t[39] \tC. L. \tChang \tand \tR. C. Lee, \tSymbolic Logic \tand Mechanical \tTheo- \trem \tProving. \tNew \tYork:  Academic, \t1973. \t[40] \tC. C. Chang \tand \tH. J. Keisler, \tModel \tTheory. \tAmsterdam, \tThe \tNetherlands.  1977. \t[41] \tC. Chang, \u201cOn \tthe evaluation \tof queries containing derived \trela- \ttions \tin relational \tdatabases,\u201d \tin Advances  in \tDatabase \tTheory, \tVol. \tI, H. \tGallaire, \tJ. Minker, \tand \tJ. M. \tNicholas, \tEds. \tNew \tYork:  Plenum,  1981. \t[42] \tC. L. Chang \tand \tA. \tWalker, \u201cPROSQL: \tA Prolog programming \tinterface \twith \tSQLiDS,\u201d \tin Proc. \tFirst \tWorkshop \tExpert  Data- \tbase \tSyst., \tKiawah Island, \tSC, \tOct. \t1984; \tand \tin Expert Database \tSystems, \tL. Kerschberg, \tEd. \tMenlo \tPark, CA: \tBenjamin-Cum- \tmings,  1986. \t[43] \tD. \tChimenti, \tT. O\u2019Hare,  R. \tKrishnamurthy, \tS. Naqvi, \tS. Tsur, \tC. \tWest, \tand \tC. \tZaniolo, \t\u201cAn \toverview  of  the \tLDL \tsystem,\u201d \tSpecial \tIssue \ton \tDatabases \tand \tLogic, \tIEEE \tData \tEngineering, \tvol. \tIO, \tDec.  1987. \t[44] \tK. L. Clark, \u201cNegation \tas failure,\u201d \tin Logicand \tDatabases. \tNew \tYork:  Plenum,  1978. \t[45] \tW. \tF. Clocksin and \tC. S. Mellish, \tProgramming \tin Prolog. \tNew \tYork: Springer-Verlag, \t1981. \t[46] \tF. Cuppens and \tR. Demolombe, \t\u201cA \tPROLOG-relational  DBMS \tinterface using delayed \tevaluation,\u201d \tpresented  at \tthe Workshop on \tIntegration \tof Logic  Programming  and  Databases,  Venice,  Dec. \t1986. \t[47] \tP. Devanbu  and \tR. \tAgrawal, \t\u201cMoving \tselections  into \tfixpoint \tqueries,\u201d \tmanuscript, \tBell \tLabs, \tMurray \tHill, \tNJ, \t1986. \t[48] \tP. Fischer  and \tS. Thomas, \t\u201cOperators \tfor \tnon-first-normal-form \trelations,\u201d \tin Proc. \t7th \tInt. \tComput. \tSofrware \tAppl. \tConf., \tChi- \tcago, \tIL, \t1983. \t[49] \tM. \tFitting, \t\u201cA \tKripke-Kleene semantics \tfor \tlogic programs,\u201d \tJ. \tLogic Programming, \tvol. \t2, pp.  295-312, \t1985. \t[50] \tK. Fuchi, \t\u201cRevisiting  original philosophy \tof fifth \tgeneration com- \tputer \tproject,\u201d \tpresented  at \tthe \tInt. Conf. \tFifth Generation Com- \tput. \tSyst., \t1984. \t[SI] \tH. Gallaire and \tJ. Minker, \tEds, \tLogic \tand \tDatabases. \tNew \tYork: \tPlenum,  1978. \t[52] \tG. Gardarin and \tC. De \tMaindreville,  \u201cEvaluation  of database \tre- \tcursive \tlogic \tprograms as \trecurrent function \tseries,\u2019\u2019 \tin Proc. \tACM- \tSIGMOD \tConf., \tWashington, \tDC, \tMay \t1986. \t[53] \tG. Gardarin and \tE.  Simon, \u201cLes \tsystemes \tde gestion \tde bases \tse \tdonnees \tdeductives,\u201d \tTechnique  et Science \tInformatiques, \tvol. \t6, \t1987. \t[54] \tG. Gardarin,  \u201cMagic  functions: \tA technique \tto optimize extended \tdatalog recursive \tprograms,\u201d \tin Proc. \t13th \tConf. \tVery Large Da- \ttabases, \tBrighton, \tU.K., \t1987. \t[55] \tG. \tGentzen, \t\u201cDie \tWiederspruchsfreiheit \tder \treinen \tZahlentheo- \trie,\u201d \tMath. \tAnnalen, \tvol.  112, \tpp. 493-565,  1936. \t[56]  K. \tGodel,  \u201cDie \tVollstandigkeit \tder \tAxiome \tdes \tlogischen Funk- \ttionenkalkiils,\u201d \tMonatsheftefur \tMathematik \tund \tPhysik, \tvol. \t37, \t<br \/>\n[57] \t-, \t\u201cUber \tformal  unentscheidbare  Satze \tder \tPrincipia  Mathe- \tmatica  und \tvenvandter  Systeme \tI,\u201d \tMonatshefte \tfur \tMathematik \tund \tPhysik, \tvol. \t38, \tpp. \t173-198,  1931. \t[58] \tA. Guttman, \u201cNew \tfeatures for \trelational \tdatabase systems \tto sup- \tport \tCAD \tapplications,\u201d \tPh.D. \tdissertation,  Dep.  Comput. \tSci., \tUniv.  California, Berkeley, \tJune \t1984. \t[59] \tL. J. Henschen and \tS. A. Naqvi, \t\u201cOn \tcompiling queries \tin recur- \tsive \tfirst \torder \tdatabases,\u201d \tJ. ACM, \tvol. \t31, \tno. \t1, 1984. \t[60] \tD. Hilbert, \u201cAxiomatisches \tDenken,\u201d \tMathematische Annalen \t78, \tpp.  405-415,  1918. <br \/>\npp. \t349-360, \t1930. \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\nCERl \ter al.: \tDATALOG \t165 \t<br \/>\n(611 \tY. E. Ioannidis \tand \tE. Wong, \u201cAn \talgebraic \tapproach \tto recursive \tinference,\u201d  Electron.  Res. \tLab., \tUniv. \tCalifomia, \tBerkeley, \tInt. \tRep.  UCB\/ERL \tM85\/92, \t1985. \t[62] \tY. E. Ioannidis, \t\u201cOn \tthe \tcomputation  of  the  transitive \tclosure \tof \trelational \toperators,\u201d \tin Proc. \t12th  Int. \tCon$ \tVery Large \tData- \tbases, \tKyoto, \tJapan,  1986. \t[63] \tY. E. Ioannidis \tand \tE. Wong,  \u201cTransforming \tnon-linear  recursion \tinto linear recursion,\u201d  manuscript,  1987. \t[64] \tY. E. Ioannidis, \tJ. Chen, \tM. \tA. Friedman, \tand \tM. \tM. \tTsangaris, \t\u201cBERMUDA-An  architectural \tperspective \ton  interfacing  Prolog \tto a database \tmachine,\u201d \tDep. \tComput. \tSci.,  Univ. \tWisconsin, \tTech.  Rep. \t723, Oct. \t1987. \t[65] \tH. \tItoh, \t\u201cResearch  and development \ton  knowledge \tbase \tsystems \tat ICOT,\u201d \tin Proc. \t12th \tInt. Conf.  Very \tLarge \tData Bases, \tKyoto, \tJapan, \tAug. \t1986. \t[66]  B. \tJaeschke and \tH. J. Schek, \u201cRemarks \ton the algebra \tof non  first \tnormal \tform \trelations,\u201d \tin \tProc. \tACM \tSICMOD-SICACT \tSymp. \tPrinciples \tDatabase \tSyst., \tLos  Angeles, \tCA, \t1982, \t<br \/>\n[67] \tJ. F. Jenq and \tS. Sahni, \t\u201cAll \tpairs \tshortest  paths  on  a  hypercube \tmultiprocessor,\u201d \tin Proc. \tIEEE \tInf. \tCon$ \tParallel \tProcessing, \tAug. \t1987. \t[68] \tM. \tKifer  and \tE. L. \tLozinskii,  \u201cFiltering \tdata \tflow \tin deductive \tdatabases,\u201d \tin Proc. \t1st \tInt. \tCon$ \tDatabase \tTheory, \tRoma, \tSept. \t1986. \t[69] \tW. \tKim, \tD. \tS. Reiner, \tand \tD. \tS. Batory, \tQuery  Processing \tin \tDatabase  Systems. \t[70] \tJ. King, \t\u201cQuist: \tA system \tfor  semantic \tquery \toptimization \tin re- \tlational \tdatabases,\u201d \tin Proc. \t7th \tInt. \tCon$ \tVery \tLarge \tData \tBases, \tCannes, \t1981. \t[71] \tPh.G. \tKolaitis, \t\u201cOn \tthe \texpressive \tpower \tof stratified \tdatalog \tpro- \tgrams,\u201d \tStanford \tUniv., \tStanford, \tCA, \tpreprint, \tNov. \t1987. \t[72] \tPh.G. \tKolaitis \tand \tCh.H. \tPapadimitriou, \t\u201cWhy \tnot \tnegation \tby \tfixpoint?,\u201d \tin Proc. \tACM \tSIGMOD-SICACT Symp. Principles \tDa- \ttabase \tSyst., \t1988, pp. \t231-239. \t[73] \tR. Krishnamurthy \tand \tC. Zaniolo, \t\u201cOptimization \tin a  logic based \tlanguage \tfor \tknowledge \tand \tdata \tintensive \tapplications,\u201d \tin Proc. \tInt. \tCon5  Extending \tDatabase \tTechnol. \t(EDBT88). \tVenice,  1988; \tand \tLNCS, \tNo. 303. \t1741 \tK. Kunen,  \u201cNegation \tin logic \tprogramming,\u201d \tJ. Logic  Program- \tming, \tvol. \t4, pp. \t289-308, \t1987. \t[75] \tG.  M. \tKuper,  \u201cLogic  programming \twith \tsets,\u201d \tin Proc. \tACM \tSICMOD-SICACT Symp. Principles \tDatabase \tSyst., \t1987, pp. \t11- \t20. \t[76] \tE.  Lambrichts, \tP. Nees, \tJ. Paredaens, \tP. Peelman, \tand \tL. Tanca, \t\u201cMilAnt: \tAn extension \tof datalog \twith \tcomplex \tobjects, functions \tand \tnegation,\u201d \tDep. \tComput.  Sci., \tUniv. \tAntwerp, \tInt. \tRep., \t1988. \t[77] \tE.  Lambrichts, \tP. Nees, \tJ. Paredaens, \tP. Peelman, \tand \tL. Tanca, \t\u201cIntegration \tof functions \tin the \tfixpoint \tsemantics \tof  rule  based \tsystems,\u201d \tin Proc. \t2nd \tSymp. \tMath. \tFundamentals \tDatabase \tThe- \tory, \tVisegrad, \tHungary,  June \t1989; \tand \tLNCS. \tNew  York: \tSpringer-Verlag,  1989. \t[78] \tJ.  W. \tLloyd, \tFoundations \tof Logic  Programming, \t2nd  extended \t<br \/>\n1791 \tL. Lowenheim, \t\u201cUber \tMoglichkeiten \tim Relativkalkul,\u201d \tMathe- \tmathische \tAnnalen, \tvol. \t76, \tpp. \t447-470, \t1915. \t[80] \tD. \tW. \tLoveland, \tAutomafed \tTheorem Proving:  A  Logical \tBasis. \tAmsterdam, \tThe \tNetherlands: \tNorth-Holland, \t1978. \t[81] \tE.  Lozinskii, \t\u201cEvaluating \tqueries \tin deductive databases \tby gen- \terating,\u201d \tin Proc. \tInt. \tJoint \tCon$ \tArt$cial \tIntell., \t1985. \t[82] \tD. Maier and \tD. S. Warren, \tComputing  with \tLogic. \tMenlo \tPark, \tCA: Benjamin-Cummings, \t1988. \t[83] \tA. \tI. Malcev, \tAlgebraic \tSystems. \tNew  York: \tSpringer-Verlag, \t1973. \t[84] \tV.  Mannino, \tP. Chu, \tand \tT. Sager, \t\u201cStatistical  profile \testimation \tin database \tsystems,\u201d \tACM \tComput. Surveys, \tvol. \t20, \tSept.  1988. \t[85] \tD. \tMcKay \tand \tS. Shapiro, \t\u201cUsing \tactive \tconnection \tgraphs  for \treasoning  with \trecursive rules,\u2019\u2019 \tin Proc. \t7th Int.  Joint \tCon$ \tAr- \t@cia1 \tInfell., \t1981. \t[86] \tG. Marque-Pucheu, \u201cAlgebraic \tstructure \tof answers \tin a recursive \tlogic data-base,\u2019\u2019 \tActa \tInform., \t1983. \t[87] \tG. \tMarque-Pucheu, \tJ. M. \tGallausiaux, \tand \tG. Jomier;  \u201clnterfac- \ting \tProlog \tand \trelational \tdatabase \tmanagement \tsystems,\u201d \tNew \tAp- \tplications \tofDatabases, \tGardarin \tand \tGelenbe, \tEds. \tNew  York: \tAcademic, \t1984. \t[88] \tK. Morris, \tJ. D. \tUllman,  and A. \tVan \tGelder, \t\u201cDesign \toverview \t<br \/>\npp. \t124-138. \t<br \/>\nNew \tYork: \tSpringer-Verlag,  1985. \t<br \/>\nNew \tYork: \tSpringer-Verlag,  1988. \t<br \/>\ned. \tNew  York: \tSpringer-Verlag,  1987. \tI \t<br \/>\nof the \tNail! \tsystem,\u201d \tin Proc. \tInt. \tCon5 \tLogic  Programming. \tNew \tYork: Academic, \t1986. \t[89] \tK. Moms, \tJ. Naughton, Y. \tSaraiya, \tJ. Ullman, \tand \tA. Van \tGelder, \t\u201cYAWN! \t(Yet \tanother \twindow \ton NAIL!).\u201d  Special \tIssue \ton Da- \ttabases \tand \tLogic, \tIEEE \tDatu \tEng., \tvol.  10, \tDec. \t1987. \t[90]  K. \tA. \tMorris, \t\u201cAn \talgorithm \tfor \tordering subgoals \tin Nail!,\u201d \tin \tProc. \tACM  SIGNOD-SIGACT \tSymp.  Principles  Database \tSyst., \tAustin, \tTX, \t1988. \t[91] \tK. \tMurakami, \tT. \tKakuta,  N.  Miyazaki, \tS. Shibayama, \tand \tH. \tYokota,  \u201cA \trelational \tdatabase \tmachine, \tfirst \tstep  to \tknowledge \tbase \tmachine,\u201d \tin Proc. \tloth \tSymp. \tComput. \tArchitecfure, \tJune \t1983. \t[92] \tS. Naqvi \tand \tS. Tsur, \tA  Logical \tLanguage \tfor \tData \tund \tKnowl- \tedge Bases. \t[93] \tW. \tNejdl, \t\u201cRecursive \tstrategies \tfor \tanswering \trecursive queries- \tthe \tRQA\/FQI \tstrategy,\u201d \tin Proc. \t13th \tInt. \tCon5  Very \tLarge Data \tBases, \tBrighton, \tU.K., \tSept.  1987. \t[94] \tE. Neuhold  and \tM. \tStonebraker, \t\u201cFuture \tdirections \tin DBMS \tre- \tsearch,\u201d  Int. \tComput. \tSci. \tInst., \tBerkeley, \tCA, \tTR \t88-01, \tMay \t1988. \t1951 \tT. Przymusinski, \t\u201cOn \tthe \tsemantics \tof stratified \tdeductive \tdata- \tbases,\u201d \tin Proc. \tWorkshop \tFoundations \tDeductive  Databases \tLogic Programming, \tWashington, \tDC, \t1986, \tpp, \t433-443. \t[96] \t-, \t\u201cEvery \tlogic \tprogram \thas \ta  natural  stratification \tand  an \tit- \terated  least \tfixed \tpoint \tmodel,\u201d \tin 8th \tACM \tSymp. \tPrinciples \tDa- \ttabase \tSyst. \t(PODS), \tMar. \t1989, \tpp. \t11-21, \t1971 \tQuintus \tComputer  Systems \tInc., \tMountain \tView, \tCA, \tQuinrus \tProlog \tData \tBase \tInterface \tManual, \tversion \t1, June \t29.  1987. \t[98] \tR. \tRamakrishnan, \tC. \tBeeri, \tand \tR. \tKnshnamurty.  \u201cOptimizing \texistential \tDatalog \tqueries,\u201d \tin Proc. \tACM \tSIGMOD-SICACT \tSymp. \tPrinciples \tof Database \tSyst., \tAustin, \tTX, \tMar. \t1988. \t[99] \t-, \t\u201cMagic \ttemplates, \tA spellbinding \tapproach \tto logic \tevalua- \ttion,\u201d \tin Proc. \tLogic \tProgramming \tConf., \tAug. \t1988. \t[IOO] \tD. \tRoelants, \t\u201cRecursive \trules \tin logic \tdatabases,\u201d \tPhilips \tRes. \tLab., \tBruxelles, \tRep. \tR513, Mar. \t1987, \tsubmitted \tfor \tpublication. \t[loll \tJ. Rohmer, \tR.  Lescoeur, \tand \tJ. M. \tKerisit, \t\u201cThe \tAlexander \tmethod: \ttechnique \tfor \tthe  processing  of  recursive \taxioms \tin de- \tductive \tdatabases,\u201d \tin New  Generation Computing, \tvol. \t4. \tNew \tYork: \tSpringer-Verlag,  1986. \t[I021 \tA. \tRosenthal, \tS. Heiler, \tU. \tDayal, \tand \tF. Manola, \t\u201cTraversal \trecursion:  A  practical \tapproach \tto supporting  recursive \tapplica- \ttions,\u201d \tin Proc. \tACM \tSIGMOD \t1986 \tInt. \tCotif: \tManagement \tof \tData, \tWashington, \tDC, \tMay \t1986. \t[IO31 \tK. A. Ross, \t\u201cA \tprocedural \tsemantics for \twell founded  negation \tin \tlogic \tprograms,\u201d \tin 8th \tACM \tSymp.  Principles \tDatabase \tSyst. \t(PODS), \tMar.  1989, \tpp. \t22-32. \t[IO41 \tD. \tSacca\u2019 \tand \tC. \tZaniolo, \t\u201cOn \tthe  implementation \tof a  simple \tclass \tof  logic \tqueries \tfor \tdatabases,\u201d \tin Proc. \tACM \tI986 \tSIC- \tMOD-SIGACT Symp.  Principles \tDatabase \tSyst., \tCambridge. \tMA, \tMar. \t1986. \t[ 1051 \t-, \t\u201cImplementing \trecursive  logic \tqueries \twith  function \tsym- \tbols,\u201d \tMCC Tech.  Rep.  DB-401-86, Dec. \t1986. \t[lo61 \t-, \t\u201cMagic \tcounting \tmethods,\u201d \tin Pror. \tACM-SIGMOD \tCon.f., \tSan \tFrancisco, \tCA, \tMay \t1987. \t[106a] \tD. Sacca, \tM. \tDispinziezi, \tA. Mecchia, \tC. Pizzuti, \tC. Del \tGracco, \tand \tP. Naggar, \t\u201cThe \tadvanced \tdatabase \tenvironment \tof the KIWI \tsystem,\u201d \tSpecial \tIssue \ton \tDatabases \tand \tLogic, \tIEEE \tData \tEn- \tgineering, \tvol. \t10, \tno. \t4, Dec. \t1987. \t[lo71 \tY. \tSagiv, \t\u201cOptimizing  Datalog programs,\u201d \tin Proc. \tACM \t1987 \tSIGMOD-SIGACT \tSymp. \tPrinciples \tDatabase \tSyst., \tSan \tDiego, \tCA, \tMar. \t1987. \t[IO81 \tL. Schmitz, \t\u201cAn \timproved \ttransitive  closure  algorithm,\u201d \tCom- \tput., \tvol. \t30, \t1983. \t[I091 \tC. \tP. Schnorr, \t\u201cAn \talgorithm \tfor \ttransitive \tclosure \twith \tlinear \texpected \ttime,\u201d \tSIAM \tJ. Comput., \tvol. \t7, May \t1978. \t[110] \tJ. C. \tShepherdson, \u201cNegation \tas Failure \t11,\u201d \tJ. Logic \tPrograrn- \tming, \tvol. \t2, no. \t3, pp. \t185-202, \t1985. \t[Ill] \t-, \t\u201cNegation \tin logic \tprogramming,\u201d \tin Foundations \tof De- \tductive Databases  and Logic  Programming, \tJ. Minker, \tEd. \tLos \t<br \/>\n[I121 \t0. \tShmueli, and \tSh. \tNaqvi,  \u201cSet  grouping \tand  layering \tin Horn \tclause \tprograms,\u201d \tin Proc. \tInt. \tCon5 \tLogic Programming, \t1987, \t<br \/>\n[I 131 \tT. \tSkolem, \t\u201cLogisch-kombinatorische \tUntersuchungen \tiiber \tdie \tErfullbarkeit \toder \tBeweisbarkeit \tmathematischer \tSatze \tnebst \tei- \tnem \tTheoreme uber \tdichte \tMengen,\u201d \tSkrifter \tutRit \tav Videnskaps- \tselskapet  i \tKristiana, \tI, Mat-Nut. \tKlasse, \tno. \t4, 1920. \t<br \/>\nNew \tYork: Computer \tScience Press,  1989. \t<br \/>\nAltos, \tCA, \t1988, pp. \t19-88. \t<br \/>\npp. \t152-177. \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.<br \/>\n<br \/>\n166 \tIEEE \tTRANSACTIONS \tON  KNOWLEDGE \tAND \tDATA \tENGINEERING, \tVOL. \tI, NO. \tI, MARCH \t1989 \t<br \/>\n[I141 \tD. E. Smith, \tM. \tR. Genesereth, \tand \tM. L. \tGinsberg,  \u201cControlling \trecursive \tinference,\u201d \tArti$cal \tIntell., \tvol. \t30, \tno. \t3, 1986. \t[I 151 \tL. Sterling and \tE. Shapiro, \tThe \tArt \tof Prolog. \tCambridge, MA: \tM.I.T. \tPress,  1986. \t[I 161 \tM. \tE.  Stickel, \t\u201cA \tunification \talgorithm for associative \tcommuta- \ttive \tfunctions,\u201d \tJACM, \tvol. \t28, \tJuly \t1981. \t(1171 \tM. \tStonebraker \tand \tL. A. \tRowe, \tEds., \t\u201cThe \tPostgres \tpapers,\u201d \tUniv.  Calif, \tBerkeley, \tMem.  UCB\/ERL \tM86\/86, \tJune  1987 \t(re- \tvised \tversion). \t[I 181 \tL. \tTanca,  \u201cOptimization \tof recursive \tlogic  queries  to \trelational \tdatabases\u201d \t(in \tItalian) \tPh.D. \tdissertation, \tPolitecnico \tdi Milano \tand \tUniversita\u2019 \tdi Napoli, \t1988. \t[I 191 \tS. Tsur \tand \tC. \tZaniolo, \t\u201cLDL: \tA  logic-based  query \tlanguage,\u201d \tin Proc.  12th \tInt. \tCon6 \tVery \tLarge  Data \tBases, \tKyoto, \tJapan, \t1986. \t[I201 \tJ. D. \tUllman,  \u201cImplementation \tof  logic  query  languages \tfor \tda- \ttabases,\u201d \tACM  Trans. \tDatabase \tSyst., \tvol. \t10, \tno. \t3, 1985. \t11211 \tJ. D. \tUllman  and \tA. \tVan \tGelder, \t\u201cTesting \tapplicability \tof top- \tdown \tcapture  rules,\u201d \tStanford \tUniv.,  Stanford, \tCA, \tInt. \tRep. \tSTAN-CS-85.1046;  and \tACM-J., \tto be published. \t[ 1221 \tJ. D. \tUllman, \tPrinciples \tof Databases \tand \tKnowledge-Base Sys- \ttems, \tVolume \tI. \t[I231 \tP.  Valduriez  and \tH. Boral, \t\u201cEvaluation \tof recursive queries using \tjoin  indices,\u201d \tin Proc. \t1st \tInt. \tCon5 \tExpert  Database \tSyst., \tCharleston, \tSC. \t1986. \t[I241 \tP. Valduriez  and \tS. Khoshafian, \t\u201cParallel \tevaluation  of  the  tran- \tsitive  closure  of  a \tdatabase  relation.\u201d \tInt. \tJ. Parallel  Program- \tming, \tvol. \t17, \tFeb.  1988. \t[ 1251 \tM. \tVan \tEmden and \tR. Kowalski, \t\u201cThe \tsemantics \tof predicate logic \tas a programming \tlanguage,\u201d \tJ. ACM, \tvol. \t4, Oct.  1976. \t[I261 \tA.  Van \tGelder, \tA. \tRoss, \tand \tJ. S. Schlipf, \t\u201cThe \twell-founded \tsemantics \tfor \tgeneral  logic \tprograms,\u201d \tin 7th \tACM  Symp. \tPrin- \tciples Database  Syst. \t(PODS), \tMar. \t1988, pp. \t221-230. \t[I271 \tA. \tVan \tGelder. \t\u201cThe \talternating  fixpoint of  logic \tprograms \twith \tnegation,\u201d \tin 8th ACM  Symp.  Principles  Darabase Syst. \t(PODS), \tMar. \t1989, \tpp.  1-10, \t[I281 \tL. Vieille, \t\u201cRecursive axioms \tin deductive databases: \tThe \tQuery- \tSubquery \tapproach,\u201d \tin Proc. \tInt. \tCon6 \tExpert  Database \tSyst., \tL. Kerschberg, \tEd., \tCharleston,  1986. \t[129] \t-, \t\u201cA \tdatabase \tcomplete \tproof  procedure  based  on \tSLD \treso- \tlution,\u201d \tECRC,  Munich, \tWest \tGermany, \tInt. \tRep. \tIR-KB-40, \tNov. \t1986. \t[I301 \t-, \t\u201cFrom \tQSQ \tto QoSaQ:  Global \toptimization \tof recursive \tqueries,\u201d \tin Proc. \t2nd \tInt. \tCon5 \tExpert  Database \tSyst., \tL. \tKerschberg, \tEd., \tTyson \tCorner,  1988. \t11311 \tH. \tS. Warren, \t\u201cA \tmodification \tof Warshall\u2019s \talgorithm \tfor  the \ttransitive \tclosure \tof binary \trelations,\u201d \tCommun. \tACM, \tvol. \t18, \tApr.  1975. \t[I321 \tS. Warshall, \t\u201cA \ttheorem  on  boolean \tmatrices,\u201d \tJ. ACM, \tvol. \t9, \tJune \t1962. \t[I331 \tC. \tZaniolo, \t\u201cThe \trepresentation  and \tdeductive \tretrieval \tof com- \tplex \tobjects,\u201d \tin Proc. \t11th  Int.  Con6 \tVery \tLarge  Data \tBases, \tAug. \t1985. \t[I341 \t-, \tSpecial Issue on  Databases  and \tLogic, \tIEEE Data \tEng., \tvol. \t10, \tDec. \t1987. \t[I351 \tC. \tZaniolo \tand \tD. \tSacca\u2019, \t\u201cRule \trewriting  methods \tfor \tefficient \tevaluation \tof Horn \tlogic,\u201d \tMCC Tech.  Rep. DB-084-87, \t1987. \t[I361 \tM. \tM. \tZloof, \t\u201cQuery-by-example: \tOperations \ton  the  transitive \tclosure,\u201d \tIBM, \tYorktown \tHeights, \tNY, \tRC \t5526, \t1975. \t<br \/>\nPotomac, MD: Computer \tScience,  1988. \t<br \/>\nStefano \tCeri \tIS a Professor \tof Computer \tScience \tat the \tDipartimento \tdi Matematica, \tUniversity \tof \tModena, Modena, \tItaly, \tand \tVisiting \tProfessor \tat \tStanford \tUniversity,  Stanford, \tCA, \tduring  the \tsummer \tterms \tUntil  1986 he was with the Dipar- \ttimento \tdi  Elettronica, \tPolitecnico \tdi \tMilano, \tMilan,  Italy. \tHis  research  interests  include  dis- \ttributed \tdatabases,  deductive \tand  object-oriented \tdatabases,  database  design, \tmedical \tdatabases, \tand \tthe \tuse \tof databases \tin software engineering \tHP \tis the author of numerous articles \tin these areas \tand \tcoauthor \tof the \tbook, \tDistributed \tDatabases, \tPrinciples \tand \tSystems \t(New \tYork:  McGraw-Hill). He \thas  been  an \tactive \tparticipant  of  several \tjoint \tprojects  between \tthe \tuniversity  and \tindustry, \tsponsored \tby the \tNa- \ttional  Research  Council  of \tItaly, \tthe \tNSF, \tand \tEsprit \t(EEC) \tDr \tCeri \tis a member \tof ACM, \tVice-chairman \tof the \tIEEE \tTechnical \tCommittee on \tData \tEngineering, and Associate \tEditor \tof the \tjournals, \tACM \tTrunm-trons \ton Database  Svstems \tand \tDistributed  Computing \t<br \/>\nGeorg \tGottlob \treceived \tthe \tDipl.-Ing. \tdegree and \tthe \tDoctorate \tin computer \tscience  from the \tTech- \tnical \tUniversity \tof Vienna, Vienna, \tAustria. \tFrom \t1982 \tto 1984 he served \tas a Research \tAs- \tsociate \tat the \tPolitecnico \tdi Milano, \tMilan, Italy, \tand \tfrom \t1985 \tto 1988 he  was a \tStaff\u2018 \tScientist \tat \tthe Institute \tfor \tApplied \tMathematics, \tItalian \tNa- \ttional Research Council \t(C.N.R.), \tGenoa. \tDuring \tthe \tSummers \tof 1985 \tand \t1987, \the  lectured  and \tperformed  research \tat Stanford \tUniversity, \tStan- \tford. \tCA. \tPresentlv he \tis a Professor of \tComputer \tScience \tat the \tTechnical \tUniversity  of \tVienna, \twhere  he \tdirects \tthe  Data- \tbase \tand \tExpert-System \tSubdivision. \tHis \tresearch interests \tare \tin the \tfields \tof databases,  expert-systems, \tand \tapplied  mathematical \tlogic. \tDr.  Gottlob \tis a member \tof ACM, \tthe \tIEEE \tComputer \tSociety, \tand the \tKurt \tGoedel \tSociety. \t<br \/>\nLetizia \tTanca \treceived  the \tPh.D. \tdegree \tin ap- \tplied  mathematics  and \tcomputer \tscience from the \tPolitecnico di \tMilano, \tMilan,  Italy, \tin 1988. \tPrior to \tworking towards \tthe \tPh.D. \tdegree, she \tworked \tin industry \tas a  Software  Designer. \tAt \tpresent \tshe \tis Postdoctorate  Fellow \tat the \tPolitec- \tnico \tdi Milano.  Her \tmain  research  interests \tare \tthe \ttreatment \tof  negative \tand \tfunctional  informa- \ttion \tin relational  and \tdeductive databases \tand \tex- \ttensions \tto  the \trelational  data \tmodel \t(in \tMilan) \twithin  the  proiect  Algres \t(an \textended  relational \t-- \tenvironment \tfor \tmanipulating \tcomplex \tobjects). \tShe \tis also  collaborating \twith  Prof.  Paredaens  of  the  University \tof Antwerp (UIA), \ton  the  project \tof a  rule-based \tlanguage \tfor \tmanipulating \tcomplex  objects  and \tfunctions. \tShe \thas written a book on the integration \tof Relational Databases \tandLogic \tProgramming, \tcoauthored \twith \tS. Ceri \tand \tG. Gottlob  (New \tYork: \tSprin- \tger-Verlag, to \tbe published). \t<br \/>\nAuthorized licensed use limited to: University of Minnesota. Downloaded on February 17,2010 at 18:13:49 EST from IEEE Xplore.  Restrictions apply.","created_at":"2019-05-26T22:38:35+0000","updated_at":"2019-05-26T22:38:35+0000","published_at":"2013-01-02T15:01:33+0000","published_by":["IEEE"],"annotations":[],"mimetype":"application\/pdf","reading_time":105,"domain_name":"www.lifl.fr","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["datalog"],"is_public":false,"id":960,"title":"frankmcsherry\/blog","url":"https:\/\/github.com\/frankmcsherry\/blog\/blob\/master\/posts\/2018-05-19.md","content":"<p>In this post we will build up a relatively concise <a href=\"https:\/\/en.wikipedia.org\/wiki\/Datalog\" rel=\"nofollow\">Datalog<\/a> engine: <a href=\"https:\/\/github.com\/frankmcsherry\/datafrog\">DataFrog<\/a>.<\/p>\n<p>I initially wrote \"simple\", and I think that is nearly the case, but I'll have to let you be the judge. One of the design goals is that simple or not, there should be little enough in the way of someone coming to understand how it all works, with enough time. No complicated runtime, or weird abstractions and such. Or, less weird, anyhow.<\/p>\n<h3><a id=\"user-content-acks\" class=\"anchor\" aria-hidden=\"true\" href=\"#acks\"><\/a>Acks<\/h3>\n<p>This project got put together rather suddenly, in response to <a href=\"https:\/\/github.com\/rust-lang-nursery\/polonius\/pull\/36#issuecomment-390393717\">some work the Rust folks are doing<\/a> on their new and improved borrow checker. They had been using <a href=\"https:\/\/github.com\/frankmcsherry\/differential-dataflow\">differential dataflow<\/a>, which was super brave of them, but it seemed like a more stream-lined Datalog implementation might suit them better for now.<\/p>\n<p>Thanks go out to <a href=\"https:\/\/twitter.com\/lqd\" rel=\"nofollow\">@lqd<\/a>, <a href=\"https:\/\/twitter.com\/nikomatsakis\" rel=\"nofollow\">@nikomatsakis<\/a>, and <a href=\"https:\/\/twitter.com\/qmx\" rel=\"nofollow\">@qmx<\/a> for moving things along (and testing, and listening, and explaining how it could be better, etc).<\/p>\n<h3><a id=\"user-content-datalog\" class=\"anchor\" aria-hidden=\"true\" href=\"#datalog\"><\/a>Datalog<\/h3>\n<p>Datalog is a recently re-popularized language from decades backs, in which you start with a few collections of \"tuples\", often called \"facts\" in Datalog, and repeatedly apply a set of rules that derive new tuples from existing tuples. These rules all have a restricted form, where you can insist that some tuple values in different collections be the same, and only when they match do they produce a new tuple.<\/p>\n<p>For example, perhaps we have two collections <code>nodes<\/code> and <code>edges<\/code>, where <code>nodes<\/code> contains <code>usize<\/code> tuples and <code>edges<\/code> contains <code>(usize, usize)<\/code> tuples. We might write a rule<\/p>\n<pre>nodes(y) &lt;- nodes(x), edges(x,y)\n<\/pre>\n<p>which says that whenever we have an element of <code>nodes<\/code> and elements of <code>edges<\/code> that match in their first coordinate, we can add the second coordinate to <code>nodes<\/code>.<\/p>\n<h3><a id=\"user-content-relations\" class=\"anchor\" aria-hidden=\"true\" href=\"#relations\"><\/a>Relations<\/h3>\n<p>Our engine is going to start with a definition of a <code>Relation&lt;Tuple&gt;<\/code>, which will be a sorted list of distinct elements of type <code>Tuple<\/code>.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>\/\/\/ A sorted list of distinct tuples.\npub struct Relation&lt;Tuple: Ord&gt; {\n    elements: Vec&lt;Tuple&gt;\n}<\/pre><\/div>\n<p>This type is not much different than a <code>Vec&lt;Tuple&gt;<\/code>, which is just a list of <code>Tuple<\/code> elements. The main difference is that it will guarantee that its elements are sorted and distinct. We could ensure this by providing a very limited constructor:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>impl&lt;Tuple: Ord&gt; From&lt;Vec&lt;Tuple&gt;&gt; for Relation&lt;Tuple&gt; {\n    fn from(mut elements: Vec&lt;Tuple&gt;) -&gt; Self {\n        elements.sort_unstable();\n        elements.dedup();\n        Relation { elements }\n    }\n}<\/pre><\/div>\n<p>This explains how to create a relation from a <code>Vec&lt;Tuple&gt;<\/code>, which is essentially the same thing minus the guarantee about ordering and distinctness. As you can see, we perform the sorting and deduplication. Fortunately, if the input is already sorted and deduplicated this is relatively cheap.<\/p>\n<p>In fact we will generalize this a bit, so that we can take in input types other than just <code>Vec&lt;Tuple&gt;<\/code>. We should be able to take anything that be enumerated as a list of <code>Tuple<\/code> elements, which we can express using the <code>IntoIterator<\/code> trait as a constraint:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>impl&lt;Tuple: Ord, I: IntoIterator&lt;Item=Tuple&gt;&gt; From&lt;I&gt; for Relation&lt;Tuple&gt; {\n    fn from(iterator: I) -&gt; Self {\n        let mut elements: Vec&lt;Tuple&gt; = iterator.into_iter().collect();\n        elements.sort_unstable();\n        elements.dedup();\n        Relation { elements }\n    }\n}<\/pre><\/div>\n<p>This implementation will allow us to use not only vectors, but also other iterators that might transform or restrict a set of tuples.<\/p>\n<p>In the interest of completeness, we will have one more method for the <code>Relation<\/code> type, a method <code>merge<\/code> that consumes two relations and produces their union:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>impl&lt;Tuple: Ord&gt; Relation&lt;Tuple&gt; {\n    \/\/\/ Merges two relations into their union.\n    pub fn merge(self, other: Self) -&gt; Self {\n        let mut elements = Vec::with_capacity(self.elements.len() + other.elements.len());\n        elements.extend(self.elements.into_iter());\n        elements.extend(other.elements.into_iter());\n        elements.into()\n    }\n}<\/pre><\/div>\n<p>The <code>into()<\/code> method is provided by Rust for any type implementing <code>From<\/code> (as defined just above). This fairly primitive method just concatenates the two relations and the sorts and deduplicates them (in <code>from()<\/code>, by way of <code>into()<\/code>).<\/p>\n<h3><a id=\"user-content-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#variables\"><\/a>Variables<\/h3>\n<p>While the <code>Relation<\/code> type may be great for static sets of tuples, in Datalog we expect our relations to grow. As we apply rules we must add more facts, and then these new facts may lead to even more facts, until we eventually reach some limit (one always reaches a limit in Datalog, by virtue of not being able to create new values for tuple coordinates).<\/p>\n<p>Our framework is going to repeatedly apply a set of rules to our growing relations, each of whose tuples are either \"recent\" meaning new as of the most recent application of rules, or \"stable\" meaning they have been present for multiple rounds rule applications (and the rules have had a chance to react to their presence).<\/p>\n<p>To accommodate growing relations, we introduce a new type <code>Variable&lt;Tuple&gt;<\/code>, which is similar in spirit to a <code>Relation&lt;Tuple&gt;<\/code> except that we pay some attention to which <code>Tuple<\/code> elements are brand new and which are not.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>pub struct Variable&lt;Tuple: Ord&gt; {\n    \/\/\/ A list of already processed tuples.\n    stable: Vec&lt;Relation&lt;Tuple&gt;&gt;,\n    \/\/\/ A list of recently added but unprocessed tuples.\n    recent: Relation&lt;Tuple&gt;,\n    \/\/\/ A list of tuples yet to be introduced.\n    to_add: Vec&lt;Relation&lt;Tuple&gt;&gt;,\n}<\/pre><\/div>\n<p>In this type tuples can be found in one of three places.<\/p>\n<ol><li>\n<p>A new tuple is first added to the <code>self.to_add<\/code> list, which indicates that we think the tuple should be in the relation that the variable represents. This list is mostly a \"to-do\" list, and we aren't going to look at its contents other than to add tuples.<\/p>\n<\/li>\n<li>\n<p>Tuples in the <code>to_add<\/code> list are eventually promoted to the <code>self.recent<\/code> relation, if it turns out that they are novel (they haven't already been added to the variable). The <code>recent<\/code> relation contains those tuples that have just been admitted, and should be re-considered by each of the rules that we might apply.<\/p>\n<\/li>\n<li>\n<p>Once they have been considered by all rules, recent tuples are moved into <code>self.stable<\/code>, which contains the full list of processed tuples. As we will see, some rules will want access to even old tuples, and so we will want to keep them around, even if they are not driving new rule applications.<\/p>\n<\/li>\n<\/ol><p>Our rules will be able to look at a <code>Variable<\/code> and see which tuples are new and should be reacted to (<code>self.recent<\/code>), if needed consult tuples that already exist (<code>self.stable<\/code>), and totally ignore tuples that may be added in the next iteration (<code>self.to_add<\/code>).<\/p>\n<p>You might be wondering why <code>stable<\/code> is a list of relations rather than just one, and it has to do with how we will efficiently maintain the contents. We want the flexibility to build <code>stable<\/code> out of several lists of geometrically varying size, so that when we add a single tuple we need not rebuild the entire list.<\/p>\n<h3><a id=\"user-content-rules\" class=\"anchor\" aria-hidden=\"true\" href=\"#rules\"><\/a>Rules<\/h3>\n<p>What do rules look like? We had an example up above,<\/p>\n<pre>nodes(y) &lt;- nodes(x), edges(x,y)\n<\/pre>\n<p>which looks for matches in <code>nodes<\/code> and <code>edges<\/code> and adds them to <code>nodes<\/code>. We could naively implement this by looking at all of <code>nodes<\/code> and <code>edges<\/code> each iteration, both <code>stable<\/code> and <code>recent<\/code>, and adding every matching result we find to <code>nodes<\/code>. That would be a lot of work, and fortunately there are simpler ways to accomplish the same thing.<\/p>\n<p>In each round, we assume that we have already added to the output any tuples that derive from <code>stable<\/code> tuples in the input variables. This means that we only need to look at matches that involve at least one <code>recent<\/code> tuple. In the case up above with just two relations, that means we need to look for matches between:<\/p>\n<ul><li><code>nodes.recent<\/code> and <code>edges.stable<\/code>, and<\/li>\n<li><code>nodes.stable<\/code> and <code>edges.recent<\/code>, and<\/li>\n<li><code>nodes.recent<\/code> and <code>edges.recent<\/code>.<\/li>\n<\/ul><p>If we can come up with an efficient way to find matches between two input <code>Relation<\/code> types, we can just use that three times with the associated classes of tuple.<\/p>\n<h3><a id=\"user-content-helpers\" class=\"anchor\" aria-hidden=\"true\" href=\"#helpers\"><\/a>Helpers<\/h3>\n<p>Let's develop a helpful method that we will use to join together relations, be they <code>recent<\/code> or <code>stable<\/code>. We can't process any two relations, because they need to have a certain structure to find common keys. We will insist that their <code>Tuple<\/code> types have the form <code>(Key, Val1)<\/code> and <code>(Key, Val2)<\/code>, just by defining an implementation on relations of that type (meaning it is applicable to such pair tuple types, and not otherwise).<\/p>\n<p>We also aren't exactly sure what to do with matches yet, so let's leave that open-ended. We will also ask for a function that we can apply to matching <code>Key<\/code>, <code>Val1<\/code>, and <code>Val2<\/code> triples, and we will plan to apply that function to each match rather than populate some output list or anything like that. As we will see, we can stash the output in a list using the function, and be a bit more flexible at the same time.<\/p>\n<p>Here is a prototype for the method:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>fn join_helper&lt;Key: Ord, Val1, Val2&gt;(\n    input1: &amp;Relation&lt;(Key,Val1)&gt;,\n    input2: &amp;Relation&lt;(Key,Val2)&gt;,\n    mut result: impl FnMut(&amp;Key, &amp;Val1, &amp;Val2)) {\n    \/\/ do some stuff probably.\n}<\/pre><\/div>\n<p>What will we do in such a method? Well, we know that each relation is sorted, so we really just need to do a merge between the two relations. We swing through each of them looking for matching keys, and when the keys don't match we advance the relation with the smaller key.<\/p>\n<p>To march through the relations, we will recast them as \"slices\", which is how Rust calls a contiguous hunk of valid typed memory. In other languages we might use indices, or just raw pointers, but this approach is one way to convince Rust's type system that all of our data is valid; we will not be dereferencing random garbage memory.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    \/\/ represent the relations as slices.\n    let mut slice1 = &amp;input1.elements[..];\n    let mut slice2 = &amp;input2.elements[..];\n    while !slice1.is_empty() &amp;&amp; !slice2.is_empty() {\n        use std::cmp::Ordering;\n        \/\/ If the keys match call `result`, else advance the smaller key until they might.\n        match slice1[0].0.cmp(&amp;slice2[0].0) {\n            Ordering::Less =&gt; {\n                slice1 = gallop(slice1, |x| x.0 &lt; slice2[0].0);\n            },\n            Ordering::Equal =&gt; {\n                \/\/ Determine the number of matching keys in each slice.\n                let count1 = slice1.iter().take_while(|x| x.0 == slice1[0].0).count();\n                let count2 = slice2.iter().take_while(|x| x.0 == slice2[0].0).count();\n                \/\/ Produce results from the cross-product of matches.\n                for index1 in 0 .. count1 {\n                    for index2 in 0 .. count2 {\n                        result(&amp;slice1[0].0, &amp;slice1[index1].1, &amp;slice2[index2].1);\n                    }\n                }\n                \/\/ Advance slices past this key.\n                slice1 = &amp;slice1[count1..];\n                slice2 = &amp;slice2[count2..];\n            }\n            Ordering::Greater =&gt; {\n                slice2 = gallop(slice2, |x| x.0 &lt; slice1[0].0);\n            }\n        }\n    }<\/pre><\/div>\n<p>I think a lot of this makes sense, except for the part where I call <code>gallop<\/code>. You probably don't know what that method does, but roughly it slides forward through its first argument until the second argument (a predicate) is no longer true. Instead of doing this one element at a time, it repeatedly doubles its step size until it finds a violation, and then repeatedly halves the step size until it is right up against the limit. For completeness, it looks like this:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>pub fn gallop&lt;T&gt;(mut slice: &amp;[T], mut cmp: impl FnMut(&amp;T)-&gt;bool) -&gt; &amp;[T] {\n    \/\/ if empty slice, or already &gt;= element, return\n    if slice.len() &gt; 0 &amp;&amp; cmp(&amp;slice[0]) {\n        let mut step = 1;\n        while step &lt; slice.len() &amp;&amp; cmp(&amp;slice[step]) {\n            slice = &amp;slice[step..];\n            step = step &lt;&lt; 1;\n        }\n        step = step &gt;&gt; 1;\n        while step &gt; 0 {\n            if step &lt; slice.len() &amp;&amp; cmp(&amp;slice[step]) {\n                slice = &amp;slice[step..];\n            }\n            step = step &gt;&gt; 1;\n        }\n        slice = &amp;slice[1..]; \/\/ advance one, as we always stayed &lt; value\n    }\n    return slice;\n}<\/pre><\/div>\n<h3><a id=\"user-content-rules-redux\" class=\"anchor\" aria-hidden=\"true\" href=\"#rules-redux\"><\/a>Rules (redux)<\/h3>\n<p>Now that we have a helper in hand, what might a rule look like?<\/p>\n<p>Recall that we had those three cases we needed to process,<\/p>\n<ul><li><code>nodes.recent<\/code> and <code>edges.stable<\/code>, and<\/li>\n<li><code>nodes.stable<\/code> and <code>edges.recent<\/code>, and<\/li>\n<li><code>nodes.recent<\/code> and <code>edges.recent<\/code>.<\/li>\n<\/ul><p>Let's write a method that takes two input variables and adds new results to an output variable. Rather than require that the output variable contain tuples exactly matching the key and values of the inputs, we will let the user supply a transformation from those three types to whatever tuple type it manages.<\/p>\n<p>Remember that <code>stable<\/code> contains a list of relations, so we need to swing through each of them.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>pub fn join_into&lt;Key: Ord, Val1: Ord, Val2: Ord, Result: Ord&gt;(\n    input1: &amp;Variable&lt;(Key, Val1)&gt;,\n    input2: &amp;Variable&lt;(Key, Val2)&gt;,\n    output: &amp;mut Variable&lt;Result&gt;,\n    mut logic: impl FnMut(&amp;Key, &amp;Val1, &amp;Val2)-&gt;Result) {\n    let mut results = Vec::new();\n    \/\/ input1.recent and input2.stable.\n    for batch2 in input2.stable.iter() {\n        join_helper(&amp;input1.recent, &amp;batch2, |k,v1,v2| results.push(logic(k,v1,v2)));\n    }\n    \/\/ input1.stable and input2.recent.\n    for batch1 in input1.stable.iter() {\n        join_helper(&amp;batch1, &amp;input2.recent, |k,v1,v2| results.push(logic(k,v1,v2)));\n    }\n    \/\/ input1.recent and input2.recent.\n    join_helper(&amp;input1.recent, &amp;input2.recent, |k,v1,v2| results.push(logic(k,v1,v2)));\n    output.insert(results.into());\n}<\/pre><\/div>\n<p>That's it for the <code>join_into<\/code> rule! If we apply that over and over and over, we will populate <code>output<\/code> with all of the results of matching up <code>input1<\/code> and <code>input2<\/code> on their shared key, allowing <code>output<\/code> to pick out whichever tuple fields it is most interested in keeping.<\/p>\n<h3><a id=\"user-content-distinctness\" class=\"anchor\" aria-hidden=\"true\" href=\"#distinctness\"><\/a>Distinctness<\/h3>\n<p>There is one last really important thing to do: advance the tuples in each <code>Variable<\/code>.<\/p>\n<p>Remember that we have <code>to_add<\/code>, <code>recent<\/code>, and <code>stable<\/code>, and we want to migrate <code>recent<\/code> into <code>stable<\/code> and then <code>to_add<\/code> into <code>recent<\/code>. This is also a great time to check if <code>recent<\/code> is non-empty, because we will probably want to stop iterating once all of our variables have empty recent relations (because there should be nothing new to do).<\/p>\n<p>Our method to do this is a bit scary, so we will break it down into two parts: first migrate <code>recent<\/code> into <code>stable<\/code>, and second migrate <code>to_add<\/code> into <code>recent<\/code>. The overall signature of the method we are about to write is:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>impl&lt;Tuple: Ord&gt; Variable&lt;Tuple&gt; {\n    fn changed(&amp;mut self) -&gt; bool {\n        \/\/ 1. Merge self.recent into self.stable.\n        unimplemented!()\n        \/\/ 2. Move self.to_add into self.recent.\n        unimplemented!()\n        \/\/ Return true iff recent is non-empty.\n        !self.recent.is_empty()\n    }<\/pre><\/div>\n<p>Let's start with the first step: merging <code>self.recent<\/code> and <code>self.stable<\/code>. The plan here is that we are maintaining a list of relations, each of which are distinct from each other. We will also make sure that <code>recent<\/code> is distinct from everything in <code>stable<\/code> (in just a moment), so all we really need to do is add <code>recent<\/code> to the <code>stable<\/code> list.<\/p>\n<p>Except, if we just add to the stable list, we may end up with a long list of small relations, and that is a pain. Instead, we will merge stable relations that have the same size, ensuring that no two are within a factor of two of each other. This ensures we have at most a number of relations logarithmic in the total number of tuples, and that we perform an amortized bounded amount of work for each tuple we add (again, logarithmic in the total number of tuples).<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>        \/\/ 1. Merge self.recent into self.stable.\n        if !self.recent.is_empty() {\n            let mut recent = ::std::mem::replace(&amp;mut self.recent, Vec::new().into());\n            while self.stable.last().map(|x| x.len() &lt;= 2 * recent.len()) == Some(true) {\n                let last = self.stable.pop().unwrap();\n                recent = recent.merge(last);\n            }\n            self.stable.push(recent);\n        }<\/pre><\/div>\n<p>This isn't the very best implementation possible, but it works great for now. We end up with an empty <code>self.recent<\/code>, and a <code>self.stable<\/code> which has only so many relations in it.<\/p>\n<p>The next step is to turn <code>self.to_add<\/code> into the new <code>self.recent<\/code>, where we also want to remove any elements that are already present in our updated <code>self.stable<\/code>. This removal is what ensures we eventually converge: if we eventually add no new tuples, then our next <code>self.recent<\/code> will be empty and we can finally rest.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>        \/\/ 2. Move self.to_add into self.recent.\n        if let Some(mut to_add) = self.to_add.pop() {\n            \/\/ 2a. Merge all newly added relations.\n            while let Some(to_add_more) = self.to_add.pop() {\n                to_add = to_add.merge(to_add_more);\n            }\n            \/\/ 2b. Restrict `to_add` to tuples not in `self.stable`.\n            for batch in self.stable.iter() {\n                let mut slice = &amp;batch[..];\n                to_add.elements.retain(|x| {\n                    slice = join::gallop(slice, |y| y &lt; x);\n                    slice.len() == 0 || &amp;slice[0] != x\n                });\n            }\n            self.recent = to_add;\n        }<\/pre><\/div>\n<p>The two steps here are first to coalesce all of the newly added tuples into a <code>Relation<\/code>, which will be sorted and distinct, and then to swing through each relation in <code>self.stable<\/code> and retaining new elements only when they cannot be found in the stable relation. We use that neat <code>gallop<\/code> thing again.<\/p>\n<p>We've now completed the <code>changed()<\/code> method, which tells us whether there have been any changes in the variable since last we look, and which maintains all of the data internally in a reasonable representation.<\/p>\n<p>That's just about all of the engine!<\/p>\n<h3><a id=\"user-content-caveats\" class=\"anchor\" aria-hidden=\"true\" href=\"#caveats\"><\/a>Caveats<\/h3>\n<p>Ok, actually there are a bunch of other things. For various ergonomic reasons, several fields are wrapped in <code>Rc&lt;RefCell&lt;_&gt;&gt;<\/code> layers, which provide run-time ownership tests. Roughly, we have users create an instance of a <code>Iteration<\/code> type, from which they create variables. This iteration thing wraps all of the <code>changed()<\/code> logic together and tries to prevent you from accidentally not stepping your variables along. Unfortunately this shares some ownership around, and means that we need reference counted ref-cells.<\/p>\n<p>There are also several places that I lied about the implementations, in that I actually wrote something a bit more complicated in the interest of performance (e.g. <code>Relation::merge()<\/code>). These do make the code work a bit better, but you should get just fine performance without them (a few more copies in the case of <code>merge()<\/code>).<\/p>\n<p>You can check out <a href=\"https:\/\/github.com\/frankmcsherry\/datafrog\">the DataFrog repository<\/a> to see the real story, but the presented structure is roughly accurate, I think.<\/p>\n<h3><a id=\"user-content-an-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#an-example\"><\/a>An example<\/h3>\n<p>Let's write an example!<\/p>\n<p>This example is from our <code>nodes<\/code> and <code>edges<\/code> example up above, but actually applied to a real problem. At least, there was a paper published in ASPLOS 2017 that used this as one of their benchmark computations. We are going to reproduce their results!<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>extern crate datafrog;\nuse datafrog::Iteration;\nfn main() {\n    let timer = ::std::time::Instant::now();\n    \/\/ ELIDED: load `nodes` and `edges` here.\n    println!(\"{:?}\\tData loaded\", timer.elapsed());\n    \/\/ Create a new iteration context, ...\n    let mut iteration = Iteration::new();\n    \/\/ .. some variables, ..\n    let variable1 = iteration.variable::&lt;(u32,u32)&gt;(\"nodes\");\n    let variable2 = iteration.variable::&lt;(u32,u32)&gt;(\"edges\");\n    \/\/ .. load them with some initial values, ..\n    variable1.insert(nodes.into());\n    variable2.insert(edges.into());\n    \/\/ .. and then start iterating rules!\n    while iteration.changed() {\n        \/\/ N(c,a) &lt;-  N(b,a), E(b,c)\n        variable1.from_join(&amp;variable1, &amp;variable2, |_b, &amp;a, &amp;c| (c,a));\n    }\n    let reachable = variable1.complete();\n    println!(\"{:?}\\tComputation complete (nodes_final: {})\", timer.elapsed(), reachable.len());\n}<\/pre><\/div>\n<p>Woo! That is a pretty cool program. I have some code up at the top that I've elided that reads the initial collections in from disk. We still time all of that, even though most of it is silliness like parsing text.<\/p>\n<p>It's worth pointing out that the rule as described is a bit weird. The <code>variable1<\/code> variable actually tracks the <em>transpose<\/em> of the <code>N<\/code> relation, which is why a join between <code>variable1<\/code> and <code>variable2<\/code> finds a common value <code>b<\/code> rather than matching up <code>a<\/code> and <code>b<\/code>, which .. isn't what we want. It occasionally takes some head-scratching to make sure you have the right query.<\/p>\n<p>So, does it perform well?<\/p>\n<pre>Echidnatron% cargo run --release --bin graspan1 -- ~\/Desktop\/graspan\/httpd_df.dms\n   Compiling datafrog v0.1.0 (file:\/\/\/Users\/mcsherry\/Projects\/datafrog)\n    Finished release [optimized] target(s) in 1.43 secs\n     Running `target\/release\/graspan1 \/Users\/mcsherry\/Desktop\/graspan\/httpd_df.dms`\nDuration { secs: 1, nanos: 564660489 }  Data loaded\nDuration { secs: 4, nanos: 521095814 }  Computation complete (nodes_final: 9393283)\nEchidnatron%\n<\/pre>\n<p>About 1.5s to load the data, and then three more seconds to run the computation. Maybe add another 1.5s if you want to count the program compilation (seems legit, in some circumstances). How does that compare to the ASPLOS paper? Pretty well, it turns out. Well enough that I'm still blocked on them confirming that we've actually computed the same thing, before getting too sassy.<\/p>\n<h3><a id=\"user-content-next-steps\" class=\"anchor\" aria-hidden=\"true\" href=\"#next-steps\"><\/a>Next steps<\/h3>\n<p>Ergonomics seem to be a bit of a pain here. If you want to have more complicated rules, you need to define a variable for each binary join that you do, which results in a wall of definitions relatively far removed from their use. That is annoying. On the other hand, I'm hoping that a lot of this ends up being automatoable from simpler Datalog syntax.<\/p>\n<p>I would expect the ergonomics might improve in the future, as some folks do seem interested in using this in anger. The more eyeballs we get on it, and elbows applied to it, the better I see it getting.<\/p>\n<p>If nothing else, I hope it is an informative and educational tool about how you too could totally build a Datalog engine. If there are any parts you still aren't confident about, give a holler and I can try and clarify out those aspects.<\/p>\n<h3><a id=\"user-content-addendum-2018-05-21-treefrog-leapjoin\" class=\"anchor\" aria-hidden=\"true\" href=\"#addendum-2018-05-21-treefrog-leapjoin\"><\/a>Addendum 2018-05-21: Treefrog Leapjoin<\/h3>\n<p>The Rust folks have a benchmark query they are interested in, related to a more sophisticated set of rules for their borrow checker (the part of Rust that does the static analysis on your program to see if it can rule out racy memory references). It is a lot more complicated than<\/p>\n<pre>nodes(y) &lt;- nodes(x), edges(x,y)\n<\/pre>\n<p>If you take the naive form of their analysis, the rules look like<\/p>\n<pre>subset(R1, R2, P) &lt;- outlives(R1, R2, P).\nsubset(R1, R3, P) &lt;- subset(R1, R2, P), subset(R2, R3, P).\nsubset(R1, R2, Q) &lt;- subset(R1, R2, P), cfg_edge(P, Q), region_live_at(R1, Q), region_live_at(R2, Q).\nrequires(R, B, P) &lt;- borrow_region(R, B, P).\nrequires(R2, B, P) &lt;- requires(R1, B, P), subset(R1, R2, P).\nrequires(R, B, Q) &lt;- requires(R, B, P), !killed(B, P), cfg_edge(P, Q), region_live_at(R, Q).\nborrow_live_at(B, P) &lt;- requires(R, B, P), region_live_at(R, P)\n<\/pre>\n<p>The current bestest form of the analysis does not use these simple rules, but we will be able to do something instructive with them nonetheless. Let's focus on the rule<\/p>\n<pre>subset(R1, R2, Q) &lt;- subset(R1, R2, P), cfg_edge(P, Q), region_live_at(R1, Q), region_live_at(R2, Q).\n<\/pre>\n<p>This rule happens to extend the <code>subset<\/code> relation with a new value <code>Q<\/code> by joining with three other relations. All of that work and we just get one additional <code>Q<\/code> out of it (and we discard the <code>P<\/code>). Let's see how it is written using the relatively simple framework up above:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    \/\/ subset_1 and subset_2 are temporary relations; subset_p is the real deal.\n    subset_1.from_join(&amp;subset_p, &amp;cfg, |&amp;_p, &amp;(r1, r2), &amp;q| ((r1, q), r2));\n    subset_2.from_join(&amp;subset_1, &amp;rla, |&amp;(r1, q), &amp;r2, &amp;()| ((r2, q), r1));\n    subset_p.from_join(&amp;subset_2, &amp;rla, |&amp;(r2, q), &amp;r1, &amp;()| (q, (r1, r2)));<\/pre><\/div>\n<p>This isn't the whole story for the computation, but the fragment shows us that we start with a <code>subset_p<\/code> relation that we join with a <code>cfg<\/code>, then twice with <code>rla<\/code>. The <code>subset_p<\/code> relation is <code>subset<\/code> indexed by <code>P<\/code>, so that a join with <code>cfg<\/code> adds in a <code>Q<\/code> field; we then join twice with the <code>(R,Q)<\/code> tuples in <code>rla<\/code> to restrict to tuples where both <code>(R1,Q)<\/code> and <code>(R2,Q)<\/code> are in <code>rla<\/code>.<\/p>\n<p>For each <code>(r1, r2, p)<\/code> we are doing a fair bit of work just to get out the new <code>q<\/code> values. The <code>cfg<\/code> relation proposes them all, we then write them all down and re-order so that <code>rla<\/code> can test each of them, twice. Is there perhaps a better way?<\/p>\n<h4><a id=\"user-content-the-treefrog-awakens\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-treefrog-awakens\"><\/a>The Treefrog Awakens<\/h4>\n<p>There is a neat algorithm called <a href=\"https:\/\/arxiv.org\/pdf\/1210.0481v5.pdf\" rel=\"nofollow\">\"leapfrog triejoin\"<\/a>, the intellectual property of LogicBlox, Inc. The algorithm we are going to develop, \"treefrog leapjoin\", may seem similar, but it is a purely superficial connection. They are both connected to the study of <a href=\"https:\/\/cs.stanford.edu\/people\/chrismre\/papers\/paper49.Ngo.pdf\" rel=\"nofollow\">worst-case optimal joins<\/a>, and I think TFLJ is something of a dual to LFTJ.<\/p>\n<p>I should also stress that while the name itself is novel, TFLJ the algorithm is not; it's totally just out of that WCOJ paper up there.<\/p>\n<p>Here is the rough idea:<\/p>\n<p>Imagine that someone provides us with a tuple <code>(r1, r2, p)<\/code> and asks for the list of <code>q<\/code> that should extend it. We have a few relations (<code>cfg<\/code>, <code>rla<\/code>, and <code>rla<\/code> again) that have opinions on which <code>q<\/code> should extend it, so let's maybe ask them.<\/p>\n<p>Imagine each relation had a method<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    fn propose(&amp;self, tuple: &amp;Tuple) -&gt; Vec&lt;Q&gt;;<\/pre><\/div>\n<p>that would tell us what values it would propose for any instance of <code>Tuple<\/code>. We could take these lists and intersect them, and then list out the matching <code>q<\/code> values all in one fell swoop. No intermediate <code>subset_1<\/code> and <code>subset_2<\/code> relations, no sorting them by different keys or anything like that.<\/p>\n<p>It is pretty easy to write a <code>propose<\/code> method like the above, beause for any <code>tuple: &amp;Tuple<\/code> each relation just needs to know how to pick out the fields of <code>tuple<\/code> that are bound to non-<code>Q<\/code> variables in the relation. For example, <code>cfg<\/code> needs to be smart enough to pick out the <code>P<\/code> and then use it to look up <code>Q<\/code> values. Each of the uses of <code>rla<\/code> need to pick out either <code>R1<\/code> or <code>R2<\/code> (based on which use it is) and use it to look up <code>Q<\/code> values.<\/p>\n<p>The relation itself will not be able to do this, which we can see from the two uses of <code>rla<\/code> which must do different things. Instead, we can <em>wrap<\/em> the relation in some type that will provide the logic we need perhaps just by associating a <code>Fn(&amp;Tuple)-&gt;Key<\/code> function which extracts the relation's key from any supplied tuple.<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    pub struct ExtendWith&lt;'a, Key: Ord+'a, Val: Ord+'a, Tuple: Ord, Func: Fn(&amp;Tuple)-&gt;Key&gt; {\n        relation: &amp;'a Relation&lt;(Key, Val)&gt;,\n        key_func: Func,\n        phantom: ::std::marker::PhantomData&lt;Tuple&gt;,\n    }<\/pre><\/div>\n<p>If you are still pondering Rust and all of its idosyncracies, there might be a bit to digest here. This type has several parameters, including the mysteriously ominous <code>'a<\/code> parameter. You can approximately view this struct as pairing a reference to a <code>Relation&lt;(Key, Val)&gt;<\/code> with a specific function <code>Fn(&amp;Tuple)-&gt;Key<\/code>. The <code>'a<\/code> parameter is Rust's way of acknowledging that while this reference may not be around forever, when we talk about other references to things perhaps we can say that they will be around for roughly as long.<\/p>\n<p>Here is a <code>propose()<\/code> method for <code>ExtendWith<\/code> (where we apparently populate a <code>Vec&lt;&amp;'a Val&gt;<\/code> instead of returning one).<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    fn propose(&amp;mut self, prefix: &amp;Tuple) -&gt; Vec&lt;&amp;'a Val&gt; {\n        let key = (self.key_func)(prefix);\n        let start = binary_search(&amp;self.relation[..], |x| &amp;x.0 &lt; &amp;key);\n        let slice = &amp;self.relation[self.start ..];\n        let end = slice.len() - gallop(slice, |x| &amp;x.0 &lt;= &amp;key).len();\n        slice[start .. end].iter().map(|&amp;(_, ref val)| val).collect()\n    }<\/pre><\/div>\n<p>This would probably work pretty well, using <code>propose<\/code> for each relation wrapped with an <code>ExtendWith<\/code>, but we are going to do something a bit smarter and more flexible.<\/p>\n<h4><a id=\"user-content-enter-the-treefrog\" class=\"anchor\" aria-hidden=\"true\" href=\"#enter-the-treefrog\"><\/a>Enter the Treefrog<\/h4>\n<p>The <code>propose<\/code> method is a solid way to start out eliciting information from each of the participating relations, but we can do a bit better. Imagine that each participating relation provided three methods:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    \/\/\/ Methods to support treefrog leapjoin.\n    pub trait Leaper&lt;'a, Tuple, Val&gt; {\n        \/\/\/ Estimates the number of proposed values.\n        fn count(&amp;mut self, prefix: &amp;Tuple) -&gt; usize;\n        \/\/\/ Populates `values` with proposed values.\n        fn propose(&amp;mut self, prefix: &amp;Tuple, values: &amp;mut Vec&lt;&amp;'a Val&gt;);\n        \/\/\/ Restricts `values` to proposed values.\n        fn intersect(&amp;mut self, prefix: &amp;Tuple, values: &amp;mut Vec&lt;&amp;'a Val&gt;);\n    }<\/pre><\/div>\n<p>The <code>propose<\/code> method is familiar, if slightly different, but what do the other two methods do? The <code>count<\/code> method estimates a number of values that would be proposed, and just needs to avoid being zero if the relation would actually propose some values. The <code>intersect<\/code> method takes a list of proposed values and restricts them down to those that the relation would propose; this can be a lot more efficient than actually proposing the values themselves (imagine a small <code>values<\/code> and a large list of proposals).<\/p>\n<p>Using just this trait, let's write a <code>leapjoin_into<\/code> method that extends any <code>Variable&lt;Tuple&gt;<\/code> using a list of these <code>Leaper<\/code> implementors:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>\/\/\/ Performs treefrog leapjoin using a list of leapers.\npub fn leapjoin_into&lt;'a, Tuple: Ord, Val: Ord+'a, Result: Ord&gt;(\n    source: &amp;Variable&lt;Tuple&gt;,\n    leapers: &amp;mut [&amp;mut Leaper&lt;'a, Tuple, Val&gt;],\n    output: &amp;Variable&lt;Result&gt;,\n    mut logic: impl FnMut(&amp;Tuple, &amp;Val)-&gt;Result) {\n    let mut result = Vec::new();    \/\/ temp output storage.\n    let mut values = Vec::new();    \/\/ temp value storage.\n    for tuple in source.recent.borrow().iter() {\n        \/\/ 1. Determine which leaper would propose the fewest values.\n        \/\/ 2. Have the least-proposing leaper propose their values.\n        \/\/ 3. Have the other leapers restrict the proposals.\n        \/\/ 4. Call `logic` on each value, push into `result`.\n    }\n    output.insert(result.into());\n}<\/pre><\/div>\n<p>These four steps are not so complicated, I just wanted to avoid a wall of text. Let's look at each of them individually:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>        \/\/ 1. Determine which leaper would propose the fewest values.\n        let mut min_index = usize::max_value();\n        let mut min_count = usize::max_value();\n        for index in 0 .. leapers.len() {\n            let count = leapers[index].count(tuple);\n            if min_count &gt; count {\n                min_count = count;\n                min_index = index;\n            }\n        }<\/pre><\/div>\n<p>Great! Super easy. We just ask each leaper, and they tell us. The rest is pretty easy too:<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>        if min_count &gt; 0 {\n            \/\/ 2. Have the least-proposing leaper propose their values.\n            leapers[min_index].propose(tuple, &amp;mut values);\n            \/\/ 3. Have the other leapers restrict the proposals.\n            for index in 0 .. leapers.len() {\n                if index != min_index {\n                    leapers[index].intersect(tuple, &amp;mut values);\n                }\n            }\n            \/\/ 4. Call `logic` on each value, push into `result`.\n            for val in values.drain(..) {\n                result.push(logic(tuple, val));\n            }\n        }<\/pre><\/div>\n<p>That is the whole <code>leapjoin_into<\/code> implementation!<\/p>\n<h4><a id=\"user-content-other-flavors-of-treefrog\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-flavors-of-treefrog\"><\/a>Other flavors of TreeFrog<\/h4>\n<p>While one of the reasons for the three methods (<code>count<\/code>, <code>propose<\/code>, and <code>intersect<\/code>) is efficiency, another is flexibility. To see what this is about, let's consider an query fragment from Rust's <a href=\"https:\/\/github.com\/rust-lang-nursery\/polonius\/blob\/master\/src\/output\/datafrog_opt.rs#L164-L179\">\"more optimized\" version<\/a> of their borrow checker logic.<\/p>\n<pre>dead_region_requires(R, B, P, Q) :-\n    requires(R, B, P),\n    !killed(B, P),\n    cfg_edge(P, Q),\n    !region_live_at(R, Q).\n<\/pre>\n<p>This method starts with triples <code>(r, b, p)<\/code> from the <code>requires<\/code> variable, and does various restrictions to introduce a new <code>q<\/code>.<\/p>\n<p>However, these constraints aren't just a matter of <em>proposing<\/em> values for <code>q<\/code>. The <code>!killed(B, P)<\/code> constraint says that <em>no<\/em> <code>q<\/code> are appropriate for any <code>(b, p)<\/code> in <code>killed<\/code>. The <code>!rla(R, Q)<\/code> says that only <code>q<\/code> <em>not<\/em> present in <code>rla<\/code> are appropriate.<\/p>\n<p>These two negative constraints would be very hard to write in terms of <code>propose<\/code>. We can't list out the values that would be satisfactory, because there could be an unbounded number of them. When <code>(b, p)<\/code> is not in <code>killed<\/code>, literally any value of <code>q<\/code> is just fine. The <code>rla<\/code> relation would have to list the values <em>not<\/em> associated with <code>r<\/code>, which while not <em>any<\/em> value of <code>q<\/code> are still most of them.<\/p>\n<p>Fortunately, even for these two classes of constraint we can implement the <code>Leaper<\/code> trait, with methods<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    fn count(&amp;mut self, prefix: &amp;Tuple) -&gt; usize;\n    fn propose(&amp;mut self, prefix: &amp;Tuple, values: &amp;mut Vec&lt;&amp;'a Val&gt;);\n    fn intersect(&amp;mut self, prefix: &amp;Tuple, values: &amp;mut Vec&lt;&amp;'a Val&gt;);<\/pre><\/div>\n<p>A bit of pondering, and you may conclude that you can implement each of these methods for the two types of constraints above. Where we called our first <code>Leaper<\/code> type <code>ExtendWith<\/code>, I'm going to call these two <code>FilterAnti<\/code> and <code>ExtendAnti<\/code> for the <code>killed<\/code> and <code>rla<\/code> uses respectively.<\/p>\n<p>The <code>FilterAnti<\/code> implementation should report a count of either <code>0<\/code> or <code>usize::max_value()<\/code>, depending on whether the tuple's key and value are present in the relation or not. It is welcome to panic in its <code>propose<\/code> implementation (something has gone wrong if we are there), and its <code>intersect<\/code> implementation can be empty (we could only be there if we reported a count of <code>usize::max_value()<\/code>).<\/p>\n<p>The <code>ExtendAnti<\/code> implementation shourd report a count of <code>usize::max_value()<\/code>, panic in <code>propose()<\/code>, and perform roughly the same logic for <code>intersect<\/code> as does <code>ExtendWith<\/code> but complementing its decisions (retaining elements when they are absent from the relation).<\/p>\n<p>Notice that it is important that <code>leapjoin_into()<\/code> does nothing when a <code>min_count<\/code> of zero comes back. That is a lot cheaper than asking the <code>FilterAnti<\/code> and <code>ExtendAnti<\/code> wrappers to double check that the tuple is absent and propose nothing.<\/p>\n<h4><a id=\"user-content-consequences\" class=\"anchor\" aria-hidden=\"true\" href=\"#consequences\"><\/a>Consequences<\/h4>\n<p>Let's check out the performance implications on <a href=\"https:\/\/github.com\/rust-lang-nursery\/polonius\/\">Rust's prototype borrow checker<\/a>.<\/p>\n<p>We can grab and run the current version (which uses DataFrog, but not the treefrog leapjoin):<\/p>\n<pre>Echidnatron% git clone https:\/\/github.com\/rust-lang-nursery\/polonius\/\nEchidnatron% cd polonius\nEchidnatron% cargo +nightly run --release -- -a DatafrogOpt inputs\/clap-rs\/app-parser-{{impl}}-add_defaults\/ --skip-tuples\n...\nTime: 7.692s\nEchidnatron%\n<\/pre>\n<p>Before getting too excited, we should update <code>Cargo.toml<\/code> to point at the current <a href=\"https:\/\/github.com\/frankmcsherry\/datafrog\">datafrog master<\/a>, which has <a href=\"https:\/\/github.com\/frankmcsherry\/datafrog\/pull\/2\">some recent improvements<\/a>.<\/p>\n<pre>Echidnatron% nano Cargo.toml\nEchidnatron% cargo +nightly run --release -- -a DatafrogOpt inputs\/clap-rs\/app-parser-{{impl}}-add_defaults\/ --skip-tuples\n...\nTime: 6.985s\nEchidnatron%\n<\/pre>\n<p>Now let's do a bit of editing in <code>datafrog_opt.rs<\/code>, which is where their logic lives. I went and did this by hand, and it's not pretty (but arguably not less pretty than before). You end up with a bunch of code that looks like<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    \/\/ dead_region_requires((R, P, Q), B) :-\n    \/\/   requires(R, B, P),\n    \/\/   !killed(B, P),\n    \/\/   cfg_edge(P, Q),\n    \/\/   !region_live_at(R, Q).\n    leapjoin_into(\n        &amp;requires_rp,           \/\/ dynamic source\n        &amp;mut [\n            &amp;mut killed.filter_anti(|&amp;((_r,p),b)| (b,p)),\n            &amp;mut cfg_edge_rel.extend_with(|&amp;((_r,p),_b)| p),\n            &amp;mut region_live_at_rel.extend_anti(|&amp;((r,_p),_b)| r),\n        ],\n        &amp;dead_region_requires,  \/\/ destination\n        |&amp;((r,p),b),&amp;q| ((r,p,q),b)\n    );<\/pre><\/div>\n<p>instead of<\/p>\n<div class=\"highlight highlight-source-rust\"><pre>    \/\/ dead_region_requires((R, P, Q), B) :-\n    \/\/   requires(R, B, P),\n    \/\/   !killed(B, P),\n    \/\/   cfg_edge(P, Q),\n    \/\/   !region_live_at(R, Q).\n    dead_region_requires_1.from_antijoin(&amp;requires_bp, &amp;killed, |&amp;(b, p), &amp;r| (p, (b, r)));\n    dead_region_requires_2.from_join(\n        &amp;dead_region_requires_1,\n        &amp;cfg_edge_p,\n        |&amp;p, &amp;(b, r), &amp;q| ((r, q), (b, p)),\n    );\n    dead_region_requires.from_antijoin(\n        &amp;dead_region_requires_2,\n        &amp;region_live_at_rel,\n        |&amp;(r, q), &amp;(b, p)| ((r, p, q), b),\n    );<\/pre><\/div>\n<p>In any case, you also end up improving the time from 6.985 seconds to<\/p>\n<pre>Echidnatron% nano src\/output\/datafrog_opt.rs\nEchidnatron% cargo +nightly run --release -- -a DatafrogOpt inputs\/clap-rs\/app-parser-{{impl}}-add_defaults\/ --skip-tuples\n...\nTime: 5.275s\nEchidnatron%\n<\/pre>\n<p>This isn't the world's greatest speed-up, but it is pretty solid. There are also several further avenues to explore, including better ways to perform intersection (right now we always use <code>gallop<\/code>) and replacing <code>Vec&lt;&amp;'a Val&gt;<\/code> lists of pointers with just <code>Vec&lt;Val&gt;<\/code> because all of our <code>Val<\/code> types are just integers.<\/p>\n<p>For <em>our<\/em> purposes, we have ideally learned a bit more about how to implement clever join strategies, while additionally supporting various flavors of negative constraints (though only in static relations, rather than variables).<\/p>","created_at":"2019-05-26T20:42:08+0000","updated_at":"2019-05-26T20:42:10+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":30,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/5741500?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":959,"title":"books","url":"https:\/\/postmake.io\/books","created_at":"2019-05-26T00:54:37+0000","updated_at":"2019-05-26T00:54:37+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"postmake.io"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":958,"title":"The structure of musical harmony as an ordered phase of sound: A statistical mechanics approach to music theory","url":"https:\/\/advances.sciencemag.org\/content\/5\/5\/eaav8490.full","content":"<div id=\"sec-4\" class=\"subsection\"><h3>Mean field approximation<\/h3><p id=\"p-13\">We first apply a mean field approximation and study the equilibrium distribution of pitches. In the limit <em>N<sub>t<\/sub><\/em> \u2192 \u221e and all <em>a<sub>ij<\/sub><\/em> = <em>a<\/em><sub>0<\/sub> constant, the pitches within the set are described only by their probability distribution <em>P<\/em>(<em>x<\/em>), where the pitch <em>x<\/em> = log<sub>2<\/sub><em>f<\/em>\/<em>f<\/em><sub>ref<\/sub> is specified by the ratio of its frequency <em>f<\/em> with an arbitrary reference frequency <em>f<\/em><sub>ref<\/sub>. <em>P<\/em>(<em>x<\/em>) represents the probability with which a pitch <em>x<\/em> is used in a system of music. We make two further simplifying assumptions: (i) We take a fixed value of <em>w<sub>c<\/sub><\/em>, so <em>D<\/em> is a function of <em>x<\/em> only, and (ii) we assume the periodicity <em>P<\/em>(<em>x<\/em>) = <em>P<\/em>(<em>x<\/em> + 1). That is, the pitch distribution is the same in every octave. With these assumptions, we define Dp(x)=\u2211n=\u2212\u221e\u221eD(x+n) and restrict our attention to a single octave <em>x<\/em> = [0,1). Integrating over <em>P<\/em>(<em>x<\/em>), we obtain the total dissonanceDtot=12\u222b01\u222b01P(x)Dp(x\u2212y)P(y)dydx(1)and entropyS=\u2212\u222b01P(x)ln\u00a0P(x)dx(2)<\/p><p id=\"p-14\">We now wish to find the distribution <em>P<\/em>(<em>x<\/em>) that minimizes the free energy <em>F<\/em> = <em>D<\/em><sub>tot<\/sub> \u2212 <em>TS<\/em>, subject to the constraint that <em>P<\/em>(<em>x<\/em>) is normalized. Using a variational method, we obtain an integral equation for the extremum condition and find stable minima using an iterative approach (see Methods) (<a id=\"xref-ref-13-1\" class=\"xref-bibr\" href=\"#ref-13\"><em>13<\/em><\/a>).<\/p><p id=\"p-15\"><a id=\"xref-fig-2-1\" class=\"xref-fig\" href=\"#F2\">Figure 2A<\/a> shows the solutions <em>P<\/em>(<em>x<\/em>) for several values of <em>T<\/em> using the first 10 partials of the sawtooth timbre and <em>w<sub>c<\/sub><\/em> = 0.03 [see fig. S1A for a plot of <em>P<\/em>(<em>x<\/em>) over a continuous range of <em>T<\/em>]. At high <em>T<\/em>, <em>P<\/em>(<em>x<\/em>) = 1 is the stable solution, and the system is completely disordered. At low <em>T<\/em>, <em>P<\/em>(<em>x<\/em>) approaches a single delta-function spike at an arbitrary pitch. As <em>T<\/em> increases from zero, additional peaks emerge, eventually leading to a transition to a <em>P<\/em>(<em>x<\/em>) with 12 equal peaks. We will see below that the high- and low-<em>T<\/em> behaviors are separated by phase transitions from random, disordered sound to the ordered, discrete tuning systems used in music.<\/p><figure class=\"fig pos-float type-figure  odd figure figure--data\" id=\"F2\">\n    <figcaption class=\"fig-caption\" id=\"F2-caption\">\n    Fig. 2 Mean field results.(<strong>A<\/strong>) Solutions <em>P<\/em>(<em>x<\/em>) to <a id=\"xref-disp-formula-4-1\" class=\"xref-disp-formula\" href=\"#disp-formula-4\">Eq. 4<\/a> for a range of <em>T<\/em> and <em>w<sub>c<\/sub><\/em> = 0.03. (<strong>B<\/strong>) Modulus \u2223<em>p<sub>k<\/sub><\/em>\u2223 of order parameters versus temperature <em>T<\/em>. The dashed line indicates the predicted <em>T<\/em><sub><em>c<\/em><sub>1<\/sub><\/sub> = \u2212 <em>d<\/em><sub>12<\/sub>\/2.  <\/figcaption><\/figure><p id=\"p-17\">Insight can be gained using the Fourier representations P(x)=12\u2211k=0\u221epke2\u03c0ikx+pk*e\u22122\u03c0ikx and Dp(x)=\u2211k=0\u221edkcos(2\u03c0kx). For example, the real coefficients <em>d<sub>k<\/sub><\/em> are shown in <a id=\"xref-fig-1-5\" class=\"xref-fig\" href=\"#F1\">Fig. 1C<\/a> for the <em>D<sub>p<\/sub><\/em>(<em>x<\/em>) calculated from (<a id=\"xref-fig-1-6\" class=\"xref-fig\" href=\"#F1\">Fig. 1B<\/a>). Normalization requires <em>p<\/em><sub>0<\/sub> = 1, with the other <em>p<sub>k<\/sub><\/em> representing order parameters that describe the degree of <em>k<\/em>-fold periodic order.<\/p><p id=\"p-18\"><a id=\"xref-fig-2-2\" class=\"xref-fig\" href=\"#F2\">Figure 2B<\/a> shows \u2223<em>p<sub>k<\/sub><\/em>(<em>T<\/em>)\u2223 for <em>k<\/em> = 0 to 20 for the data shown in (<a id=\"xref-fig-1-7\" class=\"xref-fig\" href=\"#F1\">Fig. 1A<\/a>). As expected, the high temperature solution has <em>p<\/em><sub>0<\/sub> = 1 and all other <em>p<sub>k<\/sub><\/em> = 0. As <em>T<\/em> decreases, a single component <em>p<\/em><sub>12<\/sub> is the first to become nonzero at <em>T<\/em><sub><em>c<\/em>1<\/sub> = 20.2. After a further decrease of <em>T<\/em>, the other <em>p<sub>k<\/sub><\/em> become nonzero at <em>T<\/em> &lt; <em>T<\/em><sub><em>c<\/em>2<\/sub> = 16.2.<\/p><p id=\"p-19\">The transition between the high-T disordered solution and the 12-fold solution can be understood from Landau theory. In the neighborhood of <em>T<\/em><sub><em>c<\/em>1<\/sub>, we can expand <em>F<\/em> in powers of a single order parameter <em>p<\/em><sub><em>k<\/em><sub>0<\/sub><\/sub> (see Methods). We find that there is a second-order phase transition at <em>T<\/em><sub><em>c<\/em>1<\/sub> = \u2212 <em>d<\/em><sub><em>k<\/em><sub>0<\/sub><\/sub>\/2 from the disordered phase to a phase with <em>k<\/em><sub>0<\/sub>-fold order, where <em>d<\/em><sub><em>k<\/em>0<\/sub> is the most negative Fourier coefficient. <em>d<\/em><sub>12<\/sub> is the most negative Fourier coefficient in <a id=\"xref-fig-1-8\" class=\"xref-fig\" href=\"#F1\">Fig. 1C<\/a>, with the transition to the 12-fold phase occurring at <em>T<\/em> = \u2212 <em>d<\/em><sub>12<\/sub>\/2, as indicated by the dashed line in <a id=\"xref-fig-2-3\" class=\"xref-fig\" href=\"#F2\">Fig. 2B<\/a>.<\/p><p id=\"p-20\">The free energy only depends on the modulus of <em>p<\/em><sub><em>k<\/em><sub>0<\/sub><\/sub>, with the complex phase randomly chosen by spontaneous symmetry breaking. This symmetry breaking is seen in <a id=\"xref-fig-2-4\" class=\"xref-fig\" href=\"#F2\">Fig. 2<\/a> as the continuous symmetry of the disordered phase changes to the 12-fold translational symmetry below <em>T<\/em><sub><em>c<\/em>1<\/sub>. The randomly chosen phase of <em>p<\/em><sub>12<\/sub> gives rise to the particular positions of the 12 peaks in <em>P<\/em>(<em>x<\/em>). In musical terms, a benchmark pitch is chosen arbitrarily and is fixed only by convention (e.g., A440 at 440 Hz). In this phase, all 12 peaks are equivalent and equally spaced, calling to mind the equal temperament (ET) tuning that has been commonly used in Western music for the last several hundred years (<a id=\"xref-ref-14-1\" class=\"xref-bibr\" href=\"#ref-14\"><em>14<\/em><\/a>).<\/p><p id=\"p-21\">When more than one <em>p<\/em><sub><em>k<\/em> \u2265 1<\/sub> \u2260 0, there are additional terms in the free energy arising from interactions between the modes. The coupling yields more complex behavior where multiple modes undergo transitions at the same temperature <em>T<\/em><sub><em>c<\/em>2<\/sub>, reducing the symmetry to only the assumed octave-wise (\u201conefold\u201d) symmetry. The spontaneous symmetry breaking randomly chooses one of the 12 peaks to become dominant over the others. This symmetry breaking calls to mind a tuning system that favors a particular key, as was common in medieval music through the time of J. S. Bach. For example, \u201cjust intonation\u201d (JI) tuning places pitches at small-integer ratios relative to the root pitch. In <a id=\"xref-fig-3-1\" class=\"xref-fig\" href=\"#F3\">Fig. 3<\/a>, we compare the maxima of <em>P<\/em>(<em>x<\/em>) from the mean field model with the pitches used in the JI and ET systems. Above <em>T<\/em><sub><em>c<\/em>2<\/sub>, all pitches are treated equally by symmetry, so we must have the even 12-fold division of ET. When the symmetry is broken below <em>T<\/em><sub><em>c<\/em>2<\/sub>, a single pitch can be designated as the root, and the pitches tend to align with the JI scheme. Just below <em>T<\/em><sub><em>c<\/em>2<\/sub>, the peaks reflect a compromise between ET and JI, which is similar to tuning systems that have also been used historically, such as the sixth-comma meantone system (<a id=\"xref-ref-14-2\" class=\"xref-bibr\" href=\"#ref-14\"><em>14<\/em><\/a>).<\/p><figure class=\"fig pos-float type-figure  odd figure figure--data\" id=\"F3\">\n    <figcaption class=\"fig-caption\" id=\"F3-caption\">\n    Fig. 3 Comparison of peak positions of <em>P<\/em>(<em>x<\/em>) versus <em>T<\/em> from the mean field model with ET or JI tuning.Two values are shown for the tritone (the pitch near 600 cents) in JI, corresponding to an augmented fourth or diminished fifth. The mean field result shows equal peaks at both positions.  <\/figcaption><\/figure><p id=\"p-23\">The dependence of <em>P<\/em>(<em>x<\/em>) on both <em>T<\/em> and the fixed value of <em>w<sub>c<\/sub><\/em> can be visualized via a color map of <em>d<\/em>\u2223<em>p<sub>k<\/sub><\/em>\u2223\/<em>dT<\/em>, with different values of <em>k<\/em> represented by superimposing different colors (see Methods). The resulting plot is essentially a phase diagram with transitions made apparent by large values of <em>d<\/em>\u2223<em>p<sub>k<\/sub><\/em>\u2223\/<em>dT<\/em>. <a id=\"xref-fig-4-1\" class=\"xref-fig\" href=\"#F4\">Figure 4<\/a> shows the phase diagram for the sawtooth timbre used above, with <em>T<\/em> swept down at each value of <em>w<sub>c<\/sub><\/em>. [See fig. S1 for plots of <em>P<\/em>(<em>x<\/em>) versus <em>T<\/em> at several values of <em>w<sub>c<\/sub><\/em>, as well as a comparison of the phase diagrams with <em>T<\/em> swept up and down. As expected for a second-order transition, <em>T<\/em><sub><em>c<\/em>1<\/sub> is the same for <em>T<\/em> swept up and down, up to a small computational error. The lower transition shows significant hysteresis in <em>T<\/em><sub><em>c<\/em>2<\/sub>, illustrating that mode coupling leads to a first-order transition]. The lines on the plot show \u2212<em>d<sub>k<\/sub><\/em>\/2 versus <em>w<sub>c<\/sub><\/em>, with the value of <em>k<\/em> indicated by the same color scale. When <em>T<\/em> is swept down from the disordered phase, the first transition is well described by the single-mode, mean field prediction. The single value of <em>k<\/em><sub>0<\/sub> can be seen clearly just below <em>T<\/em><sub><em>c<\/em>1<\/sub> by a distinct color. As more modes become mixed in, the color fades, suddenly becoming bright white after the transition at <em>T<\/em><sub><em>c<\/em>2<\/sub>. While the 12-fold phase that appears from 0.022 &lt; <em>w<sub>c<\/sub><\/em> &lt; 0.042 reproduces the 12-fold division of the octave used in Western music, other ranges of <em>w<sub>c<\/sub><\/em> exhibit phases with other values of <em>k<\/em><sub>0<\/sub>, such as 5, 7, 19, and 31. Suggestively, these values of <em>k<\/em><sub>0<\/sub> are among those used in some non-Western music traditions (<a id=\"xref-ref-15-1\" class=\"xref-bibr\" href=\"#ref-15\"><em>15<\/em><\/a>, <a id=\"xref-ref-16-1\" class=\"xref-bibr\" href=\"#ref-16\"><em>16<\/em><\/a>) or in modern music (<a id=\"xref-ref-17-1\" class=\"xref-bibr\" href=\"#ref-17\"><em>17<\/em><\/a>). The particular values of <em>k<\/em><sub>0<\/sub> that appear, and their relative stabilities, depend on the choice of timbre (see fig. S1, C and D, for phase diagrams of timbres featured in non-Western music that do not use a 12-fold octave division). This mean field model therefore provides an avenue for optimizing the timbre of instruments to achieve a desired stability (or instability) for an existing or new musical system.<\/p><figure class=\"fig pos-float type-figure  odd figure figure--data\" id=\"F4\">\n    <figcaption class=\"fig-caption\" id=\"F4-caption\">\n    Fig. 4 Phase diagram showing significant order parameters \u2223<em>p<sub>k<\/sub><\/em>\u2223 versus <em>T<\/em> and <em>w<sub>c<\/sub><\/em>.<em>d<\/em>\u2223<em>p<sub>k<\/sub><\/em>\u2223\/<em>dT<\/em> is plotted with color indicated by the color bar, superimposing all values of <em>k<\/em>. Regions of a single color indicate phases with one or few dominant <em>p<sub>k<\/sub><\/em>; white regions indicate phases with many significant <em>p<sub>k<\/sub><\/em>. The lines show \u2212<em>d<sub>k<\/sub><\/em>\/2 versus <em>w<sub>c<\/sub><\/em> for values of <em>k<\/em> corresponding to the color bar. <em>w<sub>c<\/sub><\/em> corresponding to pitches from C4 (middle C) to C8 are indicated.  <\/figcaption><\/figure><p id=\"p-25\">The mean field model assumes a fixed value of <em>w<sub>c<\/sub><\/em>, although experimentally <em>w<sub>c<\/sub><\/em> is found to depend on the root pitch (lower of the two pitches). The values of <em>w<sub>c<\/sub><\/em> that correspond to root pitches from middle C (C4) to the highest C on the piano (C8) are indicated on <a id=\"xref-fig-4-2\" class=\"xref-fig\" href=\"#F4\">Fig. 4<\/a>. Each ordered phase occupies a fairly small range of root pitches, with the 12-fold phase occurring only at relatively high pitch, likely because of the gross simplifications of the mean field approximation.<\/p><p id=\"p-26\">So far, we have considered a model with a simplified dissonance function and assumed octave-wise periodicity, treated in the mean field approximation. One can explore other assumptions for the periodicity as well, simply by redefining <em>x<\/em> = log<em><sub>b<\/sub>f<\/em>\/<em>f<\/em><sub>ref<\/sub> for <em>b<\/em> other than 2. For example, one finds that for tones with odd harmonic partials (\u03d5<em><sub>n<\/sub><\/em> = 1,3,5, \u2026 ), taking <em>b<\/em> = 3 (so \u201coctaves\u201d are separated by a factor of 3 in frequency) yields a 13-fold phase known as the Bohlen-Pierce scale (<a id=\"xref-ref-18-1\" class=\"xref-bibr\" href=\"#ref-18\"><em>18<\/em><\/a>) with lower free energy than with <em>b<\/em> = 2 (see fig. S2). One could simplify further in a Potts-type model (<a id=\"xref-ref-19-1\" class=\"xref-bibr\" href=\"#ref-19\"><em>19<\/em><\/a>), in which the tones <em>t<sub>i<\/sub><\/em> may take on only a discrete set of <em>q<\/em> pitches, <em>x<sub>i<\/sub><\/em> = <em>n<\/em>\/<em>q<\/em>, <em>n<\/em> = 0,1, \u2026 , <em>q<\/em> \u2212 1. This model allows further analytic calculation and lends itself to other methods from statistical mechanics. These models serve to demonstrate that phase transitions can occur from disordered sound to ordered arrangements of pitches that bear notable resemblance to familiar musical systems.<\/p><\/div><div id=\"sec-5\" class=\"subsection\"><h3>The tone lattice<\/h3><p id=\"p-27\">To go beyond the mean field model, we turn to numerical simulation of a system of interacting tones. Using this method, we will be able to relax the unrealistic assumption that each tone interacts equally with all other tones. Instead, each tone will interact with a subset of the other tones. This can be accomplished by placing the tones on a lattice, with interaction only with a set of nearest neighbors. This lattice can be interpreted as existing in an abstract space, where tones on nearby lattice sites have a stronger harmonic relationship than more distant ones. Although many lattices with different dimension and different interactions may be studied, here, I choose one of the simplest cases that can be expected to produce rich behavior: tones on a two-dimensional (2D) square lattice, with only short-range interactions. Given octave-wise periodicity of interactions, this system can be described using the well-studied XY model (<a id=\"xref-ref-20-1\" class=\"xref-bibr\" href=\"#ref-20\"><em>20<\/em><\/a>, <a id=\"xref-ref-21-1\" class=\"xref-bibr\" href=\"#ref-21\"><em>21<\/em><\/a>), where the pitch <em>x<\/em> stands in for the angle \u03b8. The XY model on a 2D lattice is known to produce particularly rich behavior. As in superfluid thin films (<a id=\"xref-ref-22-1\" class=\"xref-bibr\" href=\"#ref-22\"><em>22<\/em><\/a>) and nematic liquid crystals (<a id=\"xref-ref-23-1\" class=\"xref-bibr\" href=\"#ref-23\"><em>23<\/em><\/a>), one expects to observe a Kosterlitz-Thouless transition from a disordered state, with free vortices to an ordered state exhibiting quasi\u2013long-range order (<a id=\"xref-ref-24-1\" class=\"xref-bibr\" href=\"#ref-24\"><em>24<\/em><\/a>\u2013<a id=\"xref-ref-26-1\" class=\"xref-bibr\" href=\"#ref-26\"><em>26<\/em><\/a>). A vortex in this system is a region surrounding a topological defect, where the pitches along a closed path enclosing the defect traverse one or more octaves. Following a quench of the system from the disordered to ordered state, metastable states are observed with bound vortices and antivortices (<a id=\"xref-ref-27-1\" class=\"xref-bibr\" href=\"#ref-27\"><em>27<\/em><\/a>\u2013<a id=\"xref-ref-29-1\" class=\"xref-bibr\" href=\"#ref-29\"><em>29<\/em><\/a>). In this case, moreover, interactions tend not only to align neighboring pitches but also to favor sudden jumps by particular intervals \u0394<em>x<\/em> \u2248 3\/12,4\/12,5\/12,7\/12,9\/12. This type of XY model, where the energy has multiple minima, has found application, for example, in describing stacking domains in bilayer graphene (<a id=\"xref-ref-30-1\" class=\"xref-bibr\" href=\"#ref-30\"><em>30<\/em><\/a>), ferroelectric domains in hexagonal manganites (<a id=\"xref-ref-31-1\" class=\"xref-bibr\" href=\"#ref-31\"><em>31<\/em><\/a>), and axion-based cosmological models (<a id=\"xref-ref-32-1\" class=\"xref-bibr\" href=\"#ref-32\"><em>32<\/em><\/a>). For sufficiently small vortices, we expect not a continuous variation of <em>x<\/em> around a vortex but, instead, a number of domain boundaries where <em>x<\/em> changes by one of the more consonant intervals.<\/p><p id=\"p-28\">To study the 2D XY behavior that emerges from a set of tones on a lattice, we consider interactions only with a set of nearest neighbors and simulate the resulting stochastic dynamics (see Methods). As before, tones with pitches <em>f<sub>i<\/sub><\/em> and <em>f<sub>j<\/sub><\/em> interact because of their mutual dissonance <em>D<\/em>(<em>f<sub>i<\/sub><\/em>, <em>f<sub>j<\/sub><\/em>). For maximum generality, we make no assumption about the periodicity of the pitch distribution and allow the dissonance width <em>w<sub>c<\/sub><\/em> to vary with the root pitch min(<em>f<sub>i<\/sub><\/em>, <em>f<sub>j<\/sub><\/em>) so that intervals involving lower pitches have broader dissonance curves, as observed experimentally. [Figure S3 shows <em>D<\/em>(\u0394<em>x<\/em>) for several root pitches]. We then use a combination of Langevin dynamics and Metropolis Monte Carlo (MC) to simulate the behavior of the tone lattice at temperature <em>T<\/em>.<\/p><p id=\"p-29\"><a id=\"xref-fig-5-1\" class=\"xref-fig\" href=\"#F5\">Figure 5A<\/a> shows the histogram of pitch classes <em>x<sub>i<\/sub><\/em> mod 1 following a simulated quench from a disordered state to selected <em>T<\/em> (see fig. S4A for the full pitch histograms). At <em>T<\/em> = 5.5 and above, no long-range order of the pitches is observed. At <em>T<\/em> = 5 and below, the pitches show clear ordering. At <em>T<\/em> = 5 and <em>T<\/em> = 4, we see five- and sevenfold division of the octave, as was also seen in the mean field result. At <em>T<\/em> = 3.5, the 12-fold division of the octave has emerged, reproducing the Western musical system (for example, see fig. S4B showing the 12-fold division of the octave in both the frequency spectrum of the <em>T<\/em> = 3.5 tone lattice and the spectrum of Bach\u2019s <em>Prelude in D major, BWV850<\/em>). Now, we can delve deeper by studying the spatial arrangement of these pitches on the lattice. [See fig. S5 for plots of the distribution <em>C<\/em>(\u0394<em>x<\/em>, <em>r<\/em>) of intervals \u0394<em>x<\/em> = <em>x<sub>i<\/sub><\/em> \u2212 <em>x<sub>j<\/sub><\/em> versus distance <em>r<\/em> between sites <em>i<\/em> and <em>j<\/em>, at <em>T<\/em> = 5.5 and <em>T<\/em> = 3.5].<\/p><figure class=\"fig pos-float type-figure  odd figure figure--data\" id=\"F5\">\n    <figcaption class=\"fig-caption\" id=\"F5-caption\">\n    Fig. 5 Tone lattice simulation results.(<strong>A<\/strong>) Histograms of pitches <em>x<\/em> mod 1 in a metastable configuration of the tone lattice following a quench to temperature <em>T<\/em>. At <em>T<\/em> = 6, no ordering is observed. At <em>T<\/em> = 5,4, and 3.5, ordering is observed, with the octave divided into 5, 7, and 12, respectively. (<strong>B<\/strong>) Pitches on the tone lattice at <em>T<\/em> = 3.5. Pitch domains are labeled with pitch indices 0 to 11. Major or minor triads are marked with triangles. Junctions of more than three pitch domains are marked with circles. (<strong>C<\/strong>) The Tonnetz (edges shown in blue) with connections between neighboring pitch domains at <em>T<\/em> = 3.5 shown in red.  <\/figcaption><\/figure><p id=\"p-31\"><a id=\"xref-fig-5-2\" class=\"xref-fig\" href=\"#F5\">Figure 5B<\/a> shows a visualization of the pitches on the lattice in a metastable configuration following a quench to <em>T<\/em> = 3.5. Each pixel corresponds to one lattice site colored according to a hue\/saturation\/value triplet. The hue (ranging from blue to red, as shown on the color bar) represents the pitch class <em>x<sub>i<\/sub><\/em> mod 1, with the saturation and value adjusted to represent \u230a<em>x<sub>i<\/sub><\/em>\u230b, the octave in which that pitch lies. Darker (lighter) colors represent lower (higher) octaves. The image is characterized by domains of the same hue, indicating a single pitch class within that domain (pitch classes are labeled by indices 0 to 11). Repeated runs of the simulation result in final states that differ in the details but show the same general behavior, such as the same division of the octave, and similar correlations. Figure S6 shows simulation results at <em>T<\/em> = 4, 5, and 6.<\/p><p id=\"p-32\">The quenched tone lattice can be viewed as a metastable configuration of bound vortices and antivortices, as expected for the 2D XY model. Because pitches can vary continuously only at a high cost of dissonance, it is more favorable to find domains of nearly constant pitch class separated by domain boundaries, where the pitch class changes by a consonant interval. The most common interval across a boundary is a fourth or fifth (change in pitch index of 5 or 7 mod 12). Major and minor thirds are also common (change in pitch index of 3 or 4 mod 12). Places where two or more domain boundaries meet at a point constitute a topological defect, with the surrounding domains forming a vortex around it. The simplest vortex consists of three domains that meet at a point (see points marked with a triangle in <a id=\"xref-fig-5-3\" class=\"xref-fig\" href=\"#F5\">Fig. 5B<\/a>). The pitches of these domains consist of a root pitch, a major or minor third above the root, and a fifth above the root. This combination of three pitches forms a major or minor triad\u2014the type of chord that forms the backbone of Western music. Vortices with more domain boundaries are indicated by circles. (See the Supplementary Materials for sound clips of selected regions of the tone lattice.)<\/p><p id=\"p-33\">The arrangement of pitch domains on the lattice shown in <a id=\"xref-fig-5-4\" class=\"xref-fig\" href=\"#F5\">Fig. 5B<\/a> reflects elements of musical harmony. The Tonnetz (<a id=\"xref-fig-5-5\" class=\"xref-fig\" href=\"#F5\">Fig. 5C<\/a>), originally described by Euler (<a id=\"xref-ref-33-1\" class=\"xref-bibr\" href=\"#ref-33\"><em>33<\/em><\/a>, <a id=\"xref-ref-34-1\" class=\"xref-bibr\" href=\"#ref-34\"><em>34<\/em><\/a>), is a graph in which the nodes are pitch indices and the edges (shown in blue) represent fifths, major thirds, and minor thirds connecting those pitch classes. In the ET approximation, the Tonnetz can be viewed as a graph on a torus, with the connections around the torus indicated by the gray circles. Moving continuously to the right, we increase by fifths, cycling around the \u201ccircle of fifths.\u201d Moving along one diagonal direction changes by minor thirds, and the other diagonal direction changes by major thirds. Each set of three adjacent nodes forming a triangle represents a major or minor triad. The pitch classes on the tone lattice in <a id=\"xref-fig-5-6\" class=\"xref-fig\" href=\"#F5\">Fig. 5B<\/a> that share a border can be represented by a similar graph. We construct this graph first by identifying all pitch domains, defined as contiguous regions with at least five lattice sites having the same pitch, rounded to the nearest of the 12 pitch classes. Pitch domains that share a border are then identified as those that have at least two adjacent lattice sites in one domain bordering two adjacent lattice sites in the other domain. The resulting graph of all neighboring domains on the tone lattice (red lines in <a id=\"xref-fig-5-7\" class=\"xref-fig\" href=\"#F5\">Fig. 5C<\/a>) maps directly onto the Tonnetz. All possible fifths are present (for example, the ellipse in <a id=\"xref-fig-5-8\" class=\"xref-fig\" href=\"#F5\">Fig. 5B<\/a> traverses the circle of fifths), with many of the thirds present as well.<\/p><p id=\"p-34\">The major or natural minor diatonic scales used in traditional Western music consist of seven contiguous pitch classes around the circle of fifths (or a trapezoid on the Tonnetz), whereas the anhemitonic pentatonic scales used in other musical traditions consist of five contiguous pitches on the circle of fifths. Therefore, all such scales may be constructed from contiguous regions of the tone lattice, with major and minor triads within these scales appearing in these regions. Furthermore, as the tones on a finite lattice relax through the quenched metastable states toward equilibrium, some of the 12 pitch classes begin to grow and others shrink, possibly eventually disappearing. Because the domains are most often separated by fifths or fourths, it is likely that the domains that grow and do not disappear will be contiguous on the circle of fifths. Therefore, the pitches on the tone lattice are expected to converge toward a major or minor diatonic scale (and, upon further relaxation, to an anhemitonic pentatonic scale). <a id=\"xref-fig-6-1\" class=\"xref-fig\" href=\"#F6\">Figure 6A<\/a> shows a histogram of pitch classes from the <em>T<\/em> = 3.5 tone lattice, arranged in order of the circle of fifths. One can see that the first seven pitch classes are all fairly common (in the shaded box), and the three least common pitch classes are contiguous on the circle of fifths and outside the box. This type of pitch distribution is common in Western music, where many of the pitches are drawn from a major or minor diatonic scale (and thus from seven contiguous pitch classes on the circle of fifths), but with some exceptions. For example, the pitch distribution from Bach\u2019s <em>BWV850<\/em> is shown in <a id=\"xref-fig-6-2\" class=\"xref-fig\" href=\"#F6\">Fig. 6B<\/a>, in which all 12 pitch classes are present, but with the first 7 pitch classes (corresponding to the D major scale) occurring more frequently than the latter 5. In other runs of the simulation, or at different times within one run, one might observe peaks in the pitch distribution with more equal or less equal height (or some missing), which would agree better with other styles of music.<\/p><figure class=\"fig pos-float type-figure  odd figure figure--data\" id=\"F6\">\n    <figcaption class=\"fig-caption\" id=\"F6-caption\">\n    Fig. 6 Comparison of the pitch class distribution on the tone lattice and in a selected piece of music.(<strong>A<\/strong>) Histogram of pitch classes 12<em>x<\/em> mod 12 rounded to the nearest pitch class on the <em>T<\/em> = 3.5 tone lattice and arranged by ascending fifths. (<strong>B<\/strong>) Histogram of pitch classes appearing in Bach\u2019s <em>Prelude and Fugue in D major, BWV850<\/em>, arranged by ascending fifths. The shaded area corresponds to the notes of a diatonic scale.  <\/figcaption><\/figure><p id=\"p-36\">The arrangement of triads on the tone lattice gives rise to the same elements of music theory that have been extracted from the Tonnetz, such as a geometric interpretation of tonality (<a id=\"xref-ref-35-1\" class=\"xref-bibr\" href=\"#ref-35\"><em>35<\/em><\/a>), and progressions of triads that produce parsimonious voice leading, the concept that allows the combination of melody and harmony known as counterpoint (<a id=\"xref-ref-36-1\" class=\"xref-bibr\" href=\"#ref-36\"><em>36<\/em><\/a>). By varying the quench temperature <em>T<\/em>, pitch distributions with different numbers of pitch classes are observed. Inspecting the arrangement of pitch domains on lattices at different <em>T<\/em>, one observes certain commonly occurring intervals across domain boundaries. One can then propose \u201cTonnetzes\u201d in these musical systems by analogy to the usual Tonnetz with 12 pitch classes, where the most commonly occurring interval forms the horizontal edges and the less common intervals are represented by diagonal edges (see fig. S6).<\/p><p id=\"p-37\">The fact that the behavior of the tone lattice in two dimensions replicates many features of traditional Western music raises the question: What happens in other dimensions? No phase transition is expected in 1D (unless the interactions are extended to long range). In 3D, the expected phase transition of a system with XY symmetry has been studied extensively in the context of cosmology. In this case, the phase transition is not a Kosterlitz-Thouless transition but, instead, a normal first- or second-order transition. Still, though, one expects the emergence of topological defects in the ordered phase if the cooling takes place at a finite rate by the so-called Kibble-Zurek mechanism (<a id=\"xref-ref-37-1\" class=\"xref-bibr\" href=\"#ref-37\"><em>37<\/em><\/a>). In these topological defects, the point-like vortex cores seen in the 2D lattice are now extended along 1D paths, originally referred to as cosmic strings, and later observed in condensed matter systems (<a id=\"xref-ref-31-2\" class=\"xref-bibr\" href=\"#ref-31\"><em>31<\/em><\/a>, <a id=\"xref-ref-38-1\" class=\"xref-bibr\" href=\"#ref-38\"><em>38<\/em><\/a>). It is possible that the extension of the tone lattice to three dimensions may give rise to a 3D Tonnetz-like structure, as has been previously proposed in the spiral array model (<a id=\"xref-ref-35-2\" class=\"xref-bibr\" href=\"#ref-35\"><em>35<\/em><\/a>), or in generalized Tonnetzes (<a id=\"xref-ref-3-2\" class=\"xref-bibr\" href=\"#ref-3\"><em>3<\/em><\/a>).<\/p><p id=\"p-38\">The results presented here demonstrate that many patterns in music can emerge from this statistical mechanics framework. Undoubtedly, further insights into the structure of music can be gained by studying related models and bringing all of the tools of statistical mechanics to bear. The historical development of thermodynamics provides a parallel to the present case. Starting in the 17th century, empirical laws (Boyle\u2019s law, Charles\u2019s law, etc.) were postulated to explain observations of the behavior of gases. This top-down approach to thermodynamics proved quite useful. However, the development of statistical mechanics yielded a bottom-up theory, and this fundamental understanding led to an array of new discoveries in the 20th century and beyond. Likewise, top-down music theory has proven quite useful for composing and understanding music, but a bottom-up theory provides a foundational understanding, perhaps leading to new ways of composing, understanding, and enjoying music.<\/p><\/div>","created_at":"2019-05-25T17:14:50+0000","updated_at":"2019-05-25T17:14:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":19,"domain_name":"advances.sciencemag.org","preview_picture":"https:\/\/advances.sciencemag.org\/sites\/default\/files\/highwire\/advances\/5\/5.cover-source.gif","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["wearables"],"is_public":false,"id":957,"title":"Twiddler 3 by Tek Gear","url":"https:\/\/twiddler.tekgear.com\/","content":"Twiddler 3 by Tek Gear\n\n<div id=\"transcroller-body\" class=\"aos-all\"><p><a href=\"http:\/\/forum.tekgear.com\/t\/looking-for-beta-testers-twiddler-firmware-12u-and-new-tuner\/232\">Firmware Beta Available<\/a><\/p><\/div>\n\n<section id=\"contact\" class=\"v3\"><div class=\"container\"><div class=\"row\"><div class=\"col-lg-12\"><p>We\u2019ve put together some additional background for you on the Twiddler.<\/p><ul><li>What People are Saying: <a href=\"https:\/\/twiddler.tekgear.com\/pdf\/aboutTwiddler3.pdf\">about Twiddler 3<\/a><\/li>\n        <li>World's Most Mobile Keyboard: <a href=\"http:\/\/www.prweb.com\/releases\/Twiddler3\/keyboard\/prweb12702182.htm\">News Release<\/a><\/li>\n        <li>Introducing the Twiddler3: <a href=\"https:\/\/twiddler.tekgear.com\/pdf\/TwiddlerMediaKit-Intro.pdf\">Your tiny keyboard-to-go<\/a><\/li>\n        <li>The design behind the Twiddler3: <a href=\"https:\/\/twiddler.tekgear.com\/pdf\/TwiddlerMediaKit-design.pdf\">Ergonomic design<\/a> <\/li>\n        <li>Working with wearables: <a href=\"https:\/\/twiddler.tekgear.com\/pdf\/TwiddlerMediaKit-wearables.pdf\">Connecting with any Bluetooth device<\/a><\/li>\n        <li>How to use the Twiddler: <a href=\"https:\/\/twiddler.tekgear.com\/pdf\/TwiddlerMediaKit-howTo.pdf\">A primer on chording<\/a><\/li>\n      <\/ul><\/div><\/div><\/div><\/section>","created_at":"2019-05-23T19:08:55+0000","updated_at":"2019-05-23T19:08:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"twiddler.tekgear.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["wearables"],"is_public":false,"id":956,"title":"Vufine","url":"https:\/\/store.vufine.com\/","content":"<main class=\"main-content\" id=\"MainContent\" role=\"main\"><div class=\"main-home\"><div id=\"shopify-section-home-page-banner\" class=\"shopify-section\"><div class=\"main-home-banner-wrap\"><p>\n    <a href=\"https:\/\/store.vufine.com\/products\/vufine-wearable-display-2\" class=\"main-home-banner-button\">\n      Learn More\n      <i class=\"fa fa-chevron-right\">\n    <\/i><\/a><\/p><\/div><\/div><div class=\"main-home-center\"><div class=\"main-home-container\"><div class=\"main-home-center-top\"><p class=\"main-home-center-text\">\n                    Vufine+ is a high definition, wearable display that seamlessly integrates with your technology. Whether for your smartphone, laptop or drone, Vufine+ connects via HDMI\n                    to provide a second monitor or mirrored display. Use Vufine+ anywhere you need an extra screen without taking your eyes off the world around you.\n                <\/p><\/div><div class=\"main-home-center-wrap\"><div class=\"main-home-center-left\"><div class=\"main-home-service-block\"><i class=\"fa fa-tv\">Optimized Display ModesFor the largest display across all your devices<\/i><\/div><div class=\"main-home-service-block\"><h3 class=\"main-home-service-title\">HDMI Compatible<\/h3><p class=\"main-home-service-text\">Compatible with any device capable of outputting a 720p HDMI signal<\/p><\/div><div class=\"main-home-service-block\"><i class=\"fa fa-bicycle\">Highly PortableSmall and incredibly tough<\/i><\/div><\/div><div class=\"main-home-center-mid\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1629\/7313\/t\/9\/assets\/Vufine_Assembly.jpg?2\" alt=\"image\" \/><\/div><div class=\"main-home-center-right\"><div class=\"main-home-service-block\"><i class=\"fa fa-magnet\">Magnetic Docking StationEasily adjustable and incredibly stable<\/i><\/div><div class=\"main-home-service-block\"><i class=\"fa fa-battery-full\">Internal BatteryFor a 90 minute runtime<\/i><\/div><div class=\"main-home-service-block\"><i class=\"fa fa-star\">Sleek DesignAn elegant, comfortable design<\/i><\/div><\/div><\/div><\/div><\/div><div class=\"main-home-bottom\"><div class=\"main-home-bottom-wrap\"><div class=\"main-home-bottom-block\"><h3>Streamlined Functionality<\/h3><p>\n                    Paring Vufine down to its most essential functionality allowed us to avoid privacy\n                    concerns, a confusing new interface and prohibitive pricing.\n                <\/p><\/div><div class=\"main-home-bottom-block main-home-bottom-block-mid\"><h3>Exceptional Convenience<\/h3><p>\n                    Whether for your smartphone, laptop, or drone, Vufine is a simple and highly accessible\n                    companion device that redefines your experience with technology.\n                <\/p><\/div><div class=\"main-home-bottom-block\"><h3>Endless Opportunity<\/h3><p>\n                    Vufine facilitates your engagement with technology across a near limitless set of use\n                    cases so, no matter the scenario, your hands are free to do the real work.\n                <\/p><\/div><\/div><div class=\"main-home-container\"><div class=\"main-home-bottom-down\"><div class=\"left-wrap\"><h3>Now Available to Order for $199<\/h3><p>\n                        Even if you missed out on our highly successful Kickstarter campaign, Vufine+ is shipping and available to order worldwide.\n                    <\/p><\/div><a href=\"https:\/\/store.vufine.com\/collections\/all\" class=\"main-home-bottom-button\">Learn More<\/a><\/div><\/div><\/div><\/div>\n    <\/main>","created_at":"2019-05-23T19:08:36+0000","updated_at":"2019-05-23T19:08:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"store.vufine.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":955,"title":"David Beazley - Lambda Calculus from the Ground Up - PyCon 2019","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=pkCLMl0e_0k","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/pkCLMl0e_0k?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-05-23T02:34:40+0000","updated_at":"2019-05-23T02:34:40+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/pkCLMl0e_0k\/maxresdefault.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":954,"title":"Builtin Superheroes (Screencast)","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=j6VSAsKAj98","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/j6VSAsKAj98?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-05-23T02:34:27+0000","updated_at":"2019-05-23T02:34:27+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/j6VSAsKAj98\/maxresdefault.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["lisp","racket","scheme"],"is_public":false,"id":953,"title":"Why Racket? Why Lisp? | Butterick\u2019s Practical Typography","url":"https:\/\/practicaltypography.com\/why-racket-why-lisp.html","content":"<p>This book was made pos\u00adsi\u00adble by a pub\u00adlish\u00ading sys\u00adtem called <a href=\"http:\/\/pollenpub.com\">Pollen.<\/a> I cre\u00adated Pollen with the <a href=\"http:\/\/racket-lang.org\">Racket<\/a> pro\u00adgram\u00adming lan\u00adguage. Racket is a de\u00adscen\u00addant of <a href=\"http:\/\/schemers.org\/\">Scheme,<\/a> which in turn is a de\u00adscen\u00addant of\u00a0Lisp.<\/p><aside><p>An <a href=\"https:\/\/beautifulracket.com\/appendix\/why-racket-why-lisp.html\">up\u00addated ver\u00adsion<\/a> of this piece is part of <a href=\"https:\/\/beautifulracket.com\/\">Beau\u00adti\u00adful Racket,<\/a> my book about mak\u00ading pro\u00adgram\u00adming lan\u00adguages with Racket.<\/p><\/aside><p>So while Racket is not Lisp (in the spe\u00adcific Com\u00admon Lisp sense), it is <em>a<\/em> Lisp (in the fa\u00admil\u00adial sense) which means that its core ideas\u2014and core virtues\u2014are shared with Lisp. So talk\u00ading about Racket means talk\u00ading about\u00a0Lisp.<\/p><p>In prac\u00adti\u00adcal pro\u00adgram\u00adming projects, Lisps are rare, and Racket es\u00adpe\u00adcially so. Thus, be\u00adfore I em\u00adbarked on my Lisp ad\u00adven\u00adture, I wanted to un\u00adder\u00adstand the costs &amp; ben\u00ade\u00adfits of us\u00ading a Lisp. Why do Lisps have such a great rep\u00adu\u00adta\u00adtion, but so few users? Was I see\u00ading some\u00adthing every\u00adone else missed? Or did they know some\u00adthing I didn\u2019t? To find out, I read what\u00adever I could find about Lisps, in\u00adclud\u00ading Paul Gra\u00adham\u2019s <em>Hack\u00aders &amp; Painters<\/em> and Pe\u00adter Seibel\u2019s <em>Prac\u00adti\u00adcal Com\u00admon Lisp<\/em>. (OK, parts. It\u2019s a big book.)<\/p><p>What I found was plenty of Lisp flat\u00adtery from ex\u00adpert Lisp pro\u00adgram\u00admers. (Also plenty of Lisp kvetch\u00adery from its de\u00adtrac\u00adtors.) What I didn\u2019t find were sim\u00adple, per\u00adsua\u00adsive ar\u00adgu\u00adments in its fa\u00advor. So here\u2019s why Racket was the right tool for this project, and what I see as the prac\u00adti\u00adcal virtues of Lisps in general.<\/p><p><a href=\"#consider-the-source\">Consider the source<\/a><\/p><p>I didn\u2019t study com\u00adputer sci\u00adence in col\u00adlege (though I was a math ma\u00adjor for two years, be\u00adfore switch\u00ading to de\u00adsign). I\u2019ve never held an of\u00adfi\u00adcial job as a pro\u00adgram\u00admer. Rather, pro\u00adgram\u00adming has been a sec\u00adondary skill I\u2019ve used in my work as a web de\u00adsigner, type de\u00adsigner, and writer.<\/p><p>These days, I spend an in\u00adcreas\u00ading amount of my time pro\u00adgram\u00adming. This pro\u00adgram\u00adming gen\u00ader\u00adates in\u00adcome. So by the sim\u00adplest de\u00adf\u00adi\u00adn\u00adi\u00adtion\u2014does the skill make you money?\u2014I sup\u00adpose I qual\u00adify as a pro\u00adfes\u00adsional pro\u00adgram\u00admer. And since most of my pro\u00adgram\u00adming ef\u00adforts are in Racket, I qual\u00adify as a pro\u00adfes\u00adsional Racket programmer.<\/p><p>Mind you, I\u2019m not claim\u00ading that I\u2019m a <em>good<\/em> pro\u00adgram\u00admer. (You can <a href=\"http:\/\/pollenpub.com\">de\u00adcide for your\u00adself.<\/a>) Among the Racket com\u00admu\u00adnity, which is laden with com\u00adputer-sci\u00adence Ph\u00adDs &amp; pro\u00adfes\u00adsors, I (have no choice but to) em\u00adbrace my rel\u00ada\u00adtive ig\u00adno\u00adrance. Hence the ti\u00adtle of my talk at <a href=\"http:\/\/con.racket-lang.org\/\">Racket\u00adCon 2014:<\/a> <em>Like a Blind Squir\u00adrel in a Fer\u00adrari<\/em>.<\/p><p>Yet de\u00adspite my flaws as a pro\u00adgram\u00admer, with Racket I\u2019ve been able to ren\u00adder big\u00adger ideas into pro\u00adgrams more quickly, and with fewer bugs, than any lan\u00adguage I\u2019ve used be\u00adfore (and there have been many\u2014Ba\u00adsic, C, C++, Perl, Java, Java\u00adScript, Python, and oth\u00aders). Since I haven\u2019t got\u00adten a brain trans\u00adplant re\u00adcently, there must be some\u00adthing spe\u00adcial about Racket as a language.<\/p><p><a href=\"#if-youre-not-familiar-with-lisp\">If you\u2019re not familiar with Lisp<\/a><\/p><p>Lisp is a lan\u00adguage most pro\u00adgram\u00admers have heard of, for two rea\u00adsons. First, it\u2019s one of the old\u00adest com\u00adputer lan\u00adguages, in use since 1958. Sec\u00adond, it\u2019s ac\u00adcrued a rep\u00adu\u00adta\u00adtion as a lan\u00adguage for braini\u00adacs. Orig\u00adi\u00adnally this rep\u00adu\u00adta\u00adtion arose from its as\u00adso\u00adci\u00ada\u00adtion with the field of ar\u00adti\u00adfi\u00adcial in\u00adtel\u00adli\u00adgence. Since then, this rep\u00adu\u00adta\u00adtion has been main\u00adtained by pe\u00adri\u00adodic en\u00addorse\u00adments from re\u00adspected pro\u00adgram\u00admers (lat\u00adterly, <a href=\"http:\/\/www.catb.org\/esr\/faqs\/hacker-howto.html\">Eric Ray\u00admond<\/a> and <a href=\"http:\/\/www.paulgraham.com\/avg.html\">Paul Gra\u00adham<\/a>) and the en\u00addur\u00ading fame of the text\u00adbook used in in\u00adtro\u00adduc\u00adtory com\u00adputer-sci\u00adence courses at MIT, <em><a href=\"http:\/\/mitpress.mit.edu\/sicp\/\">Struc\u00adture and In\u00adter\u00adpre\u00adta\u00adtion of Com\u00adputer Pro\u00adgrams<\/a><\/em> (which uses Scheme, and that one I did read start to finish).<\/p><p>But as main\u00adstream pro\u00adgram\u00adming tools, Lisp and its de\u00adscen\u00addants have been largely ig\u00adnored. Pop\u00adu\u00adlar\u00adity of pro\u00adgram\u00adming lan\u00adguages is tricky to mea\u00adsure, but here\u2019s a sim\u00adple proxy\u2014let\u2019s count the num\u00adber of projects cur\u00adrently hosted on <a href=\"http:\/\/github.com\">Git\u00adHub.<\/a> One could quib\u00adble about the ac\u00adcu\u00adracy of this method, ex\u00adcept that the re\u00adsults aren\u2019t even close:<\/p><table><thead><tr><th class=\"no-hyphens\">Language<\/th><th class=\"no-hyphens\">GitHub projects<\/th><\/tr><\/thead><tbody><tr><td>JavaScript<\/td><td>885,467<\/td><\/tr><tr><td>Java<\/td><td>644,711<\/td><\/tr><tr><td>Ruby<\/td><td>622,088<\/td><\/tr><tr><td>Python<\/td><td>432,533<\/td><\/tr><tr><td>PHP<\/td><td>420,913<\/td><\/tr><tr><td>C<\/td><td>218,287<\/td><\/tr><tr><td>Clojure<\/td><td>21,108<\/td><\/tr><tr><td>Common Lisp<\/td><td>6883<\/td><\/tr><tr><td>Scheme<\/td><td>5128<\/td><\/tr><tr><td>Racket<\/td><td>2727<\/td><\/tr><\/tbody><\/table><p>The last four lan\u00adguages are Lisps, and to\u00adgether ac\u00adcount for only 35,846 projects. Racket it\u00adself only ac\u00adcounts for a small frac\u00adtion of this small fraction.<\/p><aside><p>\u201c[Java\u00adScript] has a lot of stu\u00adpid in it \u2026\u00a0<br \/>The good parts of [Java\u00adScript] go back to Scheme and Self\u201d<br \/>\u2014Bren\u00addan Eich,<br \/><a href=\"http:\/\/www.jwz.org\/blog\/2010\/10\/every-day-i-learn-something-new-and-stupid\/#comment-1020\">here<\/a> and <a href=\"http:\/\/www.jwz.org\/blog\/2010\/10\/every-day-i-learn-something-new-and-stupid\/#comment-1089\">here<\/a><\/p><\/aside><p>Pop\u00adu\u00adlar pro\u00adgram\u00adming lan\u00adguages aren\u2019t nec\u00ades\u00adsar\u00adily good\u2014look what\u2019s at the top of that list\u2014but un\u00adpop\u00adu\u00adlar lan\u00adguages of\u00adten have fa\u00adtal flaws that pre\u00advent wider adop\u00adtion. As I was con\u00adsid\u00ader\u00ading lan\u00adguages, Racket had a lot to rec\u00adom\u00admend it. But was there a fa\u00adtal flaw I was over\u00adlook\u00ading? And by com\u00admit\u00adting to a Lisp, would I be paint\u00ading my\u00adself into a cor\u00adner? I wanted to un\u00adder\u00adstand the risks and benefits.<\/p><p><a href=\"#flattery-gets-you-nowhere\">Flattery gets you nowhere<\/a><\/p><p>I said above that Lisp flat\u00adtery is easy to find. The prob\u00adlem with Lisp flat\u00adtery is that it makes sense only to ex\u00adpe\u00adri\u00adenced Lisp pro\u00adgram\u00admers. To oth\u00aders\u2014es\u00adpe\u00adcially those who are try\u00ading to de\u00adcide whether to learn and use a Lisp\u2014it just comes across as un\u00adsub\u00adstan\u00adti\u00adated hoodoo.<\/p><p>For ex\u00adam\u00adple, in his es\u00adsay <em><a href=\"http:\/\/www.catb.org\/esr\/faqs\/hacker-howto.html\">How to Be\u00adcome a Hacker,<\/a><\/em> Eric Ray\u00admond says \u201cLisp is worth learn\u00ading for \u2026 the pro\u00adfound en\u00adlight\u00aden\u00adment ex\u00adpe\u00adri\u00adence you will have when you fi\u00adnally get it. That ex\u00adpe\u00adri\u00adence will make you a bet\u00adter pro\u00adgram\u00admer for the rest of your days, even if you never ac\u00adtu\u00adally use Lisp it\u00adself a lot.\u201d Un\u00adfor\u00adtu\u00adnately Ray\u00admond doesn\u2019t fol\u00adlow up this claim with a de\u00adscrip\u00adtion of the \u201cen\u00adlight\u00aden\u00adment ex\u00adpe\u00adri\u00adence,\u201d nor why it\u2019s \u201cpro\u00adfound,\u201d nor how it will im\u00adprove your pro\u00adgram\u00adming skills generally.<\/p><p>To be fair, Ray\u00admond\u2019s es\u00adsay is not fo\u00adcused on Lisp. But com\u00adpare <em><a href=\"http:\/\/www.paulgraham.com\/avg.html\">Beat\u00ading the Av\u00ader\u00adages,<\/a><\/em> by Paul Gra\u00adham, which is. Gra\u00adham starts off by cit\u00ading Ray\u00admond\u2019s com\u00adpli\u00adment to Lisp and seems ready to make the claim concrete.<\/p><p>In\u00adstead, he breaks it into smaller chunks of flat\u00adtery. \u201cWe knew Lisp was a really good lan\u00adguage for writ\u00ading soft\u00adware quickly.\u201d Be\u00adcause of what char\u00adac\u00adter\u00adis\u00adtics? He doesn\u2019t say, but then de\u00adscribes Lisp as his \u201cse\u00adcret weapon.\u201d OK, so what\u2019s the se\u00adcret? He says \u201cpro\u00adgram\u00adming lan\u00adguages vary in power.\u201d Fine, but what ex\u00adactly makes Lisp more powerful?<\/p><p>Gra\u00adham of\u00adfers one con\u00adcrete ex\u00adam\u00adple: Lisp\u2019s macro fa\u00adcil\u00adity, which he de\u00adscribes as its abil\u00adity to make \u201cpro\u00adgrams that write pro\u00adgrams.\u201d Af\u00adter 18 months us\u00ading a Lisp lan\u00adguage, I\u2019d agree with Gra\u00adham that macros are great when you need them. But for some\u00adone new to Lisp lan\u00adguages, they\u2019re not a bread-and-but\u00adter ben\u00ade\u00adfit. (They only make it to #8 on my list of fa\u00advorite fea\u00adtures below.)<\/p><p>I was hope\u00adful when I opened Pe\u00adter Seibel\u2019s <em><a href=\"http:\/\/www.gigamonkeys.com\/book\">Prac\u00adti\u00adcal Com\u00admon Lisp<\/a><\/em> and saw that the <a href=\"http:\/\/www.gigamonkeys.com\/book\/introduction-why-lisp.html\">in\u00adtro\u00adduc\u00adtion<\/a> was sub\u00adti\u00adtled \u201cWhy Lisp?\u201d Yes, tell me! Seibel echoes Gra\u00adham\u2019s claim: \u201cYou\u2019ll get more done, faster, us\u00ading [Lisp] than you would us\u00ading pretty much any other lan\u00adguage.\u201d OK, but how? Seibel won\u00adders whether \u201cI like Lisp be\u00adcause of some quirk in the way my brain is wired. It could even be ge\u00adnetic, since my dad has it too.\u201d That\u2019s not en\u00adcour\u00adag\u00ading to those of us out\u00adside your fam\u00adily. Ul\u00adti\u00admately, he sums up the ap\u00adpeal of Lisp by de\u00adscrib\u00ading it as \u201cthe pro\u00adgram\u00adma\u00adble pro\u00adgram\u00adming lan\u00adguage.\u201d But I\u2019ve never\u00a0used a pro\u00adgram\u00adma\u00adble pro\u00adgram\u00adming lan\u00adguage. Why should I start?<\/p><p>And by the way, when do I get the speed and power you keep promising?<\/p><p>In short\u2014<em>what\u2019s in it for me,\u00a0now?<\/em><\/p><p>This is the fun\u00adda\u00admen\u00adtal ques\u00adtion that Lisp ad\u00advo\u00adcates have to an\u00adswer for new users. But more of\u00adten, it\u2019s side\u00adstepped. I\u2019m not pick\u00ading on Ray\u00admond or Gra\u00adham or Seibel. They\u2019re ex\u00adcel\u00adlent writ\u00aders, and as pro\u00adgram\u00admers, they\u2019re way out of my league. As I learn more about Lisps, I re\u00adturn to these pieces and they make more sense.<\/p><p>But these pieces are also em\u00adblem\u00adatic of a gen\u00aderal weak\u00adness of mes\u00adsag\u00ading about Lisp. I say that not as a rank\u00ading mem\u00adber of the Lisp com\u00admu\u00adnity, but rather as some\u00adone who spent a lot of time seek\u00ading an an\u00adswer to that fun\u00adda\u00admen\u00adtal ques\u00adtion. I never got\u00a0it.<\/p><p>Seibel is pass\u00ading the buck when he says that to un\u00adder\u00adstand the ben\u00ade\u00adfits of Lisp, \u201cyou\u2019re go\u00ading to have to learn some Lisp and see for your\u00adself.\u201d Sure, this method works\u2014us\u00ading Racket for a few months fi\u00adnally made the ben\u00ade\u00adfits of Lisp clear to me. But it also re\u00adquires an in\u00advest\u00adment of about 100\u2013200 hours.<\/p><aside><p>For more on the perils of taxing reader patience, see<br \/><a href=\"https:\/\/practicaltypography.com\/why-does-typography-matter.html\" class=\"xref\">why does typography matter<\/a>.<\/p><\/aside><p>That\u2019s ask\u00ading too much. If Lisp lan\u00adguages are so great, then it should be pos\u00adsi\u00adble to sum\u00adma\u00adrize their ben\u00ade\u00adfits in con\u00adcise, prac\u00adti\u00adcal terms. If Lisp ad\u00advo\u00adcates refuse to do this, then we shouldn\u2019t be sur\u00adprised when these lan\u00adguages re\u00admain stuck near the bot\u00adtom of the charts.<\/p><p><a href=\"#so-really--whats-in-it-for-me-now\">So really\u2014what\u2019s in it for me, now?<\/a><\/p><p>In a word, <em>ex\u00adpres\u00adsive\u00adness:<\/em> the mea\u00adsure of how easy it is to put your ideas into code. For in\u00adstance, an ex\u00adpres\u00adsive lan\u00adguage like Racket lets you write the \u201cHello world\u201d pro\u00adgram like\u00a0this:<\/p><p><code>\"Hello world\"<\/code><\/p><p>Whereas a less ex\u00adpres\u00adsive lan\u00adguage\u2014I won\u2019t name names\u2014re\u00adquires\u00a0this:<\/p><p><code>public class HelloWorld {<br \/>\u00a0\u00a0\u00a0\u00a0public static void main(String[] args) {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0System.out.println(\"Hello world\");<br \/>\u00a0\u00a0\u00a0\u00a0}<br \/>}<\/code><\/p><p>Con\u00adci\u00adsion is valu\u00adable, but ex\u00adpres\u00adsive\u00adness also em\u00adbod\u00adies other qual\u00adi\u00adties: pre\u00adci\u00adsion, read\u00adabil\u00adity, flex\u00adi\u00adbil\u00adity, po\u00adten\u00adtial for abstraction.<\/p><aside><p>Me, be\u00adfore Racket.<\/p><\/aside><p class=\"image\"><img src=\"https:\/\/practicaltypography.com\/images\/racetrack.jpg\" alt=\"image\" \/><\/p><p>Com\u00adpared to other lan\u00adguages, Lisps are tremen\u00addously ex\u00adpres\u00adsive. Like the over\u00adpow\u00adered Japan\u00adese mo\u00adtor\u00adcy\u00adcle I once owned, they go where you want, very quickly, with a min\u00adi\u00admum of in\u00adput. If you\u2019ve rid\u00adden a mo\u00adtor\u00adcy\u00adcle, then you know what I mean. If you haven\u2019t, good news\u2014Lisps are cheaper and safer.<\/p><p>Here\u2019s my rank\u00ading of the lan\u00adguage fea\u00adtures that have of\u00adfered the most im\u00adme\u00addi\u00adate value to me, a pro\u00adgram\u00admer new to the Lisp world. For each, I\u2019ve noted whether it\u2019s a fea\u00adture of Racket specif\u00adi\u00adcally, or Lisps generally.<\/p><ol><li><p><strong>Every\u00adthing is an ex\u00adpres\u00adsion.<\/strong> [Lisps] Most pro\u00adgram\u00adming lan\u00adguages are a com\u00adbi\u00adna\u00adtion of two syn\u00adtac\u00adti\u00adcally dis\u00adtinct in\u00adgre\u00addi\u00adents: <em>ex\u00adpres\u00adsions<\/em> (things that are eval\u00adu\u00adated to pro\u00adduce a value) and <em>state\u00adments<\/em> (things that ex\u00adpress an ac\u00adtion). For in\u00adstance, in Python, <code>x = 1<\/code> is a state\u00adment, and <code>(x + 1)<\/code> is an expression.<\/p><p>State\u00adments and ex\u00adpres\u00adsions are dis\u00adtinct be\u00adcause while ex\u00adpres\u00adsions can be nat\u00adu\u00adrally nested with each other, state\u00adments and ex\u00adpres\u00adsions can\u00adnot. For in\u00adstance, in Python, this is still a valid expression:<\/p><p><code>(x + (y +\u00a01))<\/code><\/p><p>but this is\u00a0not:<\/p><p><code>(x + (if is_true(): 1 else:\u00a02))<\/code><\/p><p>be\u00adcause the if\u2013else con\u00addi\u00adtional is a state\u00adment, and can only be used in cer\u00adtain positions.<\/p><p>By mak\u00ading every\u00adthing an ex\u00adpres\u00adsion, how\u00adever, Lisps re\u00admove this lim\u00adi\u00adta\u00adtion. Since ex\u00adpres\u00adsions are nestable, any\u00adthing in the lan\u00adguage can be com\u00adbined with nearly any\u00adthing else. For in\u00adstance, be\u00adcause an if\u2013else con\u00addi\u00adtional is an ex\u00adpres\u00adsion, you can use it in place of a value:<\/p><aside><p>I\u2019m us\u00ading in\u00adfix no\u00adta\u00adtion to make a vi\u00adsual anal\u00adogy. Lisps typ\u00adi\u00adcally use pre\u00adfix no\u00adta\u00adtion, with the func\u00adtion at the front: <code>(+ x (+ y\u00a01))<\/code><\/p><\/aside><p><code>(x . + . (if (is_true) 1\u00a02)<\/code><\/p><p>You could also use a con\u00addi\u00adtional in place of an operator:<\/p><p><code>(x . (if (wants_sum) + *) .\u00a01)<\/code><\/p><p>You could even nest an\u00adother con\u00addi\u00adtional within\u00a0that:<\/p><p><code>(x . (if ((if (cond) cond_1 cond_2))\u00a0...)<\/code><\/p><p>And so forth. This is a syn\u00adthetic ex\u00adam\u00adple. The point is not that you\u2019d nec\u00ades\u00adsar\u00adily want to do this, but that Lisps per\u00admit it. As a pro\u00adgram\u00admer, this both sim\u00adpli\u00adfies your work (be\u00adcause every\u00adthing snaps to\u00adgether eas\u00adily) and ex\u00adpands your pos\u00adsi\u00adbil\u00adi\u00adties (be\u00adcause you can com\u00adbine parts of the lan\u00adguage in un\u00adusual ways if you feel like\u00a0it).<\/p><p>It\u2019s sim\u00adi\u00adlar to the ba\u00adsic idea be\u00adhind Le\u00adgos. Other build\u00ading sets of\u00adfer spe\u00adcial\u00adized pieces that can only fit to\u00adgether cer\u00adtain ways. But by shar\u00ading uni\u00adform mea\u00adsure\u00adments, Lego bricks of\u00adfer max\u00adi\u00admum pos\u00adsi\u00adbil\u00adi\u00adties for com\u00adbi\u00adna\u00adtions. This ends up be\u00ading more flex\u00adi\u00adble &amp; more\u00a0fun.<\/p><p>So it is with an ex\u00adpres\u00adsion-based lan\u00adguage. If you find this idea ex\u00adcit\u00ading, con\u00adgrat\u00adu\u00adla\u00adtions\u2014you might be a Lisp pro\u00adgram\u00admer. (If you find this idea weird and scary, this is a good mo\u00adment to <a href=\"http:\/\/www.buzzfeed.com\/search?q=puppies\">bail out.<\/a>)<\/p><\/li><li><p><strong>Every ex\u00adpres\u00adsion is ei\u00adther a sin\u00adgle value or a list.<\/strong> [Lisps] Sin\u00adgle val\u00adues are things like num\u00adbers and strings and hash ta\u00adbles. (In Lisps, they\u2019re some\u00adtimes called <em>atoms<\/em>.) That part is no big\u00a0deal.<\/p><aside><p>The name \u201cLisp\u201d is de\u00adrived from \u201clist processing.\u201d<\/p><\/aside><p>The list part, how\u00adever, is a big deal. In a lan\u00adguage like Python, the list is one data type within the lan\u00adguage. But in Lisps, the list is more like an or\u00adga\u00adniz\u00ading prin\u00adci\u00adple for every\u00adthing that hap\u00adpens. So yes, you can use the list as a data type. But a func\u00adtion call is also a list. In fact, the source code for the func\u00adtion is a list. Ac\u00adtu\u00adally, the rest of the pro\u00adgram is too. Lists are every\u00adwhere. (The fancy CS term for this prop\u00aderty is <em>ho\u00admoiconic\u00adity<\/em>.)<\/p><p>The ben\u00ade\u00adfits of lists are sim\u00adi\u00adlar to that of ex\u00adpres\u00adsions. By bring\u00ading more of the lan\u00adguage into a con\u00adsis\u00adtent form, more pos\u00adsi\u00adbil\u00adi\u00adties arise for how pieces can be com\u00adbined and manipulated.<\/p><p>Seibel de\u00adscribes Lisp as a tool for get\u00adting \u201cmore done, faster.\u201d Here, you can start to see why this is so. Lisp lan\u00adguages are im\u00admensely flex\u00adi\u00adble and per\u00admis\u00adsive in how their pieces can be con\u00adnected. This means that the way you <em>think<\/em> about a pro\u00adgram\u00adming prob\u00adlem can be quite close to the way you ac\u00adtu\u00adally pro\u00adgram it. (This is also why Lisps have tra\u00addi\u00adtion\u00adally ex\u00adcelled for pro\u00adto\u00adtypes and ex\u00adploratory work.)<\/p><p>To be fair, get\u00adting the most out of a Lisp means learn\u00ading to think more in the Lisp id\u00adiom of lists and ex\u00adpres\u00adsions. So I can see why Seibel says that try\u00ading it your\u00adself is the best way to be con\u00advinced of the ben\u00ade\u00adfits. As you get a feel for lists and ex\u00adpres\u00adsions, it does pay in\u00adcreas\u00ading div\u00adi\u00addends through\u00adout the lan\u00adguage. You see how tiny lines of code can pro\u00adduce epic amounts of work. You also start to ap\u00adpre\u00adci\u00adate that even in a well-de\u00adsigned lan\u00adguage like Python, you\u2019re spend\u00ading a lot of time shap\u00ading your ideas to fit its lim\u00adi\u00adta\u00adtions, like shav\u00ading an in\u00advis\u00adi\u00adble\u00a0yak.<\/p><\/li><li><p><strong>Func\u00adtional pro\u00adgram\u00adming.<\/strong> [Lisps] Yes, I know that other lan\u00adguages of\u00adfer func\u00adtional-pro\u00adgram\u00adming fea\u00adtures, and that Lisps aren\u2019t con\u00adsid\u00adered pure func\u00adtional lan\u00adguages. But many pro\u00adgram\u00admers haven\u2019t been ex\u00adposed to this id\u00adiom, and thus tend to un\u00adder\u00adrate its ben\u00ade\u00adfits. I know I was in that category.<\/p><p>Func\u00adtional pro\u00adgram\u00adming doesn\u2019t mean pro\u00adgram\u00adming with func\u00adtions. Every\u00adbody does that. Func\u00adtional pro\u00adgram\u00adming refers to a stricter style where func\u00adtions re\u00adceive cer\u00adtain data as in\u00adput, process only that data, and re\u00adturn a re\u00adsult. In func\u00adtional pro\u00adgram\u00adming, func\u00adtions avoid two habits com\u00admon in other lan\u00adguages: <em>data mu\u00adta\u00adtion<\/em> (= chang\u00ading data in-place rather than re\u00adturn\u00ading a value) and re\u00adly\u00ading on <em>state<\/em> (= ex\u00adtra con\u00adtext that\u2019s not pro\u00advided as in\u00adput, for in\u00adstance global variables).<\/p><p>\u201cWait\u2014I love state and data mu\u00adta\u00adtion. Why would you take them away?\u201d Be\u00adcause they\u2019re false friends. They con\u00adtra\u00addict the es\u00adsen\u00adtial con\u00adcept of a func\u00adtion, which is to en\u00adcap\u00adsu\u00adlate data and al\u00adgo\u00adrithms. When a func\u00adtion re\u00adlies on state or mu\u00adta\u00adtion, it\u2019s op\u00ader\u00adat\u00ading out\u00adside those bound\u00adaries. You ei\u00adther take on an in\u00adcreas\u00ading house\u00adkeep\u00ading bur\u00adden to keep track of how func\u00adtions af\u00adfect each other, or watch the pro\u00adgram sink into a swamp of mys\u00adte\u00adri\u00adous, com\u00adpli\u00adcated\u00a0bugs.<\/p><p>Pro\u00adgram\u00adming in a func\u00adtional style takes more ef\u00adfort at the out\u00adset. But it en\u00adcour\u00adages you to struc\u00adture the pro\u00adgram in a clean, com\u00adpart\u00admen\u00adtal\u00adized way. This pays off im\u00adme\u00addi\u00adately in pro\u00adgrams that are eas\u00adier to test and de\u00adbug. It\u2019s also more likely to lead to reusable com\u00adpo\u00adnents, since func\u00adtions are truly independent.<\/p><p>This bite-the-bul\u00adlet as\u00adpect of func\u00adtional pro\u00adgram\u00adming is an\u00adother rea\u00adson why you can get \u201cmore done, faster\u201d with a Lisp. The dif\u00adfer\u00adence be\u00adtween pro\u00adto\u00adtype and pro\u00adduc\u00adtion code of\u00adten ends up be\u00ading small, be\u00adcause you don\u2019t take as many short\u00adcuts at the start. The pro\u00adgram grows and evolves more smoothly be\u00adcause it\u2019s easy to change one part with\u00adout caus\u00ading rip\u00adple ef\u00adfects elsewhere.<\/p><\/li><li><p><strong>Li\u00adbraries &amp; doc\u00adu\u00admen\u00adta\u00adtion.<\/strong> [Racket] This might not look like a com\u00adpet\u00adi\u00adtive dif\u00adfer\u00aden\u00adtia\u00adtor\u2014doesn\u2019t every pro\u00adgram\u00adming lan\u00adguage have li\u00adbraries &amp; documentation?<\/p><p>Yes, but prob\u00ada\u00adbly not like this. As a con\u00adse\u00adquence of be\u00ading used in re\u00adsearch set\u00adtings for many years\u2014Racket\u2019s core de\u00advel\u00adop\u00adment team is mostly CS pro\u00adfes\u00adsors\u2014Racket\u2019s li\u00adbraries &amp; docs are more like a trans\u00admis\u00adsion from a highly evolved alien intelligence.<\/p><p>You get the es\u00adsen\u00adtials, of course:\u00a0<a href=\"http:\/\/docs.racket-lang.org\/web-server-internal\/index.html\">web server,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/json\/index.html\">JSON,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/xml\/index.html\">XML,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/draw\/index.html\">draw\u00ading,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/foreign\/index.html\">for\u00adeign-func\u00adtion in\u00adter\u00adface,<\/a> and so on. Then you no\u00adtice pack\u00adages you maybe didn\u2019t ex\u00adpect:\u00a0<a href=\"http:\/\/docs.racket-lang.org\/framework\/index.html\">GUI ap\u00adpli\u00adca\u00adtion frame\u00adwork,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/plot\/index.html\">math plot\u00adting,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/pkg\/index.html\">pack\u00adage-dis\u00adtri\u00adbu\u00adtion sys\u00adtem,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/rackunit\/index.html\">unit tester.<\/a> Be\u00adyond that, your face starts to melt a lit\u00adtle bit: <a href=\"http:\/\/docs.racket-lang.org\/redex\/index.html\">se\u00adman\u00adtics en\u00adgi\u00adneer\u00ading?<\/a> <a href=\"http:\/\/docs.racket-lang.org\/future-visualizer\/index.html\">Fu\u00adtures visualizer?<\/a><\/p><p>I won\u2019t pre\u00adtend to know what all this shit does. A lot of it is over my head. But I like that. Each week I use Racket, I end up ex\u00adplor\u00ading a new part of the li\u00adbrary, and learn\u00ading some\u00adthing new. As op\u00adposed to other lan\u00adguages that seem to kill brain cells on con\u00adtact (= pretty much any\u00adthing named *Script, I find).<\/p><aside><p>If you don\u2019t like the de\u00adsign of the docs, blame me\u2014that\u2019s my ma\u00adjor con\u00adtri\u00adbu\u00adtion to Racket thus\u00a0far.<\/p><\/aside><p>This learn\u00ading is only pos\u00adsi\u00adble be\u00adcause of Racket\u2019s truly out\u00adstand\u00ading doc\u00adu\u00admen\u00adta\u00adtion. It\u2019s vast, thor\u00adough, pre\u00adcise, and ap\u00adproach\u00adable. <a href=\"http:\/\/docs.racket-lang.org\/\">See for yourself.<\/a><\/p><\/li><li><p><strong>Dr\u00adRacket.<\/strong> [Racket] Yes, I know how to use a com\u00admand line. But Racket in\u00adcludes a cross-plat\u00adform graph\u00adi\u00adcal IDE called <a href=\"http:\/\/docs.racket-lang.org\/drracket\/index.html\">Dr\u00adRacket<\/a> that\u2019s pretty great. Dr\u00adRacket lets you edit, run, and de\u00adbug Racket source files (or any other lan\u00adguage based on Racket\u2014see item #9 on this list.)<\/p><p>No, it doesn\u2019t have the Ginsu-level search-and-re\u00adplace fa\u00adcil\u00adi\u00adties of some\u00adthing like Sub\u00adlime Text. But it does have help\u00adful edit\u00ading fea\u00adtures op\u00adti\u00admized for Racket code (for in\u00adstance, you can right-click on a sym\u00adbol name and re\u00adname it through\u00adout the file, or jump from a func\u00adtion to its documentation).<\/p><p>More\u00adover, the com\u00admand line within Dr\u00adRacket doesn\u2019t just show plain text\u2014it can show stacked frac\u00adtions, <a href=\"http:\/\/docs.racket-lang.org\/teachpack\/2htdpimage.html\">draw\u00adings,<\/a> <a href=\"http:\/\/docs.racket-lang.org\/plot\/renderer3d.html?q=plot\">math plots,<\/a> and other un\u00adex\u00adpected guests. If your com\u00admand line does all that, by all means keep us\u00ading\u00a0it.<\/p><\/li><li><p><strong>X-ex\u00adpres\u00adsions.<\/strong> [Lisps] This choice is some\u00adwhat bi\u00adased by my work with Racket, which mostly in\u00advolves doc\u00adu\u00adment pro\u00adcess\u00ading and type\u00adset\u00adting. But re\u00adlated top\u00adics arise in most web pro\u00adgram\u00adming. An X-ex\u00adpres\u00adsion is a spe\u00adcial na\u00adtive data struc\u00adture that Lisps use to rep\u00adre\u00adsent HTML and other XML-ish\u00a0data.<\/p><p>Well, not \u201cspe\u00adcial\u201d in a Lispy sense\u2014keep\u00ading with the usual pol\u00adicy, an X-ex\u00adpres\u00adsion is just an\u00adother list\u2014but spe\u00adcial in the sense that other pro\u00adgram\u00adming lan\u00adguages don\u2019t have it. Usu\u00adally your choice is to rep\u00adre\u00adsent HTML ei\u00adther as a string or as a full XML tree. A string is wrong be\u00adcause it doesn\u2019t cap\u00adture the struc\u00adture of the HTML, as de\u00adfined by its tags and at\u00adtrib\u00adutes. An XML tree shows this struc\u00adture, but con\u00adceals the se\u00adquen\u00adtial na\u00adture of the data el\u00ade\u00adments, and is un\u00adwieldy to work\u00a0with.<\/p><p>An X-ex\u00adpres\u00adsion ends up be\u00ading an ideal hy\u00adbrid be\u00adtween a string and a tree. More\u00adover, be\u00adcause it\u2019s just an\u00adother list-based ex\u00adpres\u00adsion in the lan\u00adguage, you have a lot of op\u00adtions for pro\u00adcess\u00ading it. Trans\u00adlat\u00ading an X-ex\u00adpres\u00adsion to or from a text rep\u00adre\u00adsen\u00adta\u00adtion us\u00ading an\u00adgle brack\u00adets is triv\u00adial and fast. (<a href=\"http:\/\/pkg-build.racket-lang.org\/doc\/pollen\/second-tutorial.html#%28part._.X-expressions%29\">De\u00adtails.<\/a>)<\/p><p>Given the close kin\u00adship be\u00adtween XML-ish data struc\u00adtures and Lisp lan\u00adguages, I have no ex\u00adpla\u00adna\u00adtion why, dur\u00ading the in\u00adter\u00adnet era, they\u2019ve not been paired more of\u00adten. They\u2019re like peanut but\u00adter and jelly.<\/p><\/li><li><p><strong>Scrib\u00adble.<\/strong> [Racket] Pollen wouldn\u2019t have been pos\u00adsi\u00adble with\u00adout <a href=\"http:\/\/docs.racket-lang.org\/scribble\/index.html\">Scrib\u00adble,<\/a> so for me, this has been the stone-cold killer fea\u00adture of Racket. But that won\u2019t be true for every\u00adone, so I\u2019m mov\u00ading it down on the\u00a0list.<\/p><aside><p>Scrib\u00adble was orig\u00adi\u00adnally cre\u00adated to serve as Racket\u2019s <a href=\"https:\/\/www.cs.utah.edu\/plt\/publications\/icfp09-fbf.pdf\">doc\u00adu\u00admen\u00adta\u00adtion lan\u00adguage<\/a> (a job it does well).<\/p><\/aside><p>Scrib\u00adble is a di\u00adalect of Racket that in\u00adverts the or\u00addi\u00adnary re\u00adla\u00adtion\u00adship of plain text and code: rather than em\u00adbed\u00adding text strings within source, a Scrib\u00adble doc\u00adu\u00adment con\u00adsists of code ex\u00adpres\u00adsions em\u00adbed\u00added within plain\u00a0text.<\/p><p>\u201cSo it\u2019s like an HTML tem\u00adplate lan\u00adguage.\u201d Yes, in the sense that a tem\u00adplate lan\u00adguage al\u00adlows code to be em\u00adbed\u00added in text. But also no, be\u00adcause a tem\u00adplate lan\u00adguage is usu\u00adally a pid\u00adgin ver\u00adsion of a real pro\u00adgram\u00adming lan\u00adguage. Scrib\u00adble, by con\u00adtrast, lets you in\u00advoke any Racket code sim\u00adply by adding a com\u00admand char\u00adac\u00adter to the front. In keep\u00ading with the theme al\u00adready es\u00adtab\u00adlished, this ap\u00adproach is both sim\u00adpler (be\u00adcause there\u2019s al\u00admost noth\u00ading new to learn) and more pow\u00ader\u00adful (be\u00adcause you can in\u00advoke any\u00adthing in Racket).<\/p><p>In its com\u00adbi\u00adna\u00adtion of text and code, Scrib\u00adble has more kin\u00adship with La\u00adTeX. While it doesn\u2019t have the type\u00adset\u00adting fa\u00adcil\u00adi\u00adties of La\u00adTeX, the pro\u00adgram\u00adming fa\u00adcil\u00adi\u00adties are much better.<\/p><\/li><li><p><strong>Macros.<\/strong> [Racket] Also known in Racket as <em>syn\u00adtax trans\u00adfor\u00adma\u00adtions<\/em>. Se\u00adri\u00adous Rack\u00ade\u00adteers some\u00adtimes pre\u00adfer that term be\u00adcause a Racket <a href=\"http:\/\/docs.racket-lang.org\/guide\/macros.html\">macro<\/a> can be far more so\u00adphis\u00adti\u00adcated than the usual Com\u00admon Lisp macro.<\/p><p>A macro in Com\u00admon Lisp is a func\u00adtion that runs at com\u00adpile-time, ac\u00adcept\u00ading sym\u00adbols as in\u00adput and in\u00adject\u00ading them into a tem\u00adplate to pro\u00adduce new code. Syn\u00adtax trans\u00adfor\u00adma\u00adtions in Racket, on the other hand, rely on the con\u00adcept of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hygienic_macro\">hy\u00adgiene<\/a> and can of\u00adfer Com\u00admon Lisp-style macros, but also more elab\u00ado\u00adrate syn\u00adtax rearrangements.<\/p><p>But for\u00adget that\u2014what\u2019s in it for you? As a pro\u00adgram\u00admer, you end up get\u00adting two bites at the ap\u00adple every time you run a file: Racket runs the syn\u00adtax trans\u00adfor\u00adma\u00adtions (which al\u00adter the source code), and then the source code itself.<\/p><p>Un\u00adlike some\u00adthing like the C pre\u00adproces\u00adsor, which is ba\u00adsi\u00adcally a sep\u00ada\u00adrate mini-lan\u00adguage, Racket syn\u00adtax trans\u00adfor\u00adma\u00adtions are them\u00adselves Racket func\u00adtions that give you ac\u00adcess to every\u00adthing in Racket. Like lists and ex\u00adpres\u00adsions, syn\u00adtax trans\u00adfor\u00adma\u00adtions add an\u00adother layer of ex\u00adpres\u00adsive possibilities.<\/p><\/li><li><p><strong>Cre\u00adate new pro\u00adgram\u00adming lan\u00adguages.<\/strong> [Racket] When I first read that Racket could be used to <a href=\"http:\/\/docs.racket-lang.org\/guide\/languages.html\">cre\u00adate new lan\u00adguages,<\/a> I had two thoughts\u2014<em>are they se\u00adri\u00adous?<\/em> and <em>would I really want to do that?<\/em> The an\u00adswers were <em>yes<\/em> and <em>oh\u00a0hell\u00a0yes.<\/em><\/p><p>Be\u00adtween ex\u00adpres\u00adsions, lists, and syn\u00adtax trans\u00adfor\u00adma\u00adtions, Racket gives you a huge amount of se\u00adman\u00adtic flex\u00adi\u00adbil\u00adity. But on top of that, it also adds syn\u00adtac\u00adtic flex\u00adi\u00adb\u00adlity, in that you can de\u00adfine a <em><a href=\"http:\/\/docs.racket-lang.org\/guide\/hash-reader.html\">reader<\/a><\/em> that con\u00adverts sur\u00adface syn\u00adtax into stan\u00addard Racket.<\/p><aside><p>Paul Gra\u00adham\u2019s pro\u00adgram\u00adming lan\u00adguage <a href=\"http:\/\/arclanguage.org\/\">Arc,<\/a> a di\u00adalect of Lisp, was built on top of Racket.<\/p><\/aside><p>You can use this fa\u00adcil\u00adity to make <a href=\"http:\/\/docs.racket-lang.org\/ts-guide\/index.html\">spe\u00adcial\u00adized di\u00adalects<\/a> of Racket. Or im\u00adple\u00adment <a href=\"http:\/\/docs.racket-lang.org\/r6rs\/index.html\">ear\u00adlier lan\u00adguages.<\/a> Or cre\u00adate en\u00adtirely <a href=\"http:\/\/docs.racket-lang.org\/datalog\/datalog.html\">new lan\u00adguages<\/a> with their own rules. You can use any of these lan\u00adguages within Dr\u00adRacket to code new projects. (These spe\u00adcial lan\u00adguages are some\u00adtimes called <em>do\u00admain-spe\u00adcific lan\u00adguages<\/em>, or <em>DSLs<\/em>.) Scrib\u00adble is a DSL based on Racket; Pollen is a set of DSLs based on Scribble.<\/p><p>If you\u2019re like most pro\u00adgram\u00admers, you\u2019ve never had a tool for mak\u00ading a new lan\u00adguage, so you\u2019ve not con\u00adsid\u00adered it a re\u00adal\u00adis\u00adtic ap\u00adproach to a prob\u00adlem. And maybe you\u2019ll never use it. But when you do, it is awe\u00adsome, in both the new and old senses of that\u00a0word.<\/p><\/li><li><p><strong>Op\u00adpor\u00adtu\u00adni\u00adties to par\u00adtic\u00adi\u00adpate.<\/strong> [Racket] In the\u00adory, open-source soft\u00adware projects cre\u00adate the op\u00adpor\u00adtu\u00adnity for groups of de\u00advel\u00adop\u00aders to join to\u00adgether and make bet\u00adter things in col\u00adlab\u00ado\u00adra\u00adtion than they could separately.<\/p><p>In prac\u00adtice, I\u2019ve found that open-source projects sort into a bi\u00admodal dis\u00adtri\u00adbu\u00adtion: over here, the un\u00adder\u00addoc\u00adu\u00admented solo projects that sput\u00adter along fit\u00adfully (if at all); over there, the ma\u00adture, pop\u00adu\u00adlar projects that can be in\u00adtim\u00adi\u00addat\u00ading for new contributors.<\/p><p>As an open-source project, <a href=\"https:\/\/github.com\/plt\/racket\">Racket<\/a> is po\u00adsi\u00adtioned at a happy medium. The core de\u00advel\u00adop\u00adment team has been work\u00ading to\u00adgether for years, and the com\u00admits re\u00admain <a href=\"https:\/\/github.com\/plt\/racket\/commits\/master\">fast &amp; fu\u00adri\u00adous.<\/a> But they\u2019re friendly sci\u00aden\u00adtists, not Shire-dwelling ego\u00adtists, and re\u00admain re\u00adcep\u00adtive to im\u00adprove\u00adments across the whole sys\u00adtem. If you have a bet\u00adter idea, they\u2019ll lis\u00adten; if you code it up to their stan\u00addards and make a pull re\u00adquest, they\u2019ll take\u00a0it.<\/p><\/li><\/ol><p>The point of this list has been to tell you about the pos\u00adi\u00adtives. That doesn\u2019t mean there aren\u2019t neg\u00ada\u00adtives. The small pool of Racket pro\u00adgram\u00admers means that when you hit a pot\u00adhole, it\u2019s pos\u00adsi\u00adble no one\u2019s ever seen your prob\u00adlem (= the in\u00adverse of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Linus's_Law\">Li\u00adnus\u2019s Law<\/a>). If I wanted to hire a Racket pro\u00adgram\u00admer, the op\u00adtions would be\u00a0few.<\/p><p>Still, why shouldn\u2019t I be en\u00adthu\u00adsi\u00adas\u00adtic? What I\u2019ve been able to ac\u00adcom\u00adplish so far with Racket has been tremen\u00addously use\u00adful, ed\u00adu\u00adca\u00adtional, and fun\u2014the most fun I\u2019ve had in 25+ years of programming.<\/p><p>If you think I sound like a fan\u00adboy or cult mem\u00adber, I can live with that. But those are peo\u00adple whose en\u00adthu\u00adsi\u00adasm is dis\u00adpro\u00adpor\u00adtion\u00adate to re\u00adal\u00adity. Here, I\u2019ve tried to stay out of the clouds (and the weeds) and ex\u00adplain the con\u00adcrete, prac\u00adti\u00adcal fea\u00adtures that have made Racket such a plea\u00adsure in my own\u00a0work.<\/p><p>As al\u00adways, your mileage may vary. But if I per\u00adsuade a few peo\u00adple to <a href=\"http:\/\/racket-lang.org\/download\/\">down\u00adload Racket<\/a> and try it, I\u2019ll be happy. In fact, if you try it and <em>don\u2019t<\/em> like it, I in\u00advite you to <a href=\"https:\/\/practicaltypography.com\/contact.html\" class=\"xref\">con\u00adtact<\/a> me, be\u00adcause I\u2019m al\u00adways cu\u00adri\u00adous to hear dis\u00adsent\u00ading opinions.<\/p><p>I will end by tak\u00ading on the big kahuna\u2014<\/p><p><a href=\"#does-profound-enlightenment-await\">Does \u201cprofound enlightenment\u201d await?<\/a><\/p><p>I won\u2019t claim I\u2019ve reached the top of the moun\u00adtain. But I can tell you what the view looks like so\u00a0far.<\/p><p>There\u2019s a sense in which Lisp and its de\u00adscen\u00addants are more than pro\u00adgram\u00adming lan\u00adguages. They\u2019re tools in the broader in\u00adtel\u00adlec\u00adtual in\u00adquiry into the the\u00adory of com\u00adpu\u00adta\u00adtion. Lisp\u2019s in\u00adven\u00adtor, John Mc\u00adCarthy, orig\u00adi\u00adnally <a href=\"http:\/\/www-formal.stanford.edu\/jmc\/history\/lisp.ps\">con\u00adsid\u00adered Lisp<\/a> a \u201cway of de\u00adscrib\u00ading com\u00adputable func\u00adtions much neater than the Tur\u00ading ma\u00adchines,\u201d adapt\u00ading the no\u00adta\u00adtion of lambda cal\u00adcu\u00adlus to do so. Racket, like\u00adwise, has grown out of sci\u00aden\u00adtific re\u00adsearch and exploration.<\/p><p>The the\u00adory of com\u00adpu\u00adta\u00adtion is just one of many great sci\u00aden\u00adtific dis\u00adcov\u00ader\u00adies in the last 100 years. But I don\u2019t get to use quan\u00adtum me\u00adchan\u00adics or rel\u00ada\u00adtiv\u00adity or DNA se\u00adquenc\u00ading in my daily work. When I\u2019m pro\u00adgram\u00adming, how\u00adever, I\u2019m us\u00ading computation.<\/p><p>Racket, as a Lisp di\u00adalect, has many prac\u00adti\u00adcal ben\u00ade\u00adfits. But it also opens a win\u00addow onto this vast the\u00ado\u00adret\u00adi\u00adcal world that un\u00adder\u00adlies every\u00adthing we can do with pro\u00adgrams. I\u2019m not a brainiac com\u00adputer sci\u00aden\u00adtist. But some days, through that win\u00addow, I can start to see a bit of what they see\u2014some math, some sci\u00adence, a lot of truth, and more than a lit\u00adtle beauty\u00a0and mystery.<\/p><p>Paul Gra\u00adham calls Lisp a \u201cse\u00adcret weapon.\u201d I would clar\u00adify: Lisp it\u00adself isn\u2019t the se\u00adcret weapon. Rather, <em>you<\/em> are\u2014be\u00adcause a Lisp lan\u00adguage of\u00adfers you the chance to dis\u00adcover your po\u00adten\u00adtial as a pro\u00adgram\u00admer and a thinker, and thereby raise your ex\u00adpec\u00adta\u00adtions for what you can accomplish.<\/p><p>If that\u2019s not a step to\u00adward en\u00adlight\u00aden\u00adment, I don\u2019t know what\u00a0is.<\/p><p><a href=\"#further-reading\">Further reading<\/a><\/p><p>The afore\u00admen\u00adtioned <em><a href=\"http:\/\/mitpress.mit.edu\/sicp\/\">Struc\u00adture and In\u00adter\u00adpre\u00adta\u00adtion of Com\u00adputer Pro\u00adgrams<\/a><\/em> de\u00adserves every com\u00adpli\u00adment it gets. It\u2019s mind-bend\u00adingly great, and ac\u00adces\u00adsi\u00adble to any\u00adone with a mild cu\u00adrios\u00adity about soft\u00adware. By us\u00ading Scheme to il\u00adlus\u00adtrate meta-top\u00adics about pro\u00adgrams, it im\u00adpliedly makes a great case for Lisps.<\/p><p>Like\u00adwise, Pe\u00adter Norvig\u2019s <em><a href=\"http:\/\/norvig.com\/paip.html\">Par\u00ada\u00addigms of Ar\u00adti\u00adfi\u00adcial In\u00adtel\u00adli\u00adgence Pro\u00adgram\u00adming<\/a><\/em> is not in\u00adtended as a gen\u00aderal in\u00adtro\u00adduc\u00adtion to Lisp, but it ends up do\u00ading that job well, by show\u00ading how so\u00adphis\u00adti\u00adcated pro\u00adgrams can arise from small con\u00adstruc\u00adtions of Lisp\u00a0code.<\/p><p>As I\u2019ve al\u00adready men\u00adtioned, Racket\u2019s doc\u00adu\u00admen\u00adta\u00adtion is con\u00adsis\u00adtently ex\u00adcel\u00adlent. Sim\u00adply read\u00ading through <em><a href=\"http:\/\/docs.racket-lang.org\/guide\/index.html\">The Racket Guide<\/a><\/em> and try\u00ading some ex\u00adam\u00adples is rewarding.<\/p><p>I ex\u00adplain more about why Racket was es\u00adsen\u00adtial for Pollen in the <a href=\"http:\/\/pkg-build.racket-lang.org\/doc\/pollen\/\">Pollen documentation.<\/a><\/p><p><a href=\"#getting-racket\">Getting Racket<\/a><\/p><p>If you want to try Racket, you can down\u00adload <a href=\"http:\/\/racket-lang.org\/download\/\">the cur\u00adrent sta\u00adble ver\u00adsion<\/a> (easy), <a href=\"http:\/\/www.cs.utah.edu\/plt\/snapshots\/\">a nightly build<\/a> (easy), or <a href=\"https:\/\/github.com\/plt\/racket\">com\u00adpile it from source<\/a> (slightly daunt\u00ading the first time but then easy).<\/p><p>\u2014Matthew But\u00adt\u00ader\u00adick<br \/>20 Aug\u00a02014<\/p><div class=\"btw\"><p>by the way<\/p><ul><li><p>At Racket\u00adCon in Sep\u00adtem\u00adber 2014, I gave a short talk about Pollen, ty\u00adpog\u00adra\u00adphy, and dig\u00adi\u00adtal book\u00admak\u00ading. Here\u2019s <a href=\"http:\/\/youtu.be\/IMz09jYOgoc\">the video.<\/a><\/p><\/li><\/ul><\/div>","created_at":"2019-05-17T20:07:42+0000","updated_at":"2019-05-17T20:07:55+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":31,"domain_name":"practicaltypography.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":952,"title":"Use a bought SSL certificate","url":"https:\/\/docs.iredmail.org\/use.a.bought.ssl.certificate.html","content":"Use a bought SSL certificate<div class=\"admonition note\"><p class=\"admonition-title\">This tutorial is available in other languages. <a href=\"https:\/\/bitbucket.org\/zhb\/iredmail-docs\/src\">Help translate more<\/a><\/p><p><a href=\"https:\/\/docs.iredmail.org\/use.a.bought.ssl.certificate-zh_CN.html\">\u7b80\u4f53\u4e2d\u6587<\/a> \/<\/p><\/div><div class=\"toc\"><ul><li><a href=\"#use-a-bought-ssl-certificate\">Use a bought SSL certificate<\/a><ul><li><a href=\"#get-a-ssl-certificate\">Get a SSL certificate<\/a><ul><li><a href=\"#request-a-free-cert-from-lets-encrypt\">Request a free cert from Let's Encrypt<\/a><\/li>\n<li><a href=\"#buy-from-a-trusted-ssl-vendor\">Buy from a trusted SSL vendor<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#use-the-bought-cert\">Use the bought cert<\/a><ul><li><a href=\"#replace-cert-files\">Replace cert files<\/a><\/li>\n<li><a href=\"#restart-network-services\">Restart network services<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#verify-the-cert\">Verify the cert<\/a><\/li>\n<li><a href=\"#see-also\">See Also<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/div>\n<p>iRedMail generates a self-signed SSL certificate during installation, it's\nfine if you just want to secure the network connections (POP3\/IMAP\/SMTP over\nTLS, HTTPS), but mail clients or web browsers will promot a annoying message\nto warn you this self-signed certificate is not trusted. To avoid this\nannoying message, you have to buy a SSL certificate from SSL certificate\nprovider. Search <code>buy ssl certificate<\/code> in Google will give you many SSL\nproviders, choose the one you prefer.<\/p>\n<h2 id=\"get-a-ssl-certificate\">Get a SSL certificate<\/h2>\n<h3 id=\"request-a-free-cert-from-lets-encrypt\">Request a free cert from Let's Encrypt<\/h3>\n<p>We have another tutorial to show you to request a free cert from Let's Encrypt:\n<a href=\"https:\/\/docs.iredmail.org\/letsencrypt.html\">Request a free cert from Let's Encrypt<\/a>.<\/p>\n<h3 id=\"buy-from-a-trusted-ssl-vendor\">Buy from a trusted SSL vendor<\/h3>\n<p>To buy ssl cert from a trusted vendor, you need to generate a new SSL\nkey and signing request file on your server with <code>openssl<\/code> command:<\/p>\n<div class=\"admonition warning\"><p class=\"admonition-title\">Warning<\/p><p>Do NOT use key length smaller than <code>2048<\/code> bit, it's insecure.<\/p><\/div>\n<pre># openssl req -new -newkey rsa:2048 -nodes -keyout privkey.pem -out server.csr\n<\/pre>\n<p>This command will generate two files:<\/p>\n<ul><li><code>privkey.pem<\/code>: the private key for the decryption of your SSL certificate.<\/li>\n<li><code>server.csr<\/code>: the certificate signing request (CSR) file used to apply\n  for your SSL certificate. <strong>This file is required by SSL certificate\n  provider.<\/strong><\/li>\n<\/ul><p>The openssl command will prompt for the following X.509 attributes of the\ncertificate:<\/p>\n<ul><li><code>Country Name (2 letter code)<\/code>: Use the two-letter code without punctuation\n  for country. for example: US, CA, CN.<\/li>\n<li><code>State or Province Name (full name)<\/code>: Spell out the state completely; do not\n  abbreviate the state or province name, for example: California.<\/li>\n<li><code>Locality Name (eg, city)<\/code>: City or town name, for example: Berkeley.<\/li>\n<li><code>Organization Name (eg, company)<\/code>: Your company name.<\/li>\n<li><code>Organizational Unit Name (eg, section)<\/code>: The name of the department or \n  organization unit making the request.<\/li>\n<li><code>Common Name (e.g. server FQDN or YOUR name)<\/code>: server FQDN or your name.<\/li>\n<li><code>Email Address []<\/code>: your full email address.<\/li>\n<li><code>A challenge password []<\/code>: type a password for this ssl certificate.<\/li>\n<li><code>An optional company name []<\/code>: an optional company name.<\/li>\n<\/ul><p><strong>NOTE<\/strong>: Some certificates can only be used on web servers using the <code>Common Name<\/code>\nspecified during enrollment. For example, a certificate for the domain\n<code>domain.com<\/code> will receive a warning if accessing a site named <code>www.domain.com<\/code>\nor <code>secure.domain.com<\/code>, because <code>www.domain.com<\/code> and <code>secure.domain.com<\/code> are\ndifferent from <code>domain.com<\/code>.<\/p>\n<p>Now you have two files: <code>privkey.pem<\/code> and <code>server.csr<\/code>. Go to the website of\nyour preferred SSL privider, it will ask you to upload <code>server.csr<\/code> file to\nissue an SSL certificate.<\/p>\n<p>Usually, SSL provider will give you 2 files:<\/p>\n<ul><li><code>cert.pem<\/code><\/li>\n<li><code>fullchain.pem<\/code> (some SSL providers use name <code>server.ca-bundle<\/code>)<\/li>\n<\/ul><p>We need above 2 files, and <code>privkey.pem<\/code>. Upload them to your server, you can\nstore them in any directory you like, recommended directories are:<\/p>\n<ul><li>on RHEL\/CentOS: <code>cert.pem<\/code> and <code>fullchain.pem<\/code> should be placed under\n  <code>\/etc\/pki\/tls\/certs\/<\/code>, <code>privkey.pem<\/code> should be <code>\/etc\/pki\/tls\/private\/<\/code>.<\/li>\n<li>on Debian\/Ubuntu, FreeBSD: <code>cert.pem<\/code> and <code>fullchain.pem<\/code> should be\n  placed under <code>\/etc\/ssl\/certs\/<\/code>, <code>privkey.pem<\/code> should be <code>\/etc\/ssl\/private\/<\/code>.<\/li>\n<li>on OpenBSD: <code>\/etc\/ssl\/<\/code>.<\/li>\n<\/ul><h2 id=\"use-the-bought-cert\">Use the bought cert<\/h2>\n<p>The easiest and quickest way to use the bought cert is replacing\nthe self-signed SSL cert generated by iRedMail installer, then\nrestart services which use the cert files.<\/p>\n<h3 id=\"replace-cert-files\">Replace cert files<\/h3>\n<div class=\"admonition warning\"><p class=\"admonition-title\">Warning<\/p><p>If you deployed iRedMail with the iRedMail Easy platform, ssl cert files\nare stored under <code>\/opt\/iredmail\/ssl\/<\/code>:<\/p><ul><li><code>key.pem<\/code>: private key<\/li>\n<li><code>cert.pem<\/code>: certificate<\/li>\n<li><code>combined.pem<\/code>: full chain<\/li>\n<\/ul><\/div>\n<ul><li>On RHEL\/CentOS:<\/li>\n<\/ul><pre>mv \/etc\/pki\/tls\/certs\/iRedMail.crt{,.bak}       # Backup. Rename iRedMail.crt to iRedMail.crt.bak\nmv \/etc\/pki\/tls\/private\/iRedMail.key{,.bak}     # Backup. Rename iRedMail.key to iRedMail.key.bak\ncp fullchain.pem \/etc\/pki\/tls\/certs\/iRedMail.crt\ncp privkey.pem \/etc\/pki\/tls\/private\/iRedMail.key\n<\/pre>\n<ul><li>On Debian\/Ubuntu, FreeBSD and OpenBSD:<\/li>\n<\/ul><pre>mv \/etc\/ssl\/certs\/iRedMail.crt{,.bak}       # Backup. Rename iRedMail.crt to iRedMail.crt.bak\nmv \/etc\/ssl\/private\/iRedMail.key{,.bak}     # Backup. Rename iRedMail.key to iRedMail.key.bak\ncp fullchain.pem \/etc\/ssl\/certs\/iRedMail.crt\ncp privkey.pem \/etc\/ssl\/private\/iRedMail.key\n<\/pre>\n<h3 id=\"restart-network-services\">Restart network services<\/h3>\n<p>Required services:<\/p>\n<ul><li>Postfix<\/li>\n<li>Dovecot<\/li>\n<li>Nginx or Apache<\/li>\n<\/ul><p>Depends on the backend you chose during iRedMail installation, you may need to\nrestart:<\/p>\n<ul><li>MySQL or MariaDB<\/li>\n<li>PostgreSQL<\/li>\n<li>OpenLDAP<\/li>\n<\/ul><h2 id=\"verify-the-cert\">Verify the cert<\/h2>\n<ul><li>To verify ssl cert used in Postfix (SMTP server) and Dovecot, please launch a\n  mail client application (MUA, e.g. Outlook, Thunderbird) and create an email\n  account, make sure you correctly configured the MUA to connect to mail\n  server. If SSL cert is not valid, MUA will warn you.<\/li>\n<li>For Apache \/ Nginx web server, you can access your website with favourite web\n  browser, the browser should show you the ssl cert status. Or, use other\n  website to help test it, for example:\n  <a href=\"https:\/\/www.ssllabs.com\/ssltest\/index.html\">https:\/\/www.ssllabs.com\/ssltest\/index.html<\/a> (input your web host name, then\n  submit and wait for a result).<\/li>\n<\/ul><h2 id=\"see-also\">See Also<\/h2>\n<ul><li><a href=\"https:\/\/docs.iredmail.org\/letsencrypt.html\">Request a free cert from Let's Encrypt<\/a><\/li>\n<\/ul>","created_at":"2019-05-17T03:17:27+0000","updated_at":"2019-05-17T03:17:27+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"docs.iredmail.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["gamedev","lisp","cl"],"is_public":false,"id":951,"title":"Creating a (Non-Trivial) Lisp Game in 2018 \u2013 (defun games ())","url":"https:\/\/defungames.com\/2018\/12\/creating-a-non-trivial-lisp-game-in-2018\/","content":"<div class=\"entry-content\"><div class=\"su-box su-box-style-noise feature-box\"><div class=\"su-box-title\">Spycursion is now on Patreon!<\/div><div class=\"su-box-content su-clearfix\"><h3>We\u2019d be honored to have your support on <a href=\"https:\/\/patreon.com\/spycursion\">Patreon<\/a>.<\/h3><p>Just want to follow progress? You can join our mailing list here:<\/p>&#13;\n\t<section id=\"yikes-mailchimp-container-1\" class=\"yikes-mailchimp-container yikes-mailchimp-container-1\"><form id=\"feature-box-1\" class=\"yikes-easy-mc-form yikes-easy-mc-form-1 yikes-mailchimp-form-inline\" method=\"post\" data-attr-form-id=\"1\" action=\"action\"><div>&#13;\n&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"yikes-easy-mc-form-1-EMAIL\" class=\"label-inline EMAIL-label yikes-mailchimp-field-required\">&#13;\n&#13;\n\t\t\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\t\t\t<input id=\"yikes-easy-mc-form-1-EMAIL\" name=\"email\" placeholder=\"i_love_spycursion@example.com\" class=\"yikes-easy-mc-email field-no-label\" required=\"required\" type=\"email\" value=\"\" \/><\/label>&#13;\n\t\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t&#13;\n\t\t\t\t<input type=\"hidden\" name=\"yikes-mailchimp-honeypot\" id=\"yikes-mailchimp-honeypot-1\" value=\"\" \/><input type=\"hidden\" name=\"yikes-mailchimp-associated-list-id\" id=\"yikes-mailchimp-associated-list-id-1\" value=\"e89f75b9d0\" \/><input type=\"hidden\" name=\"yikes-mailchimp-submitted-form\" id=\"yikes-mailchimp-submitted-form-1\" value=\"1\" \/><label class=\"empty-form-inline-label submit-button-inline-label\">\u00a0<button type=\"submit\" class=\"yikes-easy-mc-submit-button yikes-easy-mc-submit-button-1 btn btn-primary\"> Subscribe!<\/button><\/label>\t\t\t\t&#13;\n\t\t\t\t<input type=\"hidden\" id=\"yikes_easy_mc_new_subscriber_1\" name=\"yikes_easy_mc_new_subscriber\" value=\"9ef996eded\" \/><input type=\"hidden\" name=\"_wp_http_referer\" value=\"\/2018\/12\/creating-a-non-trivial-lisp-game-in-2018\/\" \/><\/div><\/form>&#13;\n\t\t\t&#13;\n&#13;\n\t\t\t<\/section><\/div><\/div><\/div><div class=\"entry-content\">\n\t\t<p>Lisp programmers are a small group. According to the <a href=\"https:\/\/www.tiobe.com\/tiobe-index\/\">TIOBE index<\/a>, Lisp currently sits at #32 on their ranking of programming language popularity. (<em>Common Lisp<\/em> specifically is somewhere between 51 and 100.) You can certainly imagine that Lisp <em>game<\/em> programmers are an even smaller group. And then when you consider <em>non-trivial<\/em> games, the games made by teams, the ones that are significant enough in scope to potentially fund the studio building them\u2026 well, I don\u2019t think it\u2019s much of a leap to say that you can probably count the number of those games on one hand.<\/p>\n<p>Suffice it to say that not a whole lot has been written on the art of developing games in Lisp, relative to other languages. As a leader of one of the aforementioned five-or-fewer teams building a sizeable Lisp game (Common Lisp, in our case), I want to contribute to that collective knowledge base. But first, a caveat: It would be premature to consider either (defun games ()), as a studio, or myself, personally, as any kind of \u201cauthoritative voice\u201d in the Lisp game community. Spycursion is the first game that we as a studio are producing, and it hasn\u2019t even been funded yet. (<strong>Hint hint<\/strong><strong>, Kickstarter, Feb. 26, hint hint!<\/strong>) And although I was a hobbyist Common Lisp developer for about seven years before starting work on Spycursion last year, there\u2019s still a lot about this language I don\u2019t know, and there are certainly people who could write better\/faster Lisp game code than me\u2026 I\u2019m just the only one who was insane enough to start this project. \ud83d\ude09 We also recently picked up another developer (he hasn\u2019t introduced himself yet, but he will) whose opinions may differ from mine. Taken altogether, what I\u2019m trying to say is: I\u2019m not a guru. I\u2019m just one guy trying to find his way on this bizarre-yet-beautiful landscape, just like everyone else!<\/p>\n<p>Aaaanyway, Lisp games. The first thing you should know, before you go tell all your friends \u201cHey, they wrote Spycursion in Lisp, let\u2019s create our Overwatch\/Pokemon\/Dark Souls cross-over in Lisp, too!\u201d is that Spycursion is <em>weird<\/em>. The game itself is weird, its architecture is weird, its developers are weird\u2026 Point is, we\u2019re working with some parameters that make Lisp not only suitable for Spycursion, but perhaps even the ideal language for it. Your Overwatch\/Pokemon\/Dark Souls cross-over may not fit the bill. (Or it may. I\u2019ll let you know after we release Game #2.) It\u2019s also important to remember that Spycursion has a client and a server component, both of which are written in CL. This was not always the case for the client; I\u2019ll share that story in a minute. But let\u2019s talk about the server first.<\/p>\n<p>As we\u2019ve mentioned before, Spycursion has <a href=\"https:\/\/defungames.com\/2018\/07\/speaking-slang\/\">its own programming language<\/a>. You may have read about how domain-specific languages (DSL\u2019s) in Lisp are grand, how they\u2019ll make your coffee for you, blah blah blah. In my experience, a lot of the conversation about this can get pretty academic, but here\u2019s a summary of what happens when you compile and run a Slang program in the world of Spycursion:<\/p>\n<p>1. Compilation: The server parses your program (assuming it is syntactically correct) and creates a tree of Lisp forms from it.<br \/>\n2. Runtime: Some contextualization happens with that tree of Lisp forms, and then it is evaluated within the context of an in-game device.<br \/>\n3. (Optional): If the program needs to communicate with another program on another in-game device, via in-game network, then some magic happens to determine where that program is, and runs it, too.<\/p>\n<p>This is of course a lot more complicated than I\u2019m making it sound, but it\u2019s still far less complicated than in probably most, if not <em>all<\/em>, other mainstream languages. Lisp allows you, as a player, to run your own code within someone else\u2019s online game. (And yes, we <em>do<\/em> run security checks on your code before we parse it. We\u2019re not n00bs, thankyouverymuch.) <em>This is not merely simulated.<\/em> There are plenty of places within Spycursion where we \u201ccheat\u201d to make things easier on ourselves \u2014 hey, we\u2019re an indie, cut us some slack \u2014 but Slang is not one of those places. Just thinking about trying to do all of this in C++ gives me an icky feeling. (I eagerly await the e-mail from someone telling me how you <em>can<\/em> do it in C++, if you just jump through these 27 hoops\u2026)<\/p>\n<p>So, yes, Lisp is awesome as a game <em>server<\/em>, but I have a feeling that\u2019s not what most of you are here for. You want to know why we\u2019re using Lisp for the Spycursion client, and how you can write your own amazing thing in Lisp too, and clearly the answers are <span style=\"text-decoration: line-through;\">because I\u2019m insane and you should stop listening to me<\/span> because it\u2019s <em>fun<\/em>, dadgum it, and by being a <span style=\"text-decoration: line-through;\">masochist<\/span> early adopter and maybe having a fairly simple game, graphically. Sarcasm aside, these questions deserve longer answers:<\/p>\n<p><strong>Why Lisp (for the client)?<\/strong> Okay, story time. Long, long ago, when the Spycursion server was just Slang and a WebSocket interface and not much else, I set about creating a text-only testing client. I did this in Python, figuring it would be relatively quick and easy, which it was. At this point I\u2019d not completely decided on how the client would be implemented, but I reasoned that, despite Python not being my best language, it would be easy to find developers for \u2014 so I could handle doing all the server-side Lisp as long as I had someone else dedicated to the client.<\/p>\n<p>But there was a problem: Writing in Python just wasn\u2019t <em>fun<\/em>. Some of you may not understand this, but it\u2019s a phenomenon I\u2019ve heard about before \u2014 once you get \u201cspoiled\u201d by Lisp, you have trouble getting motivated to code in any other \u201clesser\u201d language. This was before we existed as a studio, when I was doing this project entirely by myself, so I didn\u2019t have anyone else to shove the Python work off to, but I knew it was important to get <em>something<\/em> visual produced so that I could begin to show people (not simply tell them) what Spycursion was about.<\/p>\n<p>So I chucked the Python client, rewrote the thing in CL, and around that time, started getting ideas of what I felt this fledgling studio could be. In my opinion, the reasons \u201cnobody programs in Lisp\u201d mostly boil down to, well\u2026 because \u201cnobody programs in Lisp.\u201d The community is small, so there are few tools available, so the community doesn\u2019t grow because people don\u2019t use Lisp for major projects due to a lack of tools. This is even more true for Lisp games specifically, but I believe that this chicken-and-egg problem can be helped by creating a well-known game in Lisp, if only to point to it as a success and prove that it can be done. Additionally, by open-sourcing the client code (which we\u2019ve not done yet, mostly due to legal questions that need answering), we show new and aspiring Lispers <em>how<\/em> it can be done. Again, to avoid getting ahead of ourselves \u2014 this thing <em>could still fail<\/em>, perhaps in spectacular fashion. But at least if it does, you\u2019ll get a front-row seat. \ud83d\ude09 And I guarantee you that it won\u2019t be because of the programming language we chose.<\/p>\n<p><strong>Why do you think it will work?<\/strong> As I mentioned before, Spycursion is weird. There are very experienced people within the Lisp community who think that Lisp, in its current state, is not ready for serious game programming. I have no basis on which to disagree with them, but I do think it depends on the type of game you\u2019re creating. Spycursion has several things going for it that make it different from what you might imagine as a \u201ctypical game,\u201d especially a AAA game:<\/p>\n<ul><li><strong>There\u2019s no twitch action.<\/strong> An ill-timed garbage collection isn\u2019t going to be the difference between life and death. As of yet, we\u2019ve made no attempt to customize GC or really push for low-level performance in ways that might be challenging to do in Lisp. We might, at some point, but I\u2019m guessing we won\u2019t even need to.<\/li>\n<li><strong>Much of the tricky stuff is offloaded to the server.<\/strong> Don\u2019t underestimate the advantage of having a server (read: one single platform, of your choice) running parts of your game. That means fewer shared library headaches and fewer client performance woes \u2014 assuming you do everything asynchronously \u2014 for starters. (Oh, but please don\u2019t be that developer who forces a network connection for your single-player game. That\u2019s just rude.)<\/li>\n<li><strong>We don\u2019t need to build for consoles.<\/strong> Ever tried to run Lisp on the Switch? I haven\u2019t\u2026 and I don\u2019t want to.<\/li>\n<li><strong>Physics? We don\u2019t need no stinkin\u2019 physics.<\/strong> Or at least no physics that can\u2019t be calculated based solely on what I remember from 9th grade Geometry.<\/li>\n<li><strong>We keep graphics simple.<\/strong> I don\u2019t really know how to quantify that, but the game is meant to be played 3rd-person, from a distanced zoom level. (We let you zoom in, but I wouldn\u2019t necessarily recommend it.) The assets will be relatively low-poly and there will be relatively few of them. Note that \u201csimple\u201d does <em>not<\/em> mean ugly. We\u2019re aiming to make it all look good while still putting most of our effort into gameplay \u2014 graphics are one of the places we \u201ccheat.\u201d You\u2019ll be able to judge how we\u2019re doing once we finish the Kickstarter trailer.<\/li>\n<\/ul><p>Put all of the above together, and you get a picture of a fairly minimalistic game client that really gets out of the way and lets you enjoy playing. (Or at least that\u2019s the picture I hope you get!) And to be clear, I am <em>not<\/em> saying that fancier, AAA-level games can\u2019t be made with Lisp. I\u2019m saying I really don\u2019t know, but this is our attempt to get as close as we can with the resources we have.<\/p>\n<p><strong>How are you doing it?<\/strong> I would say \u201cpain, sweat, and toil,\u201d but that kind of describes <em>all<\/em> game development, so\u2026<\/p>\n<ul><li><strong>The engine.<\/strong> Oh, just kidding, we aren\u2019t using one. In 2018, I know of at least three Common Lisp 3D game engines in varying stages of development: <a href=\"https:\/\/github.com\/HackerTheory\/first-light\">First Light<\/a>, <a href=\"https:\/\/github.com\/Shirakumo\/trial\">Trial<\/a>, and <a href=\"https:\/\/github.com\/borodust\/cl-bodge\">cl-bodge<\/a>. In 2017, when I was researching these engines, they were all either too early in development or otherwise just didn\u2019t fit our needs. They might now, but I haven\u2019t followed their progress enough to know for sure. This brings me to what is maybe the most important thing to know if you\u2019re looking to write a non-trivial Lisp game in 2018: <strong>You have to be okay with writing a lot of things yourself.<\/strong> Lisp makes this fun to do, and might still even save you time overall depending on just how much you need to write. But it\u2019s a very different mindset from being, say, a plug-and-play JavaScript developer.<\/li>\n<li><a href=\"https:\/\/github.com\/cbaggers\/cepl\"><strong>CEPL<\/strong><\/a><strong>.<\/strong> CEPL is an abstraction layer around OpenGL which has been a joy to work with, once I figured out how. (Tip: See the author\u2019s <a href=\"https:\/\/www.youtube.com\/user\/CBaggers\/videos\">video stream<\/a>, it\u2019s great.) Its goal seems to be to make OpenGL lispy and fun to work with, and I really think it has achieved that. Spycursion uses CEPL extensively, and while I\u2019m sure we\u2019re doing some things suboptimally, performance has been fine so far. We\u2019re also using cepl.sdl2 which brings in cl-sdl2.<\/li>\n<li><strong><a href=\"https:\/\/github.com\/Shinmera\/qtools\/\">QTools<\/a>.<\/strong> QTools is a lispy wrapper around Qt 4. Most games probably won\u2019t want the heft of Qt, but our <a href=\"https:\/\/defungames.com\/wp-content\/uploads\/2018\/07\/ui_screenshot_20180617.png\">device UI<\/a> actually needs something a bit more featureful, so it\u2019s worth mentioning. We have run into some build problems with QTools, particularly on Linux. Add in that there doesn\u2019t seem to be a Qt5 update on the horizon, and we may well end up switching away from it (to what, I can\u2019t say yet). But it\u2019s currently working well, for what we need it for.<\/li>\n<li><strong>Other libraries<\/strong>\n<ul><li><strong>Assets:<\/strong> <a href=\"https:\/\/github.com\/3b\/classimp\">classimp<\/a>, <a href=\"https:\/\/github.com\/cbaggers\/dirt\/\">dirt<\/a><\/li>\n<li><strong>Fonts:<\/strong> <a href=\"https:\/\/shirakumo.github.io\/cl-fond\/\">cl-fond<\/a><\/li>\n<li><strong>Input:<\/strong> <a href=\"https:\/\/github.com\/cbaggers\/skitter\">skitter<\/a><\/li>\n<li><strong>Network:<\/strong> <a href=\"https:\/\/github.com\/fukamachi\/websocket-driver\/\">websocket-driver-client<\/a><\/li>\n<li>\u2026 and some other less-interesting stuff.<\/li>\n<\/ul><\/li>\n<\/ul><p>And that\u2019s it! Lisp game development isn\u2019t for everyone, but I stand by my assertion that it can be done, and done <em>well<\/em>, given the right game \u2014 and perhaps the right amount of patience.<\/p>\n<p><strong>TL;DR:<\/strong><\/p>\n<ul><li>Use CEPL. Or one of the engines, if they fit. Or maybe both.<\/li>\n<li>Yes, you should do it, even if you fail. That\u2019s how the community grows.<\/li>\n<li>Be prepared to write a lot of stuff yourself. This is not necessarily a bad thing. It\u2019s fun!\n<ul><li>Except skeletal animation. Don\u2019t write that yourself.<\/li>\n<\/ul><\/li>\n<li>Support our Kickstarter starting on February 26th!<\/li>\n<\/ul><\/div>","created_at":"2019-05-16T20:28:05+0000","updated_at":"2019-05-16T20:28:20+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":11,"domain_name":"defungames.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","android"],"is_public":false,"id":950,"title":"running-emacs-on-android.html","url":"http:\/\/endlessparentheses.com\/running-emacs-on-android.html","created_at":"2019-05-15T01:30:34+0000","updated_at":"2019-05-15T01:30:39+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"endlessparentheses.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":949,"title":"forum-explorer","url":"https:\/\/www.mcnutt.in\/forum-explorer\/","created_at":"2019-05-14T23:26:04+0000","updated_at":"2019-05-14T23:26:04+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.mcnutt.in"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":948,"title":"going-critical","url":"https:\/\/www.meltingasphalt.com\/interactive\/going-critical\/","created_at":"2019-05-14T15:21:51+0000","updated_at":"2019-05-14T15:21:51+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.meltingasphalt.com"},{"is_archived":0,"is_starred":0,"tags":["procgen"],"is_public":false,"id":947,"title":"Show HN: Eigengrau's Generator \u2013 A procedural city generator for tabletop gaming","url":"https:\/\/news.ycombinator.com\/item?id=19904605","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D19904605\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"19904605\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/github.com\/ryceg\/Eigengrau-s-Essential-Establishment-Generator\" class=\"storylink\">Show HN: Eigengrau's Generator \u2013 A procedural city generator for tabletop gaming<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=github.com\">github.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        136 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=wishingonawendy\" class=\"hnuser\">wishingonawendy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=19904605\">16 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=19904605&amp;goto=item%3Fid%3D19904605\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Show%20HN%3A%20Eigengrau's%20Generator%20%E2%80%93%20A%20procedural%20city%20generator%20for%20tabletop%20gaming&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Show%20HN%3A%20Eigengrau's%20Generator%20%E2%80%93%20A%20procedural%20city%20generator%20for%20tabletop%20gaming\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=19904605&amp;auth=dcb1ffa7243c4ef83f04e60a5f95419e21a8e426\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=19904605\">15\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-05-14T14:47:23+0000","updated_at":"2019-05-14T14:47:26+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":946,"title":"Learn to change history with git rebase!","url":"https:\/\/git-rebase.io\/","content":"<div class=\"content\"><h2>Setting up a sandbox<\/h2><p>\n          We don't want to mess up any of your actual repositories, so\n          throughout this guide we'll be working with a sandbox repo. Run these\n          commands to get started:\n        <\/p><pre>git init \/tmp\/rebase-sandbox\ncd \/tmp\/rebase-sandbox\ngit commit --allow-empty -m\"Initial commit\" <a id=\"ref-1\" href=\"#note-1\"><\/a><\/pre><sup># 1<\/sup><p>\n          If you run into trouble, just run <code>rm -rf\n            \/tmp\/rebase-sandbox<\/code> and run these steps again to start over.\n          Each step of this guide can be run on a fresh sandbox, so it's not\n          necessary to re-do every task.\n        <\/p><\/div><div class=\"content\"><h2 id=\"amending\">Amending your last commit<\/h2><p>\n          Let's start with something simple: fixing your most recent commit.\n          Let's add a file to our sandbox - and make a mistake:\n        <\/p><pre>echo \"Hello wrold!\" &gt;greeting.txt\ngit add greeting.txt\ngit commit -m\"Add greeting.txt\"<\/pre><p>\n          Fixing this mistake is pretty easy. We can just edit the file and\n          commit with <code>--amend<\/code>, like so:\n        <\/p><pre>echo \"Hello world!\" &gt;greeting.txt\ngit commit -a --amend<\/pre><p>\n          Specifying <code>-a<\/code> automatically stages (i.e. <code>git\n            add<\/code>'s) all files that git already knows about, and\n          <code>--amend<\/code> will <em>squash<\/em> the changes into the most\n          recent commit. Save and quit your editor (you have a chance to change\n          the commit message now if you'd like). You can see the fixed commit\n          by running <code>git show<\/code>:\n        <\/p><pre>commit f5f19fbf6d35b2db37dcac3a55289ff9602e4d00 (HEAD -&gt; master)\nAuthor: Drew DeVault \nDate:   Sun Apr 28 11:09:47 2019 -0400\n    Add greeting.txt\n<strong>diff --git a\/greeting.txt b\/greeting.txt\nnew file mode 100644\nindex 0000000..cd08755\n--- \/dev\/null\n+++ b\/greeting.txt<\/strong>\n@@ -0,0 +1 @@\n+Hello world!<\/pre><h2 id=\"fixup\">Fixing up older commits<\/h2><p>\n          Amending only works for the most recent commit. What happens if you\n          need to correct an older commit? Let's start by setting up our sandbox\n          accordingly:\n        <\/p><pre>echo \"Hello!\" &gt;greeting.txt\ngit add greeting.txt\ngit commit -m\"Add greeting.txt\"\necho \"Goodbye world!\" &gt;farewell.txt\ngit add farewell.txt\ngit commit -m\"Add farewell.txt\"<\/pre><p>\n          Looks like <code>greeting.txt<\/code> is missing \"world\". Let's write a\n          commit normally which fixes that:\n        <\/p><pre>echo \"Hello world!\" &gt;greeting.txt\ngit commit -a -m\"fixup greeting.txt\"<\/pre><p>\n          So now the files look correct, but our history could be better - let's\n          use the new commit to \"fixup\" the last one. For this, we need to\n          introduce a new tool: the interactive rebase. We're going to edit the\n          last three commits this way, so we'll run <code>git rebase -i\n            HEAD~3<\/code> (<code>-i<\/code> for interactive). This'll open your\n          text editor with something like this:\n        <\/p><pre>pick 8d3fc77 Add greeting.txt\npick 2a73a77 Add farewell.txt\npick 0b9d0bb fixup greeting.txt\n# Rebase f5f19fb..0b9d0bb onto f5f19fb (3 commands)\n#\n# Commands:\n# p, pick &lt;commit&gt; = use commit\n# f, fixup &lt;commit&gt; = like \"squash\", but discard this commit's log message<\/pre><p>\n          This is the rebase plan, and by editing this file you can instruct\n          git on how to edit history. I've trimmed the summary to just the\n          details relevant to this part of the rebase guide, but feel free to\n          skim the full summary in your text editor.\n        <\/p><p>\n          When we save and close our editor, git is going to remove all of these\n          commits from its history, then execute each line one at a time. By\n          default, it's going to <em>pick<\/em> each commit, summoning it from\n          the heap and adding it to the branch. If we don't edit this file at\n          all, we'll end up right back where we started, picking every commit\n          as-is. We're going to use one of my favorite features now: fixup. Edit\n          the third line to change the operation from \"pick\" to \"fixup\" and move\n          it to immediately after the commit we want to \"fix up\":\n        <\/p><pre>pick 8d3fc77 Add greeting.txt\nfixup 0b9d0bb fixup greeting.txt\npick 2a73a77 Add farewell.txt<\/pre><p><strong>Tip<\/strong>: We can also abbreviate this with just \"f\" to\n          speed things up next time.<\/p><p>\n          Save and quit your editor - git will run these commands. We can check\n          the log to verify the result:\n        <\/p><pre>$ git log -2 --oneline\nfcff6ae (HEAD -&gt; master) Add farewell.txt\na479e94 Add greeting.txt<\/pre><h2 id=\"squash\">Squashing several commits into one<\/h2><p>\n          As you work, you may find it useful to write lots of commits as you\n          reach small milestones or fix bugs in previous commits. However, it\n          may be useful to \"squash\" these commits together, to make a cleaner\n          history before merging your work into master. For this, we'll use the\n          \"squash\" operation. Let's start by writing a bunch of commits - just\n          copy and paste this if you want to speed it up:\n        <\/p><pre>git checkout -b squash\nfor c in H e l l o , ' ' w o r l d; do\n    echo \"$c\" &gt;&gt;squash.txt\n    git add squash.txt\n    git commit -m\"Add '$c' to squash.txt\"\ndone<\/pre><p>\n          That's a lot of commits to make a file that says \"Hello, world\"! Let's\n          start another interactive rebase to squash them together. Note that we\n          checked out a branch to try this on, first. Because of that, we can\n          quickly rebase all of the commits since we branched by using <code>git\n            rebase -i master<\/code>. The result:\n        <\/p><pre>pick 1e85199 Add 'H' to squash.txt\npick fff6631 Add 'e' to squash.txt\npick b354c74 Add 'l' to squash.txt\npick 04aaf74 Add 'l' to squash.txt\npick 9b0f720 Add 'o' to squash.txt\npick 66b114d Add ',' to squash.txt\npick dc158cd Add ' ' to squash.txt\npick dfcf9d6 Add 'w' to squash.txt\npick 7a85f34 Add 'o' to squash.txt\npick c275c27 Add 'r' to squash.txt\npick a513fd1 Add 'l' to squash.txt\npick 6b608ae Add 'd' to squash.txt\n# Rebase 1af1b46..6b608ae onto 1af1b46 (12 commands)\n#\n# Commands:\n# p, pick &lt;commit&gt; = use commit\n# s, squash &lt;commit&gt; = use commit, but meld into previous commit<\/pre><div class=\"tip\"><strong>Tip<\/strong>: your local master branch evolves independently\n          of the remote master branch, and git stores the remote branch as<code>origin\/master<\/code>. Combined with this trick,<code>git\n            rebase -i origin\/master<\/code>is often a very convenient way to\n          rebase all of the commits which haven't been merged upstream yet!<\/div><p>\n          We're going to squash all of these changes into the first commit. To\n          do this, change every \"pick\" operation to \"squash\", except for the\n          first line, like so:\n        <\/p><pre>pick 1e85199 Add 'H' to squash.txt\nsquash fff6631 Add 'e' to squash.txt\nsquash b354c74 Add 'l' to squash.txt\nsquash 04aaf74 Add 'l' to squash.txt\nsquash 9b0f720 Add 'o' to squash.txt\nsquash 66b114d Add ',' to squash.txt\nsquash dc158cd Add ' ' to squash.txt\nsquash dfcf9d6 Add 'w' to squash.txt\nsquash 7a85f34 Add 'o' to squash.txt\nsquash c275c27 Add 'r' to squash.txt\nsquash a513fd1 Add 'l' to squash.txt\nsquash 6b608ae Add 'd' to squash.txt<\/pre><p>\n          When you save and close your editor, git will think about this for a\n          moment, then open your editor again to revise the final commit\n          message. You'll see something like this:\n        <\/p><pre># This is a combination of 12 commits.\n# This is the 1st commit message:\nAdd 'H' to squash.txt\n# This is the commit message #2:\nAdd 'e' to squash.txt\n# This is the commit message #3:\nAdd 'l' to squash.txt\n# This is the commit message #4:\nAdd 'l' to squash.txt\n# This is the commit message #5:\nAdd 'o' to squash.txt\n# This is the commit message #6:\nAdd ',' to squash.txt\n# This is the commit message #7:\nAdd ' ' to squash.txt\n# This is the commit message #8:\nAdd 'w' to squash.txt\n# This is the commit message #9:\nAdd 'o' to squash.txt\n# This is the commit message #10:\nAdd 'r' to squash.txt\n# This is the commit message #11:\nAdd 'l' to squash.txt\n# This is the commit message #12:\nAdd 'd' to squash.txt\n# Please enter the commit message for your changes. Lines starting\n# with '#' will be ignored, and an empty message aborts the commit.\n#\n# Date:      Sun Apr 28 14:21:56 2019 -0400\n#\n# interactive rebase in progress; onto 1af1b46\n# Last commands done (12 commands done):\n#    squash a513fd1 Add 'l' to squash.txt\n#    squash 6b608ae Add 'd' to squash.txt\n# No commands remaining.\n# You are currently rebasing branch 'squash' on '1af1b46'.\n#\n# Changes to be committed:\n#\tnew file:   squash.txt\n#<\/pre><p>\n          This defaults to a combination of all of the commit messages which\n          were squashed, but leaving it like this is almost always not what you\n          want. The old commit messages may be useful for reference when writing\n          the new one, though.\n        <\/p><p><strong>Tip<\/strong>: the \"fixup\" command you learned about in the\n          previous section can be used for this purpose, too - but it discards\n          the messages of the squashed commits.<\/p><p>\n          Let's delete everything and replace it with a better commit message,\n          like this:\n        <\/p><pre>Add squash.txt with contents \"Hello, world\"\n# Please enter the commit message for your changes. Lines starting\n# with '#' will be ignored, and an empty message aborts the commit.\n#\n# Date:      Sun Apr 28 14:21:56 2019 -0400\n#\n# interactive rebase in progress; onto 1af1b46\n# Last commands done (12 commands done):\n#    squash a513fd1 Add 'l' to squash.txt\n#    squash 6b608ae Add 'd' to squash.txt\n# No commands remaining.\n# You are currently rebasing branch 'squash' on '1af1b46'.\n#\n# Changes to be committed:\n#\tnew file:   squash.txt\n#<\/pre><p>\n          Save and quit your editor, then examine your git log - success!\n        <\/p><pre>commit c785f476c7dff76f21ce2cad7c51cf2af00a44b6 (HEAD -&gt; squash)\nAuthor: Drew DeVault \nDate:   Sun Apr 28 14:21:56 2019 -0400\n    Add squash.txt with contents \"Hello, world\"\n<\/pre><p>\n          Before we move on, let's pull our changes into the master branch and\n          get rid of this scratch one. We can use <code>git rebase<\/code> like\n          we use <code>git merge<\/code>, but it avoids making a merge commit:\n        <\/p><pre>git checkout master\ngit rebase squash\ngit branch -D squash<\/pre><p>\n          We generally prefer to avoid using git merge unless we're actually\n          merging unrelated histories. If you have two divergent branches, a\n          git merge is useful to have a record of when they were... merged. In\n          the course of your normal work, rebase is often more appropriate.\n        <\/p><h2 id=\"edit\">Splitting one commit into several<\/h2><p>\n          Sometimes the opposite problem happens - one commit is just too big.\n          Let's look into splitting it up. This time, let's write some actual\n          code. Start with a simple C program<sup><a href=\"#note-2\" id=\"ref-2\">2<\/a><\/sup> (you can still copy+paste this snippet into\n          your shell to do this quickly):\n        <\/p><pre>cat &lt;&lt;EOF &gt;main.c\nint <strong>main<\/strong>(int argc, char *argv[]) {\n    return 0;\n}\nEOF<\/pre><p>We'll commit this first.<\/p><pre>git add main.c\ngit commit -m\"Add C program skeleton\"<\/pre><p>\n          Next, let's extend the program a bit:\n        <\/p><pre>cat &lt;&lt;EOF &gt;main.c\n#include &amp;ltstdio.h&gt;\nconst char *<strong>get_name<\/strong>() {\n    static char buf[128];\n    scanf(\"%s\", buf);\n    return buf;\n}\nint <strong>main<\/strong>(int argc, char *argv[]) {\n    printf(\"What's your name? \");\n    const char *name = get_name();\n    printf(\"Hello, %s!\\n\", name);\n    return 0;\n}\nEOF<\/pre><p>\n          After we commit this, we'll be ready to learn how to split it up.\n        <\/p><pre>git commit -a -m\"Flesh out C program\"<\/pre><p>\n          The first step is to start an interactive rebase. Let's rebase both\n          commits with <code>git rebase -i HEAD~2<\/code>, giving us this rebase\n          plan:\n        <\/p><pre>pick 237b246 Add C program skeleton\npick b3f188b Flesh out C program\n# Rebase c785f47..b3f188b onto c785f47 (2 commands)\n#\n# Commands:\n# p, pick &lt;commit&gt; = use commit\n# e, edit &lt;commit&gt; = use commit, but stop for amending<\/pre><p>\n          Change the second commit's command from \"pick\" to \"edit\", then save\n          and close your editor. Git will think about this for a second, then\n          present you with this:\n        <\/p><pre>Stopped at b3f188b...  Flesh out C program\nYou can amend the commit now, with\n  git commit --amend\nOnce you are satisfied with your changes, run\n  git rebase --continue<\/pre><p>\n          We could follow these instructions to <em>add<\/em> new changes to the\n          commit, but instead let's do a \"soft reset\"<sup><a href=\"#note-3\" id=\"ref-3\">3<\/a><\/sup> by running <code>git\n            reset HEAD^<\/code>. If you run <code>git status<\/code> after this,\n          you'll see that it un-commits the latest commit and adds its changes\n          to the working tree:\n        <\/p><pre>\nLast commands done (2 commands done):\n   pick 237b246 Add C program skeleton\n   edit b3f188b Flesh out C program\nNo commands remaining.\nYou are currently splitting a commit while rebasing branch 'master' on 'c785f47'.\n  (Once your working directory is clean, run \"git rebase --continue\")\nChanges not staged for commit:\n  (use \"git add ...\" to update what will be committed)\n  (use \"git checkout -- ...\" to discard changes in working directory)\n  modified:   main.c\nno changes added to commit (use \"git add\" and\/or \"git commit -a\")<\/pre><p>\n          To split this up, we're going to do an <em>interactive commit<\/em>.\n          This allows us to selectively commit only specific changes from the\n          working tree. Run <code>git commit -p<\/code> to start this process,\n          and you'll be presented with the following prompt:\n        <\/p><pre><strong>diff --git a\/main.c b\/main.c\nindex b1d9c2c..3463610 100644\n--- a\/main.c\n+++ b\/main.c<\/strong>\n@@ -1,3 +1,14 @@\n+#include &amp;ltstdio.h&gt;\n+\n+const char *get_name() {\n+    static char buf[128];\n+    scanf(\"%s\", buf);\n+    return buf;\n+}\n+\n int main(int argc, char *argv[]) {\n+    printf(\"What's your name? \");\n+    const char *name = get_name();\n+    printf(\"Hello, %s!\\n\", name);\n     return 0;\n }\n<strong>Stage this hunk [y,n,q,a,d,s,e,?]? <\/strong><\/pre><p>\n          Git has presented you with just one \"hunk\" (i.e. a single change) to\n          consider committing. This one is too big, though - let's use the \"s\"\n          command to \"split\" up the hunk into smaller parts.\n        <\/p><pre><strong>Split into 2 hunks.<\/strong>\n@@ -1 +1,9 @@\n+#include &lt;stdio.h&gt;\n+\n+const char *get_name() {\n+    static char buf[128];\n+    scanf(\"%s\", buf);\n+    return buf;\n+}\n+\n int main(int argc, char *argv[]) {\n<strong>Stage this hunk [y,n,q,a,d,j,J,g,\/,e,?]? <\/strong><\/pre><p><strong>Tip<\/strong>: If you're curious about the other options, press\n          \"?\" to summarize them.<\/p><p>\n          This hunk looks better - a single, self-contained change. Let's hit\n          \"y\" to answer the question (and stage that \"hunk\"), then \"q\" to \"quit\"\n          the interactive session and proceed with the commit. Your editor will\n          pop up to ask you to enter a suitable commit message.\n        <\/p><pre>\nAdd get_name function to C program\n# Please enter the commit message for your changes. Lines starting\n# with '#' will be ignored, and an empty message aborts the commit.\n#\n# interactive rebase in progress; onto c785f47\n# Last commands done (2 commands done):\n#    pick 237b246 Add C program skeleton\n#    edit b3f188b Flesh out C program\n# No commands remaining.\n# You are currently splitting a commit while rebasing branch 'master' on 'c785f47'.\n#\n# Changes to be committed:\n#\tmodified:   main.c\n#\n# Changes not staged for commit:\n#\tmodified:   main.c\n#<\/pre><p>\n          Save and close your editor, then we'll make the second commit. We\n          could do another interactive commit, but since we just want to include\n          the rest of the changes in this commit we'll just do this:\n        <\/p><pre>git commit -a -m\"Prompt user for their name\"\ngit rebase --continue<\/pre><p>\n          That last command tells git that we're done editing this commit, and\n          to continue to the next rebase command. That's it! Run <code>git\n          log<\/code> to see the fruits of your labor:\n        <\/p><pre>$ git log -3 --oneline\nfe19cc3 (HEAD -&gt; master) Prompt user for their name\n659a489 Add get_name function to C program\n237b246 Add C program skeleton<\/pre><h2 id=\"reorder\">Reordering commits<\/h2><p>\n          This one is pretty easy. Let's start by setting up our sandbox:\n        <\/p><pre>echo \"Goodbye now!\" &gt;farewell.txt\ngit add farewell.txt\ngit commit -m\"Add farewell.txt\"\necho \"Hello there!\" &gt;greeting.txt\ngit add greeting.txt\ngit commit -m\"Add greeting.txt\"\necho \"How're you doing?\" &gt;inquiry.txt\ngit add inquiry.txt\ngit commit -m\"Add inquiry.txt\"<\/pre><p>\n          The git log should now look like this:\n        <\/p><pre>\nf03baa5 (HEAD -&gt; master) Add inquiry.txt\na4cebf7 Add greeting.txt\n90bb015 Add farewell.txt<\/pre><p>\n          Clearly, this is all out of order. Let's do an interactive rebase of\n          the past 3 commits to resolve this. Run <code>git rebase -i\n            HEAD~3<\/code> and this rebase plan will appear:\n        <\/p><pre>pick 90bb015 Add farewell.txt\npick a4cebf7 Add greeting.txt\npick f03baa5 Add inquiry.txt\n# Rebase fe19cc3..f03baa5 onto fe19cc3 (3 commands)\n#\n# Commands:\n# p, pick &lt;commit&gt; = use commit\n#\n# These lines can be re-ordered; they are executed from top to bottom.<\/pre><p>\n          The fix is now straightforward: just reorder these lines in the order\n          you wish for the commits to appear. Should look something like this:\n        <\/p><pre>pick a4cebf7 Add greeting.txt\npick f03baa5 Add inquiry.txt\npick 90bb015 Add farewell.txt<\/pre><p>\n          Save and close your editor and git will do the rest for you. Note that\n          it's possible to end up with conflicts when you do this in practice -\n          <a href=\"#conflicts\">click here for help resolving conflicts<\/a>.\n        <\/p><h2 id=\"pull\">git pull --rebase<\/h2><p>\n          If you've been writing some commits on a branch which has been updated\n          upstream, normally <code>git pull<\/code> will create a merge commit.\n          In this respect, <code>git pull<\/code>'s behavior by default is\n          equivalent to:\n        <\/p><pre>git fetch origin\ngit merge origin\/master<\/pre><p>\n          There's another option, which is often more useful and leads to a much\n          cleaner history: <code>git pull --rebase<\/code>. Unlike the merge\n          approach, this is equivalent to the following:\n        <\/p><pre>git fetch origin\ngit rebase origin\/master<\/pre><p>\n          The merge approach is simpler and easier to understand, but the rebase\n          approach is almost always what you want to do if you understand how to\n          use git rebase. If you like, you can set it as the default behavior\n          like so:\n        <\/p><pre>git config --global pull.rebase true<\/pre><p>\n          When you do this, technically you're applying the procedure we\n          discuss in the next section... so let's explain what it means to do\n          that deliberately, too.\n        <\/p><h2 id=\"rebase\">Using git rebase to... rebase<\/h2><p>\n          Ironically, the feature of git rebase that I use the least is the one\n          it's named for: rebasing branches. Say you have the following\n          branches:\n        <\/p><pre>o--o--o--o--&gt; master\n   \\--o--o--&gt; feature-1\n      \\--o--&gt; feature-2<\/pre><p>\n          It turns out feature-2 doesn't depend on any of the changes in\n          feature-1, so you can just base it off of master. The fix is thus:\n        <\/p><pre>git checkout feature-2\ngit rebase master<\/pre><p>\n          The non-interactive rebase does the default operation for all\n          implicated commits (\"pick\")<a id=\"ref-4\" href=\"#note-4\"><sup>4<\/sup><\/a>, which simply rolls your history\n          back to the last common anscestor and replays the commits from both\n          branches. Your history now looks like this:\n        <\/p><pre>o--o--o--o--&gt; master\n   |     \\--o--&gt; feature-2\n   \\--o--o--&gt; feature-1<\/pre><h2 id=\"conflicts\">Resolving conflicts<\/h2>The details on resolving merge conflicts are beyond the scope of\n          this guide - keep your eye out for another guide for this in the\n          future. Assuming you're familiar with resolving conflicts in general,\n          here are the specifics that apply to rebasing.<p>\n          Sometimes you'll get a merge conflict when doing a rebase, which you\n          can handle just like any other merge conflict. Git will set up the\n          conflict markers in the affected files, <code>git status<\/code> will\n          show you what you need to resolve, and you can mark files as resolved\n          with <code>git add<\/code> or <code>git rm<\/code>. However, in the\n          context of a git rebase, there are two options you should be aware of.\n        <\/p><p>\n          The first is how you complete the conflict resolution. Rather than\n          <code>git commit<\/code> like you'll use when addressing conflicts that\n          arise from <code>git merge<\/code>, the appropriate command for\n          rebasing is <code>git rebase --continue<\/code>. However, there's\n          another option available to you: <code>git rebase --skip<\/code>. This\n          will skip the commit you're working on, and it won't be included in\n          the rebase. This is most common when doing a non-interactive rebase,\n          when git doesn't realize that a commit it's pulled from the \"other\"\n          branch is an updated version of the commit that it conflicts with on\n          \"our\" branch.\n        <\/p><h2 id=\"reflog\">Help! I broke it!<\/h2><p>\n          No doubt about it - rebasing can be hard sometimes. If you've made a\n          mistake and in so doing lost commits which you needed, then <code>git\n            reflog<\/code> is here to save the day. Running this command will\n          show you every operation which changed a <em>ref<\/em>, or reference -\n          that is, branches and tags. Each line shows you what the old reference\n          pointed to, and you can <code>git cherry-pick<\/code>, <code>git\n            checkout<\/code>, <code>git show<\/code>, or use any other operation\n          on git commits once thought lost.\n        <\/p><\/div>","created_at":"2019-05-10T21:35:55+0000","updated_at":"2019-05-10T21:35:55+0000","annotations":[],"mimetype":"text\/html","reading_time":16,"domain_name":"git-rebase.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":945,"title":"Parenthetical Prose to Bewildering Bytecode","url":"https:\/\/blog.ndk.io\/clojure-compilation.html","content":"<p>How does Clojure compilation work? What black magic is needed to transform elegant parenthetical prose into binary instructions executed by a processor?<\/p><p>Let\u2019s start with a simple question. Is Clojure compiled or interpreted? Someone asked this same question <a href=\"http:\/\/stackoverflow.com\/questions\/5669933\/is-clojure-compiled-or-interpreted\">on Stack Overflow<\/a>. The answer? It\u2019s complicated. Clojure code can be either dynamically loaded or AOT (ahead of time) compiled to Java bytecode. For loading Clojure code dynamically, the process looks like this:<\/p><p><img class=\"blog-img\" src=\"https:\/\/blog.ndk.io\/img\/dynamic_compilation.svg\" alt=\"image\" \/><\/p><p>Clojure Dynamic Compilation<\/p><p>For AOT compilation on the JVM, it looks more like this:<\/p><p><img class=\"blog-img\" src=\"https:\/\/blog.ndk.io\/img\/AOT_compilation.svg\" alt=\"image\" \/><\/p><p>Clojure AOT Compilation<\/p><p>So AOT compilation looks a lot like normal compilation: the code is compiled and then the compiled code is executed. For dynamic compilation the picture is a little hazier. It turns out it also uses an intermediate bytecode representation, but this is generated and used at run time without being saved to a file.<\/p><p>In many respects the process seems to be the same, but let\u2019s ignore dynamic compilation for the time being and look at AOT compilation. We\u2019re also talking only about Clojure compilation for the JVM. For AOT compilation the compiler must produce bytecode files, which are executed on the Java Virtual Machine. What do these bytecode files look like and how does the JVM execute them? Let\u2019s take a look.<\/p><h3 id=\"java-compilation\">Java Compilation<\/h3><p>The Java Virtual Machine does not understand Java but only Java bytecode. What is Java bytecode? Java bytecode is a collection of class files. Each class file contains a description of a class and the methods defined in the class, which have been translated to Java bytecode instructions. It looks like this:<\/p><p><img class=\"blog-img\" src=\"https:\/\/blog.ndk.io\/img\/class_file.svg\" alt=\"image\" \/><\/p><p>Java Class File<\/p><p>Magic is the hexadecimal string \u201c0xCAFEBABE\u201d, which just provides an easy way to see that the file is a Java class file. The next items tell the Java class file version, such as J2SE 7. The constant pool is a data structure containing references to classes, interfaces, and other constants used in the class. Following the constant pool is information about the class such as whether it is a class or an interface, whether it is public or private, its name, and the name of classes or interfaces it extends or implements. Following this are the fields (class or instance variables), methods, and attributes of the class. The code for methods is represented with Java bytecode instructions.<\/p><p>The gory details are described in the <a href=\"http:\/\/docs.oracle.com\/javase\/specs\/jvms\/se7\/html\/jvms-4.html\">JVM specification<\/a>.<\/p><p>So how does the JVM load and execute this code? The JVM <em>loads<\/em>, <em>links<\/em>, and <em>initializes<\/em> class files. Note that the class file information does not have to be stored in a file, but can also be downloaded over the network or generated on the fly. The last point is important for dynamic Clojure compilation.<\/p><p><img class=\"blog-img\" src=\"https:\/\/blog.ndk.io\/img\/loading_linking_init.gif\" alt=\"image\" \/><\/p><p>Class Loading Process<\/p><div style=\"text-align: center;\">Source: [Inside the Java Virtual Machine](http:\/\/www.artima.com\/insidejvm\/ed2\/lifetypeP.html)<\/div><p><em>Loading<\/em> is finding a binary representation of a class and creating an internal JVM representation of the class (or interface).<\/p><p><em>Linking<\/em> is adding this class to the current state of the JVM runtime so that it can be executed. The linking process also verifies that the representation of the class or interface is structurally correct and prepares static fields by creating them and setting them to default values. The linking step may also resolve symbolic references to the actual classes or interfaces that contain them.<\/p><p><em>Initialization<\/em> is executing the &lt;clinit&gt; method of the class so that the class is ready to be used. The &lt;clinit&gt; method is the class or interface initialization method, kind of like class constructors (labeled in the JVM as &lt;init&gt;) are instance initialization methods.<\/p><p>Classes are initialized only when needed. A class is initialized at JVM startup if it is the main specified class. Classes are also initialized when they are referenced using JVM instructions like <code class=\"highlighter-rouge\">new<\/code> or static method invocations, or when a sub-class is initialized. When a class needs to be initialized, it must first be loaded and linked, though these steps may occur at other times.<\/p><h3 id=\"java-compilation-hello-world\">Java Compilation Hello World<\/h3><p>That\u2019s all nice, but what does it mean? What does this look like in practice? Suppose we have the following Hello World example in the Java file Hello.java:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class Hello {\n    public static void main(String[] args){\n        System.out.println(\"Hello world\");\n    }\n}\n<\/pre><\/div><\/div><p>If we compile this using <code class=\"highlighter-rouge\">javac<\/code> it generates a single class file Hello.class with the following contents. Here we interpret the generated code with <code class=\"highlighter-rouge\">javap<\/code>.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>public class Hello extends java.lang.Object\nminor version: 0\nmajor version: 50\nConstant pool:\nconst #1 = Method\t#6.#15;         \/\/ java\/lang\/Object.\"&lt;init&gt;\":()V\nconst #2 = Field\t#16.#17;\t    \/\/ java\/lang\/System.out:Ljava\/io\/PrintStream;\nconst #3 = String\t#18;\t        \/\/ Hello world\nconst #4 = Method\t#19.#20;\t    \/\/ java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\nconst #5 = class\t#21;            \/\/ Hello\nconst #6 = class\t#22;            \/\/ java\/lang\/Object\nconst #7 = Asciz\t&lt;init&gt;;\nconst #8 = Asciz\t()V;\nconst #9 = Asciz\tCode;\nconst #10 = Asciz\tLineNumberTable;\nconst #11 = Asciz\tmain;\nconst #12 = Asciz\t([Ljava\/lang\/String;)V;\nconst #13 = Asciz\tSourceFile;\nconst #14 = Asciz\tHello.java;\nconst #15 = NameAndType\t#7:#8;      \/\/ \"&lt;init&gt;\":()V\nconst #16 = class\t#23;\t        \/\/ java\/lang\/System\nconst #17 = NameAndType\t#24:#25;    \/\/ out:Ljava\/io\/PrintStream;\nconst #18 = Asciz\tHello world;\nconst #19 = class\t#26;            \/\/ java\/io\/PrintStream\nconst #20 = NameAndType\t#27:#28;    \/\/ println:(Ljava\/lang\/String;)V\nconst #21 = Asciz\tHello;\nconst #22 = Asciz\tjava\/lang\/Object;\nconst #23 = Asciz\tjava\/lang\/System;\nconst #24 = Asciz\tout;\nconst #25 = Asciz\tLjava\/io\/PrintStream;;\nconst #26 = Asciz\tjava\/io\/PrintStream;\nconst #27 = Asciz\tprintln;\nconst #28 = Asciz\t(Ljava\/lang\/String;)V;\n{\n  public Hello();\n    aload_0\n    invokespecial\t#1; \/\/ Method java\/lang\/Object.\"&lt;init&gt;\":()V\n    return\n  public static void main(java.lang.String[]);\n    getstatic #2;     \/\/ Field java\/lang\/System.out:Ljava\/io\/PrintStream;\n    ldc\t#3;           \/\/ String Hello world\n    invokevirtual #4; \/\/ Method java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n    return\n}\n<\/pre><\/div><\/div><p>Or fitting this back to our diagram:<\/p><p><img class=\"blog-img\" src=\"https:\/\/blog.ndk.io\/img\/java_helloworld.svg\" alt=\"image\" \/><\/p><p>Java Hello World Class File<\/p><p>In this example our main method was compiled to the bytecode equivalent of the following:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public static void main(java.lang.String[]);\n  getstatic #2;     \/\/ Get static field out from java.lang.System\n  ldc #3;           \/\/ Push constant \"Hello world\" to stack\n  invokevirtual #4; \/\/ Invoke println method of java.io.PrintStream\n  return\n<\/pre><\/div><\/div><p>The JVM uses a stack to hold operands for instructions. In this example to print \u201cHello world\u201d we fetch the the static field <a href=\"http:\/\/docs.oracle.com\/javase\/7\/docs\/api\/java\/lang\/System.html#out\">java.lang.System.out<\/a>, which returns an object of type PrintStream, and push its value to the stack. Then we push our constant string \u201cHello world\u201d to the stack. Finally we call invokevirtual with constant \u201cprintln\u201d to execute <a href=\"http:\/\/docs.oracle.com\/javase\/7\/docs\/api\/java\/io\/PrintStream.html#println%28java.lang.String%29\">PrintStream.println<\/a> and print our message.<\/p><h3 id=\"java-compilation-summary\">Java Compilation Summary<\/h3><p>The JVM loads and executes class files. For Java each class file is a straightforward representation of a Java class. The JVM loads, links, and initializes these classes at run time when needed. The execution itself usually uses a form of just-in-time (JIT) compilation, where Java bytecode is binary translated to machine code just before it is needed.<\/p><p>A Clojure compiler must do a similar trick as a Java compiler. It must take a set of Clojure source code files and produce a corresponding set of class files that the JVM understands.<\/p><h3 id=\"aot-clojure-compilation-hello-world\">AOT Clojure Compilation: Hello World<\/h3><p>How does the Clojure compiler convert Clojure source to Java bytecode? Let\u2019s take a look at an example. Suppose we define a hello world Clojure file as follows:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(ns hello.core)\n(defn -main [&amp; args]\n  (println \"Hello world\"))\n<\/pre><\/div><\/div><p>This does basically the same thing as our Java example. It defines a namespace hello.core containing a function -main, which calls println to greet the world. What Java bytecode output might you expect for this? You might expect something like the Java class file above: a single class file called core.class, named after the namespace, with a single static main method that does the printing.<\/p><p>Let\u2019s test it. We create a Clojure application called \u201chello\u201d using Leiningen and modify the project configuration to specify AOT compilation. We run <code class=\"highlighter-rouge\">lein compile<\/code> to compile the code. What was produced?<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ lein compile\nCompiling hello.core\n$ ls -1 target\/classes\/hello\/\ncore$fn__16.class\ncore__init.class\ncore$loading__4910__auto__.class\ncore$_main.class\n<\/pre><\/div><\/div><p>This created not one class file, not two, but four class files. What is going on here? Where did all of that code come from?<\/p><h3 id=\"aot-clojure-source-code\">AOT: Clojure Source Code<\/h3><p>The first place to look is back at the Clojure source code.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(ns hello.core)\n(defn -main [&amp; args]\n  (println \"Hello world\"))\n<\/pre><\/div><\/div><p>That doesn\u2019t do much. But wait, Clojure has macros. Let\u2019s try macro-expanding the code once:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(do\n  (clojure.core\/in-ns 'hello.core)\n  (clojure.core\/with-loading-context\n    (clojure.core\/refer 'clojure.core))\n  (if (.equals 'hello.core 'clojure.core)\n    nil\n    (do\n      (clojure.core\/dosync\n        (clojure.core\/commute\n          @#'clojure.core\/*loaded-libs*\n          clojure.core\/conj\n          'hello.core))\n      nil)))\n(def -main\n  (clojure.core\/fn \n    ([&amp; args]\n      (println \"Hello world\"))))\n<\/pre><\/div><\/div><p>That didn\u2019t remove all of the macros. Here\u2019s the full macro expansion:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(do\n  (clojure.core\/in-ns 'hello.core)\n  ((fn*\n    loading__4910__auto__\n    ([]\n       (. clojure.lang.Var\n          (clojure.core\/pushThreadBindings\n           {clojure.lang.Compiler\/LOADER\n            (. (. loading__4910__auto__ getClass) getClassLoader)}))\n       (try\n         (clojure.core\/refer 'clojure.core)\n         (finally\n           (. clojure.lang.Var (clojure.core\/popThreadBindings)))))))\n  (if (. 'hello.core equals 'clojure.core)\n    nil\n    (do\n      (. clojure.lang.LockingTransaction\n         (clojure.core\/runInTransaction\n          (fn*\n           ([]\n              (clojure.core\/commute\n               @#'clojure.core\/*loaded-libs*\n               clojure.core\/conj\n               'hello.core)))))\n      nil)))\n(def -main \n  (fn* ([&amp; args]\n    (println \"Hello world\"))))\n<\/pre><\/div><\/div><p>What are all of those quotes and @ symbols? That doesn\u2019t look very Lispy. Oh yeah. Reader macros. Let\u2019s remove those, too:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(do\n  (clojure.core\/in-ns (quote hello.core))\n  ((fn*\n    loading__4910__auto__\n    ([]\n       (. clojure.lang.Var\n          (clojure.core\/pushThreadBindings\n           {clojure.lang.Compiler\/LOADER\n            (. (. loading__4910__auto__ getClass) getClassLoader)}))\n       (try\n         (clojure.core\/refer (quote clojure.core))\n         (finally\n           (. clojure.lang.Var (clojure.core\/popThreadBindings)))))))\n  (if (. (quote hello.core) equals (quote clojure.core))\n    nil\n    (do\n      (. clojure.lang.LockingTransaction\n         (clojure.core\/runInTransaction\n          (fn*\n           ([]\n              (clojure.core\/commute\n               (deref (var clojure.core\/*loaded-libs*))\n               clojure.core\/conj\n               (quote hello.core))))))\n      nil)))\n(def -main \n  (fn* ([&amp; args]\n    (println \"Hello world\"))))\n<\/pre><\/div><\/div><p>It doesn\u2019t look so simple anymore, does it?<\/p><h3 id=\"generated-files\">Generated files<\/h3><p>If we closely compare the fully expanded Clojure code and the generated files we notice a few things. Three of the generated class files correspond to functions. One is for the -main function and the other two are for the anonymous functions within the namespace macro. The final class is a loader class for the core namespace (or hello\/core.clj file, as each namespace corresponds to a file).<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>core__init.class \/\/ core.clj loader class\n  public static {}\n  public load()\n  public static void __init0()\ncore$_main.class \/\/ -main function\n  public static {}\n  public hello.core$_main()\n  public java.lang.Object doInvoke(java.lang.Object)\n  public int getRequiredArity()\ncore$fn__16.class \/\/ dosync anonymous function\n  public static {}\n  public hello.core$fn__16()\n  public java.lang.Object invoke()\ncore$loading__4910__auto__.class \/\/ with-loading-context anonymous function\n  public static {}\n  public hello.core$loading__4910__auto__()\n  public java.lang.Object invoke()\n<\/pre><\/div><\/div><p>Clojure has first-class functions, which can be passed around like any other variable. The JVM doesn\u2019t support first-class functions, but recognizes only data represented as class files. Clojure seems to use the same approach as other languages like Scala for representing functions. Each function is separated into its own class with a method to invoke the function here named <code class=\"highlighter-rouge\">doInvoke<\/code> for a normal function and <code class=\"highlighter-rouge\">invoke<\/code> for the anonymous functions.<\/p><p>We\u2019ll take a look at each of these files in turn.<\/p><h3 id=\"core__init-loader-class-for-core-namespace\">core__init: Loader class for core namespace<\/h3><p>According to <a href=\"http:\/\/clojure.org\/compilation\">Clojure.org documentation<\/a>, for each namespace a loader class with an __init suffix is created. This is the loader class for the hello.core namespace.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>package hello;\nimport clojure.lang.AFn;\nimport clojure.lang.Compiler;\nimport clojure.lang.IFn;\nimport clojure.lang.IPersistentMap;\nimport clojure.lang.Keyword;\nimport clojure.lang.LockingTransaction;\nimport clojure.lang.PersistentList;\nimport clojure.lang.RT;\nimport clojure.lang.Symbol;\nimport clojure.lang.Var;\nimport java.util.Arrays;\nimport java.util.concurrent.Callable;\npublic class core__init\n{\n  public static final Var const__0;\n  public static final AFn const__1;\n  public static final AFn const__2;\n  public static final Var const__3;\n  public static final Keyword const__4;\n  public static final Keyword const__5;\n  public static final Object const__6;\n  public static final Keyword const__7;\n  public static final Object const__8;\n  public static final Keyword const__9;\n  public static final Object const__10;\n  public static final AFn const__11;\n  public static void load() {\n    if (((Symbol)const__1).equals(const__2)){\n       tmpTernaryOp = null;\n       break label67;\n       ((IFn)new core.loading__4910__auto__()).invoke();\n    } else {\n      LockingTransaction.runInTransaction((Callable)new core.fn__16());\n    }\n    label67: tmp70_67 = const__3;\n    tmp70_67.setMeta((IPersistentMap)const__11);\n    tmp70_67.bindRoot(new core._main());\n  }\n  public static void __init0() {\n    const__0 = (Var)RT.var(\"clojure.core\", \"in-ns\");\n    const__1 = (AFn)Symbol.intern(null, \"hello.core\");\n    const__2 = (AFn)Symbol.intern(null, \"clojure.core\");\n    const__3 = (Var)RT.var(\"hello.core\", \"-main\");\n    const__4 = (Keyword)RT.keyword(null, \"file\");\n    const__5 = (Keyword)RT.keyword(null, \"column\");\n    const__6 = Integer.valueOf(1);\n    const__7 = (Keyword)RT.keyword(null, \"line\");\n    const__8 = Integer.valueOf(3);\n    const__9 = (Keyword)RT.keyword(null, \"arglists\");\n    const__10 = PersistentList.create(Arrays.asList(new Object[] {\n      RT.vector(new Object[] {\n        Symbol.intern(null, \"&amp;\"),\n        Symbol.intern(null, \"args\")\n      })\n    }));\n    const__11 = (AFn)RT.map(new Object[] {\n      RT.keyword(null, \"arglists\"),\n      PersistentList.create(Arrays.asList(new Object[] {\n        RT.vector(new Object[] {\n          Symbol.intern(null, \"&amp;\"), \n          Symbol.intern(null, \"args\")\n        })\n      })), \n      RT.keyword(null, \"column\"),\n      Integer.valueOf(1),\n      RT.keyword(null, \"line\"),\n      Integer.valueOf(3),\n      RT.keyword(null, \"file\"),\n      \"hello\/core.clj\"\n    });\n  }\n  static\n  {\n    __init0();\n    Compiler.pushNSandLoader(Class.forName(\"hello.core__init\").getClassLoader());\n    try\n    {\n      load();\n      Var.popThreadBindings();\n    }\n    finally\n    {\n      Var.popThreadBindings();\n      throw finally;\n    }\n  }\n}\n<\/pre><\/div><\/div><p>When the class is loaded, the static initializer method is called. This method in turn invokes the __init0 and load methods. The _init0 method sets up a bunch of variables with calls to <code class=\"highlighter-rouge\">RT.var<\/code>, <code class=\"highlighter-rouge\">RT.keyword<\/code>, and <code class=\"highlighter-rouge\">Symbol.intern<\/code>.<\/p><p>It also creates a map containing metadata for the -main function. The \u201cfile\u201d, \u201ccolumn\u201d, \u201cline\u201d values tell the location of the -main function in the core namespace file. The \u201carglists\u201d identifier gives the arguments that the function accepts. All of this is stored in a map, which would be represented in Clojure as follows:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>{:arglists [\"&amp;\" \"args\"]\n :column 1\n :line   3\n :keyword \"hello\/core.clj\"}\n<\/pre><\/div><\/div><p>After calling __init0, the static initializer calls the load method within a block executing <a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/master\/src\/jvm\/clojure\/lang\/Compiler.java\">Compiler.pushNSandLoader<\/a> beforehand and <a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/master\/src\/jvm\/clojure\/lang\/Var.java\">Var.popThreadBindings<\/a> afterwards. What it does is roughly equivalent to the following Clojure code, <a href=\"http:\/\/www.deepbluelambda.org\/programming\/clojure\/how-clojure-works-a-simple-namespace\">as presented by Daniel Solano G\u00f3mez<\/a>.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(bind [clojure.core\/*ns*         nil\n       clojure.core\/*fn-loader*  loader\n       clojure.core\/*read-eval*  true]\n  (greeter.hello_init\/load))\n<\/pre><\/div><\/div><p>Finally, the load method performs the actual code of the ns function i.e. it does the equivalent of the following code:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(do\n  (clojure.core\/in-ns (quote hello.core))\n  (loading__4910__auto__)\n  (if (. (quote hello.core) equals (quote clojure.core)) nil\n    (do\n      (. clojure.lang.LockingTransaction\n         (clojure.core\/runInTransaction\n          (fn__16))))\n      nil)))\n<\/pre><\/div><\/div><p>Remember that our two anonymous functions have been deferred to separate files.<\/p><h3 id=\"core_main-main-function\">core$_main: main function<\/h3><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>package hello;\nimport clojure.lang.IFn;\nimport clojure.lang.RT;\nimport clojure.lang.RestFn;\nimport clojure.lang.Var;\npublic final class core$_main extends RestFn\n{\n  public static final Var const__0 = (Var)RT.var(\"clojure.core\", \"println\");\n  public Object doInvoke(Object args) {\n    return ((IFn)const__0.getRawRoot()).invoke(\"Hello world\");\n  }\n  public int getRequiredArity()\n  {\n    return 0;\n  }\n}\n<\/pre><\/div><\/div><p>The <code class=\"highlighter-rouge\">-main<\/code> function class file is quite simple. It fetches the root binding for the <code class=\"highlighter-rouge\">clojure.core\/println<\/code> var, casts it as a function, and invokes it with our constant string <code class=\"highlighter-rouge\">\"Hello world\"<\/code>.<\/p><p>In addition it has a method <code class=\"highlighter-rouge\">getRequiredArity<\/code>. A few simple experiments show that this returns the number of fixed arguments that a variadic function take. In this case our main method does not have any fixed arguments, so it returns zero.<\/p><p>For a better comparison with the Java example, let\u2019s look at a more direct interpretation of the bytecode with <code class=\"highlighter-rouge\">javap<\/code>:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>public final class hello.core$_main extends clojure.lang.RestFn\n  minor version: 0\n  major version: 49\n  flags: ACC_PUBLIC, ACC_FINAL, ACC_SUPER\nConstant pool:\n   #1 = Utf8 hello\/core$_main\n   #2 = Class #1                   \/\/  hello\/core$_main\n   #3 = Utf8 clojure\/lang\/RestFn\n   #4 = Class #3                   \/\/  clojure\/lang\/RestFn\n   #5 = Utf8 core.clj\n   #6 = Utf8 const__0\n   #7 = Utf8 Lclojure\/lang\/Var;\n   #8 = Utf8 &lt;clinit&gt;\n   #9 = Utf8 ()V\n  #10 = Utf8 clojure.core\n  #11 = String #10                 \/\/  clojure.core\n  #12 = Utf8 println\n  #13 = String #12                 \/\/  println\n  #14 = Utf8 clojure\/lang\/RT\n  #15 = Class #14                  \/\/  clojure\/lang\/RT\n  #16 = Utf8 var\n  #17 = Utf8 (Ljava\/lang\/String;Ljava\/lang\/String;)Lclojure\/lang\/Var;\n  #18 = NameAndType #16:#17        \/\/  var:(Ljava\/lang\/String;Ljava\/lang\/String;)\n                                   \/\/  Lclojure\/lang\/Var;\n  #19 = Methodref #15.#18          \/\/  clojure\/lang\/RT.var:(Ljava\/lang\/String;\n                                   \/\/  Ljava\/lang\/String;)Lclojure\/lang\/Var;\n  #20 = Utf8 clojure\/lang\/Var\n  #21 = Class #20                  \/\/  clojure\/lang\/Var\n  #22 = NameAndType #6:#7          \/\/  const__0:Lclojure\/lang\/Var;\n  #23 = Fieldref #2.#22            \/\/  hello\/core$_main.const__0:Lclojure\/lang\/Var;\n  #24 = Utf8 &lt;init&gt;\n  #25 = NameAndType #24:#9         \/\/  \"&lt;init&gt;\":()V\n  #26 = Methodref #4.#25           \/\/  clojure\/lang\/RestFn.\"&lt;init&gt;\":()V\n  #27 = Utf8 doInvoke\n  #28 = Utf8 (Ljava\/lang\/Object;)Ljava\/lang\/Object;\n  #29 = Utf8 getRawRoot\n  #30 = Utf8 ()Ljava\/lang\/Object;\n  #31 = NameAndType #29:#30        \/\/  getRawRoot:()Ljava\/lang\/Object;\n  #32 = Methodref #21.#31          \/\/  clojure\/lang\/Var.getRawRoot:()Ljava\/lang\/Object;\n  #33 = Utf8 clojure\/lang\/IFn\n  #34 = Class #33                  \/\/  clojure\/lang\/IFn\n  #35 = Utf8 Hello world\n  #36 = String #35                 \/\/  Hello world\n  #37 = Utf8 invoke\n  #38 = NameAndType #37:#28        \/\/  invoke:(Ljava\/lang\/Object;)Ljava\/lang\/Object;\n  #39 = InterfaceMethodref #34.#38 \/\/  clojure\/lang\/IFn.invoke:(Ljava\/lang\/Object;)\n                                   \/\/  Ljava\/lang\/Object;\n  #40 = Utf8 this\n  #41 = Utf8 Ljava\/lang\/Object;\n  #42 = Utf8 args\n  #43 = Utf8 getRequiredArity\n  #44 = Utf8 ()I\n  #45 = Utf8 Code\n  #46 = Utf8 LineNumberTable\n  #47 = Utf8 LocalVariableTable\n  #48 = Utf8 SourceFile\n  #49 = Utf8 SourceDebugExtension\n{\n  public static final clojure.lang.Var const__0;\n    flags: ACC_PUBLIC, ACC_STATIC, ACC_FINAL\n  public static {};\n    flags: ACC_PUBLIC, ACC_STATIC\n    ldc           #11                 \/\/ String clojure.core\n    ldc           #13                 \/\/ String println\n    invokestatic  #19                 \/\/ Method clojure\/lang\/RT.var:(Ljava\/lang\/String;\n                                      \/\/ Ljava\/lang\/String;)Lclojure\/lang\/Var;\n    checkcast     #21                 \/\/ class clojure\/lang\/Var\n    putstatic     #23                 \/\/ Field const__0:Lclojure\/lang\/Var;\n    return        \n  public hello.core$_main();\n    flags: ACC_PUBLIC\n    aload_0       \n    invokespecial #26                 \/\/ Method clojure\/lang\/RestFn.\"&lt;init&gt;\":()V\n    return        \n  public java.lang.Object doInvoke(java.lang.Object);\n    flags: ACC_PUBLIC\n    getstatic     #23                 \/\/ Field const__0:Lclojure\/lang\/Var;\n    invokevirtual #32                 \/\/ Method clojure\/lang\/Var.getRawRoot:\n                                      \/\/ ()Ljava\/lang\/Object;\n    checkcast     #34                 \/\/ class clojure\/lang\/IFn\n    ldc           #36                 \/\/ String Hello world\n    invokeinterface #39,  2           \/\/ InterfaceMethod clojure\/lang\/IFn.invoke:\n                                      \/\/ (Ljava\/lang\/Object;)Ljava\/lang\/Object;\n    areturn       \n  public int getRequiredArity();\n    flags: ACC_PUBLIC\n    iconst_0      \n    ireturn       \n}\n<\/pre><\/div><\/div><p>We notice two key differences when comparing this to the compiled Java Hello World file:<\/p><ul><li>Clojure uses an extra level of indirection for calling the print function. Java fetches <code class=\"highlighter-rouge\">java.lang.System.out<\/code> and calls <code class=\"highlighter-rouge\">invokevirtual<\/code> to print. Clojure loads the var pointing to the function, calls <code class=\"highlighter-rouge\">invokevirtual<\/code> to get the function value, casts it to a function, and calls <code class=\"highlighter-rouge\">invokeinterface<\/code> to print.<\/li>\n  <li>Clojure does more setup work. Java just loads <code class=\"highlighter-rouge\">java.lang.System.out<\/code> directly. Clojure sets up a var with its root binding pointing to the <code class=\"highlighter-rouge\">clojure.core\/println<\/code> function beforehand in the static initializer.<\/li>\n<\/ul><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ core$loading__4910__auto__.class\npackage hello;\nimport clojure.lang.AFn;\nimport clojure.lang.AFunction;\nimport clojure.lang.Associative;\nimport clojure.lang.Compiler;\nimport clojure.lang.IFn;\nimport clojure.lang.RT;\nimport clojure.lang.Symbol;\nimport clojure.lang.Var;\npublic final class core$loading__4910__auto__ extends AFunction\n{\n  public static final Var const__0 = (Var)RT.var(\"clojure.core\", \"refer\");\n  public static final AFn const__1 = (AFn)Symbol.intern(null, \"clojure.core\");\n  public Object invoke() {\n    Var.pushThreadBindings((Associative)RT.mapUniqueKeys(new Object[] {\n      Compiler.LOADER, ((Class)getClass()).getClassLoader()\n    }));\n    Object localObject1;\n    try {\n      localObject1 = ((IFn)const__0.getRawRoot()).invoke(const__1);\n      Var.popThreadBindings();\n    } finally {\n      Var.popThreadBindings();\n    }\n    return localObject1;\n  }\n}\n<\/pre><\/div><\/div><p>This class corresponds to the expansion of <code class=\"highlighter-rouge\">with-loading-context<\/code> from the <code class=\"highlighter-rouge\">ns<\/code> macro:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>([] (. clojure.lang.Var\n       (clojure.core\/pushThreadBindings\n         {clojure.lang.Compiler\/LOADER\n           (. (. loading__4910__auto__ getClass) getClassLoader)}))\n    (try\n      (clojure.core\/refer (quote clojure.core))\n      (finally\n        (. clojure.lang.Var (clojure.core\/popThreadBindings)))))\n<\/pre><\/div><\/div><p>The compiled code seems to be a straightforward interpretation of the original code. Clojure uses a custom class loader (defined in <a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/master\/src\/jvm\/clojure\/lang\/DynamicClassLoader.java\">DynamicClassLoader<\/a>) instead of the default JVM bootstrap loader. According to the <a href=\"http:\/\/docs.oracle.com\/javase\/specs\/jvms\/se7\/html\/jvms-5.html#jvms-5.3\">JVM specification<\/a>, \u201capplications employ user-defined class loaders in order to extend the manner in which the Java Virtual Machine dynamically loads and thereby creates classes.\u201d For dynamic compilation it is clear that a custom class loader is necessary. Offhand it seems like AOT compilation could use the default bootstrap loader, but there may be good reasons for using a custom class loader that I\u2019m not familiar with.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ core$fn__16.class\npackage hello;\nimport clojure.lang.AFn;\nimport clojure.lang.AFunction;\nimport clojure.lang.IFn;\nimport clojure.lang.RT;\nimport clojure.lang.Symbol;\nimport clojure.lang.Var;\npublic final class core$fn__16 extends AFunction\n{\n  public static final Var const__0 = (Var)RT.var(\"clojure.core\", \"commute\");\n  public static final Var const__1 = (Var)RT.var(\"clojure.core\", \"deref\");\n  public static final Var const__2 = (Var)RT.var(\"clojure.core\", \"*loaded-libs*\");\n  public static final Var const__3 = (Var)RT.var(\"clojure.core\", \"conj\");\n  public static final AFn const__4 = (AFn)Symbol.intern(null, \"hello.core\");\n  public Object invoke() {\n    return ((IFn)const__0.getRawRoot())\n      .invoke(((IFn)const__1.getRawRoot()).invoke(const__2),\n              const__3.getRawRoot(),\n              const__4);\n  }\n}\n<\/pre><\/div><\/div><p>And here\u2019s the corresponding source code within the expansion of <code class=\"highlighter-rouge\">dosync<\/code> in the <code class=\"highlighter-rouge\">ns<\/code> form:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(fn* ([] (clojure.core\/commute\n           (deref (var clojure.core\/*loaded-libs*))\n           clojure.core\/conj\n           (quote hello.core))))\n<\/pre><\/div><\/div><p>This class file fetches and invokes <code class=\"highlighter-rouge\">clojure.core\/commute<\/code> with the given arguments. Note again the setup work and indirection. Four Vars are created and one Symbol. To invoke the <code class=\"highlighter-rouge\">commute<\/code> function, it gets the root binding of the <code class=\"highlighter-rouge\">commute<\/code> var declared earlier and invokes it. Similarly the root binding of <code class=\"highlighter-rouge\">deref<\/code> is fetched and invoked and the root binding of <code class=\"highlighter-rouge\">conj<\/code> is fetched.<\/p><h3 id=\"summary\">Summary<\/h3><p>This post looked at Java compilation and Clojure AOT compilation to try and understand the Clojure compilation better. What did we learn?<\/p><ul><li>The JVM understands Java class files, which consist of class file info, a constant pool, and class info and compiled methods.<\/li>\n  <li>The JVM loads, links, and initializes class files dynamically when needed.<\/li>\n  <li>Java Hello World compiles to a single, simple class file.<\/li>\n  <li>Clojure Hello World compiles to four class files with some interesting things happening in the namespace macro.<\/li>\n  <li>A few key differences between compiled Java and Clojure:\n    <ul><li>Clojure functions are compiled to separate class files.<\/li>\n      <li>Clojure uses a custom class loader.<\/li>\n      <li>Clojure does more setup work, creating Vars to point to each function a class file needs (among other things).<\/li>\n      <li>Clojure calls functions indirectly, first fetching the value of a Var and then invoking it.<\/li>\n    <\/ul><\/li>\n<\/ul>","created_at":"2019-05-10T13:26:21+0000","updated_at":"2019-05-10T13:26:21+0000","annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"blog.ndk.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["css","clojurescript","reagent"],"is_public":false,"id":944,"title":"Jarzka\/stylefy","url":"https:\/\/github.com\/Jarzka\/stylefy","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Jarzka\/stylefy\/blob\/master\/stylefy_logo_small.png\"><img src=\"https:\/\/github.com\/Jarzka\/stylefy\/raw\/master\/stylefy_logo_small.png\" alt=\"\" \/><\/a><\/p>\n<p><a href=\"https:\/\/clojars.org\/stylefy\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/1f6318a5cec760df8824b851d37960b0d656445c\/68747470733a2f2f696d672e736869656c64732e696f2f636c6f6a6172732f762f7374796c6566792e737667\" alt=\"Clojars Project\" data-canonical-src=\"https:\/\/img.shields.io\/clojars\/v\/stylefy.svg\" \/><\/a>\n<a href=\"https:\/\/circleci.com\/gh\/Jarzka\/stylefy\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/3f63ca3971b1f7abb9b74b42ea5a66da286ef239\/68747470733a2f2f636972636c6563692e636f6d2f67682f4a61727a6b612f7374796c6566792e7376673f7374796c653d737667\" alt=\"CircleCI\" data-canonical-src=\"https:\/\/circleci.com\/gh\/Jarzka\/stylefy.svg?style=svg\" \/><\/a><\/p>\n<p>ClojureScript library for styling UI components with ease.<\/p>\n<p><a href=\"https:\/\/jarzka.github.io\/stylefy\/doc\" rel=\"nofollow\">API documentation<\/a><\/p>\n<p>stylefy makes it possible to define UI component styles as Clojure data and attach them into components easily. When styles are defined as Clojure data, they can be easily transformed with Clojure's powerful functions (like merge) and parametrised. Styles are converted to CSS on-demand, and since the converted CSS is handled internally by the library, there is no need to worry about things like writing selectors, name conflicts, difficult cascading, dead CSS code etc. All in all the purpose of this library is to make writing style code easier and more maintainable.<\/p>\n<p>stylefy is designed to be used in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Single-page_application\" rel=\"nofollow\">SPA<\/a> applications along with <a href=\"https:\/\/github.com\/reagent-project\/reagent\">Reagent<\/a>. stylefy uses <a href=\"https:\/\/github.com\/noprompt\/garden\">Garden<\/a> in the background to do most of its CSS conversions.<\/p>\n<ul><li>Define styles as Clojure data, for any UI component or HTML tag<\/li>\n<li>Sub-styles help you to define a style for your component and all the elements inside of it in a single map<\/li>\n<li>Use any 3rd party CSS code (such as Bootstrap) along with stylefy<\/li>\n<li>Vendor prefixes, define which vendor prefixes are used and which properties should be prefixed (locally or globally)<\/li>\n<li>Media queries, define how your style looks on various screen sizes<\/li>\n<li>Feature queries, define how your style looks when certain CSS features are not supported by the browser (uses CSS @supports query)<\/li>\n<li>CSS pseudo-classes, pseudo-elements, keyframes &amp; font-faces are supported<\/li>\n<li>Style caching using local storage (can be turned off)<\/li>\n<li>Multi-instance support (you can run multiple apps using stylefy on the same web page)<\/li>\n<li>Manual mode for styling 3rd party components and resolving corner cases in which complex CSS selectors are needed<\/li>\n<li>Small and simple core API<\/li>\n<li>Automatic style reloading with <a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\">Figwheel<\/a><\/li>\n<li>All features are tested to work with Chrome, Firefox, Edge &amp; Internet Explorer 11<\/li>\n<\/ul>\n<h2><a id=\"user-content-how-is-this-library-different-than-garden\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-is-this-library-different-than-garden\"><\/a>How is this library different than Garden?<\/h2>\n<p>Garden is awesome, but it's \"only\" a Clojure to CSS generator. If you want to use Garden to style your page, you are pretty much going to write CSS code as usual, i.e. write classes and selectors to stylize things on the page. You also need to avoid CSS quirks like name conflicts and make sure you always handle CSS cascading process correctly. stylefy helps you with this; you just write your style definition in a map and attach it to your component in the render function by calling <strong>use-style<\/strong>. There is no need to write CSS classes or selectors, no need to worry about name conflicts etc<\/p>\n<p>Yes, it is possible to easily attach styles to components with Garden too if you use inline styles. But if you use stylefy, all your style definitions are converted to unique CSS classes automatically and the corresponding class is attached to your component. This is more effective than using inline-styles, especially if the same component exists multiple times on the same page. The style is defined only once in the CSS class, not multiple times in each component instance. Also, some CSS features like pseudoclasses (:hover etc.), media queries and feature queries are not available to use as inline styles. For stylefy, this is not a problem, as it allows you to define pseudoclasses and media queries within the style map and converts everything to CSS automatically.<\/p>\n<p>TLDR; stylefy it's like using inline CSS, but with full support for all CSS features that are not normally available when using CSS inline.<\/p>\n<h2><a id=\"user-content-any-real-projects-using-stylefy\" class=\"anchor\" aria-hidden=\"true\" href=\"#any-real-projects-using-stylefy\"><\/a>Any real projects using stylefy?<\/h2>\n<p>Yup, for example:<\/p>\n<p>Add Leiningen project dependency:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>[stylefy \"1.12.0\"]<\/pre><\/div>\n<h2><a id=\"user-content-init\" class=\"anchor\" aria-hidden=\"true\" href=\"#init\"><\/a>Init<\/h2>\n<p>Add the following <strong>style<\/strong> tags on your page's <strong>head<\/strong> tag. It is <strong>recommended<\/strong> that these tags are the last style tags in the <strong>header<\/strong> so that it is less likely that possible other styles would override them.<\/p>\n<p>The first tag is going to contain CSS definitions that are not going to change (font-face, keyframes etc.). The second will contain class definitions that are added into the DOM on-demand when components need them.<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;style id=\"_stylefy-constant-styles_\"&gt;&lt;\/style&gt;\n&lt;style id=\"_stylefy-styles_\"&gt;&lt;\/style&gt;<\/pre><\/div>\n<p>Require stylefy:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(:require [stylefy.core :as stylefy])<\/pre><\/div>\n<p>Then, call <strong>stylefy\/init<\/strong> once when your application starts:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/init)<\/pre><\/div>\n<h2><a id=\"user-content-creating--using-styles\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating--using-styles\"><\/a>Creating &amp; using styles<\/h2>\n<p>Create a style as a normal Clojure map:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def button-style {:padding \"25px\"\n                   :background-color \"#BBBBBB\"\n                   :border \"1px solid black\"})<\/pre><\/div>\n<p>To use it in a component, use the <strong>use-style<\/strong> function:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn- button [text]\n  [:div (use-style button-style)\n    text])<\/pre><\/div>\n<p><strong>use-style<\/strong> accepts HTML attributes as the second parameter:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn- button [text]\n  [:div (use-style button-style {:on-click #(.log js\/console \"Click!\")\n                                 :class \"some-3rd-party-button-class\"})\n    text])<\/pre><\/div>\n<h3><a id=\"user-content-combine--parametrise-styles\" class=\"anchor\" aria-hidden=\"true\" href=\"#combine--parametrise-styles\"><\/a>Combine &amp; parametrise styles<\/h3>\n<p>Combine or parametrise styles however you like:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def primary-button (merge generic-button {:background-color \"rgb(88, 121, 193)\"}))\n(defn button-style [background-color]\n  (merge generic-button {:background-color background-color}))<\/pre><\/div>\n<h3><a id=\"user-content-passing-styles-to-components\" class=\"anchor\" aria-hidden=\"true\" href=\"#passing-styles-to-components\"><\/a>Passing styles to components<\/h3>\n<p><strong>use-style<\/strong> is designed to be called only inside component render functions to define styles for <strong>HTML<\/strong> elements. If you need to pass styles to Reagent components, pass them as regular Clojure maps, and call <strong>use-style<\/strong> last, only for HTML elements:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn- button-with-custom-style [text style]\n   [:button (use-style style {:on-click #(.log js\/console \"Click!\"))\n     text])\n     \n(defn- button-wrapper []\n  [:div\n    ;; This is OK\n    [button-with-custom-style \"Hello\" {:padding \"25px\"\n                                       :background-color \"#BBBBBB\"\n                                       :border \"1px solid black\"}]\n    ;; This is NOT ok, because use-style would be called twice, second time with incorrect arguments\n    [button-with-custom-style \"Hello\" (use-style {:padding \"25px\"\n                                                  :background-color \"#BBBBBB\"\n                                                  :border \"1px solid black\"})]])<\/pre><\/div>\n<h3><a id=\"user-content-how-it-works-technical-details\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-it-works-technical-details\"><\/a>How it works (technical details)<\/h3>\n<p>Calling <strong>use-style<\/strong> asks stylefy to save the style (if it has not been saved already) and add it into the DOM as CSS class asynchronously (soon, but not immediately). The return value is a map containing the given style properties as inline style. It is needed until the CSS class has been generated and inserted into the DOM. When the DOM is ready, the component is forced to re-render itself and use only the CSS class definition.<\/p>\n<p>You might ask why does <strong>use-style<\/strong> work asynchronously? Consider a case when one or more components are going to be rendered and all of them are calling <strong>use-style<\/strong> very many times with different style maps. In this case, updating the DOM on every single call would slow the rendering process down. To keep the rendering fast, the idea is to collect as many style maps as possible during a single render event, convert all of them to CSS and add into the DOM at once.<\/p>\n<p>If the style contains some specific definitions that cannot be present as inline style (some specific\nor media queries), the component is going to be hidden for a few milliseconds with CSS <strong>visibility<\/strong> set to <strong>hidden<\/strong>, until the converted CSS style is added into the DOM. In most cases, this should not be a problem, but if needed, styles can also be added into the DOM synchronously (immediately) by calling <strong>prepare-styles<\/strong>. It is recommended to call this function during the <em>:component-will-mount<\/em> lifecycle method. It makes sure the given styles are completely ready to be used when the component needs them.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(r\/create-class\n  {:component-will-mount #(stylefy\/prepare-styles [style1 style2 style3])\n   :render (fn []\n             [:div (use-style style1)\n               [:div (use-style style2)]\n               [:div (use-style style3)]])})<\/pre><\/div>\n<p>It's good to keep in mind that most of the time <strong>prepare-styles<\/strong> is not needed, but calling <strong>use-style<\/strong> is enough. Also, when caching is used, the style will be ready after its CSS has been created for the first time.<\/p>\n<h2><a id=\"user-content-modes-pseudo-classes--pseudo-elements\" class=\"anchor\" aria-hidden=\"true\" href=\"#modes-pseudo-classes--pseudo-elements\"><\/a>Modes (pseudo-classes &amp; pseudo-elements)<\/h2>\n<p>Define how your style looks in different modes, such as when mouse is on top of an element using the style:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def simple-element {:background-color \"rgb(88, 121, 193)\"\n                     ::stylefy\/mode {:hover {:background-color \"rgb(98, 131, 213)\"}\n                                     :before {:content \"'CSS generated content'\"}\n                                     \"::-webkit-progress-bar\" {:-webkit-appearance \"none\"}}})<\/pre><\/div>\n<p>stylefy modes are pretty much the same thing as pseudoclasses\/pseudoelements in CSS and they simply create a new \"class:mode\" selector for your style. The reason for not using the name pseudoclass is completely self-willed; I think \"mode\" as a name is a little bit more informative than CSS pseudoclasses.<\/p>\n<h2><a id=\"user-content-sub-styles\" class=\"anchor\" aria-hidden=\"true\" href=\"#sub-styles\"><\/a>Sub-styles<\/h2>\n<p>Define a style for your component and all the elements inside of it in a single map:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def list-container-style (merge generic-container\n                                 {::stylefy\/sub-styles {:list {:margin-top \"1em\"}\n                                                        :list-item {:color \"black\"}}}))\n(defn list-in-container []\n  [:div (use-style list-container-style)\n   [:ul (use-sub-style list-container-style :list)\n    [:li (use-sub-style list-container-style :list-item) \"List element 1\"]\n    [:li (use-sub-style list-container-style :list-item) \"List element 2\"]\n    [:li (use-sub-style list-container-style :list-item) \"List element 3\"]]])<\/pre><\/div>\n<p>Another version using deeper sub-style nesting:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def list-container-style (merge generic-container\n                                 {::stylefy\/sub-styles\n                                   {:list {:margin-top \"1em\"\n                                           ::stylefy\/sub-styles {:item {:color \"black\"}}}}}))\n(defn list-in-container []\n  [:div (use-style list-container-style)\n   [:ul (use-sub-style list-container-style :list)\n    [:li (use-style (sub-style list-container-style :list :item)) \"List element 1\"]\n    [:li (use-style (sub-style list-container-style :list :item)) \"List element 2\"]\n    [:li (use-style (sub-style list-container-style :list :item)) \"List element 3\"]]])<\/pre><\/div>\n<p>Sub-styles are nothing special, they are supposed to contain the same contents as the main style map. If you wish, you can always omit them and use the regular <strong>use-style<\/strong> without any sub-styles. However, sub-styles help you to define styles that are closely related to each other in a single map, which can make style maintenance easier.<\/p>\n<h2><a id=\"user-content-vendor-prefixes\" class=\"anchor\" aria-hidden=\"true\" href=\"#vendor-prefixes\"><\/a>Vendor prefixes<\/h2>\n<p>Supported in the same way as Garden supports them.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def button {:border \"1px solid black\"\n             :background-color \"#888888\"\n             :border-radius \"5px\"\n             :color \"white\"\n             :text-align :center\n             :padding \"5px\"\n             :width \"150px\"\n             :height \"38px\"\n             ::stylefy\/vendors [\"webkit\" \"moz\" \"o\"]\n             ::stylefy\/auto-prefix #{:border-radius}})<\/pre><\/div>\n<p>You can also use globally defined vendor prefixes. These prefixes are automatically added into every style map.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/init {:global-vendor-prefixes {::stylefy\/vendors [\"webkit\" \"moz\" \"o\"]\n                                        ::stylefy\/auto-prefix #{:border-radius}}})<\/pre><\/div>\n<h2><a id=\"user-content-media-queries\" class=\"anchor\" aria-hidden=\"true\" href=\"#media-queries\"><\/a>Media queries<\/h2>\n<p>Define how your style looks on various screen sizes:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def phone-width \"414px\")\n(def column {:padding \"5px\"\n             :color \"white\"})\n(def responsive-layout {:display :flex\n                        :flex-direction :row\n                        ::stylefy\/media {{:max-width phone-width} {:flex-direction :column}}\n                        ::stylefy\/sub-styles {:column1 (merge column\n                                                              {:background-color \"#AA0000\"\n                                                               :flex 1})\n                                              :column2 (merge column\n                                                              {:background-color \"#00AA00\"\n                                                               :flex 2})\n                                              :column3 (merge column\n                                                              {:background-color \"#0000AA\"\n                                                               :flex 1})}})\n(defn responsive-layout []\n  [:div (use-style responsive-layout)\n   [:div (use-sub-style responsive-layout :column1)\n    [:p \"This is column 1\"]]\n   [:div (use-sub-style responsive-layout :column2)\n    [:p \"This is column 2\"]]\n   [:div (use-sub-style responsive-layout :column3)\n    [:p \"This is column 3\"]]])<\/pre><\/div>\n<p>You can also use modes and vendor prefixes inside media query style map.<\/p>\n<p>For syntax help, see Garden's <a href=\"https:\/\/github.com\/noprompt\/garden\/wiki\/Media-Queries\">documentation<\/a>.<\/p>\n<h2><a id=\"user-content-feature-queries\" class=\"anchor\" aria-hidden=\"true\" href=\"#feature-queries\"><\/a>Feature queries<\/h2>\n<p>Define how your style looks when certain CSS features are supported by the browser:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def grid-style {;; Default style uses Flexbox as fallback\n                 :display \"flex\"\n                 :flex-direction \"row\"\n                 :flex-wrap \"wrap\"\n                 ::stylefy\/media {{:max-width styles\/phone-width}\n                                  {:display \"block\"}}\n                 ;; Use CSS Grid style if it is supported by the browser.\n                 ;; If the browser does not support CSS Grid or feature queries at all, this\n                 ;; block is simply ignored.\n                 ::stylefy\/supports {\"display: grid\"\n                                     {:display \"grid\"\n                                      :grid-template-columns \"1fr 1fr 1fr\"\n                                      ;; Make CSS Grid responsive\n                                      ::stylefy\/media {{:max-width styles\/phone-width}\n                                                       {:grid-template-columns \"1fr\"}}}}})\n<\/pre><\/div>\n<p>You can use modes, media queries, and vendor prefixes inside feature query style map.<\/p>\n<h2><a id=\"user-content-3rd-party-classes\" class=\"anchor\" aria-hidden=\"true\" href=\"#3rd-party-classes\"><\/a>3rd party classes<\/h2>\n<p>Use 3rd party classes along with stylefy definitions:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn- bs-navbar-item [index index-atom text]\n  [:li (use-style styles\/clickable\n          {:class (when (= @index-atom index) \"active\")\n           :role \"presentation\"\n           :on-click #(reset! index-atom index)})\n   [:a text]])<\/pre><\/div>\n<p>Alternative syntax:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn- bs-navbar []\n  (let [active-index (r\/atom 0)]\n    (fn []\n      ;; Additional classes can also be attached in the name of the element\n      [:ul.nav.nav-pills (use-style styles\/boostrap-navbar-overrides)\n       [bs-navbar-item 0 active-index \"One\"]\n       [bs-navbar-item 1 active-index \"Two\"]\n       [bs-navbar-item 2 active-index \"Three\"]\n       [bs-navbar-item 3 active-index \"Four\"]])))<\/pre><\/div>\n<p>3rd party classes can also be attached directly into a style map. This means that the defined additional class names are always used with the style:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def boostrap-navbar {:background-color \"#DDDDDD\"\n                      ::stylefy\/with-classes [\"nav\" \"nav-pills\"]})\n                      \n(defn- bs-navbar []\n  (let [active-index (r\/atom 0)]\n    (fn []\n      [:ul (use-style boostrap-navbar)\n       [bs-navbar-item 0 active-index \"One\"]\n       [bs-navbar-item 1 active-index \"Two\"]\n       [bs-navbar-item 2 active-index \"Three\"]\n       [bs-navbar-item 3 active-index \"Four\"]])))<\/pre><\/div>\n<h2><a id=\"user-content-font-face\" class=\"anchor\" aria-hidden=\"true\" href=\"#font-face\"><\/a>Font-face<\/h2>\n<p>Call <strong>stylefy\/font-face<\/strong> and the given font-face is added into the DOM as CSS code asynchronously.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/font-face {:font-family \"open_sans\"\n                    :src \"url('..\/fonts\/OpenSans-Regular-webfont.woff') format('woff')\"\n                    :font-weight \"normal\"\n                    :font-style \"normal\"})<\/pre><\/div>\n<h2><a id=\"user-content-keyframes\" class=\"anchor\" aria-hidden=\"true\" href=\"#keyframes\"><\/a>Keyframes<\/h2>\n<p>Call <strong>stylefy\/keyframes<\/strong> and the given keyframes are added into the DOM as CSS code asynchronously.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/keyframes \"simple-animation\"\n                   [:from\n                    {:background-color \"red\"}]\n                   [:to\n                    {:background-color \"blue\"}])\n(def animated-box (merge simple-box\n                         {:animation-name \"simple-animation\"\n                          :animation-duration \"3s\"\n                          :animation-iteration-count \"infinite\"}))<\/pre><\/div>\n<h2><a id=\"user-content-custom-class-names\" class=\"anchor\" aria-hidden=\"true\" href=\"#custom-class-names\"><\/a>Custom class names<\/h2>\n<p>As has been told, stylefy converts style definition to unique CSS classes automatically and there is no need to worry about class names. It can, however, be useful to be able to generate custom named classes for example when working with 3rd party libraries \/ frameworks. For this purpose, call <strong>stylefy\/class<\/strong>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>;; This generates a CSS class with the name \"background-transition\" and adds it into the DOM.\n(stylefy\/class \"background-transition\"\n               {:transition \"background-color 1s\"})\n;; Use the generated class in a component like any other class\n[:div.background-transition]<\/pre><\/div>\n<h2><a id=\"user-content-custom-tag-styles\" class=\"anchor\" aria-hidden=\"true\" href=\"#custom-tag-styles\"><\/a>Custom tag styles<\/h2>\n<p>As has been told, stylefy converts style definition to unique CSS classes automatically and there is no need to worry about writing selectors for HTML tags. However, custom tag styles can be useful for setting styles on base elements, like html or body. For this purpose, call <strong>stylefy\/tag<\/strong>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>;; This generates a CSS tag selector and style for \"body\" element\n(def body-style\n  {:background-color :lightyellow\n   :padding          :5px})\n(stylefy\/tag \"body\" body-style)<\/pre><\/div>\n<h2><a id=\"user-content-style-caching\" class=\"anchor\" aria-hidden=\"true\" href=\"#style-caching\"><\/a>Style caching<\/h2>\n<p>stylefy supports style caching with HTML5 local storage. The converted CSS code is added into local storage and loaded from there when the page is reloaded.<\/p>\n<p>As from version 1.7.0, caching with local storage is turned on by default. You can turn it off in the initialisation function:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>  (stylefy\/init {:use-caching? false})<\/pre><\/div>\n<p>By default, the cache is cleared in seven days. You can also clear it manually by calling:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[stylefy.cache :as stylefy-cache])\n(stylefy-cache\/clear)<\/pre><\/div>\n<h3><a id=\"user-content-cache-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#cache-options\"><\/a>Cache options<\/h3>\n<p>Cache options support automatic cache clearing when a certain amount of time is passed.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/init {:use-caching? true\n               :cache-options {:expires (* 1 60 60 24 7)}}) ; Cache is cleared after 7 days<\/pre><\/div>\n<h2><a id=\"user-content-garden-units-and-colors\" class=\"anchor\" aria-hidden=\"true\" href=\"#garden-units-and-colors\"><\/a>Garden units and colors<\/h2>\n<p>You can use Garden's <a href=\"https:\/\/github.com\/noprompt\/garden\/wiki\/Units-%26-Arithmetic\">Unit<\/a> and <a href=\"https:\/\/github.com\/noprompt\/garden\/wiki\/Color\">Color<\/a> helpers with stylefy. You can also use Garden's <code>defcssfn<\/code> macro to create custom CSS functions (notice that <code>defcssfn<\/code> is a macro and needs <code>garden.core<\/code> dependency declaration).<\/p>\n<h2><a id=\"user-content-debugging-and-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#debugging-and-testing\"><\/a>Debugging and testing<\/h2>\n<p>If you want to test your user interface by examining CSS class names, stylefy's automatically generated class names can become a hassle. To make testing and debugging easier, you can use your own prefix in stylefy's automatically generated class names:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def my-style {:color \"red\"\n               ::stylefy\/class-prefix \"debugthis\"}<\/pre><\/div>\n<p>Notice that you need to turn custom prefixes on separately on the init function:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/init {:use-custom-class-prefix? true})<\/pre><\/div>\n<h2><a id=\"user-content-multi-instance-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#multi-instance-support\"><\/a>Multi-instance support<\/h2>\n<p>Running multiple apps using stylefy on the same web page is currently possible if every app is built separately (every app contains its own JS file). Every instance can use its own style tag, so use your app name as a suffix in the <strong>style<\/strong> tag id.<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;style id=\"_stylefy-constant-styles_myapp\"&gt;&lt;\/style&gt;\n&lt;style id=\"_stylefy-styles_myapp\"&gt;&lt;\/style&gt;<\/pre><\/div>\n<p>Then init stylefy with multi-instance support. Instance-id is a unique string (for example app name). Base-node is an optional base node for style tags (handy if you use web components).<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(stylefy\/init {:multi-instance {:base-node (dommy\/sel1 \"#myapp\")\n                                :instance-id \"myapp\"}})<\/pre><\/div>\n<h2><a id=\"user-content-manual-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#manual-mode\"><\/a>Manual mode<\/h2>\n<p>Manual mode can be used to style child elements with manually written CSS selectors using Garden syntax. It should be used only for <strong>corner cases<\/strong> in which complex CSS selectors are needed, or when you want to style some <strong>3rd party child components<\/strong> that do not take style props as parameters. The selector and the style written in manual mode will be scoped inside the element in which you use the style map with <strong>use-style<\/strong>. To avoid confusion, stylefy's special keywords do not work in manual mode. For the most part, it is recommended to use <strong>sub-styles<\/strong> for styling child elements.<\/p>\n<p>An example of such corner case is a situation in which we want to change the style of some child element when the parent element is being hovered:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def mobile-media-query {:max-width \"550px\"})\n(def hoverbox-style\n  {:width \"500px\"1\n   :height \"200px\"\n   :padding \"33px\"\n   :margin-bottom \"10px\"\n   :background-color \"#55AA55\"\n   ::stylefy\/sub-styles {:innerbox {:width \"100%\"\n                                    :height \"100%\"\n                                    :background-color \"#444444\"}}\n   ;; Change the background color of the child element when the parent element is being hovered.\n   ;; This is a corner case that stylefy cannot handle directly, so we use manual mode to resolve it.\n   ::stylefy\/manual [[:&amp;:hover [:.innerbox\n                                ;; Brighten by default\n                                {:background-color \"#999999\"}]]\n                     (at-media mobile-media-query [:&amp;:hover [:.innerbox\n                                                             ;; Darker on mobile\n                                                             {:background-color \"#666666\"}]])]\n   ::stylefy\/media {mobile-media-query\n                    {:width \"100%\"}}})\n(defn hoverbox []\n  [:div (use-style hoverbox-style)\n   [:div.innerbox (use-sub-style hoverbox-style :innerbox)]])<\/pre><\/div>\n<p>For syntax help, see Garden's <a href=\"https:\/\/github.com\/noprompt\/garden\/wiki\/Syntax\">documentation<\/a>.<\/p>\n<p>More examples available here: <a href=\"https:\/\/github.com\/Jarzka\/stylefy\/tree\/master\/examples\/src\/stylefy\/examples\">https:\/\/github.com\/Jarzka\/stylefy\/tree\/master\/examples\/src\/stylefy\/examples<\/a><\/p>\n<p>Here: <a href=\"https:\/\/github.com\/Jarzka\/stylefy\/releases\">https:\/\/github.com\/Jarzka\/stylefy\/releases<\/a><\/p>\n<ul><li>Need to namespace or unnamespace keywords in a map? Checkout my other library: <a href=\"https:\/\/github.com\/Jarzka\/namespacefy\">namespacefy<\/a><\/li>\n<li>If you also want to present SQL queries as Clojure data, checkout <a href=\"https:\/\/github.com\/tatut\/specql\">specql<\/a><\/li>\n<\/ul><\/article>","created_at":"2019-05-08T20:41:55+0000","updated_at":"2019-05-08T20:42:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":15,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/5528061?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":943,"title":"cssfx.dev","url":"https:\/\/cssfx.dev\/","created_at":"2019-05-08T20:21:31+0000","updated_at":"2019-05-08T20:21:31+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"cssfx.dev"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":942,"title":"Kenney \u2022 1-Bit Pack","url":"https:\/\/kenney.nl\/assets\/bit-pack","content":"<p>This package includes over 1,000 different tiles, characters, objects and items. It's perfect for roguelike and RPG in any setting <em>(fantasy, feudal Japan, modern, sci-fi etc.)<\/em> and even comes with tiles for UI, platformers and puzzle games. Perfect for prototyping!<\/p><p>Also includes a monochrome (black and white) version, plus versions with a transparent background.<\/p><p><strong>Additional samples:<\/strong><\/p><ul class=\"donate-list noselect\"><li>Support<\/li>\n\t\n\t<li>Donate<\/li>\n\t<li>or<\/li>\n\t<li>Patreon<\/li>\n<\/ul>","created_at":"2019-05-08T12:57:42+0000","updated_at":"2019-05-08T12:57:42+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"kenney.nl","preview_picture":"https:\/\/kenney.nl\/assets\/bit-pack\/preview.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["javascript","programming languages","parsing","grammars"],"is_public":false,"id":941,"title":"modernserf\/zebu","url":"https:\/\/github.com\/modernserf\/zebu","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><h2><a id=\"user-content-what-is-this\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-is-this\"><\/a>What is this?<\/h2>\n<p>Zebu is a JavaScript library for building <a href=\"http:\/\/staff.um.edu.mt\/afra1\/seminar\/little-languages.pdf\" rel=\"nofollow\">little languages<\/a> with <a href=\"http:\/\/2ality.com\/2016\/11\/computing-tag-functions.html\" rel=\"nofollow\">tagged template literals<\/a>. Here are some examples of little languages built with Zebu:<\/p>\n<h3><a id=\"user-content-ranges\" class=\"anchor\" aria-hidden=\"true\" href=\"#ranges\"><\/a><a href=\"https:\/\/github.com\/modernserf\/zebu\/blob\/master\/src\/examples\/range.mjs\">Ranges<\/a><\/h3>\n<div class=\"highlight highlight-source-js\"><pre>range`1,3 ... (10)` \/\/ =&gt; yields 1, 3, 5, 7, 9<\/pre><\/div>\n<h3><a id=\"user-content-data-expressions\" class=\"anchor\" aria-hidden=\"true\" href=\"#data-expressions\"><\/a><a href=\"https:\/\/github.com\/modernserf\/zebu\/blob\/master\/src\/examples\/data-expressions.mjs\">Data expressions<\/a><\/h3>\n<div class=\"highlight highlight-source-js\"><pre>const obj = { foo: { bar: 3 } }\ndx`.foo.bar`.replace(obj, 5) \/\/ =&gt; { foo: { bar: 5 } }<\/pre><\/div>\n<h3><a id=\"user-content-react-proptypes\" class=\"anchor\" aria-hidden=\"true\" href=\"#react-proptypes\"><\/a><a href=\"https:\/\/github.com\/modernserf\/zebu\/blob\/master\/src\/examples\/prop-types.mjs\">React PropTypes<\/a><\/h3>\n<div class=\"highlight highlight-source-js\"><pre>const types = propTypes`\n  className: string?\n  type: (\"select\" | \"datalist\")?\n  options: [{ id: string, label: string }]\n  value: string\n  onChange: func\n`\n\/* =&gt; {\n  className: PropTypes.string,\n  type: PropTypes.oneOf([\"select\", \"datalist\"]),\n  options: PropTypes.arrayOf([\n    PropTypes.shape({\n      id: PropTypes.string.isRequired,\n      label: PropTypes.string.isRequired,\n    }).isRequired\n  ]).isRequired,\n  value: PropTypes.string.isRequired,\n  onChange: PropTypes.func.isRequired,\n} *\/<\/pre><\/div>\n<h3><a id=\"user-content-matrix-math-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#matrix-math-operations\"><\/a><a href=\"https:\/\/github.com\/modernserf\/zebu\/blob\/master\/src\/examples\/matrix.mjs\">Matrix math operations<\/a><\/h3>\n<div class=\"highlight highlight-source-js\"><pre>const x = [\n  [7, 1],\n  [-2, 3],\n]\nmatrix`\n  [ 2 0 \n   -1 3 ] * ${x}\n`\n\/\/ =&gt; [[14, 2], [-13, 8]]<\/pre><\/div>\n<h3><a id=\"user-content-interactive-fiction\" class=\"anchor\" aria-hidden=\"true\" href=\"#interactive-fiction\"><\/a><a href=\"https:\/\/github.com\/modernserf\/zebu\/blob\/master\/src\/examples\/interactive-fiction.mjs\">Interactive fiction<\/a><\/h3>\n<div class=\"highlight highlight-source-js\"><pre>const leonAdventure = story`\n  === begin ===\n  \"You wake up, unfortunately. There is a treasure chest in the room. You do not know where it came from.\"\n  * \"open treasure chest\" -&gt; treasure\n  * \"pet cat\" -&gt; cat\n  * \"go to the bathroom\" -&gt; toilet\n  === cat ===\n  \"The cat purrs with approval at your touch.\"\n  * \"open treasure chest\" -&gt; treasure\n  * \"pet cat\" -&gt; cat\n  * \"go to the bathroom\" -&gt; toilet\n`<\/pre><\/div>\n<h3><a id=\"user-content-state-machines\" class=\"anchor\" aria-hidden=\"true\" href=\"#state-machines\"><\/a>State machines<\/h3>\n<div class=\"highlight highlight-source-js\"><pre>const traffic = machine`\n  initState: #green\n  states: #green | #yellow | #red\n  events: #timer\n  onTransition: ${(state) =&gt; console.log(state)}\n  #green  @ #timer -&gt; #yellow\n  #yellow @ #timer -&gt; #red\n  #red    @ #timer -&gt; #green\n`\ntraffic.start() \/\/ log { type: \"green\" }\ntraffic.send({ type: \"timer\" }) \/\/ log { type: \"yellow\" }<\/pre><\/div>\n<h3><a id=\"user-content-text-matching\" class=\"anchor\" aria-hidden=\"true\" href=\"#text-matching\"><\/a>Text matching<\/h3>\n<div class=\"highlight highlight-source-js\"><pre>const joinObjects = (objects) =&gt; objects.reduce((l, r) =&gt; Object.assign(l, r), {})\nconst url = text`\n  URL       = Protocol Host Path? \"\/\"? Search? Anchor?\n              : ${(protocol, host, path, _, search, anchor) =&gt; ({ protocol, host, path, search, anchor })}\n  Protocol  = ${\/[a-z]+\/} \":\/\/\"\n  Host      = ${\/[A-Za-z0-9-]+\/} ++ \".\"\n  Path      = \"\/\" (Component ++ \"\/\")  : ${(_, path) =&gt; path}\n  Search    = \"?\" Pair ++ \"&amp;\"         : ${(_, pairs) =&gt; joinObjects(pairs)}\n  Pair      = Component \"=\" Component : ${(key, _, value) =&gt; ({[key]: value})}\n  Anchor    = \"#\" Component           : ${(_, target) =&gt; target}\n  Component = ${\/[A-Za-z0-9()_\\-~]\/}  : ${decodeURIComponent}\n`\nurl.match(\"https:\/\/github.com\/modernserf\/zebu?foo=bar20baz\"\/)\n\/* =&gt; { \n  ok: true, \n  value: {\n    protocol: \"https\",\n    host: [\"github\", \"com\"],\n    path: [\"modernserf\", \"zebu\"],\n    search: { foo: \"bar baz\" },\n    anchor: null,\n  },\n} *\/<\/pre><\/div>\n<h2><a id=\"user-content-installation--requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation--requirements\"><\/a>Installation &amp; Requirements<\/h2>\n<pre>npm install zebu\n<\/pre>\n<p>Zebu uses ES2018 features -- it works in node v10 and in modern browsers via <code>&lt;script type=\"module\"&gt;<\/code>, but will likely need to be transpiled with Babel for production use.<\/p>\n<h2><a id=\"user-content-writing-a-language\" class=\"anchor\" aria-hidden=\"true\" href=\"#writing-a-language\"><\/a>Writing a language<\/h2>\n<p>Zebu exports the function <code>grammar<\/code> that is used to define the rules for interpreting a language. Let's use <code>grammar<\/code> to make a yaml-like configuration language, called \"spaml\":<\/p>\n<div class=\"highlight highlight-source-js\"><pre>import { grammar } from \"zebu\"\nconst spaml = grammar`\n  Block = Pair ** Sep           : ${fromPairs}\n  Pair  = Key (\":\" line?) Expr  : ${(key, _, value) =&gt; [key, value]}\n  Expr  = #[ Expr ** Sep ]      : ${(xs = []) =&gt; xs}\n        | #{ Block }\n        | value\n        | \"true\"                : ${() =&gt; true}\n        | \"false\"               : ${() =&gt; false}\n        | \"null\"                : ${() =&gt; null}\n  Key   = identifier | value\n  Sep   = line | \",\"\n`\nfunction fromPairs (pairs) {\n  const obj = {}\n  for (const [key, value] of pairs) {\n    obj[key] = value\n  }\n  return obj\n}<\/pre><\/div>\n<p>You can use the <code>spaml<\/code> grammar like this:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const justin = spaml`\n  name: \"Justin\"\n  twitter_handle: \"modernserf\"\n  hobbies: [\"karaoke\", \"mixology\", \"programming\"]\n`<\/pre><\/div>\n<p>which results in:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>{ \n  name: \"Justin\", \n  twitter_handle: \"modernserf\", \n  hobbies: [\"karaoke\", \"mixology\", \"programming\"],\n}<\/pre><\/div>\n<h3><a id=\"user-content-tokenizing\" class=\"anchor\" aria-hidden=\"true\" href=\"#tokenizing\"><\/a>Tokenizing<\/h3>\n<p>So how does this work? <code>grammar<\/code> and <code>spaml<\/code> are both functions that work with tagged template literals. First, the string components and interpolated values are transformed into tokens. This text:<\/p>\n<pre>  name: \"Justin\"\n  twitter_handle: \"modernserf\"\n  hobbies: [\"karaoke\", \"mixology\", \"programming\"]\n<\/pre>\n<p>is transformed into:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>[\n  { type: 'identifier', value: 'name' },\n  { type: 'operator', value: ':' },\n  { type: 'value', value: 'Justin' },\n  { type: 'line' },\n  { type: 'identifier', value: 'twitter_handle' },\n  { type: 'operator', value: ':' },\n  { type: 'value', value: 'modernserf' },\n  { type: 'line' },\n  { type: 'identifier', value: 'hobbies' },\n  { type: 'operator', value: ':' },\n  { type: 'startToken', value: '[' },\n  { type: 'value', value: 'karaoke' },\n  { type: 'operator', value: ',' },\n  { type: 'value', value: 'mixology' },\n  { type: 'operator', value: ',' },\n  { type: 'value', value: 'programming' },\n  { type: 'endToken', value: ']' },\n]<\/pre><\/div>\n<p>Some text, like tabs and spaces, are removed altogether. JS-style comments (both <code>\/\/<\/code> line comments and <code>\/* *\/<\/code> block) are also removed. If there are multiple newlines in a row, they are consolidated into a single <code>line<\/code> token; any opening and closing lines are removed as well.<\/p>\n<p>Numbers and quoted strings become <code>value<\/code> tokens; all interpolated values are inserted as <code>value<\/code> tokens as well. Words that match JavaScript's rules for identifiers become <code>identifier<\/code> tokens. Most punctuation becomes <code>operator<\/code> tokens.<\/p>\n<p>Grouping symbols <code>( ) [ ] { }<\/code> are specially handled and become <code>startToken<\/code> and <code>endToken<\/code>, and these have a special purpose in the next stage of processing.<\/p>\n<h3><a id=\"user-content-skeleton-syntax-trees\" class=\"anchor\" aria-hidden=\"true\" href=\"#skeleton-syntax-trees\"><\/a>Skeleton syntax trees<\/h3>\n<p>Next, the grouping tokens are matched, and the tokens between them are collected into a single <code>structure<\/code> token:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>[\n  { type: 'identifier', value: 'name' },\n  { type: 'operator', value: ':' },\n  { type: 'value', value: 'Justin' },\n  { type: 'line' },\n  { type: 'identifier', value: 'twitter_handle' },\n  { type: 'operator', value: ':' },\n  { type: 'value', value: 'modernserf' },\n  { type: 'line' },\n  { type: 'identifier', value: 'hobbies' },\n  { type: 'operator', value: ':' },\n  { type: 'structure', structureType: '[]', value: [\n    { type: 'value', value: 'karaoke' },\n    { type: 'operator', value: ',' },\n    { type: 'value', value: 'mixology' },\n    { type: 'operator', value: ',' },\n    { type: 'value', value: 'programming' },\n  ] }\n]<\/pre><\/div>\n<p>This step allows Zebu to target <a href=\"https:\/\/en.wikipedia.org\/wiki\/Nested_word\" rel=\"nofollow\">visibly pushdown languages<\/a> -- essentially this means that Zebu grammars can be as (computationally) simple as regular expressions, but still allow recursive structures in brackets. This also makes it easier to provide good error messages &amp; avoid some performance issues. This technique was adapted from <a href=\"https:\/\/github.com\/ianh\/owl\">Owl<\/a> and Dylan's <a href=\"https:\/\/people.csail.mit.edu\/jrb\/Projects\/dexprs.pdf\" rel=\"nofollow\">D-Expressions<\/a>.<\/p>\n<h3><a id=\"user-content-parsing\" class=\"anchor\" aria-hidden=\"true\" href=\"#parsing\"><\/a>Parsing<\/h3>\n<p>In the next step, we try to match the tokens to the top rule of the grammar.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const spaml = grammar`\n  Block = Pair ** Sep           : ${fromPairs}\n  Pair  = Key (\":\" line?) Expr  : ${(key, _, value) =&gt; [key, value]}\n  Expr  = #[ Expr ** Sep ]      : ${(xs = []) =&gt; xs}\n        | #{ Block }\n        | value\n        | \"true\"                : ${() =&gt; true}\n        | \"false\"               : ${() =&gt; false}\n        | \"null\"                : ${() =&gt; null}\n  Key   = identifier | value\n  Sep   = line | \",\"\n`<\/pre><\/div>\n<p>These parsing expressions match a single token:<\/p>\n<ul><li><code>line<\/code>, <code>value<\/code>, <code>operator<\/code>, <code>identifier<\/code> - match a token of this type<\/li>\n<li><code>\"include\"<\/code>, <code>\"+\"<\/code> - match an operator or identifier token with this value<\/li>\n<\/ul><p>These parsing expressions work similarly to regular expressions, and can refer to the rules defined below them:<\/p>\n<ul><li><code>expr1 expr2<\/code> - matches expr1 followed by <code>expr2<\/code>, returning the result of <code>expr2<\/code><\/li>\n<li><code>expr1 expr2 : ${func}<\/code>  matches <code>expr1<\/code> followed by <code>expr2<\/code>, returning <code>func(expr1, expr2)<\/code><\/li>\n<li><code>expr1 | expr1<\/code> - try matching <code>expr1<\/code>, else match <code>expr2<\/code><\/li>\n<li><code>expr+<\/code> - match one or more expr, returning a list of <code>expr<\/code> results<\/li>\n<li><code>expr*<\/code> - match zero or more expr, returning a list of <code>expr<\/code> results<\/li>\n<li><code>expr?<\/code> - match zero or one expr, returning null or <code>expr<\/code> result<\/li>\n<\/ul><p>These parsing expressions are useful for matching simple operator expressions, or lists with separators, and can refer to the rules defined below them. They both return lists of <code>expr<\/code> results.<\/p>\n<ul><li><code>expr ++ separator<\/code> - match one or more <code>expr<\/code> separated by <code>separator<\/code><\/li>\n<li><code>expr ** separator<\/code> - match zero or more <code>expr<\/code> separated by <code>separator<\/code>, with an optional trailing <code>separator<\/code><\/li>\n<\/ul><p>These parsing expressions match expressions wrapped in bracketing punctuation. Unlike the other parsing expressions, these can also refer to rules defined <em>above<\/em> them:<\/p>\n<ul><li><code>#( expr )<\/code> match <code>expr<\/code> wrapped in parentheses<\/li>\n<li><code>#[ expr ]<\/code> match <code>expr<\/code> wrapped in square brackets<\/li>\n<li><code>#{ expr }<\/code> match <code>expr<\/code> wrapped in curly braces<\/li>\n<\/ul><h3><a id=\"user-content-combining-grammars\" class=\"anchor\" aria-hidden=\"true\" href=\"#combining-grammars\"><\/a>Combining grammars<\/h3>\n<p>You can include grammars in each other with <code>include<\/code>:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const commaSeparatedList = (parent) =&gt; grammar`#[ ${parent.Expr} ** \",\" ]`\nconst lang = grammar`\n  Expr  = #(value)\n        | include ${commaSeparatedList}\n        | value\n`<\/pre><\/div>\n<h3><a id=\"user-content-operator-grammars\" class=\"anchor\" aria-hidden=\"true\" href=\"#operator-grammars\"><\/a>Operator grammars<\/h3>\n<p>Operator precedence is rather tedious to implement in this grammar, so zebu also includes a separate mini-language just for defining operators. Operators in the same clause have the same precedence level; clauses nearer the top of the list are higher in precedence.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const math = op`\n  left  \"+\"   : ${(l, r) =&gt; l + r}\n        \"-\"   : ${(l, r) =&gt; l - r}\n  left  \"*\"   : ${(l, r) =&gt; l * r}\n        \"\/\"   : ${(l, r) =&gt; l \/ r}\n        \"%\"   : ${(l, r) =&gt; l % r}\n  right \"**\"  : ${(l, r) =&gt; l ** r}\n  pre   \"-\"   : ${x =&gt; -x}\n  post  \"++\"  : ${x =&gt; x + 1}\n        \"--\"  : ${x =&gt; x - 1}\n`\nmath`3 * 4 + 5 \/ 6`\n\/\/ parses as (3 * 4) + (5 \/ 6)<\/pre><\/div>\n<p>You can <code>include<\/code> operator grammars into \"regular\" grammars:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const expr = grammar`\n  Expr = include ${parent =&gt; op`\n    left \"++\" : ${(xs, ys) =&gt; xs.concat(ys)}\n    root ${parent.RootExpr}\n  `}\n  RootExpr  = #[ Expr ** \",\" ]\n            | value\n`\nexpr`[\"foo\", \"bar\"] ++ [\"baz\"]`\n\/\/ =&gt; ['foo', 'bar', 'baz']<\/pre><\/div>\n<h2><a id=\"user-content-future-work\" class=\"anchor\" aria-hidden=\"true\" href=\"#future-work\"><\/a>Future work<\/h2>\n<p><strong>Better Documentation.<\/strong> It took me a long time to understand how parsing works, and I suspect that Zebu's documentation is quite opaque to someone who is new to parsing. How can Zebu be a good introduction to parsing?<\/p>\n<p><strong><a href=\"https:\/\/github.com\/kentcdodds\/babel-plugin-macros\">Babel macro<\/a> support.<\/strong> Zebu is designed such that most of the parsing, both in grammars and in the languages they create, can be done at compile time. Zebu doesn't currently take advantage of this, but it would be considerably more useful if it did.<\/p>\n<p><strong>Interactive parsing workbench.<\/strong> <a href=\"https:\/\/omrelli.ug\/nearley-playground\/\" rel=\"nofollow\">Nearley<\/a> and <a href=\"https:\/\/ianh.github.io\/owl\/try\/#example\" rel=\"nofollow\">Owl<\/a> both have interactive browser examples that were immensely helpful for understanding how they work. Zebu should also support online\/interactive examples.<\/p>\n<p><strong>Better error messages.<\/strong> Parsing tools are notorious for inscrutable error messages. Can Zebu do better?<\/p>\n<\/article>","created_at":"2019-05-07T17:38:20+0000","updated_at":"2019-05-07T17:38:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/938722?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":940,"title":"in which a path is charted through the coming apocalypse","url":"https:\/\/technomancy.us\/184","content":"<div id=\"logo\"><a href=\"https:\/\/technomancy.us\/list\"><img src=\"https:\/\/technomancy.us\/i\/technomancy.png\" alt=\"Technomancy\" \/><\/a><\/div><p>I've long counted myself among the grumpy old-timers who\n  grudgingly accept the shift towards web-based-everything and just\n  try to make the most of it, wistfully remembering the days when I\n  could just do everything from within Emacs. One of my core\n  survival strategies in this web-first world has been to trick my\n  browser into at least having the decency to pretend to be Emacs. I\n  accomplished this in Firefox<sup><a href=\"#fn1\">1<\/a><\/sup> with\n  the <a href=\"https:\/\/github.com\/mooz\/keysnail\/wiki\">Keysnail<\/a>\n  extension. Keysnail has remarkable flexibility in how it overrides\n  Firefox's default key bindings to match those of Emacs, and\n  everything has been more or less great.<\/p><p>Unfortunately, <a href=\"https:\/\/blog.mozilla.org\/addons\/2016\/11\/23\/add-ons-in-2017\/\">a\n  soon-to-be-released update to Firefox<\/a> will remove the\n  extension\n  mechanism <a href=\"https:\/\/github.com\/mooz\/keysnail\/issues\/222\">used\n  by Keysnail<\/a>.<\/p><img src=\"https:\/\/technomancy.us\/i\/green-lake-laptop.jpg\" alt=\"laptop at Green Lake\" \/><p>I have felt very conflicted about this, because the old state of\n  affairs is admittedly untenable. Firefox currently\n  uses <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Mozilla\/Gecko\">Gecko<\/a>,\n  a decades-old rendering engine written in C++, and like much\n  software written in C++ it has a pretty distressing security track\n  record. Version 57 of Firefox replaces parts of Gecko with\n  functionality from <a href=\"https:\/\/servo.org\">Servo<\/a>, a\n  browser engine implemented in the Rust programming language. Most\n  of the bugs in Gecko which have led to embarrassing security flaws\n  are simply impossible in Servo. The fact\n  that so much safety-critical code is still being written in C++\n  and similar languages is a sad state of affairs, and we should\n  celebrate changes that mean end users will no longer bear the penalty\n  for programmers' reluctance to move beyond the technology of the\n  1980s.<\/p><p>But on the other hand, losing the ability to shape your computing\n  environment to your whims is <em>awful<\/em>. I lost track of how\n  many times (when using Chromium or other keysnail-less browsers)\n  I've wanted to throw my laptop out the window when I held\n  down <kbd>ctrl-n<\/kbd> to scroll down and it opened seventeen new\n  windows instead. I can't remember ever wanting to open a new browser\n  window in the past <em>decade<\/em>; why should I be stuck with a\n  key bound to that command and no way to disable it?<\/p><p>Of course, the new Firefox will still have an extension\n  mechanism, but it's a pale shadow of the old one. Citing the\n  flimsy<sup><a href=\"#fn2\">2<\/a><\/sup> excuse\n  of <a href=\"https:\/\/github.com\/lusakasa\/saka-key\/issues\/53#issuecomment-332319791\">security<\/a>,\n  key bindings like <kbd>C-n<\/kbd> are hard-coded into the browser\n  and forbidden from being overridden.<\/p><img src=\"https:\/\/technomancy.us\/i\/tumwater.jpg\" alt=\"Tumwater Falls\" style=\"float: left;\" \/><p>Things were looking bleak for me, and I contemplated whether I\n  would switch to curl or just give up software development\n  altogether for a career in goat-herding. I ended up finding a\n  solution from a most unlikely place.<\/p><h4>EXWM saves the day<\/h4><p>I had heard of <a href=\"https:\/\/github.com\/ch11ng\/exwm\">EXWM<\/a>\n  a while ago, and it struck me as a quixotic curiosity. The X\n  Window System uses a network socket for its control protocol,\n  allowing a lot of flexibility including native forwarding of\n  interfaces for remote programs. The developer of EXWM had taken an\n  XML description of the specification for the network protocol and\n  written a compiler to turn it\n  into <a href=\"https:\/\/github.com\/ch11ng\/xelb\/blob\/master\/xelb.el#L36\">a\n  library of Emacs Lisp functions<\/a> which he then used to\n  implement a window manager in pure Emacs Lisp. While I admired the\n  chutzpah this must have taken, I assumed it was a novelty that\n  could never be practical.<\/p><p>Eventually the Firefox conundrum prompted me to give it a second\n  look due to a feature called Simulation\n  Keys. The <code>exwm-input-set-simulation-keys<\/code> function allows\n  you to define a translation mapping so that a certain key\n  combination will be intercepted by EXWM when a non-Emacs program\n  has focus, and a different set of key input events will be sent\n  instead. It seemed too good to be true; I could let go of Keysnail\n  and instead get the same features applied to every program I\n  use<sup><a href=\"#fn3\">3<\/a><\/sup>.<\/p><p>I'm happy to report that EXWM does actually function startlingly\n  well as a window manager. The simulation keys feature is amazing\n  and puts my Firefox-related fears at ease, and having all\n  configuration written in a single language simplifies my setup\n  dramatically. Every X window you launch is given an Emacs buffer,\n  and all your normal splits and window resizing commands work\n  great with it. With the tiling window managers I used in the\n  past, it was so unusual for me to use something other than the\n  \"one fullscreen window per display\" setup that I would often\n  forget the key bindings for splitting and rearranging\n  windows. EXWM even integrates \"system tray\" programs into the Emacs\n  echo area, so your wifi connect tool shows up unobtrusively in\n  the bottom right corner.<\/p><p>There are a handful of gotchas. Emacs Lisp lacks general-purpose\n  concurrency features, but it does allow for concurrency when\n  dealing with subprocesses and network communication. Most\n  well-written Emacs Lisp will never block the main event loop,\n  which is good because when using EXWM that means the entire window\n  manager is stuck until the blocking operation completes. I only\n  came across two exceptions to this rule. One of them\n  is <code>smtpmail-send-it<\/code>, which can be replaced by\n  the <a href=\"https:\/\/github.com\/jwiegley\/emacs-async\/blob\/master\/smtpmail-async.el\">smtpmail-async<\/a>\n  library. The other is the <code>racket-run<\/code> command, which I was\n  able\n  to <a href=\"https:\/\/github.com\/greghendershott\/racket-mode\/pull\/282\">patch\n  in about an hour<\/a> to remove the blocking\n  call<sup><a href=\"#fn4\">4<\/a><\/sup>.<\/p><p>Other folks might run into more problems if they use other\n  third-party libraries which don't take care to use the network\n  functions properly. But for my use<sup><a href=\"#fn5\">5<\/a><\/sup>,\n  it's been very smooth, and I'm thrilled to have it.<\/p><p><b>Update<\/b>: I've\n  started <a href=\"http:\/\/p.hagelb.org\/exwm-ff-tabs.html\">configuring my\n  browser to open everything in new windows instead of new tabs<\/a>,\n  which sounds crazy, but is very useful, because it means that you\n  can use Emacs's built-in buffer switching tools to change tabs,\n  which are much better than anything I've seen inside a browser.<\/p><hr \/><div class=\"footnotes\"><p>[<a name=\"fn1\" id=\"fn1\">1<\/a>] I used <a href=\"http:\/\/conkeror\">Conkeror<\/a>\n  for several years, but eventually things got to the point where\n  browsing without <a href=\"https:\/\/noscript.net\">Noscript<\/a>\n  became untenable, and I could never get the two to work well together.<\/p><p>[<a name=\"fn2\" id=\"fn2\">2<\/a>] The rationale of \"it's for security\" would\n  stand up to a little more scrutiny if it weren't for the fact that\n  extensions <em>can<\/em> rebind <kbd>C-t<\/kbd>, a key which is used\n  hundreds if not thousands of times more often than <kbd>C-n<\/kbd>.<\/p><p>[<a name=\"fn3\" id=\"fn3\">3<\/a>]\n  Granted <a href=\"http:\/\/www.gnumeric.org\/\">gnumeric<\/a> is the\n  only program I use outside the browser and Emacs, but it's still\n  greatly appreciated. I also use\n  the <a href=\"https:\/\/key.saka.io\/\">Saka Key<\/a> extension, which\n  implements Keysnail's ability to trigger links from the keyboard\n  even if they don't have text attached to them.<\/p><p>[<a name=\"fn4\" id=\"fn4\">4<\/a>] I feel that the increasing \"Emacs needs\n  concurrency!\" calls tend to overstate the problem. Yes, of course\n  it would be nicer for the programmer to code using coroutines\n  (coming in Emacs 26!) instead of callbacks, but in the end this is\n  a convenience for the author, not for the end user.<\/p><p>[<a name=\"fn5\" id=\"fn5\">5<\/a>] <a href=\"https:\/\/github.com\/technomancy\/dotfiles\/blob\/master\/.emacs.d\/phil\/wm.el\">My\n    customizations<\/a> largely revolve around replacing\n    my <a href=\"https:\/\/github.com\/technomancy\/dotfiles\/blob\/master\/.xbindkeysrc.scm\">xbindkeys<\/a>\n    config with elisp, mapping workspace numbers to physical\n    displays, and some <code>eshell<\/code> commands to give one eshell\n    buffer per workspace. EXWM has XMonad-style workspaces where you\n    can change the workspace for each display independently rather\n    than forcing you to change them all at once like many more\n    conventional WMs, and I'm very glad it does.<\/p><\/div><p id=\"khomut\">\u0e5b<\/p>","created_at":"2019-05-06T03:16:57+0000","updated_at":"2019-05-06T03:16:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"technomancy.us","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","lisp","stumpwm"],"is_public":false,"id":939,"title":"index.html","url":"https:\/\/www.kaashif.co.uk\/2015\/06\/28\/hacking-stumpwm-with-common-lisp\/index.html","created_at":"2019-05-04T15:51:10+0000","updated_at":"2019-05-04T15:51:17+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.kaashif.co.uk"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":938,"title":"Putting the I back in IDE: Towards a Github Explorer","url":"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/","content":"<p>Imagine a system for editing and reviewing code where:<\/p><ul><li>\n    <p>Every branch of every repo gets its own sandboxed directory. Your\nrevision history in each branch, including uncommitted stuff, is\npersisted, as are build artifacts. When you switch contexts, each\nproject is just as you left it.<\/p>\n  <\/li>\n  <li>\n    <p>Within your editor, you can pull up a global view of all your\nbranches, your outstanding pull requests, and the pull requests\nyou\u2019re assigned to review. It\u2019s one keystroke to view the summary\nfor a pull, and one more to start editing any of its files right\nthere under your cursor.<\/p>\n  <\/li>\n  <li>\n    <p>Code review happens entirely within the editor. You\u2019re fed a series\nof diffs: one keystroke to approve, one keystroke to start\nediting. Dive in, make your changes, leave comments for the author,\npush, and move on.<\/p>\n  <\/li>\n<\/ul><p>We\u2019ve actually developed this workflow at Jane Street, and it\u2019s been\nused daily by hundreds of engineers for about two years now. It feels\nlike what an <em>integrated<\/em> development environment is supposed to feel\nlike. Making software today is as much about collaboration as it is\nabout writing your own code. But most developers are forced to switch\nfrom their editor to a web browser to read someone else\u2019s code, and\nswitch back if they want to play with it themselves.<\/p><p>Code review that takes place in a browser isn\u2019t just more\ninconvenient, it\u2019s often shallower, too: to really understand a piece\nof code you have to build it, run it, and explore it with your own\nhands, in your own editor. The system we\u2019ve built at Jane Street is\ndesigned to make that level of engagement seamless.<\/p><p>The trouble is that it\u2019s pretty coupled to our somewhat uncommon\ntoolchain: we use Mercurial for revision control; instead of\nsubmitting pull requests to Github, we submit \u201cfeatures\u201d to\n<a href=\"https:\/\/www.janestreet.com\/tech-talks\/code-review\/\">Iron<\/a>, a code\nreview system that we wrote in OCaml; and to tie it all together, we\nuse a UI called \u201cFeature Explorer\u201d that we built on top of Emacs.<\/p><p>The goal of this post, then, is just to show off what the <em>result<\/em>\nlooks like \u2013 to show you what kind of workflow is possible \u2013 in the\nhope that it inspires similar tools in other ecosystems. The truth is\nthat after writing code at work this way for a while, you start to\nwish that you had something similar at home. But there\u2019s nothing quite\nlike it for Git and Github and editors like Vim or Textmate or VS\nCode.<\/p><h2 id=\"what-git-and-github-could-look-like-with-truly-deep-editor-integration\">What Git and Github could look like with truly deep editor integration<\/h2><p>When you walk around the office here, there\u2019s one window you\u2019ll see\nopen on every developer\u2019s screen:<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/todo.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/todo.png\" alt=\"Feature Explorer todo\" \/><\/a><\/p><p>This is your todo in Feature Explorer. At the very top are the\nfeatures you\u2019ve been assigned to review, and below that are the\nfeatures you own.<\/p><p>(Note that we call everything a \u201cfeature,\u201d but really they come in two\nflavors: features that are ready for review are the equivalent of pull\nrequests, and features that you\u2019re just privately hacking away on are\nthe equivalent of branches.)<\/p><p>It looks pretty dull but even here there\u2019s actually a lot of\npower. Notice how some of the features are indented a bit? That\u2019s\nbecause they belong to different repositories. To create a new repo,\nor a new branch within an existing repo, you need only move your\ncursor to the appropriate line and hit <code class=\"highlighter-rouge\">!c<\/code>. Navigating across repos\nis as cheap as navigating across branches.<\/p><p>When you drill down into a feature, you see a page just like the one\nyou see on Github for a pull request:<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/github.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/github.png\" alt=\"Github view of feature\" \/><\/a><\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/feature.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/feature.png\" alt=\"Feature Explorer view of feature\" \/><\/a><\/p><p><em>(What a Github pull request would look like if it were a Feature\nExplorer feature.)<\/em><\/p><p>Of course the difference is that here, that \u201cpage\u201d is just a buffer in\nyour editor. Which means you can hit Enter and, instead of getting a\nlist of inert diffs in your web browser, you get diffs that can take\nyou, in a single keystroke, to the very lines that were changed \u2013\nloaded right there in your editor.<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/drilldown.gif\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/drilldown.gif\" alt=\"Drilling down in Feature Explorer\" \/><\/a><\/p><p>What\u2019s more, the file you\u2019re looking at isn\u2019t some global copy \u2013 it\u2019s\nthe file <em>for that branch<\/em>. You didn\u2019t have to think about it but you\nare already in the feature\u2019s sandbox. Any changes you make are\nrelative to that feature\u2019s tip revision, and won\u2019t affect any other\nfeatures.<\/p><p>It\u2019s worth dwelling on that for a second. All you did was hit Enter a\nfew times, but it\u2019s as if you performed four steps at radically\ndifferent levels, all at once:<\/p><ol><li>\n    <p>You used a Github-esque code review tool to get an overall view\nof what the feature (the pull request) does, including the files it\nchanges.<\/p>\n  <\/li>\n  <li>\n    <p>You used a Git-esque revision control tool to check out (<code class=\"highlighter-rouge\">git\ncheckout<\/code>) the specific repo and branch for that pull request.<\/p>\n  <\/li>\n  <li>\n    <p>You used a file system to switch into the relevant directory\nwhere that repo &amp; branch live (<code class=\"highlighter-rouge\">cd ...<\/code>) so that you can compile,\nrun tests, manage revision history, etc., within a sandbox.<\/p>\n  <\/li>\n  <li>\n    <p>You used an editor to load the file you care about into a buffer\nwhere you can actually make changes, use code completion, and so\non.<\/p>\n  <\/li>\n<\/ol><p>These levels are blended so seamlessly that the distinctions between\nthem fall away. When you navigate from feature to feature, you don\u2019t\nhave to worry about where you are in your file system, or what revision\nyou\u2019re on. All you\u2019re thinking about is the code you\u2019re looking\nat. You get used to thinking of the world not in terms of repositories\nand branches and directories, but in terms of features, i.e., units of\nfunctionality.<\/p><p>This makes it almost trivial to switch contexts. Let\u2019s say I want to\nbring up a coworker\u2019s pull request on a completely different repo. I\nsplit my Emacs window, navigate to the global todo, cursor over to her\nfeature, Enter, Enter, and now I\u2019ve got a diff, or the underlying file\nitself, loaded in my buffer. I can make changes to the file and push\nthem, I can run the build against my coworker\u2019s changes, etc., all\nwithout affecting anything about <em>my<\/em> feature; everything I touch is\nsandboxed to the world of her pull request. When I\u2019m done, I close the\nwindow and get back to what I was doing. I literally never had to\nleave my editor.<\/p><p>You might think, hey, Atom has a\n<a href=\"https:\/\/github.atom.io\/\">slick new Github integration<\/a>! Isn\u2019t that\nthe same thing?<\/p><p>Atom\u2019s Github integration, as much an improvement as it is, is not a\nwhole lot more than a browser window that happens to be inside your\neditor. Navigating to a pull request inside of the Github window\ndoesn\u2019t change \u201cwhere you are\u201d on disk; you can\u2019t just navigate from\nrepo to repo and, right there in your editor, pull up a file in the\nexact state it\u2019s in in that pull request.<\/p><p>Once you <em>can<\/em> do that, you\u2019ll wish the functionality existed\neverywhere people write code.<\/p><h2 id=\"the-lifecycle-of-a-pull-request-in-feature-explorer\">The lifecycle of a pull request in Feature Explorer<\/h2><p>Let\u2019s go back to the main todo and create a new feature. This is like\ncreating a new branch in git and a pull request in Github at the same\ntime. Just hit <code class=\"highlighter-rouge\">!c<\/code> and give the feature a name, specifying which repo\nit\u2019s a part of.<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/create-feature.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/create-feature.png\" alt=\"Creating a feature\" \/><\/a><\/p><p>When you hit Enter, already Iron has created a \u201cworkspace\u201d for the new\nfeature. Under the hood, workspaces are managed using a Mercurial\nextension called ShareExtension; the git equivalent would be to have\nmultiple clones of the same repo on the same disk.<\/p><p>The way it works at Jane Street is that each user has a ~\/workspaces\ndirectory on their disk. In ~\/workspaces\/REPO\/+clone+\/ there\u2019s a full\nclone of the repo; then you do your work in\n~\/workspaces\/REPO\/BRANCH1\/+share+ and\n~\/workspaces\/REPO\/BRANCH2\/+share+\/ and so on.<\/p><p>Those +share+ directories have very small \/.hg directories compared to\nthe main +clone+ one, because the versions in the +share+ directories\nare mostly made up of pointers to the +clone+ directories.<\/p><p>Since each workspace is a literal directory on disk, just by <code class=\"highlighter-rouge\">cd<\/code>\u2018ing\nyou can go from one to the other and get a totally independent\nworkspace where you can create files, run the build, and so on.<\/p><p>Back in the main todo, you\u2019ll see the feature you just created in the\nlist of features you own.<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/new-feature.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/new-feature.png\" alt=\"New feature\" \/><\/a><\/p><p>There\u2019s a \u201cnext step\u201d column because a feature can be in a bunch of\ndifferent states:<\/p><ul><li>\n    <p>It might be currently under review. In this case, you\u2019ll see how\nmany reviewers are left.<\/p>\n  <\/li>\n  <li>\n    <p>It might be ready to release. All your code has been reviewed and\napproved: if you have the right permissions, just hit <code class=\"highlighter-rouge\">!rl<\/code> to\n\u201crelease\u201d (i.e., merge the pull into master).<\/p>\n  <\/li>\n  <li>\n    <p>It might be brand new, waiting for you to add code or enable\nreview. A feature where review hasn\u2019t been enabled is just like a\nbranch you\u2019re privately working on, rather than being a pull\nrequest. It isn\u2019t visible to other people. Get in there and start\nhacking.<\/p>\n  <\/li>\n<\/ul><p>Features that are red have a failing build, the yellow\nones are under review, the green ones are ready to release, and the\nwhite ones are works in progress.<\/p><p>To add files to your newly created pull request, just nav to its main\npage and use your editor\u2019s keybindings to create a new file.<\/p><p>When you do that, you\u2019ll notice that your root directory has changed\nto ~\/workspaces\/REPO\/BRANCH\/+share+\/ because you\u2019re in the sandbox.<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/workspace.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/workspace.png\" alt=\"Workspace\" \/><\/a><\/p><p>To \u201csubmit a pull request,\u201d you need only \u201cenable review,\u201d i.e., hit\n<code class=\"highlighter-rouge\">!e<\/code> on the feature\u2019s main page \u2013 suddenly, the feature will be\npublicly visible and will show up in other people\u2019s todos. What\u2019s neat\nabout this workflow is that it\u2019s just as easy to <em>disable<\/em> review, say\nif someone\u2019s comments in code review led to a major rethink of how to\napproach the feature. You can go quiet, hack hack hack, and re-enable\nonce you\u2019re ready to show your code to the world again.<\/p><h2 id=\"code-review\">Code Review<\/h2><p>We put code review at the top of your todo and color it yellow as a\nbit of a nudge. The idea is that someone else\u2019s code that\u2019s <em>almost<\/em>\nfully baked and <em>almost<\/em> ready to be released is basically a\nhair-on-fire priority compared to code that you\u2019ve just started\nwriting. It\u2019s worth interrupting you for.<\/p><p>But we want to get you in and out of code review as quickly as\npossible. Here\u2019s how it works. In Feature Explorer, doing review is a\nmatter of going to the global todo, hitting Enter on one of the\nfeatures assigned to you, and pressing <code class=\"highlighter-rouge\">r<\/code>. That\u2019ll bring up a screen\nlike this:<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/review.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/review.png\" alt=\"Code review summary\" \/><\/a><\/p><p>Each of those lines represents a patch for your review. Hit Enter\nagain to read it:<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/review-patch.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/review-patch.png\" alt=\"Code review patch\" \/><\/a><\/p><p>There\u2019s actually a lot of intelligence behind these screens \u2013 our\ncode review system, Iron,\n<a href=\"https:\/\/blog.janestreet.com\/code-review-that-isnt-boring\/\">uses a sophisticated algebra<\/a>\nto calculate which diffs to show a reviewer based on what they\u2019ve\nalready reviewed, what merges and rebases there\u2019ve been in the\nmeantime, and the like \u2013 but the point to emphasize here is just that\nit\u2019s all right there in your editor.<\/p><p>If you like what you see in a given patch, just hit <code class=\"highlighter-rouge\">!r<\/code> to approve\nit. If you don\u2019t understand it, or want to leave a comment, you can\nopen the file by pressing <code class=\"highlighter-rouge\">e<\/code>. The file will be opened in your editor\nin a feature-specific sandbox. Feature Explorer will even put your\ncursor on exactly the line affected by the patch.<\/p><p>At Jane Street, reviewers leave notes as literal comments in the\ncode. They\u2019re distinguished by the special designation \u201cCR\u201d at the\nbeginning of the comment. You can assign CRs, resolve them, etc., all\njust by editing the comment\u2019s text using a special (but simple)\nsyntax.<\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/cr1.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/cr1.png\" alt=\"CR step 1\" \/><\/a><\/p><p><a href=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/cr2.png\"><img src=\"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/cr2.png\" alt=\"CR step 2\" \/><\/a><\/p><p><em>(Code review conversations take place in the code itself using\nspecial \u201cCR\u201d comments. Pending CRs show up in the assigned user\u2019s\ntodo.)<\/em><\/p><p>This system might seem primitive compared to Github\u2019s in-browser\ncommenting UI, but it means that you can do code review entirely\nwithin your editor. And since Feature Explorer is so tightly\nintegrated with the code review system, it\u2019s trivial to bring other\npeople into the mix: if you assign someone a CR, the feature it\u2019s a\npart of will immediately show up in their todo.<\/p><h2 id=\"go-fork-and-prosper\">Go fork and prosper<\/h2><p>There\u2019s nothing special about Mercurial, Iron, or even Emacs that\nmakes this system possible. Any extensible editor \u2013 like Vim, Atom,\nVS Code, Textmate \u2013 could support the core UI. Deep integrations for\ngit already exist in most of these editors, and the Github API should\nbe powerful enough to support an Iron-like code review system.<\/p><p>The highest-impact parts are in some ways the simplest, for example,\nthe part that lets you navigate from a Github pull request down to a\nfile in a specific branch on your local disk. If you\u2019ve already got a\nclone of a repo, branching is cheap; creating sandboxed directories\nis cheap; and with a good enough Internet connection, listing pull\nrequests and the files they affect is cheap. (One of the things we\nfound with our system is that it didn\u2019t really click until we had a\nsub-100ms code review server.) The rest is mostly editor glue.<\/p><p>Our system works well for us, but as sometimes happens with a piece of\nproprietary software built over a long time, it really <em>only<\/em> works\nfor us. One hopes the idea is intriguing enough to encourage\ndevelopment elsewhere.<\/p>","created_at":"2019-05-04T14:31:12+0000","updated_at":"2019-05-04T14:31:12+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"blog.janestreet.com","preview_picture":"https:\/\/blog.janestreet.com\/putting-the-i-back-in-ide-towards-a-github-explorer\/postimage.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":937,"title":"Local-first software: You own your data, in spite of the cloud","url":"https:\/\/www.inkandswitch.com\/local-first.html","content":"<h2 id=\"motivation\">Motivation: collaboration and ownership<\/h2><p>It's amazing how easily we can collaborate online nowadays. We use Google Docs to collaborate on documents, spreadsheets and presentations; in Figma we work together on user interface designs; we communicate with colleagues using Slack; we track tasks in Trello; and so on. We depend on these and many other online services, e.g. for taking notes, planning projects or events, remembering contacts, and a whole raft of business uses.<\/p><aside class=\"move-up-6\"><p>We will call these services \u201ccloud apps,\u201d but you could just as well call them \u201c<a href=\"https:\/\/en.wikipedia.org\/wiki\/Software_as_a_service\">SaaS<\/a>\u201d or \u201cweb-based apps.\u201d What they have in common is that we typically access them through a web browser or through mobile apps, and that they store their data on a server.<\/p><\/aside><p>Today\u2019s cloud apps offer big benefits compared to earlier generations of software: seamless collaboration, and being able to access data from any device. As we run more and more of our lives and work through these cloud apps, they become more and more critical to us. The more time we invest in using one of these apps, the more valuable the data in it becomes to us.<\/p><p>However, in our research we have spoken to a lot of creative professionals, and in that process we have also learned about the downsides of cloud apps.<\/p><p>When you have put a lot of creative energy and effort into making something, you tend to have a deep emotional attachment to it. If you do creative work, this probably seems familiar. (When we say \u201ccreative work,\u201d we mean not just visual art, or music, or poetry \u2014 many other activities, such as explaining a technical topic, implementing an intricate algorithm, designing a user interface, or figuring out how to lead a team towards some goal are also creative efforts.)<\/p><aside class=\"move-up-2\"><p>Our research on software that supports the creative process is discussed further in our articles <a href=\"https:\/\/www.inkandswitch.com\/capstone-manuscript.html\">Capstone, a tablet for thinking<\/a> and <a href=\"https:\/\/medium.com\/@hirodusk\/the-ipad-as-a-fast-precise-tool-for-creativity-21384ea18659\">The iPad as a fast, precise tool for creativity<\/a>.<\/p><\/aside><p>In the process of performing that creative work, you typically produce files and data: documents, presentations, spreadsheets, code, notes, drawings, and so on. And you will want to keep that data: for reference and inspiration in the future, to include it in a portfolio, or simply to archive because you feel proud of it. It is important to <em>feel ownership<\/em> of that data, because the creative expression is something so personal.<\/p><p>Unfortunately, cloud apps are problematic in this regard. Although they let you access your data anywhere, all data access must go via the server, and you can only do the things that the server will let you do. In a sense, you don\u2019t have full ownership of that data \u2014 the cloud provider does.<\/p><aside><p>\u201cThere is no cloud, it\u2019s just someone else\u2019s computer.\u201d \u2013 <a href=\"http:\/\/www.chriswatterston.com\/blog\/my-there-is-no-cloud-sticker\">Chris Watterston<\/a><\/p><\/aside><p>Cloud apps are provided as a service; if the service is unavailable, you cannot use the software, and you can no longer access your data created with that software. If the service shuts down, even though you might be able to export your data, without the servers there is normally no way for you to continue running your own copy of that software. Thus, you are at the mercy of the company providing the service.<\/p><p>Before web apps came along, we had what we might call \u201cold-fashioned\u201d apps: programs running on your local computer, reading and writing files on the local disk. We still use a lot of applications of this type today: text editors and IDEs, Git and other version control systems, and many specialized software packages such as graphics applications or CAD software fall in this category.<\/p><aside class=\"move-up-1\"><p>The software we are talking about in this article are apps for creating documents or files (such as text, graphics, spreadsheets, CAD drawings, or music), or personal data repositories (such as notes, calendars, to-do lists, or password managers). We are not talking about implementing things like banking services, e-commerce, social networking, ride-sharing, or similar services, which are well served by centralized systems.<\/p><\/aside><p>In old-fashioned apps, the data lives in files on your local disk, so you have full agency and ownership of that data: you can do anything you like, including long-term archiving, making backups, manipulating the files using other programs, or deleting the files if you no longer want them. You don\u2019t need anybody\u2019s permission to access your files, since they are yours. You don\u2019t have to depend on servers operated by another company.<\/p><p>To sum up: the cloud gives us collaboration, but old-fashioned apps give us ownership. Can\u2019t we have the best of both worlds?<\/p><p>We would like both the convenient cross-device access and real-time collaboration provided by cloud apps, and also the personal ownership of your own data embodied by \u201cold-fashioned\u201d software.<\/p><h2 id=\"seven-ideals\">Seven ideals for local-first software<\/h2><p>We believe that data ownership and real-time collaboration are not at odds with each other. It is possible to create software that has all the advantages of cloud apps, while also allowing you to retain full ownership of the data, documents and files you create.<\/p><p>We call this type of software <strong>local-first software<\/strong>, since it prioritizes the use of local storage (the disk built into your computer) and local networks (such as your home WiFi) over servers in remote datacenters.<\/p><p>In cloud apps, the data on the server is treated as the primary, authoritative copy of the data; if a client has a copy of the data, it is merely a cache that is subordinate to the server. Any data modification must be sent to the server, otherwise it \u201cdidn\u2019t happen.\u201d In local-first applications we swap these roles: we treat the copy of the data on your local device \u2014 your laptop, tablet, or phone \u2014 as the primary copy. Servers still exist, but they hold secondary copies of your data in order to assist with access from multiple devices. As we shall see, this change in perspective has profound implications.<\/p><p>Here are seven ideals we would like to strive for in local-first software.<\/p><h3 id=\"no-spinners\">1. No spinners: your work at your fingertips<\/h3><p>Much of today\u2019s software <a href=\"https:\/\/danluu.com\/input-lag\/\">feels slower<\/a> than previous generations of software. Even though CPUs have become ever faster, there is often a perceptible delay between some user input (e.g. clicking a button, or hitting a key) and the corresponding result appearing on the display. In <a href=\"https:\/\/www.inkandswitch.com\/slow-software.html\">previous work<\/a> we measured the performance of modern software and analyzed why these delays occur.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/world-ping-times.png\" alt=\"Server-to-server round-trip times between various locations worldwide\" \/><figcaption>Server-to-server round-trip times between AWS datacenters in various locations worldwide. Data from: Peter Bailis, Aaron Davidson, Alan Fekete, et al.: \u201c<a href=\"http:\/\/arxiv.org\/pdf\/1302.0309.pdf\">Highly Available Transactions: Virtues and Limitations<\/a>,\u201d VLDB 2014.\n          <\/figcaption><\/figure><p>With cloud apps, since the primary copy of the data is on a server, all data modifications, and many data lookups, require a round-trip to a server. Depending on where you live, the server may well be located on another continent, so the speed of light places a limit on how fast the software can be.<\/p><aside class=\"move-up-5\"><p>The user interface may try to hide that latency by showing the operation as if it were complete, even though the request is still in progress \u2014 a pattern known as <a href=\"https:\/\/uxplanet.org\/optimistic-1000-34d9eefe4c05\">Optimistic UI<\/a> \u2014 but until the request is complete, there is always the possibility that it may fail (for example, due to an unstable Internet connection). Thus, an optimistic UI still sometimes exposes the latency of the network round-trip when an error occurs.<\/p><\/aside><p>Local-first software is different: because it keeps the primary copy of the data on the local device, there is never a need for the user to wait for a request to a server to complete. All operations can be handled by reading and writing files on the local disk, and data synchronization with other devices happens quietly in the background.<\/p><p>While this by itself does not guarantee that the software will be fast, we expect that local-first software has the potential to respond near-instantaneously to user input, never needing to show you a spinner while you wait, and allowing you to operate with your data at your fingertips.<\/p><h3 id=\"multi-device\">2. Your work is not trapped on one device<\/h3><p>Users today rely on several computing devices to do their work, and modern applications must support such workflows. For example, users may capture ideas on the go using their smartphone, organize and think through those ideas <a href=\"https:\/\/www.inkandswitch.com\/capstone-manuscript.html\">on a tablet<\/a>, and then type up the outcome as a document on their laptop.<\/p><p>This means that while local-first apps keep their data in local storage on each device, it is also necessary for that data to be synchronized across all of the devices on which a user does their work. Various data synchronization technologies exist, and we discuss them in detail in a <a href=\"#existing-models\">later section<\/a>.<\/p><p>Most cross-device sync services also store a copy of the data on a server, which provides a convenient off-site backup for the data. These solutions work quite well as long as each file is only edited by one person at a time. If several people edit the same file at the same time, conflicts may arise, which we discuss in the <a href=\"#collaboration\">section on collaboration<\/a>.<\/p><h3 id=\"offline\">3. The network is optional<\/h3><p>Personal mobile devices move through areas of varying network availability: unreliable coffee shop WiFi, while on a plane or on a train going through a tunnel, in an elevator or a parking garage. In developing countries or rural areas, infrastructure for Internet access is sometimes patchy. While traveling internationally, many mobile users disable cellular data due to the cost of roaming. Overall, there is plenty of need for offline-capable apps, such as for researchers or journalists who need to write while in the field.<\/p><p>\u201cOld-fashioned\u201d apps work fine without an Internet connection, but cloud apps typically don\u2019t work while offline. For several years the <a href=\"http:\/\/offlinefirst.org\/\">Offline First<\/a> movement has been encouraging developers of web and mobile apps to improve offline support, but in practice it has been difficult to retrofit offline support to cloud apps, because tools and libraries designed for a server-centric model do not easily adapt to situations in which users make edits while offline.<\/p><aside class=\"move-up-6\"><p>Although it is possible to make <a href=\"https:\/\/developers.google.com\/web\/fundamentals\/codelabs\/offline\/\">web apps work offline<\/a>, it can be difficult for a user to know whether all the necessary code and data for an application have been downloaded.<\/p><\/aside><p>Local-first applications keep the primary copy of the data in files in each device\u2019s local filesystem, so the user can read and write this data anytime, even while offline. The data is then synchronized with other devices sometime later, when a network connection is available. The data synchronization need not necessarily go via the Internet: local-first apps could also use Bluetooth or local WiFi to sync data to nearby devices.<\/p><p>Moreover, for good offline support it is desirable for the software to run as a locally installed executable on your device, rather than a tab in a web browser. For mobile apps it is already standard that the whole app is downloaded and installed before it is used.<\/p><h3 id=\"collaboration\">4. Seamless collaboration with your colleagues<\/h3><p>Collaboration typically requires that several people contribute material to a document or file. However, in old-fashioned software it is problematic for several people to work on the same file at the same time: the result is often a <em>conflict<\/em>. In text files such as source code, resolving conflicts is tedious and annoying, and the task quickly becomes very difficult or impossible for complex file formats such as spreadsheets or graphics documents. Hence, collaborators may have to agree up front who is going to edit a file, and only have one person at a time who may make changes.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/dropbox-conflict.png\" alt=\"Finder window showing conflicted files in Dropbox\" \/><figcaption>A <a href=\"https:\/\/help.dropbox.com\/syncing-uploads\/conflicted-copy\">\u201cconflicted copy\u201d<\/a> on Dropbox. The user must merge the changes manually.\n          <\/figcaption><\/figure><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/evernote-conflict.png\" alt=\"A note with conflicting edits in Evernote\" \/><figcaption>In Evernote, if a note is changed concurrently, it is moved to a <a href=\"https:\/\/discussion.evernote.com\/topic\/86113-how-do-i-resolve-sync-conflicts\/\">\u201cconflicting changes\u201d notebook<\/a>, and there is nothing to support the user in resolving the situation \u2014 not even a facility to compare the different versions of a note.\n          <\/figcaption><\/figure><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/merge-conflict.png\" alt=\"Resolving a Git merge conflict with DiffMerge\" \/><figcaption>In Git and other version control systems, several people may modify the same file in different commits. Combining those changes often results in <a href=\"https:\/\/git-scm.com\/book\/en\/v2\/Git-Branching-Basic-Branching-and-Merging#_basic_merge_conflicts\">merge conflicts<\/a>, which can be resolved using specialized tools (such as <a href=\"https:\/\/www.sourcegear.com\/diffmerge\/\">DiffMerge<\/a>, shown here). These tools are primarily designed for line-oriented text files such as source code; for other file formats, tool support is much weaker.\n          <\/figcaption><\/figure><p>On the other hand, cloud apps such as Google Docs have vastly simplified collaboration by allowing multiple users to edit a document simultaneously, without having to send files back and forth by email and without worrying about conflicts. Users have come to expect this kind of seamless real-time collaboration in a wide range of applications.<\/p><p>In local-first apps, our ideal is to support real-time collaboration that is on par with the best cloud apps today, or better. Achieving this goal is one of the biggest challenges in realizing local-first software, but we believe it is possible: in a <a href=\"#future\">later section<\/a> we discuss technologies that enable real-time collaboration in a local-first setting.<\/p><p>Moreover, we expect that local-first apps can support various workflows for collaboration. Besides having several people edit the same document in real-time, it is sometimes useful for one person to tentatively propose changes that can be reviewed and selectively applied by someone else. Google Docs supports this workflow with its <a href=\"https:\/\/support.google.com\/docs\/answer\/6033474\">suggesting mode<\/a>, and <a href=\"https:\/\/help.github.com\/en\/articles\/about-pull-requests\">pull requests<\/a> serve this purpose in Git.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/suggest-changes.png\" alt=\"Suggesting changes in Google Docs\" \/><figcaption>In Google Docs, collaborators can either edit the document directly, or they can <a href=\"https:\/\/support.google.com\/docs\/answer\/6033474\">suggest changes<\/a>, which can then be accepted or rejected by the document owner.\n          <\/figcaption><\/figure><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/pull-request.png\" alt=\"A pull request on GitHub\" \/><figcaption>The collaboration workflow on GitHub is based on <a href=\"https:\/\/help.github.com\/en\/articles\/about-pull-requests\">pull requests<\/a>. A user may change multiple source files in multiple commits, and submit them as a proposed change to a project. Other users may review and amend the pull request before it is finally merged or rejected.\n          <\/figcaption><\/figure><h3 id=\"longevity\">5. The Long Now<\/h3><p>An important aspect of data ownership is that you can continue accessing the data for a long time in the future. When you do some work with local-first software, your work should continue to be accessible indefinitely, even after the company that produced the software is gone.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/cuneiform.jpg\" alt=\"Cuneiform script on clay tablet, ca. 3000 BCE\" \/><figcaption>Cuneiform script on clay tablet, ca. 3000 BCE. Image from <a href=\"https:\/\/commons.wikimedia.org\/wiki\/File:Early_writing_tablet_recording_the_allocation_of_beer.jpg\">Wikimedia Commons<\/a><\/figcaption><\/figure><p>\u201cOld-fashioned\u201d apps continue to work forever, as long as you have a copy of the data and some way of running the software. Even if the software author goes bust, you can continue running the last released version of the software. Even if the operating system and the computer it runs on become obsolete, you can still run the software in a virtual machine or emulator. As storage media evolve over the decades, you can copy your files to new storage media and continue to access them.<\/p><aside class=\"move-up-8\"><p>The Internet Archive maintains a <a href=\"https:\/\/archive.org\/details\/softwarelibrary\">collection of historical software<\/a> that can be run using an emulator in a modern web browser; enthusiasts at the <a href=\"http:\/\/eab.abime.net\/\">English Amiga Board<\/a> share tips on running historical software.<\/p><\/aside><p>On the other hand, cloud apps depend on the service continuing to be available: if the service is unavailable, you cannot use the software, and you can no longer access your data created with that software. This means you are betting that the creators of the software will continue supporting it for a long time \u2014 at least as long as you care about the data.<\/p><aside><p><a href=\"https:\/\/ourincrediblejourney.tumblr.com\/\">Our incredible journey<\/a> is a blog that documents startup products getting shut down after an acquisition.<\/p><\/aside><p>Although there does not seem to be a great danger of Google shutting down Google Docs anytime soon, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Google_Reader\">popular products<\/a> do sometimes <a href=\"https:\/\/killedbygoogle.com\">get shut down<\/a> or <a href=\"https:\/\/www.independent.co.uk\/arts-entertainment\/music\/news\/myspace-songs-lost-website-move-migration-mp3-music-server-accounts-a8827881.html\">lose data<\/a>, so we know to be careful. And even with long-lived software there is the risk that the pricing or features change in a way you don\u2019t like, and with a cloud app, continuing to use the old version is not an option \u2014 you will be upgraded whether you like it or not.<\/p><p>Local-first software enables greater longevity because your data, and the software that is needed to read and modify your data, are all stored locally on your computer. We believe this is important not just for your own sake, but also for future historians who will want to read the documents we create today. Without longevity of our data, we risk creating <a href=\"https:\/\/www.bbc.co.uk\/news\/science-environment-31450389\">what Vint Cerf calls a \u201cdigital Dark Age<\/a>.\u201d<\/p><aside class=\"move-up-2\"><p>We have previously written about <a href=\"https:\/\/www.inkandswitch.com\/capstone-manuscript-appendices.html#c-web-page-archiving\">long-term archiving of web pages<\/a>. For an interesting discussion of long-term data preservation, see <a href=\"http:\/\/www.vpri.org\/pdf\/tr2015004_cuneiform.pdf\">\u201cThe Cuneiform Tablets of 2015\u201d<\/a>, a paper by Long Tien Nguyen and Alan Kay at <a href=\"https:\/\/2015.onward-conference.org\/track\/onward2015-essays\">Onward! 2015<\/a>.<\/p><\/aside><p>Some file formats (such as plain text, JPEG, and PDF) are so ubiquitous that they will probably be readable for centuries to come. The US Library of Congress also <a href=\"https:\/\/www.sqlite.org\/locrsf.html\">recommends XML, JSON, or SQLite<\/a> as archival formats for datasets. However, in order to read less common file formats and to preserve interactivity, you need to be able to run the original software (if necessary, in a virtual machine or emulator). Local-first software enables this.<\/p><h3 id=\"privacy\">6. Security and privacy by default<\/h3><p>One problem with the architecture of cloud apps is that they store all the data from all of their users in a centralized database. This large collection of data is an attractive target for attackers: a rogue employee, or a hacker who gains access to the company\u2019s servers, can read and tamper with all of your data. Such security breaches are sadly <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_data_breaches\">terrifyingly common<\/a>, and with cloud apps we are unfortunately at the mercy of the provider.<\/p><p>While Google has a world-class security team, the sad reality is that most companies do not. And while Google is good at defending your data against external attackers, the company internally is free to use your data in a myriad ways, such as feeding your data into its machine learning systems.<\/p><aside class=\"move-up-4\"><p>Quoting from the <a href=\"https:\/\/www.google.com\/drive\/terms-of-service\/\">Google Drive terms of service<\/a>: \u201cOur automated systems analyze your content to provide you personally relevant product features, such as customized search results, and spam and malware detection.\u201d<\/p><\/aside><p>Maybe you feel that your data would not be of interest to any attacker. However, for many professions, dealing with sensitive data is an important part of their work. For example, medical professionals handle sensitive patient data, investigative journalists handle confidential information from sources, governments and diplomatic representatives conduct sensitive negotiations, and so on. Many of these professionals cannot use cloud apps due to regulatory compliance and confidentiality obligations.<\/p><p>Local-first apps, on the other hand, have better privacy and security built in at the core. Your local devices store only your own data, avoiding the centralized cloud database holding everybody\u2019s data. Local-first apps can use <em>end-to-end encryption<\/em> so that any servers that store a copy of your files only hold encrypted data that they cannot read.<\/p><aside class=\"move-up-4\"><p>Modern messaging apps like <a href=\"https:\/\/www.usenix.org\/conference\/usenixsecurity16\/technical-sessions\/presentation\/garman\">iMessage<\/a>, <a href=\"https:\/\/www.whatsapp.com\/security\/WhatsApp-Security-Whitepaper.pdf\">WhatsApp<\/a> and <a href=\"https:\/\/www.signal.org\/docs\/\">Signal<\/a> already use end-to-end encryption, <a href=\"https:\/\/keybase.io\">Keybase<\/a> provides encrypted file sharing and messaging, and <a href=\"https:\/\/www.tarsnap.com\/crypto.html\">Tarsnap<\/a> takes this approach for backups. We hope to see this trend expand to other kinds of software as well.<\/p><\/aside><h3 id=\"user-control\">7. You retain ultimate control<\/h3><p>With cloud apps, anything you do is subject to the provider\u2019s terms of service. Most people never have any problem with this, but occasionally issues arise. For example, in October 2017, several Google Docs users were <a href=\"https:\/\/www.washingtonpost.com\/news\/the-switch\/wp\/2017\/10\/31\/a-mysterious-message-is-locking-google-docs-users-out-of-their-files\/\">locked out of their documents<\/a> because an automated system incorrectly flagged these documents as abusive. This incident served as a reminder that with cloud apps, the service provider, not you, has the ultimate say on what you are allowed to do.<\/p><p>In local-first apps, what you do is subject only to the laws of physics, not any terms of service.<\/p><p>If you are physically able to copy some data from one place to another, then you are allowed to do it. Since the data is stored on your own devices, there is nobody to stop you. You are free to do anything you like with your data: to modify it, to make backups, to write a program to process it, or to delete it if you no longer want it. You have full control.<\/p><aside class=\"move-up-6\"><p>Under the <a href=\"https:\/\/www.echr.coe.int\/Documents\/Convention_ENG.pdf\">European Convention on Human Rights<\/a>, your freedom of <em>thought<\/em> and <em>opinion<\/em> is <a href=\"https:\/\/www.echr.coe.int\/Documents\/Guide_Art_9_ENG.pdf\">unconditional<\/a> \u2014 the state may never interfere with it, since it is yours alone \u2014 whereas freedom of <em>expression<\/em> (including freedom of <em>speech<\/em>) can be restricted in certain ways, since it affects other people. Communication services like social networks convey expression, but the raw notes and unpublished work of a creative person are a way of developing thoughts and opinions, and thus <a href=\"http:\/\/data.parliament.uk\/writtenevidence\/committeeevidence.svc\/evidencedocument\/draft-investigatory-powers-bill-committee\/draft-investigatory-powers-bill\/written\/26353.pdf\">warrant unconditional protection<\/a>.<\/p><\/aside><h2 id=\"existing-models\">Existing data storage and sharing models<\/h2><p>We believe professional and creative users deserve software that realizes the local-first goals, helping them collaborate seamlessly while also allowing them to retain full ownership of their work. If we can give users these qualities in the software they use to do their most important work, we can help them be better at what they do, and potentially make a significant difference to many people\u2019s professional lives.<\/p><p>However, while the ideals of local-first software may resonate with you, you may still be wondering how achievable they are in practice. Are they just utopian thinking?<\/p><p>In the remainder of this article we discuss what it means to realize local-first software in practice. We look at a wide range of existing technologies and break down how well they satisfy the local-first ideals. In the following tables, \u2713 means the technology meets the ideal, \u2014 means it partially meets the ideal, and \u2717 means it does not meet the ideal.<\/p><p>As we shall see, many technologies satisfy some of the goals, but none are able to satisfy them all. Finally, we examine a technique from the cutting edge of computer science research that might be a foundational piece in realizing local-first software in the future.<\/p><h3 id=\"user-experience\">How application architecture affects user experience<\/h3><p>Let\u2019s start by examining software from the end user\u2019s perspective, and break down how well different software architectures meet the <a href=\"#seven-ideals\">seven goals<\/a> of local-first software. In the <a href=\"#dev-infrastructure\">next section<\/a> we compare storage technologies and APIs that are used by software engineers to build applications.<\/p><h4 id=\"files-attachments\">Files and email attachments<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Files + email attachments<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p>Viewed through the lens of our seven goals, traditional files have many desirable properties: they can be viewed and edited offline, they give full control to users, and they can readily be backed up and preserved for the long term. Software relying on local files also has the potential to be very fast.<\/p><p>However, accessing files from multiple devices is trickier. It is possible to transfer a file across devices using various technologies:<\/p><ul><li>Sending it back and forth by email;<\/li>\n            <li>Passing a USB drive back and forth;<\/li>\n            <li>Via a distributed file system such as a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Network-attached_storage\">NAS server<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Network_File_System\">NFS<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/File_Transfer_Protocol\">FTP<\/a>, or <a href=\"https:\/\/linux.die.net\/man\/1\/rsync\">rsync<\/a>;<\/li>\n            <li>Using a cloud file storage service like Dropbox, Google Drive, or OneDrive (see <a href=\"#dropbox\">later section<\/a>);<\/li>\n            <li>Using a version control system such as Git (see <a href=\"#git\">later section<\/a>).<\/li>\n        <\/ul><p>Of these, email attachments are probably the most common sharing mechanism, especially among users who are not technical experts. Attachments are easy to understand and trustworthy. Once you have a copy of a document, it does not spontaneously change: if you view an email six months later, the attachments are still there in their original form. Unlike a web app, an attachment can be opened without any additional login process.<\/p><p>The weakest point of email attachments is collaboration. Generally, only one person at a time can make changes to a file, otherwise a difficult manual merge is required. File versioning quickly becomes messy: a back-and-forth email thread with attachments often leads to filenames such as <code>Budget draft 2 (Jane's version) final final 3.xls<\/code>.<\/p><p>Nevertheless, for apps that want to incorporate local-first ideas, a good starting point is to offer an export feature that produces a widely-supported file format (e.g. plain text, PDF, PNG, or JPEG) and allows it to be shared e.g. via email attachment, Slack, or WhatsApp.<\/p><h4 id=\"web-apps\">Web apps: Google Docs, Trello, Figma, Pinterest, etc.<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Google Docs<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"okay\">\u2014<\/td>\n          <\/tr><tr><td>Trello<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n          <\/tr><tr><td>Pinterest<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p>At the opposite end of the spectrum are pure web apps, where the user\u2019s local software (web browser or mobile app) is a thin client and the data storage resides on a server. The server typically uses a large-scale database in which the data of millions of users are all mixed together in one giant collection.<\/p><p>Web apps have set the standard for real-time collaboration. As a user you can trust that when you open a document on any device, you are seeing the most current and up-to-date version. This is so overwhelmingly useful for team work that these applications have become dominant. Even traditionally local-only software like Microsoft Office is making the transition to cloud services, with <a href=\"http:\/\/fortune.com\/2017\/07\/20\/microsoft-office-365-earnings\/\">Office 365 eclipsing locally-installed Office as of 2017<\/a>.<\/p><p>With the rise of <a href=\"https:\/\/medium.com\/@anupamr\/distributed-teams-are-the-new-cloud-for-startups-14240a9822d7\">remote work and distributed teams<\/a>, real-time collaborative productivity tools are becoming even more important. Ten users on a team video call can bring up the same Trello board and each make edits on their own computer while simultaneously seeing what other users are doing.<\/p><p>The flip side to this is a total loss of ownership and control: the data on the server is what counts, and any data on your client device is unimportant \u2014 it is merely a cache. Most web apps have little or no support for offline working: if your network hiccups for even a moment, you are locked out of your work mid-sentence.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/gdocs-offline.png\" alt=\"Offline indicator in Google Docs\" \/><figcaption>If Google Docs detects that it is offline, it blocks editing of the document.\n          <\/figcaption><\/figure><p>A few of the best web apps <a href=\"https:\/\/uxplanet.org\/optimistic-1000-34d9eefe4c05\">hide the latency of server communication<\/a> using JavaScript, and try to provide limited offline support (for example, the <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/google-docs-offline\/ghbmnnjooekpmoecnnnilnnbdlolhkhi\">Google Docs offline plugin<\/a>). However, these efforts appear retrofitted to an application architecture that is fundamentally centered on synchronous interaction with a server. Users report mixed results when trying to work offline.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/gdocs-extension.png\" alt=\"A negative user review of the Google Docs offline extension\" \/><figcaption>A negative user review of the Google Docs offline extension.\n          <\/figcaption><\/figure><p>Some web apps, for example Milanote and Figma, offer installable desktop clients that are essentially repackaged web browsers. If you try to use these clients to access your work while your network is intermittent, while the vendor\u2019s servers are experiencing an outage, or after the vendor has been acquired and shut down, it becomes clear that your work was never truly yours.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/figma-offline.png\" alt=\"Offline error message in Figma\" \/><figcaption>The <a href=\"https:\/\/www.figma.com\">Figma<\/a> desktop client in action.\n          <\/figcaption><\/figure><h4 id=\"dropbox\">Dropbox, Google Drive, Box, OneDrive, etc.<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Dropbox<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p>Cloud-based file sync products like <a href=\"https:\/\/www.dropbox.com\/\">Dropbox<\/a>, <a href=\"https:\/\/www.google.com\/drive\/\">Google Drive<\/a>, <a href=\"https:\/\/www.box.com\/\">Box<\/a>, or <a href=\"https:\/\/onedrive.live.com\/\">OneDrive<\/a> make files available on multiple devices. On desktop operating systems (Windows, Linux, Mac OS) these tools work by watching a designated folder on the local file system. Any software on your computer can read and write files in this folder, and whenever a file is changed on one computer, it is automatically copied to all of your other computers.<\/p><p>As these tools use the local filesystem, they have many attractive properties: access to local files is fast, and working offline is no problem (files edited offline are synced the next time an Internet connection is available). If the sync service were shut down, your files would still remain unharmed on your local disk, and it would be easy to switch to a different syncing service. If your computer\u2019s hard drive fails, you can restore your work simply by installing the app and waiting for it to sync. This provides good longevity and control over your data.<\/p><p>However, on mobile platforms (iOS and Android), Dropbox and its cousins use a completely different model. The mobile apps do not synchronize an entire folder \u2014 instead, they are thin clients that fetch your data from a server one file at a time, and by default they do not work offline. There is a <a href=\"https:\/\/help.dropbox.com\/mobile\/access-files-offline\">\u201cMake available offline\u201d<\/a> option, but you need to remember to invoke it ahead of going offline, it is clumsy, and only works when the app is open. The <a href=\"https:\/\/www.dropbox.com\/developers\">Dropbox API<\/a> is also very server-centric.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/dropbox-mobile.jpg\" alt=\"The Dropbox mobile app showing a spinner while waiting to download a file\" \/><figcaption>Users of the Dropbox mobile app spend a lot of time looking at spinners, a stark contrast to the at-your-fingertips feeling of the Dropbox desktop product.\n          <\/figcaption><\/figure><p>The weakest point of file sync products is the lack of real-time collaboration: if the same file is edited on two different devices, the result is a conflict that needs to be merged manually, as discussed <a href=\"#collaboration\">previously<\/a>. The fact that these tools synchronize files in any format is both a strength (compatibility with any application) and a weakness (inability to perform format-specific merges).<\/p><h4 id=\"git\">Git and GitHub<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Git+GitHub<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p><a href=\"https:\/\/git-scm.com\">Git<\/a> and <a href=\"https:\/\/github.com\">GitHub<\/a> are primarily used by software engineers to collaborate on source code. They are perhaps the closest thing we have to a true local-first software package: compared to server-centric version control systems such as <a href=\"https:\/\/subversion.apache.org\">Subversion<\/a>, Git works fully offline, it is fast, it gives full control to users, and it is suitable for long-term preservation of data. This is the case because a Git repository on your local filesystem is a primary copy of the data, and is not subordinate to any server.<\/p><aside class=\"move-up-7\"><p>We focus on Git\/GitHub here as the most successful examples, but these lessons also apply to other distributed revision control tools like <a href=\"https:\/\/www.mercurial-scm.org\">Mercurial<\/a> or <a href=\"http:\/\/darcs.net\/\">Darcs<\/a>, and other repository hosting services such as <a href=\"https:\/\/about.gitlab.com\">GitLab<\/a> or <a href=\"https:\/\/bitbucket.org\">Bitbucket<\/a>. In principle it is possible to collaborate without a repository service, e.g. by <a href=\"https:\/\/drewdevault.com\/2018\/07\/02\/Email-driven-git.html\">sending patch files by email<\/a>, but the majority of Git users rely on GitHub.<\/p><\/aside><p>A repository hosting service like GitHub enables collaboration around Git repositories, accessing data from multiple devices, as well as providing a backup and archival location. Support for mobile devices is currently weak, although <a href=\"https:\/\/workingcopyapp.com\">Working Copy<\/a> is a promising Git client for iOS. GitHub stores repositories unencrypted; if stronger privacy is required, it is possible for you to run your own repository server.<\/p><p>We think the Git model points the way toward a future for local-first software. However, as it currently stands, Git has two major weaknesses:<\/p><ol><li>Git is excellent for asynchronous collaboration, especially using <a href=\"https:\/\/help.github.com\/en\/articles\/about-pull-requests\">pull requests<\/a>, which take a coarse-grained set of changes and allow them to be discussed and amended before merging them into the shared master branch. But Git has no capability for real-time, fine-grained collaboration, such as the automatic, instantaneous merging that occurs in tools like Google Docs, Trello, and Figma.<\/li>\n            <li>Git is highly optimized for code and similar line-based text files; other file formats are treated as binary blobs that cannot meaningfully be edited or merged. Despite GitHub\u2019s efforts to display and compare <a href=\"https:\/\/github.blog\/2011-03-21-behold-image-view-modes\/\">images<\/a>, <a href=\"https:\/\/github.blog\/2014-02-14-rendered-prose-diffs\/\">prose<\/a>, and <a href=\"https:\/\/github.blog\/2013-09-17-3d-file-diffs\/\">CAD files<\/a>, non-textual file formats remain second-class in Git.<\/li>\n        <\/ol><p>It\u2019s interesting to note that most software engineers have been reluctant to embrace cloud software for their editors, IDEs, runtime environments, and build tools. In theory, we might expect this demographic of sophisticated users to embrace newer technologies sooner than other types of users. But if you ask an engineer why they don\u2019t use a cloud-based editor like <a href=\"https:\/\/aws.amazon.com\/cloud9\/\">Cloud9<\/a> or <a href=\"https:\/\/repl.it\">Repl.it<\/a>, or a runtime environment like <a href=\"https:\/\/colab.research.google.com\">Colaboratory<\/a>, the answers will usually include \u201cit\u2019s too slow\u201d or \u201cI don\u2019t trust it\u201d or \u201cI want my code on my local system.\u201d These sentiments seem to reflect some of the same motivations as local-first software. If we as developers want these things for ourselves and our work, perhaps we might imagine that other types of creative professionals would want these same qualities for their own work.<\/p><h3 id=\"dev-infrastructure\">Developer infrastructure for building apps<\/h3><p>Now that we have examined the user experience of a range of applications through the lens of the local-first ideals, let\u2019s switch mindsets to that of an application developer. If you are creating an app and want to offer users some or all of the local-first experience, what are your options for data storage and synchronization infrastructure?<\/p><h4 id=\"thin-client\">Web app (thin client)<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Web apps<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p>A web app in its purest form is usually a Rails, Django, PHP, or Node.js program running on a server, storing its data in a SQL or NoSQL database, and serving web pages over HTTPS. All of the data is on the server, and the user\u2019s web browser is only a thin client.<\/p><p>This architecture offers many benefits: zero installation (just visit a URL), and nothing for the user to manage, as all data is stored and managed in one place by the engineering and DevOps professionals who deploy the application. Users can access the application from all of their devices, and colleagues can easily collaborate by logging in to the same application.<\/p><aside class=\"move-up-6\"><p>JavaScript frameworks such as <a href=\"https:\/\/www.meteor.com\">Meteor<\/a> and <a href=\"https:\/\/github.com\/share\/sharedb\">ShareDB<\/a>, and services such as <a href=\"https:\/\/pusher.com\">Pusher<\/a> and <a href=\"https:\/\/www.ably.io\">Ably<\/a>, make it easier to add real-time collaboration features to web applications, building on top of lower-level protocols such as <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/WebSockets_API\">WebSocket<\/a>.<\/p><\/aside><p>On the other hand, a web app that needs to perform a request to a server for every user action is going to be slow. It is possible to <a href=\"https:\/\/uxplanet.org\/optimistic-1000-34d9eefe4c05\">hide the round-trip times<\/a> in some cases by using client-side JavaScript, but these approaches quickly break down if the user\u2019s internet connection is unstable.<\/p><p>Despite many efforts to make web browsers more offline-friendly (<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/HTML\/Using_the_application_cache\">manifests<\/a>, <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Window\/localStorage\">localStorage<\/a>, <a href=\"https:\/\/developers.google.com\/web\/fundamentals\/primers\/service-workers\/\">service workers<\/a>, and <a href=\"https:\/\/developers.google.com\/web\/progressive-web-apps\/\">Progressive Web Apps<\/a>, among others), the architecture of web apps remains fundamentally server-centric. Offline support is an afterthought in most web apps, and the result is accordingly fragile. In many web browsers, if the user clears their cookies, <a href=\"https:\/\/stackoverflow.com\/questions\/8537112\/when-is-localstorage-cleared\">all data in local storage is also deleted<\/a>; while this is not a problem for a cache, it makes the browser\u2019s local storage unsuitable for storing data of any long-term importance.<\/p><aside class=\"move-up-6\"><p>News website <a href=\"https:\/\/www.theguardian.com\/info\/developer-blog\/2015\/nov\/04\/building-an-offline-page-for-theguardiancom\">The Guardian<\/a> documents how they used service workers to build an offline experience for their users.<\/p><\/aside><p>Relying on third-party web apps also scores poorly in terms of longevity, privacy, and user control. It is possible to improve these properties if the web app is open source and users are willing to self-host their own instances of the server. However, we believe that self-hosting is <a href=\"https:\/\/rosenzweig.io\/blog\/the-federation-fallacy.html\">not a viable option<\/a> for the vast majority of users who do not want to become system administrators; moreover, most web apps are closed source, ruling out this option entirely.<\/p><p>All in all, we speculate that web apps will never be able to provide all the local-first properties we are looking for, due to the fundamental thin-client nature of the platform. By choosing to build a web app, you are choosing the path of data belonging to you and your company, not to your users.<\/p><h4 id=\"thick-client\">Mobile app with local storage (thick client)<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Thick client<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p>iOS and Android apps are locally installed software, with the entire app binary downloaded and installed before the app is run. Many apps are nevertheless thin clients, similarly to web apps, which require a server in order to function (for example, Twitter, Yelp, or Facebook). Without a reliable Internet connection, these apps give you spinners, error messages, and unexpected behavior.<\/p><p>However, there is another category of mobile apps that are more in line with the local-first ideals. These apps store data on the local device in the first instance, using a persistence layer like <a href=\"https:\/\/sqlite.org\/\">SQLite<\/a>, <a href=\"https:\/\/developer.apple.com\/documentation\/coredata\">Core Data<\/a>, or just plain files. Some of these (such as <a href=\"https:\/\/helloclue.com\">Clue<\/a> or <a href=\"https:\/\/culturedcode.com\/things\/\">Things<\/a>) started life as a single-user app without any server, and then added a cloud backend later, as a way to sync between devices or share data with other users.<\/p><p>These thick-client apps have the advantage of being fast and working offline, because the server sync happens in the background. They generally continue working if the server is shut down. The degree to which they offer privacy and user control over data varies depending on the app in question.<\/p><p>Things get more difficult if the data may be modified on multiple devices or by multiple collaborating users. The developers of mobile apps are generally experts in end-user app development, not in distributed systems. We have seen multiple app development teams writing their own ad-hoc diffing, merging, and conflict resolution algorithms, and the resulting data sync solutions are often unreliable and brittle. A more specialized storage backend, as discussed in the next section, can help.<\/p><h4 id=\"backend-service\">Backend-as-a-Service: Firebase, CloudKit, Realm<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Firebase, CloudKit, Realm<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"poor\">\u2717<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p><a href=\"https:\/\/firebase.google.com\">Firebase<\/a> is the most successful of mobile backend-as-a-service options. It is essentially a local on-device database combined with a cloud database service and data synchronization between the two. Firebase allows sharing of data across multiple devices, and it supports <a href=\"https:\/\/firebase.google.com\/docs\/firestore\/manage-data\/enable-offline\">offline use<\/a>. However, as a proprietary hosted service, we give it a low score for privacy and longevity.<\/p><aside class=\"move-up-5\"><p>Another popular backend-as-a-service was <a href=\"https:\/\/en.wikipedia.org\/wiki\/Parse_(platform)\">Parse<\/a>, but it was acquired and then <a href=\"https:\/\/willowtreeapps.com\/ideas\/parse-shutdown-what-it-means-and-what-you-can-do\">shut down<\/a> by Facebook in 2017. Apps relying on it were forced to move to other backend services, underlining the importance of longevity.<\/p><\/aside><p>Firebase offers a great experience for you, the developer: you can view, edit, and delete data in a free-form way in the Firebase console. But the user does not have a comparable way of accessing, manipulating and managing their data, leaving the user with little ownership and control.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/firebase-console.png\" alt=\"The Firebase console, where data can be viewed and edited\" \/><figcaption>The Firebase console: great for developers, off-limits for the end user.\n          <\/figcaption><\/figure><p><a href=\"https:\/\/developer.apple.com\/icloud\/cloudkit\/\">Apple\u2019s CloudKit<\/a> offers a Firebase-like experience for apps willing to limit themselves to the iOS and Mac platforms. It is a key-value store with syncing, good offline capabilities, and it has the added benefit of being built into the platform (thereby sidestepping the clumsiness of users having to create an account and log in). It\u2019s a great choice for indie iOS developers and is used to good effect by tools like <a href=\"https:\/\/ulysses.app\">Ulysses<\/a>, <a href=\"https:\/\/bear.app\">Bear<\/a>, <a href=\"https:\/\/overcast.fm\">Overcast<\/a>, and many more.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/ulysses-sync.png\" alt=\"The preferences dialog of Ulysses, with the iCloud option checked\" \/><figcaption>With one checkbox, Ulysses syncs work across all of the user\u2019s connected devices, thanks to its use of CloudKit.\n          <\/figcaption><\/figure><p>Another project in this vein is <a href=\"https:\/\/realm.io\">Realm<\/a>. This persistence library for iOS gained popularity compared to Core Data due to its cleaner API. The client-side library for local persistence is called <em>Realm Database<\/em>, while the associated Firebase-like backend service is called <em><a href=\"https:\/\/docs.realm.io\/sync\/what-is-realm-platform\">Realm Object Server<\/a><\/em>. Notably, the object server is <a href=\"https:\/\/github.com\/realm\/realm-object-server\">open source and self-hostable<\/a>, which reduces the risk of being locked in to a service that might one day disappear.<\/p><p>Mobile apps that treat the on-device data as the primary copy (or at least more than a disposable cache), and use sync services like Firebase or iCloud, get us a good bit of the way toward local-first software.<\/p><h4 id=\"couchdb\">CouchDB<\/h4><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>CouchDB<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"poor\">\u2717<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"okay\">\u2014<\/td>\n            <td class=\"okay\">\u2014<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p><a href=\"https:\/\/couchdb.apache.org\">CouchDB<\/a> is a database that is notable for pioneering a multi-master replication approach: several machines each have a fully-fledged copy of the database, each replica can independently make changes to the data, and any pair of replicas can synchronize with each other to exchange the latest changes. CouchDB is designed for use on servers; <a href=\"https:\/\/www.ibm.com\/cloud\/cloudant\">Cloudant<\/a> provides a hosted version; <a href=\"https:\/\/pouchdb.com\">PouchDB<\/a> and <a href=\"http:\/\/hood.ie\/\">Hoodie<\/a> are sibling projects that use the same sync protocol but are designed to run on end-user devices.<\/p><p>Philosophically, CouchDB is closely aligned to the local-first principles, as evidenced in particular by the <a href=\"http:\/\/guide.couchdb.org\/\">CouchDB book<\/a>, which provides an excellent introduction to relevant topics such as <a href=\"http:\/\/guide.couchdb.org\/editions\/1\/en\/consistency.html\">distributed consistency<\/a>, <a href=\"http:\/\/guide.couchdb.org\/editions\/1\/en\/replication.html\">replication<\/a>, <a href=\"http:\/\/guide.couchdb.org\/editions\/1\/en\/notifications.html\">change notifications<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Multiversion_concurrency_control\">multiversion concurrency control<\/a>.<\/p><p>While CouchDB\/PouchDB allow multiple devices to concurrently make changes to a database, these changes lead to conflicts that need to be explicitly resolved by application code. This conflict resolution code is difficult to write correctly, making CouchDB impractical for applications with very fine-grained collaboration, like in Google Docs, where every keystroke is potentially an individual change.<\/p><p>In practice, the CouchDB model has <a href=\"https:\/\/medium.com\/offline-camp\/couchdb-pouchdb-and-hoodie-as-a-stack-for-progressive-web-apps-a6078a985f18\">not been widely adopted<\/a>. Various reasons have been cited for this: scalability problems when a separate database per user is required; difficulty embedding the JavaScript client in native apps on iOS and Android; the problem of conflict resolution; the unfamiliar MapReduce model for performing queries; and more. All in all, while we agree with much of the philosophy behind CouchDB, we feel that the implementation has not been able to realize the local-first vision in practice.<\/p><h2 id=\"future\">Towards a better future<\/h2><p>As we have shown, none of the existing data layers for application development fully satisfy the local-first ideals. Thus, three years ago, our lab set out to search for a solution that gives seven green checkmarks.<\/p><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>???<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n            <td class=\"good\">\u2713<\/td>\n          <\/tr><\/tbody><\/table><\/figure><p>We have found some technologies that appear to be promising foundations for local-first ideals. Most notably are the family of distributed systems algorithms called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Conflict-free_replicated_data_type\">Conflict-free Replicated Data Types<\/a> (CRDTs).<\/p><h3 id=\"crdts\">CRDTs as a foundational technology<\/h3><p>CRDTs emerged from academic computer science research <a href=\"https:\/\/pages.lip6.fr\/Marc.Shapiro\/papers\/RR-7687.pdf\">in 2011<\/a>. They are general-purpose data structures, like hash maps and lists, but the special thing about them is that they are multi-user from the ground up.<\/p><p>Every application needs some data structures to store its document state. For example, if your application is a text editor, the core data structure is the array of characters that make up the document. If your application is a spreadsheet, the data structure is a matrix of cells containing text, numbers, or formulas referencing other cells. If it is a vector graphics application, the data structure is a tree of graphical objects such as text objects, rectangles, lines, and other shapes.<\/p><p>If you are building a single-user application, you would maintain those data structures in memory using model objects, hash maps, lists, records\/structs and the like. If you are building a collaborative multi-user application, you can swap out those data structures for CRDTs.<\/p><p>This means that if you have a copy of a document, you can make changes to the CRDTs, and the data structures keep track of the changes. The changes can be encoded as a byte string and sent to your collaborators via any communication channel you like (e.g. via a server, over a peer-to-peer connection, by Bluetooth between local devices, or even on a USB stick).<\/p><p>Your collaborators can receive these changes and apply them to their copy of the document. Unlike a merge in Git, where you might have to resolve merge conflicts by hand, CRDTs perform merges automatically.<\/p><p>Breaking it down:<\/p><ul><li><strong>Conflict-free:<\/strong> when several collaborators make changes, they can be automatically merged together into a consistent view of the document;<\/li>\n            <li><strong>Replicated:<\/strong> every collaborator has their own copy of the data, stored on their local device;<\/li>\n            <li><strong>Data Types:<\/strong> general-purpose data structures like maps and lists.<\/li><\/ul><p>These properties map directly to our local-first ideals. Having a copy of the data locally, rather than depending on a server, enables offline working, longevity, good performance, and user control. Being able to automatically merge changes enables multi-device use and collaboration. Flexibility of communication channels allows encryption, and hence improved privacy and security.<\/p><p>The changes tracked by a CRDT can be as small as a single keystroke, enabling Google Docs-style real-time collaboration. But you could also collect a larger set of changes and send them to collaborators as a batch, more like a pull request in Git. Because the data structures are general-purpose, we can develop general-purpose tools for storage, communication, and management of CRDTs, saving us from having to re-implement those things in every single app.<\/p><p>For a more technical introduction to CRDTs we suggest:<\/p><ul><li>Alexei Baboulevitch\u2019s <a href=\"http:\/\/archagon.net\/blog\/2018\/03\/24\/data-laced-with-history\/\">Data Laced with History<\/a><\/li>\n          <li>Martin Kleppmann\u2019s <a href=\"https:\/\/www.youtube.com\/watch?v=B5NULPSiOGw\">Convergence vs Consensus<\/a> (<a href=\"https:\/\/speakerdeck.com\/ept\/convergence-versus-consensus-crdts-and-the-quest-for-distributed-consistency\">slides<\/a>)<\/li>\n          <li>Shapiro et al.\u2019s <a href=\"https:\/\/hal.inria.fr\/inria-00555588\/document\">comprehensive survey<\/a><\/li>\n          <li>Attiya et al.\u2019s <a href=\"http:\/\/software.imdea.org\/~gotsman\/papers\/editing-podc16.pdf\">specification of collaborative text editing<\/a><\/li>\n          <li>Gomes et al.\u2019s <a href=\"https:\/\/dl.acm.org\/citation.cfm?doid=3152284.3133933\">formal verification of CRDTs<\/a><\/li>\n        <\/ul><p>Ink &amp; Switch has developed an open-source, JavaScript CRDT implementation called <a href=\"https:\/\/github.com\/automerge\/automerge\">Automerge<\/a>. It is based on our earlier research on <a href=\"https:\/\/arxiv.org\/pdf\/1608.03960.pdf\">JSON CRDTs<\/a>. We have then combined Automerge with the <a href=\"https:\/\/datproject.org\">Dat networking stack<\/a> to form <a href=\"https:\/\/github.com\/automerge\/hypermerge\">Hypermerge<\/a>. We do not claim that these libraries fully realize local-first ideals \u2014 more work is still required.<\/p><p>However, based on our experience with them, we believe that CRDTs have the potential to be a foundation for a new generation of software. Just as packet switching was an enabling technology for the Internet and the web, or as capacitive touchscreens were an enabling technology for smartphones, so we think CRDTs may be the foundation for collaborative software that gives users full ownership of their data.<\/p><h3 id=\"prototypes\">Ink &amp; Switch prototypes<\/h3><p>While academic research has made good progress designing the algorithms for CRDTs and verifying their theoretical correctness, there is so far relatively little industrial use of these technologies. Moreover, most industrial CRDT use has been in server-centric computing, but we believe this technology has significant potential in client-side applications for creative work.<\/p><aside class=\"move-up-4\"><p>Server-centric systems using CRDTs include <a href=\"https:\/\/azure.microsoft.com\/en-us\/blog\/azure-cosmos-db-pushing-the-frontier-of-globally-distributed-databases\/\">Azure Cosmos DB<\/a>, <a href=\"http:\/\/christophermeiklejohn.com\/erlang\/lasp\/2019\/03\/08\/monotonicity.html\">Riak<\/a>, <a href=\"https:\/\/www.infoq.com\/presentations\/weave-mesh\">Weave Mesh<\/a>, <a href=\"https:\/\/developers.soundcloud.com\/blog\/roshi-a-crdt-system-for-timestamped-events\">SoundCloud\u2019s Roshi<\/a>, and <a href=\"https:\/\/github.com\/facebook\/openr\/blob\/master\/openr\/docs\/KvStore.md\">Facebook\u2019s OpenR<\/a>. However, we are most interested in the use of CRDTs on end-user devices.<\/p><\/aside><p>This was the motivation for <a href=\"https:\/\/www.inkandswitch.com\">our lab<\/a> to embark on a series of experimental prototypes with collaborative, local-first applications built on CRDTs. Each prototype offered an end-user experience modeled after an existing app for creative work such as Trello, Figma, or Milanote.<\/p><p>These experiments explored questions in three areas:<\/p><ul><li><strong>Technology viability.<\/strong> How close are CRDTs to being usable for working software? What do we need for network communication, or installation of the software to begin with?<\/li>\n          <li><strong>User experience.<\/strong> How does local-first software feel to use? Can we get a seamless Google Docs-like real-time collaboration experience without an authoritative centralized server? How about a Git-like, offline-friendly, asynchronous collaboration experience for data types other than source code? And generally, how are user interfaces different without a centralized server?<\/li>\n          <li><strong>Developer experience.<\/strong> For an app developer, how does the use of a CRDT-based data layer compare to existing storage layers like a SQL database, a filesystem, or Core Data? Is a distributed system harder to write software for? Do we need schemas and type checking? What will developers use for debugging and introspection of their application\u2019s data layer?<\/li>\n        <\/ul><p>We built three prototypes using <a href=\"https:\/\/electronjs.org\/\">Electron<\/a>, JavaScript, and <a href=\"https:\/\/reactjs.org\">React<\/a>. This gave us the rapid development capability of web technologies while also giving our users a piece of software they can download and install, which we discovered is an important part of the local-first feeling of ownership.<\/p><h4 id=\"trellis\">Kanban board<\/h4><p><a href=\"https:\/\/github.com\/automerge\/trellis#readme\">Trellis<\/a> is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Kanban_board\">Kanban board<\/a> modeled after the popular <a href=\"https:\/\/trello.com\/\">Trello<\/a> project management software.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/trellis.png\" alt=\"Screenshot of Trellis, a clone of Trello\" \/><figcaption>Trellis offers a Trello-like experience with local-first software. The change history on the right reflects changes made by all users active in the document.\n          <\/figcaption><\/figure><p>On this project we experimented with <a href=\"https:\/\/webrtc.org\">WebRTC<\/a> for the network communication layer.<\/p><p>On the user experience side, we designed a rudimentary \u201cchange history\u201d inspired by Git and Google Docs\u2019 \u201c<a href=\"https:\/\/support.google.com\/docs\/answer\/190843\">See New Changes<\/a>\u201d that allows users to see the operations on their Kanban board. This includes stepping back in time to view earlier states of the document.<\/p><p>Watch Trellis in action with the <a href=\"https:\/\/www.youtube.com\/watch?v=L9fdyDlhByM\">demo video<\/a> or <a href=\"https:\/\/github.com\/automerge\/trellis\/releases\">download a release<\/a> and try it yourself.<\/p><h4 id=\"pixelpusher\">Collaborative drawing<\/h4><p><a href=\"https:\/\/github.com\/automerge\/pixelpusher#readme\">PixelPusher<\/a> is a collaborative drawing program, bringing a Figma-like real-time experience to <a href=\"https:\/\/www.jvalen.com\">Javier Valencia<\/a>\u2019s <a href=\"https:\/\/www.pixelartcss.com\">Pixel Art to CSS<\/a>.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/pixelpusher.png\" alt=\"Screenshot of the PixelPusher user interface\" \/><figcaption>Drawing together in real-time. A URL at the top offers a quick way to share this document with other users. The \u201cVersions\u201d panel on the right shows all branches of the current document. The arrow buttons offer instant merging between branches.\n          <\/figcaption><\/figure><p>On this project we experimented with network communication via peer-to-peer libraries from the <a href=\"https:\/\/datproject.org\">Dat project<\/a>.<\/p><p>User experience experiments include URLs for document sharing, a visual branch\/merge facility inspired by Git, a conflict-resolution mechanism that highlights conflicted pixels in red, and basic user identity via user-drawn avatars.<\/p><p>Read the <a href=\"https:\/\/medium.com\/@pvh\/pixelpusher-real-time-peer-to-peer-collaboration-with-react-7c7bc8ecbf74\">full project report<\/a> or <a href=\"https:\/\/github.com\/automerge\/pixelpusher\/releases\">download a release<\/a> to try it yourself.<\/p><h4 id=\"pushpin\">Media canvas<\/h4><p><a href=\"https:\/\/inkandswitch.github.io\/pushpin\/\">PushPin<\/a> is a mixed media canvas workspace similar to <a href=\"https:\/\/miro.com\">Miro<\/a> or <a href=\"https:\/\/www.milanote.com\">Milanote<\/a>. As our third project built on Automerge, it\u2019s the most fully-realized of these three. Real use by our team and external test users put more strain on the underlying data layer.<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/pushpin.jpg\" alt=\"Screenshot of PushPin, showing images and text cards on a canvas\" \/><figcaption>PushPin\u2019s canvas mixes text, images, discussion threads, and web links. Users see each other via presence avatars in the toolbar, and navigate between their own documents using the URL bar.\n          <\/figcaption><\/figure><p>PushPin explored nested and connected shared documents, varied renderers for CRDT documents, a more advanced identity system that included an \u201coutbox\u201d model for sharing, and support for sharing ephemeral data such as selection highlights.<\/p><p>Watch the <a href=\"https:\/\/www.youtube.com\/watch?v=Dox3XAoTCyg\">PushPin demo video<\/a> or <a href=\"https:\/\/github.com\/inkandswitch\/pushpin\/releases\">download a release<\/a> and try it yourself.<\/p><h4 id=\"findings\">Findings<\/h4><p>We learned the following lessons from building and using these prototypes.<\/p><p><strong>CRDT technology works.<\/strong><\/p><p>From the beginning we were pleasantly surprised by the reliability of Automerge. App developers on our team were able to integrate the library with relative ease, and the automatic merging of data was almost always straightforward and seamless.<\/p><p><strong>The user experience with offline work is splendid.<\/strong><\/p><p>The process of going offline, continuing to work for as long as you want, and then reconnecting to merge changes with colleagues worked well. While other applications on the system threw up errors (\u201coffline! warning!\u201d) and blocked the user from working, the local-first prototypes function normally regardless of network status. Unlike browser-based systems, there is never any anxiety about whether the application will work or the data will be there when the user needs it. This gives the user a feeling of ownership over their tools and their work, just as we had hoped.<\/p><p><strong>Developer experience is viable when combined with <a href=\"http:\/\/people.seas.harvard.edu\/~chong\/pubs\/pldi13-elm.pdf\">Functional Reactive Programming<\/a> (FRP).<\/strong><\/p><p>The FRP model of React fits well with CRDTs. A data layer based on CRDTs means the user\u2019s document is simultaneously getting updates from the local user (e.g. as they type into a text document) but also from the network (as other users and other devices make changes to the document).<\/p><p>Because the FRP model reliably synchronizes the visible state of the application with the underlying state of the shared document, the developer is freed from the tedious work of tracking changes arriving from other users and reconciling them with the current view. Also, by ensuring all changes to the underlying state are made through a single function (a <a href=\"https:\/\/redux.js.org\/basics\/reducers\">\u201creducer\u201d<\/a>), it\u2019s easy to ensure that all relevant local changes are sent to other users.<\/p><p>The result of this model was that all of our prototypes realized real-time collaboration and full offline capability with little effort from the application developer. This is a significant benefit as it allows app developers to focus on their application rather than the challenges of data distribution.<\/p><p><strong>Conflicts are not as significant a problem as we feared.<\/strong><\/p><p>We were curious how often users would encounter conflicts in their work when collaborating with others. We were pleasantly surprised to discover that users have an intuitive sense of human collaboration and avoid creating conflicts with their collaborators. In cases where conflicts do occur, it is often arbitrary how they are resolved and satisfactory either way. For example, if two users fix a typo in the document\u2019s title, they may not do it the same way, but they often do not appear to care which outcome is selected.<\/p><p>When different users concurrently modify different parts of the document state, Automerge will merge these changes cleanly without difficulty. With the Kanban app, for example, one user could post a comment on a card and another could move it to another column, and the merged result will reflect both of these changes.<\/p><p>In cases where conflict resolution is not clear, Automerge retains the conflicting values and allows the application to resolve them. In the PixelPusher prototype we explored a user interface for conflict resolution. Combining these features with change history inspection is an open research question.<\/p><p><strong>Visualizing document history is important.<\/strong><\/p><p>In a distributed collaborative system another user can deliver any number of changes to you at any moment. Unlike centralized systems, where servers mediate change, local-first applications need to find their own solutions to these problems. Without the right tools, it can be difficult to understand how a document came to look the way it does, what versions of the document exist, or where contributions came from.<\/p><p>In the Trellis project we experimented with a \u201ctime travel\u201d interface, allowing a user to move back in time to see earlier states of a merged document, and automatically highlighting recently changed elements as changes are received from other users. The ability to traverse a potentially complex merged document history in a linear fashion helps to provide context and could become a universal tool for understanding collaboration.<\/p><p><strong>URLs are a good mechanism for sharing.<\/strong><\/p><p>We experimented with a number of mechanisms for sharing documents with other users, and found that a URL model, inspired by the web, makes the most sense to users and developers. URLs can be copied and pasted, and shared via communication channels such as email or chat. Access permissions for documents beyond secret URLs remain an open research question.<\/p><p><strong>Peer-to-peer systems are never fully \u201conline\u201d or \u201coffline\u201d and it can be hard to reason about how data moves in them.<\/strong><\/p><p>A traditional centralized system is generally \u201cup\u201d or \u201cdown,\u201d states defined by each client by their ability to maintain a steady network connection to the server. The server determines the truth of a given piece of data.<\/p><p>In a decentralized system, we can have a kaleidoscopic complexity to our data. Any user may have a different perspective on what data they either have, choose to share, or accept. For example, one user\u2019s edits to a document might be on their laptop on an airplane; when the plane lands and the computer reconnects, those changes are distributed to other users. Other users might choose to accept all, some, or none of those changes to their version of the document.<\/p><p>Different versions of a document can lead to confusion. As with a Git repository, what a particular user sees in the \u201cmaster\u201d branch is a function of the last time they communicated with other users. Newly arriving changes might unexpectedly modify parts of the document you are working on, but manually merging every change from every user is tedious. Decentralized documents enable users to be in control over their own data, but further study is needed to understand what this means in practical user-interface terms.<\/p><p><strong>CRDTs accumulate a large change history, which creates performance problems.<\/strong><\/p><p>Our team used PushPin for \u201creal\u201d documents such as sprint planning. Performance and memory\/disk usage quickly became a problem because CRDTs store all history, including character-by-character text edits. These pile up, but can\u2019t easily be truncated because it\u2019s impossible to know when someone might reconnect to your shared document after six months away and need to merge changes from that point forward.<\/p><p>We continue to optimize Automerge, but this is a major area of ongoing work.<\/p><p><strong>Network communication remains an unsolved problem.<\/strong><\/p><p>CRDT algorithms provide only for the merging of data, but say nothing about how different users\u2019 edits arrive on the same physical computer.<\/p><p>In these experiments we tried network communication via <a href=\"https:\/\/webrtc.org\">WebRTC<\/a>; a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sneakernet\">\u201csneakernet\u201d<\/a> implementation of copying files around with Dropbox and USB keys; possible use of the <a href=\"https:\/\/ipfs.io\">IPFS protocols<\/a>; and eventually settled on the <a href=\"https:\/\/github.com\/mafintosh\/hypercore\">Hypercore<\/a> peer-to-peer libraries from <a href=\"https:\/\/datproject.org\">Dat<\/a>.<\/p><p>CRDTs do not require a peer-to-peer networking layer; using a server for communication is fine for CRDTs. However, to fully realize the longevity goal of local-first software, we want applications to outlive any backend services managed by their vendors, so a decentralized solution is the logical end goal.<\/p><p>The use of P2P technologies in our prototypes yielded mixed results. On one hand, these technologies are nowhere near production-ready: <a href=\"https:\/\/tools.ietf.org\/html\/rfc5389\">NAT traversal<\/a>, in particular, is unreliable depending on the particular router or network topology where the user is currently connected. But the promise suggested by P2P protocols and the <a href=\"https:\/\/decentralizedweb.net\">Decentralized Web<\/a> community is substantial. Live collaboration between computers without Internet access feels like magic in a world that has come to depend on centralized APIs.<\/p><p><strong>Cloud servers still have their place for discovery, backup, and burst compute.<\/strong><\/p><p>A real-time collaborative prototype like PushPin lets users share their documents with other users without an intermediating server. This is excellent for privacy and ownership, but can result in situations where a user shares a document, and then closes their laptop lid before the other user has connected. If the users are not online at the same time, they cannot connect to each other.<\/p><p>Servers thus have a role to play in the local-first world \u2014 not as central authorities, but as \u201ccloud peers\u201d that support client applications without being on the critical path. For example, a cloud peer that stores a copy of the document, and forwards it to other peers when they come online, could solve the closed-laptop problem above.<\/p><aside><p><a href=\"https:\/\/hashbase.io\/\">Hashbase<\/a> is an example of a cloud peer and bridge for <a href=\"https:\/\/datproject.org\">Dat<\/a> and <a href=\"https:\/\/beakerbrowser.com\">Beaker Browser<\/a>.<\/p><\/aside><p>Similarly, cloud peers could be:<\/p><ul><li>an archival\/backup location (especially for phones or other devices with limited storage);<\/li>\n          <li>a bridge to traditional server APIs (such as weather forecasts or a stock tickers);<\/li>\n          <li>a provider of burst computing resources (like rendering a video using a powerful GPU).<\/li>\n        <\/ul><p>The key difference between traditional systems and local-first systems is not an absence of servers, but a change in their responsibilities: they are in a supporting role, not the source of truth.<\/p><h3 id=\"help\">How you can help<\/h3><p>These experiments suggest that local-first software is possible. Collaboration and ownership are not at odds with each other \u2014 we can get the best of both worlds, and users can benefit.<\/p><p>However, the underlying technologies are still a work in progress. They are good for developing prototypes, and we hope that they will evolve and stabilize in the coming years, but realistically, it is not yet advisable to replace a proven product like Firebase with an experimental project like Automerge in a production setting today.<\/p><p>If you believe in a local-first future, as we do, what can you (and all of us in the technology field) do to move us toward it? Here are some suggestions.<\/p><h4 id=\"distsys\">For distributed systems and programming languages researchers<\/h4><p>Local-first software has benefited tremendously from recent research into distributed systems, including CRDTs and peer-to-peer technologies. The current research community is making excellent progress in improving the performance and power of CRDTs and we eagerly await further results from that work. Still, there are interesting opportunities for further work.<\/p><p>Most CRDT research operates in a model where all collaborators immediately apply their edits to a single version of a document. However, practical local-first applications require more flexibility: users must have the freedom to reject edits made by another collaborator, or to make private changes to a version of the document that is not shared with others. A user might want to apply changes speculatively or reformat their change history. These concepts are well understood in the distributed source control world as \u201cbranches,\u201d \u201cforks,\u201d \u201crebasing,\u201d and so on. There is little work to date on understanding the algorithms and programming models for collaboration in situations where multiple document versions and branches exist side-by-side.<\/p><p>We see further interesting problems around types, schema migrations, and compatibility. Different collaborators may be using different versions of an application, potentially with different features. As there is no central database server, there is no authoritative \u201ccurrent\u201d schema for the data. How can we write software so that varying application versions can safely interoperate, even as data formats evolve? This question has analogues in cloud-based API design, but a local-first setting provides additional challenges.<\/p><h4 id=\"hci\">For Human-Computer Interaction (HCI) researchers<\/h4><p>For centralized systems, there are ample examples in the field today of applications that indicate their \u201csync\u201d state with a server. Decentralized systems have a whole host of interesting new opportunities to explore user interface challenges.<\/p><p>We hope researchers will consider how to communicate online and offline states, or available and unavailable states for systems where any other user may hold a different copy of data. How should we think about connectivity when everyone is a peer? What does it mean to be \u201conline\u201d when we can collaborate directly with other nodes without access to the wider Internet?<\/p><figure><img src=\"https:\/\/www.inkandswitch.com\/media\/local-first\/git-history.png\" alt=\"Example Git commit history as visualized by GitX\" \/><figcaption>The \u201crailroad track\u201d model, as used in <a href=\"https:\/\/rowanj.github.io\/gitx\/\">GitX<\/a> for visualizing the structure of source code history in a Git repository.\n          <\/figcaption><\/figure><p>When every document can develop a complex version history, simply through daily operation, an acute problem arises: how do we communicate this version history to users? How should users think about versioning, share and accept changes, and understand how their documents came to be a certain way when there is no central source of truth? Today there are two mainstream models for change management: a source-code model of diffs and patches, and a Google Docs model of suggestions and comments. Are these the best we can do? How do we generalize these ideas to data formats that are not text? We are eager to see what can be discovered.<\/p><p>While centralized systems rely heavily on access control and permissions, the same concepts do not directly apply in a local-first context. For example, any user who has a copy of some data cannot be prevented from locally modifying it; however, other users may choose whether or not to subscribe to those changes. How should users think about sharing, permissions, and feedback? If we can\u2019t remove documents from others\u2019 computers, what does it mean to \u201cstop sharing\u201d with someone?<\/p><p>We believe that the assumption of centralization is deeply ingrained in our user experiences today, and we are only beginning to discover the consequences of changing that assumption. We hope these open questions will inspire researchers to explore what we believe is an untapped area.<\/p><h4 id=\"practitioners\">For practitioners<\/h4><p>If you\u2019re a software engineer, designer, product manager, or independent app developer working on production-ready software today, how can you help? We suggest taking incremental steps toward a local-first future. Start by scoring your app:<\/p><figure class=\"wide\"><table class=\"scorecard align-center\"><tbody><tr><th><strong>1.<\/strong> Fast<\/th>\n              <th><strong>2.<\/strong> Multi-device<\/th>\n              <th><strong>3.<\/strong> Offline<\/th>\n              <th><strong>4.<\/strong> Collaboration<\/th>\n              <th><strong>5.<\/strong> Longevity<\/th>\n              <th><strong>6.<\/strong> Privacy<\/th>\n              <th><strong>7.<\/strong> User control<\/th>\n            <\/tr><tr><td>Your app<\/td>\n            \n            \n            \n            \n            \n            \n            \n          <\/tr><\/tbody><\/table><\/figure><p>Then some strategies for improving each area:<\/p><ul><li><strong>Fast.<\/strong> Aggressive caching and downloading resources ahead of time can be a way to prevent the user from seeing spinners when they open your app or a document they previously had open. Trust the local cache by default instead of making the user wait for a network fetch.<\/li>\n            <li><strong>Multi-device.<\/strong> Syncing infrastructure like Firebase and iCloud make multi-device relatively painless, although they do introduce longevity and privacy concerns. Self-hosted infrastructure like <a href=\"https:\/\/docs.realm.io\/sync\/getting-started-1\/getting-a-realm-object-server-instance\">Realm Object Server<\/a> provides an alternative trade-off.<\/li>\n            <li><strong>Offline.<\/strong> In the web world, <a href=\"https:\/\/developers.google.com\/web\/progressive-web-apps\/\">Progressive Web Apps<\/a> offer features like Service Workers and app manifests that can help. In the mobile world, be aware of <a href=\"https:\/\/developer.apple.com\/documentation\/webkit\">WebKit<\/a> frames and other network-dependent components. Test your app by turning off your WiFi, or using traffic shapers such as the <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/network\/network-conditions\">Chrome Dev Tools network condition simulator<\/a> or the <a href=\"https:\/\/nshipster.com\/network-link-conditioner\/\">iOS network link conditioner<\/a>.<\/li>\n            <li><strong>Collaboration.<\/strong> Besides CRDTs, the more established technology for real-time collaboration is <a href=\"http:\/\/www.codecommit.com\/blog\/java\/understanding-and-applying-operational-transformation\">Operational Transformation<\/a> (OT), as implemented e.g. in <a href=\"https:\/\/github.com\/share\/sharedb\">ShareDB<\/a>.<\/li>\n            <li><strong>Longevity.<\/strong> Make sure your software can easily export to flattened, standard formats like JSON or PDF. For example: mass export such as <a href=\"https:\/\/takeout.google.com\/settings\/takeout\">Google Takeout<\/a>; continuous backup into stable file formats <a href=\"https:\/\/support.goodnotes.com\/hc\/en-us\/articles\/202168425-How-should-I-backup-my-documents-\">such as in GoodNotes<\/a>; and JSON download of documents <a href=\"https:\/\/help.trello.com\/article\/747-exporting-data-from-trello-1\">such as in Trello<\/a>.<\/li>\n            <li><strong>Privacy.<\/strong> Cloud apps are fundamentally non-private, with employees of the company and governments able to peek at user data at any time. But for mobile or desktop applications, try to make clear to users when the data is stored only on their device versus being transmitted to a backend.<\/li>\n            <li><strong>User control.<\/strong> Can users easily back up, duplicate, or delete some or all of their documents within your application? Often this involves re-implementing all the basic filesystem operations, as Google Docs has done with Google Drive.<\/li>\n        <\/ul><h4 id=\"startups\">Call for startups<\/h4><p>If you are an entrepreneur interested in building developer infrastructure, all of the above suggests an interesting market opportunity: \u201cFirebase for CRDTs.\u201d<\/p><p>Such a startup would need to offer a great developer experience and a local persistence library (something like SQLite or Realm). It would need to be available for mobile platforms (iOS, Android), native desktop (Windows, Mac, Linux), and web technologies (Electron, Progressive Web Apps).<\/p><p>User control, privacy, multi-device support, and collaboration would all be baked in. Application developers could focus on building their app, knowing that the easiest implementation path would also given them top marks on the local-first scorecard. As litmus test to see if you have succeeded, we suggest: do all your customers\u2019 apps continue working in perpetuity, even if all servers are shut down?<\/p><p>We believe the \u201cFirebase for CRDTs\u201d opportunity will be huge as CRDTs come of age. We\u2019d like to hear from you if you\u2019re working on this.<\/p><h2 id=\"conclusions\">Conclusions<\/h2><p>Computers are one of the most important creative tools mankind has ever produced. Software has become the conduit through which our work is done and the repository in which that work resides.<\/p><p>In the pursuit of better tools we moved many applications to the cloud. Cloud software is in many regards superior to \u201cold-fashioned\u201d software: it offers collaborative, always-up-to-date applications, accessible from anywhere in the world. We no longer worry about what software version we are running, or what machine a file lives on.<\/p><p>However, in the cloud, ownership of data is vested in the servers, not the users, and so we became borrowers of our own data. The documents created in cloud apps are destined to disappear when the creators of those services cease to maintain them. Cloud services defy long-term preservation. No Wayback Machine can restore a sunsetted web application. The Internet Archive cannot preserve your Google Docs.<\/p><p>In this article we explored a new way forward for software of the future. We have shown that it is possible for users to retain ownership and control of their data, while also benefiting from the features we associate with the cloud: seamless collaboration and access from anywhere. It is possible to get the best of both worlds.<\/p><p>But more work is needed to realize the local-first approach in practice. Application developers can take incremental steps, such as improving offline support and making better use of on-device storage. Researchers can continue improving the algorithms, programming models, and user interfaces for local-first software. Entrepreneurs can develop foundational technologies such as CRDTs and peer-to-peer networking into mature products able to power the next generation of applications.<\/p><p>Today it is easy to create a web application in which the server takes ownership of all the data. But it is too hard to build collaborative software that respects users\u2019 ownership and agency. In order to shift the balance, we need to improve the tools for developing local-first software. We hope that you will join us.<\/p><p class=\"outro\">We welcome your thoughts, questions, or critique: <a href=\"http:\/\/twitter.com\/inkandswitch\" title=\"Tweet us @inkandswitch\">@inkandswitch<\/a> or <a href=\"https:\/\/www.inkandswitch.com\/cdn-cgi\/l\/email-protection#2e464b4242416e4740454f404a5d59475a4d46004d4143\" title=\"Send us an Email\">[email\u00a0protected]<\/a>.<\/p><h3 id=\"acknowledgments\">Acknowledgments<\/h3><p>Martin Kleppmann is supported by a grant from The Boeing Company. Thank you to our collaborators at Ink &amp; Switch who worked on the prototypes discussed above: Julia Roggatz, Orion Henry, Roshan Choxi, Jeff Peterson, Jim Pick, and Ignatius Gilfedder. Thank you also to Heidi Howard for feedback on a draft of this article.<\/p>","created_at":"2019-05-03T21:55:14+0000","updated_at":"2019-05-03T21:55:14+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":56,"domain_name":"www.inkandswitch.com","preview_picture":"https:\/\/www.inkandswitch.com\/media\/local-first\/cuneiform.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":936,"title":"A most advanced operating system","url":"https:\/\/ambrevar.xyz\/guix-advance\/","content":"<p>\n<a href=\"https:\/\/www.gentoo.org\/\">Gentoo<\/a> together with all distributions using the <a href=\"https:\/\/wiki.gentoo.org\/wiki\/Project:Portage\">Portage<\/a> package manager are\nsource-based distributions with an iconic feature: the <a href=\"https:\/\/wiki.gentoo.org\/wiki\/Project:Portage\">USE flags<\/a>, which lets the\nuser enable features across the whole system (for instance disable sound, enable\ngraphical interface support, etc.).\n<\/p><p>\nUSE flags make it trivial to toggle features that were exposed by the packagers\n(and they have the benefit of being tested).  On the other hand, Portage does\nnot let you configure features that were not thought of in advance by the\npackager.  For instance, if a package has optional audio but the packager didn\u2019t\nexpose the corresponding USE flag, the user cannot do anything about it (beside\ncreating a new package definition).\n<\/p><p>\nConversely, Guix gives you full customization for everything, albeit with a bit\nmore Scheme code. Roughly, in pseudo-code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-lisp\">(loop-over (TARGET-PACKAGES)\n  (package\n    (inherit TARGET)\n    (changes-here... including patches, build options, etc.))\n<\/pre><\/div><p>\nwill batch-define <code>TARGET-PACKAGES<\/code> with your changes.  None of the changes have\nto be carved into the package definition.  At any time, the user retains full\ncontrol over the changes that can be made to the packages.\n<\/p><p>\nI\u2019ve loved Gentoo while I was using it, but after moving to Guix it became\napparent that Portage\u2019s approach to package customization is more limited.\n<\/p><ul class=\"org-ul\"><li>The USE flag system does not let you customize unplanned, arbitrary features.<\/li>\n<li>USE flags add a whole class of complexity (see the rather complex <a href=\"https:\/\/dev.gentoo.org\/~zmedico\/portage\/doc\/man\/ebuild.5.html\">atom\nsemantic<\/a>) to describe and manager feature relationships across packages.  Guix\ncompletely nullifies this complexity layer by using Guile Scheme to program\nthe relationships between packages.<\/li>\n<\/ul><p>\nMoreover, Portage suffer from the same aforementioned issue with the lack of\nproper support for multiple versions.  Indeed, USE flags significantly increase\nthe magnitude of the issue (a frequent complaint about Portage): when\nincompatible USE flags propagate to some dependencies, the user must find a\nresolution manually.  Sometimes, it means that the desired feature won\u2019t be\napplicable (at least not without significant work on the package definitions).\n<\/p><p>\nOn the practical level, Guix provides pre-compiled packages which can be a huge\ntime-saver compared to Gentoo which does not (but Portage support binary package\ndistribution).\n<\/p><p>\nThe *BSD systems (e.g. FreeBSD) suffer from similar issues with the <code>make\nconfig<\/code> customization interface.\n<\/p>","created_at":"2019-05-03T20:22:59+0000","updated_at":"2019-05-03T20:22:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"ambrevar.xyz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":935,"title":"How do you add a new window manager to the gdm menu?","url":"https:\/\/askubuntu.com\/questions\/4645\/how-do-you-add-a-new-window-manager-to-the-gdm-menu\/4687#4687","content":"<p>By gdm options menu, do you mean the list at the bottom of the login screen ?<\/p><p>If yes, then, as far as I know, the \"correct approach\" for adding a window manager to the list of session options presented to user at login time is to add a .desktop file to \/usr\/share\/xsessions. Take an existing .desktop file there as starting point (say, gnome.desktop) and modify fields like Exec, TryExec, Icon to suit your new WM.  <\/p><p>Since this is so obvious (by that I mean covered in the tutorials of many WM's I have tried to add), I think I haven't understood your question well. Can you confirm if you have tried this ?<\/p>","created_at":"2019-05-03T18:19:16+0000","updated_at":"2019-05-03T18:19:16+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"askubuntu.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/askubuntu\/img\/apple-touch-icon@2.png?v=c492c9229955","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["linux","virtualbox","arch"],"is_public":false,"id":934,"title":"How to install Arch Linux on VirtualBox","url":"https:\/\/www.howtoforge.com\/tutorial\/install-arch-linux-on-virtualbox\/","content":"<div class=\"contributeEdit\"><div id=\"tocContainer\"><h3>On this page<\/h3><ol class=\"toc\"><li><a href=\"#arch-linux-repositories\">Arch Linux Repositories<\/a><\/li>\n<li><a href=\"#install-arch-linux-on-virtual-box\">Install Arch Linux on Virtual Box<\/a>\n<ol><li><a href=\"#download-arch-linux\">Download Arch Linux<\/a><\/li>\n<li><a href=\"#initializing-installation-with-oracle-vm-virtualbox-manager\">Initializing Installation with Oracle VM VirtualBox Manager<\/a><\/li>\n<li><a href=\"#partition-the-hard-disk\">Partition the hard disk<\/a><\/li>\n<li><a href=\"#bootstrap-arch-linux\">Bootstrap Arch Linux<\/a><\/li>\n<li><a href=\"#setup-hostname-and-networking\">Setup hostname and networking<\/a><\/li>\n<li><a href=\"#install-the-bootloader\">Install the Bootloader<\/a><\/li>\n<li><a href=\"#boot-into-the-installed-archlinux-operating-system\">Boot into the installed ArchLinux operating system<\/a>\n<\/li><\/ol><\/li>\n<\/ol><\/div><\/div><p>Arch Linux is a Linux-based operating system that is designed for i689 and x86-64 computers. Its unique package manager is responsible for providing updates to the latest software applications using \u201cpacman\u201d with complete tracking. Pacman is the package manager that is used to install, update, and remove the software packages. It is designed entirely for free and open-source software, along with the support from the Linux community.<\/p><p>Arch Linux is also popular for having a comprehensive documentation in form of the community wiki known as ArchWiki. This Linux operating system is based on binary packages that are targeted for i832, 64-bit, and 32-bit systems and optimized for the best performance on modern hardware systems.<\/p><p>You can install Arch Linux directly to your home computer by following this\u00a0guide but you can also install it on a virtual machine on your Windows computer by using VirtualBox.<\/p><h2 id=\"arch-linux-repositories\">Arch Linux Repositories<\/h2><p>To install Arch Linux on Virtual Box, you must know the basic repositories of this Linux-based operating system. A repository is a storage location from where the software packages are retrieved during the installation process. There are multiple repositories available for Arch Linux, which are accessible via pacman and maintained by package maintainers. Here is a list of some of the basic repositories used to install Arch Linux on Virtual Box:<\/p><ul><li>\n<p>The <b>core <\/b>repository contains all the packages that are needed to setup\u00a0the base system like booting Arch Linux and building packages.<\/p>\n<\/li>\n<li>\n<p>The <b>extra <\/b>repository contains extra packages that do not fit in the core involving desktop environment.<\/p>\n<\/li>\n<li>\n<p>The <b>community <\/b>repository<b> <\/b>has packages that are adopted by trusted Linux community users, and most of them will transfer to the core or extra repository.<\/p>\n<\/li>\n<li>\n<p>The <b>Multilib <\/b>repository contains 32-bit software and libraries for 32-bit application installation on 64-bit system.<\/p>\n<\/li>\n<li>\n<p>The <b>testing <\/b>repository contains packages that are destined for core or extra repositories.<\/p>\n<\/li>\n<li>\n<p>The <b>community-testing<\/b> repository is for the Linux community.<\/p>\n<\/li>\n<li>\n<p>The <b>multilib testing <\/b>repository<b> <\/b>is similar to the testing repository, but for multilib candidates.<\/p>\n<\/li>\n<li>\n<p>The <b>gnome-unstable <\/b>repository has the latest GNOME desktop environment.<\/p>\n<\/li>\n<li>\n<p>The <b>kde-unstable <\/b>repository contains the latest KDE software before they are been released.<\/p>\n<\/li>\n<\/ul><h2 id=\"install-arch-linux-on-virtual-box\">Install Arch Linux on Virtual Box<\/h2><p>To install Arch Linux on Virtual Box, you must have the latest Arch Linux version that you can download from their <a href=\"https:\/\/www.archlinux.org\/\">official website<\/a>. You can pick either the direct download option or torrent download, which is on a secure server. Before the installation, make sure you have 20 MB of free disk space and 1 GB of RAM in your system.<\/p><p><a class=\"fancybox\" id=\"img-1212\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/1212.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/1212.png\" alt=\"Download Arch Linux\" width=\"550\" height=\"246\" \/><\/a><\/p><h3 id=\"initializing-installation-with-oracle-vm-virtualbox-manager\">Initializing Installation with Oracle VM VirtualBox Manager<\/h3><p>Open the Oracle VM VirtualBox manager, click on new, and type in the name of the new operating system you want to create; in this case, it is Arch Linux. The system will automatically pick up the type and version of the Arch Linux, based on your system\u2019s configuration. Click on <i>next<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-12121\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12121.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12121.png\" alt=\"Open VirtualBox\" width=\"550\" height=\"420\" \/><\/a><\/p><p>Allocate the desired \u00a0RAM size to your new operating system, which is ideally 1024 MB. Click on <i>next <\/i>and then click on <i>create<\/i>, to create a virtual disk now.<\/p><p><a class=\"fancybox\" id=\"img-12122\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12122.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12122.png\" alt=\"Set RAM size\" width=\"550\" height=\"423\" \/><\/a><\/p><p>On the next page, you will be asked to select the type of hard disk file you want for your new operating system. Select VirtualBox Disk Image usually.<\/p><p><a class=\"fancybox\" id=\"img-12123\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12123.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12123.png\" alt=\"Select disk file type\" width=\"550\" height=\"417\" \/><\/a><\/p><p>Choose dynamically allocated and click <i>next<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-12124\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12124.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12124.png\" alt=\"Allocate disk space dynamically\" width=\"550\" height=\"413\" \/><\/a><\/p><p>Allocate 20 GB hard disk file location and size.<\/p><p><a class=\"fancybox\" id=\"img-12125\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12125.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12125.png\" alt=\"Choose a hard disk size of 20GB\" width=\"550\" height=\"427\" \/><\/a><\/p><p>Now you can see that your Arch Linux operating system is created. Now you can click <i>start<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-12126\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12126.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12126.png\" alt=\"Start the Virtual Machine\" width=\"550\" height=\"422\" \/><\/a><\/p><p>Click on 'browser' and select the startup disk, which you downloaded from the Arch Linux official website.<\/p><p><a class=\"fancybox\" id=\"img-12127\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12127.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12127.png\" alt=\"Choose Arch Linux Disk image\" width=\"550\" height=\"429\" \/><\/a><\/p><p><a class=\"fancybox\" id=\"img-12128\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12128.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12128.png\" alt=\"Browse for disk image file\" width=\"550\" height=\"390\" \/><\/a><\/p><p>Click on <i>start<\/i> and then open the full-screen view.<\/p><p><a class=\"fancybox\" id=\"img-12129\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12129.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12129.png\" alt=\"Open the full-screen view\" width=\"550\" height=\"422\" \/><\/a><\/p><p><a class=\"fancybox\" id=\"img-121210\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121210.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121210.png\" alt=\"Start the Arch Linux VM\" width=\"550\" height=\"425\" \/><\/a><\/p><h2><b>Booting to Install Arch Linux on Virtual Box<\/b><\/h2><p>You will see the first opening page of Arch Linux. Click on \u201cBoot Arch Linux (x86_64), which is for 64-bit or click on \u201cBoot Arch Linux (i686)\u201d, which is for 32-bit.<\/p><p><a class=\"fancybox\" id=\"img-121211\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121211.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121211.png\" alt=\"Choose to boot Arch Linux\" width=\"550\" height=\"414\" \/><\/a><\/p><p>As soon as you click on the first option, the system will start booting. While it completes the temporary boot, we are basically moving into the live version and are logged in as root user.<\/p><p><a class=\"fancybox\" id=\"img-121212\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121212.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121212.png\" alt=\"Booting into Arch Linux live version\" width=\"550\" height=\"311\" \/><\/a><\/p><p>Check the Internet connection by typing in the following command.<\/p><pre spellcheck=\"false\"><i>ping google.com <\/i><\/pre><p>The word ping stands for packet internet gopher. You will soon see the response that means Arch Linux has activated the Internet connection. This is essential to perform certain installation steps.<\/p><p><a class=\"fancybox\" id=\"img-121213\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121213.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121213.png\" alt=\"Test internet connectivity with ping\" width=\"550\" height=\"298\" \/><\/a><\/p><p>Clear the command by typing<\/p><pre spellcheck=\"false\"><i>clear<\/i><\/pre><p>Before we start the installation, you should partition your drive. Type <i># fdisk \u2013 l\u00a0<\/i>and you will see the current system\u2019s disk partition. Focus on the 20 GB hard drives that you allocated to Arch in the beginning.<\/p><pre spellcheck=\"false\">fsdisk -l<\/pre><p><a class=\"fancybox\" id=\"img-121214\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121214.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121214.png\" alt=\"Partition the harddisk\" width=\"550\" height=\"296\" \/><\/a><\/p><h3 id=\"partition-the-hard-disk\">Partition the hard disk<\/h3><p>We are going to partition this 20 GB space into three partitions. The first one is the primary root partition that will be of 10 GB. The second will be the\u00a0swap partition, which will be twice the initial RAM allocation that will be 2048 MB. The third will be the logical partition that will be 8 GB allocated.<\/p><p><a class=\"fancybox\" id=\"img-121215\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121215.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121215.png\" alt=\"Create 3 disk partitions\" width=\"550\" height=\"296\" \/><\/a><\/p><p>Type the command:<\/p><pre spellcheck=\"false\"><i>cfdisk<\/i><\/pre><p>\u00a0You will see gpt, dos, sgi, and sun. Select the\u00a0<i>dos <\/i>option and press <i>enter<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-121216\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121216.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121216.png\" alt=\"Create a partition of type DOS\" width=\"550\" height=\"307\" \/><\/a><\/p><p>Here you will see the main disk space, which is 20 GB. To change this, press &lt;enter&gt; on the free space and type <strong>10G<\/strong>.<\/p><p><a class=\"fancybox\" id=\"img-121217\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121217.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121217.png\" alt=\"Choose 10GB partition size\" width=\"550\" height=\"321\" \/><\/a><\/p><p>Press <i>enter <\/i>and then click on the \u201c<strong>primary<\/strong>\u201d partition.<\/p><p><a class=\"fancybox\" id=\"img-121218\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121218.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121218.png\" alt=\"image\" \/><\/a><\/p><p>Now select the \u201c<strong>bootable<\/strong>\u201d type by pressing enter on bootable.<\/p><p><a class=\"fancybox\" id=\"img-121219\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121219.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121219.png\" alt=\"Make partition bootable.\" width=\"550\" height=\"338\" \/><\/a><\/p><p>Go to <i>write <\/i>and press enter, to\u00a0write partition to disk.<\/p><p><a class=\"fancybox\" id=\"img-121220\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121220.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121220.png\" alt=\"Write partition to disk\" width=\"550\" height=\"346\" \/><\/a><\/p><p>Then type <i>yes<\/i> to make the changes, successfully.<\/p><p><a class=\"fancybox\" id=\"img-121221\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121221.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121221.png\" alt=\"Confirm changes\" width=\"550\" height=\"354\" \/><\/a><\/p><p>Now the 10 GB partition is created. Click on <i>free space <\/i>and then enter the partition size of 2048 M.<\/p><p><a class=\"fancybox\" id=\"img-121222\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121222.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121222.png\" alt=\"Create the 2GB swap partition\" width=\"550\" height=\"338\" \/><\/a><\/p><p>Now follow the same steps to create the logical partition. Then press enter on <i>quit<\/i> and clear the command by typing<\/p><pre spellcheck=\"false\"><i>clear<\/i><\/pre><p><a class=\"fancybox\" id=\"img-121223\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121223.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121223.png\" alt=\"quit cfdisk\" width=\"550\" height=\"357\" \/><\/a><\/p><p>Format the newly partitioned disk by typing:<\/p><pre spellcheck=\"false\"><i>mkfs.ext4 \/dev\/sda1<\/i><\/pre><p>Here\u00a0<i>sda1 <\/i>is the partition name. Repeat this for the\u00a0second drive by typing:<\/p><pre spellcheck=\"false\"><i>mkfs.ext4 \/dev\/sda3 <\/i><\/pre><p>and for the\u00a0swap partition,<\/p><pre spellcheck=\"false\"><i>mkswap\u00a0\/dev\/sda2<\/i><\/pre><p><a class=\"fancybox\" id=\"img-121224\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121224.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121224.png\" alt=\"Format the swap partition with mkswap\" width=\"550\" height=\"312\" \/><\/a><\/p><p>Activate the swap by typing:<\/p><pre spellcheck=\"false\"><i>swapon\u00a0\/dev\/sda2 <\/i><\/pre><p>and clear the command by typing:<\/p><pre spellcheck=\"false\"><i>clear<\/i><\/pre><p><a class=\"fancybox\" id=\"img-121225\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121225.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121225.png\" alt=\"Enable swap\" width=\"550\" height=\"327\" \/><\/a><\/p><p>Mount the primary partition to start the installation part by typing:<\/p><pre spellcheck=\"false\"><i>mount \/dev\/sda1 \/mnt<br \/>mkdir \/mnt\/home<br \/>mount \/dev\/sda3 \/mnt\/home<\/i><\/pre><p><a class=\"fancybox\" id=\"img-121226\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121226.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121226.png\" alt=\"Mount the partitions\" width=\"550\" height=\"280\" \/><\/a><\/p><h3 id=\"bootstrap-arch-linux\">Bootstrap Arch Linux<\/h3><p>Bootstrap the system by typing:<\/p><pre spellcheck=\"false\"><i>pacstrap \/mnt base base-devel<\/i><\/pre><p>You can see that it is synchronizing the data packages.<\/p><p><a class=\"fancybox\" id=\"img-121227\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121227.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121227.png\" alt=\"Bootstrap Arch Linux\" width=\"550\" height=\"304\" \/><\/a><\/p><p>The installation will start and will take a few minutes.<\/p><p><a class=\"fancybox\" id=\"img-121228\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121228.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121228.png\" alt=\"Arch Linux installation has been started\" width=\"550\" height=\"346\" \/><\/a><\/p><p>After the base installation create the fstab file by tying:<\/p><pre spellcheck=\"false\"><i>genfstab \/mnt&gt;&gt; \/mnt\/etc\/fstab <\/i><\/pre><p><a class=\"fancybox\" id=\"img-121229\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121229.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121229.png\" alt=\"Generating \/etc\/fstab\" width=\"550\" height=\"284\" \/><\/a><\/p><p>Configure locale, time and root password.<\/p><p>Change the system root to the Arch Linux installation directory by typing:<\/p><pre spellcheck=\"false\"><i>arch-chroot \/mnt \/bin\/bash<\/i><\/pre><p>Now to configure the language settings:<\/p><pre spellcheck=\"false\"><i>nano \/etc\/locale.gen<\/i><\/pre><p><a class=\"fancybox\" id=\"img-121230\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121230.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121230.png\" alt=\"Set language in Arch Linux\" width=\"550\" height=\"298\" \/><\/a><\/p><p>Select the following language configuration (en_US.UTF-8 UTF-8) by deleting the # and pressing <i>control + x, <\/i>press <i>y<\/i>, and press <i>enter<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-121231\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121231.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121231.png\" alt=\"select language\" width=\"550\" height=\"309\" \/><\/a><\/p><p>Now activate it by typing:<\/p><pre spellcheck=\"false\"><i>locale-gen<\/i><\/pre><p>and press <i>enter<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-121232\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/121232.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/121232.png\" alt=\"Generate the locales in Arch Linux\" width=\"550\" height=\"275\" \/><\/a><\/p><p>Create the \u00a0\/etc\/locale.conf file by typing:<\/p><pre spellcheck=\"false\"><i>nano \/etc\/locale.conf <\/i><\/pre><p>then press <i>enter<\/i>. Now you can add your language to the system by adding this line to the file:<\/p><pre spellcheck=\"false\">LANG=en_US.UTF-8<\/pre><p>\u00a0Then press <i>control + x <\/i>and press <i>y<\/i>, and press <i>enter<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-1\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/1.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/1.png\" alt=\"Set default language\" width=\"550\" height=\"216\" \/><\/a><\/p><p>Synchronize the zone information by typing:<\/p><pre spellcheck=\"false\">ls \/usr\/share\/zoneinfo<\/pre><p> and you will find a\u00a0list of all the zones of the world.<\/p><p><a class=\"fancybox\" id=\"img-11\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/11.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/11.png\" alt=\"Set System language default\" width=\"550\" height=\"293\" \/><\/a><\/p><p>To select your zone, type:<\/p><pre spellcheck=\"false\"><i>ln \u2013s \/usr\/share\/zoneinfo\/Asia\/Kolkata \/etc\/localtime <\/i><\/pre><p>or you can select any name from the below list.<\/p><p><a class=\"fancybox\" id=\"img-12\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/12.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/12.png\" alt=\"image\" \/><\/a><\/p><p>Set the time standard using the command.<\/p><pre spellcheck=\"false\">hwclock --systohc\u00a0--utc<\/pre><p>And the hardware clock is synchronized.<\/p><p><a class=\"fancybox\" id=\"img-13\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/13.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/13.png\" alt=\"Set time\" width=\"550\" height=\"258\" \/><\/a><\/p><p>Set the root user password by typing:<\/p><pre spellcheck=\"false\"><i>passwd <\/i><\/pre><p>And press <i>enter<\/i>. Then type your password and press enter.<\/p><p><a class=\"fancybox\" id=\"img-14\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/14.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/14.png\" alt=\"Set the root password\" width=\"550\" height=\"219\" \/><\/a><\/p><h3 id=\"setup-hostname-and-networking\">Setup hostname and networking<\/h3><p>Enter the host name of the network by typing:<\/p><pre spellcheck=\"false\"><i>nano \/etc\/hostname<\/i><\/pre><p><i> <\/i>and press <i>enter<\/i>. Now type any name and then press <i>control + x <\/i>and press <i>y<\/i>, and press <i>enter<\/i>.<\/p><p><a class=\"fancybox\" id=\"img-15\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/15.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/15.png\" alt=\"Set the hostname\" width=\"550\" height=\"325\" \/><\/a><\/p><p>Enable the dhcpcd by typing:<\/p><pre spellcheck=\"false\"><i>systemctl enable dhcpcd<\/i><\/pre><p>and it will be started at next boot time to fetch an IP address automatically.<\/p><p><a class=\"fancybox\" id=\"img-16\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/16.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/16.png\" alt=\"Enable dhcpd\" width=\"550\" height=\"217\" \/><\/a><\/p><h3 id=\"install-the-bootloader\">Install the Bootloader<\/h3><p>The final step, initiate the grub installation. Type:<\/p><pre spellcheck=\"false\"><i>pacman \u2013S grub os-prober<\/i><\/pre><p>then press <i>y <\/i>and the download part will begin.<\/p><p><a class=\"fancybox\" id=\"img-17\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/17.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/17.png\" alt=\"Configure grub\" width=\"550\" height=\"261\" \/><\/a><\/p><p>Install the grub boot loader to the hard disk by typing:<\/p><pre spellcheck=\"false\"><i>grub-install \/dev\/sda<\/i><\/pre><p>and configure it:<\/p><pre spellcheck=\"false\"><i>grub-mkconfig \u2013o \/boot\/grub\/grub.cfg<\/i><\/pre><p><a class=\"fancybox\" id=\"img-18\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/18.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/18.png\" alt=\"Install and configure grub boot loader\" width=\"550\" height=\"253\" \/><\/a><\/p><p>Finally, exit from chroot and then reboot the system by typing:<\/p><pre spellcheck=\"false\"><i>exit<br \/>reboot<\/i><\/pre><p>and press <i>enter.<\/i><\/p><p><a class=\"fancybox\" id=\"img-19\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/19.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/19.png\" alt=\"Reboot the system\" width=\"550\" height=\"249\" \/><\/a><\/p><h3 id=\"boot-into-the-installed-archlinux-operating-system\">Boot into the installed ArchLinux operating system<\/h3><p>Choose \u201cBoot Existing OS\u201d to boot Arch Linux on Virtual Box.<\/p><p><a class=\"fancybox\" id=\"img-110\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/110.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/110.png\" alt=\"Boot Arch Linux\" width=\"550\" height=\"376\" \/><\/a><\/p><p>Login with your root name and password, and you will enter your new Arch Linux operating system.<\/p><p><a class=\"fancybox\" id=\"img-111\" href=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/big\/111.png\"><img src=\"https:\/\/www.howtoforge.com\/images\/install_arch_linux_on_virtual_box\/111.png\" alt=\"Arch Linux installed successfully\" width=\"550\" height=\"246\" \/><\/a><\/p>","created_at":"2019-05-03T15:36:05+0000","updated_at":"2019-05-03T15:36:13+0000","published_by":["Dimitris"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"www.howtoforge.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":933,"title":"A packaging tutorial for Guix \u2014 2018 \u2014 Blog \u2014 GNU\u00a0Guix","url":"https:\/\/www.gnu.org\/software\/guix\/blog\/2018\/a-packaging-tutorial-for-guix\/","content":"<p>GNU Guix stands out as the <em>hackable<\/em> package manager, mostly because it uses\n<a href=\"https:\/\/www.gnu.org\/software\/guile\/\">GNU Guile<\/a>, a powerful high-level programming language, one of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Scheme_(programming_language)\">Scheme<\/a>\ndialects from the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lisp_(programming_language)\">Lisp family<\/a>.<\/p><p>Package definitions are also written in Scheme, which empowers Guix in some very\nunique ways, unlike most other package managers that use shell scripts or\nsimple languages.<\/p><ul><li><p>Use functions, structures, macros and all of Scheme expressiveness for your\npackage definitions.<\/p><\/li><li><p>Inheritance makes it easy to customize a package by inheriting from it and\nmodifying only what is needed.<\/p><\/li><li><p>Batch processing: the whole package collection can be parsed, filtered and\nprocessed.  Building a headless server with all graphical interfaces stripped\nout?  It's possible.  Want to rebuild everything from source using specific\ncompiler optimization flags?  Pass the <code>#:make-flags \"...\"<\/code> argument to the\nlist of packages.  It wouldn't be a stretch to think <a href=\"https:\/\/wiki.gentoo.org\/wiki\/USE_flag\">Gentoo USE flags<\/a> here,\nbut this goes even further: the changes don't have to be thought out\nbeforehand by the packager, they can be <em>programmed<\/em> by the user!<\/p><\/li><\/ul><p>The following tutorial covers all the basics around package creation with Guix.\nIt does not assume much knowledge of the Guix system nor of the Lisp language.\nThe reader is only expected to be familiar with the command line and to have some\nbasic programming knowledge.<\/p><p>The <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Defining-Packages.html\">\u201cDefining Packages\u201d section of the manual<\/a> introduces the basics of Guix\npackaging.  In the following section, we will partly go over those basics again.<\/p><p><code>GNU hello<\/code> is a dummy project that serves as an idiomatic example for\npackaging.  It uses the GNU build system (<code>.\/configure &amp;&amp; make &amp;&amp; make install<\/code>).\nGuix already provides a package definition which is a perfect example to start\nwith.  You can look up its declaration with <code>guix edit hello<\/code> from the\ncommand line.  Let's see how it looks:<\/p><pre>(define-public hello\n  (package\n    (name \"hello\")\n    (version \"2.10\")\n    (source (origin\n              (method url-fetch)\n              (uri (string-append \"mirror:\/\/gnu\/hello\/hello-\" version\n                                  \".tar.gz\"))\n              (sha256\n               (base32\n                \"0ssi1wpaf7plaswqqjwigppsg5fyh99vdlb9kzl7c9lng89ndq1i\"))))\n    (build-system gnu-build-system)\n    (synopsis \"Hello, GNU world: An example GNU package\")\n    (description\n     \"GNU Hello prints the message \\\"Hello, world!\\\" and then exits.  It\nserves as an example of standard GNU coding practices.  As such, it supports\ncommand-line arguments, multiple languages, and so on.\")\n    (home-page \"https:\/\/www.gnu.org\/software\/hello\/\")\n    (license gpl3+)))<\/pre><p>As you can see, most of it is rather straightforward.  But let's review the\nfields together:<\/p><ul><li><strong>name:<\/strong> The project name.  Using Scheme conventions, we prefer to keep it\nlower case, without underscore and using dash-separated words.<\/li><li><p><strong>source:<\/strong> This field contains a description of the source code origin.  The\n<code>origin<\/code> record contains these fields:<\/p><ol><li>The method, here <code>url-fetch<\/code> to download via HTTP\/FTP, but other methods\nexist, such as <code>git-fetch<\/code> for Git repositories.<\/li><li>The URI, which is typically some <code>https:\/\/<\/code> location for <code>url-fetch<\/code>.  Here\nthe special <code>mirror:\/\/gnu<\/code> refers to a set of well known locations, all of\nwhich can be used by Guix to fetch the source, should some of them fail.<\/li><li>The <code>sha256<\/code> checksum of the requested file.  This is essential to ensure\nthe source is not corrupted.  Note that Guix works with base32 strings,\nhence the call to the <code>base32<\/code> function.<\/li><\/ol><\/li><li><strong>build-system:<\/strong> This is where the power of abstraction provided by the Scheme\nlanguage really shines: in this case, the <code>gnu-build-system<\/code>\nabstracts away the famous <code>.\/configure &amp;&amp; make &amp;&amp; make install<\/code> shell invocations.  Other build systems include the\n<code>trivial-build-system<\/code> which does not do anything and requires\nfrom the packager to program all the build steps, the\n<code>python-build-system<\/code>, the <code>emacs-build-system<\/code>, <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Build-Systems.html\">and many\nmore<\/a>.<\/li><li><strong>synopsis:<\/strong> It should be a concise summary of what the package\ndoes.  For many packages a tagline from the\nproject's home page can be used as the synopsis.<\/li><li><strong>description:<\/strong> Same as for the synopsis, it's fine to re-use the project\ndescription from the homepage.  Note that Guix uses Texinfo\nsyntax.<\/li><li><strong>home-page:<\/strong> Use HTTPS if available.<\/li><li><strong>license:<\/strong> See <code>guix\/licenses.scm<\/code> in the project source for a full list.<\/li><\/ul><p>Time to build our first package!  Nothing fancy here for now: we will stick to a\ndummy \"my-hello\", a copy of the above declaration.<\/p><p>As with the ritualistic \"Hello World\" taught with most programming languages,\nthis will possibly be the most \"manual\" approach.  We will work out an ideal\nsetup later; for now we will go the simplest route.<\/p><p>Save the following to a file <code>my-hello.scm<\/code>.<\/p><pre>(use-modules (guix packages)\n             (guix download)\n             (guix build-system gnu)\n             (guix licenses))\n(package\n  (name \"my-hello\")\n  (version \"2.10\")\n  (source (origin\n            (method url-fetch)\n            (uri (string-append \"mirror:\/\/gnu\/hello\/hello-\" version\n                                \".tar.gz\"))\n            (sha256\n             (base32\n              \"0ssi1wpaf7plaswqqjwigppsg5fyh99vdlb9kzl7c9lng89ndq1i\"))))\n  (build-system gnu-build-system)\n  (synopsis \"Hello, Guix world: An example custom Guix package\")\n  (description\n   \"GNU Hello prints the message \\\"Hello, world!\\\" and then exits.  It\nserves as an example of standard GNU coding practices.  As such, it supports\ncommand-line arguments, multiple languages, and so on.\")\n  (home-page \"https:\/\/www.gnu.org\/software\/hello\/\")\n  (license gpl3+))<\/pre><p>We will explain the extra code in a moment.<\/p><p>Feel free to play with the different values of the various fields.  If you\nchange the source, you'll need to update the checksum.  Indeed, Guix refuses to\nbuild anything if the given checksum does not match the computed checksum of the\nsource code.  To obtain the correct checksum of the package declaration, we\nneed to download the source, compute the sha256 checksum and convert it to\nbase32.<\/p><p>Thankfully, Guix can automate this task for us; all we need is to provide the\nURI:<\/p><pre class=\"language-sh\">$ guix download mirror:\/\/gnu\/hello\/hello-2.10.tar.gz\nStarting download of \/tmp\/guix-file.JLYgL7\nFrom https:\/\/ftpmirror.gnu.org\/gnu\/hello\/hello-2.10.tar.gz...\nfollowing redirection to `https:\/\/mirror.ibcp.fr\/pub\/gnu\/hello\/hello-2.10.tar.gz'...\n \u202610.tar.gz  709KiB                                 2.5MiB\/s 00:00 [##################] 100.0%\n\/gnu\/store\/hbdalsf5lpf01x4dcknwx6xbn6n5km6k-hello-2.10.tar.gz\n0ssi1wpaf7plaswqqjwigppsg5fyh99vdlb9kzl7c9lng89ndq1i<\/pre><p>In this specific case the output tells us which mirror was chosen.\nIf the result of the above command is not the same as in the above snippet,\nupdate your <code>my-hello<\/code> declaration accordingly.<\/p><p>Note that GNU package tarballs come with an OpenPGP signature, so you\nshould definitely check the signature of this tarball with <code>gpg<\/code> to\nauthenticate it before going further:<\/p><pre class=\"language-sh\">$ guix download mirror:\/\/gnu\/hello\/hello-2.10.tar.gz.sig\nStarting download of \/tmp\/guix-file.03tFfb\nFrom https:\/\/ftpmirror.gnu.org\/gnu\/hello\/hello-2.10.tar.gz.sig...\nfollowing redirection to `https:\/\/ftp.igh.cnrs.fr\/pub\/gnu\/hello\/hello-2.10.tar.gz.sig'...\n \u2026.tar.gz.sig  819B                                                                                                                       1.2MiB\/s 00:00 [##################] 100.0%\n\/gnu\/store\/rzs8wba9ka7grrmgcpfyxvs58mly0sx6-hello-2.10.tar.gz.sig\n0q0v86n3y38z17rl146gdakw9xc4mcscpk8dscs412j22glrv9jf\n$ gpg --verify \/gnu\/store\/rzs8wba9ka7grrmgcpfyxvs58mly0sx6-hello-2.10.tar.gz.sig \/gnu\/store\/hbdalsf5lpf01x4dcknwx6xbn6n5km6k-hello-2.10.tar.gz\ngpg: Signature made Sun 16 Nov 2014 01:08:37 PM CET\ngpg:                using RSA key A9553245FDE9B739\ngpg: Good signature from \"Sami Kerola &lt;kerolasa@iki.fi&gt;\" [unknown]\ngpg:                 aka \"Sami Kerola (http:\/\/www.iki.fi\/kerolasa\/) &lt;kerolasa@iki.fi&gt;\" [unknown]\ngpg: WARNING: This key is not certified with a trusted signature!\ngpg:          There is no indication that the signature belongs to the owner.\nPrimary key fingerprint: 8ED3 96E3 7E38 D471 A005  30D3 A955 3245 FDE9 B739<\/pre><p>You can then happily run<\/p><pre class=\"language-sh\">$ guix package --install-from-file=my-hello.scm<\/pre><p>You should now have <code>my-hello<\/code> in your profile!<\/p><pre class=\"language-sh\">$ guix package --list-installed=my-hello\nmy-hello    2.10    out\n\/gnu\/store\/f1db2mfm8syb8qvc357c53slbvf1g9m9-my-hello-2.10<\/pre><p>We've gone as far as we could without any knowledge of Scheme.  Now is the right\ntime to introduce the minimum we need from the language before we can proceed.<\/p><p>As we've seen above, basic packages don't require much Scheme knowledge, if\nat all.  But as you progress and your desire to write more and more complex\npackages grows, it will become both necessary and empowering to hone your Lisper\nskills.<\/p><p>Since an extensive Lisp course is very much out of the scope of this tutorial,\nwe will only cover some basics here.<\/p><p>Guix uses the Guile implementation of Scheme.  To start playing with the\nlanguage, install it with <code>guix package --install guile<\/code> and start a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Read%E2%80%93eval%E2%80%93print_loop\">REPL<\/a> by\nrunning <code>guile<\/code> from the command line.<\/p><p>Alternatively you can also run <code>guix environment --ad-hoc guile -- guile<\/code> if\nyou'd rather not have Guile installed in your user profile.<\/p><p>In the following examples we use the <code>&gt;<\/code> symbol to denote the REPL prompt, that\nis, the line reserved for user input.  See <a href=\"https:\/\/www.gnu.org\/software\/guile\/manual\/en\/html_node\/Using-Guile-Interactively.html\">the Guile manual<\/a> for more details on\nthe REPL.<\/p><ul><li><p>Scheme syntax boils down to a tree of expressions (or <em>s-expression<\/em> in Lisp\nlingo).  An expression can be a literal such as numbers and strings, or a\ncompound which is a parenthesized list of compounds and literals.  <code>#t<\/code> and\n<code>#f<\/code> stand for the booleans \"true\" and \"false\", respectively.<\/p><p>Examples of valid expressions:<\/p><pre>&gt; \"Hello World!\"\n\"Hello World!\"\n&gt; 17\n17\n&gt; (display (string-append \"Hello \" \"Guix\" \"\\n\"))\n\"Hello Guix!\"<\/pre><\/li><li><p>This last example is a function call embedded in another function call.  When\na parenthesized expression is evaluated, the first term is the function and\nthe rest are the arguments passed to the function.  Every function returns the\nlast evaluated expression as value.<\/p><\/li><li><p>Anonymous functions are declared with the <code>lambda<\/code> term:<\/p><pre>&gt; (lambda (x) (* x x))\n#&lt;procedure 120e348 at &lt;unknown port&gt;:24:0 (x)&gt;<\/pre><p>The above lambda returns the square of its argument.  Since everything is an\nexpression, the <code>lambda<\/code> expression returns an anonymous function, which can\nin turn be applied to an argument:<\/p><pre>&gt; ((lambda (x) (* x x)) 3)\n9<\/pre><\/li><li><p>Anything can be assigned a global name with <code>define<\/code>:<\/p><pre>&gt; (define a 3)\n&gt; (define square (lambda (x) (* x x)))\n&gt; (square a)\n9<\/pre><\/li><li><p>Procedures can be defined more concisely with the following syntax:<\/p><pre>(define (square x) (* x x))<\/pre><\/li><li><p>A list structure can be created with the <code>list<\/code> procedure:<\/p><pre>&gt; (list 2 a 5 7)\n(2 3 5 7)<\/pre><\/li><li><p>The <em>quote<\/em> disables evaluation of a parenthesized expression: the first term\nis not called over the other terms.  Thus it effectively returns a list of\nterms.<\/p><pre>&gt; '(display (string-append \"Hello \" \"Guix\" \"\\n\"))\n(display (string-append \"Hello \" \"Guix\" \"\\n\"))\n&gt; '(2 a 5 7)\n(2 a 5 7)<\/pre><\/li><li><p>The <em>quasiquote<\/em> disables evaluation of a parenthesized expression until a\ncomma re-enables it.  Thus it provides us with fine-grained control over what\nis evaluated and what is not.<\/p><pre>&gt; `(2 a 5 7 (2 ,a 5 ,(+ a 4)))\n(2 a 5 7 (2 3 5 7))<\/pre><p>Note that the above result is a list of mixed elements: numbers, symbols (here\n<code>a<\/code>) and the last element is a list itself.<\/p><\/li><li><p>Multiple variables can be named locally with <code>let<\/code>:<\/p><pre>&gt; (define x 10)\n&gt; (let ((x 2)\n        (y 3))\n    (list x y))\n(2 3)\n&gt; x\n10\n&gt; y\nERROR: In procedure module-lookup: Unbound variable: y<\/pre><p>Use <code>let*<\/code> to allow later variable declarations to refer to earlier\ndefinitions.<\/p><pre>&gt; (let* ((x 2)\n         (y (* x 3)))\n    (list x y))\n(2 6)<\/pre><\/li><li><p>The keyword syntax is <code>#:<\/code>, it is used to create unique identifiers.  See also\nthe <a href=\"https:\/\/www.gnu.org\/software\/guile\/manual\/en\/html_node\/Keywords.html\">Keywords section in the Guile manual<\/a>.<\/p><\/li><li><p>The percentage <code>%<\/code> is typically used for read-only global variables in the\nbuild stage.  Note that it is merely a convention, like <code>_<\/code> in C.  Scheme Lisp\ntreats <code>%<\/code> exactly the same as any other letter.<\/p><\/li><li><p>Modules are created with <code>define-module<\/code>.  For instance<\/p><pre>(define-module (guix build-system ruby)\n  #:use-module (guix store)\n  #:export (ruby-build\n            ruby-build-system))<\/pre><p>defines the module <code>ruby<\/code> which must be located in\n<code>guix\/build-system\/ruby.scm<\/code> somewhere in <code>GUILE_LOAD_PATH<\/code>.  It depends on\nthe <code>(guix store)<\/code> module and it exports two symbols, <code>ruby-build<\/code> and\n<code>ruby-build-system<\/code>.<\/p><\/li><\/ul><p>For a more detailed introduction, check out <a href=\"http:\/\/www.troubleshooters.com\/codecorn\/scheme_guile\/hello.htm\">Scheme at a Glance<\/a>, by Steve Litt.<\/p><p>One of the reference Scheme books is the seminal <em>Structure and Interpretation\nof Computer Programs<\/em>, by Harold Abelson and Gerald Jay Sussman, with Julie\nSussman.  You'll find a free copy <a href=\"https:\/\/mitpress.mit.edu\/sites\/default\/files\/sicp\/index.html\">online<\/a>, together with <a href=\"https:\/\/ocw.mit.edu\/courses\/electrical-engineering-and-computer-science\/6-001-structure-and-interpretation-of-computer-programs-spring-2005\/video-lectures\/\">videos of the lectures\nby the authors<\/a>.  The book is available in Texinfo format as the <code>sicp<\/code> Guix\npackage.  Go ahead, run <code>guix package --install sicp<\/code> and start reading with\n<code>info sicp<\/code> (or with the Emacs Info reader).  An unofficial ebook <a href=\"https:\/\/sarabander.github.io\/sicp\/\">is also\navailable<\/a>.<\/p><p>You'll find more books, tutorials and other resources at <a href=\"https:\/\/schemers.org\/\">https:\/\/schemers.org\/<\/a>.<\/p><p>Now that we know some Scheme basics we can detail the different possible setups\nfor working on Guix packages.<\/p><p>There are several ways to set up a Guix packaging environment.<\/p><p>We recommend you work directly on the Guix source checkout since it makes it\neasier for everyone to contribute to the project.<\/p><p>But first, let's look at other possibilities.<\/p><h3>Local file<\/h3><p>This is what we previously did with <code>my-hello<\/code>.  With the Scheme basics we've\ncovered, we are now able to explain the leading chunks.  As stated in <code>guix package --help<\/code>:<\/p><pre class=\"language-sh\">-f, --install-from-file=FILE\n                       install the package that the code within FILE\n                       evaluates to<\/pre><p>Thus the last expression <em>must<\/em> return a package, which is the case in our\nearlier example.<\/p><p>The <code>use-modules<\/code> expression tells which of the modules we need in the file.\nModules are a collection of values and procedures.  They are commonly called\n\"libraries\" or \"packages\" in other programming languages.<\/p><h3>GUIX_PACKAGE_PATH<\/h3><p><em>Note: Starting from Guix 0.16, the more flexible Guix \"channels\" are the\npreferred way and supersede <code>GUIX_PACKAGE_PATH<\/code>.  See below.<\/em><\/p><p>It can be tedious to specify the file from the command line instead of simply\ncalling <code>guix package --install my-hello<\/code> as you would do with the official\npackages.<\/p><p>Guix makes it possible to streamline the process by adding as many \"package\ndeclaration paths\" as you want.<\/p><p>Create a directory, say <code>~.\/guix-packages<\/code> and add it to the <code>GUIX_PACKAGE_PATH<\/code>\nenvironment variable:<\/p><pre class=\"language-sh\">$ mkdir ~\/guix-packages\n$ export GUIX_PACKAGE_PATH=~\/guix-packages<\/pre><p>To add several directories, separate them with a colon (<code>:<\/code>).<\/p><p>Our previous <code>my-hello<\/code> needs some adjustments though:<\/p><pre>(define-module (my-hello)\n  #:use-module (guix licenses)\n  #:use-module (guix packages)\n  #:use-module (guix build-system gnu)\n  #:use-module (guix download))\n(define-public my-hello\n  (package\n    (name \"my-hello\")\n    (version \"2.10\")\n    (source (origin\n              (method url-fetch)\n              (uri (string-append \"mirror:\/\/gnu\/hello\/hello-\" version\n                                  \".tar.gz\"))\n              (sha256\n               (base32\n                \"0ssi1wpaf7plaswqqjwigppsg5fyh99vdlb9kzl7c9lng89ndq1i\"))))\n    (build-system gnu-build-system)\n    (synopsis \"Hello, Guix world: An example custom Guix package\")\n    (description\n     \"GNU Hello prints the message \\\"Hello, world!\\\" and then exits.  It\nserves as an example of standard GNU coding practices.  As such, it supports\ncommand-line arguments, multiple languages, and so on.\")\n    (home-page \"https:\/\/www.gnu.org\/software\/hello\/\")\n    (license gpl3+)))<\/pre><p>Note that we have assigned the package value to an exported variable name with\n<code>define-public<\/code>.  This is effectively assigning the package to the <code>my-hello<\/code>\nvariable so that it can be referenced, among other as dependency of other\npackages.<\/p><p>If you use <code>guix package --install-from-file=my-hello.scm<\/code> on the above file, it\nwill fail because the last expression, <code>define-public<\/code>, does not return a\npackage.  If you want to use <code>define-public<\/code> in this use-case nonetheless, make\nsure the file ends with an evaluation of <code>my-hello<\/code>:<\/p><pre>; ...\n(define-public my-hello\n  ; ...\n  )\nmy-hello<\/pre><p>This last example is not very typical.<\/p><p>Now <code>my-hello<\/code> should be part of the package collection like all other official\npackages.  You can verify this with:<\/p><pre class=\"language-sh\">$ guix package --show=my-hello<\/pre><h3>Guix channels<\/h3><p>Guix 0.16 features channels, which is very similar to <code>GUIX_PACKAGE_PATH<\/code> but\nprovides better integration and provenance tracking.  Channels are not\nnecessarily local, they can be maintained as a public Git repository for\ninstance.  Of course, several channels can be used at the same time.<\/p><p>See the <a href=\"http:\/\/guix.info\/manual\/en\/Channels.html\">\u201cChannels\u201d section in the manual<\/a> for setup details.<\/p><h3>Direct checkout hacking<\/h3><p>Working directly on the Guix project is recommended: it reduces the friction\nwhen the time comes to submit your changes upstream to let the community benefit\nfrom your hard work!<\/p><p>Unlike most software distributions, the Guix repository holds in one place both\nthe tooling (including the package manager) and the package definitions.  This\nchoice was made so that it would give developers the flexibility to modify the\nAPI without breakage by updating all packages at the same time.  This reduces\ndevelopment inertia.<\/p><p>Check out the official <a href=\"https:\/\/git-scm.com\/\">Git<\/a> repository:<\/p><pre class=\"language-sh\">$ git clone https:\/\/git.savannah.gnu.org\/git\/guix.git<\/pre><p>In the rest of this article, we use <code>$GUIX_CHECKOUT<\/code> to refer to the location of\nthe checkout.<\/p><p>Follow the instruction from the <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Contributing.html\">\"Contributing\" chapter<\/a> in the manual to set up the\nrepository environment.<\/p><p>Once ready, you should be able to use the package definitions from the\nrepository environment.<\/p><p>Feel free to edit package definitions found in <code>$GUIX_CHECKOUT\/gnu\/packages<\/code>.<\/p><p>The <code>$GUIX_CHECKOUT\/pre-inst-env<\/code> script lets you use <code>guix<\/code> over the package\ncollection of the repository.<\/p><ul><li><p>Search packages, such as Ruby:<\/p><pre class=\"language-sh\">$ cd $GUIX_CHECKOUT\n$ .\/pre-inst-env guix package --list-available=ruby\n    ruby    1.8.7-p374      out     gnu\/packages\/ruby.scm:119:2\n    ruby    2.1.6   out     gnu\/packages\/ruby.scm:91:2\n    ruby    2.2.2   out     gnu\/packages\/ruby.scm:39:2<\/pre><\/li><li><p>Build a package, here Ruby version 2.1:<\/p><pre class=\"language-sh\">$ .\/pre-inst-env guix build --keep-failed ruby@2.1\n\/gnu\/store\/c13v73jxmj2nir2xjqaz5259zywsa9zi-ruby-2.1.6<\/pre><\/li><li><p>Install it to your user profile:<\/p><pre class=\"language-sh\">$ .\/pre-inst-env guix package --install ruby@2.1<\/pre><\/li><li><p>Check for common mistakes:<\/p><pre class=\"language-sh\">$ .\/pre-inst-env guix lint ruby@2.1<\/pre><\/li><\/ul><p>Guix strives at maintaining a high packaging standard; when contributing to the\nGuix project, remember to<\/p><p>Once you are happy with the result, you are welcome to send your contribution to\nmake it part of Guix.  This process is also detailed in the <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Contributing.html\">manual<\/a>.<\/p><p>It's a community effort so the more join in, the better Guix becomes!<\/p><p>The above \"Hello World\" example is as simple as it goes.  Packages can be more\ncomplex than that and Guix can handle more advanced scenarios.  Let's look at\nanother, more sophisticated package (slightly modified from the source):<\/p><pre>(define-module (gnu packages version-control)\n  #:use-module ((guix licenses) #:prefix license:)\n  #:use-module (guix utils)\n  #:use-module (guix packages)\n  #:use-module (guix git-download)\n  #:use-module (guix build-system cmake)\n  #:use-module (gnu packages ssh)\n  #:use-module (gnu packages web)\n  #:use-module (gnu packages pkg-config)\n  #:use-module (gnu packages python)\n  #:use-module (gnu packages compression)\n  #:use-module (gnu packages tls))\n(define-public my-libgit2\n  (let ((commit \"e98d0a37c93574d2c6107bf7f31140b548c6a7bf\")\n        (revision \"1\"))\n    (package\n      (name \"my-libgit2\")\n      (version (git-version \"0.26.6\" revision commit))\n      (source (origin\n                (method git-fetch)\n                (uri (git-reference\n                      (url \"https:\/\/github.com\/libgit2\/libgit2\/\")\n                      (commit commit)))\n                (file-name (git-file-name name version))\n                (sha256\n                 (base32\n                  \"17pjvprmdrx4h6bb1hhc98w9qi6ki7yl57f090n9kbhswxqfs7s3\"))\n                (patches (search-patches \"libgit2-mtime-0.patch\"))\n                (modules '((guix build utils)))\n                (snippet '(begin\n                            ;; Remove bundled software.\n                            (delete-file-recursively \"deps\")\n                            #t))))\n      (build-system cmake-build-system)\n      (outputs '(\"out\" \"debug\"))\n      (arguments\n       `(#:tests? #t                            ; Run the test suite (this is the default)\n         #:configure-flags '(\"-DUSE_SHA1DC=ON\") ; SHA-1 collision detection\n         #:phases\n         (modify-phases %standard-phases\n           (add-after 'unpack 'fix-hardcoded-paths\n             (lambda _\n               (substitute* \"tests\/repo\/init.c\"\n                 ((\"#!\/bin\/sh\") (string-append \"#!\" (which \"sh\"))))\n               (substitute* \"tests\/clar\/fs.h\"\n                 ((\"\/bin\/cp\") (which \"cp\"))\n                 ((\"\/bin\/rm\") (which \"rm\")))\n               #t))\n           ;; Run checks more verbosely.\n           (replace 'check\n             (lambda _ (invoke \".\/libgit2_clar\" \"-v\" \"-Q\")))\n           (add-after 'unpack 'make-files-writable-for-tests\n               (lambda _ (for-each make-file-writable (find-files \".\" \".*\")))))))\n      (inputs\n       `((\"libssh2\" ,libssh2)\n         (\"http-parser\" ,http-parser)\n         (\"python\" ,python-wrapper)))\n      (native-inputs\n       `((\"pkg-config\" ,pkg-config)))\n      (propagated-inputs\n       ;; These two libraries are in 'Requires.private' in libgit2.pc.\n       `((\"openssl\" ,openssl)\n         (\"zlib\" ,zlib)))\n      (home-page \"https:\/\/libgit2.github.com\/\")\n      (synopsis \"Library providing Git core methods\")\n      (description\n       \"Libgit2 is a portable, pure C implementation of the Git core methods\nprovided as a re-entrant linkable library with a solid API, allowing you to\nwrite native speed custom Git applications in any language with bindings.\")\n      ;; GPLv2 with linking exception\n      (license license:gpl2))))<\/pre><p>(In those cases were you only want to tweak a few fields from a package\ndefinition, you should rely on inheritance instead of copy-pasting everything.\nSee below.)<\/p><p>Let's discuss those fields in depth.<\/p><h3><code>git-fetch<\/code> method<\/h3><p>Unlike the <code>url-fetch<\/code> method, <code>git-fetch<\/code> expects a <code>git-reference<\/code> which takes\na Git repository and a commit.  The commit can be any Git reference such as\ntags, so if the <code>version<\/code> is tagged, then it can be used directly.  Sometimes\nthe tag is prefixed with a <code>v<\/code>, in which case you'd use <code>(commit (string-append \"v\" version))<\/code>.<\/p><p>To ensure that the source code from the Git repository is stored in a unique\ndirectory with a readable name we use <code>(file-name (git-file-name name version))<\/code>.<\/p><p>Note that there is also a <code>git-version<\/code> procedure that can be used to derive the\nversion when packaging programs for a specific commit.<\/p><h3>Snippets<\/h3><p>Snippets are quoted (i.e. non-evaluated) Scheme code that are a means of patching\nthe source.  They are a Guix-y alternative to the traditional <code>.patch<\/code> files.\nBecause of the quote, the code in only evaluated when passed to the Guix daemon\nfor building.<\/p><p>There can be as many snippet as needed.<\/p><p>Snippets might need additional Guile modules which can be imported from the\n<code>modules<\/code> field.<\/p><h3>Inputs<\/h3><p>First, a syntactic comment: See the quasi-quote \/ comma syntax?<\/p><pre>(native-inputs\n `((\"pkg-config\" ,pkg-config)))<\/pre><p>is equivalent to<\/p><pre>(native-inputs\n (list (list \"pkg-config\" pkg-config)))<\/pre><p>You'll mostly see the former because it's shorter.<\/p><p>There are 3 different input types.  In short:<\/p><ul><li><strong>native-inputs:<\/strong> Required for building but not runtime \u2013 installing a package\nthrough a substitute won't install these inputs.<\/li><li><strong>inputs:<\/strong> Installed in the store but not in the profile, as well as being\npresent at build time.<\/li><li><strong>propagated-inputs:<\/strong> Installed in the store and in the profile, as well as\nbeing present at build time.<\/li><\/ul><p>See <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/package-Reference.html\">the package reference in the manual<\/a> for more details.<\/p><p>The distinction between the various inputs is important: if a dependency can be\nhandled as an <em>input<\/em> instead of a <em>propagated input<\/em>, it should be done so, or\nelse it \"pollutes\" the user profile for no good reason.<\/p><p>For instance, a user installing a graphical program that depends on a\ncommand line tool might only be interested in the graphical part, so there is no\nneed to force the command line tool into the user profile.  The dependency is a\nconcern to the package, not to the user.  <em>Inputs<\/em> make it possible to handle\ndependencies without bugging the user by adding undesired executable files (or\nlibraries) to their profile.<\/p><p>Same goes for <em>native-inputs<\/em>: once the program is installed, build-time\ndependencies can be safely garbage-collected.\nIt also matters when a substitute is available, in which case only the <em>inputs<\/em>\nand <em>propagated inputs<\/em> will be fetched: the <em>native inputs<\/em> are not required to\ninstall a package from a substitute.<\/p><h3>Outputs<\/h3><p>Just like how a package can have multiple inputs, it can also produce multiple\noutputs.<\/p><p>Each output corresponds to a separate directory in the store.<\/p><p>The user can choose which output to install; this is useful to save space or\nto avoid polluting the user profile with unwanted executables or libraries.<\/p><p>Output separation is optional.  When the <code>outputs<\/code> field is left out, the\ndefault and only output (the complete package) is referred to as <code>\"out\"<\/code>.<\/p><p>Typical separate output names include <code>debug<\/code> and <code>doc<\/code>.<\/p><p>It's advised to separate outputs only when you've shown it's worth it: if the\noutput size is significant (compare with <code>guix size<\/code>) or in case the package is\nmodular.<\/p><h3>Build system arguments<\/h3><p>The <code>arguments<\/code> is a keyword-value list used to configure the build process.<\/p><p>The simplest argument <code>#:tests?<\/code> can be used to disable the test suite when\nbuilding the package.  This is mostly useful when the package does not feature\nany test suite.  It's strongly recommended to keep the test suite on if there is\none.<\/p><p>Another  common argument is <code>:make-flags<\/code>, which specifies a list of flags to\nappend when running make, as you would from the command line.  For instance, the\nfollowing flags<\/p><pre>#:make-flags (list (string-append \"prefix=\" (assoc-ref %outputs \"out\"))\n                   \"CC=gcc\")<\/pre><p>translate into<\/p><pre class=\"language-sh\">$ make CC=gcc prefix=\/gnu\/store\/...-&lt;out&gt;<\/pre><p>This sets the C compiler to <code>gcc<\/code> and the <code>prefix<\/code> variable (the installation\ndirectory in Make parlance) to <code>(assoc-ref %outputs \"out\")<\/code>, which is a build-stage\nglobal variable pointing to the destination directory in the store (something like\n<code>\/gnu\/store\/...-my-libgit2-20180408<\/code>).<\/p><p>Similarly, it's possible to set the \"configure\" flags.<\/p><pre>#:configure-flags '(\"-DUSE_SHA1DC=ON\")<\/pre><p>The <code>%build-inputs<\/code> variable is also generated in scope.  It's an association\ntable that maps the input names to their store directories.<\/p><p>The <code>phases<\/code> keyword lists the sequential steps of the build system.  Typically\nphases include <code>unpack<\/code>, <code>configure<\/code>, <code>build<\/code>, <code>install<\/code> and <code>check<\/code>.  To know\nmore about those phases, you need to work out the appropriate build system\ndefinition in <code>$GUIX_CHECKOUT\/guix\/build\/gnu-build-system.scm<\/code>:<\/p><pre>(define %standard-phases\n  ;; Standard build phases, as a list of symbol\/procedure pairs.\n  (let-syntax ((phases (syntax-rules ()\n                         ((_ p ...) `((p . ,p) ...)))))\n    (phases set-SOURCE-DATE-EPOCH set-paths install-locale unpack\n            bootstrap\n            patch-usr-bin-file\n            patch-source-shebangs configure patch-generated-file-shebangs\n            build check install\n            patch-shebangs strip\n            validate-runpath\n            validate-documentation-location\n            delete-info-dir-file\n            patch-dot-desktop-files\n            install-license-files\n            reset-gzip-timestamps\n            compress-documentation)))<\/pre><p>Or from the REPL:<\/p><pre>&gt; (add-to-load-path \"\/path\/to\/guix\/checkout\")\n&gt; ,module (guix build gnu-build-system)\n&gt; (map first %standard-phases)\n(set-SOURCE-DATE-EPOCH set-paths install-locale unpack bootstrap\npatch-usr-bin-file patch-source-shebangs configure\npatch-generated-file-shebangs build check install patch-shebangs strip\nvalidate-runpath validate-documentation-location delete-info-dir-file\npatch-dot-desktop-files install-license-files reset-gzip-timestamps\ncompress-documentation)<\/pre><p>If you want to know more about what happens during those phases, consult the\nassociated procedures.<\/p><p>For instance, as of this writing the definition of <code>unpack<\/code> for the GNU build\nsystem is<\/p><pre>(define* (unpack #:key source #:allow-other-keys)\n  \"Unpack SOURCE in the working directory, and change directory within the\nsource.  When SOURCE is a directory, copy it in a sub-directory of the current\nworking directory.\"\n  (if (file-is-directory? source)\n      (begin\n        (mkdir \"source\")\n        (chdir \"source\")\n        ;; Preserve timestamps (set to the Epoch) on the copied tree so that\n        ;; things work deterministically.\n        (copy-recursively source \".\"\n                          #:keep-mtime? #t))\n      (begin\n        (if (string-suffix? \".zip\" source)\n            (invoke \"unzip\" source)\n            (invoke \"tar\" \"xvf\" source))\n        (chdir (first-subdirectory \".\"))))\n  #t)<\/pre><p>Note the <code>chdir<\/code> call: it changes the working directory to where the source was\nunpacked.\nThus every phase following the <code>unpack<\/code> will use the source as a working\ndirectory, which is why we can directly work on the source files.\nThat is to say, unless a later phase changes the working directory to something\nelse.<\/p><p>We modify the list of <code>%standard-phases<\/code> of the build system with the\n<code>modify-phases<\/code> macro as per the list of specified modifications, which may have\nthe following forms:<\/p><ul><li><code>(add-before PHASE NEW-PHASE PROCEDURE)<\/code>: Run <code>PROCEDURE<\/code> named <code>NEW-PHASE<\/code> before <code>PHASE<\/code>.<\/li><li><code>(add-after PHASE NEW-PHASE PROCEDURE)<\/code>: Same, but afterwards.<\/li><li><code>(replace PHASE PROCEDURE)<\/code>.<\/li><li><code>(delete PHASE)<\/code>.<\/li><\/ul><p>The <code>PROCEDURE<\/code> supports the keyword arguments <code>inputs<\/code> and <code>outputs<\/code>.  Each\ninput (whether <em>native<\/em>, <em>propagated<\/em> or not) and output directory is referenced\nby their name in those variables.  Thus <code>(assoc-ref outputs \"out\")<\/code> is the store\ndirectory of the main output of the package.  A phase procedure may look like\nthis:<\/p><pre>(lambda* (#:key inputs outputs #:allow-other-keys)\n  (let (((bash-directory (assoc-ref inputs \"bash\"))\n         (output-directory (assoc-ref outputs \"out\"))\n         (doc-directory (assoc-ref outputs \"doc\"))\n  ; ...\n  #t)<\/pre><p>The procedure must return <code>#t<\/code> on success.  It's brittle to rely on the return\nvalue of the last expression used to tweak the phase because there is no\nguarantee it would be a <code>#t<\/code>.  Hence the trailing <code>#t<\/code> to ensure the right value\nis returned on success.<\/p><h3>Code staging<\/h3><p>The astute reader may have noticed the quasi-quote and comma syntax in the\nargument field.  Indeed, the build code in the package declaration should not be\nevaluated on the client side, but only when passed to the Guix daemon.  This\nmechanism of passing code around two running processes is called <a href=\"https:\/\/arxiv.org\/abs\/1709.00833\">code staging<\/a>.<\/p><h3>\"Utils\" functions<\/h3><p>When customizing <code>phases<\/code>, we often need to write code that mimics the\nequivalent system invocations (<code>make<\/code>, <code>mkdir<\/code>, <code>cp<\/code>, etc.) commonly used during\nregular \"Unix-style\" installations.<\/p><p>Some like <code>chmod<\/code> are native to Guile.  See the <a href=\"https:\/\/www.gnu.org\/software\/guile\/manual\/guile.html\">Guile reference manual<\/a> for a\ncomplete list.<\/p><p>Guix provides additional helper functions which prove especially handy in the\ncontext of package management.<\/p><p>Some of those functions can be found in\n<code>$GUIX_CHECKOUT\/guix\/guix\/build\/utils.scm<\/code>.  Most of them mirror the behaviour\nof the traditional Unix system commands:<\/p><ul><li><strong>which:<\/strong> Like the <code>which<\/code> system command.<\/li><li><strong>find-files:<\/strong> Akin to the <code>find<\/code> system command.<\/li><li><strong>mkdir-p:<\/strong> Like <code>mkdir -p<\/code>, which creates all parents as needed.<\/li><li><strong>install-file:<\/strong> Similar to <code>install<\/code> when installing a file to a (possibly\nnon-existing) directory.  Guile has <code>copy-file<\/code> which works\nlike <code>cp<\/code>.<\/li><li><strong>copy-recursively:<\/strong> Like <code>cp -r<\/code>.<\/li><li><strong>delete-file-recursively:<\/strong> Like <code>rm -rf<\/code>.<\/li><li><strong>invoke:<\/strong> Run an executable.  This should be used instead of <code>system*<\/code>.<\/li><li><strong>with-directory-excursion:<\/strong> Run the body in a different working directory,\nthen restore the previous working directory.<\/li><li><strong>substitute*:<\/strong> A \"sed-like\" function.<\/li><\/ul><h3>Module prefix<\/h3><p>The license in our last example needs a prefix: this is because of how the\n<code>license<\/code> module was imported in the package, as <code>#:use-module ((guix licenses) #:prefix license:)<\/code>.  The <a href=\"https:\/\/www.gnu.org\/software\/guile\/manual\/en\/html_node\/Using-Guile-Modules.html\">Guile module import mechanism<\/a> gives the user full\ncontrol over namespacing: this is needed to avoid clashes between, say, the\n<code>zlib<\/code> variable from <code>licenses.scm<\/code> (a <em>license<\/em> value) and the <code>zlib<\/code> variable\nfrom <code>compression.scm<\/code> (a <em>package<\/em> value).<\/p><p>What we've seen so far covers the majority of packages using a build system\nother than the <code>trivial-build-system<\/code>.  The latter does not automate anything\nand leaves you to build everything manually.  This can be more demanding and we\nwon't cover it here for now, but thankfully it is rarely necessary to fall back\non this system.<\/p><p>For the other build systems, such as ASDF, Emacs, Perl, Ruby and many more, the\nprocess is very similar to the GNU build system except for a few specialized\narguments.<\/p><p>Learn more about build systems in<\/p><p>We can't repeat it enough: having a full-fledged programming language at hand\nempowers us in ways that reach far beyond traditional package management.<\/p><p>Let's illustrate this with some awesome features of Guix!<\/p><h3>Recursive importers<\/h3><p>You might find some build systems good enough that there is little to do at all\nto write a package, to the point that it becomes repetitive and tedious after a\nwhile.  A <em>raison d'\u00eatre<\/em> of computers is to replace human beings at those\nboring tasks.  So let's tell Guix to do this for us and create the package\ndefinition of an R package from CRAN (the output is trimmed for conciseness):<\/p><pre class=\"language-sh\">$ guix import cran --recursive walrus\n(define-public r-mc2d\n    ; ...\n    (license gpl2+)))\n(define-public r-jmvcore\n    ; ...\n    (license gpl2+)))\n(define-public r-wrs2\n    ; ...\n    (license gpl3)))\n(define-public r-walrus\n  (package\n    (name \"r-walrus\")\n    (version \"1.0.3\")\n    (source\n      (origin\n        (method url-fetch)\n        (uri (cran-uri \"walrus\" version))\n        (sha256\n          (base32\n            \"1nk2glcvy4hyksl5ipq2mz8jy4fss90hx6cq98m3w96kzjni6jjj\"))))\n    (build-system r-build-system)\n    (propagated-inputs\n      `((\"r-ggplot2\" ,r-ggplot2)\n        (\"r-jmvcore\" ,r-jmvcore)\n        (\"r-r6\" ,r-r6)\n        (\"r-wrs2\" ,r-wrs2)))\n    (home-page \"https:\/\/github.com\/jamovi\/walrus\")\n    (synopsis \"Robust Statistical Methods\")\n    (description\n      \"This package provides a toolbox of common robust statistical tests, including robust descriptives, robust t-tests, and robust ANOVA.  It is also available as a module for 'jamovi' (see &lt;https:\/\/www.jamovi.org&gt; for more information).  Walrus is based on the WRS2 package by Patrick Mair, which is in turn based on the scripts and work of Rand Wilcox.  These analyses are described in depth in the book 'Introduction to Robust Estimation &amp; Hypothesis Testing'.\")\n    (license gpl3)))<\/pre><p>The recursive importer won't import packages for which Guix already has package\ndefinitions, except for the very first.<\/p><p>Not all applications can be packaged this way, only those relying on a select\nnumber of supported systems.  Read about the full list of importers in the <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-import.html\">guix\nimport section<\/a> of the manual.<\/p><h3>Automatic update<\/h3><p>Guix can be smart enough to check for updates on systems it knows.  It can\nreport outdated package definitions with<\/p><pre class=\"language-sh\">$ guix refresh hello<\/pre><p>In most cases, updating a package to a newer version requires little more than\nchanging the version number and the checksum.  Guix can do that automatically as\nwell:<\/p><pre class=\"language-sh\">$ guix refresh hello --update<\/pre><h3>Inheritance<\/h3><p>If you've started browsing the existing package definitions, you might have\nnoticed that a significant number of them have a <code>inherit<\/code> field:<\/p><pre>(define-public adwaita-icon-theme\n  (package (inherit gnome-icon-theme)\n    (name \"adwaita-icon-theme\")\n    (version \"3.26.1\")\n    (source (origin\n              (method url-fetch)\n              (uri (string-append \"mirror:\/\/gnome\/sources\/\" name \"\/\"\n                                  (version-major+minor version) \"\/\"\n                                  name \"-\" version \".tar.xz\"))\n              (sha256\n               (base32\n                \"17fpahgh5dyckgz7rwqvzgnhx53cx9kr2xw0szprc6bnqy977fi8\"))))\n    (native-inputs\n     `((\"gtk-encode-symbolic-svg\" ,gtk+ \"bin\")))))<\/pre><p>All unspecified fields are inherited from the parent package.  This is very\nconvenient to create alternative packages, for instance with different source,\nversion or compilation options.<\/p><p>Sadly, some applications can be tough to package.  Sometimes they need a patch to\nwork with the non-standard filesystem hierarchy enforced by the store.\nSometimes the tests won't run properly.  (They can be skipped but this is not\nrecommended.)  Other times the resulting package won't be reproducible.<\/p><p>Should you be stuck, unable to figure out how to fix any sort of packaging\nissue, don't hesitate to ask the community for help.<\/p><p>See the <a href=\"https:\/\/www.gnu.org\/software\/guix\/contact\/\">Guix homepage<\/a> for information on the mailing lists, IRC, etc.<\/p><p>This tutorial was a showcase of the sophisticated package management that Guix\nboasts.  At this point we have mostly restricted this introduction to the\n<code>gnu-build-system<\/code> which is a core abstraction layer on which more advanced\nabstractions are based.<\/p><p>Where do we go from here?  Next we ought to dissect the innards of the build\nsystem by removing all abstractions, using the <code>trivial-build-system<\/code>: this\nshould give us a thorough understanding of the process before investigating some\nmore advanced packaging techniques and edge cases.<\/p><p>Other features worth exploring are the interactive editing and debugging\ncapabilities of Guix provided by the Guile REPL.<\/p><p>Those fancy features are completely optional and can wait; now is a good time\nto take a well-deserved break.  With what we've introduced here you should be\nwell armed to package lots of programs.  You can get started right away and\nhopefully we will see your contributions soon!<\/p><p><a href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/4.0\/\"><img src=\"https:\/\/i.creativecommons.org\/l\/by-sa\/4.0\/88x31.png\" alt=\"CC-BY-SA\" title=\"CC-BY-SA&#10;4.0\" \/><\/a><\/p><p>This work is licensed under a Creative Commons Attribution-ShareAlike 4.0\nInternational License.<\/p><h4>About GNU Guix<\/h4><p><a href=\"https:\/\/www.gnu.org\/software\/guix\">GNU\u00a0Guix<\/a> is a transactional package\nmanager for the GNU system.  The Guix System Distribution or GuixSD is\nan advanced distribution of the GNU system that relies on GNU Guix and\n<a href=\"https:\/\/www.gnu.org\/distros\/free-system-distribution-guidelines.html\">respects the user's\nfreedom<\/a>.<\/p><p>In addition to standard package management features, Guix supports\ntransactional upgrades and roll-backs, unprivileged package management,\nper-user profiles, and garbage collection.  Guix uses low-level\nmechanisms from the Nix package manager, except that packages are\ndefined as native <a href=\"https:\/\/www.gnu.org\/software\/guile\">Guile<\/a> modules,\nusing extensions to the <a href=\"http:\/\/schemers.org\">Scheme<\/a> language.  GuixSD\noffers a declarative approach to operating system configuration\nmanagement, and is highly customizable and hackable.<\/p><p>GuixSD can be used on an i686, x86_64, ARMv7, and AArch64 machines.  It\nis also possible to use Guix on top of an already installed GNU\/Linux\nsystem, including on mips64el and aarch64.<\/p>","created_at":"2019-05-03T14:36:59+0000","updated_at":"2019-05-03T14:36:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":29,"domain_name":"www.gnu.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":932,"title":"x11 - What desktop environment does startx run, and how can I change it? - Unix & Linux Stack Exchange","url":"https:\/\/unix.stackexchange.com\/questions\/243195\/what-desktop-environment-does-startx-run-and-how-can-i-change-it\/243210#243210","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>From many docs, I read that <code>startx<\/code> is starting LXDE in Raspbian OS. I am a little bit confused.<\/p>\n<p>Will always <code>startx<\/code> run LXDE GUI?<\/p>\n<p>Also I have seen example with using <code>startlxde<\/code> command. How is that command different and why <code>startx<\/code> and <code>startlxde<\/code> are running the same GUI(LXDE)? Or maybe it runs it because it is the default GUI?<\/p>\n<p>How can I choose default GUI if I have multiple ones?<\/p>\n<p>Could you please explain more details around the GUI in Linux systems?<\/p><\/div><h2 data-answercount=\"2\">&#13;\n                                2 Answers&#13;\n                            2&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><code>startx<\/code> runs <code>xinit<\/code> which starts an X server and a client session. The client session is <code>~\/.xinitrc<\/code> if present, and otherwise <code>\/etc\/X11\/xinit\/xinitrc<\/code> (the location may vary between distributions). What this script does varies between distributions. On Debian (including derivatives such as Raspbian), <code>\/etc\/X11\/xinit\/xinitrc<\/code> runs <code>\/etc\/X11\/Xsession<\/code> which in turn runs scripts in <code>\/etc\/X11\/Xsession.d<\/code>. The Debian scripts look for a user session in other files (<code>~\/.xsession<\/code>, <code>~\/.xsessionrc<\/code>, <code>~\/.Xsession<\/code>) and, if no user setting is applicable, runs <code>x-session-manager<\/code> (falling back to <code>x-window-manager<\/code> if no [session manager] is installed, falling back to <code>x-terminal-emulator<\/code> in the unlikely case that no window manager is installed).<\/p>\n<p>If you want control over what gets executed, you can create one of the user files, either <code>~\/.xsession<\/code> or <code>~\/.xinitrc<\/code>. The file <code>~\/.xsession<\/code> is also used if you log in on a <a href=\"https:\/\/en.wikipedia.org\/wiki\/X_display_manager_(program_type)\">display manager<\/a> (i.e. if you type your password in a GUI window). The file <code>~\/.xinitrc<\/code> is specific to <code>xinit<\/code> and <code>startx<\/code>. Using <code>~\/.xsession<\/code> goes through <code>\/etc\/X11\/Xsession<\/code> so it sets up things like input methods, resources, password agents, etc. If you use <code>.xinitrc<\/code>, you'll have to do all of these manually. Once again, I'm describing Debian here, other Unix variants might set things up differently. The use of <code>~\/.xinitrc<\/code> to specify what gets executed when you run <code>startx<\/code> or <code>xinit<\/code> is universal.<\/p>\n<p>Whether you use <code>~\/.xinitrc<\/code> or <code>~\/.xsession<\/code>, this file (usually a shell script, but it doesn't have to be if you really want to use something else) must prepare whatever needs to be prepared (e.g. keyboard settings, resources, applets that aren't started by the window manager, etc.), and then at the end run the program that manages the session. When the script ends, the session terminates. Typically, you would use <code>exec<\/code> at the end of the script, to replace the script by the session manager or window manager.<\/p>\n<p>Your system presumably has <code>\/usr\/bin\/startlxde<\/code> as the system-wide default session manager. On Debian and derivatives, you can check the available session managers with<\/p>\n<pre>update-alternatives --list x-session-manager\n<\/pre>\n<p>or get a more verbose description indicating which one is current with<\/p>\n<pre>update-alternatives --display x-session-manager\n<\/pre>\n<p>If LXDE wasn't the system-wide default and you wanted to make it the default for your account, you could use the following <code>~\/.xsession<\/code> file:<\/p>\n<pre>#!\/bin\/sh\nexec startlxde\n<\/pre>\n<p>On some Unix variants, that would only run for graphical logins, not for <code>startx<\/code>, so you'd also need to create an identical <code>~\/.xinitrc<\/code>. (Or not identical: in <code>~\/.xsession<\/code>, you might want to do other things, because that's the first file that's executed in a graphical session; for example you might put <code>. ~\/.profile<\/code> near the top, to set some environment variables.)<\/p>\n<p>If you want to try out other environments as a one-off, you can specify a different program to run on the command line of <code>startx<\/code> itself. The <code>startx<\/code> program has a quirk: you need to use the full path to the program.<\/p>\n<pre>startx \/usr\/bin\/startkde\n<\/pre>\n<p>The <code>startx<\/code> command also lets you specify arguments to pass to the server. For example, if you want to run multiple GUI sessions at the same time, you can pass a different display number each time. Pass server arguments after <code>--<\/code> on the command line of <code>startx<\/code>.<\/p>\n<pre>startx \/usr\/bin\/startkde -- :1\n<\/pre><\/div>","created_at":"2019-05-03T14:33:11+0000","updated_at":"2019-05-03T14:33:11+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"unix.stackexchange.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/unix\/img\/apple-touch-icon@2.png?v=32fb07f7ce26","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":931,"title":"Top (GNU Guix Reference Manual)","url":"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/index.html","content":"<p>This document describes GNU Guix version 1.0.0, a functional\npackage management tool written for the GNU system.\n<\/p><div>\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Introduction.html#Introduction\" accesskey=\"1\">Introduction<\/a>:What is Guix about?\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Installation.html#Installation\" accesskey=\"2\">Installation<\/a>:Installing Guix.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/System-Installation.html#System-Installation\" accesskey=\"3\">System Installation<\/a>:Installing the whole operating system.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Package-Management.html#Package-Management\" accesskey=\"4\">Package Management<\/a>:Package installation, upgrade, etc.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Development.html#Development\" accesskey=\"5\">Development<\/a>:Guix-aided software development.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Programming-Interface.html#Programming-Interface\" accesskey=\"6\">Programming Interface<\/a>:Using Guix in Scheme.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Utilities.html#Utilities\" accesskey=\"7\">Utilities<\/a>:Package management commands.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/System-Configuration.html#System-Configuration\" accesskey=\"8\">System Configuration<\/a>:Configuring the operating system.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Documentation.html#Documentation\" accesskey=\"9\">Documentation<\/a>:Browsing software user manuals.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Installing-Debugging-Files.html#Installing-Debugging-Files\">Installing Debugging Files<\/a>:Feeding the debugger.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Security-Updates.html#Security-Updates\">Security Updates<\/a>:Deploying security fixes quickly.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Bootstrapping.html#Bootstrapping\">Bootstrapping<\/a>:GNU\/Linux built from scratch.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Porting.html#Porting\">Porting<\/a>:Targeting another platform or kernel.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Contributing.html#Contributing\">Contributing<\/a>:Your help needed!\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Acknowledgments.html#Acknowledgments\">Acknowledgments<\/a>:Thanks!\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/GNU-Free-Documentation-License.html#GNU-Free-Documentation-License\">GNU Free Documentation License<\/a>:The license of this manual.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Concept-Index.html#Concept-Index\">Concept Index<\/a>:Concepts.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Programming-Index.html#Programming-Index\">Programming Index<\/a>:Data types, functions, and variables.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Managing-Software-the-Guix-Way.html#Managing-Software-the-Guix-Way\">Managing Software the Guix Way<\/a>:What\u2019s special.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/GNU-Distribution.html#GNU-Distribution\">GNU Distribution<\/a>:The packages and tools.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Binary-Installation.html#Binary-Installation\">Binary Installation<\/a>:Getting Guix running in no time!\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Requirements.html#Requirements\">Requirements<\/a>:Software needed to build and run Guix.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Running-the-Test-Suite.html#Running-the-Test-Suite\">Running the Test Suite<\/a>:Testing Guix.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Setting-Up-the-Daemon.html#Setting-Up-the-Daemon\">Setting Up the Daemon<\/a>:Preparing the build daemon\u2019s environment.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix_002ddaemon.html#Invoking-guix_002ddaemon\">Invoking guix-daemon<\/a>:Running the build daemon.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Application-Setup.html#Application-Setup\">Application Setup<\/a>:Application-specific setup.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Build-Environment-Setup.html#Build-Environment-Setup\">Build Environment Setup<\/a>:Preparing the isolated build environment.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Daemon-Offload-Setup.html#Daemon-Offload-Setup\">Daemon Offload Setup<\/a>:Offloading builds to remote machines.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/SELinux-Support.html#SELinux-Support\">SELinux Support<\/a>:Using an SELinux policy for the daemon.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Limitations.html#Limitations\">Limitations<\/a>:What you can expect.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Hardware-Considerations.html#Hardware-Considerations\">Hardware Considerations<\/a>:Supported hardware.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/USB-Stick-and-DVD-Installation.html#USB-Stick-and-DVD-Installation\">USB Stick and DVD Installation<\/a>:Preparing the installation medium.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Preparing-for-Installation.html#Preparing-for-Installation\">Preparing for Installation<\/a>:Networking, partitioning, etc.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Guided-Graphical-Installation.html#Guided-Graphical-Installation\">Guided Graphical Installation<\/a>:Easy graphical installation.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Manual-Installation.html#Manual-Installation\">Manual Installation<\/a>:Manual installation for wizards.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/After-System-Installation.html#After-System-Installation\">After System Installation<\/a>:When installation succeeded.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Installing-Guix-in-a-VM.html#Installing-Guix-in-a-VM\">Installing Guix in a VM<\/a>:Guix System playground.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Building-the-Installation-Image.html#Building-the-Installation-Image\">Building the Installation Image<\/a>:How this comes to be.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Keyboard-Layout-and-Networking-and-Partitioning.html#Keyboard-Layout-and-Networking-and-Partitioning\">Keyboard Layout and Networking and Partitioning<\/a>:Initial setup.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Proceeding-with-the-Installation.html#Proceeding-with-the-Installation\">Proceeding with the Installation<\/a>:Installing.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Features.html#Features\">Features<\/a>:How Guix will make your life brighter.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-package.html#Invoking-guix-package\">Invoking guix package<\/a>:Package installation, removal, etc.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Substitutes.html#Substitutes\">Substitutes<\/a>:Downloading pre-built binaries.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Packages-with-Multiple-Outputs.html#Packages-with-Multiple-Outputs\">Packages with Multiple Outputs<\/a>:Single source package, multiple outputs.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-gc.html#Invoking-guix-gc\">Invoking guix gc<\/a>:Running the garbage collector.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-pull.html#Invoking-guix-pull\">Invoking guix pull<\/a>:Fetching the latest Guix and distribution.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Channels.html#Channels\">Channels<\/a>:Customizing the package collection.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Inferiors.html#Inferiors\">Inferiors<\/a>:Interacting with another revision of Guix.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-describe.html#Invoking-guix-describe\">Invoking guix describe<\/a>:Display information about your Guix revision.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-archive.html#Invoking-guix-archive\">Invoking guix archive<\/a>:Exporting and importing store files.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Official-Substitute-Server.html#Official-Substitute-Server\">Official Substitute Server<\/a>:One particular source of substitutes.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Substitute-Server-Authorization.html#Substitute-Server-Authorization\">Substitute Server Authorization<\/a>:How to enable or disable substitutes.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Substitute-Authentication.html#Substitute-Authentication\">Substitute Authentication<\/a>:How Guix verifies substitutes.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Proxy-Settings.html#Proxy-Settings\">Proxy Settings<\/a>:How to get substitutes via proxy.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Substitution-Failure.html#Substitution-Failure\">Substitution Failure<\/a>:What happens when substitution fails.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/On-Trusting-Binaries.html#On-Trusting-Binaries\">On Trusting Binaries<\/a>:How can you trust that binary blob?\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-environment.html#Invoking-guix-environment\">Invoking guix environment<\/a>:Setting up development environments.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-pack.html#Invoking-guix-pack\">Invoking guix pack<\/a>:Creating software bundles.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Package-Modules.html#Package-Modules\">Package Modules<\/a>:Packages from the programmer\u2019s viewpoint.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Defining-Packages.html#Defining-Packages\">Defining Packages<\/a>:Defining new packages.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Build-Systems.html#Build-Systems\">Build Systems<\/a>:Specifying how packages are built.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/The-Store.html#The-Store\">The Store<\/a>:Manipulating the package store.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Derivations.html#Derivations\">Derivations<\/a>:Low-level interface to package derivations.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/The-Store-Monad.html#The-Store-Monad\">The Store Monad<\/a>:Purely functional interface to the store.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/G_002dExpressions.html#G_002dExpressions\">G-Expressions<\/a>:Manipulating build expressions.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-repl.html#Invoking-guix-repl\">Invoking guix repl<\/a>:Fiddling with Guix interactively.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/package-Reference.html#package-Reference\">package Reference<\/a>:The package data type.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/origin-Reference.html#origin-Reference\">origin Reference<\/a>:The origin data type.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-build.html#Invoking-guix-build\">Invoking guix build<\/a>:Building packages from the command line.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-edit.html#Invoking-guix-edit\">Invoking guix edit<\/a>:Editing package definitions.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-download.html#Invoking-guix-download\">Invoking guix download<\/a>:Downloading a file and printing its hash.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-hash.html#Invoking-guix-hash\">Invoking guix hash<\/a>:Computing the cryptographic hash of a file.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-import.html#Invoking-guix-import\">Invoking guix import<\/a>:Importing package definitions.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-refresh.html#Invoking-guix-refresh\">Invoking guix refresh<\/a>:Updating package definitions.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-lint.html#Invoking-guix-lint\">Invoking guix lint<\/a>:Finding errors in package definitions.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-size.html#Invoking-guix-size\">Invoking guix size<\/a>:Profiling disk usage.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-graph.html#Invoking-guix-graph\">Invoking guix graph<\/a>:Visualizing the graph of packages.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-publish.html#Invoking-guix-publish\">Invoking guix publish<\/a>:Sharing substitutes.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-challenge.html#Invoking-guix-challenge\">Invoking guix challenge<\/a>:Challenging substitute servers.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-copy.html#Invoking-guix-copy\">Invoking guix copy<\/a>:Copying to and from a remote store.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-container.html#Invoking-guix-container\">Invoking guix container<\/a>:Process isolation.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-weather.html#Invoking-guix-weather\">Invoking guix weather<\/a>:Assessing substitute availability.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-processes.html#Invoking-guix-processes\">Invoking guix processes<\/a>:Listing client processes.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Common-Build-Options.html#Common-Build-Options\">Common Build Options<\/a>:Build options for most commands.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Package-Transformation-Options.html#Package-Transformation-Options\">Package Transformation Options<\/a>:Creating variants of packages.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Additional-Build-Options.html#Additional-Build-Options\">Additional Build Options<\/a>:Options specific to \u2019guix build\u2019.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Debugging-Build-Failures.html#Debugging-Build-Failures\">Debugging Build Failures<\/a>:Real life packaging experience.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Using-the-Configuration-System.html#Using-the-Configuration-System\">Using the Configuration System<\/a>:Customizing your GNU system.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/operating_002dsystem-Reference.html#operating_002dsystem-Reference\">operating-system Reference<\/a>:Detail of operating-system declarations.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/File-Systems.html#File-Systems\">File Systems<\/a>:Configuring file system mounts.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Mapped-Devices.html#Mapped-Devices\">Mapped Devices<\/a>:Block device extra processing.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/User-Accounts.html#User-Accounts\">User Accounts<\/a>:Specifying user accounts.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Keyboard-Layout.html#Keyboard-Layout\">Keyboard Layout<\/a>:How the system interprets key strokes.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Locales.html#Locales\">Locales<\/a>:Language and cultural convention settings.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Services.html#Services\">Services<\/a>:Specifying system services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Setuid-Programs.html#Setuid-Programs\">Setuid Programs<\/a>:Programs running with root privileges.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/X_002e509-Certificates.html#X_002e509-Certificates\">X.509 Certificates<\/a>:Authenticating HTTPS servers.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Name-Service-Switch.html#Name-Service-Switch\">Name Service Switch<\/a>:Configuring libc\u2019s name service switch.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Initial-RAM-Disk.html#Initial-RAM-Disk\">Initial RAM Disk<\/a>:Linux-Libre bootstrapping.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Bootloader-Configuration.html#Bootloader-Configuration\">Bootloader Configuration<\/a>:Configuring the boot loader.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Invoking-guix-system.html#Invoking-guix-system\">Invoking guix system<\/a>:Instantiating a system configuration.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Running-Guix-in-a-VM.html#Running-Guix-in-a-VM\">Running Guix in a VM<\/a>:How to run Guix System in a virtual machine.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Defining-Services.html#Defining-Services\">Defining Services<\/a>:Adding new service definitions.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Base-Services.html#Base-Services\">Base Services<\/a>:Essential system services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Scheduled-Job-Execution.html#Scheduled-Job-Execution\">Scheduled Job Execution<\/a>:The mcron service.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Log-Rotation.html#Log-Rotation\">Log Rotation<\/a>:The rottlog service.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Networking-Services.html#Networking-Services\">Networking Services<\/a>:Network setup, SSH daemon, etc.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/X-Window.html#X-Window\">X Window<\/a>:Graphical display.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Printing-Services.html#Printing-Services\">Printing Services<\/a>:Local and remote printer support.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Desktop-Services.html#Desktop-Services\">Desktop Services<\/a>:D-Bus and desktop services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Sound-Services.html#Sound-Services\">Sound Services<\/a>:ALSA and Pulseaudio services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Database-Services.html#Database-Services\">Database Services<\/a>:SQL databases, key-value stores, etc.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Mail-Services.html#Mail-Services\">Mail Services<\/a>:IMAP, POP3, SMTP, and all that.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Messaging-Services.html#Messaging-Services\">Messaging Services<\/a>:Messaging services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Telephony-Services.html#Telephony-Services\">Telephony Services<\/a>:Telephony services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Monitoring-Services.html#Monitoring-Services\">Monitoring Services<\/a>:Monitoring services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Kerberos-Services.html#Kerberos-Services\">Kerberos Services<\/a>:Kerberos services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Web-Services.html#Web-Services\">Web Services<\/a>:Web servers.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Certificate-Services.html#Certificate-Services\">Certificate Services<\/a>:TLS certificates via Let\u2019s Encrypt.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/DNS-Services.html#DNS-Services\">DNS Services<\/a>:DNS daemons.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/VPN-Services.html#VPN-Services\">VPN Services<\/a>:VPN daemons.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Network-File-System.html#Network-File-System\">Network File System<\/a>:NFS related services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Continuous-Integration.html#Continuous-Integration\">Continuous Integration<\/a>:The Cuirass service.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Power-Management-Services.html#Power-Management-Services\">Power Management Services<\/a>:Extending battery life.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Audio-Services.html#Audio-Services\">Audio Services<\/a>:The MPD.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Virtualization-Services.html#Virtualization-Services\">Virtualization Services<\/a>:Virtualization services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Version-Control-Services.html#Version-Control-Services\">Version Control Services<\/a>:Providing remote access to Git repositories.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Game-Services.html#Game-Services\">Game Services<\/a>:Game servers.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Miscellaneous-Services.html#Miscellaneous-Services\">Miscellaneous Services<\/a>:Other services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Service-Composition.html#Service-Composition\">Service Composition<\/a>:The model for composing services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Service-Types-and-Services.html#Service-Types-and-Services\">Service Types and Services<\/a>:Types and services.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Service-Reference.html#Service-Reference\">Service Reference<\/a>:API reference.\n\u2022 <a href=\"https:\/\/www.gnu.org\/software\/guix\/manual\/en\/html_node\/Shepherd-Services.html#Shepherd-Services\">Shepherd Services<\/a>:A particular type of service.\n<\/div>","created_at":"2019-05-02T23:30:32+0000","updated_at":"2019-05-02T23:30:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.gnu.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","wm"],"is_public":false,"id":930,"title":"Emacs is My New Window Manager","url":"http:\/\/www.howardism.org\/Technical\/Emacs\/new-window-manager.html","content":"<header>\n<\/header><p>\nMost companies that employ me, hand me a \u201cwork laptop\u201d as I enter the\nbuilding. Of course, I do not install personal software and keep a\nclear division between my \u201cwork like\u201d and my \u201creal life.\u201d\n<\/p><p>\nHowever, I also don\u2019t like to carry two computers just to jot down\npersonal notes. My remedy is to install a virtualization system and\ncreate a \u201cpersonal\u201d virtual machine. (Building cloud software as my\nday job means I usually have a few VMs running all the time.)\n<\/p><p>\nSince I want this VM to have minimal impact on my work, I base it on a\n\u201cServer\u201d version of Ubuntu. however, I like some graphical features,\nso my most minimal <i>after market<\/i> installation approach is:<sup><a id=\"fnr.1\" class=\"footref\" href=\"#fn.1\">1<\/a><\/sup><\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">sudo apt-get install -y xinit\n<\/pre><\/div><p>\nSince most of what I do is <code>org-mode<\/code> work, the next step is:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">sudo apt-get install -y emacs\n<\/pre><\/div><p>\nI have played with a lot of window managers, and while some claim to\nbe unobtrusive and minimal<sup><a id=\"fnr.2\" class=\"footref\" href=\"#fn.2\">2<\/a><\/sup>, I really just want Emacs in full-screen\nmode (utilizing all screen estate possible).<sup><a id=\"fnr.3\" class=\"footref\" href=\"#fn.3\">3<\/a><\/sup><\/p><p>\nTo accomplish this, I create an <code>.xinitrc<\/code> file that contains only:\n<\/p><pre class=\"example\">\nexec emacs\n<\/pre><p>\nThat\u2019s right, folks, <b>Emacs is my window manager<\/b>. I add the\nfollowing to my Emacs <code>init.el<\/code> script:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-elisp\">(set-frame-parameter nil 'fullscreen 'fullboth)\n<\/pre><\/div><p>\nAnd now I can split the screen into windows, launch programs\u2013 even\nedit files\u2013 all without fondling the mouse, and since Emacs is in its\ngraphical mode, I can use my favorite fonts, decorate the fringe, etc.\n<\/p><figure><img src=\"http:\/\/www.howardism.org\/Technical\/Emacs\/new-window-manager-1.png\" alt=\"new-window-manager-1.png\" \/><\/figure><p>\nIf I need a program that doesn\u2019t run well <i>within<\/i> Emacs, I can\ncall <code>xterm<\/code> (or any other application) with <code>M-! xterm<\/code>.\n<\/p><div id=\"footnotes\"><div id=\"text-footnotes\"><div class=\"footdef\"><sup><a id=\"fn.1\" class=\"footnum\" href=\"#fnr.1\">1<\/a><\/sup><div class=\"footpara\"><p class=\"footpara\">\nSince I use my virtual machine primarily to take notes,\nyou\u2019ll need to install <code>ispell<\/code> (or one of the variants):\n<\/p><pre class=\"example\">\napt-get install -y ispell\n<\/pre><p class=\"footpara\">\nIn order to install the VirtualBox extensions, do:\n<\/p><pre class=\"example\">\nsudo apt-get install -y build-essential\nsudo mount \/dev\/cdrom \/mnt\nsudo \/mnt\/VBoxLinuxAdditions.run\n<\/pre><\/div><\/div><div class=\"footdef\"><sup><a id=\"fn.2\" class=\"footnum\" href=\"#fnr.2\">2<\/a><\/sup><div class=\"footpara\"><p class=\"footpara\">\nI always grin when I read the change log in the release notes of\nany window manager project. It begins by complaining that all other\nwindow managers are bloated, and that this one will be small and\nefficient.\n<\/p><p class=\"footpara\">\nAs you follow the history of the project, the complaints that it needs\nto do launching menus, iconify windows with translucent images, docks\nand panels, and soon the project is just as big as the competition.\n<\/p><p class=\"footpara\">\nNot that there is anything wrong with that\u2026 Writing a window manager\nis a lot of fun.\n<\/p><\/div><\/div><div class=\"footdef\"><sup><a id=\"fn.3\" class=\"footnum\" href=\"#fnr.3\">3<\/a><\/sup><div class=\"footpara\"><p class=\"footpara\">\nWhile taking notes at a conference, a person sitting behind me\nasked me after the session what window manager I was using. I said\nthis was just a normal Mac, but running Emacs in a full-screen mode\nwith a couple <code>org-mode<\/code> files, an embedded <a href=\"http:\/\/w3m.sourceforge.net\/\">w3m<\/a> browser, and an\nEmacs-based <a href=\"http:\/\/emacswiki.org\/emacs\/TwitteringMode\">twitter client<\/a>.\n<\/p><\/div><\/div><div class=\"footdef\"><sup><a id=\"fn.4\" class=\"footnum\" href=\"#fnr.4\">4<\/a><\/sup><div class=\"footpara\"><p class=\"footpara\">\nBy the way, to have EWW use Chromium, add this to your init file:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-elisp\">(setq shr-external-browser \"chromium-browser\")\n<\/pre><\/div><\/div><\/div><div class=\"footdef\"><sup><a id=\"fn.5\" class=\"footnum\" href=\"#fnr.5\">5<\/a><\/sup><div class=\"footpara\"><p class=\"footpara\">\nAlong with having the best name, the <a href=\"http:\/\/www.nongnu.org\/ratpoison\/\">Ratpoison window manager<\/a>\nis minimal and mouse-less (and even has a nice Info page), but its\ndefault <i>escape<\/i> key, <code>Control-t<\/code> is the transpose key in Emacs. To\nget around this and call Emacs\u2019 transpose function, type <code>C-t t<\/code>. This\nis <i>okay<\/i> when swapping two letters, but now swapping lines is <code>C-x\nC-t t<\/code>. Acceptable if I\u2019m swapping only two lines, but I\u2019m often\nswapping multiple lines\u2013and now it is painful.\n<\/p><p class=\"footpara\">\nTo make Ratpoison completely palatable with Emacs, make it use a\n<i>Command<\/i> key sequence (like the useless <code>Command-t<\/code> or whatever\nthat Windows key is on other keyboard). To do this, simply add the\nfollowing to your <code>~\/.ratpoisonrc<\/code> file:\n<\/p><pre class=\"example\">\nescape s-t\n<\/pre><p class=\"footpara\">\n<b>Note:<\/b> I\u2019ve received many <i>alternative<\/i> solutions to the most minimal,\nmouse-less window managers, Josuah Demangeon suggested <a href=\"http:\/\/git.suckless.org\/2wm\">2wm<\/a>. While\nI still prefer <i>ratpoison<\/i>, his journey mirrored mine, so I thought I\nwould attach his email\u2026especially, since variety is the spice of it\nall, right?\n<\/p><blockquote>\n<p>\nTL;DR - 2wm is the most smallest tiling window manager\n<\/p>\n<p>\nYour post inspired me, and I decided to look for the nimblest window manager that could make Emacs rocks on X11.\n<\/p>\n<p>\nI had a look at <a href=\"https:\/\/github.com\/mackstann\/tinywm\">tinywm<\/a>, but it is entirely mouse-driven, but hacking it would be like write a window manager from scratch.\n<\/p>\n<p>\nI had a look at <a href=\"http:\/\/www.6809.org.uk\/evilwm\/\">evilwm<\/a>, which is great, but I got difficulties to make it work as I wanted to.\n<\/p>\n<p>\nI looked at <a href=\"http:\/\/dwm.suckless.org\/\">dwm<\/a>: a very good one by Suckless, with many features. Maybe a little too much, and it wouldn\u2019t be like a bare Emacs experience with this wm. But after browsing a bit <a href=\"http:\/\/suckless.org\/other_projects\">their website<\/a> I found 2wm\n<\/p>\n<p>\n2wm is a keyboard-controlled tiling window manager that attach and detach windows on a main or side frame. It also support floating.\n<\/p>\n<p>\nFor me, 2wm is the minimal gap between Emacs and X11, as well as the ultimate window manager!\n<\/p>\n<p>\nIf you want to give it a try have a look at my <a href=\"https:\/\/github.com\/biotician\/src\/tree\/master\/2wm\">repo<\/a>. I tweaked the key-bindings to fit Emacs.\n<\/p>\n<\/blockquote><\/div><\/div><\/div><\/div>","created_at":"2019-05-02T15:17:10+0000","updated_at":"2019-05-02T15:17:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.howardism.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["sql","postgresql","joins"],"is_public":false,"id":929,"title":"PostgreSQL: Documentation: 11: 7.2.\u00a0Table Expressions","url":"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html","content":"<div class=\"toc\"><dl class=\"toc\"><dt><a href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-FROM\">7.2.1. The <code class=\"literal\">FROM<\/code> Clause<\/a><\/dt>\n      <dt><a href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-WHERE\">7.2.2. The <code class=\"literal\">WHERE<\/code> Clause<\/a><\/dt>\n      <dt><a href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-GROUP\">7.2.3. The <code class=\"literal\">GROUP BY<\/code> and <code class=\"literal\">HAVING<\/code> Clauses<\/a><\/dt>\n      <dt><a href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-GROUPING-SETS\">7.2.4. <code class=\"literal\">GROUPING SETS<\/code>, <code class=\"literal\">CUBE<\/code>, and <code class=\"literal\">ROLLUP<\/code><\/a><\/dt>\n      <dt><a href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-WINDOW\">7.2.5. Window Function Processing<\/a><\/dt>\n    <\/dl><\/div><p>A <em class=\"firstterm\">table expression<\/em> computes a table. The table expression contains a <code class=\"literal\">FROM<\/code> clause that is optionally followed by <code class=\"literal\">WHERE<\/code>, <code class=\"literal\">GROUP BY<\/code>, and <code class=\"literal\">HAVING<\/code> clauses. Trivial table expressions simply refer to a table on disk, a so-called base table, but more complex expressions can be used to modify or combine base tables in various ways.<\/p><p>The optional <code class=\"literal\">WHERE<\/code>, <code class=\"literal\">GROUP BY<\/code>, and <code class=\"literal\">HAVING<\/code> clauses in the table expression specify a pipeline of successive transformations performed on the table derived in the <code class=\"literal\">FROM<\/code> clause. All these transformations produce a virtual table that provides the rows that are passed to the select list to compute the output rows of the query.<\/p><div class=\"sect2\" id=\"QUERIES-FROM\"><div class=\"titlepage\"><div><div><h3 class=\"title\">7.2.1.\u00a0The <code class=\"literal\">FROM<\/code> Clause<\/h3><\/div><\/div><\/div><p>The <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/sql-select.html#SQL-FROM\" title=\"FROM Clause\"><code class=\"literal\">FROM<\/code> Clause<\/a> derives a table from one or more other tables given in a comma-separated table reference list.<\/p><pre class=\"synopsis\">\nFROM <em class=\"replaceable\"><code>table_reference<\/code><\/em> [, <em class=\"replaceable\"><code>table_reference<\/code><\/em> [, ...]]\n<\/pre><p>A table reference can be a table name (possibly schema-qualified), or a derived table such as a subquery, a <code class=\"literal\">JOIN<\/code> construct, or complex combinations of these. If more than one table reference is listed in the <code class=\"literal\">FROM<\/code> clause, the tables are cross-joined (that is, the Cartesian product of their rows is formed; see below). The result of the <code class=\"literal\">FROM<\/code> list is an intermediate virtual table that can then be subject to transformations by the <code class=\"literal\">WHERE<\/code>, <code class=\"literal\">GROUP BY<\/code>, and <code class=\"literal\">HAVING<\/code> clauses and is finally the result of the overall table expression.<\/p><p>When a table reference names a table that is the parent of a table inheritance hierarchy, the table reference produces rows of not only that table but all of its descendant tables, unless the key word <code class=\"literal\">ONLY<\/code> precedes the table name. However, the reference produces only the columns that appear in the named table \u2014 any columns added in subtables are ignored.<\/p><p>Instead of writing <code class=\"literal\">ONLY<\/code> before the table name, you can write <code class=\"literal\">*<\/code> after the table name to explicitly specify that descendant tables are included. There is no real reason to use this syntax any more, because searching descendant tables is now always the default behavior. However, it is supported for compatibility with older releases.<\/p><div class=\"sect3\" id=\"QUERIES-JOIN\"><div class=\"titlepage\"><div><h4 class=\"title\">7.2.1.1.\u00a0Joined Tables<\/h4><\/div><\/div><p>A joined table is a table derived from two other (real or derived) tables according to the rules of the particular join type. Inner, outer, and cross-joins are available. The general syntax of a joined table is<\/p><pre class=\"synopsis\">\n<em class=\"replaceable\"><code>T1<\/code><\/em> <em class=\"replaceable\"><code>join_type<\/code><\/em> <em class=\"replaceable\"><code>T2<\/code><\/em> [ <em class=\"replaceable\"><code>join_condition<\/code><\/em> ]\n<\/pre><p>Joins of all types can be chained together, or nested: either or both <em class=\"replaceable\"><code>T1<\/code><\/em> and <em class=\"replaceable\"><code>T2<\/code><\/em> can be joined tables. Parentheses can be used around <code class=\"literal\">JOIN<\/code> clauses to control the join order. In the absence of parentheses, <code class=\"literal\">JOIN<\/code> clauses nest left-to-right.<\/p><div class=\"variablelist\"><p class=\"title\"><strong>Join Types<\/strong><\/p><dl class=\"variablelist\"><dt>Cross join  <\/dt>\n          <dd>\n            <pre class=\"synopsis\">\n<em class=\"replaceable\"><code>T1<\/code><\/em> CROSS JOIN <em class=\"replaceable\"><code>T2<\/code><\/em>\n<\/pre>\n            <p>For every possible combination of rows from <em class=\"replaceable\"><code>T1<\/code><\/em> and <em class=\"replaceable\"><code>T2<\/code><\/em> (i.e., a Cartesian product), the joined table will contain a row consisting of all columns in <em class=\"replaceable\"><code>T1<\/code><\/em> followed by all columns in <em class=\"replaceable\"><code>T2<\/code><\/em>. If the tables have N and M rows respectively, the joined table will have N * M rows.<\/p>\n            <p><code class=\"literal\">FROM <em class=\"replaceable\"><code>T1<\/code><\/em> CROSS JOIN <em class=\"replaceable\"><code>T2<\/code><\/em><\/code> is equivalent to <code class=\"literal\">FROM <em class=\"replaceable\"><code>T1<\/code><\/em> INNER JOIN <em class=\"replaceable\"><code>T2<\/code><\/em> ON TRUE<\/code> (see below). It is also equivalent to <code class=\"literal\">FROM <em class=\"replaceable\"><code>T1<\/code><\/em>, <em class=\"replaceable\"><code>T2<\/code><\/em><\/code>.<\/p>\n            <div class=\"note\"><h3 class=\"title\">Note<\/h3><p>This latter equivalence does not hold exactly when more than two tables appear, because <code class=\"literal\">JOIN<\/code> binds more tightly than comma. For example <code class=\"literal\">FROM <em class=\"replaceable\"><code>T1<\/code><\/em> CROSS JOIN <em class=\"replaceable\"><code>T2<\/code><\/em> INNER JOIN <em class=\"replaceable\"><code>T3<\/code><\/em> ON <em class=\"replaceable\"><code>condition<\/code><\/em><\/code> is not the same as <code class=\"literal\">FROM <em class=\"replaceable\"><code>T1<\/code><\/em>, <em class=\"replaceable\"><code>T2<\/code><\/em> INNER JOIN <em class=\"replaceable\"><code>T3<\/code><\/em> ON <em class=\"replaceable\"><code>condition<\/code><\/em><\/code> because the <em class=\"replaceable\"><code>condition<\/code><\/em> can reference <em class=\"replaceable\"><code>T1<\/code><\/em> in the first case but not the second.<\/p><\/div>\n          <\/dd>\n          <dt>Qualified joins  <\/dt>\n          <dd>\n            <pre class=\"synopsis\">\n<em class=\"replaceable\"><code>T1<\/code><\/em> { [INNER] | { LEFT | RIGHT | FULL } [OUTER] } JOIN <em class=\"replaceable\"><code>T2<\/code><\/em> ON <em class=\"replaceable\"><code>boolean_expression<\/code><\/em>\n<em class=\"replaceable\"><code>T1<\/code><\/em> { [INNER] | { LEFT | RIGHT | FULL } [OUTER] } JOIN <em class=\"replaceable\"><code>T2<\/code><\/em> USING ( <em class=\"replaceable\"><code>join column list<\/code><\/em> )\n<em class=\"replaceable\"><code>T1<\/code><\/em> NATURAL { [INNER] | { LEFT | RIGHT | FULL } [OUTER] } JOIN <em class=\"replaceable\"><code>T2<\/code><\/em>\n<\/pre>\n            <p>The words <code class=\"literal\">INNER<\/code> and <code class=\"literal\">OUTER<\/code> are optional in all forms. <code class=\"literal\">INNER<\/code> is the default; <code class=\"literal\">LEFT<\/code>, <code class=\"literal\">RIGHT<\/code>, and <code class=\"literal\">FULL<\/code> imply an outer join.<\/p>\n            <p>The <em class=\"firstterm\">join condition<\/em> is specified in the <code class=\"literal\">ON<\/code> or <code class=\"literal\">USING<\/code> clause, or implicitly by the word <code class=\"literal\">NATURAL<\/code>. The join condition determines which rows from the two source tables are considered to \u201cmatch\u201d, as explained in detail below.<\/p>\n            <p>The possible types of qualified join are:<\/p>\n            <div class=\"variablelist\"><dl class=\"variablelist\"><dt><code class=\"literal\">INNER JOIN<\/code><\/dt>\n                <dd>\n                  <p>For each row R1 of T1, the joined table has a row for each row in T2 that satisfies the join condition with R1.<\/p>\n                <\/dd>\n                <dt><code class=\"literal\">LEFT OUTER JOIN<\/code>  <\/dt>\n                <dd>\n                  <p>First, an inner join is performed. Then, for each row in T1 that does not satisfy the join condition with any row in T2, a joined row is added with null values in columns of T2. Thus, the joined table always has at least one row for each row in T1.<\/p>\n                <\/dd>\n                <dt><code class=\"literal\">RIGHT OUTER JOIN<\/code>  <\/dt>\n                <dd>\n                  <p>First, an inner join is performed. Then, for each row in T2 that does not satisfy the join condition with any row in T1, a joined row is added with null values in columns of T1. This is the converse of a left join: the result table will always have a row for each row in T2.<\/p>\n                <\/dd>\n                <dt><code class=\"literal\">FULL OUTER JOIN<\/code><\/dt>\n                <dd>\n                  <p>First, an inner join is performed. Then, for each row in T1 that does not satisfy the join condition with any row in T2, a joined row is added with null values in columns of T2. Also, for each row of T2 that does not satisfy the join condition with any row in T1, a joined row with null values in the columns of T1 is added.<\/p>\n                <\/dd>\n              <\/dl><\/div>\n            <p>The <code class=\"literal\">ON<\/code> clause is the most general kind of join condition: it takes a Boolean value expression of the same kind as is used in a <code class=\"literal\">WHERE<\/code> clause. A pair of rows from <em class=\"replaceable\"><code>T1<\/code><\/em> and <em class=\"replaceable\"><code>T2<\/code><\/em> match if the <code class=\"literal\">ON<\/code> expression evaluates to true.<\/p>\n            <p>The <code class=\"literal\">USING<\/code> clause is a shorthand that allows you to take advantage of the specific situation where both sides of the join use the same name for the joining column(s). It takes a comma-separated list of the shared column names and forms a join condition that includes an equality comparison for each one. For example, joining <em class=\"replaceable\"><code>T1<\/code><\/em> and <em class=\"replaceable\"><code>T2<\/code><\/em> with <code class=\"literal\">USING (a, b)<\/code> produces the join condition <code class=\"literal\">ON <em class=\"replaceable\"><code>T1<\/code><\/em>.a = <em class=\"replaceable\"><code>T2<\/code><\/em>.a AND <em class=\"replaceable\"><code>T1<\/code><\/em>.b = <em class=\"replaceable\"><code>T2<\/code><\/em>.b<\/code>.<\/p>\n            <p>Furthermore, the output of <code class=\"literal\">JOIN USING<\/code> suppresses redundant columns: there is no need to print both of the matched columns, since they must have equal values. While <code class=\"literal\">JOIN ON<\/code> produces all columns from <em class=\"replaceable\"><code>T1<\/code><\/em> followed by all columns from <em class=\"replaceable\"><code>T2<\/code><\/em>, <code class=\"literal\">JOIN USING<\/code> produces one output column for each of the listed column pairs (in the listed order), followed by any remaining columns from <em class=\"replaceable\"><code>T1<\/code><\/em>, followed by any remaining columns from <em class=\"replaceable\"><code>T2<\/code><\/em>.<\/p>\n            <p>  Finally, <code class=\"literal\">NATURAL<\/code> is a shorthand form of <code class=\"literal\">USING<\/code>: it forms a <code class=\"literal\">USING<\/code> list consisting of all column names that appear in both input tables. As with <code class=\"literal\">USING<\/code>, these columns appear only once in the output table. If there are no common column names, <code class=\"literal\">NATURAL JOIN<\/code> behaves like <code class=\"literal\">JOIN ... ON TRUE<\/code>, producing a cross-product join.<\/p>\n            <div class=\"note\"><h3 class=\"title\">Note<\/h3><p><code class=\"literal\">USING<\/code> is reasonably safe from column changes in the joined relations since only the listed columns are combined. <code class=\"literal\">NATURAL<\/code> is considerably more risky since any schema changes to either relation that cause a new matching column name to be present will cause the join to combine that new column as well.<\/p><\/div>\n          <\/dd>\n        <\/dl><\/div><p>To put this together, assume we have tables <code class=\"literal\">t1<\/code>:<\/p><pre class=\"programlisting\">\n num | name\n-----+------\n   1 | a\n   2 | b\n   3 | c\n<\/pre><p>and <code class=\"literal\">t2<\/code>:<\/p><pre class=\"programlisting\">\n num | value\n-----+-------\n   1 | xxx\n   3 | yyy\n   5 | zzz\n<\/pre><p>then we get the following results for the various joins:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT * FROM t1 CROSS JOIN t2;<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n   1 | a    |   3 | yyy\n   1 | a    |   5 | zzz\n   2 | b    |   1 | xxx\n   2 | b    |   3 | yyy\n   2 | b    |   5 | zzz\n   3 | c    |   1 | xxx\n   3 | c    |   3 | yyy\n   3 | c    |   5 | zzz\n(9 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 INNER JOIN t2 ON t1.num = t2.num;<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n   3 | c    |   3 | yyy\n(2 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 INNER JOIN t2 USING (num);<\/code><\/strong>\n num | name | value\n-----+------+-------\n   1 | a    | xxx\n   3 | c    | yyy\n(2 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 NATURAL INNER JOIN t2;<\/code><\/strong>\n num | name | value\n-----+------+-------\n   1 | a    | xxx\n   3 | c    | yyy\n(2 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 LEFT JOIN t2 ON t1.num = t2.num;<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n   2 | b    |     |\n   3 | c    |   3 | yyy\n(3 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 LEFT JOIN t2 USING (num);<\/code><\/strong>\n num | name | value\n-----+------+-------\n   1 | a    | xxx\n   2 | b    |\n   3 | c    | yyy\n(3 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 RIGHT JOIN t2 ON t1.num = t2.num;<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n   3 | c    |   3 | yyy\n     |      |   5 | zzz\n(3 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT * FROM t1 FULL JOIN t2 ON t1.num = t2.num;<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n   2 | b    |     |\n   3 | c    |   3 | yyy\n     |      |   5 | zzz\n(4 rows)\n<\/pre><p>The join condition specified with <code class=\"literal\">ON<\/code> can also contain conditions that do not relate directly to the join. This can prove useful for some queries but needs to be thought out carefully. For example:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT * FROM t1 LEFT JOIN t2 ON t1.num = t2.num AND t2.value = 'xxx';<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n   2 | b    |     |\n   3 | c    |     |\n(3 rows)\n<\/pre><p>Notice that placing the restriction in the <code class=\"literal\">WHERE<\/code> clause produces a different result:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT * FROM t1 LEFT JOIN t2 ON t1.num = t2.num WHERE t2.value = 'xxx';<\/code><\/strong>\n num | name | num | value\n-----+------+-----+-------\n   1 | a    |   1 | xxx\n(1 row)\n<\/pre><p>This is because a restriction placed in the <code class=\"literal\">ON<\/code> clause is processed <em>before<\/em> the join, while a restriction placed in the <code class=\"literal\">WHERE<\/code> clause is processed <em>after<\/em> the join. That does not matter with inner joins, but it matters a lot with outer joins.<\/p><\/div><div class=\"sect3\" id=\"QUERIES-TABLE-ALIASES\"><div class=\"titlepage\"><div><h4 class=\"title\">7.2.1.2.\u00a0Table and Column Aliases<\/h4><\/div><\/div><p>A temporary name can be given to tables and complex table references to be used for references to the derived table in the rest of the query. This is called a <em class=\"firstterm\">table alias<\/em>.<\/p><p>To create a table alias, write<\/p><pre class=\"synopsis\">\nFROM <em class=\"replaceable\"><code>table_reference<\/code><\/em> AS <em class=\"replaceable\"><code>alias<\/code><\/em>\n<\/pre><p>or<\/p><pre class=\"synopsis\">\nFROM <em class=\"replaceable\"><code>table_reference<\/code><\/em> <em class=\"replaceable\"><code>alias<\/code><\/em>\n<\/pre><p>The <code class=\"literal\">AS<\/code> key word is optional noise. <em class=\"replaceable\"><code>alias<\/code><\/em> can be any identifier.<\/p><p>A typical application of table aliases is to assign short identifiers to long table names to keep the join clauses readable. For example:<\/p><pre class=\"programlisting\">\nSELECT * FROM some_very_long_table_name s JOIN another_fairly_long_name a ON s.id = a.num;\n<\/pre><p>The alias becomes the new name of the table reference so far as the current query is concerned \u2014 it is not allowed to refer to the table by the original name elsewhere in the query. Thus, this is not valid:<\/p><pre class=\"programlisting\">\nSELECT * FROM my_table AS m WHERE my_table.a &gt; 5;    -- wrong\n<\/pre><p>Table aliases are mainly for notational convenience, but it is necessary to use them when joining a table to itself, e.g.:<\/p><pre class=\"programlisting\">\nSELECT * FROM people AS mother JOIN people AS child ON mother.id = child.mother_id;\n<\/pre><p>Additionally, an alias is required if the table reference is a subquery (see <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-SUBQUERIES\" title=\"7.2.1.3.\u00a0Subqueries\">Section\u00a07.2.1.3<\/a>).<\/p><p>Parentheses are used to resolve ambiguities. In the following example, the first statement assigns the alias <code class=\"literal\">b<\/code> to the second instance of <code class=\"literal\">my_table<\/code>, but the second statement assigns the alias to the result of the join:<\/p><pre class=\"programlisting\">\nSELECT * FROM my_table AS a CROSS JOIN my_table AS b ...\nSELECT * FROM (my_table AS a CROSS JOIN my_table) AS b ...\n<\/pre><p>Another form of table aliasing gives temporary names to the columns of the table, as well as the table itself:<\/p><pre class=\"synopsis\">\nFROM <em class=\"replaceable\"><code>table_reference<\/code><\/em> [AS] <em class=\"replaceable\"><code>alias<\/code><\/em> ( <em class=\"replaceable\"><code>column1<\/code><\/em> [, <em class=\"replaceable\"><code>column2<\/code><\/em> [, ...]] )\n<\/pre><p>If fewer column aliases are specified than the actual table has columns, the remaining columns are not renamed. This syntax is especially useful for self-joins or subqueries.<\/p><p>When an alias is applied to the output of a <code class=\"literal\">JOIN<\/code> clause, the alias hides the original name(s) within the <code class=\"literal\">JOIN<\/code>. For example:<\/p><pre class=\"programlisting\">\nSELECT a.* FROM my_table AS a JOIN your_table AS b ON ...\n<\/pre><p>is valid SQL, but:<\/p><pre class=\"programlisting\">\nSELECT a.* FROM (my_table AS a JOIN your_table AS b ON ...) AS c\n<\/pre><p>is not valid; the table alias <code class=\"literal\">a<\/code> is not visible outside the alias <code class=\"literal\">c<\/code>.<\/p><\/div><div class=\"sect3\" id=\"QUERIES-SUBQUERIES\"><div class=\"titlepage\"><div><h4 class=\"title\">7.2.1.3.\u00a0Subqueries<\/h4><\/div><\/div><p>Subqueries specifying a derived table must be enclosed in parentheses and <em>must<\/em> be assigned a table alias name (as in <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-table-expressions.html#QUERIES-TABLE-ALIASES\" title=\"7.2.1.2.\u00a0Table and Column Aliases\">Section\u00a07.2.1.2<\/a>). For example:<\/p><pre class=\"programlisting\">\nFROM (SELECT * FROM table1) AS alias_name\n<\/pre><p>This example is equivalent to <code class=\"literal\">FROM table1 AS alias_name<\/code>. More interesting cases, which cannot be reduced to a plain join, arise when the subquery involves grouping or aggregation.<\/p><p>A subquery can also be a <code class=\"command\">VALUES<\/code> list:<\/p><pre class=\"programlisting\">\nFROM (VALUES ('anne', 'smith'), ('bob', 'jones'), ('joe', 'blow'))\n     AS names(first, last)\n<\/pre><p>Again, a table alias is required. Assigning alias names to the columns of the <code class=\"command\">VALUES<\/code> list is optional, but is good practice. For more information see <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-values.html\" title=\"7.7.\u00a0VALUES Lists\">Section\u00a07.7<\/a>.<\/p><\/div><div class=\"sect3\" id=\"QUERIES-TABLEFUNCTIONS\"><div class=\"titlepage\"><div><h4 class=\"title\">7.2.1.4.\u00a0Table Functions<\/h4><\/div><\/div><p>Table functions are functions that produce a set of rows, made up of either base data types (scalar types) or composite data types (table rows). They are used like a table, view, or subquery in the <code class=\"literal\">FROM<\/code> clause of a query. Columns returned by table functions can be included in <code class=\"literal\">SELECT<\/code>, <code class=\"literal\">JOIN<\/code>, or <code class=\"literal\">WHERE<\/code> clauses in the same manner as columns of a table, view, or subquery.<\/p><p>Table functions may also be combined using the <code class=\"literal\">ROWS FROM<\/code> syntax, with the results returned in parallel columns; the number of result rows in this case is that of the largest function result, with smaller results padded with null values to match.<\/p><pre class=\"synopsis\">\n<em class=\"replaceable\"><code>function_call<\/code><\/em> [WITH ORDINALITY] [[AS] <em class=\"replaceable\"><code>table_alias<\/code><\/em> [(<em class=\"replaceable\"><code>column_alias<\/code><\/em> [, ... ])]]\nROWS FROM( <em class=\"replaceable\"><code>function_call<\/code><\/em> [, ... ] ) [WITH ORDINALITY] [[AS] <em class=\"replaceable\"><code>table_alias<\/code><\/em> [(<em class=\"replaceable\"><code>column_alias<\/code><\/em> [, ... ])]]\n<\/pre><p>If the <code class=\"literal\">WITH ORDINALITY<\/code> clause is specified, an additional column of type <code class=\"type\">bigint<\/code> will be added to the function result columns. This column numbers the rows of the function result set, starting from 1. (This is a generalization of the SQL-standard syntax for <code class=\"literal\">UNNEST ... WITH ORDINALITY<\/code>.) By default, the ordinal column is called <code class=\"literal\">ordinality<\/code>, but a different column name can be assigned to it using an <code class=\"literal\">AS<\/code> clause.<\/p><p>The special table function <code class=\"literal\">UNNEST<\/code> may be called with any number of array parameters, and it returns a corresponding number of columns, as if <code class=\"literal\">UNNEST<\/code> (<a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/functions-array.html\" title=\"9.18.\u00a0Array Functions and Operators\">Section\u00a09.18<\/a>) had been called on each parameter separately and combined using the <code class=\"literal\">ROWS FROM<\/code> construct.<\/p><pre class=\"synopsis\">\nUNNEST( <em class=\"replaceable\"><code>array_expression<\/code><\/em> [, ... ] ) [WITH ORDINALITY] [[AS] <em class=\"replaceable\"><code>table_alias<\/code><\/em> [(<em class=\"replaceable\"><code>column_alias<\/code><\/em> [, ... ])]]\n<\/pre><p>If no <em class=\"replaceable\"><code>table_alias<\/code><\/em> is specified, the function name is used as the table name; in the case of a <code class=\"literal\">ROWS FROM()<\/code> construct, the first function's name is used.<\/p><p>If column aliases are not supplied, then for a function returning a base data type, the column name is also the same as the function name. For a function returning a composite type, the result columns get the names of the individual attributes of the type.<\/p><p>Some examples:<\/p><pre class=\"programlisting\">\nCREATE TABLE foo (fooid int, foosubid int, fooname text);\nCREATE FUNCTION getfoo(int) RETURNS SETOF foo AS $$\n    SELECT * FROM foo WHERE fooid = $1;\n$$ LANGUAGE SQL;\nSELECT * FROM getfoo(1) AS t1;\nSELECT * FROM foo\n    WHERE foosubid IN (\n                        SELECT foosubid\n                        FROM getfoo(foo.fooid) z\n                        WHERE z.fooid = foo.fooid\n                      );\nCREATE VIEW vw_getfoo AS SELECT * FROM getfoo(1);\nSELECT * FROM vw_getfoo;\n<\/pre><p>In some cases it is useful to define table functions that can return different column sets depending on how they are invoked. To support this, the table function can be declared as returning the pseudo-type <code class=\"type\">record<\/code>. When such a function is used in a query, the expected row structure must be specified in the query itself, so that the system can know how to parse and plan the query. This syntax looks like:<\/p><pre class=\"synopsis\">\n<em class=\"replaceable\"><code>function_call<\/code><\/em> [AS] <em class=\"replaceable\"><code>alias<\/code><\/em> (<em class=\"replaceable\"><code>column_definition<\/code><\/em> [, ... ])\n<em class=\"replaceable\"><code>function_call<\/code><\/em> AS [<em class=\"replaceable\"><code>alias<\/code><\/em>] (<em class=\"replaceable\"><code>column_definition<\/code><\/em> [, ... ])\nROWS FROM( ... <em class=\"replaceable\"><code>function_call<\/code><\/em> AS (<em class=\"replaceable\"><code>column_definition<\/code><\/em> [, ... ]) [, ... ] )\n<\/pre><p>When not using the <code class=\"literal\">ROWS FROM()<\/code> syntax, the <em class=\"replaceable\"><code>column_definition<\/code><\/em> list replaces the column alias list that could otherwise be attached to the <code class=\"literal\">FROM<\/code> item; the names in the column definitions serve as column aliases. When using the <code class=\"literal\">ROWS FROM()<\/code> syntax, a <em class=\"replaceable\"><code>column_definition<\/code><\/em> list can be attached to each member function separately; or if there is only one member function and no <code class=\"literal\">WITH ORDINALITY<\/code> clause, a <em class=\"replaceable\"><code>column_definition<\/code><\/em> list can be written in place of a column alias list following <code class=\"literal\">ROWS FROM()<\/code>.<\/p><p>Consider this example:<\/p><pre class=\"programlisting\">\nSELECT *\n    FROM dblink('dbname=mydb', 'SELECT proname, prosrc FROM pg_proc')\n      AS t1(proname name, prosrc text)\n    WHERE proname LIKE 'bytea%';\n<\/pre><p>The <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/contrib-dblink-function.html\" title=\"dblink\">dblink<\/a> function (part of the <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/dblink.html\" title=\"F.10.\u00a0dblink\">dblink<\/a> module) executes a remote query. It is declared to return <code class=\"type\">record<\/code> since it might be used for any kind of query. The actual column set must be specified in the calling query so that the parser knows, for example, what <code class=\"literal\">*<\/code> should expand to.<\/p><\/div><div class=\"sect3\" id=\"QUERIES-LATERAL\"><div class=\"titlepage\"><div><div><h4 class=\"title\">7.2.1.5.\u00a0<code class=\"literal\">LATERAL<\/code> Subqueries<\/h4><\/div><\/div><\/div><p>Subqueries appearing in <code class=\"literal\">FROM<\/code> can be preceded by the key word <code class=\"literal\">LATERAL<\/code>. This allows them to reference columns provided by preceding <code class=\"literal\">FROM<\/code> items. (Without <code class=\"literal\">LATERAL<\/code>, each subquery is evaluated independently and so cannot cross-reference any other <code class=\"literal\">FROM<\/code> item.)<\/p><p>Table functions appearing in <code class=\"literal\">FROM<\/code> can also be preceded by the key word <code class=\"literal\">LATERAL<\/code>, but for functions the key word is optional; the function's arguments can contain references to columns provided by preceding <code class=\"literal\">FROM<\/code> items in any case.<\/p><p>A <code class=\"literal\">LATERAL<\/code> item can appear at top level in the <code class=\"literal\">FROM<\/code> list, or within a <code class=\"literal\">JOIN<\/code> tree. In the latter case it can also refer to any items that are on the left-hand side of a <code class=\"literal\">JOIN<\/code> that it is on the right-hand side of.<\/p><p>When a <code class=\"literal\">FROM<\/code> item contains <code class=\"literal\">LATERAL<\/code> cross-references, evaluation proceeds as follows: for each row of the <code class=\"literal\">FROM<\/code> item providing the cross-referenced column(s), or set of rows of multiple <code class=\"literal\">FROM<\/code> items providing the columns, the <code class=\"literal\">LATERAL<\/code> item is evaluated using that row or row set's values of the columns. The resulting row(s) are joined as usual with the rows they were computed from. This is repeated for each row or set of rows from the column source table(s).<\/p><p>A trivial example of <code class=\"literal\">LATERAL<\/code> is<\/p><pre class=\"programlisting\">\nSELECT * FROM foo, LATERAL (SELECT * FROM bar WHERE bar.id = foo.bar_id) ss;\n<\/pre><p>This is not especially useful since it has exactly the same result as the more conventional<\/p><pre class=\"programlisting\">\nSELECT * FROM foo, bar WHERE bar.id = foo.bar_id;\n<\/pre><p><code class=\"literal\">LATERAL<\/code> is primarily useful when the cross-referenced column is necessary for computing the row(s) to be joined. A common application is providing an argument value for a set-returning function. For example, supposing that  returns the set of vertices of a polygon, we could identify close-together vertices of polygons stored in a table with:<\/p><pre class=\"programlisting\">\nSELECT p1.id, p2.id, v1, v2\nFROM polygons p1, polygons p2,\n     LATERAL vertices(p1.poly) v1,\n     LATERAL vertices(p2.poly) v2\nWHERE (v1 &lt;-&gt; v2) &lt; 10 AND p1.id != p2.id;\n<\/pre><p>This query could also be written<\/p><pre class=\"programlisting\">\nSELECT p1.id, p2.id, v1, v2\nFROM polygons p1 CROSS JOIN LATERAL vertices(p1.poly) v1,\n     polygons p2 CROSS JOIN LATERAL vertices(p2.poly) v2\nWHERE (v1 &lt;-&gt; v2) &lt; 10 AND p1.id != p2.id;\n<\/pre><p>or in several other equivalent formulations. (As already mentioned, the <code class=\"literal\">LATERAL<\/code> key word is unnecessary in this example, but we use it for clarity.)<\/p><p>It is often particularly handy to <code class=\"literal\">LEFT JOIN<\/code> to a <code class=\"literal\">LATERAL<\/code> subquery, so that source rows will appear in the result even if the <code class=\"literal\">LATERAL<\/code> subquery produces no rows for them. For example, if  returns the names of products made by a manufacturer, but some manufacturers in our table currently produce no products, we could find out which ones those are like this:<\/p><pre class=\"programlisting\">\nSELECT m.name\nFROM manufacturers m LEFT JOIN LATERAL get_product_names(m.id) pname ON true\nWHERE pname IS NULL;\n<\/pre><\/div><\/div><div class=\"sect2\" id=\"QUERIES-WHERE\"><div class=\"titlepage\"><div><div><h3 class=\"title\">7.2.2.\u00a0The <code class=\"literal\">WHERE<\/code> Clause<\/h3><\/div><\/div><\/div><p>The syntax of the <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/sql-select.html#SQL-WHERE\" title=\"WHERE Clause\"><code class=\"literal\">WHERE<\/code> Clause<\/a> is<\/p><pre class=\"synopsis\">\nWHERE <em class=\"replaceable\"><code>search_condition<\/code><\/em>\n<\/pre><p>where <em class=\"replaceable\"><code>search_condition<\/code><\/em> is any value expression (see <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/sql-expressions.html\" title=\"4.2.\u00a0Value Expressions\">Section\u00a04.2<\/a>) that returns a value of type <code class=\"type\">boolean<\/code>.<\/p><p>After the processing of the <code class=\"literal\">FROM<\/code> clause is done, each row of the derived virtual table is checked against the search condition. If the result of the condition is true, the row is kept in the output table, otherwise (i.e., if the result is false or null) it is discarded. The search condition typically references at least one column of the table generated in the <code class=\"literal\">FROM<\/code> clause; this is not required, but otherwise the <code class=\"literal\">WHERE<\/code> clause will be fairly useless.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>The join condition of an inner join can be written either in the <code class=\"literal\">WHERE<\/code> clause or in the <code class=\"literal\">JOIN<\/code> clause. For example, these table expressions are equivalent:<\/p><pre class=\"programlisting\">\nFROM a, b WHERE a.id = b.id AND b.val &gt; 5\n<\/pre><p>and:<\/p><pre class=\"programlisting\">\nFROM a INNER JOIN b ON (a.id = b.id) WHERE b.val &gt; 5\n<\/pre><p>or perhaps even:<\/p><pre class=\"programlisting\">\nFROM a NATURAL JOIN b WHERE b.val &gt; 5\n<\/pre><p>Which one of these you use is mainly a matter of style. The <code class=\"literal\">JOIN<\/code> syntax in the <code class=\"literal\">FROM<\/code> clause is probably not as portable to other SQL database management systems, even though it is in the SQL standard. For outer joins there is no choice: they must be done in the <code class=\"literal\">FROM<\/code> clause. The <code class=\"literal\">ON<\/code> or <code class=\"literal\">USING<\/code> clause of an outer join is <em>not<\/em> equivalent to a <code class=\"literal\">WHERE<\/code> condition, because it results in the addition of rows (for unmatched input rows) as well as the removal of rows in the final result.<\/p><\/div><p>Here are some examples of <code class=\"literal\">WHERE<\/code> clauses:<\/p><pre class=\"programlisting\">\nSELECT ... FROM fdt WHERE c1 &gt; 5\nSELECT ... FROM fdt WHERE c1 IN (1, 2, 3)\nSELECT ... FROM fdt WHERE c1 IN (SELECT c1 FROM t2)\nSELECT ... FROM fdt WHERE c1 IN (SELECT c3 FROM t2 WHERE c2 = fdt.c1 + 10)\nSELECT ... FROM fdt WHERE c1 BETWEEN (SELECT c3 FROM t2 WHERE c2 = fdt.c1 + 10) AND 100\nSELECT ... FROM fdt WHERE EXISTS (SELECT c1 FROM t2 WHERE c2 &gt; fdt.c1)\n<\/pre><p><code class=\"literal\">fdt<\/code> is the table derived in the <code class=\"literal\">FROM<\/code> clause. Rows that do not meet the search condition of the <code class=\"literal\">WHERE<\/code> clause are eliminated from <code class=\"literal\">fdt<\/code>. Notice the use of scalar subqueries as value expressions. Just like any other query, the subqueries can employ complex table expressions. Notice also how <code class=\"literal\">fdt<\/code> is referenced in the subqueries. Qualifying <code class=\"literal\">c1<\/code> as <code class=\"literal\">fdt.c1<\/code> is only necessary if <code class=\"literal\">c1<\/code> is also the name of a column in the derived input table of the subquery. But qualifying the column name adds clarity even when it is not needed. This example shows how the column naming scope of an outer query extends into its inner queries.<\/p><\/div><div class=\"sect2\" id=\"QUERIES-GROUP\"><div class=\"titlepage\"><div><div><h3 class=\"title\">7.2.3.\u00a0The <code class=\"literal\">GROUP BY<\/code> and <code class=\"literal\">HAVING<\/code> Clauses<\/h3><\/div><\/div><\/div><p>After passing the <code class=\"literal\">WHERE<\/code> filter, the derived input table might be subject to grouping, using the <code class=\"literal\">GROUP BY<\/code> clause, and elimination of group rows using the <code class=\"literal\">HAVING<\/code> clause.<\/p><pre class=\"synopsis\">\nSELECT <em class=\"replaceable\"><code>select_list<\/code><\/em>\n    FROM ...\n    [WHERE ...]\n    GROUP BY <em class=\"replaceable\"><code>grouping_column_reference<\/code><\/em> [, <em class=\"replaceable\"><code>grouping_column_reference<\/code><\/em>]...\n<\/pre><p>The <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/sql-select.html#SQL-GROUPBY\" title=\"GROUP BY Clause\"><code class=\"literal\">GROUP BY<\/code> Clause<\/a> is used to group together those rows in a table that have the same values in all the columns listed. The order in which the columns are listed does not matter. The effect is to combine each set of rows having common values into one group row that represents all rows in the group. This is done to eliminate redundancy in the output and\/or compute aggregates that apply to these groups. For instance:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT * FROM test1;<\/code><\/strong>\n x | y\n---+---\n a | 3\n c | 2\n b | 5\n a | 1\n(4 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT x FROM test1 GROUP BY x;<\/code><\/strong>\n x\n---\n a\n b\n c\n(3 rows)\n<\/pre><p>In the second query, we could not have written <code class=\"literal\">SELECT * FROM test1 GROUP BY x<\/code>, because there is no single value for the column <code class=\"literal\">y<\/code> that could be associated with each group. The grouped-by columns can be referenced in the select list since they have a single value in each group.<\/p><p>In general, if a table is grouped, columns that are not listed in <code class=\"literal\">GROUP BY<\/code> cannot be referenced except in aggregate expressions. An example with aggregate expressions is:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT x, sum(y) FROM test1 GROUP BY x;<\/code><\/strong>\n x | sum\n---+-----\n a |   4\n b |   5\n c |   2\n(3 rows)\n<\/pre><p>Here <code class=\"literal\">sum<\/code> is an aggregate function that computes a single value over the entire group. More information about the available aggregate functions can be found in <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/functions-aggregate.html\" title=\"9.20.\u00a0Aggregate Functions\">Section\u00a09.20<\/a>.<\/p><div class=\"tip\"><h3 class=\"title\">Tip<\/h3><p>Grouping without aggregate expressions effectively calculates the set of distinct values in a column. This can also be achieved using the <code class=\"literal\">DISTINCT<\/code> clause (see <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/queries-select-lists.html#QUERIES-DISTINCT\" title=\"7.3.3.\u00a0DISTINCT\">Section\u00a07.3.3<\/a>).<\/p><\/div><p>Here is another example: it calculates the total sales for each product (rather than the total sales of all products):<\/p><pre class=\"programlisting\">\nSELECT product_id, p.name, (sum(s.units) * p.price) AS sales\n    FROM products p LEFT JOIN sales s USING (product_id)\n    GROUP BY product_id, p.name, p.price;\n<\/pre><p>In this example, the columns <code class=\"literal\">product_id<\/code>, <code class=\"literal\">p.name<\/code>, and <code class=\"literal\">p.price<\/code> must be in the <code class=\"literal\">GROUP BY<\/code> clause since they are referenced in the query select list (but see below). The column <code class=\"literal\">s.units<\/code> does not have to be in the <code class=\"literal\">GROUP BY<\/code> list since it is only used in an aggregate expression (<code class=\"literal\">sum(...)<\/code>), which represents the sales of a product. For each product, the query returns a summary row about all sales of the product.<\/p><p>If the products table is set up so that, say, <code class=\"literal\">product_id<\/code> is the primary key, then it would be enough to group by <code class=\"literal\">product_id<\/code> in the above example, since name and price would be <em class=\"firstterm\">functionally dependent<\/em> on the product ID, and so there would be no ambiguity about which name and price value to return for each product ID group.<\/p><p>In strict SQL, <code class=\"literal\">GROUP BY<\/code> can only group by columns of the source table but PostgreSQL extends this to also allow <code class=\"literal\">GROUP BY<\/code> to group by columns in the select list. Grouping by value expressions instead of simple column names is also allowed.<\/p><p>If a table has been grouped using <code class=\"literal\">GROUP BY<\/code>, but only certain groups are of interest, the <code class=\"literal\">HAVING<\/code> clause can be used, much like a <code class=\"literal\">WHERE<\/code> clause, to eliminate groups from the result. The syntax is:<\/p><pre class=\"synopsis\">\nSELECT <em class=\"replaceable\"><code>select_list<\/code><\/em> FROM ... [WHERE ...] GROUP BY ... HAVING <em class=\"replaceable\"><code>boolean_expression<\/code><\/em>\n<\/pre><p>Expressions in the <code class=\"literal\">HAVING<\/code> clause can refer both to grouped expressions and to ungrouped expressions (which necessarily involve an aggregate function).<\/p><p>Example:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT x, sum(y) FROM test1 GROUP BY x HAVING sum(y) &gt; 3;<\/code><\/strong>\n x | sum\n---+-----\n a |   4\n b |   5\n(2 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT x, sum(y) FROM test1 GROUP BY x HAVING x &lt; 'c';<\/code><\/strong>\n x | sum\n---+-----\n a |   4\n b |   5\n(2 rows)\n<\/pre><p>Again, a more realistic example:<\/p><pre class=\"programlisting\">\nSELECT product_id, p.name, (sum(s.units) * (p.price - p.cost)) AS profit\n    FROM products p LEFT JOIN sales s USING (product_id)\n    WHERE s.date &gt; CURRENT_DATE - INTERVAL '4 weeks'\n    GROUP BY product_id, p.name, p.price, p.cost\n    HAVING sum(p.price * s.units) &gt; 5000;\n<\/pre><p>In the example above, the <code class=\"literal\">WHERE<\/code> clause is selecting rows by a column that is not grouped (the expression is only true for sales during the last four weeks), while the <code class=\"literal\">HAVING<\/code> clause restricts the output to groups with total gross sales over 5000. Note that the aggregate expressions do not necessarily need to be the same in all parts of the query.<\/p><p>If a query contains aggregate function calls, but no <code class=\"literal\">GROUP BY<\/code> clause, grouping still occurs: the result is a single group row (or perhaps no rows at all, if the single row is then eliminated by <code class=\"literal\">HAVING<\/code>). The same is true if it contains a <code class=\"literal\">HAVING<\/code> clause, even without any aggregate function calls or <code class=\"literal\">GROUP BY<\/code> clause.<\/p><\/div><div class=\"sect2\" id=\"QUERIES-GROUPING-SETS\"><div class=\"titlepage\"><div><div><h3 class=\"title\">7.2.4.\u00a0<code class=\"literal\">GROUPING SETS<\/code>, <code class=\"literal\">CUBE<\/code>, and <code class=\"literal\">ROLLUP<\/code><\/h3><\/div><\/div><\/div><p>More complex grouping operations than those described above are possible using the concept of <em class=\"firstterm\">grouping sets<\/em>. The data selected by the <code class=\"literal\">FROM<\/code> and <code class=\"literal\">WHERE<\/code> clauses is grouped separately by each specified grouping set, aggregates computed for each group just as for simple <code class=\"literal\">GROUP BY<\/code> clauses, and then the results returned. For example:<\/p><pre class=\"prompt\">=&gt; <strong class=\"userinput\"><code>SELECT * FROM items_sold;<\/code><\/strong>\n brand | size | sales\n-------+------+-------\n Foo   | L    |  10\n Foo   | M    |  20\n Bar   | M    |  15\n Bar   | L    |  5\n(4 rows)\n<code class=\"prompt\">=&gt;<\/code> <strong class=\"userinput\"><code>SELECT brand, size, sum(sales) FROM items_sold GROUP BY GROUPING SETS ((brand), (size), ());<\/code><\/strong>\n brand | size | sum\n-------+------+-----\n Foo   |      |  30\n Bar   |      |  20\n       | L    |  15\n       | M    |  35\n       |      |  50\n(5 rows)\n<\/pre><p>Each sublist of <code class=\"literal\">GROUPING SETS<\/code> may specify zero or more columns or expressions and is interpreted the same way as though it were directly in the <code class=\"literal\">GROUP BY<\/code> clause. An empty grouping set means that all rows are aggregated down to a single group (which is output even if no input rows were present), as described above for the case of aggregate functions with no <code class=\"literal\">GROUP BY<\/code> clause.<\/p><p>References to the grouping columns or expressions are replaced by null values in result rows for grouping sets in which those columns do not appear. To distinguish which grouping a particular output row resulted from, see <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/functions-aggregate.html#FUNCTIONS-GROUPING-TABLE\" title=\"Table\u00a09.56.\u00a0Grouping Operations\">Table\u00a09.56<\/a>.<\/p><p>A shorthand notation is provided for specifying two common types of grouping set. A clause of the form<\/p><pre class=\"programlisting\">\nROLLUP ( <em class=\"replaceable\"><code>e1<\/code><\/em>, <em class=\"replaceable\"><code>e2<\/code><\/em>, <em class=\"replaceable\"><code>e3<\/code><\/em>, ... )\n<\/pre><p>represents the given list of expressions and all prefixes of the list including the empty list; thus it is equivalent to<\/p><pre class=\"programlisting\">\nGROUPING SETS (\n    ( <em class=\"replaceable\"><code>e1<\/code><\/em>, <em class=\"replaceable\"><code>e2<\/code><\/em>, <em class=\"replaceable\"><code>e3<\/code><\/em>, ... ),\n    ...\n    ( <em class=\"replaceable\"><code>e1<\/code><\/em>, <em class=\"replaceable\"><code>e2<\/code><\/em> ),\n    ( <em class=\"replaceable\"><code>e1<\/code><\/em> ),\n    ( )\n)\n<\/pre><p>This is commonly used for analysis over hierarchical data; e.g. total salary by department, division, and company-wide total.<\/p><p>A clause of the form<\/p><pre class=\"programlisting\">\nCUBE ( <em class=\"replaceable\"><code>e1<\/code><\/em>, <em class=\"replaceable\"><code>e2<\/code><\/em>, ... )\n<\/pre><p>represents the given list and all of its possible subsets (i.e. the power set). Thus<\/p><pre class=\"programlisting\">\nCUBE ( a, b, c )\n<\/pre><p>is equivalent to<\/p><pre class=\"programlisting\">\nGROUPING SETS (\n    ( a, b, c ),\n    ( a, b    ),\n    ( a,    c ),\n    ( a       ),\n    (    b, c ),\n    (    b    ),\n    (       c ),\n    (         )\n)\n<\/pre><p>The individual elements of a <code class=\"literal\">CUBE<\/code> or <code class=\"literal\">ROLLUP<\/code> clause may be either individual expressions, or sublists of elements in parentheses. In the latter case, the sublists are treated as single units for the purposes of generating the individual grouping sets. For example:<\/p><pre class=\"programlisting\">\nCUBE ( (a, b), (c, d) )\n<\/pre><p>is equivalent to<\/p><pre class=\"programlisting\">\nGROUPING SETS (\n    ( a, b, c, d ),\n    ( a, b       ),\n    (       c, d ),\n    (            )\n)\n<\/pre><p>and<\/p><pre class=\"programlisting\">\nROLLUP ( a, (b, c), d )\n<\/pre><p>is equivalent to<\/p><pre class=\"programlisting\">\nGROUPING SETS (\n    ( a, b, c, d ),\n    ( a, b, c    ),\n    ( a          ),\n    (            )\n)\n<\/pre><p>The <code class=\"literal\">CUBE<\/code> and <code class=\"literal\">ROLLUP<\/code> constructs can be used either directly in the <code class=\"literal\">GROUP BY<\/code> clause, or nested inside a <code class=\"literal\">GROUPING SETS<\/code> clause. If one <code class=\"literal\">GROUPING SETS<\/code> clause is nested inside another, the effect is the same as if all the elements of the inner clause had been written directly in the outer clause.<\/p><p>If multiple grouping items are specified in a single <code class=\"literal\">GROUP BY<\/code> clause, then the final list of grouping sets is the cross product of the individual items. For example:<\/p><pre class=\"programlisting\">\nGROUP BY a, CUBE (b, c), GROUPING SETS ((d), (e))\n<\/pre><p>is equivalent to<\/p><pre class=\"programlisting\">\nGROUP BY GROUPING SETS (\n    (a, b, c, d), (a, b, c, e),\n    (a, b, d),    (a, b, e),\n    (a, c, d),    (a, c, e),\n    (a, d),       (a, e)\n)\n<\/pre><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>The construct <code class=\"literal\">(a, b)<\/code> is normally recognized in expressions as a <a class=\"link\" href=\"https:\/\/www.postgresql.org\/docs\/11\/sql-expressions.html#SQL-SYNTAX-ROW-CONSTRUCTORS\" title=\"4.2.13.\u00a0Row Constructors\">row constructor<\/a>. Within the <code class=\"literal\">GROUP BY<\/code> clause, this does not apply at the top levels of expressions, and <code class=\"literal\">(a, b)<\/code> is parsed as a list of expressions as described above. If for some reason you <em>need<\/em> a row constructor in a grouping expression, use <code class=\"literal\">ROW(a, b)<\/code>.<\/p><\/div><\/div><div class=\"sect2\" id=\"QUERIES-WINDOW\"><div class=\"titlepage\"><div><h3 class=\"title\">7.2.5.\u00a0Window Function Processing<\/h3><\/div><\/div><p>If the query contains any window functions (see <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/tutorial-window.html\" title=\"3.5.\u00a0Window Functions\">Section\u00a03.5<\/a>, <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/functions-window.html\" title=\"9.21.\u00a0Window Functions\">Section\u00a09.21<\/a> and <a class=\"xref\" href=\"https:\/\/www.postgresql.org\/docs\/11\/sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS\" title=\"4.2.8.\u00a0Window Function Calls\">Section\u00a04.2.8<\/a>), these functions are evaluated after any grouping, aggregation, and <code class=\"literal\">HAVING<\/code> filtering is performed. That is, if the query uses any aggregates, <code class=\"literal\">GROUP BY<\/code>, or <code class=\"literal\">HAVING<\/code>, then the rows seen by the window functions are the group rows instead of the original table rows from <code class=\"literal\">FROM<\/code>\/<code class=\"literal\">WHERE<\/code>.<\/p><p>When multiple window functions are used, all the window functions having syntactically equivalent <code class=\"literal\">PARTITION BY<\/code> and <code class=\"literal\">ORDER BY<\/code> clauses in their window definitions are guaranteed to be evaluated in a single pass over the data. Therefore they will see the same sort ordering, even if the <code class=\"literal\">ORDER BY<\/code> does not uniquely determine an ordering. However, no guarantees are made about the evaluation of functions having different <code class=\"literal\">PARTITION BY<\/code> or <code class=\"literal\">ORDER BY<\/code> specifications. (In such cases a sort step is typically required between the passes of window function evaluations, and the sort is not guaranteed to preserve ordering of rows that its <code class=\"literal\">ORDER BY<\/code> sees as equivalent.)<\/p><p>Currently, window functions always require presorted data, and so the query output will be ordered according to one or another of the window functions' <code class=\"literal\">PARTITION BY<\/code>\/<code class=\"literal\">ORDER BY<\/code> clauses. It is not recommended to rely on this, however. Use an explicit top-level <code class=\"literal\">ORDER BY<\/code> clause if you want to be sure the results are sorted in a particular way.<\/p><\/div>","created_at":"2019-05-02T14:26:39+0000","updated_at":"2019-05-02T14:26:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":27,"domain_name":"www.postgresql.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":928,"title":"How To: Loop Pedal With Ableton Live","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=cFQzc8bYE2g","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/cFQzc8bYE2g?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-04-30T21:01:40+0000","updated_at":"2019-04-30T21:01:40+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/cFQzc8bYE2g\/maxresdefault.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["music","ableton","looping"],"is_public":false,"id":927,"title":"Live-looping With Ableton &amp; MIDI","url":"https:\/\/www.dawsons.co.uk\/blog\/live-looping-ableton-midi","content":"<h2>Using Ableton\u2019s Looper Device As A Performance Tool<\/h2>\n<p>Many players now are exploring with the possibilities offered by looping. Even popstars\u00a0like Ed Sheeran and KT Tunstall are using loops as the centre-piece of their sound.<\/p>\n<p>For guitar players particularly, looping tends to come in the shape of a dedicated foot pedal. But what\u2019s interesting for more studio-based or electronic musicians is the functionality offered by <a href=\"https:\/\/www.dawsons.co.uk\/catalogsearch\/result\/?q=ableton\" rel=\"noopener\">Ableton Live<\/a>. Built into Ableton is a powerful, intuitive looping device which offers much of the same practicality as an external looper. Combined with a simple MIDI footswitch, you can do pretty much everything you could do with something like the <a class=\"inblog\" href=\"https:\/\/www.dawsons.co.uk\/boss-rc-3\">Boss RC-3<\/a>.<\/p>\n<p>The real beauty comes in the fact that, as well as the looping, you will also be plugged into Ableton with its incredible features. It\u2019s the best of both worlds; traditional looping, and the full capability of Live including recording, effects, instruments and sounds.<\/p>\n<p>Let\u2019s have a look at how to start live-looping with <a href=\"https:\/\/www.dawsons.co.uk\/catalogsearch\/result\/?q=ableton\" rel=\"noopener\">Ableton<\/a>.<\/p>\n<p><img class=\"aligncenter\" src=\"https:\/\/www.dawsons.co.uk\/blog\/wp-content\/uploads\/2017\/01\/LiveTech_03-0110.jpg\" alt=\"Ableton Live Looper\" width=\"450\" height=\"224\" \/><\/p>\n<h3>Getting started<\/h3>\n<p>It\u2019s worth familiarising yourself with the looper tool in Ableton before we start. You\u2019ll find it included in Ableton\u2019s audio effects panel. There are a couple of options we\u2019ll consider; looping with an instrument (guitar would be the obvious example) and looping using a MIDI instrument. If you\u2019re using a guitar, drag the looper onto the audio track into which your guitar is connected. If you\u2019re using a virtual instrument or VSTi, load that into a MIDI track then drop the looper in after it.<\/p>\n<p>You\u2019ll notice straight away the big button for recording. Before you fly into action, take a minute to get things set up the way you want them. The first priority should be mapping that big button to a MIDI controller. While you can use anything with MIDI capability, it makes sense given the circumstances to use some kind of foot pedal. A simple MIDI enabled foot pedal will do the job here, you don\u2019t need anything too fancy.<\/p>\n<p>Under its default settings, the foot pedal will start recording on the first press. You can then set it to either immediately record the next layer with the next press \u2013 great for quickly building up layers \u2013 or to stop recording but keep playing the loop. This setting is perfect for anyone looking to lay down a basic chord progression and jam over the top. You can then double-tap the switch to stop playing, and press &amp; hold to clear the recording and start afresh.<\/p>\n<h3>Behave yourself<\/h3>\n<p>The next job is to tell the looper how you want it to behave. You could have it run to a set tempo, dictated by Ableton\u2019s global tempo control; this is useful if you\u2019re using virtual instruments as you can use the in-built metronome for accuracy. Ableton will also quantize the recording, meaning it will cleverly \u2018shift\u2019 any slightly out of time notes to make them in time. Alternatively, you could set the looper to base its tempo on your playing, which it does with a decent level of accuracy.<\/p>\n<p><a href=\"https:\/\/www.dawsons.co.uk\/blog\/wp-content\/uploads\/2017\/02\/ableton-standard.jpg\"><img src=\"https:\/\/www.dawsons.co.uk\/blog\/wp-content\/uploads\/2017\/02\/ableton-standard.jpg\" alt=\"Ableton\" width=\"450\" height=\"424\" class=\"aligncenter\" \/><\/a><\/p>\n<p>From here, it\u2019s pretty easy to get up and running. But let\u2019s say you wanted to record more than one instrument. You might, for example, want to experiment with a full band sound. The process here is slightly more complex, but with the same overall principles. Just more of them.<\/p>\n<h3>Multi-instrument setup<\/h3>\n<p>To give an example, create two audio tracks and two MIDI tracks. Load your favourite drum machine into one MIDI track, and a synth bass into another. Then, in the audio tracks, route in a guitar and a microphone. Now it gets a bit more complex. You could, for ease of use, load the looper device into a bus track and route the four individual tracks solely into the bus; this would capture the \u2018overall\u2019 sound but doesn\u2019t leave much room for controlling the different components.<\/p>\n<p>Really, each track needs its own looper device, but the problem here is that the foot pedal would start recording on all four simultaneously unless you have an alternative solution. The best method to get the full benefit would be to use an external MIDI controller like the <a class=\"inblog\" href=\"https:\/\/www.dawsons.co.uk\/novation-launchkey-25-mk-2-controller-keyboard\">Novation Launchkey 25<\/a>. Using the Launchkey, you can assign the \u2018arm record\u2019 buttons on the individual tracks to buttons on the keyboard. Then, you\u2019d load individual instances of the looper device onto each track, and simply \u2018arm\u2019 the track you want to focus on at that time.<\/p>\n<p>There will always be demand for external, dedicated hardware loop pedals. They offer portability, and don\u2019t require laptops or other devices to operate. But for those of you with access to a copy of Ableton, the in-built looper device is a powerful, controllable little tool which opens up a world of new possibilities for recording. And, being a full DAW application, any great ideas you stumble across can be built up extremely quickly into full tracks.<\/p>","created_at":"2019-04-30T18:13:20+0000","updated_at":"2019-04-30T18:13:26+0000","published_at":"2017-02-19T09:21:45+0000","published_by":["Chris Corfield"],"annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"www.dawsons.co.uk","preview_picture":"https:\/\/www.dawsons.co.uk\/blog\/wp-content\/uploads\/2018\/07\/iStock-888374940.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":926,"title":"cknd\/stackprinter","url":"https:\/\/github.com\/cknd\/stackprinter","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p>This prints tracebacks \/ call stacks with code context and the values of nearby variables. It answers most of the questions I'd ask an interactive debugger: Where in the code did it happen, what's in the relevant local variables, and why was <em>that<\/em> function called with <em>those<\/em> arguments.<\/p>\n<p>It's not a fully-grown <a href=\"https:\/\/sentry.io\/welcome\/\" rel=\"nofollow\">error monitoring system<\/a>, just a more helpful version of Python's built-in crash message. I sometimes use it locally instead of a debugger, but mostly it helps me sleep when my code runs somewhere where the only debug tool is a log file.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>pip install stackprinter<\/pre><\/div>\n<h3><a id=\"user-content-before\" class=\"anchor\" aria-hidden=\"true\" href=\"#before\"><\/a>Before<\/h3>\n<pre>Traceback (most recent call last):\n  File \"demo.py\", line 10, in &lt;module&gt;\n    dangerous_function(somelist + anotherlist)\n  File \"demo.py\", line 4, in dangerous_function\n    return sorted(blub, key=lambda xs: sum(xs))\n  File \"demo.py\", line 4, in &lt;lambda&gt;\n    return sorted(blub, key=lambda xs: sum(xs))\nTypeError: unsupported operand type(s) for +: 'int' and 'str'\n<\/pre>\n<h3><a id=\"user-content-after\" class=\"anchor\" aria-hidden=\"true\" href=\"#after\"><\/a>After<\/h3>\n<pre>File demo.py, line 10, in &lt;module&gt;\n    8         somelist = [[1,2], [3,4]]\n    9         anotherlist = [['5', 6]]\n--&gt; 10        dangerous_function(somelist + anotherlist)\n    11    except:\n    ..................................................\n     somelist = [[1, 2], [3, 4]]\n     anotherlist = [['5', 6]]\n    ..................................................\nFile demo.py, line 4, in dangerous_function\n    3     def dangerous_function(blub):\n--&gt; 4         return sorted(blub, key=lambda xs: sum(xs))\n    ..................................................\n     blub = [[1, 2], [3, 4], ['5', 6]]\n    ..................................................\nFile demo.py, line 4, in &lt;lambda&gt;\n    2\n    3     def dangerous_function(blub):\n--&gt; 4         return sorted(blub, key=lambda xs: sum(xs))\n    5\n    ..................................................\n     xs = ['5', 6]\n    ..................................................\nTypeError: unsupported operand type(s) for +: 'int' and 'str'\n<\/pre>\n<p>By default, it tries to be somewhat polite about screen space. (It only shows a few source lines and the function header, and only the variables in the visible code, and only (?) 500 characters per variable). You can <a href=\"https:\/\/github.com\/cknd\/stackprinter\/blob\/master\/stackprinter\/__init__.py#L82-L127\">configure<\/a> exactly how verbose things should be. It also attempts advanced stunts like \"dot attribute lookups\", \"showing the shape of numpy arrays\" and <a href=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/darkbg.png\" rel=\"nofollow\">\"colors\"<\/a>.<\/p>\n<h2><a id=\"user-content-exception-logging\" class=\"anchor\" aria-hidden=\"true\" href=\"#exception-logging\"><\/a>Exception logging<\/h2>\n<p>To globally replace the default python crash message, call <code>set_excepthook()<\/code> somewhere. This will print any uncaught exception to stderr by default.<\/p>\n<div class=\"highlight highlight-source-python\"><pre>import stackprinter\nstackprinter.set_excepthook(style='color')<\/pre><\/div>\n<p>To see a specific exception, call <a href=\"https:\/\/github.com\/cknd\/stackprinter\/blob\/master\/stackprinter\/__init__.py#L154-L162\"><code>show()<\/code><\/a> or <a href=\"https:\/\/github.com\/cknd\/stackprinter\/blob\/master\/stackprinter\/__init__.py#L28-L137\"><code>format()<\/code><\/a> inside an <code>except<\/code> block. <code>show()<\/code> prints to stderr by default, <code>format()<\/code> just returns a string, for custom logging.<\/p>\n<div class=\"highlight highlight-source-python\"><pre>try:\n    something()\nexcept:\n    # grab the current exception, print the traceback to stderr:\n    stackprinter.show()\n    # ...or only get a string, e.g. for logging:\n    logger.error(stackprinter.format())<\/pre><\/div>\n<p>You can also pass a previously caught exception object explicitly.<\/p>\n<div class=\"highlight highlight-source-python\"><pre># or explicitely grab a particular exception\ntry:\n    something()\nexcept ValueError as e:\n    stackprinter.show(e)  # or format(e)\n<\/pre><\/div>\n<div class=\"highlight highlight-source-python\"><pre># or collect exceptions in a little jar somewhere, to log them later\ntry:\n    something()\nexcept ValueError as e:\n    errors.append(e)\n# later:\nfor err in errors:\n    message = stackprinter.format(err)\n    logging.log(message)<\/pre><\/div>\n<p>You can blacklist certain file paths, to make the stack less verbose whenever it runs through those files. For example, calling <code>show(exc, suppressed_paths=[r\"lib\/python.*\/site-packages\"])<\/code> shrinks calls within libraries to one line each.<\/p>\n<p>By default, the output is plain text, which is good for log files. For some reason, there is also a color mode \ud83c\udf08, enabled by passing <code>style='darkbg'<\/code> or <code>style='lightbg'<\/code> to any of these methods (or <code>'darkbg2'<\/code> or <code>'lightbg2'<\/code>). It's an attempt at <a href=\"https:\/\/medium.com\/@brianwill\/making-semantic-highlighting-useful-9aeac92411df\" rel=\"nofollow\">semantic highlighting<\/a>, i.e. the colors follow the different variables instead of the syntax, like so:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/darkbg.png\"><img src=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/darkbg.png\" width=\"400\" alt=\"image\" \/><\/a>   <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/notebook.png\"><img src=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/notebook.png\" width=\"400\" alt=\"image\" \/><\/a><\/p>\n<p>For more config etc, for now, <a href=\"https:\/\/github.com\/cknd\/stackprinter\/blob\/master\/stackprinter\/__init__.py#L82-L127\">see the docstring of <code>format()<\/code><\/a>.<\/p>\n<h2><a id=\"user-content-printing-the-stack-of-the-current-thread\" class=\"anchor\" aria-hidden=\"true\" href=\"#printing-the-stack-of-the-current-thread\"><\/a>Printing the stack of the current thread<\/h2>\n<p>To see your own thread's current call stack, call <code>show<\/code> or <code>format<\/code> anywhere outside of exception handling.<\/p>\n<div class=\"highlight highlight-source-python\"><pre>stackprinter.show() # or format()<\/pre><\/div>\n<h2><a id=\"user-content-printing-the-current-stack-of-another-thread\" class=\"anchor\" aria-hidden=\"true\" href=\"#printing-the-current-stack-of-another-thread\"><\/a>Printing the current stack of another thread<\/h2>\n<p>To inspect the call stack of any other running thread:<\/p>\n<div class=\"highlight highlight-source-python\"><pre>thread = threading.Thread(target=something)\nthread.start()\n# (...)\nstackprinter.show(thread) # or format(thread)<\/pre><\/div>\n<p>Basically, this is a frame formatter. For each <a href=\"https:\/\/en.wikipedia.org\/wiki\/Call_stack\" rel=\"nofollow\">frame on the call stack<\/a>, it grabs the source code to find out which source lines reference which variables. Then it displays code and variables in the neighbourhood of the last executed line.<\/p>\n<p>Since this already requires a map of where each variable occurs in the code, it was difficult not to also implement the whole semantic highlighting color thing seen in the screenshots. The colors are ANSI escape codes now, but it should be fairly straightforward\u2122 to render the underlying data without any 1980ies terminal technology. Say, a foldable and clickable HTML page with downloadable pickled variables. For now you'll have to pipe the ANSI strings through <a href=\"https:\/\/github.com\/ralphbean\/ansi2html\/\">ansi2html<\/a> or something.<\/p>\n<p>The format and everything is inspired by the excellent <a href=\"https:\/\/ipython.readthedocs.io\/en\/stable\/api\/generated\/IPython.core.ultratb.html\" rel=\"nofollow\"><code>ultratb<\/code><\/a> in IPython. One day I'd like to contribute the whole \"find out which variables in <code>locals<\/code> and <code>globals<\/code> are nearby in the source and print only those\" machine over there, after trimming its complexity a bit.<\/p>\n<p>This displays variable values as they are <em>at the time of formatting<\/em>. In\nmulti-threaded programs, variables can change while we're busy walking\nthe stack &amp; printing them. So, if nothing seems to make sense, consider that\nyour exception and the traceback messages are from slightly different times.\nSadly, there is no responsible way to freeze all other threads as soon\nas we want to inspect some thread's call stack (...or is there?)<\/p>\n<h2><a id=\"user-content-tracing-a-piece-of-code-as-it-is-executed\" class=\"anchor\" aria-hidden=\"true\" href=\"#tracing-a-piece-of-code-as-it-is-executed\"><\/a>Tracing a piece of code as it is executed<\/h2>\n<p>More for curiosity than anything else, you can watch a piece of code execute step-by-step, printing a trace of all calls &amp; returns 'live' as they are happening. Slows everything down though, of course.<\/p>\n<div class=\"highlight highlight-source-python\"><pre>with stackprinter.TracePrinter(style='darkbg'):\n    dosomething()<\/pre><\/div>\n<p>or<\/p>\n<div class=\"highlight highlight-source-python\"><pre>tp = stackprinter.TracePrinter(style='darkbg')\ntp.enable()\ndosomething()\n# (...) +1 million lines\ntp.disable()<\/pre><\/div>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/trace.png\"><img src=\"https:\/\/raw.githubusercontent.com\/cknd\/stackprinter\/master\/trace.png\" width=\"300\" alt=\"image\" \/><\/a><\/p>\n<p>*coughs*<\/p>\n<p>For now, just look at all the doc strings, <a href=\"https:\/\/github.com\/cknd\/stackprinter\/blob\/master\/stackprinter\/__init__.py#L28-L137\">e.g. those of <code>format()<\/code><\/a><\/p>\n<\/article>","created_at":"2019-04-28T20:28:31+0000","updated_at":"2019-04-28T20:28:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/486707?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":925,"title":"(ollie {:on :parens}): Clojure projects from scratch","url":"https:\/\/oli.me.uk\/2018-02-26-clojure-projects-from-scratch\/","content":"<p>This post is intended <i>primarily<\/i> for two groups of people:<\/p><ol><li>People just starting out in Clojure, who know parts of the language but don't know how to begin structuring a real project.<\/li><li>Seasoned Clojurians who wish to see how to structure a project with the new Clojure CLI + <code>deps.edn<\/code> and not <a href=\"https:\/\/leiningen.org\/\">lein<\/a> or <a href=\"http:\/\/boot-clj.com\/\">boot<\/a> (even though they're still awesome).<\/li><\/ol><p>My goal is to teach you how to go from an empty directory, to a project you can run, test, compile and (if you wish to) publish with ease. We're going to get there through a series of relatively small steps so you can understand all the tools you're using.<\/p><p>There won't be an awful lot of Clojure code here, so don't worry if you're still getting your head around the language itself. The only code example will be a \"Hello, World!\", if that helps.<\/p><blockquote><p> Note: This post assumes usage of Linux, OSX or similar, I'm afraid it isn't intended for Windows users since I just don't have the knowledge to help you there. Some of the information will apply, but you'll have to adapt things, I'm sure you can find Windows specific guides for the parts that don't fit. <\/p><\/blockquote><p>I'm not going into what editor you should use because that's a book in itself. If you're totally at a loss, check out <a href=\"https:\/\/cursive-ide.com\/\">Cursive<\/a>, although I use <a href=\"http:\/\/spacemacs.org\/\">Spacemacs<\/a> because I can't survive without good Vim emulation. There's probably a great plugin for your editor of choice and instructions on getting started, have a Google.<\/p><p>A lot of what I'm going to be talking about can be found in practice in <a href=\"https:\/\/github.com\/robert-stuttaford\/bridge\">github.com\/robert-stuttaford\/bridge<\/a>, you may want to have a peruse at some point.<\/p><h2 id=\"installing_the_clojure_cli\">Installing the Clojure CLI<\/h2><p>To run Clojure you'll need the command line tool (introduced around the time of Clojure 1.9) that manages dependencies and allows you execute code.<\/p><p>If you're on OSX, you can use <code>brew<\/code> to install the CLI.<\/p><pre class=\"bash\">$ brew install clojure\n<\/pre><p>I have found that I could install it through the Arch Linux package manager although it was slightly out of date at the time of writing, so I don't recommend this just yet. If you're on Linux you can run the manual installer easily enough.<\/p><pre class=\"bash\">$ curl -O https:\/\/download.clojure.org\/install\/linux-install-1.9.0.326.sh\n$ sudo bash linux-install-1.9.0.326.sh\n<\/pre><p>To update, use the package manager you used for the installation or find the latest Linux installer URL on the <a href=\"https:\/\/clojure.org\/guides\/getting_started\">getting started<\/a> page.<\/p><p>You should now be able to drop into a Clojure REPL with one command. You can run <code>clojure<\/code> or <code>clj<\/code> in your terminal, the latter has a slightly better editing experience but requires you to have <code>rlwrap<\/code> installed.<\/p><pre class=\"bash\">$ clj\nClojure 1.9.0\nuser=&gt; (+ 10 15)\n25\n<\/pre><h2 id=\"initial_files\">Initial files<\/h2><p>Presuming our project is called <code>hey<\/code>, let's go ahead and create these directories and files:<\/p><pre class=\"bash\">$ mkdir -p hey\/{src\/hey,test\/hey}\n$ cd hey\n$ touch src\/hey\/core.clj test\/hey\/core_test.clj\n<\/pre><p>This provides us with the following directory structure:<\/p><pre>$ tree\n.\n\u251c\u2500\u2500 src\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 hey\n\u2502\u00a0\u00a0     \u2514\u2500\u2500 core.clj\n\u2514\u2500\u2500 test\n    \u2514\u2500\u2500 hey\n        \u2514\u2500\u2500 core_test.clj\n4 directories, 2 files\n<\/pre><p>Let's insert some content into these files:<\/p><h3 id=\"src\/hey\/core.clj\">src\/hey\/core.clj<\/h3><pre class=\"clojure\">(ns hey.core)\n(defn -main []\n  (println \"Hello, World!\"))\n<\/pre><h3 id=\"test\/hey\/core_test.clj\">test\/hey\/core_test.clj<\/h3><pre class=\"clojure\">(ns hey.core-test\n  (:require [clojure.test :as t]\n            [hey.core :as sut]))\n(t\/deftest basic-tests\n  (t\/testing \"it says hello to everyone\"\n    (t\/is (= (with-out-str (sut\/-main)) \"Hello, World!\\n\"))))\n<\/pre><p>The main namespace simply prints \"Hello, World!\" when executed and the test confirms that functionality.<\/p><h2 id=\"running_your_code\">Running your code<\/h2><p>Now that we have a bare bones program and test file in our project directory, we're probably going to want to run it. We can do that with the Clojure CLI, go ahead and execute the following:<\/p><pre class=\"bash\">$ clj -m hey.core\n<\/pre><p>You should see \"Hello, World!\" printed in your terminal. Let's try jumping into a REPL so we can interact with our code directly:<\/p><pre class=\"bash\">$ clj\nClojure 1.9.0\nuser=&gt; (load \"hey\/core\")\nnil\nuser=&gt; (in-ns 'hey.core)\n#object[clojure.lang.Namespace 0x2072acb2 \"hey.core\"]\nhey.core=&gt; (-main)\nHello, World!\nnil\n<\/pre><p>If your editor supports Clojure, you can probably connect a REPL and interact with your code through there too. With spacemacs I would type <code>,'<\/code> to \"jack in\" with CIDER. I can then use <code>,ee<\/code> to evaluate expressions as I work.<\/p><h2 id=\"testing\">Testing<\/h2><p>We have a test file but no way to run it. We could create our own test runner namespace that executed <code>clojure.test\/run-all-tests<\/code>, but that requires telling it about every testing namespace we have in our project. It gets tedious after a while, so let's get something that does it for us.<\/p><p>Create a file called <code>deps.edn<\/code> at the top of your project and add the following to it:<\/p><h3 id=\"deps.edn\">deps.edn<\/h3><pre class=\"clojure\">{:deps\n {org.clojure\/clojure {:mvn\/version \"1.9.0\"}}\n :aliases\n {:test\n  {:extra-paths [\"test\"]\n   :extra-deps\n   {com.cognitect\/test-runner {:git\/url \"git@github.com:cognitect-labs\/test-runner\"\n                               :sha \"5f2b5c2efb444df76fb5252102b33f542ebf7f58\"}}\n   :main-opts [\"-m\" \"cognitect.test-runner\"]}}}\n<\/pre><p>Let's break this down:<\/p><ul><li><code>:deps<\/code> is where we specify our dependencies, right now all we're depending on is Clojure 1.9.0.<\/li><li><code>:aliases<\/code> is where we specify special overrides that we can apply with the <code>-A<\/code> argument to the CLI.<\/li><li><code>:test<\/code> is the name of our alias, it adds the <code>test<\/code> directory to the paths list and <code>com.cognitect\/test-runner<\/code> to the dependencies.<\/li><li><code>:main-opts<\/code> instructs Clojure that we want these arguments applied when the alias is active. In this case, we're using <code>-m<\/code> to specify which namespace to execute.<\/li><\/ul><p>The usage of <code>deps.edn<\/code> is documented further in <a href=\"https:\/\/clojure.org\/guides\/deps_and_cli\">the deps guide<\/a>.<\/p><p>This will discover and run our test for us, let's run it now:<\/p><pre class=\"bash\">$ clj -Atest\nRunning tests in #{\"test\"}\nTesting hey.core-test\nRan 1 tests containing 1 assertions.\n0 failures, 0 errors.\n<\/pre><p>Hopefully you see the same success message as myself. You can see that we applied the values specified in our alias with the <code>-Atest<\/code> argument.<\/p><h2 id=\"building_executable_jars\">Building executable jars<\/h2><p>In this section we're going to build an \"uberjar\" containing your application alongside all of the dependencies it requires to run.<\/p><blockquote><p> You only need to bother with uberjars if you're writing an application you wish to deploy and run somewhere. If you're building a library for others to depend on you probably won't need this. <\/p><\/blockquote><p>Compiling your project into an uberjar will involve similar steps to getting your tests running, we're going to add another alias with another dependency which does the job for us.<\/p><p>Go ahead and add this new alias to the <code>:aliases<\/code> section of your <code>deps.edn<\/code> file, next to the <code>:test<\/code> alias:<\/p><pre class=\"clojure\">:uberjar\n{:extra-deps\n {pack\/pack.alpha\n  {:git\/url \"git@github.com:juxt\/pack.alpha.git\"\n   :sha     \"e6d0691c5f58135e1ef6fb1c9dda563611d36205\"}}\n :main-opts [\"-m\" \"mach.pack.alpha.capsule\" \"deps.edn\" \"dist\/hey.jar\"]}\n<\/pre><p>We can now build a jar that we can execute directly through the <code>java<\/code> program, without the Clojure CLI:<\/p><pre class=\"bash\">$ clj -Auberjar\n$ java -jar dist\/hey.jar # Drops us into a Clojure REPL.\n$ java -jar dist\/hey.jar -m hey.core # Executes our \"Hello, World!\".\n<\/pre><p>Please note that your code has not been AOT (ahead of time) compiled, it's still just plain Clojure that's compiled as and when it's required at run time. This can mean very slightly slower startup times when you're working with a large codebase.<\/p><p>If this becomes an issue for you you'll have to work out how to perform AOT compilation as you build your uberjar. By then, pack may even support it as a core feature.<\/p><h2 id=\"publishing_to_clojars\">Publishing to Clojars<\/h2><p>In this section we're going to publish a small jar file to <a href=\"https:\/\/clojars.org\/\">Clojars<\/a> containing only your source code, we'll be using maven to perform the deploy.<\/p><blockquote><p> This is intended for libraries that others will depend on and use, you won't need to worry about this section if you're building an application you'll be running. <\/p><\/blockquote><p>First, we're going to add your Clojars login to <code>~\/.m2\/settings.xml<\/code>:<\/p><pre class=\"xml\">&lt;settings&gt;\n  &lt;servers&gt;\n    &lt;server&gt;\n      &lt;id&gt;clojars&lt;\/id&gt;\n      &lt;username&gt;username&lt;\/username&gt;\n      &lt;password&gt;password&lt;\/password&gt;\n    &lt;\/server&gt;\n  &lt;\/servers&gt;\n&lt;\/settings&gt;\n<\/pre><p>Now we're going to generate your base <code>pom.xml<\/code> file, you should run this command whenever you're going to publish so the dependencies get updated:<\/p><pre class=\"bash\">$ clj -Spom\n<\/pre><p>Here's my example version, I've annotated each section. There's some you'll want to change as well as a couple of parts you'll want to add and update:<\/p><pre class=\"xml\">&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;\n&lt;project xmlns=\"http:\/\/maven.apache.org\/POM\/4.0.0\" xmlns:xsi=\"http:\/\/www.w3.org\/2001\/XMLSchema-instance\" xsi:schemaLocation=\"http:\/\/maven.apache.org\/POM\/4.0.0 http:\/\/maven.apache.org\/xsd\/maven-4.0.0.xsd\"&gt;\n  &lt;modelVersion&gt;4.0.0&lt;\/modelVersion&gt;\n  &lt;!-- Here by default, but will need updating --&gt;\n  &lt;groupId&gt;org.clojars.olical&lt;\/groupId&gt;\n  &lt;artifactId&gt;hey&lt;\/artifactId&gt;\n  &lt;version&gt;2.1.0-SNAPSHOT&lt;\/version&gt;\n  &lt;name&gt;hey&lt;\/name&gt;\n  &lt;!-- Here by default, updated by `clj -Spom` --&gt;\n  &lt;dependencies&gt;\n    &lt;dependency&gt;\n      &lt;groupId&gt;org.clojure&lt;\/groupId&gt;\n      &lt;artifactId&gt;clojure&lt;\/artifactId&gt;\n      &lt;version&gt;1.9.0&lt;\/version&gt;\n    &lt;\/dependency&gt;\n  &lt;\/dependencies&gt;\n  &lt;build&gt;\n    &lt;sourceDirectory&gt;src&lt;\/sourceDirectory&gt;\n    &lt;!-- Essential for bundling your source files into the JAR --&gt;\n    &lt;resources&gt;\n      &lt;resource&gt;\n        &lt;directory&gt;src&lt;\/directory&gt;\n      &lt;\/resource&gt;\n    &lt;\/resources&gt;\n  &lt;\/build&gt;\n  &lt;repositories&gt;\n    &lt;repository&gt;\n      &lt;id&gt;clojars&lt;\/id&gt;\n      &lt;url&gt;https:\/\/clojars.org\/repo&lt;\/url&gt;\n    &lt;\/repository&gt;\n  &lt;\/repositories&gt;\n  &lt;!-- Essential for pushing to Clojars --&gt;\n  &lt;distributionManagement&gt;\n    &lt;repository&gt;\n      &lt;id&gt;clojars&lt;\/id&gt;\n      &lt;name&gt;Clojars repository&lt;\/name&gt;\n      &lt;url&gt;https:\/\/clojars.org\/repo&lt;\/url&gt;\n    &lt;\/repository&gt;\n  &lt;\/distributionManagement&gt;\n  &lt;!-- Optional extras for Clojars --&gt;\n  &lt;description&gt;Just a Hello, World!&lt;\/description&gt;\n  &lt;url&gt;https:\/\/github.com\/Olical\/clojure-hey-example&lt;\/url&gt;\n  &lt;licenses&gt;\n    &lt;license&gt;\n      &lt;name&gt;Unlicense&lt;\/name&gt;\n      &lt;url&gt;https:\/\/unlicense.org\/&lt;\/url&gt;\n    &lt;\/license&gt;\n  &lt;\/licenses&gt;\n  &lt;scm&gt;\n    &lt;url&gt;https:\/\/github.com\/Olical\/clojure-hey-example&lt;\/url&gt;\n  &lt;\/scm&gt;\n&lt;\/project&gt;\n<\/pre><p>We can now tell maven to build and deploy a jar to Clojars:<\/p><pre class=\"bash\">$ mvn deploy\n<\/pre><p>A lot of this information comes from <a href=\"https:\/\/github.com\/clojars\/clojars-web\/wiki\/Pushing\">Clojar's guide to pushing<\/a> and <a href=\"https:\/\/maven.apache.org\/guides\/mini\/guide-3rd-party-jars-remote.html\">Maven's guide to deploying 3rd party jars<\/a>.<\/p><p>If everything went to plan, your Clojars account should now contain a fresh new jar. Note that this is <i>not<\/i> an uberjar, it only contains your source files and dependency information, not the actual dependencies themselves. The dependencies will be resolved by a tool such as the Clojure CLI.<\/p><h2 id=\"ergonomics\">Ergonomics<\/h2><p>As it stands, to deploy our jar to Clojars we'll want to take the following steps:<\/p><ul><li>Update the version number in our <code>pom.xml<\/code>.<\/li><li>Run the tests with <code>clj -Atest<\/code>.<\/li><li>Run <code>clj -Spom<\/code> to update our <code>pom.xml<\/code> with any dependency changes.<\/li><li>Run <code>mvn deploy<\/code>. <\/li><\/ul><p>This isn't particularly catchy, so we'll wrap everything we've seen so far in a pretty little <code>Makefile<\/code>:<\/p><pre class=\"makefile\">.PHONY: run test uberjar deploy\nrun:\n\tclj -m hey.core\ntest:\n\tclj -Atest\nuberjar:\n\tclj -Auberjar\ndeploy: test\n\tclj -Spom\n\tmvn deploy\n<\/pre><p>Now all you need to do when you wish to deploy is bump the version number in your <code>pom.xml<\/code> and execute <code>make deploy<\/code>.<\/p><h2 id=\"thanks!\">Thanks!<\/h2><p>I really hope this post has helped you out! You can find the example project I built during the writing of this post at <a href=\"https:\/\/github.com\/Olical\/clojure-hey-example\">github.com\/Olical\/clojure-hey-example<\/a> and the <a href=\"https:\/\/clojars.org\/org.clojars.olical\/hey\/versions\/2.1.0-SNAPSHOT\">Clojars page here<\/a>.<\/p><p>Happy Clojuring!<\/p>","created_at":"2019-04-28T04:43:17+0000","updated_at":"2019-04-28T04:43:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"oli.me.uk","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","clojurescript","datalog","datomic","datascript"],"is_public":false,"id":924,"title":"A shallow dive into DataScript internals","url":"https:\/\/tonsky.me\/blog\/datascript-internals\/","content":"<p>This is an overview of <a href=\"https:\/\/github.com\/tonsky\/datascript\">DataScript code base<\/a>. Without going into much detail, it paints the overall picture of how code is structured, what parts it\u2019s built of and what purpose they serve. If you\u2019re interested in studying DataScript sources, this is a great place to start.<\/p><p>For those who are using DataScript this post may help to get better understanding of machinery behind public APIs and make better use of them.<\/p><h2 id=\"datoms\">Datoms<\/h2><p>Minimal piece of information in DataScript is Datom. It\u2019s defined in <code class=\"highlighter-rouge\">datascript.core<\/code> as<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>(defrecord Datom [e a v tx added])\n<\/pre><\/div><\/div><p>where <code class=\"highlighter-rouge\">[e a v]<\/code> is entity, attribute and value, <code class=\"highlighter-rouge\">tx<\/code> is a transaction id (integer) and <code class=\"highlighter-rouge\">added<\/code> is a boolean flag to differentiate additions from retractions.<\/p><p>Datom has <code class=\"highlighter-rouge\">(-hash)<\/code> and <code class=\"highlighter-rouge\">(-eqiv)<\/code> methods overloaded so only <code>[e a v]<\/code> take part in comparisons. It helps to guarantee that each datom can be added to the database only once.<\/p><h2 id=\"db\">DB<\/h2><p>Database is just an immutable, persistent collection of datoms. It\u2019s very much like a built-in ClojureScript data structure, e.g. set. Main operations supported by DB are:<\/p><ul><li>Add a datom (similar to <code class=\"highlighter-rouge\">conj<\/code>)<\/li>\n  <li>Retract a datom (similar to <code class=\"highlighter-rouge\">disj<\/code>)<\/li>\n  <li>Search for range of datoms<\/li>\n<\/ul><p>Both addition and retraction are implemented in <code class=\"highlighter-rouge\">datascript.core\/with-datom<\/code>. Searching is implemented as a part of <code class=\"highlighter-rouge\">ISearch<\/code> protocol.<\/p><p>DataScript DB contains only currently relevant datoms. There\u2019s no history tracking at DB level. When datom is removed from a DB, there\u2019s no trace of it anywhere. Retracted means gone.<\/p><p>Internally DB contains three different indexes to help with various search patterns. Each index is a sorted set of datoms (it\u2019s literally a set). They\u2019re named after sort order: <code class=\"highlighter-rouge\">EAVT<\/code>, <code class=\"highlighter-rouge\">AEVT<\/code> and <code class=\"highlighter-rouge\">AVET<\/code>. Every index contains all datoms of current DB, meaning all three indexes are the same set sorted differently. There\u2019s no other datoms storage inside the DB besides indexes.<\/p><p>DB is defined in <code class=\"highlighter-rouge\">datascript.core<\/code> as<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>(defrecord DB [schema\n               eavt aevt avet\n               max-eid\n               max-tx\n               rschema])\n<\/pre><\/div><\/div><p>Besides indexes, there\u2019s only schema (just a regular map) and some internal bookkeeping: max seen entity id, latest transaction id and reverse value-to-key schema map.<\/p><h2 id=\"btset\">BTSet<\/h2><p>Each DataScript index is stored as <code class=\"highlighter-rouge\">datascript.btset<\/code>. It\u2019s an immutable, persistent implementation of B+ tree data structure. It\u2019s a proper ClojureScript collection, with all protocols of <code class=\"highlighter-rouge\">sorted-set<\/code> implemented.<\/p><p><code class=\"highlighter-rouge\">BTSet<\/code> was needed because DataScript does a lot of range scans over indexes. Due to usage of continuous js arrays as tree nodes, range scans over <code class=\"highlighter-rouge\">BTSet<\/code> are ~3 times faster than over built-in <code class=\"highlighter-rouge\">sorted-set<\/code> which is a Red-Black binary tree.<\/p><p><code class=\"highlighter-rouge\">BTSet<\/code> uses generative testing to validate its correctness: same operations are performed at <code class=\"highlighter-rouge\">BTSet<\/code> and <code class=\"highlighter-rouge\">sorted-set<\/code>, results are compared for equality (see <code class=\"highlighter-rouge\">stresstest-...<\/code> in <code class=\"highlighter-rouge\">datascript.test.btset<\/code>).<\/p><h2 id=\"adding-data-to-db\">Adding data to DB<\/h2><p>DataScript has a lot of conventions about how to format data before adding it to the DB. Basically, you can provide vector of form <code>[op e a v]<\/code> or a map of form <code>{:db\/id e, (a v)+}<\/code>. There\u2019s also a lot of nuances like resolving temporary ids, calling <code class=\"highlighter-rouge\">:db.fn\/*<\/code> shortcuts, referencing current transaction id, using nested maps for components, understanding different attributes arity and reversing reverse references.<\/p><p>Biggest part of <code class=\"highlighter-rouge\">datascript.core<\/code>, starting from <code class=\"highlighter-rouge\">(defrecord TxReport)<\/code>, is all about solving these problems. Main work happens in <code class=\"highlighter-rouge\">transact-tx-data<\/code>, which uses recursion to reduce complexity of transaction data step-by-step, until it\u2019s reduced to a set of plain datoms:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>entity map \u2192 op vector[s] \u2192 id resolution \u2192 datom[s]\n<\/pre><\/div><\/div><p><code class=\"highlighter-rouge\">transact-tx-data<\/code> loop also builds <code class=\"highlighter-rouge\">TxReport<\/code> record along the way:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>(defrecord TxReport [db-before\n                     db-after\n                     tx-data\n                     tempids\n                     tx-meta])\n<\/pre><\/div><\/div><p><code class=\"highlighter-rouge\">TxReport<\/code> is a log of everything that has happened during transaction. It stores temporary entity ids resolution table (<code class=\"highlighter-rouge\">tempids<\/code>) and raw datoms which were used to modify DB (<code class=\"highlighter-rouge\">tx-data<\/code>). Given <code class=\"highlighter-rouge\">TxReport<\/code> and <code class=\"highlighter-rouge\">db-before<\/code>, it\u2019s trivial to replay transaction and calculate <code class=\"highlighter-rouge\">db-after<\/code>:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>db-before + tx-data = db-after\n<\/pre><\/div><\/div><p><code class=\"highlighter-rouge\">tx-data<\/code> is an end result of transaction data simplification, with all temporary and implicit ids allocated to actual entity ids, all maps expanded into vector forms, and all vectors converted to <code class=\"highlighter-rouge\">Datom<\/code>s.<\/p><p><code class=\"highlighter-rouge\">TxReport<\/code> is also the only place where you can see datoms with <code class=\"highlighter-rouge\">added == false<\/code> for datoms which were retracted.<\/p><h2 id=\"entities\">Entities<\/h2><p>Entity is just a convenient interface to <code class=\"highlighter-rouge\">EAVT<\/code> index. Basically this:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>(:person\/name (d\/entity db 1))\n<\/pre><\/div><\/div><p>is translated to this:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>(first (d\/datoms db :eavt 1 :person\/name))\n<\/pre><\/div><\/div><p>with some per-entity caching. Most of the namespace <code class=\"highlighter-rouge\">datascript.impl.entity<\/code> is spent on implementing protocols to make entities look like normal ClojureScript maps.<\/p><h2 id=\"queries\">Queries<\/h2><p>Query engine implements Datalog over DBs and collections. It is the most complicated part of DataScript so far.<\/p><p>Query engine operates on relations. Relation is just a collection of tuples. E.g. this:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>:where [?e :name ?v]\n<\/pre><\/div><\/div><p>will be resolved by querying DB to following relation:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>R1 = {:symbols {?e 0 ?v 1}\n      :tuples  [#js [0 \u201cIvan\u201d], #js [5 \u201cOleg\u201d], ...]}\n<\/pre><\/div><\/div><p>During query execution, <code class=\"highlighter-rouge\">:where<\/code> is processed clause-by-clause. Relations which share no common variables are kept separate in a <code class=\"highlighter-rouge\">context<\/code>, building a collection of relations. If <code class=\"highlighter-rouge\">:where<\/code> looks like this:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>:where [?e :name ?v]\n       [?e2 :name ?v2]\n<\/pre><\/div><\/div><p>then DataScript will create two separate relations:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>R1 = {:symbols {?e 0 ?v 1}\n      :tuples  [#js [0 \u201cIvan\u201d], #js [5 \u201cOleg\u201d], ...]}\nR2 = {:symbols {?e2 0 ?v2 1}\n      :tuples  [#js [0 \u201cIvan\u201d], #js [5 \u201cOleg\u201d], ...]}\n<\/pre><\/div><\/div><p>Now what happens if we introduce third clause that implicitly joins first two relations by referring to the same variables? Let\u2019s take <code>[?e :friend ?e2]<\/code>. DataScript will first resolve that clause to its own relation:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>R3 = {:symbols {?e 0 ?e2 1}\n      :tuples  [#js [0 5], #js [5 0], ...]}\n<\/pre><\/div><\/div><p>Now context contains three relations, but they are not truly independent: they share common variables. This is forbidden, so at this point they will be reduced to a single relation. DataScript needs to hash-join latest relation (<code class=\"highlighter-rouge\">R3<\/code>) with any conflicting relations there are. In our case, <code class=\"highlighter-rouge\">R3<\/code> has intersection with both <code class=\"highlighter-rouge\">R1<\/code> (by <code>?e<\/code>) and <code class=\"highlighter-rouge\">R2<\/code> (by <code>?e2<\/code>), so what happens is we first do cartesian product of <code class=\"highlighter-rouge\">R1<\/code> and <code class=\"highlighter-rouge\">R2<\/code>, thus getting single relation to join with, then hash-join result with <code class=\"highlighter-rouge\">R3<\/code> by <code>(?e ?e2)<\/code> compound key:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>R4 = (R1 \u00d7 R2) `hash-join` R3\n<\/pre><\/div><\/div><p>Then <code class=\"highlighter-rouge\">R4<\/code> will replace <code class=\"highlighter-rouge\">R1<\/code>, <code class=\"highlighter-rouge\">R2<\/code> and <code class=\"highlighter-rouge\">R3<\/code> in a context.<\/p><p>Full algorithm:<\/p><ol><li>Parse query (<code class=\"highlighter-rouge\">datascript.parser<\/code>) and cache parse result (planned for 0.10.0). Most of the query validation happens at parsing stage.<\/li>\n  <li>Scan sources (<code class=\"highlighter-rouge\">:in<\/code> clause) and transform them to relations. This will create initial set of relations which populate context. Constants are also resolved to relations, e.g. <code class=\"highlighter-rouge\">:in ?a<\/code> \u2192 <code class=\"highlighter-rouge\">{:symbols {?a 0}, :tuples [#js [val]]}<\/code><\/li>\n  <li>For each clause in <code class=\"highlighter-rouge\">:where<\/code>, do this:\n    <ol><li>Try to resolve as much variables as possible using context so far and substitute them to clause (planned for 0.10.0)<\/li>\n      <li>Execute clause and get result as a relation. If clause is a rule call, there\u2019s special inner loop for rule resolution here, but result is still a relation.<\/li>\n      <li>If resulting relation has intersection by variables with any existing relation, hash-join them together, fusing into one big relation<\/li>\n      <li>Put resulting relation into the context<\/li>\n      <li>Move onto the next clause<\/li>\n    <\/ol><\/li>\n  <li>Build result set based on <code class=\"highlighter-rouge\">:find<\/code> and <code class=\"highlighter-rouge\">:with<\/code> vars list: do cartesian product on all relevant relations, leave just vars that matter from them, collect them into a <code class=\"highlighter-rouge\">set<\/code><\/li>\n  <li>If there\u2019s some aggregation happening, do <code class=\"highlighter-rouge\">group-by<\/code> and run aggregation functions<\/li>\n  <li>If <code class=\"highlighter-rouge\">pull()<\/code> is used, call Pull API for each entity<\/li>\n  <li>If find specifications are used, do post-processing: unwrap inner tuples, take first element from a set, etc.<\/li>\n<\/ol><p>Query engine is implemented in <code class=\"highlighter-rouge\">datascript.query<\/code> and uses parser from <code class=\"highlighter-rouge\">datascript.parser<\/code>.<\/p><h2 id=\"pull-api\">Pull API<\/h2><p>Pull API is implemented in <code class=\"highlighter-rouge\">datascript.pull-api<\/code> by <a href=\"https:\/\/github.com\/dthume\">David Thomas Hume<\/a>. It walks all entity ids, for each entity it walks all attributes from a pull pattern, and for each recursive attribute calls itself. It\u2019s a recursive algorithm implemented as a state machine that emulates stack so it can handle unlimited recursion depth.<\/p><p>There\u2019s also <code class=\"highlighter-rouge\">datascript.pull-parser<\/code> where pull pattern is parsed and validated.<\/p><p>During implementation of DataScript pull pattern parser and query parser, we\u2019ve helped Datomic team to fix couple of inaccuracies in their query grammar. This is why having alternative implementations is usually a good thing.<\/p><h2 id=\"filtered-dbs\">Filtered DBs<\/h2><p>Given database <code class=\"highlighter-rouge\">D<\/code> and predicate <code class=\"highlighter-rouge\">P: datom, DB \u2192 boolean<\/code>, you can construct special view <code class=\"highlighter-rouge\">F = (datascript\/filter D P)<\/code> that will work like a regular database but for all needs (query, entity, pull, index access) will look like as if it only contains datoms that satisfy <code class=\"highlighter-rouge\">P<\/code>.<\/p><p>Filtered databases are implemented in <code class=\"highlighter-rouge\">datascript.core<\/code> by just extending couple of protocols: <code class=\"highlighter-rouge\">IDB<\/code>, <code class=\"highlighter-rouge\">ISearch<\/code>, <code class=\"highlighter-rouge\">IIndexAccess<\/code> \u2014 mostly by adding additional <code class=\"highlighter-rouge\">filter<\/code> post-process step.<\/p><p>For advanced uses it means that it\u2019s quite easy to pretend to be a database: just implement these protocols and rest of DataScript code will work as expected.<\/p><h2 id=\"database-mutation\">Database mutation<\/h2><p>Everything we\u2019ve discussed so far works on top of immutable, persistent DB objects. There\u2019s also a small piece of code to handle database mutations (as of 0.9.0, it\u2019s literally 25 lines).<\/p><p><code class=\"highlighter-rouge\">datascript\/create-conn<\/code> returns a new database object wrapped with an atom. There\u2019s also a <code class=\"highlighter-rouge\">listen!<\/code> machinery for listening for DB changes. It was added because regular atom watchers receive value-before and value-after, but database changes have more useful information: list of datoms added and temporary ids resolution table.<\/p><p>DataScript \u201cconnection\u201d is not a connection in any sense. I borrowed the term from Datomic, but now I\u2019m thinking about changing it to <code class=\"highlighter-rouge\">db-ref<\/code> or better, if only I could think of anything.<\/p><h2 id=\"javascript-interop\">JavaScript interop<\/h2><p>Almost from the beginning DataScript has supported JavaScript interface. You include pre-compiled JS file (270k raw\/~60k gzipped as of 0.9.0) and call <code class=\"highlighter-rouge\">datascript.empty_db()<\/code>, <code class=\"highlighter-rouge\">datascript.query()<\/code> and stuff.<\/p><p>This was made possible by <code class=\"highlighter-rouge\">datascript.js<\/code> facade: it converts data coming from JS into ClojureScript data structures, calls proper <code class=\"highlighter-rouge\">datascript.*<\/code> APIs, converts results into JS form and returns them back to JS caller. Not the most effective way to do things, but the only one manageable without writing js-native DataScript\u2019s evil twin. Even with such approach, some things have leaked into the main DataScript code: it has to take into account that attributes can be strings, for example. But for the most part JS interop is isolated.<\/p><p>And JS interop is feature-full: everything you can do from CLJS, you can do from JS. Maybe not as well battle-tested, but nonetheless. See <code class=\"highlighter-rouge\">test\/js\/tests.html<\/code> for how DataScript can be used from JS.<\/p><h2 id=\"tests\">Tests<\/h2><p>DataScript does acceptance testing almost exclusively. Everything in <code class=\"highlighter-rouge\">datascript.test.*<\/code> works with top-level interfaces of <code class=\"highlighter-rouge\">datascript<\/code> and <code class=\"highlighter-rouge\">datascript.core<\/code> namespaces. Exceptions are <code class=\"highlighter-rouge\">datascript.test.btset<\/code> (generative tests of BTSet implementation) and <code class=\"highlighter-rouge\">datascript.test.perf<\/code> (somewhat ad-hoc and messy performance tests of various parts of DataScript).<\/p><p>Having just acceptance tests works because semantics and APIs are already defined and documented by Datomic team whom DataScript is following. APIs do not change often, so extensive tests suite (2.8k loc <code class=\"highlighter-rouge\">test\/<\/code> for 3k loc <code class=\"highlighter-rouge\">src\/<\/code>) is not a dead weight and requires little maintenance.<\/p><p>Tests allowed me to change underlying implementation without rewriting a single test: I completely changed index implementation and inner DB structure once already and replaced query engine twice so far. It gives tremendous confidence to pull such refactorings without breaking anyone else\u2019s code.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>I didn\u2019t know what to write in conclusion although I felt every post needs one. So here it is: have fun, enjoy your day.<\/p>","created_at":"2019-04-27T23:06:42+0000","updated_at":"2019-04-27T23:06:57+0000","published_at":"2015-02-23T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"tonsky.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":923,"title":"PostgreSQL: Documentation: 10: Chapter\u00a011.\u00a0Indexes","url":"https:\/\/www.postgresql.org\/docs\/10\/indexes.html","content":"<div class=\"container-fluid margin\"><div class=\"row\"><div id=\"pgContentWrap\" class=\"col-11\"><div class=\"row\"><div class=\"col-md-6 mb-2\"><div class=\"row\"><div class=\"col\"><p><a href=\"https:\/\/www.postgresql.org\/docs\/\" title=\"Documentation\">Documentation<\/a> \u2192 <a href=\"https:\/\/www.postgresql.org\/docs\/10\/index.html\">PostgreSQL 10<\/a><\/p><\/div><\/div><div class=\"row\"><p>Supported Versions:\n                      \n                        \n                        \n                          <a href=\"https:\/\/www.postgresql.org\/docs\/current\/indexes.html\" title=\"PostgreSQL 11 - Chapter\u00a011.\u00a0Indexes\">Current<\/a>\n                          (<a href=\"https:\/\/www.postgresql.org\/docs\/11\/indexes.html\" title=\"PostgreSQL 11 - Chapter\u00a011.\u00a0Indexes\">11<\/a>)\n                        \n                      \n                         \/ \n                        \n                          <a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes.html\" title=\"PostgreSQL 10 - Chapter\u00a011.\u00a0Indexes\" class=\"docs-version-selected\">10<\/a>\n                        \n                      \n                         \/ \n                        \n                          <a href=\"https:\/\/www.postgresql.org\/docs\/9.6\/indexes.html\" title=\"PostgreSQL 9.6 - Chapter\u00a011.\u00a0Indexes\">9.6<\/a>\n                        \n                      \n                         \/ \n                        \n                          <a href=\"https:\/\/www.postgresql.org\/docs\/9.5\/indexes.html\" title=\"PostgreSQL 9.5 - Chapter\u00a011.\u00a0Indexes\">9.5<\/a>\n                        \n                      \n                         \/ \n                        \n                          <a href=\"https:\/\/www.postgresql.org\/docs\/9.4\/indexes.html\" title=\"PostgreSQL 9.4 - Chapter\u00a011.\u00a0Indexes\">9.4<\/a><\/p><\/div><div class=\"row\"><p>Development Versions:\n                    \n                      \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/devel\/indexes.html\" title=\"PostgreSQL devel - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">devel<\/a><\/p><\/div><div class=\"row\"><p>Unsupported versions:\n                    \n                      \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/9.3\/indexes.html\" title=\"PostgreSQL 9.3 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">9.3<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/9.2\/indexes.html\" title=\"PostgreSQL 9.2 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">9.2<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/9.1\/indexes.html\" title=\"PostgreSQL 9.1 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">9.1<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/9.0\/indexes.html\" title=\"PostgreSQL 9.0 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">9.0<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/8.4\/indexes.html\" title=\"PostgreSQL 8.4 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">8.4<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/8.3\/indexes.html\" title=\"PostgreSQL 8.3 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">8.3<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/8.2\/indexes.html\" title=\"PostgreSQL 8.2 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">8.2<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/8.1\/indexes.html\" title=\"PostgreSQL 8.1 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">8.1<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/8.0\/indexes.html\" title=\"PostgreSQL 8.0 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">8.0<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/7.4\/indexes.html\" title=\"PostgreSQL 7.4 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">7.4<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/7.3\/indexes.html\" title=\"PostgreSQL 7.3 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">7.3<\/a>\n                    \n                       \/ \n                      <a href=\"https:\/\/www.postgresql.org\/docs\/7.2\/indexes.html\" title=\"PostgreSQL 7.2 - Chapter\u00a011.\u00a0Indexes\" rel=\"nofollow\">7.2<\/a><\/p><\/div><\/div><\/div><div id=\"docContent\"><p><\/p><table summary=\"Navigation header\" style=\"width: 100%;\"><tr><th colspan=\"5\" align=\"center\">Chapter\u00a011.\u00a0Indexes<\/th>\n<\/tr><tr><td align=\"left\" style=\"width: 10%;\"><a accesskey=\"p\" href=\"https:\/\/www.postgresql.org\/docs\/10\/typeconv-select.html\" title=\"10.6.\u00a0SELECT Output Columns\">Prev<\/a>\u00a0<\/td>\n<td align=\"left\" style=\"width: 10%;\"><a accesskey=\"u\" href=\"https:\/\/www.postgresql.org\/docs\/10\/sql.html\" title=\"Part\u00a0II.\u00a0The SQL Language\">Up<\/a><\/td>\n<th align=\"center\" style=\"width: 60%;\">Part\u00a0II.\u00a0The SQL Language<\/th>\n<td align=\"right\" style=\"width: 10%;\"><a accesskey=\"h\" href=\"https:\/\/www.postgresql.org\/docs\/10\/index.html\" title=\"PostgreSQL 10.7 Documentation\">Home<\/a><\/td>\n<td align=\"right\" style=\"width: 10%;\">\u00a0<a accesskey=\"n\" href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-intro.html\" title=\"11.1.\u00a0Introduction\">Next<\/a><\/td>\n<\/tr><\/table><hr \/><div class=\"chapter\" id=\"INDEXES\"><div class=\"toc\"><p><strong>Table of Contents<\/strong><\/p><dl class=\"toc\"><dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-intro.html\">11.1. Introduction<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-types.html\">11.2. Index Types<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-multicolumn.html\">11.3. Multicolumn Indexes<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-ordering.html\">11.4. Indexes and <code class=\"literal\">ORDER BY<\/code><\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-bitmap-scans.html\">11.5. Combining Multiple Indexes<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-unique.html\">11.6. Unique Indexes<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-expressional.html\">11.7. Indexes on Expressions<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-partial.html\">11.8. Partial Indexes<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-opclass.html\">11.9. Operator Classes and Operator Families<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-collations.html\">11.10. Indexes and Collations<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-index-only-scans.html\">11.11. Index-Only Scans<\/a><\/dt>\n<dt><a href=\"https:\/\/www.postgresql.org\/docs\/10\/indexes-examine.html\">11.12. Examining Index Usage<\/a><\/dt>\n<\/dl><\/div><p>Indexes are a common way to enhance database performance. An index allows the database server to find and retrieve specific rows much faster than it could do without an index. But indexes also add overhead to the database system as a whole, so they should be used sensibly.<\/p><\/div><\/div><\/div><\/div><\/div>","created_at":"2019-04-27T20:18:31+0000","updated_at":"2019-04-27T20:18:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.postgresql.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":922,"title":"PostgreSQL Features You May Not Have Tried But Should","url":"https:\/\/pgdash.io\/blog\/postgres-features.html","content":"<p>PostgreSQL contains a slew of great features. Many of them are very well known.\nOthers can be incredibly useful but are not as widely appreciated. Here are\nour top picks for great PostgreSQL features you may not have taken a look at\nbut really should, since they can help you get code into production faster,\nmake ops easier and generally get things done with less code and sweat.<\/p><h3 id=\"pubsub-notifications\">Pub\/Sub Notifications<\/h3><p>PostgreSQL comes with a simple non-durable topic-based publish-subscribe\nnotification system. It\u2019s no Kafka, but the features do support common use cases.<\/p><p>Messages on a specific topic can be broadcast to all connected subscribers who\nare listening for that topic. The messages are <em>pushed<\/em> by the Postgres server\nto the listening clients. Polling is not required, but your database driver\nshould support delivery of notifications to the application asynchronously.<\/p><p>The notification consists of a topic name and a payload (upto about 8000\ncharacters). The payload would typically be a JSON string, but of course it can\nbe anything. You can send a notification using the <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/sql-notify.html\">NOTIFY<\/a>\ncommand:<\/p><figure class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">NOTIFY 'foo_events', '{\"userid\":42,\"action\":\"grok\"}'<\/pre><\/figure><p>or the <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/sql-notify.html\">pg_notify()<\/a>\nfunction:<\/p><figure class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">SELECT pg_notify('foo_events', '{\"userid\":42,\"action\":\"grok\"}');<\/pre><\/figure><p>The subscription happens with the <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/sql-listen.html\">LISTEN<\/a>\ncommand, but typically you\u2019ve to use driver-specific APIs. Here\u2019s the <a href=\"https:\/\/godoc.org\/github.com\/lib\/pq\/example\/listen\">Go version<\/a>\nfor example.<\/p><h3 id=\"table-inheritance\">Table Inheritance<\/h3><p>Assume you have a table called \u201cinvoices\u201d. You\u2019ve to now support \u201cgovernment\ninvoices\u201d, which are invoices but have a few additional fields. How\nwould you model this? Rather than adding nullable columns to invoices, or\na single nullable JSON column to invoices, try the inheritance feature:<\/p><figure class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">CREATE TABLE invoices (\n    invoice_number   int  NOT NULL PRIMARY KEY,\n    issued_on        date NOT NULL DEFAULT now()\n);\nCREATE TABLE government_invoices (\n    department_id    text NOT NULL\n) INHERITS (invoices);<\/pre><\/figure><p>This reflects the situation that all government invoices are invoices, but have\nan extra attribute. The \u201cgovernment_invoices\u201d table above has a total of 3 columns:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# \\d invoices\n                  Table \"public.invoices\"\n     Column     |  Type   | Collation | Nullable | Default\n----------------+---------+-----------+----------+---------\n invoice_number | integer |           | not null |\n issued_on      | date    |           | not null | now()\nIndexes:\n    \"invoices_pkey\" PRIMARY KEY, btree (invoice_number)\nNumber of child tables: 1 (Use \\d+ to list them.)\ntest=# \\d government_invoices\n            Table \"public.government_invoices\"\n     Column     |  Type   | Collation | Nullable | Default\n----------------+---------+-----------+----------+---------\n invoice_number | integer |           | not null |\n issued_on      | date    |           | not null | now()\n department_id  | text    |           | not null |\nInherits: invoices<\/pre><\/figure><p>Adding rows work as though the tables are independent:<\/p><figure class=\"highlight\"><pre class=\"language-sql\" data-lang=\"sql\">INSERT INTO invoices (invoice_number) VALUES (100);\nINSERT INTO government_invoices\n    (invoice_number, department_id) VALUES (101, 'DOD');<\/pre><\/figure><p>But watch happens when you SELECT:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# SELECT * FROM government_invoices;\n invoice_number | issued_on  | department_id\n----------------+------------+---------------\n            101 | 2018-06-19 | DOD\n(1 row)\ntest=# SELECT * FROM invoices;\n invoice_number | issued_on\n----------------+------------\n            100 | 2018-06-19\n            101 | 2018-06-19\n(2 rows)<\/pre><\/figure><p>Invoice number 101 from the child table is also listed in the parent table. This\nfacilitates algorithms that know only about invoices to work on the parent\ninvoices table while ignoring any child tables.<\/p><p>You can learn more about table inheritance in PostgreSQL <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/ddl-inherit.html\">in the docs<\/a>.<\/p><h3 id=\"foreign-data-wrappers\">Foreign Data Wrappers<\/h3><p>Did you know you can have virtual tables that actually serve data from another\nPostgreSQL instance? Or even SQLite files, MongoDB, Redis, and more? This feature\nis called <em>Foreign Data Wrappers<\/em>, which provides a standardized way to access\nand manipulate data stored externally to the Postgres server you\u2019re connecting to.\nThere are various FDW implementations available that let you connect to various\ndifferent data sources. These are typically packaged as extensions.<\/p><p>The standard Postgres distribution comes with the\n<a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/postgres-fdw.html\">postgres_fdw<\/a>\nextension, which lets you connect to other Postgres servers.\nFor example, you can move a big table to another server, and setup a virtual\ntable (the proper term is a \u201cforeign table\u201d) locally:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">-- install the extension (required only once)\nCREATE EXTENSION postgres_fdw;\n-- big_server is our big, remote server with the migrated table\nCREATE SERVER big_server FOREIGN DATA WRAPPER postgres_fdw\n  OPTIONS (host '10.1.2.3', dbname 'big_db');\n-- create a user mapping for the app user\nCREATE USER MAPPING FOR app_user SERVER big_server\n  OPTIONS (user 'remote_user', password 'remote_pass');\n-- and finally create the virtual table invoices -&gt; big_db.public.invoices\nCREATE FOREIGN TABLE invoices (\n  invoice_num int NOT NULL PRIMARY KEY\n  -- other columns omitted for brevity\n) SERVER big_server;<\/pre><\/figure><p>This <a href=\"https:\/\/wiki.postgresql.org\/wiki\/Foreign_data_wrappers\">Wiki page<\/a> has\na good list of the many FDW implementations available.<\/p><p>Apart from accessing data from other servers, FDW has also been used to implement\nalternate storage layouts, like <a href=\"https:\/\/github.com\/citusdata\/cstore_fdw\">cstore_fdw<\/a>.<\/p><p>There is also <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/dblink.html\">dblink<\/a>,\nwhich is yet another method of accessing remote PostgreSQL (only) data.<\/p><h3 id=\"partitioned-tables\">Partitioned Tables<\/h3><p>Starting with version 10, PostgreSQL natively supports sharding of a table into\nmultiple child tables, based on a partition key value computed from one or more\ncolumn values. This allows a single, large table to be physically stored as\nseparate tables, for better DML performance and storage management.<\/p><p>Here\u2019s how to create a partitioned table, with the intention of adding a new\ntable each month:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">-- the parent table\nCREATE TABLE invoices (\n    invoice_number   int  NOT NULL,\n    issued_on        date NOT NULL DEFAULT now()\n) PARTITION BY RANGE (issued_on);\n  \n-- table for the month of May 2018\nCREATE TABLE invoices_2018_05 PARTITION OF invoices\n  FOR VALUES FROM ('2018-05-01') TO ('2018-06-01');\n-- table for the month of June 2018\nCREATE TABLE invoices_2018_06 PARTITION OF invoices\n  FOR VALUES FROM ('2018-06-01') TO ('2018-07-01');<\/pre><\/figure><p>The child tables have to be manually or programmatically created, it does not\nhappen automatically.<\/p><p>You can query and insert at the parent level, and PostgreSQL will route it to\nthe appropriate child table, have a look:<\/p><p>First, we insert two rows:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# INSERT INTO invoices VALUES (10042, '2018-05-15');\nINSERT 0 1\ntest=# INSERT INTO invoices VALUES (43029, '2018-06-15');\nINSERT 0 1<\/pre><\/figure><p>We can see that the rows are actually inserted into the child tables:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# SELECT * FROM invoices_2018_05;\n invoice_number | issued_on\n----------------+------------\n          10042 | 2018-05-15\n(1 row)\ntest=# SELECT * FROM invoices_2018_06;\n invoice_number | issued_on\n----------------+------------\n          43029 | 2018-06-15\n(1 row)<\/pre><\/figure><p>But queries can be done on the parent, returning combined result:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# SELECT * FROM invoices;\n invoice_number | issued_on\n----------------+------------\n          10042 | 2018-05-15\n          43029 | 2018-06-15\n(2 rows)<\/pre><\/figure><p>Partitioning is in some ways similar to inheritance (query at parent level), but\nthere are differences too (partition parent contains no data, for example).\nYou can read more about partitioning <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/ddl-partitioning.html\">in the docs<\/a>.<\/p><p>PostgreSQL 11, currently in beta, improves upon this feature. <a href=\"https:\/\/pgdash.io\/blog\/partition-postgres-11.html\">This article<\/a>\nhas an overview of the improvements.<\/p><h3 id=\"range-types\">Range Types<\/h3><p>Ever worked with temperature ranges, calendar scheduling, price ranges and the\nlike? Working with intervals are one of those deceptively simple things that\ngently lead you into hair-pulling and late night debugging.<\/p><p>Here\u2019s a table with a range column, and some values:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">CREATE TABLE prices (\n    item  text,\n    price int4range -- int4range is a range of regular integers\n);\nINSERT INTO prices VALUES ('mouse',    '[10,16)');\nINSERT INTO prices VALUES ('keyboard', '[20,31)');\nINSERT INTO prices VALUES ('joystick', '[35,56)');<\/pre><\/figure><p>The numbers in the mismatched brackets represent a half-open interval.\nHere is the query to find all items that are in the price range $15 to $30,\nusing the <code class=\"highlighter-rouge\">&amp;&amp;<\/code> operator (range overlap):<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# SELECT * FROM prices WHERE price &amp;&amp; int4range(15,30);\n   item   |  price\n----------+---------\n mouse    | [10,16)\n keyboard | [20,31)\n(2 rows)<\/pre><\/figure><p>If you are not impressed, try writing the query without ranges. (No really,\njust try it.)<\/p><p>Range types are quite powerful \u2013 there are <a href=\"https:\/\/www.postgresql.org\/docs\/10\/static\/functions-range.html#RANGE-OPERATORS-TABLE\">more operators<\/a>,\n<a href=\"https:\/\/www.postgresql.org\/docs\/10\/static\/functions-range.html#RANGE-FUNCTIONS-TABLE\">functions<\/a>,\nyou can define your own range types, and even index them.<\/p><p>To learn more about ranges, try this <a href=\"https:\/\/wiki.postgresql.org\/images\/7\/73\/Range-types-pgopen-2012.pdf\">presentation<\/a>,\nand <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/rangetypes.html\">the docs<\/a>.<\/p><h3 id=\"array-types\">Array Types<\/h3><p>PostgreSQL has supported arrays for a long time. Array types can reduce the\namount of boilerplate application code and simplify queries. Here is a table\nthat uses an array column:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">CREATE TABLE posts (\n    title text NOT NULL PRIMARY KEY,\n    tags  text[]\n);<\/pre><\/figure><p>Assuming each row represents a blog post, each having a set of tags, here is\nhow we can list all the posts that have both \u201cpostgres\u201d and \u201cgo\u201d tags:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# SELECT title, tags FROM posts WHERE '{\"postgres\", \"go\"}' &lt;@ tags;\n               title               |          tags\n-----------------------------------+------------------------\n Writing PostgreSQL Triggers in Go | {postgres,triggers,go}\n(1 row)<\/pre><\/figure><p>The usage of the array type here makes for concise data modelling and simpler\nqueries. Postgres arrays come with <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/functions-array.html\">operators and functions<\/a>,\nincluding <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/functions-aggregate.html\">aggregate functions<\/a>.\nYou can also create indexes on array expressions. Here is\n<a href=\"https:\/\/www.opsdash.com\/blog\/postgres-arrays-golang.html\">an article<\/a> on using\narrays with <a href=\"https:\/\/golang.org\">Go<\/a>.<\/p><h3 id=\"triggers\">Triggers<\/h3><p>You can ask PostgreSQL to execute a specific function when rows are inserted,\nupdated or deleted from a table. The function can even alter the values during\nan insert. You can read more about triggers <a href=\"https:\/\/www.postgresql.org\/docs\/10\/static\/plpgsql-trigger.html\">here<\/a>.\nJust to whet your appetite, here is a trigger that sends out a notification\nand writes to an audit log when a user is created:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">-- a table of users\nCREATE TABLE users (\n  username text NOT NULL PRIMARY KEY\n);\n-- an audit log\nCREATE TABLE audit_log (\n  at          timestamptz NOT NULL DEFAULT now(),\n  description text NOT NULL\n);\n-- the actual function that is executed per insert\nCREATE FUNCTION on_user_added() RETURNS TRIGGER AS $$\nBEGIN\n  IF (TG_OP = 'INSERT') THEN\n    -- add an entry into the audit log\n    INSERT INTO audit_log (description)\n        VALUES ('new user created, username is ' || NEW.username);\n    -- send a notification\n    PERFORM pg_notify('usercreated', NEW.username);\n  END IF;\n  RETURN NULL;\nEND;\n$$ LANGUAGE plpgsql;\n-- set the function as an insert trigger\nCREATE TRIGGER on_user_added AFTER INSERT ON users\n  FOR EACH ROW EXECUTE PROCEDURE on_user_added();<\/pre><\/figure><p>Now if we try to add a new user, an audit log entry is automatically added:<\/p><figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">test=# INSERT INTO users VALUES ('alice');\nINSERT 0 1\ntest=# SELECT * FROM audit_log;\n                at             |             description\n-------------------------------+-------------------------------------\n 2018-06-19 04:00:30.672947+00 | new user created, username is alice\n(1 row)<\/pre><\/figure><h3 id=\"pg_stat_statements\">pg_stat_statements<\/h3><p><a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/pgstatstatements.html\">pg_stat_statements<\/a>\nis an extension that is present by default in your PostgreSQL distribution,\nbut is not enabled. This extension records a wealth of information about\neach statement executed, including the time taken, the memory used and disk I\/Os\ninitiated. It\u2019s indispensible for understanding and debugging query performance\nissues.<\/p><p>The overhead of installing and enabling this extension is small, it\u2019s very\nsimple to use, and there is no reason NOT to run it on your production servers!\nRead <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/pgstatstatements.html\">the docs<\/a>\nfor more info.<\/p><h3 id=\"hash-gin-and-brin-indexes\">Hash, GIN and BRIN Indexes<\/h3><p>The default index type in PostgreSQL is the B-Tree, but there are also\nother types, that are documented <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/indexes-types.html\">here<\/a>.\nOther index types are very helpful in cases that aren\u2019t actually uncommon.\nIn particular, setting up indexes of the hash, GIN and BRIN type might just\nbe the solution to your performance issues:<\/p><ul><li><em>Hash<\/em>: Unlike B-Tree indexes which have inherent ordering, hash indexes are\nunordered and can only do equality matches (lookup). However, hash indexes\noccupy much lesser space and are faster than B-Trees for equality matches.\n(Also, note that prior to PostgreSQL 10 it was not possible to replicate\nhash indexes; they were unlogged.)<\/li>\n  <li><em>GIN<\/em>: GIN is an inverted index, which essentially allows for multiple values\nfor a single key. GIN indexes are useful for indexing arrays, JSON, ranges,\nfull text search etc.<\/li>\n  <li><em>BRIN<\/em>: If your data has a specific, natural order \u2013 for example, time series\ndata \u2013 and your queries typically work only with a small range of it, then\nBRIN indexes can speed up your queries with very little overhead. BRIN indexes\nmaintain ranges per block of data, allowing the optimizer to skip over blocks\nthat contain rows that won\u2019t be selected by the query.<\/li>\n<\/ul><p>Start reading about PostgreSQL index types <a href=\"https:\/\/www.postgresql.org\/docs\/current\/static\/indexes-types.html\">here<\/a>.<\/p><p>PostgreSQL can do full text search very well, including in languages other\nthan English. Here\u2019s an\n<a href=\"https:\/\/www.opsdash.com\/blog\/postgres-full-text-search-golang.html\">article<\/a> that\nwalks you through the steps of creating a full text query search app over\nPostgres in Go.<\/p><p><a href=\"https:\/\/pgdash.io\">pgDash<\/a> is an in-depth monitoring solution designed\nspecifically for PostgreSQL deployments. pgDash shows you information and\nmetrics about every aspect of your PostgreSQL database server, collected using\nthe open-source tool <a href=\"https:\/\/pgmetrics.io\">pgmetrics<\/a>.<\/p><p><a href=\"https:\/\/pgdash.io\/blog\/images\/pgdash-repl-primary.png\"><img src=\"https:\/\/pgdash.io\/blog\/images\/pgdash-repl-primary.png\" alt=\"Monitoring with pgDash\" \/><\/a><\/p><p>pgDash provides core reporting and visualization\nfunctionality, including collecting and displaying PostgreSQL information and\nproviding time-series graphs, detailed reports, alerting and more.<\/p>","created_at":"2019-04-27T20:14:20+0000","updated_at":"2019-04-27T20:14:20+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":10,"domain_name":"pgdash.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":921,"title":"understanding-letf-and-how-it-replaces-flet.html","url":"https:\/\/endlessparentheses.com\/understanding-letf-and-how-it-replaces-flet.html","created_at":"2019-04-24T12:43:17+0000","updated_at":"2019-04-24T12:43:17+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"endlessparentheses.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":920,"title":"Foundations of Databases | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=19726520","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D19726520\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"19726520\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/webdam.inria.fr\/Alice\/\" class=\"storylink\">Foundations of Databases<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=inria.fr\">inria.fr<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        421 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=tosh\" class=\"hnuser\">tosh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=19726520\">12 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=19726520&amp;goto=item%3Fid%3D19726520\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Foundations%20of%20Databases&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Foundations%20of%20Databases\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=19726520&amp;auth=be66a70acdaf7de2e4efbff6fd4aa0aab721c9b8\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=19726520\">20\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-04-24T11:00:46+0000","updated_at":"2019-04-24T11:00:46+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":919,"title":"Graphiti","url":"https:\/\/www.graphiti.dev\/guides\/why","content":"<p style=\"text-align: center;\">\n  <img width=\"100%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/why.png\" alt=\"image\" \/><\/p><p>I\u2019ve written GraphQL and had a pleasant experience. I have enourmous respect for GraphQL. I also believe there is a fundamental flaw in GraphQL\u2019s design.<\/p><p>Let\u2019s remember <a href=\"https:\/\/dev.to\/smizell\/why-people-like-graphql-221c\">why people like GraphQL<\/a> in the first place: because it addressed common frustrations with REST APIs:<\/p><blockquote>\n  <p><em>[We] heard from integrators that our REST API also wasn\u2019t very flexible. It sometimes required two or three separate calls to assemble a complete view of a resource. It seemed like our responses simultaneously sent too much data and didn\u2019t include data that consumers needed.<\/em><\/p>\n  <p>- <a href=\"https:\/\/githubengineering.com\/the-github-graphql-api\">\u201cThe Github Graph API\u201d<\/a>, <strong>GitHub Engineering<\/strong><\/p>\n<\/blockquote><p>GraphQL solves real problems. It\u2019s flaw is that it solved these\nproblems using zero-sum thinking: we must\nabandon the existing paradigm and forge a new one. It\u2019s GraphQL\nversus REST, one or the other. <a href=\"https:\/\/medium.freecodecamp.org\/rest-apis-are-rest-in-peace-apis-long-live-graphql-d412e559d8e4\">REST is dead, long live GraphQL<\/a>.<\/p><p>Graphiti instead approaches the problem using <a href=\"http:\/\/aturon.github.io\/2018\/06\/02\/listening-part-2\">positive-sum thinking<\/a>:<\/p><blockquote>\n  <p><em>Positive-sum thinking is how we embrace pluralism while retaining a coherent vision and set of values\u2026A zero-sum view would assume that apparent oppositions are fundamental, e.g., that appealing to the JS crowd inherently hurts the C++ one. A positive-sum view starts by seeing different perspectives and priorities as legitimate and worthwhile, with a faith that<\/em> <strong>by respecting each other in this way, we can find strictly better solutions than had we optimized solely for one perspective.<\/strong><\/p>\n  <p>- <a href=\"http:\/\/aturon.github.io\/2018\/06\/02\/listening-part-2\">\u201cListening and Trust\u201d<\/a>, <strong>Aaron Turon<\/strong><\/p>\n<\/blockquote><p>GraphQL optimized around REST\u2019s shortcomings, and in doing so it dropped\nREST\u2019s advantages. There is no need for such a zero-sum tradeoff. We can take\neverything great about GraphQL and build it <strong><em>on top of<\/em><\/strong> REST (and\nHTTP!), instead of replacing it altogether. We can have our cake and eat it too.<\/p><p style=\"text-align: center;\">\n  <img width=\"100%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/rest-graffiti.jpg\" alt=\"image\" \/><\/p><p>You\u2019ll find plenty of <a href=\"https:\/\/www.howtographql.com\/basics\/1-graphql-is-the-better-rest\">GraphQL posts<\/a> that describe REST as an inflexible paradigm of a bygone era, a lack of granularity that necessarily leads to data under- and over-fetching.<\/p><p>It\u2019s true that many REST APIs work this way, but this is not REST. While there\u2019s\nendless debate around which APIs are considered \u201cRESTful\u201d, I don\u2019t think\nwe need to look much further than what the letters actually stand for:<\/p><blockquote>\n  <p><strong>Representational State Transfer<\/strong>. <em>This sentence is not only what REST stands for, it is also the tiniest possible description of what REST actually means\u2026It is not a standard, rather a style describing the act of transfering a state of something by its representation.<\/em><\/p>\n  <p><em>Let\u2019s consider this:<\/em><\/p>\n  <p><em>Marcus is a farmer. He has a ranch with 4 pigs, 12 chickens and 3 cows. He is now simulating a REST API while I am the client. If I want to request the current state of his farm using REST I just ask him: \u201cState?\u201d<\/em><\/p>\n  <p><em>Marcus answers: \u201c4 pigs, 12 chickens, 3 cows\u201d.\nThis is the most simple example of Representional State Transfer. Marcus transfered the state of his farm to me using a representation. The representation of the farm is the plain sentence: \u201c4 pigs, 12 chickens, 3 cows\u201d.<\/em><\/p>\n  <p><em>So lets get to the next level. How would I tell Marcus to add 2 cows to his farm the REST way?\nMaybe tell him: \u201cMarcus, please add 2 cows to your farm\u201d.<\/em><\/p>\n  <p><em>Do you think this was REST? Are we transfering state by its representation here? <strong>NO!<\/strong> This was calling a remote procedure. The procedure of adding 2 cows to the farm.<\/em><\/p>\n  <p><em>Marcus sadly answers: \u201c400, Bad Request. What do you mean?\u201d<\/em><\/p>\n  <p><em>So lets try this again. How would we do this the REST way? What was the representation again? It was \u201c4 pigs, 12 chickens, 3 cows\u201d. Ok. so let\u2019s try this again transfering the representation\u2026<\/em><\/p>\n  <p><em>me: \u201cMarcus, \u2026 4 pigs, 12 chickens, 5 cows \u2026 please!\u201d.<\/em>\n<em>Marcus: \u201cAlright !\u201d.<\/em>\n<em>me: \u201cMarcus, \u2026 what is your state now?\u201d.<\/em>\n<em>Marcus: \u201c4 pigs, 12 chickens, 5 cows\u201d.<\/em>\n<em>me: \u201cAhh, great!\u201d<\/em>\n<em>See? It was really not that hard and it was REST.<\/em><\/p>\n  <p>- <a href=\"http:\/\/www.beabetterdeveloper.com\/2013\/07\/why-rest-is-so-important.html\">\u201cWhy REST Is So Important\u201d<\/a>, <strong>Gregor Riegler<\/strong><\/p>\n<\/blockquote><p>In other words, <strong>this is REST<\/strong>:<\/p><p style=\"text-align: center;\">\n  <img width=\"80%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/rest1.gif\" alt=\"image\" \/><\/p><p>We\u2019re moving an object from the server to the client, possibly modifying\nthat object, then moving back to the server.<\/p><p>The other important aspect of REST is <strong>Links<\/strong>. Because all objects are\naddressable at a URL, we can use HTTP links to connect Resources. This\nallows for (among other things) <strong>lazy-loading<\/strong> - the logic connecting Resources can be\nchanged server-side without breaking clients. If we have a \u201cTop\nComments\u201d relationship, we can redefine \u201cTop\u201d without disturbing\nclients.<\/p><p style=\"text-align: center;\">\n  <img width=\"80%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/rest2.gif\" alt=\"image\" \/><\/p><p>Pretty simple right? Here\u2019s how we might implement this in GraphQL:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">type CreateEmployeeInput {\n  name: String\n  age: Int\n}\ntype CreateEmployeePayload {\n  employee: Employee\n}\ntype UpdateEmployeeInput {\n  employeeId: ID!\n  name: String\n  age: Int\n}\ntype UpdateEmployeePayload {\n  employee: Employee\n}\ntype DestroyEmployeeInput {\n  id: ID!\n}\ntype DestroyEmployeePayload {\n  employee: Employee\n}\ntype Employee {\n  id: ID!\n  name: String\n}\ncreateEmployee(input: CreateEmployeeInput!): CreateEmployeePayload\nupdateEmployee(input: UpdateEmployeeInput!): UpdateEmployeePayload\ndestroyEmployee(input: DestroyEmployeeInput!): DestroyEmployeePayload\nemployee(id: ID!): Employee<\/pre><\/figure><p>The first thing to note is that GraphQL is super badass at describing\nfields and types. The next thing to note is that fields and types are\nthe wrong abstraction.<\/p><p>Defining a schema like this allows bespoke, fine-grained detail. If we\nwanted, the <code class=\"highlighter-rouge\">CreateEmployeePayload<\/code> could be different than the\n<code class=\"highlighter-rouge\">UpdateEmployeePayload<\/code> - same for inputs like <code class=\"highlighter-rouge\">CreateEmployeeInput<\/code> and\n<code class=\"highlighter-rouge\">UpdateEmployeeInput<\/code>. If we wanted other actions, like\n<code class=\"highlighter-rouge\">promoteEmployee<\/code> or <code class=\"highlighter-rouge\">deactivateEmployee<\/code>, they would be easy to add and\nfollow the same basic constructs.<\/p><p>This is <strong>RPC<\/strong> - hand-crafted, custom requests. We have a high level of\n<strong>configuration<\/strong> but a low level of <strong>convention<\/strong>. Not only will\ndevelopers have to spend more time hand-crafting these requests, but\npatterns are likely to diverge from one API to the next, from team to\nteam, as time moves on. In fact, the above is really a best-case\nscenario with common naming convention of <code class=\"highlighter-rouge\">create\/update\/destroy<\/code> - the\nGithub API adds verbs like <code class=\"highlighter-rouge\">add<\/code>, <code class=\"highlighter-rouge\">remove<\/code>, <code class=\"highlighter-rouge\">lock<\/code>, <code class=\"highlighter-rouge\">move<\/code> and more.<\/p><p>Oh, and we dropped support for lazy-loading along the way.<\/p><p style=\"text-align: center;\">\n  <img src=\"https:\/\/www.graphiti.dev\/assets\/img\/conformity.png\" alt=\"image\" \/><\/p><p>The benefit of REST over RPC is conventions. Conventions cause increased\nproductivity and consistency (leading to fewer misunderstandiings and\nchances for bugs). Let\u2019s start thinking in REST, and see where it takes\nus.<\/p><p>In REST, we know the input and output is always the Resource:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">type Employee {\n  id: ID!\n  name: string\n}\ncreateEmployee(input: Employee): Employee\nupdateEmployee(input: Employee!): Employee\ndestroyEmployee(input: Employee!): Employee\nemployee(id: ID): Employee<\/pre><\/figure><p>OK, a little tighter. But there\u2019s actually no reason to type this out\neach time - we can assume developers are already familiar with the\nconvention.<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">type Employee {\n  id: ID!\n  name: string\n}\ncreateEmployee\nupdateEmployee\ndestroyEmployee\nemployee(id: ID)<\/pre><\/figure><p>Getting there. OK, and we know we\u2019re dealing with an Employee, and we\nknow we won\u2019t have custom verbs like <code class=\"highlighter-rouge\">promote<\/code> or <code class=\"highlighter-rouge\">remove<\/code> - we\u2019re\nmoving objects here, and nothing else (if that throws you for a mental\nloop, see <a href=\"https:\/\/www.youtube.com\/watch?v=HctYHe-YjnE\">this presentation by Derek Prior<\/a>, Engineering Manager at\nGitHub).<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">type Employee {\n  id: ID!\n  name: string\n}\ncreate\nupdate\ndestroy\nshow \/\/ employee(id: ID)\nindex \/\/ employee()<\/pre><\/figure><p>By adopting conventions, we not only removed boilerplate - we removed\nthe chance of subtle inconsistencies. This is better for both providers\nand consumers of the API.<\/p><p>We\u2019re just getting started.<\/p><p>The above schema covers basic CRUD. But we probably want to filter data, right? Let\u2019s say we want to return all employees with a given name:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">employee(id: ID, name: String)<\/pre><\/figure><p>Again, we\u2019re seeing chances for inconsistency. What\u2019s the <code class=\"highlighter-rouge\">name<\/code>\nparameter - straight equality? Case sensitive? Contains? I guess we\ncould throw a bunch of suffixes at it:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">employee(id: ID, name_eq: String, name_suffix: String, name_prefix: String, name_contains: String, name_not_eq: String, name_not_suffix: String, name_prefix: String, name_not_prefix: String)<\/pre><\/figure><p>Works, but a bit unwieldy\u2026and again, likely to diverge wildly across\nimplementations.<\/p><p>What about sorting? Should we do it similar to the <a href=\"https:\/\/developer.github.com\/v4\/explorer\/\">Github API<\/a>?:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">enum OrderDirection {\n  ASC\n  DESC\n}\nenum EmployeeOrderField {\n  ID\n  NAME\n}\ntype EmployeeOrder {\n  field: EmployeeOrderField!\n  direction: OrderDirection\n}\nemployee(orderBy: EmployeeOrder)<\/pre><\/figure><p>Or should we do it like <a href=\"https:\/\/www.howtographql.com\/graphql-js\/8-filtering-pagination-and-sorting\/\">How to\nGraphQL<\/a>?:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">enum EmployeeOrderByInput {\n  id_ASC\n  id_DESC\n  name_ASC\n  name_DESC\n}\nemployee(orderBy: EmployeeOrderByInput)<\/pre><\/figure><p>We have divergent APIs right off the bat, and neither one supports\nmultisort.<\/p><p style=\"text-align: center;\">\n  <img width=\"100%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/rethink-rest-graffiti.jpg\" alt=\"image\" \/><\/p><p>Earlier, we covered RESTfully moving an object between the server and\nthe client. Those objects connected together with Links, which allowed\nfor lazy-loading.<\/p><p>Instead of throwing away this paradigm, what if we just <strong>added\neager-loading<\/strong>?<\/p><p style=\"text-align: center;\">\n  <img width=\"80%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/rest3.gif\" alt=\"image\" \/><\/p><p>REST doesn\u2019t have a query\nspecification or built-in schema, but it does have this Resource concept. <strong>Instead of\nthinking of a bag fields and types, what if we thought in Resources<\/strong>?:<\/p><p>Resources have <strong>Attributes<\/strong>. An <code class=\"highlighter-rouge\">Employee<\/code> has a <code class=\"highlighter-rouge\">first_name<\/code> which is\na <code class=\"highlighter-rouge\">string<\/code>, an <code class=\"highlighter-rouge\">age<\/code> which is an <code class=\"highlighter-rouge\">integer<\/code>, and so forth.<\/p><p>We\u2019d probably want to filter and sort by these attributes right? We might add\nsome additional filters and sorts, we might want to opt-out of others,\nbut querying a Resource by its attributes serves as a reasonable\nbaseline.<\/p><p>If we have an attribute and it\u2019s a <code class=\"highlighter-rouge\">string<\/code>, we know we\u2019re\ntalking about operators like <code class=\"highlighter-rouge\">suffix<\/code> and <code class=\"highlighter-rouge\">prefix<\/code>, but an <code class=\"highlighter-rouge\">integer<\/code>\nattribute would want operators like <code class=\"highlighter-rouge\">greater_than<\/code> and <code class=\"highlighter-rouge\">less_than<\/code>.<\/p><p>Resources also have <strong>relationships<\/strong> to other Resources. We should be\nable to lazy-load those relationships (with <strong>Links<\/strong>), or eager-load\nthose relationships. Whether lazy or eager, the same logic should apply.<\/p><p>If we defined these Resources, with their attributes and relationships,\nthen <strong>the input and the output wouldn\u2019t actually matter<\/strong>. Or more\naccurately: we can <em><strong>accept<\/strong> and <strong>render<\/strong> whatever payload we want<\/em>. Maybe\nwe\u2019ll even dynamically serve different payloads based on <code class=\"highlighter-rouge\">Content-Type<\/code>.<\/p><p>OK, so really we don\u2019t need to define <em>inputs<\/em> and <em>outputs<\/em> - those can\nbe assumed by convention, swapped on-demand. What we really need to define is the <strong>Resource<\/strong>.<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\">class EmployeeResource &lt; ApplicationRecord\n  attribute :name, :string, sortable: true, filterable: true\n  attribute :age, :integer, writable: false\nend<\/pre><\/figure><p>With nothing but this Resource definition and some assumed conventions,\nwe get all this behavior out of the box:<\/p><ul><li>Create<\/li>\n  <li>Update<\/li>\n  <li>Delete<\/li>\n  <li>Read\n    <ul><li>Filter\n        <ul><li>String (<code class=\"highlighter-rouge\">name<\/code>)\n            <ul><li><code class=\"highlighter-rouge\">eq<\/code> (case sensitive)<\/li>\n              <li><code class=\"highlighter-rouge\">eql<\/code> (case insensitive)<\/li>\n              <li><code class=\"highlighter-rouge\">prefix<\/code><\/li>\n              <li><code class=\"highlighter-rouge\">suffix<\/code><\/li>\n              <li><code class=\"highlighter-rouge\">match<\/code><\/li>\n              <li><code class=\"highlighter-rouge\">not_*<\/code> (<code class=\"highlighter-rouge\">not_eq<\/code>, <code class=\"highlighter-rouge\">not_prefix<\/code>, etc)<\/li>\n            <\/ul><\/li>\n          <li>Dates and Numbers (<code class=\"highlighter-rouge\">age<\/code>)\n            <ul><li><code class=\"highlighter-rouge\">eq<\/code><\/li>\n              <li><code class=\"highlighter-rouge\">gt<\/code> (greater than)<\/li>\n              <li><code class=\"highlighter-rouge\">lt<\/code> (less than)<\/li>\n              <li><code class=\"highlighter-rouge\">gte<\/code> (greater than\/equal to)<\/li>\n              <li><code class=\"highlighter-rouge\">lte<\/code> (less than\/equal to)<\/li>\n            <\/ul><\/li>\n        <\/ul><\/li>\n      <li>Sort \/ Multisort<\/li>\n      <li>Paginate<\/li>\n      <li>Fieldsets<\/li>\n    <\/ul><\/li>\n<\/ul><p>Not just an API contract, but <strong>out-of-the-\ud83d\udce6 behavior<\/strong>. If we know\nthe configuration of a Resource, we can <strong>automate query and persistence\noperations<\/strong>. Obviously we\u2019ll need lots of ways to customize and\noverride, and support for any number of datastores and clients. But if\nwe have the configuration, a defined contract for querying and\npersistence data, we can build patterns around it.<\/p><p>There\u2019s more to this than a bunch of out-of-the-\ud83d\udce6 standards and\nbehavior. If we thought only in Fields and Types, we\u2019d use GraphiQL to\nsee something like:<\/p><p style=\"text-align: center;\">\n  <img width=\"30%\" class=\"drop\" src=\"https:\/\/user-images.githubusercontent.com\/55264\/52915902-78418d80-32a7-11e9-8515-021312258400.png\" alt=\"image\" \/><\/p><p>But if we thought in Resources, it\u2019s not just a big bag of fields. We\u2019re\nable to organize those fields around meaningful concepts:<\/p><p style=\"text-align: center;\">\n  <img width=\"40%\" class=\"drop\" src=\"https:\/\/user-images.githubusercontent.com\/55264\/52916024-e3d82a80-32a8-11e9-8bb6-07ac9bf988dc.png\" alt=\"image\" \/><\/p><p>This screenshot is from <a href=\"https:\/\/www.graphiti.dev\/guides\/vandal\">Vandal<\/a>, the Graphiti UI.<\/p><p>Because we started with a better abstraction, we ended with a more\nintuitive\nUI. As a marketer-turned-programmer myself, I really\nappreciate when data exploration tools like this are friendly to\nless-technical users. A user of Vandal doesn\u2019t need to know about <code class=\"highlighter-rouge\">Connection<\/code>s\nor <code class=\"highlighter-rouge\">Edge<\/code>s, they just need to click around. I like that my product owner and I can walk\nthrough the domain together, validating concepts and solidifying a shared\nunderstanding.<\/p><p>While both GraphQL and Graphiti can fetch your graph of data in a single\nrequest, GraphQL does this with a single endpoint. Graphiti keeps\nRESTful URLs, for three reasons. First, HTTP caching: <a href=\"https:\/\/philsturgeon.uk\/api\/2017\/01\/26\/graphql-vs-rest-caching\/\">Phil Sturgeon has\nan excellent comparison of GraphQL and endpoint-based caching here<\/a>.<\/p><p>Second, adding endpoints allows for customization. Maybe we want to send\na welcome email when we <code class=\"highlighter-rouge\">POST \/employees<\/code> but not when we <code class=\"highlighter-rouge\">POST\n\/admin\/employees<\/code>. Maybe <code class=\"highlighter-rouge\">\/exemplary_employees<\/code> applies additional query\nlogic by default, but wants to re-use everything else. In other words,\nResources form your graph, but Endpoints <em>expose<\/em> that graph to the\noutside world (and hold relevant logic around exposure). Read more in\nour <a href=\"https:\/\/www.graphiti.dev\/guides\/concepts\/endpoints\">Endpoints Guide<\/a>.<\/p><p>The third reason is lazy-loading with <a href=\"https:\/\/www.graphiti.dev\/graphiti\/guides\/concepts\/links\">Links<\/a>, which we\u2019ll cover in a bit.<\/p><p>We even get schema benefits. Schemas are great for tooling and\nbackwards-compatibility checks\u2026but when they are oriented around\nFields and Types, they can only tell you so much. When they are oriented\naround Resources, they can expose less-obvious concepts. Maybe we sort\nEmployees by <code class=\"highlighter-rouge\">created_at<\/code> by default:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\">{\n  name: \"EmployeeResource\",\n  type: \"employees\",\n  attributes: { ... },\n  default_sort: [{ \"created_at\": \"desc\" }],\n  ...\n}<\/pre><\/figure><p>Because this is specified in the schema, not only are clients more\ninformed, but changing this default would raise a backwards-compatibility\nerror:<\/p><figure class=\"highlight\"><pre class=\"language-error\" data-lang=\"error\">EmployeeResource: default sort changed from [{:created_at=&gt;\"desc\"}] to [{:last_name=&gt;\"asc\"}].<\/pre><\/figure><p>When developing in Graphiti, we introspect your Resources and\nautomatically generate the schema for you. Backwards-compatibility\nchecks can be done with a command-line task, or whenever your tests run.<\/p><p>You may be thinking, \u201c<em>OK, but REST only works for a single object. I\u2019ll\nhave to make multiple requests, and be right back where I started. I\nneed GraphQL to solve this problem<\/em>\u201d.<\/p><p>Not true.<\/p><p>Years before GraphQL came out, respected developers from different\ncompanies and backgrounds came together and began the discussion on how to improve REST\nAPIs. This wasn\u2019t a project pushed by a hundred-billion dollar\ncompany; it was an organic, community-driven effort. The result was the <a href=\"https:\/\/jsonapi.org\">JSON:API<\/a> standard,\nwhich tackled granular queries long ago:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\"># GET \/employees?include=positions\n{\n  data: {\n    id: \"123\",\n    type: \"employees\",\n    attributes: { name: \"Kayla Webb\" },\n    relationships: {\n      positions: {\n        data: { id: \"456\", type: \"positions\" }\n      }\n    }\n  },\n  included: [\n    {\n      id: \"456\",\n      type: \"positions\",\n      attributes: { title: \"Engineer\" }\n    }\n  ]\n}<\/pre><\/figure><p>Wonky payload, right? First of all, don\u2019t be scared. In Graphiti,\nyou can add <code class=\"highlighter-rouge\">.json<\/code> to the URL and output a more traditional flat\nstructure:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\">{\n  data: {\n    id: \"123\",\n    name: \"Jeesoo Ryoo\",\n    positions: [{ id: \"456\", title: \"Engineer\" }]\n  }\n}<\/pre><\/figure><p>You can absolutely develop in Graphiti this way, but you\u2019d be giving up\nsome smart things JSON:API does. One example is de-duplicating each node\nin the graph: if we\u2019re listing 100 <code class=\"highlighter-rouge\">Post<\/code>s and they all have the same\n<code class=\"highlighter-rouge\">Author<\/code>, you\u2019ll have to render that <code class=\"highlighter-rouge\">Author<\/code> 100 times. JSON:API would\nonly render it once.<\/p><p>Another is the <code class=\"highlighter-rouge\">type\/id<\/code> combo:<\/p><p>There\u2019s a bunch of reasons this is important, but one I\u2019ve always been\npartial to is websockets. This <code class=\"highlighter-rouge\">type\/id<\/code> combo allows us to uniquely\nidentify records once they\u2019ve been loaded in JS memory. So, every time a\nResource is saved we can push its state (JSON representation) to clients\nwith a websocket. Here I am randomly updating a bunch of backend data,\nand watching the UI update in real-time.<\/p><p style=\"text-align: center;\">\n  <img width=\"80%\" src=\"https:\/\/user-images.githubusercontent.com\/55264\/38929932-042408da-42dc-11e8-93ec-16b0f9b62da3.gif\" alt=\"image\" \/><\/p><p>This took only a handful of lines of code, all of which could be\npackaged into a library to make this automatic.<\/p><p>Another example is lazy-loading data with Links:<\/p><p>Instead of loading everying up-front, we want to defer loading for\nperformance reasons. Maybe we want to render our <code class=\"highlighter-rouge\">Employee<\/code> detail page\nsuper quick, and we don\u2019t need to list the <code class=\"highlighter-rouge\">Position<\/code>s until the user\nclicks something.<\/p><p>You can do this in GraphQL, but you need to bake logic into the client,\nwhich means changing the logic would break clients (read more about\nthis in the <a href=\"https:\/\/www.graphiti.dev\/guides\/concepts\/links\">Links Guide<\/a>).\nLuckily, REST and JSON:API are optimized for lazy-loading:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\"># ...\npositions: {\n  links: {\n    related: 'http:\/\/example.com\/api\/positions?filter[employee_id]=123'\n  }\n}\n# ...<\/pre><\/figure><p>Graphiti generates these Links between Resources automatically. If you\nchange the logic connecting Resources (which applies to eager loading as\nwell), we\u2019ll update the <a href=\"https:\/\/www.graphiti.dev\/guides\/concepts\/links\">Link<\/a> - clients can simply follow the\nlink, and we can change logic server-side with no breakage.<\/p><p>There\u2019s a bunch of other great stuff about JSON:API, but that is a\ntopic for another day.<\/p><blockquote>\n  <p>(<em>Quick aside: I\u2019ve always found the name of this project hilariously\nawkward, easy to confuse with any API outputting JSON. But you could say\nthe same about GraphQL! Just as JSON:API isn\u2019t the only API standard\noutputting JSON, GraphQL isn\u2019t the only Graph Query Language - in fact,\nyou could call JSON:API a Graph Query Language as well!<\/em>)<\/p>\n<\/blockquote><p>Anything you could do with a single Resource, you can do with multiple\nResources. In other words, we can fetch an <code class=\"highlighter-rouge\">Employee<\/code>, and their\n<code class=\"highlighter-rouge\">Positions<\/code> - but only <code class=\"highlighter-rouge\">active<\/code> positions where the title starts with\n<code class=\"highlighter-rouge\">Eng<\/code>, ordered by <code class=\"highlighter-rouge\">created_at<\/code>. This is called <strong>Deep Querying<\/strong>.<\/p><p>The point here is that REST doesn\u2019t mean multiple requests. Sure, our\nstate is now a graph of objects instead of a single object, but there\u2019s\nno need for a wholesale revamp. In fact, we don\u2019t need to change much at\nall.<\/p><p style=\"text-align: center;\">\n  <img width=\"100%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/persist.jpg\" alt=\"image\" \/><\/p><p>Just as we can <strong>query<\/strong> multiple Resources at once, we can also\n<strong>persist<\/strong> multiple objects at once.<\/p><p>When persisting in REST, we send the same object back to the server\nalongside a verb. That verb tells us if we\u2019re creating, updating (part\nor whole), or deleting. Same thing here. Let\u2019s specify the verb\nalongside the relationship:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\"># ...\npositions: {\n  data: { id: \"456\", type: \"positions\", method: \"destroy\" }\n}\n# ...<\/pre><\/figure><p>There are only 4 possible verbs:<\/p><ul><li><code class=\"highlighter-rouge\">create<\/code><\/li>\n  <li><code class=\"highlighter-rouge\">update<\/code><\/li>\n  <li><code class=\"highlighter-rouge\">destroy<\/code><\/li>\n  <li><code class=\"highlighter-rouge\">disassociate<\/code><\/li>\n<\/ul><p>Oh, and we\u2019ll run everything within a transaction, throwing in\nconventions for data validation as well:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\"># Response code 422\n{\n  code:  'unprocessable_entity',\n  status: '422',\n  title: \"Validation Error\",\n  detail: \"Name can't be blank\",\n  source: { pointer: '\/data\/attributes\/name' },\n  meta: {\n    attribute: :name,\n    message: \"can't be blank\",\n    code: :blank\n  }\n}<\/pre><\/figure><p>Earlier, we covered how REST was optimized for lazy-loading. Graphiti\nbuilds on top of REST to add eager-loading, and \u201ceager-persisting\u201d. In\nother words, we can both read and write a graph of data in a single\nrequest.<\/p><blockquote>\n  <p><em>You should think hard before breaking up <a href=\"https:\/\/m.signalvnoise.com\/the-majestic-monolith\">a Magestic Monolith<\/a>; beware the tradeoffs. Still, if you need it, Graphiti has your back.<\/em><\/p>\n<\/blockquote><p>We now have a consistent interface for queries and relationships. We\nalso covered how Resources connect together with Links. Put two and\ntwo together, and you\u2019ll see a Resource doesn\u2019t need to be local to same\napplication. We can have cross-API, remote Resources as well.<\/p><p>The popular GraphQL platform Apollo does something similar with <a href=\"https:\/\/www.apollographql.com\/docs\/graphql-tools\/schema-stitching.html\">Schema Stitching<\/a>:<\/p><p style=\"text-align: center;\">\n  <img width=\"50%\" src=\"https:\/\/user-images.githubusercontent.com\/55264\/52920133-506a1e00-32d7-11e9-8986-23795dbedc2c.png\" alt=\"image\" \/><\/p><p>You won\u2019t need to write code like this in Graphiti. Because we have\nconventions, we can automate this stuff. Just supply a URL:<\/p><figure class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\">class EmployeeResource &lt; ApplicationResource\n  has_many :positions, remote: 'http:\/\/example.com\/api\/positions'\nend<\/pre><\/figure><p>That\u2019s it. Everything works the same. We can fetch an <code class=\"highlighter-rouge\">Employee<\/code> and her <code class=\"highlighter-rouge\">Position<\/code>s in a single\nrequest, add additional local <em>or<\/em> remote Resources to the request, and\nDeep Query. If you use Vandal, you\u2019ll think it\u2019s all the same API.<\/p><p>Microservices \ud83c\udf89!<\/p><p style=\"text-align: center;\">\n  <img width=\"80%\" class=\"drop\" src=\"https:\/\/user-images.githubusercontent.com\/55264\/48570112-9034fa00-e8d1-11e8-8f0e-f392207db355.jpg\" alt=\"image\" \/><\/p><p>OK, not that DDD. <a href=\"https:\/\/www.amazon.com\/Domain-Driven-Design-Tackling-Complexity-Software\/dp\/0321125215\">This DDD<\/a>:<\/p><p style=\"text-align: center;\">\n  <img width=\"20%\" class=\"drop\" src=\"https:\/\/camo.githubusercontent.com\/7812f718ee453d3153b82242f7373d6aab2d630c\/68747470733a2f2f696d616765732d6e612e73736c2d696d616765732d616d617a6f6e2e636f6d2f696d616765732f492f3531735a573837736c524c2e5f53583337355f424f312c3230342c3230332c3230305f2e6a7067\" alt=\"image\" \/><\/p><p>This book can be a little dry, but the concepts are excellent. Let\u2019s\ncheck <a href=\"https:\/\/stackoverflow.com\/questions\/1222392\/can-someone-explain-domain-driven-design-ddd-in-plain-english-please\/1222488#1222488\">StackOverflow for a quick summary of DDD<\/a>:<\/p><blockquote>\n  <p><em>DDD is about trying to make your software a model of a real-world system or process. In using DDD, you are meant to work closely with a domain expert who can explain how the real-world system works. For example, if you\u2019re developing a system that handles the placing of bets on horse races, your domain expert might be an experienced bookmaker.<\/em><\/p>\n  <p><em>Between yourself and the domain expert, you build a ubiquitous language (UL), which is basically a conceptual description of the system. The idea is that you should be able to write down what the system does in a way that the domain expert can read it and verify that it is correct. In our betting example, the ubiquitous language would include the definition of words such as \u2018race\u2019, \u2018bet\u2019, \u2018odds\u2019 and so on.<\/em><\/p>\n  <p><em>The concepts described by the UL will form the basis of your object-oriented design. DDD provides some clear guidance\u2026recommends several patterns\u2026<\/em><\/p>\n  <p>- <strong>Rob Knight<\/strong><\/p>\n<\/blockquote><p>I trim the quote because the prescriptive patterns of DDD often\novershadow its core concepts. <a href=\"https:\/\/www.youtube.com\/watch?v=R2IAgnpkBck\">The creator of DDD, Eric Evans covered this in his\nkeynote last year<\/a>.<\/p><p>REST and Resources <a href=\"https:\/\/www.youtube.com\/watch?v=GFhoSMD6idk\">force a developer to think deeply about their\ndomain, and this leads to better object-oriented code<\/a>.\nRather than a bag of fields and types, Graphiti pushes you to think\ncritically about your domain.<\/p><p>A Resource is what DDD calls an <a href=\"https:\/\/lostechies.com\/jimmybogard\/2008\/05\/21\/entities-value-objects-aggregates-and-roots\/\">Entity<\/a>. REST is about moving Domain Entities from client to server and back again.<\/p><p>You don\u2019t <strong><em>need<\/em><\/strong> to know anything about DDD to develop in Graphiti.\nBut don\u2019t sleep on it.<\/p><p>You can use any HTTP client with Graphiti. If you\u2019re using JS, a simple\n<code class=\"highlighter-rouge\">fetch<\/code> will do.<\/p><p>But you may be looking for a more robust client, one that takes\nadvantage of Graphiti conventions. Look no further than\n<a href=\"https:\/\/www.graphiti.dev\/js\">Spraypaint<\/a>, our\nofficial client heavily inspired by ActiveRecord. Query your API the\nsame way you query your database:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">\/\/ All of this is chainable\nlet { data } = await Employee\n  .where({ name: { prefix: \"Jane\" } })\n  .order({ created_at: \"desc\" })\n  .page(2).per(10)\n  .select(['name', 'age'])\n  .stat({ total: 'count' })\n  .includes({ positions: 'department' })\n  .all()\nlet record = data[0]\nrecord.positions[0].department.name = 'Updated!'\nawait record.save({ with: { positions: 'department' } })<\/pre><\/figure><p>By relying on conventions we can move the URL, request, and response\nunder the hood, allowing you to focus on what matters - your domain.<\/p><p>OK, let\u2019s come full circle. Let\u2019s say some of these conventions resonate\nwith you, but nevertheless you\u2019d like to develop a GraphQL API. I <strong><em>still<\/em><\/strong> think\nGraphiti is your best bet, because Graphiti supports GraphQL.<\/p><p>Remember, we started with this long-hand RPC code:<\/p><figure class=\"highlight\"><pre class=\"language-typescript\" data-lang=\"typescript\">type CreateEmployeeInput {\n  name: String\n  age: Int\n}\ntype CreateEmployeePayload {\n  employee: Employee\n}\ntype UpdateEmployeeInput {\n  employeeId: ID!\n  name: String\n  age: Int\n}\ntype UpdateEmployeePayload {\n  employee: Employee\n}\ntype DestroyEmployeeInput {\n  id: ID!\n}\ntype DestroyEmployeePayload {\n  employee: Employee\n}\ntype Employee {\n  id: ID!\n  name: String\n}\ncreateEmployee(input: CreateEmployeeInput!): CreateEmployeePayload\nupdateEmployee(input: UpdateEmployeeInput!): UpdateEmployeePayload\ndestroyEmployee(input: DestroyEmployeeInput!): DestroyEmployeePayload\nemployee(id: ID!): Employee<\/pre><\/figure><p>Graphiti does not need all this boilerplate. But\nif we have the short-hand, that means <strong>we can automatically generate\nthe long-hand<\/strong>. We can introspect the Graphiti Resources, and use\n<a href=\"https:\/\/github.com\/rmosolgo\/graphql-ruby\">graphql-ruby<\/a> to automatically generate GraphQL code.<\/p><p>That project is <a href=\"https:\/\/github.com\/wadetandy\/graphiti-graphql\">graphiti-graphql<\/a>. While still more of an experiment at this stage, it\u2019s shaping up nicely. The main blockers are the conventions missing from any GraphQL API - how should we render validation errors, which sorting standard should we adopt, etc. But we have proof that if there\u2019s a target to hit, we can autogenerate it.<\/p><p>So we <strong><em>can<\/em><\/strong> generate GraphQL via Graphiti, but should we? Maybe! Not\nan unreasonable pursuit. If you\u2019d like to go down this route, I\u2019d love\nto hear from you!<\/p><p>Still, let\u2019s be clear what we\u2019re missing: HTTP caching, error codes, lazy-loading, and more.<\/p><blockquote>\n  <p><em>Having that level of consistency, and working with that for\njust a little while means that you can start to forget about it. And\nthat\u2019s the power of conventions in general\u2026it used to be something\nyou had to think about and make a decision. Well, decisions are bad.\nDecisions take up your brain power, and it requires brain cycles to\nconsider which or the other. The more decisions you can take out of\nthe whole thing, the more brain power you can free up to consider the\nreally important things.<\/em><\/p>\n  <p><em>If everybody is doing the same thing in the same way, it means that\nyou can easily go from one application to the other, and expect the\nsame things to happen.<\/em><\/p>\n  <p>- <a href=\"https:\/\/www.youtube.com\/watch?v=GFhoSMD6idk\">\u201cResources on Rails\u201d<\/a>, <strong>David Heinemeier Hansson<\/strong><\/p>\n<\/blockquote><p>When conventions form your programming foundation, you end up with these\nhigh-level abstractions where a few lines of code hide the underlying\ncomplexity. Often, this is flippantly referred to as \u2728\u201d<strong>Magic<\/strong>\u201c\ud83d\udd2e<\/p><p>The thing about magic is, once you learn the trick it often comes down\nto something simple: a mirror, a trick deck, a quick hand.<\/p><p>An object, moving back and forth.<\/p><p style=\"text-align: center;\">\n  <img width=\"80%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/rest1.gif\" alt=\"image\" \/><\/p><p>Simple concepts can often be the most powerful, and I will never get\ntired of seeing this trick performed.<\/p><p>There\u2019s plenty more we haven\u2019t even touched on. Check out the\n<a href=\"https:\/\/www.graphiti.dev\/guides\">Guides<\/a>, or dive right in with the <a href=\"https:\/\/www.graphiti.dev\/quickstart\">Quickstart<\/a> or <a href=\"https:\/\/www.graphiti.dev\/tutorial\">Tutorial<\/a>. If\nthings get tricky, ask for help in our <a href=\"https:\/\/join.slack.com\/t\/graphiti-api\/shared_invite\/enQtMjkyMTA3MDgxNTQzLWVkMDM3NTlmNTIwODY2YWFkMGNiNzUzZGMzOTY3YmNmZjBhYzIyZWZlZTk4YmI1YTI0Y2M0OTZmZGYwN2QxZjg\">Slack Chat<\/a>. Reach out to me\ndirectly at <a href=\"mailto:richmolj@gmail.com\">richmolj@gmail.com<\/a> or <a href=\"https:\/\/twitter.com\/richmolj\">@richmolj<\/a> on Twitter.<\/p><p>I can\u2019t believe I just wrote all that. I swear, I never set out to do any of this.<\/p><p>There are a lot of big ideas here, but this project is not some ideological crusade. Three years ago I had a bug in my application, and I found it surprisingly tricky to fix. I kept pulling that string, I connected with other developers pulling their own strings, one thing led to another, and here we are. This project is nothing more than a good-faith effort to build quality software, based on my own personal experience.<\/p><p>Let\u2019s be clear: <strong>GraphQL is awesome<\/strong>. It solves real problems in a\nuser-friendly manner, has sexy tooling and amazing libraries created\nby kickass developers. There are many advantages of GraphQL not covered\nhere, and plenty more great ideas I\u2019d like to steal. If you find some of\nthe framing here annoying, if you find yourself thinking \u201c<em>\u2026but what\nabout X and Y?<\/em>\u201d, just know I feel the same way whenever a GraphQL\npost talks about REST. It will be a continual goal of this project to\nraise the level of debate.<\/p><p>Same for RPC. I work for a company that made billions off of RPC. I tend\nto think REST is a better fit for the web, and RPC is a better fit for\nthings like financial trading systems or video games, but plenty of\nreasonable people can disagree. Conventions and abstractions come with\ntheir own cost: a learning curve, indirection, lack of flexibility. If\nyou prefer something more low-level, I completely understand. I\u2019ll\ncontinue trying to learn from you.<\/p><p>Beware projects and ideologies that claim to be a solution to all your\nproblems. Everything has tradeoffs. The best we can do is to avoid\nzero-sum thinking, and instead think about how we can take the best\nideas from a variety of solutions and experiences, striving to\ncontinually improve. It is hard, it is exhausting, it is the right thing\nto do.<\/p><p>Unless you like XML. Screw XML.<\/p><p style=\"text-align: center;\">\n  <img width=\"100%\" src=\"https:\/\/www.graphiti.dev\/assets\/img\/love-graffiti.jpg\" alt=\"image\" \/><\/p>","created_at":"2019-04-23T17:46:04+0000","updated_at":"2019-04-23T17:46:04+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":21,"domain_name":"www.graphiti.dev","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":918,"title":"ArchiveBox","url":"https:\/\/archivebox.io\/","content":"<header><h2>\ud83d\uddc3 The open source self-hosted web archive. Takes browser history\/bookmarks\/Pocket\/Pinboard\/etc., saves HTML, JS, PDFs, media, and more...<\/h2>\n          \n          \n          \n        \n      <\/header><section id=\"main_content\"><div style=\"margin: auto;\"><img src=\"https:\/\/i.imgur.com\/4nkFjdv.png\" height=\"80px\" alt=\"image\" \/>\u25b6\ufe0f<a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Quickstart\">Quickstart<\/a>|<a href=\"https:\/\/archive.sweeting.me\">Demo<\/a>|<a href=\"https:\/\/github.com\/pirate\/ArchiveBox\">Github<\/a>|<a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\">Documentation<\/a>|<a href=\"#background--motivation\">Info &amp; Motivation<\/a>|<a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community\">Community<\/a>|<a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Roadmap\">Roadmap<\/a><pre>\n\"Your own personal internet archive\" (\u7f51\u7ad9\u5b58\u6863 \/ \u722c\u866b)\n<\/pre><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/blob\/master\/LICENSE\"><img src=\"https:\/\/img.shields.io\/badge\/Open_source-MIT-green.svg?logo=git&amp;logoColor=green\" alt=\"image\" \/><\/a><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/commits\/dev\"><img src=\"https:\/\/img.shields.io\/github\/last-commit\/pirate\/ArchiveBox.svg?logo=Sublime+Text&amp;logoColor=green&amp;label=Active\" alt=\"image\" \/><\/a><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\"><img src=\"https:\/\/img.shields.io\/github\/stars\/pirate\/ArchiveBox.svg?logo=github&amp;label=Stars&amp;logoColor=blue\" alt=\"image\" \/><\/a><a href=\"https:\/\/test.pypi.org\/project\/archivebox\/\"><img src=\"https:\/\/img.shields.io\/badge\/Python-%3E%3D3.5-yellow.svg?logo=python&amp;logoColor=yellow\" alt=\"image\" \/><\/a><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Install#dependencies\"><img src=\"https:\/\/img.shields.io\/badge\/Chromium-%3E%3D59-orange.svg?logo=Google+Chrome&amp;logoColor=orange\" alt=\"image\" \/><\/a><a href=\"https:\/\/hub.docker.com\/r\/nikisweeting\/archivebox\"><img src=\"https:\/\/img.shields.io\/badge\/Docker-all%20platforms-lightblue.svg?logo=docker&amp;logoColor=lightblue\" alt=\"image\" \/><\/a><hr \/>*\ud83d\udca5 Attention: Big API changes are coming soon! Check out [v0.4.0](https:\/\/github.com\/pirate\/ArchiveBox\/pull\/207) and help us test it! \ud83d\udca5*<\/div><p><strong>ArchiveBox takes a list of website URLs you want to archive, and creates a local, static, browsable HTML clone of the content from those websites (it saves HTML, JS, media files, PDFs, images and more).<\/strong><\/p><p>You can use it to preserve access to websites you care about by storing them locally offline.  ArchiveBox imports lists of URLs, renders the pages in a headless, autheticated, user-scriptable browser, and then archives the content in multiple redundant common formats (HTML, PDF, PNG, WARC) that will last long after the originals disappear off the internet.  It automatically extracts assets and media from pages and saves them in easily-accessible folders, with out-of-the-box support for extracting git repositories, audio, video, subtitles, images, PDFs, and more.<\/p><h4 id=\"how-does-it-work\">How does it work?<\/h4><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre>echo 'http:\/\/example.com' | .\/archive\n<\/pre><\/div><\/div><p>After installing the dependencies, just pipe some new links into the <code class=\"highlighter-rouge\">.\/archive<\/code> command to start your archive.<\/p><p>ArchiveBox is written in Python 3.5 and uses wget, Chrome headless, youtube-dl, pywb, and other common unix tools to save each page you add in multiple redundant formats.  It doesn\u2019t require a constantly running server or backend, just open the generated <code class=\"highlighter-rouge\">output\/index.html<\/code> in a browser to view the archive. It can import and export links as JSON (among other formats), so it\u2019s easy to script or hook up to other APIs.  If you run it on a schedule and import from browser history or bookmarks regularly, you can sleep soundly knowing that the slice of the internet you care about will be automatically preserved in multiple, durable long-term formats that will be accessible for decades (or longer).<\/p><div style=\"margin: auto;\"><img src=\"https:\/\/i.imgur.com\/3tBL7PU.png\" width=\"30%\" alt=\"CLI Screenshot\" style=\"text-align: top;\" \/><img src=\"https:\/\/i.imgur.com\/viklZNG.png\" width=\"30%\" alt=\"Desktop index screenshot\" style=\"text-align: top;\" \/><img src=\"https:\/\/i.imgur.com\/RefWsXB.jpg\" width=\"30%\" alt=\"Desktop details page Screenshot\" \/><br \/><sup><a href=\"https:\/\/archive.sweeting.me\/\">Demo<\/a> | <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Usage\">Usage<\/a> | <a href=\"#screenshots\">Screenshots<\/a><\/sup><br \/><sub>. . . . . . . . . . . . . . . . . . . . . . . . . . . .<\/sub><\/div><h2 id=\"quickstart\">Quickstart<\/h2><p>ArchiveBox has <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Install#dependencies\">3 main dependencies<\/a> beyond <code class=\"highlighter-rouge\">python3<\/code>: <code class=\"highlighter-rouge\">wget<\/code>, <code class=\"highlighter-rouge\">chromium<\/code>, and <code class=\"highlighter-rouge\">youtube-dl<\/code>.\nTo get started, you can <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Install\">install them manually<\/a> using your system\u2019s package manager, use the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Quickstart\">automated helper script<\/a>, or use the official <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Docker\">Docker<\/a> container.  All three dependencies are optional if <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Configuration#archive-method-toggles\">disabled<\/a> in settings.<\/p><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre># 1. Install dependencies (use apt on ubuntu, brew on mac, or pkg on BSD)\napt install python3 python3-pip git curl wget youtube-dl chromium-browser\n# 2. Download ArchiveBox\ngit clone https:\/\/github.com\/pirate\/ArchiveBox.git &amp;&amp; cd ArchiveBox\n# 3. Add your first links to your archive\necho 'https:\/\/example.com' | .\/archive                  # pass URLs to archive via stdin\n.\/archive https:\/\/getpocket.com\/users\/example\/feed\/all  # or import an RSS\/JSON\/XML\/TXT feed\n<\/pre><\/div><\/div><p>Once you\u2019ve added your first links, open <code class=\"highlighter-rouge\">output\/index.html<\/code> in a browser to view the archive.  <a href=\"https:\/\/archive.sweeting.me\">DEMO: archive.sweeting.me<\/a><br \/>For more information, see the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Quickstart\">full Quickstart guide<\/a>, <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Usage\">Usage<\/a>, and <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Configuration\">Configuration<\/a> docs.<\/p><p><em>(<code class=\"highlighter-rouge\">pip install archivebox<\/code> will be available in the near future, follow our <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Roadmap\">Roadmap<\/a> for progress)<\/em><\/p><hr \/><div style=\"margin: auto;\"><img src=\"https:\/\/i.imgur.com\/PVO88AZ.png\" width=\"80%\" alt=\"image\" \/><\/div><p>Because modern websites are complicated and often rely on dynamic content, \nArchiveBox archives the sites in <strong>several different formats<\/strong> beyond what public \narchiving services like Archive.org and Archive.is are capable of saving. Using multiple \nmethods and the market-dominant browser to execute JS ensures we can save even the most \ncomplex, finicky websites in at least a few high-quality, long-term data formats.<\/p><p>ArchiveBox imports a list of URLs from stdin, remote URL, or file, then adds the pages to a local archive folder using wget to create a browsable HTML clone, youtube-dl to extract media, and a full instance of Chrome headless for PDF, Screenshot, and DOM dumps, and more\u2026<\/p><p>Running <code class=\"highlighter-rouge\">.\/archive<\/code> adds only new, unique links into <code class=\"highlighter-rouge\">output\/<\/code> on each run. Because it will ignore duplicates and only archive each link the first time you add it, you can schedule it to <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Scheduled-Archiving\">run on a timer<\/a> and re-import all your feeds multiple times a day. It will run quickly even if the feeds are large, because it\u2019s only archiving the newest links since the last run.  For each link, it runs through all the archive methods. Methods that fail will save <code class=\"highlighter-rouge\">None<\/code> and be automatically retried on the next run, methods that succeed save their output into the data folder and are never retried\/overwritten by subsequent runs.  Support for saving multiple snapshots of each site over time will be <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/issues\/179\">added soon<\/a> (along with the ability to view diffs of the changes between runs).<\/p><p>All the archived links are stored by date bookmarked in <code class=\"highlighter-rouge\">output\/archive\/&lt;timestamp&gt;<\/code>, and everything is indexed nicely with JSON &amp; HTML files. The intent is for all the content to be viewable with common software in 50 - 100 years without needing to run ArchiveBox in a VM.<\/p><h4 id=\"can-import-links-from-many-formats\">Can import links from many formats:<\/h4><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre>echo 'http:\/\/example.com' | .\/archive\n.\/archive ~\/Downloads\/firefox_bookmarks_export.html\n.\/archive https:\/\/example.com\/some\/rss\/feed.xml\n<\/pre><\/div><\/div><ul><li><img src=\"https:\/\/nicksweeting.com\/images\/bookmarks.png\" height=\"22px\" alt=\"image\" \/> Browser history or bookmarks exports (Chrome, Firefox, Safari, IE, Opera, and more)<\/li>\n  <li><img src=\"https:\/\/nicksweeting.com\/images\/rss.svg\" height=\"22px\" alt=\"image\" \/> RSS, XML, JSON, CSV, SQL, HTML, Markdown, TXT, or any other text-based format<\/li>\n  <li><img src=\"https:\/\/getpocket.com\/favicon.ico\" height=\"22px\" alt=\"image\" \/> Pocket, Pinboard, Instapaper, Shaarli, Delicious, Reddit Saved Posts, Wallabag, Unmark.it, OneTab, and more<\/li>\n<\/ul><p>See the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Usage#CLI-Usage\">Usage: CLI<\/a> page for documentation and examples.<\/p><h4 id=\"saves-lots-of-useful-stuff-for-each-imported-link\">Saves lots of useful stuff for each imported link:<\/h4><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre> ls output\/archive\/&lt;timestamp&gt;\/\n<\/pre><\/div><\/div><ul><li><strong>Index:<\/strong> <code class=\"highlighter-rouge\">index.html<\/code> &amp; <code class=\"highlighter-rouge\">index.json<\/code> HTML and JSON index files containing metadata and details<\/li>\n  <li><strong>Title:<\/strong> <code class=\"highlighter-rouge\">title<\/code> title of the site<\/li>\n  <li><strong>Favicon:<\/strong> <code class=\"highlighter-rouge\">favicon.ico<\/code> favicon of the site<\/li>\n  <li><strong>WGET Clone:<\/strong> <code class=\"highlighter-rouge\">example.com\/page-name.html<\/code> wget clone of the site, with .html appended if not present<\/li>\n  <li><strong>WARC:<\/strong> <code class=\"highlighter-rouge\">warc\/&lt;timestamp&gt;.gz<\/code> gzipped WARC of all the resources fetched while archiving<\/li>\n  <li><strong>PDF:<\/strong> <code class=\"highlighter-rouge\">output.pdf<\/code> Printed PDF of site using headless chrome<\/li>\n  <li><strong>Screenshot:<\/strong> <code class=\"highlighter-rouge\">screenshot.png<\/code> 1440x900 screenshot of site using headless chrome<\/li>\n  <li><strong>DOM Dump:<\/strong> <code class=\"highlighter-rouge\">output.html<\/code> DOM Dump of the HTML after rendering using headless chrome<\/li>\n  <li><strong>URL to Archive.org:<\/strong> <code class=\"highlighter-rouge\">archive.org.txt<\/code> A link to the saved site on archive.org<\/li>\n  <li><strong>Audio &amp; Video:<\/strong> <code class=\"highlighter-rouge\">media\/<\/code> all audio\/video files + playlists, including subtitles &amp; metadata with youtube-dl<\/li>\n  <li><strong>Source Code:<\/strong> <code class=\"highlighter-rouge\">git\/<\/code> clone of any repository found on github, bitbucket, or gitlab links<\/li>\n  <li><em>More coming soon! See the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Roadmap\">Roadmap<\/a>\u2026<\/em><\/li>\n<\/ul><p>It does everything out-of-the-box by default, but you can disable or tweak <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Configuration\">individual archive methods<\/a> via environment variables or config file.<\/p><p>If you\u2019re importing URLs with secret tokens in them (e.g Google Docs, CodiMD notepads, etc), you may want to disable some of these methods to avoid leaking private URLs to 3rd party APIs during the archiving process.  See the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Security-Overview#stealth-mode\">Security Overview<\/a> page for more details.<\/p><h2 id=\"key-features\">Key Features<\/h2><ul><li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/blob\/master\/LICENSE\"><strong>Free &amp; open source<\/strong><\/a>, doesn\u2019t require signing up for anything, stores all data locally<\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Install#dependencies\"><strong>Few dependencies<\/strong><\/a> and <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Usage#CLI-Usage\">simple command line interface<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\"><strong>Comprehensive documentation<\/strong><\/a>, <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Roadmap\">active development<\/a>, and <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community\">rich community<\/a><\/li>\n  <li><strong>Doesn\u2019t require a constantly-running server<\/strong>, proxy, or native app<\/li>\n  <li>Easy to set up <strong><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Scheduled-Archiving\">scheduled importing<\/a> from multiple sources<\/strong><\/li>\n  <li>Uses common, <strong>durable, <a href=\"#saves-lots-of-useful-stuff-for-each-imported-link\">long-term formats<\/a><\/strong> like HTML, JSON, PDF, PNG, and WARC<\/li>\n  <li><strong>Suitable for paywalled \/ <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Configuration#chrome_user_data_dir\">authenticated content<\/a><\/strong> (can use your cookies)<\/li>\n  <li>Can <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/issues\/51\"><strong>run scripts during archiving<\/strong><\/a> to <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/issues\/80\">scroll pages<\/a>, <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/issues\/175\">close modals<\/a>, expand comment threads, etc.<\/li>\n  <li>Can also <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Configuration#submit_archive_dot_org\"><strong>mirror content to 3rd-party archiving services<\/strong><\/a> automatically for redundancy<\/li>\n<\/ul><h2 id=\"background--motivation\">Background &amp; Motivation<\/h2><p>Vast treasure troves of knowledge are lost every day on the internet to link rot.  As a society, we have an imperative to preserve some important parts of that treasure, just like we preserve our books, paintings, and music in physical libraries long after the originals go out of print or fade into obscurity.<\/p><p>Whether it\u2019s to resist censorship by saving articles before they get taken down or edited, or\njust to save a collection of early 2010\u2019s flash games you love to play, having the tools to \narchive internet content enables to you save the stuff you care most about before it disappears.<\/p><div style=\"margin: auto;\"><img src=\"https:\/\/i.imgur.com\/bC6eZcV.png\" width=\"50%\" alt=\"image\" \/><br \/><sup><i>Image from <a href=\"https:\/\/digiday.com\/media\/wtf-link-rot\/\">WTF is Link Rot?<\/a>...<\/i><br \/><\/sup><\/div><p>The balance between the permanence and ephemeral nature of content on the internet is part of what makes it beautiful. \nI don\u2019t think everything should be preserved in an automated fashion, making all content permanent and never removable, but I do think people should be able to decide for themselves and effectively archive specific content that they care about.<\/p><h2 id=\"comparison-to-other-projects\">Comparison to Other Projects<\/h2><p>\u25b6 <strong>Check out our <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community\">community page<\/a> for an index of web archiving initiatives and projects.<\/strong><\/p><p><img src=\"https:\/\/i.imgur.com\/4nkFjdv.png\" width=\"10%\" alt=\"image\" style=\"float: left;\" \/> The aim of ArchiveBox is to go beyond what the Wayback Machine and other public archiving services can do, by adding a headless browser to replay sessions accurately, and by automatically extracting all the content in multiple redundant formats that will survive being passed down to historians and archivists through many generations.<\/p><h4 id=\"user-interface--intended-purpose\">User Interface &amp; Intended Purpose<\/h4><p>ArchiveBox differentiates itself from <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community#Web-Archiving-Projects\">similar projects<\/a> by being a simple, one-shot CLI inferface for users to ingest bulk feeds of URLs over extended periods, as opposed to being a backend service that ingests individual, manually-submitted URLs from a web UI.<\/p><p>An alternative tool <a href=\"https:\/\/github.com\/webrecorder\/pywb\">pywb<\/a> allows you to run a browser through an always-running archiving proxy which records the traffic to WARC files.  ArchiveBox intends to support this style of live proxy-archiving using <code class=\"highlighter-rouge\">pywb<\/code> in the future, but for now it only ingests lists of links at a time via browser history, bookmarks, RSS, etc.<\/p><p>Unlike crawler software that starts from a seed URL and works outwards, or public tools like Archive.org designed for users to manually submit links from the public internet, ArchiveBox tries to be a set-and-forget archiver suitable for archiving your entire browsing history, RSS feeds, or bookmarks, including private\/authenticated content that you wouldn\u2019t otherwise share with a centralized service.  Also by having each user store their own content locally, we can save much larger portions of everyone\u2019s browsing history than a shared centralized service would be able to handle.<\/p><h4 id=\"storage-requirements\">Storage Requirements<\/h4><p>Because ArchiveBox is designed to ingest a firehose of browser history and bookmark feeds to a local disk, it can be much more disk-space intensive than a centralized service like the Internet Archive or Archive.today.  However, as storage space gets cheaper and compression improves, you should be able to use it continuously over the years without having to delete anything.  In my experience, ArchiveBox uses about 5gb per 1000 articles, but your milage may vary depending on which options you have enabled and what types of sites you\u2019re archiving. By default, it archives everything in as many formats as possible, meaning it takes more space than a using a single method, but more content is accurately replayable over extended periods of time. Storage requirements can be reduced by using a compressed\/deduplicated filesystem like ZFS\/BTRFS, or by setting <code class=\"highlighter-rouge\">FETCH_MEDIA=False<\/code> to skip audio &amp; video files.<\/p><h2 id=\"learn-more\">Learn more<\/h2><p>Whether you want learn which organizations are the big players in the web archiving space, want to find a specific open source tool for your web archiving need, or just want to see where archivists hang out online, our Community Wiki page serves as an index of the broader web archiving community.  Check it out to learn about some of the coolest web archiving projects and communities on the web!<\/p><p><img src=\"https:\/\/i.imgur.com\/0ZOmOvN.png\" width=\"14%\" alt=\"image\" style=\"float: right;\" \/><\/p><ul><li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community\">Community Wiki<\/a>\n    <ul><li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community#The-Master-Lists\">The Master Lists<\/a><br \/><em>Community-maintained indexes of archiving tools and institutions.<\/em><\/li>\n      <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community#Web-Archiving-Projects\">Web Archiving Software<\/a><br \/><em>Open source tools and projects in the internet archiving space.<\/em><\/li>\n      <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community#Reading-List\">Reading List<\/a><br \/><em>Articles, posts, and blogs relevant to ArchiveBox and web archiving in general.<\/em><\/li>\n      <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community#Communities\">Communities<\/a><br \/><em>A collection of the most active internet archiving communities and initiatives.<\/em><\/li>\n    <\/ul><\/li>\n  <li>Check out the ArchiveBox <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Roadmap\">Roadmap<\/a> and <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Changelog\">Changelog<\/a><\/li>\n  <li>Learn why archiving the internet is important by reading the \u201c<a href=\"https:\/\/parameters.ssrc.org\/2018\/09\/on-the-importance-of-web-archiving\/\">On the Importance of Web Archiving<\/a>\u201d blog post.<\/li>\n  <li>Or reach out to me for questions and comments via <a href=\"https:\/\/twitter.com\/thesquashSH\">@theSquashSH<\/a> on Twitter.<\/li>\n<\/ul><hr \/><p><img src=\"https:\/\/read-the-docs-guidelines.readthedocs-hosted.com\/_images\/logo-dark.png\" width=\"13%\" alt=\"image\" style=\"float: right;\" \/><\/p><p>We use the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\">Github wiki system<\/a> for documentation.<\/p><p>You can also access the docs locally by looking in the <a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Home\"><code class=\"highlighter-rouge\">ArchiveBox\/docs\/<\/code><\/a> folder.<\/p><h2 id=\"getting-started\">Getting Started<\/h2><ul><li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Quickstart\">Quickstart<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Install\">Install<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Docker\">Docker<\/a><\/li>\n<\/ul><h2 id=\"reference\">Reference<\/h2><ul><li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Usage\">Usage<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Configuration\">Configuration<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Quickstart#2-get-your-list-of-urls-to-archive\">Supported Sources<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki#can-save-these-things-for-each-site\">Supported Outputs<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Scheduled-Archiving\">Scheduled Archiving<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Publishing-Your-Archive\">Publishing Your Archive<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Install-Chromium\">Chromium Install<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Security-Overview\">Security Overview<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Troubleshooting\">Troubleshooting<\/a><\/li>\n<\/ul><ul><li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Roadmap\">Roadmap<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Changelog\">Changelog<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Donations\">Donations<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox#background--motivation\">Background &amp; Motivation<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/pirate\/ArchiveBox\/wiki\/Web-Archiving-Community\">Web Archiving Community<\/a><\/li>\n<\/ul><hr \/><hr \/><div style=\"margin: auto;\"><p><img src=\"https:\/\/raw.githubusercontent.com\/Monadical-SAS\/redux-time\/HEAD\/examples\/static\/jeremy.jpg\" height=\"40px\" alt=\"image\" \/><br \/><sub><i>This project is maintained mostly in <a href=\"https:\/\/nicksweeting.com\/blog#About\">my spare time<\/a> with the help from generous contributors.<\/i><\/sub><\/p><p>Contributor Spotlight:<\/p><p><a href=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/links\/0\"><img src=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/images\/0\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/links\/1\"><img src=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/images\/1\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/links\/2\"><img src=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/images\/2\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/links\/3\"><img src=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/images\/3\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/links\/4\"><img src=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/images\/4\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/links\/5\"><img src=\"https:\/\/sourcerer.io\/fame\/pirate\/pirate\/ArchiveBox\/images\/5\" alt=\"image\" \/><\/a>\n<br \/><a href=\"https:\/\/www.patreon.com\/theSquashSH\"><img src=\"https:\/\/img.shields.io\/badge\/Donate_to_support_development-via_Patreon-%23DD5D76.svg?style=flat\" alt=\"image\" \/><\/a>\n<\/p><p><a href=\"https:\/\/twitter.com\/thesquashSH\"><img src=\"https:\/\/img.shields.io\/badge\/Tweet-%40theSquashSH-blue.svg?style=flat\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/github.com\/pirate\/ArchiveBox\"><img src=\"https:\/\/img.shields.io\/github\/stars\/pirate\/ArchiveBox.svg?style=flat&amp;label=Star+on+Github\" alt=\"image\" \/><\/a>\n\n<\/p><\/div><\/section>","created_at":"2019-04-23T16:43:13+0000","updated_at":"2019-04-23T16:43:13+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":10,"domain_name":"archivebox.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["startup","career","engineering","software"],"is_public":false,"id":917,"title":"Senior Engineers Reduce Risk","url":"https:\/\/hackernoon.com\/senior-engineers-reduce-risk-5ab2adc13c97?gi=879998caf0de","content":"<p id=\"fd6a\" class=\"graf graf--p graf-after--h3\">If you read articles on career development in software, you\u2019ll already know there are many things a senior engineer is <em class=\"markup--em markup--p-em\">not<\/em>. They are not, for instance, just someone with ten years of experience. They are not just someone who maintains a popular open source project. They are not just someone who wrote the initial prototype for their company\u2019s flagship product.<\/p><p id=\"05f8\" class=\"graf graf--p graf-after--p\">Instead, senior engineers possess many disparate skills, the exact composition of which varies from author to author. These laundry lists are <a href=\"http:\/\/www.kitchensoap.com\/2012\/10\/25\/on-being-a-senior-engineer\/\" data-href=\"http:\/\/www.kitchensoap.com\/2012\/10\/25\/on-being-a-senior-engineer\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">detailed<\/a>, <a href=\"http:\/\/frontside.io\/blog\/2016\/07\/07\/the-conjoined-triangles-of-senior-level-development.html\" data-href=\"http:\/\/frontside.io\/blog\/2016\/07\/07\/the-conjoined-triangles-of-senior-level-development.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">nuanced<\/a>, and typically leave the reader feeling as if they have a long way to go.<\/p><p id=\"6851\" class=\"graf graf--p graf-after--p\">Empirically, though, there are many people who have the title \u201csenior engineer\u201d who do not possess all these skills. Unless all of them can be explained by title inflation or poor management, this is a serious disconnect with reality. We have to treat these articles as aspirational: a description of the industry as it should be. For someone wanting to make a significant, recognized impact on their company, today, these articles are more distracting than helpful.<\/p><p id=\"c794\" class=\"graf graf--p graf-after--p\">Senior engineers reduce risk, in every sense. Often, \u201crisk\u201d is used to describe technical risk, which is that the software doesn\u2019t function properly, or is never completed at all. But there are other issues that can prevent the business atop the software from succeeding\u200a\u2014\u200arisks around process, or product design, or sales, or the company\u2019s culture. A senior engineer understands these risks, and mitigates them where possible.<\/p><h4 id=\"104c\" class=\"graf graf--h4 graf-after--p\">senior engineers affect the larger\u00a0system<\/h4><p id=\"76bd\" class=\"graf graf--p graf-after--h4\">Most startups die because they build the wrong product. The core risks are rarely technical; if no one wants the product, building it well won\u2019t change the outcome. The most important thing an engineer can do during this time is either speed up the search, or narrow the search space. Building prototypes, gaining domain expertise, and generating metrics all lower the chance the company will run out of money.<\/p><p id=\"8c24\" class=\"graf graf--p graf-after--p\">Once a company finds product-market fit, the risks change. The product has to work, scale, and easily adapt to anything learned by the customer-facing employees. New users introduce technical risk but, more importantly, new hires introduce risk into the company\u2019s processes and culture. If the company can\u2019t make new hires productive, it will collapse under the weight of its payroll.<\/p><p id=\"19aa\" class=\"graf graf--p graf-after--p\">The software will need to be broken into manageable pieces, or else new engineers will need to hold the entire growing system in their head to be productive. Where before there may have been one engineer who handled \u201ctalking to customers\u201d, the planning process may now need to include salespeople. Above all, each new hire dilutes any clarity that existed around what the company values, and what it doesn\u2019t. Left alone, the company will become whatever people were used to at their last job.<\/p><p id=\"be6a\" class=\"graf graf--p graf-after--p\">As the company continues to grow, any new project will require collaboration across teams. After a few projects get delayed because of one missing piece, the leadership will start to value predictable results above all else. Processes will be established, recurring \u201csync\u201d meetings will be scheduled. Variance in productivity, above or below the mean, will be minimized. The largest risk is perceived to be a small component which is never completed, obviating all the other related work. In this environment, a senior engineer is an engineer who works within the processes, and even refines them, to ensure reliable output.<\/p><p id=\"fa63\" class=\"graf graf--p graf-after--p\">This narrative arc describes the average growing software startup, but no company is exactly average. The precise risks will vary depending on who founded the company, what it does, and who they hire. The senior engineers may reduce risk deliberately, or simply have the right skill set at the right time. A fresh graduate with the right domain expertise can have a huge impact on the larger system without ever understanding it.<\/p><h4 id=\"2c3d\" class=\"graf graf--h4 graf-after--p\">senior engineers find\u00a0leverage<\/h4><p id=\"bba8\" class=\"graf graf--p graf-after--h4\">Impact can be serendipitous, but the greatest impact comes from a deliberate attempt to effect change with the least effort. If production is unstable, precluding a common failure mode will have a more immediate impact than volunteering for pager duty. If the company is rapidly hiring, improving and standardizing interview practices will have a broader impact than conducting a dozen phone screens every week. If there are any junior engineers, spending an hour to talk over unfamiliar topics will have a longer lasting impact than writing more code.<\/p><p id=\"b28a\" class=\"graf graf--p graf-after--p\">Where possible, solving the general form of any problem is preferable to solving a single instance. However, this is only possible if the problem space is well understood. Premature abstraction, in code or elsewhere, introduces risk. If the failure modes in production are poorly understood, volunteering for pager duty may be the only way to really help.<\/p><h4 id=\"ab4b\" class=\"graf graf--h4 graf-after--p\">senior engineers are storytellers<\/h4><p id=\"9b5d\" class=\"graf graf--p graf-after--h4\">Most risks, especially where managed effectively, never come to pass. For this reason, it\u2019s often impossible to tell the difference between real risk and perceived risk. By the same token, it can be difficult to tell the difference between real impact and perceived impact. If a senior engineer identifies a significant risk, they have to be able to concisely explain and prioritize it for a non-expert audience.<\/p><p id=\"e39c\" class=\"graf graf--p graf-after--p\">Some see this as a form of corporate politics; a skilled storyteller can protect their company from non-existent risks, and be seen as a hero. But the fact that storytelling can be misused doesn\u2019t make it invalid, just something that should be approached with a critical eye.<\/p><h4 id=\"3ec0\" class=\"graf graf--h4 graf-after--p\">senior engineers choose companies with the right\u00a0risks<\/h4><p id=\"5d18\" class=\"graf graf--p graf-after--h4\">Every company has different risks, and so every company expects something different from their senior engineers. An engineer who has spent the last five years making small, continuous improvements to the processes in a larger company may not enjoy or even understand the sort of role expected by a three person startup. The expectation that \u201csenior\u201d is a fungible title is both widespread and harmful, leading to unrealistic expectations from both engineers and companies.<\/p><p id=\"e25b\" class=\"graf graf--p graf-after--p\">Through trial and error, engineers can identify the sorts of problems they enjoy solving. A senior engineer should find a company which both has those problems, and knows it. Anything else will lead to frustration, and eventually apathy.<\/p><p id=\"b262\" class=\"graf graf--p graf-after--p\">Likewise, companies should try to communicate their risks, early and often. If we were to judge by what\u2019s asked in interviews, most companies believe their only risks are technical, but that\u2019s absurd. The interviewers know the real problems facing the company, but the polite fiction is that they\u2019re only temporary, and soon everyone will be able to focus entirely on the algorithms, which are what really matter. Since most companies tell the same story, candidates have to read between the lines to see if there\u2019s a good fit. Honesty would make things much simpler.<\/p><h4 id=\"0e8f\" class=\"graf graf--h4 graf-after--p\">senior engineers know titles don\u2019t mean\u00a0much<\/h4><p id=\"b13f\" class=\"graf graf--p graf-after--h4\">Without context, knowing someone was a \u201csenior engineer\u201d tells you almost nothing. Titles can matter a lot in some environments, but in those cases titles are just another tool for reducing risk; if being a senior, or staff, or principal engineer is the only way to make your voice heard, then pursuing those titles is worthwhile. More often, though, it\u2019s just a distraction.<\/p><blockquote id=\"1c98\" class=\"graf graf--blockquote graf-after--p\"><a href=\"http:\/\/bit.ly\/Hackernoon\" data-href=\"http:\/\/bit.ly\/Hackernoon\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow noopener\" target=\"_blank\">Hacker Noon<\/a><div> is how hackers start their afternoons. We\u2019re a part of the <\/div><a href=\"http:\/\/bit.ly\/atAMIatAMI\" data-href=\"http:\/\/bit.ly\/atAMIatAMI\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener nofollow noopener\" target=\"_blank\">@AMI<\/a><div>family. We are now <\/div><a href=\"http:\/\/bit.ly\/hackernoonsubmission\" data-href=\"http:\/\/bit.ly\/hackernoonsubmission\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener nofollow noopener\" target=\"_blank\">accepting submissions<\/a><div> and happy to <\/div><a href=\"mailto:partners@amipublications.com\" data-href=\"mailto:partners@amipublications.com\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">discuss advertising &amp;sponsorship<\/a><div> opportunities.<\/div><\/blockquote><blockquote id=\"dca4\" class=\"graf graf--blockquote graf-after--blockquote\"><div>To learn more, <a href=\"https:\/\/goo.gl\/4ofytp\" data-href=\"https:\/\/goo.gl\/4ofytp\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">read our about page<\/a>, <a href=\"http:\/\/bit.ly\/HackernoonFB\" data-href=\"http:\/\/bit.ly\/HackernoonFB\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">like\/message us on Facebook<\/a>, or simply, <a href=\"https:\/\/goo.gl\/k7XYbx\" data-href=\"https:\/\/goo.gl\/k7XYbx\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">tweet\/DM @HackerNoon.<\/a><\/div><\/blockquote><blockquote id=\"708a\" class=\"graf graf--blockquote graf-after--blockquote\"><div>If you enjoyed this story, we recommend reading our <a href=\"http:\/\/bit.ly\/hackernoonlatestt\" data-href=\"http:\/\/bit.ly\/hackernoonlatestt\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener nofollow noopener\" target=\"_blank\">latest tech stories<\/a> and <a href=\"https:\/\/hackernoon.com\/trending\" data-href=\"https:\/\/hackernoon.com\/trending\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener nofollow\" target=\"_blank\">trending tech stories<\/a>. Until next time, don\u2019t take the realities of the world for granted!<\/div><\/blockquote><figure id=\"d811\" class=\"graf graf--figure graf-after--blockquote graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/bit.ly\/2LLdEDb\" data-href=\"http:\/\/bit.ly\/2LLdEDb\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"noopener\" target=\"_blank\"><img class=\"graf-image\" data-image-id=\"1*Xv3XV-DgnRK9WjRs90xGMA@2x.png\" data-width=\"1400\" data-height=\"774\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Xv3XV-DgnRK9WjRs90xGMA@2x.png\" alt=\"image\" \/><\/a><\/div><\/figure>","created_at":"2019-04-22T13:44:30+0000","updated_at":"2019-04-22T13:44:42+0000","published_at":"2016-07-18T14:38:56+0000","published_by":["Zach Tellman"],"annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"hackernoon.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":916,"title":"cool-RR\/PySnooper","url":"https:\/\/github.com\/cool-RR\/pysnooper","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/cool-RR\/PySnooper\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/630fa0aa2d0fef0cd70eb552933eeee67d139948\/68747470733a2f2f696d672e736869656c64732e696f2f7472617669732f636f6f6c2d52522f5079536e6f6f7065722f6d61737465722e737667\" alt=\"Travis CI\" data-canonical-src=\"https:\/\/img.shields.io\/travis\/cool-RR\/PySnooper\/master.svg\" \/><\/a><\/p>\n<p><strong>PySnooper<\/strong> is a poor man's debugger.<\/p>\n<p>You're trying to figure out why your Python code isn't doing what you think it should be doing. You'd love to use a full-fledged debugger with breakpoints and watches, but you can't be bothered to set one up right now.<\/p>\n<p>You want to know which lines are running and which aren't, and what the values of the local variables are.<\/p>\n<p>Most people would use <code>print<\/code> lines, in strategic locations, some of them showing the values of variables.<\/p>\n<p><strong>PySnooper<\/strong> lets you do the same, except instead of carefully crafting the right <code>print<\/code> lines, you just add one decorator line to the function you're interested in. You'll get a play-by-play log of your function, including which lines ran and   when, and exactly when local variables were changed.<\/p>\n<p>What makes <strong>PySnooper<\/strong> stand out from all other code intelligence tools? You can use it in your shitty, sprawling enterprise codebase without having to do any setup. Just slap the decorator on, as shown below, and redirect the output to a dedicated log file by specifying its path as the first argument.<\/p>\n<p>We're writing a function that converts a number to binary, by returing a list of bits. Let's snoop on it by adding the <code>@pysnooper.snoop()<\/code> decorator:<\/p>\n<pre>import pysnooper\n@pysnooper.snoop()\ndef number_to_bits(number):\n    if number:\n        bits = []\n        while number:\n            number, remainder = divmod(number, 2)\n            bits.insert(0, remainder)\n        return bits\n    else:\n        return [0]\nnumber_to_bits(6)\n<\/pre>\n<p>The output to stderr is:<\/p>\n<pre>Starting var:.. number = 6\n21:14:32.099769 call         3 @pysnooper.snoop()\n21:14:32.099769 line         5     if number:\n21:14:32.099769 line         6         bits = []\nNew var:....... bits = []\n21:14:32.099769 line         7         while number:\n21:14:32.099769 line         8             number, remainder = divmod(number, 2)\nNew var:....... remainder = 0\nModified var:.. number = 3\n21:14:32.099769 line         9             bits.insert(0, remainder)\nModified var:.. bits = [0]\n21:14:32.099769 line         7         while number:\n21:14:32.099769 line         8             number, remainder = divmod(number, 2)\nModified var:.. number = 1\nModified var:.. remainder = 1\n21:14:32.099769 line         9             bits.insert(0, remainder)\nModified var:.. bits = [1, 0]\n21:14:32.099769 line         7         while number:\n21:14:32.099769 line         8             number, remainder = divmod(number, 2)\nModified var:.. number = 0\n21:14:32.099769 line         9             bits.insert(0, remainder)\nModified var:.. bits = [1, 1, 0]\n21:14:32.099769 line         7         while number:\n21:14:32.099769 line        10         return bits\n21:14:32.099769 return      10         return bits\n<\/pre>\n<p>If stderr is not easily accessible for you, you can redirect the output to a file:<\/p>\n<pre>@pysnooper.snoop('\/my\/log\/file.log')\n<\/pre>\n<p>See values of some variables that aren't local variables:<\/p>\n<pre>@pysnooper.snoop(variables=('foo.bar', 'self.whatever'))\n<\/pre>\n<p>Show snoop lines for functions that your function calls:<\/p>\n<pre>@pysnooper.snoop(depth=2)\n<\/pre>\n<p>Start all snoop lines with a prefix, to grep for them easily:<\/p>\n<pre>@pysnooper.snoop(prefix='ZZZ ')\n<\/pre>\n<p>Use <code>pip<\/code>:<\/p>\n<pre>pip install pysnooper\n<\/pre>\n<p>Copyright (c) 2019 Ram Rachum, released under the MIT license.<\/p>\n<p>I provide\n<a href=\"https:\/\/chipmunkdev.com\" rel=\"nofollow\">Development services in Python and Django<\/a> and I give workshops to teach people Python and related topics.<\/p>\n<\/article>","created_at":"2019-04-22T13:23:27+0000","updated_at":"2019-04-22T13:23:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/56778?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["os","open-source","community","coc"],"is_public":false,"id":915,"title":"Why Optimistic Merging Works Better","url":"http:\/\/hintjens.com\/blog:106","content":"Why Optimistic Merging Works Better - Hintjens.com\n<div id=\"skrollr-body\"><div id=\"container-wrap-wrap\"><div id=\"container-wrap\"><div id=\"container\"><div id=\"content-wrap\"><div id=\"side-bar\"><a href=\"http:\/\/hintjens.wdfiles.com\/local--files\/nav:side\/culture.jpg\"><img src=\"http:\/\/hintjens.wdfiles.com\/local--resized-images\/nav:side\/culture.jpg\/small.jpg\" alt=\"culture.jpg\" class=\"image\" \/><\/a><p><strong>Pieter Hintjens<\/strong><\/p><p><a href=\"http:\/\/hintjens.com\/main:biography\">Biography<\/a> | <a href=\"http:\/\/hintjens.com\/main:portfolio\">Portfolio<\/a> | <a href=\"https:\/\/github.com\/hintjens\">GitHub<\/a><br \/><a href=\"http:\/\/hintjens.com\/books\">Books<\/a> | <a href=\"https:\/\/vimeo.com\/user10099130\/videos\">Videos<\/a> | <a href=\"http:\/\/www.slideshare.net\/pieterh\">Slides<\/a> | <a href=\"http:\/\/fiction.hintjens.com\/\">Stories<\/a> | <a href=\"http:\/\/wiki.hintjens.com\/\">Wiki<\/a><br \/><a href=\"http:\/\/hintjens.com\/contact\">Contact<\/a><\/p><p><strong>Thank you for the gifts<\/strong><\/p><p><a href=\"http:\/\/paypal.me\/Hintjens\">Kind people already sent donations<\/a> that will make my childrens' lives easier, when I'm not there any more. Thank you enormously! :-)<\/p><p>Podcasts\/Interviews<\/p><p><a href=\"http:\/\/softwareengineeringdaily.com\/2016\/06\/23\/death-distributed-systems-pieter-hintjens\/\">Software Engineering Daily<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=ApqI9XLRk4k\">Interview by Adam Dymitruk<\/a><br \/><a href=\"http:\/\/devchat.tv\/ruby-rogues\/188-rr-community-building-with-pieter-hintjens\">Ruby Rogues #188, Community Building<\/a><\/p><p>Talks on video<\/p><p><a href=\"http:\/\/deredactie.be\/cm\/vrtnieuws\/videozone\/programmas\/deafspraak\/2.44331\">\"De afspraak\", VRT, May 2016<\/a> (panel discussion)<br \/><a href=\"http:\/\/www.rtl.be\/rtltvi\/video\/581260.aspx\">RTL TV, May 2016<\/a> (panel discussion)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=uzxcILudFWM\">DomCode 2015, Nov 2015<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=7HECD3eLoVo\">Coding Serbia, Oct 2015<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=O8CbzKREAj4\">Euroscipy, Cambridge, Aug 2015<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=FKu09-_AaeM&amp;index=4&amp;list=PLFxMLyzsWeqr4SCyx0ZmFkVz2vZJk_CI5\">PyGrunn, Groningen, May 2015<\/a><br \/><a href=\"http:\/\/www.ustream.tv\/recorded\/61450822\">Craft Conf, Budapest, Apr 2015<\/a><br \/><a href=\"http:\/\/www.infoq.com\/presentations\/protocol-design-2015\">QCon London, Mar 2015<\/a><br \/><a href=\"http:\/\/www.infoq.com\/presentations\/protocol-design\">Code Mesh London, Nov 2014<\/a><br \/><a href=\"http:\/\/www.infoq.com\/presentations\/trick-better-software\">Build Stuff, Vilnius 2014<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=_QF9sFJGJuc\">Build Stuff, Vilnius, Nov 2014<\/a> (interview)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=xFVDNTXIC_Y\">Coding Serbia, Nov 2014<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=DAOFZfSlFs0\">Devnology, Leusden, Oct 2014<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=QB7VXSc5H3A\">EuroPython, Berlin, Jul 2014<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=qTNqgwN5gzA&amp;feature=youtu.be\">Build Stuff, Vilnius, Dec 2013<\/a> (interview)<br \/><a href=\"http:\/\/www.infoq.com\/presentations\/digital-revolution-freedom\">Build Stuff, Vilnius, Dec 2013<\/a> (keynote)<br \/><a href=\"http:\/\/vimeo.com\/82175816\">Code Mesh London, Dec 2013<\/a><br \/><a href=\"http:\/\/vimeo.com\/79378301\">DevOps Days, London, Nov 2013<\/a><br \/><a href=\"https:\/\/vimeo.com\/68236487\">NDC Oslo, Jun 2013<\/a><br \/><a href=\"https:\/\/vimeo.com\/69666055\">CERN - Geneva, Jun 2013<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=uj-li0LO_2g\">Tech Mesh, London, Dec 2012<\/a><br \/><a href=\"http:\/\/www.infoq.com\/presentations\/Architecture-Scale-ZeroMQ\">Strange Loop, St.Louis, Oct 2012<\/a><br \/><a href=\"https:\/\/vimeo.com\/37854675\">ZeroMQ@PDX part 2 - Portland, Feb 2012<\/a><br \/><a href=\"https:\/\/vimeo.com\/38117896\">ZeroMQ@PDX part 1 - Portland, Feb 2012<\/a><br \/><a href=\"http:\/\/www.youtube.com\/watch?gl=BE&amp;v=4II94h6-IFY\">FLOSS Weekly, Petaliuma CA, Dec 2011<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=CCBYzKfmQ4U\">FOSDEM 2011, Brussels<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=h0facLxJPUE\">Berlin Buzzwords 2010<\/a> (keynote)<br \/><a href=\"https:\/\/www.youtube.com\/watch?v=bQgPArfbXm4\">FOSDEM 2009, Brussels<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=fJo1jmuOuqQ\">Hellenic FOSS Conference 2008, Athens part 1<\/a><br \/><a href=\"https:\/\/www.youtube.com\/watch?v=Hx6cS1VLhoo\">Hellenic FOSS Conference 2008, Athens part 2<\/a><\/p><p>Email me at moc.snejtnih|reteip#moc.snejtnih|reteip if you'd like me to present at your event, or in your organization.<\/p><a href=\"http:\/\/hintjens.com\/books\"><img src=\"https:\/\/www.createspace.com\/Img\/T448\/T45\/T21\/BookCoverImage.jpg?random=0.4431380270844333\" width=\"100\" alt=\"BookCoverImage.jpg?random=0.4431380270844333\" class=\"image\" \/><\/a><a href=\"http:\/\/hintjens.com\/books\"><img src=\"http:\/\/hintjens.wdfiles.com\/local--files\/nav:side\/socarch.png\" width=\"100\" alt=\"socarch.png\" class=\"image\" \/><\/a><a href=\"http:\/\/hintjens.com\/books\"><img src=\"http:\/\/hintjens.wdfiles.com\/local--files\/nav:side\/confessions.jpg\" width=\"100\" alt=\"confessions.jpg\" class=\"image\" \/><\/a><a href=\"http:\/\/hintjens.com\/books\"><img src=\"http:\/\/hintjens.wdfiles.com\/local--files\/nav:side\/psccover.jpg\" width=\"100\" alt=\"psccover.jpg\" class=\"image\" \/><\/a><a href=\"http:\/\/hintjens.com\/books\"><img src=\"http:\/\/akamaicovers.oreilly.com\/images\/0636920026136\/cat.gif\" width=\"100\" alt=\"cat.gif\" class=\"image\" \/><\/a><a href=\"http:\/\/hintjens.com\/books\"><img src=\"http:\/\/hintjens.wdfiles.com\/local--files\/nav:side\/scalablec.png\" width=\"100\" alt=\"scalablec.png\" class=\"image\" \/><\/a><a href=\"http:\/\/hintjens.com\/books\"><img src=\"https:\/\/www.createspace.com\/Img\/T409\/T85\/T11\/BookCoverImage.jpg?random=0.7859812398986378\" width=\"100\" alt=\"BookCoverImage.jpg?random=0.7859812398986378\" class=\"image\" \/><\/a><p><strong><a href=\"http:\/\/hintjens.com\/books\">Download and read online<\/a><\/strong><\/p><p><strong><a href=\"http:\/\/hintjens.com\/legal:start\">Legal<\/a> | <a href=\"http:\/\/hintjens.com\/about\">About<\/a> | <a href=\"http:\/\/hintjens.com\/contact\">Contact<\/a><\/strong><\/p><\/div><div id=\"main-content\"><p>Why Optimistic Merging Works Better<\/p><p><a href=\"http:\/\/hintjens.com\/blog:_start\">All Articles<\/a> \u00bb Why Optimistic Merging Works Better<\/p><div id=\"page-content\"><div><p><a href=\"http:\/\/www.wikidot.com\/user:info\/pieterh\"><img class=\"small\" src=\"http:\/\/www.wikidot.com\/avatar.php?userid=99&amp;amp;size=small&amp;amp;timestamp=1473156446\" alt=\"pieterh\" \/><\/a><a href=\"http:\/\/www.wikidot.com\/user:info\/pieterh\">pieterh<\/a> wrote on 16 Nov 2015 15:14<\/p><div class=\"image-container floatleft\"><a href=\"http:\/\/hintjens.wdfiles.com\/local--files\/blog:106\/user_add.png\"><img src=\"http:\/\/hintjens.wdfiles.com\/local--resized-images\/blog:106\/user_add.png\/small.jpg\" alt=\"user_add.png\" class=\"image\" \/><\/a><\/div><p>I spoke at <a href=\"http:\/\/conference.domcode.org\/\">DomCode<\/a> in November 2015 (excellent conference, small and beautiful city!) explaining my top rules for building open source communities. One person asked me later to explain why I recommend to merge quickly, without waiting for Continuous Integration testing to finish, and without review of the code. I'm going to call this strategy Optimistic Merging or OM. Here's the reasoning behind OM.<\/p><p>Standard practice (Pessimistic Merging, or PM) is to wait until CI is done, then do a code review, then test the patch on a branch, and then provide feedback to the author. The author can then fix the patch and the test\/review cycle starts again. At this stage the maintainer can (and often does) make value judgments such as \"I don't like how you do this\" or \"this doesn't fit with our project vision.\"<\/p><p>In the worst case, patches can wait for weeks, or months, to be accepted. Or they are never accepted. Or, they are rejected with various excuses and argumentation.<\/p><p>PM is how most projects work, and I believe most projects get it wrong. Let me start by listing the problems PM creates:<\/p><ul><li><em>It tells new contributors, \"guilty until proven innocent,\"<\/em> which is a negative message that creates negative emotions. Contributors who feel unwelcome will always look for alternatives. Driving away contributors is bad. Making slow, quiet enemies is worse.<\/li>\n<\/ul><ul><li><em>It gives maintainers power over new contributors<\/em>, which many maintainers abuse. This abuse can be subconscious. Yet it is widespread. Maintainers inherently strive to remain important in their project. If they can keep out potential competitors by delaying and blocking their patches, they will.<\/li>\n<\/ul><ul><li><em>It opens the door to discrimination<\/em>. One can argue, a project belongs to its maintainers, so they can choose who they want to work with. My response is: projects that are not aggressively inclusive will die, and deserve to die.<\/li>\n<\/ul><ul><li><em>It slows down the learning cycle<\/em>. Innovation demands rapid experiment-failure-success cycles. Someone identifies a problem or inefficiency in a product. Someone proposes a fix. The fix is tested and works or fails. We have learned something new. The faster this cycle happens, the faster and more accurately the project can move.<\/li>\n<\/ul><ul><li><em>It gives outsiders the chance to troll the project<\/em>. It is a simple as raising an objection to a new patch. \"I don't like this code.\" Discussions over details can use up much more effort than writing code. It is far cheaper to attack a patch than to make one. These economics favor the trolls and punish the honest contributors.<\/li>\n<\/ul><ul><li><em>It puts the burden of work on individual contributors<\/em>, which is ironic and sad for open source. We want to work together yet we're told to fix our work alone.<\/li>\n<\/ul><p>Now let's see how this works when we use Optimistic Merge. To start with, understand that not all patches nor all contributors are the same. We see at least four main cases in our open source projects:<\/p><ol><li>Good contributors who know the rules and write excellent, perfect patches.<\/li>\n<li>Good contributors who make mistakes, and who write useful yet broken patches.<\/li>\n<li>Mediocre contributors who make patches that no-one notices or cares about.<\/li>\n<li>Trollish contributors who ignore the rules, and who write toxic patches.<\/li>\n<\/ol><p>PM assumes all patches are toxic until proven good (4). Whereas in reality most patches tend to be useful, and worth improving (2).<\/p><p>Let's see how each scenario works, with PM and OM:<\/p><ol><li>PM: depending on unspecified, arbitrary criteria, patch may be merged rapidly or slowly. At least sometimes, a good contributor will be left with bad feelings. OM: good contributors feel happy and appreciated, and continue to provide excellent patches until they are done using the project.<\/li>\n<li>PM: contributor retreats, fixes patch, comes back somewhat humiliated. OM: second contributor joins in to help first fix their patch. We get a short, happy patch party. New contributor now has a coach and friend in the project.<\/li>\n<li>PM: we get a flamewar and everyone wonders why the community is so hostile. OM: the mediocre contributor is largely ignored. If patch needs fixing, it'll happen rapidly. Contributor loses interest and eventually the patch is reverted.<\/li>\n<li>PM: we get a flamewar which troll wins by sheer force of argument. Community explodes in fight-or-flee emotions. Bad patches get pushed through. OM: existing contributor immediately reverts the patch. There is no discussion. Troll may try again, and eventually may be banned. Toxic patches remain in git history forever.<\/li>\n<\/ol><p>In each case, OM has a better outcome than PM.<\/p><p>In the majority case (patches that need further work), Optimistic Merge creates the conditions for mentoring and coaching. And indeed this is what we see in ZeroMQ projects, and is one of the reasons they are such fun to work on.<\/p><p>References:<\/p><ul><li><a href=\"http:\/\/rfc.zeromq.org\/spec:22\">ZeroMQ RFC 22<\/a>, C4.1: the Collective Code Construction Contract.<\/li>\n<\/ul><p>My top tips were, for what it's worth:<\/p><ul><li><em>People before code<\/em>: build the right community and it will build the right code.<\/li>\n<li><em>Progress before consensus<\/em>: look for processes that work without upfront consensus (except on rules).<\/li>\n<li><em>Problems before solutions<\/em>: use a problem-driven process.<\/li>\n<li><em>Contracts before internals<\/em>: use contracts to test behavior, not inspection of internals.<\/li>\n<li><em>Rules before hope<\/em>: write down your development process or use C4.1.<\/li>\n<li><em>Merit before power<\/em>: treat everyone fairly and equitably.<\/li>\n<li><em>Market before product<\/em>: aim for a market of competing and interoperating projects, not a single product.<\/li>\n<\/ul><\/div><\/div><p><a href=\"\/system:page-tags\/tag\/community#pages\">community<\/a><\/p><\/div><\/div><p>Unless otherwise stated, the content of this page is licensed under <a rel=\"license\" href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/3.0\/\">Creative Commons Attribution-ShareAlike 3.0 License<\/a><\/p><\/div><\/div><\/div><\/div>\n    \n    \n    \n    \n    <noscript>\n        <\/noscript>\n<div id=\"page-options-bottom-tips\"><p>Click here to edit contents of this page.<\/p><\/div>\n<div id=\"page-options-bottom-2-tips\"><p>Click here to toggle editing of individual sections of the page (if possible).         Watch headings for an \"edit\" link when available.<\/p><p>Append content without editing the whole page source.<\/p><p>Check out how this page has evolved in the past.<\/p><p>If you want to discuss contents of this page - this is the easiest way to do it.<\/p><p>View and manage file attachments for this page.<\/p><p>A few useful tools to manage this Site.<\/p><p>See pages that link to and include this page.<\/p><p>Change the name (also URL address, possibly the category) of the page.<\/p><p>View wiki source for this page without editing.<\/p><p>View\/set parent page (used for creating breadcrumbs and structured layout).<\/p><p>Notify administrators if there is objectionable content in this page.<\/p><p>Something does not work as expected? Find out what you can do.<\/p><p>General Wikidot.com documentation and help section.<\/p><p>Wikidot.com Terms of Service - what you can, what you should not etc.<\/p><p>Wikidot.com Privacy Policy.<\/p><\/div>","created_at":"2019-04-22T12:54:58+0000","updated_at":"2019-04-22T12:55:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"hintjens.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["devops","k8s","minikube"],"is_public":false,"id":914,"title":"Learn Kubernetes Basics","url":"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/","content":"Learn Kubernetes Basics - Kubernetes<header><section id=\"hero\" class=\"light-text no-sub\"><div id=\"vendorStrip\" class=\"light-text\"><ul><li><a href=\"https:\/\/kubernetes.io\/docs\/home\/\">HOME<\/a><\/li><li><a href=\"https:\/\/kubernetes.io\/docs\/setup\/\">SETUP<\/a><\/li><li><a href=\"https:\/\/kubernetes.io\/docs\/concepts\/\">CONCEPTS<\/a><\/li><li><a href=\"https:\/\/kubernetes.io\/docs\/tasks\/\">TASKS<\/a><\/li><li><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/\" class=\"YAH\">TUTORIALS<\/a><\/li><li><a href=\"https:\/\/kubernetes.io\/docs\/reference\/\">REFERENCE<\/a><\/li><li><a href=\"https:\/\/kubernetes.io\/docs\/contribute\/\">CONTRIBUTE<\/a><\/li><\/ul><\/div><\/section><section id=\"encyclopedia\"><div id=\"docsContent\"><p><a href=\"https:\/\/github.com\/kubernetes\/website\/edit\/master\/content\/en\/docs\/tutorials\/kubernetes-basics\/_index.html\" id=\"editPageButton\" target=\"_blank\">Edit This Page<\/a><\/p><div class=\"layout\" id=\"top\"><main class=\"content\"><div class=\"row\"><div class=\"col-md-9\"><h2>Kubernetes Basics<\/h2><p>This tutorial provides a walkthrough of the basics of the Kubernetes cluster orchestration system. Each module contains some background information on major Kubernetes features and concepts, and includes an interactive online tutorial. These interactive tutorials let you manage a simple cluster and its containerized applications for yourself.<\/p><p>Using the interactive tutorials, you can learn to:<\/p><ul><li>Deploy a containerized application on a cluster<\/li><li>Scale the deployment<\/li><li>Update the containerized application with a new software version<\/li><li>Debug the containerized application<\/li><\/ul><p>The tutorials use Katacoda to run a virtual terminal in your web browser that runs Minikube, a small-scale local deployment of Kubernetes that can run anywhere. There's no need to install any software or configure anything; each interactive tutorial runs directly out of your web browser itself.<\/p><\/div><\/div><br \/><div class=\"row\"><div class=\"col-md-9\"><h2>What can Kubernetes do for you?<\/h2><p>With modern web services, users expect applications to be available 24\/7, and developers expect to deploy new versions of those applications several times a day. Containerization helps package software to serve these goals, enabling applications to be released and updated in an easy and fast way without downtime. Kubernetes helps you make sure those containerized applications run where and when you want, and helps them find the resources and tools they need to work. Kubernetes is a production-ready, open source platform designed with Google's accumulated experience in container orchestration, combined with best-of-breed ideas from the community.<\/p><\/div><\/div><div id=\"basics-modules\" class=\"content__modules\"><h2>Kubernetes Basics Modules<\/h2><div class=\"row\"><div class=\"col-md-12\"><div class=\"row\"><div class=\"col-md-4\"><div class=\"thumbnail\"><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/create-cluster\/cluster-intro\/\"><img src=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/public\/images\/module_01.svg?v=1469803628347\" alt=\"\" \/><\/a><p><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/create-cluster\/cluster-intro\/\"><h5>1. Create a Kubernetes cluster<\/h5><\/a><\/p><\/div><\/div><div class=\"col-md-4\"><div class=\"thumbnail\"><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/deploy-app\/deploy-intro\/\"><img src=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/public\/images\/module_02.svg?v=1469803628347\" alt=\"\" \/><\/a><p><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/deploy-app\/deploy-intro\/\"><h5>2. Deploy an app<\/h5><\/a><\/p><\/div><\/div><div class=\"col-md-4\"><div class=\"thumbnail\"><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/explore\/explore-intro\/\"><img src=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/public\/images\/module_03.svg?v=1469803628347\" alt=\"\" \/><\/a><p><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/explore\/explore-intro\/\"><h5>3. Explore your app<\/h5><\/a><\/p><\/div><\/div><\/div><\/div><div class=\"col-md-12\"><div class=\"row\"><div class=\"col-md-4\"><div class=\"thumbnail\"><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/expose\/expose-intro\/\"><img src=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/public\/images\/module_04.svg?v=1469803628347\" alt=\"\" \/><\/a><p><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/expose\/expose-intro\/\"><h5>4. Expose your app publicly<\/h5><\/a><\/p><\/div><\/div><div class=\"col-md-4\"><div class=\"thumbnail\"><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/scale\/scale-intro\/\"><img src=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/public\/images\/module_05.svg?v=1469803628347\" alt=\"\" \/><\/a><p><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/scale\/scale-intro\/\"><h5>5. Scale up your app<\/h5><\/a><\/p><\/div><\/div><div class=\"col-md-4\"><div class=\"thumbnail\"><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/update\/update-intro\/\"><img src=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/public\/images\/module_06.svg?v=1469803628347\" alt=\"\" \/><\/a><p><a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/update\/update-intro\/\"><h5>6. Update your app<\/h5><\/a><\/p><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"row\"><p><a class=\"btn btn-lg btn-success\" href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/create-cluster\/cluster-intro\/\" role=\"button\">Start the tutorial\u203a<\/a><\/p><\/div><\/main><\/div><h2>Feedback<\/h2><p class=\"feedback--prompt\">Was this page helpful?<\/p><\/div><\/section>\n<\/header>","created_at":"2019-04-18T18:36:10+0000","updated_at":"2019-04-18T18:36:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"kubernetes.io","preview_picture":"https:\/\/kubernetes.io\/images\/kubernetes-horizontal-color.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":913,"title":"Immutable.js","url":"Immutable.js","created_at":"2019-04-17T13:43:36+0000","updated_at":"2019-04-17T13:43:36+0000","published_by":[""],"annotations":[],"reading_time":0},{"is_archived":0,"is_starred":0,"tags":["graphs","rdf","jena","triplestore"],"is_public":false,"id":912,"title":"Apache Jena - An Introduction to RDF and the Jena RDF API","url":"https:\/\/jena.apache.org\/tutorials\/rdf_api.html","content":"<h2>Preface<\/h2><p>This is a tutorial introduction to both W3C's Resource Description Framework\n(RDF) and Jena, a Java API for RDF.  It is written for the programmer who is\nunfamiliar with RDF and who learns best by prototyping, or, for other\nreasons, wishes to move quickly to implementation. Some familiarity\nwith both XML and Java is assumed.<\/p><p>Implementing too quickly, without first understanding the RDF data model,\nleads to frustration and disappointment. Yet studying the data model\nalone is dry stuff and often leads to tortuous metaphysical conundrums.  It\nis better to approach understanding both the data model and how to use it in\nparallel. Learn a bit of the data model and try it out. Then\nlearn a bit more and try that out. Then the theory informs the practice\nand the practice the theory. The data model is quite simple, so this\napproach does not take long.<\/p><p>RDF has an XML syntax and many who are familiar with XML will think of RDF\nin terms of that syntax. This is a mistake. RDF should be\nunderstood in terms of its data model. RDF data can be represented in\nXML, but understanding the syntax is secondary to understanding the data\nmodel.<\/p><p>An implementation of the Jena API, including the working source code for\nall the examples used in this tutorial can be downloaded from\n<a href=\"https:\/\/jena.apache.org\/download\/index.cgi\"><code>jena.apache.org\/download\/index.cgi<\/code><\/a>.<\/p><h2>Table of Contents<\/h2><ol><li><a href=\"#ch-Introduction\">Introduction<\/a><\/li>\n  <li><a href=\"#ch-Statements\">Statements<\/a><\/li>\n  <li><a href=\"#ch-Writing RDF\">Writing RDF<\/a><\/li>\n  <li><a href=\"#ch-Reading RDF\">Reading RDF<\/a><\/li>\n  <li><a href=\"#ch-Prefixes\">Controlling Prefixes<\/a><\/li>\n  <li><a href=\"#ch-Jena RDF Packages\">Jena RDF Packages<\/a><\/li>\n  <li><a href=\"#ch-Navigating a Model\">Navigating a Model<\/a><\/li>\n  <li><a href=\"#ch-Querying a Model\">Querying a Model<\/a><\/li>\n  <li><a href=\"#ch-Operations on Models\">Operations on Models<\/a><\/li>\n  <li><a href=\"#ch-Containers\">Containers<\/a><\/li>\n  <li><a href=\"#ch-More about Literals and Datatypes\">More about Literals and\n    Datatypes<\/a><\/li>\n  <li><a href=\"#ch-Glossary\">Glossary<\/a><\/li>\n<\/ol><h2>The Resource Description Framework (RDF) is a standard (technically a W3C\nRecommendation) for describing resources. What is a resource?  That is\nrather a deep question and the precise definition is still the subject of\ndebate.  For our purposes we can think of it as anything we can identify.\nYou are a resource, as is your home page, this tutorial, the number one and\nthe great white whale in Moby Dick.Our examples in this tutorial will be about people.  They use an <a href=\"http:\/\/www.w3.org\/TR\/vcard-rdf\/\">RDF representation of VCARDS<\/a>.  RDF\nis best thought of in the form of node and arc diagrams.  A simple vcard\nmight look like this in RDF:\n<img border=\"0\" src=\"https:\/\/jena.apache.org\/tutorials\/figures\/fig1.png\" alt=\"figure 1\" width=\"240\" height=\"180\" \/>The <a href=\"#glos-Resource\"><i>resource<\/i><\/a>, John Smith, is shown as\nan ellipse and is identified by a Uniform Resource Identifier (URI)<span><sup><a href=\"#fn-01\">1<\/a><\/sup><\/span>,  in this case\n\"http:\/\/...\/JohnSmith\".  If you try to access that resource using your\nbrowser, you are unlikely to be successful; April the first jokes not\nwithstanding, you would be rather surprised if your browser were able to\ndeliver John Smith to your desk top. If you are unfamiliar with URI's,\nthink of them simply as rather strange looking names.Resources have <a href=\"#glos-Property\"><i>properties<\/i><\/a>.  In these\nexamples we are interested in the sort of properties that would appear on\nJohn Smith's business card. Figure 1 shows only one property, John\nSmith's full name.  A property is represented by an arc, labeled with the\nname of a property.  The name of a property is also a URI, but as URI's are\nrather long and cumbersome, the diagram shows it in XML qname form.   The\npart before the ':' is called a namespace prefix and represents a\nnamespace. The part after the ':' is called a local name and represents\na name in that namespace. Properties are usually represented in this\nqname form when written as RDF XML and it is a convenient shorthand for\nrepresenting them in diagrams and in text. Strictly, however,\nproperties are identified by a URI. The nsprefix:localname form is a\nshorthand for the URI of the namespace concatenated with the localname.\nThere is no requirement that the URI of a property resolve to anything when\naccessed by a browser.Each property has a value. In this case the value is a <a href=\"#glos-Literal\"><i>literal<\/i><\/a>, which for now we can think of as a\nstrings of characters<span><sup><a href=\"#fn-02\">2<\/a><\/sup><\/span>.\nLiterals are shown in rectangles.Jena is a Java API which can be used to create and manipulate RDF graphs\nlike this one. Jena has object classes to represent graphs, resources,\nproperties and literals. The interfaces representing resources,\nproperties and literals are called Resource, Property and Literal\nrespectively. In Jena, a graph is called a model and is represented by the Model\ninterface.The code to create this graph, or model, is simple:\n  \/\/ some definitions\nstatic String personURI    = \"http:\/\/somewhere\/JohnSmith\";\nstatic String fullName     = \"John Smith\";\n\/\/ create an empty Model\nModel model = ModelFactory.createDefaultModel();\n\/\/ create the resource\nResource johnSmith = model.createResource(personURI);\n\/\/ add the property\n johnSmith.addProperty(VCARD.FN, fullName);\nIt begins with some constant definitions and then creates an empty Model\nor model, using the <code>ModelFactory<\/code> method <code>createDefaultModel()<\/code>\nto create a memory-based model. Jena contains other implementations\nof the Model interface, e.g one which uses a relational database: these\ntypes of Model are also available from ModelFactory.The John Smith resource is then created and a property added to it.\nThe property is provided by a \"constant\" class VCARD which holds objects\nrepresenting all the definitions in the VCARD schema.  Jena provides constant\nclasses for other well known schemas, such as RDF and RDF schema themselves,\nDublin Core and OWL.The working code for this example can be found in the \/src-examples directory of\nthe Jena distribution as \n<a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial01.java\">tutorial\n1<\/a>.\nAs an exercise, take this code and modify it to create a simple VCARD for\nyourself.The code to create the resource and add the property, can be more\ncompactly written in a cascading style:\n  Resource johnSmith =\n        model.createResource(personURI)\n             .addProperty(VCARD.FN, fullName);\nNow let's add some more detail to the vcard, exploring some more features\nof RDF and Jena.In the first example, the property value was a literal. RDF\nproperties can also take other resources as their value.  Using a common RDF\ntechnique, this example shows how to represent the different parts of John\nSmith's name:\n<img border=\"0\" src=\"https:\/\/jena.apache.org\/tutorials\/figures\/fig2.png\" alt=\"figure 2\" width=\"360\" height=\"240\" \/>Here we have added a new property, vcard:N, to represent the structure of\nJohn Smith's name.  There are several things of interest about this\nModel. Note that the vcard:N property takes a resource as its value.\nNote also that the ellipse representing the compound name has no URI.\nIt is known as an <i><a href=\"#glos-blank node\">blank Node<\/a>.<\/i>The Jena code to construct this example, is again very simple. First\nsome declarations and the creation of the empty model.\n  \/\/ some definitions\nString personURI    = \"http:\/\/somewhere\/JohnSmith\";\nString givenName    = \"John\";\nString familyName   = \"Smith\";\nString fullName     = givenName + \" \" + familyName;\n\/\/ create an empty Model\nModel model = ModelFactory.createDefaultModel();\n\/\/ create the resource\n\/\/   and add the properties cascading style\nResource johnSmith\n  = model.createResource(personURI)\n         .addProperty(VCARD.FN, fullName)\n         .addProperty(VCARD.N,\n                      model.createResource()\n                           .addProperty(VCARD.Given, givenName)\n                           .addProperty(VCARD.Family, familyName));\nThe working code for this example can be found as <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial02.java\">tutorial 2<\/a> in the \/src-examples directory\nof the Jena distribution.Each arc in an RDF Model is called a <i><a href=\"#glos-Statement\">statement<\/a><\/i>. Each statement asserts a fact\nabout a resource. A statement has three parts:the <i><a href=\"#glos-Subject\"> subject<\/a><\/i> is the resource from\n    which the arc leaves\n  the <i><a href=\"#glos-Predicate\"> predicate<\/a><\/i> is the property\n    that labels the arc\n  the <i><a href=\"#glos-Object\"> object<\/a><\/i> is the resource or\n    literal pointed to by the arc\nA statement is sometimes called a <a href=\"#glos-Triple\">triple<\/a>,\nbecause of its three parts.An RDF Model is represented as a <i>set<\/i> of statements.  Each call of\n<code>addProperty<\/code> in tutorial2 added another statement to the Model.\n(Because a Model is set of statements, adding a duplicate of a statement has no\neffect.)  The Jena model interface defines a <code>listStatements()<\/code>\nmethod which returns an <code>StmtIterator<\/code>, a subtype of Java's\n<code>Iterator<\/code> over all the statements in a Model.\n<code>StmtIterator<\/code> has a method <code>nextStatement()<\/code>\nwhich returns the next statement from the iterator (the same one that\n<code>next()<\/code> would deliver, already cast to <code>Statement<\/code>).\nThe <code>Statement<\/code> interface provides accessor\nmethods to the subject, predicate and object of a statement.Now we will use that interface to extend tutorial2 to list all the\nstatements created and print them out.  The complete code for this can be\nfound in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial03.java\">tutorial 3<\/a>.\n\/\/ list the statements in the Model\nStmtIterator iter = model.listStatements();\n\/\/ print out the predicate, subject and object of each statement\nwhile (iter.hasNext()) {\n    Statement stmt      = iter.nextStatement();  \/\/ get next statement\n    Resource  subject   = stmt.getSubject();     \/\/ get the subject\n    Property  predicate = stmt.getPredicate();   \/\/ get the predicate\n    RDFNode   object    = stmt.getObject();      \/\/ get the object\n    System.out.print(subject.toString());\n    System.out.print(\" \" + predicate.toString() + \" \");\n    if (object instanceof Resource) {\n       System.out.print(object.toString());\n    } else {\n        \/\/ object is a literal\n        System.out.print(\" \\\"\" + object.toString() + \"\\\"\");\n    }\n    System.out.println(\" .\");\n} \nSince the object of a statement can be either a resource or a literal, the\n<code>getObject()<\/code> method returns an\nobject typed as <code>RDFNode<\/code>, which is a\ncommon superclass of both <code>Resource<\/code>\nand <code>Literal<\/code>.  The underlying object\nis of the appropriate type, so the code uses <code>instanceof<\/code> to\ndetermine which and\nprocesses it accordingly.When run, this program should produce output resembling:http:\/\/somewhere\/JohnSmith http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#N 413f6415-c3b0-4259-b74d-4bd6e757eb60 .\n413f6415-c3b0-4259-b74d-4bd6e757eb60 http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#Family  \"Smith\" .\n413f6415-c3b0-4259-b74d-4bd6e757eb60 http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#Given  \"John\" .\nhttp:\/\/somewhere\/JohnSmith http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#FN  \"John Smith\" .Now you know why it is clearer to draw Models.  If you look carefully, you\nwill see that each line consists of three fields representing the subject,\npredicate and object of each statement.  There are four arcs in the Model, so\nthere are four statements.  The \"14df86:ecc3dee17b:-7fff\" is an internal\nidentifier generated by Jena.  It is not a URI and should not be confused\nwith one.  It is simply an internal label used by the Jena implementation.The W3C <a href=\"http:\/\/www.w3.org\/2001\/sw\/RDFCore\/\">RDFCore Working\nGroup<\/a> have defined a similar simple notation called <a href=\"http:\/\/www.w3.org\/TR\/rdf-testcases\/#ntriples\">N-Triples<\/a>.  The name\nmeans \"triple notation\".  We will see in the next section that Jena has an\nN-Triples writer built in.Jena has methods for reading and writing RDF as XML. These can be\nused to save an RDF model to a file and later read it back in again.Tutorial 3 created a model and wrote it out in triple form. <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial04.java\">Tutorial 4<\/a> modifies tutorial 3 to write the\nmodel in RDF XML form to the standard output stream. The code again, is\nvery simple: <code>model.write<\/code> can take an <code>OutputStream<\/code>\nargument.\n  \/\/ now write the model in XML form to a file\nmodel.write(System.out);\nThe output should look something like this:\n  &lt;rdf:RDF\n  xmlns:rdf='http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#'\n  xmlns:vcard='http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#'\n &gt;\n  &lt;rdf:Description rdf:about='http:\/\/somewhere\/JohnSmith'&gt;\n    &lt;vcard:FN&gt;John Smith&lt;\/vcard:FN&gt;\n    &lt;vcard:N rdf:nodeID=\"A0\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:nodeID=\"A0\"&gt;\n    &lt;vcard:Given&gt;John&lt;\/vcard:Given&gt;\n    &lt;vcard:Family&gt;Smith&lt;\/vcard:Family&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;\nThe RDF specifications specify how to represent RDF as XML. The RDF\nXML syntax is quite complex.  The reader is referred to the <a href=\"http:\/\/www.w3.org\/TR\/rdf-primer\/\">primer<\/a> being developed by the\nRDFCore WG for a more detailed introduction.  However, let's take a quick look\nat how to interpret the above.RDF is usually embedded in an &lt;rdf:RDF&gt; element.  The element is\noptional if there are other ways of knowing that some XML is RDF, but it is\nusually present.  The RDF element defines the two namespaces used in the\ndocument.  There is then an &lt;rdf:Description&gt; element which describes\nthe resource whose URI is \"http:\/\/somewhere\/JohnSmith\".  If the rdf:about\nattribute was missing, this element would represent a blank node.The &lt;vcard:FN&gt; element describes a property of the resource.  The\nproperty name is the \"FN\" in the vcard namespace.  RDF converts this to a URI\nreference by concatenating the URI reference for the namespace prefix and\n\"FN\", the local name part of the name.  This gives a URI reference of\n\"http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#FN\".  The value of the property\nis the literal \"John Smith\".The &lt;vcard:N&gt; element is a resource.  In this case the resource is\nrepresented by a relative URI reference.  RDF converts this to an absolute\nURI reference by concatenating it with the base URI of the current\ndocument.There is an error in this RDF XML; it does not exactly represent the Model\nwe created.  The blank node in the Model has been given a URI reference.  It\nis no longer blank.  The RDF\/XML syntax is not capable of representing all\nRDF Models; for example it cannot represent a blank node which is the object\nof two statements.  The 'dumb' writer we used to write this RDF\/XML  makes no\nattempt to write correctly the subset of Models which can be written\ncorrectly.  It gives a URI to each blank node, making it no longer blank.Jena has an extensible interface which allows new writers for different\nserialization languages for RDF to be easily plugged in.  The above call\ninvoked the standard 'dumb' writer.  Jena also includes a more sophisticated\nRDF\/XML writer which can be invoked by specifying another argument to the\n<code>write()<\/code> method call:\n  \/\/ now write the model in XML form to a file\nmodel.write(System.out, \"RDF\/XML-ABBREV\");\n \nThis writer, the so called PrettyWriter, takes advantage of features of\nthe RDF\/XML abbreviated syntax to write a Model more compactly.  It is also\nable to preserve blank nodes where that is possible.  It is however, not\nsuitable for writing very large Models, as its performance is unlikely to be\nacceptable.  To write large files and preserve blank nodes, write in\nN-Triples format:\n  \/\/ now write the model in N-TRIPLES form to a file\nmodel.write(System.out, \"N-TRIPLES\");\n  \nThis will produce output similar to that of tutorial 3 which conforms to\nthe N-Triples specification.<a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial05.java\">Tutorial 5<\/a> demonstrates reading the\nstatements recorded in RDF XML form into a model. With this tutorial,\nwe have provided <a href=\"https:\/\/jena.apache.org\/tutorials\/sparql_data\/vc-db-1.rdf\">a small database of vcards in RDF\/XML form<\/a>.  The following\ncode will read it in and write it out. <em>Note that for this application to\nrun, the input file must be in the current directory.<\/em>\n  \n \/\/ create an empty model\n Model model = ModelFactory.createDefaultModel();\n \/\/ use the FileManager to find the input file\n InputStream in = FileManager.get().open( inputFileName );\nif (in == null) {\n    throw new IllegalArgumentException(\n                                 \"File: \" + inputFileName + \" not found\");\n}\n\/\/ read the RDF\/XML file\nmodel.read(in, null);\n\/\/ write it to standard out\nmodel.write(System.out);\n      \nThe second argument to the <code>read()<\/code> method call is the URI which will\nbe used for resolving relative URI's.  As there are no relative URI\nreferences in the test file, it is allowed to be empty. When run, <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial05.java\"> tutorial 5<\/a> will produce XML output which\nlooks like:\n  &lt;rdf:RDF\n  xmlns:rdf='http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#'\n  xmlns:vcard='http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#'\n &gt;\n  &lt;rdf:Description rdf:nodeID=\"A0\"&gt;\n    &lt;vcard:Family&gt;Smith&lt;\/vcard:Family&gt;\n    &lt;vcard:Given&gt;John&lt;\/vcard:Given&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about='http:\/\/somewhere\/JohnSmith\/'&gt;\n    &lt;vcard:FN&gt;John Smith&lt;\/vcard:FN&gt;\n    &lt;vcard:N rdf:nodeID=\"A0\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about='http:\/\/somewhere\/SarahJones\/'&gt;\n    &lt;vcard:FN&gt;Sarah Jones&lt;\/vcard:FN&gt;\n    &lt;vcard:N rdf:nodeID=\"A1\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about='http:\/\/somewhere\/MattJones\/'&gt;\n    &lt;vcard:FN&gt;Matt Jones&lt;\/vcard:FN&gt;\n    &lt;vcard:N rdf:nodeID=\"A2\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:nodeID=\"A3\"&gt;\n    &lt;vcard:Family&gt;Smith&lt;\/vcard:Family&gt;\n    &lt;vcard:Given&gt;Rebecca&lt;\/vcard:Given&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:nodeID=\"A1\"&gt;\n    &lt;vcard:Family&gt;Jones&lt;\/vcard:Family&gt;\n    &lt;vcard:Given&gt;Sarah&lt;\/vcard:Given&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:nodeID=\"A2\"&gt;\n    &lt;vcard:Family&gt;Jones&lt;\/vcard:Family&gt;\n    &lt;vcard:Given&gt;Matthew&lt;\/vcard:Given&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about='http:\/\/somewhere\/RebeccaSmith\/'&gt;\n    &lt;vcard:FN&gt;Becky Smith&lt;\/vcard:FN&gt;\n    &lt;vcard:N rdf:nodeID=\"A3\"\/&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;\nControlling Prefixes<\/h2><h3>Explicit prefix definitions<\/h3><p>In the previous section, we saw that the output XML declared a namespace\nprefix <code>vcard<\/code> and used that prefix to abbreviate URIs. While RDF\nuses only the full URIs, and not this shortened form, Jena provides ways\nof controlling the namespaces used on output with its <i>prefix mappings<\/i>.\nHere's a simple example.<\/p><blockquote>\n<pre>\n Model m = ModelFactory.createDefaultModel();\n String nsA = \"http:\/\/somewhere\/else#\";\n String nsB = \"http:\/\/nowhere\/else#\";\n Resource root = m.createResource( nsA + \"root\" );\n Property P = m.createProperty( nsA + \"P\" );\n Property Q = m.createProperty( nsB + \"Q\" );\n Resource x = m.createResource( nsA + \"x\" );\n Resource y = m.createResource( nsA + \"y\" );\n Resource z = m.createResource( nsA + \"z\" );\n m.add( root, P, x ).add( root, P, y ).add( y, Q, z );\n System.out.println( \"# -- no special prefixes defined\" );\n m.write( System.out );\n System.out.println( \"# -- nsA defined\" );\n m.setNsPrefix( \"nsA\", nsA );\n m.write( System.out );\n System.out.println( \"# -- nsA and cat defined\" );\n m.setNsPrefix( \"cat\", nsB );\n m.write( System.out );\n<\/pre>\n<\/blockquote><p>The output from this fragment is three lots of RDF\/XML, with\nthree different prefix mappings. First the default, with no\nprefixes other than the standard ones:<\/p><blockquote>\n<pre>\n# -- no special prefixes defined\n&lt;rdf:RDF\n    xmlns:j.0=\"http:\/\/nowhere\/else#\"\n    xmlns:rdf=\"http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#\"\n    xmlns:j.1=\"http:\/\/somewhere\/else#\" &gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/else#root\"&gt;\n    &lt;j.1:P rdf:resource=\"http:\/\/somewhere\/else#x\"\/&gt;\n    &lt;j.1:P rdf:resource=\"http:\/\/somewhere\/else#y\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/else#y\"&gt;\n    &lt;j.0:Q rdf:resource=\"http:\/\/somewhere\/else#z\"\/&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;\n<\/pre>\n<\/blockquote><p>We see that the rdf namespace is declared automatically, since it\nis required for tags such as <code>&lt;rdf:RDF&gt;<\/code> and\n<code>&lt;rdf:resource&gt;<\/code>. XML namespace declarations are also\nneeded for using the two properties P and Q, but since their\nprefixes have not been introduced to the model in this example,\nthey get invented namespace names: <code>j.0<\/code> and <code>j.1<\/code>.<\/p><p>The method <code>setNsPrefix(String prefix, String URI)<\/code>\ndeclares that the namespace <code>URI<\/code> may be abbreviated\nby <code>prefix<\/code>. Jena requires that <code>prefix<\/code> be\na legal XML namespace name, and that <code>URI<\/code> ends with a\nnon-name character. The RDF\/XML writer will turn these prefix\ndeclarations into XML namespace declarations and use them in its\noutput:\n<\/p><blockquote>\n<pre>\n# -- nsA defined\n&lt;rdf:RDF\n    xmlns:j.0=\"http:\/\/nowhere\/else#\"\n    xmlns:rdf=\"http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#\"\n    xmlns:nsA=\"http:\/\/somewhere\/else#\" &gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/else#root\"&gt;\n    &lt;nsA:P rdf:resource=\"http:\/\/somewhere\/else#x\"\/&gt;\n    &lt;nsA:P rdf:resource=\"http:\/\/somewhere\/else#y\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/else#y\"&gt;\n    &lt;j.0:Q rdf:resource=\"http:\/\/somewhere\/else#z\"\/&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;\n<\/pre>\n<\/blockquote>The other namespace still gets the constructed name, but the nsA name\nis now used in the property tags. There's no need for the prefix name\nto have anything to do with the variables in the Jena code:<blockquote>\n<pre>\n# -- nsA and cat defined\n&lt;rdf:RDF\n    xmlns:cat=\"http:\/\/nowhere\/else#\"\n    xmlns:rdf=\"http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#\"\n    xmlns:nsA=\"http:\/\/somewhere\/else#\" &gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/else#root\"&gt;\n    &lt;nsA:P rdf:resource=\"http:\/\/somewhere\/else#x\"\/&gt;\n    &lt;nsA:P rdf:resource=\"http:\/\/somewhere\/else#y\"\/&gt;\n  &lt;\/rdf:Description&gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/else#y\"&gt;\n    &lt;cat:Q rdf:resource=\"http:\/\/somewhere\/else#z\"\/&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;\n<\/pre>\n<\/blockquote>Both prefixes are used for output, and no generated prefixes are needed.<h3>Implicit prefix definitions<\/h3>As well as prefix declarations provided by calls to<code>setNsPrefix<\/code>,\nJena will remember the prefixes that were used in input to<code>model.read()<\/code>.<p>Take the output produced by the previous fragment, and paste it into\nsome file, with URL <b>file:\/tmp\/fragment.rdf<\/b> say. Then run the\ncode:\n<\/p><blockquote>\n<pre>\nModel m2 = ModelFactory.createDefaultModel();\nm2.read( \"file:\/tmp\/fragment.rdf\" );\nm2.write( System.out );\n<\/pre>\n<\/blockquote>You'll see that the prefixes from the input are preserved in the output.\nAll the prefixes are written, even if they're not used anywhere. You can\nremove a prefix with<code>removeNsPrefix(String prefix)<\/code>if you\ndon't want it in the output.<p>Since NTriples doesn't have any short way of writing URIs, it takes\nno notice of prefixes on output and doesn't provide any on input. The\nnotation <b>N3<\/b>, also supported by Jena, does have short prefixed names,\nand records them on input and uses them on output.\n<\/p><p>Jena has further operations on the prefix mappings that a model holds,\nsuch as extracting a Java <code>Map<\/code> of the exiting mappings, or\nadding a whole group of mappings at once; see the documentation for\n<code>PrefixMapping<\/code> for details.\n<\/p><h2>Jena RDF Packages<\/h2><p>Jena is a Java API for semantic web applications.  The key RDF package for\nthe application developer is\n<code>org.apache.jena.rdf.model<\/code>. The API has been defined\nin terms of interfaces so that application code can work with different\nimplementations without change. This package contains interfaces for\nrepresenting models, resources, properties, literals, statements and all the\nother key concepts of RDF, and a ModelFactory for creating models.  So that\napplication code remains independent of\nthe implementation, it is best if it uses interfaces wherever possible, not\nspecific class implementations.<\/p><p>The <code>org.apache.jena.tutorial<\/code> package contains the\nworking source code for all the examples used in this tutorial.<\/p><p>The <code>org.apache.jena...impl<\/code> packages contains\nimplementation classes which may be common to many implementations. For\nexample, they defines classes <code>ResourceImpl<\/code>,\n<code>PropertyImpl<\/code>, and <code>LiteralImpl<\/code> which may be\nused directly or subclassed by different implementations. Applications\nshould rarely, if ever, use these classes directly. For example, rather\nthan creating a new instance of <code>ResourceImpl<\/code>, it is better to\nuse the <code>createResource<\/code> method of whatever model is being\nused. That way, if the model implementation has used an optimized\nimplementation of <code>Resource<\/code>, then no conversions between the two\ntypes will be necessary.<\/p><p>So far, this tutorial has dealt mainly with creating, reading and writing\nRDF Models. It is now time to deal with accessing information held in a\nModel.<\/p><p>Given the URI of a resource, the resource object can be retrieved from a\nmodel using the <code>Model.getResource(String uri)<\/code> method. This\nmethod is defined to return a Resource object if one exists in the model, or\notherwise to create a new one. For example, to retrieve the John Smith\nresource from the model read in from the file in tutorial 5:<\/p><blockquote>\n  <pre>\/\/ retrieve the John Smith vcard resource from the model\nResource vcard = model.getResource(johnSmithURI);\n  <\/pre>\n<\/blockquote><p>The Resource interface defines a number of methods for accessing the\nproperties of a resource. The <code>Resource.getProperty(Property\np)<\/code> method accesses a property of the resource. This method does\nnot follow the usual Java accessor convention in that the type of the object\nreturned is <code>Statement,<\/code> not the <code>Property<\/code> that you\nmight have expected.  Returning the whole statement allows the application to\naccess the value of the property using one of its accessor methods which\nreturn the object of the statement. For example to retrieve the\nresource which is the value of the <code>vcard:N<\/code> property:<\/p><blockquote>\n  <pre>\/\/ retrieve the value of the N property\nResource name = (Resource) vcard.getProperty(VCARD.N)\n                                .getObject();<\/pre>\n<\/blockquote><p>In general, the object of a statement could be a resource or a literal, so\nthe application code, knowing the value must be a resource, casts the\nreturned object. One of the things that Jena tries to do is to provide\ntype specific methods so the application does not have to cast and type\nchecking can be done at compile time. The code fragment above, can be\nmore conveniently written:<\/p><blockquote>\n  <pre>\/\/ retrieve the value of the N property\nResource name = vcard.getProperty(VCARD.N)\n                     .getResource();<\/pre>\n<\/blockquote><p>Similarly, the literal value of a property can be retrieved:<\/p><blockquote>\n  <pre>\/\/ retrieve the given name property\nString fullName = vcard.getProperty(VCARD.FN)\n                        .getString();<\/pre>\n<\/blockquote><p>In this example, the vcard resource has only one <code>vcard:FN<\/code> and\none <code>vcard:N<\/code> property. RDF permits a resource to repeat a\nproperty; for example Adam might have more than one nickname.  Let's give him\ntwo:<\/p><blockquote>\n  <pre>\/\/ add two nickname properties to vcard\nvcard.addProperty(VCARD.NICKNAME, \"Smithy\")\n     .addProperty(VCARD.NICKNAME, \"Adman\");<\/pre>\n<\/blockquote><p>As noted before, Jena represents an RDF Model as <em>set <\/em>of\nstatements, so adding a statement with the subject, predicate and object as\none already in the Model will have no effect.  Jena does not define which of\nthe two nicknames present in the Model will be returned.  The result of\ncalling <code>vcard.getProperty(VCARD.NICKNAME)<\/code> is indeterminate. Jena\nwill return one of the values, but there is no guarantee even that two\nconsecutive calls will return the same value.<\/p><p>If it is possible that a property may occur more than once, then the\nResource.listProperties(Property p) method can be used to return an iterator\nwhich will list them all.  This method returns an iterator which returns\nobjects of type <code>Statement<\/code>.  We can list the nicknames like\nthis:<\/p><blockquote>\n  <pre>\/\/ set up the output\nSystem.out.println(\"The nicknames of \\\"\"\n                      + fullName + \"\\\" are:\");\n\/\/ list the nicknames\nStmtIterator iter = vcard.listProperties(VCARD.NICKNAME);\nwhile (iter.hasNext()) {\n    System.out.println(\"    \" + iter.nextStatement()\n                                    .getObject()\n                                    .toString());\n}<\/pre>\n<\/blockquote><p>This code can be found in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial06.java\"> tutorial 6<\/a>.\nThe statement iterator <code>iter<\/code> produces each and every statement\nwith subject <code>vcard<\/code> and predicate <code>VCARD.NICKNAME<\/code>,\nso looping over it allows us to fetch each statement by using\n<code>nextStatement()<\/code>, get the object field, and convert it to\na string.\nThe code produces the following output when run:<\/p><blockquote>\n  <pre>The nicknames of \"John Smith\" are:\n    Smithy\n    Adman<\/pre>\n<\/blockquote><p>All the properties of a resource can be listed by using the\n<code>listProperties()<\/code> method without an argument.\n<\/p><h2>Querying a Model<\/h2><p>The previous section dealt with the case of navigating a model from a\nresource with a known URI. This section deals with searching a\nmodel. The core Jena API supports only a limited query primitive.  The\nmore powerful query facilities of SPARQL are described elsewhere.<\/p><p>The <code>Model.listStatements()<\/code> method, which lists all the\nstatements in a model, is perhaps the crudest way of querying a model.\nIts use is not recommended on very large Models.\n<code>Model.listSubjects()<\/code> is similar, but returns an iterator over\nall resources that have properties, <i>ie<\/i> are the subject of some\nstatement.<\/p><p><code>Model.listSubjectsWithProperty(Property p, RDFNode\no)<\/code> will return an iterator over all the resources which\nhave property <code>p<\/code> with value <code>o<\/code>. If we assume that\nonly vcard resources\nwill have <code>vcard:FN<\/code> property, and that in our data, all such\nresources have such a property, then we can find all the vcards like this:<\/p><blockquote>\n  <pre>\/\/ list vcards\nResIterator iter = model.listSubjectsWithProperty(VCARD.FN);\nwhile (iter.hasNext()) {\n    Resource r = iter.nextResource();\n    ...\n}<\/pre>\n<\/blockquote><p>All these query methods are simply syntactic sugar over a primitive query\nmethod <code>model.listStatements(Selector s)<\/code>.  This method returns an\niterator over all the statements in the model 'selected' by <code>s<\/code>.\nThe selector interface is designed to be extensible, but for now, there is\nonly one implementation of it, the class <code>SimpleSelector<\/code> from the\npackage <code>org.apache.jena.rdf.model<\/code>.  Using\n<code>SimpleSelector <\/code>is one of the rare occasions in Jena when it is\nnecessary to use a specific class rather than an interface.  The\n<code>SimpleSelector<\/code> constructor takes three arguments:<\/p><blockquote>\n  <pre>Selector selector = new SimpleSelector(subject, predicate, object)\n  <\/pre>\n<\/blockquote><p>This selector will select all statements with a subject that matches\n<code>subject<\/code>, a predicate that matches <code>predicate<\/code> and an\nobject that matches <code>object<\/code>.  If a <code>null<\/code> is supplied\nin any of the positions, it matches anything; otherwise they match corresponding\nequal resources or literals. (Two resources are equal if they have equal URIs\nor are the same blank node; two literals are the same if all their components\nare equal.) Thus:<\/p><blockquote>\n  <pre>Selector selector = new SimpleSelector(null, null, null);\n  <\/pre>\n<\/blockquote><p>will select all the statements in a Model.<\/p><blockquote>\n  <pre>Selector selector = new SimpleSelector(null, VCARD.FN, null);\n  <\/pre>\n<\/blockquote><p>will select all the statements with VCARD.FN as their predicate, whatever\nthe subject or object. As a special shorthand,\n<\/p><blockquote><div><code>listStatements( S, P, O )\n<\/code><\/div><\/blockquote>is equivalent to<blockquote><div><code>listStatements( new SimpleSelector( S, P, O ) )\n<\/code><\/div><\/blockquote><p>\nThe following code, which can be found in full in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial07.java\">tutorial 7<\/a> lists the full names on all the\nvcards in the database.<\/p><blockquote>\n  <pre>\/\/ select all the resources with a VCARD.FN property\nResIterator iter = model.listSubjectsWithProperty(VCARD.FN);\nif (iter.hasNext()) {\n    System.out.println(\"The database contains vcards for:\");\n    while (iter.hasNext()) {\n        System.out.println(\"  \" + iter.nextResource()\n                                      .getProperty(VCARD.FN)\n                                      .getString());\n    }\n} else {\n    System.out.println(\"No vcards were found in the database\");\n}\n            <\/pre>\n<\/blockquote><p>This should produce output similar to the following:<\/p><blockquote>\n  <pre>The database contains vcards for:\n  Sarah Jones\n  John Smith\n  Matt Jones\n  Becky Smith\n  <\/pre>\n<\/blockquote><p>Your next exercise is to modify this code to use <code>SimpleSelector\n<\/code>instead of <code>listSubjectsWithProperty<\/code>.<\/p><p>Lets see how to implement some finer control over the statements selected.\n<code>SimpleSelector<\/code> can be subclassed and its selects method modified\nto perform further filtering:<\/p><blockquote>\n  <pre>\/\/ select all the resources with a VCARD.FN property\n\/\/ whose value ends with \"Smith\"\nStmtIterator iter = model.listStatements(\n    new SimpleSelector(null, VCARD.FN, (RDFNode) null) {\n        public boolean selects(Statement s)\n            {return s.getString().endsWith(\"Smith\");}\n    });<\/pre>\n<\/blockquote><p>This sample code uses a neat Java technique of overriding a method\ndefinition inline when creating an instance of the class.  Here the\n<code>selects(...)<\/code> method checks to ensure that the full name ends\nwith \"Smith\".  It is important to note that filtering based on the subject,\npredicate and object arguments takes place before the\n<code>selects(...)<\/code> method is called, so the extra test will only be\napplied to matching statements.<\/p><p>The full code can be found in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial08.java\">tutorial\n8<\/a> and produces output like this:<\/p><blockquote>\n  <pre>The database contains vcards for:\n  John Smith\n  Becky Smith<\/pre>\n<\/blockquote><p>You might think that:<\/p><blockquote>\n  <pre>\/\/ do all filtering in the selects method\nStmtIterator iter = model.listStatements(\n  new\n      SimpleSelector(null, null, (RDFNode) null) {\n          public boolean selects(Statement s) {\n              return (subject == null   || s.getSubject().equals(subject))\n                  &amp;&amp; (predicate == null || s.getPredicate().equals(predicate))\n                  &amp;&amp; (object == null    || s.getObject().equals(object)) ;\n          }\n     }\n     });<\/pre>\n<\/blockquote><p>is equivalent to:<\/p><blockquote>\n  <pre>StmtIterator iter =\n  model.listStatements(new SimpleSelector(subject, predicate, object)<\/pre>\n<\/blockquote><p>Whilst functionally they may be equivalent, the first form will list all\nthe statements in the Model and test each one individually, whilst the second\nallows indexes maintained by the implementation to improve performance.  Try\nit on a large Model and see for yourself, but make a cup of coffee first.<\/p><h2>Operations on Models<\/h2><p>Jena provides three operations for manipulating Models as a whole.  These\nare the common set operations of union, intersection and difference.<\/p><p>The union of two Models is the union of the sets of statements which\nrepresent each Model.  This is one of the key operations that the design of\nRDF supports.  It enables data from disparate data sources to be merged.\nConsider the following two Models:<\/p><p>\n<img alt=\"figure 4\" src=\"https:\/\/jena.apache.org\/tutorials\/figures\/fig4.png\" width=\"240\" height=\"240\" \/>and\n<img alt=\"figure 5\" src=\"https:\/\/jena.apache.org\/tutorials\/figures\/fig5.png\" width=\"240\" height=\"240\" \/><\/p><p>When these are merged, the two http:\/\/...JohnSmith nodes are merged into\none and the duplicate <code>vcard:FN<\/code> arc is dropped to produce:<\/p><p>\n<img alt=\"figure 6\" src=\"https:\/\/jena.apache.org\/tutorials\/figures\/fig6.png\" width=\"540\" height=\"240\" \/><\/p><p>Lets look at the code to do this (the full code is in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial09.java\">tutorial 9<\/a>) and see what happens.<\/p><blockquote>\n  <pre>\/\/ read the RDF\/XML files\nmodel1.read(new InputStreamReader(in1), \"\");\nmodel2.read(new InputStreamReader(in2), \"\");\n\/\/ merge the Models\nModel model = model1.union(model2);\n\/\/ print the Model as RDF\/XML\nmodel.write(system.out, \"RDF\/XML-ABBREV\");<\/pre>\n<\/blockquote><p>The output produced by the pretty writer looks like this:<\/p><blockquote>\n  <pre>&lt;rdf:RDF\n    xmlns:rdf=\"<a href=\"http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#\">http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#<\/a>\"\n    xmlns:vcard=\"http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#\"&gt;\n  &lt;rdf:Description rdf:about=\"http:\/\/somewhere\/JohnSmith\/\"&gt;\n    &lt;vcard:EMAIL&gt;\n      &lt;vcard:internet&gt;\n        &lt;rdf:value&gt;John@somewhere.com&lt;\/rdf:value&gt;\n      &lt;\/vcard:internet&gt;\n    &lt;\/vcard:EMAIL&gt;\n    &lt;vcard:N rdf:parseType=\"Resource\"&gt;\n      &lt;vcard:Given&gt;John&lt;\/vcard:Given&gt;\n      &lt;vcard:Family&gt;Smith&lt;\/vcard:Family&gt;\n    &lt;\/vcard:N&gt;\n    &lt;vcard:FN&gt;John Smith&lt;\/vcard:FN&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;<\/pre>\n  <pre> <\/pre>\n<\/blockquote><p>Even if you are unfamiliar with the details of the RDF\/XML syntax, it\nshould be reasonably clear that the Models have merged as expected.  The\nintersection and difference of the Models can be computed in a similar\nmanner, using the methods <code>.intersection(Model)<\/code> and\n<code>.difference(Model)<\/code>; see the\n<a href=\"https:\/\/jena.apache.org\/documentation\/javadoc\/jena\/org\/apache\/jena\/rdf\/model\/Model.html#difference(org.apache.jena.rdf.model.Model)\">difference<\/a>\nand\n<a href=\"https:\/\/jena.apache.org\/documentation\/javadoc\/jena\/org\/apache\/jena\/rdf\/model\/Model.html#intersection(org.apache.jena.rdf.model.Model)\">intersection<\/a>\nJavadocs for more details.\n<\/p><h2 id=\"ch-Containers\">Containers<\/h2><p>RDF defines a special kind of resources for representing collections of\nthings. These resources are called <i>containers<\/i>. The members of a\ncontainer can be either literals or resources. There are three kinds of\ncontainer:<\/p><ul><li>a BAG is an unordered collection<\/li>\n  <li>an ALT is an unordered collection intended to represent\n  alternatives<\/li>\n  <li>a SEQ is an ordered collection<\/li>\n<\/ul><p>A container is represented by a resource. That resource will have an\nrdf:type property whose value should be one of rdf:Bag, rdf:Alt or rdf:Seq,\nor a subclass of one of these, depending on the type of the container.\nThe first member of the container is the value of the container's rdf:_1\nproperty; the second member of the container is the value of the container's\nrdf:_2 property and so on. The rdf:_nnn properties are known as the\n<i>ordinal properties<\/i>.<\/p><p>For example, the Model for a simple bag containing the vcards of the\nSmith's might look like this:<\/p><p style=\"text-align: center;\">\n<img border=\"0\" src=\"https:\/\/jena.apache.org\/tutorials\/figures\/fig3.png\" alt=\"figure 3\" width=\"720\" height=\"420\" \/><\/p><p style=\"text-align: left;\">Whilst the members of the bag are represented by the\nproperties rdf:_1, rdf:_2 etc the ordering of the properties is not\nsignificant. We could switch the values of the rdf:_1 and rdf:_2\nproperties and the resulting Model would represent the same information.<\/p><p style=\"text-align: left;\">Alt's are intended to represent alternatives. For\nexample, lets say a resource represented a software product. It might\nhave a property to indicate where it might be obtained from. The value\nof that property might be an Alt collection containing various sites from\nwhich it could be downloaded. Alt's are unordered except that the\nrdf:_1 property has special significance. It represents the default\nchoice.<\/p><p style=\"text-align: left;\">Whilst containers can be handled using the basic machinery of\nresources and properties, Jena has explicit interfaces and implementation\nclasses to handle them. It is not a good idea to have an object\nmanipulating a container, and at the same time to modify the state of that\ncontainer using the lower level methods.<\/p><p style=\"text-align: left;\">Let's modify tutorial 8 to create this bag:<\/p><blockquote>\n  <pre>\/\/ create a bag\nBag smiths = model.createBag();\n\/\/ select all the resources with a VCARD.FN property\n\/\/ whose value ends with \"Smith\"\nStmtIterator iter = model.listStatements(\n    new SimpleSelector(null, VCARD.FN, (RDFNode) null) {\n        public boolean selects(Statement s) {\n                return s.getString().endsWith(\"Smith\");\n        }\n    });\n\/\/ add the Smith's to the bag\nwhile (iter.hasNext()) {\n    smiths.add(iter.nextStatement().getSubject());\n}<\/pre>\n<\/blockquote><p>If we write out this Model, it contains something like the following:<\/p><blockquote>\n  <pre>&lt;rdf:RDF\n  xmlns:rdf='http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#'\n  xmlns:vcard='http:\/\/www.w3.org\/2001\/vcard-rdf\/3.0#'\n &gt;\n...\n  &lt;rdf:Description rdf:nodeID=\"A3\"&gt;\n    &lt;rdf:type rdf:resource='http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#Bag'\/&gt;\n    &lt;rdf:_1 rdf:resource='http:\/\/somewhere\/JohnSmith\/'\/&gt;\n    &lt;rdf:_2 rdf:resource='http:\/\/somewhere\/RebeccaSmith\/'\/&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;<\/pre>\n<\/blockquote><p>which represents the Bag resource.<\/p><p>The container interface provides an iterator to list the contents of a\ncontainer:<\/p><blockquote>\n  <pre>\/\/ print out the members of the bag\nNodeIterator iter2 = smiths.iterator();\nif (iter2.hasNext()) {\n    System.out.println(\"The bag contains:\");\n    while (iter2.hasNext()) {\n        System.out.println(\"  \" +\n            ((Resource) iter2.next())\n                            .getProperty(VCARD.FN)\n                            .getString());\n    }\n} else {\n    System.out.println(\"The bag is empty\");\n}<\/pre>\n<\/blockquote><p>which produces the following output:<\/p><blockquote>\n  <pre>The bag contains:\n  John Smith\n  Becky Smith<\/pre>\n<\/blockquote><p>Executable example code can be found in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial10.java\">\ntutorial 10<\/a>, which glues together the fragments above into a complete\nexample.<\/p><p>The Jena classes offer methods for manipulating containers including\nadding new members, inserting new members into the middle of a container and\nremoving existing members.  The Jena container classes currently ensure that\nthe list of ordinal properties used starts at rdf:_1 and is contiguous.\nThe RDFCore WG have relaxed this constraint, which allows partial\nrepresentation of containers.  This therefore is an area of Jena may be\nchanged in the future.<\/p><p>RDF literals are not just simple strings.  Literals may have a language\ntag to indicate the language of the literal.  The literal \"chat\" with an\nEnglish language tag is considered different to the literal \"chat\" with a\nFrench language tag.  This rather strange behaviour is an artefact of the\noriginal RDF\/XML syntax.<\/p><p>Further there are really two sorts of Literals.  In one, the string\ncomponent is just that, an ordinary string.  In the other the string\ncomponent is expected to be a well balanced fragment of XML.  When an RDF\nModel is written as RDF\/XML a special construction using a\nparseType='Literal' attribute is used to represent it.<\/p><p>In Jena, these attributes of a literal may be set when the literal is\nconstructed, e.g. in <a href=\"https:\/\/github.com\/apache\/jena\/tree\/master\/jena-core\/src-examples\/jena\/examples\/rdf\/Tutorial11.java\">tutorial 11<\/a>:<\/p><blockquote>\n  <pre>\/\/ create the resource\nResource r = model.createResource();\n\/\/ add the property\nr.addProperty(RDFS.label, model.createLiteral(\"chat\", \"en\"))\n .addProperty(RDFS.label, model.createLiteral(\"chat\", \"fr\"))\n .addProperty(RDFS.label, model.createLiteral(\"&lt;em&gt;chat&lt;\/em&gt;\", true));\n\/\/ write out the Model\nmodel.write(system.out);<\/pre>\n<\/blockquote><p>produces<\/p><blockquote>\n  <pre>&lt;rdf:RDF\n  xmlns:rdf='http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#'\n  xmlns:rdfs='http:\/\/www.w3.org\/2000\/01\/rdf-schema#'\n &gt;\n  &lt;rdf:Description rdf:nodeID=\"A0\"&gt;\n    &lt;rdfs:label xml:lang='en'&gt;chat&lt;\/rdfs:label&gt;\n    &lt;rdfs:label xml:lang='fr'&gt;chat&lt;\/rdfs:label&gt;\n    &lt;rdfs:label rdf:parseType='Literal'&gt;&lt;em&gt;chat&lt;\/em&gt;&lt;\/rdfs:label&gt;\n  &lt;\/rdf:Description&gt;\n&lt;\/rdf:RDF&gt;<\/pre>\n<\/blockquote><p>For two literals to be considered equal, they must either both be XML\nliterals or both be simple literals.  In addition, either both must have no\nlanguage tag, or if language tags are present they must be equal.  For simple\nliterals the strings must be equal.  XML literals have two notions of\nequality.  The simple notion is that the conditions previously mentioned are\ntrue and the strings are also equal.  The other notion is that they can be\nequal if the canonicalization of their strings is equal.<\/p><p>Jena's interfaces also support typed literals. The old-fashioned way\n(shown below) treats typed literals as shorthand for strings: typed\nvalues are converted in the usual Java way to strings and these strings\nare stored in the Model. For example, try (noting that for\nsimple literals, we can omit the <code>model.createLiteral<\/code>(...)\ncall):<\/p><blockquote>\n  <pre>\/\/ create the resource\nResource r = model.createResource();\n\/\/ add the property\nr.addProperty(RDFS.label, \"11\")\n .addProperty(RDFS.label, 11);\n\/\/ write out the Model\nmodel.write(system.out, \"N-TRIPLE\");<\/pre>\n<\/blockquote><p>The output produced is:<\/p><blockquote>\n  <pre>_:A... &lt;http:\/\/www.w3.org\/2000\/01\/rdf-schema#label&gt; \"11\" .<\/pre>\n<\/blockquote><p>Since both literals are really just the string \"11\", then only one\nstatement is added.<\/p><p>The RDFCore WG has defined mechanisms for supporting datatypes in RDF.\nJena supports these using the <i>typed literal<\/i> mechanisms; they are\nnot discussed in this tutorial.<\/p><h2 id=\"ch-Glossary\">Glossary<\/h2><dl><dt><a>Blank Node<\/a><\/dt>\n<dd>Represents a resource, but does not indicate a URI for the\nresource.  Blank nodes act like existentially qualified variables in first\norder logic.<\/dd>\n<dt><a>Dublin Core<\/a><\/dt>\n<dd>A standard for metadata about web resources. Further\ninformation can be found at the <a href=\"http:\/\/purl.oclc.org\/dc\/\">Dublin Core\nweb site<\/a>.<\/dd>\n<dt><a id=\"glos-Literal\">Literal<\/a><\/dt>\n<dd>A string of characters which can be the value of a property.<\/dd>\n<dt><a id=\"glos-Object\">Object<\/a><\/dt>\n<dd>The part of a triple which is the value of the statement.<\/dd>\n<dt><a id=\"glos-Predicate\">Predicate<\/a><\/dt>\n<dd>The property part of a triple.<\/dd>\n<dt><a id=\"glos-Property\">Property<\/a><\/dt>\n<dd>A property is an attribute of a resource. For example\nDC.title is a property, as is RDF.type.\n<\/dd><dt><a id=\"glos-Resource\">Resource<\/a><\/dt>\n<dd>Some entity. It could be a web resource such as web page, or\nit could be a concrete physical thing such as a tree or a car. It could be an\nabstract idea such as chess or football. Resources are named by URI's.<\/dd>\n<dt><a id=\"glos-Statement\">Statement<\/a><\/dt>\n<dd>An arc in an RDF Model, normally interpreted as a fact.<\/dd>\n<dt><a id=\"glos-Subject\">Subject<\/a><\/dt>\n<dd>The resource which is the source of an arc in an RDF Model<\/dd>\n<dt><a id=\"glos-Triple\">Triple<\/a><\/dt>\n<dd>A structure containing a subject, a predicate and an object.\nAnother term for a statement.<\/dd>\n<\/dl><h2>Footnotes<\/h2><ol><li>The identifier of an RDF resource can\n    include a fragment identifier, e.g.\n    http:\/\/hostname\/rdf\/tutorial\/#ch-Introduction, so, strictly speaking, an\n    RDF resource is identified by a URI reference.<\/li>\n  <li>As well as being a string of characters,\n    literals also have an optional language encoding to represent the\n    language of the string. For example the literal \"two\" might have a\n    language encoding of \"en\" for English and the literal \"deux\" might have a\n    language encoding of \"fr\" for France.<\/li>\n<\/ol>[1]: sparql_data\/vc-db-1.rdf","created_at":"2019-04-17T03:41:14+0000","updated_at":"2019-04-17T03:41:24+0000","annotations":[],"mimetype":"text\/html","reading_time":34,"domain_name":"jena.apache.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","metaprogramming"],"is_public":false,"id":911,"title":"PHP like a Boss: The Magic of Metaprogramming & Reflection - Part 1 - Goodheads","url":"http:\/\/goodheads.io\/2016\/03\/13\/php-like-a-boss-the-magic-of-metaprogramming-reflection-part-1\/","content":"<p><em>PHP is a Shitty language.<\/em><\/p><p><em>PHP is garbage.<\/em><\/p>\n<p><em>PHP is f**ked up!<\/em><\/p>\n<p>I have been constantly molested verbally by developers of other languages like Python and Ruby that think PHP is just outrightly stupid but as a PHP Ninja that I am, I know the good sides, I know the dark sides, I have also explored the ugly sides and overtime I have known how to wield superpowers in PHP that can do what these other languages can &amp; also better.<\/p>\n<p>Metaprogramming is the ability of a program\/code to be able to create another piece of code dynamically at runtime and also manipulate itself if need be. Ikem has given a good explanation about it in his post: <a href=\"https:\/\/goodheads-prosper.rhcloud.com\/2016\/02\/27\/ruby-on-steroids-the-magic-of-metaprogramming-part-1\/\">Ruby on Steroids: The Magic of MetaProgramming \u2013 An Unexpected Journey<\/a>\u00a0so you can actually just check that out. So much ado about Metaprogramming but PHP can handle this technique pretty well too.<\/p>\n<p>According to <em><strong>Serge Smerting<\/strong><\/em>, Metaprogramming :<\/p>\n<ul><li>is how programs are self-aware<\/li>\n<li>is the art of operating metadata<\/li>\n<li>is black magic processed at build step<\/li>\n<li>is making the code shorter<\/li>\n<\/ul><p>Let\u2019s get started with some of the techniques:<\/p>\n<p><strong>1. eval<\/strong><\/p>\n<p>There is a powerful PHP function that can evaluate a piece of PHP code represented in string format\u00a0which means that it is possible for one to accept user input or read code off of a file and get it evaluated from within your program.<\/p>\n\t\t<div id=\"crayon-59654f99d78c6712553055\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\neval(\"function getName() { echo 'this is prosper'; } getName();\");\n#=&gt; this is prosper<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78c6712553055-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78c6712553055-2\">2<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78c6712553055-1\">eval(\"function getName() { echo 'this is prosper'; } getName();\");<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78c6712553055-2\">#=&gt; this is prosper<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>Let\u2019s change <code>getName<\/code> to <code>getSomething<\/code>, now that\u2019s a function that doesn\u2019t exist so it should throw an error:<\/p>\n\t\t<div id=\"crayon-59654f99d78d5593105515\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\neval(\"function getName() { echo 'this is prosper'; } getSomething();\");\n#=&gt; Fatal error: Call to undefined function getSomething() in \/Users\/unicodeveloper\/source\/php-workspace\/reflection\/index.php(9) : eval()'d code on line 1<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78d5593105515-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78d5593105515-2\">2<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78d5593105515-1\">eval(\"function getName() { echo 'this is prosper'; } getSomething();\");<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78d5593105515-2\">#=&gt; Fatal error: Call to undefined function getSomething() in \/Users\/unicodeveloper\/source\/php-workspace\/reflection\/index.php(9) : eval()'d code on line 1<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>A typical real world example is WordPress plugins and online code editors. These editors allow you write PHP code and it processes that code to give you the desired output.<\/p>\n<p><strong>Tip:<\/strong> You can use <strong>file_get_contents()<\/strong> to read off source code from a file and pass it onto the <strong>eval()<\/strong> function to process.<\/p>\n<p>This <strong>eval()<\/strong>\u00a0function is very dangerous because it allows running of any form of PHP code, so be very careful and properly validate any user input that is passed onto this function.<\/p>\n<p><strong>2. __set and __get magic methods<\/strong><\/p>\n<p>These methods enable you to dynamically create and retrieve properties in a class that never existed during runtime. <strong>__get<\/strong> is used to read the value of an undefined property of a class while<strong> __set<\/strong> is used to set the value of a new propery of a class at runtime. Let\u2019s get to the code for better understanding.<\/p>\n\t\t<div id=\"crayon-59654f99d78df363404522\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Student\n{\n    public $information = array();\n    public $school = 'Babcock University';\n    public $country = 'Nigeria';\n    public function __get($item) {\n        return $this-&gt;information[$item];\n    }\n    public function __set($item, $value) {\n      $this-&gt;information[$item] = $value;\n    }\n}\n$student = new Student;\n$student-&gt;name = 'Prosper Otemuyiwa';\n$student-&gt;email = 'prosper@goodheads.io';\necho $student-&gt;name; #=&gt; Prosper Otemuyiwa\necho $student-&gt;email; #=&gt; prosper@goodheads.io\necho $student-&gt;country; #=&gt; Nigeria<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-17\">17<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-18\">18<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-19\">19<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-20\">20<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-21\">21<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78df363404522-22\">22<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78df363404522-23\">23<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-1\">class Student<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-3\">\u00a0\u00a0\u00a0\u00a0public $information = array();<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-4\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-5\">\u00a0\u00a0\u00a0\u00a0public $school = 'Babcock University';<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-6\">\u00a0\u00a0\u00a0\u00a0public $country = 'Nigeria';<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-7\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-8\">\u00a0\u00a0\u00a0\u00a0public function __get($item) {<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-9\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return $this-&gt;information[$item];<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-10\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-11\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-12\">\u00a0\u00a0\u00a0\u00a0public function __set($item, $value) {<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-13\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$this-&gt;information[$item] = $value;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-14\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-15\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-16\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-17\">$student = new Student;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-18\">$student-&gt;name = 'Prosper Otemuyiwa';<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-19\">$student-&gt;email = 'prosper@goodheads.io';<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-20\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-21\">echo $student-&gt;name; #=&gt; Prosper Otemuyiwa<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78df363404522-22\">echo $student-&gt;email; #=&gt; prosper@goodheads.io<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78df363404522-23\">echo $student-&gt;country; #=&gt; Nigeria<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>The <code>name<\/code> and <code>email<\/code> properties don\u2019t exist in the class, We created them on the fly, yet we were able to assign values to them and also retrieve those values. How sweet! <img src=\"http:\/\/goodheads.io\/wp-includes\/images\/smilies\/simple-smile.png\" alt=\":smile:\" class=\"wp-smiley\" \/><\/p>\n<p>In a Real World, A very good use case for this are ORMs, In a typical ORM you can do something like:<\/p>\n\t\t<div id=\"crayon-59654f99d78e8039250090\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\n&lt;?php\n$user = new User;\n$user-&gt;name = 'unicodeveloper';\n$user-&gt;bestFood = 'Bread and beans';\n$user-&gt;save();\n#=&gt; This saves the information in the name and bestFood columns of the user table<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78e8039250090-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78e8039250090-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78e8039250090-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78e8039250090-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78e8039250090-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78e8039250090-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78e8039250090-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78e8039250090-8\">8<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78e8039250090-1\">&lt;?php<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78e8039250090-2\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78e8039250090-3\">$user = new User;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78e8039250090-4\">$user-&gt;name = 'unicodeveloper';<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78e8039250090-5\">$user-&gt;bestFood = 'Bread and beans';<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78e8039250090-6\">$user-&gt;save();<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78e8039250090-7\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78e8039250090-8\">#=&gt; This saves the information in the name and bestFood columns of the user table<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>The underlying code makes use of these magic methods to be able to dynamically set any property for the model and check if the names of those properties correspond to column names in the table,\u2026if they do, it saves successfully, if they don\u2019t it throws an appropriate exception. Same applies for also retrieving data from the database.<\/p>\n<p>3. <strong>__invoke<\/strong><\/p>\n<p>This magic method allows you to call an object as if it were a function or even as a closure. How can something be an object and also a function?. Magic right? A trivial example is:<\/p>\n\t\t<div id=\"crayon-59654f99d78ed563428873\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Student\n{\n    public function __invoke()\n    {\n        echo \"This is a student\";\n    }\n}\n$student = new Student;\n$student(); #=&gt; This is a student<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78ed563428873-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78ed563428873-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78ed563428873-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78ed563428873-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78ed563428873-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78ed563428873-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78ed563428873-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78ed563428873-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78ed563428873-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78ed563428873-10\">10<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78ed563428873-1\">class Student<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78ed563428873-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78ed563428873-3\">\u00a0\u00a0\u00a0\u00a0public function __invoke()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78ed563428873-4\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78ed563428873-5\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0echo \"This is a student\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78ed563428873-6\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78ed563428873-7\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78ed563428873-8\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78ed563428873-9\">$student = new Student;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78ed563428873-10\">$student(); #=&gt; This is a student<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>So the object was invoked as a function. This magic method came with PHP 5.3<\/p>\n<p>A practical use case of this is:<\/p>\n\t\t<div id=\"crayon-59654f99d78f2008750933\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Student\n{\n    public function __invoke()\n    {\n        echo \"I am a student\";\n    }\n}\nClass College\n{\n    public static function attends(callable $callback)\n    {\n        $callback();\n        echo \" I attend Rutherford's College\";\n    }\n}\n$student = new Student();\nCollege::attends($student); #=&gt; I am a student I attend Rutherford's College<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-17\">17<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f2008750933-18\">18<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f2008750933-19\">19<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-1\">class Student<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-3\">\u00a0\u00a0\u00a0\u00a0public function __invoke()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-4\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-5\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0echo \"I am a student\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-6\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-7\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-8\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-9\">Class College<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-10\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-11\">\u00a0\u00a0\u00a0\u00a0public static function attends(callable $callback)<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-12\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-13\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$callback();<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-14\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0echo \" I attend Rutherford's College\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-15\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-16\">}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-17\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f2008750933-18\">$student = new Student();<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f2008750933-19\">College::attends($student); #=&gt; I am a student I attend Rutherford's College<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>The object was invoked as a callback into <code>attends<\/code> method of the <code>College<\/code> class.<\/p>\n<p><strong>4. __toString()<\/strong><\/p>\n<p>This magic method allows us to change the default behaviour of an object when it is been echoed. It returns a string representation. Example:<\/p>\n\t\t<div id=\"crayon-59654f99d78f7133899907\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div><\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Student\n{\n    private $name;\n    public function __construct($name)\n    {\n        $this-&gt;name = $name;\n    }\n    public function __toString()\n    {\n        return \"My name is {$this-&gt;name}. I am a Student\";\n    }\n}\n$student = new Student('Prosper Otemuyiwa');\necho $student; #=&gt; My name is Prosper Otemuyiwa. I am a Student<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78f7133899907-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78f7133899907-17\">17<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-1\">class Student<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-3\">\u00a0\u00a0\u00a0\u00a0private $name;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-4\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-5\">\u00a0\u00a0\u00a0\u00a0public function __construct($name)<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-6\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-7\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$this-&gt;name = $name;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-8\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-9\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-10\">\u00a0\u00a0\u00a0\u00a0public function __toString()<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-11\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-12\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"My name is {$this-&gt;name}. I am a Student\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-13\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-14\">}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-15\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78f7133899907-16\">$student = new Student('Prosper Otemuyiwa');<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78f7133899907-17\">echo $student; #=&gt; My name is Prosper Otemuyiwa. I am a Student<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>Here we are echoing the object directly as a string. It\u2019s a very nice shortcut that can shorten the amount of lines of code in your project.<\/p>\n<p><strong>5. __call and __callStatic<\/strong><\/p>\n<p>With these methods, you can handle situations when methods that don\u2019t exist\/publicly accessible are called on an object. Example:<\/p>\n\t\t<div id=\"crayon-59654f99d78fc512804333\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Person\n{\n    public function eat()\n    {\n        return \"I can eat\";\n    }\n    public function drive()\n    {\n        return \"I can drive\";\n    }\n    public function sleep()\n    {\n        return \"I can sleep\";\n    }\n}\n$person = new Person();\n$person-&gt;sleep(); #=&gt; I can sleep;\n$person-&gt;missisipi(); #=&gt; Fatal error: Call to undefined method UnicodeveloperSeasonPerson::missisipi() in \/Users\/unicodeveloper\/source\/php-workspace\/reflection\/index.php on line 12<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-17\">17<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-18\">18<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-19\">19<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d78fc512804333-20\">20<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d78fc512804333-21\">21<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-1\">class Person<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-3\">\u00a0\u00a0\u00a0\u00a0public function eat()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-4\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-5\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can eat\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-6\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-7\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-8\">\u00a0\u00a0\u00a0\u00a0public function drive()<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-9\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-10\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can drive\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-11\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-12\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-13\">\u00a0\u00a0\u00a0\u00a0public function sleep()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-14\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-15\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can sleep\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-16\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-17\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-18\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-19\">$person = new Person();<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d78fc512804333-20\">$person-&gt;sleep(); #=&gt; I can sleep;<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d78fc512804333-21\">$person-&gt;missisipi(); #=&gt; Fatal error: Call to undefined method UnicodeveloperSeasonPerson::missisipi() in \/Users\/unicodeveloper\/source\/php-workspace\/reflection\/index.php on line 12<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>Calling a missisipi method that doesn\u2019t exist in that class throws a FATAL error<\/p>\n<p>Now, let\u2019s take care of methods that don\u2019t exist. Example:<\/p>\n\t\t<div id=\"crayon-59654f99d7901799645233\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Person\n{\n    public function eat()\n    {\n        return \"I can eat\";\n    }\n    public function drive()\n    {\n        return \"I can drive\";\n    }\n    public function sleep()\n    {\n        return \"I can sleep\";\n    }\n    public function __call($name, $args)\n    {\n        if (method_exists($this, $name)) {\n            return $this-&gt;$name();\n        }\n        if ($name === 'destroy') {\n            return \"You can't call this method on this Class. Too dangerous my friend\";\n        }\n        $message = \"Hey, you just called the {$name} method but unfortunately there is nothing like that.\";\n    }\n}\n$person = new Person;\n$person-&gt;destroy(); #=&gt; You can't call this method on this Class. Too dangerous my friend\n$person-&gt;missisipi(); #=&gt; Hey, you just called the missisipi method but unfortunately there is nothing like that.<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-17\">17<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-18\">18<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-19\">19<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-20\">20<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-21\">21<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-22\">22<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-23\">23<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-24\">24<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-25\">25<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-26\">26<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-27\">27<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-28\">28<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-29\">29<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-30\">30<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-31\">31<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-32\">32<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7901799645233-33\">33<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7901799645233-34\">34<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-1\">class Person<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-3\">\u00a0\u00a0\u00a0\u00a0public function eat()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-4\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-5\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can eat\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-6\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-7\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-8\">\u00a0\u00a0\u00a0\u00a0public function drive()<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-9\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-10\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can drive\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-11\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-12\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-13\">\u00a0\u00a0\u00a0\u00a0public function sleep()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-14\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-15\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can sleep\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-16\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-17\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-18\">\u00a0\u00a0\u00a0\u00a0public function __call($name, $args)<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-19\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-20\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0if (method_exists($this, $name)) {<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-21\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return $this-&gt;$name();<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-22\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-23\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-24\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0if ($name === 'destroy') {<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-25\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"You can't call this method on this Class. Too dangerous my friend\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-26\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-27\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-28\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$message = \"Hey, you just called the {$name} method but unfortunately there is nothing like that.\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-29\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-30\">}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-31\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-32\">$person = new Person;<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7901799645233-33\">$person-&gt;destroy(); #=&gt; You can't call this method on this Class. Too dangerous my friend<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7901799645233-34\">$person-&gt;missisipi(); #=&gt; Hey, you just called the missisipi method but unfortunately there is nothing like that.<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>Here, in the <strong>__call<\/strong> method, we are explicitly checking with <strong>method_exists<\/strong> if the method that is called on the object exists, if it does, go ahead and call the method.<\/p>\n<p>If it was a <strong>destroy()<\/strong> method that was called, it should print out <code>You can't call this method on this Class. Too dangerous my friend.<\/code><\/p>\n<p>If the method doesn\u2019t exist in that class at all, then report with a custom message that there is no method like that.<\/p>\n<p><strong>__callStatic<\/strong> is used for static functions like so:<\/p>\n\t\t<div id=\"crayon-59654f99d7906415680337\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Person\n{\n    public static function eat()\n    {\n        return \"I can eat\";\n    }\n    public static function drive()\n    {\n        return \"I can drive\";\n    }\n    public static function sleep()\n    {\n        return \"I can sleep\";\n    }\n    public static function __callStatic($name, $args)\n    {\n        if (method_exists(new static, $name)) {\n            return $this-&gt;$name();\n        }\n        if ($name === 'destroy') {\n            return \"You can't call this method on this Class. Too dangerous my friend\";\n        }\n        $message = \"Hey, you just called the {$name} method but unfortunately there is nothing like that.\";\n        return $message;\n    }\n}\nPerson::destroy(); #=&gt; You can't call this method on this Class. Too dangerous my friend\nPerson::missisipi(); #=&gt; Hey, you just called the missisipi method but unfortunately there is nothing like that.<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-17\">17<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-18\">18<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-19\">19<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-20\">20<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-21\">21<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-22\">22<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-23\">23<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-24\">24<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-25\">25<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-26\">26<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-27\">27<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-28\">28<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-29\">29<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-30\">30<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-31\">31<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-32\">32<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7906415680337-33\">33<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7906415680337-34\">34<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-1\">class Person<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-3\">\u00a0\u00a0\u00a0\u00a0public static function eat()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-4\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-5\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can eat\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-6\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-7\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-8\">\u00a0\u00a0\u00a0\u00a0public static function drive()<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-9\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-10\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can drive\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-11\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-12\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-13\">\u00a0\u00a0\u00a0\u00a0public static function sleep()<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-14\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-15\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"I can sleep\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-16\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-17\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-18\">\u00a0\u00a0\u00a0\u00a0public static function __callStatic($name, $args)<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-19\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-20\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0if (method_exists(new static, $name)) {<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-21\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return $this-&gt;$name();<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-22\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-23\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-24\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0if ($name === 'destroy') {<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-25\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return \"You can't call this method on this Class. Too dangerous my friend\";<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-26\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-27\">\u00a0<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-28\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$message = \"Hey, you just called the {$name} method but unfortunately there is nothing like that.\";<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-29\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return $message;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-30\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-31\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-32\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7906415680337-33\">Person::destroy(); #=&gt; You can't call this method on this Class. Too dangerous my friend<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7906415680337-34\">Person::missisipi(); #=&gt; Hey, you just called the missisipi method but unfortunately there is nothing like that.<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p><strong>6. call_user_func and call_user_func_array<\/strong><\/p>\n<p>Sometimes you are not sure what methods would be called, in your case, a method call might be determined by a user\u2019s input..how do you handle that dynamically?<\/p>\n<p>So <strong>call_user_func<\/strong> calls functions whose name you don\u2019t know ahead of time.<\/p>\n\t\t<div id=\"crayon-59654f99d790b032016122\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nclass Arithmetic\n{\n    public static function operations(callable $callback)\n    {\n        return call_user_func($callback);\n    }\n}\n$number = 7;\n$multResult = Arithmetic::operations(function() use ($number) {\n    return $number * $number;\n});\n$addResult = Arithmetic::operations(function() use ($number) {\n    return $number + $number;\n});\necho $multResult; #=&gt; 49\necho $addResult; #=&gt; 14<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-5\">5<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-6\">6<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-7\">7<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-8\">8<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-9\">9<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-10\">10<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-11\">11<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-12\">12<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-13\">13<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-14\">14<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-15\">15<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-16\">16<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-17\">17<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-18\">18<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d790b032016122-19\">19<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d790b032016122-20\">20<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-1\">class Arithmetic<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-2\">{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-3\">\u00a0\u00a0\u00a0\u00a0public static function operations(callable $callback)<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-4\">\u00a0\u00a0\u00a0\u00a0{<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-5\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0return call_user_func($callback);<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-6\">\u00a0\u00a0\u00a0\u00a0}<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-7\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-8\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-9\">$number = 7;<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-10\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-11\">$multResult = Arithmetic::operations(function() use ($number) {<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-12\">\u00a0\u00a0\u00a0\u00a0return $number * $number;<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-13\">});<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-14\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-15\">$addResult = Arithmetic::operations(function() use ($number) {<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-16\">\u00a0\u00a0\u00a0\u00a0return $number + $number;<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-17\">});<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-18\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d790b032016122-19\">echo $multResult; #=&gt; 49<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d790b032016122-20\">echo $addResult; #=&gt; 14<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n<p>Here we passed callbacks ( anonymous functions), one for addition and another for multiplication and it performed those operations and gave us results because <strong>call_user_func<\/strong> was able to call those functions.<\/p>\n<p><strong><em>call_user_func_array()<\/em><\/strong> is very helpful for functions that have complex and variable parameter requirements like so:<\/p>\n\t\t<div id=\"crayon-59654f99d7910696345474\" class=\"crayon-syntax crayon-theme-classic crayon-font-monaco crayon-os-pc print-yes notranslate\" data-settings=\"minimize scroll-mouseover\">\n\t\t\t<div class=\"crayon-toolbar\" data-settings=\"mouseover overlay hide delay\">\n\t\t\t<div class=\"crayon-tools\"><div class=\"crayon-button crayon-nums-button\" title=\"Toggle Line Numbers\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-plain-button\" title=\"Toggle Plain Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-wrap-button\" title=\"Toggle Line Wrap\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-expand-button\" title=\"Expand Code\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-copy-button\" title=\"Copy\"><div class=\"crayon-button-icon\"><\/div><div class=\"crayon-button crayon-popup-button\" title=\"Open Code In New Window\"><div class=\"crayon-button-icon\"><\/div>PHP<\/div><\/div>\n\t\t\t<div class=\"crayon-info\">\n\t\t\t<div class=\"crayon-plain-wrap\"><textarea wrap=\"soft\" class=\"crayon-plain print-no\" data-settings=\"dblclick\" readonly=\"readonly\" rows=\"10\" cols=\"50\">\nfunction add($num, $num2) {\n    echo  $num + $num2;\n}\ncall_user_func_array(\"add\", array(1,2));<\/textarea><\/div>\n\t\t\t<div class=\"crayon-main\">\n\t\t\t\t<table class=\"crayon-table\"><tr class=\"crayon-row\"><td class=\"crayon-nums\" data-settings=\"show\">\n\t\t\t\t\t<div class=\"crayon-nums-content\"><div class=\"crayon-num\" data-line=\"crayon-59654f99d7910696345474-1\">1<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7910696345474-2\">2<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7910696345474-3\">3<\/div><div class=\"crayon-num crayon-striped-num\" data-line=\"crayon-59654f99d7910696345474-4\">4<\/div><div class=\"crayon-num\" data-line=\"crayon-59654f99d7910696345474-5\">5<\/div><\/div>\n\t\t\t\t<\/td>\n\t\t\t\t\t\t<td class=\"crayon-code\"><div class=\"crayon-pre\"><div class=\"crayon-line\" id=\"crayon-59654f99d7910696345474-1\">function add($num, $num2) {<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7910696345474-2\">\u00a0\u00a0\u00a0\u00a0echo\u00a0\u00a0$num + $num2;<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7910696345474-3\">}<\/div><div class=\"crayon-line crayon-striped-line\" id=\"crayon-59654f99d7910696345474-4\">\u00a0<\/div><div class=\"crayon-line\" id=\"crayon-59654f99d7910696345474-5\">call_user_func_array(\"add\", array(1,2));<\/div><\/div><\/td>\n\t\t\t\t\t<\/tr><\/table><\/div>\n\t\t<\/div>\n\n\n<p><strong>Conclusion<\/strong><\/p>\n<p>You have seen some of the magic you can perform on methods and how you can manipulate your programs during runtime. You have also seen how to use code to create code on the fly. We\u2019ll be talking about the Reflection Class and it\u2019s sister classes in the next part of this series.<\/p>\n<p>You\u2019ll receive more POWER as a PHP developer! <img src=\"http:\/\/goodheads.io\/wp-includes\/images\/smilies\/simple-smile.png\" alt=\":smile:\" class=\"wp-smiley\" \/><\/p>\n<p>Please let me know if you have any questions or observations in the comments section.<\/p>\n<div id=\"shr_canvas2\" class=\"shareaholic-canvas shareaholic-ui shareaholic-resolved-canvas ng-scope\" data-app-id=\"24010768\" data-app=\"share_buttons\" data-title=\"Ruby on Steroids: The Magic of MetaProgramming - Method Spells\" data-link=\"https:\/\/goodheads-prosper.rhcloud.com\/2016\/03\/12\/ruby-on-steroids-the-magic-of-metaprogramming-method-spells\/\" data-summary=\"\">\n\n<div class=\"shareaholic-canvas\" data-app-id=\"24052053\" data-app=\"share_buttons\" data-title=\"PHP like a Boss: The Magic of Metaprogramming &amp;amp; Reflection - Part 1\" data-link=\"http:\/\/goodheads.io\/2016\/03\/13\/php-like-a-boss-the-magic-of-metaprogramming-reflection-part-1\/\" data-summary=\"\"><div class=\"shareaholic-canvas\" data-app-id=\"24052061\" data-app=\"recommendations\" data-title=\"PHP like a Boss: The Magic of Metaprogramming &amp;amp; Reflection - Part 1\" data-link=\"http:\/\/goodheads.io\/2016\/03\/13\/php-like-a-boss-the-magic-of-metaprogramming-reflection-part-1\/\" data-summary=\"\"><div id=\"wpautbox-below\"><ul class=\"a-tab-nav\"><li class=\"a-tab-active\"><a href=\"#wpautbox_about\"><i class=\"el-icon-user wpautbox-icon\"> About the Author<\/i><\/a><\/li><li><a href=\"#wpautbox_latest-post\"><i class=\"el-icon-list wpautbox-icon\"> Latest Posts<\/i><\/a><\/li><\/ul><div class=\"a-tab-container\"><div class=\"a-tab-content\" id=\"wpautbox_about\"><div class=\"wpautbox-avatar\"><img alt=\"PROSPER OTEMUYIWA\" src=\"http:\/\/goodheads.io\/wp-content\/uploads\/2016\/03\/PROSPER-OTEMUYIWA_avatar_1458048032-80x80.png\" class=\"avatar avatar-80 photo\" height=\"80\" width=\"80\" \/><\/div><div class=\"wpautbox-author-meta\"><h4 class=\"wpautbox-name\">About PROSPER OTEMUYIWA<\/h4>Food Ninja, Code Slinger, Technical Trainer, Accidental Writer, Open Source Advocate and Developer Evangelist.<\/div><\/div><div class=\"a-tab-content\" id=\"wpautbox_latest-post\"><ul class=\"wpautbox-post_type-list wpautbox-latest-post\"><li><a href=\"http:\/\/goodheads.io\/2016\/08\/01\/build-youtube-part-10\/\">How to build your own Youtube \u2013 Part 10<\/a> - August 1, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/07\/25\/build-youtube-part-9\/\">How to build your own Youtube \u2013 Part 9<\/a> - July 25, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/07\/23\/build-youtube-part-8\/\">How to build your own Youtube \u2013 Part 8<\/a> - July 23, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/07\/06\/build-youtube-part-6\/\">How to build your own Youtube \u2013 Part 6<\/a> - July 6, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/07\/03\/introducing-laravel-password-v1-0\/\">Introducing Laravel Password v1.0<\/a> - July 3, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/06\/28\/build-youtube-part-5\/\">How to build your own Youtube \u2013 Part 5<\/a> - June 28, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/06\/23\/build-youtube-part-4\/\">How to build your own Youtube \u2013 Part 4<\/a> - June 23, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/06\/15\/build-youtube-part-3\/\">How to build your own Youtube \u2013 Part 3<\/a> - June 15, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/06\/08\/build-youtube-part-2\/\">How to build your own Youtube \u2013 Part 2<\/a> - June 8, 2016<\/li><li><a href=\"http:\/\/goodheads.io\/2016\/06\/01\/build-youtube-part-1\/\">How to build your own Youtube \u2013 Part 1<\/a> - June 1, 2016<\/li><\/ul><\/div><div class=\"wpautbox-socials wpautbox-socials-square wpautbox-socials-colored\"><\/div><\/div>\t\t\t\t\t\t\t<\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2019-04-17T03:19:56+0000","updated_at":"2019-04-17T03:19:59+0000","published_at":"2016-03-13T20:48:40+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"goodheads.io","preview_picture":"http:\/\/goodheads.io\/wp-content\/uploads\/2016\/03\/magic.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["json-ld","activitypub","php","rdf"],"is_public":false,"id":910,"title":"lanthaler\/JsonLD","url":"https:\/\/github.com\/lanthaler\/JsonLD\/blob\/master\/README.md","content":"<p>JsonLD is a fully conforming <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/\" rel=\"nofollow\">JSON-LD<\/a>\nprocessor written in PHP. It is extensively tested and passes the\n<a href=\"https:\/\/github.com\/json-ld\/tests\">official JSON-LD test suite<\/a>.<\/p>\n<p>There's an <a href=\"http:\/\/www.markus-lanthaler.com\/jsonld\/playground\/\" rel=\"nofollow\">online playground<\/a>\nwhere you can evaluate the processor's basic functionality.<\/p>\n<p>Additionally to the features defined by the <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/\" rel=\"nofollow\">JSON-LD API specification<\/a>,\nJsonLD supports <a href=\"http:\/\/json-ld.org\/spec\/latest\/json-ld-framing\/\" rel=\"nofollow\">framing<\/a>\n(including <a href=\"https:\/\/github.com\/json-ld\/json-ld.org\/issues\/110\">value matching<\/a>,\n<a href=\"https:\/\/github.com\/json-ld\/json-ld.org\/issues\/110\">deep-filtering<\/a>,\n<a href=\"https:\/\/github.com\/json-ld\/json-ld.org\/issues\/119\">aggressive re-embedding<\/a>, and\n<a href=\"https:\/\/github.com\/json-ld\/json-ld.org\/issues\/118\">named graphs<\/a>) and an experimental\n<a href=\"https:\/\/github.com\/lanthaler\/JsonLD\/issues\/15\">object-oriented interface for JSON-LD documents<\/a>.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>The easiest way to install <code>JsonLD<\/code> is by requiring it with <a href=\"https:\/\/getcomposer.org\/\" rel=\"nofollow\">Composer<\/a>.<\/p>\n<pre>composer require ml\/json-ld\n<\/pre>\n<p>... and including Composer's autoloader to your project<\/p>\n<div class=\"highlight highlight-text-html-php\"><pre>require('vendor\/autoload.php');<\/pre><\/div>\n<p>Of course, you can also download JsonLD as\n<a href=\"https:\/\/github.com\/lanthaler\/JsonLD\/releases\">ZIP archive<\/a> from Github.<\/p>\n<p>JsonLD requires PHP 5.3 or later.<\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>The library supports the official <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/\" rel=\"nofollow\">JSON-LD API<\/a> as\nwell as a object-oriented interface for JSON-LD documents (not fully implemented yet,\nsee <a href=\"https:\/\/github.com\/lanthaler\/JsonLD\/issues\/15\">issue #15<\/a> for details).<\/p>\n<p>All classes are extensively documented. Please have a look at the source code.<\/p>\n<div class=\"highlight highlight-text-html-php\"><pre>\/\/ Official JSON-LD API\n$expanded = JsonLD::expand('document.jsonld');\n$compacted = JsonLD::compact('document.jsonld', 'context.jsonld');\n$framed = JsonLD::frame('document.jsonld', 'frame.jsonld');\n$flattened = JsonLD::flatten('document.jsonld');\n$quads = JsonLD::toRdf('document.jsonld');\n\/\/ Output the expanded document (pretty print)\nprint JsonLD::toString($expanded, true);\n\/\/ Serialize the quads as N-Quads\n$nquads = new NQuads();\n$serialized = $nquads-&gt;serialize($quads);\nprint $serialized;\n\/\/ And parse them again to a JSON-LD document\n$quads = $nquads-&gt;parse($serialized);\n$document = JsonLD::fromRdf($quads);\nprint JsonLD::toString($document, true);\n\/\/ Node-centric API\n$doc = JsonLD::getDocument('document.jsonld');\n\/\/ get the default graph\n$graph = $doc-&gt;getGraph();\n\/\/ get all nodes in the graph\n$nodes = $graph-&gt;getNodes();\n\/\/ retrieve a node by ID\n$node = $graph-&gt;getNode('http:\/\/example.com\/node1');\n\/\/ get a property\n$node-&gt;getProperty('http:\/\/example.com\/vocab\/name');\n\/\/ add a new blank node to the graph\n$newNode = $graph-&gt;createNode();\n\/\/ link the new blank node to the existing node\n$node-&gt;addPropertyValue('http:\/\/example.com\/vocab\/link', $newNode);\n\/\/ even reverse properties are supported; this returns $newNode\n$node-&gt;getReverseProperty('http:\/\/example.com\/vocab\/link');\n\/\/ serialize the graph and convert it to a string\n$serialized = JsonLD::toString($graph-&gt;toJsonLd());<\/pre><\/div>\n<h2><a id=\"user-content-commercial-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#commercial-support\"><\/a>Commercial Support<\/h2>\n<p>Commercial support is available on request.<\/p>","created_at":"2019-04-16T12:27:05+0000","updated_at":"2019-04-20T14:20:14+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/456407?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["databases","prolog","postgresql","datalog","logic"],"is_public":false,"id":909,"title":"ekoontz\/psqlog","url":"https:\/\/github.com\/ekoontz\/psqlog","content":"<div class=\"plain\"><pre>Psqlog - a Prolog implementation in PostgreSQL\n---------------------\n- Table setup\nThere are only 2 tables in psqlog: \n--preds:\nA row in the pred table is simply an id, a character string and a (possibly NULL) antec_of. A non-NULL antec_of \nmeans that the row in question is right side of a rule, and the antec_of is an index into the same table for the \nrow of the left side (the consequent) of the rule. For example,\nin the prolog rule:\n(1) father(X,Y) :- male(X), child(X,Y)\nThe preds male and child are antecedents of the pred father. Therefore, the pred table would look something like:\n id   |   name   | antec_of \n------+----------+----------\n 100  | male     | 102         \n 101  | child    | 102\n 102  | father   | \nSo preds 100 and 101 both have their antec_of set to the id of their precedent predicate, 'father', with id 102. This latter \nrow has a NULL antec_of because it is the consequent. Facts, which are simple rules with no antecedents, also \nhave a NULL antec_of. For example, the fact:\n(2) male(joe)\nwould be represented as a row similar to 102 above. Unlike left sides of rules, however, a fact has no rows which have\nthe fact's id as their antec_of value.\nNote that there as many rows in the pred table as there are predicates in rules and facts; thus another rule such as:\n(3) son(A,B) :- male(A),child(A,B)\nwould have separate rows for 'male' and 'child' from those used in the 'father' rule above in (1).\n--args:\nThis table describes the arguments of predicates according to their position and name. An arg may be either a \nconstant, which is encoded by the 'type' column's values of 'c' or 'v', respectively. The arg's pred is given by the \npred_id column, and the position has a column as well. Note that it would be an error for 2 rows to have \nboth the same pred_id and the same value for their position column. For example, the relevant five rows \nin the args table for the rule in (1) would be (with possiby different values for the id column):\n  id   | type |   name   | pred_id | position \n-------+------+----------+---------+----------\n 16623 | V    | X        |     100 |        1\n 16624 | V    | Y        |     100 |        2\n 16625 | V    | X        |     101 |        1\n 16625 | V    | X        |     102 |        1\n 16625 | V    | Y        |     102 |        2\n- Integrating with existing tables\nPsqlog can be applied to existing table structures by making use of views. SQL views allow us to map \nan existing domain (a set of tables and relations) onto Psqlog's very minimal set of tables.\nAs mentioned above, Psqlog has only two tables : preds and args. However, by using views, we can convert existing\ntables to the pred and arg structure used by Psqlog.\nExample: one-place predicate.\n=============================\nSuppose we want to have a table of employees. One column 'telecommutes' is a boolean flag\nthat indicates whether the given employee is a telecommuter or not. We want to map this \nrelation to the predicate \"telecommutes(X)\" - that is, in prolog, we want to ask our system:\n:- telecommutes(Bob)\nAnd receive a response \n\"Yes\"\nif Bob is a telecommuter, and \"No\" otherwise.\nOur domain-specific table looks like:\nCREATE TABLE employee (\n       employee_id integer PRIMARY KEY,\n       name varchar(256),\n       telecommutes boolean\n);\n           Table \"public.employee\"\n Column        |          Type          | Modifiers \n---------------+------------------------+-----------\n name          | character varying(128) | \n telecommutes  | boolean                | \n id            | integer                | not null\nWe can generate a 'telecommutes_pred' and 'telecommutes_arg' based on this as follows:\n  CREATE VIEW telecommutes_pred AS \n    SELECT \n        id,'telecommutes' as telecommutes,'' AS antec_of \n    FROM employee WHERE telecommutes='t';\n  CREATE VIEW telecommutes_arg AS \n    SELECT \n        'c' AS type,employee.name,employee.id AS pred_id,1 AS position \n    FROM employee WHERE telecommutes='t';\nGiven the following sample values in employee:\n id | name | telecommutes \n----+------+---------\n  1 | Bob  | t\n  2 | John | t\n  3 | Bill | f\n  3 | Andy | f\n   (Figure 1: sample values for employee table)\nThe above views yield:\nSELECT * FROM telecommutes_pred;\n id | telecommutes | antec_of \n----+--------------+----------\n  1 | telecommutes | \n  2 | telecommutes | \nSELECT * FROM telecommutes_arg; \n type | name | pred_id | position \n------+------+---------+----------\n c    | Bob  |       1 |        1\n c    | John |       2 |        1\nExample: two-place predicate.\n=============================\nSuppose we have the following table to express how employees can report to each other:\n Table \"public.management\" \n Column  |  Type   | Modifiers \n---------+---------+-----------\n managed | integer | \n manager | integer | \n id      | integer | primary key\nWe then create the following views:\nCREATE VIEW boss_pred AS \n     SELECT management.id,'manages' AS manages, '' AS antec_of FROM management;\nCREATE VIEW boss_arg AS \n   SELECT \n     'c' AS type,\n     boss.name AS name,\n     management.id AS pred_id,\n     1 AS position \n   FROM \n     employee AS boss, employee AS worker,management \n   WHERE worker.id = management.managed AND boss.id=management.manager \n UNION \n   SELECT 'c' AS type,\n     worker.name AS name,\n     management.id AS pred_id,\n     2 AS position \n  FROM \n    employee AS boss, employee AS worker,management \n  WHERE worker.id = management.managed AND boss.id=management.manager;\nFor example, suppose we have the following row in the management table:\n(note: managed and manager column values are indices into the employee table,\nas dumped above in Figure 1.\npsqlog=# select * from management;\n managed | manager | id \n---------+---------+----\n       2 |       3 |  1\nNow we can select from the above views and derive the following:\npsqlog=# SELECT * FROM boss_pred;\n id | manages | antec_of \n----+---------+----------\n  1 | manages | \n(1 row)\npsqlog=# SELECT * FROM boss_arg; \n type | name | pred_id | position \n------+------+---------+----------\n c    | Bill |       1 |        1\n c    | John |       1 |        2\n(2 rows)\n<\/pre><\/div>","created_at":"2019-04-15T14:39:47+0000","updated_at":"2019-04-15T14:40:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/101289?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":908,"title":"Rich Hickey - Talks, Podcasts, and more - PurelyFunctional.tv","url":"https:\/\/purelyfunctional.tv\/programmer-profiles\/rich-hickey\/","content":"<div class=\"pa4 measure\"><img class=\"w-100 fr-ns w-25-ns ml2 mb2\" src=\"https:\/\/purelyfunctional.tv\/wp-content\/uploads\/2016\/09\/Rich-Hickey-610x343.jpg\" alt=\"image\" \/><p>Rich Hickey is the creator of Clojure and Datomic. He's a prolific\nspeaker and has a lot of great ideas about programming. His impact on\nthe future of programming is ongoing and extends well beyond\nClojure. This is a collection of his talks, interviews, and articles.<\/p><\/div><section class=\"br4 bg-light-gray\" id=\"filter-group\"><div class=\"filter-group-open flex\"><h3 class=\"w-90 pa4 ma0\"> Filter<\/h3><\/div><div class=\"filter-stuff pa4 dn\"><div class=\"cf\"><div class=\"filters media_types w-100 w-33-ns pa2 fl\"><h4 class=\"tc\">Media types<\/h4><\/div><div class=\"filters resource_types w-100 w-33-ns pa2 fl\"><h4 class=\"tc\">Resource types<\/h4><\/div><\/div><\/div><\/section><section class=\"mw8 center mt5 resources\"><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3713\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=VSdnJDO-xdg\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/VSdnJDO-xdg\/maxresdefault.jpg\" data-youtube_id=\"VSdnJDO-xdg\" alt=\"image\" \/><\/a><\/div><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"databases,datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3732\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=Cym4TZwTCNU\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/Cym4TZwTCNU\/maxresdefault.jpg\" data-youtube_id=\"Cym4TZwTCNU\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Hickey discusses the architecture of Datomic, his new database\nsystem. The interesting thing about Datomic is that it pulls apart the\npieces of the database (storage, indexing, querying, etc.) in order to\nchoose different tradeoffs from what traditional database systems have\nmade. In addition, Datomic is a true record-keeping system: nothing is\never overwritten.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 databases \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3733\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=4iaIwiemqfo\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/4iaIwiemqfo\/maxresdefault.jpg\" data-youtube_id=\"4iaIwiemqfo\" alt=\"image\" \/><\/a><\/div><\/div><p>Video \u00b7 Presentation \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure,datomic\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3734\"><div><a href=\"https:\/\/channel9.msdn.com\/posts\/Expert-to-Expert-Erik-Meijer-and-Rich-Hickey-Clojure-and-Datomic\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/sec.ch9.ms\/ch9\/3acd\/8aaa8b2d-2d34-4c5d-bfd6-10c8ebc83acd\/E2EMeijerHickeyClojureGOTOChicago_512.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Erik Meijer interviews Rich Hickey about Clojure. He briefly talks\nabout Datomic near the end.<\/p><\/div><p>Video \u00b7 Interview \u00b7 clojure \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Text\" data-resource_type=\"Article\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3736\"><div><a href=\"https:\/\/www.infoq.com\/articles\/Architecture-Datomic\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/articles\/Architecture-Datomic\/en\/resources\/large-DatomicPhysicalModel1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>An article by Rich Hickey discussing the motivations and broad architecture of Datomic.<\/p><\/div><p>Text \u00b7 Article \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3737\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/Datomic\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/Datomic\/en\/slides\/1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>A nice talk about the architecture and implementation of Datomic.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic,immutability\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3738\"><div class=\"description\"><p>An interview with Rich Hickey about Datomic.<\/p><\/div><p>Video \u00b7 Interview \u00b7 datomic \u00b7 immutability<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic,immutability\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3739\"><div class=\"description\"><p>Another interview with Rich Hickey about Datomic.<\/p><\/div><p>Video \u00b7 Interview \u00b7 datomic \u00b7 immutability<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"codeq,datomic\" data-media_type=\"Audio\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3740\"><p><\/p><div class=\"description\"><p>Codeq is a system that reads in a Git repo's commit history and puts\nit into Datomic. You can then query it. Craig Andera interviewed Rich\nHickey about it.<\/p><\/div><p>Audio \u00b7 Interview \u00b7 codeq \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3741\"><div class=\"description\"><p>Another interview with Rich Hickey about Datomic.<\/p><\/div><p>Video \u00b7 Interview \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3742\"><div class=\"relative\"><div class=\"vimeo-image\"><a href=\"https:\/\/vimeo.com\/45136212\" target=\"_blank\"> <img data-vimeo_id=\"45136212\" src=\"src\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey does a deep dive into the design of Datomic. What problems\nwere they addressing? What decisions did they have to make?<\/p><\/div><p>Video \u00b7 Presentation \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Audio\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3743\"><p><\/p><div class=\"description\"><p>Craig Andera interviewed Rich Hickey shortly after Datomic was released.<\/p><\/div><p>Audio \u00b7 Interview \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"simplicity\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3747\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=rI8tNMsozo0\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/rI8tNMsozo0\/maxresdefault.jpg\" data-youtube_id=\"rI8tNMsozo0\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Given at RailsConf, this talk dives deep into the meaning of the word\nsimple. Often confused with easy, which is subjective, simple has an\nobjective definition. Simple means not tied together.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 simplicity<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"problem solving,programming\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3749\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=f84n5oFoZBc\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/f84n5oFoZBc\/maxresdefault.jpg\" data-youtube_id=\"f84n5oFoZBc\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>The title is a play on popular \"programming methodologies\". In this\ntalk, Hickey explains how thinking more and thinking better about a\nproblem can actually develop a better design than structured\nmethodologies like TDD. Problem solving and deep thinking are things\nhumans do well. Hickey dives into how to best use your brain to write\nprograms.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 problem solving \u00b7 programming<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3751\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/design-composition-performance-keynote\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/design-composition-performance-keynote\/en\/slides\/sl1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>I was lucky to be in attendance at Clojure\/West when Rich Hickey gave\nthis talk. The title is a double-entendre, but I\u2019ll let you gather the\nreal meaning when you watch it. Just know that it\u2019s one of Rich\u2019s\nexcellent philosophical talks. Sam Garrett has <a href=\"https:\/\/github.com\/sinemetu1\/clj-west-2013-notes\/blob\/master\/3-18-2013\/design-composition-and-performance.md\">some notes<\/a><\/p><p>Another recording: <a href=\"https:\/\/www.infoq.com\/presentations\/Design-Composition-Performance\">Design, Composition, and Performance<\/a><\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"jvm,time\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3753\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/Are-We-There-Yet-Rich-Hickey\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/Are-We-There-Yet-Rich-Hickey\/en\/slides\/1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Rich Hickey presents the idea that our software needs to explicitly\nmodel time. We should be explicit about what will change and how. To\ndo this, we need to make values immutable by default. This model of\ntime is baked into Clojure: immutable data structures plus a model\nstructured model of state change.<\/p><p>A <a href=\"https:\/\/github.com\/matthiasn\/talk-transcripts\/blob\/master\/Hickey_Rich\/AreWeThereYet.md\">transcript<\/a> is available.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 jvm \u00b7 time<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"time\" data-media_type=\"Text\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3754\"><div class=\"description\"><p>An interview with Rich Hickey about time.<\/p><\/div><p>Text \u00b7 Interview \u00b7 time<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3758\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=P76Vbsk_3J0\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/P76Vbsk_3J0\/maxresdefault.jpg\" data-youtube_id=\"P76Vbsk_3J0\" alt=\"image\" \/><\/a><\/div><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3761\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=7mbcYxHO0nM\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/7mbcYxHO0nM\/maxresdefault.jpg\" data-youtube_id=\"7mbcYxHO0nM\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>There was a time when there were more Common Lisp programmers than\nClojure programmers. This talk explains to the Lispers what made\nClojure different and why.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Audio\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3766\"><p><\/p><div class=\"description\"><p>Rich Hickey interviewed about Clojure.<\/p><\/div><p>Audio \u00b7 Interview \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3767\"><div class=\"description\"><p>Rich Hickey in discussion about Clojure.<\/p><\/div><p>Video \u00b7 Interview \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure,concurrency\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3768\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=dGVqrGmwOAw\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/dGVqrGmwOAw\/maxresdefault.jpg\" data-youtube_id=\"dGVqrGmwOAw\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>The famous \"Ant Colony\" demo where Rich Hickey builds a graphical\nrepresentation of an ant colony to demonstrate Refs. Each cell in the\ngrid is a Ref.<\/p><p>You can get the <a href=\"https:\/\/github.com\/dimhold\/clojure-concurrency-rich-hickey\/blob\/master\/ClojureConcurrencyTalk.pdf?raw=true\">slides<\/a> and see the\n(<a href=\"https:\/\/github.com\/juliangamble\/clojure-ants-simulation\">https:\/\/github.com\/juliangamble\/clojure-ants-simulation<\/a>).<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure \u00b7 concurrency<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure,clojurescript,edn\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3772\"><div class=\"description\"><p>In Clojure 1.4, <a href=\"https:\/\/github.com\/edn-format\/edn\">edn<\/a> was introduced. This talk dives into the motivations and how it compares with other data formats.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure \u00b7 clojurescript \u00b7 edn<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure,protocols\" data-media_type=\"\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3774\"><div class=\"description\"><p>Clojure 1.3 introduced protocols. This is another interview.<\/p><\/div><p>Interview \u00b7 clojure \u00b7 protocols<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3776\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/hickey-clojure\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/hickey-clojure\/en\/slides\/1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>In this very early talk, Rich Hickey goes through a veritable laundry list of Clojure features. It just shows you how far we've come as an industry.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure.spec\" data-media_type=\"Audio\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3777\"><p><\/p><div class=\"description\"><p>Clojure 1.9 will have a new feature called clojure.spec. Craig Andera\ninterviews Rich Hickey about its motivations.<\/p><\/div><p>Audio \u00b7 Interview \u00b7 clojure.spec<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3778\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=DZxvH_iFAXE\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/DZxvH_iFAXE\/maxresdefault.jpg\" data-youtube_id=\"DZxvH_iFAXE\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey interviewed about Java as a host, the open source development process, and what attracts people to Clojure.<\/p><\/div><p>Video \u00b7 Interview \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure,complexity,productivity\" data-media_type=\"Text\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3779\"><div class=\"description\"><p>In this article, Rich Hickey is interviewed about programmer productivity, complexity, and Clojure.<\/p><\/div><p>Text \u00b7 Interview \u00b7 clojure \u00b7 complexity \u00b7 productivity<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"reducers\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3785\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=sf-lrFERkvs\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/sf-lrFERkvs\/maxresdefault.jpg\" data-youtube_id=\"sf-lrFERkvs\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey interviewed about Reducers.<\/p><\/div><p>Video \u00b7 Interview \u00b7 reducers<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"transducers\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3788\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=4KqUvG8HPYo\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/4KqUvG8HPYo\/maxresdefault.jpg\" data-youtube_id=\"4KqUvG8HPYo\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Implementation details about Transducers and what it takes to make a custom one.<\/p><p>A <a href=\"https:\/\/github.com\/matthiasn\/talk-transcripts\/blob\/master\/Hickey_Rich\/InsideTransducers.md\">transcript<\/a> is available.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 transducers<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"core.async\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3789\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/core-async-clojure\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/core-async-clojure\/en\/slides\/sl1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Rich Hickey explains the design and motivation of core.async.<\/p><p>Another recording his <a href=\"https:\/\/www.infoq.com\/presentations\/clojure-core-async\">here<\/a>.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 core.async<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"core.async\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3790\"><div class=\"relative\"><div class=\"vimeo-image\"><a href=\"https:\/\/vimeo.com\/100518968\" target=\"_blank\"> <img data-vimeo_id=\"100518968\" src=\"src\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>If you've ever wondered how core.async works under the hood, this video is for you. Short answer: lots of locks. :)<\/p><p>A <a href=\"https:\/\/github.com\/matthiasn\/talk-transcripts\/blob\/master\/Hickey_Rich\/ImplementationDetails.md\">transcript<\/a> is available.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 core.async<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"core.async\" data-media_type=\"Audio\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3791\"><p><\/p><div class=\"description\"><p>Rich Hickey introduces core.async in the ThinkRelevance podcast (now the Cognicast). He talks about what it does, how it works, and why some of the choices were made.<\/p><\/div><p>Audio \u00b7 Presentation \u00b7 core.async<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"immutability\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3748\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=-6BsiVyC1kM\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/-6BsiVyC1kM\/maxresdefault.jpg\" data-youtube_id=\"-6BsiVyC1kM\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>In this talk, Rich Hickey evangelizes the use of immutable values\ninstead of mutable objects. He opposes the fundamental idea of Object\nOriented Programming, which is to couple data with functionality. Data\nis cross-platform, stable, and comparable.<\/p><p>Rich Hickey is a force of nature. The mark he will leave on programming will reach much farther than the Clojure programming language. The agenda he is setting forth will have as much impact on the notion of state as McCarthy conditionals had on structured programming.<\/p><p>Please listen closely. The main idea of this talk goes far deeper than \"immutable good, mutable bad\".<\/p><p>It is wonderful to be a part of this.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 immutability<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"harmonikit\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3792\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=bhkdyCPYgLs\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/bhkdyCPYgLs\/maxresdefault.jpg\" data-youtube_id=\"bhkdyCPYgLs\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey explains the design of an audio patch creator.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 harmonikit<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"transducers\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3787\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=6mTbuzafcII\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/6mTbuzafcII\/maxresdefault.jpg\" data-youtube_id=\"6mTbuzafcII\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>A nice talk about what transducers are and a little about how they work.<\/p><p>A <a href=\"https:\/\/github.com\/matthiasn\/talk-transcripts\/blob\/master\/Hickey_Rich\/Transducers.md\">transcript<\/a> is available.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 transducers<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"reducers\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3786\"><div class=\"relative\"><div class=\"vimeo-image\"><a href=\"https:\/\/vimeo.com\/45561411\" target=\"_blank\"> <img data-vimeo_id=\"45561411\" src=\"src\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey at EuroClojure explaining reducers.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 reducers<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"reducers\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3783\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/Clojure-Reducers\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/clojure-reducers\/en\/slides\/sl1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Rich Hickey explains the motivation and implementation of reducers.<\/p><p>A <a href=\"https:\/\/github.com\/matthiasn\/talk-transcripts\/blob\/master\/Hickey_Rich\/Reducers.md\">transcript<\/a> is available.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 reducers<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3782\"><div><a href=\"https:\/\/channel9.msdn.com\/blogs\/charles\/emerging-langs-clojure-and-f\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/sec.ch9.ms\/ecn\/ch9\/3931\/563931\/ELangsClojureFSharpHickeyPamer_512_ch9.png\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Rich Hickey is interviewed along with Joe Pamer (an F# compiler developer).<\/p><\/div><p>Video \u00b7 Interview \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Text\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3781\"><div class=\"description\"><p>Another interview of Rich Hickey about Clojure.<\/p><\/div><p>Text \u00b7 Interview \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"immutability\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3775\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/Value-Identity-State-Rich-Hickey\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/Value-Identity-State-Rich-Hickey\/en\/slides\/1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Rich Hickey explains his conceptions of value, identity, and state. A great talk about the basics of functional programming in a changing world.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 immutability<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojurescript\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3771\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=tVooR-dF_Ag\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/tVooR-dF_Ag\/maxresdefault.jpg\" data-youtube_id=\"tVooR-dF_Ag\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey introducing ClojureScript. Back then it was just a budding\nlittle language! But this talk goes deep into the motivations for\nClojureScript, its relationship to Google Closure, and what parts of\nClojure are possible to implement.<\/p><p><a href=\"http:\/\/cloud.github.com\/downloads\/clojure\/clojurescript\/clojurescript%20slides.pdf\">Slides<\/a> are available, as well as a nice <a href=\"http:\/\/clojure.com\/blog\/2011\/07\/22\/introducing-clojurescript.html\">writeup by Stuart Sierra<\/a>.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojurescript<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Text\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3770\"><div class=\"description\"><p>Rich Hickey interviewed about STM, metadata, multimethods, and more.<\/p><\/div><p>Text \u00b7 Interview \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3764\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=sp2Zv7KFQQ0\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/sp2Zv7KFQQ0\/maxresdefault.jpg\" data-youtube_id=\"sp2Zv7KFQQ0\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey gave a screencast, largely at the REPL, exploring the data\nstructures of Clojure. It includes numbers, symbols, keywords,\nstrings, and the collections.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3762\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=ketJlzX-254\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/ketJlzX-254\/maxresdefault.jpg\" data-youtube_id=\"ketJlzX-254\" alt=\"image\" \/><\/a><\/div><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3760\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=cPNkH-7PRTk\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/cPNkH-7PRTk\/maxresdefault.jpg\" data-youtube_id=\"cPNkH-7PRTk\" alt=\"image\" \/><\/a><\/div><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3759\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=hb3rurFxrZ8\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/hb3rurFxrZ8\/maxresdefault.jpg\" data-youtube_id=\"hb3rurFxrZ8\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Way back in the early days of Clojure, Rich Hickey gave classroom\nstyle lectures to introduce Clojure to Java programmers. Be careful,\nsome of the language has changed.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"Distributed Systems\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3750\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=ROor6_NGIWU\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/ROor6_NGIWU\/maxresdefault.jpg\" data-youtube_id=\"ROor6_NGIWU\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Starting the talk with an obviously dry humor title, Rick Hickey\nexplores the idea of software systems -- multiple programs\ncommunicating -- and how the architecture of those systems compares\nwith the architecture of a single program. Within programs, we take a\nlot of things for granted. Those same things are often the source of\nerrors once machines start failing, networks go down, or versions\nchange. What are the tradeoffs for data formats for communicating?\nWhat characteristics should we strive for in the individual pieces of\nour systems? These questions and more are explored.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 Distributed Systems<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"databases,datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3744\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=P-NZei5ANaQ\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/P-NZei5ANaQ\/maxresdefault.jpg\" data-youtube_id=\"P-NZei5ANaQ\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>A high-level presentation of the architectural design of databases.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 databases \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Text\" data-resource_type=\"Article\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3735\"><div class=\"description\"><p>An article by Rich Hickey about the data model in Datomic.<\/p><\/div><p>Text \u00b7 Article \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3731\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=EKdV1IgAaFc\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/EKdV1IgAaFc\/maxresdefault.jpg\" data-youtube_id=\"EKdV1IgAaFc\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey discusses how Datomic revisits the notion of the\ndatabase. He explores the benefits of making the database an immutable\nvalue.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"3717\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=RKcqYZZ9RDY\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/RKcqYZZ9RDY\/maxresdefault.jpg\" data-youtube_id=\"RKcqYZZ9RDY\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>A quick intro to Datomic by Rich Hickey.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"simplicity\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"6332\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/Simple-Made-Easy\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/Simple-Made-Easy\/en\/slides\/sl1.jpg\" alt=\"image\" \/><\/a><\/div><p>Video \u00b7 Presentation \u00b7 simplicity<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure.spec,optionalitiy\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30509\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=YR5WdGrpoug\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/YR5WdGrpoug\/maxresdefault.jpg\" data-youtube_id=\"YR5WdGrpoug\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey talks about why clojure.spec has failed with optionality: values that may or may not be there. He also discusses why Maybe (or Option) in many statically typed languages is not a solution to the problem he is trying to solve.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure.spec \u00b7 optionalitiy<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic,datomic ions\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30510\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=thpzXjmYyGk\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/thpzXjmYyGk\/maxresdefault.jpg\" data-youtube_id=\"thpzXjmYyGk\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Datomic Cloud has the ability to run Clojure applications inside the same JVM as your database. It's called Ions and in this talk, Rich Hickey talks about how it works and why.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 datomic \u00b7 datomic ions<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure.spec,dependencies\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30512\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=oyLBGkS5ICk\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/oyLBGkS5ICk\/maxresdefault.jpg\" data-youtube_id=\"oyLBGkS5ICk\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey picks apart the problems with how library dependencies are thought about in the industry. He talks about how Clojure Spec can fit in.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure.spec \u00b7 dependencies<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30511\"><div class=\"relative\"><div class=\"youtube-image\"><a href=\"https:\/\/www.youtube.com\/watch?v=2V1FtfBDsLU\" target=\"_blank\"> <img class=\"w-100\" src=\"http:\/\/img.youtube.com\/vi\/2V1FtfBDsLU\/maxresdefault.jpg\" data-youtube_id=\"2V1FtfBDsLU\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey talks about the reasons he created Clojure and the design problems Clojure was built for.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure.spec\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30513\"><div class=\"relative\"><div class=\"vimeo-image\"><a href=\"https:\/\/vimeo.com\/195711510\" target=\"_blank\"> <img data-vimeo_id=\"195711510\" src=\"src\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"description\"><p>Rich Hickey presents Clojure Spec.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 clojure.spec<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"datomic\" data-media_type=\"Video\" data-resource_type=\"Presentation\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30514\"><div><a href=\"https:\/\/www.infoq.com\/presentations\/datomic-functional-database\" target=\"_blank\"> <img class=\"w-100\" src=\"https:\/\/res.infoq.com\/presentations\/datomic-functional-database\/en\/slides\/sl1.jpg\" alt=\"image\" \/><\/a><\/div><div class=\"description\"><p>Rich Hickey discusses the design of a database using functional principles.<\/p><\/div><p>Video \u00b7 Presentation \u00b7 datomic<\/p><\/div><\/div><div class=\"pv4\" data-tag=\"clojure,clojure.spec,dependencies,problem solving\" data-media_type=\"Audio\" data-resource_type=\"Interview\"><div class=\"cf w-100 pa4 bg-white shadow-5 br4\" id=\"30515\"><p><\/p><p>Audio \u00b7 Interview \u00b7 clojure \u00b7 clojure.spec \u00b7 dependencies \u00b7 problem solving<\/p><\/div><\/div><\/section>","created_at":"2019-04-15T11:45:59+0000","updated_at":"2019-04-15T11:45:59+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"purelyfunctional.tv","preview_picture":"https:\/\/purelyfunctional.tv\/wp-content\/uploads\/2016\/09\/Rich-Hickey.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":907,"title":"Microsub-spec - IndieWeb","url":"https:\/\/indieweb.org\/Microsub-spec","content":"<p>The <b><dfn>Microsub specification<\/dfn><\/b> provides a standardized way for clients to consume and interact with feeds collected by a server.\n<\/p><ul><li> See the <b><a href=\"https:\/\/indieweb.org\/Microsub\" title=\"Microsub\">main Microsub page<\/a><\/b> for current clients (web, iOS, Android) to start using Microsub today!<\/li><\/ul><p>The Microsub server is responsible for managing the accounts you follow, retrieving updates from them, and the Microsub endpoint provides the feed entries in a normalized format for easy consumption by clients.\n<\/p><dl><dt>Status<\/dt>\n<dd>This is an early <b>Editor's Draft<\/b>, and feedback is encouraged.<\/dd>\n<dt>Participate<\/dt>\n<dd>Wiki (<a href=\"#Feedback\">Feedback<\/a>)<\/dd>\n<dd>GitHub: <a class=\"external text\" href=\"https:\/\/github.com\/indieweb\/microsub\/issues\">issues and brainstorming<\/a><\/dd>\n<dd><a class=\"external text\" href=\"https:\/\/chat.indieweb.org\/dev\">discussion on #indieweb-dev<\/a> (<a class=\"external text\" href=\"irc:\/\/irc.freenode.net\/indieweb-dev\">on Freenode IRC<\/a>)<\/dd>\n<dt>Editor<\/dt>\n<dd><img src=\"https:\/\/aaronparecki.com\/photo.jpg\" class=\"\" alt=\"\" \/><a href=\"https:\/\/indieweb.org\/User:Aaronparecki.com\" title=\"User:Aaronparecki.com\">Aaron Parecki<\/a><\/dd>\n<dt>License<\/dt>\n<dd><small>Per <a class=\"external text\" href=\"https:\/\/creativecommons.org\/publicdomain\/zero\/1.0\/\">CC0<\/a>, to the extent possible under law, the editor(s) and contributors have waived all copyright and related or neighboring rights to this work. In addition, as of 2019-04-14, the editor(s) and contributors (2017-04-09 onward) have made this specification available under the <a class=\"external text\" href=\"http:\/\/www.openwebfoundation.org\/legal\/the-owf-1-0-agreements\/owfa-1-0\">Open Web Foundation Agreement Version 1.0<\/a>.<\/small><\/dd><\/dl><div><div id=\"toc\" class=\"toc\"><p><\/p><h2>Contents<\/h2><ul><li class=\"toclevel-1 tocsection-1\"><a href=\"#Design_Goals\">1 Design Goals<\/a><\/li>\n<li class=\"toclevel-1 tocsection-2\"><a href=\"#Endpoints\">2 Endpoints<\/a>\n<ul><li class=\"toclevel-2 tocsection-3\"><a href=\"#Discovery\">2.1 Discovery<\/a><\/li>\n<li class=\"toclevel-2 tocsection-4\"><a href=\"#Authentication_and_Authorization\">2.2 Authentication and Authorization<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-5\"><a href=\"#Channels\">3 Channels<\/a><\/li>\n<li class=\"toclevel-1 tocsection-6\"><a href=\"#Users\">4 Users<\/a><\/li>\n<li class=\"toclevel-1 tocsection-7\"><a href=\"#Actions\">5 Actions<\/a>\n<ul><li class=\"toclevel-2 tocsection-8\"><a href=\"#Timelines\">5.1 Timelines<\/a>\n<ul><li class=\"toclevel-3 tocsection-9\"><a href=\"#Retrieve_Entries_in_a_Channel\">5.1.1 Retrieve Entries in a Channel<\/a><\/li>\n<li class=\"toclevel-3 tocsection-10\"><a href=\"#Mark_Entries_Read\">5.1.2 Mark Entries Read<\/a><\/li>\n<li class=\"toclevel-3 tocsection-11\"><a href=\"#Remove_Entry_from_a_Channel\">5.1.3 Remove Entry from a Channel<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-12\"><a href=\"#Search\">5.2 Search<\/a>\n<ul><li class=\"toclevel-3 tocsection-13\"><a href=\"#Searching_for_Content\">5.2.1 Searching for Content<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-14\"><a href=\"#Preview\">5.3 Preview<\/a><\/li>\n<li class=\"toclevel-2 tocsection-15\"><a href=\"#Following\">5.4 Following<\/a><\/li>\n<li class=\"toclevel-2 tocsection-16\"><a href=\"#Unfollowing\">5.5 Unfollowing<\/a><\/li>\n<li class=\"toclevel-2 tocsection-17\"><a href=\"#Muting\">5.6 Muting<\/a><\/li>\n<li class=\"toclevel-2 tocsection-18\"><a href=\"#Unmute\">5.7 Unmute<\/a><\/li>\n<li class=\"toclevel-2 tocsection-19\"><a href=\"#Blocking\">5.8 Blocking<\/a><\/li>\n<li class=\"toclevel-2 tocsection-20\"><a href=\"#Channels_2\">5.9 Channels<\/a>\n<ul><li class=\"toclevel-3 tocsection-21\"><a href=\"#Create_a_Channel\">5.9.1 Create a Channel<\/a><\/li>\n<li class=\"toclevel-3 tocsection-22\"><a href=\"#Update_a_Channel\">5.9.2 Update a Channel<\/a><\/li>\n<li class=\"toclevel-3 tocsection-23\"><a href=\"#Delete_a_Channel\">5.9.3 Delete a Channel<\/a><\/li>\n<li class=\"toclevel-3 tocsection-24\"><a href=\"#Set_Channel_Order\">5.9.4 Set Channel Order<\/a>\n<ul><li class=\"toclevel-4 tocsection-25\"><a href=\"#Order_Algorithm\">5.9.4.1 Order Algorithm<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-26\"><a href=\"#Types_of_Feeds\">6 Types of Feeds<\/a><\/li>\n<li class=\"toclevel-1 tocsection-27\"><a href=\"#Objects\">7 Objects<\/a>\n<ul><li class=\"toclevel-2 tocsection-28\"><a href=\"#Posts\">7.1 Posts<\/a>\n<ul><li class=\"toclevel-3 tocsection-29\"><a href=\"#entry\">7.1.1 entry<\/a><\/li>\n<li class=\"toclevel-3 tocsection-30\"><a href=\"#card\">7.1.2 card<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-31\"><a href=\"#Channels_3\">7.2 Channels<\/a><\/li>\n<li class=\"toclevel-2 tocsection-32\"><a href=\"#Feeds\">7.3 Feeds<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-33\"><a href=\"#Paging\">8 Paging<\/a>\n<ul><li class=\"toclevel-2 tocsection-34\"><a href=\"#Example_Paging_Workflow\">8.1 Example Paging Workflow<\/a><\/li>\n<li class=\"toclevel-2 tocsection-35\"><a href=\"#Limiting_Results\">8.2 Limiting Results<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-36\"><a href=\"#Authentication_and_Authorization_Details\">9 Authentication and Authorization Details<\/a>\n<ul><li class=\"toclevel-2 tocsection-37\"><a href=\"#Scopes\">9.1 Scopes<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-38\"><a href=\"#Design_Considerations\">10 Design Considerations<\/a>\n<ul><li class=\"toclevel-2 tocsection-39\"><a href=\"#Why_a_single_endpoint_instead_of_individual_endpoints_for_each_operation\">10.1 Why a single endpoint instead of individual endpoints for each operation<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-40\"><a href=\"#Feedback\">11 Feedback<\/a><\/li>\n<li class=\"toclevel-1 tocsection-41\"><a href=\"#Brainstorming\">12 Brainstorming<\/a><\/li>\n<\/ul><\/div><\/div><h2>Design Goals<\/h2><p>The goal of Microsub is to simplify the process of building a <a href=\"https:\/\/indieweb.org\/reader\" title=\"reader\">reader<\/a>, since there are many moving parts when consuming external content. \n<\/p><p>In general, when subscribing to a feed, a reader should use <a href=\"https:\/\/indieweb.org\/WebSub\" title=\"WebSub\">WebSub<\/a> if the feed is enabled with it, but may need to fall back to polling if not. Depending on the format of the feed, there can be many variations in the actual data available at the feed. For example, there are several different ways an <a href=\"https:\/\/indieweb.org\/h-entry\" title=\"h-entry\">h-entry<\/a> can represent the author of the entry, described at <a href=\"https:\/\/indieweb.org\/authorship\" title=\"authorship\">authorship<\/a>. There are also multiple ways a list of <a href=\"https:\/\/indieweb.org\/h-entry\" title=\"h-entry\">h-entrys<\/a> can appear in an <a href=\"https:\/\/indieweb.org\/h-feed\" title=\"h-feed\">h-feed<\/a>.\n<\/p><p>The role of the Microsub server is to normalize the data in the wild and turn it into a simpler format for displaying in clients. Clients should never have to second guess or doubt any data they receive from the Microsub server. The assumption is that the server has done all the verification and normalizing of the data, and it is ready to display to the user. The fewer checks and conditionals that clients have to write the better.\n<\/p><p>Where HTML or Microformats may provide multiple ways of representing content, Microsub instead has just one way of representing data. For example, in Microformats the \"author\" property may be a string URL or an <code>h-card<\/code>, but in Microsub it will always be a jf2 card.\n<\/p><h2>Endpoints<\/h2><p>The Microsub endpoint is where the client will make all API requests. All API requests require authentication with a Bearer access token that the client needs to obtain. If the client does not have a preexisting relationship with the server, then the following method of discovery and authorization should be used to obtain an access token and discover the Microsub endpoint.\n<\/p><p>It is possible for a client to be pre-configured with a Microsub endpoint, or to use other methods of obtaining an access token if there is a preexisting relationship between the client and the server providing the Microsub endpoint.\n<\/p><h3>Discovery<\/h3><p>The client first performs discovery on the user's profile URL to find the Microsub endpoint and authorization endpoint. Given a user's profile URL, perform an HTTP GET request and look for either a <code>&lt;link rel=\"microsub\"&gt;<\/code> or HTTP <code>Link<\/code> header with a <code>rel<\/code> value of <code>microsub<\/code>. Additionally, look for links with rel values <code>authorization_endpoint<\/code> and <code>token_endpoint<\/code>.\n<\/p><pre>\nLink: &lt;https:\/\/aaronpk.example\/auth&gt;; rel=\"authorization_endpoint\"\nLink: &lt;https:\/\/aaronpk.example\/token&gt;; rel=\"token_endpoint\"\nLink: &lt;https:\/\/aaronpk.example\/microsub&gt;; rel=\"microsub\"\n<\/pre><pre>\n&lt;link rel=\"authorization_endpoint\" href=\"https:\/\/aaronpk.example\/auth\"&gt;\n&lt;link rel=\"token_endpoint\" href=\"https:\/\/aaronpk.example\/token\"&gt;\n&lt;link rel=\"microsub\" href=\"https:\/\/aaronpk.example\/microsub\"&gt;\n<\/pre><p>The Microsub endpoint URL MUST NOT contain a fragment, and MAY contain query string components. If the URL contains a query string, then any GET requests MUST properly append the additional parameters to the query string, and POST requests MUST NOT send the query string properties in the post body. e.g. making a GET request with the additional query string component \"action=config\" to the endpoint \"\/endpoints?type=microsub\" would result in a URL of \"\/endpoints?type=microsub&amp;action=config\"\n<\/p><p>(Note: The client will likely want to also find the <a href=\"https:\/\/indieweb.org\/Micropub\" title=\"Micropub\">Micropub<\/a> endpoint for the user so that the client can post replies and other interactions to the user's website.)\n<\/p><h3>Authentication and Authorization<\/h3><p>Authorization is handled the same way as <a href=\"https:\/\/indieweb.org\/Micropub\" title=\"Micropub\">Micropub<\/a>, using <a href=\"https:\/\/indieweb.org\/IndieAuth\" title=\"IndieAuth\">IndieAuth<\/a> to <a href=\"https:\/\/indieweb.org\/obtaining-an-access-token\" title=\"obtaining-an-access-token\">obtain an access token<\/a>.\n<\/p><h2>Channels<\/h2><p>Channels are described by the following properties:\n<\/p><ul><li> <code>uid<\/code> - a string representation of a user-specific unique ID for the channel. This uid will be unique for each user, but may be duplicated across different users. Some implementations will use constant strings such as \"example\", while others may use numeric database IDs such as \"15029932\", a random string like \"1NYnmUVYR5qBVXYBzt\", or a URL such as \"http:\/\/user.example.com\/channel\/foo\". The valid characters for a uid are any URL-safe character. <\/li>\n<li> <code>name<\/code> - the display name for the channel. This may include any valid UTF-8 sequence. The client should use this name when displaying the name of the channel in the interface.<\/li><\/ul><pre>\n{\n  \"uid\": \"indieweb\",\n  \"name\": \"IndieWeb\"\n}\n<\/pre><p>Servers must always have a channel with the uid <code>notifications<\/code>, and must always have at least one other channel for a user.\n<\/p><p>Some actions may want to apply to every channel, so the uid of <code>global<\/code> is reserved for this purpose. Actions such as <code>mute<\/code> that want to mute a user across every channel use the channel uid of <code>global<\/code>.\n<\/p><h2>Users<\/h2><p>All users are identified by profile URLs, with some constraints. User profile URLs MUST use either the <code>http<\/code> or <code>https<\/code> scheme, and MAY contain path and query string components, and MUST NOT contain fragments.\n<\/p><h2>Actions<\/h2><p>All operations in Microsub are considered \"actions\", and are specified with a query string or form body parameter of <code>action<\/code>.\n<\/p><ul><li> channels<\/li>\n<li> search<\/li>\n<li> preview<\/li>\n<li> follow \/ unfollow<\/li>\n<li> timeline<\/li>\n<li> mute \/ unmute<\/li>\n<li> block \/ unblock<\/li><\/ul><p>Actions that operate within the context of a channel can accept a query string or form body parameter of <code>channel<\/code> specifying the <code>uid<\/code> of the channel to use.\n<\/p><h3>Timelines<\/h3><p><code>action=timeline<\/code>\n<\/p><h4>Retrieve Entries in a Channel<\/h4><p><b>GET<\/b>\n<\/p><p>Retrieve the entries in a given channel.\n<\/p><p>Parameters:\n<\/p><ul><li> action=timeline<\/li>\n<li> channel={uid}<\/li>\n<li> after={cursor}<\/li>\n<li> before={cursor}<\/li><\/ul><p>The response will include a property <code>items<\/code> with an array of post objects. See <a href=\"#Posts\">#Posts<\/a> for documentation on the format of items.\n<\/p><pre>\n{\n  \"items\": [\n    { ... },\n    { ... }\n  ],\n  \"paging\": {\n    \"after\": \"xxxxx\",\n    \"before\": \"xxxxx\"\n  }\n}\n<\/pre><p>See <a href=\"#Paging\">#Paging<\/a> for more details on the paging mechanism.\n<\/p><h4>Mark Entries Read<\/h4><p><b>POST<\/b>\n<\/p><p>To mark one or more individual entries as read or unread: \n<\/p><p>Parameters:\n<\/p><ul><li> action=timeline<\/li>\n<li> method=mark_read (or mark_unread)<\/li>\n<li> channel={uid}<\/li>\n<li> entry={entry-id} or entry[]={entry-id}<\/li><\/ul><p>To mark an entry read as well as everything before it in the timeline: (does not apply to mark_unread)\n<\/p><ul><li> action=timeline<\/li>\n<li> method=mark_read<\/li>\n<li> channel={uid}<\/li>\n<li> last_read_entry={entry-id}<\/li><\/ul><p>These commands only affect the entry within the context of the given channel's timeline, and do not affect other timelines.\n<\/p><h4>Remove Entry from a Channel<\/h4><p><b>POST<\/b>\n<\/p><p>Parameters:\n<\/p><ul><li> action=timeline<\/li>\n<li> method=remove<\/li>\n<li> channel={uid}<\/li>\n<li> entry={entry-id} or entry[]={entry-id}<\/li><\/ul><p>The client can pass one or more entry IDs (found via the <code>_id<\/code> property on an entry) to remove it from the channel. The server may decide whether to actually delete the entry or just hide it from view. The entry will no longer show up in the given channel's timeline, and this command will not affect the given entry in other channels.\n<\/p><p>Servers will likely use their own generated IDs for the entry ID, so clients must not make assumptions about the uniqueness of these values.\n<\/p><h3>Search<\/h3><p><code>action=search<\/code>\n<\/p><p>The \"search\" action exists to provide a UI for the server to respond with the full URL of possible things to subscribe to. For example, a user should not be expected to type the exact URL of a feed to subscribe to, but instead should be able to enter partial matches, e.g. entering <code>aaronparecki.com<\/code> should return the full URL of <code><a class=\"external free\" href=\"https:\/\/aaronparecki.com\/\">https:\/\/aaronparecki.com\/<\/a><\/code>.\n<\/p><p>Using the \"search\" action, a client can provide a single text field where the user can enter either partial URLs or even arbitrary search terms, and the server can reply with a list of URLs that can actually be subscribed to. This also provides the ability to have a confirmation step where users can see a preview of what they will be subscribing to before they actually do so.\n<\/p><p>If the search term is a URL or partial URL, the Microsub server SHOULD fetch the URL if not already known, and discover any feeds at that URL that can be subscribed to. The server may also return feeds that are already known that match the search term, for example if another user on the server has previously subscribed to a matching URL.\n<\/p><p><b>POST<\/b>\n<\/p><p>Request \n<\/p><ul><li> action=search<\/li>\n<li> query=<\/li><\/ul><pre>POST \/microsub\nContent-type: application\/x-www-form-urlencoded\naction=search&amp;query=aaronparecki.com\n<\/pre><p>Response\n<\/p><pre>HTTP\/1.1 200 Ok\nContent-type: application\/json\n{\n  \"results\": [\n    {\n      \"type\": \"feed\",\n      \"url\": \"https:\/\/aaronparecki.com\/\",\n      \"name\": \"Aaron Parecki\",\n      \"photo\": \"https:\/\/aaronparecki.com\/images\/profile.jpg\",\n      \"description\": \"Aaron Parecki's home page\"\n    },\n    {\n      \"type\": \"feed\",\n      \"url\": \"https:\/\/percolator.today\/\",\n      \"name\": \"Percolator\",\n      \"photo\": \"https:\/\/percolator.today\/images\/cover.jpg\",\n      \"description\": \"A Microcast by Aaron Parecki\",\n      \"author\": {\n        \"name\": \"Aaron Parecki\",\n        \"url\": \"https:\/\/aaronparecki.com\/\",\n        \"photo\": \"https:\/\/aaronparecki.com\/images\/profile.jpg\"\n      }\n    },\n    { ... }\n  ]\n}<\/pre><p>The response results are a <a href=\"https:\/\/indieweb.org\/jf2\" title=\"jf2\">jf2<\/a> version of the <a href=\"https:\/\/indieweb.org\/h-feed\" title=\"h-feed\">h-feed<\/a> vocabulary. The only required properies are the type and URL.\n<\/p><ul><li> type - (required) <code>feed<\/code><\/li>\n<li> url - (required) the URL of the result which can be used in a follow action to follow the feed<\/li>\n<li> name - the display name of the feed<\/li>\n<li> photo - a photo or icon representing the feed<\/li>\n<li> description - a description of the feed, either as reported by the feed or generated by the Microsub server<\/li>\n<li> author - a <a href=\"https:\/\/indieweb.org\/jf2\" title=\"jf2\">jf2<\/a> <a href=\"https:\/\/indieweb.org\/h-card\" title=\"h-card\">h-card<\/a> describing the author of the feed. For multi-author feeds this may be an h-card representing the organization, or may be omitted<\/li><\/ul><p>TBD: It might be nice to show an approximate frequency of posts, e.g. <a href=\"https:\/\/indieweb.org\/File:feedly-search-for-url.png\" title=\"File:feedly-search-for-url.png\">Feedly<\/a>, and <a class=\"external free\" href=\"https:\/\/aaronparecki.com\/follow\">https:\/\/aaronparecki.com\/follow<\/a>\n<\/p><h4>Searching for Content<\/h4><p>TBD: implementing a search API for searching past posts that the Microsub server has indexed.\n<\/p><ul><li> action=search<\/li>\n<li> channel={uid}<\/li>\n<li> query={term}<\/li><\/ul><p>The presence of the \"channel\" parameter indicates to the server that the client wants to search for posts rather than search for feeds. The channel parameter can be set to an individual channel uid, or \"global\" to search across all channels.\n<\/p><h3>Preview<\/h3><p><code>action=preview<\/code>\n<\/p><p><b>POST<\/b>\n<\/p><ul><li> action=preview<\/li>\n<li> url={url}<\/li><\/ul><p>The \"preview\" action exists so that the client can display a preview of a URL to the user before the user wants to create a subscription for it. The preview should show as much about the URL as the server can determine, such as basic profile information about the user, and a few recent entries by the user. There should be no permanent side effects created by previewing a URL, and as much as possible, the URL being previewed should not be provided with identifying information of the user who is previewing the URL.\n<\/p><p>The response includes the list of items in the feed if available, in the same format as returned by the <a href=\"#Timelines\">#Timelines<\/a> API call.\n<\/p><h3>Following<\/h3><p><code>action=follow<\/code>\n<\/p><p><b>GET<\/b>\n<\/p><ul><li> action=follow<\/li>\n<li> channel={uid}<\/li><\/ul><p>Retrieve the list of feeds being followed in the given channel.\n<\/p><p>Response\n<\/p><pre>{\n  \"items\": [\n    {\n      \"type\": \"feed\",\n      \"url\": \"http:\/\/tantek.com\/\"\n    }\n  ]\n}<\/pre><p>The feed items may contain additional properties that describe the feed, see <a class=\"external free\" href=\"http:\/\/microformats.org\/wiki\/h-feed\">http:\/\/microformats.org\/wiki\/h-feed<\/a>.\n<\/p><p><br \/><b>POST<\/b>\n<\/p><p>Follow a new URL in a channel. \n<\/p><ul><li> action=follow<\/li>\n<li> channel={uid}<\/li>\n<li> url={url}<\/li><\/ul><pre>POST \/microsub\nContent-type: application\/x-www-form-urlencoded\naction=follow&amp;channel=indieweb&amp;url=http:\/\/tantek.com\/\n<\/pre><p>When a request to the follow endpoint is made, the Microsub server registers the follow action, and begins delivering content at that URL into the channel. The Microsub server can subscribe to the target URL via any mechanism available, but most often will attempt a <a href=\"https:\/\/indieweb.org\/WebSub\" title=\"WebSub\">WebSub<\/a> subscription for its HTML+Microformats, or Atom\/RSS feed, and fall back to polling if that fails.\n<\/p><p>New entries at the followed URL will appear in the channel when fetched from the channel's timeline. The Microsub server may fetch existing entries in the feed and add them to the channel at its discretion.\n<\/p><p>The response is a JSON representation of the channel, the same as is returned when listing all feeds followed in a channel.\n<\/p><pre>{\n  \"type\": \"feed\",\n  \"url\": \"http:\/\/tantek.com\/\"\n}<\/pre><h3>Unfollowing<\/h3><p><code>action=unfollow<\/code>\n<\/p><p><b>POST<\/b>\n<\/p><ul><li> action=unfollow<\/li>\n<li> channel={uid}<\/li>\n<li> url={url}<\/li><\/ul><p>Unfollow a feed from a channel.\n<\/p><pre>POST \/microsub\nContent-type: application\/x-www-form-urlencoded\naction=unfollow&amp;channel=indieweb&amp;url=http:\/\/tantek.com\/\n<\/pre><p>The Microsub server may remove all of this feed's items from the channel, or may leave them in place, at its discretion. If you are used to treating these channels as an IRC or Slack timeline, it would be more appropriate to leave the old items in the channel, just stop delivering new ones. However if you are more used to treating these channels as a Twitter or Facebook feed, then you may want the server to remove them from the channel.\n<\/p><p>TODO: Should there be another parameter for the client to specify whether to remove previous entries or leave them? \n<\/p><h3>Muting<\/h3><p><code>action=mute<\/code>\n<\/p><p>Clients should provide a \"mute\" option in the interface. This allows the user to mute someone's profile, hiding all posts with the muted user's profile URL as the author from being displayed.\n<\/p><p>Muting users will cause all posts by the muted user to be hidden from display. The server MAY still store the posts internally, so that un-muting the user will cause past entries to appear again.\n<\/p><p>Any side effect at the server SHOULD NOT cause the muted user to know they have been muted. Muting users SHOULD NOT have any externally visible side effects.\n<\/p><p>For example, in the context of the <a href=\"https:\/\/indieweb.org\/Salmention\" title=\"Salmention\">Salmention<\/a> spec, the server should still behave as if the muted user was not muted.\n<\/p><p><b>GET<\/b>\n<\/p><ul><li> action=mute<\/li>\n<li> channel={uid}<\/li><\/ul><p>Retrieve the list of users that are muted in the given channel.\n<\/p><p>Request\n<\/p><pre>\nGET \/microsub?action=mute&amp;channel=indieweb HTTP\/1.1\n<\/pre><p>Response\n<\/p><pre>\nHTTP\/1.1 200 Ok\nContent-type: application\/json\n{\n  \"items\": [\n    {\n      \"type\": \"card\",\n      \"url\": \"http:\/\/annoying.example.com\/\",\n      \"name\": \"Annoying Person\",\n      \"photo\": \"http:\/\/annoying.example.com\/photo.jpg\"\n    },\n    {\n      ...\n    }\n  ]\n}\n<\/pre><p><b>POST<\/b>\n<\/p><ul><li> action=mute<\/li>\n<li> channel={uid}<\/li>\n<li> url={url}<\/li><\/ul><p>Mute a user in a channel, or with the uid <code>global<\/code> mutes the user across every channel.\n<\/p><p>Request\n<\/p><pre>\nPOST \/microsub HTTP\/1.1\nContent-type: application\/x-www-form-urlencoded\naction=mute&amp;channel=indieweb&amp;url=https:\/\/annoying.example.com\/\n<\/pre><p>Response\n<\/p><pre>\nHTTP\/1.1 200 Ok\n<\/pre><h3>Unmute<\/h3><p><b>POST<\/b>\n<\/p><p>To unmute a user, use <code>action=unmute<\/code> and provide the URL of the account to unmute. Unmuting an account that was previously not muted has no effect and should not be considered an error.\n<\/p><h3>Blocking<\/h3><p><code>action=block<\/code>\n<\/p><p>Blocking users will cause all previous posts by the blocked user to be hidden or deleted, and future posts by that user should not be stored. Additionally, the server SHOULD NOT produce any content or side effect that would notify the blocked user about a post. It is acceptable for the blocked user to know they have been blocked.\n<\/p><p>For example, in the context of the Webmention spec, the server should not send webmentions even if the user mentions the blocked user in a post. In the context of the Salmention spec, the server should stop sending follow-up webmentions to the blocked user. \n<\/p><p><b>GET<\/b>\n<\/p><ul><li> action=block<\/li>\n<li> channel={uid}<\/li><\/ul><p>Retrieve the list of users that are blocked in the given channel.\n<\/p><p>TODO: document the response format\n<\/p><p><b>POST<\/b>\n<\/p><ul><li> action=block<\/li>\n<li> channel={uid}<\/li>\n<li> url={url}<\/li><\/ul><p>Block a user in a channel, or with the uid <code>global<\/code> blocks the user across every channel.\n<\/p><h3>Channels<\/h3><p><code>action=channels<\/code>\n<\/p><p><b>GET<\/b>\n<\/p><ul><li> action=channels<\/li><\/ul><p>Retrieve the list of channels for the user.\n<\/p><p>The response will contain a <code>channels<\/code> property with the list of channel uids and names. The <code>uid=notifications<\/code> channel must always be the first in the list, as clients are expected to treat it separately and not show it in the channel list.\n<\/p><pre>GET \/microsub?action=channels\nAuthorization: Bearer xxxxxxxxx\nHTTP\/1.1 200 Ok\nContent-type: application\/json\n{\n  \"channels\": [\n    {\n      \"uid\": \"notifications\",\n      \"name\": \"Notifications\",\n      \"unread\": 0\n    },\n    {\n      \"uid\": \"31eccfe322d6c48c50dea2c84efc74ff\",\n      \"name\": \"IndieWeb\"\n      \"unread\": 9\n    },\n    {\n      \"uid\": \"1870e67e924856dc7e4c37732b303b45\",\n      \"name\": \"W3C\",\n      \"unread\": true\n    }\n  ]\n}<\/pre><p>Servers <i>should<\/i> support tracking the read state of items in channels, and return the number of unread items, or a boolean, when the channel list is queried. If the server does not support read state tracking, then the server <i>must not<\/i> return the <code>unread<\/code> property in the response.\n<\/p><p>For channels that support full unread count tracking, the value of the <code>unread<\/code> property will be an integer. For channels that support only boolean unread tracking (whether there are new entries or not), the value will be <code>true<\/code> or <code>false<\/code>.\n<\/p><p><b>POST<\/b>\n<\/p><p>To create, update, or delete channels, the client sends a POST request with the <code>channels<\/code> action.\n<\/p><p>To create a channel, the client includes the name of the channel to create. The uid of the channel will be assigned by the server.\n<\/p><ul><li> action=channels<\/li>\n<li> name={channel name}<\/li><\/ul><p>To update a channel, the client includes the uid of the channel to update in the <code>channel<\/code> parameter, and includes the new name of the channel. Changing the name of the channel MAY change the <code>uid<\/code> of the channel, and if it does, the server MUST return the new <code>uid<\/code> in the response.\n<\/p><ul><li> action=channels<\/li>\n<li> channel={uid}<\/li>\n<li> name={channel name}<\/li><\/ul><p>To delete a channel, the client includes an additional parameter <code>method=delete<\/code>. Note that the <code>notifications<\/code> channel cannot be deleted.\n<\/p><ul><li> action=channels<\/li>\n<li> method=delete<\/li>\n<li> channel={uid}<\/li><\/ul><p>Both creating and updating a channel MUST return the <code>uid<\/code> and <code>name<\/code> properties for the channel.\n<\/p><h4>Create a Channel<\/h4><p>Request\n<\/p><ul><li> action=channels<\/li>\n<li> name={channel name}<\/li><\/ul><pre>\nPOST \/microsub HTTP\/1.1\nContent-type: application\/x-www-form-urlencoded\naction=channels&amp;name=Coworkers\n<\/pre><p>Response\n<\/p><pre>\nHTTP\/1.1 200 Ok\nContent-type: application\/json\n{\n  \"uid\": \"2c5e5d7c4d57da68cb03c972846e827845af974c9b\",\n  \"name\": \"Coworkers\"\n}\n<\/pre><h4>Update a Channel<\/h4><p>Request\n<\/p><ul><li> action=channels<\/li>\n<li> channel={uid}<\/li>\n<li> name={channel name}<\/li><\/ul><pre>\nPOST \/microsub HTTP\/1.1\nContent-type: application\/x-www-form-urlencoded\naction=channels\n&amp;channel=2c5e5d7c4d57da68cb03c972846e827845af974c9b\n&amp;name=Friends\n<\/pre><p>Response\n<\/p><pre>\nHTTP\/1.1 200 Ok\nContent-type: application\/json\n{\n  \"uid\": \"2c5e5d7c4d57da68cb03c972846e827845af974c9b\",\n  \"name\": \"Friends\"\n}\n<\/pre><h4>Delete a Channel<\/h4><p>Request\n<\/p><ul><li> action=channels<\/li>\n<li> method=delete<\/li>\n<li> channel={uid}<\/li><\/ul><pre>\nPOST \/microsub HTTP\/1.1\nContent-type: application\/x-www-form-urlencoded\naction=channels\n&amp;method=delete\n&amp;channel=2c5e5d7c4d57da68cb03c972846e827845af974c9b\n<\/pre><p>Response\n<\/p><pre>\nHTTP\/1.1 200 Ok\nContent-type: application\/json\n<\/pre><h4>Set Channel Order<\/h4><p>Channels are ordered according to the user's preference. When returning the list of channels, the <code>notifications<\/code> channel must always be the first in the list, as clients are expected to treat it separately such as showing it as a separate icon, not in the main channel list.\n<\/p><p>To modify the order of channels, the client sends a POST request with a list of channel IDs in the new ordering.\n<\/p><p>Only the order of the channel IDs specified will be changed. While this command works equally well specifying two, three or more channels, this provides two primary methods for adjusting channel orders. \n<\/p><ul><li> To move a channel up or down, the client can include the two IDs of adjacent channels in the new desired order. <\/li>\n<li> The client can specify a new ordering for all the channels by providing the full list of channel IDs in the new order.<\/li><\/ul><p>Request\n<\/p><ul><li> action=channels<\/li>\n<li> method=order<\/li>\n<li> channels[]={uid}<\/li>\n<li> channels[]={uid}<\/li>\n<li> ...<\/li><\/ul><pre>\nPOST \/microsub HTTP\/1.1\nContent-type: application\/x-www-form-urlencoded\naction=channels\n&amp;method=order\n&amp;channels[]=2c5e5d7c4d57da68cb03\n&amp;channels[]=c972846e827845af974c\n<\/pre><p>Response\n<\/p><pre>\nHTTP\/1.1 200 Ok\nContent-type: application\/json\n<\/pre><h5>Order Algorithm<\/h5><p>From the server's perspective, the below is a description of how to process the order command.\n<\/p><p>Given a list of channels with IDs:\n<\/p><pre>[a b c d e f g h]<\/pre><p>and a command to set the order of the following in the list: <code>[d a c g]<\/code>\n<\/p><p>Assign the items in the initial list a numeric index:\n<\/p><pre>[a b c d e f g h]\n[1 2 3 4 5 6 7 8]<\/pre><p>Build a new map with the input items and their existing numeric order:\n<\/p><pre>[\n  1 =&gt; a\n  3 =&gt; c\n  4 =&gt; d\n  7 =&gt; g\n]<\/pre><p>For each item in the input list in the given order, set the value in the map:\n<\/p><pre>[\n  1 =&gt; d\n  3 =&gt; a\n  4 =&gt; c\n  7 =&gt; g\n]<\/pre><p>For each item in the new map, change the value at the corresponding numeric index of the list:\n<\/p><pre>[d b a c e f g h]\n[1 2 3 4 5 6 7 8]<\/pre><h2>Types of Feeds<\/h2><p>The specific types and formats of feeds that can be followed is out of scope of Microsub. Instead, it's up to the Microsub server to support whichever feed formats it wishes. Typically, Microsub servers will prefer a Microformats 2 feed such as an <code>h-feed<\/code> or list of <code>h-entry<\/code>s, and will then fall back to finding an Atom or RSS feed. Other types of feeds may be supported, but clients should not make any assumptions about which formats are supported, and should make use of the \"preview\" action so that users have an indication of whether a subscription will succeed.\n<\/p><h2>Objects<\/h2><h3>Posts<\/h3><p>Posts are the basic object used in the API. Posts can be short status updates, photos, videos, podcast episodes, checkins, and many other content types. Post objects returned in the \"items\" array MUST be valid <a href=\"https:\/\/indieweb.org\/jf2\" title=\"jf2\">jf2<\/a> post objects. \n<\/p><p>(Note that the representations below should be treated as canonical examples, whereas the jf2 spec itself may provide multiple variations of values. For example, as of 2019-02-25, jf2 says that if there is one value of a property it should be returned as a string, but Microsub will always include the \"photo\" property as an array.)\n<\/p><p>An example of a simple jf2 post in the context of a channel timeline is below.\n<\/p><pre>{\n    \"type\": \"entry\",\n    \"published\": \"2017-04-28T11:58:35-07:00\",\n    \"url\": \"https:\/\/aaronparecki.com\/2017\/04\/28\/9\/p3k-http\",\n    \"author\": {\n        \"type\": \"card\",\n        \"name\": \"Aaron Parecki\",\n        \"url\": \"https:\/\/aaronparecki.com\/\",\n        \"photo\": \"https:\/\/aaronparecki.com\/images\/profile.jpg\"\n    },\n    \"category\": [\n        \"http\",\n        \"p3k\",\n        \"library\",\n        \"code\",\n        \"indieweb\"\n    ],\n    \"photo\": [\n        \"https:\/\/aaronparecki.com\/2017\/04\/28\/9\/photo.png\"\n    ],\n    \"content\": {\n        \"text\": \"Finally packaged up my HTTP functions into a library! https:\/\/github.com\/aaronpk\/p3k-http Previously I had been copy+pasting these around to quite a few projects. Happy to have consolidated these finally!\",\n        \"html\": \"Finally packaged up my HTTP functions into a library! &lt;a href=\\\"https:\/\/github.com\/aaronpk\/p3k-http\\\"&gt;https:\/\/github.com\/aaronpk\/p3k-http&lt;\/a&gt; Previously I had been copy+pasting these around to quite a few projects. Happy to have consolidated these finally!\"\n    },\n    \"_id\": \"abc987\",\n    \"_is_read\": true\n}<\/pre><p>All properties beginning with an underscore <code>_<\/code> are set by the Microsub server and are not part of the post's contents. \n<\/p><ul><li> <code>_id<\/code> (string) - a unique internal ID for this entry in the given channel, used to mark an entry as read or remove it from a channel<\/li>\n<li> <code>_is_read<\/code> (boolean) - indicates whether this entry is read or unread so that clients can visually indicate the state to the user<\/li><\/ul><p>Neither the Microsub spec nor jf2 defines the vocabularies used in posts. The vocabularies are defined by <a href=\"https:\/\/indieweb.org\/Microformats\" class=\"mw-redirect\" title=\"Microformats\">Microformats<\/a>, and represented as <a href=\"https:\/\/indieweb.org\/jf2\" title=\"jf2\">jf2<\/a> posts. Clients may also find the <a class=\"external text\" href=\"https:\/\/www.w3.org\/TR\/post-type-discovery\/\">Post Type Discovery<\/a> spec useful when consuming posts. A non-comprehensive list of properties and their expected values is below, based on the most common kinds of posts encountered.\n<\/p><h4>entry<\/h4><p>Entries are the base post type for most kinds of posts including <a href=\"https:\/\/indieweb.org\/articles\" class=\"mw-redirect\" title=\"articles\">articles<\/a>, <a href=\"https:\/\/indieweb.org\/notes\" class=\"mw-redirect\" title=\"notes\">notes<\/a>, <a href=\"https:\/\/indieweb.org\/photos\" class=\"mw-redirect\" title=\"photos\">photos<\/a>, <a href=\"https:\/\/indieweb.org\/checkins\" class=\"mw-redirect\" title=\"checkins\">checkins<\/a>, and more.\n<\/p><p>Properties:\n<\/p><ul><li> <code>type=entry<\/code><\/li>\n<li> <code>published<\/code> (string) - ISO8601 timestamp<\/li>\n<li> <code>url<\/code> (string) - the permalink of the post<\/li>\n<li> <code>uid<\/code> (string) - a unique ID for the post, often the same as the url<\/li>\n<li> <code>name<\/code> (string) - the name of the post, such as if the post is an <a href=\"https:\/\/indieweb.org\/article\" title=\"article\">article<\/a><\/li>\n<li> <code>content<\/code> (object)\n<ul><li> <code>text<\/code> (string) - a plaintext representation of the post content. If a client displays this text in a browser, it is expected to escape any HTML entities itself.<\/li>\n<li> <code>html<\/code> (string) - an HTML representation of the post content. The Microsub server should do its own HTML sanitization, but clients may also wish to sanitize the HTML themselves according to their own security requirements before displaying the HTML.<\/li><\/ul><\/li>\n<li> <code>author<\/code> (card object) - a card representing the author of the post<\/li>\n<li> <code>category<\/code> (array of strings) - a list of categories (tags) of the post<\/li>\n<li> <code>photo<\/code> (array of strings) - one or more URLs to images associated with the post<\/li>\n<li> <code>video<\/code> (array of strings) - one or more URLs to video files associated with the post<\/li>\n<li> <code>audio<\/code> (array of strings) - one or more URLs to audio files associated with the post, e.g. podcasts<\/li>\n<li> <code>like-of<\/code>, <code>repost-of<\/code>, <code>bookmark-of<\/code>, <code>in-reply-to<\/code> (array of strings) - each of these properties indicates a separate post that this post is in response to. See <a href=\"https:\/\/indieweb.org\/like\" title=\"like\">like<\/a>, <a href=\"https:\/\/indieweb.org\/repost\" title=\"repost\">repost<\/a>, <a href=\"https:\/\/indieweb.org\/bookmark\" title=\"bookmark\">bookmark<\/a>, and <a href=\"https:\/\/indieweb.org\/reply\" title=\"reply\">reply<\/a> for more information.<\/li>\n<li> <code>syndication<\/code> (array of strings) - one or more URLs where this post has been syndicated to<\/li>\n<li> <code>checkin<\/code> (card object) - indicates this post is a checkin, and the value is a card representing the location of the checkin<\/li>\n<li> <code>_id<\/code> - a unique ID for this post in this channel. Used to reference this post when doing things like removing it from view. Note that the underscore prefix indicates this is not a property from the main post, but instead is a property generated by the Microsub server.<\/li><\/ul><h4>card<\/h4><p>Cards represent people or places and are used as the values of certain properties of entries.\n<\/p><p>People:\n<\/p><ul><li> <code>type=card<\/code><\/li>\n<li> <code>name<\/code> (string) - the full name of this person<\/li>\n<li> <code>url<\/code> (string) - the profile URL of this person, used as the identifier for this profile<\/li>\n<li> <code>photo<\/code> (string) - the profile photo of this person<\/li><\/ul><p>Places:\n<\/p><ul><li> <code>type=card<\/code><\/li>\n<li> <code>name<\/code> (string) - the full name of the place<\/li>\n<li> <code>url<\/code> (string) - the URL of the place, used as the identifier for this profile<\/li>\n<li> <code>latitude<\/code> and <code>longitude<\/code> (string) - the GPS coordinates of this location<\/li>\n<li> <code>street-address<\/code>, <code>locality<\/code>, <code>region<\/code>, <code>country<\/code> (string) - the location of this place<\/li><\/ul><h3>Channels<\/h3><h3>Feeds<\/h3><h2>Paging<\/h2><p>Microsub uses a cursor-based pagination model, which provides the most flexibility to server implementations while still being easy to use by clients.\n<\/p><p>For API requests that paginate their results (e.g. <a href=\"#Timeline\">#Timeline<\/a>), there is an additional object returned with the values needed to page through those results. \n<\/p><pre>{\n  \"items\": [ ... ],\n  \"paging\": {\n    \"before\": \"xxxxx\",\n    \"after\": \"xxxxx\"\n  }\n}<\/pre><ul><li> If there are any items returned in the response, the server MUST return a \"before\" value that will retrieve items before all the returned items in the list. <\/li>\n<li> If there are no items returned in the response, the server MUST NOT return a \"before\" value.<\/li>\n<li> If there are additional items available that were not returned in the response, the server MUST return an \"after\" value that will retrieve the next page of items.<\/li>\n<li> If there are no more items available, the server MUST NOT return an \"after\" value.<\/li><\/ul><p>As far as the client is concerned, the \"before\" and \"after\" values are arbitrary strings. This allows the server to internally use whatever specific implementation is most appropriate for its backend technology. (Often this will be either a timestamp or a unique ID identifying the first and last items in the returned list.)\n<\/p><p>To make a timeline request for the next page of results, the client adds the \"after=xxxxx\" parameter to the query string. This allows easy navigation through the whole list of items in the channel.\n<\/p><p>While the user is reading the timeline, the client will likely also want to poll the timeline to see if new posts have been added since it was originally requested. Since not all Microsub servers will support streaming, the client needs an efficient way to poll for new items. The client adds \"before=xxxxx\" to the query string to request items that come before the first item returned in the previous request. This way the client can poll that until new items appear, and only the new items will be returned.\n<\/p><h3>Example Paging Workflow<\/h3><ul><li> The user loads the client and makes a request for the timeline for the default channel:\n<ul><li> <code>\/microsub?action=timeline<\/code><\/li>\n<li> The server replies with the newest 20 items, and includes <code>before=5a1713e55a171588<\/code> and <code>after=5a1713e55a17136c<\/code><\/li><\/ul><\/li>\n<li> The user scrolls to the bottom and clicks \"load more\". The client makes a request for the next set of results:\n<ul><li> <code>\/microsub?action=timeline&amp;after=5a1713e55a17136c<\/code><\/li>\n<li> The server replies with 3 more items, and does not include an \"after\" paging cursor in the response, indicating that there are no more items in the timeline.<\/li><\/ul><\/li>\n<li> Meanwhile, in the background, the client polls the timeline to find newer items by using the first \"before\" cursor that was returned in the initial request:\n<ul><li> <code>\/microsub?action=timeline&amp;before=5a1713e55a171588<\/code><\/li>\n<li> The server replies with an empty <code>items<\/code> list and an empty <code>paging<\/code> object indicating there are no new items<\/li><\/ul><\/li>\n<li> After some interval, the client polls for new items again\n<ul><li> <code>\/microsub?action=timeline&amp;before=5a1713e55a171588<\/code><\/li>\n<li> In the time between the two polls, there have been 25 new items added to the timeline, more than one page of results. The server replies with the newest 20 <code>items<\/code> list, and includes new before and after values, <code>before=5a1724ad5a171599<\/code> and <code>after=5a1722ea5a171280<\/code><\/li><\/ul><\/li>\n<li> Since the client sees there is an \"after\" cursor, it immediately fetches the next page of results using the original \"before\" and the new \"after\" value:\n<ul><li> <code>\/microsub?action=timeline&amp;before=5a1713e55a171588&amp;after=5a1722ea5a171280<\/code><\/li>\n<li> The server replies with the remaining 5 items, and no new cursors since there is no more data missing.<\/li><\/ul><\/li>\n<li> The client then continues polling with the latest \"before\" cursor it received<\/li><\/ul><p>See the <a class=\"external text\" href=\"https:\/\/github.com\/aaronpk\/Aperture\/blob\/master\/aperture\/tests\/Feature\/PagingTest.php#L19\">tests in Aperture<\/a> for an example of the API requests made.\n<\/p><h3>Limiting Results<\/h3><p>Microsub servers SHOULD set a default limit on the number of items returned in lists. A reasonable default limit is 20 items. Microsub servers SHOULD support an additional query parameter <code>limit<\/code> which clients can use to indicate the requested limit of number items returned. Microsub servers MAY set an upper or lower bound on the values they accept for the <code>limit<\/code>, and MAY return a different number of items in the list than the client requets, for any reason. Clients should not expect the number of results returned to exactly match the number of results requested.\n<\/p><h2>Authentication and Authorization Details<\/h2><p>The client builds an <a href=\"https:\/\/indieweb.org\/IndieAuth\" title=\"IndieAuth\">IndieAuth<\/a> authorization request URL at the <a href=\"https:\/\/indieweb.org\/authorization-endpoint\" title=\"authorization-endpoint\">authorization endpoint<\/a>, and directs the user's browser there. In a native client, the client should use a system-native browser, rather than using a web view embedded in the application. See <a class=\"external text\" href=\"https:\/\/tools.ietf.org\/html\/draft-ietf-oauth-native-apps\">OAuth 2.0 for Native Apps<\/a> for more details.\n<\/p><p>Build a URL with the following query parameters:\n<\/p><ul><li> <code>me={the user's profile URL}<\/code> - the URL that the user entered at which the Microsub endpoint was found<\/li>\n<li> <code>response_type=code<\/code><\/li>\n<li> <code>client_id={the client's URL, e.g. its home page}<\/code><\/li>\n<li> <code>state={random state}<\/code> - the client should generate a unique state value, and verify that it matches when the user is redirected<\/li>\n<li> <code>redirect_uri={the client's redirect URI}<\/code> - for native apps, this may include a custom URL scheme<\/li>\n<li> <code>scope={requested scope}<\/code> - a space-separated list of scopes that the client is requesting<\/li><\/ul><h4>Scopes<\/h4><p>Microsub defines the following scopes:\n<\/p><ul><li> <code>read<\/code> - this is the minimum scope clients should request. this allows clients to have read access to channels.<\/li>\n<li> <code>follow<\/code> - allows the client to manage the following list<\/li>\n<li> <code>mute<\/code> - allows the client to mute and unmute users<\/li>\n<li> <code>block<\/code> - allows the client to block and unblock users<\/li>\n<li> <code>channels<\/code> - allows the client to manage channels<\/li><\/ul><p>Additionally, the client may request Micropub scopes, in order for the user to be able to reply or like posts from within the client.\n<\/p><ul><li> <code>create<\/code><\/li>\n<li> <code>update<\/code><\/li>\n<li> <code>delete<\/code><\/li><\/ul><p>The recommended set of scopes to request is <code>read follow mute block create<\/code>, which enables a rich set of interaction on the client, while also protecting the security of the user by default.\n<\/p><p>The user will visit the authorization endpoint, and if they approve the request, their browser will be redirected back to the client's redirect URI with a <code>code<\/code> and <code>state<\/code> in the URL.\n<\/p><pre>\nHTTP\/1.1 302 Found\nLocation: https:\/\/client.example\/redirect?code=xxxxxxxx\n                                         &amp;state=1234567890\n<\/pre><p>The client verifies the state value matches the state it generated for the initial request, and can then exchange the authorization code for an access token.\n<\/p><p>The client makes a POST request to the token endpoint initially discovered, with the following parameters:\n<\/p><ul><li> <code>grant_type=authorization_code<\/code><\/li>\n<li> <code>code=xxxxxxxxx<\/code><\/li>\n<li> <code>redirect_uri={the client's redirect URI}<\/code><\/li>\n<li> <code>client_id={the client's URL}<\/code><\/li><\/ul><p>The response will be a JSON object with the following keys:\n<\/p><pre>\n{\n  \"access_token\": \"XXXXXXXXXXX\",\n  \"scope\": \"read follow mute block create\",\n  \"me\": \"https:\/\/aaronpk.example\/\"\n}\n<\/pre><p>The <code>me<\/code> value returned MAY be different from the original <code>me<\/code> value input, but MUST have a matching host name. This enables support for multi-user websites, and allows the user's server to normalize profile URLs, e.g. it will always return <code>https:\/\/aaronpk.example\/<\/code> even if the user initially enters <code>http:\/\/aaronpk.example<\/code>.\n<\/p><p>The scopes returned MAY be different from what the client requested, based on whether the user choose to deny certain scopes, or grant additional scopes during the authorization request.\n<\/p><h2>Design Considerations<\/h2><h3>Why a single endpoint instead of individual endpoints for each operation<\/h3><p>Many similar APIs such as the <a href=\"https:\/\/indieweb.org\/Twitter_API\" class=\"mw-redirect\" title=\"Twitter API\">Twitter API<\/a> or <a href=\"https:\/\/indieweb.org\/wiki\/index.php?title=Wordpress_API&amp;action=edit&amp;redlink=1\" class=\"new\" title=\"Wordpress API (page does not exist)\">Wordpress API<\/a> use unique URLs for each type of operation: following, muting, fetching posts, etc. Microsub instead takes an <a href=\"https:\/\/indieweb.org\/RPC\" class=\"mw-redirect\" title=\"RPC\">RPC<\/a> style approach, where all requests are made against a single endpoint, with the operation is specified with a query or form parameter.\n<\/p><p>This allows more flexibility in the design of the server, since the spec is not imposing a URL design on the server, each can choose a URL for the Microsub endpoint that makes sense for itself. \n<\/p><p>This also makes clients easier to write, since all requests are made against one base URL rather than needing to either keep track of a fixed URL pattern, or have a configurable URL pattern.\n<\/p><h2>Feedback<\/h2><p>Feel free to leave notes here or open an <a class=\"external text\" href=\"https:\/\/github.com\/indieweb\/microsub\/issues\">issue on GitHub<\/a>.\n<\/p><p>Client developers should be aware of the potential inclusion of a notifications channel when reordering channels. If it is hidden from the ui it should also be hidden from any potential reordering.\n<\/p><h2>Brainstorming<\/h2><p>See <a class=\"external text\" href=\"https:\/\/github.com\/indieweb\/microsub\/issues?q=is%3Aissue+is%3Aopen+label%3Abrainstorming\">brainstorming issues on GitHub<\/a>.\n<\/p>","created_at":"2019-04-14T00:53:55+0000","updated_at":"2019-04-14T00:53:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":29,"domain_name":"indieweb.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["databases"],"is_public":false,"id":906,"title":"How Does a Database Work?","url":"https:\/\/cstack.github.io\/db_tutorial\/","content":"How Does a Database Work? | Let\u2019s Build a Simple Database\n    <div class=\"wrapper\"><header>\n        <p>Writing a sqlite clone from scratch in C<\/p>\n        <p class=\"view\"><a href=\"https:\/\/cstack.github.io\/db_tutorial\/\">Overview<\/a><\/p>\n        \n          <p class=\"view\"><a href=\"https:\/\/github.com\/cstack\/db_tutorial\">View on GitHub (pull requests welcome)<\/a><\/p>\n        \n        \n        \n      <\/header><section><ul><li>What format is data saved in? (in memory and on disk)<\/li>\n  <li>When does it move from memory to disk?<\/li>\n  <li>Why can there only be one primary key per table?<\/li>\n  <li>How does rolling back a transaction work?<\/li>\n  <li>How are indexes formatted?<\/li>\n  <li>When and how does a full table scan happen?<\/li>\n  <li>What format is a prepared statement saved in?<\/li>\n<\/ul><p>In short, how does a database <strong>work<\/strong>?<\/p><p>I\u2019m building a clone of <a href=\"https:\/\/www.sqlite.org\/arch.html\">sqlite<\/a> from scratch in C in order to understand, and I\u2019m going to document my process as I go.<\/p><ul><li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part1.html\">Part 1 - Introduction and Setting up the REPL<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part2.html\">Part 2 - World\u2019s Simplest SQL Compiler and Virtual Machine<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part3.html\">Part 3 - An In-Memory, Append-Only, Single-Table Database<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part4.html\">Part 4 - Our First Tests (and Bugs)<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part5.html\">Part 5 - Persistence to Disk<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part6.html\">Part 6 - The Cursor Abstraction<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part7.html\">Part 7 - Introduction to the B-Tree<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part8.html\">Part 8 - B-Tree Leaf Node Format<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part9.html\">Part 9 - Binary Search and Duplicate Keys<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part10.html\">Part 10 - Splitting a Leaf Node<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part11.html\">Part 11 - Recursively Searching the B-Tree<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part12.html\">Part 12 - Scanning a Multi-Level B-Tree<\/a><\/li>\n  <li><a href=\"https:\/\/cstack.github.io\/db_tutorial\/parts\/part13.html\">Part 13 - Updating Parent Node After a Split<\/a><\/li>\n<\/ul><blockquote>\n  <p>\u201cWhat I cannot create, I do not understand.\u201d \u2013 <a href=\"https:\/\/en.m.wikiquote.org\/wiki\/Richard_Feynman\">Richard Feynman<\/a><\/p>\n<\/blockquote><table class=\"image\"><caption style=\"text-align: bottom;\">sqlite architecture (https:\/\/www.sqlite.org\/arch.html)<\/caption>\n<tr><td><a href=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/arch2.gif\"><img src=\"https:\/\/cstack.github.io\/db_tutorial\/assets\/images\/arch2.gif\" alt=\"sqlite architecture (https:\/\/www.sqlite.org\/arch.html)\" \/><\/a><\/td><\/tr><\/table><\/section><\/div>","created_at":"2019-04-11T11:47:18+0000","updated_at":"2019-04-11T11:47:22+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"cstack.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":905,"title":"About This Website - Gwern.net","url":"https:\/\/www.gwern.net\/About","content":"<p>This page is about <code>gwern.net<\/code>; for information about me, see <a href=\"https:\/\/www.gwern.net\/Links\">Links<\/a>.<\/p><section id=\"the-content\" class=\"level1\"><div class=\"epigraph\"><blockquote>\n<p>\u201cAh! let not Censure term our fate our choice, \/ The stage but echoes back the public\u2019s voice; \/ The drama\u2019s laws the drama\u2019s patrons give, \/ For we that live to please must please to live.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Samuel%20Johnson\" title=\"Wikipedia: Samuel Johnson\">Samuel Johnson<\/a> (<a href=\"http:\/\/en.wikiquote.org\/wiki\/Samuel_Johnson#Prologue_at_the_Opening_of_Drury_Lane_Theatre_.281747.29\">\u201cPrologue at the Opening of Drury Lane Theatre\u201d<\/a>)<\/p>\n<\/blockquote><\/div><p>The content here varies from <a href=\"https:\/\/www.gwern.net\/Google-shutdowns\">statistics<\/a> to <a href=\"https:\/\/www.gwern.net\/DNB-FAQ\">psychology<\/a> to <a href=\"https:\/\/www.gwern.net\/Zeo\">self-experiments<\/a>\/<a href=\"https:\/\/www.gwern.net\/Weather\">Quantified Self<\/a> to <a href=\"https:\/\/www.gwern.net\/Culture-is-not-about-Esthetics\">philosophy<\/a> to <a href=\"https:\/\/www.gwern.net\/fiction\/Brave-poem\">poetry<\/a> to <a href=\"https:\/\/www.gwern.net\/haskell\/Wikipedia-RSS-Archive-Bot\">programming<\/a> to <a href=\"https:\/\/www.gwern.net\/otaku\">anime<\/a> to investigations of <a href=\"https:\/\/www.gwern.net\/Silk-Road\">online drug markets<\/a> or <a href=\"https:\/\/www.gwern.net\/Death-Note-script\">leaked movie scripts<\/a> (or two topics at once: <a href=\"https:\/\/www.gwern.net\/Hafu\">anime &amp; statistics<\/a> or <a href=\"https:\/\/www.gwern.net\/Death-Note-Ending\">anime &amp; criticism<\/a> or heck <a href=\"https:\/\/www.gwern.net\/Death-Note-Anonymity\">anime &amp; statistics &amp; criticism<\/a>!). It is everything I felt worth writing for the past few years that didn\u2019t fit somewhere like Wikipedia or was already written\u2014\u201c\u2026I realised that I wanted to read about them what I myself knew. More than this\u2014what only I knew. Deprived of this possibility, I decided to write about them. Hence this book.\u201d<a href=\"#fn1\" class=\"footnote-ref\" id=\"fnref1\" role=\"doc-noteref\"><sup>1<\/sup><\/a> I never expected to write so much, but I discovered that once I had a hammer, nails were everywhere, and that <a href=\"https:\/\/en.wikipedia.org\/wiki\/Say%27s%20Law\" title=\"Wikipedia: Say's Law\">supply creates its own demand<\/a><a href=\"#fn2\" class=\"footnote-ref\" id=\"fnref2\" role=\"doc-noteref\"><sup>2<\/sup><\/a>. I believe that someone who has been well-educated will think of something worth writing at least once a week; to a surprising extent, this has been true. (I added ~130 documents to this repository over the first 3 years.) There are many benefits to keeping notes as they allow one to accumulate confirming and especially contradictory evidence<a href=\"#fn3\" class=\"footnote-ref\" id=\"fnref3\" role=\"doc-noteref\"><sup>3<\/sup><\/a>, and even drafts can be useful so you <a href=\"https:\/\/en.wikipedia.org\/wiki\/Don%27t%20Repeat%20Yourself\" title=\"Wikipedia: Don\u2019t Repeat Yourself\">Don\u2019t Repeat Yourself<\/a> or simply decently respect the opinions of mankind:<\/p><div class=\"epigraph\"><blockquote>\n<p>\u201cSpecial knowledge can be a terrible disadvantage if it leads you too far along a path you cannot explain anymore.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Brian%20Herbert\" title=\"Wikipedia: Brian Herbert\">Brian Herbert<\/a> (<em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Dune%3A%20House%20Harkonnen\" title=\"Wikipedia: Dune: House Harkonnen\">Dune: House Harkonnen<\/a><\/em>)<\/p>\n<\/blockquote><\/div><p>One of my personal interests is applying the idea of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Long%20Now\" title=\"Wikipedia: Long Now\">Long Now<\/a>. What and how do you write a personal site with the long-term in mind? We live most of our lives in the future, and the actuarial tables give me until the 2070-2080s, excluding any benefits from <a href=\"https:\/\/en.wikipedia.org\/wiki\/caloric%20restriction\" title=\"Wikipedia: caloric restriction\">caloric restriction<\/a>\/<a href=\"https:\/\/en.wikipedia.org\/wiki\/intermittent%20fasting\" title=\"Wikipedia: intermittent fasting\">intermittent fasting<\/a> or projects like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Strategies%20for%20Engineered%20Negligible%20Senescence\" title=\"Wikipedia: Strategies for Engineered Negligible Senescence\">SENS<\/a>. It is a common-place in science fiction<a href=\"#fn4\" class=\"footnote-ref\" id=\"fnref4\" role=\"doc-noteref\"><sup>4<\/sup><\/a> that longevity would cause widespread risk aversion. But on the other hand, it could do the opposite: the longer you live, the more long-shots you can afford to invest in. Someone with a timespan of 70 years has reason to protect against black swans\u2014but also time to look for them.<a href=\"#fn5\" class=\"footnote-ref\" id=\"fnref5\" role=\"doc-noteref\"><sup>5<\/sup><\/a> It\u2019s worth noting that old people make many short-term choices, as reflected in increased suicide rates and reduced investment in education or new hobbies, and this is not due solely to the ravages of age but the proximity of death\u2014the HIV-infected (but otherwise in perfect health) act similarly short-term.<a href=\"#fn6\" class=\"footnote-ref\" id=\"fnref6\" role=\"doc-noteref\"><sup>6<\/sup><\/a><\/p><p>What sort of writing could you create if you worked on it (be it ever so rarely) for the next 60 years? What could you do if you started <em>now<\/em>?<a href=\"#fn7\" class=\"footnote-ref\" id=\"fnref7\" role=\"doc-noteref\"><sup>7<\/sup><\/a><\/p><blockquote>\n<p>\"Of all the books I have delivered to the presses, none, I think, is as personal as the straggling collection mustered for this hodgepodge, precisely because it abounds in reflections and interpolations. Few things have happened to me, and I have read a great many. Or rather, few things have happened to me more worth remembering than Schopenhauer\u2019s thought or the music of England\u2019s words.<\/p>\n<p>A man sets himself the task of portraying the world. Through the years he peoples a space with images of provinces, kingdoms, mountains, bays, ships, islands, fishes, rooms, instruments, stars, horses, and people. Shortly before his death, he discovers that that patient labyrinth of lines traces the image of his face.\"<a href=\"#fn8\" class=\"footnote-ref\" id=\"fnref8\" role=\"doc-noteref\"><sup>8<\/sup><\/a><\/p>\n<\/blockquote><section id=\"long-site\" class=\"level2\"><div class=\"epigraph\"><blockquote>\n<p>\u201cThe Internet is self destructing paper. A place where anything written is soon destroyed by rapacious competition and the only preservation is to forever copy writing from sheet to sheet faster than they can burn. If it\u2019s worth writing, it\u2019s worth keeping. If it can be kept, it might be worth writing\u2026If you store your writing on a third party site like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Blogger\" title=\"Wikipedia: Blogger\">Blogger<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Livejournal\" title=\"Wikipedia: Livejournal\">Livejournal<\/a> or even on your own site, but in the complex format used by blog\/wiki software du jour you will <em>lose it forever<\/em> as soon as hypersonic wings of Internet labor flows direct people\u2019s energies elsewhere. For most information published on the Internet, perhaps that is <em>not a moment too soon<\/em>, but how can the muse of originality soar when immolating transience brushes every feather?\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Julian%20Assange\" title=\"Wikipedia: Julian Assange\">Julian Assange<\/a> (<a href=\"http:\/\/web.archive.org\/web\/20071020051936\/http:\/\/iq.org\/\">\u201cSelf destructing paper\u201d<\/a>, 5 December 2006)<\/p>\n<\/blockquote><\/div><p>Keeping the site running that long is a challenge, and leads to the recommendations for <a href=\"https:\/\/www.gwern.net\/Resilient-Haskell-Software\">Resilient Haskell Software<\/a>: 100% <a href=\"https:\/\/en.wikipedia.org\/wiki\/FLOSS\" title=\"Wikipedia: FLOSS\">FLOSS<\/a> software<a href=\"#fn9\" class=\"footnote-ref\" id=\"fnref9\" role=\"doc-noteref\"><sup>9<\/sup><\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Open%20format\" title=\"Wikipedia: Open format\">open standards<\/a> for data, <a href=\"http:\/\/catb.org\/~esr\/writings\/taoup\/html\/ch05s01.html\">textual<\/a> human-readability, avoiding external dependencies<a href=\"#fn10\" class=\"footnote-ref\" id=\"fnref10\" role=\"doc-noteref\"><sup>10<\/sup><\/a><a href=\"#fn11\" class=\"footnote-ref\" id=\"fnref11\" role=\"doc-noteref\"><sup>11<\/sup><\/a>, and staticness<a href=\"#fn12\" class=\"footnote-ref\" id=\"fnref12\" role=\"doc-noteref\"><sup>12<\/sup><\/a>.<\/p><p>Preserving the content is another challenge. Keeping the content in a <a href=\"https:\/\/en.wikipedia.org\/wiki\/DVCS\" title=\"Wikipedia: DVCS\">DVCS<\/a> like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Git%20%28software%29\" title=\"Wikipedia: Git (software)\">git<\/a> protects against file corruption and makes it easier to mirror the content; regular backups<a href=\"#fn13\" class=\"footnote-ref\" id=\"fnref13\" role=\"doc-noteref\"><sup>13<\/sup><\/a> help. I have taken additional measures: <a href=\"https:\/\/en.wikipedia.org\/wiki\/WebCitation\" title=\"Wikipedia: WebCitation\">WebCitation<\/a> has archived most pages and almost all external links; the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Internet%20Archive\" title=\"Wikipedia: Internet Archive\">Internet Archive<\/a> is also archiving pages &amp; external links<a href=\"#fn14\" class=\"footnote-ref\" id=\"fnref14\" role=\"doc-noteref\"><sup>14<\/sup><\/a>. (For details, read <a href=\"https:\/\/www.gwern.net\/Archiving-URLs\">Archiving URLs<\/a>.)<\/p><p>One could continue in this vein, devising ever more powerful &amp; robust storage methods (perhaps combine the DVCS with <a href=\"https:\/\/en.wikipedia.org\/wiki\/forward%20error%20correction\" title=\"Wikipedia: forward error correction\">forward error correction<\/a> through <a href=\"https:\/\/en.wikipedia.org\/wiki\/PAR2\" title=\"Wikipedia: PAR2\">PAR2<\/a>, a la <a href=\"http:\/\/lwn.net\/Articles\/380983\/\">bup<\/a>), but what is one to fill the storage with?<\/p><\/section><section id=\"long-content\" class=\"level2\"><div class=\"epigraph\"><blockquote>\n<p>\u201cWhat has been done, thought, written, or spoken is not culture; culture is only that fraction which is <em>remembered<\/em>.\u201d<\/p>\n<p>Gary Taylor (<a href=\"http:\/\/www.amazon.com\/Clock-Long-Now-Responsibility-Computer\/dp\/0465007805\/?tag=gwernnet-20\"><em>The Clock of the Long Now<\/em><\/a>; emphasis added)<a href=\"#fn15\" class=\"footnote-ref\" id=\"fnref15\" role=\"doc-noteref\"><sup>15<\/sup><\/a><\/p>\n<\/blockquote><\/div><p>\u2018Blog posts\u2019 might be the answer. But I have read blogs for many years and most blog posts are the triumph of the hare over the tortoise. They are meant to be read by a few people on a weekday in 2004 and never again, and are <a href=\"http:\/\/www.nytimes.com\/2009\/06\/07\/fashion\/07blogs.html\" title=\"Blogs Falling in an Empty Forest\">quickly<\/a> <a href=\"https:\/\/web.archive.org\/web\/20140126035203\/http:\/\/www.caslon.com.au\/weblogprofile1.htm\" title=\"blog statistics and demographics\">abandoned<\/a>\u2014and perhaps as Assange says, not a moment too soon. (But isn\u2019t that sad? Isn\u2019t it a terrible <a href=\"https:\/\/en.wikipedia.org\/wiki\/Rate%20of%20return\" title=\"Wikipedia: Rate of return\">ROI<\/a> for one\u2019s time?) On the other hand, the best blogs always seem to be building something: they are rough drafts\u2014works in progress<a href=\"#fn16\" class=\"footnote-ref\" id=\"fnref16\" role=\"doc-noteref\"><sup>16<\/sup><\/a>. So I did not wish to write a blog. Then what? More than just \u201cevergreen content\u201d, what would constitute <em>Long<\/em> Content as opposed to the existing culture of Short Content? How does one live in a Long Now sort of way?<a href=\"#fn17\" class=\"footnote-ref\" id=\"fnref17\" role=\"doc-noteref\"><sup>17<\/sup><\/a><\/p><blockquote>\n<p>It\u2019s shocking to find how many people do not believe they can learn, and how many more believe learning to be difficult. Muad\u2019Dib knew that every experience carries its lesson.<a href=\"#fn18\" class=\"footnote-ref\" id=\"fnref18\" role=\"doc-noteref\"><sup>18<\/sup><\/a><\/p>\n<\/blockquote><p>My answer is that one uses such a framework to work on projects that are too big to work on normally or too tedious. (Conscientiousness is often lacking online or in volunteer communities<a href=\"#fn19\" class=\"footnote-ref\" id=\"fnref19\" role=\"doc-noteref\"><sup>19<\/sup><\/a> and many useful things go undone.) Knowing your site <em>will<\/em> survive for decades to come gives you the mental wherewithal to tackle long-term tasks like gathering information for years, and such persistence can be useful<a href=\"#fn20\" class=\"footnote-ref\" id=\"fnref20\" role=\"doc-noteref\"><sup>20<\/sup><\/a>\u2014if one holds onto every glimmer of genius for years, then even the dullest person may look a bit like a genius himself<a href=\"#fn21\" class=\"footnote-ref\" id=\"fnref21\" role=\"doc-noteref\"><sup>21<\/sup><\/a>. (Even experienced professionals can only write at their peak for a few hours a day\u2014usually <a href=\"https:\/\/www.gwern.net\/Morning-writing\">first thing in the morning<\/a>, it seems.) Half the challenge of fighting procrastination is the <a href=\"http:\/\/lesswrong.com\/lw\/3kv\/working_hurts_less_than_procrastinating_we_fear\/\">pain of starting<\/a>\u2014I find when I actually get <a href=\"http:\/\/lesswrong.com\/lw\/3kv\/working_hurts_less_than_procrastinating_we_fear\/\">into the swing<\/a> of working on even dull tasks, it\u2019s not so bad. So this suggests a solution: never start. Merely have perpetual drafts, which one tweaks from time to time. And the rest takes care of itself. I have a few examples of this:<\/p><ol type=\"1\"><li><p><a href=\"https:\/\/www.gwern.net\/DNB-FAQ\">DNB FAQ<\/a>:<\/p>\n<p>When I read in <em>Wired<\/em> in 2008 that the obscure working memory exercise called dual n-back (DNB) had been found to increase IQ substantially, I was shocked. IQ is one of the most stubborn properties of one\u2019s mind, one of the most fragile<a href=\"#fn22\" class=\"footnote-ref\" id=\"fnref22\" role=\"doc-noteref\"><sup>22<\/sup><\/a>, the hardest to affect positively<a href=\"#fn23\" class=\"footnote-ref\" id=\"fnref23\" role=\"doc-noteref\"><sup>23<\/sup><\/a>, but also one of the most valuable traits one could have<a href=\"#fn24\" class=\"footnote-ref\" id=\"fnref24\" role=\"doc-noteref\"><sup>24<\/sup><\/a>; if the technique panned out, it would be <em>huge<\/em>. Unfortunately, DNB requires a major time investment (as in, half an hour daily); which would be a bargain\u2014if it delivers. So, to do DNB or not?<\/p>\n<p>Questions of great import like this are worth studying carefully. The wheels of academia grind exceeding slow, and only a fool expects unanimous answers from fields like psychology. Any attempt to answer the question \u2018is DNB worthwhile?\u2019 will require years and cover a breadth of material. This FAQ on DNB is my attempt to cover that breadth over those years.<\/p><\/li>\n<li><p><a href=\"https:\/\/www.gwern.net\/otaku\"><em>Neon Genesis Evangelion<\/em> notes<\/a> and <a href=\"https:\/\/www.gwern.net\/otaku-essay\">essay draft<\/a>:<\/p>\n<p>I have been discussing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Neon%20Genesis%20Evangelion\" title=\"Wikipedia: Neon Genesis Evangelion\">NGE<\/a> since 2004. The task of interpreting Eva is very difficult; the source works themselves are a major time-sink<a href=\"#fn25\" class=\"footnote-ref\" id=\"fnref25\" role=\"doc-noteref\"><sup>25<\/sup><\/a>, and there are thousands of primary, secondary, and tertiary works to consider\u2014personal essays, interviews, reviews, etc. The net effect is that many Eva fans \u2018know\u2019 certain things about Eva, such as <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/End%20of%20Evangelion\" title=\"Wikipedia: End of Evangelion\">End of Evangelion<\/a><\/em> not being a grand \u2018screw you\u2019 statement by Hideaki Anno or that the TV series was censored, but they no longer have <em>proof<\/em>. Because each fan remembers a different subset, they have irreconcilable interpretations. (Half the value of the page for me is having a place to store things I\u2019ve said in countless fora which I can eventually turn into something more systematic.)<\/p>\n<p>To compile claims from all those works, to dig up forgotten references, to scroll through microfilms, buy issues of defunct magazines\u2014all this is enough work to shatter <a href=\"https:\/\/en.wikipedia.org\/wiki\/Karoshi\" title=\"Wikipedia: Karoshi\">the heart<\/a> of the stoutest salaryman. Which is why I began years ago and expect not to finish for years to come. (Finishing by 2020 seems like a good <a href=\"http:\/\/predictionbook.com\/predictions\/1951\">prediction<\/a>.)<\/p><\/li>\n<li><p><a href=\"https:\/\/www.gwern.net\/fiction\/Cloud-Nine\"><em>Cloud Nine<\/em><\/a>: Years ago I was reading the papers of the economist <a href=\"https:\/\/en.wikipedia.org\/wiki\/Robin%20Hanson\" title=\"Wikipedia: Robin Hanson\">Robin Hanson<\/a>. I recommend his work highly; even if they are wrong, they are imaginative and some of the finest speculative fiction I have read. (Except they were non-fiction.) One night I had a dream in which I saw in a flash a medieval city run in part on Hansonian grounds; a <a href=\"https:\/\/en.wikipedia.org\/wiki\/steampunk\" title=\"Wikipedia: steampunk\">steampunk<\/a> version of his <a href=\"https:\/\/en.wikipedia.org\/wiki\/futarchy\" title=\"Wikipedia: futarchy\">futarchy<\/a>. A city must have another city as a rival, and soon I had remembered the strange \u201990s idea of <a href=\"https:\/\/en.wikipedia.org\/wiki\/assassination%20market\" title=\"Wikipedia: assassination market\">assassination market<\/a>s, which was easily tweaked to work in a medieval setting. Finally, between them, was one of my favorite proposals, Buckminster Fuller\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/cloud%20nine\" title=\"Wikipedia: cloud nine\">cloud nine<\/a> megastructure.<\/p>\n<p>I wrote several drafts but always lost them. Sad<a href=\"#fn26\" class=\"footnote-ref\" id=\"fnref26\" role=\"doc-noteref\"><sup>26<\/sup><\/a> and discouraged, I abandoned it for years. This fear leads straight into the next example.<\/p><\/li>\n<li><p>A Book reading list:<\/p>\n<p>Once, I didn\u2019t have to keep reading lists. I simply went to the school library shelf where I left off and grabbed the next book. But then I began reading harder books, and they would cite other books, and sometimes would even have horrifying lists of hundreds of other books I ought to read (\u2018bibliographies\u2019). I tried remembering the most important ones but quickly forgot. So I began keeping a book list on paper. I thought I would throw it away in a few months when I read them all, but somehow it kept growing and growing. I didn\u2019t trust computers to store it before<a href=\"#fn27\" class=\"footnote-ref\" id=\"fnref27\" role=\"doc-noteref\"><sup>27<\/sup><\/a>, but now I do, and it lives on in digital form (currently on <a href=\"https:\/\/www.gwern.net\/Links#websites\">Goodreads<\/a>\u2014because they have export functionality). With it, I can track how my interests evolved over time<a href=\"#fn28\" class=\"footnote-ref\" id=\"fnref28\" role=\"doc-noteref\"><sup>28<\/sup><\/a>, and what I was reading at the time. I sometimes wonder if I will read them all even by 2070.<\/p><\/li>\n<\/ol><p>What is next? So far the pages will persist through time, and they will gradually improve over time. But a truly Long Now approach would be to make them be improved <em>by<\/em> time\u2014make them more valuable the more time passes. (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Stewart%20Brand\" title=\"Wikipedia: Stewart Brand\">Stewart Brand<\/a> remarks in <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/The%20Clock%20of%20the%20Long%20Now\" title=\"Wikipedia: The Clock of the Long Now\">The Clock of the Long Now<\/a><\/em> that a group of monks carved thousands of scriptures into stone, hoping to preserve them for posterity\u2014but posterity would value far more a carefully preserved collection of monk feces, which would tell us countless valuable things about important phenomenon like global warming.)<\/p><p>One idea I am exploring is adding long-term predictions like the ones I make on <a href=\"http:\/\/predictionbook.com\/users\/gwern\">PredictionBook.com<\/a>. Many<a href=\"#fn29\" class=\"footnote-ref\" id=\"fnref29\" role=\"doc-noteref\"><sup>29<\/sup><\/a> pages explicitly or implicitly make predictions about the future. As time passes, predictions would be validated or falsified, providing feedback on the ideas.<a href=\"#fn30\" class=\"footnote-ref\" id=\"fnref30\" role=\"doc-noteref\"><sup>30<\/sup><\/a><\/p><p>For example, the Evangelion essay\u2019s paradigm implies many things about the future movies in <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Rebuild%20of%20Evangelion\" title=\"Wikipedia: Rebuild of Evangelion\">Rebuild of Evangelion<\/a><\/em><a href=\"#fn31\" class=\"footnote-ref\" id=\"fnref31\" role=\"doc-noteref\"><sup>31<\/sup><\/a>; <a href=\"https:\/\/www.gwern.net\/The-Melancholy-of-Kyon\">The Melancholy of Kyon<\/a> is an extended prediction<a href=\"#fn32\" class=\"footnote-ref\" id=\"fnref32\" role=\"doc-noteref\"><sup>32<\/sup><\/a> of future plot developments in <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/The%20Melancholy%20of%20Haruhi%20Suzumiya\" title=\"Wikipedia: The Melancholy of Haruhi Suzumiya\">The Melancholy of Haruhi Suzumiya<\/a><\/em> series; <a href=\"https:\/\/www.gwern.net\/haskell\/Summer-of-Code\">Haskell Summer of Code<\/a> has suggestions about what makes good projects, which could be turned into predictions by applying them to predict success or failure when the next Summer of Code choices are announced. And so on.<\/p><p>I don\u2019t think \u201cLong Content\u201d is simply for working on things which are equivalent to a \u201c<a href=\"https:\/\/en.wikipedia.org\/wiki\/monograph\" title=\"Wikipedia: monograph\">monograph<\/a>\u201d (a work which attempts to be an exhaustive exposition of all that is known\u2014and what has been recently discovered\u2014on a single topic), although monographs clearly would benefit from such an approach. If I write a short essay cynically remarking on, say, Al Gore and predicting he\u2019d sell out and registered some predictions and came back 20 years later to see how it worked out, I would consider this \u201cLong Content\u201d (it gets more interesting with time, as the predictions reach maturation); but one couldn\u2019t consider this a \u201cmonograph\u201d in any ordinary sense of the word.<\/p><p>One of the ironies of this approach is that as a <a href=\"https:\/\/en.wikipedia.org\/wiki\/transhumanist\" title=\"Wikipedia: transhumanist\">transhumanist<\/a>, I assign non-trivial probability to the world undergoing massive change during the 21st century due to any of a number of technologies such as artificial intelligence (such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/mind%20uploading\" title=\"Wikipedia: mind uploading\">mind uploading<\/a><a href=\"#fn33\" class=\"footnote-ref\" id=\"fnref33\" role=\"doc-noteref\"><sup>33<\/sup><\/a>) or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Molecular%20assembler\" title=\"Wikipedia: Molecular assembler\">nanotechnology<\/a>; yet here I am, planning as if I and the world were immortal.<\/p><p>I personally believe that one should \u201cthink Less Wrong and act Long Now\u201d, if you follow me. I diligently do my daily <a href=\"https:\/\/www.gwern.net\/Spaced-repetition\">spaced-repetition review<\/a> and n-backing; I carefully design my website and writings to last decades, actively think about how to write material that improves with time, and work on writings that will not be finished for years (if ever). It\u2019s a bit schizophrenic since both are totalized worldviews with drastically conflicting recommendations about where to invest my time. It\u2019s a case of high <a href=\"https:\/\/en.wikipedia.org\/wiki\/discount%20rates\" title=\"Wikipedia: discount rates\">discount rates<\/a> versus low discount rates; and one could fairly accuse me of committing the <a href=\"https:\/\/en.wikipedia.org\/wiki\/sunk%20cost%20fallacy\" title=\"Wikipedia: sunk cost fallacy\">sunk cost fallacy<\/a>, but then, I\u2019m not sure that <a href=\"https:\/\/www.gwern.net\/Sunk-cost\">sunk cost fallacy is a fallacy<\/a> (certainly, I have more to show for my wasted time than most people).<\/p><p>The Long Now views its proposals like the Clock and the Long Library and <a href=\"https:\/\/en.wikipedia.org\/wiki\/seedbanks\" title=\"Wikipedia: seedbanks\">seedbanks<\/a> as insurance\u2014in case the future turns out to be surprisingly <em>unsurprising<\/em>. I view these writings similarly. If <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ray%20Kurzweil\" title=\"Wikipedia: Ray Kurzweil\">Ray Kurzweil<\/a>\u2019s most ambitious predictions turn out right and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Technological%20singularity\" title=\"Wikipedia: Technological singularity\">Singularity<\/a> happens by 2050 or so, then much of my writings will be moot, but I will have all the benefits of said Singularity; if the Singularity never happens or ultimately pays off in a very disappointing way, then my writings will be valuable to me. By working on them, I hedge my bets.<\/p><\/section><section id=\"finding-my-ideas\" class=\"level2\"><p>To the extent I personally have any method for \u2018getting started\u2019 on writing something, it\u2019s to pay attention to anytime you find yourself thinking, \u201chow irritating that there\u2019s no good webpage\/Wikipedia article on <em>X<\/em>\u201d or \u201cI wonder if <em>Y<\/em>\u201d or \u201chas anyone done <em>Z<\/em>\u201d or \u201chuh, I just realized that <em>A<\/em>!\u201d or \u201cthis is the third time I\u2019ve had to explain this, jeez.\u201d<\/p><p>The DNB FAQ started because I was irritated people were repeating themselves on the dual n-back mailing list; the <a href=\"https:\/\/www.gwern.net\/Modafinil\">modafinil<\/a> article started because it was a pain to figure out where one <em>could<\/em> order modafinil; the trio of <em>Death Note<\/em> articles (<a href=\"https:\/\/www.gwern.net\/Death-Note-Anonymity\">Anonymity<\/a>, <a href=\"https:\/\/www.gwern.net\/Death-Note-Ending\">Ending<\/a>, <a href=\"https:\/\/www.gwern.net\/Death-Note-script\">Script<\/a>) all started because I had an amusing thought about information theory; the <a href=\"https:\/\/www.gwern.net\/Silk-Road\">Silk Road 1<\/a> page was commissioned after I growsed about how deeply sensationalist &amp; shallow &amp; ill-informed all the mainstream media articles on the Silk Road drug marketplace were (similarly for <a href=\"https:\/\/www.gwern.net\/Bitcoin-is-Worse-is-Better\">Bitcoin is Worse is Better<\/a>); my <a href=\"https:\/\/www.gwern.net\/Google-shutdowns\">Google survival analysis<\/a> was based on thinking it was a pity that Arthur\u2019s <em>Guardian<\/em> analysis was trivially &amp; fatally flawed; and so on and so forth.<\/p><p>None of these seems special to me. Anyone could\u2019ve compiled the DNB FAQ; anyone could\u2019ve kept a list of online pharmacies where one could buy modafinil; someone tried something similar to my Google shutdown analysis before me (and the fancier statistics were all standard tools). If I have done anything meritorious with them, it was perhaps simply putting more work into them than someone else would have; to <a href=\"http:\/\/www.smithsonianmag.com\/arts-culture\/Teller-Reveals-His-Secrets.html?c=y&amp;story=fullstory\" title=\"Teller Reveals His Secrets: The smaller, quieter half of the magician duo Penn &amp; Teller writes about how magicians manipulate the human mind\">quote Teller<\/a>:<\/p><blockquote>\n<p>\u201cI think you\u2019ll see what I mean if I teach you a few principles magicians employ when they want to alter your perceptions\u2026Make the secret a lot more trouble than the trick seems worth. You will be fooled by a trick if it involves more time, money and practice than you (or any other sane onlooker) would be willing to invest.\u201d<\/p>\n<p>\u201cMy partner, Penn, and I once produced 500 live cockroaches from a top hat on the desk of talk-show host David Letterman. To prepare this took weeks. We hired an entomologist who provided slow-moving, camera-friendly cockroaches (the kind from under your stove don\u2019t hang around for close-ups) and taught us to pick the bugs up without screaming like preadolescent girls. Then we built a secret compartment out of foam-core (one of the few materials cockroaches can\u2019t cling to) and worked out a devious routine for sneaking the compartment into the hat. More trouble than the trick was worth? To you, probably. But not to magicians.\u201d<\/p>\n<\/blockquote><p>Besides that, I think after a while writing\/research can be a virtuous circle or autocatalytic. If one were to look at <a href=\"https:\/\/www.gwern.net\/docs\/gwern.net-gitstats\/index.html\">my repo statistics<\/a>, you see that I haven\u2019t always been writing as much. What seems to happen is that as I write more:<\/p><ul><li><p>I learn more tools<\/p>\n<p>eg. I learned basic <a href=\"https:\/\/en.wikipedia.org\/wiki\/meta-analysis\" title=\"Wikipedia: meta-analysis\">meta-analysis<\/a> in R to answer what all the positive &amp; negative <a href=\"https:\/\/www.gwern.net\/DNB-FAQ\">n-back studies<\/a> <a href=\"https:\/\/www.gwern.net\/DNB-meta-analysis\">summed to<\/a>, but then I was able to use it for <a href=\"https:\/\/www.gwern.net\/Iodine#meta-analysis\">iodine<\/a>; I learned linear models for <a href=\"https:\/\/www.gwern.net\/hpmor#analysis\">analyzing <em>MoR<\/em> reviews<\/a> but now I can use them anywhere I want to, like in my <a href=\"https:\/\/www.gwern.net\/Touhou\">Touhou draft material<\/a>.<\/p>\n<p>The \u201cFeynman method\u201d has been facetiously described as \u201cfind a problem; think very hard; write down the answer\u201d, but <a href=\"http:\/\/alumni.media.mit.edu\/~cahn\/life\/gian-carlo-rota-10-lessons.html\" title=\"'Ten Lessons I Wish I Had Been Taught', Rota 1996\">Gian-Carlo Rota<\/a> gives the real one:<\/p>\n<blockquote>\n<p>Richard Feynman was fond of giving the following advice on how to be a genius. You have to keep a dozen of your favorite problems constantly present in your mind, although by and large they will lay in a dormant state. Every time you hear or read a new trick or a new result, test it against each of your twelve problems to see whether it helps. Every once in a while there will be a hit, and people will say: \u201cHow did he do it? He must be a genius!\u201d<\/p>\n<\/blockquote><\/li>\n<li><p>I internalize a habit of noticing interesting questions that flit across my brain<\/p>\n<p>eg. in March 2013 while meditating: \u201cI wonder if more doujin music gets released when unemployment goes up and people may have more spare time or fail to find jobs? Hey! That giant Touhou music torrent I downloaded, with its 45000 songs all tagged with release year, could probably answer that!\u201d (One could argue that these questions probably <em>should<\/em> be ignored and not investigated in depth\u2014Teller again\u2014nevertheless, this is how things work for me.)<\/p><\/li>\n<li><p>if you aren\u2019t writing, you\u2019ll ignore useful links or quotes; but if you stick them in small asides or footnotes as you notice them, eventually you\u2019ll have something bigger.<\/p>\n<p>I grab things I see on Google Alerts &amp; Scholar, Pubmed, Reddit, Hacker News, my RSS feeds, books I read, and note them somewhere until they amount to something. (An example would be my slowly accreting citations on <a href=\"https:\/\/www.gwern.net\/Embryo-selection#iqincome-bibliography\">IQ and economics<\/a>.)<\/p><\/li>\n<li><p>people leave comments, ping me on IRC, send me emails, or leave anonymous messages, all of which help<\/p>\n<p>Some examples of this come from my most popular page, on Silk Road 1:<\/p>\n<ol type=\"1\"><li>an anonymous message led me to <a href=\"https:\/\/www.gwern.net\/Silk-Road#a-mole\">investigate a vendor in depth and ponder the accusation leveled against them<\/a>; I wrote it up and gave my opinions and thus I got another short essay to add to my SR page which I would not have had otherwise (and I think there\u2019s a <a href=\"http:\/\/predictionbook.com\/predictions\/16142\">&lt;20% chance<\/a> that in a few years this will pay off and become a very interesting essay).<\/li>\n<li>CMU\u2019s Nicholas Christin, who <a href=\"http:\/\/arxiv.org\/abs\/1207.7139\" title=\"Traveling the Silk Road: A measurement analysis of a large anonymous online marketplace\">wrote a paper<\/a> by scraping SR for many months and giving all sorts of overall statistics, emailed me to point out I was citing inaccurate figures from the first version of his paper. I thanked him for the correction and while I was replying, mentioned I had a hard time believing his paper\u2019s claims about the extreme rarity of scams on SR as estimated through buyer feedback. After some back and forth and suggesting specific mechanisms how the estimates could be positively biased, he was able to check his database and confirmed that there was at least one very large omission of scams in the scraped data and there was probably a general undersampling; so now I have a more accurate feedback estimate for my SR page (important for estimating risk of ordering) and he said he\u2019ll acknowledge me in the\/a paper, which is nice.<\/li>\n<\/ol><\/li>\n<\/ul><\/section><section id=\"information-organizing\" class=\"level2\"><p>Occasionally people ask how I manage information and read things.<\/p><ol type=\"1\"><li><p>For quotes or facts which are very important, I employ <a href=\"https:\/\/www.gwern.net\/Spaced-repetition\">spaced repetition<\/a> by adding them to my Mnemosyne<\/p><\/li>\n<li><p>I keep web clippings in Evernotes; I also excerpt from research papers &amp; books, and miscellaneous sources. This is useful for targeted searches when I remember a fact but not where I learned it, and for storing things which I don\u2019t want to memorize but which have no logical home in my website or LW or elsewhere. It is also helpful for writing my <a href=\"https:\/\/www.gwern.net\/Book-reviews\">book reviews<\/a> and the <a href=\"https:\/\/tinyletter.com\/gwern\">monthly newsletter<\/a>, as I can read through my book excerpts to remind myself of the highlights and at the end of the month review clippings from papers\/webpages to find good things to reshare which I was too busy at the time to do so or was unsure of its importance. I don\u2019t make any use of more complex Evernote features.<\/p>\n<p>I periodically back up my Evernote using the <a href=\"http:\/\/nevernote.sourceforge.net\/\">Linux client Nixnote\u2019s<\/a> export feature. (I made sure there was a working export method before I began using Evernote, and use it only as long as Nixnote continues to work.)<\/p>\n<p>My workflow for dealing with PDFs, as of late 2014, is:<\/p>\n<ol type=\"1\"><li>if necessary, jailbreak the paper using Libgen or a university proxy, then upload a copy to Dropbox, named <code>year-author.pdf<\/code><\/li>\n<li>read the paper, making excerpts as I go<\/li>\n<li>store the metadata &amp; excerpts in Evernote, possibly copying over to my Google+ account if I think it\u2019s good or otherwise interesting<\/li>\n<li>if useful, integrate into <code>gwern.net<\/code> with its title\/year\/author metadata, adding a local fulltext copy if the paper had to be jailbroken, otherwise rely on my custom archiving setup to preserve the remote URL<\/li>\n<li>hence, any future searches for the filename \/ title \/ key contents should result in hits either in my Evernote or <code>gwern.net<\/code><\/li>\n<\/ol><\/li>\n<li><p>Web pages are archived &amp; backed up by <a href=\"https:\/\/www.gwern.net\/Archiving-URLs\">my custom archiving setup<\/a>. This is intended mostly for fixing dead links (eg to recover the fulltext of the original URL of an Evernote clipping).<\/p><\/li>\n<li><p>I don\u2019t have any special book reading techniques. For really good books I excerpt from each chapter and stick the quotes into Evernote.<\/p><\/li>\n<li><p>I store insights and thoughts in various pages as parenthetical comments, footnotes, and appendices. If they don\u2019t fit anywhere, I dump them in <a href=\"https:\/\/www.gwern.net\/Notes\">Notes<\/a>.<\/p><\/li>\n<li><p>Larger masses of citations and quotes typically get turned into pages.<\/p><\/li>\n<li><p>I make heavy use of RSS subscriptions for news. For that, I am currently using <a href=\"https:\/\/en.wikipedia.org\/wiki\/Liferea\" title=\"Wikipedia: Liferea\">Liferea<\/a>. (Not that I\u2019m hugely thrilled about it. Google Reader was much better.)<\/p><\/li>\n<li><p>For projects and followups, I use reminders in Google Calendar.<\/p><\/li>\n<li><p>For recording personal data, I automate as much as possible (eg Zeo and <a href=\"http:\/\/arbtt.nomeata.de\/\">arbtt<\/a>) and I make a habit of the rest\u2014getting up in the morning is a great time to build a habit of recording data because it\u2019s a time of habits like eating breakfast and getting dressed.<\/p><\/li>\n<\/ol><p>Hence, to refind information, I use a combination of Google, Evernote, <code>grep<\/code> (on the <code>gwern.net<\/code> files), occasionally Mnemosyne, and a good visual memory.<\/p><p>As far as writing goes, I do not use note-taking software or things like <a href=\"https:\/\/en.wikipedia.org\/wiki\/FreeMind\" title=\"Wikipedia: FreeMind\">FreeMind<\/a> or <a href=\"https:\/\/en.wikipedia.org\/wiki\/org-mode\" title=\"Wikipedia: org-mode\">org-mode<\/a>\u2014not that I think they are useless but I am worried about whether they would ever repay the large upfront investments of learning\/tweaking or interfere with other things. Instead, I occasionally compile outlines of articles from comments on LW\/Reddit\/IRC, keep editing them with stuff as I remember them, search for relevant parts, allow little thoughts to bubble up while meditating, and pay attention to when I am irritated at people being wrong or annoyed that a particular topic hasn\u2019t been written down yet.<\/p><\/section><section id=\"writing-checklist\" class=\"level2\"><p>It turns out that writing essays (technical or philosophical) is a lot like writing code\u2014there are so many ways to err that you need a process with as much automation as possible. My current checklist for finishing an essay:<\/p><ul><li><p>syntax:<\/p>\n<ul><li><p><a href=\"http:\/\/www.emacswiki.org\/emacs\/MarkdownMode#toc1\">balanced brackets &amp; quotes check<\/a><\/p><\/li>\n<li><p>do a Pandoc\/Firefox preview for visible major formatting problems<\/p><\/li>\n<li><p><a href=\"#markdown-checker\">Markdown lint-checker<\/a><\/p><\/li>\n<li><p><a href=\"http:\/\/wummel.github.io\/linkchecker\/\"><code>linkchecker<\/code><\/a> for<\/p>\n<ol type=\"1\"><li>dead links<\/li>\n<li>inserting into <a href=\"https:\/\/www.gwern.net\/Archiving-URLs\">archive queue<\/a><\/li>\n<\/ol><\/li>\n<\/ul><\/li>\n<li><p>references:<\/p>\n<ul><li>for academic hyperlinks, include a <a href=\"https:\/\/en.wikipedia.org\/wiki\/tooltip\" title=\"Wikipedia: tooltip\">tooltip<\/a> with the title &amp; author metadata<\/li>\n<li>for any papers cited: either link full text, provide a local full text, or submit a request on <a href=\"http:\/\/www.reddit.com\/r\/Scholar\/\">\/r\/Scholar<\/a> or <a href=\"http:\/\/lesswrong.com\/lw\/ji3\/lesswrong_help_desk_free_paper_downloads_and_more\/\" title=\"Free research help, editing and article downloads for LessWrong\">LessWrong<\/a>; for books, insert an Amazon link (one day I will be able to dare to host ebooks\u2026)<\/li>\n<\/ul><\/li>\n<li><p>language:<\/p>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ispell\" title=\"Wikipedia: Ispell\">spellcheck<\/a><\/li>\n<li>check readability level (eg. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Flesch-Kincaid\" title=\"Wikipedia: Flesch-Kincaid\">Flesch-Kincaid<\/a>)<\/li>\n<li><a href=\"#confidence-tags\">probability word checklist<\/a><\/li>\n<li>check for use of the word \u201csignificant\u201d\/\u201csignificance\u201d and insert \u201c[statistically]\u201d as appropriate (to disambiguate between <a href=\"https:\/\/en.wikipedia.org\/wiki\/effect%20sizes\" title=\"Wikipedia: effect sizes\">effect sizes<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/statistical%20significance\" title=\"Wikipedia: statistical significance\">statistical significance<\/a>; this common confusion is one reason for <a href=\"http:\/\/lesswrong.com\/lw\/g13\/against_nhst\/\" title=\"Against NHST\">\u201cstatistical-significance considered harmful\u201d<\/a>)<\/li>\n<li>convert English units to metric<\/li>\n<li><a href=\"https:\/\/github.com\/amperser\/proselint\/\">proselint<\/a> checks<\/li>\n<\/ul><\/li>\n<li><p>content:<\/p>\n<ul><li><p>mention any use of PredictionBook in <a href=\"https:\/\/www.gwern.net\/Prediction-markets\">my essay on forecasting &amp; prediction markets<\/a><\/p><\/li>\n<li><p>mention any use of Fermi estimates in <a href=\"https:\/\/www.gwern.net\/Notes#fermi-calculations\">Fermi calculations<\/a><\/p><\/li>\n<li><p>arrange for notifications of future results (when deciding to do long-term followups, eg dual n-back research):<\/p>\n<ol type=\"1\"><li><a href=\"http:\/\/www.google.com\/alerts\">Google Alerts<\/a><\/li>\n<li><a href=\"http:\/\/googlescholar.blogspot.com\/2010\/06\/google-scholar-alerts.html\">Google Scholar Alerts<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed\" title=\"Wikipedia: PubMed\">PubMed<\/a> alerts<\/li>\n<\/ol><\/li>\n<li><p>if using statistics:<\/p>\n<ul><li><p>rerun all code to verify reproducibility<\/p><\/li>\n<li><p>use reporting\/quality checklists:<\/p>\n<ul><li>cross-sectional &amp; other non-randomized analyses: <a href=\"http:\/\/www.plosmedicine.org\/article\/info%3Adoi%2F10.1371%2Fjournal.pmed.0040297\" title=\"'Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and Elaboration', Vandenbroucke et al 2007\">STROBE<\/a><\/li>\n<li>randomized experiments: <a href=\"http:\/\/annals.org\/data\/Journals\/AIM\/20207\/7TT1.jpeg\">CONSORT<\/a> (<a href=\"http:\/\/annals.org\/data\/Journals\/AIM\/20207\/0000605-201006010-00007.pdf\" title=\"'CONSORT 2010 Statement: Updated Guidelines for Reporting Parallel Group Randomized Trials', Schulz et al 2010\">statement<\/a>)<\/li>\n<li>meta-analyses: <a href=\"http:\/\/www.plosmedicine.org\/article\/info%3Adoi%2F10.1371%2Fjournal.pmed.1000097\" title=\"'Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement', Moher et al 2009\">PRISMA<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><p>after publication, publicize; currently:<\/p>\n<ol type=\"1\"><li>Hacker News<\/li>\n<li>Reddit<\/li>\n<li>LessWrong (and further sites as appropriate)<\/li>\n<li>Google+ (link to previous submissions)<\/li>\n<li>Twitter<\/li>\n<\/ol><\/li>\n<\/ul><section id=\"markdown-checker\" class=\"level3\"><h3><a href=\"#markdown-checker\" title=\"Link to section: 'Markdown checker'\">Markdown checker<\/a><\/h3><p>I\u2019ve noticed that when I make Markdown syntax errors, they tend to be predictable and show up either in the original Markdown source, or in the rendered HTML. Two common source errors:<\/p><pre>\"(www\"\n\")www\"<\/pre><p>And the following should <em>rarely<\/em> show up in the final rendered HTML:<\/p><pre>\"\\frac\"\n\"\\times\"\n\"(http\"\n\")http\"\n\"[http\"\n\"]http\"\n\" _ \"\n\"[^\"\n\"^]\"\n\"&lt;!--\"\n\"--&gt;\"\n\"&lt;-- \"\n\"&lt;-\"\n\"-&gt;\"\n\"$title$\"\n\"$description$\"\n\"$author$\"\n\"$tags$\"\n\"$category$\"<\/pre><p>Another static warning is checking for too-long lines which will cause browsers to use scrollbars, for which I\u2019ve written a <a href=\"https:\/\/www.gwern.net\/haskell\/markdown-length-checker.hs\">Pandoc script<\/a>, and one for a bad habit of mine\u2014<a href=\"https:\/\/www.gwern.net\/haskell\/markdown-footnote-length.hs\">too-long footnotes<\/a>. Combining checks for all these defects gives me a <a href=\"https:\/\/en.wikipedia.org\/wiki\/lint%20%28software%29\" title=\"Wikipedia: lint (software)\">\u201clint\u201d<\/a> shell script for individual Markdown files which checks the page source and then the rendered HTML in various ways (I suggest adding <code>--color=auto<\/code> to your default <code>grep<\/code> options).<\/p><p>The shell script can be found at <a href=\"https:\/\/www.gwern.net\/markdown-lint.sh\"><code>markdown-lint.sh<\/code><\/a>.<\/p><\/section><section id=\"anonymous-feedback\" class=\"level3\"><h3><a href=\"#anonymous-feedback\" title=\"Link to section: 'Anonymous feedback'\">Anonymous feedback<\/a><\/h3><p>Back in November 2011, lukeprog posted <a href=\"http:\/\/lesswrong.com\/lw\/8bt\/tell_me_what_you_think_of_me\/\">\u201cTell me what you think of me\u201d<\/a> where he described his use of a Google Docs form for anonymous receipt of textual feedback or comments. Typically, most forms of communication are non-anonymous, or if they are anonymous, they\u2019re public. One can set up pseudonyms and use those for private contact, but it\u2019s not always that easy, and is definitely a series of <a href=\"http:\/\/lesswrong.com\/lw\/f1\/beware_trivial_inconveniences\/\">trivial inconveniences<\/a> (if anonymous feedback is not solicited, one has to feel it\u2019s important enough to do and violate implicit norms against anonymous messages; one has to set up an identity; one has to compose and send off the message, etc).<\/p><p>I thought it was a good idea to try out, and on 8 November 2011, I set up my own anonymous feedback form and stuck it in the footer of all pages on <a href=\"https:\/\/www.gwern.net\/\">gwern.net<\/a> where it remains to this day. I did wonder if anyone would use the form, especially since I am easy to contact via email, use multiple sites like Reddit or Lesswrong, and even my Disqus comments allow anonymous comments\u2014so who, if anyone, would be using this form? I scheduled a followup in 2 years on 30 November 2013 to review how the form fared.<\/p><p>754 days, 2.884m page views, and 1.350m unique visitors later, I have received 116 pieces of feedback (mean of 24.8k visits per feedback). I categorize them as follows in descending order of frequency:<\/p><ul><li>Corrections, problems (technical or otherwise), suggested edits: 34<\/li>\n<li>Praise: 31<\/li>\n<li>Question\/request (personal, tech support, etc): 22<\/li>\n<li>Misc (eg gibberish, socializing, Japanese): 13<\/li>\n<li>Criticism: 9<\/li>\n<li>News\/suggestions: 5<\/li>\n<li>Feature request: 4<\/li>\n<li>Request for cybering: 1<\/li>\n<li>Extortion: 1 (see my <a href=\"https:\/\/www.gwern.net\/Blackmail#september\">blackmail page<\/a> dealing with the September 2013 incident)<\/li>\n<\/ul><p>Some submissions cover multiple angles (they can be quite long), sometimes people double-submitted or left it blank, etc, so the numbers won\u2019t sum to 116.<\/p><p>In general, a lot of the corrections were usable and fixed issues of varying importance, from typos to the entire site\u2019s CSS being broken due to being uploaded with the wrong MIME type. One of the news\/suggestion feedbacks was very valuable, as it lead to writing the Silk Road mini-essay <a href=\"https:\/\/www.gwern.net\/Silk-Road#a-mole\">\u201cA Mole?\u201d<\/a> A lot of the questions were a waste of my time; I\u2019d say half related to Tor\/Bitcoin\/Silk-Road. (I also got an irritating number of emails from people asking me to, say, buy LSD or heroin off SR for them.) The feature requests were usually for a better RSS feed, which I tried to oblige by starting the <a href=\"https:\/\/www.gwern.net\/Changelog\">Changelog<\/a> page. The cybering and extortion were amusing, if nothing else. The praise was good for me mentally, as I don\u2019t interact much with people.<\/p><p>I consider the anonymous feedback form to have been a success, I\u2019m glad lukeprog brought it up on LW, and I plan to keep the feedback form indefinitely.<\/p><section id=\"feedback-causes\" class=\"level4\"><h4><a href=\"#feedback-causes\" title=\"Link to section: 'Feedback causes'\">Feedback causes<\/a><\/h4><p>One thing I wondered is whether feedback was purely a function of traffic (the more visits, the more people who could see the link in the footer and decide to leave a comment), or more related to time (perhaps people returning regularly and eventually being emboldened or noticing something to comment on). So I compiled daily hits, combined with the feedback dates, and looked at a graph of hits:<\/p><figure><img alt=\"Hits over time for gwern.net\" height=\"385\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2013-11-30-gwernnet-hitsovertime.png\" width=\"754\" \/><figcaption>Hits over time for <code>gwern.net<\/code><\/figcaption><\/figure><p>The hits are heavily skewed by Hacker News &amp; Reddit traffic spikes, and probably should be log transformed. Then I did a logistic regression on hits, log hits, and a simple time index:<\/p><div class=\"sourceCode\" id=\"cb4\"><pre class=\"sourceCode r\"><a class=\"sourceLine\" id=\"cb4-1\" title=\"1\">feedback &lt;- read.csv(\"https:\/\/www.gwern.net\/docs\/traffic\/2013-gwernnet-anonymousfeedback.csv\",<\/a>\n<a class=\"sourceLine\" id=\"cb4-2\" title=\"2\">                     colClasses=c(\"Date\",\"logical\",\"integer\"))<\/a>\n<a class=\"sourceLine\" id=\"cb4-3\" title=\"3\">plot(Visits ~ Day, data=feedback)<\/a>\n<a class=\"sourceLine\" id=\"cb4-4\" title=\"4\">feedback$Time &lt;- 1:nrow(feedback)<\/a>\n<a class=\"sourceLine\" id=\"cb4-5\" title=\"5\">summary(step(glm(Feedback ~ log(Visits) + Visits + Time, family=binomial, data=feedback)))<\/a>\n<a class=\"sourceLine\" id=\"cb4-6\" title=\"6\"># ...<\/a>\n<a class=\"sourceLine\" id=\"cb4-7\" title=\"7\"># Coefficients:<\/a>\n<a class=\"sourceLine\" id=\"cb4-8\" title=\"8\">#              Estimate Std. Error z value Pr(&gt;|z|)<\/a>\n<a class=\"sourceLine\" id=\"cb4-9\" title=\"9\"># (Intercept) -7.363507   1.311703   -5.61  2.0e-08<\/a>\n<a class=\"sourceLine\" id=\"cb4-10\" title=\"10\"># log(Visits)  0.749730   0.173846    4.31  1.6e-05<\/a>\n<a class=\"sourceLine\" id=\"cb4-11\" title=\"11\"># Time        -0.000881   0.000569   -1.55     0.12<\/a>\n<a class=\"sourceLine\" id=\"cb4-12\" title=\"12\">#<\/a>\n<a class=\"sourceLine\" id=\"cb4-13\" title=\"13\"># (Dispersion parameter for binomial family taken to be 1)<\/a>\n<a class=\"sourceLine\" id=\"cb4-14\" title=\"14\">#<\/a>\n<a class=\"sourceLine\" id=\"cb4-15\" title=\"15\">#     Null deviance: 578.78  on 753  degrees of freedom<\/a>\n<a class=\"sourceLine\" id=\"cb4-16\" title=\"16\"># Residual deviance: 559.94  on 751  degrees of freedom<\/a>\n<a class=\"sourceLine\" id=\"cb4-17\" title=\"17\"># AIC: 565.9<\/a><\/pre><\/div><p>The logged hits works out better than regular hits, and survives to the simplified model. And the traffic influence seems much larger than the time variable (which is, curiously, negative).<\/p><\/section><\/section><\/section><\/section><section id=\"technical-aspects\" class=\"level1\"><section id=\"popularity\" class=\"level2\"><section id=\"october-2010february-2011\" class=\"level3\"><h3><a href=\"#october-2010february-2011\" title=\"Link to section: 'October 2010\u2013February 2011'\">October 2010\u2013February 2011<\/a><\/h3><p>My editing activity, as generated by <a href=\"https:\/\/hackage.haskell.org\/package\/darcs-graph\" title=\"Hackage: darcs-graph\">darcs-graph<\/a>: <\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): October to February\" height=\"186\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2011-february-darcs-history.png\" width=\"640\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): October to February<\/figcaption><\/figure><section id=\"traffic\" class=\"level4\"><h4><a href=\"#traffic\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><div class=\"epigraph\"><blockquote>\n<p>\u201cAn audience, even an audience of one, is always to be treasured and respected.\u201d<\/p>\n<p><a href=\"http:\/\/starwars.wikia.com\/wiki\/Adalric_Cessius_Brandl\">Adalric Brandl<\/a> (<a href=\"http:\/\/starwars.wikia.com\/wiki\/Uhl_Eharl_Khoehng_%28short_story%29\">\u201cUhl Eharl Khoehng\u201d<\/a>, Patricia A. Jackson)<\/p>\n<\/blockquote><\/div><p>Popularity-wise, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Google%20Analytics\" title=\"Wikipedia: Google Analytics\">Google Analytics<\/a> <a href=\"https:\/\/www.gwern.net\/docs\/traffic\/201010-201102-gwern.net-analytics.pdf\">reports that<\/a> over the 150 days between 1 October 2010 and 28 February 2011,  there were 4,346 page-views (average 30\/day):<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis)\" height=\"130\" src=\"https:\/\/www.gwern.net\/images\/traffic\/201010-201102-traffic-history.png\" width=\"1209\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis)<\/figcaption><\/figure><p>The most popular pages were<a href=\"#fn34\" class=\"footnote-ref\" id=\"fnref34\" role=\"doc-noteref\"><sup>34<\/sup><\/a>:<\/p><ol type=\"1\"><li><a href=\"https:\/\/www.gwern.net\/DNB-FAQ\">DNB FAQ<\/a>: 1,180<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Modafinil\">Modafinil<\/a>: 644<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/haskell\/Summer-of-Code\">Haskell Summer of Code<\/a>: 241<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Nootropics\">Nootropics experiments log<\/a>: 108<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/The-Melancholy-of-Kyon\">The Melancholy of Kyon<\/a>: 104<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Spaced-repetition\">Spaced repetition<\/a>: 101<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Links\">Links<\/a>: 96<\/li>\n<\/ol><p>The rankings are not as I would <em>prefer<\/em> (I imagine Internet archivist <a href=\"https:\/\/en.wikipedia.org\/wiki\/Jason%20Scott%20Sadofsky\" title=\"Wikipedia: Jason Scott Sadofsky\">Jason Scott Sadofsky<\/a> feels much the same way about <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sockington\" title=\"Wikipedia: Sockington\">Sockington<\/a>), but it\u2019s pretty clear that people enjoy my more practical articles the most.<\/p><\/section><\/section><section id=\"february-2011july-2011\" class=\"level3\"><h3><a href=\"#february-2011july-2011\" title=\"Link to section: 'February 2011\u2013July 2011'\">February 2011\u2013July 2011<\/a><\/h3><p><code>darcs-graph<\/code> for this period:<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): Repository creation to July\" height=\"182\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2011-july-darcs-history.png\" width=\"640\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): Repository creation to July<\/figcaption><\/figure><section id=\"traffic-1\" class=\"level4\"><h4><a href=\"#traffic-1\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><div class=\"epigraph\"><blockquote>\n<p>\u201cStreaming in the wind \/ the smoke from Fuji \/ vanishes in the sky; \/ I know not where \/ these thoughts of mine go, either.\u201d<\/p>\n<p>the monk <a href=\"https:\/\/en.wikipedia.org\/wiki\/Saigy%C3%B4\" title=\"Wikipedia: Saigy\u00f4\">Saigy\u00f4<\/a> (<em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Shin%20Kokin%20Wakashu\" title=\"Wikipedia: Shin Kokin Wakashu\">Shin Kokin Wakashu<\/a><\/em> XVII: <a href=\"http:\/\/www.temcauley.staff.shef.ac.uk\/waka1495.shtml\">#1615<\/a>)<\/p>\n<\/blockquote><\/div><p>Google Analytics <a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20110228-20110702-gwern.net-analytics.pdf\">reports that<\/a> over the 124 days between 28 February 2011 and 2 July 2011,  there were 42,410 page-views (average 342\/day):<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis)\" height=\"186\" src=\"https:\/\/www.gwern.net\/images\/traffic\/201102-201107-traffic-history.png\" width=\"1186\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis)<\/figcaption><\/figure><p>The most popular pages ranking changed considerably; while the DNB FAQ maintained its pre-eminent popularity, 3 new pages bumped out \u2018Links\u2019, \u2018Spaced repetition\u2019, \u2018The Melancholy of Kyon\u2019, and \u2018Haskell Summer of Code\u2019. I am a little surprised that my 2 <em>Death Note<\/em> essays seemed to\u2019ve struck a chord, and even more surprised that my sloppy &amp; random &amp; un-rigorous notes about nootropics would be consistently popular:<\/p><ol type=\"1\"><li>DNB FAQ: 10,406<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/index\">home\/main page<\/a>: 4,189<\/li>\n<li>Modafinil: 3,231<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Girl-Scouts-and-good-governance\">Girl Scouts and good governance<\/a>: 2,633<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Death-Note-Ending\"><em>Death Note<\/em>\u2019s Ending<\/a>: 1,779<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Death-Note-Anonymity\"><em>Death Note<\/em> Anonymity<\/a>: 1,366<\/li>\n<li>Nootropics: 2,056<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/haskell\/Archiving-GitHub\">Archiving GitHub<\/a>: 706<\/li>\n<\/ol><\/section><section id=\"promotion\" class=\"level4\"><h4><a href=\"#promotion\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><div class=\"epigraph\"><blockquote>\n<p>\u201cThey accumulate \/ but there are none to buy them\u2014 \/ these leaves of words \/ piling up like wares for sale \/ beneath the Sumiyoshi Pine.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sh%C5%8Dtetsu\" title=\"Wikipedia: Sh\u014dtetsu\">Sh\u014dtetsu<\/a>; (#180, \u201cFamous Market Town\u201d, <a href=\"https:\/\/www.gwern.net\/docs\/japanese\/1997-carter-shotetsu-unforgottendreams.pdf\" title=\"Poems by the Zen monk Sh\u014dtetsu\"><em>Unforgotten Dreams<\/em><\/a>; trans. Steven D. Carter)<a href=\"#fn35\" class=\"footnote-ref\" id=\"fnref35\" role=\"doc-noteref\"><sup>35<\/sup><\/a><\/p>\n<\/blockquote><\/div><p>As a writer, I desire feedback. I also want to feel that my work has been of use to people. So while it would be nice if the world beat a path to my website, I recognize that I have to put some effort into marketing my work. I\u2019ve tried a number of methods.<\/p><ol type=\"1\"><li><a href=\"https:\/\/en.bitcoin.it\/wiki\/Witcoin\">Witcoin<\/a>: I submitted any number of fairly popular articles but my total Witcoin traffic over this period was 132 visits\u2014a traffic total I could have gotten with one slightly popular link on Reddit or a few links in comments. While I didn\u2019t lose any Bitcoins (because my registration was funded by Kiba\u2019s donation of \u0e3f1) and actually profited \u0e3f2.77, I have spent at least 6 hours figuring out how to use Witcoin, submitting articles, responding to comments, and voting. Not the best use of time.<\/li>\n<li>Google <a href=\"https:\/\/en.wikipedia.org\/wiki\/AdWords\" title=\"Wikipedia: AdWords\">AdWords<\/a>: initially disappointing, with after 3010 impression, there were still no clicks! It was funded by the $100 coupon for signing up for Google\u2019s Webmaster Tools. Interface is decent given complexity of task, but deeply frustrating to have to wait many weeks for the DNB FAQ and Modafinil ads to be approved or rejected. Finally, almost in June, the DNB FAQ ads were approved and the modafinil ads rejected. From 23 March to 2 July 2011, I paid $21.07 for 98,900 impressions yielding 63 clicks through. (Those visitors only spent an average of &lt;1.5 minutes on the site, too.) Again, not a great investment of time.<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/StumbleUpon\" title=\"Wikipedia: StumbleUpon\">StumbleUpon<\/a>: with just 3 articles \u2018stumbled\u2019 (included in the database), specifically DNB FAQ, <a href=\"https:\/\/www.gwern.net\/In-Defense-Of-Inclusionism\">In Defense Of Inclusionism<\/a> &amp; Nootropics, StumbleUpon was responsible for 161 visits or 2.77% of all traffic in the period I looked at. How much traffic could I expect with 30 or 40 articles stumbled? Quite a bit. SU has no \u2018front page\u2019 like other social news aggregators so traffic is more of a trickle than flood; nevertheless, \u201c<em>Death Note<\/em> Ending\u201d clicked with SU readers and I got &gt;500 readers out of it in a day or two. In total over this time, SU drove 2,257 visits. SU tended to give a pretty steady 30\u201350 visits a day with rare spikes when an article clicked. The downside is that after looking at SU comments and at how much time they spend on pages<a href=\"#fn36\" class=\"footnote-ref\" id=\"fnref36\" role=\"doc-noteref\"><sup>36<\/sup><\/a>, I have to agree with Arvind Narayanan\u2019s <a href=\"http:\/\/arvindn.livejournal.com\/133249.html\">\u201cStumbleUpon Considered Harmful\u201d<\/a>\u2014SUers do not want quality content but quick content, for the dopamine boost.<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Hacker%20News\" title=\"Wikipedia: Hacker News\">Hacker News<\/a>: \u201cGirl Scouts and good governance\u201d made it to the front page, resulting in 1,727 visits &amp; setting <code>gwern.net<\/code> traffic records (it is that giant spike in the traffic graph), but apparently minimal viewing of other pages. Further, while I seem to get a modest amount of Reddit traffic from even unsuccessful submissions, HN submissions will sink without a trace. Kiba calls Hacker News a \u2018lottery\u2019, but it seems to be one worth playing.<\/li>\n<li>LessWrong is a natural place to <a href=\"http:\/\/lesswrong.com\/user\/gwern\/submitted\/\">post many of my writings<\/a>. And perhaps unsurprisingly, LW is my second-largest source of traffic, coming in after SU with 1,857 visits. While few of my submissions get upvoted all that highly, most of them drove a fair amount of traffic even in the Discussion ghetto. (Linking in comments also drives a surprising amount of traffic over long periods to my practical articles like on n-back or melatonin.) At some point I hope to have a good Article and see how much of a disparity there is.<\/li>\n<\/ol><\/section><\/section><section id=\"july-2011december-2011\" class=\"level3\"><h3><a href=\"#july-2011december-2011\" title=\"Link to section: 'July 2011\u2013December 2011'\">July 2011\u2013December 2011<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cRes audita perit, litera scripta manet.\u201d<\/p>\n<p>Proverb<\/p>\n<\/blockquote><\/div><p><code>darcs-graph<\/code> for this period (including 1 January 2012): <\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2011 to 2 January 2012\" height=\"181\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2011-december-darcs-history.png\" width=\"640\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2011 to 2 January 2012<\/figcaption><\/figure><p>I ran into a <a href=\"https:\/\/web.archive.org\/web\/20170112021012\/http:\/\/chrisdone.com\/posts\/programmer-at-work-statistics\" title=\"Programmer at work: statistics from a two year project (gitstats)\">cool post by Christopher Done<\/a> on a tool that does detailed analysis of patch patterns on a Git repository, <a href=\"http:\/\/gitstats.sourceforge.net\/\">GitStats<\/a>, and this spurred me to create a Git mirror of <code>gwern.net<\/code> using <a href=\"https:\/\/github.com\/purcell\/darcs-to-git\">darcs-to-git<\/a>. GitStats produces a <a href=\"https:\/\/www.gwern.net\/docs\/gwern.net-gitstats\/index.html\">whole bundle<\/a> of graphs and figures, some of which I found surprising. (I did not expect to see a large spike on Wednesday and relatively few patches on Saturday, or a spike around 5 PM, as opposed to the early morning.) I think I will update the GitStats output with each output, as a (large) adjunct to the <code>darcs-graph<\/code> plots.<\/p><section id=\"traffic-2\" class=\"level4\"><h4><a href=\"#traffic-2\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><div class=\"epigraph\"><blockquote>\n<p>\u201c\u2026prompt no more the follies you decry, \/ As tyrants doom their tools of guilt to die; \/ \u2019Tis yours this night to bid the reign commence \/ Of rescu\u2019d Nature, and reviving Sense; \/ \u2026Bid scenic Virtue form the rising age, \/ And Truth diffuse her radiance from the stage.\u201d<\/p>\n<p>Samuel Johnson (\u201cPrologue\u201d)<\/p>\n<\/blockquote><\/div><p>Google Analytics <a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20110702-20120102-gwern.net-analytics.pdf\">reports<\/a>, that over the 185 days between 2 July 2011 and 2 January 2012,  there were 191,015 page-views (average 1,032\/day) by 79,346 visitors for a total of 115,585 visits (average 624\/day). This is better than I expected and makes me wonder about <a href=\"http:\/\/predictionbook.com\/predictions\/4895\" title=\"gwern.net total traffic will average &lt;2000 daily visits in the week a year from now\">my prediction<\/a> for &lt;2000 average daily visits by 2013 (but it still seems unlikely traffic will triple over the next year).<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis)\" height=\"174\" src=\"https:\/\/www.gwern.net\/images\/traffic\/201107-201201-traffic-history.png\" width=\"1190\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis)<\/figcaption><\/figure><p>The main change in page popularity did not surprise me; when I was writing \u201cSilk Road 1\u201d, I knew it would almost certainly be popular given how very popular the Gawker article was but also how lacking in practical details it was, and I also suspected that \u201cBitcoin is Worse is Better\u201d would be fairly popular as it argued an interesting and controversial thesis (the original and promoted version is hosted on <code>BitcoinWeekly.com<\/code>, so its hit-count ought to be low). I\u2019m surprised at how much the page <code>gwern.net<\/code> still gets; a surprising number of people must either visit the main page after reading another article or click on my various blog comments.<\/p><ol type=\"1\"><li>Silk Road: 62,167<\/li>\n<li>DNB FAQ: 25,541<\/li>\n<li>home\/main page: 16,967<\/li>\n<li>Modafinil: 10,437<\/li>\n<li>Nootropics: 10,219<\/li>\n<li>Spaced repetition: 6,570<\/li>\n<li>Bitcoin is Worse is Better: 4,297<\/li>\n<li>Links: 3,490<\/li>\n<\/ol><p>More interesting is the other signals of popularity: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Zeo%20Inc.\" title=\"Wikipedia: Zeo Inc.\">Zeo Inc.<\/a> gave me a free set of headbands (worth ~$50) because they liked my <a href=\"https:\/\/www.gwern.net\/Zeo\">Zeo self-experiments<\/a>, a software engineer\/manager contacted me to see about recruiting me, ThinkGum offered me some of their eponymous product for my Nootropics page, and my request for Bitcoin donations has paid off a little with a few donations \u0e3f0.1\u20131 and one generous donation of \u0e3f20 (worth a bit upwards of $100 at the time; I spent it on modafinil). (This is all intrinsically helpful but I value it mostly because money speaks louder than words.)<\/p><\/section><section id=\"promotion-1\" class=\"level4\"><h4><a href=\"#promotion-1\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>I\u2019ve done relatively little in this period compared with the <a href=\"#promotion\">previous period<\/a>:<\/p><ol type=\"1\"><li>I abandoned Witcoin not long after my experiment with it; and now Witcoin is dead, pending a possible open-sourcing of the codebase.<\/li>\n<li>My AdWords credit is mostly expired. For some reason, my click-through rates kept dropping.<\/li>\n<li>StumbleUpon remains a good traffic source (1,430 visits). I continue to \u2018stumble\u2019 my new articles when I remember to do so.<\/li>\n<li>Hacker News was responsible for a great deal of my traffic in this period (4,175 visits). Most of it was not my doing, however\u2014whenever I submit links, they do poorly.<\/li>\n<li>LessWrong remains a major traffic driver (6,961 visits); I continue to see a lot of referrals from old posts and comments. Nor do all the links seem to be perceived negatively or as self-promotion by the LW community: I <a href=\"http:\/\/lesswrong.com\/lw\/8kv\/recent_updates_to_gwernnet\/\">posted an article<\/a> describing site updates and the article was received well to my surprise, eliciting very favorable reviews of my writings in general. That was nice.<\/li>\n<li>For this period, I did spend a little more effort submitting stuff to Reddit; and I was handsomely rewarded with the Silk Road submission skyrocketing and become one of the all-time most popular articles in the Bitcoin subreddit. Between that and my nootropics articles, Reddit sent me 20,842 visits.<\/li>\n<\/ol><p>The largest traffic sources are Google at 36,625 visits and direct\/no-referrals at 24,118 visits. As I have no idea how to improve these two figures, I ignore it. I write good content, submit it places, supply metadata, and abide by my hackerly principles; I hope that that is all the SEO I need.<\/p><\/section><\/section><section id=\"january-2012july-2012\" class=\"level3\"><h3><a href=\"#january-2012july-2012\" title=\"Link to section: 'January 2012\u2013July 2012'\">January 2012\u2013July 2012<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cUproot your questions from their ground and the dangling roots will be seen. More questions!\u201d<\/p>\n<p>Frank Herbert (\u201cMentat Zensufi admonition\u201d, <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Chapterhouse%20Dune\" title=\"Wikipedia: Chapterhouse Dune\">Chapterhouse Dune<\/a><\/em>)<\/p>\n<\/blockquote><\/div><p><code>darcs-graph<\/code> for this period (3 January 2012-2 July 2012): <\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January 2012 to 2 July 2012\" height=\"183\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2012-january-darcs-history.png\" width=\"640\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January 2012 to 2 July 2012<\/figcaption><\/figure><section id=\"traffic-3\" class=\"level4\"><h4><a href=\"#traffic-3\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><div class=\"epigraph\"><blockquote>\n<p>\u201cIf it were not for the intellectual snobs who pay - in solid cash - the tribute which philistinism owes to culture, the arts would perish with their starving practitioners. Let us thank heaven for hypocrisy.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Aldous%20Huxley\" title=\"Wikipedia: Aldous Huxley\">Aldous Huxley<\/a> (<a href=\"https:\/\/archive.org\/details\/jestingpilateint00huxl\/page\/156\">\u201cIndian &amp; Burma\u201d, <em>Jesting Pilate<\/em><\/a>, 1926)<\/p>\n<\/blockquote><\/div><p>The <a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20120103-20120702-gwern.net-analytics.pdf\">Analytics report<\/a> records traffic over those 182 days as being substantially increased: to 570,997 page-views (average 3,137\/day) - almost 5 times the previous six month period - by 268,031 unique visitors in 366,301 visits (average 2012\/day). If this rate continues, I will likely lose my traffic prediction (which doesn\u2019t bother me very much). In particular, the lifetime total page-views is now at 809,000, which would seem to imply that I will <a href=\"http:\/\/predictionbook.com\/predictions\/7457\">break the million page-view<\/a> mark by the next update! I would be very pleased by that milestone.<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis), early 2012\" height=\"186\" src=\"https:\/\/www.gwern.net\/images\/traffic\/201201-201207-traffic-history.png\" width=\"1319\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis), early 2012<\/figcaption><\/figure><p>The popularity ranking remains mainly the same. 2 differences from the past stand out: the sudden popularity of the Zeo and \u201cDeath Note Anonymity\u201d articles. Both owe their inclusion to 2 successful front-page appearances on Hacker News. DNA was submitted by someone I don\u2019t know, but I submitted the Zeo page at the conclusion of my first <a href=\"https:\/\/www.gwern.net\/Zeo#vitamin-d\">Vitamin D experiment<\/a> where I concluded that that Vitamin D consumed in the evening did indeed damage my sleep (I then followed up with <a href=\"https:\/\/www.gwern.net\/Zeo#vitamin-d-at-morn-helps\">a second experiment<\/a> which found that Vitamin D consumed in the morning did not damage my sleep.) I am proud of these two experiments and so I was gratified that Hacker News found them worthwhile too.<\/p><ol type=\"1\"><li>Silk Road: 308,895<\/li>\n<li>DNB FAQ: 36,488<\/li>\n<li>home\/main page: 32,259<\/li>\n<li>Modafinil: 23,801<\/li>\n<li>Nootropics: 21,970<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Death-Note-Anonymity\"><em>Death Note<\/em> Anonymity<\/a>: 15,860<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Zeo\">Zeo sleep self-experiments<\/a>: 15,841<\/li>\n<li>Spaced repetition: 8,628<\/li>\n<li>Links: 7,431<\/li>\n<\/ol><p>Donation-wise, I received ~\u0e3f2, and a number of Paypal donations: $5, $4, $5, and $100 from a particularly generous LessWronger who wished me to backup my files more securely &amp; remotely. Alexandra Carmichael was impressed by my Zeo experiments and asked me to write an ebook on sleep for an upcoming Quantified Self series of O\u2019Reilly ebooks; we will see how that goes.<\/p><\/section><section id=\"promotion-2\" class=\"level4\"><h4><a href=\"#promotion-2\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>Hacker News, Reddit, and LessWrong remain major referral drivers. For recent new pages, I\u2019ve been trying a checklist which includes submission to SU, Google+, Hacker News, Reddit, and LessWrong as appropriate; I am not sure how well it is working since popularity seems very random.<\/p><\/section><\/section><section id=\"july-2012january-2013\" class=\"level3\"><h3><a href=\"#july-2012january-2013\" title=\"Link to section: 'July 2012\u2013January 2013'\">July 2012\u2013January 2013<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cAll I say is by way of discourse, and nothing by way of advice. I should not speak so boldly if it were my due to be believed.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Michel%20de%20Montaigne\" title=\"Wikipedia: Michel de Montaigne\">Michel de Montaigne<\/a> (<a href=\"http:\/\/oll.libertyfund.org\/title\/1750\/91301\">\u201cOf Cripples\u201d<\/a>, <em>Essays<\/em>)<\/p>\n<\/blockquote><\/div><p><code>darcs-graph<\/code> for this period (3 July 2012-2 January 2013): <\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2012 to 2 January 2013\" height=\"171\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2012-july-darcs-history.png\" width=\"627\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2012 to 2 January 2013<\/figcaption><\/figure><section id=\"traffic-4\" class=\"level4\"><h4><a href=\"#traffic-4\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20120703-20130102-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 758,843 page views by 366,028 unique visitors over the 184 days for a daily average of 4,124.1 page-views, which is double the previous half-year average of 2,012 daily page-views; traffic growth is clearly slowing, though, since the previous half-year had quadruple the traffic compared to <em>its<\/em> predecessor. My prediction of breaking the million page-view mark came true, by a very large margin: the lifetime total page-views is now 1,568,957 page-views.<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis), late 2012\" height=\"281\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20120703-20130102-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis), late 2012<\/figcaption><\/figure><p>Popularity rankings have changed a bit: the <em>Death Note<\/em> essay and my sleep experiments have fallen out of the top 10 (the former because not many people are still linking it, and the latter probably because my latest experiments were relatively boring), replaced by a sidebar link and one of my terrorism-related essays:<\/p><ol type=\"1\"><li>Silk Road: 491,934<\/li>\n<li>home\/main page: 37,853<\/li>\n<li>Modafinil: 31,047<\/li>\n<li>DNB FAQ: 27,433<\/li>\n<li>Nootropics: 22,015<\/li>\n<li>Drug heuristics\/Algernon\u2019s Law: 18,991<\/li>\n<li>Spaced repetition: 11,693<\/li>\n<li>Links: 8,123<\/li>\n<li>About: 8,075<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Terrorism-is-not-about-Terror\">\u201cTerrorism is not about Terror\u201d<\/a>: 7,369<\/li>\n<\/ol><p>I am quite surprised that my <a href=\"https:\/\/www.gwern.net\/Slowing-Moores-Law\">Slowing Moore\u2019s Law<\/a> essay does not even make the top <em>50<\/em> pages, given that it deals with an novel thesis on which there\u2019s many interesting things to think and which is easily misunderstood.<\/p><p>Donations: the ebook fell through when O\u2019Reilly decided to cancel the entire series, which was a disappointment; Carmichael had finished her book on mood and is self-publishing, but I haven\u2019t seen tremendous interest in sleep and will probably just roll my draft material into the existing Zeo page. Paypal donations performed outstandingly: I received $10, $25, $10, $15, $200, &amp; $10 ($270). Bitcoiners were not so generous: \u0e3f0.25, \u0e3f0.32, &amp; \u0e3f1 (\u0e3f1.57, or $21 at the 2 January 2013 Mt.Gox exchange rate).<\/p><\/section><section id=\"promotion-3\" class=\"level4\"><h4><a href=\"#promotion-3\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>To Hacker News, Reddit, and LessWrong, I can add as a major referrer Wikipedia\u2014primarily to the Silk Road article, but also to a few <em>Evangelion<\/em>-related pages. StumbleUpon has declined to the 10th largest referrer:<\/p><ol type=\"1\"><li>HN: 16,886<\/li>\n<li>Reddit: 15,219<\/li>\n<li>Wikipedia: 7,531<\/li>\n<li>LessWrong: 6,332<\/li>\n<li>Facebook: 3,952<\/li>\n<li><code>brainworkshop.sourceforge.net<\/code>: 3,733<\/li>\n<li><code>google.com<\/code>: 2,649<\/li>\n<li><code>youtube.com<\/code>: 1,798<\/li>\n<li><code>mainstreamlos.tumblr.com<\/code>: 1,694<\/li>\n<li><code>stumbleupon.com<\/code>: 1,239<\/li>\n<\/ol><p>I haven\u2019t spent much time promoting my content, but improving the site with metadata &amp; writing new content: for example, I tripled the size of my <a href=\"https:\/\/www.gwern.net\/Hafu\">anime-manga hafu census<\/a> database.<\/p><\/section><\/section><section id=\"january-2013july-2013\" class=\"level3\"><h3><a href=\"#january-2013july-2013\" title=\"Link to section: 'January 2013\u2013July 2013'\">January 2013\u2013July 2013<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cI was decimated. To program any more would be pointless. My programs would never live as long as <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/The%20Trial\" title=\"Wikipedia: The Trial\">The Trial<\/a><\/em>. A computer will never live as long as <em>The Trial<\/em>. \u2026What if <a href=\"https:\/\/en.wikipedia.org\/wiki\/Amerika%20%28novel%29\" title=\"Wikipedia: Amerika (novel)\"><em>Amerika<\/em><\/a> was only written for <a href=\"https:\/\/en.wikipedia.org\/wiki\/PowerPC%2332-bit%20PowerPC\" title=\"Wikipedia: PowerPC#32-bit PowerPC\">32-bit PowerPC<\/a>?\u201d<\/p>\n<p>_why the lucky stiff (<a href=\"http:\/\/kevinw.github.io\/2013\/04\/30\/why-did-why-the-lucky-stiff-quit\/\" title=\"Why Did why the lucky stiff Quit? ...or, is this time well spent?\"><em>CLOSURE<\/em><\/a>)<\/p>\n<\/blockquote><\/div><p><code>darcs-graph<\/code> for this period (3 January 2013\u20132 July 2013): <\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January 2013 to 2 July 2013\" height=\"171\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2013-july-darcs-history.png\" width=\"628\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January 2013 to 2 July 2013<\/figcaption><\/figure><section id=\"traffic-5\" class=\"level4\"><h4><a href=\"#traffic-5\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20130103-20130702-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 832,415 page-views by 422,285 unique visitors over 181 days, or 4598 page-views per day. (The lifetime total page-views has thus reached 2,403,807.) This represents ~11% growth in traffic compared to the previous 6-month period of 4,124 dailies, continuing the slowdown trend\u2014possibly the next half-year won\u2019t see any growth, or a decline.<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis), early-mid 2013\" height=\"200\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20130103-20130702-traffic-history.png\" width=\"1302\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis), early-mid 2013<\/figcaption><\/figure><p>Popularity rankings have changed: Silk Road lost a lot of Wikipedia traffic when, during an editing dispute, an administrator deleted it from their article on Silk Road, and it was only restored on 2 July. 2 new statistical essays (<a href=\"https:\/\/www.gwern.net\/Death-Note-script\"><em>Death Note<\/em> script<\/a> &amp; <a href=\"https:\/\/www.gwern.net\/Google-shutdowns\">Google survival analysis<\/a>) enter the top 10, but neither were able to dethrone Silk Road as that darknet market continued to thrive &amp; receive media coverage:<\/p><ol type=\"1\"><li>Silk Road: 396,993<\/li>\n<li>Modafinil: 58,856<\/li>\n<li>Google shutdowns: 49,907<\/li>\n<li>DNB FAQ: 25,979<\/li>\n<li>Death Note script: 25,792<\/li>\n<li>Nootropics: 22,702<\/li>\n<li>home\/main page: 20,051<\/li>\n<li>Links: 16,789<\/li>\n<li>About: 14,746<\/li>\n<li>Spaced repetition: 11,711<\/li>\n<\/ol><p>Financial: this period was a remarkable period. My contracting work dried up (largely my own fault) and pressed for money, I began exploring alternatives:<\/p><ol type=\"1\"><li>the easiest strategy was to turn all my Amazon links for books &amp; nootropics &amp; miscellaneous into affiliate links; they are already there, do not affect readers, and coding it up was as simple as appending a string to links in <code>hakyll.hs<\/code> by a simple modification of my existing code for easily linking to Wikipedia. This worked as far as it went, which was not very far ($47 in Q1 2013 and $130 Q2 2013).<\/li>\n<li>more painfully, I decided to try Google\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/AdSense\" title=\"Wikipedia: AdSense\">AdSense<\/a>. While AdWords hadn\u2019t been the most pleasant experience in the world, it was fairly decent, and Google\u2019s ads always seemed reasonable to me in the search engine (before I learned of AdBlock). AdSense has scary language in its terms of service forbidding detailed discussion of revenue, but I should be safe when I say that the mean CTR was 0.17% &amp; the mean CPC was $0.57, ~40% of visitors did <em>not<\/em> have ad filtering enabled, and so over the 80 days AdSense was enabled, I earned ~$260. Helpful, but not rent-paying.<\/li>\n<li>donations made the difference. After \u201cGoogle shutdowns\u201d <a href=\"https:\/\/news.ycombinator.com\/item?id=5653748\">hit the front page<\/a> of Hacker News, an acquaintance <a href=\"https:\/\/news.ycombinator.com\/item?id=5654008\">mentioned I could use some money<\/a>. <a href=\"https:\/\/news.ycombinator.com\/item?id=5653874\">I received<\/a> $408 in 14 Paypal donations; through Bitcoin, 1.62btc (then $189) in 4 donations (<a href=\"https:\/\/news.ycombinator.com\/item?id=5660220\">followup<\/a>). I am grateful to all the HNers who donated.<\/li>\n<li>Related to that, I was contacted in May by <a href=\"http:\/\/www.sgenomics.co.uk\/\">Nava Whiteford<\/a> with an offer: he had heard my financial target of $300\/month and offered that (in bitcoins) in exchange for a sponsorship\/banner linking to his graphical terminal <a href=\"http:\/\/41j.com\/hterm\/\">HTerm<\/a>\u2014a combination purchase\/donation. This was both larger &amp; more stable than AdSense, so I accepted, and 3 months have thus far passed.<\/li>\n<\/ol><\/section><section id=\"promotion-4\" class=\"level4\"><h4><a href=\"#promotion-4\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>Wikipedia declines in this period due to the aforementioned removal, and more obscure sites become chief referrers:<\/p><ol type=\"1\"><li><code>news.ycombinator.com<\/code>: 45,186<\/li>\n<li><code>reddit.com<\/code>: 34,225<\/li>\n<li><code>en.wikipedia.org<\/code>: 16,527<\/li>\n<li><code>lesswrong.com<\/code>: 11,177<\/li>\n<li>Twitter: 6,717<\/li>\n<li><code>facebook.com<\/code>: 5,675<\/li>\n<li><code>mainstreamlos.tumblr.com<\/code>: 5,576<\/li>\n<li><code>boingboing.net<\/code>: 4,982<\/li>\n<li><code>motherboard.vice.com<\/code>: 4,496<\/li>\n<li><code>bulletproofexec.com<\/code>: 4,336<\/li>\n<\/ol><\/section><\/section><section id=\"july-2013january-2014\" class=\"level3\"><h3><a href=\"#july-2013january-2014\" title=\"Link to section: 'July 2013\u2013January 2014'\">July 2013\u2013January 2014<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cThe great globe reels in the solar fire, \/ Spinning the trivial and unique away. \/ (How all things flash! How all things flare!) \/ \u2026Time is the school in which we learn \/ Time is the fire in which we burn\u2026\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Delmore%20Schwartz\" title=\"Wikipedia: Delmore Schwartz\">Delmore Schwartz<\/a> (<a href=\"http:\/\/www.poetryfoundation.org\/poem\/171344\">\u201cCalmly We Walk Through This April\u2019s Day\u201d<\/a>)<\/p>\n<\/blockquote><\/div><p><code>darcs-graph<\/code> for this period (3 July 2013\u20132 January 2014): <\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2013 to 2 January 2014\" height=\"171\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2014-january-darcs-history.png\" width=\"626\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2013 to 2 January 2014<\/figcaption><\/figure><section id=\"traffic-6\" class=\"level4\"><h4><a href=\"#traffic-6\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20130703-20140102-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 770,264 page-views by 340,104 unique visitors over 184 days, or 4186 page-views per day. (Lifetime total: 3,173,172 page-views by 1,472,606 unique visitors.) Since the previous period was 4598\/day, that means traffic fell by ~9%. (Traffic would have fallen even more if I hadn\u2019t happened to be engaged in a successful campaign of Hacker News submissions at the time.) My guess is that this is largely attributable to the fall of Silk Road in early October rendering my page of much less interest, and that was one of my most popular pages. It was inevitable, really, but I had hoped SR would last more than &lt;3 years. Traffic may fall or increase during the next period.<\/p><figure><img alt=\"Plot of page-hits (y-axis) versus date (x-axis), mid-late 2013\" height=\"193\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20130703-20140102-traffic-history.png\" width=\"1320\" \/><figcaption>Plot of page-hits (y-axis) versus date (x-axis), mid-late 2013<\/figcaption><\/figure><p>This rebalancing is evident in the traffic rankings, where SR is still #1 and bigger than #2 (Modafinil), but by a much smaller factor this time (1.26x rather than 6.66x):<\/p><ol type=\"1\"><li>Silk Road: 80,233<\/li>\n<li>Modafinil: 63,966<\/li>\n<li>home\/main page: 65,203<\/li>\n<li>Melatonin: 50,952<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Blackmail\">Blackmail attempts<\/a>: 36,573<\/li>\n<li>Spaced repetition: 35768<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/LSD-microdosing\">LSD microdosing self-experiment<\/a>: 26,825<\/li>\n<li>Nicotine: 25,370<\/li>\n<li>DNB FAQ: 23,169<\/li>\n<li>Nootropics: 20,883<\/li>\n<\/ol><p>My finances continued to improve in part due to Whiteford, additional donations (I explored Flattr &amp; Gratipay; the latter seems to be working out better), another more-targeted banner on the modafinil page, and a major appreciation in Bitcoin. (Takes a real load off me.) I thank all my donators. I think the results show in my <a href=\"https:\/\/www.gwern.net\/Changelog\">Changelog<\/a>.<\/p><\/section><section id=\"promotion-5\" class=\"level4\"><h4><a href=\"#promotion-5\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><ol type=\"1\"><li><code>reddit.com<\/code>: 148,940<\/li>\n<li><code>news.ycombinator.com<\/code>: 84,163<\/li>\n<li><code>lesswrong.com<\/code>: 41,511<\/li>\n<li><code>en.wikipedia.org<\/code>: 31,967<\/li>\n<li><code>facebook.com<\/code>: 20,394<\/li>\n<li><code>t.co<\/code> [Twitter]: 19,628<\/li>\n<li><code>brainworkshop.sourceforge.net<\/code>: 16,808<\/li>\n<li><code>bulletproofexec.com<\/code>: 11,135<\/li>\n<li><code>google.com<\/code> [Google+?]: 10,154<\/li>\n<li><code>mainstreamlos.tumblr.com<\/code>: 10,128<\/li>\n<\/ol><p>Interestingly, despite my Hacker News experiment (which resulted in dozens of my pages reaching its main page), <code>news.ycombinator.com<\/code> is <em>still<\/em> beaten out traffic-wise by Reddit, by a huge factor (almost double).<\/p><\/section><\/section><section id=\"januaryjuly-2014\" class=\"level3\"><h3><a href=\"#januaryjuly-2014\" title=\"Link to section: 'January\u2013July 2014'\">January\u2013July 2014<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cA courteous handsome young man \/ well-versed in the Classics and Histories \/ people address him <em>sir<\/em> \/ everyone calls him a scholar \/ but he hasn\u2019t found a position yet \/ and he doesn\u2019t know how to farm \/ in winter he wears a tattered robe \/ this is how books fool us\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Han-Shan\" title=\"Wikipedia: Han-Shan\">Han-Shan<\/a> (#28, <a href=\"http:\/\/www.amazon.com\/Collected-Mountain-Mandarin-Chinese-English\/dp\/1556591403?tag=gwernnet-20\"><em>The Collected Songs of Cold Mountain<\/em><\/a>)<\/p>\n<\/blockquote><\/div><p>Activity graph<a href=\"#fn37\" class=\"footnote-ref\" id=\"fnref37\" role=\"doc-noteref\"><sup>37<\/sup><\/a> for this period (3 January 2014\u20132 July 2014):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January to July 2014\" height=\"751\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2014-july-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January to July 2014<\/figcaption><\/figure><section id=\"traffic-7\" class=\"level4\"><h4><a href=\"#traffic-7\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20140103-20140702-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 370,592 page-views by 162,032 unique visitors over 181 days, or 2047 page-views per day. (Lifetime total: 3,544,489 page-views by 1,628,173 unique visitors.) Since the previous period was 4186\/day, that means traffic fell by a disastrous ~50%, defeating my hopes that traffic might recover post-Silk-Road &amp; Hacker News experiment<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), early-mid 2014\" height=\"179\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20140103-20140702-traffic-history.png\" width=\"1286\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), early-mid 2014<\/figcaption><\/figure><p>The traffic rankings show a continue in the plummet of SR, and its dethroning by Modafinil and even the index page, while hosting some early Satoshi Nakamoto drew considerable interest &amp; my <a href=\"https:\/\/www.gwern.net\/Mistakes\">\u201cIntellectual Mistakes I Have Made\u201d<\/a> page was surprisingly linked on Tyler Cowen\u2019s <a href=\"http:\/\/marginalrevolution.com\/\">Marginal Revolution<\/a> blog:<\/p><ol type=\"1\"><li>Modafinil: 54,992<\/li>\n<li>home\/main page: 43,984<\/li>\n<li>Silk Road: 27,998<\/li>\n<li>Spaced repetition: 27,680<\/li>\n<li>DNB FAQ: 17,833<\/li>\n<li>Nootropics: 15,935<\/li>\n<li>Mistakes: 15,267<\/li>\n<li>Nicotine: 11,936<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/docs\/bitcoin\/2008-nakamoto\">Wei Dai\/Satoshi Nakamoto emails<\/a>: 9,827<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Links\">Links<\/a>: 9,391<\/li>\n<\/ol><p>On the plus side, it\u2019s hard to see how traffic could fall much more.<\/p><p>Finance-wise, Gratipay has worked out much better than I expected. I thank all my donators.<\/p><\/section><section id=\"promotion-6\" class=\"level4\"><h4><a href=\"#promotion-6\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>Hacker News turns out to no longer show up in the referrals analytics due to stripping referrals (apparently related to HTTPS use?), which is unfortunate; I suspect other sites are also \u201cgoing dark\u201d, given that the numbers are more than halving:<\/p><ol type=\"1\"><li><code>reddit.com<\/code>: 24,176<\/li>\n<li><code>marginalrevolution.com<\/code>: 5,695<\/li>\n<li><code>bulletproofexec.com<\/code>: 5,504<\/li>\n<li><code>disqus.com<\/code>: 4,051<\/li>\n<li><code>lesswrong.com<\/code>: 3,954<\/li>\n<li><code>longecity.org<\/code>: 3,051<\/li>\n<li>Twitter: 2,771<\/li>\n<li><code>boldanddetermined.com<\/code>: 2,54<\/li>\n<li><code>vice.com<\/code>: 1,812<\/li>\n<li><code>brainworkshop.sourceforge.net<\/code>: 1,625<\/li>\n<\/ol><p>I also began <a href=\"https:\/\/tinyletter.com\/gwern\">a monthly mailing list<\/a>; with 300 people on it, it doesn\u2019t seem to make a big difference to traffic but it\u2019s a good way to curate my links and non-<code>gwern.net<\/code> writings, and may be quite useful.<\/p><\/section><\/section><section id=\"july-2014january-2015\" class=\"level3\"><h3><a href=\"#july-2014january-2015\" title=\"Link to section: 'July 2014\u2013January 2015'\">July 2014\u2013January 2015<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cWriting is in itself a joy, \/ Yet saints and sages have long since held it in awe.\u201d<\/p>\n<p>\u201cFor it is being, created from a void; \/ It is sound rung out of profound silence. \/ In a sheet of paper is contained the infinite, \/ And, evolved from an inch-sized heart, an endless panorama.\u201d<\/p>\n<p>Lu Chi\u2019s <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Wen%20fu\" title=\"Wikipedia: Wen fu\">Wen fu<\/a><\/em> (section 4, <a href=\"http:\/\/web.mnstate.edu\/gracyk\/courses\/web%20publishing\/LuChi.htm\">\u2018The Joy of Writing\u2019, of \u201cThe Art of Writing\u201d<\/a>).<\/p>\n<\/blockquote><\/div><p>Activity graph<a href=\"#fn38\" class=\"footnote-ref\" id=\"fnref38\" role=\"doc-noteref\"><sup>38<\/sup><\/a> for this period (3 July 2014\u20132 January 2015):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2014 to January 2015\" height=\"872\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2015-january-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2014 to January 2015<\/figcaption><\/figure><section id=\"traffic-8\" class=\"level4\"><h4><a href=\"#traffic-8\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20140703-20150103-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 347,482 page-views by 154,599 unique visitors over the 184 days, or 1888 page-views per day for a daily drop of 159 compared to the previous six months. (Lifetime total: 3,891,971 page-views by 1,778,197 unique visitors.) This drop is not nearly as bad before, and may represent a bottom.<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), late 2014\" height=\"159\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20140703-20150103-traffic-history.png\" width=\"1274\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), late 2014<\/figcaption><\/figure><p>Page rankings:<\/p><ol type=\"1\"><li>Modafinil: 44,420<\/li>\n<li>home\/main page: 41,499<\/li>\n<li>Nootropics: 41,499<\/li>\n<li>Silk Road: 21,863<\/li>\n<li>Spaced repetition: 17,620<\/li>\n<li>Causality: 15,396<\/li>\n<li>DNB FAQ: 14,463<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Self-decrypting-files\">Time-lock crypto<\/a>: 11,313<\/li>\n<li>Nicotine: 9,804<\/li>\n<li>Links: 8,823<\/li>\n<\/ol><p>Silk Road declines a little more; the Wei Dai\/Satoshi Nakamoto emails &amp; Mistakes pages drop out after their moments in the limelight; and we see 2 new entries: I wrote a statistical essay on quantifying how often correlation\u2260causality (a topic of burning importance in almost any health-related research), and an older cryptography essay became abruptly more relevant when one of my proposals was implemented using Bitcoin.<\/p><p>Gratipay continues to work well, and I am grateful for all my donators.<\/p><\/section><section id=\"promotion-7\" class=\"level4\"><h4><a href=\"#promotion-7\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>I did minimal active promotion of my content, letting my AdWords credit be used up and not spending much effort on Reddit or Hacker News,<\/p><p>Confusingly, HN is back in my referrals:<\/p><ol type=\"1\"><li><code>reddit.com<\/code>: 27,182<\/li>\n<li><code>news.ycombinator.com<\/code>: 18,420<\/li>\n<li><code>lesswrong.com<\/code>: 4,583<\/li>\n<li>Twitter: 3,559<\/li>\n<li><a href=\"http:\/\/slatestarcodex.com\/\"><code>slatestarcodex.com<\/code><\/a>: 2,327<\/li>\n<li>Longecity: 2,143<\/li>\n<li><code>bulletproofexec.com<\/code>: 2,123<\/li>\n<li><code>boldanddetermined.com<\/code>: 1,488<\/li>\n<li><code>brainworkshop.sourceforge.net<\/code>: 1,346<\/li>\n<li><code>disqus.com<\/code>: 1,276<\/li>\n<\/ol><p>My <a href=\"https:\/\/tinyletter.com\/gwern\">monthly mailing list<\/a> continued, having reached 456 subscribers.<\/p><\/section><\/section><section id=\"january-2015july-2015\" class=\"level3\"><h3><a href=\"#january-2015july-2015\" title=\"Link to section: 'January 2015\u2013July 2015'\">January 2015\u2013July 2015<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u2018\u201cThe horror of that moment,\u201d the King went on, \u201cI shall never <em>never<\/em> forget!\u201d \u201cYou will, though\u201d, the Queen said, \u201cif you don\u2019t make a memorandum of it.\u201d\u2019<\/p>\n<p>Lewis Carroll (<em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Through%20the%20Looking-Glass%2C%20and%20What%20Alice%20Found%20There\" title=\"Wikipedia: Through the Looking-Glass, and What Alice Found There\">Through the Looking-Glass, and What Alice Found There<\/a><\/em>)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (3 January 2015\u20132 July 2015):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January 2015 to July 2015\" height=\"900\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2015-july-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January 2015 to July 2015<\/figcaption><\/figure><section id=\"traffic-9\" class=\"level4\"><h4><a href=\"#traffic-9\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20140703-20150103-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 378,580 page-views by 175,303 unique visitors over the 181 days, or 2091 page-views per day for a daily increase of +203 compared to the previous six months. (Lifetime total: 4,270,551 page-views by 1,949,700 unique visitors.) This is better than the previous drop but may be due mostly to some additional spikes from HN.<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), early 2015\" height=\"178\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20150103-20150702-traffic-history.png\" width=\"1276\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), early 2015<\/figcaption><\/figure><p>Page rankings:<\/p><ol type=\"1\"><li>Modafinil: 48,259<\/li>\n<li>home\/main page: 48,614<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/DNM-arrests\">Darknet Market (DNM) arrests<\/a>: 20,312<\/li>\n<li>Spaced repetition: 17,522<\/li>\n<li>DNB FAQ: 16,076<\/li>\n<li>Silk Road: 15,116<\/li>\n<li>LSD microdosing: 13,673<\/li>\n<li>Nicotine: 9,804<\/li>\n<li>Links: 9,234<\/li>\n<li><em>Death Note<\/em> Anonymity: 9,234<\/li>\n<\/ol><p>The new entry here, \u201cDNM arrests\u201d, is a long-term project in which I compile all the several hundred known and inferable arrests\/prosecutions related to Silk Road 1 and its successor DNMs; it can be very difficult to find details for individual cases (often requiring hours reading court documents for a key detail) or keep track of them, and only recently was I able to clear out my backlog and present it to the community. But it\u2019s useful in giving people an authoritative idea of the risks and what people are arrested for, and I\u2019m pleased to see that it\u2019s being read &amp; consulted. \u201cLSD microdosing\u201d seems to be getting more traffic due to some recent media coverage of the idea and people linking to my writeup as (still) the only blinded self-experiment on the topic.<\/p><p>Gratipay, unfortunately, has been forced to change due to financial regulations and is no longer useful for donations. It had a good run but I will probably be switching to Patreon.<\/p><\/section><section id=\"promotion-8\" class=\"level4\"><h4><a href=\"#promotion-8\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>As before, I did minimal active promotion of my content.<\/p><ol type=\"1\"><li><code>reddit.com<\/code>: 27,783<\/li>\n<li><code>news.ycombinator.com<\/code>: 11,538<\/li>\n<li><code>lesswrong.com<\/code>: 3,264<\/li>\n<li>Twitter (<code>t.co<\/code>): 2,700<\/li>\n<li><a href=\"http:\/\/slatestarcodex.com\/\">Slate Star Codex<\/a>: 2,357<\/li>\n<li>Longecity: 1,871<\/li>\n<li><code>brainworkshop.sourceforge.net<\/code>: 1,403<\/li>\n<li>Deep Dot Web: 1,226<\/li>\n<li><code>rs.io<\/code>: 1,197<\/li>\n<li>Facebook: 1,148<\/li>\n<\/ol><p>My monthly mailing list has reached 606 subscribers.<\/p><\/section><\/section><section id=\"july-2015january-2016\" class=\"level3\"><h3><a href=\"#july-2015january-2016\" title=\"Link to section: 'July 2015\u2013January 2016'\">July 2015\u2013January 2016<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cWe will never run out of things to program as long as there is a single program around.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Alan%20Perlis\" title=\"Wikipedia: Alan Perlis\">Alan Perlis<\/a> (<a href=\"https:\/\/www.gwern.net\/docs\/cs\/1982-perlis.pdf\">\u201cEpigrams in Programming\u201d<\/a> 1982)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (3 July 2015\u20133 January 2016):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2015 to January 2016\" height=\"1025\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2016-january-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2015 to January 2016<\/figcaption><\/figure><section id=\"traffic-10\" class=\"level4\"><h4><a href=\"#traffic-10\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20150703-20160103-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 472,244 page-views by 400,305 unique visitors over the 185 days, or 2552 page-views per day for a daily increase of +461 compared to the previous six months. (Lifetime total: 4,742,795 page-views by 2,191,138 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), late 2015\" height=\"151\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20150703-20160103-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), late 2015<\/figcaption><\/figure><p>Page rankings:<\/p><ol type=\"1\"><li>Modafinil: 67,152<\/li>\n<li>home\/main page: 50708<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/DNM-archives\">DNM archives 2011-2015<\/a>: 44,164<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Turing-complete\">Surprisingly Turing-complete systems<\/a>: 23,879<\/li>\n<li>LSD microdosing: 22,078<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/AB-testing\"><code>char-rnn<\/code> for mimicking CSS<\/a>: 20,581<\/li>\n<li>DNM arrests: 18,524<\/li>\n<li>Nootropics: 15,709<\/li>\n<li>Terrorism is not about Terror: 15,223<\/li>\n<li>Spaced repetition: 14,903<\/li>\n<li>DNB FAQ: 13,954<\/li>\n<li>Silk Road: 11,606<\/li>\n<\/ol><p>The new entry here, \u201cDNM archives\u201d, is the culmination of my DNM work: mirrors\/scrapes of dozens of markets, along with all sorts of archives and datasets related to the DNMs, covering years of market activity (1.2tb in total, to give an idea of scope). The other 2 surprises were a tutorial on how to set up <code>char-rnn<\/code> on an Amazon EC2 instance and train it to generate remarkably valid CSS (suggesting that it may be possible to train a reinforcement-learning deep neural network to design the visual appearance of websites and optimize the appearance based on visitor experiments), and an old little compilation I made of things exhibiting Turing-completeness which I found bizarre or surprising or disturbing (which naturally led to a lot of nitpicking and debate over definitions).<\/p><p><a href=\"https:\/\/www.patreon.com\/gwern\">My Patreon profile<\/a>, set up after I was forced out of Gratipay, has been a smashing success, with generally ~$300\/month in donations.<\/p><\/section><section id=\"promotion-9\" class=\"level4\"><h4><a href=\"#promotion-9\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>As before, I did minimal active promotion of my content.<\/p><ol type=\"1\"><li><code>news.ycombinator.com<\/code>: 51,256<\/li>\n<li><code>reddit.com<\/code>: 25,479<\/li>\n<li>Twitter (<code>t.co<\/code>): 7,748<\/li>\n<li>Slate Star Codex: 3,166<\/li>\n<li>Facebook: 3,044<\/li>\n<li><code>lesswrong.com<\/code>: 3,035<\/li>\n<li>Disqus: 2,042<\/li>\n<li>Bullet Proof Exec: 1,931<\/li>\n<li>Longecity: 1,887<\/li>\n<li><em>Wired<\/em>: 1,854<\/li>\n<\/ol><p>(I have no idea why Facebook traffic increased so much, since the referrers are useless.)<\/p><p>My monthly mailing list has reached 857 subscribers.<\/p><\/section><\/section><section id=\"january-2016july-2016\" class=\"level3\"><h3><a href=\"#january-2016july-2016\" title=\"Link to section: 'January 2016\u2013July 2016'\">January 2016\u2013July 2016<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cIt is easy enough to say that man is immortal simply because he will endure: that when the last ding-dong of doom has clanged and faded from the last worthless rock hanging tideless in the last red and dying evening, that even then there will still be one more sound: that of his puny inexhaustible voice, still talking. I refuse to accept this. I believe that man will not merely endure: he will prevail.\u201d<\/p>\n<p>William Faulkner (<a href=\"https:\/\/www.nobelprize.org\/prizes\/literature\/1949\/faulkner\/speech\/\">1950 Nobel Prize speech<\/a>)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (4 January 2016\u20133 July 2016):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January 2016 to July 2016\" height=\"1023\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2016-july-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January 2016 to July 2016<\/figcaption><\/figure><section id=\"traffic-11\" class=\"level4\"><h4><a href=\"#traffic-11\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20160104-20160703-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 294,009 page-views by 141,795 unique visitors over the 182 days, or 1615 page-views per day for a daily decrease of 937 compared to the previous six months. (Lifetime total: 5,036,804 page-views by 2,329,713 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), early 2016\" height=\"158\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20160104-20160703-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), early 2016<\/figcaption><\/figure><p>Page rankings:<\/p><ol type=\"1\"><li>homepage\/index: 35,673<\/li>\n<li>Modafinil: 33,928<\/li>\n<li>LSD microdosing: 21,712<\/li>\n<li>DNM arrests: 14,322<\/li>\n<li>Spaced repetition: 14,044<\/li>\n<li>Nootropics: 12,117<\/li>\n<li>DNB FAQ: 10,983<\/li>\n<li>Silk Road: 7,654<\/li>\n<li>About: 6,642<\/li>\n<li>Links: 6,410<\/li>\n<\/ol><p><a href=\"https:\/\/www.patreon.com\/gwern\">My Patreon profile<\/a> continues to be a success.<\/p><\/section><section id=\"promotion-10\" class=\"level4\"><h4><a href=\"#promotion-10\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>As before, I did minimal active promotion of my content.<\/p><ol type=\"1\"><li>Reddit: 17,792<\/li>\n<li>Facebook: 4119<\/li>\n<li>Twitter: 2,572<\/li>\n<li>HN: 2,518<\/li>\n<li>Slate Star Codex: 2,467<\/li>\n<li>LessWrong: 2,135<\/li>\n<li>Longecity: 1,555<\/li>\n<li><a href=\"http:\/\/rs.io\/\">rs.io<\/a>: 1,440<\/li>\n<li><code>brainworkshop.sourceforge.net<\/code>: 1,035<\/li>\n<li>MetaFilter: 1,018<\/li>\n<\/ol><p>My monthly mailing list has reached 1053 subscribers.<\/p><\/section><\/section><section id=\"july-2016january-2017\" class=\"level3\"><h3><a href=\"#july-2016january-2017\" title=\"Link to section: 'July 2016\u2013January 2017'\">July 2016\u2013January 2017<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cWith the people of the past \/ How I wish I could share the beauty \/ Of these cherry blossoms. \/\/ Have each of us across the years \/ Left behind this very thought?\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fujiwara%20no%20Teika\" title=\"Wikipedia: Fujiwara no Teika\">Fujiwara no Teika<\/a> (\u201cShogaku Hyakushu\u201d, 1181 AD, trans. <a href=\"https:\/\/www.gwern.net\/docs\/japanese\/1990-bundy-poeticapprenticeship:fujiwarateikasshogakuhyakushu.pdf\" title=\"Poetic Apprenticeship: Fujiwara Teika's 'Shogaku Hyakushu'\">Bundy 1990<\/a>)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (4 July 2016\u20133 January 2017):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2016\u2013January 2017\" height=\"1010\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2017-january-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2016\u2013January 2017<\/figcaption><\/figure><section id=\"traffic-12\" class=\"level4\"><h4><a href=\"#traffic-12\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20160704-20170103-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 339,529 page-views by 172,203 unique visitors over the 184 days, or 1845 page-views per day for a daily increase of 230 compared to the previous six months. (Lifetime total: 5,376,333 page-views by 2,499,949 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), late 2016\" height=\"177\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20160704-20170103-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), late 2016<\/figcaption><\/figure><p>Page rankings:<\/p><ol type=\"1\"><li>Nootropics: 43,266<\/li>\n<li>index\/main page: 39,161<\/li>\n<li>Spaced repetition: 26,965<\/li>\n<li>Modafinil: 26,849<\/li>\n<li>DNM arrests: 16,279<\/li>\n<li>LSD microdosing: 12,943<\/li>\n<li>DNB FAQ: 10,501<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Tool-AI\">Why Tool AIs Want to Be Agent AIs<\/a>: 8,026<\/li>\n<li>In Defense of Inclusionism: 7,037<\/li>\n<li>Links: 6,677<\/li>\n<\/ol><p><a href=\"https:\/\/www.patreon.com\/gwern\">My Patreon profile<\/a> has largely leveled off at a constant $310.<\/p><\/section><section id=\"promotion-11\" class=\"level4\"><h4><a href=\"#promotion-11\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>As before, I did minimal active promotion of my content.<\/p><ol type=\"1\"><li><p>HN: 42,341<\/p>\n<p>Traffic was driven primarily by a few big HN submissions by other people.<\/p><\/li>\n<li><p>Reddit: 16,151<\/p><\/li>\n<li><p>Twitter: 4,915<\/p><\/li>\n<li><p>Facebook: 4,702<\/p><\/li>\n<li><p><code>rs.io<\/code>: 1,585<\/p><\/li>\n<li><p>Longecity: 1,175<\/p><\/li>\n<li><p><code>thethirdwave.co<\/code>: 1,171<\/p><\/li>\n<li><p>DuckDuckGo search: 990<\/p><\/li>\n<li><p><code>brainworkshop.sourceforge.net<\/code>: 912<\/p><\/li>\n<li><p>Wikipedia: 794<\/p><\/li>\n<\/ol><p>My monthly mailing list has reached 1212 subscribers, up ~200.<\/p><\/section><\/section><section id=\"january-2017july-2017\" class=\"level3\"><h3><a href=\"#january-2017july-2017\" title=\"Link to section: 'January 2017\u2013July 2017'\">January 2017\u2013July 2017<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cAuthors write things down so as to have to think of them less.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Alain%20de%20Botton\" title=\"Wikipedia: Alain de Botton\">Alain de Botton<\/a><\/p>\n<\/blockquote><\/div><p>Activity graph for this period (4 January 2017\u20133 July 2017):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January 2017\u2013July 2017\" height=\"965\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2017-july-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January 2017\u2013July 2017<\/figcaption><\/figure><section id=\"traffic-13\" class=\"level4\"><h4><a href=\"#traffic-13\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20170104-20170703-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 228,652 page-views by 112,814 unique visitors over the 181 days, or 1263 page-views per day for a daily decrease of 582 visits compared to the previous six months. (Lifetime total: 5,604,985 page-views by 2,609,657 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), early 2017\" height=\"238\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20170104-20170703-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), early 2017<\/figcaption><\/figure><p>Page rankings:<\/p><ol type=\"1\"><li>Modafinil: 16,980<\/li>\n<li>Spaced repetition: 14,484<\/li>\n<li>LSD microdosing: 13,394<\/li>\n<li>DNB FAQ: 10,145<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Story-Of-Your-Life\">\u201cTed Chiang\u2019s \u2018Story Of Your Life\u2019 Is Not A Time-Travel Story\u201d<\/a>: 9,760<\/li>\n<li>Nootropics: 9,309<\/li>\n<li>DNM arrests: 8,148<\/li>\n<li>Silk Road: 6,456<\/li>\n<li>Links: 5,668<\/li>\n<li>About: 5,176<\/li>\n<\/ol><p>Aside from a critical exegesis of \u201cStory Of Your Life\u201d (prompted by its successful movie adaptation), I spent much of early 2017 tweaking <code>gwern.net<\/code> and working on unpopular statistics projects (particularly <a href=\"https:\/\/www.gwern.net\/Coin-flip\">\u201cThe Kelly Coin-Flipping Game: Exact Solutions via Decision Trees\u201d<\/a>, which though thorough and highly educational for myself, like <a href=\"https:\/\/www.gwern.net\/Mail-delivery\" title=\"Bayesian decision-theoretic analysis of local mail delivery times: modeling deliveries as survival analysis, model comparison, optimizing check times with a loss function, and optimal data collection.\">\u201cWhen Should I Check The Mail?\u201d<\/a>, will never be popular reading). In late 2017, I hope to finish up many draft or in-progress works, which may help maintain traffic.<\/p><p>After rewriting &amp; including social proof &amp; tweeting, <a href=\"https:\/\/www.patreon.com\/gwern\">my Patreon profile<\/a> has increased to $600.<\/p><\/section><section id=\"promotion-12\" class=\"level4\"><h4><a href=\"#promotion-12\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>As before, I did minimal active promotion of my content. Referrals:<\/p><ol type=\"1\"><li>HN: 10,881 (\u201cStory Of Your Life\u201d was an unexpected hit)<\/li>\n<li>Reddit: 9,538<\/li>\n<li>Twitter: 3,702<\/li>\n<li>Facebook: 2,274<\/li>\n<li>LessWrong: 1,568<\/li>\n<li>Slate Star Codex: 1,462<\/li>\n<li><code>rs.io<\/code>: 1,315<\/li>\n<li>Longecity: 1,071<\/li>\n<li>Brain Workshop: 861<\/li>\n<li>DuckDuckGo: 820<\/li>\n<\/ol><p>My monthly mailing list has reached 1419 subscribers, up ~207.<\/p><\/section><\/section><section id=\"july-2017january-2018\" class=\"level3\"><h3><a href=\"#july-2017january-2018\" title=\"Link to section: 'July 2017\u2013January 2018'\">July 2017\u2013January 2018<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cIt is only the attempt to write down your ideas that enables them to develop.\u201d<\/p>\n<p>Wittgenstein to Drury (<a href=\"https:\/\/www.amazon.com\/Recollections-Wittgenstein-Hermine\/dp\/0192876287?tag=gwernnet-20\"><em>Recollections of Wittgenstein<\/em><\/a>, Rhees 1984)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (4 July 2017\u20133 January 2018):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2017\u2013January 2018\" height=\"965\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2018-january-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2017\u2013January 2018<\/figcaption><\/figure><section id=\"traffic-14\" class=\"level4\"><h4><a href=\"#traffic-14\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20170704-20180103-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 326,852 page-views by 155,532 unique visitors over the 184 days, or 1776 page-views per day for a daily increase of 513 visits compared to the previous six months. (Lifetime total: 5,931,837 page-views by 2,763,402 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), late 2017\" height=\"335\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20170704-20180103-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), late 2017<\/figcaption><\/figure><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20170704-20180103-gwern.net-analytics-pages.pdf\">Page rankings<\/a> (somewhat scrambled due to the big renaming and GA not merging analytics across URLs):<\/p><ol type=\"1\"><li>index: 43,863<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Ads\" title=\"9 months of daily A\/B-testing of Google AdSense banner ads on gwern.net indicates banner ads decrease total traffic substantially, possibly due to spillover effects in reader engagement and resharing\">Banner Ads Considered Harmful<\/a>: 20,013<\/li>\n<li>Modafinil: 14,340<\/li>\n<li>DNM arrests: 14,115<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Tanks\" title=\"AI folklore tells a story about a neural network trained to detect tanks which instead learned to detect time of day; investigating, this probably never happened\">The Neural Net Tank Urban Legend<\/a>: 13,808<\/li>\n<li>DNB FAQ: 13,650<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Bitcoin-is-Worse-is-Better\" title=\"Bitcoin's long gestation and early opposition indicates it is an example of the 'Worse is Better' paradigm in which an ugly complex design with few attractive theoretical properties compared to purer competitors nevertheless successfully takes over a niche, survives, and becomes gradually refined.\">Bitcoin Is Worse Is Better<\/a>: 11,401<\/li>\n<li>Spaced repetition: 12,428<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Greenland\" title=\"Denmark turned down 100m USD from the USA; I discuss how this was a bad idea\">Reasons of State: Why Didn\u2019t Denmark Sell Greenland?<\/a>: 11,150<\/li>\n<li>Nootropics: 7,615<\/li>\n<\/ol><p>My A\/B test of banner ads turned in an unexpected result of substantial harm, which was of considerable general interest (especially to anyone who runs an ad-supported website), so no surprise that had a lot of traffic. I wrote off the obscure neural net\/tank story years ago after some cursory investigation showed it had all the classic traits of an urban legend; but it has kept coming up, and at Anders Sandberg\u2019s request, I wrote it up in considerably more detail, and was surprised that people took such an interest in it and even defended it as probably genuine, in the absence of all evidence\u2014I guess people like debunkings? The final oddity was the resurgence of my old essay on an extremely obscure bit of US historical trivia, that it had tried to acquire Greenland in the early Cold War, where I noted that Denmark\u2019s refusal was a costly one; I don\u2019t know why anyone was reading it, but I was annoyed at the idiotic objections people made, typically either pure fantasy, already refuted like the many suggestions that Greenland oil would yet pay off, or illogical virtue signaling\u2014ironically demonstrating my point. (Considerable effort invested in defining hundreds of redirects for 404 errors didn\u2019t pay off as people kept discovering ever new ways to misspell URLs, but it did allow me to rename pages without destroying traffic by breaking most incoming links, switching from spaces to hyphens in order to idiot-proof them.)<\/p><p><a href=\"https:\/\/www.patreon.com\/gwern\">My Patreon profile<\/a> has increased to $700.<\/p><\/section><section id=\"promotion-13\" class=\"level4\"><h4><a href=\"#promotion-13\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>As before, I did minimal active promotion of my content. Referrals:<\/p><ol type=\"1\"><li>HN: 26,112<\/li>\n<li>Reddit: 10,715<\/li>\n<li>Marginal Revolution: 8,064<\/li>\n<li>Twitter: 7,111<\/li>\n<li>Slate Star Codex: 5,906<\/li>\n<li>Facebook: 2228<\/li>\n<li>LessWrong: 1,547<\/li>\n<li>The Browser: 1,057 (this would have been higher if The Browser had not carelessly broken its URL due to the space-escaping issue, which prompted my site-wide renaming)<\/li>\n<li>Duck Duck Go: 782<\/li>\n<li>Longecity: 777<\/li>\n<\/ol><p>My monthly mailing list has reached 1786 subscribers, up ~367.<\/p><\/section><\/section><section id=\"january-2018july-2018\" class=\"level3\"><h3><a href=\"#january-2018july-2018\" title=\"Link to section: 'January 2018\u2013July 2018'\">January 2018\u2013July 2018<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cPeople will have their miracles, their stories, their heroes and heroines and saints and martyrs and divinities to exercise their gifts of affection, admiration, wonder, and worship, and their Judases and devils to enable them to be angry and yet feel that they do well to be angry. Every one of these legends is the common heritage of the human race; and there is only one inexorable condition attached to their healthy enjoyment, which is that no one shall believe them literally. The reading of stories and delighting in them made Don Quixote a gentleman: the believing them literally made him a madman who slew lambs instead of feeding them.\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/George%20Bernard%20Shaw\" title=\"Wikipedia: George Bernard Shaw\">George Bernard Shaw<\/a> (<a href=\"https:\/\/www.gutenberg.org\/files\/13084\/13084.txt\"><em>A Touchstone For Dogma<\/em><\/a>)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (4 January 2018\u20133 July 2018):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): January 2018\u2013July 2018\" height=\"690\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2018-july-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): January 2018\u2013July 2018<\/figcaption><\/figure><section id=\"traffic-15\" class=\"level4\"><h4><a href=\"#traffic-15\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20180104-20180703-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 329,057 page-views by 148,481 unique visitors over the 181 days, or 1818 page-views per day for a daily increase of 42 pageviews a day compared to the previous six months. (Lifetime total: 6,236,345 page-views by 2,945,757 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), early 2018\" height=\"181\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20180104-20180703-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), early 2018<\/figcaption><\/figure><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20180104-20180703-gwern.net-analytics-pages.pdf\">Page rankings<\/a>:<\/p><ol type=\"1\"><li>index: 43,535<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Complement\" title=\"A classic pattern in technology economics, identified by Joel Spolsky, is layers of the stack attempting to become monopolies while turning other layers into perfectly-competitive markets which are commoditized, in order to harvest most of the consumer surplus; discussion and examples.\">\u201cLaws of Tech: Commoditize Your Complement\u201d<\/a>: 29,755<\/li>\n<li>Modafinil: 16,339<\/li>\n<li>DNB FAQ: 15,922<\/li>\n<li>Spaced repetition: 15,599<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Danbooru2017\" title=\"Danbooru2017 is a large-scale anime image database with 2.9m+ images annotated with 77.5m+ tags; it can be useful for machine learning purposes such as image recognition and generation.\">\u201cDanbooru2017: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset\u201d<\/a>: 10,531<\/li>\n<li>Nootropics: 8,020<\/li>\n<li>Links: 7,105<\/li>\n<li>DNM arrests: 6,707<\/li>\n<li>Death Note Anonymity: 6,464<\/li>\n<\/ol><p>\u201cLaws of Tech: Commoditize Your Complement\u201d is based on an old Joel Spolsky essay I found insightful and to explain a lot of puzzling behaviors in Silicon Valley, so I was surprised when, during my SF trip, no one who I mentioned it to in discussions of technology had heard of it; I returned resolved to try to promote it some more by re-emphasizing it and collating a list of examples so no one could fail to see the pattern, and I believe I succeeded. The launch of Danbooru2017 was less satisfying. I kept running into difficulties and delays, most associated with how to format the metadata (provided in a not-too-useful JSON format)\u2014indeed, I had expected to launch it in April 2017 and build my deep learning workstation in December 2017. Instead, it wasn\u2019t until February 2018 that I gave up on the metadata, reasoning that I was letting the perfect be the enemy of better, and until July 2018 that I finished building my workstation with 2 GPUs (after 2 dead motherboards, among other problems). There have been somewhere around 200 downloads of the dataset as far as I can extrapolate from the torrent client upload &amp; rsync logs, but thus far I only know of 1 serious use of it (a CycleGAN paper). I remain hopeful that uses will gradually increase over the next few years, the way they did for my darknet market dataset; in any case, I\u2019ve begun working on some uses of Danbooru2017 myself.<\/p><p><a href=\"https:\/\/www.patreon.com\/gwern\">My Patreon profile<\/a> decreased to $670.<\/p><\/section><section id=\"promotion-14\" class=\"level4\"><h4><a href=\"#promotion-14\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>Referrals:<\/p><ol type=\"1\"><li>HN: 13,533<\/li>\n<li>Reddit: 6,686<\/li>\n<li>Twitter: 6,529<\/li>\n<li>Slate Star Codex: 6,246<\/li>\n<li>Facebook: 2,830<\/li>\n<li><code>adme.ru<\/code>: 2,414<\/li>\n<li>Android Google Search: 788<\/li>\n<li>The Browser: 562<\/li>\n<li>4chan: 523<\/li>\n<li>Duck Duck Go: 521<\/li>\n<\/ol><p>My monthly mailing list has reached 2183 subscribers, up ~307. By crossing the 2000 subscriber mark, Mailchimp disabled my account and began charging ~$30 a month to send out a single newsletter; as it\u2019s a little hard for me to justify spending anywhere up to $400\/year on my newsletter, which I don\u2019t directly make any money on, I had to stop using Mailchimp. after looking around some, I switched to <a href=\"https:\/\/www.tinyletter.com\/\">TinyLetter<\/a>, which oddly enough is <em>also<\/em> run by Mailchimp but doesn\u2019t have that arbitrary subscriber limit (though it lacks many features of Mailchimp proper).<\/p><\/section><\/section><section id=\"july-2018january-2019\" class=\"level3\"><h3><a href=\"#july-2018january-2019\" title=\"Link to section: 'July 2018\u2013January 2019'\">July 2018\u2013January 2019<\/a><\/h3><div class=\"epigraph\"><blockquote>\n<p>\u201cIs not a patron, my lord, one who looks with unconcern on a man struggling for life in the water, and when he has reached ground, encumbers him with help?\u201d<\/p>\n<p>Samuel Johnson (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Letter%20to%20Chesterfield\" title=\"Wikipedia: Letter to Chesterfield\">Letter to Chesterfield<\/a>, 1755)<\/p>\n<\/blockquote><\/div><p>Activity graph for this period (4 July 2018\u20133 January 2019):<\/p><figure><img alt=\"Plot of patch creations (y-axis) versus date (x-axis): July 2018\u2013January 2019\" height=\"690\" src=\"https:\/\/www.gwern.net\/images\/traffic\/2019-january-git-history.png\" width=\"1400\" \/><figcaption>Plot of patch creations (y-axis) versus date (x-axis): July 2018\u2013January 2019<\/figcaption><\/figure><section id=\"traffic-16\" class=\"level4\"><h4><a href=\"#traffic-16\" title=\"Link to section: 'Traffic'\">Traffic<\/a><\/h4><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20180704-20190103-gwern.net-analytics.pdf\">Analytics traffic<\/a> records 407,395 page-views by 185,141 unique visitors over the 184 days, or 2214 page-views per day for an increase of 396 pageviews a day compared to the previous six months. (Lifetime total: 6,638,722 page-views by 3,113,460 unique visitors.)<\/p><figure><img alt=\"Plot of page-views (y-axis) versus date (x-axis), late 2018\" height=\"216\" src=\"https:\/\/www.gwern.net\/images\/traffic\/20180704-20190103-traffic-history.png\" width=\"1400\" \/><figcaption>Plot of page-views (y-axis) versus date (x-axis), late 2018<\/figcaption><\/figure><p><a href=\"https:\/\/www.gwern.net\/docs\/traffic\/20180704-20190103-gwern.net-analytics-pages.pdf\">Page rankings<\/a>:<\/p><ol type=\"1\"><li>index: 50,849<\/li>\n<li><a href=\"https:\/\/www.gwern.net\/Morning-writing\" title=\"Many writers anecdotally report they write best first thing early in the morning, apparently even if they are not morning people. Do they, and why?\">\u201cWhat Is The Morning Writing Effect?\u201d<\/a>: 27,241<\/li>\n<li>Turing-complete: 26,662<\/li>\n<li>Notes: (for <a href=\"https:\/\/www.gwern.net\/Questions\">open questions<\/a>\/<a href=\"https:\/\/www.gwern.net\/Littlewood\" title=\"Selection effects in media become increasingly strong as populations and media increase, meaning that rare datapoints driven by unusual processes such as the mentally ill or hoaxers are increasingly unreliable as evidence of anything at all and must be ignored. At scale, anything that can happen will happen a small but nonzero times.\">Littlewood\u2019s Law<\/a>): 17,706<\/li>\n<li>Spaced repetition: 14,471<\/li>\n<li>Modafinil: 13,958<\/li>\n<li>\u201cBitcoin Is Worse Is Better\u201d: 13,169<\/li>\n<li>DNB FAQ: 10,344<\/li>\n<li>Links: 9,729<\/li>\n<li>Drug heuristics\/Algernon\u2019s Law: 9,724<\/li>\n<li>And honorable mentions for the new pages <a href=\"https:\/\/www.gwern.net\/Search\" title=\"A description of tips and tricks for effective Internet search of papers\/books for research.\">\u201cInternet Search Tips\u201d<\/a> which <em>almost<\/em> made it to the top 10 (#13) despite being released mid-way December 2018, and <a href=\"https:\/\/www.gwern.net\/Cat-Sense\"><em>Cat Sense<\/em>, Bradshaw 2013: Are We Good Owners?<\/a> (on cat psychology, genetics, domestication, &amp; dysgenics) at #21.<\/li>\n<\/ol><p>I was disappointed that two of my new ones had little or no interest(<a href=\"https:\/\/www.gwern.net\/MLP\" title=\"A review of the cartoon series 'My Little Pony: Friendship Is Magic', focusing on fandom, plot, development, and meaning of bronydom.\">\u201c<em>My Little Pony<\/em>: Immanetizing The Equestrian\u201d<\/a> &amp; <a href=\"https:\/\/www.gwern.net\/docs\/statistics\/order\/beanmachine-multistage\/index.html\" title=\"An interactive JS visualization of Galton quincunxes demonstrating the compounding power\/ratchet effect of iterative selection rather than a single selection step.\">\u201cMulti-Stage Selection Bean Machine JS Demo\u201d<\/a>), and that Danbooru2017, despite its use in the impressive <a href=\"https:\/\/github.com\/lllyasviel\/style2paints\">style2paints v4<\/a>, saw little uptake. Maybe Danbooru2018 will do better.<\/p><p><a href=\"https:\/\/www.patreon.com\/gwern\">My Patreon profile<\/a> increased to $835.<\/p><\/section><section id=\"promotion-15\" class=\"level4\"><h4><a href=\"#promotion-15\" title=\"Link to section: 'Promotion'\">Promotion<\/a><\/h4><p>Referrals saw some unusual changes:<\/p><ol type=\"1\"><li>HN: 37,603<\/li>\n<li>Twitter: 10,909<\/li>\n<li>Reddit: 6,046<\/li>\n<li>Slate Star Codex: 4,479<\/li>\n<li>Facebook: 1,389<\/li>\n<li>Github: 1,209<\/li>\n<li>CoinMarketCap: 1,115<\/li>\n<li>TinyLetter: 1,089<\/li>\n<li><a href=\"https:\/\/guzey.com\/\">Alexey Guzey<\/a>: 795<\/li>\n<li>LessWrong: 782<\/li>\n<\/ol><p>My monthly mailing list has reached 2753 subscribers, up ~570.<\/p><\/section><\/section><\/section><section id=\"colophon\" class=\"level2\"><section id=\"hosting\" class=\"level3\"><h3><a href=\"#hosting\" title=\"Link to section: 'Hosting'\">Hosting<\/a><\/h3><p><code>gwern.net<\/code> is served by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Amazon%20S3\" title=\"Wikipedia: Amazon S3\">Amazon S3<\/a> through the <a href=\"https:\/\/en.wikipedia.org\/wiki\/CloudFlare\" title=\"Wikipedia: CloudFlare\">CloudFlare<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Content%20delivery%20network\" title=\"Wikipedia: Content delivery network\">CDN<\/a>. (Amazon charges less for bandwidth and disk space than NFSN, although one loses all the capabilities offered by Apache\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/.htaccess\" title=\"Wikipedia: .htaccess\">.htaccess<\/a>, and Gzip compression is difficult so must be handled by CloudFlare; total costs may turn out to be a wash and I will consider the switch to Amazon S3 a success if it can bring my monthly bill to &lt;$10 or &lt;$120 a year.)<\/p><p>From October 2010 to June 2012, the site was hosted on <a href=\"https:\/\/www.nearlyfreespeech.net\">NearlyFreeSpeech.net<\/a>, an old hosting company; its specific niche is controversial material and activist-friendly pricing. Its libertarian owners cast a jaundiced eye on <a href=\"https:\/\/en.wikipedia.org\/wiki\/takedown%20request\" title=\"Wikipedia: takedown request\">takedown request<\/a>s, and pricing is <a href=\"https:\/\/en.wikipedia.org\/wiki\/pay-as-you-go\" title=\"Wikipedia: pay-as-you-go\">pay-as-you-go<\/a>. I like the former aspect, but the latter sold me on NFSN. Before I stumbled on NFSN (someone mentioned it offhandedly while chatting), I was getting ready to pay $10-15 a month ($120 yearly) to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Linode\" title=\"Wikipedia: Linode\">Linode<\/a>. Linode\u2019s offerings are overkill since I do not run dynamic websites or something like <a href=\"http:\/\/www.haskell.org\">Haskell.org<\/a> (with wikis and mailing lists and <a href=\"https:\/\/en.wikipedia.org\/wiki\/darcs\" title=\"Wikipedia: darcs\">darcs<\/a> repositories), but I didn\u2019t know a good alternative. NFSN\u2019s pricing meant that I paid for usage rather than large flat fees. I put in $32 to cover registering <code>gwern.net<\/code> until 2014, and then another $10 to cover bandwidth &amp; storage price. DNS aside, I was billed $8.27 for October-December 2010; DNS included, January-April 2011 cost $10.09. $10 covered months of <code>gwern.net<\/code> for what I would have paid Linode in 1 month! In total, my 2010 costs were $39.44 (<a href=\"https:\/\/www.gwern.net\/docs\/traffic\/2010-gwernnet-nfsncosts.maff\">bill archive<\/a>); my 2011 costs were $118.32 ($9.86 a month; <a href=\"https:\/\/www.gwern.net\/docs\/traffic\/2011-gwernnet-nfsncosts.maff\">archive<\/a>); and my 2012 costs through June were $112.54 ($21 a month; <a href=\"https:\/\/www.gwern.net\/docs\/traffic\/2012-gwernnet-nfsncosts.maff\">archive<\/a>); sum total: $270.3.<\/p><p>The switch to Amazon S3 hosting is complicated by my simultaneous addition of CloudFlare as a CDN; my total June 2012 Amazon bill is $1.62, with $0.19 for storage. CloudFlare claims it covered 17.5GB of 24.9GB total bandwidth, so the $1.41 represents 30% of my total bandwidth; multiply 1.41 by 3 is 4.30, and my hypothetical non-CloudFlare S3 bill is ~$4.5. Even at $10, this was well below the $21 monthly cost at NFSN. (The traffic graph indicates that June 2012 was a relatively quiet period, but I don\u2019t think this eliminates the factor of 5.) From July 2012 to June 2013, my Amazon bills totaled $60, which is reasonable except for the steady increase ($1.62\/$3.27\/$2.43\/$2.45\/$2.88\/$3.43\/$4.12\/$5.36\/$5.65\/$5.49\/$4.88\/$8.48\/$9.26), being primarily driven by out-bound bandwidth (in June 2013, the $9.26 was largely due to the 75GB transferred\u2014and that was <em>after<\/em> CloudFlare dealt with 82GB); $9.26 is much higher than I would prefer since that would be &gt;$110 annually. This was probably due to all the graphics I included in the \u201cGoogle shutdowns\u201d analysis, since it returned to a more reasonable $5.14 on 42GB of traffic in August. September, October, November and December 2013 saw high levels maintained at $7.63\/$12.11\/$5.49\/$8.75, so it\u2019s probably a new normal. 2014 entailed new costs related to EC2 instances &amp; S3 bandwidth spikes due to hosting a multi-gigabyte scientific dataset, so bills ran $8.51\/$7.40\/$7.32\/$9.15\/$26.63\/$14.75\/$7.79\/$7.98\/$8.98\/$7.71\/$7\/$5.94. 2015 &amp; 2016 were similar: $5.94\/$7.30\/$8.21\/$9.00\/$8.00\/$8.30\/$10.00\/$9.68\/$14.74\/$7.10\/$7.39\/$8.03\/$8.20\/$8.31\/$8.25\/$9.04\/$7.60\/$7.93\/$7.96\/$9.98\/$9.22\/$11.80\/$9.01\/$8.87. 2017 saw costs increase due to one of my side-projects, aggressively increasing fulltexting of <code>gwern.net<\/code> by providing more papers &amp; scanning cited books, only partially offset by changes like lossy optimization of images &amp; converting GIFs to WebMs: $12.49\/$10.68\/$11.02\/$12.53\/$11.05\/$10.63\/$9.04\/$11.03\/$14.67\/$15.52\/$13.12\/$12.23 (total: $144.01). In 2018, I continued fulltexting: $13.08\/$14.85\/$14.14\/$18.73\/$18.88\/$15.92\/$15.64\/$15.27\/$16.66\/$22.56\/$23.59\/$25.91\/(total: $213).<\/p><\/section><section id=\"source\" class=\"level3\"><h3><a href=\"#source\" title=\"Link to section: 'Source'\">Source<\/a><\/h3><p>The revision history is kept in git; individual page sources can be read by appending \u201c.page\u201d to their URL.<\/p><section id=\"size\" class=\"level4\"><h4><a href=\"#size\" title=\"Link to section: 'Size'\">Size<\/a><\/h4><p>As of 4 January 2019, the source of <code>gwern.net<\/code> is composed of &gt;321 text files with &gt;3,208,000 words or 23MB; this includes my writings &amp; documents I have transcribed into Markdown, but excludes images, PDFs, archives, infrastructure, and the revision history.  With those included and everything compiled to the static<a href=\"#fn39\" class=\"footnote-ref\" id=\"fnref39\" role=\"doc-noteref\"><sup>39<\/sup><\/a> HTML, the site is &gt;14.2GB. The source repository contains &gt;11,801 patches (this is an under-count as the creation of the repository in 26 September 2008 included already-written material).<\/p><section id=\"benfords-law\" class=\"level5\"><h5><a href=\"#benfords-law\" title=\"Link to section: 'Benford\u2019s law'\">Benford\u2019s law<\/a><\/h5><p>In March 2013 I wondered, upon seeing a mention of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Benford%27s%20law\" title=\"Wikipedia: Benford\u2019s law\">Benford\u2019s law<\/a>: \u201cif I extracted all the numbers from everything I\u2019ve written on <code>gwern.net<\/code>, would it satisfy Benford\u2019s law?\u201d It seems the answer is\u2026 <em>almost<\/em>. I generate the list of numbers by running a Haskell program to parse digits, commas, and periods; and then I process it with shell utilities.<a href=\"#fn40\" class=\"footnote-ref\" id=\"fnref40\" role=\"doc-noteref\"><sup>40<\/sup><\/a> This can then be read in R to run a <a href=\"https:\/\/en.wikipedia.org\/wiki\/chi-squared%20test\" title=\"Wikipedia: chi-squared test\">chi-squared test<\/a> confirming lack of fit (<em>p<\/em>=~0) and generate this comparison of the data &amp; Benford\u2019s law<a href=\"#fn41\" class=\"footnote-ref\" id=\"fnref41\" role=\"doc-noteref\"><sup>41<\/sup><\/a>:<\/p><figure><img alt=\"Histogram\/barplot of parsed numbers vs predicted\" height=\"427\" src=\"https:\/\/www.gwern.net\/images\/2013-benfords-law.png\" width=\"734\" \/><figcaption>Histogram\/barplot of parsed numbers vs predicted<\/figcaption><\/figure><p>There\u2019s a clear resemblance for everything but the digit \u20182\u2019, which then blows the fit to heck. I have no idea why 2 is overrepresented\u2014it may be due to all the citations to recent academic papers which would involve numbers starting with \u20182\u2019 (2002, 2010, 2013\u2026) and cause a double-count in both the citation and filename, since if I look in the <code>docs\/<\/code> fulltext folder, I see 160 files starting with \u20181\u2019 but 326 starting with \u20182\u2019. But this can\u2019t be the entire explanation since \u20182\u2019 has 20.3k entries while to fit Benford, it needs to be just 11.5k\u2014leaving a gap of ~10k numbers unexplained. A mystery.<\/p><\/section><\/section><\/section><section id=\"tools\" class=\"level3\"><h3><a href=\"#tools\" title=\"Link to section: 'Tools'\">Tools<\/a><\/h3><p>Software tools &amp; libraries used in the site as a whole:<\/p><ul><li><p>The source files are written in <a href=\"http:\/\/johnmacfarlane.net\/pandoc\/\">Pandoc<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Markdown\" title=\"Wikipedia: Markdown\">Markdown<\/a><\/p><\/li>\n<li><p>math is written in <a href=\"https:\/\/en.wikipedia.org\/wiki\/LaTeX\" title=\"Wikipedia: LaTeX\">LaTeX<\/a> compiled to <a href=\"https:\/\/en.wikipedia.org\/wiki\/MathML\" title=\"Wikipedia: MathML\">MathML<\/a><\/p><\/li>\n<li><p>the site is compiled with the <a href=\"https:\/\/github.com\/jaspervdj\/Hakyll\/\">Hakyll<\/a> (v4) static site generator, used to generate <code>gwern.net<\/code>, written in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Haskell%20%28programming%20language%29\" title=\"Wikipedia: Haskell (programming language)\">Haskell<\/a>; for the gory details, see <a href=\"https:\/\/www.gwern.net\/hakyll.hs\"><code>hakyll.hs<\/code><\/a> which implements the compilation, RSS feed generation, &amp; parsing of interwiki links as well.<\/p>\n<p>My preferred method of use is to browse &amp; edit locally using Emacs, and then distribute using Hakyll. To use Hakyll, you <code>cd<\/code> into your repository and <code>runhaskell hakyll.hs build<\/code> (with <code>hakyll.hs<\/code> having whatever options you like). Hakyll will build a static HTML\/CSS hierarchy inside <code>_site\/<\/code>; you can then do something like <code>firefox _static\/index<\/code>. (Because HTML extensions are not specified in the interest of <a href=\"http:\/\/www.w3.org\/TR\/cooluris\/\">cool URIs<\/a>, you cannot use the Hakyll <code>watch<\/code> webserver as of January 2014.) Hakyll\u2019s main advantage for me is relatively straightforward integration with the Pandoc Markdown libraries; Hakyll is not that easy to use, and so I do not recommend use of Hakyll as a general static site generator unless one is already adept with Haskell.<\/p><\/li>\n<li><p>the CSS is borrowed from a motley of sources but stems primarily from the <a href=\"http:\/\/jaspervdj.be\/hakyll\/\">Hakyll homepage<\/a> &amp; <a href=\"https:\/\/github.com\/jgm\/gitit\">Gitit<\/a>; for specifics, see the unminified <code>static\/css\/default.css<\/code> in the source repository.<\/p><\/li>\n<li><p>JavaScript:<\/p>\n<ul><li>Comments are outsourced to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Disqus\" title=\"Wikipedia: Disqus\">Disqus<\/a> (since I am not interested in writing a dynamic system to do it, and their anti-spam techniques are much better than mine).<\/li>\n<li>the floating footnotes are via <a href=\"http:\/\/ignorethecode.net\/blog\/2010\/04\/20\/footnotes\/\"><code>footnotes.js<\/code><\/a><\/li>\n<li>the HTML tables are sortable via <a href=\"http:\/\/tablesorter.com\/docs\/\">tablesorter<\/a><\/li>\n<li>the MathML is rendered using <a href=\"https:\/\/en.wikipedia.org\/wiki\/MathJax\" title=\"Wikipedia: MathJax\">MathJax<\/a><\/li>\n<li>analytics are handled by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Google%20Analytics\" title=\"Wikipedia: Google Analytics\">Google Analytics<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/A%2FB%20testing\" title=\"Wikipedia: A\/B testing\">A\/B testing<\/a> is done using <a href=\"https:\/\/github.com\/danmaz74\/ABalytics\">ABalytics<\/a> (hooks into Google Analytics; see current <a href=\"https:\/\/www.gwern.net\/AB-testing\">testing notes<\/a>)<\/li>\n<\/ul><\/li>\n<li><p>Book affiliate links are through an <a href=\"https:\/\/en.wikipedia.org\/wiki\/Amazon%20Affiliates\" title=\"Wikipedia: Amazon Affiliates\">Amazon Affiliates<\/a> tag appended in the <code>hakyll.hs<\/code><\/p><\/li>\n<\/ul><p>These tools encourage a minimalist site; I believe that <a href=\"https:\/\/en.wikipedia.org\/wiki\/minimalism\" title=\"Wikipedia: minimalism\">minimalism<\/a> helps one focus on the content. Anything besides the content is <a href=\"http:\/\/www.jwz.org\/gruntle\/design.html\">distraction and not design<\/a>. \u2018Attention!\u2019, as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ikkyu\" title=\"Wikipedia: Ikkyu\">Ikkyu<\/a> would say<a href=\"#fn42\" class=\"footnote-ref\" id=\"fnref42\" role=\"doc-noteref\"><sup>42<\/sup><\/a>.<\/p><\/section><section id=\"license\" class=\"level3\"><h3><a href=\"#license\" title=\"Link to section: 'License'\">License<\/a><\/h3><p>This site is licensed under the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Creative%20Commons\" title=\"Wikipedia: Creative Commons\">Creative Commons<\/a> <a href=\"http:\/\/creativecommons.org\/about\/cc0\">public domain (CC-0)<\/a> license.<\/p><p>I believe the public domain license reduces <a href=\"https:\/\/en.wikipedia.org\/wiki\/Fear%2C%20uncertainty%20and%20doubt\" title=\"Wikipedia: Fear, uncertainty and doubt\">FUD<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/dead-weight%20loss\" title=\"Wikipedia: dead-weight loss\">dead-weight loss<\/a><a href=\"#fn43\" class=\"footnote-ref\" id=\"fnref43\" role=\"doc-noteref\"><sup>43<\/sup><\/a>, encourages copying (<a href=\"https:\/\/en.wikipedia.org\/wiki\/LOCKSS\" title=\"Wikipedia: LOCKSS\">LOCKSS<\/a>), gives back (however little) to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Free%20Software\" title=\"Wikipedia: Free Software\">Free Software<\/a>\/<a href=\"https:\/\/en.wikipedia.org\/wiki\/Free%20Content\" title=\"Wikipedia: Free Content\">Free Content<\/a>, and costs me nothing<a href=\"#fn44\" class=\"footnote-ref\" id=\"fnref44\" role=\"doc-noteref\"><sup>44<\/sup><\/a>.<\/p><\/section><\/section><\/section><section class=\"footnotes\" role=\"doc-endnotes\"><hr \/><ol><li id=\"fn1\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gennadi%20Sosonko\" title=\"Wikipedia: Gennadi Sosonko\">Gennadi Sosonko<\/a>, pg 19 of <a href=\"http:\/\/www.amazon.com\/Russian-Silhouettes-Genna-Sosonko\/dp\/9056912933\/?tag=gwernnet-20\"><em>Russian Silhouettes<\/em><\/a>, on why he wrote his book of biographical sketches of great Soviet chess players. (As Richardson asks (<em>Vectors 1.0<\/em>, 2001): \u201c25. Why would we write if we\u2019d already heard what we wanted to hear?\u201d)<a href=\"#fnref1\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn2\" role=\"doc-endnote\"><p>\u201cIt is only the attempt to write down your ideas that enables them to develop.\u201d \u2013Wittgenstein (pg 109, <a href=\"http:\/\/www.amazon.com\/Recollections-Wittgenstein-Hermine\/dp\/0192876287\/?tag=gwernnet-20\"><em>Recollections of Wittgenstein<\/em><\/a>); \u201cI thought a little [while in the isolation tank], and then I stopped thinking altogether\u2026incredible how idleness of body leads to idleness of mind. After 2 days, I\u2019d turned into an idiot. That\u2019s the reason why, during a flight, astronauts are always kept busy.\u201d \u2013Oriana Fallaci, <a href=\"http:\/\/www.johndcook.com\/blog\/2010\/12\/11\/after-two-days-id-turned-into-an-idiot\/\">quoted<\/a> in <a href=\"http:\/\/www.amazon.com\/Rocket-Men-Epic-Story-First\/dp\/B002VPE85K\/?tag=gwernnet-20\"><em>Rocket Men: The Epic Story of the First Men on the Moon<\/em><\/a> by Craig Nelson.<a href=\"#fnref2\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn3\" role=\"doc-endnote\"><p>One danger of such an approach is that you will simply engage in <a href=\"https:\/\/en.wikipedia.org\/wiki\/confirmation%20bias\" title=\"Wikipedia: confirmation bias\">confirmation bias<\/a>, and build up an impressive-looking wall of citations that is completely wrong but effective in brainwashing yourself. The only solution is to be diligent to include criticism\u2014so even if you do not escape brainwashing, at least your readers have a chance. <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/The%20Autobiography%20of%20Charles%20Darwin\" title=\"Wikipedia: The Autobiography of Charles Darwin\">The Autobiography of Charles Darwin<\/a><\/em>, 1902:<\/p>\n<blockquote>\n<p>I had, also, during many years followed a golden rule, namely, that whenever a published fact, a new observation or thought came across me, which was opposed to my general results, to make a memorandum of it without fail and at once; for I had found by experience that such facts and thoughts were far more apt to escape from the memory than favourable ones. Owing to this habit, very few objections were raised against my views which I had not at least noticed and attempted to answer.<\/p>\n<\/blockquote>\n<a href=\"#fnref3\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn4\" role=\"doc-endnote\"><p>Such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Larry%20Niven\" title=\"Wikipedia: Larry Niven\">Larry Niven<\/a>\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/Known%20Space\" title=\"Wikipedia: Known Space\">Known Space<\/a> universe; consider the introduction to the chronologically last story in that setting, \u201cSafe at Any Speed\u201d (<em>Tales of Known Space<\/em>).<a href=\"#fnref4\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn5\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Erich%20Fromm\" title=\"Wikipedia: Erich Fromm\">Erich Fromm<\/a>:<\/p>\n<blockquote>\n<p>\u201cIf the individual lived five hundred or one thousand years, this clash (between his interests and those of society) might not exist or at least might be considerably reduced. He then might live and harvest with joy what he sowed in sorrow; the suffering of one historical period which will bear fruit in the next one could bear fruit for him too.\u201d<\/p>\n<\/blockquote>\n<a href=\"#fnref5\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn6\" role=\"doc-endnote\"><p>From <a href=\"https:\/\/en.wikipedia.org\/wiki\/Richard%20Posner\" title=\"Wikipedia: Richard Posner\">Richard Posner<\/a>\u2019s <a href=\"http:\/\/www.amazon.com\/Aging-Old-Age-Richard-Posner\/dp\/0226675688\/?tag=gwernnet-20\"><em>Aging and Old Age<\/em><\/a>:<\/p>\n<blockquote>\n<p>One way to distinguish empirically between aging effects and proximity-to-death effects would be to compare, with respect to choice of occupation, investment, education, leisure activities, and other activities, elderly people on the one hand with young or middle-aged people who have truncated life expectancies but are in apparent good health, on the other. For example, a person newly infected with the AIDS virus (HIV) has roughly the same life expectancy as a 65-year-old and is unlikely to have, as yet, [major] symptoms. The conventional human-capital model implies that, after correction for differences in income and for other differences between such persons and elderly persons who have the same life expectancy (a big difference is that the former will not have pension entitlements to fall back upon), the behavior of the two groups will be similar. It does appear to be similar, so far as investing in human capital is concerned; the truncation of the payback period causes disinvestment. And there is a high suicide rate among HIV-infected persons (even before they have reached the point in the progression of the disease at which they are classified as persons with AIDS), just as there is, as we shall see in chapter 6, among elderly persons.<\/p>\n<\/blockquote>\n<a href=\"#fnref6\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn7\" role=\"doc-endnote\"><p><a href=\"https:\/\/millercenter.org\/the-presidency\/presidential-speeches\/march-23-1962-address-university-california-berkeley#dp-expandable-text\" title=\"Address at the University of California, Berkeley (March 23, 1962)\">John F. Kennedy<\/a>, 1962:<\/p>\n<blockquote>\n<p>I am reminded of the story of the great French Marshal Lyautey, who once asked his gardener to plant a tree. The gardener objected that the tree was slow-growing and would not reach maturity for a hundred years. The Marshal replied, \u201cIn that case, there is no time to lose, plant it this afternoon.\u201d<\/p>\n<\/blockquote>\n<a href=\"#fnref7\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn8\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Jorge%20Luis%20Borges\" title=\"Wikipedia: Jorge Luis Borges\">Jorge Luis Borges<\/a>, <em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Dreamtigers\" title=\"Wikipedia: Dreamtigers\">Dreamtigers<\/a><\/em> <a href=\"http:\/\/thefloatinglibrary.com\/2008\/12\/09\/dreamtigers-epiloge-j-l-borges\/\">Epilogue<\/a><a href=\"#fnref8\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn9\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mark%20Pilgrim\" title=\"Wikipedia: Mark Pilgrim\">Mark Pilgrim<\/a>, <a href=\"http:\/\/web.archive.org\/web\/20110726001925\/http:\/\/diveintomark.org\/archives\/2004\/05\/14\/freedom-0\">\u201cFreedom 0\u201d<\/a>:<\/p>\n<blockquote>\n<p>In the long run, the utility of all non-Free software approaches zero. All non-Free software is a dead end.<\/p>\n<\/blockquote>\n<a href=\"#fnref9\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn10\" role=\"doc-endnote\"><p>These dependencies can be subtle. Computer archivist Jason Scott <a href=\"http:\/\/web.archive.org\/web\/20110716201019\/http:\/\/www.archiveteam.org\/archives\/media\/The%20Spendiferous%20Story%20of%20Archive%20Team%20-%20Jason%20Scott%20-%20PDA2011.txt\">writes<\/a> of <a href=\"https:\/\/en.wikipedia.org\/wiki\/URL%20shortening%23Shortcomings\" title=\"Wikipedia: URL shortening#Shortcomings\">URL shortening<\/a> services that:<\/p>\n<blockquote>\n<p>URL shorteners may be one of the worst ideas, one of the most backward ideas, to come out of the last five years. In very recent times, per-site shorteners, where a website registers a smaller version of its hostname and provides a single small link for a more complicated piece of content within it\u2026 those are fine. But these general-purpose URL shorteners, with their shady or fragile setups and utter dependence upon them, well. If we lose <a href=\"https:\/\/en.wikipedia.org\/wiki\/TinyURL\" title=\"Wikipedia: TinyURL\">TinyURL<\/a> or <a href=\"https:\/\/en.wikipedia.org\/wiki\/bit.ly\" title=\"Wikipedia: bit.ly\">bit.ly<\/a>, millions of weblogs, essays, and non-archived tweets lose their meaning. Instantly. To someone in the future, it\u2019ll be like everyone from a certain era of history, say ten years of the 18th century, started speaking in a one-time pad of cryptographic pass phrases. We\u2019re doing our best to stop it. Some of the shorteners have been helpful, others have been hostile. A number have died. We\u2019re going to release torrents on a regular basis of these spreadsheets, these code breaking spreadsheets, and we hope others do too.<\/p>\n<\/blockquote>\n<a href=\"#fnref10\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn11\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Joshua%20Schachter\" title=\"Wikipedia: Joshua Schachter\">Joshua Schachter<\/a> <a href=\"https:\/\/web.archive.org\/web\/20131016131625\/http:\/\/joshua.schachter.org\/2009\/04\/on-url-shorteners.html\">remarks<\/a> (and the comments provide even more examples) further on URL shorteners:<\/p>\n<blockquote>\n<p>But the biggest burden falls on the clicker, the person who follows the links. The extra layer of indirection slows down browsing with additional DNS lookups and server hits. A new and <a href=\"http:\/\/ask.slashdot.org\/article.pl?sid=07\/11\/18\/1319201\">potentially unreliable middleman<\/a> now sits between the link and its destination. And the long-term archivability of the hyperlink now depends on the health of a third party. The shortener may decide a link is a Terms Of Service violation and delete it. If the shortener <a href=\"http:\/\/ma.gnolia.com\/\">accidentally erases a database<\/a>, forgets to renew its domain, or just <a href=\"http:\/\/6uold.blogspot.com\/2008\/06\/long-list-of-url-shorteners.html\">disappears<\/a>, the link will break. If a top-level domain <a href=\"http:\/\/workbench.cadenhead.org\/news\/3503\/bitly-builds-business-libya-domain\">changes its policy on commercial use<\/a>, the link will break. If the shortener gets hacked, every link becomes a potential phishing attack.<\/p>\n<\/blockquote>\n<a href=\"#fnref11\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn12\" role=\"doc-endnote\"><p>A static text-source site has many advantages for Long Content that I consider use almost a no-brainer.<\/p>\n<ul><li>By nature, they compile most content down to flat standalone textual files, which allow recovery of content even if the original site software has bitrotted or the source files have been lost or the compiled versions cannot be directly used in new site software: one can parse them with XML tools or with quick hacks or by eye.<\/li>\n<li>Site compilers generally require dependencies to be declared up front, and the approach makes explicitness and content easy, but dynamic interdependent components difficult, all of which discourages creeping complexity and hidden state.<\/li>\n<li>A static site can be archived into a tarball of files which will be readable as long as web browsers exist (or afterwards if the HTML is reasonably clean), but it could be difficult to archive a CMS like WordPress or Blogspot (the latter doesn\u2019t even provide the content in HTML\u2014it only provides a rat\u2019s-nest of inscrutable JavaScript files which then download the content from <em>somewhere<\/em> and display it <em>somehow<\/em>; indeed, I\u2019m not sure how I would automate archiving of such a site if I had to; I would need some sort of headless browser to run the JS and serialize the final resulting DOM, possibly with some scripting of mouse\/keyboard actions).<\/li>\n<li>The content is often not available locally, or is stored in opaque binary formats rather than text (if one is lucky, it will at least be a database), both of which make it difficult to port content to other website software; you won\u2019t have the necessary pieces, or they will be in wildly incompatible formats.<\/li>\n<li>Static sites are usually written in a reasonably standardized markup language such as Markdown or Latex, in distinction to blogs which force one through WYSIWYG editors or invent their own markup conventions, which is yet another barrier: parsing a possibly ill-defined language.<\/li>\n<li>The lowered sysadmin efforts (who wants to be constantly cleaning up spam or hacks on their WordPress blog?) are a final advantage: lower running costs make it more likely that a site will stay up rather than cease to be worth the hassle.<\/li>\n<\/ul><p>Static sites are not appropriate for many kinds of websites, but they are appropriate for websites which are content-oriented, do not need interactivity, expect to migrate website software several times over coming decades, want to enable archiving by oneself or third parties (\u201clots of copies keeps stuff safe\u201d), and to gracefully degrade after loss or bitrot.<a href=\"#fnref12\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn13\" role=\"doc-endnote\"><p>Such as burning the occasional copy onto read-only media like DVDs.<a href=\"#fnref13\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn14\" role=\"doc-endnote\"><p>One can\u2019t be sure; the IA is fed by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Alexa\" title=\"Wikipedia: Alexa\">Alexa<\/a>, and Alexa doesn\u2019t guarantee pages will be <a href=\"https:\/\/en.wikipedia.org\/wiki\/Web%20crawler\" title=\"Wikipedia: Web crawler\">spidered<\/a> &amp; preserved if one goes through their request form.<a href=\"#fnref14\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn15\" role=\"doc-endnote\"><p>I am diligent in backing up my files, in periodically copying my content from the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cloud%20computing\" title=\"Wikipedia: Cloud computing\">cloud<\/a>, and in preserving viewed Internet content; why do I do all this? Because I want to believe that my memories are precious, that the things I saw and said are valuable; \u201cI want to meet them again, because I believe my feelings at that time were real.\u201d My past is not trash to me, used up &amp; discarded.<a href=\"#fnref15\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn16\" role=\"doc-endnote\"><p>Examples of such blogs:<\/p>\n<ol type=\"1\"><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Eliezer%20Yudkowsky\" title=\"Wikipedia: Eliezer Yudkowsky\">Eliezer Yudkowsky<\/a>\u2019s contributions to <a href=\"http:\/\/www.lesswrong.com\">LessWrong<\/a> were the rough draft of a philosophy book (or two)<\/li>\n<li>John Robb\u2019s <a href=\"http:\/\/globalguerrillas.typepad.com\/\">Global Guerrillas<\/a> lead to his <a href=\"http:\/\/www.amazon.com\/exec\/obidos\/ASIN\/0471780790\/?tag=gwernnet-20\"><em>Brave New War: The Next Stage of Terrorism and the End of Globalization<\/em><\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Kevin%20Kelly\" title=\"Wikipedia: Kevin Kelly\">Kevin Kelly<\/a>\u2019s <a href=\"http:\/\/www.kk.org\/thetechnium\/\">Technium<\/a> was turned into <a href=\"http:\/\/www.amazon.com\/What-Technology-Wants-Kevin-Kelly\/dp\/0670022152\/?tag=gwernnet-20\"><em>What Technology Wants<\/em><\/a>.<\/li>\n<\/ol><p>An example of how <em>not<\/em> to do it would be <a href=\"https:\/\/en.wikipedia.org\/wiki\/Robin%20Hanson\" title=\"Wikipedia: Robin Hanson\">Robin Hanson<\/a>\u2018s <a href=\"http:\/\/www.overcomingbias.com\/\">Overcoming Bias<\/a> blog; it is stuffed with fascinating citations &amp; sketches of ideas, but they never go anywhere with the exception of his mind emulation economy posts which were eventually published in 2016 as <em>The Age of Em<\/em>. Just his posts on <a href=\"http:\/\/www.overcomingbias.com\/tag\/medicine\">medicine<\/a> would make a fascinating essay or just list\u2014but he has never made one. (<a href=\"https:\/\/www.gwern.net\/docs\/economics\/2007-hanson.pdf\">\u201cShowing That You Care: The Evolution of Health Altruism\u201d<\/a> would be a natural home for many of his posts\u2019 contents, but will never be updated.)<a href=\"#fnref16\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn17\" role=\"doc-endnote\"><p><a href=\"http:\/\/interviews.slashdot.org\/story\/11\/09\/06\/1458254\/Kevin-Kelly-Answers-Your-Questions\">\u201cKevin Kelly Answers Your Questions\u201d<\/a>, 6 September 2011:<\/p>\n<blockquote>\n<p>[Question:] \u201cOne purpose of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Long%20Now%20Clock\" title=\"Wikipedia: Long Now Clock\">Long Now Clock<\/a> is to encourage long-term thinking. Aside from the Clock, though, what do you think people can do in their everyday lives to adopt or promote long-term thinking?\u201d<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Kevin%20Kelly%20%28editor%29\" title=\"Wikipedia: Kevin Kelly (editor)\">KK<\/a>: \"The 10,000-year Clock we are building <a href=\"http:\/\/10000yearclock.net\/\">in the hills of west Texas<\/a> is meant to remind us to think long-term, but learning how to do that as in individual is difficult. Part of the difficulty is that as individuals we constrained to short lives, and are inherently not long-term. So part of the skill in thinking long-term is to place our values and energies in ways that transcend the individual \u2013 either in generational projects, or in social enterprises.<\/p>\n<p>As a start I recommend engaging in a project that will not be complete in your lifetime. Another way is to require that your current projects exhibit some payoff that is not immediate; perhaps some small portion of it pays off in the future. A third way is to create things that get better, or run up in time, rather than one that decays and runs down in time. For instance a seedling grows into a tree, which has seedlings of its own. A program like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Heifer%20Project\" title=\"Wikipedia: Heifer Project\">Heifer Project<\/a> which gives breeding pairs of animals to poor farmers, who in turn must give one breeding pair away themselves, is an exotropic scheme, growing up over time.\"<\/p>\n<\/blockquote>\n<a href=\"#fnref17\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn18\" role=\"doc-endnote\"><p>\u2018Princess Irulan\u2019, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Frank%20Herbert\" title=\"Wikipedia: Frank Herbert\">Frank Herbert<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Dune%20%28novel%29\" title=\"Wikipedia: Dune (novel)\"><em>Dune<\/em><\/a><a href=\"#fnref18\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn19\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/GiveWell\" title=\"Wikipedia: GiveWell\">GiveWell<\/a> reports in <a href=\"http:\/\/blog.givewell.org\/2011\/07\/13\/a-good-volunteer-is-hard-to-find\/\">\u201cA good volunteer is hard to find\u201d<\/a> that of volunteers motivated enough to email them asking to help, something like &lt;20% will complete the GiveWell test assignment and render meaningful help. Such persons would have been well-advised to have simply donated some money. I have long noted that many of the most popular pages on <code>gwern.net<\/code> could have been written by anyone and drew on no unique talents of mine; I have on several occasions received offers to help with the DNB FAQ\u2014none of which have resulted in <em>actual<\/em> help.<a href=\"#fnref19\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn20\" role=\"doc-endnote\"><p>An old sentiment; consider \u201cA drop hollows out the stone\u201d (Ovid, <em>Epistles<\/em>) or Thomas Carlyle\u2019s \u201cThe weakest living creature, by concentrating his powers on a single object, can accomplish something. The strongest, by dispensing his over many, may fail to accomplish anything. The drop, by continually falling, bores its passage through the hardest rock. The hasty torrent rushes over it with hideous uproar, and leaves no trace behind.\u201d (<em>The life of Friedrich Schiller<\/em>, 1825)<a href=\"#fnref20\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn21\" role=\"doc-endnote\"><p><a href=\"http:\/\/alumni.media.mit.edu\/~cahn\/life\/gian-carlo-rota-10-lessons.html#feynmann\">\u201cTen Lessons I wish I had been Taught\u201d<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gian-Carlo%20Rota\" title=\"Wikipedia: Gian-Carlo Rota\">Gian-Carlo Rota<\/a>:<\/p>\n<blockquote>\n<p>Richard Feynman was fond of giving the following advice on how to be a genius. You have to keep a dozen of your favorite problems constantly present in your mind, although by and large they will lay in a dormant state. Every time you hear or read a new trick or a new result, test it against each of your twelve problems to see whether it helps. Every once in a while there will be a hit, and people will say: \u2018<em>How did he do it? He must be a genius<\/em>!\u2019<\/p>\n<\/blockquote>\n<a href=\"#fnref21\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn22\" role=\"doc-endnote\"><p>IQ is sometimes used as a proxy for health, like height, because it sometimes seems like any health problem will damage IQ. Didn\u2019t get much protein as a kid? Congratulations, your nerves will lack <a href=\"https:\/\/en.wikipedia.org\/wiki\/myelination\" title=\"Wikipedia: myelination\">myelination<\/a> and you will literally think slower. Missing some <a href=\"https:\/\/www.gwern.net\/Iodine\">iodine<\/a>? Say good bye to &lt;10 points! If you\u2019re anemic or iron-deficient, that might increase to &lt;15 points. Have tapeworms? There go some more points, and maybe centimeters off your adult height, thanks to the worms stealing nutrients from you. Have a rough birth and suffer a spot of <a href=\"https:\/\/en.wikipedia.org\/wiki\/hypoxia\" title=\"Wikipedia: hypoxia\">hypoxia<\/a> before you began breathing on your own? Tough luck, old bean. It is very easy to <em>lower<\/em> IQ; you can do it with a baseball bat. It\u2019s the other way around that\u2019s nearly impossible.<a href=\"#fnref22\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn23\" role=\"doc-endnote\"><p>And America has tried pretty hard over the past 60 years to affect IQ. The whole nature\/nurture <a href=\"https:\/\/en.wikipedia.org\/wiki\/Race%20and%20intelligence\" title=\"Wikipedia: Race and intelligence\">black-white IQ debate<\/a> would be moot if there were some nutrient or educational system which could add even 10 points on average, because then we would use it on all the blacks. But it seems that I\u2019m constantly reading about programs like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Headstart\" title=\"Wikipedia: Headstart\">Headstart<\/a> which boost IQ for a little while\u2026 and do nothing in the long run.<a href=\"#fnref23\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn24\" role=\"doc-endnote\"><p>For details on the many valuable correlates of the Conscientiousness personality factor, see <a href=\"https:\/\/www.gwern.net\/Conscientiousness-and-online-education#conscientiousness\">Conscientiousness and online education<\/a>.<a href=\"#fnref24\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn25\" role=\"doc-endnote\"><p>25 episodes, 6 movies, &gt;11 manga volumes\u2014just to stick to the core works.<a href=\"#fnref25\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn26\" role=\"doc-endnote\"><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mibu%20no%20Tadamine\" title=\"Wikipedia: Mibu no Tadamine\">Mibu no Tadamine<\/a>, <a href=\"http:\/\/www.temcauley.staff.shef.ac.uk\/waka0627.shtml\"><em>KKS<\/em> XII: 609<\/a>:<\/p>\n<pre>More than my life\nWhat I most regret\nIs\nA dream unfinished\nAnd awakening.<\/pre>\n<a href=\"#fnref26\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn27\" role=\"doc-endnote\"><p>As with <em>Cloud Nine<\/em>; I accidentally erased everything on a routine basis while messing around with Windows.<a href=\"#fnref27\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn28\" role=\"doc-endnote\"><p>For example, I notice I am no longer deeply interested in the occult. Hopefully this is because I have grown mentally and recognize it as rubbish; I would be embarrassed if when I died it turned out my youthful self had a better grasp on the real world.<a href=\"#fnref28\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn29\" role=\"doc-endnote\"><p>Some pages don\u2019t have any connection to predictions. It\u2019s possible to make predictions for some border cases like the terrorism essays (death tolls, achievements of particular groups\u2019 policy goals), but what about the short stories or poems? My imagination fails there.<a href=\"#fnref29\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn30\" role=\"doc-endnote\"><p>Thinking of predictions is good mental discipline; we should always be able to <a href=\"http:\/\/lesswrong.com\/lw\/i3\/making_beliefs_pay_rent_in_anticipated_experiences\/\">cash out<\/a> our beliefs in terms of the real world, or know why we cannot. Unfortunately, humans being humans, we need to actually track our predictions\u2014<a href=\"https:\/\/en.wikipedia.org\/wiki\/Confirmation%20bias\" title=\"Wikipedia: Confirmation bias\"><em>all<\/em> of them<\/a>\u2014lest our predicting degenerate into <a href=\"http:\/\/lesswrong.com\/lw\/hi\/futuristic_predictions_as_consumable_goods\/\">entertainment<\/a> like political punditry.<a href=\"#fnref30\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn31\" role=\"doc-endnote\"><p>Dozens of theories have been put forth. I have been collecting &amp; making predictions; and am up to 219. It will be interesting to see how the movies turn out.<a href=\"#fnref31\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn32\" role=\"doc-endnote\"><p>I have 2 predictions registered about the thesis on PB.com: <a href=\"http:\/\/predictionbook.com\/predictions\/1833\">1 reviewer will accept my theory by 2016<\/a> and <a href=\"http:\/\/predictionbook.com\/predictions\/1832\">the light novels will finish by 2015<\/a>.<a href=\"#fnref32\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn33\" role=\"doc-endnote\"><p>See Robin Hanson, <a href=\"http:\/\/hanson.gmu.edu\/uploads.html\">\u201cIf Uploads Come First\u201d<\/a><a href=\"#fnref33\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn34\" role=\"doc-endnote\"><p>Anecdotally, the rankings seem correct. When I went to a <a href=\"http:\/\/lesswrong.com\/\">LessWrong<\/a> meetup in California, many knew of or had read the DNB FAQ, some had read or used my modafinil price-chart, and very few remembered reading anything else.<a href=\"#fnref34\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn35\" role=\"doc-endnote\"><p>I am sometimes reminded of another waka, by <a href=\"http:\/\/thegreenleaf.co.uk\/hp\/Ikkyu\/00haiku.htm\">Ikkyu<\/a>:<\/p>\n<blockquote>\n<p>To write something and leave it behind us,<br \/>Is but a dream.<br \/>When we awake we know<br \/>There is not even anyone to read it.<\/p>\n<\/blockquote>\n<a href=\"#fnref35\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn36\" role=\"doc-endnote\"><p>They spend an average of 27 seconds; in comparison, my <em>second<\/em> largest source of traffic, LessWrongers, average 3 minutes and 29 seconds; even my third largest traffic source, Redditers, manage almost 2 minutes. Even random people coming from Google manage to spend 44 seconds on their visit!<a href=\"#fnref36\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn37\" role=\"doc-endnote\"><p>Since I converted the repo to use git, <code>darcs-graph<\/code> was no longer available, so I wrote my own script in R using splines rather than a 7-day moving average:<\/p>\n<div class=\"sourceCode\" id=\"cb5\"><pre class=\"sourceCode r\"><a class=\"sourceLine\" id=\"cb5-1\" title=\"1\">start &lt;- as.Date(\"2014-01-02\")<\/a>\n<a class=\"sourceLine\" id=\"cb5-2\" title=\"2\">end &lt;- as.Date(\"2014-07-02\")<\/a>\n<a class=\"sourceLine\" id=\"cb5-4\" title=\"4\">patches &lt;- as.Date(system(paste0(\"git log --after=\", start, \" --before=\", end,<\/a>\n<a class=\"sourceLine\" id=\"cb5-5\" title=\"5\">                           \" --format='%ad' --date=short master\"),<\/a>\n<a class=\"sourceLine\" id=\"cb5-6\" title=\"6\">                   intern=TRUE))<\/a>\n<a class=\"sourceLine\" id=\"cb5-7\" title=\"7\">patchFrame &lt;- aggregate( patches , by = list(patches) , length )<\/a>\n<a class=\"sourceLine\" id=\"cb5-9\" title=\"9\">days &lt;- seq(start, end, \"day\")<\/a>\n<a class=\"sourceLine\" id=\"cb5-10\" title=\"10\">dayFrame &lt;- aggregate( days , by = list(days) , length )<\/a>\n<a class=\"sourceLine\" id=\"cb5-12\" title=\"12\">bar &lt;- rbind(dayFrame, patchFrame)<\/a>\n<a class=\"sourceLine\" id=\"cb5-14\" title=\"14\">all &lt;- aggregate(bar$x, by=list(bar$Group.1), FUN=function (x) { sum(as.vector(x));})<\/a>\n<a class=\"sourceLine\" id=\"cb5-16\" title=\"16\">plot(x ~ Group.1, data=all, xlab=\"Date (Jan--July 2014)\", ylab=\"gwern.net daily patches\")<\/a>\n<a class=\"sourceLine\" id=\"cb5-17\" title=\"17\">library(splines); lines(predict(interpSpline(days, all$x)))<\/a><\/pre><\/div>\n<a href=\"#fnref37\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn38\" role=\"doc-endnote\"><p>I rewrote the script a little to use the nicer plots of <a href=\"http:\/\/ggplot2.org\/\">ggplot2<\/a> and switches to LOESS for the smoothed curve:<\/p>\n<div class=\"sourceCode\" id=\"cb6\"><pre class=\"sourceCode r\"><a class=\"sourceLine\" id=\"cb6-1\" title=\"1\">start &lt;- as.Date(\"2014-07-03\")<\/a>\n<a class=\"sourceLine\" id=\"cb6-2\" title=\"2\">end &lt;- as.Date(\"2015-01-02\")<\/a>\n<a class=\"sourceLine\" id=\"cb6-4\" title=\"4\">patches &lt;- as.Date(system(paste0(\"git --git-dir=\/home\/gwern\/wiki\/.git\/ log --after=\",<\/a>\n<a class=\"sourceLine\" id=\"cb6-5\" title=\"5\">                          start, \" --before=\", end,<\/a>\n<a class=\"sourceLine\" id=\"cb6-6\" title=\"6\">                          \" --format='%ad' --date=short master\"), intern=TRUE))<\/a>\n<a class=\"sourceLine\" id=\"cb6-7\" title=\"7\">patchFrame &lt;- aggregate( patches , by = list(patches) , length )<\/a>\n<a class=\"sourceLine\" id=\"cb6-9\" title=\"9\">days &lt;- seq(start, end, \"day\")<\/a>\n<a class=\"sourceLine\" id=\"cb6-10\" title=\"10\">dayFrame &lt;- aggregate( days , by = list(days) , length )<\/a>\n<a class=\"sourceLine\" id=\"cb6-12\" title=\"12\">bar &lt;- rbind(dayFrame, patchFrame)<\/a>\n<a class=\"sourceLine\" id=\"cb6-14\" title=\"14\">all &lt;- aggregate(bar$x, by=list(bar$Group.1), FUN=function (x) { sum(as.vector(x));})<\/a>\n<a class=\"sourceLine\" id=\"cb6-16\" title=\"16\">library(ggplot2)<\/a>\n<a class=\"sourceLine\" id=\"cb6-17\" title=\"17\">qplot(Group.1, x, data=all) +<\/a>\n<a class=\"sourceLine\" id=\"cb6-18\" title=\"18\"> theme_bw() +<\/a>\n<a class=\"sourceLine\" id=\"cb6-19\" title=\"19\"> xlab(\"Date (2014)\") + ylab(\"Patch count\") +<\/a>\n<a class=\"sourceLine\" id=\"cb6-20\" title=\"20\"> theme(legend.title=element_blank()) +<\/a>\n<a class=\"sourceLine\" id=\"cb6-21\" title=\"21\"> stat_smooth()<\/a><\/pre><\/div>\n<a href=\"#fnref38\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn39\" role=\"doc-endnote\"><p>I like the static site approach to things; it tends to be harder to use and more restrictive, but in exchange it yields <a href=\"http:\/\/inessential.com\/2011\/03\/16\/a_plea_for_baked_weblogs\">better performance<\/a> &amp; leads to fewer <a href=\"http:\/\/www.aaronsw.com\/weblog\/000404\">hassles or runtime issues<\/a>.<a href=\"#fnref39\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn40\" role=\"doc-endnote\"><p>We write a short Haskell program as part of a pipeline:<\/p>\n<div class=\"sourceCode\" id=\"cb7\"><pre class=\"sourceCode bash\"><a class=\"sourceLine\" id=\"cb7-1\" title=\"1\">echo '{-# LANGUAGE OverloadedStrings #-};<\/a>\n<a class=\"sourceLine\" id=\"cb7-2\" title=\"2\">      import Data.Text as T;<\/a>\n<a class=\"sourceLine\" id=\"cb7-3\" title=\"3\">      main = interact (T.unpack . T.unlines . Prelude.filter (\/=\"\") .<\/a>\n<a class=\"sourceLine\" id=\"cb7-4\" title=\"4\">                       T.split (not . (`elem` \"0123456789,.\")) . T.pack)' &gt; ~\/number.hs &amp;&amp;<\/a>\n<a class=\"sourceLine\" id=\"cb7-5\" title=\"5\">find ~\/wiki\/ -type f -name \"*.page\" -exec cat \"{}\" \\; | runhaskell ~\/number.hs |<\/a>\n<a class=\"sourceLine\" id=\"cb7-6\" title=\"6\"> sort | tr -d ',' | tr -d '.' | cut -c 1 | sed -e 's\/0$\/\/' -e '\/^$\/d' &gt; ~\/number.txt<\/a><\/pre><\/div>\n<a href=\"#fnref40\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn41\" role=\"doc-endnote\"><p>Graph then test:<\/p>\n<div class=\"sourceCode\" id=\"cb8\"><pre class=\"sourceCode r\"><a class=\"sourceLine\" id=\"cb8-1\" title=\"1\">numbers &lt;- read.table(\"number.txt\")<\/a>\n<a class=\"sourceLine\" id=\"cb8-2\" title=\"2\">ta &lt;- table(numbers$V1); ta<\/a>\n<a class=\"sourceLine\" id=\"cb8-4\" title=\"4\">#     1     2     3     4     5     6     7     8     9<\/a>\n<a class=\"sourceLine\" id=\"cb8-5\" title=\"5\"># 20550 20356  7087  5655  3900  2508  2075  2349  2068<\/a>\n<a class=\"sourceLine\" id=\"cb8-6\" title=\"6\">## cribbing exact R code from http:\/\/www.math.utah.edu\/~treiberg\/M3074BenfordEg.pdf<\/a>\n<a class=\"sourceLine\" id=\"cb8-7\" title=\"7\">sta &lt;- sum(ta)<\/a>\n<a class=\"sourceLine\" id=\"cb8-8\" title=\"8\">pb &lt;- sapply(1:9, function(x) log10(1+1\/x)); pb<\/a>\n<a class=\"sourceLine\" id=\"cb8-9\" title=\"9\">m &lt;- cbind(ta\/sta,pb)<\/a>\n<a class=\"sourceLine\" id=\"cb8-10\" title=\"10\">colnames(m)&lt;- c(\"Observed Prop.\", \"Theoretical Prop.\")<\/a>\n<a class=\"sourceLine\" id=\"cb8-11\" title=\"11\">barplot( rbind(ta\/sta,pb\/sum(pb)), beside = T, col = rainbow(7)[c(2,5)],<\/a>\n<a class=\"sourceLine\" id=\"cb8-12\" title=\"12\">              xlab = \"First Digit\")<\/a>\n<a class=\"sourceLine\" id=\"cb8-13\" title=\"13\">title(\"Benford's Law Compared to Writing Data\")<\/a>\n<a class=\"sourceLine\" id=\"cb8-14\" title=\"14\">legend(16,.28, legend = c(\"From Page Data\", \"Theoretical\"),<\/a>\n<a class=\"sourceLine\" id=\"cb8-15\" title=\"15\">       fill = rainbow(7)[c(2,5)],bg=\"white\")<\/a>\n<a class=\"sourceLine\" id=\"cb8-16\" title=\"16\">chisq.test(ta,p=pb)<\/a>\n<a class=\"sourceLine\" id=\"cb8-17\" title=\"17\">#<\/a>\n<a class=\"sourceLine\" id=\"cb8-18\" title=\"18\">#     Chi-squared test for given probabilities<\/a>\n<a class=\"sourceLine\" id=\"cb8-19\" title=\"19\">#<\/a>\n<a class=\"sourceLine\" id=\"cb8-20\" title=\"20\"># data:  ta<\/a>\n<a class=\"sourceLine\" id=\"cb8-21\" title=\"21\"># X-squared = 9331, df = 8, p-value &lt; 2.2e-16<\/a><\/pre><\/div>\n<a href=\"#fnref41\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn42\" role=\"doc-endnote\"><p>Paraphrased from <em>Dialogues of the Zen Masters<\/em> as quoted in pg 11 of the Editor\u2019s Introduction to <a href=\"http:\/\/www.amazon.com\/Three-Pillars-Zen-Teaching-Enlightenment\/dp\/0385260938\/?tag=gwernnet-20\"><em>Three Pillars of Zen<\/em><\/a>:<\/p>\n<blockquote>\n<p>One day a man of the people said to Master Ikkyu: \u201cMaster, will you please write for me maxims of the highest wisdom?\u201d Ikkyu immediately brushed out the word \u2018Attention\u2019. \u201cIs that all? Will you not write some more?\u201d Ikkyu then brushed out twice: \u2018Attention. Attention.\u2019 The man remarked irritably that there wasn\u2019t much depth or subtlety to that. Then Ikkyu wrote the same word 3 times running: \u2018Attention. Attention. Attention.\u2019 Half-angered, the man demanded: \u201cWhat does \u2018Attention\u2019 mean anyway?\u201d And Ikkyu answered gently: \u201cAttention means attention.\u201d<\/p>\n<\/blockquote>\n<a href=\"#fnref42\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/li>\n<li id=\"fn43\" role=\"doc-endnote\"><p>PD increases economic efficiency through\u2014if nothing else\u2014making works easier to find. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Tim%20O%27Reilly\" title=\"Wikipedia: Tim O\u2019Reilly\">Tim O\u2019Reilly<\/a> says that <a href=\"http:\/\/www.openp2p.com\/pub\/a\/p2p\/2002\/12\/11\/piracy.html\" title=\"'Piracy is Progressive Taxation, and Other Thoughts on the Evolution of Online Distribution', O`Reilly 2002\">\u201cObscurity is a far greater threat to authors and creative artists than piracy.\u201d<\/a> If that is so, then that means that difficulty of finding works reduces the welfare of artists <em>and<\/em> consumers, because both forgo a beneficial trade (the artist loses any revenue and the consumer loses any enjoyment). Even small increases in inconvenience make <a href=\"https:\/\/www.gwern.net\/In-Defense-Of-Inclusionism#new-regimes\">big differences<\/a>.<a href=\"#fnref43\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn44\" role=\"doc-endnote\"><p>Not that I <em>could<\/em> sell anything on this wiki; and if I could, I would polish it as much as possible, giving me fresh copyright.<a href=\"#fnref44\" class=\"footnote-back\" role=\"doc-backlink\">\u21a9<\/a><\/p><\/li>\n<\/ol><\/section>","created_at":"2019-04-09T18:31:22+0000","updated_at":"2019-04-09T18:31:22+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":90,"domain_name":"www.gwern.net","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":904,"title":"Tufte CSS","url":"https:\/\/edwardtufte.github.io\/tufte-css\/","content":"<p class=\"subtitle\">Dave Liepmann<\/p>\n<section><p>Tufte CSS provides tools to style web articles using the ideas demonstrated by Edward Tufte\u2019s books and handouts. Tufte\u2019s style is known for its simplicity, extensive use of sidenotes, tight integration of graphics with text, and carefully chosen typography.<\/p><p>Tufte CSS was created by <a href=\"http:\/\/www.daveliepmann.com\">Dave Liepmann<\/a> and is now an Edward Tufte project. The original idea was cribbed from <a href=\"https:\/\/tufte-latex.github.io\/tufte-latex\/\">Tufte-LaTeX<\/a> and <a href=\"http:\/\/rmarkdown.rstudio.com\/tufte_handout_format.html\">R Markdown\u2019s Tufte Handout format<\/a>. We give hearty thanks to all the people who have contributed to those projects.<\/p><p>If you see anything that Tufte CSS could improve, we welcome your contribution in the form of an issue or pull request on the GitHub project: <a href=\"https:\/\/github.com\/edwardtufte\/tufte-css\">tufte-css<\/a>. Please note the <a href=\"https:\/\/github.com\/edwardtufte\/tufte-css#contributing\">contribution guidelines<\/a>.<\/p><p>Finally, a reminder about the goal of this project. The web is not print. Webpages are not books. Therefore, the goal of Tufte CSS is not to say \u201cwebsites should look like this interpretation of Tufte\u2019s books\u201d but rather \u201chere are some techniques Tufte developed that we\u2019ve found useful in print; maybe you can find a way to make them useful on the web\u201d. Tufte CSS is merely a sketch of one way to implement this particular set of ideas. It should be a starting point, not a design goal, because any project should present their information as best suits their particular circumstances.<\/p><\/section><section><h2 id=\"getting-started\">Getting Started<\/h2><p>To use Tufte CSS, copy <code>tufte.css<\/code> and the <code>et-book<\/code> directory of font files to your project directory, then add the following to your HTML document\u2019s <code>head<\/code> block:<\/p><pre class=\"code\">&lt;link rel=\"stylesheet\" href=\"tufte.css\"\/&gt;<\/pre><p>Now you just have to use the provided CSS rules, and the Tufte CSS conventions described in this document. For best results, View Source and Inspect Element frequently.<\/p><\/section><section><h2 id=\"fundamentals\">Fundamentals<\/h2><p>Organize your document with an <code>article<\/code> element inside your <code>body<\/code> tag. Inside that, use <code>section<\/code> tags around each logical grouping of text and headings.<\/p><p>Tufte CSS uses <code>h1<\/code> for the document title, <code>p<\/code> with class <code>subtitle<\/code> for the document subtitle, <code>h2<\/code> for section headings, and <code>h3<\/code> for low-level headings. More specific headings are not supported. If you feel the urge to reach for a heading of level 4 or greater, consider redesigning your document:<\/p><blockquote cite=\"http:\/\/www.edwardtufte.com\/bboard\/q-and-a-fetch-msg?msg_id=0000hB\">\n<p>[It is] notable that the Feynman lectures (3 volumes) write about all of physics in 1800 pages, using only 2 levels of hierarchical headings: chapters and A-level heads in the text. It also uses the methodology of <em>sentences<\/em> which then cumulate sequentially into <em>paragraphs<\/em>, rather than the grunts of bullet points. Undergraduate Caltech physics is very complicated material, but it didn\u2019t require an elaborate hierarchy to organize.<\/p>\n<footer><a href=\"http:\/\/www.edwardtufte.com\/bboard\/q-and-a-fetch-msg?msg_id=0000hB\">Edward Tufte, forum post, \u2018Book design: advice and examples\u2019 thread<\/a><\/footer><\/blockquote><p>As a bonus, this excerpt regarding the use of headings provides an example of block quotes. In Tufte CSS they are just lightly styled, semantically correct HTML using <code>blockquote<\/code> and <code>footer<\/code> elements. See page 20 of <a href=\"https:\/\/www.edwardtufte.com\/tufte\/books_vdqi\">The Visual Display of Quantitative Information<\/a> for an example in print.<\/p><p>In his later books<label for=\"sn-in-his-later-books\" class=\"margin-toggle sidenote-number\"><a href=\"http:\/\/www.edwardtufte.com\/tufte\/books_be\"><em>Beautiful Evidence<\/em><\/a>, Tufte starts each section with a bit of vertical space, a non-indented paragraph, and the first few words of the sentence set in small caps. For this we use a span with the class <code>newthought<\/code>, as demonstrated at the beginning of this paragraph. Vertical spacing is accomplished separately through <code>&lt;section&gt;<\/code> tags. Be consistent: though we do so in this paragraph for the purpose of demonstration, do not alternate use of header elements and the <code>newthought<\/code> technique. Pick one approach and stick to it.<\/label><\/p><h3 id=\"fundamentals--text\">Text<\/h3><p>Although paper handouts obviously have a pure white background, the web is better served by the use of slightly off-white and off-black colors. Tufte CSS uses <code>#fffff8<\/code> and <code>#111111<\/code> because they are nearly indistinguishable from their \u2018pure\u2019 cousins, but dial down the harsh contrast. We stick to the greyscale for text, reserving color for specific, careful use in figures and images.<\/p><p>In print, Tufte has used the proprietary Monotype Bembo<label for=\"sn-proprietary-monotype-bembo\" class=\"margin-toggle sidenote-number\">See Tufte\u2019s comment in the <a href=\"http:\/\/www.edwardtufte.com\/bboard\/q-and-a-fetch-msg?msg_id=0000Vt\">Tufte book fonts<\/a> thread. font. A similar effect is achieved in digital formats with the now open-source <a href=\"https:\/\/github.com\/edwardtufte\/et-book\">ETBook<\/a>, which Tufte CSS supplies with a <code>@font-face<\/code> reference to a .ttf file. In case ETBook somehow doesn\u2019t work, Tufte CSS shifts gracefully to other serif fonts like Palatino and Georgia.<\/label><\/p><p>Also notice how Tufte CSS includes separate font files for bold (strong) and italic (emphasis), instead of relying on the browser to mechanically transform the text. This is typographic best practice.<\/p><p class=\"sans\">If you prefer sans-serifs, use the <code>sans<\/code> class. It relies on Gill Sans, Tufte\u2019s sans-serif font of choice.<\/p><p>Links in Tufte CSS match the body text in color and do not change on mouseover or when clicked. Here is a <a href=\"#\">dummy example<\/a> that goes nowhere. These links are underlined, since this is the most widely recognized indicator of clickable text. <label for=\"mn-blue-links\" class=\"margin-toggle\">\u2295<\/label>Blue text, while also a widely recognizable clickable-text indicator, is crass and distracting. Luckily, it is also rendered unnecessary by the use of underlining. However, because most browsers\u2019 default underlining does not clear descenders and is so thick and distracting, the underline effect is instead achieved using CSS trickery involving background gradients instead of standard <code>text-decoration<\/code>. Credit goes to Adam Schwartz for that technique.<\/p><p>As always, these design choices are merely one approach that Tufte CSS provides by default. Other approaches, such as changing color on click or mouseover, or using highlighting or color instead of underlining to denote links, could also be made to work. The goal is to make sentences readable without interference from links, as well as to make links immediately identifiable even by casual web users.<\/p><\/section><section><h2 id=\"epigraphs\">Epigraphs<\/h2><div class=\"epigraph\"><blockquote>\n<p>The English language . . . becomes ugly and inaccurate because our thoughts are foolish, but the slovenliness of our language makes it easier for us to have foolish thoughts.<\/p>\n<footer>George Orwell, \u201cPolitics and the English Language\u201d<\/footer><\/blockquote><blockquote>\n<p>For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled.<\/p>\n<footer>Richard P. Feynman, <cite>\u201cWhat Do You Care What Other People Think?\u201d<\/cite><\/footer><\/blockquote><blockquote>\n<div>I do not paint things, I paint only the differences between things.<footer>Henri Matisse, <cite>Henri Matisse Dessins: th\u00e8mes et variations<\/cite> (Paris, 1943), 37<\/footer><\/div>\n<\/blockquote><\/div><p>If you\u2019d like to introduce your page or a section of your page with some quotes, use epigraphs. Modeled after chapter epigraphs in Tufte\u2019s books (particularly <em>Beautiful Evidence<\/em>), these are <code>blockquote<\/code> elements with a bit of specialized styling. Quoted text is italicized. The source goes in a <code>footer<\/code> element inside the <code>blockquote<\/code>. We have provided three examples in the epigraph of this section, demonstrating shorter and longer quotes, with and without a paragraph tag, and showing how multiple quotes within an epigraph fit together with the use of a wrapper class.<\/p><\/section><section><h2 id=\"sidenotes\">Sidenotes: Footnotes and Marginal Notes<\/h2><p>One of the most distinctive features of Tufte\u2019s style is his extensive use of sidenotes.<label for=\"sn-extensive-use-of-sidenotes\" class=\"margin-toggle sidenote-number\">This is a sidenote. Sidenotes are like footnotes, except they don\u2019t force the reader to jump their eye to the bottom of the page, but instead display off to the side in the margin. Perhaps you have noticed their use in this document already. You are very astute.<\/label><\/p><p>Sidenotes are a great example of the web not being like print. On sufficiently large viewports, Tufte CSS uses the margin for sidenotes, margin notes, and small figures. On smaller viewports, elements that would go in the margin are hidden until the user toggles them into view. The goal is to present related but not necessary information such as asides or citations <em>as close as possible<\/em> to the text that references them. At the same time, this secondary information should stay out of the way of the eye, not interfering with the progression of ideas in the main text.<\/p><p>Sidenotes consist of two elements: a superscript reference number that goes inline with the text, and a sidenote with content. To add the former, just put a label and dummy checkbox into the text where you want the reference to go, like so:<\/p><pre class=\"code\">\n&lt;label for=\"sn-demo\"\n       class=\"margin-toggle sidenote-number\"&gt;\n&lt;\/label&gt;\n&lt;input type=\"checkbox\"\n       id=\"sn-demo\"\n       class=\"margin-toggle\"\/&gt;<\/pre><p>You must manually assign a reference <code>id<\/code> to each side or margin note, replacing \u201csn-demo\u201d in the <code>for<\/code> and the <code>id<\/code> attribute values with an appropriate descriptor. It is useful to use prefixes like <code>sn-<\/code> for sidenotes and <code>mn-<\/code> for margin notes.<\/p><p>Immediately adjacent to that sidenote reference in the main text goes the sidenote content itself, in a <code>span<\/code> with class <code>sidenote<\/code>. This tag is also inserted directly in the middle of the body text, but is either pushed into the margin or hidden by default. Make sure to position your sidenotes correctly by keeping the sidenote-number label close to the sidenote itself.<\/p><p>If you want a sidenote without footnote-style numberings, then you want a margin note. <label for=\"mn-demo\" class=\"margin-toggle\">\u2295<\/label> This is a margin note. Notice there isn\u2019t a number preceding the note. On large screens, a margin note is just a sidenote that omits the reference number. This lessens the distracting effect taking away from the flow of the main text, but can increase the cognitive load of matching a margin note to its referent text. However, on small screens, a margin note is like a sidenote except its viewability-toggle is a symbol rather than a reference number. This document currently uses the symbol \u2295 (<code>&amp;#8853;<\/code>), but it\u2019s up to you.<\/p><p>Margin notes are created just like sidenotes, but with the <code>marginnote<\/code> class for the content and the <code>margin-toggle<\/code> class for the label and dummy checkbox. For instance, here is the code for the margin note used in the previous paragraph:<\/p><pre class=\"code\">\n&lt;label for=\"mn-demo\" class=\"margin-toggle\"&gt;&amp;#8853;&lt;\/label&gt;\n&lt;input type=\"checkbox\" id=\"mn-demo\" class=\"margin-toggle\"\/&gt;\n&lt;span class=\"marginnote\"&gt;\n  This is a margin note. Notice there isn\u2019t a number preceding the note.\n&lt;\/span&gt;<\/pre><p>Figures in the margin are created as margin notes, as demonstrated in the next section.<\/p><\/section><section><h2 id=\"figures\">Figures<\/h2><p>Tufte emphasizes tight integration of graphics with text. Data, graphs, and figures are kept with the text that discusses them. In print, this means they are not relegated to a separate page. On the web, that means readability of graphics and their accompanying text without extra clicks, tab-switching, or scrolling.<\/p><p>Figures should try to use the <code>figure<\/code> element, which by default are constrained to the main column. Don\u2019t wrap figures in a paragraph tag. Any label or margin note goes in a regular margin note inside the figure. For example, most of the time one should introduce a figure directly into the main flow of discussion, like so:<\/p><figure><label for=\"mn-exports-imports\" class=\"margin-toggle\">\u2295<\/label>From Edward Tufte, <em>Visual Display of Quantitative Information<\/em>, page 92. <img src=\"https:\/\/edwardtufte.github.io\/tufte-css\/img\/exports-imports.png\" alt=\"Exports and Imports to and from Denmark &amp; Norway from 1700 to 1780\" referrerpolicy=\"no-referrer\" \/><\/figure><p><label for=\"mn-figure-1\" class=\"margin-toggle\">\u2295<\/label><img src=\"https:\/\/edwardtufte.github.io\/tufte-css\/img\/rhino.png\" alt=\"Image of a Rhinoceros\" referrerpolicy=\"no-referrer\" \/>F.J. Cole, \u201cThe History of Albrecht D\u00fcrer\u2019s Rhinoceros in Zooological Literature,\u201d <em>Science, Medicine, and History: Essays on the Evolution of Scientific Thought and Medical Practice<\/em> (London, 1953), ed. E. Ashworth Underwood, 337-356. From page 71 of Edward Tufte\u2019s <em>Visual Explanations<\/em>. But tight integration of graphics with text is central to Tufte\u2019s work even when those graphics are ancillary to the main body of a text. In many of those cases, a margin figure may be most appropriate. To place figures in the margin, just wrap an image (or whatever) in a margin note inside a <code>p<\/code> tag, as seen to the right of this paragraph.<\/p><p>If you need a full-width figure, give it the <code>fullwidth<\/code> class. Make sure that\u2019s inside an <code>article<\/code>, and it will take up (almost) the full width of the screen. This approach is demonstrated below using Edward Tufte\u2019s English translation of the Napoleon\u2019s March data visualization. From <em>Beautiful Evidence<\/em>, page 122-124.<\/p><figure class=\"fullwidth\"><img src=\"https:\/\/edwardtufte.github.io\/tufte-css\/img\/napoleons-march.png\" alt=\"Figurative map of the successive losses of the French Army in the Russian campaign, 1812-1813\" referrerpolicy=\"no-referrer\" \/><\/figure><p>One obstacle to creating elegant figures on the web is the difficulty of handling different screen sizes, especially on the fly. Embedded <code>iframe<\/code> elements are particularly troublesome. For these instances we provide a helper class, <code>iframe-wrapper<\/code>, the most common use for which is probably YouTube videos, e.g.<\/p><pre class=\"code\">\n&lt;figure class=\"iframe-wrapper\"&gt;\n  &lt;iframe width=\"853\" height=\"480\" src=\"https:\/\/www.youtube.com\/embed\/YslQ2625TR4\" frameborder=\"0\" allowfullscreen&gt;&lt;\/iframe&gt;\n&lt;\/figure&gt;\n        <\/pre><figure class=\"iframe-wrapper\"><iframe width=\"853\" height=\"480\" src=\"https:\/\/www.youtube.com\/embed\/YslQ2625TR4\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/figure><p>You can use this class on a <code>div<\/code> instead of a <code>figure<\/code>, with slightly different results but the same general effect. Experiment and choose depending on your application.<\/p><\/section><section><h2 id=\"code\">Code<\/h2><p>Technical jargon, programming language terms, and code samples are denoted with the <code>code<\/code> class, as I\u2019ve been using in this document to denote HTML. Code needs to be monospace for formatting purposes and to aid in code analysis, but it must maintain its readability. To those ends, Tufte CSS follows GitHub\u2019s font selection, which shifts gracefully along the monospace spectrum from the elegant but rare Consolas all the way to good old reliable Courier.<\/p><p>Extended code examples should use a <code>pre<\/code> tag with class <code>code<\/code>. This adds control over indentation and overflow as well:<\/p><pre class=\"code\">;; Some code examples in Clojure. This is a comment.\n;; applying a function to every item in the collection\n(map tufte-css blog-posts)\n;;;; if unfamiliar, see http:\/\/www.lispcast.com\/annotated-map\n;; side-effecty loop (unformatted, causing text overflow) - from https:\/\/clojuredocs.org\/clojure.core\/doseq\n(doseq [[[a b] [c d]] (map list (sorted-map :1 1 :2 2) (sorted-map :3 3 :4 4))] (prn (* b d)))\n;; that same side-effecty loop, formatted\n(doseq [[[a b] [c d]] (map list\n                           (sorted-map :1 1 :2 2)\n                           (sorted-map :3 3 :4 4))]\n  (prn (* b d)))\n;; If this proselytizing has worked, check out:\n;; http:\/\/howistart.org\/posts\/clojure\/1<\/pre><\/section><section><h2 id=\"imagequilts\">ImageQuilts<\/h2><p>Tufte CSS provides support for Edward Tufte and Adam Schwartz\u2019s <a href=\"http:\/\/imagequilts.com\/\">ImageQuilts<\/a>. See the <a href=\"http:\/\/www.edwardtufte.com\/bboard\/q-and-a-fetch-msg?msg_id=0003wk\">ET forum announcement thread<\/a> for more on quilts. Some have ragged edges, others straight. Include these images just as you would any other <code>figure<\/code>.<\/p><p>This is an ImageQuilt surveying Chinese calligraphy, placed in a full-width figure to accomodate its girth:<\/p><figure class=\"fullwidth\"><img src=\"https:\/\/edwardtufte.github.io\/tufte-css\/img\/imagequilt-chinese-calligraphy.png\" alt=\"Image of Chinese Calligraphy\" referrerpolicy=\"no-referrer\" \/><\/figure><p>Here is an ImageQuilt of 47 animal sounds over and over, in a figure constrained to the main text region. This quilt has ragged edges, but the image itself is of course still rectangular.<\/p><figure><img src=\"https:\/\/edwardtufte.github.io\/tufte-css\/img\/imagequilt-animal-sounds.png\" alt=\"Image of animal sounds\" referrerpolicy=\"no-referrer\" \/><\/figure><\/section><section><h2 id=\"epilogue\">Epilogue<\/h2><p>Many thanks go to Edward Tufte for leading the way with his work. It is only through his kind and careful editing that this project accomplishes what it does. All errors of implementation are of course mine.<\/p><\/section>","created_at":"2019-04-09T18:30:42+0000","updated_at":"2020-03-09T21:23:28+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"edwardtufte.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":903,"title":"Eric Lawler - Servant-leader CTO","url":"https:\/\/lawler.io\/","content":"<div><header>\n                    <p>Written April 08, 2019<\/p>\n                        <p>Filed under \u201c<a href=\"https:\/\/lawler.io\/categories\/dross\/\" rel=\"tag\">Dross<\/a>\u201d<\/p>\n                <\/header><section><h2>Or, how to build an anti-Medium.com<\/h2><p>Two curiously-related pieces of content were shared on Hacker News today. The first was a short article by <a href=\"https:\/\/jrl.ninja\/etc\/1\">Joshua Li<\/a>, explaining how a few basic CSS styles can make for a great-looking website\u2014across varying devices. The second was a longer post by Marco Fioretti on how <a href=\"http:\/\/stop.zona-m.net\/2018\/01\/indeed-it-seems-that-google-is-forgetting-the-old-web\/\">Google is \u201cforgetting\u201d the old web<\/a>.<\/p><p>The latter post kicked up the usual fierce debates <label for=\"mn-seo\" class=\"margin-toggle\">\u2295<\/label>\u2026and people arguing about whether a search engine should actually return results matching what you type in. around whether the internet is still an open place if <em>one<\/em> gatekeeper determines whether you can find and read certain content.<\/p><p>The former article was berated for having too much CSS <label for=\"mn-ornot\" class=\"margin-toggle\">\u2295<\/label>\u2026or not enough CSS. Or not including styles for printers. or for using the CSS \u201crem\u201d unit, instead of ems or pixels or vws.<\/p><p>But what really drew my eye was a comment from <a href=\"https:\/\/news.ycombinator.com\/item?id=19610271\">Marcus Holmes<\/a>:\n<\/p><blockquote><p>Is it just me, then, that hates the \u201cnarrow strip of text down the middle of my large monitor\u201d school of web design?<\/p>\n<p>I don\u2019t understand why I\u2019m being forced to scroll when there\u2019s all this blank space to the sides.<\/p><footer>Marcus Holmes, comment on news.ycombinator.com<\/footer><\/blockquote><p>Great question, Marcus. <label for=\"mn-actually\" class=\"margin-toggle\">\u2295<\/label><em>Actually<\/em>, it\u2019s so I can use these sick side notes and margin notes! Thank you, Tufte CSS. Why <em>doesn\u2019t<\/em> this website fill 100% of your 1080p or 4K monitor\u2019s screen with text? There\u2019s a very specific reason, and that is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Line_length\"><em>text measure<\/em><\/a>.<\/p><blockquote><p>A block of text or paragraph has a maximum line length that fits a determined design. If the lines are too short then the text becomes disjointed; if they are too long the content loses rhythm as the reader searches for the start of each line.<\/p><footer>Wikipedia entry on \u201cMeasure (typography)\u201d, 2019<\/footer><\/blockquote><p>The <label for=\"mn-characters\" class=\"margin-toggle\">\u2295<\/label>In both typography- and computer-speak, a <em>character<\/em> is an individual letter or typographic mark. optimal width of a paragraph of text is anywhere from 50\u201380 characters per line, depending on which of the many studies you cite. If a paragraph is <em>too<\/em> wide, your eye loses its place when <a href=\"https:\/\/en.wikipedia.org\/wiki\/Eye_movement_in_reading#Saccades\">tracking back<\/a> to the beginning of the next line. This decreases reading comprehension\u2014and increases my irritation when I realize I\u2019m re-reading the same line. \ud83e\udd26<\/p><h3>So what\u2019s the point?<\/h3><p>For <label for=\"mn-firefox\" class=\"margin-toggle\">\u2295<\/label>Firefox offers a \u201cReader\u201d mode that will automatically constrain the page to the same ~68 character maximum width. Again, this is not a coincidence! this website, I chose a text measure that is approximately 70 characters wide, with average line widths of 65-68 characters. After your eye adjusts to the \u201cconstrained\u201d width for reading long articles, Wikipedia begins to tire you out.<\/p><p>It took several iterations to strike <a href=\"https:\/\/github.com\/Eiriksmal\/lawler-dot-io\/commit\/68f052e9bf8bbe9c835d3b6683d036971a1d6d82\">the right balance<\/a> in this design between too-wide and too-narrow. <label for=\"mn-duh\" class=\"margin-toggle\">\u2295<\/label>Given how many web typography dissertations I read in the course of refining this design, I shouldn\u2019t have been surprised the professionals use consistent text measures\u2014across publishers. My final measurement was to drag a bunch of books off the shelf, type out a line from each of them, and arrive at my own ideal width based on what the professionals chose. To my surprise, regardless of the book\u2019s physical page sizes and typeface size, all of them were <em>exactly<\/em> 66 characters wide.<\/p><\/section><section><h2>What\u2019s this have to do with the other article?<\/h2><p>The Fioretti article references a serious problem with the most-used search engine on the planet: It literally will not answer search results for sites that are \u201ctoo old.\u201d This could be a post written in 2007, or 2012. No one knows exactly how the Google algorithms rank and prune pages.<\/p><p>Lawler.io was inspired by <a href=\"https:\/\/www.gwern.net\/About\">Gwern<\/a>, a self-proclaimed \u201cfreelance writer &amp; researcher who lives in Virginia.\u201d He wrote something concerning his own website that runs deeply counter to the current online culture: Gwern.net is home to \u201cstable long-term essays which improve over time.\u201d<\/p><h3>Stable, long-term<\/h3><p>A key ingredient to Gwern\u2019s website is his decision to use incredibly robust computer-y bits behind the scenes, to ensure his website isn\u2019t going anywhere for a long, long time.<\/p><p>While I am not going to the same extremes as Gwern to keep lawler.io online for decades, it <em>is<\/em> very appealing to create a simple, easy-to-maintain website that can persist for years to come.<\/p><p>Much has been written about <a href=\"https:\/\/en.wikipedia.org\/wiki\/Link_rot\">link rot<\/a>, the phenomena that most of what goes on the internet disappear after a few weeks or months. Frankly, a lot of the content deserves to rot and disappear. The world has no need for another listicle of 13 Must-Know Viral Memes<label for=\"sn-believe\" class=\"margin-toggle sidenote-number\">You won\u2019t believe #7!.<\/label><\/p><p>The first challenge is putting content on that website that <em>deserves<\/em> to persist. The second, easier<label for=\"sn-easier\" class=\"margin-toggle sidenote-number\">Easier for this hacker-generalist CTO, at least. challenge, is keeping it online and living at the same web address.<\/label><\/p><h3>Kaizen: Continuous improvement over time<\/h3><p>Journalists are paid to write articles. Buzzfeed-esque <label for=\"sn-buzz\" class=\"margin-toggle sidenote-number\">I\u2019m going to use \u201cBuzzfeed\u201d as the Kleenex of \u201cshort web articles with outlandish titles and a splashy lead image to get people to click to open the article and share with all their friends but otherwise leave no impact on the \u2018reader\u2019 whatsoever.\u201d I understand Buzzfeed News has \u201creal\u201d journalists now and does \u201cserious\u201d work. This does not change the impact they had on online content. listicles come from \u201ccontent farms\u201d that pay writers to\u2026 write. Your writing must hit certain key metrics or your paychecks will dry up. The most important metric? How many people open up your article, of course!<\/label><\/p><p>This incentivizes the creation of short articles that draw lots of clicks<label for=\"sn-engage\" class=\"margin-toggle sidenote-number\">\u201cEngagement\u201d in social media marketing terms., but don\u2019t leave much of an impact on the reader.<\/label><\/p><p>For this kind of content, it almost makes sense that Google would stop surfacing Buzzfeed articles from 2012. It wasn\u2019t worth reading then and it <em>certainly<\/em> isn\u2019t worth reading now.<\/p><h3>Medium.com<\/h3><p>Medium.com exploded in popularity because it, too, used to respect the typography rules we discussed above. <label for=\"mn-now\" class=\"margin-toggle\">\u2295<\/label>Today, it\u2019s a bit too wide for my personal taste: ~77 characters It used simple black-ish text on a white-ish background. A straight-forward design that got out of your way and <em>let you read.<\/em> And read we did.<\/p><p>But Medium needed to make money. As more authors poured onto the site, as more companies chose to host their blogs on white-labeled Medium.com domains, Medium\u2019s design evolved. Nowadays, it\u2019s cluttered with crap. Here\u2019s a screenshot of a \u201cRecommended\u201d article from tonight:<\/p><figure class=\"fullwidth\"><img src=\"https:\/\/lawler.io\/images\/2019\/medium.com.png\" alt=\"Medium.com is a mess\" \/>Screenshot of an article linked to from Medium.com's homepage. (The thin typeface got squished when the image was resized)<\/figure><p>It\u2019s cluttered with a huge call to action to join the website, a bunch of nav on the top, and a social widget that follows you down the page during the entire read. Of course, when you hit the bottom, the clutter maximizes with <em>more<\/em> call-to-actions, obligatory comment boxes<label for=\"sn-engagement\" class=\"margin-toggle sidenote-number\">Never forget your engagement KPIs. and other content you might want to read.<\/label><\/p><p>I digress. The real thing that Medium.com has changed is expectations on what \u201clong\u201d content is. The pictured article has an impressive 5-minute reading estimate<label for=\"sn-squint\" class=\"margin-toggle sidenote-number\">Squint and you can see it under the author\u2019s name and link to follow her writing on Medium.. Most of the content I see on Medium are 2-3 minute reads. Any more than that and you <em>seriously<\/em> risk losing readers. <\/label><\/p><p>In fact, that\u2019s the second point in the quote I led off with. \u201cI don\u2019t understand why I\u2019m being forced to scroll.\u201d Though we can be assured Mr. Holmes is willing to scroll on a computer to read long articles, most of the people on the internet can\u2019t be bothered. Buzzfeed knows this. Medium.com knows this, too.<\/p><p>But is there another form of consumer? Is there a form of content on the internet that takes longer than 5 minutes to read? Is anyone writing longer articles that repulse the casual surfer, mindlessly flicking through links on their iPhone X? Apparently, such a fiend does exist! Quietly typing away, hundreds of individuals are crafting articles designed to be truly thought-provoking, to stand the test of time and Deserve To Be Read decades later.<\/p><p>Gwern\u2019s <label for=\"mn-metoo\" class=\"margin-toggle\">\u2295<\/label>I, too, quietly tidy up these <del>humble<\/del> scrivings, tightening them over time. Full edit histories, with explanatory commit messages, are <a href=\"https:\/\/github.com\/Eiriksmal\/lawler-dot-io\/tree\/main\/content\">in GitHub<\/a> for interested parties. taken this approach to the next level\u2014he continuously updates his posts as he learns new information on the topic. His experiments will have periodic updates for years. Years!<\/p><h3>Aspirations<\/h3><p>My aspiration is to create some substantive content that Deserves To Be Read. While this piece is certainly not one of them, it\u2019s a start. A way of painfully filing off the rust from my long-neglected long-form writing skills. A way of unblocking writers block and getting the juices flowing.<\/p><\/section><section><h2>The answer<\/h2><p>What do these two articles have to do with one another? Simple: My website was designed to fit into both categories. An aesthetically-pleasing, easy-to-read website with no tracking codes. A home for long-form content that repulses article skimmers. A place for writing that Google will certainly ignore in 5 years.<\/p><\/section><footer class=\"de-emphasize\"><p>Finalized at 10:48 PM.\nTidied up on April 09, 2019 @ 10:42 AM.                    <\/p><p>Tagged with\n                     and                     .                <\/p><\/footer><\/div><div><header>\n                    <p>Written April 02, 2019<\/p>\n                        <p>Filed under \u201c<a href=\"https:\/\/lawler.io\/categories\/business\/\" rel=\"tag\">Business<\/a>\u201d<\/p>\n                <\/header><section><h2>King of business books<\/h2><p><em>The Goal<\/em> remains one of the best business books. Facts are true. Their truth does not change with the passage of time. No matter how our hairstyles evolve and computers continue to shrink, the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Theory_of_constraints\" target=\"_blank\" rel=\"noopener\">theory of constraints<\/a> remains a powerful explanation for how work is accomplished. Very few business books have stood this same test of time.<\/p><p>30 years after his initial book was published, an anniversary edition was released containing praise from Kevin Behr, a co-author of <em>The Phoenix Project<\/em><label for=\"sn-phoenix\" class=\"margin-toggle sidenote-number\">Mr. Behr\u2019s book is an attempt to rewrite <em>The Goal<\/em> for computer-y types\u2014ditching the manufacturing setting for the dull vinyl flooring of a Fortune 1000 company\u2019s IT department.. <\/label><\/p><blockquote><p>Now more than ever, <em>The Goal<\/em> is my first recommendation for CIOs, CTOs and DevOps to better see What to Change, What to Change to and just How to Cause the Change. Thank you, Eli!<\/p><footer>Kevin Behr, co-author of <em>The Phoenix Project<\/em><\/footer><\/blockquote><p>Despite taking more words to describe the same concepts, <em>The Phoenix Project<\/em> is landing on numerous \u201cTop 10 Books Every New And Experienced Engineering Manager Should Read\u201d listicles written in the past year. I heartily recommend every CTO or CIO read the <em>original<\/em> dissertation on the Theory of Constraints before reading Mr. Behr\u2019s less-useful, devops-heavy application of ToC.<\/p><\/section><section><h2>Key takeaways<\/h2><p>While <del>stuck at jury duty<\/del> chosen to serve the lovely people of beautiful San Diego County, I took the following notes. <label for=\"mn-meta\" class=\"margin-toggle\">\u2295<\/label>Anything in quotes is a direct quote. Everything else is my synthesis.\n<\/p><div class=\"epigraph\"><blockquote><p>Productivity is meaningless unless you know what your goal is.<\/p><footer>Goldratt, 1984, p. 32<\/footer><\/blockquote><\/div><h3>Building the foundation (pages 32-59)<\/h3><p>The goal is to <strong>make money<\/strong>. Profit, ROI, and cash flow measure this. Efficiencies, good people, high tech solutions, selling quality products, capturing market share, etc.\u2014these merely <em>enable<\/em> a company to make money, but should never be confused with The Goal.<\/p><p>Net profit and a solid ROI is not enough to stay in business. Cash flow is vital\u2014are we bringing in enough money to meet our expenses each month?<\/p><p><strong>THE GOAL<\/strong> is <label for=\"mn-thegoal\" class=\"margin-toggle\">\u2295<\/label>All three must increase at the same time\u2014no cheating for short-term pops in share price. \u201cto make money by increasing net profit, while simultaneously increasing return on investment, while simultaneously increasing cash flow.\u201d (page 49)<\/p><h3>Defining the terms (pages 60-75)<\/h3><p>Measurements for daily operation of a manufacturing plant. <label for=\"mn-expl\" class=\"margin-toggle\">\u2295<\/label>Put another way, throughput is money coming in. Inventory is money circulating inside the system. Opex is money paid <em>out<\/em> to make that throughput money come <em>in<\/em>. These are a form of expressing \u201cThe\u201d Goal in terms a plant understands (Mr. Behr does something similar for IT work in <em>The Phoenix Project<\/em>). See if you can pick up the common element:<\/p><ul><li>Throughput: The rate at which the system generates <u>money<\/u> through sales.<\/li>\n<li>Inventory: All the <u>money<\/u> that the system has invested in purchasing things it intends to sell.<\/li>\n<li>Operational expense<label for=\"sn-opex\" class=\"margin-toggle sidenote-number\">Opex, in a non-consulting company, includes <em>all<\/em> employee time. Working? Idle? No matter what the employee is doing, it\u2019s still opex.: All the <u>money<\/u> the system spends in order to turn inventory into throughput.<\/label><\/li>\n<\/ul><p>To make money, the value of the product\u2014and the amount charged!\u2014has to be greater than the combination of the investment in inventory and the total operational expense per <em>sold<\/em> unit. Units waiting to sell do not make you money.<\/p><p>Even buildings and machines are inventory<label for=\"sn-invest\" class=\"margin-toggle sidenote-number\">When you see \u201cinvestment\u201d, think \u201cinventory.\u201d. You can depreciate part of them as opex, but whatever value is left can theoretically be sold as throughput. Making the machine\u2026inventory!<\/label><\/p><h3>Understanding flow (pages 80-101)<\/h3><p>Trimming production capacity to precisely meet marketing demand is doomed to failure.<\/p><ul><li>Dependent events and statistical fluctuations necessitate having a buffer on capacity.<\/li>\n<li>Keeping everyone \u201cbusy\u201d in a plant creates excessive inventory.<\/li>\n<\/ul><p>Statistical fluctuations are the possible range of time a task can take. If the <em>average<\/em> time to solder a wire takes 4.3 minutes, but the range is 2.1 -&gt; 6.4 minutes, no one can predict whether any individual wire will take 2.1 or 4.3 or even 7 minutes to complete<label for=\"sn-max\" class=\"margin-toggle sidenote-number\">Setting a new maximum time for soldering..<\/label><\/p><p>(Uh, sorry. This needs serious context, but is just for my personal recall.) Even with all the boys <em>walking<\/em> the same speed, even minor <em>pauses<\/em><label for=\"sn-pause\" class=\"margin-toggle sidenote-number\">This is critical. Do you see? You have a bunch of kids walking the same speed, but they\u2019re all randomly <em>pausing<\/em> the \u201cwork,\u201d causing the line to stretch out down the trail\u2026\ud83d\ude35 to adjust backpacks creates additional length in the line. But the length can\u2019t <em>decrease<\/em> because you can\u2019t walk in front of the kid you\u2019re following! Only the line leader has a pace that can be independently set.<\/label><\/p><ul><li>\u201cDependency limits the opportunities for higher fluctuations. That can improve throughput.\u201d (page 101)<\/li>\n<li>Throughput is the rate at which the <em>last<\/em> dependent event happens\u2014not the first.<\/li>\n<\/ul><h3>Taking it to the real world (pages 110+)<\/h3><div class=\"epigraph\"><blockquote><p>We always run, never stop; the other option, having some workers idle, is taboo.<\/p><footer>Goldratt, 1984, p. 113<\/footer><\/blockquote><\/div><p>Local maxima are useless. <label for=\"mn-morelater\" class=\"margin-toggle\">\u2295<\/label>(More to come as I finish digesting the final portion of the book) The <em>entire<\/em> system must be optimized. All processing centers <em>cannot<\/em> have the same capacity. Downstream centers need <em>additional<\/em> capacity to absorb the statistical fluctuations from upstream centers.<\/p><\/section><section><h2>Curious what to read next?<\/h2><p>Ken Cone, CTO of Radeus Labs <label for=\"mn-ken\" class=\"margin-toggle\">\u2295<\/label>CTO, co-owner, and lean manufacturing wizard!, encourages anyone interested in the ToC to turn to Dr. Goldratt\u2019s followup, <em>Critical Chain<\/em>. It provides a detailed explanation of the drum-buffer-rope methodology for controlling the flow of materials through bottlenecks.<\/p><p>If you like this kind of un-orthodox thinking, and are involved in software development, you should also read <em>Kanban: Successful Evolutionary Change for Your Technology Business<\/em> by David J. Anderson, inventor of the Kanban (capital K) agile software method. He takes a similarly counter-intuitive principle\u2014limiting how many simultaneous tasks development teams can work on\u2014and builds on it to create flexible solutions that scale from 2-man teams all the way up to Microsoft-levels of bureaucracy.<\/p><\/section><footer class=\"de-emphasize\"><p>Finalized at 10:33 PM.\nTidied up on April 04, 2019 @ 9:04 PM.                    <\/p><p>Tagged with\n                    ,                     ,                     ,                     , and                     .                <\/p><\/footer><\/div><div><header>\n                    <p>Written March 20, 2019<\/p>\n                        <p>Filed under \u201c<a href=\"https:\/\/lawler.io\/categories\/dross\/\" rel=\"tag\">Dross<\/a>\u201d<\/p>\n                <\/header><section><h2>The Beginning<\/h2><p>When learning computer programming, it\u2019s traditional to begin one\u2019s journey towards mastering the syntax by instructing the computer to print <a href=\"https:\/\/en.wikipedia.org\/wiki\/%22Hello,_World!%22_program#History\">Hello, World!<\/a> on the screen.<\/p><p>This is Python\u2019s <em>Hello, World!<\/em>:<\/p><div class=\"highlight\"><pre>#!\/usr\/bin\/env python\nprint(\"Hello, World!\")\n<\/pre><\/div><p>And this is PHP\u2019s:<\/p><div class=\"highlight\"><pre>&lt;?php\necho \"Hello, World!\";\n?&gt;\n<\/pre><\/div><p>My <a href=\"https:\/\/github.com\/Eiriksmal\/gameboy\"><em>Hello, World!<\/em> for the GameBoy<\/a>, written in Z80 assembly, is an impressive 179 lines long\u2014a sharp contrast to Python and PHP.<\/p><\/section><section><h2>What\u2019s the point?<\/h2><p>This brief missive is lawler.io\u2019s <em>Hello, World!<\/em>, if you will. In the weeks to come, this lonely piece of content will be joined by more meaningful thoughts.<\/p><blockquote><p>Productivity is meaningless unless you know what your goal is.<\/p>\n<footer>Eliahu M. Goldratt, <cite>The Goal<\/cite> (1984), pg 32<\/footer><\/blockquote><p>My goals for this website are straightforward:<\/p><ol><li>Allow me to keep my writing skills sharp, lest they rust in the withering deluge of texts, emails, and social media lols.<\/li>\n<li>Give me a space to better understand my own thoughts by typing them out.<\/li>\n<li>Serve as a resource for others. When someone asks me for help, I can write a thoughtful response and throw it into the tangled snake of wires that is the world wide web\u2014maybe some brave soul will stumble across the response and lessen their own pain.<\/li>\n<\/ol><p>Here\u2019s to a great 2019.<\/p><\/section><footer class=\"de-emphasize\"><p>Finalized at 1:11 AM.\n                    <\/p><p>Tagged with\n                    .                <\/p><\/footer><\/div>","created_at":"2019-04-09T18:29:25+0000","updated_at":"2019-04-09T18:29:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"lawler.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":902,"title":"Web Design in 4 minutes","url":"https:\/\/jgthms.com\/web-design-in-4-minutes\/","content":"<section id=\"start\"><p>Let's say you have a product, a portfolio, or just an idea you want to share with everyone on your <em>own<\/em> website. Before you publish it on the internet, you want to make it look attractive, professional, or at least <em>decent<\/em> to look at.<\/p><p>What is the <a class=\"step\" data-step=\"0\" href=\"#content\">first thing<\/a> you need to work on?<\/p><\/section><section id=\"content\"><h2>Content<\/h2><p>The purpose of <strong>design<\/strong> is to enhance the presentation of the content it's applied to. It might sound obvious, but content being the <strong>primary<\/strong> element of a website, it should not be established as an afterthought.<\/p><p>Written content, like the paragraph you're currently reading, makes up for more than 90% of the Web. Styling this textual content will go a long way.<\/p><p>Let's assume you've already finalised the content you want to publish and just created an empty <code>style.css<\/code> file, what is the <a class=\"step\" data-step=\"1\" href=\"#centering\">first rule<\/a> you can write?<\/p><\/section><section id=\"centering\"><h2>Centering<\/h2><p>Long lines of text can be hard to parse, and thus hard to <strong>read<\/strong>. Setting a limit of characters per line greatly enhances the readability and appeal of a wall of text.<\/p><pre>body {\n  margin: 0 auto;\n  max-width: 50em;\n}<\/pre><p>After styling the text <em>blocks<\/em>, what about styling the <a class=\"step\" data-step=\"2\" href=\"#font-family\">text itself<\/a>?<\/p><\/section><section id=\"font-family\"><h2>Font family<\/h2><p>The browser's font defaults to <code>\"Times\"<\/code>, which can look unappealing (mostly because it is the \"unstyled\" font). Switching to a <strong>sans-serif<\/strong> font like <code>\"Helvetica\"<\/code> or <code>\"Arial\"<\/code> can vastly improve the look of your page.<\/p><pre>body {\n  font-family: \"Helvetica\", \"Arial\", sans-serif;\n}<\/pre><p><em>If you want to stick with a serif font, try <code>\"Georgia\"<\/code>.<\/em><\/p><p>While this makes the text more <em>appealing<\/em>, let's also make it <a class=\"step\" data-step=\"3\" href=\"#spacing\">more readable<\/a>.<\/p><\/section><section id=\"spacing\"><h2>Spacing<\/h2><p>When a page looks \"broken\" to a user, it's usually a <strong>spacing<\/strong> issue. Providing space both <em>around<\/em> and <em>within<\/em> your content can increase the appeal of your page.<\/p><pre>body {\n  line-height: 1.5;\n  padding: 4em 1em;\n}\nh2 {\n  margin-top: 1em;\n  padding-top: 1em;\n}<\/pre><p>While the layout has greatly improved so far, let's apply more <a class=\"step\" data-step=\"4\" href=\"#color-contrast\">subtle changes<\/a>.<\/p><\/section><section id=\"color-contrast\"><h2>Color &amp; contrast<\/h2><p>Black text on a white background can be harsh on the eyes. Opting for a softer shade of black for body text makes the page more <strong>comfortable<\/strong> to read.<\/p><pre>body {\n  color: #555;\n}<\/pre><p>And in order to keep a decent level of <strong>contrast<\/strong>, let's pick a darker shade for <strong>important<\/strong> words<\/p><pre>h1,\nh2,\nstrong {\n  color: #333;\n}<\/pre><p>While most of the page has been improved visually, some elements (like the code snippets) still seem <a class=\"step\" data-step=\"5\" href=\"#balance\">out of place<\/a>.<\/p><\/section><section id=\"balance\"><h2>Balance<\/h2><p>It only takes a few additional touches to correct the balance of the page:<\/p><pre>code,\npre {\n  background: #eee;\n}\ncode {\n  padding: 2px 4px;\n  vertical-align: text-bottom;\n}\npre {\n  padding: 1em;\n}<\/pre><p>At this point, you might want to make your page stand out and give it <a class=\"step\" data-step=\"6\" href=\"#primary-color\">identity<\/a>.<\/p><\/section><section id=\"primary-color\"><h2>Primary color<\/h2><p>Most brands have a <strong>primary color<\/strong> that acts as a visual accent. On a website, this accent can be used to provide emphasis on interactive elements, like <strong>links<\/strong>.<\/p><pre>a {\n  color: #e81c4f;\n}<\/pre><p>But to keep the balance, we will need some <a class=\"step\" data-step=\"7\" href=\"#secondary-colors\">additional colors<\/a>.<\/p><\/section><section id=\"secondary-colors\"><h2>Secondary colors<\/h2><p>The accent color can be complemented with more <em>subtle<\/em> shades, to be used on borders, backgrounds, or even the body text.<\/p><pre>body {\n  color: #566b78;\n}\ncode,\npre {\n  background: #f5f7f9;\n  border-bottom: 1px solid #d8dee9;\n  color: #a7adba;\n}\npre {\n  border-left: 2px solid #69c;\n}<\/pre><p>Having changed the shades, why not change the <a class=\"step\" data-step=\"8\" href=\"#custom-font\">shapes<\/a>...<\/p><\/section><section id=\"custom-font\"><h2>Custom font<\/h2><p>Since text is the main content of a webpage, using a <strong>custom font<\/strong> gives the page even more noticeable identity.<\/p><p>While you can embed your own webfont or use an online service like <a href=\"https:\/\/typekit.com\" target=\"_blank\">Typekit<\/a>, let's use <code>\"Roboto\"<\/code> from the free <a href=\"https:\/\/fonts.google.com\" target=\"_blank\">Google Fonts<\/a> service:<\/p><pre>@import 'https:\/\/fonts.googleapis.com\/css?family=Roboto:300,400,500';\nbody {\n  font-family: \"Roboto\", \"Helvetica\", \"Arial\", sans-serif;\n}<\/pre><p>After enhancing your identity through <em>text<\/em>, how about adding a <a class=\"step\" data-step=\"9\" href=\"#images\">thousand words<\/a>...<\/p><\/section><section id=\"images\"><h2>\n          <img src=\"https:\/\/jgthms.com\/web-design-in-4-minutes\/images.png\" alt=\"Spongebob rainbow meme saying 'Images'\" \/><\/h2><p>Graphics and icons can be used either as ornaments to support your content, or actively take part in the message you want to convey.<\/p><p>Let's enhance our header with a nice <strong>background image<\/strong> from <a href=\"https:\/\/unsplash.com\/photos\/qH36EgNjPJY\" target=\"_blank\">Unsplash<\/a><\/p><pre>header {\n  background-color: #263d36;\n  background-image: url(\"header.jpg\");\n  background-position: center top;\n  background-repeat: no-repeat;\n  background-size: cover;\n  line-height: 1.2;\n  padding: 10vw 2em;\n  text-align: center;\n}<\/pre><p>Let's also add a <strong>logo<\/strong><\/p><pre>header img {\n  display: inline-block;\n  height: 120px;\n  vertical-align: top;\n  width: 120px;\n}<\/pre><p>Let's take that opportunity to enhance the text styles.<\/p><pre>header h1 {\n  color: white;\n  font-size: 2.5em;\n  font-weight: 300;\n}\nheader a {\n  border: 1px solid #e81c4f;\n  border-radius: 290486px;\n  color: white;\n  font-size: 0.6em;\n  letter-spacing: 0.2em;\n  padding: 1em 2em;\n  text-transform: uppercase;\n  text-decoration: none;\n  transition: none 200ms ease-out;\n  transition-property: color, background;\n}\nheader a:hover {\n  background:  #e81c4f;\n  color: white;\n}<\/pre><p><a href=\"#header\">And voil\u00e0<\/a>!<\/p><p>We've designed a decent page in just a few minutes, following basic principles of web design. There's only <a class=\"step\" data-step=\"10\" href=\"#share\">one last thing<\/a> left to do...<\/p><\/section><footer><a class=\"button\" href=\"https:\/\/jgthms.com\" target=\"_blank\">Made by <strong>@jgthms<\/strong><\/a>\n      <\/footer>","created_at":"2019-04-08T18:15:46+0000","updated_at":"2019-04-08T18:15:46+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"jgthms.com","preview_picture":"https:\/\/jgthms.com\/web-design-in-4-minutes\/learn-web-design-in-4-minutes.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":901,"title":"Behind the scenes on retrofitting DCSS\u2019 seeding \u00ab Dungeon Crawl Stone Soup","url":"http:\/\/crawl.develz.org\/wordpress\/behind-the-scenes-on-retrofitting-dcss-seeding","content":"<div class=\"post-date\">\nSunday, March 31st, 2019 21:57<\/div>\n\n<div class=\"post-4467 post type-post status-publish format-standard hentry category-development-2\" id=\"post-4467\">\n<div class=\"post-title\">\n<h1><a href=\"http:\/\/crawl.develz.org\/wordpress\/behind-the-scenes-on-retrofitting-dcss-seeding\" rel=\"bookmark\" title=\"Behind the scenes on retrofitting DCSS\u2019 seeding\">Behind the scenes on retrofitting DCSS\u2019 seeding<\/a><\/h1>\n<\/div>\n\n<div class=\"post-meta\">\nby <\/div>\n\n<div class=\"post-content\">\n<h3>What is seeding?<\/h3>\n<p>The 0.23 release introduced a major and ongoing overhaul in how DCSS approaches random seeding from the player\u2019s perspective. If you\u2019re not familiar with this idea from other games with random procedural generation like Minecraft, Spelunky and brogue, here\u2019s how it\u2019s supposed to work: if you enter a specific code (the \u201cseed\u201d) into the game when starting it, you\u2019ll get a repeatable form of randomness that encompasses dungeon layout, items, and monsters. Any dungeon generated with that seed should be identical. This has many applications: from Spelunky or Brogue (and many others) you might be familiar with the concept of a daily, weekly, or monthly run. In these, all players enter the same seed and over some time interval explore the same dungeon. Here are a few ways crawl community has been experimenting with this since the release:<\/p>\n<ul><li>On <a href=\"https:\/\/crawl.project357.org\/\">the CPO server<\/a>, run by chequers, there is an experimental weekly online challenge with a share dungeon.<\/li>\n<li>Ultraviolent4 has started running an offline <a href=\"https:\/\/www.reddit.com\/r\/dcssgotm\/\">Game of the Month<\/a> challenge. (Edit: The next challenge begins on Apr 7, <a href=\"https:\/\/www.reddit.com\/r\/dcss\/comments\/b7wtfy\/behind_the_scenes_on_retrofitting_dcss_seeding\/ejupd5i\/\">per uv4<\/a>.)<\/li>\n<\/ul><p>I\u2019m excited to see what else develops. But what does it take for this to actually work? This post goes into the background on what has been involved in doing this retrofitting for DCSS, where it is at, and what the remaining problems are.<\/p>\n<h3>Randomness and procedural generation<\/h3>\n<p>Procedural level generation happens via constrained randomness. Some process repeatedly draws on a random number generator according to a fixed, predefined set of constraints and rules, making choices that (ideally) generate a coherent, explorable, locale. At the heart of this is a random number generator, which usually takes the form of a function that the programmer can repeatedly call, generating a sequence of numbers.<\/p>\n<p>Now, \u201ctrue randomness\u201d is more of an ideal than anything that can be implemented, so RNGs are really only \u201cpseudo-random\u201d. What this means is that they aim to (in addition to matching some overall distribution) be <em>unpredictable<\/em>, which a viewer or adversary will perceive as random. <a title=\"Blum, Blum and Shub 1986 - A simple unpredictable pseudo-random number generator\" href=\"https:\/\/pdfs.semanticscholar.org\/c19b\/91cdc1da67c52e606cd4752472ce0db83131.pdf\" target=\"_blank\">Unpredictability<\/a> can be defined in precise mathematical terms that I won\u2019t go into here, but it doesn\u2019t actually mean complicated: some of the best rng algorithms are remarkably simple. They basically all work by initializing the rng state with a \u201cseed\u201d, and repeatedly transforming that state, and the heart of the one we use, PCG, is just <a title=\"pcg implementations\" href=\"http:\/\/www.pcg-random.org\/download.html\" target=\"_blank\">5 lines of code<\/a>. For cases where there <em>is<\/em> an adversary, you want this initial seed to itself be unpredictable, and so common strategies are to use some external state of the program (e.g. exact time, <a title=\"NIST randomness beacon\" href=\"https:\/\/www.nist.gov\/programs-projects\/nist-randomness-beacon\" target=\"_blank\">external devices<\/a>) to do this initialization: then, if the adversary doesn\u2019t know the initial state, and the algorithm is unpredictable in the right way, they\u2019re out of luck.<\/p>\n<p>Video games flip this idea on its head. For some kinds of RNGs, if you initialize the generator in the same way, multiple instances will generate the same sequence of numbers in the same order, <em>while still behaving unpredictably and well-distributed<\/em> as long as the seed is unknown. This means that if the rules and constraints that draw on the RNG are implemented correctly (we\u2019ll come back to this), the behavior given that seed acts random but is completely replicable. You can think of the RNG as, given a seed, providing a \u201ccode book\u201d of integers that acts random, but as long as you have the code book and the correct rules\/constraints (giving you position in the list), you can decode the behavior of the game.<\/p>\n<p>There are, then, two main ways to make this go wrong: mess up the initial conditions, or cause the rng to get <em>out of sync<\/em>. This latter case is when two processes are using the same sequence, but make decisions starting from different points in the sequence. Because of unpredicatability, their behavior will diverge extremely rapidly (more on this later). In general, if RNG behavior diverges at all, it tends to diverge <a title=\"butterfly effect\" href=\"https:\/\/en.wikipedia.org\/wiki\/Butterfly_effect\" target=\"_blank\">thoroughly and quickly<\/a>.<\/p>\n<p>DCSS has long used <a title=\"PCG RNG\" href=\"http:\/\/www.pcg-random.org\/\" target=\"_blank\">the PCG RNG<\/a> which (like the Mersenne Twister, xorshift, and many others) supports this reproducibility; retired dev bh <a title=\"PCG in crawl\" href=\"https:\/\/github.com\/crawl\/crawl\/commit\/415d666aa2f0\" target=\"_blank\">added this in 2015<\/a>, replacing a hand-rolled implementation of an algorithm from the 80s that was more predictable. However, using a good RNG algorithm is only one piece of the picture, and turns out to be probably the least complicated part of this project.<\/p>\n<h3>Why player-facing seeding was (and is) hard for DCSS<\/h3>\n<p>Crawl wasn\u2019t remotely designed with seeding as a player-facing feature in mind; though many roguelikes have used specifiable seeds for testing, the mainstream use as a player feature seems to be quite recent. Actually, I couldn\u2019t find out much of the history of using seeding for <em>completely<\/em> repeatable procedural generation as a player-oriented feature, but it doesn\u2019t seem to have been very mainstream before games like Minecraft and Spelunky; an earlier example, pointed out to me by Roguelikes discord user esc, is Civ 3. Seed-based pseudo-randomness, internal to the game, is much older, and for example Moria allowed choosing a seed for testing purposes. But getting seeding to work as a <em>player<\/em> feature is a whole different project.<\/p>\n<p>Here are four challenges that DCSS faced (and still faces) when trying to get player-facing seeding to work well:<\/p>\n<ol><li>Mixing of gameplay \/ levelgen RNG<\/li>\n<li>Arbitrary interactions in level generation between levels.<\/li>\n<li>Extremely complex level generation with a huge diversity of procedural generation techniques, implemented in a mix of C++ and Lua code.<\/li>\n<li>Diversity of build targets and settings: support for windows, mac, linux, with 32 or 64 bit builds. Builds on non-x86 architectures such as ARM.<\/li>\n<\/ol><p>When starting the project, I was aware of factors 3-4, but thought 1-2 would be the biggest challenges, and so it was there that I began. In retrospect, 4 is actually the biggest problem.<\/p>\n<p>Before 0.23, generating a level and entering a level were completely intertwined: levels were not generated until you entered them. \u201cGreedy\u201d level generation goes back to Linley\u2019s Dungeon Crawl as far as I\u2019m aware, and on its own terms is a fairly simple and clean approach, spreading the CPU cost of level generation across the game in a way that hides it from the player. However, it can lead to a problem illustrated by the following toy example: suppose that you have a separate RNG for each level, the player is on level 2, and for both levels 3 and 4 Sigmund is poised to generate given the RNG state. If the player takes stairs to level 3, then he\u2019ll generate there; if the player takes a shaft to level 4, then he\u2019ll generate there. Since Sigmund is a unique monster, he cannot then regenerate if the player enters levels 4 and 3 respectively, and so the RNG diverges right away \u2014 the builder may place another unique, or do something different altogether, and after that initial decision will tend to be completely different.<\/p>\n<p>Not only that, this first assumption wasn\u2019t made: DCSS before 0.23 used a single RNG for gameplay and levelgen decisions, so any different action the player took might lead to completely different level generation.<\/p>\n<h3>Part 1: pregeneration and RNG separation<\/h3>\n<p>This problem has long been known, and <a title=\"tavern discussion of seeding\" href=\"https:\/\/crawl.develz.org\/tavern\/viewtopic.php?p=302759\" target=\"_blank\">comes up regularly<\/a> in discussions of seeding. I adopted two solutions from these discussions: separate out levelgen RNG from gameplay, and allow pregeneration of the dungeon. The latter solves almost all of problems 1-2 on its own, and had several past implementations by community members. The most developed of these was <a title=\"bel's seeded crawl\" href=\"https:\/\/github.com\/belrogue\/crawl\/compare\/master...belrogue:seeded_crawl\" target=\"_blank\">due to tavern user bel<\/a> in the above linked thread, and just needed some adjustments so that level generation was better separated from level entry. The initial seeding patch can be seen <a title=\"initial seed PR\" href=\"https:\/\/github.com\/crawl\/crawl\/pull\/968\" target=\"_blank\">here<\/a>.<\/p>\n<p>Incidentally, if you\u2019re designing a game from scratch, your life is going to be easier if you take a different strategy for these sorts of non-independent decisions. For example, you could place uniques independently of levelgen (something we do now only for one unique, Boris, who respawns), perhaps deciding where they go at the beginning of the game. However, it\u2019s probably too late for DCSS in this respect.<\/p>\n<h3>Part 2: testing and fixing many, many seed stability bugs<\/h3>\n<p>This patch, it turns out, was only the beginning, opening things up to a whole world of ways in which a repeatable RNG can get out of sync based on problems 3-4.<\/p>\n<p>To start with, I wrote a lua test that, given a seed, prints the vault list for that seed, re-runs generation, and compares the result. While true seed stability bugs on a single device have been fairly rare, this second part helped identify a tremendous amount of <a title=\"code to reset global state\" href=\"https:\/\/github.com\/crawl\/crawl\/blob\/f2055891f03d71e7da4261c832e1d593d7ba4fe8\/crawl-ref\/source\/dungeon.cc#L589\" target=\"_blank\">global state that can matter for seeding<\/a>, scattered all over crawlcode. You can see an example of the output <a href=\"https:\/\/travis-ci.org\/rawlins\/crawl\/jobs\/501737499\">here<\/a> if you scroll down to \u201cRunning test #31: \u2018vault_catalog.lua\u2019.\u201d<\/p>\n<p>The real challenge has been seed stability across devices. Here is a sampling of problems that have come up:<\/p>\n<ul><li>If you\u2019re drawing randomly from a list, and something permutes the underlying list, it doesn\u2019t matter if your rng draw is the same! For example, if the OS provides a directory listing in a different order.<\/li>\n<li>If the order of some calls isn\u2019t determined by the language\/library standard, then it will probably be implementation dependent. This is a problem that has come up again and again:<\/li>\n<li>C++ (and C) doesn\u2019t define the order of multiple function calls that make up an arithmetic expression, and in practice their order varies. You might think this would go by order of operations, but this would be wrong. They all have to be <a title=\"sequence random calls\" href=\"https:\/\/github.com\/crawl\/crawl\/commit\/cdddf7c89e99\" target=\"_blank\">explicitly sequenced<\/a>.<\/li>\n<li>Lua iteration order, both through the C interface, and via the \u201cpairs\u201d function, is not <a title=\"lua iteration fix 1\" href=\"https:\/\/github.com\/crawl\/crawl\/commit\/18b65154750a\" target=\"_blank\">guaranteed<\/a> to be <a title=\"Lua iteration fix 2\" href=\"https:\/\/github.com\/crawl\/crawl\/commit\/05a21ba7acd5\" target=\"_blank\">stable<\/a>.<\/li>\n<li>Anything in C++ that sorts may have implementation-specific behavior if a comparison operation indicates two objects are equal. This is fine if your comparison operation is fine-grained enough, but <a title=\"comparison fix\" href=\"https:\/\/github.com\/crawl\/crawl\/commit\/d121a6d4c08e\" target=\"_blank\">sometimes it wasn\u2019t<\/a>.<\/li>\n<\/ul>A large set of bugs that were introduced with the initial pregeneration patch because the level builder (and related code) assumed the player was about to enter the level; something that could be designed around from scratch but was fairly baked into dcss as it stood.\nDCSS caches parsed maps, and some cases of randomness in the cache were causing divergence.\nCases where the builder made use of player state: the most prominent of these were <a href=\"https:\/\/github.com\/crawl\/crawl\/commit\/60e86920ebe8\">vaults that generated items<\/a> using the acquirement code. (This still can happen in troves, but only there.)\nFloating point issues. First, Lua 5.1 uses different Number to int conversion techniques on different build targets, and these have different truncation behavior. Then, even controlling for this it turns out that e.g. SSE2 behavior for non-exactly-represented integers can lead to different ints than x87, for example, (1 \/ 100 * .7). I doubt all of these have been found yet, and for the 0.23.2 release I still needed to manually force the windows build to not use x87. I\u2019m looking at you layout_delve\u2026\nCases where unexpected code was drawn upon by the builder. For example, levelgen tries to ensure connectivity, and it turned out the code that does this was being <a title=\"explore options\" href=\"https:\/\/github.com\/crawl\/crawl\/commit\/56e1b8fd0ee2\" target=\"_blank\">affected by certain rc options<\/a>! Another example is that slime connectivity in particular uses a <em>different<\/em> algorithm (because of the wall damage) that led to an underspecification bug mentioned above.\n<p>In effectively all of these examples, the result of the bug was immediate loss of seed synchronization across devices \u2014 and many of these were things that wouldn\u2019t be considered bugs at all without the goal of seed stability. Debugging them was uniformly extremely difficult, because it required both replicating the divergence and then finding the exact point of RNG divergence. Without VMs and docker, I would not have been able to do most of this.<\/p>\n<p>Some of these problems can be programmed defensively against, with experience, but I doubt I\u2019ve found all the seed stability bugs in crawlcode. At this point, I\u2019m mostly reliant on people testing seeds on various platforms \u2014 Ultraviolent4\u2032s <a title=\"gotm\" href=\"https:\/\/www.reddit.com\/r\/dcssgotm\/comments\/ax2vg7\/dcss_game_of_the_month_1_announcement_thread\/\" target=\"_blank\">seeded Game of The Month<\/a> series has so far been invaluable for finding seed\/pregen bugs. (Incidentally, the challenge of localizing seed bugs, together with the challenges of developing for seed stability, are why I would strongly recommend fellow roguelike devs against using seed-based save formats\u2026)<\/p>\n<h3>Part 3: next steps<\/h3>\n<p>There are still many improvements to be made, and I expect a lot of my effort for 0.24 will go towards finding and quashing more seed stability bugs. Here are a few improvements I hope to include in the next version:<\/p>\n<ul><li>Incremental pregeneration: generate the dungeon in a predetermined order, but only do as much as you need to do to play catchup on level changes. This will allow hopefully removing the distinction between pregenerated and non-pregenerated games, along with allowing seeded play online. (Right now, the main barrier is that up-front level generation is very hard on a server\u2019s cpu.)<\/li>\n<li>More RNG improvements for repeatability: apply various techniques to make rng de-synchronization less impactful. Remove floating point code as much as possible, and improve some of the lua numeric issues further.<\/li>\n<li>Non-numeric (readable) seeds, a la <a href=\"https:\/\/www.gridsagegames.com\/blog\/2017\/05\/working-seeds\/\">cogmind<\/a>.<\/li>\n<li>A seed catalog of some kind.<\/li>\n<li>What seed-related features would you like to see?<\/li>\n<\/ul><p>The DCSS community has been extremely helpful in finding and reporting bugs, and I doubt we are done. Please go play one of the seeding based challenges, and let us know if you find problems!<\/p>\n<\/div>\n\n\n<\/div>\n\n<div class=\"post-nav\">\n<div class=\"newer\">\n\u00ab <a href=\"http:\/\/crawl.develz.org\/wordpress\/0-23-2-and-0-22-2-bugfix-releases\" rel=\"next\">0.23.2 and 0.22.2 bugfix releases<\/a>  \u00ab newer<\/div>\n<div class=\"older\">\nolder \u00bb  <a href=\"http:\/\/crawl.develz.org\/wordpress\/0-23-1-bugfix-release\" rel=\"prev\">0.23.1 Bugfix Release<\/a> \u00bb<\/div>\n<\/div>","created_at":"2019-04-07T19:40:36+0000","updated_at":"2019-04-07T19:40:36+0000","published_by":["advil"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":11,"domain_name":"crawl.develz.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":900,"title":"Citybound","url":"https:\/\/aeplay.org\/citybound","content":"Citybound<div class=\"w-layout-grid grid\"><div id=\"w-node-5749400518d6-bee9e583\" data-w-id=\"4844a770-8c4c-131d-d033-5749400518d6\" class=\"screencap\"><img src=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476.png\" srcset=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476-p-500.png 500w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476-p-800.png 800w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476-p-1080.png 1080w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476-p-1600.png 1600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476-p-2000.png 2000w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476-p-2600.png 2600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c24208c35af3ace29d8ff05_Screenshot%202018-12-26%20at%2014.39.29.476.png 3104w\" alt=\"\" class=\"screenshot\" \/><\/div><p id=\"w-node-356016d93eb7-bee9e583\" class=\"grid-p\">See and inspect all aspects of city life, across many scales of time and space.<br \/><\/p><p id=\"w-node-356016d93eb3-bee9e583\" class=\"grid-p\">Shape your city by sketching detailed project plans for infrastructure and zones.<\/p><p id=\"w-node-356016d93eb5-bee9e583\" class=\"grid-p\">Implement projects, decide on budgets and make government decisions.<\/p><p id=\"w-node-356016d93ebf-bee9e583\" class=\"mainmissionp\">Citybound has an ambitious vision, but is still actively developed as a largely one-man <a href=\"https:\/\/github.com\/citybound\/citybound\" target=\"_blank\">open-source effort<\/a>. Follow my creation process on the <a href=\"https:\/\/aeplay.org\/citybound-devblog\">DevBlog<\/a> and the <a href=\"https:\/\/reddit.com\/r\/Citybound\" target=\"_blank\">Citybound community on Reddit<\/a>. If you are brave, try out the <a href=\"https:\/\/aeplay.org\/citybound-livebuilds\">experimental live builds<\/a>. Citybound is proudly community-funded.<\/p><div id=\"w-node-356016d93eb0-bee9e583\"><div class=\"div-block-2\"><p>Support Citybound<\/p><p><a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=FZ9N9DKGE63RQ&amp;source=url\" target=\"_blank\" class=\"link-button patreon paypal w-button\">DONATE\u00a0ON\u00a0<\/a><\/p><\/div><\/div><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-971a9ac007e8-bee9e583\" class=\"grid-p\"><strong>In Citybound, every single household (a family or a business) is individually simulated.<\/strong> Each household has a home, and the daily activities of its family members or employees are driven by the resources the household needs and offers. Each household maintains a precise real-time inventory of its resources, which not only include tangible goods like groceries, raw materials, or money, but also intangible concepts like sleep, health, recreation, work-force, touristic interests or business services.<\/p><div id=\"w-node-9ed862ae1ce3-bee9e583\" data-w-id=\"ef65bb58-f5ff-e018-37d7-9ed862ae1ce3\" class=\"screencap\"><img src=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot.png\" srcset=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot-p-500.png 500w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot-p-800.png 800w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot-p-1080.png 1080w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot-p-1600.png 1600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot-p-2000.png 2000w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot-p-2600.png 2600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430cb35af3a5e2bd90d16_microeconomy%20screenshot.png 3104w\" alt=\"\" class=\"screenshot\" \/><\/div><p id=\"w-node-8d40f2487722-bee9e583\" class=\"grid-p\"><strong>Economic patterns arise directly from household interactions. <\/strong>To fulfil its human or business needs, each household tries to find attractive \"trade-partners\"\u00a0for each resource, factoring in price and quality differences, but also local reachability based on transport conditions. Households re-prioritise their needs throughout the day, leading to natural patterns of commute, shopping or logistical transport trips.<\/p><p id=\"w-node-909667354bf1-bee9e583\" class=\"grid-p\"><strong>Future work: <\/strong>Pre-simulating local trade conditions to determine attractiveness of real estate, thus influencing immigration, development and city growth. Scaling the system to high-density buildings and large populations in the millions.<\/p><div id=\"w-node-858d2b1c08ed-bee9e583\" class=\"references\"><p data-w-id=\"b5964bbb-20aa-37fa-4e4a-8ece7ab47a80\" class=\"grid-p-references-button\"><strong>Details &amp;\u00a0Inspiration \u25be<\/strong><\/p><div class=\"grid-p-references-inner w-richtext\"><ul><li>The original inspiration for simulating cities at the level of individuals and their interactions (and starting Citybound in the first place)\u00a0came from <a href=\"https:\/\/en.wikipedia.org\/wiki\/SimCity_(2013_video_game)\">SimCity (2013)<\/a>, which, however, only implemented a very superficial version of that (modelling people more like fluids with sources and drains)<\/li><\/ul><p>Papers:<\/p><ul><li>Kim, D. (2012). <em>Modelling Urban Growth\u202f: Towards an Agent Based Microeconomic Approach to Urban Dynamics and Spatial Policy Simulation<\/em><\/li><li>Koomen, E., Diogo, V., Dekkers, J., &amp; Rietveld, P. (2015). <em>A utility-based suitability framework for integrated local-scale land-use modelling.<\/em><\/li><li>Waddell, P., Borning, A., Noth, M., Freier, N., Becke, M., &amp; Ulfarsson, G. F. (2003). <em>Microsimulation of Urban Development and Location Choices: Design and Implementation of UrbanSim.<\/em> <\/li><li>Park, I. K., &amp; von Rabenau, B. (2011). <em>Disentangling agglomeration economies: Agents, sources, and spatial dependence.<\/em><\/li><li>Tiglao, N. C. C. (2005). <em>Modeling Households and Location Choices in Metro Manila.<\/em><\/li><li>Weidner, T., Moeckel, R., &amp; Brinckerhoff, P. (2011). <em>SILO: A Land-\u00adUse Model For Integrated Modeling<\/em>.<\/li><li>Beckmann, K. J., Br\u00fcggemann, U., Gr\u00e4fe, J., Huber, F., Meiners, H., Mieth, P., \u2026 Wegener, M. (2007). <em>ILUMASS Integrated Land-Use Modelling and Transportation System Simulation.<\/em><\/li><\/ul><\/div><\/div><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-94732a0999c0-bee9e583\" class=\"grid-p\">The efficient flow of traffic through transport infrastructure is vital for a city's health. Traffic is spontaneously determined by the micro-economic goals of its participants - and chaotic events are at the heart of traffic. That's why <strong>in Citybound every single trip of every person is simulated microscopically.<\/strong> Hundreds of thousands of cars physically move along roads and have to break, accelerate and change lanes in traffic to safely get to their destination.<\/p><p id=\"w-node-94732a0999c5-bee9e583\" class=\"grid-p\"><strong>Future work: <\/strong>Other modes of transport (pedestrians, light &amp;\u00a0heavy rail, airports, etc.). Multi-modal pathfinding (combining walking, public transport, taxis and driving to reach destinations).<\/p><div id=\"w-node-89ec280bd438-bee9e583\" data-w-id=\"ca7450a4-638c-6699-c195-89ec280bd438\" class=\"screencap\"><img src=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot.png\" srcset=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot-p-500.png 500w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot-p-800.png 800w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot-p-1080.png 1080w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot-p-1600.png 1600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot-p-2000.png 2000w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot-p-2600.png 2600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2436713c78e546ad8c1eff_traffic%20screenshot.png 3104w\" alt=\"\" class=\"screenshot\" \/><\/div><div id=\"w-node-c29aee41d8c9-bee9e583\" class=\"references\"><p data-w-id=\"cf7ad52f-0af4-1def-859c-c29aee41d8ca\" class=\"grid-p-references-button\"><strong>Details &amp;\u00a0Inspiration \u25be<\/strong><\/p><div class=\"grid-p-references-inner w-richtext\"><ul><li>Microscopic traffic simulation was the first part of Citybound that I\u00a0implemented - my strategy was to show of that I\u00a0could make this very computation-intensive part of city simulation scale to city sizes of millions of inhabitants. With already early prototypes being able to simulate up to 400,000 cars in real-time, this would suffice for cities of population ~4 million.<\/li><li>Because the road network is subject to frequent change in Citybound, many static pathfinding methods that require long pre-computation but offer fast run-time lookup are unsuitable. Pathfinding information in Citybound is stored in the road network itself and is dynamically updated and propagated through the network, similar to internet routing tables.<\/li><\/ul><p>Papers:<\/p><ul><li><em>Blatnig, S. (2008). Microscopic Traffic Simulation with Intelligent Agents.<\/em><\/li><li>Liberti, L. (2014). <em>14th Workshop on Algorithmic Approaches for Transportation Modelling, Optimization, and Systems.<\/em><\/li><li><em>Sturtevant, N., &amp; Buro, M. (2004). Partial Pathfinding Using Map Abstraction and Refinement.<\/em><\/li><\/ul><\/div><\/div><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-192f8534ea7d-bee9e583\" class=\"grid-p\"><strong>The core principle of interacting with the game world in Citybound is through planning. <\/strong>This gives you the ability to preview effects of even large-scale projects in your city, to undo and redo and to combine different ideas. Smart paint-gestures become beautifully curved roads and perfectly shaped zones, offering the power and expressiveness of professional CAD tools while being much more intuitive and fun to use.<\/p><div id=\"w-node-84cb352641ac-bee9e583\" data-w-id=\"c01f4aee-36c7-c0d7-529d-84cb352641ac\" class=\"screencap\"><img src=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot.png\" srcset=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot-p-500.png 500w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot-p-800.png 800w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot-p-1080.png 1080w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot-p-1600.png 1600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot-p-2000.png 2000w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot-p-2600.png 2600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2430404fdbba0859bdaada_planning%20screenshot.png 3104w\" alt=\"\" class=\"screenshot\" \/><\/div><p id=\"w-node-cfaf9a838e42-bee9e583\" class=\"grid-p\">Precise management of plan updates not only enables smooth interactive editing of even huge cities, but also makes <strong>collaborative multiplayer planning <\/strong>possible, where several people can play in and work on the same city together in real time.<\/p><p id=\"w-node-c3ac6f065de2-bee9e583\" class=\"grid-p\"><strong>Future work: <\/strong>Planning with 3D\u00a0features like bridges and tunnels. Preview of financial and political impact of planned projects.<\/p><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-306a82458648-bee9e583\" class=\"grid-p\">To realise Citybound's vision of simulating large-scale cities and streaming them to browsers for interaction, I\u00a0had to invent and implement quite an unusual computation architecture. Actors and message passing are the key concept in <a href=\"https:\/\/github.com\/aeickhoff\/kay\" target=\"_blank\">Kay<\/a>, my actor-system framework that elevates Citybound's simulation logic from its transparently networked and easily scalable underpinnings. Implemented in Rust, it allows type-safe and high-performance actor updates and inter-actor communication, with low-level optimisations for cache-locality while offering high-level features like distributed dynamic dispatch and broadcast messages.<\/p><div id=\"w-node-15db3e5b7a1d-bee9e583\" class=\"references\"><p data-w-id=\"14c106ee-63fc-2b26-91c2-15db3e5b7a1e\" class=\"grid-p-references-button\"><strong>Details &amp;\u00a0Inspiration \u25be<\/strong><\/p><div class=\"grid-p-references-inner w-richtext\"><ul><li>The main inspiration for my approach was my experience with the <a href=\"https:\/\/www.erlang.org\">Erlang<\/a> programming language for distributed systems.<\/li><li>Kay has relatively few features and abstractions in comparison, but that allows a focus on high-performance, with minimal actor and message footprints in memory.<\/li><li>The consolidated layout of actor state in memory and the mechanisms for handling broadcast messages that are received by up to millions of actors are heavily inspired by <a href=\"http:\/\/gamesfromwithin.com\/data-oriented-design\">Data Oriented Game Engine Design<\/a>, the core ingredient in the highly successful pattern of Entity-Component Systems.<\/li><\/ul><\/div><\/div><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-41d9364cc348-bee9e583\" class=\"grid-p\">Compiling Kay to WebAssembly allows running a fully fledged actor system node in the browser, which can communicate with the simulation backend through type-safe high-performance messages and do computation-intensive preparation work for rendering, before efficiently serialising and passing rendering-relevant information as JavaScript objects to a React app that implements the rest of the UI. The React app can call Rust hooks that were exported to JavaScript, to send Kay messages back to the simulation backend for influencing the game world based on player input.<\/p><img src=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2454af35af3a0f6cd927dd_architecture.png\" srcset=\"https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2454af35af3a0f6cd927dd_architecture-p-1080.png 1080w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2454af35af3a0f6cd927dd_architecture-p-1600.png 1600w, https:\/\/uploads-ssl.webflow.com\/58ef666b690e1f31ad12a43a\/5c2454af35af3a0f6cd927dd_architecture.png 1612w\" id=\"w-node-25e96e05a330-bee9e583\" alt=\"\" \/><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-9dd3cc3ec6b3-bee9e583\" class=\"grid-p\">Defining a combined 2D\u00a0and 3D UI\u00a0to interact with Citybound's quite complex planning and simulation objects is a difficult task. Reactive, component-based frameworks with clear information flow heavily simplified 2D\u00a0UI\u00a0development in recent years. I\u00a0extend React with my <a href=\"https:\/\/github.com\/aeickhoff\/monet\" target=\"_blank\">Monet<\/a> library, applying React's principle (wrapping stateful updates in a declarative, functional API) to WebGL\u00a0rendering and GPU\u00a0resource management. In addition, I\u00a0define a new kind of React component representing an interactive shape in a 3D\u00a0world. Taken together, this allows me to fully define the 2D and 3D appearance and interactivity of a game object in a single React component.<\/p><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-bee6d4fbaa16-bee9e583\" class=\"grid-p\">Almost every object in the Citybound world can be arbitrarily shaped, be it roads, zones or building lots - and their exact shape typically results from complex geometric operations between other objects. My <a href=\"https:\/\/github.com\/aeickhoff\/descartes\" target=\"_blank\">Descartes<\/a> library provides everything from simple path operations to vast polygon-overlay tasks, with a focus on performance and topological robustness despite approximation and inaccuracies.<\/p><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-aa447e19e88c-bee9e583\" class=\"grid-p\">To provide real-time interactive editing of large city plans and to minimise bandwidth usage between UI\u00a0and server while streaming plan result previews, it is vital to have a precise formalisation of how changes in plan input affect plan results, and then communicating only the differences. Currently implemented ad-hoc in a monolithical planning manager, this approach will be generalised and translated into a distributed framework to allow plans to scale beyond one backend server.<\/p><div id=\"w-node-593298d4196d-bee9e583\" class=\"references\"><p data-w-id=\"87017114-683c-cd03-f5d7-593298d4196e\" class=\"grid-p-references-button\"><strong>Details &amp;\u00a0Inspiration \u25be<\/strong><\/p><div class=\"grid-p-references-inner w-richtext\"><p>Papers:<\/p><ul><li>Matthew A. Hammer, Joshua Dunfield, Kyle Headley, Nicholas Labich, Jeffrey S. Foster, Michael W. Hicks, David Van Horn. <em>Incremental computation with names<\/em> - also see<a href=\"http:\/\/adapton.org\"> adapton.org<\/a><\/li><\/ul><\/div><\/div><\/div><div class=\"w-layout-grid grid\"><p id=\"w-node-a620d79e6462-bee9e583\" class=\"grid-p\">The fact that almost all geometry in Citybound is procedural is both an opportunity and a challenge. Early work has begun in my <a href=\"https:\/\/github.com\/aeickhoff\/michelangelo\" target=\"_blank\">Michelangelo<\/a> library that will provide expressive shape-grammars to describe architecture styles or natural shapes based on extrusions, which can then be evaluated on demand to generate optimally-batched meshes and instance-arrays with dynamic level of detail and animated components.<\/p><div id=\"w-node-dc9aa66dff59-bee9e583\" class=\"references\"><p data-w-id=\"ef35c612-a3e4-0d99-e45b-dc9aa66dff5a\" class=\"grid-p-references-button\"><strong>Details &amp;\u00a0Inspiration \u25be<\/strong><\/p><div class=\"grid-p-references-inner w-richtext\"><p>Papers:<\/p><ul><li>Kelly, T., &amp; Wonka, P. (2011). Interactive architectural modeling with procedural extrusions.<\/li><li>Kelly, T. (2013). Unwritten Procedural Modeling with Skeletons.<\/li><\/ul><\/div><\/div><\/div>","created_at":"2019-04-05T16:02:42+0000","updated_at":"2019-04-05T16:02:42+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"aeplay.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":899,"title":"citybound\/citybound","url":"https:\/\/github.com\/citybound\/citybound","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/cb.png\"><img src=\"https:\/\/github.com\/citybound\/citybound\/raw\/master\/cb.png\" alt=\"Citybound\" width=\"192\" \/><\/a><\/p>\n<p>Citybound is a city building game with a focus on realism, collaborative planning and simulation of microscopic details. It is independently developed, open source and funded through Patreon.<\/p>\n<ul><li><a href=\"http:\/\/cityboundsim.com\" rel=\"nofollow\">Homepage<\/a><\/li>\n<li><a href=\"http:\/\/aeplay.co\/citybound-livebuilds\" rel=\"nofollow\">Latest live builds<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/LICENSE.txt\">LICENSE<\/a> (AGPL)<\/li>\n<li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/CONTRIBUTING.md\">Contributing &amp; Development<\/a><\/li>\n<\/ul>\n<p><strong>This is a living and open document, find out <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/CONTRIBUTING.md\">how to contribute your suggestions<\/a>.<\/strong><\/p>\n<p>Citybound is a city building game with a focus on realism, collaborative planning and simulation of microscopic details.<\/p>\n<h2><a id=\"user-content-subtopics\" class=\"anchor\" aria-hidden=\"true\" href=\"#subtopics\"><\/a>Subtopics<\/h2>\n<ul><li>Simulated World\n<ul><li><a href=\"#simulation-core\">Simulation Core<\/a><\/li>\n<li><a href=\"#transport\">Transport<\/a><\/li>\n<li><a href=\"#economy-&amp;-household-behaviour\">Economy &amp; Household Behaviour<\/a><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/construction\/README.md\">Construction<\/a><\/del><\/li>\n<li><a href=\"#environment\">Environment<\/a><\/li>\n<\/ul><\/li>\n<li>Player Interaction\n<ul><li><a href=\"#planning\">Planning<\/a><\/li>\n<li><a href=\"#governance\">Governance<\/a><\/li>\n<\/ul><\/li>\n<li>Engine and Support Libraries\n<ul><li><a href=\"https:\/\/github.com\/aeickhoff\/kay\">kay<\/a> - enabling distributed simulation &amp; multiplayer<\/li>\n<li><a href=\"https:\/\/github.com\/aeickhoff\/descartes\">descartes<\/a> - computational geometry and GIS algorithms<\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-inspiration\" class=\"anchor\" aria-hidden=\"true\" href=\"#inspiration\"><\/a>Inspiration<\/h2>\n<ul><li>Games\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/SimCity\" rel=\"nofollow\">SimCity Series<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/SimAnt\" rel=\"nofollow\">SimAnt<\/a><\/li>\n<li><a href=\"https:\/\/www.old-games.com\/download\/6261\/lego-loco\" rel=\"nofollow\">LEGO Loco<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/A-Train#A-Train_III\" rel=\"nofollow\">A-Train (Take the A-Train III)<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/RollerCoaster_Tycoon\" rel=\"nofollow\">Rollercoaster Tycoon Series<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Factorio\" rel=\"nofollow\">Factorio<\/a><\/li>\n<\/ul><\/li>\n<li>Books\/Ideas\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/A_Pattern_Language\" rel=\"nofollow\">A Pattern Language<\/a> (Christopher Alexander)<\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-philosophy\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy\"><\/a>Philosophy<\/h2>\n<ul><li>Citybound is a game about:\n<ul><li><strong>The beauty of emergent complexity,<\/strong> in the familiar example of a city. (<code>EMERGENCE<\/code>)<\/li>\n<li><strong>The challenge of understanding <code>INTERDEPENDENT SYSTEMS<\/code>,<\/strong> that form a whole.<\/li>\n<li><strong>The experience of <code>PLANNING<\/code>,<\/strong> i.e. having a complex idea, developing it, carefully preparing it and making it real.<\/li>\n<li><strong>The experience of <code>COLLABORATION<\/code>,<\/strong> the way humans work together to achieve great things.<\/li>\n<li>The interplay between the <strong>player's conscious action and the cities' organic reaction<\/strong> (<code>CITY AS AN ORGANISM<\/code>)<\/li>\n<\/ul><\/li>\n<li>Citybound aspires to:\n<ul><li><code>INTERACTION BY PAINTING<\/code>, i.e. through powerful but simple tools with haptic qualities<\/li>\n<li><code>CLARITY OF CAUSE AND EFFECT<\/code>, making patterns of behavior and consequences of one\u2019s actions obvious- in the small and in the large<\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-gameplay-decisions\" class=\"anchor\" aria-hidden=\"true\" href=\"#gameplay-decisions\"><\/a>Gameplay Decisions<\/h2>\n<ul><li>Cities are built in large continuous regions with several million inhabitants (<code>VASTNESS<\/code>)<\/li>\n<li>Time progresses in <code>OVERLAID TIME-SCALES<\/code>: Daily and yearly activities happen at a similar pace. (<code>MULTI-SCALE PHENOMENA<\/code>)<\/li>\n<li>Cities consist of persistent <code>UNIQUE INDIVIDUALS<\/code>, i.e. people &amp; businesses<\/li>\n<\/ul><h2><a id=\"user-content-implementation-philosophy\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>Develop systems from first principles, with complex behaviors emerging from simple microscopic interactions <code>(emergence)<\/code><\/li>\n<li>Solve problems generally, do not restrict thinking in specialized problem instances <code>(actual problem)<\/code><\/li>\n<li>Be brave to do things in new ways, where necessary <code>(radicality)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions\"><\/a>Implementation Decisions<\/h2>\n<ul><li>Use Rust as the main programming language for high performance and a strong type system to rely on<\/li>\n<li>Use an actor-based model for isolation, dynamic dispatch and simple parallelization and networking<\/li>\n<li>Use a Rust &amp; JavaScript combination to implement a cross-platform, WebGL based UI that communicates with a simulation backend<\/li>\n<\/ul><h2><a id=\"user-content-hierarchy-of-player-skills\" class=\"anchor\" aria-hidden=\"true\" href=\"#hierarchy-of-player-skills\"><\/a>Hierarchy of Player Skills<\/h2>\n<ul><li>\"Building a city\"\n<ul><li>\"Observation &amp; Inspection\"\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/inspection\/stats\/README.md\">\"Reading Statistics &amp; Maps\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/inspection\/households\/README.md\">\"Inspecting Individual Households\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/inspection\/stories\/README.md\">\"Reading Stories\"<\/a><\/del><\/li>\n<\/ul><\/li>\n<li>\"Planning\"\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/transport\/planning\/README.md\">\"Transport Infrastructure Planning\"<\/a><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/utilities\/README.md\">\"Utility Infrastructure Planning\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/zoning\/README.md\">\"Zoning\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/services\/README.md\">\"Services Planning\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/environment\/terraforming\/README.md\">\"Terraforming\"<\/a><\/del><\/li>\n<\/ul><\/li>\n<li>\"Execution &amp; Finances\"\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/projects\/README.md\">\"Projects &amp; Grants\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/finances\/README.md\">\"Budgeting\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/finances\/README.md\">\"Taxation\"<\/a><\/del><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul>\n<p>The simulation core takes care of the progression and scales of in-game time by frequently updating continuous processes as well as sending one-off notifications and timeouts<\/p>\n<h2><a id=\"user-content-prior-art--inspiration\" class=\"anchor\" aria-hidden=\"true\" href=\"#prior-art--inspiration\"><\/a>Prior Art &amp; Inspiration<\/h2>\n<h2><a id=\"user-content-inspiration-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#inspiration-1\"><\/a>Inspiration<\/h2>\n<ul><li>Compressed and overlaid timescales to make the player experience short-term and long-term phenomena at once <code>(multi-scale)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-decisions\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions\"><\/a>Decisions<\/h2>\n<h2><a id=\"user-content-implementation-philosophy-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-1\"><\/a>Implementation Philosophy<\/h2>\n<h2><a id=\"user-content-implementation-decisions-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-1\"><\/a>Implementation Decisions<\/h2>\n<h2><a id=\"user-content-subtopics-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#subtopics-1\"><\/a>Subtopics<\/h2>\n<ul><li>Planning\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/transport_planning\/README.md\">Transport Planning<\/a> (40% alpha)<\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Rail &amp; Tram Planning<\/a><\/del> (0% alpha)<\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Channel\/Waterway Planning<\/a><\/del> (0% alpha)<\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Airport Planning<\/a><\/del> (0% alpha)<\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Public Transport Planning<\/a><\/del> (0% alpha)<\/li>\n<\/ul><\/li>\n<li>Construction\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/construction\/README.md\">Road Construction<\/a> (50% alpha)<\/li>\n<\/ul><\/li>\n<li>Road Trips &amp; Pathfinding\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/pathfinding\/README.md\">Road Trips &amp; Pathfinding<\/a> (10% alpha)<\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Shared &amp; Scheduled Vehicles<\/a><\/del> (0% alpha)<\/li>\n<\/ul><\/li>\n<li>Traffic\n<ul><li>Pedestrian Traffic (0% alpha)\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Built &amp; Organic Paths<\/a><\/del>, <del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Pedestrian Path Microtraffic<\/a><\/del><\/li>\n<\/ul><\/li>\n<li>Road Traffic (50% alpha)\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/lane\/README.md\">Road Lanes<\/a> &amp; <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/microtraffic\/README.md\">Car Microtraffic<\/a><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/rendering\/README.md\">Rendering<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Parking<\/a><\/del> (0% alpha)<\/li>\n<\/ul><\/li>\n<li>Rail Traffic (0% alpha)\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Train Microtraffic<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Stations<\/a><\/del><\/li>\n<\/ul><\/li>\n<li>Water Traffic (0% beta)\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Vessel Microtraffic<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Ports<\/a><\/del><\/li>\n<\/ul><\/li>\n<li>Air Traffic (0% beta)\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Airplane &amp; Taxiing Microtraffic<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Airports<\/a><\/del><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-philosophy-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-1\"><\/a>Philosophy<\/h2>\n<ul><li>Traffic is the pulse of a city <code>(city = organism)<\/code><\/li>\n<li>Traffic is a chaotic emergent phenomenon (<code>EMERGENCE<\/code>)<\/li>\n<li>Traffic shapes the city and the city shapes traffic <code>(interdependent systems)<\/code><\/li>\n<li>Traffic shapes lives and lives shape traffic <code>(interdependent systems)<\/code><\/li>\n<li>People and goods use a multitude\/combinations of transport systems <code>(diverse individuals)<\/code><\/li>\n<li>Pedestrians &amp; vehicles move \"like different kinds of swarms\/liquids\" <code>(diverse individuals)<\/code> <code>(city = anthill)<\/code> <code>(vastness)<\/code> <code>(flows)<\/code>\n<ul><li>individuals are the most visible when they are on the move<\/li>\n<li>crowds look antsy, sprawling, colorful<\/li>\n<li>cars can flow nicely but also get really pushy, different types\/sizes of cars have obviously different silhouettes and movement patterns<\/li>\n<li>trains and boats are heavy and careful, they have impressive momentum<\/li>\n<li>planes are graceful and precise but need a lot of space<\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-decisions-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-1\"><\/a>Decisions<\/h2>\n<ul><li>Always model each trip individually and microscopically (never use statistical models) <code>(actual problem)<\/code> <code>(emergence)<\/code>\n<ul><li>Physical motion through time\/space<\/li>\n<li>Realistic use of potentially several modes of transport in sequence<\/li>\n<li>Counteract time-compression with generous behaviour parameters (but never cheat, i.e. teleport) <code>(actual problem) (time compression)<\/code><\/li>\n<\/ul><\/li>\n<li>Pedestrians, road traffic, public transport and freight transport of all kinds are all equally important<\/li>\n<li>Parking is a core problem &amp; component of road traffic<\/li>\n<\/ul><h2><a id=\"user-content-implementation-philosophy-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-2\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>Simulation goal: trips are successful often enough and efficient enough to power all economic transaction realistically<\/li>\n<li>All kinds of traffic are \"vessels\/bodies moving along the edges of a transport network graph\"<\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-2\"><\/a>Implementation Decisions<\/h2>\n<ul><li>Optimal granularity of traffic simulation: simulate the most strongly locally interacting traffic entities (e.g. cars on one lane) all at once within one actor (one lane = one actor)<\/li>\n<\/ul><h2><a id=\"user-content-gameplay--skills\" class=\"anchor\" aria-hidden=\"true\" href=\"#gameplay--skills\"><\/a>Gameplay &amp; Skills<\/h2>\n<ul><li>\"Building a city\"\n<ul><li>\"Planning\"\n<ul><li>\"Transport Infrastructure Planning\"\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/planning\/README.md#road-planning-skill\">\"Road Planning\"<\/a><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">\"Rail &amp; Tram Planning\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">\"Channel\/Waterway Planning\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">\"Airport Planning\"<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">\"Public Transport Planning\"<\/a><\/del><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul>\n<h2><a id=\"user-content-philosophy--decisions\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy--decisions\"><\/a>Philosophy &amp; Decisions<\/h2>\n<ul><li>See <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/README.md\">Transport<\/a><\/li>\n<\/ul><h2><a id=\"user-content-implementation-philosophy-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-3\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>Keep car interactions as local as possible computationally<\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-3\"><\/a>Implementation Decisions<\/h2>\n<ul><li><strong>Cars themselves are not actors, one lane is the atomic actor,<\/strong> it updates all the cars on it in one go<\/li>\n<li>Cars mostly move along a lane like a train, only on switch lanes they can move sideways to merge between normal lanes<\/li>\n<li>Cars react to general obstacles, which are either the next car on their own lane, or cars on interacting lanes, whose position is mapped onto the current lane\n<ul><li>Interactions are distinguished as parallel\/opposing\/merging<\/li>\n<li>Interacting lanes after each update exchange their cars as obstacles for the other lane<\/li>\n<\/ul><\/li>\n<li>Use well-known 1D acceleration and breaking model and extend it to multi-lane behaviour\n<ul><li>\"Intelligent Driver Model\", see <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/intelligent_acceleration.rs\">intelligent_acceleration.rs<\/a><\/li>\n<\/ul><\/li>\n<li>Cars on switch lanes have more complex behaviour, reacting to cars on their left &amp; right lanes and other cars merging between them.\n<ul><li>Based on the perceived risk\/required breaking, merging cars can decide to abort their merge<\/li>\n<\/ul><\/li>\n<li>Also see <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/lane\/README.md\">Road Lanes<\/a><\/li>\n<\/ul>\n<h2><a id=\"user-content-philosophy-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-2\"><\/a>Philosophy<\/h2>\n<ul><li>The complex fabric of real road geometry goes beyond \"streets\" and \"intersections\", which at best can be understood as loose groups of related lanes (<code>EMERGENCE<\/code>, <code>ACTUAL PROBLEM<\/code>)<\/li>\n<\/ul><h2><a id=\"user-content-design-decision\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-decision\"><\/a>Design Decision<\/h2>\n<ul><li>The atomic unit of road geometry in Citybound is the individual lane.<\/li>\n<li>Roads and intersections are just collections of lanes <code>(emergence)<\/code> <code>(flexibility)<\/code><\/li>\n<\/ul><h2><a id=\"user-content-implementation-philosophy-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-4\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>Really make individual lanes the main abstraction, only group them into higher abstractions when absolutely necessary.<\/li>\n<li>If possible, only handle a group of \"nearby\/related\" lanes temporarily while using a tool<\/li>\n<li>Also see <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/planning\/README.md\">Road Planning<\/a><\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-4\"><\/a>Implementation Decisions<\/h2>\n<ul><li>Each lane keeps track of\n<ul><li>it's connectivity to preceding\/following lanes (connected to it's start or end)<\/li>\n<li>lanes it overlaps, distinguishing between nearly-parallel overlaps and opposing overlaps<\/li>\n<\/ul><\/li>\n<li>Special case: two parallel lines that can be merged between are not directly connected\n<ul><li>instead, an invisible \"switch lane\" is placed along their range of overlap<\/li>\n<li>the lanes on either side have a special parallel overlap relation ship to the switch lane only, not to each other.<\/li>\n<li>the reason for this is that merging car behaviour is a more complex special case<\/li>\n<\/ul><\/li>\n<li>Also see <a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/microtraffic\/README.md\">Car Microtraffic<\/a><\/li>\n<\/ul>\n<h2><a id=\"user-content-prior-art--inspiration-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#prior-art--inspiration-1\"><\/a>Prior Art &amp; Inspiration<\/h2>\n<ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=xfbVPDNl7V4\" rel=\"nofollow\">Prof. Philip Mirowski: Should Economists be Experts in Markets or Experts in \"Human Nature\"?<\/a><\/li>\n<li>Roller Coaster Tycoon as the gold standard for individual, grouped, localised thoughts as a key tool to inform game decisions<\/li>\n<\/ul><h2><a id=\"user-content-philosophy-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-3\"><\/a>Philosophy<\/h2>\n<ul><li>\n<p>All macro-economic behaviours arise from interactions of individual households (families, companies) <code>(emergence)<\/code><\/p>\n<\/li>\n<li>\n<p>Members of households have both selfish and shared goals, resulting in a trade-off between individuality and cooperation<\/p>\n<\/li>\n<li>\n<p>Locality of households and interactions is essential and enables all kinds of logistics and distributed industries<\/p>\n<ul><li>All resources that have to physically moved in real life have to also be moved in Citybound<\/li>\n<\/ul><\/li>\n<li>\n<p>The real estate market is a huge factor in shaping a city<\/p>\n<\/li>\n<li>\n<p>The character of a city is largely determined by the sum of its industries<\/p>\n<\/li>\n<li>\n<p>The vast majority of consumed resources and services depend on complex chains of earlier resources, these resource chains\/trees are at the core of shaping industries, the required logistics and thus also a city<\/p>\n<\/li>\n<li>\n<p>The player is not supposed to manage supply\/demand\/resource chains, but to enable them to be as effective as possible, through the means of infrastructure planning and zoning\/governance<\/p>\n<\/li>\n<li>\n<p>Even though the amount of information presented in the game might go way beyond realistically knowable things, this contrast in availability of information compared to the real world has an educational and almost call-to-action aspect to it<\/p>\n<\/li>\n<li>\n<p>Individuals act as a lens to see the game world from, other than the players birds-eye-view<\/p>\n<\/li>\n<li>\n<p>Overall \"macro\" appearance of the city hints at general issues which might inspire you to investigate in detail<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-decisions-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-2\"><\/a>Decisions<\/h2>\n<ul><li>Provide both omniscient-level information about individuals as well as realistic large-scale indicators<\/li>\n<li>Actually model the real estate market with ownership and businesses, including city government ownership<\/li>\n<li>All complex resources can only be a) traded or b) created from its component parts, never \"out of nowhere\"<\/li>\n<\/ul><h2><a id=\"user-content-implementation-philosophy-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-5\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>Having to rely on absolute amounts for resources makes balancing very hard and amplifies bugs and makes the system potentially very unstable<\/li>\n<li>The amount of different resource types and how fine-grained they are is to be chosen as minimal as possible, while still producing interesting and prevalent resource chains<\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-5\"><\/a>Implementation Decisions<\/h2>\n<ul><li>Resources represent both tangible goods as well as intangible (psychological, physiological, social) metrics<\/li>\n<li>All of these are measured using scalar values, from -infinity to infinity\n<ul><li>these values represent how well a household is doing on a particular resource (0 representing ideal), not necessarily absolute stockpile values<\/li>\n<\/ul><\/li>\n<li>One household and its associated resources are always tied to one specific building\/location - if this is one location of a larger company, these cooperate mostly as if they were separate businesses<\/li>\n<li>\"Crafting recipes\" to determine both how resource combine, at which costs, and which businesses do that<\/li>\n<\/ul><h2><a id=\"user-content-parts\" class=\"anchor\" aria-hidden=\"true\" href=\"#parts\"><\/a>Parts<\/h2>\n<ul><li>Households (20% alpha)\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\/households\/family\">Families &amp; Persons<\/a><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Businesses \/ Industries<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Neighboring Cities<\/a><\/del><\/li>\n<\/ul><\/li>\n<li>Markets\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Market for Goods<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Real Estate Market<\/a><\/del><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-skills\" class=\"anchor\" aria-hidden=\"true\" href=\"#skills\"><\/a>Skills<\/h2>\n<ul><li>\"Building a city\"\n<ul><li>\"Observation &amp; Inspection\"\n<ul><li>\"Inspecting Resource Chains\"\n<ul><li>Godmode-style knowledge\n<ul><li>\"Inspect all consumers of the products of a business\"<\/li>\n<li>\"Inspect all the source producers of input resources of a business\"<\/li>\n<li>\"Inspect the full journey\/lifecycle of an end product, including becoming waste\/...\"<\/li>\n<li>\"Business surplus\/deficits per resource category\"<\/li>\n<li>\"Infer bottlenecks\/problems by showing all related \"nodes\" in the economic network of one particular business, including paths\"<\/li>\n<\/ul><\/li>\n<li>Realistic knowledge\n<ul><li>\"Infer industry composition based on evident architectural styles\"<\/li>\n<li>\"Directly show all businesses of a precise resource category\"<\/li>\n<li>\"Infer causes of homelessness\"<\/li>\n<li>\"Learn from protests\"<\/li>\n<li>\"Learn from topical advisors\" ?<\/li>\n<li>\"Learn about a neighborhood by looking at it\"<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li>\"Planning\"<\/li>\n<li>\"Execution &amp; Finances\"<\/li>\n<\/ul><\/li>\n<\/ul>\n<h2><a id=\"user-content-prior-art--inspiration-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#prior-art--inspiration-2\"><\/a>Prior Art &amp; Inspiration<\/h2>\n<h2><a id=\"user-content-philosophy-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-4\"><\/a>Philosophy<\/h2>\n<h2><a id=\"user-content-decisions-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-3\"><\/a>Decisions<\/h2>\n<h2><a id=\"user-content-implementation-philosophy-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-6\"><\/a>Implementation Philosophy<\/h2>\n<h2><a id=\"user-content-implementation-decisions-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-6\"><\/a>Implementation Decisions<\/h2>\n<h2><a id=\"user-content-parts-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#parts-1\"><\/a>Parts<\/h2>\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Zoning &amp; Spatial Decrees<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Taxation &amp; Budgeting<\/a><\/del><\/li>\n<\/ul>\n<h2><a id=\"user-content-prior-art--inspiration-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#prior-art--inspiration-3\"><\/a>Prior Art &amp; Inspiration<\/h2>\n<ul><li>Code version control systems and file difference visualizers<\/li>\n<li>Blueprints\/ghost parts in Factorio<\/li>\n<\/ul><h2><a id=\"user-content-philosophy-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-5\"><\/a>Philosophy<\/h2>\n<ul><li>In reality it is not only physically impossible to instantaneously build infrastructure, but there are good reasons why planning is employed before even deciding what to build <code>(planning)<\/code>\n<ul><li>Plans are easy to revise and iterate on, changing existing structure is extremely costly<\/li>\n<li>You can start planning something even when you do not have the permission or resources to implement the plan yet. In fact, you might need to have a plan to show in order to acquire the necessary permission and resources<\/li>\n<li>Planning should be as flexible and forgiving as possible<\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-decisions-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-4\"><\/a>Decisions<\/h2>\n<ul><li>The player only ever interacts directly with plans. The only way to affect construction of infrastructure is through implementation of plans.<\/li>\n<li>Plans have full undo\/redo history.<\/li>\n<li>Plans clearly show structures to be added and to be removed relative to what exists. <code>(clarity)<\/code><\/li>\n<li>Modifications of existing structures seamlessly become part of a plan, this should feel as tangible as creating new structures.<\/li>\n<\/ul><h2><a id=\"user-content-implementation-philosophy-7\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-7\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>\"Materialized Reality\" and plans exists in completely different \"spheres\"<\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions-7\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-7\"><\/a>Implementation Decisions<\/h2>\n<ul><li>Instead of keeping a perfect, complicated two-way mapping between plans and implemented reality at all times, they are kept separate<\/li>\n<li>When new structures are added to a plan or previously planned and implemented structures are modified within the plan, they are compared geometrically in detail with the materialized reality to calculate the actual diff of what would change when implementing the new plan<\/li>\n<\/ul>\n<h2><a id=\"user-content-philosophy-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-6\"><\/a>Philosophy<\/h2>\n<h2><a id=\"user-content-decisions-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-5\"><\/a>Decisions<\/h2>\n<h2><a id=\"user-content-implementation-philosophy-8\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-8\"><\/a>Implementation Philosophy<\/h2>\n<ul><li>Keep pathfinding information local and dynamically changeable<\/li>\n<li>Make use of the \"highway effect\" (for most of each trip, only the rough source and rough destination determine the typically important main connections to choose)<\/li>\n<li>Be able to create a trip from household to household (not necessarily precise addresses)<\/li>\n<\/ul><h2><a id=\"user-content-implementation-decisions-8\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-8\"><\/a>Implementation Decisions<\/h2>\n<ul><li>Trips are a persistent representation of each ongoing trip (since cars themselves are not actors, but exist on road actors, or are exchanged as messages)\n<ul><li>Trips can be created from and to \"rough\" locations (such as households), which are then first resolved to precise locations on the road network<\/li>\n<\/ul><\/li>\n<li>Use a simple two-layer landmark model for the pathfinding itself:\n<ul><li>\"nodes\" (individual lanes) and \"landmarks\" (groups of lanes, represented as one rough destination)<\/li>\n<li>nodes store closest-next-hop information to <em>all<\/em> landmarks, but only to nodes up to a certain distance<\/li>\n<li>if a node is too far away, first navigate towards its parent landmark, until a direct next-hop is known<\/li>\n<li>lanes organically join\/leave landmarks, keeping landmarks continuous each, and of roughly equal group size<\/li>\n<li>lanes exchange information about changed\/updated next-hops with their local neighbors (similar to router table updates), forming an eventually converging network<\/li>\n<\/ul><\/li>\n<\/ul>\n<h2><a id=\"user-content-prior-art--inspiration-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#prior-art--inspiration-4\"><\/a>Prior Art &amp; Inspiration<\/h2>\n<h2><a id=\"user-content-philosophy-7\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-7\"><\/a>Philosophy<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/citybound\/citybound\/planning\/README.md\">Planning in general<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/lane\/README.md\">Lanes as the atomic unit of roads &amp; intersections<\/a><\/li>\n<\/ul><h2><a id=\"user-content-decisions-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-6\"><\/a>Decisions<\/h2>\n<h2><a id=\"user-content-implementation-philosophy-9\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-9\"><\/a>Implementation Philosophy<\/h2>\n<h2><a id=\"user-content-implementation-decisions-9\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-9\"><\/a>Implementation Decisions<\/h2>\n<h2><a id=\"user-content-actions---effects\" class=\"anchor\" aria-hidden=\"true\" href=\"#actions---effects\"><\/a>Actions -&gt; Effects<\/h2>\n<h2><a id=\"user-content-effects---feedback\" class=\"anchor\" aria-hidden=\"true\" href=\"#effects---feedback\"><\/a>Effects -&gt; Feedback<\/h2>\n<h2><a id=\"user-content-feedback---learning\" class=\"anchor\" aria-hidden=\"true\" href=\"#feedback---learning\"><\/a>Feedback -&gt; Learning<\/h2>\n<h2><a id=\"user-content-prior-art--inspiration-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#prior-art--inspiration-5\"><\/a>Prior Art &amp; Inspiration<\/h2>\n<h2><a id=\"user-content-philosophy-8\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-8\"><\/a>Philosophy<\/h2>\n<h2><a id=\"user-content-decisions-7\" class=\"anchor\" aria-hidden=\"true\" href=\"#decisions-7\"><\/a>Decisions<\/h2>\n<h2><a id=\"user-content-implementation-philosophy-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-philosophy-10\"><\/a>Implementation Philosophy<\/h2>\n<h2><a id=\"user-content-implementation-decisions-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-decisions-10\"><\/a>Implementation Decisions<\/h2>\n<h2><a id=\"user-content-parts-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#parts-2\"><\/a>Parts<\/h2>\n<ul><li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Terrain<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Water &amp; Weather<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Vegetation &amp; Pollution<\/a><\/del><\/li>\n<li><del><a href=\"https:\/\/github.com\/citybound\/citybound\/blob\/master\">Culture &amp; Political Context<\/a><\/del><\/li>\n<\/ul>\n<h2><a id=\"user-content-philosophy-9\" class=\"anchor\" aria-hidden=\"true\" href=\"#philosophy-9\"><\/a>Philosophy<\/h2>\n<ul><li>Not only the exchange of physical goods influences the decision-making of families, but intangible psychological, physiological, social metrics play an equally if not more important role<\/li>\n<li>There is no central happiness metric, but only individual diverse motivations that vary both over time as well as from person to person - one can only get a rough overview of \"how well someone is doing\" by looking at all of them<\/li>\n<\/ul><\/article>","created_at":"2019-04-05T16:01:29+0000","updated_at":"2019-04-05T16:01:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/15174518?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","3d-printing"],"is_public":false,"id":898,"title":"3D Printing With Clojure","url":"https:\/\/adereth.github.io\/blog\/2014\/04\/09\/3d-printing-with-clojure\/","content":"<p>I\u2019ve been doing some 3D printing for <a href=\"https:\/\/twitter.com\/adereth\/status\/444145229109555200\/photo\/1\">my next keyboard project<\/a> and I\u2019ve got a workflow that I\u2019m pretty happy with that I\u2019d like to share.<\/p><p>When I first started trying to make models a month ago, I tried <a href=\"http:\/\/www.blender.org\/\">Blender<\/a>.  It\u2019s an amazing beast, but after a few hours of tutorials it was clear that it would take a while to get proficient with it.  Also, it is really designed for interactive modeling and I need something that I can programmatically tweak.<\/p>\n<h2>OpenSCAD<\/h2>\n<p><img src=\"https:\/\/adereth.github.io\/images\/openscad.gif\" alt=\"OpenSCAD Screenshot\" \/><\/p>\n<p>A couple of friends suggested <a href=\"http:\/\/www.openscad.org\/\">OpenSCAD<\/a>, which is touted as \u201cthe programmers\u2019 solid 3D CAD modeler.\u201d  It provides a power set of primitive shapes and operations, but the language itself leaves a bit to be desired.  This isn\u2019t a beat-up-on-SCAD post, but a few of the things that irked me were:<\/p>\n<ul><li>Strange function application syntax (parameters in parens after the function name with an expression or block following the closing paren)<\/li>\n<li>Unclear variable binding rules (multiple passes are made over the code and the results of changing a variable may affect things earlier in the code unexpectedly)<\/li>\n<li>No package\/namespace management<\/li>\n<li>Multiple looping constructs that depend on what you are going to do with the results, not on how you want to loop<\/li>\n<\/ul><h2>scad-clj<\/h2>\n<p>Fortunately, <a href=\"https:\/\/github.com\/farrellm\">Matt Farrell<\/a> has written <a href=\"https:\/\/github.com\/farrellm\/scad-clj\">scad-clj<\/a>, an OpenSCAD DSL in Clojure.  It addresses every issue I had with OpenSCAD and lends itself to a really nice workflow with the Clojure REPL.<\/p>\n<p>To get started using it, add the dependency on <a href=\"https:\/\/clojars.org\/scad-clj\"><code>[scad-clj \"0.1.0\"]<\/code><\/a> to your <code>project.clj<\/code> and fire up your REPL.<\/p>\n<p>All of the functions for creating 3D models live in the <code>scad-clj.model<\/code> namespace.  There\u2019s no documentation yet, so in the beginning you\u2019ll have to look at the <a href=\"https:\/\/github.com\/farrellm\/scad-clj\/blob\/master\/src\/scad_clj\/model.clj\">source for <code>model.clj<\/code><\/a> and occassionally the <a href=\"http:\/\/www.openscad.org\/documentation.html\">OpenSCAD documentation<\/a>.  Fortunately, there really isn\u2019t much to learn and it\u2019s quite a revelation to discover that almost everything you\u2019ll want to do can be done with a handful of functions.<\/p>\n<p>Here\u2019s a simple model that showcases each of the primitive shapes:<\/p>\n<figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n(def primitives\n  (union\n   (cube 100 100 100)\n   (sphere 110)\n   (cylinder 10 150)))\n<\/figcaption><\/figure><p>Evaluating this gives us a data structure that can be converted into an .scad file using <code>scad-clj.scad\/write-scad<\/code> to generate a string and <code>spit<\/code>.<\/p>\n<figure class=\"code\"><figcaption>1\n2\n(spit \"post-demo.scad\"\n      (write-scad primitives))\n<\/figcaption><\/figure><p>We\u2019re going to use OpenSCAD to view the results.  One feature of OpenSCAD that is super useful for this workflow is that it watches opened files and automatically refreshes the rendering when the file is updated.  This means that we can just re-evaluate our Clojure code and see the results immediately in another window:<\/p>\n<p><img src=\"https:\/\/adereth.github.io\/images\/scad-primitives.png\" alt=\"Primitives Screenshot\" \/><\/p>\n<p>scad-clj makes all new primitive shapes centered at the origin.  We can use the shape operator functions to move them around and deform them:<\/p>\n<figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n(def primitives\n  (union\n   (-&gt;&gt; (cube 100 100 100)\n        (rotate (\/ Math\/PI 4) [1 1 1])\n        (translate [150 0 0]))\n   (-&gt;&gt; (sphere 70)\n        (scale [1\/2 1\/2 2])\n        (translate [-150 0 0]))\n   (cylinder 10 160)))\n<\/figcaption><\/figure><p><img src=\"https:\/\/adereth.github.io\/images\/scad-operators.png\" alt=\"Operator Screenshot\" \/><\/p>\n<p>I snuck <code>union<\/code> into those examples.  Shapes can also be combined using <code>intersection<\/code>, <code>difference<\/code>, and <code>hull<\/code>.  It\u2019s pretty incredible how much can be done with just these.  For example, here\u2019s the latest iteration of my keyboard design built using clj-scad:<\/p>\n<p><img src=\"https:\/\/adereth.github.io\/images\/scad-keyboard.png\" alt=\"Keyboard\" \/><\/p>\n<h2>3D Printing<\/h2>\n<p>Once your design is complete, you can use OpenSCAD to export it as an STL file which can then be imported to software like <a href=\"http:\/\/replicat.org\/\">ReplicatorG<\/a> or <a href=\"https:\/\/www.makerbot.com\/makerware\/\">Makerware<\/a> for processing into an .x3g file that can be printed:<\/p>\n<p><img src=\"https:\/\/adereth.github.io\/images\/printed.JPG\" alt=\"Keyboard\" \/><\/p>","created_at":"2019-04-05T13:26:10+0000","updated_at":"2019-04-05T13:26:18+0000","published_at":"2014-04-09T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"adereth.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":897,"title":"LisaDziuba\/Awesome-Design-Tools","url":"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><div>\n<p style=\"text-align: center;\">\n    <a href=\"https:\/\/flawlessapp.io\/feedback?github\" rel=\"nofollow\">\n        <img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Awesome-design-tools-cover.png\" alt=\"Awesome-Design-Tools\" \/><\/a>\n<\/p>\n<p style=\"text-align: center;\">\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/13c4e50d88df7178ae1882a203ed57b641674f94\/68747470733a2f2f63646e2e7261776769742e636f6d2f73696e647265736f726875732f617765736f6d652f643733303566333864323966656437386661383536353265336136336531353464643865383832392f6d656469612f62616467652e737667\"><img alt=\"awesome\" src=\"https:\/\/camo.githubusercontent.com\/13c4e50d88df7178ae1882a203ed57b641674f94\/68747470733a2f2f63646e2e7261776769742e636f6d2f73696e647265736f726875732f617765736f6d652f643733303566333864323966656437386661383536353265336136336531353464643865383832392f6d656469612f62616467652e737667\" data-canonical-src=\"https:\/\/cdn.rawgit.com\/sindresorhus\/awesome\/d7305f38d29fed78fa85652e3a63e154dd8e8829\/media\/badge.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/d4e0f63e9613ee474a7dfdc23c240b9795712c96\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f5052732d77656c636f6d652d627269676874677265656e2e737667\"><img alt=\"PRs Welcome\" src=\"https:\/\/camo.githubusercontent.com\/d4e0f63e9613ee474a7dfdc23c240b9795712c96\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f5052732d77656c636f6d652d627269676874677265656e2e737667\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/PRs-welcome-brightgreen.svg\" \/><\/a>\n<\/p>\n<h2><a id=\"user-content-how-to-use-and-contribute\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-use-and-contribute\"><\/a>How to Use and Contribute<\/h2>\n<ul><li>Go through the Table of Contents and choose the tool.<\/li>\n<li>Search for a keyword or phrase (for example, \"animation\", \"prototyping\").<\/li>\n<li>Ask <a href=\"https:\/\/twitter.com\/LisaDziuba\" rel=\"nofollow\">Lisa<\/a> on Twitter.<\/li>\n<\/ul><p>If you found some great design tool, just send a Pull Request with respect to our <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Contribution_Guidelines.md\">Contribution Guidelines<\/a> (they're very simple, please take a look). We use such labels for free <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a>, open source <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a> and Mac only <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a> tools, don't forget to add them. Now I'd love to see your suggestions!<\/p>\n<\/div>\n<div>\n<h2><a id=\"user-content-table-of-content\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-content\"><\/a>Table of Content<\/h2>\n<ul><li><a href=\"#accessibility-tools\">Accessibility Tools<\/a><\/li>\n<li><a href=\"#animation-tools\">Animation Tools<\/a><\/li>\n<li><a href=\"#augmented-reality\">Augmented Reality<\/a><\/li>\n<li><a href=\"#collaboration-tools\">Collaboration Tools<\/a><\/li>\n<li><a href=\"#color-picker-tools\">Color Picker Tools<\/a><\/li>\n<li><a href=\"#design-feedback-tools\">Design Feedback Tools<\/a><\/li>\n<li><a href=\"#design-handoff-tools\">Design Handoff Tools<\/a><\/li>\n<li><a href=\"#design-inspiration\">Design Inspiration<\/a><\/li>\n<li><a href=\"#design-system-tools\">Design System Tools<\/a><\/li>\n<li><a href=\"#design-to-code-tools\">Design to Code Tools<\/a><\/li>\n<li><a href=\"#design-version-control\">Design Version Control<\/a><\/li>\n<li><a href=\"#experience-monitoring\">Experience Monitoring<\/a><\/li>\n<li><a href=\"#font-tools\">Font Tools<\/a><\/li>\n<li><a href=\"#gradient-tools\">Gradient Tools<\/a><\/li>\n<li><a href=\"#icons-tools\">Icons Tools<\/a><\/li>\n<li><a href=\"#illustrations\">Illustrations<\/a><\/li>\n<li><a href=\"#information-architecture\">Information Architecture<\/a><\/li>\n<li><a href=\"#logo-design\">Logo Design<\/a><\/li>\n<li><a href=\"#mockup-tools\">Mockup Tools<\/a><\/li>\n<li><a href=\"#no-code-tools\">No Code Tools<\/a><\/li>\n<li><a href=\"#pixel-art-tools\">Pixel Art Tools<\/a><\/li>\n<li><a href=\"#prototyping-tools\">Prototyping Tools<\/a><\/li>\n<li><a href=\"#screenshot-software\">Screenshot Software<\/a><\/li>\n<li><a href=\"#sketching-tools\">Sketching Tools<\/a><\/li>\n<li><a href=\"#sound-design\">Sound Design<\/a><\/li>\n<li><a href=\"#stock-photos-tools\">Stock Photos Tools<\/a><\/li>\n<li><a href=\"#stock-videos\">Stock Videos<\/a><\/li>\n<li><a href=\"#tools-for-learning-design\">Tools for Learning Design<\/a><\/li>\n<li><a href=\"#ui-design-tools\">UI Design Tools<\/a><\/li>\n<li><a href=\"#user-flow-tools\">User Flow Tools<\/a><\/li>\n<li><a href=\"#user-research-tools\">User Research Tools<\/a><\/li>\n<li><a href=\"#visual-debugging-tools\">Visual Debugging Tools<\/a><\/li>\n<li><a href=\"#wireframing-tools\">Wireframing Tools<\/a><\/li>\n<li><a href=\"#three-d-modeling-software\">3D Modeling Software<\/a><\/li>\n<\/ul><\/div>\n<p style=\"text-align: center;\">\n    <a href=\"https:\/\/flawlessapp.io\/feedback?github\" rel=\"nofollow\">\n        <img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/flawless-feedback.png\" alt=\"Awesome-Design-Tools\" \/><\/a>\n<\/p>\n<h3><a id=\"user-content-accessibility-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#accessibility-tools\"><\/a>Accessibility Tools<\/h3>\n<p>Accessibility is the practice of creating websites and apps usable for all people, including individuals with visual, motor, auditory, speech, or cognitive disabilities. Here you will find web accessibility tools, accessibility testing tools, and accessibility apps both for developers and designers:<\/p>\n<ul><li><a href=\"https:\/\/www.ssa.gov\/accessibility\/andi\/help\/install.html\" rel=\"nofollow\">ANDI<\/a> \u2014 accessibility testing tool for web content (bookmarklet). It will automatically detect accessibility issues, give suggestions to improve online accessibility and check 508 compliance. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/SSAgov\/ANDI\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.deque.com\/axe\/\" rel=\"nofollow\">Axe<\/a> \u2014 accessibility engine designed to work on all modern browsers and with whatever tools, frameworks, libraries and environments you use today. It's automated accessibility testing tool for developers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/dequelabs\/axe-core\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/www.colorbox.io\/\" rel=\"nofollow\">ColorBox<\/a> \u2014 a web app that algorithmically builds accessible color systems. Done by Lyft Design team. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/contrasteapp.com\/\" rel=\"nofollow\">Contraste<\/a> \u2014 an app for checking the accessibility of text against the Web Content Accessibility Guidelines (WCAG). <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/usecontrast.com\/\" rel=\"nofollow\">Contrast<\/a> \u2014 a macOS app for designers and developers to get quick access to WCAG color contrast ratios. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/hexnaw.com\/\" rel=\"nofollow\">Hex Naw<\/a> \u2014 helps you to test entire color systems for contrast and accessibility. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/pa11y.org\/\" rel=\"nofollow\">PA11Y<\/a> \u2014 accessibility testing tool to find issues with your web pages. It runs HTML CodeSniffer from the command line for programmatic accessibility reporting. It's accessibility developer tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/michelf.ca\/projects\/sim-daltonism\/\" rel=\"nofollow\">Sim Daltonism<\/a> \u2014 a color blindness simulator for macOS and iOS that lets you visualize colors as they are perceived with various types of color blindness. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/www.github.com\/michelf\/sim-daltonism\/\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/khan.github.io\/tota11y\/\" rel=\"nofollow\">tota11y<\/a> \u2014 an accessibility visualization toolkit. Interesting fact: inside the tool name you can see \"a11y\". It's an abbreviation of accessibility as \"a\" followed by 11 more letters, followed by \"y\". <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/wave.webaim.org\/\" rel=\"nofollow\">WAVE<\/a> \u2014 allows you to evaluate web content for accessibility issues directly within Chrome and Firefox. It's a web accessibility checker. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>Accessibility checking feature is also available in VisBug from <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools#visual-debugging-tools\">Visual Debugging Tools<\/a> categoty.<\/p>\n<\/div>\n<h3><a id=\"user-content-animation-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#animation-tools\"><\/a>Animation Tools<\/h3>\n<p>Animations guide people through the product friendly and smoothly. Live interactive UI makes users feel delighted with instant feedback and emotional touch. These free and paid tools are designed to make animation creation easier. If you plan to make animated transitions, micro-interactions or scroll-based animations, go through these tools:<\/p>\n<ul><li><a href=\"https:\/\/www.adobe.com\/products\/aftereffects.html\" rel=\"nofollow\">After Effects<\/a> \u2014 a digital visual effects, motion graphics, and compositing application developed by Adobe.<\/li>\n<li><a href=\"https:\/\/beatflyer.com\/\" rel=\"nofollow\">BeatFlyer<\/a> \u2014 a web tool that lets you create very quickly loopy and catchy animations from your multi-layer compositions.<\/li>\n<li><a href=\"https:\/\/www.2dimensions.com\/about-flare\" rel=\"nofollow\">Flare<\/a> \u2014 a powerful design and animation tool, which allows designers and developers to create animations for their app or game. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/createwithflow.com\/\" rel=\"nofollow\">Flow<\/a> \u2014 a professional animation tool for Sketch designs that also exports production ready code for iOS, Web or SVG files. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.haikuforteams.com\/\" rel=\"nofollow\">Haiku Animator<\/a> \u2014 keyframe-based animation connecting your UI tools to code that developers can easily use.<\/li>\n<li><a href=\"https:\/\/www.keyshapeapp.com\/\" rel=\"nofollow\">Keyshape<\/a> \u2014 2D animation tool that allows for animated SVG export.\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/kiteapp.co\/\" rel=\"nofollow\">Kite Compositor<\/a> \u2014 a powerful animation and prototyping application for Mac &amp; iOS. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/uselightbox.com\/\" rel=\"nofollow\">LightBox<\/a> \u2014 2D, Hand-Drawn animation package. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/airbnb.io\/lottie\/\" rel=\"nofollow\">Lottie<\/a> \u2014 a mobile library for Android and iOS that parses Adobe After Effects animations exported as json with Bodymovin and renders them natively on mobile and on the web. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/jeremyckahn.github.io\/mantra\/\" rel=\"nofollow\">Mantra<\/a> \u2014 a web-based timeline animation tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/meteoritic.io\/\" rel=\"nofollow\">Meteoritic<\/a> \u2014 animation delights library for iOS and Android. Choose from a variety of animations that suits different app styles. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/offeo.com\/\" rel=\"nofollow\">OFFEO<\/a> \u2014 an online video maker for small businesses to create video ads like the Big Brands without the big budget. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/sliderapp.co\" rel=\"nofollow\">Slider<\/a> \u2014 create stylish animated posts for Instagram and Facebook in just a few taps, all from your mobile phone. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/spiritapp.io\/\" rel=\"nofollow\">Spirit<\/a> \u2014 the animation tool that helps you to create and manage the animations real time in the browser. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/jeremyckahn.github.io\/stylie\/\" rel=\"nofollow\">Stylie<\/a> \u2014 a web-based graphical CSS3 animation tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/tumult.com\/hype\/\" rel=\"nofollow\">Tumult Hype<\/a> \u2014 the HTML5 animation\/interactives creation app for macOS. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>Principle, Framer, Invision Studio, Flinto are also among UI &amp; UX animation apps and software animation tools.<\/p>\n<\/div>\n<h3><a id=\"user-content-augmented-reality\" class=\"anchor\" aria-hidden=\"true\" href=\"#augmented-reality\"><\/a>Augmented Reality<\/h3>\n<p>Augmented Reality is a technology that upgrades our real world, by adding a layer of digital information to it. The use of AR increases day by day with dozens of new AR apps, development kits, and AR frameworks. So in this section is collected different augmented reality tools for creating, projecting and prototyping apps focused on AR:<\/p>\n<ul><li><a href=\"https:\/\/daqri.com\/\" rel=\"nofollow\">Daqri<\/a> \u2014 delivers professional-grade augmented reality for the workforce by overlaying digital information onto the physical environment.<\/li>\n<li><a href=\"https:\/\/www.easyar.com\/\" rel=\"nofollow\">EasyAR<\/a> \u2014 a mobile application and an Augmented Reality Engine, making AR easier.<\/li>\n<li><a href=\"https:\/\/www.hpreveal.com\/\" rel=\"nofollow\">HP Reveal<\/a> \u2014 adding value to printed content through visual interactivity.<\/li>\n<li><a href=\"https:\/\/www.layar.com\/\" rel=\"nofollow\">Layar<\/a> \u2014 find various items based upon augmented reality technology.<\/li>\n<li><a href=\"https:\/\/lensstudio.snapchat.com\/\" rel=\"nofollow\">Lens Studio<\/a> \u2014 create, publish, and share your own AR experiences for Snapchat with or without code. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/lightform.com\/\" rel=\"nofollow\">Lightform<\/a> \u2014 a design tool for projected AR.<\/li>\n<li><a href=\"https:\/\/www.sparkar.com\" rel=\"nofollow\">Spark AR Studio<\/a> \u2014 \u0441reate AR experiences for Instagram without code. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.torch.app\/\" rel=\"nofollow\">Torch<\/a> \u2014 a cloud-based 3D design and prototyping app focused on mobile AR. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/unity.com\/\" rel=\"nofollow\">Unity<\/a> \u2014 build high-quality 3D and 2D games, deploy them across mobile, desktop and VR\/AR.<\/li>\n<li><a href=\"https:\/\/www.vectary.com\/\" rel=\"nofollow\">Vectary<\/a> \u2014 \u0441reate 3D and AR for your website.<\/li>\n<li><a href=\"https:\/\/www.vuforia.com\/\" rel=\"nofollow\">Vuforia<\/a> \u2014 SDK for mobile devices that enables the creation of augmented reality applications.<\/li>\n<li><a href=\"https:\/\/www.wikitude.com\/\" rel=\"nofollow\">Wikitude<\/a> \u2014 empowers your iOS, Android &amp; Smart Glasses apps with Image &amp; Object Tracking, Instant tracking (SLAM), Geo AR, and more.<\/li>\n<\/ul><h3><a id=\"user-content-collaboration-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#collaboration-tools\"><\/a>Collaboration Tools<\/h3>\n<p>Looking to try some tools for agile design and development collaboration? Good idea because such tools make the lives of designers and developers much easier, save time and improve productivity. Well, you know all that in Slack-driven era. Here you'll find the best collaboration tools for product teams:<\/p>\n<ul><li><a href=\"https:\/\/airtable.com\/\" rel=\"nofollow\">Airtable<\/a> \u2014 part spreadsheet, part database, and entirely flexible, teams use Airtable to organize their work, their way.<\/li>\n<li><a href=\"https:\/\/appear.in\/\" rel=\"nofollow\">Appear.in<\/a> \u2014 a simple video collaboration tool.<\/li>\n<li><a href=\"https:\/\/asana.com\/\" rel=\"nofollow\">Asana<\/a> \u2014 the work management platform teams use to stay focused on the goals, projects, and daily tasks that grow business.<\/li>\n<li><a href=\"https:\/\/basecamp.com\/\" rel=\"nofollow\">Basecamp<\/a> \u2014 the project management suite designed to organize employees, delegate tasks, and monitor progress right from one place.<\/li>\n<li><a href=\"https:\/\/filestage.io\/\" rel=\"nofollow\">Filestage<\/a> \u2014 an agile content approval software that helps you to review videos, designs and documents with clients and co-workers.<\/li>\n<li><a href=\"https:\/\/material.io\/tools\/gallery\/\" rel=\"nofollow\">Gallery<\/a> \u2014 a collaborative tool for uploading design work, getting feedback, and tracking revisions. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.hivedesk.com\/\" rel=\"nofollow\">HiveDesk<\/a> \u2014 an automatic time tracking for remote employees with screenshots. It makes it easier to monitor productivity of remote employees and manage projects.<\/li>\n<li><a href=\"https:\/\/www.atlassian.com\/software\/jira\" rel=\"nofollow\">Jira<\/a> \u2014 software development tool used by agile teams.<\/li>\n<li><a href=\"https:\/\/keybase.io\" rel=\"nofollow\">Keybase<\/a> \u2014 a free collaboration app with built-in end-to-end encrypted chat, file sharing, git repositories, and more <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/mattermost.com\/\" rel=\"nofollow\">Mattermost<\/a> \u2014 a flexible, open source messaging platform that meets even the most demanding privacy and security standards. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/milanote.com\/\" rel=\"nofollow\">Milanote<\/a> \u2014 an easy-to-use, collaborative tool to organize your ideas and projects into visual boards.<\/li>\n<li><a href=\"https:\/\/mixed.io\" rel=\"nofollow\">Mixed<\/a> \u2014 real-time whiteboard and collaboration tools for distributed teams. ![free.svg]<\/li>\n<li><a href=\"https:\/\/mural.co\/\" rel=\"nofollow\">MURAL<\/a> \u2014 think and collaborate visually. Anywhere, anytime.<\/li>\n<li><a href=\"https:\/\/www.notion.so\" rel=\"nofollow\">Notion<\/a> \u2014 write, plan, collaborate, and get organized. Notion is all you need \u2014 in one tool.<\/li>\n<li><a href=\"https:\/\/www.proofhub.com\/\" rel=\"nofollow\">ProofHub<\/a> \u2014 the one place for all your projects, teams and communications.<\/li>\n<li><a href=\"https:\/\/www.realtimeboard.com\/\" rel=\"nofollow\">RealtimeBoard (Miro)<\/a> \u2014 whiteboarding platform for cross-functional team collaboration. It was recently renaimed to Miro.<\/li>\n<li><a href=\"https:\/\/slack.com\/\" rel=\"nofollow\">Slack<\/a> \u2014 a collaboration hub for work, no matter what work you do. It\u2019s a place where conversations happen, decisions are made, and information is always at your fingertips.<\/li>\n<li><a href=\"https:\/\/sunsama.com\/\" rel=\"nofollow\">Sunsama<\/a> \u2014 a beautifully designed, team oriented task manager that consolidates Trello\/Jira\/Asana tickets into a single, calendar-oriented view.<\/li>\n<li><a href=\"https:\/\/taskade.com\/\" rel=\"nofollow\">Taskade<\/a> \u2014 team productivity made simple, fun and designed for remote teams. Taskade is collaborative and syncs in real-time across web, mobile and desktop.<\/li>\n<li><a href=\"https:\/\/trello.com\/\" rel=\"nofollow\">Trello<\/a> \u2014 a web-based project management application that enables you to organize and prioritize your projects in a fun, flexible and rewarding way.<\/li>\n<li><a href=\"https:\/\/www.witeboard.com\/\" rel=\"nofollow\">Witeboard<\/a> \u2014 simple real-time whiteboard for collaboration work. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.workzone.com\/\" rel=\"nofollow\">Workzone<\/a> \u2014 a simple, powerful project management software.<\/li>\n<li><a href=\"https:\/\/www.wrike.com\/\" rel=\"nofollow\">Wrike<\/a> \u2014 an online project management software that gives you full visibility and control over your tasks.<\/li>\n<li><a href=\"https:\/\/zenkit.com\/\" rel=\"nofollow\">Zenkit<\/a> \u2014 a tool to schedule meetings, track project\u2019s progress and brainstorm new ideas.<\/li>\n<\/ul><div>\n<p>You can also have design collaboration into <a href=\"https:\/\/www.invisionapp.com\/feature\/freehand\" rel=\"nofollow\">InVision Freehand<\/a> tool.<\/p>\n<\/div>\n<h3><a id=\"user-content-color-picker-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#color-picker-tools\"><\/a>Color Picker Tools<\/h3>\n<p>If you are looking for an eyedropper tool, color identifier or color capture, check this section. With the color pickers mentioned here, you will be able to create new color combinations and define really great ones. Almost all of these tools are free:<\/p>\n<ul><li><a href=\"https:\/\/chroma.spencerhamm.com\/\" rel=\"nofollow\">Chroma<\/a> \u2014 free web app which allows you to identify the color. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/color.adobe.com\/explore\/?filter=most-popular&amp;time=month\" rel=\"nofollow\">Color<\/a> \u2014 create color schemes with the color wheel or browse thousands of color combinations from the Color community. Made by Adobe. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/colourco.de\/\" rel=\"nofollow\">Colourcode<\/a> \u2014 online designer tool, which allows you to easily and intuitively combine colours. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/color.obscuredetour.com\/\" rel=\"nofollow\">Color Deck<\/a> \u2014 a HSL driven color palette collection. It's also open-source. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/obscuredetour\/color-deck\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.colorhexa.com\/\" rel=\"nofollow\">Color Hexa<\/a> \u2014 free color tool providing information about any color and generating matching color palettes for designs.<\/li>\n<li><a href=\"https:\/\/colorpicker.crea-th.at\/\" rel=\"nofollow\">Colorpicker<\/a> \u2014 a complete and open-source colors manipulation tool with picking. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/cloudflare.design\/color\/\" rel=\"nofollow\">Color by Cloudflare<\/a> \u2014 a tool to preview palettes against UI elements, cycle through accessible color combos and reate palettes manually, via URL import, or generatively.<\/li>\n<li><a href=\"https:\/\/www.instagram.com\/colours.cafe\/\" rel=\"nofollow\">Colour Cafe<\/a> \u2014 colours inspiration that have selected modern colour palette in Instagram. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/coolors.co\/\" rel=\"nofollow\">Coolors<\/a> \u2014 the super-fast color scheme generator. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/culrs.com\/#\/\" rel=\"nofollow\">Culrs<\/a> \u2014 thoughtfully crafted and easy-to-use color palettes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/geenes.app\" rel=\"nofollow\">Geenes<\/a> \u2014 a tool to create harmonious color themes with code and sketch file export.<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/instant-eyedropper.com\/\" rel=\"nofollow\">Instant Eyedropper<\/a> \u2014 a minimalist eyedropper tool for windows, supports multiple color formats for both designers and developers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/annystudio.com\/software\/colorpicker\/\" rel=\"nofollow\">Just Color Picker<\/a> \u2014 Free portable offline colour picker and colour editor. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/khroma.co\/\" rel=\"nofollow\">Khroma<\/a> \u2014 AI-based tool to generate color palettes based on your preferences.<\/li>\n<li><a href=\"https:\/\/github.com\/BafS\/Material-Colors-native\">Material Colors Native<\/a> \u2014 Google's material design color palette in the form of a mac app. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/paletton.com\" rel=\"nofollow\">Paletton<\/a> \u2014 a designer tool for creating color combinations that work together well. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/picular.co\/\" rel=\"nofollow\">Picular<\/a> \u2014 a rocket fast primary color generator using Google's image search. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/itunes.apple.com\/us\/app\/pikka-color-picker\/id1195076754\" rel=\"nofollow\">Pikka<\/a> \u2014 color picker &amp; color schemes generator for macOS. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/casesandberg.github.io\/react-color\/\" rel=\"nofollow\">React Color<\/a> \u2014 a collection of color pickers from Sketch, Photoshop, Chrome.  It\u2018s free and open-source. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/casesandberg\/react-color\/\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/sipapp.io\/\" rel=\"nofollow\">Sip<\/a> \u2014 a better way to collect, organize and share colors. Global eyedropper tool for MacOS. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/bjango.com\/mac\/skalacolor\/\" rel=\"nofollow\">Skala Color<\/a> \u2014 works with a huge variety of formats, covering everything you\u2019re likely to need for web, iOS, Android, and macOS development. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/swatchesapp.io\/\" rel=\"nofollow\">Swatches<\/a> \u2014 iOS app that lets you collect, inspect and share the colors that inspire you in your daily life. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-design-feedback-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-feedback-tools\"><\/a>Design Feedback Tools<\/h3>\n<p>How do you provide and get feedback during the development process? This process is usually pretty messy for many product teams, which caused product delays and time waste on back-and-forth communication. So if you wish to get fast and structural feedback on UI issues or visual bugs \u2014 take a look at this tool:<\/p>\n<ul><li><a href=\"https:\/\/flawlessapp.io\/feedback\" rel=\"nofollow\">Flawless Feedback<\/a> \u2014 review and annotate iOS apps then share your feedback in Jira or Trello. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-design-handoff-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-handoff-tools\"><\/a>Design Handoff Tools<\/h3>\n<p>Design handoff happens when designers finish the work and need to deliver designs to developers with all specs and assets. Design handoff tools allow to automatically generate styleguide, comment on the design, inspect elements. These tools minimize guesswork and improve the effectiveness of the design process.<\/p>\n<ul><li><a href=\"https:\/\/avocode.com\" rel=\"nofollow\">Avocode<\/a> \u2014 open designs without design tools. Export images without preparing layers. Click on layers to get code.<\/li>\n<li><a href=\"https:\/\/flawlessapp.io\/\" rel=\"nofollow\">Flawless App<\/a> \u2014 a tool to compare the expected design and the real app right inside iOS Simulator (useful for designers who use Xcode or mobile developers). <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.invisionapp.com\/feature\/inspect\/\" rel=\"nofollow\">Inspect<\/a> \u2014 prepare designs for development. Made by Invision.<\/li>\n<li><a href=\"https:\/\/github.com\/utom\/sketch-measure\">Sketch Measure<\/a> \u2014 Sketch redline plugin. Annotate distance and size of elements. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/specctr.com\" rel=\"nofollow\">Specctr<\/a> \u2014 in the unfortunate event you're designing UI in PS, AI, or ID this tool creates redline annotations.<\/li>\n<li><a href=\"https:\/\/sympli.io\" rel=\"nofollow\">Sympli<\/a> \u2014 automated specs and assets handoff from Sketch, Photoshop and Adobe XD. Integrated with Jira, Xcode and Android Studio.<\/li>\n<li><a href=\"https:\/\/zeplin.io\/\" rel=\"nofollow\">Zeplin<\/a> \u2014 handoff designs and styleguides with accurate specs, assets, code snippets automatically.<\/li>\n<\/ul><h3><a id=\"user-content-design-inspiration\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-inspiration\"><\/a>Design Inspiration<\/h3>\n<p>The creative process can be tough. So if ideas don\u2019t come to your mind right away, try to search for inspiration in the work of other designers. These design inspiration sites feature design patterns, user flows, email markups and creative solutions of popular companies and great products:<\/p>\n<ul><li><a href=\"https:\/\/codemyui.com\/\" rel=\"nofollow\">CodeMyUI<\/a> \u2014 handpicked collection of web design &amp; UI inspiration with code snippets. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/collectui.com\/\" rel=\"nofollow\">Collect UI<\/a> \u2014 a platform for your daily inspiration collected from daily ui &amp; beyond. Based on Dribbble shots, 14419 hand picked designs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/www.creative-portfolios.com\/\" rel=\"nofollow\">Creative Portfolios<\/a> \u2014 a curation of the most creative portfolios made by designers &amp; developers.<\/li>\n<li><a href=\"https:\/\/dribbble.com\/\" rel=\"nofollow\">Dribbble<\/a> \u2014 an online community for showcasing user-made artwork. Great resource to get inspired from others work.<\/li>\n<li><a href=\"http:\/\/inspirationui.com\/\" rel=\"nofollow\">Inspiration UI<\/a> \u2014 community that aims to provide the best design resources for those who create for the web. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/hoverstat.es\/\" rel=\"nofollow\">Hover States<\/a> \u2014 showcase great work from the bleeding edge of digital culture for, and with, the creative community.<\/li>\n<li><a href=\"https:\/\/httpster.net\/\" rel=\"nofollow\">Httpster<\/a> \u2014 an inspiration resource showcasing totally rocking websites made by people from all over the world. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.h69.design\/\" rel=\"nofollow\">H69.Design<\/a> \u2014 landing page colletions &amp; Free resources for designers.<\/li>\n<li><a href=\"https:\/\/www.lapa.ninja\/\" rel=\"nofollow\">Lapa Ninja<\/a> \u2014 landing page design inspiration from around the web. 1800+ landing page examples and updated daily.<\/li>\n<li><a href=\"https:\/\/www.microcopyinspirations.com\" rel=\"nofollow\">Microcopy Inspirations<\/a> \u2014 curated examples of good microcopy on websites and apps.<\/li>\n<li><a href=\"https:\/\/www.mobile-patterns.com\/\" rel=\"nofollow\">Mobile Patterns<\/a> \u2014 a design inspirational library featuring finest UI UX Patterns (iOS and Android).<\/li>\n<li><a href=\"https:\/\/mobbin.design\/\" rel=\"nofollow\">Mobbin<\/a> \u2014 browse mobile design patterns for popular apps. Explore common user flows such as onboarding, account sign up and log in, core functionality, and more. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/owwly.com\/feed\/inspirations\" rel=\"nofollow\">Owwly<\/a> \u2014 home for digital products crafted with passion to design.<\/li>\n<li><a href=\"https:\/\/pageflows.com\/\" rel=\"nofollow\">Page Flows<\/a> \u2014 user flow videos and screenshots to inspire you when you're stuck. Discover mobile and web design patterns for over 160 different tasks. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pttrns.com\/\" rel=\"nofollow\">pttrns<\/a> \u2014 design patterns for popular mobile apps.<\/li>\n<li><a href=\"https:\/\/reallygoodemails.com\/\" rel=\"nofollow\">Really Good Emails<\/a> \u2014 4,150+ handpicked email design. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.reallygoodux.io\/\" rel=\"nofollow\">ReallyGoodUX<\/a> \u2014 screenshots and examples of great UX from real mobile and web products. Discover the best UX examples\u2014including onboarding tours and walkthroughs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/thedesignteam.io\/\" rel=\"nofollow\">The Design Team<\/a> \u2014 comics about a design team for a tech startup in Silicon Valley.<\/li>\n<li><a href=\"https:\/\/www.typewolf.com\/\" rel=\"nofollow\">Typewolf<\/a> \u2014 helps designers choose the perfect font combination for their next design project.<\/li>\n<li><a href=\"https:\/\/uigarage.net\/\" rel=\"nofollow\">UI Garage<\/a> \u2014 the one stop shop for designers, developers and marketer to find inspiration, tools and the best resources.<\/li>\n<li><a href=\"https:\/\/www.uisources.com\/\" rel=\"nofollow\">UI Sources<\/a> \u2014 over 500+ interactions from the best designed and top grossing apps on the App Store today.  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/uirecipes.com\/\" rel=\"nofollow\">UI Recipes<\/a> \u2014 weekly 15 min lessons on UI design from the hottest apps.<\/li>\n<li><a href=\"http:\/\/uxarchive.com\/\" rel=\"nofollow\">UX Archive<\/a> \u2014 browse more than 400 user flows across 120 mobile apps. Examine tasks such as booking, logging in, onboarding, purchasing, seaching, and more. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.webdesignmuseum.org\/\" rel=\"nofollow\">Web Design Museum<\/a> \u2014 over 1,200 carefully selected web sites that show web design trends between the years 1994 and 2006.<\/li>\n<\/ul><h3><a id=\"user-content-design-system-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-system-tools\"><\/a>Design System Tools<\/h3>\n<p>I bet you heard about Design Systems, as it\u2019s a pretty trendy topic. Design systems provide consistent, robust design patterns to keep design and development in sync. They are essentially collections of rules, constraints, and principles, implemented in design and code. Here you can find tools to build, maintain and organize your own design system.<\/p>\n<ul><li><a href=\"https:\/\/cabanadesignsystem.com\/\" rel=\"nofollow\">Cabana<\/a> \u2014 a Premium Design System for Sketch that helps you create amazing products faster than ever before.<\/li>\n<li><a href=\"https:\/\/www.invisionapp.com\/design-system-manager\/\" rel=\"nofollow\">Design System Manager<\/a> \u2014 Invision's design system manager.<\/li>\n<li><a href=\"https:\/\/www.eosdesignsystem.com\" rel=\"nofollow\">EOS Design System<\/a> \u2014 an open source and customizable built on top of Bootstrap following the Atomic Design concept. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/frontify.com\/\" rel=\"nofollow\">Frontify<\/a> \u2014 create graphical guidelines, patterns libraries, design systems.<\/li>\n<li><a href=\"https:\/\/www.lingoapp.com\/\" rel=\"nofollow\">Lingo<\/a> \u2014 create a shared asset library with your entire team. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/lucid.style\/\" rel=\"nofollow\">Lucid<\/a> \u2014 tool for creating, managing and sharing a design system. From a simple component library through to detailed descriptions of your styles.<\/li>\n<li><a href=\"https:\/\/www.primedesignsystem.com\" rel=\"nofollow\">Prime Design System Kit<\/a> \u2014 it includes device templates, charts, UI kit and even illustration kit. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.specifyapp.com\/\" rel=\"nofollow\">Specify<\/a> \u2014 a tool to create, scale and maintain a design system.<\/li>\n<li><a href=\"https:\/\/symboldesign.co\/\" rel=\"nofollow\">Symbol Design System<\/a> \u2014 design System for Sketch based on atomic elements.<\/li>\n<li><a href=\"https:\/\/www.zeroheight.com\/\" rel=\"nofollow\">Zeroheight<\/a> \u2014 styleguides created by designers, extended by developers, and editable by everyone.<\/li>\n<\/ul><div>\n<p>We can also add to this in Sketch, Figma, UXPin and Framer X (<a href=\"https:\/\/www.framer.com\/design-systems\/\" rel=\"nofollow\">Framer X Team Store<\/a>).<\/p>\n<\/div>\n<h3><a id=\"user-content-design-to-code-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-to-code-tools\"><\/a>Design to Code Tools<\/h3>\n<p>Everyone can learn development but it takes time and effort. If you need a website or an app right now and you don\u2019t want to hire a developer, pay attention to the website builders. Such design to code tools will help you to make a portfolio, simple landing or an app pretty fast and beautiful.<\/p>\n<ul><li><a href=\"https:\/\/blocsapp.com\/\" rel=\"nofollow\">Blocs<\/a> \u2014 a fast, easy to use and powerful visual web design tool, that lets you create responsive websites without writing code. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/bootstrapstudio.io\/\" rel=\"nofollow\">Bootstrap Studio<\/a> \u2014 a powerful web design tool for creating responsive websites using the Bootstrap framework.<\/li>\n<li><a href=\"https:\/\/grapedrop.com\/\" rel=\"nofollow\">Grapedrop<\/a> \u2014 design your components, web projects and publish them instantly online, with an easy to use editor.<\/li>\n<li><a href=\"https:\/\/www.trymason.com\/core\" rel=\"nofollow\">Mason<\/a> \u2014 is a feature-building platform where modern product teams design, build, and deploy front-end features together.<\/li>\n<li><a href=\"https:\/\/mobirise.com\/\" rel=\"nofollow\">Mobirise<\/a> \u2014 an offline drag and drop website builder software that is based on Bootstrap 3\/4 and \\ AMP. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pinegrow.com\/\" rel=\"nofollow\">Pinegrow<\/a> \u2014 a professional visual editor for CSS Grid, Bootstrap 4 and 3, Foundation, responsive design, HTML, and CSS.<\/li>\n<li><a href=\"https:\/\/www.pxdiv.com\" rel=\"nofollow\">px.div<\/a> \u2014 Proper site build tool for developers and designers.<\/li>\n<li><a href=\"https:\/\/readymag.com\/\" rel=\"nofollow\">Readymag<\/a> \u2014 a visually-pleasing tool for designing on the web from landing pages to multimedia long-reads, presentations and portfolios.<\/li>\n<li><a href=\"https:\/\/sparkleapp.com\/\" rel=\"nofollow\">Sparkle<\/a> \u2014 the easiest way to make a real website, no code, no jargon. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/studio.design\/\" rel=\"nofollow\">STUDIO<\/a> \u2014 design from scratch, collaborate in real time and publish websites.<\/li>\n<li><a href=\"https:\/\/tilda.cc\/\" rel=\"nofollow\">Tilda<\/a> \u2014 create a website, landing page or online store for free with the help of Tilda modules and publish it on the same day.<\/li>\n<li><a href=\"https:\/\/www.wix.com\/\" rel=\"nofollow\">Wix<\/a> \u2014 the easiest and fullest-featured website builder, that allows you to create your own highly customized site.<\/li>\n<li><a href=\"https:\/\/webflow.com\/\" rel=\"nofollow\">Webflow<\/a> \u2014 build responsive websites in your browser, then host with us or export your code to host wherever.<\/li>\n<\/ul><h3><a id=\"user-content-design-version-control\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-version-control\"><\/a>Design Version Control<\/h3>\n<p>Developers actively use version control tools for a long time, probably since 2005 (Git first release). Using a version control system is no brainer for dev teams, while design version control system appeared only recently. This market is rapidly developing and we expect to see even more in version control for designers:<\/p>\n<ul><li><a href=\"https:\/\/www.abstract.com\/\" rel=\"nofollow\">Abstract<\/a> \u2014 a platform for design teams to version, manage and collaborate on Sketch files.<\/li>\n<li><a href=\"http:\/\/folioformac.com\/\" rel=\"nofollow\">Folio<\/a> \u2014 a simple version control system for design teams based on Git. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/kactus.io\/\" rel=\"nofollow\">Kactus<\/a> \u2014 design version control without changing your tools. Manage changes, document work and keep your team in sync. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/plantapp.io\/\" rel=\"nofollow\">Plant<\/a> \u2014 Mac app and Sketch plugin offering complete version control for designers and teams. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/relayforfigma.com\/\" rel=\"nofollow\">Relay for Figma<\/a> \u2014 Relay pushes your latest assets from Figma directly to your codebase (not released yet).<\/li>\n<li><a href=\"https:\/\/versions.sympli.io\" rel=\"nofollow\">Versions<\/a> \u2014 a version control tool for designers with visual diff, merge and conflict resolution. Works with GitHub, Bitbucket, GitLab and Azure Devops.<\/li>\n<\/ul><h3><a id=\"user-content-experience-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#experience-monitoring\"><\/a>Experience Monitoring<\/h3>\n<p>Listening to users is important but seeing the real usage is even more crucial. For these, you need to install different analytic tools, experience monitoring software, and user behavior apps. Just use those analytics solutions with respect to users data:<\/p>\n<ul><li><a href=\"https:\/\/amplitude.com\/\" rel=\"nofollow\">Amplitude<\/a> \u2014 understand your users, rapidly release better product experiences, and grow your business.<\/li>\n<li><a href=\"https:\/\/www.optimalworkshop.com\/chalkmark\" rel=\"nofollow\">Chalkmark<\/a> \u2014 increase your customer conversion and task oriented success rates by getting quick feedback on designs before you implemen.<\/li>\n<li><a href=\"https:\/\/usefathom.com\/\" rel=\"nofollow\">Fathom<\/a> \u2014 provides simple, useful websites stats without tracking or storing personal data of your users.<\/li>\n<li><a href=\"https:\/\/www.fullstory.com\/\" rel=\"nofollow\">FullStory<\/a> \u2014 an app that captures all your customer experience data in one powerful, easy-to-use platform.<\/li>\n<li><a href=\"https:\/\/analytics.google.com\/analytics\/web\/\" rel=\"nofollow\">Google Analytics<\/a> \u2014 measure your advertising ROI and track your sites and applications. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/heapanalytics.com\/\" rel=\"nofollow\">Heap<\/a> \u2014 automatically capture every web, mobile, and cloud interaction then retroactively analyze your data without writing code.<\/li>\n<li><a href=\"https:\/\/www.hotjar.com\/\" rel=\"nofollow\">Hotjar<\/a> \u2014 see how visitors are really using your website and collect user feedback.<\/li>\n<li><a href=\"https:\/\/www.inspectlet.com\/\" rel=\"nofollow\">Inspectlet<\/a> \u2014 record videos of your visitors as they use your site, so you can see everything they do.<\/li>\n<li><a href=\"https:\/\/www.logrocket.com\/\" rel=\"nofollow\">LogRocket<\/a> \u2014 see what users do on your site, helping you reproduce bugs and fix issues faster.<\/li>\n<li><a href=\"https:\/\/mixpanel.com\/\" rel=\"nofollow\">Mixpanel<\/a> \u2014 get insights across all of your user-centric data to make smarter decisions and act faster based on how customers use your product or website.<\/li>\n<li><a href=\"https:\/\/mouseflow.com\/\" rel=\"nofollow\">Mouseflow<\/a> \u2014 record mouse cursor movements.<\/li>\n<li><a href=\"https:\/\/www.pendo.io\/\" rel=\"nofollow\">Pendo<\/a> \u2014 measure and elevate the customer experience within their applications.<\/li>\n<\/ul><h3><a id=\"user-content-font-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#font-tools\"><\/a>Font Tools<\/h3>\n<p>Fonts are commonly used for making the web a more beautiful place. It\u2019s an essential part of any design. In this section, you\u2019ll find fonts generators &amp; font finder tools that allow you to manage and work with fonts:<\/p>\n<ul><li><a href=\"https:\/\/fontba.se\/\" rel=\"nofollow\">Fontbase<\/a> \u2014 font management made easy. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/fontpair.co\/\" rel=\"nofollow\">FontPair<\/a> \u2014 a simple tool that helps you pair Google Fonts together. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/fontface.ninja\/\" rel=\"nofollow\">Fontface Ninja<\/a> \u2014 browser extension to discover what fonts are being used on any website.<\/li>\n<li><a href=\"https:\/\/fonts.adobe.com\/fonts\" rel=\"nofollow\">Fonts<\/a> \u2014 thousands of beautiful fonts by Adobe. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/fontspark.app\/\" rel=\"nofollow\">FontSpark<\/a> \u2014 a simple tool to help designers quickly find the best fonts for their projects. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.fontsquirrel.com\/fonts\/acherus-grotesque\" rel=\"nofollow\">Font Squirrel<\/a> \u2014 download free fonts with wide collections.<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/fonts.google.com\/\" rel=\"nofollow\">Google Fonts<\/a> \u2014 making the web more beautiful, fast, and open through great typography. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/google-webfonts-helper.herokuapp.com\/\" rel=\"nofollow\">Google Webfonts Helper<\/a> \u2014 a hassle way to self-host Google Fonts. It\u2018s free and open-source. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/majodev\/google-webfonts-helper\/\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/rightfontapp.com\/\" rel=\"nofollow\">RightFont<\/a> \u2014 font managing app, helps preview, sync, install and organize fonts over iCloud, Dropbox or Google Drive. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/sansforgetica.rmit\/\" rel=\"nofollow\">Sans Forgetica<\/a> \u2014 a downloadable font that is scientifically designed to help you remember your study notes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/typefaceapp.com\/\" rel=\"nofollow\">Typeface<\/a> \u2014 font manager that improves your design workflow with live font previews and flexible tagging. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/wakamaifondue.com\/\" rel=\"nofollow\">Wakamai Fondue<\/a> \u2014 the tool that answers the question \"what can my font do?\". <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/webfontapp.com\/\" rel=\"nofollow\">Webfont<\/a> \u2014 create and maintain custom SVG icon fonts, made a shared library of icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>You can also handle fonts with Specify, which is mentioned in the <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools#design-system-tools\">Design System Tools<\/a> section.<\/p>\n<\/div>\n<h3><a id=\"user-content-gradient-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#gradient-tools\"><\/a>Gradient Tools<\/h3>\n<p>You can see gradient colors are everywhere \u2014 UI, branding, illustration, typography. A gradient is created by using two or more different colors to paint one element while gradually fading between them. It might look as memorable, fresh and unique color. To make such gradient for your design, use these gradient color palettes.<\/p>\n<ul><li><a href=\"http:\/\/www.colinkeany.com\/blend\/\" rel=\"nofollow\">Blend<\/a> \u2014 create and customize beautiful CSS3 gradients. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.colorffy.com\/\" rel=\"nofollow\">Colorffy<\/a> \u2014 get css, images, ui previews or create your own gradients. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/mycolor.space\/\" rel=\"nofollow\">ColorSpace<\/a> \u2014 the perfect matching color scheme for your next project! Generate nice color palettes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/colorspark.app\/gradient\" rel=\"nofollow\">ColorSpark<\/a> \u2014 helps designers find unique colors and striking gradient combinations. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/coolbackgrounds.io\/\" rel=\"nofollow\">Cool Backgrounds<\/a> \u2014 explore a beautifully curated selection of cool backgrounds that you can add to blogs, websites.<\/li>\n<li><a href=\"https:\/\/webkul.github.io\/coolhue\/\" rel=\"nofollow\">CoolHue<\/a> \u2014 handpicked gradient palette and swatch for your next super amazing stuff. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/cssgradient.io\/\" rel=\"nofollow\">CSS Gradient<\/a> \u2014 a free css gradient generator tool, this website lets you create a colorful gradient background. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.eggradients.com\/\" rel=\"nofollow\">Egg Gradients<\/a> \u2014 a color collection page which includes shades of 12 different colors. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.grabient.com\/\" rel=\"nofollow\">Grabient<\/a> \u2014 beautiful and simple UI for generating web gradients. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/cssgradient.io\/gradient-backgrounds\/\" rel=\"nofollow\">Gradient Backgrounds<\/a> \u2014 discover the best gradient backgrounds from a curated collection of the ultimate list of gradient sites. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/gradienthunt.com\/\" rel=\"nofollow\">Gradient Hunt<\/a> \u2014 a free and open platform for colour inspiration with thousands of trendy hand-made colour gradients. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/gradientsguru.com\/\" rel=\"nofollow\">Gradients Guru<\/a> \u2014 blend mode background images &amp; beautiful background gradients for your UI. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/htmlcolors.com\/\" rel=\"nofollow\">HtmlColours<\/a> \u2014 realize how easy and beautiful is to create gradients with this gradients generator. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/uigradients.com\/\" rel=\"nofollow\">UI Gradients<\/a> \u2014 handpicked collection of beautiful color gradients for designers and developers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/webgradients.com\/\" rel=\"nofollow\">Web Gradients<\/a> \u2014 a free collection of 180 linear gradients that you can use as content backdrops in any part of your website. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-icons-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#icons-tools\"><\/a>Icons Tools<\/h3>\n<p>As well as fonts, icons are used in every design. These basic elements support and guide many user actions inside the product. Without a doubt, icons are a vital element in user navigation. While making those small design elements is hard and time-consuming, you can get thousands of vector icons for personal and commercial use.<\/p>\n<ul><li><a href=\"http:\/\/animaticons.co\/\" rel=\"nofollow\">Animaticons<\/a> \u2014 animaticons are a growing set of beautiful, high-resolution, animated GIF icons that you can customize.<\/li>\n<li><a href=\"http:\/\/digitalnomadicons.com\/\" rel=\"nofollow\">Digital Nomad Icons<\/a> \u2014 lifestyle icon &amp; emoji pack for your next project. 25 high resolution flat icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/interface.market\/essential\/basic-pack\" rel=\"nofollow\">Essential Glyphs<\/a> \u2014 created to cover your needs in perfect-shaped icons. Adapted for small and large sizes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/feathericons.com\/\" rel=\"nofollow\">Feather Icons<\/a> \u2014 each icon is designed on a 24x24 grid with an emphasis on simplicity, consistency and readability. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.flaticon.com\/\" rel=\"nofollow\">Flaticon<\/a> \u2014 1593000+ vector icons in SVG, PSD, PNG, EPS format or as icon font.<\/li>\n<li><a href=\"https:\/\/fontawesome.com\/\" rel=\"nofollow\">Font Awesome<\/a> \u2014 the web's most popular icon set and toolkit, also it's open source. <a href=\"https:\/\/github.com\/FortAwesome\/Font-Awesome\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/fontello.com\/\" rel=\"nofollow\">Fontello<\/a> \u2014 tool to build custom fonts with icons, also open source. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/fontello\/fontello\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/iconscout.com\/\" rel=\"nofollow\">Iconscout<\/a> \u2014 get high-quality Icons, Illustrations and Stock photos at one place.<\/li>\n<li><a href=\"https:\/\/www.iconfinder.com\" rel=\"nofollow\">Iconfinder<\/a> \u2014 a marketplace for vector icons (SVG). Icon sets available in IconJar format.<\/li>\n<li><a href=\"https:\/\/www.geticonjar.com\/\" rel=\"nofollow\">IconJar<\/a> \u2014 store all your icons in one icon manager. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/iconmonstr.com\/\" rel=\"nofollow\">Iconmonstr<\/a> \u2014 discover 4412+ free icons in 305 collections.Big and continuously growing source of simple icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/iconstore.co\/\" rel=\"nofollow\">Icon Store<\/a> \u2014 a library of free vector icons for personal and commercial projects, designed by first-class designers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/iconsvg.xyz\/\" rel=\"nofollow\">ICONSVG<\/a> \u2014 a tool to simplify the process of finding and generating common icons for your project. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/icons8.com\/icons\" rel=\"nofollow\">Icons8<\/a> \u2014 free iOS, Android and Windows styled icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.ikonate.com\/\" rel=\"nofollow\">Ikonate<\/a> \u2014 customise, adjust and download free vector icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/illustrio.com\/\" rel=\"nofollow\">illustrio<\/a> \u2014 a smarter icon library. Build something great with 100% customizable icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/ionicons.com\/\" rel=\"nofollow\">Ionicons<\/a> \u2014 beautifully crafted open source icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/materialdesignicons.com\/\" rel=\"nofollow\">Material Design Icons<\/a> \u2014 free material design icons made possible by open source contributons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/icons.pixsellz.io\/\" rel=\"nofollow\">Material Icons Library<\/a> \u2014 free collection of 1000+ icons for popular graphics tools. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/thenounproject.com\/\" rel=\"nofollow\">Noun Project<\/a> \u2014 icons for everything. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/nucleoapp.com\/\" rel=\"nofollow\">Nucleo<\/a> \u2014 a Mac and Windows app to collect, customize and export all your icons. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/simpleicons.org\/\" rel=\"nofollow\">Simple Icons<\/a> \u2014 free SVG icons for popular brands. Simply click the icon you want, and the download should start automatically. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/emoji.streamlineicons.com\/\" rel=\"nofollow\">Streamline Emoji<\/a> \u2014 a free collection of 850 + vector emoji. The style is inspired by the japanese \u201cKawaii\u201d (cute) style. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/www.svgs.us\/\" rel=\"nofollow\">Svgsus<\/a> \u2014 SVG icon management tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/tilda.cc\/free-icons\/\" rel=\"nofollow\">Tilda Icons<\/a> \u2014 download free icons for landing pages. More than 700 vector icons, collected in 43 sets for business. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/iconscout.com\/unicons\" rel=\"nofollow\">Unicons<\/a> \u2014 1000+ pixel-perfect vector icons and Iconfont for your next project. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/visualpharm.com\/\" rel=\"nofollow\">VisualPharm<\/a> \u2014 free SVG Icons with super-fast search and free Coke. Built for fun by Icons8. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.xicons.co\/\" rel=\"nofollow\">Xicons<\/a> \u2014 free (do whatever you want) vector icons. A new icon pack every 10 days. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.zwicon.com\/\" rel=\"nofollow\">Zwicon<\/a> \u2014 handcrafted iconset for your next project. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>You can also handle icons with Specify, which is mentioned in the <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools#design-system-tools\">Design System Tools<\/a> section.<\/p>\n<\/div>\n<h3><a id=\"user-content-illustrations\" class=\"anchor\" aria-hidden=\"true\" href=\"#illustrations\"><\/a>Illustrations<\/h3>\n<p>Illustrations can be used on your landing page, blog posts, inside your app or email campaign. They make your design live and playful. While drawing good illustration is a task of a skilled graphic designer, you can grab free SVG images &amp; illustrations from very kind people in our community:<\/p>\n<ul><li><a href=\"https:\/\/absurd.design\" rel=\"nofollow\">Absurd Design<\/a> \u2014 free surrealist illustrations for your projects. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.drawkit.io\/\" rel=\"nofollow\">DrawKit<\/a> \u2014 a collection of fully customisable SVG illustrations, in two styles, available to use on your next project, app, or website. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/lukaszadam.com\/illustrations\" rel=\"nofollow\">LukaszAdam<\/a> \u2014  free vector art illustrations and icons. They are available for personal and commercial use. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/icons8.com\/ouch\" rel=\"nofollow\">Ouch<\/a> \u2014 vector illustrations to class up your project. Free for both personal and commercial use. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/undraw.co\" rel=\"nofollow\">unDraw<\/a> \u2014 collection of beautiful SVG images. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-information-architecture\" class=\"anchor\" aria-hidden=\"true\" href=\"#information-architecture\"><\/a>Information Architecture<\/h3>\n<p>Information architecture helps designers organize and structure content inside websites, mobile apps, and other software. So users will understand product functionality and will find everything needed. These information architecture tools should allow you to create visual sitemaps and to improve your website content structure:<\/p>\n<ul><li><a href=\"https:\/\/dynomapper.com\/\" rel=\"nofollow\">DYNO Mapper<\/a> \u2014 organize website projects using visual sitemaps, content inventory, content audit, content planning, daily keyword tracking, and website accessibility testing.<\/li>\n<li><a href=\"https:\/\/www.omnigroup.com\/omnigraffle\/\" rel=\"nofollow\">OmniGraffle<\/a> \u2014 \u0441reate beautiful diagrams and designs with this powerful and easy to use app.<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.optimalworkshop.com\/optimalsort\" rel=\"nofollow\">OptimalSort<\/a> \u2014 card sorting tool that allows you to understand how people categorize content in your UI.<\/li>\n<li><a href=\"https:\/\/www.optimalworkshop.com\/treejack\" rel=\"nofollow\">Treejack<\/a> \u2014 upload your proposed sitemap to see how a user would move through your site.<\/li>\n<\/ul><h3><a id=\"user-content-logo-design\" class=\"anchor\" aria-hidden=\"true\" href=\"#logo-design\"><\/a>Logo Design<\/h3>\n<p>A logo is the starting point of your brand identity. It reflects the product mission, functionality and brand message. Ideally, the logo creates a strong connection between your product and the users. Logo design is an art, as well as many other design disciplines. With the right logo design tools, this art can be done right a bit faster.<\/p>\n<ul><li><a href=\"https:\/\/www.designevo.com\/\" rel=\"nofollow\">Design Evo<\/a> \u2014 a large collection of vector icons and shapes help you design custom logos easily. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.freelogodesign.org\/\" rel=\"nofollow\">Free Logo Design<\/a> \u2014 create stunning professional logos in seconds. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.graphicsprings.com\/\" rel=\"nofollow\">Graphic Springs<\/a> \u2014 choose from a gallery of vector format logo templates, categorized by industry. Edit the graphics and texts of your template, then download.<\/li>\n<li><a href=\"https:\/\/www.logodesignteam.com\/\" rel=\"nofollow\">Logo Design Team<\/a> \u2014 offers affordable professional custom logo design and brand design services.<\/li>\n<li><a href=\"https:\/\/logojoy.com\/\" rel=\"nofollow\">Logojoy<\/a> \u2014 uses artificial intelligence to instantly generate unique logo ideas for your business.<\/li>\n<li><a href=\"https:\/\/logolab.app\/home\" rel=\"nofollow\">Logo Lab<\/a> \u2014 put your logo to the test with automated visual experiments. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.ucraft.com\/free-logo-maker\" rel=\"nofollow\">Logo Maker<\/a> \u2014 allows you to make a free logo within seconds. Start with a blank canvas, add text and an icon, and then customize. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.designhill.com\/tools\/logo-maker\" rel=\"nofollow\">Logo Maker<\/a> by Designhill \u2014 generate 1000s of unique logos in only few minutes, without design expertise.<\/li>\n<li><a href=\"https:\/\/logomakr.com\/\" rel=\"nofollow\">Logo Makr<\/a> \u2014 design a logo for free, choose from hundreds of fonts and icons, then save a logo on the computer. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.logomyway.com\/\" rel=\"nofollow\">Logo My Way<\/a> \u2014 choose from 600 custom logo designs from creative designers worldwide.<\/li>\n<li><a href=\"http:\/\/thelogopackage.com\" rel=\"nofollow\">Logo Package Express<\/a> \u2014 export and package dozens of logo variations and file types for your clients automatically in under 5 minutes.<\/li>\n<li><a href=\"https:\/\/brandmark.io\/logo-rank\/\" rel=\"nofollow\">Logo Rank<\/a> \u2014 upload your logo to see how it rates on a number of objective measures. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>You can also do your logo design with Adobe Photoshop, GIMP, Inkscape, Krita and Vectr which are mentioned in <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools#ui-design-tools\">UI design tools<\/a>.<\/p>\n<\/div>\n<h3><a id=\"user-content-mockup-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#mockup-tools\"><\/a>Mockup Tools<\/h3>\n<p>A mockup is a visual way of representing the product. While a wireframe mostly represents a product\u2019s structure, a mockup shows how the product is going to look like. These mockup tools that help you create and collaborate on mockups, wireframes, diagrams, and prototypes:<\/p>\n<ul><li><a href=\"https:\/\/artboard.studio\/\" rel=\"nofollow\">Artboard Studio<\/a> \u2014 is world's the first online graphic design application mainly focused on product mockups.<\/li>\n<li><a href=\"https:\/\/cleanmock.com\" rel=\"nofollow\">Cleanmock<\/a> \u2014 create stunning mockups using latest device frames like iPhone &amp; custom backgrounds.<\/li>\n<li><a href=\"https:\/\/craftwork.design\/\" rel=\"nofollow\">Craftwork Design<\/a> \u2014 free and premium high quality digital products which make your work faster and easier.<\/li>\n<li><a href=\"https:\/\/facebook.design\/devices\" rel=\"nofollow\">Devices by Facebook<\/a> \u2014 images and Sketch files of popular devices. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/dimmy.club\/\" rel=\"nofollow\">Dimmy.club<\/a> \u2014 device mockup generator for your website and app screenshots.<\/li>\n<li><a href=\"https:\/\/frrames.com\/\" rel=\"nofollow\">Frrames<\/a> \u2014 Frrames mockups is perfectly crafted responsive windows mockups for your ideal presentation.<\/li>\n<li><a href=\"https:\/\/www.ls.graphics\/\" rel=\"nofollow\">Lstore Graphics<\/a> \u2014 free and premium mockups, UI\/UX tools, scene creators for busy designers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.mockflow.com\/\" rel=\"nofollow\">Mockflow<\/a> \u2014 the quickest way to brainstorm user interface ideas.<\/li>\n<li><a href=\"https:\/\/www.mockupworld.co\/\" rel=\"nofollow\">Mockup World<\/a> \u2014 tons of free and legal, fully layered, easily customizable photo realistic PSDs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/mockupsforfree.com\/\" rel=\"nofollow\">Mockups For Free<\/a> \u2014 free design resources, PSD files for graphic and web designers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/mockuuups.studio\/\" rel=\"nofollow\">Mockuuups<\/a> \u2014 drag-and-drop tool for creating beautiful app previews or any marketing materials.<\/li>\n<li><a href=\"https:\/\/www.mock.video\/\" rel=\"nofollow\">Mock Video<\/a> \u2014 instantly create mockups by adding a device frame to your videos.<\/li>\n<li><a href=\"https:\/\/moqups.com\/\" rel=\"nofollow\">Moqups<\/a> \u2014 helps you create and collaborate in real-time on wireframes, mockups, diagrams and prototypes.<\/li>\n<li><a href=\"https:\/\/overframe.xyz\/\" rel=\"nofollow\">Overframe for Mac<\/a> \u2014 record your prototype &amp; app with device frame overlay. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pixelbuddha.net\/\" rel=\"nofollow\">PixelBuddha<\/a> \u2014 free and premium high quality resources for web designers and developers.<\/li>\n<li><a href=\"https:\/\/store.ramotion.com\/\" rel=\"nofollow\">Ramotion Store<\/a> \u2014 carefully crafted Apple and Android mockups for Sketch and Photoshop.<\/li>\n<li><a href=\"https:\/\/www.rotato.xyz\/\" rel=\"nofollow\">Rotato<\/a> \u2014 animated 3D mockups for your app designs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.screely.com\/\" rel=\"nofollow\">Screely<\/a> \u2014 quickly frame web page designs into a minimalist browser window. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.screenspace.io\/\" rel=\"nofollow\">ScreenSpace<\/a> \u2014 3D devices videos for motion designer.<\/li>\n<li><a href=\"https:\/\/smartmockups.com\/\" rel=\"nofollow\">Smartmockups<\/a> \u2014 create stunning product mockups with just a few clicks.<\/li>\n<li><a href=\"https:\/\/shotsnapp.com\/\" rel=\"nofollow\">shotsnapp<\/a> \u2014 create beautiful device mockup presentation for your digital app and website design.<\/li>\n<li><a href=\"https:\/\/themockup.club\/\" rel=\"nofollow\">The Mockup Club<\/a> \u2014 a directory of the best free design mockups for Photoshop, Sketch, Figma and InVision Studio. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/threed.io\/\" rel=\"nofollow\">Threed<\/a> \u2014 generate custom 3D Device Mockups in your Browser.<\/li>\n<li><a href=\"https:\/\/www.uistore.design\/\" rel=\"nofollow\">UI Store Design<\/a> \u2014 200+ free handpicked design resources for Sketch, Figma, Adobe XD, InVision Studio, Photoshop, Illustrator CC. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/ui8.net\/category\/freebies\" rel=\"nofollow\">UI8 - Design Freebies<\/a> \u2014 epic design freebies to get you started. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/mockups.kkuistore.com\/\" rel=\"nofollow\">Vector Mockups Library<\/a> \u2014 free collection of presentation Mockups for Sketch, Figma &amp; Photoshop. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-no-code-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#no-code-tools\"><\/a>No Code Tools<\/h3>\n<p>With a rise of \u201cno code tools\u201d, everyone with a laptop can build and launch a project. These tools help designers and makers create websites, apps, and even games. No code tools allow to automate routine tasks and can be used without a development background. Take a look at the tools here and if you need more \u2014 check Design to Code section.<\/p>\n<ul><li><a href=\"https:\/\/bubble.is\/\" rel=\"nofollow\">Bubble<\/a> \u2014 build and host web applications without having to write code or hire a team of engineers.<\/li>\n<li><a href=\"https:\/\/carrd.co\/\" rel=\"nofollow\">Carrd<\/a> \u2014 simple, free, fully responsive one-page sites for pretty much anything. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/coda.io\" rel=\"nofollow\">Coda<\/a> \u2014 a new type of document that blends the flexibility of documents, the power of spreadsheets, and the utility of apps into a single new canvas. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pagexl.com\/\" rel=\"nofollow\">PageXL<\/a> \u2014 a simple one-page website builder for quickly creating landing pages. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.remove.bg\/\" rel=\"nofollow\">Remove.bg<\/a> \u2014 is a free service to remove the background of any photo. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/tryretool.com\/\" rel=\"nofollow\">Retool<\/a> \u2014 gives you building blocks and you can build tools much faster.<\/li>\n<li><a href=\"https:\/\/www.sheet2site.com\/\" rel=\"nofollow\">Sheet2Site<\/a> \u2014 create a website from Google Sheets without writing code.<\/li>\n<li><a href=\"https:\/\/www.shopify.com\/\" rel=\"nofollow\">Shopify<\/a> \u2014 one platform with all the ecommerce and point of sale features you need to start, run, and grow your business.<\/li>\n<li><a href=\"https:\/\/thunkable.com\/\" rel=\"nofollow\">Thunkable<\/a> \u2014 a drag-and-drop tool for anyone to build native mobile apps.<\/li>\n<\/ul><h3><a id=\"user-content-pixel-art-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#pixel-art-tools\"><\/a>Pixel Art Tools<\/h3>\n<p>Pixel art is a digital art form where color is applied to individual pixels to create an image. The pixel art can be used to create everything from intricate scenes and game backgrounds to character designs or emoji. If you feel curious to try, check this pixel art software for both macOS and Windows:<\/p>\n<ul><li><a href=\"https:\/\/www.aseprite.org\/\" rel=\"nofollow\">Aseprite<\/a> \u2014 animated sprite editor &amp; pixel art tool.<\/li>\n<li><a href=\"https:\/\/github.com\/gmattie\/Data-Pixels\">Data Pixels<\/a> \u2014 create pixel art programmatically. Includes DataPixels.js and Data Pixels Playground desktop app. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/gitlab.com\/GrafX2\/grafX2\" rel=\"nofollow\">Grafx2<\/a> \u2014 a bitmap paint program inspired by the Amiga programs \u200bDeluxe Paint and Brilliance. Specialized in 256-color drawing. <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\">open-source.svg<\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/graphicsgale.com\/us\/\" rel=\"nofollow\">GraphicsGale<\/a> \u2014 has many pixel-art specific features, including palette control, selectively erasing colors, and tools for quickly replacing and trying new colors. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/jackschaedler.github.io\/goya\/\" rel=\"nofollow\">Goya<\/a> \u2014 a pixel-art editor built on ClojureScript and Om. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/make8bitart.com\/\" rel=\"nofollow\">Make 8 bit art<\/a> \u2014 an open source and in-browser pixel art drawing tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/jennschiffer\/make8bitart\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.piskelapp.com\/\" rel=\"nofollow\">Piskel<\/a> \u2014 online editor for animated sprites &amp; pixel art. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.pixelartcss.com\/\" rel=\"nofollow\">Pixel art to CSS<\/a> \u2014 pixel art animation and drawing web app powered by React. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/app.poxi.rocks\/\" rel=\"nofollow\">Poxi<\/a> \u2014 a flat pixel art editor. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pyxeledit.com\/\" rel=\"nofollow\">Pyxel Edit<\/a> \u2014 a pixel art editor designed to make it fun and easy to make tilesets, levels and animations. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-prototyping-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#prototyping-tools\"><\/a>Prototyping Tools<\/h3>\n<p>A prototype is a simple experimental design of a proposed solution. It should help to test ideas, design assumptions, and hypothesis in a fast and cheap way. Prototyping tools allow designers and clients to see how the product would function in the real world and collaborate on this solution. Many modern prototyping tools can use for wireframing, prototyping, and collaboration:<\/p>\n<ul><li><a href=\"https:\/\/www.meetalva.io\/\" rel=\"nofollow\">Alva<\/a> \u2014 create living prototypes with code components. It's also open source. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/meetalva\/alva\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.axure.com\/\" rel=\"nofollow\">Axure RP<\/a> \u2014 wireframing, prototyping, collaboration &amp; specifications generation.<\/li>\n<li><a href=\"https:\/\/balsamiq.com\/\" rel=\"nofollow\">Balsamiq<\/a> \u2014 wireframing and collaboration.<\/li>\n<li><a href=\"https:\/\/creolabs.com\/\" rel=\"nofollow\">Creo<\/a> \u2014 prototyping, code exporting and built-in mobile app builder. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.invisionapp.com\/\" rel=\"nofollow\">InVision<\/a> \u2014 prototyping, collaboration &amp; workflow platform. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.flinto.com\/\" rel=\"nofollow\">Flinto<\/a> \u2014 a Mac app for creating interactive and animated prototypes of app designs.<\/li>\n<li><a href=\"https:\/\/framer.com\/\" rel=\"nofollow\">Framer X<\/a> \u2014 a tool to visually design realistic interactive prototypes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.apple.com\/keynote\/\" rel=\"nofollow\">Keynote<\/a> \u2014 the built-in Mac app for creating presentations that can also be used for quick prototyping (see how Apple designers <a href=\"https:\/\/developer.apple.com\/videos\/play\/wwdc2014\/223\/\" rel=\"nofollow\">use it<\/a> to verify design concepts). <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/marvelapp.com\/\" rel=\"nofollow\">Marvel App<\/a> \u2014 the collaborative design platform. Wireframe, prototype, design online and create design specs in one place.<\/li>\n<li><a href=\"https:\/\/maze.design\/\" rel=\"nofollow\">Maze<\/a> \u2014 a tool for designers and developers that gives analytical results with actionable KPIs for your Invision prototypes.<\/li>\n<li><a href=\"https:\/\/pencil.evolus.vn\/\" rel=\"nofollow\">Pencil<\/a> \u2014 prototyping tool with many built-in components that people can easily install and use to create mockups in popular desktop platforms. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/principleformac.com\/\" rel=\"nofollow\">Principle<\/a> \u2014 makes it easy to design animated and interactive user interfaces. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.protopie.io\/\" rel=\"nofollow\">ProtoPie<\/a> \u2014 piece hi-fi interactions together, build sensor-aided prototypes and share your amazing creations in a matter of minutes.<\/li>\n<li><a href=\"https:\/\/proto.io\/\" rel=\"nofollow\">Proto.io<\/a> \u2014 a tool to create fully-interactive high-fidelity prototypes that look and work exactly as your app should.<\/li>\n<li><a href=\"https:\/\/www.uxpin.com\/\" rel=\"nofollow\">UXPin<\/a> \u2014 build prototypes that feel real, with powers of code components, logic, states and design systems. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>You can also do prototyping with Figma, Adobe XD, Sketch and InVision Studio, which are mentioned in the <a href=\"#ui-design-tools\">UI Design Tools section<\/a>.<\/p>\n<\/div>\n<p style=\"text-align: center;\">\n    <a href=\"https:\/\/flawlessapp.io\/feedback?github\" rel=\"nofollow\">\n        <img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/flawless-feedback.png\" alt=\"Awesome-Design-Tools\" \/><\/a>\n<\/p>\n<h3><a id=\"user-content-screenshot-software\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshot-software\"><\/a>Screenshot Software<\/h3>\n<p>Taking screenshots is a typical task in the design &amp; development workflow. So these free and full-featured screenshot apps can help you capture a screen with ease. Some screen captures are macOS only, while others support both OS:<\/p>\n<ul><li><a href=\"https:\/\/www.techsmith.com\/video-editor.html\" rel=\"nofollow\">Camtasia<\/a> \u2014 a screen recorder that comes with full-blown built in editor.<\/li>\n<li><a href=\"https:\/\/getcleanshot.com\" rel=\"nofollow\">CleanShot<\/a> \u2014 capture your screen in a superior way with built-in annotation tool and Quick Access Overlay. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.getcloudapp.com\/\" rel=\"nofollow\">CloudApp<\/a> \u2014 record videos, webcam, annotate screenshots, create GIFs.<\/li>\n<li><a href=\"https:\/\/giphy.com\/apps\/giphycapture\" rel=\"nofollow\">Giphy Capture<\/a> \u2014 capture parts of your screen and export as gif or mp4. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/getgreenshot.org\/\" rel=\"nofollow\">Greenshot<\/a> \u2014 take a screenshot of a selected region, window or fullscreen. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/getkap.co\" rel=\"nofollow\">Kap<\/a> \u2014 open source screen recorder with options to export to GIF, MP4, WebM and APNG. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/wulkano\/kap\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/itunes.apple.com\/us\/app\/lightshot-screenshot\/id526298438\" rel=\"nofollow\">Lighshot<\/a> \u2014 taking quick captures of your screen. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/itunes.apple.com\/us\/app\/monosnap\/id540348655?mt=12\" rel=\"nofollow\">Monosnap<\/a> \u2014 create screenshots, annotate and upload them to the cloud. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/obsproject.com\/\" rel=\"nofollow\">OBS<\/a> \u2014 open source software for video recording and live streaming. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/support.apple.com\/quicktime\" rel=\"nofollow\">Quicktime<\/a> \u2014 a video player that you can use to record your screen. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.thnkdev.com\/Screenie\/\" rel=\"nofollow\">Screenie<\/a> \u2014 filter and search through images, change screenshot filetypes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.screentogif.com\/\" rel=\"nofollow\">ScreenToGif<\/a> \u2014 record a gif of part of the screen. Only available for Windows. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/wulkano\/kap\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/getsharex.com\/\" rel=\"nofollow\">ShareX<\/a> \u2014 screen capture, file sharing and productivity tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/support.microsoft.com\/en-us\/help\/13776\/windows-use-snipping-tool-to-capture-screenshots\" rel=\"nofollow\">Snipping Tool<\/a> \u2014 Windows free screenshot tool. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/snappy-app.com\/\" rel=\"nofollow\">Snappy<\/a> \u2014 takes quick shots and organizes them for you into collections. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/teampaper.me\/snap\/\" rel=\"nofollow\">Teampaper Snap<\/a> \u2014 allows you to take screenshots of a selected area. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-sketching-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#sketching-tools\"><\/a>Sketching Tools<\/h3>\n<p>Sometimes you need just a pencil and paper to start creating your app or website. So here are you can find online sketching tools with great sketch sheet templates to speed up your creative process:<\/p>\n<ul><li><a href=\"https:\/\/zurb.com\/playground\/responsive-sketchsheets\" rel=\"nofollow\">Responsive Sketchsheets<\/a> \u2014 offers responsive pre-design templates. These will automatically adapt their layout to any screen size and include neat features. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/sketchsheets.com\/\" rel=\"nofollow\">Sketchsheets<\/a> \u2014 an open source project dedicated to providing free printable templates of the latest devices and platforms for wireframing designs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/rnarrkus\/sketchsheets\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/sneakpeekit.com\/\" rel=\"nofollow\">Sneakpeekit<\/a> \u2014 print note taking grids and devices frames. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/sketchize.com\/\" rel=\"nofollow\">Sketchize<\/a> \u2014 just choose any of the sketch sheets that fits your project, print it out and start to sketch your dream. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><div>\n<p>You can also do some sketching with Sketch mentioned in <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools#ui-design-tools\">UI design tools<\/a>.<\/p>\n<\/div>\n<h3><a id=\"user-content-sound-design\" class=\"anchor\" aria-hidden=\"true\" href=\"#sound-design\"><\/a>Sound Design<\/h3>\n<p>Sound design is an art of creating a soundscape for a site, app, movie, game or any other product. The sound has great potential for transforming the way people connect with your product. Some sound design software is very advanced and can be used mostly by sound designers, while others are good for beginners.<\/p>\n<ul><li><a href=\"https:\/\/appsounds.pro\/\" rel=\"nofollow\">Appsounds<\/a> \u2014 UI Sound packs for apps, games, and any product.<\/li>\n<li><a href=\"https:\/\/audiojungle.net\/\" rel=\"nofollow\">AudioJungle<\/a> \u2014 836,206 tracks and sounds from the community of musicians and sound engineers.<\/li>\n<li><a href=\"https:\/\/www.bensound.com\/\" rel=\"nofollow\">Bensound<\/a> \u2014 download creative commons music, royalty free music for free and use it in your project. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/freesound.org\/\" rel=\"nofollow\">Freesound<\/a> \u2014 a collaborative database of Creative Commons Licensed sounds. Browse, download and share sounds. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/icons8.com\/music\" rel=\"nofollow\">Fugue Music<\/a> \u2014 download free background music for your videos. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/cycling74.com\/products\/max\/\" rel=\"nofollow\">Max<\/a> \u2014 connect objects with virtual patch cords to create interactive sounds, graphics, and custom effects.<\/li>\n<li><a href=\"http:\/\/reaper.fm\/\" rel=\"nofollow\">Reaper<\/a> \u2014 import any audio and MIDI, synthesize, sample, compose, arrange, edit, mix, and master songs or any other audio projects.<\/li>\n<li><a href=\"https:\/\/sonic-pi.net\/\" rel=\"nofollow\">Sonic Pi<\/a> \u2014 a live coding music synth. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/soundkit.io\/\" rel=\"nofollow\">SoundKit<\/a> \u2014 a UI sound library designed for all of your interface needs.<\/li>\n<li><a href=\"https:\/\/uisounds.prototypr.io\/\" rel=\"nofollow\">UI Sounds<\/a> \u2014 learn sound design for user interfaces by example.<\/li>\n<li><a href=\"https:\/\/wistia.com\/resources\/music\" rel=\"nofollow\">Wistia Music<\/a> \u2014 download some free background tracks to add energy and emotion to your videos. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.wowa.me\/\" rel=\"nofollow\">WOWA<\/a> \u2014 download royalty free music for YouTube videos, Podcasts and Apps. No copyright CC0. Music inspired by Unsplash. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/audiolibrary\/music\" rel=\"nofollow\">YouTube Audio Library<\/a> \u2014 browse and download free music for your project. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-stock-photos-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#stock-photos-tools\"><\/a>Stock Photos Tools<\/h3>\n<p>Need a high-quality photo for iOS app or new banner? You can always shoot it yourself or borrow from the stock photo sites. Luckily, there are hundreds of beautiful, free stock photos and royalty-free images that you can use for any project:<\/p>\n<ul><li><a href=\"https:\/\/burst.shopify.com\/\" rel=\"nofollow\">Burst<\/a> \u2014 free stock photos for websites and commercial use. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/duotone.shapefactory.co\/\" rel=\"nofollow\">Duotone<\/a> \u2014 free duotone images to use in any project, or make custom duotone images. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/foodiesfeed.com\/\" rel=\"nofollow\">FoodiesFeed<\/a> \u2014 thousands of beautiful realistic free food pictures in high resolution. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/freephotos.cc\/\" rel=\"nofollow\">FreePhotos.cc<\/a> \u2014 free stock photos for commercial use. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/freestocks.org\/\" rel=\"nofollow\">Freestocks.org<\/a> \u2014 high quality photos all released under Creative Commons CC0. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/gratisography.com\" rel=\"nofollow\">Gratisography<\/a> \u2014 collection of free high-resolution pictures. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/jaymantri.com\/\" rel=\"nofollow\">Jay Mantri<\/a> \u2014 7 new photos released every Thursday under the Creative Commons CC0. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/kaboompics.com\/\" rel=\"nofollow\">Kaboom Pics<\/a> \u2014 stock photos including abstract, city\/architecture, fashion, food &amp; landscapes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/landingstock.com\/\" rel=\"nofollow\">LandingStock<\/a> \u2014 a collection of free images for your landing page. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.lifeofpix.com\/\" rel=\"nofollow\">Life of Pix<\/a> \u2014 free high-resolution photos, created by the LEEROY team. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/magdeleine.co\/\" rel=\"nofollow\">Magdeleine<\/a> \u2014 free high-quality stock photos for your inspiration. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/photos.icons8.com\/\" rel=\"nofollow\">Moose<\/a> \u2014 don't search for stock photos, create them. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/mmtstock.com\/\" rel=\"nofollow\">MMT STock<\/a> \u2014 high resolution photos provided by Jeffrey Betts with Creative Commons CC0. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/nos.twnsnd.co\/\" rel=\"nofollow\">New Old Stock<\/a> \u2014 a vintage photos from the public archives free of known copyright restrictions. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.pexels.com\/\" rel=\"nofollow\">Pexels<\/a> \u2014 an aggregate of many free stock photo resources. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/photos.icons8.com\/creator\" rel=\"nofollow\">Photo Creator<\/a> \u2014 a free tool for creating realistic stock photos on your demand. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/picography.co\/\" rel=\"nofollow\">Picography<\/a> \u2014 free stock photos by Dave Meier and various other photographers with Creative Commons CC0. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pixabay.com\/\" rel=\"nofollow\">Pixabay<\/a> \u2014 sharing photos, illustrations, vector graphics, and film footage under a proprietary license. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/picjumbo.com\/\" rel=\"nofollow\">Picjumbo<\/a> \u2014 a collection of totally free photos for your commercial &amp; personal works. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pngtree.com\/\" rel=\"nofollow\">Pngtree<\/a> \u2014 millions of PNG images, backgrounds and vectors for free download. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pxhere.com\/\" rel=\"nofollow\">pxhere<\/a> \u2014 free image stock. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.reshot.com\/\" rel=\"nofollow\">Reshot<\/a> \u2014 a massive library of handpicked free stock photos you won\u2019t find elsewhere. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/shotstash.com\/\" rel=\"nofollow\">ShotStash<\/a> \u2014 thousands of free high-resolution CC0 photos and videos. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/skitterphoto.com\/\" rel=\"nofollow\">SkitterPhoto<\/a> \u2014 a wide variety of stock photos and are released under Creative Commons CC0. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/startupstockphotos.com\/\" rel=\"nofollow\">Startup Stock Photos<\/a> \u2014 free photos for startups, bloggers and publishers. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/stocksnap.io\/\" rel=\"nofollow\">StockSnap.io<\/a> \u2014 a large selection of free stock photos and high resolution images. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/styledstock.co\/\" rel=\"nofollow\">StyledStock<\/a> \u2014 free feminine stock photography for every woman entrepreneur. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/uifaces.co\/\" rel=\"nofollow\">UI Faces<\/a> \u2014 aggregator that indexes various free avatar sources that you can use in your design mockups. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/unsplash.com\" rel=\"nofollow\">Unsplash<\/a> \u2014 stock photos free to use. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.flickr.com\/photos\/wocintechchat\" rel=\"nofollow\">#WOCinTech Chat Photos<\/a> \u2014 free stock photos of women technologists of diverse backgrounds. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/www.zoommyapp.com\" rel=\"nofollow\">Zoommy<\/a> \u2014 helps you find awesome free stock photos for your creative product or inspiration. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-tools-for-learning-design\" class=\"anchor\" aria-hidden=\"true\" href=\"#tools-for-learning-design\"><\/a>Tools for Learning Design<\/h3>\n<h3><a id=\"user-content-stock-videos\" class=\"anchor\" aria-hidden=\"true\" href=\"#stock-videos\"><\/a>Stock Videos<\/h3>\n<p>If you work with video-content, you will enjoy these high-quality, hand-curated stock videos. You'll find many good and free stock video sites below, which you can use on your website, in your ad campaigns or social media:<\/p>\n<ul><li><a href=\"https:\/\/coverr.co\/\" rel=\"nofollow\">Coverr<\/a> \u2014 beautiful, free stock video footage for your homepage. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/www.lifeofvids.com\/\" rel=\"nofollow\">Life of Vids<\/a> \u2014 provides free stock videos, clips, and loops. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"http:\/\/mazwai.com\" rel=\"nofollow\">Mazwai<\/a> \u2014 great collection of free creative commons HD video clips &amp; footages. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/mixkit.co\/\" rel=\"nofollow\">Mixkit<\/a> \u2014 extraordinary Free HD videos. For you to use, completely free of charge. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.motionplaces.com\/\" rel=\"nofollow\">Motion Places<\/a> \u2014 a curated collection of beautiful footage you can use for your projects. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pixabay.com\/\" rel=\"nofollow\">Pixabay<\/a> \u2014 it has more than 1.5 million royalty-free stock videos and photos shared by their community of creators. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.pexels.com\/videos\/\" rel=\"nofollow\">Pexels Videos<\/a> \u2014 makes it easy to find free stock footage for your website, promo video or anything else. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.videezy.com\/\" rel=\"nofollow\">Videezy<\/a> \u2014 download millions of free and premium stock footage and motion graphics instantly.<\/li>\n<li><a href=\"https:\/\/www.videvo.net\/\" rel=\"nofollow\">Videvo<\/a> \u2014 videvo offers a large selection of HD video clips, motion graphics and free stock footage. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.123rf.com\/\" rel=\"nofollow\">123RF<\/a> \u2014 over 100,000 daily new content &amp; 24\/7 online support.<\/li>\n<\/ul><h3><a id=\"user-content-ui-design-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#ui-design-tools\"><\/a>UI Design Tools<\/h3>\n<p>What are the best UI design tools in 2019? You can pick any of the tools below and it will allow you to do dozens of design tasks \u2014 UI for site or mobile, wireframe, prototype, animation, logo. These are great and fully-functional tools for UX &amp; UI designers:<\/p>\n<ul><li><a href=\"https:\/\/www.adobe.com\/products\/xd.html\" rel=\"nofollow\">Adobe XD<\/a> \u2014 design, prototype, and share any user experience, from websites and mobile apps to voice interactions. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/affinity.serif.com\/en-gb\/designer\/\" rel=\"nofollow\">Affinity Designer<\/a> \u2014 a vector graphics editor for macOS, iOS, and Microsoft Windows.<\/li>\n<li><a href=\"https:\/\/www.botmock.com\" rel=\"nofollow\">Botmock<\/a> \u2014 design, prototype, and test voice and text conversational apps. Supports multiple platforms.<\/li>\n<li><a href=\"https:\/\/www.cleverbrush.com\/\" rel=\"nofollow\">CleverBrush<\/a> \u2014 a browser-based online vector editor and digital publishing tool which is possible to integrate to page as JS component.<\/li>\n<li><a href=\"https:\/\/www.figma.com\/\" rel=\"nofollow\">Figma<\/a> \u2014 a design tool based in the browser, that allows to design and prototype with real-time collaboration.<\/li>\n<li><a href=\"https:\/\/www.gimp.org\/\" rel=\"nofollow\">GIMP<\/a> \u2014 a free &amp; open source imaging and graphic design software. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/designer.io\/\" rel=\"nofollow\">Gravit<\/a> \u2014 a free vector design app, available for macOS, Windows, Linux, ChromeOS and browser. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.adobe.com\/products\/illustrator.html\" rel=\"nofollow\">Illustrator<\/a> \u2014 create logos, icons, drawings, typography, and illustrations for print, web, video, and mobile. Made by Adobe.<\/li>\n<li><a href=\"https:\/\/inkscape.org\/\" rel=\"nofollow\">Inkscape<\/a> \u2014 a free and open-source vector graphics editor. It has ability to create or edit vector graphics such as illustrations, diagrams, line arts, charts, logos and complex paintings. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/krita.org\/en\/\" rel=\"nofollow\">Krita<\/a> \u2014 a free painting and graphic design software considered a good alternative to Adobe Photoshop. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/icons8.com\/lunacy\" rel=\"nofollow\">Lunacy<\/a> \u2014 a free native windows app that works offline and supports .sketch files. Flexible and light weighed. Intuitive and easy to use. Speedups and empowers UI and UX designers. The must have to produce stunning designs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.photopea.com\/\" rel=\"nofollow\">Photopea<\/a> \u2014 a free browser-based graphic design app alternative to Photoshop. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.adobe.com\/products\/photoshop.html\" rel=\"nofollow\">Photoshop<\/a> \u2014 imaging and graphic design software developed by Adobe.<\/li>\n<li><a href=\"https:\/\/studio.pixelixe.com\" rel=\"nofollow\">Pixelixe<\/a> \u2014 a graphic design tool built for marketers, bloggers and small businesses that needs to create stunning and unique images, graphics or static webpages in minutes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.sketchapp.com\/\" rel=\"nofollow\">Sketch<\/a> \u2014 a design toolkit built for Mac. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.invisionapp.com\/studio\" rel=\"nofollow\">Studio<\/a> \u2014 combines design, prototyping, and collaboration into one harmonious workflow. Made by Invision.<\/li>\n<li><a href=\"https:\/\/vectr.com\/\" rel=\"nofollow\">Vectr<\/a> \u2014 a simple web and desktop cross-platform tool to create vector graphics easily and intuitively. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-user-flow-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-flow-tools\"><\/a>User Flow Tools<\/h3>\n<p>User flow is a series of steps a user takes to achieve a meaningful goal. It's a fast way to plan customer journey paths and improve UX. So if you need to make a user flow diagram, user flow map or a sitemap, take a look at these tools:<\/p>\n<ul><li><a href=\"https:\/\/www.draw.io\/\" rel=\"nofollow\">Draw.io<\/a> \u2014 a free online diagram software for making flowcharts, process diagrams, org charts, UML, ER and network diagrams. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/flowmapp.com\/\" rel=\"nofollow\">Flowmapp<\/a> \u2014 an online tool for creating sitemaps and user flows that helps you to effectively design and plan user experience.<\/li>\n<li><a href=\"https:\/\/docs.google.com\/drawings\/\" rel=\"nofollow\">Google Drawings<\/a> \u2014 create diagrams and charts, for free; all within Google Docs. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.lucidchart.com\/\" rel=\"nofollow\">Lucidchart<\/a> \u2014 online tool for creating diagrams, flow charts, sitemaps, and more.<\/li>\n<li><a href=\"https:\/\/mindnode.com\/\" rel=\"nofollow\">MindNode<\/a> \u2014 a mind mapping app that makes brainstorming simple and easy. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/ninjamock.com\/\" rel=\"nofollow\">NinjaMock<\/a> \u2014 wireframe and user flow online tool. Link your views and create logic flow prototype. All with freehand visual style.<\/li>\n<li><a href=\"https:\/\/www.omnigroup.com\/omnigraffle\/\" rel=\"nofollow\">OmniGraffle<\/a> \u2014 a diagramming and digital illustration application for macOS and iOS. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/overflow.io\/\" rel=\"nofollow\">Overflow<\/a> \u2014 turn your designs into playable user flow diagrams that tell a story.<\/li>\n<li><a href=\"https:\/\/sketch.systems\/\" rel=\"nofollow\">Sketch.systems<\/a> \u2014 ui and flow design with interactive state machines. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/uxmisfit.com\/squid\/\" rel=\"nofollow\">SQUID<\/a> \u2014 create beautiful User Flows in Sketch in just minutes. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.websequencediagrams.com\/\" rel=\"nofollow\">WebSequenceDiagrams<\/a> \u2014 a simple webapp to work out object interactions arranged in time sequence. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/whimsical.co\/\" rel=\"nofollow\">Whimsical<\/a> \u2014 easy to create flow charts, wireframes and sticky notes.<\/li>\n<li><a href=\"http:\/\/wireflow.co\" rel=\"nofollow\">Wireflow<\/a> \u2014 free, online and open source tool for creating beautiful user flow prototypes. <a href=\"https:\/\/github.com\/vanila-io\/wireflow\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.xmind.net\/zen\/\" rel=\"nofollow\">XMind: ZEN<\/a> \u2014 a brainstorming and mind mapping tool that can switch between outline and tree-chart. Link topics with other charts.<\/li>\n<li><a href=\"https:\/\/www.yworks.com\/products\/yed\" rel=\"nofollow\">yEd - Graph Editor<\/a> \u2014 free desktop tool for making diagrams. Usable for wide variety of use cases. Auto-layout helps a lot when making flowcharts. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<\/ul><h3><a id=\"user-content-user-research-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-research-tools\"><\/a>User Research Tools<\/h3>\n<p>User research helps you understand user behaviors, needs, and motivations through various qualitative and quantitative methods (interviews, observation, forms, etc). These user research tools can be useful for you:<\/p>\n<ul><li><a href=\"https:\/\/calendly.com\/\" rel=\"nofollow\">Calendly<\/a> \u2014 Calendly helps you schedule meetings without the back-and-forth emails.<\/li>\n<li><a href=\"https:\/\/crowdsignal.com\/\" rel=\"nofollow\">Crowd Signal<\/a> \u2014 collect, organize and analyze data from a variety of sources, including social media and mobile.<\/li>\n<li><a href=\"https:\/\/www.gotomeeting.com\/en-gb\" rel=\"nofollow\">GoToMeeting<\/a> \u2014 a simple, extraordinarily powerful web conferencing service.<\/li>\n<li><a href=\"https:\/\/www.feedbacklite.com\/\" rel=\"nofollow\">Feedback Lite<\/a> \u2014 collect high quality customer feedback using Voice of Customer solutions designed to improve your website performance and boost customer engagement.<\/li>\n<li><a href=\"https:\/\/www.airsquirrels.com\/reflector\" rel=\"nofollow\">Reflector<\/a> \u2014 Reflector is a basic screen-mirroring and recording tool so you can conduct user tests remotely, using any existing wireframes or prototypes.<\/li>\n<li><a href=\"https:\/\/www.optimalworkshop.com\/reframer\" rel=\"nofollow\">Reframer<\/a> \u2014 a research tool that helps your team to capture, tag (code) and identify patterns in qualitative data across multiple research participants.<\/li>\n<li><a href=\"https:\/\/www.surveymonkey.com\/\" rel=\"nofollow\">Survey Monkey<\/a> \u2014 online survey tool to capture the voices and opinions of the people who matter most to you.<\/li>\n<li><a href=\"https:\/\/www.typeform.com\/\" rel=\"nofollow\">Typeform<\/a> \u2014 use a simple drag-and-drop interface to create any kind of form, survey, or questionnaire, and even take credit card payments.<\/li>\n<li><a href=\"https:\/\/www.wufoo.com\" rel=\"nofollow\">Wufoo<\/a> \u2014 \u0441reate forms, collect data and payments, and automate your workflows.<\/li>\n<li><a href=\"https:\/\/www.yesinsights.com\/\" rel=\"nofollow\">YesInsights<\/a> \u2014 simple one question and NPS surveys to improve your business.<\/li>\n<li><a href=\"https:\/\/userbitapp.com\/\" rel=\"nofollow\">UserBit<\/a> \u2014 a platform of real-time research tools for your team. Tag\/code interviews and feedback, capture insights, create personas, visual sitemaps and more.<\/li>\n<li><a href=\"https:\/\/www.userinterviews.com\/\" rel=\"nofollow\">User Interviews<\/a> \u2014 recruit participants from a community of 125,000 members based on profession, demographics, geography and more, for any kind of research.<\/li>\n<\/ul><p style=\"text-align: center;\">\n    <a href=\"https:\/\/flawlessapp.io\/feedback?github\" rel=\"nofollow\">\n        <img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/flawless-feedback.png\" alt=\"Awesome-Design-Tools\" \/><\/a>\n<\/p>\n<h3><a id=\"user-content-visual-debugging-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#visual-debugging-tools\"><\/a>Visual Debugging Tools<\/h3>\n<ul><li><a href=\"https:\/\/logrocket.com\/\" rel=\"nofollow\">LogRocket<\/a> \u2014 a tool to record what users do on your site so you can reproduce bugs and fix issues faster.<\/li>\n<li><a href=\"https:\/\/getpixelsnap.com\" rel=\"nofollow\">PixelSnap<\/a> \u2014 the magical way to measure everything on your screen. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/mac.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/mac.svg?sanitize=true\" alt=\"mac.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/github.com\/GoogleChromeLabs\/ProjectVisBug\">VisBug<\/a> \u2014 on any webpage: visually debug, quick inspect style and a11y, change CSS visually, click and drag elements around, and much more. <a href=\"https:\/\/github.com\/GoogleChromeLabs\/ProjectVisBug\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.canvasflip.com\/visual-inspector\/\" rel=\"nofollow\">Visual Inspector<\/a> \u2014 collaboration tool for website feedback and fixing design bugs.<\/li>\n<\/ul><h3><a id=\"user-content-wireframing-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#wireframing-tools\"><\/a>Wireframing Tools<\/h3>\n<p>A wireframe is a visual mockup that outlines the basic structure of the site or an app. It contains the most essential elements and the content, helping you easily explain ideas on design. Wireframes are a low-fidelity way of showing a design. In this section is presented the best wireframing tools for a variety of use cases.<\/p>\n<ul><li><a href=\"https:\/\/balsamiq.cloud\/\" rel=\"nofollow\">Balsamiq Cloud<\/a> \u2014 effortless UI sketching. Capture your ideas, collaborate on your designs, get everyone on board.<\/li>\n<li><a href=\"https:\/\/cssgrid.cc\/css-grid-builder.html\" rel=\"nofollow\">CSS Grid Builder<\/a> \u2014 with this tool there is no HTML part to the grid. The grid only becomes apparent when content (elements) are added and positioned according to the grid rules. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.gliffy.com\/\" rel=\"nofollow\">Gliffy<\/a> \u2014 a tool for creating the framework, UML diagrams, flowcharts, wireframes and more.<\/li>\n<li><a href=\"http:\/\/gridzzly.com\/\" rel=\"nofollow\">Gridzzly<\/a> \u2014 the easiest way to create custom grid paper printable. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/grid.layoutit.com\/\" rel=\"nofollow\">Layoutit<\/a> \u2014 a interface builder for CSS Grid &amp; Bootstrap that wants to be the kick-off for your front-end developments. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/pidoco.com\/en\" rel=\"nofollow\">Pidoco<\/a> \u2014 software that lets you quickly create click-through wireframes and fully interactive UX prototypes.<\/li>\n<li><a href=\"https:\/\/wireframe.cc\/\" rel=\"nofollow\">Wireframe.cc<\/a> \u2014 a simple wireframing tool that doesn't get in your way. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/whimsical.co\/wireframes\/\" rel=\"nofollow\">Whimsical Wireframes<\/a> \u2014 instant wireframing at your fingertips. Rich library of configurable elements like buttons, inputs, checkboxes.<\/li>\n<li><a href=\"https:\/\/1200px.com\/\" rel=\"nofollow\">1200px Grid System<\/a> \u2014 helps you build a grid system for much wider website designs than the 960-pixel style.<\/li>\n<\/ul><h3><a id=\"user-content-three-d-modeling-software\" class=\"anchor\" aria-hidden=\"true\" href=\"#three-d-modeling-software\"><\/a>Three D Modeling Software<\/h3>\n<p>3D graphics are used in gaming, film making, architecture, engineering, and 3D printing. 3D artists &amp; designers use specific 3D modeling software, mentioned in this section.<\/p>\n<ul><li><a href=\"https:\/\/www.autodesk.com\/products\/fusion-360\/overview\" rel=\"nofollow\">Autodesk<\/a> \u2014 integrated CAD, CAM, and CAE software. Unify design, engineering, and manufacturing into a single platform.<\/li>\n<li><a href=\"https:\/\/www.blender.org\/\" rel=\"nofollow\">Blender<\/a> \u2014 free and open source 3D Creation Software. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.freecadweb.org\/\" rel=\"nofollow\">FreeCAD<\/a> \u2014 a free and opensource multiplatform 3D parametric modeler. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a href=\"https:\/\/github.com\/FreeCAD\/FreeCAD\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.matterhackers.com\/store\/l\/mattercontrol\/sk\/MKZGTDW6\" rel=\"nofollow\">MatterControl<\/a> \u2014 a free, open-source, all-in-one software package that lets you design, slice, organize, and manage your 3D prints. <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/free.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/free.svg?sanitize=true\" alt=\"free.svg\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Media\/open-source.svg\"><img src=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/raw\/master\/Media\/open-source.svg?sanitize=true\" alt=\"open-source.svg\" \/><\/a><\/li>\n<li><a href=\"https:\/\/www.autodesk.com\/products\/maya\/overview\" rel=\"nofollow\">Maya<\/a> \u2014 make animations, environments, motion graphics, virtual reality, and character creation, all in one toolset.<\/li>\n<li><a href=\"https:\/\/www.onshape.com\/\" rel=\"nofollow\">Onshape<\/a> \u2014 a modeling software specially oriented to design technical and spare parts, was the first full-cloud 3D software created.<\/li>\n<li><a href=\"https:\/\/www.rhino3d.com\/\" rel=\"nofollow\">Rhino<\/a> \u2014 a curve-based 3D modeling software that creates mathematically-precise models of 3D surfaces.<\/li>\n<li><a href=\"https:\/\/www.sketchup.com\" rel=\"nofollow\">SketchUp<\/a> \u2014 3D design software that truly makes 3D modeling for everyone, with a simple to learn yet robust toolset.<\/li>\n<li><a href=\"https:\/\/www.vectary.com\/\" rel=\"nofollow\">Vectary<\/a> \u2014 create beautiful 3D models with our drag and drop 3D modeling tool.<\/li>\n<li><a href=\"https:\/\/www.3dslash.net\" rel=\"nofollow\">3D Slash<\/a> \u2014 3D modeling tool, available on all devices and all OS, online and offline.<\/li>\n<\/ul><h3><a id=\"user-content-addendum-reference--inspiration\" class=\"anchor\" aria-hidden=\"true\" href=\"#addendum-reference--inspiration\"><\/a>Addendum (Reference &amp; Inspiration)<\/h3>\n<p>Awesome Design Tools is curated by <a href=\"https:\/\/twitter.com\/LisaDziuba\" rel=\"nofollow\">Lisa Dziuba<\/a> &amp; <a href=\"https:\/\/twitter.com\/ValiaHavryliuk\" rel=\"nofollow\">Valia Havruliyk<\/a> from Flawless team. And it was hugely inspired by articles from the design community and Prototypr.io Toolbox made by our good friend Graeme Fulton. If you found some great design tool, please <a href=\"https:\/\/github.com\/LisaDziuba\/Awesome-Design-Tools\/blob\/master\/Contribution_Guidelines.md\">suggest it<\/a>. Thanks for making this project awesome :)<\/p>\n<\/article>","created_at":"2019-04-05T11:04:05+0000","updated_at":"2019-04-05T11:04:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":41,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/23191656?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":896,"title":"Beyond Exception Handling: Conditions and Restarts","url":"http:\/\/www.gigamonkeys.com\/book\/beyond-exception-handling-conditions-and-restarts.html","content":"Beyond Exception Handling: Conditions and Restarts<p>Copyright \u00a9 2003-2005, Peter Seibel<\/p><p>One of Lisp's great features is its <i>condition<\/i> system. It serves a\nsimilar purpose to the exception handling systems in Java, Python, and\nC++ but is more flexible. In fact, its flexibility extends beyond\nerror handling--conditions are more general than exceptions in that a\ncondition can represent any occurrence during a program's execution\nthat may be of interest to code at different levels on the call stack.\nFor example, in the section \"Other Uses for Conditions,\" you'll see\nthat conditions can be used to emit warnings without disrupting\nexecution of the code that emits the warning while allowing code\nhigher on the call stack to control whether the warning message is\nprinted. For the time being, however, I'll focus on error handling.<\/p><p>The condition system is more flexible than exception systems because\ninstead of providing a two-part division between the code that\nsignals an error<sup>1<\/sup> and the code that handles it,<sup>2<\/sup> the condition system splits the\nresponsibilities into three parts--<i>signaling<\/i> a condition,\n<i>handling<\/i> it, and <i>restarting<\/i>. In this chapter, I'll describe\nhow you could use conditions in part of a hypothetical application\nfor analyzing log files. You'll see how you could use the condition\nsystem to allow a low-level function to detect a problem while\nparsing a log file and signal an error, to allow mid-level code to\nprovide several possible ways of recovering from such an error, and\nto allow code at the highest level of the application to define a\npolicy for choosing which recovery strategy to use.<\/p><p>To start, I'll introduce some terminology: <i>errors<\/i>, as I'll use\nthe term, are the consequences of Murphy's law. If something can go\nwrong, it will: a file that your program needs to read will be\nmissing, a disk that you need to write to will be full, the server\nyou're talking to will crash, or the network will go down. If any of\nthese things happen, it may stop a piece of code from doing what you\nwant. But there's no bug; there's no place in the code that you can\nfix to make the nonexistent file exist or the disk not be full.\nHowever, if the rest of the program is depending on the actions that\nwere going to be taken, then you'd better deal with the error somehow\nor you <i>will<\/i> have introduced a bug. So, errors aren't caused by\nbugs, but neglecting to handle an error is almost certainly a bug.<\/p><p>So, what does it mean to handle an error? In a well-written program,\neach function is a black box hiding its inner workings. Programs are\nthen built out of layers of functions: high-level functions are built\non top of the lower-level functions, and so on. This hierarchy of\nfunctionality manifests itself at runtime in the form of the call\nstack: if <code>high<\/code> calls <code>medium<\/code>, which calls <code>low<\/code>,\nwhen the flow of control is in <code>low<\/code>, it's also still in\n<code>medium<\/code> and <code>high<\/code>, that is, they're still on the call\nstack.<\/p><p>Because each function is a black box, function boundaries are an\nexcellent place to deal with errors. Each function--<code>low<\/code>, for\nexample--has a job to do. Its direct caller--<code>medium<\/code> in this\ncase--is counting on it to do its job. However, an error that\nprevents it from doing its job puts all its callers at risk:\n<code>medium<\/code> called <code>low<\/code> because it needs the work done that\n<code>low<\/code> does; if that work doesn't get done, <code>medium<\/code> is in\ntrouble. But this means that <code>medium<\/code>'s caller, <code>high<\/code>, is\nalso in trouble--and so on up the call stack to the very top of the\nprogram. On the other hand, because each function is a black box, if\nany of the functions in the call stack can somehow do their job\ndespite underlying errors, then none of the functions above it needs\nto know there was a problem--all those functions care about is that\nthe function they called somehow did the work expected of it.<\/p><p>In most languages, errors are handled by returning from a failing\nfunction and giving the caller the choice of either recovering or\nfailing itself. Some languages use the normal function return\nmechanism, while languages with exceptions return control by\n<i>throwing<\/i> or <i>raising<\/i> an exception. Exceptions are a vast\nimprovement over using normal function returns, but both schemes\nsuffer from a common flaw: while searching for a function that can\nrecover, the stack unwinds, which means code that might recover has\nto do so without the context of what the lower-level code was trying\nto do when the error actually occurred.<\/p><p>Consider the hypothetical call chain of <code>high<\/code>, <code>medium<\/code>,\n<code>low<\/code>. If <code>low<\/code> fails and <code>medium<\/code> can't recover, the\nball is in <code>high<\/code>'s court. For <code>high<\/code> to handle the error,\nit must either do its job without any help from <code>medium<\/code> or\nsomehow change things so calling <code>medium<\/code> will work and call it\nagain. The first option is theoretically clean but implies a lot of\nextra code--a whole extra implementation of whatever it was\n<code>medium<\/code> was supposed to do. And the further the stack unwinds,\nthe more work that needs to be redone. The second option--patching\nthings up and retrying--is tricky; for <code>high<\/code> to be able to\nchange the state of the world so a second call into <code>medium<\/code>\nwon't end up causing an error in <code>low<\/code>, it'd need an unseemly\nknowledge of the inner workings of both <code>medium<\/code> and <code>low<\/code>,\ncontrary to the notion that each function is a black box.<\/p><a name=\"the-lisp-way\"><h2>The Lisp Way<\/h2><\/a><p>Common Lisp's error handling system gives you a way out of this\nconundrum by letting you separate the code that actually recovers\nfrom an error from the code that decides how to recover. Thus, you\ncan put recovery code in low-level functions without committing to\nactually using any particular recovery strategy, leaving that\ndecision to code in high-level functions.<\/p><p>To get a sense of how this works, let's suppose you're writing an\napplication that reads some sort of textual log file, such as a Web\nserver's log. Somewhere in your application you'll have a function to\nparse the individual log entries. Let's assume you'll write a\nfunction, <code>parse-log-entry<\/code>, that will be passed a string\ncontaining the text of a single log entry and that is supposed to\nreturn a <code>log-entry<\/code> object representing the entry. This\nfunction will be called from a function, <code>parse-log-file<\/code>, that\nreads a complete log file and returns a list of objects representing\nall the entries in the file.<\/p><p>To keep things simple, the <code>parse-log-entry<\/code> function will not\nbe required to parse incorrectly formatted entries. It will, however,\nbe able to detect when its input is malformed. But what should it do\nwhen it detects bad input? In C you'd return a special value to\nindicate there was a problem. In Java or Python you'd throw or raise\nan exception. In Common Lisp, you signal a condition.<\/p><a name=\"conditions\" id=\"conditions\"><h2>Conditions<\/h2><\/a><p>A <i>condition<\/i> is an object whose class indicates the general nature\nof the condition and whose instance data carries information about\nthe details of the particular circumstances that lead to the\ncondition being signaled.<sup>3<\/sup> In this\nhypothetical log analysis program, you might define a condition\nclass, <code>malformed-log-entry-error<\/code>, that <code>parse-log-entry<\/code>\nwill signal if it's given data it can't parse.<\/p><p>Condition classes are defined with the <code><b>DEFINE-CONDITION<\/b><\/code> macro,\nwhich works essentially the same as <code><b>DEFCLASS<\/b><\/code> except that the\ndefault superclass of classes defined with <code><b>DEFINE-CONDITION<\/b><\/code> is\n<code><b>CONDITION<\/b><\/code> rather than <code><b>STANDARD-OBJECT<\/b><\/code>. Slots are specified\nin the same way, and condition classes can singly and multiply\ninherit from other classes that descend from <code><b>CONDITION<\/b><\/code>. But for\nhistorical reasons, condition classes aren't required to be instances\nof <code><b>STANDARD-OBJECT<\/b><\/code>, so some of the functions you use with\n<code><b>DEFCLASS<\/b><\/code>ed classes aren't required to work with conditions. In\nparticular, a condition's slots can't be accessed using\n<code><b>SLOT-VALUE<\/b><\/code>; you must specify either a <code>:reader<\/code> option or\nan <code>:accessor<\/code> option for any slot whose value you intend to\nuse. Likewise, new condition objects are created with\n<code><b>MAKE-CONDITION<\/b><\/code> rather than <code><b>MAKE-INSTANCE<\/b><\/code>.\n<code><b>MAKE-CONDITION<\/b><\/code> initializes the slots of the new condition based\non the <code>:initarg<\/code>s it's passed, but there's no way to further\ncustomize a condition's initialization, equivalent to\n<code><b>INITIALIZE-INSTANCE<\/b><\/code>.<sup>4<\/sup><\/p><p>When using the condition system for error handling, you should define\nyour conditions as subclasses of <code><b>ERROR<\/b><\/code>, a subclass of\n<code><b>CONDITION<\/b><\/code>. Thus, you might define\n<code>malformed-log-entry-error<\/code>, with a slot to hold the argument\nthat was passed to <code>parse-log-entry<\/code>, like this:<\/p><pre>(define-condition malformed-log-entry-error (error)\n  ((text :initarg :text :reader text)))<\/pre><a name=\"condition-handlers\"><h2>Condition Handlers<\/h2><\/a><p>In <code>parse-log-entry<\/code> you'll signal a\n<code>malformed-log-entry-error<\/code> if you can't parse the log entry.\nYou signal errors with the function <code><b>ERROR<\/b><\/code>, which calls the\nlower-level function <code><b>SIGNAL<\/b><\/code> and drops into the debugger if the\ncondition isn't handled. You can call <code><b>ERROR<\/b><\/code> two ways: you can\npass it an already instantiated condition object, or you can pass it\nthe name of the condition class and any initargs needed to construct\na new condition, and it will instantiate the condition for you. The\nformer is occasionally useful for resignaling an existing condition\nobject, but the latter is more concise. Thus, you could write\n<code>parse-log-entry<\/code> like this, eliding the details of actually\nparsing a log entry:<\/p><pre>(defun parse-log-entry (text)\n  (if (well-formed-log-entry-p text)\n    (make-instance 'log-entry ...)\n    (error 'malformed-log-entry-error :text text)))<\/pre><p>What happens when the error is signaled depends on the code above\n<code>parse-log-entry<\/code> on the call stack. To avoid landing in the\ndebugger, you must establish a <i>condition handler<\/i> in one of the\nfunctions leading to the call to <code>parse-log-entry<\/code>. When a\ncondition is signaled, the signaling machinery looks through a list\nof active condition handlers, looking for a handler that can handle\nthe condition being signaled based on the condition's class. Each\ncondition handler consists of a type specifier indicating what types\nof conditions it can handle and a function that takes a single\nargument, the condition. At any given moment there can be many active\ncondition handlers established at various levels of the call stack.\nWhen a condition is signaled, the signaling machinery finds the most\nrecently established handler whose type specifier is compatible with\nthe condition being signaled and calls its function, passing it the\ncondition object.<\/p><p>The handler function can then choose whether to handle the condition.\nThe function can decline to handle the condition by simply returning\nnormally, in which case control returns to the <code><b>SIGNAL<\/b><\/code> function,\nwhich will search for the next most recently established handler with\na compatible type specifier. To handle the condition, the function\nmust transfer control out of <code><b>SIGNAL<\/b><\/code> via a <i>nonlocal exit<\/i>. In\nthe next section, you'll see how a handler can choose where to\ntransfer control. However, many condition handlers simply want to\nunwind the stack to the place where they were established and then\nrun some code. The macro <code><b>HANDLER-CASE<\/b><\/code> establishes this kind of\ncondition handler. The basic form of a <code><b>HANDLER-CASE<\/b><\/code> is as\nfollows:<\/p><pre>(handler-case <i>expression<\/i>\n  <i>error-clause<\/i>*)<\/pre><p>where each <i>error-clause<\/i> is of the following form:<\/p><pre>(<i>condition-type<\/i> ([<i>var<\/i>]) <i>code<\/i>)<\/pre><p>If the <i>expression<\/i> returns normally, then its value is returned by\nthe <code><b>HANDLER-CASE<\/b><\/code>. The body of a <code><b>HANDLER-CASE<\/b><\/code> must be a\nsingle expression; you can use <code><b>PROGN<\/b><\/code> to combine several\nexpressions into a single form. If, however, the expression signals a\ncondition that's an instance of any of the <i>condition-type<\/i>s\nspecified in any <i>error-clause<\/i>, then the code in the appropriate\nerror clause is executed and its value returned by the\n<code><b>HANDLER-CASE<\/b><\/code>. The <i>var<\/i>, if included, is the name of the\nvariable that will hold the condition object when the handler code is\nexecuted. If the code doesn't need to access the condition object,\nyou can omit the variable name.<\/p><p>For instance, one way to handle the <code>malformed-log-entry-error<\/code>\nsignaled by <code>parse-log-entry<\/code> in its caller,\n<code>parse-log-file<\/code>, would be to skip the malformed entry. In the\nfollowing function, the <code><b>HANDLER-CASE<\/b><\/code> expression will either\nreturn the value returned by <code>parse-log-entry<\/code> or return\n<code><b>NIL<\/b><\/code> if a <code>malformed-log-entry-error<\/code> is signaled. (The\n<code>it<\/code> in the <code><b>LOOP<\/b><\/code> clause <code>collect it<\/code> is another\n<code><b>LOOP<\/b><\/code> keyword, which refers to the value of the most recently\nevaluated conditional test, in this case the value of <code>entry<\/code>.)<\/p><pre>(defun parse-log-file (file)\n  (with-open-file (in file :direction :input)\n    (loop for text = (read-line in nil nil) while text\n       for entry = (handler-case (parse-log-entry text)\n                     (malformed-log-entry-error () nil))\n       when entry collect it)))<\/pre><p>When <code>parse-log-entry<\/code> returns normally, its value will be\nassigned to <code>entry<\/code> and collected by the <code><b>LOOP<\/b><\/code>. But if\n<code>parse-log-entry<\/code> signals a <code>malformed-log-entry-error<\/code>,\nthen the error clause will return <code><b>NIL<\/b><\/code>, which won't be collected.<\/p><p>JAVA-STYLE EXCEPTON HANDLING<\/p><p>This version of <code>parse-log-file<\/code> has one serious deficiency:\nit's doing too much. As its name suggests, the job of\n<code>parse-log-file<\/code> is to parse the file and produce a list of\n<code>log-entry<\/code> objects; if it can't, it's not its place to decide\nwhat to do instead. What if you want to use <code>parse-log-file<\/code> in\nan application that wants to tell the user that the log file is\ncorrupted or one that wants to recover from malformed entries by\nfixing them up and re-parsing them? Or maybe an application is fine\nwith skipping them but only until a certain number of corrupted\nentries have been seen.<\/p><p>You could try to fix this problem by moving the <code><b>HANDLER-CASE<\/b><\/code> to\na higher-level function. However, then you'd have no way to implement\nthe current policy of skipping individual entries--when the error was\nsignaled, the stack would be unwound all the way to the higher-level\nfunction, abandoning the parsing of the log file altogether. What you\nwant is a way to provide the current recovery strategy without\nrequiring that it always be used.<\/p><a name=\"restarts\" id=\"restarts\"><h2>Restarts<\/h2><\/a><p>The condition system lets you do this by splitting the error handling\ncode into two parts. You place code that actually recovers from\nerrors into <i>restarts<\/i>, and condition handlers can then handle a\ncondition by invoking an appropriate restart. You can place restart\ncode in mid- or low-level functions, such as <code>parse-log-file<\/code> or\n<code>parse-log-entry<\/code>, while moving the condition handlers into the\nupper levels of the application.<\/p><p>To change <code>parse-log-file<\/code> so it establishes a restart instead\nof a condition handler, you can change the <code><b>HANDLER-CASE<\/b><\/code> to a\n<code><b>RESTART-CASE<\/b><\/code>. The form of <code><b>RESTART-CASE<\/b><\/code> is quite similar to\na <code><b>HANDLER-CASE<\/b><\/code> except the names of restarts are just names, not\nnecessarily the names of condition types. In general, a restart name\nshould describe the action the restart takes. In\n<code>parse-log-file<\/code>, you can call the restart <code>skip-log-entry<\/code>\nsince that's what it does. The new version will look like this:<\/p><pre>(defun parse-log-file (file)\n  (with-open-file (in file :direction :input)\n    (loop for text = (read-line in nil nil) while text\n       for entry = (restart-case (parse-log-entry text)\n                     (skip-log-entry () nil))\n       when entry collect it)))<\/pre><p>If you invoke this version of <code>parse-log-file<\/code> on a log file\ncontaining corrupted entries, it won't handle the error directly;\nyou'll end up in the debugger. However, there among the various\nrestarts presented by the debugger will be one called\n<code>skip-log-entry<\/code>, which, if you choose it, will cause\n<code>parse-log-file<\/code> to continue on its way as before. To avoid\nending up in the debugger, you can establish a condition handler that\ninvokes the <code>skip-log-entry<\/code> restart automatically.<\/p><p>The advantage of establishing a restart rather than having\n<code>parse-log-file<\/code> handle the error directly is it makes\n<code>parse-log-file<\/code> usable in more situations. The higher-level\ncode that invokes <code>parse-log-file<\/code> doesn't have to invoke the\n<code>skip-log-entry<\/code> restart. It can choose to handle the error at a\nhigher level. Or, as I'll show in the next section, you can add\nrestarts to <code>parse-log-entry<\/code> to provide other recovery\nstrategies, and then condition handlers can choose which strategy\nthey want to use.<\/p><p>But before I can talk about that, you need to see how to set up a\ncondition handler that will invoke the <code>skip-log-entry<\/code> restart.\nYou can set up the handler anywhere in the chain of calls leading to\n<code>parse-log-file<\/code>. This may be quite high up in your application,\nnot necessarily in <code>parse-log-file<\/code>'s direct caller. For\ninstance, suppose the main entry point to your application is a\nfunction, <code>log-analyzer<\/code>, that finds a bunch of logs and\nanalyzes them with the function <code>analyze-log<\/code>, which eventually\nleads to a call to <code>parse-log-file<\/code>. Without any error handling,\nit might look like this:<\/p><pre>(defun log-analyzer ()\n  (dolist (log (find-all-logs))\n    (analyze-log log)))<\/pre><p>The job of <code>analyze-log<\/code> is to call, directly or indirectly,\n<code>parse-log-file<\/code> and then do something with the list of log\nentries returned. An extremely simple version might look like this:<\/p><pre>(defun analyze-log (log)\n  (dolist (entry (parse-log-file log))\n    (analyze-entry entry)))<\/pre><p>where the function <code>analyze-entry<\/code> is presumably responsible for\nextracting whatever information you care about from each log entry and\nstashing it away somewhere.<\/p><p>Thus, the path from the top-level function, <code>log-analyzer<\/code>, to\n<code>parse-log-entry<\/code>, which actually signals an error, is as\nfollows:<\/p><p><img class=\"figure\" src=\"http:\/\/www.gigamonkeys.com\/book\/figures\/restart-call-stack.png\" alt=\"image\" \/><\/p><p>Assuming you always want to skip malformed log entries, you could\nchange this function to establish a condition handler that invokes\nthe <code>skip-log-entry<\/code> restart for you. However, you can't use\n<code><b>HANDLER-CASE<\/b><\/code> to establish the condition handler because then the\nstack would be unwound to the function where the <code><b>HANDLER-CASE<\/b><\/code>\nappears. Instead, you need to use the lower-level macro\n<code><b>HANDLER-BIND<\/b><\/code>. The basic form of <code><b>HANDLER-BIND<\/b><\/code> is as follows:<\/p><pre>(handler-bind (<i>binding<\/i>*) <i>form<\/i>*)<\/pre><p>where each binding is a list of a condition type and a handler\nfunction of one argument. After the handler bindings, the body of the\n<code><b>HANDLER-BIND<\/b><\/code> can contain any number of forms. Unlike the handler\ncode in <code><b>HANDLER-CASE<\/b><\/code>, the handler code must be a function\nobject, and it must accept a single argument. A more important\ndifference between <code><b>HANDLER-BIND<\/b><\/code> and <code><b>HANDLER-CASE<\/b><\/code> is that\nthe handler function bound by <code><b>HANDLER-BIND<\/b><\/code> will be run without\nunwinding the stack--the flow of control will still be in the call to\n<code>parse-log-entry<\/code> when this function is called. The call to\n<code><b>INVOKE-RESTART<\/b><\/code> will find and invoke the most recently bound\nrestart with the given name. So you can add a handler to\n<code>log-analyzer<\/code> that will invoke the <code>skip-log-entry<\/code>\nrestart established in <code>parse-log-file<\/code> like this:<sup>5<\/sup><\/p><pre>(defun log-analyzer ()\n  (handler-bind ((malformed-log-entry-error\n                  #'(lambda (c)\n                      (invoke-restart 'skip-log-entry))))\n    (dolist (log (find-all-logs))\n      (analyze-log log))))<\/pre><p>In this <code><b>HANDLER-BIND<\/b><\/code>, the handler function is an anonymous\nfunction that invokes the restart <code>skip-log-entry<\/code>. You could\nalso define a named function that does the same thing and bind it\ninstead. In fact, a common practice when defining a restart is to\ndefine a function, with the same name and taking a single argument,\nthe condition, that invokes the eponymous restart. Such functions are\ncalled <i>restart functions<\/i>. You could define a restart function for\n<code>skip-log-entry<\/code> like this:<\/p><pre>(defun skip-log-entry (c)\n  (invoke-restart 'skip-log-entry))<\/pre><p>Then you could change the definition of <code>log-analyzer<\/code> to this:<\/p><pre>(defun log-analyzer ()\n  (handler-bind ((malformed-log-entry-error #'skip-log-entry))\n    (dolist (log (find-all-logs))\n      (analyze-log log))))<\/pre><p>As written, the <code>skip-log-entry<\/code> restart function assumes that a\n<code>skip-log-entry<\/code> restart has been established. If a\n<code>malformed-log-entry-error<\/code> is ever signaled by code called from\n<code>log-analyzer<\/code> without a <code>skip-log-entry<\/code> having been\nestablished, the call to <code><b>INVOKE-RESTART<\/b><\/code> will signal a\n<code><b>CONTROL-ERROR<\/b><\/code> when it fails to find the <code>skip-log-entry<\/code>\nrestart. If you want to allow for the possibility that a\n<code>malformed-log-entry-error<\/code> might be signaled from code that\ndoesn't have a <code>skip-log-entry<\/code> restart established, you could\nchange the <code>skip-log-entry<\/code> function to this:<\/p><pre>(defun skip-log-entry (c)\n  (let ((restart (find-restart 'skip-log-entry)))\n    (when restart (invoke-restart restart))))<\/pre><p><code><b>FIND-RESTART<\/b><\/code> looks for a restart with a given name and returns\nan object representing the restart if the restart is found and\n<code><b>NIL<\/b><\/code> if not. You can invoke the restart by passing the restart\nobject to <code><b>INVOKE-RESTART<\/b><\/code>. Thus, when <code>skip-log-entry<\/code> is\nbound with <code><b>HANDLER-BIND<\/b><\/code>, it will handle the condition by\ninvoking the <code>skip-log-entry<\/code> restart if one is available and\notherwise will return normally, giving other condition handlers,\nbound higher on the stack, a chance to handle the condition.<\/p><a name=\"providing-multiple-restarts\"><h2>Providing Multiple Restarts<\/h2><\/a><p>Since restarts must be explicitly invoked to have any effect, you can\ndefine multiple restarts, each providing a different recovery\nstrategy. As I mentioned earlier, not all log-parsing applications\nwill necessarily want to skip malformed entries. Some applications\nmight want <code>parse-log-file<\/code> to include a special kind of object\nrepresenting malformed entries in the list of <code>log-entry<\/code>\nobjects; other applications may have some way to repair a malformed\nentry and may want a way to pass the fixed entry back to\n<code>parse-log-entry<\/code>.<\/p><p>To allow more complex recovery protocols, restarts can take arbitrary\narguments, which are passed in the call to <code><b>INVOKE-RESTART<\/b><\/code>. You\ncan provide support for both the recovery strategies I just mentioned\nby adding two restarts to <code>parse-log-entry<\/code>, each of which takes\na single argument. One simply returns the value it's passed as the\nreturn value of <code>parse-log-entry<\/code>, while the other tries to\nparse its argument in the place of the original log entry.<\/p><pre>(defun parse-log-entry (text)\n  (if (well-formed-log-entry-p text)\n    (make-instance 'log-entry ...)\n    (restart-case (error 'malformed-log-entry-error :text text)\n      (use-value (value) value)\n      (reparse-entry (fixed-text) (parse-log-entry fixed-text)))))<\/pre><p>The name <code><b>USE-VALUE<\/b><\/code> is a standard name for this kind of restart.\nCommon Lisp defines a restart function for <code><b>USE-VALUE<\/b><\/code> similar to\nthe <code>skip-log-entry<\/code> function you just defined. So, if you\nwanted to change the policy on malformed entries to one that created\nan instance of <code>malformed-log-entry<\/code>, you could change\n<code>log-analyzer<\/code> to this (assuming the existence of a\n<code>malformed-log-entry<\/code> class with a <code>:text<\/code> initarg):<\/p><pre>(defun log-analyzer ()\n  (handler-bind ((malformed-log-entry-error\n                  #'(lambda (c)\n                      (use-value\n                       (make-instance 'malformed-log-entry :text (text c))))))\n    (dolist (log (find-all-logs))\n      (analyze-log log))))<\/pre><p>You could also have put these new restarts into <code>parse-log-file<\/code>\ninstead of <code>parse-log-entry<\/code>. However, you generally want to put\nrestarts in the lowest-level code possible. It wouldn't, though, be\nappropriate to move the <code>skip-log-entry<\/code> restart into\n<code>parse-log-entry<\/code> since that would cause <code>parse-log-entry<\/code>\nto sometimes return normally with <code><b>NIL<\/b><\/code>, the very thing you\nstarted out trying to avoid. And it'd be an equally bad idea to\nremove the <code>skip-log-entry<\/code> restart on the theory that the\ncondition handler could get the same effect by invoking the\n<code>use-value<\/code> restart with <code><b>NIL<\/b><\/code> as the argument; that would\nrequire the condition handler to have intimate knowledge of how the\n<code>parse-log-file<\/code> works. As it stands, the <code>skip-log-entry<\/code>\nis a properly abstracted part of the log-parsing API.<\/p><a name=\"other-uses-for-conditions\"><h2>Other Uses for Conditions<\/h2><\/a><p>While conditions are mainly used for error handling, they can be used\nfor other purposes--you can use conditions, condition handlers, and\nrestarts to build a variety of protocols between low- and high-level\ncode. The key to understanding the potential of conditions is to\nunderstand that merely signaling a condition has no effect on the\nflow of control.<\/p><p>The primitive signaling function <code><b>SIGNAL<\/b><\/code> implements the mechanism\nof searching for an applicable condition handler and invoking its\nhandler function. The reason a handler can decline to handle a\ncondition by returning normally is because the call to the handler\nfunction is just a regular function call--when the handler returns,\ncontrol passes back to <code><b>SIGNAL<\/b><\/code>, which then looks for another,\nless recently bound handler that can handle the condition. If\n<code><b>SIGNAL<\/b><\/code> runs out of condition handlers before the condition is\nhandled, it also returns normally.<\/p><p>The <code><b>ERROR<\/b><\/code> function you've been using calls <code><b>SIGNAL<\/b><\/code>. If the\nerror is handled by a condition handler that transfers control via\n<code><b>HANDLER-CASE<\/b><\/code> or by invoking a restart, then the call to\n<code><b>SIGNAL<\/b><\/code> never returns. But if <code><b>SIGNAL<\/b><\/code> returns, <code><b>ERROR<\/b><\/code>\ninvokes the debugger by calling the function stored in\n<code><b>*DEBUGGER-HOOK*<\/b><\/code>. Thus, a call to <code><b>ERROR<\/b><\/code> can never return\nnormally; the condition must be handled either by a condition handler\nor in the debugger.<\/p><p>Another condition signaling function, <code><b>WARN<\/b><\/code>, provides an example\nof a different kind of protocol built on the condition system. Like\n<code><b>ERROR<\/b><\/code>, <code><b>WARN<\/b><\/code> calls <code><b>SIGNAL<\/b><\/code> to signal a condition. But if\n<code><b>SIGNAL<\/b><\/code> returns, <code><b>WARN<\/b><\/code> doesn't invoke the debugger--it prints\nthe condition to <code><b>*ERROR-OUTPUT*<\/b><\/code> and returns <code><b>NIL<\/b><\/code>, allowing\nits caller to proceed. <code><b>WARN<\/b><\/code> also establishes a restart,\n<code><b>MUFFLE-WARNING<\/b><\/code>, around the call to <code><b>SIGNAL<\/b><\/code> that can be used\nby a condition handler to make <code><b>WARN<\/b><\/code> return without printing\nanything. The restart function <code><b>MUFFLE-WARNING<\/b><\/code> finds and invokes\nits eponymous restart, signaling a <code><b>CONTROL-ERROR<\/b><\/code> if no such\nrestart is available. Of course, a condition signaled with <code><b>WARN<\/b><\/code>\ncould also be handled in some other way--a condition handler could\n\"promote\" a warning to an error by handling it as if it were an\nerror.<\/p><p>For instance, in the log-parsing application, if there were ways a\nlog entry could be slightly malformed but still parsable, you could\nwrite <code>parse-log-entry<\/code> to go ahead and parse the slightly\ndefective entries but to signal a condition with <code><b>WARN<\/b><\/code> when it\ndid. Then the larger application could choose to let the warning\nprint, to muffle the warning, or to treat the warning like an error,\nrecovering the same way it would from a\n<code>malformed-log-entry-error<\/code>.<\/p><p>A third error-signaling function, <code><b>CERROR<\/b><\/code>, provides yet another\nprotocol. Like <code><b>ERROR<\/b><\/code>, <code><b>CERROR<\/b><\/code> will drop you into the\ndebugger if the condition it signals isn't handled. But like\n<code><b>WARN<\/b><\/code>, it establishes a restart before it signals the condition.\nThe restart, <code><b>CONTINUE<\/b><\/code>, causes <code><b>CERROR<\/b><\/code> to return normally--if\nthe restart is invoked by a condition handler, it will keep you out\nof the debugger altogether. Otherwise, you can use the restart once\nyou're in the debugger to resume the computation immediately after\nthe call to <code><b>CERROR<\/b><\/code>. The function <code><b>CONTINUE<\/b><\/code> finds and invokes\nthe <code><b>CONTINUE<\/b><\/code> restart if it's available and returns <code><b>NIL<\/b><\/code>\notherwise.<\/p><p>You can also build your own protocols on <code><b>SIGNAL<\/b><\/code>--whenever\nlow-level code needs to communicate information back up the call\nstack to higher-level code, the condition mechanism is a reasonable\nmechanism to use. But for most purposes, one of the standard error or\nwarning protocols should suffice.<\/p><p>You'll use the condition system in future practical chapters, both\nfor regular error handling and, in Chapter 25, to help in handling a\ntricky corner case of parsing ID3 files. Unfortunately, it's the fate\nof error handling to always get short shrift in programming\ntexts--proper error handling, or lack thereof, is often the biggest\ndifference between illustrative code and hardened, production-quality\ncode. The trick to writing the latter has more to do with adopting a\nparticularly rigorous way of thinking about software than with the\ndetails of any particular programming language constructs. That said,\nif your goal is to write that kind of software, you'll find the\nCommon Lisp condition system is an excellent tool for writing robust\ncode and one that fits quite nicely into Common Lisp's incremental\ndevelopment style.<\/p><p>Writing Robust Software<\/p><p>In the next chapter I'll give a quick overview of some of the 25\nspecial operators you haven't had a chance to use yet, at least not\ndirectly.\n<\/p><hr \/><div class=\"notes\"><p><sup>1<\/sup><i>Throws<\/i> or <i>raises<\/i> an exception in\nJava\/Python terms<\/p><p><sup>2<\/sup><i>Catches<\/i> the\nexception in Java\/Python terms<\/p><p><sup>3<\/sup>In this respect, a condition is a lot\nlike an exception in Java or Python except not all conditions\nrepresent an error or <i>exceptional<\/i> situation.<\/p><p><sup>4<\/sup>In some Common Lisp implementations,\nconditions are defined as subclasses of <code><b>STANDARD-OBJECT<\/b><\/code>, in\nwhich case <code><b>SLOT-VALUE<\/b><\/code>, <code><b>MAKE-INSTANCE<\/b><\/code>, and\n<code><b>INITIALIZE-INSTANCE<\/b><\/code> will work, but it's not portable to rely on\nit.<\/p><p><sup>5<\/sup>The\ncompiler may complain if the parameter is never used. You can silence\nthat warning by adding a declaration <code>(declare (ignore c))<\/code> as\nthe first expression in the <code><b>LAMBDA<\/b><\/code> body.<\/p><\/div>","created_at":"2019-04-04T23:46:59+0000","updated_at":"2019-04-04T23:46:59+0000","annotations":[],"mimetype":"text\/html","reading_time":24,"domain_name":"www.gigamonkeys.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":895,"title":"provide-input-to-the-compilation-buffer.html","url":"http:\/\/endlessparentheses.com\/provide-input-to-the-compilation-buffer.html","created_at":"2019-04-04T15:06:40+0000","updated_at":"2019-04-04T15:06:42+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"endlessparentheses.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":894,"title":"Nomad by HashiCorp","url":"https:\/\/www.nomadproject.io\/","content":"Nomad by HashiCorp\n    \n    \n    \n<header><div class=\"container hero\"><div class=\"row\"><div class=\"col-md-offset-2 col-md-8\"><a class=\"button primary\" href=\"https:\/\/www.nomadproject.io\/intro\/index.html\">Get Started<\/a><a class=\"button\" href=\"https:\/\/www.nomadproject.io\/downloads.html\">Download 0.8.7<\/a><\/div><\/div><\/div>\n<\/header><section id=\"features\" class=\"marketing green\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-12\"><h2>Simple and Lightweight<\/h2><p class=\"lead\">\n          HashiCorp Nomad is a single binary that schedules applications and services on\n          Linux, Windows, and Mac. It is an open source scheduler that uses a\n          declarative job file for scheduling virtualized, containerized, and\n          standalone applications.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-6\"><h3>1. Declare Jobs<\/h3><p>\n          Users compose and submit high-level job files. Nomad handles the\n          scheduling and upgrading of the applications over time.\n        <\/p><p>\n          This flexibility makes it easy to deploy one container, dozens of\n          containers, or even <a href=\"https:\/\/www.hashicorp.com\/c1m\/\">millions<\/a>.\n        <\/p><\/div><div class=\"col-sm-6\"><h3>2. Plan Changes<\/h3><p>\n          With built-in dry-run execution, Nomad shows what scheduling decisions\n          it will take before it takes them. Operators can approve or deny these\n          changes to create a safe and reproducible workflow.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-6\"><h3>3. Run Applications<\/h3><p>\n          Nomad runs applications and ensures they keep running in failure\n          scenarios. In addition to long-running services, Nomad can schedule\n          batch jobs, distributed cron jobs, and parameterized jobs.\n        <\/p><\/div><div class=\"col-sm-6\"><h3>4. Monitor Progress<\/h3><p>\n          Stream logs, send signals, and interact with the file system of\n          scheduled applications. These operator-friendly commands bring the\n          familiar debugging tools to a scheduled world.\n        <\/p><\/div><\/div><\/div><\/section><section id=\"hybrid-multi-cloud\" class=\"marketing\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-12\">Combine<h2>Hybrid &amp; Multi Cloud<\/h2><p class=\"lead\">\n          As more organizations adopt cloud technologies, the desire to run in\n          multiple datacenters and multiple regions becomes critical. Nomad can\n          span public and private clouds and treat all your infrastructure as a\n          pool of resources.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-6\"><h3>Hybrid Cloud<\/h3><p>\n          Nomad bridges the gap between the public and private cloud, creating a\n          unified interface for developers to run any application on any\n          infrastructure. Easily run services across your internal private cloud\n          and external public cloud transparently to application developers.\n        <\/p><\/div><div class=\"col-sm-6\"><h3>Multi Cloud<\/h3><p>\n          As more organizations adopt cloud technologies, the desire to run\n          applications simultaneously across multiple clouds increases. Nomad\n          combines the power of all cloud providers into a single, unified\n          interface for deploying applications and running services.\n        <\/p><\/div><\/div><\/div><\/section><section id=\"simplify\" class=\"marketing green\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-12\">Maintain<h2>Simplify Operations<\/h2><p class=\"lead\">\n          Nomad simplifies operations by supporting several deployment\n          strategies to safely upgrade jobs, automatically handling machine\n          failures, and providing a single workflow to deploy applications.\n        <\/p><\/div><\/div><div class=\"row\"><p><\/p><h3>Rolling Deploys<\/h3><\/div><div class=\"row\"><div class=\"col-sm-5\" id=\"feature-rolling-deploys\">rolling...9\/12done!12\/12<\/div><div class=\"col-sm-7\"><p>\n          In order to update an application while reducing downtime, Nomad\n          provides a built-in mechanism for rolling upgrades. Operators specify\n          the rate at which they would like to upgrade their service in the\n          declarative job file, plan the changes, and submit the new job. Nomad\n          then updates the service to the newest version using task health and\n          Consul health check information to ensure replacement allocations are\n          healthy. With full control over the update strategy, rolling upgrades\n          make application deployments simple and safe.\n        <\/p><\/div><\/div><div class=\"row\"><p><\/p><h3>Blue\/Green Deployments<\/h3><\/div><div class=\"row\"><div class=\"col-sm-7\"><p>\n          Nomad supports native blue\/green and canary deployments through the\n          declarative job file syntax. Instead of doing a rolling upgrade of the\n          existing allocations, the new version of the group is deployed along\n          side the existing set. While this duplicates the resources required\n          during the upgrade process, it allows very safe deployments as the\n          original version of the group is untouched till the operator promotes\n          the new version.\n        <\/p><\/div><div class=\"col-sm-5\">web-app:1.0group.blueweb-app:1.1group.green8\/102\/10<\/div><\/div><div class=\"row\"><p><\/p><h3>Automatic Machine Failures<\/h3><\/div><div class=\"row\"><div class=\"col-sm-7\"><p>\n          Because it knows the state of all nodes and applications, if a\n          node becomes unhealthy, Nomad will automatically reschedule the\n          applications that were previously running on the unhealthy host\n          onto a new, healthy host if capacity exists. As new clients are\n          added to the cluster, they can share the responsibility of\n          running applications and services.\n        <\/p><\/div><\/div><\/div><\/section><section id=\"flexible-workloads\" class=\"marketing\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-12\">Support<h2>Flexible Workloads<\/h2><p class=\"lead\">\n          In addition to supporting Linux, Windows, and Mac, Nomad has\n          extensible support for containerized, virtualized, and standalone\n          applications. Easily start Docker containers, VMs, or application\n          runtimes like Java.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-6\"><h3>Diverse Technologies<\/h3><p>\n          By supporting a wide range of technologies such as Docker, rkt, and\n          LXC, Nomad does not force you into a single technology. Easily use\n          multiple container or virtualization runtimes simultaneously.\n          Applications with their own runtime such as Java or Go can be\n          scheduled directly without the overhead of a container.\n        <\/p><\/div><div class=\"col-sm-6\"><h3>Multi OS<\/h3><p>\n          The Nomad client runs on Linux, Windows, and Mac OS. This diverse\n          operating system support enables using the same application scheduler\n          for all your scheduling and runtime needs.\n        <\/p><\/div><\/div><\/div><\/section><section id=\"binpacking\" class=\"marketing black\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-12\">Maximize<h2>Increase Utilization &amp; Reduce Costs<\/h2><p class=\"lead\">\n          Nomad uses bin packing to optimize application placement onto servers\n          to maximize resource utilization, increase density, and help reduce\n          costs.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-6\"><h3>Increase Density<\/h3><p>\n          Nomad places applications and services based on a bin packing\n          algorithm, which is used to optimize the resource utilization and\n          density of applications. Nomad automatically augments the bin-packing\n          algorithm with anti-affinity rules to avoid colocating instances of\n          the same service to reduce the probability of correlated failures.\n        <\/p><\/div><div class=\"col-sm-6\"><h3>Reduce Costs<\/h3><p>\n          By maximizing resource utilization, Nomad can help remove unused or\n          under-utilized machines in the cluster, reducing overall costs.\n          Additionally, Nomad enables organizations to utilize larger, more\n          cost-effective compute devices which can be less expensive at scale.\n        <\/p><\/div><\/div><\/div><\/section><section id=\"examples\" class=\"marketing gray\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-12\"><h2>Examples<\/h2><h3>Create and Submit Jobs<\/h3><p>\n          Easily create, validate, submit, and check the status of jobs using\n          the Nomad CLI.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-12\"><div class=\"terminal\"><p>admin@hashicorp.com:\n              nomad init\n            \n            Example job file written to example.nomad\n            \n            \n              admin@hashicorp.com:\n              nomad job validate example.nomad\n            \n            Job validation successful\n            \n            \n              admin@hashicorp.com:\n              nomad job run example.nomad\n            \n            ==&gt; Monitoring evaluation \"feb23392\"\n                Evaluation triggered by job \"example\"\n                Evaluation within deployment: \"7ffa7ed1\"\n                Allocation \"a2f24fcf\" created: node \"3714a84a\", group \"cache\"\n                Evaluation status changed: \"pending\" -&gt; \"complete\"\n            ==&gt; Evaluation \"feb23392\" finished with status \"complete\"\n            \n            \n              admin@hashicorp.com:\n              nomad status\n            \n            ID       Type     Priority  Status   Submit Date\n            example  service  50        running  04\/03\/19 14:50:49 UTC<\/p><\/div><\/div><\/div><div class=\"row\"><div class=\"col-sm-12\"><h3>Plan Changes<\/h3><p>\n          Perform a dry-run to check application placement, scheduling\n          decisions, and visualize failures before they happen.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-12\"><div class=\"terminal\"><p>admin@hashicorp.com:\n              nomad job plan example.nomad\n            \n            +\/- Job: \"example\"\n            +\/- Task Group: \"cache\" (2 create, 1 in-place update)\n              +\/- Count: \"1\" =&gt; \"3\" (forces create)\n                  Task: \"redis\"\n            \n            Scheduler dry-run:\n            - All tasks successfully allocated.\n            \n            Job Modify Index: 7\n            \n            \n              admin@hashicorp.com:\n              nomad job run -check-index 7 example.nomad\n            \n            ==&gt; Monitoring evaluation \"e338a6ae\"\n                Evaluation triggered by job \"example\"\n                Evaluation within deployment: \"43310206\"\n                Allocation \"991a0d89\" created: node \"c5916a99\", group \"cache\"\n                Allocation \"daee8be1\" created: node \"c5916a99\", group \"cache\"\n                Allocation \"83f29312\" modified: node \"c5916a99\", group \"cache\"\n                Evaluation status changed: \"pending\" -&gt; \"complete\"\n            ==&gt; Evaluation \"e338a6ae\" finished with status \"complete\"<\/p><\/div><\/div><\/div><div class=\"row\"><div class=\"col-sm-12\"><h3>View Application Logs<\/h3><p>\n          Stream application logs directly in the terminal to help analyze and\n          debug applications.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"col-sm-12\"><div class=\"terminal\"><p>admin@hashicorp.com:\n              nomad alloc logs -tail -job example\n            \n            03 Apr 14:50:49 # Server started, Redis version 3.2.9\n            03 Apr 14:50:49 * The server is now ready to accept connections on port 6379<\/p><\/div><\/div><\/div><div class=\"row\"><div class=\"col-sm-12 col-lg-8 col-lg-offset-2\"><p class=\"lead\" style=\"text-align: center;\">\n          The introduction contains a walkthrough guide, glossary, and a range\n          of examples for learning and experimenting with Nomad.\n        <\/p><p style=\"text-align: center;\">\n          <a href=\"https:\/\/www.nomadproject.io\/intro\/index.html\" class=\"button primary\">Get Started<\/a>\n        <\/p><\/div><\/div><\/div><\/section>","created_at":"2019-04-04T12:45:51+0000","updated_at":"2019-04-04T12:45:51+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.nomadproject.io","preview_picture":"https:\/\/www.nomadproject.io\/assets\/images\/og-image-2b63cc11.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":893,"title":"Best way to model Graph data in postgresql","url":"https:\/\/stackoverflow.com\/questions\/20776718\/best-way-to-model-graph-data-in-postgresql","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>How would one go about storing and querying sparse directed or undirected graphs in Postgresql. There is something like <a href=\"http:\/\/pgfoundry.org\/projects\/pggraph\" rel=\"noreferrer\">pggraph<\/a>, but that is still in planning stage.<\/p>\n<p>I realize dedicated graph databases like Neo4J are best suited for this. However is there way to implement same within Postgresql, by using extension or a data type, which would avoid adding another database engine.dtata<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/20776718\/revisions\" title=\"show all edits to this post\">edited Dec 25 '13 at 21:06<\/a>&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        jethar&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Dec 25 '13 at 20:59&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/631775\/jethar\">jethar<\/a>jethar&#13;\n        <div class=\"-flair\">&#13;\n            86021217&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"5\">&#13;\n                                5 Answers&#13;\n                            5&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Question is, I think, way too vague and broad to give a precise answer...<\/p>\n<p>In essence, though, there are some techniques to efficiently query graph data within an SQL database, that apply to highly specialized scenarios. You could opt to maintain a GRIPP index, for instance, if your interests lie in finding shortest paths. (It basically works a bit like pre-ordered tree index, applied to graphs.) To the best of my knowledge, none of these techniques are standardized yet.<\/p>\n<p>With that being said, and seeing your comment that mentions social networks, the odds are that each of them will be overkill. If your interest primarily lies in fetching data related to a user's friends, or something equivalent in the sense that it amounts to querying a node's neighborhood, the number of nodes you'll need to traverse in joins is so tiny that there is no need for specialized tools, data structures, etc.: simply use recursive CTEs.<\/p>\n<p><a href=\"http:\/\/www.postgresql.org\/docs\/current\/static\/queries-with.html\" rel=\"noreferrer\">http:\/\/www.postgresql.org\/docs\/current\/static\/queries-with.html<\/a><\/p>\n<p>For optimal performance when using the latter, shift as many <code>where<\/code> conditions within the <code>with (...)<\/code> part of the query, so as to eliminate nodes early.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/20782576\/revisions\" title=\"show all edits to this post\">edited Dec 26 '13 at 9:57<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Dec 26 '13 at 9:47&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/417194\/denis-de-bernardy\">Denis de Bernardy<\/a>Denis de Bernardy&#13;\n        <div class=\"-flair\">&#13;\n            58.1k890121&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Use PostgreSQL for the underlying storage and use networkX or iGraph via PL\/Python for the processing engine.<\/p>\n<p>In their book \"<a href=\"http:\/\/graphdatabases.com\" rel=\"noreferrer\">Graph Databases<\/a>\", Ian Robinson, Jim Webber, and Emil Eifrem make a distinction between the underlying storage and the processing engine. If you look at the answer I followed in a recent problem (see <a href=\"https:\/\/stackoverflow.com\/questions\/27967093\/how-to-aggregate-matching-pairs-into-connected-components-in-python\">here<\/a>), you will see that I'm using PostgreSQL for the underlying storage and networkX as  the processing engine. The performance gain relative to my original solution was huge (and similar to the ones described in the \"Graph Databases\" book) and implementing it was very easy.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/27997453\/revisions\" title=\"show all edits to this post\">edited May 23 '17 at 12:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jan 17 '15 at 7:57&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/745737\/ian-gow\">Ian Gow<\/a>Ian Gow&#13;\n        <div class=\"-flair\">&#13;\n            1,25211323&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Since the question is generic, I would add a solution that can work for mostly planar graphs like street networks, PostgreSQL offers an excellent solution through Postgis <a href=\"https:\/\/postgis.net\/docs\/Topology.html\" rel=\"nofollow noreferrer\">topology<\/a>. \nPostgis topology stores geometries as edges, nodes and faces and their relative relationships. This means that from the geometry of a street network you can select edges and their starting and ending nodes and from this easily build a graph in the processing engine of your choice (networkx or graph-tool for Pyhton are examples). <\/p>\n<p>As I said, though, Postgresql\/Postgis topology works when we want to study geometries like street networks from the perspective of graph analysis.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jun 17 '18 at 12:22&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5276212\/duccio-a\">Duccio A<\/a>Duccio A&#13;\n        <div class=\"-flair\">&#13;\n            434313&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>At this point I would recommend experimenting with <a href=\"https:\/\/bitnine.net\" rel=\"nofollow noreferrer\">AgensGraph, a promising multimodel distribution of PostgreSQL<\/a> that offers first class graph databases and queries from both SQL and Cypher. Note that it is a full server, and not an extension like PostGIS, although PostgreSQL extensions can be added to it.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/52026777\/revisions\" title=\"show all edits to this post\">edited Sep 25 '18 at 2:04<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 26 '18 at 13:35&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5281749\/eron-lloyd\">Eron Lloyd<\/a>Eron Lloyd&#13;\n        &#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Use AgensGraph from bitnine.net\n<a href=\"https:\/\/bitnine.net\/\" rel=\"nofollow noreferrer\">https:\/\/bitnine.net\/<\/a><\/p>\n<p>and if you want to visualise your graph you can use AgensBrowser also from bitnine<\/p>\n<p>Thanks\nJoe<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 21 at 16:07&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/10945822\/joe\">Joe<\/a>Joe&#13;\n        &#13;<\/div>","created_at":"2019-04-03T20:51:29+0000","updated_at":"2019-04-03T20:51:29+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":892,"title":"Get Started","url":"https:\/\/neanderthal.uncomplicate.org\/articles\/getting_started.html","content":"<p>Neanderthal\u2019s default option is to use use native libraries, so it is very important that you do not skip any part of this guide.<\/p><h2 id=\"how-to-get-started\">How to Get Started<\/h2><ul><li>Walk through this guide, set up your development environment, and try the examples.<\/li>\n<li>Familiarize yourself with Neanderthal\u2019s <a href=\"https:\/\/neanderthal.uncomplicate.org\/articles\/guides.html\">more detailed tutorials<\/a> and <a href=\"https:\/\/neanderthal.uncomplicate.org\/codox\">API documentation<\/a>.<\/li>\n<\/ul><h2 id=\"usage\">Usage<\/h2><p>This is what you\u2019ll be able to do after the <a href=\"#installation\">installation<\/a>:<\/p><p>First <code class=\"highlighter-rouge\">use<\/code> or <code class=\"highlighter-rouge\">require<\/code> <code class=\"highlighter-rouge\">uncomplicate.neanderthal.core<\/code> and <code class=\"highlighter-rouge\">uncomplicate.neanderthal.native<\/code> in your namespace, and you\u2019ll be able to call appropriate functions from the Neanderthal library.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(ns example\n  (:use [uncomplicate.neanderthal core native]))\n<\/pre><\/div><\/div><p>Now you can create neanderthal vectors and matrices, and do computations with them.<\/p><p>Here is how we create two double-precision vectors and compute their dot product:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def x (dv 1 2 3))\n(def y (dv 10 20 30))\n(dot x y)\n<\/pre><\/div><\/div><p>This is one of the ways to multiply matrices:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def a (dge 3 2 [1 2 3 4 5 6]))\n(def b (dge 2 3 [10 20 30 40 50 60]))\n(mm a b)\n<\/pre><\/div><\/div><h2 id=\"overview-and-features\">Overview and Features<\/h2><p>Neanderthal is a Clojure library for fast matrix and linear algebra computations that supports pluggable engines:<\/p><ul><li>The <strong>native engine<\/strong> is based on a highly optimized native <a href=\"https:\/\/https:\/\/software.intel.com\/en-us\/intel-mkl\">Intel\u2019s MKL<\/a> library of <a href=\"https:\/\/netlib.org\/blas\/\">BLAS<\/a> and <a href=\"https:\/\/www.netlib.org\/lapack\/\">LAPACK<\/a> computation routines (MKL is not open-source, but it is free to use and redistribute since 2016).<\/li>\n<li>The <strong>CUDA GPU engine<\/strong> is based on cuBLAS and supports all modern Nvidia GPUs. It uses <a href=\"https:\/\/clojurecuda.uncomplicate.org\">ClojureCUDA<\/a> and <a href=\"https:\/\/jcuda.org\">JCuda<\/a> libraries. Check out <a href=\"https:\/\/clojurecuda.uncomplicate.org\">Uncomplicate ClojureCUDA<\/a>.<\/li>\n<li>The <strong>OpenCL GPU engine<\/strong> is based on OpenCL BLAS routines from <a href=\"https:\/\/github.com\/CNugteren\/CLBlast\">CLBlast<\/a> library for even more computational power when needed. It uses <a href=\"https:\/\/clojurecl.uncomplicate.org\">ClojureCL<\/a> and <a href=\"https:\/\/jocl.org\">JOCL<\/a> libraries. Check out <a href=\"https:\/\/clojurecl.uncomplicate.org\">Uncomplicate ClojureCL<\/a>.<\/li>\n<\/ul><h3 id=\"implemented-features\">Implemented Features<\/h3><ul><li>Data structures: double and single precision vectors, dense matrices (GE), triangular matrices (TR), symmetric matrices (SY), banded, diagonal, etc.;<\/li>\n<li>BLAS Level 1, 2, and 3 routines;<\/li>\n<li>Lots of (but not all) LAPACK routines;<\/li>\n<li>Various Clojure vector and matrix functions (transpositions, submatrices etc.);<\/li>\n<li>Easy and efficient data mapping and transfer to and from GPUs;<\/li>\n<li>Fast map, reduce and fold implementations for the provided structures.<\/li>\n<li>OpenCL GPU support<\/li>\n<li>CUDA GPU support<\/li>\n<\/ul><h3 id=\"on-the-todo-list\">On the TODO List<\/h3><ul><li>\u201cTensors\u201d (as defined in various deep learning libraries)<\/li>\n<li>Sparse matrices;<\/li>\n<li>Support for complex numbers;<\/li>\n<\/ul><h2 id=\"installation\">Installation<\/h2><ol><li>Add Neanderthal jars to your classpath (<a href=\"https:\/\/clojars.org\/uncomplicate\/neanderthal\">from the Clojars<\/a>).\n1a. If you are running on Java 9 or higher, you need to enable the <code class=\"highlighter-rouge\">java.base<\/code> module. Add this to your JVM options (:jvm-opts in leiningen): <code class=\"highlighter-rouge\">\"--add-opens=java.base\/jdk.internal.ref=ALL-UNNAMED\"<\/code>.<\/li>\n<li>To use the native engine: install Intel\u2019s MKL on your system following <a href=\"#the-native-library-used-by-neanderthals-native-engine\">Native Engine Requirements<\/a>).<\/li>\n<li>To use the GPU engine: install the drivers and a CUDA and\/or OpenCL platform software provided by the vendor of your graphic card (you probably already have that; see <a href=\"#gpu-drivers-for-the-gpu-engine\">GPU Engine Requirements<\/a>).<\/li>\n<\/ol><h3 id=\"with-leiningen\">With Leiningen<\/h3><p>The most straightforward way to include Neanderthal in your project is with Leiningen. Add the following dependency to your <code class=\"highlighter-rouge\">project.clj<\/code>, just like in <a href=\"https:\/\/github.com\/uncomplicate\/neanderthal\/blob\/master\/examples\/hello-world\/project.clj\">the Hello World project<\/a>:<\/p><p><img src=\"https:\/\/clojars.org\/uncomplicate\/neanderthal\/latest-version.svg\" alt=\"\" \/><\/p><h2 id=\"requirements\">Requirements<\/h2><p>You need at least Java 8.<\/p><p>Neanderthal\u2019s data structures are written in Clojure, so many functions work even without native engines. However, you probably need Neanderthal because of its fast BLAS native or GPU engines. Here is how to make sure they are available.<\/p><h3 id=\"the-native-library-used-by-neanderthals-native-engine\">The native library used by Neanderthal\u2019s native engine<\/h3><ul><li>Works on Linux, OS X, and Windows!<\/li>\n<\/ul><p>Neanderthal <strong>uses the native Intel MKL library and expects that you make it available on your system, typically as shared xyz.so, xyz.dll, or xyz.dylib files<\/strong>. Intel MKL is highly optimized for various architectures; its installation comes with many optimized binaries for all supported architectures, that are then selected during runtime according to the hardware at hand. Neanderthal has been built and tested with <strong>Intel MKL 2018<\/strong>; please make sure that you use a compatible MKL version.<\/p><p><strong>You do not need to compile or tune anything yourself.<\/strong><\/p><p>There are two main steps to how to make MKL available on your system; either:<\/p><ol><li>(Optional) Install MKL using a <a href=\"https:\/\/software.intel.com\/en-us\/intel-mkl\">GUI installer provided by Intel<\/a> free of charge. In case you use this method, you may <a href=\"https:\/\/software.intel.com\/en-us\/node\/528500\">set environment variables as explained in this guide<\/a>, but it is probably not required, since you do <strong>not<\/strong> need to compile anything.<\/li>\n<li>Put all required binary files (that you installed with MKL installer or copied from wherever you acquired them) in a directory that is available from your <code class=\"highlighter-rouge\">LD_LIBRARY_PATH<\/code> (Linux), <code class=\"highlighter-rouge\">DYLD_LIBRARY_PATH<\/code> (OSX) or, <code class=\"highlighter-rouge\">PATH<\/code> (windows). Those binary files are available from anyone who installed MKL and have the (free) license to redistribute it with a project.<\/li>\n<\/ol><p>For a Leiningen project, <a href=\"https:\/\/github.com\/athos\/lein-with-env-vars\">lein-with-env-vars<\/a> is useful for passing the path of the MKL libraries to the native linker environment variables, see <a href=\"https:\/\/github.com\/viesti\/lin-alg-app\">lin-alg-app<\/a> example project.<\/p><p>Either way, Neanderthal does not care how MKL has been provided, as long as it is on the path of your OS. When it comes to distributing the software you build using Neanderthal, I guess the most convenient option is that you include the required MKL binary files in the uberjar or other package that you use to ship the product. Then, it would not require any additional action from your users.<\/p><p>This is the list of MKL files that Neanderthal requires:<\/p><ul><li>Mandatory MKL libraries (this is what I use on my Intel Core i7 4790k):\n<ul><li><code class=\"highlighter-rouge\">libmkl_rt.so<\/code><\/li>\n<li><code class=\"highlighter-rouge\">libmkl_core.so<\/code><\/li>\n<li><code class=\"highlighter-rouge\">libmkl_intel_lp64.so<\/code><\/li>\n<li><code class=\"highlighter-rouge\">libiomp5.so<\/code><\/li>\n<li><code class=\"highlighter-rouge\">libmkl_intel_thread<\/code><\/li>\n<li><code class=\"highlighter-rouge\">libmkl_avx2.so<\/code> (if your CPU supports AVX2, which it probably does)<\/li>\n<\/ul><\/li>\n<li>Optionally, your processor might support additional set of instructions that may require an additional file from MKL. See the MKL guide and browse all available files once you install MKL to discover this. For example, I guess that you might also need <code class=\"highlighter-rouge\">libmkl_avx512.so<\/code> if you have a Xeon processor.<\/li>\n<\/ul><p>Please note that, if you use Windows or OS X, the binary file extensions are not <code class=\"highlighter-rouge\">.so<\/code>, but <code class=\"highlighter-rouge\">.dll<\/code> and <code class=\"highlighter-rouge\">dylib<\/code> respectively.<\/p><p><strong>Note for OSX users:<\/strong> MKL installation on my OSX 10.11 placed <code class=\"highlighter-rouge\">libiomp5.dylib<\/code> in a different folder than the rest of the MKL libraries. In such case, copy that file where it is visible, or don\u2019t rely on the MKL installation, but select the needed library files and put them somewhere on the <code class=\"highlighter-rouge\">DYLD_LIBRARY_PATH<\/code>. In newer versions of OSX, you\u2019d have to configure the \u201csystem integrity protection (SIP)\u201d settings for <code class=\"highlighter-rouge\">DYLD_LIBRARY_PATH<\/code> to be respected by the system <a href=\"https:\/\/github.com\/uncomplicate\/neanderthal\/issues\/31\">see more here<\/a>. If you want a quick &amp; dirty solution without much fuss, copying the <code class=\"highlighter-rouge\">dylib<\/code> files and pasting them into <code class=\"highlighter-rouge\">\/usr\/local\/lib<\/code> has been reported to work by multiple users.<\/p><p><strong>Another important thing on Mac<\/strong>: CUDA GPU code that Neanderthal uses could not have been compiled for macOS since it is difficult to find a Mac with Nvidia hardware. That means that you have to <em>explicitly exclude jcuda natives<\/em> in your project. See how to do that in the <a href=\"https:\/\/github.com\/uncomplicate\/neanderthal\/blob\/master\/examples\/hello-world\/project.clj\">Hello World project example<\/a>. <strong>If you have a Mac with an Nvidia GPU and can help with the macOS build, please drop me a note.<\/strong> This will be permanently fixed in the following versions of Neanderthal.<\/p><p><strong>Note for Windows users:<\/strong> MKL installation on my Windows 10 keeps all required <code class=\"highlighter-rouge\">.dll<\/code> files in the <code class=\"highlighter-rouge\">&lt;install dir&gt;\\redist<\/code> folder. The usual folders that keep <code class=\"highlighter-rouge\">.so<\/code> and <code class=\"highlighter-rouge\">dylib<\/code> on Linux and OSX, keep <code class=\"highlighter-rouge\">.lib<\/code> files on Windows - you do not need those. Add the folder that contains the <code class=\"highlighter-rouge\">dll<\/code>s into the <code class=\"highlighter-rouge\">PATH<\/code> environment variable, and you\u2019re ready to go. Some Windows users reported that <code class=\"highlighter-rouge\">libiomp5.dll<\/code> too is in another folder; see the note for OSX users and take the equivalent Windows action.<\/p><h3 id=\"gpu-drivers-for-the-opencl-gpu-engine\">GPU drivers for the OpenCL GPU engine<\/h3><p>Everything will magically work (no need to compile anything) on Nvidia, AMD, and Intel\u2019s GPUs and CPUs as long as you have appropriate GPU drivers.<\/p><p>Works on Linux, Windows, and OS X!<\/p><p>Follow the <a href=\"https:\/\/clojurecl.uncomplicate.org\/articles\/getting_started.html\">ClojureCL getting started guide<\/a> for the links for the GPU platform that you use and more detailed info.<\/p><p><strong>If you use a pre-2.0 OpenCL platform (Nvidia and\/or OS X), you\u2019ll have to use <code class=\"highlighter-rouge\">command-queue-1<\/code> and\/or <code class=\"highlighter-rouge\">with-default-1<\/code> instead of <code class=\"highlighter-rouge\">command-queue<\/code> and <code class=\"highlighter-rouge\">with-default<\/code> that are used in the examples.<\/strong><\/p><h3 id=\"gpu-drivers-for-the-cuda-gpu-engine\">GPU drivers for the CUDA GPU engine<\/h3><p>Everything will magically work (no need to compile anything) on Nvidia, provided that you <strong>installed the latest Nvidia\u2019s CUDA Toolkit<\/strong>.<\/p><p>Follow the <a href=\"https:\/\/clojurecuda.uncomplicate.org\/articles\/getting_started.html\">ClojureCUDA getting started guide<\/a> for the links for the GPU platform that you use and more detailed info.<\/p><h2 id=\"where-to-go-next\">Where to Go Next<\/h2><p>Hopefully this guide got you started and now you\u2019d like to learn more. I expect to build a comprehensive base of articles and references for exploring this topic, so please check the <a href=\"https:\/\/neanderthal.uncomplicate.org\/articles\/guides.html\">All Guides<\/a> page from time to time. Of course, you should also check the <a href=\"https:\/\/neanderthal.uncomplicate.org\/codox\">Neanderthal API<\/a> for specific details, and feel free to take a glance at <a href=\"https:\/\/github.com\/uncomplicate\/neanderthal\">the source<\/a> while you are there.<\/p><p>It is also a good idea to follow <a href=\"https:\/\/dragan.rocks\">my blog at dragan.rocks<\/a> since I\u2019ll write about Neanderthal there.<\/p><h2 data-toc-skip=\"\" id=\"tell_us_what_you_think\">Tell Us What You Think!<\/h2><p>Please take some time to tell us about your experience with the library and this site. <a href=\"https:\/\/neanderthal.uncomplicate.org\/articles\/community.html\">Let us know<\/a> what we should be explaining or is not clear enough. If you are willing to contribute improvements, even better!<\/p>","created_at":"2019-04-03T15:14:40+0000","updated_at":"2019-04-03T15:14:40+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"neanderthal.uncomplicate.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":891,"title":"The future of presenting starts here.","url":"https:\/\/pitch.com\/","content":"<section class=\"wow animateIn\"><p class=\"teaser narrow\">Presentations have defined entire industries and architected how organizations work. Yet for decades we\u2019ve been stuck with the same tools. Pitch is a new beginning. An open platform for presentations and content collaboration. A new paradigm to improve the way presentations are crafted and knowledge is shared.<\/p><\/section><section class=\"small center\"><div class=\"features\"><div class=\"feature wow animateIn\"><img alt=\"Teams\" src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/7984e4ddb4991139e57471b2676586474545ec5f\/07cde\/images\/features\/teams.png\" width=\"360\" \/><h3>Designed for everyone.<br \/>Made for teams.<\/h3><p>No more missing fonts, broken layouts or manual versioning of presentations. For teams and individuals, collaboration has never been easier.<\/p><\/div><div class=\"feature wow animateIn\" data-wow-delay=\".2s\"><img alt=\"Collaboration\" src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/f4c1bc4fe260124a487327e818604f88c66f4fb5\/8a05e\/images\/features\/collab.png\" width=\"360\" \/><h3>Work anywhere.<br \/>On all your devices.<\/h3><p>Collaborate seamlessly across devices from wherever you may be \u2013 whether you\u2019re in the office, on the go, or making last-minute changes before your next meeting.<\/p><\/div><div class=\"feature wow animateIn\"><img alt=\"Templates\" src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/b304c53ec496028d3fc207fad9d7a750e463fa2f\/a616e\/images\/features\/templates.png\" width=\"360\" \/><h3>Powerful templates.<br \/>Next-level layouts.<\/h3><p>Poorly designed presentations are a thing of the past. Create beautiful and high-quality content that is aligned with your brand, in just a few clicks.<\/p><\/div><div class=\"feature wow animateIn\" data-wow-delay=\".2s\"><img alt=\"Integrations\" src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/f950147a990f7cfb79790bdedd330c1138982538\/b52a5\/images\/features\/integrations.png\" width=\"360\" \/><h3>Data from everywhere.<br \/>Updated in real time.<\/h3><p>Manually updating your business reports and sales pitches is tedious and painful. With Pitch, connect to external data sources to turn your presentations into living documents.<\/p><\/div><\/div><\/section>","created_at":"2019-03-30T02:34:04+0000","updated_at":"2019-03-30T02:34:04+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"pitch.com","preview_picture":"https:\/\/www.pitch.com\/images\/meta-image.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":890,"title":"kanaka\/mal","url":"https:\/\/github.com\/kanaka\/mal","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/kanaka\/mal\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/28d41556cf726e46c06fc2d472dc0b20d1ac413d\/68747470733a2f2f7472617669732d63692e6f72672f6b616e616b612f6d616c2e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/kanaka\/mal.svg?branch=master\" \/><\/a><\/p>\n<h2><a id=\"user-content-description\" class=\"anchor\" aria-hidden=\"true\" href=\"#description\"><\/a>Description<\/h2>\n<p><strong>1. Mal is a Clojure inspired Lisp interpreter<\/strong><\/p>\n<p><strong>2. Mal is implemented in 75 languages (76 implementations total)<\/strong><\/p>\n<p><strong>3. Mal is a learning tool<\/strong><\/p>\n<p>Each implementation of mal is separated into\n11 incremental, self-contained (and testable) steps that demonstrate\ncore concepts of Lisp. The last step is capable of self-hosting\n(running the mal implementation of mal). See the <a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md\">make-a-lisp process\nguide<\/a>.<\/p>\n<p>The make-a-lisp steps are:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step0\">step0_repl<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step1\">step1_read_print<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step2\">step2_eval<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step3\">step3_env<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step4\">step4_if_fn_do<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step5\">step5_tco<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step6\">step6_file<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step7\">step7_quote<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step8\">step8_macros<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#step9\">step9_try<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#stepA\">stepA_mal<\/a><\/li>\n<\/ul><p>Each make-a-lisp step has an associated architectural diagram. That elements\nthat are new for that step are highlighted in red.\nHere is the final diagram for <a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md#stepA\">step A<\/a>:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/stepA_mal.png\"><img src=\"https:\/\/github.com\/kanaka\/mal\/raw\/master\/process\/stepA_mal.png\" alt=\"stepA_mal architecture\" \/><\/a><\/p>\n<p>If you are interesting in creating a mal implementation (or just\ninterested in using mal for something), please drop by the #mal\nchannel on freenode. In addition to the <a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/process\/guide.md\">make-a-lisp process\nguide<\/a> there is also a <a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/docs\/FAQ.md\">mal\/make-a-lisp\nFAQ<\/a> where I attempt to answer some common questions.<\/p>\n<h2><a id=\"user-content-presentations\" class=\"anchor\" aria-hidden=\"true\" href=\"#presentations\"><\/a>Presentations<\/h2>\n<p>Mal was presented publicly for the first time in a lightning talk at\nClojure West 2014 (unfortunately there is no video). See\nexamples\/clojurewest2014.mal for the presentation that was given at the\nconference (yes, the presentation is a mal program).<\/p>\n<p>At Midwest.io 2015, Joel Martin gave a presentation on Mal titled\n\"Achievement Unlocked: A Better Path to Language Learning\".\n<a href=\"https:\/\/www.youtube.com\/watch?v=lgyOAiRtZGw\" rel=\"nofollow\">Video<\/a>,\n<a href=\"http:\/\/kanaka.github.io\/midwest.io.mal\/\" rel=\"nofollow\">Slides<\/a>.<\/p>\n<p>More recently Joel gave a presentation on \"Make Your Own Lisp Interpreter\nin 10 Incremental Steps\" at LambdaConf 2016:\n<a href=\"https:\/\/www.youtube.com\/watch?v=jVhupfthTEk\" rel=\"nofollow\">Part 1<\/a>,\n<a href=\"https:\/\/www.youtube.com\/watch?v=X5OQBMGpaTU\" rel=\"nofollow\">Part 2<\/a>,\n<a href=\"https:\/\/www.youtube.com\/watch?v=6mARZzGgX4U\" rel=\"nofollow\">Part 3<\/a>,\n<a href=\"https:\/\/www.youtube.com\/watch?v=dCO1SYR5kDU\" rel=\"nofollow\">Part 4<\/a>,\n<a href=\"http:\/\/kanaka.github.io\/lambdaconf\/\" rel=\"nofollow\">Slides<\/a>.<\/p>\n<h2><a id=\"user-content-buildingrunning-implementations\" class=\"anchor\" aria-hidden=\"true\" href=\"#buildingrunning-implementations\"><\/a>Building\/running implementations<\/h2>\n<p>The simplest way to run any given implementation is to use docker.\nEvery implementation has a docker image pre-built with language\ndependencies installed. You can launch the REPL using a convenient\ntarget in the top level Makefile (where IMPL is the implementation\ndirectory name and stepX is the step to run):<\/p>\n<pre>make DOCKERIZE=1 \"repl^IMPL^stepX\"\n    # OR stepA is the default step:\nmake DOCKERIZE=1 \"repl^IMPL\"\n<\/pre>\n<h3><a id=\"user-content-ada\" class=\"anchor\" aria-hidden=\"true\" href=\"#ada\"><\/a>Ada<\/h3>\n<p>The Ada implementation was developed with GNAT 4.9 on debian. It also\ncompiles unchanged on windows if you have windows versions of git,\nGNAT and (optionally) make.  There are no external dependencies\n(readline not implemented).<\/p>\n<pre>cd ada\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-ada2\" class=\"anchor\" aria-hidden=\"true\" href=\"#ada2\"><\/a>Ada.2<\/h3>\n<p>The second Ada implementation was developed with GNAT 8 and links with\nthe GNU readline library.<\/p>\n<pre>cd ada\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-gnu-awk\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnu-awk\"><\/a>GNU awk<\/h3>\n<p>The GNU awk implementation of mal has been tested with GNU awk 4.1.1.<\/p>\n<pre>cd gawk\ngawk -O -f stepX_YYY.awk\n<\/pre>\n<h3><a id=\"user-content-bash-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#bash-4\"><\/a>Bash 4<\/h3>\n<pre>cd bash\nbash stepX_YYY.sh\n<\/pre>\n<h3><a id=\"user-content-basic-c64-and-qbasic\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-c64-and-qbasic\"><\/a>BASIC (C64 and QBasic)<\/h3>\n<p>The BASIC implementation uses a preprocessor that can generate BASIC\ncode that is compatible with both C64 BASIC (CBM v2) and QBasic. The\nC64 mode has been tested with\n<a href=\"https:\/\/github.com\/kanaka\/cbmbasic\">cbmbasic<\/a> (the patched version is\ncurrently required to fix issues with line input) and the QBasic mode\nhas been tested with <a href=\"http:\/\/www.qb64.net\/\" rel=\"nofollow\">qb64<\/a>.<\/p>\n<p>Generate C64 code and run it using cbmbasic:<\/p>\n<pre>cd basic\nmake stepX_YYY.bas\nSTEP=stepX_YYY .\/run\n<\/pre>\n<p>Generate QBasic code and load it into qb64:<\/p>\n<pre>cd basic\nmake MODE=qbasic stepX_YYY.bas\n.\/qb64 stepX_YYY.bas\n<\/pre>\n<p>Thanks to <a href=\"https:\/\/github.com\/sjsyrek\">Steven Syrek<\/a> for the original\ninspiration for this implementation.<\/p>\n<h3><a id=\"user-content-c\" class=\"anchor\" aria-hidden=\"true\" href=\"#c\"><\/a>C<\/h3>\n<p>The C implementation of mal requires the following libraries (lib and\nheader packages): glib, libffi6, libgc, and either the libedit or GNU readline\nlibrary.<\/p>\n<pre>cd c\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-c-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#c-1\"><\/a>C++<\/h3>\n<p>The C++ implementation of mal requires g++-4.9 or clang++-3.5 and\na readline compatible library to build. See the <code>cpp\/README.md<\/code> for\nmore details:<\/p>\n<pre>cd cpp\nmake\n    # OR\nmake CXX=clang++-3.5\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-c-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#c-2\"><\/a>C#<\/h3>\n<p>The C# implementation of mal has been tested on Linux using the Mono\nC# compiler (mcs) and the Mono runtime (version 2.10.8.1). Both are\nrequired to build and run the C# implementation.<\/p>\n<pre>cd cs\nmake\nmono .\/stepX_YYY.exe\n<\/pre>\n<h3><a id=\"user-content-chuck\" class=\"anchor\" aria-hidden=\"true\" href=\"#chuck\"><\/a>ChucK<\/h3>\n<p>The ChucK implementation has been tested with ChucK 1.3.5.2.<\/p>\n<pre>cd chuck\n.\/run\n<\/pre>\n<h3><a id=\"user-content-clojure\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure\"><\/a>Clojure<\/h3>\n<p>For the most part the Clojure implementation requires Clojure 1.5,\nhowever, to pass all tests, Clojure 1.8.0-RC4 is required.<\/p>\n<pre>cd clojure\nlein with-profile +stepX trampoline run\n<\/pre>\n<h3><a id=\"user-content-coffeescript\" class=\"anchor\" aria-hidden=\"true\" href=\"#coffeescript\"><\/a>CoffeeScript<\/h3>\n<pre>sudo npm install -g coffee-script\ncd coffee\ncoffee .\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-common-lisp\" class=\"anchor\" aria-hidden=\"true\" href=\"#common-lisp\"><\/a>Common Lisp<\/h3>\n<p>The implementation has been tested with SBCL, CCL, CMUCL, GNU CLISP, ECL and\nAllegro CL on Ubuntu 16.04 and Ubuntu 12.04, see\nthe <a href=\"https:\/\/github.com\/kanaka\/mal\/blob\/master\/common-lisp\/README.org\">README<\/a> for more details. Provided you have the\ndependencies mentioned installed, do the following to run the implementation<\/p>\n<pre>cd common-lisp\nmake\n.\/run\n<\/pre>\n<h3><a id=\"user-content-crystal\" class=\"anchor\" aria-hidden=\"true\" href=\"#crystal\"><\/a>Crystal<\/h3>\n<p>The Crystal implementation of mal has been tested with Crystal 0.26.1.<\/p>\n<pre>cd crystal\ncrystal run .\/stepX_YYY.cr\n    # OR\nmake   # needed to run tests\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-d\" class=\"anchor\" aria-hidden=\"true\" href=\"#d\"><\/a>D<\/h3>\n<p>The D implementation of mal was tested with GDC 4.8.  It requires the GNU\nreadline library.<\/p>\n<pre>cd d\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-dart\" class=\"anchor\" aria-hidden=\"true\" href=\"#dart\"><\/a>Dart<\/h3>\n<p>The Dart implementation has been tested with Dart 1.20.<\/p>\n<pre>cd dart\ndart .\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-emacs-lisp\" class=\"anchor\" aria-hidden=\"true\" href=\"#emacs-lisp\"><\/a>Emacs Lisp<\/h3>\n<p>The Emacs Lisp implementation of mal has been tested with Emacs 24.3\nand 24.5.  While there is very basic readline editing (<code>&lt;backspace&gt;<\/code>\nand <code>C-d<\/code> work, <code>C-c<\/code> cancels the process), it is recommended to use\n<code>rlwrap<\/code>.<\/p>\n<pre>cd elisp\nemacs -Q --batch --load stepX_YYY.el\n# with full readline support\nrlwrap emacs -Q --batch --load stepX_YYY.el\n<\/pre>\n<h3><a id=\"user-content-elixir\" class=\"anchor\" aria-hidden=\"true\" href=\"#elixir\"><\/a>Elixir<\/h3>\n<p>The Elixir implementation of mal has been tested with Elixir 1.0.5.<\/p>\n<pre>cd elixir\nmix stepX_YYY\n# Or with readline\/line editing functionality:\niex -S mix stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-elm\" class=\"anchor\" aria-hidden=\"true\" href=\"#elm\"><\/a>Elm<\/h3>\n<p>The Elm implementation of mal has been tested with Elm 0.18.0<\/p>\n<pre>cd elm\nmake stepX_YYY.js\nSTEP=stepX_YYY .\/run\n<\/pre>\n<h3><a id=\"user-content-erlang\" class=\"anchor\" aria-hidden=\"true\" href=\"#erlang\"><\/a>Erlang<\/h3>\n<p>The Erlang implementation of mal requires <a href=\"http:\/\/www.erlang.org\/download.html\" rel=\"nofollow\">Erlang\/OTP R17<\/a>\nand <a href=\"https:\/\/github.com\/rebar\/rebar\">rebar<\/a> to build.<\/p>\n<pre>cd erlang\nmake\n    # OR\nMAL_STEP=stepX_YYY rebar compile escriptize # build individual step\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-es6-ecmascript-2015\" class=\"anchor\" aria-hidden=\"true\" href=\"#es6-ecmascript-2015\"><\/a>ES6 (ECMAScript 2015)<\/h3>\n<p>The ES6 \/ ECMAScript 2015 implementation uses the\n<a href=\"https:\/\/babeljs.io\" rel=\"nofollow\">babel<\/a> compiler to generate ES5 compatible\nJavaScript. The generated code has been tested with Node 0.12.4.<\/p>\n<pre>cd es6\nmake\nnode build\/stepX_YYY.js\n<\/pre>\n<h3><a id=\"user-content-f\" class=\"anchor\" aria-hidden=\"true\" href=\"#f\"><\/a>F#<\/h3>\n<p>The F# implementation of mal has been tested on Linux using the Mono\nF# compiler (fsharpc) and the Mono runtime (version 3.12.1). The mono C#\ncompiler (mcs) is also necessary to compile the readline dependency. All are\nrequired to build and run the F# implementation.<\/p>\n<pre>cd fsharp\nmake\nmono .\/stepX_YYY.exe\n<\/pre>\n<h3><a id=\"user-content-factor\" class=\"anchor\" aria-hidden=\"true\" href=\"#factor\"><\/a>Factor<\/h3>\n<p>The Factor implementation of mal has been tested with Factor 0.97\n(<a href=\"http:\/\/factorcode.org\" rel=\"nofollow\">factorcode.org<\/a>).<\/p>\n<pre>cd factor\nFACTOR_ROOTS=. factor -run=stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-fantom\" class=\"anchor\" aria-hidden=\"true\" href=\"#fantom\"><\/a>Fantom<\/h3>\n<p>The Fantom implementation of mal has been tested with Fantom 1.0.70.<\/p>\n<pre>cd fantom\nmake lib\/fan\/stepX_YYY.pod\nSTEP=stepX_YYY .\/run\n<\/pre>\n<h3><a id=\"user-content-forth\" class=\"anchor\" aria-hidden=\"true\" href=\"#forth\"><\/a>Forth<\/h3>\n<pre>cd forth\ngforth stepX_YYY.fs\n<\/pre>\n<h3><a id=\"user-content-gnu-guile-21\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnu-guile-21\"><\/a>GNU Guile 2.1+<\/h3>\n<pre>cd guile\nguile -L .\/ stepX_YYY.scm\n<\/pre>\n<h3><a id=\"user-content-gnu-smalltalk\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnu-smalltalk\"><\/a>GNU Smalltalk<\/h3>\n<p>The Smalltalk implementation of mal has been tested with GNU Smalltalk 3.2.91.<\/p>\n<pre>cd gnu-smalltalk\n.\/run\n<\/pre>\n<h3><a id=\"user-content-go\" class=\"anchor\" aria-hidden=\"true\" href=\"#go\"><\/a>Go<\/h3>\n<p>The Go implementation of mal requires that go is installed on on the\npath. The implementation has been tested with Go 1.3.1.<\/p>\n<pre>cd go\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-groovy\" class=\"anchor\" aria-hidden=\"true\" href=\"#groovy\"><\/a>Groovy<\/h3>\n<p>The Groovy implementation of mal requires Groovy to run and has been\ntested with Groovy 1.8.6.<\/p>\n<pre>cd groovy\nmake\ngroovy .\/stepX_YYY.groovy\n<\/pre>\n<h3><a id=\"user-content-haskell\" class=\"anchor\" aria-hidden=\"true\" href=\"#haskell\"><\/a>Haskell<\/h3>\n<p>The Haskell implementation requires the ghc compiler version 7.10.1 or\nlater and also the Haskell parsec and readline (or editline) packages.<\/p>\n<pre>cd haskell\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-haxe-neko-python-c-and-javascript\" class=\"anchor\" aria-hidden=\"true\" href=\"#haxe-neko-python-c-and-javascript\"><\/a>Haxe (Neko, Python, C++ and JavaScript)<\/h3>\n<p>The Haxe implementation of mal requires Haxe version 3.2 to compile.\nFour different Haxe targets are supported: Neko, Python, C++, and\nJavaScript.<\/p>\n<pre>cd haxe\n# Neko\nmake all-neko\nneko .\/stepX_YYY.n\n# Python\nmake all-python\npython3 .\/stepX_YYY.py\n# C++\nmake all-cpp\n.\/cpp\/stepX_YYY\n# JavaScript\nmake all-js\nnode .\/stepX_YYY.js\n<\/pre>\n<h3><a id=\"user-content-hy\" class=\"anchor\" aria-hidden=\"true\" href=\"#hy\"><\/a>Hy<\/h3>\n<p>The Hy implementation of mal has been tested with Hy 0.13.0.<\/p>\n<pre>cd hy\n.\/stepX_YYY.hy\n<\/pre>\n<h3><a id=\"user-content-io\" class=\"anchor\" aria-hidden=\"true\" href=\"#io\"><\/a>Io<\/h3>\n<p>The Io implementation of mal has been tested with Io version 20110905.<\/p>\n<pre>cd io\nio .\/stepX_YYY.io\n<\/pre>\n<h3><a id=\"user-content-java-17\" class=\"anchor\" aria-hidden=\"true\" href=\"#java-17\"><\/a>Java 1.7<\/h3>\n<p>The Java implementation of mal requires maven2 to build.<\/p>\n<pre>cd java\nmvn compile\nmvn -quiet exec:java -Dexec.mainClass=mal.stepX_YYY\n    # OR\nmvn -quiet exec:java -Dexec.mainClass=mal.stepX_YYY -Dexec.args=\"CMDLINE_ARGS\"\n<\/pre>\n<h3><a id=\"user-content-javascriptnode\" class=\"anchor\" aria-hidden=\"true\" href=\"#javascriptnode\"><\/a>JavaScript\/Node<\/h3>\n<pre>cd js\nnpm update\nnode stepX_YYY.js\n<\/pre>\n<h3><a id=\"user-content-julia\" class=\"anchor\" aria-hidden=\"true\" href=\"#julia\"><\/a>Julia<\/h3>\n<p>The Julia implementation of mal requires Julia 0.4.<\/p>\n<pre>cd julia\njulia stepX_YYY.jl\n<\/pre>\n<h3><a id=\"user-content-kotlin\" class=\"anchor\" aria-hidden=\"true\" href=\"#kotlin\"><\/a>Kotlin<\/h3>\n<p>The Kotlin implementation of mal has been tested with Kotlin 1.0.<\/p>\n<pre>cd kotlin\nmake\njava -jar stepX_YYY.jar\n<\/pre>\n<h3><a id=\"user-content-livescript\" class=\"anchor\" aria-hidden=\"true\" href=\"#livescript\"><\/a>LiveScript<\/h3>\n<p>The LiveScript implementation of mal has been tested with LiveScript 1.5.<\/p>\n<pre>cd livescript\nmake\nnode_modules\/.bin\/lsc stepX_YYY.ls\n<\/pre>\n<h3><a id=\"user-content-logo\" class=\"anchor\" aria-hidden=\"true\" href=\"#logo\"><\/a>Logo<\/h3>\n<p>The Logo implementation of mal has been tested with UCBLogo 6.0.<\/p>\n<pre>cd logo\nlogo stepX_YYY.lg\n<\/pre>\n<h3><a id=\"user-content-lua\" class=\"anchor\" aria-hidden=\"true\" href=\"#lua\"><\/a>Lua<\/h3>\n<p>The Lua implementation of mal has been tested with Lua 5.2. The\nimplementation requires that luarocks and the lua-rex-pcre library\nare installed.<\/p>\n<pre>cd lua\nmake  # to build and link linenoise.so\n.\/stepX_YYY.lua\n<\/pre>\n<h3><a id=\"user-content-mal\" class=\"anchor\" aria-hidden=\"true\" href=\"#mal\"><\/a>Mal<\/h3>\n<p>Running the mal implementation of mal involves running stepA of one of\nthe other implementations and passing the mal step to run as a command\nline argument.<\/p>\n<pre>cd IMPL\nIMPL_STEPA_CMD ..\/mal\/stepX_YYY.mal\n<\/pre>\n<h3><a id=\"user-content-gnu-make-381\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnu-make-381\"><\/a>GNU Make 3.81<\/h3>\n<pre>cd make\nmake -f stepX_YYY.mk\n<\/pre>\n<h3><a id=\"user-content-nasm\" class=\"anchor\" aria-hidden=\"true\" href=\"#nasm\"><\/a>NASM<\/h3>\n<p>The NASM implementation of mal is written for x86-64 Linux, and has been tested\nwith Linux 3.16.0-4-amd64 and NASM version 2.11.05.<\/p>\n<pre>cd nasm\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-nim-0170\" class=\"anchor\" aria-hidden=\"true\" href=\"#nim-0170\"><\/a>Nim 0.17.0<\/h3>\n<p>The Nim implementation of mal has been tested with Nim 0.17.0.<\/p>\n<pre>cd nim\nmake\n  # OR\nnimble build\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-object-pascal\" class=\"anchor\" aria-hidden=\"true\" href=\"#object-pascal\"><\/a>Object Pascal<\/h3>\n<p>The Object Pascal implementation of mal has been built and tested on\nLinux using the Free Pascal compiler version 2.6.2 and 2.6.4.<\/p>\n<pre>cd objpascal\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-objective-c\" class=\"anchor\" aria-hidden=\"true\" href=\"#objective-c\"><\/a>Objective C<\/h3>\n<p>The Objective C implementation of mal has been built and tested on\nLinux using clang\/LLVM 3.6. It has also been built and tested on OS\nX using XCode 7.<\/p>\n<pre>cd objc\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-ocaml-4010\" class=\"anchor\" aria-hidden=\"true\" href=\"#ocaml-4010\"><\/a>OCaml 4.01.0<\/h3>\n<pre>cd ocaml\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-matlab-gnu-octave-and-matlab\" class=\"anchor\" aria-hidden=\"true\" href=\"#matlab-gnu-octave-and-matlab\"><\/a>MATLAB (GNU Octave and MATLAB)<\/h3>\n<p>The MatLab implementation has been tested with GNU Octave 4.2.1.\nIt has also been tested with MATLAB version R2014a on Linux. Note that\nMATLAB is a commercial product.<\/p>\n<pre>cd matlab\n.\/stepX_YYY\noctave -q --no-gui --no-history --eval \"stepX_YYY();quit;\"\nmatlab -nodisplay -nosplash -nodesktop -nojvm -r \"stepX_YYY();quit;\"\n    # OR with command line arguments\noctave -q --no-gui --no-history --eval \"stepX_YYY('arg1','arg2');quit;\"\nmatlab -nodisplay -nosplash -nodesktop -nojvm -r \"stepX_YYY('arg1','arg2');quit;\"\n<\/pre>\n<h3><a id=\"user-content-minimal\" class=\"anchor\" aria-hidden=\"true\" href=\"#minimal\"><\/a>miniMAL<\/h3>\n<p><a href=\"https:\/\/github.com\/kanaka\/miniMAL\">miniMAL<\/a> is small Lisp interpreter\nimplemented in less than 1024 bytes of JavaScript. To run the miniMAL\nimplementation of mal you need to download\/install the miniMAL\ninterpreter (which requires Node.js).<\/p>\n<pre>cd miniMAL\n# Download miniMAL and dependencies\nnpm install\nexport PATH=`pwd`\/node_modules\/minimal-lisp\/:$PATH\n# Now run mal implementation in miniMAL\nminiMAL .\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-perl-58\" class=\"anchor\" aria-hidden=\"true\" href=\"#perl-58\"><\/a>Perl 5.8<\/h3>\n<p>For readline line editing support, install Term::ReadLine::Perl or\nTerm::ReadLine::Gnu from CPAN.<\/p>\n<pre>cd perl\nperl stepX_YYY.pl\n<\/pre>\n<h3><a id=\"user-content-perl-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#perl-6\"><\/a>Perl 6<\/h3>\n<p>The Perl 6 implementation was tested on Rakudo Perl 6 2016.04.<\/p>\n<pre>cd perl6\nperl6 stepX_YYY.pl\n<\/pre>\n<h3><a id=\"user-content-php-53\" class=\"anchor\" aria-hidden=\"true\" href=\"#php-53\"><\/a>PHP 5.3<\/h3>\n<p>The PHP implementation of mal requires the php command line interface\nto run.<\/p>\n<pre>cd php\nphp stepX_YYY.php\n<\/pre>\n<h3><a id=\"user-content-picolisp\" class=\"anchor\" aria-hidden=\"true\" href=\"#picolisp\"><\/a>Picolisp<\/h3>\n<p>The Picolisp implementation requires libreadline and Picolisp 3.1.11\nor later.<\/p>\n<pre>cd picolisp\n.\/run\n<\/pre>\n<h3><a id=\"user-content-plpgsql-postgres-sql-procedural-language\" class=\"anchor\" aria-hidden=\"true\" href=\"#plpgsql-postgres-sql-procedural-language\"><\/a>PL\/pgSQL (Postgres SQL Procedural Language)<\/h3>\n<p>The PL\/pgSQL implementation of mal requires a running Postgres server\n(the \"kanaka\/mal-test-plpgsql\" docker image automatically starts\na Postgres server). The implementation connects to the Postgres server\nand create a database named \"mal\" to store tables and stored\nprocedures. The wrapper script uses the psql command to connect to the\nserver and defaults to the user \"postgres\" but this can be overridden\nwith the PSQL_USER environment variable. A password can be specified\nusing the PGPASSWORD environment variable. The implementation has been\ntested with Postgres 9.4.<\/p>\n<pre>cd plpgsql\n.\/wrap.sh stepX_YYY.sql\n    # OR\nPSQL_USER=myuser PGPASSWORD=mypass .\/wrap.sh stepX_YYY.sql\n<\/pre>\n<h3><a id=\"user-content-plsql-oracle-sql-procedural-language\" class=\"anchor\" aria-hidden=\"true\" href=\"#plsql-oracle-sql-procedural-language\"><\/a>PL\/SQL (Oracle SQL Procedural Language)<\/h3>\n<p>The PL\/pgSQL implementation of mal requires a running Oracle DB\nserver (the \"kanaka\/mal-test-plsql\" docker image automatically\nstarts an Oracle Express server). The implementation connects to the\nOracle server to create types, tables and stored procedures. The\ndefault SQL<em>Plus logon value (username\/password@connect_identifier) is\n\"system\/oracle\" but this can be overridden with the ORACLE_LOGON\nenvironment variable. The implementation has been tested with Oracle\nExpress Edition 11g Release 2. Note that any SQL<\/em>Plus connection\nwarnings (user password expiration, etc) will interfere with the\nability of the wrapper script to communicate with the DB.<\/p>\n<pre>cd plsql\n.\/wrap.sh stepX_YYY.sql\n    # OR\nORACLE_LOGON=myuser\/mypass@ORCL .\/wrap.sh stepX_YYY.sql\n<\/pre>\n<h3><a id=\"user-content-postscript-level-23\" class=\"anchor\" aria-hidden=\"true\" href=\"#postscript-level-23\"><\/a>Postscript Level 2\/3<\/h3>\n<p>The Postscript implementation of mal requires ghostscript to run. It\nhas been tested with ghostscript 9.10.<\/p>\n<pre>cd ps\ngs -q -dNODISPLAY -I.\/ stepX_YYY.ps\n<\/pre>\n<h3><a id=\"user-content-powershell\" class=\"anchor\" aria-hidden=\"true\" href=\"#powershell\"><\/a>PowerShell<\/h3>\n<p>The PowerShell implementation of mal requires the PowerShell script\nlanguage. It has been tested with PowerShell 6.0.0 Alpha 9 on Linux.<\/p>\n<pre>cd powershell\npowershell .\/stepX_YYY.ps1\n<\/pre>\n<h3><a id=\"user-content-python-2x-and-3x\" class=\"anchor\" aria-hidden=\"true\" href=\"#python-2x-and-3x\"><\/a>Python (2.X and 3.X)<\/h3>\n<pre>cd python\npython stepX_YYY.py\n<\/pre>\n<h3><a id=\"user-content-rpython\" class=\"anchor\" aria-hidden=\"true\" href=\"#rpython\"><\/a>RPython<\/h3>\n<p>You must have <a href=\"https:\/\/rpython.readthedocs.org\/\" rel=\"nofollow\">rpython<\/a> on your path\n(included with <a href=\"https:\/\/bitbucket.org\/pypy\/pypy\/\" rel=\"nofollow\">pypy<\/a>).<\/p>\n<pre>cd rpython\nmake        # this takes a very long time\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-r\" class=\"anchor\" aria-hidden=\"true\" href=\"#r\"><\/a>R<\/h3>\n<p>The R implementation of mal requires R (r-base-core) to run.<\/p>\n<pre>cd r\nmake libs  # to download and build rdyncall\nRscript stepX_YYY.r\n<\/pre>\n<h3><a id=\"user-content-racket-53\" class=\"anchor\" aria-hidden=\"true\" href=\"#racket-53\"><\/a>Racket (5.3)<\/h3>\n<p>The Racket implementation of mal requires the Racket\ncompiler\/interpreter to run.<\/p>\n<pre>cd racket\n.\/stepX_YYY.rkt\n<\/pre>\n<h3><a id=\"user-content-rexx\" class=\"anchor\" aria-hidden=\"true\" href=\"#rexx\"><\/a>Rexx<\/h3>\n<p>The Rexx implementation of mal has been tested with Regina Rexx 3.6.<\/p>\n<pre>cd rexx\nmake\nrexx -a .\/stepX_YYY.rexxpp\n<\/pre>\n<h3><a id=\"user-content-ruby-19\" class=\"anchor\" aria-hidden=\"true\" href=\"#ruby-19\"><\/a>Ruby (1.9+)<\/h3>\n<pre>cd ruby\nruby stepX_YYY.rb\n<\/pre>\n<h3><a id=\"user-content-rust-100-nightly\" class=\"anchor\" aria-hidden=\"true\" href=\"#rust-100-nightly\"><\/a>Rust (1.0.0 nightly)<\/h3>\n<p>The rust implementation of mal requires the rust compiler and build\ntool (cargo) to build.<\/p>\n<pre>cd rust\ncargo run --release --bin stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-scala\" class=\"anchor\" aria-hidden=\"true\" href=\"#scala\"><\/a>Scala<\/h3>\n<p>Install scala and sbt (<a href=\"http:\/\/www.scala-sbt.org\/0.13\/tutorial\/Installing-sbt-on-Linux.html\" rel=\"nofollow\">http:\/\/www.scala-sbt.org\/0.13\/tutorial\/Installing-sbt-on-Linux.html<\/a>):<\/p>\n<pre>cd scala\nsbt 'run-main stepX_YYY'\n    # OR\nsbt compile\nscala -classpath target\/scala*\/classes stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-scheme-r7rs\" class=\"anchor\" aria-hidden=\"true\" href=\"#scheme-r7rs\"><\/a>Scheme (R7RS)<\/h3>\n<p>The Scheme implementation of mal has been tested with Chibi-Scheme\n0.7.3, Kawa 2.4, Gauche 0.9.5, CHICKEN 4.11.0, Sagittarius 0.8.3,\nCyclone 0.6.3 (Git version) and Foment 0.4 (Git version).  You should\nbe able to get it running on other conforming R7RS implementations\nafter figuring out how libraries are loaded and adjusting the\n<code>Makefile<\/code> and <code>run<\/code> script accordingly.<\/p>\n<pre>cd scheme\nmake symlinks\n# chibi\nscheme_MODE=chibi .\/run\n# kawa\nmake kawa\nscheme_MODE=kawa .\/run\n# gauche\nscheme_MODE=gauche .\/run\n# chicken\nmake chicken\nscheme_MODE=chicken .\/run\n# sagittarius\nscheme_MODE=sagittarius .\/run\n# cyclone\nmake cyclone\nscheme_MODE=cyclone .\/run\n# foment\nscheme_MODE=foment .\/run\n<\/pre>\n<h3><a id=\"user-content-skew\" class=\"anchor\" aria-hidden=\"true\" href=\"#skew\"><\/a>Skew<\/h3>\n<p>The Skew implementation of mal has been tested with Skew 0.7.42.<\/p>\n<pre>cd skew\nmake\nnode stepX_YYY.js\n<\/pre>\n<h3><a id=\"user-content-swift\" class=\"anchor\" aria-hidden=\"true\" href=\"#swift\"><\/a>Swift<\/h3>\n<p>The Swift implementation of mal requires the Swift 2.0 compiler (XCode\n7.0) to build. Older versions will not work due to changes in the\nlanguage and standard library.<\/p>\n<pre>cd swift\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-swift-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#swift-3\"><\/a>Swift 3<\/h3>\n<p>The Swift 3 implementation of mal requires the Swift 3.0 compiler. It\nhas been tested with Swift 3 Preview 3.<\/p>\n<pre>cd swift3\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-swift-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#swift-4\"><\/a>Swift 4<\/h3>\n<p>The Swift 4 implementation of mal requires the Swift 4.0 compiler. It\nhas been tested with Swift 4.2.3 release.<\/p>\n<pre>cd swift4\nmake\n.\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-tcl-86\" class=\"anchor\" aria-hidden=\"true\" href=\"#tcl-86\"><\/a>Tcl 8.6<\/h3>\n<p>The Tcl implementation of mal requires Tcl 8.6 to run.  For readline line\nediting support, install tclreadline.<\/p>\n<pre>cd tcl\ntclsh .\/stepX_YYY.tcl\n<\/pre>\n<h3><a id=\"user-content-typescript\" class=\"anchor\" aria-hidden=\"true\" href=\"#typescript\"><\/a>TypeScript<\/h3>\n<p>The TypeScript implementation of mal requires the TypeScript 2.2 compiler.\nIt has been tested with Node.js v6.<\/p>\n<pre>cd ts\nmake\nnode .\/stepX_YYY.js\n<\/pre>\n<h3><a id=\"user-content-vhdl\" class=\"anchor\" aria-hidden=\"true\" href=\"#vhdl\"><\/a>VHDL<\/h3>\n<p>The VHDL implementation of mal has been tested with GHDL 0.29.<\/p>\n<pre>cd vhdl\nmake\n.\/run_vhdl.sh .\/stepX_YYY\n<\/pre>\n<h3><a id=\"user-content-vimscript\" class=\"anchor\" aria-hidden=\"true\" href=\"#vimscript\"><\/a>Vimscript<\/h3>\n<p>The Vimscript implementation of mal requires Vim 8.0 to run.<\/p>\n<pre>cd vimscript\n.\/run_vimscript.sh .\/stepX_YYY.vim\n<\/pre>\n<h3><a id=\"user-content-visual-basicnet\" class=\"anchor\" aria-hidden=\"true\" href=\"#visual-basicnet\"><\/a>Visual Basic.NET<\/h3>\n<p>The VB.NET implementation of mal has been tested on Linux using the Mono\nVB compiler (vbnc) and the Mono runtime (version 2.10.8.1). Both are\nrequired to build and run the VB.NET implementation.<\/p>\n<pre>cd vb\nmake\nmono .\/stepX_YYY.exe\n<\/pre>\n<h3><a id=\"user-content-webassembly-wasm\" class=\"anchor\" aria-hidden=\"true\" href=\"#webassembly-wasm\"><\/a>WebAssembly (wasm)<\/h3>\n<p>The WebAssembly implementation is written in\n<a href=\"https:\/\/github.com\/kanaka\/wam\">Wam<\/a> (WebAssembly Macro language) and\nruns under the <a href=\"https:\/\/github.com\/kanaka\/wac\">wac\/wace<\/a> WebAssembly\nruntime.<\/p>\n<pre>cd wasm\nmake\nwace .\/stepX_YYY.wasm\n<\/pre>\n<h3><a id=\"user-content-yorick\" class=\"anchor\" aria-hidden=\"true\" href=\"#yorick\"><\/a>Yorick<\/h3>\n<p>The Yorick implementation of mal was tested on Yorick 2.2.04.<\/p>\n<pre>cd yorick\nyorick -batch .\/stepX_YYY.i\n<\/pre>\n<h2><a id=\"user-content-running-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-tests\"><\/a>Running tests<\/h2>\n<p>The top level Makefile has a number of useful targets to assist with\nimplementation development and testing. The <code>help<\/code> target provides\na list of the targets and options:<\/p>\n<pre>make help\n<\/pre>\n<h3><a id=\"user-content-functional-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#functional-tests\"><\/a>Functional tests<\/h3>\n<p>The are over 600 generic functional tests (for all implementations)\nin the <code>tests\/<\/code> directory. Each step has a corresponding test file\ncontaining tests specific to that step. The <code>runtest.py<\/code> test harness\nlaunches a Mal step implementation and then feeds the tests one at\na time to the implementation and compares the output\/return value to\nthe expected output\/return value.<\/p>\n<ul><li>To run all the tests across all implementations (be prepared to wait):<\/li>\n<\/ul><pre>make test\n<\/pre>\n<ul><li>To run all tests against a single implementation:<\/li>\n<\/ul><pre>make \"test^IMPL\"\n# e.g.\nmake \"test^clojure\"\nmake \"test^js\"\n<\/pre>\n<ul><li>To run tests for a single step against all implementations:<\/li>\n<\/ul><pre>make \"test^stepX\"\n# e.g.\nmake \"test^step2\"\nmake \"test^step7\"\n<\/pre>\n<ul><li>To run tests for a specific step against a single implementation:<\/li>\n<\/ul><pre>make \"test^IMPL^stepX\"\n# e.g\nmake \"test^ruby^step3\"\nmake \"test^ps^step4\"\n<\/pre>\n<h3><a id=\"user-content-self-hosted-functional-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#self-hosted-functional-tests\"><\/a>Self-hosted functional tests<\/h3>\n<ul><li>To run the functional tests in self-hosted mode, you specify <code>mal<\/code>\nas the test implementation and use the <code>MAL_IMPL<\/code> make variable\nto change the underlying host language (default is JavaScript):<\/li>\n<\/ul><pre>make MAL_IMPL=IMPL \"test^mal^step2\"\n# e.g.\nmake \"test^mal^step2\"   # js is default\nmake MAL_IMPL=ruby \"test^mal^step2\"\nmake MAL_IMPL=python \"test^mal^step2\"\n<\/pre>\n<h3><a id=\"user-content-starting-the-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#starting-the-repl\"><\/a>Starting the REPL<\/h3>\n<ul><li>To start the REPL of an implementation in a specific step:<\/li>\n<\/ul><pre>make \"repl^IMPL^stepX\"\n# e.g\nmake \"repl^ruby^step3\"\nmake \"repl^ps^step4\"\n<\/pre>\n<ul><li>If you omit the step, then <code>stepA<\/code> is used:<\/li>\n<\/ul><pre>make \"repl^IMPL\"\n# e.g\nmake \"repl^ruby\"\nmake \"repl^ps\"\n<\/pre>\n<ul><li>To start the REPL of the self-hosted implementation, specify <code>mal<\/code> as the\nREPL implementation and use the <code>MAL_IMPL<\/code> make variable to change the\nunderlying host language (default is JavaScript):<\/li>\n<\/ul><pre>make MAL_IMPL=IMPL \"repl^mal^stepX\"\n# e.g.\nmake \"repl^mal^step2\"   # js is default\nmake MAL_IMPL=ruby \"repl^mal^step2\"\nmake MAL_IMPL=python \"repl^mal\"\n<\/pre>\n<h3><a id=\"user-content-performance-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance-tests\"><\/a>Performance tests<\/h3>\n<p>Warning: These performance tests are neither statistically valid nor\ncomprehensive; runtime performance is a not a primary goal of mal. If\nyou draw any serious conclusions from these performance tests, then\nplease contact me about some amazing oceanfront property in Kansas\nthat I'm willing to sell you for cheap.<\/p>\n<ul><li>To run performance tests against a single implementation:<\/li>\n<\/ul><pre>make \"perf^IMPL\"\n# e.g.\nmake \"perf^js\"\n<\/pre>\n<ul><li>To run performance tests against all implementations:<\/li>\n<\/ul><pre>make \"perf\"\n<\/pre>\n<h3><a id=\"user-content-generating-language-statistics\" class=\"anchor\" aria-hidden=\"true\" href=\"#generating-language-statistics\"><\/a>Generating language statistics<\/h3>\n<ul><li>To report line and byte statistics for a single implementation:<\/li>\n<\/ul><pre>make \"stats^IMPL\"\n# e.g.\nmake \"stats^js\"\n<\/pre>\n<ul><li>To report line and bytes statistics for general Lisp code (env, core\nand stepA):<\/li>\n<\/ul><pre>make \"stats-lisp^IMPL\"\n# e.g.\nmake \"stats-lisp^js\"\n<\/pre>\n<h2><a id=\"user-content-dockerized-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#dockerized-testing\"><\/a>Dockerized testing<\/h2>\n<p>Every implementation directory contains a Dockerfile to create\na docker image containing all the dependencies for that\nimplementation. In addition, the top-level Makefile contains support\nfor running the tests target (and perf, stats, repl, etc) within\na docker container for that implementation by passing <em>\"DOCKERIZE=1\"<\/em>\non the make command line. For example:<\/p>\n<pre>make DOCKERIZE=1 \"test^js^step3\"\n<\/pre>\n<p>Existing implementations already have docker images built and pushed\nto the docker registry. However, if\nyou wish to build or rebuild a docker image locally, the toplevel\nMakefile provides a rule for building docker images:<\/p>\n<pre>make \"docker-build^IMPL\"\n<\/pre>\n<p><strong>Notes<\/strong>:<\/p>\n<ul><li>Docker images are named <em>\"kanaka\/mal-test-IMPL\"<\/em><\/li>\n<li>JVM-based language implementations (Groovy, Java, Clojure, Scala):\nyou will probably need to run this command once manually\nfirst <code>make DOCKERIZE=1 \"repl^IMPL\"<\/code> before you can run tests because\nruntime dependencies need to be downloaded to avoid the tests timing\nout. These dependencies are downloaded to dot-files in the \/mal\ndirectory so they will persist between runs.<\/li>\n<\/ul><h2><a id=\"user-content-external-implementations\" class=\"anchor\" aria-hidden=\"true\" href=\"#external-implementations\"><\/a>External Implementations<\/h2>\n<p>The following implementations are maintained as separate projects:<\/p>\n<h3><a id=\"user-content-holyc\" class=\"anchor\" aria-hidden=\"true\" href=\"#holyc\"><\/a>HolyC<\/h3>\n<h3><a id=\"user-content-rust\" class=\"anchor\" aria-hidden=\"true\" href=\"#rust\"><\/a>Rust<\/h3>\n<ul><li><a href=\"https:\/\/github.com\/seven1m\/mal-rust\">by Tim Morgan<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/vi\/mal-rust-vi\">by vi<\/a> - using <a href=\"https:\/\/pest.rs\/\" rel=\"nofollow\">Pest<\/a> grammar, not using typical Mal infrastructure (cargo-ized steps and built-in converted tests).<\/li>\n<\/ul><h2><a id=\"user-content-other-mal-projects\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-mal-projects\"><\/a>Other mal Projects<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/dubek\/malc\">malc<\/a> - Mal (Make A Lisp) compiler. Compiles a Mal program to LLVM assembly language, then binary.<\/li>\n<li><a href=\"https:\/\/git.sr.ht\/~tim\/malcc\" rel=\"nofollow\">malcc<\/a> (@seven1m) - malcc is an incremental compiler implementation for the Mal language. It uses the Tiny C Compiler as the compiler backend and has full support for the Mal language, including macros, tail-call elimination, and even run-time eval. <a href=\"https:\/\/mpov.timmorgan.org\/i-built-a-lisp-compiler\/\" rel=\"nofollow\">\"I Built a Lisp Compiler\"<\/a> post about the process.<\/li>\n<li><a href=\"https:\/\/github.com\/chr15m\/frock\">frock<\/a> - Clojure-flavoured PHP. Uses mal\/php to run programs.<\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Mal (make-a-lisp) is licensed under the MPL 2.0 (Mozilla Public\nLicense 2.0). See LICENSE.txt for more details.<\/p>\n<\/article>","created_at":"2019-03-28T14:26:07+0000","updated_at":"2019-03-28T14:26:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":17,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/70127?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":889,"title":"YAAC: Another Awesome CV - Overleaf, Online LaTeX Editor","url":"https:\/\/www.overleaf.com\/latex\/templates\/awesome-source-cv\/wrdjtkkytqcw","content":"<p>YAAC: Another Awesome CV is a template that use FontAwesome and Source Sans Pro Font. It also use the Lualatex engine. <\/p><p>This template also use elements from <a href=\"https:\/\/www.overleaf.com\/latex\/templates\/altacv-template\/trgqjpwnmtgv#.WB0G6BKLRZh\" rel=\"nofollow\">AltaCV<\/a> like cv tags and cv skills. <\/p><b>Changelog: <\/b><ul><li><u>Version 1.9.2<\/u>: New <code>project<\/code> section\n<\/li><li><u>Version 1.9<\/u>: Add <code>monochrome<\/code> option, add <code>showLinks<\/code>option<\/li>\n<li><u>Version 1.8<\/u>: Add macros to handle referees and keywords<\/li>\n<li><u>Version 1.7<\/u>: Renaming and minor changes<\/li>\n<li><u>Version 1.6.2<\/u>: Add fonction to put two columns on a section.<\/li>\n<li><u>Version 1.6<\/u>: Use Lualatex as default engine.<\/li>\n<li><u>Version 1.5<\/u>: New alternative option to change the way the header is generated (<code>\\documentclass[localFont,alternative]{awesome-source-cv}<\/code>) <\/li>\n<\/ul><b>More infos:<\/b><p>Full source and documentation can be downloaded on \n<a href=\"https:\/\/github.com\/darwiin\/yaac-another-awesome-cv\" rel=\"nofollow\">Github<\/a><\/p>","created_at":"2019-03-25T14:23:28+0000","updated_at":"2019-03-25T14:23:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.overleaf.com","preview_picture":"https:\/\/cdn.sharelatex.com\/img\/ol-brand\/overleaf_og_logo.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":888,"title":"Harmopark","url":"https:\/\/www.harmopark.app\/","content":"<header><main id=\"main\" role=\"main\" class=\"container-fluid\"><div id=\"loading\"><div class=\"center\"><h5 class=\"text-center text-danger\">Loading<\/h5><\/div><\/div>&#13;\n\t\t<div id=\"htToolBar\" class=\"d-flex justify-content-center align-items-center flex-wrap invisible\" role=\"toolbar\" aria-label=\"Harmonic table toolbar\"><div class=\"d-flex justify-content-center flex-wrap align-items-center\"><div class=\"d-flex\"><div class=\"modal fade\" id=\"keyHarmony\" tabindex=\"-1\" role=\"dialog\" aria-labelledby=\"keyHarmonyLabel\" aria-hidden=\"true\"><div class=\"modal-dialog modal-lg\" role=\"document\"><div class=\"modal-content\"><p>&#13;\n\t\t\t\t\t\t\t\t\t<\/p><h5 class=\"modal-title\" id=\"keyHarmonyLabel\">Diatonic harmony for <\/h5>&#13;\n\t\t\t\t\t\t\t\t\t&#13;<div class=\"modal-body\"><table class=\"table table-hover\" summary=\"Key chords\"><caption>These are chords built off of each degree of the scale.<br \/>This table is not exhaustive; there may be other chords available.<\/caption>\n\t\t\t\t\t\t\t\t\t\t<tbody><tr><th scope=\"row\">Roman numerals<\/th><\/tr><tr><th scope=\"row\">Triads<\/th><\/tr><tr><th scope=\"row\">Tetrachords<\/th><\/tr><tr><th scope=\"row\">Pentachords<\/th><\/tr><tr><th scope=\"row\">Hexachords<\/th><\/tr><tr><th scope=\"row\">Heptachords<\/th><\/tr><tr class=\"d-none\"><th scope=\"row\">Octachords<\/th><\/tr><tr class=\"d-none\"><th scope=\"row\">Nonachords<\/th><\/tr><tr class=\"d-none\"><th scope=\"row\">Decachords<\/th><\/tr><tr class=\"d-none\"><th scope=\"row\">Undecachords<\/th><\/tr><tr class=\"d-none\"><th scope=\"row\">Dodecachords<\/th><\/tr><\/tbody><\/table><p>Alternative names for this mode: <\/p><\/div><\/div><\/div><\/div><div class=\"modal fade\" id=\"soundSettings\" tabindex=\"-1\" role=\"dialog\" aria-labelledby=\"soundSettingsLabel\" aria-hidden=\"true\"><div class=\"modal-dialog\" role=\"document\"><div class=\"modal-content\"><p>&#13;\n\t\t\t\t\t\t\t\t\t<\/p><h5 class=\"modal-title\" id=\"soundSettingsLabel\">Sound settings<\/h5>&#13;\n\t\t\t\t\t\t\t\t\t&#13;<div class=\"modal-body\"><form action=\"action\"><div>&#13;\n\t\t\t\t\t\t\t\t\t\t<div class=\"tab-content\"><div class=\"tab-pane fade show active\" id=\"keyboardSettings\" role=\"tabpanel\" aria-labelledby=\"keyboard-tab\"><div class=\"form-row d-flex justify-content-around bg-light rounded mt-2 mb-2\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"htVolume\">Volume<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"synthPreset\">Synth preset<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><div id=\"htFX\"><label>FX<\/label><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"addChorus\">Chorus<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"addTremolo\">Tremolo<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"addDelay\">Delay<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"addReverb\">Reverb<\/label>&#13;<\/p><\/div><\/div><div id=\"chorusSettings\"><h6 class=\"text-center\">Chorus<\/h6><div class=\"form-row d-flex justify-content-around bg-light mb-2\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"chorusFrequency\">Frequency<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"chorusDelayTime\">Delay<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"chorusDepth\">Depth<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"chorusWet\">Wet mix<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><div id=\"tremoloSettings\"><h6 class=\"text-center\">Tremolo<\/h6><div class=\"form-row d-flex justify-content-around bg-light rounded mb-2\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"tremoloDepth\">Frequency<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"tremoloType\">Osc. type<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"tremoloDepth\">Depth<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"tremoloWet\">Wet mix<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><div id=\"delaySettings\"><h6 class=\"text-center\">Delay<\/h6><div class=\"form-row d-flex justify-content-around bg-light rounded mb-2\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"delayTime\">Time<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"delayFeedback\">Feedback<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"delayWet\">Wet mix<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><div id=\"reverbSettings\"><h6 class=\"text-center\">Reverb<\/h6><div class=\"form-row d-flex justify-content-around bg-light rounded mb-2\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"reverbPreDelay\">Pre-delay<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"reverbDecay\">Decay<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"reverbWet\">Wet mix<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"tab-pane fade\" id=\"progressionsSettings\" role=\"tabpanel\" aria-labelledby=\"progressions-tab\"><div class=\"form-row d-flex justify-content-around bg-light rounded mt-2 mb-2\"><div class=\"form-group\"><label for=\"chordDuration\">Chord duration<\/label><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"chordDurationWhole\">Whole note<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"chordDurationHalf\">Half note<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"chordDurationQuarter\">Quarter note<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"chordDurationEighth\">Eighth note<\/label>&#13;<\/p><\/div><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"progressionTempo\">Playback tempo<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><div class=\"form-group d-flex flex-column justify-content-center\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"arpeggiateChords\">Arpeggiate chords<\/label>&#13;<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label class=\"custom-control-label\" for=\"addBeat\">Add drum beat<\/label>&#13;<\/p><\/div><\/div><div id=\"beatSettings\"><h6 class=\"text-center\">Drum beat volume<\/h6><div class=\"form-row d-flex justify-content-around bg-light mb-2\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"kickVolume\">Kick volume<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"snareVolume\">Snare volume<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<label for=\"hiHatVolume\">Hi-hat volume<\/label>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><\/div>&#13;\n\t\t\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\t\t<\/div><\/form><\/div><\/div><\/div><\/div><\/div><p>&#13;\n\t\t\t\t\t<label for=\"selectionSwitch\" class=\"unselectable\"><em>Play mode<\/em><\/label>&#13;\n\t\t\t\t\t<label class=\"switch ml-2\">&#13;\n\t\t\t\t\t\t<\/label>&#13;<\/p><\/div><\/div>&#13;\n\t\t<div id=\"htArea\"><div id=\"htSelection\" class=\"card shadow-sm invisible\"><p>&#13;\n\t  \t\t\t\t<\/p><h6 class=\"unselectable\">Selection info<\/h6>&#13;\n\t  \t\t\t\t&#13;\n\t  \t\t\t\t\t\t\t\t\t&#13;<div class=\"card-body text-secondary text-center\"><p>&#13;\n    \t\t\t\t\t<\/p><h6 id=\"pcTitle\" class=\"card-title text-center unselectable\">Pitch constellation<\/h6>&#13;<div><h6 class=\"card-title text-center unselectable\">Selected pitches<\/h6><p id=\"selectionVoicing\" class=\"text-danger unselectable\">&#13;\n\t    \t\t\t\t\t<em>none<\/em>&#13;\n\t    \t\t\t\t<\/p><h6 class=\"card-title text-center unselectable\">Chord<\/h6><div id=\"selectionNames\" class=\"carousel slide\" data-interval=\"false\"><a class=\"carousel-control-prev d-none\" href=\"#selectionNames\" role=\"button\" data-slide=\"prev\">&#13;\n\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\tPrevious&#13;\n\t\t\t\t\t\t\t<\/a><a class=\"carousel-control-next d-none\" href=\"#selectionNames\" role=\"button\" data-slide=\"next\">&#13;\n\t\t\t\t\t\t\t\t&#13;\n\t\t\t\t\t\t\t\tNext&#13;\n\t\t\t\t\t\t\t<\/a><\/div><p>&#13;\n\t    \t\t\t\t\t<em class=\"unselectable pl-2\">To build a chord, right-click on the desired keys in the Harmonic Table.<\/em>&#13;<\/p><\/div><\/div><\/div><\/div>&#13;\n\t\t<div id=\"chordProgressions\"><div id=\"parallelProgression\" class=\"progression border-success shadow-sm mt-4 d-none\"><h6 class=\"title unselectable\">Parallel progression<\/h6><\/div><div id=\"negativeProgression\" class=\"progression border-secondary shadow-sm mt-4 d-none\"><h6 class=\"title unselectable\">Negative progression<\/h6><div id=\"negativeProgressionFields\" class=\"d-flex\"><div id=\"negativeProgressionAxisInputGroup\" class=\"input-group input-group-sm\"><p>&#13;\n\t\t\t\t\t\t\t<label class=\"input-group-text text-white bg-danger\" for=\"negativeProgressionAxis\">Symmetry Axis<\/label>&#13;<\/p><\/div><\/div><\/div><\/div>&#13;\n\t\t&#13;\n\t<\/main><\/header>","created_at":"2019-03-25T13:24:42+0000","updated_at":"2019-03-25T13:24:42+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.harmopark.app","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure"],"is_public":false,"id":887,"title":"Boilerplate for running Clojure as a shebang script","url":"https:\/\/gist.github.com\/ericnormand\/6bb4562c4bc578ef223182e3bb1e72c5?__s=vrd97cronavsf4qzuzhx","content":"<div class=\"application-main\" data-commit-hovercards-enabled=\"\"><div itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Code\"><main id=\"gist-pjax-container\" class=\"gist-content-wrapper\" data-pjax-container=\"\">\n<div class=\"container-lg px-3 new-discussion-timeline experiment-repo-nav\"><div class=\"repository-content gist-content\"><div><p>Boilerplate for running Clojure as a shebang script<\/p><div class=\"js-gist-file-update-container js-task-list-container file-box\"><div id=\"file-00_script-clj\" class=\"file\"><p><\/p><table class=\"highlight tab-size js-file-line-container\" data-tab-size=\"8\"><tr><td id=\"file-00_script-clj-L1\" class=\"blob-num js-line-number\" data-line-number=\"1\">\n        <\/td><td id=\"file-00_script-clj-LC1\" class=\"blob-code blob-code-inner js-file-line\">#!\/bin\/sh<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L2\" class=\"blob-num js-line-number\" data-line-number=\"2\">\n        <\/td><td id=\"file-00_script-clj-LC2\" class=\"blob-code blob-code-inner js-file-line\">#_(<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L3\" class=\"blob-num js-line-number\" data-line-number=\"3\">\n        <\/td><td id=\"file-00_script-clj-LC3\" class=\"blob-code blob-code-inner js-file-line\">\n<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L4\" class=\"blob-num js-line-number\" data-line-number=\"4\">\n        <\/td><td id=\"file-00_script-clj-LC4\" class=\"blob-code blob-code-inner js-file-line\">   #_DEPS is same format as deps.edn. Multiline is okay.<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L5\" class=\"blob-num js-line-number\" data-line-number=\"5\">\n        <\/td><td id=\"file-00_script-clj-LC5\" class=\"blob-code blob-code-inner js-file-line\">   DEPS='<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L6\" class=\"blob-num js-line-number\" data-line-number=\"6\">\n        <\/td><td id=\"file-00_script-clj-LC6\" class=\"blob-code blob-code-inner js-file-line\">   {:deps {clj-time {:mvn\/version \"0.14.2\"}}}<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L7\" class=\"blob-num js-line-number\" data-line-number=\"7\">\n        <\/td><td id=\"file-00_script-clj-LC7\" class=\"blob-code blob-code-inner js-file-line\">   '<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L8\" class=\"blob-num js-line-number\" data-line-number=\"8\">\n        <\/td><td id=\"file-00_script-clj-LC8\" class=\"blob-code blob-code-inner js-file-line\">\n<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L9\" class=\"blob-num js-line-number\" data-line-number=\"9\">\n        <\/td><td id=\"file-00_script-clj-LC9\" class=\"blob-code blob-code-inner js-file-line\">   #_You can put other options here<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L10\" class=\"blob-num js-line-number\" data-line-number=\"10\">\n        <\/td><td id=\"file-00_script-clj-LC10\" class=\"blob-code blob-code-inner js-file-line\">   OPTS='<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L11\" class=\"blob-num js-line-number\" data-line-number=\"11\">\n        <\/td><td id=\"file-00_script-clj-LC11\" class=\"blob-code blob-code-inner js-file-line\">   -J-Xms256m -J-Xmx256m -J-client<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L12\" class=\"blob-num js-line-number\" data-line-number=\"12\">\n        <\/td><td id=\"file-00_script-clj-LC12\" class=\"blob-code blob-code-inner js-file-line\">   '<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L13\" class=\"blob-num js-line-number\" data-line-number=\"13\">\n        <\/td><td id=\"file-00_script-clj-LC13\" class=\"blob-code blob-code-inner js-file-line\">\n<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L14\" class=\"blob-num js-line-number\" data-line-number=\"14\">\n        <\/td><td id=\"file-00_script-clj-LC14\" class=\"blob-code blob-code-inner js-file-line\">exec clojure $OPTS -Sdeps \"$DEPS\" \"$0\" \"$@\"<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L15\" class=\"blob-num js-line-number\" data-line-number=\"15\">\n        <\/td><td id=\"file-00_script-clj-LC15\" class=\"blob-code blob-code-inner js-file-line\">\n<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L16\" class=\"blob-num js-line-number\" data-line-number=\"16\">\n        <\/td><td id=\"file-00_script-clj-LC16\" class=\"blob-code blob-code-inner js-file-line\">)<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L17\" class=\"blob-num js-line-number\" data-line-number=\"17\">\n        <\/td><td id=\"file-00_script-clj-LC17\" class=\"blob-code blob-code-inner js-file-line\">\n<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L18\" class=\"blob-num js-line-number\" data-line-number=\"18\">\n        <\/td><td id=\"file-00_script-clj-LC18\" class=\"blob-code blob-code-inner js-file-line\">(println \"Hello!\")<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L19\" class=\"blob-num js-line-number\" data-line-number=\"19\">\n        <\/td><td id=\"file-00_script-clj-LC19\" class=\"blob-code blob-code-inner js-file-line\">\n<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L20\" class=\"blob-num js-line-number\" data-line-number=\"20\">\n        <\/td><td id=\"file-00_script-clj-LC20\" class=\"blob-code blob-code-inner js-file-line\">(require '[clj-time.core :as t])<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L21\" class=\"blob-num js-line-number\" data-line-number=\"21\">\n        <\/td><td id=\"file-00_script-clj-LC21\" class=\"blob-code blob-code-inner js-file-line\">\n      <\/td><\/tr><tr><td id=\"file-00_script-clj-L22\" class=\"blob-num js-line-number\" data-line-number=\"22\">\n        <\/td><td id=\"file-00_script-clj-LC22\" class=\"blob-code blob-code-inner js-file-line\">(prn (str (t\/now)))<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L23\" class=\"blob-num js-line-number\" data-line-number=\"23\">\n        <\/td><td id=\"file-00_script-clj-LC23\" class=\"blob-code blob-code-inner js-file-line\">\n      <\/td><\/tr><tr><td id=\"file-00_script-clj-L24\" class=\"blob-num js-line-number\" data-line-number=\"24\">\n        <\/td><td id=\"file-00_script-clj-LC24\" class=\"blob-code blob-code-inner js-file-line\">(prn *command-line-args*)<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L25\" class=\"blob-num js-line-number\" data-line-number=\"25\">\n        <\/td><td id=\"file-00_script-clj-LC25\" class=\"blob-code blob-code-inner js-file-line\">\n      <\/td><\/tr><tr><td id=\"file-00_script-clj-L26\" class=\"blob-num js-line-number\" data-line-number=\"26\">\n        <\/td><td id=\"file-00_script-clj-LC26\" class=\"blob-code blob-code-inner js-file-line\">(println (.. (Runtime\/getRuntime)<\/td>\n      <\/tr><tr><td id=\"file-00_script-clj-L27\" class=\"blob-num js-line-number\" data-line-number=\"27\">\n        <\/td><td id=\"file-00_script-clj-LC27\" class=\"blob-code blob-code-inner js-file-line\">             totalMemory))<\/td>\n      <\/tr><\/table><\/div><\/div><\/div><\/div><\/div>\n    <\/main><\/div><\/div>\n        \n<div class=\"footer container-lg width-full px-3\" role=\"contentinfo\"><div class=\"position-relative d-flex flex-justify-between pt-6 pb-2 mt-6 f6 text-gray border-top border-gray-light\"><ul class=\"list-style-none d-flex flex-wrap\"><li class=\"mr-3\">\u00a9 2019 GitHub, Inc.<\/li>\n        <li class=\"mr-3\"><a data-ga-click=\"Footer, go to terms, text:terms\" href=\"https:\/\/github.com\/site\/terms\">Terms<\/a><\/li>\n        <li class=\"mr-3\"><a data-ga-click=\"Footer, go to privacy, text:privacy\" href=\"https:\/\/github.com\/site\/privacy\">Privacy<\/a><\/li>\n        <li class=\"mr-3\"><a data-ga-click=\"Footer, go to security, text:security\" href=\"https:\/\/github.com\/security\">Security<\/a><\/li>\n        <li class=\"mr-3\"><a href=\"https:\/\/githubstatus.com\/\" data-ga-click=\"Footer, go to status, text:status\">Status<\/a><\/li>\n        <li><a data-ga-click=\"Footer, go to help, text:help\" href=\"https:\/\/help.github.com\">Help<\/a><\/li>\n    <\/ul><ul class=\"list-style-none d-flex flex-wrap\"><li class=\"mr-3\"><a data-ga-click=\"Footer, go to contact, text:contact\" href=\"https:\/\/github.com\/contact\">Contact GitHub<\/a><\/li>\n        <li class=\"mr-3\"><a href=\"https:\/\/github.com\/pricing\" data-ga-click=\"Footer, go to Pricing, text:Pricing\">Pricing<\/a><\/li>\n      <li class=\"mr-3\"><a href=\"https:\/\/developer.github.com\" data-ga-click=\"Footer, go to api, text:api\">API<\/a><\/li>\n      <li class=\"mr-3\"><a href=\"https:\/\/training.github.com\" data-ga-click=\"Footer, go to training, text:training\">Training<\/a><\/li>\n        <li class=\"mr-3\"><a href=\"https:\/\/github.blog\" data-ga-click=\"Footer, go to blog, text:blog\">Blog<\/a><\/li>\n        <li><a data-ga-click=\"Footer, go to about, text:about\" href=\"https:\/\/github.com\/about\">About<\/a><\/li>\n    <\/ul><\/div><\/div>\n  <div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n    \n  \n  <details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark\" open=\"open\"><summary aria-haspopup=\"dialog\" aria-label=\"Close dialog\">\n      \n    <\/summary><\/details>","created_at":"2019-03-25T13:16:09+0000","updated_at":"2019-03-25T13:16:12+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"gist.github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/116172?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":886,"title":"valeriansaliou\/sonic","url":"https:\/\/github.com\/valeriansaliou\/sonic","content":"<p><a href=\"https:\/\/travis-ci.org\/valeriansaliou\/sonic\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/febc15a023970a72892151c3b3b4c1789206227e\/68747470733a2f2f7472617669732d63692e6f72672f76616c657269616e73616c696f752f736f6e69632e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/valeriansaliou\/sonic.svg?branch=master\" \/><\/a> <a href=\"https:\/\/deps.rs\/repo\/github\/valeriansaliou\/sonic\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/620601563e7418cea51909969a68b6dad09ef4aa\/68747470733a2f2f646570732e72732f7265706f2f6769746875622f76616c657269616e73616c696f752f736f6e69632f7374617475732e737667\" alt=\"Dependency Status\" data-canonical-src=\"https:\/\/deps.rs\/repo\/github\/valeriansaliou\/sonic\/status.svg\" \/><\/a> <a href=\"https:\/\/www.buymeacoffee.com\/valeriansaliou\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/cd005dca0ef55d7725912ec03a936d3a7c8de5b5\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6275792532306d6525323061253230636f666665652d646f6e6174652d79656c6c6f772e737667\" alt=\"Buy Me A Coffee\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/buy%20me%20a%20coffee-donate-yellow.svg\" \/><\/a><\/p><p><strong>Sonic is a fast, lightweight and schema-less search backend. It ingests search texts and identifier tuples, that can then be queried against in microseconds time.<\/strong><\/p><p>Sonic can be used as a simple alternative to super-heavy and full-featured search backends such as Elasticsearch in some use-cases. It is capable of normalizing natural language search queries, auto-completing a search query and providing the most relevant results for a query. Sonic is an identifier index, rather than a document index; when queried, it returns IDs that can then be used to refer to the matched documents in an external database.<\/p><p>A strong attention to performance and code cleanliness has been given when designing Sonic. It aims at being crash-free, super-fast and puts minimum strain on server resources (our measurements have shown that Sonic - when under load - responds to search queries in the \u03bcs range, eats ~30MB RAM and has a low CPU footprint; <a href=\"https:\/\/github.com\/valeriansaliou\/sonic#how-fast--lightweight-is-it\">see our benchmarks<\/a>).<\/p><p><strong>\ud83c\uddeb\ud83c\uddf7 Crafted in Nantes, France.<\/strong><\/p><p><strong>\ud83d\udcf0 The Sonic project was initially announced in <a href=\"https:\/\/journal.valeriansaliou.name\/announcing-sonic-a-super-light-alternative-to-elasticsearch\/\" rel=\"nofollow\">a post on my personal journal<\/a>.<\/strong><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/bcc95ee95ed8b384ce091330ebc06f7c89910166\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f62616e6e65722e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/bcc95ee95ed8b384ce091330ebc06f7c89910166\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f62616e6e65722e6a7067\" alt=\"Sonic\" data-canonical-src=\"https:\/\/valeriansaliou.github.io\/sonic\/images\/banner.jpg\" \/><\/a><\/p><blockquote>\n<p><em>\u00ab Sonic \u00bb is the mascot of the Sonic project. I drew it to look like a psychedelic hipster hedgehog.<\/em><\/p>\n<\/blockquote><h2>Who uses it?<\/h2><table><tbody><tr><td align=\"center\"><a href=\"https:\/\/crisp.chat\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d8130493704c28c97f38ec64980e94397b5db6af\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f6c6f676f2d63726973702e706e67\" height=\"64\" data-canonical-src=\"https:\/\/valeriansaliou.github.io\/sonic\/images\/logo-crisp.png\" alt=\"image\" \/><\/a><\/td>\n<\/tr><tr><td align=\"center\">Crisp<\/td>\n<\/tr><\/tbody><\/table><p><em>\ud83d\udc4b You use Sonic and you want to be listed there? <a href=\"https:\/\/valeriansaliou.name\/\" rel=\"nofollow\">Contact me<\/a>.<\/em><\/p><h2>Demo<\/h2><p>Sonic is integrated in all Crisp search products on the <a href=\"https:\/\/crisp.chat\/\" rel=\"nofollow\">Crisp<\/a> platform. It is used to index half a billion objects on a $5\/mth 1-vCPU SSD cloud server (as of 2019). Crisp users use it to search in their messages, conversations, contacts, helpdesk articles and more.<\/p><p><strong>You can test Sonic live on: <a href=\"https:\/\/help.crisp.chat\/\" rel=\"nofollow\">Crisp Helpdesk<\/a>, and get an idea of the speed and relevance of Sonic search results.<\/strong><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/8588780586f080dc84e80f6af67ce0d64dbbc779\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f63726973702d7365617263682d64656d6f2e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/8588780586f080dc84e80f6af67ce0d64dbbc779\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f63726973702d7365617263682d64656d6f2e676966\" alt=\"Demo on Crisp Helpdesk search\" data-canonical-src=\"https:\/\/valeriansaliou.github.io\/sonic\/images\/crisp-search-demo.gif\" \/><\/a><\/p><blockquote>\n<p><em>Sonic fuzzy search in helpdesk articles at its best. Lookup for any word or group of terms, get results instantly.<\/em><\/p>\n<\/blockquote><h2>Features<\/h2><ul><li><strong>Search terms are stored in collections, organized in buckets<\/strong>; you may use a single bucket, or a bucket per user on your platform if you need to search in separate indexes.<\/li>\n<li><strong>Search results return object identifiers<\/strong>, that can be resolved from an external database if you need to enrich the search results. This makes Sonic a simple word index, that points to identifier results. Sonic doesn't store any direct textual data in its index, but it still holds a word graph for auto-completion and typo corrections.<\/li>\n<li><strong>Search query typos are corrected<\/strong> if there are not enough exact-match results for a given word in a search query, Sonic tries to correct the word and tries against alternate words. You're allowed to make mistakes when searching.<\/li>\n<li><strong>Insert and remove items in the index<\/strong>; index-altering operations are light and can be commited to the server while it is running. A background tasker handles the job of consolidating the index so that the entries you have pushed or popped are quickly made available for search.<\/li>\n<li><strong>Auto-complete any word<\/strong> in real-time via the suggest operation. This helps build a snappy word suggestion feature in your end-user search interface.<\/li>\n<li><strong>Full Unicode compatibility<\/strong> on 80+ most spoken languages in the word. Sonic removes useless stop words from any text (eg. 'the' in English), after guessing the text language. This ensures any searched or ingested text is clean before it hits the index; <a href=\"https:\/\/github.com\/valeriansaliou\/sonic#which-text-languages-are-supported\">see languages<\/a>.<\/li>\n<li><strong>Networked channel interface (Sonic Channel)<\/strong>, that let you search your index, manage data ingestion (push in the index, pop from the index, flush a collection, flush a bucket, etc.) and perform administrative actions. The Sonic Channel protocol was designed to be lightweight on resources and simple to integrate with (the protocol is specified in the sections below); <a href=\"https:\/\/github.com\/valeriansaliou\/sonic\/blob\/master\/PROTOCOL.md\">read protocol specification<\/a>.<\/li>\n<li><strong>Easy-to-use libraries<\/strong>, that let you connect to Sonic Channel from your apps; <a href=\"https:\/\/github.com\/valeriansaliou\/sonic#-sonic-channel-libraries\">see libraries<\/a>.<\/li>\n<\/ul><h2>How to use it?<\/h2><h3>Installation<\/h3><p>Sonic is built in Rust. To install it, either download a version from the <a href=\"https:\/\/github.com\/valeriansaliou\/sonic\/releases\">Sonic releases<\/a> page, use <code>cargo install<\/code> or pull the source code from <code>master<\/code>.<\/p><p><strong>\ud83d\udc49 Install from source:<\/strong><\/p><p>If you pulled the source code from Git, you can build it using <code>cargo<\/code>:<\/p><div class=\"highlight highlight-source-shell\"><pre>cargo build --release<\/pre><\/div><p>You can find the built binaries in the <code>.\/target\/release<\/code> directory.<\/p><p><em>Install <code>clang<\/code> to be able to compile the required RocksDB dependency.<\/em><\/p><p><strong>\ud83d\udc49 Install from Cargo:<\/strong><\/p><p>You can install Sonic directly with <code>cargo install<\/code>:<\/p><div class=\"highlight highlight-source-shell\"><pre>cargo install sonic-server<\/pre><\/div><p>Ensure that your <code>$PATH<\/code> is properly configured to source the Crates binaries, and then run Sonic using the <code>sonic<\/code> command.<\/p><p><strong>\ud83d\udc49 Install from Docker Hub:<\/strong><\/p><p>You might find it convenient to run Sonic via Docker. You can find the pre-built Sonic image on Docker Hub as <a href=\"https:\/\/hub.docker.com\/r\/valeriansaliou\/sonic\/\" rel=\"nofollow\">valeriansaliou\/sonic<\/a>.<\/p><p>First, pull the <code>valeriansaliou\/sonic<\/code> image:<\/p><div class=\"highlight highlight-source-shell\"><pre>docker pull valeriansaliou\/sonic:v1.1.0<\/pre><\/div><p>Then, seed it a configuration file and run it (replace <code>\/path\/to\/your\/sonic\/config.cfg<\/code> with the path to your configuration file):<\/p><div class=\"highlight highlight-source-shell\"><pre>docker run -p 1491:1491 -v \/path\/to\/your\/sonic\/config.cfg:\/etc\/sonic.cfg -v \/path\/to\/your\/sonic\/store\/:\/var\/lib\/sonic\/store\/ valeriansaliou\/sonic:v1.1.0<\/pre><\/div><p>In the configuration file, ensure that:<\/p><ul><li><code>channel.inet<\/code> is set to <code>0.0.0.0:1491<\/code> (this lets Sonic Channel be reached from outside the container)<\/li>\n<li><code>store.kv.path<\/code> is set to <code>\/var\/lib\/sonic\/store\/kv\/<\/code> (this lets the external KV store directory be reached by Sonic)<\/li>\n<li><code>store.fst.path<\/code> is set to <code>\/var\/lib\/sonic\/store\/fst\/<\/code> (this lets the external FST store directory be reached by Sonic)<\/li>\n<\/ul><p>Sonic Channel will be reachable from <code>tcp:\/\/localhost:1491<\/code>.<\/p><h3>Configuration<\/h3><p>Use the sample <a href=\"https:\/\/github.com\/valeriansaliou\/sonic\/blob\/master\/config.cfg\">config.cfg<\/a> configuration file and adjust it to your own environment.<\/p><p><em>If you are looking to fine-tune your configuration, you may read our <a href=\"https:\/\/github.com\/valeriansaliou\/sonic\/blob\/master\/CONFIGURATION.md\">detailed configuration documentation<\/a>.<\/em><\/p><h3>Run Sonic<\/h3><p>Sonic can be run as such:<\/p><p><code>.\/sonic -c \/path\/to\/config.cfg<\/code><\/p><h2>Perform searches and manage objects<\/h2><p>Both searches and object management (ie. data ingestion) is handled via the Sonic Channel protocol only. As we want to keep things simple with Sonic (similarly to how Redis does), connecting to Sonic Channel is the way to go when you need to interact with the Sonic search database.<\/p><p>Sonic distributes official libraries, that let you integrate Sonic to your apps easily. Click on a library below to see library integration documentation and code.<\/p><p><em>If you are looking for details on the raw Sonic Channel TCP-based protocol, you can read our <a href=\"https:\/\/github.com\/valeriansaliou\/sonic\/blob\/master\/PROTOCOL.md\">detailed protocol documentation<\/a>. It can prove handy if you are looking to code your own Sonic Channel library.<\/em><\/p><h3>\ud83d\udce6 Sonic Channel Libraries<\/h3><p>Sonic distributes official Sonic Channel bindings for your programming language:<\/p><ul><li><strong>NodeJS<\/strong>: <strong><a href=\"https:\/\/www.npmjs.com\/package\/sonic-channel\" rel=\"nofollow\">node-sonic-channel<\/a><\/strong><\/li>\n<\/ul><p><em>\ud83d\udc49 Cannot find the library for your programming language? Build your own and be referenced here! (<a href=\"https:\/\/valeriansaliou.name\/\" rel=\"nofollow\">contact me<\/a>)<\/em><\/p><h2>Which text languages are supported?<\/h2><p>Sonic supports a wide range of languages in its lexing system. If a language is not in this list, you will still be able to push this language to the search index, but stop-words will not be eluded, which could lead to lower-quality search results.<\/p><p><strong>The languages supported by the lexing system are:<\/strong><\/p><ul><li>\ud83c\uddff\ud83c\udde6 Afrikaans<\/li>\n<li>\ud83c\uddf8\ud83c\udde6 Arabic<\/li>\n<li>\ud83c\udde6\ud83c\uddff Azerbaijani<\/li>\n<li>\ud83c\udde7\ud83c\udde9 Bengali<\/li>\n<li>\ud83c\udde7\ud83c\uddec Bulgarian<\/li>\n<li>\ud83c\uddf2\ud83c\uddf2 Burmese<\/li>\n<li>\ud83c\udde8\ud83c\uddf3 Chinese (Mandarin)<\/li>\n<li>\ud83c\udded\ud83c\uddf7 Croatian<\/li>\n<li>\ud83c\udde8\ud83c\uddff Czech<\/li>\n<li>\ud83c\udde9\ud83c\uddf0 Danish<\/li>\n<li>\ud83c\uddf3\ud83c\uddf1 Dutch<\/li>\n<li>\ud83c\uddfa\ud83c\uddf8 English<\/li>\n<li>\ud83c\udff3 Esperanto<\/li>\n<li>\ud83c\uddea\ud83c\uddea Estonian<\/li>\n<li>\ud83c\uddeb\ud83c\uddee Finnish<\/li>\n<li>\ud83c\uddeb\ud83c\uddf7 French<\/li>\n<li>\ud83c\udde9\ud83c\uddea German<\/li>\n<li>\ud83c\uddec\ud83c\uddf7 Greek<\/li>\n<li>\ud83c\uddf3\ud83c\uddec Hausa<\/li>\n<li>\ud83c\uddee\ud83c\uddf1 Hebrew<\/li>\n<li>\ud83c\uddee\ud83c\uddf3 Hindi<\/li>\n<li>\ud83c\udded\ud83c\uddfa Hungarian<\/li>\n<li>\ud83c\uddee\ud83c\udde9 Indonesian<\/li>\n<li>\ud83c\uddee\ud83c\uddf9 Italian<\/li>\n<li>\ud83c\uddef\ud83c\uddf5 Japanese<\/li>\n<li>\ud83c\uddf0\ud83c\udded Khmer<\/li>\n<li>\ud83c\uddf0\ud83c\uddf7 Korean<\/li>\n<li>\ud83c\udff3 Kurdish<\/li>\n<li>\ud83c\uddf1\ud83c\uddfb Latvian<\/li>\n<li>\ud83c\uddf1\ud83c\uddf9 Lithuanian<\/li>\n<li>\ud83c\uddee\ud83c\uddf3 Marathi<\/li>\n<li>\ud83c\uddf3\ud83c\uddf5 Nepali<\/li>\n<li>\ud83c\uddee\ud83c\uddf7 Persian<\/li>\n<li>\ud83c\uddf5\ud83c\uddf1 Polish<\/li>\n<li>\ud83c\uddf5\ud83c\uddf9 Portuguese<\/li>\n<li>\ud83c\uddee\ud83c\uddf3 Punjabi<\/li>\n<li>\ud83c\uddf7\ud83c\uddfa Russian<\/li>\n<li>\ud83c\uddf8\ud83c\uddee Slovene<\/li>\n<li>\ud83c\uddf8\ud83c\uddf4 Somali<\/li>\n<li>\ud83c\uddea\ud83c\uddf8 Spanish<\/li>\n<li>\ud83c\uddf8\ud83c\uddea Swedish<\/li>\n<li>\ud83c\uddf5\ud83c\udded Tagalog<\/li>\n<li>\ud83c\uddee\ud83c\uddf3 Tamil<\/li>\n<li>\ud83c\uddf9\ud83c\udded Thai<\/li>\n<li>\ud83c\uddf9\ud83c\uddf7 Turkish<\/li>\n<li>\ud83c\uddfa\ud83c\udde6 Ukrainian<\/li>\n<li>\ud83c\uddf5\ud83c\uddf0 Urdu<\/li>\n<li>\ud83c\uddfb\ud83c\uddf3 Vietnamese<\/li>\n<li>\ud83c\uddee\ud83c\uddf1 Yiddish<\/li>\n<li>\ud83c\uddf3\ud83c\uddec Yoruba<\/li>\n<li>\ud83c\uddff\ud83c\udde6 Zulu<\/li>\n<\/ul><h2>How fast &amp; lightweight is it?<\/h2><p>Sonic was built for <a href=\"https:\/\/crisp.chat\/\" rel=\"nofollow\">Crisp<\/a> from the start. As Crisp was growing and indexing more and more search data into a full-text search SQL database, we decided it was time to switch to a proper search backend system. When reviewing Elasticsearch (ELS) and others, we found those were full-featured heavyweight systems that did not scale well with Crisp's freemium-based cost structure.<\/p><p>At the end, we decided to build our own search backend, designed to be simple and lightweight on resources.<\/p><p>You can run function-level benchmarks with the command: <code>cargo bench --features benchmark<\/code><\/p><h3>\ud83d\udc69\u200d\ud83d\udd2c Benchmark #1<\/h3><h4>\u27a1\ufe0f Scenario<\/h4><p>We performed an extract of all messages from the Crisp team used for <a href=\"https:\/\/crisp.chat\/\" rel=\"nofollow\">Crisp<\/a> own customer support.<\/p><p>We want to import all those messages into a clean Sonic instance, and then perform searches on the index we built. We will measure the time that Sonic spent executing each operation (ie. each <code>PUSH<\/code> and <code>QUERY<\/code> commands over Sonic Channel), and group results per 1,000 operations (this outputs a mean time per 1,000 operations).<\/p><h4>\u27a1\ufe0f Context<\/h4><p><strong>Our benchmark is ran on the following computer:<\/strong><\/p><ul><li><strong>Device<\/strong>: MacBook Pro (Retina, 15-inch, Mid 2014)<\/li>\n<li><strong>OS<\/strong>: MacOS 10.14.3<\/li>\n<li><strong>Disk<\/strong>: 512GB SSD (formatted under the AFS file system)<\/li>\n<li><strong>CPU<\/strong>: 2.5 GHz Intel Core i7<\/li>\n<li><strong>RAM<\/strong>: 16 GB 1600 MHz DDR3<\/li>\n<\/ul><p><strong>Sonic is compiled as following:<\/strong><\/p><ul><li><strong>Sonic version<\/strong>: 1.0.1<\/li>\n<li><strong>Rustc version<\/strong>: <code>rustc 1.35.0-nightly (719b0d984 2019-03-13)<\/code><\/li>\n<li><strong>Compiler flags<\/strong>: <code>release<\/code> profile (<code>-03<\/code> with <code>lto<\/code>)<\/li>\n<\/ul><p><strong>Our dataset is as such:<\/strong><\/p><ul><li><strong>Number of objects<\/strong>: ~1,000,000 messages<\/li>\n<li><strong>Total size<\/strong>: ~100MB of raw message text (this does not account for identifiers and other metas)<\/li>\n<\/ul><h4>\u27a1\ufe0f Scripts<\/h4><p><strong>The scripts we used to perform the benchmark are:<\/strong><\/p><ol><li><strong>PUSH script<\/strong>: <a href=\"https:\/\/gist.github.com\/valeriansaliou\/e5ab737b28601ebd70483f904d21aa09\">sonic-benchmark_batch-push.js<\/a><\/li>\n<li><strong>QUERY script<\/strong>: <a href=\"https:\/\/gist.github.com\/valeriansaliou\/3ef8315d7282bd173c2cb9eba64fa739\">sonic-benchmark_batch-query.js<\/a><\/li>\n<\/ol><h4>\u23ec Results<\/h4><p><strong>Our findings:<\/strong><\/p><ul><li>We imported ~1,000,000 messages of dynamic length (some very long, eg. emails);<\/li>\n<li>Once imported, the search index weights 20MB (KV) + 1.4MB (FST) on disk;<\/li>\n<li>CPU usage during import averaged 75% of a single CPU core;<\/li>\n<li>RAM usage for the Sonic process peaked at 28MB during our benchmark;<\/li>\n<li>We used a single Sonic Channel TCP connection, which limits the import to a single thread (we could have load-balanced this across as many Sonic Channel connections as there are CPUs);<\/li>\n<li>We get an import RPS approaching 4,000 operations per second (per thread);<\/li>\n<li>We get a search query RPS approaching 1,000 operations per second (per thread);<\/li>\n<li>On the hyper-threaded 4-cores CPU used, we could have parallelized operations to 8 virtual cores, thus theoretically increasing the import RPS to 32,000 operations \/ second, while the search query RPS would be increased to 8,000 operations \/ second (we may be SSD-bound at some point though);<\/li>\n<\/ul><p><strong>Compared results per operation (on a single object):<\/strong><\/p><p>We took a sample of 8 results from our batched operations, which produced a total of 1,000 results (1,000,000 items, with 1,000 items batched per measurement report).<\/p><p><em>This is not very scientific, but it should give you a clear idea of Sonic performances.<\/em><\/p><p><strong>Time spent per operation:<\/strong><\/p><table><thead><tr><th>Operation<\/th>\n<th>Average<\/th>\n<th>Best<\/th>\n<th>Worst<\/th>\n<\/tr><\/thead><tbody><tr><td>PUSH<\/td>\n<td>275\u03bcs<\/td>\n<td>190\u03bcs<\/td>\n<td>363\u03bcs<\/td>\n<\/tr><tr><td>QUERY<\/td>\n<td>880\u03bcs<\/td>\n<td>852\u03bcs<\/td>\n<td>1ms<\/td>\n<\/tr><\/tbody><\/table><p><strong>Batch PUSH results as seen from our terminal (from initial index of: 0 objects):<\/strong><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/8e8c2acd0917b60d6a8897c076e7608dd7a62560\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f62656e63686d61726b2d62617463682d707573682e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/8e8c2acd0917b60d6a8897c076e7608dd7a62560\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f62656e63686d61726b2d62617463682d707573682e706e67\" alt=\"Batch PUSH benchmark\" data-canonical-src=\"https:\/\/valeriansaliou.github.io\/sonic\/images\/benchmark-batch-push.png\" \/><\/a><\/p><p><strong>Batch QUERY results as seen from our terminal (on index of: 1,000,000 objects):<\/strong><\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/5798a819dd65abf94a5e1d72e890d261207fad80\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f62656e63686d61726b2d62617463682d71756572792e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/5798a819dd65abf94a5e1d72e890d261207fad80\/68747470733a2f2f76616c657269616e73616c696f752e6769746875622e696f2f736f6e69632f696d616765732f62656e63686d61726b2d62617463682d71756572792e706e67\" alt=\"Batch QUERY benchmark\" data-canonical-src=\"https:\/\/valeriansaliou.github.io\/sonic\/images\/benchmark-batch-query.png\" \/><\/a><\/p><h2>Limitations<\/h2><ul><li><strong>Indexed data limits<\/strong>: Sonic is designed for large search indexes split over thousands of search buckets per collection. An IID (ie. Internal-ID) is stored in the index as a 32 bits number, which theoretically allow up to ~4.2 billion objects to be indexed (ie. OID) per bucket. We've observed storage savings of 30% to 40%, which justifies the trade-off on large databases (versus Sonic using 64 bits IIDs). Also, Sonic only keeps the N most recently pushed results for a given word, in a sliding window way (the sliding window width can be configured).<\/li>\n<li><strong>Search query limits<\/strong>: Sonic Natural Language Processing system (NLP) does not work at the sentence-level, for storage compactness reasons (we keep the FST graph shallow as to reduce time and space complexity). It works at the word-level, and is thus able to search per-word and can predict a word based on user input, though it is unable to predict the next word in a sentence.<\/li>\n<li><strong>Real-time limits<\/strong>: the FST needs to be re-built every time a word is pushed or popped from the bucket graph. As this is quite heavy, Sonic batches rebuild cycles. If you have just pushed a new word to the index and you are not seeing it in the <code>SUGGEST<\/code> command yet, wait for the next rebuild cycle to kick-in, or force it with <code>TRIGGER consolidate<\/code> in a <code>control<\/code> channel.<\/li>\n<li><strong>Interoperability limits<\/strong>: Sonic Channel protocol is the only way to read and write search entries to the Sonic search index. Sonic does not expose any HTTP API. Sonic Channel has been built with performance and minimal network footprint in mind. If you need to access Sonic Channel from an unsupported programming language, you can either <a href=\"https:\/\/github.com\/valeriansaliou\/sonic\/issues\/new\">open an issue<\/a> or look at the reference <a href=\"https:\/\/github.com\/valeriansaliou\/node-sonic-channel\">node-sonic-channel<\/a> implementation and build it in your target programming language.<\/li>\n<li><strong>Hardware limits<\/strong>: Sonic performs the search on the file-system directly; ie. it does not fit the index in RAM. A search query results in a lot of random accesses on the disk, which means that it will be quite slow on old-school HDDs and super-fast on newer SSDs. Do store the Sonic database on SSD-backed file systems only.<\/li>\n<\/ul><h2>\ud83d\udd25 Report A Vulnerability<\/h2><p>If you find a vulnerability in Sonic, you are more than welcome to report it directly to <a href=\"https:\/\/github.com\/valeriansaliou\">@valeriansaliou<\/a> by sending an encrypted email to <a href=\"mailto:valerian@valeriansaliou.name\">valerian@valeriansaliou.name<\/a>. Do not report vulnerabilities in public GitHub issues, as they may be exploited by malicious people to target production servers running an unpatched Sonic instance.<\/p><p><strong>\u26a0\ufe0f You must encrypt your email using <a href=\"https:\/\/github.com\/valeriansaliou\">@valeriansaliou<\/a> GPG public key: <a href=\"https:\/\/valeriansaliou.name\/files\/keys\/valeriansaliou.gpg.pub.asc\" rel=\"nofollow\">\ud83d\udd11valeriansaliou.gpg.pub.asc<\/a>.<\/strong><\/p><p><strong>\ud83c\udf81 Based on the severity of the vulnerability, I may offer a $100 (US) bounty to whomever reported it.<\/strong><\/p>","created_at":"2019-03-24T02:00:51+0000","updated_at":"2019-03-24T13:04:09+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/1451907?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":885,"title":"Three protocols and a future of the decentralized internet","url":"https:\/\/blog.datproject.org\/2019\/03\/22\/three-protocols-and-a-future-of-the-decentralized-internet\/","content":"<p>Hi, I'm Darius Kazemi and I'm Code for Science &amp; Society's <a href=\"https:\/\/blog.mozilla.org\/blog\/2018\/08\/21\/mozilla-announces-25-new-fellows-in-openness-science-and-tech-policy\/\">Mozilla Fellow<\/a>. I'm halfway through my fellowship and we thought now would be a good time for me to share some of the things that I've been researching, building, and planning.<\/p><p>I knew coming into this fellowship that I wanted my work to be about the decentralized web as a whole, and not just Dat. CS&amp;S was totally receptive to that and so I've been spending time learning about Dat, but also about <a href=\"https:\/\/www.scuttlebutt.nz\/\">Secure ScuttleButt<\/a> and <a href=\"https:\/\/activitypub.rocks\/\">ActivityPub<\/a>.<\/p><p>If what we broadly want is for power to be redistributed from the hands of large centralized corporate interests and back into the hands of individual people and small groups, then we already have the technical protocols to do this. No one protocol will enable this shift in power; each protocol has its strengths and weaknesses. What is needed is software that utilizes these protocols for their strengths, which is to say what is needed is developers and designers who are willing to act as bridges between the communities of practice surrounding the protocols.<\/p><h2 id=\"complementaryprotocolsworkingtogether\">Complementary protocols, working together<\/h2><p>I recently attended ScuttleCamp outside of Wellington, New Zealand. A bunch of Secure ScuttleButt community members met up in order to talk about what works and what doesn't with SSB, and where the community would like it to go in the future. It was a great event that I would estimate consisted of equal parts core SSB project team, active users of SSB, and interested outsiders (like me). It was in \"unconference\" format, so the attendees came up with the schedule on the fly. I was pleasantly surprised to see two Dat sessions pop up, including one on \"Bridging Dat and SSB\". There seemed to be a lot of understanding from the SSB team and even users that SSB has its strengths and weaknesses, and it would make sense to use multiple decentralized\/distributed protocols to provide great, useful, people-powered software.<\/p><p>In addition to working with and meeting people from the Dat and SSB community, I've been <a href=\"https:\/\/github.com\/dariusk\/express-activitypub\">developing<\/a> <a href=\"https:\/\/github.com\/dariusk\/rss-to-activitypub\/\">software<\/a> that uses ActivityPub. ActivityPub is yet another protocol that tries to reduce our reliance on tech giants to do the kind of things we want to do with the internet. I <a href=\"https:\/\/hacks.mozilla.org\/2018\/11\/decentralizing-social-interactions-with-activitypub\/\">wrote an introduction to ActivityPub for the Mozilla Hacks blog<\/a> that might be a helpful read.<\/p><p>In the course of learning about these three protocols and being inspired by conversations with the people that use them, I've come to see them as fulfilling different, complementary roles.<\/p><p>I'll try and sketch them out here <em>extremely<\/em> briefly and broadly, and conclude with my thoughts on how we can design better software by bridging these technologies. (I know there are more technologies out there than these three, but these are the three that I know best.)<\/p><h3 id=\"dat\">Dat<\/h3><p>If you're reading this blog, chances are you already know about Dat. Dat was invented to allow researchers to share very large data sets with each other. It's a protocol that powers software like <a href=\"https:\/\/beakerbrowser.com\/\">Beaker Browser<\/a>, a web browser that lets you create and publish your own websites without needing a web hosting service.<\/p><p>My basic bullet points:<\/p><ul><li>Peer to peer protocol, can share data via your laptop or phone without connecting to some company's server, no out-of-the-box web browser support<\/li>\n<li>All your content is its own hidden island until its URL is sent to someone else; that person can share the URL or even post it somewhere public and now your island is on the map<\/li>\n<li>Good for sharing large files, your friends can consent to re-sharing them if they want<\/li>\n<li>Sends blobs of unstructured data around (you can add your own structure)<\/li>\n<li>Discovering new stuff requires manually sharing links<\/li>\n<\/ul><h3 id=\"securescuttlebutt\">Secure ScuttleButt<\/h3><p>Secure ScuttleButt (SSB) was invented, at least in part, to remove distinctions between frontend and backend software, enabling the \"backend\" to be carried around by a user, who then owns their own data. SSB is a protocol that, like Dat, powers <a href=\"https:\/\/www.scuttlebutt.nz\/applications\">a bunch of different decentralized applications<\/a>. Chief among them is <a href=\"https:\/\/www.scuttlebutt.nz\/getting-started\">Patchwork<\/a>, which is a kind of decentralized forum software that lets people have and follow conversations about different topics. SSB is also used for things like <a href=\"https:\/\/github.com\/Happy0\/ssb-chess-mithril\/blob\/master\/README.md\">decentralized chess<\/a> and <a href=\"https:\/\/github.com\/noffle\/git-ssb-intro\">decentralized GitHub-style code collaboration<\/a>.<\/p><p>My basic bullet points:<\/p><ul><li>Peer to peer protocol, can share data via your laptop or phone without connecting to some company's server, no out-of-the-box web browser support<\/li>\n<li>Automatic publication of content to your friends<\/li>\n<li>Bad for sharing large files; among other things you don't want to force all your friends to store all your large files<\/li>\n<li>Sends blobs of unstructured data around (you can add your own structure)<\/li>\n<li>Discovering new stuff is somewhat automatic; once you are initially connected to a network then you can find friends-of-friends and keep connecting to more people<\/li>\n<\/ul><h3 id=\"activitypub\">ActivityPub<\/h3><p>ActivityPub was designed to solve the problem of \"walled gardens\" on the internet by providing a common language for websites and apps to speak to each other. Quoting <a href=\"https:\/\/hacks.mozilla.org\/2018\/11\/decentralizing-social-interactions-with-activitypub\/\">my previous writing on the topic<\/a>, ActivityPub is a protocol that \"enables a decentralized social web, where a network of servers interact with each other on behalf of individual users\/clients, very much like email operates at a macro level.\" It enables Twitter-style services like <a href=\"https:\/\/joinmastodon.org\">Mastodon<\/a> to talk to Instagram-style services like <a href=\"https:\/\/pixelfed.social\/\">PixelFed<\/a> and YouTube-style services like <a href=\"https:\/\/joinpeertube.org\/en\/\">PeerTube<\/a>. (Brief nerd note: I am referring to the server-to-server portion of the protocol here.)<\/p><p>My basic bullet points:<\/p><ul><li>Web protocol, so it works in all desktop and mobile web browsers without special software<\/li>\n<li>Not peer to peer, relies on special network nodes (servers) and an expert class of systems administrators to operate<\/li>\n<li>Since content is \"federated\" (replicated) between known servers, large files can fill up server hard drives pretty fast<\/li>\n<li>Designed around sending highly structured data<\/li>\n<li>Discovering new stuff can be pretty easy, it's not too different from existing centralized services that way<\/li>\n<\/ul><h2 id=\"whatthismeans\">What this means<\/h2><p>I've started thinking of these three protocols as each solving certain problems in the universe of software design, and when we understand where the protocols fit into the design space, we can start to think of solutions to seemingly intractable problems.<\/p><ul><li>\n<p>Dat is great for sharing big files (or bundles of small files).<\/p>\n<\/li>\n<li>\n<p>Secure ScuttleButt is great for spreading messages organically between connected networks of people or computers.<\/p>\n<\/li>\n<li>\n<p>ActivityPub is great for coordinating messages between networks of always-on servers that are actively maintained by specialist systems administrators.<\/p>\n<\/li>\n<\/ul><p>Looking at the software this way gives us a framework to design complex decentralized services.<\/p><p>For example: how can we have a Google Drive style system where a bunch of people can just connect to each other's computers and share files that way, removing the giant tech company in the middle? Maybe we serve the files over Dat, we have a small network of always-on Dat pinning servers that coordinate with each other over ActivityPub to share file serving responsibilities, and perhaps discovery of what files my friends have been posting happens over Secure ScuttleButt. Maybe every pinning server is also a SSB pub, and <em>these<\/em> coordinate over ActivityPub so there is robust delivery of social data to every user too. That's a very rough example and of course there are lots of problems that would need to be solved but it gives us a starting point for thinking about how to make this thing happen.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>All of this points to two major recommendations that I can make.<\/p><p>First, that software developers should not be afraid to mix, match, and layer protocols. There is no rule that says you can't do this, yet I've noticed people picking one protocol and sticking to it out of some kind of loyalty. When really, as a developer, your loyalty should be to your values of an open, decentralized internet, whatever those are.  If we view these protocols as tools in a toolbox rather than monolithic solutions or ecosystems, we can build decentralized software that rivals and surpasses in utility anything the big centralized tech giants have to offer.<\/p><p>Second, communities that build and support these protocols need to keep their specific strengths and weaknesses in mind. Instead of trying to make a \"kitchen sink\" protocol that solves a wide variety of problems, we should acknowledge the limitations of our protocols (and our own resources) and instead support and foster interoperability and collaboration between our communities.<\/p><p>The internet has always been built on a bunch of different technologies filling in the gaps for each other. For example, while web browsers mainly use the TCP protocol for loading web pages, they switch to using UDP when it's convenient for things like video chat. Decentralized web protocols should be no exception to this pattern. The only way we're going to make the internet better is by working together to help each other.<\/p><p><em>Thanks to Joe Hand and <a href=\"http:\/\/eight45.net\/\">noffle<\/a> for feedback, and to <a href=\"https:\/\/coolguy.website\/\">Zach Mandeville<\/a> and <a href=\"https:\/\/www.cameralibre.cc\/\">Sam Muirhead<\/a> for getting me thinking about this in the first place.<\/em><\/p>","created_at":"2019-03-23T00:47:33+0000","updated_at":"2019-03-23T00:47:33+0000","published_at":"2019-03-22T17:39:17+0000","published_by":["Darius Kazemi"],"annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"blog.datproject.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":884,"title":"How I'm able to take notes in mathematics lectures using LaTeX and Vim","url":"https:\/\/castel.dev\/post\/lecture-notes-1\/","content":"<p><em>A while back I an\u00adswered a ques\u00adtion on Quora: <a href=\"https:\/\/www.quora.com\/Can-people-actually-keep-up-with-note-taking-in-Mathematics-lectures-with-LaTeX\/answer\/Gilles-Castel-1\">Can peo\u00adple ac\u00adtu\u00adal\u00adly keep up with note-\u200btaking in Math\u00ade\u00admat\u00adics lec\u00adtures with LaTeX<\/a>.\nThere, I ex\u00adplained my work\u00adflow of tak\u00ading lec\u00adture notes in LaTeX using Vim and how I draw fig\u00adures in Inkscape.\nHow\u00adev\u00ader, a lot has changed since then and I\u2019d like to write a few blog posts ex\u00adplain\u00ading my work\u00adflow.<\/em><\/p><p>I start\u00aded using LaTeX to write lec\u00adture notes in the sec\u00adond se\u00admes\u00adter of my bach\u00ade\u00adlor in math\u00ade\u00admat\u00adics, and I\u2019ve been using it ever since,\nwhich makes for a total of more than 1700 pages of notes.\nTo give you an idea of what those notes look like, here are some ex\u00adam\u00adples:<\/p><p>\n  <a class=\"gatsby-resp-image-link\" href=\"https:\/\/castel.dev\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/9ccd8\/ca1.png\" target=\"_blank\" rel=\"noopener\">\n  \n  \n    \n      <img class=\"gatsby-resp-image-image\" alt=\"ca1\" title=\"\" src=\"https:\/\/castel.dev\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/14883\/ca1.png\" srcset=\"\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/3f2fd\/ca1.png 170w, &#10;\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/dd39f\/ca1.png 340w, &#10;\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/14883\/ca1.png 680w, &#10;\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/5ce4e\/ca1.png 1020w, &#10;\/static\/ba66c52b01c8fe6b31c97af52b5eb7f7\/9ccd8\/ca1.png 1173w\" \/><\/a>\n    \n  <a class=\"gatsby-resp-image-link\" href=\"https:\/\/castel.dev\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/0c252\/ca2.png\" target=\"_blank\" rel=\"noopener\">\n  \n  \n    \n      <img class=\"gatsby-resp-image-image\" alt=\"ca2\" title=\"\" src=\"https:\/\/castel.dev\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/14883\/ca2.png\" srcset=\"\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/3f2fd\/ca2.png 170w, &#10;\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/dd39f\/ca2.png 340w, &#10;\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/14883\/ca2.png 680w, &#10;\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/5ce4e\/ca2.png 1020w, &#10;\/static\/1a393ac73dbb96f6fbd98b258e8a325a\/0c252\/ca2.png 1163w\" \/><\/a>\n    \n  <a class=\"gatsby-resp-image-link\" href=\"https:\/\/castel.dev\/static\/cb323a2ec665d98f93f78e13d3972e0e\/d5ff5\/ca3.png\" target=\"_blank\" rel=\"noopener\">\n  \n  \n    \n      <img class=\"gatsby-resp-image-image\" alt=\"ca3\" title=\"\" src=\"https:\/\/castel.dev\/static\/cb323a2ec665d98f93f78e13d3972e0e\/14883\/ca3.png\" srcset=\"\/static\/cb323a2ec665d98f93f78e13d3972e0e\/3f2fd\/ca3.png 170w, &#10;\/static\/cb323a2ec665d98f93f78e13d3972e0e\/dd39f\/ca3.png 340w, &#10;\/static\/cb323a2ec665d98f93f78e13d3972e0e\/14883\/ca3.png 680w, &#10;\/static\/cb323a2ec665d98f93f78e13d3972e0e\/5ce4e\/ca3.png 1020w, &#10;\/static\/cb323a2ec665d98f93f78e13d3972e0e\/d5ff5\/ca3.png 1155w\" \/><\/a>\n    <\/p><p>These lec\u00adture notes \u2014 in\u00adclud\u00ading fig\u00adures \u2014 are made while at\u00adtend\u00ading the lec\u00adture and have not been edit\u00aded af\u00adter\u00adwards.\nTo make note tak\u00ading using LaTeX vi\u00adable, I had four goals in mind:<\/p><ul><li>Writ\u00ading text and math\u00ade\u00admat\u00adi\u00adcal for\u00admu\u00adlas in LaTeX should be as fast as the lec\u00adtur\u00ader writ\u00ading on a black\u00adboard: no delay is ac\u00adcept\u00adable.<\/li>\n<li>Draw\u00ading fig\u00adures should be al\u00admost as fast as the lec\u00adtur\u00ader.<\/li>\n<li>Man\u00adag\u00ading notes, i.e. adding a note, com\u00adpil\u00ading all my notes, com\u00adpil\u00ading the last two lec\u00adtures, search\u00ading in notes, etc. should be easy and quick.<\/li>\n<li>An\u00adno\u00adtat\u00ading pdf doc\u00adu\u00adments using LaTeX should be pos\u00adsi\u00adble for when I want to write notes along\u00adside a pdf doc\u00adu\u00adment.<\/li>\n<\/ul><p>This blog post will focus on the first item: writ\u00ading LaTeX.<\/p><h2>Vim and LaTeX<\/h2><p>For writ\u00ading text and math\u00ade\u00admat\u00adi\u00adcal for\u00admu\u00adlas in LaTeX, I use Vim.\nVim is a pow\u00ader\u00adful gen\u00ader\u00adal pur\u00adpose text ed\u00adi\u00adtor that\u2019s very ex\u00adten\u00adsi\u00adble.\nI use it for writ\u00ading code, LaTeX, mark\u00addown, \u2026 ba\u00adsi\u00adcal\u00adly every\u00adthing that\u2019s text-\u200bbased.\nIt has a fair\u00adly steep learn\u00ading curve, but once you\u2019ve got the ba\u00adsics down, it\u2019s hard to get back to an ed\u00adi\u00adtor with\u00adout Vim key\u00adbind\u00adings.\nHere\u2019s what my screen looks like when I\u2019m edit\u00ading LaTeX:<\/p><p>\n  <a class=\"gatsby-resp-image-link\" href=\"https:\/\/castel.dev\/static\/4282d15d03150c89182ab56634d2b4b0\/daa9e\/vimzathura.png\" target=\"_blank\" rel=\"noopener\">\n  \n  \n    \n      <img class=\"gatsby-resp-image-image\" alt=\"Vim and Zathura\" title=\"\" src=\"https:\/\/castel.dev\/static\/4282d15d03150c89182ab56634d2b4b0\/14883\/vimzathura.png\" srcset=\"\/static\/4282d15d03150c89182ab56634d2b4b0\/3f2fd\/vimzathura.png 170w, &#10;\/static\/4282d15d03150c89182ab56634d2b4b0\/dd39f\/vimzathura.png 340w, &#10;\/static\/4282d15d03150c89182ab56634d2b4b0\/14883\/vimzathura.png 680w, &#10;\/static\/4282d15d03150c89182ab56634d2b4b0\/5ce4e\/vimzathura.png 1020w, &#10;\/static\/4282d15d03150c89182ab56634d2b4b0\/55fdf\/vimzathura.png 1360w, &#10;\/static\/4282d15d03150c89182ab56634d2b4b0\/daa9e\/vimzathura.png 1919w\" \/><\/a>\n    <\/p><p>On the left you see Vim and on the right my pdf view\u00ader, <a href=\"https:\/\/pwmt.org\/projects\/zathura\/\">Za\u00adthu\u00adra<\/a>, which also has Vim-\u200blike key\u00adbind\u00adings.\nI\u2019m using Ubun\u00adtu with <a href=\"https:\/\/github.com\/baskerville\/bspwm\">bspwm<\/a> as my win\u00addow man\u00adag\u00ader.\nThe LaTeX plu\u00adg\u00adin I\u2019m using in Vim is <a href=\"https:\/\/github.com\/lervag\/vimtex\">vim\u00adtex<\/a>.\nIt pro\u00advides syn\u00adtax high\u00adlight\u00ading, table of con\u00adtents view, sync\u00adtex, etc.\nUsing <a href=\"https:\/\/github.com\/junegunn\/vim-plug\">vim-\u200bplug<\/a>, I con\u00adfig\u00adured it as fol\u00adlows:<\/p><div class=\"gatsby-highlight\" data-language=\"vim\"><pre class=\"language-vim\">Plug 'lervag\/vimtex'\nlet g:tex_flavor='latex'\nlet g:vimtex_view_method='zathura'\nlet g:vimtex_quickfix_mode=0\nset conceallevel=1\nlet g:tex_conceal='abdmg'<\/pre><\/div><p>The last two lines con\u00adfig\u00adure the con\u00adceal\u00adment.\nThis is a fea\u00adture where LaTeX code is re\u00adplaced or made in\u00advis\u00adi\u00adble when your cur\u00adsor is not on that line.\nBy mak\u00ading <code class=\"language-text\">\\[<\/code>, <code class=\"language-text\">\\]<\/code>, <code class=\"language-text\">$<\/code> in\u00advis\u00adi\u00adble, they\u2019re less ob\u00adtru\u00adsive which gives you a bet\u00adter overview of the doc\u00adu\u00adment. This fea\u00adture also re\u00adplaces <code class=\"language-text\">\\bigcap<\/code> by by <code class=\"language-text\">\u2229<\/code>, <code class=\"language-text\">\\in<\/code> by <code class=\"language-text\">\u2208<\/code> etc.\nThe fol\u00adlow\u00ading an\u00adi\u00adma\u00adtion should make that clear.<\/p><p><img src=\"https:\/\/castel.dev\/conceal-d362e0b837be308b0f8897482a331ec6.gif\" alt=\"conceal\" \/><\/p><p>With this set up, I come to the crux of this blog post: writ\u00ading LaTeX as fast as the lec\u00adtur\u00ader can write on the black\u00adboard.\nThis is where snip\u00adpets come into play.<\/p><h2>Snip\u00adpets<\/h2><h3>What\u2019s a snip\u00adpet?<\/h3><p>A snip\u00adpet is a short reusable piece of text that can be trig\u00adgered by some other text.\nFor ex\u00adam\u00adple, when I type <code class=\"language-text\">sign<\/code> and press <kbd>Tab<\/kbd>, the word <code class=\"language-text\">sign<\/code> will be ex\u00adpand\u00aded to a sig\u00adna\u00adture:<\/p><p><img src=\"https:\/\/castel.dev\/sign-5ab28b1ad6114b68f4ce459d4f89eeae.gif\" alt=\"sign\" \/><\/p><p>Snip\u00adpets can also be dy\u00adnam\u00adic: when I type <code class=\"language-text\">today<\/code> and press <kbd>Tab<\/kbd>, the word <code class=\"language-text\">today<\/code> will be re\u00adplaced by the cur\u00adrent date, and <code class=\"language-text\">box<\/code> <kbd>Tab<\/kbd> be\u00adcomes a box that au\u00adto\u00admat\u00adi\u00adcal\u00adly grows in size.<\/p><p><img src=\"https:\/\/castel.dev\/today-1b6301a6561d104b4bc79654cb75470d.gif\" alt=\"today\" \/><\/p><p><img src=\"https:\/\/castel.dev\/box-8b7caefc472eae0941f41365d16e8d2d.gif\" alt=\"box\" \/><\/p><p>You can even use one snip\u00adpet in\u00adside an\u00adoth\u00ader:<\/p><p><img src=\"https:\/\/castel.dev\/todaybox-cdd1fded48947d163c2e5e6b974cfec0.gif\" alt=\"todaybox\" \/><\/p><h3>Using Ul\u00adtiSnips to cre\u00adate snip\u00adpets<\/h3><p>I use the plu\u00adg\u00adin <a href=\"https:\/\/github.com\/SirVer\/ultisnips\">Ul\u00adtiSnips<\/a> to man\u00adage my snip\u00adpets. My con\u00adfig\u00adu\u00adra\u00adtion is<\/p><div class=\"gatsby-highlight\" data-language=\"vim\"><pre class=\"language-vim\">Plug 'sirver\/ultisnips'\nlet g:UltiSnipsExpandTrigger = '&lt;tab&gt;'\nlet g:UltiSnipsJumpForwardTrigger = '&lt;tab&gt;'\nlet g:UltiSnipsJumpBackwardTrigger = '&lt;s-tab&gt;'<\/pre><\/div><p>The code for the <code class=\"language-text\">sign<\/code> snip\u00adpet is the fol\u00adlow\u00ading:<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet sign \"Signature\"\nYours sincerely,\nGilles Castel\nendsnippet<\/pre><\/div><p>For dy\u00adnam\u00adic snip\u00adpets, you can put code be\u00adtween back\u00adticks <code class=\"language-text\">``<\/code> which will be run when the snip\u00adpet is ex\u00adpand\u00aded.\nHere, I\u2019ve used bash to for\u00admat the cur\u00adrent date: <code class=\"language-text\">date + %F<\/code>.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet today \"Date\"\n`date +%F`\nendsnippet<\/pre><\/div><p>You can also use Python in\u00adside a <code class=\"language-text\">`!p ... `<\/code> block.\nHave a look at the code for the <code class=\"language-text\">box<\/code> snip\u00adpet:<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet box \"Box\"\n`!p snip.rv = '\u250c' + '\u2500' * (len(t[1]) + 2) + '\u2510'`\n\u2502 $1 \u2502\n`!p snip.rv = '\u2514' + '\u2500' * (len(t[1]) + 2) + '\u2518'`\n$0\nendsnippet<\/pre><\/div><p>These Python code blocks will be re\u00adplaced by the value of the vari\u00adable <code class=\"language-text\">snip.rv<\/code>.\nIn\u00adside these blocks, you have ac\u00adcess to the cur\u00adrent state of the snip\u00adpet, e.g. <code class=\"language-text\">t[1]<\/code> con\u00adtains the first tab stop, <code class=\"language-text\">fn<\/code> the cur\u00adrent file\u00adname, \u2026<\/p><h2>LaTeX snip\u00adpets<\/h2><p>Using snip\u00adpets, writ\u00ading LaTeX is a lot faster than writ\u00ading it by hand.\nEs\u00adpe\u00adcial\u00adly some of the more com\u00adplex snip\u00adpets can save you a lot of time and frus\u00adtra\u00adtion.  Let\u2019s begin with some sim\u00adple snip\u00adpets.<\/p><h3>En\u00advi\u00adron\u00adments<\/h3><p>To in\u00adsert an en\u00advi\u00adron\u00adment, all I have to do is type <code class=\"language-text\">beg<\/code> at the be\u00adgin\u00adning of a line.\nThen I type the name of the en\u00advi\u00adron\u00adment, which is mir\u00adrored in the <code class=\"language-text\">\\end{}<\/code> com\u00admand. Press\u00ading <kbd>Tab<\/kbd> places the cur\u00adsor in\u00adside the newly cre\u00adat\u00aded en\u00advi\u00adron\u00adment.<\/p><p><img src=\"https:\/\/castel.dev\/beginend-d5ed641a9cb568d13dd56d98fec6376b.gif\" alt=\"beginend\" \/><\/p><p>The code for this snip\u00adpet is the fol\u00adlow\u00ading.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet beg \"begin{} \/ end{}\" bA\n\\begin{$1}\n\t$0\n\\end{$1}\nendsnippet<\/pre><\/div><p>The <code class=\"language-text\">b<\/code> means that this snip\u00adpet will only be ex\u00adpand\u00aded at the be\u00adgin\u00adning of a line and <code class=\"language-text\">A<\/code> stands for auto ex\u00adpand, which means I do not have to press <kbd>Tab<\/kbd> to ex\u00adpand the snip\u00adpet. Tab stops \u2014 i.e. places you can jump to by press\u00ading <kbd>Tab<\/kbd> and <kbd>Shift<\/kbd>+<kbd>Tab<\/kbd> \u2014 are rep\u00adre\u00adsent\u00aded by <code class=\"language-text\">$1<\/code>, <code class=\"language-text\">$2<\/code>, \u2026 and the last one with <code class=\"language-text\">$0<\/code>.<\/p><h3>In\u00adline and dis\u00adplay math<\/h3><p>Two of my most fre\u00adquent\u00adly used snip\u00adpets are <code class=\"language-text\">mk<\/code> and <code class=\"language-text\">dm<\/code>.\nThey\u2019re the snip\u00adpets re\u00adspon\u00adsi\u00adble for start\u00ading math mode.\nThe first one is a snip\u00adpet for in\u00adline math, the sec\u00adond one for dis\u00adplayed math.<\/p><p><img src=\"https:\/\/castel.dev\/mkdm-12a16b6a90c4d85a887c7cde0f18bb8c.gif\" alt=\"mkdm\" \/><\/p><p>The snip\u00adpet for in\u00adline math is \u2018smart\u2019: it knows when to in\u00adsert a space after the dol\u00adlar sign.\nWhen I start typ\u00ading a word di\u00adrect\u00adly be\u00adhind the clos\u00ading <code class=\"language-text\">$<\/code>, it adds a space.\nHow\u00adev\u00ader, when I type a non-\u200bword char\u00adac\u00adter, it does not add a space, which would be pre\u00adferred for ex\u00adam\u00adple in the case of \u2018$p$-\u200bvalue\u2019.<\/p><p><img src=\"https:\/\/castel.dev\/mk_space-1172a1f556fd36b42f7c3642b3952e2c.gif\" alt=\"mk_space\" \/><\/p><p>The code for this snip\u00adpet is the fol\u00adlow\u00ading.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet mk \"Math\" wA\n$${1}$`!p\nif t[2] and t[2][0] not in [',', '.', '?', '-', ' ']:\n    snip.rv = ' '\nelse:\n    snip.rv = ''\n`$2\nendsnippet<\/pre><\/div><p>The <code class=\"language-text\">w<\/code> at the end of the first line means that this snip\u00adpet will ex\u00adpand at word bound\u00adaries, so e.g. <code class=\"language-text\">hellomk<\/code> won\u2019t ex\u00adpand, but <code class=\"language-text\">hello mk<\/code> will.<\/p><p>The snip\u00adpet for dis\u00adplayed math is more sim\u00adple, but it also is quite handy; it makes me never for\u00adget end\u00ading equa\u00adtions with a pe\u00adri\u00adod.<\/p><p><img src=\"https:\/\/castel.dev\/dm-a682e97365ecfe19fe0c5bc08bacfec6.gif\" alt=\"dm\" \/><\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet dm \"Math\" wA\n\\[\n$1\n.\\] $0\nendsnippet<\/pre><\/div><h3>Sub- and su\u00adper\u00adscripts<\/h3><p>An\u00adoth\u00ader use\u00adful snip\u00adpet is one for sub\u00adscripts.\nIt changes changes <code class=\"language-text\">a1<\/code> to <code class=\"language-text\">a_1<\/code> and <code class=\"language-text\">a_12<\/code> to <code class=\"language-text\">a_{12}<\/code>.<\/p><p><img src=\"https:\/\/castel.dev\/subscripts-f26633cb822c6b52d346dfbba6e1a1bd.gif\" alt=\"subscripts\" \/><\/p><p>The code for this snip\u00adpet uses a reg\u00adu\u00adlar ex\u00adpres\u00adsion for its trig\u00adger.\nIt ex\u00adpands the snip\u00adpet when you type a char\u00adac\u00adter fol\u00adlowed by a digit, which en\u00adcod\u00aded by <code class=\"language-text\">[A-Za-z]\\d<\/code>, or a char\u00adac\u00adter fol\u00adlowed by <code class=\"language-text\">_<\/code> and two dig\u00adits: <code class=\"language-text\">[A-Za-z]_\\d\\d<\/code>.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet '([A-Za-z])(\\d)' \"auto subscript\" wrA\n`!p snip.rv = match.group(1)`_`!p snip.rv = match.group(2)`\nendsnippet\nsnippet '([A-Za-z])_(\\d\\d)' \"auto subscript2\" wrA\n`!p snip.rv = match.group(1)`_{`!p snip.rv = match.group(2)`}\nendsnippet<\/pre><\/div><p>When you wrap parts of a reg\u00adu\u00adlar ex\u00adpres\u00adsion in a group using paren\u00adthe\u00adsis, e.g. <code class=\"language-text\">(\\d\\d)<\/code>, you can use them in the ex\u00adpan\u00adsion of the snip\u00adpet via <code class=\"language-text\">match.group(i)<\/code> in Python.<\/p><p>As for su\u00adper\u00adscripts, I use <code class=\"language-text\">td<\/code>, which be\u00adcomes <code class=\"language-text\">^{}<\/code>.  How\u00adev\u00ader, for squared, cubed, com\u00adple\u00adment and a hand\u00adful of other com\u00admon ones, I use ded\u00adi\u00adcat\u00aded snip\u00adpets such as <code class=\"language-text\">sr<\/code>, <code class=\"language-text\">cb<\/code> and <code class=\"language-text\">comp<\/code>.<\/p><p><img src=\"https:\/\/castel.dev\/superscripts-da855cd9e67c48f508fab5ddc8c877fe.gif\" alt=\"superscripts\" \/><\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet sr \"^2\" iA\n^2\nendsnippet\nsnippet cb \"^3\" iA\n^3\nendsnippet\nsnippet compl \"complement\" iA\n^{c}\nendsnippet\nsnippet td \"superscript\" iA\n^{$1}$0\nendsnippet<\/pre><\/div><h3>Frac\u00adtions<\/h3><p>One of my most con\u00adve\u00adnient snip\u00adpets is one for frac\u00adtions. This makes the fol\u00adlow\u00ading ex\u00adpan\u00adsions:<\/p><table><thead><tr><th align=\"right\">\n<\/th><th align=\"left\">\n<\/th><\/tr><\/thead><tbody><tr><td align=\"right\"><code class=\"language-text\">\/\/<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\frac{}{}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">3\/<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\frac{3}{}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">4\\pi^2\/<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\frac{4\\pi^2}{}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">(1 + 2 + 3)\/<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\frac{1 + 2 + 3}{}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">(1+(2+3)\/)<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">(1 + \\frac{2+3}{})<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">(1 + (2+3))\/<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\frac{1 + (2+3)}{}<\/code><\/td>\n<\/tr><\/tbody><\/table><p><img src=\"https:\/\/castel.dev\/frac-3f83f1bdc3078aa16382e80a276f199f.gif\" alt=\"frac\" \/><\/p><p>The code for the first one is easy:<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet \/\/ \"Fraction\" iA\n\\\\frac{$1}{$2}$0\nendsnippet<\/pre><\/div><p>The sec\u00adond and third ex\u00adam\u00adples are made pos\u00adsi\u00adble using reg\u00adu\u00adlar ex\u00adpres\u00adsions to match for ex\u00adpres\u00adsions like <code class=\"language-text\">3\/<\/code>, <code class=\"language-text\">4ac\/<\/code>, <code class=\"language-text\">6\\pi^2\/<\/code>, <code class=\"language-text\">a_2\/<\/code>, etc.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet '((\\d+)|(\\d*)(\\\\)?([A-Za-z]+)((\\^|_)(\\{\\d+\\}|\\d))*)\/' \"Fraction\" wrA\n\\\\frac{`!p snip.rv = match.group(1)`}{$1}$0\nendsnippet<\/pre><\/div><p>As you can see, reg\u00adu\u00adlar ex\u00adpres\u00adsions can be\u00adcome quite over\u00adwhelm\u00ading, but here\u2019s a di\u00ada\u00adgram that should ex\u00adplain it:<\/p><p><img src=\"https:\/\/castel.dev\/regex-908df5ec254f9cd5b08b99b8a54318d0.svg\" alt=\"Diagram of the regular expression\" \/><\/p><p>In the fourth and fifth cases, it tries to find the match\u00ading paren\u00adthe\u00adsis. As this isn\u2019t pos\u00adsi\u00adble using the reg\u00adu\u00adlar ex\u00adpres\u00adsion en\u00adgine of Ul\u00adtiSnips, I re\u00adsort\u00aded to using Python:<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">priority 1000\nsnippet '^.*\\)\/' \"() Fraction\" wrA\n`!p\nstripped = match.string[:-1]\ndepth = 0\ni = len(stripped) - 1\nwhile True:\n\tif stripped[i] == ')': depth += 1\n\tif stripped[i] == '(': depth -= 1\n\tif depth == 0: break;\n\ti -= 1\nsnip.rv = stripped[0:i] + \"\\\\frac{\" + stripped[i+1:-1] + \"}\"\n`{$1}$0\nendsnippet<\/pre><\/div><p>The last snip\u00adpet con\u00adcern\u00ading frac\u00adtions I\u2019d like to share is one that uses your se\u00adlec\u00adtion to make a frac\u00adtion.  You can use it by first se\u00adlect\u00ading some text, then press\u00ading <kbd>Tab<\/kbd>, typ\u00ading <code class=\"language-text\">\/<\/code> and press\u00ading <kbd>Tab<\/kbd> again.<\/p><p><img src=\"https:\/\/castel.dev\/visualfrac-0da01b9566dc437fabf0c563ed14941f.gif\" alt=\"visualfrac\" \/><\/p><p>The code makes use of the <code class=\"language-text\">${VISUAL}<\/code> vari\u00adable that rep\u00adre\u00adsents your se\u00adlec\u00adtion.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet \/ \"Fraction\" iA\n\\\\frac{${VISUAL}}{$1}$0\nendsnippet<\/pre><\/div><h3>Sympy and Math\u00ade\u00admat\u00adi\u00adca<\/h3><p>An\u00adoth\u00ader cool \u2014 but less used \u2014 snip\u00adpet is one that uses <a href=\"https:\/\/www.sympy.org\/\">sympy<\/a> to eval\u00adu\u00adate math\u00ade\u00admat\u00adi\u00adcal ex\u00adpres\u00adsions. For ex\u00adam\u00adple: <code class=\"language-text\">sympy<\/code> <kbd>Tab<\/kbd> ex\u00adpands to <code class=\"language-text\">sympy | sympy<\/code>, and <code class=\"language-text\">sympy 1 + 1 sympy<\/code> <kbd>Tab<\/kbd> ex\u00adpands to <code class=\"language-text\">2<\/code>.<\/p><p><img src=\"https:\/\/castel.dev\/sympy-8c8f143fdcdc2f3e6de9853cf7ea3def.gif\" alt=\"sympy\" \/><\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet sympy \"sympy block \" w\nsympy $1 sympy$0\nendsnippet\npriority 10000\nsnippet 'sympy(.*)sympy' \"evaluate sympy\" wr\n`!p\nfrom sympy import *\nx, y, z, t = symbols('x y z t')\nk, m, n = symbols('k m n', integer=True)\nf, g, h = symbols('f g h', cls=Function)\ninit_printing()\nsnip.rv = eval('latex(' + match.group(1).replace('\\\\', '') \\\n    .replace('^', '**') \\\n    .replace('{', '(') \\\n    .replace('}', ')') + ')')\n`\nendsnippet<\/pre><\/div><p>For the Math\u00ade\u00admat\u00adi\u00adca users out there, you can do some\u00adthing sim\u00adi\u00adlar:<\/p><p><img src=\"https:\/\/castel.dev\/mathematica-82f7f187a685a52ed7665fded563f936.gif\" alt=\"mathematica\" \/><\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">priority 1000\nsnippet math \"mathematica block\" w\nmath $1 math$0\nendsnippet\npriority 10000\nsnippet 'math(.*)math' \"evaluate mathematica\" wr\n`!p\nimport subprocess\ncode = 'ToString[' + match.group(1) + ', TeXForm]'\nsnip.rv = subprocess.check_output(['wolframscript', '-code', code])\n`\nendsnippet<\/pre><\/div><h3>Post\u00adfix snip\u00adpets<\/h3><p>Some other snip\u00adpets I find worth shar\u00ading are post\u00adfix snip\u00adpets.\nEx\u00adam\u00adples of such snip\u00adpets are <code class=\"language-text\">phat<\/code> \u2192 <code class=\"language-text\">\\hat{p}<\/code> and <code class=\"language-text\">zbar<\/code> \u2192 <code class=\"language-text\">\\overline{z}<\/code>.\nA sim\u00adi\u00adlar snip\u00adpet is a post\u00adfix vec\u00adtor, for ex\u00adam\u00adple <code class=\"language-text\">v,.<\/code> \u2192 <code class=\"language-text\">\\vec{v}<\/code> and <code class=\"language-text\">v.,<\/code> \u2192 <code class=\"language-text\">\\vec{v}<\/code>.\nThe order of <code class=\"language-text\">,<\/code> and <code class=\"language-text\">.<\/code> doesn\u2019t mat\u00adter, so I can press them both at the same time.\nThese snip\u00adpets are a real time-\u200bsaver, be\u00adcause you can type in the same order the lec\u00adtur\u00ader writes on the black\u00adboard.<\/p><p><img src=\"https:\/\/castel.dev\/barhatvec-7c2e62aae292bc4d65eae363c816e206.gif\" alt=\"barhatvec\" \/><\/p><p>Note that I can still use <code class=\"language-text\">bar<\/code> and <code class=\"language-text\">hat<\/code> pre\u00adfix too, as I\u2019ve added them with a lower pri\u00ador\u00adi\u00adty.\nThe code for those snip\u00adpets is:<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">priority 10\nsnippet \"bar\" \"bar\" riA\n\\overline{$1}$0\nendsnippet\npriority 100\nsnippet \"([a-zA-Z])bar\" \"bar\" riA\n\\overline{`!p snip.rv=match.group(1)`}\nendsnippet<\/pre><\/div><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">priority 10\nsnippet \"hat\" \"hat\" riA\n\\hat{$1}$0\nendsnippet\npriority 100\nsnippet \"([a-zA-Z])hat\" \"hat\" riA\n\\hat{`!p snip.rv=match.group(1)`}\nendsnippet<\/pre><\/div><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet \"(\\\\?\\w+)(,\\.|\\.,)\" \"Vector postfix\" riA\n\\vec{`!p snip.rv=match.group(1)`}\nendsnippet <\/pre><\/div><h3>Other snip\u00adpets<\/h3><p>I have about 100 other com\u00admon\u00adly used snip\u00adpets.\nThey are avail\u00adable <a href=\"https:\/\/castel.dev\/tex-e0c2e8b64036f77db00411d562750c12.snippets\">here<\/a>.\nMost of them are quite sim\u00adple.\nFor ex\u00adam\u00adple, <code class=\"language-text\">!&gt;<\/code> be\u00adcomes <code class=\"language-text\">\\mapsto<\/code>, <code class=\"language-text\">-&gt;<\/code> be\u00adcomes <code class=\"language-text\">\\to<\/code>, etc.<\/p><p><img src=\"https:\/\/castel.dev\/complex5-4ee36ef41c17f1ded1fffe13ba32b60d.gif\" alt=\"complex5\" \/><\/p><p><code class=\"language-text\">fun<\/code> be\u00adcomes <code class=\"language-text\">f: \\R \\to \\R :<\/code>, <code class=\"language-text\">!&gt;<\/code> \u2192 <code class=\"language-text\">\\mapsto<\/code>, <code class=\"language-text\">-&gt;<\/code> \u2192 <code class=\"language-text\">\\to<\/code>, <code class=\"language-text\">cc<\/code> \u2192 <code class=\"language-text\">\\subset<\/code>.<\/p><p><img src=\"https:\/\/castel.dev\/fun3-a45092c95d91059b779f74cedda840e1.gif\" alt=\"fun3\" \/><\/p><p><code class=\"language-text\">lim<\/code> be\u00adcomes <code class=\"language-text\">\\lim_{n \\to \\infty}<\/code>, <code class=\"language-text\">sum<\/code> \u2192 <code class=\"language-text\">\\sum_{n = 1}^{\\infty}<\/code>, <code class=\"language-text\">ooo<\/code> \u2192 <code class=\"language-text\">\\infty<\/code><\/p><p><img src=\"https:\/\/castel.dev\/sum4-2be9c99e6ab5787fc5196bbf4ce525a7.gif\" alt=\"sum4\" \/><\/p><p><img src=\"https:\/\/castel.dev\/bazel2-3be522f5e6e5ca7f871138732ba1b14d.gif\" alt=\"bazel2\" \/><\/p><h3>Course spe\u00adcif\u00adic snip\u00adpets<\/h3><p>Be\u00adside my com\u00admon\u00adly used snip\u00adpets, I also have course spe\u00adcif\u00adic snip\u00adpets.\nThese are loaded by adding the fol\u00adlow\u00ading to my <code class=\"language-text\">.vimrc<\/code>:<\/p><div class=\"gatsby-highlight\" data-language=\"vim\"><pre class=\"language-vim\">set rtp+=~\/current_course<\/pre><\/div><p>where <code class=\"language-text\">current_course<\/code> is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Symbolic_link\">sym\u00adlink<\/a> to my cur\u00adrent\u00adly ac\u00adti\u00advat\u00aded course (more about that in an\u00adoth\u00ader blog post).\nIn that fold\u00ader, I have a file <code class=\"language-text\">~\/current_course\/UltiSnips\/tex.snippets<\/code> in which I in\u00adclude course spe\u00adcif\u00adic snip\u00adpets. For ex\u00adam\u00adple, for quan\u00adtum me\u00adchan\u00adics, I have snip\u00adpets for bra\/ket no\u00adta\u00adtion.<\/p><table><thead><tr><th align=\"right\">\n<\/th><th align=\"left\">\n<\/th><\/tr><\/thead><tbody><tr><td align=\"right\"><code class=\"language-text\">&lt;a|<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\bra{a}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">&lt;q|<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\bra{\\psi}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">|a&gt;<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\ket{a}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">|q&gt;<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\ket{\\psi}<\/code><\/td>\n<\/tr><tr><td align=\"right\"><code class=\"language-text\">&lt;a|b&gt;<\/code><\/td>\n<td>\u2192<\/td>\n<td align=\"left\"><code class=\"language-text\">\\braket{a}{b}<\/code><\/td>\n<\/tr><\/tbody><\/table><p>As <code class=\"language-text\">\\psi<\/code> is used a lot in quan\u00adtum me\u00adchan\u00adics, I re\u00adplace all in\u00adstances of <code class=\"language-text\">q<\/code> in a braket with <code class=\"language-text\">\\psi<\/code> when ex\u00adpand\u00aded.<\/p><p><img src=\"https:\/\/castel.dev\/braket-268d331fdb83b6fa3563b6f2b0baac66.gif\" alt=\"braket\" \/><\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">snippet \"\\&lt;(.*?)\\|\" \"bra\" riA\n\\bra{`!p snip.rv = match.group(1).replace('q', f'\\psi').replace('f', f'\\phi')`}\nendsnippet\nsnippet \"\\|(.*?)\\&gt;\" \"ket\" riA\n\\ket{`!p snip.rv = match.group(1).replace('q', f'\\psi').replace('f', f'\\phi')`}\nendsnippet\nsnippet \"(.*)\\\\bra{(.*?)}([^\\|]*?)\\&gt;\" \"braket\" riA\n`!p snip.rv = match.group(1)`\\braket{`!p snip.rv = match.group(2)`}{`!p snip.rv = match.group(3).replace('q', f'\\psi').replace('f', f'\\phi')`}\nendsnippet<\/pre><\/div><h3>Con\u00adtext<\/h3><p>One thing to con\u00adsid\u00ader when writ\u00ading these snip\u00adpets is, \u2018will these snip\u00adpets col\u00adlide with usual text?\u2019\nFor ex\u00adam\u00adple, ac\u00adcord\u00ading to my dic\u00adtio\u00adnary, there are about 72 words in Eng\u00adlish and 2000 words in Dutch that con\u00adtain <code class=\"language-text\">sr<\/code>, which means that while I\u2019m typ\u00ading the word <code class=\"language-text\">disregard<\/code>, the <code class=\"language-text\">sr<\/code> would ex\u00adpand to <code class=\"language-text\">^2<\/code>, giv\u00ading me <code class=\"language-text\">di^2egard<\/code>.<\/p><p>The so\u00adlu\u00adtion to this prob\u00adlem is adding a <em>con\u00adtext<\/em> to snip\u00adpets. Using the syn\u00adtax high\u00adlight\u00ading of Vim, it can be de\u00adter\u00admined whether or not Ul\u00adtiSnips should ex\u00adpand the snip\u00adpet de\u00adpend\u00ading if you\u2019re in math or text.  I came up with the fol\u00adlow\u00ading:<\/p><div class=\"gatsby-highlight\" data-language=\"python\"><pre class=\"language-python\">global !p\ntexMathZones = ['texMathZone'+x for x in ['A', 'AS', 'B', 'BS', 'C',\n'CS', 'D', 'DS', 'E', 'ES', 'F', 'FS', 'G', 'GS', 'H', 'HS', 'I', 'IS',\n'J', 'JS', 'K', 'KS', 'L', 'LS', 'DS', 'V', 'W', 'X', 'Y', 'Z']]\ntexIgnoreMathZones = ['texMathText']\ntexMathZoneIds = vim.eval('map('+str(texMathZones)+\", 'hlID(v:val)')\")\ntexIgnoreMathZoneIds = vim.eval('map('+str(texIgnoreMathZones)+\", 'hlID(v:val)')\")\nignore = texIgnoreMathZoneIds[0]\ndef math():\n\tsynstackids = vim.eval(\"synstack(line('.'), col('.') - (col('.')&gt;=2 ? 1 : 0))\")\n\ttry:\n\t\tfirst = next(\n            i for i in reversed(synstackids)\n            if i in texIgnoreMathZoneIds or i in texMathZoneIds\n        )\n\t\treturn first != ignore\n\texcept StopIteration:\n\t\treturn False\nendglobal<\/pre><\/div><p>Now you can add <code class=\"language-text\">context \"math()\"<\/code> to the snip\u00adpets you\u2019d only want to ex\u00adpand in a math\u00ade\u00admat\u00adi\u00adcal con\u00adtext.<\/p><div class=\"gatsby-highlight\" data-language=\"snip\"><pre class=\"language-snip\">context \"math()\"\nsnippet sr \"^2\" iA\n^2\nendsnippet<\/pre><\/div><p>Note that a \u2018math\u00ade\u00admat\u00adi\u00adcal con\u00adtext\u2019 is a sub\u00adtle thing.\nSome\u00adtimes you add some text in\u00adside a math en\u00advi\u00adron\u00adment by using <code class=\"language-text\">\\text{...}<\/code>.\nIn that case, you do not want snip\u00adpets to ex\u00adpand.\nHow\u00adev\u00ader, in the fol\u00adlow\u00ading case: <code class=\"language-text\">\\[ \\text{$...$} \\]<\/code>, they <em>should<\/em> ex\u00adpand.\nThis is why the code for the <code class=\"language-text\">math<\/code> con\u00adtext is a bit com\u00adpli\u00adcat\u00aded.\nThe fol\u00adlow\u00ading an\u00adi\u00adma\u00adtion il\u00adlus\u00adtrates these sub\u00adtleties.<\/p><p><img src=\"https:\/\/castel.dev\/syntaxtree-e1b3ca2dd4fc1f5fe39679fb6ba38aad.gif\" alt=\"syntaxtree\" \/><\/p><h2>Cor\u00adrect\u00ading spelling mis\u00adtakes on the fly<\/h2><p>While in\u00adsert\u00ading math\u00ade\u00admat\u00adics is an im\u00adpor\u00adtant part of my note-\u200btaking setup, most of the time I\u2019m typ\u00ading Eng\u00adlish. At about 80 words per minute, my typ\u00ading skills are not bad, but I still make a lot of typos.\nThis is why I added a key\u00adbind\u00ading to Vim that cor\u00adrects the spelling mis\u00adtakes, with\u00adout in\u00adter\u00adrupt\u00ading my flow. When I press <kbd>Ctrl+L<\/kbd> while I\u2019m typ\u00ading, the pre\u00advi\u00adous spelling mis\u00adtake is cor\u00adrect\u00aded. It looks like this:<\/p><p><img src=\"https:\/\/castel.dev\/spellcheck-0add705683b79cb94498104251a6d5b8.gif\" alt=\"spellcheck\" \/><\/p><p>My con\u00adfig\u00adu\u00adra\u00adtion for spell check is the fol\u00adlow\u00ading:<\/p><div class=\"gatsby-highlight\" data-language=\"vim\"><pre class=\"language-vim\">setlocal spell\nset spelllang=nl,en_gb\ninoremap &lt;C-l&gt; &lt;c-g&gt;u&lt;Esc&gt;[s1z=`]a&lt;c-g&gt;u<\/pre><\/div><p>It ba\u00adsi\u00adcal\u00adly jumps to the pre\u00advi\u00adous spelling mis\u00adtake <code class=\"language-text\">[s<\/code>, then picks the first sug\u00adges\u00adtion <code class=\"language-text\">1z=<\/code>, and then jumps back <code class=\"language-text\">`]a<\/code>. The <code class=\"language-text\">&lt;c-g&gt;u<\/code> in the mid\u00addle make it pos\u00adsi\u00adble to undo the spelling cor\u00adrec\u00adtion quick\u00adly.<\/p><h2>In con\u00adclu\u00adsion<\/h2><p>Using snip\u00adpets in Vim, writ\u00ading LaTeX is no longer an an\u00adnoy\u00adance, but rather a plea\u00adsure.\nIn com\u00adbi\u00adna\u00adtion with spell check on the fly, it al\u00adlows for a com\u00adfort\u00adable math\u00ade\u00admat\u00adi\u00adcal note-\u200btaking setup.\nA few pieces are miss\u00ading though, for ex\u00adam\u00adple draw\u00ading fig\u00adures dig\u00adi\u00adtal\u00adly and em\u00adbed\u00adding them in a LaTeX doc\u00adu\u00adment. This is a topic I\u2019d like to tack\u00adle in a fu\u00adture blog post.<\/p>","created_at":"2019-03-21T13:00:46+0000","updated_at":"2019-03-21T13:00:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"castel.dev","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["wordpress","oauth","auth"],"is_public":false,"id":883,"title":"WP-API\/OAuth2","url":"https:\/\/github.com\/WP-API\/OAuth2","content":"<article class=\"markdown-body entry-content p-5\" itemprop=\"text\">\n<p>Connect applications to your WordPress site without ever giving away your password.<\/p>\n<p>This plugin uses the OAuth 2 protocol to allow delegated authorization; that is, to allow applications to access a site using a set of secondary credentials. This allows server administrators to control which applications can access the site, as well as allowing users to control which applications have access to their data.<\/p>\n<p>This plugin only supports WordPress &gt;= 4.8.<\/p>\n<h2><a id=\"user-content-contributors-welcome\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributors-welcome\"><\/a>Contributors Welcome!<\/h2>\n<p>This plugin works and is in use in several production environments, but the user experience and documentation could be substantially improved. We welcome input and contributions to make this tool better!<\/p>\n<h2><a id=\"user-content-credits\" class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h2>\n<p>This plugin is licensed under the GNU General Public License v2 or later:<\/p>\n<blockquote>\n<p>Copyright 2017 by the contributors.<\/p>\n<p>This program is free software; you can redistribute it and\/or modify\nit under the terms of the GNU General Public License as published by\nthe Free Software Foundation; either version 2 of the License, or\n(at your option) any later version.<\/p>\n<p>This program is distributed in the hope that it will be useful,\nbut WITHOUT ANY WARRANTY; without even the implied warranty of\nMERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\nGNU General Public License for more details.<\/p>\n<p>You should have received a copy of the GNU General Public License\nalong with this program; if not, write to the Free Software\nFoundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA<\/p>\n<\/blockquote>\n<p>Thanks to the contributors at the WCEU 2017 Contributor Day who were responsible for getting this plugin off the ground and into a usable state: @almirbi, @richardsweeney, @tfrommen.<\/p>\n<\/article>","created_at":"2019-03-20T14:19:28+0000","updated_at":"2019-03-20T14:19:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/5706008?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":882,"title":"Pixelfed \u2013 An alternative to centralized image sharing platforms","url":"https:\/\/news.ycombinator.com\/item?id=19420212","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D19420212\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"19420212\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/pixelfed.social\/site\/about\" class=\"storylink\">Pixelfed \u2013 An alternative to centralized image sharing platforms<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=pixelfed.social\">pixelfed.social<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        252 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=colinprince\" class=\"hnuser\">colinprince<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=19420212\">20 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=19420212&amp;goto=item%3Fid%3D19420212\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Pixelfed%20%E2%80%93%20An%20alternative%20to%20centralized%20image%20sharing%20platforms&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Pixelfed%20%E2%80%93%20An%20alternative%20to%20centralized%20image%20sharing%20platforms\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=19420212&amp;auth=c9d89cd282836c02ce2cd5c556a0315c10764066\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=19420212\">96\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/www.ycombinator.com\/apply\/\">\n        Applications are open for YC Summer 2019\n      <\/a><\/div><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-03-20T13:48:36+0000","updated_at":"2019-03-20T13:48:36+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["linux","kernel","vfs","filesystems","os"],"is_public":false,"id":881,"title":"Virtual filesystems in Linux: Why we need them and how they work","url":"https:\/\/opensource.com\/article\/19\/3\/virtual-filesystems-linux","content":"<p>What is a filesystem? According to early Linux contributor and author <a href=\"https:\/\/www.pearson.com\/us\/higher-education\/program\/Love-Linux-Kernel-Development-3rd-Edition\/PGM202532.html\" target=\"_blank\">Robert Love<\/a>, \"A filesystem is a hierarchical storage of data adhering to a specific structure.\" However, this description applies equally well to VFAT (Virtual File Allocation Table), Git, and <a href=\"http:\/\/cassandra.apache.org\/\" target=\"_blank\">Cassandra<\/a> (a <a href=\"https:\/\/en.wikipedia.org\/wiki\/NoSQL\" target=\"_blank\">NoSQL database<\/a>). So what distinguishes a filesystem?<\/p><h2 id=\"filesystem-basics\">Filesystem basics<\/h2><p>The Linux kernel requires that for an entity to be a filesystem, it must also implement the <strong>open()<\/strong>, <strong>read()<\/strong>, and <strong>write()<\/strong> methods on persistent objects that have names associated with them. From the point of view of <a href=\"http:\/\/lwn.net\/Articles\/444910\/\" target=\"_blank\">object-oriented programming<\/a>, the kernel treats the generic filesystem as an abstract interface, and these big-three functions are \"virtual,\" with no default definition. Accordingly, the kernel's default filesystem implementation is called a virtual filesystem (VFS).<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"If we can open(), close(), read() and write(), it is a file as this console session shows.\" data-delta=\"1\" data-fid=\"425686\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_1-console.png\" title=\"Console\" width=\"680\" height=\"202\" \/><figcaption><em>If we can open(), read(), and write(), it is a file as this console session shows.<\/em><\/figcaption><\/figure><p>VFS underlies the famous observation that in Unix-like systems \"everything is a file.\" Consider how weird it is that the tiny demo above featuring the character device <em>\/dev\/console<\/em> actually works. The image shows an interactive Bash session on a virtual teletype (tty). Sending a string into the virtual console device makes it appear on the virtual screen. VFS has other, even odder properties. For example, it's <a href=\"https:\/\/lwn.net\/Articles\/22355\/\" target=\"_blank\">possible to seek in them<\/a>.<\/p>The familiar filesystems like ext4, NFS, and \/proc all provide definitions of the big-three\u00a0functions in a C-language data structure called<a href=\"https:\/\/git.kernel.org\/pub\/scm\/linux\/kernel\/git\/torvalds\/linux.git\/tree\/include\/linux\/fs.h\">file_operations<\/a>. In addition, particular filesystems extend and override the VFS functions in the familiar object-oriented way. As Robert Love points out, the abstraction of VFS enables Linux users to blithely copy files to and from foreign operating systems or abstract entities like pipes without worrying about their internal data format. On behalf of userspace, via a system call, a process can copy from a file into the kernel's data structures with the read() method of one filesystem, then use the write() method of another kind of filesystem to output the data.<p>The function definitions that belong to the VFS base type itself are found in the <a href=\"https:\/\/git.kernel.org\/pub\/scm\/linux\/kernel\/git\/torvalds\/linux.git\/tree\/fs\" target=\"_blank\">fs\/*.c files<\/a> in kernel source, while the subdirectories of fs\/ contain the specific filesystems. The kernel also contains filesystem-like entities such as cgroups, \/dev, and tmpfs, which are needed early in the boot process and are therefore defined in the kernel's init\/ subdirectory. Note that cgroups, \/dev, and tmpfs do not call the file_operations big-three\u00a0functions, but directly read from and write to memory instead.<\/p><p>The diagram below roughly illustrates how userspace accesses various types of filesystems commonly mounted on Linux systems. Not shown are constructs like pipes, dmesg, and POSIX clocks that also implement struct file_operations and whose accesses therefore pass through the VFS layer.<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"How userspace accesses various types of filesystems\" data-delta=\"2\" data-fid=\"425691\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_2-shim-layer.png\" title=\"How userspace accesses various types of filesystems\" width=\"689\" height=\"388\" \/><figcaption><em>VFS are a \"shim layer\" between system calls and implementors of specific file_operations like ext4 and procfs. The file_operations functions can then communicate either with device-specific drivers or with memory accessors. tmpfs, devtmpfs and cgroups don't make use of file_operations but access memory directly.<\/em><\/figcaption><\/figure><p>VFS's existence promotes code reuse, as the basic methods associated with filesystems need not be re-implemented by every filesystem type. Code reuse is a widely accepted software engineering best practice! Alas, if the reused code <a href=\"https:\/\/lwn.net\/Articles\/774114\/\" target=\"_blank\">introduces serious bugs<\/a>, then all the implementations that inherit the common methods suffer from them.<\/p><h2 id=\"tmp-a-simple-tip\">\/tmp: A simple tip<\/h2><p>An easy way to find out what VFSes are present on a system is to type <strong>mount | grep -v sd | grep -v :\/<\/strong>, which will list all mounted filesystems that are not resident on a disk and not NFS on most computers. One of the listed VFS mounts will assuredly be \/tmp, right?<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Man with shocked expression\" data-delta=\"3\" data-fid=\"425701\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_3-crazy.jpg\" title=\"Man with shocked expression\" width=\"500\" height=\"375\" \/><figcaption><em>Everyone knows that keeping \/<\/em><em>tmp<\/em><em> on a physical storage device is crazy! credit: <a href=\"https:\/\/tinyurl.com\/ybomxyfo\">https:\/\/tinyurl.com\/ybomxyfo<\/a><\/em><\/figcaption><\/figure><p>Why is keeping \/tmp on storage inadvisable? Because the files in \/tmp are temporary(!), and storage devices are slower than memory, where tmpfs are created. Further, physical devices are more subject to wear from frequent writing than memory is. Last, files in \/tmp may contain sensitive information, so having them disappear at every reboot is a feature.<\/p><p>Unfortunately, installation scripts for some Linux distros still create \/tmp on a storage device by default. Do not despair should this be the case with your system. Follow simple instructions on the always excellent <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Tmpfs\" target=\"_blank\">Arch Wiki<\/a> to fix the problem, keeping in mind that memory allocated to tmpfs is not available for other purposes. In other words, a system with a gigantic tmpfs with large files in it can run out of memory and crash. Another tip: when editing the \/etc\/fstab file, be sure to end it with a newline, as your system will not boot otherwise. (Guess how I know.)<\/p><h2 id=\"proc-and-sys\">\/proc and \/sys<\/h2><p>Besides \/tmp, the VFSes with which\u00a0most Linux users are most familiar are \/proc and \/sys. (\/dev relies on shared memory and has no file_operations). Why two flavors? Let's have a look in more detail.<\/p><p>The procfs offers a snapshot into the instantaneous state of the kernel and the processes that it controls for userspace. In \/proc, the kernel publishes information about the facilities it provides, like interrupts, virtual memory, and the scheduler. In addition, \/proc\/sys is where the settings that are configurable via the <a href=\"http:\/\/man7.org\/linux\/man-pages\/man8\/sysctl.8.html\" target=\"_blank\">sysctl command<\/a> are accessible to userspace. Status and statistics on individual processes are reported in \/proc\/&lt;PID&gt; directories.<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Console\" data-delta=\"4\" data-fid=\"425706\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_4-proc-meminfo.png\" title=\"Console\" width=\"628\" height=\"477\" \/><figcaption><em>\/proc\/<\/em><em>meminfo<\/em><em> is an empty file that nonetheless contains valuable information.<\/em><\/figcaption><\/figure><p>The behavior of \/proc files illustrates how unlike on-disk filesystems VFS can be. On the one hand, \/proc\/meminfo contains the information presented by the command <strong>free<\/strong>. On the other hand, it's also empty! How can this be? The situation is reminiscent of a famous article written by Cornell University physicist N. David Mermin in 1985 called \"<a href=\"http:\/\/www-f1.ijs.si\/~ramsak\/km1\/mermin.moon.pdf\" target=\"_blank\">Is the moon there when nobody looks?<\/a> Reality and the quantum theory.\" The truth is that the kernel gathers statistics about memory when a process requests them from \/proc, and there actually <em>is<\/em> nothing in the files in \/proc when no one is looking. As <a href=\"https:\/\/en.wikiquote.org\/wiki\/David_Mermin\" target=\"_blank\">Mermin said<\/a>, \"It is a fundamental quantum doctrine that a measurement does not, in general, reveal a preexisting value of the measured property.\" (The answer to the question about the moon is left as an exercise.)<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Full moon\" data-delta=\"5\" data-fid=\"425711\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_5-moon.jpg\" title=\"Full moon\" width=\"500\" height=\"504\" \/><figcaption><em>The files in \/proc are empty when no process accesses them. (<a href=\"https:\/\/commons.wikimedia.org\/wiki\/Moon#\/media\/File:Full_Moon_Luc_Viatour.jpg\">Source<\/a>)<\/em><\/figcaption><\/figure><p>The apparent emptiness of procfs makes sense, as the information available there is dynamic. The situation with sysfs is different. Let's compare how many files of at least one byte in size there are in \/proc versus \/sys.<\/p><div class=\"media media-element-container media-default media-wysiwyg-align-center\"><div id=\"file-425716\" class=\"file file-image file-image-png\"><div class=\"content\"><img alt=\"Console\" title=\"Console\" class=\"media-element file-default\" data-delta=\"6\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_6-filesize.png\" width=\"626\" height=\"114\" \/><\/div><\/div><\/div><p>Procfs has precisely one, namely the exported kernel configuration, which is an exception since it needs to be generated only once per boot. On the other hand, \/sys has lots of larger files, most of which comprise one page of memory. Typically, sysfs files contain exactly one number or string, in contrast to the tables of information produced by reading files like \/proc\/meminfo.<\/p><p>The purpose of sysfs is to expose the readable and writable properties of what the kernel calls \"kobjects\" to userspace. The only purpose of kobjects is reference-counting: when the last reference to a kobject is deleted, the system will reclaim the resources associated with it. Yet, \/sys constitutes most of the kernel's famous \"<a href=\"https:\/\/git.kernel.org\/pub\/scm\/linux\/kernel\/git\/torvalds\/linux.git\/tree\/Documentation\/ABI\/stable\" target=\"_blank\">stable ABI to userspace<\/a>\" which <a href=\"https:\/\/lkml.org\/lkml\/2012\/12\/23\/75\" target=\"_blank\">no one may ever, under any circumstances, \"break.\"<\/a> That doesn't mean the files in sysfs are static, which would be contrary to reference-counting of volatile objects.<\/p><p>The kernel's stable ABI instead constrains what <em>can<\/em> appear in \/sys, not what is actually present at any given instant. Listing the permissions on files in sysfs gives an idea of how the configurable, tunable parameters of devices, modules, filesystems, etc. can be set or read. Logic compels the conclusion that procfs is also part of the kernel's stable ABI, although the kernel's <a href=\"https:\/\/git.kernel.org\/pub\/scm\/linux\/kernel\/git\/torvalds\/linux.git\/tree\/Documentation\/ABI\/stable\" target=\"_blank\">documentation<\/a> doesn't state so explicitly.<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Console\" data-delta=\"7\" data-fid=\"425721\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_7-sysfs.png\" title=\"Console\" width=\"680\" height=\"231\" \/><figcaption><em>Files in <\/em><em>sysfs<\/em><em> describe exactly one property each for an entity and may be readable, writable or both. The \"0\" in the file reveals that the SSD is not removable.<\/em><\/figcaption><\/figure><h2 id=\"snooping-on-vfs-with-ebpf-and-bcc-tools\">Snooping on VFS with eBPF and bcc tools<\/h2><p>The easiest way to learn how the kernel manages sysfs files is to watch it in action, and the simplest way to watch on ARM64 or x86_64 is to use eBPF. eBPF (extended Berkeley Packet Filter) consists of a <a href=\"https:\/\/events.linuxfoundation.org\/sites\/events\/files\/slides\/bpf_collabsummit_2015feb20.pdf\" target=\"_blank\">virtual machine running inside the kernel<\/a> that privileged users can query from the command line. Kernel source tells the reader what the kernel <em>can<\/em> do; running eBPF tools on a booted system shows instead what the kernel actually <em>does<\/em>.<\/p><p>Happily, getting started with eBPF is pretty easy via the <a href=\"https:\/\/github.com\/iovisor\/bcc\" target=\"_blank\">bcc<\/a> tools, which are available as <a href=\"https:\/\/github.com\/iovisor\/bcc\/blob\/master\/INSTALL.md\" target=\"_blank\">packages from major Linux distros<\/a> and have been <a href=\"http:\/\/brendangregg.com\/ebpf.html\" target=\"_blank\">amply documented<\/a> by Brendan Gregg. The bcc tools are Python scripts with small embedded snippets of C, meaning anyone who is comfortable with either language can readily modify them. At this count, <a href=\"https:\/\/github.com\/iovisor\/bcc\/tree\/master\/tools\" target=\"_blank\">there are 80 Python scripts in bcc\/tools<\/a>, making it highly likely that a system administrator or developer will find an existing one relevant to her\/his needs.<\/p><p>To get a very crude idea about what work VFSes are performing on a running system, try the simple <a href=\"https:\/\/github.com\/iovisor\/bcc\/blob\/master\/tools\/vfscount_example.txt\" target=\"_blank\">vfscount<\/a> or <a href=\"https:\/\/github.com\/iovisor\/bcc\/blob\/master\/tools\/vfsstat.py\" target=\"_blank\">vfsstat<\/a>, which show that dozens of calls to vfs_open() and its friends occur every second.<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Console - vfsstat.py\" data-delta=\"8\" data-fid=\"425726\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_8-vfsstat.png\" title=\"Console - vfsstat.py\" width=\"585\" height=\"219\" \/><figcaption><em>vfsstat.py is a Python script with an embedded C snippet that simply counts VFS function calls.<\/em><\/figcaption><\/figure><p>For a less trivial example, let's watch what happens in sysfs when a USB stick is inserted on a running system.<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Console when USB is inserted\" data-delta=\"9\" data-fid=\"425731\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_9-ebpf.png\" title=\"Console when USB is inserted\" width=\"680\" height=\"372\" \/><figcaption><em>Watch with eBPF what happens in \/sys when a USB stick is inserted, with simple and complex examples.<\/em><\/figcaption><\/figure><p>In the first simple example above, the <a href=\"https:\/\/github.com\/iovisor\/bcc\/blob\/master\/tools\/trace_example.txt\" target=\"_blank\">trace.py<\/a> bcc tools script prints out a message whenever the sysfs_create_files() command runs. We see that sysfs_create_files() was started by a kworker thread in response to the USB stick insertion, but what file was created? The second example illustrates the full power of eBPF. Here, trace.py is printing the kernel backtrace (-K option) plus the name of the file created by sysfs_create_files(). The snippet inside the single quotes is some C source code, including an easily recognizable format string, that the provided Python script <a href=\"https:\/\/events.static.linuxfound.org\/sites\/events\/files\/slides\/bpf_collabsummit_2015feb20.pdf\" target=\"_blank\">induces a LLVM just-in-time compiler<\/a>\u00a0to compile and execute inside an in-kernel virtual machine. The full sysfs_create_files() function signature must be reproduced in the second command so that the format string can refer to one of the parameters. Making mistakes in this C snippet results in recognizable C-compiler errors. For example, if the <strong>-I<\/strong> parameter is omitted, the result is \"Failed to compile BPF text.\" Developers who are conversant with either C or Python will find the bcc tools easy to extend and modify.<\/p><p>When the USB stick is inserted, the kernel backtrace appears showing that PID 7711 is a kworker thread that created a file called \"events\" in sysfs. A corresponding invocation with sysfs_remove_files() shows that removal of the USB stick results in removal of the events file, in keeping with the idea of reference counting. Watching sysfs_create_link() with eBPF during USB stick insertion (not shown) reveals that no fewer than 48 symbolic links are created.<\/p><p>What is the purpose of the events file anyway? Using <a href=\"http:\/\/northstar-www.dartmouth.edu\/doc\/solaris-forte\/manuals\/c\/user_guide\/cscope.html\" target=\"_blank\">cscope<\/a> to find the function <a href=\"https:\/\/git.kernel.org\/pub\/scm\/linux\/kernel\/git\/torvalds\/linux.git\/tree\/block\/genhd.c#n665\" target=\"_blank\">__device_add_disk()<\/a> reveals that it calls disk_add_events(), and either \"media_change\" or \"eject_request\" may be written to the events file. Here, the kernel's block layer is informing userspace about the appearance and disappearance of the \"disk.\"\u00a0Consider how quickly informative this method of investigating how USB stick insertion works is compared to trying to figure out the process solely from the source.<\/p><p>Assuredly, no one shuts down a server or desktop system by pulling out the power plug. Why? Because mounted filesystems on the physical storage devices may have pending writes, and the data structures that record their state may become out of sync with what is written on the storage. When that happens, system owners will have to wait at next boot for the <a href=\"http:\/\/www.man7.org\/linux\/man-pages\/man8\/fsck.8.html\" target=\"_blank\">fsck filesystem-recovery tool<\/a> to run and, in the worst case, will actually lose data.<\/p><p>Yet, aficionados will have heard that many IoT and embedded devices like routers, thermostats, and automobiles now run Linux. Many of these devices almost entirely lack a user interface, and there's no way to \"unboot\" them cleanly. Consider jump-starting a car with a dead battery where the power to the <a href=\"https:\/\/wiki.automotivelinux.org\/_media\/eg-rhsa\/agl_referencehardwarespec_v0.1.0_20171018.pdf\" target=\"_blank\">Linux-running head unit<\/a> goes up and down repeatedly. How is it that the system boots without a long fsck when the engine finally starts running? The answer is that embedded devices rely on <a href=\"https:\/\/elinux.org\/images\/1\/1f\/Read-only_rootfs.pdf\" target=\"_blank\">a read-only root fileystem<\/a> (ro-rootfs for short).<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Photograph of a console\" data-delta=\"10\" data-fid=\"425736\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_10-code.jpg\" title=\"Photograph of a console\" width=\"500\" height=\"375\" \/><figcaption><em>ro-<\/em><em>rootfs<\/em><em> are why embedded systems don't frequently need to fsck. Credit (with permission): <a href=\"https:\/\/tinyurl.com\/yxoauoub\">https:\/\/tinyurl.com\/yxoauoub<\/a><\/em><\/figcaption><\/figure><p>A ro-rootfs offers many advantages that are less obvious than incorruptibility. One is that malware cannot write to \/usr or \/lib if no Linux process can write there. Another is that a largely immutable filesystem is critical for field support of remote devices, as support personnel possess local systems that are nominally identical to those in the field. Perhaps the most important (but also most subtle) advantage is that ro-rootfs forces developers to decide during a project's design phase which system objects will be immutable. Dealing with ro-rootfs may often be inconvenient or even painful, as <a href=\"https:\/\/www.meetup.com\/ACCU-Bay-Area\/events\/drpmvfytlbqb\/\" target=\"_blank\">const variables in programming languages<\/a> often are, but the benefits easily repay the extra overhead.<\/p><p>Creating a read-only rootfs does require some additional amount of effort for embedded developers, and that's where VFS comes in. Linux needs files in \/var to be writable, and in addition, many popular applications that embedded systems run will try to create configuration dot-files in $HOME. One solution for configuration files in the home directory is typically to pregenerate them and build them into the rootfs. For \/var, one approach is to mount it on a separate writable partition while \/ itself is mounted as read-only. Using bind or overlay mounts is another popular alternative.<\/p><h2 id=\"bind-and-overlay-mounts-and-their-use-by-containers\">Bind and overlay mounts and their use by containers<\/h2><p>Running <strong><a href=\"http:\/\/man7.org\/linux\/man-pages\/man8\/mount.8.html\" target=\"_blank\">man mount<\/a><\/strong> is the best place to learn about bind and overlay mounts, which give embedded developers and system administrators the power to create a filesystem in one path location and then provide it to applications at a second one. For embedded systems, the implication is that it's possible to store the files in \/var on an unwritable flash device but overlay- or bind-mount a path in a tmpfs onto the \/var path at boot so that applications can scrawl there to their heart's delight. At next power-on, the changes in \/var will be gone. Overlay mounts provide a union between the tmpfs and the underlying filesystem and allow apparent modification to an existing file in a ro-rootfs, while bind mounts can make new empty tmpfs directories show up as writable at ro-rootfs paths. While overlayfs is a proper filesystem type, bind mounts are implemented by the <a href=\"https:\/\/git.kernel.org\/pub\/scm\/linux\/kernel\/git\/torvalds\/linux.git\/tree\/Documentation\/filesystems\/sharedsubtree.txt\" target=\"_blank\">VFS namespace facility<\/a>.<\/p><p>Based on the description of overlay and bind mounts, no one will be surprised that <a href=\"https:\/\/coreos.com\/os\/docs\/latest\/kernel-modules.html\" target=\"_blank\">Linux containers<\/a> make heavy use of them. Let's spy on what happens when we employ <a href=\"https:\/\/www.freedesktop.org\/software\/systemd\/man\/systemd-nspawn.html\" target=\"_blank\">systemd-nspawn<\/a> to start up a container by running bcc's mountsnoop tool:<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Console - system-nspawn invocation\" data-delta=\"11\" data-fid=\"425741\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_11-system-nspawn.png\" title=\"Console - system-nspawn invocation\" width=\"615\" height=\"113\" \/><figcaption><em>The system-<\/em><em>nspawn<\/em><em> invocation fires up the container while mountsnoop.py runs.<\/em><\/figcaption><\/figure><p>And let's see what happened:<\/p><figure class=\"image media-wysiwyg-align-center file-default media-element\"><img alt=\"Console - Running mountsnoop\" data-delta=\"12\" data-fid=\"425746\" data-media-element=\"1\" src=\"https:\/\/opensource.com\/sites\/default\/files\/uploads\/virtualfilesystems_12-mountsnoop.png\" title=\"Console - Running mountsnoop\" width=\"680\" height=\"236\" \/><figcaption><em>Running <\/em><em>mountsnoop<\/em><em> during the container \"boot\" reveals that the container runtime relies heavily on bind mounts. (Only the beginning of the lengthy output is displayed)<\/em><\/figcaption><\/figure><p>Here, systemd-nspawn is providing selected files in the host's procfs and sysfs to the container at paths in its rootfs. Besides the MS_BIND flag that sets bind-mounting, some of the other flags that the \"mount\" system call invokes determine the relationship between changes in the host namespace and in the container.\u00a0 For example, the bind-mount can either propagate changes in \/proc and \/sys to the container, or hide them, depending on the invocation.<\/p><h2 id=\"summary\">Summary<\/h2><p>Understanding Linux internals can seem an impossible task, as the kernel itself contains a gigantic amount of code, leaving aside Linux userspace applications and the system-call interface in C libraries like glibc.\u00a0One way to make progress is to read the source code of one kernel subsystem with an emphasis on understanding the userspace-facing system calls and headers plus major kernel internal interfaces, exemplified here by the file_operations table. The file operations are what makes \"everything is a file\" actually work, so getting a handle on them is particularly satisfying. The kernel C source files in the top-level fs\/ directory constitute its implementation of virtual filesystems, which are the shim layer that enables broad and relatively straightforward interoperability of popular filesystems and storage devices. Bind and overlay mounts via Linux namespaces are the VFS magic that makes containers and read-only root filesystems possible. In combination with a study of source code, the eBPF kernel facility and its bcc interface makes probing the kernel simpler than ever before.<\/p><p><em>Much thanks to <a href=\"http:\/\/shallowsky.com\/\">Akkana Peck<\/a> and <a href=\"http:\/\/eagercon.com\/\">Michael Eager<\/a> for comments and corrections.\u00a0<\/em><\/p><hr \/><p><em>Alison Chaiken will present <a href=\"https:\/\/www.socallinuxexpo.org\/scale\/17x\/presentations\/virtual-filesystems-why-we-need-them-and-how-they-work\" target=\"_blank\">Virtual filesystems: why we need them and how they work<\/a> at the 17th annual Southern California Linux Expo (<a href=\"https:\/\/www.socallinuxexpo.org\/\" target=\"_blank\">SCaLE 17x<\/a>) March 7-10 in Pasadena, Calif.<\/em><\/p>","created_at":"2019-03-19T13:12:41+0000","updated_at":"2019-03-19T13:12:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"opensource.com","preview_picture":"https:\/\/opensource.com\/sites\/default\/files\/lead-images\/documents_papers_file_storage_work.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":880,"title":"a Toolkit for Computer-Aided Musicology","url":"https:\/\/web.mit.edu\/music21\/","content":"The first release of music21 version 4 was released today (August 6, 2017). \u00a0This is the first major release of\u00a0music21\u00a0in a year, and it brings with it a wealth of new tools for analyzing music with a computer and performing digital musicology, music theory, and composition.<p>Download from\u00a0<a href=\"https:\/\/github.com\/cuthbertLab\/music21\/releases\">https:\/\/github.com\/cuthbertLab\/music21\/releases<\/a>\u00a0or from the terminal, type:<\/p><p>\u00a0 \u00a0 pip3 install --upgrade music21<\/p><p>(or without the \"3\" if you are using Python 2)<\/p><p>Version 4 is the last version of music21 that will support Python 2.7. \u00a0If you run Version 4 on Python 2.7, you will see a warning that it's time to move up to the brilliance that is Python 3.6.<\/p><p>As with all new \"X\" release names, v.4 has backward incompatible behaviors that I think are worth it for the great new features.\u00a0<\/p><p>Among the 272 commits since v.3.1:<\/p><p><em>Major new features:<\/em><\/p><ul><li>Graphing rewrite!<\/li><\/ul><div>.plot() and Graphing has always been one of the most powerful parts of\u00a0<code>music21<\/code>\u00a0since long before v.1.0 (mad props, Christopher Ariza!) but it's also been one of the most daunting aspects of using\u00a0<code>music21<\/code>. It shouldn't be any more. The code has gone through a major rewrite to improve the simplicity of doing easy things and the power when doing difficult things. The easy things are documented in\u00a0<a href=\"http:\/\/web.mit.edu\/music21\/doc\/usersGuide\/usersGuide_22_graphing.html\">Chapter 22<\/a>\u00a0and the hard things in\u00a0<a href=\"http:\/\/web.mit.edu\/music21\/doc\/usersGuide\/usersGuide_44_advancedGraphing.html\">Chapter 44<\/a>.<\/div><ul><li>Local Corpora are great!<\/li><\/ul><p>There's been a major rewrite of the corpus.corpora.LocalCorpus() function that makes it a fantastic way to work quickly with files you are working on from your own hard drive. \u00a0Set up your own local corpus, add paths to it, set the cacheFilePath to somewhere to store the metadata and you'll be able to search for pieces with particular features and metadata without needing to parse the score.<\/p><ul><li>Style!<\/li><\/ul><p>the all new style module and style.Style object handles aspects of a note or other object's visual display that are not (usually) semantic. This class has allowed a major increase in the ability to properly preserve MusicXML visual formatting on input and export.<\/p><p>Style objects are created only when needed, so el.hasStyleInformation() allows for checking for the presence of a .style object without creating one.<\/p><p>(To be documented more soon)<\/p><ul><li>Major rewrite of TinyNotation allows for easy extensibility.<\/li><\/ul><p>Documented in the\u00a0<a href=\"http:\/\/web.mit.edu\/music21\/doc\/usersGuide\/usersGuide_16_tinyNotation.html\">User's Guide<\/a>! Check it out!<\/p><ul><li>Always improving docs<\/li><\/ul><p>The User's Guide goes up to chapter 24 now, with major new examples in\u00a0<a href=\"http:\/\/web.mit.edu\/music21\/doc\/usersGuide\/usersGuide_20_examples2.html\">Chapter 20<\/a>\u00a0along with rewritten chapters on keys, time signatures, sorting, and so on and so on. Plus all examples are now Retina quality for viewing fine details of scores.<\/p><p>Documentation has been moved out of the music21 directory into the root directory -- it is no longer installed with music21 from pip -- this change was necessitated by the move to retina quality graphics, but reduces the installation size from 90MB to 15MB for the full corpus version and 6MB for the no-corpus version. \u00a0All documents are now tested with nbval, to ensure they stay up to date.<\/p><p><em>Other new features<\/em><\/p><ul><li>Stream.measures() now optionally allows for indexes (where 0 always is the first measure; -1 is the last, and so on) making .getElementsByClass('Measure') not necessary in most cases.<\/li><li>much better metadata processing in musicxml, humdrum, and braille<\/li><li>improved braille translation (tuplets) -- thanks Bo-Chen<\/li><li>better beaming, meter, and tuplets in ABC<\/li><li>output directly to PDF if MuseScore is installed.<\/li><li>Nested Tuplets! including in MusicXML.<\/li><li>Non-traditional key signatures<\/li><li>New works by Clara Schumann in the corpus.<\/li><li>stream.iterators.OffsetIterator() -- iterate groups of objects by offset.<\/li><li>improvements to analysis.discrete<\/li><li>demos\/build_melody shows how to build MidiFile directly (thanks PeterMitrano!)<\/li><li>corpus paths are now searchable in corpus.search()<\/li><li>matplotlib and musescore graphics in Jupyter notebook are now retina quality.<\/li><li>Chord.add() and Chord.remove() allow for direct manipulation of chords.<\/li><li>Improvements to parallel processing in music21.<\/li><li>Ottava spanners now come in two types, transposing and non-transposing -- reflecting whether the pitches under the spanner already reflect the transposition (non-transposing) or not.<\/li><li>Palestrina humdrum has been reorganized and parses completely.<\/li><li>Many improvements to spanners and RomanNumerals.<\/li><li>Beams work much better and transfer in and out of MusicXML more completely.<\/li><li>Every MusicXML 3.0 articulation is now supported.<\/li><li>Core routines in stream.core have now been exposed publicly. \u00a0They are dangerous to use, but for anyone working on their own parsers, they can speed up insertions and appends by an order of magnitude.<\/li><\/ul><p>Others, including bugs squashed:<\/p><ul><li>Warning on Python 2 that music21 v. 4 is the last version to support Py 2.<\/li><li>Stream.template() is a great way to get an empty stream that otherwise matches the current Stream. \u00a0Replaces the obsolete .measureTemplate()<\/li><li>ABC key signature and mode error fixed.<\/li><li>RecursiveIterator gets a .currentOffsetInHierarchy, which can let even more places remove the dependence on .flat. \u00a0In fact, .flat uses the .recurse() method internally, because recurse() is now so fast.<\/li><li>AudioSearch bugs fixed (thanks jjrob13)<\/li><li>Chord.normalOrderString (thanks emzhang)<\/li><li>Removed lots of old crutches including the \"analysisData\" on Stream, Note.editorial, and others. Style fixes most of this.<\/li><li>fix to Bach BWV 386 (thanks alexcoplan) and to Beethoven Opus 59 no 3, movement 1.<\/li><li>Note.pitches returns a tuple not list, just like Chord.pitches<\/li><li>Converter can deal with some wrong file extensions now.<\/li><li>Instrument reprs are fixed<\/li><li>configure finds many more notation programs.<\/li><li>ties are imported better between elements in and out of voices in musicxml<\/li><li>configure works on macOS when user directory contains spaces.<\/li><li>Bugs in ending and restarting a recursiveIterator fixed.<\/li><li>doc errors fixed (thanks Andrew Sanchez)<\/li><li>bestClef() has been moved to the clef module where it belongs.<\/li><li>MusicXML sound tag now is placed properly (thanks Almog Cohen)<\/li><li>MIDI output from transposed scores now plays in concert pitch.<\/li><li>Breves are now acceptable as full measure rests.<\/li><\/ul><p><em>Deprecations and deprecated elements removed<\/em><\/p><ul><li>(this list does not contain changes to the alpha\/ directory which can change at any time)<\/li><li>Note.ps, and Note.accidental, Note.pitchClass, Note.pitchClassString, Note.diatonicNoteNum, and Note.microtone are all deprecated. Use Note.pitch.ps, etc. instead.<\/li><li>Chord.normalForm is deprecated because it gave the wrong answer. use normalOrder instead; same with normalFormString<\/li><li>SpannerBundle.list is deprecated; use list(SpannerBundle) instead<\/li><li>with the advent of .style, el.color is deprecated, use el.style.color instead<\/li><li>Stream.stream() is deprecated -- now that the transition to iterators is done, there should be no need for this.<\/li><li><em>REMOVED<\/em>\u00a0stream.getOffsetByElement; use s.elementOffset(el) instead.<\/li><li><em>REMOVED<\/em>\u00a0stream.haveBeamsBeenMade; use stream.streamStatus.haveBeamsBeenMade<\/li><li><em>REMOVED<\/em>\u00a0stream.makeTupletBrackets(); use stream.makeNotation.makeTupletBrackets(s)<\/li><li><em>REMOVED<\/em>\u00a0stream.realizeOrnaments; use stream.makeNotation.realizeOrnaments(s)<\/li><li><i>REMOVED <\/i>VirtualCorpus -- it may return at some point but with a lot more features.<\/li><li>nbconvert is no longer packaged with music21<\/li><li>.exe files are no longer generated -- they were rarely used and pip is a better choice for Windows users now.<\/li><\/ul><p>As the last version of music21 to run on Python 2, version 4 will have a longer support period for security patches and major bug fixes that render large parts of the system unusable for multiple users. This LTS (Long-term support) will run until the <a href=\"https:\/\/pythonclock.org\/\" target=\"_blank\">clock on python 2<\/a> runs out. (Currently two more years and eight months). \u00a0After that time, Python 2 will no longer be supported by the Python Foundation, and thus, by music21 either. \u00a0We're looking forward to joining the glorious Python 3-only feature and finally get to use some great features to make development faster and more stable. \u00a0Music21 v. 5 will be released at about the same time Django v. 2 (Py3 only) and a year after IPython\/Jupyter made their first Python 3 only release. \u00a0Matplotlib will soon follow. \u00a0<a href=\"http:\/\/www.python3statement.org\/\" target=\"_blank\">Python 3 is the future<\/a> of humanistic and scientific programming.<\/p><p>As always, we thank the community for great support. \u00a0We'd always love to hear how you use music21 via the Google Groups mailing list. \u00a0Music21 was made possible by grants from the Seaver Institute, the National Endowment for the Humanities, and the School of Humanities, Arts, and Social Sciences \/ Music and Theater Arts Section at MIT.<\/p>","created_at":"2019-03-19T12:20:40+0000","updated_at":"2019-03-19T12:20:40+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"web.mit.edu","http_status":"200"},{"is_archived":0,"is_starred":1,"tags":[],"is_public":false,"id":879,"title":"x775 - Introduction to Datalog","url":"https:\/\/x775.net\/2019\/03\/18\/Introduction-to-Datalog.html","content":"<header><section id=\"post-body\"><p><em>The following post will serve as a brief introduction to the declarative programming paradigm through a closer examination of the logic programming language Datalog. I am writing this as I go, hoping to both further my own understanding of the paradigm and language whilst, hopefully, providing others with the easy, practically-oriented introduction I myself have missed. Please consider everything in this post work-in-progress and, consequently, subject to change. Any questions, comments, or other feedback, please get in touch.<\/em><\/p><p>In contrast to the imperative programming paradigm, declarative programming expresses the logic of a computation without explicitly specifying its control flow. That is, declarative programming focuses on the <em>what<\/em>, not the <em>how<\/em>. If you have tentatively researched declarative programming before, you have likely read something akin to the above description already. If, on the other hand, this is your first time reading anything about declarative programming, the idea of developing an application <em>without<\/em> focusing on the <em>how<\/em> likely sounds alien \u2013 it sure did for me!<\/p><p>However, chances are that you have already worked declaratively without realising! Indeed, ubiquitous, domain-specific languages like HTML, CSS, and SQL all belong to the declarative paradigm. HTML, for instance, merely describes <em>what<\/em> should appear on a given webpage without regard to the order in which the various elements should be rendered. Similarly, a SQL query just describes <em>what<\/em> to return, not <em>how<\/em> to compute the return value. This, in essence, is what people mean when they say declarative programming focuses on the <em>what<\/em>.<\/p><p>It is worth accentuating how the term <em>declarative programming<\/em> is really an umbrella term which, in addition to the domain-specific languages mentioned above, also encapsulates the subparadigm <em>logic programming<\/em>. This is the subparadigm which we are going to concern ourselves with throughout this tutorial.<\/p><hr \/><p><strong>Logic programming<\/strong> is based on formal logic. This means that any program written in a logic programming language, e.g. Datalog or Prolog, is nothing more than a set of sentences in logical form, expressing facts and rules about a given domain. This allows logic programming to boast some nice features that other paradigms lack, namely:<\/p><ol><li>no side-effects,<\/li>\n  <li>no explicit control flow, and<\/li>\n  <li>strong guarantees about termination. \n<p>However, the fact that logic programming relies on sentences in logical form also affects its expressive power\u2014that is, which algorithms can be expressed using which logic programming languages. The below illustration seeks to visualise this.\n<\/p><\/li>\n<\/ol><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/datalog.png\" alt=\"drawing\" width=\"65%\" \/><\/div><p><br \/>The innermost circle encapsulates <a href=\"https:\/\/en.wikipedia.org\/wiki\/Relational_algebra\">relational algebra<\/a>\u2014something which you might be familiar with if you have SQL-experience\u2014and shows Datalog as an immediate extension. We consider Datalog an immediate extension of relational algebra because whilst it, unlike relational algrebra, supports recursion, it can do little more. If you are unfamiliar with relational algebra there is no need to worry; you are not going to be writing any relational algebra in this tutorial. This is all about expressive power and putting things into perspective.<\/p><p>The next circle extends Datalog with negation. This allows us to implement any polynomial time algorithm using what is called <em>stratified<\/em> Datalog. It is worth noting that Datalog without negation, like relational algebra, is hardly useful in the real-world. All references to <em>Datalog<\/em> will thus hereinafter refer to the version supporting negation unless explicitly stated otherwise. If we take some precautions when extending Datalog with negation, we continue to retain strong guarantees about termination.<\/p><p>If we wish to express algorithms beyond those of polynomial time, however, we have to extend Datalog yet again. This time we move to the outermost circle, and effectively transform Datalog to Prolog in the process. Prolog is a Turing-complete logic programming language which enables us to express any algorithm we so desire. As a direct consequence of this newly gained expressive power, however, we lose our nice guarantees about termination. I will briefly return to Prolog towards the end of this tutorial.<\/p><h4 id=\"a-new-way-of-thinking\">A new way of thinking<\/h4><p>It is with the above in mind perhaps not surprising that Datalog, at its core, is a very simple language. The challenge with Datalog is thus not learning an extensive syntax, fighting your compiler, or juggling numerous libraries. The challenge, rather, lies, as with any new programming paradigm, in rewiring your brain to understand the concept. It <em>is<\/em> a new way of thinking!<\/p><p>We have already established how any program written in a logic programming language is written as a collection of <em>logic constraints<\/em>\u2014that is, sentences in logical form. The compiler uses this collection of constraints to compute a solution. Read that again. Yes, really! In fact, the compiler can freely choose which algorithms and data structures to employ in order to compute a solution when prompted. We have no guarantees that the compiler will always choose the same approach, and it might even decide to do things in parallel! Given the same input, however, the compiler will <em>always<\/em> produce the same output. Fortunately, since the compiler takes care of this behind the scenes, all we need to concern ourselves with is writing proper constraints.<\/p><h4 id=\"facts-and-rules\">Facts and rules<\/h4><p>All logic programs\u2014that is, our collection of logic constraints\u2014consist of a finite set of <em>facts<\/em> and <em>rules<\/em>, and Datalog programs are no exception. The concept of facts and rules is vaguely equivalent to that of classes and functions in other languages. In short:<\/p><ul><li><em>facts<\/em> = assertions of the world in which we operate.<\/li>\n  <li><em>rules<\/em> = sentences allowing us to infer new facts from existing ones.\n<p>So far, so good\u2026 except, how do we actually write a Datalog program? Easy enough. We start by introducing some facts. Imagine, for instance, that we want to model our shelf with movies neatly organised by their main genres. This is straightforward:\n<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Movie(\"Pulp Fiction\", \"Crime\").\nMovie(\"The Hateful Eight\", \"Crime\"). \nMovie(\"Reservoir Dogs\", \"Thriller\").\n<\/pre><\/div><\/div>\n  <\/li>\n<\/ul><p>The above says that we have three facts, each holding two values and ended by a <code class=\"highlighter-rouge\">.<\/code> (period). In this specific case, the two values are the title of the movie and the genre of the movie, respectively. The name\u2014that is, the predicate\u2014of each fact is <code class=\"highlighter-rouge\">Movie<\/code>. If we then want to expand our world with, say, the actors from each movie:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>StarringIn(\"Pulp Fiction\", \"Tim Roth\").\nStarringIn(\"The Hateful Eight\", \"Tim Roth\").\nStarringIn(\"Reservoir Dogs\", \"Harvey Keitel\").\n...\n<\/pre><\/div><\/div><p>Similarly, if we want to include the name of the directors:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>DirectedBy(\"Pulp Fiction\", \"Quentin Tarantino\").\nDirectedBy(\"The Hateful Eight\", \"Quentin Tarantino\").\nDirectedBy(\"Reservoir Dogs\", \"Quentin Tarantino\").\n<\/pre><\/div><\/div><p>And so on so forth!<\/p><p>It is generally a good idea to introduce your facts at the beginning of your program to enforce some concept of structure and order. However, if you prefer a different structure\u2014perhaps you want things alphabetically\u2014you can freely do so! Indeed, the lack of explicit control flow means that there is no pre-defined structure which you ought to obey. Additionally, when making facts, the order of your constants does not matter. That is:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>DirectedBy(\"Reservoir Dogs\", \"Quentin Tarantino\").\n<\/pre><\/div><\/div><p>is as valid as<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>DirectedBy(\"Quentin Tarantino\", \"Reservoir Dogs\").\n<\/pre><\/div><\/div><p>You should thus simply opt for the approach which makes the most sense to <em>you<\/em>, and otherwise be consistent! The exact order of your constants only matters when writing rules, as we will see shortly.<\/p><p>Now, having modelled our movie collection, imagine a friend who really enjoys crime movies visits. He has a\u2014suspiciously\u2014specific preference in movies, and especially appreciates those directed by Quentin Tarantino and which stars Tim Roth. Which movie(s) should we put on?<\/p><p>We could obviously answer this by reading through the above facts, but imagine our collection contained tens of thousands of titles\u2014and thus exponentially more facts. Or, since few people are likely to have that many movies, imagine you are Netflix and need to recommend new movies based on what a given user has already expressed an interest in. The concept here is the same, and in both scenarios, opting for a manual approach quickly becomes mindbogglingly tedious, if not impossible.<\/p><p>Introducing <em>rules<\/em>!<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>ExtraAwesomeMovie(title) :-\n    Movie(title, \"Crime\"),\n    StarringIn(title, \"Tim Roth\"),\n    DirectedBy(title, \"Quentin Tarantino\").\n<\/pre><\/div><\/div><p>This rule reads as follows:<\/p><blockquote>\n  <p><em>Iff there is a movie with <strong>title<\/strong>, and the genre of that <strong>title<\/strong> is <code class=\"highlighter-rouge\">Crime<\/code>, and the <strong>title<\/strong> stars <code class=\"highlighter-rouge\">Tim Roth<\/code>, and the <strong>title<\/strong> is directed by <code class=\"highlighter-rouge\">Quentin Tarantino<\/code><\/em>.<\/p>\n<\/blockquote><p>We know we are dealing with a rule due to the <code class=\"highlighter-rouge\">:-<\/code> part at the end of the first line. The <code class=\"highlighter-rouge\">:-<\/code> means <em>if and only if<\/em>, or iff. Note that rules are checked against every available fact in our program, but only return the fact(s)\u2014in this case movie(s)\u2014for which the rule in question holds true. That is, we <em>infer<\/em> that a movie is an <code class=\"highlighter-rouge\">ExtraAwesomeMovie<\/code> if and only if the above rule holds true.<\/p><p>If our example program had contained the fact <code class=\"highlighter-rouge\">FirstName(\"Jennifer\").<\/code>, this fact would, consequently, also have been checked in the above rule, but since there exists no <code class=\"highlighter-rouge\">Movie<\/code>-fact with title \u201cJennifer\u201d and genre \u201cCrime\u201d, the compiler would simply have continued to the next available fact. Of course, had there existed a <code class=\"highlighter-rouge\">Movie<\/code>-fact with title \u201cJennifer\u201d and genre \u201cCrime\u201d, it would still need to star Tim Roth and be directed by Quentin Tarantino to be returned.<\/p><p>Furthermore, whilst the order of your constants in your facts does not matter, you ultimately do need to obey the chosen order when writing rules. That is, <code class=\"highlighter-rouge\">Movie(title, \"Crime\")<\/code> and <code class=\"highlighter-rouge\">Movie(\"Crime\", title)<\/code> are <em>not<\/em> the same! The first would with our set of facts match crime movies and return their titles, whereas the second would exclusively match movies called \u201cCrime\u201d and return their genres.<\/p><p>The result of the above <code class=\"highlighter-rouge\">ExtraAwesomeMovie<\/code>-rule is:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>ExtraAwesomeMovie(\"Pulp Fiction\").\nExtraAwesomeMovie(\"The Hateful Eight\").\n<\/pre><\/div><\/div><p>The above shows the two new facts of type <code class=\"highlighter-rouge\">ExtraAwesomeMovie<\/code> which we have inferred from our original set of facts. Given the preferences of our friend, which we used to design our rule, we can safely select either of the two movies since we know for a fact (pun intended!) that both are <code class=\"highlighter-rouge\">ExtraAwesomeMovie<\/code>\u2019s! Having introduced one example, we will now backtrack slightly and establish proper terminology and syntax.<\/p><h3 id=\"datalog-syntax\">Datalog Syntax<\/h3><p>In Datalog, both facts and rules are represented as Horn clauses\u2014the specifics of which we will not be covering in this post, though if the reader is interested, additional information may be found <a href=\"https:\/\/en.wikipedia.org\/wiki\/Horn_clause\">here<\/a>. It is sufficient to know that a Horn clause is a clause which contain at most one positive literal, and that there exists three types of Horn clauses: (1) facts (or <em>unit clauses<\/em>), expressing unconditional knowledge, (2) rules, expressing conditional knowledge, and (3) goal clauses. We have already introduced the two former, and we will return to the third later.<\/p><p>The complete grammar of Datalog is:<\/p><p>Where<br \/>$p$ = finite set of predicate symbols.<br \/>$k$ = finite set of variable symbols.<br \/>$x$ = finite set of constants.\n<\/p><p>In other words: A Datalog program $P$ consists of a sequence of constraints $c_1, \u2026, c_n$ \u2013 a design which means that the shortest possible Datalog program is the empty sequence of constraints. A constraint $c$ is a Horn clause consisting of a <em>head atom<\/em> $a_0$, and an optional sequence of <em>body atoms<\/em> $a_1, \u2026, a_n$. If the body is empty, we are dealing with a <em>fact<\/em>, otherwise it is a <em>rule<\/em>. An atom $a$ consists of a predicate symbol\u2014that is, a name\u2014and a sequence of terms $t_1, \u2026, t_n$. A term is either a constant (e.g. a string like <code class=\"highlighter-rouge\">\"Crime\"<\/code>, or an integer like <code class=\"highlighter-rouge\">31<\/code>), or a variable (e.g. <code class=\"highlighter-rouge\">title<\/code>).<\/p><p>Or, explained visually:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/datalogsyntax.png\" alt=\"drawing\" width=\"70%\" \/><\/div><p>Additionally, as already have noted, the <code class=\"highlighter-rouge\">.<\/code> (period) at the end of a line concludes either that specific fact or rule. However, if we are writing a rule which contains multiple lines, we need to separate each but the last line using <code class=\"highlighter-rouge\">,<\/code> (comma) as shown in the initial example above.<\/p><blockquote>\n  <p><em>Different sources employ different terminology. Specifically, the concept of a constraint (which I refer to as either being a <code class=\"highlighter-rouge\">fact<\/code> or a <code class=\"highlighter-rouge\">rule<\/code>) is also sometimes referred to as a <code class=\"highlighter-rouge\">clause<\/code>, or just <code class=\"highlighter-rouge\">rule<\/code>. Further, the concept of an atom is also sometimes referred to as a <code class=\"highlighter-rouge\">literal<\/code>. Some papers will mention which terminology they employ if it is not already obvious from the context.<\/em><\/p>\n<\/blockquote><p>This is it though. The above is the entire Datalog syntax.<\/p><p>Moving on!<\/p><h4 id=\"datalog-groundness-and-safety\">Datalog Groundness and Safety<\/h4><p>An atom is said to be <em>ground<\/em> if all its terms are constants, e.g:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>NonGroundedFact(\"Walter White\", actor).\n<\/pre><\/div><\/div><p>is not grounded (due to the <em>actor<\/em>-variable), whereas<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>GroundedFact(\"Walter White\", \"Bryan Cranston\").\n<\/pre><\/div><\/div><p>is grounded because both terms are constants. Similary, a rule is said to be <em>ground<\/em> if all its atoms are ground, for instance:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>NonGroundedRule(\"Walter White\", actor) :-\n    MainCharacter(\"Walter White\", \"Breaking Bad\"),\n    StarringIn(\"Breaking Bad\", actor).\n<\/pre><\/div><\/div><p>is not grounded (due to the two <em>actor<\/em>-variables), whereas<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>GroundedRule(\"Walter White\", \"Bryan Cranston\") :-\n    MainCharacter(\"Walter White\", \"Breaking Bad\"),\n    StarringIn(\"Breaking Bad\", \"Bryan Cranston\").\n<\/pre><\/div><\/div><p>is grounded.<\/p><p>In order to ensure that we produce meaningful results, we introduce the notion of <strong>safety<\/strong>. We consider a given program $P$ <strong>safe<\/strong> iff every fact of $P$ is ground, and every variable $x$ that occurs in the head of a rule also occurs in the body of the rule (e.g. like how the <code class=\"highlighter-rouge\">actor<\/code> variable in the abovementioned <code class=\"highlighter-rouge\">NonGroundRule<\/code> appears both in the head and body). This is called the <em>range restriction property<\/em>.<\/p><blockquote>\n  <p>It is important to accentuate that we do not need to have grounded rules for our Datalog program to be safe; we merely need to ensure what appears in the head also at some point appears in the body at least once!<\/p>\n<\/blockquote><p>These conditions ensure that the set of facts which can be inferred from our Datalog program is finite\u2014something which in turn guarantees termination. This is likely not something which you need to actively think about whilst programming considering how the vast majority of Datalog implementations outright reject unsafe programs, but now you know why rejection happens!<\/p><h4 id=\"databases\">Databases<\/h4><p>We consider the facts in our program our <em>database<\/em>. We employ this database to answer queries\u2014both by reading what is already in our database, and by using the existing knowledge to infer new knowledge liked we did with <code class=\"highlighter-rouge\">ExtraAwesomeMovie<\/code> above. In the context of Datalog, a <em>query<\/em> is really just asking whether something exists one way or the other. We will return to queries in greater detail later.<\/p><p>In Datalog, we specifically operate with two distinct types of databases: (1) the <em>extensional database (EDB)<\/em>, and (2) the <em>intensional database (IDB)<\/em>. The EDB is the set of facts already in our Datalog program $P$, and the IDB is the set of facts we can <em>infer<\/em> from $P$. In other words, we can consider the EDB the <em>input<\/em> of our Datalog program, and the IDB the <em>output<\/em> of our Datalog program.<\/p><p>For instance, consider the program:\n<br \/><\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Edge(\"a\", \"b\").\nEdge(\"b\", \"c\").\nPath(x, y) :-\n    Edge(x, y).\nPath(x, z) :-\n    Path(x, y),\n    Edge(y, z).\n<\/pre><\/div><\/div><p>The above <code class=\"highlighter-rouge\">Path<\/code>-rule states that iff there is a path from $x \\rightarrow y$, and an edge from $y \\rightarrow z$, then there is also an path from $x \\rightarrow  z$. At the moment, our EDB consists of two facts, namely:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Edge(\"a\", \"b\").     extensional\nEdge(\"b\", \"c\").     extensional\n<\/pre><\/div><\/div><p>However, by running the above program, we can infer three additional facts:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Edge(\"a\", \"b\").     extensional\nEdge(\"b\", \"c\").     extensional\nPath(\"a\", \"b\").     intensional\nPath(\"b\", \"c\").     intensional\nPath(\"a\", \"c\").     intensional\n<\/pre><\/div><\/div><p>Note how the output of a rule, in this case <code class=\"highlighter-rouge\">Path<\/code>, consists of new ground atoms\u2014facts\u2014of the same name with terms satisfying the rule. The above should also make it clear why we can consider the EDB our input and the IDB our output. A Datalog program semantically is thus, in essence, an union between the EDB and the IDB.<\/p><blockquote>\n  <p><em>What I refer to as \u201cdatabase\u201d is also sometimes referred to as \u201cschema\u201d. That is, some papers operate with an \u201cextensional schema\u201d and an \u201cintensional schema\u201d. The abbreviations <strong>EDB<\/strong> and <strong>IDB<\/strong>, respectively, remain the same, however.<\/em><\/p>\n<\/blockquote><h3 id=\"datalog-semantics\">Datalog Semantics<\/h3><p>In the introduction I alluded to the magic which transpire when the compiler computes a solution. It is not entirely true that we are completely in the dark when it comes to how the compiler operates. Indeed, there exists three different, equivalent Datalog semantics, namely:<\/p><ol><li>Model-theoretic semantics<\/li>\n  <li>Fixed-point semantics<\/li>\n  <li>Proof-theoretic semantics\n<p>These three semantics collectively hold a key to understanding Datalog programs. That is, by covering these three semantics, we attain the understanding and knowledge necessary to adequately reason about the solution of a Datalog program.\n<\/p><h4 id=\"model-theoretic-semantics\">Model-theoretic semantics<\/h4>\n    <p>Model-theoretic semantics defines the meaning of a Datalog program in terms of <em>interpretations<\/em> and <em>models<\/em>. An interpretation is a set of facts, and a model is a specific interpretation which satisfy all constraints in our Datalog program. The smallest of such models is known as the <em>minimal model<\/em>, which is what we consider the solution. The intuition for model-theoretic semantics is that whilst a solution may be computed using different means, we can easily agree on the meaning of the query (e.g. find <code class=\"highlighter-rouge\">ExtraAwesomeMovie<\/code>) as well as check whether a proposed solution is valid (e.g. if our friend is satisfied with the movie selection).<\/p>\n  <\/li>\n<\/ol><p>To properly understand model-theoretic semantics, we need to introduce some additional definitions and concepts:<\/p><ul><li>Herbrand Universe<\/li>\n  <li>Herbrand Base<\/li>\n  <li>Truth<\/li>\n  <li>Models<\/li>\n  <li>Minimality\n<p>To assist our understanding of the above, consider the program:<\/p><\/li>\n<\/ul><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Sansa Stark\", \"Eddard Stark\").\nParent(\"Arya Stark\", \"Eddard Stark\").\nParent(\"Eddard Stark\", \"Rickard Stark\").\nGrandParent(x, z) :-\n    Parent(x, y),\n    Parent(y, z).\n<\/pre><\/div><\/div><p>The <strong>Herbrand Universe<\/strong>, $U$, is the set of <em>all constants<\/em> appearing <em>anywhere<\/em> in our program. Note that if some constant is introduced in a rule, this, too, will be included in $U$. In the above example,<\/p><p>$U$ = <code class=\"highlighter-rouge\">{\"Sansa Stark, \"Arya Stark\", \"Eddard Stark\", \"Rickard Stark\"}<\/code>.<\/p><p>The <strong>Herbrand Base<\/strong>, $B$, is the set of <em>all ground atoms<\/em> with predicate symbols drawn from your program and terms drawn from the Herbrand Universe $U$. Intuitively, you can think of $B$ as the EDB (the existing facts), the IDB (the facts we can infer), and the un-inferrable facts (meaningless facts) all mixed together.<\/p><p>Consider $B$ of the above example:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Arya Stark\", \"Arya Stark\").         GrandParent(\"Arya Stark\", \"Arya Stark\").\nParent(\"Arya Stark\", \"Sansa Stark\").        GrandParent(\"Arya Stark\", \"Sansa Stark\").\nParent(\"Arya Stark\", \"Eddard Stark\").       GrandParent(\"Arya Stark\", \"Eddard Stark\").\nParent(\"Arya Stark\", \"Rickard Stark\").      GrandParent(\"Arya Stark\", \"Rickard Stark\").\nParent(\"Sansa Stark\", \"Sansa Stark\").       GrandParent(\"Sansa Stark\", \"Sansa Stark\").\nParent(\"Sansa Stark\", \"Arya Stark\").        GrandParent(\"Sansa Stark\", \"Arya Stark\").\nParent(\"Sansa Stark\", \"Eddard Stark\").      GrandParent(\"Sansa Stark\", \"Eddard Stark\").\nParent(\"Sansa Stark\", \"Rickard Stark\").     GrandParent(\"Sansa Stark\", \"Rickard Stark\").\nParent(\"Eddard Stark\", \"Arya Stark\").       GrandParent(\"Eddard Stark\", \"Arya Stark\").\nParent(\"Eddard Stark\", \"Sansa Stark\").      GrandParent(\"Eddard Stark\", \"Sansa Stark\").\nParent(\"Eddard Stark\", \"Eddard Stark\").     GrandParent(\"Eddard Stark\", \"Eddard Stark\").\nParent(\"Eddard Stark\", \"Rickard Stark\").    GrandParent(\"Eddard Stark\", \"Rickard Stark\").\nParent(\"Rickard Stark\", \"Arya Stark\").      GrandParent(\"Rickard Stark\", \"Arya Stark\").\nParent(\"Rickard Stark\", \"Sansa Stark\").     GrandParent(\"Rickard Stark\", \"Sansa Stark\").\nParent(\"Rickard Stark\", \"Eddard Stark\").    GrandParent(\"Rickard Stark\", \"Eddard Stark\").\nParent(\"Rickard Stark\", \"Rickard Stark\").   GrandParent(\"Rickard Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div><p>You will quickly notice how, as already hinted, many of the above atoms make absolutely no sense; e.g. <code class=\"highlighter-rouge\">GrandParent(\"Arya Stark\", \"Arya Stark\").<\/code>. This is why we need to reintroduce the notion of interpretations. As already mentioned, an interpretation is a set of facts. More specifically, however, a given interpretation is a <em>subset<\/em> of the Herbrand Base.<\/p><blockquote>\n  <p><em>An interpretation is also sometimes referred to as a database instance.<\/em><\/p>\n<\/blockquote><p>For example,<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Rickard Stark\", \"Eddard Stark\").    GrandParent(\"Rickard Stark\", \"Eddard Stark\").\nParent(\"Rickard Stark\", \"Rickard Stark\").   GrandParent(\"Rickard Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div><p>is an interpretation. However, as should be evident, not every interpretation is going to be true. This does not mean that we cannot say anything about the truth given an interpretation, though! Remember, all the correct facts are <em>also<\/em> part of our Herbrand Base. Given any interpretation $I$, we can determine the <em>truth<\/em> of a constraint as follows:<\/p><ul><li>A ground atom $a = p(k_1, \u2026, k_n)$ is <strong>true<\/strong> for $I$ if $a \\in I$.<\/li>\n  <li>A conjunction of atoms, $a_1, \u2026, a_n$ is <strong>true<\/strong> for $I$ if each $a$ is true in $I$.<\/li>\n  <li>A ground rule $a_0 \\Leftarrow a_1, \u2026, a_n$ is <strong>true<\/strong> for $I$ if either the body conjunction is <em>false<\/em>, or the head atom is <em>true<\/em>.\n<p>That is, given an interpretation $I$:\n<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Arya Stark\", \"Eddard Stark\").\nParent(\"Eddard Stark\", \"Rickard Stark\").\nGrandParent(\"Arya Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div>\n  <\/li>\n<\/ul><p>The ground atom<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Arya Stark\", \"Eddard Stark\").\n<\/pre><\/div><\/div><p>is <strong>true<\/strong> because it appears in $I$. However, the ground atom<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Sansa Stark\", \"Eddard Stark\").\n<\/pre><\/div><\/div><p>is <strong>false<\/strong> because it does not appear in $I$. Again, this does not mean the fact itself is wrong; we are exclusively considering this in the context of the given interpretation.<\/p><p>The ground rule<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>GrandParent(\"Arya Stark\", \"Rickard Stark\") :-\n    Parent(\"Arya Stark\", \"Eddard Stark\").\n    Parent(\"Eddark Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div><p>is <strong>true<\/strong>. Why? Again, for rules, either the body is false, or, if the body is true, the head has to be true. The first <code class=\"highlighter-rouge\">Parent<\/code> atom (<code class=\"highlighter-rouge\">Parent(\"Arya Stark\", \"Eddard Stark\")<\/code>) is in $I$\u2014as is the second <code class=\"highlighter-rouge\">Parent<\/code> atom (<code class=\"highlighter-rouge\">Parent(\"Eddard Stark\", \"Rickard Stark\")<\/code>). This means that the body is true, which means the head ground atom must be in the interpretation $I$, which it is. In other words, given the above interpretation $I$, the above ground rule is true.<\/p><blockquote>\n  <p><em>Just because an interpretation is true does not mean it is also true in the real world. It is important to accentuate how the notion of truth is wholly reliant on the specific domain which we have modelled and operate in. Incorrect data will, therefore, inherently, lead to incorrect answers\u2014much like how a tainted SQL database will result in incorrect or misleading answers.<\/em><\/p>\n<\/blockquote><p>This leads us to the notion of <strong>models<\/strong>. In short, a model $M$ of a Datalog program $P$ is an interpretation $I$ that makes each ground instance of each rule in $P$ true. A <em>ground instance<\/em> of a rule is obtained by replacing every variable in a rule with a constant from the Herbrand Universe.<\/p><p>In our above example, this would translate to the interpretation:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Arya Stark\", \"Eddard Stark\").\nParent(\"Sansa Stark\", \"Eddard Stark\").\nParent(\"Eddard Stark\", \"Rickard Stark\").\nGrandParent(\"Arya Stark\", \"Rickard Stark\").\nGrandParent(\"Sansa Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div><p>In other words, a model is an interpretation for which when we instantiate all ground rules in our program, they are satisfied by the set of facts in the interpretation.<\/p><p>A model $M$ is <strong>minimal<\/strong> if and only if there is no other model which contains fewer facts. What this means is that <strong>the minimal model is the intersection of all models<\/strong>. A direct consequence hereof is that the minimal model is <em>unique<\/em>. This is what allows any Datalog program to always have a unique solution\u2014the minimal model.<\/p><p><strong>In short<\/strong>, model-theoretic semantics allows us to check a given solution and verify whether it is the minimal model. We first confirm whether all ground rule instances are true, and then, if they are, we know we have a model. We then check to see if that model is minimal by checking whether any facts can be removed.<\/p><p>A natural question, then, is <em>how do we actually find the minimal model?<\/em><\/p><h4 id=\"fixed-point-semantics\">Fixed-point semantics<\/h4><p>Using the above principles of model-theoretic semantics, we develop a method for finding the minimal model. We start by defining the notion of an <em>immediate consequence operator<\/em>, $T_p$, as the head atoms of each ground rule instance satisfied by a given interpretation $I$. That was quite a mouthful, so consider the following interpretation from our <em>Game of Thrones<\/em> inspired example instead:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Parent(\"Arya Stark\", \"Eddard Stark\").\nParent(\"Eddard Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div><p>The above facts satisfy our <code class=\"highlighter-rouge\">GroundParent<\/code>-rule and allows us to ground it:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>GrandParent(\"Arya Stark\", \"Rickard Stark\").\n<\/pre><\/div><\/div><p>We now have a new fact which we can continue working with. $T_p$ can intuitively be thought of as iteratively computing the set of facts which can be inferred in a single step from a given interpretation $I$. Or, in other words:<\/p><p>Where $\u00d8$ is empty set of facts and where $Iteration0$ translates to our EDB\u2014that is, the facts already in your program. Please note how $T_p^{2}(\u00d8) = T_p(T_p(\u00d8))$, which is why it suffices to write $Iteration_{n}$ the way we do above. What really happens above though is that we repeatedly apply $T_p$ to itself starting from the empty set. We continue to do this until we stop inferring new facts. This process is called <em>computing the least-fixed point of $T_p$<\/em>.<\/p><p>Consider this example:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Edge(\"a\", \"b\").\nEdge(\"b\", \"c\").\nEdge(\"c\", \"d\").\nEdge(\"d\", \"e\").\nPath(x, y) :-\n    Edge(x, y).\nPath(x, z) :-\n    Path(x, y),\n    Edge(y, z).\n<\/pre><\/div><\/div><p>$Iteration0$ =<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ facts                iteration\nEdge(\"a\", \"b\").         0\nEdge(\"b\", \"c\").         0\nEdge(\"c\", \"d\").         0\nEdge(\"d\", \"e\").         0\n<\/pre><\/div><\/div><p>Now we have the four edges which appear in the program itself. We cannot use the second <code class=\"highlighter-rouge\">Path<\/code>-rule because that requires both a path and an edge. We can, however, use the first <code class=\"highlighter-rouge\">Path<\/code>-rule, so we will do just that.<\/p><p>$Iteration1$ =<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ facts                iteration\nPath(\"a\", \"b\").         1\nPath(\"b\", \"c\").         1\nPath(\"c\", \"d\").         1\nPath(\"d\", \"e\").         1\n<\/pre><\/div><\/div><p>We now have four edges and four paths. We can now use the second <code class=\"highlighter-rouge\">Path<\/code> rule to infer even more paths.<\/p><p>$Iteration2$ =<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ facts                iteration\nPath(\"a\", \"c\").         2\nPath(\"b\", \"d\").         2\nPath(\"c\", \"e\").         2\n<\/pre><\/div><\/div><p>And again, with the new paths.<\/p><p>$Iteration3$ =<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ facts                iteration\nPath(\"a\", \"d\").         3\nPath(\"b\", \"e\").         3\n<\/pre><\/div><\/div><p>And again.<\/p><p>$Iteration4$ =<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ facts                iteration\nPath(\"a\", \"e\").         4\n<\/pre><\/div><\/div><p>At this point we have exhausted how many new facts we can infer, and we have a final result. Notice that any further application of the $T_p$-function will not alter the result\u2014meaning that we have reached the fixed-point, hence the name. A fixed-point for a function $f$ is a value $x$, such that $f(x) = x$. Additional information about fixed-point arithmetics can be found <a href=\"https:\/\/en.wikipedia.org\/wiki\/Fixed-point_arithmetic\">here<\/a>.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ facts                iteration\nEdge(\"a\", \"b\").         0\nEdge(\"b\", \"c\").         0\nEdge(\"c\", \"d\").         0\nEdge(\"d\", \"e\").         0\nPath(\"a\", \"b\").         1\nPath(\"b\", \"c\").         1\nPath(\"c\", \"d\").         1\nPath(\"d\", \"e\").         1\nPath(\"a\", \"c\").         2\nPath(\"b\", \"d\").         2\nPath(\"c\", \"e\").         2\nPath(\"a\", \"d\").         3\nPath(\"b\", \"e\").         3\nPath(\"a\", \"e\").         4\n<\/pre><\/div><\/div><p>The above is the minimal model of our program.<\/p><p>Using the immediate consequence operator to compute the minimal model of a given Datalog program is called <strong>bottom-up evaluation<\/strong>. This description makes sense when considering how we start from nothing\u2014the empty set of facts\u2014and iteratively move up by computing new facts using the previous ones. The example covered here employs a strategy called <em>na\u00efve evaluation<\/em>, though a better, far more efficient strategy used in practice is called <em>semi-na\u00efve evaluation<\/em>. The essential idea of this approach is to keep track of the inferred facts and only re-evaluate the rules affected by the new facts using only the new facts. We will not be covering the specifics as it is sufficient to know that this approach exists (your Datalog engine most likely uses it!), and what bottom-up evaluation is in general.<\/p><h4 id=\"proof-theoretic-semantics\">Proof-theoretic semantics<\/h4><p>So far, so good. However, what if we only really want to know whether there exists a path from $a \\rightarrow d$? We could naturally compute all paths in the program, like we just did, and check once done. This is fine, but it is not really super optimal either! Proof-theoretic semantics seeks to solve (read: optimise) this. We can now, finally, properly introduce the third type of Horn clauses previously mentioned; <strong>goals<\/strong>.<\/p><blockquote>\n  <p><em>A goal is also sometimes called a \u201cquery\u201d, which makes sense when discussing its practical usage as I have done thus far. However, in the context of this section, I will opt for using the goal-term. You may use them interchangeably, though.<\/em><\/p>\n<\/blockquote><p>A goal is an atom which we want to determine whether exists in the minimal model. A goal can be ground or have free variables\u2014in the case of the latter, we would want to determine the values of those variables before returning. Where we end facts and rules with a <code class=\"highlighter-rouge\">.<\/code> (period), goals are ended with a <code class=\"highlighter-rouge\">?<\/code> (question mark) as though we were posing a question\u2014which, in many ways, is exactly what we are! The output of a goal is both a simple true \/ false answer, but also a <em>proof tree<\/em> which shows how the goal was actually inferred. Because of the proof tree, proof-semantic semantics is called <strong>top-down evaluation<\/strong>.<\/p><p>If we return to our previous example once again:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Edge(\"a\", \"b\").\nEdge(\"b\", \"c\").\nEdge(\"c\", \"d\").\nEdge(\"d\", \"e\").\nPath(x, y) :-\n    Edge(x, y).\nPath(x, z) :-\n    Path(x, y),\n    Edge(y, z).\n<\/pre><\/div><\/div><p>And say we want to know whether there exists a path $a \\rightarrow d$?. In practice, the query would be <code class=\"highlighter-rouge\">Path(\"a\", \"d\")?<\/code>, though this might differ slightly depending on your Datalog engine (some handle queries as filters, for instance). In the above program, however, we have two means of deriving a path; either<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ rule 1\nPath(x, y) :-\n    Edge(x, y).\n<\/pre><\/div><\/div><p>or<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ rule 2\nPath(x, z) :-\n    Path(x, y),\n    Edge(y, z).\n<\/pre><\/div><\/div><p>If we then insert our <code class=\"highlighter-rouge\">\"a\"<\/code> and <code class=\"highlighter-rouge\">\"d\"<\/code>, we get<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ rule 1\nPath(\"a\", \"d\") :-\n    Edge(\"a\", \"d\").\n<\/pre><\/div><\/div><p>and<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ rule 2\nPath(\"a\", \"d\") :-\n    Path(\"a\", y),\n    Edge(y, \"d\").\n<\/pre><\/div><\/div><p>The first rule can obviously not be used; we have no <code class=\"highlighter-rouge\">Edge(\"a\", \"d\").<\/code>. However, the second rule can be used. Indeed, since the second rule initially contains three variables, we now effectively have two <strong>subgoals<\/strong>, namely:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Path(\"a\", y)?\nEdge(y, \"d\")?\n<\/pre><\/div><\/div><p>Notice how the above two subgoals are taken directly from the body of the second rule. Of these, the second subgoal, <code class=\"highlighter-rouge\">Edge(y, \"d\")?<\/code>, is quickly solved by checking our EDB: We have an <code class=\"highlighter-rouge\">Edge(y, \"d\").<\/code> if <code class=\"highlighter-rouge\">y = \"c\"<\/code>. This leaves just one subgoal. However, since we just solved <code class=\"highlighter-rouge\">Edge(y, \"d\")?<\/code> as <code class=\"highlighter-rouge\">Edge(\"c\", \"d\").<\/code>, we need to replace <code class=\"highlighter-rouge\">y<\/code> with <code class=\"highlighter-rouge\">\"c\"<\/code> in the remaining subgoal as well because this, too, relies on <code class=\"highlighter-rouge\">y<\/code>. We thus recurse once again, seeking this time to derive <code class=\"highlighter-rouge\">Path(\"a\", \"c\")?<\/code>:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ rule 1\nPath(\"a\", \"c\") :-\n    Edge(\"a\", \"c\").\n<\/pre><\/div><\/div><p>The first rule is, once again, useless, so we immediately turn our attention to the second rule:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ rule 2\nPath(\"a\", \"c\") :-\n    Path(\"a\", y),\n    Edge(y, \"c\").\n<\/pre><\/div><\/div><p>\u2026 and so on, so forth.<\/p><p>The ultimate output, beyond a true \/ false to our initial goal or query, is a proof-tree as illustrated below:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/prooftree.png\" alt=\"drawing\" width=\"70%\" \/><\/div><h3 id=\"datalog-with-negation\">Datalog with Negation<\/h3><p>We have now established how to determine whether a specific path exists. However, what if what we are really interested in finding is the set of pairs for which <em>there is <strong>no<\/strong> path<\/em> from $x \\rightarrow y$? Introducing <strong>negation<\/strong>! If you recall the very first illustration, we are now moving to the third innermost circle, just below Prolog.<\/p><p>As an example, let us update our previous path program:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Vertex(\"a\").\nVertex(\"b\").\nVertex(\"c\").\nVertex(\"d\").\nVertex(\"e\").\nEdge(\"a\", \"b\").\nEdge(\"b\", \"c\").\nEdge(\"c\", \"d\").\nPath(x, y) :-\n    Edge(x, y).\nPath(x, z) :-\n    Path(x, y),\n    Edge(y, z).\nUnconnected(x, y) :-\n    Vertex (x),\n    Vertex (y),\n    not Path(x, y).\n<\/pre><\/div><\/div><p>The above <code class=\"highlighter-rouge\">Unconnected<\/code> rule reads <em>iff there is a vertex <strong>x<\/strong>, and a vertex <strong>y<\/strong>, and there is no path from <strong>x $\\rightarrow$ y<\/strong><\/em>. However, before we proceed, we need to update our grammar to properly accommodate negation. It is a minor change as we really only need to distinguish between head and body atoms, as well as allow atoms in the body to be negated.<\/p><p>The complete grammar of Datalog with negation is:<\/p><p>Where <br \/>$p$ = finite set of predicate symbols.<br \/>$k$ = finite set of variable symbols.<br \/>$x$ = finite set of constants.\n<\/p><p>This change also means we need to update our notion of <em>truth<\/em> with regards to a given interpretation. Given an interpretation $I$, we now determine the <em>truth<\/em> of a constraint as follows:<\/p><ul><li>A <strong>positive<\/strong> ground atom $b$ is <strong>true<\/strong> for $I$ if $b \\in I$.<\/li>\n  <li>A <strong>negative<\/strong> ground atom $b$ is <strong>true<\/strong> for $I$ if $b \\notin I$.<\/li>\n  <li>A conjunction of atoms, $b_1, \u2026, b_n$ is <strong>true<\/strong> for $I$ if each $b$ is true in $I$.<\/li>\n  <li>A ground rule $b_0 \\Leftarrow a_1, \u2026, b_n$ is <strong>true<\/strong> for $I$ if either the body conjunction is <em>false<\/em>, or the head atom is <em>true<\/em>.<\/li>\n<\/ul><p><br \/>Note how the only change is distinguishing between positive and negative atoms.<\/p><h4 id=\"problems-with-negation\">Problems with negation<\/h4><p>Nothing is ever simple.<\/p><p>Consider the program:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Actor(x) :- not Actress(x).\nActress(x) :- not Actor(x).\n<\/pre><\/div><\/div><p>Assume that the Herbrand Universe, $U$, is the constant <code class=\"highlighter-rouge\">\"Alex\"<\/code>. In such a scenario, the above program has <em>two<\/em> \u201cminimal\u201d models, namely: $M_1 = {Actor(\u201cAlex\u201d)}$ and $M_2 = {Actress(\u201cAlex\u201d)}$. If we look at $M_1$ first, and replace the variable <code class=\"highlighter-rouge\">x<\/code> with a constant from $U$, we have as our first rule:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Actor(\"Alex\") :- not Actress(\"Alex\").\n<\/pre><\/div><\/div><p>Since <code class=\"highlighter-rouge\">Actress(\"Alex\").<\/code> does not exist in $M_1$, the body atom is true. This, as per the rules regarding truth above, must mean that the head atom is true\u2014that is, <code class=\"highlighter-rouge\">Actor(\"Alex\").<\/code> must be true.<\/p><p>If we then look at the second rule:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Actress(\"Alex\") :- not Actor(\"Alex\").\n<\/pre><\/div><\/div><p>Since <code class=\"highlighter-rouge\">Actor(\"Alex\").<\/code> <em>is<\/em> in $M_1$, the body of the <code class=\"highlighter-rouge\">Actress<\/code>-rule is false. However, since a false body does not say anything about the head, we effectively conclude that $M_1$ is a valid model. Following this logic, we can also conclude that $M_2$ is a valid model. This is a problem because neither of the two is the minimal model\u2014there can, by definition, only be one minimal model!<\/p><h4 id=\"stratified-negation\">Stratified negation<\/h4><p>If we <em>disallow<\/em> recursion through negation, we can circumvent the abovementioned issues whilst still accommodating negation. In essence, the idea is to separate our Datalog program $P$ into a sequence of smaller, more confined programs $P_{1}, \u2026, P_{n}$, and then compute them one by one as such:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/stratifieddatalog.png\" alt=\"drawing\" width=\"70%\" \/><\/div><p>The arrow symbolises how the computed facts\u2014the IDB $\\cup$ EDB\u2014of a given subprogram $p_{i}$ becomes the facts\u2014the EDB\u2014of the next subprogram $p_{i+1}$. The solution will then be the facts computed by the final subprogram, $p_{n}$.<\/p><p>A <em>stratification<\/em> of a Datalog program $P$ is a <em>partition<\/em> of the predicate symbols subject to the following two conditions: If there is a rule<\/p><p>(1) $A(\u2026) \\Leftarrow \u2026, B(\u2026), \u2026$ in $P$, where the predicate $A$ is in stratum $p_{i}$, and the predicate $B$ is in stratum $p_{j}$, then $i \\geq j$.<\/p><p>(2) $A(\u2026) \\Leftarrow \u2026, not \\enspace B(\u2026), \u2026$ in $P$ where the predicate $A$ is in stratum $p_{i}$, and the predicate $B$ is in stratum $p_{j}$, then $i &gt; j$.<\/p><p>If we look at the second rule, what this really says is that in order to compute $A$, we must know what $B$ is. In other words: $B$ must belong to a lower strata such that it can be completely computed before we attempt to compute $A$. It is in this regard worth noting that it is possible to write programs which do not obey the above two rules. These programs need to be rejected, and most Datalog engines will do so accordingly. In conclusion though, a given Datalog program $P$ is said to be <em>stratifiable<\/em> if it has a stratification (and thus obeys the above two rules). Once we have a stratified Datalog program, it must satisfy an additional safety property, namely: <em>Every variable which occurs in a negative atom must also occur in a positive atom in the body.<\/em><\/p><p>That is, the following<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Unconnected(x, y) :- \n    not Path(x, y).\n<\/pre><\/div><\/div><p>is considered <em>unsafe<\/em> because <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> are not positively bound, whereas<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Unconnected(x, y) :-\n    Vertex(x),\n    Vertex(y),\n    not Path(x, y).\n<\/pre><\/div><\/div><p>is <em>safe<\/em> because the two negated variables appear in positive atoms in the body. This means that if we look at our Datalog program once again:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Path(x, y) :-\n    Edge(x, y).\nPath(x, z) :-\n    Path(x, y),\n    Edge(y, z).\nUnconnected(x, y) :-\n    Vertex (x),\n    Vertex (y),\n    not Path(x, y).\n<\/pre><\/div><\/div><p>We can stratify it by partitioning it as such:<\/p><p>$p_{0}$ = {$Edge, Path, Vertex$}<br \/>$p_{1}$ = {$Unconnected$}<\/p><p>The reason why we partition as above\u2014and in the process ensure we can compute a minimal model\u2014is that $Unconnected$ is the only rule in which negation is employed. It is, moreover, evident that we need to know what $Path$ is before we can say anything about what $Path$ is not. This is fortunately possible; as is saying something about $Edge$ and $Vertex$. As already alluded to, however, we have no guarantee that a given Datalog program is actually stratifiable\u2026 so how we do determine whether that is the case?<\/p><h4 id=\"precedence-graphs\">Precedence graphs<\/h4><p>Given a Datalog program $P$, we define the <em>precedence graph<\/em> $g$ as follows. If there is a rule:<\/p><ul><li>$A(\u2026) \\Leftarrow \u2026, B(\u2026), \u2026$, then there is an edge $B \\xrightarrow{} A$.<\/li>\n  <li>$A(\u2026) \\Leftarrow \u2026, not \\enspace B(\u2026), \u2026$, then there is an edge $B \\xrightarrow{-} A$.\n<\/li>\n<\/ul><p>A Datalog program is stratifiable iff its precendece graph contains <em>no cycles<\/em> labelled with $-$. This means that even if we had a massive precedence graph with thousands of edges forming a cycle, if just one of the edges had a $-$ label, the Datalog program would not be stratifiable. In most cases, it will be possible to rewrite the program in such a manner that the meaning and inteded result are preserved.<\/p><p>Consider the program:<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Husband(x) :-\n    Man(x),\n    Married(x).\nBachelor(x) :-\n    Man(x),\n    not Husband(x).\n<\/pre><\/div><\/div><p>A <code class=\"highlighter-rouge\">Man<\/code> is a <code class=\"highlighter-rouge\">Bachelor<\/code> if he is not a <code class=\"highlighter-rouge\">Husband<\/code>, and a <code class=\"highlighter-rouge\">Man<\/code> is a <code class=\"highlighter-rouge\">Husband<\/code> if he is <code class=\"highlighter-rouge\">Married<\/code>. If we draw its precedence graph, we get:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/precedencegraph1.png\" alt=\"drawing\" height=\"300px\" width=\"70%\" \/><\/div><p>This seems fine; no cycles! If, on the other hand, our program looked like<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Husband(x) :-\n    Man(x),\n    not Bachelor(x).\nBachelor(x) :-\n    Man(x),\n    not Husband(x)\n<\/pre><\/div><\/div><p>and we now consider a <code class=\"highlighter-rouge\">Husband<\/code> a <code class=\"highlighter-rouge\">Man<\/code> who is not a <code class=\"highlighter-rouge\">Bachelor<\/code>, and a <code class=\"highlighter-rouge\">Bachelor<\/code> a <code class=\"highlighter-rouge\">Man<\/code> who is not a <code class=\"highlighter-rouge\">Husband<\/code>, our precedence graph would look as follows:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/precedencegraph2.png\" alt=\"drawing\" height=\"300px\" width=\"70%\" \/><\/div><p>Observe how we now have a cycle with a negated edge. This version of our program is, consequently, not stratifiable. It is important to accentuate that there need not be a $-$ on <em>every<\/em> edge in the cycle for the program to not be stratifiable. Indeed, the following precedence graph would be equally problematic:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/precedencegraph3.png\" alt=\"drawing\" height=\"300px\" width=\"70%\" \/><\/div><p>With a proper precedence graph $g$ in hand, we can compute the strata by first computing the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Strongly_connected_component\">strongly connected components<\/a> of $g$ and then conducting a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Topological_sorting\">topological sort<\/a> of these components. The order of the strata for the above program, for instance, would be:<\/p><div><img src=\"https:\/\/x775.net\/assets\/images\/datalog\/strata.png\" alt=\"drawing\" width=\"70%\" \/><\/div><p>Indeed, since <code class=\"highlighter-rouge\">Married<\/code> and <code class=\"highlighter-rouge\">Man<\/code> are independent, we can start with those two. However, because the first rule states $i \\geq j$, we can also include <code class=\"highlighter-rouge\">Husband<\/code> in this partition. That is, $p_{0}$ = {$Married, Man, Husband$}. Unlike every other node, <code class=\"highlighter-rouge\">Bachelor<\/code> depends on what <code class=\"highlighter-rouge\">Husband<\/code> <em>is not<\/em>. We thus need to know what <code class=\"highlighter-rouge\">Husband<\/code> is <em>first<\/em> before we can say what it is not\u2014note how the second rule states $i &gt; j$. This makes $p_{2}$ = {$Bachelor$}. Each partition corresponds to a different strata.<\/p><h3 id=\"prolog\">Prolog<\/h3><p>Prolog is like Datalog, but features two, seemingly simple extensions; (1) a fact may contain variables, and (2) a term may be a function application. However simple these changes might appear, the impact of the above two extensions is profound. Prolog is a Turing-complete language, but it lacks some of Datalog\u2019s neat properties, namely:<\/p><ol><li>No termination guarantee<\/li>\n  <li>No (finite) minimal model<\/li>\n  <li>No fixed-point semantics \n<p>This means that where we in Datalog deterministically can enumerate all facts with a bottom-up evaluation, or search for a specific fact using top-down evaluation, we are limited to searching with top-down evaluation without guarantees of termination in Prolog. However, on the other hand, Prolog now offers all the expressive power we need to express any given algorithm! There are many strategies one can employ whilst programming in Prolog to avoid most pitfalls, though unlike Datalog, these now need to be actively accounted for.<\/p><\/li>\n<\/ol><p><em>Thanks to <a href=\"https:\/\/github.com\/bargsteen\">@Bargsteen<\/a> for providing feedback and encouraging continued work!<\/em><\/p><\/section><\/header>","created_at":"2019-03-18T17:35:13+0000","updated_at":"2019-04-19T12:59:28+0000","starred_at":"2019-04-19T12:59:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":33,"domain_name":"x775.net","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["procgen","generative","music"],"is_public":false,"id":878,"title":"generative-music\/generative.fm","url":"https:\/\/github.com\/generative-music\/generative.fm","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>A platform for playing <a href=\"https:\/\/medium.com\/@metalex9\/introduction-to-generative-music-91e00e4dba11\" rel=\"nofollow\">generative music<\/a> in the browser.<\/p>\n<p><a href=\"https:\/\/generative.fm\" rel=\"nofollow\">https:\/\/generative.fm<\/a><\/p>\n<h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2>\n<h3><a id=\"user-content-audio\" class=\"anchor\" aria-hidden=\"true\" href=\"#audio\"><\/a>Audio<\/h3>\n<p>This site uses <a href=\"https:\/\/tonejs.github.io\/\" rel=\"nofollow\">Tone.js<\/a> to play audio in the browser through the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Web_Audio_API\" rel=\"nofollow\">Web Audio API<\/a>. Most of the samples are provided by the <a href=\"https:\/\/vis.versilstudios.com\/vsco-community.html\" rel=\"nofollow\">Versilian Studios Chamber Orchestra 2 Community Edition<\/a>. Other samples provided by the <a href=\"https:\/\/github.com\/peastman\/sso\">Sonatina Symphonic Orchestra<\/a>. The <a href=\"https:\/\/kasper.bandcamp.com\/album\/singing-bowls\" rel=\"nofollow\">singing bowl samples<\/a> were provided by <a href=\"https:\/\/kasper.bandcamp.com\/\" rel=\"nofollow\">Kasper<\/a>. All other samples were specially recorded for the site.<\/p>\n<h3><a id=\"user-content-music\" class=\"anchor\" aria-hidden=\"true\" href=\"#music\"><\/a>Music<\/h3>\n<p>Source code for the music generators can be found in the <a href=\"https:\/\/github.com\/generative-music\/pieces-alex-bainter\">generative-music\/pieces-alex-bainter<\/a> repository.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>You will need <a href=\"https:\/\/nodejs.org\/en\/\" rel=\"nofollow\">Node.js<\/a> installed on your machine.<\/p>\n<ol><li>Clone to your machine.<\/li>\n<li>Navigate to the project's directory.<\/li>\n<li>Run <code>npm i<\/code>.<\/li>\n<li>Run <code>npm start<\/code>.<\/li>\n<li>Navigate to <a href=\"http:\/\/localhost:9999\" rel=\"nofollow\">localhost:9999<\/a>.<\/li>\n<\/ol><h2><a id=\"user-content-issues-and-feedback\" class=\"anchor\" aria-hidden=\"true\" href=\"#issues-and-feedback\"><\/a>Issues and Feedback<\/h2>\n<p>Problems or general feedback can be reported <a href=\"https:\/\/github.com\/generative-music\/site\/issues\">here on Github<\/a> or by sending an email to <a href=\"mailto:alex@alexbainter.com\">alex@alexbainter.com<\/a>. Thanks!<\/p>\n<\/article>","created_at":"2019-03-18T12:42:46+0000","updated_at":"2019-03-18T12:42:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/42454998?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":877,"title":"Vim anti-patterns | Arabesque","url":"https:\/\/sanctum.geek.nz\/arabesque\/vim-anti-patterns\/","content":"<p>The benefits of getting to grips with Vim are immense in terms of editing speed\nand maintaining your \u201cflow\u201d when you\u2019re on a roll, whether writing code,\npoetry, or prose, but because the learning curve is so steep for a text editor,\nit\u2019s very easy to retain habits from your time learning the editor that stick\nwith you well into mastery. Because Vim makes you so fast and fluent, it\u2019s\nespecially hard to root these out because you might not even notice them, but\nit\u2019s worth it. Here I\u2019ll list some of the more common ones.<\/p>\n\n<h2>Moving one line at a time<\/h2>\n\n<p>If you have to move more than a couple of lines, moving one line at a time by\nholding down <code>j<\/code> or <code>k<\/code> is inefficient. There are many more <a href=\"http:\/\/www.benorenstein.com\/blog\/jumping-long-distances-in-vim\">ways to move\nvertically<\/a> in Vim. I find that the two most useful are moving by paragraph\nand by screenful, but this depends on how far and how precisely you have to\nmove.<\/p>\n\n<ul><li><code>{<\/code> \u2014 Move to start of previous paragraph or code block.<\/li>\n<li><code>}<\/code> \u2014 Move to end of next paragraph or code block.<\/li>\n<li><code>Ctrl+F<\/code> \u2014 Move forward one screenful.<\/li>\n<li><code>Ctrl+B<\/code> \u2014 Move backward one screenful.<\/li>\n<\/ul><p>If you happen to know precisely where you want to go, navigating by searching\nis the way to go, searching forward with <code>\/<\/code> and backward with <code>?<\/code>.<\/p>\n\n<p>It\u2019s always useful to jump <em>back<\/em> to where you were, as well, which is easily\nenough done with two backticks, or <code>gi<\/code> to go to the last place you inserted\ntext. If you like, you can even go back and forth through your entire change\nlist of positions with <code>g;<\/code> and <code>g,<\/code>.<\/p>\n\n<h2>Moving one character at a time<\/h2>\n\n<p>Similarly, moving one character at a time with <code>h<\/code> and <code>l<\/code> is often a waste\nwhen you have <code>t<\/code> and <code>f<\/code>:<\/p>\n\n<ul><li><code>t&lt;char&gt;<\/code> \u2014 Move <em>forward until<\/em> the next occurrence of the\u00a0character.<\/li>\n<li><code>f&lt;char&gt;<\/code> \u2014 Move <em>forward over<\/em> the next occurrence of the character.<\/li>\n<li><code>T&lt;char&gt;<\/code> \u2014 Move <em>backward until<\/em> the previous occurrence of the\ncharacter.<\/li>\n<li><code>F&lt;char&gt;<\/code> \u2014 Move <em>backward over<\/em> the previous occurrence of the character.<\/li>\n<\/ul><p>Moving wordwise with <code>w<\/code>, <code>W<\/code>, <code>b<\/code>, <code>B<\/code>, <code>e<\/code>, and <code>E<\/code> is better, too. Again,\nsearching to navigate is good here, and don\u2019t forget you can <em>yank<\/em>,\u00a0<em>delete<\/em>\nor <em>change<\/em> forward or backward to a search result:<\/p>\n\n<pre><code>y\/search&lt;Enter&gt;\ny?search&lt;Enter&gt;\nd\/search&lt;Enter&gt;\nd?search&lt;Enter&gt;\nc\/search&lt;Enter&gt;\nc?search&lt;Enter&gt;\n<\/code><\/pre>\n\n<h2>Searching for the word under the cursor<\/h2>\n\n<p>Don\u2019t bother typing it, or yanking\/pasting it; just use <code>*<\/code> or <code>#<\/code>. It\u2019s\ndizzying how much faster this feels when you use it enough for it to become\nautomatic.<\/p>\n\n<h2>Deleting, then inserting<\/h2>\n\n<p>Deleting text with intent to replace it by entering insert mode immediately\nafterward isn\u2019t necessary:<\/p>\n\n<pre><code>d2wi\n<\/code><\/pre>\n\n<p>It\u2019s quicker and tidier to use <code>c<\/code> for change:<\/p>\n\n<pre><code>c2w\n<\/code><\/pre>\n\n<p>This has the added benefit of making the entire operation repeatable with the\n<code>.<\/code> command.<\/p>\n\n<h2>Using the arrow keys<\/h2>\n\n<p>Vim lets you use the arrow keys to move around in both insert and normal mode,\nbut once you\u2019re used to using <code>hjkl<\/code> to navigate, moving to the arrow keys to\nmove around in text feels clumsy; you should be able to spend the vast majority\nof a Vim session with your hands firmly centered around <a href=\"https:\/\/en.wikipedia.org\/wiki\/Touch_typing\">home row<\/a>.\nSimilarly, while the Home and End keys work the same way they do in most\neditors, there\u2019s no particular reason to use them when functional equivalents\nare closer to home in <code>^<\/code> and <code>$<\/code>.<\/p>\n\n<p>So wean yourself off the arrow keys, by the simple expedient of disabling them\nentirely, at least temporarily:<\/p>\n\n<pre><code>noremap &lt;Up&gt; &lt;nop&gt;\nnoremap &lt;Down&gt; &lt;nop&gt;\nnoremap &lt;Left&gt; &lt;nop&gt;\nnoremap &lt;Right&gt; &lt;nop&gt;\n<\/code><\/pre>\n\n<p>The benefits of sticking to home row aren\u2019t simply in speed; it <em>feels<\/em> nicer\nto be able to rest your wrists in front of the keyboard and not have to move\nthem too far, and for some people it has even <a href=\"http:\/\/www.reddit.com\/r\/vim\/comments\/mgt9q\/vim_can_even_fix_my_body\/\">helped prevent repetitive strain\ninjury<\/a>.<\/p>\n\n<h2>Moving in insert mode<\/h2>\n\n<p>There\u2019s an additional benefit to the above in that it will ease you into\nthinking less about insert mode as a mode in which you move around; that\u2019s what\nnormal mode is for. You should, in general, spend as little time in insert mode\nas possible. When you want to move, you\u2019ll get in the habit of leaving insert\nmode, and moving around far more efficiently in normal mode instead. This\ndistinction also helps to keep your insert operations more <em>atomic<\/em>, and hence\nmore useful to repeat.<\/p>\n\n<h2>Moving to Escape<\/h2>\n\n<p>The Escape key on modern keyboards is a lot further from home row than it was\non <a href=\"https:\/\/en.wikipedia.org\/wiki\/File:KB_Terminal_ADM3A.svg\">Bill Joy\u2019s keyboard<\/a> back when he designed vi. Hitting Escape is usually\nunnecessary; Ctrl+[ is a lot closer, and more comfortable. It doesn\u2019t take\nlong using this combination instead to make reaching for Escape as you did when\nyou were a newbie feel very awkward. You might also consider mapping the\notherwise pretty useless Caps Lock key to be another Escape key in your\noperating system, or even mapping uncommon key combinations like <code>jj<\/code> to\nEscape. I feel this is a bit drastic, but it works well for a lot of people:<\/p>\n\n<pre><code>inoremap jj &lt;Esc&gt;\n<\/code><\/pre>\n\n<h2>Moving to the start or end of the line, then inserting<\/h2>\n\n<p>Just use <code>I<\/code> and <code>A<\/code>. Again, these make the action repeatable for other lines\nwhich might need the same operation.<\/p>\n\n<h2>Entering insert mode, then opening a new line<\/h2>\n\n<p>Just use <code>o<\/code> and <code>O<\/code> to open a new line below and above respectively, and enter\ninsert mode on it at the same time.<\/p>\n\n<h2>Entering insert mode to delete text<\/h2>\n\n<p>This is a pretty obvious contradiction. Instead, delete the text by moving to\nit and using <code>d<\/code> with an appropriate motion or text object. Again, this is\nrepeatable, and means you\u2019re not holding down Backspace. In general, if you\u2019re\nholding down a key in Vim, there\u2019s probably a faster way.<\/p>\n\n<h2>Repeating commands or searches<\/h2>\n\n<p>Just type <code>@:<\/code> for commands or <code>n<\/code>\/<code>N<\/code> for searches; Vim doesn\u2019t forget what\nyour last search was as soon as you stop flicking through results. If it wasn\u2019t\nyour most recent command or search but it\u2019s definitely in your history, just\ntype <code>q:<\/code> or <code>q\/<\/code>, <a href=\"https:\/\/sanctum.geek.nz\/arabesque\/vim-command-window\/\">find it in the list<\/a>, and hit Enter.<\/p>\n\n<h2>Repeating substitutions<\/h2>\n\n<p>Just type <code>&amp;<\/code> to repeat the last substitution on the current line. You can\nrepeat it on <em>all<\/em> lines by typing <code>g&amp;<\/code>.<\/p>\n\n<h2>Repeating macro calls<\/h2>\n\n<p>Just type <code>@@<\/code>.<\/p>\n\n<p>These are really only just a few of the common traps to avoid to increase your\nspeed and general efficiency with the editor without requiring plugins or\nsubstantial remappings. Check out the <a href=\"http:\/\/vim.wikia.com\/wiki\/Vim_Tips_Wiki\">Vim Tips wiki<\/a> for some other really\nhelpful examples.<\/p>\n<div class=\"related-posts\">\n<div id=\"related-posts-MRP_all\" class=\"related-posts-type\">\n<h2>Related Posts<\/h2>\n<ul><li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/advanced-vim-macros\/\">Advanced Vim macros<\/a><\/li>\n<li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/advanced-vim-registers\/\">Advanced Vim registers<\/a><\/li>\n<li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/avoiding-the-arrow-keys\/\">Avoiding the arrow keys<\/a><\/li>\n<li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/elegant-awk-usage\/\">Elegant Awk usage<\/a><\/li>\n<li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/faster-vim-search-and-replace\/\">Faster Vim search\/replace<\/a><\/li>\n<li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/unix-as-ide-editing\/\">Unix as IDE: Editing<\/a><\/li>\n<li><a href=\"https:\/\/sanctum.geek.nz\/arabesque\/vim-misconceptions\/\">Vim misconceptions<\/a><\/li>\n<\/ul><\/div>\n<\/div>","created_at":"2019-03-17T21:11:53+0000","updated_at":"2019-03-17T21:11:53+0000","published_by":["Tom Ryder"],"annotations":[],"mimetype":"text\/html","language":"en_NZ","reading_time":5,"domain_name":"sanctum.geek.nz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":876,"title":"Write yourself a Git!","url":"https:\/\/wyag.thb.lt\/","content":"<p>\nNow that we have repositories, putting things inside them is in order.\nAlso, repositories are boring, and writing a Git implementation\nshouldn\u2019t be just a matter of writing a bunch of <code>mkdir<\/code>.  Let\u2019s talk\nabout <b>objects<\/b>, and let\u2019s implement <code>git hash-object<\/code> and <code>git cat-file<\/code>.\n<\/p><p>\nMaybe you don\u2019t know these two commands \u2014 they\u2019re not exactly part\nof an everyday git toolbox, and they\u2019re actually quite low-level\n(\u201cplumbing\u201d, in git parlance).  What they do is actually very simple:\n<code>hash-object<\/code> converts an existing file into a git object, and <code>cat-file<\/code>\nprints an existing git object to the standard output.\n<\/p><p>\nNow, <b>what actually is a Git object?<\/b>  At its core, Git is a\n\u201ccontent-addressed filesystem\u201d.  That means that unlike regular\nfilesystems, where the name of a file is arbitrary and unrelated to\nthat file\u2019s contents, the names of files as stored by Git are\nmathematically derived from their contents.  This has a very important\nimplication: if a single byte of, say, a text file, changes, its\ninternal name will change, too.  To put it simply: you don\u2019t <i>modify<\/i> a\nfile, you create a new file in a different location.  Objects are just\nthat: files in the git repository, whose path is determined by their\ncontents.\n<\/p><div class=\"warning\"><p>\n<b>Git is not (really) a key-value store<\/b>\n<\/p><p>\nSome documentation, including the excellent <a href=\"https:\/\/git-scm.com\/book\/id\/v2\/Git-Internals-Git-Objects\">Pro Git<\/a>, call Git a\n\u201ckey-value store\u201d.  This is not incorrect, but may be misleading.\nRegular filesystems are actually closer to a key-value store than Git\nis.  Because it computes keys from data, Git should rather be called a\n<i>value-value store<\/i>.\n<\/p><\/div><p>\nGit uses objects to store quite a lot of things: first and foremost,\nthe actual files it keeps in version control \u2014 source code, for\nexample.  Commit are objects, too, as well as tags.  With a few\nnotable exceptions (which we\u2019ll see later!), almost everything, in\nGit, is stored as an object.\n<\/p><p>\nThe path is computed by calculating the <a href=\"https:\/\/en.wikipedia.org\/wiki\/SHA-1\">SHA-1<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cryptographic_hash_function\">hash<\/a> of its contents.\nMore precisely, Git renders the hash as a lowercase hexadecimal\nstring, and splits it in two parts: the first two characters, and the\nrest.  It uses the first part as a directory name, the rest as the\nfile name (this is because most filesystems hate having too many files\nin a single directory and would slow down to a crawl.  Git\u2019s method\ncreates 256 possible intermediate directories, hence dividing the\naverage number of files per directory by 256)\n<\/p><div class=\"note\"><p>\n<b>What is a hash function?<\/b>\n<\/p><p>\nSimply put, a hash function is a kind of unidirectional mathematical\nfunction: it is easy to compute the hash of a value, but there\u2019s no\nway to compute which value produced a hash.  A very simple example of\na hash function is the <code>strlen<\/code> function.  It\u2019s really easy to compute\nthe length of a string, and the length of a given string will never\nchange (unless the string itself changes, of course!) but it\u2019s\nimpossible to retrieve the original string, given only its length.\n<i>Cryptographic<\/i> hash functions are just a much more complex version of\nthe same, with the added property that computing an input meant to\nproduce a given hash is hard enough to be\npractically impossible.  (With <code>strlen<\/code>, producing an input <code>i<\/code> with\n<code>strlen(i) == 12<\/code>, you just have to type twelve random characters.  With\nalgorithms such as SHA-1. it would take much, much longer \u2014 long\nenough to be practically impossible<sup><a id=\"fnr.1\" class=\"footref\" href=\"#fn.1\">1<\/a><\/sup>.\n<\/p><\/div><p>\nBefore we start implementing the object storage system, we must\nunderstand their exact storage format.  A object start by an header\nthat specify its type: <code>blob<\/code>, <code>commit<\/code>, <code>tag<\/code> or <code>tree<\/code>.  This header is\nfollowed by an ASCII space (0x20), then the size of the object in bytes as an\nASCII number, then null (0x00) (the null byte), then the contents of the\nobject.  The first 48 bytes of a commit object in Wyag\u2019s repo look like this:\n<\/p><pre class=\"example\">\n00000000  63 6f 6d 6d 69 74 20 31  30 38 36 00 74 72 65 65  |commit 1086.tree|\n00000010  20 32 39 66 66 31 36 63  39 63 31 34 65 32 36 35  | 29ff16c9c14e265|\n00000020  32 62 32 32 66 38 62 37  38 62 62 30 38 61 35 61  |2b22f8b78bb08a5a|\n<\/pre><p>\nIn the first line, we see the type header, a space (<code>0x20<\/code>), the size in\nASCII (1086) and the null separator <code>0x00<\/code>.  The last four bytes on the\nfirst line are the beginning of that object\u2019s contents, the word\n\u201ctree\u201d \u2014 we\u2019ll discuss that further when we\u2019ll talk about commits.\n<\/p><p>\nThe objects (headers and contents) are stored compressed with <code>zlib<\/code>.\n<\/p>","created_at":"2019-03-14T11:57:33+0000","updated_at":"2019-03-14T11:57:33+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"wyag.thb.lt","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":875,"title":"Datopia","url":"https:\/\/datopia.io\/","content":"<div class=\"row\"><div class=\"col-lg-12 text-center\"><img src=\"https:\/\/datopia.io\/images\/datopia.png\" class=\"logo\" alt=\"image\" \/><\/div><\/div><div class=\"row\"><div class=\"col-lg-8 offset-lg-2 pitch-cont\"><p>\n      On-chain governance, voting analytics, supply chain management \u2014\n      processes many suspect will yield to blockchains \u2014 consume and translate\n      <i>vast<\/i> quantities of information in all but the simplest cases.\n      Platforms encumbered by astronomical storage costs \u2014 or\n      inexpressive data modelling primitives \u2014 are unlikely to\n      provide us with sustainable, cost-effective or <i>transparent<\/i> solutions\n      to data-hungry problems.  Our inaugural post,\n      <a href=\"https:\/\/blog.datopia.io\/2018\/11\/03\/blockchain-information-system\/\">Blockchains\n      as Information Systems<\/a>, covers these ideas in more detail.\n    <\/p><img src=\"https:\/\/datopia.io\/images\/divider.svg\" class=\"divider\" alt=\"image\" \/><p>\n      At Datopia, we're building an information system influenced by commercial\n      and academic work on immutable storage: a blockchain\n      database capable of trustlessly and inferentially interrogating <i>histories<\/i>\n      of structured <i>facts<\/i>.  Our ideas and approach will continue to be\n      articulated in detail on our <a href=\"https:\/\/blog.datopia.io\">blog<\/a>.\n    <\/p><\/div><\/div>","created_at":"2019-03-13T22:49:40+0000","updated_at":"2019-03-13T22:49:40+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"datopia.io","preview_picture":"https:\/\/datopia.io\/images\/og-fb-logo.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","elisp"],"is_public":false,"id":874,"title":"skeeto\/emacs-aio","url":"https:\/\/github.com\/skeeto\/emacs-aio","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><code>aio<\/code> is to Emacs Lisp as <a href=\"https:\/\/docs.python.org\/3\/library\/asyncio.html\" rel=\"nofollow\"><code>asyncio<\/code><\/a> is to Python. This\npackage builds upon Emacs 25 generators to provide functions that\npause while they wait on asynchronous events. They do not block any\nthread while paused.<\/p>\n<p>See also: <a href=\"https:\/\/nullprogram.com\/blog\/2019\/03\/10\/\" rel=\"nofollow\">An Async \/ Await Library for Emacs Lisp<\/a><\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>An async function is defined using <code>aio-defun<\/code> or <code>aio-lambda<\/code>. The\nbody of such functions can use <code>aio-await<\/code> to pause the function and\nwait on a given promise. The function continues with the promise's\nresolved value when it's ready. The package provides a number of\nfunctions that return promises, and every async function returns a\npromise representing its future return value.<\/p>\n<p>For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(aio-defun foo (url)\n  (aio-await (aio-sleep 3))\n  (message \"Done sleeping. Now fetching %s\" url)\n  (let* ((result (aio-await (aio-url-retrieve url)))\n         (contents (with-current-buffer (cdr result)\n                     (prog1 (buffer-string)\n                       (kill-buffer)))))\n    (message \"Result: %s\" contents)))<\/pre><\/div>\n<p>If an uncaught signal terminates an asynchronous function, that signal\nis captured by its return value promise and propagated into any\nfunction that awaits on that function.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(aio-defun arith-error-example ()\n  (\/ 0 0))\n(aio-defun consumer-example ()\n  (condition-case error\n      (aio-await (arith-error-example))\n    (arith-error (message \"Caught %S\" error))))\n(consumer-example)\n;; =&gt; #s(aio-promise nil nil)\n;; *Messages*: Caught (arith-error)<\/pre><\/div>\n<p>Converting a callback-based function into a promise-returning,\nasync-friendly function is simple. Create a new promise object with\n<code>aio-promise<\/code>, then <code>aio-resolve<\/code> that promise in your callback. To\nchain onto an existing promise, use <code>aio-listen<\/code> to attach a new\ncallback.<\/p>\n<h2><a id=\"user-content-promise-returning-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#promise-returning-functions\"><\/a>Promise-returning functions<\/h2>\n<p>Here are some useful promise-returning \u2014 i.e. awaitable \u2014 functions\ndefined by this package.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(aio-sleep seconds &amp;optional result)\n;; Return a promise that is resolved after SECONDS with RESULT.\n(aio-url-retrieve url &amp;optional silent inhibit-cookies)\n;; Wraps `url-retrieve' in a promise.\n(aio-process-sentinel process)\n;; Return a promise representing the sentinel of PROCESS.\n(aio-process-filter process)\n;; Return a promise representing the filter of PROCESS.\n(aio-select promises)\n;; Return a promise that resolves when any in PROMISES resolves.\n(aio-all promises)\n;; Return a promise that resolves when all PROMISES are resolved.\"<\/pre><\/div>\n<\/article>","created_at":"2019-03-11T13:24:18+0000","updated_at":"2019-03-11T13:24:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/29988?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":873,"title":"The Essence of Datalog (2018)","url":"https:\/\/news.ycombinator.com\/item?id=19351385","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/front\">past<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D19351385\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"19351385\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/dodisturb.me\/posts\/2018-12-25-The-Essence-of-Datalog.html\" class=\"storylink\">The Essence of Datalog (2018)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=dodisturb.me\">dodisturb.me<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        104 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=bibyte\" class=\"hnuser\">bibyte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=19351385\">8 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=19351385&amp;goto=item%3Fid%3D19351385\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=The%20Essence%20of%20Datalog&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=The%20Essence%20of%20Datalog\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=19351385&amp;auth=7f8e1529b6e42118db6065a0ccf998333cea00ef\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=19351385\">19\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/www.ycombinator.com\/apply\/\">\n        Applications are open for YC Summer 2019\n      <\/a><\/div><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-03-10T19:42:28+0000","updated_at":"2019-03-10T19:42:28+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["datalog","haskell","logic-programming"],"is_public":false,"id":872,"title":"Do disturb me | The Essence of Datalog","url":"https:\/\/dodisturb.me\/posts\/2018-12-25-The-Essence-of-Datalog.html","content":"<p class=\"in-which\">In which we implement a simple Datalog engine in not many lines of Haskell to understand its semantics.<\/p><div class=\"post\"><p>Datalog is arguably the simplest logic programming language there is. Depending on your background, you can see it as a principled SQL or a Prolog with manners. It has recently become popular among the scalable program analysis crowd. Semantically, it is simple and clean, so it doesn\u2019t take many lines of Haskell to produce an (inefficient) implementation. I\u2019ll follow Richard Feynman\u2019s maxim: \u201cWhat I cannot create, I do not understand.\u201d Hopefully, this post should give you some feel for logic and relational programming and also demonstrate that \u03bb-calculus is not the only language you can implement with ease in Haskell.<\/p><p>We start by looking at a simple program and implement an evaluator for it. Along with the implementation, we give a lightweight discussion of semantic concerns, in particular about its termination. Later, we try quenching the thirst for efficiency a little by giving some avenues for optimisation and conclude with where to go next for those interested. There is a copy of the full implementation at the end of the post.<\/p><p>I want this post to be accessible to everyone, not only to the programming language people. For that reason, all uses of theory terms are auxiliary. If you feel that\u2019s not the case at any point, give me a shout on Twitter or email me (both available in <a href=\"https:\/\/dodisturb.me\">my homepage<\/a>) and I\u2019ll fix it.<\/p><h2 id=\"a-crash-course-in-logic-programming\">A crash course in logic programming<\/h2><p>The litmus test for any logic programming language is to be able to compute a variation of the following ancestry program.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">adviser(\"Andrew Rice\",     \"Mistral Contrastin\").\nadviser(\"Dominic Orchard\", \"Andrew Rice\").\nadviser(\"Andy Hopper\",     \"Andrew Rice\").\nadviser(\"Alan Mycroft\",    \"Dominic Orchard\").\nadviser(\"David Wheeler\",   \"Andy Hopper\").\nadviser(\"Rod Burstall\",    \"Alan Mycroft\").\nadviser(\"Robin Milner\",    \"Alan Mycroft\").\nacademicAncestor(X,Y) :- adviser(X,Y).\nacademicAncestor(X,Z) :- adviser(X,Y), academicAncestor(Y,Z).<\/pre><\/div><p>This program encodes some facts about my academic genealogy encoded in the <code>adviser<\/code> relation and has some basic rules that define what it means to be an <code>academicAncestor<\/code>. The facts effectively constitute a database table. Rules allow us to deduce new knowledge. The first rule says that an adviser is an ancestor. The second says the adviser of a known ancestor is also an ancestor. This ability to use recursion to specify relationships in data was one of the original selling points of Datalog.<\/p><p>We want to query the data that is represented explicitly and implicitly. Here are some example queries:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">?- academicAncestor(\"Robin Milner\", Intermediate),\n   academicAncestor(Intermediate, \"Mistral Contrastin\").\n?- academicAncestor(\"Alan Turing\", \"Mistral Contrastin\").\n?- academicAncestor(\"David Wheeler\", \"Mistral Contrastin\").<\/pre><\/div><p>The first one says \u201cis there an academic ancestorial connection between Robin Milner and I and if so who?\u201d. The second one says \u201cis Alan Turing my academic ancestor?\u201d and the third one is the same question but for David Wheeler.<\/p><p>Another appeal of Datalog can be seen in these queries. Datalog is mostly relations (hence the name relational programming). This means there are no inherent inputs and outputs. That\u2019s why we can fill in or leave out either parameter of <code>academicAncestor<\/code>. You leave a variable in place of the knowledge you want to know and the system fills it for you. This also means you can write a program for a particular input and output relationship in mind and you get the opposite program for free.<\/p><h2 id=\"evaluating-a-datalog\">Evaluating a Datalog<\/h2><p>Datalog is similar to \u03bb-calculus in the sense that there are as many variants of both as there are sands in a beach.<a href=\"#fn1\" class=\"footnoteRef\" id=\"fnref1\"><sup>1<\/sup><\/a> The variation we are going to implement today is the simplest one there is. The bits and pieces you see in the previous section are all you get.<\/p><p>We\u2019ll implement the engine the way we evaluate Datalog programs: <em>initially, one rule at a time, then all at once<\/em>. Since Datalog is deeply rooted in database theory, it is usually implemented as relational algebra concepts such as joins and selections. I have a programming language background, so I\u2019ll use substitutions instead. This will be entirely equivalent, but also lead to simpler code.<\/p><h3 id=\"representing-datalog\">Representing Datalog<\/h3><p>A rule consists of a head and a body. A head is an atom and so are the comma separated items in the body. An atom then consists of a predicate name like <code>adviser<\/code> and a list of terms. We have two kinds of terms: variables and symbols. This corresponds to few simple datatype declarations.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">data Rule = Rule { _head :: Atom, _body :: [ Atom ] }\ndata Atom = Atom { _predSym :: String, _terms :: [ Term ] } deriving Eq\ndata Term = Var String | Sym String deriving Eq<\/pre><\/div><p>The following rule<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">academicAncestor(X,Z) :- adviser(X,Y), academicAncestor(Y,Z).<\/pre><\/div><p>corresponds to<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">Rule (Atom \"academicAncestor\" [ Var \"X\", Var \"Z\" ])\n  [ Atom \"adviser\"          [ Var \"X\", Var \"Y\" ]\n  , Atom \"academicAncestor\" [ Var \"Y\", Var \"Z\" ] ]<\/pre><\/div><p>Facts are just bodiless rules. For example,<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">adviser(\"Andrew Rice\", \"Mistral Contrastin\").<\/pre><\/div><p>corresponds to<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">Rule (Atom \"adviser\" [ Sym \"Andrew Rice\", Sym \"Mistral Contrastin\" ]) []<\/pre><\/div><p>You might be thinking but what about the queries, the <code>Rule<\/code> datatype doesn\u2019t allow headless bodies. That\u2019s true, but the dirty secret about queries is that they get rewritten as rules with the head atom generated from the query. The predicate symbol for each query is fresh and the terms of the head are just the variables in the body. For example,<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">?- academicAncestor(\"Robin Milner\", Intermediate),\n   academicAncestor(Intermediate, \"Mistral Contrastin\").<\/pre><\/div><p>gets fresh head atom <code>query1(Intermediate)<\/code> and can be thought of as<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">query1(Intermediate) :-\n  academicAncestor(\"Robin Milner\", Intermediate),\n  academicAncestor(Intermediate, \"Mistral Contrastin\").<\/pre><\/div><p>An entire Datalog program is nothing but a collection of rules.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">type Program = [ Rule ]<\/pre><\/div><p>We also need a way of representing the known facts about the universe. We can call this our <em>knowledge base<\/em>. For simplicity, we use a horribly inefficient list of atoms.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">type KnowledgeBase = [ Atom ]<\/pre><\/div><p>Since we\u2019ll use substitutions to evaluate atoms, we can define that here too.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">type Substitution = [ (Term, Term) ]<\/pre><\/div><p>Our substitutions are simpler than those that are used in \u03bb-calculus. A Haskell tuple <code>(x,c)<\/code> represents the substitution <code>[c\/x]<\/code> meaning we replace <code>c<\/code> for <code>x<\/code>. Here, <code>x<\/code> is always a variable and <code>c<\/code> is always a constant (symbol). The variable restriction is usual; it is just not reflected in the type. The constant restriction, however, is not. Datalog doesn\u2019t have function symbols and substituting a variable is not necessary for its evaluation. Like the variable restriction, the constant restriction is also not reflected in the type.<\/p><p>Applying a substitution to an atom or a variable (potentially) makes it <em>ground<\/em> that is it replaces the variable with a constant.<\/p><p>We can also define the only constant that will be used in the evaluator: the empty substitution.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">emptySubstitution :: Substitution\nemptySubstitution = []<\/pre><\/div><p>Mind-blowing, I know.<\/p><p>Let\u2019s focus on evaluating a single rule from our example program.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode prolog\">academicAncestor(X,Z) :- adviser(X,Y), academicAncestor(Y,Z).<\/pre><\/div><p>We know some facts about the world and using this rule, we want to know some more. There are two ways. One is to gather everything we know about <code>adviser<\/code> and <code>academicAncestor<\/code> separately and <em>join<\/em> them together making sure the second parameter (<code>Y<\/code>) of <code>adviser<\/code> matches the first parameter (<code>Y<\/code>) of <code>academicAncestor<\/code>. Another way is to look at any one of the atoms and find assignments to its variables and substitute them in the other atom, only then we look for assignments to the remaining variables of this second atom. This approach would work for the example above as follows: find possible assignments to <code>X<\/code> and <code>Y<\/code> through <code>adviser<\/code>, substitute the values of <code>Y<\/code> in <code>academicAncestor<\/code>, and finally look for values of <code>academicAncestor<\/code> in the knowledge base that agree on the newly substituted value of <code>Y<\/code> to obtain values for <code>Z<\/code>.<\/p><p>In the context of databases, the first is called <em>join-before-select<\/em> and the second, not so surprisingly, <em>select-before-join<\/em>. We\u2019ll implement the latter because it is simpler.<a href=\"#fn2\" class=\"footnoteRef\" id=\"fnref2\"><sup>2<\/sup><\/a><\/p><p>Since we need unification to obtain substitutions from facts &amp; body atoms and substitution to ground further atoms. We start by implementing those.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">substitute :: Atom -&gt; Substitution -&gt; Atom\nsubstitute atom substitution = atom { _terms = map go (_terms atom) }\n  where\n  go sym@Sym{} = sym\n  go var@Var{} = fromMaybe var (var `lookup` substitution)<\/pre><\/div><p>Substitution is pretty much what you would expect looking up what to substitute for variables and do it whenever it can find a binding otherwise leaving the variable alone.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">unify :: Atom -&gt; Atom -&gt; Maybe Substitution\nunify (Atom predSym ts) (Atom predSym' ts')\n  | predSym == predSym' = go $ zip ts ts'\n  | otherwise           = Nothing\n  where\n  go :: [ (Term, Term) ] -&gt; Maybe Substitution\n  go []                           = Just emptySubstitution\n  go ((s@Sym{}, s'@Sym{}) : rest) = if s == s' then go rest else Nothing\n  go ((v@Var{}, s@Sym{})  : rest) = do\n    incompleteSubstitution &lt;- go rest\n    case v `lookup` incompleteSubstitution of\n      Just s' | s \/= s'   -&gt; Nothing\n      _                   -&gt; return $ (v,s) : incompleteSubstitution\n  go ((_, Var{}) : _) = error \"The second atom is assumed to be ground.\"<\/pre><\/div><p>Unification is also simple, in fact, too simple. For one thing, we cheat and unify whenever we have two atoms that have the same predicate symbol. In Datalog, a predicate is determined by its predicate symbol <em>and<\/em> arity (the number of terms). Here, we assume each predicate symbol determines the arity. More importantly, we throw an error when a term from the second atom is a variable. The reason is unification only occurs between a body atom and a fact. Since facts cannot have variables, this is safe. This is consistent with our earlier assumptions about the form of substitutions.<\/p><p>We take special care of unification of atoms with repeated variables. It requires failing in case of a contradictory variable assignment. Consider unifying <code>p(X,X)<\/code> with <code>p(\"a\",\"b\")<\/code>.<\/p><p>Next we evaluate an atom into a list of substitutions by finding facts that fit the template provided by a body atom and capturing the assignments to its variables. Those assignments to variables are just the substitutions we are looking for.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">evalAtom :: KnowledgeBase -&gt; Atom -&gt; [ Substitution ] -&gt; [ Substitution ]\nevalAtom kb atom substitutions = do\n  substitution &lt;- substitutions\n  let downToEarthAtom = substitute atom substitution\n  extension &lt;- mapMaybe (unify downToEarthAtom) kb\n  return $ substitution &lt;&gt; extension<\/pre><\/div><p>Here we do exactly what is described above but build the substitutions by accumulation. This means we have substitutions that we use on the body atom to ground its variables, but then unifying this more down to earth atom (it\u2019s more ground, get it?) with the facts we know gives us extensions to the substitution we started with. Thus, we extend and accumulate substitutions.<\/p><p>Now all we need to do is to walk the body and accumulate substitutions, then using these substitutions we deduce new facts based on the head of the rule.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">walk :: KnowledgeBase -&gt; [ Atom ] -&gt; [ Substitution ]\nwalk kb = foldr (evalAtom kb) [ emptySubstitution ]\nevalRule :: KnowledgeBase -&gt; Rule -&gt; KnowledgeBase\nevalRule kb (Rule head body) = map (substitute head) (walk kb body)<\/pre><\/div><p>Here\u2019s something to think about. We said the facts we deduce will be ground, but <code>evalRule<\/code> does not check if the substitution has a binding for all the variables that appear in the head. Does that mean we are potentially concluding non-ground facts? Don\u2019t worry if you\u2019re not sure, we\u2019ll come back to it <a href=\"#range-restriction\">below<\/a>.<\/p><h3 id=\"all-at-once\">All at once<\/h3><p>All we need now is to evaluate all of the rules together. We\u2019ll do one of the simplest possible things. We\u2019ll evaluate each rule independently and then concatenate the newly derived facts together and bundle them with what we already know. Then repeat the process until we can\u2019t produce new facts any more.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">immediateConsequence :: Program -&gt; KnowledgeBase -&gt; KnowledgeBase\nimmediateConsequence rules kb =\n  nub . (kb &lt;&gt;) . concatMap (evalRule kb) $ rules\nsolve :: Program -&gt; KnowledgeBase\nsolve rules =\n  if all isRangeRestricted rules\n    then fix step []\n    else error \"The input program is not range-restricted.\"\n  where\n  step :: (KnowledgeBase -&gt; KnowledgeBase)\n       -&gt; (KnowledgeBase -&gt; KnowledgeBase)\n  step f currentKB | nextKB &lt;- immediateConsequence rules currentKB =\n    if nextKB == currentKB\n      then currentKB\n      else f nextKB<\/pre><\/div><p>The <code>immediateConsequence<\/code> function does the bundling step and <code>solve<\/code> computes the fixpoint from an empty set of facts.<\/p><p>The only thing left unexplained is the <code>isRangeRestricted<\/code> predicate which ensures the program is well-formed.<\/p><h3 id=\"semantic-considerations\">Semantic considerations<\/h3><p>Now that we have something that almost compiles we can talk about semantics. Starting with the last missing piece: the range restriction predicate followed by termination.<\/p><h4 id=\"range-restriction-and-domain-independence\">Range restriction and domain independence<\/h4><p>In a rule, if every variable in the head appears somewhere in the body, we call the rule range-restricted. We check if every rule is range-restricted before evaluating them in <code>solve<\/code>.<a href=\"#fn3\" class=\"footnoteRef\" id=\"fnref3\"><sup>3<\/sup><\/a> This is why we didn\u2019t have to check if there were any variables left in the head after substitution in <code>evalRule<\/code>. If we can get to the stage of substituting into head, we are guaranteed to find values for each head variable.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">isRangeRestricted :: Rule -&gt; Bool\nisRangeRestricted Rule{..} =\n  vars _head `isSubsequenceOf` concatMap vars _body\n  where\n  vars Atom{..} = nub $ filter (\\case {Var{} -&gt; True; _ -&gt; False}) _terms<\/pre><\/div><p>There still remains the question of why we need range restriction at all? After all, the ability to deduce generic facts seems useful. For example, stating <code>p(X,X)<\/code> as a fact is a compact way of saying <code>p<\/code> is reflexive.<\/p><p>The problem is something called <em>domain independence<\/em>. It basically means if the set of values a variable can take changes from one database to another (but not the instance itself), queries still compute the same thing. In other words, it prevents your program\u2019s result to change under your feet if the values in your database are the same. Since checking domain independence in general is undecidable, range restriction is a safe syntactic approximation.<\/p><p>In this implementation what this means is if I change the definition of datatype <code>Term<\/code> so that the symbols do not use <code>String<\/code> but instead a type for strings up to length 10, if my initial ground facts were all strings of length up to 10, then the results to all queries remain the same. We can\u2019t guarantee the same thing without domain independence.<\/p><p>Another problem is that when a query is posed with free variables, we expect values to be filled for that variable. If we can deduce a generic fact such as <code>p(X,X)<\/code>, then asking for the values of <code>X<\/code> would enumerate the set of infinite strings.<\/p><p>Bear in mind, there are Datalog variants that lift this restriction. In fact, it wouldn\u2019t be too much work to implement it here. We just need a notion of \u03b1-equivalence, a store that can handle non-ground values, and a unification algorithm that handles variables.<\/p><h4 id=\"every-good-thing-must-come-to-an-end\">Every good thing must come to an end<\/h4><p>Does this procedure terminate? Yes, it does. When? So long as you don\u2019t have infinite programs. It would be a cheeky thing to do though. One of the appeals of Datalog is that it is not Turing-complete and every query to the system terminates.<\/p><p>The termination argument is pretty simple. Immediate consequence function is <em>monotone<\/em> that is the facts it produces encompasses the facts it starts with. Further, the number of facts that can be produced is <em>bounded<\/em> if we start with a finite database. Our initial set of facts used to kick-start <code>solve<\/code> is empty. Hence, as long as our program is finite, we have finite number of facts.<a href=\"#fn4\" class=\"footnoteRef\" id=\"fnref4\"><sup>4<\/sup><\/a> As an upper bound, if we have <em>N<\/em> constants throughout the program, then for each relation of arity <em>k<\/em>, we can have at most <em>N<\/em><sup><em>k<\/em><\/sup> facts. So the number of facts we can produce is also finite.<\/p><p>One place I\u2019m being a bit loose is the monotone bit. What I said makes sense with sets, but we\u2019re working here with lists. Although <code>[1,2,3]<\/code> and <code>[2,1,3]<\/code> encompass each other, they don\u2019t compare equal using <code>==<\/code>. The reason this implementation never gets in a cycle is because <code>nub<\/code> function called on the amalgamation of facts in <code>immediateConsequence<\/code> preserves the first occurrence of each element in the list. Since we prepend the already known facts before calling <code>nub<\/code>, <code>==<\/code> behaves as if it is set equality.<\/p><p>If you want to sound smart explaining all of this and reduce the size of your audience to a group of people who would understand this without you mentioning anyway, you can just say the following. We have a non-empty finite lattice which means it is complete and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Knaster&#x2013;Tarski_theorem\">Knaster-Tarski theorem<\/a> (which is a very cool theorem \u2764\ufe0f) applied to the immediate consequence monotone function over this complete lattice implies that it has a least fixpoint.<\/p><h2 id=\"quality-assurance\">Quality assurance<\/h2><p>Time for the litmus test. We can now translate the ancestry program into Haskell.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">ancestor :: Program\nancestor =\n  -- Facts\n  fmap (\\terms -&gt; Rule (Atom \"adviser\" terms) [])\n    [ [ Sym \"Andrew Rice\",     Sym \"Mistral Contrastin\" ]\n    , [ Sym \"Dominic Orchard\", Sym \"Mistral Contrastin\" ]\n    , [ Sym \"Andy Hopper\",     Sym \"Andrew Rice\" ]\n    , [ Sym \"Alan Mycroft\",    Sym \"Dominic Orchard\" ]\n    , [ Sym \"David Wheeler\",   Sym \"Andy Hopper\" ]\n    , [ Sym \"Rod Burstall\",    Sym \"Alan Mycroft\" ]\n    , [ Sym \"Robin Milner\",    Sym \"Alan Mycroft\" ]\n    ] &lt;&gt;\n  -- Actual rules\n  [ Rule (Atom \"academicAncestor\" [ Var \"X\", Var \"Y\" ])\n      [ Atom \"adviser\" [ Var \"X\", Var \"Y\" ] ]\n  , Rule (Atom \"academicAncestor\" [ Var \"X\", Var \"Z\" ])\n      [ Atom \"adviser\"          [ Var \"X\", Var \"Y\" ]\n      , Atom \"academicAncestor\" [ Var \"Y\", Var \"Z\" ] ]\n  ] &lt;&gt;\n  -- Queries\n  [ Rule (Atom \"query1\" [ Var \"Intermediate\" ])\n      (fmap (Atom \"academicAncestor\")\n        [ [ Sym \"Robin Milner\", Var \"Intermediate\" ]\n        , [ Var \"Intermediate\", Sym \"Mistral Contrastin\" ] ])\n  , Rule (Atom \"query2\" [ ])\n      [ Atom \"academicAncestor\"\n          [ Sym \"Alan Turing\", Sym \"Mistral Contrastin\" ] ]\n  , Rule (Atom \"query3\" [ ])\n      [ Atom \"academicAncestor\"\n          [ Sym \"David Wheeler\", Sym \"Mistral Contrastin\" ] ]\n  ]<\/pre><\/div><p>We can make querying a bit more pleasant with a function that returns possible bindings.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">query :: String -&gt; Program -&gt; [ Substitution ]\nquery predSym pr =\n  case queryVarsL of\n    [ queryVars ] -&gt; zip queryVars &lt;$&gt; relevantKnowledgeBaseSyms\n    [] -&gt; error $ \"The query '\" ++ predSym ++ \"' doesn't exist.\"\n    _  -&gt; error $ \"The query '\" ++ predSym ++ \"' has multiple clauses.\"\n  where\n  relevantKnowledgeBase = filter ((== predSym) . _predSym) $ solve pr\n  relevantKnowledgeBaseSyms = _terms &lt;$&gt; relevantKnowledgeBase\n  queryRules = filter ((== predSym) . _predSym . _head) pr\n  queryVarsL = _terms . _head &lt;$&gt; queryRules<\/pre><\/div><p>All it does it to run the program, find the named query and return the bindings to its arguments. Let\u2019s execute all three queries.<\/p><pre>&gt; query \"query1\" ancestor\n[[(Intermediate,\"Dominic Orchard\")],[(Intermediate,\"Alan Mycroft\")]]\n&gt; query \"query2\" ancestor\n[]\n&gt; query \"query3\" ancestor\n[[]]<\/pre><p>The empty list of variable bindings, <code>[]<\/code>, mean there are no possible assignments that satisfy the query, whereas the singleton empty binding, <code>[[]]<\/code>, means the query is satisfied. It just didn\u2019t have any variables that need to be bound to values.<\/p><p>All three queries return the expected results. Dominic Orchard and Alan Mycroft are between me and Robin Milner. I am not an academic descendant of Alan Turing, but I am of David Wheeler.<\/p><p>Since these three queries run fine, we can safely conclude that the evaluator here is bug-free.<\/p><h2 id=\"addressing-the-turtle-in-the-room\">Addressing the turtle in the room<\/h2><p>As I mentioned before, this evaluator is inefficient. Discussing the reasons why and what we can do to make it better is not only a software engineering exercise, but also another way of highlighting why Datalog is a good language. The optimisations I discuss here are quite major, but they are not a complete list. There are many other optimisations that draw both from programming languages and database theory literatures.<\/p><h3 id=\"optimise-for-the-query\">Optimise for the query<\/h3><p>Basically, all I told you about Datalog semantics is incomplete. You\u2019ll struggle to find a resource that discusses Datalog evaluation the way I do because the semantics are always defined with respect to a query. So what we really need is a program query pair. What we compute here is much stronger and is also unnecessary. For example, if the queries were interested only in the <code>adviser<\/code> relation, computing <code>academicAncestors<\/code> would be a waste of time.<\/p><p>One way of dealing with this is to use a top-down evaluator which starts from the query and uses <em>resolution<\/em> which is a proof technique. This allows only the relevant facts to be derived. Most Prolog interpreters use a similar top-down evaluation strategy. It is not preferred in Datalog for various reasons. One of which is that the na\u00efve implementation of it brings non-termination.<\/p><p>We can achieve the same in a bottom-up evaluator (like ours) using <em>magic set transformation<\/em>. The very long story short, based on the input program, it generates magic rules based on the dataflow of the program and inserts atoms defined by these rules into the original rule bodies to restrict what can be derived. Its effectiveness varies from program to program and there are alternatives to it. If you want to learn more about it, you can look at <a href=\"https:\/\/www.sciencedirect.com\/science\/article\/pii\/074310669190038Q\">On the Power of Magic<\/a> (you must admit, it is a catchy title) by Beeri and Ramakrishnan.<\/p><h3 id=\"semi-na&#xef;ve-evaluation\">Semi-na\u00efve evaluation<\/h3><p>While discussing <a href=\"#range-restriction\">range restriction<\/a>, we mentioned that this implementation rederives all the known facts in each iteration. This is awful and is why it is called the na\u00efve evaluation. There is also a modestly named semi-na\u00efve evaluation. It exploits a simple observation. In order to derive a new fact from a rule, at least one new fact of the previous iteration needs to be used.<\/p><p>The way this gets implemented is that we maintain a delta of facts as well as an accumulator and rewrite the rules to make versions of them that use the deltas.<\/p><p>This is particularly effective for the so called <em>linear rules<\/em> such as the recursive <code>academicAncestor<\/code> rule which only has one derived predicate in its body. In that case, the semi-na\u00efve evaluation turns a quadratic computation into a linear one.<\/p><h3 id=\"incremental-evaluation\">Incremental evaluation<\/h3><p>Datalog is amenable to incremental evaluation. Even in this version of the evaluator, you can see that when a fixpoint is reached, we can enrich the resulting knowledge base with additional facts and apply the fixpoint algorithm again. This performs at least as good as starting from scratch and often much much better.<\/p><p>The situation gets more complicated when there is negation in the language because additional facts invalidate facts that depend on the negation of those facts. This is a particular instance of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Non-monotonic_logic\"><em>non-monotonic reasoning<\/em><\/a>. However, Datalog, overall, is still a good language if you\u2019re after incremental evaluation.<\/p><p>Furthermore, the idea of incremental evaluation is intricately connected with maintaining a delta of facts. If your evaluator uses semi-na\u00efve evaluation already, having an incremental evaluator is not much extra effort.<\/p><h3 id=\"data-structures\">Data structures<\/h3><p>Using lists to compute over sets is a bad idea. Depending on the application, hash tables, proper databases, in memory key-value stores, or at the very least of <code>Set<\/code> and <code>Map<\/code> modules from the <code>containers<\/code> package will certainly make things better. Use of sets in general is a good idea because then we don\u2019t have to rely on <code>nub<\/code> function\u2019s internals for termination as discussed <a href=\"#every-good-thing-must-come-to-an-end\">earlier<\/a>.<\/p><h3 id=\"dependency-graphs\">Dependency graphs<\/h3><p>We evaluate all rules in one pot but that is also very inefficient. We can partition the program by determining dependencies between predicates. This allows us to treat evaluated dependencies as static knowledge, so we have to deal with a smaller collection of changing facts at any one time.<\/p><p>For example in the ancestor program, we would have a graph with <code>adviser<\/code> and <code>academicAncestor<\/code> nodes where the former points to the latter and the latter loops around. Meaning we can compute the fixpoint for <code>adviser<\/code> rules first, then forget about those rules and compute the fixpoint of <code>academicAncestor<\/code>.<\/p><p>Also if your Datalog variant has <em>stratified negation<\/em> which is a popular way of incorporating negated atoms, you already have to do the dependency analysis and partition your program. So engineering-wise, this optimisation comes for free.<\/p><h3 id=\"parallelisation-distributed-computation\">Parallelisation &amp; distributed computation<\/h3><p>Datalog evaluation is great for data-parallel computation. Even by inspecting our implementation, it is evident that evaluating all the rules in a given iteration is an embarrassingly parallel problem. The dependency graph can be used to further parallelise the evaluation using work queues. Similarly, the workload can be separated over multiple machines, a bit like <a href=\"https:\/\/en.wikipedia.org\/wiki\/MapReduce\">MapReduce<\/a>.<\/p><p>Short program, long prose is the theme. I hope I managed to give some ideas about how Datalog works. I tried to squeeze in as much semantics and tips for more efficient implementation as possible which is probably more useful than a short program.<\/p><p>If you want to look at Datalog further. <a href=\"https:\/\/souffle-lang.github.io\">Souffl\u00e9<\/a> is a modern variant geared towards program analysis and is blazingly fast.<\/p><p>If you are interested in a more formal treatment of Datalog as well as some of the optimisations I mentioned chapters 12 to 15 of <a href=\"http:\/\/webdam.inria.fr\/Alice\/\">Foundations of Databases<\/a> by Abiteboul, Hull, and Vianu are probably the best resources which collect everything together in such detail and with modern exposition.<\/p><h2 id=\"full-program\">Full program<\/h2><p>Here\u2019s the full program for your convenience.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode haskell\">{-# LANGUAGE LambdaCase #-}\n{-# LANGUAGE RecordWildCards #-}\nmodule SimpleDatalog where\nimport Data.Function (fix)\nimport Data.List (nub, intercalate, isSubsequenceOf)\nimport Data.Maybe (mapMaybe, fromMaybe, isNothing)\ntype Program = [ Rule ]\ndata Rule = Rule { _head :: Atom, _body :: [ Atom ] }\ndata Atom = Atom { _predSym :: String, _terms :: [ Term ] } deriving Eq\ndata Term = Var String | Sym String deriving Eq\ntype KnowledgeBase = [ Atom ]\ntype Substitution = [ (Term, Term) ]\nemptySubstitution :: Substitution\nemptySubstitution = []\nsolve :: Program -&gt; KnowledgeBase\nsolve rules =\n  if all isRangeRestricted rules\n    then fix step []\n    else error \"The input program is not range-restricted.\"\n  where\n  step :: (KnowledgeBase -&gt; KnowledgeBase)\n       -&gt; (KnowledgeBase -&gt; KnowledgeBase)\n  step f currentKB | nextKB &lt;- immediateConsequence rules currentKB =\n    if nextKB == currentKB\n      then currentKB\n      else f nextKB\nisRangeRestricted :: Rule -&gt; Bool\nisRangeRestricted Rule{..} =\n  vars _head `isSubsequenceOf` concatMap vars _body\n  where\n  vars Atom{..} = nub $ filter (\\case {Var{} -&gt; True; _ -&gt; False}) _terms\nimmediateConsequence :: Program -&gt; KnowledgeBase -&gt; KnowledgeBase\nimmediateConsequence rules kb =\n  nub . (kb &lt;&gt;) . concatMap (evalRule kb) $ rules\nevalRule :: KnowledgeBase -&gt; Rule -&gt; KnowledgeBase\nevalRule kb (Rule head body) = map (substitute head) (walk kb body)\nwalk :: KnowledgeBase -&gt; [ Atom ] -&gt; [ Substitution ]\nwalk kb = foldr (evalAtom kb) [ emptySubstitution ]\nevalAtom :: KnowledgeBase -&gt; Atom -&gt; [ Substitution ] -&gt; [ Substitution ]\nevalAtom kb atom substitutions = do\n  substitution &lt;- substitutions\n  let downToEarthAtom = substitute atom substitution\n  extension &lt;- mapMaybe (unify downToEarthAtom) kb\n  return $ substitution &lt;&gt; extension\nsubstitute :: Atom -&gt; Substitution -&gt; Atom\nsubstitute atom substitution = atom { _terms = map go (_terms atom) }\n  where\n  go sym@Sym{} = sym\n  go var@Var{} = fromMaybe var (var `lookup` substitution)\nunify :: Atom -&gt; Atom -&gt; Maybe Substitution\nunify (Atom predSym ts) (Atom predSym' ts')\n  | predSym == predSym' = go $ zip ts ts'\n  | otherwise           = Nothing\n  where\n  go :: [ (Term, Term) ] -&gt; Maybe Substitution\n  go []                           = Just emptySubstitution\n  go ((s@Sym{}, s'@Sym{}) : rest) = if s == s' then go rest else Nothing\n  go ((v@Var{}, s@Sym{})  : rest) = do\n    incompleteSubstitution &lt;- go rest\n    case v `lookup` incompleteSubstitution of\n      Just s' | s \/= s'   -&gt; Nothing\n      _                   -&gt; return $ (v,s) : incompleteSubstitution\n  go ((_, Var{}) : _) = error \"The second atom is assumed to be ground.\"\n{-\n- adviser(\"Andrew Rice\",     \"Mistral Contrastin\").\n- adviser(\"Dominic Orchard\", \"Andrew Rice\").\n- adviser(\"Andy Hopper\",     \"Andrew Rice\").\n- adviser(\"Alan Mycroft\",    \"Dominic Orchard\").\n- adviser(\"David Wheeler\",   \"Andy Hopper\").\n- adviser(\"Rod Burstall\",    \"Alan Mycroft\").\n- adviser(\"Robin Milner\",    \"Alan Mycroft\").\n-\n- academicAncestor(X,Y) :- adviser(X,Y).\n- academicAncestor(X,Z) :- adviser(X,Y), academicAncestor(Y,Z).\n-\n- ?- academicAncestor(\"Robin Milner\", Intermediate),\n-    academicAncestor(Intermediate, \"Mistral Contrastin\").\n- ?- academicAncestor(\"Alan Turing\", \"Mistral Contrastin\").\n- ?- academicAncestor(\"David Wheeler\", \"Mistral Contrastin\").\n-}\nancestor :: Program\nancestor =\n  -- Facts\n  fmap (\\terms -&gt; Rule (Atom \"adviser\" terms) [])\n    [ [ Sym \"Andrew Rice\",     Sym \"Mistral Contrastin\" ]\n    , [ Sym \"Dominic Orchard\", Sym \"Mistral Contrastin\" ]\n    , [ Sym \"Andy Hopper\",     Sym \"Andrew Rice\" ]\n    , [ Sym \"Alan Mycroft\",    Sym \"Dominic Orchard\" ]\n    , [ Sym \"David Wheeler\",   Sym \"Andy Hopper\" ]\n    , [ Sym \"Rod Burstall\",    Sym \"Alan Mycroft\" ]\n    , [ Sym \"Robin Milner\",    Sym \"Alan Mycroft\" ]\n    ] &lt;&gt;\n  -- Actual rules\n  [ Rule (Atom \"academicAncestor\" [ Var \"X\", Var \"Y\" ])\n      [ Atom \"adviser\" [ Var \"X\", Var \"Y\" ] ]\n  , Rule (Atom \"academicAncestor\" [ Var \"X\", Var \"Z\" ])\n      [ Atom \"adviser\"          [ Var \"X\", Var \"Y\" ]\n      , Atom \"academicAncestor\" [ Var \"Y\", Var \"Z\" ] ]\n  ] &lt;&gt;\n  -- Queries\n  [ Rule (Atom \"query1\" [ Var \"Intermediate\" ])\n      (fmap (Atom \"academicAncestor\")\n        [ [ Sym \"Robin Milner\", Var \"Intermediate\" ]\n        , [ Var \"Intermediate\", Sym \"Mistral Contrastin\" ] ])\n  , Rule (Atom \"query2\" [ ])\n      [ Atom \"academicAncestor\"\n          [ Sym \"Alan Turing\", Sym \"Mistral Contrastin\" ] ]\n  , Rule (Atom \"query3\" [ ])\n      [ Atom \"academicAncestor\"\n          [ Sym \"David Wheeler\", Sym \"Mistral Contrastin\" ] ]\n  ]\nquery :: String -&gt; Program -&gt; [ Substitution ]\nquery predSym pr =\n  case queryVarsL of\n    [ queryVars ] -&gt; zip queryVars &lt;$&gt; relevantKnowledgeBaseSyms\n    [] -&gt; error $ \"The query '\" ++ predSym ++ \"' doesn't exist.\"\n    _  -&gt; error $ \"The query '\" ++ predSym ++ \"' has multiple clauses.\"\n  where\n  relevantKnowledgeBase = filter ((== predSym) . _predSym) $ solve pr\n  relevantKnowledgeBaseSyms = _terms &lt;$&gt; relevantKnowledgeBase\n  queryRules = filter ((== predSym) . _predSym . _head) pr\n  queryVarsL = _terms . _head &lt;$&gt; queryRules<\/pre><\/div><div class=\"footnotes\"><hr \/><ol><li id=\"fn1\"><p>My impression is, as time goes on, \u03bb-calculus variants try to tame \u03bb-calculus (simply typed, linear, dependent, etc.) while Datalog variants (relaxed range-restriction, functional symbols, explicit quantification, etc.) let it run wild.<a href=\"#fnref1\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn2\"><p>As it happens, select-before-join is usually (but certainly not always) the faster approach in real world. However, this doesn\u2019t affect us at all because the performance benefit requires an indexed database for the columns that are grounded by substitution.<a href=\"#fnref2\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn3\"><p>This is a simple check, but if we wanted, we could make it even neater. Range-restriction is something that can be baked into the <code>Rule<\/code> datatype using some dependent types. We would need to modify <code>Atom<\/code> to keep track of its variables, but it is doable. While at it we could also make substitutions keep track of their variables, which would enable us to enforce the assumptions about our substitutions. You see, this is my favourite kind of dependent types. It doesn\u2019t give you full correctness, but eliminates a whole class of software bugs and potentially give you some performance boost. This may very well become a future blog post (famous last words on the topic).<a href=\"#fnref3\">\u21a9<\/a><\/p><\/li>\n<li id=\"fn4\"><p>Infinite programs are an interesting concept. I can\u2019t think of any use for it though. Also evaluating them is tricky as you can\u2019t traverse all your program statements, rules, etc. Notice that our Datalog evaluator wouldn\u2019t be terminating not only because of a potential infinite facts problem, but because we try to evaluate each rule.<a href=\"#fnref4\">\u21a9<\/a><\/p><\/li>\n<\/ol><\/div><\/div>","created_at":"2019-03-10T19:42:14+0000","updated_at":"2019-04-17T03:32:13+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":24,"domain_name":"dodisturb.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":871,"title":"The Emacs Package Developer's Handbook","url":"https:\/\/alphapapa.github.io\/emacs-package-dev-handbook\/","content":"<p>\nThese macros make comparing multiple forms easy:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-elisp\">;;;###autoload\n(cl-defmacro bench-multi (&amp;key (times 1) forms ensure-equal raw)\n  \"Return Org table as a list with benchmark results for FORMS.\n Runs FORMS with `benchmark-run-compiled' for TIMES iterations.\n When ENSURE-EQUAL is non-nil, the results of FORMS are compared,\n and an error is raised if they aren't `equal'. If the results\n are sequences, the difference between them is shown with\n `seq-difference'.\n When RAW is non-nil, the raw results from\n `benchmark-run-compiled' are returned instead of an Org table\n list.\n If the first element of a form is a string, it's used as the\n form's description in the bench-multi-results; otherwise, forms\n are numbered from 0.\n Before each form is run, `garbage-collect' is called.\"\n  ;; MAYBE: Since `bench-multi-lexical' byte-compiles the file, I'm not sure if\n  ;; `benchmark-run-compiled' is necessary over `benchmark-run', or if it matters.\n  (declare (indent defun))\n  (let*((keys (gensym \"keys\"))\n        (result-times (gensym \"result-times\"))\n        (header '((\"Form\" \"x faster than next\" \"Total runtime\" \"# of GCs\" \"Total GC runtime\")\n                  hline))\n        ;; Copy forms so that a subsequent call of the macro will get the original forms.\n        (forms (copy-list forms))\n        (descriptions (cl-loop for form in forms\n                               for i from 0\n                               collect (if (stringp (car form))\n                                           (prog1 (car form)\n                                             (setf (nth i forms) (cadr (nth i forms))))\n                                         i))))\n    `(unwind-protect\n         (progn\n           (defvar bench-multi-results nil)\n           (let* ((bench-multi-results (make-hash-table))\n                  (,result-times (sort (list ,@(cl-loop for form in forms\n                                                        for i from 0\n                                                        for description = (nth i descriptions)\n                                                        collect `(progn\n                                                                   (garbage-collect)\n                                                                   (cons ,description\n                                                                         (benchmark-run-compiled ,times\n                                                                           ,(if ensure-equal\n                                                                                `(puthash ,description ,form bench-multi-results)\n                                                                              form))))))\n                                       (lambda (a b)\n                                         (&lt; (second a) (second b))))))\n             ,(when ensure-equal\n                `(cl-loop with ,keys = (hash-table-keys bench-multi-results)\n                          for i from 0 to (- (length ,keys) 2)\n                          unless (equal (gethash (nth i ,keys) bench-multi-results)\n                                        (gethash (nth (1+ i) ,keys) bench-multi-results))\n                          do (if (sequencep (gethash (car (hash-table-keys bench-multi-results)) bench-multi-results))\n                                 (let* ((k1) (k2)\n                                        ;; If the difference in one order is nil, try in other order.\n                                        (difference (or (setq k1 (nth i ,keys)\n                                                              k2 (nth (1+ i) ,keys)\n                                                              difference (seq-difference (gethash k1 bench-multi-results)\n                                                                                         (gethash k2 bench-multi-results)))\n                                                        (setq k1 (nth (1+ i) ,keys)\n                                                              k2 (nth i ,keys)\n                                                              difference (seq-difference (gethash k1 bench-multi-results)\n                                                                                         (gethash k2 bench-multi-results))))))\n                                   (user-error \"Forms' bench-multi-results not equal: difference (%s - %s): %S\"\n                                               k1 k2 difference))\n                               ;; Not a sequence\n                               (user-error \"Forms' bench-multi-results not equal: %s:%S %s:%S\"\n                                           (nth i ,keys) (nth (1+ i) ,keys)\n                                           (gethash (nth i ,keys) bench-multi-results)\n                                           (gethash (nth (1+ i) ,keys) bench-multi-results)))))\n             ;; Add factors to times and return table\n             (if ,raw\n                 ,result-times\n               (append ',header\n                       (bench-multi-process-results ,result-times)))))\n       (unintern 'bench-multi-results nil))))\n(defun bench-multi-process-results (results)\n  \"Return sorted RESULTS with factors added.\"\n  (setq results (sort results (-on #'&lt; #'second)))\n  (cl-loop with length = (length results)\n           for i from 0 to (1- length)\n           for description = (car (nth i results))\n           for factor = (if (&lt; i (1- length))\n                            (format \"%.2f\" (\/ (second (nth (1+ i) results))\n                                              (second (nth i results))))\n                          \"slowest\")\n           collect (append (list description factor)\n                           (list (format \"%.6f\" (second (nth i results)))\n                                 (third (nth i results))\n                                 (if (&gt; (fourth (nth i results)) 0)\n                                     (format \"%.6f\" (fourth (nth i results)))\n                                   0)))))\n<\/pre><\/div><p>\nUsed like:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-elisp\">(bench-multi\n  :forms ((\"org-map-entries\" (sort (org-map-entries (lambda ()\n                                                      (nth 4 (org-heading-components)))\n                                                    \"\/+MAYBE\" 'agenda)\n                                   #'string&lt;))\n          (\"regexp\" (sort (-flatten\n                           (-non-nil\n                            (mapcar (lambda (file)\n                                      (let ((case-fold-search t))\n                                        (with-current-buffer (find-buffer-visiting file)\n                                          (org-with-wide-buffer\n                                           (goto-char (point-min))\n                                           (cl-loop with regexp = (format org-heading-keyword-regexp-format \"MAYBE\")\n                                                    while (re-search-forward regexp nil t)\n                                                    collect (nth 4 (org-heading-components)))))))\n                                    (org-agenda-files))))\n                          #'string&lt;))))\n<\/pre><\/div><table border=\"2\" cellpadding=\"6\" rules=\"groups\" frame=\"hsides\" style=\"border-spacing: 0px;\"><colgroup><col class=\"org-left\" \/><col class=\"org-left\" \/><col class=\"org-right\" \/><col class=\"org-right\" \/><col class=\"org-right\" \/><\/colgroup><thead><tr><th scope=\"col\" class=\"org-left\">Form<\/th>\n<th scope=\"col\" class=\"org-left\">x faster than next<\/th>\n<th scope=\"col\" class=\"org-right\">Total runtime<\/th>\n<th scope=\"col\" class=\"org-right\"># of GCs<\/th>\n<th scope=\"col\" class=\"org-right\">Total GC runtime<\/th>\n<\/tr><\/thead><tbody><tr><td class=\"org-left\">regexp<\/td>\n<td class=\"org-left\">61.97<\/td>\n<td class=\"org-right\">0.042440<\/td>\n<td class=\"org-right\">0<\/td>\n<td class=\"org-right\">0<\/td>\n<\/tr><tr><td class=\"org-left\">org-map-entries<\/td>\n<td class=\"org-left\">slowest<\/td>\n<td class=\"org-right\">2.630019<\/td>\n<td class=\"org-right\">0<\/td>\n<td class=\"org-right\">0<\/td>\n<\/tr><\/tbody><\/table><p>\nIt can also help catch bugs by ensuring that each form returns the same results.  For example, the benchmark above contains a subtle bug: because <code>case-fold-search<\/code> in the <code>regexp<\/code> form is non-nil, the regexp is compared case-insensitively, so it matches Org headings which start with <code>Maybe<\/code> rather than only ones which start with <code>MAYBE<\/code>.  Using the <code>:ensure-equal t<\/code> argument to <code>bench-multi<\/code> compares the results and raises an error showing the difference between the two sequences the forms evaluate to:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-elisp\">(bench-multi :ensure-equal t\n  :forms ((\"org-map-entries\" (sort (org-map-entries (lambda ()\n                                                      (nth 4 (org-heading-components)))\n                                                    \"\/+MAYBE\" 'agenda)\n                                   #'string&lt;))\n          (\"regexp\" (sort (-flatten\n                           (-non-nil\n                            (mapcar (lambda (file)\n                                      (let ((case-fold-search t))\n                                        (with-current-buffer (find-buffer-visiting file)\n                                          (org-with-wide-buffer\n                                           (goto-char (point-min))\n                                           (cl-loop with regexp = (format org-heading-keyword-regexp-format \"MAYBE\")\n                                                    while (re-search-forward regexp nil t)\n                                                    collect (nth 4 (org-heading-components)))))))\n                                    (org-agenda-files))))\n                          #'string&lt;))))\n<\/pre><\/div><pre class=\"example\">\nuser-error: Forms\u2019 results not equal: difference (regexp - org-map-entries): (\"Maybe this is not the case?\")\n<\/pre><p>\nFixing the error, by setting <code>case-fold-search<\/code> to <code>nil<\/code>, not only makes the forms give the same result but, in this case, doubles the performance of the faster form:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-elisp\">(bench-multi :ensure-equal t\n  :forms ((\"org-map-entries\" (sort (org-map-entries (lambda ()\n                                                      (nth 4 (org-heading-components)))\n                                                    \"\/+MAYBE\" 'agenda)\n                                   #'string&lt;))\n          (\"regexp\" (sort (-flatten\n                           (-non-nil\n                            (mapcar (lambda (file)\n                                      (let ((case-fold-search nil))\n                                        (with-current-buffer (find-buffer-visiting file)\n                                          (org-with-wide-buffer\n                                           (goto-char (point-min))\n                                           (cl-loop with regexp = (format org-heading-keyword-regexp-format \"MAYBE\")\n                                                    while (re-search-forward regexp nil t)\n                                                    collect (nth 4 (org-heading-components)))))))\n                                    (org-agenda-files))))\n                          #'string&lt;))))\n<\/pre><\/div><table border=\"2\" cellpadding=\"6\" rules=\"groups\" frame=\"hsides\" style=\"border-spacing: 0px;\"><colgroup><col class=\"org-left\" \/><col class=\"org-left\" \/><col class=\"org-right\" \/><col class=\"org-right\" \/><col class=\"org-right\" \/><\/colgroup><thead><tr><th scope=\"col\" class=\"org-left\">Form<\/th>\n<th scope=\"col\" class=\"org-left\">x faster than next<\/th>\n<th scope=\"col\" class=\"org-right\">Total runtime<\/th>\n<th scope=\"col\" class=\"org-right\"># of GCs<\/th>\n<th scope=\"col\" class=\"org-right\">Total GC runtime<\/th>\n<\/tr><\/thead><tbody><tr><td class=\"org-left\">regexp<\/td>\n<td class=\"org-left\">134.09<\/td>\n<td class=\"org-right\">0.019550<\/td>\n<td class=\"org-right\">0<\/td>\n<td class=\"org-right\">0<\/td>\n<\/tr><tr><td class=\"org-left\">org-map-entries<\/td>\n<td class=\"org-left\">slowest<\/td>\n<td class=\"org-right\">2.621377<\/td>\n<td class=\"org-right\">0<\/td>\n<td class=\"org-right\">0<\/td>\n<\/tr><\/tbody><\/table><p>\nSo this macro showed which code is faster and helped catch a subtle bug.\n<\/p>","created_at":"2019-03-08T23:56:11+0000","updated_at":"2019-03-08T23:56:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"alphapapa.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":870,"title":"Game Streaming - Play Games With Friends | Parsec","url":"https:\/\/parsecgaming.com\/","content":"Game Streaming - Play Games With Friends | Parsec\n<section class=\"page hero on-index light-text js-page\"><section class=\"page coplay reverse js-page\"><div class=\"column coplay-copy\">Play Games with Your Friends No Matter Where They Live<h2>Forget what you know about screen sharing, Netplay, and difficult-to-use VPNs.<\/h2><p class=\"page-copy\">Parsec lets you invite your friends to play games with you online even if the game doesn't have online multiplayer. You can play together, watch them play, or take over the controller to show them how it's done.<\/p><\/div><div class=\"column\"><img class=\"hero-image\" src=\"https:\/\/s3.amazonaws.com\/parsec-static\/img\/controllers.svg\" alt=\"image\" \/><\/div><\/section><section class=\"page performance light-text js-page\"><div class=\"column performance-copy\">Matchup with friends or meet new people<h2>Low Latency Online Multiplayer<\/h2><p class=\"page-copy\">You might get falcon kicked, but it won't be because of lag or frame drops. Play local multiplayer games online, watch your friends beat that final boss, or pass the controller with Parsec.<\/p><p><a href=\"https:\/\/parsecgaming.com\/signup\" class=\"button\">Signup Now<\/a><\/p><\/div><p><\/p><\/section><section class=\"page cloud light-text js-page js-asteroids\"><div class=\"column\">Matchmaking For Any Game<h2>Head to Party Finder to discover new games to play<\/h2><p class=\"page-copy\">You can jump into any game and meet new people to play games with through Parsec. It's just like heading to your local arcade and finding someone to matchup against in Street Fighter.<\/p><p><a href=\"https:\/\/parsecgaming.com\/parties\" class=\"button\">Join A Party<\/a><\/p><\/div><div class=\"column\"><img class=\"column-image\" src=\"https:\/\/s3.amazonaws.com\/parsec-static\/img\/ship.svg\" alt=\"image\" \/><\/div><\/section><div class=\"start light-text\"><p>Ready to start gaming?\n<\/p><\/div>\n <noscript>\n<\/noscript><\/section>","created_at":"2019-03-06T17:21:56+0000","updated_at":"2019-03-06T17:21:56+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"parsecgaming.com","preview_picture":"https:\/\/s3.amazonaws.com\/parsec-static\/img\/parsec-social.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":869,"title":"Model Metropolis","url":"https:\/\/logicmag.io\/06-model-metropolis\/","content":"<h3 id=\"by-kevin-t-baker\">by Kevin T. Baker<\/h3><p><em>Behind one of the most iconic computer games of all time is a theory of how cities die\u2014one that has proven dangerously influential.<\/em><\/p><figure class=\"kg-card kg-image-card\"><img src=\"https:\/\/logicmag.io\/content\/images\/2019\/01\/simcity-1.png\" class=\"kg-image\" alt=\"image\" \/><\/figure><p>In 1984, the developer Will Wright had just finished work on his first video game, a shoot-em-up called <em>Raid on Bungeling Bay<\/em>. In it, the player controls a helicopter dropping bombs on enemy targets on a series of islands. Wright was happy with the game, which was a commercial and critical success, but even after it was released, he continued tinkering with the terrain editor he had used to design Raid\u2019s levels. \u201cI found out,\u201d Wright later told the <em>Onion AV Club<\/em>, \u201cthat I was having a lot more fun doing that part than just playing the game and going around bombing stuff.\u201d Enthralled by the islands he was making, Wright kept adding features to his level editor, adding complex elements like cars, people, and houses. He became fascinated with the idea of making these islands behave more like cities, and kept tinkering with ways to make the world \u201ccome alive and be more dynamic.\u201d <\/p><p>Looking to understand how real cities worked, Wright came across a 1969 book by Jay Forrester called <em>Urban Dynamics<\/em>. Forrester was an electrical engineer who had launched a second career as an expert on computer simulation; <em>Urban Dynamics<\/em> deployed his simulation methodology to offer a controversial theory of how cities grew and declined. Wright used Forrester\u2019s theories to transform the cities he was designing in his level editor from static maps of buildings and roads into vibrant models of a growing metropolis. Eventually, Wright became convinced that his \u201cguinea-pig city\u201d was an entertaining, open-ended video game. Released in 1989, the game became wildly popular, selling millions of copies, winning dozens of awards, and spawning an entire franchise of successors and dozens of imitators. It was called <em>SimCity<\/em>.<\/p><p>Almost as soon as <em>SimCity<\/em> came out, journalists, academics, and other critics began to speculate on the effects that the game might have on real-world planning and politics. Within a few years of its release, instructors at universities across the country began to integrate <em>SimCity<\/em> into their urban planning and political science curriculums. Commentators like the sociologist Paul Starr worried that the game\u2019s underlying code was an \u201cunreachable black box\u201d which could \u201cseduce\u201d players into accepting its assumptions, like the fact that low taxes promoted growth in this virtual world. \u201cI became a total Republican playing this game,\u201d one <em>SimCity<\/em> fan told the <em>Los Angeles Times<\/em> in 1992. \u201cAll I wanted was for my city to grow, grow, grow.\u201d <\/p><p>Despite all this attention, few writers looked closely at the work which sparked Wright\u2019s interest in urban simulation in the first place. Largely forgotten now, Jay Forrester\u2019s <em>Urban Dynamics<\/em> put forth the controversial claim that the overwhelming majority of American urban policy was not only misguided but that these policies aggravated the very problems that they were intended to solve. In place of Great Society-style welfare programs, Forrester argued that cities should take a less interventionist approach to the problems of urban poverty and blight, and instead encourage revitalization indirectly through incentives for businesses and for the professional class. Forrester\u2019s message proved popular among conservative and libertarian writers, Nixon Administration officials, and other critics of the Great Society for its hands-off approach to urban policy. This outlook, supposedly backed up by computer models, remains highly influential among establishment pundits and policymakers today.<\/p><h3 id=\"150-equations-200-parameters\">150 Equations, 200 Parameters<\/h3><p>Jay Wright Forrester was one of the most important figures in the history of computing, but he is also one of the least understood. He trained at Gordon Brown\u2019s Servomechanisms Laboratory at MIT, spending World War II designing automatic stabilizers for the U.S. Navy\u2019s radars. After the war, he led the development of the Whirlwind computer, arguably the most important computer project of the early postwar period. This machine, after humble beginnings as a flight simulator, morphed into a general-purpose computer which stood at the heart of the Semi-Automatic Ground Environment (SAGE), a multibillion-dollar network of computers and radars that promised to computerize the U.S. Air Force\u2019s response to a Soviet nuclear attack by streamlining the detection of incoming bombers and automatically deploying fighters to intercept them.<\/p><p>In 1956, with the SAGE system not yet finished, Forrester abruptly changed careers, shifting his gaze from electronic systems to human ones. From the unlikely setting of the MIT Sloan School of Management, he founded a discipline called \u201cindustrial dynamics\u201d (later rechristened \u201csystem dynamics\u201d). At first, this field focused on creating computer simulations of production and distribution problems in industrial firms. But Forrester and his cadre of graduate students later expanded it into a general methodology for understanding social, economic, and environmental systems. The most famous example of this group\u2019s work was the \u201cdoomsday\u201d World 3 model that stood at the center of the landmark environmental text <em>The Limits to Growth<\/em>, a book that warned of a potential collapse of industrial civilization by 2050.<\/p><p><em>Urban Dynamics<\/em> was Forrester\u2019s first attempt to apply his methodology outside of the corporate boardroom. He came up with the idea of tackling the problems of cities after meeting John F. Collins, a conservative Democratic politician and the outgoing mayor of Boston, who had recently taken a position at MIT. Listening to Collins\u2019s stories about his time as mayor in the 1960s, Forrester became convinced industrial dynamics could be used to study the poverty and capital flight associated with the US\u2019s ongoing \u201curban crisis.\u201d Even though Forrester had no expertise in urban affairs\u2014or in social science more generally\u2014Collins agreed that a collaboration could prove fruitful.<\/p><p>Over the course of 1968, Forrester devoted about twenty-five hours a week to his project with Collins. During that time, he met with the former mayor and his team of advisors and constructed a sprawling flowchart representing the relationships between different aspects of the city\u2019s structure. Forrester translated this flowchart into the group\u2019s house simulation language, DYNAMO. After a staff secretary or a graduate student punched the DYNAMO equations onto cards, it could be loaded on a machine. From there, the computer would generate a running version of the model, and output line-plots and tabular data representing the decade-by-decade evolution of the simulated city. <\/p><p>Forrester spent months tinkering with this model, tested and corrected it for errors, and ran a \u201chundred or more system experiments to explore the effects of various policies on the revival of a city that has aged into economic decline.\u201d Six months after beginning the project, and over 2000 pages of teletype printouts later, Forrester declared that he had reduced the problems of the city to a series of 150 equations and 200 parameters.<\/p><h3 id=\"death-spirals\">Death Spirals<\/h3><p>At the start of a standard 250-year run in Forrester\u2019s model, the simulated city is empty. No land is occupied, there is no economic activity, and there is little incentive for construction. As the city gradually develops, increases in housing, population, and industry all serve to reinforce one another, and the city enters a period of sustained economic and population growth. During this period, people are attracted to the city, and new housing and businesses are constructed at a brisk pace.<\/p><p>But as the city ages and its land area reaches full occupancy, growth slows. Areas which were considered \u201cattractive and useful\u201d have already been occupied. New construction takes place on more marginal land, and because this land is less attractive, the pace of construction slows. When no greenfield parcels of land are left for development, new construction becomes impossible, and new housing and firms can only be built when old ones are demolished. New migrants, once a boon to the city\u2019s industry, continue to flow into the metropolis, causing overcrowding and underemployment, dampening economic vitality, and sending the city into a death spiral of blight and decay.<\/p><p>The arc of this story reflected the simplified, and sometimes entirely fictional, assumptions of Forrester\u2019s model. On its most basic level, <em>Urban Dynamics<\/em> modeled the relationship between population, housing stock, and industrial buildings against a background of government policies. The city inside Forrester\u2019s model was a highly abstracted one. There were no neighborhoods, no parks, no roads, no suburbs, and no racial or ethnic conflicts. (In fact, the people inside the model didn\u2019t belong to racial, ethnic, or gender categories at all.) Economic and political life in the outside world had no effect on the simulated city. To the extent that the world outside the model existed, it served only as a source for migrants into the city, and a place for them to flee to if the city became inhospitable.<\/p><p>The residents of Forrester\u2019s simulated city belonged to one of three class categories, \u201cmanagerial-professional,\u201d \u201cworker,\u201d and \u201cunderemployed.\u201d As one moved down the class ladder in the urban dynamics model, classist assumptions about the urban poor piled up: birth rates were higher, tax contributions were lower, and the use of public expenditures increased. This meant that the urban poor served as a massive drag on the health of the simulated city: they did not add to its economic life, they had large families which strained public services, and they contributed only paltry amounts to the city\u2019s coffers.<\/p><p>Forrester was cagey about how much this mapped onto real life. He cautioned that his model was a \u201cmethod of analysis\u201d and that it was unwise to take its conclusions as applicable without first ensuring that the model's assumptions fit a specific city\u2019s situation. At the same time, Forrester used the simulation as a stand-in for cities in general, making sweeping claims about the failure of what he regarded as \u201ccounterproductive\u201d urban policies.<\/p><p>To Forrester, low-income housing was an especially egregious example of a \u201ccounterproductive\u201d urban program. According to the model, these programs increased the local tax burden, attracted underemployed people into the city, and occupied land which might otherwise have been put to more economically healthy uses. Housing programs aimed at improving the condition of the underemployed, Forrester warned, \u201cincreased unemployment and reduced upward economic mobility\u201d and would condemn the underemployed to lifelong poverty. This idea wouldn\u2019t have seemed new to anyone steeped in the conservative or libertarian tradition, but Forrester\u2019s technical approach helped update it for the digital age.<\/p><h3 id=\"the-perversity-thesis\">The Perversity Thesis<\/h3><p>When we consider the social effects of computers in political and social life, we usually think in terms of expanded power and new possibilities. This perspective on computation permeates even our critical visions of technology. But we should also be attentive to the power that computers and the accompanying language of \u201csystems\u201d and \u201ccomplexity\u201d have to narrow our conception of the politically possible.<\/p><p>Forrester thought that the basic problem of urban planning\u2014and making social policy in general\u2014was that \u201cthe human mind is not adapted to interpreting how social systems behave.\u201d In a paper serialized in two early issues of <em>Reason<\/em>, the libertarian magazine founded in 1968, Forrester argued that for most of human history, people have only needed to understand basic cause-and-effect relationships, but that our social systems are governed by complex processes that unfold over long periods of time. He claimed that our \u201cmental models,\u201d the cognitive maps we have of the world, are ill-suited to help us navigate the web of \u00a0interrelationships that make up the structure of our society. <\/p><p>For him, this complexity meant that policy interventions could, and usually would, have very different social effects than those imagined by policymakers. This led him to make the stark assertion that \u201cthe intuitive solutions to the problems of complex social systems\u201d are \u201cwrong most of the time.\u201d In essence, anything we do to try to improve society will backfire and make things even worse.<\/p><p>In this respect, Forrester\u2019s approach to the problems of American cities mirrored the \u201cbenign neglect\u201d outlook of influential Nixon adviser Daniel Patrick Moynihan and the rest of the administration. Indeed, Moynihan was an enthusiastic proponent of Forrester\u2019s work and recommended <em>Urban Dynamics<\/em> to his fellow White House officials. Forrester\u2019s arguments enabled the Nixon Administration to claim that its plans to <em>slash<\/em> programs created to help the urban poor and people of color would actually, counterintuitively, <em>help<\/em> these people.<\/p><p>Forrester\u2019s central claim about complexity wasn\u2019t a new one; it has a long history on the political right. In a 1991 book, <em>Rhetoric of Reaction<\/em>, the development economist and economic historian Albert O. Hirschman identified this style of argument as an example of what he called the \u201cperversity thesis.\u201d This kind of attack, which Hirschman traced back to Edmund Burke\u2019s writings on the French Revolution, amounts to a kind of concern trolling. Using this rhetorical tactic, the conservative speaker can claim that they share your social goal, but simultaneously argue that the means you are using to achieve it will only make matters worse. When commentators claim \u201cno-platforming will only make more Nazis,\u201d that welfare programs lock recipients into a \u201ccycle of dependency,\u201d or that economic planning will lead a society down a \u201croad to serfdom,\u201d they\u2019re making this kind of perversity argument.<\/p><p>What Forrester did was give the perversity thesis a patina of scientific and computational respectability. Hirschman himself makes specific reference to <em>Urban Dynamics<\/em> and argues that the \u201cspecial, sophisticated attire\u201d of Forrester\u2019s models helped reintroduce this kind of argument \u201cinto polite company.\u201d In the nearly fifty years since it has come out, Forrester\u2019s \u201ccounterintuitive\u201d style of thinking has become the default way of analyzing policy for mainstream wonks. For many, \u201ccounterintuitivity\u201d is the new intuition.<\/p><p>Expert knowledge, of course, has an important place in democratic deliberation, but it can also cut people out of the policy process, dampen the urgency of moral claims, and program a sense of powerlessness into our public discourse. Appeals to a social system\u2019s \u201ccomplexity\u201d and the potential for \u201cperverse outcomes\u201d can be enough to sink transformative social programs that are still on the drawing board. This might not matter in the context of a virtual environment like that of \u00a0<em>Urban Dynamics<\/em> or <em>SimCity<\/em>, but we have decades of real-world evidence that demonstrates the disastrous costs of the \u201ccounterintuitive\u201d anti-welfare agenda. Straightforward solutions to poverty and economic misery\u2014redistribution and \u00a0the provision of public services\u2014have both empirical backing and moral force. Maybe it\u2019s time we start listening to our intuition again. \u00a0<\/p><p><em>Kevin T. Baker is historian of science who works on the history of prediction, computer simulation, and environmental and economic modeling.<\/em><\/p><hr \/><p><em>This piece appears in<\/em> Logic'<em>s sixth issue, \"Play.\" To order the issue, head on over to our <a href=\"https:\/\/store.logicmag.io\/\">store<\/a>. To receive future issues, <a href=\"https:\/\/logicmag.io\/subscribe\">subscribe<\/a>.<\/em><\/p><hr \/><p><a href=\"https:\/\/logicmag.io\/06-play\/\">&lt; Back to the Table of Contents<\/a><\/p>","created_at":"2019-03-06T03:52:15+0000","updated_at":"2019-03-06T03:52:15+0000","published_at":"2019-01-30T06:36:05+0000","published_by":["The Editors"],"annotations":[],"mimetype":"text\/html","reading_time":12,"domain_name":"logicmag.io","preview_picture":"https:\/\/logicmag.io\/content\/images\/2019\/01\/simcity-2.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["datascience"],"is_public":false,"id":868,"title":"Data Science Getting Started Guide","url":"https:\/\/www.datascienceweekly.org\/data-science-guides\/data-science-getting-started-guide","content":"<p class=\"quote_start\">\n      You ask:\n      <br \/>\"How do I get a data science job?\"\n    <\/p><p class=\"quote_start\">\n      You get:\n      <br \/>\"First you have to learn: Linear Algebra, Convex Optimization, Differential Equations, Calculus, Algorithms, Distributed Computing, Databases (SQL &amp; NoSQL), Machine Learning, Probabilistic Modeling, Deep Learning, Natural Language Processing, Data Visualization, and  don't forget Scala for functional programming, and Hadoop, and Big Data, and Experimental Design, and Functional Analysis, and......\"\n    <\/p><p class=\"quote_start\">\n      Your eyes glaze over as you see yourself going back to a dimly-lit lecture hall for the rest of your life.\n    <\/p><hr \/><br \/><p class=\"content-paragraph\">\n      You just read yet another a list of books you MUST COMPREHEND, MOOC's you MUST TAKE, and statistical programming libraries you MUST MASTER to get started in data science.\n      <\/p><p>With the plethora of information online, it's immensely unmotivating, overwhelming, and perplexing to navigate a huge list of resources without any context for how they fit together, why they matter, or how they will help you get a data science job.\n      <\/p><p>Do you really need to know statistical learning theory and functional analysis at a graduate school level to get a data science job?\n      <\/p><p>You end up spending more time debating what to learn than actually making progress.\n      <\/p><p>Given your specific educational background and work experience, you still don't know what knowledge gaps you have, which you need to close, and which you can ignore.\n      <\/p><p>You don't have time, money, or patience to waste learning things that don't matter.\n      <\/p><p>A data science job (with its high profile and lucrative salary) seems more and more out of your reach with each passing day.\n      <\/p><p>You wish someone would put you our of your misery by telling you specifically what you need to learn (down to the specific key words and topics), how to put together your data science portfolio, and how to best write your resume to become a data scientist.\n    <\/p><br \/><p class=\"content-paragraph\">\n      You would be 100% confident on which MOOC's to take, which books to read, and which statistical programming libraries to use.\n      <\/p><p>You'd know the highly specific knowledge gaps you must close given your educational and professional background.\n      <\/p><p>Every data science project you completed would make hiring managers sit up and notice due to your topic choices, how you communicate, and what you accomplished.\n      <\/p><p>Your resume would be at the top of the \"must-talk-to-candidates\" pile for every data science job you applied to.\n      <\/p><p>Your money, time, and effort would be spent wisely and to great effect in attaining your dream data science job.\n      <\/p><p>You'd be motivated, sure, and enthusiastic about having the right data science position within your reach.\n      <\/p><p>You'd be a high-profile data scientist with a lucrative salary in no time!\n    <\/p><br \/><p class=\"content-paragraph\">\n      Done are the days of sitting at a table being surrounded by books on books on books and endless amounts of data science resources.\n      <\/p><p>Throw out that giant stack of textbooks.\n      <\/p><p>You don't have to learn most of the stuff people say you need to learn.\n      <\/p><p>The <strong><em>Get A Data Science Job Course<\/em><\/strong>\n      contains three guides you will love because of how much they will help you:\n    <\/p><ol><li>\u00a0\u00a0Data Science Getting Started Guide<\/li>\n        <li>\u00a0\u00a0Data Science Project Portfolio Guide<\/li>\n        <li>\u00a0\u00a0Data Science Resume Guide<\/li>\n      <\/ol><p class=\"content-paragraph\">\n      Each guide will introduce you to tons of fail-safe processes you can follow to the letter that will help you get that data science job as fast as possible.\n      <\/p><p>The first guide, <strong><em>Data Science Getting Started Guide<\/em><\/strong>, will teach you:\n    <\/p><ul><li>How to figure out the knowledge gaps that MUST be closed in order for your to become a data scientist<\/li>\n        <li>How to develop your perfectly personalized \"learning data science\" action plan<\/li>\n        <li>What kind of data scientist you are<\/li>\n        <li>How to work backwards from a data science job to your current situation to build your plan<\/li>\n        <li>Which knowledge gaps matter and which you can safely ignore<\/li>\n        <li>Which \"become a data scientist\" resources you can trust<\/li>\n        <li>How to build the right data science skill sets from day one so that you don't waste time or effort<\/li>\n      <\/ul><p class=\"content-paragraph\">\n      <br \/>The second guide, <strong><em>Data Science Project Portfolio Guide<\/em><\/strong>, will teach you:\n    <\/p><ul><li>What A Data Science Project Portfolio Is And What Does It Do<\/li>\n        <li>How To Build Your Portfolio Quickly And Effectively With The Right Goals And Direction<\/li>\n        <li>How To Structure Your Data Science Project Portfolio Landing Page<\/li>\n        <li>How To Choose The Right Projects For Your Portfolio<\/li>\n        <li>Project Specific Write-up For Your Data Science Project Portfolio<\/li>\n      <\/ul><p class=\"content-paragraph\">\n      <br \/>The third guide, <strong><em>Data Science Resume Guide<\/em><\/strong>, will teach you:\n    <\/p><ol><li>\n        Figuring Out What\u2019s Been Going Wrong\n        <ul><li>What\u2019s the purpose of your Data Science Resume<\/li>\n          <li>Six Ways Your Resume Could be Sabotaging You<\/li>\n        <\/ul><\/li>\n      <li>\n        Reading the Hiring Manager\u2019s Mind\n        <ul><li>How to Figure Out What the Job Posting Means<\/li>\n          <li>How to Research Key Issues Facing the Company<\/li>\n          <li>What Are Data Scientists Doing There Today?<\/li>\n          <li>How to Assess the Hiring Manager\u2019s Real HOPES<\/li>\n        <\/ul><\/li>\n      <li>\n        Identifying YOU as the Perfect Candidate\n        <ul><li>How to Identify Your Complete Set of Strengths<\/li> \n          <li>How to Communicate Your Unique SPEC <\/li>\n        <\/ul><\/li>\n      <li>\n        Making Your Resume Look the Part\n        <ul><li>What Sections to Include (and Not) <\/li>\n          <li>How to Order Your Resume (Case-Specific)<\/li>\n          <li>How to Improve Your Format (including Templates)<\/li>\n          <li>Top 10 Formatting FAQs <\/li>\n        <\/ul><\/li>\n      <li>\n        Doing Your Skills Justice\n        <ul><li>3 Principles for How to Structure the Skills section<\/li>\n          <li>How to Identify What Skills to List (and Not List)<\/li> \n          <li>Which Projects to include in the Projects section <\/li>\n          <li>How to describe your Project Work <\/li>\n        <\/ul><\/li>\n      <li>\n        Making Your Experience Come to Life\n        <ul><li>What Experiences to List  (and Not List) <\/li>\n          <li>How to Structure the Experiences Section <\/li>\n        <\/ul><\/li>\n      <li>\n        Showcasing Your Education\n        <ul><li>What Education\/Courses to List (and Not List) <\/li>\n          <li>Guidelines for How Much Detail to Share<\/li>\n        <\/ul><\/li>\n      <li>\n        Nailing the Cover letter\n        <ul><li>What Template to Follow<\/li>\n          <li>How to Tailor Your Cover Letter<\/li>\n          <li>How to Say Something New <\/li>\n          <li>How to Toot Your Own Horn<\/li>\n          <li>How to Make it Readable (by a 10th Grader)<\/li>\n          <li>How to Write to Someone (Ideally Hiring Manager)<\/li>\n        <\/ul><\/li>\n      <li>\n        Pulling it All Together\n        <ul><li>Three Proven Techniques for Self-Review<\/li>\n          <li>7 Questions to Guide Peer Review <\/li>\n          <li>Final, Pre-Submission Checklist <\/li>\n        <\/ul><\/li>\n      <li>\n        Hitting Submit with Confidence\n      <\/li>\n    <\/ol><p class=\"content-paragraph\">\n      <br \/>The <strong><em>Get A Data Science Job Course<\/em><\/strong> will direct you in constructing your own highly personalized plan for what you need to learn and what you can safely ignore - saving you time, effort, and worry.\n    <\/p><p>And most importantly, the <strong><em>Get A Data Science Job Course<\/em><\/strong> will give you the straight truth of what matters and what you can ignore to achieve your dream of becoming a data scientist with a high profile and lucrative salary faster than you've ever imagined.\n      <\/p><p>It will be like future you (an awesome data scientist) travelled back in time to tell current you all the right things to do and focus on.\n      <\/p><p>Become a data scientist faster than you think possible!\n    <\/p><br \/><div class=\"col-lg-12\"><div class=\"well\"><p class=\"content-paragraph\">\n        How would you feel if you woke up tomorrow with perfect knowledge, given your background and education, for how to become a data scientist as fast as possible?\n        <\/p><p>Entry level Data Scientists earn $80-$100k per year. The average US Data Scientist makes $118K. Some Senior Data Scientists make between $200,000 to $300,000 per year.\n        <\/p><p>Kickstart your data science career today for the cost of a nice dinner. Just <strong><em>$9.99<\/em><\/strong> gives you THREE guides that show you tons of fail-safe processes to figure out specifically what you need to learn (down to the specific key words and topics), how to construct your data science portfolio, and how to best write your resume to become a data scientist.\n        <\/p><p>Download instantly, read on any device you have, and set yourself up for data science success.\n        <\/p><p><a href=\"https:\/\/datascienceweekly.dpdcart.com\/cart\/add?product_id=177618&amp;method_id=191676\" rel=\"nofollow\" type=\"button\" class=\"btn btn-danger btn-lg\">\t\t\n        $9.99<br \/>Buy the \"Get A Data Science Job Course\" and Start Today\t\t\n        <\/a>\n        <\/p><p><strong>Buy with confidence and without fear<\/strong>. This guide was written to help you. If you're not happy, email <a href=\"mailto:team@datascienceweekly.org\">team@datascienceweekly.org<\/a> and you'll receive a refund.\n      <\/p><\/div><p class=\"content-paragraph\">\n      <img alt=\"Data Science Getting Started Guide Author - Sebastian Gutierrez\" class=\"img-rounded\" src=\"https:\/\/s3.amazonaws.com\/datascienceweekly\/data-science-website-images\/sebastian_gutierrez_271x296.png\" width=\"271\" height=\"296\" \/><\/p><p>Sebastian Gutierrez is the author of \"Data Scientists at Work\", an interview-based book covering the background, work, and thoughts of 16 of the world's leading data scientists. Sebastian co-runs the Data Science Weekly email newsletter which proudly reaches ~24k subscribers a week. Sebastian is one of four moderators for the Data Science subreddit whose community numbers ~30k members. Sebastian has spoken in Data-related conferences in the US, the UK, and Europe. Sebastian Gutierrez holds a BS in Mathematics from MIT and an MA in Economics from the University of San Francisco.\n    <\/p><p class=\"content-paragraph\">\n      You are awesome and you will be an awesome data scientist. Get started today!\n    <\/p><\/div>","created_at":"2019-03-01T15:02:53+0000","updated_at":"2019-03-01T15:03:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"www.datascienceweekly.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":867,"title":"r\/orgmode - Org-babel with python virtual environments","url":"https:\/\/www.reddit.com\/r\/orgmode\/comments\/4nffsk\/orgbabel_with_python_virtual_environments\/","content":"<p><a class=\"s1bd5ppi-4 imJoxG\" href=\"https:\/\/www.reddit.com\/r\/orgmode\/\">r\/orgmode<\/a><\/p><div class=\"s1bd5ppi-8 gjbPXG\"><div class=\"s1bd5ppi-9 lgpBtg\"><p class=\"s1bd5ppi-10 jmHofS\">9<\/p><p class=\"s1bd5ppi-11 dAmide\" id=\"IdCard--CurrentlyViewing--undefined--t5_2s5ak\">Online<\/p><\/div><\/div><p class=\"s1bd5ppi-13 bczYis\">[Org-mode](http:\/\/orgmode.org\/) is an Emacs major mode for taking notes, managing to-do lists, planning projects, authoring documents, literate programming and devops, and more, using a fast and effective plain-text system.<\/p><p><a class=\"s1bd5ppi-16 jjLcIO s1f6o7cs-2 eqioci\" href=\"https:\/\/www.reddit.com\/r\/orgmode\/submit\">Create Post<\/a><\/p>","created_at":"2019-03-01T13:41:27+0000","updated_at":"2019-03-01T13:41:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.reddit.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":866,"title":"org mode - Using org-babel to enable virtualenv - Emacs Stack Exchange","url":"https:\/\/emacs.stackexchange.com\/questions\/24267\/using-org-babel-to-enable-virtualenv","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I hoped to enable a virtualenv with org-babel source block before executing some python. <\/p>\n<p>I wanted to do something like <a href=\"https:\/\/github.com\/jorgenschaefer\/pyvenv\" rel=\"nofollow\">pyenv<\/a> pyvenv-workon before it evaulates my code block. I didn't want to have to manually enable the virtualenv.<\/p>\n<p>Something like,<\/p>\n<pre>#+NAME: elasticsearch_dsl_test\n#+BEGIN_SRC: python :prologue \"(pyvenv-workon 'elasticsearch)\"\nfrom elasticsearch import Elasticsearch\nfrom elasticsearch_dsl import Search, Q\nclient = Elasticsearch()\nsearch = Search(using=client, index=\"localdev-news-newsentry\")\nresponse = search.execute()\nprint [\"No response\", \"Was successfull\"][int(response.success())]\nprint \"Response took {0}ms.\".format(response.took)\n#+END_SRC\n<\/pre>\n<p>I couldn't quite get it working.<\/p><\/div><h2 data-answercount=\"1\">&#13;\n                                1 Answer&#13;\n                            1&#13;<\/h2>","created_at":"2019-03-01T13:38:55+0000","updated_at":"2019-03-01T13:38:55+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"emacs.stackexchange.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/emacs\/img\/apple-touch-icon@2.png?v=dbead6deb8b0","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":864,"title":"Welcome to the Mulgara Project!","url":"http:\/\/www.mulgara.org\/","content":"<div id=\"columnLeft\"><p>Mulgara hasn't been getting a lot of love from the devs lately, but we're still here. Paula has been starting to port elements over to Clojure, with the intention of reimplementing everything in that language. See <a href=\"https:\/\/github.com\/quoll\/kiara\">Kiara<\/a>, <a href=\"https:\/\/github.com\/quoll\/crg-turtle\">crg-turtle<\/a> and <a href=\"https:\/\/github.com\/quoll\/crg-turtle\">tree<\/a> for recent developments there. Meanwhile, any updates needed in the Mulgara code can still be requested. The codebase has been moved to Github to help facilitate this, as well as avoid potential problems with the mulgara.org server.February 10, 2014<\/p><p><a href=\"http:\/\/www.mulgara.org\/news.html#date20120110\">Mulgara 2.1.13 released<\/a>. A minor release to meet government auditing requirements. See the <a href=\"http:\/\/www.mulgara.org\/release.html#v2.1.13\">release notes<\/a>. <a href=\"http:\/\/www.mulgara.org\/download.html\">Download from here<\/a>. - January 10, 2012<\/p><p><a href=\"http:\/\/www.mulgara.org\/news.html#date20111214\">Mulgara 2.1.12 released<\/a>. A minor release with the introduction of transitive predicates in SPARQL. See the <a href=\"http:\/\/www.mulgara.org\/release.html#v2.1.12\">release notes<\/a>. <a href=\"http:\/\/www.mulgara.org\/download.html\">Download from here<\/a>. - December 14, 2011<\/p><p><a href=\"http:\/\/www.mulgara.org\/news.html#date20110929\">Mulgara 2.1.11 released<\/a>. This has been a cleanup and minor bugfix release to help with commercial deployment. See the <a href=\"http:\/\/www.mulgara.org\/release.html#v2.1.11\">release notes<\/a>. <a href=\"http:\/\/www.mulgara.org\/download.html\">Download from here<\/a>. - September 29, 2011<\/p><p><a href=\"http:\/\/www.mulgara.org\/news.html#date20110719\">After some time away, Mulgara v2.1.10 has been released<\/a>. This release deals with bug related to rolling back large transactions using the XA11 string pool, and some rollback errors that occur exclusively in Windows. See the <a href=\"http:\/\/www.mulgara.org\/release.html#v2.1.10\">release notes<\/a>. <a href=\"http:\/\/www.mulgara.org\/download.html\">Download from here<\/a>. - July 19, 2011<\/p><br \/>Updated by<a href=\"https:\/\/twitter.com\/quoll\">Paula<\/a>on February 14, 2014<\/div><div id=\"columnRight\" style=\"margin: auto;\">Source code on<a href=\"https:\/\/github.com\/quoll\/mulgara\">Github<\/a><\/div>","created_at":"2019-02-27T00:39:50+0000","updated_at":"2019-02-27T00:39:50+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.mulgara.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":865,"title":"threatgrid\/naga","url":"https:\/\/github.com\/threatgrid\/naga","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Datalog based rules engine.<\/p>\n<p><a href=\"http:\/\/clojars.org\/org.clojars.quoll\/naga\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/651bd985aa1f76561842c50e0cfeb040223b6946\/687474703a2f2f636c6f6a6172732e6f72672f6f72672e636c6f6a6172732e71756f6c6c2f6e6167612f6c61746573742d76657273696f6e2e737667\" alt=\"Clojars Project\" data-canonical-src=\"http:\/\/clojars.org\/org.clojars.quoll\/naga\/latest-version.svg\" \/><\/a><\/p>\n<p>Naga allows users to load data, and define rules to entailed new data. Once rules have been\nexecuted, the database will be populated with new inferences which can be queried.<\/p>\n<p>Naga can use its internal database, or wrap an external graph database. A command line\nutility to demonstrate Naga will load rules, into memory, run them, and print all the\ninferred results.<\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>To run Naga on a Datalog script, provide the script via stdin or in a filename argument.<\/p>\n<p>While Naga is still in development, the easiest way to run it is with the\n<a href=\"http:\/\/leiningen.org\" rel=\"nofollow\">Leiningen<\/a> build tool.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>  lein run pabu\/family.lg<\/pre><\/div>\n<p>This runs the program found in the <code>family.lg<\/code> file. It loads the data specified in the file,\nexecutes the rules, and finally prints the final results database, without the input data.<\/p>\n<h3><a id=\"user-content-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#options\"><\/a>Options<\/h3>\n<ul><li><code>--init<\/code> Initialization data for the configured storage.<\/li>\n<li><code>--json<\/code> Input path\/url for JSON to be loaded and processed.<\/li>\n<li><code>--out<\/code> Output file when processing JSON data (ignored when JSON not used).<\/li>\n<li><code>--uri<\/code> URI describing a database to connect to. (default: <code>mem<\/code>. Datomic supported).<\/li>\n<\/ul><h2><a id=\"user-content-programs\" class=\"anchor\" aria-hidden=\"true\" href=\"#programs\"><\/a>Programs<\/h2>\n<p>The language being implemented is called \"Pabu\" and it strongly resembles Prolog. It is\nsimply a series of statements, which are of two types: assertions and rules.<\/p>\n<h3><a id=\"user-content-assertions\" class=\"anchor\" aria-hidden=\"true\" href=\"#assertions\"><\/a>Assertions<\/h3>\n<p>To declare facts, specify a unary or binary predicate.<\/p>\n<pre>  man(fred).\n  friend(fred,barney).\n<\/pre>\n<p>The first statement declares that <code>fred<\/code> is of type <code>man<\/code>. The second declares that <code>fred<\/code> has\na <code>friend<\/code> who is <code>barney<\/code>.<\/p>\n<p>Nothing needs to be declared about <code>man<\/code> or <code>friend<\/code>. The system doesn't actually care what\nthey mean, just that they can be used in this way. The use of these predicates is all the\ndeclaration that they need.<\/p>\n<h3><a id=\"user-content-rules\" class=\"anchor\" aria-hidden=\"true\" href=\"#rules\"><\/a>Rules<\/h3>\n<p>Rules are declared in 2 parts.<\/p>\n<p><em>head<\/em> :- <em>body<\/em> .<\/p>\n<p>The <em>body<\/em> of the rule, defines the data that causes the rule\nto be executed. This is a comma separated series of predicates, each typically containing\none or more variables. The predicate itself can also be variable\n(this is very unusual in logic systems).<\/p>\n<p>The <em>head<\/em> of the rule uses some of the variables in the <em>body<\/em> to declare new information\nthat the rule will create. It is comprised of a single predicate.<\/p>\n<p><em>Variables<\/em> are words that begin with a capital letter (yes, Prolog really does look like this).<\/p>\n<p>Here is a rule that will infer an <code>uncle<\/code> relationship from existing data:<\/p>\n<pre>  uncle(Nibling,Uncle) :- parent(Nibling,Parent), brother(Parent,Uncle).\n<\/pre>\n<p>In the above statement, <em>Nibling<\/em>, <em>Parent<\/em>, and <em>Uncle<\/em> are all variables. Once variables\nhave been found to match the predicates after the <strong>:-<\/strong> symbol, then they can be substituted\ninto the <code>uncle<\/code> predicate in the head of the rule.<\/p>\n<h3><a id=\"user-content-other-syntax\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-syntax\"><\/a>Other Syntax<\/h3>\n<p>Both assertions and rules end with a period.<\/p>\n<p>Pabu (and Prolog) uses \"C\" style comments:<\/p>\n<pre>  \/* This is a comment *\/\n<\/pre>\n<p>Any element can be given a namespace by using a colon separator. Only 1 colon may appear in an identifier.<\/p>\n<pre>  owl:SymmetricProperty(sibling).\n<\/pre>\n<p>To see this in use, look in pabu\/family-2nd-ord.lg, and try running it:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>  lein run pabu\/family-2nd-ord.lg<\/pre><\/div>\n<h2><a id=\"user-content-apis\" class=\"anchor\" aria-hidden=\"true\" href=\"#apis\"><\/a>APIs<\/h2>\n<p>Naga defines a data access API to talk to storage. This is a Clojure protocol or Java interface\ncalled <code>Storage<\/code>, found in <code>naga.store<\/code>. It should be possible to wrap most graph database APIs\nin the <code>Storage<\/code> API.<\/p>\n<p>For the moment, the only configured implementation is an in-memory store.<\/p>\n<h2><a id=\"user-content-in-memory-database\" class=\"anchor\" aria-hidden=\"true\" href=\"#in-memory-database\"><\/a>In Memory Database<\/h2>\n<p>Naga implements its own graph database. It has a relatively capable query planner, and internal\noperations for inner joins and projection. More operations are in the works.<\/p>\n<p>Queries may be executed directly against the database, but for the moment they require API access.\nThis uses the same <code>Storage<\/code> API described above.<\/p>\n<p>We also have some partial implementations for on-disk storage, which we hope to use.\nThese are based on the same architecture as the indexes in the\n<a href=\"http:\/\/github.com\/quoll\/mulgara\">Mulgara Database<\/a>.<\/p>\n<p>This may eventually be split out into its own project.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2016-2017 Cisco Systems<\/p>\n<p>Copyright \u00a9 2011-2016 Paula Gearon<\/p>\n<p>Distributed under the Eclipse Public License either version 1.0 or (at\nyour option) any later version.<\/p>\n<\/article>","created_at":"2019-02-27T00:39:50+0000","updated_at":"2019-02-27T00:39:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/337733?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":863,"title":"Unison File Synchronizer","url":"https:\/\/www.cis.upenn.edu\/~bcpierce\/unison\/","content":"&#13;&#13;    Unison &#13;      is a file-synchronization tool for OSX, Unix, and Windows. It allows two replicas &#13;      of a collection of files and directories to be stored on different hosts &#13;      (or different disks on the same host), modified separately, and then brought &#13;      up to date by propagating the changes in each replica to the other.<p>Unison shares a number of features with tools such as configuration management &#13;      packages (<a href=\"http:\/\/www.cyclic.com\/\">CVS<\/a>, <a href=\"http:\/\/www.XCF.Berkeley.EDU\/~jmacd\/prcs.html\">PRCS<\/a>, &#13;      <a href=\"http:\/\/subversion.tigris.org\/\">Subversion<\/a>, <a href=\"http:\/\/www.bitkeeper.com\/\">BitKeeper<\/a>, &#13;      etc.), distributed filesystems (<a href=\"http:\/\/www.coda.cs.cmu.edu\/\">Coda<\/a>, &#13;      etc.), uni-directional mirroring utilities (<a href=\"http:\/\/samba.anu.edu.au\/rsync\/\">rsync<\/a>, &#13;      etc.), and other synchronizers (<a href=\"http:\/\/www.pumatech.com\">Intellisync<\/a>, &#13;      <a href=\"http:\/\/www.merl.com\/reports\/TR99-14\/\">Reconcile<\/a>, etc). However, &#13;      there are several points where it differs: &#13;      <\/p><ul><li> Unison runs on both Windows and many flavors of Unix (Solaris, Linux, &#13;          OS X, etc.) systems. Moreover, Unison works <em>across<\/em> platforms, &#13;          allowing you to synchronize a Windows laptop with a Unix server, for &#13;          example. &#13;        <\/li><li>Unlike simple mirroring or backup utilities, Unison can deal with &#13;          updates to <i>both<\/i> replicas of a distributed directory structure. &#13;          Updates that do not conflict are propagated automatically. Conflicting &#13;          updates are detected and displayed. &#13;        <\/li><li>Unlike a distributed filesystem, Unison is a user-level program: there &#13;          is no need to modify the kernel or to have superuser privileges on either &#13;          host. &#13;        <\/li><li>Unison works between any pair of machines connected to the internet, &#13;          communicating over either a direct socket link or tunneling over an &#13;          encrypted ssh connection. It is careful &#13;          with network bandwidth, and runs well over slow links such as PPP connections. &#13;          Transfers of small updates to large files are optimized using a compression &#13;          protocol similar to rsync. &#13;        <\/li><li>Unison is resilient to failure. It is careful to leave the replicas &#13;          and its own private structures in a sensible state at all times, even &#13;          in case of abnormal termination or communication failures. &#13;        <\/li><li>Unison has a clear and precise specification. &#13;        <\/li><li>Unison is free; full source code is available under the GNU Public &#13;          License. &#13;      <\/li><\/ul><div style=\"text-align: center;\">&#13;&#13;    <p>&#13;      <a href=\"https:\/\/www.cis.upenn.edu\/~bcpierce\/unison\/download.html\">Download<\/a>&#13;<\/p>&#13;     <table style=\"background-color: #888888;\"><tr><td>&#13;    <table class=\"announce\" style=\"width: 400px;\"><tr><td id=\"announce\">&#13;    <b>Donations<\/b> to&#13;    the Unison project are gratefully accepted.  The link to&#13;    the left leads to a PayPal donation page.<\/td><\/tr><\/table><\/td><\/tr><\/table><\/div> &#13;      &#13;","created_at":"2019-02-26T21:42:34+0000","updated_at":"2019-02-26T21:42:34+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.cis.upenn.edu","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":862,"title":"A Raspberry Pi Grimoire For The Command Line Wizard","url":"https:\/\/hackaday.com\/2019\/02\/25\/a-raspberry-pi-grimoire-for-the-command-line-wizard\/","content":"<p>Who says there\u2019s no such thing as magic? Not anyone who knows what a Unix pipe is, that\u2019s for sure. If you do some of your best incantations at a blinking cursor, this <a href=\"https:\/\/hackaday.io\/project\/164091-spellbook\" target=\"_blank\">scratch-built Raspberry Pi Zero \u201cSpellbook\u201d laptop created by [Calvin]<\/a> might be just what the apothecary ordered. Lucky for us, he was kind enough to document the design and construction of this penguin-powered tome for anyone else who wishes to dabble in the GNU Dark Arts.<\/p>\n<p><a href=\"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg\" target=\"_blank\"><img data-attachment-id=\"346616\" data-permalink=\"https:\/\/hackaday.com\/2019\/02\/25\/a-raspberry-pi-grimoire-for-the-command-line-wizard\/pibook_detail\/\" data-orig-file=\"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg\" data-orig-size=\"954,765\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"pibook_detail\" data-image-description=\"\" data-medium-file=\"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=400&amp;h=321\" data-large-file=\"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=779\" class=\"alignright size-medium wp-image-346616\" src=\"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=400&amp;h=321\" alt=\"\" width=\"400\" height=\"321\" srcset=\"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=400&amp;h=321 400w, https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=800&amp;h=642 800w, https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=250&amp;h=200 250w, https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=768&amp;h=616 768w, https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_detail.jpg?w=779&amp;h=625 779w\" \/><\/a>In the series of videos after the break, viewers have the opportunity to watch a project go from idea to final product. The first video was uploaded nearly a month before the project was completed, and goes over some of the design elements of the project as well as different ideas [Calvin] had in terms of things like component placement. Throughout the video, he illustrates his ideas in TinkerCAD, which might not have been <em>our<\/em> first choice for a project this complex, but it does go to show what\u2019s possible in the free web-based CAD package.<\/p>\n<p>By the second video, [Calvin] has printed some parts and now has the hardware coming together. The general idea is that the outside panels of the \u201cbook\u201d are made out of steel cut from the side panel of an old computer, with the 3D printed components taking the form of spacers between the electronic components. These plastic \u201cpages\u201d are not only easier and faster to print than a complete case, but help sell the appearance of the book when viewed from the sides.<\/p>\n<p>[Calvin] has shared his TinkerCAD design so that others can print out the necessary components for the book, though you\u2019ll have to source your own steel plates. He also breaks down all the principle components he used and gives links to where you can buy them, from the display and keyboard down to the screws and standoffs. He went with the Pi Zero and sticks to mainly console work, but if you want something with enough power to throw around a graphical environment, he says there\u2019s room in the case for a Pi 3.<\/p>\n<p>Hackers seem to enjoy hiding hardware inside of books, PLA or otherwise. We\u2019ve recently <a href=\"https:\/\/hackaday.com\/2018\/09\/22\/wonderful-ipad-mini-2-home-made-spiral-notebook-case\/\">seen an iPad nestled snugly into a notebook<\/a>, and of course no house would be complete without <a href=\"https:\/\/hackaday.com\/2018\/03\/10\/secret-book-light-switch\/\">a book doubling as a hidden switch<\/a>.<\/p>\n<p>\n<\/p><p><iframe class=\"youtube-player\" width=\"800\" height=\"480\" src=\"https:\/\/www.youtube.com\/embed\/YzPa3XTZbz4?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>\n<p><iframe class=\"youtube-player\" width=\"800\" height=\"480\" src=\"https:\/\/www.youtube.com\/embed\/kjdEVaSI5tQ?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>\n<p><iframe class=\"youtube-player\" width=\"800\" height=\"480\" src=\"https:\/\/www.youtube.com\/embed\/9PDkytWfMA0?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>","created_at":"2019-02-26T21:22:21+0000","updated_at":"2019-02-26T21:22:21+0000","published_at":"2019-02-26T06:00:40+0000","published_by":["Tom Nardi"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"hackaday.com","preview_picture":"https:\/\/hackadaycom.files.wordpress.com\/2019\/02\/pibook_feat.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["programming languages"],"is_public":false,"id":861,"title":"tour-de-babel - steveyegge2","url":"https:\/\/sites.google.com\/site\/steveyegge2\/tour-de-babel","content":"<p> This is my whirlwind languages tour \u2014 the one I was going to write for the Amazon Developers Journal this month, but couldn't find a way to do it that was... presentable. <\/p><p> For one thing, I lapse occasionally into coarseness and profanity here, so it wasn't appropriate for an official-ish Amazon publication. Instead, I'm stuffing it into my blog, which nobody reads. Except for you. Yep, just you. Hiya. <\/p><p> For another, it's really a work in progress, just a few snippets here and there, not polished at all. Another great reason to make it a blog entry. Doesn't have to be good, or complete. It's just what I've got today. Enjoy! <\/p><p>My whirlwind tour will cover C, C++, Lisp, Java, Perl, (all languages we use at Amazon), Ruby (which I just plain <em>like<\/em>), and Python, which is in there because \u2014 well, no sense getting ahead of ourselves, now. <\/p><h3>C<\/h3><p> You just have to know C. Why? Because for all practical purposes, every computer in the world you'll ever use is a von Neumann machine, and C is a lightweight, expressive syntax for the von Neumann machine's capabilities. <\/p><p> The von Neumann architecture is the standard computer architecture you use every day: a CPU, RAM, a disk, a bus. Multi-CPU doesn't really change it that much. The von Neumann machine is a convenient, cost-effective, 1950s realization of a Turing Machine, which is a famous abstract model for performing computations. <\/p><p> There are other kinds of machines too. For instance, there are Lisp Machines, which are convenient 1950s realizations of Lisp, a programming language notation based on the lambda calculus, which is another model for performing computations. Unlike Turing machines, the lambda calculus can be read and written by humans. But the two models are equivalent in power. They both model precisely what computers are capable of. <\/p><p> Lisp Machines aren't very common though, except at garage sales. von Neumann machines won the popularity race. There are various other kinds of computers, such as convenient realizations of neural networks or cellular automata, but they're nowhere as popular either, at least not yet. <\/p><p> So you have to know C. <\/p><p> You also have to know C because it's the language that Unix is written in, and happens also to be the language that Windows and virtually all other operating systems are written in, because they're OSes for von Neumann machines, so what else would you use? Anything significantly different from C is going to be too far removed from the actual capabilities of the hardware to perform well enough, at least for an OS \u2014 at least in the last century, which is when they were all written. <\/p><p> You should also know Lisp. You don't have to use it for real work, although it comes in quite handy for lots of GNU applications. In particular, you should learn Scheme, which is a small, pure dialect of Lisp. The GNU version is called <a href=\"http:\/\/www.gnu.org\/software\/guile\/guile.html\" rel=\"nofollow\">Guile.<\/a> <\/p><p> They teach Scheme at MIT and Berkeley to new students for a semester or two, and the students have absolutely no clue as to why they're learning this weird language. It's a lousy first language, to be honest, and probably a lousy second one too. You should learn it, eventually, and not as your first or second language. <\/p><p>It's hard, though. It's a big jump. It's not sufficient to learn how to write C-like programs in Lisp. That's pointless. C and Lisp stand at opposite ends of the spectrum; they're each great at what the other one sucks at. <\/p>If C is the closest language to modeling how computers work, Lisp is the closest to modeling how<em>computation<\/em>works. You don't need to know a lot of Lisp, really. Stick with Scheme, since it's the simplest and cleanest. Other Lisps have grown into big, complex programming environments, just like C++ and Java have, with libraries and tools and stuff.<em>That<\/em>, you don't need to know. But you should be able to write programs in Scheme. If you can make your way through all the exercises in<u>The Little Schemer<\/u>and<a href=\"http:\/\/www.amazon.com\/o\/asin\/026256100X\" rel=\"nofollow\">The Seasoned Schemer<\/a>, you'll know enough, I think.<p> But you choose a language for day-to-day programming based on its libraries, documentation, tools support, OS integration, resources, and a host of other things that have very little to do with how computers work, and a whole lot to do with how <em>people<\/em> work. <\/p><p> People still write stuff in straight C. Lots of stuff. You should know it! <\/p><h3>C++<\/h3><p> C++ is the dumbest language on earth, in the very real sense of being the least sentient. It doesn't know about itself. It is not <em>introspective<\/em>. Neither is C, but C isn't \"Object-Oriented\", and object orientation is in no small measure about making your programs know about themselves. Objects are actors. So OO languages need to have runtime reflection and typing. C++ doesn't, not really, not that you'd ever use. <\/p><p> As for C: it's so easy to write a C compiler that you can build tools on top of C that act like introspection. C++, on the other hand, is essentially un-parseable, so if you want to write smart tools that can, for example, tell you the signatures of your virtual functions, or refactor your code for you, you're stuck using someone else's toolset, since <em>you<\/em> sure as heck aren't gonna parse it. And all the toolsets for parsing C++ out there just plain suck. <\/p><p> C++ is dumb, and you can't write smart systems in a dumb language. Languages shape the world. Dumb languages make for dumb worlds. <\/p><p> All of computing is based on <em>abstractions<\/em>. You build higher-level things on lower-level ones. You don't try to build a city out of molecules. Trying to use too low-level an abstraction gets you into trouble. <\/p><p> We are in trouble. <\/p><p> The biggest thing you can reasonably write in C is an operating system, and they're not very big, not really. They look big because of all their apps, but kernels are small. <\/p><p> The biggest thing you can write in C++ is... also an operating system. Well, maybe a little bigger. Let's say three times bigger. Or even ten times. But operating system kernels are at most, what, maybe a million lines of code? So I'd argue the biggest system you can reasonably write in C++ is maybe 10 million lines, and then it starts to break down and become this emergent thing that you have no hope of controlling, like the plant in Little Shop of Horrors. Feeeeeed meeeeeee... <\/p><p> If you can get it to compile by then, that is. <\/p><p> We have 50 million lines of C++ code. No, it's more than that now. I don't know what it is anymore. It was 50 million last Christmas, nine months ago, and was expanding at 8 million lines a quarter. The expansion rate was increasing as well. Ouch. <\/p><p> Stuff takes forever to do around here. An Amazon engineer once described our code base as \"a huge mountain of poop, the biggest mountain you've ever seen, and your job is to crawl into the very center of it, every time you need to fix something.\" <\/p><p> That was four years ago, folks. That engineer has moved on to greener pastures. Too bad; he was really good. <\/p><p> It's all C++'s fault. Don't argue. It is. We're using the dumbest language in the world. That's kind of meta-dumb, don't you think? <\/p><p> With that said, it <em>is<\/em> obviously possible to write nice C++ code, by which I mean, code that's mostly C, with some C++ features mixed in tastefully and minimally. But it almost never happens. C++ is a vast playground, and makes you feel smart when you know all of it, so you're always tempted to use all of it. But that's really, really hard to do <em>well<\/em>, because it's such a crap language to begin with. In the end, you just make a mess, even if you're good. <\/p><p> I know, this is Heresy, with a capital-'H'. Whatever. I loved C++ in college, because it's all I knew. When I heard that my languages prof, Craig Chambers, absolutely detested C++, I thought: \"Why? I like it just fine.\" And when I heard that the inventor of STL was on record as saying he hated OOP, I thought he was cracked. How could anyone hate OOP, especially the inventor of STL? <\/p><p> Familiarity breeds contempt in most cases, but not with computer languages. You have to become an expert with a <em>better<\/em> language before you can start to have contempt for the one you're most familiar with. <\/p><p> So if you don't like what I'm saying about about C++, go become an expert at a better language (I recommend Lisp), and then you'll be armed to disagree with me. You won't, though. I'll have tricked you. You won't like C++ anymore, and you might be irked that I tricked you into disliking your ex-favorite language. So maybe you'd better just forget about all this. C++ is great. Really. It's just ducky. Forget what I said about it. It's fine. <\/p><h3>Lisp<\/h3><p> <em>(I'm betting this next section will astonish you, even if you've been here a while.)<\/em> <\/p><p> When Amazon got its start, we had brilliant engineers. I didn't know all of them, but I knew some of them. <\/p><p> Examples? Shel Kaphan. Brilliant. Greg Linden. Brilliant. Eric Benson. Independently famous in his own right, before he ever even came to Amazon. Also brilliant. <\/p><p> They wrote the Obidos webserver. Obidos made Amazon successful. It was only later that poop-making engineers and web devs, frontend folks mostly \u2014 schedule-driven people who could make their managers happy by delivering crap fast \u2014 it was only later that these people made Obidos bad. Clogged the river, so to speak. But Obidos was a key cornerstone of Amazon's initial success. <\/p><p> The original brilliant guys and gals here only allowed two languages in Amazon's hallowed source repository: C and Lisp. <\/p><p> Go figure. <\/p><p> They all used Emacs, of course. Hell, Eric Benson was one of the <em>authors<\/em> of XEmacs<sup><a href=\"#note1\">1<\/a><\/sup>. All of the greatest engineers in the world use Emacs. The world-changer types. Not the great gal in the cube next to you. Not Fred, the amazing guy down the hall. I'm talking about the greatest software developers of our profession, the ones who changed the face of the industry. The James Goslings, the Donald Knuths, the Paul Grahams<sup><a href=\"#note2\">2<\/a><\/sup>, the Jamie Zawinskis, the Eric Bensons. Real engineers use Emacs. You have to be way smart to use it well, and it makes you incredibly powerful if you can master it. Go look over Paul Nordstrom's shoulder while he works sometime, if you don't believe me. It's a real eye-opener for someone who's used Visual Blub .NET-like IDEs their whole career. <\/p><p> Emacs is the 100-year editor. <\/p><p> Shel, Eric, Greg, and others like them that I wasn't fortunate enough to work with directly: they didn't allow C++ here, and they didn't allow Perl. (Or Java, for that matter). They knew better. <\/p><p> Now C++, Java and Perl are all we write in. The elders have moved on to greener pastures too.<\/p><p> Shel wrote Mailman in C, and Customer Service wrapped it in Lisp. Emacs-Lisp. You don't know what Mailman is. Not unless you're a longtime Amazon employee, probably non-technical, and you've had to make our customers happy. Not indirectly, because some bullshit feature you wrote broke (because it was in C++) and pissed off our customers, so you had to go and fix it to restore happiness. No, I mean directly; i.e., you had to <em>talk<\/em> to them. Our lovely, illiterate, eloquent, well-meaning, hopeful, confused, helpful, angry, happy customers, the real ones, the ones buying stuff from us, our <em>customers<\/em>. Then you know Mailman. <\/p><p> Mailman was the Customer Service customer-email processing application for ... four, five years? A long time, anyway. It was written in Emacs. Everyone loved it. <\/p><p> People <em>still love it<\/em>. To this very day, I still have to listen to long stories from our non-technical folks about how much they miss Mailman. I'm not shitting you. Last Christmas I was at an Amazon party, some party I have no idea how I got invited to, filled with business people, all of them much prettier and more charming than me and the folks I work with here in the Furnace, the Boiler Room of Amazon. Four young women found out I was in Customer Service, cornered me, and talked for <em>fifteen minutes<\/em> about how much they missed Mailman and Emacs, and how Arizona (the JSP replacement we'd spent years developing) still just wasn't doing it for them. <\/p><p> It was truly surreal. I think they may have spiked the eggnog. <\/p><p> Shel's a genius. Emacs is a genius. Even non-technical people love Emacs. I'm typing in Emacs right now. I'd never voluntarily type anywhere else. It's more than just a productivity boost from having great typing shortcuts and text-editing features found nowhere else on the planet. I type 130 to 140 WPM, error-free, in Emacs, when I'm doing free-form text. I've timed it, with a typing-test Emacs application I wrote. But it's <em>more<\/em> than that. <\/p><p> Emacs has the Quality Without a Name. <\/p><p> We retired Mailman. That's because we have the Quality <em>With<\/em> a Name \u2014 namely, Suckiness. We suck. We couldn't find anyone who was good enough at Emacs-Lisp to make it work. Nowadays it would be easy; Amazon's filled up with Emacs Lisp hackers, but back then, CS Apps couldn't get the time of day from anyone, so they did what they could with what they had, and there weren't enough Emacs-Lisp folks. For a while, they even had Bob Glickstein on contract, the guy who wrote the O'Reilly \"giraffe\" book <u>Writing Gnu Emacs Extensions<\/u>, sitting there writing Gnu Emacs Extensions for Mailman in this little office in the Securities building. <\/p><p> CS Apps was Amazon's first 2-pizza team, you know. They're completely autonomous \u2014 then and now. Nobody talks to them, nobody helps them, they build everything themselves. They don't have web devs, they don't have support engineers, they don't have squat, except for rock-solid engineers and a mentoring culture. And that's all they've ever needed. <\/p><p> But they had to retire Mailman. Alas. Alackaday. And I still get to hear about how much people miss it. At parties, even. <\/p><p> I think there may still be more Lisp hackers, per capita, in CS Apps than in any other group at Amazon. Not that they get to use it much, but as Eric Raymond said, even if you don't program in it much, learning Lisp will be a profound experience that will make you a better engineer for the rest of your life. <\/p><p> Religion isn't the opiate of the masses anymore, Karl. IDEs are. <\/p><h3>Java<\/h3><p> Java is simultaneously the best and the worst thing that has happened to computing in the past 10 years. <\/p><p> On the one hand, Java frees you up from many mundane and error-prone details of C++ coding. No more bounds errors, no more core dumps. Exceptions thrown point you to the exact line of code that erred, and are right 99% of the time. Objects print themselves intelligently on demand. Etc., etc. <\/p><p> On the other hand, in addition to being a language, a virtual machine, a huge set of class libraries, a security model, and a portable bytecode format, Java is a religion. So you can't trust anyone who loves it too much. It's a tricky business to hire good Java programmers. <\/p><p> But Java really has been a big step forward for software engineering in general. <\/p><p> Going from C++ to Java isn't just changing syntax. It's a shocking paradigm shift that takes a while to sink in. It's like suddenly getting your own Executive Assistant. You know how VPs always seem to have all this time to be in meetings, and know how the company's running, and write cool documents, and stuff like that? VPs tend to forget that they're actually TWO full-time people: their self and their EA. Having an EA frees you up to think about the <em>problems<\/em> you need to solve; not having one forces you to spend half your time on mundane tasks. Switching to Java turns you into two programmers \u2014 one taking care of all this <em>stuff<\/em> that you no longer have to think much about, and another one focused on the problem domain. It's a staggering difference, and one you can get used to in a real hurry. <\/p><p> As Jamie Zawinski said in his famous \"java sucks\" article: \"First the good stuff: Java doesn't have free(). I have to admit right off that, after that, all else is gravy. That one point makes me able to forgive just about anything else, no matter how egregious. Given this one point, everything else in this document fades nearly into insignificance.\" <\/p><p> Jamie's article was written in 1997, which in Java years is a <em>long<\/em> time ago, and Java has improved a lot since he wrote it; some of the things he complains about are even fixed now. <\/p><p> Most of them aren't. Java does still kind of suck, as a language. But as Jamie points out, it's \"the best language going today, which is to say, it's the marginally acceptable one among the set of complete bagbiting loser languages that we have to work with out here in the real world.\" <\/p><p> Really, you should <a href=\"http:\/\/www.jwz.org\/doc\/java.html\" rel=\"nofollow\">read it<\/a>. <\/p><p> Java is truly wonderful along almost every dimension except for the language itself, which is mostly what JWZ was griping about. But that's a <em>lot<\/em> to gripe about. Libraries can only get you so far if your language sucks. Trust me: you may know many, many things better than I do, but I know that libraries can't really save a sucky language. Five years of assembly-language hell at Geoworks taught me that. <\/p><p> Compared to C++, Java as a language is about even. Well, scratch that, it's a lot better, because it has <em>strings<\/em>, oh man, how can you use a language with lousy string support. <\/p><p>But Java's missing some nice features from C++, such as pass-by-reference(-to-stack-object), typedefs, macros, and operator overloading. Stuff that comes in handy now and again. <\/p><p>Oh, and multiple inheritance, which now I've come to appreciate in my old age. If you think my <a href=\"https:\/\/sites.google.com\/site\/steveyegge2\/when-polymorphism-fails\">Opinionated Elf<\/a> was a good counterpoint to polymorphism dogma, I've got several brilliant examples of why you <em>need<\/em> multiple inheritance, or at least Ruby-style mixins or automatic delegation. Ask me about the Glowing Sword or Cloak of Thieving sometime. Interfaces suck. <\/p><p>Gosling even said, a few years ago, that if he had to do it all over again, he wouldn't have used interfaces. <\/p><p> But that's <em>just exactly<\/em> what the problem with Java is. When James said that, people were shocked. I could feel the shock waves, could feel the marketing and legal folks at Sun maneuvering to hush him up, brush it off, say it wasn't so. <\/p><p> The problem with Java is that people are blinded by the marketing hype. That's the problem with C++, with Perl, with any language that's popular, and it's a serious one, because languages can't become popular <em>without<\/em> hype. So if the language designer suggests innocently that the language might not have been designed perfectly, it's time to shoot the language designer full of horse tranquilizers and shut down the conference. <\/p><p> Languages need hype to survive; I just wish people didn't have to be <em>blinded<\/em> by it. <\/p><p> I drank the OOP Kool-Aid, I regurgitated the hype myself. When I started at Amazon, I could recite for you all the incantations, psalms, and voodoo chants that I'd learned, all in lieu of intelligence or experience, the ones that told me Multiple Inheritance is Evil 'cuz Everyone Says So, and Operator Overloading Is Evil, and so on. I even vaguely sort of knew why, but not really. Since then I've come to realize that it's not MI that sucks, it's developers who suck. <em>I<\/em> sucked, and I still do, although hopefully less every year. <\/p><p> I had an interview candidate last week tell me that MI is Evil because, for instance, you could make a Human class that multiply-inherits from Head, Arm, Leg, and Torso. He was both right and wrong. That MI situation was evil, sure, but it was all him. Stupid from a distance, evil if he'd made it in through the front door. <\/p><p> Bad developers, who constitute the majority of all developers worldwide, can write bad code in any language you throw at them. <\/p><p> That said, though, MI is no picnic; mixins seem to be a better solution, but <em>nobody<\/em> has solved the problem perfectly yet. But I'll still take Java over C++, even without MI, because I know that no matter how good my intentions are, I will at some point be surrounded by people who don't know how to code, and they will do far less damage with Java than with C++. <\/p><p> Besides, there's way more to Java than the core language. And even the language is evolving, albeit glacially, so there's hope. It's what we should be using at Amazon. <\/p><p> You just have to be careful, because as with any other language, you can easily find people who know a lot about the language environment, and very little about taste, computing, or anything else that's important. <\/p><p> When in doubt, hire Java programmers who are polyglots, who detest large spongy frameworks like J2EE and EJB, and who use Emacs. All good rules of thumb. <\/p><h3>Perl<\/h3><p> Perl. Where to start? <\/p><p> Perl is an old friend. Perl and I go way back. I started writing Perl stuff in maybe 1995, and it's served me well for nearly a decade. <\/p><p> It's like that old bicycle you've put 100k or 200k miles on, and you'll always have a warm fuzzy spot for it, even though you've since moved on to a more modern bike that weighs 5 lbs and doesn't make your ass hurt so much. <\/p><p> Perl is popular for three reasons: <\/p><ol><li>\n<p> You can <em>get stuff done<\/em> really fast in Perl, which is what really matters, in the end. <\/p>\n<\/li>\n<li>\n<p> Perl has the best marketing in the world. You could write a book about how brilliant their marketing is. Sun has marketed Java with money, and Perl is almost keeping up, popularity-wise, purely on the on sheer marketing brilliance of Larry Wall and his buddies. Folks at Harvard Business School should study Perl's marketing. It's astonishing. <\/p>\n<\/li>\n<li>\n<p> Until roughly, oh, <em>now<\/em>, it had no real competitors. <\/p>\n<\/li>\n<\/ol><p> There are \"better\" languages than Perl \u2014 hell, there are lots of them, if you define \"better\" as \"not being insane\". Lisp, Smalltalk, Python, gosh, I could probably name 20 or 30 languages that are \"better\" than Perl, inasmuch as they don't look like that Sperm Whale that exploded in the streets of Taiwan over the summer. Whale guts everywhere, covering cars, motorcycles, pedestrians. That's Perl. It's charming, really. <\/p><p> But Perl has many, many things going for it that, until recently, no other language had, and they compensated for its exo-intestinal qualities. You can make all sorts of useful things out of exploded whale, including perfume. It's quite useful. And so is Perl. <\/p><p> While all those other languages (Lisp and Smalltalk being particularly noteworthy offenders) tried to pretend that operating systems don't exist, and that lists (for Lisp) or objects (for Smalltalk) are the be-all, end-all of getting shit done, Perl did exactly the opposite. Larry said: Unix and string processing are the be-all, end-all of getting shit done. <\/p><p> And for many tasks, he was absolutely right. So Perl is better at Unix integration and string processing than any language on the planet, save one, and that one only arrived on the scene recently, from the land of Godzilla. I'll get to that one later. <\/p><p> Sadly, Larry focused sooooo hard on Unix integration and string processing that he totally forgot about lists and objects until it was far too late to implement them properly. In fact, a few key mistakes he made early on in Perl's... well, I hesitate to use the word \"design\" for whale guts, but let's call it Perl's \"lifecycle\" \u2014 those mistakes made it so hard to do lists and objects correctly that Perl has evolved into a genuine Rube Goldberg machine, at least if you want to use lists or objects. <\/p><p> Lists and objects are pretty farging important too, Larry! <\/p><p> Perl can't do lists because Larry made the tragically stupid decision early on to flatten them automatically. So (1, 2, (3, 4)) magically becomes (1, 2, 3, 4). Not that you <em>ever<\/em> want it to work this way. But Larry happened to be working on some problem for which it was convenient on that particular day, and Perl's data structures have been pure exploded whale ever since. <\/p><p> Now you can't read a book or tutorial or PowerPoint on Perl without spending at least a third of your time learning about \"references\", which are Larry's pathetic, broken, Goldbergian fix for his list-flattening insanity. But Perl's marketing is so incredibly good that it makes you feel as if references are the best thing that ever happened to you. You can take a reference to anything! It's fun! Smells good, too! <\/p><p> Perl can't do objects because Larry never r<i>eeee<\/i>ally believed in them. Maybe that's OK; I'm still not quite sure if I believe in them either. But then why did he try adding them? Perl's OO is a halfhearted add-on that never caught on with the Perl community. It's just not as inspired as the string-processing or Unix integration stuff. <\/p><p> And of course, Perl has plenty of other crackpot design features. Take its \"contexts\", for instance, which are a horrid outgrowth of Larry's comical decision to have N variable namespaces, dereferenced by sigils, which he sort of copied from shell-script. In Perl, every operator, every function, every <em>operation<\/em> in the language behaves randomly in one of six different ways, depending on the current \"context\". There are no rules or heuristics governing how a particular operation will behave in a given context. You just have to commit it all to memory. <\/p><p> Need an example? Here's one: accessing a hash in a scalar context gives you a string containing a fraction whose numerator is the number of allocated keys, and the denominator is the number of buckets. Whale guts, I'm telling you. <\/p><p> Like I said, though \u2014 until recently, nothing could get the job done like Perl could. <\/p><h3>Ruby<\/h3><p> Every 15 years or so, languages are replaced with better ones. C was replaced by C++, at least for large-scale application development by people who needed performance but desperately wanted data types too. C++ is being replaced by Java, and Java will doubtless be replaced with something better in seven years \u2014 well, seven years after it finishes replacing C++, which evidently hasn't fully happened yet, mostly because Microsoft was able to stall it before it became ubiquitous on the desktop. But for server-side applications, C++ is basically on its way out. <\/p><p> Perl will be gone soon, too. That's because a new language called Ruby has finally been translated into English. Yep, it was invented in Japan, of all places \u2014 everyone else was as surprised as you are, since Japan's known for its hardware and manufacturing, but not for its software development. Why, is anyone's guess, but I'm thinking it's the whole typing thing; I just can't imagine they were able to type fast enough before, what with having an alphabet with ten thousand characters in it. But Emacs got multibyte support a few years ago, so I can imagine they're pretty dang fast with it now. (And yes, they use Emacs \u2014 in fact Japanese folks did the majority of the Mule [multibyte] support for Emacs, and it's rock-solid.) <\/p><p> Anyway, Ruby stole <em>everything<\/em> good from Perl; in fact, Matz, Ruby's author (Yukihiro Matsumoto, if I recall correctly, but he goes by \"Matz\"), feels he may have stolen a little too much from Perl, and got some whale guts on his shoes. But only a little. <\/p><p> For the most part, Ruby took Perl's string processing and Unix integration <em>as-is<\/em>, meaning the syntax is identical, and so right there, before anything else happens, you already have the Best of Perl. And that's a great start, especially if you don't take the Rest of Perl. <\/p><p> But then Matz took the best of list processing from Lisp, and the best of OO from Smalltalk and other languages, and the best of iterators from CLU, and pretty much the best of everything from everyone. <\/p><p> And he somehow made it all work together so well that you don't even notice that it <em>has<\/em> all that stuff. I learned Ruby faster than any other language, out of maybe 30 or 40 total; it took me about 3 days before I was more comfortable using Ruby than I was in Perl, after eight years of Perl hacking. It's so consistent that you start being able to guess how things will work, and you're right most of the time. It's beautiful. And fun. And practical. <\/p><p> If languages are bicycles, then Awk is a pink kiddie bike with a white basket and streamers coming off the handlebars, Perl is a beach cruiser (remember how cool they were? Gosh.) and Ruby is a $7,500 titanium mountain bike. The leap from Perl to Ruby is as significant as the leap from C++ to Java, but without any of the downsides, because Ruby's essentially a proper superset of Perl's functionality, whereas Java took some things away that people missed, and didn't offer real replacements for them. <\/p><p> I'll write more about Ruby sometime. I need to be inspired first. Read <a href=\"http:\/\/poignantguide.net\/ruby\/\" rel=\"nofollow\">Why the Lucky Stiff's (poignant) guide to Ruby<\/a>. <em>That<\/em> is an inspired piece of work. Really. Read it. It's amazing. I don't understand the kind of mind that could produce it, but it's funny, and poignant, and all about Ruby. Sort of. You'll see. <\/p><h3>Python<\/h3><p> Well gosh, what about Python, a nice language that has patiently been waiting in the wings for all these years? The Python community has long been the refuge for folks who finally took the red pill and woke up from the Perl Matrix. <\/p><p> Well, they're just like the Smalltalk folks, who waited forever to replace C++, and then Java came along and screwed them royally, and permanently. Oops. Ruby's doing exactly that to Python, right now, today. Practically overnight. <\/p><p> Python <em>would<\/em> have taken over the world, but it has two fatal flaws: the whitespace thing, and the permafrost thing. <\/p><p> The whitespace thing is simply that Python uses indentation to determine block nesting. It forces you to indent everything a certain way, and they do this so that everyone's code will look the same. A surprising number of programmers <em>hate<\/em> this, because it feels to them like their freedom is being taken away; it feels as if Python is trampling their constitutional right to use shotgun formatting and obfuscated one-liners.<sup><a href=\"#note3\">3<\/a><\/sup><\/p><p> Python's author, Guido Van Rossum, also made some boneheaded technical blunders early on \u2014 none quite as extravagant as Larry's blunders, but a few were real doozies nonetheless. For instance, Python originally had no lexical scoping. But it didn't have dynamic scoping either, and dynamic scoping may have its share of problems, but it at least sort of works. Python had NOTHING except for global and local (function) scope, so even though it had a \"real\" OO system, classes couldn't even access their own damned instance variables. You have to pass a \"self\" parameter to EVERY instance method and then get to your instance data by accessing it through self. So everything in Python is self, selfself, selfselfself, selfSELFselfSELF__SELF__, and it drives you frigging nuts, even if you don't mind the whitespace thing. <\/p><p>Etc.<\/p><p> But in my opinion, it's really the frost thing that killed Python, and has prevented it from ever achieving its wish to be the premier scripting language, or the premier anything language, for that matter. Heck, people still use Tcl as an embedded interpreter, even though Python is far superior to Tcl in every conceivable way \u2014 except, that is, for the frost thing. <\/p><p> What's the frost thing, you ask? Well, I used to have a lot of exceptionally mean stuff written here, but since Python's actually quite pleasant to work with (if you can overlook its warts), I no longer think it's such a great idea to bash on Pythonistas. The \"frost thing\" is just that they used to have a tendency to be a bit, well, frosty. Why? <\/p><p>Because they were so tired of hearing about the whitespace thing!<\/p><p> I think that's why Python never reached Perl's level of popularity, but maybe I'm just imagining things. <\/p><h3>Coda<\/h3><p> That was the ADJ article I really wanted to write. Or at least something like it. For some reason, though, my true feelings only seem to come out during insomniac attacks between 3am and 6am. Time for bed! 2 hours 'til my next meeting. <\/p><p>(Published September 2004. Minor updates on 3\/28\/2006)<\/p><hr \/><p> <b> Notes <\/b> <\/p><p><sup><a name=\"note1\" id=\"note1\">1<\/a><\/sup> Eric tells me it was actually almost all Jamie Zawinski, when they worked at Lucid together.<\/p><p> <sup><a name=\"note2\" id=\"note2\">2<\/a><\/sup> It's been pointed out many times since I wrote this that Paul Graham actually uses vi. Go figure!<\/p><p> <sup><a name=\"note3\" id=\"note3\">3<\/a><\/sup> For the record, I personally don't mind the whitespace thing at all. I think it's silly to dislike Python for that reason. What I'm saying is that a surprising percentage of *other* programmers hate it. <\/p>","created_at":"2019-02-26T19:33:59+0000","updated_at":"2019-02-26T19:34:06+0000","annotations":[],"mimetype":"text\/html","reading_time":28,"domain_name":"sites.google.com","preview_picture":"https:\/\/sites.google.com\/site\/steveyegge2\/stevey-email.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":860,"title":"Weaving a budget with Org & ledger","url":"https:\/\/orgmode.org\/worg\/org-tutorials\/weaving-a-budget.html","content":"<p>\nThis is where Org, and particularly babel\u2019s noweb features come in.\nBut first, let\u2019s look at how you can use Org to generate some useful\nreports. Here is how I generate last month\u2019s expenses:\n<\/p><pre class=\"example\">\n* Last month's expenses\n#+begin_src ledger :cmdline bal -p \"last month\" ^Expenses: :results output :exports results\n!include \/path\/to\/ledger.lgr\n#+end_src\n<\/pre><p>\n(You may need to ensure that <code>ob-ledger<\/code> is loaded into your Org;\nsee the <a href=\"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/languages\/ob-doc-ledger.html\">ob-ledger doc<\/a> for details.) Now, if you type <code>C-c C-c<\/code> on this\nentry, it will generate a report of your expenses last month, which\nwill be placed after the source block, and will look like:\n<\/p><pre class=\"example\">\n#+RESULTS:\n#+begin_example\n            $150.00  Expenses\n            $100.00    Food\n             $50.00    Bars\n--------------------\n            $150.00\n#+end_example\n<\/pre><p>\nBy using babel, you can generate a lot of useful reports and keep them\nup-to-date, exporting them to HTML or PDF for printing, etc.\n<\/p><p>\nBut some of the really useful features come when you use Org\u2019s noweb\nto weave a budget together. noweb allows, among other things, you to\ninclude source blocks in other blocks. First, you\u2019ll set up your base\nbudget, which will be included in every budget. Say you know that\nevery month you will spend $500 on food. You create a ledger source\nblock with a name, <code>base-budget<\/code>:\n<\/p><pre class=\"example\">\n#+name: base-budget\n#+begin_src ledger\n  ; :BUDGET:\n  Budgeted:Expenses:Food  $500\n  Budgeted:Savings\n#+end_src\n<\/pre><p>\nThis means that you want $500 to go to food, and the balance to\nsavings. (The first line <code>; :BUDGET:<\/code> is a tag to let you identify\nthis as a budget; this helps with some reports.) Now, you can set up\nyour budget block:\n<\/p><pre class=\"example\">\n#+name: budget\n#+begin_src ledger :noweb yes\n= expr (date&gt;=[2014] and account =~ \/^Expenses:\/)\n  Budgeted:$account        -1\n  Budget  1\n2014\/01\/01 * Budget\n &lt;&lt;base-budget&gt;&gt;\n Budget  -$1000\n#+end_src\n<\/pre><p>\nWhat you have done here is set up your automated transaction, as\nabove, and set up your first month\u2019s budget. You have assumed that\nyour first month\u2019s income is $1000, and you have included your base\nbudget. When this is woven by noweb, <code>&lt;&lt;base-budget&gt;&gt;<\/code> will be\nreplaced with the <code>base-budget<\/code> source block, and Org will pass the\nfollowing on to ledger:\n<\/p><pre class=\"example\">\n2014\/01\/01 * Budget\n  Budgeted:Expenses:Food  $500\n  Budgeted:Savings\n  Budget                -$1000\n<\/pre><p>\nThis will set up $500 in your food budget, and the balance ($500) for\nsavings.\n<\/p><p>\nNow, to generate a report, you can use the following:\n<\/p><pre class=\"example\">\n#+begin_src ledger :cmdline bal -p \"this year\" ^Budgeted:Expenses -E :results output :exports results :noweb yes\n&lt;&lt;budget&gt;&gt;\n!include \/path\/to\/ledger.lgr\n#+end_src\n<\/pre><p>\nRunning this report (<code>C-c C-c<\/code>) will tell you your budget balance for\neach expense. If you budgeted $500 for food and spent $600, your\nbalance will be -$100. If you spend $300, your balance will be $200.\n<\/p><p>\nEach month, month, you will create a new budget entry in your <code>budget<\/code>\nsource block, with the <code>Budget<\/code> account equaling your income that\nmonth. You will include your <code>&lt;&lt;base-budget&gt;&gt;<\/code>, and any adjustments\nyou want to make.\n<\/p><p>\nFor example, say that in June you take up cycling. You want to budget\n$100 per month for this, but you want to start in June, not January.\nYou will add the following entry to your <code>budget<\/code> source block:\n<\/p><pre class=\"example\">\n2014\/06\/01 * Budget\n &lt;&lt;base-budget&gt;&gt;\n Budgeted:Expenses:Cycling  $100\n Budget                   -$1000\n<\/pre><p>\nThis means that for June your budget will be $500 for food, $100 for\ncycling, and $400 for savings.\n<\/p><p>\nNow, you will need to keep adding the cycling budget line every time\nfrom now on, so you might want, at some point, to get complicated. You\ncould define a new source block with the name <code>base-budget-new<\/code>,\ninclude your old <code>base-budget<\/code>, using <code>&lt;&lt;base-budget&gt;&gt;<\/code>, and then\ninclude that in your budget entries going forward, to avoid duplicate\ntyping. noweb should allow you to structure your budget entries\nhowever you like.\n<\/p><p>\nI\u2019ve found this solution to be flexible enough to allow me to manage a\nbudget that is contantly being adjusted, and easy to generate\nprintable reports. ledger and Org make an excellent pair.<\/p>","created_at":"2019-02-26T15:06:27+0000","updated_at":"2019-02-26T15:06:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":859,"title":"Using Ledger for Accounting in Org-mode with Babel","url":"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/languages\/ob-doc-ledger.html","content":"<p>\nThe easiest, albeit possibly less useful, way in which to use Ledger\nwithin an org file is to use a single source block to record all\nLedger entries.  The following is an example source block:\n<\/p><pre class=\"example\">\n#+name: allinone\n#+begin_src ledger\n2010\/01\/01 * Starting balance\n  assets:bank:savings  \u00a31300.00\n  income:starting balances\n2010\/07\/22 * Got paid\n  assets:bank:chequing  \u00a31000.00\n  income:salary\n2010\/07\/23 Rent\n  expenses:rent  \u00a3500.00\n  assets:bank:chequing\n2010\/07\/24 Food\n  expenses:food  \u00a3150.00\n  assets:bank:chequing\n2010\/07\/31 * Interest on bank savings\n  assets:bank:savings  \u00a33.53\n  income:interest\n2010\/07\/31 * Transfer savings\n  assets:bank:savings  \u00a3250.00\n  assets:bank:chequing\n2010\/08\/01 got paid again\n  assets:bank:chequing  \u00a31000.00\n  income:salary\n#+end_src\n<\/pre><p>\nIn this example, we have combined both expenses and income into one\nset of Ledger entries.  We can now generate register and balance\nreports (as well as many other types of reports) using babel to invoke\nLedger with specific arguments.  The arguments are passed to Ledger\nusing the <code>:cmdline<\/code> header argument.  In the code block above, there is\nno such argument so the system takes the default.  For Ledger code\nblocks, the default <code>:cmdline<\/code> argument is <code>bal<\/code> and the result of\nevaluating this code block (<code>C-c C-c<\/code>) would be:\n<\/p><pre class=\"example\">\n#+results: allinone()\n:            \u00a32653.53  assets\n:             \u00a3650.00  expenses\n:           \u00a3-3303.53  income\n<\/pre><p>\nIf, instead, you wished to generate a register of all the\ntransactions, you would change the <code>#+begin_src<\/code> line for the code block\nto include the required command line option:\n<\/p><pre class=\"example\">\n#+begin_src ledger :cmdline reg\n<\/pre><p>\nEvaluating the code block again would generate a different report.\n<\/p><p>\nHaving to change the actual directive on the code block and\nre-evaluate makes it difficult to have more than one view of your\ntransactions and financial state.  Eventually, babel will support\npassing arguments to <code>#+call<\/code> evaluations of code blocks but this\nsupport is missing currently.  Instead, we can use the concepts of\nliterary programming, as implemented by the <code>noweb<\/code> features of babel,\nto help us.\n<\/p>","created_at":"2019-02-26T14:59:01+0000","updated_at":"2019-02-26T14:59:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["twitch","streaming","livecoding"],"is_public":false,"id":858,"title":"My Twitch Live Coding Setup","url":"https:\/\/medium.com\/@suzhinton\/my-twitch-live-coding-setup-b2516672fb21","content":"<div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"1742\" class=\"graf graf--h3 graf--leading graf--title\">My Twitch Live Coding\u00a0Setup<\/h1><div class=\"uiScale uiScale-ui--regular uiScale-caption--regular u-flexCenter u-marginVertical24 u-fontSize15 js-postMetaLockup\"><div class=\"u-flex0\"><a class=\"link u-baseColor--link avatar\" href=\"https:\/\/medium.com\/@suzhinton?source=post_header_lockup\" data-action=\"show-user-card\" data-action-source=\"post_header_lockup\" data-action-value=\"89a233b79704\" data-action-type=\"hover\" data-user-id=\"89a233b79704\" dir=\"auto\"><img src=\"https:\/\/cdn-images-1.medium.com\/fit\/c\/100\/100\/1*FydmiVPSjGgJr9_AtRN_Ig.jpeg\" class=\"avatar-image u-size50x50\" alt=\"Go to the profile of Suz Hinton\" \/><\/a><\/div><div class=\"u-flex1 u-paddingLeft15 u-overflowHidden\"><div class=\"u-paddingBottom3\"><a class=\"ds-link ds-link--styleSubtle ui-captionStrong u-inlineBlock link link--darken link--darker\" href=\"https:\/\/medium.com\/@suzhinton\" data-action=\"show-user-card\" data-action-value=\"89a233b79704\" data-action-type=\"hover\" data-user-id=\"89a233b79704\" dir=\"auto\">Suz Hinton<\/a><button class=\"button button--smallest u-noUserSelect button--withChrome u-baseColor--buttonNormal button--withHover button--unblock js-unblockButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-block-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/@suzhinton\/my-twitch-live-coding-setup-b2516672fb21\" data-action-source=\"post_header_lockup\">BlockedUnblock<\/button><button class=\"button button--primary button--smallest button--dark u-noUserSelect button--withChrome u-accentColor--buttonDark button--follow js-followButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-subscribe-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/_\/subscribe\/user\/89a233b79704\" data-action-source=\"post_header_lockup-89a233b79704-------------------------follow_byline\">FollowFollowing<\/button><\/div><div class=\"ui-caption u-noWrapWithEllipsis js-testPostMetaInlineSupplemental\"><time datetime=\"2019-02-20T17:40:42.592Z\">Feb 20<\/time><\/div><\/div><\/div><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"97d7\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*obShOzDcFnjR1jXMyL9Flw.jpeg\" data-width=\"4022\" data-height=\"2156\" data-is-featured=\"true\" data-action=\"zoom\" data-action-value=\"1*obShOzDcFnjR1jXMyL9Flw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*obShOzDcFnjR1jXMyL9Flw.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"1df8\" class=\"graf graf--p graf-after--figure\">This article is an updated review of tools and software that I use to run my <a href=\"https:\/\/www.twitch.tv\/noopkat\/\" data-href=\"https:\/\/www.twitch.tv\/noopkat\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">live coding stream<\/a> each Sunday. If you\u2019re interested in the <a href=\"https:\/\/medium.freecodecamp.org\/lessons-from-my-first-year-of-live-coding-on-twitch-41a32e2f41c1\" data-href=\"https:\/\/medium.freecodecamp.org\/lessons-from-my-first-year-of-live-coding-on-twitch-41a32e2f41c1\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">original article<\/a>, which offers some advice outside of tools and software, feel free to read it as well! When I wrote the original post, it was intended to be a \u201clessons learned\u201d post, and I didn\u2019t expect folks to use it as a real guide. Reflecting upon that, I\u2019m going to share my current setup with more attention to detail this time, as I still get questions on a weekly basis from folks looking for more information.<\/p><p id=\"f45d\" class=\"graf graf--p graf-after--p\">A reminder that this is not intended to be a \u201cget started\u201d guide, rather it\u2019s written to share how one particular established streamer runs their show. Keep in mind that everyone runs theirs differently.<\/p><p id=\"1fa9\" class=\"graf graf--p graf-after--p\">I\u2019ll breakdown everything I do and use into sections for easier consumption:<\/p><ul class=\"postList\"><li id=\"891d\" class=\"graf graf--li graf-after--p\">Hardware<\/li><li id=\"8baa\" class=\"graf graf--li graf-after--li\">Stream design in OBS<\/li><li id=\"3e78\" class=\"graf graf--li graf-after--li\">The thing where I\u2019m not capturing all of my desktop<\/li><li id=\"3672\" class=\"graf graf--li graf-after--li\">Technical settings in OBS<\/li><li id=\"7ac5\" class=\"graf graf--li graf-after--li\">Twitch channel page design<\/li><li id=\"e680\" class=\"graf graf--li graf-after--li\">Automation<\/li><\/ul><p id=\"15f5\" class=\"graf graf--p graf-after--li\"><strong class=\"markup--strong markup--p-strong\">A caveat:<\/strong> I stream on macOS, so there is going to be a bias towards tools and techniques for this operating system. Most tools mentioned by gaming streamers are Windows focused, so I want to share how I stream effectively on macOS.<\/p><h3 id=\"7b92\" class=\"graf graf--h3 graf-after--p\">Hardware<\/h3><p id=\"29f5\" class=\"graf graf--p graf-after--h3\">Not much has changed, but I\u2019ll list the hardware I use below:<\/p><p id=\"5c1f\" class=\"graf graf--p graf-after--p\">The decisions I made around each item are based upon finding gear that is not too expensive, but still had good reviews online when I researched. I\u2019m simply not interested in the \u201cbest of the best\u201d, as I believe that hardware can only do so much for your stream. You do you.<\/p><ul class=\"postList\"><li id=\"c4ee\" class=\"graf graf--li graf-after--p\">13\" Macbook Pro 3.3 GHz i7 (late 2016)<\/li><li id=\"9538\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/Acer-K272HUL-bmiidp-27-inch-Widescreen\/dp\/B00JB6HCIC\" data-href=\"https:\/\/www.amazon.com\/Acer-K272HUL-bmiidp-27-inch-Widescreen\/dp\/B00JB6HCIC\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Acer K272HUL<\/a> 27\" monitor at 2560 x 1440 resolution<\/li><li id=\"fb1a\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/LX-Desk-Mount-LCD-Arm\/dp\/B00358RIRC\" data-href=\"https:\/\/www.amazon.com\/LX-Desk-Mount-LCD-Arm\/dp\/B00358RIRC\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Ergotron monitor arm<\/a><\/li><li id=\"3312\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/Perixx-PERIMICE-712B-Wireless-Mouse-Laptop\/dp\/B0083K9I8Y\" data-href=\"https:\/\/www.amazon.com\/Perixx-PERIMICE-712B-Wireless-Mouse-Laptop\/dp\/B0083K9I8Y\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Perixx PERIMICE-712B<\/a><\/li><li id=\"bd4a\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/kbdfans.cn\/products\/gk64-mechanical-keyboard-64key\" data-href=\"https:\/\/kbdfans.cn\/products\/gk64-mechanical-keyboard-64key\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">GK64 Keyboard<\/a> with Kailh Pro Purple switches and Godspeed DSA keycaps<\/li><li id=\"8b42\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/Blue-Yeti-USB-Microphone-Silver\/dp\/B0170NWLWY\" data-href=\"https:\/\/www.amazon.com\/Blue-Yeti-USB-Microphone-Silver\/dp\/B0170NWLWY\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Blue Yeti Microphone<\/a> in Space Grey<\/li><li id=\"5c9d\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/Blue-Radius-II-Microphone-Improved\/dp\/B00TTQLA50\" data-href=\"https:\/\/www.amazon.com\/Blue-Radius-II-Microphone-Improved\/dp\/B00TTQLA50\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">BLUE Radius II Shock Mount for Yeti\/Yeti Pro Microphone<\/a><\/li><li id=\"3ab5\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/Logitech-BRIO-Conferencing-Recording-Streaming\/dp\/B01N5UOYC4\" data-href=\"https:\/\/www.amazon.com\/Logitech-BRIO-Conferencing-Recording-Streaming\/dp\/B01N5UOYC4\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Logitech BRIO Ultra HD webcam<\/a><\/li><li id=\"59d3\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/gp\/product\/B018G8E6BG\/ref=oh_aui_detailpage_o04_s00?ie=UTF8&amp;psc=1\" data-href=\"https:\/\/www.amazon.com\/gp\/product\/B018G8E6BG\/ref=oh_aui_detailpage_o04_s00?ie=UTF8&amp;psc=1\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">iQunix laptop stand<\/a><\/li><li id=\"3cef\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/gp\/product\/B00HL7Z46K\/ref=oh_aui_search_detailpage?ie=UTF8&amp;psc=1\" data-href=\"https:\/\/www.amazon.com\/gp\/product\/B00HL7Z46K\/ref=oh_aui_search_detailpage?ie=UTF8&amp;psc=1\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">12 port powered USB hub<\/a><\/li><li id=\"39fa\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.amazon.com\/Programmable-KVM-Keypads-Keyboards-Gray\/dp\/B01AC6OYCG\/\" data-href=\"https:\/\/www.amazon.com\/Programmable-KVM-Keypads-Keyboards-Gray\/dp\/B01AC6OYCG\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">XKeys 16 key programmable keypad<\/a><\/li><\/ul><p id=\"4366\" class=\"graf graf--p graf-after--li\">This is my desk these days (I upgraded from my $30 IKEA table to an Uplift motorised standing desk):<\/p><figure id=\"f4bc\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*obShOzDcFnjR1jXMyL9Flw.jpeg\" data-width=\"4022\" data-height=\"2156\" data-is-featured=\"true\" data-action=\"zoom\" data-action-value=\"1*obShOzDcFnjR1jXMyL9Flw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*obShOzDcFnjR1jXMyL9Flw.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><h3 id=\"4f82\" class=\"graf graf--h3 graf-after--figure\">Stream design in\u00a0OBS<\/h3><p id=\"2ac8\" class=\"graf graf--p graf-after--h3\">I use <a href=\"https:\/\/obsproject.com\/\" data-href=\"https:\/\/obsproject.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Open Broadcasting Software<\/a> to composite my audio and video for streaming to Twitch. There are other packages out there, but I started with this one and still prefer it after trying a couple of other options out briefly. OBS has been solid for me since I began streaming over two and a half years ago.<\/p><p id=\"d324\" class=\"graf graf--p graf-after--p\">A few people have asked me about <a href=\"https:\/\/streamlabs.com\/\" data-href=\"https:\/\/streamlabs.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs OBS <\/a>(SLOBS), however because I stream on macOS as an operating system SLOBS is not available to me as an option.<\/p><p id=\"0427\" class=\"graf graf--p graf-after--p\">Let\u2019s dig a little deeper into how I have OBS set up!<\/p><h4 id=\"8c6b\" class=\"graf graf--h4 graf-after--p\">My OBS\u00a0scenes<\/h4><p id=\"ea83\" class=\"graf graf--p graf-after--h4\">I meticulously organise my scenes in order to make them easy to manage and update. I name everything with prefixes, and delete any sources within them that I am no longer using. I have four main scenes that I use: Standby, Self, Desktop, and Offline. I\u2019ll detail each one.<\/p><p id=\"58aa\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Standby scene<\/strong><\/p><p id=\"9a10\" class=\"graf graf--p graf-after--p\">This scene contains an image telling my stream viewers that I\u2019m about to start. A scrolling text box was added to show that the stream is not \u201cfrozen\u201d or buffering. A small comforting detail for those waiting. I often also ensure my music is playing in the background too. The illustration is by <a href=\"https:\/\/twitter.com\/emilywithcurls\" data-href=\"https:\/\/twitter.com\/emilywithcurls\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Emily Griffin<\/a>.<\/p><figure id=\"76b4\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*zw5r8pIRLNZ08WbAP3tRaA.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*zw5r8pIRLNZ08WbAP3tRaA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*zw5r8pIRLNZ08WbAP3tRaA.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"1b64\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Offline scene<\/strong><\/p><p id=\"0e06\" class=\"graf graf--p graf-after--p\">Similar to the Standby scene, this contains an image and nothing else. I switch to it when my stream is ending in a few seconds, and I\u2019m getting ready to stop everything on the OBS side. My Twitch offline poster uses the same image, so it\u2019s seamless on the channel when the livestream stops. The illustration is by <a href=\"https:\/\/twitter.com\/emilywithcurls\" data-href=\"https:\/\/twitter.com\/emilywithcurls\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Emily Griffin<\/a>.<\/p><figure id=\"e1b2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*hl-J75gSmnCF7_yTF72g8g.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*hl-J75gSmnCF7_yTF72g8g.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*hl-J75gSmnCF7_yTF72g8g.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"6189\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Self scene<\/strong><\/p><p id=\"b67c\" class=\"graf graf--p graf-after--p\">I call this the \u201cself\u201d or \u201cselfie\u201d scene mostly because it shows myself in most of the video frame (right). I use this scene whenever I want to chit-chat with my viewers candidly (say if a test suite is running and we\u2019re waiting), to open the stream, and also when I\u2019m running through the \u201cshow and tell\u201d segment of my stream, so that I can show items up close. I include a chat stream on the left as well, so that the conversation context is not missing should I choose to share this video somewhere other than on Twitch. The chat box is a widget offered by <a href=\"https:\/\/streamlabs.com\" data-href=\"https:\/\/streamlabs.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs<\/a>.<\/p><figure id=\"62f4\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*B907dw_o24qeu7FeoPWGrw.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*B907dw_o24qeu7FeoPWGrw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*B907dw_o24qeu7FeoPWGrw.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"c0b1\" class=\"graf graf--p graf-after--figure\">How it looks live:<\/p><figure id=\"68a7\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*LnNvp77mhXcWpvC5tLeqag.gif\" data-width=\"500\" data-height=\"280\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*LnNvp77mhXcWpvC5tLeqag.gif\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"3677\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Desktop scene<\/strong><\/p><p id=\"094f\" class=\"graf graf--p graf-after--p\">This is by far the most involved and complicated scene I use! It\u2019s the one that is used for the majority of each stream. It has a lot of pieces, but on a basic level it shows my desktop ie. what I\u2019m working on, a webcam feed of me, and a chat box so that context is preserved in the videos when viewed off the Twitch platform. The chat box also serves as an indicator to folks as to the stream delay and how long it takes for me to see and respond their chat message.<\/p><figure id=\"81ee\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*qKpJW1jrJ9TwyQsUxaWfZQ.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*qKpJW1jrJ9TwyQsUxaWfZQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*qKpJW1jrJ9TwyQsUxaWfZQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"ccfa\" class=\"graf graf--p graf-after--figure\">Here is a list of everything in this scene:<\/p><ol class=\"postList\"><li id=\"f43c\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Header and footer.<\/strong> These show details about viewer interactions, the temperature of my study, what music is playing, and my social media handles. I will explain these in more detail later on.<\/li><li id=\"808d\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Desktop capture. <\/strong>I only capture a portion of my entire screen, so that I can have private areas dedicated to managing and moderating my stream live. More on this later. I use vim on stream for my code editor as it looks nice and clean \/ minimal when side by side with the other parts of the scene layout. In the friendliest tone possible, I literally don\u2019t care what code editor you use, and I don\u2019t care what you think about my choice. If you\u2019re interested, <a href=\"https:\/\/github.com\/noopkat\/dotfiles\" data-href=\"https:\/\/github.com\/noopkat\/dotfiles\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">my dotfiles are on Github<\/a>.<\/li><li id=\"2a9f\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Chat box.<\/strong> I use an overlay widget for this, offered by <a href=\"https:\/\/streamlabs.com\" data-href=\"https:\/\/streamlabs.com\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs<\/a>. It\u2019s added via a \u2018browser source\u2019 within OBS. I have taken the time to style the chat box to match the aesthetics of my stream overall. The colours and font also match my code editor theme, which is <a href=\"https:\/\/draculatheme.com\/\" data-href=\"https:\/\/draculatheme.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Dracula<\/a> and Inconsolata. <a href=\"https:\/\/medium.com\/@suzhinton\/customizing-your-livestream-chat-widget-2ff9de49e83f\" data-href=\"https:\/\/medium.com\/@suzhinton\/customizing-your-livestream-chat-widget-2ff9de49e83f\" class=\"markup--anchor markup--li-anchor\" target=\"_blank\">I wrote an article on how I did this<\/a> if you want to know more detail. I used <a href=\"https:\/\/www.pixelmator.com\/pro\/\" data-href=\"https:\/\/www.pixelmator.com\/pro\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pixelmator<\/a> to draw the blue frame that appears around the chat box, and then imported that as an image source into OBS.<\/li><li id=\"d722\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Webcam capture.<\/strong> It\u2019s fun to watch a stream featuring the person who is streaming, as it builds more intimacy with viewers. You can see me laughing at your joke, or cursing at a bug I\u2019m encountering in my code. I used <a href=\"https:\/\/www.pixelmator.com\/pro\/\" data-href=\"https:\/\/www.pixelmator.com\/pro\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pixelmator<\/a> to draw the blue frame around the webcam source, and then imported that as an image source into OBS.<\/li><li id=\"5f07\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Stream alerts.<\/strong> Whenever someone follows my channel, subscribes, hosts me, or cheers Bits, I want to celebrate that moment. In order to do this, I use the <a href=\"https:\/\/streamlabs.com\" data-href=\"https:\/\/streamlabs.com\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs<\/a> Alert Box overlay widget imported as a browser source within OBS. In a nutshell, it plays an animation over the top of the stream with the person\u2019s Twitch name in it. I have this source as the very top layer within the scene so that it when it shows an alert, everyone can see it clearly.<\/li><li id=\"eac7\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Stream Events.<\/strong> The alerts mentioned above are momentary, so I have the last 5 events listed in another <a href=\"https:\/\/streamlabs.com\" data-href=\"https:\/\/streamlabs.com\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs<\/a> widget called Event List. This is really not super necessary and doesn\u2019t add much, so I\u2019m thinking of removing it.<\/li><li id=\"ef79\" class=\"graf graf--li graf--startsWithSingleQuote graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\u2018Secret\u2019 overlay.<\/strong> This is an image sized and positioned to cover up the desktop capture region of this scene. I use it whenever I have to do something private on my stream, such as rotating API keys, or entering one time passwords on websites. It is set to hidden by default, and I have it linked to a keyboard shortcut to toggle its visibility.<\/li><\/ol><p id=\"cbe8\" class=\"graf graf--p graf-after--li\">How it looks live:<\/p><figure id=\"dafc\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*RlqUrhPFaMjnZkublWOKdQ.gif\" data-width=\"500\" data-height=\"280\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*RlqUrhPFaMjnZkublWOKdQ.gif\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"f87b\" class=\"graf graf--p graf-after--figure\">With a Streamlabs alert popping up:<\/p><figure id=\"948e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*SulOYkrfhx7gH8YnTc8IeQ.gif\" data-width=\"500\" data-height=\"280\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*SulOYkrfhx7gH8YnTc8IeQ.gif\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"a245\" class=\"graf graf--p graf-after--figure\">What the secret overlay looks like when visible:<\/p><figure id=\"817a\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*edwRt5gr5w4NJvJZVp3ijQ.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*edwRt5gr5w4NJvJZVp3ijQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*edwRt5gr5w4NJvJZVp3ijQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"8987\" class=\"graf graf--p graf-after--figure\">The adorable \u2018secret\u2019 illustration is by <a href=\"https:\/\/twitter.com\/emilywithcurls\" data-href=\"https:\/\/twitter.com\/emilywithcurls\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Emily Griffin<\/a>.<\/p><p id=\"d2d2\" class=\"graf graf--p graf-after--p\">You might have also noticed that I have a bunch of scenes showing in the OBS screenshots I have been sharing. There is a very handy tip that I want to share with you which is to use \u2018child scenes\u2019 within your main scenes, which OBS allows you to do. Why?<\/p><p id=\"224e\" class=\"graf graf--p graf-after--p\">There\u2019ll be certain sources that you\u2019ll want to group together as one unit, and the easiest method to do this is to create a new scene for them. Doing this will give you two major advantages\u200a\u2014\u200ayou can reuse the sources as a group in multiple scenes without having to duplicate individual sources one by one, and also when you update the child scene, all of your main scenes that use it will instantly get those updates at once. You can also drag and position multiple sources at once if they\u2019re contained in a child scene together. Child scenes aren\u2019t different from main scenes as far as OBS is concerned\u200a\u2014\u200aI\u2019m just calling them this to differentiate them from each other in functionality.<\/p><p id=\"0500\" class=\"graf graf--p graf-after--p\">My header and footer child scenes look like this:<\/p><figure id=\"d351\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*PcR3VJ65Hn1S-qK4Qo9NmQ.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*PcR3VJ65Hn1S-qK4Qo9NmQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*PcR3VJ65Hn1S-qK4Qo9NmQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><figure id=\"5e46\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*4Lo3f8ngwZzp29LYiBQtzw.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*4Lo3f8ngwZzp29LYiBQtzw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*4Lo3f8ngwZzp29LYiBQtzw.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"8144\" class=\"graf graf--p graf-after--figure\">I used <a href=\"https:\/\/www.pixelmator.com\/pro\/\" data-href=\"https:\/\/www.pixelmator.com\/pro\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pixelmator<\/a> to design the header.<\/p><p id=\"fa8d\" class=\"graf graf--p graf-after--p\">Here is a list of everything in these two scenes and how I do the magical live updating text:<\/p><ol class=\"postList\"><li id=\"fea8\" class=\"graf graf--li graf-after--p\">Heart \/ musical note emoji png image sources for aesthetic reasons<\/li><li id=\"e404\" class=\"graf graf--li graf-after--li\">Live text labels for last follower, last subscriber, and last donation. I run a desktop app from <a href=\"https:\/\/streamlabs.com\" data-href=\"https:\/\/streamlabs.com\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs<\/a> called \u201cStream Labels\u201d, which writes each individual stat from your stream to a text file, continually updating upon new stats. When you create a text source in OBS in your scene, you can configure the source to read (and live update) from a text file. I point OBS at the correct text file for each label I\u2019ve chosen to show. It\u2019s a bit tedious to set up, but doing this in a child scene makes this a set and forget endeavor if you want to show them in every scene.<\/li><li id=\"2031\" class=\"graf graf--li graf-after--li\">The music playing text label is a <a href=\"https:\/\/gist.github.com\/noopkat\/ea98f5a6ef197e636394ff7b270a16a0\" data-href=\"https:\/\/gist.github.com\/noopkat\/ea98f5a6ef197e636394ff7b270a16a0\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">custom AppleScript I wrote<\/a> to regularly dump the currently playing song in iTunes to a text file. I set it up to run as a launchd daemon. It\u2019s shown in OBS with the same text file source method as the other stream labels explained above. If you want to know why I use iTunes or how I find and choose my music, I cover that later on in this post.<\/li><li id=\"2207\" class=\"graf graf--li graf-after--li\">The study temperature text label was a joke back when my study would get really hot in the summers of NYC. I was visibly sweating and my coding ability suffered as the room would grow hotter during each stream. I did this with a <a href=\"https:\/\/github.com\/noopkat\/study-temp\" data-href=\"https:\/\/github.com\/noopkat\/study-temp\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">browser app that I made<\/a> that shows telemetry being sent from a micro-controller sitting on my desk. I am using this web page in OBS as a browser source, but it could also very easily just be another daemon writing to a text file. I should probably make time to convert it one day. If you want to know more about the micro-controller and sensors used, hit me up as that\u2019s a whole other thing not relevant here.<\/li><\/ol><p id=\"91dd\" class=\"graf graf--p graf-after--li\">The webcam and chat box child scenes look like this:<\/p><figure id=\"add0\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*fvhKvNIEav5fTPDi0y6NqQ.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*fvhKvNIEav5fTPDi0y6NqQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*fvhKvNIEav5fTPDi0y6NqQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><figure id=\"c82d\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*b3tUzugXVDTLdr-K1jjzXg.png\" data-width=\"1199\" data-height=\"1074\" data-action=\"zoom\" data-action-value=\"1*b3tUzugXVDTLdr-K1jjzXg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*b3tUzugXVDTLdr-K1jjzXg.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"b618\" class=\"graf graf--p graf-after--figure\">Another trick you might have noticed is I have created a couple of empty scenes named \u2018\u200a\u2014\u200a\u2014\u200a\u2014\u200a\u2014 \u2018 to act as separators for my scene list. It\u2019s silly but it works really well to separate your main scenes from your child scenes. I have prefixed the names of my main scenes with \u2018scene: \u2018 and the child scenes are prefixed with \u2018group: \u2018 so that I can identify them very easily.<\/p><p id=\"3982\" class=\"graf graf--p graf-after--p\">I have use handy prefixes on the names of all of my sources. This just makes them easier to spot by type and makes everything look neat.<\/p><p id=\"c924\" class=\"graf graf--p graf-after--p\">This is all very optional! You don\u2019t have to over architect your scenes in OBS. But I found that these pro tips have really helped me stay organised in OBS. It\u2019s pretty easy to make a mess when you\u2019re first playing around with scenes and sources. Be as organised or disorganised as you want to be\u200a\u2014\u200aI\u2019m not judging you.<\/p><h4 id=\"315a\" class=\"graf graf--h4 graf-after--p\">The thing where I\u2019m not capturing all of my\u00a0desktop<\/h4><p id=\"4f0c\" class=\"graf graf--p graf-after--h4\">I don\u2019t capture all of my desktop like a lot of streamers do. This is for one specific reason\u200a\u2014\u200aI only want to use one monitor screen. Why?<\/p><p id=\"f077\" class=\"graf graf--p graf-after--p\">For a couple of reasons, I have found it to be challenging to use an additional monitor screen to monitor Twitch chat messages, keep an eye on new follow events, etc. My peripheral vision only helps me so much, so having to remember to watch an additional screen while trying to focus on my code showing on my main monitor is a struggle. When I experimented with this setup, I just didn\u2019t look over at the other monitor as often as I should to keep up on viewer activity. I also noticed that every time I looked over to acknowledge a new follower or a funny chat message, my face would be turned away from the webcam, which breaks the personal connection with the viewers I was talking to. It looked kinda disconnected and strange. Imagine if someone physically turned their body away from you every time they talked to you in real life. It feels awkward!<\/p><p id=\"dd26\" class=\"graf graf--p graf-after--p\">Capturing only part of my desktop did not make things feel squished, and allowed me to keep all of those viewer activity helpers close enough for me to notice and read very quickly. Your milage may vary, and almost no other streamer agrees with me on this, but this is what works really well for me.<\/p><p id=\"13eb\" class=\"graf graf--p graf-after--p\">So what do I have in the \u2018secret\u2019 area of my desktop?<\/p><ol class=\"postList\"><li id=\"3d3c\" class=\"graf graf--li graf-after--p\">Twitch chat, popped out into a separate browser window<\/li><li id=\"dbe3\" class=\"graf graf--li graf-after--li\">The Streamlabs Alerts Box popped out into a browser window<\/li><li id=\"11e5\" class=\"graf graf--li graf-after--li\">The Streamlabs Event List popped out into a browser window<\/li><li id=\"2569\" class=\"graf graf--li graf-after--li\">iTunes mini player<\/li><\/ol><p id=\"44e4\" class=\"graf graf--p graf-after--li\">Here is what my entire desktop looks like when streaming:<\/p><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"d86a\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*fqcI9dByN9PVx44dTaNjGA.png\" data-width=\"2560\" data-height=\"1440\" data-action=\"zoom\" data-action-value=\"1*fqcI9dByN9PVx44dTaNjGA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*fqcI9dByN9PVx44dTaNjGA.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"72f2\" class=\"graf graf--p graf-after--figure\">My ideal setup would use one of those ultra-Wide monitors so that I have more room in the non-capture zone, but this is fine for now. A girl can dream, and maybe even budget for a new monitor soon.<\/p><h3 id=\"1bed\" class=\"graf graf--h3 graf-after--p\">My technical settings within\u00a0OBS<\/h3><p id=\"e11e\" class=\"graf graf--p graf-after--h3\">I\u2019m going to list the important video and output settings that I have found success with below.<\/p><h4 id=\"e695\" class=\"graf graf--h4 graf-after--p\">Video settings<\/h4><ul class=\"postList\"><li id=\"a63c\" class=\"graf graf--li graf-after--h4\">Base (Canvas) Resolution: 2560x1440<\/li><li id=\"eed8\" class=\"graf graf--li graf-after--li\">Output (Scaled) Resolution: 1280x720<\/li><li id=\"9bb3\" class=\"graf graf--li graf-after--li\">Downscale Filter: Bicubic (16 samples)<\/li><li id=\"4765\" class=\"graf graf--li graf-after--li\">Common FPS Values: 30<\/li><\/ul><h4 id=\"9870\" class=\"graf graf--h4 graf-after--li\">Output settings<\/h4><ul class=\"postList\"><li id=\"36fc\" class=\"graf graf--li graf-after--h4\">Encoder: x264 (but checked \u2018enforce streaming service encoder settings\u2019)<\/li><li id=\"3a47\" class=\"graf graf--li graf-after--li\">Do not rescale output<\/li><li id=\"a494\" class=\"graf graf--li graf-after--li\">Rate Control: CBR (Twitch performs better with consistent vs. variable)<\/li><li id=\"0be0\" class=\"graf graf--li graf-after--li\">Bitrate: 2000<\/li><li id=\"4338\" class=\"graf graf--li graf-after--li\">Keyframe interval: 0<\/li><li id=\"500a\" class=\"graf graf--li graf-after--li\">CPU usage: veryfast<\/li><li id=\"66fa\" class=\"graf graf--li graf-after--li\">Profile: main<\/li><li id=\"6395\" class=\"graf graf--li graf-after--li\">Tune: (none)<\/li><\/ul><p id=\"8c43\" class=\"graf graf--p graf-after--li\">So why the low output resolution and bitrate, I hear you ask? The answer is very blunt\u200a\u2014\u200ayour stream is not about you, it\u2019s about your viewers. Some of your viewers do not have blazing fast, modern internet connections to watch you in crisp high resolution. I was born and raised in Australia, so I am painfully aware of this and remember waiting 20 minutes just to finish pushing code to a remote SVN service. As long as they can read your code on your video feed, you\u2019re good. Keep things as low as possible in order to prevent buffering delays and intermittent dropped frames on your side as well. Your viewers will thank you by sticking around and returning each stream.<\/p><h4 id=\"8efa\" class=\"graf graf--h4 graf-after--p\">Audio<\/h4><p id=\"f6c9\" class=\"graf graf--p graf-after--h4\">This one is going to get complicated. I am streaming on macOS, which introduces some snags when working with audio sources in OBS. If you use Windows to stream\u200a\u2014\u200afeel free to disregard this entire section.<\/p><p id=\"bc2e\" class=\"graf graf--p graf-after--p\">The TLDR is: if you want music and other desktop sounds to play on your stream, use <a href=\"https:\/\/rogueamoeba.com\/freebies\/soundflower\/\" data-href=\"https:\/\/rogueamoeba.com\/freebies\/soundflower\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Soundflower<\/a> for a really pared back solution to get started with. Soundflower will allow you to use a virtual device to route your desktop sounds to. OBS can then use the Soundflower device as an audio source.<\/p><p id=\"b61c\" class=\"graf graf--p graf-after--p\">I personally have a more comprehensive setup, where I route audio to several virtual devices in order to use OBS as an audio mixer for my needs. <strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">It is entirely optional, please do not feel that this is at all necessary.<\/em><\/strong><\/p><p id=\"dfdb\" class=\"graf graf--p graf-after--p\">I use <a href=\"https:\/\/rogueamoeba.com\/loopback\/\" data-href=\"https:\/\/rogueamoeba.com\/loopback\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Loopback<\/a>, which is an audio routing tool that uses software only to help you move audio through virtual devices from various sources. I use two virtual devices in Loopback; one for my music (iTunes), and another for the sound effects I sometimes play on my stream.<\/p><p id=\"40d5\" class=\"graf graf--p graf-after--p\">The music related device pipes the music playing from iTunes into a virtual device I call \u2018Twitch Music\u2019. Here is a screengrab of what it looks like in Loopback:<\/p><figure id=\"d4c5\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*WHKCYbFfQ5T185JDIu2LAw.png\" data-width=\"2224\" data-height=\"1224\" data-action=\"zoom\" data-action-value=\"1*WHKCYbFfQ5T185JDIu2LAw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*WHKCYbFfQ5T185JDIu2LAw.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"e264\" class=\"graf graf--p graf-after--figure\">The sound effect related device is a simple \u2018pass-thru\u2019\u200a\u2014\u200ait\u2019ll accept any input audio you give to it. I cover how I play my sound effects via this pass-thru in the automation section of this post later on. However, here is a screengrab of what it looks like in Loopback:<\/p><figure id=\"cc2b\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*Ov9yZksVrQnAx17lY5WPKQ.png\" data-width=\"2224\" data-height=\"1224\" data-action=\"zoom\" data-action-value=\"1*Ov9yZksVrQnAx17lY5WPKQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Ov9yZksVrQnAx17lY5WPKQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"ba28\" class=\"graf graf--p graf-after--figure\">In OBS I set up an audio source for each of these virtual devices, including Soundflower. Soundflower is capturing the remaining sound that isn\u2019t music or a sound effect (such as Streamlabs tool alert sounds, error sounds from my computer programs, etc).<\/p><p id=\"10a2\" class=\"graf graf--p graf-after--p\">Here is a screengrab of the audio settings in OBS:<\/p><figure id=\"c0cb\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*K6c57BZIeqQnPEzv3OKyog.png\" data-width=\"1093\" data-height=\"882\" data-action=\"zoom\" data-action-value=\"1*K6c57BZIeqQnPEzv3OKyog.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*K6c57BZIeqQnPEzv3OKyog.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"baa6\" class=\"graf graf--p graf-after--figure\">Setting up these sources separately allows me to control each one individually, just like a mixer in real life. I can individually adjust the volume of my mic, music, sound effects, and misc. desktop sounds at will, which is really handy to have as a feature when streaming.<\/p><p id=\"7408\" class=\"graf graf--p graf-after--p\">Here\u2019s the mixer interface in OBS to show you what I mean:<\/p><figure id=\"952a\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*mCeIOUoVZZzJx6B5PUgqHw.png\" data-width=\"304\" data-height=\"374\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*mCeIOUoVZZzJx6B5PUgqHw.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"bd77\" class=\"graf graf--p graf-after--figure\">Last but not least\u200a\u2014\u200aall of this does not help me to also hear what I need to hear while streaming. I do not want to hear the music, but I do want to hear computer error sounds, and also Streamlabs follow and subscription alerts which help me to notice them and call them out. This is where I use the Audio MIDI Setup application (built into macOS) to solve this problem. I create a multi-output device, and check the boxes to include the audio sources I do want to hear. Just before I stream, I switch to this multi-output device as the output source for my laptop\u2019s audio. This routes the helpful sounds to me, and filters out the sounds I do not want. I call this device \u2018Twitch Personal\u2019 so that I remember it\u2019s for my own local use.<\/p><p id=\"ea40\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">If you want to conduct remote video calls with fellow programmers to \u201cpair\u201d with them on your stream, you will absolutely need to do this, or you won\u2019t be able to hear them speaking (only your viewers will hear them).<\/strong><\/p><p id=\"605e\" class=\"graf graf--p graf-after--p\">Here\u2019s what my multi-output device looks like:<\/p><figure id=\"de5b\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*x6bA4VU8fiGpa0RELkYDHw.png\" data-width=\"1486\" data-height=\"1148\" data-action=\"zoom\" data-action-value=\"1*x6bA4VU8fiGpa0RELkYDHw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*x6bA4VU8fiGpa0RELkYDHw.png\" alt=\"image\" \/><\/div><\/div><\/figure><h4 id=\"d8c5\" class=\"graf graf--h4 graf-after--figure\">Music<\/h4><p id=\"e8e4\" class=\"graf graf--p graf-after--h4\">I randomly source royalty free, creative commons licensed music from sites such as <a href=\"http:\/\/freemusicarchive.org\/\" data-href=\"http:\/\/freemusicarchive.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Free Music Archive<\/a> and <a href=\"https:\/\/bandcamp.com\/\" data-href=\"https:\/\/bandcamp.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Bandcamp<\/a>. There are services out there like <a href=\"https:\/\/www.pretzel.rocks\/\" data-href=\"https:\/\/www.pretzel.rocks\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pretzel Rocks<\/a>, but I want a specific music type for my channel that other streaming music resources don\u2019t offer.<\/p><p id=\"d6a0\" class=\"graf graf--p graf-after--p\">I use iTunes to play the music on stream because it supports AppleScript, and I have written scripts that automate the music and the song identification.<\/p><h3 id=\"8ac3\" class=\"graf graf--h3 graf-after--p\">Twitch channel page aesthetics and\u00a0emotes<\/h3><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"85b5\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*MVPLj0PlbaTLLK_rQ5RWGQ.png\" data-width=\"1783\" data-height=\"1349\" data-action=\"zoom\" data-action-value=\"1*MVPLj0PlbaTLLK_rQ5RWGQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*MVPLj0PlbaTLLK_rQ5RWGQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"a8b5\" class=\"graf graf--p graf-after--figure\">I have re-designed my Twitch channel page since I wrote my original post about streaming. The objectives were:<\/p><ul class=\"postList\"><li id=\"e958\" class=\"graf graf--li graf-after--p\">Bring a more cohesive look and feel to my channel<\/li><li id=\"71a8\" class=\"graf graf--li graf-after--li\">Include myself in some of the assets to create a personal touch<\/li><li id=\"22b8\" class=\"graf graf--li graf-after--li\">Make the channel page look friendly and unique<\/li><\/ul><p id=\"2c3a\" class=\"graf graf--p graf-after--li\">I hired <a href=\"https:\/\/twitter.com\/emilywithcurls\" data-href=\"https:\/\/twitter.com\/emilywithcurls\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Emily Griffin<\/a> to create all of the illustrations, panel banners, and emotes that you see on my twitch channel page. It was 100% worth the investment. I knew of Emily because we ran in similar social circles, and I\u2019d also seen her work online so I was certain that her style would be a great fit for that I wanted my channel to become.<\/p><p id=\"7088\" class=\"graf graf--p graf-after--p\">Put some effort into your panels below the channel video frame. Potential new viewers will often visit your channel\u2019s page when you\u2019re offline, so having a good set of panels to describe your channel and letting them know when your next stream is sets a great first impression and is generally helpful for them when deciding whether to follow up when you\u2019re live. They also help a lot when viewers stumble upon your stream when you\u2019re live, and are not sure what you\u2019re working on or what your stream is about.<\/p><h3 id=\"172f\" class=\"graf graf--h3 graf-after--p\">Stream moderation and\u00a0privacy<\/h3><h4 id=\"5d9d\" class=\"graf graf--h4 graf-after--h3\">AutoMod<\/h4><p id=\"e9b2\" class=\"graf graf--p graf-after--h4\">Twitch has a feature called AutoMod, which is really helpful to have moderating your channel\u2019s chat room. There is an entire moderation settings page on Twitch dedicated to different approaches you can use. Here are my main settings that I know to have made a positive difference to chat quality.<\/p><ul class=\"postList\"><li id=\"6b86\" class=\"graf graf--li graf-after--p\">I have my AutoMod set to Level 4 (ie. a lot of moderation).<\/li><li id=\"7e20\" class=\"graf graf--li graf-after--li\">I have a 4 second chat delay instated so that my human moderators (bless them I am not worthy) have a chance to remove any nasty or inappropriate content before everyone else sees it.<\/li><li id=\"7d83\" class=\"graf graf--li graf-after--li\">All folks wanting to leave chat messages need to have an email address registered with their account first.<\/li><li id=\"96a9\" class=\"graf graf--li graf-after--li\">You have to have been following my channel for a minimum of 30 minutes before you\u2019re allowed to chat in the chat room.<\/li><\/ul><p id=\"f72b\" class=\"graf graf--p graf-after--li\">My chat rules are:<\/p><ul class=\"postList\"><li id=\"b830\" class=\"graf graf--li graf-after--p\">Follow the Twitch Community Guidelines.<\/li><li id=\"6379\" class=\"graf graf--li graf-after--li\">Be kind!<\/li><li id=\"b596\" class=\"graf graf--li graf-after--li\">Respect other humans.<\/li><li id=\"0054\" class=\"graf graf--li graf-after--li\">Include everyone.<\/li><li id=\"ad91\" class=\"graf graf--li graf-after--li\">No language bashing.<\/li><li id=\"0d50\" class=\"graf graf--li graf-after--li\">Listen to the moderation team.<\/li><\/ul><p id=\"a0b8\" class=\"graf graf--p graf-after--li\">I also use <a href=\"https:\/\/beta.nightbot.tv\/\" data-href=\"https:\/\/beta.nightbot.tv\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Nightbot<\/a> for some of their moderation features, but have turned most of them off since Twitch introduced AutoMod.<\/p><h4 id=\"73be\" class=\"graf graf--h4 graf-after--p\">Human Mods\u00a0\u2764\ufe0f<\/h4><p id=\"8494\" class=\"graf graf--p graf-after--h4\">I do not know what I\u2019d do without my friendly and cherished human moderators. While not critical when first streaming, once you have built a meaningful following real people as moderators are very key. They are the backbone of the positive community that I have grown around my stream channel. I currently have 15 moderators officially. I\u2019d hazard a guess that on average 3\u20135 of them are present during my streams, and it varies in numbers and the people depending on the week.<\/p><p id=\"3c02\" class=\"graf graf--p graf-after--p\">I won\u2019t mention them here in fear of putting them on blast, but I want to thank them so much for their very hard and sometimes stressful work when weird \/ nasty stuff goes down (which is not super often, in all honestly and relief). My AutoMod settings are aggressively set to cut them some slack, and so they only need to step in if things get super weird \u2764<\/p><h4 id=\"e013\" class=\"graf graf--h4 graf-after--p\">Privacy<\/h4><p id=\"1d22\" class=\"graf graf--p graf-after--h4\">I try to keep a pretty locked down ship privacy-wise. Inevitably, I\u2019ll do something by accident that I shouldn\u2019t have from time to time, but what\u2019s most important is that I mitigate the risks as much as possible, and plan for when something goes wrong.<\/p><p id=\"5af0\" class=\"graf graf--p graf-after--p\">I use my secret blocker when needing to enter plain text secrets anywhere in my code as I mentioned earlier in this article. If I accidentally show a secret, such as running git diff or cat on a file I shouldn\u2019t have, I immediately rotate the secret \/ API key and move on after having a laugh at myself.<\/p><p id=\"1ef6\" class=\"graf graf--p graf-after--p\">I am also aware that there are certain techniques folks might use to phish personal information from me on stream. This can be something as innocent as sending me an Amazon product link. On that Amazon product link page, my shipping address zip code is displayed if I\u2019m logged in. This is just one example, but there are many ways for me to be baited into showing details I did not volunteer to.<\/p><p id=\"b0d9\" class=\"graf graf--p graf-after--p\">One option is to have a separate browser profile where I\u2019m logged out of everything. In all honesty this hasn\u2019t been a super practical option for me. So instead, I use a browser extension called <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/block-site-website-blocke\/eiimnmioipafcokbfikbljfdeojpcgbh?hl=en\" data-href=\"https:\/\/chrome.google.com\/webstore\/detail\/block-site-website-blocke\/eiimnmioipafcokbfikbljfdeojpcgbh?hl=en\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Block Site<\/a> to schedule blocking of certain websites during the time window that I am streaming in. This extension is actually designed for productivity reasons, but it\u2019s been helpful for me to have peace of mind. I add sites to it regularly when I think of them. I have all of the major shopping\/ordering sites covered, as well as email accounts, and the settings pages of many technical tools and websites I use online.<\/p><p id=\"1042\" class=\"graf graf--p graf-after--p\">If I\u2019m worried about the contents of a link posted in chat, I just don\u2019t click on it, and ask one of my mods to check it out for me.<\/p><p id=\"f4f5\" class=\"graf graf--p graf-after--p\">If you want a more detailed discussion, I talked about <a href=\"https:\/\/www.youtube.com\/watch?v=HSpFT3tZzyM\" data-href=\"https:\/\/www.youtube.com\/watch?v=HSpFT3tZzyM\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">privacy and security while streaming with Tanya Janca on her DevSlop channel<\/a>.<\/p><h3 id=\"5de8\" class=\"graf graf--h3 graf-after--p\">Automation<\/h3><p id=\"8aab\" class=\"graf graf--p graf-after--h3\">In truth, I don\u2019t have a lot of automation set up to streamline the streaming process, but I\u2019ll outline what I do have because it helps a lot. A few of these automated processes have been mentioned in previous sections of this post.<\/p><h4 id=\"b21e\" class=\"graf graf--h4 graf-after--p\">Chat automation<\/h4><p id=\"a973\" class=\"graf graf--p graf-after--h4\">I use <a href=\"https:\/\/beta.nightbot.tv\/\" data-href=\"https:\/\/beta.nightbot.tv\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Nightbot<\/a> not just for moderation, but for setting up special chat commands that will autoreply with commonly asked questions while I\u2019m streaming. This saves me a lot of time and distraction when folks drop in asking things such as \u201cwhat are you working on?\u201d, \u201cwhat is that terminal theme you\u2019re using?\u201d, and \u201ccan you share your dotfiles?\u201d. FYI, the chat commands set up for these questions are \u2018!whatamidoing\u2019, \u2018!editor\u2019, and \u2018!dotfiles\u2019 respectively.<\/p><p id=\"5a16\" class=\"graf graf--p graf-after--p\">Here is what it looks like in the chat:<\/p><figure id=\"5314\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*exDQjeCqvMKVH3O1J3nCMA.png\" data-width=\"351\" data-height=\"113\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*exDQjeCqvMKVH3O1J3nCMA.png\" alt=\"image\" \/><\/div><\/div><\/figure><h4 id=\"1c33\" class=\"graf graf--h4 graf-after--figure\">Keyboard shortcuts<\/h4><p id=\"518d\" class=\"graf graf--p graf-after--h4\">I have keyboard shortcuts set up for the following:<\/p><ul class=\"postList\"><li id=\"b106\" class=\"graf graf--li graf-after--p\">Switching to specific scenes in OBS<\/li><li id=\"d9cb\" class=\"graf graf--li graf-after--li\">Toggling the \u2018secret blocker\u2019 on the desktop scene in OBS<\/li><li id=\"8418\" class=\"graf graf--li graf-after--li\">Playing sound effects<\/li><li id=\"43c6\" class=\"graf graf--li graf-after--li\">Running an automated setup script before streaming<\/li><\/ul><p id=\"587c\" class=\"graf graf--p graf-after--li\">I use <a href=\"https:\/\/www.keyboardmaestro.com\/main\/\" data-href=\"https:\/\/www.keyboardmaestro.com\/main\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Keyboard Maestro<\/a> for all of these shortcuts, including sending shortcuts directly to OBS because there\u2019s a global keyboard shortcut bug in OBS on macOS so this is how I get around it.<\/p><p id=\"da06\" class=\"graf graf--p graf-after--p\">Remember earlier where I detailed a separate virtual device for sending sound effects to? Here is an example of a macro to do that in Keyboard Maestro:<\/p><figure id=\"bbce\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*icdYqfNhEjWSEQA9v75gLw.png\" data-width=\"1044\" data-height=\"629\" data-action=\"zoom\" data-action-value=\"1*icdYqfNhEjWSEQA9v75gLw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*icdYqfNhEjWSEQA9v75gLw.png\" alt=\"image\" \/><\/div><\/div><\/figure><h4 id=\"43f0\" class=\"graf graf--h4 graf-after--figure\">One giant AppleScript<\/h4><p id=\"5dca\" class=\"graf graf--p graf-after--h4\">I recently captured the entire pre-stream setup process into one AppleScript (and some companion bash scripts). This script is a true time saver every Sunday morning since I made it. This AppleScript does the following for me:<\/p><ol class=\"postList\"><li id=\"9af5\" class=\"graf graf--li graf-after--p\">Opens all of the browser windows I need (streamlabs alerts\/events, Twitch chat, etc) after minimising my personal browser window with all my private tabs<\/li><li id=\"a910\" class=\"graf graf--li graf-after--li\">Opens a new iTerm custom window arrangement, perfectly sized and aligned for my Twitch stream, after minimising my personal iTerm window with my day-to-day work in it.<\/li><li id=\"5ed3\" class=\"graf graf--li graf-after--li\">Switches my audio output source to \u2018Twitch Personal\u2019<\/li><li id=\"3281\" class=\"graf graf--li graf-after--li\">Starts up iTunes, positions it in the bottom right corner, and starts playing my custom Twitch playlist<\/li><li id=\"c621\" class=\"graf graf--li graf-after--li\">Starts up the Stream Labels app and minimises it<\/li><li id=\"0960\" class=\"graf graf--li graf-after--li\">Pops up a dialog asking to update my\u00a0!whatamidoing Nighbot command reply content.<\/li><\/ol><p id=\"7b5d\" class=\"graf graf--p graf-after--li\">The last step which updates my Nightbot command does the following:<\/p><ol class=\"postList\"><li id=\"48de\" class=\"graf graf--li graf-after--p\">Hits the Nightbot API and gets the last command reply I used from last week<\/li><li id=\"4d09\" class=\"graf graf--li graf-after--li\">Pre-populates the input field on the dialog with the last reply it fetched, and if I want to change it I can<\/li><li id=\"90dd\" class=\"graf graf--li graf-after--li\">Once I hit \u2018Save\u2019, it hits the Nighbot API again to set the command reply to the new one I typed in<\/li><li id=\"126a\" class=\"graf graf--li graf-after--li\">It\u2019ll then show a macOS notification with the JSON of the Nightbot API response so I can check if it worked ok.<\/li><\/ol><p id=\"36e0\" class=\"graf graf--p graf-after--li\">I assign a keyboard shortcut to run the AppleScript in Keyboard Maestro:<\/p><figure id=\"68ee\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*9tu8Jv7TK7J01I7Nciy8GA.png\" data-width=\"2132\" data-height=\"1210\" data-action=\"zoom\" data-action-value=\"1*9tu8Jv7TK7J01I7Nciy8GA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*9tu8Jv7TK7J01I7Nciy8GA.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"7313\" class=\"graf graf--p graf-after--figure\">Here is the Nightbot prompt:<\/p><figure id=\"95ca\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*Xg5dfuDm43L4c-BQPVn4hQ.png\" data-width=\"532\" data-height=\"271\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Xg5dfuDm43L4c-BQPVn4hQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"bf18\" class=\"graf graf--p graf-after--figure\">And the response:<\/p><figure id=\"af38\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*nnrsf6bNOk9XMsrqngmnyA.png\" data-width=\"916\" data-height=\"406\" data-action=\"zoom\" data-action-value=\"1*nnrsf6bNOk9XMsrqngmnyA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*nnrsf6bNOk9XMsrqngmnyA.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"f905\" class=\"graf graf--p graf-after--figure\">The <a href=\"https:\/\/github.com\/noopkat\/twitch-scripts\/blob\/master\/setup-twitch.applescript\" data-href=\"https:\/\/github.com\/noopkat\/twitch-scripts\/blob\/master\/setup-twitch.applescript\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">final AppleScript is here<\/a> if you\u2019d like to read it.<\/p><p id=\"4a4e\" class=\"graf graf--p graf-after--p\">And I have hosted all of my scripts including this one in a <a href=\"https:\/\/github.com\/noopkat\/twitch-scripts\" data-href=\"https:\/\/github.com\/noopkat\/twitch-scripts\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">repository on Github<\/a>.<\/p><h3 id=\"d40c\" class=\"graf graf--h3 graf-after--p\">Round Up<\/h3><p id=\"0479\" class=\"graf graf--p graf-after--h3\">I have collected all of the tools and resources mentioned in this post into one big list for your convenience:<\/p><ul class=\"postList\"><li id=\"c9ea\" class=\"graf graf--li graf-after--p\"><a href=\"https:\/\/obsproject.com\/\" data-href=\"https:\/\/obsproject.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Open Broadcasting Software<\/a><\/li><li id=\"234c\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/twitter.com\/emilywithcurls\" data-href=\"https:\/\/twitter.com\/emilywithcurls\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Emily Griffin<\/a><\/li><li id=\"355f\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/streamlabs.com\/\" data-href=\"https:\/\/streamlabs.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Streamlabs<\/a><\/li><li id=\"4154\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/draculatheme.com\/\" data-href=\"https:\/\/draculatheme.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Dracula Theme<\/a><\/li><li id=\"e5fa\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/medium.com\/@suzhinton\/customizing-your-livestream-chat-widget-2ff9de49e83f\" data-href=\"https:\/\/medium.com\/@suzhinton\/customizing-your-livestream-chat-widget-2ff9de49e83f\" class=\"markup--anchor markup--li-anchor\" target=\"_blank\">Customising chat widget article<\/a><\/li><li id=\"f9a5\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.pixelmator.com\/pro\/\" data-href=\"https:\/\/www.pixelmator.com\/pro\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pixelmator<\/a><\/li><li id=\"7ddd\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/rogueamoeba.com\/freebies\/soundflower\/\" data-href=\"https:\/\/rogueamoeba.com\/freebies\/soundflower\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Soundflower<\/a><\/li><li id=\"e2ad\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/rogueamoeba.com\/loopback\/\" data-href=\"https:\/\/rogueamoeba.com\/loopback\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Loopback<\/a><\/li><li id=\"ef42\" class=\"graf graf--li graf-after--li\"><a href=\"http:\/\/freemusicarchive.org\/\" data-href=\"http:\/\/freemusicarchive.org\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Free Music Archive<\/a><\/li><li id=\"4086\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/bandcamp.com\/\" data-href=\"https:\/\/bandcamp.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Bandcamp<\/a><\/li><li id=\"ba17\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.pretzel.rocks\/\" data-href=\"https:\/\/www.pretzel.rocks\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pretzel Rocks<\/a><\/li><li id=\"7dd8\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/beta.nightbot.tv\/\" data-href=\"https:\/\/beta.nightbot.tv\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Nightbot<\/a><\/li><li id=\"544a\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/block-site-website-blocke\/eiimnmioipafcokbfikbljfdeojpcgbh?hl=en\" data-href=\"https:\/\/chrome.google.com\/webstore\/detail\/block-site-website-blocke\/eiimnmioipafcokbfikbljfdeojpcgbh?hl=en\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Block Site<\/a><\/li><li id=\"cf70\" class=\"graf graf--li graf-after--li\">P<a href=\"https:\/\/www.youtube.com\/watch?v=HSpFT3tZzyM\" data-href=\"https:\/\/www.youtube.com\/watch?v=HSpFT3tZzyM\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">rivacy and security stream with Tanya Janca on DevSlop<\/a><\/li><li id=\"e91e\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.keyboardmaestro.com\/main\/\" data-href=\"https:\/\/www.keyboardmaestro.com\/main\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Keyboard Maestro<\/a><\/li><li id=\"9be5\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/gist.github.com\/noopkat\/ea98f5a6ef197e636394ff7b270a16a0\" data-href=\"https:\/\/gist.github.com\/noopkat\/ea98f5a6ef197e636394ff7b270a16a0\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">AppleScript for polling current song playing in iTunes<\/a><\/li><li id=\"e3ba\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/github.com\/noopkat\/twitch-scripts\/blob\/master\/setup-twitch.applescript\" data-href=\"https:\/\/github.com\/noopkat\/twitch-scripts\/blob\/master\/setup-twitch.applescript\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Ultimate setup AppleScript<\/a><\/li><li id=\"8a45\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/github.com\/noopkat\/study-temp\" data-href=\"https:\/\/github.com\/noopkat\/study-temp\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Study temperature browser app<\/a><\/li><li id=\"20e4\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/github.com\/noopkat\/twitch-scripts\" data-href=\"https:\/\/github.com\/noopkat\/twitch-scripts\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Collective Twitch scripts repository<\/a><\/li><\/ul><h3 id=\"1c17\" class=\"graf graf--h3 graf-after--li\">Conclusion<\/h3><p id=\"9591\" class=\"graf graf--p graf-after--h3 graf--trailing\">I hope this has been informative for you. I have tried to share as much detail as possible and address the most common questions I get about my setup. If you have additional questions, let me know. Just keep in mind\u200a\u2014\u200athis is what works for me. What works for you may be entirely different. Happy streaming!<\/p><\/div><\/div>","created_at":"2019-02-25T03:04:28+0000","updated_at":"2019-02-25T18:26:36+0000","published_at":"2019-02-20T17:40:42+0000","published_by":["Suz Hinton"],"annotations":[],"mimetype":"text\/html","reading_time":23,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*obShOzDcFnjR1jXMyL9Flw.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":857,"title":"threatgrid\/asami","url":"https:\/\/github.com\/threatgrid\/asami\/blob\/master\/README.md","content":"<p>An implementation of the Naga storage protocol in memory.<\/p>\n<p><a href=\"http:\/\/clojars.org\/org.clojars.quoll\/asami\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/37a2eafcc12d17109d097fd0bdba1171702b8b67\/687474703a2f2f636c6f6a6172732e6f72672f6f72672e636c6f6a6172732e71756f6c6c2f6173616d692f6c61746573742d76657273696f6e2e737667\" alt=\"Clojars Project\" data-canonical-src=\"http:\/\/clojars.org\/org.clojars.quoll\/asami\/latest-version.svg\" \/><\/a><\/p>\n<p>This has a query API that looks very similar to a simplified Datomic.<\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>Create a store with:<\/p>\n<p><code>(asami.core\/create-store config)<\/code><\/p>\n<p>Where <em>config<\/em> is a standard Naga config, but will be ignored.<\/p>\n<p>Alternatively, just use the <code>asami.core\/empty-store<\/code> object.<\/p>\n<p>Add data with:\n<code>naga.store\/assert-data<\/code><\/p>\n<p>Query using:\n<code>naga.store\/query<\/code><\/p>\n<p>Look at the <code>naga.store\/Storage<\/code> protocol (in <a href=\"https:\/\/github.com\/threatgrid\/naga-store\">threatgrid\/naga-store<\/a> ) for the full suite of functions to apply to the storage.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2018 Cisco<\/p>\n<p>Distributed under the Eclipse Public License either version 1.0 or (at\nyour option) any later version.<\/p>","created_at":"2019-02-24T18:14:45+0000","updated_at":"2019-02-24T18:14:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/337733?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":856,"title":"A 3D engine built using HTML and CSS 3D transforms by Keith Clark","url":"https:\/\/keithclark.co.uk\/labs\/css-fps\/","content":"A 3D engine built using HTML and CSS 3D transforms by Keith Clark\n<h2 id=\"please-select-a-demo\">Please select a demo:<\/h2>\n<div class=\"demos\"><a href=\"https:\/\/keithclark.co.uk\/labs\/css-fps\/desktop\/\" class=\"demo\">\n<img src=\"https:\/\/keithclark.co.uk\/labs\/css-fps\/screen1.jpg\" width=\"400\" height=\"282\" alt=\"image\" \/><h2 id=\"mouse-amp-keyboard\">Mouse &amp; keyboard<\/h2>\n<p>Runs best in a desktop browser<\/p>\n<\/a><a href=\"https:\/\/keithclark.co.uk\/labs\/css-fps\/mobile\/\" class=\"demo\">\n<img src=\"https:\/\/keithclark.co.uk\/labs\/css-fps\/screen2.jpg\" width=\"400\" height=\"282\" alt=\"image\" \/><h2 id=\"touch-screen\">Touch screen<\/h2>\n<p>Runs best on a mobile browser<\/p>\n<\/a><a href=\"https:\/\/keithclark.co.uk\/labs\/css-fps\/nojs\/\" class=\"demo\">\n<img src=\"https:\/\/keithclark.co.uk\/labs\/css-fps\/screen3.jpg\" width=\"400\" height=\"282\" alt=\"image\" \/><h2 id=\"pure-css-demo\">Pure CSS demo<\/h2>\n<p>Just sit back and watch<\/p>\n<\/a><\/div>\n<p>This is demo of a CSS powered 3D environment. Geometry is created with HTML elements and CSS transforms. Textures and lightmaps are composed by layering multiple background-images and colour is applied using CSS blend-modes.<\/p>\n<p>A <a href=\"https:\/\/keithclark.co.uk\/\">Keith Clark<\/a> experiment.<\/p>","created_at":"2019-02-24T15:47:07+0000","updated_at":"2019-02-24T15:47:07+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"keithclark.co.uk","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","gamedev","clojurescript"],"is_public":false,"id":855,"title":"Game Development with Clojure\/ClojureScript","url":"https:\/\/lambdaisland.com\/blog\/08-12-2016-game-development-with-clojure-clojurescript","content":"<p>This weekend it\u2019s <a shape=\"rect\" href=\"https:\/\/ldjam.com\/\">Ludum Dare<\/a> again, the world\u2019s longest\nrunning game jam. The idea is that, alone or with a team, you build a game in a\nweekend based on a certain theme.<\/p><p>We got a little team together here in Berlin, and so I\u2019ve been reviewing what\noptions there are for someone wanting to build a game in Clojure or\nClojurescript.<\/p><p>The good news is there are plenty of options, as you\u2019ll see from the list below.\nYou can do desktop games, browser based games with canvas or webgl, and you can\neven create <a shape=\"rect\" href=\"https:\/\/unity3d.com\/\">Unity 3D<\/a> games, all from your comfortable\nClojure parentheses.<\/p><p>That\u2019s more or less where the good news ends. Most of these frameworks and\nlibraries are still alpha-level software, many have a \u201cuse at your own risk,\nyou\u2019re on your own\u201d kind of warning on the project page. And the\ndocumentation\u2026 well\u2026 let\u2019s just say that you\u2019re likely better off studying\n<a shape=\"rect\" href=\"https:\/\/en.wikipedia.org\/wiki\/Oracle_bone\">cracked turtle bones<\/a> to figure out\nhow these libraries work.<\/p><p>Disclaimer: most of these I haven\u2019t actually tried, certainly not in anger, so\nmuch of what follows is from what I could gather looking at the project page and\nperusing some code samples and (if present) docs. Feel free to \u201cwell actually\u201d\nme <a shape=\"rect\" href=\"https:\/\/twitter.com\/lambdaisland\">on Twitter<\/a> if you feel I\u2019m\nmisrepresenting things.<\/p><h2 id=\"desktopmobile-offerings\">Desktop\/mobile offerings<\/h2><h3 id=\"quilhttpquilinfo\"><a shape=\"rect\" href=\"http:\/\/quil.info\/\">Quil<\/a><\/h3><p>Platforms: desktop, web<br style=\"clear: none;\" \/>\nBased on: Processing \/ Processing.js<br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc74<br style=\"clear: none;\" \/>\nDocumentation: \ud83d\ude0a<br style=\"clear: none;\" \/><\/p><p>Quil is Clojure\u2019s version of the graphics programming toolkit Processing. It\u2019s\nbeginner friendly, and really lends itself to exploration, which is also why\nClojureBridge chose it for the official curriculum.<\/p><p>It\u2019s been around for a while and \u201cjust works\u201d. The project site has decent API\ndocs, there\u2019s a wiki with useful info, and you can find plenty of examples\nfloating around, like the ones I created last year for\n<a shape=\"rect\" href=\"http:\/\/landofquil.we-do-fp.berlin\/\">The Land of Quil<\/a>, or the ones that are\nbeing worked on as part of the\n<a shape=\"rect\" href=\"https:\/\/github.com\/clojurebridge-berlin\/study-group-curriculum\">ClojureBridge Berlin Study Group Curriculum<\/a>.<\/p><p>It\u2019s also the only offering on this list that\u2019s actually cross platform. You do\nhave to keep a few small things in mind, but it\u2019s very doable to have a single\ncode base work for both desktop and in the browser, which is pretty cool.<\/p><p>The only downside is that for game development Quil is pretty limited. Its sweet\nspot lies more in generative art and interactive graphics. You get a bunch of\ndrawing primitives, and some functions for mouse and keyboard input, but beyond\nthat you\u2019re on your own. No sprites, physics, collision detection, scenes,\nlayers, sound, tweening, or event system. It\u2019s still a lot of fun to play around\nwith, but if you only have a weekend then maybe tone down your ambitions.<\/p><h3 id=\"play-cljhttpsgithubcomoakesplay-clj\"><a shape=\"rect\" href=\"https:\/\/github.com\/oakes\/play-clj\">Play-clj<\/a><\/h3><p>Platforms: desktop, Android<br style=\"clear: none;\" \/>\nBased on: <a shape=\"rect\" href=\"https:\/\/libgdx.badlogicgames.com\/\">libGDX<\/a><br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc69<br style=\"clear: none;\" \/>\nDocumentation: \ud83d\ude0a<br style=\"clear: none;\" \/><\/p><p>Play-clj is Zach Oakes game framework for Clojure. There\u2019s a big warning on the\nREADME that he\u2019s focusing his efforts now on Play-cljs, which I\u2019ll cover further\ndown. Zach is the author of <a shape=\"rect\" href=\"https:\/\/sekao.net\/nightcode\/\">Nightcode<\/a>, a\nbeginner friendly editor\/IDE for Clojure. He\u2019s also created a Nightcode spin-off\ncalled <a shape=\"rect\" href=\"https:\/\/sekao.net\/nightmod\/\">Nightmod<\/a>, specifically for game\ndevelopment using Play-clj. It calls itself \u201ca tool for making live-moddable\ngames\u201d. The idea is clear: bring REPL-like interactive development to game dev.<\/p><p>Despite the warning Play-clj seems pretty solid. It\u2019s the only offering in this\nlist that explicitly mentions mobile as a target. Play-clj games use a\n<a shape=\"rect\" href=\"https:\/\/en.wikipedia.org\/wiki\/Entity%E2%80%93component%E2%80%93system\">Entity-Component System<\/a>\nwhich is pretty standard in game development nowadays. It\u2019s based on libGDX\nwhich is a feature rich game dev framework, and as such provides many features\nyou\u2019d expect, like sprites or physics simulation.<\/p><p>There\u2019s also comprehensive documentation, not just API docs but also (bless the\nauthor) and actual step-by-step tutorial. It doesn\u2019t seem like much to ask but\nit\u2019s a rare offering in the documentation wasteland you\u2019ll encounter further\ndown.<\/p><p>Oh and Zach (I can say Zach, right?) spoke about Play-clj at Clojure\/conj!\n<a shape=\"rect\" href=\"https:\/\/www.youtube.com\/watch?v=0GzzFeS5cMc\">Check out the talk: Making Games at Runtime With Clojure<\/a>.<\/p><h3 id=\"arcadiahttpsgithubcomarcadia-unityarcadia\"><a shape=\"rect\" href=\"https:\/\/github.com\/arcadia-unity\/Arcadia\">Arcadia<\/a><\/h3><p>Platforms: desktop<br style=\"clear: none;\" \/>\nBased on: <a shape=\"rect\" href=\"https:\/\/unity3d.com\/\">Unity<\/a><br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc66<br style=\"clear: none;\" \/>\nDocumentation: \ud83e\udd14<br style=\"clear: none;\" \/><\/p><p>Arcadia is definitely the odd in one out in this list, but it\u2019s such a cool and\npromising project that it wouldn\u2019t stand not to mention it. Most readers of this\nblog will know that there is a JVM and a Javascript based version of Clojure,\nbut there\u2019s a lesser known implementation called ClojureCLR, targeting\nMicrosoft\u2019s Common Language Runtime, the virtual machine that powers .NET.<\/p><p>This is what enabled Arcadia to integrate Clojure with Unity, a bona fide\nA-level professional game development environment. This also means the workflow\nis a bit other than usual. Unity comes with its own coding environment, and to\nget started with Arcadia you clone the Arcadia source into your Unity project,\nalthough this might change as the project matures.<\/p><p>Unity is a commercial offering, it is not free software\/open source, but there\u2019s\na free plan for hobby use.<\/p><p>Documentation is still limited, but there are a handful of wiki pages that do\nseem to contain all you need to get rolling.<\/p><p>Arcadia is development by <a shape=\"rect\" href=\"https:\/\/twitter.com\/timsgardner\">Tim Gardner<\/a> and\n<a shape=\"rect\" href=\"https:\/\/twitter.com\/ra\">Ramsey Nasser<\/a> from\n<a shape=\"rect\" href=\"http:\/\/kitchentablecoders.com\/about\/\">Kitchen Table Coders<\/a>, the same group\nthat <a shape=\"rect\" href=\"https:\/\/twitter.com\/swannodette\">David Nolen<\/a> is part of.<\/p><h2 id=\"browser-based\">Browser based<\/h2><p>Now we come to the web part, games that are playable directly in your browser.\nNote that we already mentioned Quil at the top. Since it\u2019s cross platform it\nreally belongs in both categories.<\/p><p>To evaluate these libraries it\u2019s good to first have a look at what powers them\nunder the hood. Whatever superpowers they have will be inherited directly from\nwhatever javascript library is driving them.<\/p><p>Some of these only provide a thin wrapper for Clojurescript, meaning they don\u2019t\nbring anything new to the table, but merely provide a more idiomatic syntax,\navoiding lots of interop calls. Others try to add value of their own.<\/p><h3 id=\"play-cljshttpsoakesgithubioplay-cljs\"><a shape=\"rect\" href=\"https:\/\/oakes.github.io\/play-cljs\/\">Play-cljs<\/a><\/h3><p>Platforms: browser<br style=\"clear: none;\" \/>\nBased on: <a shape=\"rect\" href=\"http:\/\/p5js.org\/\">p5.js<\/a>\nMaturity: \ud83d\udc76<br style=\"clear: none;\" \/>\nDocumentation: \ud83e\udd14<br style=\"clear: none;\" \/><\/p><p>P5.js is a reimagination of Processing.js, the library Quil is based on. It\nwants to \u201cmake coding accessible for artists, designers, educators, and\nbeginners\u201d. That\u2019s a great goal, but just like with Quil this means that handy\nfeatures specifically for game development are a bit thin on the ground.<\/p><p>Play-cljs seems to mostly be a thin wrapper around p5.js, although it\u2019s possible\nthe author plans to build more gaming features in. Note that despite the one\nletter difference, play-clj and play-cljs are completely different animals, with\ncompletely different feature sets and APIs.<\/p><p>Zach tends to do a pretty good job at documenting things, although there\u2019s no\nstep-by-step tutorial in this case. You do get a bunch of example games, and\ngenerated API docs, which for such a minimal library really might be enough.<\/p><h3 id=\"chocolatierhttpsgithubcomalexkehayiaschocolatier\"><a shape=\"rect\" href=\"https:\/\/github.com\/alexkehayias\/chocolatier\">Chocolatier<\/a><\/h3><p>Platforms: browser<br style=\"clear: none;\" \/>\nBased on: <a shape=\"rect\" href=\"http:\/\/www.pixijs.com\/\">Pixi.js<\/a><br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc76<br style=\"clear: none;\" \/>\nDocumentation: \ud83d\ude20<br style=\"clear: none;\" \/><\/p><p>Chocolatier seems pretty impressive, it\u2019s based on Pixi.js which prizes itself\nat being \u201cthe fastest, most flexible 2D WebGL renderer\u201d. It has few hundred\nGithub stars, and it seems a lot of thought has gone into its design. The author\neven did a\n<a shape=\"rect\" href=\"https:\/\/www.youtube.com\/watch?v=TW1ie0pIO_E\">conj talk about Functional Game Design<\/a>,\nand the README talks a lot about its \u201cmodified entity component system\u201d, but I\u2019m\nreally missing a gentler introduction on how to actually <em>use<\/em> the thing.<\/p><p>There\u2019s a big example in the README that\u2019s more or less provided without\ncomment. There\u2019s a documentation folder, with a single file <code>Intro.md<\/code>,\ncontaining a single line: \u201cTODO: write great documentation\u201d.<\/p><p>It\u2019s really unfortunate, I feel like if I had a day or two to sink my teeth into\nthis, preferrably well rested and with few distractions, I could get to a point\nwhere I could have a lot of fun with it. Sigh\u2026 oh well, I guess as a creator\nof accessible technical content it\u2019s good to know there\u2019s still plenty of work\nto be done.<\/p><h3 id=\"phzrhttpsgithubcomdparisphzr\"><a shape=\"rect\" href=\"https:\/\/github.com\/dparis\/phzr\">Phzr<\/a><\/h3><p>Platforms: browser<br style=\"clear: none;\" \/>\nBased on: <a shape=\"rect\" href=\"http:\/\/phaser.io\/\">Phaser<\/a><br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc76<br style=\"clear: none;\" \/>\nDocumentation: \ud83e\udd14<br style=\"clear: none;\" \/><\/p><p>Phzr is just a thin wrapper around Phaser, one of the more popular Javascript\ngame engines. Documentation is limited, and instead you\u2019re simply referred to\nthe Phaser docs. This of course is the benefit of thin wrappers, as long as the\nunderlying library is well documented you\u2019re all golden. So I think this\nactually would make a pretty good choice, although the question is if you really\nneed a wrapper then.<\/p><h3 id=\"re-framehttpsgithubcomday8re-framewiki--svg\"><a shape=\"rect\" href=\"https:\/\/github.com\/Day8\/re-frame\/wiki\">re-frame<\/a> + SVG<\/h3><p>Platforms: browser<br style=\"clear: none;\" \/>\nBased on: Reagent\/React<br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc66<br style=\"clear: none;\" \/>\nDocumentation: \ud83e\udd14<br style=\"clear: none;\" \/><\/p><p>Not a game library as such, but still something I would consider a valid option.\nBrowser support for SVG has improved rapidly in recent years, and React hasn\u2019t\nmissed the boat either. The result is that you get drawing primitives right in\nyour DOM, which can then be styled and animated, and which generate DOM events that\nyou can respond to.<\/p><p>In that sense it\u2019s no less basic than Quil or play-cljs. You could use and React\nwrapper like Reagent, Om, Quiescent, or Rum, but my favorite by far is re-frame,\nand I believe its event system is uniquely suited for use with games.<\/p><p>And re-frame has some pretty decent docs. It\u2019s a bit of a weird style, and the\nREADME is a bit of work to get through, but it does explain well how it works,\nand there are wiki pages explaining all the core concepts. Some even say the\nre-frame docs are like the hidden documentation for Reagent, the library\nre-frame is built on. Some Reagent concepts are only properly explained here.<\/p><p>You might have to be a bit more performance-conscious than with other options,\nbut make good use of React\/Reagent\u2019s optimizations to avoid unnecessary rendering, and keep in mind that\n<a shape=\"rect\" href=\"https:\/\/www.html5rocks.com\/en\/tutorials\/speed\/high-performance-animations\/\">some things are cheaper to animate than others<\/a>.<\/p><p>Finally if I would go this route I would combine it with the excellent\n<a shape=\"rect\" href=\"https:\/\/github.com\/thi-ng\/geom\/\">thi.ng\/geom<\/a> 2D\/3D geometry toolkit for\nClojure\/Clojurescript. The same goes for Quil or Play-cljs.<\/p><h3 id=\"javascript-engines\">Javascript engines<\/h3><p>Platforms: browser<br style=\"clear: none;\" \/>\nMaturity: \ud83d\udc75<br style=\"clear: none;\" \/>\nDocumentation: \ud83d\ude03<br style=\"clear: none;\" \/><\/p><p>Finally a very compelling option is to simply forget about Clojurescript\nspecific libraries, and instead pick any of the popular Javascript game engines.\nThere are\n<a shape=\"rect\" href=\"https:\/\/html5gameengine.com\/\">literally dozens of high quality libraries out there<\/a>.\nForget about being idiomatic, just interop the shit out of it!<\/p><p>The benefits are obvious: you get a battle tested foundation backed by a large\ncommunity, and actual documentation. Just use it as is, and wrap as you go where\nit makes sense.<\/p>","created_at":"2019-02-22T20:43:41+0000","updated_at":"2019-02-22T20:43:51+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"lambdaisland.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["javascript","audio"],"is_public":false,"id":854,"title":"goldfire\/howler.js","url":"https:\/\/github.com\/goldfire\/howler.js","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/howlerjs.com\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/98b571ff7230cf5d95e1d41d5d8ea0620c20a79d\/68747470733a2f2f73332e616d617a6f6e6177732e636f6d2f686f776c65722e6a732f686f776c65722d6c6f676f2e706e67\" alt=\"howler.js\" title=\"howler.js\" data-canonical-src=\"https:\/\/s3.amazonaws.com\/howler.js\/howler-logo.png\" \/><\/a><\/p>\n<p><a href=\"https:\/\/howlerjs.com\" rel=\"nofollow\">howler.js<\/a> is an audio library for the modern web. It defaults to <a href=\"http:\/\/webaudio.github.io\/web-audio-api\/\" rel=\"nofollow\">Web Audio API<\/a> and falls back to <a href=\"https:\/\/html.spec.whatwg.org\/multipage\/embedded-content.html#the-audio-element\" rel=\"nofollow\">HTML5 Audio<\/a>. This makes working with audio in JavaScript easy and reliable across all platforms.<\/p>\n<p>Additional information, live demos and a user showcase are available at <a href=\"https:\/\/howlerjs.com\" rel=\"nofollow\">howlerjs.com<\/a>.<\/p>\n<h3><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h3>\n<ul><li>Single API for all audio needs<\/li>\n<li>Defaults to Web Audio API and falls back to HTML5 Audio<\/li>\n<li>Handles edge cases and bugs across environments<\/li>\n<li>Supports all codecs for full cross-browser support<\/li>\n<li>Automatic caching for improved performance<\/li>\n<li>Control sounds individually, in groups or globally<\/li>\n<li>Playback of multiple sounds at once<\/li>\n<li>Easy sound sprite definition and playback<\/li>\n<li>Full control for fading, rate, seek, volume, etc.<\/li>\n<li>Easily add 3D spatial sound or stereo panning<\/li>\n<li>Modular - use what you want and easy to extend<\/li>\n<li>No outside dependencies, just pure JavaScript<\/li>\n<li>As light as 7kb gzipped<\/li>\n<\/ul><h3><a id=\"user-content-browser-compatibility\" class=\"anchor\" aria-hidden=\"true\" href=\"#browser-compatibility\"><\/a>Browser Compatibility<\/h3>\n<p>Tested in the following browsers\/versions:<\/p>\n<ul><li>Google Chrome 7.0+<\/li>\n<li>Internet Explorer 9.0+<\/li>\n<li>Firefox 4.0+<\/li>\n<li>Safari 5.1.4+<\/li>\n<li>Mobile Safari 6.0+ (after user input)<\/li>\n<li>Opera 12.0+<\/li>\n<li>Microsoft Edge<\/li>\n<\/ul><h3><a id=\"user-content-live-demos\" class=\"anchor\" aria-hidden=\"true\" href=\"#live-demos\"><\/a>Live Demos<\/h3>\n<ul><li><a href=\"https:\/\/howlerjs.com\/#player\" rel=\"nofollow\">Audio Player<\/a><\/li>\n<li><a href=\"https:\/\/howlerjs.com\/#radio\" rel=\"nofollow\">Radio<\/a><\/li>\n<li><a href=\"https:\/\/howlerjs.com\/#spatial\" rel=\"nofollow\">Spatial Audio<\/a><\/li>\n<li><a href=\"https:\/\/howlerjs.com\/#sprite\" rel=\"nofollow\">Audio Sprites<\/a><\/li>\n<\/ul><h3><a id=\"user-content-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#contents\"><\/a>Contents<\/h3>\n<ul><li><a href=\"#quick-start\">Quick Start<\/a><\/li>\n<li><a href=\"#examples\">Examples<\/a><\/li>\n<li><a href=\"#core\">Core<\/a>\n<ul><li><a href=\"#options\">Options<\/a><\/li>\n<li><a href=\"#methods\">Methods<\/a><\/li>\n<li><a href=\"#global-options\">Global Options<\/a><\/li>\n<li><a href=\"#global-methods\">Global Methods<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#plugin-spatial\">Plugin: Spatial<\/a>\n<ul><li><a href=\"#options-1\">Options<\/a><\/li>\n<li><a href=\"#methods-1\">Methods<\/a><\/li>\n<li><a href=\"#global-methods-1\">Global Methods<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#mobilechrome-playback\">Mobile Playback<\/a><\/li>\n<li><a href=\"#dolby-audio-playback\">Dolby Audio Playback<\/a><\/li>\n<li><a href=\"#facebook-instant-games\">Facebook Instant Games<\/a><\/li>\n<li><a href=\"#format-recommendations\">Format Recommendations<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h3><a id=\"user-content-quick-start\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-start\"><\/a>Quick Start<\/h3>\n<p>Several options to get up and running:<\/p>\n<ul><li>Clone the repo: <code>git clone https:\/\/github.com\/goldfire\/howler.js.git<\/code><\/li>\n<li>Install with <a href=\"https:\/\/www.npmjs.com\/package\/howler\" rel=\"nofollow\">npm<\/a>: <code>npm install howler<\/code><\/li>\n<li>Install with <a href=\"https:\/\/yarnpkg.com\/en\/package\/howler\" rel=\"nofollow\">Yarn<\/a>: <code>yarn add howler<\/code><\/li>\n<li>Install with <a href=\"http:\/\/bower.io\/\" rel=\"nofollow\">Bower<\/a>: <code>bower install howler<\/code><\/li>\n<li>Hosted CDN: <a href=\"https:\/\/cdnjs.com\/libraries\/howler\" rel=\"nofollow\"><code>cdnjs<\/code><\/a> <a href=\"https:\/\/www.jsdelivr.com\/projects\/howler.js\" rel=\"nofollow\"><code>jsDelivr<\/code><\/a><\/li>\n<\/ul><p>In the browser:<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;script src=\"\/path\/to\/howler.js\"&gt;&lt;\/script&gt;\n&lt;script&gt;\n    var sound = new Howl({\n      src: ['sound.webm', 'sound.mp3']\n    });\n&lt;\/script&gt;<\/pre><\/div>\n<p>As a dependency:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>import {Howl, Howler} from 'howler';<\/pre><\/div>\n<div class=\"highlight highlight-source-js\"><pre>const {Howl, Howler} = require('howler');<\/pre><\/div>\n<h3><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h3>\n<h5><a id=\"user-content-most-basic-play-an-mp3\" class=\"anchor\" aria-hidden=\"true\" href=\"#most-basic-play-an-mp3\"><\/a>Most basic, play an MP3:<\/h5>\n<div class=\"highlight highlight-source-js\"><pre>var sound = new Howl({\n  src: ['sound.mp3']\n});\nsound.play();<\/pre><\/div>\n<h5><a id=\"user-content-more-playback-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-playback-options\"><\/a>More playback options:<\/h5>\n<div class=\"highlight highlight-source-js\"><pre>var sound = new Howl({\n  src: ['sound.webm', 'sound.mp3', 'sound.wav'],\n  autoplay: true,\n  loop: true,\n  volume: 0.5,\n  onend: function() {\n    console.log('Finished!');\n  }\n});<\/pre><\/div>\n<h5><a id=\"user-content-define-and-play-a-sound-sprite\" class=\"anchor\" aria-hidden=\"true\" href=\"#define-and-play-a-sound-sprite\"><\/a>Define and play a sound sprite:<\/h5>\n<div class=\"highlight highlight-source-js\"><pre>var sound = new Howl({\n  src: ['sounds.webm', 'sounds.mp3'],\n  sprite: {\n    blast: [0, 3000],\n    laser: [4000, 1000],\n    winner: [6000, 5000]\n  }\n});\n\/\/ Shoot the laser!\nsound.play('laser');<\/pre><\/div>\n<h5><a id=\"user-content-listen-for-events\" class=\"anchor\" aria-hidden=\"true\" href=\"#listen-for-events\"><\/a>Listen for events:<\/h5>\n<div class=\"highlight highlight-source-js\"><pre>var sound = new Howl({\n  src: ['sound.webm', 'sound.mp3']\n});\n\/\/ Clear listener after first call.\nsound.once('load', function(){\n  sound.play();\n});\n\/\/ Fires when the sound finishes playing.\nsound.on('end', function(){\n  console.log('Finished!');\n});<\/pre><\/div>\n<h5><a id=\"user-content-control-multiple-sounds\" class=\"anchor\" aria-hidden=\"true\" href=\"#control-multiple-sounds\"><\/a>Control multiple sounds:<\/h5>\n<div class=\"highlight highlight-source-js\"><pre>var sound = new Howl({\n  src: ['sound.webm', 'sound.mp3']\n});\n\/\/ Play returns a unique Sound ID that can be passed\n\/\/ into any method on Howl to control that specific sound.\nvar id1 = sound.play();\nvar id2 = sound.play();\n\/\/ Fade out the first sound and speed up the second.\nsound.fade(1, 0, 1000, id1);\nsound.rate(1.5, id2);<\/pre><\/div>\n<h5><a id=\"user-content-es6\" class=\"anchor\" aria-hidden=\"true\" href=\"#es6\"><\/a>ES6:<\/h5>\n<div class=\"highlight highlight-source-js\"><pre>import {Howl, Howler} from 'howler';\n\/\/ Setup the new Howl.\nconst sound = new Howl({\n  src: ['sound.webm', 'sound.mp3']\n});\n\/\/ Play the sound.\nsound.play();\n\/\/ Change global volume.\nHowler.volume(0.5);<\/pre><\/div>\n<p>More in-depth examples (with accompanying live demos) can be found in the <a href=\"https:\/\/github.com\/goldfire\/howler.js\/tree\/master\/examples\">examples directory<\/a>.<\/p>\n<h2><a id=\"user-content-core\" class=\"anchor\" aria-hidden=\"true\" href=\"#core\"><\/a>Core<\/h2>\n<h3><a id=\"user-content-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#options\"><\/a>Options<\/h3>\n<h4><a id=\"user-content-src-arraystring--required\" class=\"anchor\" aria-hidden=\"true\" href=\"#src-arraystring--required\"><\/a>src <code>Array\/String<\/code> <code>[]<\/code> <em><code>required<\/code><\/em><\/h4>\n<p>The sources to the track(s) to be loaded for the sound (URLs or base64 data URIs). These should be in order of preference, howler.js will automatically load the first one that is compatible with the current browser. If your files have no extensions, you will need to explicitly specify the extension using the <code>format<\/code> property.<\/p>\n<h4><a id=\"user-content-volume-number-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#volume-number-10\"><\/a>volume <code>Number<\/code> <code>1.0<\/code><\/h4>\n<p>The volume of the specific track, from <code>0.0<\/code> to <code>1.0<\/code>.<\/p>\n<h4><a id=\"user-content-html5-boolean-false\" class=\"anchor\" aria-hidden=\"true\" href=\"#html5-boolean-false\"><\/a>html5 <code>Boolean<\/code> <code>false<\/code><\/h4>\n<p>Set to <code>true<\/code> to force HTML5 Audio. This should be used for large audio files so that you don't have to wait for the full file to be downloaded and decoded before playing.<\/p>\n<h4><a id=\"user-content-loop-boolean-false\" class=\"anchor\" aria-hidden=\"true\" href=\"#loop-boolean-false\"><\/a>loop <code>Boolean<\/code> <code>false<\/code><\/h4>\n<p>Set to <code>true<\/code> to automatically loop the sound forever.<\/p>\n<h4><a id=\"user-content-preload-boolean-true\" class=\"anchor\" aria-hidden=\"true\" href=\"#preload-boolean-true\"><\/a>preload <code>Boolean<\/code> <code>true<\/code><\/h4>\n<p>Automatically begin downloading the audio file when the <code>Howl<\/code> is defined.<\/p>\n<h4><a id=\"user-content-autoplay-boolean-false\" class=\"anchor\" aria-hidden=\"true\" href=\"#autoplay-boolean-false\"><\/a>autoplay <code>Boolean<\/code> <code>false<\/code><\/h4>\n<p>Set to <code>true<\/code> to automatically start playback when sound is loaded.<\/p>\n<h4><a id=\"user-content-mute-boolean-false\" class=\"anchor\" aria-hidden=\"true\" href=\"#mute-boolean-false\"><\/a>mute <code>Boolean<\/code> <code>false<\/code><\/h4>\n<p>Set to <code>true<\/code> to load the audio muted.<\/p>\n<h4><a id=\"user-content-sprite-object-\" class=\"anchor\" aria-hidden=\"true\" href=\"#sprite-object-\"><\/a>sprite <code>Object<\/code> <code>{}<\/code><\/h4>\n<p>Define a sound sprite for the sound. The offset and duration are defined in milliseconds. A third (optional) parameter is available to set a sprite as looping. An easy way to generate compatible sound sprites is with <a href=\"https:\/\/github.com\/tonistiigi\/audiosprite\">audiosprite<\/a>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>{\n  key: [offset, duration, (loop)]\n}<\/pre><\/div>\n<h4><a id=\"user-content-rate-number-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#rate-number-10\"><\/a>rate <code>Number<\/code> <code>1.0<\/code><\/h4>\n<p>The rate of playback. 0.5 to 4.0, with 1.0 being normal speed.<\/p>\n<h4><a id=\"user-content-pool-number-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#pool-number-5\"><\/a>pool <code>Number<\/code> <code>5<\/code><\/h4>\n<p>The size of the inactive sounds pool. Once sounds are stopped or finish playing, they are marked as ended and ready for cleanup. We keep a pool of these to recycle for improved performance. Generally this doesn't need to be changed. It is important to keep in mind that when a sound is paused, it won't be removed from the pool and will still be considered active so that it can be resumed later.<\/p>\n<h4><a id=\"user-content-format-array-\" class=\"anchor\" aria-hidden=\"true\" href=\"#format-array-\"><\/a>format <code>Array<\/code> <code>[]<\/code><\/h4>\n<p>howler.js automatically detects your file format from the extension, but you may also specify a format in situations where extraction won't work (such as with a SoundCloud stream).<\/p>\n<h4><a id=\"user-content-xhrwithcredentials-boolean-false\" class=\"anchor\" aria-hidden=\"true\" href=\"#xhrwithcredentials-boolean-false\"><\/a>xhrWithCredentials <code>Boolean<\/code> <code>false<\/code><\/h4>\n<p>Whether or not to enable the <code>withCredentials<\/code> flag on XHR requests used to fetch audio files when using Web Audio API (<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/XMLHttpRequest\/withCredentials\" rel=\"nofollow\">see reference<\/a>).<\/p>\n<h4><a id=\"user-content-onload-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onload-function\"><\/a>onload <code>Function<\/code><\/h4>\n<p>Fires when the sound is loaded.<\/p>\n<h4><a id=\"user-content-onloaderror-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onloaderror-function\"><\/a>onloaderror <code>Function<\/code><\/h4>\n<p>Fires when the sound is unable to load. The first parameter is the ID of the sound (if it exists) and the second is the error message\/code.<\/p>\n<h4><a id=\"user-content-onplayerror-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onplayerror-function\"><\/a>onplayerror <code>Function<\/code><\/h4>\n<p>Fires when the sound is unable to play. The first parameter is the ID of the sound and the second is the error message\/code.<\/p>\n<h4><a id=\"user-content-onplay-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onplay-function\"><\/a>onplay <code>Function<\/code><\/h4>\n<p>Fires when the sound begins playing. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onend-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onend-function\"><\/a>onend <code>Function<\/code><\/h4>\n<p>Fires when the sound finishes playing (if it is looping, it'll fire at the end of each loop). The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onpause-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onpause-function\"><\/a>onpause <code>Function<\/code><\/h4>\n<p>Fires when the sound has been paused. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onstop-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onstop-function\"><\/a>onstop <code>Function<\/code><\/h4>\n<p>Fires when the sound has been stopped. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onmute-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onmute-function\"><\/a>onmute <code>Function<\/code><\/h4>\n<p>Fires when the sound has been muted\/unmuted. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onvolume-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onvolume-function\"><\/a>onvolume <code>Function<\/code><\/h4>\n<p>Fires when the sound's volume has changed. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onrate-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onrate-function\"><\/a>onrate <code>Function<\/code><\/h4>\n<p>Fires when the sound's playback rate has changed. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onseek-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onseek-function\"><\/a>onseek <code>Function<\/code><\/h4>\n<p>Fires when the sound has been seeked. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onfade-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onfade-function\"><\/a>onfade <code>Function<\/code><\/h4>\n<p>Fires when the current sound finishes fading in\/out. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onunlock-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onunlock-function\"><\/a>onunlock <code>Function<\/code><\/h4>\n<p>Fires when audio has been automatically unlocked through a touch\/click event.<\/p>\n<h3><a id=\"user-content-methods\" class=\"anchor\" aria-hidden=\"true\" href=\"#methods\"><\/a>Methods<\/h3>\n<h4><a id=\"user-content-playspriteid\" class=\"anchor\" aria-hidden=\"true\" href=\"#playspriteid\"><\/a>play([sprite\/id])<\/h4>\n<p>Begins playback of a sound. Returns the sound id to be used with other methods. Only method that can't be chained.<\/p>\n<ul><li><strong>sprite\/id<\/strong>: <code>String\/Number<\/code> <code>optional<\/code> Takes one parameter that can either be a sprite or sound ID. If a sprite is passed, a new sound will play based on the sprite's definition. If a sound ID is passed, the previously played sound will be played (for example, after pausing it). However, if an ID of a sound that has been drained from the pool is passed, nothing will play.<\/li>\n<\/ul><h4><a id=\"user-content-pauseid\" class=\"anchor\" aria-hidden=\"true\" href=\"#pauseid\"><\/a>pause([id])<\/h4>\n<p>Pauses playback of sound or group, saving the <code>seek<\/code> of playback.<\/p>\n<ul><li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all sounds in group are paused.<\/li>\n<\/ul><h4><a id=\"user-content-stopid\" class=\"anchor\" aria-hidden=\"true\" href=\"#stopid\"><\/a>stop([id])<\/h4>\n<p>Stops playback of sound, resetting <code>seek<\/code> to <code>0<\/code>.<\/p>\n<ul><li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all sounds in group are stopped.<\/li>\n<\/ul><h4><a id=\"user-content-mutemuted-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#mutemuted-id\"><\/a>mute([muted], [id])<\/h4>\n<p>Mutes the sound, but doesn't pause the playback.<\/p>\n<ul><li><strong>muted<\/strong>: <code>Boolean<\/code> <code>optional<\/code> True to mute and false to unmute.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all sounds in group are stopped.<\/li>\n<\/ul><h4><a id=\"user-content-volumevolume-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#volumevolume-id\"><\/a>volume([volume], [id])<\/h4>\n<p>Get\/set volume of this sound or the group. This method optionally takes 0, 1 or 2 arguments.<\/p>\n<ul><li><strong>volume<\/strong>: <code>Number<\/code> <code>optional<\/code> Volume from <code>0.0<\/code> to <code>1.0<\/code>.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all sounds in group have volume altered relative to their own volume.<\/li>\n<\/ul><h4><a id=\"user-content-fadefrom-to-duration-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#fadefrom-to-duration-id\"><\/a>fade(from, to, duration, [id])<\/h4>\n<p>Fade a currently playing sound between two volumes. Fires the <code>fade<\/code> event when complete.<\/p>\n<ul><li><strong>from<\/strong>: <code>Number<\/code> Volume to fade from (<code>0.0<\/code> to <code>1.0<\/code>).<\/li>\n<li><strong>to<\/strong>: <code>Number<\/code> Volume to fade to (<code>0.0<\/code> to <code>1.0<\/code>).<\/li>\n<li><strong>duration<\/strong>: <code>Number<\/code> Time in milliseconds to fade.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all sounds in group will fade.<\/li>\n<\/ul><h4><a id=\"user-content-raterate-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#raterate-id\"><\/a>rate([rate], [id])<\/h4>\n<p>Get\/set the rate of playback for a sound. This method optionally takes 0, 1 or 2 arguments.<\/p>\n<ul><li><strong>rate<\/strong>: <code>Number<\/code> <code>optional<\/code> The rate of playback. 0.5 to 4.0, with 1.0 being normal speed.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, playback rate of all sounds in group will change.<\/li>\n<\/ul><h4><a id=\"user-content-seekseek-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#seekseek-id\"><\/a>seek([seek], [id])<\/h4>\n<p>Get\/set the position of playback for a sound. This method optionally takes 0, 1 or 2 arguments.<\/p>\n<ul><li><strong>seek<\/strong>: <code>Number<\/code> <code>optional<\/code> The position to move current playback to (in seconds).<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, the first sound will seek.<\/li>\n<\/ul><h4><a id=\"user-content-looploop-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#looploop-id\"><\/a>loop([loop], [id])<\/h4>\n<p>Get\/set whether to loop the sound or group. This method can optionally take 0, 1 or 2 arguments.<\/p>\n<ul><li><strong>loop<\/strong>: <code>Boolean<\/code> <code>optional<\/code> To loop or not to loop, that is the question.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all sounds in group will have their <code>loop<\/code> property updated.<\/li>\n<\/ul><h4><a id=\"user-content-state\" class=\"anchor\" aria-hidden=\"true\" href=\"#state\"><\/a>state()<\/h4>\n<p>Check the load status of the <code>Howl<\/code>, returns a <code>unloaded<\/code>, <code>loading<\/code> or <code>loaded<\/code>.<\/p>\n<h4><a id=\"user-content-playingid\" class=\"anchor\" aria-hidden=\"true\" href=\"#playingid\"><\/a>playing(id)<\/h4>\n<p>Check if a sound is currently playing or not, returns a <code>Boolean<\/code>. If no sound ID is passed, check if any sound in the <code>Howl<\/code> group is playing.<\/p>\n<ul><li><strong>id<\/strong>: <code>Number<\/code> The sound ID to check.<\/li>\n<\/ul><h4><a id=\"user-content-durationid\" class=\"anchor\" aria-hidden=\"true\" href=\"#durationid\"><\/a>duration([id])<\/h4>\n<p>Get the duration of the audio source. Will return 0 until after the <code>load<\/code> event fires.<\/p>\n<ul><li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID to check. Passing an ID will return the duration of the sprite being played on this instance; otherwise, the full source duration is returned.<\/li>\n<\/ul><h4><a id=\"user-content-onevent-function-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#onevent-function-id\"><\/a>on(event, function, [id])<\/h4>\n<p>Listen for events. Multiple events can be added by calling this multiple times.<\/p>\n<ul><li><strong>event<\/strong>: <code>String<\/code> Name of event to fire\/set (<code>load<\/code>, <code>loaderror<\/code>, <code>playerror<\/code>, <code>play<\/code>, <code>end<\/code>, <code>pause<\/code>, <code>stop<\/code>, <code>mute<\/code>, <code>volume<\/code>, <code>rate<\/code>, <code>seek<\/code>, <code>fade<\/code>, <code>unlock<\/code>).<\/li>\n<li><strong>function<\/strong>: <code>Function<\/code> Define function to fire on event.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> Only listen to events for this sound id.<\/li>\n<\/ul><h4><a id=\"user-content-onceevent-function-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#onceevent-function-id\"><\/a>once(event, function, [id])<\/h4>\n<p>Same as <code>on<\/code>, but it removes itself after the callback is fired.<\/p>\n<ul><li><strong>event<\/strong>: <code>String<\/code> Name of event to fire\/set (<code>load<\/code>, <code>loaderror<\/code>, <code>playerror<\/code>, <code>play<\/code>, <code>end<\/code>, <code>pause<\/code>, <code>stop<\/code>, <code>mute<\/code>, <code>volume<\/code>, <code>rate<\/code>, <code>seek<\/code>, <code>fade<\/code>, <code>unlock<\/code>).<\/li>\n<li><strong>function<\/strong>: <code>Function<\/code> Define function to fire on event.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> Only listen to events for this sound id.<\/li>\n<\/ul><h4><a id=\"user-content-offevent-function-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#offevent-function-id\"><\/a>off(event, [function], [id])<\/h4>\n<p>Remove event listener that you've set. Call without parameters to remove all events.<\/p>\n<ul><li><strong>event<\/strong>: <code>String<\/code> Name of event (<code>load<\/code>, <code>loaderror<\/code>, <code>playerror<\/code>, <code>play<\/code>, <code>end<\/code>, <code>pause<\/code>, <code>stop<\/code>, <code>mute<\/code>, <code>volume<\/code>, <code>rate<\/code>, <code>seek<\/code>, <code>fade<\/code>, <code>unlock<\/code>).<\/li>\n<li><strong>function<\/strong>: <code>Function<\/code> <code>optional<\/code> The listener to remove. Omit this to remove all events of type.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> Only remove events for this sound id.<\/li>\n<\/ul><h4><a id=\"user-content-load\" class=\"anchor\" aria-hidden=\"true\" href=\"#load\"><\/a>load()<\/h4>\n<p>This is called by default, but if you set <code>preload<\/code> to false, you must call <code>load<\/code> before you can play any sounds.<\/p>\n<h4><a id=\"user-content-unload\" class=\"anchor\" aria-hidden=\"true\" href=\"#unload\"><\/a>unload()<\/h4>\n<p>Unload and destroy a Howl object. This will immediately stop all sounds attached to this sound and remove it from the cache.<\/p>\n<h3><a id=\"user-content-global-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#global-options\"><\/a>Global Options<\/h3>\n<h4><a id=\"user-content-usingwebaudio-boolean\" class=\"anchor\" aria-hidden=\"true\" href=\"#usingwebaudio-boolean\"><\/a>usingWebAudio <code>Boolean<\/code><\/h4>\n<p><code>true<\/code> if the Web Audio API is available.<\/p>\n<h4><a id=\"user-content-noaudio-boolean\" class=\"anchor\" aria-hidden=\"true\" href=\"#noaudio-boolean\"><\/a>noAudio <code>Boolean<\/code><\/h4>\n<p><code>true<\/code> if no audio is available.<\/p>\n<h4><a id=\"user-content-autounlock-boolean-true\" class=\"anchor\" aria-hidden=\"true\" href=\"#autounlock-boolean-true\"><\/a>autoUnlock <code>Boolean<\/code> <code>true<\/code><\/h4>\n<p>Automatically attempts to enable audio on mobile (iOS, Android, etc) devices and desktop Chrome\/Safari.<\/p>\n<h4><a id=\"user-content-html5poolsize-number-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#html5poolsize-number-10\"><\/a>html5PoolSize <code>Number<\/code> <code>10<\/code><\/h4>\n<p>Each HTML5 Audio object must be unlocked individually, so we keep a global pool of unlocked nodes to share between all <code>Howl<\/code> instances. This pool gets created on the first user interaction and is set to the size of this property.<\/p>\n<h4><a id=\"user-content-autosuspend-boolean-true\" class=\"anchor\" aria-hidden=\"true\" href=\"#autosuspend-boolean-true\"><\/a>autoSuspend <code>Boolean<\/code> <code>true<\/code><\/h4>\n<p>Automatically suspends the Web Audio AudioContext after 30 seconds of inactivity to decrease processing and energy usage. Automatically resumes upon new playback. Set this property to <code>false<\/code> to disable this behavior.<\/p>\n<h4><a id=\"user-content-ctx-boolean-web-audio-only\" class=\"anchor\" aria-hidden=\"true\" href=\"#ctx-boolean-web-audio-only\"><\/a>ctx <code>Boolean<\/code> <em><code>Web Audio Only<\/code><\/em><\/h4>\n<p>Exposes the <code>AudioContext<\/code> with Web Audio API.<\/p>\n<h4><a id=\"user-content-mastergain-boolean-web-audio-only\" class=\"anchor\" aria-hidden=\"true\" href=\"#mastergain-boolean-web-audio-only\"><\/a>masterGain <code>Boolean<\/code> <em><code>Web Audio Only<\/code><\/em><\/h4>\n<p>Exposes the master <code>GainNode<\/code> with Web Audio API. This can be useful for writing plugins or advanced usage.<\/p>\n<h3><a id=\"user-content-global-methods\" class=\"anchor\" aria-hidden=\"true\" href=\"#global-methods\"><\/a>Global Methods<\/h3>\n<p>The following methods are used to modify all sounds globally, and are called from the <code>Howler<\/code> object.<\/p>\n<h4><a id=\"user-content-mutemuted\" class=\"anchor\" aria-hidden=\"true\" href=\"#mutemuted\"><\/a>mute(muted)<\/h4>\n<p>Mute or unmute all sounds.<\/p>\n<ul><li><strong>muted<\/strong>: <code>Boolean<\/code> True to mute and false to unmute.<\/li>\n<\/ul><h4><a id=\"user-content-volumevolume\" class=\"anchor\" aria-hidden=\"true\" href=\"#volumevolume\"><\/a>volume([volume])<\/h4>\n<p>Get\/set the global volume for all sounds, relative to their own volume.<\/p>\n<ul><li><strong>volume<\/strong>: <code>Number<\/code> <code>optional<\/code> Volume from <code>0.0<\/code> to <code>1.0<\/code>.<\/li>\n<\/ul><h4><a id=\"user-content-codecsext\" class=\"anchor\" aria-hidden=\"true\" href=\"#codecsext\"><\/a>codecs(ext)<\/h4>\n<p>Check supported audio codecs. Returns <code>true<\/code> if the codec is supported in the current browser.<\/p>\n<ul><li><strong>ext<\/strong>: <code>String<\/code> File extension. One of: \"mp3\", \"mpeg\", \"opus\", \"ogg\", \"oga\", \"wav\", \"aac\", \"caf\", m4a\", \"mp4\", \"weba\", \"webm\", \"dolby\", \"flac\".<\/li>\n<\/ul><h4><a id=\"user-content-unload-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#unload-1\"><\/a>unload()<\/h4>\n<p>Unload and destroy all currently loaded Howl objects. This will immediately stop all sounds and remove them from cache.<\/p>\n<h2><a id=\"user-content-plugin-spatial\" class=\"anchor\" aria-hidden=\"true\" href=\"#plugin-spatial\"><\/a>Plugin: Spatial<\/h2>\n<h3><a id=\"user-content-options-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#options-1\"><\/a>Options<\/h3>\n<h4><a id=\"user-content-orientation-array-1-0-0\" class=\"anchor\" aria-hidden=\"true\" href=\"#orientation-array-1-0-0\"><\/a>orientation <code>Array<\/code> <code>[1, 0, 0]<\/code><\/h4>\n<p>Sets the direction the audio source is pointing in the 3D cartesian coordinate space. Depending on how directional the sound is, based on the <code>cone<\/code> attributes, a sound pointing away from the listener can be quiet or silent.<\/p>\n<h4><a id=\"user-content-stereo-number-null\" class=\"anchor\" aria-hidden=\"true\" href=\"#stereo-number-null\"><\/a>stereo <code>Number<\/code> <code>null<\/code><\/h4>\n<p>Sets the stereo panning value of the audio source for this sound or group. This makes it easy to setup left\/right panning with a value of <code>-1.0<\/code> being far left and a value of <code>1.0<\/code> being far right.<\/p>\n<h4><a id=\"user-content-pos-array-null\" class=\"anchor\" aria-hidden=\"true\" href=\"#pos-array-null\"><\/a>pos <code>Array<\/code> <code>null<\/code><\/h4>\n<p>Sets the 3D spatial position of the audio source for this sound or group relative to the global listener.<\/p>\n<h4><a id=\"user-content-pannerattr-object\" class=\"anchor\" aria-hidden=\"true\" href=\"#pannerattr-object\"><\/a>pannerAttr <code>Object<\/code><\/h4>\n<p>Sets the panner node's attributes for a sound or group of sounds. See the <code>pannerAttr<\/code> method for all available options.<\/p>\n<h4><a id=\"user-content-onstereo-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onstereo-function\"><\/a>onstereo <code>Function<\/code><\/h4>\n<p>Fires when the current sound has the stereo panning changed. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onpos-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onpos-function\"><\/a>onpos <code>Function<\/code><\/h4>\n<p>Fires when the current sound has the listener position changed. The first parameter is the ID of the sound.<\/p>\n<h4><a id=\"user-content-onorientation-function\" class=\"anchor\" aria-hidden=\"true\" href=\"#onorientation-function\"><\/a>onorientation <code>Function<\/code><\/h4>\n<p>Fires when the current sound has the direction of the listener changed. The first parameter is the ID of the sound.<\/p>\n<h3><a id=\"user-content-methods-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#methods-1\"><\/a>Methods<\/h3>\n<h4><a id=\"user-content-stereopan-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#stereopan-id\"><\/a>stereo(pan, [id])<\/h4>\n<p>Get\/set the stereo panning of the audio source for this sound or all in the group.<\/p>\n<ul><li><strong>pan<\/strong>: <code>Number<\/code> A value of <code>-1.0<\/code> is all the way left and <code>1.0<\/code> is all the way right.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all in group will be updated.<\/li>\n<\/ul><h4><a id=\"user-content-posx-y-z-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#posx-y-z-id\"><\/a>pos(x, y, z, [id])<\/h4>\n<p>Get\/set the 3D spatial position of the audio source for this sound or group relative to the global listener.<\/p>\n<ul><li><strong>x<\/strong>: <code>Number<\/code> The x-position of the audio source.<\/li>\n<li><strong>y<\/strong>: <code>Number<\/code> The y-position of the audio source.<\/li>\n<li><strong>z<\/strong>: <code>Number<\/code> The z-position of the audio source.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all in group will be updated.<\/li>\n<\/ul><h4><a id=\"user-content-orientationx-y-z-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#orientationx-y-z-id\"><\/a>orientation(x, y, z, [id])<\/h4>\n<p>Get\/set the direction the audio source is pointing in the 3D cartesian coordinate space. Depending on how directional the sound is, based on the <code>cone<\/code> attributes, a sound pointing away from the listener can be quiet or silent.<\/p>\n<ul><li><strong>x<\/strong>: <code>Number<\/code> The x-orientation of the source.<\/li>\n<li><strong>y<\/strong>: <code>Number<\/code> The y-orientation of the source.<\/li>\n<li><strong>z<\/strong>: <code>Number<\/code> The z-orientation of the source.<\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all in group will be updated.<\/li>\n<\/ul><h4><a id=\"user-content-pannerattro-id\" class=\"anchor\" aria-hidden=\"true\" href=\"#pannerattro-id\"><\/a>pannerAttr(o, [id])<\/h4>\n<p>Get\/set the panner node's attributes for a sound or group of sounds.<\/p>\n<ul><li><strong>o<\/strong>: <code>Object<\/code> All values to update.\n<ul><li><strong>coneInnerAngle<\/strong> <code>360<\/code> A parameter for directional audio sources, this is an angle, in degrees, inside of which there will be no volume reduction.<\/li>\n<li><strong>coneOuterAngle<\/strong> <code>360<\/code> A parameter for directional audio sources, this is an angle, in degrees, outside of which the volume will be reduced to a constant value of <code>coneOuterGain<\/code>.<\/li>\n<li><strong>coneOuterGain<\/strong> <code>0<\/code> A parameter for directional audio sources, this is the gain outside of the <code>coneOuterAngle<\/code>. It is a linear value in the range <code>[0, 1]<\/code>.<\/li>\n<li><strong>distanceModel<\/strong> <code>inverse<\/code> Determines algorithm used to reduce volume as audio moves away from listener. Can be <code>linear<\/code>, <code>inverse<\/code> or `exponential. You can find the implementations of each in the <a href=\"https:\/\/webaudio.github.io\/web-audio-api\/#idl-def-DistanceModelType\" rel=\"nofollow\">spec<\/a>.<\/li>\n<li><strong>maxDistance<\/strong> <code>10000<\/code> The maximum distance between source and listener, after which the volume will not be reduced any further.<\/li>\n<li><strong>refDistance<\/strong> <code>1<\/code> A reference distance for reducing volume as source moves further from the listener. This is simply a variable of the distance model and has a different effect depending on which model is used and the scale of your coordinates. Generally, volume will be equal to 1 at this distance.<\/li>\n<li><strong>rolloffFactor<\/strong> <code>1<\/code> How quickly the volume reduces as source moves from listener. This is simply a variable of the distance model and can be in the range of <code>[0, 1]<\/code> with <code>linear<\/code> and <code>[0, \u221e]<\/code> with <code>inverse<\/code> and <code>exponential<\/code>.<\/li>\n<li><strong>panningModel<\/strong> <code>HRTF<\/code> Determines which spatialization algorithm is used to position audio. Can be <code>HRTF<\/code> or <code>equalpower<\/code>.<\/li>\n<\/ul><\/li>\n<li><strong>id<\/strong>: <code>Number<\/code> <code>optional<\/code> The sound ID. If none is passed, all in group will be updated.<\/li>\n<\/ul><h3><a id=\"user-content-global-methods-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#global-methods-1\"><\/a>Global Methods<\/h3>\n<h4><a id=\"user-content-stereopan\" class=\"anchor\" aria-hidden=\"true\" href=\"#stereopan\"><\/a>stereo(pan)<\/h4>\n<p>Helper method to update the stereo panning position of all current <code>Howls<\/code>. Future <code>Howls<\/code> will not use this value unless explicitly set.<\/p>\n<ul><li><strong>pan<\/strong>: <code>Number<\/code> A value of -1.0 is all the way left and 1.0 is all the way right.<\/li>\n<\/ul><h4><a id=\"user-content-posx-y-z\" class=\"anchor\" aria-hidden=\"true\" href=\"#posx-y-z\"><\/a>pos(x, y, z)<\/h4>\n<p>Get\/set the position of the listener in 3D cartesian space. Sounds using 3D position will be relative to the listener's position.<\/p>\n<ul><li><strong>x<\/strong>: <code>Number<\/code> The x-position of the listener.<\/li>\n<li><strong>y<\/strong>: <code>Number<\/code> The y-position of the listener.<\/li>\n<li><strong>z<\/strong>: <code>Number<\/code> The z-position of the listener.<\/li>\n<\/ul><h4><a id=\"user-content-orientationx-y-z-xup-yup-zup\" class=\"anchor\" aria-hidden=\"true\" href=\"#orientationx-y-z-xup-yup-zup\"><\/a>orientation(x, y, z, xUp, yUp, zUp)<\/h4>\n<p>Get\/set the direction the listener is pointing in the 3D cartesian space. A front and up vector must be provided. The front is the direction the face of the listener is pointing, and up is the direction the top of the listener is pointing. Thus, these values are expected to be at right angles from each other.<\/p>\n<ul><li><strong>x<\/strong>: <code>Number<\/code> The x-orientation of listener.<\/li>\n<li><strong>y<\/strong>: <code>Number<\/code> The y-orientation of listener.<\/li>\n<li><strong>z<\/strong>: <code>Number<\/code> The z-orientation of listener.<\/li>\n<li><strong>xUp<\/strong>: <code>Number<\/code> The x-orientation of the top of the listener.<\/li>\n<li><strong>yUp<\/strong>: <code>Number<\/code> The y-orientation of the top of the listener.<\/li>\n<li><strong>zUp<\/strong>: <code>Number<\/code> The z-orientation of the top of the listener.<\/li>\n<\/ul><h3><a id=\"user-content-mobilechrome-playback\" class=\"anchor\" aria-hidden=\"true\" href=\"#mobilechrome-playback\"><\/a>Mobile\/Chrome Playback<\/h3>\n<p>By default, audio on mobile browsers and Chrome\/Safari is locked until a sound is played within a user interaction, and then it plays normally the rest of the page session (<a href=\"https:\/\/developer.apple.com\/library\/safari\/documentation\/audiovideo\/conceptual\/using_html5_audio_video\/PlayingandSynthesizingSounds\/PlayingandSynthesizingSounds.html\" rel=\"nofollow\">Apple documentation<\/a>). The default behavior of howler.js is to attempt to silently unlock audio playback by playing an empty buffer on the first <code>touchend<\/code> event. This behavior can be disabled by calling:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Howler.autoUnlock = false;<\/pre><\/div>\n<p>If you try to play audio automatically on page load, you can listen to a <code>playerror<\/code> event and then wait for the <code>unlock<\/code> event to try and play the audio again:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var sound = new Howl({\n  src: ['sound.webm', 'sound.mp3'],\n  onplayerror: function() {\n    sound.once('unlock', function() {\n      sound.play();\n    });\n  }\n});\nsound.play();<\/pre><\/div>\n<h3><a id=\"user-content-dolby-audio-playback\" class=\"anchor\" aria-hidden=\"true\" href=\"#dolby-audio-playback\"><\/a>Dolby Audio Playback<\/h3>\n<p>Full support for playback of the Dolby Audio format (currently support in Edge and Safari) is included. However, you must specify that the file you are loading is <code>dolby<\/code> since it is in a <code>mp4<\/code> container.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var dolbySound = new Howl({\n  src: ['sound.mp4', 'sound.webm', 'sound.mp3'],\n  format: ['dolby', 'webm', 'mp3']\n});<\/pre><\/div>\n<h3><a id=\"user-content-facebook-instant-games\" class=\"anchor\" aria-hidden=\"true\" href=\"#facebook-instant-games\"><\/a>Facebook Instant Games<\/h3>\n<p>Howler.js provides audio support for the new <a href=\"https:\/\/developers.facebook.com\/docs\/games\/instant-games\/engine-recommendations\" rel=\"nofollow\">Facebook Instant Games<\/a> platform. If you encounter any issues while developing for Instant Games, open an issue with the tag <code>[IG]<\/code>.<\/p>\n<h3><a id=\"user-content-format-recommendations\" class=\"anchor\" aria-hidden=\"true\" href=\"#format-recommendations\"><\/a>Format Recommendations<\/h3>\n<p>Howler.js supports a wide array of audio codecs that have varying browser support (\"mp3\", \"opus\", \"ogg\", \"wav\", \"aac\", \"m4a\", \"mp4\", \"webm\", ...), but if you want full browser coverage you still need to use at least two of them. If your goal is to have the best balance of small filesize and high quality, based on extensive production testing, your best bet is to default to <code>webm<\/code> and fallback to <code>mp3<\/code>. <code>webm<\/code> has nearly full browser coverage with a great combination of compression and quality. You'll need the <code>mp3<\/code> fallback for Internet Explorer.<\/p>\n<p>It is important to remember that howler.js selects the first compatible sound from your array of sources. So if you want <code>webm<\/code> to be used before <code>mp3<\/code>, you need to put the sources in that order.<\/p>\n<p>If you want your <code>webm<\/code> files to be seekable in Firefox, be sure to encode them with the cues element. One way to do this is by using the <code>dash<\/code> flag in <a href=\"https:\/\/www.ffmpeg.org\/\" rel=\"nofollow\">ffmpeg<\/a>:<\/p>\n<pre>ffmpeg -i sound1.wav -dash 1 sound1.webm\n<\/pre>\n<h3><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h3>\n<p>Copyright (c) 2013-2018 <a href=\"https:\/\/twitter.com\/GoldFireStudios\" rel=\"nofollow\">James Simpson<\/a> and <a href=\"http:\/\/goldfirestudios.com\" rel=\"nofollow\">GoldFire Studios, Inc.<\/a><\/p>\n<p>Released under the <a href=\"https:\/\/github.com\/goldfire\/howler.js\/blob\/master\/LICENSE.md\">MIT License<\/a>.<\/p>\n<\/article>","created_at":"2019-02-22T20:10:00+0000","updated_at":"2019-02-22T20:10:16+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":17,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/882005?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":853,"title":"Plain text","url":"http:\/\/tank.hyperfiddle.com\/:dustingetz!timothy-baldridge-datomic-datalog\/","content":"<pre>&lt;!DOCTYPE html&gt;\nHyperfiddle<link rel=\"icon\" type=\"image\/png\" href=\"denied:data:image\/png;base64,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\" \/><link rel=\"stylesheet\" href=\"\/static\/1adc76b75928c03a44b5cf4c7a8e55dc177bcd21\/styles.css\" \/>1adc76b75928c03a44b5cf4c7a8e55dc177bcd21Hanging out with Timothy Baldridge, talking about Datomic, Datalog, GraphQL, APIs, etc (2018)\n<em>In this transcript recorded Jan 2019, Timothy Baldridge and Dustin Getz talk about Clojure and databases. Timothy Baldridge is a software engineer and databases specialist at Cisco, and before that he was a consultant at Cognitect creating web applications for clients using Clojure\/Script Datomic. Dustin Getz is the founder of Hyperfiddle, a fast way to make custom web apps using Datomic.<\/em>\n\nTable of Contents\n\n<iframe src=\"https:\/\/www.youtube.com\/embed\/PJkV-weX3XY?origin=http:\/\/hyperfiddle.net&amp;amp;iv_load_policy=3&amp;amp;modestbranding=1&amp;amp;playsinline=1&amp;amp;showinfo=0&amp;amp;rel=0&amp;amp;enablejsapi=1\"><\/iframe>\nWhat Datomic is great at, and what it isn&#x27;t\nTIMOTHY: A lot of the work I&#x27;ve been doing recently, both at Cognitect as well as now here at Cisco is more around logic engines and query engines and that sort of thing. We have a lot of data stored in different places and a lot of data stores have a lot of different properties. You do a lot of work at Datomic, you&#x27;re well familiar with how Datomic has some things it&#x27;s really good at and kind of some holes in other areas. Maybe stuff like... What&#x27;s the word I&#x27;m looking for here? Materialized views and if you have a query you wanna run... You wanna get triggers set whenever some data matches a query in the database, that sort of thing.\nTIMOTHY: Those are areas that Datomic maybe isn&#x27;t so good at. I do a lot of work with Elasticsearch now and Elasticsearch has no support for joins, whatsoever. It&#x27;s not relational at all, [chuckle] but it&#x27;s really fast and it can... You can store a ton of data in it. And so a lot of the work I&#x27;m working on is kind of just providing tools for people to be able to query that amount of data efficiently, so they can hand us a query and we&#x27;ll analyze the query, optimize it, and find the most efficient way to talk to Elasticsearch. And that uses a lot of the logic engine stuff I&#x27;ve done some talks on, and things in the past. Kind of our own version of Datalog, but even Datalog has its... Datalog is a set-based thing.\nTIMOTHY: So in Datomic if you say, &quot;Give me these results, but I only want the first three,&quot; all Datalogs are like this, by the way, it&#x27;ll find all the results, throw away all of them but the first three and hand you the first three. And so that&#x27;s not so good when you&#x27;re working with what could be billions of results. You can&#x27;t find them all, you don&#x27;t want that, you want a more lazy approach to find your results. And so a lot of the stuff I&#x27;ve worked on really in the past three years or so, is a lot of that, of just saying, &quot;I know I want the first three results, can we get those efficiently?&quot; And then if I can come back later and say, &quot;I want the next set and the next set.&quot; And doing that across data stores that maybe don&#x27;t have any concept of joins, which is fairly common: Cassandra, Elasticsearch, Riak; although I don&#x27;t know if that&#x27;s still a thing or not. Those are all data stores that... Databases that have no concept of joins or relations or, that sort of thing. So, that kind of overlaps with a lot of the interest I have in programming languages, and that sort of thing, and that these are kind of optimizing logic engines that have a programming language interface, the user is gonna say, &quot;Find this where this is true&quot; and you gotta kind of interpret that. And so there&#x27;s a lot of overlap between those two.\nWhat <em>can&#x27;t<\/em> you model with Datalog?\nDUSTIN: That&#x27;s really interesting. With Hyperfiddle, one of the things I think about a lot is... Datalog can express a lot by itself, especially simpler CRUD apps. But what type of things <em>can&#x27;t<\/em> it express? And what types of things that we might not be thinking about could you express with say some other additional query language, that we can&#x27;t express today?\nTIMOTHY: Yeah. So as long as your data set&#x27;s small enough, Datalog is gonna be way faster than a Prolog approach or miniKanren or core.logic or something based on those things. It&#x27;s kind of a lazy... And all at once set-based approach is faster if you want all the results. But the one thing that I would love to see people investigate more, and there&#x27;s been some very small attempts in this, is that there&#x27;s a whole &#x27;nother set of engines, logic engines, if you wanna call them that, called RETE networks or rules engines... Where you build the query once and it sits there and the query doesn&#x27;t change but as you put new data into the engine, it&#x27;ll trigger new outputs so... So this is the classic like... If you ever say, &quot;I want to know when this query would return an additional result.&quot;\nDUSTIN: Have you looked into what Niko presented at the Conj?\nTIMOTHY: I don&#x27;t know that I have. That was the one, the talk you did earlier that you sent me, right? I have not seen that yet, but yeah, it&#x27;s a lot of the same sort of thing of... It kind of runs in the engine. It&#x27;s interesting because <span class=\"highlight\">Datalog is this idea of, our query&#x27;s gonna change and relative to that the data&#x27;s not gonna change that much. Incremental Dataflows is the opposite. The data&#x27;s always changing and your queries aren&#x27;t really gonna change.<\/span> \nDo situated data applications need anything beyond Datomic Datalog + Incremental Datalog?\nDUSTIN: So here&#x27;s the big question then from my perspective at least, is if you&#x27;ve got Datalog on the one hand \u2014 pull based \u2014 and then you&#x27;ve got the incremental streaming-based RETE network or incremental dataflows or whatever, what are the holes in that? If you can assume both of those are good? What else do we need?\nTIMOTHY: I think that&#x27;s... Yeah, that&#x27;s pretty much it, aside from maybe a streaming result, a lazy... Something that allows you to lazily produce results from the system and go back later. Prolog had that back in the day, that you could say... You could tell it, &quot;Get me a result for this query,&quot; and it runs until it finds one result. And then you can say, &quot;Get me the next one, now get me the next one.&quot; \nSorting problem\nTIMOTHY: Unfortunately with that, one of the downsides is any sort of sorting is really hard to do...\nDUSTIN: It&#x27;s impossible, right?\nTIMOTHY: Yeah, it really is. Unless you do a whole lot of tricks, which I&#x27;ve done in some systems where it&#x27;s like the backend. If the backend knows about sorting you can kind of let that flow through the whole system. So if Elasticsearch or whatever, can give me the results in a sorted manner, pre-sorted, then I can kind of flow that through the system and my results coming out the other end are sorted. But yeah, it&#x27;s sorting is... If someone said sorting is no longer a thing in this world that we know of, I would be a very happy person. [laughter]\nPagination\u00a0\u2013 is it even needed?\nDUSTIN: One thing that seems interesting, if you look at bigger systems like say Facebook, like Facebook comments on... The President&#x27;s page has millions of them, they just have the linked list approach of like, &quot;Hey we&#x27;re just gonna traverse one node at a time and the order is inherent than that.&quot; Does that cover all the use cases that people want pagination?\nTIMOTHY: That&#x27;s interesting. I actually haven&#x27;t thought about that. I think that one of the problems is just people expect arbitrary sorting. That is a sorting based on time that each person... If you said, &quot;I want all the comments made about the President sorted by name of the poster&quot;, that&#x27;s a little bit harder to do. You&#x27;d have to then have two lists.\nDUSTIN: But that&#x27;s not an interesting query because there&#x27;s 10 million of them.\nTIMOTHY: Yes and no. In that context, no, Maybe it&#x27;s not interesting but in a different context, depending what the data set is. A lot of the stuff we work with would be things like... I may want to see all the viruses that we&#x27;ve seen recently sorted by a threat value that we&#x27;ve assigned the virus or then sorted by the domain address that they talked to. I don&#x27;t know, I&#x27;m pulling stuff out of my head here. [laughter] But yeah, there&#x27;s cases like that. And the true part of the story that I&#x27;m sure you&#x27;re familiar with is that if you&#x27;re doing UI work, everybody just expects, you click the header at the top and it sorts, [laughter] and you click it again and it inverse sorts. That&#x27;s just the way the world works. [laughter] Yeah.\nDUSTIN: We&#x27;re gonna find out. [laughter] What you said about sorting by threat value is interesting... Oh, because if you have the incremental data flow, it&#x27;s like aggregates. You can materialize it and do the sorting in that and view incrementally. If you keep inserting to the right place, you can actually sort it.\nIncremental rules engines fix queries at compile time\nTIMOTHY: Yes. You can. Now of course the problem there with a rules engine... I have to look at the incremental data flow, but I imagine it&#x27;s kind of the same sort of approach. The problem with that is that if you want to sort by a new thing or you have a new piece of information, you have to change the network and backflow any...\nDUSTIN: You have to know compile time...\nTIMOTHY: Yeah. Yeah. Exactly. And there are systems where you can add a new rule at runtime, but a lot of those also involve then re-ingesting history into the network to repopulate anything they... &#x27;Cause a lot of these systems to save memory, will just throw away any information you&#x27;re not interested in, from compiled query. That&#x27;s actually a problem that I&#x27;ve seen in some systems is... Okay, great. If we built the system, whenever we add a new rule, we would have to shut down the server and restart it or throw away the data flow or just the fact of ingesting 10 months worth of data would just take a long time. Yeah. Nothing&#x27;s ever for free. But I&#x27;ve often wanted that sort of thing of being able to have a engine that kind of, can do both. The same query language could express both pull and push-based, because there&#x27;s almost always a time in a system where we&#x27;ll say... Someone will say, &quot;Hey, this is great. We can query your system. Can I give you a query and you give me a data feed?&quot;\nTIMOTHY: And you&#x27;re just pushed to this end point whenever you get a new match, and that&#x27;s hard to do [laughter] without writing them.\nDUSTIN: It&#x27;s interesting that Datalog makes it possible because to some extent, this is all a moving target just like, <span class=\"highlight\">&quot;What is the world gonna be like in five years and what are the capabilities people are gonna ask for in five years?&quot; And given the operational transparency of the way the Cognitect technology works, I can see a path forward where there&#x27;s incremental data flows running with Datomic parity, where you can run the same query both ways.<\/span>\nAggregates\nTIMOTHY: Yeah, yeah. There&#x27;s nothing stopping that from happening with Datomic&#x27;s Datalog language. You could use the same syntax for both. I worked on that for a while and I thought there was a couple things... Oh, it&#x27;s aggregates. Aggregates and the like are real interesting when you get into a rules engine, but there&#x27;s no real way of expressing those in Datalog, and there&#x27;s really no reason to because you do those in Clojure code, currently. There is a little bit of that, but you could definitely extend Datomic&#x27;s Datalog to do more.\nDUSTIN: Or just re-implement it in different ways, with different operations and characteristics.\nOptimizing\nTIMOTHY: Backends. Yeah. And that was even the question when I was working on that, was, &quot;Do you then target the Datalog or do you provide an abstraction above it to say, &#x27;Hey you&#x27;re gonna write in this top level language that understands aggregates and all this, and we&#x27;ll compile the Datalog or we&#x27;ll compile to a Clara rules or whatever.&#x27;&quot; And I just never got down to writing that. Mostly because these things just take a long time to perfect and build, and I&#x27;m still tweaking what I&#x27;ve written for Cisco and I&#x27;ll probably be tweaking it for years. [laughter] There&#x27;s always something to improve or change. What&#x27;s fascinating to me about it though too, is that there&#x27;s a lot of business logic that you can bake into an engine when you write it on your own. So the classic thing here with Datomic is Datomic does not provide an optimizer for their query language. Because Datomic doesn&#x27;t store statistics about attributes and there&#x27;s just not a whole lot that an optimizer could do. And as Rich has said, &quot;The optimizers work great until they don&#x27;t and then, you&#x27;re sunk.&quot; You know?\nDUSTIN: Right.\nTIMOTHY: But in a business situation where you know the data you have, you can leverage that a lot more. And I&#x27;ve done that in some projects where, you know, &quot;I know, yeah, this could blow up or this could be highly inefficient or we could have gigabytes of data in this attribute, but we won&#x27;t and so, we can pre-calculate these values or this thing doesn&#x27;t change often enough that we care.&quot; Statistics would be a good example of that. So I wrote a system that was for a bank and a part of the calculation was optimized based upon the... What was it? The average size of a portfolio of trades or something like that. And that&#x27;s a value that will change over time. But the assumption was, if the system&#x27;s been running for years, the average size of a portfolio isn&#x27;t gonna change that often between server restarts, so you can catch that at the start of the server, when the server starts up, and you&#x27;re gonna push new code in two weeks anyway and blow it all away.\nDUSTIN: So at what layer do you bake in these optimizations?\nTIMOTHY: We were putting these in... We would have a layer that the user would use. Kinda like, we were using Instaparse to parse whatever code they gave us that would come out into an AST. And then, we&#x27;d run a nanopass compiler over that, so a bunch of passes over the AST to refine it. And part of that, what I was just mentioning, was like a partial evaluator of saying, you know, &quot;We&#x27;re looking up a portfolio by an ID.&quot; So, we could go hit Datomic or we could cache all the IDs in memory of all of our portfolios, &#x27;cause there&#x27;s only X amount of them, and then we can replace that ID lookup, so that just removes one join from your Datalog query, right? And if you look that up, maybe the next thing that you are going to look up in the chain, you could also partial evaluate. So, there were some times when you would run the partial evaluator and come to find out through the entire query, you&#x27;re only ever dealing with one result. And that&#x27;s actually what a partial evaluator would do. It would run a clause and say, &quot;Do we think this clause is only gonna return one result or 10,000?&quot; it would try to look up one result and if it found one result, and that was the only one... It would optimize it away and then we&#x27;re going up through, through, through. So, there were times you&#x27;d run the optimizer and it&#x27;s just like, yeah, the result is a constant, it&#x27;s just three or whatever, you know? [chuckle]\nFacebook&#x27;s GraphQL infrastructure\nDUSTIN: Facebook has this layer where they... This layer in their GraphQL implementation where you define the queries, but since there are known at compile time before you deploy, you can use that as an interface to inject hand-coded implementations of the query or hand optimized [0:19:57], but without exposing the UI developer to caring that there&#x27;s an optimized path.\nTIMOTHY: Oh yeah, interesting. In fact that&#x27;s something... I&#x27;ve been doing a lot of work with GraphQL and that&#x27;s something I want to try to do on a, somewhat of an automated basis. To say, &quot;Okay, you gave me a GraphQL query, I have an AST, I&#x27;ve optimized it, I know what I&#x27;m going to do to the servers for that query.&quot; I would love to be able to make that a stored procedure, which GraphQL kind of, mostly allows for, where I could have variables. So you can plug in your variables and now we can just ignore that optimization pass and run the same query for different variables. As it turns out though, for a lot of that it&#x27;s like Clojure is really fast and in a lot of these systems that optimization pass is 20 milliseconds. And at that point, it&#x27;s like, &quot;What&#x27;s the point?&quot; [laughter] I mean, until you get to the problem of, hey 20 milliseconds isn&#x27;t fast enough, so much more in the system, response cycle is higher. It&#x27;s gonna take longer to JSON and code the data, than run that optimizer pass.\nDUSTIN: But that&#x27;s what&#x27;s cool about doing it after the fact, is you can just forget about it until it&#x27;s a problem and then you can start injecting optimizations by toggling a check box and bypass it in this other way, and that&#x27;s not expensive to do.\nTIMOTHY: Yeah, yeah, exactly, that sort of abstraction. And that is the goal with GraphQL in a lot of these systems as well. It&#x27;s like, &quot;Hey, now the developers can change the back-end without changing the interface the users use.&quot; Yeah.\nGraphQL - Does Clojure need it?\nDUSTIN: What do you think about GraphQL overall? Is this something that you&#x27;ve been asked before?\nTIMOTHY: No, no, I haven&#x27;t. Well, not publicly. It&#x27;s great. For what it&#x27;s trying to do, it&#x27;s great. <span class=\"highlight\">I don&#x27;t think Clojure needs GraphQL, to be honest with you.<\/span>\nDUSTIN: Tell me why.\nTIMOTHY: Because it&#x27;s... I don&#x27;t know what the languages they wrote GraphQL for but... What is that? It&#x27;s not JSON, what is that syntax? No, I think the biggest thing is it&#x27;s not data. The response is data. The response is JSON or preferably EDN or Transit or something, but the query language itself is kind of weird. And I actually, <span class=\"highlight\">there&#x27;s two areas of things that I&#x27;ve hacked around GraphQL because it doesn&#x27;t support it properly. One is OR, there&#x27;s no logical OR in the language. So there&#x27;s no way of saying, &quot;Get me people whose name is Tim or their name is Bill.&quot; You have to either do like an array, but then you can&#x27;t do arbitrary binary logic. So you can&#x27;t say, &quot;His name is Bill OR Tim&quot; OR &quot;They live in Chicago&quot; because there&#x27;s no way to group those OR clauses in the way I just said it.<\/span> And even the way I just said it doesn&#x27;t make sense in English, but you get the point. [chuckle]\nDUSTIN: But you could <em>name<\/em> a collection and then have relational logic on an implementation of it?\nTIMOTHY: Name a collection... Yeah or what you can do, what I&#x27;ve actually ended up doing is, I have a input variable that&#x27;s JSON, and it&#x27;s a AST as JSON. And it&#x27;s gross, and it&#x27;s ugly, but hey, it works. The other thing too, is there&#x27;s no way to express... So if you have a GraphQL [0:23:25] tree, you can provide parameters to the children, so I can filter the parent and I can filter the children. But there&#x27;s no way to filter the parent by the children. So let&#x27;s say you have people, and you have vehicles that they own. You can say, &quot;Find me all people named Tim and list their cars that are made by Ford.&quot; <span class=\"highlight\">But what you can&#x27;t say is, &quot;Find me all people who own a car made by Ford.&quot; You actually would have to invert that in GraphQL and say, &quot;Find me all cars made by Ford, now find me the people that own those,&quot; Which is not what the original query wanted either.<\/span> You want a way of saying like &quot;people.cars.make equals Ford&quot; at the top level, and have that filter the people that are returned.\nDUSTIN: But surely there&#x27;s a way to just use GraphQL like a pull language, but not a filtering language.\nTIMOTHY: Yes, and once again, that&#x27;s also what I&#x27;ve done. I have an attribute just called Query and inside that&#x27;s a string and you can give arbitrary... Well, JSON or you can do Solr, whatever syntax you want. And yeah, that works, it&#x27;s just, it&#x27;s funny because all of that is rectified by just using EDN or JSON as the input. And so, that&#x27;s the one thing I&#x27;m not sure that I agree with, with GraphQL is that I think they should have just had JSON be the input. And I know they want it type-checked and all of this sort of stuff, but you could have type-checked JSON as well. But whatever. [chuckle] I think part of it is GraphQL does look cleaner than a blob of JSON in that the names, your labels, the keys and your maps aren&#x27;t in quotes, that sort of thing. That you can say, Person: Name. But hey, machines write these things. Not humans in production. So what&#x27;s the point?\nDUSTIN: So I&#x27;m trying to understand what you meant when you said Clojure doesn&#x27;t need it.\nTIMOTHY: Yeah, so Clojure doesn&#x27;t need the GraphQL syntax... Like the actual syntax. Clojure&#x27;s interface could literally be EDN in and EDN out or Transit in and Transit out.\nGraphQL and clojure.spec\nTIMOTHY: And you don&#x27;t even necessarily need a... Like spec, spec would work. Whenever we get data specs \u2014 like specs is data, where you could actually download the specs. That&#x27;s something I think at that point, you basically don&#x27;t need GraphQL at all, because what you have then is... You could have an end point that says, &quot;Give me the specs that you&#x27;re aware of,&quot; you download the specs into the client, adjust them, process them, whatever, and now you can type-check your request and then your request is just entity name, person, fields that you&#x27;re requesting, list them out in array of values, what&#x27;s the filter? That sort of thing. It becomes an AST. It&#x27;s a trio...\nDUSTIN: When you say spec, you mean spec and also probably the schema... If you translate this into Datomic which is more EDN-like, you&#x27;ve got a schema and you&#x27;ve got a spec and they are kind of the same thing.\nTIMOTHY: Yeah, Datomic schema is much more data-dependent. You don&#x27;t necessarily say in Datomic, &quot;This needs to be a string of five characters where the first character is a capital T.&quot;\nDUSTIN: But you do know it&#x27;s a string and then you&#x27;ve got specs to EDN...\nTIMOTHY: Yeah, exactly.\nClojure specs as data\nTIMOTHY: <span class=\"highlight\">The perfect future I would love to see is where we have Clojure specs as data, so you can introspect them and generate them and transfer them across the wire.<\/span> And then you could derive your Datomic schema from those specs. So you could just add a new attribute to an entity and that tells you then what needs to go into the database, update the database accordingly. And then yeah, you can check your inputs based on that too. And so it&#x27;s a... It&#x27;s fascinating because a lot of what Rich talked about at the Conj, resonated with me as far as the features he wants to add to spec, and I&#x27;ve felt a lot of those pain points of; well, a person at the web level is different than a person at the database level. But there&#x27;s a two-field difference, but it&#x27;s more or less the same thing, I don&#x27;t wanna keep creating types at every layer, as well as just a lot of the relationship stuff, he kinda hinted on this in the talk, but the example I give to people is currently in spec, it&#x27;s really hard to spec HTTP. If you wanna write a spec that checks the request and response of a REST endpoint, that&#x27;s really hard because the type of a child is dependent upon the error code of the parents. So if I get a 500, then I expect the body to be this. That&#x27;s really kinda hard to do, that sort of polymorphic. Because the parent is an HTTP response, which is [0:28:47] body and this sort of thing. Yeah, and so hopefully that sort of stuff is improving, sounds like it will, but we&#x27;ll see when it comes up.\nDUSTIN: That&#x27;d be really cool. It seems to me like there&#x27;s a lot of... The industry is figuring out a lot of things that weren&#x27;t figured out before, like in the macro sense. Even functional programming in the last 10 years, and how to properly architect apps isn&#x27;t a solved problem. <span class=\"highlight\">It seems like everywhere outside of Clojure there&#x27;s a lot of ad hoc incremental advances stacked on top of each other but Clojure is the only place I know of that has a thought-out and principled derivation of them in a way that fits together cleanly.<\/span>\nTIMOTHY: Yeah. I mean, with some holes right? I think we see that within Cognitect products perhaps. Well, mostly between Clojure and Datomic, but I mean, I don&#x27;t know that we found the end-all be-all for like how functional app should be written, like the web stack should be written, right? The ring still revolves around a lot of, &quot;Well it&#x27;s a function, it takes a single argument.&quot; There&#x27;s a lot of side-effecting and that sort of thing that&#x27;s still happening. But yeah, I think Datomic is really the realization of what would happen if you put a data structure on the disk, a persons data structure on the disk.\nDUSTIN: When you say web stack, you&#x27;re talking about like lower level implementation stuff, right?\nTIMOTHY: Yeah. What I&#x27;m saying is that like as far as... You look at Datomic and you look at Clojure and they fit together very well. It&#x27;s still that succession time model that there&#x27;s a very clean connection between the two, you can tell they&#x27;re both written by the same person, right?\nDUSTIN: Yeah, <span><\/span>[I&#x27;d go so far as to say that they [Clojure and Datomic] are the same thing.]{.highlight}\nTIMOTHY: Yeah. Yeah. Yeah exactly. But but we don&#x27;t see that through the entire Clojure ecosystem... I would go so far as to say, I think we can see a glimpse of what a perfect Clojure app would be, and to be honest with you I&#x27;ve seen that more with... Pedestal&#x27;s got its problems, but for all of that there are some very cool stuff you could do when you&#x27;ve basically... <span class=\"highlight\">What Pedestal does is reify the call stack of the application, what ring has in a bunch of handlers and middleware where you walk up the stack and back down and so, &quot;Hey I work with Liberator and stuff,&quot; where you know, you get an error and your call stack is like 200 frames deep of decision trees in Liberator, in Pedestal that&#x27;s all in data, it&#x27;s all in a giant hash-map that basically says, &#x27;Here&#x27;s the steps we will execute on our way out.&quot; And so it&#x27;s very possible to just take the entire call stack of a Pedestal application and just do... Like just save it somewhere, you can run it on... You can go somewhere else. It&#x27;s kind of like, it&#x27;s the entire web request as data, including the frames, the stack frames. And it has interesting properties for like testing and introspection debugging, that sort of thing.<\/span>\nA future web stack for people who aren&#x27;t professional software engineers?\nDUSTIN: Even that&#x27;s all pretty low level for me. Hyperfiddle is after more of a vision where like, what would it take for my mom or my future kids to be programmers and yet not be professional software engineers. All the web stack stuff is all like a layer underneath what matters, because if you&#x27;ve got specs, you&#x27;ve got datalog \u2014\u00a0both types of datalog, the pull one and the incremental one \u2014 Nowhere in there does it matter that it went over an HTTP request or not?\nTIMOTHY: Yeah, it&#x27;s interesting, it&#x27;s interesting question is when I think that&#x27;s been asked for years...\nDUSTIN: In fact, <span class=\"highlight\">why can&#x27;t you like press a button to switch out your implementation to have it go over a websocket now or to have it whatever else there is in 20 years, same application running on a different like IO implementation strategy?<\/span>\nTIMOTHY: Yeah. It&#x27;s fascinating because on one hand I kind of agree. I remember learning programming back when... I was doing HTTP and... HTML and PHP and that sort of thing, and that was something... I bought a book on PHP, fun story, and I learned C before this and all this and I have friends that are like, &quot;What is up with the [0:33:33].&quot;\nDUSTIN: When were you like 15? You seem like you&#x27;re about my age.\nTIMOTHY: Yeah, yeah, roughly about that time, right. But I knew C, I knew C beforehand. So I pull up this book on learning PHP. I literally spent three hours reading the book and I&#x27;m like, &quot;This is stupid... &quot; [laughter] &quot;I know it all now. Literally I can set this book down and write a webpage.&quot; Because all the names were just, people were like, &quot;Why is it &#x27;strstr&#x27;.&quot; To take a sub-string of a string in PHP, it&#x27;s like what was in C, it makes sense. But yeah, so all that to say, there is a problem in programming today in that... I learned to program in GW Basic and QBasic and that sort of stuff and, there&#x27;s Scratch today, there&#x27;s whatever the, different Scratch that uses the Lego things? I forget. I don&#x27;t know. But a lot of this stuff is just really, really complex and I think it&#x27;s really hard to write something that&#x27;s both powerful and useful for the average person.\nCan the Javascript-GraphQL stack ever get there?\nDUSTIN: I don&#x27;t think that JavaScript, GraphQL, Facebook, whatever, I don&#x27;t think that stack is ever gonna get there. I think there&#x27;s too many holes, as you described, like GraphQL gets like 80% or something, but it&#x27;ll never get all of it.\nTIMOTHY: No, no, no. And I don&#x27;t think that&#x27;s the... Maybe they claim that&#x27;s the point. I don&#x27;t think it&#x27;s that. <span class=\"highlight\">For me, with GraphQL, it&#x27;s more about just like inter-op with other departments, that GraphQL literally runs on anything... Any language out there that can talk to the web, there&#x27;s some GraphQL client for. And so it&#x27;s that idea of just being able to say, &quot;Hey, yeah. You got a GraphQL client? Go download the schema, tell me if you have any problems.&quot;<\/span> But...\nClojure has not yet reached critical mass\nDUSTIN: &#x27;Cause if you try to extrapolate what could happen with Clojure, I don&#x27;t think it&#x27;s reached a critical mass yet where all the benefits that we&#x27;ve been talking about have really reached a point where anyone can achieve these benefits. <span class=\"highlight\">Actually achieving the benefits of Clojure takes a lot of work right now.<\/span>\nTIMOTHY: Yeah, and I think there is... There&#x27;s something I&#x27;ve seen develop in Clojure recently, and that is a lot of the recent developments in Clojure are very enterprise-focused and as far as the Datomic Ions. I have friends... Hey, giving an example, I have a brother who&#x27;s a mechanical engineer, I mean an EE-engineer, kind of rides the line between the two of them. And he knows enough programming to be dangerous, right. Like if he needs to test some circuitry, he can write something that targets an Adreno board or whatever, does all this stuff. He can care less about Clojure, and it&#x27;s one of those things where it&#x27;s like Python does anything and everything he would ever want, and I see that as one of those areas where I don&#x27;t think Clojure is pushing into the accessibility, at least I haven&#x27;t seen it push more into the accessibility world.\nDUSTIN: It may not be time yet.\nTIMOTHY: What was that?\nDUSTIN: There&#x27;s an order in which you have to do things and it may not be time yet.\nTIMOTHY: Yeah, that could be. That could be. I don&#x27;t know what the world&#x27;s gonna be like too, but the area I see Clojure moving into is more the enterprise place where...\nDUSTIN: But you have to go to enterprise first, right? &#x27;Cause if they have something that&#x27;s easy it also first has to be proven <em>complete<\/em> and to prove it complete, it means people are using it in real world circumstances, and validating that it is better.\nTIMOTHY: I don&#x27;t know that I would agree with that &#x27;cause other languages have started at the other end of things. We could say what we want about decisions Ruby makes or Python makes or even PHP. PHP is still the most popular programming language on the planet.\nDUSTIN: But it was complete. Python, PHP, all these were still all imperative languages and they didn&#x27;t take any risks relative to... They were a little better than the thing that came before them, or even a lot better even, Python was a lot better than C.\nTIMOTHY: Right, right. But what they focused on was ease of use and accessibility over all else and that&#x27;s kind of what I&#x27;m saying is that I don&#x27;t know that Clojure... Maybe someday Clojure will focus on those things. What I see it focusing on and has focused on for the last five years is very much more solving hard problems faced by enterprises, running Clojure on ions or even Datomic, all the new Datomic stuff runs mostly in AWS and... That&#x27;s the running problem, is it&#x27;s like, when I was a kid learning programming in C or whatever, it was like, &quot;Where can I find a free C compiler?&quot; We&#x27;re gonna use DJ CPP, which was GPP or G++ with DOS, [chuckle] And it was like... You get all that, that&#x27;s kind of where your thing is. If the barrier of entry is, you need a system that can support a two gig runtime, just to boot whatever you&#x27;re running Clojure on, and it&#x27;s about two gigs. It doesn&#x27;t even matter what... If you&#x27;re using Cursive or Emacs or whatever, you need about two gigs of free RAM to just get Clojure up and running. And then all the cool stuff, Datomic and that sort of stuff is ions, in AWS, a dollar a day is dirt cheap. But that&#x27;s still more than a hobbyist will pay. Now the world is setting up something free and... Even all of that like...\nDUSTIN: What if you didn&#x27;t have to set it up?\nTIMOTHY: Well, that&#x27;s the thing, what if you didn&#x27;t have to set it up? What if it was completely free? What if you didn&#x27;t have to pay? What if it was open source? It&#x27;s all the same sort of thing of just like, you&#x27;re still kind of in that world of it&#x27;s more work to just get it running. It&#x27;s not just, download this one executable, click &quot;Run&quot; and now you can be a programmer. I&#x27;m not saying all this disparagingly towards Clojure. I think there&#x27;s something that someone mentioned to me, although it was just pure economics one, it made a lot of sense to me. Is that optimizing for the ease of use has very little return at the end of the day. Here&#x27;s the example. I should make this very clear, I do not work for Datomic, whatever, I do not speak for them. From my side of things, it&#x27;s like if Datomic was free, what would that gain Cognitect? You&#x27;d get more users maybe, but in the end, that&#x27;s more people to support that probably are just gonna putz around for a while and then not use it. [chuckle] And that&#x27;s the problem with any product. It&#x27;s like if I give it to you for free and you want help with it, now I&#x27;m spending my time for free, supporting something that I&#x27;m getting nothing out of. It&#x27;s not economical. And even providing a free trial or an inexpensive version or &quot;Here&#x27;s a version for educational use&quot; or whatever. It&#x27;s an economics of what...\nDUSTIN: To put the paradox... Tech markets are winner takes all, so it <em>is<\/em> a paradox. You can give it away for free, and then maybe you didn&#x27;t win, and then you wasted your time and your money. So, you need to find a balance. And then maybe flip a switch, where you reach the point where, &quot;Okay, now it&#x27;s free.&quot;\nTIMOTHY: Yeah. And there&#x27;s multiple of ways of looking at that. I have never heard of this officially, but the old adage that I&#x27;ve heard for years, is that Adobe never prosecutes students in college who pirate their software, because they know the first thing they do when they leave, are they&#x27;re gonna go and demand that employers buy it. [chuckle] So yeah, there&#x27;s always that aspect of, if you can get people hooked on the software. But I guess what I&#x27;m saying is, too, there&#x27;s a whole area... The whole area of enterprise is based on this idea that discs are cheap, RAM is even cheaper, I guess, to some extent. Processors are cheap. The thing that I tell new programmers that I come in contact with, it&#x27;s just like, &quot;Yeah, you have to get over the idea that writing more performant code saves the company money \u2013 it probably doesn&#x27;t.&quot; And there&#x27;s times when I have to sit back and say, &quot;Wow, this is inefficient. I can optimize this code.&quot; But why? It&#x27;s cheaper, it&#x27;s actually cheaper... For the amount that the average developer in America is paid, pay them for three weeks or just buy a new server. And that&#x27;s the unfortunate aspect. But that dynamic is different in embedded systems, in game systems, in any small desktop application, anything like that. And I think that&#x27;s why... I did some work with JavaFX and stuff too, and that&#x27;s really the problem with a lot of that is it&#x27;s like, &quot;I&#x27;m gonna go write an applicat... &quot; I wrote an application in JavaFX with Clojure. Right? It takes like, it all boots up. Once the whole thing&#x27;s up and running, the JVM&#x27;s using like a gig of RAM or something. Then I rewrote most of that in C# and it uses 10 megs of RAM. Just because C# is optimized for this desktop environment, reduced RAM usage. It gives memory back to the operating system as soon as it&#x27;s done with it. There&#x27;s just... It was C# versus Clojure, mutable versus immutable and that sort of stuff. I don&#x27;t know, I don&#x27;t think the benefits for a... Are necessarily there for running desktop applications. But yeah. The whole market&#x27;s shifting. Who knows. I don&#x27;t even know what people would write desktop applications for anymore.\nFunctional Programming for domains other than web apps?\nTIMOTHY: That&#x27;s something I actually I mentioned on Twitter a while back, I would love to see somebody write something in a functional language that&#x27;s NOT a web app, nor a backend service, nor an enterprise thing.\nDUSTIN: What you need... Yeah... The way the system works is too imperative. The one stack has so many layers, it almost gives you more opportunities to insert abstractions.\nTIMOTHY: Yeah. Yeah. Exactly. And I would love to see... You know... Hey tax season&#x27;s coming. What would a tax application look like written in a functional language? You know whatever those things are written in, it&#x27;s gotta be horrific. [laughter] And then there&#x27;s also to apps on the desktop. But the most of what I see... Or actually I think what I actually said was, I would love to see a non-CRUD app written in a functional language &#x27;cause there just aren&#x27;t that many of &#x27;em. Most of them are...\nDUSTIN: Well most stack&#x27;s also they&#x27;re about information processing, whereas the desktop apps are more about effects like...\nTIMOTHY: Yeah.\nDUSTIN: What&#x27;s on the screen or something.\nTIMOTHY: And presentation and that sort of thing. Yeah. But as we&#x27;ve seen with React, you can put a functional veneer over that effect.\nDUSTIN: React is pretty slow in the small. You couldn&#x27;t... You don&#x27;t write games in it? Right?\nTIMOTHY: Right. Right. Yeah.\nTIMOTHY: Exactly. So shoutout here to Arcadia. That&#x27;s what they&#x27;re trying to do with a lot of their... Their stuff... The people behind the.NET Clojure implementation targeting Unity.\nDUSTIN: Yeah. I tried to use that two years ago.\nTIMOTHY: Yeah. So I think they finally got funding, so now there&#x27;s full-time people working on it, and I think it&#x27;ll get better in the near future. But... And then if... I always have to mention this, too. There&#x27;s a 2013 QuakeCon. John Carmack got waxing eloquent about Haskell and functional programming. He&#x27;s the guy that wrote Doom, for anyone that&#x27;s not aware, and he started talking about how you could write a functional language, write a Doom-like first-person shooter in a functional language. It&#x27;s about 40 minutes in... I&#x27;m sure there&#x27;s a link to it online. You know, Carmack talks about Haskell at QuakeCon 2013.\nTIMOTHY: But he describes like a garbage collector and all of this, all the mechanics you would need to build the system, &#x27;cause his whole thing was... Yeah... At the end of the day an FPS is tracking like 8 megabytes of mutable data, everything else is assets and textures and all that, and that it&#x27;s really... It&#x27;s very much feasible to write that sort of thing in a functional language. Except there&#x27;s still problems, which is... Things you would never have thought of. The one example he gave is, in a purely functional world, two people come up to a door, and now they both wanna walk into the doorway. So you&#x27;re here, next frame you&#x27;re both trying to walk into the doorway. That&#x27;s a problem that requires communication between the two things, they both can&#x27;t walk through the doorway at once. And that&#x27;s actually fairly hard to do in a functional language.\nDUSTIN: I wonder, it seems to me that if you look at the way software evolves, it can evolve for stupid reasons, like OpenGL is written in an imperative style API because the hardware&#x27;s imperative. It busts the layer above OpenGL. Well, because OpenGL is imperative, it takes a really smart person to figure out how to turn that into like a garbage collector of virtual DOM or something, to commit that [0:47:29]. It takes a person who&#x27;s paid to do the work, in order to make another imperative layer on top of it. It doesn&#x27;t require brilliance. So you get these stack of things that are just like written in terms of the things below it.\nTIMOTHY: Yeah. Yeah. I think that&#x27;s definitely the case.\nDUSTIN: So I would wanna see... I don&#x27;t know. I would wanna see turtles all the way down.\nTIMOTHY: Exactly. Yeah. I would love to see that too.div.hyperfiddle-user { \n  padding-top: 3em; \n  max-width: 60em;\n  padding-bottom: 300px;\n}\n\nh2 { \n  margin-top: 1em; \n}\n\nimg { \n  width: 100%; \n}\n\nfigure .figure-caption {\n  font-size: 1.5em;\n}\n\n.highlight { \n  background-color: #ffe599; \n}\n!function(){var analytics=window.analytics=window.analytics||[];if(!analytics.initialize)if(analytics.invoked)window.console&amp;&amp;console.error&amp;&amp;console.error(\"Segment snippet included twice.\");else{analytics.invoked=!0;analytics.methods=[\"trackSubmit\",\"trackClick\",\"trackLink\",\"trackForm\",\"pageview\",\"identify\",\"reset\",\"group\",\"track\",\"ready\",\"alias\",\"debug\",\"page\",\"once\",\"off\",\"on\"];analytics.factory=function(t){return function(){var e=Array.prototype.slice.call(arguments);e.unshift(t);analytics.push(e);return analytics}};for(var t=0;t[\"^ \",\"~:host-env\",[\"~#HostEnvironment\",[\"^ \",\"~:hostname\",\"tank.hyperfiddle.com\",\"~:service-uri\",\"~rhttp:\/\/tank.hyperfiddle.com\",\"~:active-ide?\",false,\"~:domain-eid\",[\"~:domain\/ident\",\"tank\"],\"~:auth\/root\",\"hyperfiddle.com\",\"~:ide\/root\",\"hyperfiddle.com\",\"~:build\",\"1adc76b75928c03a44b5cf4c7a8e55dc177bcd21\"]],\"~:sentry\",[\"^ \",\"~:dsn\",\"https:\/\/7d76c67be46b481cbe941a0d520477f2@sentry.io\/1301595\",\"~:environment\",\"prod\",\"~:release\",\"1adc76b75928c03a44b5cf4c7a8e55dc177bcd21\"],\"~:hyperfiddle.bootstrap\/init-level\",7][\"^ \",\"~:hyperfiddle.runtime\/auto-transact\",[\"^ \",\"~rdatomic:free:\/\/datomic:4334\/tank\",false,\"~rdatomic:free:\/\/datomic:4334\/~dustin.getz@hyperfiddle.net+beta\",false,\"~rdatomic:free:\/\/datomic:4334\/clojurians-log\",false,\"~rdatomic:free:\/\/datomic:4334\/mbrainz\",false],\"~:pressed-keys\",[\"~#set\",[]],\"~:hyperfiddle.runtime\/domain\",[\"^ \",\"~:db\/id\",17592186046634,\"~:domain\/home-route\",\"[:hfnet.tank\/index]\",\"~:domain\/databases\",[[\"^ \",\"~:domain.database\/name\",\"$\",\"~:domain.database\/record\",[\"^ \",\"~:database\/uri\",\"~rdatomic:free:\/\/datomic:4334\/tank\",\"~:database.custom-security\/client\",\"hyperfiddle.security.client\/entity-ownership\",\"~:database\/write-security\",[\"^ \",\"~:db\/ident\",\"~:hyperfiddle.security\/custom\"],\"~:hyperfiddle\/owners\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]],[\"^ \",\"^;\",\"$beta\",\"^&lt;\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/~dustin.getz@hyperfiddle.net+beta\",\"^?\",[\"^ \",\"^@\",\"~:hyperfiddle.security\/owner-only\"],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]],[\"^ \",\"^;\",\"$clojurians\",\"^&lt;\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/clojurians-log\",\"^?\",[\"^ \",\"^@\",\"^C\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]],[\"^ \",\"^;\",\"$mbrainz\",\"^&lt;\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/mbrainz\",\"^?\",[\"^ \",\"^@\",\"^C\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]]],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"],\"~:domain\/fiddle-database\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/tank\",\"^&gt;\",\"hyperfiddle.security.client\/entity-ownership\",\"^?\",[\"^ \",\"^@\",\"^A\"],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:domain\/code\",\"(defn youtube [val ctx props]\\n  [:div.plyr__video-embed\\n   [:iframe\\n    ; val is like: https:\/\/www.youtube.com\/embed\/eQ_5KRndbyU\\n    {:src (str val \\\"?origin=http:\/\/hyperfiddle.net&amp;iv_load_policy=3&amp;modestbranding=1&amp;playsinline=1&amp;showinfo=0&amp;rel=0&amp;enablejsapi=1\\\")\\n     :allowfullscreen true\\n     :allowtransparency true\\n     :allow \\\"autoplay\\\"}]])\\n\\n(defn vimeo [val ctx props]\\n  [:div.plyr__video-embed\\n   [:iframe\\n    ; val is like: https:\/\/www.youtube.com\/embed\/eQ_5KRndbyU\\n    {:src (str val \\\"https:\/\/player.vimeo.com\/video\/76979871?loop=false&amp;byline=false&amp;portrait=false&amp;title=false&amp;speed=true&amp;transparent=0&amp;gesture=media\\\")\\n     :allowfullscreen true\\n     :allowtransparency true\\n     :allow \\\"autoplay\\\"}]])\\n\\n(defn hash-color [v &amp; [l]]\\n  (let [seed 0.3100632204946232\\n        golden-ratio 0.618033988749895]\\n    (cuerdas.core\/format\\n     \\\"hsl(%s, %s%, %s%)\\\"\\n     (* 360 (mod (+ seed (* (hash v) golden-ratio)) 1))\\n     55  #_\\\"Too bright hurts the eyes\\\"\\n     (or l 70) #_\\\"Medium gray (50) can be read on white and black backgrounds\\\")))\\n\\n(defn scale-opacity [v v-max]\\n  (\/ (js\/Math.sqrt v) (js\/Math.sqrt v-max)))\",\"~:domain\/ident\",\"tank\"],\"~:staging\/selected-uri\",null,\"~:hyperfiddle.runtime\/partitions\",[\"^ \",\"~_\",[\"^ \",\"~:tempid-lookups\",[\"^ \",\"~rdatomic:free:\/\/datomic:4334\/domains\",[\"~#right\",null],\"^1\",[\"^K\",null]],\"~:stage\",[\"^ \"],\"~:ptm\",[\"~#cmap\",[[\"~#EReq\",[[\"^F\",\"hyperfiddle\"],[\"~#DbVal\",[\"~rdatomic:free:\/\/datomic:4334\/domains\",null]],[\"^8\",\"^B\",\"~:domain\/aliases\",[\"^ \",\"^:\",[\"^;\",[\"^ \",\"^&lt;\",[\"^=\",\"^&gt;\",[\"^ \",\"^?\",[\"^@\"]],\"^B\"]]],\"^D\",[\"^=\",\"^&gt;\",[\"^ \",\"^?\",[\"^@\"]],\"^B\"]],\"~:domain\/disable-javascript\",\"~:domain\/environment\",\"^F\",\"~:domain\/router\",\"^9\",\"^E\",\"~:domain\/css\"]]],[\"^ \",\"^8\",17592186045517,\"^9\",\"[:hyperfiddle.ide\/home]\",\"^:\",[[\"^ \",\"^;\",\"$domains\",\"^&lt;\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/domains\",\"^&gt;\",\"hyperfiddle.security.domains\/client\",\"^?\",[\"^ \",\"^@\",\"^A\"],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]],[\"^ \",\"^;\",\"$users\",\"^&lt;\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/hyperfiddle-users\",\"^?\",[\"^ \",\"^@\",\"~:hyperfiddle.security\/allow-anonymous\"],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]],[\"^ \",\"^;\",\"$\",\"^&lt;\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/root\",\"^?\",[\"^ \",\"^@\",\"^C\"],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]]]],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\",\"~uca192cc8-4ccb-48c6-853f-6fc7dcdd1810\"],\"^D\",[\"^ \",\"^=\",\"~rdatomic:free:\/\/datomic:4334\/root\",\"^?\",[\"^ \",\"^@\",\"^C\"],\"^B\",[\"~u87108fa3-e7d5-4ed5-a87a-81a6eb6e1aae\",\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"^U\",\"\/* Not th \u2013 that hits fiddle shortcuts *\/\\ndiv.hyperfiddle.ui div.hyperfiddle.field.-fiddle-pull,\\ndiv.hyperfiddle.ui div.hyperfiddle.field.-fiddle-query,\\ndiv.hyperfiddle.ui div.hyperfiddle.field.-fiddle-cljs-ns,\\ndiv.hyperfiddle.ui div.hyperfiddle.field.-fiddle-renderer,\\ndiv.hyperfiddle.ui div.hyperfiddle.field.-fiddle-css,\\ndiv.hyperfiddle.ui div.hyperfiddle.field.-fiddle-markdown { display: block !important; }\",\"^F\",\"hyperfiddle\",\"^S\",\"{:auth0 {\\\"hyperfiddle.net\\\" {:domain \\\"https:\/\/hypercrud.auth0.com\\\"\\n                            :client-id \\\"HKjJJXteIwACAPO2y7AFtJCkuQj1y2yt\\\"}\\n         \\\"hfmaster.xyz\\\" {:domain \\\"https:\/\/hypercrud-master.auth0.com\\\"\\n                         :client-id \\\"IGkBUHFdxiO1uvc52B5InIvNFoGgY229\\\"}\\n         \\\"hfmaster.net\\\" {:domain \\\"https:\/\/hypercrud-master.auth0.com\\\"\\n                         :client-id \\\"IGkBUHFdxiO1uvc52B5InIvNFoGgY229\\\"}\\n         \\\"hffeature.xyz\\\" {:domain \\\"https:\/\/hypercrud-master.auth0.com\\\"\\n                          :client-id \\\"IGkBUHFdxiO1uvc52B5InIvNFoGgY229\\\"}\\n         \\\"hfdemo.xyz\\\" {:domain \\\"https:\/\/hypercrud-master.auth0.com\\\"\\n                       :client-id \\\"IGkBUHFdxiO1uvc52B5InIvNFoGgY229\\\"}\\n         \\\"hyperfiddle.site\\\" {:domain \\\"https:\/\/hypercrud-dev.auth0.com\\\"\\n                             :client-id \\\"FV2T2B6QDSdakM0Hg57VygzdvZctn7oj\\\"}\\n         }}\"],[\"^O\",[[\"^F\",\"tank\"],[\"^P\",[\"~rdatomic:free:\/\/datomic:4334\/domains\",null]],[\"^8\",\"^E\",\"^U\"]]],[\"^ \",\"^8\",17592186046634,\"^E\",\"(defn youtube [val ctx props]\\n  [:div.plyr__video-embed\\n   [:iframe\\n    ; val is like: https:\/\/www.youtube.com\/embed\/eQ_5KRndbyU\\n    {:src (str val \\\"?origin=http:\/\/hyperfiddle.net&amp;iv_load_policy=3&amp;modestbranding=1&amp;playsinline=1&amp;showinfo=0&amp;rel=0&amp;enablejsapi=1\\\")\\n     :allowfullscreen true\\n     :allowtransparency true\\n     :allow \\\"autoplay\\\"}]])\\n\\n(defn vimeo [val ctx props]\\n  [:div.plyr__video-embed\\n   [:iframe\\n    ; val is like: https:\/\/www.youtube.com\/embed\/eQ_5KRndbyU\\n    {:src (str val \\\"https:\/\/player.vimeo.com\/video\/76979871?loop=false&amp;byline=false&amp;portrait=false&amp;title=false&amp;speed=true&amp;transparent=0&amp;gesture=media\\\")\\n     :allowfullscreen true\\n     :allowtransparency true\\n     :allow \\\"autoplay\\\"}]])\\n\\n(defn hash-color [v &amp; [l]]\\n  (let [seed 0.3100632204946232\\n        golden-ratio 0.618033988749895]\\n    (cuerdas.core\/format\\n     \\\"hsl(%s, %s%, %s%)\\\"\\n     (* 360 (mod (+ seed (* (hash v) golden-ratio)) 1))\\n     55  #_\\\"Too bright hurts the eyes\\\"\\n     (or l 70) #_\\\"Medium gray (50) can be read on white and black backgrounds\\\")))\\n\\n(defn scale-opacity [v v-max]\\n  (\/ (js\/Math.sqrt v) (js\/Math.sqrt v-max)))\"],[\"^O\",[[\"~:fiddle\/ident\",\"~:dustingetz\/timothy-baldridge-datomic-datalog\"],[\"^P\",[\"~rdatomic:free:\/\/datomic:4334\/tank\",null]],[\"^8\",\"~:fiddle\/css\",\"^W\",\"~:fiddle\/uuid\",[\"^ \",\"~:fiddle\/links\",[\"^8\",\"~:link\/class\",[\"^ \",\"~:link\/fiddle\",[\"^8\",\"^W\",\"~:fiddle\/query\",\"~:fiddle\/type\"]],\"~:link\/formula\",\"~:link\/ident\",\"~:link\/path\",\"~:link\/tx-fn\"]],\"~:fiddle\/markdown\",\"~:fiddle\/pull\",\"~:fiddle\/pull-database\",\"^12\",\"~:fiddle\/cljs-ns\",\"~:fiddle\/renderer\",\"^13\",\"~:fiddle\/hydrate-result-as-fiddle\",\"~$*\"]]],[\"^ \",\"^8\",17592186047689,\"^1;\",\"(defn youtube [val ctx props]\\n  [:div.plyr__video-embed\\n   [:iframe\\n    ; val is like: https:\/\/www.youtube.com\/embed\/eQ_5KRndbyU\\n    {:src (str val \\\"?origin=http:\/\/hyperfiddle.net&amp;iv_load_policy=3&amp;modestbranding=1&amp;playsinline=1&amp;showinfo=0&amp;rel=0&amp;enablejsapi=1\\\")\\n     :allowfullscreen true\\n     :allowtransparency true\\n     :allow \\\"autoplay\\\"}]])\\n\\n(defn vimeo [val ctx props]\\n  [:div.plyr__video-embed\\n   [:iframe\\n    ; val is like: https:\/\/www.youtube.com\/embed\/eQ_5KRndbyU\\n    {:src (str val \\\"https:\/\/player.vimeo.com\/video\/76979871?loop=false&amp;byline=false&amp;portrait=false&amp;title=false&amp;speed=true&amp;transparent=0&amp;gesture=media\\\")\\n     :allowfullscreen true\\n     :allowtransparency true\\n     :allow \\\"autoplay\\\"}]])\",\"^Y\",\"div.hyperfiddle-user { \\n  padding-top: 3em; \\n  max-width: 60em;\\n  padding-bottom: 300px;\\n}\\n\\nh2 { \\n  margin-top: 1em; \\n}\\n\\nimg { \\n  width: 100%; \\n}\\n\\nfigure .figure-caption {\\n  font-size: 1.5em;\\n}\\n\\n.highlight { \\n  background-color: #ffe599; \\n}\",\"^W\",\"^X\",\"^18\",\"# Hanging out with Timothy Baldridge, talking about Datomic, Datalog, GraphQL, APIs, etc (2018)\\n\\n*In this transcript recorded Jan 2019, Timothy Baldridge and Dustin Getz talk about Clojure and databases. Timothy Baldridge is a software engineer and databases specialist at Cisco, and before that he was a consultant at Cognitect creating web applications for clients using Clojure\/Script Datomic. Dustin Getz is the founder of Hyperfiddle, a fast way to make custom web apps using Datomic.*\\n\\n***\\n\\n## Table of Contents\\n\\n##\\n\\n![js\/user.youtube](https:\/\/www.youtube.com\/embed\/PJkV-weX3XY)\\n\\n## What Datomic is great at, and what it isn't\\n\\nTIMOTHY: A lot of the work I've been doing recently, both at Cognitect as well as now here at Cisco is more around logic engines and query engines and that sort of thing. We have a lot of data stored in different places and a lot of data stores have a lot of different properties. You do a lot of work at Datomic, you're well familiar with how Datomic has some things it's really good at and kind of some holes in other areas. Maybe stuff like... What's the word I'm looking for here? Materialized views and if you have a query you wanna run... You wanna get triggers set whenever some data matches a query in the database, that sort of thing.\\n\\nTIMOTHY: Those are areas that Datomic maybe isn't so good at. I do a lot of work with Elasticsearch now and Elasticsearch has no support for joins, whatsoever. It's not relational at all, [chuckle] but it's really fast and it can... You can store a ton of data in it. And so a lot of the work I'm working on is kind of just providing tools for people to be able to query that amount of data efficiently, so they can hand us a query and we'll analyze the query, optimize it, and find the most efficient way to talk to Elasticsearch. And that uses a lot of the logic engine stuff I've done some talks on, and things in the past. Kind of our own version of Datalog, but even Datalog has its... Datalog is a set-based thing.\\n\\nTIMOTHY: So in Datomic if you say, \\\"Give me these results, but I only want the first three,\\\" all Datalogs are like this, by the way, it'll find all the results, throw away all of them but the first three and hand you the first three. And so that's not so good when you're working with what could be billions of results. You can't find them all, you don't want that, you want a more lazy approach to find your results. And so a lot of the stuff I've worked on really in the past three years or so, is a lot of that, of just saying, \\\"I know I want the first three results, can we get those efficiently?\\\" And then if I can come back later and say, \\\"I want the next set and the next set.\\\" And doing that across data stores that maybe don't have any concept of joins, which is fairly common: Cassandra, Elasticsearch, Riak; although I don't know if that's still a thing or not. Those are all data stores that... Databases that have no concept of joins or relations or, that sort of thing. So, that kind of overlaps with a lot of the interest I have in programming languages, and that sort of thing, and that these are kind of optimizing logic engines that have a programming language interface, the user is gonna say, \\\"Find this where this is true\\\" and you gotta kind of interpret that. And so there's a lot of overlap between those two.\\n\\n## What *can't* you model with Datalog?\\n\\nDUSTIN: That's really interesting. With Hyperfiddle, one of the things I think about a lot is... Datalog can express a lot by itself, especially simpler CRUD apps. But what type of things *can't* it express? And what types of things that we might not be thinking about could you express with say some other additional query language, that we can't express today?\\n\\nTIMOTHY: Yeah. So as long as your data set's small enough, Datalog is gonna be way faster than a Prolog approach or miniKanren or core.logic or something based on those things. It's kind of a lazy... And all at once set-based approach is faster if you want all the results. But the one thing that I would love to see people investigate more, and there's been some very small attempts in this, is that there's a whole 'nother set of engines, logic engines, if you wanna call them that, called RETE networks or rules engines... Where you build the query once and it sits there and the query doesn't change but as you put new data into the engine, it'll trigger new outputs so... So this is the classic like... If you ever say, \\\"I want to know when this query would return an additional result.\\\"\\n\\nDUSTIN: Have you looked into what Niko presented at the Conj?\\n\\nTIMOTHY: I don't know that I have. That was the one, the talk you did earlier that you sent me, right? I have not seen that yet, but yeah, it's a lot of the same sort of thing of... It kind of runs in the engine. It's interesting because !span[Datalog is this idea of, our query's gonna change and relative to that the data's not gonna change that much. Incremental Dataflows is the opposite. The data's always changing and your queries aren't really gonna change.]{.highlight} \\n\\n## Do situated data applications need anything beyond Datomic Datalog + Incremental Datalog?\\n\\nDUSTIN: So here's the big question then from my perspective at least, is if you've got Datalog on the one hand \u2014 pull based \u2014 and then you've got the incremental streaming-based RETE network or incremental dataflows or whatever, what are the holes in that? If you can assume both of those are good? What else do we need?\\n\\nTIMOTHY: I think that's... Yeah, that's pretty much it, aside from maybe a streaming result, a lazy... Something that allows you to lazily produce results from the system and go back later. Prolog had that back in the day, that you could say... You could tell it, \\\"Get me a result for this query,\\\" and it runs until it finds one result. And then you can say, \\\"Get me the next one, now get me the next one.\\\" \\n\\n## Sorting problem\\n\\nTIMOTHY: Unfortunately with that, one of the downsides is any sort of sorting is really hard to do...\\n\\nDUSTIN: It's impossible, right?\\n\\nTIMOTHY: Yeah, it really is. Unless you do a whole lot of tricks, which I've done in some systems where it's like the backend. If the backend knows about sorting you can kind of let that flow through the whole system. So if Elasticsearch or whatever, can give me the results in a sorted manner, pre-sorted, then I can kind of flow that through the system and my results coming out the other end are sorted. But yeah, it's sorting is... If someone said sorting is no longer a thing in this world that we know of, I would be a very happy person. [laughter]\\n\\n## Pagination\u00a0\u2013 is it even needed?\\n\\nDUSTIN: One thing that seems interesting, if you look at bigger systems like say Facebook, like Facebook comments on... The President's page has millions of them, they just have the linked list approach of like, \\\"Hey we're just gonna traverse one node at a time and the order is inherent than that.\\\" Does that cover all the use cases that people want pagination?\\n\\nTIMOTHY: That's interesting. I actually haven't thought about that. I think that one of the problems is just people expect arbitrary sorting. That is a sorting based on time that each person... If you said, \\\"I want all the comments made about the President sorted by name of the poster\\\", that's a little bit harder to do. You'd have to then have two lists.\\n\\nDUSTIN: But that's not an interesting query because there's 10 million of them.\\n\\nTIMOTHY: Yes and no. In that context, no, Maybe it's not interesting but in a different context, depending what the data set is. A lot of the stuff we work with would be things like... I may want to see all the viruses that we've seen recently sorted by a threat value that we've assigned the virus or then sorted by the domain address that they talked to. I don't know, I'm pulling stuff out of my head here. [laughter] But yeah, there's cases like that. And the true part of the story that I'm sure you're familiar with is that if you're doing UI work, everybody just expects, you click the header at the top and it sorts, [laughter] and you click it again and it inverse sorts. That's just the way the world works. [laughter] Yeah.\\n\\nDUSTIN: We're gonna find out. [laughter] What you said about sorting by threat value is interesting... Oh, because if you have the incremental data flow, it's like aggregates. You can materialize it and do the sorting in that and view incrementally. If you keep inserting to the right place, you can actually sort it.\\n\\n## Incremental rules engines fix queries at compile time\\n\\nTIMOTHY: Yes. You can. Now of course the problem there with a rules engine... I have to look at the incremental data flow, but I imagine it's kind of the same sort of approach. The problem with that is that if you want to sort by a new thing or you have a new piece of information, you have to change the network and backflow any...\\n\\nDUSTIN: You have to know compile time...\\n\\nTIMOTHY: Yeah. Yeah. Exactly. And there are systems where you can add a new rule at runtime, but a lot of those also involve then re-ingesting history into the network to repopulate anything they... 'Cause a lot of these systems to save memory, will just throw away any information you're not interested in, from compiled query. That's actually a problem that I've seen in some systems is... Okay, great. If we built the system, whenever we add a new rule, we would have to shut down the server and restart it or throw away the data flow or just the fact of ingesting 10 months worth of data would just take a long time. Yeah. Nothing's ever for free. But I've often wanted that sort of thing of being able to have a engine that kind of, can do both. The same query language could express both pull and push-based, because there's almost always a time in a system where we'll say... Someone will say, \\\"Hey, this is great. We can query your system. Can I give you a query and you give me a data feed?\\\"\\n\\nTIMOTHY: And you're just pushed to this end point whenever you get a new match, and that's hard to do [laughter] without writing them.\\n\\nDUSTIN: It's interesting that Datalog makes it possible because to some extent, this is all a moving target just like, !span[\\\"What is the world gonna be like in five years and what are the capabilities people are gonna ask for in five years?\\\" And given the operational transparency of the way the Cognitect technology works, I can see a path forward where there's incremental data flows running with Datomic parity, where you can run the same query both ways.]{.highlight}\\n\\n## Aggregates\\n\\nTIMOTHY: Yeah, yeah. There's nothing stopping that from happening with Datomic's Datalog language. You could use the same syntax for both. I worked on that for a while and I thought there was a couple things... Oh, it's aggregates. Aggregates and the like are real interesting when you get into a rules engine, but there's no real way of expressing those in Datalog, and there's really no reason to because you do those in Clojure code, currently. There is a little bit of that, but you could definitely extend Datomic's Datalog to do more.\\n\\nDUSTIN: Or just re-implement it in different ways, with different operations and characteristics.\\n\\n## Optimizing\\n\\nTIMOTHY: Backends. Yeah. And that was even the question when I was working on that, was, \\\"Do you then target the Datalog or do you provide an abstraction above it to say, 'Hey you're gonna write in this top level language that understands aggregates and all this, and we'll compile the Datalog or we'll compile to a Clara rules or whatever.'\\\" And I just never got down to writing that. Mostly because these things just take a long time to perfect and build, and I'm still tweaking what I've written for Cisco and I'll probably be tweaking it for years. [laughter] There's always something to improve or change. What's fascinating to me about it though too, is that there's a lot of business logic that you can bake into an engine when you write it on your own. So the classic thing here with Datomic is Datomic does not provide an optimizer for their query language. Because Datomic doesn't store statistics about attributes and there's just not a whole lot that an optimizer could do. And as Rich has said, \\\"The optimizers work great until they don't and then, you're sunk.\\\" You know?\\n\\nDUSTIN: Right.\\n\\nTIMOTHY: But in a business situation where you know the data you have, you can leverage that a lot more. And I've done that in some projects where, you know, \\\"I know, yeah, this could blow up or this could be highly inefficient or we could have gigabytes of data in this attribute, but we won't and so, we can pre-calculate these values or this thing doesn't change often enough that we care.\\\" Statistics would be a good example of that. So I wrote a system that was for a bank and a part of the calculation was optimized based upon the... What was it? The average size of a portfolio of trades or something like that. And that's a value that will change over time. But the assumption was, if the system's been running for years, the average size of a portfolio isn't gonna change that often between server restarts, so you can catch that at the start of the server, when the server starts up, and you're gonna push new code in two weeks anyway and blow it all away.\\n\\nDUSTIN: So at what layer do you bake in these optimizations?\\n\\nTIMOTHY: We were putting these in... We would have a layer that the user would use. Kinda like, we were using Instaparse to parse whatever code they gave us that would come out into an AST. And then, we'd run a nanopass compiler over that, so a bunch of passes over the AST to refine it. And part of that, what I was just mentioning, was like a partial evaluator of saying, you know, \\\"We're looking up a portfolio by an ID.\\\" So, we could go hit Datomic or we could cache all the IDs in memory of all of our portfolios, 'cause there's only X amount of them, and then we can replace that ID lookup, so that just removes one join from your Datalog query, right? And if you look that up, maybe the next thing that you are going to look up in the chain, you could also partial evaluate. So, there were some times when you would run the partial evaluator and come to find out through the entire query, you're only ever dealing with one result. And that's actually what a partial evaluator would do. It would run a clause and say, \\\"Do we think this clause is only gonna return one result or 10,000?\\\" it would try to look up one result and if it found one result, and that was the only one... It would optimize it away and then we're going up through, through, through. So, there were times you'd run the optimizer and it's just like, yeah, the result is a constant, it's just three or whatever, you know? [chuckle]\\n\\n## Facebook's GraphQL infrastructure\\n\\nDUSTIN: Facebook has this layer where they... This layer in their GraphQL implementation where you define the queries, but since there are known at compile time before you deploy, you can use that as an interface to inject hand-coded implementations of the query or hand optimized [0:19:57], but without exposing the UI developer to caring that there's an optimized path.\\n\\nTIMOTHY: Oh yeah, interesting. In fact that's something... I've been doing a lot of work with GraphQL and that's something I want to try to do on a, somewhat of an automated basis. To say, \\\"Okay, you gave me a GraphQL query, I have an AST, I've optimized it, I know what I'm going to do to the servers for that query.\\\" I would love to be able to make that a stored procedure, which GraphQL kind of, mostly allows for, where I could have variables. So you can plug in your variables and now we can just ignore that optimization pass and run the same query for different variables. As it turns out though, for a lot of that it's like Clojure is really fast and in a lot of these systems that optimization pass is 20 milliseconds. And at that point, it's like, \\\"What's the point?\\\" [laughter] I mean, until you get to the problem of, hey 20 milliseconds isn't fast enough, so much more in the system, response cycle is higher. It's gonna take longer to JSON and code the data, than run that optimizer pass.\\n\\nDUSTIN: But that's what's cool about doing it after the fact, is you can just forget about it until it's a problem and then you can start injecting optimizations by toggling a check box and bypass it in this other way, and that's not expensive to do.\\n\\nTIMOTHY: Yeah, yeah, exactly, that sort of abstraction. And that is the goal with GraphQL in a lot of these systems as well. It's like, \\\"Hey, now the developers can change the back-end without changing the interface the users use.\\\" Yeah.\\n\\n## GraphQL - Does Clojure need it?\\n\\nDUSTIN: What do you think about GraphQL overall? Is this something that you've been asked before?\\n\\nTIMOTHY: No, no, I haven't. Well, not publicly. It's great. For what it's trying to do, it's great. !span[I don't think Clojure needs GraphQL, to be honest with you.]{.highlight}\\n\\nDUSTIN: Tell me why.\\n\\nTIMOTHY: Because it's... I don't know what the languages they wrote GraphQL for but... What is that? It's not JSON, what is that syntax? No, I think the biggest thing is it's not data. The response is data. The response is JSON or preferably EDN or Transit or something, but the query language itself is kind of weird. And I actually, !span[there's two areas of things that I've hacked around GraphQL because it doesn't support it properly. One is OR, there's no logical OR in the language. So there's no way of saying, \\\"Get me people whose name is Tim or their name is Bill.\\\" You have to either do like an array, but then you can't do arbitrary binary logic. So you can't say, \\\"His name is Bill OR Tim\\\" OR \\\"They live in Chicago\\\" because there's no way to group those OR clauses in the way I just said it.]{.highlight} And even the way I just said it doesn't make sense in English, but you get the point. [chuckle]\\n\\nDUSTIN: But you could *name* a collection and then have relational logic on an implementation of it?\\n\\nTIMOTHY: Name a collection... Yeah or what you can do, what I've actually ended up doing is, I have a input variable that's JSON, and it's a AST as JSON. And it's gross, and it's ugly, but hey, it works. The other thing too, is there's no way to express... So if you have a GraphQL [0:23:25] tree, you can provide parameters to the children, so I can filter the parent and I can filter the children. But there's no way to filter the parent by the children. So let's say you have people, and you have vehicles that they own. You can say, \\\"Find me all people named Tim and list their cars that are made by Ford.\\\" !span[But what you can't say is, \\\"Find me all people who own a car made by Ford.\\\" You actually would have to invert that in GraphQL and say, \\\"Find me all cars made by Ford, now find me the people that own those,\\\" Which is not what the original query wanted either.]{.highlight} You want a way of saying like \\\"people.cars.make equals Ford\\\" at the top level, and have that filter the people that are returned.\\n\\nDUSTIN: But surely there's a way to just use GraphQL like a pull language, but not a filtering language.\\n\\nTIMOTHY: Yes, and once again, that's also what I've done. I have an attribute just called Query and inside that's a string and you can give arbitrary... Well, JSON or you can do Solr, whatever syntax you want. And yeah, that works, it's just, it's funny because all of that is rectified by just using EDN or JSON as the input. And so, that's the one thing I'm not sure that I agree with, with GraphQL is that I think they should have just had JSON be the input. And I know they want it type-checked and all of this sort of stuff, but you could have type-checked JSON as well. But whatever. [chuckle] I think part of it is GraphQL does look cleaner than a blob of JSON in that the names, your labels, the keys and your maps aren't in quotes, that sort of thing. That you can say, Person: Name. But hey, machines write these things. Not humans in production. So what's the point?\\n\\nDUSTIN: So I'm trying to understand what you meant when you said Clojure doesn't need it.\\n\\nTIMOTHY: Yeah, so Clojure doesn't need the GraphQL syntax... Like the actual syntax. Clojure's interface could literally be EDN in and EDN out or Transit in and Transit out.\\n\\n## GraphQL and clojure.spec\\n\\nTIMOTHY: And you don't even necessarily need a... Like spec, spec would work. Whenever we get data specs \u2014 like specs is data, where you could actually download the specs. That's something I think at that point, you basically don't need GraphQL at all, because what you have then is... You could have an end point that says, \\\"Give me the specs that you're aware of,\\\" you download the specs into the client, adjust them, process them, whatever, and now you can type-check your request and then your request is just entity name, person, fields that you're requesting, list them out in array of values, what's the filter? That sort of thing. It becomes an AST. It's a trio...\\n\\nDUSTIN: When you say spec, you mean spec and also probably the schema... If you translate this into Datomic which is more EDN-like, you've got a schema and you've got a spec and they are kind of the same thing.\\n\\nTIMOTHY: Yeah, Datomic schema is much more data-dependent. You don't necessarily say in Datomic, \\\"This needs to be a string of five characters where the first character is a capital T.\\\"\\n\\nDUSTIN: But you do know it's a string and then you've got specs to EDN...\\n\\nTIMOTHY: Yeah, exactly.\\n\\n## Clojure specs as data\\n\\nTIMOTHY: !span[The perfect future I would love to see is where we have Clojure specs as data, so you can introspect them and generate them and transfer them across the wire.]{.highlight} And then you could derive your Datomic schema from those specs. So you could just add a new attribute to an entity and that tells you then what needs to go into the database, update the database accordingly. And then yeah, you can check your inputs based on that too. And so it's a... It's fascinating because a lot of what Rich talked about at the Conj, resonated with me as far as the features he wants to add to spec, and I've felt a lot of those pain points of; well, a person at the web level is different than a person at the database level. But there's a two-field difference, but it's more or less the same thing, I don't wanna keep creating types at every layer, as well as just a lot of the relationship stuff, he kinda hinted on this in the talk, but the example I give to people is currently in spec, it's really hard to spec HTTP. If you wanna write a spec that checks the request and response of a REST endpoint, that's really hard because the type of a child is dependent upon the error code of the parents. So if I get a 500, then I expect the body to be this. That's really kinda hard to do, that sort of polymorphic. Because the parent is an HTTP response, which is [0:28:47] body and this sort of thing. Yeah, and so hopefully that sort of stuff is improving, sounds like it will, but we'll see when it comes up.\\n\\nDUSTIN: That'd be really cool. It seems to me like there's a lot of... The industry is figuring out a lot of things that weren't figured out before, like in the macro sense. Even functional programming in the last 10 years, and how to properly architect apps isn't a solved problem. !span[It seems like everywhere outside of Clojure there's a lot of ad hoc incremental advances stacked on top of each other but Clojure is the only place I know of that has a thought-out and principled derivation of them in a way that fits together cleanly.]{.highlight}\\n\\nTIMOTHY: Yeah. I mean, with some holes right? I think we see that within Cognitect products perhaps. Well, mostly between Clojure and Datomic, but I mean, I don't know that we found the end-all be-all for like how functional app should be written, like the web stack should be written, right? The ring still revolves around a lot of, \\\"Well it's a function, it takes a single argument.\\\" There's a lot of side-effecting and that sort of thing that's still happening. But yeah, I think Datomic is really the realization of what would happen if you put a data structure on the disk, a persons data structure on the disk.\\n\\nDUSTIN: When you say web stack, you're talking about like lower level implementation stuff, right?\\n\\nTIMOTHY: Yeah. What I'm saying is that like as far as... You look at Datomic and you look at Clojure and they fit together very well. It's still that succession time model that there's a very clean connection between the two, you can tell they're both written by the same person, right?\\n\\nDUSTIN: Yeah, !span[I'd go so far as to say that they [Clojure and Datomic] are the same thing.]{.highlight}\\n\\nTIMOTHY: Yeah. Yeah. Yeah exactly. But but we don't see that through the entire Clojure ecosystem... I would go so far as to say, I think we can see a glimpse of what a perfect Clojure app would be, and to be honest with you I've seen that more with... Pedestal's got its problems, but for all of that there are some very cool stuff you could do when you've basically... !span[What Pedestal does is reify the call stack of the application, what ring has in a bunch of handlers and middleware where you walk up the stack and back down and so, \\\"Hey I work with Liberator and stuff,\\\" where you know, you get an error and your call stack is like 200 frames deep of decision trees in Liberator, in Pedestal that's all in data, it's all in a giant hash-map that basically says, 'Here's the steps we will execute on our way out.\\\" And so it's very possible to just take the entire call stack of a Pedestal application and just do... Like just save it somewhere, you can run it on... You can go somewhere else. It's kind of like, it's the entire web request as data, including the frames, the stack frames. And it has interesting properties for like testing and introspection debugging, that sort of thing.]{.highlight}\\n\\n## A future web stack for people who aren't professional software engineers?\\n\\nDUSTIN: Even that's all pretty low level for me. Hyperfiddle is after more of a vision where like, what would it take for my mom or my future kids to be programmers and yet not be professional software engineers. All the web stack stuff is all like a layer underneath what matters, because if you've got specs, you've got datalog \u2014\u00a0both types of datalog, the pull one and the incremental one \u2014 Nowhere in there does it matter that it went over an HTTP request or not?\\n\\nTIMOTHY: Yeah, it's interesting, it's interesting question is when I think that's been asked for years...\\n\\nDUSTIN: In fact, !span[why can't you like press a button to switch out your implementation to have it go over a websocket now or to have it whatever else there is in 20 years, same application running on a different like IO implementation strategy?]{.highlight}\\n\\nTIMOTHY: Yeah. It's fascinating because on one hand I kind of agree. I remember learning programming back when... I was doing HTTP and... HTML and PHP and that sort of thing, and that was something... I bought a book on PHP, fun story, and I learned C before this and all this and I have friends that are like, \\\"What is up with the [0:33:33].\\\"\\n\\nDUSTIN: When were you like 15? You seem like you're about my age.\\n\\nTIMOTHY: Yeah, yeah, roughly about that time, right. But I knew C, I knew C beforehand. So I pull up this book on learning PHP. I literally spent three hours reading the book and I'm like, \\\"This is stupid... \\\" [laughter] \\\"I know it all now. Literally I can set this book down and write a webpage.\\\" Because all the names were just, people were like, \\\"Why is it 'strstr'.\\\" To take a sub-string of a string in PHP, it's like what was in C, it makes sense. But yeah, so all that to say, there is a problem in programming today in that... I learned to program in GW Basic and QBasic and that sort of stuff and, there's Scratch today, there's whatever the, different Scratch that uses the Lego things? I forget. I don't know. But a lot of this stuff is just really, really complex and I think it's really hard to write something that's both powerful and useful for the average person.\\n\\n## Can the Javascript-GraphQL stack ever get there?\\n\\nDUSTIN: I don't think that JavaScript, GraphQL, Facebook, whatever, I don't think that stack is ever gonna get there. I think there's too many holes, as you described, like GraphQL gets like 80% or something, but it'll never get all of it.\\n\\nTIMOTHY: No, no, no. And I don't think that's the... Maybe they claim that's the point. I don't think it's that. !span[For me, with GraphQL, it's more about just like inter-op with other departments, that GraphQL literally runs on anything... Any language out there that can talk to the web, there's some GraphQL client for. And so it's that idea of just being able to say, \\\"Hey, yeah. You got a GraphQL client? Go download the schema, tell me if you have any problems.\\\"]{.highlight} But...\\n\\n## Clojure has not yet reached critical mass\\n\\nDUSTIN: 'Cause if you try to extrapolate what could happen with Clojure, I don't think it's reached a critical mass yet where all the benefits that we've been talking about have really reached a point where anyone can achieve these benefits. !span[Actually achieving the benefits of Clojure takes a lot of work right now.]{.highlight}\\n\\nTIMOTHY: Yeah, and I think there is... There's something I've seen develop in Clojure recently, and that is a lot of the recent developments in Clojure are very enterprise-focused and as far as the Datomic Ions. I have friends... Hey, giving an example, I have a brother who's a mechanical engineer, I mean an EE-engineer, kind of rides the line between the two of them. And he knows enough programming to be dangerous, right. Like if he needs to test some circuitry, he can write something that targets an Adreno board or whatever, does all this stuff. He can care less about Clojure, and it's one of those things where it's like Python does anything and everything he would ever want, and I see that as one of those areas where I don't think Clojure is pushing into the accessibility, at least I haven't seen it push more into the accessibility world.\\n\\nDUSTIN: It may not be time yet.\\n\\nTIMOTHY: What was that?\\n\\nDUSTIN: There's an order in which you have to do things and it may not be time yet.\\n\\nTIMOTHY: Yeah, that could be. That could be. I don't know what the world's gonna be like too, but the area I see Clojure moving into is more the enterprise place where...\\n\\nDUSTIN: But you have to go to enterprise first, right? 'Cause if they have something that's easy it also first has to be proven *complete* and to prove it complete, it means people are using it in real world circumstances, and validating that it is better.\\n\\nTIMOTHY: I don't know that I would agree with that 'cause other languages have started at the other end of things. We could say what we want about decisions Ruby makes or Python makes or even PHP. PHP is still the most popular programming language on the planet.\\n\\nDUSTIN: But it was complete. Python, PHP, all these were still all imperative languages and they didn't take any risks relative to... They were a little better than the thing that came before them, or even a lot better even, Python was a lot better than C.\\n\\nTIMOTHY: Right, right. But what they focused on was ease of use and accessibility over all else and that's kind of what I'm saying is that I don't know that Clojure... Maybe someday Clojure will focus on those things. What I see it focusing on and has focused on for the last five years is very much more solving hard problems faced by enterprises, running Clojure on ions or even Datomic, all the new Datomic stuff runs mostly in AWS and... That's the running problem, is it's like, when I was a kid learning programming in C or whatever, it was like, \\\"Where can I find a free C compiler?\\\" We're gonna use DJ CPP, which was GPP or G++ with DOS, [chuckle] And it was like... You get all that, that's kind of where your thing is. If the barrier of entry is, you need a system that can support a two gig runtime, just to boot whatever you're running Clojure on, and it's about two gigs. It doesn't even matter what... If you're using Cursive or Emacs or whatever, you need about two gigs of free RAM to just get Clojure up and running. And then all the cool stuff, Datomic and that sort of stuff is ions, in AWS, a dollar a day is dirt cheap. But that's still more than a hobbyist will pay. Now the world is setting up something free and... Even all of that like...\\n\\nDUSTIN: What if you didn't have to set it up?\\n\\nTIMOTHY: Well, that's the thing, what if you didn't have to set it up? What if it was completely free? What if you didn't have to pay? What if it was open source? It's all the same sort of thing of just like, you're still kind of in that world of it's more work to just get it running. It's not just, download this one executable, click \\\"Run\\\" and now you can be a programmer. I'm not saying all this disparagingly towards Clojure. I think there's something that someone mentioned to me, although it was just pure economics one, it made a lot of sense to me. Is that optimizing for the ease of use has very little return at the end of the day. Here's the example. I should make this very clear, I do not work for Datomic, whatever, I do not speak for them. From my side of things, it's like if Datomic was free, what would that gain Cognitect? You'd get more users maybe, but in the end, that's more people to support that probably are just gonna putz around for a while and then not use it. [chuckle] And that's the problem with any product. It's like if I give it to you for free and you want help with it, now I'm spending my time for free, supporting something that I'm getting nothing out of. It's not economical. And even providing a free trial or an inexpensive version or \\\"Here's a version for educational use\\\" or whatever. It's an economics of what...\\n\\nDUSTIN: To put the paradox... Tech markets are winner takes all, so it *is* a paradox. You can give it away for free, and then maybe you didn't win, and then you wasted your time and your money. So, you need to find a balance. And then maybe flip a switch, where you reach the point where, \\\"Okay, now it's free.\\\"\\n\\nTIMOTHY: Yeah. And there's multiple of ways of looking at that. I have never heard of this officially, but the old adage that I've heard for years, is that Adobe never prosecutes students in college who pirate their software, because they know the first thing they do when they leave, are they're gonna go and demand that employers buy it. [chuckle] So yeah, there's always that aspect of, if you can get people hooked on the software. But I guess what I'm saying is, too, there's a whole area... The whole area of enterprise is based on this idea that discs are cheap, RAM is even cheaper, I guess, to some extent. Processors are cheap. The thing that I tell new programmers that I come in contact with, it's just like, \\\"Yeah, you have to get over the idea that writing more performant code saves the company money \u2013 it probably doesn't.\\\" And there's times when I have to sit back and say, \\\"Wow, this is inefficient. I can optimize this code.\\\" But why? It's cheaper, it's actually cheaper... For the amount that the average developer in America is paid, pay them for three weeks or just buy a new server. And that's the unfortunate aspect. But that dynamic is different in embedded systems, in game systems, in any small desktop application, anything like that. And I think that's why... I did some work with JavaFX and stuff too, and that's really the problem with a lot of that is it's like, \\\"I'm gonna go write an applicat... \\\" I wrote an application in JavaFX with Clojure. Right? It takes like, it all boots up. Once the whole thing's up and running, the JVM's using like a gig of RAM or something. Then I rewrote most of that in C# and it uses 10 megs of RAM. Just because C# is optimized for this desktop environment, reduced RAM usage. It gives memory back to the operating system as soon as it's done with it. There's just... It was C# versus Clojure, mutable versus immutable and that sort of stuff. I don't know, I don't think the benefits for a... Are necessarily there for running desktop applications. But yeah. The whole market's shifting. Who knows. I don't even know what people would write desktop applications for anymore.\\n\\n## Functional Programming for domains other than web apps?\\n\\nTIMOTHY: That's something I actually I mentioned on Twitter a while back, I would love to see somebody write something in a functional language that's NOT a web app, nor a backend service, nor an enterprise thing.\\n\\nDUSTIN: What you need... Yeah... The way the system works is too imperative. The one stack has so many layers, it almost gives you more opportunities to insert abstractions.\\n\\nTIMOTHY: Yeah. Yeah. Exactly. And I would love to see... You know... Hey tax season's coming. What would a tax application look like written in a functional language? You know whatever those things are written in, it's gotta be horrific. [laughter] And then there's also to apps on the desktop. But the most of what I see... Or actually I think what I actually said was, I would love to see a non-CRUD app written in a functional language 'cause there just aren't that many of 'em. Most of them are...\\n\\nDUSTIN: Well most stack's also they're about information processing, whereas the desktop apps are more about effects like...\\n\\nTIMOTHY: Yeah.\\n\\nDUSTIN: What's on the screen or something.\\n\\nTIMOTHY: And presentation and that sort of thing. Yeah. But as we've seen with React, you can put a functional veneer over that effect.\\n\\nDUSTIN: React is pretty slow in the small. You couldn't... You don't write games in it? Right?\\n\\nTIMOTHY: Right. Right. Yeah.\\n\\nTIMOTHY: Exactly. So shoutout here to Arcadia. That's what they're trying to do with a lot of their... Their stuff... The people behind the.NET Clojure implementation targeting Unity.\\n\\nDUSTIN: Yeah. I tried to use that two years ago.\\n\\nTIMOTHY: Yeah. So I think they finally got funding, so now there's full-time people working on it, and I think it'll get better in the near future. But... And then if... I always have to mention this, too. There's a 2013 QuakeCon. John Carmack got waxing eloquent about Haskell and functional programming. He's the guy that wrote Doom, for anyone that's not aware, and he started talking about how you could write a functional language, write a Doom-like first-person shooter in a functional language. It's about 40 minutes in... I'm sure there's a link to it online. You know, Carmack talks about Haskell at QuakeCon 2013.\\n\\nTIMOTHY: But he describes like a garbage collector and all of this, all the mechanics you would need to build the system, 'cause his whole thing was... Yeah... At the end of the day an FPS is tracking like 8 megabytes of mutable data, everything else is assets and textures and all that, and that it's really... It's very much feasible to write that sort of thing in a functional language. Except there's still problems, which is... Things you would never have thought of. The one example he gave is, in a purely functional world, two people come up to a door, and now they both wanna walk into the doorway. So you're here, next frame you're both trying to walk into the doorway. That's a problem that requires communication between the two things, they both can't walk through the doorway at once. And that's actually fairly hard to do in a functional language.\\n\\nDUSTIN: I wonder, it seems to me that if you look at the way software evolves, it can evolve for stupid reasons, like OpenGL is written in an imperative style API because the hardware's imperative. It busts the layer above OpenGL. Well, because OpenGL is imperative, it takes a really smart person to figure out how to turn that into like a garbage collector of virtual DOM or something, to commit that [0:47:29]. It takes a person who's paid to do the work, in order to make another imperative layer on top of it. It doesn't require brilliance. So you get these stack of things that are just like written in terms of the things below it.\\n\\nTIMOTHY: Yeah. Yeah. I think that's definitely the case.\\n\\nDUSTIN: So I would wanna see... I don't know. I would wanna see turtles all the way down.\\n\\nTIMOTHY: Exactly. Yeah. I would love to see that too.\",\"^1&lt;\",\"(let [{:keys [:hypercrud.browser\/fiddle]} ctx]\\n  [:div.container props\\n   [contrib.loader\/Loader! [\\\"\/\/cdn.plyr.io\/3.2.4\/plyr.js\\\" \\\"\/\/cdn.plyr.io\/3.2.4\/plyr.css\\\"] #(js\/Plyr. \\\".plyr__video-embed\\\")]\\n   [hyperfiddle.ui\/markdown (:fiddle\/markdown @fiddle) ctx]\\n   [hyperfiddle.ui\/result val ctx {}]])\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\",\"~ue40605db-fa1e-483d-9054-9f6dd0ee75ca\"],\"~:dustingetz.post\/hidden\",true],[\"~#QReq\",[[\"~:find\",[[\"~#list\",[\"~$pull\",\"~$?attr\",[\"~:attribute\/ident\",\"~:attribute\/renderer\"]]],\"~$...\"],\"~:where\",[\"^1C\",\"^1D\"]],[[\"^P\",[\"~rdatomic:free:\/\/datomic:4334\/tank\",null]]]]],[[\"^ \",\"^1D\",\"^18\",\"^1E\",\"hyperfiddle.ui.controls\/markdown-editor\"]]]],\"~:route\",[\"^X\"],\"~:schemas\",[\"^ \",\"^1\",[\"^ \",\"~:dustingetz.post\/related\",[\"^ \",\"^8\",107,\"^@\",\"^1J\",\"~:db\/valueType\",[\"^ \",\"^@\",\"~:db.type\/ref\"],\"~:db\/cardinality\",[\"^ \",\"^@\",\"~:db.cardinality\/many\"],\"~:db\/doc\",\"This post links to these posts, e.g. a table of contents, or \\\"if you liked this you might also like\\\"\"],\"~:db\/excise\",[\"^ \",\"^8\",15,\"^@\",\"^1P\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"~:db.cardinality\/one\"]],\"~:dustingetz.reg\/email\",[\"^ \",\"^8\",95,\"^@\",\"^1R\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/string\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"~:db\/unique\",[\"^ \",\"^@\",\"~:db.unique\/identity\"],\"~:db\/index\",true],\"~:helloworld\/email\",[\"^ \",\"^8\",94,\"^@\",\"^1W\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"hello world email \"],\"~:district\/region\",[\"^ \",\"^8\",93,\"^@\",\"^1X\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A district region enum value\"],\"~:db\/fn\",[\"^ \",\"^8\",52,\"^@\",\"^1Y\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/fn\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A function-valued attribute for direct use by transactions and queries.\"],\"~:person\/parents\",[\"^ \",\"^8\",141,\"^@\",\"^1[\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"^1V\",[\"^ \",\"^8\",44,\"^@\",\"^1V\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/boolean\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create an AVET index for the attribute. Defaults to false.\"],\"~:dustingetz.reg\/age\",[\"^ \",\"^8\",97,\"^@\",\"^21\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/long\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.streaker\/yoga\",[\"^ \",\"^8\",126,\"^@\",\"^23\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:community\/neighborhood\",[\"^ \",\"^8\",86,\"^@\",\"^24\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A community's neighborhood\"],\"^1T\",[\"^ \",\"^8\",42,\"^@\",\"^1T\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If value is :db.unique\/value, then attribute value is unique to each entity. Attempts to insert a duplicate value for a temporary entity id will fail. If value is :db.unique\/identity, then attribute value is unique, and upsert is enabled. Attempting to insert a duplicate value for a temporary entity id will cause all attributes associated with that temporary id to be merged with the entity already in the database. Defaults to nil.\"],\"~:dustingetz.streaker\/date\",[\"^ \",\"^8\",118,\"^@\",\"^25\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/instant\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:community\/type\",[\"^ \",\"^8\",89,\"^@\",\"^27\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"Community type enum values\"],\"~:dustingetz.post\/published-date\",[\"^ \",\"^8\",108,\"^@\",\"^28\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:community\/orgtype\",[\"^ \",\"^8\",88,\"^@\",\"^29\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A community orgtype enum value\"],\"~:db.excise\/beforeT\",[\"^ \",\"^8\",17,\"^@\",\"^2:\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1:\",[\"^ \",\"^8\",72,\"^@\",\"^1:\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Argument to `datomic.api\/pull`, defaults to $\"],\"^1;\",[\"^ \",\"^8\",65,\"^@\",\"^1;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"ClojureScript `user` namespace, available in :fiddle\/renderer.\"],\"~:dustingetz.streaker.lifting\/bench-1rm\",[\"^ \",\"^8\",127,\"^@\",\"^2;\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"One rep max\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.task\/title\",[\"^ \",\"^8\",110,\"^@\",\"^2&lt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^Y\",[\"^ \",\"^8\",66,\"^@\",\"^Y\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Fiddle CSS. \\n\\nWarning: CSS is not scoped, please write targetted CSS\"],\"^19\",[\"^ \",\"^8\",71,\"^@\",\"^19\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Datomic pull expression for the entity addressed by the URL\"],\"~:community\/name\",[\"^ \",\"^8\",84,\"^@\",\"^2=\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"~:db\/fulltext\",true,\"^1O\",\"A community's name\"],\"~:hyperfiddle\/archived\",[\"^ \",\"^8\",144,\"^@\",\"^2?\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.reg\/birthdate\",[\"^ \",\"^8\",98,\"^@\",\"^2@\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:db.sys\/reId\",[\"^ \",\"^8\",9,\"^@\",\"^2A\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute for an id e in the log that has been changed to id v in the index\"],\"^1K\",[\"^ \",\"^8\",40,\"^@\",\"^1K\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute that specifies the attribute's value type. Built-in value types include, :db.type\/keyword, :db.type\/string, :db.type\/ref, :db.type\/instant, :db.type\/long, :db.type\/bigdec, :db.type\/boolean, :db.type\/float, :db.type\/uuid, :db.type\/double, :db.type\/bigint,  :db.type\/uri.\"],\"~:db\/txInstant\",[\"^ \",\"^8\",50,\"^@\",\"^2B\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Attribute whose value is a :db.type\/instant. A :db\/txInstant is recorded automatically with every transaction.\"],\"~:community\/category\",[\"^ \",\"^8\",87,\"^@\",\"^2C\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^2&gt;\",true,\"^1O\",\"All community categories\"],\"~:dustingetz.streaker.lifting\/squat-1rm\",[\"^ \",\"^8\",128,\"^@\",\"^2D\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"One rep max\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.streaker\/teatotal\",[\"^ \",\"^8\",122,\"^@\",\"^2E\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"^14\",[\"^ \",\"^8\",79,\"^@\",\"^14\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"deprecated; function hook to influence target query inputs wiring, this is fully managed now\"],\"^W\",[\"^ \",\"^8\",68,\"^@\",\"^W\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/keyword\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Fiddle identifier used in URLs. Warning: changing this breaks fiddle URLs.\"],\"^1D\",[\"^ \",\"^8\",63,\"^@\",\"^1D\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"FK to schema, they can't be directly on $ schema because attribute renderers are a \\\"source code\\\" concern. TODO: move these off domain and into the fiddle repo.\"],\"^1&lt;\",[\"^ \",\"^8\",74,\"^@\",\"^1&lt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Reagent expression for the fiddle view\"],\"^1E\",[\"^ \",\"^8\",64,\"^@\",\"^1E\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Default attribute renderer, a CLJS var like `hyperfiddle.ui.controls\/code`.\"],\"~:db\/noHistory\",[\"^ \",\"^8\",45,\"^@\",\"^2G\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, past values of the attribute are not retained after indexing. Defaults to false.\"],\"^B\",[\"^ \",\"^8\",83,\"^@\",\"^B\",\"^1K\",[\"^ \",\"^@\",\"~:db.type\/uuid\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"Entity owners uuids, used by ACLs\"],\"^1&gt;\",[\"^ \",\"^8\",104,\"^@\",\"^1&gt;\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Omit from user-facing view but show in admin dashboards\"],\"~:dustingetz.post\/sort-index\",[\"^ \",\"^8\",102,\"^@\",\"^2I\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.streaker\/exercise\",[\"^ \",\"^8\",138,\"^@\",\"^2J\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:neighborhood\/district\",[\"^ \",\"^8\",91,\"^@\",\"^2K\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A neighborhood's district\"],\"~:dustingetz.task\/parent\",[\"^ \",\"^8\",136,\"^@\",\"^2L\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^12\",\"[:find\\n (pull ?e [:dustingetz.task\/completed\\n           :dustingetz.task\/title\\n           :db\/id])\\n :where\\n [?e :dustingetz.task\/title]\\n [?e :dustingetz.task\/parent [:fiddle\/ident :dustingetz\/games]]]\",\"^13\",\"~:query\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:db\/isComponent\",[\"^ \",\"^8\",43,\"^@\",\"^2N\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of attribute whose vtype is :db.type\/ref. If true, then the attribute is a component of the entity referencing it. When you query for an entire entity, components are fetched automatically. Defaults to nil.\"],\"~:dustingetz.streaker\/food-log\",[\"^ \",\"^8\",129,\"^@\",\"^2O\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.streaker\/weight-6am\",[\"^ \",\"^8\",137,\"^@\",\"^2P\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:db\/lang\",[\"^ \",\"^8\",46,\"^@\",\"^2Q\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Attribute of a data function. Value is a keyword naming the implementation language of the function. Legal values are :db.lang\/java and :db.lang\/clojure\"],\"~:dustingetz.post\/draft-placeholder\",[\"^ \",\"^8\",105,\"^@\",\"^2R\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Show in rendered view, but disabled (no link)\"],\"^2&gt;\",[\"^ \",\"^8\",51,\"^@\",\"^2&gt;\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create a fulltext search index for the attribute. Defaults to false.\"],\"~:putterson.todone\/done\",[\"^ \",\"^8\",143,\"^@\",\"^2S\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~u469401bc-9881-491c-a2b1-0aa1198ad3cf\"]],\"~:dustingetz.streaker\/gym\",[\"^ \",\"^8\",125,\"^@\",\"^2T\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Generally PIT\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.reg\/name\",[\"^ \",\"^8\",96,\"^@\",\"^2U\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz\/next\",[\"^ \",\"^8\",112,\"^@\",\"^2V\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.storm\/ident\",[\"^ \",\"^8\",114,\"^@\",\"^2W\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"]],\"~:dustingetz.slack-converter\/raw-source\",[\"^ \",\"^8\",111,\"^@\",\"^2X\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.streaker\/mood\",[\"^ \",\"^8\",121,\"^@\",\"^2Y\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz\/prev\",[\"^ \",\"^8\",113,\"^@\",\"^2Z\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.post\/published\",[\"^ \",\"^8\",103,\"^@\",\"^2[\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.task\/completed\",[\"^ \",\"^8\",109,\"^@\",\"^30\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:didier\/team-name\",[\"^ \",\"^8\",130,\"^@\",\"^31\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.streaker\/weight\",[\"^ \",\"^8\",117,\"^@\",\"^32\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Should be decimal, but js\/jvm coercion issues\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:putterson.todone\/title\",[\"^ \",\"^8\",142,\"^@\",\"^33\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~u469401bc-9881-491c-a2b1-0aa1198ad3cf\"]],\"~:dustingetz.nav\/children\",[\"^ \",\"^8\",106,\"^@\",\"^34\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"Each nav item has optional children\"],\"~:dustingetz.streaker\/comment\",[\"^ \",\"^8\",120,\"^@\",\"^35\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:zzz.fiddle\/request\",[\"^ \",\"^8\",75,\"^@\",\"^36\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"No longer used\"],\"~:person\/name\",[\"^ \",\"^8\",140,\"^@\",\"^37\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"^Z\",[\"^ \",\"^8\",133,\"^@\",\"^Z\",\"^1K\",[\"^ \",\"^@\",\"^2H\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"For naming anonymous fiddles\"],\"~:dustingetz\/foo\",[\"^ \",\"^8\",131,\"^@\",\"^38\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.post\/slug\",[\"^ \",\"^8\",115,\"^@\",\"^39\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Human readable unique identifier for this post \",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.streaker\/eating\",[\"^ \",\"^8\",123,\"^@\",\"^3:\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"^17\",[\"^ \",\"^8\",82,\"^@\",\"^17\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"names a hyperfiddle.api\/tx-fn multimethod which builds a transaction\"],\"^1M\",[\"^ \",\"^8\",41,\"^@\",\"^1M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. Two possible values: :db.cardinality\/one for single-valued attributes, and :db.cardinality\/many for many-valued attributes. Defaults to :db.cardinality\/one.\"],\"^1O\",[\"^ \",\"^8\",62,\"^@\",\"^1O\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"Documentation string for an entity.\"],\"^11\",[\"^ \",\"^8\",78,\"^@\",\"^11\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"link target\"],\"~:link\/rel\",[\"^ \",\"^8\",81,\"^@\",\"^3;\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"archived\"],\"^16\",[\"^ \",\"^8\",80,\"^@\",\"^16\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"todo rename; specifies the attribute for which this link is valid\"],\"^10\",[\"^ \",\"^8\",77,\"^@\",\"^10\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"semantic selector, like html css classes\"],\"~:dustingetz.streaker\/miles-ran\",[\"^ \",\"^8\",119,\"^@\",\"^3&lt;\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Miles run, should be float.\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:db.excise\/before\",[\"^ \",\"^8\",18,\"^@\",\"^3=\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^@\",[\"^ \",\"^8\",10,\"^@\",\"^@\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Attribute used to uniquely name an entity.\"],\"~:db\/code\",[\"^ \",\"^8\",47,\"^@\",\"^3&gt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"String-valued attribute of a data function that contains the function's source code.\"],\"^13\",[\"^ \",\"^8\",76,\"^@\",\"^13\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Which Datomic query API\"],\"~:dustingetz.gist\/src-clojure\",[\"^ \",\"^8\",135,\"^@\",\"^3?\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:dustingetz.post\/title\",[\"^ \",\"^8\",101,\"^@\",\"^3@\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^[\",[\"^ \",\"^8\",69,\"^@\",\"^[\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^2N\",true,\"^1O\",\"Links to other fiddles that are available from this fiddle\"],\"~:dustingetz.reg\/gender\",[\"^ \",\"^8\",99,\"^@\",\"^3A\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:dustingetz.streaker\/intention\",[\"^ \",\"^8\",146,\"^@\",\"^3B\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:db.install\/valueType\",[\"^ \",\"^8\",12,\"^@\",\"^3C\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a value type.\"],\"~:neighborhood\/name\",[\"^ \",\"^8\",90,\"^@\",\"^3D\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"A unique neighborhood name (upsertable)\"],\"~:db.alter\/attribute\",[\"^ \",\"^8\",19,\"^@\",\"^3E\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will alter the definition of existing attribute v.\"],\"~:district\/name\",[\"^ \",\"^8\",92,\"^@\",\"^3F\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"A unique district name (upsertable)\"],\"~:db.install\/function\",[\"^ \",\"^8\",14,\"^@\",\"^3G\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a data function.\"],\"~:hyperfiddle\/deprecated\",[\"^ \",\"^8\",145,\"^@\",\"^3H\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:db.install\/partition\",[\"^ \",\"^8\",11,\"^@\",\"^3I\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a partition.\"],\"~:dustingetz\/foo1\",[\"^ \",\"^8\",132,\"^@\",\"^3J\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^18\",[\"^ \",\"^8\",70,\"^@\",\"^18\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Markdown expression for fiddle view, optional\"],\"~:dustingetz.streaker\/pt\",[\"^ \",\"^8\",124,\"^@\",\"^3K\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"personal trainer\",\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"^1=\",[\"^ \",\"^8\",67,\"^@\",\"^1=\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Experimental. When set, data-sync will interpret this fiddle's result as a fiddle - like a higher order fiddle - this is a recursion mechanic.\"],\"~:dustingetz.streaker\/diet\",[\"^ \",\"^8\",139,\"^@\",\"^3L\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^B\",[\"~uacd054a8-4e36-4d6c-a9ec-95bdc47f0d39\"]],\"~:db.install\/attribute\",[\"^ \",\"^8\",13,\"^@\",\"^3M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as an attribute.\"],\"~:db.excise\/attrs\",[\"^ \",\"^8\",16,\"^@\",\"^3N\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"]],\"~:dustingetz.reg\/shirt-size\",[\"^ \",\"^8\",100,\"^@\",\"^3O\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:hyperfiddle\/starred\",[\"^ \",\"^8\",134,\"^@\",\"^3P\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:fressian\/tag\",[\"^ \",\"^8\",39,\"^@\",\"^3Q\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Keyword-valued attribute of a value type that specifies the underlying fressian type used for serialization.\"],\"^12\",[\"^ \",\"^8\",73,\"^@\",\"^12\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Datomic query datalog. \\n\\nWarning: no support yet for rules, d\/history, d\/log or other datomic API access.\"],\"~:dustingetz.storm\/channel\",[\"^ \",\"^8\",116,\"^@\",\"^3R\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:community\/url\",[\"^ \",\"^8\",85,\"^@\",\"^3S\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A community's url\"],\"~:db.sys\/partiallyIndexed\",[\"^ \",\"^8\",8,\"^@\",\"^3T\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute set to true for transactions not fully incorporated into the index\"]],\"^2\",[\"^ \",\"~:link\/create?\",[\"^ \",\"^8\",79,\"^@\",\"^3U\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Create a new record (tempid entity).\"],\"^1P\",[\"^ \",\"^8\",15,\"^@\",\"^1P\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:hypercrud\/props\",[\"^ \",\"^8\",78,\"^@\",\"^3V\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Map of extra user props for the renderer, e.g. `:disabled` `:label-fn` `:tooltip`\"],\"~:hfnet.beta\/real-name\",[\"^ \",\"^8\",92,\"^@\",\"^3W\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1Y\",[\"^ \",\"^8\",52,\"^@\",\"^1Y\",\"^1K\",[\"^ \",\"^@\",\"^1Z\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A function-valued attribute for direct use by transactions and queries.\"],\"^1V\",[\"^ \",\"^8\",44,\"^@\",\"^1V\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create an AVET index for the attribute. Defaults to false.\"],\"^1T\",[\"^ \",\"^8\",42,\"^@\",\"^1T\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If value is :db.unique\/value, then attribute value is unique to each entity. Attempts to insert a duplicate value for a temporary entity id will fail. If value is :db.unique\/identity, then attribute value is unique, and upsert is enabled. Attempting to insert a duplicate value for a temporary entity id will cause all attributes associated with that temporary id to be merged with the entity already in the database. Defaults to nil.\"],\"^2:\",[\"^ \",\"^8\",17,\"^@\",\"^2:\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1:\",[\"^ \",\"^8\",73,\"^@\",\"^1:\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Argument to `datomic.api\/pull`\"],\"^1;\",[\"^ \",\"^8\",66,\"^@\",\"^1;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Undocumented, pending cleanup Q3'18\"],\"~:fiddle\/request\",[\"^ \",\"^8\",76,\"^@\",\"^3X\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^Y\",[\"^ \",\"^8\",67,\"^@\",\"^Y\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"CSS for this fiddle\"],\"^19\",[\"^ \",\"^8\",72,\"^@\",\"^19\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Datomic pull expression. Drives form fields and order. If you omit `:db\/id` the form will be read-only.\"],\"^2A\",[\"^ \",\"^8\",9,\"^@\",\"^2A\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute for an id e in the log that has been changed to id v in the index\"],\"~:hfnet.beta\/accepted-on\",[\"^ \",\"^8\",89,\"^@\",\"^3Y\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1K\",[\"^ \",\"^8\",40,\"^@\",\"^1K\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute that specifies the attribute's value type. Built-in value types include, :db.type\/keyword, :db.type\/string, :db.type\/ref, :db.type\/instant, :db.type\/long, :db.type\/bigdec, :db.type\/boolean, :db.type\/float, :db.type\/uuid, :db.type\/double, :db.type\/bigint,  :db.type\/uri.\"],\"^2B\",[\"^ \",\"^8\",50,\"^@\",\"^2B\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Attribute whose value is a :db.type\/instant. A :db\/txInstant is recorded automatically with every transaction.\"],\"^14\",[\"^ \",\"^8\",83,\"^@\",\"^14\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"CLJC expression which specifies target fiddle's query input arguments as a value or list of values. Hyperfiddle provides a default formula where sensible.\"],\"^W\",[\"^ \",\"^8\",69,\"^@\",\"^W\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Identifies this fiddle and used in fiddle URLs.\"],\"^1D\",[\"^ \",\"^8\",63,\"^@\",\"^1D\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Named schema attribute\"],\"^1&lt;\",[\"^ \",\"^8\",75,\"^@\",\"^1&lt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Reagent hiccup expression for the view. Clear this to restore default.\"],\"^1E\",[\"^ \",\"^8\",64,\"^@\",\"^1E\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Fully qualified var like `hyperfiddle.ui.controls\/code`. This is not the actual Datomic schema entity because \\\"source code\\\" concerns like renderers are not schema.\"],\"^2G\",[\"^ \",\"^8\",45,\"^@\",\"^2G\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, past values of the attribute are not retained after indexing. Defaults to false.\"],\"^2N\",[\"^ \",\"^8\",43,\"^@\",\"^2N\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of attribute whose vtype is :db.type\/ref. If true, then the attribute is a component of the entity referencing it. When you query for an entire entity, components are fetched automatically. Defaults to nil.\"],\"^2Q\",[\"^ \",\"^8\",46,\"^@\",\"^2Q\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Attribute of a data function. Value is a keyword naming the implementation language of the function. Legal values are :db.lang\/java and :db.lang\/clojure\"],\"~:link\/render-inline?\",[\"^ \",\"^8\",87,\"^@\",\"^3Z\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Embed target fiddle (like an iframe)\"],\"^2&gt;\",[\"^ \",\"^8\",51,\"^@\",\"^2&gt;\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create a fulltext search index for the attribute. Defaults to false.\"],\"~:hfnet.beta\/on-deck\",[\"^ \",\"^8\",91,\"^@\",\"^3[\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:link\/dependent?\",[\"^ \",\"^8\",80,\"^@\",\"^40\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Indicates the target fiddle has a dependency on data from this fiddle. Only set if necessary because it makes the query more expensive.\"],\"~:google.analytics\/client-id\",[\"^ \",\"^8\",97,\"^@\",\"^41\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"]],\"~:hfnet.beta\/slack-handle\",[\"^ \",\"^8\",90,\"^@\",\"^42\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1V\",true],\"^17\",[\"^ \",\"^8\",88,\"^@\",\"^17\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"~`:button` click handler, a function that can modify the form value before staging as a Datomic transaction. managed+create links have a default `:tx-fn` implementing a parent-child reference.\"],\"^1M\",[\"^ \",\"^8\",41,\"^@\",\"^1M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. Two possible values: :db.cardinality\/one for single-valued attributes, and :db.cardinality\/many for many-valued attributes. Defaults to :db.cardinality\/one.\"],\"^1O\",[\"^ \",\"^8\",62,\"^@\",\"^1O\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"Documentation string for an entity.\"],\"^11\",[\"^ \",\"^8\",82,\"^@\",\"^11\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Link to other fiddles\"],\"^3;\",[\"^ \",\"^8\",86,\"^@\",\"^3;\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Names the semantic relationship between this fiddle and the target fiddle e.g. `:options`. Widget renderers may use this hint to render views intelligently.\"],\"^16\",[\"^ \",\"^8\",85,\"^@\",\"^16\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Associate this link in a specific position of the form, e.g. `\\\"0 :reg\/gender\\\"` where 0 is the find-element index.\"],\"~:hfnet.beta\/comment\",[\"^ \",\"^8\",93,\"^@\",\"^43\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^3=\",[\"^ \",\"^8\",18,\"^@\",\"^3=\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^@\",[\"^ \",\"^8\",10,\"^@\",\"^@\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Attribute used to uniquely name an entity.\"],\"^3&gt;\",[\"^ \",\"^8\",47,\"^@\",\"^3&gt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"String-valued attribute of a data function that contains the function's source code.\"],\"^13\",[\"^ \",\"^8\",77,\"^@\",\"^13\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Which Datomic API to use for data fetching, blank means nothing\"],\"~:user\/user-id\",[\"^ \",\"^8\",96,\"^@\",\"^44\",\"^1K\",[\"^ \",\"^@\",\"^2H\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"FK to $users\"],\"^[\",[\"^ \",\"^8\",70,\"^@\",\"^[\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^2N\",true,\"^1O\",\"Links to other fiddles. This is a more advanced feature.\"],\"^3C\",[\"^ \",\"^8\",12,\"^@\",\"^3C\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a value type.\"],\"^3E\",[\"^ \",\"^8\",19,\"^@\",\"^3E\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will alter the definition of existing attribute v.\"],\"^3G\",[\"^ \",\"^8\",14,\"^@\",\"^3G\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a data function.\"],\"~:fiddle\/bindings\",[\"^ \",\"^8\",65,\"^@\",\"^45\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Deprecated internal hook `(fn [ctx] (assoc ctx ...))`\"],\"^3I\",[\"^ \",\"^8\",11,\"^@\",\"^3I\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a partition.\"],\"^18\",[\"^ \",\"^8\",71,\"^@\",\"^18\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Convenient place to store markdown, `:fiddle\/renderer` must render this.\"],\"^1=\",[\"^ \",\"^8\",68,\"^@\",\"^1=\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Experimental\"],\"~:hfnet.beta\/archived\",[\"^ \",\"^8\",95,\"^@\",\"^46\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:link\/disabled?\",[\"^ \",\"^8\",81,\"^@\",\"^47\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Disabled links are not shown and do not load data\"],\"^3M\",[\"^ \",\"^8\",13,\"^@\",\"^3M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as an attribute.\"],\"^3N\",[\"^ \",\"^8\",16,\"^@\",\"^3N\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"]],\"^3Q\",[\"^ \",\"^8\",39,\"^@\",\"^3Q\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Keyword-valued attribute of a value type that specifies the underlying fressian type used for serialization.\"],\"^12\",[\"^ \",\"^8\",74,\"^@\",\"^12\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Datomic query. Database inputs are resolved by name through the `:domain\/environment`. Pull `:db\/id` for editable forms. Currently no support yet for rules, d\/history or d\/log.\"],\"~:zzz.google.analytics\/client-id\",[\"^ \",\"^8\",94,\"^@\",\"^48\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"]],\"^3T\",[\"^ \",\"^8\",8,\"^@\",\"^3T\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute set to true for transactions not fully incorporated into the index\"],\"~:link\/managed?\",[\"^ \",\"^8\",84,\"^@\",\"^49\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Allocate a \\\"branch\\\" (`datomic.api\/with`) for this form. This lets a `:link\/tx-fn` control the transaction.\"]],\"^3\",[\"^ \",\"~:user-profile\/display-name-normalized\",[\"^ \",\"^8\",83,\"^@\",\"^4:\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:message\/user\",[\"^ \",\"^8\",66,\"^@\",\"^4;\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"User who posted the message\"],\"^1P\",[\"^ \",\"^8\",15,\"^@\",\"^1P\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1Y\",[\"^ \",\"^8\",52,\"^@\",\"^1Y\",\"^1K\",[\"^ \",\"^@\",\"^1Z\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A function-valued attribute for direct use by transactions and queries.\"],\"~:user\/profile\",[\"^ \",\"^8\",78,\"^@\",\"^4&lt;\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Reference to a user's profile, which contains email, avatar, display name, etc.\"],\"^1V\",[\"^ \",\"^8\",44,\"^@\",\"^1V\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create an AVET index for the attribute. Defaults to false.\"],\"^1T\",[\"^ \",\"^8\",42,\"^@\",\"^1T\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If value is :db.unique\/value, then attribute value is unique to each entity. Attempts to insert a duplicate value for a temporary entity id will fail. If value is :db.unique\/identity, then attribute value is unique, and upsert is enabled. Attempting to insert a duplicate value for a temporary entity id will cause all attributes associated with that temporary id to be merged with the entity already in the database. Defaults to nil.\"],\"~:message\/key\",[\"^ \",\"^8\",63,\"^@\",\"^4=\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Key consisting of channel id + ts. Useful for upserts and refs.\"],\"^2:\",[\"^ \",\"^8\",17,\"^@\",\"^2:\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user\/owner?\",[\"^ \",\"^8\",77,\"^@\",\"^4&gt;\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user\/name\",[\"^ \",\"^8\",74,\"^@\",\"^4?\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"A user's public handle.\"],\"~:message\/text\",[\"^ \",\"^8\",64,\"^@\",\"^4@\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Message text (markdown)\"],\"~:user-profile\/image-original\",[\"^ \",\"^8\",88,\"^@\",\"^4A\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:channel\/creator\",[\"^ \",\"^8\",98,\"^@\",\"^4B\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^2A\",[\"^ \",\"^8\",9,\"^@\",\"^2A\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute for an id e in the log that has been changed to id v in the index\"],\"^1K\",[\"^ \",\"^8\",40,\"^@\",\"^1K\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute that specifies the attribute's value type. Built-in value types include, :db.type\/keyword, :db.type\/string, :db.type\/ref, :db.type\/instant, :db.type\/long, :db.type\/bigdec, :db.type\/boolean, :db.type\/float, :db.type\/uuid, :db.type\/double, :db.type\/bigint,  :db.type\/uri.\"],\"^2B\",[\"^ \",\"^8\",50,\"^@\",\"^2B\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Attribute whose value is a :db.type\/instant. A :db\/txInstant is recorded automatically with every transaction.\"],\"~:message\/ts\",[\"^ \",\"^8\",67,\"^@\",\"^4C\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Message timestamp (seconds since epoch up to 6 decimals). Stored as string because it is used by slack as a kind of identifier. Unique per channel.\"],\"~:message\/day\",[\"^ \",\"^8\",70,\"^@\",\"^4D\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The day this message is categorized under, e.g. 2017-11-20.\"],\"~:user-profile\/real-name-normalized\",[\"^ \",\"^8\",85,\"^@\",\"^4E\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/avatar-hash\",[\"^ \",\"^8\",86,\"^@\",\"^4F\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/email\",[\"^ \",\"^8\",79,\"^@\",\"^4G\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^2G\",[\"^ \",\"^8\",45,\"^@\",\"^2G\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, past values of the attribute are not retained after indexing. Defaults to false.\"],\"~:user-profile\/first-name\",[\"^ \",\"^8\",80,\"^@\",\"^4H\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^2N\",[\"^ \",\"^8\",43,\"^@\",\"^2N\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of attribute whose vtype is :db.type\/ref. If true, then the attribute is a component of the entity referencing it. When you query for an entire entity, components are fetched automatically. Defaults to nil.\"],\"~:channel\/name\",[\"^ \",\"^8\",96,\"^@\",\"^4I\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"]],\"~:user-profile\/image-32\",[\"^ \",\"^8\",90,\"^@\",\"^4J\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^2Q\",[\"^ \",\"^8\",46,\"^@\",\"^2Q\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Attribute of a data function. Value is a keyword naming the implementation language of the function. Legal values are :db.lang\/java and :db.lang\/clojure\"],\"^2&gt;\",[\"^ \",\"^8\",51,\"^@\",\"^2&gt;\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create a fulltext search index for the attribute. Defaults to false.\"],\"~:channel\/created\",[\"^ \",\"^8\",97,\"^@\",\"^4K\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/image-48\",[\"^ \",\"^8\",91,\"^@\",\"^4L\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:channel\/slack-id\",[\"^ \",\"^8\",95,\"^@\",\"^4M\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"]],\"~:message\/thread-ts\",[\"^ \",\"^8\",68,\"^@\",\"^4N\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Thread parent message timestamp (seconds since epoch up to 6 decimals). Stored as string because it is used by slack as a kind of identifier. Unique per channel.\"],\"~:user-profile\/display-name\",[\"^ \",\"^8\",82,\"^@\",\"^4O\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/image-72\",[\"^ \",\"^8\",92,\"^@\",\"^4P\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:message\/channel\",[\"^ \",\"^8\",65,\"^@\",\"^4Q\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Channel the message was posted in.\"],\"~:event\/subtype\",[\"^ \",\"^8\",71,\"^@\",\"^4R\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1M\",[\"^ \",\"^8\",41,\"^@\",\"^1M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. Two possible values: :db.cardinality\/one for single-valued attributes, and :db.cardinality\/many for many-valued attributes. Defaults to :db.cardinality\/one.\"],\"~:user-profile\/image-192\",[\"^ \",\"^8\",93,\"^@\",\"^4S\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^1O\",[\"^ \",\"^8\",62,\"^@\",\"^1O\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"Documentation string for an entity.\"],\"~:event\/ts\",[\"^ \",\"^8\",72,\"^@\",\"^4T\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user\/real-name\",[\"^ \",\"^8\",75,\"^@\",\"^4U\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/title\",[\"^ \",\"^8\",87,\"^@\",\"^4V\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^3=\",[\"^ \",\"^8\",18,\"^@\",\"^3=\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:message\/thread-inst\",[\"^ \",\"^8\",69,\"^@\",\"^4W\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Same as :message\/thread-ts, but parsed to java.util.Date. This is lossy: Date has milisecond precision, ts has microseconds.\"],\"^@\",[\"^ \",\"^8\",10,\"^@\",\"^@\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Attribute used to uniquely name an entity.\"],\"^3&gt;\",[\"^ \",\"^8\",47,\"^@\",\"^3&gt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"String-valued attribute of a data function that contains the function's source code.\"],\"~:user\/slack-id\",[\"^ \",\"^8\",73,\"^@\",\"^4X\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Internal user identifier used by slack. Alphanumeric, starts with U.\"],\"~:user-profile\/last-name\",[\"^ \",\"^8\",81,\"^@\",\"^4Y\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/image-512\",[\"^ \",\"^8\",94,\"^@\",\"^4Z\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:user-profile\/real-name\",[\"^ \",\"^8\",84,\"^@\",\"^4[\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^3C\",[\"^ \",\"^8\",12,\"^@\",\"^3C\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a value type.\"],\"^3E\",[\"^ \",\"^8\",19,\"^@\",\"^3E\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will alter the definition of existing attribute v.\"],\"^3G\",[\"^ \",\"^8\",14,\"^@\",\"^3G\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a data function.\"],\"^3I\",[\"^ \",\"^8\",11,\"^@\",\"^3I\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a partition.\"],\"^3M\",[\"^ \",\"^8\",13,\"^@\",\"^3M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as an attribute.\"],\"~:user\/admin?\",[\"^ \",\"^8\",76,\"^@\",\"^50\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"^3N\",[\"^ \",\"^8\",16,\"^@\",\"^3N\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"]],\"^3Q\",[\"^ \",\"^8\",39,\"^@\",\"^3Q\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Keyword-valued attribute of a value type that specifies the underlying fressian type used for serialization.\"],\"^3T\",[\"^ \",\"^8\",8,\"^@\",\"^3T\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute set to true for transactions not fully incorporated into the index\"],\"~:user-profile\/image-24\",[\"^ \",\"^8\",89,\"^@\",\"^51\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]]],\"^4\",[\"^ \",\"~:release\/script\",[\"^ \",\"^8\",97,\"^@\",\"^52\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The script used in the release\"],\"~:label\/type\",[\"^ \",\"^8\",86,\"^@\",\"^53\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Enum, one of :label.type\/distributor, :label.type\/holding,\\n  :label.type\/production, :label.type\/originalProduction,\\n  :label.type\/bootlegProduction, :label.type\/reissueProduction, or\\n  :label.type\/publisher.\"],\"~:artist\/sortName\",[\"^ \",\"^8\",68,\"^@\",\"^54\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The artist's name for use in alphabetical sorting, e.g. Beatles, The\"],\"~:abstractRelease\/artists\",[\"^ \",\"^8\",81,\"^@\",\"^55\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"The set of artists contributing to the abstract release\"],\"^1P\",[\"^ \",\"^8\",15,\"^@\",\"^1P\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:release\/packaging\",[\"^ \",\"^8\",102,\"^@\",\"^56\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The type of packaging used in the release, an enum, one\\n  of: :release.packaging\/jewelCase, :release.packaging\/slimJewelCase, :release.packaging\/digipak, :release.packaging\/other\\n  , :release.packaging\/keepCase, :release.packaging\/none,\\n  or :release.packaging\/cardboardPaperSleeve\"],\"~:abstractRelease\/type\",[\"^ \",\"^8\",80,\"^@\",\"^57\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Enum, one\\n  of: :release.type\/album, :release.type\/single, :release.type\/ep, :release.type\/audiobook,\\n  or :release.type\/other\"],\"~:release\/name\",[\"^ \",\"^8\",100,\"^@\",\"^58\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^2&gt;\",true,\"^1O\",\"The name of the release\"],\"^1Y\",[\"^ \",\"^8\",52,\"^@\",\"^1Y\",\"^1K\",[\"^ \",\"^@\",\"^1Z\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"A function-valued attribute for direct use by transactions and queries.\"],\"~:release\/artists\",[\"^ \",\"^8\",107,\"^@\",\"^59\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"The set of artists contributing to the release\"],\"~:artist\/name\",[\"^ \",\"^8\",67,\"^@\",\"^5:\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^2&gt;\",true,\"^1O\",\"The artist's name\"],\"^1V\",[\"^ \",\"^8\",44,\"^@\",\"^1V\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create an AVET index for the attribute. Defaults to false.\"],\"~:label\/endYear\",[\"^ \",\"^8\",91,\"^@\",\"^5;\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The year the label stopped business\"],\"~:medium\/position\",[\"^ \",\"^8\",112,\"^@\",\"^5&lt;\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The position of this medium in the release relative to the other media, i.e. disc 1\"],\"~:artist\/type\",[\"^ \",\"^8\",69,\"^@\",\"^5=\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Enum, one of :artist.type\/person, :artist.type\/other, :artist.type\/group.\"],\"^1T\",[\"^ \",\"^8\",42,\"^@\",\"^1T\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If value is :db.unique\/value, then attribute value is unique to each entity. Attempts to insert a duplicate value for a temporary entity id will fail. If value is :db.unique\/identity, then attribute value is unique, and upsert is enabled. Attempting to insert a duplicate value for a temporary entity id will cause all attributes associated with that temporary id to be merged with the entity already in the database. Defaults to nil.\"],\"~:abstractRelease\/name\",[\"^ \",\"^8\",79,\"^@\",\"^5&gt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The name of the abstract release\"],\"~:label\/startMonth\",[\"^ \",\"^8\",89,\"^@\",\"^5?\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The month the label started business\"],\"^2:\",[\"^ \",\"^8\",17,\"^@\",\"^2:\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:script\/name\",[\"^ \",\"^8\",65,\"^@\",\"^5@\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"~:db.unique\/value\"],\"^1O\",\"Name of written character set, e.g. Hebrew, Latin, Cyrillic\"],\"~:abstractRelease\/gid\",[\"^ \",\"^8\",78,\"^@\",\"^5B\",\"^1K\",[\"^ \",\"^@\",\"^2H\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"The globally unique MusicBrainz ID for the abstract release\"],\"~:track\/name\",[\"^ \",\"^8\",118,\"^@\",\"^5C\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^2&gt;\",true,\"^1O\",\"The track name\"],\"~:track\/artists\",[\"^ \",\"^8\",115,\"^@\",\"^5D\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"The artists who contributed to the track\"],\"~:release\/country\",[\"^ \",\"^8\",95,\"^@\",\"^5E\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The country where the recording was released\"],\"~:artist\/country\",[\"^ \",\"^8\",71,\"^@\",\"^5F\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The artist's country of origin\"],\"~:artist\/gid\",[\"^ \",\"^8\",66,\"^@\",\"^5G\",\"^1K\",[\"^ \",\"^@\",\"^2H\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"The globally unique MusicBrainz ID for an artist\"],\"^2A\",[\"^ \",\"^8\",9,\"^@\",\"^2A\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute for an id e in the log that has been changed to id v in the index\"],\"^1K\",[\"^ \",\"^8\",40,\"^@\",\"^1K\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute that specifies the attribute's value type. Built-in value types include, :db.type\/keyword, :db.type\/string, :db.type\/ref, :db.type\/instant, :db.type\/long, :db.type\/bigdec, :db.type\/boolean, :db.type\/float, :db.type\/uuid, :db.type\/double, :db.type\/bigint,  :db.type\/uri.\"],\"~:artist\/startDay\",[\"^ \",\"^8\",74,\"^@\",\"^5H\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The day the artist started actively recording\"],\"~:release\/gid\",[\"^ \",\"^8\",94,\"^@\",\"^5I\",\"^1K\",[\"^ \",\"^@\",\"^2H\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"The globally unique MusicBrainz ID for the release\"],\"~:artist\/endDay\",[\"^ \",\"^8\",77,\"^@\",\"^5J\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The day the artist stopped actively recording\"],\"~:label\/country\",[\"^ \",\"^8\",87,\"^@\",\"^5K\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The country where the record label is located\"],\"^2B\",[\"^ \",\"^8\",50,\"^@\",\"^2B\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Attribute whose value is a :db.type\/instant. A :db\/txInstant is recorded automatically with every transaction.\"],\"~:artist\/startYear\",[\"^ \",\"^8\",72,\"^@\",\"^5L\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The year the artist started actively recording\"],\"~:release\/labels\",[\"^ \",\"^8\",96,\"^@\",\"^5M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"The label on which the recording was released\"],\"~:artist\/endMonth\",[\"^ \",\"^8\",76,\"^@\",\"^5N\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The month the artist stopped actively recording\"],\"~:release\/language\",[\"^ \",\"^8\",98,\"^@\",\"^5O\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The language used in the release\"],\"~:medium\/format\",[\"^ \",\"^8\",111,\"^@\",\"^5P\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The format of the medium. An enum with lots of possible values\"],\"~:release\/day\",[\"^ \",\"^8\",105,\"^@\",\"^5Q\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The day of the release\"],\"^2G\",[\"^ \",\"^8\",45,\"^@\",\"^2G\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, past values of the attribute are not retained after indexing. Defaults to false.\"],\"~:release\/status\",[\"^ \",\"^8\",109,\"^@\",\"^5R\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The status of the release\"],\"~:label\/sortName\",[\"^ \",\"^8\",85,\"^@\",\"^5S\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The name of the record label for use in alphabetical sorting\"],\"^2N\",[\"^ \",\"^8\",43,\"^@\",\"^2N\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of attribute whose vtype is :db.type\/ref. If true, then the attribute is a component of the entity referencing it. When you query for an entire entity, components are fetched automatically. Defaults to nil.\"],\"^2Q\",[\"^ \",\"^8\",46,\"^@\",\"^2Q\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Attribute of a data function. Value is a keyword naming the implementation language of the function. Legal values are :db.lang\/java and :db.lang\/clojure\"],\"^2&gt;\",[\"^ \",\"^8\",51,\"^@\",\"^2&gt;\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. If true, create a fulltext search index for the attribute. Defaults to false.\"],\"~:language\/name\",[\"^ \",\"^8\",64,\"^@\",\"^5T\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^5A\"],\"^1O\",\"The name of the written and spoken language\"],\"~:artist\/endYear\",[\"^ \",\"^8\",75,\"^@\",\"^5U\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The year the artist stopped actively recording\"],\"~:release\/month\",[\"^ \",\"^8\",104,\"^@\",\"^5V\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The month of the release\"],\"~:release\/artistCredit\",[\"^ \",\"^8\",106,\"^@\",\"^5W\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"The string represenation of the artist(s) to be credited on the release\"],\"~:label\/gid\",[\"^ \",\"^8\",83,\"^@\",\"^5X\",\"^1K\",[\"^ \",\"^@\",\"^2H\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"The globally unique MusicBrainz ID for the record label\"],\"~:country\/name\",[\"^ \",\"^8\",63,\"^@\",\"^5Y\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^5A\"],\"^1O\",\"The name of the country\"],\"~:medium\/name\",[\"^ \",\"^8\",113,\"^@\",\"^5Z\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"The name of the medium itself, distinct from the name of the release\"],\"~:abstractRelease\/artistCredit\",[\"^ \",\"^8\",82,\"^@\",\"^5[\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"The string represenation of the artist(s) to be credited on the abstract release\"],\"~:release\/abstractRelease\",[\"^ \",\"^8\",108,\"^@\",\"^60\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"This release is the physical manifestation of the\\n  associated abstract release, e.g. the the 1984 US vinyl release of\\n  \\\"The Wall\\\" by Columbia, as opposed to the 2000 US CD release of\\n  \\\"The Wall\\\" by Capitol Records.\"],\"~:label\/startDay\",[\"^ \",\"^8\",90,\"^@\",\"^61\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The day the label started business\"],\"^1M\",[\"^ \",\"^8\",41,\"^@\",\"^1M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Property of an attribute. Two possible values: :db.cardinality\/one for single-valued attributes, and :db.cardinality\/many for many-valued attributes. Defaults to :db.cardinality\/one.\"],\"^1O\",[\"^ \",\"^8\",62,\"^@\",\"^1O\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"Documentation string for an entity.\"],\"^3=\",[\"^ \",\"^8\",18,\"^@\",\"^3=\",\"^1K\",[\"^ \",\"^@\",\"^26\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"]],\"~:artist\/startMonth\",[\"^ \",\"^8\",73,\"^@\",\"^62\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The month the artist started actively recording\"],\"~:track\/duration\",[\"^ \",\"^8\",119,\"^@\",\"^63\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The duration of the track in msecs\"],\"^@\",[\"^ \",\"^8\",10,\"^@\",\"^@\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1T\",[\"^ \",\"^@\",\"^1U\"],\"^1O\",\"Attribute used to uniquely name an entity.\"],\"^3&gt;\",[\"^ \",\"^8\",47,\"^@\",\"^3&gt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"String-valued attribute of a data function that contains the function's source code.\"],\"~:artist\/gender\",[\"^ \",\"^8\",70,\"^@\",\"^64\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"Enum, one of :artist.gender\/male, :artist.gender\/female, or :artist.gender\/other.\"],\"~:track\/position\",[\"^ \",\"^8\",117,\"^@\",\"^65\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The position of the track relative to the other tracks on the medium\"],\"~:medium\/trackCount\",[\"^ \",\"^8\",114,\"^@\",\"^66\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The total number of tracks on the medium\"],\"~:release\/year\",[\"^ \",\"^8\",103,\"^@\",\"^67\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The year of the release\"],\"^3C\",[\"^ \",\"^8\",12,\"^@\",\"^3C\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a value type.\"],\"~:label\/endMonth\",[\"^ \",\"^8\",92,\"^@\",\"^68\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The month the label stopped business\"],\"^3E\",[\"^ \",\"^8\",19,\"^@\",\"^3E\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will alter the definition of existing attribute v.\"],\"^3G\",[\"^ \",\"^8\",14,\"^@\",\"^3G\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a data function.\"],\"~:release\/media\",[\"^ \",\"^8\",101,\"^@\",\"^69\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^2N\",true,\"^1O\",\"The various media (CDs, vinyl records, cassette tapes, etc.) included in the release.\"],\"~:label\/endDay\",[\"^ \",\"^8\",93,\"^@\",\"^6:\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The day the label stopped business\"],\"^3I\",[\"^ \",\"^8\",11,\"^@\",\"^3I\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as a partition.\"],\"~:label\/startYear\",[\"^ \",\"^8\",88,\"^@\",\"^6;\",\"^1K\",[\"^ \",\"^@\",\"^22\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"The year the label started business\"],\"~:label\/name\",[\"^ \",\"^8\",84,\"^@\",\"^6&lt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^2&gt;\",true,\"^1O\",\"The name of the record label\"],\"^3M\",[\"^ \",\"^8\",13,\"^@\",\"^3M\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^1O\",\"System attribute with type :db.type\/ref. Asserting this attribute on :db.part\/db with value v will install v as an attribute.\"],\"^3N\",[\"^ \",\"^8\",16,\"^@\",\"^3N\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"]],\"~:medium\/tracks\",[\"^ \",\"^8\",110,\"^@\",\"^6=\",\"^1K\",[\"^ \",\"^@\",\"^1L\"],\"^1M\",[\"^ \",\"^@\",\"^1N\"],\"^2N\",true,\"^1O\",\"The set of tracks found on this medium\"],\"^3Q\",[\"^ \",\"^8\",39,\"^@\",\"^3Q\",\"^1K\",[\"^ \",\"^@\",\"^2F\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1V\",true,\"^1O\",\"Keyword-valued attribute of a value type that specifies the underlying fressian type used for serialization.\"],\"~:release\/barcode\",[\"^ \",\"^8\",99,\"^@\",\"^6&gt;\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"The barcode on the release packaging\"],\"^3T\",[\"^ \",\"^8\",8,\"^@\",\"^3T\",\"^1K\",[\"^ \",\"^@\",\"^20\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^1O\",\"System-assigned attribute set to true for transactions not fully incorporated into the index\"],\"~:track\/artistCredit\",[\"^ \",\"^8\",116,\"^@\",\"^6?\",\"^1K\",[\"^ \",\"^@\",\"^1S\"],\"^1M\",[\"^ \",\"^@\",\"^1Q\"],\"^2&gt;\",true,\"^1O\",\"The artists who contributed to the track\"]]],\"~:hyperfiddle.runtime\/branch-aux\",[\"^ \",\"~:hyperfiddle.ide\/foo\",\"page\"],\"~:hydrate-id\",null,\"~:error\",null,\"~:local-basis\",[\"^ \",\"^J\",2407,\"^3\",18827,\"^4\",148253,\"^1\",3348,\"^2\",1133],\"~:popovers\",[\"^6\",[]]]],\"~:hyperfiddle.runtime\/fatal-error\",null,\"~:hyperfiddle.runtime\/global-basis\",[\"^ \",\"~:domain\",[\"^ \",\"^J\",2407],\"~:ide\",[\"^ \"],\"~:user\",[\"^ \",\"^2\",1133,\"^4\",148253,\"^1\",3348,\"^3\",18827]],\"~:display-mode\",\"~:hypercrud.browser.browser-ui\/user\",\"~:hyperfiddle.runtime\/user\",[\"^ \"],\"~:hyperfiddle.runtime\/user-id\",null]<\/pre>","created_at":"2019-02-22T17:04:05+0000","updated_at":"2019-02-22T17:04:05+0000","annotations":[],"mimetype":"text\/plain","reading_time":119,"domain_name":"tank.hyperfiddle.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":852,"title":"Sharing files and directories in Ubuntu Server","url":"https:\/\/idiallo.com\/blog\/file-sharing-on-ubuntu-web-server?rel=159","content":"<p>If you read my post about <a href=\"http:\/\/idiallo.com\/blog\/creating-your-own-cdn-with-nginx\">creating your own CDN<\/a> you must have wondered how to share files across servers. I use one server to host my code and another to host my image files or assets. This could be problematic for most simple set ups because you have to upload your files into a different server.<\/p><p>The solution is to use shared folders. You create a folder on your CDN server, and make it accessible through the application server. It's actually very easy.<\/p>\n<p>You need NFS (network file system). The CDN will have the server application and our code server will have the client application. <\/p>\n<h3>Server (CDN)<\/h3>\n<p>Install the nfs server program:<\/p>\n<pre>sudo apt-get install nfs-kernel-server portmap\n<\/pre>\n<p>Half the work is already complete. All you have to do now is setup the directories you want to share. <\/p>\n<p>Let's say the folder we want to share is called <code>\/var\/assets<\/code>. In my case I didn't have to change the ownership but if it fails to share, do change the owner to <code>nobody<\/code> and the group to <code>nogroup<\/code>.<\/p>\n<pre>sudo chown nobody:nogroup \/var\/assets\n<\/pre>\n<p>Now you folder is ready to be shared. We can add an entry inside <code>\/etc\/exports<\/code> to let the machine know that this folder is to be exported.<\/p>\n<pre>sudo vim \/etc\/exports\n<\/pre>\n<p>You can find your machine IP address using the command <code>ifconfig<\/code> right after <code>inet addr:&lt;ipaddress&gt;<\/code>. Let's assume <code>10.0.0.2<\/code> for our CDN server and <code>10.0.0.1<\/code> for our Application server.<\/p>\n<p>Inside the export file add this entry:<\/p>\n<pre>\/var\/assets       10.0.0.1(rw,sync,no_root_squash,no_subtree_check)\n<\/pre>\n<p>Here we are telling our server that a client with ip <code>10.0.0.1<\/code> will have access to this folder.<\/p>\n<p>Save and confirm these entries using the <code>exportfs<\/code> command.<\/p>\n<pre>exportfs -a\n<\/pre>\n<p>Now our server application is ready to broadcast the shared folders to the client. Now let's setup the client machine.<\/p>\n<h3>Client (Application)<\/h3>\n<p>Install the nfs client program:<\/p>\n<pre>sudo apt-get install nfs-common portmap\n<\/pre>\n<p>All we need to do now is create the directories where we want the shared folders to appear. The standard is to have the directory in the <code>\/mnt<\/code> directory since what we are doing is called mounting technically.<\/p>\n<p>Let's create the folders:<\/p>\n<pre>mkdir -p \/mnt\/nfs\/var\/assets\n<\/pre>\n<p>Now we can mount our shared folder at this location:<\/p>\n<pre>mount 10.0.0.2:\/var\/assets \/mnt\/nfs\/var\/assets\n<\/pre>\n<p>That's it! Note that this is the IP of our CDN server. You can run the <code>df<\/code> command to ensure that it works correctly. <\/p>\n<pre>df -h\nFilesystem             Size  Used Avail Use% Mounted on\n\/dev\/sda                20G  120M   29G   5% \/\nudev                   119M  4.0K  102M   1% \/dev\ntmpfs                   49M  405K   49M   1% \/run\nnone                   6.0M     0  6.0M   0% \/run\/lock\n10.0.0.2:\/var\/assets    20G  948M   19G   5% \/mnt\/nfs\/var\/assets\n<\/pre>\n<p>You can also create a file from the client and check if it is available in the server.<\/p>\n<pre>echo \"hello world\" &gt; \/mnt\/nfs\/var\/assets\/test.txt\n<\/pre>\n<p>One thing I like to do is create a symlink of the shared folder this way I don't have to type all this path in my application when uploading a file.<\/p>\n<pre>ln -s \/mnt\/nfs\/var\/assets \/var\/www\/assets\n<\/pre>\n<p>This makes it part of my web application without really being there.<\/p>\n<p>Also if you want to unmount a folder all you need is use the <code>umount<\/code> command.<\/p>\n<pre>sudo umount \/mnt\/nfs\/var\/assets\n<\/pre>\n<p>Last but not least, if your client server were to be rebooted for some reason, you would lose the connection. You would have to re-mount the folder. You can solve this by making sure that your shared folder is mounted at start time.<\/p>\n<p>Add this entry inside <code>\/etc\/fstab<\/code><\/p>\n<pre>10.0.0.2:\/var\/assets \/mnt\/nfs\/var\/assets   nfs     auto,noatime,nolock,bg,nfsvers=3,intr,tcp,actimeo=1800 0 0\n<\/pre>\n<p>That's it now you know you can enjoy uploading files from your we application and serving those files through your CDN.<\/p>\n<p><strong>Ressources:<\/strong><\/p>\n<hr class=\"article__end\" \/><aside class=\"prev-next\"><div class=\"prev-next__right\">\n\t\t\tNext: <a class=\"prev-next__link\" href=\"https:\/\/idiallo.com\/blog\/endeavor\">Endeavor<\/a>\n\t\t<\/div>\n\t\t\t\t\t\t<div class=\"prev-next__left\">\n\t\t\tPrevious: <a class=\"prev-next__link\" href=\"https:\/\/idiallo.com\/blog\/revisiting-old-code\">Revisiting old code<\/a>\n\t\t<\/div>\n\t\t\t\t\n\t<\/aside>","created_at":"2019-02-21T16:03:45+0000","updated_at":"2019-02-21T16:03:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"idiallo.com","preview_picture":"http:\/\/cdn.idiallo.com\/images\/assets\/201\/thumb.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":851,"title":"EmptyEpsilon - Multiplayer Spaceship Bridge Simulator","url":"https:\/\/daid.github.io\/EmptyEpsilon\/","content":"<ul><li><a href=\"#scripting-tabs-0\">0: Tools &amp; basics<\/a><\/li>\n\t\t\t\t<li><a href=\"#scripting-tabs-1\">1: Building a universe<\/a><\/li>\n\t\t\t\t<li><a href=\"#scripting-tabs-2\">2: Death &amp; destruction<\/a><\/li>\n\t\t\t\t<li><a href=\"#scripting-tabs-3\">3: We are on a mission!<\/a><\/li>\n\t\t\t<\/ul><div id=\"scripting-tabs-0\"><h2>About this series<\/h2><p>This series documents the development of the EmptyEpsilon scenario <b>Beacon of Light<\/b>, which is included with the game.<\/p><p>EmptyEpsilon uses scenarios scripted in the <a href=\"http:\/\/www.lua.org\">Lua scripting language<\/a>, which allows for many types of scenarios: basic sandboxes, story-driven campaigns, randomized enemies, and even dynamic branching dialogue. These scenarios are located in the \"scripts\" folder where EmptyEpsilon is installed.<\/p><p>The Beacon of Light scenario begins with instructions for the players to follow, then a series of challenges for them to overcome.<\/p><p>The scenario scripting system can be used to create even more complex scenarios, but this guide won't cover such features. All of the scenarios are viewable and editable, however, and you can look at the \"Waves\" scenario in the scripts folder to see an example of complex randomized events. If you downloaded the official EmptyEpsilon build or compiled it yourself, it also comes with a scripting guide that lists the available functions.<\/p><p>You can follow the first parts of this guide even if you have little or no programming knowledge. Later sections cover more advanced topics.<\/p><h2>Tools<\/h2><p>For scenario development, you need a basic text editor. While Windows Notepad can work, I highly recommend something more advanced.<\/p><p>There are many options, including these free editors:<\/p><ul><li><a href=\"http:\/\/www.pnotepad.org\/\">Programmer's Notepad<\/a> (Windows)<\/li>\n\t\t\t\t\t<li><a href=\"http:\/\/www.sublimetext.com\/\">Sublime Text<\/a> (Windows, OS X, Linux)<\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/atom.io\/\">Atom<\/a> (Windows, OS X, Linux)<\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/notepad-plus-plus.org\/\">Notepad++<\/a> (Windows)<\/li>\n\t\t\t\t<\/ul><p>I use Programmer's Notepad, but any of these editors will do.<\/p><h2>1. Setting up the scenario file<\/h2><p>We'll start with an empty scenario as a template.<\/p><p>Open your text editor and enter the following code:<\/p><p class=\"code\">\n-- Name: Beacon of light series\n-- Description: The beacon of light scenario, build from the series at EmptyEpsilon.org.\n--- Near the far outpost of Orion-5, Exuari attacks are increasing. A diplomat went missing, and your mission will start with recovering him.\n-- Type: Mission\n-- Init is run when the scenario is started. Create your initial world\nfunction init()\n    -- Create the main ship for the players.\n    player = PlayerSpaceship():setFaction(\"Human Navy\"):setTemplate(\"Atlantis\")\nend\nfunction update(delta)\nend\n\t\t\t\t<\/p><p>Save this file as <b>scenario_99_tutorial.lua<\/b> in the folder <b>EmptyEpsilon\/scripts<\/b>.<\/p><p>Open EmptyEpsilon and start a new server. Your scenario should be at the bottom of the list. Select the scenario and start it, and notice that there a ship for the players is already present.<\/p><h2>Tips &amp; Tricks<\/h2><ul><li><p>The file must start with <b>scenario_<\/b> and must be located in the <b>scripts<\/b> folder for EmptyEpsilon to recognize it as a scenario.<\/p><\/li>\n\t\t\t\t    <li><p>The number in the scenario filename sets the order in which EmptyEpsilon shows the scenarios. The final part of the filename is for reference only.<\/p><\/li>\n\t\t\t\t    <li><p>The scenario name and description displayed in the game are set by the <b>-- Name:<\/b> and <b>-- Description:<\/b> lines at the beginning of the scenario file. For multi-line descriptions, each line should immediately follow the previous line and start with three dashes ---.<\/p><\/li>\n\t\t\t\t    <li><p>Other lines starting with a double dash -- are comment lines in the scenario script that you can use for your own reference. The game ignores them.<\/p><\/li>\n\t\t\t\t    <li><p>To quickly examine the scenario, you can use the Game Master screen listed with the Alternative Options on the ship selection menu. This lets you view and modify the entire game world with ease.<\/p><\/li>\n\t\t\t\t<\/ul><\/div><div id=\"scripting-tabs-1\"><h2>Building a universe<\/h2><p>Now that you have setup a basic script and know how to save and run it, let's populate the universe a bit more.<\/p><p>First, let's fill it with some stations, and then add some nebulae and asteroids.<\/p><h2>Adding stations<\/h2><p>The only thing in our universe is the player ship. This doesn't make for a very interesting universe, so let's add some space stations that they can dock and interact with.<\/p><p>When looking at your scenario from the Game Master screen, the starting point is located at the top left corner of sector F5. These coordinates are \"0, 0\", with the X coordinate first and the Y coordinate second. If you click on the player ship on the Game Master screen, you can see its coordinates in the top left corner.<\/p><p>Note as you move up\/north\/on heading 0, your Y coordinate <b>decreases<\/b>; locations north of the starting point have a <b>negative<\/b> Y coordinate. Moving right\/east\/on heading 90 <b>increases<\/b> your X coordinate. Also note that in-game distances are measured in units (U), while scripts measure distances in milliunits -- 1U in the game is equal to 1000 in a script. See <a href=\"https:\/\/daid.github.io\/EmptyEpsilon\/info_images\/sector_positions.png\">this image<\/a> for an example.<\/p><p>Add the following code after the creation of the player ship:<\/p><p class=\"code\">\nSpaceStation():setTemplate(\"Small Station\"):setFaction(\"Human Navy\"):setPosition(23500, 16100)\n\t\t\t\t<\/p><p>This line places a small Human Navy station 23.5U (23500) to the right of the starting point at sector F5, and 16.1U (16100) down from the starting point. Since each sector is a 20U square, this places the station in sector F6. The player ship also belongs to the Human Navy faction, so the station will be friendly to the players.<\/p><p>Next, add these lines to add three more stations:<\/p><p class=\"code\">\nSpaceStation():setTemplate(\"Medium Station\"):setFaction(\"Human Navy\"):setPosition(-25200, 32200)\nSpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\"):setPosition(-45600, -15800)\nSpaceStation():setTemplate(\"Small Station\"):setFaction(\"Independent\"):setPosition(9100,-35400)\n\t\t\t\t<\/p><p>This adds another friendly station in sector G3 (25.2U left and 32.2U down from the starting point), a large enemy Exuari station in sector E2 (45.6U left and 15.8U up from the starting point), and a small neutral station in sector D5 (9.1U right and 35.4U up from the starting point).<\/p><p>Now, we want the player to start closer to the small friendly station. So we set the player position near the small station.<\/p><p class=\"code\">\nPlayerSpaceship():setFaction(\"Human Navy\"):setShipTemplate(\"Atlantis\"):setPosition(22400, 16200)\n\t\t\t\t<\/p><h2>Creating nebulae<\/h2><p>Nebulae are interesting additions to a universe. They block a 5U circle on long-range radar as well as the space behind it, which gives players and enemies room to hide and provide interesting tactical options to the game.<\/p><p>Let's surround the enemy station with some nebulae so that it's shielded from sensors.<\/p><p class=\"code\">\n--Nebula that hide the enemy station.\nNebula():setPosition(-43300,  2200)\nNebula():setPosition(-34000,  -700)\nNebula():setPosition(-32000,-10000)\nNebula():setPosition(-24000,-14300)\nNebula():setPosition(-28600,-21900)\n\t\t\t\t<\/p><p>Such a cloud of nebulae clearly hides something, so let's also add a few random nebulae.<\/p><p class=\"code\">\n--Random nebulae in the system\nNebula():setPosition( -8000,-38300)\nNebula():setPosition( 24000,-30700)\nNebula():setPosition( 42300,  3100)\nNebula():setPosition( 49200, 10700)\nNebula():setPosition(  3750, 31250)\nNebula():setPosition(-39500, 18700)\n\t\t\t\t<\/p><h2>Creating asteroids<\/h2><p>You can add an asteroid in a similar manner as nebulae and stations:<\/p><p class=\"code\">\nAsteroid():setPosition(-1000, -1000)\n\t\t\t\t<\/p><p>However, because you usually want belts of 50, or even 100, asteroids, placing them all by hand would take quite a bit of time!<\/p><p>Fortunately, we can use Lua functions to help add many asteroids quickly. First, let's use the <b>random()<\/b> function to place an asteroid in a random location of sector E5. Because sector E5 is one sector to the left of sector F5, the range is from 0 to -20000 on the X axis and 0 to 2000 on the Y axis.<\/p><p class=\"code\">\nAsteroid():setPosition(random(0, 20000), random(-20000, 0))\n\t\t\t\t<\/p><p>Now let's use a loop to repeat this 100 times..<\/p><p class=\"code\">\n--Create 100 asteroids\nfor asteroid_counter=1,100 do\n    Asteroid():setPosition(random(0, 20000), random(-20000, 0))\nend\n\t\t\t\t<\/p><p>This is a <b>for loop<\/b>. It counts from 1 to 100 and performs the action inside once for each count. In this case, it adds an asteroid to a random position in sector E5.<\/p><p>Which... results in a pretty square asteroid field. Not that pretty.<\/p><p>So right now, let's tweak the parameters to create a field from sector E4 to G4 in a straight line.<\/p><p class=\"code\">\n--Create 50 asteroids\nfor asteroid_counter=1,50 do\n    Asteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\nend\n<\/p><p>That's already much better. But from a 3D view, this asteroid belt looks very flat. To fix that, add some VisualAsteroids as well. These asteroids are above or below the ship's plane and can't be hit. They're only there for visuals.<\/p><p class=\"code\">\n--Create 100 asteroids\nfor asteroid_counter=1,50 do\n    Asteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n    VisualAsteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\nend\n\t\t\t\t<\/p><h2>Spice it up a bit<\/h2><p>The stations have automatically assigned callsigns, such as DS-7 and other DS-? numbers. We can change these through the scenario scripts.<\/p><p>By calling the <b>setCallSign()<\/b> function on the stations, we can set a different name for each of them. Keep the callsign short.<\/p><p class=\"code\">\nSpaceStation():setTemplate(\"Small Station\"):setFaction(\"Human Navy\"):setPosition(23500, 16100):setCallSign(\"Research-1\")\nSpaceStation():setTemplate(\"Medium Station\"):setFaction(\"Human Navy\"):setPosition(-25200, 32200):setCallSign(\"Orion-5\")\nSpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\"):setPosition(-45600, -15800):setCallSign(\"Omega\")\nSpaceStation():setTemplate(\"Small Station\"):setFaction(\"Independent\"):setPosition(9100,-35400):setCallSign(\"Refugee-X\")\n\t\t\t\t<\/p><p>We'll call the small station a research station, which gives it a bit more character. The enemy station is called Omega, as it's big. And the neutral station is called Refugee-X, which makes it sound like it could house refugees. These few simple names give the whole universe more character.<\/p><p>Finally, let's christen the player ship the <b>Epsilon<\/b>.<\/p><p class=\"code\">\nPlayerSpaceship():setFaction(\"Human Navy\"):setShipTemplate(\"Atlantis\"):setPosition(22400, 18200):setCallSign(\"Epsilon\")\n\t\t\t\t<\/p><h2>Tips &amp; Tricks<\/h2><p>In addition to asteroids, stations, and nebulae, you also have other options to make your universe more exciting:<\/p><p class=\"code\">\nBlackHole() -- Creates a 5U-radius black hole that drags anything that comes toward its center and destroys it.\nWormHole():setTargetPosition(20000, -20000) -- Creates a 5U-radius wormhole that leads to another point in space.\nMine() -- A simple mine causes a 1U-radius explosion when a ship comes too close.\nWarpJammer():setFaction(\"Exuari\") -- A warp jammer, which jams warp and jump drives of other factions in a 5U radius. Can be destroyed.\nSupplyDrop():setFaction(\"Human Navy\") -- A supply drop, which a friendly player ship can pick up by flying over it.\n-- Supply drops are empty by default but can be filled with energy and missile weapons through additional script parameters.\n\t\t\t\t<\/p><h2>Compiled example<\/h2><p>This is the script so far:<\/p><p class=\"code\">\n-- Name: Beacon of light - Day 1\n-- Description: The beacon of light scenario, build from the series at EmptyEpsilon.org\n--- Near the far outpost of Orion-5, Exuari attacks are increasing. A diplomat went missing, and your mission will start with recovering him.\n-- Type: Mission\n-- Init is run when the scenario is started. Create your initial world\nfunction init()\n    -- Create the main ship for the players.\n    PlayerSpaceship():setFaction(\"Human Navy\"):setShipTemplate(\"Atlantis\"):setPosition(22400, 18200):setCallSign(\"Epsilon\")\n    SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Human Navy\"):setPosition(23500, 16100):setCallSign(\"Research-1\")\n    SpaceStation():setTemplate(\"Medium Station\"):setFaction(\"Human Navy\"):setPosition(-25200, 32200):setCallSign(\"Orion-5\")\n    SpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\"):setPosition(-45600, -15800):setCallSign(\"Omega\")\n    SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Independent\"):setPosition(9100,-35400):setCallSign(\"Refugee-X\")\n    --Nebulae that hide the enemy station.\n    Nebula():setPosition(-43300,  2200)\n    Nebula():setPosition(-34000,  -700)\n    Nebula():setPosition(-32000,-10000)\n    Nebula():setPosition(-24000,-14300)\n    Nebula():setPosition(-28600,-21900)\n    --Random nebulae in the system\n    Nebula():setPosition( -8000,-38300)\n    Nebula():setPosition( 24000,-30700)\n    Nebula():setPosition( 42300,  3100)\n    Nebula():setPosition( 49200, 10700)\n    Nebula():setPosition(  3750, 31250)\n    Nebula():setPosition(-39500, 18700)\n    --Create 100 asteroids\n    for asteroid_counter=1,50 do\n        Asteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n        VisualAsteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n    end\nend\nfunction update(delta)\nend\n\t\t\t    <\/p><\/div><div id=\"scripting-tabs-2\"><h2>Making friends and enemies<\/h2><p>Now we'll add some things that really spice things up: other ships.<\/p><h2>About other ships<\/h2><p>When working with other ships, remember a few basic things:<\/p><ul><li><b>Factions:<\/b> A ship's faction defines its relationship with other ships. For example, ships of the Exuari faction attack ships and stations of the Human Navy faction, and vice versa.<\/li>\n\t\t\t\t    <li><b>Orders:<\/b> Every computer-controlled ship has orders that define how the ship behaves. A ship can be ordered to stay still, attack a target, defend a location, or seek out targets on its own.<\/li>\n\t\t\t\t    <li><b>Templates:<\/b> Every ship's appearance, equipment, and base statistics are drawn from a template. A script can override nearly any trait in a template, except for the ship's model.<\/li>\n\t\t\t\t<\/ul><h2>Your first enemy<\/h2><p>Let's create an enemy.<\/p><p class=\"code\">\nCpuShip():setShipTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(1000, 1000)\n\t\t\t\t<\/p><p>Fight it! I'm sure you'll win -- notice that it doesn't do anything, not even respond to your attacks. This is because the default orders for a CpuShip are to stay idle and do nothing.<\/p><p>Let's give it one of the most basic orders: to seek new targets and attack them.<\/p><p class=\"code\">\nCpuShip():setShipTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(1000, 1000):orderRoaming()\n\t\t\t\t<\/p><p>Now you have an enemy to fight!<\/p><p>Let's start setting up the scenario's opposition with 2 cruisers at the lower nebula. They will fly out and seek the player quite well on their own.<\/p><p class=\"code\">\n--Small Exuari strike team, starting in the nebula at G5.\nCpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3550, 31250):orderRoaming()\nCpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3950, 31250):orderRoaming()\n\t\t\t\t<\/p><h2>Defending the station<\/h2><p>The Exuari have a large station in this area, so let's set up some defence.<\/p><p>We'll create 3 ships to defend Omega station. This means we need a way to refer to the station in our script. Edit the lines for Omega station to <b>assign<\/b> it to a <b>variable<\/b>.<\/p><p class=\"code\">\n-- Grab a reference to Omega station.\nenemy_station = SpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\")\nenemy_station:setPosition(-45600, -15800):setCallSign(\"Omega\")\n\t\t\t\t<\/p><p>Note how with a variable as a reference, we can move the setPosition() and setCallSign() functions to a new line. Let's also use this reference to assign new ships to its defense.<\/p><p class=\"code\">\n-- Create the defense for the station\nCpuShip():setTemplate(\"Starhammer II\"):setFaction(\"Exuari\"):setPosition(-44000, -14000):orderDefendTarget(enemy_station)\nCpuShip():setTemplate(\"Phobos T3\"):setFaction(\"Exuari\"):setPosition(-47000, -14000):orderDefendTarget(enemy_station)\nCpuShip():setTemplate(\"Atlantis X23\"):setFaction(\"Exuari\"):setPosition(-46000, -18000):orderDefendTarget(enemy_station)\n\t\t\t\t<\/p><p>With the DefendTarget orders, the ships will circle around the target and attack any enemies in range. You can set ships to defend any object, including another ship.<\/p><p>For example, we can setup two extra fighters which follow and defend the Atlantis X23.<\/p><p class=\"code\">\nenemy_dreadnought = CpuShip():setShipTemplate(\"Atlantis X23\"):setFaction(\"Exuari\")\nenemy_dreadnought:setPosition(-46000, -18000):orderDefendTarget(enemy_station)\nCpuShip():setShipTemplate(\"MT52 Hornet\"):setFaction(\"Exuari\"):setPosition(-46000, -18100):orderDefendTarget(enemy_dreadnought)\nCpuShip():setShipTemplate(\"MT52 Hornet\"):setFaction(\"Exuari\"):setPosition(-46000, -18200):orderDefendTarget(enemy_dreadnought)<\/p><h2>You lose! (Or, you should be able to)<\/h2><p>When the player ship is destroyed, or when all enemies are destroyed, nothing happens. Let's fix this.<\/p><p>First, we want to end the game with defeat if we lose the player ship. We need to edit the lines that add the player ship to create a reference to it.<\/p><p class=\"code\">\nplayer = PlayerSpaceship():setFaction(\"Human Navy\"):setShipTemplate(\"Atlantis\")\nplayer:setPosition(22400, 18200):setCallSign(\"Epsilon\")<\/p><p>Next, we'll finally use the <b>update()<\/b> function that has been empty so far. This function is special: the game runs any code in the update function every game <b>tick<\/b>, or about 60 times per second. Let's use that to constantly check whether the player ship is destroyed, and if it is, then end the game.<\/p><p>The <b>isValid()<\/b> function checks an object and returns a value of either true or false: true if the object exists, false if it doesn't. With a <b>conditional<\/b> statement (also called an <b>if<\/b> statement), we can ask, \"Is the player ship not a valid object?\", and if the response is true (the player doesn't exist), we can tell the game to declare victory for the Exuari. Doing so in the update function asks that question repeatedly.<\/p><p class=\"code\">\nfunction update(delta)\n    --When the player ship is destroyed, call it a victory for the Exuari.\n    if not player:isValid() then\n        victory(\"Exuari\")\n    end\nend\n\t\t\t\t<\/p><p>Pretty simple, right? And naturally, you can setup victory for the player the same way by checking Omega station.<\/p><p class=\"code\">\n--When Omega station is destroyed, call it a victory for the Human Navy.\nif not enemy_station:isValid() then\n    victory(\"Human Navy\")\nend\n\t\t\t\t<\/p><h2>Tips &amp; Tricks - More about ships<\/h2><ul><li>For all possible ship templates, check the files starting with <b>shipTemplates_<\/b> in the <b>scripts<\/b> folder. These files contain all of the standard ship definitions.<\/li>\n\t\t\t\t    <li>In addition to orderRoaming and orderDefendTarget, CpuShips can also have these orders:\n\t\t\t\t    <ul><li><b>orderIdle():<\/b> Default state. Do not move or attack.<\/li>\n\t\t\t\t\t    <li><b>orderStandGround():<\/b> Hold this position, but attack enemies when they are in range. The ship will fire missiles as well as beam weapons.<\/li>\n\t\t\t\t\t    <li><b>orderDefendLocation(x, y):<\/b> Fly toward the given position and defend it from attacks.<\/li>\n\t\t\t\t\t    <li><b>orderFlyFormation(target, offset_x, offset_y):<\/b> Fly in formation with the target by keeping a certain offset. When enemies are near the target, engage them. Gives fleet behaviour.<\/li>\n\t\t\t\t\t    <li><b>orderFlyTowards(x, y):<\/b> Fly toward a target position and attack enemies when they come too close.<\/li>\n\t\t\t\t\t    <li><b>orderFlyTowardsBlind(x, y):<\/b> Fly toward a target position and ignore everything else.<\/li>\n\t\t\t\t\t    <li><b>orderAttack(target):<\/b> Attack a specific target.<\/li>\n\t\t\t\t\t    <li><b>orderDock(target):<\/b> Dock with a target. Can be used to make computer-controlled ships dock with stations.<\/li>\n\t\t\t\t    <\/ul><\/li>\n\t\t\t\t<\/ul><h2>Compiled example<\/h2><p>This is the script so far:<\/p><p class=\"code\">\n-- Name: Beacon of light series - Day 2\n-- Description: The beacon of light scenario, build from the series at EmptyEpsilon.org.\n--- Near the far outpost of Orion-5, Exuari attacks are increasing. A diplomat went missing, and your mission will start with recovering him.\n-- Type: Mission\n    --Create the main ship for the players.\n    player = PlayerSpaceship():setFaction(\"Human Navy\"):setTemplate(\"Atlantis\")\n    player:setPosition(22400, 18200):setCallSign(\"Epsilon\")\n    SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Human Navy\"):setPosition(23500, 16100):setCallSign(\"Research-1\")\n    SpaceStation():setTemplate(\"Medium Station\"):setFaction(\"Human Navy\"):setPosition(-25200, 32200):setCallSign(\"Orion-5\")\n    --Grab a reference to Omega station.\n    enemy_station = SpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\")\n    enemy_station:setPosition(-45600, -15800):setCallSign(\"Omega\")\n    SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Independent\"):setPosition(9100,-35400):setCallSign(\"Refugee-X\")\n    --Nebula that hide the enemy station.\n    Nebula():setPosition(-43300,  2200)\n    Nebula():setPosition(-34000,  -700)\n    Nebula():setPosition(-32000,-10000)\n    Nebula():setPosition(-24000,-14300)\n    Nebula():setPosition(-28600,-21900)\n    --Random nebulae in the system.\n    Nebula():setPosition( -8000,-38300)\n    Nebula():setPosition( 24000,-30700)\n    Nebula():setPosition( 42300,  3100)\n    Nebula():setPosition( 49200, 10700)\n    Nebula():setPosition(  3750, 31250)\n    Nebula():setPosition(-39500, 18700)\n    --Create 100 asteroids.\n    for asteroid_counter=1,50 do\n        Asteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n        VisualAsteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n    end\n    --Create the defense for the station.\n    CpuShip():setTemplate(\"Starhammer II\"):setFaction(\"Exuari\"):setPosition(-44000, -14000):orderDefendTarget(enemy_station)\n    CpuShip():setTemplate(\"Phobos T3\"):setFaction(\"Exuari\"):setPosition(-47000, -14000):orderDefendTarget(enemy_station)\n    enemy_dreadnought = CpuShip():setTemplate(\"Atlantis X23\"):setFaction(\"Exuari\")\n    enemy_dreadnought:setPosition(-46000, -18000):orderDefendTarget(enemy_station)\n    CpuShip():setTemplate(\"MT52 Hornet\"):setFaction(\"Exuari\"):setPosition(-46000, -18100):orderDefendTarget(enemy_dreadnought)\n    CpuShip():setTemplate(\"MT52 Hornet\"):setFaction(\"Exuari\"):setPosition(-46000, -18200):orderDefendTarget(enemy_dreadnought)\n    --Small Exuari strike team in the nebula at G5.\n    CpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3550, 31250)\n    CpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3950, 31250)\nend\nfunction update(delta)\n    --When the player ship is destroyed, call it a victory for the Exuari.\n    if not player:isValid() then\n        victory(\"Exuari\")\n    end\n    --When Omega station is destroyed, call it a victory for the Human Navy.\n    if not enemy_station:isValid() then\n        victory(\"Human Navy\")\n    end\nend\n\t\t\t\t<\/p><\/div><div id=\"scripting-tabs-3\"><h2>Ready for your orders?<\/h2><p>Now that we know how to create a basic universe and fill it with some enemies, let's start creating a mission.<\/p><p>To kick off this mission, we'll send a transmission from <b>Research-1<\/b> to the player, indicating that a transport ship carrying a diplomat was lost en route from Research-1 station to Orion-X.<\/p><p>This involves more code, and we'll explain it in less detail than the previous steps. Keep the Lua reference and script reference handy in case you need help.<\/p><h2>Your first order!<\/h2><p>Start by creating references for the rest of the stations, as we'll need them later on.<\/p><p class=\"code\">\nresearch_station = SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Human Navy\")\nresearch_station:setPosition(23500, 16100):setCallSign(\"Research-1\")\nmain_station = SpaceStation():setTemplate(\"Medium Station\"):setFaction(\"Human Navy\")\nmain_station:setPosition(-25200, 32200):setCallSign(\"Orion-5\")\nenemy_station = SpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\")\nenemy_station:setPosition(-45600, -15800):setCallSign(\"Omega\")\nneutral_station = SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Independent\")\nneutral_station:setPosition(9100,-35400):setCallSign(\"Refugee-X\")\n\t\t\t\t<\/p><p>At the end of the init function, send a transmission to the player by using the <b>sendCommsMessage()<\/b> function. Note how we use double brackets [[ to wrap multiple lines of text.<\/p><p class=\"code\">\n--Start off the mission by sending a transmission to the player\nresearch_station:sendCommsMessage(player, [[Epsilon, please come in?\nWe lost contact with our transport RT-4, who was transporting a diplomat from our research station to Orion-X.\nLast contact was before RT-4 entered the nebula at G5.\nPlease investigate and recover the diplomat if possible!]])\n\t\t\t\t<\/p><h2>Improving the enemy strike team<\/h2><p>We still have the Exuari strike team set up in the nebula at G5. However, they have roaming orders and will fly out of the nebula to engage enemies. Instead, we want the strike team to hide in the nebula. We also need a transport (designated RT-4, as in our comms message) in the nebula.<\/p><p>By default, friendly ships have a basic <b>comms script<\/b> (comms_ship.lua in the scripts folder) that responds to player hails and takes basic orders. We don't want RT-4 to respond, so we can set the comms script to nothing (\"\").<\/p><p class=\"code\">\ntransport_RT4 = CpuShip():setTemplate(\"Flavia\"):setFaction(\"Human Navy\"):setPosition(3750, 31250)\ntransport_RT4:orderIdle():setCallSign(\"RT-4\"):setCommsScript(\"\")\n\t\t\t\t<\/p><p>We want the Exuari strike team to destroy RT-4 in front of the players' eyes, so we assign them references and set their orders to Idle so they won't move until ordered. We also set RT-4's hull and shields to 1 so that it can be destroyed in a single shot.<\/p><p class=\"code\">\ntransport_RT4:setHull(1):setShieldsMax(1, 1)\n--Small Exuari strike team, guarding RT-4 in the nebula at G5.\nexuari_RT4_guard1 = CpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3550, 31250):setRotation(0)\nexuari_RT4_guard2 = CpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3950, 31250):setRotation(180)\nexuari_RT4_guard1:orderIdle()\nexuari_RT4_guard2:orderIdle()\n\t\t\t\t<\/p><p>Next, we'll script the mission itself. First, set the initial mission state at the end of the init() function. We'll do that with a variable to track the mission's current state, and we'll increase it each time the players trigger a new event.<\/p><p class=\"code\">\n--Set the initial mission state\nmission_state = 1\n\t\t\t\t<\/p><p>Change the update() function. We still want the mission to end with an Exuari victory when the player is destroyed, and we also want the Exuari to win if Research-1 is destroyed. For now, however, remove the Human Navy victory condition.<\/p><p class=\"code\">\nfunction update(delta)\n    --When the player ship or research station is destroyed, call it a victory for the Exuari.\n    if not player:isValid() or not research_station:isValid() then\n        victory(\"Exuari\")\n    end\n\t\t\t\t<\/p><p>Now we'll set up what happens during our initial mission state of 1. When the player gets within 5U of transport RT-5, order the Exuari ships to roam around. This causes them to target and destroy RT-4, then target the player ship.<\/p><p class=\"code\">\n    if mission_state == 1 then\n        if distance(player, transport_RT4) &lt; 5000 then\n            exuari_RT4_guard1:orderRoaming()\n            exuari_RT4_guard2:orderRoaming()\n            mission_state = 2\n        end\n    end\nend\n\t\t\t\t<\/p><p>Finally, outside of the update() function, add a new utility function to measure the distance between two objects. This function uses the <b>getPosition()<\/b> function to get an object's X and Y coordinates. The <b>local<\/b> keyword is a Lua feature that prevents these variables from existing outside of this function, which is safer.<\/p><p class=\"code\">\nfunction distance(obj1, obj2)\n    local x1, y1 = obj1:getPosition()\n    local x2, y2 = obj2:getPosition()\n    local xd, yd = (x1 - x2), (y1 - y2)\n    return math.sqrt(xd * xd + yd * yd)\nend\n\t\t\t\t<\/p><h2>Get the Exuari scum! (Oh, and save the diplomat.)<\/h2><p>We can quite safely assume that the Exuari will destroy RT-4 during mission_state 2. In this new state, we'll wait until RT-4 is destroyed to spawn an empty supply drop as an escape pod. We'll also transmit a message to the Epsilon to tell the players what to do. Back in the update() function, add the following:<\/p><p class=\"code\">\nif mission_state == 2 then\n    if not transport_RT4:isValid() then\n        -- RT-4 destroyed, send a transmission to the player, create a supply drop to indicate an escape pod.\n        mission_state = 3\n        transport_RT4_drop = SupplyDrop():setFaction(\"Human Navy\"):setPosition(3750, 31250)\n        research_station:sendCommsMessage(player, [[RT-4 has been destroyed! But an escape pod is ejected from the ship.\nLifesigns detected in the pod, please pick up the pod to see if the diplomat made it. His death would be a great blow to the peace negotiations in the region.\nAnd destroy those Exuari scum while you are at it!]])\nend\n\t\t\t\t<\/p><p>Once again, we increase mission_state to trigger the next event. We also create the supply drop and keep a reference to it, and we send new instructions to the player ship.<\/p><p>We want the escape pod to be picked up, but we also want the Exuari ships to be destroyed before we continue. After we pick up the escape pod, we want the players to deliver it to the station and finish the mission.<\/p><p class=\"code\">\nif mission_state == 3 then\n    if not transport_RT4_drop:isValid() and not exuari_RT4_guard1:isValid() and not exuari_RT4_guard2:isValid() then\n\t-- Escape pod picked up, all Exuari ships destroyed.\n\tmission_state = 4\n    research_station:sendCommsMessage(player, [[The diplomat is safely picked up! Thanks for that.\nPlease deliver the diplomat to Orion-5 in sector G3. Do this by docking with the station.]])\n    end\nend\n\t\t\t\t<\/p><p>With the diplomat aboard the Epsilon, we can set up the Human Navy victory condition: his safe delivery to Orion-5.<\/p><p class=\"code\">\nif mission_state == 4 then\n    if player:isDocked(main_station) then\n\t-- Docked and delivered the diplomat.\n\tvictory(\"Human Navy\")\n    end\nend\n\t\t\t\t<\/p><h2>Tips &amp; tricks<\/h2><p>There is a LOT you can do with mission scripts, even with the limited tools at your disposal. Mission progression doesn't have to be as linear as in this example, and you could create branching plotlines or randomized conditions. For example, you could make RT-4 survive the assault long enough for the players to have a chance of destroying the Exuari ships first, which could yield more rewards or lead to a different mission path.<\/p><p>If you want to try some things for yourself but aren't sure where to go from here, try setting up a retaliatory Exuari strike after rescuing the diplomat. Spawn some ships near the edge of the nebulae at Omega station and order them to attack the player ship. You can also send a transmission from the new enemy ships to the players to let them know why they are being attacked.<\/p><h2>Compiled example<\/h2><p>This is the script so far:<\/p><p class=\"code\">\n-- Name: Beacon of light series\n-- Description: The beacon of light scenario, build from the series at EmptyEpsilon.org.\n--- Near the far outpost of Orion-5, Exuari attacks are increasing. A diplomat went missing, and your mission will start with recovering him.\n-- Type: Mission\n-- Init is run when the scenario is started. Create your initial world\nfunction init()\n    --Create the main ship for the players.\n    player = PlayerSpaceship():setFaction(\"Human Navy\"):setTemplate(\"Atlantis\")\n    player:setPosition(22400, 18200):setCallSign(\"Epsilon\")\n    research_station = SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Human Navy\")\n    research_station:setPosition(23500, 16100):setCallSign(\"Research-1\")\n    main_station = SpaceStation():setTemplate(\"Medium Station\"):setFaction(\"Human Navy\")\n    main_station:setPosition(-25200, 32200):setCallSign(\"Orion-5\")\n    enemy_station = SpaceStation():setTemplate(\"Large Station\"):setFaction(\"Exuari\")\n    enemy_station:setPosition(-45600, -15800):setCallSign(\"Omega\")\n    neutral_station = SpaceStation():setTemplate(\"Small Station\"):setFaction(\"Independent\")\n    neutral_station:setPosition(9100,-35400):setCallSign(\"Refugee-X\")\n    \n    --Nebula that hide the enemy station.\n    Nebula():setPosition(-43300,  2200)\n    Nebula():setPosition(-34000,  -700)\n    Nebula():setPosition(-32000,-10000)\n    Nebula():setPosition(-24000,-14300)\n    Nebula():setPosition(-28600,-21900)\n    --Random nebulae in the system\n    Nebula():setPosition( -8000,-38300)\n    Nebula():setPosition( 24000,-30700)\n    Nebula():setPosition( 42300,  3100)\n    Nebula():setPosition( 49200, 10700)\n    Nebula():setPosition(  3750, 31250)\n    Nebula():setPosition(-39500, 18700)\n    --Create 100 asteroids.\n    for asteroid_counter=1,50 do\n        Asteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n        VisualAsteroid():setPosition(random(-10000, -5000), random(-10000, 30000))\n    end\n    -- Create the defense for the station\n    CpuShip():setTemplate(\"Starhammer II\"):setFaction(\"Exuari\"):setPosition(-44000, -14000):orderDefendTarget(enemy_station)\n    CpuShip():setTemplate(\"Phobos T3\"):setFaction(\"Exuari\"):setPosition(-47000, -14000):orderDefendTarget(enemy_station)\n    enemy_dreadnought = CpuShip():setTemplate(\"Atlantis X23\"):setFaction(\"Exuari\")\n    enemy_dreadnought:setPosition(-46000, -18000):orderDefendTarget(enemy_station)\n    CpuShip():setTemplate(\"MT52 Hornet\"):setFaction(\"Exuari\"):setPosition(-46000, -18100):orderDefendTarget(enemy_dreadnought)\n    CpuShip():setTemplate(\"MT52 Hornet\"):setFaction(\"Exuari\"):setPosition(-46000, -18200):orderDefendTarget(enemy_dreadnought)\n    transport_RT4 = CpuShip():setTemplate(\"Flavia\"):setFaction(\"Human Navy\"):setPosition(3750, 31250)\n    transport_RT4:orderIdle():setCallSign(\"RT-4\"):setCommsScript(\"\")\n    transport_RT4:setHull(1):setShieldsMax(1, 1)\n    --Small Exuari strike team, guarding RT-4 in the nebula at G5.\n    exuari_RT4_guard1 = CpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3550, 31250):setRotation(0)\n    exuari_RT4_guard2 = CpuShip():setTemplate(\"Adder MK5\"):setFaction(\"Exuari\"):setPosition(3950, 31250):setRotation(180)\n    exuari_RT4_guard1:orderIdle()\n    exuari_RT4_guard2:orderIdle()\n    --Start off the mission by sending a transmission to the player\n    research_station:sendCommsMessage(player, [[Epsilon, please come in?\nWe lost contact with our transport RT-4, who was transporting a diplomat from our research station to Orion-X.\nLast contact was before RT-4 entered the nebula at G5.\nPlease investigate and recover the diplomat if possible!]])\n    --Set the initial mission state\n    mission_state = 1\nend\nfunction update(delta)\n    --When the player ship or the research station is destroyed, call it a victory for the Exuari.\n    if not player:isValid() or not research_station:isValid() then\n        victory(\"Exuari\")\n    end\n    if mission_state == 1 then\n        if distance(player, transport_RT4) &lt; 5000 then\n            exuari_RT4_guard1:orderRoaming()\n            exuari_RT4_guard2:orderRoaming()\n            mission_state = 2\n        end\n    end\n    if mission_state == 2 then\n        if not transport_RT4:isValid() then\n            -- RT-4 destroyed, send a transmission to the player, create a supply drop to indicate an escape pod.\n            mission_state = 3\n            transport_RT4_drop = SupplyDrop():setFaction(\"Human Navy\"):setPosition(3750, 31250)\n            research_station:sendCommsMessage(player, [[RT-4 has been destroyed! But an escape pod is ejected from the ship.\nLifesigns detected in the pod, please pick up the pod to see if the diplomat made it.\nAnd destroy those Exuari scum while you are at it!]])\n        end\n    end\n    if mission_state == 3 then\n        if not transport_RT4_drop:isValid() and not exuari_RT4_guard1:isValid() and not exuari_RT4_guard2:isValid() then\n            -- Escape pod picked up, all Exuari ships destroyed.\n            mission_state = 4\n            research_station:sendCommsMessage(player, [[The diplomat is safely picked up! Thanks for that.\nPlease deliver the diplomat to Orion-5 in sector G3. Do this by docking with the station.]])\n        end\n    end\n    if mission_state == 4 then\n        if player:isDocked(main_station) then\n            -- Docked and delivered the diplomat.\n            victory(\"Human Navy\")\n        end\n    end\nend\nfunction distance(obj1, obj2)\n    local x1, y1 = obj1:getPosition()\n    local x2, y2 = obj2:getPosition()\n    local xd, yd = (x1 - x2), (y1 - y2)\n    return math.sqrt(xd * xd + yd * yd)\nend\n\t\t\t    <\/p><\/div>","created_at":"2019-02-21T02:39:46+0000","updated_at":"2019-02-21T02:39:46+0000","annotations":[],"mimetype":"text\/html","reading_time":25,"domain_name":"daid.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":850,"title":"Globulation2","url":"https:\/\/globulation2.org\/wiki\/Main_Page","content":"<table id=\"language-bar\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td align=\"center\">\n<p><strong class=\"selflink\">English<\/strong> | \n<a href=\"https:\/\/globulation2.org\/wiki\/Fr:Main_Page\" title=\"Fr:Main Page\">Fran\u00e7ais<\/a> | \n<a href=\"https:\/\/globulation2.org\/wiki\/De:Main_Page\" title=\"De:Main Page\">Deutsch<\/a> | \n<a href=\"https:\/\/globulation2.org\/wiki\/Es:Main_Page\" title=\"Es:Main Page\">Espa\u00f1ol<\/a> | \n<a href=\"https:\/\/globulation2.org\/wiki\/It:Main_Page\" title=\"It:Main Page\">Italiano<\/a> |\n<a href=\"https:\/\/globulation2.org\/wiki\/Cs:Main_Page\" title=\"Cs:Main Page\">\u010cesky<\/a> |\n<a href=\"https:\/\/globulation2.org\/wiki\/Ro:Main_Page\" title=\"Ro:Main Page\">Rom\u00e2n\u0103<\/a> |\n<a href=\"https:\/\/globulation2.org\/index.php?title=Ru:Main_Page&amp;action=edit&amp;redlink=1\" class=\"new\" title=\"Ru:Main Page (page does not exist)\">\u0420\u0443\u0441\u0441\u043a\u0438\u0439<\/a> |\n<a href=\"https:\/\/globulation2.org\/wiki\/Ca:Main_Page\" title=\"Ca:Main Page\">Catal\u00e0<\/a> |\n<a href=\"https:\/\/globulation2.org\/wiki\/Fi:Main_Page\" title=\"Fi:Main Page\">Suomi<\/a> |\n<a href=\"https:\/\/globulation2.org\/wiki\/Ja:Main_Page\" title=\"Ja:Main Page\">Japanese<\/a> |\n<a href=\"https:\/\/globulation2.org\/index.php?title=Ru:Main_Page&amp;action=edit&amp;redlink=1\" class=\"new\" title=\"Ru:Main Page (page does not exist)\">\u0420\u0443\u0441\u0441\u043a\u0438\u0439<\/a> |\n<a href=\"https:\/\/globulation2.org\/wiki\/Zh:Main_Page\" title=\"Zh:Main Page\">Chinese<\/a>\n<\/p>\n<\/td><\/tr><\/table><div><a href=\"https:\/\/globulation2.org\/wiki\/File:Alpha23-attack.jpg\" class=\"image\"><img alt=\"Alpha23-attack.jpg\" src=\"https:\/\/globulation2.org\/wiki\/images\/thumb\/7\/71\/Alpha23-attack.jpg\/200px-Alpha23-attack.jpg\" width=\"200\" height=\"151\" srcset=\"\/wiki\/images\/thumb\/7\/71\/Alpha23-attack.jpg\/300px-Alpha23-attack.jpg 1.5x, \/wiki\/images\/thumb\/7\/71\/Alpha23-attack.jpg\/400px-Alpha23-attack.jpg 2x\" \/><\/a><\/div><p>Globulation 2 is an innovative Real-Time Strategy (RTS) game which reduces micro-management by automatically assigning tasks to units. Take a look at the <a href=\"https:\/\/globulation2.org\/wiki\/Screenshots\" title=\"Screenshots\">screenshots<\/a> or just <a href=\"https:\/\/globulation2.org\/wiki\/Download_and_Install\" title=\"Download and Install\">download<\/a> and try it right now. Globulation 2 is Free Software, licensed and distributed under the <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.gnu.org\/copyleft\/gpl.html\">GNU General Public License (version 3)<\/a>.\n<\/p><p>Globulation 2 brings a new type of gameplay to RTS games. The player chooses the number of units to assign to various tasks, and the units do their best to satisfy the requests. This allows players to manage more units and focus on strategy rather than on micro-management. Globulation 2 also <a href=\"https:\/\/globulation2.org\/wiki\/AI\" title=\"AI\">features AI<\/a>, allowing single-player games or any possible combination of human-computer teams. The game also includes a <a href=\"https:\/\/globulation2.org\/wiki\/Map_editor_guide\" title=\"Map editor guide\">scripting language<\/a> for versatile gameplay or tutorials and an integrated <a href=\"https:\/\/globulation2.org\/wiki\/Map_editor_guide\" title=\"Map editor guide\">map editor<\/a>. You can play Globulation 2 in single player mode, through your local network, or over the Internet with <a href=\"https:\/\/globulation2.org\/wiki\/YOG\" title=\"YOG\">Ysagoon Online Gaming<\/a> (or <i>YOG<\/i> for short). -- <a href=\"https:\/\/globulation2.org\/wiki\/Team\" title=\"Team\">Globulation 2 team<\/a>\n<\/p><table cellpadding=\"10\"><tr><td valign=\"top\">\n<h2>You Can Help!<\/h2>\n<p>Globulation 2 isn't just a bunch of code - it needs artwork, playable maps, campaigns and tutorials, sound effects and documentation to make a fun game. If you want to see your name in Globulation 2's credits, jump right in!\n<\/p>\n<dl><dt><a href=\"https:\/\/globulation2.org\/wiki\/Players\" title=\"Players\">Players<\/a><\/dt>\n<dd> The more Players, the more fun in Globulation 2 online games. <\/dd>\n<dt>Programmer<\/dt>\n<dd> There's a wide variety of topics that need contribution.  You name it - we need it.<br \/>If you want to program right away, you can do so by forking this code for example at <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/bitbucket.org\/giszmo\/glob2\">BitBucket<\/a> and ask for a merge once you think it is worth merging\u00a0;)<\/dd>\n<dt>Map and Campaign Designer<\/dt>\n<dd> To give the game a good selection of maps, and a new, longer, in depth campaign<\/dd>\n<dt>2D and 3D Artist<\/dt>\n<dd> To add campaign specific units, to do an opening cinematic.<\/dd>\n<dt><a href=\"https:\/\/globulation2.org\/wiki\/SoundFX\" title=\"SoundFX\" class=\"mw-redirect\">Sound Artist<\/a><\/dt>\n<dd> To do in game sound effects for events like attacking, moving, ...<\/dd>\n<dt>Tester<\/dt>\n<dd> The more testers; the better gameplay and the less bugs.<\/dd>\n<dt>Documenter<\/dt>\n<dd> We always welcome people who can write documentation on playing the game, to go along with the tutorial<\/dd>\n<dt>Translators<\/dt>\n<dd> We are also always in need of people who can translate the in game texts to new languages or maintain existing translations.<\/dd><\/dl><p>Should you be interested, the first steps would be to have a look on the <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.globulation2.org\/forums\">forums<\/a>, join the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/savannah.nongnu.org\/mail\/?group=glob2\">development Mailing list<\/a>, drop into the developers' <a href=\"https:\/\/globulation2.org\/wiki\/IRC\" title=\"IRC\">IRC<\/a> channel (#glob2-dev and #glob2 on irc.globulation2.org) and look around this site.\n<\/p><p>You also can help by simply <b>telling your friends about Globulation 2<\/b> or inform us of any game reporting site that does not list Globulation 2.\n<\/p>\n<\/td>\n<td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n<h2>News<\/h2>\n<dl><dt>23.04.2017<\/dt>\n<dd> As planned, the YOG server has just been moved to helium.progval.net as well. Users and maps databases have been preserved.<\/dd>\n<dt>05.03.2017<\/dt>\n<dd> The wiki and forums have been moved to a new server (helium.progval.net) and upgraded. YOG will follow in a near future.<\/dd>\n<dt>26.05.2013<\/dt>\n<dd> We had to delete a lot of wiki accounts because of spammers. If your account has been deleted, contact ProgVal on #glob2 to restore it.<\/dd>\n<dt>30.06.2012<\/dt>\n<dd> Forum, wiki and YOG server has been moved to <a href=\"https:\/\/globulation2.org\/index.php?title=User:ProgVal&amp;action=edit&amp;redlink=1\" class=\"new\" title=\"User:ProgVal (page does not exist)\">ProgVal<\/a>'s server (hydrogen.progval.net) and the YOG user database has been cleared.<\/dd><\/dl><dl><dd>More news in the <a href=\"https:\/\/globulation2.org\/wiki\/News_Archive\" title=\"News Archive\">news archive<\/a><\/dd><\/dl><\/td><\/tr><\/table>","created_at":"2019-02-21T02:39:00+0000","updated_at":"2019-02-21T02:39:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"globulation2.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":849,"title":"lisphp\/lisphp","url":"https:\/\/github.com\/lisphp\/lisphp","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"http:\/\/travis-ci.org\/lisphp\/lisphp\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d6aa062527d58e2544bbbfae3c657a2a0f8f53bf\/68747470733a2f2f7365637572652e7472617669732d63692e6f72672f6c69737068702f6c69737068702e706e673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/secure.travis-ci.org\/lisphp\/lisphp.png?branch=master\" \/><\/a><\/p>\n<p>Lisphp is a Lisp dialect written in PHP. It was created to be embedded in\nweb services or to be distributed within web applications. For that reason,\nit implements sandbox environment for security issues and multiple\nenvironment instances.<\/p>\n<h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<p>It requires PHP 5.3.0 or higher version. It also requires <a href=\"http:\/\/kr.php.net\/manual\/en\/book.spl.php\" rel=\"nofollow\">SPL<\/a> which is\navailable and compiled by default.<\/p>\n<h2><a id=\"user-content-standalone-command-line-interface\" class=\"anchor\" aria-hidden=\"true\" href=\"#standalone-command-line-interface\"><\/a>Standalone command line interface<\/h2>\n<p>There is <code>bin\/lisphp<\/code>, a standalone command line interface. It can take one\nparameter, the filename of the Lisphp program to be executed.<\/p>\n<pre>$ bin\/lisphp program.lisphp\n<\/pre>\n<p>You can run the program in sandbox with the option <code>-s<\/code>.<\/p>\n<pre>$ bin\/lisphp -s program.lisphp\n<\/pre>\n<h2><a id=\"user-content-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#repl\"><\/a>REPL<\/h2>\n<p>If there is no filename in arguments to <code>bin\/lisphp<\/code>, it enters REPL mode.<\/p>\n<pre>$ bin\/lisphp\n&gt;&gt;&gt; (form to evaluate)\n<\/pre>\n<p>Similarly you can specify <code>-s<\/code> to enter REPL mode in sandbox.<\/p>\n<pre>$ bin\/lisphp -s\n<\/pre>\n<ul><li><code>&gt;&gt;&gt;<\/code> is a prompt.<\/li>\n<li><code>==&gt;<\/code> is a returned value of evaluation.<\/li>\n<li><code>!!!<\/code> is a thrown exception.<\/li>\n<\/ul><h2><a id=\"user-content-simple-tutorial\" class=\"anchor\" aria-hidden=\"true\" href=\"#simple-tutorial\"><\/a>Simple tutorial<\/h2>\n<pre>&gt;&gt;&gt; (+ 12 34)\n==&gt; 46\n&gt;&gt;&gt; (- 1 2)\n==&gt; -1\n&gt;&gt;&gt; (* 5 6)\n==&gt; 30\n&gt;&gt;&gt; (\/ 30 5)\n==&gt; 6\n&gt;&gt;&gt; (\/ 30 4)\n==&gt; 7.5\n&gt;&gt;&gt; (% 30 4)\n==&gt; 2\n&gt;&gt;&gt; (. \"hello\" \"world\")\n==&gt; 'helloworld'\n&gt;&gt;&gt; (define pi 3.14)\n==&gt; 3.14\n&gt;&gt;&gt; pi\n==&gt; 3.14\n&gt;&gt;&gt; (float? pi)\n==&gt; true\n&gt;&gt;&gt; (string? \"abc\")\n==&gt; true\n&gt;&gt;&gt; (* pi 10 10)\n==&gt; 314\n<\/pre>\n<h2><a id=\"user-content-embed-in-your-app\" class=\"anchor\" aria-hidden=\"true\" href=\"#embed-in-your-app\"><\/a>Embed in your app<\/h2>\n<p>In order to execute a Lisphp program, an <em>environment instance<\/em> is required.\nEnvironment represents a global state for the program. It includes global\nsymbols, built-in functions and macros. A program to be executed starts from the\ninitialized environment. You can initialize the environment with\n<code>Lisphp_Environment<\/code> class.<\/p>\n<pre>require 'vendor\/autoload.php';\n$env = Lisphp_Environment::sandbox();\n$program = new Lisphp_Program($lisphpCode);\n$program-&gt;execute($env);\n<\/pre>\n<p>There are two given environment sets in <code>Lisphp_Environment<\/code>. One is the\nsandbox, which is created with the method <code>Lisphp_Environment::sandbox()<\/code>.\nIn the sandbox mode, programs cannot access the file system, IO, etc.\nThe other set is the full environment of Lisphp, which is initialized with\n<code>Lisphp_Environment::full()<\/code>. This environment provides <code>use<\/code> macro for\nimporting native PHP functions and classes. File system, IO, socket, etc.\ncan be accessed in this full environment. Following code touches file a.txt and\nwrites some text.<\/p>\n<pre>(use fopen fwrite fclose)\n{let ([fp (fopen \"a.txt\" \"w\")])\n     (fwrite fp \"some text\")\n     (fclose fp)}\n<\/pre>\n<h2><a id=\"user-content-macro-use-and-from\" class=\"anchor\" aria-hidden=\"true\" href=\"#macro-use-and-from\"><\/a>Macro <code>use<\/code> and <code>from<\/code><\/h2>\n<p>The full environment of Lisphp provides <code>use<\/code> macro. It can import native PHP\nfunctions and classes.<\/p>\n<pre>(use strrev array_sum array-product [substr substring])\n<\/pre>\n<p>It imports by taking function identifiers. Hyphens in identifiers are replaced\nby underscores. If you supply a list as an argument to <code>use<\/code> macro, the second\nsymbol becomes the alias of the first function identifier.<\/p>\n<pre>(strrev \"hello\")                #=&gt; \"olleh\"\n(array_sum [array 1 2 3])       #=&gt; 6\n(array-product [array 4 5 6])   #=&gt; 120\n(substring \"world\" 2)           #=&gt; \"rld\"\n<\/pre>\n<p>Wrap identifiers with angle brackets in order to import class. According to the\n<a href=\"http:\/\/pear.php.net\/manual\/en\/standards.naming.php\" rel=\"nofollow\">PEAR naming convention<\/a> for classes, slashes are treated as hierarchical\nseparators, so it gets replaced by underscores.<\/p>\n<pre>(use &lt;PDO&gt; Lisphp\/&lt;Program&gt;)\n<\/pre>\n<p>Imported classes are applicable and act as instantiating functions.\nStatic methods in imported classes are also imported as well.<\/p>\n<pre>(&lt;PDO&gt; \"mysql:dbname=testdb;host=127.0.0.1\" \"dbuser\" \"dbpass\")\n(Lisphp\/&lt;Program&gt;\/load \"program.lisphp\")\n<\/pre>\n<p>There's also a macro called <code>from<\/code>. It simplifies the step of importing objects\nand resolving their names.<\/p>\n<pre>(from Lisphp [&lt;Program&gt; &lt;Scope&gt;])\n<\/pre>\n<p>It has the same behavior as the following code which utilizes <code>use<\/code>.<\/p>\n<pre>(use Lisphp\/&lt;Program&gt; Lisphp\/&lt;Scope&gt;)\n(define &lt;Program&gt; Lisphp\/&lt;Program&gt;)\n(define &lt;Scope&gt; Lisphp\/&lt;Scope&gt;)\n(define Lisphp\/&lt;Program&gt; nil)\n(define Lisphp\/&lt;Scope&gt; nil)\n<\/pre>\n<h2><a id=\"user-content-define-custom-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#define-custom-functions\"><\/a>Define custom functions<\/h2>\n<p>There is a macro <code>lambda<\/code> that creates a new function. It takes parameter\nlist as its first argument, and then a trailing function body.<\/p>\n<pre>(lambda (a b) (+ a b))\n<\/pre>\n<p>Functions are also values, so in order to name the function, use <code>define<\/code>.<\/p>\n<pre>(define fibonacci\n        {lambda [n]\n                (if (= n 0) 0\n                    {if (\/= n 1)\n                        (+ (fibonacci (- n 1))\n                           (fibonacci (- n 2)))\n                        1})})\n<\/pre>\n<p>Following code defines the same function.<\/p>\n<pre>(define (fibonacci n)\n        (if (= n 0) 0\n            {if (\/= n 1)\n                (+ (fibonacci (- n 1))\n                   (fibonacci (- n 2)))\n                1}))\n<\/pre>\n<p>Function body can contain one or more forms. All forms are evaluated\nsequentially then the evaluated value of the last form is returned.<\/p>\n<p>Plus, of course, it implements lexical scope (that is also known as closure)\nalso.<\/p>\n<pre>(define (adder n)\n        {lambda [x]\n                (setf! n (+ n x))\n                n})\n<\/pre>\n<p>Special form <code>define<\/code> defines global variables (and functions),\nbut <code>setf!<\/code> modifies local variables. See also <code>let<\/code> form.<\/p>\n<h2><a id=\"user-content-define-custom-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#define-custom-macros\"><\/a>Define custom macros<\/h2>\n<p>The built-in macros in Lisphp such as <code>eval<\/code>, <code>define<\/code>, <code>lambda<\/code>, <code>let<\/code>, <code>if<\/code>,\n<code>and<\/code>, <code>or<\/code> do not evaluate the form of their arguments'. For example, <code>define<\/code>\ntakes the name to define as its first argument, but it does not evaluate the\nname. In the same way, <code>if<\/code> takes three forms as arguments, but always\nevaluates only two of those arguments and ignores the other. It is impossible\nto implement <code>if<\/code> as a function, because then every argument would have to be\nevaluated. If you have a case like this, you can try defining a <code>macro<\/code>.<\/p>\n<pre>(define if*\n        {macro [let {(cond (eval (car #arguments)\n                                 #scope))}\n                    (eval (at #arguments (or (and cond 1) 2))\n                          #scope)]})\n(define quote*\n        [macro (car #arguments)])\n<\/pre>\n<h2><a id=\"user-content-quote\" class=\"anchor\" aria-hidden=\"true\" href=\"#quote\"><\/a>Quote<\/h2>\n<p>There are two ways to quote a form in Lisphp. First is the macro <code>quote<\/code>, and\nthe other is quote syntax <code>:<\/code>. (You cannot use the traditional single quotation\nbecause it is already being used as a string literal.)<\/p>\n<pre>(quote abc)\n:abc\n(quote (+ a b))\n:(+ a b)\n<\/pre>\n<h2><a id=\"user-content-playing-with-objects\" class=\"anchor\" aria-hidden=\"true\" href=\"#playing-with-objects\"><\/a>Playing with objects<\/h2>\n<p>In order to get an attribute of an object, use <code>-&gt;<\/code> macro. It takes an object as\nits first argument, and the name of the attribute follows.<\/p>\n<pre>(use [dir &lt;dir&gt;])\n(define directory (&lt;dir&gt; \"\/tmp\"))\n(define handle (-&gt; directory handle))\n<\/pre>\n<p>There is also a syntactic sugar for object attribute chaining.<\/p>\n<pre>(-&gt; object attribute names go here)\n<\/pre>\n<p>This form is equivalent to the following PHP expression.<\/p>\n<pre>$object-&gt;attribute-&gt;names-&gt;go-&gt;here\n<\/pre>\n<p>Instance methods can also be invoked by <code>-&gt;<\/code>.<\/p>\n<pre>((-&gt; object method) method arguments)\n<\/pre>\n<p>This form is equivalent to the following PHP expression.<\/p>\n<pre>$object-&gt;method($method, $arguments)\n<\/pre>\n<p>Because <code>-&gt;<\/code> does not call but gets method as a function object, the expression\nabove is equivalent to the following code.<\/p>\n<pre>call_user_func(array($object, 'method'), $method, $arguments)\n<\/pre>\n<h2><a id=\"user-content-about-lists-and-nil\" class=\"anchor\" aria-hidden=\"true\" href=\"#about-lists-and-nil\"><\/a>About lists and <code>nil<\/code><\/h2>\n<p>Lisphp implements lists in primitive, but it has some differences between\noriginal Lisp. In original Lisp, lists are made from <a href=\"http:\/\/en.wikipedia.org\/wiki\/Cons\" rel=\"nofollow\">cons<\/a> pairs. But lists\nin Lisphp is just an instance of <code>Lisphp_List<\/code> class, a subclass of\n<a href=\"http:\/\/php.net\/manual\/en\/class.arrayobject.php\" rel=\"nofollow\"><code>ArrayObject<\/code><\/a>. So it is not exactly a linked list but is similar\nto an array. In the same manner, <code>nil<\/code> is  also not an empty list in Lisphp\nunlike in the original Lisp. It is a just synonym for PHP <a href=\"http:\/\/php.net\/manual\/en\/language.types.null.php\" rel=\"nofollow\"><code>null<\/code><\/a> value.<\/p>\n<h2><a id=\"user-content-about-value-types-and-reference-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#about-value-types-and-reference-types\"><\/a>About value types and reference types<\/h2>\n<p>In PHP, primitive types such as boolean, integer, float, string, and array\nbehave as value types. They are always copied when they are passed as arguments\nor returned from a called function. For example, <code>arr<\/code> is empty from the\nbeginning to the end in the following code.<\/p>\n<pre>(define arr (array))\n(set-at! arr \"element\")\n<\/pre>\n<p>Such behavior is not a problem for scalar types like boolean, integer, float,\nand string because they are immutable. Yet this can be problematic for native\narrays.<\/p>\n<p>In PHP, objects behave as reference types, and there exists class <code>ArrayObject<\/code>\nwhich has the same interface as PHP's native array. <code>Lisphp_List<\/code> is a subclass\nof <code>ArrayObject<\/code> and you can use these classes instead of arrays.<\/p>\n<pre>(use &lt;ArrayObject&gt;)\n(define arr (&lt;ArrayObject&gt;))\n(set-at! arr \"element\")\n(define lis (list))\n(set-at! arr \"element\")\n<\/pre>\n<h2><a id=\"user-content-mailing-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#mailing-list\"><\/a>Mailing list<\/h2>\n<p>There is the mailing list to discuss about Lisphp: <a href=\"mailto:lisphp@googlegroups.com\">lisphp@googlegroups.com<\/a>.<\/p>\n<p>The web archive for this mailing list: <a href=\"http:\/\/groups.google.com\/group\/lisphp\" rel=\"nofollow\">http:\/\/groups.google.com\/group\/lisphp<\/a>.<\/p>\n<h2><a id=\"user-content-author-and-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#author-and-license\"><\/a>Author and license<\/h2>\n<p>Lisphp was written by Hong Minhee <a href=\"http:\/\/dahlia.kr\/\" rel=\"nofollow\">http:\/\/dahlia.kr\/<\/a>.<\/p>\n<p>Lisphp is distributed under the MIT license.<\/p>\n<\/article>","created_at":"2019-02-20T23:08:08+0000","updated_at":"2019-02-20T23:08:08+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/2313007?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":848,"title":"andirueckel\/undefined-medium","url":"https:\/\/github.com\/andirueckel\/undefined-medium","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/undefined-medium.com\" rel=\"nofollow\">undefined medium<\/a> is a free and open-source pixel grid-based monospace typeface suitable for programming, writing, and whatever else you can think of \u2026 it\u2019s pretty undefined.<\/p>\n<p>It is inspired by many 5\u00d77 pixel grid typefaces, especially Gilles Boccon-Gibod\u2019s <a href=\"https:\/\/www.bok.net\/MonteCarlo\" rel=\"nofollow\">MonteCarlo<\/a>, which is unfortunately, since 2010, no longer maintained.<\/p>\n<p>The 400+ character set fully supports Latin Extended-A with Western European, Central European, and South Eastern European languages.<\/p>\n<p>In addition, there\u2019s many typographic symbols, mathematical symbols, superscripts, double-spaced fractions, a copyleft symbol, and a capital sharp s character.<\/p>\n<h2><a id=\"user-content-download\" class=\"anchor\" aria-hidden=\"true\" href=\"#download\"><\/a>download<\/h2>\n<p>undefined medium is available in OTF and TTF fonts for desktop use, as well as in WOFF and WOFF2 fonts for web projects.<\/p>\n<p>You can download the compiled fonts from the <a href=\"https:\/\/github.com\/andirueckel\/undefined-medium\/releases\">releases<\/a> section of this repository.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>license<\/h2>\n<p>This Font Software is licensed under the SIL Open Font License, Version 1.1.<\/p>\n<p>This license is stored in this repository\u2019s <a href=\"https:\/\/github.com\/andirueckel\/undefined-medium\/blob\/master\/OFL.txt\">OFL.txt<\/a>, and is also available with a FAQ at: <a href=\"https:\/\/scripts.sil.org\/OFL\" rel=\"nofollow\">https:\/\/scripts.sil.org\/OFL<\/a><\/p>\n<\/article>","created_at":"2019-02-20T22:22:27+0000","updated_at":"2019-02-20T22:22:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/46488911?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":847,"title":"Explaining Code using ASCII Art \u2013 Embedded in Academia","url":"https:\/\/blog.regehr.org\/archives\/1653","content":"<p>People tend to be visual: we use pictures to understand problems. Mainstream programming languages, on the other hand, operate in an almost completely different kind of abstract space, leaving a big gap between programs and pictures. This piece is about pictures drawn using a text character set and then embedded in source code. I love these! The other day I asked around on Twitter for more examples and <a href=\"https:\/\/twitter.com\/johnregehr\/status\/1095018518737637376\">the responses<\/a> far exceeded expectations (thanks everyone!). There are a ton of great examples in the thread; here I\u2019ve categorized a few of them. Click on images go to the repositories.<\/p>\n<h2>Data Structures<\/h2>\n<p>One of the most common kinds of ASCII art in code is illustrating the shape of a data structure.<\/p>\n<p>The example I started with comes from LLVM:<br \/><a href=\"https:\/\/t.co\/FGDDGsTOsW\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/llvm1.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>The layout of a data structure in the Jikes RVM:<br \/><a href=\"https:\/\/github.com\/JikesRVM\/JikesRVM\/blob\/master\/rvm\/src\/org\/jikesrvm\/objectmodel\/TIBLayoutConstants.java#L19\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/jikes.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>A tree rotate in Musl:<br \/><a href=\"https:\/\/git.musl-libc.org\/cgit\/musl\/tree\/src\/search\/tsearch.c?id=v1.1.21\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/musl.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Double-ended queue from the Rust library:<br \/><a href=\"https:\/\/github.com\/rust-lang\/rust\/blob\/57d7cfc3cf50f0c427ad3043ff09eaef20671320\/src\/liballoc\/collections\/vec_deque.rs#L1399\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/rust-deq.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Swift compiler internals:<br \/><a href=\"https:\/\/github.com\/apple\/swift\/blob\/master\/lib\/AST\/NameLookup.cpp#L1521\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/swift.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Malloc header layout:<br \/><a href=\"https:\/\/code.woboq.org\/userspace\/glibc\/malloc\/malloc.c.html#1075\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/malloc.jpg\" alt=\"image\" \/><\/a><\/p>\n<h2>State Machines<\/h2>\n<p>JavaScript profiling:<br \/><a href=\"https:\/\/searchfox.org\/mozilla-central\/rev\/00c0d068ece99717bea7475f7dc07e61f7f35984\/tools\/profiler\/core\/RegisteredThread.h#293-301\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/js-profile.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>RPCs in Cloud Spanner:<br \/><a href=\"https:\/\/github.com\/googleapis\/google-cloud-go\/blob\/master\/spanner\/read.go#L353\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/go-rpc.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>I\/O stream states:<br \/><a href=\"https:\/\/github.com\/carllerche\/h2\/blob\/8a0b7ff64f5fe26324ee9f540722549231ac1ad6\/src\/proto\/streams\/state.rs#L13-L50\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/stream.jpg\" alt=\"image\" \/><\/a><\/p>\n<h2>Logical Structure in the Problem Domain<\/h2>\n<p>Control flow in an NWScript program being decompiled:<br \/><a href=\"https:\/\/github.com\/xoreos\/xoreos-tools\/blob\/master\/src\/nwscript\/controlflow.cpp#L328\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/decompile.png\" alt=\"image\" \/><\/a><\/p>\n<p>ECC internals:<br \/><a href=\"https:\/\/github.com\/dalek-cryptography\/curve25519-dalek\/blob\/develop\/src\/backend\/vector\/avx2\/edwards.rs#L123\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/dalek.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Formatting numbers:<br \/><a href=\"https:\/\/searchfox.org\/mozilla-central\/source\/js\/src\/builtin\/intl\/NumberFormat.cpp#592\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/number.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>A quantum circuit:<br \/><a href=\"https:\/\/github.com\/Microsoft\/QuantumKatas\/blob\/master\/PhaseEstimation\/Tasks.qs#L117\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/quantum.png\" alt=\"image\" \/><\/a><\/p>\n<p>Balancing memory management objectives in an OS kernel:<br \/><a href=\"https:\/\/github.com\/torvalds\/linux\/blob\/v4.20\/mm\/page-writeback.c#L828\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/paging.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Subtyping relations (this is a very cool special case where the ASCII art is also code):<br \/><a href=\"https:\/\/docs.racket-lang.org\/2d\/index.html\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/racket.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>The format of a DBF file:<br \/><a href=\"https:\/\/github.com\/SocialExplorer\/FastDBF\/blob\/master\/FastDBF\/DbfHeader.cs#L35\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/dbf.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>A lookup table for image processing:<br \/><a href=\"https:\/\/github.com\/libvips\/libvips\/blob\/master\/libvips\/convolution\/sharpen.c#L412\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/sharpen.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Shape of a color function:<br \/><a href=\"https:\/\/github.com\/chriscalo\/design-system-utils\/blob\/master\/src\/color-scales.js\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/color.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Structure of a URI:<br \/><a href=\"https:\/\/searchfox.org\/mozilla-central\/rev\/00c0d068ece99717bea7475f7dc07e61f7f35984\/netwerk\/base\/nsIURI.idl#9-36\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/uri.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>A very quick tutorial on undo systems from emacs:<br \/><a href=\"https:\/\/github.com\/emacsmirror\/undo-tree\/blob\/master\/undo-tree.el#L242-L751\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/undo.jpg\" alt=\"image\" \/><\/a><\/p>\n<h2>Geometry<\/h2>\n<p>Attitude control in the Apollo Guidance Computer (!!!):<br \/><a href=\"https:\/\/github.com\/chrislgarry\/Apollo-11\/blob\/master\/Comanche055\/AUTOMATIC_MANEUVERS.agc#L193\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/apollo.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Image tiling:<br \/><a href=\"https:\/\/searchfox.org\/mozilla-central\/rev\/00c0d068ece99717bea7475f7dc07e61f7f35984\/gfx\/wr\/webrender\/src\/image.rs#247\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/image1.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Boomerang trajectories in Nethack:<br \/><a href=\"https:\/\/github.com\/NetHack\/NetHack\/blob\/1a8a774719abe55b5a57f63cc43f183339dc6433\/src\/zap.c#L3461-L3472\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/boomerang.png\" alt=\"image\" \/><\/a><\/p>\n<p>Rendering CSS borders:<br \/><a href=\"https:\/\/searchfox.org\/mozilla-central\/rev\/00c0d068ece99717bea7475f7dc07e61f7f35984\/layout\/painting\/nsCSSRenderingBorders.cpp#918-956\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/css.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Quadtrees:<br \/><a href=\"https:\/\/github.com\/libfive\/libfive\/blob\/master\/libfive\/src\/render\/brep\/simplex\/simplex_tree.cpp#L688-L717\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/quadtrees.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Speed control in a milling machine:<br \/><a href=\"https:\/\/github.com\/grbl\/grbl\/blob\/b237ad566ad25a1503e14fa5f487c34f88bec01f\/grbl\/stepper.c#L147\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/milling.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>Scrolling web pages:<br \/><a href=\"https:\/\/github.com\/stipsan\/compute-scroll-into-view\/blob\/d6447854d04031ee3942c6d5654a8477b6bb928f\/src\/index.ts#L124-L168\"><img src=\"https:\/\/blog.regehr.org\/extra_files\/ascii\/scrolling.jpg\" alt=\"image\" \/><\/a><\/p>\n<p>I hope you enjoyed these as much as I did!<\/p>","created_at":"2019-02-18T22:49:11+0000","updated_at":"2019-02-18T22:49:11+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"blog.regehr.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":846,"title":"How to deploy Clojure applications","url":"https:\/\/purelyfunctional.tv\/article\/jvm-deployment-options\/","content":"<p>If you're lucky, there comes a time in your Clojure app's life when it\nwill need to leave the comforts of the development environment and to\nopen its ports to the world. As a mature platform, the JVM has a lot\nof server deployment options. While options, in general, are good,\nthey are intimidating. Some of the deployment options are of the\n\"multimillion-dollar enterprise certification\" variety.<\/p><p>When you're working at a company, you usually inherit their deployment\nsystem. And that's great because then you just do what they do. But\nwhat if you are on your own? What are the options for deploying a\nClojure server?<\/p><h2>Executables<\/h2><p>Deployment options fall into two broad categories. The first is the\nsimplest: the <strong>application runs as an executable<\/strong>, just like other\nservers you're familiar with. You can use all of the operating system\nstuff you're used to, like turning it into a service that starts at\nboot.<\/p><p>This is how I run most of my apps. I compile an executable uberjar and\nuse something like <a href=\"http:\/\/supervisord.org\/\">supervisord<\/a> to keep them running.<\/p><h2>Containers<\/h2><p>The second option is a little more complex. You can actually deploy\nyour application to what's called a <em>container<\/em>. For instance, <a href=\"http:\/\/wildfly.org\/\">Wildfly<\/a> is an <em>application server container<\/em>. Tomcat is a <em>servlet\ncontainer<\/em>.<\/p><p>Containers do two things that make them important.<\/p><ol><li><p>They run <strong>multiple applications in a single JVM<\/strong>.<\/p><p>The container\nisolates the different applications so they can load different\nversions of the same class and can't access each other's\nmemory. You can write your application as if it were running in its\nown JVM, without the downsides of requiring a lot of memory per\napplication.<\/p><\/li><li><p><strong>They provide services<\/strong> such as web servers, database access, queues,\nand schedulers.<\/p><p>Your applications can talk to each other using the\nqueues, share database connections, and tap into high-performance\nweb servers.<\/p><\/li><\/ol><p>Now I'll go over some options and how they fit into this scheme. This\nis a non-exhaustive list of my recommendations.<\/p><h2>Heroku<\/h2><p><a href=\"https:\/\/devcenter.heroku.com\/articles\/clojure-support\">Heroku<\/a> works out of the box with Clojure. It can detect if you're\nusing Leiningen. It will build an uberjar and deploy it to their cloud\nas an executable. That means you'll want a built-in web server like <a href=\"http:\/\/www.eclipse.org\/jetty\/\">Jetty<\/a>, <a href=\"http:\/\/netty.io\/\">Netty<\/a>, <a href=\"http:\/\/www.http-kit.org\/\">http-kit<\/a>, or <a href=\"https:\/\/github.com\/ztellman\/aleph\">aleph<\/a>. There's a <a href=\"https:\/\/github.com\/boot-clj\/boot\/wiki\/Running-Boot-on-Heroku\">Boot buildpack<\/a>.<\/p><p>Heroku has a free tier and is affordable from there. You can also get <a href=\"https:\/\/elements.heroku.com\/addons\">other services<\/a> like queues,\ndatabases, email sending, etc. Deploying to Heroku is as easy as <code>git push heroku master<\/code>. It's very easy to get something running.<\/p><h2>AWS Elastic Beanstalk<\/h2><p>Amazon Web Services provides a container service called <a href=\"https:\/\/aws.amazon.com\/elasticbeanstalk\/\">Elastic Beanstalk<\/a>. You\nupload a WAR (Web Application Archive)--basically a JAR file with\ndeployment information in it--and AWS will deploy it to <a href=\"http:\/\/tomcat.apache.org\/\">Tomcat<\/a> for\nyou. With Elastic Beanstalk, you only pay for the AWS resources you\nuse. You're running on EC2 and you can connect to the other services\nthey provide.<\/p><p>A few years ago I used <a href=\"https:\/\/github.com\/ktgit\/lein-elastic-beanstalk\">lein-elastic-beanstalk<\/a> but it <a href=\"https:\/\/victorjcheng.wordpress.com\/2016\/02\/02\/deploying-a-clojure-app-using-elastic-beanstalk\/\">looks a little out of date now<\/a>. You\nmight want to try the <a href=\"https:\/\/aws.amazon.com\/cli\/\">AWS CLI<\/a>.<\/p><h2>Hosting on a server<\/h2><p>You can run your own server. This gives you complete control at the\ncost of having to maintain a server. Servers are not that easy! There\nare security concerns all the time.<\/p><p>I've had good luck running on Ubuntu using <a href=\"http:\/\/supervisord.org\/\">supervisord<\/a> for\nexecutable applications. People often use <a href=\"https:\/\/github.com\/systemd\/systemd\">systemd<\/a>. I've also used <a href=\"http:\/\/wildfly.org\/\">Wildfly<\/a> as an application server.<\/p><h2>Microservices<\/h2><p>If you're going to be developing a microservices application, you\nprobably want an application container. JVMs take a lot of memory\neach. Being able to share that across services is important. Also,\ncontainers often have a messaging system that you can use out of the\nbox.<\/p><h2>Immutant<\/h2><p><a href=\"http:\/\/immutant.org\/\">Immutant<\/a> is a full-fledge, enterprise grade application server\ncontainer for Clojure apps. It leverages <a href=\"http:\/\/wildfly.org\/\">Wildfly<\/a> and provides\nClojure wrappers for a host of services offered directly in the\ncontainer. I highly recommend it. The setup is easy and the developers\nare very responsive.<\/p><h2>Docker<\/h2><p>You can run Clojure apps in Docker containers. It works, and certainly\nif you've only got a small number of JVM apps, this may be a great\noption. But I'm a little worried about memory usage when you start\ndeploying many dockerized JVMs. <a href=\"https:\/\/wtfleming.github.io\/2014\/11\/29\/dockerize-clojure-compojure-http-kit-web-application\/\">Here is<\/a> a good description of the setup for a web app. There are Docker images for lots of Clojure systems, including <a href=\"https:\/\/hub.docker.com\/r\/jboss\/immutant\/\">Immutant<\/a>.<\/p><h2>OSv<\/h2><p>I've never run <a href=\"http:\/\/osv.io\/\">OSv<\/a>, but it could be an interesting option if you're\nwilling to go bleeding edge. OSv is a tiny OS that can run JVM\napplications. OSv runs directly on the hypervisor, so it removes a\nhuge layer of abstraction, meaning it's leaner and faster than running\na JVM on top of Linux.<\/p><p>There are <a href=\"http:\/\/blog.osv.io\/blog\/2014\/04\/22\/riemann-on-osv\/\">some instructions<\/a>. Also <a href=\"https:\/\/github.com\/cloudius-systems\/osv\/wiki\/Porting-Java-applications-to-OSv\">here<\/a> and <a href=\"https:\/\/github.com\/cloudius-systems\/osv\/wiki\/Porting-Clojure-applications-to-OSv\">here<\/a>.<\/p><h2>Other resources<\/h2><p><a href=\"http:\/\/www.luminusweb.net\/docs\/deployment.md\">Luminus has deployment docs<\/a> with many options.<\/p><h2>Conclusions<\/h2><p>The many deployment options can be overwhelming, but really the\nabundant choices are a blessing. The JVM is a mature platform that\nfinds itself in a wide variety of situations, from small hobby\nprojects to billion-dollar enterprise applications.<\/p><p>I've seen a lot of people turned off of Clojure because of the JVM,\nand I think that's a real shame. The JVM does have its warts, and it\nhas a history, but on the whole it's a great platform. That's why I\ncreated <a href=\"http:\/\/pftv.wpengine.com\/courses\/jvm-clojure\/\"><em>JVM Fundamentals for Clojure<\/em><\/a>. Please check it\nout if you're considering avoiding Clojure because of the JVM.<\/p><p>One of the cool things about the JVM is the Just-in-Time compiler\n(JIT). It's a way to optimize code at runtime based on the abundant\ninformation that running code gives you. We'll look at it next time!<\/p>","created_at":"2019-02-18T19:51:45+0000","updated_at":"2019-02-18T19:51:45+0000","published_at":"2017-02-24T16:59:34+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"purelyfunctional.tv","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["python","profiling"],"is_public":false,"id":845,"title":"Profiling Python Like a Boss","url":"https:\/\/zapier.com\/engineering\/profiling-python-boss\/","content":"<blockquote>\n<p>We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. ~Donald Knuth<\/p>\n<\/blockquote><p>It is unwise to embark on an optimization quest without first considering the famous quote by Knuth. However, sometimes that quick and hacky O(N^2) code you wrote (and promptly forgot about) to get some feature out the door rears its ugly head and demands more of your attention: this article is for that day.<\/p><p>So what follows are some useful tools and patterns for quick n' dirty profiling in Python. The goal here is simple: <strong>identify bottlenecks as quickly as possible, fix them and then confirm you've fixed them.<\/strong><\/p><h2>Write a Test<\/h2><p>Before you embark on any of this journey, write a quick high-level test that demonstrates the slowness. You may need to introduce some minimum set of data to reproduce a significant enough slowdown. Usually a run time of second or two is good enough to get a handle on an appreciable improvement when you find it.<\/p><p>It is also quite useful to have some standard tests to ensure your optimizations don't modify the behavior of the slow code. You might also be able to repurpose those tests as a bit of a \"benchmark\" by running them many times when profiling\/benchmarking.<\/p><p>The timing of this test case leads us to our first quick n' dirty profiling tool\u2026<\/p><h2>Simple Timers<\/h2><p>Timers are simple, flexible ways to measure execution time. You can drop them in just about anywhere and they have very minimal side-effects. It is quite easy to roll your own timer and even customize it to work exactly how you like. For example, a simple timing decorator might look like:<\/p><div class=\"codehilite\"><pre>import time\ndef timefunc(f):\n    def f_timer(*args, **kwargs):\n        start = time.time()\n        result = f(*args, **kwargs)\n        end = time.time()\n        print f.__name__, 'took', end - start, 'time'\n        return result\n    return f_timer\ndef get_number():\n    for x in xrange(5000000):\n        yield x\n@timefunc\ndef expensive_function():\n    for x in get_number():\n        i = x ^ x ^ x\n    return 'some result!'\n# prints \"expensive_function took 0.72583088875 seconds\"\nresult = expensive_function()\n<\/pre><\/div><p>Of course, you can get even more clever with by doing something with a context manager, giving you checkpoints and other good bits to work with:<\/p><div class=\"codehilite\"><pre>import time\nclass timewith():\n    def __init__(self, name=''):\n        self.name = name\n        self.start = time.time()\n    @property\n    def elapsed(self):\n        return time.time() - self.start\n    def checkpoint(self, name=''):\n        print '{timer} {checkpoint} took {elapsed} seconds'.format(\n            timer=self.name,\n            checkpoint=name,\n            elapsed=self.elapsed,\n        ).strip()\n    def __enter__(self):\n        return self\n    def __exit__(self, type, value, traceback):\n        self.checkpoint('finished')\n        pass\ndef get_number():\n    for x in xrange(5000000):\n        yield x\ndef expensive_function():\n    for x in get_number():\n        i = x ^ x ^ x\n    return 'some result!'\n# prints something like:\n# fancy thing done with something took 0.582462072372 seconds\n# fancy thing done with something else took 1.75355315208 seconds\n# fancy thing finished took 1.7535982132 seconds\nwith timewith('fancy thing') as timer:\n    expensive_function()\n    timer.checkpoint('done with something')\n    expensive_function()\n    expensive_function()\n    timer.checkpoint('done with something else')\n# or directly\ntimer = timewith('fancy thing')\nexpensive_function()\ntimer.checkpoint('done with something')\n<\/pre><\/div><p>Timers require you to do some digging. Start wrapping a few of the higher level functions and confirm where the bottleneck is, then drill down into that function, repeating as you go. When you find the disproportionately slow bit of code, fix it, and work your way back out confirming that it is fixed.<\/p><p><em>Handy tip: Don't forget the handy <a href=\"http:\/\/docs.python.org\/2\/library\/timeit.html\">timeit<\/a> module! It tends to be more useful for benchmarking small pieces of code than for doing the actual investigation.<\/em><\/p><ul><li><strong>Timer Pros:<\/strong> easy to understand and implement. Also very simple to compare before and after fixes. Works across many languages.<\/li>\n<li><strong>Timer Cons:<\/strong> sometimes a little <em>too<\/em> simplistic for extremely complex codebases, you might spend more time placing and replacing boilerplate code than you will fixing the problem!<\/li>\n<\/ul><h2>Built-in Profiler<\/h2><p>Breaking out the built-in profiler is akin to breaking out the big guns. It is extremely powerful, but a little unwieldy and sometimes a bit difficult to interpret and act on.<\/p><p>You can read a bit more about the built-in <a href=\"http:\/\/docs.python.org\/2\/library\/profile.html\">profile module<\/a> but the basics are quite simple: you enable and disable the profiler and it logs all function calls and execution times. It can then compile and print the output for you. A quick decorator simplifies this:<\/p><div class=\"codehilite\"><pre>import cProfile\ndef do_cprofile(func):\n    def profiled_func(*args, **kwargs):\n        profile = cProfile.Profile()\n        try:\n            profile.enable()\n            result = func(*args, **kwargs)\n            profile.disable()\n            return result\n        finally:\n            profile.print_stats()\n    return profiled_func\ndef get_number():\n    for x in xrange(5000000):\n        yield x\n@do_cprofile\ndef expensive_function():\n    for x in get_number():\n        i = x ^ x ^ x\n    return 'some result!'\n# perform profiling\nresult = expensive_function()\n<\/pre><\/div><p>In the case of the above code block, you should see something printed to the console that looks something like this:<\/p><div class=\"codehilite\"><pre>         5000003 function calls in 1.626 seconds\n   Ordered by: standard name\n   ncalls  tottime  percall  cumtime  percall filename:lineno(function)\n  5000001    0.571    0.000    0.571    0.000 timers.py:92(get_number)\n        1    1.055    1.055    1.626    1.626 timers.py:96(expensive_function)\n        1    0.000    0.000    0.000    0.000 {method 'disable' of '_lsprof.Profiler' objects}\n<\/pre><\/div><p>As you can see, it gives good numbers on various function calls but it omits a vital piece of information: what is it about a function that makes it so slow?<\/p><p>However, it is a great start to basic profiling. Sometimes it can even point you to the solution with very little fuss. I often use it as a gut check to start the debugging process before I dig deeper into the specific functions that are either slow are called way too often.<\/p><ul><li><strong>Built-in Pros:<\/strong> No external dependencies and quite fast. Useful for quick high-level checks.<\/li>\n<li><strong>Built-in Cons:<\/strong> Rather limited information that usually requires deeper debugging; reports are a bit unintuitive, especially for complex codebases.<\/li>\n<\/ul><h2>Line Profiler<\/h2><p>If the built-in profile was a big gun, consider the line profiler an Ion cannon. It is quite heavy and powerful (and a lot of fun to use!).<\/p><p>For this example, we'll use the great <a href=\"http:\/\/pythonhosted.org\/line_profiler\/\">kernprof<\/a> line-profiler available as a handy <a href=\"https:\/\/pypi.python.org\/pypi\/line_profiler\/\">line_profiler PyPi package<\/a>. For ease of use, we'll wrap it once again in a decorator that simplifies usage as well as provides a way to protect us from leaving it in production (because it is dog slow!).<\/p><div class=\"codehilite\"><pre>try:\n    from line_profiler import LineProfiler\n    def do_profile(follow=[]):\n        def inner(func):\n            def profiled_func(*args, **kwargs):\n                try:\n                    profiler = LineProfiler()\n                    profiler.add_function(func)\n                    for f in follow:\n                        profiler.add_function(f)\n                    profiler.enable_by_count()\n                    return func(*args, **kwargs)\n                finally:\n                    profiler.print_stats()\n            return profiled_func\n        return inner\nexcept ImportError:\n    def do_profile(follow=[]):\n        \"Helpful if you accidentally leave in production!\"\n        def inner(func):\n            def nothing(*args, **kwargs):\n                return func(*args, **kwargs)\n            return nothing\n        return inner\ndef get_number():\n    for x in xrange(5000000):\n        yield x\n@do_profile(follow=[get_number])\ndef expensive_function():\n    for x in get_number():\n        i = x ^ x ^ x\n    return 'some result!'\nresult = expensive_function()\n<\/pre><\/div><p>If you run the above chunk of code, you should see a report that looks something like this:<\/p><div class=\"codehilite\"><pre>Timer unit: 1e-06 s\nFile: test.py\nFunction: get_number at line 43\nTotal time: 4.44195 s\nLine #      Hits         Time  Per Hit   % Time  Line Contents\n==============================================================\n    43                                           def get_number():\n    44   5000001      2223313      0.4     50.1      for x in xrange(5000000):\n    45   5000000      2218638      0.4     49.9          yield x\nFile: test.py\nFunction: expensive_function at line 47\nTotal time: 16.828 s\nLine #      Hits         Time  Per Hit   % Time  Line Contents\n==============================================================\n    47                                           def expensive_function():\n    48   5000001     14090530      2.8     83.7      for x in get_number():\n    49   5000000      2737480      0.5     16.3          i = x ^ x ^ x\n    50         1            0      0.0      0.0      return 'some result!'\n<\/pre><\/div><p>As you can see, these are highly detailed reports that give unprecedented insight into code execution. Unlike the built-in cProfiler, it can break out time spent core language features like loops or imports and gives cumulative numbers of time spent on various lines.<\/p><p>This sort of detail makes it much easier to get quick understanding of a functions internal. Plus, if you need to dive down into some third party library, you can just import the third party library and pass it directly into the decorator.<\/p><p><em>Handy tip: Just decorate your test function and pass the problem function in the follow argument!<\/em><\/p><ul><li><strong>Line Profiler Pros:<\/strong> has intuitive and finely detailed reports. Can follow functions in third party libraries.<\/li>\n<li><strong>Line Profiler Cons:<\/strong> because of the overhead it can be an order of magnitude <em>slower<\/em> than real execution time, so don't use it for benchmarking. It is also an external requirement.<\/li>\n<\/ul><h2>Conclusion &amp; Best Practices<\/h2><p>You should use some of the simpler tools (like timers or the built-in profiler) to perform some gut checks around test cases (especially around some code you know quite well) and switch to a few of the slower but more detailed tools like <code>line_profiler<\/code> to get a close look at the internals of functions.<\/p><p>Nine times out of ten you'll notice a silly mistake where repeatedly calling something inside a loop or using an incorrect data structure is chewing up some 90%+ of the time spent in a function. A couple quick (and satisfying) tweaks and you'll be on your way.<\/p><p>If you are still too slow and notice yourself doing esoteric things like comparing attribute accessing techniques or tweaking equality checking techniques, you're likely past the point of diminishing returns. You might have to either:<\/p><ol><li>Live with the slowness or just pre-calculate\/cache it.<\/li>\n<li>Rethink the entire implementation.<\/li>\n<li>Use more optimized data structures (via Numpy, Pandas, etc.).<\/li>\n<li>Write a <a href=\"http:\/\/docs.python.org\/2\/extending\/extending.html\">C extension<\/a>.<\/li>\n<\/ol><p>Beware, optimizing code can be a guilty pleasure! It is pretty fun finding interesting ways to speed up Python, but be careful not to obfuscate what is going on under the hood with arcane speedups. <strong>Readable code is generally better than slightly faster code.<\/strong> It is almost always simpler to just cache it and move on.<\/p><p>That said, enjoy! :-)<\/p><p><em>PS: check out the code examples in https:\/\/github.com\/zapier\/profiling-python-like-a-boss. Also, learn how to <a href=\"https:\/\/zapier.com\/engineering\/debugging-python-boss\/\">Debug Python Like a Boss<\/a>.<\/em><\/p>","created_at":"2019-02-18T15:33:01+0000","updated_at":"2019-02-18T15:33:04+0000","published_at":"2013-12-02T13:13:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"zapier.com","preview_picture":"https:\/\/zapier.com\/engineering\/profiling-python-boss\/images\/logos\/social.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["activitypub","fediverse","magenc","p2p","magnet"],"is_public":false,"id":844,"title":"WebOfTrustInfo\/rwot7-toronto","url":"https:\/\/github.com\/WebOfTrustInfo\/rwot7-toronto\/blob\/master\/topics-and-advance-readings\/magenc.md","content":"<p>Author: <a href=\"https:\/\/dustycloud.org\/\" rel=\"nofollow\">Christopher Lemmer Webber<\/a><\/p>\n<h2><a id=\"user-content-introduction\" class=\"anchor\" aria-hidden=\"true\" href=\"#introduction\"><\/a>Introduction<\/h2>\n<p>The web has been a great tool for the wide distribution of content.\nUnfortunately, reliance on HTTP based URIs has meant that content is\nliable to disappear or change unexpectedly.\nAdditionally, access control for private content is difficult to make\nsecure on the contemporary web.<\/p>\n<p>Content Addressed Storage allows for us to refer to objects by their\nhash without binding that object to a specific location, and may be\nreferred to with such addresses as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Uniform_Resource_Name\" rel=\"nofollow\">URNs<\/a> or <a href=\"https:\/\/ipfs.io\/\" rel=\"nofollow\">IPFS<\/a> addresses.\nThese are popularly combined with peer to peer distribution using the\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Magnet_URI_scheme\" rel=\"nofollow\">magnet URI scheme<\/a>.\nFor example, a file with the text \"foo bar baz\" may be hashed with the\nsha1 algorithm to get the address\n<code>urn:sha1:4e4ced5ee12c698209feaab89fd33e93fb7890dd<\/code>\nwhich is the content addressed URI for that file.\nThe magnet URI scheme allows for attaching additional metadata about\nwhere that file may be retrieved such as over the Bittorrent protocol.\nSuch a magnet URI might look like:<\/p>\n<pre>magnet:?xt=urn:sha1:4e4ced5ee12c698209feaab89fd33e93fb7890dd&amp;tr=https:\/\/tracker.example\/abc1\n<\/pre>\n<p>Additional metadata, such as a suggested filename, a suggested HTTP\nURL location that may hold the file, or the size of the file are also\npossible to add as query parameters to the magnet scheme:<\/p>\n<pre>magnet:?xt=urn:sha1:4e4ced5ee12c698209feaab89fd33e93fb7890dd&amp;as=https:\/\/filestore.example\/foo.txt&amp;xs=11\n<\/pre>\n<p>(Some liberties have been taken for the sake of visual clarity over\ncorrectness; future magnet URIs in this article will correctly\npercent-encode the query parameters.)<\/p>\n<p>Unfortunately this leaves content distributed in cleartext.\nThis can be especially disastrous if we would like to distribute\nprivate content while still taking advantage of peer to peer systems.\nWorse yet, participants in a peer to peer system who may volunteering\ntheir resources to help the network might be at risk of liability for\ndistributing or hosting content in the interest of helping the\nprotocol while having no knowledge of the content.<\/p>\n<p>In this paper we introduce an extension to Magnet URIs we call\n\"magenc\" by adding two more additional query parameters, the key and\nthe encryption algorithm used.\nMagenc is inspired by such systems as <a href=\"https:\/\/tahoe-lafs.org\/trac\/tahoe-lafs\" rel=\"nofollow\">Tahoe-LAFS<\/a>, <a href=\"http:\/\/www.nongnu.org\/libchop\/\" rel=\"nofollow\">libchop<\/a>, and\ncapability systems such as <a href=\"http:\/\/waterken.sourceforge.net\/\" rel=\"nofollow\">Waterken<\/a> and the <a href=\"http:\/\/erights.org\/\" rel=\"nofollow\">E programming language<\/a>.\nWe take advantage of magnet's URI composition by allowing for content\nto be \"sealed\" and \"unsealed\": when content is distributed via its\ncontent addressed representation without knowledge of the decryption\nkey, it is sealed.\nParticipants who are intended to be made knowledgeable of the file's\ncontents are given the magenc-extended magnet URI which contains the\nkey and algorithm information necessary to unseal its contents.\nAdditionally, we have provided mitigations against inferring content\nbased on file length via chunking and have provided optional file\ndeduplication via convergent encryption.<\/p>\n<p>This paper addresses immutable content only; mutable content must be\nadded as a separate layer, but may use the designs in this\nspecification as part of its design.<\/p>\n<h2><a id=\"user-content-magenc-by-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#magenc-by-example\"><\/a>Magenc by example<\/h2>\n<p>Magenc is best understood by example.\nThankfully, a demo implementation exists.\nFirst, install the Racket programming language.\nThen, at the command line run:<\/p>\n<pre>$ raco install magenc\n<\/pre>\n<p>This Magenc package comes with a demo content addressed storage\nserver.\nIt is not intended for production use, as it stores content in\nmemory only.\nLet's start it up:<\/p>\n<pre>$ raco cas-server\nYour Web application is running at http:\/\/localhost:8000.\nStop this program at any time to terminate the Web Server.\n<\/pre>\n<p>Note that though we use an HTTP server in this demo, any sort of\nstorage system that knows how to dereference the eXactTopic links can\nbe used, from a local offline database to a full on peer to peer\nnetwork.\nAlso note that the magnet URI itself will never be provided to the\nstorage system, only the eXactTopic parameter.<\/p>\n<p>This demo server only has two operations\u2026 a POST request stores\nthe content and returns its content-addressed URN:<\/p>\n<pre>$ curl -d \"Hello CAS store\" -X POST http:\/\/localhost:8000\nurn:sha256:y7y84K0IO8apO0FA9CWNPU7jqzpHFrR1W4YLChshm2w\n<\/pre>\n<p>Submitting it again will give us the same URN, since it is stored\nbased on its contents:<\/p>\n<pre>$ curl -d \"Hello CAS store\" -X POST http:\/\/localhost:8000\nurn:sha256:y7y84K0IO8apO0FA9CWNPU7jqzpHFrR1W4YLChshm2w\n<\/pre>\n<p>We can then get the content back by requesting its location based on\nthis URN, which is its eXactTopic:<\/p>\n<pre>$ curl http:\/\/localhost:8000?xt=urn:sha256:y7y84K0IO8apO0FA9CWNPU7jqzpHFrR1W4YLChshm2w\nHello CAS store\n<\/pre>\n<p>Unfortunately, this is a fairly naive CAS store and no encryption is\nperformed directly.<\/p>\n<p>Let's put ourselves in the feet of Carlos.\nCarlos would like to write a sappy love letter to his significant\nother Bob, but before he sends it he would like his friend Alice to\nreview.\nCarlos pens a rough draft:<\/p>\n<pre>$ echo \"Dear Bob, my love for you is greater than the sum of stars. -- Carlos\" &gt; sappy-love-letter.txt\n<\/pre>\n<p>Perhaps we would like to pen a sappy love letter and send it to our\nsignificant other.\nBut what if the administrator of the storage server could see it?\nCarlos might be terribly embarrassed\u2026 too embarrassed to ever post it.<\/p>\n<p>Thankfully Carlos can seal the contents:<\/p>\n<pre>$ raco magenc --upload sappy-love-letter.txt\nmagnet:?xt=urn%3Asha256%3AX74UbU3NoLTA_Nupi8DhaJ_oQpQ95KFukMAkJJotKgo&amp;ek=eekxqfiZIcEnc8cpR-sD_3X3qLaTzQW-KnovArMkGP0&amp;es=aes-ctr\n<\/pre>\n<p>We can then get the contents back:<\/p>\n<pre>$ raco magenc --get \"magnet:?xt=urn%3Asha256%3AX74UbU3NoLTA_Nupi8DhaJ_oQpQ95KFukMAkJJotKgo&amp;ek=eekxqfiZIcEnc8cpR-sD_3X3qLaTzQW-KnovArMkGP0&amp;es=aes-ctr\"\nDear Bob, my love for you is greater than the sum of stars. -- Carlos\n<\/pre>\n<p>(Astute readers will note that we haven't mentioned signing contents\nso that Alice and Bob know that Carlos really signed it.\nCarlos could always distribute a file that signs the object\nalongside or embedded in the file, but that is left as an exercise\nfor the reader.)<\/p>\n<p>Let's unpack that Magnet URI.\nIt has three different query parameters:<\/p>\n<ul><li><strong>xt<\/strong> is the eXactTopic, which gives the hash which refers to the\nrelevant <em>encrypted<\/em> file.<\/li>\n<li><strong>ek<\/strong> is the EncryptionKey used to encrypt\/decrypt the file\nand is base64 encoded.\nThis is a Magenc extension to the <code>magnet:<\/code> uri scheme.<\/li>\n<li><strong>es<\/strong> is the EncryptionSuite, the encryption algorithm used.\nThis is a Magenc extension to the <code>magnet:<\/code> uri scheme.<\/li>\n<\/ul><p>Carlos is suddenly worried that maybe the administrator is able to see\nthis content.\nWe can see exactly what file it uploaded by adding the <code>--verbose<\/code> flag.<\/p>\n<pre>$ raco magenc --verbose --upload sappy-love-letter.txt\nDEBUG: posted urn:sha256:yWTt2U5hvAn4jGAByzzHrtMasN98xadFqjKLhA4EpXM\nmagnet:?xt=urn%3Asha256%3AyWTt2U5hvAn4jGAByzzHrtMasN98xadFqjKLhA4EpXM&amp;ek=Jb79KiJpHvxkF36hzQcvCthZX0r0GEfWoa4nLAf28Vg&amp;es=aes-ctr\n<\/pre>\n<p>Ok, it looks like the file we uploaded is the same file as at the\neXactTopic.\nLet's see what happens if we download that file from the CAS store\ndirectly:<\/p>\n<pre>$ curl http:\/\/localhost:8000?xt=urn:sha256:yWTt2U5hvAn4jGAByzzHrtMasN98xadFqjKLhA4EpXM\n&lt;screen fills with garbage&gt;\n<\/pre>\n<p>Yikes!\nIt spit out a bunch of binary junk\u2026 Carlos can't tell what that is,\nand thankfully neither can the storage administrator.<\/p>\n<p>However, we could write out those contents out to a file and decrypt\nit manually.\nIf we did, it would look like this:<\/p>\n<pre>(3:raw70:Dear Bob, my love for you is greater than the sum of stars. -- Carlos\\n)\n<\/pre>\n<p>\u2026 followed by a bunch of bytes of whitespace.\nThe whitespace is easy to explain: to prevent side channel attacks\nwhere content can be inferred by file length, additional padding is\nadded so that files fit within the relevant \"chunk size\".\n(We are defaulting to a chunk size of 32 Kilobytes.)\nIf this file were larger, we would split it into multiple chunks\nand the first object retrieved would be a <code>manifest<\/code> pointing at\nthe rest of the chunks.\nSince this file is so small we can include it in a single chunk\nwithout requiring a manifest, so it is a <code>raw<\/code> file.<\/p>\n<p>Now, what's with the rest of the stuff in the file?\nFirst of all, since the \"raw\" file is an optimization, we need\na way to distinguish between the raw file and a \"manifest\" file.\nSecond, while there isn't much to the raw file, manifest files\nneed more structure.\nSo we are using the world's simplest structured data format,\n<a href=\"https:\/\/people.csail.mit.edu\/rivest\/Sexp.txt\" rel=\"nofollow\">canonical s-expressions<\/a>.\nCanonical s-expressions have the advantage of being small and\neasy to implement (an implementation can take as little time\nas an hour or two for a seasoned programmer).\nCanonical s-expressions combine two concepts: netstrings and\ns-expressions.<\/p>\n<p>Netstrings are the world's simplest non-structured data format.\nThe netstring for the word \"cat\" is <code>3:cat<\/code>.\nNumber of characters in base-10 of the object's size in bytes,\nthen the bytes.\nThis also has the advantage that binary data is easy to encode.<\/p>\n<p>S-expressions are lists of items, delimited by parentheses.\nSo we could describe a zoo full of animals (where an animal\nis described by its name, what it says, and what it eats) like:<\/p>\n<pre>(zoo (animal cat meow kibble)\n     (animal monkey ooh-ooh bananas))\n<\/pre>\n<p>(This is also more or less the \"advanced encoding\" of\ncanonical s-expressions.)<\/p>\n<p>Canonical s-expressions replace the items in the lists with\nnetstrings.\nSo the canonical s-exp (in canonical form) for our zoo would look like:<\/p>\n<pre>(3:zoo(6:animal3:cat4:meow6:kibble)(6:animal6:monkey7:ooh-ooh7:bananas))\n<\/pre>\n<p>Magenc implementations MUST only use the canonical implementation of\ncanonical s-expressions on the wire, so the implementation overhead\nhere is very small.<\/p>\n<p>Knowing that, the format for a \"raw\" object in Magenc looks like:<\/p>\n<pre>(raw &lt;raw-data-goes-here&gt;)\n<\/pre>\n<p>And that's it, hence the contents of the encrypted object.\nBut in general, users do not need to know about this.<\/p>\n<p>Carlos sends the Magenc-extended Magnet URI to Alice.\nShe says that the letter is very nice, but that Carlos is such a nice\nartist, so why not do a digital painting of a galaxy with the words\nsuperimposed?\nCarlos immediately agrees that this is a good and not at all\ntacky improvement to the design and gets to work.<\/p>\n<p>Carlos has finished the painting and uploads it.\nThis time he is curious about how many objects it will upload, so\nhe runs with <code>--verbose<\/code><\/p>\n<pre>$ raco magenc --verbose --upload sappy-love-painting.png \nDEBUG: posted urn:sha256:7gfqd3hDTf56FEb9i_x9_cxwgVwjUNDwldJtC9v1T8o\nDEBUG: posted urn:sha256:E7XwTkJO2ufGW7cCc9qk5rNJPh2xTbxxDN0HMQHiP4s\nDEBUG: posted urn:sha256:JVdTXIHBBkF2dHvc_FW8wmKN_2ow9lSq3orWF0V0G4o\nDEBUG: posted urn:sha256:hClheQii8FoomnWNLiLc8E2A8HYtCWbVtA22xWblMCs\nDEBUG: posted urn:sha256:fwFIj8TIXaeiViqbH252V5L0mY3EOb5pPXhQg-Xci1c\nmagnet:?xt=urn%3Asha256%3AfwFIj8TIXaeiViqbH252V5L0mY3EOb5pPXhQg-Xci1c&amp;ek=IEbic8nd8V82n2K6huAyI6T_fq0vKVj5NDbBpn4JhiI&amp;es=aes-ctr\n<\/pre>\n<p>Huh, that's interesting!\nThis time it uploaded more chunks than last time.\nLet's decrypt the eXactTopic again.\nMinus the padding (manifest files are padded to the length of the\nnearest available multiple of the chunk size), it looks like:<\/p>\n<pre>(8:manifest5:327686:12528654:urn:sha256:7gfqd3hDTf56FEb9i_x9_cxwgVwjUNDwldJtC9v1T8o54:urn:sha256:E7XwTkJO2ufGW7cCc9qk5rNJPh2xTbxxDN0HMQHiP4s54:urn:sha256:JVdTXIHBBkF2dHvc_FW8wmKN_2ow9lSq3orWF0V0G4o54:urn:sha256:hClheQii8FoomnWNLiLc8E2A8HYtCWbVtA22xWblMCs)\n<\/pre>\n<p>That's pretty hard to read, so here it is pretty-printed:<\/p>\n<pre>(manifest \"32768\" \"125286\"\n \"urn:sha256:7gfqd3hDTf56FEb9i_x9_cxwgVwjUNDwldJtC9v1T8o\"\n \"urn:sha256:E7XwTkJO2ufGW7cCc9qk5rNJPh2xTbxxDN0HMQHiP4s\"\n \"urn:sha256:JVdTXIHBBkF2dHvc_FW8wmKN_2ow9lSq3orWF0V0G4o\"\n \"urn:sha256:hClheQii8FoomnWNLiLc8E2A8HYtCWbVtA22xWblMCs\")\n<\/pre>\n<p>Manifest files look like:<\/p>\n<pre>(manifest &lt;chunk-size&gt; &lt;file-size&gt; &lt;chunk1&gt; &lt;chunk2&gt; ...)\n<\/pre>\n<p>So based on that, we can tell that the sappy love painting's manifest\nindeed contains all the chunks we saw from the DEBUG messages of the\nupload stage.<\/p>\n<p>Alice downloads the file on her end:<\/p>\n<pre>$ raco magenc --verbose --get \"magnet:?xt=urn%3Asha256%3AfwFIj8TIXaeiViqbH252V5L0mY3EOb5pPXhQg-Xci1c&amp;ek=IEbic8nd8V82n2K6huAyI6T_fq0vKVj5NDbBpn4JhiI&amp;es=aes-ctr\" --output carlos-love-painting.png\nDEBUG: got urn:sha256:fwFIj8TIXaeiViqbH252V5L0mY3EOb5pPXhQg-Xci1c\nDEBUG: got urn:sha256:7gfqd3hDTf56FEb9i_x9_cxwgVwjUNDwldJtC9v1T8o\nDEBUG: got urn:sha256:E7XwTkJO2ufGW7cCc9qk5rNJPh2xTbxxDN0HMQHiP4s\nDEBUG: got urn:sha256:JVdTXIHBBkF2dHvc_FW8wmKN_2ow9lSq3orWF0V0G4o\nDEBUG: got urn:sha256:hClheQii8FoomnWNLiLc8E2A8HYtCWbVtA22xWblMCs\n<\/pre>\n<p>This makes sense\u2026 the first object that Alice's client retrieved was\nthe manifest, so that's the first object outputted in the DEBUG\nmessages, and then every file after that was listed as a chunk in the\nmanifest.\nAlice's client takes care of details so she doesn't have to, from\ndecrypting the manifest and each chunk, to verifying that each chunk\nmatches the chunk size calculating how much to read from the last\nchunk based on the total file size (the rest is padding).\nBut Alice doesn't need to worry about any of that.<\/p>\n<p>Alice looks at the image and thinks it looks quite good.\nShe says as such to Carlos, who enthusiastically then passes the\nmagenc-extended magnet url along to Bob.<\/p>\n<p>Bob really appreciates the painting, and Carlos, feeling relieved that\nBob enjoys it so much, consents to Bob's request that they share the\npainting far and wide across their group of friends.<\/p>\n<p>Bob and Carlos' friend group all use the same shared magenc-enabled\nbackup service and pay for the costs of that shared storage.\nWhy pay the additional storage cost for the same file?\nWe can see the potential problem when we try running the upload\ncommand multiple times:<\/p>\n<pre>$ raco magenc --upload sappy-love-painting.png\nmagnet:?xt=urn%3Asha256%3AFY9rimPzXjqw8u3AfM-ErrGp072W4gYPE9fCJ8MhXuE&amp;ek=NLeRbCuKcrPoQE8G9OWyDukaiDqQS_nhKZHkc4oCKzk&amp;es=aes-ctr\n$ raco magenc --upload sappy-love-painting.png\nmagnet:?xt=urn%3Asha256%3AV5FXelScKY2bv1Rl4emN6qFjHAojfExBkyMV-HMAvUI&amp;ek=0Hf-JsD_olN7TESvaHBy-DyDqQUrzneIOc53nZdXXmc&amp;es=aes-ctr\n$ raco magenc --upload sappy-love-painting.png\nmagnet:?xt=urn%3Asha256%3AoYOvXseaJ3OdWtQ6EUiS83V5FRuG5VXejAgps-bWGIM&amp;ek=EPf3E03lx1qpeLBlMMAOVjosSac8IVbyLUDTMhhu_Y8&amp;es=aes-ctr\n<\/pre>\n<p>Uhoh\u2026 each of those uploads creates a new random key, and thus\na new object.\nFortunately we can take advantage of a concept called \"convergent\nencryption\" where instead we use the hash of the decrypted file as\nthe key (note that this will be different than the hash of the\nencrypted objects, so having access to the hash of the encrypted\nobject used in the eXactTopic will not expose the contents).\n(Note that convergent encryption should be fine in this case but\nmay have security risks for <a href=\"https:\/\/tahoe-lafs.org\/hacktahoelafs\/drew_perttula.html\" rel=\"nofollow\">certain data-sharing situations<\/a>.)\nLet's see what happens now if we try uploading the same file\nmultiple times:<\/p>\n<pre>$ raco magenc --upload sappy-love-painting.png --convergent\nmagnet:?xt=urn%3Asha256%3AH9dDY4Y0y39K2rziDY5Ul9iBDfP51Fg9EAHncrIINDU&amp;ek=rGVdQ2K4dGhEOFd-50eoRwNRYPGf1w0yIi_E5_bE5WY&amp;es=aes-ctr\n$ raco magenc --upload sappy-love-painting.png --convergent\nmagnet:?xt=urn%3Asha256%3AH9dDY4Y0y39K2rziDY5Ul9iBDfP51Fg9EAHncrIINDU&amp;ek=rGVdQ2K4dGhEOFd-50eoRwNRYPGf1w0yIi_E5_bE5WY&amp;es=aes-ctr\n$ raco magenc --upload sappy-love-painting.png --convergent\nmagnet:?xt=urn%3Asha256%3AH9dDY4Y0y39K2rziDY5Ul9iBDfP51Fg9EAHncrIINDU&amp;ek=rGVdQ2K4dGhEOFd-50eoRwNRYPGf1w0yIi_E5_bE5WY&amp;es=aes-ctr\n<\/pre>\n<p>As we can see, this time all of the magnet URIs are exactly the same!\nUsing convergent encryption, Bob and Carlos' friend group can all\nsafely and back up the same file to the shared hosting service without\nwasting space.\nThat will be extra helpful after all the photos and videos will be\nshared amongst each other from their upcoming wedding!<\/p>\n<h2><a id=\"user-content-conclusions\" class=\"anchor\" aria-hidden=\"true\" href=\"#conclusions\"><\/a>Conclusions<\/h2>\n<p>Using content addressed storage is a great way to distribute content\nin a decentralized system which is robust against failing nodes, but\nby default is unsuitable for anything but public content and puts a\ngreat burden of responsibility for hosting providers and peer to peer\nnodes that want to distribute content for the sake of the network's\noperation but which do not want to be held responsible for the files'\ncontents.\nAdding symmetric encryption provides privacy for participants sharing\ncontent while also allowing general distribution of that content\nwithout revealing information except to participants who have been\nintentionally chosen to receive it.<\/p>\n<p>The magnet URI scheme already composes other URIs with additional\nmetadata and is in wide use for many peer to peer networks.\nThis makes it a reasonable place to add metadata for the encryption\nkey and suite (although clients which are not aware of the magenc\nextensions will receive the encrypted information without being\naware of how to interpret it).\nAlthough this demo used URNs, since the magnet URI scheme does not\nspecify a specific URI scheme for the eXactTopic, other URI schemes\nsuch as IPFS can be used to the same effect.<\/p>\n<p>The decision to type the first file retrieved means that for\nparticularly small files, no more than one request is needed to\nretrieve the entire file since the file may be tagged as \"raw\".\nFor larger files, the decision to use a single manifest file over a\nmerkle tree has the upside of ease of implementation and that the\nnumber of requests is <code>n+1<\/code> where <code>n<\/code> is the number of chunks, whereas\na merkle tree would be <code>O(log(n))+n<\/code> requests.\nHowever when a manifest file exceeds the chunk size it is rounded up\nto the nearest multiple of the chunk size which leaks both that this\nfile is a manifest file and also gives some coarse grained information\nabout the approximate range of the file it references in megabytes.\nHowever the author believes this tradeoff is worth it; one additional\nbenefit of the system described in this paper is that given a decent\nprogramming language and a cryptography library, the entire system can\nbe written in one weekend, including the canonical s-expressions\nlibrary.\nThe author of this article wrote the demo tools used in this article\nin a single day (although they had already written the canonical\ns-expressions library on a different day).<\/p>\n<p>Details about garbage collection, which nodes will host content,\nand how to provide mutability are out of the scope of this paper\nbut may be composed separately in conjunction with its design.\nFor example, a gossip protocol plus a signed git-like merkle tree\ncould be used to allow entities to provide updates while referencing\nimmutable objects referenced by the magenc-extended magnet URIs\ndescribed in this paper.\nIf an enforced linear structure to those updates is necessary,\na consensus algorithm could be used.\nAs for garbage collection, entities may choose to pin or store the\nobjects they particularly care about, or services may use such\na policy such as deleting objects which have not been recently\nrequested.<\/p>\n<p>Local tests against this not-particularly-optimized demo\nimplementation indicate that the general design is reasonably fast\nfor many use cases.\nUploading a 38 megabyte video file takes .8 seconds (excluding startup\ntime) against the demo http server CAS store, and against an\nin-process memory store takes .6 seconds.\nFetching the same uploaded file takes 1.2 seconds against the HTTP\nserver and a mere .2 seconds against the memory store, indicating that\nthe primary overhead on retrieval is the HTTP server implementation\nitself.<\/p>\n<h2><a id=\"user-content-related-work-and-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#related-work-and-thanks\"><\/a>Related work and thanks<\/h2>\n<p><a href=\"https:\/\/tahoe-lafs.org\/trac\/tahoe-lafs\" rel=\"nofollow\">Tahoe-LAFS<\/a> and <a href=\"http:\/\/www.nongnu.org\/libchop\/\" rel=\"nofollow\">Libchop<\/a> are the primary sources of inspiration for\nthe ideas in this paper.\nBoth of those have the fundamental idea of distributing encrypted,\nchunked copies of file to storage providers that are unaware of\nthe files' contents.\nMuch of this design came from reading the <a href=\"https:\/\/gnunet.org\/sites\/default\/files\/lafs.pdf\" rel=\"nofollow\">Tahoe-LAFS paper<\/a> and\n<a href=\"https:\/\/hal.archives-ouvertes.fr\/hal-00187069\/en\/\" rel=\"nofollow\">this paper about libchop<\/a>.\nBoth Tahoe-LAFS and libchop provide some features that are not\nin this paper for the sake of keeping things simple to implement,\nsuch as compression and versioning.<\/p>\n<p><a href=\"http:\/\/waterken.sourceforge.net\/\" rel=\"nofollow\">Waterken<\/a> and <a href=\"http:\/\/sebsauvage.net\/wiki\/doku.php?id=php:zerobin\" rel=\"nofollow\">Zerobin<\/a> both provided the idea of having a URI that\ncontains a secret which the source holding the content cannot see.\nBoth of these use the fragment identifier part of a URI to accomplish\nthis, but this would not work for us for two reasons: both of these\nuse a dynamic client (such as one using javascript) to interpret\nincoming data, and also because motivating examples for this paper\ninclude linked data systems where the fragment is already highly\noverloaded and needed for such things as referring to vocabulary\nterms.<\/p>\n<p>Special thanks to Ludovic Court\u00e8s for discussing these ideas\nand to Zooko for trying to convince Christopher Lemmer Webber\nto take Tahoe-LAFS seriously for their projects, even if it took\nChris nearly a decade later to do so.\nThanks also to the capability community in general for listening\nto and answering many questions, such as what the heck does a\nsealer \/ unsealer pair mean anyway.<\/p>","created_at":"2019-02-18T15:11:06+0000","updated_at":"2019-02-18T15:11:22+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/14255940?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":843,"title":"Will Gutenberg Replace Page Builders Like Elementor, Divi or Beaver Builder?","url":"https:\/\/www.codeinwp.com\/blog\/gutenberg-replace-page-builders\/?amp","content":"<p><img src=\"https:\/\/mk0codeinwp10tp0961a.kinstacdn.com\/wp-content\/uploads\/2018\/08\/gutenberg-replace-page-builders.jpg\" alt=\"Will Gutenberg Replace Page Builders?\" width=\"1450\" height=\"740\" class=\"aligncenter size-full wp-image-38751 amp-wp-enforced-sizes\" \/><\/p><p>As we keep rolling towards WordPress 5.0, a lot of people are wondering if we\u2019ll see Gutenberg replace page builders like Elementor, Divi or Beaver Builder once it\u2019s officially merged into the core.<\/p><p>In this post, I\u2019ll shed some light on this topic and analyze what\u2019s likely to come next for page builders in the \u201cGutenberg age\u201d.<\/p><div class=\"su-row\"><div class=\"su-column su-column-size-1-2\"><p>First off, it\u2019s a tough thing to cover because <em>some<\/em> of the answer is going to hinge on the far off plan for Gutenberg to \u201cgo beyond the post into page templates and ultimately, full site customization\u201d. Right now, we don\u2019t have any real frame of reference for how that will work, which naturally makes detailed prognostications difficult.<\/p><\/div><\/div><p><strong>Here\u2019s the TL;DR for where it stands based on what we know today, though:<\/strong><\/p><p>There\u2019s little chance of Gutenberg universally replacing page builders <strong><em>for all types of users<\/em><\/strong>. At least not anytime soon. But page builders shouldn\u2019t start celebrating quite yet because <strong>Gutenberg will likely steal casual users away from page builders.<\/strong><\/p><p>Keep reading for more on how we think it will play out\u2026<\/p><h2>A quick catch-up on how Gutenberg works<\/h2><p>I won\u2019t belabor this point because you probably already know a bit about Gutenberg if you\u2019re reading this post.<\/p><p>But, for the uninitiated, the Gutenberg Editor is the upcoming replacement for the current WordPress TinyMCE editor. Here\u2019s Gutenberg in a few bullet points:<\/p><ul><li>It uses individual \u201cblocks\u201d for design, rather than a single text field like the current TinyMCE editor.<\/li> <li>A block can be anything. Text and images are two core elements, but a block can also be a button, a quote, embedded content, or lots more.<\/li> <li>There\u2019s <em>very limited<\/em> column support (in beta).<\/li> <li>You can use drag-and-drop to arrange the order of your blocks. But again, it\u2019s more limited than a page builder.<\/li> <\/ul><p>It all looks something like this:<\/p><p>If you want to learn more, check out <a href=\"https:\/\/www.codeinwp.com\/blog\/wordpress-gutenberg-guide\/\">our full Gutenberg editor guide<\/a>.<\/p><h2>Gutenberg isn\u2019t anywhere close to a page builder right now (but that won\u2019t matter for everyone)<\/h2><p>If you\u2019ve played around with the Gutenberg editor, you know that <em>in its current incarnation<\/em>, it\u2019s not even close to a full replacement for page builders. <em>If you haven\u2019t, <\/em><a href=\"https:\/\/frontenberg.tomjn.com\/\" target=\"_blank\" rel=\"noopener\"><em>go play around with it here<\/em><\/a><em> (no need to install anything).<\/em><\/p><p>I could make a lengthy list of reasons why we won\u2019t see Gutenberg replace page builders in all aspects right now, but here are the biggies:<\/p><ul><li><strong>Front-end visual editing<\/strong> \u2013 Gutenberg lacks this, and <a href=\"https:\/\/github.com\/WordPress\/gutenberg\/issues\/1315#issuecomment-313900765\" target=\"_blank\" rel=\"noopener\">that\u2019s intentional<\/a> (<em>though there are some cool alternatives in the works<\/em>).<\/li> <li><strong>Style\/design options<\/strong> \u2013 while Gutenberg lets you change something basic like a button color, the overall styling options aren\u2019t even close.<\/li> <li><strong>Columns<\/strong> \u2013 Gutenberg now has a beta column block (<em>that\u2019s pretty buggy right now<\/em>)\u2026but nothing to match how page builders do it.<\/li> <li><strong>No specialty blocks <\/strong>\u2013 page builders include specialty blocks like testimonials and pricing tables, which Gutenberg doesn\u2019t have by default (<em>but you can get them!<\/em>)<\/li> <\/ul><p>If you look at the <a href=\"https:\/\/www.codeinwp.com\/blog\/elementor-vs-divi-builder-vs-beaver-builder\/\">most popular page builders<\/a>, not only do they do all that stuff, <em>they do it really well<\/em>.<\/p><p>For a lot of users, the lack of those features makes Gutenberg a complete non-starter in terms of a viable page builder replacement.<\/p><h3>People who build WordPress sites with page builders need this functionality<\/h3><p>If you\u2019re using a page builder to actually build your site (<em>rather than just basic blog post layouts<\/em>), this functionality is 100% essential.<\/p><p>And that\u2019s significant \u2013 there are huge communities of people building complete sites based on page builders \u2013 us included! We just <a href=\"https:\/\/elementor.com\/codeinwp-case-study\/\" target=\"_blank\" rel=\"noopener\">rebuilt this very site with Elementor<\/a>. And we\u2019re not alone\u2026<\/p><p>Just a quick perusal on Facebook turns up:<\/p><ul><li><strong>Elementor Community<\/strong> \u2013 22,000+ members<\/li> <li><strong>Beaver Builders<\/strong> \u2013 12,000+ members<\/li> <li><strong>Divi Theme Users<\/strong> \u2013 28,000+ members<\/li> <\/ul><p>And that\u2019s just scratching the surface!<\/p><p><strong>Those are freelancers and agencies who\u2019ve built their entire workflows around using page builders. <\/strong>That\u2019s tens of thousands of client sites built with page builders.<\/p><p>And right now, Gutenberg isn\u2019t doing anything that\u2019s going to make those workflows change.<\/p><p>Here\u2019s the thing, too:<\/p><p>It\u2019s not enough for Gutenberg to <em>come close<\/em> (or even match those page builders). Gutenberg will need to actually surpass all of these tools to incentivize people to switch. And that\u2019s a tall order because\u2026<\/p><h3>Page builders aren\u2019t just sitting around doing nothing, either<\/h3><p>While Gutenberg is still chugging along in beta, we\u2019ve seen page builders, notably Elementor, keep pushing out more and more features that further distance them from Gutenberg.<\/p><p>For example, Elementor Pro 2.0 now lets users design their entire sites \u2013 header, footer, etc. using Elementor. And Elementor also keeps adding deeper integrations for custom fields and other features to help people <a href=\"https:\/\/www.codeinwp.com\/blog\/best-drag-and-drop-theme-builders\/\">build more complex sites<\/a>.<\/p><p>Here\u2019s, for instance, a new feature in Elementor that allows you to create single page templates and then replace the default ones used in your theme:<\/p><p><img class=\"aligncenter size-full wp-image-38363 amp-wp-enforced-sizes\" src=\"https:\/\/mk0codeinwp10tp0961a.kinstacdn.com\/wp-content\/uploads\/2018\/08\/gutenberg-replace-page-builders2.png\" alt=\"Elementor theme building\" width=\"989\" height=\"591\" \/><\/p><p>Other big names like Divi Builder, Beaver Builder, etc. also keep pushing forward with more control and better user interfaces.<\/p><p>Will Gutenberg get there? <em>Maybe<\/em>. But you have to remember that Gutenberg is always going to be more constrained because it has to be the default editor <strong>for everyone<\/strong>.<\/p><p>Page builders can afford to offer specialized site building workflows. Gutenberg has to work for everyone no matter what.<\/p><h2>Gutenberg won\u2019t get implementers, but it will grab casual users<\/h2><p>While implementers and builders are certainly a big part of the WordPress ecosystem, they aren\u2019t the only people who helped Elementor wrack up <strong>over 1,000,000 active installs<\/strong> in ~2 years.<\/p><p>Casual users rely on page builders, too. And <strong>that\u2019s where we\u2019ll see Gutenberg replace page builders.<\/strong><\/p><h3>Gutenberg eliminates most casual users\u2019 editor pain point<\/h3><p>Right now, the difference between the TinyMCE editor and a page builder is<strong> night and day.<\/strong><\/p><p>Unless you just want to write a basic text\/image post, you\u2019re probably going to seek out an alternative.<\/p><p>And <em>right now<\/em>, that alternative is\u2026<strong>a page builder!<\/strong><\/p><p>For example, there are heaps of affiliate sites who want to style their long-form content and reviews in a way to optimize conversions. And <em>right now<\/em>, most of those webmasters opt for a page builder because <strong>that\u2019s the only way.<\/strong><\/p><p>They don\u2019t care about all the features implementers care about \u2013 they just want good-looking buttons, columns\u2026simple layout stuff:<br \/><\/p><figure id=\"attachment_38362\" class=\"wp-caption aligncenter amp-wp-inline-26943f3a8f9f2b2048c3121525a80414\"><img class=\"size-full wp-image-38362 amp-wp-enforced-sizes\" src=\"https:\/\/mk0codeinwp10tp0961a.kinstacdn.com\/wp-content\/uploads\/2018\/08\/gutenberg-replace-page-builders.png\" alt=\"affiliate examples\" width=\"1138\" height=\"807\" \/><figcaption class=\"wp-caption-text\">This is what affiliates use page builders for \u2013 Gutenberg can do this.<\/figcaption><\/figure><p>With Gutenberg, that pain point largely disappears (<em>especially if Gutenberg can figure out columns<\/em>). Those types of casual users will be able to create better-looking content without needing to turn to an outside page builder. And even if the page builder <em>could <\/em>let them do more, the difference won\u2019t be big enough to pull most people away from the default.<\/p><p>Instead of being night and day. It will be like\u2026dawn and day.<\/p><h3>Third-party Gutenberg add-ons will help pick up the slack<\/h3><p>Gutenberg isn\u2019t even out yet and there\u2019s already a huge marketplace for third-party blocks at WordPress.org.<\/p><p>These add-ons can help fill in the gaps in Gutenberg and make its widget selection equal to, or maybe even better than, the big page builders.<\/p><p>These add-ons won\u2019t be able to fully replicate a page builder, but they\u2019ll give casual users more options and further eliminate the pain point\u2026which gives those people even less of a reason to try and learn a new page builder interface.<\/p><p>Even with Gutenberg in beta, these blocks are doing some neat things. For example, here\u2019s a pricing table block from the Stackable plugin:<\/p><p>Since writing our <a href=\"https:\/\/www.codeinwp.com\/blog\/build-for-gutenberg-how-others-do-it\/\">post on how developers are building for Gutenberg<\/a>, I\u2019ve already seen several new block plugins spring up, as well.<\/p><h3>Other plugin developers have to support Gutenberg \u2013 page builders don\u2019t get that luxury<\/h3><p>Beyond third-party block plugins specifically designed for Gutenberg, Gutenberg also has another big advantage:<\/p><p><strong>Even totally unrelated plugins have to support Gutenberg<\/strong>. That means users will have seamless integrations inside Gutenberg for many of their favorite plugins (<a href=\"https:\/\/www.codeinwp.com\/blog\/build-for-gutenberg-how-others-do-it\/\">some examples here<\/a>).<\/p><p>This is a level of support that page builders just can\u2019t achieve. And because it will eliminate the abstraction of shortcodes, I think it\u2019s just one more reason that casual users will no longer gravitate towards page builders.<\/p><h2>So will Gutenberg replace page builders? It depends on the use<\/h2><p>In the end, I think it comes down to this\u2026<\/p><p>I don\u2019t think it\u2019s a huge step to think that <strong>we\u2019ll see Gutenberg replace page builders for casual users not long after launching.<\/strong><\/p><p>For those people, especially people who are new to WordPress, Gutenberg, along with all the associated third-party plugins and integrations, will do the job well enough that they won\u2019t feel the need to seek out a page builder.<\/p><p>Where I <strong>don\u2019t <\/strong>think we\u2019ll see Gutenberg replace page builders is with site builders and implementers (<em>at least not any time soon<\/em>).<\/p><p>For heavy-duty users, page builders like <a href=\"https:\/\/www.codeinwp.com\/blog\/elementor-vs-divi-builder-vs-beaver-builder\/\">Elementor, Beaver Builder, and Divi<\/a> already have a massive head start. And they aren\u2019t just sitting there waiting for Gutenberg to catch up \u2013 they keep pushing things with site builders, custom fields integrations, etc.<\/p><p>Because of that, I don\u2019t think we see Gutenberg replace page builders with builders and implementers any time soon.<\/p><p><strong><em>Now over to you \u2013 I\u2019m sure people have some strong opinions about this topic. So what about you \u2013 will Gutenberg replace page builders? Why or why not?<\/em><\/strong><\/p><p>Don\u2019t forget to join our crash course on speeding up your WordPress site. With some simple fixes, you can reduce your loading time by even 50-80%:<\/p><div class=\"cwp-ad\"><a href=\"https:\/\/app.monstercampaigns.com\/c\/egsy1own8bpecyi7poxw\" target=\"_blank\"> <img src=\"https:\/\/mk0codeinwp10tp0961a.kinstacdn.com\/wp-content\/themes\/codeinwp2018\/img\/cwp-banner-speed.jpg\" width=\"600\" height=\"306\" alt=\"image\" \/><\/a><\/div>","created_at":"2019-02-17T03:19:26+0000","updated_at":"2019-02-17T03:19:26+0000","published_at":"2018-09-06T14:37:20+0000","published_by":["Colin Newcomer"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"www.codeinwp.com","preview_picture":"https:\/\/mk0codeinwp10tp0961a.kinstacdn.com\/wp-content\/uploads\/2018\/08\/gutenberg-replace-page-builders.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":842,"title":"dakshshah96\/awesome-startup-credits","url":"https:\/\/github.com\/dakshshah96\/awesome-startup-credits","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/twitter.com\/intent\/tweet?text=Check%20out%20Awesome%20Startup%20Credits%E2%80%94A%20collection%20of%20awesome%20companies%20offering%20free\/discounted%20plans%20for%20eligible%20startups!%20https%3A\/\/github.com\/dakshshah96\/awesome-startup-credits\" rel=\"nofollow\">Share on Twitter<\/a><\/p>\n<p>A collection of awesome companies offering free\/discounted plans for eligible startups.<\/p>\n<p><em>In order to be eligible for free credits or a discounted startup plan, most companies require your startup to be a part of an accelerator\/incubator.<\/em><\/p>\n<h2><a id=\"user-content-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#contents\"><\/a>Contents<\/h2>\n<ul><li><a href=\"#advertising\">Advertising<\/a><\/li>\n<li><a href=\"#analytics\">Analytics<\/a><\/li>\n<li><a href=\"#business-suites\">Business Suites<\/a><\/li>\n<li><a href=\"#cloud-computing\">Cloud Computing<\/a><\/li>\n<li><a href=\"#cloud-database\">Cloud Database<\/a><\/li>\n<li><a href=\"#cloud-telephony\">Cloud Telephony<\/a><\/li>\n<li><a href=\"#email-delivery\">Email Delivery<\/a><\/li>\n<li><a href=\"#payments\">Payments<\/a><\/li>\n<li><a href=\"#marketing-and-sales\">Marketing and Sales<\/a><\/li>\n<li><a href=\"#messaging\">Messaging<\/a><\/li>\n<li><a href=\"#miscellaneous\">Miscellaneous<\/a><\/li>\n<\/ul><h3><a id=\"user-content-advertising\" class=\"anchor\" aria-hidden=\"true\" href=\"#advertising\"><\/a>Advertising<\/h3>\n<ul><li><a href=\"https:\/\/developers.snapchat.com\/accelerate\/\" rel=\"nofollow\">Snap Accelerate<\/a> - Snap Accelerate provides companies with the resources and support they need to scale their marketing and customer acquisition strategies with Snap.<\/li>\n<\/ul><h3><a id=\"user-content-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#analytics\"><\/a>Analytics<\/h3>\n<ul><li><a href=\"https:\/\/mixpanel.com\/startups\/\" rel=\"nofollow\">Mixpanel for Startups<\/a> - With access to Mixpanel\u2019s most advanced features, startups can do complex behavioral analysis in seconds and  collaborate across teams to unlock faster growth.<\/li>\n<li><a href=\"https:\/\/segment.com\/industry\/startups\/\" rel=\"nofollow\">Segment Startup Program<\/a> - Eligible startups get full access to Segment's Team Plan (up to $25,000 value) free for up to two years.<\/li>\n<\/ul><h3><a id=\"user-content-business-suites\" class=\"anchor\" aria-hidden=\"true\" href=\"#business-suites\"><\/a>Business Suites<\/h3>\n<ul><li><a href=\"https:\/\/www.zoho.com\/one\/startups.html\" rel=\"nofollow\">Zoho One for Startups<\/a> - Zoho One is a bundle of 40 different applications, for both web and mobile. You can pick any of the apps from this assorted suite, and be business sorted.<\/li>\n<li><a href=\"https:\/\/www.zendesk.com\/startups\/\" rel=\"nofollow\">Zendesk<\/a> - Get a year of Zendesk for free while you get your startup off the ground.<\/li>\n<\/ul><h3><a id=\"user-content-cloud-computing\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloud-computing\"><\/a>Cloud Computing<\/h3>\n<ul><li><a href=\"https:\/\/www.digitalocean.com\/hatch\/\" rel=\"nofollow\">Hatch by DigitalOcean<\/a> - Hatch helps developers create, launch, and scale their startups. This includes a range of perks to get you started, including 12 months of DigitalOcean credit to easily deploy and scale your infrastructure.<\/li>\n<li><a href=\"https:\/\/aws.amazon.com\/activate\/\" rel=\"nofollow\">AWS Activate<\/a> - The AWS Activate program provides startups with the resources they need to quickly get started on AWS \u2013 including credits, training, and support.<\/li>\n<li><a href=\"https:\/\/developer.ibm.com\/startups\/\" rel=\"nofollow\">Startup with IBM<\/a> - With $120,000 in free IBM Cloud credits, Startup with IBM can\nput your business on the path to transformative growth. Disrupt your industry with IBM.<\/li>\n<li><a href=\"https:\/\/cloud.google.com\/developers\/startups\/\" rel=\"nofollow\">Google Cloud for Startups<\/a> - Get up and running quickly and easily with mentorship, training, and Cloud credits from the Google Cloud for Startups program.<\/li>\n<li><a href=\"https:\/\/startups.microsoft.com\/en-us\/\" rel=\"nofollow\">Microsoft for Startups<\/a> - Over the next two years, Microsoft has pledged 500 million dollars to support startups with access to technology, community spaces, and sales opportunities.<\/li>\n<li><a href=\"https:\/\/www.alibabacloud.com\/startup\" rel=\"nofollow\">Create@Alibaba Cloud<\/a> - Create@Alibaba Cloud is a global program focused on accelerating business success for startups.<\/li>\n<li><a href=\"https:\/\/www.oracle.com\/startup\/\" rel=\"nofollow\">Oracle Global Startup Ecosystem<\/a> - The Oracle Global Startup Ecosystem works with startups at various stages, from founders in university-affiliated incubators to early stage to scaleups.<\/li>\n<\/ul><h3><a id=\"user-content-cloud-database\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloud-database\"><\/a>Cloud Database<\/h3>\n<ul><li><a href=\"https:\/\/www.mongodb.com\/startup-accelerator\" rel=\"nofollow\">MongoDB Startup Accelerator<\/a> - The MongoDB Startup Accelerator offers access to MongoDB Atlas, the fully-managed global cloud database, and support to help startups build and scale their businesses.<\/li>\n<li><a href=\"https:\/\/scalegrid.io\/pricing\/offers\/startup-program.html\" rel=\"nofollow\">ScaleGrid Startup Program<\/a> - Save up to $30,000 for a year with 90% off on ScaleGrid's MongoDB, Redis, MySQL and PostgreSQL 'Bring Your Own Cloud' database management plans.<\/li>\n<\/ul><h3><a id=\"user-content-cloud-telephony\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloud-telephony\"><\/a>Cloud Telephony<\/h3>\n<ul><li><a href=\"https:\/\/www.knowlarity.com\/startups\/\" rel=\"nofollow\">Knowlarity for Startups<\/a> - Unicorn dreams? Knowlarity is happy to offer its communications suite products \u2014 SuperReceptionist and SuperConference free of cost for 6 months.<\/li>\n<li><a href=\"https:\/\/ahoy.twilio.com\/startup\" rel=\"nofollow\">Twilio for Startups<\/a> - Twilio has partnered with a variety of accelerators to provide Twilio credits for participating startups.<\/li>\n<li><a href=\"https:\/\/myoperator.co\/startup-program\" rel=\"nofollow\">MyOperator for Startups [INDIA ONLY]<\/a> - MyOperator\u2019s Startup Program will transform the way you and your team communicates with leads and customers. India\u2019s first end-to-end call management solution comes with exclusive startup program products &amp; features.<\/li>\n<li><a href=\"https:\/\/exotel.com\/cloud-telephony-for-startups\/\" rel=\"nofollow\">Exotel for Startups [INDIA ONLY]<\/a> - Exotel for Startups is a special account for startups that gives you access to all of Exotel\u2019s features.<\/li>\n<li><a href=\"https:\/\/msg91.com\/startups\/\" rel=\"nofollow\">MSG91 for Startups [INDIA ONLY]<\/a> - A startup for startups. Grow your startup with MSG91's SMS for startups. First few transactional bulk SMS free every month.<\/li>\n<\/ul><h3><a id=\"user-content-email-delivery\" class=\"anchor\" aria-hidden=\"true\" href=\"#email-delivery\"><\/a>Email Delivery<\/h3>\n<ul><li><a href=\"https:\/\/postmarkapp.com\/for\/bootstrapped-startups#pricing\" rel=\"nofollow\">Postmark<\/a> - If you\u2019ve launched your product, are charging for it, and haven\u2019t taken outside investment, contact the Postmark support team and they'll give you $75 account credit to help with your email costs.<\/li>\n<li><a href=\"https:\/\/sendgrid.com\/accelerate\/\" rel=\"nofollow\">SendGrid Accelerate<\/a> - SendGrid Accelerate is a program to help startups succeed. Startups participating in the SendGrid Accelerate program receive 12 months of product credit, access to mentoring, and networking opportunities exclusive to program participants.<\/li>\n<\/ul><h3><a id=\"user-content-payments\" class=\"anchor\" aria-hidden=\"true\" href=\"#payments\"><\/a>Payments<\/h3>\n<ul><li><a href=\"https:\/\/docs.google.com\/forms\/d\/e\/1FAIpQLScIxDEITVh8no-_tFtaakI8Jql-Q9SXWEJ_iKW16DVWqfbeBw\/viewform\" rel=\"nofollow\">Razorpay [INDIA ONLY]<\/a> - Razorpay presents the ideal plan for startups and budding businesses to kickstart their online payments.<\/li>\n<li><a href=\"https:\/\/www.chargebee.com\/launch\/\" rel=\"nofollow\">Chargebee Launch Plan<\/a> - Chargebee makes managing subscription billing easier for you. With Chargebee\u2019s Launch Plan, you get your first $50K in invoices for free.<\/li>\n<\/ul><h3><a id=\"user-content-marketing-and-sales\" class=\"anchor\" aria-hidden=\"true\" href=\"#marketing-and-sales\"><\/a>Marketing and Sales<\/h3>\n<ul><li><a href=\"https:\/\/www.hubspot.com\/startups\" rel=\"nofollow\">HubSpot<\/a> - Discover how HubSpot can help your startup grow with their CRM, sales, marketing, and customer service software, plus education and support.<\/li>\n<li><a href=\"https:\/\/www.drift.com\/startups\/\" rel=\"nofollow\">Drift<\/a> - Drift is a leading conversational marketing platform and they're giving 86% off to early stage startups.<\/li>\n<\/ul><h3><a id=\"user-content-messaging\" class=\"anchor\" aria-hidden=\"true\" href=\"#messaging\"><\/a>Messaging<\/h3>\n<ul><li><a href=\"https:\/\/www.intercom.com\/early-stage\" rel=\"nofollow\">Intercom for Early Stage<\/a> - Intercom\u2019s user-friendly products help you acquire, onboard, engage and support your customers in a personal way. Eligible companies get all of Intercom's pro products for a flat rate of $49 a month for up to one year.<\/li>\n<\/ul><h3><a id=\"user-content-miscellaneous\" class=\"anchor\" aria-hidden=\"true\" href=\"#miscellaneous\"><\/a>Miscellaneous<\/h3>\n<ul><li><a href=\"https:\/\/www.esri.com\/en-us\/about\/esri-partner-network\/our-partners\/esri-startup-program\" rel=\"nofollow\">Esri Startup Program<\/a> - The Esri Startup Program is a global three-year program that helps startups build mapping and location intelligence into their products and businesses.<\/li>\n<li><a href=\"https:\/\/www.solidworks.com\/solution\/business-segments\/entrepreneurs-startups\" rel=\"nofollow\">SOLIDWORKS<\/a> - For early-stage hardware startups that meet the eligibility criteria, the SOLIDWORKS for Entrepreneurs program offers software, training, and co-marketing resources at no cost to help you succeed.<\/li>\n<li><a href=\"https:\/\/in.mathworks.com\/products\/startups.html\" rel=\"nofollow\">MathWorks<\/a> - Eligible early stage technology startups can get MATLAB, Simulink, and more than 90 industry-specific toolboxes at a startup-friendly price.<\/li>\n<\/ul><h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>Please feel free to make a PR if you know of any other awesome companies which offer free\/discounted plans for startups. Check out <a href=\"https:\/\/github.com\/dakshshah96\/awesome-startup-credits\/blob\/master\/CONTRIBUTING.md\">CONTRIBUTING.md<\/a> for more information.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p><a href=\"https:\/\/creativecommons.org\/publicdomain\/zero\/1.0\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/60561947585c982aee67ed3e3b25388184cc0aa3\/687474703a2f2f6d6972726f72732e6372656174697665636f6d6d6f6e732e6f72672f70726573736b69742f627574746f6e732f38387833312f7376672f63632d7a65726f2e737667\" alt=\"CC0\" data-canonical-src=\"http:\/\/mirrors.creativecommons.org\/presskit\/buttons\/88x31\/svg\/cc-zero.svg\" \/><\/a><\/p>\n<p>To the extent possible under law, <a href=\"https:\/\/daksh.me\" rel=\"nofollow\">Daksh Shah<\/a> has waived all copyright and related or neighboring rights to this work.<\/p>\n<\/article>","created_at":"2019-02-16T22:12:09+0000","updated_at":"2019-02-16T22:12:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/7896438?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":841,"title":"Why Google Needed a Graph Serving System - Dgraph Blog","url":"https:\/\/blog.dgraph.io\/post\/why-google-needed-graph-serving-system\/","content":"<p>When I introduce myself and explain what we are building at Dgraph Labs, I am typically asked if I\nworked at Facebook, or if what I\u2019m building is inspired by Facebook. A lot of people know about the\nefforts at Facebook to serve their social graph, because they have published <a href=\"https:\/\/www.facebook.com\/notes\/facebook-engineering\/tao-the-power-of-the-graph\/10151525983993920\/\">multiple<\/a>\n<a href=\"https:\/\/code.fb.com\/data-infrastructure\/dragon-a-distributed-graph-query-engine\/\">articles<\/a> about the graph infrastructure they put together.<\/p><p>Word from Google has been limited to serving Knowledge Graph, but nothing has been said about the\ninternal infrastructure which makes this happen. There are specialized systems in place at Google to\nserve the knowledge graph.  In fact, we (at Google) placed big bets on graph serving system. I\nmyself put at least two promotions at stake to go work on the new graph thing back in 2010, and see\nwhat we could build.<\/p><p><strong>Google needed to build a graph serving system<\/strong> to serve not just the complex relationships in the\nKnowledge Graph data, but also all the <a href=\"https:\/\/en.ryte.com\/wiki\/Google_OneBox\">OneBoxes<\/a> which had access to structured data. The\nserving system needed to traverse facts, with high enough throughput and low enough latency to be\nhit by a good chunk of <a href=\"https:\/\/google.com\">web search<\/a> queries. No available system or database\nwas able to do all three.<\/p><p>Now that I\u2019ve answered the why, I\u2019ll take the rest of the blog post to walk you through my journey,\nintertwined with Google\u2019s, of building the graph system to serve the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Knowledge_Graph\">Knowledge Graph<\/a> and\n<a href=\"https:\/\/en.ryte.com\/wiki\/Google_OneBox\">OneBoxes<\/a>.<\/p><p>I\u2019ll quickly introduce myself. I worked at Google from 2006 to 2013. First as an intern, then as a\nsoftware engineer in Web Search Infrastructure. In 2010, Google <a href=\"https:\/\/techcrunch.com\/2010\/07\/16\/google-acquires-metaweb-to-make-search-smarter\/\">acquired Metaweb<\/a> and\nmy team had just launched <a href=\"https:\/\/googleblog.blogspot.com\/2010\/06\/our-new-search-index-caffeine.html\">Caffeine<\/a>. I wanted to do something different and started\nworking with Metaweb folks (in SF), splitting my time between San Francisco and Mountain View. My\naim was to figure out how we could use the knowledge graph to improve web search.<\/p><p>There were projects at Google which predate my entrance into graphs. Notably, a project called\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Google_Squared\">Squared<\/a> was built out of the NY office, and there was some talk of Knowledge Cards. These were\nsporadic efforts by individuals\/small teams, but at this time there was no established chain of\ncommand in place, something that would eventually cause me to leave Google. But we\u2019ll get to that\nlater.<\/p><p>As mentioned before, Google acquired Metaweb in <a href=\"https:\/\/techcrunch.com\/2010\/07\/16\/google-acquires-metaweb-to-make-search-smarter\/\">2010<\/a>. Metaweb had built a high-quality\nknowledge graph using multiple techniques, including crawling and parsing Wikipedia, and using a\nWikipedia-like crowd-sourced curation run via <a href=\"https:\/\/developers.google.com\/freebase\/\">Freebase<\/a>. All of this was powered by a graph\ndatabase they had built in-house called Graphd \u2013 a graph daemon (now <a href=\"https:\/\/github.com\/google\/graphd\">published<\/a> on GitHub).<\/p><p>Graphd had some pretty typical properties. Like a daemon, it ran on a single server and all the data\nwas in memory. The entire Freebase website was being run out of Graphd. After the acquisition, one\nof the challenges Google had was to continue to run Freebase.<\/p><p><strong>Google has built an empire on commodity hardware and distributed software.<\/strong> A single server\ndatabase could have never housed the crawling, indexing and serving for Search. Google built SSTable\nand then Bigtable, which could scale horizontally to hundreds or thousands of machines, working\ntogether to serve petabytes of data.  Machines were allocated using Borg (<a href=\"https:\/\/kubernetes.io\/\">K8s<\/a> came out of that\nexperience), communicated using Stubby (<a href=\"https:\/\/grpc.io\/\">gRPC<\/a> came out of that), resolved IP addresses via Borg\nname service (BNS, baked into K8s), and housed their data on Google File System (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Google_File_System\">GFS<\/a>, think\nHadoop FS).  <em>Processes can die and machines can crash, but the system just keeps humming.<\/em><\/p><p>That was the environment in which Graphd found itself. The idea of a database serving an entire\nwebsite running on a single server was alien to Google (myself included).  In particular,\nGraphd needed 64GB or more memory just to function. If you are sneering at the memory requirement,\nnote that this was back in 2010. Most Google servers were maxed at 32GB. In fact, Google had to\nprocure special machines with enough RAM to serve Graphd as it stood.<\/p><h4 id=\"replacement-for-graphd\">Replacement for Graphd<\/h4><p>Ideas were thrown around about how Graphd could be moved or rewritten to work in a distributed way.\nBut see, graphs are hard. They are not key-value databases, where one can just take a chunk of data,\nmove it to another server and when asked for the key, serve it. Graphs promise efficient joins and\ntraversals, which require software to be built in a particular way.<\/p><p>One of the ideas was to use a project called MindMeld (IIRC). The promise was that memory from\nanother server could be accessed much faster via network hardware. This was supposedly faster than\ndoing normal RPCs, fast enough to pseudo-replicate direct memory accesses required by an in-memory\ndatabase. <em>That idea didn\u2019t go very far.<\/em><\/p><p>Another idea, which actually became a project, was to build a truly distributed graph serving system.\nSomething which can not only replace Graphd for Freebase but also serve all knowledge efforts in\nthe future. <strong>This was named, Dgraph \u2013 distributed graph, a spin on Graphd \u2013 graph daemon.<\/strong><\/p><p>If you are now wondering, the answer is <em>yes.<\/em> Dgraph Labs, the company and Dgraph, the open source\nproject are named after this project at Google<sup class=\"footnote-ref\" id=\"fnref:1\"><a href=\"#fn:1\">1<\/a><\/sup>.<\/p><p>For the majority of this blog post, when I refer to Dgraph, I am referring to the project internal\nto Google, not the open source project that we built. But more on <em>that<\/em> later.<\/p><h3 id=\"the-story-of-cerebro-a-knowledge-engine\">The Story of Cerebro: A Knowledge Engine<\/h3><blockquote>\n<p><strong>Inadvertently building a graph serving system.<\/strong><\/p>\n<\/blockquote><p>While I was generally aware of the Dgraph effort to replace Graphd, my goal was to build something\nto improve web search. I came across a research engineer, DH, at Metaweb, who had built\n<a href=\"https:\/\/blog.dgraph.io\/refs\/freebase-cubed.pdf\">Cubed<\/a>.<\/p><p>As I mentioned before, a ragtag of engineers at Google NY had built Google <a href=\"https:\/\/en.wikipedia.org\/wiki\/Google_Squared\">Squared<\/a>. So, doing a\none up, DH built Cubed. While Squared was a dud, Cubed was <em>very<\/em> impressive. I started thinking\nabout how I could build that at Google. Google had bits and pieces of the puzzle that I could\nreadily use.<\/p><p>The first piece of the puzzle was a search project which provided a way to understand which words\nbelonged together with a high degree of accuracy. For example, when you see a phrase like <code>[tom\nhanks movies]<\/code>, it can tell you that <code>[tom]<\/code> and <code>[hanks]<\/code> belong together. Similarly, from <code>[san\nfrancisco weather]<\/code>, <code>[san]<\/code> and <code>[francisco]<\/code> belong together. These are obvious things for humans,\nnot so obvious to machines.<\/p><p>The second piece of the puzzle was to understand grammar. When a query asks for <code>[books by french\nauthors]<\/code>, a machine could interpret it as <code>[books]<\/code> by <code>[french authors]<\/code> (i.e.  books by those\nauthors who are French). But it could also interpret it as <code>[french books]<\/code> by <code>[authors]<\/code> (i.e.\nFrench language books by any author). I used a Part-Of-Speech (POS) tagger by <a href=\"https:\/\/nlp.stanford.edu\/software\/tagger.shtml\">Stanford<\/a> to\nunderstand grammar better and build a tree.<\/p><p>The third piece of the puzzle was to understand entities. <code>[french]<\/code> can mean many things. It could be\nthe country (the region), the nationality (referring to French people), the cuisine (referring to\nthe food), or the language. There was another project I could use to get a list of entities that the\nword or phrase could correspond to.<\/p><p>The fourth piece of the puzzle was to understand the relationship between entities. Now that I knew\nhow to associate the words into phrases, the order in which phrases should be executed, i.e. their\ngrammar, and what entities they could correspond to, I needed a way to find relationships between\nthese entities to create machine interpretations. For example, a query says <code>[books by french\nauthors]<\/code> and POS tells us it is <code>[books]<\/code> by <code>[french authors]<\/code>. We have a few entities for\n<code>[french]<\/code> and few for <code>[authors]<\/code>, and the algorithm needs to determine how they are connected.\nThey could be connected by birth, i.e. authors who were born in France (but could be writing in\nEnglish), or authors who are French nationals, or authors who speak or write French (but could be\nunrelated to France, the country), or authors who just enjoy French food.<\/p><p>To determine if and how the entities are connected, I needed a graph system. Graphd was never going\nto scale to Google levels, and I understood web search. Knowledge Graph data was formatted in\n<a href=\"https:\/\/www.w3.org\/TR\/rdf-concepts\/#section-triples\">Triples<\/a>, i.e. every fact was represented by three pieces, subject (entity), predicate\n(relationship), and object (another entity). Queries had to go from <code>[S P]<\/code> \u2192 <code>[O]<\/code>, from <code>[P\nO]<\/code> \u2192 <code>[S]<\/code>, and sometimes from <code>[S O]<\/code> \u2192 <code>[P]<\/code>.<\/p><p><img src=\"https:\/\/blog.dgraph.io\/images\/google-spo-indexing.png\" alt=\"image\" \/><\/p><p><strong>I used Google\u2019s search index system,<\/strong> assigned a docid to each triple and built three indices,\none each for S, P and O. In addition, the index allowed attachments, so I attached the type\ninformation about each entity (i.e. actor, book, person, etc).<\/p><blockquote>\n<p>I built this graph serving system, with the knowledge that it had the join depth problem (explained\nbelow) and was unsuited for any complex graph queries. In fact, when someone from the Metaweb team\nasked me to make it generally accessible to other teams, I flat-out rejected the idea.<\/p>\n<\/blockquote><p>Now to determine relationships, I\u2019d do queries to see how many results each yielded. Does <code>[french]<\/code>\nand <code>[author]<\/code> yield any results? Pick those results and see how they are connected to <code>[books]<\/code>,\nand so on.  This resulted in multiple machine interpretations of the query. For example, when you\nrun <code>[tom hanks movies]<\/code>, it would produce interpretations like <code>[movies directed by tom hanks]<\/code>,\n<code>[movies starring tom hanks]<\/code>, <code>[movies produced by tom hanks]<\/code>, and automatically reject\ninterpretations like <code>[movies named tom hanks]<\/code>.<\/p><p><img src=\"https:\/\/blog.dgraph.io\/images\/google-spo-intersections.png\" alt=\"image\" \/><\/p><p>With each interpretation, it would generate a list of results \u2013 valid entities in the graph \u2013 and\nit would also return their types (present in attachments). This was extremely powerful because\nunderstanding the type of the results allowed capabilities like filtering, sorting or further\nexpansion. For movie results, you could sort the movies by year of release, length of the\nmovie (short, long), language, awards won, etc.<\/p><p>This project seemed so smart, that we (DH was loosely involved as the knowledge graph expert) named\nit <em>Cerebro,<\/em> after the device of the same name appearing in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cerebro\">X-Men<\/a>.<\/p><p><strong>Cerebro would often reveal very interesting facts<\/strong> that one didn\u2019t originally search for. When\nyou\u2019d run a query like <code>[us presidents]<\/code>, Cerebro would understand that presidents are humans, and\nhumans have height.  Therefore, it would allow you to sort presidents by height and show that\n<em>Abraham Lincoln<\/em> is the tallest US president. It would also let people be filtered by nationality.\nIn this case, it showed America and Britain in the list, because US had one British president,\nnamely <em>George Washington.<\/em> (<em>Disclaimer: Results based on the state of KG at the time; can\u2019t vouch\nfor the correctness of these results.<\/em>)<\/p><h3 id=\"blue-links-vs-knowledge\">Blue Links vs Knowledge<\/h3><p>Cerebro had a chance to truly understand user queries. Provided we had data in the graph, we could\ngenerate machine interpretations of the query, generate the list of results and know a lot about the\nresults to support further exploration. As explained above, once you understand you\u2019re dealing with\nmovies or humans or books, etc., you can enable specific filtering and sorting abilities. You could\nalso do edge traversals to show connected data, going from <code>[us presidents]<\/code> to <code>[schools they went\nto]<\/code>, or <code>[children they fathered]<\/code>. This ability to jump from one list of results to another was\ndemonstrated by DH in another project he\u2019d built called <a href=\"https:\/\/vimeo.com\/1513562\">Parallax<\/a>.<\/p><p>Cerebro was very impressive<sup class=\"footnote-ref\" id=\"fnref:2\"><a href=\"#fn:2\">2<\/a><\/sup> and the leadership at Metaweb was supportive of it. Even the graph\nserving part of it was performant and functional. I referred to it as a <strong>Knowledge Engine<\/strong> (an\nupgrade from search engine). But there was no leadership for knowledge in place at Google. My\nmanager had little interest in it, and after being told to talk to this person and that, I got a\nchance to show it to a very senior leader in search.<\/p><p><strong>The response was not what I was hoping for.<\/strong> For the demo of <code>[books by french authors]<\/code>, the\nsenior search leader showed me Google search results for the query, which showed ten blue links, and\nasserted that Google can do the same thing. Also, they do not want to take traffic away from\nwebsites, because then the owners would be pissed.<\/p><p><em>If you are thinking he was right, consider this:<\/em> When Google does a web search, it does not truly\nunderstand the query. It looks for the right keywords, at the right relative position, the rank of\nthe page, and so on. It is a very complex and extremely sophisticated system, but it does not\ntruly understand either the query or the results. It is up to the user to read, parse, and extract\nthe pieces of information they need from the results, and make further searches to put the full list\nof results together.<\/p><p>For example, for <code>[books by french authors]<\/code>, first one needs to put together an exhaustive list\nwhich might not be available in a single web page. Then sort these books by year of publication, or\nfilter by publication houses, etc. \u2014 all those would require a lot of link following, further\nsearches, and human aggregation of results. Cerebro had the power to cut down on all that effort and\nmake the user interaction simple and flawless.<\/p><p>However, this was the typical approach towards knowledge back then. Management wasn\u2019t sure of the\nutility of the knowledge graph, or how and in what capacity that should be involved with\nsearch.  <strong>This new way of approaching knowledge was not easily digestible by an organization that\nhad been so massively successful by providing users with links to web pages.<\/strong><\/p><p>After butting heads with the management for a year, I eventually lost interest in continuing. A\nmanager from the Google Shanghai office reached out to me, and I handed the project over to him in June\n2011. He put a team of 15 engineers on the project. I spent a week in Shanghai transferring whatever\nI had built and learned to the engineers. DH was involved as well and he guided the team for the\nlong term.<\/p><h3 id=\"join-depth-problem\">Join-Depth Problem<\/h3><p>The graph serving system I had built for Cerebro had a join-depth problem. A join is executed when\nthe result set from an earlier portion of the query is needed to execute a later portion of it. A\ntypical join would involve some <code>SELECT<\/code>, i.e. a filter in certain results from the universal data\nset, then using these results to filter against another portion of the dataset. I\u2019ll illustrate with\nan example.<\/p><p>Say, you want to know <code>[people in SF who eat sushi]<\/code>. Data is sharded by people, and has information\nabout who lives in which city and what food they eat.<\/p><p>The above query is a single-level join. If an application external to a database was executing this,\nit would do one query to execute the first step. Then execute multiple queries (one query for each\nresult), to figure out what each person eats, picking only those who eat sushi.<\/p><p>The second step suffers from a fan-out problem. If the first step has a million results (population\nof San Francisco), then the second step would need to put each result into a query, retrieving their\neating habit, followed by a filter.<\/p><p><img src=\"https:\/\/blog.dgraph.io\/images\/google-fan-out.png\" alt=\"image\" \/><\/p><p>Distributed system engineers typically solve this by doing a <em>broadcast.<\/em> They would batch up the\nresults, corresponding to their sharding function, and make a query to each server in the cluster.\nThis would give them a join, but cause query latency issues.<\/p><p>Broadcasts in a distributed system are bad. This issue is best explained by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Jeff_Dean_(computer_scientist)\">Jeff Dean<\/a> of\nGoogle in his \u201cAchieving Rapid Response Times in Large Online Services\u201d talk (<a href=\"https:\/\/www.youtube.com\/watch?v=1-3Ahy7Fxsc\">video<\/a>,\n<a href=\"https:\/\/static.googleusercontent.com\/media\/research.google.com\/en\/\/pubs\/archive\/44875.pdf\">slides<\/a>). The overall latency of a query is always greater than the latency of the slowest\ncomponent. <strong>Small blips on individual machines cause delays, and touching more machines per query\nincreases the likelihood of delays dramatically.<\/strong><\/p><blockquote>\n<p>Consider a server with 50th-%ile latency of 1ms, but 99th-%ile latency of 1s. If a query only\ntouches one server, only 1% of requests take over a second. But if the query touches 100 of these\nservers, 63% of requests would take over a second.<\/p>\n<\/blockquote><p>Thus, broadcasts to execute one query are bad for query latency. Now consider if two, three or more\njoins need to happen. They would get too slow for real-time (OLTP) execution.<\/p><p>This high fan-out broadcast problem is shared by most non-native graph databases, including <a href=\"https:\/\/github.com\/JanusGraph\/janusgraph\">Janus\ngraph<\/a>, <a href=\"https:\/\/blog.twitter.com\/engineering\/en_us\/a\/2010\/introducing-flockdb.html\">Twitter\u2019s FlockDB<\/a>, and <a href=\"https:\/\/www.facebook.com\/notes\/facebook-engineering\/tao-the-power-of-the-graph\/10151525983993920\/\">Facebook\u2019s TAO<\/a>.<\/p><p>Distributed joins are a hard problem. Existing native graph databases avoid this problem by keeping\nthe universal dataset within one machine (standalone DB), and doing all the joins without touching\nother servers, e.g. <a href=\"https:\/\/github.com\/neo4j\/neo4j\">Neo4j<\/a>.<\/p><h3 id=\"enter-dgraph-an-arbitrary-depth-join-engine\">Enter Dgraph: An Arbitrary Depth Join Engine<\/h3><p>After wrapping up Cerebro, and having experience building a graph serving system, I got involved\nwith the Dgraph project, becoming one of the three tech leads on the project. The concepts involved\nin Dgraph design were novel and solved the join-depth problem.<\/p><p>In particular, Dgraph sharded the graph data in a way where each join can be executed entirely by\none machine. Going back to <code>subject-predicate-object<\/code> (SPO), each instance of Dgraph would hold all\nthe subjects and objects corresponding to each predicate in that instance. Multiple predicates would\nbe stored on the instance, each predicate being stored in its entirety.<\/p><p><strong>This allowed the queries to execute arbitrary depth joins<\/strong> while avoiding the fan-out broadcast\nproblem. Taking the query <code>[people in SF who eat sushi]<\/code>, this would result in <em>at most two network\ncalls<\/em> within the database irrespective of the size of the cluster. The first call would find all\npeople who live in SF. The second call would send this list of people and intersect with all the\npeople who eat sushi. We can then add more constraints or expansions, each step would still involve\nat most one more network call.<\/p><p><img src=\"https:\/\/blog.dgraph.io\/images\/google-dgraph-query.png\" alt=\"image\" \/><\/p><p>This introduces the problem of very large predicates located on a single server, but that can be\nsolved by further splitting up a predicate among two or more instances as the size grows. Even then,\na single predicate split across the entire cluster would be the worst-case behavior in only the most\nextreme cases where all the data corresponds to just one predicate. In the rest of the cases, the\ndesign of sharding data by predicates works better to achieve much faster query latency in\nreal-world systems.<\/p><p>Sharding was not the only innovation in Dgraph. All objects were assigned <code>integer<\/code> IDs and were\nsorted and stored in a posting list structure to allow for quick intersections of posting lists.\nThis would allow fast filtering during joins, finding common references, etc. Ideas from Google\u2019s\nweb serving system were involved.<\/p><h3 id=\"uniting-oneboxes-via-plasma\">Uniting OneBoxes via Plasma<\/h3><p><strong>Dgraph at Google was not a database.<\/strong> It was a serving system, the equivalent of Google\u2019s web\nsearch serving system. In addition, it was meant to react to live updates. As a real-time updating\nserving system, it needed a real-time graph indexing system. I had a lot of experience with\nreal-time incremental indexing systems having worked on <a href=\"https:\/\/googleblog.blogspot.com\/2010\/06\/our-new-search-index-caffeine.html\">Caffeine<\/a>.<\/p><p>I started a project to unite all Google OneBoxes under this graph indexing system, which involved\nweather, flights, events, and so on. You might not know the term, but you definitely have seen them.\n<a href=\"https:\/\/en.ryte.com\/wiki\/Google_OneBox\">OneBox<\/a> is a separate display box which gets shown when certain types of queries are run,\nwhere Google can return richer information. To see one in action, try [<a href=\"https:\/\/www.google.com\/search?q=weather+in+sf\">weather in\nsf<\/a>].<\/p><p>Before this project, each OneBox was being run by standalone backends and maintained by different\nteams. <strong>There was a rich set of structured data, but no data was being shared between the boxes.<\/strong>\nNot only was maintaining all these backends a lot of work operationally but also the lack of\nknowledge share limited the kind of queries Google could respond to.<\/p><p>For example, <code>[events in SF]<\/code> can show the events, and <code>[weather in SF]<\/code> can show the weather. But if\n<code>[events in SF]<\/code> could understand that the weather is rainy and know whether events are indoors or\noutdoors, it can filter (or at least sort) the events based on weather (<em>in a heavy rainstorm, maybe a\nmovie or a symphony is the best option<\/em>).<\/p><p>Along with help from the Metaweb team, we started converting all this data into the SPO format and\nindexing it under one system. I named the system <strong>Plasma, a real-time graph indexing system<\/strong> for the\ngraph serving system, Dgraph.<\/p><h3 id=\"management-shuffle\">Management Shuffle<\/h3><p>Like Cerebro, Plasma was an under-funded project but kept on gaining steam.  Eventually, when\nmanagement came to realize that OneBoxes were imminently moving to this project, they needed the\n<em>\u201cright people\u201d<\/em> in charge of Knowledge. In the midst of that game of politics, I saw three different\nmanagement changes, each with zero prior experience with graphs.<\/p><p>During this shuffle, Dgraph was considered <em>too complex<\/em> by management who supported <a href=\"https:\/\/en.wikipedia.org\/wiki\/Spanner_(database)\">Spanner<\/a>,\na globally distributed SQL database which needs GPS clocks to ensure global consistency. <em>The irony\nof this is still mind-boggling.<\/em><\/p><p>Dgraph got canceled, Plasma survived. And a\nnew team under a new leader was put in charge, with a hierarchy in place reporting to the CEO.\nThe new team \u2013 with little understanding or knowledge of graph issues \u2013 decided to build a serving\nsystem based on Google\u2019s existing search index (<em>like I had done for Cerebro<\/em>). I suggested using the\nsystem I had already built for Cerebro, but that was rejected. I changed Plasma to crawl and\nexpand each knowledge topic out a few levels, so this system could treat it like a web document.\nThey called it TS <em>(name abbreviated)<\/em>.<\/p><p><strong>This meant that the new serving system would not be able to do any deep joins.<\/strong> A curse of a decision\nthat I see in many companies because engineers start with the wrong idea that <em>\u201cgraphs are a simple\nproblem that can be solved by just building a layer on top of another system.\u201d<\/em><\/p><p>After a few more months, I left Google in May 2013, having worked on Dgraph\/Plasma for just about\ntwo years.<\/p><h3 id=\"after-story\">After Story<\/h3><ul><li><p>A few years later, Web Search Infrastructure was renamed to Web Search and Knowledge Graph\nInfrastructure and the leader to whom I had demoed Cerebro came to lead the Knowledge effort,\ntalking about how they intend to replace blue links with knowledge \u2014 directly answering user\nqueries as much as possible.<\/p><\/li>\n<li><p>When the team in Shanghai working on Cerebro was close to getting it into production, the project\ngot pulled from under their feet and moved to the Google NY office. Eventually, it got launched as\nKnowledge Strip. If you search for [<a href=\"https:\/\/www.google.com\/search?q=tom+hanks+movies\">tom hanks\nmovies<\/a>], you\u2019ll see it at the top. It has sort of\nimproved since the initial launch, but still does not support the level of filtering and sorting\noffered by Cerebro.<\/p><\/li>\n<li><p>All the three tech leads (including me) who worked on Dgraph eventually left Google. To the best of\nmy knowledge, the other two leads are now working at Microsoft and LinkedIn.<\/p><\/li>\n<li><p>I did manage to get the two promotions and was due for a third one when I left Google as a Senior\nSoftware Engineer.<\/p><\/li>\n<li><p>Based on some anecdotal knowledge, the current version of TS is actually very close to Cerebro\u2019s\ngraph system design, with subject, predicate and object each having an index. <strong>It, therefore,\ncontinues to suffer from the join-depth problem.<\/strong><\/p><\/li>\n<li><p>Plasma has since been rewritten and renamed, but still continues to act as a real-time graph\nindexing system, supporting TS. Together, they continue to host and serve all structured data\nat Google, including Knowledge Graph.<\/p><\/li>\n<li><p><strong>Google\u2019s inability to do deep joins is visible in many places.<\/strong> For one, we still don\u2019t see a\nmarriage of the various data feeds of OneBoxes: <code>[cities by most rain in asia]<\/code> does not produce a\nlist of city entities despite that weather and KG data are readily available (instead, the result is\na quote from a web page); <code>[events in SF]<\/code> cannot be filtered based on weather; <code>[US presidents]<\/code>\nresults cannot be further sorted, filtered, or expanded to their children or schools they attended.\nI suspect this was also one of the reasons to discontinue <a href=\"https:\/\/developers.google.com\/freebase\/\">Freebase<\/a>.<\/p><\/li>\n<\/ul><h3 id=\"dgraph-a-phoenix\">Dgraph: A Phoenix<\/h3><p>Two years after leaving Google, I decided to <a href=\"https:\/\/github.com\/dgraph-io\/dgraph\">build Dgraph<\/a>. Outside of Google, I witnessed\nthe same indecision about graph systems as I did inside. The graph space had a lot of half-baked\nsolutions, in particular, a lot of custom solutions, hastily put together on top of relational or\nNoSQL databases, or as one of the many features of multi-model databases. If a native solution\nexisted, it suffered from scalability issues.<\/p><p>Nothing I saw had a coherent story with a performant, scalable design at heart. <strong>Building a\nhorizontally scalable, low-latency graph database with arbitrary-depth joins is an extremely hard\nproblem<\/strong> and I wanted to ensure that we built <a href=\"https:\/\/github.com\/dgraph-io\/dgraph\">Dgraph<\/a> the right way.<\/p><p>The <a href=\"https:\/\/github.com\/dgraph-io\/dgraph\">Dgraph<\/a> team has spent the last three years \u2014 not only learning from my own\nexperience but also putting a lot of original research into the design \u2014 building a graph\ndatabase unparalleled in the market. So companies have the choice of using a robust, scalable, and\nperformant solution instead of a yet another half baked one.<\/p><p>Thanks for reading this post. If you liked this post, show us some love and give us <a href=\"https:\/\/github.com\/dgraph-io\/dgraph\">a star on GitHub<\/a>.<\/p><hr \/><div class=\"engage\"><p><strong>We are building a fast, transactional and distributed graph database. If you like what\n      you read above, <a href=\"https:\/\/dgraph.io\/about\">come join<\/a> the team!<\/strong><\/p><table><thead><tr><\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/docs.dgraph.io\" target=\"_blank\">Get started with Dgraph<\/a><\/td>\n      <td><a href=\"https:\/\/docs.dgraph.io\" target=\"_blank\">https:\/\/docs.dgraph.io<\/a><\/td>\n    <\/tr><tr><td><a href=\"https:\/\/github.com\/dgraph-io\/dgraph\" target=\"_blank\">Star us on GitHub<\/a><\/td>\n      <td><a href=\"https:\/\/github.com\/dgraph-io\/dgraph\" target=\"_blank\">https:\/\/github.com\/dgraph-io\/dgraph<\/a><\/td>\n    <\/tr><tr><td><a href=\"https:\/\/discuss.dgraph.io\" target=\"_blank\">Ask us questions<\/a><\/td>\n      <td><a href=\"https:\/\/discuss.dgraph.io\" target=\"_blank\">https:\/\/discuss.dgraph.io<\/a><\/td>\n    <\/tr><\/tbody><\/table><strong>Fortune 500 companies are using Dgraph in production. If you need production support,\n    <a href=\"https:\/\/blog.dgraph.io\/cdn-cgi\/l\/email-protection#86ebe7e8eff5eec6e2e1f4e7f6eea8efe9\">talk to us.<\/a><\/strong><\/div><p><em>P.S. Special thanks to Daniel Mai for helping put together the graphics in the blog post.<\/em><\/p><div class=\"footnotes\"><hr \/><ol><li id=\"fn:1\"><p>Dgraph is a trademark of Dgraph Labs, Inc. (<em>FWIW, Google does not release projects\n  externally with the same name as they are referred to internally.<\/em>)<\/p>\n <a class=\"footnote-return\" href=\"#fnref:1\"><sup>[return]<\/sup><\/a><\/li>\n<li id=\"fn:2\"><p>Some more queries powered by Cerebro:\n  <code>[female african american politicians]<\/code>,\n  <code>[bollywood actors married to politicians]<\/code>,\n  <code>[children of us presidents]<\/code>,\n  <code>[movies starring tom hanks released in the 90s]<\/code><\/p>\n <a class=\"footnote-return\" href=\"#fnref:2\"><sup>[return]<\/sup><\/a><\/li>\n<\/ol><\/div>","created_at":"2019-02-16T16:07:41+0000","updated_at":"2019-02-16T16:07:41+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":21,"domain_name":"blog.dgraph.io","preview_picture":"https:\/\/blog.dgraph.io\/images\/google-kg2.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["slack","irc"],"is_public":false,"id":840,"title":"Slirc: IRC interface to Slack","url":"https:\/\/www.dlbeer.co.nz\/articles\/slirc\/index.html","content":"<p>15 May 2018<\/p><p>If you need to communicate with other people via Slack, but would prefer to avoid the web interface and Electron application, they offer an IRC gateway. Unfortunately, the gateway has some shortcomings that make it a fairly unsatisfactory solution:<\/p><ul><li><p>The away status of other users is not shown in <code>\/WHOIS<\/code> queries.<\/p><\/li>\n<li><p>Channels that you are a member of (as shown in the web interface) are not joined automatically, and if you don't join them explicitly, you won't see messages from them. Other users who are trying to talk to you may not expect this.<\/p><\/li>\n<li><p>Pasting is cumbersome: the normal way that web interface users will paste text snippets leads to you receiving a message containing only a URL directing you to the Slack web interface in order to view the pasted text.<\/p><\/li>\n<li><p>If you have email notifications enabled, they will continue to come through while you're connected, and will stack up endlessly, even if you are not marked as away. This is because messages are not marked as read until you disconnect the IRC client. The rationale given for this is that they have no way of knowing if you've actually seen the message in the IRC client (but then why mark them as read once when the client disconnects?).<\/p><\/li>\n<li><p>The IRC gateway is not available for all users.<\/p><\/li>\n<\/ul><p>I ended up implementing my own gateway in the form of a locally-running Perl script that communicates with Slack's RTM interface. This has numerous advantages over the official interface:<\/p><ul><li><p>Away status is reported correctly in <code>\/WHOIS<\/code> and <code>\/WHO<\/code> queries.<\/p><\/li>\n<li><p>Channels that you're in are force-joined on connection. If you leave a channel in IRC, you leave it via the API too (making the change visible to other users).<\/p><\/li>\n<li><p>Pasted text snippets are fetched and displayed inline as ordinary messages.<\/p><\/li>\n<li><p>Messages are marked as read on receipt in the normal way.<\/p><\/li>\n<li><p>It doesn't require any intervention on the part of the workspace owner to use. It requires a Slack API key, but you can obtain a usable key on your own, in spite of what the Slack documentation claims.<\/p><\/li>\n<li><p>Being a self-contained script, it's easy to customize if you don't like the way it works.<\/p><\/li>\n<\/ul><h2 id=\"installation-and-setup\">Installation and setup<\/h2><p>Download the script here:<\/p><p>To run it, you will first need to install the prerequisite modules from CPAN. These are:<\/p><ul><li><code>AnyEvent<\/code><\/li>\n<li><code>AnyEvent::HTTP<\/code><\/li>\n<li><code>AnyEvent::Socket<\/code><\/li>\n<li><code>AnyEvent::WebSocket::Client<\/code><\/li>\n<li><code>URI::Encode<\/code><\/li>\n<li><code>Data::Dumper<\/code><\/li>\n<li><code>JSON<\/code><\/li>\n<\/ul><p>On most Linux systems, you will be able to install each of these with:<\/p><pre>sudo cpan &lt;module-list&gt;<\/pre><p>Make sure you have OpenSSL development headers installed before attempting to install the socket module.<\/p><p>You then need to create a configuration file. This should contain the following lines:<\/p><pre>slack_token=&lt;API key&gt;\npassword=&lt;Password for IRC server&gt;\nport=&lt;Port for IRC server&gt;<\/pre><p>Save your configuration file and run the gateway with:<\/p><pre>.\/slirc-20180515.pl &lt;file.conf&gt;<\/pre><p>Since 15 May 2018, you can also listen on a Unix domain socket. To use this feature, add the following line to your config file:<\/p><pre>unix_socket=&lt;path&gt;<\/pre><p>This feature was added by <a href=\"https:\/\/www.chiark.greenend.org.uk\/~cjwatson\/\">Colin Watson<\/a>.<\/p><p>You should then be able to connect using the IRC client of your choice on the configured port using the password specified in the configuration file. The IRC server does not support TLS, and so doesn't bind to any address other than 127.0.0.1. If you need access to it across the network, you should use a TLS proxy or a TLS-capable IRC bouncer on the same host.<\/p><h3 id=\"obtaining-an-api-key\">Obtaining an API key<\/h3><p>You can obtain a \"legacy\" API key by requesting it via the web interface from your workspace owner. However, you can also obtain one yourself, since the web interface uses the RTM API.<\/p><p>To do this, log in via the web interface with your web browser's network request tracing enabled and wait. Eventually, the chat interface will be ready, and you should see a series of requests to the API. Pick any of them and look in the POST body for a parameter called <code>token<\/code>. As long as you don't explicitly log out (clearing cookies is fine), this key will remain usable.<\/p><p>A simple Perl script using <code>LWP::UserAgent<\/code> automates the process of extracting API keys from non-SSO login credentials:<\/p><h2 id=\"using-the-gateway\">Using the gateway<\/h2><p>Multiple IRC clients can be connected simultaneously. Messages from any client will be sent via the Slack API. Messages and other events received from the Slack API will be broadcast to all connected IRC clients.<\/p><p>If the RTM connection goes down, all IRC clients will be disconnected and will need to reconnect. IRC clients that connect before the RTM connection is available will not be sent welcome messages until the connection is up.<\/p><p>A ChanServ-like user called X is present on the server. It supports the following commands:<\/p><dl><dt><code>\/msg X newgroup &lt;name&gt;<\/code><\/dt>\n<dd>Create a new private group. Users can be invited to the group via the usual <code>\/INVITE<\/code> command.\n<\/dd>\n<dt><code>\/msg X newchan &lt;name&gt;<\/code><\/dt>\n<dd>Create a new public channel.\n<\/dd>\n<dt><code>\/msg X archive &lt;name&gt;<\/code><\/dt>\n<dd>Archive the given group or channel.\n<\/dd>\n<dt><code>\/msg X cat &lt;file-id&gt;<\/code><\/dt>\n<dd>Fetch the given file and display it inline in the form of <code>NOTICE<\/code> messages.\n<\/dd>\n<dt><code>\/msg X disconnect<\/code><\/dt>\n<dd>Disconnect and reconnect the RTM client.\n<\/dd>\n<\/dl><p>Mentions of the form <code>&lt;@nickname&gt;<\/code> or <code>&lt;#channel&gt;<\/code> will be translated and appear correctly to clients using the web interface.<\/p><h2 id=\"copyright\">Copyright<\/h2><p>Copyright (C) 2017 Daniel Beer &lt;<a href=\"mailto:dlbeer@gmail.com\">dlbeer@gmail.com<\/a>&gt;<\/p><p>Permission to use, copy, modify, and\/or distribute this software for any purpose with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies.<\/p><p>THE SOFTWARE IS PROVIDED \"AS IS\" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.<\/p>","created_at":"2019-02-16T15:20:55+0000","updated_at":"2019-02-16T15:20:59+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"www.dlbeer.co.nz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["fediverse","matrix","riot","slack"],"is_public":false,"id":839,"title":"Using Matrix to connect to Slack from an IRC client on Ubuntu \u2013 Andy Balaam's Blog","url":"https:\/\/www.artificialworlds.net\/blog\/2018\/06\/08\/using-matrix-to-connect-to-slack-from-an-irc-client-on-ubuntu\/","content":"<p><strong>Update:<\/strong> I found a better solution <a href=\"https:\/\/www.artificialworlds.net\/blog\/2018\/06\/15\/connecting-to-slack-from-an-irc-client-using-slirc\/\">using slirc<\/a>.<\/p>\n<p>I like using <a href=\"https:\/\/hexchat.github.io\/\">HexChat<\/a> to talk to my colleagues, like <a href=\"https:\/\/xkcd.com\/1782\/\">one other guy<\/a>.  It is fast, and it pops up a new window when someone sends me a direct\/private message.<\/p>\n<p>Recently, <a href=\"https:\/\/drewdevault.com\/2015\/11\/01\/Please-stop-using-slack.html\">Slack<\/a> shut down their <a href=\"https:\/\/www.theregister.co.uk\/2018\/03\/09\/slack_cuts_ties_to_irc_and_xmpp\/\">IRC gateway<\/a>, forcing me to use their slow UI, and making me unable to be responsive to direct messages without spending my entire life checking said UI, or being disturbed by notifications.<\/p>\n<p>So, I decided to set up a <a href=\"https:\/\/matrix.org\">Matrix<\/a> bridge, because how hard can it be?<\/p>\n<p>My system is Ubuntu MATE 18.04.<\/p>\n<h2>What I did<\/h2>\n<h3>Install Synapse<\/h3>\n<p>Install <a href=\"https:\/\/github.com\/matrix-org\/synapse\">Synapse<\/a>, a Matrix homeserver.<\/p>\n<pre>wget -qO - https:\/\/matrix.org\/packages\/debian\/repo-key.asc | sudo apt-key add -&#13;\nsudo apt-add-repository http:\/\/matrix.org\/packages\/debian\/&#13;\nsudo apt update&#13;\nsudo apt install matrix-synapse<\/pre>\n<p>Edit \/etc\/matrix-synapse\/homeserver.yaml to change the line containing \u201cregistration_shared_secret\u201d to look something like:<\/p>\n<pre>registration_shared_secret: \"some secret password you made up\"<\/pre>\n<p>(Replacing \u201csome secret password you made up\u201d with something secret.)<\/p>\n<p>Make a new user:<\/p>\n<pre>register_new_matrix_user -c \/etc\/matrix-synapse\/homeserver.yaml http:\/\/localhost:8008<\/pre>\n<p>Enter a username and password, and take a note of them.<\/p>\n<h3>Install Riot<\/h3>\n<p>(Technically, this step is optional, but I definitely recommend it to check everything is working.)<\/p>\n<p>Install <a href=\"https:\/\/riot.im\/\">Riot<\/a>, a client that can talk to a homeserver.<\/p>\n<pre>curl https:\/\/riot.im\/packages\/debian\/repo-key.asc | sudo apt-key add -&#13;\nsudo apt-add-repository https:\/\/riot.im\/packages\/debian\/&#13;\nsudo apt update&#13;\nsudo apt install riot-web<\/pre>\n<p>Start Riot from the menu or by typing <code>riot-web<\/code> in a console.<\/p>\n<p>Choose to log in to a Custom Server (not create a new user), and enter the Home Server URL as <code>http:\/\/localhost:8008<\/code>, leave the Identity Server URL as it is, and enter the username and password you entered in the previous step.<\/p>\n<p>Riot should be able to connect to the Synapse server, even though it won\u2019t have any rooms in it yet.<\/p>\n<p>If this doesn\u2019t work, try running <code>riot-web<\/code> in a console, and check the Synapse log file at <code>\/var\/log\/matrix-synapse\/homeserver.log<\/code><\/p>\n<h3>Install matrix-puppet-slack<\/h3>\n<p>(Note: after this step, the homeserver can only be for you now, because it is logged into Slack as you.  If you want to make a new Matrix server for lots of people, I found that <a href=\"https:\/\/github.com\/matrix-org\/matrix-appservice-slack\">matrix-appservice-slack<\/a> works, but this only connects channels, not direct messages.)<\/p>\n<p>Get a <a href=\"https:\/\/api.slack.com\/custom-integrations\/legacy-tokens\">Slack legacy token<\/a> which matrix-puppet-slack will use to connect to Slack as you.  Follow the instructions given at that link, and you should end up with a token that looks something like \u201cabcd-123456768-ETC-ETC\u201d.  Keep a note of it.<\/p>\n<p>Install <a href=\"https:\/\/github.com\/matrix-hacks\/matrix-puppet-slack\">matrix-puppet-slack<\/a> which allows you to connect a homeserver to a Slack instance, pretending it is you:<\/p>\n<pre>cd&#13;\ngit clone https:\/\/github.com\/matrix-hacks\/matrix-puppet-slack.git&#13;\ncd matrix-puppet-slack\/&#13;\nnpm install&#13;\ncp config.sample.json config.json<\/pre>\n<p>Edit config.json so it looks something like this:<\/p>\n<pre>{&#13;\n  \"slack\": [{&#13;\n    \"team_name\": \"Name of my team\",&#13;\n    \"user_access_token\": \"abcd-123456768-ETC-ETC\"&#13;\n  }],&#13;\n  \"registrationPath\": \"slack-registration.yaml\",&#13;\n  \"port\": 8090,&#13;\n  \"bridge\": {&#13;\n    \"homeserverUrl\":\"http:\/\/localhost:8008\",&#13;\n    \"domain\": \"localhost\",&#13;\n    \"registration\": \"slack-registration.yaml\"&#13;\n  }&#13;\n}<\/pre>\n<p>Now generate a config file that Synapse will use, called slack-registration.yaml, by typing this command (still in the matrix-puppet-slack directory):<\/p>\n<pre>node index.js -r -u \"http:\/\/localhost:8090\"<\/pre>\n<p>When asked, type in the username and password you entered in the first step (when you ran the register_new_matrix_user command).<\/p>\n<p>This creates a config file telling Synapse that matrix-puppet-slack is running on the local machine on port 8090.<\/p>\n<p>Now, run it:<\/p>\n<pre>node index.js<\/pre>\n<p>and edit Synapse\u2019s config file <code>\/etc\/matrix-synapse\/homeserver.yaml<\/code> to point at the new config file you just made, by editing the line mentioning app_service_config_files so that it looks like this:<\/p>\n<pre>app_service_config_files: [&#13;\n    \"\/path\/to\/matrix-puppet-slack\/slack-registration.yaml\"&#13;\n]<\/pre>\n<p>Make sure you replace \u201c\/path\/to\/matrix-puppet-slack\u201d in the above with the directory where you put matrix-puppet-slack, which is \/home\/yourusername\/matrix-puppet-slack if you followed the instructions exactly.<\/p>\n<p>Restart Synapse, and when you go back into Riot, you should see your Slack channels appear gradually (as people talk in them):<\/p>\n<pre>sudo service matrix-synapse restart<\/pre>\n<p>If not, check the output from the <code>node index.js<\/code> command, and the Synapse log file at <code>\/var\/log\/matrix-synapse\/homeserver.log<\/code>.<\/p>\n<h3>Install matrix-ircd<\/h3>\n<p>To allow connecting to Matrix from an IRC client, install <a href=\"https:\/\/github.com\/matrix-org\/matrix-ircd\">matrix-ircd<\/a>:<\/p>\n<pre>cd&#13;\ngit clone https:\/\/github.com\/matrix-org\/matrix-ircd.git&#13;\ncd matrix-ircd&#13;\ncargo build&#13;\ncargo run -- --url \"http:\/\/localhost:8008\"<\/pre>\n<p>(Note this works on Ubuntu 18.04, but operating systems with older versions of Rust may need to get the latest first \u2013 see the <a href=\"https:\/\/github.com\/matrix-org\/matrix-ircd\">matrix-ircd<\/a> page for details.)<\/p>\n<h3>Connect from an IRC client<\/h3>\n<p>Now use an IRC client to connect to matrix-ird.  I recommend <a href=\"https:\/\/hexchat.github.io\/\">HexChat<\/a> if you\u2019re looking for one.<\/p>\n<p>matrix-ircd runs on port 5999 by default, so you should be able to connect to it from your IRC client by setting the server to <code>localhost\/5999<\/code> and using the username and password from the first step (when you ran the register_new_matrix_user command).  Use plain username\/password authentication.<\/p>\n<p>If it works, you should see your Slack channels appear in your IRC client.  If not check the logs mentioned before, and the error messages and logs from your IRC client.<\/p>\n<h2>What works and doesn\u2019t work<\/h2>\n<ul><li>Messages typed by me and others into IRC, Riot and Slack appear in all the other places, including direct messages.<\/li>\n<li>Synapse and Riot are installed fully, so Synapse will be running after a reboot, and Riot is available in the menus, but matrix-puppet-slack and matrix-ircd are just running in a console, and have to be manually started.  It should be reasonably simple to make custom systemd files to start them automatically.<\/li>\n<li>matrix-puppet-slack crashed once and I had to restart it, so this may not be very reliable.  I logged <a href=\"https:\/\/github.com\/matrix-hacks\/matrix-puppet-slack\/issues\/40\">an issue<\/a>.<\/li>\n<li>The Slack channel names look terrible in my IRC client.  I think matrix-ircd needs to find the pretty names (which are displayed correctly in Riot) and use them instead of the coded names used under the cover in Matrix.<\/li>\n<li>My first ever direct message to someone does not work, even though I can choose them and attempt to send a message.  As soon as they say something to me, they appear as a channel in Riot and IRC, and I can talk back and forth no problem from then on.<\/li>\n<li>Direct messages look like normal channels in Matrix, which means I can\u2019t use HexChat to pop up notifications for them, so this was all pointless.<\/li>\n<\/ul><p>Please leave comments below if you know how I can fix any of these problems.<\/p>\n<h2>Conclusions<\/h2>\n<ul><li>Matrix seems really cool<\/li>\n<li>It is basically possible to bridge IRC-Matrix-Slack as I wanted<\/li>\n<li>But, some remaining bugs mean I have to keep using Slack\u2019s UI for now :-(<\/li>\n<li>Please comment telling me how to do this better<\/li>\n<\/ul>","created_at":"2019-02-16T15:19:03+0000","updated_at":"2019-02-16T15:19:10+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"www.artificialworlds.net","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":838,"title":"matrix-org\/matrix-appservice-irc","url":"https:\/\/github.com\/matrix-org\/matrix-appservice-irc\/wiki\/End-user-FAQ","content":"<p><em>Frequently asked questions by people who aren't running the IRC bridge by themselves.<\/em><\/p>\n<h2>\n<a id=\"user-content-what-do-i-do-if-i-already-closed-direct-chat-with-the-appservice\" class=\"anchor\" href=\"#what-do-i-do-if-i-already-closed-direct-chat-with-the-appservice\" aria-hidden=\"true\"><\/a>What do I do if I already closed direct chat with the appservice?<\/h2>\n<p>Simply open it again, the appservice will join on invite. You can find it from user list of any bridged room, for example the freenode one is <code>@appservice-irc:matrix.org<\/code>.<\/p>\n<h2>\n<a id=\"user-content-how-do-i-change-my-irc-side-nick\" class=\"anchor\" href=\"#how-do-i-change-my-irc-side-nick\" aria-hidden=\"true\"><\/a>How do I change my IRC side nick?<\/h2>\n<p>Say <code>!nick YourNewNicknameHere<\/code> in direct chat with the <a href=\"https:\/\/github.com\/matrix-org\/matrix-appservice-irc\/wiki\/Bridged-IRC-networks\">appservice user<\/a>.<\/p>\n<h2>\n<a id=\"user-content-how-do-i-join-password-protected-rooms-with-a-key\" class=\"anchor\" href=\"#how-do-i-join-password-protected-rooms-with-a-key\" aria-hidden=\"true\"><\/a>How do I join password protected rooms with a key?<\/h2>\n<p>Say <code>!join #channel key<\/code> in direct chat with the <a href=\"https:\/\/github.com\/matrix-org\/matrix-appservice-irc\/wiki\/Bridged-IRC-networks\">appservice user<\/a>.<\/p>\n<h3>\n<a id=\"user-content-i-keep-getting-err_banonchan\" class=\"anchor\" href=\"#i-keep-getting-err_banonchan\" aria-hidden=\"true\"><\/a>I keep getting <code>err_banonchan<\/code>\n<\/h3>\n<p>You are banned from the channel on the IRC side of the bridge, so the appservice kicks you out of the corresponding Matrix room.<\/p>\n<ul><li>If the channel simply bans all unidentified users, you need to identify yourself explicitly with your nickname (see below (<em>How do I register\/identify to NickServ?<\/em>))<\/li>\n<li>If the channel bans you specifically or all Matrix users, you may have to start leaving the rooms you suspect to potentially be the cause or leave all the channels you have joined (you can quit every channel by sending <code>!quit<\/code> to the <a href=\"https:\/\/github.com\/matrix-org\/matrix-appservice-irc\/wiki\/Bridged-IRC-networks\">appservice user<\/a> which will also disconnect your nick from the network).<\/li>\n<\/ul><h2>\n<a id=\"user-content-why-dont-i-get-notifications-from-irc-pings\" class=\"anchor\" href=\"#why-dont-i-get-notifications-from-irc-pings\" aria-hidden=\"true\"><\/a>Why don't I get notifications from IRC pings?<\/h2>\n<p><em>This section only applies for matrix-appservice-irc versions older than 0.11.0 released on 2018-08-XX.<\/em><\/p>\n<p>Currently you have to add your IRC nick to your <em>keywords<\/em>. You can do this by opening <em>Settings<\/em> in <em>Riot Web<\/em> and under <em>NOTIFICATIONS<\/em> there is<\/p>\n<blockquote>\n<p>Messages containing keywords<\/p>\n<\/blockquote>\n<p>The word <em>keywords<\/em> is a link and by clicking it you can add\/edit your keywords and add your IRC nick there to get pinged by it in the future.<\/p>\n<h2>\n<a id=\"user-content-i-am-getting-kicked-out-of-a-bridged-room-with-the--err_needreggednick-error\" class=\"anchor\" href=\"#i-am-getting-kicked-out-of-a-bridged-room-with-the--err_needreggednick-error\" aria-hidden=\"true\"><\/a>I am getting kicked out of a bridged room with the  <code>err_needreggednick<\/code> error<\/h2>\n<p>You need to register your username on the bridged IRC network that room (see next section). The name of the appservice account that kicked you can be used to tell <a href=\"https:\/\/github.com\/matrix-org\/matrix-appservice-irc\/wiki\/Bridged-IRC-networks\">which network<\/a>.<\/p>\n<h2>\n<a id=\"user-content-how-do-i-registeridentify-to-nickserv\" class=\"anchor\" href=\"#how-do-i-registeridentify-to-nickserv\" aria-hidden=\"true\"><\/a>How do I register\/identify to NickServ?<\/h2>\n<p>Start by opening a direct chat with NickServ. It's generally <code>@_&lt;network&gt;_NickServ:homeserver.example.org<\/code>, though some networks omit the leading <code>_<\/code> like freenode: <code>@freenode_NickServ:matrix.org<\/code>. You can figure out the network prefix from <a href=\"https:\/\/github.com\/matrix-org\/matrix-appservice-irc\/wiki\/Bridged-IRC-networks\">Bridged IRC networks<\/a>.<\/p>\n<p>Then you can register\/identify normally depending on the syntax of your network which you can generally check by saying \"help register\" or \"help identify\".<\/p>\n<p>For example freenode uses the syntaxes:<\/p>\n<ul><li>\n<code>identify [&lt;nickname&gt;] &lt;password&gt;<\/code>\n<ul><li>where nickname is optional assuming you are identifying for your current nickname<\/li>\n<\/ul><\/li>\n<li><code>register &lt;password&gt; &lt;email&gt;<\/code><\/li>\n<\/ul><p>In case you would already be registered to NickServ and possibly have IRC client running, instead of <code>register<\/code> you would use <code>identify &lt;nickname&gt; &lt;password&gt;<\/code> and then <code>group<\/code> to group your current nick under your existing account.<\/p>\n<h3>\n<a id=\"user-content-and-do-it-automatically\" class=\"anchor\" href=\"#and-do-it-automatically\" aria-hidden=\"true\"><\/a>And do it automatically?<\/h3>\n<p>Again in direct chat with the <code>@appservice-irc<\/code> you can use <code>!storepass &lt;string&gt;<\/code> command to save a string that is sent in PASS command when your ghost (=your IRC-presence) connects if your network sends the PASS command to NickServ.<\/p>\n<p>With freenode you can <code>!storepass nickname:password<\/code>, some only accept the <code>!storepass password<\/code> and some just ignore the command.<\/p>\n<h2>\n<a id=\"user-content-why-i-dont-get-messages-from-unregistered-to-irc-irc-users\" class=\"anchor\" href=\"#why-i-dont-get-messages-from-unregistered-to-irc-irc-users\" aria-hidden=\"true\"><\/a>Why I don't get messages from unregistered-to-IRC IRC users?<\/h2>\n<p>The bridges on matrix.org automatically sets umode +R (+G at OFTC) for all ghosts instructing the IRC server to not send messages from unidentified users.<\/p>\n<p>When an IRC user sends a private message to Matrix user, the bridge converts it to a room and invites the Matrix user there. In the past, there have been PM spam bots causing 2000+ room invites in a short time and crashing Synapse (the Matrix homeserver software) on matrix.org taking the whole homeserver down and the best way to prevent this from happening again is requiring registration which should be available for all IRC users.<\/p>\n<p>In order to allow messages from unidentified users, run e.g. <code>!cmd chat.freenode.net MODE YOURNICK -R<\/code>, you can get YOURNICK and the server with <code>!nick<\/code>.<\/p>\n<h2>\n<a id=\"user-content-how-do-i-change-the-topic\" class=\"anchor\" href=\"#how-do-i-change-the-topic\" aria-hidden=\"true\"><\/a>How do I change the topic?<\/h2>\n<h3>\n<a id=\"user-content-if-you-are-a-moderator-at-matrix-and-op-at-irc\" class=\"anchor\" href=\"#if-you-are-a-moderator-at-matrix-and-op-at-irc\" aria-hidden=\"true\"><\/a>If you are a moderator at Matrix and op at IRC<\/h3>\n<p>Simply set the room topic at Matrix and it will automatically sync to IRC.<\/p>\n<h3>\n<a id=\"user-content-if-you-arent-op-at-irc-or-the-channel-is--t\" class=\"anchor\" href=\"#if-you-arent-op-at-irc-or-the-channel-is--t\" aria-hidden=\"true\"><\/a>If you aren't op at IRC or the channel is -t<\/h3>\n<p>Use the raw <code>TOPIC<\/code> command. Don't forget to prepend the <code>:<\/code> character before the sentence you will set as topic. For example:<\/p>\n<pre>!cmd TOPIC #foo :Welcome to the #foo channel!\n<\/pre>\n<p>will set <code>Welcome to the #foo channel!<\/code> as the topic for the <code>#foo<\/code> IRC channel.<\/p>\n<h2>\n<a id=\"user-content-i-am-a-chanop-and-have-a-spam-problem-how-can-i-fix-it-without-affecting-matrix-users\" class=\"anchor\" href=\"#i-am-a-chanop-and-have-a-spam-problem-how-can-i-fix-it-without-affecting-matrix-users\" aria-hidden=\"true\"><\/a>I am a chanop and have a spam problem, how can I fix it without affecting Matrix users?<\/h2>\n<p>You have likely attempted to use <code>mode +r<\/code> or similar which prevents unidentified users from joining the channel also affecting Matrix users. It's not the only way and you may wish to think about some other modes:<\/p>\n<ul><li>\n<p><code>MODE +S<\/code> requires SSL\/TLS secured connection for joining and has for now also cut spambots. However it may require your non-Matrix users to reconfigure their clients to connect to port 6697 and set checkbox for using SSL\/TLS.<\/p>\n<\/li>\n<li>\n<p><code>MODE +be $~a *!*@gateway\/shell\/matrix.org\/*<\/code> bans unidentified users, while making Matrix users exempt from spam. However the drawback is that to ban a Matrix user, you would need to connect to Matrix to do it.<\/p>\n<\/li>\n<li>\n<p><code>MODE +q $~a<\/code> with the exemption <code>+e *!*@gateway\/shell\/matrix.org\/*<\/code> (can be set in one command <code>MODE #channel +qe $~a *!*@gateway\/shell\/matrix.org\/*<\/code>) will prevent anyone unregistered speaking in the channel (or changing nick) but will allow them to join. (While exempting the matrix bridge users). This allows people to read the topic and the channel (to get nick registration instructions) but not to post.<\/p>\n<\/li>\n<li>\n<p><code>MODE +q $~a<\/code> would quiet all unidentified users preventing them from talking and <code>\/cs flags #channel *!*@gateway\/shell\/matrix.org\/* +V<\/code> would give all Matrix users automatic voice on join overriding this. The drawback would be that you cannot quiet Matrix users anymore (note that the matrix-appservice-bridge doesn't understand quiets anyway) and it may look ugly\/weird.<\/p>\n<\/li>\n<li>\n<p><code>MODE +m<\/code> &amp; setting a bot to autovoice everyone with some delay from joining to the channel. This would likely require you to host something by yourself and it would also prevent using of quiet instead of ban, but would be the easiest solution for your users.<\/p>\n<\/li>\n<\/ul>","created_at":"2019-02-15T23:29:48+0000","updated_at":"2019-02-15T23:29:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/8418310?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":837,"title":"The state of Python Packaging - package types","url":"https:\/\/www.bernat.tech\/pep-517-and-python-packaging\/","content":"<p>pip 19.0 has been released on 22nd January 2019. On the feature list most notably it now supports PEP-517, which by default is turned <strong>on<\/strong> when that the project has a <code>pyproject.toml<\/code> at the root folder. The PEP in question has been created in 2015 and accepted in 2017. Even though it took a while until pip implemented it, the release and the issues that followed it up confirmed that many people are not at all familiar with it. Read on if you want to get a picture of how the Python packaging ecosystem evolved as of today and where we hope to see it down the line. We hope even though the introduction of aforementioned python enhancement proposal may cause some discomfort at the moment, we will benefit from it in the long term.<\/p><p>I've joined the Python open source community around three years ago (though used it for more like 8 years). From the early days I've remarked that the Python packaging has a reputation of a kinda black box. There's a lot of unknown, and people mostly get by with just copying other projects build configurations and roll with it. On my path to understand better this black box and try to improve it I've became the maintainer of both the virtualenv and tox project, with occasionally contributing to both <code>setuptools<\/code> and <code>pip<\/code>.<\/p><p>As I hope to give an exhausting (hopefully still kinda high level) overview of the subject I decided to split it up in three posts. In this first post I will give a broad overview of how Python packaging works along with the type of packages it has. In a subsequent post I will present in detail how the installation of packages works, and how PEP-517\/518 tries to improve on it. Finally, I dedicate a whole other post to present some of the painful lessons we learned while introducing these improvements. A heads up, I will focus mostly on the Python Packaging Authorities systems (<code>pip<\/code>, <code>setuptools<\/code>, so no <code>conda<\/code> or operating system specific packagers).<\/p><figure class=\"kg-card kg-image-card\"><img src=\"https:\/\/images.unsplash.com\/photo-1509037837816-6d0b98fcdf08?ixlib=rb-1.2.1&amp;q=80&amp;fm=jpg&amp;crop=entropy&amp;cs=tinysrgb&amp;w=1080&amp;fit=max&amp;ixid=eyJhcHBfaWQiOjExNzczfQ\" class=\"kg-image\" alt=\"shallow focus photography of adult fawn pug\" \/><figcaption>Photo by <a href=\"https:\/\/unsplash.com\/@marcuslcramer?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Marcus Cramer<\/a> \/ <a href=\"https:\/\/unsplash.com\/?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Unsplash<\/a> - peoples face when they gaze into Python packaging first time<\/figcaption><\/figure><p>To tell this story, I'll need to tell the story of how to distribute python packages; more specifically explain how package installation worked in the past and how we hope it will work in the future. To have a concrete example on what to demonstrate, let me introduce my wonderful example library: <code>pugs<\/code>. This library reasonably simple: it just generates a single package called <code>pugs<\/code> that contains only a single module called <code>logic<\/code>. The logic is to generate random quotes, you guessed right, about <em>pugs<\/em>. Here's a simple example structure of it viewed as a source tree (also available under <a href=\"https:\/\/github.com\/gaborbernat\/pugs\">gaborbernat\/pugs<\/a>):<\/p><pre>pugs-project\n\u251c\u2500\u2500 README.rst\n\u251c\u2500\u2500 setup.cfg\n\u251c\u2500\u2500 setup.py\n\u251c\u2500\u2500 LICENSE.txt\n\u251c\u2500\u2500 src\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 pugs\n\u2502\u00a0\u00a0     \u251c\u2500\u2500 __init__.py\n\u2502\u00a0\u00a0     \u2514\u2500\u2500 logic.py\n\u251c\u2500\u2500 tests\n\u2502\u00a0\u00a0 \u251c\u2500\u2500 test_init.py\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 test_logic.py\n\u251c\u2500\u2500 tox.ini\n\u2514\u2500\u2500 azure-pipelines.yml<\/pre><p>We have four distinct content type here:<\/p><p>What would it mean for our <code>pugs<\/code> package to be available on a user machine's interpreter? Ideally, the user should be able to import it and call functions from it once it starts up the interpreter:<\/p><ul><li>the business logic code (what's inside the <code>src<\/code> folder),<\/li>\n<li>the test code (<code>tests<\/code> folder and <code>tox.ini<\/code>),<\/li>\n<li>the packaging code and metadata (<code>setup.py<\/code>, <code>setup.cfg<\/code>, <code>LICENSE.txt<\/code>, <code>README.rst<\/code> - note we use nowadays the de facto standard packaging tool <a href=\"https:\/\/pypi.org\/project\/setuptools\">setuptools<\/a>),<\/li>\n<li>files helping with project management and maintenance:\n<ul><li>continuous integration (<code>azure-pipelines.yml<\/code>)<\/li>\n<li>version control (<code>.git<\/code>)<\/li>\n<li>project management (for example a potential <code>.github<\/code> folder).<\/li>\n<\/ul><\/li>\n<\/ul><pre>Python 3.7.2 (v3.7.2:9a3ffc0492, Dec 24 2018, 02:44:43)\n[Clang 6.0 (clang-600.0.57)] on darwin\nType \"help\", \"copyright\", \"credits\" or \"license\" for more information.\n&gt;&gt;&gt; import pugs\n&gt;&gt;&gt; pugs.do_tell()\n\"An enlightened pug knows how to make the best of whatever he has to work with - A Pug's Guide to Dating -  Gemma Correll\"<\/pre><figure class=\"kg-card kg-image-card\"><img src=\"https:\/\/images.unsplash.com\/photo-1536351392281-79d89dd69869?ixlib=rb-1.2.1&amp;q=80&amp;fm=jpg&amp;crop=entropy&amp;cs=tinysrgb&amp;w=1080&amp;fit=max&amp;ixid=eyJhcHBfaWQiOjExNzczfQ\" class=\"kg-image\" alt=\"depth photo of fawn pug\" \/><figcaption>Photo by <a href=\"https:\/\/unsplash.com\/@ryanantooa?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Ryan Antooa<\/a> \/ <a href=\"https:\/\/unsplash.com\/?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Unsplash<\/a> - let's get started, excited!<\/figcaption><\/figure><p>How does Python knows what's available or not? The short answer is it does not. Not upfront, at least, that is. Instead, it will try to load and see if it succeeds dynamically. From where does it load it? There are many possible locations, but in most cases, we are talking about loading it from a folder on the file system. Where is this folder? For a given module, one can print out the representation of the module to find out: <\/p><pre>&gt;&gt;&gt; import pugs\n&gt;&gt;&gt; pugs\n&lt;module 'pugs' from '\/Users\/bernat\/Library\/Python\/3.7\/lib\/python\/site-packages\/pugs\/__init__.py'&gt;<\/pre><p>The folder under you'll find it depends on:<\/p><ul><li> the type of the package it is (third-party or built-in\/aka part of the standard library)<\/li><li>if it's globally or just for the current user available (see <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0370\/\">PEP-370<\/a>),<\/li><li>and if it's a system python or a virtual environment. <\/li><\/ul><p>Generally speaking though, for a given python interpreter, one can find a list of possible directories by printing out the <code>sys.path<\/code> variables content, for example on my MacOS:<\/p><pre>&gt;&gt;&gt; import sys\n&gt;&gt;&gt; print('\\n'.join(sys.path))\n\/Library\/Frameworks\/Python.framework\/Versions\/3.7\/lib\/python37.zip\n\/Library\/Frameworks\/Python.framework\/Versions\/3.7\/lib\/python3.7\n\/Library\/Frameworks\/Python.framework\/Versions\/3.7\/lib\/python3.7\/lib-dynload\n\/Users\/bernat\/Library\/Python\/3.7\/lib\/python\/site-packages\n\/Library\/Frameworks\/Python.framework\/Versions\/3.7\/lib\/python3.7\/site-packages<\/pre><p>For third-party packages, it's going to be some <code>site-packages<\/code> folder. Note in the above example how there's a system-wide and a user-specific example of this. How does packages end up inside this folder? It must be put there by some installer.<\/p><p>The following diagram displays how most of the time things go:<\/p><ol><li>The developer writes some python code inside a folder (referred to as the source tree).<\/li><li>Some tool (such as <code>setuptools<\/code>) then takes the source tree and packages it up for redistribution.<\/li><li>The generated package is uploaded via another tool (twine) to a central repository (usually <a href=\"https:\/\/pypi.org\">https:\/\/pypi.org<\/a>) the end user machine has access to. <\/li><li> The end user machine uses some installer that discovers, downloads and installs the package in question. The installation operation boils down by creating the correct directory structure and metadata inside the <code>site-packages<\/code> folder. <\/li><\/ol><figure class=\"kg-card kg-image-card\"><img src=\"https:\/\/images.unsplash.com\/photo-1538131688925-7e0eb2e7828b?ixlib=rb-1.2.1&amp;q=80&amp;fm=jpg&amp;crop=entropy&amp;cs=tinysrgb&amp;w=1080&amp;fit=max&amp;ixid=eyJhcHBfaWQiOjExNzczfQ\" class=\"kg-image\" alt=\"image\" \/><figcaption>Photo by <a href=\"https:\/\/unsplash.com\/@pinho?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Pinho .<\/a> \/ <a href=\"https:\/\/unsplash.com\/?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Unsplash<\/a> - discovering new things<\/figcaption><\/figure><p>A package during installation must generate at least two types of content to be put in the site package: a metadata folder about the package contents the <code>{package}-{version}.dist-info<\/code> and the business logic files.<\/p><pre>\/Users\/bgabor8\/Library\/Python\/3.7\/lib\/python\/site-packages\/pugs\n\u251c\u2500\u2500 __init__.py\n\u251c\u2500\u2500 __pycache__\n\u2502   \u251c\u2500\u2500 __init__.cpython-37.pyc\n\u2502   \u2514\u2500\u2500 logic.cpython-37.pyc\n\u2514\u2500\u2500 logic.py\n\/Users\/bgabor8\/Library\/Python\/3.7\/lib\/python\/site-packages\/pugs-0.0.1.dist-info\n\u251c\u2500\u2500 INSTALLER\n\u251c\u2500\u2500 LICENSE.txt\n\u251c\u2500\u2500 METADATA\n\u251c\u2500\u2500 RECORD\n\u251c\u2500\u2500 WHEEL\n\u251c\u2500\u2500 top_level.txt\n\u2514\u2500\u2500 zip-safe\n<\/pre><p>The distribution info folder describes the package: what installer was used to put it there, what license the package has attached to it, what files were created as part of the installation process, what is the top level python package, what entry points the package exposes and so on. A good description of each file can be found inside <a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0427\/#id14\">PEP-427<\/a>. <\/p><p>How do we get from our source tree to these two content type? We have two distinct paths in front of us:<\/p><ol><li>Generate this directory structure and metadata from our source tree, compress it into a single file and publish it to the central package repository. In this case, the installer has to download the package and extract it to the site-packages folder. We refer to this type of package as a <code>wheel<\/code> package.<\/li><li>Alternatively you can create an archive containing the package source, build scripts and metadata required to generate the installable directory structure, then upload that to the central repository. This is called a source distribution or <code>sdist<\/code>. In this case the installer has a lot more work to do, needs to extract the archive, run the builder and only then copy it over.<\/li><\/ol><p>The difference between the two options is mainly about where the package <em>compile\/build<\/em> operation happens: on the developers' machine or the end users machine. If it happens on the developers' side (such as in the case of wheels), the installation process is very lightweight. Everything has been already done on the developer machine side. The user machine side operation is just a simple download and decompress. <\/p><p>In our case, we used <code>setuptools<\/code> as our builder (the part that generates from the source tree the content to be put inside the site-packages folder). \u00a0Therefore, in order to perform the build operation on the user machine we would need to ensure that an <em><u>appropriate<\/u><\/em> version of <code>setuptools<\/code> is available on the user machine (if you're using a feature from version <code>40.6.0<\/code> you must guarantee that the user has at least that version).<\/p><p>Another use case to consider is that Python offers the ability to access C\/C++ libraries from within python (to get that extra performance where you need it). Packages that do so are referred to as C-extension packages, as they take advantage of the C-extension API CPython offers. <\/p><p>Such extensions to work though need to C\/C++ compile their functionality against both the C\/C++ library they interact with and the current Python interpreters C-API library. In these cases the build operation actually involves calling a binary compiler, not just metadata and folder structure generation as was the case with pure python packages (such as our pugs library was).<\/p><p>If the build happens on the user machine one needs to ensure that the correct libraries and compilers are also available at build time. This is now a much harder job, as these are platform specific binaries that are distributed via the platforms packaging tools. The lack or version miss-match of these libraries often trigger cryptic errors during the build, that leaves users frustrated and puzzled. <\/p><p>Therefore, if possible, always prefer packaging your package as a <strong><em>wheel<\/em><\/strong>. This will altogether avoid the problem of users not having all the correct build dependencies (both pure python types such as setuptools or binary ones as is the C\/C++ compiler). \u00a0Even if those build dependencies are easy to provision (such as in case of pure python builders - e.g. setuptools) you can save install time by avoiding this step entirely. <\/p><p>That being said they are still two use cases that warrant providing source distributions (even when you provide a wheel):<\/p><ol><li>C-extension source distributions \u00a0tend to be more auditable, as one can read the source code and thus offer greater transparency in what it offers: many big corporate environments prefer using these over wheels for this sole reason (they generally extend the rule to pure python wheels too, mostly to avoid the need of categorizing what is pure python wheel and what is not).<\/li><li>You might not be able to provide a wheel for every possible platform out there (especially true in case of c-extension packages), in this case, the source distribution may allow these platforms to generate the wheel themselves.<\/li><\/ol><p><strong>Conclusion<\/strong><\/p><p>The difference between a source tree, a source distribution and a wheel:<\/p><ul><li><strong>source tree<em> - <\/em><\/strong>contains all project files available on the developers' machine\/repository (business logic, tests, packaging data, CI files, IDE files, SVC etc.) - for example, see example project above.<\/li><li><strong>source distribution<\/strong> - <em> <\/em>contains code files required to build a wheel (business logic + packaging data + often also the unit tests files to validate the build; notably lacks developer environment content such as CI\/IDE\/version control files) - format: <code>pugs-0.0.1.tar.gz<\/code>.<\/li><li><strong>wheel <\/strong>- contains the package metadata and source files to be put into the site packages folder - format: <code>pugs-0.0.1-py2.py3-none-any.whl<\/code>. <\/li><\/ul><figure class=\"kg-card kg-image-card\"><img src=\"https:\/\/images.unsplash.com\/photo-1517423568366-8b83523034fd?ixlib=rb-1.2.1&amp;q=80&amp;fm=jpg&amp;crop=entropy&amp;cs=tinysrgb&amp;w=1080&amp;fit=max&amp;ixid=eyJhcHBfaWQiOjExNzczfQ\" class=\"kg-image\" alt=\"dog wearing shirt with hat\" \/><figcaption>Photo by <a href=\"https:\/\/unsplash.com\/@charlesdeluvio?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Charles \ud83c\uddf5\ud83c\udded<\/a> \/ <a href=\"https:\/\/unsplash.com\/?utm_source=ghost&amp;utm_medium=referral&amp;utm_campaign=api-credit\">Unsplash<\/a> - hmmm<\/figcaption><\/figure><p><a href=\"https:\/\/www.bernat.tech\/pep-517-518\/\">Read the next post of the series here<\/a> to find out what happens when we install a package. \u00a0Thanks for reading!<\/p>","created_at":"2019-02-13T20:00:41+0000","updated_at":"2019-02-13T20:00:41+0000","published_at":"2019-02-07T13:40:54+0000","published_by":["Bernat Gabor"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"www.bernat.tech","preview_picture":"https:\/\/images.unsplash.com\/photo-1474314170901-f351b68f544f?ixlib=rb-1.2.1&q=80&fm=jpg&crop=entropy&cs=tinysrgb&w=1080&fit=max&ixid=eyJhcHBfaWQiOjExNzczfQ","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":836,"title":"How to stop using TLS-SNI-01 with Certbot","url":"https:\/\/community.letsencrypt.org\/t\/how-to-stop-using-tls-sni-01-with-certbot\/83210?_escaped_fragment_=","content":"<div class=\"post\" itemprop=\"articleBody\"><p>Let\u2019s Encrypt is removing support for domain validation with TLS-SNI-01. If you\u2019re using Certbot and received an email titled \u201cAction required: Let\u2019s Encrypt certificate renewals\u201d, here\u2019s how to fix the problem. It\u2019s possible you\u2019ve upgraded Certbot in the time since the last TLS-SNI validation mentioned in the email, in which case you\u2019re fine. These instructions tell you how to check.<\/p>\n<ol><li>\n<p>Confirm your Certbot version is 0.28 or higher:<\/p>\n<pre>certbot --version || \/path\/to\/certbot-auto --version\n<\/pre>\n<\/li>\n<\/ol><p>If the version is less than 0.28, you need to upgrade your Certbot. Visit <a href=\"https:\/\/certbot.eff.org\/\">https:\/\/certbot.eff.org\/<\/a> and follow the instructions for your webserver and OS.<\/p>\n<ol start=\"2\"><li>\n<p>Remove any explicit references to tls-sni-01 in your renewal configuration:<\/p>\n<pre>sudo sh -c \"sed -i.bak -e 's\/^\\(pref_challs.*\\)tls-sni-01\\(.*\\)\/\\1http-01\\2\/g' \/etc\/letsencrypt\/renewal\/*; rm -f \/etc\/letsencrypt\/renewal\/*.bak\"\n<\/pre>\n<\/li>\n<li>\n<p>Do a full renewal dry run:<\/p>\n<pre>sudo certbot renew --dry-run\n<\/pre>\n<\/li>\n<\/ol><p>If the dry run succeeds, and your Certbot version is 0.28 or higher, you\u2019re good to go! No further action should be required to deal with the end of TLS-SNI-01 support. If it fails, fix the validation problems you see and try again.<\/p>\n<p>If you get a connection refused or connection timeout, you may have a firewall blocking port 80. tls-sni-01 used port 443, but http-01 uses port 80. Ideally your web server should allow both ports. If that\u2019s not possible, for instance because your ISP blocks port 80, you\u2019ll need to switch to the dns-01 challenge, or use an ACME client that supports tls-alpn-01.<\/p>\n<p>Note: if you installed Certbot in late 2015 or early 2016, it may be called <code>letsencrypt<\/code> or <code>letsencrypt-auto<\/code> (the project was renamed). Follow the instructions at <a href=\"https:\/\/certbot.eff.org\">https:\/\/certbot.eff.org<\/a> to install the latest version.<\/p>\n<p>Credit to <a class=\"mention\" href=\"https:\/\/community.letsencrypt.org\/u\/_az\">@_az<\/a> for the suggestion to write more step-by-step instructions and <a class=\"mention\" href=\"https:\/\/community.letsencrypt.org\/u\/jsha\">@jsha<\/a> for rewriting these instructions with that suggestion in mind.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><div class=\"post\" itemprop=\"articleBody\"><p>A post was split to a new topic: <a href=\"https:\/\/community.letsencrypt.org\/t\/using-port-443-for-renewal-after-tls-sni-is-disabled\/83262\">Using port 443 for renewal after TLS-SNI is disabled<\/a><\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><\/div><\/div>","created_at":"2019-02-12T21:46:49+0000","updated_at":"2019-02-12T21:46:49+0000","published_at":"2019-01-18T22:10:33+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"community.letsencrypt.org","preview_picture":"https:\/\/discourse-cdn-sjc1.com\/letsencrypt\/uploads\/default\/original\/3X\/c\/a\/ca6c06ea1ea201324bba7048c6841ce60236468d.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["gnu","mailman","email"],"is_public":false,"id":835,"title":"Mailman Suite Installation \u2014 Mailman Suite 3.0 documentation","url":"http:\/\/docs.mailman3.org\/en\/latest\/prodsetup.html","content":"<p>This is a step-by-step installation guide for Mailman Suite, also sometimes\nreferred as Mailman 3 Suite or just Mailman 3. There are more than one ways to\nhave Mailman 3 running on your machine.<\/p>\n<div class=\"section\" id=\"mailman-3-in-docker\"><h2>Mailman 3 in Docker<\/h2><p>Abhilash Raj maintains container images for Mailman 3 which you can use directly\nwithout having to go through all the steps to download dependencies and\nconfiguring Mailman. You can find the <a class=\"reference external\" href=\"https:\/\/asynchronous.in\/docker-mailman\/\">detailed documentation<\/a> on how to use\nthem. If you have any issues using the container images, please report them\ndirectly on the <a class=\"reference external\" href=\"https:\/\/github.com\/maxking\/docker-mailman\">Github repo for docker-mailman<\/a>.<\/p><\/div>\n<div class=\"section\" id=\"distribution-packages\"><h2>Distribution Packages<\/h2><p>Debian provides Mailman 3 packages. The meta-package <cite>mailman3-full<\/cite> depends on\nall components of a complete Mailman 3 suite. If you want to split the\ninstallation of the Mailman 3 core mailinglist delivery daemon and the Mailman 3\nDjango web suite with Postorius and Hyperkitty, take a look at the packages\n<cite>mailman3<\/cite> and <cite>mailman3-web<\/cite>.<\/p><p>The Debian Mailman 3 packages can be found here:\n<a class=\"reference external\" href=\"https:\/\/packages.debian.org\/search?keywords=mailman3\">https:\/\/packages.debian.org\/search?keywords=mailman3<\/a><\/p><p>Other distribution specific packages for Mailman 3 are not yet available. If you\nwould like to help package Mailman 3 for your favorite Linux distro, please get\nin touch at <a class=\"reference external\" href=\"mailto:mailman-developers%40python.org\">mailman-developers@python.org<\/a>.<\/p><\/div>\n<div class=\"section\" id=\"dependencies\"><h2>Dependencies<\/h2><blockquote>\n<div><ul><li><p class=\"first\">Python3.5+.<\/p>\n<\/li>\n<li><p class=\"first\">MTA: Mailman 3 in theory would support any MTA or mail server which can send\nemails to Mailman over LMTP. Officially, there are configurations for\n<a class=\"reference external\" href=\"http:\/\/www.postfix.org\/\">Postfix<\/a>, <a class=\"reference external\" href=\"http:\/\/www.exim.org\/\">Exim4<\/a>, <a class=\"reference external\" href=\"http:\/\/cr.yp.to\/qmail.html\">qmail<\/a> and <a class=\"reference external\" href=\"https:\/\/www.proofpoint.com\/us\/sendmail-open-source\">sendmail<\/a>. Mailman Core has a fairly elaborate\ndocumentation on <a class=\"reference external\" href=\"https:\/\/mailman.readthedocs.io\/en\/latest\/src\/mailman\/docs\/mta.html\">setting up your MTA<\/a>. Look below at <a href=\"#id3\">`setting up mailman\ncore`_<\/a> to find out the location of configuration file <code class=\"docutils literal notranslate\">mailman.cfg<\/code> which\nis mentioned in the documentation above.<\/p>\n<p>The Web Front-end is based on a Python web framework called Django. For\nemail verification and sending out error logs, Django also must be\nconfigured to send out emails.<\/p>\n<\/li>\n<li><p class=\"first\">A <a class=\"reference external\" href=\"http:\/\/sass-lang.com\/\">sass compiler<\/a>. Syntactically Awesome Stylesheets or Sass is an\nextension of CSS, which is used to style webpages, that adds more power to\nthe old CSS syntax. It allows using variables, nested rules, inline imports\netc. which CSS originally doesn\u2019t support. Hyperkitty uses this to generate\nCSS styles.<\/p>\n<p>You can either use the Ruby implementation or <a class=\"reference external\" href=\"http:\/\/sass-lang.com\/libsass\">C\/C++ implementation<\/a>. Please look at the <a class=\"reference external\" href=\"http:\/\/sass-lang.com\/install\">installation guide for\nsass<\/a> to see how you can get one.<\/p>\n<p>For apt based systems:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo apt install ruby-sass\n<\/pre><\/div><\/div>\n<p>For yum based systems try:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo yum install rubygem-sass\n<\/pre><\/div><\/div>\n<p>or:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo dnf install rubygem-sass\n<\/pre><\/div><\/div>\n<p>After installing this, you\u2019d have to configure Django to use these\ncompilers. A basic configuration would look like:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre># To be added to Django's settings.py\nCOMPRESS_PRECOMPILERS = (\n  ('text\/x-scss', 'sass -t compressed {infile} {outfile}'),\n  ('text\/x-sass', 'sass -t compressed {infile} {outfile}'),\n)\n<\/pre><\/div><\/div>\n<p>You can replace <code class=\"docutils literal notranslate\">sass<\/code> above with whatever is the name of the binary\nis. For the Ruby implementation it is generally <code class=\"docutils literal notranslate\">sass<\/code>, for C\/C++ it is\n<code class=\"docutils literal notranslate\">sassc<\/code>.<\/p>\n<\/li>\n<li><p class=\"first\">A full text search engine like <a class=\"reference external\" href=\"https:\/\/pythonhosted.org\/Whoosh\/\">Whoosh<\/a> or <a class=\"reference external\" href=\"https:\/\/xapian.org\/\">Xapian<\/a>. Whoosh is the easiest\nto setup. First, install Whoosh:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo pip3 install whoosh\n<\/pre><\/div><\/div>\n<p>Then configure Django to use it:<\/p>\n<div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre># To be added to Django's settings.py\nHAYSTACK_CONNECTIONS = {\n      'default': {\n      'ENGINE': 'haystack.backends.whoosh_backend.WhooshEngine',\n      'PATH': os.path.join(BASE_DIR, \"fulltext_index\"),\n      },\n}\n<\/pre><\/div><\/div>\n<p>For setup instruction on other search backends, look at the documentation\nfor <a class=\"reference external\" href=\"https:\/\/django-haystack.readthedocs.io\/en\/master\/backend_support.html\">django-haystack<\/a>.<\/p>\n<\/li>\n<li><p class=\"first\">An HTML to plaintext converter like <code class=\"docutils literal notranslate\">lynx<\/code> is required by Mailman Core if you\nhave configured it to convert emails to plaintext.<\/p>\n<\/li>\n<\/ul><\/div><\/blockquote><\/div>\n<div class=\"section\" id=\"installing-mailman-core\"><h2>Installing Mailman Core<\/h2><p>Mailman Core is responsible for sending and receiving emails. It exposes a REST\nAPI that different clients can use to interact with over an HTTP protocol. The\nAPI itself is an administrative API and it is recommended that you don\u2019t expose\nit to outside of your host or trusted network. To install Core run:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo pip3 install mailman\n<\/pre><\/div><\/div><p>This _should_ install the latest release of Mailman Core, which is 3.1.0 as of\nwriting of this document. After this, create a configuration file at\n<code class=\"docutils literal notranslate\">\/etc\/mailman.cfg<\/code> for Mailman Core.<\/p><p>After this, running <code class=\"docutils literal notranslate\">mailman info<\/code> will give you an output which looks\nsomething like below:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ mailman info\nGNU Mailman 3.1.0 (Between The Wheels)\nPython 3.5.3 (default, Jan 19 2017, 14:11:04)\n[GCC 6.3.0 20170118]\nconfig file: \/etc\/mailman.cfg\ndb url: sqlite:\/\/\/\/var\/lib\/mailman\/data\/mailman.db\ndevmode: DISABLED\nREST root url: http:\/\/localhost:8001\/3.1\/\nREST credentials: restadmin:restpass\n<\/pre><\/div><\/div><\/div>\n<div class=\"section\" id=\"installing-web-ui\"><h2>Installing Web UI<\/h2><p>Postorius and Hyperkitty are Mailman\u2019s official Web UI and Archiver. Both of\nthem are Django based apps and can be integrated into an existing Django\ninstallation\/website.<\/p><div class=\"section\" id=\"installing-postorius-hyperkitty\"><h3>Installing Postorius &amp; Hyperkitty<\/h3><p>Postorius and Hyperkitty are Django \u201capps\u201d and can be directly installed using\nthe commands below:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo pip3 install postorius\n$ sudo pip3 install hyperkitty\n<\/pre><\/div><\/div><p>mailman-hyperkitty plugin enables interaction between Core and Hyperkitty. You\ncan install it using:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo pip3 install mailman-hyperkitty\n<\/pre><\/div><\/div><p>Note that mailman-hyperkitty is a plugin for Mailman Core and not Django.<\/p><\/div><div class=\"section\" id=\"setting-up-django-project\"><h3>Setting up Django Project<\/h3><p>If you have absolutely no idea about Django, just clone\/download this\n<a class=\"reference external\" href=\"https:\/\/gitlab.com\/mailman\/mailman-suite\/tree\/master\">mailman-suite repo<\/a> . It should have a documented <code class=\"docutils literal notranslate\">settings.py<\/code> and a\npre-configured <code class=\"docutils literal notranslate\">urls.py<\/code> which you can use to run new Django installation. If\nyou find any setting that is not documented, please look at <a class=\"reference external\" href=\"https:\/\/docs.djangoproject.com\/en\/1.11\/ref\/settings\/\">Django\u2019s settings\nreference<\/a> for all available settings.<\/p><p>Exact commands would look something like this:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre># Download and install the latest release of django.\n$ sudo pip3 install Django&gt;=1.11\n# Clone the repo locally.\n$ git clone https:\/\/gitlab.com\/mailman\/mailman-suite.git\n$ cd mailman-suite\/mailman-suite_project\/\n# Create the tables in the database and load fixtures.\n$ python3 manage.py migrate\n# Copy all the static files to one single location.\n$ python3 manage.py collectstatic\n# Run the Django's \"development\" server at localhost:8000\n$ python3 manage.py runserver\n<\/pre><\/div><\/div><\/div><\/div>\n<div class=\"section\" id=\"setting-up-a-wsgi-server\"><h2>Setting up a WSGI server<\/h2><p>These instructions are to setup your Django website behind a webserver. We are\nusing a <a class=\"reference external\" href=\"http:\/\/uwsgi-docs.readthedocs.io\/\">uwsgi<\/a> as the wsgi server to\ncommunicate between the webserver and Django. To install uwsgi, run:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ sudo pip3 install uwsgi\n<\/pre><\/div><\/div><div class=\"admonition note\"><p class=\"first admonition-title\">Note<\/p><p class=\"last\">The configuration below doesn\u2019t serve static files, so if you are just\n\u201ctrying-it-out\u201d and want static files to be served, you need to add\nsome additional configuration and steps. See <a class=\"reference external\" href=\"https:\/\/uwsgi-docs.readthedocs.io\/en\/latest\/StaticFiles.html\">serving static files\nwith uwsgi<\/a>.<\/p><\/div><p>Then you can configure it using the following configuration file:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre># uwsgi.ini\n#\n[uwsgi]\n# Port on which uwsgi will be listening.\nhttp-socket = 0.0.0.0:8000\n# Move to the directory wher the django files are.\nchdir = \/path-to-django-project-directory\/\n# Use the wsgi file provided with the django project.\nwsgi-file = wsgi.py\n# Setup default number of processes and threads per process.\nmaster = true\nprocess = 2\nthreads = 2\n# Drop privielges and don't run as root.\nuid = 1000\ngid = 1000\n# Setup the django_q related worker processes.\nattach-daemon = .\/manage.py qcluster\n# Setup the request log.\nreq-logger = file:\/path-to-logs\/logs\/uwsgi.log\n# Log cron seperately.\nlogger = cron file:\/path-to-logs\/logs\/uwsgi-cron.log\nlog-route = cron uwsgi-cron\n# Log qcluster commands seperately.\nlogger = qcluster file:\/path-to-logs\/logs\/uwsgi-qcluster.log\nlog-route = qcluster uwsgi-daemons\n# Last log and it logs the rest of the stuff.\nlogger = file:\/path-to-logs\/logs\/uwsgi-error.log\n<\/pre><\/div><\/div><p>You can run uwsgi using the following command:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>$ uwsgi --ini \/path\/to\/uwsgi.ini\n<\/pre><\/div><\/div><p>Note that in the above configuration, there is a command called <code class=\"docutils literal notranslate\">python\nmanage.py qcluster<\/code> which run the <code class=\"docutils literal notranslate\">django-q<\/code> processes. You can remove this\nfrom here if you want to manage this yourself via some other init process.<\/p><p>Have a look at <a href=\"#id5\">`uwsgi`_<\/a> documentation to learn more about different\nconfiguration options. One minor optimization that can be done is to replace:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>http-socket = 0.0.0.0:8000\n<\/pre><\/div><\/div><p>with a more performant option:<\/p><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>uwsgi-socket = 0.0.0.0:8000\n<\/pre><\/div><\/div><p>However, this requires support for uwsgi protocol in the webserver. Nginx and\nApache, the two most popular web-servers have the support and the Nginx\nconfiguration below uses <code class=\"docutils literal notranslate\">uwsgi-socket<\/code>.<\/p><\/div>\n<div class=\"section\" id=\"nginx-configuration\"><h2>Nginx Configuration<\/h2><p>You can reverse proxy the requests to uwsgi server using Nginx. Uwsgi has a\nspecial protocol called uwsgi protocol that is available in Nginx as a\nplugin. Add the following configuration to your sites-availbale in Nginx<\/p><div class=\"admonition note\"><p class=\"first admonition-title\">Note<\/p><p>This configuration relies on uwsgi listening on a <code class=\"docutils literal notranslate\">uwsgi-socket<\/code>\ninstead of a <code class=\"docutils literal notranslate\">http-socket<\/code>. Please make sure before using this\nconfiguration.<\/p><p class=\"last\">Also, if you must use the <code class=\"docutils literal notranslate\">http-socket<\/code> option for some reason,\nreplace the line below with <code class=\"docutils literal notranslate\">uwsgi_pass 0.0.0.0:8000;<\/code> with\n<code class=\"docutils literal notranslate\">proxy_pass http:\/\/0.0.0.0:8000;<\/code>.<\/p><\/div><div class=\"highlight-default notranslate\"><div class=\"highlight\"><pre>server {\n   listen 443 ssl default_server;\n   listen [::]:443 ssl default_server;\n   server_name MY_SERVER_NAME;\n   location \/static\/ {\n        alias \/path-to-djang-staticdir\/static\/;\n   }\n   ssl_certificate \/path-to-ssl-certs\/cert.pem;\n   ssl_certificate_key \/path-to-ssl-certs\/privkey.pem;\n   location \/ {\n           include uwsgi_params;\n           uwsgi_pass 0.0.0.0:8000;\n   }\n}\n<\/pre><\/div><\/div><p>Fill in the appropriate paths above in the configuration before using it.<\/p><\/div>","created_at":"2019-02-12T18:07:56+0000","updated_at":"2019-02-12T18:08:08+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"docs.mailman3.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","stackexchange"],"is_public":false,"id":834,"title":"vermiculus\/sx.el","url":"https:\/\/github.com\/vermiculus\/sx.el","content":"<p><a href=\"https:\/\/travis-ci.org\/vermiculus\/sx.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a3f91825453e1151d4ed72e619b5d5689d987cd6\/68747470733a2f2f7472617669732d63692e6f72672f7665726d6963756c75732f73782e656c2e7376673f6272616e63683d6d6173746572\" alt=\"https:\/\/travis-ci.org\/vermiculus\/sx.el.svg?branch=master\" data-canonical-src=\"https:\/\/travis-ci.org\/vermiculus\/sx.el.svg?branch=master\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/melpa.org\/#\/sx\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/8e33ab2521e37d0792d9d5294518484d9f287b33\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f73782d62616467652e737667\" alt=\"http:\/\/melpa.org\/packages\/sx-badge.svg\" data-canonical-src=\"http:\/\/melpa.org\/packages\/sx-badge.svg\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"http:\/\/stable.melpa.org\/#\/sx\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/b9f723ba0cc67116c1cf2cc8b550d899b0869c27\/687474703a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f73782d62616467652e737667\" alt=\"http:\/\/stable.melpa.org\/packages\/sx-badge.svg\" data-canonical-src=\"http:\/\/stable.melpa.org\/packages\/sx-badge.svg\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/gitter.im\/vermiculus\/sx.el?utm_source=badge&amp;utm_medium=badge&amp;utm_campaign=pr-badge&amp;utm_content=badge\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/80bf7c9122b2d8f4dbadfc434c9074c1754267ae\/68747470733a2f2f6261646765732e6769747465722e696d2f4a6f696e20436861742e737667\" alt=\"https:\/\/badges.gitter.im\/Join Chat.svg\" data-canonical-src=\"https:\/\/badges.gitter.im\/Join Chat.svg\" referrerpolicy=\"no-referrer\" \/><\/a> <a href=\"https:\/\/www.waffle.io\/vermiculus\/sx.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7978524ca428e74c42bc345df8c145917f1e28b6\/68747470733a2f2f62616467652e776166666c652e696f2f7665726d6963756c75732f73782e656c2e737667\" alt=\"https:\/\/badge.waffle.io\/vermiculus\/sx.el.svg\" data-canonical-src=\"https:\/\/badge.waffle.io\/vermiculus\/sx.el.svg\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>SX is a full-featured Stack Exchange mode for GNU Emacs 24+. Using the official API, it provides a versatile experience for the Stack Exchange network within Emacs itself.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/vermiculus\/sx.el\/blob\/master\/list-and-question.png\"><img src=\"https:\/\/github.com\/vermiculus\/sx.el\/raw\/master\/list-and-question.png\" alt=\"list-and-question.png\" referrerpolicy=\"no-referrer\" \/><\/a><\/p>\n<p>As always, <code>C-h m<\/code> is the definitive resource for functionality, but here are a few highlights:<\/p>\n<h2>Viewing Questions<\/h2>\n<h3>Question Lists<\/h3>\n<p>View questions with one of the <code>sx-tab-<\/code> commands. These translate roughly to the different \u2018tabs\u2019 that you can view on the official site. Implemented tabs include:<\/p>\n<dl><dt><code>sx-tab-all-questions<\/code><\/dt>\n<dd>All questions.<\/dd>\n<dt><code>sx-tab-unanswered<\/code><\/dt>\n<dd><code>u<\/code> Unanswered questions.<\/dd>\n<dt><code>sx-tab-unanswered-my-tags<\/code><\/dt>\n<dd><code>U<\/code> Unanswered questions in your followed tags.<\/dd>\n<dt><code>sx-tab-featured<\/code><\/dt>\n<dd><code>f<\/code> Featured questions.<\/dd>\n<dt><code>sx-tab-starred<\/code><\/dt>\n<dd><code>*<\/code> Your starred questions.<\/dd>\n<dt><code>sx-tab-frontpage<\/code><\/dt>\n<dd><code>h<\/code> The frontpage of the site.<\/dd>\n<dt><code>sx-tab-meta-or-main<\/code><\/dt>\n<dd><code>m<\/code> Toggle between the meta and main sites.<\/dd>\n<dt><code>sx-tab-newest<\/code><\/dt>\n<dd><code>n<\/code> Newest questions first.<\/dd>\n<dt><code>sx-tab-topvoted<\/code><\/dt>\n<dd><code>v<\/code> Highest-scoring questions first.<\/dd>\n<dt><code>sx-tab-week<\/code><\/dt>\n<dd><code>w<\/code> Hot questions of the week.<\/dd>\n<\/dl><p>Inside SX buffers, you can switch to other tabs by typing <code>s<\/code> followed by the key listed with the tab above. You can also use <code>sx-tab-switch<\/code> (<code>s t<\/code>) to switch between tabs with completion. Each of these opens up a list of questions, and you can further customize the ordering of the list with <code>O<\/code>. Other keys include:<\/p>\n<dl><dt><code>n<\/code> and <code>p<\/code><\/dt>\n<dd>Navigate the list.<\/dd>\n<dt><code>j<\/code> and <code>k<\/code><\/dt>\n<dd>Navigate while viewing the question in a separate buffer.<\/dd>\n<dt><code>s s<\/code><\/dt>\n<dd>Switch site.<\/dd>\n<dt><code>v<\/code><\/dt>\n<dd>Visit the <code>thing-at-point<\/code> in your browser.<\/dd>\n<dt><code>w<\/code><\/dt>\n<dd>Copy the <code>thing-at-point<\/code> (usually a link).<\/dd>\n<dt><code>u<\/code> and <code>d<\/code><\/dt>\n<dd>Upvote and downvote.<\/dd>\n<dt><code>RET<\/code><\/dt>\n<dd>Open the question buffer.<\/dd>\n<\/dl><h3>Questions and Answers<\/h3>\n<dl><dt><code>u<\/code> and <code>d<\/code><\/dt>\n<dd>Upvote and downvote the question, answer, or comment at point.<\/dd>\n<dt><code>a<\/code><\/dt>\n<dd>Add an answer.<\/dd>\n<dt><code>e<\/code><\/dt>\n<dd>Edit the question or answer (or comment if it\u2019s yours; the usual rules apply)<\/dd>\n<dt><code>*<\/code><\/dt>\n<dd>Star the question.<\/dd>\n<dt><code>K<\/code><\/dt>\n<dd>Delete the question\/answer\/comment if you are able.<\/dd>\n<dt><code>w<\/code><\/dt>\n<dd>Copy code at point.<\/dd>\n<\/dl><h2>Your Inbox<\/h2>\n<dl><dt><code>sx-inbox<\/code><\/dt>\n<dd>View your inbox<\/dd>\n<dt><code>RET<\/code><\/dt>\n<dd>Visit inbox item<\/dd>\n<\/dl><p>SX is now available on MELPA! Both the stable release and the development version can be found there. Install it via the Package Menu or just run<\/p>\n<pre lang=\"text\" xml:lang=\"text\">M-x package-install RET sx RET\n<\/pre>\n<h2>Authenticating<\/h2>\n<p>If you are going to be doing any asking\/answering\/commenting\/upvoting\/downvoting\/ <em>etc.<\/em>, you must use <code>sx-authenticate<\/code> to provide SX with an authentication token to act on your behalf.<\/p>\n<p>After authentication, you will be redirected to the project page. This page will prominently display your authentication token. <strong>Keep this secret!<\/strong> It\u2019s as good as your password as far as StackExchange is concerned. Copy and paste the token into the prompt in Emacs.<\/p>\n<blockquote>\n<p>SX will store this authentication token in plain text in the <code>sx<\/code> folder of your <code>.emacs.d<\/code>. Please take any and all steps necessary to protect the security of your account. This token is as good as a password.<\/p>\n<\/blockquote>\n<h2>Sample Keybindings<\/h2>\n<h3>With <code>use-package<\/code><\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'use-package)\n(use-package sx\n  :config\n  (bind-keys :prefix \"C-c s\"\n             :prefix-map my-sx-map\n             :prefix-docstring \"Global keymap for SX.\"\n             (\"q\" . sx-tab-all-questions)\n             (\"i\" . sx-inbox)\n             (\"o\" . sx-open-link)\n             (\"u\" . sx-tab-unanswered-my-tags)\n             (\"a\" . sx-ask)\n             (\"s\" . sx-search)))<\/pre><\/div>\n<h3>Standard (With Prefix Key)<\/h3>\n<p>Shamelessly stolen from <a href=\"http:\/\/endlessparentheses.com\/sx-el-announcement-and-more-launcher-map.html\" rel=\"nofollow\">Endless Parentheses<\/a>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-prefix-command 'launcher-map)\n(global-set-key (kbd \"s-l\") 'launcher-map)\n(define-key launcher-map \"qq\" #'sx-tab-all-questions)\n(define-key launcher-map \"qi\" #'sx-inbox)\n(define-key launcher-map \"qo\" #'sx-open-link)\n(define-key launcher-map \"qu\" #'sx-tab-unanswered-my-tags)\n(define-key launcher-map \"qa\" #'sx-ask)\n(define-key launcher-map \"qs\" #'sx-search)<\/pre><\/div>\n<p>Please help contribute! Doing any of the following will help us immensely:<\/p>\n<p>For a better view of all of the open issues, take a look at our lovely <a href=\"http:\/\/www.waffle.io\/vermiculus\/sx.el\" rel=\"nofollow\">Waffle board<\/a>. Feel free to take the torch on anything in <code>backlog<\/code> or <code>ready<\/code>. If you have thoughts on any other issues, don\u2019t hesitate to chime in!<\/p>\n<p>See also <code>CONTRIBUTING.org<\/code>.<\/p>\n<ul><li><a href=\"http:\/\/www.gnu.org\/software\/emacs\/\" rel=\"nofollow\">GNU Emacs<\/a><\/li>\n<li><a href=\"https:\/\/api.stackexchange.com\/docs\" rel=\"nofollow\">Stack Exchange API v2.2<\/a><\/li>\n<\/ul><h2>Icons<\/h2>\n<p>SX has no explicit <em>need<\/em> for an icon, although standard SVG files have been gathered in <code>resources\/<\/code> if anyone would fancy a crack at it.<\/p>","created_at":"2019-02-12T13:45:27+0000","updated_at":"2020-01-07T15:53:43+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/2082195?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":833,"title":"emacs-lsp\/lsp-mode","url":"https:\/\/github.com\/emacs-lsp\/lsp-mode","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/melpa.org\/#\/lsp-mode\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/9af67d09707b47dd26115a48505d481f9dcdd805\/68747470733a2f2f6d656c70612e6f72672f7061636b616765732f6c73702d6d6f64652d62616467652e737667\" alt=\"https:\/\/melpa.org\/packages\/lsp-mode-badge.svg\" data-canonical-src=\"https:\/\/melpa.org\/packages\/lsp-mode-badge.svg\" \/><\/a>\n  <a href=\"https:\/\/stable.melpa.org\/#\/lsp-mode\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7a2cc5cd5fb1a171dde8e45435a353f849853123\/68747470733a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f6c73702d6d6f64652d62616467652e737667\" alt=\"https:\/\/stable.melpa.org\/packages\/lsp-mode-badge.svg\" data-canonical-src=\"https:\/\/stable.melpa.org\/packages\/lsp-mode-badge.svg\" \/><\/a>\n  <a href=\"https:\/\/gitter.im\/emacs-lsp\/lsp-mode\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/40a1fe7a717b706a101e6e1d917dfdb3f422a8c3\/68747470733a2f2f6261646765732e6769747465722e696d2f656d6163732d6c73702f6c73702d6d6f64652e737667\" alt=\"https:\/\/badges.gitter.im\/emacs-lsp\/lsp-mode.svg\" data-canonical-src=\"https:\/\/badges.gitter.im\/emacs-lsp\/lsp-mode.svg\" \/><\/a>\n  <a href=\"https:\/\/travis-ci.org\/emacs-lsp\/lsp-mode\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5674691682ca175709f1682ede1931f44a188aef\/68747470733a2f2f7472617669732d63692e6f72672f656d6163732d6c73702f6c73702d6d6f64652e7376673f6272616e63683d6d6173746572\" alt=\"https:\/\/travis-ci.org\/emacs-lsp\/lsp-mode.svg?branch=master\" data-canonical-src=\"https:\/\/travis-ci.org\/emacs-lsp\/lsp-mode.svg?branch=master\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/blob\/master\/examples\/head.png\"><img src=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/raw\/master\/examples\/head.png\" alt=\"examples\/head.png\" \/><\/a><\/p>\n<h2><a id=\"user-content-table-of-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-contents\"><\/a>Table of Contents<\/h2>\n<ul><li><a href=\"#language-server-protocol-support-for-emacs\">Language Server Protocol Support for Emacs<\/a>\n    <ul><li><a href=\"#why\">Why?<\/a><\/li>\n      <li><a href=\"#overview\">Overview<\/a><\/li>\n      <li><a href=\"#installation\">Installation<\/a>\n        <ul><li><a href=\"#install-via-melpa\">Install via melpa<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#configuration\">Configuration<\/a>\n        <ul><li><a href=\"#install-language-server\">Install language server<\/a><\/li>\n          <li><a href=\"#configure-lsp-mode\">Configure lsp-mode<\/a>\n            <ul><li><a href=\"#vanilla-emacs\">Vanilla Emacs<\/a><\/li>\n              <li><a href=\"#spacemacs\">Spacemacs<\/a><\/li>\n              <li><a href=\"#use-package\">use-package<\/a><\/li>\n            <\/ul><\/li>\n          <li><a href=\"#how-it-works\">How it works?<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#supported-languages\">Supported languages<\/a><\/li>\n      <li><a href=\"#commands\">Commands<\/a><\/li>\n      <li><a href=\"#settings\">Settings<\/a><\/li>\n      <li><a href=\"#screenshots\">Screenshots<\/a><\/li>\n      <li><a href=\"#extensions\">Extensions<\/a>\n        <ul><li><a href=\"#tramp-experimental\">TRAMP (Experimental)<\/a>\n            <ul><li><a href=\"#how-does-it-work\">How does it work?<\/a><\/li>\n              <li><a href=\"#sample-configuration\">Sample configuration<\/a><\/li>\n            <\/ul><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#troubleshooting\">Troubleshooting<\/a><\/li>\n      <li><a href=\"#adding-support-for-languages\">Adding support for languages<\/a><\/li>\n      <li><a href=\"#faq\">FAQ<\/a><\/li>\n      <li><a href=\"#see-also\">See also<\/a><\/li>\n    <\/ul><\/li>\n<\/ul><h2><a id=\"user-content-why\" class=\"anchor\" aria-hidden=\"true\" href=\"#why\"><\/a>Why?<\/h2>\n<ul><li>\u2764\ufe0f<b>Community Driven<\/b><\/li>\n  <li>\ud83d\udc8e<b>Fully featured<\/b> - supports LSP core and Language Server non-standard extensions<\/li>\n  <li>\ud83d\ude80<b>Easy to configure<\/b> - works out of the box and automatically upgrades if additional packages are present<\/li>\n  <li>\ud83c\udf1f<b>Flexible<\/b> - could be configured as full-blown IDE with flashy UI or minimal distraction free<\/li>\n<\/ul><h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2>\n<p>Client for <a href=\"https:\/\/github.com\/Microsoft\/language-server-protocol\/\">Language Server Protocol<\/a> (v3.14). <a href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\">lsp-mode<\/a> aims to provide IDE-like experience by providing optional integration with the most popular Emacs packages like <code>company<\/code>, <code>flycheck<\/code> and <code>projectile<\/code>.<\/p>\n<ul><li>Non-blocking asynchronous calls<\/li>\n  <li>Real-time Diagnostics\/linting (via builtin <code>flymake<\/code> when Emacs &gt; 26 or <a href=\"https:\/\/github.com\/flycheck\/flycheck\">flycheck<\/a>\/<a href=\"https:\/\/github.com\/emacs-lsp\/lsp-ui\">lsp-ui<\/a>)<\/li>\n  <li>Code completion - using <a href=\"https:\/\/github.com\/tigersoldier\/company-lsp\">company-lsp<\/a> or builtin <code>complete-at-point<\/code><\/li>\n  <li>Hovers - using <a href=\"https:\/\/github.com\/emacs-lsp\/lsp-ui\">lsp-ui<\/a><\/li>\n  <li>Code actions - using <code>lsp-execute-code-action<\/code> or <a href=\"https:\/\/github.com\/emacs-lsp\/lsp-ui\">lsp-ui<\/a><\/li>\n  <li>Code outline - using builtin <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Imenu.html\" rel=\"nofollow\">imenu<\/a> or <code>helm-imenu<\/code><\/li>\n  <li>Code navigation - using builtin <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Xref.html\" rel=\"nofollow\">xref<\/a><\/li>\n  <li>Code lens (references\/implementations) - using builtin <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Xref.html\" rel=\"nofollow\">xref<\/a><\/li>\n  <li>Highlights<\/li>\n  <li>Formatting<\/li>\n  <li>Debugger - <a href=\"https:\/\/github.com\/yyoncho\/dap-mode\/\">dap-mode<\/a><\/li>\n  <li>Helm integration - <a href=\"https:\/\/github.com\/yyoncho\/helm-lsp\/\">helm-lsp<\/a><\/li>\n<\/ul><h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<h3><a id=\"user-content-install-via-melpa\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-via-melpa\"><\/a>Install via melpa<\/h3>\n<p>The recommended way to install <code>lsp-mode<\/code> is via <code>package.el<\/code> - the built-in package manager in Emacs. <code>lsp-mode<\/code> is available on the two major <code>package.el<\/code> community maintained repos - <a href=\"http:\/\/stable.melpa.org\" rel=\"nofollow\">MELPA Stable<\/a> and <a href=\"http:\/\/melpa.org\" rel=\"nofollow\">MELPA<\/a>.<\/p>\n<p>M-x <code>package-install<\/code> [RET] <code>lsp-mode<\/code> [RET]<\/p>\n<h2><a id=\"user-content-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration\"><\/a>Configuration<\/h2>\n<h3><a id=\"user-content-install-language-server\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-language-server\"><\/a>Install language server<\/h3>\n<p>Check the table bellow with the list of supported servers and the corresponding instructions on how to install the server.<\/p>\n<h3><a id=\"user-content-configure-lsp-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#configure-lsp-mode\"><\/a>Configure lsp-mode<\/h3>\n<h4><a id=\"user-content-vanilla-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#vanilla-emacs\"><\/a>Vanilla Emacs<\/h4>\n<p>You could go minimal and use <code>lsp-mode<\/code> as it is without external packages with the built-in <code>flymake<\/code> and <code>completion-at-point<\/code> or you could install the following extensions for better experience:<\/p>\n<ul><li>install <a href=\"https:\/\/github.com\/emacs-lsp\/lsp-ui\">lsp-ui<\/a> for <a href=\"https:\/\/github.com\/flycheck\/flycheck\">flycheck<\/a> integration and higher level UI modules.<\/li>\n  <li>install <a href=\"https:\/\/github.com\/tigersoldier\/company-lsp\">company-lsp<\/a> if you want to use <code>company-mode<\/code> for completion.<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'lsp-mode)\n(add-hook 'XXX-mode-hook #'lsp)<\/pre><\/div>\n<p>where <code>XXX<\/code> could be major mode like <code>python<\/code>, <code>java<\/code>, <code>c++<\/code>. Alternatively, if you want to minimize you configuration you may use <code>prog-mode-hook<\/code>. In case you do that, <code>lsp<\/code> will try to start for each programming mode and echo a message when there is no client registered for the current mode or if the corresponding server is not present. In addition, <code>lsp-mode<\/code> will automatically detect and configure <a href=\"https:\/\/github.com\/emacs-lsp\/lsp-ui\">lsp-ui<\/a> and <a href=\"https:\/\/github.com\/tigersoldier\/company-lsp\">company-lsp<\/a>. To turn off that behavior you could set <code>lsp-auto-configure<\/code> to <code>nil<\/code>.<\/p>\n<h4><a id=\"user-content-spacemacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#spacemacs\"><\/a>Spacemacs<\/h4>\n<p><a href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\">lsp-mode<\/a> is included in spacemacs develop branch. Add <code>lsp<\/code> to <code>dotspacemacs-configuration-layers<\/code> and configure the language that you want to use to be backed by <code>lsp<\/code> backend.<\/p>\n<h4><a id=\"user-content-use-package\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-package\"><\/a>use-package<\/h4>\n<p>Replace <code>(require 'lsp-mode)<\/code> with the following if you use use-package.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lsp-mode\n  :commands lsp\n  :init\n  (setq ...))\n(use-package lsp-ui :commands lsp-ui-mode)\n(use-package company-lsp :commands company-lsp)<\/pre><\/div>\n<h3><a id=\"user-content-how-it-works\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-it-works\"><\/a>How it works?<\/h3>\n<p><code>lsp-mode<\/code> has predefined list of server configurations (located in <code>lsp-clients.el<\/code>) containing a mapping from <code>major-mode<\/code> to the server configuration. In addition to the default server configuration located in <code>lsp-clients.el<\/code> there are few languages servers which require separate package(check <a href=\"#supported-languages\">Supported languages<\/a>). When you open a file from a particular project <code>lsp-mode<\/code> and call <code>lsp<\/code> command <code>lsp-mode<\/code> will look for server registrations able to handle current <code>major-mode<\/code>. If there is such client <code>lsp-mode<\/code> will look for the project root. If you open a file from the project for the first time you will be prompted the project. Once the project is imported project root is saved in <code>lsp-session<\/code> file and it will be loaded the next time you start Emacs so you no longer will be asked for a project root when you open a file from that project. Later if you want to change the project root you may use <code>lsp-workspace-folder-remove<\/code> to remove the project and call <code>lsp-workspace-folder-add<\/code> to add the root. If you want to force starting a particular language server in a file you may use <code>C-u<\/code> <code>M-x<\/code> <code>lsp<\/code> which will prompt you to select language server to start.<\/p>\n<h2><a id=\"user-content-supported-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-languages\"><\/a>Supported languages<\/h2>\n<p>Some of the servers are directly supported by <code>lsp-mode<\/code> by requiring\n  <code>lsp-clients.el<\/code> while other require installing additional package which provide\n  server specific functionality.<\/p>\n<h2><a id=\"user-content-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#commands\"><\/a>Commands<\/h2>\n<ul><li><code>lsp-describe-session<\/code> - Display session folders and running servers.<\/li>\n  <li><code>lsp-describe-thing-at-point<\/code> - Display help for the thing at point.<\/li>\n  <li><code>lsp-execute-code-action<\/code> - Execute code action.<\/li>\n  <li><code>lsp-format-buffer<\/code> - Format current buffer<\/li>\n  <li><code>lsp-goto-implementation<\/code> - Go to implementation<\/li>\n  <li><code>lsp-goto-type-definition<\/code> - Go to type definition<\/li>\n  <li><code>lsp-rename<\/code> - Rename symbol at point<\/li>\n  <li><code>lsp-restart-workspace<\/code> - Restart project<\/li>\n  <li><code>lsp-symbol-highlight<\/code> - Highlight all relevant references to the symbol under point.<\/li>\n  <li><code>lsp-workspace-folders-add<\/code> - Add workspace folder<\/li>\n  <li><code>lsp-workspace-folders-remove<\/code> - Remove workspace folder<\/li>\n  <li><code>lsp-workspace-folders-switch<\/code> - Switch workspace folder<\/li>\n  <li><code>imenu<\/code> or <code>helm-imenu<\/code> - display document structure.<\/li>\n  <li><code>completion-at-point<\/code> - display completion using built-in emacs <code>completion-at-point<\/code> framework.<\/li>\n  <li><code>lsp-find-definition<\/code> - to find the definition for the symbol under point.<\/li>\n  <li><code>lsp-find-references<\/code> - Find references for the symbol under point.<\/li>\n  <li><code>lsp-lens-show<\/code> - Show lenses in the current file<\/li>\n  <li><code>lsp-lens-hide<\/code> - Hide lenses in the current file<\/li>\n  <li><code>lsp-lens-mode<\/code> (experimental) - Turn on\/off lenses in the current file.<\/li>\n<\/ul><h2><a id=\"user-content-settings\" class=\"anchor\" aria-hidden=\"true\" href=\"#settings\"><\/a>Settings<\/h2>\n<ul><li><code>lsp-print-io<\/code> - If non-nil, print all messages to and from the language server to <code>*lsp-log*<\/code>.<\/li>\n  <li><code>lsp-print-performance<\/code> - If non-nil, print performance info. to <code>*lsp-log*<\/code>.<\/li>\n  <li><code>lsp-inhibit-message<\/code> - If non-nil, inhibit the message echo via <code>inhibit-message<\/code>.<\/li>\n  <li><code>lsp-report-if-no-buffer<\/code> - If non nil the errors will be reported even when the file is not open.<\/li>\n  <li><code>lsp-keep-workspace-alive<\/code> - If non nil keep workspace alive when the last workspace buffer is closed.<\/li>\n  <li><code>lsp-enable-snippet<\/code> - Enable\/disable snippet completion support.<\/li>\n  <li><code>lsp-auto-guess-root<\/code> - Automatically guess the project root using projectile\/project.<\/li>\n  <li><code>lsp-restart<\/code> - Defines how server exited event must be handled.<\/li>\n  <li><code>lsp-session-file<\/code> - Automatically guess the project root using projectile\/project.<\/li>\n  <li><code>lsp-auto-configure<\/code> - Auto configure <code>lsp-mode<\/code>. When set to t <code>lsp-mode<\/code> will auto-configure <code>lsp-ui<\/code> and <code>company-lsp<\/code>.<\/li>\n  <li><code>lsp-document-sync-method<\/code> - How to sync the document with the language server.<\/li>\n  <li><code>lsp-auto-execute-action<\/code> - Auto-execute single action.<\/li>\n  <li><code>lsp-eldoc-render-all<\/code> - Define whether all of the returned by <code>document\/onHover<\/code> will be displayed. If <code>lsp-markup-display-all<\/code> is set to nil <code>eldoc<\/code> will show only the symbol information.<\/li>\n  <li><code>lsp-enable-completion-at-point<\/code> - Enable <code>completion-at-point<\/code> integration.<\/li>\n  <li><code>lsp-enable-xref<\/code> - Enable xref integration.<\/li>\n  <li><code>lsp-prefer-flymake<\/code> - If you prefer flycheck and <code>lsp-ui-flycheck<\/code> is available, <code>(setq lsp-prefer-flymake nil)<\/code>. If set to <code>:none<\/code> neither of two will be enabled.<\/li>\n  <li><code>lsp-enable-indentation<\/code> - Indent regions using the file formatting functionality provided by the language server.<\/li>\n  <li><code>lsp-enable-on-type-formatting<\/code> - Enable <code>textDocument\/onTypeFormatting<\/code> integration.<\/li>\n  <li><code>lsp-before-save-edits<\/code> - If non-nil, <code>lsp-mode<\/code> will apply edits suggested by the language server before saving a document.<\/li>\n  <li><code>lsp-imenu-show-container-name<\/code> - Display the symbol\u2019s container name in an imenu entry.<\/li>\n  <li><code>lsp-imenu-container-name-separator<\/code> - Separator string to use to separate the container name from the symbol while displaying imenu entries.<\/li>\n  <li><code>lsp-imenu-sort-methods<\/code> - How to sort the imenu items. The value is a list of <code>kind<\/code>, <code>name<\/code> or <code>position<\/code>. Priorities are determined by the index of the element.<\/li>\n  <li><code>lsp-response-timeout<\/code> - Number of seconds to wait for a response from the language server before timing out.<\/li>\n<\/ul><h2><a id=\"user-content-screenshots\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshots\"><\/a>Screenshots<\/h2>\n<ul><li>RUST Completion with company-lsp\n    <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/blob\/master\/examples\/completion.png\"><img src=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/raw\/master\/examples\/completion.png\" alt=\"examples\/completion.png\" \/><\/a><\/li>\n  <li>Typescript references using lsp-ui\n    <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/blob\/master\/examples\/references.png\"><img src=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/raw\/master\/examples\/references.png\" alt=\"examples\/references.png\" \/><\/a><\/li>\n  <li>Debugging Python using dap-mode\n    <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/blob\/master\/examples\/python_debugging.png\"><img src=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/raw\/master\/examples\/python_debugging.png\" alt=\"examples\/python_debugging.png\" \/><\/a><\/li>\n  <li>Call hierarchy via ccls\n    <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/blob\/master\/examples\/call-hierarchy-ccls.png\"><img src=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/raw\/master\/examples\/call-hierarchy-ccls.png\" alt=\"examples\/call-hierarchy-ccls.png\" \/><\/a><\/li>\n<\/ul><h2><a id=\"user-content-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#extensions\"><\/a>Extensions<\/h2>\n<h3><a id=\"user-content-tramp-experimental\" class=\"anchor\" aria-hidden=\"true\" href=\"#tramp-experimental\"><\/a>TRAMP (Experimental)<\/h3>\n<p>LSP mode has support for tramp buffers with the following requirements.<\/p>\n<ul><li>The language server have to be present on the remote server since it is running on the remote machine.<\/li>\n  <li>Having multi folder language server (like <a href=\"https:\/\/github.com\/eclipse\/eclipse.jdt.ls\">Eclipse JDT LS<\/a>) cannot have local and remote servers.<\/li>\n<\/ul><h4><a id=\"user-content-how-does-it-work\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-does-it-work\"><\/a>How does it work?<\/h4>\n<p><code>lsp-mode<\/code> detects whether a particular file is located on remote machine and looks for a client which can handle the current major mode and it is marked as <code>:remote?<\/code> t. Then <code>lsp-mode<\/code> starts the client through tramp.<\/p>\n<h4><a id=\"user-content-sample-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#sample-configuration\"><\/a>Sample configuration<\/h4>\n<p>Here it is example how you can configure python language server to work when using <code>TRAMP<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(lsp-register-client\n (make-lsp-client :new-connection (lsp-tramp-connection \"pyls\")\n                  :major-modes '(python-mode)\n                  :remote? t\n                  :server-id 'pyls))<\/pre><\/div>\n<h2><a id=\"user-content-troubleshooting\" class=\"anchor\" aria-hidden=\"true\" href=\"#troubleshooting\"><\/a>Troubleshooting<\/h2>\n<ul><li>set <code>lsp-print-io<\/code> to <code>t<\/code> to inspect communication between client and the server.<\/li>\n  <li><code>lsp-describe-session<\/code> will show the current projects roots + the started severs and allows inspecting the server capabilities.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/blob\/master\/examples\/describe.png\"><img src=\"https:\/\/github.com\/emacs-lsp\/lsp-mode\/raw\/master\/examples\/describe.png\" alt=\"examples\/describe.png\" \/><\/a><\/p>\n<h2><a id=\"user-content-adding-support-for-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#adding-support-for-languages\"><\/a>Adding support for languages<\/h2>\n<p>Here it is the minimal configuration that is needed for new language server registration. Refer to the documentation of <code>lsp-client.el<\/code> for the additional settings supported on registration time. <code>lsp-language-id-configuration<\/code> must be updated to contain the corresponding mode -&gt; language id - in this case <code>(python-mode . \"python\")<\/code><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(lsp-register-client\n (make-lsp-client :new-connection (lsp-stdio-connection \"pyls\")\n                  :major-modes '(python-mode)\n                  :server-id 'pyls))<\/pre><\/div>\n<h2><a id=\"user-content-faq\" class=\"anchor\" aria-hidden=\"true\" href=\"#faq\"><\/a>FAQ<\/h2>\n<ul><li>How to configure a server with local variables?\n    Add <code>lsp<\/code> server call to <code>hack-local-variables-hook<\/code> which runs right after the local variables are loaded.\n    <div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'hack-local-variables-hook\n          (lambda () (when (derived-mode-p 'XXX-mode) (lsp))))\n    <\/pre><\/div>\n  <\/li>\n  <li>I have multiple language servers registered for language FOO. Which one will be used when opening a project?\n    The one with highest priority wins. <code>lsp-clients.el<\/code> predefined servers have\n    priority -1, lower than external packages (priority 0 if unspecified). If a\n    server is registered with <code>:add-on?<\/code> flag set to <code>t<\/code> it will be started in\n    parallel to the other servers that are registered for the current mode.<\/li>\n<\/ul><h2><a id=\"user-content-see-also\" class=\"anchor\" aria-hidden=\"true\" href=\"#see-also\"><\/a>See also<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/yyoncho\/dap-mode\">dap-mode<\/a> - Debugger integration for <code>lsp-mode<\/code>.<\/li>\n  <li><a href=\"https:\/\/github.com\/joaotavora\/eglot\">eglot<\/a> - An alternative minimal LSP implementation.<\/li>\n<\/ul><\/article>","created_at":"2019-02-12T13:43:33+0000","updated_at":"2019-02-12T13:43:33+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/27148249?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":832,"title":"kiwanami\/emacs-calfw","url":"https:\/\/github.com\/kiwanami\/emacs-calfw\/blob\/master\/readme.md","content":"<h2><a id=\"user-content-what-is-calfw\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-is-calfw\"><\/a>What is calfw?<\/h2>\n<p>This program displays a calendar view in the Emacs buffer.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/bcf0ca5fd834e2ea161099edfcb8bc969edebbc1\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d39453545302e706e673f77696474683d363030\"><img src=\"https:\/\/camo.githubusercontent.com\/bcf0ca5fd834e2ea161099edfcb8bc969edebbc1\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d39453545302e706e673f77696474683d363030\" alt=\"Calfw image\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/OnjKgBHat0kHs0xp-9E5E0.png?width=600\" \/><\/a><\/p>\n<h3><a id=\"user-content-screenshots\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshots\"><\/a>Screenshots<\/h3>\n<p>Currently, calfw has 4 views, month, 1week, 2week and day view.\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/7fe05e2e0f9c31eb5cc8f48143bc8d932dbe3bfd\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d46333735362e706e673f77696474683d363030\"><img src=\"https:\/\/camo.githubusercontent.com\/7fe05e2e0f9c31eb5cc8f48143bc8d932dbe3bfd\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d46333735362e706e673f77696474683d363030\" alt=\"Views\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/OnjKgBHat0kHs0xp-F3756.png?width=600\" \/><\/a><\/p>\n<p>Pushing SPC key, the detail buffer pops up. Pushing SPC key again, the buffer is closed.\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/75b096ab19bbf3865d9edd8e9388a8859f11debc\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d38334338302e706e673f77696474683d363030\"><img src=\"https:\/\/camo.githubusercontent.com\/75b096ab19bbf3865d9edd8e9388a8859f11debc\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d38334338302e706e673f77696474683d363030\" alt=\"Pop up details\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/OnjKgBHat0kHs0xp-83C80.png?width=600\" \/><\/a><\/p>\n<p>Many information items are displayed in the Emacs buffer.\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/61db8f948f27ab1b2254ab2e5fae07da3aa0432d\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d42393631422e706e673f77696474683d363030\"><img src=\"https:\/\/camo.githubusercontent.com\/61db8f948f27ab1b2254ab2e5fae07da3aa0432d\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f4f6e6a4b6742486174306b48733078702d42393631422e706e673f77696474683d363030\" alt=\"View details\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/OnjKgBHat0kHs0xp-B961B.png?width=600\" \/><\/a><\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>To use this program, locate this file to load-path directory,\nand add the following code to your .emacs.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'calfw)<\/pre><\/div>\n<p>Executing the command <code>cfw:open-calendar-buffer<\/code>, switch to the calendar buffer.\nYou can navigate the date like calendar.el.<\/p>\n<p>Schedule data which are shown in the calendar view, are collected by a\nlist of the struct <code>cfw:source<\/code> objects through the named argument\nvariables <code>:contents-sources<\/code> and <code>:annotation-sources<\/code>. The former\nsource defines schedule contents. The later one does date\nannotations like the moon phases.<\/p>\n<p>This program gets the holidays using the function\n<code>calendar-holiday-list<\/code>. See the document for the holidays.el and the Info text.<\/p>\n<h2><a id=\"user-content-key-bindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#key-bindings\"><\/a>Key bindings<\/h2>\n<p>In the calendar buffer and region, you can use following key bindings:<\/p>\n<table><thead><tr><th>Navigation<\/th>\n<\/tr><\/thead><tbody><tr><td>[left], b, h<\/td>\n<td>Previous day<\/td>\n<\/tr><tr><td>[right], f, l<\/td>\n<td>Next day<\/td>\n<\/tr><tr><td>[up], p, k<\/td>\n<td>Previous week<\/td>\n<\/tr><tr><td>[down], n, j<\/td>\n<td>Next week<\/td>\n<\/tr><tr><td>^<\/td>\n<td>Week begin<\/td>\n<\/tr><tr><td>$<\/td>\n<td>Week end<\/td>\n<\/tr><tr><td>[home]<\/td>\n<td>First date in this month<\/td>\n<\/tr><tr><td>[end]<\/td>\n<td>Last date in this month<\/td>\n<\/tr><tr><td>M-v, [PgUp], &lt;<\/td>\n<td>Previous month<\/td>\n<\/tr><tr><td>C-v, [PgDown], &gt;<\/td>\n<td>Next month<\/td>\n<\/tr><tr><td>t<\/td>\n<td>Today<\/td>\n<\/tr><tr><td>g<\/td>\n<td>Absolute date (YYYY\/MM\/DD)<\/td>\n<\/tr><tr><td>TAB<\/td>\n<td>Next item in a day<\/td>\n<\/tr><\/tbody><\/table><table><thead><tr><th>Changing View<\/th>\n<\/tr><\/thead><tbody><tr><td>M<\/td>\n<td>Month view<\/td>\n<\/tr><tr><td>W<\/td>\n<td>1 Week view<\/td>\n<\/tr><tr><td>T<\/td>\n<td>2 Week view<\/td>\n<\/tr><tr><td>D<\/td>\n<td>Day view<\/td>\n<\/tr><\/tbody><\/table><table><thead><tr><th>Operation<\/th>\n<\/tr><\/thead><tbody><tr><td>r<\/td>\n<td>Refresh data and re-draw contents<\/td>\n<\/tr><tr><td>SPC<\/td>\n<td>Pop-up detail buffer (like Quicklook in Mac)<\/td>\n<\/tr><tr><td>RET, [click]<\/td>\n<td>Jump (howm, orgmode)<\/td>\n<\/tr><tr><td>q<\/td>\n<td>Bury buffer<\/td>\n<\/tr><\/tbody><\/table><p>The buttons on the toolbar can be clicked.<\/p>\n<h2><a id=\"user-content-add-ons\" class=\"anchor\" aria-hidden=\"true\" href=\"#add-ons\"><\/a>Add-ons:<\/h2>\n<p>Following programs are also useful:<\/p>\n<ul><li>calfw-howm.el : Display howm schedules (<a href=\"http:\/\/howm.sourceforge.jp\/index.html\" rel=\"nofollow\">http:\/\/howm.sourceforge.jp\/index.html<\/a>)<\/li>\n<li>calfw-ical.el : Display schedules of the iCalendar format, such as the google calendar.<\/li>\n<li>calfw-org.el  : Display org schedules (<a href=\"http:\/\/orgmode.org\/\" rel=\"nofollow\">http:\/\/orgmode.org\/<\/a>)<\/li>\n<li>calfw-cal.el  : Display diary schedules.<\/li>\n<\/ul><h2><a id=\"user-content-setting-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#setting-example\"><\/a>Setting example:<\/h2>\n<h3><a id=\"user-content-for-howm-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-howm-users\"><\/a>For howm users:<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(eval-after-load \"howm-menu\" '(progn\n  (require 'calfw-howm)\n  (cfw:install-howm-schedules)\n  (define-key howm-mode-map (kbd \"M-C\") 'cfw:open-howm-calendar)\n))<\/pre><\/div>\n<p>If you are using Elscreen, here is useful.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key howm-mode-map (kbd \"M-C\") 'cfw:elscreen-open-howm-calendar)<\/pre><\/div>\n<p>You can display a calendar in your howm menu file.<\/p>\n<pre>%here%(cfw:howm-schedule-inline)\n<\/pre>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/1416e559ea0d2e276327c1a2a2c429c1131ead9c\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f7672536349344b32516c6d44417066642d31463934312e706e673f77696474683d343530\"><img src=\"https:\/\/camo.githubusercontent.com\/1416e559ea0d2e276327c1a2a2c429c1131ead9c\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f7672536349344b32516c6d44417066642d31463934312e706e673f77696474683d343530\" alt=\"howm menu embedding\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/vrScI4K2QlmDApfd-1F941.png?width=450\" \/><\/a><\/p>\n<h3><a id=\"user-content-for-org-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-org-users\"><\/a>For org users:<\/h3>\n<pre>(require 'calfw-org)\n<\/pre>\n<p>Then, M-x <code>cfw:open-org-calendar<\/code>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/585347cc96320b0ccff66bb9cff4e0320b42f984\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5336614a6e744736676947733434596e2d38394342322e706e673f77696474683d343530\"><img src=\"https:\/\/camo.githubusercontent.com\/585347cc96320b0ccff66bb9cff4e0320b42f984\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5336614a6e744736676947733434596e2d38394342322e706e673f77696474683d343530\" alt=\"org-agenda and calfw-org\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/S6aJntG6giGs44Yn-89CB2.png?width=450\" \/><\/a><\/p>\n<h4><a id=\"user-content-filtering-agenda-items\" class=\"anchor\" aria-hidden=\"true\" href=\"#filtering-agenda-items\"><\/a>Filtering agenda items<\/h4>\n<p>You can choose agenda items with <code>cfw:org-agenda-schedule-args<\/code>, like following code:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cfw:org-agenda-schedule-args '(:timestamp))<\/pre><\/div>\n<p>This setting restricts items containing a date stamp or date range matching the selected date.\nIf <code>cfw:org-agenda-schedule-args<\/code> is <code>nil<\/code>, the default customize variable <code>org-agenda-entry-types<\/code> is used. For the further information, please refer the orgmode document.<\/p>\n<ul><li><a href=\"http:\/\/orgmode.org\/worg\/org-tutorials\/org-custom-agenda-commands.html#sec-5\" rel=\"nofollow\">Worg: Speeding up custom agenda commands<\/a><\/li>\n<\/ul><h4><a id=\"user-content-orgmode-like-key-bindng\" class=\"anchor\" aria-hidden=\"true\" href=\"#orgmode-like-key-bindng\"><\/a>Orgmode like key bindng<\/h4>\n<p>You can use another key binding like org agenda buffer, setting <code>cfw:org-overwrite-default-keybinding<\/code> to non-nil, like following code:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cfw:org-overwrite-default-keybinding t)<\/pre><\/div>\n<p>Then, following key bindings are overwritten:<\/p>\n<table><thead><tr><th>key<\/th>\n<th>function<\/th>\n<\/tr><\/thead><tbody><tr><td>g<\/td>\n<td>Refresh data and re-draw contents (cfw:refresh-calendar-buffer)<\/td>\n<\/tr><tr><td>j<\/td>\n<td>Goto the specified date (cfw:org-goto-date)<\/td>\n<\/tr><tr><td>k<\/td>\n<td>org-capture<\/td>\n<\/tr><tr><td>x<\/td>\n<td>Close calfw and other buffers opened by calfw-org (cfw:org-clean-exit)<\/td>\n<\/tr><tr><td>d<\/td>\n<td>Day view (cfw:change-view-day)<\/td>\n<\/tr><tr><td>v d<\/td>\n<td>Day view (cfw:change-view-day)<\/td>\n<\/tr><tr><td>v w<\/td>\n<td>1 Week view (cfw:change-view-week)<\/td>\n<\/tr><tr><td>v m<\/td>\n<td>Month View (cfw:change-view-month)<\/td>\n<\/tr><\/tbody><\/table><h4><a id=\"user-content-synchronization-with-google-calendar\" class=\"anchor\" aria-hidden=\"true\" href=\"#synchronization-with-google-calendar\"><\/a>Synchronization with google calendar<\/h4>\n<p>Here is the program which helps synchronization schedule items between org and google calendar, and also collaborates with calfw.<\/p>\n<h3><a id=\"user-content-for-ical-google-calendar-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-ical-google-calendar-users\"><\/a>For iCal (Google Calendar) users:<\/h3>\n<p>Here is a minimum sample code:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'calfw-ical)\n(cfw:open-ical-calendar \"http:\/\/www.google.com\/calendar\/ical\/...\/basic.ics\")<\/pre><\/div>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/95fea69b30f65b836d74b815a3f4fceb1dfd6aa4\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f7672536349344b32516c6d44417066642d35453830382e706e673f77696474683d343530\"><img src=\"https:\/\/camo.githubusercontent.com\/95fea69b30f65b836d74b815a3f4fceb1dfd6aa4\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f7672536349344b32516c6d44417066642d35453830382e706e673f77696474683d343530\" alt=\"Google Calendar and calfw-ical\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/vrScI4K2QlmDApfd-5E808.png?width=450\" \/><\/a><\/p>\n<p>Here is the add-on program which communicate with google calendar via API:<\/p>\n<ul><li><a href=\"http:\/\/sheephead.homelinux.org\/2011\/01\/18\/6559\/\" rel=\"nofollow\">calfw\u304b\u3089Google\u30ab\u30ec\u30f3\u30c0\u30fc\u3092\u7de8\u96c6\u3059\u308bcalfw-gcal.el\u3092\u66f8\u3044\u3066\u307f\u305f<\/a>\n<\/li>\n<\/ul><h3><a id=\"user-content-for-diary-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-diary-users\"><\/a>For diary users:<\/h3>\n<p>Here is a minimum sample code:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'calfw-cal)<\/pre><\/div>\n<p>Then, M-x <code>cfw:open-diary-calendar<\/code>.<\/p>\n<p>If you see a blank entry for each day, set the variable <code>diary-list-include-blanks<\/code> to nil.<\/p>\n<h3><a id=\"user-content-general-setting\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-setting\"><\/a>General setting<\/h3>\n<p>The calfw view can display many schedule items, gathering some schedule sources.\nUsing the function <code>cfw:open-calendar-buffer<\/code> is the general way to display the schedules.<\/p>\n<p>Here is the sample code:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'calfw-cal)\n(require 'calfw-ical)\n(require 'calfw-howm)\n(require 'calfw-org)\n(defun my-open-calendar ()\n  (interactive)\n  (cfw:open-calendar-buffer\n   :contents-sources\n   (list\n    (cfw:org-create-source \"Green\")  ; orgmode source\n    (cfw:howm-create-source \"Blue\")  ; howm source\n    (cfw:cal-create-source \"Orange\") ; diary source\n    (cfw:ical-create-source \"Moon\" \"~\/moon.ics\" \"Gray\")  ; ICS source1\n    (cfw:ical-create-source \"gcal\" \"https:\/\/....\/basic.ics\" \"IndianRed\") ; google calendar ICS\n   ))) <\/pre><\/div>\n<p>The function <code>cfw:open-calendar-buffer<\/code> receives schedules sources via\nthe named argument <code>:contents-sources<\/code>.<\/p>\n<p>One can customize the keymap on the calendar buffer with the named\nargument <code>:custom-map<\/code> of <code>cfw:open-calendar-buffer<\/code>.<\/p>\n<h2><a id=\"user-content-customize\" class=\"anchor\" aria-hidden=\"true\" href=\"#customize\"><\/a>Customize<\/h2>\n<h3><a id=\"user-content-holidays\" class=\"anchor\" aria-hidden=\"true\" href=\"#holidays\"><\/a>Holidays<\/h3>\n<p>The calfw collects holidays from the customize variable\n<code>calendar-holidays<\/code> which belongs to holidays.el in the Emacs. See the\ndocument and source of holidays.el for details.<\/p>\n<h3><a id=\"user-content-format-of-month-and-week-days\" class=\"anchor\" aria-hidden=\"true\" href=\"#format-of-month-and-week-days\"><\/a>Format of month and week days<\/h3>\n<p>The calfw uses some customization variables in the calendar.el.<\/p>\n<p>Here is a customization code:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Month\n(setq calendar-month-name-array\n  [\"January\" \"February\" \"March\"     \"April\"   \"May\"      \"June\"\n   \"July\"    \"August\"   \"September\" \"October\" \"November\" \"December\"])\n;; Week days\n(setq calendar-day-name-array\n      [\"Sunday\" \"Monday\" \"Tuesday\" \"Wednesday\" \"Thursday\" \"Friday\" \"Saturday\"])\n;; First day of the week\n(setq calendar-week-start-day 0) ; 0:Sunday, 1:Monday<\/pre><\/div>\n<h3><a id=\"user-content-faces\" class=\"anchor\" aria-hidden=\"true\" href=\"#faces\"><\/a>Faces<\/h3>\n<p>One can customize the faces.<\/p>\n<p>Here is a template code for face customization:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(custom-set-faces\n '(cfw:face-title ((t (:foreground \"#f0dfaf\" :weight bold :height 2.0 :inherit variable-pitch))))\n '(cfw:face-header ((t (:foreground \"#d0bf8f\" :weight bold))))\n '(cfw:face-sunday ((t :foreground \"#cc9393\" :background \"grey10\" :weight bold)))\n '(cfw:face-saturday ((t :foreground \"#8cd0d3\" :background \"grey10\" :weight bold)))\n '(cfw:face-holiday ((t :background \"grey10\" :foreground \"#8c5353\" :weight bold)))\n '(cfw:face-grid ((t :foreground \"DarkGrey\")))\n '(cfw:face-default-content ((t :foreground \"#bfebbf\")))\n '(cfw:face-periods ((t :foreground \"cyan\")))\n '(cfw:face-day-title ((t :background \"grey10\")))\n '(cfw:face-default-day ((t :weight bold :inherit cfw:face-day-title)))\n '(cfw:face-annotation ((t :foreground \"RosyBrown\" :inherit cfw:face-day-title)))\n '(cfw:face-disable ((t :foreground \"DarkGray\" :inherit cfw:face-day-title)))\n '(cfw:face-today-title ((t :background \"#7f9f7f\" :weight bold)))\n '(cfw:face-today ((t :background: \"grey10\" :weight bold)))\n '(cfw:face-select ((t :background \"#2f2f2f\")))\n '(cfw:face-toolbar ((t :foreground \"Steelblue4\" :background \"Steelblue4\")))\n '(cfw:face-toolbar-button-off ((t :foreground \"Gray10\" :weight bold)))\n '(cfw:face-toolbar-button-on ((t :foreground \"Gray50\" :weight bold))))<\/pre><\/div>\n<h3><a id=\"user-content-grid-frame\" class=\"anchor\" aria-hidden=\"true\" href=\"#grid-frame\"><\/a>Grid frame<\/h3>\n<p>Users can have nice unicode grid frame. However, in the some environment, the Emacs can not display the grid characters correctly. Please try following settings.<\/p>\n<p>Grid setting example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Default setting\n(setq cfw:fchar-junction ?+\n      cfw:fchar-vertical-line ?|\n      cfw:fchar-horizontal-line ?-\n      cfw:fchar-left-junction ?+\n      cfw:fchar-right-junction ?+\n      cfw:fchar-top-junction ?+\n      cfw:fchar-top-left-corner ?+\n      cfw:fchar-top-right-corner ?+ )\n;; Unicode characters\n(setq cfw:fchar-junction ?\u254b\n      cfw:fchar-vertical-line ?\u2503\n      cfw:fchar-horizontal-line ?\u2501\n      cfw:fchar-left-junction ?\u2523\n      cfw:fchar-right-junction ?\u252b\n      cfw:fchar-top-junction ?\u252f\n      cfw:fchar-top-left-corner ?\u250f\n      cfw:fchar-top-right-corner ?\u2513)\n      \n;; Another unicode chars\n(setq cfw:fchar-junction ?\u256c\n      cfw:fchar-vertical-line ?\u2551\n      cfw:fchar-horizontal-line ?\u2550\n      cfw:fchar-left-junction ?\u2560\n      cfw:fchar-right-junction ?\u2563\n      cfw:fchar-top-junction ?\u2566\n      cfw:fchar-top-left-corner ?\u2554\n      cfw:fchar-top-right-corner ?\u2557)<\/pre><\/div>\n<h3><a id=\"user-content-line-breaking\" class=\"anchor\" aria-hidden=\"true\" href=\"#line-breaking\"><\/a>Line breaking<\/h3>\n<p>If a content string is longer than the cell width, the calfw breaks into the multiple lines.\nIn the current implementation, the Calfw has 3 strategies: none, simple and wordwrap. The variable <code>cfw:render-line-breaker<\/code> selects the strategy to break lines.<\/p>\n<ul><li><code>cfw:render-line-breaker-none<\/code>\n<ul><li>Never breaks lines. Longer contents are truncated.<\/li>\n<\/ul><\/li>\n<li><code>cfw:render-line-breaker-simple<\/code> (default)\n<ul><li>This strategy breaks lines with rigid width. This may be not so beautiful, but In the most cases it looks good.<\/li>\n<\/ul><\/li>\n<li><code>cfw:render-line-breaker-wordwrap<\/code>\n<ul><li>This strategy breaks lines with the emacs function <code>fill-region<\/code>. Although, the line breaking algorithm of the Emacs is not so smart as more complicated ones, such as Knuth\/Plass algorithm, this strategy is better than the simple one.<\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-calfw-framework-details\" class=\"anchor\" aria-hidden=\"true\" href=\"#calfw-framework-details\"><\/a>Calfw framework details<\/h2>\n<p>In this section, I would explain how to add a new calendar source and how to embed the calfw component in the other applications.<\/p>\n<h3><a id=\"user-content-how-to-add-a-new-calendar-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-add-a-new-calendar-source\"><\/a>How to add a new calendar source?<\/h3>\n<p>Defining the <code>cfw:source<\/code> object, one can extend calfw calendar source.<\/p>\n<h4><a id=\"user-content-struct-cfwsource-details\" class=\"anchor\" aria-hidden=\"true\" href=\"#struct-cfwsource-details\"><\/a>struct cfw:source details<\/h4>\n<p>The struct <code>cfw:source<\/code> is a simple data type defined by cl-defstruct.<\/p>\n<p>Here is the details of the slot members of <code>cfw:source<\/code>.<\/p>\n<table><thead><tr><th>slot name<\/th>\n<th>description<\/th>\n<\/tr><\/thead><tbody><tr><td>name<\/td>\n<td>[required] Source name. This name is shown at the status bar.<\/td>\n<\/tr><tr><td>data<\/td>\n<td>[required] Data function which returns calendar contents. The function details are described in the next section.<\/td>\n<\/tr><tr><td>update<\/td>\n<td>[option] Update function. Calfw calls this function when this source needs to refresh the data.<\/td>\n<\/tr><tr><td>color<\/td>\n<td>[option] Color string for this source.  Color names those are shown by <code>M-x list-colors-display<\/code> or RGB hex format like \"#abcdef\".<\/td>\n<\/tr><tr><td>period-fgcolor<\/td>\n<td>[option] Foreground color for period items. The default color is white or black.<\/td>\n<\/tr><tr><td>period-bgcolor<\/td>\n<td>[option] Background color for period items. The default color is <code>cfw:source-color<\/code>.<\/td>\n<\/tr><tr><td>opt-face<\/td>\n<td>[option] Additional options for the normal item face.  Ex. <code>:opt-face '(:weight bold)<\/code><\/td>\n<\/tr><tr><td>opt-period-face<\/td>\n<td>[option] Additional options for the period item face.<\/td>\n<\/tr><\/tbody><\/table><p>Only <code>name<\/code> and <code>data<\/code> slots are essential. Many slots are visual options.<\/p>\n<p>In many cases, one has to specify only the <code>color<\/code> slot for visual,\nbecause the calfw chooses appropriate colors for the rest color options.<\/p>\n<h4><a id=\"user-content-cfwsource-data-details\" class=\"anchor\" aria-hidden=\"true\" href=\"#cfwsource-data-details\"><\/a>cfw:source-data details<\/h4>\n<p>This section explains what objects the function-slot <code>cfw:source-data<\/code> should return.<\/p>\n<p>The function-slot <code>cfw:source-data<\/code> receives two arguments, start and\nend date of the query period, and returns a list of instances of <code>cfw:event<\/code> struct.<\/p>\n<p>Here is a simple example.<\/p>\n<p><code>cfw:source-data example1:<\/code><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; cfw:source-data example\n(defun sample-data1 (b e)\n  (list\n    (make-cfw:event :title \"item1\"   :start-date  (cfw:date 1 1 2011))\n    (make-cfw:event :title \"item2-1\" :start-date  (cfw:date 1 10 2011))\n    (make-cfw:event :title \"item2-2\" :start-date  (cfw:date 1 10 2011))\n    ))\n(cfw:open-calendar-buffer\n  :date (cfw:date 1 1 2011)\n  :contents-sources\n   (list \n     (make-cfw:source\n      :name \"test1\" :data 'sample-data1)))<\/pre><\/div>\n<p>Evaluating this code in the scratch buffer, following result is displayed.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/a43a34a5e48d8cb5d445a67e9b75b00beaf4be43\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d35303331302e706e673f77696474683d343530\"><img src=\"https:\/\/camo.githubusercontent.com\/a43a34a5e48d8cb5d445a67e9b75b00beaf4be43\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d35303331302e706e673f77696474683d343530\" alt=\"Simple source example\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/P6baUrxEQj4NYheV-50310.png?width=450\" \/><\/a><\/p>\n<p>The date is specified by <code>cfw:date<\/code> type, <code>([month] [day] [year])<\/code>. This format is commonly used in calendar.el and orgmode.\n(I diagrammed the exchange ways for some time and date formats in Emacs, <a href=\"https:\/\/cacoo.com\/diagrams\/lsA64PTazlLTbSwR\" rel=\"nofollow\">here<\/a>.)<\/p>\n<p>Period items are little different. One period item is specified by\n<code>:start-date<\/code> and <code>:end-date<\/code>, and the nested list which has the symbol <code>periods<\/code> at the head collects them, like the following code.<\/p>\n<p><code>cfw:source-data example2:<\/code><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; cfw:source-data period items\n(defun sample-data2 (b e)\n  (list\n    (make-cfw:event :title \"Item1\"\n          :start-date  (cfw:date 1 15 2011))\n    (list 'periods\n      (make-cfw:event :title \"Period item\"\n          :start-date (cfw:date 1 8 2011)\n          :end-date   (cfw:date 1 9 2011)\n          :description \"Period item description\")\n      (make-cfw:event :title \"Next item\"\n          :start-date (cfw:date 1 11 2011)\n          :end-date   (cfw:date 1 12 2011)\n          :description \"Next item description\"))))\n  \n(cfw:open-calendar-buffer\n  :date (cfw:date 1 1 2011)\n  :contents-sources\n   (list \n     (make-cfw:source\n      :name \"test2\" :data 'sample-data2)))<\/pre><\/div>\n<p>Evaluating this code in the scratch buffer, following result is displayed.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/9df9ed9f86fd39689ca48b0151762a5ecb4b850e\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d34303331352e706e673f77696474683d343530\"><img src=\"https:\/\/camo.githubusercontent.com\/9df9ed9f86fd39689ca48b0151762a5ecb4b850e\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d34303331352e706e673f77696474683d343530\" alt=\"Range items example\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/P6baUrxEQj4NYheV-40315.png?width=450\" \/><\/a><\/p>\n<p>Here are other detailed specifications.<\/p>\n<ul><li>The both start and end date are included by the query period.<\/li>\n<li>The items those aren't included in the query period are ignored.<\/li>\n<li><code>cfw:source-data<\/code> should return a value as fast as possible, because users are waiting for the result. Caching is good idea.<\/li>\n<li>Schedule items don't have to be ordered. Duplicated items may be gathered.<\/li>\n<li>In the day cell, the items are sorted by <code>string-lessp<\/code>, i.e. numerical and alphabetical order.\n<ul><li>The ordering function can be customized by the named argument <code>:sorter<\/code> of the component construction.<\/li>\n<\/ul><\/li>\n<\/ul><p>In the above examples, the dates of the schedule items are fixed. The actual sources generate result values by the programs. The codes of calfw add-ons may be helpful for your implementation.<\/p>\n<h5><a id=\"user-content-cfwevent-struct-detail\" class=\"anchor\" aria-hidden=\"true\" href=\"#cfwevent-struct-detail\"><\/a>cfw:event struct detail<\/h5>\n<p>The <code>cfw:event<\/code> struct:<\/p>\n<table><thead><tr><th>slot name<\/th>\n<th>description<\/th>\n<\/tr><\/thead><tbody><tr><td><code>title<\/code><\/td>\n<td>event title [string]<\/td>\n<\/tr><tr><td><code>start-date<\/code><\/td>\n<td>start date of the event [cfw:date]<\/td>\n<\/tr><tr><td><code>start-time<\/code><\/td>\n<td>start time of the event (optional)<\/td>\n<\/tr><tr><td><code>end-date<\/code><\/td>\n<td>end date of the event [cfw:date] (optional)<\/td>\n<\/tr><tr><td><code>end-time<\/code><\/td>\n<td>end of the event (optional)<\/td>\n<\/tr><tr><td><code>description<\/code><\/td>\n<td>event description [string] (optional)<\/td>\n<\/tr><tr><td><code>location<\/code><\/td>\n<td>location [string] (optional)<\/td>\n<\/tr><tr><td><code>source<\/code><\/td>\n<td>[internal] source of the event<\/td>\n<\/tr><\/tbody><\/table><h5><a id=\"user-content-event-formatting\" class=\"anchor\" aria-hidden=\"true\" href=\"#event-formatting\"><\/a>Event formatting<\/h5>\n<p>The framework has several formatting functions for <code>cfw:event<\/code> instances.\nThe functions are used by the calfw plugins (cal,ical, etc) to display in a common way.<\/p>\n<table><thead><tr><th>Format function<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><code>cfw:event-overview<\/code><\/td>\n<td>To get an overview of the event (month, 2-week &amp; week view)<\/td>\n<\/tr><tr><td><code>cfw:event-days-overview<\/code><\/td>\n<td>Overview in day-view.<\/td>\n<\/tr><tr><td><code>cfw:event-period-overview<\/code><\/td>\n<td>Overview of periods (same for all views)<\/td>\n<\/tr><tr><td><code>cfw:event-detail<\/code><\/td>\n<td>Detailed information of the event for the detail-view<\/td>\n<\/tr><\/tbody><\/table><p>The formatting can be customized by the user with several formatting strings:<\/p>\n<ul><li><code>cfw:event-format-overview<\/code><\/li>\n<li><code>cfw:event-format-days-overview<\/code><\/li>\n<li><code>cfw:event-format-period-overview<\/code><\/li>\n<li><code>cfw:event-format-detail<\/code><\/li>\n<li><code>cfw:event-format-title<\/code><\/li>\n<li><code>cfw:event-format-start-date<\/code><\/li>\n<li><code>cfw:event-format-start-time<\/code><\/li>\n<li><code>cfw:event-format-end-date<\/code><\/li>\n<li><code>cfw:event-format-end-time<\/code><\/li>\n<li><code>cfw:event-format-location<\/code><\/li>\n<li><code>cfw:event-format-description<\/code><\/li>\n<\/ul><h4><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h4>\n<ul><li><a href=\"https:\/\/gist.github.com\/kiwanami\/d77d9669440f3336bb9d\">calfw-git.el<\/a>\n<ul><li>Displaying git commit history items in calfw calendar view<\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/gist.github.com\/kiwanami\/1fd257fc1e8907d4d92e\">calfw-syobocal.el<\/a>\n<ul><li>Retrieving schedule items via Web API and displaying them in calfw calendar view<\/li>\n<\/ul><\/li>\n<\/ul><h4><a id=\"user-content-another-way-to-define-schedule-items-legacy-method\" class=\"anchor\" aria-hidden=\"true\" href=\"#another-way-to-define-schedule-items-legacy-method\"><\/a>Another way to define schedule items (legacy method)<\/h4>\n<p><em>This subsection explains legacy method to define schedule items, so as for users to read old source codes. We should not use this method in the future.<\/em><\/p>\n<p>The function-slot <code>cfw:source-data<\/code> receives two arguments, start and\nend date of the query period, and returns an alist that consists of\n([date] . ([item1] [item2] ... )).<\/p>\n<p>Here is a simple example.<\/p>\n<p><code>cfw:source-data example1:<\/code><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; cfw:source-data example\n(defun sample-data1 (b e)\n  '(\n    ((1  1 2011) . (\"item1\"))\n    ((1 10 2011) . (\"item2-1\" \"item2-2\"))\n    ))\n(cfw:open-calendar-buffer\n  :date (cfw:date 1 1 2011)\n  :contents-sources\n   (list \n     (make-cfw:source\n      :name \"test1\" :data 'sample-data1)))<\/pre><\/div>\n<p>Period items are little different. One period item is specified by\n<code>([start date] [end date] [content])<\/code> and the <code>periods<\/code> record of the\nalist collects them as a list, like the following code.<\/p>\n<p><code>cfw:source-data example2:<\/code><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; cfw:source-data period items\n(defun sample-data2 (b e)\n  '(\n    ((1  8 2011) . (\"item1\"))\n     (periods\n      ((1 8 2011) (1 9 2011) \"period item\")\n      ((1 11 2011) (1 12 2011) \"next item\"))\n    ))\n;; (A . (B C) ) is equivalent to (A B C)\n(cfw:open-calendar-buffer\n  :date (cfw:date 1 1 2011)\n  :contents-sources\n   (list \n     (make-cfw:source\n      :name \"test2\" :data 'sample-data2)))<\/pre><\/div>\n<h3><a id=\"user-content-how-to-embed-the-calfw-component-in-the-other-applications\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-embed-the-calfw-component-in-the-other-applications\"><\/a>How to embed the calfw component in the other applications?<\/h3>\n<p>In this section, the details of calfw components would be explained so as for users to extend calfw in themselves.<\/p>\n<p>Calfw is built on the MVC architecture, using simple structure objects and modules employed by naming rules.<\/p>\n<h4><a id=\"user-content-calfw-component\" class=\"anchor\" aria-hidden=\"true\" href=\"#calfw-component\"><\/a>Calfw component<\/h4>\n<p>Calfw has three destination components to display the calendar.<\/p>\n<ul><li>Independent buffer<\/li>\n<li>Region in the other buffer<\/li>\n<li>Text output<\/li>\n<\/ul><h5><a id=\"user-content-buffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer\"><\/a>Buffer<\/h5>\n<p>The 'buffer' destination displays the calendar view as ordinary Emacs applications do.<\/p>\n<p>The function <code>cfw:open-calendar-buffer<\/code> makes a new calendar buffer (calfw buffer) and displays it by <code>switch-to-buffer<\/code>. The major mode of the calfw buffer is <code>cfw:calendar-mode<\/code> and the keymap <code>cfw:calendar-mode-map<\/code> is bound.<\/p>\n<p>This destination is easy to use for applications and users, because the buffer is usual application boundary and users know how to use buffers.<\/p>\n<h5><a id=\"user-content-region\" class=\"anchor\" aria-hidden=\"true\" href=\"#region\"><\/a>Region<\/h5>\n<p>The 'Region' destination embeds the calendar view in the buffer which is managed by the other applications. This destination can give the other applications a nice calendar view. See the howm embedding for example.<\/p>\n<p>Let's try a demonstration. Evaluate this code in your scratch buffer.<\/p>\n<p>Region destination example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Evaluate this code in the scratch buffer\n(require 'calfw)\n(cfw:create-calendar-component-region :height 10)<\/pre><\/div>\n<p>Then, the calendar view will be embedded in the scratch buffer like the following screenshot. You can navigate the calfw view in the buffer. Undoing for the some times, you can remove the calfw view.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/20c2837bcd3a81924de7027b8025130503047715\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d42393634392e706e673f77696474683d363030\"><img src=\"https:\/\/camo.githubusercontent.com\/20c2837bcd3a81924de7027b8025130503047715\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d42393634392e706e673f77696474683d363030\" alt=\"calfw in the scratch buffer\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/P6baUrxEQj4NYheV-B9649.png?width=600\" \/><\/a><\/p>\n<p>Because this destination never interacts anything out of the region and has its own key-binds as a text property, users can easily embed a calendar view in the other applications.<\/p>\n<h5><a id=\"user-content-text\" class=\"anchor\" aria-hidden=\"true\" href=\"#text\"><\/a>Text<\/h5>\n<p>The 'text' destination generates just a text which represent calfw view. The function <code>cfw:get-calendar-text<\/code> returns the text.<\/p>\n<h5><a id=\"user-content-destination-and-view\" class=\"anchor\" aria-hidden=\"true\" href=\"#destination-and-view\"><\/a>Destination and View<\/h5>\n<p>Three destinations are explained as mentioned above. Although they have different appearance, the application can operate the calfw component in the same way.<\/p>\n<p>Let us call them 'destination', it is the abstraction of UI components.<\/p>\n<p>The similar word 'view' means in which form the calfw displays the contents, for example, monthly form, two-weeks and weekly one and etc.<\/p>\n<h4><a id=\"user-content-calfw-objects\" class=\"anchor\" aria-hidden=\"true\" href=\"#calfw-objects\"><\/a>Calfw objects<\/h4>\n<h5><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h5>\n<p>The calfw consists of four objects:<\/p>\n<ul><li><code>cfw:component<\/code> that gathers following objects up.<\/li>\n<li><code>cfw:model<\/code> that manages calendar contents.<\/li>\n<li><code>cfw:source<\/code> that defines schedule items.<\/li>\n<li><code>cfw:dest<\/code> that is abstraction of destinations.<\/li>\n<\/ul><p>The relations between the objects are displayed as UML class diagram (<a href=\"http:\/\/astah.change-vision.com\/ja\/:title=Astah\" rel=\"nofollow\">Diagrammed by astah<\/a>).<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/031531dc1c2ca155b798f4c2a9aea120b78dea44\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d45433843362e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/031531dc1c2ca155b798f4c2a9aea120b78dea44\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d45433843362e706e67\" alt=\"Overview for calfw objects\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/P6baUrxEQj4NYheV-EC8C6.png\" \/><\/a><\/p>\n<p><code>cfw:component<\/code> acts as Controller of MVC. It connects model object and destination one, and controls all events. It also gives the interface of calfw objects for the other applications.<\/p>\n<p><code>cfw:model<\/code> and <code>cfw:source<\/code> act as Model of MVC. They manage the schedule contents and calendar logic.<\/p>\n<p><code>cfw:dest<\/code> acts as View of MVC. It abstracts the common interface from UI destinations.<\/p>\n<h5><a id=\"user-content-cfwcomponent\" class=\"anchor\" aria-hidden=\"true\" href=\"#cfwcomponent\"><\/a>cfw:component<\/h5>\n<p>The object <code>cfw:component<\/code> controls calfw objects and events.<\/p>\n<p>The object has following information:<\/p>\n<ul><li>References to <code>cfw:dest<\/code> object and <code>cfw:model<\/code> one.<\/li>\n<li>Selected date on the calfw component.<\/li>\n<li>View style.<\/li>\n<li>Hooks\n<ul><li><code>update-hooks<\/code><\/li>\n<li><code>selection-change-hooks<\/code><\/li>\n<li><code>click-hooks<\/code>.<\/li>\n<\/ul><\/li>\n<\/ul><p>The object has following operations:<\/p>\n<ul><li>Getting object references to <code>cfw:dest<\/code>, <code>cfw:model<\/code>, belonging buffer and so on.<\/li>\n<li>Getting and setting the selected date (<code>get-selected-date<\/code> \/ <code>set-selected-date<\/code>).<\/li>\n<li>Getting and setting the view style (<code>get-view<\/code> \/ <code>set-view<\/code>).\n<ul><li>The view style is a symbol, such as <code>month<\/code>, <code>two-weeks<\/code>, <code>week<\/code> and <code>day<\/code>.<\/li>\n<\/ul><\/li>\n<li>Resizing and refreshing the view (<code>resize<\/code> \/ <code>update<\/code>).<\/li>\n<li>Managing hooks (<code>add-xxx-hook<\/code> \/ <code>remove-xxx-hook<\/code>)<\/li>\n<\/ul><p>After construction of the calfw component, the destination object can not be changed.<\/p>\n<p>The views are defined as a function and dispatched by the function <code>cfw:cp-dispatch-view-impl<\/code>.<\/p>\n<p>The instance of the calfw component is stored at following places:<\/p>\n<ul><li><code>buffer<\/code> destination: the buffer-local variable <code>cfw:component<\/code><\/li>\n<li><code>region<\/code> destination: the text property <code>cfw:component<\/code><\/li>\n<li><code>text<\/code> destination: N\/A<\/li>\n<\/ul><p>Calling the utility function <code>cfw:cp-get-component<\/code>, one can obtain the calfw instance at the appropriate places. The stateless functions, such as simple event handler functions, can use this function to get the instance.<\/p>\n<p>The applications those have the state-full operations, however, should hold their own calfw instance for the safety object reference.<\/p>\n<h5><a id=\"user-content-cfwmodel\" class=\"anchor\" aria-hidden=\"true\" href=\"#cfwmodel\"><\/a>cfw:model<\/h5>\n<p>The object <code>cfw:model<\/code> gathers schedule sources and gives a common interface for view functions to access the contents.<\/p>\n<p>The object has following information:<\/p>\n<ul><li>contents source objects (<code>contents-sources<\/code>)<\/li>\n<li>annotation source objects (<code>annotation-sources<\/code>)<\/li>\n<li>sorting function (<code>sorter<\/code>)<\/li>\n<\/ul><p>The model object has no information of views and destinations, just manages schedule contents.<\/p>\n<p>The holidays are retrieved from the global function <code>calendar-holiday-list<\/code> of calendar.el.<\/p>\n<p>The schedule contents are modified through the model object after the component construction.<\/p>\n<p>(In the current implementation, the model object is build by alist. Then, view functions adds some data as view model. I think it is not good solution, so the implementation may be modified in future.)<\/p>\n<h5><a id=\"user-content-cfwdest\" class=\"anchor\" aria-hidden=\"true\" href=\"#cfwdest\"><\/a>cfw:dest<\/h5>\n<p>The object <code>cfw:dest<\/code> abstracts rendering destinations and gives a common interface of rendering operation to view functions.<\/p>\n<p>The object has following information:<\/p>\n<ul><li>destination buffer object (<code>buffer<\/code>)<\/li>\n<li>region functions (<code>min-func<\/code>, <code>max-func<\/code>)<\/li>\n<li>reference size (<code>width<\/code>, <code>height<\/code>)<\/li>\n<li>clearing function (<code>clear-func<\/code>)<\/li>\n<li>advice functions (<code>before-update-func<\/code>, <code>after-update-func<\/code>)<\/li>\n<li>overlay data (<code>select-ol<\/code>, <code>today-ol<\/code>)<\/li>\n<\/ul><p>In the current implementation, <code>cfw:dest<\/code> has three forms, buffer, region and text, mentioned above. Actually, the region destination is what I want. One buffer can have some destination objects, because all data (including local-variables and keymaps) are packed in the <code>cfw:dest<\/code> object.<\/p>\n<h4><a id=\"user-content-application-design\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-design\"><\/a>Application design<\/h4>\n<p>In this section, I would describe a simple guide line of application design using calfw.<\/p>\n<p>One can use calfw as an application UI (like calfw-howm) or dialog UI for selecting a date (like calendar.el). The user application can choose the destination style: buffer or region. Switching between them is very easy.<\/p>\n<p>The data presentation can be achieved by defining <code>cfw:source<\/code> object. It may be straightforward.<\/p>\n<p>The input events by the user can be caught by hooks in the <code>cfw:component<\/code>. Then, the selected date is obtained by the function <code>cfw:cursor-to-nearest-date<\/code> or <code>cfw:cursor-to-date<\/code>. The current implementation, calfw can not treat a range on the calendar.<\/p>\n<p>Generally, any events can be caught by the custom keymap which is given by the named argument <code>:custom-map<\/code> with component construction. Furthermore, because calfw reserves the text properties (face, keymap and so on) on the text that is returned by <code>cfw:source<\/code> objects, one can control event handling at each characters.<\/p>\n<p>Once the model is modified, update function of the <code>cfw:component<\/code> object should be called to refresh the view.<\/p>\n<p>The summary diagram is here.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/9edf73c222af3274d92fb051455eff7db7757e70\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d34363544342e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/9edf73c222af3274d92fb051455eff7db7757e70\/68747470733a2f2f6361636f6f2e636f6d2f6469616772616d732f5036626155727845516a344e596865562d34363544342e706e67\" alt=\"Summary of application design\" data-canonical-src=\"https:\/\/cacoo.com\/diagrams\/P6baUrxEQj4NYheV-465D4.png\" \/><\/a><\/p>\n<p>See the calfw-howm.el code for more details.<\/p>\n<h2><a id=\"user-content-history\" class=\"anchor\" aria-hidden=\"true\" href=\"#history\"><\/a>History<\/h2>\n<ul><li>2015\/09\/24 ver 1.5 : Fixed bugs and added some customize variables.<\/li>\n<li>2015\/02\/27 ver 1.4 : Introduced cfw:event struct, improved some functions, fixed some bugs.<\/li>\n<li>2011\/10\/10 ver 1.3 : Improved visual and navigation: multi-line, moving items in a day, diary mode and so on.<\/li>\n<li>2011\/07\/20 ver 1.2 : Merged many patches and improved many and bug fixed.<\/li>\n<li>2011\/07\/05 ver 1.0 : Refactored the whole implementation and design. Improved UI and views.<\/li>\n<li>2011\/01\/07 ver 0.2.1 : Supporting org-agenda schedules.<\/li>\n<li>2011\/01\/07 ver 0.1 : First release. Supporting howm and iCal schedules.<\/li>\n<\/ul><hr \/><p>SAKURAI, Masashi\nm.sakurai atmark kiwanami.net<\/p>\n<p>Time-stamp: &lt;2015-09-24 11:47:57 sakurai&gt;<\/p>","created_at":"2019-02-12T13:41:15+0000","updated_at":"2019-02-12T13:41:15+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/158801?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["rabbitmq","kafka"],"is_public":false,"id":831,"title":"Understanding When to use RabbitMQ or Apache Kafka","url":"https:\/\/content.pivotal.io\/blog\/understanding-when-to-use-rabbitmq-or-apache-kafka","content":"<p dir=\"ltr\">How do humans make decisions? In everyday life, emotion is often the circuit-breaking factor in pulling the trigger on a complex or overwhelming decision. \u00a0But for experts making complex decisions that have long term consequences, it can\u2019t be pure impulse. \u00a0High performers typically use the circuit breaker of \u201cinstinct\u201d, \u201cgut feel\u201d or other emotions only once their expert, unconscious mind has absorbed all the facts required to make a decision.<\/p><p dir=\"ltr\">Today there are dozens of messaging technologies, countless ESBs, and nearly 100 iPaaS vendors in market. \u00a0\u00a0Naturally, this leads to questions about how to choose the right messaging technology for your needs - particularly for those already invested in a particular choice. \u00a0Do we switch wholesale? \u00a0Just use the right tool for the right job? Have we correctly framed the job at hand for the business need? Given that, what is the right tool for me? \u00a0Worse, an exhaustive market analysis might never finish, but due diligence is critical given the average lifespan of integration code.<\/p><p dir=\"ltr\">This post endeavors give the unconscious, expert mind some even handed treatment to consider, starting with the most modern, popular choices today: RabbitMQ and Apache Kafka. \u00a0Each has it\u2019s own origin story, design intent, uses cases where it shines, integration capabilities and developer experience. Origins are revealing about the overall design intent for any piece of software, and make good starting point. \u00a0However it\u2019s important to note that in this article, my aim is to compare the two around the overlapping use case of message broker, less the \u201cevent store \/ event sourcing\u201d use case, where Kafka excels today.<\/p><p>Origins<\/p><p dir=\"ltr\"><a href=\"http:\/\/rabbitmq.com\">RabbitMQ<\/a> is a \u201ctraditional\u201d message broker that implements variety of messaging protocols. It was one of the first open source message brokers to achieve a reasonable level of features, client libraries, dev tools, and quality documentation. RabbitMQ was originally developed to implement <a href=\"http:\/\/www.amqp.org\/product\/overview\">AMQP<\/a>, an open wire protocol for messaging with powerful routing features. While Java has messaging standards like JMS, it\u2019s not helpful for non-Java applications that need distributed messaging which is severely limiting to any integration scenario, microservice or monolithic. With the advent of AMQP, cross-language flexibility became real for open source message brokers.<\/p><p dir=\"ltr\">Apache Kafka is developed in Scala and <a href=\"https:\/\/engineering.linkedin.com\/distributed-systems\/log-what-every-software-engineer-should-know-about-real-time-datas-unifying\">started out at LinkedIn<\/a> as a way to connect different internal systems. At the time, LinkedIn was moving to a more distributed architecture and needed to reimagine capabilities like data integration and realtime stream processing, breaking away from previously monolithic approaches to these problems. Kafka is well adopted today within the Apache Software Foundation ecosystem of products and is particularly useful in event-driven architecture.\u00a0<\/p><p dir=\"ltr\">Architecture and Design<\/p><p dir=\"ltr\">RabbitMQ is designed as a general purpose message broker, employing several variations of point to point, request\/reply and pub-sub communication styles patterns. \u00a0It uses a smart broker \/ dumb consumer model, focused on consistent delivery of messages to consumers that consume at a roughly similar pace as the broker keeps track of consumer state. \u00a0It is mature, performs well when configured correctly, is well supported (client libraries Java, .NET, node.js, Ruby, PHP and many more languages) and has dozens of plugins available that extend it to more use cases and integration scenarios.<\/p><p dir=\"ltr\"><img src=\"https:\/\/lh3.googleusercontent.com\/TmA6flkGzB1yc1xK6lGbJZ0YYqO__39trLIPxM62VUjsr09wClmbv9mT3WX4F0cuDssmkiHkekWR6AvXY0iVScuksmLxyM27FaJGYbgPezCIjRs-l8Ct3MfuUU3bRbpfWT6dhVBO\" alt=\"image\" \/><\/p><p dir=\"ltr\"><em>Figure 1 - Simplified overall RabbitMQ architecture. Source:\u00a0http:\/\/kth.diva-portal.org\/smash\/get\/diva2:813137\/FULLTEXT01.pdf\u00a0<\/em><\/p><p dir=\"ltr\">Communication in RabbitMQ can be either synchronous or asynchronous as needed. \u00a0Publishers send messages to exchanges, and consumers retrieve messages from queues. Decoupling producers from queues via exchanges ensures that producers aren't burdened with hardcoded routing decisions. \u00a0RabbitMQ also offers a number of distributed deployment scenarios (and does require all nodes be able to resolve hostnames). It can be setup for multi-node clusters to cluster federation and does not have dependencies on external services (but some cluster formation plugins can use AWS APIs, DNS, Consul, etcd). \u00a0<\/p><p dir=\"ltr\">Apache Kafka is designed for high volume publish-subscribe messages and streams, meant to be durable, fast, and scalable. At its essence, Kafka provides a durable message store, similar to a log, run in a server cluster, that stores streams of records in categories called topics.<\/p><p dir=\"ltr\"><img src=\"https:\/\/lh4.googleusercontent.com\/7b_8lfPpsZ0OPFuRWOVxcmJ1pZw0Bv8jHT-afdmZ8R31pkF0UmNetRtB989k8E7li7jAvL82sa_iB_uNO_scfWZjrAWkgTmSrjELcezMbva3IxqkQ2jrwPlw-S4h55HH_MHDxPUb\" alt=\"image\" \/><\/p><p dir=\"ltr\"><em>Figure 2 - Global Apache Kafka architecture (with 1 topic, 1 partition, replication factor 4). Source:\u00a0http:\/\/kth.diva-portal.org\/smash\/get\/diva2:813137\/FULLTEXT01.pdf\u00a0<\/em><\/p><p dir=\"ltr\">Every message consists of a key, a value, and a timestamp. \u00a0Nearly the opposite of RabbitMQ, Kafka employs a dumb broker and uses smart consumers to read its buffer. \u00a0Kafka does not attempt to track which messages were read by each consumer and only retain unread messages; rather, Kafka retains all messages for a set amount of time, and consumers are responsible to track their location in each log (consumer state). Consequently, with the right developer talent creating the consumer code, Kafka can support a large number of consumers and retain large amounts of data with very little overhead. \u00a0\u00a0As the diagram above shows, Kafka does require external services to run - in this case Apache Zookeeper, which is often regarded as non-trivial to understand, setup and operate.\u00a0 \u00a0\u00a0<\/p><h2 dir=\"ltr\">Requirements and Use Cases<\/h2><p dir=\"ltr\">Many developers begin exploring messaging when they realize they have to connect lots of things together, and other integration patterns such as shared databases are not feasible or too dangerous.<\/p><p dir=\"ltr\">Apache Kafka includes the broker itself, which is actually the best known and the most popular part of it, and has been designed and prominently marketed towards stream processing scenarios. In addition to that, Apache Kafka has recently added Kafka Streams which positions itself as an alternative to streaming platforms such as Apache Spark, Apache Flink, Apache Beam\/Google Cloud Data Flow and Spring Cloud Data Flow. The documentation does a good job of discussing popular <a href=\"https:\/\/kafka.apache.org\/uses\">use cases<\/a> like Website Activity Tracking, Metrics, Log Aggregation, Stream Processing, Event Sourcing and Commit logs. One of those use cases it describes is messaging, which can generate some confusion. \u00a0So let\u2019s unpack that a bit and get some clarity on which messaging scenarios are best for Kafka for, like:<\/p><ul dir=\"ltr\"><li>Stream from A to B without complex routing, with maximal throughput (100k\/sec+), delivered in partitioned order at least once.<\/li>\n\t<li>When your application needs access to stream history, delivered in partitioned order at least once. \u00a0Kafka is a durable message store and clients can get a \u201creplay\u201d of the event stream on demand, as opposed to more traditional message brokers where once a message has been delivered, it is removed from the queue.<\/li>\n\t<li>Stream Processing<\/li>\n\t<li>Event Sourcing<\/li>\n<\/ul><p dir=\"ltr\">RabbitMQ is a general purpose messaging solution, often used to allow web servers to respond to requests quickly instead of being forced to perform resource-heavy procedures while the user waits for the result. It\u2019s also good for distributing a message to multiple recipients for consumption or for balancing loads between workers under high load (20k+\/sec). \u00a0When your requirements extend beyond throughput, RabbitMQ has a lot to offer: <a href=\"http:\/\/www.rabbitmq.com\/confirms.html\">features for reliable delivery<\/a>, routing, federation, HA, security, management tools and <a href=\"http:\/\/www.rabbitmq.com\/features.html\">other features<\/a>. \u00a0Let\u2019s examine some scenarios best for RabbitMQ, like:<\/p><ul dir=\"ltr\"><li>Your application needs to work with any combination of existing protocols like AMQP 0-9-1, STOMP, MQTT, AMQP 1.0.<\/li>\n\t<li>You need a finer-grained consistency control\/guarantees on a per-message basis (dead letter queues, etc.) However, Kafka has recently added better support for <a href=\"https:\/\/medium.com\/@jaykreps\/exactly-once-support-in-apache-kafka-55e1fdd0a35f\">transactions<\/a>.\u00a0<\/li>\n\t<li>Your application needs variety in point to point, request \/ reply, and publish\/subscribe messaging<\/li>\n\t<li>Complex routing to consumers, integrate multiple services\/apps with non-trivial routing logic<\/li>\n<\/ul><p dir=\"ltr\">RabbitMQ can also effectively address several of Kafka\u2019s strong uses cases above, but with the help of additional software. RabbitMQ is often used with Apache Cassandra when application needs access to stream history, or with the LevelDB plugin for applications that need an \u201cinfinite\u201d queue, but neither feature ships with RabbitMQ itself.<\/p><p dir=\"ltr\">For a deeper dive on microservice - specific use cases with Kafka and RabbitMQ, head over to the Pivotal blog and <a href=\"https:\/\/content.pivotal.io\/blog\/messaging-patterns-for-event-driven-microservices\">read this short post by Fred Melo<\/a>.<\/p><h2 dir=\"ltr\">Developer Experience<\/h2><p dir=\"ltr\">RabbitMQ officially supports Java, Spring, .NET, PHP, Python, Ruby, JavaScript, Go, Elixir, Objective-C, Swift - with many other <a href=\"https:\/\/www.rabbitmq.com\/devtools.html\">clients and devtools<\/a> via community plugins. The RabbitMQ client libraries are mature and well documented.<\/p><p dir=\"ltr\">Apache Kafka has made strides in this area, and while it only ships a Java client, there is a growing catalog of community <a href=\"https:\/\/cwiki.apache.org\/confluence\/display\/KAFKA\/Clients\">open source clients<\/a>, <a href=\"https:\/\/cwiki.apache.org\/confluence\/display\/KAFKA\/Ecosystem\">ecosystem projects<\/a>, and well as an adapter SDK allowing you to build your own system integration. \u00a0Much of the configuration is done via .properties files or programmatically.<\/p><p dir=\"ltr\">The popularity of these two options has a strong influence on many other software providers who \u00a0make sure that RabbitMQ and Kafka work well with or on their technology.<\/p><p dir=\"ltr\">As for developer experience...it's worth mentioning\u00a0 the support that we provide in <a href=\"http:\/\/projects.spring.io\/spring-kafka\" target=\"_blank\">Spring Kafka<\/a>, <a href=\"https:\/\/cloud.spring.io\/spring-cloud-stream\/\" target=\"_blank\">Spring Cloud Stream<\/a>, etc.\u00a0<\/p><h2 dir=\"ltr\">Security and Operations<\/h2><p dir=\"ltr\">Both are strengths of RabbitMQ. \u00a0RabbitMQ management plugin provides an HTTP API, a browser-based UI for management and monitoring, plus CLI tools for operators. External tools like CollectD, Datadog, or New Relic are required for longer term monitoring data storage. RabbitMQ also provides API and tools for monitoring, audit and application troubleshooting. \u00a0Besides support for TLS, RabbitMQ ships with RBAC backed by a built-in data store, LDAP or external HTTPS-based providers and supports authentication using x509 certificate instead of username\/password pairs. Additional authentication methods can be fairly straightforwardly developed with plugins.<\/p><p dir=\"ltr\">These domains pose a challenge for Apache Kafka. \u00a0On the security front, the recent Kafka 0.9 release added TLS, JAAS role based access control and kerberos\/plain\/scram auth, using a CLI to manage security policy. \u00a0This made a substantial improvement on earlier versions where you could only lock down access at the network level, which didn\u2019t work well for sharing or multi-tenancy.<\/p><p dir=\"ltr\">Kafka uses a management CLI comprised of shell scripts, property files and specifically formatted JSON files. Kafka Brokers, Producers and Consumers emit metrics via Yammer\/JMX but do not maintain any history, which pragmatically means using a 3rd party monitoring system. \u00a0Using these tools, operations is able manage partitions and topics, check consumer offset position, and use the HA and FT capabilities that Apache Zookeeper provides for Kafka. While many view the requirement for Zookeeper with a high degree of skepticism, it does confer clustering benefits for Kafka users.<\/p><p dir=\"ltr\">For example, a 3-node Kafka cluster the system is functional even after 2 failures. \u00a0However if you want to support as many failures in Zookeeper you need an additional 5 Zookeeper nodes as Zookeeper is a quorum based system and can only tolerate N\/2+1 failures. These obviously should not be co-located with the Kafka nodes - so to stand up a 3 node Kafka system you need ~ 8 servers. \u00a0Operators must take the properties of the ZK cluster into account when reasoning about the availability of any Kafka system, both in terms of resource consumption and design.<\/p><h2 dir=\"ltr\">Performance<\/h2><p dir=\"ltr\">Kafka shines here by design: 100k\/sec performance is often a key driver for people choosing Apache Kafka. \u00a0<\/p><p dir=\"ltr\">Of course, message per second rates are tricky to state and quantify since they depend on so much including your environment and hardware, the nature of your workload, which delivery guarantees are used (e.g. persistent is costly, mirroring even more so), etc.<\/p><p dir=\"ltr\">20K messages per second is easy to push through a single Rabbit queue, indeed rather more than that isn't hard, with not much demanded in the way of guarantees. The queue is backed by a single Erlang lightweight thread that gets cooperatively scheduled on a pool of native OS threads - so it \u00a0becomes a natural choke point or bottleneck as a single queue is never going to do more work than it can get CPU cycles to work in.<\/p><p dir=\"ltr\">Increasing the messages per second often comes down to properly exploiting the parallelism available in one's environment by doing such things as breaking traffic across multiple queues via clever routing (so that different queues can be running concurrently). When RabbitMQ achieved <a href=\"https:\/\/content.pivotal.io\/blog\/rabbitmq-hits-one-million-messages-per-second-on-google-compute-engine\">1 million message per second <\/a>, this use case basically came down entirely to doing that judiciously - but was achieved using lot of resources, around 30 RabbitMQ nodes. \u00a0Most RabbitMQ users enjoy excellent performance with clusters made up of anywhere from three to seven RabbitMQ nodes.<\/p><h2 dir=\"ltr\">Making the call<\/h2><p dir=\"ltr\">Absorb some research on a few of the other top options on the market. If you want to go deeper with the most popular options, a <a href=\"http:\/\/kth.diva-portal.org\/smash\/get\/diva2:813137\/FULLTEXT01.pdf\">master\u2019s thesis from Nicolas Nannoni<\/a>\u00a0inspired this article and it features a side-by-side comparison table in Section 4.4 (page 39) but is a bit dated at this point.\u00a0If you feel like plunking down $15.00 USD, this ACM <a href=\"https:\/\/dl.acm.org\/citation.cfm?id=3093908\" target=\"_blank\">report<\/a> is also excellent and more recent.<\/p><p dir=\"ltr\">While researching, loop back with the stakeholders and the business as often as possible. Understanding the business use case is the single largest factor in making the right choice for your situation. \u00a0Then, <a href=\"https:\/\/www.amazon.com\/How-We-Decide-Jonah-Lehrer\/dp\/0547247990\">if you are pop psychology fan<\/a>, your best bet is sleep on it, let it percolate, and let your instincts take over. \u00a0You got this.<\/p><p dir=\"ltr\">\u00a0<\/p><hr \/><h2 dir=\"ltr\">Learn More<\/h2>","created_at":"2019-02-11T22:13:08+0000","updated_at":"2019-02-11T22:13:13+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":10,"domain_name":"content.pivotal.io","preview_picture":"https:\/\/content.cdntwrk.com\/files\/aHViPTYzOTc1JmNtZD1pdGVtZWRpdG9yaW1hZ2UmZmlsZW5hbWU9aXRlbWVkaXRvcmltYWdlXzU5MDBlMGNkOWEyN2EucG5nJnZlcnNpb249MDAwMCZzaWc9MDcyYTk0MzhlYWI5ZGMyZjkzZWIyZDEwMWI3NDFkYTU%253D","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["x11","xquartz"],"is_public":false,"id":830,"title":"Using a different X11 window manager with XQuartz - Tony Finch's blog","url":"https:\/\/fanf.livejournal.com\/142372.html","content":"<h3>(In particular, <a href=\"https:\/\/i3wm.org\" rel=\"nofollow\">i3<\/a> running on a remote machine!)<\/h3><p>I have been rebuilding my workstation, and taking the opportunity to\ndo some major housecleaning (including moving some old home directory\nstuff from CVS to git!)<\/p><p>Since 1999 I have used <a href=\"http:\/\/fvwm.org\" rel=\"nofollow\">fvwm<\/a> as my X11 window\nmanager. I have a weird configuration which makes it work a lot like a\ntiling window manager - I have no title bars, very thin window\nborders, keypresses to move windows to predefined locations. The main\nannoyance is that this configuration is not at all\nresolution-independent and a massive faff to redo for different screen\nlayouts - updating the predefined locations and the corresponding\nkeypresses.<\/p><p>I have heard good things about <a href=\"https:\/\/i3wm.org\" rel=\"nofollow\">i3<\/a> so I thought I\nwould try it. Out of the box it works a lot like my fvwm\nconfiguration, so I decided to make the switch. It was probably about\nthe same faff to configure i3 the way I wanted (40 workspaces) but\nfuture updates should be much easier!<\/p><h3>XQuartz and quartz-wm<\/h3><p>I did some of this configuration at home after work, using\n<a href=\"http:\/\/www.xquartz.org\" rel=\"nofollow\">XQuartz<\/a> on my MacBook as the X11 server.\nXQuartz comes with its own Mac-like window manager called <code>quartz-wm<\/code>.<\/p><p>You can't just switch window managers by starting another one - X only\nallows one window manager at a time, and other window managers will\nrefuse to start if one is already running. So you have to configure\nyour X session if you want to use a different window manager.<\/p><h3>X session startup<\/h3><p>Traditionally, X stashes a lot of configuration stuff in\n<code>\/usr\/X11R6\/lib\/X11<\/code>. I use <code>xdm<\/code> which has a subdirectory of ugly\nsession management scripts; there is also an <code>xinit<\/code> subdirectory for\nsimpler setups. Debian sensibly moves a lot of this gubbins into\n<code>\/etc\/X11<\/code>; XQuartz puts them in <code>\/opt\/X11\/lib\/X11<\/code>.<\/p><p>As well as their locations, the contents of the session scripts vary\nfrom one version of X11 to another. So if you want to configure your\nsession, be prepared to read some shell scripts. Debian has sensibly\nunified them into a shared <code>Xsession<\/code> script which even has a man page!<\/p><h3>XQuartz startup<\/h3><p>XQuartz does not use a display manager; it uses <code>startx<\/code>, so the\nrelevant script is <code>\/opt\/X11\/lib\/X11\/xinit\/xinitrc<\/code>. This has a nice\n<code>run-parts<\/code> style directory, inside which is the script we care about,\n<code>\/opt\/X11\/lib\/X11\/xinit\/xinitrc.d\/98-user.s<wbr \/>h<\/code>. This in turn invokes\nscripts in a per-user <code>run-parts<\/code> directory, <code>~\/.xinitrc.d<\/code>.<\/p><h3>Choose your own window manager<\/h3><p>So, what you do is,<\/p><pre>\n    <b>$<\/b> mkdir .xinitrc.d\n    <b>$<\/b> cat &gt;.xinitrc.d\/99-wm.sh\n    #!\/bin\/sh\n    exec twm\n    <b>^D<\/b>\n    <b>$<\/b> chmod +x .xinitrc.d\/99-wm.sh\n    <b>$<\/b> open \/Applications\/Utilities\/XQuartz.app\n<\/pre><p>(The <code>.sh<\/code> and the <code>chmod<\/code> are necessary.)<\/p><p>This should cause an <code>xterm<\/code> to appear with <code>twm<\/code> decoration instead\nof <code>quartz-wm<\/code> decoration.<\/p><h3>My supid trick<\/h3><p>Of course, X is a network protocol, so (like any other X application)\nyou don't have to run the window manager on the same machine as the X\nserver. My <code>99-wm.sh<\/code> was roughly,<\/p><pre>\n    #!\/bin\/sh\n    exec ssh -AY <i>workstation<\/i> i3\n<\/pre><p>And with XQuartz configured to run fullscreen this was good enough to\nhave a serious hack at <code>.i3\/config<\/code> :-)<\/p>","created_at":"2019-02-11T14:01:10+0000","updated_at":"2019-02-11T14:01:19+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"fanf.livejournal.com","preview_picture":"https:\/\/l-stat.livejournal.net\/img\/sign.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":829,"title":"r\/math - A monad is a monoid in the category of endofunctors, what's the problem?","url":"https:\/\/www.reddit.com\/r\/math\/comments\/ap25mr\/a_monad_is_a_monoid_in_the_category_of\/","content":"<p class=\"s90z9tc-10 fHRkcP\">This phrase appears in the hilarious <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"http:\/\/james-iry.blogspot.com\/2009\/05\/brief-incomplete-and-mostly-wrong.html\"><em class=\"s90z9tc-15 kRPwUl\">A Brief, Incomplete, and Mostly Wrong History of Programming Languages<\/em><\/a>, making fun of the obscure mathematical concepts in the functional programming language Haskell.<\/p><p class=\"s90z9tc-10 fHRkcP\">I've studied some category theory and I've programmed in Haskell, and I've seen this statement many times before. But I've never actually taken the time to break it down and see why it's true. I just spent a couple of hours refreshing my rusty category theory knowledge and figuring it out. I wanted to write it down to make sure I actually understood it well. So... here goes.<\/p><p class=\"s90z9tc-10 fHRkcP\">A [category](<a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/en.wikipedia.org\/wiki\/Category_(mathematics)\">https:\/\/en.wikipedia.org\/wiki\/Category_(mathematics)<\/a>) is simply a collection of 'objects' (or 'points') with 'morphisms' (or 'arrows') between them, satisfying two very simple rules:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">You can compose an arrow <code class=\"s90z9tc-7 klSgyG\">f: A -&gt; B<\/code> with an arrow <code class=\"s90z9tc-7 klSgyG\">g: B -&gt; C<\/code> to get a new arrow <code class=\"s90z9tc-7 klSgyG\">g . f: A -&gt; C<\/code>, and this composition is associative (i.e. <code class=\"s90z9tc-7 klSgyG\">h . (g . f) = (h . g) . f<\/code>).<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">For every object <code class=\"s90z9tc-7 klSgyG\">A<\/code> there exists an <em class=\"s90z9tc-15 kRPwUl\">identity arrow<\/em> <code class=\"s90z9tc-7 klSgyG\">id_A: A -&gt; A<\/code>, such that for every arrow <code class=\"s90z9tc-7 klSgyG\">f: A -&gt; B<\/code> we have <code class=\"s90z9tc-7 klSgyG\">id_B . f = f . id_A = f<\/code>.<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">The classical example in mathematics is the category <code class=\"s90z9tc-7 klSgyG\">Set<\/code>, whose objects are sets, and whose arrows are functions between these sets. In the world of Haskell, we have the category <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/wiki.haskell.org\/Hask\"><code class=\"s90z9tc-7 klSgyG\">Hask<\/code><\/a>, whose objects are Haskell types and whose arrows are functions between these types. So, for example, <code class=\"s90z9tc-7 klSgyG\">Float<\/code> and <code class=\"s90z9tc-7 klSgyG\">Int<\/code> are objects, and <code class=\"s90z9tc-7 klSgyG\">round:: Float -&gt; Int<\/code> is an arrow.<\/p><p class=\"s90z9tc-10 fHRkcP\">In category theory a <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/en.wikipedia.org\/wiki\/Functor\">functor<\/a> is a map between two categories. So if <code class=\"s90z9tc-7 klSgyG\">C<\/code> and <code class=\"s90z9tc-7 klSgyG\">D<\/code> are categories, then a functor <code class=\"s90z9tc-7 klSgyG\">F: C -&gt; D<\/code> will map objects in <code class=\"s90z9tc-7 klSgyG\">C<\/code> to objects in <code class=\"s90z9tc-7 klSgyG\">D<\/code>, and arrows in <code class=\"s90z9tc-7 klSgyG\">C<\/code> to arrows in <code class=\"s90z9tc-7 klSgyG\">D<\/code>. It does this in a 'nice' way. That is:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">The starts and ends of arrows are mapped nicely: if <code class=\"s90z9tc-7 klSgyG\">f<\/code> is an arrow in <code class=\"s90z9tc-7 klSgyG\">C<\/code> from object <code class=\"s90z9tc-7 klSgyG\">A<\/code> to <code class=\"s90z9tc-7 klSgyG\">B<\/code>, then <code class=\"s90z9tc-7 klSgyG\">F(f)<\/code> is an arrow in <code class=\"s90z9tc-7 klSgyG\">D<\/code> from <code class=\"s90z9tc-7 klSgyG\">F(A)<\/code> to <code class=\"s90z9tc-7 klSgyG\">F(B)<\/code>.<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Identities are preserved: <code class=\"s90z9tc-7 klSgyG\">F(id_A) = id_F(A)<\/code>.<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Composition is preserved: <code class=\"s90z9tc-7 klSgyG\">F(g . f) = F(g) . F(f)<\/code>.<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">Note that this concept is a bit more general than the concept of functors in Haskell (see below).<\/p><h2 class=\"s90z9tc-1 hCXhWq\">What's an endofunctor?<\/h2><p class=\"s90z9tc-10 fHRkcP\">An endofunctor is simply a functor from a category to itself. So in the above, assume that <code class=\"s90z9tc-7 klSgyG\">C = D<\/code>. Note, that doesn't mean that the endofunctor <code class=\"s90z9tc-7 klSgyG\">F<\/code> doesn't <em class=\"s90z9tc-15 kRPwUl\">do<\/em> anything. Just like a function from the real numbers to the real numbers might still <em class=\"s90z9tc-15 kRPwUl\">change<\/em> the numbers, the functor <code class=\"s90z9tc-7 klSgyG\">F<\/code> might still change the objects and arrows that are fed through it in some way.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">What's a Haskell Functor?<\/h2><p class=\"s90z9tc-10 fHRkcP\">What is known as a <code class=\"s90z9tc-7 klSgyG\">Functor<\/code> in Haskell is actually an endofunctor on the category <code class=\"s90z9tc-7 klSgyG\">Hask<\/code>. Recall that <code class=\"s90z9tc-7 klSgyG\">Hask<\/code> has types as its objects and functions between these types as its arrows. So an endofunctor on <code class=\"s90z9tc-7 klSgyG\">Hask<\/code> will map a type to some other type, and will also map functions to functions in some nice way.<\/p><p class=\"s90z9tc-10 fHRkcP\">This is the <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"http:\/\/hackage.haskell.org\/package\/base-4.12.0.0\/docs\/src\/GHC.Base.html#Functor\">definition of <code class=\"s90z9tc-7 klSgyG\">Functor<\/code> in Haskell<\/a>:<\/p><pre class=\"s90z9tc-7 klSgyG\">{- | The 'Functor' class is used for types that can be mapped over.\nInstances of 'Functor' should satisfy the following laws:\n&gt; fmap id  ==  id\n&gt; fmap (f . g)  ==  fmap f . fmap g\nThe instances of 'Functor' for lists, 'Data.Maybe.Maybe' and 'System.IO.IO'\nsatisfy these laws.\n-}\nclass  Functor f  where\n    fmap        :: (a -&gt; b) -&gt; f a -&gt; f b\n<\/pre><p class=\"s90z9tc-10 fHRkcP\">See the symmetry with the mathematical definition of functors above? The type constructor <code class=\"s90z9tc-7 klSgyG\">f<\/code> fulfills the role of the functor's action on types (objects), while <code class=\"s90z9tc-7 klSgyG\">fmap<\/code> fulfills the role of the functor's action on functions (arrows).<\/p><p class=\"s90z9tc-10 fHRkcP\">The two classical examples of <code class=\"s90z9tc-7 klSgyG\">Functor<\/code>s in Haskell are lists (<code class=\"s90z9tc-7 klSgyG\">[]<\/code>) and <code class=\"s90z9tc-7 klSgyG\">Maybe<\/code>.<\/p><h3 class=\"s90z9tc-2 bQaRpt\">List is a Functor<\/h3><p class=\"s90z9tc-10 fHRkcP\">List is a type constructor that, given some type <code class=\"s90z9tc-7 klSgyG\">a<\/code> will give you a new type <code class=\"s90z9tc-7 klSgyG\">[a]<\/code>, namely lists consisting of values of type <code class=\"s90z9tc-7 klSgyG\">a<\/code>. This is the type (object) mapping part. The function mapping part is <code class=\"s90z9tc-7 klSgyG\">fmap<\/code> (also known simply as <code class=\"s90z9tc-7 klSgyG\">map<\/code>). Given some function <code class=\"s90z9tc-7 klSgyG\">f:: a -&gt; b<\/code> it will give you a function <code class=\"s90z9tc-7 klSgyG\">fmap f:: [a] -&gt; [b]<\/code>, namely the function that applies <code class=\"s90z9tc-7 klSgyG\">f<\/code> to every element of a list of values of type <code class=\"s90z9tc-7 klSgyG\">a<\/code>. You can see that <code class=\"s90z9tc-7 klSgyG\">fmap id<\/code> is indeed the identity: doing nothing to every element of a list is the same as doing nothing to the list. The other law, <code class=\"s90z9tc-7 klSgyG\">fmap (f . g) = fmap f . fmap g<\/code>, is also easy to see: doing <code class=\"s90z9tc-7 klSgyG\">g<\/code> and then <code class=\"s90z9tc-7 klSgyG\">f<\/code> to every element of a list, is the same as first doing <code class=\"s90z9tc-7 klSgyG\">g<\/code> to every element of a list and then doing <code class=\"s90z9tc-7 klSgyG\">f<\/code> to every element of the resulting list.<\/p><h3 class=\"s90z9tc-2 bQaRpt\">Maybe is a Functor<\/h3><p class=\"s90z9tc-10 fHRkcP\"><code class=\"s90z9tc-7 klSgyG\">Maybe<\/code> is a type constructor that given some type <code class=\"s90z9tc-7 klSgyG\">a<\/code> will give you a new type <code class=\"s90z9tc-7 klSgyG\">Maybe a<\/code>. The values of this type are <code class=\"s90z9tc-7 klSgyG\">Just x<\/code> for any <code class=\"s90z9tc-7 klSgyG\">x<\/code> of type <code class=\"s90z9tc-7 klSgyG\">a<\/code>, and <code class=\"s90z9tc-7 klSgyG\">Nothing<\/code>. The function mapping part <code class=\"s90z9tc-7 klSgyG\">fmap<\/code> will take some function <code class=\"s90z9tc-7 klSgyG\">f:: a -&gt; b<\/code> and give you a new function <code class=\"s90z9tc-7 klSgyG\">fmap f:: Maybe a -&gt; Maybe b<\/code>. It will take <code class=\"s90z9tc-7 klSgyG\">Just x<\/code> to <code class=\"s90z9tc-7 klSgyG\">Just (f x)<\/code>, and <code class=\"s90z9tc-7 klSgyG\">Nothing<\/code> to <code class=\"s90z9tc-7 klSgyG\">Nothing<\/code>. Can you prove the laws?<\/p><p class=\"s90z9tc-10 fHRkcP\">Ok, now things are going to get a bit more tricky. First we need to talk about natural transformations, which are basically arrows between functors. Then we'll use this to build a category of endofunctors, and finally we'll look at examples in Haskell.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">What are natural transformations?<\/h2><p class=\"s90z9tc-10 fHRkcP\">It turns out that if you have categories <code class=\"s90z9tc-7 klSgyG\">C<\/code> and <code class=\"s90z9tc-7 klSgyG\">D<\/code> and functors <code class=\"s90z9tc-7 klSgyG\">F, G: C -&gt; D<\/code> between them, you can sometimes find so called '<a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/en.wikipedia.org\/wiki\/Natural_transformation\">natural transformations<\/a>' between the functors <code class=\"s90z9tc-7 klSgyG\">F<\/code> and <code class=\"s90z9tc-7 klSgyG\">G<\/code>. A natural transformation <code class=\"s90z9tc-7 klSgyG\">t: F =&gt; G<\/code> is a family of arrows in <code class=\"s90z9tc-7 klSgyG\">D<\/code> that satisfies two requirements:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">The arrows <code class=\"s90z9tc-7 klSgyG\">t<\/code> go from the results of <code class=\"s90z9tc-7 klSgyG\">F<\/code> to the results of <code class=\"s90z9tc-7 klSgyG\">G<\/code>. More precisly, for every object <code class=\"s90z9tc-7 klSgyG\">A<\/code> in <code class=\"s90z9tc-7 klSgyG\">C<\/code> we have an arrow <code class=\"s90z9tc-7 klSgyG\">t_A: F(A) -&gt; G(A)<\/code> in <code class=\"s90z9tc-7 klSgyG\">D<\/code>. This is the 'component' of <code class=\"s90z9tc-7 klSgyG\">t<\/code> at <code class=\"s90z9tc-7 klSgyG\">A<\/code>.<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">For every arrow <code class=\"s90z9tc-7 klSgyG\">f<\/code> in <code class=\"s90z9tc-7 klSgyG\">C<\/code>, applying <code class=\"s90z9tc-7 klSgyG\">F(f)<\/code> first and then <code class=\"s90z9tc-7 klSgyG\">t<\/code> is the same as applying <code class=\"s90z9tc-7 klSgyG\">t<\/code> first and then <code class=\"s90z9tc-7 klSgyG\">G(f)<\/code>. This is the 'natural' part of a 'natural transformation'. More precisely, for every <code class=\"s90z9tc-7 klSgyG\">f: X -&gt; Y<\/code> in <code class=\"s90z9tc-7 klSgyG\">C<\/code> we have <code class=\"s90z9tc-7 klSgyG\">t_Y . F(f) = G(f) . t_X<\/code>. Graphically, it means that going around <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f2\/Natural_transformation.svg\/1267px-Natural_transformation.svg.png\">this diagram<\/a> in either direction does the same thing ('the diagram commutes').<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">Of course there's a special case where <code class=\"s90z9tc-7 klSgyG\">F<\/code> and <code class=\"s90z9tc-7 klSgyG\">G<\/code> are endofunctors (so <code class=\"s90z9tc-7 klSgyG\">C = D<\/code>). There's also nothing stopping us from setting <code class=\"s90z9tc-7 klSgyG\">F = G<\/code>, so then we're looking for natural transformations from a functor to itself. Just as with endofunctors these natural transformations may still <em class=\"s90z9tc-15 kRPwUl\">do something<\/em>.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">The category of endofunctors<\/h2><p class=\"s90z9tc-10 fHRkcP\">Now, for some category <code class=\"s90z9tc-7 klSgyG\">C<\/code> we have a bunch of endofunctors on <code class=\"s90z9tc-7 klSgyG\">C<\/code> and we have natural transformations between these endofunctors. We can make this into a category! So let's introduce <code class=\"s90z9tc-7 klSgyG\">Endo(C)<\/code>, the category whose objects are endofunctors on <code class=\"s90z9tc-7 klSgyG\">C<\/code>, and whose arrows are natural transformations between these endofunctors. You can check that composition of natural transformations is indeed associative, and that there is an identity natural transformation from every endofunctor to itself. But that's not super relevant here.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">The category of endofunctors in Haskell<\/h2><p class=\"s90z9tc-10 fHRkcP\">Lost? Let's think about what this would look like in the world of Haskell: what does <code class=\"s90z9tc-7 klSgyG\">Endo(Hask)<\/code> look like? Well, its objects are endofunctors on <code class=\"s90z9tc-7 klSgyG\">Hask<\/code>, which are simply <code class=\"s90z9tc-7 klSgyG\">Functor<\/code>s. Its arrows are 'transformations' from one <code class=\"s90z9tc-7 klSgyG\">Functor<\/code> to another. So if <code class=\"s90z9tc-7 klSgyG\">f<\/code> and <code class=\"s90z9tc-7 klSgyG\">g<\/code> are functors, then we're looking for some set of functions <code class=\"s90z9tc-7 klSgyG\">t<\/code>, such that for every type <code class=\"s90z9tc-7 klSgyG\">a<\/code> we have a function <code class=\"s90z9tc-7 klSgyG\">t:: f a -&gt; g a<\/code>.<\/p><p class=\"s90z9tc-10 fHRkcP\">Let's pick <code class=\"s90z9tc-7 klSgyG\">f = Maybe<\/code> and <code class=\"s90z9tc-7 klSgyG\">g = []<\/code>. So we're looking for a set of functions <code class=\"s90z9tc-7 klSgyG\">t:: Maybe a -&gt; [a]<\/code>. Well here's one example:<\/p><pre class=\"s90z9tc-7 klSgyG\">maybeToList            :: Maybe a -&gt; [a]\nmaybeToList  Nothing   = []\nmaybeToList  (Just x)  = [x]\n<\/pre><p class=\"s90z9tc-10 fHRkcP\">Ok, so <code class=\"s90z9tc-7 klSgyG\">maybeToList<\/code> is a transformation from the endofunctor <code class=\"s90z9tc-7 klSgyG\">Maybe<\/code> to the endofunctor <code class=\"s90z9tc-7 klSgyG\">[]<\/code>. But is it natural? Well, let's take some arbitrary function <code class=\"s90z9tc-7 klSgyG\">f: a-&gt; b<\/code>. If <code class=\"s90z9tc-7 klSgyG\">maybeToList<\/code> is natural, it must satisfy <code class=\"s90z9tc-7 klSgyG\">maybeToList . fmap f = map f . maybeToList<\/code> (note that I've filled in <code class=\"s90z9tc-7 klSgyG\">map<\/code> for the list's <code class=\"s90z9tc-7 klSgyG\">fmap<\/code> to avoid confusion). Well this is pretty easy to check:<\/p><pre class=\"s90z9tc-7 klSgyG\">(maybeToList . fmap f) Just x = maybeToList (Just (f x)) = [f x]\n(map f . maybeToList ) Just x = map f [x]                = [f x]\n(maybeToList . fmap f) Nothing = maybeToList Nothing = []\n(map f . maybeToList ) Nothing = map f []            = []\n<\/pre><p class=\"s90z9tc-10 fHRkcP\">So yes, they're the same! So <code class=\"s90z9tc-7 klSgyG\">maybeToList<\/code> is a natural transformation from the endofunctor <code class=\"s90z9tc-7 klSgyG\">Maybe<\/code> to the endofunctor <code class=\"s90z9tc-7 klSgyG\">[]<\/code>.<\/p><p class=\"s90z9tc-10 fHRkcP\">Another interesting natural transformation is <code class=\"s90z9tc-7 klSgyG\">concat:: [[a]] -&gt; [a]<\/code>. It is a natural transformation from <code class=\"s90z9tc-7 klSgyG\">[[]]<\/code> (i.e. applying the endofunctor <code class=\"s90z9tc-7 klSgyG\">[]<\/code> twice), to <code class=\"s90z9tc-7 klSgyG\">[]<\/code>. The naturality condition is <code class=\"s90z9tc-7 klSgyG\">concat . map (map f) = map f . concat<\/code>.<\/p><p class=\"s90z9tc-10 fHRkcP\">Ok, this one is gonna be tricky again. There's a few different layers to this:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">First we'll look at the classical concept of a monoid in set theory. We'll also look at an example of such a monoid in Haskell.<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Then we'll try to generalize the concept of monoids from set theory to category theory. We'll realize that we're missing a category-theoretic ingredient. Confusingly, this ingredient is called a 'monoidal category'.<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Armed with monoidal categories, we can understand a monoid in a category.<\/p><\/li><\/ul><h2 class=\"s90z9tc-1 hCXhWq\">Monoids in set theory<\/h2><p class=\"s90z9tc-10 fHRkcP\">In classical set theory, a <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/en.wikipedia.org\/wiki\/Monoid\">monoid<\/a> is a set <code class=\"s90z9tc-7 klSgyG\">M<\/code> with some binary operation <code class=\"s90z9tc-7 klSgyG\">\u2022: M \u00d7 M -&gt; M<\/code> satisfying the following properties:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Associativity: for all <code class=\"s90z9tc-7 klSgyG\">a, b, c<\/code> in <code class=\"s90z9tc-7 klSgyG\">M<\/code>, we have <code class=\"s90z9tc-7 klSgyG\">(a \u2022 b) \u2022 c = a \u2022 (b \u2022 c)<\/code>. Informally: \"it doesn't matter where you put the brackets.\"<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Identity element: there is some <code class=\"s90z9tc-7 klSgyG\">e<\/code> in <code class=\"s90z9tc-7 klSgyG\">M<\/code> such that for every <code class=\"s90z9tc-7 klSgyG\">a<\/code> in <code class=\"s90z9tc-7 klSgyG\">M<\/code> we have <code class=\"s90z9tc-7 klSgyG\">e \u2022 a = a \u2022 e = a<\/code>.<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">An example of a monoid is the set of all finite strings over some alphabet with string concatenation as the monoid operation. It's easy to check that associativity holds and that the empty string acts as an identity element.<\/p><p class=\"s90z9tc-10 fHRkcP\">This also gives us our example in Haskell: for every type <code class=\"s90z9tc-7 klSgyG\">a<\/code> the type <code class=\"s90z9tc-7 klSgyG\">[a]<\/code> is a monoid with <code class=\"s90z9tc-7 klSgyG\">++<\/code> (concatenation) as the monoid operation and the empty list as identity element. (Small note: technically the type <code class=\"s90z9tc-7 klSgyG\">[a]<\/code> also includes infinite lists, for which concatenation isn't well defined. Because of this, lists in Haskell technically aren't monoids, nor are they monads. We will ignore this technicality in what follows.)<\/p><h2 class=\"s90z9tc-1 hCXhWq\">Monoids in category theory<\/h2><p class=\"s90z9tc-10 fHRkcP\">Category theorists are alergic to sets. If they see a definition of a mathematical object that includes the word 'set', they'll immediately start thinking about how they can rewrite this definition without using the word 'set'. Or, more respectfully, they'll try to generalize the definition such that it holds in categories other than <code class=\"s90z9tc-7 klSgyG\">Set<\/code> too.<\/p><p class=\"s90z9tc-10 fHRkcP\">It's obvious how to begin: the set <code class=\"s90z9tc-7 klSgyG\">M<\/code> should be replaced by an arbitrary object <code class=\"s90z9tc-7 klSgyG\">M<\/code> in some category <code class=\"s90z9tc-7 klSgyG\">C<\/code>. The binary operation <code class=\"s90z9tc-7 klSgyG\">\u2022<\/code> should be replaced by some arrow <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code> in the category. Similarly, since we can't 'crack open' an object in a category (we'd have to say what it is, and category theorists hate that), we have to replace the identity element with some arrow <code class=\"s90z9tc-7 klSgyG\">\u03b7<\/code> going into <code class=\"s90z9tc-7 klSgyG\">M<\/code>.<\/p><p class=\"s90z9tc-10 fHRkcP\">The arrow <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code> should clearly end at <code class=\"s90z9tc-7 klSgyG\">M<\/code>, but where should it start? We need some way to construct a 'product object', similar to the cartesian product for sets. Turns out there's a few different ways you can do that, but the one that's useful for monoids is the concept of a 'tensor product' in a 'monoidal category'.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">Monoidal categories<\/h2><p class=\"s90z9tc-10 fHRkcP\">The aim here is to define some operation <code class=\"s90z9tc-7 klSgyG\">\u2297<\/code> that will allow us to combine two objects <code class=\"s90z9tc-7 klSgyG\">A, B<\/code> in some category <code class=\"s90z9tc-7 klSgyG\">C<\/code> into a new object <code class=\"s90z9tc-7 klSgyG\">A \u2297 B<\/code> in that same category. The natural way to map objects in categories is through functors. But in this case we would have to map two objects into one object. We need a functor that can take two arguments.<\/p><p class=\"s90z9tc-10 fHRkcP\">The typical way of solving this is by instead introducing a product space that encodes both arguments as one. So we need to have an object <code class=\"s90z9tc-7 klSgyG\">(A, B)<\/code> in some product category <code class=\"s90z9tc-7 klSgyG\">C \u00d7 C<\/code>. At this point you may think I'm going crazy: in order to define monoids in categories we need to define products of objects, but to do that we need to define monoidal categories, but in order to that we need to define products of whole categories?!<\/p><p class=\"s90z9tc-10 fHRkcP\">That's right. Luckily, it's really quite simple. For categories <code class=\"s90z9tc-7 klSgyG\">C<\/code> and <code class=\"s90z9tc-7 klSgyG\">D<\/code> the product category <code class=\"s90z9tc-7 klSgyG\">C \u00d7 D<\/code> is simply the category whose objects are pairs <code class=\"s90z9tc-7 klSgyG\">(A, B)<\/code> of objects <code class=\"s90z9tc-7 klSgyG\">A<\/code> in <code class=\"s90z9tc-7 klSgyG\">C<\/code> and <code class=\"s90z9tc-7 klSgyG\">B<\/code> in <code class=\"s90z9tc-7 klSgyG\">D<\/code>, and whose arrows are pairs <code class=\"s90z9tc-7 klSgyG\">f, g: (A, B) -&gt; (X, Y)<\/code> of arrows <code class=\"s90z9tc-7 klSgyG\">f: A -&gt; X<\/code> and <code class=\"s90z9tc-7 klSgyG\">g: B -&gt; Y<\/code> in <code class=\"s90z9tc-7 klSgyG\">C<\/code> and <code class=\"s90z9tc-7 klSgyG\">D<\/code>, respectively. Composition works straight-forwardly: <code class=\"s90z9tc-7 klSgyG\">(f, g) . (k, l) = (f . k, g . l)<\/code>, with associativity easy to check. For any object <code class=\"s90z9tc-7 klSgyG\">(A, B)<\/code> the identity arrow is simply given by <code class=\"s90z9tc-7 klSgyG\">(id_A, id_B)<\/code>.<\/p><p class=\"s90z9tc-10 fHRkcP\">Ok, so now that we defined product categories, the machinery of functors is available to us again, and we can start thinking about a functor <code class=\"s90z9tc-7 klSgyG\">\u2297: C \u00d7 C -&gt; C<\/code> that combines objects (and arrows). What properties would we want this functor to posses? Well, we would certainly want it to be associative in some way: we want <code class=\"s90z9tc-7 klSgyG\">A \u2297 (B \u2297 C)<\/code> to be 'similar' to <code class=\"s90z9tc-7 klSgyG\">(A \u2297 B) \u2297 C<\/code>. This can be made precise by saying that there is a natural transformation whose components are isomorphisms <code class=\"s90z9tc-7 klSgyG\">\u03b1: A \u2297 (B \u2297 C) -&gt; (A \u2297 B) \u2297 C<\/code>. Similarly, we'd like there to be some identity element for <code class=\"s90z9tc-7 klSgyG\">\u2297<\/code>, let's call it <code class=\"s90z9tc-7 klSgyG\">I<\/code>: we want <code class=\"s90z9tc-7 klSgyG\">I \u2297 A<\/code> and <code class=\"s90z9tc-7 klSgyG\">A \u2297 I<\/code> to be 'similar' to <code class=\"s90z9tc-7 klSgyG\">A<\/code>. This can again be made precise by saying that there are natural transformations whose components are isomorphisms <code class=\"s90z9tc-7 klSgyG\">\u03bb: I \u2297 A -&gt; A<\/code> and <code class=\"s90z9tc-7 klSgyG\">\u03c1: A \u2297 I -&gt; A<\/code>. There's some additional conditions about these natural transformations playing nice with each other (the 'coherence conditions'), but we'll skip over that here.<\/p><p class=\"s90z9tc-10 fHRkcP\">If you compare this section to the one about monoids in set theory, you can probably spot some similarities. Instead of elements of a set, we have objects in a category; instead of a binary operation <code class=\"s90z9tc-7 klSgyG\">\u2022<\/code> we have the functor <code class=\"s90z9tc-7 klSgyG\">\u2297<\/code>; instead of an identity element we have a unit object; and instead of equations involving elements of a set, we have natural transformations between functors built using <code class=\"s90z9tc-7 klSgyG\">\u2297<\/code>. This correspondence is why the term 'monoid' is reused, and why this concept is called a <a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/en.wikipedia.org\/wiki\/Monoidal_category\">monoidal category<\/a>.<\/p><p class=\"s90z9tc-10 fHRkcP\">If we want to make the monoidal structure of a category <code class=\"s90z9tc-7 klSgyG\">C<\/code> explicit, we will refer to it as the monoidal category <code class=\"s90z9tc-7 klSgyG\">(C, \u2297, I)<\/code><\/p><h2 class=\"s90z9tc-1 hCXhWq\">Monoids in a monoidal category<\/h2><p class=\"s90z9tc-10 fHRkcP\">Above we defined a concept analogous to monoids. However, it's not quite what we're looking for; a monoidal category is a category with some additional properties. What we're looking for is an <em class=\"s90z9tc-15 kRPwUl\">object in a category<\/em> satisfing certain properties. Namely, a [monoid in a monoidal category](<a target=\"_blank\" class=\"s90z9tc-27 fvqGYz\" href=\"https:\/\/en.wikipedia.org\/wiki\/Monoid_(category_theory)\">https:\/\/en.wikipedia.org\/wiki\/Monoid_(category_theory)<\/a>).<\/p><p class=\"s90z9tc-10 fHRkcP\">We start with the definition of a monoid, and then swap out set-theoretic concepts for category-theoretic ones. This is what we get:<\/p><p class=\"s90z9tc-10 fHRkcP\">A monoid in a monoidal category <code class=\"s90z9tc-7 klSgyG\">(C, \u2297, I)<\/code> is an object <code class=\"s90z9tc-7 klSgyG\">M<\/code> with an arrow <code class=\"s90z9tc-7 klSgyG\">\u03bc: M \u2297 M -&gt; M<\/code> ('multiplication') and an arrow <code class=\"s90z9tc-7 klSgyG\">\u03b7: I -&gt; M<\/code> ('unit') satisfying the two monoid axioms:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Associativity: if you have an object <code class=\"s90z9tc-7 klSgyG\">M \u2297 (M \u2297 M) \u2245 (M \u2297 M) \u2297<\/code>, it doesn't matter if you first multiply the left side or first multiply the right side. This is made exact with a commutative diagram involving <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code> and the natural isomorphism <code class=\"s90z9tc-7 klSgyG\">\u03b1<\/code> (the associator).<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Unit is an identity with multiplication: if you have an object <code class=\"s90z9tc-7 klSgyG\">I \u2297 M \u2245 M<\/code>, then 'forgetting' the <code class=\"s90z9tc-7 klSgyG\">I<\/code> using the left unitor <code class=\"s90z9tc-7 klSgyG\">\u03bb<\/code> is the same as first mapping to <code class=\"s90z9tc-7 klSgyG\">M \u2297 M<\/code> using <code class=\"s90z9tc-7 klSgyG\">\u03b7<\/code> and then to <code class=\"s90z9tc-7 klSgyG\">M<\/code> using <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code>. A symmetric equation must hold for <code class=\"s90z9tc-7 klSgyG\">M \u2297 I<\/code> using the right unitor <code class=\"s90z9tc-7 klSgyG\">\u03c1<\/code>.<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">Ok, now that we know what a monoid in a (monoidal) category is, we can start thinking about a monoid in the category of endofunctors.<\/p><p class=\"s90z9tc-10 fHRkcP\">As we found out above, we first need to find a monoidal structure in the category of endofunctors. Turns out that's pretty easy: composition of endofunctors satisfies all the properties of a tensor product! The tensor product identity <code class=\"s90z9tc-7 klSgyG\">I<\/code> is simply the identity functor <code class=\"s90z9tc-7 klSgyG\">id<\/code>. It takes a little bit of work to show that the natural transformations behave the right way, but we'll skip over that here.<\/p><p class=\"s90z9tc-10 fHRkcP\">So then a monoid in the category of endofunctors is some endofunctor <code class=\"s90z9tc-7 klSgyG\">M<\/code>, a natural transformation <code class=\"s90z9tc-7 klSgyG\">\u03bc: M . M -&gt; M<\/code> ('multiplication'), and a natural transformation <code class=\"s90z9tc-7 klSgyG\">\u03b7: id -&gt; M<\/code> ('unit'), that satisfy the monoid axioms.<\/p><p class=\"s90z9tc-10 fHRkcP\">How can we make sense of this definition? Well, here's one way that will carry over quite nicely to Haskell: we can think of our endofunctors as introducing some additional structure on the objects they're applied to. The action of the endofunctors on arrows is to 'lift' the arrow into the structure. The endofunctor <code class=\"s90z9tc-7 klSgyG\">M<\/code> then, is simply some chosen structure. The transformation <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code> ('multiplication') can be seen as 'flattening' the duplicate structure <code class=\"s90z9tc-7 klSgyG\">M . M<\/code> down to just the structure <code class=\"s90z9tc-7 klSgyG\">M<\/code>. The transformation <code class=\"s90z9tc-7 klSgyG\">\u03b7<\/code> can be thought of as injecting unstructured data into the structure <code class=\"s90z9tc-7 klSgyG\">M<\/code>.<\/p><p class=\"s90z9tc-10 fHRkcP\">What about the monoid axioms?<\/p><h2 class=\"s90z9tc-1 hCXhWq\">Associativity<\/h2><p class=\"s90z9tc-10 fHRkcP\">If you have a triplicate structure <code class=\"s90z9tc-7 klSgyG\">M . M . M<\/code>, it doesn't matter if you flatten it down to <code class=\"s90z9tc-7 klSgyG\">M<\/code> as <code class=\"s90z9tc-7 klSgyG\">(M . M) . M -&gt; M . M -&gt; M<\/code> or as <code class=\"s90z9tc-7 klSgyG\">M . (M . M) -&gt; M . M -&gt; M<\/code>. This can be symbolically stated as for every <code class=\"s90z9tc-7 klSgyG\">X<\/code>, <code class=\"s90z9tc-7 klSgyG\">\u03bc_X . M(\u03bc_X) = \u03bc_X . \u03bc_M(X)<\/code>. Or intuitively: there's a unique way to flatten layered structure.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">Unit is identity with multiplication<\/h2><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">If you have strucured data (<code class=\"s90z9tc-7 klSgyG\">M<\/code>), and then use <code class=\"s90z9tc-7 klSgyG\">\u03b7<\/code> to inject this into <code class=\"s90z9tc-7 klSgyG\">M<\/code> again, you obtain data in <code class=\"s90z9tc-7 klSgyG\">M . M<\/code>. You can then use <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code> to flatten this back down to <code class=\"s90z9tc-7 klSgyG\">M<\/code>. This is the same as doing nothing. Symbolically, for every <code class=\"s90z9tc-7 klSgyG\">X<\/code>, <code class=\"s90z9tc-7 klSgyG\">\u03bc_X . \u03b7_M(X) = id_M(X)<\/code><\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">If you have structured data (<code class=\"s90z9tc-7 klSgyG\">M<\/code>), and then use <code class=\"s90z9tc-7 klSgyG\">M(\u03b7)<\/code> give structure to the 'contents' of your structure, you again obtain data in <code class=\"s90z9tc-7 klSgyG\">M . M<\/code>. You can again use <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code> to flatten this back down to <code class=\"s90z9tc-7 klSgyG\">M<\/code>. This is, again, the same as doing nothing. Symbolically, for every <code class=\"s90z9tc-7 klSgyG\">X<\/code>, <code class=\"s90z9tc-7 klSgyG\">\u03bc_X . M(\u03b7_X) = id_M(X)<\/code>.<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">Intuitively: injecting structure and flattening it cancel each other out.<\/p><h2 class=\"s90z9tc-1 hCXhWq\">A monoid in Endo(Hask)<\/h2><p class=\"s90z9tc-10 fHRkcP\">Let's move to the world of Haskell again. As before, we look at <code class=\"s90z9tc-7 klSgyG\">Endo(Hask)<\/code>, with its familiar endofunctors of <code class=\"s90z9tc-7 klSgyG\">[]<\/code> and <code class=\"s90z9tc-7 klSgyG\">Maybe<\/code> and others. Let's choose <code class=\"s90z9tc-7 klSgyG\">M = []<\/code>. What should we pick for <code class=\"s90z9tc-7 klSgyG\">\u03bc<\/code>? It should be a natural transformation <code class=\"s90z9tc-7 klSgyG\">\u03bc:: [[a]] -&gt; [a]<\/code>. We've seen one before: <code class=\"s90z9tc-7 klSgyG\">concat<\/code> (also known as <code class=\"s90z9tc-7 klSgyG\">join<\/code>)! How about <code class=\"s90z9tc-7 klSgyG\">\u03b7<\/code>? It should be a natural transformation <code class=\"s90z9tc-7 klSgyG\">\u03b7:: id(a) -&gt; [a]<\/code>, which is simply <code class=\"s90z9tc-7 klSgyG\">\u03b7:: a -&gt; [a]<\/code>. Well it's pretty simple to come up with one:<\/p><pre class=\"s90z9tc-7 klSgyG\">inject   :: a -&gt; [a]\ninject x = [x]\n<\/pre><p class=\"s90z9tc-10 fHRkcP\">This function is also known as <code class=\"s90z9tc-7 klSgyG\">pure<\/code> or <code class=\"s90z9tc-7 klSgyG\">return<\/code> in Haskell.<\/p><p class=\"s90z9tc-10 fHRkcP\">What about the monoid axioms?<\/p><p class=\"s90z9tc-10 fHRkcP\">Well, for associativity we have to check that if we have some 3d list, that flattening the outer dimensions first will give the same result as flattening the inner dimensions. So: <code class=\"s90z9tc-7 klSgyG\">concat . concat = concat . (map concat)<\/code>. It shouldn't be too hard to convince yourself that this is true.<\/p><p class=\"s90z9tc-10 fHRkcP\">But does our choice of unit play nice with the multiplication? We need to check:<\/p><ul class=\"s90z9tc-12 drWKKm\"><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Injecting a list into a list and then concatenating is the same as doing nothing to that list: <code class=\"s90z9tc-7 klSgyG\">concat . inject = id :: [a] -&gt; [a]<\/code>. That sounds pretty reasonable.<\/p><\/li><li class=\"s90z9tc-11 irlrft\"><p class=\"s90z9tc-10 fHRkcP\">Injecting every element of a list into a list and then concatenating is the same as doing nothing to that list: <code class=\"s90z9tc-7 klSgyG\">concat . map inject = id :: [a] -&gt; [a]<\/code>. Again, that seems clear.<\/p><\/li><\/ul><p class=\"s90z9tc-10 fHRkcP\">So <code class=\"s90z9tc-7 klSgyG\">[]<\/code> is a monoid in the <code class=\"s90z9tc-7 klSgyG\">Endo(Hask)<\/code>. Or equivalently, <code class=\"s90z9tc-7 klSgyG\">[]<\/code> is a monad in <code class=\"s90z9tc-7 klSgyG\">Hask<\/code>. Hooray!<\/p>","created_at":"2019-02-11T03:21:43+0000","updated_at":"2019-02-11T03:21:43+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"www.reddit.com","preview_picture":"https:\/\/external-preview.redd.it\/NA720cYN9m6m3cCNiC30R_SLCtsDGAPJewwFgs_cLI0.png?auto=webp&s=6ba0e565b3dc7d0b66be963eefd6a90a8c578e74","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":828,"title":"No title found","url":"https:\/\/stumpwm.github.io\/","content":"<p>\nStumpWM is a tiling, keyboard driven X11 Window Manager written\nentirely in Common Lisp.\n<\/p><p>\nIf you're tired of flipping through themes like channel-surfing, and\ngoing from one perfect-except-for-just-one-thing window manager to\nanother even-more-broken-in-some-other-way then perhaps StumpWM can\nhelp.\n<\/p><p>\nStumpWM attempts to be customizable yet visually minimal. There are no\nwindow decorations, no icons, and no buttons. It does have various\nhooks to attach your personal customizations, and variables to tweak.\n<\/p><ul class=\"org-ul\"><li>Hack the good hack\n<\/li>\n<li>Debug your good hack\n<\/li>\n<li>Customize your window manager\n<\/li>\n<\/ul><p>\n<b>While it's running<\/b> That's right. With a 100% Common Lisp window\nmanager there's no stopping the hacks. Just re-eval and GO!\n<\/p>","created_at":"2019-02-10T20:32:27+0000","updated_at":"2019-02-10T20:32:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"stumpwm.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":827,"title":"ch11ng\/exwm","url":"https:\/\/github.com\/ch11ng\/exwm","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>EXWM (Emacs X Window Manager) is a full-featured tiling X window manager\nfor Emacs built on top of <a href=\"https:\/\/github.com\/ch11ng\/xelb\">XELB<\/a>.\nIt features:<\/p>\n<ul><li>Fully keyboard-driven operations<\/li>\n<li>Hybrid layout modes (tiling &amp; stacking)<\/li>\n<li>Dynamic workspace support<\/li>\n<li>ICCCM\/EWMH compliance<\/li>\n<li>(Optional) RandR (multi-monitor) support<\/li>\n<li>(Optional) Builtin system tray<\/li>\n<li>(Optional) Builtin input method<\/li>\n<\/ul><p>Please check out the\n<a href=\"https:\/\/github.com\/ch11ng\/exwm\/wiki\/Screenshots\">screenshots<\/a>\nto get an overview of what EXWM is capable of,\nand the <a href=\"https:\/\/github.com\/ch11ng\/exwm\/wiki\">user guide<\/a>\nfor a detailed explanation of its usage.<\/p>\n<p><strong>Note<\/strong>: If you install EXWM from source, it's recommended to install\nXELB also from source (otherwise install both from GNU ELPA).<\/p>\n<\/article>","created_at":"2019-02-10T19:40:38+0000","updated_at":"2019-02-10T19:40:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/13297330?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":826,"title":"poignant.guide","url":"http:\/\/poignant.guide\/","created_at":"2019-02-10T19:40:37+0000","updated_at":"2019-02-10T19:40:37+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"poignant.guide"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":825,"title":"Make your site\u2019s pages instant in 1 minute","url":"https:\/\/instant.page\/","content":"<p>Amazon and others found that 100 milliseconds of latency is responsible for 1% in sales.  But latency on the web is hard to overcome.<\/p><div class=\"footnote-content\"><p><a href=\"https:\/\/web.archive.org\/web\/20081117195303if_\/http:\/\/home.blarg.net\/~glinden\/StanfordDataMining.2006-11-29.ppt\">Amazon (PowerPoint, slide #15)<\/a>: 100\u00a0ms of latency resulted in 1% less sales.<\/p><p><a href=\"https:\/\/youtu.be\/6x0cAzQ7PVs?t=936\">Google (video)<\/a>: 500\u00a0ms caused a 20% drop in traffic.<\/p><p><a href=\"https:\/\/www.slideshare.net\/devonauerswald\/walmart-pagespeedslide\">Walmart (slide #46)<\/a>: a 100ms improvement brought up to 1% incremental revenue<\/p><p><a href=\"https:\/\/blog.mozilla.org\/metrics\/2010\/04\/05\/firefox-page-load-speed-%E2%80%93-part-ii\/\">Mozilla<\/a>: Shaving 2.2 seconds off page load time increased downloads by 15.4%<\/p><p><a href=\"https:\/\/www.slideshare.net\/stubbornella\/designing-fast-websites-presentation\/23-1_Create_a_component_library\">Yahoo<\/a>: 400\u00a0ms resulted in a 5 to 9% drop in traffic<\/p><\/div>\n<h2>Cheating latency<\/h2>\n<p>instant.page uses <em>just-in-time preloading<\/em> \u2014 it preloads a page right before a user clicks on it.<\/p>\n<p>Before a user clicks on a link, they hover their mouse over that link. When a user has hovered for 65\u00a0ms there is one chance out of two that they will click on that link, so instant.page starts preloading at this moment, leaving on average over 300\u00a0ms for the page to preload.<\/p>\n<p>On mobile, a user starts touching their display before releasing it, leaving on average 90\u00a0ms to preload the page.<\/p>\n<p>You can also click the menu at the bottom on the right to experience it.<\/p>\n<h2>Cheating the brain<\/h2>\n<p>The humain brain perceives actions taking less than 100\u00a0ms as instant.  As a result, instant.page makes your site feel instant even on 3G (assuming your pages are fast to render).<\/p>\n<div class=\"footnote-content\"><p>Jakob Nielsen: <a href=\"https:\/\/www.nngroup.com\/articles\/response-times-3-important-limits\/\">Response Times: The 3 Important Limits<\/a>:<\/p><blockquote class=\"nielsen-quote\">\n      <div><strong>0.1 second<\/strong> is about the limit for having the user feel that the system is <strong>reacting instantaneously<\/strong>, meaning that no special feedback is necessary except to display the result.\n  <\/div><\/blockquote><\/div>\n<h2>Easy on your server and your user\u2019s data plan<\/h2>\n<p>Pages are preloaded only when there\u2019s a good chance that a user is going to visit them, and it preloads only the HTML of that page, being respectful of your users\u2019 and servers\u2019 bandwidth and CPU. It\u2019s 1\u00a0kB and loads after everything else. And it\u2019s free and open source (MIT license).<\/p>","created_at":"2019-02-09T23:03:47+0000","updated_at":"2019-02-09T23:03:47+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"instant.page","preview_picture":"https:\/\/instant.page\/twitter_summary_image_v2.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":824,"title":"Build me a LISP \u2014 kirit.com","url":"https:\/\/kirit.com\/Build%20me%20a%20LISP","content":"<p>Normally programming language blog posts get started with grammars and syntax and parsing and all sorts of stuff like that. I can't be bothered with all of that, so instead, let's start with the (maybe more interesting aspect) of the actual evaluation of the code. Here we'll build an evaluation engine for LISP like languages in less than 50 lines of C++ \u2014 <a class=\"External\" href=\"https:\/\/gist.github.com\/KayEss\/45a2e88675832228f57e2d598afc02ae\">and literate C++ at that!<\/a><\/p><p>Lisp uses something called a \"cons cell\" to store each part of both the program and the data. Let's just say we have something called a cell for now, and we'll come back to defining what it is later on.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>struct cell;\n<\/pre><\/div>\n<\/td><\/tr><\/table><h2>S-Expressions<\/h2><p>A Lisp program is actually just something called an s-expression. It's a fancy term for a list of cells, but there is a special way to interpret these s-expressions. We'll come back to that later on.<\/p><p>For now, we just need to know that s-expressions describe our LISP program, and to run a LISP program we simply evaluate the s-expression.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>#include &lt;vector&gt;\nusing sexpr = std::vector&lt;cell&gt;;\n<\/pre><\/div>\n<\/td><\/tr><\/table><h2>Cells<\/h2><p>Now we can talk about what a cell is. Our simple LISP will only have a single data type, strings, so our cells will either contain a string or an s-expression.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>4\n5\n6\n7<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>#include &lt;string&gt;\n#include &lt;variant&gt;\nstruct cell {\n    std::variant&lt;std::string, sexpr&gt; content;\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>We do need to be able to make cells, and they can be made from either strings or s-expressions.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>8\n9\n10<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>    cell(std::string s) : content{std::move(s)} {}\n    cell(sexpr s) : content{std::move(s)} {}\n};\n<\/pre><\/div>\n<\/td><\/tr><\/table><h2>Evaluation<\/h2><p>There are three different sorts of data structure that we have so far:<\/p><ol><li>Strings<\/li><li>S-expressions<\/li><li>Cells<\/li><\/ol><p>Because running our programs is going to be a matter of evaluating these structures, we'll need to consider what \"evaluation\" means for each of these.<\/p><p>First of all, strings are just themselves.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>std::string eval(std::string s) { return s; }\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>The s-expressions are a bit more interesting, but for now we just have to promise to implement them so that we can start with the simpler cells (and because we have to implement the evaluation of the s-expressions in terms of evaluating cells)\u2026<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>std::string eval(sexpr const &amp;);\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>Evaluating a cell is pretty simple. It's either going to be a string or a s-expression, so depending on what we find in the cell content we can just forward to evaluating whichever one we find.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>13\n14\n15<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>std::string eval(cell const &amp;c) {\n    return std::visit([](auto const &amp;c) { return eval(c); }, c.content);\n}\n<\/pre><\/div>\n<\/td><\/tr><\/table><h2>The standard library<\/h2><p>Although an s-expression is just a vector of cells, the first item in it is special. To evaluate an s-expression we take the first item in it to be the name of a function that is to be executed, and the remaining s-expression cells are then taken as that function's arguments (some of which may be s-expressions that require evaluation themselves).<\/p><p>So if you're used to writing your programs like this:<\/p><div class=\"highlight\"><pre>foo(bar, baz)\n<\/pre><\/div><p>The s-expression form would look more like:<\/p><div class=\"highlight\"><pre>(foo bar baz)\n<\/pre><\/div><p>We're going to implement our library in C++, and we're going to look functions up by name. Because the s-expression can have any number of arguments, it makes sense to evaluate them based on a couple of iterators (<code>marker<\/code>s), one to the first argument and one to the end of the s-expression.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>using marker = sexpr::const_iterator;\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>Our built in library functions take the two iterators representing their arguments and return a string which is the result of running them.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>#include &lt;functional&gt;\nusing builtin = std::function&lt;auto(marker, marker)-&gt;std::string&gt;;\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>We now need a simple data structure that will hold these for us and allow look-ups by name.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>#include &lt;map&gt;\nstd::map&lt;std::string, builtin&gt; library;\n<\/pre><\/div>\n<\/td><\/tr><\/table><h2>Evaluating S-Expressions<\/h2><p>The first item in an s-expression is a function name that we're going to call. This means that an empty s-expression is an error.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>21\n22\n23\n24<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>#include &lt;stdexcept&gt;\nstd::string eval(sexpr const &amp;s) {\n    if (s.begin() == s.end()) {\n        throw std::domain_error(\"Empty sexpr\");\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>An alternative design would be to say that an empty s-expression evaluates to an empty string, but this is good for us.<\/p><p>Now, assuming our s-expression contains a first item, we need to fetch the function name. Remember that our s-expression is comprised of cells (which can contain strings <i>or<\/i> s-expressions) so we need to turn our cell into a string. Luckily we already have a function that does that: <code>eval<\/code>.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>    } else {\n        auto const fname = eval(*s.begin());\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>Using <code>eval<\/code> here (rather than just fetching the string out of the cell directly) has some important consequences to what this language can do.<\/p><p>Although we've said so far that the first item in an s-expression has to be the function name, by evaluating the cell to fetch that string we can actually have an s-expression as the first cell! And evaluating that s-expression gives the function name that we will run. Don't worry about this for now, there will be an example at the end.<\/p><p>Now that we have a function name we only need to look it up and call it, and of course take into account the possibility that it might not be found.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>27\n28\n29\n30\n31\n32\n33<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>        if (auto const fp = library.find(fname); fp != library.end()) {\n            return fp-&gt;second(++s.begin(), s.end());\n        } else {\n            throw std::invalid_argument{fname + \" not in the library\"};\n        }\n    }\n}\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>And that's actually it for the whole evaluation engine. Luckily we have some lines left over to try it out with.<\/p><h2>Running a program<\/h2><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>34\n35\n36<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>#include &lt;iostream&gt;\nint main() {\n    try {\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>As with all programming languages we have to make sure to say hello to the whole world.<\/p><div class=\"highlight\"><pre>(cat \"Hello\" \"world!\")\n<\/pre><\/div><p>Because we didn't bother with a parser we have to write the data structure for the program directly into the C++ source code. Thankfully it's not too hard.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>        using namespace std::string_literals;\n        auto const prog = sexpr{\"cat\"s, \"Hello \"s, \"world!\"s};\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>Of course we didn't put anything into our library yet, so let's make sure that we do at least have an implementation of <code>cat<\/code> we can call.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>39\n40\n41\n42\n43<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>        library[\"cat\"] = [](marker pos, marker end) {\n            std::string ret;\n            while (pos != end) ret += eval(*pos++);\n            return ret;\n        };\n<\/pre><\/div>\n<\/td><\/tr><\/table><p>Now we can run our program and show the result.<\/p><table class=\"highlighttable\"><tr><td><\/td><td class=\"code\"><div class=\"highlight\"><pre>        std::cout &lt;&lt; eval(prog) &lt;&lt; std::endl;\n<\/pre><\/div>\n<\/td><\/tr><\/table><p><a class=\"External\" href=\"https:\/\/gist.github.com\/KayEss\/45a2e88675832228f57e2d598afc02ae\">If you grab this C++ file you can compile and run it using something like (the file is available as a Gist on Github)<\/a>:<\/p><div class=\"highlight\"><pre>clang++ lisp.cpp -std=c++17 -o lisp\n.\/lisp\n<\/pre><\/div><p>And you'll see some output:<\/p><div class=\"highlight\"><pre>Hello world!\n<\/pre><\/div><h2>Errors\u2026<\/h2><p>The very last thing we should do is to print out some sort of error message if a run-time error came up.<\/p><table class=\"highlighttable\"><tr><td><div class=\"linenodiv\"><pre>45\n46\n47\n48\n49<\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>    } catch (std::exception &amp;e) {\n        std::cerr &lt;&lt; \"Error: \" &lt;&lt; e.what() &lt;&lt; std::endl;\n        return 1;\n    }\n}\n<\/pre><\/div>\n<\/td><\/tr><\/table><h2>More fun<\/h2><p>Remember there was the bit about the evaluation of the lead term in the s-expression? Because the first term can itself be an s-expression you can write a program to generate the function name you want to call.<\/p><p>If you do try this on your machine, try to change the program to the following and see what happens.<\/p><div class=\"highlight\"><pre>auto const prog = sexpr{\n    sexpr{\"cat\"s, \"c\"s, \"a\"s, \"t\"s},\n        \"Hello \"s, \"world!\"s};\n<\/pre><\/div><p>When you run this it will still print \"Hello world!\" because the s-expression as the first term gets evaluated and produces the name <code>cat<\/code> which then gets called to produce the output.<\/p><h3>Order fun<\/h3><p>Remember the type we had for <code>builtin<\/code> took <code>marker<\/code>s to the s-expressions of the arguments? This means that the argument values are evaluated inside of the library function definition. Look again at the loop in our <code>cat<\/code> function:<\/p><div class=\"highlight\"><pre>while (pos != end) ret += eval(*pos++);\n<\/pre><\/div><p>It would also be a reasonable choice to move that evaluation outside of the library function and instead have our <code>builtin<\/code> type look like this (a function that takes a vector of strings and produces a string):<\/p><div class=\"highlight\"><pre>using builtin = std::function&lt;auto(std::vector&lt;std::string&gt;)-&gt;std::string&gt;;\n<\/pre><\/div><p>This difference in where the expressions are evaluated is called <a class=\"External\" href=\"https:\/\/en.wikipedia.org\/wiki\/Evaluation_strategy\"><i>Evaluation\nStrategy<\/i><\/a> and the choice we made above is called <i>normal-order<\/i>. This is of course the opposite of what you're used to from pretty much every other programming language, which normally use <i>applicative-order<\/i>.<\/p><p>More commonly, normal-order is referred to as <a class=\"External\" href=\"https:\/\/en.wikipedia.org\/wiki\/Lazy_evaluation\"><i>lazy\nevaluation<\/i><\/a> and applicative-order as <a class=\"External\" href=\"https:\/\/en.wikipedia.org\/wiki\/Eager_evaluation\"><i>eager\nevaluation<\/i><\/a>.<\/p><h3><code>if<\/code> isn't (an eager) function<\/h3><p>In LISP terms our <code>builtin<\/code>s are all <a class=\"External\" href=\"http:\/\/www.lispworks.com\/documentation\/HyperSpec\/Body\/03_ababa.htm\"><i>special\nforms<\/i><\/a>. This is important in order to be able to correctly implement <i>forms<\/i> like <code>if<\/code>.<\/p><p>Let's say we want to implement <code>if<\/code> so that it can be used something like this:<\/p><div class=\"highlight\"><pre>(if (condition) (true-case) (false-case)) -&gt; string\n<\/pre><\/div><p>We will always need to evaluate the <code>condition<\/code> expression, and, depending on what happens, we are for sure going to return either the <code>true-case<\/code> or the <code>false-case<\/code>. But more importantly we only want to <b>evaluate<\/b> one of them.<\/p><p>If we have to implement <code>if<\/code> in a language that only has <i>eager evaluation<\/i> then we will evaluate both the <code>true-case<\/code> and the <code>false-case<\/code> even if we only return the result of one of them.<\/p><p>But with <i>lazy evaluation<\/i> we can save ourselves some processing, and this control of the evaluation that we give to the library functions is critical if you want to take this toy language and turn it into something real.<\/p><p>C++ itself is an eager language and this is why we cannot short-circuit in user-defined implementations of <code>&amp;&amp;<\/code> and <code>||<\/code>, something that can cause us some pain. <a class=\"External\" href=\"http:\/\/www.open-std.org\/jtc1\/sc22\/wg21\/docs\/papers\/2018\/p0927r2.pdf\">But there is hope<\/a>.<\/p>","created_at":"2019-02-09T13:27:08+0000","updated_at":"2019-02-09T13:27:08+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"kirit.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":823,"title":"wiki.xxiivv.com","url":"https:\/\/wiki.xxiivv.com\/#about","created_at":"2019-02-09T02:47:11+0000","updated_at":"2019-02-09T02:47:11+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"wiki.xxiivv.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":822,"title":"hundredrabbits\/Orca","url":"https:\/\/github.com\/hundredrabbits\/Orca","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/hundredrabbits\/Orca\/master\/resources\/logo.png\"><img src=\"https:\/\/raw.githubusercontent.com\/hundredrabbits\/Orca\/master\/resources\/logo.png\" width=\"600\" alt=\"image\" \/><\/a><\/p>\n<p><strong>Each letter of the alphabet is an operation<\/strong>, lowercase letters operate on bang(<code>*<\/code>), uppercase letters operate each frame. Have a look at some project created with <a href=\"https:\/\/twitter.com\/hashtag\/ORC%CE%9B\" rel=\"nofollow\">#ORC\u039b<\/a>, or some <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/tree\/master\/examples\">example files<\/a>. Here's an <a href=\"https:\/\/www.youtube.com\/watch?v=RaI_TuISSJE\" rel=\"nofollow\">introduction video<\/a>. You can see the <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/DESIGN.md\">design notes<\/a> for specs and upcoming features. If you need some help, visit the <a href=\"https:\/\/talk.lurk.org\/channel\/orca\" rel=\"nofollow\">chatroom<\/a>.<\/p>\n<p>For a portable version of Orca, built entirely in C, visit <a href=\"http:\/\/github.com\/hundredrabbits\/Orca-c\">Orca-c<\/a>.<\/p>\n<h2><a id=\"user-content-install--run\" class=\"anchor\" aria-hidden=\"true\" href=\"#install--run\"><\/a>Install &amp; Run<\/h2>\n<p>You can download <a href=\"https:\/\/hundredrabbits.itch.io\/orca\" rel=\"nofollow\">builds<\/a> for <strong>OSX, Windows and Linux<\/strong>, or if you wish to build it yourself, follow these steps:<\/p>\n<pre>git clone https:\/\/github.com\/hundredrabbits\/Orca.git\ncd Orca\/desktop\/\nnpm install\nnpm start\n<\/pre>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/hundredrabbits\/Orca\/master\/resources\/preview.jpg\"><img src=\"https:\/\/raw.githubusercontent.com\/hundredrabbits\/Orca\/master\/resources\/preview.jpg\" width=\"600\" alt=\"image\" \/><\/a><\/p>\n<h2><a id=\"user-content-operators\" class=\"anchor\" aria-hidden=\"true\" href=\"#operators\"><\/a>Operators<\/h2>\n<ul><li><code>A<\/code> <strong>add<\/strong>(a, b): Outputs the sum of inputs.<\/li>\n<li><code>B<\/code> <strong>bool<\/strong>(val): Bangs if input is not empty, or 0.<\/li>\n<li><code>C<\/code> <strong>clock<\/strong>('rate, mod): Outputs a constant value based on the runtime frame.<\/li>\n<li><code>D<\/code> <strong>delay<\/strong>('rate, mod): Bangs on a fraction of the runtime frame.<\/li>\n<li><code>E<\/code> <strong>east<\/strong>: Moves eastward, or bangs.<\/li>\n<li><code>F<\/code> <strong>if<\/strong>(a, b): Bangs if both inputs are equal.<\/li>\n<li><code>G<\/code> <strong>generator<\/strong>('x, 'y, 'len): Writes distant operators with offset.<\/li>\n<li><code>H<\/code> <strong>halt<\/strong>: Stops southward operators from operating.<\/li>\n<li><code>I<\/code> <strong>increment<\/strong>(min, max): Increments southward operator.<\/li>\n<li><code>J<\/code> <strong>jumper<\/strong>(val): Outputs the northward operator.<\/li>\n<li><code>K<\/code> <strong>konkat<\/strong>('len): Outputs multiple variables.<\/li>\n<li><code>L<\/code> <strong>loop<\/strong>('len): Loops a number of eastward operators.<\/li>\n<li><code>M<\/code> <strong>modulo<\/strong>(val, mod): Outputs the modulo of input.<\/li>\n<li><code>N<\/code> <strong>north<\/strong>: Moves Northward, or bangs.<\/li>\n<li><code>O<\/code> <strong>offset<\/strong>('x, 'y, val): Reads a distant operator with offset.<\/li>\n<li><code>P<\/code> <strong>push<\/strong>('len, 'key, val): Writes an eastward operator with offset.<\/li>\n<li><code>Q<\/code> <strong>query<\/strong>('x, 'y, 'len): Reads distant operators with offset.<\/li>\n<li><code>R<\/code> <strong>random<\/strong>(min, max): Outputs a random value.<\/li>\n<li><code>S<\/code> <strong>south<\/strong>: Moves southward, or bangs.<\/li>\n<li><code>T<\/code> <strong>track<\/strong>('len, 'key, val): Reads an eastward operator with offset.<\/li>\n<li><code>U<\/code> <strong>uturn<\/strong>('n, 'e, 's, 'w): Reverses movement of inputs.<\/li>\n<li><code>V<\/code> <strong>variable<\/strong>('write, read): Reads and write globally available variables.<\/li>\n<li><code>W<\/code> <strong>west<\/strong>: Moves westward, or bangs.<\/li>\n<li><code>X<\/code> <strong>teleport<\/strong>('x, 'y, val): Writes a distant operator with offset.<\/li>\n<li><code>Y<\/code> <strong>jymper<\/strong>(val): Outputs the westward operator.<\/li>\n<li><code>Z<\/code> <strong>zoom<\/strong>: Moves eastwardly, respawns west on collision.<\/li>\n<li><code>*<\/code> <strong>bang<\/strong>: Bangs neighboring operators.<\/li>\n<li><code>#<\/code> <strong>comment<\/strong>: Comments a line, or characters until the next hash.<\/li>\n<li><code>:<\/code> <strong>midi<\/strong>('channel, 'octave, 'note, velocity, length): Sends a MIDI note.<\/li>\n<li><code>;<\/code> <strong>udp<\/strong>: Sends a UDP message.<\/li>\n<li><code>=<\/code> <strong>osc<\/strong>(path): Sends a OSC message.<\/li>\n<\/ul><h2><a id=\"user-content-controls\" class=\"anchor\" aria-hidden=\"true\" href=\"#controls\"><\/a>Controls<\/h2>\n<h3><a id=\"user-content-terminal-controls\" class=\"anchor\" aria-hidden=\"true\" href=\"#terminal-controls\"><\/a>Terminal Controls<\/h3>\n<ul><li><code>enter<\/code> toggle insert\/write.<\/li>\n<li><code>space<\/code> toggle play\/pause.<\/li>\n<li><code>&gt;<\/code> increase BPM.<\/li>\n<li><code>&lt;<\/code> decrease BPM.<\/li>\n<li><code>shift+arrowKey<\/code> Expand cursor.<\/li>\n<li><code>ctrl+arrowKey<\/code> Leap cursor.<\/li>\n<li><code>alt+arrowKey<\/code> Move selection.<\/li>\n<\/ul><h3><a id=\"user-content-edit\" class=\"anchor\" aria-hidden=\"true\" href=\"#edit\"><\/a>Edit<\/h3>\n<ul><li><code>ctrl+c<\/code> copy selection.<\/li>\n<li><code>ctrl+x<\/code> cut selection.<\/li>\n<li><code>ctrl+v<\/code> paste selection.<\/li>\n<li><code>ctrl+z<\/code> undo.<\/li>\n<li><code>ctrl+shift+z<\/code> redo.<\/li>\n<\/ul><h3><a id=\"user-content-grid-controls\" class=\"anchor\" aria-hidden=\"true\" href=\"#grid-controls\"><\/a>Grid Controls<\/h3>\n<ul><li><code>]<\/code> increase grid size vertically.<\/li>\n<li><code>[<\/code> decrease grid size vertically.<\/li>\n<li><code>}<\/code> increase grid size horizontally.<\/li>\n<li><code>{<\/code> decrease grid size horizontally.<\/li>\n<li><code>ctrl\/meta+]<\/code> increase program size vertically.<\/li>\n<li><code>ctrl\/meta+[<\/code> decrease program size vertically.<\/li>\n<li><code>ctrl\/meta+}<\/code> increase program size horizontally.<\/li>\n<li><code>ctrl\/meta+{<\/code> decrease program size horizontally.<\/li>\n<\/ul><h3><a id=\"user-content-window\" class=\"anchor\" aria-hidden=\"true\" href=\"#window\"><\/a>Window<\/h3>\n<ul><li><code>ctrl+=<\/code> Zoom In.<\/li>\n<li><code>ctrl+-<\/code> Zoom Out.<\/li>\n<li><code>ctrl+0<\/code> Zoom Reset.<\/li>\n<li><code>tab<\/code> Toggle interface.<\/li>\n<li><code>backquote<\/code> Toggle background.<\/li>\n<\/ul><p>To open the console, press <code>ctrl+.<\/code>.<\/p>\n<h2><a id=\"user-content-midi\" class=\"anchor\" aria-hidden=\"true\" href=\"#midi\"><\/a>MIDI<\/h2>\n<p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/MIDI\" rel=\"nofollow\">MIDI<\/a> operator <code>:<\/code> takes up to 5 inputs('channel, 'octave, 'note, velocity, length).<\/p>\n<p>For example, <code>:25C<\/code>, is a <strong>C note, on the 5th octave, through the 3rd MIDI channel<\/strong>, <code>:04c<\/code>, is a <strong>C# note, on the 4th octave, through the 1st MIDI channel<\/strong>. Velocity is an optional value from <code>0<\/code>(0\/127) to <code>f<\/code>(127\/127). Note length is a value from <code>0<\/code>(1\/16) to <code>f<\/code>(16\/16), which is a ratio of a full bar, <em>f<\/em> being <code>16\/16<\/code>(a full bar), <em>8<\/em> being <code>1\/2<\/code>(half), <em>4<\/em> being <code>1\/4<\/code>(quarter). See it in action with <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/examples\/_midi.orca\">midi.orca<\/a>.<\/p>\n<h4><a id=\"user-content-list-midi-devices\" class=\"anchor\" aria-hidden=\"true\" href=\"#list-midi-devices\"><\/a>List Midi Devices<\/h4>\n<p>In console, type <code>terminal.io.midi.list()<\/code> to see the list of available midi devices.<\/p>\n<h4><a id=\"user-content-select-midi-device\" class=\"anchor\" aria-hidden=\"true\" href=\"#select-midi-device\"><\/a>Select Midi Device<\/h4>\n<p>In console, type <code>terminal.io.midi.select(1)<\/code> to select the second midi device.<\/p>\n<h2><a id=\"user-content-udp\" class=\"anchor\" aria-hidden=\"true\" href=\"#udp\"><\/a>UDP<\/h2>\n<p>The <a href=\"https:\/\/nodejs.org\/api\/dgram.html#dgram_socket_send_msg_offset_length_port_address_callback\" rel=\"nofollow\">UDP<\/a> operator <code>;<\/code> locks each consecutive eastwardly ports. For example, <code>;hello<\/code>, will send the string \"hello\", on bang, to the port <code>49160<\/code> on <code>localhost<\/code><\/p>\n<p>You can use the <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/listener.js\">listener.js<\/a> to test UDP messages. See it in action with <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/examples\/_udp.orca\">udp.orca<\/a>.<\/p>\n<h4><a id=\"user-content-select-udp-port\" class=\"anchor\" aria-hidden=\"true\" href=\"#select-udp-port\"><\/a>Select UDP Port<\/h4>\n<p>In console, type <code>terminal.io.udp.select(49160)<\/code> to select the <strong>49160<\/strong> udp port.<\/p>\n<h2><a id=\"user-content-osc\" class=\"anchor\" aria-hidden=\"true\" href=\"#osc\"><\/a>OSC<\/h2>\n<p>The <a href=\"https:\/\/github.com\/MylesBorins\/node-osc\">OSC<\/a> operator <code>=<\/code> locks each consecutive eastwardly ports.<\/p>\n<p>First char is used for path, nexts are sent as integers using <a href=\"https:\/\/github.com\/hundredrabbits\/Orca#base36-table\">base36 Table<\/a>. For example, <code>=1abc<\/code> will send <code>10<\/code>, <code>11<\/code> and <code>12<\/code> to <code>\/1<\/code>, via the port <code>49162<\/code> on <code>localhost<\/code>; <code>=a123<\/code> will send <code>1<\/code>, <code>2<\/code> and <code>3<\/code>, to the path <code>\/a<\/code>.  You can use the <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/examples\/listener.js\">listener.js<\/a> to test OSC messages. See it in action with <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/examples\/_osc.orca\">osc.orca<\/a> or try it with <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/TUTORIAL.md#sonicpi\">SonicPi<\/a>.<\/p>\n<h4><a id=\"user-content-select-osc-port\" class=\"anchor\" aria-hidden=\"true\" href=\"#select-osc-port\"><\/a>Select OSC Port<\/h4>\n<p>In console, type <code>terminal.io.osc.select(49162)<\/code> to select the <strong>49162<\/strong> osc port.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/hundredrabbits\/Orca\/master\/resources\/preview.hardware.jpg\"><img src=\"https:\/\/raw.githubusercontent.com\/hundredrabbits\/Orca\/master\/resources\/preview.hardware.jpg\" width=\"600\" alt=\"image\" \/><\/a><\/p>\n<h2><a id=\"user-content-base36-table\" class=\"anchor\" aria-hidden=\"true\" href=\"#base36-table\"><\/a>Base36 Table<\/h2>\n<p>Orca operates on a base of 36 increments. Operators using numeric values will typically also operate on letters and convert them into values as per the following table. For instance <code>Dp<\/code> will bang every <em>24th frame<\/em>.<\/p>\n<table><thead><tr><th align=\"center\">0<\/th>\n<th align=\"center\">1<\/th>\n<th align=\"center\">2<\/th>\n<th align=\"center\">3<\/th>\n<th align=\"center\">4<\/th>\n<th align=\"center\">5<\/th>\n<th align=\"center\">6<\/th>\n<th align=\"center\">7<\/th>\n<th align=\"center\">8<\/th>\n<th align=\"center\">9<\/th>\n<th align=\"center\">A<\/th>\n<th align=\"center\">B<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"center\">0<\/td>\n<td align=\"center\">1<\/td>\n<td align=\"center\">2<\/td>\n<td align=\"center\">3<\/td>\n<td align=\"center\">4<\/td>\n<td align=\"center\">5<\/td>\n<td align=\"center\">6<\/td>\n<td align=\"center\">7<\/td>\n<td align=\"center\">8<\/td>\n<td align=\"center\">9<\/td>\n<td align=\"center\">10<\/td>\n<td align=\"center\">11<\/td>\n<\/tr><tr><td align=\"center\"><strong>C<\/strong><\/td>\n<td align=\"center\"><strong>D<\/strong><\/td>\n<td align=\"center\"><strong>E<\/strong><\/td>\n<td align=\"center\"><strong>F<\/strong><\/td>\n<td align=\"center\"><strong>G<\/strong><\/td>\n<td align=\"center\"><strong>H<\/strong><\/td>\n<td align=\"center\"><strong>I<\/strong><\/td>\n<td align=\"center\"><strong>J<\/strong><\/td>\n<td align=\"center\"><strong>K<\/strong><\/td>\n<td align=\"center\"><strong>L<\/strong><\/td>\n<td align=\"center\"><strong>M<\/strong><\/td>\n<td align=\"center\"><strong>N<\/strong><\/td>\n<\/tr><tr><td align=\"center\">12<\/td>\n<td align=\"center\">13<\/td>\n<td align=\"center\">14<\/td>\n<td align=\"center\">15<\/td>\n<td align=\"center\">16<\/td>\n<td align=\"center\">17<\/td>\n<td align=\"center\">18<\/td>\n<td align=\"center\">19<\/td>\n<td align=\"center\">20<\/td>\n<td align=\"center\">21<\/td>\n<td align=\"center\">22<\/td>\n<td align=\"center\">23<\/td>\n<\/tr><tr><td align=\"center\"><strong>O<\/strong><\/td>\n<td align=\"center\"><strong>P<\/strong><\/td>\n<td align=\"center\"><strong>Q<\/strong><\/td>\n<td align=\"center\"><strong>R<\/strong><\/td>\n<td align=\"center\"><strong>S<\/strong><\/td>\n<td align=\"center\"><strong>T<\/strong><\/td>\n<td align=\"center\"><strong>U<\/strong><\/td>\n<td align=\"center\"><strong>V<\/strong><\/td>\n<td align=\"center\"><strong>W<\/strong><\/td>\n<td align=\"center\"><strong>X<\/strong><\/td>\n<td align=\"center\"><strong>Y<\/strong><\/td>\n<td align=\"center\"><strong>Z<\/strong><\/td>\n<\/tr><tr><td align=\"center\">24<\/td>\n<td align=\"center\">25<\/td>\n<td align=\"center\">26<\/td>\n<td align=\"center\">27<\/td>\n<td align=\"center\">28<\/td>\n<td align=\"center\">29<\/td>\n<td align=\"center\">30<\/td>\n<td align=\"center\">31<\/td>\n<td align=\"center\">32<\/td>\n<td align=\"center\">33<\/td>\n<td align=\"center\">34<\/td>\n<td align=\"center\">35<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-extras\" class=\"anchor\" aria-hidden=\"true\" href=\"#extras\"><\/a>Extras<\/h2>\n<ul><li>This application supports the <a href=\"https:\/\/github.com\/hundredrabbits\/Themes\">Ecosystem Theme<\/a>.<\/li>\n<li>Support this project through <a href=\"https:\/\/patreon.com\/100\" rel=\"nofollow\">Patreon<\/a>.<\/li>\n<li>See the <a href=\"https:\/\/github.com\/hundredrabbits\/Orca\/blob\/master\/LICENSE.md\">License<\/a> file for license rights and limitations (MIT).<\/li>\n<li>Pull Requests are welcome!<\/li>\n<\/ul><\/article>","created_at":"2019-02-09T02:32:42+0000","updated_at":"2019-02-09T02:32:42+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/15803194?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":821,"title":"Open Design: Freeware tools for Designers","url":"https:\/\/uxplanet.org\/open-design-freeware-tools-for-designers-f7bdde99f2e0?_branch_match_id=622421798263164882&gi=607c298cbc0a","content":"<section class=\"section section--body section--first\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"4c30\" class=\"graf graf--p graf-after--figure\">Sketch is the industry standard for High-fidelity UI\/UX design, It is also my favorite tool for the past 4 years, I love the plugins, and the integration with other tools (InVision, Zepplin, etc). But as a UX designer, I also use Adobe products (for design related tasks), and Axure RP (for low-fidelity wireframes, UX architecture, and mockups).<\/p><p id=\"59ed\" class=\"graf graf--p graf-after--p\">But when it comes to Adobe Products, just like many designers out there, I\u2019ve decided to ditch their Eco-System because of their Creative Cloud subscription model. And I have to admit, it won\u2019t be easy to ditch Photoshop, particularly because I worked with this amazing tool ever since I was 12 Years old (I think from version 5\u00a0;) ). In the recent years I have tried Affinity design, and so many Free and Open-Source alternatives to Commercial design tools, and I was really impressed! I have also found out on the web that I\u2019m not alone, many freelance designers, small businesses, and Startups tried open source alternatives, not only because they needed cheap or free alternatives, but because they needed Cross-Platform solution, doesn\u2019t matter if they are freelance designers, small design studios, or even Lean Startups with designer-developer workflows. But can a professional designer rely on <a href=\"https:\/\/en.wikipedia.org\/wiki\/Free_and_open-source_software\" data-href=\"https:\/\/en.wikipedia.org\/wiki\/Free_and_open-source_software\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Free and Open-Source software<\/a>?<\/p><p id=\"a630\" class=\"graf graf--p graf-after--p\">Well, I have talked with some of my friends and colleagues, and I was surprised by their answers\u2026 Most of them are UX and UI designers, so just like me, they are Sketch addicted. They also told me that they wish there was Sketch alternative for Windows, and even though I have suggested them to try Adobe XD for Windows (which I kinda like), they refused to do so because they don\u2019t want to go back to Adobe, and some of them are still on Windows 7 &amp; 8.1 (Adobe XD works only on Windows 10)\u2026 And then we talked about some free and Open-Source alternatives, but while there are plenty of alternatives to Adobe Products, unfortunately there aren\u2019t many alternatives to Sketch, which is only for Mac. But there are some open source UI design tools coming up, not to mention professional online tools that are already available, and some of them are even for free.<\/p><p id=\"1476\" class=\"graf graf--p graf-after--p\">After so many years as a Mac user, I want to switch from Mac to Linux (that\u2019s a story for different day). In fact, Linux is my secondary OS on my Mac, and on my PC as well. Hence there are so many Linux Distros for designers (Yes! Linux for Designers!), my favorites are <a href=\"https:\/\/ubuntustudio.org\/tour\/\" data-href=\"https:\/\/ubuntustudio.org\/tour\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Ubuntu Studio<\/a>, <a href=\"https:\/\/ubuntu-mate.org\/\" data-href=\"https:\/\/ubuntu-mate.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Ubuntu Mate<\/a> (with Cupertino UI), and if you are coming from Mac, you\u2019ll definitely love <a href=\"https:\/\/elementary.io\/\" data-href=\"https:\/\/elementary.io\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Elementary OS<\/a>! The point is, when it comes to Free and Open-Source Software, it\u2019s truly Cross-platform and you can use it on any platform, (Mac, Windows, and Linux), so in case you\u2019ve tried Inkscape on a Mac, you\u2019ll be able to use it also on Windows and Linux without any learning curves, and just because it\u2019s FOSS there are no subscription models!<\/p><p id=\"5d67\" class=\"graf graf--p graf-after--p graf--trailing\">So, are there decent (and free) alternatives for Adobe CC, Sketch, and Axure for the Cross-Platform Eco-systems (Mac, Windows, and Even Linux)?<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"c5f8\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">Let\u2019s start with Free and Open-Source Software (FOSS) alternatives to Adobe\u00a0CC:<\/strong><\/h3><p id=\"214d\" class=\"graf graf--p graf-after--h3\"><a href=\"https:\/\/krita.org\/en\/\" data-href=\"https:\/\/krita.org\/en\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Krita<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Photoshop<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"0650\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*G-zHlz60aQiE9Hsj.png\" data-width=\"1280\" data-height=\"776\" data-action=\"zoom\" data-action-value=\"0*G-zHlz60aQiE9Hsj.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*G-zHlz60aQiE9Hsj.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Krita.org<\/figcaption><\/figure><p id=\"b7e8\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.gimp.org\/\" data-href=\"https:\/\/www.gimp.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Gimp<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Photoshop<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"44a2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*SeCqBwt7NrFGbfJP.jpg\" data-width=\"960\" data-height=\"586\" data-action=\"zoom\" data-action-value=\"0*SeCqBwt7NrFGbfJP.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*SeCqBwt7NrFGbfJP.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by\u00a0Gimp.org<\/figcaption><\/figure><p id=\"c787\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/inkscape.org\/en\/\" data-href=\"https:\/\/inkscape.org\/en\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Inkscape<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Illustrator<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"5bd1\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*tKWGMM1pPhwOJyh6.png\" data-width=\"1680\" data-height=\"1050\" data-action=\"zoom\" data-action-value=\"0*tKWGMM1pPhwOJyh6.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*tKWGMM1pPhwOJyh6.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Inkscape.org<\/figcaption><\/figure><p id=\"0aa1\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.darktable.org\/install\/\" data-href=\"https:\/\/www.darktable.org\/install\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">DarkTable<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe LightRoom <\/strong>(Available for Windows, Mac, and Linux)<\/p><figure id=\"2338\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*IXzGAc50r_X7LMAu.jpg\" data-width=\"1920\" data-height=\"1080\" data-action=\"zoom\" data-action-value=\"0*IXzGAc50r_X7LMAu.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*IXzGAc50r_X7LMAu.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by darktable.org<\/figcaption><\/figure><p id=\"6363\" class=\"graf graf--p graf-after--figure\"><a href=\"http:\/\/lightzoneproject.org\/\" data-href=\"http:\/\/lightzoneproject.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">LightZone<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe LightRoom<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"dc2a\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*orTDtvVq89YvrbFy.jpg\" data-width=\"1920\" data-height=\"1200\" data-action=\"zoom\" data-action-value=\"0*orTDtvVq89YvrbFy.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*orTDtvVq89YvrbFy.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by lightzoneproject.org<\/figcaption><\/figure><p id=\"0b6b\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.scribus.net\/\" data-href=\"https:\/\/www.scribus.net\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Scribus<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Indesign<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"3d7c\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*mEKbnLIK_9oqPFdb.png\" data-width=\"1200\" data-height=\"675\" data-action=\"zoom\" data-action-value=\"0*mEKbnLIK_9oqPFdb.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*mEKbnLIK_9oqPFdb.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Wikipedia<\/figcaption><\/figure><p id=\"7f40\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/ardour.org\/\" data-href=\"https:\/\/ardour.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Ardour<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Audition<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"a136\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*9EzfWjTP8DKpED_Q.png\" data-width=\"2000\" data-height=\"1190\" data-action=\"zoom\" data-action-value=\"0*9EzfWjTP8DKpED_Q.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*9EzfWjTP8DKpED_Q.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by ardour.org<\/figcaption><\/figure><p id=\"aff1\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.audacityteam.org\/\" data-href=\"https:\/\/www.audacityteam.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Audacity<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Audition<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"1354\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*zBrSR9AhtEhXwLYi.png\" data-width=\"921\" data-height=\"585\" data-action=\"zoom\" data-action-value=\"0*zBrSR9AhtEhXwLYi.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*zBrSR9AhtEhXwLYi.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by audacityteam.org<\/figcaption><\/figure><p id=\"5d40\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.synfig.org\/\" data-href=\"https:\/\/www.synfig.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Synfig Studio<\/strong><\/a><strong class=\"markup--strong markup--p-strong\">\u200a<\/strong>\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Animate <\/strong>\/ Flash Pro (Available for Windows, Mac, and Linux).<\/p><figure id=\"d61f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*ASI5quXgUq-y9_a2.png\" data-width=\"1140\" data-height=\"510\" data-action=\"zoom\" data-action-value=\"0*ASI5quXgUq-y9_a2.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*ASI5quXgUq-y9_a2.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by synfig.org<\/figcaption><\/figure><p id=\"b9e3\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.blackmagicdesign.com\/products\/fusion\/\" data-href=\"https:\/\/www.blackmagicdesign.com\/products\/fusion\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Fusion<\/strong><\/a><strong class=\"markup--strong markup--p-strong\">\u200a<\/strong>\u2014\u200aProfessional and Free alternative to <strong class=\"markup--strong markup--p-strong\">Adobe After Effects <\/strong>(Available for Windows, Mac, and Linux).<\/p><figure id=\"40e2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*fVJY00I0iikzBMag.\" data-width=\"1120\" data-height=\"672\" data-action=\"zoom\" data-action-value=\"0*fVJY00I0iikzBMag.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*fVJY00I0iikzBMag.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by blackmagicdesign.com<\/figcaption><\/figure><p id=\"aeb0\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/natron.fr\/\" data-href=\"https:\/\/natron.fr\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Natron<\/strong><\/a><strong class=\"markup--strong markup--p-strong\">\u200a<\/strong>\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe After Effects<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"923f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*4Y_9uXqjovByG9G5z_iq9A.png\" data-width=\"968\" data-height=\"462\" data-action=\"zoom\" data-action-value=\"1*4Y_9uXqjovByG9G5z_iq9A.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*4Y_9uXqjovByG9G5z_iq9A.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by natron.fr<\/figcaption><\/figure><p id=\"cf79\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/kdenlive.org\/\" data-href=\"https:\/\/kdenlive.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Kdenlive<\/strong><\/a><strong class=\"markup--strong markup--p-strong\">\u200a<\/strong>\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Premiere<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"ddd3\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*l78U-xsDQ04zajnc.png\" data-width=\"1057\" data-height=\"595\" data-action=\"zoom\" data-action-value=\"0*l78U-xsDQ04zajnc.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*l78U-xsDQ04zajnc.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by kdenlive.org<\/figcaption><\/figure><p id=\"4f94\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.blackmagicdesign.com\/products\/davinciresolve\/\" data-href=\"https:\/\/www.blackmagicdesign.com\/products\/davinciresolve\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">DaVinci Resolve<\/strong><\/a>\u200a\u2014\u200aProfessional and free alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Premiere <\/strong>(Available for Windows, Mac, and Linux).<\/p><figure id=\"645f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*ESvKqT8wn5WYprXZ.\" data-width=\"1440\" data-height=\"800\" data-action=\"zoom\" data-action-value=\"0*ESvKqT8wn5WYprXZ.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*ESvKqT8wn5WYprXZ.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by blackmagicdesign.com<\/figcaption><\/figure><p id=\"5332\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.openshot.org\/\" data-href=\"https:\/\/www.openshot.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">OpenShot<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe Premiere<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"d5c4\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*an-odt8p5ajIRF2D.jpg\" data-width=\"975\" data-height=\"637\" data-action=\"zoom\" data-action-value=\"0*an-odt8p5ajIRF2D.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*an-odt8p5ajIRF2D.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by openshot.org<\/figcaption><\/figure><p id=\"227e\" class=\"graf graf--p graf-after--figure\"><a href=\"http:\/\/www.aptana.com\/index.html\" data-href=\"http:\/\/www.aptana.com\/index.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Aptana Studio<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe DreamWeaver<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"48fe\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*8uguD3UWIxUAHXmh.\" data-width=\"1202\" data-height=\"731\" data-action=\"zoom\" data-action-value=\"0*8uguD3UWIxUAHXmh.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*8uguD3UWIxUAHXmh.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by alternativeto.net<\/figcaption><\/figure><p id=\"6adb\" class=\"graf graf--p graf-after--figure\"><a href=\"http:\/\/brackets.io\/\" data-href=\"http:\/\/brackets.io\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Brackets<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Adobe DreamWeaver<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"abbb\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*VzxgdmN0OjFHBCPv.png\" data-width=\"1126\" data-height=\"589\" data-action=\"zoom\" data-action-value=\"0*VzxgdmN0OjFHBCPv.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*VzxgdmN0OjFHBCPv.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Ubuntufree.com<\/figcaption><\/figure><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"222c\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">3D Design (Free and Open-Source):<\/strong><\/h3><p id=\"7844\" class=\"graf graf--p graf-after--h3\"><a href=\"https:\/\/www.blender.org\/\" data-href=\"https:\/\/www.blender.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Blender<\/strong><\/a>\u200a\u2014\u200aProfessional and FOSS alternative to <strong class=\"markup--strong markup--p-strong\">3D Studio \/ Maya<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"e4bb\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*kwP-25Tq7uBZJ-Gs.\" data-width=\"1024\" data-height=\"621\" data-action=\"zoom\" data-action-value=\"0*kwP-25Tq7uBZJ-Gs.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*kwP-25Tq7uBZJ-Gs.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by blender.org<\/figcaption><\/figure><p id=\"6b8a\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.freecadweb.org\/\" data-href=\"https:\/\/www.freecadweb.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">FreeCad<\/strong><\/a><strong class=\"markup--strong markup--p-strong\">\u200a<\/strong>\u2014\u200aFOSS alternative to<strong class=\"markup--strong markup--p-strong\"> AutoCad <\/strong>(Available for Windows, Mac, and Linux).<\/p><figure id=\"891d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*G0G-DFIza1nnns6_.png\" data-width=\"800\" data-height=\"450\" data-action=\"zoom\" data-action-value=\"0*G0G-DFIza1nnns6_.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*G0G-DFIza1nnns6_.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by freecadweb.org<\/figcaption><\/figure><p id=\"629c\" class=\"graf graf--p graf-after--figure\"><a href=\"http:\/\/librecad.org\/cms\/home.html\" data-href=\"http:\/\/librecad.org\/cms\/home.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">LibreCad<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">AutoCad<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"11de\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*_jKV-kDe4D42Gjx4.png\" data-width=\"800\" data-height=\"500\" data-action=\"zoom\" data-action-value=\"0*_jKV-kDe4D42Gjx4.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*_jKV-kDe4D42Gjx4.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by freesoftwarefiles.com<\/figcaption><\/figure><p id=\"9ea6\" class=\"graf graf--p graf-after--figure\"><a href=\"http:\/\/www.sweethome3d.com\/\" data-href=\"http:\/\/www.sweethome3d.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Sweet Home 3D<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Sketchup 3D<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"f274\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*8bdgnLZPPgumGpGN.jpg\" data-width=\"1200\" data-height=\"614\" data-action=\"zoom\" data-action-value=\"0*8bdgnLZPPgumGpGN.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*8bdgnLZPPgumGpGN.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by sweethome3d.com<\/figcaption><\/figure><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"f21b\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">UX and Wire-framing tools:<\/strong><\/h3><p id=\"ad09\" class=\"graf graf--p graf-after--h3\"><a href=\"https:\/\/pencil.evolus.vn\/\" data-href=\"https:\/\/pencil.evolus.vn\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Pencil<\/strong><\/a>\u200a\u2014\u200aThe only FOSS alternative to <strong class=\"markup--strong markup--p-strong\">Axure RP<\/strong> (Available for Windows, Mac, and Linux).<\/p><figure id=\"b60f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*_G3VOL8yNYydx8pw.\" data-width=\"1920\" data-height=\"1048\" data-action=\"zoom\" data-action-value=\"0*_G3VOL8yNYydx8pw.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*_G3VOL8yNYydx8pw.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by alternativeto.net<\/figcaption><\/figure><p id=\"e73b\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/jdittrich.github.io\/quickMockup\/\" data-href=\"https:\/\/jdittrich.github.io\/quickMockup\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">quickMockup<\/strong><\/a>\u200a\u2014\u200aFOSS alternative to <strong class=\"markup--strong markup--p-strong\">Balsamiq <\/strong>\/<strong class=\"markup--strong markup--p-strong\"> Axure RP<\/strong> \/ <strong class=\"markup--strong markup--p-strong\">Invision<\/strong> (Available online).<\/p><figure id=\"0301\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*8yBgux68R3H2elxHIHOqbQ.png\" data-width=\"1414\" data-height=\"890\" data-action=\"zoom\" data-action-value=\"1*8yBgux68R3H2elxHIHOqbQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*8yBgux68R3H2elxHIHOqbQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Avi-B.com<\/figcaption><\/figure><p id=\"1c06\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/balsamiq.com\/\" data-href=\"https:\/\/balsamiq.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Balsamiq<\/strong><\/a>\u200a\u2014\u200aFree alternative to <strong class=\"markup--strong markup--p-strong\">Axure RP<\/strong> (Available for Windows, Mac, and it works on Linux via Wine).<\/p><figure id=\"031b\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*xayzl4214F-rlVNA.jpg\" data-width=\"959\" data-height=\"711\" data-action=\"zoom\" data-action-value=\"0*xayzl4214F-rlVNA.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*xayzl4214F-rlVNA.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by\u00a0Balsamiq<\/figcaption><\/figure><p id=\"9c74\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/wireframe.cc\/\" data-href=\"https:\/\/wireframe.cc\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Wireframe<\/strong><\/a>\u200a\u2014\u200aFree alternative to Axure RP (Non Open-Source, Available online).<\/p><figure id=\"b591\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*ov1wrc6q_UkcSePG8H76sw.png\" data-width=\"1737\" data-height=\"844\" data-action=\"zoom\" data-action-value=\"1*ov1wrc6q_UkcSePG8H76sw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ov1wrc6q_UkcSePG8H76sw.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by wireframe.cc<\/figcaption><\/figure><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"7959\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">Free and Open-Source UI design Tools, So are there any alternatives to\u00a0Sketch?<\/strong><\/h3><p id=\"562c\" class=\"graf graf--p graf-after--h3\">Sketch itself is only available for Mac, and it\u2019s not a Free and Open-Source. But it has a unique Eco-System full with plugins unlike any other tool, Sketch also has integration with other tools and it\u2019s so powerful! But unfortunately there are not many Open-Source alternatives to Sketch yet, but there\u2019s a hope!<\/p><p id=\"c1af\" class=\"graf graf--p graf-after--p\"><a href=\"https:\/\/www.designer.io\/\" data-href=\"https:\/\/www.designer.io\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Gravit Designer<\/strong><\/a>\u200a\u2014\u200aMy favorite alternative to <strong class=\"markup--strong markup--p-strong\">Sketch<\/strong>, although it\u2019s not FOSS (Free and Open-Source) anymore, but at least it\u2019s free of charge and available for the Cross-Platform: Web, Mac, Windows, and Linux! It supports\u00a0.sketch file format, and it\u2019s not only web based! They don\u2019t force you to subscribe, you can download it and use it just like Sketch, and it even works offline, in case you are working on some Secret \/ NDA Startup projects and you don\u2019t want to use cloud \/ web based products (such as InVision, Figma, Marvel, etc).<\/p><figure id=\"2082\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*RkuPrLxKVKB6IB-QWhADGg.png\" data-width=\"1920\" data-height=\"823\" data-action=\"zoom\" data-action-value=\"1*RkuPrLxKVKB6IB-QWhADGg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*RkuPrLxKVKB6IB-QWhADGg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Avi-B.com<\/figcaption><\/figure><p id=\"27b8\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/vectr.com\/\" data-href=\"https:\/\/vectr.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Vectr<\/strong><\/a>\u200a\u2014\u200aa simple yet powerful cross-platform graphics editor to bring your designs into reality, it\u2019s free but not Open-Source, and available for the Cross-Platform: Web, Mac, Windows, and Linux. In my opinion it\u2019s not powerful like Sketch, or even as Gravit designer, but it has its advantages, and worth checking out.<\/p><figure id=\"035f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*wATr6F3CzuBD35y7PDPK1g.png\" data-width=\"968\" data-height=\"462\" data-action=\"zoom\" data-action-value=\"1*wATr6F3CzuBD35y7PDPK1g.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*wATr6F3CzuBD35y7PDPK1g.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Avi-B.com<\/figcaption><\/figure><p id=\"cede\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/meetalva.io\/\" data-href=\"https:\/\/meetalva.io\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Alva<\/strong><\/a>\u200a\u2014\u200aIt\u2019s Free and Open-Source (FOSS) and works with your production frontend components, adding up-to-date, responsive, and interactive designs to your living styleguides. It\u2019s not the best alternative to Sketch and Gravit, but I guess it has its advantages.<\/p><figure id=\"0598\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*AK5ffqlBvzFgmJtD.png\" data-width=\"2000\" data-height=\"1093\" data-action=\"zoom\" data-action-value=\"0*AK5ffqlBvzFgmJtD.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*AK5ffqlBvzFgmJtD.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by meetalva.io<\/figcaption><\/figure><p id=\"4eef\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/github.com\/Alecaddd\/Akira\" data-href=\"https:\/\/github.com\/Alecaddd\/Akira\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Akira<\/strong><\/a>\u200a\u2014\u200aThe only Free and Open-Source (FOSS) <strong class=\"markup--strong markup--p-strong\">Sketch Alternative<\/strong> out there! Currently in development, you can follow the development <a href=\"https:\/\/forum.alecaddd.com\/d\/50-akira-development\" data-href=\"https:\/\/forum.alecaddd.com\/d\/50-akira-development\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">here<\/a>, and you are also welcome to help the developers since the project is open source!<\/p><figure id=\"ec04\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*2j5ePMlmfihhXz2j.\" data-width=\"1440\" data-height=\"1024\" data-action=\"zoom\" data-action-value=\"0*2j5ePMlmfihhXz2j.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*2j5ePMlmfihhXz2j.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by\u00a0<a href=\"https:\/\/github.com\/Alecaddd\" data-href=\"https:\/\/github.com\/Alecaddd\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener\" target=\"_blank\">Alecaddd<\/a><\/figcaption><\/figure><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"67ae\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">More free and Online UI Design Tools (Commercial and professional alternatives to\u00a0Sketch):<\/strong><\/h3><p id=\"9565\" class=\"graf graf--p graf-after--h3\"><a href=\"https:\/\/www.figma.com\/\" data-href=\"https:\/\/www.figma.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Figma<\/strong><\/a>\u200a\u2014\u200aIt\u2019s the best alternative to <strong class=\"markup--strong markup--p-strong\">Sketch<\/strong>, and just like Google Docs, it\u2019s a web based and powerful cloud solution! It\u2019s a collaborative interface design tool. Enables you to keep the team focused and on the same page with real-time communication and collaboration featuring commercial quality. You have to sign-up in order to use Figma.<\/p><figure id=\"c40f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*o_aFoOkninmHh2YDCIeRjg.png\" data-width=\"968\" data-height=\"462\" data-action=\"zoom\" data-action-value=\"1*o_aFoOkninmHh2YDCIeRjg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*o_aFoOkninmHh2YDCIeRjg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by Figma.com<\/figcaption><\/figure><p id=\"e757\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/marvelapp.com\/features\/\" data-href=\"https:\/\/marvelapp.com\/features\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Marvel<\/strong><\/a>\u200a\u2014\u200aMarvel is a free design and prototyping tool that transforms static design files (Sketch, Photoshop files, etc) and wire-frames into interactive, sharable prototypes that look and feel like real apps and websites.<\/p><figure id=\"bf0a\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*VERYKnO8mrBXnH6K.jpg\" data-width=\"792\" data-height=\"534\" data-action=\"zoom\" data-action-value=\"0*VERYKnO8mrBXnH6K.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*VERYKnO8mrBXnH6K.jpg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by marvelapp.com<\/figcaption><\/figure><p id=\"85a4\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/www.invisionapp.com\/\" data-href=\"https:\/\/www.invisionapp.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">InVision<\/strong><\/a>\u200a\u2014\u200aGreat prototyping tool, lets you create interactive wire-frames and also high-fidelity prototypes. You can easily share your work and get instant feedback right on-screen. If you are not on Linux, you should definitely check <a href=\"https:\/\/www.invisionapp.com\/studio\" data-href=\"https:\/\/www.invisionapp.com\/studio\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">InVision Studio<\/strong><\/a>, which is awesome and powerful tool for Mac and Windows users (similar to Adobe XD).<\/p><figure id=\"cff7\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*Yu1z2t8oOo2w7Se6.png\" data-width=\"1280\" data-height=\"800\" data-action=\"zoom\" data-action-value=\"0*Yu1z2t8oOo2w7Se6.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*Yu1z2t8oOo2w7Se6.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Image by invisionapp.com<\/figcaption><\/figure><h3 id=\"fa4e\" class=\"graf graf--h3 graf-after--figure\">In summary:<\/h3><p id=\"a80a\" class=\"graf graf--p graf-after--h3\">Finally, when it comes to Free and Open-Source Software and products, it\u2019s so good to see that there are so many designers involved, I remember when I tried Open-Source products 10\u201315 years ago, everything was so ugly and with a poor design, even most of the Linux distros I\u2019ve tried. But now, there are even <a href=\"http:\/\/beautifulopen.com\/\" data-href=\"http:\/\/beautifulopen.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Showcase websites for open source projects<\/a>, for your inspiration\u2026<\/p><p id=\"63e1\" class=\"graf graf--p graf-after--p graf--trailing\">I\u2019ll keep exploring other alternatives and will update this post from time to time, and you are welcome to <strong class=\"markup--strong markup--p-strong\">let me know in the comments<\/strong>, which free and open source product you use as an alternative to any Commercial products.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h4 id=\"9009\" class=\"graf graf--h4 graf--leading\">If you enjoyed this, read my other articles:<\/h4><p id=\"8ba1\" class=\"graf graf--p graf-after--h4\"><a href=\"https:\/\/medium.com\/adventures-in-consumer-technology\/iphone-vs-macbook-pro-9902c7fc64d3\" data-href=\"https:\/\/medium.com\/adventures-in-consumer-technology\/iphone-vs-macbook-pro-9902c7fc64d3\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">iPhone Vs. MacBook Pro<\/a><br \/><a href=\"https:\/\/medium.com\/adventures-in-consumer-technology\/smartphones-will-replace-pcs-soon-541b5c8a4f48\" data-href=\"https:\/\/medium.com\/adventures-in-consumer-technology\/smartphones-will-replace-pcs-soon-541b5c8a4f48\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Your Smartphone vs. Your PC<\/a><a href=\"https:\/\/uxplanet.org\/apple-its-time-to-merge-ios-with-macos-4f04c947c19d\" data-href=\"https:\/\/uxplanet.org\/apple-its-time-to-merge-ios-with-macos-4f04c947c19d\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><br \/>Apple, it\u2019s time to merge iOS with macOS<\/a><br \/><a href=\"https:\/\/uxplanet.org\/the-differences-between-vr-ar-mr-27012ea1c5\" data-href=\"https:\/\/uxplanet.org\/the-differences-between-vr-ar-mr-27012ea1c5\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">The differences between VR, AR &amp; MR<\/a><br \/><a href=\"https:\/\/medium.com\/@4Barel\/apple-is-going-to-reinvent-the-modular-computer-powered-by-an-iphone-1ccf758514df\" data-href=\"https:\/\/medium.com\/@4Barel\/apple-is-going-to-reinvent-the-modular-computer-powered-by-an-iphone-1ccf758514df\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Apple is going to reinvent the modular computer<\/a><br \/><a href=\"https:\/\/uxplanet.org\/cross-platform-guidelines-1c7675ec1d0c\" data-href=\"https:\/\/uxplanet.org\/cross-platform-guidelines-1c7675ec1d0c\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Cross-Platform UX Guidelines (for Devs &amp; Designers)<\/a><br \/><a href=\"https:\/\/medium.com\/@4Barel\/ux-vs-cybersecurity-3eedf77ed6e7\" data-href=\"https:\/\/medium.com\/@4Barel\/ux-vs-cybersecurity-3eedf77ed6e7\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">UX Vs. Cyber-security\u200a\u2014\u200aWhere we draw the line?<\/a><br \/><a href=\"https:\/\/medium.com\/@4Barel\/design-thinking-as-a-framework-for-innovation-e18e24cd4920\" data-href=\"https:\/\/medium.com\/@4Barel\/design-thinking-as-a-framework-for-innovation-e18e24cd4920\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Design Thinking as a framework for Innovation<\/a><br \/><a href=\"https:\/\/medium.com\/@4Barel\/lifehacks-for-productive-reading-9d265b97c444\" data-href=\"https:\/\/medium.com\/@4Barel\/lifehacks-for-productive-reading-9d265b97c444\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Lifehacks for Productive reading!<\/a><\/p><blockquote id=\"91bc\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">Help others<\/strong> to find my article on Medium by <strong class=\"markup--strong markup--blockquote-strong\">Clapping <\/strong>\ufe0fbelow.<br \/>And follow me on: <a href=\"https:\/\/twitter.com\/4Barel\" data-href=\"https:\/\/twitter.com\/4Barel\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener nofollow noopener noopener\" target=\"_blank\"><strong class=\"markup--strong markup--blockquote-strong\">Twitter<\/strong><\/a>, <a href=\"https:\/\/medium.com\/@4Barel\" data-href=\"https:\/\/medium.com\/@4Barel\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\"><strong class=\"markup--strong markup--blockquote-strong\">Medium<\/strong><\/a><strong class=\"markup--strong markup--blockquote-strong\">, <\/strong><a href=\"https:\/\/www.facebook.com\/startux.net\/\" data-href=\"https:\/\/www.facebook.com\/startux.net\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--blockquote-strong\">Facebook<\/strong><\/a><strong class=\"markup--strong markup--blockquote-strong\"> <\/strong>and <a href=\"https:\/\/www.linkedin.com\/in\/abarel\/\" data-href=\"https:\/\/www.linkedin.com\/in\/abarel\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener nofollow noopener noopener\" target=\"_blank\"><strong class=\"markup--strong markup--blockquote-strong\">LinkedIn<\/strong><\/a><\/div><\/blockquote><p id=\"39c1\" class=\"graf graf--p graf-after--blockquote graf--trailing\"><em class=\"markup--em markup--p-em\">Originally published by Avi Barel at <\/em><a href=\"http:\/\/www.startux.net\/open-design-freeware-tools-for-designers\/\" data-href=\"http:\/\/www.startux.net\/open-design-freeware-tools-for-designers\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">StartUX<\/em><\/a><em class=\"markup--em markup--p-em\"> on Feb 7, 2018.<\/em><\/p><\/div><\/div><\/section><section class=\"section section--body\"><section class=\"section section--body section--last\"><\/section><\/section>","created_at":"2019-02-08T14:12:28+0000","updated_at":"2019-02-08T14:12:28+0000","published_at":"2018-02-07T00:38:31+0000","published_by":["Avi Barel"],"annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"uxplanet.org","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*vx4OLoRuBxE1VZ7rRp0MLg.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":820,"title":"feathericons.com","url":"https:\/\/feathericons.com\/","created_at":"2019-02-08T13:37:04+0000","updated_at":"2019-02-08T13:37:04+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"feathericons.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":819,"title":"7 Rules for Creating Gorgeous UI (Part 1)","url":"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-1-559d4e805cda?_branch_match_id=622401978163552507","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"8386\" class=\"graf graf--figure graf--layoutOutsetCenter graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*X68DND6Lh8GFvM3i49KOhA.png\" data-width=\"8000\" data-height=\"3612\" data-action=\"zoom\" data-action-value=\"1*X68DND6Lh8GFvM3i49KOhA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*X68DND6Lh8GFvM3i49KOhA.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"title\" class=\"graf graf--h2 graf-after--figure graf--title\">7 Rules for Creating Gorgeous UI (Updated for\u00a02019)<\/h1><div class=\"uiScale uiScale-ui--regular uiScale-caption--regular u-flexCenter u-marginVertical24 u-fontSize15 js-postMetaLockup\"><div class=\"u-flex0\"><a class=\"link u-baseColor--link avatar\" href=\"https:\/\/medium.com\/@erikdkennedy?source=post_header_lockup\" data-action=\"show-user-card\" data-action-source=\"post_header_lockup\" data-action-value=\"a25a101a5693\" data-action-type=\"hover\" data-user-id=\"a25a101a5693\" dir=\"auto\"><img src=\"https:\/\/cdn-images-1.medium.com\/fit\/c\/100\/100\/0*6Y-XULF_0OzPokWU.png\" class=\"avatar-image u-size50x50\" alt=\"Go to the profile of Erik D. Kennedy\" \/><\/a><\/div><div class=\"u-flex1 u-paddingLeft15 u-overflowHidden\"><div class=\"u-paddingBottom3\"><a class=\"ds-link ds-link--styleSubtle ui-captionStrong u-inlineBlock link link--darken link--darker\" href=\"https:\/\/medium.com\/@erikdkennedy\" data-action=\"show-user-card\" data-action-value=\"a25a101a5693\" data-action-type=\"hover\" data-user-id=\"a25a101a5693\" dir=\"auto\">Erik D. Kennedy<\/a><button class=\"button button--smallest u-noUserSelect button--withChrome u-baseColor--buttonNormal button--withHover button--unblock js-unblockButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-block-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-1-559d4e805cda?_branch_match_id=622401978163552507\" data-action-source=\"post_header_lockup\">BlockedUnblock<\/button><button class=\"button button--primary button--smallest button--dark u-noUserSelect button--withChrome u-accentColor--buttonDark button--follow js-followButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-subscribe-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/_\/subscribe\/user\/a25a101a5693\" data-action-source=\"post_header_lockup-a25a101a5693-------------------------follow_byline\">FollowFollowing<\/button><\/div><div class=\"ui-caption u-noWrapWithEllipsis js-testPostMetaInlineSupplemental\"><time datetime=\"2014-11-13T21:59:07.598Z\">Nov 13, 2014<\/time><\/div><\/div><\/div><blockquote id=\"subtitle\" class=\"graf graf--pullquote graf-after--h2\"><div>A non-artsy guide to creating beautiful apps and\u00a0sites<\/div><\/blockquote><h3 id=\"d598\" class=\"graf graf--h3 graf-after--pullquote\">Introduction<\/h3><p id=\"bed2\" class=\"graf graf--p graf-after--h3\">OK, first things first. This guide is not for everyone. Who is this guide for?<\/p><ul class=\"postList\"><li id=\"fbae\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Developers <\/strong>who want to be able to design their own good-looking UI in a pinch.<\/li><li id=\"2b46\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">UX designers <\/strong>who want their portfolio to look better than a Pentagon PowerPoint. Or UX designers who know they can sell an awesome UX better in a pretty UI package.<\/li><\/ul><p id=\"f9ce\" class=\"graf graf--p graf-after--li\">If you went to art school or consider yourself a UI designer already, you will likely find this guide some combination of a.) boring, b.) wrong, and c.) irritating. That\u2019s fine. All your criticisms are right. Close the tab, move along.<\/p><p id=\"3bde\" class=\"graf graf--p graf-after--p\">Let me tell you what you\u2019ll find in this guide.<\/p><p id=\"4b87\" class=\"graf graf--p graf-after--p\">First, I was a UX designer with no UI skills. I <em class=\"markup--em markup--p-em\">love<\/em> designing UX, but I wasn\u2019t doing it for long before I realized there were a bunch of good reasons to learn how to make an interface look nice:<\/p><ul class=\"postList\"><li id=\"8c45\" class=\"graf graf--li graf-after--p\">My portfolio looked like crap, reflecting poorly on my work and thought process<\/li><li id=\"ed4a\" class=\"graf graf--li graf-after--li\">My UX consulting clients would rather buy someone\u2019s skills if their expertise extended to more than just <em class=\"markup--em markup--li-em\">sketching boxes and arrows<\/em><\/li><li id=\"611e\" class=\"graf graf--li graf-after--li\">Did I want to work for an early-stage startup at some point? Best to be a sweeper<\/li><\/ul><p id=\"415d\" class=\"graf graf--p graf-after--li\">I had my excuses. <em class=\"markup--em markup--p-em\">I don\u2019t know crap about aesthetics. I majored in engineering \u2013 it\u2019s almost a badge of pride to build something that looks awful.<\/em><\/p><blockquote id=\"ff7d\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div><em class=\"markup--em markup--pullquote-em\">\u201cI majored in engineering\u200a\u2014\u200ait\u2019s almost a badge of<br \/>pride to build something that looks\u00a0awful.\u201d<\/em><\/div><\/blockquote><p id=\"f533\" class=\"graf graf--p graf-after--pullquote\">In the end, I learned the aesthetics of apps the same way I\u2019ve learned any creative endeavor: <em class=\"markup--em markup--p-em\">cold, hard analysis<\/em>. And shameless copying of what works. I\u2019ve worked 10 hours on a UI project and billed for 1. The other 9 were the wild flailing of learning. Desperately searching Google and Pinterest and Dribbble for something to copy from.<\/p><p id=\"9571\" class=\"graf graf--p graf-after--p\">These \u201crules\u201d are the lessons from those hours.<\/p><p id=\"e4aa\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">So word to the nerds: if I\u2019m any good at designing UI now, it\u2019s because I\u2019ve <em class=\"markup--em markup--p-em\">analyzed <\/em>stuff\u200a\u2014\u200anot because I came out the chute with an intuitive understanding of beauty and balance.<\/strong><\/p><p id=\"3e11\" class=\"graf graf--p graf-after--p\">This article is not theory. This article is pure application. You won\u2019t see anything about golden ratios. I don\u2019t even mention color theory. Only what I\u2019ve learned from being bad and then <a href=\"http:\/\/calnewport.com\/blog\/2010\/01\/06\/the-grandmaster-in-the-corner-office-what-the-study-of-chess-experts-teaches-us-about-building-a-remarkable-life\/\" data-href=\"http:\/\/calnewport.com\/blog\/2010\/01\/06\/the-grandmaster-in-the-corner-office-what-the-study-of-chess-experts-teaches-us-about-building-a-remarkable-life\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">deliberately practicing<\/a>.<\/p><p id=\"796a\" class=\"graf graf--p graf-after--p\">Think of it this way: Judo was developed based on centuries of Japanese martial and philosophical traditions. You take judo lessons, and in addition to fighting, you\u2019ll hear a lot about energy and flow and harmony. That sort of stuff.<\/p><p id=\"2cda\" class=\"graf graf--p graf-after--p\">Krav Maga, on the other hand, was invented by some tough Jews who were fighting Nazis in the streets of 1930s Czechoslovakia. There is no <em class=\"markup--em markup--p-em\">art<\/em> to it. In Krav Maga lessons, you learn how to stab someone\u2019s eye with a pen and book it.<\/p><p id=\"bef2\" class=\"graf graf--p graf-after--p\">This is the Krav Maga of screens.<\/p><h4 id=\"8ec7\" class=\"graf graf--h4 graf-after--p\">The Rules<\/h4><p id=\"5451\" class=\"graf graf--p graf-after--h4\">Here they are:<\/p><ol class=\"postList\"><li id=\"6dd4\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Light comes from the sky<\/strong><\/li><li id=\"4481\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Black and white first<\/strong><\/li><li id=\"e8a6\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Double your whitespace<\/strong><\/li><li id=\"8484\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Learn the methods of overlaying text on images <\/strong>(see <a href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#e969\" data-href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#e969\" class=\"markup--anchor markup--li-anchor\" target=\"_blank\">Part 2<\/a>)<\/li><li id=\"4428\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Make text pop\u200a\u2014\u200aand un-pop <\/strong>(see <a href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#9b87\" data-href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#9b87\" class=\"markup--anchor markup--li-anchor\" target=\"_blank\">Part 2<\/a>)<\/li><li id=\"0c37\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Only use good fonts <\/strong>(see <a href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#6ce2\" data-href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#6ce2\" class=\"markup--anchor markup--li-anchor\" target=\"_blank\">Part 2<\/a>)<\/li><li id=\"6650\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Steal like an artist <\/strong>(see <a href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#ff26\" data-href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96#ff26\" class=\"markup--anchor markup--li-anchor\" target=\"_blank\">Part 2<\/a>)<\/li><\/ol><p id=\"f93a\" class=\"graf graf--p graf-after--li\">Let\u2019s get to it.<\/p><h3 id=\"b2a5\" class=\"graf graf--h3 graf-after--p\">Rule 1: Light Comes From the\u00a0Sky<\/h3><p id=\"6795\" class=\"graf graf--p graf-after--h3\"><em class=\"markup--em markup--p-em\">Shadows are invaluable cues in telling the human brain what user interface elements we\u2019re looking at.<\/em><\/p><p id=\"b5a7\" class=\"graf graf--p graf-after--p\">This is perhaps the <strong class=\"markup--strong markup--p-strong\">most important non-obvious thing <\/strong>to learn about UI design: <em class=\"markup--em markup--p-em\">light comes from the sky. <\/em>Light comes from the sky so frequently and consistently that for it to come from below actually looks <em class=\"markup--em markup--p-em\">freaky.<\/em><\/p><figure id=\"1614\" class=\"graf graf--figure graf--layoutInsetLeft graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*eFJGYuA67SIzu9pB1MZFKQ.jpeg\" data-width=\"600\" data-height=\"450\" data-action=\"zoom\" data-action-value=\"1*eFJGYuA67SIzu9pB1MZFKQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/800\/1*eFJGYuA67SIzu9pB1MZFKQ.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">WoooOOOooo<\/figcaption><\/div><\/figure><p id=\"c173\" class=\"graf graf--p graf-after--figure\">When light comes from the sky, it illuminates the tops of things and casts shadows below them. The tops of stuff are lighter, the bottoms are darker.<\/p><p id=\"8d36\" class=\"graf graf--p graf-after--p\">You wouldn\u2019t <em class=\"markup--em markup--p-em\">think <\/em>of people\u2019s lower eyelids as being particularly shaded, but shine some light on those suckers and all of a sudden it\u2019s demon girl at your front door.<\/p><p id=\"2f09\" class=\"graf graf--p graf-after--p\">Well, the same is true for UI. Just as we have little shadows on all the undersides of all our facial features, there are shadows on the undersides of <em class=\"markup--em markup--p-em\">tons<\/em> of UI elements. <strong class=\"markup--strong markup--p-strong\">Our screens are flat, but we\u2019ve invested a great amount of art into making so many elements on them <em class=\"markup--em markup--p-em\">appear be 3-D<\/em><\/strong>.<\/p><figure id=\"0b2d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*FbusHycLlFPn1Hv1N88fzg.png\" data-width=\"478\" data-height=\"490\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*FbusHycLlFPn1Hv1N88fzg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">My favorite part of this image is the poker finger in the lower-right.<\/figcaption><\/div><\/figure><p id=\"5c22\" class=\"graf graf--p graf-after--figure\">Take buttons. Even with this relatively \u201cflat\u201d button, there are still a handful of light-related details:<\/p><ol class=\"postList\"><li id=\"05a3\" class=\"graf graf--li graf-after--p\">The unpushed button (top) has a <strong class=\"markup--strong markup--li-strong\">darkened bottom edge<\/strong>. Sun don\u2019t shine there, son.<\/li><li id=\"2baf\" class=\"graf graf--li graf-after--li\">The unpushed button is <strong class=\"markup--strong markup--li-strong\">slightly brighter at the top<\/strong> than at the bottom. This is because it imitates a slightly curved surface. Just as how you\u2019d need to tilt a mirror held in front of you up to see the sun in it, surfaces that are tilted up reflect a <em class=\"markup--em markup--li-em\">biiiiit<\/em> more of the sun\u2019s light towards you.<\/li><li id=\"e1d6\" class=\"graf graf--li graf-after--li\">The unpushed button casts a <strong class=\"markup--strong markup--li-strong\">subtle shadow <\/strong>\u2013 perhaps easier to see in the magnified section.<\/li><li id=\"d196\" class=\"graf graf--li graf-after--li\">The pushed button, while still darker at the bottom than at the top, is <strong class=\"markup--strong markup--li-strong\">overall darker <\/strong>\u2013 this is because it\u2019s at the plane of the screen and the sun can\u2019t hit it as easily. One could argue that all the pushed buttons we see in real life are darker too, because our hands are blocking the light.<\/li><\/ol><p id=\"5fee\" class=\"graf graf--p graf-after--li\">That was just a button, and yet there are these 4 little light effects present. That\u2019s the lesson here. Now we just apply it to <em class=\"markup--em markup--p-em\">everything.<\/em><\/p><figure id=\"b183\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*4FCAIgmJa8BuildjlnsDeA.png\" data-width=\"993\" data-height=\"420\" data-action=\"zoom\" data-action-value=\"1*4FCAIgmJa8BuildjlnsDeA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*4FCAIgmJa8BuildjlnsDeA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">iOS 6 is old news, but it makes a good case study in light behavior.<\/figcaption><\/div><\/figure><p id=\"9d24\" class=\"graf graf--p graf-after--figure\">Here is a pair of iOS 6 settings\u200a\u2014\u200a\u201cDo Not Disturb\u201d and \u201cNotifications\u201d. NBD, right? But look how many light effects are going on with them.<\/p><ul class=\"postList\"><li id=\"e95a\" class=\"graf graf--li graf-after--p\">The top lip of the inset control panel casts a small shadow<\/li><li id=\"0639\" class=\"graf graf--li graf-after--li\">The \u201cON\u201d slider track is also immediately set in a bit<\/li><li id=\"cafd\" class=\"graf graf--li graf-after--li\">The \u201cON\u201d slider track is concave and the bottom reflects more light<\/li><li id=\"7713\" class=\"graf graf--li graf-after--li\">The icons are set <em class=\"markup--em markup--li-em\">out <\/em>a bit. See the bright border around the top of them? This represents a surface perpendicular to the light source, hence receiving a lot of light, hence bouncing a lot of light into your eyes.<\/li><li id=\"ef9e\" class=\"graf graf--li graf-after--li\">The divider notch is shadowed where angled away from the sun and vice versa<\/li><\/ul><figure id=\"9b5a\" class=\"graf graf--figure graf-after--li\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*gWuSN3QN9dSeVwSP2LZVow.png\" data-width=\"903\" data-height=\"318\" data-action=\"zoom\" data-action-value=\"1*gWuSN3QN9dSeVwSP2LZVow.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*gWuSN3QN9dSeVwSP2LZVow.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><em class=\"markup--em markup--figure-em\">A close-up of a divider notch, from an old concept of\u00a0mine.<\/em><\/figcaption><\/div><\/figure><p id=\"123d\" class=\"graf graf--p graf-after--figure\">Elements that are generally <strong class=\"markup--strong markup--p-strong\">inset<\/strong>:<\/p><ul class=\"postList\"><li id=\"c505\" class=\"graf graf--li graf-after--p\">Text input fields<\/li><li id=\"465a\" class=\"graf graf--li graf-after--li\">Pressed buttons<\/li><li id=\"d867\" class=\"graf graf--li graf-after--li\">Slider tracks<\/li><li id=\"f094\" class=\"graf graf--li graf-after--li\">Radio button (unselected)<\/li><li id=\"3be1\" class=\"graf graf--li graf-after--li\">Checkboxes<\/li><\/ul><p id=\"2d7e\" class=\"graf graf--p graf-after--li\">Elements that are generally <strong class=\"markup--strong markup--p-strong\">outset<\/strong>:<\/p><ul class=\"postList\"><li id=\"2ce6\" class=\"graf graf--li graf-after--p\">Buttons (unpressed)<\/li><li id=\"a3a4\" class=\"graf graf--li graf-after--li\">Slider buttons<\/li><li id=\"2a56\" class=\"graf graf--li graf-after--li\">Dropdown controls<\/li><li id=\"2c0f\" class=\"graf graf--li graf-after--li\">Cards<\/li><li id=\"833a\" class=\"graf graf--li graf-after--li\">The <em class=\"markup--em markup--li-em\">button <\/em>part of a selected radio button<\/li><li id=\"c6b1\" class=\"graf graf--li graf-after--li\">Popups<\/li><\/ul><p id=\"a0cb\" class=\"graf graf--p graf-after--li\">Now that you know, you\u2019ll notice it everywhere. You\u2019re welcome, kid.<\/p><h4 id=\"9536\" class=\"graf graf--h4 graf-after--p\">Wait, what about flat design,\u00a0Erik?<\/h4><p id=\"3401\" class=\"graf graf--p graf-after--h4\">We\u2019re 5 years into a trend of \u201cflat design\u201d, which is, well, <em class=\"markup--em markup--p-em\">exactly what it sounds like<\/em>. Flat design is a visual style in which the <strong class=\"markup--strong markup--p-strong\">elements <em class=\"markup--em markup--p-em\">lack <\/em>simulated indentations or protrusions. <\/strong>They\u2019re just lines and shapes of solid color.<\/p><figure id=\"ea89\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*LlZbvia9k5Qc74h9Z55dBQ.png\" data-width=\"478\" data-height=\"490\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*LlZbvia9k5Qc74h9Z55dBQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"6ac5\" class=\"graf graf--p graf-after--figure\">I love <em class=\"markup--em markup--p-em\">clean and simple<\/em> as much as the next guy, but I don\u2019t think this is a long-term trend here. The subtle simulation of 3-D in our interfaces seems far too intuitive to give up entirely.<\/p><p id=\"6565\" class=\"graf graf--p graf-after--p\">Five years ago, I predicted we\u2019d see a rise of \u201cflatty design\u201d, and, at least of 2019, that\u2019s where we are \u2013 elements of flat\u2019s clean appearance, with a dusting of shadows to help us figure out what we\u2019re looking at.<\/p><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"844a\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*eEdLH0G3UsUbi8MTIE-7ow.png\" data-width=\"1020\" data-height=\"342\" data-action=\"zoom\" data-action-value=\"1*eEdLH0G3UsUbi8MTIE-7ow.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*eEdLH0G3UsUbi8MTIE-7ow.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"3a1d\" class=\"graf graf--p graf-after--figure\">We live in a golden age, here in this happy medium between difficult-to-get-right skeuomorphism and usability-challenged flat design. Flatty is still clean, still simple, but you\u2019ll have <em class=\"markup--em markup--p-em\">some<\/em> shadows and cues for what to tap\/slide\/click.<\/p><figure id=\"d510\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*gWvCSNxqNjyYaq4IF31ZhQ.png\" data-width=\"690\" data-height=\"416\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*gWvCSNxqNjyYaq4IF31ZhQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">OS X Yosemite\u2014 flatty, not\u00a0flat.<\/figcaption><\/div><\/figure><p id=\"1064\" class=\"graf graf--p graf-after--figure\">Or, another example of flatty design: Google\u2019s <a href=\"https:\/\/material.io\/design\/\" data-href=\"https:\/\/material.io\/design\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Material Design language<\/a>.<\/p><figure id=\"a04c\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*FoHxa-1Wjz9OuqOUAdsC8Q.png\" data-width=\"1520\" data-height=\"1520\" data-action=\"zoom\" data-action-value=\"1*FoHxa-1Wjz9OuqOUAdsC8Q.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*FoHxa-1Wjz9OuqOUAdsC8Q.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"3826\" class=\"graf graf--p graf-after--figure\">This is the sort of thing I see sticking around.<\/p><p id=\"6754\" class=\"graf graf--p graf-after--p\">It uses subtle real-world cues to convey information. <strong class=\"markup--strong markup--p-strong\">Key word, <em class=\"markup--em markup--p-em\">subtle<\/em><\/strong><em class=\"markup--em markup--p-em\">.<\/em><\/p><p id=\"443c\" class=\"graf graf--p graf-after--p\">You can\u2019t say it doesn\u2019t imitate the real-world, but it\u2019s also not the web of 2006. There are no textures, no sheens, and the only gradients are purely decorative.<\/p><p id=\"13d9\" class=\"graf graf--p graf-after--p\">Flatty is the way of the future, I think. Flat? Psh, just a thing of the past.<\/p><figure id=\"3bf5\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*Zqcjyz-oIqZZojyYyWVl2Q.png\" data-width=\"466\" data-height=\"307\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Zqcjyz-oIqZZojyYyWVl2Q.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">That flat design looks so hot right\u00a0now!<\/figcaption><\/div><\/figure><h3 id=\"aaea\" class=\"graf graf--h3 graf-after--figure\">Rule 2: Black and White\u00a0First<\/h3><p id=\"9c2b\" class=\"graf graf--p graf-after--h3\"><em class=\"markup--em markup--p-em\">Designing in grayscale before adding color simplifies the most complex element of visual design\u2013 and forces you to focus on spacing and laying out elements.<\/em><\/p><p id=\"d89e\" class=\"graf graf--p graf-after--p\">UX designers are really into designing \u201cmobile-first\u201d these days. That means you think about how pages and interactions work on a phone <em class=\"markup--em markup--p-em\">before<\/em> imagining them on your zillion-pixel Retina monitor.<\/p><p id=\"97b8\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">That sort of constraint is great. It clarifies thinking<\/strong>. You start with the harder problem (usable app on a teeny-weeny screen), then adopt the solution to the easier problem (usable app on a large screen).<\/p><p id=\"0b72\" class=\"graf graf--p graf-after--p\">Well here\u2019s another similar constraint: design <em class=\"markup--em markup--p-em\">black and white first<\/em>. Start with the harder problem of making the app beautiful and usable in every way, but without the aid of color. <strong class=\"markup--strong markup--p-strong\">Add color last, and even then, only with purpose<\/strong>.<\/p><figure id=\"5592\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*qheNNhQhjjwxMeJ5XGocsA.png\" data-width=\"1145\" data-height=\"750\" data-action=\"zoom\" data-action-value=\"1*qheNNhQhjjwxMeJ5XGocsA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*qheNNhQhjjwxMeJ5XGocsA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/ueno.co\/\" data-href=\"http:\/\/ueno.co\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Haraldur Thorleifsson<\/a>\u2019s grayscale wireframes look as good as lesser designer\u2019s finished\u00a0sites.<\/figcaption><\/div><\/figure><p id=\"eb37\" class=\"graf graf--p graf-after--figure\">This is a reliable and easy way to keep apps looking \u201cclean\u201d and \u201csimple\u201d. <strong class=\"markup--strong markup--p-strong\">Having too many colors in too many places is a really easy way to screw up clean\/simple<\/strong>. B&amp;WF forces you to focus on things like spacing, sizes, and layout first. And those are the primary concerns of a clean and simple design.<\/p><\/div><div class=\"section-inner sectionLayout--outsetRow\" data-paragraph-count=\"3\"><figure id=\"a840\" class=\"graf graf--figure graf--layoutOutsetRow is-partialWidth graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*YxV7C-nHHir-PSbJ4-jqhQ.png\" data-width=\"1213\" data-height=\"671\" data-action=\"zoom\" data-action-value=\"1*YxV7C-nHHir-PSbJ4-jqhQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*YxV7C-nHHir-PSbJ4-jqhQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><figure id=\"8d27\" class=\"graf graf--figure graf--layoutOutsetRowContinue is-partialWidth graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*RckBhZxKQfveClU7rwGuyg.jpeg\" data-width=\"800\" data-height=\"600\" data-action=\"zoom\" data-action-value=\"1*RckBhZxKQfveClU7rwGuyg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/800\/1*RckBhZxKQfveClU7rwGuyg.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><figure id=\"7bb0\" class=\"graf graf--figure graf--layoutOutsetRowContinue is-partialWidth graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*EnbssykGOuXeXMV3AQFyjw.png\" data-width=\"400\" data-height=\"300\" data-action=\"zoom\" data-action-value=\"1*EnbssykGOuXeXMV3AQFyjw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/800\/1*EnbssykGOuXeXMV3AQFyjw.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Classy grayscale.<\/figcaption><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"b24a\" class=\"graf graf--p graf-after--figure\">There are some cases where B&amp;WF isn\u2019t as useful. Designs that have a strong specific attitude\u2014 \u201csporty\u201d, \u201cflashy\u201d, \u201ccartoony\u201d, etc.\u200a\u2014\u200aneed a designer who can use color extremely well. But<strong class=\"markup--strong markup--p-strong\"> most apps don\u2019t have a specific attitude except <em class=\"markup--em markup--p-em\">clean and simple<\/em><\/strong>. Those that do are admittedly much harder to design.<\/p><\/div><div class=\"section-inner sectionLayout--outsetRow\" data-paragraph-count=\"2\"><figure id=\"9b9c\" class=\"graf graf--figure graf--layoutOutsetRow is-partialWidth graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*OraO1vxtkxYteZyE4CXrOQ.png\" data-width=\"544\" data-height=\"598\" data-action=\"zoom\" data-action-value=\"1*OraO1vxtkxYteZyE4CXrOQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*OraO1vxtkxYteZyE4CXrOQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><figure id=\"5ac2\" class=\"graf graf--figure graf--layoutOutsetRowContinue is-partialWidth graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*JsbQFaIY6g697PMeEuMwvA.png\" data-width=\"873\" data-height=\"679\" data-action=\"zoom\" data-action-value=\"1*JsbQFaIY6g697PMeEuMwvA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*JsbQFaIY6g697PMeEuMwvA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Flashy and vibrant designs by <a href=\"https:\/\/twitter.com\/JulienRenvoye\" data-href=\"https:\/\/twitter.com\/JulienRenvoye\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Julien Renvoye<\/a> (left) and <a href=\"https:\/\/twitter.com\/CosminCapitanu\" data-href=\"https:\/\/twitter.com\/CosminCapitanu\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Cosmin Capitanu<\/a> (right). Harder than it\u00a0looks.<\/figcaption><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"5d6b\" class=\"graf graf--p graf-after--figure\">For all the rest, there\u2019s B&amp;WF.<\/p><h4 id=\"2e15\" class=\"graf graf--h4 graf-after--p\">Step 2: How to add\u00a0color<\/h4><p id=\"85cb\" class=\"graf graf--p graf-after--h4\">The simplest color to add is one color.<\/p><figure id=\"e15f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*YxV7C-nHHir-PSbJ4-jqhQ.png\" data-width=\"1213\" data-height=\"671\" data-action=\"zoom\" data-action-value=\"1*YxV7C-nHHir-PSbJ4-jqhQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*YxV7C-nHHir-PSbJ4-jqhQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"833d\" class=\"graf graf--p graf-after--figure\">Adding one color to a grayscale site draws the eye simply and effectively.<\/p><figure id=\"0af4\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*pds21170RP-6ZIkuSxgI2Q.png\" data-width=\"882\" data-height=\"723\" data-action=\"zoom\" data-action-value=\"1*pds21170RP-6ZIkuSxgI2Q.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*pds21170RP-6ZIkuSxgI2Q.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"2e8c\" class=\"graf graf--p graf-after--figure\">You can also take it one step up. Grayscale + <em class=\"markup--em markup--p-em\">two<\/em> colors, or grayscale + multiple colors of a single hue.<\/p><blockquote id=\"1a43\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">Color codes in practice\u200a\u2014\u200ai.e. wait, what\u2019s a hue?<\/strong><\/div><\/blockquote><blockquote id=\"6b93\" class=\"graf graf--blockquote graf-after--blockquote\"><div>The web by and large talks about color as RGB hex codes. It\u2019s most useful to ignore those. RGB is not a good framework for coloring designs. Much more useful is <a href=\"https:\/\/learnui.design\/blog\/the-hsb-color-system-practicioners-primer.html\" data-href=\"https:\/\/learnui.design\/blog\/the-hsb-color-system-practicioners-primer.html\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">HSB<\/a> (which is synonymous with HSV, and similar to HSL).<\/div><\/blockquote><blockquote id=\"9712\" class=\"graf graf--blockquote graf-after--blockquote\"><div>HSB is better than RGB because it fits with the way we think about color naturally, and you can predict how changes to the HSB values will affect the color you\u2019re looking at.<\/div><\/blockquote><blockquote id=\"182c\" class=\"graf graf--blockquote graf-after--blockquote\"><div>If this is news to you, here\u2019s <a href=\"https:\/\/learnui.design\/blog\/the-hsb-color-system-practicioners-primer.html\" data-href=\"https:\/\/learnui.design\/blog\/the-hsb-color-system-practicioners-primer.html\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">a good primer on HSB colors<\/a>.<\/div><\/blockquote><figure id=\"8588\" class=\"graf graf--figure graf-after--blockquote\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*tZRxO2DReDduBqOwgqd_yw.jpeg\" data-width=\"500\" data-height=\"150\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*tZRxO2DReDduBqOwgqd_yw.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Single-hue gold theme from <a href=\"http:\/\/www.smashingmagazine.com\/2010\/02\/08\/color-theory-for-designer-part-3-creating-your-own-color-palettes\/\" data-href=\"http:\/\/www.smashingmagazine.com\/2010\/02\/08\/color-theory-for-designer-part-3-creating-your-own-color-palettes\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Smashing Magazine<\/a>.<\/figcaption><\/div><\/figure><figure id=\"0a35\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*-rbrbh20EHL_Ue_IDxl_0A.jpeg\" data-width=\"500\" data-height=\"150\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*-rbrbh20EHL_Ue_IDxl_0A.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Single-hue blue theme from <a href=\"http:\/\/www.smashingmagazine.com\/2010\/02\/08\/color-theory-for-designer-part-3-creating-your-own-color-palettes\/\" data-href=\"http:\/\/www.smashingmagazine.com\/2010\/02\/08\/color-theory-for-designer-part-3-creating-your-own-color-palettes\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Smashing Magazine<\/a>.<\/figcaption><\/div><\/figure><p id=\"ca30\" class=\"graf graf--p graf-after--figure\">By modifying the <strong class=\"markup--strong markup--p-strong\">saturation<\/strong> and <strong class=\"markup--strong markup--p-strong\">brightness<\/strong> of a single hue, you can generate multiple colors\u2014 darks, lights, backgrounds, accents, eye-catchers\u2014 but it\u2019s not overwhelming on the eye.<\/p><p id=\"c7fc\" class=\"graf graf--p graf-after--p\">Using multiple colors from one or two base hues is the <strong class=\"markup--strong markup--p-strong\">most reliable way to accentuate and neutralize elements without making the design messy<\/strong>.<\/p><figure id=\"82bf\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*_fM8VVYx7hMgdJ_Wy24AXg.png\" data-width=\"400\" data-height=\"300\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*_fM8VVYx7hMgdJ_Wy24AXg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Countdown timer by <a href=\"http:\/\/kerem.co\/\" data-href=\"http:\/\/kerem.co\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Kerem\u00a0Suer<\/a><\/figcaption><\/div><\/figure><p id=\"4fe5\" class=\"graf graf--p graf-after--figure\">In my experience, adjusting your theme color to fit the use-cases your UI calls for is the <em class=\"markup--em markup--p-em\">most important skill <\/em>in using color, and I\u2019ve written a practical guide to it <a href=\"https:\/\/learnui.design\/blog\/color-in-ui-design-a-practical-framework.html\" data-href=\"https:\/\/learnui.design\/blog\/color-in-ui-design-a-practical-framework.html\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><h4 id=\"acb9\" class=\"graf graf--h4 graf-after--p\">A few other notes on\u00a0color<\/h4><p id=\"fbd6\" class=\"graf graf--p graf-after--h4\">Color is the most complicated area of visual design. And while a lot of stuff on color is obtuse and not practical for finishing the design in front of you, I\u2019ve seen some really good stuff out there.<\/p><p id=\"bf5e\" class=\"graf graf--p graf-after--p\">A small toolbox:<\/p><ul class=\"postList\"><li id=\"c15a\" class=\"graf graf--li graf-after--p\"><a href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" data-href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">Learn UI Design<\/strong><\/a>. Shameless plug: this is a course I\u2019ve created, and it contains 3 <em class=\"markup--em markup--li-em\">hours<\/em> of video about designing with color (and over 20 hours of total videos on UI design topics). Check it out at <a href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" data-href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\"><em class=\"markup--em markup--li-em\">learnui.design<\/em><\/a>.<\/li><li id=\"a57c\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/learnui.design\/blog\/color-in-ui-design-a-practical-framework.html\" data-href=\"https:\/\/learnui.design\/blog\/color-in-ui-design-a-practical-framework.html\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">Color in UI Design: A (Practical) Framework<\/strong><\/a>. If you liked this section, but want to hear more about <em class=\"markup--em markup--li-em\">color<\/em> (as opposed to just black and white), this is your article. And guess who wrote it!<\/li><li id=\"0a92\" class=\"graf graf--li graf-after--li\"><a href=\"http:\/\/ianstormtaylor.com\/design-tip-never-use-black\/\" data-href=\"http:\/\/ianstormtaylor.com\/design-tip-never-use-black\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">Never Use Black<\/strong><\/a> (Ian Storm Taylor). Talks about how totally flat grays almost never appear in the real-world, and how saturating your shades of gray\u2013 especially your darker shades\u2013 adds a visual richness to your designs. Plus, saturated grays more closely mimic the real-world, which is its own virtue.<\/li><li id=\"72cd\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/color.adobe.com\" data-href=\"https:\/\/color.adobe.com\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">Adobe Color CC<\/strong><\/a>. An awesome tool for finding, modifying, and creating color schemes.<\/li><li id=\"d50b\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/dribbble.com\/colors\/BADA55\" data-href=\"https:\/\/dribbble.com\/colors\/BADA55\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">Dribbble search-by-color<\/strong><\/a><strong class=\"markup--strong markup--li-strong\">.<\/strong> Another awesome way to find what works with a particular color. Talk about practical. If you already have one color decided, come look at what the world\u2019s best designers are doing\/matching with that color.<\/li><\/ul><h3 id=\"b7c4\" class=\"graf graf--h3 graf-after--li\">Rule 3: Double your whitespace<\/h3><p id=\"7eee\" class=\"graf graf--p graf-after--h3\"><em class=\"markup--em markup--p-em\">To make UI that looks designed, add a lot of breathing room.<\/em><\/p><p id=\"5b7a\" class=\"graf graf--p graf-after--p\">In Rule 2, I said that B&amp;WF forces designers to think about <em class=\"markup--em markup--p-em\">spacing and layout<\/em> before considering color, and how that\u2019s a good thing. Well, it\u2019s time we talk about spacing and layout.<\/p><p id=\"a29e\" class=\"graf graf--p graf-after--p\">If you\u2019ve coded HTML from scratch, you\u2019re probably familiar with the way HTML is, by default, laid out on the page.<\/p><figure id=\"59ff\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*fS6ixQIk88MJlEmph7PeJA.png\" data-width=\"891\" data-height=\"788\" data-action=\"zoom\" data-action-value=\"1*fS6ixQIk88MJlEmph7PeJA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*fS6ixQIk88MJlEmph7PeJA.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"abcb\" class=\"graf graf--p graf-after--figure\">Basically, everything is smashed towards the top of the screen. The fonts are small; there\u2019s absolutely no space between lines. There\u2019s a <em class=\"markup--em markup--p-em\">biiit<\/em> of space between paragraphs, but it isn\u2019t much. The paragraphs just stretch on to the end of the page, whether that\u2019s 100 px or 10,000 px.<\/p><p id=\"db73\" class=\"graf graf--p graf-after--p\">Aesthetically speaking, that\u2019s <em class=\"markup--em markup--p-em\">awful<\/em>. <strong class=\"markup--strong markup--p-strong\">If you want to make UI that looks <em class=\"markup--em markup--p-em\">designed<\/em>, you need to add in a lot of breathing room<\/strong>.<\/p><p id=\"779b\" class=\"graf graf--p graf-after--p\">Sometimes a ridiculous amount.<\/p><blockquote id=\"5f00\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">Whitespace, HTML, and CSS<\/strong><\/div><\/blockquote><blockquote id=\"6c0b\" class=\"graf graf--blockquote graf-after--blockquote\"><div>If you, like me, are used to formatting with CSS, where the <strong class=\"markup--strong markup--blockquote-strong\">default is no whitespace<\/strong>, it\u2019s time to untrain yourself of those bad habits. Start thinking of whitespace as the default\u2014 everything starts as whitespace, until you take it away by adding a page element.<\/div><\/blockquote><blockquote id=\"3a5c\" class=\"graf graf--blockquote graf-after--blockquote\"><div>Sound zen? I think it\u2019s a big reason people still sketch this stuff out on paper.<\/div><\/blockquote><blockquote id=\"2059\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">Starting with a blank page means starting with nothing but whitespace<\/strong>. You think of margins and spacing right from the very beginning. Everything you draw is a conscious whitespace-removing decision.<\/div><\/blockquote><blockquote id=\"427b\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">Starting with a bunch of unstyled HTML means starting with content<\/strong>. Spacing is the afterthought. It has to be explicitly stated.<\/div><\/blockquote><p id=\"a5ff\" class=\"graf graf--p graf-after--blockquote\">Here\u2019s an illustrative music player concept by <a href=\"http:\/\/www.piotrkwiatkowski.co.uk\/\" data-href=\"http:\/\/www.piotrkwiatkowski.co.uk\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Piotr Kwiatkowski<\/a>.<\/p><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"be66\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*qFwXZ_05pRv2OtiaJHIp6Q.jpeg\" data-width=\"1600\" data-height=\"1200\" data-action=\"zoom\" data-action-value=\"1*qFwXZ_05pRv2OtiaJHIp6Q.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*qFwXZ_05pRv2OtiaJHIp6Q.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"941e\" class=\"graf graf--p graf-after--figure\">Pay particular attention to the menu on the left.<\/p><figure id=\"22e5\" class=\"graf graf--figure graf--layoutInsetLeft graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*jSC64LYfVYlMHaI_B7xfKQ.png\" data-width=\"330\" data-height=\"558\" src=\"https:\/\/cdn-images-1.medium.com\/max\/800\/1*jSC64LYfVYlMHaI_B7xfKQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Left menu<\/figcaption><\/div><\/figure><p id=\"749a\" class=\"graf graf--p graf-after--figure\">The vertical space between the menu items is fully <em class=\"markup--em markup--p-em\">twice<\/em><strong class=\"markup--strong markup--p-strong\"> <\/strong>the height of the text itself. You\u2019re looking at 12px font with just as much padding above and below it.<\/p><p id=\"e215\" class=\"graf graf--p graf-after--p graf--trailing\">Or take a look at the list titles. <strong class=\"markup--strong markup--p-strong\">There\u2019s a 15px space between the word \u201cPLAYLISTS\u201d and its own underline. That\u2019s more than the <\/strong><a href=\"http:\/\/en.wikipedia.org\/wiki\/Cap_height\" data-href=\"http:\/\/en.wikipedia.org\/wiki\/Cap_height\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">cap height<\/strong><\/a><strong class=\"markup--strong markup--p-strong\"> of the font itself!<\/strong> And that\u2019s to say nothing of the 25 pixels between the lists.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><figure id=\"7a6a\" class=\"graf graf--figure graf--layoutInsetLeft graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*43qoikq5esyOer2PpETX_Q.png\" data-width=\"333\" data-height=\"482\" src=\"https:\/\/cdn-images-1.medium.com\/max\/800\/1*43qoikq5esyOer2PpETX_Q.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><em class=\"markup--em markup--figure-em\">More space in the top nav bar. The text \u201cSearch all music\u201d is 20% of the height of the bar. The icons are similarly proportioned.<\/em><\/figcaption><\/div><\/figure><p id=\"5fb6\" class=\"graf graf--p graf-after--figure graf--leading\">The left sidebar shows generous spacing in between lines of text, and more.<\/p><p id=\"d287\" class=\"graf graf--p graf-after--p graf--trailing\">Piotr was conscientious about putting in extra space here, and it paid off. While this is just a concept he put together for the fun of it (as far as I know), as far as aesthetics go, it\u2019s beautiful enough to compete with the best music UIs out there.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"1cde\" class=\"graf graf--p graf--leading\">Good, generous whitespace can make some of the messiest interfaces look inviting and simple\u2014 like forums.<\/p><figure id=\"6fd5\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*g6m0YZVyMEVMuLXzO512gg.png\" data-width=\"800\" data-height=\"600\" data-action=\"zoom\" data-action-value=\"1*g6m0YZVyMEVMuLXzO512gg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*g6m0YZVyMEVMuLXzO512gg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Forum design concept by <a href=\"http:\/\/sis.to\/\" data-href=\"http:\/\/sis.to\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Matt\u00a0Sisto<\/a><\/figcaption><\/div><\/figure><p id=\"98bd\" class=\"graf graf--p graf-after--figure\">Or Wikipedia.<\/p><figure id=\"4daf\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*SVtl39B-dSsHo3HFI0h4FA.png\" data-width=\"800\" data-height=\"600\" data-action=\"zoom\" data-action-value=\"1*SVtl39B-dSsHo3HFI0h4FA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*SVtl39B-dSsHo3HFI0h4FA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Wikipedia design concept by <a href=\"https:\/\/dribbble.com\/aureliensalomon\" data-href=\"https:\/\/dribbble.com\/aureliensalomon\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Aur\u00e9lien\u00a0Salomon<\/a><\/figcaption><\/div><\/figure><p id=\"bdb5\" class=\"graf graf--p graf-after--figure\">You can find plenty of argument that, say, the Wikipedia redesign leaves out key functionality to using the site. But you can\u2019t say it\u2019s not a good way to learn!<\/p><p id=\"0f01\" class=\"graf graf--p graf-after--p\">Put space between your lines.<\/p><p id=\"7af6\" class=\"graf graf--p graf-after--p\">Put space between your elements.<\/p><p id=\"ebe5\" class=\"graf graf--p graf-after--p\">Put space between your groups of elements.<\/p><p id=\"fd7a\" class=\"graf graf--p graf-after--p graf--trailing\"><strong class=\"markup--strong markup--p-strong\">Analyze what works<\/strong>.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"070d\" class=\"graf graf--p graf--leading\"><em class=\"markup--em markup--p-em\">OK, that wraps up Part 1. Thanks for sticking around!<\/em><\/p><p id=\"e0aa\" class=\"graf graf--p graf-after--p\">In <a href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96\" data-href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Part 2<\/a>, I\u2019ll be talking about the last 4 rules:<\/p><blockquote id=\"5d8f\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">4. Learn the methods of overlaying text on images<\/strong><\/div><\/blockquote><blockquote id=\"a1d7\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">5. Make text pop\u2014 and un-pop<\/strong><\/div><\/blockquote><blockquote id=\"10d8\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">6. Only use good fonts<\/strong><\/div><\/blockquote><blockquote id=\"76f8\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">7. Steal like an artist<\/strong><\/div><\/blockquote><p id=\"555f\" class=\"graf graf--p graf-after--blockquote\">If you learned something useful in Part 1, <a href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96\" data-href=\"https:\/\/medium.com\/@erikdkennedy\/7-rules-for-creating-gorgeous-ui-part-2-430de537ba96\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">read Part 2<\/a>.<\/p><h3 id=\"d377\" class=\"graf graf--h3 graf-after--p\">Still can\u2019t get enough? Introducing\u2026 Learn UI\u00a0Design<\/h3><p id=\"c41d\" class=\"graf graf--p graf-after--h3\">I\u2019ve been plugging away to create the <strong class=\"markup--strong markup--p-strong\">single most comprehensive <\/strong>online video course for learning the <strong class=\"markup--strong markup--p-strong\">practical skills of interface design<\/strong>.<\/p><figure id=\"e92a\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"http:\/\/learnui.design?UTM_SOURCE=medium&amp;UTM_MEDIUM=content&amp;UTM_CAMPAIGN=7-rules-part-1\" data-href=\"http:\/\/learnui.design?UTM_SOURCE=medium&amp;UTM_MEDIUM=content&amp;UTM_CAMPAIGN=7-rules-part-1\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"nofollow noopener\" target=\"_blank\"><img class=\"graf-image\" data-image-id=\"1*ih8R489EfuZb58WMR2x8Zg.png\" data-width=\"600\" data-height=\"300\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ih8R489EfuZb58WMR2x8Zg.png\" alt=\"image\" \/><\/a><\/div><\/div><\/figure><p id=\"847f\" class=\"graf graf--p graf-after--figure\">Learn UI Design covers all kinds of topics:<\/p><ul class=\"postList\"><li id=\"6032\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Color<\/strong>. Picking good colors, adjusting them for different purposes in your interface, developing a palette, fixing clashing colors, and more.<\/li><li id=\"406c\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Typography<\/strong>. Choosing fonts, styling text, pairing fonts, and more.<\/li><li id=\"5113\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Fundamentals of design<\/strong>. Spacing, alignment, grids, layout, lighting, shadows, and more.<\/li><li id=\"7e46\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">User interface elements.<\/strong> Form controls (like buttons and text boxes), icons, imagery (like photography and illustrations), tables and lists (hugely important for mobile app design), and more.<\/li><\/ul><p id=\"4f80\" class=\"graf graf--p graf-after--li\">Here\u2019s an idea of how much content there is. The topics you\u2019ve already heard about\u200a\u2014\u200alighting, use of gray, and spacing\u200a\u2014\u200aeach have their <em class=\"markup--em markup--p-em\">own<\/em> videos, totaling <strong class=\"markup--strong markup--p-strong\">1 hr, 52 min<\/strong>.<\/p><p id=\"7d5a\" class=\"graf graf--p graf-after--p\">There is a <em class=\"markup--em markup--p-em\">ton<\/em> of content in this course: <strong class=\"markup--strong markup--p-strong\">20+ hours of video<\/strong> in <strong class=\"markup--strong markup--p-strong\">40+ video lessons<\/strong>.<\/p><figure id=\"aedf\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" data-href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"nofollow noopener\" target=\"_blank\"><img class=\"graf-image\" data-image-id=\"1*bptkBKJtVrx5-yO0-ePJWQ.png\" data-width=\"1552\" data-height=\"985\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*bptkBKJtVrx5-yO0-ePJWQ.png\" alt=\"image\" \/><\/a><\/div><\/div><\/figure><p id=\"1f73\" class=\"graf graf--p graf-after--figure\">In almost every video, you\u2019ll <em class=\"markup--em markup--p-em\">watch me design in Sketch<\/em>. This is important. I\u2019m not teaching some theoretical framework here. Instead, everything I show you is the tips, tricks, and frameworks I use to design interfaces every day. Think of it as watching over my shoulder as I teach you everything I know.<\/p><p id=\"cdd4\" class=\"graf graf--p graf-after--p\">For example, in the \u201cGray\u201d video (yes: 27 minutes <em class=\"markup--em markup--p-em\">all about using a single color<\/em>), I start by showing you a practical technique to make your grays match your theme colors, no matter what those colors are. Then I do another demo to show off gray\u2019s versatility, explaining ways in which its <em class=\"markup--em markup--p-em\">subtlety<\/em> is actually a <em class=\"markup--em markup--p-em\">strength<\/em>. This includes some tips that took me <em class=\"markup--em markup--p-em\">years<\/em> to notice.<\/p><figure id=\"d586\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*3hmtgx-9aY4YA18fnVYUDA.png\" data-width=\"846\" data-height=\"152\" data-action=\"zoom\" data-action-value=\"1*3hmtgx-9aY4YA18fnVYUDA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*3hmtgx-9aY4YA18fnVYUDA.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"b8e0\" class=\"graf graf--p graf-after--figure\">Finally, I explain why gray is the classiest color, and used by so many luxury and fashion brands. Then I explain the homework for the lesson and wrap up. By the end, we\u2019ve done 3 live demos, explained a few key tricks for how to add gray to your designs, and looked at 10 well-designed real-world examples (including 1 designed by me).<\/p><p id=\"54d1\" class=\"graf graf--p graf-after--p\">Here\u2019s what folks are saying about Learn UI Design:<\/p><blockquote id=\"46f5\" class=\"graf graf--blockquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cLearn UI Design is like learning how to fly a plane by actually sitting in the cockpit with the pilot\u200a\u2014\u200aErik is constantly designing\/redesigning <strong class=\"markup--strong markup--blockquote-strong\">real-world examples<\/strong> right in front of you, explaining why X is good or bad, and how to go about making it even better.\u201d<br \/>\u200a\u2014\u200aMudassir Ali, Frontend Engineer, Canva<\/div><\/blockquote><blockquote id=\"5ab9\" class=\"graf graf--blockquote graf--startsWithDoubleQuote graf-after--blockquote\"><div>\u201cLearn UI Design\u2019s straightforward approach, illustrated with real-life examples and tutorials, was extremely helpful and eye-opening. I would <strong class=\"markup--strong markup--blockquote-strong\">highly recommend<\/strong> this course for UX designers wanting to add UI design to their toolkit.\u201d<br \/>\u200a\u2014\u200aSarah Kim, UX Design Lead<\/div><\/blockquote><blockquote id=\"28ab\" class=\"graf graf--blockquote graf--startsWithDoubleQuote graf-after--blockquote\"><div>\u201cErik\u2019s <strong class=\"markup--strong markup--blockquote-strong\">pragmatic approach<\/strong> to design has taught me infinitely more than what reading any design books ever did! Away with the books, and one more video please.\u201d<br \/>\u200a\u2014\u200aAnders Nysom, Freelance Developer<\/div><\/blockquote><p id=\"d8b6\" class=\"graf graf--p graf-after--blockquote\">If you\u2019re a dev, UX designer, or PM, and want to add visual design to your skillset, this course is tailor-made for you. Hop on over to <a href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" data-href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">learnui.design<\/a> for more.<\/p><figure id=\"e71b\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" data-href=\"http:\/\/learnui.design\/?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"nofollow noopener\" target=\"_blank\"><img class=\"graf-image\" data-image-id=\"1*a-FGKWDARSKKkJRF5Jtd8A.png\" data-width=\"640\" data-height=\"240\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*a-FGKWDARSKKkJRF5Jtd8A.png\" alt=\"image\" \/><\/a><\/div><\/div><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"910d\" class=\"graf graf--p graf--leading\">Oh, and finally: I have a <strong class=\"markup--strong markup--p-strong\">Design Newsletter<\/strong> to which I intentionally send out <em class=\"markup--em markup--p-em\">very few<\/em> but <em class=\"markup--em markup--p-em\">very high-value<\/em> design articles (like the one above). If you want to see stuff like this every month (or 6) in your inbox (basically all of it unpublished anywhere else), you should take a moment and <a href=\"http:\/\/learnui.design\/newsletter.html?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" data-href=\"http:\/\/learnui.design\/newsletter.html?utm_source=medium&amp;utm_medium=content&amp;utm_campaign=7-rules-part-1\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">sign up<\/a>:<\/p><figure id=\"1a3f\" class=\"graf graf--figure graf--iframe graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe data-width=\"800\" data-height=\"400\" width=\"700\" height=\"350\" src=\"https:\/\/medium.com\/media\/47085b36741abcb3a0452f7ea203871e?postId=559d4e805cda\" data-media-id=\"47085b36741abcb3a0452f7ea203871e\" data-thumbnail=\"https:\/\/i.embed.ly\/1\/image?url=https%3A%2F%2Fe.enpose.co%2F%3Fkey%3DdRXnS9Gplk%26w%3D700%26h%3D425%26url%3Dhttps%253A%252F%252Fupscri.be%252F615126-2%252F%253Fenpose&amp;key=a19fcc184b9711e1b4764040d3dc5c07\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><\/div><\/figure><\/div><\/div><\/section>","created_at":"2019-02-08T12:53:42+0000","updated_at":"2019-02-08T12:53:42+0000","published_at":"2014-11-13T21:59:07+0000","published_by":["Erik D. Kennedy"],"annotations":[],"mimetype":"text\/html","reading_time":16,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*X68DND6Lh8GFvM3i49KOhA.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":818,"title":"Shinmera\/tooter","url":"https:\/\/github.com\/Shinmera\/tooter","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><h2><a id=\"user-content-about-tooter\" class=\"anchor\" aria-hidden=\"true\" href=\"#about-tooter\"><\/a>About Tooter<\/h2>\n<p>This is a client library implementing the full v1 REST API protocol for <a href=\"https:\/\/joinmastodon.org\/\" rel=\"nofollow\">Mastodon<\/a>.<\/p>\n<h2><a id=\"user-content-how-to\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to\"><\/a>How To<\/h2>\n<p>Before you start, you'll need an account on some Mastodon instance. For the purposes of this documentation we'll simply pick <a href=\"https:\/\/mastodon.social\" rel=\"nofollow\">mastodon.social<\/a>. Once you've got Tooter loaded you'll first need to create a <code>client<\/code> instance.<\/p>\n<pre>(defvar *client* (make-instance 'tooter:client\n                   :base \"https:\/\/mastodon.social\"\n                   :name \"My Tooter Test App\"))\n<\/pre>\n<p>Once you have a client set up, you need to authorise it against your account so that you can make requests on your behalf.<\/p>\n<pre>(tooter:authorize *client*)\n<\/pre>\n<p>This will return an URL as its secondary value. Visit this URL in your browser, click on the authorize button, and copy the code it displays. Then call <code>authorize<\/code> again using this code.<\/p>\n<pre>(tooter:authorize *client* \"...\")\n<\/pre>\n<p>If everything goes well this should return a fully authorised client instance.<\/p>\n<pre>(tooter:account *client*)\n<\/pre>\n<p>From here on out you can make use of the full API Mastodon API. See the definition index below.<\/p>\n<pre>(tooter:make-status *client* \"Tooter works, woah nelly!\")\n<\/pre>\n<h2><a id=\"user-content-persisting-client-settings\" class=\"anchor\" aria-hidden=\"true\" href=\"#persisting-client-settings\"><\/a>Persisting Client Settings<\/h2>\n<p>Once you have authorised your client, you will probably want to save the information somewhere so that the user doesn't need to re-authorise every time. To do this, simply save the <code>key<\/code>, <code>secret<\/code>, and <code>access-token<\/code> values from the client instance. If those are set, the <code>authorize<\/code> steps are not needed.<\/p>\n<h2><a id=\"user-content-writing-an-application-library\" class=\"anchor\" aria-hidden=\"true\" href=\"#writing-an-application-library\"><\/a>Writing an Application Library<\/h2>\n<p>If you write an application on top of Mastodon and would like to offer this as another re-usable library, you should create a subclass of <code>client<\/code> that automatically provides the proper initargs for the <code>:name<\/code> and so forth. This ensures that your application is known under the correct name, but still lets the user specify the correct <code>:base<\/code>.<\/p>\n<\/article>","created_at":"2019-02-07T21:07:51+0000","updated_at":"2019-02-07T21:07:51+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/4833204?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["keyboard","dvorak"],"is_public":false,"id":817,"title":"learn.dvorak.nl","url":"https:\/\/learn.dvorak.nl\/","created_at":"2019-02-07T14:22:27+0000","updated_at":"2019-02-07T14:23:28+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"learn.dvorak.nl"},{"is_archived":0,"is_starred":0,"tags":["wordpress"],"is_public":false,"id":816,"title":"Key WordPress Statistics","url":"https:\/\/wordpress.org\/about\/stats\/","content":"<p>Here are some charts showing what sorts of systems people are running WordPress on. (You\u2019ll need JavaScript enabled to see them.)<\/p>\n\t\t\t\t\t<div id=\"wp_versions\" class=\"wporg-stats-chart loading\">\n\t\t\t\t\t<div id=\"php_versions\" class=\"wporg-stats-chart loading\">\n\t\t\t\t\t<div id=\"mysql_versions\" class=\"wporg-stats-chart loading\">\n\t\t\t\t\t<div id=\"locales\" class=\"wporg-stats-chart loading\">\n\t\t\t\t<\/div><\/div><\/div><\/div>","created_at":"2019-02-07T02:52:19+0000","updated_at":"2019-02-14T02:51:13+0000","published_at":"2018-03-28T03:08:37+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"wordpress.org","preview_picture":"https:\/\/s.w.org\/images\/backgrounds\/wordpress-bg-medblue.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["launchd","osx"],"is_public":false,"id":815,"title":"A launchd Tutorial","url":"http:\/\/www.launchd.info\/","content":"<div class=\"content\" id=\"page-Introduction\"><div class=\"section\"><p><a href=\"http:\/\/en.wikipedia.org\/wiki\/Launchd\">Wikipedia<\/a> defines launchd as \"a unified, open-source service management framework for starting, stopping and managing daemons, applications, processes, and scripts. Written and designed by Dave Zarzycki at Apple, it was introduced with Mac OS X Tiger and is licensed under the Apache License.\"<\/p><\/div><div class=\"section\"><p>A daemon is a program running in the background without requiring user input. A typical daemon might for instance perform daily maintenance tasks or scan a device for malware when it is connected.<\/p><\/div><div class=\"section\"><p>launchd differentiates between agents and daemons. The main difference is that an agent is run on behalf of the logged in user while a daemon runs on behalf of the root user or any user you specify with the UserName key.<\/p><\/div><div class=\"section\"><p>The behavior of a daemon\/agent is specified in a special XML file called a property list. Depending on where it is stored it will be treated as a daemon or an agent.<\/p><p>Job definitions crucial for the operation of the operating system are stored below \/System\/Library. You should never need to create a daemon or agent in these directories. Third-Party definitions which are relevant for every user are stored below \/Library. Job definitions for a specific user are stored below the respective user's Library directory.<\/p><table><tr><th>Type<\/th>\n\t\t\t\t\t\t<th>Location<\/th>\n\t\t\t\t\t\t<th>Run on behalf of<\/th>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>User Agents<\/td>\n\t\t\t\t\t\t<td>~\/Library\/LaunchAgents<\/td>\n\t\t\t\t\t\t<td>Currently logged in user<\/td>\n\t\t\t\t\t<\/tr><tr class=\"odd\"><td>Global Agents<\/td>\n\t\t\t\t\t\t<td>\/Library\/LaunchAgents<\/td>\n\t\t\t\t\t\t<td>Currently logged in user<\/td>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Global Daemons<\/td>\n\t\t\t\t\t\t<td>\/Library\/LaunchDaemons<\/td>\n\t\t\t\t\t\t<td>root or the user specified with the key UserName<\/td>\n\t\t\t\t\t<\/tr><tr class=\"odd\"><td>System Agents<\/td>\n\t\t\t\t\t\t<td>\/System\/Library\/LaunchAgents<\/td>\n\t\t\t\t\t\t<td>Currently logged in user<\/td>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>System Daemons<\/td>\n\t\t\t\t\t\t<td>\/System\/Library\/LaunchDaemons<\/td>\n\t\t\t\t\t\t<td>root or the user specified with the key UserName<\/td>\n\t\t\t\t\t<\/tr><\/table><p class=\"note\">The ~ character stands for the users home directory.<\/p><p>The following example shows a complete job definition with only three keys:<\/p><ul><li>Label This key is required for every job definition. It identifies the job and has to be unique for the launchd instance. Theoretically it is possible for an agent to have the same label as a daemon, as daemons are loaded by the root launchd whereas agents are loaded by a user launchd, but it is not recommended.<\/li>\n\t\t\t\t\t<li>Program This key defines <em>what<\/em> to start, in this case a shell script \/Users\/Me\/Scripts\/cleanup.sh.<\/li>\n\t\t\t\t\t<li>RunAtLoad This is one of several optional keys specifying <em>when<\/em> the job should be run, in this case right after it has been loaded.<\/li>\n\t\t\t\t<\/ul><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;\n&lt;!DOCTYPE plist PUBLIC \"-\/\/Apple\/\/DTD PLIST 1.0\/\/EN\" \"http:\/\/www.apple.com\/DTDs\/PropertyList-1.0.dtd\"&gt;\n&lt;plist version=\"1.0\"&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Label&lt;\/key&gt;\n\t\t&lt;string&gt;com.example.app&lt;\/string&gt;\n\t\t&lt;key&gt;Program&lt;\/key&gt;\n\t\t&lt;string&gt;\/Users\/Me\/Scripts\/cleanup.sh&lt;\/string&gt;\n\t\t&lt;key&gt;RunAtLoad&lt;\/key&gt;\n\t\t&lt;true\/&gt;\n\t&lt;\/dict&gt;\n&lt;\/plist&gt;<\/p><\/div><\/div><p>launchd supports more than 36 different configuration keys. Most of them are explained in the Configuration-Section of this primer.<\/p><\/div><div class=\"section\"><p>Upon system start the root launchd process will scan the daemon directories \/System\/Library\/LaunchDaemons and \/Library\/LaunchDaemons for job definitions and load them depending on the existence\/value of the Disabled key and the contents of the <a target=\"The_override_database\">override database<\/a>.<\/p><p>When a user logs in a new launchd process will be started for this user. This launchd process will scan the agent directories \/System\/Library\/LaunchAgents,  \/Library\/LaunchAgents and ~\/Library\/LaunchAgents for job definitions and load them depending on the existence\/value of the Disabled key and the contents of the <a target=\"The_override_database\">override database<\/a>.<\/p><\/div><div class=\"section\"><p>Loading a job definition does not necessarily mean to start the job. When a job is started is determined by the job definition. In fact, only when RunAtLoad or KeepAlive have been specified,launchd will start the job unconditionally when it has been loaded.<\/p><\/div><\/div><div class=\"content\" id=\"page-Configuration\"><p>Content<\/p><div class=\"section\"><p>Configuring is done using a text editor of your choice or using a 3rd-Party tool like <a href=\"http:\/\/www.soma-zone.com\/LaunchControl\">LaunchControl<\/a>. While it is not difficult to create or edit job definitions using a text editor this method is prone to errors and finding a problem in your configuration can be a daunting task. <a href=\"http:\/\/www.soma-zone.com\/LaunchControl\">LaunchControl<\/a> will make sure you always create valid job definitions and highlight errors even before you load the job.<\/p><p>All examples provided in this tutorial are shown as plain text along with the corresponding configuration section in <a href=\"http:\/\/www.soma-zone.com\/LaunchControl\">LaunchControl<\/a>.<\/p><p class=\"note\">When using a text editor, make sure to save your configuration as plain text.<\/p><\/div><div class=\"section\"><p>This key is required for every job definition. It identifies the job and has to be unique for the launchd instance. Theoretically it is possible for an agent to have the same label as a daemon, as daemons are loaded by the root launchd whereas agents are loaded by a user launchd, but it is not recommended.<\/p><p>By convention job names are written in <a href=\"http:\/\/en.wikipedia.org\/wiki\/Reverse_domain_name_notation\">reverse domain notation<\/a>. The helper process for our program LaunchControl, for instance, is named com.soma-zone.LaunchControl.Helper. For private agents the domain local is a good choice: local.cleanup<\/p><\/div><div class=\"section\"><p>A valid job definition requires at least one of these keys: Program and ProgramArguments. The first one takes a single string as its value, the latter one requires a list of strings.<\/p><\/div><div class=\"section\"><h2>Program<\/h2><p>This is as simple as it gets. Just specify the complete path to your executable:<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;Program&lt;\/key&gt;\n&lt;string&gt;\/path\/to\/program&lt;\/string&gt;<\/p><\/div><\/div><p class=\"note\">An OSX application is not an executable, just a directory. Use <a target=\"Automatically_restart_application\">this<\/a> recipe to start an application.<\/p><\/div><div class=\"section\"><h2>ProgramArguments<\/h2><p>Use this one if your executable requires command line options:<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;ProgramArguments&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;string&gt;\/usr\/bin\/rsync&lt;\/string&gt;\n\t&lt;string&gt;--archive&lt;\/string&gt;\n\t&lt;string&gt;--compress-level=9&lt;\/string&gt;\n\t&lt;string&gt;\/Volumes\/Macintosh HD&lt;\/string&gt;\n\t&lt;string&gt;\/Volumes\/Backup&lt;\/string&gt;\n&lt;\/array&gt;<\/p><\/div><\/div><p>This is equivalent to the shell command<\/p><p>\/usr\/bin\/rsync --archive --compress-level=9 \"\/Volumes\/Macintosh HD\" \"\/Volumes\/Backup\"<\/p><p class=\"note\">The first string in the list is <em>not<\/em> the first argument but the path to the executable you want to run.<\/p><\/div><div class=\"section\"><h2>Program and ProgramArguments<\/h2><p>When you provide both keys, the value of Program is the executable to be started.  The first string in ProgramArguments will show up as argv[0] in the executable. If this makes no sense to you, just think that launchd will ignore it.<\/p><\/div><div class=\"section\"><p>Use the keys in this section to customize the environment the program runs in. Set environment variables, the working directory, standard in- and output, enable core file creation or limit the CPU time the executable gets.<\/p><\/div><div class=\"section\"><h2>Setting environment variables: EnvironmentVariables<\/h2><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;EnvironmentVariables&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;PATH&lt;\/key&gt;\n\t&lt;string&gt;\/bin:\/usr\/bin:\/usr\/local\/bin&lt;\/string&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><p class=\"note\">Shell globbing and variable expansion do not work for this key.<\/p><p class=\"note\">You can set environment variables per launchd instance by issuing the launchctl command setenv. Logging out (in case of a user launchd) or rebooting (in case of the root launchd) will revert those changes. To make these changes persistent you'll have to placing this command in \/etc\/launchd-user.conf (in case of a user launchd) or \/etc\/launchd.conf (for all launchd processes).<\/p><\/div><div class=\"section\"><h2>Redirecting input and output:\n\t\t\t\t\tStandardInPath\n\t\t\t\t\tStandardOutPath\n\t\t\t\t\tStandardErrorPath\n\t\t\t\t<\/h2><p>All these keys take a string as argument. Provide absolute paths when possible. Relative paths are interpreted relative to RootDirectory or \/.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;StandardInPath&lt;\/key&gt;\n&lt;string&gt;\/tmp\/test.stdin&lt;\/string&gt;\n&lt;key&gt;StandardOutPath&lt;\/key&gt;\n&lt;string&gt;\/tmp\/test.stdout&lt;\/string&gt;\n&lt;key&gt;StandardErrorPath&lt;\/key&gt;\n&lt;string&gt;\/tmp\/test.stderr&lt;\/string&gt;<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>These input fields support auto-completion by pressing Esc. Clicking on the <img class=\"icon\" src=\"http:\/\/www.launchd.info\/img\/NSRevealFreestandingTemplate.png\" alt=\"image\" \/> button lets you select the file. The <img class=\"icon\" src=\"http:\/\/www.launchd.info\/img\/ConsoleTemplate.png\" alt=\"image\" \/> button opens the selected file in Console.app.<\/p><\/div><\/div><p class=\"note\">The keys StandardOutPath and StandardErrorPath are very important when it comes to debugging a job. Always configure these keys should a job not behave as expected. Check the resulting files for errors.<\/p><\/div><div class=\"section\"><h2>Setting the working directory: WorkingDirectory<\/h2><p>Use this key to set your programs working directory. Every relative path the executable accesses will be relative to its working directory.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;WorkingDirectory&lt;\/key&gt;\n&lt;string&gt;\/tmp&lt;\/string&gt;<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>This input field supports auto-completion by pressing Esc. Clicking on the <img class=\"icon\" src=\"http:\/\/www.launchd.info\/img\/NSRevealFreestandingTemplate.png\" alt=\"image\" \/> button lets you select the directory.<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Constraining your job: SoftResourceLimit HardResourceLimit<\/h2><p>Sometimes it is desirable to constrain a program to prevent resource starvation or enhance system performance.<\/p><p>A resource limit is specified as a soft limit and a hard limit.  When a soft limit is exceeded a process may receive a signal (for example, if the cpu time or file size is exceeded), but it will be allowed to continue execution until it reaches the hard limit (or modifies its resource limit).<\/p><p>Only the super-user may raise the hard limits. Other users may only alter soft limits within the range from 0 to the corresponding hard limit or (irreversibly) lower the hard limit.<\/p><p class=\"note\">The biggest value you can specify is the maximum value of a 64-bit signed integer: 9223372036854775807<\/p><h3>CPU<\/h3><p>The maximum amount of CPU time in seconds the process can use. If it runs for longer than this, it receives signal SIGXCPU.<\/p><h3>FileSize<\/h3><p>The program won't be able to create files larger than this. Trying to create a larger file will result in signal SIGXFSZ being sent.<\/p><p>The following example sets the soft limit to 512kiB, the hard limit to 1MiB.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;HardResourceLimits&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;FileSize&lt;\/key&gt;\n\t&lt;integer&gt;1048576&lt;\/integer&gt;\n&lt;\/dict&gt;\n&lt;key&gt;SoftResourceLimits&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;FileSize&lt;\/key&gt;\n\t&lt;integer&gt;524288&lt;\/integer&gt;\n&lt;\/dict&gt;<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Values in grey represent the current default values of the launchd instance.<\/p><\/div><\/div><h3>NumberOfFiles<\/h3><p>The maximum number of files the program can have opened at the same time. Trying to open more files will result in error EMFILE.<\/p><p class=\"note\">The maximum value for this subkey is OPEN_MAX as defined in sys\/syslimits.h: 10240<\/p><h3>Core<\/h3><p>The maximum size core file that this process can create. If the process terminates and would dump a core file larger than this, then no core file is created, so setting this limit to zero prevents core files from ever being created.<\/p><p class=\"note\">Core files will be written to the directory \/cores.<\/p><h3>Data<\/h3><p>The maximum size of data memory in bytes for the process. If the process tries to allocate data memory beyond this amount, the allocation function fails.<\/p><h3>MemoryLock<\/h3><p>The maximum amount of memory in bytes that can be locked into physical memory. Locked memory will never be paged out.<\/p><h3>NumberOfProcesses<\/h3><p>The maximum number of processes that can be created with the same user ID. If you have reached the limit for your user ID, fork will fail with error EAGAIN.<\/p><h3>ResidentSetSize<\/h3><p>The maximum amount of physical memory in bytes that this process should get. This parameter is a guide for the system's scheduler and memory allocator; the system may give the process more memory when there is a surplus.<\/p><h3>Stack<\/h3><p>The maximum stack size for the process. If the process tries to extend its stack past this size, it gets a SIGSEGV signal.<\/p><\/div><div class=\"section\"><p>A job definition is perfectly valid just with a Label and the Program or ProgramArguments key, but it won't do anything. You need to tell launchd <em>when<\/em> to start it.<\/p><\/div><div class=\"section\"><h2>RunAtLoad: Running the job when it is loaded<\/h2><p>This key is used to start the job as soon as it has been loaded. For daemons this means execution at boot time, for agents execution at login.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;RunAtLoad&lt;\/key&gt;\n&lt;true\/&gt;<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Starting a job every <em>n<\/em> seconds: StartInterval<\/h2><p>If you have a job that you want to execute every <em>n<\/em> seconds, this is for you. To run your job once every hour specify:<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;StartInterval&lt;\/key&gt;\n&lt;integer&gt;3600&lt;\/integer&gt;<\/p><\/div><\/div><p class=\"note\">If the system is asleep, the job will be started the next time the computer wakes up. If multiple intervals transpire before the computer is woken, those events will be coalesced into one event upon wake from sleep.<\/p><\/div><div class=\"section\"><h2>Starting a job at a specific time\/date: StartCalendarInterval<\/h2><p>Use this key to run your job at a specific time, say every day at 3:00 AM:<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;StartCalendarInterval&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;Hour&lt;\/key&gt;\n\t&lt;integer&gt;3&lt;\/integer&gt;\n\t&lt;key&gt;Minute&lt;\/key&gt;\n\t&lt;integer&gt;0&lt;\/integer&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><p>Available keys are<\/p><table><tr><th>Key<\/th>\n\t\t\t\t\t\t<th>Type<\/th>\n\t\t\t\t\t\t<th>Values<\/th>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Month<\/td>\n\t\t\t\t\t\t<td>Integer<\/td>\n\t\t\t\t\t\t<td>Month of year (1..12, 1 being January)<\/td>\n\t\t\t\t\t<\/tr><tr class=\"odd\"><td>Day<\/td>\n\t\t\t\t\t\t<td>Integer<\/td>\n\t\t\t\t\t\t<td>Day of month (1..31)<\/td>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Weekday<\/td>\n\t\t\t\t\t\t<td>Integer<\/td>\n\t\t\t\t\t\t<td>Day of week (0..7, 0 and 7 being Sunday)<\/td>\n\t\t\t\t\t<\/tr><tr class=\"odd\"><td>Hour<\/td>\n\t\t\t\t\t\t<td>Integer<\/td>\n\t\t\t\t\t\t<td>Hour of day (0..23)<\/td>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Minute<\/td>\n\t\t\t\t\t\t<td>Integer<\/td>\n\t\t\t\t\t\t<td>Minute of hour (0..59)<\/td>\n\t\t\t\t\t<\/tr><\/table><p class=\"note\">Omitted keys are interpreted as a wildcard. If we omitted the Minute key in our previous example the job would run every minute of the third hour of every day.<\/p><p>You may specify several calendar intervals by providing a list of dictionaries. The following job is run every day at 3:00 AM and every full hour on Sundays.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;StartCalendarInterval&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Hour&lt;\/key&gt;\n\t\t&lt;integer&gt;3&lt;\/integer&gt;\n\t\t&lt;key&gt;Minute&lt;\/key&gt;\n\t\t&lt;integer&gt;0&lt;\/integer&gt;\n\t&lt;\/dict&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Minute&lt;\/key&gt;\n\t\t&lt;integer&gt;0&lt;\/integer&gt;\n\t\t&lt;key&gt;Weekday&lt;\/key&gt;\n\t\t&lt;integer&gt;0&lt;\/integer&gt;\n\t&lt;\/dict&gt;\n&lt;\/array&gt;<\/p><\/div><\/div><h3>Cron-style time specifications<\/h3><p>Some intervals are very tedious to specify in launchd. Users familiar with cron know that there is a more elegant way to specify such intervals. A quick example: You want to run a program every five minutes between 20:00 and 23:00. In launchd you have to list all 36 matching timestamps like this:<\/p><p>&lt;key&gt;StartCalendarInterval&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Hour&lt;\/key&gt;\n\t\t&lt;integer&gt;20&lt;\/integer&gt;\n\t\t&lt;key&gt;Minute&lt;\/key&gt;\n\t\t&lt;integer&gt;0&lt;\/integer&gt;\n\t&lt;\/dict&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Hour&lt;\/key&gt;\n\t\t&lt;integer&gt;20&lt;\/integer&gt;\n\t\t&lt;key&gt;Minute&lt;\/key&gt;\n\t\t&lt;integer&gt;5&lt;\/integer&gt;\n\t&lt;\/dict&gt;\n\t.\n\t.\n\t.\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Hour&lt;\/key&gt;\n\t\t&lt;integer&gt;22&lt;\/integer&gt;\n\t\t&lt;key&gt;Minute&lt;\/key&gt;\n\t\t&lt;integer&gt;55&lt;\/integer&gt;\n\t&lt;\/dict&gt;\n&lt;\/array&gt;<\/p><p>LaunchControl allows you to generate this list from the equivalent cron-style specification: *\/5 20-22 * * *<\/p><p>The following description of the cron time specification was taken from the cron manual page:<\/p><table><tr><th>Field<\/th>\n\t\t\t\t\t\t<th>Supported values<\/th>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Minutes<\/td>\n\t\t\t\t\t\t<td>0-59<\/td>\n\t\t\t\t\t<\/tr><tr class=\"odd\"><td>Hours<\/td>\n\t\t\t\t\t\t<td>0-23<\/td>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Day of month<\/td>\n\t\t\t\t\t\t<td>1-31<\/td>\n\t\t\t\t\t<\/tr><tr class=\"odd\"><td>Month<\/td>\n\t\t\t\t\t\t<td>1-12<\/td>\n\t\t\t\t\t<\/tr><tr class=\"even\"><td>Day of week<\/td>\n\t\t\t\t\t\t<td>0-7 (0 and 7 is Sunday)<\/td>\n\t\t\t\t\t<\/tr><\/table><p>A field may be an asterisk (*), which always stands for \"first-last\".<\/p><p>Ranges of numbers are allowed.  Ranges are two numbers separated with a hyphen.  The specified range is inclusive.  For example, 8-11 for an \"hours\" entry specifies execution at hours 8, 9, 10 and 11.<\/p><p>Lists are allowed.  A list is a set of numbers (or ranges) separated by commas.  Examples: \"1,2,5,9\", \"0-4,8-12\".<\/p><p>Step values can be used in conjunction with ranges.  Following a range with \"\/&lt;number&gt;\" specifies skips of the number's value through the range.  For example, \"0-23\/2\" can be used in the hours field to specify command execution every other hour.  Steps are also permitted after an asterisk, so if you want to say \"every two hours\", just use \"*\/2\".<\/p><p class=\"note\">Unlike cron which skips job invocations when the computer is asleep, launchd will start the job the next time the computer wakes up.  If multiple intervals transpire before the computer is woken, those events will be coalesced into one event upon wake from sleep.<\/p><\/div><div class=\"section\"><h2>Run when a device has been mounted: StartOnMount<\/h2><p>A job with this key set to true will be started whenever a device is mounted, in other words, when you insert a CD\/DVD, connect an external hard disk or mount a virtual filesystem.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;StartOnMount&lt;\/key&gt;\n&lt;true\/&gt;<\/p><\/div><\/div><p class=\"note\">launchd does not report <em>which<\/em> device has been mounted.<\/p><\/div><div class=\"section\"><h2>Run when a path has been modified: WatchPaths<\/h2><p>This key takes a list of strings, each string representing a file or directory.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;WatchPaths&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;string&gt;\/path\/to\/directory_or_file&lt;\/string&gt;\n&lt;\/array&gt;<\/p><\/div><\/div><p>If the path points to a file, creating, removing and writing to this file will be start the job.<\/p><p>If the path points to a directory, creating and removing this directory, as well as creating, removing and writing files in this directory will start the job. Actions performed in subdirectories of this directory will not be detected.<\/p><p class=\"note\">launchd does not report <em>which<\/em> path has been modified.<\/p><\/div><div class=\"section\"><h2>Run when files are available for processing: QueueDirectories<\/h2><p>This key takes a list of strings, each string representing a directory.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;QueueDirectories&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;string&gt;\/path\/to\/directory&lt;\/string&gt;\n&lt;\/array&gt;<\/p><\/div><\/div><p>Whenever one of the directories specified is <em>not<\/em> empty, the job will be started. It is the responsibility of the job to remove each processed file, otherwise the job will be restarted after ThrottleInterval seconds.<\/p><p class=\"note\">launchd does not report in <em>which<\/em> directory  new files have been found nor what their names are.<\/p><\/div><div class=\"section\"><p>launchd may be used to make sure that a job is running depending on certain conditions. <\/p><\/div><div class=\"section\"><h2>...no Matter What<\/h2><p>This is the most basic setting. A value of true will run the job as soon as the job definition is loaded and restart it should it ever go down. launchd will wait ThrottleInterval seconds between restarts.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;true\/&gt;<\/p><\/div><\/div><p class=\"note\">If you configure your job to be kept alive 'no matter what', launchd will run the job immediately after loading it.<\/p><\/div><div class=\"section\"><h2>...depending on the Last Exit Status:\n\t\t\t\t\tSuccessfulExit<\/h2><p>When a job terminates it can set a numeric return value, its exit status. By convention an exit status of 0 is used when the program exited successfully. Any other number informs about a problem.<\/p><p>The subkey SuccessfulExit can be used to restart your program depending on its previous exit status. When set to true the job will be restarted until it fails. Setting this value to false will restart the job  until it succeeds.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;SuccessfulExit&lt;\/key&gt;\n\t&lt;true\/&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><p class=\"note\">If this subkey is specified, launchd will run the job immediately after loading. To avoid this behaviour set the subkey AfterInitialDemand.<\/p><\/div><div class=\"section\"><h2>...depending on Crashes:\n\t\t\t\t\tCrashed<\/h2><p>The subkey Crashed can be used to restart your program after it has crashed. When set to true the job will be restarted after it crashed. Setting this value to false will restart the job unless it has crashed.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;Crashed&lt;\/key&gt;\n\t&lt;true\/&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><p class=\"note\">If this subkey is specified, launchd will run the job immediately after loading. To avoid this behaviour set the subkey AfterInitialDemand.<\/p><\/div><div class=\"section\"><h2>...depending on Network Availability: NetworkState 10.5 10.10<\/h2><p>Setting this subkey to true will start the job when\/while any network is\/becomes available. Setting this subkey to false will start the job when\/while all network connections are down.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;NetworkState&lt;\/key&gt;\n\t&lt;true\/&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><p class=\"note\">If this subkey is specified, launchd will run the job immediately after loading. To avoid this behaviour set the subkey AfterInitialDemand.<\/p><\/div><div class=\"section\"><h2>...depending on the Existence of a Path: PathState<\/h2><p>Use this subkey to keep a job alive as long as a given path exists (true) or does not exist (false).<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;PathState&lt;\/key&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;\/tmp\/runJob&lt;\/key&gt;\n\t\t&lt;true\/&gt;\n\t&lt;\/dict&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><p class=\"note\">If this subkey is specified, launchd will run the job immediately after loading. To avoid this behaviour set the subkey AfterInitialDemand.<\/p><\/div><div class=\"section\"><h2>...depending on Other Jobs: OtherJobEnabled<\/h2><p>The name of this subkey is somewhat misleading. It does not check if a certain job is enabled, but if a job is <em>loaded<\/em>. The following job will be started (and restarted) when a job with label local.otherJob is not loaded. It will be terminated when a job with label local.otherJob is loaded.<\/p><p class=\"note\">Always specify the Label of the job you are referring to, not the name of the job definition file.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;OtherJobEnabled&lt;\/key&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;local.otherJob&lt;\/key&gt;\n\t\t&lt;false\/&gt;\n\t&lt;\/dict&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Delaying the start: AfterInitialDemand<\/h2><p>This key affects other keys related to run conditions like RunAtLoad, StartCalendarInterval, WatchPaths or KeepAlive. If set, these keys will be ignored until the job has been started manually.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;KeepAlive&lt;\/key&gt;\n&lt;dict&gt;\n\t&lt;key&gt;AfterInitialDemand&lt;\/key&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;local.otherJob&lt;\/key&gt;\n\t\t&lt;true\/&gt;\n\t&lt;\/dict&gt;\n&lt;\/dict&gt;<\/p><\/div><\/div><\/div><p><\/p><h2>Permissions and Security<\/h2><div class=\"section\"><h2>UserName GroupName InitGroups<\/h2><p>Using the keys UserName and GroupName you can specify as which user the job should run. Both keys expect a string (the name of the user\/group, not the id) as value.<\/p><p>The key InitGroups specifies if launchd should call the function initgroups(3) before starting the job.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;UserName&lt;\/key&gt;\n&lt;string&gt;nobody&lt;\/string&gt;\n&lt;key&gt;GroupName&lt;\/key&gt;\n&lt;string&gt;nobody&lt;\/string&gt;\n&lt;key&gt;InitGroups&lt;\/key&gt;\n&lt;true\/&gt;<\/p><\/div><\/div><p class=\"note\">These keys will be ignored when specified for an agent.<\/p><\/div><div class=\"section\"><h2>Umask<\/h2><p>Using this key you can specify which permissions a file or folder will have by default when it is created by this job. The value is expected to be a decimal number.<\/p><p>In the following example we specify that new files should be readable and writable by the user owning the file. In addition new directories should be searchable by the user. Group members should be able to read new files and search created directories. Others should not be able to access new files or directories at all.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>First you have to come up with three digits between 0 and 7. The first one defines user access, the second one group access and the third one world access.<\/p><table class=\"wikitable\"><tr><th>Digit<\/th>\n\t\t\t\t\t\t\t\t<th>Granted permissions<\/th>\n\t\t\t\t\t\t\t<\/tr><tr><td>0<\/td>\n\t\t\t\t\t\t\t\t<td>read, write and execute\/search<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>1<\/td>\n\t\t\t\t\t\t\t\t<td>read and write<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>2<\/td>\n\t\t\t\t\t\t\t\t<td>read and execute\/search<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>3<\/td>\n\t\t\t\t\t\t\t\t<td>read only<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>4<\/td>\n\t\t\t\t\t\t\t\t<td>write and execute<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>5<\/td>\n\t\t\t\t\t\t\t\t<td>write only<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>6<\/td>\n\t\t\t\t\t\t\t\t<td>execute\/search only<\/td>\n\t\t\t\t\t\t\t<\/tr><tr><td>7<\/td>\n\t\t\t\t\t\t\t\t<td>none<\/td>\n\t\t\t\t\t\t\t<\/tr><\/table><p>We want read, write and search permissions for the user: 0. We want read and search permissions for the group: 2. We want no access at all for everyone else: 7. Together: 027<\/p><p>The last step is to convert this octal number to its decimal representation. We'll use the bc command line tool to do that:<\/p><p>host:~ user$ <b>echo \"obase=10;ibase=8; <em>027<\/em>\" | bc<\/b>\n23<\/p><p>The number we want is 23:<\/p><p>&lt;key&gt;Umask&lt;\/key&gt;\n&lt;integer&gt;23&lt;\/integer&gt;<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>LaunchControl saves you the trouble of coming up with a bit mask and converting it to decimal. Just select the bits you'd like to set. Every selected bit is a revoked permission. To make things more clear the actual folder\/file permissions are displayed along with the octal representation.\n\t\t\t\t\t\t<\/p><\/div><\/div><p class=\"note\">This key will be ignored when specified for an agent.<\/p><p class=\"note\">New files will never be created with the execute bit set.<\/p><\/div><div class=\"section\"><h2>RootDirectory<\/h2><p>Using this key you can execute the job in a chroot(2) jail. As far as the job is concerned, the directory provided with this key is the root of the filesystem. This implies that you need copies of all files required for the job inside this directory. This includes libraries, frameworks, command line tools and configuration files.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;RootDirectory&lt;\/key&gt;\n&lt;string&gt;\/var\/jail&lt;\/string&gt;<\/p><\/div><\/div><p class=\"note\">This key will be ignored when specified for an agent.<\/p><\/div><div class=\"section\"><h2>AbandonProcessGroup<\/h2><p>When launchd wants to terminate a job it sends a SIGTERM signal which will be propagated to all child processes of the job as well. Setting the value of this key to true will stop this propagation, allowing the child processes to survive their parents.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;AbandonProcessGroup&lt;\/key&gt;\n&lt;true\/&gt;<\/p><\/div><\/div><\/div><div class=\"section\"><h2>ExitTimeOut<\/h2><p>launchd stops a job by sending the signal SIGTERM to the process. Should the job not terminate within ExitTimeOut seconds (20 seconds by default), launchd will send signal SIGKILL to force-quit it.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;ExitTimeOut&lt;\/key&gt;\n&lt;integer&gt;30&lt;\/integer&gt;<\/p><\/div><\/div><\/div><div class=\"section\"><h2>TimeOut 10.10<\/h2><p>The suggested idle time in seconds before the job should quit.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;TimeOut&lt;\/key&gt;\n&lt;integer&gt;30&lt;\/integer&gt;<\/p><\/div><\/div><p class=\"note\">launchd will not stop the job. This is the responsibility of the job itself.<\/p><\/div><div class=\"section\"><h2>ThrottleInterval<\/h2><p>Time in seconds to wait between program invocations.<\/p><p>This key may me used in combination with KeepAlive to run a job every n seconds while a certain condition applies.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;ThrottleInterval&lt;\/key&gt;\n&lt;integer&gt;30&lt;\/integer&gt;<\/p><\/div><\/div><\/div><div class=\"section\"><h2>LegacyTimers 10.9.2<\/h2><p>This key controls the behavior of timers created by the job. By default on OS X Mavericks version 10.9 and later, timers created by launchd jobs are coalesced. Batching the firing of timers with similar deadlines improves the overall energy efficiency of the system. If this key is set to true, timers created by the job will opt into less efficient but more precise behavior and not be coalesced with other timers.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;LegacyTimers&lt;\/key&gt;\n&lt;true\/&gt;<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Nice<\/h2><p>Run a job at an altered scheduling priority. Possible values range from -20 to 20. The default value is 0. Lower nice values cause more favorable scheduling.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;key&gt;Nice&lt;\/key&gt;\n&lt;integer&gt;-5&lt;\/integer&gt;<\/p><\/div><\/div><p class=\"note\">Nice values lower than zero are honored by daemons only.<\/p><\/div><\/div><div class=\"content\" id=\"page-Operation\"><div class=\"section\"><p>This section explains how to acquire information about launchd jobs and how to load, unload, start and stop them.<\/p><p>All of this can be accomplished using the command line utility launchctl, but you will see that it is far easier using <a href=\"http:\/\/www.soma-zone.com\/LaunchControl\">LaunchControl<\/a>. To <a href=\"http:\/\/nathangrigg.net\/2014\/03\/launchcontrol-for-launchd\/\">quote<\/a> Nathan Grigg:<\/p><p><i>\"<a href=\"http:\/\/www.soma-zone.com\/LaunchControl\">LaunchControl<\/a> is a Mac app by soma-zone that helps manage launchd lobs. It aims to do \u201cone thing well\u201d and succeeds spectacularly. Whether you are new to writing launchd agents or you already have some system in place, go buy LaunchControl now.\"<\/i><\/p><p class=\"note\">All examples assume that you are logged in as a regular (non-privileged) user.<\/p><p class=\"note\">All examples refer to agents. Should you want to work with daemons you have to use sudo launchctl instead of launchctl. You will be prompted for the admin password when working with daemons.<\/p><p class=\"note\"><b>Bold<\/b> text is input, regular text is output.<\/p><\/div><div class=\"section\"><h2>Getting information about a job<\/h2><p>We will get some information about a job with Label com.example.app.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"launchctl\"><p>We issue the command launchctl list, which returns a long list of currently loaded jobs. To search for our specific job with the label com.example.app we send the output to the command grep using the pipe character |.<\/p><p>host:~ user$ <b>launchctl list | grep com.example.app<\/b>\n-\t2\tcom.example.app<\/p><ul><li>We received one line of output, which means that our job is currently <b>loaded<\/b>.<\/li>\n\t\t\t\t\t\t\t<li>The first colum contains the current process id. We received a - instead of a number, which means that while the job is loaded it is currently not <b>running<\/b>. Any numeric value in this column would indicate a running job.<\/li>\n\t\t\t\t\t\t\t<li>The second column displays the last <b>exit code<\/b>. A value of 0 indicates that the job finished successfully, a positive number that the job has reported an error, a negative number that the process was terminated because it received a signal.<\/li>\n\t\t\t\t\t\t<\/ul><p class=\"note\">launchctl also accepts the Label of a job as an optional argument. This results in a more detailed output in a JSON-like format.<\/p><p>To find out if a job's Disabled key has been <b>overridden<\/b> we query the <a target=\"The_override_database\">override database<\/a> using the xpath command line utility. Its first argument is the name of the file to query, its second argument the query itself. You have to replace the job label in the query.<\/p><p>host:~ user$ <b>xpath \/var\/db\/launchd.db\/com.apple.launchd.peruser.`id -u`\/overrides.plist \\\n\"\/plist\/dict\/key[text()='<em>com.example.app<\/em>']\/following-sibling::*[1]\/key\/following-sibling::*[1]\"<\/b>\nFound 1 nodes:\n-- NODE --\n&lt;true \/&gt;<\/p><p class=\"note\">The input has been split across two lines, hence the \\ at the end of the first line. You can enter the command as a single line by ommitting this character and the newline.<\/p><p>We get a single hit, which means that the Disabled key is overridden. Its value is true. We would receive No nodes found if the key had not been overriden.<\/p><p>The database for daemons is \/var\/db\/launchd.db\/com.apple.launchd\/overrides.plist (OSX 10.5 \u2013 OSX 10.9) or  \/var\/db\/com.apple.xpc.launchd\/disabled.plist (OSX 10.10 and later).<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>This is the same job as seen using LaunchControl.<\/p><p>All information concerning the job status can be found in the left section of the main window.<\/p><ul><li>It is <b>loaded<\/b> (Status not equal to \"Unloaded\")<\/li>\n\t\t\t\t\t\t\t<li>It is currently <b>not running<\/b> (Status not equal to \"Running\")<\/li>\n\t\t\t\t\t\t\t<li>It failed with <b>exit code<\/b> 2 (Status \"Error 2\")<\/li>\n\t\t\t\t\t\t\t<li>A <b>description<\/b> of \"Error 2\": The program we tried to start could not be found.<\/li>\n\t\t\t\t\t\t\t<li>The job has been <b>disabled<\/b> (the checkbox is unchecked)<\/li>\n\t\t\t\t\t\t\t<li>The Disabled key is <b>not overridden<\/b> (grey checkbox; a green checkbox indicates that the job is permanently enabled, a red one that the job is permanently disabled)<\/li>\n\t\t\t\t\t\t<\/ul><p>The problematic key would have been highlighted even if the job had not been loaded\/run yet.<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Loading a Job<\/h2><div class=\"tabset\"><div class=\"tab\" title=\"launchctl\"><p>We have already learned that launchd loads jobs automatically, daemons at boot time, agents when the user logs in. launchctl allows you to load jobs manually. The following example loads the agent com.example.error. It is an user agent, so we know its definition will be in ~\/Library\/LaunchAgents. <\/p><p>host:~ user$ <b>launchctl load ~\/Library\/LaunchAgents\/com.example.app.plist<\/b><\/p><p>If everything went ok you will see no output. But you may receive this line:<\/p><p>host:~ user$ <b>launchctl load ~\/Library\/LaunchAgents\/com.example.app.plist<\/b>\nnothing found to load<\/p><p>If this happens the job has either set the Disabled key to true or the job has been disabled in the <a target=\"The_override_database\">override database<\/a>. In either case you can force launchctl to load the job like this:\n\t\t\t\t\t\t<\/p><p>host:~ user$ <b>launchctl load -F ~\/Library\/LaunchAgents\/com.example.app.plist<\/b><\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Loading a job with LaunchControl is a matter of clicking the Load button in the upper right of the main section. Should the job be disabled or should the Disabled key be overridden in the <a target=\"The_override_database\">override database<\/a> the button will have its label changed to \"Force Load\" automatically.<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Unloading a Job<\/h2><div class=\"tabset\"><div class=\"tab\" title=\"launchctl\"><p>launchctl allows you to unload jobs manually. The following example unloads the agent com.example.app. It is an user agent, so we know its definition will be in ~\/Library\/LaunchAgents.<\/p><p>host:~ user$ <b>launchctl unload ~\/Library\/LaunchAgents\/com.example.app.plist<\/b><\/p><p>If everything went ok you will see no output.<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Unloading a job with LaunchControl is similar to loading a job. When the currently selected job is loaded, the \"Load\" button will have its label changed to \"Unload\":<\/p><\/div><\/div><\/div><div class=\"section\"><h2>Starting a Job<\/h2><p>Sometimes you may want to start a job regardless of its run conditions. This section explains how to do that.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"launchctl\"><p>This command expects the job Label as its argument.<\/p><p>host:~ user$ <b>launchctl start com.example.app<\/b><\/p><p>If everything went ok you will see no output. But you may get something like this:<\/p><p>host:~ user$ <b>launchctl start com.example.app<\/b>\nlaunchctl start error: No such process<\/p><p>You either misspelled the Label of the job or forgot to load it in the first place.<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Select the job and select Job&gt;Start (\u2318\u21e7T) from the menu.<\/p><\/div><\/div><p class=\"note\">A job has to be loaded before it can be started.<\/p><\/div><div class=\"section\"><h2>Stopping a Job<\/h2><div class=\"tabset\"><div class=\"tab\" title=\"launchctl\"><p>This command expects the job Label as its argument.<\/p><p>host:~ user$ <b>launchctl stop com.example.app<\/b><\/p><p>If everything went ok you will see no output. But you may get something like this:<\/p><p>host:~ user$ <b>launchctl stop com.example.app<\/b>\nlaunchctl stop error: No such process<\/p><p>You either misspelled the Label of the job or forgot to load it in the first place.<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Select the job and select Job&gt;Stop (\u2318\u21e7T) from the menu.<\/p><\/div><\/div><p>Stopping a job will send the signal SIGTERM to the process. Should this not stop the process launchd will wait ExitTimeOut seconds (20 seconds by default) before sending SIGKILL.<\/p><p class=\"note\">The job may be intantly restarted when its run conditions are met.<\/p><\/div><div class=\"section\"><h2 id=\"The_override_database\">The override database<\/h2><p>When launchd is about to load a job it will check if it has the Disabled key set. Disabled jobs will not be loaded. But the value of this key can be overridden.<\/p><h3>Permanently disabling a job<\/h3><div class=\"tabset\"><div class=\"tab\" title=\"Terminal\"><p>host:~ user$ <b>launchctl unload -w ~\/Library\/LaunchAgents\/com.example.app.plist<\/b><\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Select the job and select Job&gt;Override Disabled Key&gt;Always True from the menu.<\/p><\/div><\/div><p class=\"note\">OSX 10.10 Yosemite and later: While you will always be able to switch a job's permanent status between enabled and disabled, you cannot (trivially) remove it from the override database again.<\/p><h3>Permanently enabling a job<\/h3><div class=\"tabset\"><div class=\"tab\" title=\"Terminal\"><p>host:~ user$ <b>launchctl load -w ~\/Library\/LaunchAgents\/com.example.app.plist<\/b><\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Select the job and select Job&gt;Override Disabled Key&gt;Always False from the menu.<\/p><\/div><\/div><p class=\"note\">OSX 10.10 Yosemite and later: While you will always be able to switch a job's permanent status between enabled and disabled, you cannot (trivially) remove it from the override database again.<\/p><h3>Removing a job from the override database (OSX 10.5 Leopard - OSX 10.9 Mavericks)<\/h3><div class=\"tabset\"><div class=\"tab\" title=\"Terminal\"><p>The procedure is different for agents and daemons. You'll need the label of the job you want to remove. In this example: local.job. For agents type:<\/p><p>host:~ user$ <b>sudo \/usr\/libexec\/Plistbuddy \/var\/db\/launchd.db\/com.apple.launchd.peruser.`echo $UID`\/overrides.plist -c Delete:<em>local.job<\/em><\/b><\/p><p>For daemons the command reads<\/p><p>host:~ user$ <b>sudo \/usr\/libexec\/Plistbuddy \/var\/db\/launchd.db\/com.apple.launchd\/overrides.plist -c Delete:<em>local.job<\/em><\/b><\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>Select the job and select Job&gt;Override Disabled Key&gt;Don't Override from the menu.<\/p><\/div><\/div><h3>Removing a job from the override database (OSX 10.10 Yosemite and later)<\/h3><p>Starting with OSX 10.10 Yosemite the override database was moved to a different location. Additionally the database is read only when launchd starts. Modifications to the database will be overwritten when launchd quits. Because of this modifications have to be done in recovery mode. LaunchControl can not be used to remove jobs from the override database. On OSX 10.11 El Capitan things are even more complicated due to System Integrity Protection, a facility which prevents modification of system files by any user including root.<\/p><h4>Boot into Recovery Mode<\/h4><ul><li>Restart your Mac<\/li>\n\t\t\t\t\t\t<li>Hold down \u2318R before OSX starts up. Hold these keys until the Apple logo appears. After your computer finishes starting up, you should see a desktop with an OS X menu bar and an OS X Utilities window<\/li>\n\t\t\t\t\t\t<li>Select Terminal from the Utilities menu<\/li>\n\t\t\t\t\t<\/ul><h4>Disabling System Integrity Protection (OSX 10.11 El Capitan and later)<\/h4><ul><li>Type csrutil disable and press the \u23ce key.<\/li>\n\t\t\t\t\t\t<li>You should see a message that System Integrity Protection has been disabled<\/li>\n\t\t\t\t\t<\/ul><h4>Remove the job from the override database<\/h4><p>The procedure is different for agents and daemons. You'll need the label of the job you want to remove. In this example: local.job. For agents type:<\/p><p>host:~ user$ <b>\/usr\/libexec\/Plistbuddy \/var\/db\/com.apple.xpc.launchd\/disabled.`id -u`.plist -c Delete:<em>local.job<\/em><\/b><\/p><p>For daemons the command reads<\/p><p>host:~ user$ <b>\/usr\/libexec\/Plistbuddy \/var\/db\/com.apple.xpc.launchd\/disabled.plist -c Delete:<em>local.job<\/em><\/b><\/p><h4>Restore System Integrity Protection (OSX 10.11 El Capitan and later)<\/h4><ul><li>Type csrutil enable and press the \u23ce key.<\/li>\n\t\t\t\t\t\t<li>You should see a message that System Integrity Protection has been enabled<\/li>\n\t\t\t\t\t<\/ul><h4>Reboot<\/h4><ul><li>Type reboot and press the \u23ce key.<\/li>\n\t\t\t\t\t<\/ul><\/div><\/div><div class=\"content\" id=\"page-Troubleshooting\"><div class=\"section\"><p>Finding out why a launchd job does not perform as expected can be a time consuming task if you have to do it manually. Using <a href=\"http:\/\/www.soma-zone.com\/LaunchControl\">LaunchControl<\/a> you can troubleshoot a job in a matter of seconds. Most of the time just selecting the job will highlight the problem and give advice on how to fix it.<\/p><p class=\"note\">Remember to always (re)load a job definition after changing it. Just saving it will leave launchd oblivious to the change.<\/p><\/div><div class=\"section\"><h2>Common mistakes<\/h2><h3>launchd(8) is not init(8)<\/h3><p>It is important to remember that launchd is responsible for turning a process into a background process. When referencing a shell script in Program or ProgramArguments make sure the script does not put the actual program into the background. Some programs require special command line arguments (something like --foreground) to prevent them from daemonizing themselves. If in doubt try to run the program\/script from the terminal. The program\/script should only return after the process is done.<\/p><p>The same goes for resource limits, redirection of standard input\/output\/error, re-nicing and so on. launchd supports a wealth of configuration options. Use them and don't try to emulate their effects.<\/p><h3>launchd(8) is no shell<\/h3><p>Pipes (|), redirection (&lt;,&gt;) and variable expansion ($VAR) are only three examples for shell constructs that won't be interpreted by launchd(8). If you want to write shell code, give it to the shell. Instead of<\/p><p>&lt;key&gt;ProgramArguments&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;string&gt;\/bin\/cat&lt;\/string&gt;\n\t&lt;string&gt;\/etc\/passwd&lt;\/string&gt;\n\t&lt;string&gt;|&lt;\/string&gt;\n\t&lt;string&gt;sort&lt;\/string&gt;\n&lt;\/array&gt;<\/p><p>write<\/p><p>&lt;key&gt;ProgramArguments&lt;\/key&gt;\n&lt;array&gt;\n\t&lt;string&gt;\/bin\/sh&lt;\/string&gt;\n\t&lt;string&gt;-c&lt;\/string&gt;\n\t&lt;string&gt;cat \/etc\/passwd | sort&lt;\/string&gt;\n&lt;\/array&gt;<\/p><h3>Don't use sudo in agents<\/h3><p>The unix command sudo is used to run unix commands as a different user. If executed by an unprivileged (non-root) user it requires the user to enter a password. Obviously, when used by a LaunchAgent the password cannot be entered. Solution: Convert the agent to a daemon.<\/p><\/div><div class=\"section\"><h2>My job is not loaded on login\/boot<\/h2><div class=\"tabset\"><div class=\"tab\" title=\"Terminal\"><p>Load the job manually and see what launchd has to say. The following example did not load because the job definition had the wrong permissions.<\/p><p>host:~ user$ <b>launchctl load ~\/Library\/LaunchAgents\/com.example.app.plist<\/b>\nlaunchctl: Dubious permissions on file (skipping): \/Users\/rpaehlig\/Library\/LaunchAgents\/com.example.app.plist\nnothing found to load<\/p><p class=\"note\">Make sure to load your job definition into the correct launchd instance. The example above loads an agent definition into the user launchd instance. Execute this command with root privileges (sudo) to load a daemon definition into the system launchd instance.<\/p><p>Checklist:<\/p><ul><li>Check the job definition has the proper file extension (.plist (Error: Dubious file. Not of type .plist)<\/li>\n\t\t\t\t\t\t\t<li>Check the job definition for wellformedness (Error: no plist was returned for)\n\t\t\t\t\t\t\t<p>host:~ user$ <b>plutil -lint ~\/Library\/LaunchAgents\/com.example.app.plist<\/b>\n\/Users\/Guest\/Library\/LaunchAgents\/com.example.app.plist: OK<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t<li>Check the Disabled key. Disabled jobs won't be automatically loaded.<\/li>\n\t\t\t\t\t\t\t<li>Check if the job has been disabled in the <a target=\"overridedb_check\">override database<\/a><\/li>\n\t\t\t\t\t\t\t<li>Check the permissions of the job definition. Job definitions have to be readable by the user. Job definitions must not be writable by group or other. Suggested privileges: 0644. (Error: Dubious permissions on file)<\/li>\n\t\t\t\t\t\t\t<li>Check the permissions of the directory containing the job definition. It has to be searchable by the user and must not be writable by group or other. Suggested privileges: 0755<\/li>\n\t\t\t\t\t\t\t<li>Check who owns the job definition. Agents must be owned by their respective user, daemons by root. (Error: Dubious ownership on file)<\/li>\n\t\t\t\t\t\t\t<li>Make sure the Label you provided is unique for this particular launchd instance.<\/li>\n\t\t\t\t\t\t<\/ul><\/div><div class=\"tab\" title=\"LaunchControl\"><p>LaunchControl will mark invalid jobs automatically. Select any invalid job to get further details. All issues (except wellformedness) can be auto-corrected with the click of a button.<\/p><\/div><\/div><\/div><div class=\"section\"><h2>My job does not start automatically after loading<\/h2><p>Make sure you defined a <a target=\"When_to_Start\">run condition<\/a>. LaunchControl will display a warning when no run condition has been specified.<\/p><\/div><div class=\"section\"><h2>My job does not start<\/h2><div class=\"tabset\"><div class=\"tab\" title=\"Terminal\"><p>Open two windows in Terminal.app. In the first one we will monitor the system log file for errors of our launchd instance. Replace the argument of the second grep with the Label of the job refusing to start. If you want to debug an agent enter:<\/p><p>host:~ user$ <b>tail -F \/var\/log\/system.log | grep --line-buffered \"com.apple.launchd.peruser.`id -u`\\[\" | grep \"com.example.app\"<\/b><\/p><p>If you want to debug a daemon enter:<\/p><p>host:~ user$ <b>sudo tail -F \/var\/log\/system.log | grep --line-buffered \"com.apple.launchd\\[\" | grep \"com.example.app\"<\/b><\/p><p>The last command asks for your super user password. Now start the job manually in the second window and see what launchd has to say. The following example did not start because the script did not define an executing interpreter:<\/p><p>Sep 19 12:01:52 host com.apple.launchd.peruser.201[235] (com.example.app[38643]): Job failed to exec(3) for weird reason: 8<\/p><p>launchd tells us that it was unable to execute the program specified in the job definition. To find out what \"weird reason: 8\" is, type this command:<\/p><p>host:~ user$ <b>perl -E 'say $!=shift' 8<\/b>\nExec format error<\/p><p>But you may also get a line like this:<\/p><p>Sep 19 12:01:52 host com.apple.launchd.peruser.201[235] (com.example.app[38643]): Exited with code: 8<\/p><p>While the code is the same, the problem is a different one. launchd has executed the program specified in our job definition just fine, but the program exited with exit code 8. Unfortunately there is no way of knowing what this code means, as a programmer can choose arbitrary exit codes if he wants. Maybe the program uses standard exit codes, in which case you can use the Perl one-liner to get a description. If you are lucky the user manual of the program contains a list of all exit codes with an explanation. Otherwise ask the author about the meaning of the error code.<\/p><p>Fortunately exit codes are not the only way programs can communicate with the user. Add the keys StandardOutPath and StandardErrorPath to your job definition. Reload it and try to start it again. Now examine the files you specified for those keys. It is very likely that you will find the problem description in one of those files.<\/p><p class=\"note\">Some programs accept special arguments to set the location of log files. Make sure to examine those files too.<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><ul><li><p>Select the job and fix any errors\/warnings that LaunchControl points out<\/p><\/li>\n\t\t\t\t\t\t\t<li>Check the system log\n\t\t\t\t\t\t\t\t<p>Select the job, bring up the log panel by selecting View&gt;Toggle Log (\u2318\u2325L) from the menu. Start log monitoring by clicking the <img class=\"icon\" src=\"http:\/\/www.launchd.info\/img\/NSRightFacingTriangleTemplate.png\" alt=\"image\" \/> button. Now start the job by selecting Job&gt;Start (\u2318\u21e7T) from the menu.<\/p>\n\t\t\t\t\t\t\t\t<p>If you get a message Job failed to exec(3) for weird reason then you can get a description by moving your mouse cursor over the error code in the job list (See screenshot). If you get a message Exite with code then launchd started the job just fine. Its just that the program specified in the job definition exited with an error code. Unfortunately there is no way of knowing what this code means, as a programmer can choose arbitrary exit codes if he wants. Maybe the program uses standard exit codes, in which case you can trust the description LaunchControl provides. If you are lucky the user manual of the program contains a list of all exit codes with an explanation. Otherwise ask the author about the meaning of the error code.<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t<li>Check the program output\n\t\t\t\t\t\t\t\t<p>Fortunately exit codes are not the only way programs can communicate with the user. Add the keys StandardOutPath and StandardErrorPath to your job definition and start it again by selecting Job&gt;Start (\u2318\u21e7T) from the menu. Now examine the files you specified for those keys by clicking the <img class=\"icon\" src=\"http:\/\/www.launchd.info\/img\/NSRevealFreestandingTemplate.png\" alt=\"image\" \/> button. It is very likely that you will find the problem description in one of those files.<\/p>\n\t\t\t\t\t\t\t\t<p class=\"note\">Some programs accept special arguments to set the location of log files. Make sure to examine those files too.<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t<\/ul><\/div><\/div><\/div><\/div><div class=\"content\" id=\"page-Cookbook\"><div class=\"section\"><p>This is a collection of recipes serving as examples and inspiration.<\/p><\/div><div class=\"section\" id=\"rcp-restart-application\"><p>The trick here is that an OSX application is a folder <em>containing<\/em> (among other things) the executable file. The name of the executable is stored in a file called Info.plist inside this directory. Fortunately the command line utility open knows how to deal with application bundles.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;\n&lt;!DOCTYPE plist PUBLIC \"-\/\/Apple\/\/DTD PLIST 1.0\/\/EN\" \"http:\/\/www.apple.com\/DTDs\/PropertyList-1.0.dtd\"&gt;\n&lt;plist version=\"1.0\"&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;KeepAlive&lt;\/key&gt;\n\t\t&lt;true\/&gt;\n\t\t&lt;key&gt;Label&lt;\/key&gt;\n\t\t&lt;string&gt;local.Safari.keepAlive&lt;\/string&gt;\n\t\t&lt;key&gt;ProgramArguments&lt;\/key&gt;\n\t\t&lt;array&gt;\n\t\t\t&lt;string&gt;\/usr\/bin\/open&lt;\/string&gt;\n\t\t\t&lt;string&gt;-W&lt;\/string&gt;\n\t\t\t&lt;string&gt;\/Applications\/Safari.app&lt;\/string&gt;\n\t\t&lt;\/array&gt;\n\t&lt;\/dict&gt;\n&lt;\/plist&gt;<\/p><\/div><div class=\"tab\" title=\"LaunchControl\"><p>By far the easiest war to create a job definition like this is to drag and drop the icon of the application onto the LaunchControl icon (in Dock or in Finder). This will create a job definition like this:<\/p><\/div><\/div><\/div><div class=\"section\" id=\"rcp-ThrottleInterval-KeepAlive\"><p>We want to perform a backup of a database but this backup should only be performed while the database is running. We will use the PID file to find out if it is running or not.<\/p><p>Using the subkey PathState of the key KeepAlive we can make sure that our job is run (and kept alive) while our PID file exists. We will delay restarts using the ThrottleInterval key.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;\n&lt;!DOCTYPE plist PUBLIC \"-\/\/Apple\/\/DTD PLIST 1.0\/\/EN\" \"http:\/\/www.apple.com\/DTDs\/PropertyList-1.0.dtd\"&gt;\n&lt;plist version=\"1.0\"&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Label&lt;\/key&gt;\n\t\t&lt;string&gt;com.example.app&lt;\/string&gt;\n\t\t&lt;key&gt;Program&lt;\/key&gt;\n\t\t&lt;string&gt;\/Users\/Me\/Scripts\/backup.sh&lt;\/string&gt;\n\t\t&lt;key&gt;KeepAlive&lt;\/key&gt;\n\t\t&lt;dict&gt;\n\t\t\t&lt;key&gt;PathState&lt;\/key&gt;\n\t\t\t&lt;dict&gt;\n\t\t\t\t&lt;key&gt;\/var\/log\/mysql.pid&lt;\/key&gt;\n\t\t\t\t&lt;true\/&gt;\n\t\t\t&lt;\/dict&gt;\n\t\t&lt;\/dict&gt;\n\t\t&lt;key&gt;ThrottleInterval&lt;\/key&gt;\n\t\t&lt;integer&gt;3600&lt;\/integer&gt;\n\t&lt;\/dict&gt;\n&lt;\/plist&gt;<\/p><\/div><\/div><\/div><div class=\"section\" id=\"rcp-creating-a-network-server\"><p>launchd makes it very easy to create servers. In fact, launchd can turn any program reading from standard input into a server.<\/p><p>In this recipe we will turn Unix cat(1), a program which copies STDIN to STDOUT, into an network server listening on port 8888 on our machine, copying everything it receives to the file \/tmp\/test.stdout.<\/p><div class=\"tabset\"><div class=\"tab\" title=\"Plain Text\"><p>&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;\n&lt;!DOCTYPE plist PUBLIC \"-\/\/Apple\/\/DTD PLIST 1.0\/\/EN\" \"http:\/\/www.apple.com\/DTDs\/PropertyList-1.0.dtd\"&gt;\n&lt;plist version=\"1.0\"&gt;\n&lt;dict&gt;\n\t&lt;key&gt;Label&lt;\/key&gt;\n\t&lt;string&gt;com.example.srv&lt;\/string&gt;\n\t&lt;key&gt;Program&lt;\/key&gt;\n\t&lt;string&gt;\/bin\/cat&lt;\/string&gt;\n\t&lt;key&gt;Sockets&lt;\/key&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;sock1&lt;\/key&gt;\n\t\t&lt;dict&gt;\n\t\t\t&lt;key&gt;SockNodeName&lt;\/key&gt;\n\t\t\t&lt;string&gt;localhost&lt;\/string&gt;\n\t\t\t&lt;key&gt;SockPassive&lt;\/key&gt;\n\t\t\t&lt;true\/&gt;\n\t\t\t&lt;key&gt;SockServiceName&lt;\/key&gt;\n\t\t\t&lt;string&gt;8888&lt;\/string&gt;\n\t\t&lt;\/dict&gt;\n\t&lt;\/dict&gt;\n\t&lt;key&gt;StandardOutPath&lt;\/key&gt;\n\t&lt;string&gt;\/tmp\/test.stdout&lt;\/string&gt;\n\t&lt;key&gt;inetdCompatibility&lt;\/key&gt;\n\t&lt;dict&gt;\n\t\t&lt;key&gt;Wait&lt;\/key&gt;\n\t\t&lt;false\/&gt;\n\t&lt;\/dict&gt;\n&lt;\/dict&gt;\n&lt;\/plist&gt;<\/p><\/div><\/div><p>After loading our job definition we will do a quick test.<\/p><ol><li>Open Terminal.app and establish a network connection to port 8888: telnet localhost 8888<\/li>\n\t\t\t\t\t<li>Open another Terminal.app window and have a look at our job's STDOUT: tail -f \/tmp\/test.out<\/li>\n\t\t\t\t\t<li>Type some text in the first window, press enter.<\/li>\n\t\t\t\t\t<li>The text you entered in the first window appears in the second one.<\/li>\n\t\t\t\t<\/ol><\/div><\/div>","created_at":"2019-02-07T00:45:32+0000","updated_at":"2019-02-07T00:45:37+0000","annotations":[],"mimetype":"text\/html","reading_time":40,"domain_name":"www.launchd.info","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":814,"title":"graphql\/graphiql","url":"https:\/\/github.com\/graphql\/graphiql\/tree\/master\/example","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>This example uses the version of GraphiQL found in the parent directory, rather\nthan depending on npm, so that it is easier to test new changes. In order to use\nthe compiled version of GraphiQL, first run build in the parent directory before\ninstalling and starting the example.<\/p>\n<ol><li>Run <code>npm install &amp;&amp; npm run build<\/code> in the parent directory<\/li>\n<li>Navigate to this directory (example) in Terminal<\/li>\n<li><code>npm install<\/code><\/li>\n<li><code>npm start<\/code><\/li>\n<li>Open your browser to the address listed in your console. e.g. <code>Started on http:\/\/localhost:49811\/<\/code><\/li>\n<\/ol><\/article>","created_at":"2019-02-06T22:03:06+0000","updated_at":"2019-02-06T22:03:06+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/12972006?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["equity","finance"],"is_public":false,"id":813,"title":"The Basics of Stock Options Explained for Startup Employees | Carta","url":"https:\/\/carta.com\/blog\/equity-101-stock-option-basics\/","content":"<p><em>The following post explains the basics of Startup\u00a0Stock Options from <a href=\"https:\/\/carta.com\" target=\"_blank\" rel=\"noreferrer noopener\">Carta\u2019s<\/a> presentation: \u201cUnderstand Your Options: Equity 101 for Startup Employees\u201d. The remaining sections are covered in the rest of this 3-part blog post. You can follow us on Twitter <a href=\"https:\/\/twitter.com\/cartainc?lang=en\" target=\"_blank\" rel=\"noreferrer noopener\">@cartainc<\/a> for updates on upcoming events.<\/em><\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1_C6wnf9EvMzoDxNA46D_dqQ-1521137021749.png\" alt=\"Startup Stock Options 101\" \/><\/figure><p>If you work for a startup, stock option grants are your way of sharing in the company\u2019s success. But very few startup employees understand what stock options really are and how they work. By the end of this post, you\u2019ll understand stock options better than most of your coworkers, and in some cases, your employers.<\/p><hr class=\"wp-block-separator\" \/><h3 id=\"part1stockoptionbasics\">Part 1: Stock Option for Startups 101<\/h3><p>We\u2019ll be using Kerri as our example. She just joined a new startup called Meetly, and she was promised equity stock options in her offer letter. This is a picture of Meetly\u2019s offer letter:<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1-VK8pxkToDMRpUzsc26BIZA.png\" alt=\"Stock Options Offer Letter\" \/><\/figure><p>Like most <a href=\"https:\/\/carta.com\/blog\/a-better-offer-letter\/\" target=\"_blank\" rel=\"noreferrer noopener\">offer letters<\/a>, it only tells Kerri the number of stock options she was granted. As you can see, Meetly has promised Kerri 100 shares of stock option grants. The letter does not tell her what to do with those option grants, what kind of option grants she has, or how much they are worth. There\u2019s really no transparency when it comes to employee stock options, and unfortunately, this is how most companies do it today.<\/p><p>There are four things that Kerri should consider when she receives this letter. These four things are what every startup employee should make sure they know about stock options when joining a company:<\/p><ol><li>Types of Stock Options for Startups<\/li><li>Stock Option Agreement<\/li><li>Vesting<\/li><li>Termination<\/li><\/ol><h3>Types of stock options for startups<\/h3><p>A stock option is the right to buy a set number of shares at a fixed price. In Kerri\u2019s case, that fixed price is $1. The hope is that the value of the stock goes up, but the price stays the same, which means you make money on the difference.<\/p><p>There are 2 types of employee stock options: ISOs and NSOs. ISOs is short for incentive stock options, and NSOs is short of non-qualified stock options. If you\u2019re an employee like Kerri, you probably have ISOs. If you\u2019re a consultant, you\u2019ll have NSOs. The main difference here is that ISOs have tax benefits.<\/p><h3>Stock option agreement\u00a0<\/h3><p>Now for Kerri\u2019s stock option agreement. The big distinction to make here is that an offer letter is NOT legally binding. We see a lot of companies come to Carta to issue stock option grants to their employees, and sometimes those options were promised to the employees months ago. Until you\u2019ve received and signed an option agreement, you technically don\u2019t own stock options.<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1-zl0CjXjcbl_N4ecnyEpVdw.png\" alt=\"img 7\" \/><\/figure><p>This is what Meetly\u2019s Option Agreement looks like. Yours probably looks very similar. If you didn\u2019t receive one, you can ask your company and they should be able to provide it. If you just joined in the last month or two, it is possible that the board has not approved your options yet, in which case you should receive this shortly after the next board meeting.<\/p><p>It\u2019s also worth mentioning that if your company is on <a href=\"https:\/\/carta.com\">Carta<\/a>, you won\u2019t receive a paper version of the stock option agreement. Don\u2019t panic! Instead, you checked a box agreeing to the terms when you accepted online. You can log into your portfolio to view and print the actual agreement.<\/p><h3>Vesting<\/h3><p><a href=\"https:\/\/carta.com\/blog\/what-is-stock-vesting\/\" target=\"_blank\" rel=\"noopener\">Vesting means<\/a> that you have to earn your employee stock options over time. Companies do this because they want employees that will stay with them and contribute to the company\u2019s success over many years. If they didn\u2019t have vesting, an employee could accept an offer, work at the company for a month, buy a bunch of the company\u2019s stock, and then quit. Vesting prevents that.<\/p><p>Kerri\u2019s vesting schedule is outlined in her offer letter and her option agreement. She has what\u2019s called traditional vesting. There all kinds of vesting schedules, but this is by far the most common.<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g.png\" alt=\"Stock Option Vesting Schedule\" class=\"wp-image-5379\" srcset=\"https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g.png 1000w, https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g-300x225.png 300w, https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g-768x575.png 768w, https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g-24x18.png 24w, https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g-48x36.png 48w, https:\/\/carta.com\/wp-content\/uploads\/2015\/12\/1-6pIthXMlD-2Rjk8PpiC8g-72x54.png 72w\" \/><\/figure><p>There are two parts to a traditional vesting schedule. The first part is called a \u201ccliff.\u201d A cliff is really just the first chunk of shares that vest, and you gain the right to buy. Kerri has a one year cliff, which is standard. After one year of working at Meetly, she can buy \u00bc of her options, or 25 shares.<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1-c1trUFMM1_6ekxBBcldLxQ.png\" alt=\"img10\" \/><\/figure><p>The second part is the total length of the schedule. This outlines how often, and for how long, Kerri\u2019s shares will vest. After Kerri reaches her cliff, her remaining shares will continue to vest for 36 months, or 2 shares each month.<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1-czW3Q9ys-KoxO4PVKI7jpg.png\" alt=\"img11\" \/><\/figure><p>If she leaves the company, her shares will stop vesting immediately and she will only have the right to buy the shares that have vested as of that date.<\/p><h3>Termination<\/h3><p>What happens if Kerri leaves the company? This is known as a termination.<\/p><p>The two most common scenarios are if Kerri quits or she is fired by the company. Startups generally have a higher turnover rate than larger companies, so terminations are fairly common.<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1-pfF8l77WEpKb02L5wcy-7Q.png\" alt=\"img13\" \/><\/figure><p>Remember that if Kerri leaves, she only has the right to buy her vested options; she only maintains this right for a set window of time. In the cases of both voluntary and involuntary termination (quitting or getting fired), that window is typically 30 days to 90 days.<\/p><p>The other two scenarios we see outlined in option plans are death and disability. If Kerri dies or leaves the company due to a disability, the window is generally extended to 6 months or 1 year.<\/p><p>It\u2019s really important to not forget about these windows because your company is not obligated to remind you when you leave\u200a\u2014\u200athey usually only tell you in your option agreement when you first join.<\/p><p>Below is a quick recap of what we\u2019ve covered:<\/p><figure class=\"wp-block-image\"><img src=\"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1-NKT45zqQQFg0gfw7ohUk0g.png\" alt=\"img14\" \/><\/figure><ol><li>We described the two kinds of employee stock options\u200a\u2014\u200aISOs and NSOs. Again, ISOs are issued to employees and have tax benefits.<\/li><li>We detailed Stock Option Agreements: the one document you want to make sure you receive and sign.<\/li><li>We covered vesting schedules and how companies incentivize employees to stay longer.<\/li><li>Finally, we provided information about termination and remembering to exercise in the 30-to-90-day window after leaving a company.<\/li><\/ol><p>These are the four things that every startup employee should think about when they receive their offer letter and join a new company. We cover Stock Option Economics and how to think about what your options are actually worth in Part 2.<\/p><p>Read <a href=\"https:\/\/carta.com\/blog\/equity-101-stock-economics\/\">Equity 101: Part 2<\/a><\/p><p>Read <a href=\"https:\/\/carta.com\/blog\/equity-101-exercising-and-taxes\/\" target=\"_blank\">Equity 101: Part 3<\/a><\/p><hr class=\"wp-block-separator\" \/>","created_at":"2019-02-06T19:50:20+0000","updated_at":"2019-02-06T19:50:28+0000","published_at":"2015-12-17T18:21:00+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"carta.com","preview_picture":"https:\/\/carta.com\/wp-content\/uploads\/2016\/12\/1_C6wnf9EvMzoDxNA46D_dqQ-1521137021749.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":812,"title":"Journalism generated by machine is on the rise","url":"https:\/\/news.ycombinator.com\/item?id=19084260","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D19084260\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"19084260\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/www.nytimes.com\/2019\/02\/05\/business\/media\/artificial-intelligence-journalism-robots.html\" class=\"storylink\">Journalism generated by machine is on the rise<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=nytimes.com\">nytimes.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        111 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=pseudolus\" class=\"hnuser\">pseudolus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=19084260\">14 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=19084260&amp;goto=item%3Fid%3D19084260\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Journalism%20generated%20by%20machine%20is%20on%20the%20rise&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Journalism%20generated%20by%20machine%20is%20on%20the%20rise\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=19084260&amp;auth=37b3ca57740f6c7bc8c880ffea85203a88d633a5\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=19084260\">49\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/www.ycombinator.com\/apply\/\">\n        Applications are open for YC Summer 2019\n      <\/a><\/div><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-02-06T01:13:23+0000","updated_at":"2019-02-06T01:13:23+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":811,"title":"Pioneer - A home for the ambitious outsiders of the world","url":"https:\/\/pioneer.app\/","content":"<section class=\"hero pb-2\"><div class=\"container hero-container\"><div class=\"row\"><div class=\"col-xl-8 col-lg-7 pr-lg-5 header-column\"><div class=\"logo-and-wordmark d-inline-flex align-items-center\"><div class=\"logo bg-yellow\">Created with Sketch.<\/div><div class=\"wordmark\">Created with Sketch.<\/div><\/div><p>A home for the ambitious outsiders of the world<\/p><p>We\u2019re building a community of creative young people working on interesting projects around the globe.<\/p><div class=\"pt-3 font-weight-400\"><a class=\"text-gray-E d-flex align-items-start\" href=\"https:\/\/pioneer.app\/blog\/meet-the-pioneers-take-2\/\"><p>New<\/p>\n<p>Meet the winners of the second Pioneer Tournament \u00bb<\/p>\n<\/a><\/div><\/div><div class=\"col-lg-5 col-xl-4 mt-3 mt-lg-0 pl-lg-4\"><div class=\"pl-lg-4\"><p>Join the tournament<\/p><p>Apply with any project you need help with. After 30 days, the best players will receive $1,000, a round-trip ticket to Silicon Valley, and access to mentorship from some of the world\u2019s most successful individuals.<\/p><a href=\"https:\/\/pioneer.app\/tournament\"><p>Get started<\/p>\n<\/a><p>Deadline February 10th 11:59pm PST<\/p><\/div><\/div><\/div><\/div><\/section><section class=\"value-props bg-yellow pb-2\"><div class=\"container py-4\"><div class=\"row no-gutters\"><div class=\"col-md-6 d-md-flex\"><div class=\"pb-5 pl-md-5 pr-md-1 pt-5 value-prop\"><p>Learn from the best<\/p><p>Pioneer has a network of experts across different fields.\nExperts both review applications and provide advice to Pioneers.\nAfter 3 months Pioneers present their work to experts on a global stage.<\/p><\/div><\/div><\/div><div class=\"mt-md-4 text-md-center\"><a href=\"https:\/\/pioneer.app\/tournament\"><p>Learn more<\/p>\n<\/a><\/div><\/div><\/section><section class=\"experts\"><div class=\"container home-shared-experts\"><div class=\"row\"><div class=\"col-lg-6 col-md-12 pt-2 pb-5\"><p>Learn from the world\u2019s experts<\/p><p>Get advice from industry experts.\nPioneer has a network of experts across every vertical.\nExperts review applications and provide advice to Pioneers.<\/p><\/div><div class=\"col-lg-3 col-md-4 expert px-md-4 mb-md-5 mb-4 pt-md-0 pt-1\"><img srcset=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/hollis-robbins@2x-be5617c3c82ff6b3d241b00877a26f9e9c313a5c44dd0b6c965f4a0f875a1a1c.jpg 2x\" src=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/hollis-robbins-a337216ec5bab5587ee231c84789e4a9c8b876c313d15d30d53718823530c558.jpg\" alt=\"image\" \/><div class=\"description d-md-block d-inline-block mt-md-3 ml-md-0 ml-3\"><p><a target=\"_blank\" class=\"text-black link--dotted\" href=\"https:\/\/en.wikipedia.org\/wiki\/Hollis_Robbins\">Hollis Robbins<\/a><\/p><p>Art<\/p><p>Dean of Arts &amp; Humanities, Sonoma State University<\/p><\/div><\/div><div class=\"col-lg-3 col-md-4 expert px-md-4 mb-md-5 mb-4 pt-md-0 pt-1\"><img srcset=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/tyler-cowen@2x-78e545a0738639216184be003cfc973e44e3b0d25ea718d24dd7736aae97b701.jpg 2x\" src=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/tyler-cowen-0797d0888097c7579fa3912f428dbc8c3cbf7e7bb9ac3481ee65f6f3dbfba040.jpg\" alt=\"image\" \/><div class=\"description d-md-block d-inline-block mt-md-3 ml-md-0 ml-3\"><p><a target=\"_blank\" class=\"text-black link--dotted\" href=\"https:\/\/marginalrevolution.com\/\">Tyler Cowen<\/a><\/p><p>Economics<\/p><p>Economics Professor, George Mason University<\/p><\/div><\/div><div class=\"col-lg-3 col-md-4 expert px-md-4 mb-md-5 mb-4 pt-md-0 pt-1\"><img srcset=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/patrick-collison@2x-b7248bd49dca32a865f3b573ce04037b7ddbcbb88023b6fa2bde6f4e3fd1637b.jpg 2x\" src=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/patrick-collison-8ab345e4e8c66861f71ae7d2b90ae36a2430a6de3e987bb28be3d4d66471f4be.jpg\" alt=\"image\" \/><div class=\"description d-md-block d-inline-block mt-md-3 ml-md-0 ml-3\"><p><a target=\"_blank\" class=\"text-black link--dotted\" href=\"https:\/\/twitter.com\/patrickc\">Patrick Collison<\/a><\/p><p>Fintech<\/p><p>Cofounder &amp; CEO, Stripe<\/p><\/div><\/div><div class=\"col-lg-3 col-md-4 expert px-md-4 mb-md-5 mb-4 pt-md-0 pt-1\"><img srcset=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/laura-deming@2x-c5e4a789ea9151dc8506b190cb5bca87fca7655e6b40490fcddb52dc354d053a.jpg 2x\" src=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/laura-deming-e81b0fa62d6819ef85f8e61a5b5f1545101ff67c6d6e81346b49894a1d7e1314.jpg\" alt=\"image\" \/><div class=\"description d-md-block d-inline-block mt-md-3 ml-md-0 ml-3\"><p><a target=\"_blank\" class=\"text-black link--dotted\" href=\"https:\/\/www.ldeming.com\/\">Laura Deming<\/a><\/p><p>Longevity<\/p><p>Founder, The Longevity Fund<\/p><\/div><\/div><div class=\"col-lg-3 col-md-4 expert px-md-4 mb-md-5 mb-4 pt-md-0 pt-1\"><img srcset=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/marc-andreessen@2x-bbe17126422a5df77774db64d9170e2d382dc8fad6b377c819543aa14551dd30.jpg 2x\" src=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/marc-andreessen-e85f881d97b81280427a51bac9bd03f44c4382f84f35f6c01ffd6250997e135e.jpg\" alt=\"image\" \/><div class=\"description d-md-block d-inline-block mt-md-3 ml-md-0 ml-3\"><p><a target=\"_blank\" class=\"text-black link--dotted\" href=\"https:\/\/twitter.com\/pmarca\">Marc Andreessen<\/a><\/p><p>Consumer<\/p><p>Cofounder, Andreessen Horowitz<\/p><\/div><\/div><div class=\"col-lg-3 col-md-4 expert px-md-4 mb-md-5 mb-4 pt-md-0 pt-1\"><img srcset=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/balaji-srinivasan@2x-8f9c51cb2caf5484892302188f8352596e66b324e882de3370c54a06b34b1a12.jpg 2x\" src=\"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/views\/landing\/experts\/balaji-srinivasan-411e8dcb3e2a2f211788eb1f60a727281c64c06245f160121432b315f256ad13.jpg\" alt=\"image\" \/><div class=\"description d-md-block d-inline-block mt-md-3 ml-md-0 ml-3\"><p><a target=\"_blank\" class=\"text-black link--dotted\" href=\"https:\/\/twitter.com\/balajis\">Balaji Srinivasan<\/a><\/p><p>Blockchain &amp; Cryptocurrency<\/p><p>CTO, Coinbase<\/p><\/div><\/div><\/div><div class=\"text-md-center\"><a class=\"link--inherit-color\" href=\"https:\/\/pioneer.app\/experts\"><p>See all<\/p>\n<\/a><\/div><\/div><\/section><section class=\"mission py-2\"><section class=\"mailing-list\"><div class=\"js-home-landing-mailing-list container js-view\"><div class=\"row\"><div class=\"col-lg-8 col-md-7\"><p>Stay up to date<\/p><p>Not likely to apply but interested in how this program fares?\nEnter your email to receive our updates.<\/p><\/div><\/div><\/div><\/section><\/section>","created_at":"2019-02-05T12:54:56+0000","updated_at":"2019-02-05T12:54:56+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"pioneer.app","preview_picture":"https:\/\/d1zhjck9imo0be.cloudfront.net\/assets\/twitter-image-daca5d7cfe40083ed2b9095f5461fd58ced052ef3ed433cad462bfea3d9de64e.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":810,"title":"How to use matrix.org and vector.im as an IRC client","url":"https:\/\/gist.github.com\/fstab\/ce805d3001600ac147b79d413668770d","content":"<h1><a id=\"user-content-how-to-use-matrixorg-and-vectorim-as-an-irc-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-use-matrixorg-and-vectorim-as-an-irc-client\"><\/a>How to use <a href=\"http:\/\/matrix.org\/\" rel=\"nofollow\">matrix.org<\/a> and <a href=\"https:\/\/vector.im\/\" rel=\"nofollow\">vector.im<\/a> as an <a href=\"https:\/\/webchat.freenode.net\/\" rel=\"nofollow\">IRC<\/a> client<\/h1>\n<h2><a id=\"user-content-how-to-join-an-irc-channel-on-matrixorg\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-join-an-irc-channel-on-matrixorg\"><\/a>How to Join an IRC Channel on matrix.org<\/h2>\n<p>Join the room <code>#freenode_&lt;#channel&gt;:matrix.org<\/code>, replacing <code>&lt;#channel&gt;<\/code> with the name of the IRC channel. For example, in order to join the <code>#prometheus<\/code> IRC channel, join the room <code>#freenode_#prometheus:matrix.org<\/code> on matrix.org.<\/p>\n<p>In vector.im, rooms can be joined with the directory symbol on the bottom left.<\/p>\n<h2><a id=\"user-content-how-to-use-a-freenode-registered-nickname-on-matrixorg\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-use-a-freenode-registered-nickname-on-matrixorg\"><\/a>How to Use a Freenode-Registered Nickname on matrix.org<\/h2>\n<p>As described <a href=\"https:\/\/groups.google.com\/d\/topic\/prometheus-developers\/2Prtj6ALoPA\/discussion\" rel=\"nofollow\">here<\/a>: If you have a Freenode-registered nickname and want to use it from Matrix, you can do it like this:<\/p>\n<ol><li>Leave with your nick on IRC<\/li>\n<li>On Matrix, open a new chat (with vector.im this is the \"+\" sign on the bottom left)<\/li>\n<li>Invite <code>@appservice-irc:matrix.org<\/code> to the new chat<\/li>\n<li>Say <code>!nick &lt;yournick&gt;<\/code><\/li>\n<li>A request for new chat with <code>NickServ<\/code> will pop up<\/li>\n<li>Accept it and identify yourself with your Freenode password (<code>identify &lt;password&gt;<\/code>)<\/li>\n<\/ol><p>Voila, you'll be using your Freenode-registered nick through Matrix, being able to use multiple devices simultaneously with shared history and all!<\/p>\n<h2><a id=\"user-content-how-to-start-private-irc-chats-like-msg-or-query\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-start-private-irc-chats-like-msg-or-query\"><\/a>How to Start Private IRC Chats, like \/msg or \/query<\/h2>\n<ol><li>On Matrix, open a new chat (with vector.im this is the \"+\" sign on the bottom left)<\/li>\n<li>Invite <code>@freenode_&lt;nick&gt;:matrix.org<\/code> to the new chat (replace <code>&lt;nick&gt;<\/code> with the nick name. For example, to start a chat with <code>NickServ<\/code>, invite <code>@freenode_NickServ:matrix.org<\/code>)<\/li>\n<\/ol><p>This works for nicks that are known to matrix.org, i.e. if the nick has never joined an IRC channel that was bridged to matrix it won't work.<\/p>","created_at":"2019-02-05T02:57:55+0000","updated_at":"2019-02-05T02:57:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"gist.github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/330535?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":809,"title":"Learn Svelte","url":"https:\/\/svelte.technology\/guide","content":"<section id=\"introduction\" class=\"svelte-1kk9xxo\"><h3 id=\"what-is-svelte-\">What is Svelte?<\/h3><p>Svelte is a tool for building fast web applications.<\/p><p>It is similar to JavaScript frameworks such as React, Angular, Vue, and Ractive, which all share a goal of making it easy to build slick interactive user interfaces.<\/p><p>But there's a crucial difference: Svelte converts your app into ideal JavaScript at <em>build time<\/em>, rather than interpreting your application code at <em>run time<\/em>. This means you don't pay the performance cost of the framework's abstractions, and you don't incur a penalty when your app first loads.<\/p><p>You can build your entire app with Svelte, or you can add it incrementally to an existing codebase. You can also ship components as standalone packages that work anywhere, without the overhead of a dependency on a conventional framework.<\/p><p><a href=\"https:\/\/svelte.technology\/blog\/frameworks-without-the-framework\">Read the introductory blog post<\/a> to learn more about Svelte's goals and philosophy.<\/p><h3 id=\"understanding-svelte-components\">Understanding Svelte components<\/h3><p>In Svelte, an application is composed from one or more <em>components<\/em>. A component is a reusable self-contained block of code that encapsulates markup, styles and behaviours that belong together.<\/p><p>Like Ractive and Vue, Svelte promotes the concept of <em>single-file components<\/em>: a component is just an <code>.html<\/code> file. Here's a simple example:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1igzsq8\">REPL<\/a>App.html<pre>&lt;h1&gt;Hello {name}!&lt;\/h1&gt;<\/pre><\/div><blockquote>\n<p>Wherever you see <strong>REPL<\/strong> links, click through for an interactive example<\/p>\n<\/blockquote><p>Svelte turns this into a JavaScript module that you can import into your app:<\/p><div class=\"code-block named\">main.js<pre>import App from '.\/App.html';\nconst app = new App({\n\ttarget: document.querySelector('main'),\n\tdata: { name: 'world' }\n});\n\/\/ change the data associated with the template\napp.set({ name: 'everybody' });\n\/\/ detach the component and clean everything up\napp.destroy();<\/pre><\/div><p>Congratulations, you've just learned about half of Svelte's API!<\/p><h3 id=\"getting-started\">Getting started<\/h3><p>Normally, this is the part where the instructions might tell you to add the framework to your page as a <code>&lt;script&gt;<\/code> tag. But because Svelte runs at build time, it works a little bit differently.<\/p><p>The best way to use Svelte is to integrate it into your build system \u2013 there are plugins for Rollup, Browserify, Gulp and others, with more on the way. See <a href=\"https:\/\/github.com\/sveltejs\/svelte\/#svelte\">here<\/a> for an up-to-date list.<\/p><blockquote>\n<p>You will need to have <a href=\"https:\/\/nodejs.org\/en\/\">Node.js<\/a> installed, and have some familiarity with the command line<\/p>\n<\/blockquote><h4 id=\"getting-started-using-the-repl\">Getting started using the REPL<\/h4><p>Going to the <a href=\"https:\/\/svelte.technology\/repl\">REPL<\/a> and pressing the <em>download<\/em> button on any of the examples will give you a .zip file containing everything you need to run that example locally. Just unzip it, <code>cd<\/code> to the directory, and run <code>npm install<\/code> and <code>npm run dev<\/code>. See <a href=\"https:\/\/svelte.technology\/blog\/the-easiest-way-to-get-started\">this blog post<\/a> for more information.<\/p><h4 id=\"getting-started-using-degit\">Getting started using degit<\/h4><p><a href=\"https:\/\/github.com\/Rich-Harris\/degit\">degit<\/a> is a tool for creating projects from templates stored in git repos. Install it globally...<\/p><div class=\"code-block\"><pre>npm install -g degit<\/pre><\/div><p>...then you can use it to spin up a new project:<\/p><div class=\"code-block\"><pre>degit sveltejs\/template my-new-project\ncd my-new-project\nnpm install\nnpm run dev<\/pre><\/div><p>You can use any git repo you like \u2014 these are the 'official' templates:<\/p><ul><li><a href=\"https:\/\/github.com\/sveltejs\/template\">sveltejs\/template<\/a> \u2014 this is what you get by downloading from the REPL<\/li>\n<li><a href=\"https:\/\/github.com\/sveltejs\/template-webpack\">sveltejs\/template-webpack<\/a> \u2014 similar, but uses <a href=\"https:\/\/webpack.js.org\/\">webpack<\/a> instead of <a href=\"https:\/\/rollupjs.org\/guide\/en\">Rollup<\/a><\/li>\n<\/ul><h4 id=\"getting-started-using-the-cli\">Getting started using the CLI<\/h4><p>Svelte also provides a Command Line Interface, but it's not recommended for production use. The CLI will compile your components to standalone JavaScript files, but won't automatically recompile them when they change, and won't deduplicate code shared between your components. Use one of the above methods instead.<\/p><p>If you've installed <code>svelte<\/code> globally, you can use <code>svelte --help<\/code> for a complete list of options. Some examples of the more common operations are:<\/p><div class=\"code-block\"><pre># Generate a JavaScript module from MyComponent.html\nsvelte compile MyComponent.html &gt; MyComponent.js\nsvelte compile -i MyComponent.html -o MyComponent.js\n# Generate a UMD module from MyComponent.html, inferring its name from the filename ('MyComponent')\nsvelte compile -f umd MyComponent.html &gt; MyComponent.js\n# Generate a UMD module, specifying the name\nsvelte compile -f umd -n CustomName MyComponent.html &gt; MyComponent.js\n# Compile all .html files in a directory\nsvelte compile -i src\/components -o build\/components<\/pre><\/div><blockquote>\n<p>You can also use <a href=\"https:\/\/medium.com\/@maybekatz\/introducing-npx-an-npm-package-runner-55f7d4bd282b\">npx<\/a> to use the CLI without installing Svelte globally \u2014 just prefix your command with <code>npx<\/code>: <code>npx svelte compile ...<\/code><\/p>\n<\/blockquote><\/section><section id=\"component-api\" class=\"svelte-1kk9xxo\"><p>As we saw above, you create a component instance with the <code>new<\/code> keyword:<\/p><div class=\"code-block named\">main.js<pre>import MyComponent from '.\/MyComponent.html';\nconst component = new MyComponent({\n\t\/\/ `target` is the only required option. This is the\n\t\/\/ DOM element your component will be appended to\n\ttarget: document.querySelector('main'),\n\t\/\/ `anchor` is optional.\n\t\/\/ The component is inserted immediately before this\n\t\/\/ DOM element, which must be a child of `target`\n\tanchor: document.querySelector('main #child'),\n\t\/\/ `data` is optional.\n\t\/\/ A component can have default data \u2013 we'll learn about that later.\n\tdata: {\n\t\tquestions: [\n\t\t\t'life',\n\t\t\t'the universe',\n\t\t\t'everything'\n\t\t],\n\t\tanswer: 42\n\t}\n});<\/pre><\/div><p>Every Svelte component instance has a small number of methods you can use to control it, in addition to any <a href=\"https:\/\/svelte.technology\/guide#custom-methods\">custom methods<\/a> you add.<\/p><h3 id=\"component-set-state-\">component.set(state)<\/h3><p>This updates the component's state with the new values provided and causes the DOM to update. <code>state<\/code> must be a plain old JavaScript object (POJO). Any properties <em>not<\/em> included in <code>state<\/code> will remain as they were.<\/p><div class=\"code-block\"><pre>component.set({\n\tquestions: [\n\t\t'why is the sky blue?',\n\t\t'how do planes fly?',\n\t\t'where do babies come from?'\n\t],\n\tanswer: 'ask your mother'\n});<\/pre><\/div><h3 id=\"component-get-\">component.get()<\/h3><p>Returns the component's current state:<\/p><div class=\"code-block\"><pre>const { questions, answer } = component.get();\nconsole.log(answer); \/\/ 'ask your mother'<\/pre><\/div><p>This will also retrieve the value of <a href=\"https:\/\/svelte.technology\/guide#computed-properties\">computed properties<\/a>.<\/p><blockquote>\n<p>Previous versions of Svelte allowed you to specify a key to retrieve a specific value \u2014 this was removed in version 2.<\/p>\n<\/blockquote><h3 id=\"component-on-eventname-callback-\">component.on(eventName, callback)<\/h3><p>Allows you to respond to <em>events<\/em>:<\/p><div class=\"code-block\"><pre>const listener = component.on('thingHappened', event =&gt; {\n\tconsole.log(`A thing happened: ${event.thing}`);\n});\n\/\/ some time later...\nlistener.cancel();<\/pre><\/div><p>Each component has three built-in events, corresponding to their <a href=\"https:\/\/svelte.technology\/guide#lifecycle-hooks\">lifecycle hooks<\/a>:<\/p><div class=\"code-block\"><pre>component.on('state', ({ changed, current, previous }) =&gt; {\n\tconsole.log('state changed', current);\n});\ncomponent.on('update', ({ changed, current, previous }) =&gt; {\n\tconsole.log('DOM updated after state change', current);\n});\ncomponent.on('destroy', () =&gt; {\n\tconsole.log('this component is being destroyed');\n});<\/pre><\/div><h3 id=\"component-fire-eventname-event-\">component.fire(eventName, event)<\/h3><p>The companion to <code>component.on(...)<\/code>:<\/p><div class=\"code-block\"><pre>component.fire('thingHappened', {\n\tthing: 'this event was fired'\n});<\/pre><\/div><p>At first glance <code>component.on(...)<\/code> and <code>component.fire(...)<\/code> aren't particularly useful, but it'll become more so when we learn about <a href=\"https:\/\/svelte.technology\/guide#nested-components\">nested components<\/a> and <a href=\"https:\/\/svelte.technology\/guide#component-events\">component events<\/a>.<\/p><h3 id=\"component-destroy-\">component.destroy()<\/h3><p>Removes the component from the DOM and removes any event listeners that were created. This will also fire a <code>destroy<\/code> event:<\/p><div class=\"code-block\"><pre>component.on('destroy', () =&gt; {\n\talert('goodbye!'); \/\/ please don't do this\n});\ncomponent.destroy();<\/pre><\/div><h3 id=\"component-options\">component.options<\/h3><p>The options used to instantiate the component are available in <code>component.options<\/code>.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=3tlfym\">REPL<\/a>App.html<pre>Check the console.\n&lt;script&gt;\n\texport default {\n\t\toncreate() {\n\t\t\tconsole.log(this.options);\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>This gives you access to standard options like <code>target<\/code> and <code>data<\/code>, but can also be used to access any other custom options you may choose to implement for your component.<\/p><h3 id=\"component-root\">component.root<\/h3><p>In <a href=\"https:\/\/svelte.technology\/guide#nested-components\">nested components<\/a>, each component has a <code>root<\/code> property pointing to the top-level root component \u2013 that is, the one instantiated with <code>new MyComponent({...})<\/code>.<\/p><blockquote>\n<p>Earlier versions of Svelte had a <code>component.observe(...)<\/code> method. This was removed in version 2, in favour of the <code>onstate<\/code> <a href=\"https:\/\/svelte.technology\/guide#lifecycle-hooks\">lifecycle hook<\/a>, but is still available via <a href=\"https:\/\/github.com\/sveltejs\/svelte-extras\">svelte-extras<\/a>.<\/p>\n<\/blockquote><\/section><section id=\"template-syntax\" class=\"svelte-1kk9xxo\"><p>Rather than reinventing the wheel, Svelte templates are built on foundations that have stood the test of time: HTML, CSS and JavaScript. There's very little extra stuff to learn.<\/p><blockquote>\n<p>Svelte version 1 had a slightly different template syntax. You can upgrade older components automatically using <a href=\"https:\/\/github.com\/sveltejs\/svelte-upgrade\">svelte-upgrade<\/a>.<\/p>\n<\/blockquote><p>Tags allow you to bind data to your template. Whenever your data changes (for example after <code>component.set(...)<\/code>), the DOM updates automatically. You can use any JavaScript expression in templates, and it will also automatically update:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=z2qzuf\">REPL<\/a>App.html<pre>&lt;p&gt;{a} + {b} = {a + b}&lt;\/p&gt;<\/pre><\/div><p>You can also use tags in attributes:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=me2otu\">REPL<\/a>App.html<pre>&lt;h1 style=\"color: {color};\"&gt;{color}&lt;\/h1&gt;\n&lt;p hidden={hideParagraph}&gt;You can hide this paragraph.&lt;\/p&gt;<\/pre><\/div><p><a href=\"https:\/\/www.w3.org\/TR\/html5\/infrastructure.html#sec-boolean-attributes\">Boolean attributes<\/a> like <code>hidden<\/code> will be omitted if the tag expression evaluates to false. Attributes will be removed from the element if their value is <code>undefined<\/code> or <code>null<\/code>.<\/p><h3 id=\"html\">HTML<\/h3><p>Ordinary tags render expressions as plain text. If you need your expression interpreted as HTML, wrap it in a special <code>@html<\/code> tag:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=9xfu65\">REPL<\/a>App.html<pre>&lt;p&gt;This HTML: {content}&lt;\/p&gt;\n&lt;p&gt;Renders as: {@html content}&lt;\/p&gt;<\/pre><\/div><p>As with regular tags, you can use any JavaScript expression in HTML tags, and it will automatically update the document when your data changes.<\/p><blockquote>\n<p>HTML is <strong>not<\/strong> sanitized before it is rendered! If you are displaying user input, you are responsible for first sanitizing it. Not doing so potentially opens you up to XSS attacks.<\/p>\n<\/blockquote><h3 id=\"if-blocks\">If blocks<\/h3><p>Control whether or not part of your template is rendered by wrapping it in an if block.<\/p><div class=\"code-block named\">App.html<pre>{#if user.loggedIn}\n\t&lt;a href=\"\/logout\"&gt;log out&lt;\/a&gt;\n{\/if}\n{#if !user.loggedIn}\n\t&lt;a href=\"\/login\"&gt;log in&lt;\/a&gt;\n{\/if}<\/pre><\/div><p>You can combine the two blocks above with <code>{:else}<\/code>:<\/p><div class=\"code-block named\">App.html<pre>{#if user.loggedIn}\n\t&lt;a href=\"\/logout\"&gt;log out&lt;\/a&gt;\n{:else}\n\t&lt;a href=\"\/login\"&gt;log in&lt;\/a&gt;\n{\/if}<\/pre><\/div><p>You can also use <code>{:elseif ...}<\/code>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1os996k\">REPL<\/a>App.html<pre>{#if x &gt; 10}\n\t&lt;p&gt;{x} is greater than 10&lt;\/p&gt;\n{:elseif 5 &gt; x}\n\t&lt;p&gt;{x} is less than 5&lt;\/p&gt;\n{:else}\n\t&lt;p&gt;{x} is between 5 and 10&lt;\/p&gt;\n{\/if}<\/pre><\/div><h3 id=\"each-blocks\">Each blocks<\/h3><p>Iterate over lists of data:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=scyln8\">REPL<\/a>App.html<pre>&lt;h1&gt;Cats of YouTube&lt;\/h1&gt;\n&lt;ul&gt;\n\t{#each cats as cat}\n\t\t&lt;li&gt;&lt;a target=\"_blank\" href={cat.video}&gt;{cat.name}&lt;\/a&gt;&lt;\/li&gt;\n\t{:else}\n\t\t&lt;li&gt;No cats :(&lt;\/li&gt;\n\t{\/each}\n&lt;\/ul&gt;<\/pre><\/div><p>Else is triggered when the list is empty.<\/p><p>You can access the index of the current element with <em>expression<\/em> as <em>name<\/em>, <em>index<\/em>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=19g2aa8\">REPL<\/a>App.html<pre>&lt;div class=\"grid\"&gt;\n\t{#each rows as row, y}\n\t\t&lt;div class=\"row\"&gt;\n\t\t\t{#each columns as column, x}\n\t\t\t\t&lt;code class=\"cell\"&gt;\n\t\t\t\t\t{x + 1},{y + 1}:\n\t\t\t\t\t&lt;strong&gt;{row[column]}&lt;\/strong&gt;\n\t\t\t\t&lt;\/code&gt;\n\t\t\t{\/each}\n\t\t&lt;\/div&gt;\n\t{\/each}\n&lt;\/div&gt;<\/pre><\/div><blockquote>\n<p>By default, if the list <code>a, b, c<\/code> becomes <code>a, c<\/code>, Svelte will <em>remove<\/em> the third block and <em>change<\/em> the second from <code>b<\/code> to <code>c<\/code>, rather than removing <code>b<\/code>. If that's not what you want, use a <a href=\"https:\/\/svelte.technology\/guide#keyed-each-blocks\">keyed each block<\/a>.<\/p>\n<\/blockquote><p>You can use destructuring patterns on the elements of the array:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=t2mztn\">REPL<\/a>App.html<pre>&lt;h1&gt;It's the cats of YouTube again&lt;\/h1&gt;\n&lt;ul&gt;\n\t{#each cats as {name, video} }\n\t\t&lt;li&gt;&lt;a target=\"_blank\" href={video}&gt;{name}&lt;\/a&gt;&lt;\/li&gt;\n\t{\/each}\n&lt;\/ul&gt;<\/pre><\/div><p>If you want to iterate over an object you can use <code>Object.entries(object)<\/code> which returns the object's properties as <code>[key, value]<\/code> pairs:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1hpzynj\">REPL<\/a>App.html<pre>&lt;h1&gt;Cats and Dogs&lt;\/h1&gt;\n{#each Object.entries(animals) as [animal, names]}\n\t&lt;p&gt;{animal}: {names.join(\" and \")}&lt;\/p&gt;\n{\/each}<\/pre><\/div><h3 id=\"await-blocks\">Await blocks<\/h3><p>You can represent the three states of a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Global_Objects\/Promise\">Promise<\/a> \u2014 pending, fulfilled and rejected \u2014 with an <code>await<\/code> block:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=54phb8\">REPL<\/a>App.html<pre>{#await promise}\n\t&lt;p&gt;wait for it...&lt;\/p&gt;\n{:then answer}\n\t&lt;p&gt;the answer is {answer}!&lt;\/p&gt;\n{:catch error}\n\t&lt;p&gt;well that's odd&lt;\/p&gt;\n{\/await}\n&lt;script&gt;\n\texport default {\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\tpromise: new Promise(fulfil =&gt; {\n\t\t\t\t\tsetTimeout(() =&gt; fulfil(42), 3000);\n\t\t\t\t})\n\t\t\t};\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>If the expression in <code>{#await expression}<\/code> <em>isn't<\/em> a promise, Svelte skips ahead to the <code>then<\/code> section.<\/p><h3 id=\"directives\">Directives<\/h3><p>Directives allow you to add special instructions for adding <a href=\"https:\/\/svelte.technology\/guide#event-handlers\">event handlers<\/a>, <a href=\"https:\/\/svelte.technology\/guide#bindings\">bindings<\/a>, <a href=\"https:\/\/svelte.technology\/guide#refs\">referencing elements<\/a> and so on. We'll cover each of those in later stages of this guide \u2013 for now, all you need to know is that directives can be identified by the <code>:<\/code> character:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1th74dm\">REPL<\/a>App.html<pre>&lt;p&gt;Count: {count}&lt;\/p&gt;\n&lt;button on:click=\"set({ count: count + 1 })\"&gt;+1&lt;\/button&gt;<\/pre><\/div><blockquote>\n<p>Technically, the <code>:<\/code> character is used to denote namespaced attributes in HTML. These will <em>not<\/em> be treated as directives, if encountered.<\/p>\n<\/blockquote><p>To inspect data as it changes and flows through your app, use a <code>{@debug ...}<\/code> tag:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=ljrijs\">REPL<\/a>App.html<pre>&lt;input bind:value=name&gt;\n{@debug name}\n&lt;h1&gt;Hello {name}!&lt;\/h1&gt;<\/pre><\/div><p>This will log the value of <code>name<\/code> whenever it changes. If your devtools are open, changing <code>name<\/code> will pause execution and open the debugger.<\/p><p>You can debug multiple values simultaneously (<code>{@debug foo, bar, baz}<\/code>), or use <code>{@debug}<\/code> to pause execution whenever the surrounding markup is updated.<\/p><blockquote>\n<p>Debug tags only have an effect when compiling with the <code>dev: true<\/code> compiler option.<\/p>\n<\/blockquote><\/section><section id=\"scoped-styles\" class=\"svelte-1kk9xxo\"><p>One of Svelte's key tenets is that components should be self-contained and reusable in different contexts. Because of that, it has a mechanism for <em>scoping<\/em> your CSS, so that you don't accidentally clobber other selectors on the page.<\/p><h3 id=\"adding-styles\">Adding styles<\/h3><p>Your component template can have a <code>&lt;style&gt;<\/code> tag, like so:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1r1x4wi\">REPL<\/a>App.html<pre>&lt;div class=\"foo\"&gt;\n\tBig red Comic Sans\n&lt;\/div&gt;\n&lt;style&gt;\n\t.foo {\n\t\tcolor: red;\n\t\tfont-size: 2em;\n\t\tfont-family: 'Comic Sans MS';\n\t}\n&lt;\/style&gt;<\/pre><\/div><h3 id=\"how-it-works\">How it works<\/h3><p>Open the example above in the REPL and inspect the element to see what has happened \u2013 Svelte has added a <code>svelte-[uniqueid]<\/code> class to the element, and transformed the CSS selector accordingly. Since no other element on the page can share that selector, anything else on the page with <code>class=\"foo\"<\/code> will be unaffected by our styles.<\/p><p>This is vastly simpler than achieving the same effect via <a href=\"http:\/\/caniuse.com\/#search=shadow%20dom\">Shadow DOM<\/a> and works everywhere without polyfills.<\/p><blockquote>\n<p>Svelte will add a <code>&lt;style&gt;<\/code> tag to the page containing your scoped styles. Dynamically adding styles may be impossible if your site has a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/HTTP\/CSP\">Content Security Policy<\/a>. If that's the case, you can use scoped styles by <a href=\"https:\/\/svelte.technology\/guide#rendering-css\">server-rendering your CSS<\/a> and using the <code>css: false<\/code> compiler option (or <code>--no-css<\/code> with the CLI).<\/p>\n<\/blockquote><h3 id=\"cascading-rules\">Cascading rules<\/h3><p>Styles will <em>only<\/em> apply to the current component, unless you opt in to cascading with the <code>:global(...)<\/code> modifier:<\/p><div class=\"code-block named\">App.html<pre>&lt;div&gt;\n\t&lt;Widget\/&gt;\n&lt;\/div&gt;\n&lt;style&gt;\n\tp {\n\t\t\/* this block will be disregarded, since\n\t\t   there are no &lt;p&gt; elements here *\/\n\t\tcolor: red;\n\t}\n\tdiv :global(p) {\n\t\t\/* this block will be applied to any &lt;p&gt; elements\n\t\t   inside the &lt;div&gt;, i.e. in &lt;Widget&gt; *\/\n\t\tfont-weight: bold;\n\t}\n&lt;\/style&gt;\n&lt;script&gt;\n\timport Widget from '.\/Widget.html';\n\texport default {\n\t\tcomponents: { Widget }\n\t};\n&lt;\/script&gt;<\/pre><\/div><blockquote>\n<p>Scoped styles are <em>not<\/em> dynamic \u2013 they are shared between all instances of a component. In other words you can't use <code>{tags}<\/code> inside your CSS.<\/p>\n<\/blockquote><h3 id=\"unused-style-removal\">Unused style removal<\/h3><p>Svelte will identify and remove styles that are not being used in your app. It will also emit a warning so that you can remove them from the source.<\/p><p>For rules <em>not<\/em> to be removed, they must apply to the component's markup. As far as Svelte is concerned <code>.active<\/code> is unused in the following code and should be removed:<\/p><div class=\"code-block named\">App.html<pre>&lt;div&gt;\n\t&lt;p ref:paragraph&gt;this text is not bold&lt;\/p&gt;\n&lt;\/div&gt;\n&lt;style&gt;\n\t.active {\n\t\tcolor: bold;\n\t}\n&lt;\/style&gt;\n&lt;script&gt;\n\texport default {\n\t\toncreate() {\n\t\t\tconst { active } = this.get();\n\t\t\tif (active) this.refs.paragraph.classList.add('active');\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Instead of manually manipulating the DOM, you should always use the <code>class<\/code> attribute (or the <a href=\"https:\/\/svelte.technology\/guide#classes\">class directive<\/a>):<\/p><div class=\"code-block named\">App.html<pre>&lt;div&gt;\n\t&lt;p class=\"{active ? 'active' : ''}\"&gt;this text is bold&lt;\/p&gt;\n&lt;\/div&gt;<\/pre><\/div><p>If that's impossible for some reason, you can use <code>:global(...)<\/code>:<\/p><div class=\"code-block named\">App.html<pre>&lt;style&gt;\n\tdiv :global(.active) {\n\t\tcolor: bold;\n\t}\n&lt;\/style&gt;<\/pre><\/div><p>The same applies to the contents of <code>{@html ...}<\/code> tags.<\/p><h3 id=\"special-selectors\">Special selectors<\/h3><p>If you have a <a href=\"https:\/\/svelte.technology\/guide#refs\">ref<\/a> on an element, you can use it as a CSS selector. The <code>ref:*<\/code> selector has the same specificity as a class or attribute selector.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=15ebt9o\">REPL<\/a>App.html<pre>&lt;div ref:foo&gt;\n\tyeah!\n&lt;\/div&gt;\n&lt;style&gt;\n\tref:foo {\n\t\tcolor: magenta;\n\t\tfont-size: 5em;\n\t}\n&lt;\/style&gt;<\/pre><\/div><\/section><section id=\"behaviour\" class=\"svelte-1kk9xxo\"><p>As well as scoped styles and a template, components can encapsulate <em>behaviours<\/em>. For that, we add a <code>&lt;script&gt;<\/code> element and export an object:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1scjij\">REPL<\/a>App.html<pre>&lt;div&gt;\n\t&lt;!-- template goes here --&gt;\n&lt;\/div&gt;\n&lt;script&gt;\n\texport default {\n\t\t\/\/ behaviours go here\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"default-data\">Default data<\/h3><p>Often, it makes sense for a component to have default data. This should be expressed as a function that returns a plain JavaScript object:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1qygqxy\">REPL<\/a>App.html<pre>&lt;p&gt;Count: {count}&lt;\/p&gt;\n&lt;button on:click=\"set({ count: count + 1 })\"&gt;+1&lt;\/button&gt;\n&lt;script&gt;\n\texport default {\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\tcount: 0\n\t\t\t};\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Data supplied at instantiation takes priority over defaults. In other words, if we instantiated the component above like so...<\/p><div class=\"code-block\"><pre>const counter = new Counter({\n\tdata: {\n\t\tcount: 99\n\t}\n});<\/pre><\/div><p>...then <code>{count}<\/code>, or <code>counter.get().count<\/code>, would initially be 99 rather than 0.<\/p><h3 id=\"computed-properties\">Computed properties<\/h3><p>Often, your program will use values that depend on other values \u2013 for example, you might have a filtered list, which depends on both the list <em>and<\/em> the filter. Normally in JavaScript you'd have to add logic to update the dependent property when <em>any<\/em> of the dependencies change. This is a frequent source of bugs, and it gets worse as your application grows.<\/p><p>Svelte allows you to express these dependencies in computed properties, which are recalculated whenever those dependencies change:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1hblb9y\">REPL<\/a>App.html<pre>&lt;p&gt;\n\tThe time is\n\t&lt;strong&gt;{hours}:{minutes}:{seconds}&lt;\/strong&gt;\n&lt;\/p&gt;\n&lt;script&gt;\n\texport default {\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\ttime: new Date()\n\t\t\t};\n\t\t},\n\t\tcomputed: {\n\t\t\thours:   ({ time }) =&gt; time.getHours(),\n\t\t\tminutes: ({ time }) =&gt; time.getMinutes(),\n\t\t\tseconds: ({ time }) =&gt; time.getSeconds()\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Each function is passed the component's current state object. Because we're using destructuring syntax, the compiler knows that <code>hours<\/code>, <code>minutes<\/code> and <code>seconds<\/code> only need to re-run when <code>time<\/code> changes, and not when any other values change. There's no costly dependency tracking involved \u2013 the dependency graph is resolved at compile time.<\/p><blockquote>\n<p><code>computed<\/code> must be an object literal, and the properties must be function expressions or arrow function expressions. Any external functions used in computed must be wrapped <em>here<\/em>:<\/p>\n<\/blockquote><div class=\"code-block\"><pre>import externalFunc from '_external_file';\nexport default {\n\tcomputed: {\n\t\texternalFunc: ({ dep1, dep2 }) =&gt; externalFunc(dep1, dep2);\n\t}\n}<\/pre><\/div><p>Computed properties can of course return functions. For example, we could dynamically generate a filter function for a list of items:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1kuacvf\">REPL<\/a>App.html<pre>&lt;input bind:value=search&gt;\n{#each items.filter(predicate) as word}\n\t&lt;p&gt;&lt;strong&gt;{word.slice(0, search.length)}&lt;\/strong&gt;{word.slice(search.length)}&lt;\/p&gt;\n{:else}\n\t&lt;p&gt;no matches!&lt;\/p&gt;\n{\/each}\n&lt;script&gt;\n\texport default {\n\t\tcomputed: {\n\t\t\tpredicate: ({ search }) =&gt; {\n\t\t\t\tsearch = search.toLowerCase();\n\t\t\t\treturn word =&gt; word.startsWith(search);\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"lifecycle-hooks\">Lifecycle hooks<\/h3><p>There are four 'hooks' provided by Svelte for adding control logic \u2014 <code>oncreate<\/code>, <code>ondestroy<\/code>, <code>onstate<\/code> and <code>onupdate<\/code>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=ov8jor\">REPL<\/a>App.html<pre>&lt;p&gt;\n\tThe time is\n\t&lt;strong&gt;{hours}:{minutes}:{seconds}&lt;\/strong&gt;\n&lt;\/p&gt;\n&lt;script&gt;\n\texport default {\n\t\tonstate({ changed, current, previous }) {\n\t\t\t\/\/ this fires before oncreate, and on every state change.\n\t\t\t\/\/ the first time it runs, `previous` is undefined\n\t\t\tif (changed.time) {\n\t\t\t\tconsole.log(`time changed: ${previous &amp;&amp; previous.time} -&gt; ${current.time}`);\n\t\t\t}\n\t\t},\n\t\toncreate() {\n\t\t\t\/\/ this fires when the component has been\n\t\t\t\/\/ rendered to the DOM\n\t\t\tconsole.log('in oncreate');\n\t\t\tthis.interval = setInterval(() =&gt; {\n\t\t\t\tthis.set({ time: new Date() });\n\t\t\t}, 1000);\n\t\t},\n\t\tonupdate({ changed, current, previous }) {\n\t\t\t\/\/ this fires after oncreate, and after every state change\n\t\t\t\/\/ once the DOM is synchronised with the data\n\t\t\tconsole.log(`The DOM has been updated`);\n\t\t},\n\t\tondestroy() {\n\t\t\t\/\/ this fires when the component is\n\t\t\t\/\/ removed from the DOM\n\t\t\tconsole.log('in ondestroy');\n\t\t\tclearInterval(this.interval);\n\t\t},\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\ttime: new Date()\n\t\t\t};\n\t\t},\n\t\tcomputed: {\n\t\t\thours:   ({ time }) =&gt; time.getHours(),\n\t\t\tminutes: ({ time }) =&gt; time.getMinutes(),\n\t\t\tseconds: ({ time }) =&gt; time.getSeconds()\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><blockquote>\n<p>You can add event listeners corresponding to <code>onstate<\/code>, <code>onupdate<\/code> and <code>ondestroy<\/code> programmatically \u2014 see <a href=\"https:\/\/svelte.technology\/guide#component-on-eventname-callback-\">component.on<\/a><\/p>\n<\/blockquote><h3 id=\"helpers\">Helpers<\/h3><p>Helpers are simple functions that are used in your template. In the example above, we want to ensure that <code>minutes<\/code> and <code>seconds<\/code> are preceded with a <code>0<\/code> if they only have one digit, so we add a <code>leftPad<\/code> helper:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=w7mzjq\">REPL<\/a>App.html<pre>&lt;p&gt;\n\tThe time is\n\t&lt;strong&gt;{hours}:{leftPad(minutes, 2, '0')}:{leftPad(seconds, 2, '0')}&lt;\/strong&gt;\n&lt;\/p&gt;\n&lt;script&gt;\n\timport leftPad from 'left-pad';\n\texport default {\n\t\thelpers: {\n\t\t\tleftPad\n\t\t},\n\t\toncreate() {\n\t\t\tthis.interval = setInterval(() =&gt; {\n\t\t\t\tthis.set({ time: new Date() });\n\t\t\t}, 1000);\n\t\t},\n\t\tondestroy() {\n\t\t\tclearInterval(this.interval);\n\t\t},\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\ttime: new Date()\n\t\t\t};\n\t\t},\n\t\tcomputed: {\n\t\t\thours:   ({ time }) =&gt; time.getHours(),\n\t\t\tminutes: ({ time }) =&gt; time.getMinutes(),\n\t\t\tseconds: ({ time }) =&gt; time.getSeconds()\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Of course, you could use <code>leftPad<\/code> inside the computed properties instead of in the template. There's no hard and fast rule about when you should use expressions with helpers versus when you should use computed properties \u2013 do whatever makes your component easier for the next developer to understand.<\/p><blockquote>\n<p>Helper functions should be <em>pure<\/em> \u2013 in other words, they should not have side-effects, and their returned value should only depend on their arguments.<\/p>\n<\/blockquote><h3 id=\"custom-methods\">Custom methods<\/h3><p>In addition to the <a href=\"https:\/\/svelte.technology\/guide#component-api\">built-in methods<\/a>, you can add methods of your own:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=9yymyy\">REPL<\/a>App.html<pre>&lt;p&gt;Try calling &lt;code&gt;app.say('hello!')&lt;\/code&gt; from the console&lt;\/p&gt;\n&lt;script&gt;\n\texport default {\n\t\tmethods: {\n\t\t\tsay(message) {\n\t\t\t\talert(message); \/\/ again, please don't do this\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>These become part of the component's API:<\/p><div class=\"code-block\"><pre>import MyComponent from '.\/MyComponent.html';\nvar component = new MyComponent({\n\ttarget: document.querySelector('main')\n});\ncomponent.say('\ud83d\udc4b');<\/pre><\/div><p>Methods (whether built-in or custom) can also be called inside <a href=\"https:\/\/svelte.technology\/guide#event-handlers\">event handlers<\/a>:<\/p><div class=\"code-block named\">App.html<pre>&lt;button on:click=\"say('hello')\"&gt;say hello!&lt;\/button&gt;<\/pre><\/div><h3 id=\"custom-event-handlers\">Custom event handlers<\/h3><p>Soon, we'll learn about <a href=\"https:\/\/svelte.technology\/guide#event-handlers\">event handlers<\/a> \u2013 if you want, skip ahead to that section first then come back here!<\/p><p>Most of the time you can make do with the standard DOM events (the sort you'd add via <code>element.addEventListener<\/code>, such as <code>click<\/code>) but sometimes you might need custom events to handle gestures, for example.<\/p><p>Custom events are just functions that take a node and a callback as their argument, and return an object with a <code>destroy<\/code> method that gets called when the element is removed from the page:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=16c31bk\">REPL<\/a>App.html<pre>&lt;button on:longpress=\"set({ done: true })\"&gt;click and hold&lt;\/button&gt;\n{#if done}\n\t&lt;p&gt;clicked and held&lt;\/p&gt;\n{\/if}\n&lt;script&gt;\n\texport default {\n\t\tevents: {\n\t\t\tlongpress(node, callback) {\n\t\t\t\tfunction onmousedown(event) {\n\t\t\t\t\tconst timeout = setTimeout(() =&gt; callback(event), 1000);\n\t\t\t\t\tfunction cancel() {\n\t\t\t\t\t\tclearTimeout(timeout);\n\t\t\t\t\t\tnode.removeEventListener('mouseup', cancel, false);\n\t\t\t\t\t}\n\t\t\t\t\tnode.addEventListener('mouseup', cancel, false);\n\t\t\t\t}\n\t\t\t\tnode.addEventListener('mousedown', onmousedown, false);\n\t\t\t\treturn {\n\t\t\t\t\tdestroy() {\n\t\t\t\t\t\tnode.removeEventListener('mousedown', onmousedown, false);\n\t\t\t\t\t}\n\t\t\t\t};\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"namespaces\">Namespaces<\/h3><p>Components are assumed to be in the HTML namespace. If your component is designed to be used inside an <code>&lt;svg&gt;<\/code> element, you need to specify the namespace:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=xlwpb0\">REPL<\/a>App.html<pre>&lt;svg viewBox=\"0 0 1000 1000\" style=\"width: 100%; height: 100%;\"&gt;\n\t&lt;SmileyFace x=70 y=280 size=100 fill=\"#f4d9c7\"\/&gt;\n\t&lt;SmileyFace x=800 y=250 size=150 fill=\"#40250f\"\/&gt;\n\t&lt;SmileyFace x=150 y=700 size=110 fill=\"#d2aa7a\"\/&gt;\n\t&lt;SmileyFace x=875 y=730 size=130 fill=\"#824e2e\"\/&gt;\n\t&lt;SmileyFace x=450 y=500 size=240 fill=\"#d2b198\"\/&gt;\n&lt;\/svg&gt;\n&lt;script&gt;\n\timport SmileyFace from '.\/SmileyFace.html';\n\texport default {\n\t\tcomponents: { SmileyFace }\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">SmileyFace.html<pre>&lt;!-- CC-BY-SA \u2014 https:\/\/commons.wikimedia.org\/wiki\/File:718smiley.svg --&gt;\n&lt;g transform=\"translate({x},{y}) scale({size \/ 366.5})\"&gt;\n\t&lt;circle r=366.5\/&gt;\n\t&lt;circle r=336.5 fill={fill}\/&gt;\n\t&lt;path d=\"m-41.5 298.5c-121-21-194-115-212-233v-8l-25-1-1-18h481c6 13 10 27 13 41 13 94-38 146-114 193-45 23-93 29-142 26z\"\/&gt;\n\t&lt;path d=\"m5.5 280.5c52-6 98-28 138-62 28-25 46-56 51-87 4-20 1-57-5-70l-423-1c-2 56 39 118 74 157 31 34 72 54 116 63 11 2 38 2 49 0z\" fill=\"#871945\"\/&gt;\n\t&lt;path d=\"m-290.5 -24.5c-13-26-13-57-9-85 6-27 18-52 35-68 21-20 50-23 77-18 15 4 28 12 39 23 18 17 30 40 36 67 4 20 4 41 0 60l-6 21z\"\/&gt;\n\t&lt;path d=\"m-132.5 -43.5c5-6 6-40 2-58-3-16-4-16-10-10-14 14-38 14-52 0-15-18-12-41 6-55 3-3 5-5 5-6-1-4-22-8-34-7-42 4-57.6 40-66.2 77-3 17-1 53 4 59h145.2z\" fill=\"#fff\"\/&gt;\n\t&lt;path d=\"m11.5 -23.5c-2-3-6-20-7-29-5-28-1-57 11-83 15-30 41-52 72-60 29-7 57 0 82 15 26 17 45 49 50 82 2 12 2 33 0 45-1 10-5 26-8 30z\"\/&gt;\n\t&lt;path d=\"m198.5 -42.5c4-5 5-34 4-50-2-14-6-24-8-24-1 0-3 2-6 5-17 17-47 13-58-9-7-16-4-31 8-43 4-4 7-8 7-9 0 0-4-2-8-3-51-17-105 20-115 80-3 15 0 43 3 53z\" fill=\"#fff\"\/&gt;\n\t&lt;path d=\"m137.5 223.5s-46 40-105 53c-66 15-114-7-114-7s14-76 93-95c76-18 126 49 126 49z\" fill=\"#f9bedd\"\/&gt;\n&lt;\/g&gt;\n&lt;script&gt;\n\texport default {\n\t\t\/\/ you can either use the shorthand 'svg', or the full\n\t\t\/\/ namespace: 'http:\/\/www.w3.org\/2000\/svg'. (I know\n\t\t\/\/ which one I prefer.)\n\t\tnamespace: 'svg',\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\tx: 100,\n\t\t\t\ty: 100,\n\t\t\t\tsize: 100\n\t\t\t};\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><\/section><section id=\"nested-components\" class=\"svelte-1kk9xxo\"><p>As well as containing elements (and <code>if<\/code> blocks and <code>each<\/code> blocks), Svelte components can contain <em>other<\/em> Svelte components.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=53iz9p\">REPL<\/a>App.html<pre>&lt;div class='widget-container'&gt;\n\t&lt;Widget\/&gt;\n&lt;\/div&gt;\n&lt;script&gt;\n\timport Widget from '.\/Widget.html';\n\texport default {\n\t\tcomponents: {\n\t\t\tWidget\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Widget.html<pre>&lt;p&gt;I am a nested component&lt;\/p&gt;<\/pre><\/div><p>That's similar to doing this...<\/p><div class=\"code-block\"><pre>import Widget from '.\/Widget.html';\nconst widget = new Widget({\n\ttarget: document.querySelector('.widget-container')\n});<\/pre><\/div><p>...except that Svelte takes care of destroying the child component when the parent is destroyed, and keeps the two components in sync with <em>props<\/em>.<\/p><blockquote>\n<p>Component names must be capitalised, following the widely-used JavaScript convention of capitalising constructor names. It's also an easy way to distinguish components from elements in your template.<\/p>\n<\/blockquote><h3 id=\"props\">Props<\/h3><p>Props, short for 'properties', are the means by which you pass data down from a parent to a child component \u2014 in other words, they're just like attributes on an element:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=16qktdb\">REPL<\/a>App.html<pre>&lt;div class='widget-container'&gt;\n\t&lt;Widget foo bar=\"static\" baz={dynamic}\/&gt;\n&lt;\/div&gt;\n&lt;script&gt;\n\timport Widget from '.\/Widget.html';\n\texport default {\n\t\tcomponents: {\n\t\t\tWidget\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Widget.html<pre>&lt;p&gt;foo: {foo}&lt;\/p&gt;\n&lt;p&gt;bar: {bar}&lt;\/p&gt;\n&lt;p&gt;baz: {baz}&lt;\/p&gt;<\/pre><\/div><p>As with element attributes, prop values can contain any valid JavaScript expression.<\/p><p>Often, the name of the property will be the same as the value, in which case we can use a shorthand:<\/p><div class=\"code-block named\">App.html<pre>&lt;!-- these are equivalent --&gt;\n&lt;Widget foo={foo}\/&gt;\n&lt;Widget {foo}\/&gt;<\/pre><\/div><blockquote>\n<p>Note that props are <em>one-way<\/em> \u2014 to get data from a child component into a parent component, use <a href=\"https:\/\/svelte.technology\/guide#bindings\">bindings<\/a>.<\/p>\n<\/blockquote><h3 id=\"composing-with-slot\">Composing with <code>&lt;slot&gt;<\/code><\/h3><p>A component can contain a <code>&lt;slot&gt;&lt;\/slot&gt;<\/code> element, which allows the parent component to inject content:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1fz2fc6\">REPL<\/a>App.html<pre>&lt;Box&gt;\n\t&lt;h2&gt;Hello!&lt;\/h2&gt;\n\t&lt;p&gt;This is a box. It can contain anything.&lt;\/p&gt;\n&lt;\/Box&gt;\n&lt;script&gt;\n\timport Box from '.\/Box.html';\n\texport default {\n\t\tcomponents: { Box }\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Box.html<pre>&lt;div class=\"box\"&gt;\n\t&lt;slot&gt;&lt;!-- content is injected here --&gt;&lt;\/slot&gt;\n&lt;\/div&gt;\n&lt;style&gt;\n\t.box {\n\t\tborder: 2px solid black;\n\t\tpadding: 0.5em;\n\t}\n&lt;\/style&gt;<\/pre><\/div><p>The <code>&lt;slot&gt;<\/code> element can contain 'fallback content', which will be used if no children are provided for the component:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=n202gz\">REPL<\/a>App.html<pre>&lt;Box&gt;&lt;\/Box&gt;\n&lt;script&gt;\n\timport Box from '.\/Box.html';\n\texport default {\n\t\tcomponents: { Box }\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Box.html<pre>&lt;div class=\"box\"&gt;\n\t&lt;slot&gt;\n\t\t&lt;p class=\"fallback\"&gt;the box is empty!&lt;\/p&gt;\n\t&lt;\/slot&gt;\n&lt;\/div&gt;\n&lt;style&gt;\n\t.box {\n\t\tborder: 2px solid black;\n\t\tpadding: 0.5em;\n\t}\n\t.fallback {\n\t\tcolor: #999;\n\t}\n&lt;\/style&gt;<\/pre><\/div><p>You can also have <em>named<\/em> slots. Any elements with a corresponding <code>slot<\/code> attribute will fill these slots:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=p0k5ow\">REPL<\/a>App.html<pre>&lt;ContactCard&gt;\n\t&lt;span slot=\"name\"&gt;P. Sherman&lt;\/span&gt;\n\t&lt;span slot=\"address\"&gt;42 Wallaby Way, Sydney&lt;\/span&gt;\n&lt;\/ContactCard&gt;\n&lt;script&gt;\n\timport ContactCard from '.\/ContactCard.html';\n\texport default {\n\t\tcomponents: { ContactCard }\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">ContactCard.html<pre>&lt;div class=\"contact-card\"&gt;\n\t&lt;h2&gt;&lt;slot name=\"name\"&gt;&lt;\/slot&gt;&lt;\/h2&gt;\n\t&lt;slot name=\"address\"&gt;Unknown address&lt;\/slot&gt;\n\t&lt;br&gt;\n\t&lt;slot name=\"email\"&gt;Unknown email&lt;\/slot&gt;\n&lt;\/div&gt;\n&lt;style&gt;\n\t.contact-card {\n\t\tborder: 2px solid black;\n\t\tpadding: 0.5em;\n\t}\n&lt;\/style&gt;<\/pre><\/div><h3 id=\"shorthand-imports\">Shorthand imports<\/h3><p>As an alternative to using an <code>import<\/code> declaration...<\/p><div class=\"code-block named\">App.html<pre>&lt;script&gt;\n\timport Widget from '.\/Widget.html';\n\texport default {\n\t\tcomponents: { Widget }\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>...you can write this:<\/p><div class=\"code-block named\">App.html<pre>&lt;script&gt;\n\texport default {\n\t\tcomponents: {\n\t\t\tWidget: '.\/Widget.html'\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><\/section><section id=\"special-components\" class=\"svelte-1kk9xxo\"><p>Svelte includes a handful of built-in elements with special behaviour.<\/p><h3 id=\"svelte-self\"><code>&lt;svelte:self&gt;<\/code><\/h3><p>Sometimes, a component needs to embed itself recursively \u2014 for example if you have a tree-like data structure. In Svelte, that's accomplished with the <code>&lt;svelte:self&gt;<\/code> tag:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1gqg5bs\">REPL<\/a>App.html<pre>{#if countdown &gt; 0}\n\t&lt;p&gt;{countdown}&lt;\/p&gt;\n\t&lt;svelte:self countdown=\"{countdown - 1}\"\/&gt;\n{:else}\n\t&lt;p&gt;liftoff!&lt;\/p&gt;\n{\/if}<\/pre><\/div><h3 id=\"svelte-component\"><code>&lt;svelte:component&gt;<\/code><\/h3><p>If you don't know what kind of component to render until the app runs \u2014 in other words, it's driven by state (aka a dynamic component) \u2014 you can use <code>&lt;svelte:component&gt;<\/code>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1mkkz48\">REPL<\/a>App.html<pre>&lt;input type=checkbox bind:checked=foo&gt; foo\n&lt;svelte:component this=\"{foo ? Red : Blue}\" name=\"thing\"\/&gt;\n&lt;script&gt;\n\timport Red from '.\/Red.html';\n\timport Blue from '.\/Blue.html';\n\texport default {\n\t\tdata() {\n\t\t\treturn { Red, Blue }\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><blockquote>\n<p>Note that <code>Red<\/code> and <code>Blue<\/code> are items in <code>data<\/code>, <em>not<\/em> <code>components<\/code>, unlike if we were doing <code>&lt;Red&gt;<\/code> or <code>&lt;Blue&gt;<\/code>.<\/p>\n<\/blockquote><p>The expression inside the <code>this=\"{...}\"<\/code> can be any valid JavaScript expression. For example, it could be a <a href=\"https:\/\/svelte.technology\/guide#computed-properties\">computed property<\/a>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=ku7y5t\">REPL<\/a>App.html<pre>&lt;label&gt;&lt;input type=radio bind:group=size value=small&gt; small&lt;\/label&gt;\n&lt;label&gt;&lt;input type=radio bind:group=size value=medium&gt; medium&lt;\/label&gt;\n&lt;label&gt;&lt;input type=radio bind:group=size value=large&gt; large&lt;\/label&gt;\n&lt;svelte:component this={Size}\/&gt;\n&lt;script&gt;\n\timport Small from '.\/Small.html';\n\timport Medium from '.\/Medium.html';\n\timport Large from '.\/Large.html';\n\texport default {\n\t\tcomputed: {\n\t\t\tSize: ({size}) =&gt; {\n\t\t\t\tif (size === 'small') return Small;\n\t\t\t\tif (size === 'medium') return Medium;\n\t\t\t\treturn Large;\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Small.html<pre>&lt;p style=\"font-size: 12px\"&gt;small&lt;\/p&gt;<\/pre><\/div><div class=\"code-block named\">Medium.html<pre>&lt;p style=\"font-size: 18px\"&gt;medium&lt;\/p&gt;<\/pre><\/div><div class=\"code-block named\">Large.html<pre>&lt;p style=\"font-size: 32px\"&gt;LARGE&lt;\/p&gt;<\/pre><\/div><h3 id=\"svelte-window\"><code>&lt;svelte:window&gt;<\/code><\/h3><p>The <code>&lt;svelte:window&gt;<\/code> tag gives you a convenient way to declaratively add event listeners to <code>window<\/code>. Event listeners are automatically removed when the component is destroyed.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1qi98o0\">REPL<\/a>App.html<pre>&lt;svelte:window on:keydown=\"set({ key: event.key, keyCode: event.keyCode })\"\/&gt;\n{#if key}\n\t&lt;p&gt;&lt;kbd&gt;{key === ' ' ? 'Space' : key}&lt;\/kbd&gt; (code {keyCode})&lt;\/p&gt;\n{:else}\n\t&lt;p&gt;click in this window and press any key&lt;\/p&gt;\n{\/if}\n&lt;style&gt;\n\tkbd {\n\t\tbackground-color: #eee;\n\t\tborder: 2px solid #f4f4f4;\n\t\tborder-right-color: #ddd;\n\t\tborder-bottom-color: #ddd;\n\t\tfont-size: 2em;\n\t\tmargin: 0 0.5em 0 0;\n\t\tpadding: 0.5em 0.8em;\n\t\tfont-family: Inconsolata;\n\t}\n&lt;\/style&gt;<\/pre><\/div><p>You can also bind to certain values \u2014 so far <code>innerWidth<\/code>, <code>outerWidth<\/code>, <code>innerHeight<\/code>, <code>outerHeight<\/code>, <code>scrollX<\/code>, <code>scrollY<\/code> and <code>online<\/code>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=e07a2t\">REPL<\/a>App.html<pre>&lt;svelte:window bind:scrollY=y\/&gt;\n&lt;div class=\"background\"&gt;&lt;\/div&gt;\n&lt;p class=\"fixed\"&gt;user has scrolled {y} pixels&lt;\/p&gt;\n&lt;style&gt;\n\t.background {\n\t\tposition: absolute;\n\t\tleft: 0;\n\t\ttop: 0;\n\t\twidth: 100%;\n\t\theight: 9999px;\n\t\tbackground: linear-gradient(to bottom, #7db9e8 0%,#0a1d33 100%);\n\t}\n\t.fixed {\n\t\tposition: fixed;\n\t\ttop: 1em;\n\t\tleft: 1em;\n\t\tcolor: white;\n\t}\n&lt;\/style&gt;<\/pre><\/div><h3 id=\"svelte-document\"><code>&lt;svelte:document&gt;<\/code><\/h3><p>The <code>&lt;svelte:document&gt;<\/code> tag, just like <code>&lt;svelte:window&gt;<\/code>, gives you a convenient way to declaratively add event listeners to the <code>document<\/code> object. This is useful for listening to events that don't fire on <code>window<\/code>, such as <code>mouseenter<\/code> and <code>mouseleave<\/code>.<\/p><p>If you're building an application with Svelte \u2014 particularly if you're using <a href=\"https:\/\/sapper.svelte.technology\">Sapper<\/a> \u2014 then it's likely you'll need to add some content to the <code>&lt;head&gt;<\/code> of your page, such as adding a <code>&lt;title&gt;<\/code> element.<\/p><p>You can do that with the <code>&lt;svelte:head&gt;<\/code> tag:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=fvcn1e\">REPL<\/a>App.html<pre>&lt;svelte:head&gt;\n\t&lt;title&gt;{post.title} \u2022 My blog&lt;\/title&gt;\n&lt;\/svelte:head&gt;<\/pre><\/div><p>When <a href=\"https:\/\/svelte.technology\/guide#server-side-rendering\">server rendering<\/a>, the <code>&lt;head&gt;<\/code> contents can be extracted separately to the rest of the markup.<\/p><\/section><section id=\"element-directives\" class=\"svelte-1kk9xxo\"><p>Directives are element or component-level instructions to Svelte. They look like attributes, except with a <code>:<\/code> character.<\/p><h3 id=\"event-handlers\">Event handlers<\/h3><p>In most applications, you'll need to respond to the user's actions. In Svelte, this is done with the <code>on:[event]<\/code> directive.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1th74dm\">REPL<\/a>App.html<pre>&lt;p&gt;Count: {count}&lt;\/p&gt;\n&lt;button on:click=\"set({ count: count + 1 })\"&gt;+1&lt;\/button&gt;<\/pre><\/div><p>When the user clicks the button, Svelte calls <code>component.set(...)<\/code> with the provided arguments. You can call any method belonging to the component (whether <a href=\"https:\/\/svelte.technology\/guide#component-api\">built-in<\/a> or <a href=\"https:\/\/svelte.technology\/guide#custom-methods\">custom<\/a>), and any data property (or computed property) that's in scope:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=10bob12\">REPL<\/a>App.html<pre>&lt;p&gt;Select a category:&lt;\/p&gt;\n{#each categories as category}\n\t&lt;button on:click=\"select(category)\"&gt;select {category}&lt;\/button&gt;\n{\/each}\n&lt;script&gt;\n\texport default {\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\tcategories: [\n\t\t\t\t\t'animal',\n\t\t\t\t\t'vegetable',\n\t\t\t\t\t'mineral'\n\t\t\t\t]\n\t\t\t}\n\t\t},\n\t\tmethods: {\n\t\t\tselect(name) {\n\t\t\t\talert(`selected ${name}`); \/\/ seriously, please don't do this\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>You can also access the <code>event<\/code> object in the method call:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=wwzvp5\">REPL<\/a>App.html<pre>&lt;div on:mousemove=\"set({ x: event.clientX, y: event.clientY })\"&gt;\n\tcoords: {x},{y}\n&lt;\/div&gt;\n&lt;style&gt;\n\tdiv {\n\t\tborder: 1px solid purple;\n\t\twidth: 100%;\n\t\theight: 100%;\n\t}\n&lt;\/style&gt;<\/pre><\/div><p>The target node can be referenced as <code>this<\/code>, meaning you can do this sort of thing:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=ku4c9n\">REPL<\/a>App.html<pre>&lt;input on:focus=\"this.select()\" value=\"click to select\"&gt;<\/pre><\/div><h3 id=\"event-handler-modifiers\">Event handler modifiers<\/h3><p>While you can invoke methods like <code>event.stopPropagation<\/code> directly...<\/p><div class=\"code-block named\">App.html<pre>&lt;div on:click=\"event.stopPropagation()\"&gt;...&lt;\/div&gt;<\/pre><\/div><p>...it gets annoying if you want to combine that with some other behaviour:<\/p><div class=\"code-block named\">App.html<pre>&lt;div on:click=\"setFoo(event)\"&gt;...&lt;\/div&gt;\n&lt;script&gt;\n\texport default {\n\t\tmethods: {\n\t\t\tsetFoo(event) {\n\t\t\t\tevent.stopPropagation();\n\t\t\t\tevent.preventDefault();\n\t\t\t\tthis.set({ foo: true });\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>For that reason, Svelte lets you use <em>event modifiers<\/em>:<\/p><ul><li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Event\/preventDefault\"><code>preventDefault<\/code><\/a><\/li>\n<li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Event\/stopPropagation\"><code>stopPropagation<\/code><\/a><\/li>\n<li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/EventTarget\/addEventListener#Parameters\"><code>passive<\/code><\/a> \u2014 improves scrolling performance on touch\/wheel events (Svelte will add it automatically where it's safe to do so)<\/li>\n<li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/EventTarget\/addEventListener#Parameters\"><code>once<\/code><\/a> \u2014 removes the listener after the first invocation<\/li>\n<li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/EventTarget\/addEventListener#Parameter\"><code>capture<\/code><\/a><\/li>\n<\/ul><blockquote>\n<p><code>passive<\/code> and <code>once<\/code> are not implemented in <code>legacy<\/code> mode<\/p>\n<\/blockquote><p>The example above can be achieved with modifiers \u2014 no need for a custom method:<\/p><div class=\"code-block named\">App.html<pre>&lt;div on:click|stopPropagation|preventDefault=\"set({ foo: true })\"&gt;...&lt;\/div&gt;<\/pre><\/div><h3 id=\"custom-events\">Custom events<\/h3><p>You can define your own custom events to handle complex user interactions like dragging and swiping. See the earlier section on <a href=\"https:\/\/svelte.technology\/guide#custom-event-handlers\">custom event handlers<\/a> for more information.<\/p><h3 id=\"component-events\">Component events<\/h3><p>Events are an excellent way for <a href=\"https:\/\/svelte.technology\/guide#nested-components\">nested components<\/a> to communicate with their parents. Let's revisit our earlier example, but turn it into a <code>&lt;CategoryChooser&gt;<\/code> component:<\/p><div class=\"code-block named\">CategoryChooser.html<pre>&lt;p&gt;Select a category:&lt;\/p&gt;\n{#each categories as category}\n\t&lt;button on:click=\"fire('select', { category })\"&gt;select {category}&lt;\/button&gt;\n{\/each}\n&lt;script&gt;\n\texport default {\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\tcategories: [\n\t\t\t\t\t'animal',\n\t\t\t\t\t'vegetable',\n\t\t\t\t\t'mineral'\n\t\t\t\t]\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>When the user clicks a button, the component will fire a <code>select<\/code> event, where the <code>event<\/code> object has a <code>category<\/code> property. Any component that nests <code>&lt;CategoryChooser&gt;<\/code> can listen for events like so:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1y2ccg8\">REPL<\/a>App.html<pre>&lt;CategoryChooser on:select=\"playTwentyQuestions(event.category)\"\/&gt;\n&lt;script&gt;\n\timport CategoryChooser from '.\/CategoryChooser.html';\n\texport default {\n\t\tcomponents: {\n\t\t\tCategoryChooser\n\t\t},\n\t\tmethods: {\n\t\t\tplayTwentyQuestions(category) {\n\t\t\t\talert(`ok! you chose ${category}`);\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Just as <code>this<\/code> in an element's event handler refers to the element itself, in a component event handler <code>this<\/code> refers to the component firing the event.<\/p><p>There is also a shorthand for listening for and re-firing an event unchanged.<\/p><div class=\"code-block named\">App.html<pre>&lt;!-- these are equivalent --&gt;\n&lt;Widget on:foo=\"fire('foo', event)\"\/&gt;\n&lt;Widget on:foo\/&gt;<\/pre><\/div><p>Since component events do not propagate as DOM events do, this can be used to pass events through intermediate components. This shorthand technique also applies to element events (<code>on:click<\/code> is equivalent to <code>on:click=\"fire('click', event)\"<\/code>).<\/p><h3 id=\"refs\">Refs<\/h3><p>Refs are a convenient way to store a reference to particular DOM nodes or components. Declare a ref with <code>ref:[name]<\/code>, and access it inside your component's methods with <code>this.refs.[name]<\/code>:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1m63aev\">REPL<\/a>App.html<pre>&lt;canvas ref:canvas width=200 height=200&gt;&lt;\/canvas&gt;\n&lt;script&gt;\n\timport createRenderer from '.\/createRenderer.js';\n\texport default {\n\t\toncreate() {\n\t\t\tconst canvas = this.refs.canvas;\n\t\t\tconst ctx = canvas.getContext('2d');\n\t\t\tconst renderer = createRenderer(canvas, ctx);\n\t\t\tthis.on('destroy', renderer.stop);\n\t\t}\n\t}\n&lt;\/script&gt;<\/pre><\/div><blockquote>\n<p>Since only one element or component can occupy a given <code>ref<\/code>, don't use them in <code>{#each ...}<\/code> blocks. It's fine to use them in <code>{#if ...}<\/code> blocks however.<\/p>\n<\/blockquote><p>Note that you can use refs in your <code>&lt;style&gt;<\/code> blocks \u2014 see <a href=\"https:\/\/svelte.technology\/guide#special-selectors\">Special selectors<\/a>.<\/p><h3 id=\"transitions\">Transitions<\/h3><p>Transitions allow elements to enter and leave the DOM gracefully, rather than suddenly appearing and disappearing.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=rlu570\">REPL<\/a>App.html<pre>&lt;input type=checkbox bind:checked=visible&gt; visible\n{#if visible}\n\t&lt;p transition:fade&gt;fades in and out&lt;\/p&gt;\n{\/if}\n&lt;script&gt;\n\timport { fade } from 'svelte-transitions';\n\texport default {\n\t\ttransitions: { fade }\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Transitions can have parameters \u2014 typically <code>delay<\/code> and <code>duration<\/code>, but often others, depending on the transition in question. For example, here's the <code>fly<\/code> transition from the <a href=\"https:\/\/github.com\/sveltejs\/svelte-transitions\">svelte-transitions<\/a> package:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1fkspjw\">REPL<\/a>App.html<pre>&lt;input type=checkbox bind:checked=visible&gt; visible\n{#if visible}\n\t&lt;p transition:fly=\"{y: 200, duration: 1000}\"&gt;flies 200 pixels up, slowly&lt;\/p&gt;\n{\/if}\n&lt;script&gt;\n\timport { fly } from 'svelte-transitions';\n\texport default {\n\t\ttransitions: { fly }\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>An element can have separate <code>in<\/code> and <code>out<\/code> transitions:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1ehyjn2\">REPL<\/a>App.html<pre>&lt;input type=checkbox bind:checked=visible&gt; visible\n{#if visible}\n\t&lt;p in:fly=\"{y: 50}\" out:fade&gt;flies up, fades out&lt;\/p&gt;\n{\/if}\n&lt;script&gt;\n\timport { fade, fly } from 'svelte-transitions';\n\texport default {\n\t\ttransitions: { fade, fly }\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Transitions are simple functions that take a <code>node<\/code> and any provided <code>parameters<\/code> and return an object with the following properties:<\/p><ul><li><code>duration<\/code> \u2014 how long the transition takes in milliseconds<\/li>\n<li><code>delay<\/code> \u2014 milliseconds before the transition starts<\/li>\n<li><code>easing<\/code> \u2014 an <a href=\"https:\/\/github.com\/rollup\/eases-jsnext\">easing function<\/a><\/li>\n<li><code>css<\/code> \u2014 a function that accepts an argument <code>t<\/code> between 0 and 1 and returns the styles that should be applied at that moment<\/li>\n<li><code>tick<\/code> \u2014 a function that will be called on every frame, with the same <code>t<\/code> argument, while the transition is in progress<\/li>\n<\/ul><p>Of these, <code>duration<\/code> is required, as is <em>either<\/em> <code>css<\/code> or <code>tick<\/code>. The rest are optional. Here's how the <code>fade<\/code> transition is implemented, for example:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=jwzdjt\">REPL<\/a>App.html<pre>&lt;input type=checkbox bind:checked=visible&gt; visible\n{#if visible}\n\t&lt;p transition:fade&gt;fades in and out&lt;\/p&gt;\n{\/if}\n&lt;script&gt;\n\texport default {\n\t\ttransitions: {\n\t\t\tfade(node, { delay = 0, duration = 400 }) {\n\t\t\t\tconst o = +getComputedStyle(node).opacity;\n\t\t\t\treturn {\n\t\t\t\t\tdelay,\n\t\t\t\t\tduration,\n\t\t\t\t\tcss: t =&gt; `opacity: ${t * o}`\n\t\t\t\t};\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><blockquote>\n<p>If the <code>css<\/code> option is used, Svelte will create a CSS animation that runs efficiently off the main thread. Therefore if you can achieve an effect using <code>css<\/code> rather than <code>tick<\/code>, you should.<\/p>\n<\/blockquote><h3 id=\"bindings\">Bindings<\/h3><p>As we've seen, data can be passed down to elements and components with attributes and <a href=\"https:\/\/svelte.technology\/guide#props\">props<\/a>. Occasionally, you need to get data back up; for that we use bindings.<\/p><h4 id=\"component-bindings\">Component bindings<\/h4><p>Component bindings keep values in sync between a parent and a child:<\/p><div class=\"code-block named\">App.html<pre>&lt;Widget bind:childValue=parentValue\/&gt;<\/pre><\/div><p>Whenever <code>childValue<\/code> changes in the child component, <code>parentValue<\/code> will be updated in the parent component and vice versa.<\/p><p>If the names are the same, you can shorten the declaration:<\/p><div class=\"code-block named\">App.html<pre>&lt;Widget bind:value\/&gt;<\/pre><\/div><blockquote>\n<p>Use component bindings judiciously. They can save you a lot of boilerplate, but will make it harder to reason about data flow within your application if you overuse them.<\/p>\n<\/blockquote><h4 id=\"element-bindings\">Element bindings<\/h4><p>Element bindings make it easy to respond to user interactions:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1ezd3ds\">REPL<\/a>App.html<pre>&lt;h1&gt;Hello {name}!&lt;\/h1&gt;\n&lt;input bind:value=name&gt;<\/pre><\/div><p>Some bindings are <em>one-way<\/em>, meaning that the values are read-only. Most are <em>two-way<\/em> \u2014 changing the data programmatically will update the DOM. The following bindings are available:<\/p><table><thead><tr><th>Name<\/th>\n<th>Applies to<\/th>\n<th>Kind<\/th>\n<\/tr><\/thead><tbody><tr><td><code>value<\/code><\/td>\n<td><code>&lt;input&gt;<\/code> <code>&lt;textarea&gt;<\/code> <code>&lt;select&gt;<\/code><\/td>\n<td>Two-way<\/td>\n<\/tr><tr><td><code>checked<\/code> <code>indeterminate<\/code><\/td>\n<td><code>&lt;input type=checkbox&gt;<\/code><\/td>\n<td>Two-way<\/td>\n<\/tr><tr><td><code>group<\/code> (see note)<\/td>\n<td><code>&lt;input type=checkbox&gt;<\/code> <code>&lt;input type=radio&gt;<\/code><\/td>\n<td>Two-way<\/td>\n<\/tr><tr><td><code>currentTime<\/code> <code>paused<\/code> <code>played<\/code> <code>volume<\/code><\/td>\n<td><code>&lt;audio&gt;<\/code> <code>&lt;video&gt;<\/code><\/td>\n<td>Two-way<\/td>\n<\/tr><tr><td><code>buffered<\/code> <code>duration<\/code> <code>seekable<\/code><\/td>\n<td><code>&lt;audio&gt;<\/code> <code>&lt;video&gt;<\/code><\/td>\n<td>One-way<\/td>\n<\/tr><tr><td><code>offsetWidth<\/code> <code>offsetHeight<\/code> <code>clientWidth<\/code> <code>clientHeight<\/code><\/td>\n<td>All block-level elements<\/td>\n<td>One-way<\/td>\n<\/tr><tr><td><code>scrollX<\/code> <code>scrollY<\/code><\/td>\n<td><code>&lt;svelte:window&gt;<\/code><\/td>\n<td>Two-way<\/td>\n<\/tr><tr><td><code>online<\/code> <code>innerWidth<\/code> <code>innerHeight<\/code> <code>outerWidth<\/code> <code>outerHeight<\/code><\/td>\n<td><code>&lt;svelte:window&gt;<\/code><\/td>\n<td>One-way<\/td>\n<\/tr><\/tbody><\/table><blockquote>\n<p>'group' bindings allow you to capture the current value of a <a href=\"https:\/\/svelte.technology\/repl?demo=binding-input-radio\">set of radio inputs<\/a>, or all the selected values of a <a href=\"https:\/\/svelte.technology\/repl?demo=binding-input-checkbox-group\">set of checkbox inputs<\/a>.<\/p>\n<\/blockquote><p>Here is a complete example of using two way bindings with a form:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1fsqm69\">REPL<\/a>App.html<pre>&lt;form on:submit=\"handleSubmit(event)\"&gt;\n\t&lt;input bind:value=name type=text&gt;\n\t&lt;button type=submit&gt;Say hello&lt;\/button&gt;\n&lt;\/form&gt;\n&lt;script&gt;\n\texport default {\n\t\tmethods: {\n\t\t\thandleSubmit(event) {\n\t\t\t\t\/\/ prevent the page from reloading\n\t\t\t\tevent.preventDefault();\n\t\t\t\tconst { name } = this.get();\n\t\t\t\talert(`Hello ${name}!`);\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><blockquote>\n<p>'two way' bindings allow you to update a value in a nested property as seen in <a href=\"https:\/\/svelte.technology\/repl?demo=binding-input-checkbox\">checkbox input<\/a>.<\/p>\n<\/blockquote><h3 id=\"actions\">Actions<\/h3><p>Actions let you decorate elements with additional functionality. Actions are functions which may return an object with lifecycle methods, <code>update<\/code> and <code>destroy<\/code>. The action will be called when its element is added to the DOM.<\/p><p>Use actions for things like:<\/p><ul><li>tooltips<\/li>\n<li>lazy loading images as the page is scrolled, e.g. <code>&lt;img use:lazyload data-src='giant-photo.jpg'\/&gt;<\/code><\/li>\n<li>capturing link clicks for your client router<\/li>\n<li>adding drag and drop<\/li>\n<\/ul><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1r7zre4\">REPL<\/a>App.html<pre>&lt;button on:click=\"toggleLanguage()\" use:tooltip=\"translations[language].tooltip\"&gt;\n\t{language}\n&lt;\/button&gt;\n&lt;script&gt;\n\texport default {\n\t\tactions: {\n\t\t\ttooltip(node, text) {\n\t\t\t\tconst tooltip = document.createElement('div');\n\t\t\t\ttooltip.textContent = text;\n\t\t\t\tObject.assign(tooltip.style, {\n\t\t\t\t\tposition: 'absolute',\n\t\t\t\t\tbackground: 'black',\n\t\t\t\t\tcolor: 'white',\n\t\t\t\t\tpadding: '0.5em 1em',\n\t\t\t\t\tfontSize: '12px',\n\t\t\t\t\tpointerEvents: 'none',\n\t\t\t\t\ttransform: 'translate(5px, -50%)',\n\t\t\t\t\tborderRadius: '2px',\n\t\t\t\t\ttransition: 'opacity 0.4s'\n\t\t\t\t});\n\t\t\t\tfunction position() {\n\t\t\t\t\tconst { top, right, bottom } = node.getBoundingClientRect();\n\t\t\t\t\ttooltip.style.top = `${(top + bottom) \/ 2}px`;\n\t\t\t\t\ttooltip.style.left = `${right}px`;\n\t\t\t\t}\n\t\t\t\tfunction append() {\n\t\t\t\t\tdocument.body.appendChild(tooltip);\n\t\t\t\t\ttooltip.style.opacity = 0;\n\t\t\t\t\tsetTimeout(() =&gt; tooltip.style.opacity = 1);\n\t\t\t\t\tposition();\n\t\t\t\t}\n\t\t\t\tfunction remove() {\n\t\t\t\t\ttooltip.remove();\n\t\t\t\t}\n\t\t\t\tnode.addEventListener('mouseenter', append);\n\t\t\t\tnode.addEventListener('mouseleave', remove);\n\t\t\t\treturn {\n\t\t\t\t\tupdate(text) {\n\t\t\t\t\t\ttooltip.textContent = text;\n\t\t\t\t\t\tposition();\n\t\t\t\t\t},\n\t\t\t\t\tdestroy() {\n\t\t\t\t\t\ttooltip.remove();\n\t\t\t\t\t\tnode.removeEventListener('mouseenter', append);\n\t\t\t\t\t\tnode.removeEventListener('mouseleave', remove);\n\t\t\t\t\t}\n\t\t\t\t}\n\t\t\t}\n\t\t},\n\t\tmethods: {\n\t\t\ttoggleLanguage() {\n\t\t\t\tconst { language } = this.get();\n\t\t\t\tthis.set({\n\t\t\t\t\tlanguage: language === 'english' ? 'latin' : 'english'\n\t\t\t\t});\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"classes\">Classes<\/h3><p>Classes let you toggle element classes on and off. To use classes add the directive <code>class<\/code> followed by a colon and the class name you want toggled (<code>class:the-class-name=\"anExpression\"<\/code>). The expression inside the directive's quotes will be evaluated and toggle the class on and off depending on the truthiness of the expression's result. You can only add class directives to elements.<\/p><p>This example adds the class <code>active<\/code> to <code>&lt;li&gt;<\/code> elements when the <code>url<\/code> property matches the path their links target.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=161vsu\">REPL<\/a>App.html<pre>&lt;ul class=\"links\"&gt;\n\t&lt;li class:active=\"url === '\/'\"&gt;&lt;a href=\"\/\" on:click=\"goto(event)\"&gt;Home&lt;\/a&gt;&lt;\/li&gt;\n\t&lt;li class:active=\"url.startsWith('\/blog')\"&gt;&lt;a href=\"\/blog\/\" on:click=\"goto(event)\"&gt;Blog&lt;\/a&gt;&lt;\/li&gt;\n\t&lt;li class:active=\"url.startsWith('\/about')\"&gt;&lt;a href=\"\/about\/\" on:click=\"goto(event)\"&gt;About&lt;\/a&gt;&lt;\/li&gt;\n&lt;\/ul&gt;\n&lt;script&gt;\n\texport default {\n\t\tmethods: {\n\t\t\tgoto(event) {\n\t\t\t\tevent.preventDefault();\n\t\t\t\tthis.set({ url: event.target.pathname });\n\t\t\t}\n\t\t}\n\t}\n&lt;\/script&gt;\n&lt;style&gt;\n\t.links {\n\t\tlist-style: none;\n\t}\n\t.links li {\n\t\tfloat: left;\n\t\tpadding: 10px;\n\t}\n\t\/* classes added this way are processed with encapsulated styles, no need for :global() *\/\n\t.active {\n\t\tbackground: #eee;\n\t}\n&lt;\/style&gt;<\/pre><\/div><p>Classes will work with an existing class attribute on your element. If you find yourself adding multiple ternary statements inside a class attribute, classes can simplify your component. Classes are recognized by the compiler and <a href=\"https:\/\/svelte.technology\/guide#scoped-styles\">scoped correctly<\/a>.<\/p><p>If your class name is the same as a property in your component's state, you can use the shorthand of a class binding which drops the expression (<code>class:myProp<\/code>).<\/p><p>Note that class names with dashes in them do not usually make good shorthand classes since the property will also need a dash in it. The example below uses a computed property to make working with this easier, but it may be easier to not use the shorthand in cases like this.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=bz2ba6\">REPL<\/a>App.html<pre>&lt;div class:active class:is-selected class:isAdmin&gt;\n\t&lt;p&gt;Active? {active}&lt;\/p&gt;\n\t&lt;p&gt;Selected? {isSelected}&lt;\/p&gt;\n&lt;\/div&gt;\n&lt;button on:click=\"set({ active: !active })\"&gt;Toggle Active&lt;\/button&gt;\n&lt;button on:click=\"set({ isSelected: !isSelected })\"&gt;Toggle Selected&lt;\/button&gt;\n&lt;button on:click=\"set({ isAdmin: !isAdmin })\"&gt;Toggle Admin&lt;\/button&gt;\n&lt;script&gt;\nexport default {\n\tcomputed: {\n\t\t\/\/ Because shorthand relfects the var name, you must use component.set({ \"is-selected\": true }) or use a computed\n\t\t\/\/ property like this. It might be better to avoid shorthand for class names which are not valid variable names.\n\t\t\"is-selected\": ({ isSelected }) =&gt; isSelected\n\t}\n}\n&lt;\/script&gt;\n&lt;style&gt;\n\tdiv {\n\t\twidth: 300px;\n\t\tborder: 1px solid #ccc;\n\t\tbackground: #eee;\n\t\tmargin-bottom: 10px;\n\t}\n\t.active {\n\t\tbackground: #fff;\n\t}\n\t.is-selected {\n\t\tborder-color: #99bbff;\n\t\tbox-shadow: 0 0 6px #99bbff;\n\t}\n\t.isAdmin {\n\t\toutline: 2px solid red;\n\t}\n&lt;\/style&gt;<\/pre><\/div><\/section><section id=\"plugins\" class=\"svelte-1kk9xxo\"><p>Svelte can be extended with plugins and extra methods.<\/p><h3 id=\"transition-plugins\">Transition plugins<\/h3><p>The <a href=\"https:\/\/github.com\/sveltejs\/svelte-transitions\">svelte-transitions<\/a> package includes a selection of officially supported transition plugins, such as <a href=\"https:\/\/github.com\/sveltejs\/svelte-transitions-fade\">fade<\/a>, <a href=\"https:\/\/github.com\/sveltejs\/svelte-transitions-fly\">fly<\/a> and <a href=\"https:\/\/github.com\/sveltejs\/svelte-transitions-slide\">slide<\/a>. You can include them in a component like so:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=12xwjjo\">REPL<\/a>App.html<pre>&lt;label&gt;\n\t&lt;input type=checkbox bind:checked=visible&gt; visible\n&lt;\/label&gt;\n{#if visible}\n\t&lt;!-- use `in`, `out`, or `transition` (bidirectional) --&gt;\n\t&lt;div transition:fly=\"{y:20}\"&gt;hello!&lt;\/div&gt;\n{\/if}\n&lt;script&gt;\n\timport { fly } from 'svelte-transitions';\n\texport default {\n\t\ttransitions: { fly }\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>The <a href=\"https:\/\/github.com\/sveltejs\/svelte-extras\">svelte-extras<\/a> package includes a handful of methods for tweening (animating), manipulating arrays and so on.<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=kub6h0\">REPL<\/a>App.html<pre>&lt;input bind:value=newTodo placeholder=\"buy milk\"&gt;\n&lt;button on:click=\"push('todos', newTodo)\"&gt;add todo&lt;\/button&gt;\n&lt;ul&gt;\n\t{#each todos as todo, i}\n\t\t&lt;li&gt;\n\t\t\t&lt;button on:click=\"splice('todos', i, 1)\"&gt;x&lt;\/button&gt;\n\t\t\t{todo}\n\t\t&lt;\/li&gt;\n\t{\/each}\n&lt;\/ul&gt;\n&lt;style&gt;\n\tul {\n\t\tlist-style: none;\n\t\tpadding: 0;\n\t}\n\tli button {\n\t\tcolor: rgb(200,0,0);\n\t\tbackground: rgba(200,0,0,0.1);\n\t\tborder-color: rgba(200,0,0,0.2);\n\t\tpadding: 0.2em 0.5em;\n\t}\n&lt;\/style&gt;\n&lt;script&gt;\n\timport { push, splice } from 'svelte-extras';\n\texport default {\n\t\tdata() {\n\t\t\treturn {\n\t\t\t\tnewTodo: '',\n\t\t\t\ttodos: []\n\t\t\t};\n\t\t},\n\t\tmethods: {\n\t\t\tpush,\n\t\t\tsplice\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><\/section><section id=\"static-properties\" class=\"svelte-1kk9xxo\"><h3 id=\"setup\">Setup<\/h3><p>In some situations, you might want to add static properties to your component constructor. For that, we use the <code>setup<\/code> property:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=rsdt2r\">REPL<\/a>App.html<pre>&lt;p&gt;check the console!&lt;\/p&gt;\n&lt;script&gt;\n\texport default {\n\t\tsetup(MyComponent) {\n\t\t\t\/\/ someone importing this component will be able\n\t\t\t\/\/ to access any properties or methods defined here:\n\t\t\t\/\/\n\t\t\t\/\/   import MyComponent from '.\/MyComponent.html';\n\t\t\t\/\/   console.log(MyComponent.ANSWER); \/\/ 42\n\t\t\tMyComponent.ANSWER = 42;\n\t\t},\n\t\toncreate() {\n\t\t\tconsole.log('the answer is', this.constructor.ANSWER);\n\t\t\tconsole.dir(this.constructor);\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"preload\">preload<\/h3><p>If your component definition includes a <code>preload<\/code> function, it will be attached to the component constructor as a static method. It doesn't change the behaviour of your component in any way \u2014 instead, it's a convention that allows systems like <a href=\"https:\/\/sapper.svelte.technology\">Sapper<\/a> to delay rendering of a component until its data is ready.<\/p><p>See the section on <a href=\"https:\/\/sapper.svelte.technology\/guide#preloading\">preloading<\/a> in the Sapper docs for more information.<\/p><\/section><section id=\"server-side-rendering\" class=\"svelte-1kk9xxo\"><p>So far, we've discussed creating Svelte components <em>on the client<\/em>, which is to say the browser. But you can also render Svelte components in Node.js. This can result in better perceived performance as it means the application starts rendering while the page is still downloading, before any JavaScript executes. It also has SEO advantages in some cases, and can be beneficial for people running older browsers that can't or won't run your JavaScript for whatever reason.<\/p><h3 id=\"using-the-compiler\">Using the compiler<\/h3><p>If you're using the Svelte compiler, whether directly or via an integration like <a href=\"https:\/\/github.com\/rollup\/rollup-plugin-svelte\">rollup-plugin-svelte<\/a> or <a href=\"https:\/\/github.com\/sveltejs\/svelte-loader\">svelte-loader<\/a>, you can tell it to generate a server-side component by passing the <code>generate: 'ssr'<\/code> option:<\/p><div class=\"code-block\"><pre>const { js } = svelte.compile(source, {\n\tgenerate: 'ssr' \/\/ as opposed to 'dom', the default\n});<\/pre><\/div><h3 id=\"registering-svelte\">Registering Svelte<\/h3><p>Alternatively, an easy way to use the server-side renderer is to <em>register<\/em> it:<\/p><div class=\"code-block\"><pre>require('svelte\/ssr\/register');<\/pre><\/div><p>Now you can <code>require<\/code> your components:<\/p><div class=\"code-block\"><pre>const Thing = require('.\/components\/Thing.html');<\/pre><\/div><p>If you prefer to use a different file extension, you can pass options like so:<\/p><div class=\"code-block\"><pre>require('svelte\/ssr\/register')({\n\textensions: ['.svelte']\n});<\/pre><\/div><h3 id=\"server-side-api\">Server-side API<\/h3><p>Components have a different API in Node.js \u2013 rather than creating instances with <code>set(...)<\/code> and <code>get()<\/code> methods, a component is an object with a <code>render(data, options)<\/code> method:<\/p><div class=\"code-block\"><pre>require('svelte\/ssr\/register');\nconst Thing = require('.\/components\/Thing.html');\nconst data = { answer: 42 };\nconst { html, css, head } = Thing.render(data);<\/pre><\/div><p>All your <a href=\"https:\/\/svelte.technology\/guide#default-data\">default data<\/a>, <a href=\"https:\/\/svelte.technology\/guide#computed-properties\">computed properties<\/a>, <a href=\"https:\/\/svelte.technology\/guide#helpers\">helpers<\/a> and <a href=\"https:\/\/svelte.technology\/guide#nested-components\">nested components<\/a> will work as expected.<\/p><p>Any <code>oncreate<\/code> functions or component methods will <em>not<\/em> run \u2014 these only apply to client-side components.<\/p><blockquote>\n<p>The SSR compiler will generate a CommonJS module for each of your components \u2013 meaning that <code>import<\/code> and <code>export<\/code> statements are converted into their <code>require<\/code> and <code>module.exports<\/code> equivalents. If your components have non-component dependencies, they must also work as CommonJS modules in Node. If you're using ES2015 modules, we recommend the <a href=\"https:\/\/github.com\/standard-things\/esm\"><code>esm<\/code><\/a> module for automatically converting them to CommonJS.<\/p>\n<\/blockquote><h4 id=\"using-stores\">Using stores<\/h4><p>If your components use <a href=\"https:\/\/svelte.technology\/guide#state-management\">stores<\/a>, use the second argument:<\/p><div class=\"code-block\"><pre>const { Store } = require('svelte\/store.umd.js');\nconst { html } = Thing.render(data, {\n\tstore: new Store({\n\t\tfoo: 'bar'\n\t})\n});<\/pre><\/div><h4 id=\"rendering-styles\">Rendering styles<\/h4><p>You can also extract any <a href=\"https:\/\/svelte.technology\/guide#scoped-styles\">scoped styles<\/a> that are used by the component or its children:<\/p><div class=\"code-block\"><pre>const { css } = Thing.render(data);<\/pre><\/div><p>You could put the resulting <code>css<\/code> in a separate stylesheet, or include them in the page inside a <code>&lt;style&gt;<\/code> tag. If you do this, you will probably want to prevent the client-side compiler from including the CSS again. For the CLI, use the <code>--no-css<\/code> flag. In build tool integrations like <code>rollup-plugin-svelte<\/code>, pass the <code>css: false<\/code> option.<\/p><h4 id=\"rendering-head-contents\">Rendering <code>&lt;head&gt;<\/code> contents<\/h4><p>If your component, any of its children, use the <code>&lt;svelte:head&gt;<\/code> <a href=\"https:\/\/svelte.technology\/guide#-head-tags\">component<\/a>, you can extract the contents:<\/p><div class=\"code-block\"><pre>const { head } = Thing.render(data);<\/pre><\/div><\/section><section id=\"state-management\" class=\"svelte-1kk9xxo\"><p>Svelte components have built-in state management via the <code>get<\/code> and <code>set<\/code> methods. But as your application grows beyond a certain size, you may find that passing data between components becomes laborious.<\/p><p>For example, you might have an <code>&lt;Options&gt;<\/code> component inside a <code>&lt;Sidebar&gt;<\/code> component that allows the user to control the behaviour of a <code>&lt;MainView&gt;<\/code> component. You could use bindings or events to 'send' information up from <code>&lt;Options&gt;<\/code> through <code>&lt;Sidebar&gt;<\/code> to a common ancestor \u2014 say <code>&lt;App&gt;<\/code> \u2014 which would then have the responsibility of sending it back down to <code>&lt;MainView&gt;<\/code>. But that's cumbersome, especially if you decide you want to break <code>&lt;MainView&gt;<\/code> up into a set of smaller components.<\/p><p>Instead, a popular solution to this problem is to use a <em>global store<\/em> of data that cuts across your component hierarchy. React has <a href=\"https:\/\/redux.js.org\/\">Redux<\/a> and <a href=\"https:\/\/mobx.js.org\/index.html\">MobX<\/a> (though these libraries can be used anywhere, including with Svelte), and Vue has <a href=\"https:\/\/vuex.vuejs.org\/en\/\">Vuex<\/a>.<\/p><p>Svelte has <code>Store<\/code>. <code>Store<\/code> can be used in any JavaScript app, but it's particularly well-suited to Svelte apps.<\/p><h3 id=\"the-basics\">The basics<\/h3><p>Import <code>Store<\/code> from <code>svelte\/store.js<\/code> (remember to include the curly braces, as it's a <em>named import<\/em>), then create a new store with some (optional) data:<\/p><div class=\"code-block\"><pre>import { Store } from 'svelte\/store.js';\nconst store = new Store({\n\tname: 'world'\n});<\/pre><\/div><p>Each instance of <code>Store<\/code> has <code>get<\/code>, <code>set<\/code>, <code>on<\/code> and <code>fire<\/code> methods that behave identically to their counterparts on a Svelte component:<\/p><div class=\"code-block\"><pre>const { name } = store.get(); \/\/ 'world'\nstore.on('state', ({ current, changed, previous }) =&gt; {\n\tconsole.log(`hello ${current.name}`);\n});\nstore.set({ name: 'everybody' }); \/\/ 'hello everybody'<\/pre><\/div><h3 id=\"creating-components-with-stores\">Creating components with stores<\/h3><p>Let's adapt our <a href=\"https:\/\/svelte.technology\/guide#understanding-svelte-components\">very first example<\/a>:<\/p><div class=\"code-block named\">App.html<pre>&lt;h1&gt;Hello {$name}!&lt;\/h1&gt;\n&lt;Greeting\/&gt;\n&lt;script&gt;\n\timport Greeting from '.\/Greeting.html';\n\texport default {\n\t\tcomponents: { Greeting }\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Greeting.html<pre>&lt;p&gt;It's nice to see you, {$name}&lt;\/p&gt;<\/pre><\/div><div class=\"code-block named\">main.js<pre>import App from '.\/App.html';\nimport { Store } from 'svelte\/store.js';\nconst store = new Store({\n\tname: 'world'\n});\nconst app = new App({\n\ttarget: document.querySelector('main'),\n\tstore\n});\nwindow.store = store; \/\/ useful for debugging!<\/pre><\/div><p>There are three important things to notice:<\/p><ul><li>We're passing <code>store<\/code> to <code>new App(...)<\/code> instead of <code>data<\/code><\/li>\n<li>The template refers to <code>$name<\/code> instead of <code>name<\/code>. The <code>$<\/code> prefix tells Svelte that <code>name<\/code> is a <em>store property<\/em><\/li>\n<li>Because <code>&lt;Greeting&gt;<\/code> is a child of <code>&lt;App&gt;<\/code>, it also has access to the store. Without it, <code>&lt;App&gt;<\/code> would have to pass the <code>name<\/code> property down as a component property (<code>&lt;Greeting name={name}\/&gt;<\/code>)<\/li>\n<\/ul><p>Components that depend on store properties will re-render whenever they change.<\/p><h3 id=\"declarative-stores\">Declarative stores<\/h3><p>As an alternative to adding the <code>store<\/code> option when instantiating, the component itself can declare a dependency on a store:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=1kfmpbp\">REPL<\/a>App.html<pre>&lt;h1&gt;Hello {$name}!&lt;\/h1&gt;\n&lt;Greeting\/&gt;\n&lt;script&gt;\n\timport Greeting from '.\/Greeting.html';\n\timport store from '.\/store.js';\n\texport default {\n\t\tstore: () =&gt; store,\n\t\tcomponents: { Greeting }\n\t};\n&lt;\/script&gt;<\/pre><\/div><div class=\"code-block named\">Greeting.html<pre>&lt;p&gt;It's nice to see you, {$name}&lt;\/p&gt;<\/pre><\/div><div class=\"code-block named\">store.js<pre>import { Store } from 'svelte\/store.js';\nexport default new Store({ name: 'world' });<\/pre><\/div><p>Note that the <code>store<\/code> option is a function that <em>returns<\/em> a store, rather than the store itself \u2014 this provides greater flexibility.<\/p><h3 id=\"computed-store-properties\">Computed store properties<\/h3><p>Just like components, stores can have computed properties:<\/p><div class=\"code-block\"><pre>store = new Store({\n\twidth: 10,\n\theight: 10,\n\tdepth: 10,\n\tdensity: 3\n});\nstore.compute(\n\t'volume',\n\t['width', 'height', 'depth'],\n\t(width, height, depth) =&gt; width * height * depth\n);\nstore.get().volume; \/\/ 1000\nstore.set({ width: 20 });\nstore.get().volume; \/\/ 2000\nstore.compute(\n\t'mass',\n\t['volume', 'density'],\n\t(volume, density) =&gt; volume * density\n);\nstore.get().mass; \/\/ 6000<\/pre><\/div><p>The first argument is the name of the computed property. The second is an array of <em>dependencies<\/em> \u2014 these can be data properties or other computed properties. The third argument is a function that recomputes the value whenever the dependencies change.<\/p><p>A component that was connected to this store could reference <code>{$volume}<\/code> and <code>{$mass}<\/code>, just like any other store property.<\/p><h3 id=\"accessing-the-store-inside-components\">Accessing the store inside components<\/h3><p>Each component gets a reference to <code>this.store<\/code>. This allows you to attach behaviours in <code>oncreate<\/code>...<\/p><div class=\"code-block named\">App.html<pre>&lt;script&gt;\n\texport default {\n\t\toncreate() {\n\t\t\tconst listener = this.store.on('state', ({ current }) =&gt; {\n\t\t\t\t\/\/ ...\n\t\t\t});\n\t\t\t\/\/ listeners are not automatically removed \u2014 cancel\n\t\t\t\/\/ them to prevent memory leaks\n\t\t\tthis.on('destroy', listener.cancel);\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>...or call store methods in your event handlers, using the same <code>$<\/code> prefix as data properties:<\/p><div class=\"code-block named\">App.html<pre>&lt;button on:click=\"$set({ muted: true })\"&gt;\n\tMute audio\n&lt;\/button&gt;<\/pre><\/div><h3 id=\"custom-store-methods\">Custom store methods<\/h3><p><code>Store<\/code> doesn't have a concept of <em>actions<\/em> or <em>commits<\/em>, like Redux and Vuex. Instead, state is always updated with <code>store.set(...)<\/code>.<\/p><p>You can implement custom logic by subclassing <code>Store<\/code>:<\/p><div class=\"code-block\"><pre>class TodoStore extends Store {\n\taddTodo(description) {\n\t\tconst todo = {\n\t\t\tid: generateUniqueId(),\n\t\t\tdone: false,\n\t\t\tdescription\n\t\t};\n\t\tconst todos = this.get().todos.concat(todo);\n\t\tthis.set({ todos });\n\t}\n\ttoggleTodo(id) {\n\t\tconst todos = this.get().todos.map(todo =&gt; {\n\t\t\tif (todo.id === id) {\n\t\t\t\treturn {\n\t\t\t\t\tid,\n\t\t\t\t\tdone: !todo.done,\n\t\t\t\t\tdescription: todo.description\n\t\t\t\t};\n\t\t\t}\n\t\t\treturn todo;\n\t\t});\n\t\tthis.set({ todos });\n\t}\n}\nconst store = new TodoStore({\n\ttodos: []\n});\nstore.addTodo('Finish writing this documentation');<\/pre><\/div><p>Methods can update the store asynchronously:<\/p><div class=\"code-block\"><pre>class NasdaqTracker extends Store {\n\tasync fetchStockPrices(ticker) {\n\t\tconst token = this.token = {};\n\t\tconst prices = await fetch(`\/api\/prices\/${ticker}`).then(r =&gt; r.json());\n\t\tif (token !== this.token) return; \/\/ invalidated by subsequent request\n\t\tthis.set({ prices });\n\t}\n}\nconst store = new NasdaqTracker();\nstore.fetchStockPrices('AMZN');<\/pre><\/div><p>You can call these methods in your components, just like the built-in methods:<\/p><div class=\"code-block named\">App.html<pre>&lt;input\n\tplaceholder=\"Enter a stock ticker\"\n\ton:change=\"$fetchStockPrices(this.value)\"\n&gt;<\/pre><\/div><h3 id=\"store-bindings\">Store bindings<\/h3><p>You can bind to store values just like you bind to component values \u2014 just add the <code>$<\/code> prefix:<\/p><div class=\"code-block named\">App.html<pre>&lt;!-- global audio volume control --&gt;\n&lt;input bind:value=$volume type=range min=0 max=1 step=0.01&gt;<\/pre><\/div><h3 id=\"using-store-properties-in-computed-properties\">Using store properties in computed properties<\/h3><p>Just as in templates, you can access store properties in component computed properties by prefixing them with <code>$<\/code>:<\/p><div class=\"code-block named\">App.html<pre>{#if isVisible}\n\t&lt;div class=\"todo {todo.done ? 'done': ''}\"&gt;\n\t\t{todo.description}\n\t&lt;\/div&gt;\n{\/if}\n&lt;script&gt;\n\texport default {\n\t\tcomputed: {\n\t\t\t\/\/ `todo` is a component property, `$filter` is\n\t\t\t\/\/ a store property\n\t\t\tisVisible: ({ todo, $filter }) =&gt; {\n\t\t\t\tif ($filter === 'all') return true;\n\t\t\t\tif ($filter === 'done') return todo.done;\n\t\t\t\tif ($filter === 'pending') return !todo.done;\n\t\t\t}\n\t\t}\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"built-in-optimisations\">Built-in optimisations<\/h3><p>The Svelte compiler knows which store properties your components are interested in (because of the <code>$<\/code> prefix), and writes code that only listens for changes to those properties. Because of that, you needn't worry about having many properties on your store, even ones that update frequently \u2014 components that don't use them will be unaffected.<\/p><\/section><section id=\"preprocessing\" class=\"svelte-1kk9xxo\"><p>Some developers like to use non-standard languages such as <a href=\"https:\/\/pugjs.org\/api\/getting-started.html\">Pug<\/a>, <a href=\"http:\/\/sass-lang.com\/\">Sass<\/a> or <a href=\"http:\/\/coffeescript.org\/\">CoffeeScript<\/a>.<\/p><p>It's possible to use these languages, or anything else that can be converted to HTML, CSS and JavaScript, using <em>preprocessors<\/em>.<\/p><h3 id=\"svelte-preprocess\">svelte.preprocess<\/h3><p>Svelte exports a <code>preprocess<\/code> function that takes some input source code and returns a Promise for a standard Svelte component, ready to be used with <code>svelte.compile<\/code>:<\/p><div class=\"code-block\"><pre>const svelte = require('svelte');\nconst input = fs.readFileSync('App.html', 'utf-8');\nsvelte.preprocess(input, {\n\tfilename: 'App.html', \/\/ this is passed to each preprocessor\n\tmarkup: ({ content, filename }) =&gt; {\n\t\treturn {\n\t\t\tcode: '&lt;!-- some HTML --&gt;',\n\t\t\tmap: {...}\n\t\t};\n\t},\n\tstyle: ({ content, attributes, filename }) =&gt; {\n\t\treturn {\n\t\t\tcode: '\/* some CSS *\/',\n\t\t\tmap: {...}\n\t\t};\n\t},\n\tscript: ({ content, attributes, filename }) =&gt; {\n\t\treturn {\n\t\t\tcode: '\/\/ some JavaScript',\n\t\t\tmap: {...}\n\t\t};\n\t}\n}).then(preprocessed =&gt; {\n\tfs.writeFileSync('preprocessed\/App.html', preprocessed.toString());\n\tconst { js } = svelte.compile(preprocessed);\n\tfs.writeFileSync('compiled\/App.js', js.code);\n});<\/pre><\/div><p>The <code>markup<\/code> preprocessor, if specified, runs first. The <code>content<\/code> property represents the entire input string.<\/p><p>The <code>style<\/code> and <code>script<\/code> preprocessors receive the contents of the <code>&lt;style&gt;<\/code> and <code>&lt;script&gt;<\/code> elements respectively, along with any <code>attributes<\/code> on those elements (e.g. <code>&lt;style lang='scss'&gt;<\/code>).<\/p><p>All three preprocessors are optional. Each should return a <code>{ code, map }<\/code> object or a Promise that resolves to a <code>{ code, map }<\/code> object, where <code>code<\/code> is the resulting string and <code>map<\/code> is a sourcemap representing the transformation.<\/p><blockquote>\n<p>The returned <code>map<\/code> objects are not currently used by Svelte, but will be in future versions<\/p>\n<\/blockquote><h3 id=\"using-build-tools\">Using build tools<\/h3><p>Many build tool plugins, such as <a href=\"https:\/\/github.com\/rollup\/rollup-plugin-svelte\">rollup-plugin-svelte<\/a> and <a href=\"https:\/\/github.com\/sveltejs\/svelte-loader\">svelte-loader<\/a>, allow you to specify <code>preprocess<\/code> options, in which case the build tool will do the grunt work.<\/p><\/section><section id=\"advanced\" class=\"svelte-1kk9xxo\"><h3 id=\"keyed-each-blocks\">Keyed each blocks<\/h3><p>Associating a <em>key<\/em> with a block allows Svelte to be smarter about how it adds and removes items to and from a list. To do so, add an <code>(expression)<\/code> that uniquely identifies each member of the list:<\/p><div class=\"code-block named\">App.html<pre>{#each people as person (person.name)}\n\t&lt;div&gt;{person.name}&lt;\/div&gt;\n{\/each}<\/pre><\/div><p>It's easier to show the effect of this than to describe it. Open the following example in the REPL:<\/p><div class=\"code-block named\"><a class=\"open-in-repl\" href=\"https:\/\/svelte.technology\/repl?demo=bqtyjg\">REPL<\/a>App.html<pre>&lt;button on:click=\"update()\"&gt;update&lt;\/button&gt;\n&lt;section&gt;\n\t&lt;h2&gt;Keyed&lt;\/h2&gt;\n\t{#each people as person (person.name)}\n\t\t&lt;div transition:slide&gt;{person.name}&lt;\/div&gt;\n\t{\/each}\n&lt;\/section&gt;\n&lt;section&gt;\n\t&lt;h2&gt;Non-keyed&lt;\/h2&gt;\n\t{#each people as person}\n\t\t&lt;div transition:slide&gt;{person.name}&lt;\/div&gt;\n\t{\/each}\n&lt;\/section&gt;\n&lt;style&gt;\n\tbutton {\n\t\tdisplay: block;\n\t}\n\tsection {\n\t\twidth: 10em;\n\t\tfloat: left;\n\t}\n&lt;\/style&gt;\n&lt;script&gt;\n\timport { slide } from 'svelte-transitions';\n\tvar people = ['Alice', 'Barry', 'Cecilia', 'Douglas', 'Eleanor', 'Felix', 'Grace', 'Horatio', 'Isabelle'];\n\tfunction random() {\n\t\treturn people\n\t\t\t.filter(() =&gt; Math.random() &lt; 0.5)\n\t\t\t.map(name =&gt; ({ name }))\n\t}\n\texport default {\n\t\tdata() {\n\t\t\treturn { people: random() };\n\t\t},\n\t\tmethods: {\n\t\t\tupdate() {\n\t\t\t\tthis.set({ people: random() });\n\t\t\t}\n\t\t},\n\t\ttransitions: { slide }\n\t};\n&lt;\/script&gt;<\/pre><\/div><h3 id=\"hydration\">Hydration<\/h3><p>If you're using <a href=\"https:\/\/svelte.technology\/guide#server-side-rendering\">server-side rendering<\/a>, it's likely that you'll need to create a client-side version of your app <em>on top of<\/em> the server-rendered version. A naive way to do that would involve removing all the existing DOM and rendering the client-side app in its place:<\/p><div class=\"code-block\"><pre>import App from '.\/App.html';\nconst target = document.querySelector('#element-with-server-rendered-html');\n\/\/ avoid doing this!\ntarget.innerHTML = '';\nnew App({\n\ttarget\n});<\/pre><\/div><p>Ideally, we want to reuse the existing DOM instead. This process is called <em>hydration<\/em>. First, we need to tell the compiler to include the code necessary for hydration to work by passing the <code>hydratable: true<\/code> option:<\/p><div class=\"code-block\"><pre>const { js } = svelte.compile(source, {\n\thydratable: true\n});<\/pre><\/div><p>(Most likely, you'll be passing this option to <a href=\"https:\/\/github.com\/rollup\/rollup-plugin-svelte\">rollup-plugin-svelte<\/a> or <a href=\"https:\/\/github.com\/sveltejs\/svelte-loader\">svelte-loader<\/a>.)<\/p><p>Then, when we instantiate the client-side component, we tell it to use the existing DOM with <code>hydrate: true<\/code>:<\/p><div class=\"code-block\"><pre>import App from '.\/App.html';\nconst target = document.querySelector('#element-with-server-rendered-html');\nnew App({\n\ttarget,\n\thydrate: true\n});<\/pre><\/div><blockquote>\n<p>It doesn't matter if the client-side app doesn't perfectly match the server-rendered HTML \u2014 Svelte will repair the DOM as it goes.<\/p>\n<\/blockquote><h3 id=\"immutable\">Immutable<\/h3><p>Because arrays and objects are <em>mutable<\/em>, Svelte must err on the side of caution when deciding whether or not to update things that refer to them.<\/p><p>But if all your data is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Immutable_object\">immutable<\/a>, you can use the <code>{ immutable: true }<\/code> compiler option to use strict object comparison (using <code>===<\/code>) everywhere in your app. If you have one component that uses immutable data you can set it to use the strict comparison for just that component.<\/p><p>In the example below, <code>searchResults<\/code> would normally be recalculated whenever <code>items<\/code> <em>might<\/em> have changed, but with <code>immutable: true<\/code> it will only update when <code>items<\/code> has <em>definitely<\/em> changed. This can improve the performance of your app.<\/p><div class=\"code-block named\">App.html<pre>{#each searchResults as item}\n\t&lt;div&gt;{item.name}&lt;\/div&gt;\n{\/each}\n&lt;script&gt;\n\timport FuzzySearch from 'fuzzy-search';\n\texport default {\n\t\timmutable: true,\n\t\tcomputed: {\n\t\t\tsearchResults: ({ searchString, items }) =&gt; {\n\t\t\t\tif (!searchString) return items;\n\t\t\t\tconst searcher = new FuzzySearch(items, ['name', 'location']);\n\t\t\t\treturn searcher.search(searchString);\n\t\t\t}\n\t\t}\n\t}\n&lt;\/script&gt;<\/pre><\/div><p><a href=\"https:\/\/svelte.technology\/repl?demo=immutable\">Here's a live example<\/a> showing the effect of <code>immutable: true<\/code>.<\/p><\/section><section id=\"custom-elements\" class=\"svelte-1kk9xxo\"><p><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Web_Components\/Custom_Elements\">Custom elements<\/a> are an emerging web standard for creating DOM elements that encapsulate styles and behaviours, much like Svelte components. They are part of the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Web_Components\">web components<\/a> family of specifications.<\/p><blockquote>\n<p>Most browsers need <a href=\"https:\/\/www.webcomponents.org\/polyfills\">polyfills<\/a> for custom elements. See <a href=\"https:\/\/caniuse.com\/#feat=custom-elementsv1\">caniuse<\/a> for more details<\/p>\n<\/blockquote><p>Svelte components can be used as custom elements by doing the following:<\/p><ol><li>Declaring a <code>tag<\/code> name. The value must contain a hyphen (<code>hello-world<\/code> in the example below)<\/li>\n<li>Specifying <code>customElement: true<\/code> in the compiler configuration<\/li>\n<\/ol><div class=\"code-block named\">HelloWorld.html<pre>&lt;h1&gt;Hello {name}!&lt;\/h1&gt;\n&lt;script&gt;\n\texport default {\n\t\ttag: 'hello-world'\n\t};\n&lt;\/script&gt;<\/pre><\/div><p>Importing this file will now register a globally-available <code>&lt;hello-world&gt;<\/code> custom element that accepts a <code>name<\/code> property:<\/p><div class=\"code-block\"><pre>import '.\/HelloWorld.html';\ndocument.body.innerHTML = `&lt;hello-world name=\"world\"\/&gt;`;\nconst el = document.querySelector('hello-world');\nel.name = 'everybody';<\/pre><\/div><p>See <a href=\"http:\/\/svelte-custom-elements.surge.sh\/\">svelte-custom-elements.surge.sh<\/a> (<a href=\"https:\/\/github.com\/sveltejs\/template-custom-element\">source here<\/a>) for a larger example.<\/p><p>The compiled custom elements are still full-fledged Svelte components and can be used as such:<\/p><div class=\"code-block\"><pre>el.get().name === el.name; \/\/ true\nel.set({ name: 'folks' }); \/\/ equivalent to el.name = 'folks'<\/pre><\/div><p>One crucial difference is that styles are <em>fully encapsulated<\/em> \u2014 whereas Svelte will prevent component styles from leaking <em>out<\/em>, custom elements use <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Web_Components\/Shadow_DOM\">shadow DOM<\/a> which also prevents styles from leaking <em>in<\/em>.<\/p><h3 id=\"using-slot\">Using <code>&lt;slot&gt;<\/code><\/h3><p>Custom elements can use <a href=\"https:\/\/svelte.technology\/guide#composing-with-slot\">slots<\/a> to place child elements, just like regular Svelte components.<\/p><h3 id=\"firing-events\">Firing events<\/h3><p>You can dispatch events inside custom elements to pass data out:<\/p><div class=\"code-block\"><pre>\/\/ inside a component method\nconst event = new CustomEvent('message', {\n\tdetail: 'Hello parent!',\n\tbubbles: true,\n\tcancelable: true,\n\tcomposed: true \/\/ makes the event jump shadow DOM boundary\n});\nthis.dispatchEvent(event);<\/pre><\/div><p>Other parts of the application can listen for these events with <code>addEventListener<\/code>:<\/p><div class=\"code-block\"><pre>const el = document.querySelector('hello-world');\nel.addEventListener('message', event =&gt; {\n\talert(event.detail);\n});<\/pre><\/div><blockquote>\n<p>Note the <code>composed: true<\/code> attribute of the custom event. It enables the custom DOM event to cross the shadow DOM boundary and enter into main DOM tree.<\/p>\n<\/blockquote><h3 id=\"observing-properties\">Observing properties<\/h3><p>Svelte will determine, from the template and <code>computed<\/code> values, which properties the custom element has \u2014 for example, <code>name<\/code> in our <code>&lt;hello-world&gt;<\/code> example. You can specify this list of properties manually, for example to restrict which properties are 'visible' to the rest of your app:<\/p><div class=\"code-block\"><pre>export default {\n\ttag: 'my-thing',\n\tprops: ['foo', 'bar']\n};<\/pre><\/div><h3 id=\"compiler-options\">Compiler options<\/h3><p>Earlier, we saw the use of <code>customElement: true<\/code> to instruct the Svelte compiler to generate a custom element using the <code>tag<\/code> and (optional) <code>props<\/code> declared inside the component file.<\/p><p>Alternatively, you can pass <code>tag<\/code> and <code>props<\/code> direct to the compiler:<\/p><div class=\"code-block\"><pre>const { js } = svelte.compile(source, {\n\tcustomElement: {\n\t\ttag: 'overridden-tag-name',\n\t\tprops: ['yar', 'boo']\n\t}\n});<\/pre><\/div><p>These options will override the component's own settings, if any.<\/p><h3 id=\"transpiling\">Transpiling<\/h3><ul><li><p>Custom elements use ES2015 classes (<code>MyThing extends HTMLElement<\/code>). Make sure you don't transpile the custom element code to ES5, and use a ES2015-aware minifier such as <a href=\"https:\/\/www.npmjs.com\/package\/uglify-es\">uglify-es<\/a>.<\/p>\n<\/li>\n<li><p>If you do need ES5 support, make sure to use <code>Reflect.construct<\/code> aware transpiler plugin such as <a href=\"https:\/\/github.com\/WebReflection\/babel-plugin-transform-builtin-classes\">babel-plugin-transform-builtin-classes<\/a> and a polyfill like <a href=\"https:\/\/github.com\/webcomponents\/webcomponentsjs#custom-elements-es5-adapterjs\">custom-elements-es5-adapterjs<\/a>.<\/p>\n<\/li>\n<\/ul><\/section><section id=\"miscellaneous\" class=\"svelte-1kk9xxo\"><h3 id=\"noscript\"><code>&lt;noscript&gt;<\/code><\/h3><p>If you use <code>&lt;noscript&gt;<\/code> tags in a component, Svelte will only render them in SSR mode. The DOM compiler will strip them out, since you can't create the component without JavaScript, and <code>&lt;noscript&gt;<\/code> has no effect if JavaScript is available.<\/p><\/section><section id=\"still-to-come\" class=\"svelte-1kk9xxo\"><p>This documentation is still a work-in-progress, like Svelte itself. If there are particular things that are missing or could be improved, then <a href=\"https:\/\/github.com\/sveltejs\/svelte.technology\">please raise an issue on GitHub<\/a>!<\/p><\/section>","created_at":"2019-02-04T03:51:17+0000","updated_at":"2019-02-04T03:51:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":56,"domain_name":"svelte.technology","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":808,"title":"Dotfile madness","url":"https:\/\/0x46.net\/thoughts\/2019\/02\/01\/dotfile-madness\/","content":"<p>We are no longer in control of our home directories.<\/p><p>My own home directory contains 25 ordinary files and 144 hidden files. The\ndotfiles contain data that doesn't belong to me: it belongs to the programmers\nwhose programs decided to hijack the primary location designed as a storage for\nmy personal files. I can't place those dotfiles anywhere else and they will\nappear again if I try to delete them.  All I can do is sit here knowing that in\nthe darkness, behind the scenes, they are there. Waiting in silence. Some of\nthose programmers decided to additionally place some normal files and\ndirectories in the same place. Those are clearly visible every time I execute\n<code>ls<\/code> in my home directory. It is beyond me why my home directory ended up up\ncontaining a <code>node_modules<\/code> directory, <code>package-lock.json<\/code>, a <code>yarn.lock<\/code> file\n(I have never even consciously used yarn!), some 2 strange log files\norigination from some Java software clearly using an H2 database, and a\n<code>Desktop<\/code> directory. That last one has been created by Steam, which is quite\nunfortunate as I simply do not have a desktop or a desktop environment on my\nmachine. I dread the day in which I will hear a loud knock on my door and one\nof those programmers will barge in informing me that he is going to store a\npiece of his furniture in the middle of my living room, If I don't mind.<\/p><p>To those of you reading this: I beg you. Avoid creating files or directories of\nany kind in your user's <code>$HOME<\/code> directory in order to store your configuration\nor data. This practice is bizarre at best and it is time to end it. I am sorry\nto say that many (if not most) programs are guilty of doing this while there\nare significantly better places that can be used for storing per-user program\ndata.<\/p><p>Even if we will never be able to solve this problem - due to the historical\nbaggage, backward compatibility, software no longer receiving updates or\nprogramming villains storing the files wherever they want just to see the world\nburn - we can at least try to follow some sane practices. While the <a href=\"https:\/\/plus.google.com\/+RobPikeTheHuman\/posts\/R58WgWwN9jp\">mistake of\nintroducing the mere concept of a \"hidden\" file<\/a> can't\nbe undone anymore, we can at least try to mitigate its results.<\/p><p>This particular problem has been noticed and solved a long time ago with the\ncreation of <a href=\"https:\/\/specifications.freedesktop.org\/basedir-spec\/basedir-spec-latest.html\">XDG Base Directory Specification<\/a>. The\nspecification defines a set of environment variables pointing programs to a\ndirectory in which their data or configuration should be stored. It is up to\nthe user to set those variables so if the variables are not available the\nprograms are expected to default to a directory defined by the standard and not\nthe user's home directory.<\/p><h2>User-level variables<\/h2><h3><code>$XDG_DATA_HOME<\/code><\/h3><blockquote>\n<p><code>$XDG_DATA_HOME<\/code> defines the base directory relative to which user specific\ndata files should be stored. If <code>$XDG_DATA_HOME<\/code> is either not set or empty, a\ndefault equal to <code>$HOME\/.local\/share<\/code> should be used. <\/p>\n<\/blockquote><p>Example usage: storing plugins downloaded by the user, databases created by\nyour program, user's input history, bookmarks, emails, and so on.<\/p><h3><code>$XDG_CONFIG_HOME<\/code><\/h3><blockquote>\n<p><code>$XDG_CONFIG_HOME<\/code> defines the base directory relative to which user specific\nconfiguration files should be stored. If <code>$XDG_CONFIG_HOME<\/code> is either not set\nor empty, a default equal to <code>$HOME\/.config<\/code> should be used. <\/p>\n<\/blockquote><p>You should use this directory to store user-specific configuration files for\nyour program. You will most likely want to create a default configuration file\nwith sane and sensible default values the first time your program is executed.<\/p><h3><code>$XDG_CACHE_HOME<\/code><\/h3><blockquote>\n<p><code>$XDG_CACHE_HOME<\/code> defines the base directory relative to which user specific\nnon-essential data files should be stored. If <code>$XDG_CACHE_HOME<\/code> is either not\nset or empty, a default equal to $HOME\/.cache should be used.<\/p>\n<\/blockquote><p>Example: caching thumbnails generated by your file manager, songs that the user\nof a music straming software is often listening to, and so on. Your program\nshould continue to function without any problems if this directory is removed\nby the user. Ensure that the files that are no longer needed are properly\nremoved. Remember that exceeding the sensible amount of storage space used by\nyour files will most likely upset the user who will quickly track down your\nprogram as a culprit.<\/p><h3><code>$XDG_RUNTIME_DIR<\/code><\/h3><blockquote>\n<p><code>$XDG_RUNTIME_DIR<\/code> defines the base directory relative to which user-specific\nnon-essential runtime files and other file objects (such as sockets, named\npipes, ...) should be stored.<\/p>\n<\/blockquote><p>The specification lists a series of requirements that have to be fulfilled by\nthis directory. As described by the specification you should use this directory\nto store sockets and similar files used for communication.<\/p><h2>System-level variables<\/h2><h3><code>$XDG_CONFIG_DIRS<\/code><\/h3><blockquote>\n<p><code>$XDG_CONFIG_DIRS<\/code> defines the preference-ordered set of base directories to\nsearch for configuration files in addition to the <code>$XDG_CONFIG_HOME<\/code> base\ndirectory. The directories in <code>$XDG_CONFIG_DIRS<\/code> should be seperated with a\ncolon ':'.<\/p>\n<p>If <code>$XDG_CONFIG_DIRS<\/code> is either not set or empty, a value equal to <code>\/etc\/xdg<\/code>\nshould be used. <\/p>\n<\/blockquote><p>This directory should be used for storing system-wide configuration files that\ncan be overwriten by user-specific configuration files. This directory would\nmost likely be populated during the installation process.<\/p><h3><code>$XDG_DATA_DIRS<\/code><\/h3><blockquote>\n<p><code>$XDG_DATA_DIRS<\/code> defines the preference-ordered set of base directories to\nsearch for data files in addition to the <code>$XDG_DATA_HOME<\/code> base directory. The\ndirectories in <code>$XDG_DATA_DIRS<\/code> should be seperated with a colon ':'.<\/p>\n<p>If <code>$XDG_DATA_DIRS<\/code> is either not set or empty, a value equal to\n<code>\/usr\/local\/share\/:\/usr\/share\/<\/code> should be used.<\/p>\n<\/blockquote><p>Example: storing plugins or themes that can be used by all users of your\nprogram. This directory would most likely be populated during the installation\nprocess.<\/p><h2>How does this work in practice?<\/h2><p>Using the standard is very simple. Read the relevant environment variable and\nuse the default paths defined by the standard if it is missing. You should then\nappend a program-specific directory name to it and create the entire directory\ntree to store your data.<\/p><p>As an example if you were to store configuration files you should use\n<code>$XDG_CONFIG_HOME\/your-program<\/code> as your base configuraion directory instead of\njust storing your files directly in <code>$XDG_CONFIG_HOME<\/code>. Remember to never\nhardcode the directories to the default values defined by the standard. Read\nthe environment variable first to allow the user to move those directories if\nneeded.<\/p><p>You can easily migrate your existing programs to use this standard. In order to\ndo so start using the standard when creating new files but keep also checking\nthe old location of the files when reading them. This will allow you to migrate\nwithout breaking the program for the users with the configuration or data files\nthat were created by a previous version of your program.<\/p><p>Read the standard to find out more and have a look at the directory hierarchy\nwhich is almost certain to be already present in your home directory. In\nreality a cross-platform library allowing you to get a directory for storing\nyour data will be available for your programming language of choice. On Linux\nand similar systems, this library will surely use the XDG Base Directory\nSpecification.<\/p><p>2019-02-01<\/p>","created_at":"2019-02-02T16:44:29+0000","updated_at":"2019-02-02T16:44:29+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"0x46.net","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":807,"title":"How I negotiated a $300,000 offer in Silicon Valley","url":"https:\/\/blog.usejournal.com\/how-i-negotiated-a-software-engineer-offer-in-silicon-valley-f11590f5c656?gi=1a5d91f1db25","content":"<section class=\"section section--body section--first\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h2 id=\"5711\" class=\"graf graf--h4 graf-after--h3 graf--subtitle\">In the span of a week, I had job offers to work as a software engineer at Google, Facebook, Amazon, Apple, LinkedIn, and Yelp. Here\u2019s how I negotiated them.<\/h2><figure id=\"9bd3\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*TjFEp_SQMG3xh41Ly9cKig.png\" data-width=\"662\" data-height=\"503\" data-is-featured=\"true\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*TjFEp_SQMG3xh41Ly9cKig.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">What if I told you\u2026 that negotiating isn\u2019t as scary as it\u00a0sounds?<\/figcaption><\/figure><p id=\"bae5\" class=\"graf graf--p graf-after--figure\">You\u2019re <em class=\"markup--em markup--p-em\">almost<\/em> finished with the long grind of interviewing. You\u2019re <em class=\"markup--em markup--p-em\">almost <\/em>free of the stress associated with interviewing at big tech companies. You\u2019re practically envisioning sending off a final \u201cI\u2019ll sign\u201d email.<\/p><p id=\"b6dc\" class=\"graf graf--p graf-after--p\">But you\u2019re not <em class=\"markup--em markup--p-em\">quite <\/em>there yet.<\/p><p id=\"409d\" class=\"graf graf--p graf-after--p\">There\u2019s one last hurdle to clear: negotiating an offer.<\/p><p id=\"bfc9\" class=\"graf graf--p graf-after--p\">You\u2019ve shown that you can interview, but how well can you close a deal?<\/p><p id=\"53c8\" class=\"graf graf--p graf-after--p graf--trailing\"><a href=\"https:\/\/blog.usejournal.com\/i-interviewed-at-six-top-companies-in-silicon-valley-in-six-days-and-stumbled-into-six-job-offers-fe9cc7bbc996\" data-href=\"https:\/\/blog.usejournal.com\/i-interviewed-at-six-top-companies-in-silicon-valley-in-six-days-and-stumbled-into-six-job-offers-fe9cc7bbc996\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Late last year I interviewed at six top companies in Silicon Valley in six days, and stumbled into six job offers<\/a>. I had one last task to make sure my work wasn\u2019t in vain. Here are the principles I had, the rules I followed, and what I did to negotiate an offer worth $100,000\/year more than I planned on.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"2606\" class=\"graf graf--h3 graf--leading\">Introduction<\/h3><p id=\"c522\" class=\"graf graf--p graf-after--h3\">First, some assumptions on my part:<\/p><ul class=\"postList\"><li id=\"49d7\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Your mileage may vary. <\/strong>I\u2019m obviously not stating that everything I went through is easily reproduced. I <em class=\"markup--em markup--li-em\">am<\/em> using it as a backdrop to color some advice that I think anyone totally new to this to keep in mind.<\/li><li id=\"135e\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">You have offers, plural.<\/strong> This can be two or this can be ten, but you need to have more than one offer <em class=\"markup--em markup--li-em\">concurrently<\/em>. I\u2019ve negotiated on behalf of (many) friends in the past with no leverage, but I\u2019m not going to focus on that scenario.<\/li><li id=\"d317\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">You\u2019re not afraid to negotiate. <\/strong>The night before my Google interview I spoke with another candidate in my hotel. His strategy was to low-ball himself to increase his chances of getting an offer. That\u2019s not the path that I took, it\u2019s not a path I\u2019d advocate, and it\u2019s not one I\u2019m going to even entertain. You should always be willing to see what you can get.<\/li><li id=\"e853\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">You won\u2019t practice the Dark Arts of Negotiation\u2122. <\/strong>I\u2019m pretty sure you can negotiate by lying your way through the process. From a moral perspective, though, c\u2019mon. From a practical perspective, getting caught in a lie is not going to help you.<\/li><li id=\"7ba4\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">You won\u2019t renege on an offer. <\/strong>Could I have squeezed a bit more out of the eventual offer I signed if I had? Maybe. But I\u2019m not comfortable accepting an offer and later rejecting it. I\u2019m sure other candidates do it and I don\u2019t think it gets you banned from the Valley. It\u2019s just not something I wanted to do.<\/li><li id=\"cf9c\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">You\u2019re willing to work at any of the companies if the price is right. <\/strong>I definitely had a personal ranking of which companies I\u2019d prefer over each other, but each company had a chance. No one was just there to beef up offers.<\/li><li id=\"66bf\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">I\u2019m going to both anonymize the companies I negotiated with and fuzz their numbers. <\/strong>My goal is not to answer \u201chow much does MegaCorp offer if you have competing offers.\u201d My goal is to answer \u201chow do I respectfully ask for a larger comp package?\u201d for people that don\u2019t know where to start. The total compensation for each offer is accurate (within 5%).<\/li><li id=\"1805\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">You\u2019re starting at square one.<\/strong> I give my personal experience in this article and some examples of conversations I\u2019ve had when negotiating. I don\u2019t draw on decades of experience with very in-depth detail. I link some pieces at the end that are more suited towards that type of read.<\/li><\/ul><h3 id=\"ac08\" class=\"graf graf--h3 graf-after--li\">Before the\u00a0Offer<\/h3><p id=\"6c2c\" class=\"graf graf--p graf-after--h3\">Congrats! You have multiple interviews. I\u2019m sure you\u2019re nervous. I\u2019m sure you\u2019re telling yourself you don\u2019t deserve it, that you got lucky. You think you\u2019re going to fail miserably. The last thing on your mind is how to talk numbers with recruiters.<\/p><p id=\"d440\" class=\"graf graf--p graf-after--p\">Chill out. We all go through that! \ud83d\ude42<\/p><p id=\"fe15\" class=\"graf graf--p graf-after--p\">It\u2019s understandable. <strong class=\"markup--strong markup--p-strong\">But you don\u2019t want to shoot your future self in the foot<\/strong>. There\u2019s only a few things you need to keep in mind at this point, anyway.<\/p><p id=\"91e9\" class=\"graf graf--p graf-after--p\">My first major piece of advice: <strong class=\"markup--strong markup--p-strong\">do not make this process adversarial.<\/strong><\/p><p id=\"1fe3\" class=\"graf graf--p graf-after--p\">I\u2019ve read a lot of people that view this very cynically.<\/p><p id=\"6122\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">It\u2019s us against them.<\/em><\/p><p id=\"ea45\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">Candidate versus recruiter.<\/em><\/p><p id=\"bbcd\" class=\"graf graf--p graf-after--p\">You\u2019re trying to extract the absolute most out of a company and they\u2019re trying to cheat you out of the most they can. Right?<\/p><figure id=\"a1ce\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*pvIy0e97X_ZNFsVSu5ZoUw.png\" data-width=\"824\" data-height=\"402\" data-action=\"zoom\" data-action-value=\"1*pvIy0e97X_ZNFsVSu5ZoUw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*pvIy0e97X_ZNFsVSu5ZoUw.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">I bet you didn\u2019t know that Google recruiters wore capes, did\u00a0you?<\/figcaption><\/figure><p id=\"ab9e\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Wrong. <\/strong>This may be the case sometimes, I guess, but that mindset will <em class=\"markup--em markup--p-em\">not<\/em> lend itself to a healthy negotiating phase. I had very good relationships with my recruiters and in the end it paid off. Maybe it would have gone better if I were more aggressive or lied to them. I won\u2019t ever know how that alternative universe played out. But I do know that I have no regrets about it all. I\u2019m confident I could ping any of those recruiters to kick off the process again. Also, recruiters are people, too. \ud83d\ude05 Be nice!<\/p><p id=\"94f5\" class=\"graf graf--p graf-after--p\">Next, <strong class=\"markup--strong markup--p-strong\">make it clear <em class=\"markup--em markup--p-em\">throughout<\/em> the process that you\u2019re interviewing elsewhere. <\/strong>This is critical to make sure that your offers are overlapping. I had to deal with one exploding offer. I had to pass on because the company refused to move the deadline. It took a piece of leverage away from me. If I didn\u2019t let everyone know about my situation I may have lost even more.<\/p><p id=\"7b69\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">It\u2019s up to you if you want to disclose <em class=\"markup--em markup--p-em\">where<\/em> you\u2019re interviewing.<\/strong> I would share it you think it\u2019s \u201cimpressive\u201d. Keep in mind that you should be proud of interviewing <em class=\"markup--em markup--p-em\">anywhere<\/em>. You\u2019ll have more leverage, though, if you don\u2019t remove the <em class=\"markup--em markup--p-em\">possibility<\/em> of a Google. My recruiters were very respectful about asking. Companies understand that who you talk to is your business and only theirs if you decide it is. With that said\u200a\u2014\u200athe <em class=\"markup--em markup--p-em\">threat<\/em> of getting (and signing) an offer elsewhere is fantastic leverage for everything from scheduling to getting an offer.<\/p><h3 id=\"57d0\" class=\"graf graf--h3 graf-after--p\">The Initial\u00a0Offers<\/h3><p id=\"ddc0\" class=\"graf graf--p graf-after--h3\">Alright, it\u2019s the moment of truth. You\u2019re going to hear how you did. \ud83d\ude31<\/p><p id=\"8351\" class=\"graf graf--p graf-after--p\">My recruiters all opted to deliver the news over the phone. Nearly every call followed the same pattern:<\/p><blockquote id=\"4059\" class=\"graf graf--blockquote graf-after--p\"><div>Do you have any updates from other companies? Do you have a range you\u2019re expecting from us? How does something in the neighborhood of $package sound? What do you think?<\/div><\/blockquote><p id=\"1bff\" class=\"graf graf--p graf-after--blockquote\">Pretty much anything you read will say <em class=\"markup--em markup--p-em\">never give a range.<\/em> I <em class=\"markup--em markup--p-em\">mostly<\/em> agree with that, but sometimes it makes sense to give <em class=\"markup--em markup--p-em\">something<\/em>. It takes some experience and familiarity with comp packages to do so comfortably. I only gave the first numbers to one or two companies and it was because I was familiar enough with the companies to know they would come in around $50k\/yr lower than other offers I had.<\/p><p id=\"b48f\" class=\"graf graf--p graf-after--p\">I would never give a range to a company that sends out sky-high offers. With that said, <strong class=\"markup--strong markup--p-strong\">I\u2019m not going to waste a back-and-forth cycle with a recruiter when I <em class=\"markup--em markup--p-em\">know they\u2019re not going to come close to an existing offer<\/em><\/strong>. Recruiters won\u2019t want to go back-and-forth and back-and-forth writing up new offers for you. If they won\u2019t be close, let them know where the range is going to <em class=\"markup--em markup--p-em\">start<\/em>. Wasting time waiting on a non-competitive offer isn\u2019t advantageous to anyone.<\/p><p id=\"6ea9\" class=\"graf graf--p graf-after--p\">That being said, if you <em class=\"markup--em markup--p-em\">are<\/em> dealing with a company whose budget is infinite, be smart. <a href=\"http:\/\/www.levels.fyi\" data-href=\"http:\/\/www.levels.fyi\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Levels.fyi<\/a> is the best resource I\u2019ve seen (far better than Glassdoor) when it comes to learning what\u2019s typically offered. <strong class=\"markup--strong markup--p-strong\">Shoot for the sky and <em class=\"markup--em markup--p-em\">never, ever assume you can\u2019t get a top offer<\/em>.<\/strong> Watch the situation play out and be willing to adapt to it. I ended up getting far, far more than I ever though a company would offer me, and a lot of that is due to never anchoring myself mentally to a \u201ctop value\u201d. \ud83d\udc4d<\/p><figure id=\"13b2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*bfb4Q9_258ToidJL9AiEKg.png\" data-width=\"1182\" data-height=\"752\" data-action=\"zoom\" data-action-value=\"1*bfb4Q9_258ToidJL9AiEKg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*bfb4Q9_258ToidJL9AiEKg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Trust me, recruiters will do their best to limit your offer. You don\u2019t need to help them\u00a0out!<\/figcaption><\/figure><p id=\"eb37\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Never sound \u201cgiddy\u201d<\/strong>. Remain measured, even if you\u2019re confident you\u2019d accept an offer. It\u2019s a lot to take in on the phone\u200a\u2014\u200aespecially if it\u2019s an offer that surpasses what you ever expected\u200a\u2014\u200aand you\u2019re probably not thinking straight. Always think on any package for awhile.<\/p><p id=\"3106\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Make it clear you won\u2019t accept an offer until you hear back from everyone. <\/strong>In my experience, recruiters understood. I made it clear I would like to go to any of the companies I was interviewing at. I definitely had a dream company throughout the process, but that wasn\u2019t really anyone\u2019s business but my own. \ud83d\ude42<\/p><p id=\"d42b\" class=\"graf graf--p graf-after--p\">I don\u2019t see actual examples of things to actually <em class=\"markup--em markup--p-em\">say<\/em> in these situations. I\u2019ll provide here what it sounded like for me. That should help those that don\u2019t know where to start. Don\u2019t follow this as a script (because that\u2019s weird and unnatural), but this is the sort of thing I would say to remain respectful and open to negotiate. Let\u2019s assume the offer you\u2019ve heard is something you might sign (that is, it meets your bar to go work there):<\/p><blockquote id=\"c47f\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">Recruiter<\/strong>: So, what do you think?<\/div><\/blockquote><blockquote id=\"9bde\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">You<\/strong>: That\u2019s an interesting offer. I think it\u2019s probably in the ballpark, but it\u2019s not something I\u2019m going to sign right now. I\u2019m going to wait until I hear back from the other places I\u2019m speaking to.<\/div><\/blockquote><p id=\"0d3f\" class=\"graf graf--p graf-after--blockquote\">In the above, you let your recruiter know that they\u2019re \u201cin the running\u201d, but that it\u2019s not enough to blow you away. You\u2019re respectful but firm. This is the blueprint for how your discussions should go.<\/p><p id=\"ba66\" class=\"graf graf--p graf-after--p\">Don\u2019t give away more information than you need to, don\u2019t be aggressive, and don\u2019t give numbers if it doesn\u2019t make sense. The above is all honest, too, so you don\u2019t need to worry about getting caught in a lie. Essentially, you\u2019re trying to maintain good standing with your recruiters and avoid getting into the nitty-gritty until you have all of your offers on the table. That\u2019s the fun part!<\/p><h3 id=\"b73a\" class=\"graf graf--h3 graf-after--p\">The Counter-Offers<\/h3><p id=\"d9ef\" class=\"graf graf--p graf-after--h3\">So, you\u2019ve got numbers from everyone that you\u2019re considering. Now you can start counter-offering based on what information you have. <strong class=\"markup--strong markup--p-strong\">When I counter-offered, I never used solid numbers\u200a\u2014\u200aI gave ranges. <\/strong>I don\u2019t like the idea of committing to a solid number until the very end.<\/p><p id=\"d56b\" class=\"graf graf--p graf-after--p\">There\u2019s a few reasons for this:<\/p><ul class=\"postList\"><li id=\"a0f1\" class=\"graf graf--li graf-after--p\">For one, giving a solid number feels weird; why is $120,000 acceptable, but $119,000 is a bit low? <strong class=\"markup--strong markup--li-strong\">I\u2019d much rather present that as \u201csomewhere between $120,000 and $140,000\u201d.<\/strong> It also leaves room for a company to move upward if you keep things a bit in the dark. Note that once you have numbers from a company it makes a lot more sense to be more explicit in what you\u2019re looking for.<\/li><li id=\"761c\" class=\"graf graf--li graf-after--li\">To me, there\u2019s an implication when giving a solid number that it\u2019s \u201cthe one.\u201d If a company can hit \u201cyour number\u201d, you\u2019ve lost out on the chance to negotiate much more unless more offers coming in. <strong class=\"markup--strong markup--li-strong\">Companies can still hit your number if they offer the top of your range; keep that in mind.<\/strong> If you give a $5,000 range, don\u2019t be surprised if companies meet your demands and expect to get a signed offer.<\/li><\/ul><p id=\"90d2\" class=\"graf graf--p graf-after--li\">Some may object to how I presented the above, as recruiters have a lot of flexibility. You can always give a number, have them match it, and then ask for higher. I\u2019m not denying that. But it was a goal throughout my negotiations to keep on very good standing with my recruiters and it paid off very well in the end.<\/p><figure id=\"dd7d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*rMAvqPNXHPvuqDNI0Q13_w.png\" data-width=\"2388\" data-height=\"1486\" data-action=\"zoom\" data-action-value=\"1*rMAvqPNXHPvuqDNI0Q13_w.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*rMAvqPNXHPvuqDNI0Q13_w.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">I definitely looked this cool while negotiating.<\/figcaption><\/figure><p id=\"ae10\" class=\"graf graf--p graf-after--figure\">My initial counter-offers sounded something like this (I made up the numbers here and provide a few examples):<\/p><blockquote id=\"6f53\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">Bar, Inc.<\/strong>: We were thinking of offering a base salary of around $120,000 with a four-year RSU package of $150,000 and a signing bonus of $10,000. What do you think of that?<\/div><\/blockquote><blockquote id=\"194d\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">Me<\/strong>: How flexible is Bar, Inc. on equity? I\u2019m really looking to stay and grow at a company, and equity is something I really interested in wherever I go. I\u2019m comfortable with something in the neighborhood of $120,000 on base, but I\u2019d definitely like to sign an offer with a larger equity package.<\/div><\/blockquote><blockquote id=\"2948\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">FooCorp<\/strong>: We were thinking of offering a base salary of around $100,000 with a four-year RSU package of $150,000 and a signing bonus of $10,000. What do you think of that?<\/div><\/blockquote><blockquote id=\"df69\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">Me<\/strong>: I think the base is a little bit lower than what I\u2019d like based on my offer from Bar, Inc. for $120,000, and base salary is really important to me. As for the equity package, I was looking for something closer to $200,000. The signing bonus is around what I was looking for.<\/div><\/blockquote><p id=\"3361\" class=\"graf graf--p graf-after--blockquote\"><strong class=\"markup--strong markup--p-strong\">There\u2019s a lot of actionable information above for a recruiter without you giving any implicit commitment to any number. <\/strong>FooCorp knows they need to get to $120,000 to meet your other offer. Bar knows that equity is important to you, and equity is something that companies tend to move a lot more willingly. You now have two companies moving in tandem to get a better offer to you and neither of them is talking to each other. This is pretty much how you want this entire process to go.<\/p><p id=\"c0e3\" class=\"graf graf--p graf-after--p\">It makes sense to change an established range once another company ups the ante. This is how that discussion went for me:<\/p><blockquote id=\"55e5\" class=\"graf graf--blockquote graf-after--p\"><div><strong class=\"markup--strong markup--blockquote-strong\">Me<\/strong>: Hey, I finally heard back from Baz LLC. Their offer came in a lot higher than I was expecting. They offered $150,000 in base, $200,000 in stock, and $40,000 in signing. That being said I think the work FooCorp aligns more closely with my interests, so I\u2019d really like to make something work with you guys. Can we do what we can to get to those numbers? I\u2019d love to figure this out.<\/div><\/blockquote><blockquote id=\"c4f4\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">FooCorp<\/strong>: Wow, that\u2019s a good offer. Keep in mind FooCorp is {making the world a better place \/ has amazing potential \/ is better for your career} and I would hope you wouldn\u2019t be deciding based on a paycheck. I don\u2019t know if we can match that offer exactly, but I\u2019ll see what we can do.<\/div><\/blockquote><p id=\"36e2\" class=\"graf graf--p graf-after--blockquote\">I went through the above sample conversations so many times and they all went pretty similarly.<\/p><p id=\"1fc5\" class=\"graf graf--p graf-after--p\">They always sounded very doubtful about moving their numbers. They always said it was more than they would typically give up.<\/p><p id=\"b2f8\" class=\"graf graf--p graf-after--p\">They always explained why their company\u2019s mission and culture made up for the difference in comp.<\/p><p id=\"9aeb\" class=\"graf graf--p graf-after--p\">But almost always, they upped the package.<\/p><p id=\"c4d1\" class=\"graf graf--p graf-after--p graf--trailing\">Sometimes they\u2019ll match it. Sometimes they won\u2019t. Sometimes they\u2019ll come close enough that you\u2019ll pull the trigger because it\u2019s such an exciting opportunity. You should always give it a shot, though.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"fb5d\" class=\"graf graf--h3 graf--leading\">My Offers<\/h3><figure id=\"2a2d\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*9WmHqsTsOaynjq68ouj84g.png\" data-width=\"829\" data-height=\"264\" data-action=\"zoom\" data-action-value=\"1*9WmHqsTsOaynjq68ouj84g.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*9WmHqsTsOaynjq68ouj84g.png\" alt=\"image\" \/><\/div><\/figure><p id=\"9b51\" class=\"graf graf--p graf-after--figure\">I\u2019m including this because lots of people have pinged me about it. I don\u2019t know how much people can take away from it, but I\u2019ll put it out there anyway. I\u2019ll include what tips I can.<\/p><p id=\"f760\" class=\"graf graf--p graf-after--p\">I was able to get ballpark numbers from every company before I disclosed any numbers. Here they are, anonymized and randomized. The format is base salary+bonus \/ 4yr stock package \/ signing bonus.<\/p><blockquote id=\"3613\" class=\"graf graf--blockquote graf-after--p\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Alpha<\/strong><\/code>: 180k \/ 150k \/ 0k (218k\/yr)<\/div><\/blockquote><blockquote id=\"83f6\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: 180k \/ 200k \/ 30k (238k\/yr)<\/div><\/blockquote><blockquote id=\"a861\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Charlie<\/strong><\/code>: 140k \/ 180k \/ 30k (193k\/yr)<\/div><\/blockquote><blockquote id=\"5eef\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Delta<\/strong><\/code>: 160k \/ 220k \/ 50k (228k\/yr)<\/div><\/blockquote><blockquote id=\"efef\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Echo<\/strong><\/code>: 160k \/ 200k \/ 30k (218k\/yr)<\/div><\/blockquote><blockquote id=\"49c1\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Foxtrot<\/strong><\/code>: 165k \/ 210k \/ 30k (225k\/yr)<\/div><\/blockquote><p id=\"2dd7\" class=\"graf graf--p graf-after--blockquote\">I thought it was pretty interesting how closely most of the numbers aligned despite not being able to see each other\u2019s initial offers. That must be the going rate\u2026 or something.<\/p><p id=\"26c8\" class=\"graf graf--p graf-after--p\">Anyway, <strong class=\"markup--strong markup--p-strong\">once I had all of the offers, it was time for constant communication for me<\/strong>. I called my least-favorite offer and told them what the best offer was and asked them to close the gap as much as possible. Once those numbers moved, I talked to more companies. I usually had daily or every-other-day update calls with my recruiters, and they sounded like this:<\/p><blockquote id=\"259b\" class=\"graf graf--blockquote graf-after--p\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: Any updates from your end?<\/div><\/blockquote><blockquote id=\"14f4\" class=\"graf graf--blockquote graf-after--blockquote\"><div><strong class=\"markup--strong markup--blockquote-strong\">Me<\/strong>: Yeah, I was able to get some preliminary numbers from <code class=\"markup--code markup--blockquote-code\">Foxtrot<\/code>. They were thinking somewhere in the ballpark of $225k TC, but <code class=\"markup--code markup--blockquote-code\">Delta<\/code> says they will comfortably beat that. I wasn\u2019t blown away by the offer, but given <code class=\"markup--code markup--blockquote-code\">Delta<\/code>\u2019s position on it, that\u2019s probably the low end of what I\u2019ll end up signing.<\/div><\/blockquote><blockquote id=\"c60e\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: Great, thanks so much for sharing that. I\u2019ll talk to the team and see what I can do. We\u2019d love to have you!<\/div><\/blockquote><p id=\"27c2\" class=\"graf graf--p graf-after--blockquote\"><code class=\"markup--code markup--p-code\">Alpha<\/code> refused to negotiate and refused to move their deadline, so they were out of the running pretty quickly.<\/p><figure id=\"6123\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*G4F1_K7Rm_ycez7uZTBztg.png\" data-width=\"1172\" data-height=\"588\" data-action=\"zoom\" data-action-value=\"1*G4F1_K7Rm_ycez7uZTBztg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*G4F1_K7Rm_ycez7uZTBztg.png\" alt=\"image\" \/><\/div><\/figure><p id=\"1bef\" class=\"graf graf--p graf-after--figure\">Once everyone had a chance to move their numbers, it looked like this:<\/p><blockquote id=\"e813\" class=\"graf graf--blockquote graf-after--p\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: 180k \/ 220k \/ 75k (254k\/yr; +16k)<\/div><\/blockquote><blockquote id=\"f38c\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Charlie<\/strong><\/code>: 140k \/ 200k \/ 30k (198k\/yr; +5k)<\/div><\/blockquote><blockquote id=\"16de\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Delta<\/strong><\/code>: 160k \/ 220k \/ 50k (228k\/yr; no movement)<\/div><\/blockquote><blockquote id=\"117b\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Echo<\/strong><\/code>: 160k \/ 250k \/ 30k (230k\/yr; +12k)<\/div><\/blockquote><blockquote id=\"7fb6\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Foxtrot<\/strong><\/code>: 180k \/ 240k \/ 30k (248k\/yr; +18k)<\/div><\/blockquote><p id=\"040f\" class=\"graf graf--p graf-after--blockquote\">More rinse and repeat. My goal was to whittle down the offers once I had things moving in the direction I wanted. It was a logistical nightmare keeping up with six companies during this. It was also becoming clearer that there was a gap that some companies couldn\u2019t overcome.<\/p><p id=\"c925\" class=\"graf graf--p graf-after--p\">I also had a lot more time to think about company culture and perks\/benefits, so certain companies began to pull away from each other. <strong class=\"markup--strong markup--p-strong\">I didn\u2019t formally reject any offers until I signed my final offer<\/strong>, but mentally, I wrote off a few of the companies and stopped negotiating with them. I just asked them for a few more days to decide.<\/p><p id=\"3e32\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Recruiters constantly told me they couldn\u2019t up their packages anymore, but inevitably, they always did.<\/strong> I\u2019m sure a ceiling exists and I\u2019m sure it\u2019s rarer to get number past a certain point, but it was dangerously close to a lie to continuously tell me they couldn\u2019t go any higher (\u201cyou don\u2019t have the experience\u201d) and then watch them up their offer by 10%. Part of doing business, I guess. \ud83e\udd37\u200d\ufe0f<\/p><p id=\"bce9\" class=\"graf graf--p graf-after--p\">After another round of negotiation (and a few days), it was down to three:<\/p><blockquote id=\"0bc2\" class=\"graf graf--blockquote graf-after--p\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: 180k \/ 250k \/ 75k (261k\/yr; +23k\/yr from initial)<\/div><\/blockquote><blockquote id=\"1a4a\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Delta<\/strong><\/code>: 170k \/ 250k \/ 50k (245k\/yr; +17k\/yr from initial)<\/div><\/blockquote><blockquote id=\"9a8d\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Foxtrot<\/strong><\/code>: 180k \/ 250k \/ 50k (248k\/yr; +18k\/yr from initial)<\/div><\/blockquote><p id=\"4b3d\" class=\"graf graf--p graf-after--blockquote\">At this point, it had been something like a week and a half of negotiation, and this after months of interviews\/prep. I was <em class=\"markup--em markup--p-em\">so <\/em>ready to be done. I had decided based on product, culture, pay, perks, benefits, reputation that it was between <code class=\"markup--code markup--p-code\">Bravo<\/code> and <code class=\"markup--code markup--p-code\">Delta<\/code>.<\/p><p id=\"eaab\" class=\"graf graf--p graf-after--p\"><code class=\"markup--code markup--p-code\">Delta<\/code> wanted me to give a final number and commit to it. They were <em class=\"markup--em markup--p-em\">extremely <\/em>insistent that I provide one.<\/p><p id=\"bc63\" class=\"graf graf--p graf-after--p\">I told them to give me a few hours and called my recruiter from <code class=\"markup--code markup--p-code\">Bravo<\/code>.<\/p><p id=\"d142\" class=\"graf graf--p graf-after--p\">I explained that <code class=\"markup--code markup--p-code\">Bravo<\/code> had a lot that I respected more than <code class=\"markup--code markup--p-code\">Delta<\/code>, but that <code class=\"markup--code markup--p-code\">Delta<\/code> wanted me to verbally commit to a number to move forward.<\/p><p id=\"bb73\" class=\"graf graf--p graf-after--p\">I was going to work at <code class=\"markup--code markup--p-code\">Bravo<\/code> or <code class=\"markup--code markup--p-code\">Delta<\/code>. They were offering similar packages, but I didn\u2019t know how high I could push <code class=\"markup--code markup--p-code\">Delta<\/code> on the last counter. I also wasn\u2019t comfortable playing that game.<\/p><figure id=\"d2a9\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*yQqAlTR2MuSZwjFPWSxI6A.png\" data-width=\"824\" data-height=\"468\" data-action=\"zoom\" data-action-value=\"1*yQqAlTR2MuSZwjFPWSxI6A.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*yQqAlTR2MuSZwjFPWSxI6A.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Break in case of emergency<\/figcaption><\/figure><p id=\"7593\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Everyone gets one last ask. <\/strong>You\u2019ve been building up to a point where you get to launch your most powerful salvo\u200a\u2014\u200ause it wisely.<\/p><p id=\"3f1e\" class=\"graf graf--p graf-after--p\">For me, this was <em class=\"markup--em markup--p-em\">that<\/em> moment. It was time.<\/p><p id=\"8e8e\" class=\"graf graf--p graf-after--p\">I asked <code class=\"markup--code markup--p-code\">Bravo<\/code> to up their package by a meager $5k yearly in stock (a <em class=\"markup--em markup--p-em\">very<\/em> small ask in retrospect\u2026), something companies are far more flexible on. I said if they could do that, I would sign and be confident I didn\u2019t leave too much on the table. He said he\u2019d get back to me and that he was confident we could work something out.<\/p><h3 id=\"36c0\" class=\"graf graf--h3 graf-after--p\">My Final\u00a0Offer<\/h3><p id=\"1aa2\" class=\"graf graf--p graf-after--h3\"><code class=\"markup--code markup--p-code\">Bravo<\/code>\u2019s recruiter emailed back and said he worked out a new package for me and wanted to run it by me over the phone. I already knew at this point that if he met my ask (180k \/ 270k \/ 70k, TC 265k) that I would sign right then and there. I\u2019d thought long and hard about this last little dance.<\/p><blockquote id=\"49c2\" class=\"graf graf--blockquote graf-after--p\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: I heard back and got approval for a new offer and I wanted to see what you thought. We\u2019d like to offer you a {salary + annual bonus} of $180,000. In addition to that, we would like to offer $400,000 in RSUs that vest over four years. Lastly, I was able to get approval for a $90,000 {signing bonus + relocation expenses} package. <strong class=\"markup--strong markup--blockquote-strong\">That would put your four-year compensation package at around $1,210,000, or an annual compensation package of around $300,000<\/strong>. What do you think?<\/div><\/blockquote><figure id=\"632f\" class=\"graf graf--figure graf-after--blockquote\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*mMcYHod_XDNGwbM5TkmvdA.png\" data-width=\"1174\" data-height=\"338\" data-action=\"zoom\" data-action-value=\"1*mMcYHod_XDNGwbM5TkmvdA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*mMcYHod_XDNGwbM5TkmvdA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Uh, what do I think? What do you think I\u00a0think?<\/figcaption><\/figure><blockquote id=\"702c\" class=\"graf graf--blockquote graf-after--figure\"><div><strong class=\"markup--strong markup--blockquote-strong\"><em class=\"markup--em markup--blockquote-em\">Me: <\/em><\/strong>Yeah, I\u2019ll sign that.<\/div><\/blockquote><blockquote id=\"05c3\" class=\"graf graf--blockquote graf-after--blockquote\"><div><code class=\"markup--code markup--blockquote-code\"><strong class=\"markup--strong markup--blockquote-strong\">Bravo<\/strong><\/code>: Really? Awesome!<\/div><\/blockquote><p id=\"9b3b\" class=\"graf graf--p graf-after--blockquote graf--trailing\">And that was that. No, there\u2019s no embellishment here. <strong class=\"markup--strong markup--p-strong\">My final ask was for $265k\/yr and I was offered around $300k\/yr.<\/strong> I don\u2019t know if that was due to the goodwill I had built up with the recruiter or if they were concerned I\u2019d defect. Whatever it was, I\u2019m fine with it.\u00a0:)<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"260b\" class=\"graf graf--h3 graf--leading\">The Takeaways<\/h3><p id=\"e2cd\" class=\"graf graf--p graf-after--h3\">I\u2019m going to summarize a lot of the above in a more digestible way:<\/p><ul class=\"postList\"><li id=\"cb0c\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">\ud83e\udd11 Get multiple offers.<\/strong> Obviously this is <em class=\"markup--em markup--li-em\">way<\/em> easier said than done, but even if it adds a lot of stress to your life, it certainly pays off. With interviews being quite a bit of a crapshoot, it\u2019s tough. If you want to get a new job, though, don\u2019t put your eggs in one basket and do what you can to try to get at least two offers.<\/li><li id=\"f9c9\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83d\ude4a Don\u2019t lie. <\/strong>I wasn\u2019t <em class=\"markup--em markup--li-em\">personally <\/em>asked to provide evidence of other offers, but my friends definitely have been. It\u2019s not worth it to lie, especially when your candidate profile is saved at various companies. Good luck getting into a company that catches you.<\/li><li id=\"eb17\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83e\udd17 Be kind and respectful.<\/strong> Recruiters are people, too. They\u2019re also the people that act as the path for you getting into your dream company. You don\u2019t need to schmooze them, but be friendly. Help them help you! You may have competing philosophies, but both of you want the same thing in the end: a signed offer from their company.<\/li><li id=\"8fb2\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83d\udde3 Be (selectively) communicative.<\/strong> Make it very clear to everyone what your timelines are and if you\u2019re waiting on other companies. Keep people in the loop. We hate getting ghosted by recruiters and I\u2019m sure recruiters hate being ghosted by us. Let them know if you\u2019ve signed another offer or if you\u2019re no longer considering theirs. Let them know if other companies have sent you offers and share the names if you think it\u2019ll help your case.<\/li><li id=\"15d4\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83d\udc4e Don\u2019t give a minimum (unless it makes sense)<\/strong>. You don\u2019t want to give a lower number than a company <em class=\"markup--em markup--li-em\">would<\/em> have offered you, but you also don\u2019t want to waste anyone\u2019s time. If you <em class=\"markup--em markup--li-em\">know<\/em> a company won\u2019t be competitive at first, let them know what the other company\u2019s range is. If they\u2019re out, they\u2019re out. Your time is better spent talking to companies you\u2019d actually go to.<\/li><\/ul><figure id=\"e458\" class=\"graf graf--figure graf-after--li\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*KUTgrNa8tDcqDf4QefTSHw.png\" data-width=\"592\" data-height=\"483\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*KUTgrNa8tDcqDf4QefTSHw.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">I imagine this is what whiteboard interviews looked like in the\u00a01950s<\/figcaption><\/figure><ul class=\"postList\"><li id=\"a8cb\" class=\"graf graf--li graf-after--figure\"><strong class=\"markup--strong markup--li-strong\">\ud83c\udf20 Shoot for the sky. <\/strong>Transparently, my goal was to get a total comp of around $180k. That was the number where I\u2019d quit my current job. I ended up getting over 50% more than that. I didn\u2019t think it was <em class=\"markup--em markup--li-em\">possible<\/em> for someone with two years of experience to do so. Don\u2019t set a limit in your head of what you can think you can get or it may discourage negotiating.<\/li><li id=\"dc18\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83d\udeab Don\u2019t reject an offer until you\u2019ve signed another one. <\/strong>You never know what could happen. You don\u2019t want to leave someone hanging on your decision when <em class=\"markup--em markup--li-em\">there is no way on earth you\u2019d sign their offer.<\/em> But even if an offer isn\u2019t competitive, you\u2019d still sign it if the other companies lost their headcount. Keep your options open. You earned it.<\/li><li id=\"aa3b\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83e\udd14 Think a lot and whittle down your offers.<\/strong> Over the course of your negotiations you\u2019ll converge on the one or two offers you think are \u201cbest\u201d. As that happens, focus your time on the ones that make sense, and don\u2019t waste brain cycles on the ones that don\u2019t.<\/li><li id=\"dddc\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83e\udd26\ud83c\udffb\u200d\u2642\ufe0f Bonus: don\u2019t tell teams you\u2019re disinterested in them. <\/strong>I\u2019ve seen people \u201cnegotiate\u201d by asking for a larger packages by saying they\u2019re less interested in the company. This sounds like a recipe for disaster\u200a\u2014\u200awhy would you tell a company you don\u2019t want to work for them? Find a new slant.<\/li><li id=\"c8f2\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">\ud83c\udf89 Everything above worked. <\/strong>The final offer I signed was ~$300,000 in total yearly compensation (over four years, including a variable annual bonus, and amortizing the signing\/relo package). I negotiated it up from ~$225,000*, an increase of 33%.<\/li><\/ul><h3 id=\"e05e\" class=\"graf graf--h3 graf-after--li\">Wrap-Up<\/h3><p id=\"ad1a\" class=\"graf graf--p graf-after--h3\">After my previous post, I primarily saw three questions:<\/p><ul class=\"postList\"><li id=\"6073\" class=\"graf graf--li graf-after--p\">How did I negotiate?<\/li><li id=\"0981\" class=\"graf graf--li graf-after--li\">How much was I offered?<\/li><li id=\"0857\" class=\"graf graf--li graf-after--li\">Which offer did I take?<\/li><\/ul><p id=\"1f70\" class=\"graf graf--p graf-after--li\">I answered two of those three here. Maybe you can piece together the rest. I\u2019d rather provide value than advertise for my employer. I don\u2019t want the focus here to be on the decision I made for <em class=\"markup--em markup--p-em\">my<\/em> life. I\u2019d rather help everyone <em class=\"markup--em markup--p-em\">else<\/em> get to make the same decision.<\/p><p id=\"aa02\" class=\"graf graf--p graf-after--p\">There\u2019s also much more in-depth, less-anecdotal pieces I\u2019d recommend reading. There\u2019s <a href=\"https:\/\/www.kalzumeus.com\/2012\/01\/23\/salary-negotiation\/\" data-href=\"https:\/\/www.kalzumeus.com\/2012\/01\/23\/salary-negotiation\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">this legendary one<\/a> from <a href=\"https:\/\/medium.com\/@patio11\" data-href=\"https:\/\/medium.com\/@patio11\" data-anchor-type=\"2\" data-user-id=\"22a4f31382e5\" data-action-value=\"22a4f31382e5\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Patrick McKenzie<\/a> and <a href=\"https:\/\/medium.freecodecamp.org\/ten-rules-for-negotiating-a-job-offer-ee17cccbdab6\" data-href=\"https:\/\/medium.freecodecamp.org\/ten-rules-for-negotiating-a-job-offer-ee17cccbdab6\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">another<\/a> from <a href=\"https:\/\/medium.com\/@hosseeb\" data-href=\"https:\/\/medium.com\/@hosseeb\" data-anchor-type=\"2\" data-user-id=\"8bc4e5f8b505\" data-action-value=\"8bc4e5f8b505\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Haseeb Qureshi<\/a>. What you see here is more of a collection of my personal thoughts (and a response to all of the questions I got regarding numbers).<\/p><p id=\"daea\" class=\"graf graf--p graf-after--p\"><a href=\"https:\/\/www.twitter.com\/bayareabell\" data-href=\"https:\/\/www.twitter.com\/bayareabell\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Please follow (and chat with!) me on Twitter<\/strong><\/a> if you want to provide feedback or if you have any questions. I\u2019d love to hear if there are other topics you\u2019d like to read about. I\u2019ve been getting a ton of DMs on Twitter and Discord and I\u2019ve tried to help out where possible. I vastly prefer responding on Twitter, though, so if you want to chat, try there first.<\/p><p id=\"09e4\" class=\"graf graf--p graf-after--p\">Happy negotiating. \ud83d\ude0a<\/p><ul class=\"postList\"><li id=\"8310\" class=\"graf graf--li graf-after--p graf--trailing\"><em class=\"markup--em markup--li-em\">the initial offer didn\u2019t talk about a relocation package; my guess is I would have received it anyway, so the starting value was likely higher<\/em><\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><figure id=\"6751\" class=\"graf graf--figure graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/usejournal.com\/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_image\" data-href=\"https:\/\/usejournal.com\/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_image\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"noopener\" target=\"_blank\"><img class=\"graf-image\" data-image-id=\"1*f2IVAl0TbsfES9cFGYr40g.png\" data-width=\"2000\" data-height=\"400\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*f2IVAl0TbsfES9cFGYr40g.png\" alt=\"image\" \/><\/a><\/div><\/figure><p id=\"23b5\" class=\"graf graf--p graf-after--figure graf--trailing\">\ud83d\udcdd Read this story later in <a href=\"https:\/\/usejournal.com\/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_read_later_text\" data-href=\"https:\/\/usejournal.com\/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_read_later_text\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Journal<\/a>. Wake up every Sunday morning to the week\u2019s most noteworthy Tech stories, opinions, and news waiting in your inbox: <a href=\"https:\/\/usejournal.com\/newsletter\/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_text\" data-href=\"https:\/\/usejournal.com\/newsletter\/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_text\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Get the noteworthy newsletter &gt;<\/a><\/p><\/div><\/div><\/section>","created_at":"2019-02-02T02:20:03+0000","updated_at":"2019-02-02T02:20:03+0000","published_at":"2019-01-28T13:01:01+0000","published_by":["Bay Area Belletrist"],"annotations":[],"mimetype":"text\/html","reading_time":22,"domain_name":"blog.usejournal.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*TjFEp_SQMG3xh41Ly9cKig.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["graphql","python","flask"],"is_public":false,"id":806,"title":"GraphQL server up and running with 50 lines of python","url":"https:\/\/medium.com\/@fasterpancakes\/graphql-server-up-and-running-with-50-lines-of-python-85e8ada9f637","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"de13\" class=\"graf graf--figure graf--layoutFillWidth graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*qpUWqfJQSuTO-Bc4DuHBcw.jpeg\" data-width=\"2500\" data-height=\"800\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2600\/1*qpUWqfJQSuTO-Bc4DuHBcw.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"5518\" class=\"graf graf--h3 graf-after--figure graf--trailing graf--title\">GraphQL server up and running with 50 lines of\u00a0python<\/h1><div class=\"uiScale uiScale-ui--regular uiScale-caption--regular u-flexCenter u-marginVertical24 u-fontSize15 js-postMetaLockup\"><div class=\"u-flex0\"><a class=\"link u-baseColor--link avatar\" href=\"https:\/\/medium.com\/@fasterpancakes?source=post_header_lockup\" data-action=\"show-user-card\" data-action-source=\"post_header_lockup\" data-action-value=\"628ee58ac2ef\" data-action-type=\"hover\" data-user-id=\"628ee58ac2ef\" dir=\"auto\"><img src=\"https:\/\/cdn-images-1.medium.com\/fit\/c\/100\/100\/0*ytpnaYvimOyNhdb5.jpg\" class=\"avatar-image u-size50x50\" alt=\"Go to the profile of Cameron Moss\" \/><\/a><\/div><div class=\"u-flex1 u-paddingLeft15 u-overflowHidden\"><div class=\"u-paddingBottom3\"><a class=\"ds-link ds-link--styleSubtle ui-captionStrong u-inlineBlock link link--darken link--darker\" href=\"https:\/\/medium.com\/@fasterpancakes\" data-action=\"show-user-card\" data-action-value=\"628ee58ac2ef\" data-action-type=\"hover\" data-user-id=\"628ee58ac2ef\" dir=\"auto\">Cameron Moss<\/a><button class=\"button button--smallest u-noUserSelect button--withChrome u-baseColor--buttonNormal button--withHover button--unblock js-unblockButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-block-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/@fasterpancakes\/graphql-server-up-and-running-with-50-lines-of-python-85e8ada9f637\" data-action-source=\"post_header_lockup\">BlockedUnblock<\/button><button class=\"button button--primary button--smallest button--dark u-noUserSelect button--withChrome u-accentColor--buttonDark button--follow js-followButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-subscribe-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/_\/subscribe\/user\/628ee58ac2ef\" data-action-source=\"post_header_lockup-628ee58ac2ef-------------------------follow_byline\">FollowFollowing<\/button><\/div><div class=\"ui-caption postMetaInline js-testPostMetaInlineSupplemental\"><time datetime=\"2017-08-26T01:39:55.955Z\">Aug 25, 2017<\/time><\/div><\/div><\/div><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><blockquote id=\"3be3\" class=\"graf graf--blockquote graf--leading\"><div>Prefix: I actually wrote this before choosing to use the <a href=\"http:\/\/absinthe-graphql.org\/\" data-href=\"http:\/\/absinthe-graphql.org\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Elixir project<\/a> instead, but decided to release this post anyway because it was actually the fastest to set up.<\/div><\/blockquote><p id=\"8e64\" class=\"graf graf--p graf-after--blockquote\">Node.js is far from the only server implementation out there for graphQL. I recently looked a C# and python implementations to see how quickly I could attach a graphQL server to our database here at <a href=\"http:\/\/hyr.work\" data-href=\"http:\/\/hyr.work\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">hyr<\/a> and was amazed at the great options I found.<\/p><p id=\"4f76\" class=\"graf graf--p graf-after--p\">I love well-documented projects, and found the\u00a0<a href=\"https:\/\/github.com\/graphql-dotnet\/graphql-dotnet\" data-href=\"https:\/\/github.com\/graphql-dotnet\/graphql-dotnet\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">.NET<\/a> project a bit lacking in terms of setup instructions and good examples. The python project, on the other hand, consisted of a full documentation site and an decent example to get you started. Furthermore, there\u2019s something beautiful about the simple exchange of resources between python files.<\/p><p id=\"308e\" class=\"graf graf--p graf-after--p\">I\u2019m going to show how to get the server going on top of an existing SQL database, which is an awesome. I\u2019ll be using <a href=\"http:\/\/graphene-python.org\/\" data-href=\"http:\/\/graphene-python.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Graphene<\/a>, which is a graphQL library that works with my things like Django. But for simplicity I\u2019ll be using <a href=\"http:\/\/www.sqlalchemy.org\/\" data-href=\"http:\/\/www.sqlalchemy.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">SQLAlchemy <\/a>as the ORM. It\u2019s very similar to Django. For the web server I\u2019ll be using the trusty <a href=\"http:\/\/flask.pocoo.org\/docs\/0.12\/\" data-href=\"http:\/\/flask.pocoo.org\/docs\/0.12\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Flask<\/a>.<\/p><h3 id=\"d5b3\" class=\"graf graf--h3 graf-after--p\">Moving from REST: reflect the\u00a0database<\/h3><p id=\"06c2\" class=\"graf graf--p graf-after--h3\">If you already have an existing database, perhaps with a RESTful Node.js API, it can be a pain to construct new models in a different language. That\u2019s where database reflection comes in. It\u2019s going to do this work for you.<\/p><p id=\"91eb\" class=\"graf graf--p graf-after--p\">If you\u2019re using postgres, you should know about the <a href=\"https:\/\/github.com\/calebmer\/postgraphql\" data-href=\"https:\/\/github.com\/calebmer\/postgraphql\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">postgraphQL<\/a> (Node.js) library. It similarly reflects (collects information from) your database and constructs not only the ORM(models), but also the graphQL web interface(schema). It\u2019s a completely automated add-on resulting in highly opinionated structure for both queries and mutations. In this example, using graphene and SQLAlchemy, you\u2019ll find almost the same ease of setup, but you\u2019ll have infinitely more room for customization!<\/p><p id=\"cc4c\" class=\"graf graf--p graf-after--p\">The file structure for the graphene server app (modeled after <a href=\"https:\/\/github.com\/graphql-python\/graphene-sqlalchemy\/tree\/master\/examples\/flask_sqlalchemy\" data-href=\"https:\/\/github.com\/graphql-python\/graphene-sqlalchemy\/tree\/master\/examples\/flask_sqlalchemy\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">this example<\/a>) looks like this:<\/p><pre id=\"b1df\" class=\"graf graf--pre graf-after--p\">graphene_example\/<br \/>    |-app.py<br \/>    |-database.py<br \/>    |-models.py<br \/>    |-schema.py<\/pre><p id=\"393a\" class=\"graf graf--p graf-after--pre\">Two of these files are simply configurations, <code class=\"markup--code markup--p-code\">database.py<\/code> and <code class=\"markup--code markup--p-code\">app.py<\/code> which you can think of as establishing the back and front-facing connections respectively. Let see how that works.<\/p><figure id=\"1337\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*36aUCxi8DraT88xxqWh86A.jpeg\" data-width=\"960\" data-height=\"528\" data-action=\"zoom\" data-action-value=\"1*36aUCxi8DraT88xxqWh86A.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*36aUCxi8DraT88xxqWh86A.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">From db to\u00a0client<\/figcaption><\/div><\/figure><p id=\"e4ae\" class=\"graf graf--p graf-after--figure\">The <code class=\"markup--code markup--p-code\">database.py<\/code> simply creates a connection to the database and a session object which will be used for queries. The session object is then attached to the <code class=\"markup--code markup--p-code\">Base<\/code> class that will be inherited each of the models in <code class=\"markup--code markup--p-code\">models.py<\/code>.<\/p><figure id=\"b4b5\" class=\"graf graf--figure graf--iframe graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe width=\"700\" height=\"250\" src=\"https:\/\/medium.com\/media\/338e9dd1f4c87ed78b03f133fe5ce2e9?postId=85e8ada9f637\" data-media-id=\"338e9dd1f4c87ed78b03f133fe5ce2e9\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><figcaption class=\"imageCaption\">database.py<\/figcaption><\/div><\/figure><p id=\"bd7c\" class=\"graf graf--p graf-after--figure\">Two things here. Depending on your setup the connection string may vary a bit. But for the most part, it will look something like this<\/p><pre id=\"c580\" class=\"graf graf--pre graf-after--p\">'mysql:\/\/scott:tiger@localhost\/foo'<\/pre><p id=\"9ea4\" class=\"graf graf--p graf-after--pre\">or for postgres:<\/p><pre id=\"0339\" class=\"graf graf--pre graf-after--p\">'postgresql:\/\/scott:tiger@localhost\/mydatabase\u2019<\/pre><p id=\"a1b6\" class=\"graf graf--p graf-after--pre\">Second thing is <code class=\"markup--code markup--p-code\">DeferredRefection.<\/code> What this does is it allows field definitions from your database to be inferred for each class after all of the model classes are declared. You don\u2019t have to list out each of the fields for a given db table. Let\u2019s say we already have a database with two existing tables. We connect the model to the table with <code class=\"markup--code markup--p-code\">__tablename__<\/code>. Then we only have to define primary and foreign keys and their relationships.<\/p><figure id=\"b484\" class=\"graf graf--figure graf--iframe graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe width=\"700\" height=\"250\" src=\"https:\/\/medium.com\/media\/3a4d4e8f2d13c990646b6e49b90dd85d?postId=85e8ada9f637\" data-media-id=\"3a4d4e8f2d13c990646b6e49b90dd85d\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><figcaption class=\"imageCaption\">models.py<\/figcaption><\/div><\/figure><p id=\"0241\" class=\"graf graf--p graf-after--figure\">Note I did not define a <code class=\"markup--code markup--p-code\">name<\/code> or <code class=\"markup--code markup--p-code\">age<\/code> field for the <code class=\"markup--code markup--p-code\">PersonModel. <\/code>The db reflection takes care of this unless your db field doesn\u2019t map to a SQLAlchemy type. For example, if you are using an MSSQL <code class=\"markup--code markup--p-code\">money<\/code> type you will need to add a <code class=\"markup--code markup--p-code\">Column<\/code> for it with an appropriate type.<\/p><pre id=\"2cde\" class=\"graf graf--pre graf-after--p\">rate = Column(Float)<\/pre><p id=\"f2c7\" class=\"graf graf--p graf-after--pre\">That\u2019s it! The majority of the work here is defining relationships.<\/p><h3 id=\"77f9\" class=\"graf graf--h3 graf-after--p\">Creating a GraphQL\u00a0endpoint<\/h3><p id=\"5d81\" class=\"graf graf--p graf-after--h3\">On the other side of things is the actual graphQL part. Let\u2019s start by covering the actual flask web server.<\/p><figure id=\"081e\" class=\"graf graf--figure graf--iframe graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe width=\"700\" height=\"250\" src=\"https:\/\/medium.com\/media\/83d3812229b7fd86dd6069832a0a501b?postId=85e8ada9f637\" data-media-id=\"83d3812229b7fd86dd6069832a0a501b\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><figcaption class=\"imageCaption\">app.py<\/figcaption><\/div><\/figure><p id=\"8f2c\" class=\"graf graf--p graf-after--figure\">Most of this is fairly boilerplate. You create a single route that is handled by a graphQL view function. That\u2019s all graphQL is\u200a\u2014\u200aa single endpoint! Not much else to say here. Let\u2019s look at the <code class=\"markup--code markup--p-code\">schema.py<\/code><\/p><figure id=\"e45e\" class=\"graf graf--figure graf--iframe graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe width=\"700\" height=\"250\" src=\"https:\/\/medium.com\/media\/47fbf24406b54a20d91112819b7841e3?postId=85e8ada9f637\" data-media-id=\"47fbf24406b54a20d91112819b7841e3\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><figcaption class=\"imageCaption\">schema.py<\/figcaption><\/div><\/figure><p id=\"2673\" class=\"graf graf--p graf-after--figure\">For most python developers this is perhaps the only new thing. Anyone not familiar with graphQL concepts, I highly recommend reading the <a href=\"http:\/\/graphql.org\/\" data-href=\"http:\/\/graphql.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">graphQL docs<\/a> and watching <a href=\"https:\/\/www.youtube.com\/watch?v=oPSuvaYmXBY\" data-href=\"https:\/\/www.youtube.com\/watch?v=oPSuvaYmXBY\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Greg Hurrell\u2019s deep dive video<\/a>.<\/p><p id=\"46fb\" class=\"graf graf--p graf-after--p\">The first half of <code class=\"markup--code markup--p-code\">schema.py<\/code> merely imports the models and creates new classes that are usable by graphene. It establishes the inheritance from the <code class=\"markup--code markup--p-code\">SQLAlchemyObjectType<\/code> and a node interface which is used internally by graphene (and Relay if you choose to use it) to establish a global id reference for any instances of that type.<\/p><p id=\"0589\" class=\"graf graf--p graf-after--p\">At it\u2019s most simple implementation (read-only graphQL) the schema object needs one thing: a Query. Your Query defines the entry point, or the root (or roots) of all data retrieved by your app. Think of this as the perspective or the viewer of your app\u2019s data (a common Relay pattern actually uses \u201cViewer\u201d as the root). In this case the <code class=\"markup--code markup--p-code\">Person<\/code> is the viewer so I\u2019ve defined the <code class=\"markup--code markup--p-code\">person<\/code> as the root query name.<\/p><p id=\"4008\" class=\"graf graf--p graf-after--p\">All root queries besides <code class=\"markup--code markup--p-code\">node<\/code> must have an accompanying resolve function which allows access to the parameters given to the query.<\/p><blockquote id=\"db56\" class=\"graf graf--blockquote graf-after--p\"><div>An aside here about Relay (and Apollo). Adding the <code class=\"markup--code markup--blockquote-code\">node<\/code> root field is an requirement for Relay. It enables the Relay client store to query the graphQL endpoint and ask for a node by a single globally unique Relay <code class=\"markup--code markup--blockquote-code\">id<\/code>\u00a0. Your calls will use the <code class=\"markup--code markup--blockquote-code\">person<\/code> root while behind the scene, Relay will use calls to <code class=\"markup--code markup--blockquote-code\">node<\/code> to refetch data only when it needs to\u200a\u2014\u200alike to update a single data point in a node. It\u2019s a priceless optimization.<\/div><\/blockquote><p id=\"3fcc\" class=\"graf graf--p graf-after--blockquote\">Let\u2019s look at an example query.<\/p><pre id=\"411a\" class=\"graf graf--pre graf-after--p\">query{<br \/>  person(uuid:1){<br \/>    name<br \/>    age<br \/>  }<br \/>}  <\/pre><p id=\"5642\" class=\"graf graf--p graf-after--pre\">Notice I provide a parameter to the root query of <code class=\"markup--code markup--p-code\">person<\/code>. That parameter <code class=\"markup--code markup--p-code\">uuid<\/code> must be defined in the Query and implemented in <code class=\"markup--code markup--p-code\">resolve_person<\/code> to return the person object using a call to the database or in our case, an SQLAlchemy query.<\/p><pre id=\"0a91\" class=\"graf graf--pre graf-after--p\">query{<br \/>  person(uuid:1){<br \/>    name<br \/>    age<br \/>    Articles(first:5){<br \/>      edges{<br \/>        node{<br \/>          publishedOn<br \/>          text<br \/>        }<br \/>      }<br \/>    }<br \/>  }<br \/>}<\/pre><p id=\"b5a8\" class=\"graf graf--p graf-after--pre\">Here is a query that includes a relationship or, in graphQL terms, a connection. The <code class=\"markup--code markup--p-code\">edges<\/code> field will return an array of nodes, each of which can be referenced in a client app like so:<\/p><p id=\"ec67\" class=\"graf graf--p graf-after--p\"><code class=\"markup--code markup--p-code\">res.data.person.Articles.edges.map(a =&gt; &lt;Text&gt;{a.node.text}&lt;\/Text&gt;)<\/code><\/p><p id=\"a054\" class=\"graf graf--p graf-after--p\">This somewhat naively displays the text of the first five articles associated with person whose <code class=\"markup--code markup--p-code\">uuid<\/code> is 1 in React Native given the response data provided by this call as <code class=\"markup--code markup--p-code\">res<\/code>. It is my convention to use capitalized words for connections. Other examples will use <code class=\"markup--code markup--p-code\">ArticleConnection<\/code>.<\/p><h3 id=\"f66c\" class=\"graf graf--h3 graf-after--p\">Mutations<\/h3><p id=\"d065\" class=\"graf graf--p graf-after--h3\">The most difficult and controversial part to any client-side graphQL implementation is the mutation configuration\u200a\u2014\u200athe write. It\u2019s why the Relay community is anxiously awaiting the release of Relay2. On the server side, things aren\u2019t that bad. I\u2019ll leave it as an exercise for the reader to implement a CREATE mutation. Here is slightly more complex UPDATE:<\/p><figure id=\"0c61\" class=\"graf graf--figure graf--iframe graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe width=\"700\" height=\"250\" src=\"https:\/\/medium.com\/media\/6efe8d116cc5b1a50a153ae0ae0d3ed6?postId=85e8ada9f637\" data-media-id=\"6efe8d116cc5b1a50a153ae0ae0d3ed6\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><\/div><\/figure><p id=\"d033\" class=\"graf graf--p graf-after--figure\">And it is called like so:<\/p><pre id=\"f53a\" class=\"graf graf--pre graf-after--p\">mutation{<br \/>  updatePersonName(uuid:1, name:\"Bob\"){<br \/>    person{<br \/>      name<br \/>    }<br \/>  }<br \/>}<\/pre><p id=\"c742\" class=\"graf graf--p graf-after--pre\">You\u2019ll notice that the mutation has a query written into it. That\u2019s the response. Every mutation is a PUT followed by a GET. That should be enough to get you setup with a fairly simple graphQL backend. Please excuse my bad Python.<\/p><p id=\"6d56\" class=\"graf graf--p graf-after--p graf--trailing\">One last note. SQL calls are not batched so all calls are n+1. Leave a comment please if you know of a good batching solution.<\/p><\/div><\/div><\/section>","created_at":"2019-01-31T14:05:37+0000","updated_at":"2019-01-31T14:05:41+0000","published_at":"2017-08-26T01:39:55+0000","published_by":["Cameron Moss"],"annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*qpUWqfJQSuTO-Bc4DuHBcw.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":805,"title":"The Mastermind","url":"https:\/\/magazine.atavist.com\/the-mastermind","content":"<div class=\"notification-box-content\" data-content_type=\"paywall-exhausted\">You\u2019ve read the 3 free articles available to you this month.<h2>To read more, join <em>The Atavist Magazine<\/em>.<\/h2><p>You\u2019ve read the 3 complimentary stories available to you. Get unlimited access to <em>The Atavist Magazine<\/em> and enjoy any story from our award-winning catalog, along with a new title every month.<\/p><a href=\"https:\/\/magazine.atavist.com\/subscribe\" class=\"paywall-purchase-button-custom\">Join<\/a><h3><a class=\"atavist-login\" href=\"https:\/\/magazine.atavist.com\/login\">Already have an account?<\/a><\/h3><\/div><div class=\"notification-box-content\" data-content_type=\"subscription-promo\" data-subscription_promo_behavior=\"ten_seconds\"><h2>For the love of stories...<\/h2><img src=\"https:\/\/magazine.atavist.com\/data\/files\/organization\/1\/image\/raw\/covers-1499986067-4.png\" alt=\"image\" \/><p>Support great writers and their work by becoming an Atavist member today. <\/p><p><a href=\"https:\/\/magazine.atavist.com\/subscribe?utm_source=paywall&amp;utm_medium=subscribe&amp;utm_campaign=5818f9e597e74\" class=\"paywall-purchase-button\">Join<\/a>\n  \t\t\t\t\t\t\t  \t\t\t\t\t\t\t<\/p><\/div>","created_at":"2019-01-31T13:13:32+0000","updated_at":"2019-01-31T13:13:32+0000","published_at":"2016-04-28T16:55:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"magazine.atavist.com","preview_picture":"https:\/\/magazine.atavist.com\/data\/files\/organization\/1\/image\/derivative\/scale~1200x1200~ap508237068-1478290418-26.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":804,"title":"How to teach programming to your loved ones","url":"https:\/\/media.ccc.de\/v\/35c3-9800-how_to_teach_programming_to_your_loved_ones","content":"<p class=\"persons\">\n<a href=\"https:\/\/media.ccc.de\/search?q=Mike+Sperber\">Mike Sperber<\/a>\n<\/p>Playlists:<a href=\"https:\/\/media.ccc.de\/v\/35c3-9800-how_to_teach_programming_to_your_loved_ones\/playlist\">'35c3' videos starting here<\/a>\/<a data-method=\"get\" href=\"https:\/\/media.ccc.de\/v\/35c3-9800-how_to_teach_programming_to_your_loved_ones\/audio\">audio<\/a>\/<a href=\"https:\/\/media.ccc.de\/v\/35c3-9800-how_to_teach_programming_to_your_loved_ones\/related\">related events<\/a><p>Teaching beginners how to program is often <i>hard<\/i>.  We love building programs, and seeing our loved ones struggle with this is painful.  Showing them how to copy-paste a few example programs and change a few parameters is easy, but bridging from there to building substantial programs is a different game entirely.  This talk is about how to teach programming successfully, through comprehensible <i>design recipes<\/i>, which anyone can follow, using languages and tools designed for beginners.  This approach is probably different from how you learned how to program, or how you're used to teaching.  It is more effective, however, as it teaches more material successfully to a broader spectrum of people.  It is also more enjoyable.<\/p><p>The talk is based on many years of research by the <a href=\"https:\/\/programbydesign.org\/\">Program by Design<\/a>, <a href=\"http:\/\/www.deinprogramm.de\">DeinProgramm<\/a>, and <a href=\"http:\/\/www.bootstrapworld.org\/\">Bootstrap<\/a> educational projects, as well as over 30 years of personal teaching experience in school, university and industrial contexts.  A word of warning: The resulting approach is radically different from most teaching approaches used in universities and schools.  In particular, it avoids teaching purely through examples and expecting students to develop the skills to arrive at the solutions on their own.  Instead, it eaches explicit methodology that enables students to solve problems of surprising complexity on their own, whether they are 11 or 55, whether in a classroom, a training facility, or your home.  Extensive documentation, material, and software to support this methodology is available for free.\n<\/p><h3>Download<\/h3><div class=\"download\"><div class=\"downloads\"><div class=\"row multiaudio-warning\"><div class=\"col-sm-10 col-md-11\"><div class=\"well\"><h4>These files contain multiple languages.<\/h4><p>\nThis Talk was translated into multiple languages. The files available\nfor download contain all languages as separate audio-tracks. Most\ndesktop video players allow you to choose between them.\n<\/p><p>\nPlease look for \"audio tracks\" in your desktop video player.\n<\/p><\/div><\/div><\/div><div class=\"row slides\"><p><\/p><h4>Slides<\/h4><\/div><\/div><\/div><h3>Related<\/h3><h3>Tags<\/h3><p><a href=\"https:\/\/media.ccc.de\/tags\/35c3\" rel=\"tag\" class=\"label label-default\">35c3<\/a>\n<a href=\"https:\/\/media.ccc.de\/tags\/9800\" rel=\"tag\" class=\"label label-default\">9800<\/a>\n<a href=\"https:\/\/media.ccc.de\/tags\/Hardware%20&amp;%20Making\" rel=\"tag\" class=\"label label-default\">Hardware &amp; Making<\/a><\/p>","created_at":"2019-01-31T12:31:38+0000","updated_at":"2019-01-31T12:31:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"media.ccc.de","preview_picture":"https:\/\/static.media.ccc.de\/media\/congress\/2018\/9800-hd_preview.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","themes"],"is_public":false,"id":803,"title":"Dracula \u00b7 Emacs","url":"https:\/\/draculatheme.com\/emacs\/","content":"Dracula \u2014 A dark theme for Emacs and 50+ apps<div class=\"container-fluid\"><main class=\"wrapper\"><img class=\"preview\" src=\"https:\/\/draculatheme.com\/assets\/img\/screenshots\/emacs.png\" alt=\"Theme Preview\" width=\"949\" height=\"698\" \/><div class=\"instructions\"><h3><a href=\"https:\/\/www.gnu.org\/software\/emacs\/\">Emacs<\/a><\/h3><h4>Install using MELPA<\/h4><pre>M-x package-install &lt;RET&gt; dracula-theme<\/pre><h4>Install using Homebrew<\/h4><pre>brew install dunn\/emacs\/helm<\/pre><h4>Install manually<\/h4><p>Add the emacs theme files to <code>~\/.emacs.d\/themes<\/code>.<\/p><p>To load a theme add the following to your init.el<\/p><pre>(add-to-list 'custom-theme-load-path \"~\/.emacs.d\/themes\")\n(load-theme 'dracula t)<\/pre><h3>Contributors<\/h3><p>This theme wouldn't exist without these people. Wanna help too? Check the <a class=\"cyan\" href=\"https:\/\/github.com\/dracula\/emacs\" target=\"blank\">repository on GitHub<\/a>.<\/p><\/div>\n\t<\/main><\/div><p class=\"credits\">Made with \u2665 by <a class=\"green\" href=\"https:\/\/zenorocha.com\" target=\"blank\">Zeno Rocha<\/a><br \/>under <a class=\"orange\" href=\"http:\/\/zenorocha.mit-license.org\/\" target=\"blank\">MIT license<\/a><\/p>","created_at":"2019-01-30T14:42:37+0000","updated_at":"2019-01-30T14:42:43+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"draculatheme.com","preview_picture":"https:\/\/dracula.github.io\/assets\/img\/facebook.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["sql"],"is_public":false,"id":802,"title":"What is the difference between \"INNER JOIN\" and \"OUTER JOIN\"?","url":"https:\/\/stackoverflow.com\/questions\/38549\/what-is-the-difference-between-inner-join-and-outer-join","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>Also how do <code>LEFT JOIN<\/code>, <code>RIGHT JOIN<\/code> and <code>FULL JOIN<\/code> fit in?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/38549\/revisions\" title=\"show all edits to this post\">edited Jan 25 '17 at 0:13<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2856868\/matt-rowland\">Matt Rowland<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            3,93431930&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Sep 1 '08 at 22:36&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3836\/chris-de-vries\">Chris de Vries<\/a>Chris de Vries&#13;\n        <div class=\"-flair\">&#13;\n            24.3k52527&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"27\">&#13;\n                                27 Answers&#13;\n                            27&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Assuming you're joining on columns with no duplicates, which is a very common case:<\/p>\n<ul><li><p>An inner join of A and B gives the result of A intersect B, i.e. the inner part of a <a href=\"http:\/\/en.wikipedia.org\/wiki\/Venn_diagram\" rel=\"noreferrer\">Venn diagram<\/a> intersection.<\/p><\/li>\n<li><p>An outer join of A and B gives the results of A union B, i.e. the outer parts of a Venn diagram union.<\/p><\/li>\n<\/ul><p><strong>Examples<\/strong><\/p>\n<p>Suppose you have two tables, with a single column each, and data as follows:<\/p>\n<pre>A    B\n-    -\n1    3\n2    4\n3    5\n4    6\n<\/pre>\n<p>Note that (1,2) are unique to A, (3,4) are common, and (5,6) are unique to B.<\/p>\n<p><strong>Inner join<\/strong><\/p>\n<p>An inner join using either of the equivalent queries gives the intersection of the two tables, i.e. the two rows they have in common.<\/p>\n<pre>select * from a INNER JOIN b on a.a = b.b;\nselect a.*, b.*  from a,b where a.a = b.b;\na | b\n--+--\n3 | 3\n4 | 4\n<\/pre>\n<p><strong>Left outer join<\/strong><\/p>\n<p>A left outer join will give all rows in A, plus any common rows in B.<\/p>\n<pre>select * from a LEFT OUTER JOIN b on a.a = b.b;\nselect a.*, b.*  from a,b where a.a = b.b(+);\na |  b\n--+-----\n1 | null\n2 | null\n3 |    3\n4 |    4\n<\/pre>\n<p><strong>Right outer join<\/strong><\/p>\n<p>A right outer join will give all rows in B, plus any common rows in A.<\/p>\n<pre>select * from a RIGHT OUTER JOIN b on a.a = b.b;\nselect a.*, b.*  from a,b where a.a(+) = b.b;\na    |  b\n-----+----\n3    |  3\n4    |  4\nnull |  5\nnull |  6\n<\/pre>\n<p><strong>Full outer join<\/strong><\/p>\n<p>A full outer join will give you the union of A and B, i.e. all the rows in A and all the rows in B. If something in A doesn't have a corresponding datum in B, then the B portion is null, and vice versa.<\/p>\n<pre>select * from a FULL OUTER JOIN b on a.a = b.b;\n a   |  b\n-----+-----\n   1 | null\n   2 | null\n   3 |    3\n   4 |    4\nnull |    6\nnull |    5\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/38578\/revisions\" title=\"show all edits to this post\">edited Sep 28 '17 at 19:52<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5441\/darryl-hein\">Darryl Hein<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            68.6k83187245&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 1 '08 at 22:59&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/116\/mark-harrison\">Mark Harrison<\/a>Mark Harrison&#13;\n        <div class=\"-flair\">&#13;\n            172k102271393&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Also you can consider the following schema for different join types;<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/1UKp7.png\" alt=\"visual explanation of joins\" \/><\/p>\n<p>Source: <strong><a href=\"http:\/\/www.codeproject.com\/Articles\/33052\/Visual-Representation-of-SQL-Joins\">Visual-Representation-of-SQL-Joins<\/a><\/strong> explained in detail by <strong><a href=\"http:\/\/www.codeproject.com\/script\/Membership\/View.aspx?mid=5909363\">C.L. Moffatt<\/a><\/strong><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/16598900\/revisions\" title=\"show all edits to this post\">edited Aug 6 '14 at 14:12<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/662581\/michel-ayres\">Michel Ayres<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            3,81274682&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered May 16 '13 at 23:03&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/929902\/teoman-shipahi\">Teoman shipahi<\/a>Teoman shipahi&#13;\n        <div class=\"-flair\">&#13;\n            34.9k87593&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I recommend <a href=\"http:\/\/blog.codinghorror.com\/a-visual-explanation-of-sql-joins\/\" rel=\"noreferrer\">Jeff's blog article<\/a>. The best description I've ever seen, plus there is a visualization, e.g.:<\/p>\n<p><strong>Inner Join:<\/strong><\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/ZxrNi.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/ZxrNi.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>Full Outer Join:<\/strong><\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/nFPw3.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/nFPw3.png\" alt=\"enter image description here\" \/><\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1353724\/revisions\" title=\"show all edits to this post\">edited Jan 21 '16 at 20:15<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/73226\/martin-smith\">Martin Smith<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            344k58577686&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 30 '09 at 11:52&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/29430\/ya23\">ya23<\/a>ya23&#13;\n        <div class=\"-flair\">&#13;\n            11.1k43840&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The Venn diagrams don't really do it for me. <\/p>\n<p>They don't show any distinction between a cross join and an inner join, for example, or more generally show any distinction between different types of join predicate or provide a framework for reasoning about how they will operate.<\/p>\n<p>There is no substitute for understanding the logical processing and it is relatively straightforward to grasp anyway. <\/p>\n<ol><li>Imagine a cross join.<\/li>\n<li>Evaluate the <code>on<\/code> clause against all rows from step 1 keeping those where the predicate evaluates to <code>true<\/code><\/li>\n<li>(For outer joins only) add back in any outer rows that were lost in step 2.<\/li>\n<\/ol><p>(NB: In practice the query optimiser may find more efficient ways of executing the query than the purely logical description above but the final result must be the same)<\/p>\n<p>I'll start off with an animated version of a <strong>full outer join<\/strong>. Further explanation follows.<\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/VUkfU.gif\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/VUkfU.gif\" alt=\"enter image description here\" \/><\/a><\/p>\n<hr \/>\n<p><strong>Source Tables<\/strong><\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/LVYKx.png\" alt=\"enter link description here\" \/><\/p>\n<p>First start with a <code>CROSS JOIN<\/code> (AKA Cartesian Product). This does not have an <code>ON<\/code> clause and simply returns every combination of rows from the two tables.<\/p>\n<p><strong>SELECT A.Colour, B.Colour FROM A CROSS JOIN B<\/strong><\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/cv3t6.png\" alt=\"enter link description here\" \/><\/p>\n<p>Inner and Outer joins have an \"ON\" clause predicate. <\/p>\n<ul><li><strong>Inner Join.<\/strong> Evaluate the condition in the \"ON\" clause for all rows in the cross join result. If true return the joined row. Otherwise discard it.<\/li>\n<li><strong>Left Outer Join.<\/strong> Same as inner join then for any rows in the left table that did not match anything output these with NULL values for the right table columns.<\/li>\n<li><strong>Right Outer Join.<\/strong> Same as inner join then for any rows in the right table that did not  match anything output these with NULL values for the left table columns.<\/li>\n<li><strong>Full Outer Join.<\/strong> Same as inner join then preserve left non matched rows as in left outer join and right non matching rows as per right outer join.<\/li>\n<\/ul>\n<p><strong>SELECT A.Colour, B.Colour FROM A INNER JOIN B ON A.Colour = B.Colour<\/strong><\/p>\n<p>The above is the classic equi join.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/a8IHd.png\" alt=\"Inner Join\" \/><\/p>\n<h2>Animated Version<\/h2>\n<p><a href=\"https:\/\/i.stack.imgur.com\/kZcvR.gif\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/kZcvR.gif\" alt=\"enter image description here\" \/><\/a><\/p>\n<h3>SELECT A.Colour, B.Colour FROM A INNER JOIN B ON A.Colour NOT IN ('Green','Blue')<\/h3>\n<p>The inner join condition need not necessarily be an equality condition and it need not reference columns from both (or even either) of the tables. Evaluating <code>A.Colour NOT IN ('Green','Blue')<\/code> on each row of the cross join returns.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/ZwoCi.png\" alt=\"inner 2\" \/><\/p>\n<p><strong>SELECT A.Colour, B.Colour FROM A INNER JOIN B ON 1 =1<\/strong><\/p>\n<p>The join condition evaluates to true for all rows in the cross join result so this is just the same as a cross join. I won't repeat the picture of the 16 rows again.<\/p>\n<h3>SELECT A.Colour, B.Colour FROM A LEFT OUTER JOIN B ON A.Colour = B.Colour<\/h3>\n<p>Outer Joins are logically evaluated in the same way as inner joins except that if a row from the left table (for a left join) does not join with any rows from the right hand table at all it is preserved in the result with <code>NULL<\/code> values for the right hand columns.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/4bzv2.png\" alt=\"LOJ\" \/><\/p>\n<h3>SELECT A.Colour, B.Colour FROM A LEFT OUTER JOIN B ON A.Colour = B.Colour WHERE B.Colour IS NULL<\/h3>\n<p>This simply restricts the previous result to only return the rows where <code>B.Colour IS NULL<\/code>. In this particular case these will be the rows that were preserved as they had no match in the right hand table and the query returns the single red row not matched in table <code>B<\/code>. This is known as an anti semi join.<\/p>\n<p>It is important to select a column for the <code>IS NULL<\/code> test that is either not nullable or for which the join condition ensures that any <code>NULL<\/code> values will be excluded in order for this pattern to work correctly and avoid just bringing back rows which happen to have a <code>NULL<\/code> value for that column in addition to the un matched rows.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/d6CVF.png\" alt=\"loj is null\" \/><\/p>\n<h3>SELECT A.Colour, B.Colour FROM A RIGHT OUTER JOIN B ON A.Colour = B.Colour<\/h3>\n<p>Right outer joins act similarly to left outer joins except they preserve non matching rows from the right table and null extend the left hand columns.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/LIOW4.png\" alt=\"ROJ\" \/><\/p>\n<h3>SELECT A.Colour, B.Colour FROM A FULL OUTER JOIN B ON A.Colour = B.Colour<\/h3>\n<p>Full outer joins combine the behaviour of left and right joins and preserve the non matching rows from both the left and the right tables.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/iVoqu.png\" alt=\"FOJ\" \/><\/p>\n<h3>SELECT A.Colour, B.Colour FROM A FULL OUTER JOIN B ON 1 = 0<\/h3>\n<p>No rows in the cross join match the <code>1=0<\/code> predicate. All rows from both sides are preserved using normal outer join rules with NULL in the columns from the table on the other side. <\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/gtIhf.png\" alt=\"FOJ 2\" \/><\/p>\n<h3>SELECT COALESCE(A.Colour, B.Colour) AS Colour FROM A FULL OUTER JOIN B ON 1 = 0<\/h3>\n<p>With a minor amend to the preceding query one could simulate a <code>UNION ALL<\/code> of the two tables. <\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/WPu9W.png\" alt=\"UNION ALL\" \/><\/p>\n<h3>SELECT A.Colour, B.Colour FROM A LEFT OUTER JOIN B ON A.Colour = B.Colour WHERE B.Colour = 'Green'<\/h3>\n<p>Note that the <code>WHERE<\/code> clause (if present) logically runs after the join. One common error is to perform a left outer join and then include a WHERE clause with a condition on the right table that ends up excluding the non matching rows. The above ends up performing the outer join...<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/4bzv2.png\" alt=\"LOJ\" \/><\/p>\n<p>... And then the \"Where\" clause runs. <code>NULL= 'Green'<\/code> does not evaluate to true so the row preserved by the outer join ends up discarded (along with the blue one) effectively converting the join back to an inner one.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/tRHdf.png\" alt=\"LOJtoInner\" \/><\/p>\n<p>If the intention was to include only rows from B where Colour is Green and all rows from A regardless the correct syntax would be<\/p>\n<h3>SELECT A.Colour, B.Colour FROM A LEFT OUTER JOIN B ON A.Colour = B.Colour AND B.Colour = 'Green'<\/h3>\n<p><img src=\"https:\/\/i.stack.imgur.com\/cvJ1s.png\" alt=\"enter image description here\" \/><\/p>\n<h2>SQL Fiddle<\/h2>\n<p>See these examples <a href=\"http:\/\/sqlfiddle.com\/#!17\/10d3d\/29\" rel=\"noreferrer\">run live at SQLFiddle.com<\/a>.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/27458534\/revisions\" title=\"show all edits to this post\">edited Aug 3 '18 at 7:59<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/8962019\/konsta-peltoniemi\">Konsta Peltoniemi<\/a>&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Dec 13 '14 at 11:58&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/73226\/martin-smith\">Martin Smith<\/a>Martin Smith&#13;\n        <div class=\"-flair\">&#13;\n            344k58577686&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>Joins<\/strong> are used to combine the data from two tables, with the result being a new, temporary table. Joins are performed based on something called a predicate, which specifies the condition to use in order to perform a join.  The difference between an inner join and an outer join is that an inner join will return only the rows that actually match based on the join predicate.\nLets consider Employee and Location table:<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/fZ2IZ.png\" alt=\"enter image description here\" \/><\/p>\n<p><strong><em>Inner Join:-<\/em><\/strong>\nInner join creates a new result table by combining column values of two tables (<em>Employee<\/em> and <em>Location<\/em>) based upon the join-predicate. The query compares each row of <em>Employee<\/em> with each row of <em>Location<\/em> to find all pairs of rows which satisfy the join-predicate. When the join-predicate is satisfied by matching non-NULL values, column values for each matched pair of rows of <em>Employee<\/em> and <em>Location<\/em> are combined into a result row.\nHere\u2019s what the SQL for an inner join will look like:<\/p>\n<pre>select  * from employee inner join location on employee.empID = location.empID\nOR\nselect  * from employee, location where employee.empID = location.empID\n<\/pre>\n<p>Now, here is what the result of running that SQL would look like:\n<img src=\"https:\/\/i.stack.imgur.com\/RmdJE.png\" alt=\"enter image description here\" \/><a href=\"https:\/\/i.stack.imgur.com\/LgreJ.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/LgreJ.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>Outer Join:-<\/strong>\nAn outer join does not require each record in the two joined tables to have a matching record. The joined table retains each record\u2014even if no other matching record exists. Outer joins subdivide further into left outer joins and right outer joins, depending on which table's rows are retained (left or right).<\/p>\n<p><strong><em>Left Outer Join:-<\/em><\/strong>\nThe result of a left outer join (or simply left join) for tables <em>Employee<\/em> and <em>Location<\/em> always contains all records of the \"left\" table (<em>Employee<\/em>), even if the join-condition does not find any matching record in the \"right\" table (<em>Location<\/em>).\nHere is what the SQL for a left outer join would look like, using the tables above:<\/p>\n<pre>select  * from employee left outer join location on employee.empID = location.empID;\n\/\/Use of outer keyword is optional\n<\/pre>\n<p>Now, here is what the result of running this SQL would look like:\n<img src=\"https:\/\/i.stack.imgur.com\/U9lNP.png\" alt=\"enter image description here\" \/><a href=\"https:\/\/i.stack.imgur.com\/drdng.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/drdng.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong><em>Right Outer Join:-<\/em><\/strong>\nA right outer join (or right join) closely resembles a left outer join, except with the treatment of the tables reversed. Every row from the \"right\" table (<em>Location<\/em>) will appear in the joined table at least once. If no matching row from the \"left\" table (<em>Employee<\/em>) exists, NULL will appear in columns from <em>Employee<\/em> for those records that have no match in <em>Location<\/em>.\nThis is what the SQL looks like:<\/p>\n<pre>select * from employee right outer join location  on employee.empID = location.empID;\n\/\/Use of outer keyword is optional\n<\/pre>\n<p>Using the tables above, we can show what the result set of a right outer join would look like:<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/c8Jtn.png\" alt=\"enter image description here\" \/><a href=\"https:\/\/i.stack.imgur.com\/buKqD.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/buKqD.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>Full Outer Joins:-<\/strong>\nFull Outer Join or Full Join is to retain the nonmatching information by including nonmatching rows in the results of a join, use a full outer join. It includes all rows from both tables, regardless of whether or not the other table has a matching value.\n<a href=\"https:\/\/i.stack.imgur.com\/23AMx.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/23AMx.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><a href=\"http:\/\/www.sql-join.com\/sql-join-types\/\" rel=\"noreferrer\">Image Source<\/a><\/p>\n<p><a href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/join.html\" rel=\"noreferrer\">MySQL 8.0 Reference Manual - Join Syntax<\/a><\/p>\n<p><a href=\"https:\/\/docs.oracle.com\/javadb\/10.8.3.0\/ref\/rrefsqlj29840.html\" rel=\"noreferrer\">Oracle Join operations<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/27540786\/revisions\" title=\"show all edits to this post\">edited Aug 7 '18 at 14:38<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Dec 18 '14 at 6:54&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2778527\/ajitksharma\">ajitksharma<\/a>ajitksharma&#13;\n        <div class=\"-flair\">&#13;\n            2,90311533&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<h3>Inner Join<\/h3>\n<p>Retrieve the matched rows only, that is, <code>A intersect B<\/code>.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/Zkk3I.jpg\" alt=\"Enter image description here\" \/><\/p>\n<pre>SELECT *\nFROM dbo.Students S\nINNER JOIN dbo.Advisors A\n    ON S.Advisor_ID = A.Advisor_ID\n<\/pre>\n<hr \/><h3>Left Outer Join<\/h3>\n<p>Select all records from the first table, and any records in the second\ntable that match the joined keys.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/Z584b.jpg\" alt=\"Enter image description here\" \/><\/p>\n<pre>SELECT *\nFROM dbo.Students S\nLEFT JOIN dbo.Advisors A\n    ON S.Advisor_ID = A.Advisor_ID\n<\/pre>\n<hr \/><h3>Full Outer Join<\/h3>\n<p>Select all records from the second table, and any records in the first\ntable that match the joined keys.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/c1QF3.jpg\" alt=\"Enter image description here\" \/><\/p>\n<pre>SELECT *\nFROM dbo.Students S\nFULL JOIN dbo.Advisors A\n    ON S.Advisor_ID = A.Advisor_ID\n<\/pre>\n<hr \/><h3>References<\/h3>\n<ul><li><p><em><a href=\"http:\/\/www.datamartist.com\/sql-inner-join-left-outer-join-full-outer-join-examples-with-syntax-for-sql-server\" rel=\"noreferrer\">Inner and outer joins SQL examples and the Join block<\/a><\/em><\/p><\/li>\n<li><p><em><a href=\"http:\/\/www.techonthenet.com\/sql\/joins.php\" rel=\"noreferrer\">SQL: JOINS<\/a><\/em><\/p><\/li>\n<\/ul><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/21380648\/revisions\" title=\"show all edits to this post\">edited Jul 21 '14 at 15:38<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/63550\/peter-mortensen\">Peter Mortensen<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            13.6k1984111&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jan 27 '14 at 12:16&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2224265\/tushar-gupta\">Tushar Gupta<\/a>Tushar Gupta&#13;\n        <div class=\"-flair\">&#13;\n            42.3k188088&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>In simple words:<\/p>\n<p>An <strong>inner join<\/strong> retrieve the matched rows only.<\/p>\n<p>Whereas an <strong>outer join<\/strong> retrieve the matched rows from one table and all rows in other table ....the result depends on which one you are using:<\/p>\n<ul><li><p><strong>Left<\/strong>: Matched rows in the right table and all rows in the left table<\/p><\/li>\n<li><p><strong>Right<\/strong>: Matched rows in the left table and all rows in the right table or <\/p><\/li>\n<li><p><strong>Full<\/strong>: All rows in all tables. It doesn't matter if there is a match or not<\/p><\/li>\n<\/ul><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/14292739\/revisions\" title=\"show all edits to this post\">edited Jul 21 '14 at 15:35<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/63550\/peter-mortensen\">Peter Mortensen<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            13.6k1984111&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jan 12 '13 at 11:07&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1885368\/vidyadhar\">vidyadhar<\/a>vidyadhar&#13;\n        <div class=\"-flair\">&#13;\n            2,05951427&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>A inner join only shows rows if there is a matching record on the other (right) side of the join.<\/p>\n<p>A (left) outer join shows rows for each record on the left hand side, even if there are no matching rows on the other (right) side of the join. If there is no matching row, the columns for the other (right) side would show NULLs.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/38552\/revisions\" title=\"show all edits to this post\">edited Jan 5 '15 at 1:14<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 1 '08 at 22:38&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3146\/1800-information\">1800 INFORMATION<\/a>1800 INFORMATION&#13;\n        <div class=\"-flair\">&#13;\n            98.6k23137223&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Inner joins require that a record with a related ID exist in the joined table.<\/p>\n<p>Outer joins will return records for the left side even if nothing exists for the right side.<\/p>\n<p>For instance, you have an Orders and an OrderDetails table. They are related by an \"OrderID\".<\/p>\n<p><strong>Orders<\/strong><\/p>\n<ul><li>OrderID<\/li>\n<li>CustomerName<\/li>\n<\/ul><p><strong>OrderDetails<\/strong><\/p>\n<ul><li>OrderDetailID<\/li>\n<li>OrderID<\/li>\n<li>ProductName<\/li>\n<li>Qty<\/li>\n<li>Price<\/li>\n<\/ul><p>The request<\/p>\n<pre>SELECT Orders.OrderID, Orders.CustomerName\n  FROM Orders \n INNER JOIN OrderDetails\n    ON Orders.OrderID = OrderDetails.OrderID\n<\/pre>\n<p>will only return Orders that also have something in the OrderDetails table.<\/p>\n<p>If you change it to OUTER LEFT JOIN<\/p>\n<pre>SELECT Orders.OrderID, Orders.CustomerName\n  FROM Orders \n  LEFT JOIN OrderDetails\n    ON Orders.OrderID = OrderDetails.OrderID\n<\/pre>\n<p>then it will return records from the Orders table even if they have no OrderDetails records.<\/p>\n<p>You can use this to find Orders that do not have any OrderDetails indicating a possible orphaned order by adding a where clause like <code>WHERE OrderDetails.OrderID IS NULL<\/code>.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/38565\/revisions\" title=\"show all edits to this post\">edited Sep 10 '18 at 8:25<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5188804\/sergeyur\">SergeyUr<\/a>&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 1 '08 at 22:47&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3747\/brian-boatright\">Brian Boatright<\/a>Brian Boatright&#13;\n        <div class=\"-flair\">&#13;\n            15.9k3174100&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>In simple words : <\/p>\n<p><strong>Inner join<\/strong> -&gt; Take ONLY common records from parent and child tables WHERE primary key of Parent table matches Foreign key in Child table.<\/p>\n<p><strong>Left join<\/strong> -&gt; <\/p>\n<p>pseudo code<\/p>\n<pre>1.Take All records from left Table\n2.for(each record in right table,) {\n    if(Records from left &amp; right table matching on primary &amp; foreign key){\n       use their values as it is as result of join at the right side for 2nd table.\n    } else {\n       put value NULL values in that particular record as result of join at the right side for 2nd table.\n    }\n  }\n<\/pre>\n<p><strong>Right join<\/strong> : Exactly opposite of left join . Put name of table in LEFT JOIN at right side in Right join , you get same output as LEFT JOIN.<\/p>\n<p><strong>Outer join<\/strong> : Show all records in Both tables <code>No matter what<\/code>. If records in Left table are not matching to right table based on Primary , Forieign key , use NULL value as result of join . <\/p>\n<p><strong>Example :<\/strong><\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/pCErn.png\" alt=\"Example\" \/><\/p>\n<p>Lets assume now for 2 tables <\/p>\n<p><code>1.employees  , 2.phone_numbers_employees<\/code><\/p>\n<pre>employees : id , name \nphone_numbers_employees : id , phone_num , emp_id   \n<\/pre>\n<p>Here , employees table is Master table , phone_numbers_employees is child table(it contains <code>emp_id<\/code> as foreign key which connects <code>employee.id<\/code> so its child table.)<\/p>\n<p><strong>Inner joins<\/strong> <\/p>\n<p>Take the records of 2 tables <strong>ONLY IF Primary key of employees table(its id) matches Foreign key of Child table phone_numbers_employees(emp_id)<\/strong>.<\/p>\n<p>So query would be :<\/p>\n<pre>SELECT e.id , e.name , p.phone_num FROM employees AS e INNER JOIN phone_numbers_employees AS p ON e.id = p.emp_id;\n<\/pre>\n<p>Here take only matching rows on primary key = foreign key as explained above.Here non matching rows on primary key = foreign key are skipped as result of join.<\/p>\n<p><strong>Left joins<\/strong> :<\/p>\n<p>Left join retains all rows of the left table, regardless of whether there is a row that matches on the right table.<\/p>\n<pre>SELECT e.id , e.name , p.phone_num FROM employees AS e LEFT JOIN phone_numbers_employees AS p ON e.id = p.emp_id;\n<\/pre>\n<p><strong>Outer joins<\/strong> :<\/p>\n<pre>SELECT e.id , e.name , p.phone_num FROM employees AS e OUTER JOIN phone_numbers_employees AS p ON e.id = p.emp_id;\n<\/pre>\n<p><strong>Diagramatically it looks like :<\/strong><\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/hMKKt.jpg\" alt=\"Diagram\" \/><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/28598795\/revisions\" title=\"show all edits to this post\">edited Feb 19 '15 at 5:11<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 19 '15 at 4:50&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3326275\/pratik-c-joshi\">Pratik C Joshi<\/a>Pratik C Joshi&#13;\n        <div class=\"-flair\">&#13;\n            9,01552858&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>You use <strong><code>INNER JOIN<\/code><\/strong> to return all rows from both tables where there is a match. i.e. In the resulting table all the rows and columns will have values.<\/p>\n<p>In <strong><code>OUTER JOIN<\/code><\/strong> the resulting table may have empty columns. Outer join may be either <code>LEFT<\/code> or <code>RIGHT<\/code>.<\/p>\n<p><strong><code>LEFT OUTER JOIN<\/code><\/strong> returns all the rows from the first table, even if there are no matches in the second table.<\/p>\n<p><strong><code>RIGHT OUTER JOIN<\/code><\/strong> returns all the rows from the second table, even if there are no matches in the first table.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/12616294\/revisions\" title=\"show all edits to this post\">edited Jun 9 '15 at 3:21<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4519059\/sha-t\">shA.t<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            12.9k43665&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 27 '12 at 7:33&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1126071\/vijikumar\">vijikumar<\/a>vijikumar&#13;\n        <div class=\"-flair\">&#13;\n            1,5591016&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><a href=\"https:\/\/i.stack.imgur.com\/zjKep.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/zjKep.png\" alt=\"This is a good explanation for joins\" \/><\/a><\/p>\n<p>This is a good diagrammatic explanation for all kind of joins<\/p>\n<p>source: <a href=\"http:\/\/ssiddique.info\/understanding-sql-joins-in-easy-way.html\" rel=\"noreferrer\">http:\/\/ssiddique.info\/understanding-sql-joins-in-easy-way.html<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/36143190\/revisions\" title=\"show all edits to this post\">edited Oct 18 '16 at 18:48<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Mar 21 '16 at 22:57&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2194364\/raghu-k-nair\">Raghu K Nair<\/a>Raghu K Nair&#13;\n        <div class=\"-flair\">&#13;\n            2,4741934&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><code>INNER JOIN<\/code> requires there is at least a match in comparing the two tables. For example, table A and table B which implies A \u0668 B (A intersection B).<\/p>\n<p><code>LEFT OUTER JOIN<\/code> and <code>LEFT JOIN<\/code> are the same. It gives all the records matching in both tables and all possibilities of the left table.<\/p>\n<p>Similarly, <code>RIGHT OUTER JOIN<\/code> and <code>RIGHT JOIN<\/code> are the same. It gives all the records matching in both tables and all possibilities of the right table.<\/p>\n<p><code>FULL JOIN<\/code> is the combination of <code>LEFT OUTER JOIN<\/code> and <code>RIGHT OUTER JOIN<\/code> without duplication.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/3625912\/revisions\" title=\"show all edits to this post\">edited Sep 26 '13 at 21:25<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2711316\/a-pointar\">A_Pointar<\/a>&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 2 '10 at 9:49&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/437795\/naga\">naga<\/a>naga&#13;\n        &#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The answer is in the meaning of each one, so in the results.<\/p>\n<blockquote>\n  <p><strong>Note :<\/strong><br \/>In <code>SQLite<\/code> there is no <code>RIGHT OUTER JOIN<\/code> or <code>FULL OUTER JOIN<\/code>.<br \/>And also in <code>MySQL<\/code> there is no <code>FULL OUTER JOIN<\/code>.<\/p>\n<\/blockquote>\n<p>My answer is based on above <strong>Note<\/strong>.<\/p>\n<p>When you have two tables like these: <\/p>\n<pre>--[table1]               --[table2]\nid | name                id | name\n---+-------              ---+-------\n1  | a1                  1  | a2\n2  | b1                  3  | b2\n<\/pre>\n<hr \/><p><strong>CROSS JOIN \/ OUTER JOIN :<\/strong><br \/>You can have all of those tables data with <code>CROSS JOIN<\/code> or just with <code>,<\/code> like this:<\/p>\n<pre>SELECT * FROM table1, table2\n--[OR]\nSELECT * FROM table1 CROSS JOIN table2\n--[Results:]\nid | name | id | name \n---+------+----+------\n1  | a1   | 1  | a2\n1  | a1   | 3  | b2\n2  | b1   | 1  | a2\n2  | b1   | 3  | b2\n<\/pre>\n<hr \/><p><strong>INNER JOIN :<\/strong><br \/>When you want to add a filter to above results based on a relation like <code>table1.id = table2.id<\/code> you can use <code>INNER JOIN<\/code>:<\/p>\n<pre>SELECT * FROM table1, table2 WHERE table1.id = table2.id\n--[OR]\nSELECT * FROM table1 INNER JOIN table2 ON table1.id = table2.id\n--[Results:]\nid | name | id | name \n---+------+----+------\n1  | a1   | 1  | a2\n<\/pre>\n<hr \/><p><strong>LEFT [OUTER] JOIN :<\/strong><br \/>When you want to have all rows of one of tables in the above result -with same relation- you can use <code>LEFT JOIN<\/code>:<br \/>(For <strong>RIGHT JOIN<\/strong> just change place of tables)<\/p>\n<pre>SELECT * FROM table1, table2 WHERE table1.id = table2.id \nUNION ALL\nSELECT *, Null, Null FROM table1 WHERE Not table1.id In (SELECT id FROM table2)\n--[OR]\nSELECT * FROM table1 LEFT JOIN table2 ON table1.id = table2.id\n--[Results:]\nid | name | id   | name \n---+------+------+------\n1  | a1   | 1    | a2\n2  | b1   | Null | Null\n<\/pre>\n<hr \/><p><strong>FULL OUTER JOIN :<\/strong><br \/>When you also want to have all rows of the other table in your results you can use <code>FULL OUTER JOIN<\/code>:<\/p>\n<pre>SELECT * FROM table1, table2 WHERE table1.id = table2.id\nUNION ALL\nSELECT *, Null, Null FROM table1 WHERE Not table1.id In (SELECT id FROM table2)\nUNION ALL\nSELECT Null, Null, * FROM table2 WHERE Not table2.id In (SELECT id FROM table1)\n--[OR] (recommended for SQLite)\nSELECT * FROM table1 LEFT JOIN table2 ON table1.id = table2.id\nUNION ALL\nSELECT * FROM table2 LEFT JOIN table1 ON table2.id = table1.id\nWHERE table1.id IS NULL\n--[OR]\nSELECT * FROM table1 FULL OUTER JOIN table2 On table1.id = table2.id\n--[Results:]\nid   | name | id   | name \n-----+------+------+------\n1    | a1   | 1    | a2\n2    | b1   | Null | Null\nNull | Null | 3    | b2\n<\/pre>\n<p>Well, as your need you choose each one that covers your need ;).<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/29606109\/revisions\" title=\"show all edits to this post\">edited Sep 12 '17 at 8:50<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Apr 13 '15 at 13:07&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4519059\/sha-t\">shA.t<\/a>shA.t&#13;\n        <div class=\"-flair\">&#13;\n            12.9k43665&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>Inner join.<\/strong> <\/p>\n<p>A join is combining the rows from two tables. An <em>inner join<\/em> attempts to match up the two tables based on the criteria you specify in the query, and only returns the rows that match. If a row from the first table in the join matches two rows in the second table, then two rows will be returned in the results. If there\u2019s a row in the first table that doesn\u2019t match a row in the second, it\u2019s not returned; likewise, if there\u2019s a row in the second table that doesn\u2019t match a row in the first, it\u2019s not returned.<\/p>\n<p><strong>Outer Join.<\/strong> <\/p>\n<p>A <em>left join<\/em> attempts to find match up the rows from the first table to rows in the second table. If it can\u2019t find a match, it will return the columns from the first table and leave the columns from the second table blank (null).<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/23008416\/revisions\" title=\"show all edits to this post\">edited Feb 2 '16 at 11:14<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1018966\/sandesh\">Sandesh<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,01621535&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Apr 11 '14 at 9:18&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1859906\/kanwar-singh\">Kanwar Singh <\/a>Kanwar Singh &#13;\n        <div class=\"-flair\">&#13;\n            541918&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I don't see much details about performance and optimizer in the other answers.<\/p>\n<p>Sometimes it is good to know that only <code>INNER JOIN<\/code> is associative which means the optimizer has the most option to play with it. It can reorder the join order to make it faster keeping the same result. The optimizer can use the most join modes.<\/p>\n<p>Generally it is a good practice to try to use <code>INNER JOIN<\/code> instead of the different kind of joins. (Of course if it is possible considering the expected result set.)<\/p>\n<p>There are a couple of good examples and explanation here about this strange associative behavior:<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/20030933\/revisions\" title=\"show all edits to this post\">edited May 23 '17 at 12:34<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Nov 17 '13 at 12:59&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1665673\/lajos-veres\">Lajos Veres<\/a>Lajos Veres&#13;\n        <div class=\"-flair\">&#13;\n            12.3k73051&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Having criticized the much-loved red-shaded Venn diagram, I thought it only fair to post my own attempt.<\/p>\n<p>Although @Martin Smith's answer is the best of this bunch by a long way, his only shows the key column from each table, whereas I think ideally non-key columns should also be shown.<\/p>\n<p>The best I could do in the half hour allowed, I still don't think it adequately shows that the nulls are there due to absence of key values in <code>TableB<\/code> or that <code>OUTER JOIN<\/code> is actually a union rather than a join:<\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/fzwkg.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/fzwkg.png\" alt=\"enter image description here\" \/><\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 22 '16 at 15:23&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/15354\/onedaywhen\">onedaywhen<\/a>onedaywhen&#13;\n        <div class=\"-flair\">&#13;\n            43.4k1078121&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The precise algorithm for <code>INNER JOIN<\/code>, <code>LEFT\/RIGHT OUTER JOIN<\/code> are as following: <\/p>\n<ol><li>Take each row from the first table: <code>a<\/code><\/li>\n<li>Consider all rows from second table beside it: <code>(a, b[i])<\/code><\/li>\n<li>Evaluate the <code>ON ...<\/code> clause against each pair: <code>ON( a, b[i] ) = true\/false?<\/code>\n<ul><li>When the condition evaluates to <code>true<\/code>, return that combined row  <code>(a, b[i])<\/code>.<\/li>\n<li>When reach end of second table without any match, and this is an <code>Outer Join<\/code> then return a <em>(virtual)<\/em> pair using <code>Null<\/code> for all columns of other table: <code>(a, Null)<\/code> for LEFT outer join or <code>(Null, b)<\/code> for RIGHT outer join. This is to ensure all rows of first table exists in final results.<\/li>\n<\/ul><\/li>\n<\/ol><p><strong>Note:<\/strong> the condition specified in <code>ON<\/code> clause could be anything, it is not required to use <em>Primary Keys<\/em> (and you don't need to always refer to Columns from both tables)! For example:<\/p>\n<ul><li><code>... ON T1.title = T2.title AND T1.version &lt; T2.version<\/code> ( =&gt; see this post as a sample usage: <a href=\"https:\/\/stackoverflow.com\/a\/7745635\/2803565\">Select only rows with max value on a column<\/a>)<\/li>\n<li><code>... ON T1.y IS NULL<\/code><\/li>\n<li><code>... ON 1 = 0<\/code> (just as sample)<\/li>\n<\/ul><p><a href=\"https:\/\/i.stack.imgur.com\/0dWzY.png\" rel=\"nofollow noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/0dWzY.png\" alt=\"Inner Join vs. Left Outer Join\" \/><\/a><\/p>\n<hr \/><p><a href=\"https:\/\/i.stack.imgur.com\/Bsocp.png\" rel=\"nofollow noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/Bsocp.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>Note:<\/strong> Left Join = Left Outer Join, Right Join = Right Outer Join.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/40486913\/revisions\" title=\"show all edits to this post\">edited Oct 11 '18 at 14:16<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Nov 8 '16 at 12:25&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2803565\/s-serpooshan\">S.Serpooshan<\/a>S.Serpooshan&#13;\n        <div class=\"-flair\">&#13;\n            4,04821637&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>Simplest Definitions<\/strong><\/p>\n<p>Inner Join: Returns <strong>matched records<\/strong> from both tables.<\/p>\n<p>Full Outer Join: Returns matched and <strong>unmatched records<\/strong> from both tables with null for unmatched records from <strong>Both Tables<\/strong>.<\/p>\n<p>Left Outer Join: Returns matched and unmatched records only from table on <strong>Left Side<\/strong>.<\/p>\n<p>Right Outer Join: Returns matched and unmatched records only from table on <strong>Right Side<\/strong>.<\/p>\n<p>In-Short<\/p>\n<p>Matched + Left Unmatched + Right Unmatched = <strong>Full Outer Join<\/strong><\/p>\n<p>Matched + Left Unmatched = <strong>Left Outer Join<\/strong><\/p>\n<p>Matched + Right Unmatched = <strong>Right Outer Join<\/strong><\/p>\n<p>Matched = <strong>Inner Join<\/strong><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Apr 28 '16 at 9:10&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2541634\/akshay-khale\">Akshay Khale<\/a>Akshay Khale&#13;\n        <div class=\"-flair\">&#13;\n            3,94633341&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><a href=\"https:\/\/i.stack.imgur.com\/TBMzF.jpg\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/TBMzF.jpg\" alt=\"enter image description here\" \/><\/a><\/p>\n<ul><li><code>INNER JOIN<\/code> most typical join for two or more tables.\nIt returns data match on both table ON primarykey and forignkey relation.<\/li>\n<li><code>OUTER JOIN<\/code> is same as <code>INNER JOIN<\/code>, but it also include <code>NULL<\/code> data on ResultSet.\n<ul><li><code>LEFT JOIN<\/code> = <code>INNER JOIN<\/code> + Unmatched data of left table with Null match on right table.<\/li>\n<li><code>RIGHT JOIN<\/code> = <code>INNER JOIN<\/code> + Unmatched data of right table with Null match on left table.<\/li>\n<li><code>FULL JOIN<\/code> = <code>INNER JOIN<\/code> + Unmatched data on both right and left tables with null matches.<\/li>\n<\/ul><\/li>\n<li>Self join is not a keyword in SQL, when a table references data in itself knows as self join. Using <code>INNER JOIN<\/code> and <code>OUTER JOIN<\/code> we can write self join queries.<\/li>\n<\/ul><p>For example:<\/p>\n<pre>SELECT * \nFROM   tablea a \n       INNER JOIN tableb b \n               ON a.primary_key = b.foreign_key \n       INNER JOIN tablec c \n               ON b.primary_key = c.foreign_key \n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/47981412\/revisions\" title=\"show all edits to this post\">edited Mar 25 '18 at 7:34<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Dec 26 '17 at 16:49&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1697099\/premraj\">Premraj<\/a>Premraj&#13;\n        <div class=\"-flair\">&#13;\n            29.9k11156117&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<blockquote>\n  <p>What is the difference between \u201cINNER JOIN\u201d and \u201cOUTER JOIN\u201d?<\/p>\n<\/blockquote>\n<p>They are the most commonly used existential operators in SQL, where <code>INNER JOIN<\/code> is used for 'exists' and <code>LEFT OUTER JOIN<\/code> is used for 'does not exist'.<\/p>\n<p>Consider these queries:<\/p>\n<pre>users who have posted and have votes\nusers who have posted but have no badges\n<\/pre>\n<p>People who look for set-based solutions (an industry term) would recognise the respective queries as:<\/p>\n<pre>users who have posted INTERSECT users who have votes\nusers who have posted MINUS users who have badges\n<\/pre>\n<p>Translating these into standard SQL:<\/p>\n<pre>SELECT UserId FROM Posts\nINTERSECT \nSELECT UserId FROM Votes;\nSELECT UserId FROM Posts\nEXCEPT \nSELECT UserId FROM Badges;\n<\/pre>\n<p>Others will think along similar lines of set inclusion:<\/p>\n<pre>users who have posted and IN the set of users who have votes\nusers who have posted and NOT IN the set of users who have badges\n<\/pre>\n<p>Translating these into standard SQL:<\/p>\n<pre>SELECT UserId \n  FROM Posts\n WHERE UserId IN ( SELECT UserId FROM Votes );\nSELECT UserId \n  FROM Posts\n WHERE UserId NOT IN ( SELECT UserId FROM Badges );\n<\/pre>\n<p>Some will think in terms of 'existance' within sets e.g. <\/p>\n<pre>users who have posted and EXIST in the set of users who have votes\nusers who have posted and do NOT EXIST in the set of users who have badges\n<\/pre>\n<p>Translating these into standard SQL (note we now need to use range variables i.e. <code>p<\/code>, <code>v<\/code>, <code>b<\/code>):<\/p>\n<pre>SELECT p.UserId \n  FROM Posts p\n WHERE EXISTS ( SELECT *\n                  FROM Votes v\n                 WHERE v.UserId = p.UserId );\nSELECT p.UserId \n  FROM Posts p\n WHERE NOT EXISTS ( SELECT *\n                      FROM Badges b\n                     WHERE b.UserId = p.UserId );\n<\/pre>\n<p>However, I've found that the \"industry standard\" approach is to exclusively use joins. I don't know what the thinking is here (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Law_of_the_instrument\" rel=\"nofollow noreferrer\">Law of the Instrument<\/a>? <a href=\"http:\/\/c2.com\/cgi\/wiki?PrematureOptimization\" rel=\"nofollow noreferrer\">Premature optimization<\/a>?), so I'll go straight to the syntax:<\/p>\n<pre>SELECT p.UserId \n  FROM Posts p\n       INNER JOIN Votes v ON v.UserId = p.UserId;\nSELECT p.UserId \n  FROM Posts p\n       LEFT JOIN Badges b ON b.UserId = p.UserId\n WHERE b.UserId IS NULL;\n<\/pre>\n<p>Things to note:<\/p>\n<ul><li>The only projection is from <code>Users<\/code> but we still need all those range variables (<code>p<\/code>, <code>v<\/code>, <code>b<\/code>) for search conditions.<\/li>\n<li>The <code>UserId IS NULL<\/code> search condition 'belongs' to the the <code>OUTER JOIN<\/code> but is disconnected in the query.<\/li>\n<li><code>LEFT<\/code> is the industry standard: professionals will rewrite a query to avoid using <code>RIGHT<\/code>!<\/li>\n<li>The <code>OUTER<\/code> keyword from <code>LEFT OUTER JOIN<\/code> is omitted.<\/li>\n<\/ul><p>Closing remarks:<\/p>\n<p>Sometimes joins are used in queries solely to determine whether values exist or do not exists in another set. Learn to look carefully at the attributes being projected (the columns in the <code>SELECT<\/code> clause): if there are none from the joined table then they are simply being used as existential operators. Additionally for outer join, look for instances of <code>&lt;key_column&gt; IS NULL<\/code> in the <code>WHERE<\/code> clause.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/39790040\/revisions\" title=\"show all edits to this post\">edited Apr 19 '18 at 12:48<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 30 '16 at 11:04&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/15354\/onedaywhen\">onedaywhen<\/a>onedaywhen&#13;\n        <div class=\"-flair\">&#13;\n            43.4k1078121&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>In Simple Terms,<\/p>\n<p>1.<strong>INNER JOIN OR EQUI JOIN :<\/strong> Returns the resultset that matches only the condition in both the tables.<\/p>\n<p>2.<strong>OUTER JOIN :<\/strong> Returns the resultset of all the values from both the tables even if there is condition match or not. <\/p>\n<p>3.<strong>LEFT JOIN :<\/strong> Returns the resultset of all the values from left table and only rows that match the condition in right table.<\/p>\n<p>4.<strong>RIGHT JOIN :<\/strong> Returns the resultset of all the values from right table and only rows that match the condition in left table.<\/p>\n<p>5.<strong>FULL JOIN :<\/strong> Full Join and Full outer Join are same.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/40016423\/revisions\" title=\"show all edits to this post\">edited Oct 13 '16 at 9:04<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Oct 13 '16 at 8:53&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4643764\/anands23\">Anands23<\/a>Anands23&#13;\n        <div class=\"-flair\">&#13;\n            353313&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>1.<strong>Inner Join:<\/strong> Also called as Join. It returns the rows present in both the Left table, and right table only <strong>if there is a match<\/strong>. Otherwise, it returns zero records.<\/p>\n<p>Example:<\/p>\n<pre>SELECT\n  e1.emp_name,\n  e2.emp_salary    \nFROM emp1 e1\nINNER JOIN emp2 e2\n  ON e1.emp_id = e2.emp_id\n<\/pre>\n<p><a href=\"https:\/\/i.stack.imgur.com\/lWzlz.jpg\" rel=\"nofollow noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/lWzlz.jpg\" alt=\"output1\" \/><\/a><\/p>\n<p>2.<strong>Full Outer Join:<\/strong> Also called as Full Join. It returns <strong>all the rows<\/strong> present in both the Left table, and right table.<\/p>\n<p>Example:<\/p>\n<pre>SELECT\n  e1.emp_name,\n  e2.emp_salary    \nFROM emp1 e1\nFULL OUTER JOIN emp2 e2\n  ON e1.emp_id = e2.emp_id\n<\/pre>\n<p><a href=\"https:\/\/i.stack.imgur.com\/3yY0Y.jpg\" rel=\"nofollow noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/3yY0Y.jpg\" alt=\"output2\" \/><\/a><\/p>\n<p>3.<strong>Left Outer join:<\/strong> Or simply called as Left Join. It returns all the rows present in the Left table and matching rows from the right table (if any).<\/p>\n<p>4.<strong>Right Outer Join:<\/strong> Also called as Right Join. It returns matching rows from the left table (if any), and all the rows present in the Right table.<\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/UIVk7.png\" rel=\"nofollow noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/UIVk7.png\" alt=\"joins\" \/><\/a><\/p>\n<p><strong>Advantages of Joins<\/strong><\/p>\n<ol><li>Executes faster. <\/li>\n<\/ol><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/45011393\/revisions\" title=\"show all edits to this post\">edited Jul 28 '17 at 4:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jul 10 '17 at 11:56&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/6755093\/laxmi\">Laxmi<\/a>Laxmi&#13;\n        <div class=\"-flair\">&#13;\n            2,7731629&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<ul><li><p><strong>Inner join<\/strong> - An <strong>inner join<\/strong> using either of the equivalent queries gives the intersection of the two <em>tables<\/em>, i.e. the two rows they have in common.<\/p><\/li>\n<li><p><strong>Left outer join<\/strong> - A <strong>left outer join<\/strong> will give all rows in A, plus any common rows in B.<\/p><\/li>\n<li><p><strong>Full outer join<\/strong> - A <strong>full outer join<\/strong> will give you the union of A and B, i.e. All the rows in A and all the rows in B. If something in A doesn't have a corresponding datum in B, then the B portion is null, and vice versay<\/p><\/li>\n<\/ul><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/35152521\/revisions\" title=\"show all edits to this post\">edited Jun 23 '16 at 13:55<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 2 '16 at 11:43&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1018966\/sandesh\">Sandesh<\/a>Sandesh&#13;\n        <div class=\"-flair\">&#13;\n            1,01621535&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The difference between inner join and outer join is as follow:<\/p>\n<ol><li>Inner join is a join that combined tables based on matching tuples, whereas outer join is a join that combined table based on both matched and unmatched tuple.<\/li>\n<li>Inner join merges matched row from two table in where unmatched row are omitted, whereas outer join merges rows from two tables and unmatched rows fill with null value.<\/li>\n<li>Inner join is like an intersection operation, whereas outer join is like an union operation.<\/li>\n<li>Inner join is two types, whereas outer join are three types.<\/li>\n<li>Inner join is slower, whereas outer join is faster than inner join.<\/li>\n<\/ol><\/div><div class=\"user-action-time\">&#13;\n        answered Oct 17 '17 at 12:25&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/6714430\/rashedcs\">rashedcs<\/a>rashedcs&#13;\n        <div class=\"-flair\">&#13;\n            1,1171424&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><code>left join on<\/code> (aka <code>left outer join on<\/code>) returns <code>inner join on<\/code> rows <code>union all<\/code> unmatched left table rows extended by nulls.<\/p>\n<p><code>right join (on<\/code> aka  <code>right outer join on<\/code>) returns <code>inner join on<\/code> rows <code>union all<\/code> unmatched right table rows extended by nulls.<\/p>\n<p><code>full join on<\/code> (aka <code>full outer join on<\/code>) returns <code>inner join on<\/code> rows<code>union all<\/code> unmatched left table rows extended by nulls <code>union all<\/code> unmatched right table rows extended by nulls.<\/p>\n<p>(SQL Standard 2006 SQL\/Foundation 7.7 Syntax Rules 1, General Rules 1 b, 3 c &amp; d, 5 b.) <\/p>\n<p>So don't <code>outer join<\/code> until you know what underlying <code>inner join<\/code> is involved.<\/p>\n<hr \/><p>Find out <a href=\"https:\/\/stackoverflow.com\/a\/25957600\/3404097\">what rows <code>inner join<\/code> returns<\/a>.<\/p>\n<p>Read my comments there re the many confused &amp; poor answers.<\/p>\n<p>Then read my comments here re the many confused &amp; poor answers.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/46091641\/revisions\" title=\"show all edits to this post\">edited Jul 23 '18 at 5:53<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Sep 7 '17 at 8:35&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3404097\/philipxy\">philipxy<\/a>philipxy&#13;\n        <div class=\"-flair\">&#13;\n            11.6k42151&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Consider below 2 tables:<\/p>\n<p><strong>EMP<\/strong><\/p>\n<pre>empid   name    dept_id salary\n1       Rob     1       100\n2       Mark    1       300\n3       John    2       100\n4       Mary    2       300\n5       Bill    3       700\n6       Jose    6       400\n<\/pre>\n<p><strong>Department<\/strong><\/p>\n<pre>deptid  name\n1       IT\n2       Accounts\n3       Security\n4       HR\n5       R&amp;D\n<\/pre>\n<h3>Inner Join:<\/h3>\n<p>Mostly written as just <strong>JOIN<\/strong> in sql queries. It returns only the matching records between the tables.<\/p>\n<h3>Find out all employees and their department names:<\/h3>\n<pre>Select a.empid, a.name, b.name as dept_name\nFROM emp a\nJOIN department b\nON a.dept_id = b.deptid\n;\nempid   name    dept_name\n1       Rob     IT\n2       Mark    IT\n3       John    Accounts\n4       Mary    Accounts\n5       Bill    Security\n<\/pre>\n<p>As you see above, <code>Jose<\/code> is not printed from <strong>EMP<\/strong> in the output as it's dept_id <code>6<\/code> does not find a match in the Department table.  Similarly, <code>HR<\/code> and <code>R&amp;D<\/code> rows are not printed from <strong>Department<\/strong> table as they didn't find a match in the Emp table. <\/p>\n<p><strong>So, INNER JOIN or just JOIN, returns only matching rows.<\/strong><\/p>\n<h3>LEFT JOIN :<\/h3>\n<p>This returns all records from the LEFT table and only matching records from the RIGHT table.<\/p>\n<pre>Select a.empid, a.name, b.name as dept_name\nFROM emp a\nLEFT JOIN department b\nON a.dept_id = b.deptid\n;\nempid   name    dept_name\n1       Rob     IT\n2       Mark    IT\n3       John    Accounts\n4       Mary    Accounts\n5       Bill    Security\n6       Jose    \n<\/pre>\n<p>So, if you observe the above output, all records from the LEFT table(Emp) are printed with just matching records from RIGHT table.<\/p>\n<p><code>HR<\/code> and <code>R&amp;D<\/code> rows are not printed from <strong>Department<\/strong> table as they didn't find a match in the Emp table on dept_id.<\/p>\n<p><strong>So, LEFT JOIN returns ALL rows from Left table and only matching rows from RIGHT table.<\/strong> <\/p>\n<p>Can also check DEMO <a href=\"https:\/\/dbfiddle.uk\/?rdbms=postgres_11&amp;fiddle=9dbdd6cf405ae24c5e2a0798a732d389\" rel=\"nofollow noreferrer\">here<\/a>.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/53468691\/revisions\" title=\"show all edits to this post\">edited Nov 25 '18 at 15:04<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Nov 25 '18 at 14:50&#13;<\/div><div class=\"user-details\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5820814\/mayank-porwal\">Mayank Porwal<\/a>Mayank Porwal&#13;\n        <div class=\"-flair\">&#13;\n            4,8852724&#13;\n        <\/div>&#13;<\/div>","created_at":"2019-01-30T11:15:26+0000","updated_at":"2019-01-30T11:15:31+0000","annotations":[],"mimetype":"text\/html","reading_time":31,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["activitypub"],"is_public":false,"id":801,"title":"SocialCG\/ActivityPub\/MediaUpload - W3C Wiki","url":"https:\/\/www.w3.org\/wiki\/SocialCG\/ActivityPub\/MediaUpload","content":"<p>The Media Upload mechanism defines a protocol for ActivityPub servers to support uploading document types to be referenced in activites, such as images, video or other binary data. To accomplish this, a client MUST submit a <code>multipart\/form-data<\/code> message to the user's <a href=\"#uploadMedia\"><code>uploadMedia<\/code><\/a> endpoint on their ActivityStreams profile object. (Unlike most client to server interactions, using this endpoint does not involve submitting to the <a href=\"#outbox\"><code>outbox<\/code><\/a>). A client should expect that it must be properly authenticated in order to be able to upload media.\n<\/p><p>The <code>uploadMedia<\/code> endpoint is part of the ActivityPub <code>endpoints<\/code> mapping. See <a href=\"https:\/\/www.w3.org\/wiki\/ActivityPub_extensions#uploadMedia\" title=\"ActivityPub extensions\">ActivityPub extensions<\/a>.\n<\/p><p>The submitted form data should contain two parts \/ fields:\n<\/p><dl><dt> file<\/dt>\n<dd> The media file file being uploaded.<\/dd>\n<dt> object<\/dt>\n<dd> A shell of an ActivityStreams object, which will be finalized by the server.<\/dd><\/dl><p>Assuming that the server accepts the request and that the user was appropriately authorized to upload media, servers MUST respond with a <code>201 Created<\/code> if the object is immediately available or a <code>202 Accepted<\/code> if the server is still processing the submitted media. The response MUST contain a <code>Location<\/code> header pointing to the new or to-be-created object's <code>id<\/code>.\n<\/p><p>The server, having done any appropriate processing on the received <code>file<\/code> and putting it in place, transforms the <code>object<\/code> that will be retrievable by the <code>id<\/code>. In particular, servers MUST append an <code>id<\/code> property to the object, and SHOULD include the uploaded and\/or processed file paths in the object's <code>url<\/code> property. The server MAY wrap the shell object submitted by the user in a <code>Create<\/code> via [[#object-without-create|]] if appropriate and present this as the object pointed to by the forementioned <code>Location<\/code> header in the post-media-upload response.\n<\/p><p>Ben Bitdiddle just recorded a video of his friend's latest invention and wants to upload it to his ActivityPub-compatible social network to show it off. He wants to post a <code>Create<\/code> of a new <code>Video<\/code> object, but first he has to upload the video somewhere. Luckily the ActivityPub instance he is on supports the <code>uploadMedia<\/code> endpoint! In his client, he selects the video for upload and adds a title and a description and presses submit.\n<\/p><p>Ben's client takes a look at Ben's <a href=\"#actors\">actor<\/a> profile and finds the <code>uploadMedia<\/code> endpoint there.\n<\/p><pre class=\"example highlight json\">            {\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n             \"type\": \"Person\",\n             \"id\": \"https:\/\/chatty.example\/ben\/\",\n             \"name\": \"Ben Bitdiddle\",\n             \"preferredUsername\": \"ben\",\n             \"summary\": \"Constantly fidgeting with ones and zeroes.\",\n             \"inbox\": \"https:\/\/chatty.example\/ben\/inbox\/\",\n             \"outbox\": \"https:\/\/chatty.example\/ben\/collections\/outbox\/\",\n             \"followers\": \"https:\/\/chatty.example\/ben\/collections\/followers\/\",\n             \"following\": \"https:\/\/chatty.example\/ben\/collections\/following\/\",\n             \"liked\": \"https:\/\/chatty.example\/ben\/collections\/liked\/\",\n             \"endpoints\": {\n               \"id\": \"https:\/\/chatty.example\/ben\/#endpoints\",\n               \"uploadMedia\": \"https:\/\/chatty.example\/ben\/endpoints\/uploadMedia\/\"}}\n          <\/pre><p>Ben's client then constructs a shell ActivityPub object (this doesn't have the url for the video itself yet, but that's okay, because that will be added by the server) and submits it (as the <code>object<\/code>) along with the video (as the <code>file<\/code>) by <code>POST<\/code>'ing to the <code>uploadMedia<\/code> endpoint.\n<\/p><pre class=\"example\">            POST \/ben\/endpoints\/uploadMedia\/ HTTP\/1.1\n            Host: chatty.example\n            Content-Type: multipart\/form-data; boundary=bfe40faf4ef\n            Authorization: Bearer XXXXXXXXXXX\n  \n            --bfe40faf4ef\n            Content-Disposition: form-data; name=\"object\"\n            Content-Type: application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"\n  \n            {\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n             \"type\": \"Create\",\n             \"actor\": \"https:\/\/chatty.example\/ben\/\",\n             \"to\": [\"https:\/\/chatty.example\/ben\/collections\/followers\/\",\n                    \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"],\n             \"object\": {\n               \"type\": \"Video\",\n               \"name\": \"A wireworld computer!\",\n               \"attributedTo\": \"https:\/\/chatty.example\/ben\/\",\n               \"to\": [\"https:\/\/chatty.example\/ben\/collections\/followers\/\",\n                      \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"],\n               \"summary\": \"&lt;p&gt;My friend Conway and I went over to our friend\n                              Silverman's place and saw the wireworld computer\n                              he's been constructing.  Pretty crazy!&lt;\/p&gt;\n                           &lt;p&gt;Tomorrow we're going over to our friend Langton's\n                              place and see that ant farm he's been working\n                              on...&lt;\/p&gt;\"}}\n  \n            --bfe40faf4ef\n            Content-Disposition: form-data; name=\"file\"; filename=\"wireworld.webm\"\n            Content-Type: video\/webm\n  \n            ... (binary data) ...\n            --bfe40faf4ef\n          <\/pre><p>Ben's client receives an HTTP response from the server with the status code <code>202 Accepted<\/code> since the video is still being transcoded. The response contains a <code>Location<\/code> header pointing at the url <code><a rel=\"nofollow\" class=\"external free\" href=\"https:\/\/chatty.example\/ben\/posts\/d4152a7c\/\">https:\/\/chatty.example\/ben\/posts\/d4152a7c\/<\/a><\/code>. Sure enough, a few minutes later Ben's video shows up in his <code>outbox<\/code> with the rest of the fields filled in:\n<\/p><pre class=\"example highlight json\">            {\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n             \"type\": \"Create\",\n             \"id\": \"https:\/\/chatty.example\/ben\/posts\/d4152a7c\/\",\n             \"actor\": \"https:\/\/chatty.example\/ben\/\",\n             \"to\": [\"https:\/\/chatty.example\/ben\/collections\/followers\/\",\n                    \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"],\n             \"object\": {\n               \"type\": \"Video\",\n               \"id\": \"https:\/\/chatty.example\/ben\/posts\/6e69f564\/\",\n               \"url\": [\n                 {\"type\": \"Link\",\n                  \"href\": \"https:\/\/chatty.example\/media\/6e69f564\/wireworld-small.webm\",\n                  \"mediaType\": \"video\/webm; codecs=\\\"vp8, vorbis\\\"\",\n                  \"name\": \"A wireworld computer! (small)\",\n                  \"width\": 640,\n                  \"height\": 480},\n                 {\"type\": \"Link\",\n                  \"href\": \"https:\/\/chatty.example\/media\/6e69f564\/wireworld-large.webm\",\n                  \"mediaType\": \"video\/webm; codecs=\\\"vp8, vorbis\\\"\",\n                  \"name\": \"A wireworld computer! (large)\",\n                  \"width\": 1280,\n                  \"height\": 720},\n               \"name\": \"A wireworld computer!\",\n               \"attributedTo\": \"https:\/\/chatty.example\/ben\/\",\n               \"to\": [\"https:\/\/chatty.example\/ben\/collections\/followers\/\",\n                      \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"],\n               \"summary\": \"&lt;p&gt;My friend Conway and I went over to our friend\n                              Silverman's place and saw the wireworld computer\n                              he's been constructing.  Pretty crazy!&lt;\/p&gt;\n                           &lt;p&gt;Tomorrow we're going over to our friend Langton's\n                              place and see that ant farm he's been working\n                              on...&lt;\/p&gt;\"}}\n          <\/pre>","created_at":"2019-01-30T10:48:31+0000","updated_at":"2019-01-30T10:48:36+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.w3.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["amqp","rabbitmq"],"is_public":false,"id":800,"title":"Jason\u2019s .plan \u00bb Blog Archive \u00bb Rabbits and warrens.","url":"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/2009\/01\/rabbits-and-warrens\/","content":"<p><img src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/www.dkimages.com\/discover\/previews\/811\/20089825.JPG\" border=\"0\" alt=\"20089825.JPG\" width=\"384\" height=\"293\" \/><\/p>\n\n<p><strong>If you like Rabbit and Warrens checkout RabbitMQ in Action in the sidebar.<\/strong><\/p>\n\n<p>The goal was simple enough: decouple a particular type of analysis out-of-band from mainstream e-mail processing. We started down the MySQL road\u2026put the things to be digested into a table\u2026consume them in another daemon\u2026bada bing bada boom. But pretty soon, complex ugliness crept into the design phase\u2026 You want to have multiple daemons servicing the queue?\u2026no problem we\u2019ll just hard code node numbers\u2026what? you want dynamic load re-assignment when daemons join and die?<\/p>\n<p>You get the idea\u2026what was supposed to be simple (decouple something) was spinning its own <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/en.wikipedia.org\/wiki\/Gordian_knot\">Gordian knot<\/a>. It seemed like a good time to see if every problem was looking like a nail (table), because all we had were hammers (MySQL).<\/p>\n<p>A short search later, and we entered the world of message queueing. No, no\u2026we know obviously what a message queue is. Heck, we do e-mail for a living. We\u2019ve implemented all sorts of specialized, high-speed, in-memory queues for e-mail processing. What we weren\u2019t aware of was the family of off-the-shelf, generalized, message queueing (MQ) servers\u2026a language-agnostic, no-assembly required way to wire routing between applications over a network. A message queue we didn\u2019t have to write ourselves? Hold your tongue.<\/p>\n<h2>Open up your queue\u2026<\/h2>\n<p>Cutting to the chase, over the last 4 years there have been no shortage of open-source message queueing servers written. Most of them are one-offs by folks like LiveJournal to scratch a particular itch. Yeah, they don\u2019t really care what kind of messages they carry, but their design parameters are usually creator-specific (and message persistence after a crash usually isn\u2019t one of them). However, there are three in-particular, that are designed to be highly flexible message queues for their own sake:<\/p>\n<ul><li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/activemq.apache.org\/\">Apache ActiveMQ<\/a><\/li>\n<li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.zeromq.org\/\">ZeroMQ<\/a><\/li>\n<li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.rabbitmq.com\/\">RabbitMQ<\/a><\/li>\n<\/ul><p>Apache ActiveMQ gets the most press, but it appears to have some issues not losing messages. <strong>Next.<\/strong><\/p>\n<p>ZeroMQ and RabbitMQ both support an open messaging protocol called AMQP. The advantage to AMQP is that it\u2019s designed to be a highly-robust and open alternative to the two commercial message queues out there (IBM and Tibco). Muy bueno. However, ZeroMQ doesn\u2019t support message persistence across crashes reboots. No muy bueno. That leaves us with RabbitMQ. (That being said if you don\u2019t need persistence ZeroMQ is pretty darn interesting\u2026incredibly low latency and flexible topologies).<\/p>\n<h2>That leaves us with the carrot muncher\u2026<\/h2>\n<p><img src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/www.rabbitmq.com\/img\/RabbitMQLogo.png\" border=\"0\" alt=\"\" \/><\/p>\n<p>RabbitMQ pretty much sold me the minute I read \u201cwritten in Erlang\u201d. <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/en.wikipedia.org\/wiki\/Erlang_(programming_language)\">Erlang<\/a> is a highly parallel programming language developed over at Ericsson for running telco switches\u2026yeah the kind with six bazillion 9s of uptime. In Erlang, its supposedly trivial to spin off processes and then communicate between them using message passing. Seems like the ideal underpinning for a message queue no?<\/p>\n<p>Also, RabbitMQ supports persistence. Yes Virginia, if your RabbitMQ dies, your messages don\u2019t have to die an unwitting death\u2026they can be reborn in your queues on reboot. Oh\u2026and as is always desired @ DigiTar, it <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/barryp.org\/software\/py-amqplib\/\">plays nicely with python<\/a>. All that being said, RabbitMQs documentation is well\u2026horrible. Lemme rephrase, if you already understand AMQP, the docs are fine. But how many folks know AMQP? It\u2019d be like MySQL docs assuming you knew some form of SQL\u2026er\u2026nevermind.<\/p>\n<p>So, without further ado\u2026here is a reduction of a weeks\u2019 worth of reading up on AMQP and how it works in RabbitMQ\u2026and how to play with it in Python:<\/p>\n<h2>Playing telephone<\/h2>\n<p>There are four building blocks you really care about in AMQP: virtual hosts, exchanges, queues and bindings. A virtual host holds a bundle of exchanges, queues and bindings. Why would you want multiple virtual hosts? Easy. A username in RabbitMQ grants you access to a virtual host\u2026in its entirety. So the only way to keep group A from accessing group B\u2019s exchanges\/queues\/bindings\/etc. is to create a virtual host for A and one for B. Every RabbitMQ server has a default virtual host named \u201c\/\u201d. If that\u2019s all you need, you\u2019re ready to roll.<\/p>\n<h2>Exchanges, Queues and bindings\u2026oh my!<\/h2>\n<p>Here\u2019s where my railcar went off the tracks initially. How do all the parts thread together?<\/p>\n<p>Queues are where your \u201cmessages\u201d end up. They\u2019re message buckets\u2026and your messages sit there until a client (a.k.a. consumer) connects to the queue and siphons it off. However, you can configure a queue so that if there isn\u2019t a consumer ready to accept the message when it hits the queue, the message goes poof. But we digress\u2026<\/p>\n<p>The important thing to remember is that queues are created programmatically by your consumers (not via a configuration file or command line program). That\u2019s OK, because if a consumer app tries to \u201ccreate\u201d a queue that already exists, RabbitMQ pats it on the head, smiles gently and NOOPs the request. So you can keep your MQ configuration in-line with your app code\u2026what a concept.<\/p>\n<p>OK, so you\u2019ve created and attached to your queue, and your consumer app is drumming its fingers waiting for a message\u2026and drumming\u2026and drumming\u2026but alas no message. What happened? Well you gotta pump a message in first! But to do that you\u2019ve got to have an exchange\u2026<\/p>\n<p>Exchanges are routers with routing tables. That\u2019s it. End stop. Every message has what\u2019s known as a \u201crouting key\u201d, which is simply a string. The exchange has a list of bindings (routes) that say, for example, messages with routing key \u201cX\u201d go to queue \u201ctimbuktu\u201d. But we get slightly ahead of ourselves.<\/p>\n<p>Your consumer application should create your exchanges (plural). Wait? You mean you can have more than one exchange? Yes, you can, but why? Easy. Each exchange operates in its own userland process, so adding exchanges, adds processes allowing you to scale message routing capacity with the number of cores in your server. As an example, on an 8-core server you could create 5 exchanges to maximize your utilization, leaving 3 cores open for handling the queues, etc.. Similarly, in a RabbitMQ cluster, you can use the same principle to spread exchanges across the cluster members to add even more throughput.<\/p>\n<p>OK, so you\u2019ve created an exchange\u2026but it doesn\u2019t know what queues the messages go in. You need \u201crouting rules\u201d (bindings). A binding essentially says things like this: put messages that show up in exchange \u201cdesert\u201d and have routing key \u201cali-baba\u201d into the queue \u201chideout\u201d. In other words, a binding is a routing rule that links an exchange to a queue based on a routing key. It is possible for two binding rules to use the same routing key. For example, maybe messages with the routing key \u201caudit\u201d need to go <strong>both<\/strong> to the \u201clog-forever\u201d queue and the \u201calert-the-big-dude\u201d queue. To accomplish this, just create two binding rules (each one linking the exchange to one of the queues) that both trigger on routing key \u201caudit\u201d. In this case, the exchange duplicates the message and sends it to both queues. Exchanges are just routing tables containing bindings.<\/p>\n<p>Now for the curveball: there are multiple types of exchanges. They all do routing, but they accept different styles of binding \u201crules\u201d. Why not just create one type of exchange for all style of rules? Because each rule style has a different CPU cost for analyzing if a message matches the rule. For example, a \u201ctopic\u201d exchange tries to match a message\u2019s routing key against a pattern like \u201c<strong>dogs.*<\/strong>\u201d. Matching that wildcard on the end takes more CPU than simply seeing if the routing key is \u201c<strong>dogs<\/strong>\u201d or not (e.g. a \u201cdirect\u201d exchange). If you don\u2019t need the extra flexibility of a \u201ctopic\u201d exchange, you can get more messages\/sec routed if you choose the \u201cdirect\u201d exchange type. So what are the types and how do they route?<\/p>\n<p><strong>Fanout Exchange<\/strong> \u2013 No routing keys involved. You simply bind a queue to the exchange. Any message that is sent to the exchange is sent to <strong>all<\/strong> queues bound to that exchange. Think of it like a subnet broadcast. Any host on the subnet gets a copy of the packet. Fanout exchanges route messages the fastest.<\/p>\n<p><strong>Direct Exchange<\/strong> \u2013 Routing keys are involved. A queue binds to the exchange to request messages that match a particular routing key <em>exactly<\/em>. This is a straight match. If a queue binds to the exchange requesting messages with routing key \u201c<strong>dog<\/strong>\u201d, only messages labelled \u201c<strong>dog<\/strong>\u201d get sent to that queue (not \u201c<strong>dog.puppy<\/strong>\u201d, not \u201c<strong>dog.guard<\/strong>\u201c\u2026only \u201c<strong>dog<\/strong>\u201d).<\/p>\n<p><strong>Topic Exchange<\/strong> \u2013 Matches routing keys against a pattern. Instead of binding with a particular routing key, the queue binds with a pattern string. The symbol <strong>#<\/strong> matches one or more words, and the symbol <strong>*<\/strong> matches any single word (no more, no less). So \u201c<strong>audit.#<\/strong>\u201d would match \u201c<strong>audit.irs.corporate<\/strong>\u201d, but \u201c<strong>audit.*<\/strong>\u201d would only match \u201c<strong>audit.irs<\/strong>\u201d. Our friends at RedHat have put together a great image to express how topic exchanges work:<\/p>\n<div><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/uploads\/2009\/01\/chapter-2_amqp_stack.png\" rel=\"lightbox[84]\"><img class=\"size-full wp-image-141\" title=\"chapter-2_amqp_stack\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/uploads\/2009\/01\/chapter-2_amqp_stack.png\" alt=\"AMQP Stack Diagram\" width=\"480\" height=\"588\" \/><\/a><\/div>\n<p>Source: <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/manning.com\/videla\">RabbitMQ in Action<\/a> (by me and a very cool Uruguayan dude\u2026Mr. Alvaro)<\/p>\n<h2>Persistent little bugger\u2026<\/h2>\n<p>You spend all that time creating your queues, exchanges and bindings, and then BANG!\u2026the server fries faster than the griddle at McDonald\u2019s. All your queues, exchanges and bindings are there right? Oh geez\u2026what about the messages in the queues you hadn\u2019t serviced yet?<\/p>\n<p>Relax, providing you created everything with the default arguments, it\u2019s all gone\u2026poof\u2026whoosh\u2026nada\u2026nil. That\u2019s right, RabbitMQ rebooted as empty as a baby\u2019s noggin. You gotta redo everything kemosabe. How do you keep this from happening in the future?<\/p>\n<p>On your queues and your exchanges there\u2019s a creation-time flag called \u201cdurable\u201d. There\u2019s only one thing durable means in AMQP-land\u2026the queue or exchange marked durable will be re-created automatically on reboot. It <strong>does not mean<\/strong> the messages in the queues will survive the reboot. They won\u2019t. So how do we make not only our config but messages persist through a reboot?<\/p>\n<p>Well the first question is, do you really want your messages to persist? For a message to last through a reboot, it has to be written to disk, and even a simple checkpoint to disk takes time. If you value message routing speed more than the contents of the message, don\u2019t make your messages persistent. That being said, for our particular needs @ DigiTar, persistence is important.<\/p>\n<p>When you publish your message to an exchange, there\u2019s a flag called \u201cDelivery Mode\u201d. Depending on the AMQP library you\u2019re using there will be different ways of setting it (we\u2019ll cover the Python library later). But the long and the short of it is you want the \u201cDelivery Mode\u201d set to the value 2, which means \u201cpersistent\u201d. \u201cDelivery Mode\u201d usually (depending on your AMQP library) defaults to a value of 1, which means \u201cnon-persistent\u201d. So the steps for persistent messaging are:<\/p>\n<ol><li>Mark the exchange \u201cdurable\u201d.<\/li>\n<li>Mark the queue \u201cdurable\u201d.<\/li>\n<li>Set the message\u2019s \u201cdelivery mode\u201d to a value of 2<\/li>\n<\/ol><p>That\u2019s it. Not really rocket science, but enough moving parts to make a mistake and send little Sally\u2019s dental records into cyber-Nirvana.<\/p>\n<p>There may be one thing nagging you though\u2026what about the binding? We didn\u2019t mark the binding \u201cdurable\u201d when we created it. It\u2019s alright. If you bind a durable queue to a durable exchange, RabbitMQ will automatically preserve the binding. Similarly, if you delete any exchange\/queue (durable or not) any bindings that depend on it get deleted automatically.<\/p>\n<p>Two things to be aware of:<\/p>\n<ul><li>RabbitMQ will <strong>not<\/strong> allow you to bind a <strong>non-durable<\/strong> exchange to a <strong>durable<\/strong> queue, or vice-versa. Both the exchange and the queue must be durable for the binding operation to succeed.<\/li>\n<li>You cannot change the creation flags on a queue or exchange after you\u2019ve created it. For example, if you create a queue as \u201cnon-durable\u201d, and want to change it to \u201cdurable\u201d, the only way to do this is to destroy the queue and re-create it. It\u2019s a good reason to double check your declarations.<\/li>\n<\/ul><h2>Food for snakes<\/h2>\n<p>A real empty area for AMQP usage is using it in Python programs. For other languages there are plenty of references:<\/p>\n<ul><li>Java \u2013 <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.rabbitmq.com\/java-client.html\">http:\/\/www.rabbitmq.com\/java-client.html<\/a><\/li>\n<li>.NET \u2013 <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.rabbitmq.com\/releases\/rabbitmq-dotnet-client\/v1.5.0\/rabbitmq-dotnet-client-1.5.0-user-guide.pdf\">http:\/\/www.rabbitmq.com\/releases\/rabbitmq-dotnet-client\/v1.5.0\/rabbitmq-dotnet-client-1.5.0-user-guide.pdf<\/a><\/li>\n<li>Ruby \u2013 <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/somic.org\/2008\/06\/24\/ruby-amqp-rabbitmq-example\/\">http:\/\/somic.org\/2008\/06\/24\/ruby-amqp-rabbitmq-example\/<\/a><\/li>\n<\/ul><p>But for little old Python, you need to dig it out yourself. So other folks don\u2019t have to wander in the wilderness like I did, here\u2019s a little primer on using Python to do the AMQP-tasks we\u2019ve talked about:<\/p>\n<p>First, you\u2019ll need a Python AMQP library\u2026and there are two:<\/p>\n<ul><li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/barryp.org\/software\/py-amqplib\/\">py-amqplib<\/a> \u2013 General AMQP library<\/li>\n<li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/https:\/\/launchpad.net\/txamqp\">txAMQP<\/a> \u2013 An AMQP library that uses the <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.twistedmatrix.com\/\">Twisted<\/a> framework, thereby allowing asynchronous I\/O.<\/li>\n<\/ul><p>Depending on your needs, py-amqplib or txAMQP may be more to your liking. Being Twisted-based, txAMQP holds the promise of building super performing AMQP consumers that use async I\/O. But Twisted programming is a topic all its own\u2026so we\u2019re going to use py-amqplib for clarity\u2019s sake. <strong>UPDATE: Please check the comments for example code showing use of txAMQP from Esteve Fernandez.<\/strong><\/p>\n<p>AMQP supports pipelining multiple MQ communication channels over one TCP connection, where each channel is a communication stream used by your program. Every AMQP program has at least one connection and one channel:<\/p>\n<div id=\"wpshdo_1\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_1\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_1\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_1\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_1\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_1\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_1\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">from amqplib import client_0_8 as amqp\nconn = amqp.Connection(host=\"localhost:5672\u00a0\", userid=\"guest\",\npassword=\"guest\", virtual_host=\"\/\", insist=False)\nchan = conn.channel()<\/pre><\/div><\/div>\n<p>Each channel is assigned an integer channel number automatically by the <strong>.channel()<\/strong> method of the <strong>Connection()<\/strong> class. Alternately, you can specify the channel number yourself by calling <strong>.channel(x)<\/strong> , where x is the channel number you want. More often than not, its a good idea to just let the <strong>.channel()<\/strong> method auto-assign the channel number to avoid collisions.<\/p>\n<p>Now we\u2019ve got a connection and channel to talk over. At this point, our code is going to diverge into two applications that use that same bit we\u2019ve created so far: a consumer and the publisher. Let\u2019s create the consumer app by creating a queue named \u201c<strong>po_box<\/strong>\u201d and an exchange named \u201c<strong>sorting_room<\/strong>\u201d:<\/p>\n<div id=\"wpshdo_2\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_2\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_2\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_2\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_2\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_2\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_2\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">chan.queue_declare(queue=\"po_box\", durable=True,\nexclusive=False, auto_delete=False)\nchan.exchange_declare(exchange=\"sorting_room\", type=\"direct\", durable=True,\nauto_delete=False,)<\/pre><\/div><\/div>\n<p>What did that do? First, it created a queue called \u201c<strong>po_box<\/strong>\u201d that is durable (will be re-created on reboot) and will not be automatically deleted when the last consumer detaches from it (<strong>auto_delete=False<\/strong>). It\u2019s important to set <strong>auto_delete<\/strong> to false when making a queue (or exchange) durable, otherwise the queue itself will disappear when the last consumer detaches (regardless of the <strong>durable<\/strong> flag). Setting both <strong>durable<\/strong> and <strong>auto_delete<\/strong> to true, would make a queue that would be recreated only if RabbitMQ died unexpectedly with consumers still attached.<\/p>\n<p>(You may have noticed there\u2019s another flag specified called \u201c<strong>exclusive<\/strong>\u201d. If set to true, only the consumer that creates the queue will be allowed to attach to it. It\u2019s a queue that is private to the creating consumer.)<\/p>\n<p>There\u2019s also the exchange declaration for the \u201c<strong>sorting_room<\/strong>\u201d exchange. <strong>auto_delete<\/strong> and <strong>durable<\/strong> mean the same things as they do in a queue declaration. However, <strong>.exchange_declare()<\/strong> introduces an argument called <strong>type<\/strong> that defines what type of exchange you\u2019re making (as described earlier): <strong>fanout<\/strong>, <strong>direct<\/strong> or <strong>topic<\/strong>.<\/p>\n<p>At this point, you\u2019ve got a queue to receive messages and an exchange to publish them to initially\u2026but we need a binding to link the two together:<\/p>\n<div id=\"wpshdo_3\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_3\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_3\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_3\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_3\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_3\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_3\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">chan.queue_bind(queue=\"po_box\", exchange=\"sorting_room\",\nrouting_key=\"jason\")<\/pre><\/div><\/div>\n<p>The binding is pretty straight forward. Any messages arriving at the \u201c<strong>sorting_room<\/strong>\u201d exchange with the routing key \u201c<strong>jason<\/strong>\u201d gets routed to the \u201c<strong>po_box<\/strong>\u201d queue.<\/p>\n<p>Now, there\u2019s two methods of getting messages out of the queue. The first is to call <strong>chan.basic_get()<\/strong> to pull the next message off the queue (if there are no messages waiting on the queue, <strong>chan.basic_get()<\/strong> will return a None object\u2026thereby blowing up the <strong>print msg.body<\/strong> code below if not trapped) :<\/p>\n<div id=\"wpshdo_4\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_4\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_4\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_4\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_4\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_4\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_4\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">msg = chan.basic_get(\"po_box\")\nprint msg.body\nchan.basic_ack(msg.delivery_tag)<\/pre><\/div><\/div>\n<p>But what if you want your application to be notified as soon as a message is available for it? To do that, instead of <strong>chan.basic_get()<\/strong>, you need to register a callback for new messages using <strong>chan.basic_consume()<\/strong>:<\/p>\n<div id=\"wpshdo_5\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_5\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_5\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_5\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_5\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_5\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_5\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">def recv_callback(msg):\nprint 'Received: ' + msg.body\nchan.basic_consume(queue='po_box', no_ack=True,\ncallback=recv_callback, consumer_tag=\"testtag\")\nwhile True:\nchan.wait()\nchan.basic_cancel(\"testtag\")<\/pre><\/div><\/div>\n<p><strong>chan.wait()<\/strong> is looped infinitely, which is what causes the channel to wait for the next message notification from the queue. <strong>chan.basic_cancel()<\/strong> is how you unregister your message notification callback. The argument specifies the <strong>consumer_tag<\/strong> you specified in the original <strong>chan.basic_consume()<\/strong> registration (that\u2019s how it figures out which callback to unregister). In this case <strong>chan.basic_cancel()<\/strong> never gets called due to the infinite loop that precedes it\u2026but you need to know about it, so it\u2019s in the snippet.<\/p>\n<p>The one additional thing you should pay attention to in the consumer is the <strong>no_ack<\/strong> argument. It\u2019s accepted on both <strong>chan.basic_get()<\/strong> and <strong>chan.basic_consume()<\/strong> and defaults to false. When you grab a message off a queue, RabbitMQ needs you to explicitly acknowledge that you have it. If you don\u2019t, RabbitMQ will re-assign the message to another consumer on the queue after a timeout interval (or on disconnect by the consumer that initially received it  without ack\u2019ing it). If you set the <strong>no_ack<\/strong> argument to true, then <strong>py-amqplib<\/strong> will add a \u201cno_ack\u201d property to your AMQP request for the next message. That will instruct the AMQP server to not expect an acknowledgement for that get\/consume. However, in most cases, you probably want to send the acknowledgement yourself (e.g. you need to put the message contents in a database before you acknowledge). Acknowledgements are done by caling the <strong>chan.basic_ack()<\/strong> method, using the <strong>delivery_tag<\/strong> property of the message you\u2019re acknowledging as the argument (see the <strong>chan.basic_get()<\/strong> code snippet above for an example).<\/p>\n<p>That\u2019s all she wrote for the consumer. (Download: <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/code-samples\/amqp_consumer.py\">amqp_consumer.py<\/a>)<\/p>\n<p>But what good is a consumer, if nobody is sending it messages? So you need a publisher. The code below will publish a simple message to the \u201c<strong>sorting_room<\/strong>\u201d exchange and mark it with the routing key \u201c<strong>jason<\/strong>\u201d:<\/p>\n<div id=\"wpshdo_6\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_6\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_6\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_6\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_6\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_6\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_6\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">msg = amqp.Message(\"Test message!\")\nmsg.properties[\"delivery_mode\"] = 2\nchan.basic_publish(msg,exchange=\"sorting_room\",routing_key=\"jason\")<\/pre><\/div><\/div>\n<p>You may notice that we set the <strong>delivery_mode<\/strong> element of the message\u2019s properties to \u201c2\u201d. Since the queue and exchange were marked durable, this will ensure the message is sent as persistent (i.e. will survive a reboot of RabbitMQ while it is in transit to the consumer).<\/p>\n<p>The only other thing we need to do (and this needs to be done on both consumer and publisher apps), is close the channel and connection:<\/p>\n<div id=\"wpshdo_7\" class=\"wp-synhighlighter-outer\"><div id=\"wpshdt_7\" class=\"wp-synhighlighter-expanded\"><table border=\"0\" style=\"width: 100%;\"><tr><td align=\"left\" style=\"width: 80%;\"><a id=\"wpshat_7\" class=\"wp-synhighlighter-title\" href=\"#codesyntax_7\" title=\"Click to show\/hide code block\">Code block<\/a><\/td><td align=\"right\"><a href=\"#codesyntax_7\" title=\"Show code only\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/code.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"#codesyntax_7\" title=\"Print code\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/printer.png\" alt=\"image\" \/><\/a>\u00a0<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/About.html\" target=\"_blank\" title=\"Show plugin information\"><img border=\"0\" src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-content\/plugins\/wp-synhighlight\/themes\/default\/images\/info.gif\" alt=\"image\" \/><\/a>\u00a0<\/td><\/tr><\/table><\/div><div id=\"wpshdi_7\" class=\"wp-synhighlighter-inner\"><pre class=\"python\">chan.close()\nconn.close()<\/pre><\/div><\/div>\n<p>Pretty simple, no? (Download: <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/code-samples\/amqp_publisher.py\">amqp_publisher.py<\/a>)<\/p>\n<h2>Giving it a shot\u2026<\/h2>\n<p>Now we\u2019ve written our consumer and publisher, so let\u2019s give it a go. (This assumes you have RabbitMQ <a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.rabbitmq.com\/install.html\">installed<\/a> and running on <strong>localhost<\/strong>.)<\/p>\n<p>Open up the first terminal, and run <strong>python .\/amqp_consumer.py<\/strong> to get the consumer running and to create your queues, exchanges and bindings.<\/p>\n<p>Then run <strong>python .\/amqp_publisher.py \u201cAMQP rocks.\u201d<\/strong> in a second terminal. If everything went well, you should see your message printed by the consumer on the first terminal.<\/p>\n<h2>Taking it all in<\/h2>\n<p>I realize this has been a really fast run through AMQP\/RabbitMQ and using it from Python. Hopefully, it will fill in some of the holes of how all the concepts fit together and how they get used in a real Python program. If you find any errors in my write-up, I\u2019d very much appreciate it if you\u2019d please let me know (<a href=\"http:\/\/web.archive.org\/web\/20160323134044\/mailto:williamsjj@digitar.com\">williamsjj@digitar.com<\/a>). Similarly, I\u2019d be happy to answer any questions that I can. Next up\u2026.clustering! But I\u2019ve got to figure it out first. <img src=\"http:\/\/web.archive.org\/web\/20160323134044im_\/http:\/\/blogs.digitar.com\/jjww\/wp-includes\/images\/smilies\/icon_smile.gif\" alt=\":-)\" class=\"wp-smiley\" \/><\/p>\n<p>NB: Special thanks to Barry Pederson and Gordon Sims for correcting my understanding of no_ack\u2019s operation and for catching syntactically incorrect Python code I missed.<\/p>\n<p>NB: My knowledge on the subject was distilled from these sources, which are excellent further reading:<\/p>\n<ul><li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/blogs.digitar.com\/jjww\/2009\/01\/rabbits-and-warrens\/\">zeromq: Message-oriented Middleware Analysis<\/a><\/li>\n<li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/www.rabbitmq.com\/releases\/rabbitmq-dotnet-client\/v1.5.0\/rabbitmq-dotnet-client-1.5.0-user-guide.pdf\">RabbitMQ .NET Client Library User Guide <\/a><\/li>\n<li><a href=\"http:\/\/web.archive.org\/web\/20160323134044\/http:\/\/jira.amqp.org\/confluence\/download\/attachments\/720900\/amqp0-8.pdf?version=1\">Advanced Message Queuing Protocol: Protocol Specification Version 0-8<\/a><\/li>\n<\/ul>","created_at":"2019-01-29T14:14:09+0000","updated_at":"2019-01-29T14:14:14+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":17,"domain_name":"web.archive.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":799,"title":"Ask HN: Is it just me, or is CSS too damn hard?","url":"https:\/\/news.ycombinator.com\/item?id=19021719","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D19021719\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"19021719\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=19021719\" class=\"storylink\">Ask HN: Is it just me, or is CSS too damn hard?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        94 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=napsterbr\" class=\"hnuser\">napsterbr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=19021719\">4 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=19021719&amp;goto=item%3Fid%3D19021719\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Is%20it%20just%20me%2C%20or%20is%20CSS%20too%20damn%20hard%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Is%20it%20just%20me%2C%20or%20is%20CSS%20too%20damn%20hard%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=19021719&amp;auth=d7bc1189adf2ccf76846fc113c436350b160826b\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=19021719\">87\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Hey HN,<p>I'm a seasoned backend developer and systems administrator, with over 10 years of full-time programming experience. I've worked with anything from Assembly, C, Rust all the way to Python, Erlang, Elixir, PHP, Haskell, Lisp, Clojure.<\/p><p>When it comes to frontend, I have used React, Vue.js, ReasonML, Clojurescript and Elm extensively, and I know my way around web technologies in general. However, ever since I first started programming, I was NEVER able to get the gist behind CSS. I can't center my divs properly. I can't say \"hey, CSS, this is the parent div, and all child divs must obey its size\". I can't do anything basic on CSS without turning for help.<\/p><p>Incidentally, but very likely related, I always failed to have any glimpse into how design works. I have no idea which colors \"go\" with which ones, and pretty much all fonts look like the same for me.<\/p><p>I'd like to know the experience from fellow programmers. Is CSS an \"art\" thing? Is it common for other developers to have issues with it, or is it just me? Is there any material out there that made CSS \"click\" for you?<\/p><\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/www.ycombinator.com\/apply\/\">\n        Applications are open for YC Summer 2019\n      <\/a><\/div><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2019-01-29T02:50:18+0000","updated_at":"2019-01-29T02:50:18+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":798,"title":"Hacking Content Marketing: How\u00a0To\u00a0Buy Content For a Website?","url":"https:\/\/contentfly.co\/blog\/hacking-content-marketing-how-to-buy-content-for-a-website\/","content":"<figure class=\"post-full-image\"><section class=\"post-full-content\"><div class=\"post-content\"><p>For small businesses, organic creation of content can be difficult and time consuming.<br \/>It is possible for businesses to buy content for their websites.<\/p><h2 id=\"option-1-scripted\">Option 1 - Scripted<\/h2><p>One example of content, where businesses are able to buy quality work is<a href=\"https:\/\/www.scripted.com\/\"> Scripted<\/a><strong><em>.<\/em><\/strong><br \/>Scripted runs a vigorous process for its\u2019 writers, offering a number of content to purchase.<\/p><p><strong>Process<\/strong><\/p><ul><li>Sign up to a free, 1 month trial membership <\/li><li>Place an order with chosen content <\/li><li>Await the first draft<\/li><li>Confirm any revisions <\/li><\/ul><p>Scripted drafts can take five business days, with a maximum of seven days before the content is live<\/p><p><strong>Strengths<\/strong><\/p><ul><li>Free 1 month membership<\/li><li>Can communicate directly with writers<\/li><li>Ideas can be pitched FROM writers<\/li><\/ul><p><strong>Weaknesses<\/strong><\/p><ul><li>Membership ranges from $149 to $1299 monthly<\/li><li>Once a writer is approved, they remain in the network<\/li><li>For the monthly option of unlimited words, this will cost $1299 a month.<\/li><\/ul><p><br \/><strong>Pricing:<\/strong> Scripted charges a premium price for articles that are not quality-moderated post access.<\/p><p>A free monthly trial is accessible, however after one month \u2013 a monthly membership fee of $149 is billed \u2013 which covers 10,000 words (content costs are <strong>not included<\/strong> in plans - these will be charged as extra). Unlimited words will cost $1299. <\/p><p>Many small businesses do not have expendable funds upon start-up, which Scripted may price them out of. However, there is an alternative.<\/p><h2 id=\"option-2-contentfly\">Option 2 - ContentFly<\/h2><p><strong><a href=\"https:\/\/contentfly.co\/?utm_source=blog&amp;utm_campaign=hacking-marketing-buying-content\">ContentFly<\/a> <\/strong>charges a flat monthly rate, accessing large pool of writers \u2013 quality is of the highest importance. A free, 14 day trial is offered.<\/p><p><strong>Process<\/strong><\/p><ul><li>Sign up and post your first content brief through a robust dashboard.<\/li><li>The ContentFly algorithm will connect the brief and the best writer for the job. <\/li><li>The first draft can take 36 hours for a writer to submit. Afterwards, unlimited revisions are given.<\/li><\/ul><p><strong>Strengths<\/strong><\/p><ul><li>Contentfly is cheaper than Scripted \u2013 which gives businesses a lot of options to focus their spending into other channels.<\/li><li>Purely a <a href=\"https:\/\/contentfly.co\/?utm_source=blog&amp;utm_campaign=hacking-marketing-buying-content\">ghostwriting service<\/a>, giving the credit to the business.<\/li><li>Contentfly offers a minimum of 36 hours for content to be pitched, delivered and live (within weekdays).<\/li><li>Quality Control. Writers must meet a minimum of 4 stars overall score. With well-researched articles, businesses can be certain that ContentFly is the best they can use to buy unique articles online.<\/li><\/ul><p><strong>Weaknesses<\/strong><\/p><ul><li>One of the few weaknesses Contentfly has is the lack of contact between business and writer. However, multiple revisions can be requested by a business, and as the service is completely for ghostwriting \u2013 minimal contact is normal. <\/li><\/ul><p><strong>Pricing:<\/strong> Made for all budgets \u2013 and is a very affordable option to buy content for websites. <\/p><p>A flat rate of $250 a month \u2013 with access to a highly skilled writer pool, royalty free images, and up to 4000 words of content.<\/p><h2 id=\"option-3-hiring-freelancer-writers-directly\">Option 3 - hiring freelancer writers directly<\/h2><p><strong>Process<\/strong><\/p><ul><li>Contact made through platforms \u2013 job sites, forums, social media platforms<\/li><li>Idea pitched to writer for content<\/li><li>Proposed pricing option set<\/li><li>Content submitted\/revised if necessary <\/li><\/ul><p>Due to the varying number of channels in which contact can be made \u2013 the process can vary, however the skeleton remains the same.<\/p><p><strong>Strengths<\/strong><\/p><ul><li>Direct contact with writers<\/li><li>Content is tailored exactly to the needs of the business<\/li><li>Pricing can be negotiated and work ongoing<\/li><\/ul><p><strong>Weaknesses<\/strong><\/p><ul><li>Quality in work may be eccentric<\/li><li>Rights to all works is solely that of the writer<\/li><li>Pricing will vary based on experience\/specializations<\/li><\/ul><p><strong>Pricing:<\/strong> Freelance writer rates can vary \u2013 Charging 25c per word, where some will work with packages and\/or a higher word rate.<\/p><p>It is important, in this sense \u2013 for businesses to do their research to find a freelancer who can fit their budget.<\/p><\/div><\/section><\/figure>","created_at":"2019-01-28T19:36:50+0000","updated_at":"2019-01-28T19:36:50+0000","published_at":"2019-01-28T14:23:41+0000","published_by":["ContentFly"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"contentfly.co","preview_picture":"http:\/\/contentfly.co\/blog\/content\/images\/2019\/01\/andraz-lazic-1160627-unsplash.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":797,"title":"ContentFly - The \"Uber\" for Content","url":"https:\/\/contentfly.co\/","content":"<div class=\"section section-pad bg-rand\"><div class=\"container\"><div class=\"row align-items-center\"><div class=\"col-md-6\"><div class=\"img\"><img src=\"https:\/\/contentfly.co\/images\/delivery.png\" alt=\"\" \/><\/div><\/div><div class=\"col-md-5 offset-md-1\"><div class=\"text-block\"><h3>It's like Uber for Content<\/h3><p class=\"lead\">No more searching &amp; interviewing writers - <strong>we'll find the best writer<\/strong> (among thousands) for the job and take care of everything.<\/p><\/div><\/div><\/div><\/div><\/div>\n\t\n\t\n\t<div class=\"section section-pad bg-rand-alt\"><div class=\"container\"><div class=\"row align-items-center\"><div class=\"col-md-5 order-md-first order-last\"><div class=\"text-block text-right\"><h3>Premium quality at affordable rates<\/h3><p class=\"lead\">Our writers are native speakers, primarily based in North America. We leverage <strong>economies of scale<\/strong> to deliver industry-low rates.\t<\/p><\/div><\/div><div class=\"col-md-6  offset-md-1\"><div class=\"img circle-tip\"><img src=\"https:\/\/contentfly.co\/images\/outsource.png\" alt=\"\" \/><\/div><\/div><\/div><\/div><\/div>\n\t\n\t\n\t<div class=\"section section-pad bg-rand-prime\"><div class=\"circle-tip alt\"><div class=\"container\"><div class=\"row align-items-center\"><div class=\"col-lg-7 col-md-8 col-sm-8\"><div class=\"row box\"><div class=\"quote col-sm-6\"><div class=\"quote-in\"><p>ContentFly has proven to be our missing link for getting badass content out with ease and professionalism.<\/p><div class=\"quote-name\"><h6 class=\"quote-title no-mg\">Amy Randall<\/h6><p>LessDoing.com<\/p><\/div><\/div><\/div><div class=\"quote col-sm-6\"><div class=\"quote-in\"><p>ContentFly is the best writing service I\u2019ve ever used. We\u2019ve published in publications like the Wall Street Journal and Inc Magazine.<\/p><div class=\"quote-name\"><h6 class=\"quote-title no-mg\">Lauren Proctor<\/h6><p>Jobbatical.com<\/p><\/div><\/div><\/div><div class=\"quote col-sm-6\"><div class=\"quote-in\"><p>It's so unbelievably easy, and the quality is exceptional. I'm waiting for them to branch out to other areas.<\/p><div class=\"quote-name\"><h6 class=\"quote-title no-mg\">Melissa Sepp<\/h6><p>AmplerBikes.com<\/p><\/div><\/div><\/div><div class=\"quote col-sm-6\"><div class=\"quote-in\"><p>I thought \"Uber for Content\" was a bit cheeky, but that's exactly how it works. It's magic.<\/p><div class=\"quote-name\"><h6 class=\"quote-title no-mg\">Nelson Truong<\/h6><p>Sapient Nitro<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2019-01-28T19:35:46+0000","updated_at":"2019-01-28T19:35:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"contentfly.co","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":796,"title":"The Day I Fell in Love with Fuzzing \u00ab null program","url":"https:\/\/nullprogram.com\/blog\/2019\/01\/25\/","content":"<time datetime=\"2019-01-25\">\n    January 25, 2019\n  <\/time><p>nullprogram.com\/blog\/2019\/01\/25\/<\/p><p>In 2007 I wrote a pair of modding tools, <a href=\"https:\/\/github.com\/skeeto\/binitools\">binitools<\/a>, for a space\ntrading and combat simulation game named <a href=\"https:\/\/en.wikipedia.org\/wiki\/Freelancer_(video_game)\"><em>Freelancer<\/em><\/a>. The game\nstores its non-art assets in the format of \u201cbinary INI\u201d files, or \u201cBINI\u201d\nfiles. The motivation for the binary format over traditional INI files\nwas probably performance: it\u2019s faster to load and read these files than\nit is to parse arbitrary text in INI format.<\/p><p>Much of the in-game content can be changed simply by modifying these\nfiles \u2014 changing time names, editing commodity prices, tweaking ship\nstatistics, or even adding new ships to the game. The binary nature\nmakes them unsuitable to in-place modification, so the natural approach\nis to convert them to text INI files, make the desired modifications\nusing a text editor, then convert back to the BINI format and replace\nthe file in the game\u2019s installation.<\/p><p>I didn\u2019t reverse engineer the BINI format, nor was I the first person\nthe create tools to edit them. The existing tools weren\u2019t to my tastes,\nand I had my own vision for how they should work \u2014 an interface more\nclosely following <a href=\"http:\/\/www.catb.org\/esr\/writings\/taoup\/html\/\">the Unix tradition<\/a> despite the target being a\nWindows game.<\/p><p>When I got started, I had just learned how to use <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/yacc.html\">yacc<\/a> (really\n<a href=\"https:\/\/www.gnu.org\/software\/bison\/\">Bison<\/a>) and <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/lex.html\">lex<\/a> (really <a href=\"https:\/\/github.com\/westes\/flex\">flex<\/a>), as well as\nAutoconf, so I went all-in with these newly-discovered tools. It was\nexciting to try them out in a real-world situation, though I slavishly\naped the practices of other open source projects without really\nunderstanding why things were they way they were. Due to the use of\nyacc\/lex and the configure script build, compiling the project required\na full, Unix-like environment. This is all visible in <a href=\"https:\/\/github.com\/skeeto\/binitools\/tree\/original\">the original\nversion of the source<\/a>.<\/p><p>The project was moderately successful in two ways. First, I was able to\nuse the tools to modify the game. Second, other people were using the\ntools, since the binaries I built show up in various collections of\nFreelancer modding tools online.<\/p><h3 id=\"the-rewrite\">The Rewrite<\/h3><p>That\u2019s the way things were until mid-2018 when I revisited the project.\nEver look at your own old code and wonder what they heck you were\nthinking? My INI format was far more rigid and strict than necessary, I\nwas doing questionable things when writing out binary data, and the\nbuild wasn\u2019t even working correctly.<\/p><p>With an additional decade of experience under my belt, I knew I could do\n<em>way<\/em> better if I were to rewrite these tools today. So, over the course\nof a few days, I did, from scratch. That\u2019s what\u2019s visible in the master\nbranch today.<\/p><p><a href=\"https:\/\/nullprogram.com\/blog\/2017\/03\/30\/\">I like to keep things simple<\/a> which meant no more Autoconf, and\ninstead <a href=\"https:\/\/nullprogram.com\/blog\/2017\/08\/20\/\">a simple, portable Makefile<\/a>. No more yacc or lex, and\ninstead a hand-coded parser. Using only conforming, portable C. The\nresult was so simple that I can <a href=\"https:\/\/nullprogram.com\/blog\/2016\/06\/13\/\">build using Visual Studio<\/a> in a\nsingle, short command, so the Makefile isn\u2019t all that necessary. With\none small tweak (replace <code class=\"highlighter-rouge\">stdint.h<\/code> with a <code class=\"highlighter-rouge\">typedef<\/code>), I can even <a href=\"https:\/\/nullprogram.com\/blog\/2018\/04\/13\/\">build\nand run binitools in DOS<\/a>.<\/p><p>The new version is faster, leaner, cleaner, and simpler. It\u2019s far more\nflexible about its INI input, so its easier to use. But is it more\ncorrect?<\/p><h3 id=\"fuzzing\">Fuzzing<\/h3><p>I\u2019ve been interested in <a href=\"https:\/\/labs.mwrinfosecurity.com\/blog\/what-the-fuzz\/\">fuzzing<\/a> for years, especially\n<a href=\"http:\/\/lcamtuf.coredump.cx\/afl\/\">american fuzzy lop<\/a>, or <em>afl<\/em>. However, I wasn\u2019t having success\nwith it. I\u2019d fuzz some of the tools I use regularly, and it wouldn\u2019t\nfind anything of note, at least not before I gave up. I fuzzed <a href=\"https:\/\/github.com\/skeeto\/pdjson\">my\nJSON library<\/a>, and somehow it turned up nothing. Surely my\nJSON parser couldn\u2019t be <em>that<\/em> robust already, could it? Fuzzing just\nwasn\u2019t accomplishing anything for me. (As it turns out, my JSON\nlibrary <em>is<\/em> quite robust, thanks in large part to various\ncontributors!)<\/p><p>So I\u2019ve got this relatively new INI parser, and while it can\nsuccessfully parse and correctly re-assemble the game\u2019s original set of\nBINI files, it hasn\u2019t <em>really<\/em> been exercised that much. Surely there\u2019s\nsomething in here for a fuzzer to find. Plus I don\u2019t even have to write\na line of code in order to run afl against it. The tools already read\nfrom standard input by default, which is perfect.<\/p><p>Assuming you\u2019ve got the necessary tools installed (make, gcc, afl),\nhere\u2019s how easy it is to start fuzzing binitools:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ make CC=afl-gcc\n$ mkdir in out\n$ echo '[x]' &gt; in\/empty\n$ afl-fuzz -i in -o out -- .\/bini\n<\/pre><\/div><\/div><p>The <code class=\"highlighter-rouge\">bini<\/code> utility takes INI as input and produces BINI as output, so\nit\u2019s far more interesting to fuzz than its inverse, <code class=\"highlighter-rouge\">unbini<\/code>. Since\n<code class=\"highlighter-rouge\">unbini<\/code> parses relatively simple binary data, there are (probably) no\nbugs for the fuzzer to find. I did try anyway just in case.<\/p><p><img src=\"https:\/\/nullprogram.com\/img\/screenshot\/afl.png\" alt=\"\" \/><\/p><p>In my example above, I swapped out the default compiler for afl\u2019s GCC\nwrapper (<code class=\"highlighter-rouge\">CC=afl-gcc<\/code>). It calls GCC in the background, but in doing so\nadds its own instrumentation to the binary. When fuzzing, <code class=\"highlighter-rouge\">afl-fuzz<\/code>\nuses that instrumentation to monitor the program\u2019s execution path. The\n<a href=\"http:\/\/lcamtuf.coredump.cx\/afl\/technical_details.txt\">afl whitepaper<\/a> explains the technical details.<\/p><p>I also created input and output directories, placing a minimal, working\nexample into the input directory, which gives afl a starting point. As\nafl runs, it mutates a queue of inputs and observes the changes on the\nprogram\u2019s execution. The output directory contains the results and, more\nimportantly, a corpus of inputs that cause unique execution paths. In\nother words, the fuzzer output will be lots of inputs that exercise many\ndifferent edge cases.<\/p><p>The most exciting and dreaded result is a crash. The first time I ran it\nagainst binitools, <code class=\"highlighter-rouge\">bini<\/code> had <em>many<\/em> such crashes. Within minutes, afl\nwas finding a number of subtle and interesting bugs in my program, which\nwas <em>incredibly<\/em> useful. It even discovered an unlikely <a href=\"https:\/\/github.com\/skeeto\/binitools\/commit\/b695aec7d0021299cbd83c8c6983055f16d11507\">stale pointer\nbug<\/a> by exercising different orderings for various memory\nallocations. This particular bug was the turning point that made me\nrealize the value of fuzzing.<\/p><p>Not all the bugs it found led to crashes. I also combed through the\noutputs to see what sorts of inputs were succeeding, what was failing,\nand observe how my program handled various edge cases. It was rejecting\nsome inputs I thought should be valid, accepting some I thought should\nbe invalid, and interpreting some in ways I hadn\u2019t intended. So even\nafter I fixed the crashing inputs, I still made tweaks to the parser to\nfix each of these troublesome inputs.<\/p><h3 id=\"building-a-test-suite\">Building a test suite<\/h3><p>Once I combed out all the fuzzer-discovered bugs, and I agreed with the\nparser on how all the various edge cases should be handled, I turned the\nfuzzer\u2019s corpus into a test suite \u2014 though not directly.<\/p><p>I had run the fuzzer in parallel \u2014 a process that is explained in the\nafl documentation \u2014 so I had lots of redundant inputs. By redundant I\nmean that the inputs are different but have the same execution path.\nFortunately afl has a tool to deal with this: <code class=\"highlighter-rouge\">afl-cmin<\/code>, the corpus\nminimization tool. It eliminates all the redundant inputs.<\/p><p>Second, many of these inputs were longer than necessary in order to\ninvoke their unique execution path. There\u2019s <code class=\"highlighter-rouge\">afl-tmin<\/code>, the test case\nminimizer, which I used to further shrink my test corpus.<\/p><p>I sorted the valid from invalid inputs and checked them into the\nrepository. Have a look at all the wacky inputs <a href=\"https:\/\/lcamtuf.blogspot.com\/2014\/11\/pulling-jpegs-out-of-thin-air.html\">invented by<\/a> the\nfuzzer starting from my single, minimal input:<\/p><ul><li><a href=\"https:\/\/github.com\/skeeto\/binitools\/tree\/master\/tests\/valid\">valid inputs<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/skeeto\/binitools\/tree\/master\/tests\/invalid\">invalid inputs<\/a><\/li>\n<\/ul><p>This essentially locks down the parser, and the test suite ensures a\nparticular build behaves in a <em>very<\/em> specific way. This is most useful\nfor ensuring that builds on other platforms and by other compilers are\nindeed behaving identically with respect to their outputs. My test suite\neven revealed a bug in diet libc, as binitools doesn\u2019t pass the tests\nwhen linked against it. If I were to make non-trivial changes to the\nparser, I\u2019d essentially need to scrap the current test suite and start\nover, having afl generate an entire new corpus for the new parser.<\/p><p>Fuzzing has certainly proven itself to be a powerful technique. It found\na number of bugs that I likely wouldn\u2019t have otherwise discovered on my\nown. I\u2019ve since gotten more savvy on its use and have used it on other\nsoftware \u2014 not just software I\u2019ve written myself \u2014 and discovered more\nbugs. It\u2019s got a permanent slot on my software developer toolbelt.<\/p>","created_at":"2019-01-28T19:03:19+0000","updated_at":"2019-01-28T19:03:19+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"nullprogram.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":795,"title":"A Standard and Clean Series A Term Sheet","url":"https:\/\/blog.ycombinator.com\/a-standard-and-clean-series-a-term-sheet\/","content":"<p>While working with companies in <a href=\"https:\/\/blog.ycombinator.com\/category\/series-a\/\">YC\u2019s Series A program<\/a>, we\u2019ve noticed a common problem: founders don\u2019t know what \u201cgood\u201d looks like in a term sheet. This makes sense, because it is often, literally, the first time in their careers that they\u2019ve seen one. This puts founders at a significant disadvantage because VCs see term sheets all the time and know what to expect. Because we\u2019ve invested in so many founders over the years and have seen hundreds of Series A term sheets, we know what \u201cgood\u201d looks like. We work with our founders to understand where terms diverge from \u201cgood\u201d, what they can do about that divergence, and when and how it makes sense to negotiate.<\/p><p>Below is what a Series A term sheet looks like with standard and clean terms from a good Silicon Valley VC. Bracketed items (besides the names of the company and lead investor) are always or frequently negotiated. Items not in brackets are sometimes negotiated, but this has more to do with the idiosyncratic features of the company or the situation, and generally aren\u2019t terms that parties intend to heavily bargain over during the negotiation.<\/p><p>One of the critical things you\u2019ll notice is that we didn\u2019t put in standard pricing. While the lead in a Series A round generally wants 20% of the company, pricing can flex up and down depending on the leverage held by each side. We think price is an important term, but too specific to each raise to try to create a standard. We\u2019re more concerned with terms around control and structure that are less familiar to founders, and therefore more prone to cause confusion and trouble.<\/p><p>Note: this term sheet doesn\u2019t belong to any particular VC \u2014 we drafted it \u2014 but it does substantively reflect what we see most often. Founders with a lot of negotiating leverage can sometimes do better, and the converse is true too.<\/p><p>You can also download the <a href=\"https:\/\/www.ycombinator.com\/docs\/Standard_Series_A_TS.docx\">Word version of the doc here<\/a>.<\/p><p><a href=\"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-1.png\"><img src=\"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-1-790x1024.png\" alt=\"\" width=\"780\" height=\"1011\" class=\"aligncenter size-large wp-image-1103405\" srcset=\"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-1-790x1024.png 790w, https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-1-231x300.png 231w, https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-1-768x996.png 768w, https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-1.png 1234w\" \/><\/a><\/p><p><a href=\"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-2.png\"><img src=\"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-2-1024x724.png\" alt=\"\" width=\"780\" height=\"551\" class=\"aligncenter size-large wp-image-1103406\" srcset=\"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-2-1024x724.png 1024w, https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-2-300x212.png 300w, https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-2-768x543.png 768w, https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/Series-A-2.png 1582w\" \/><\/a><\/p><p>It may be surprising to see everything covered in a single page.<sup id=\"footnoteid1\"><a href=\"#footnote1\">1<\/a><\/sup> This wasn\u2019t always the case, but became common over the last decade as some investors decided to make their term sheets more user friendly by shortening the legalese as if to say, \u201cWe aren\u2019t going to get bogged down in the minutiae. We\u2019re going to make this easy, friendly, standard and fast.\u201d<\/p><p>This leads us to the most important thing to understand about the term sheet: it\u2019s another way in which your Series A investor might be telling you something. A contract allocates risks between the parties, so the terms the investor insists on can sometimes say a lot about the investor\u2019s perceived risks. These perceived risks show up in a couple of ways.<\/p><p>The first way relates to control terms. We don\u2019t mean the set of investor vetoes in the \u201cVoting Rights\u201d section, which are pretty standard fare,<sup id=\"footnoteid2\"><a href=\"#footnote2\">2<\/a><\/sup> but rather issues of board composition and the investor\u2019s ability to block or dictate operational decisions made by the board. The board structure in this term sheet is founder-friendly because the founders retain board control 2-1.<sup id=\"footnoteid3\"><a href=\"#footnote3\">3<\/a><\/sup> The way in which founders most often lose control at the Series A is with a 2-2-1 board structure, i.e. 2 founders, 2 investors and an independent board member. The loss of board control is most significant because it means the founders can be fired from their own company.<sup id=\"footnoteid4\"><a href=\"#footnote4\">4<\/a><\/sup> Another way in which founders lose some control is a term that doesn\u2019t appear in the standard example above, which is a separate provision that says the investor director\u2019s approval is required for operational decisions like setting the annual budget, hiring\/firing executives, pivoting the business, adding new lines of business, etc. When boards are set up to take power away from founders, the investor\u2019s outward justification will frequently be reasons of governance or accountability. But the more power that\u2019s taken away, the more it\u2019s undeniable that the investor is attempting to structure away a perceived risk. So when an investor says that they\u2019re committed to partnering with you for the long-term \u2013 or that they\u2019re betting everything on you \u2013 but then tells you something else with the terms that they insist on, believe the terms.<\/p><p>The other way perceived risks manifest is if a term sheet includes non-standard or \u201cdirty\u201d economic terms. Here, the term sheet example is instructive not for what it contains but what it <em>doesn\u2019t<\/em>. Examples of such terms would be:<\/p><ul><li>Liquidation preference greater than 1x \u2014 the investor gets back more than its invested capital first.<\/li>\n<li>Participating preferred \u2014 the investor double-dips by getting its money back plus its pro rata portion of exit proceeds, rather than choosing between the two.<\/li>\n<li>Cumulative dividends \u2014 the investor compounds its liquidation preference every year by X%, which increases the economic hurdle that has to be cleared before founders and employees see any value.<\/li>\n<li>Warrant coverage \u2014 the investor gets extra fully diluted ownership without paying for it at the agreed upon valuation.<\/li>\n<\/ul><p>These are all ways of adding structure to reduce typical venture risk, either directly by boosting the investor\u2019s downside economics, or indirectly by juicing the upside outcomes. The investor is essentially saying, \u201cI\u2019m sort of afraid of losing my money.\u201d It can also foreshadow how they might behave when things aren\u2019t going well, such as pushing you to sell when you don\u2019t want to, or dial back risk when it\u2019s important to take it. Good investors would rather address economic risks by negotiating valuation, and are otherwise happy to give standard terms because they know that the real money in venture is not made with structure, but by building long-term value, which they are confident in their ability to help you do.<\/p><p>The last thing to remember is that your Series A documents are a foundation and precedent for the terms of future rounds. Good foundations make the next term sheet and financing round fast and simple, as future investors just step into the same straightforward terms. Doing the opposite complicates future fundraises, such as future investors asking for the same structure-heavy terms, existing investors refusing to drop terms that subsequent investors want removed as a precondition of investing, etc. Unwinding bad terms is difficult, and oftentimes impossible.<\/p><p>That said, the point is to get a clean deal, not to cycle a lot to get the perfect deal. No one ever built an enduring company just by winning their Series A negotiation. Also, even if you can\u2019t get everything right or the way you want it, you always have the power to execute. If you do that, the value you build can outrun suboptimal terms or establish leverage to renegotiate later. So don\u2019t lose sight of the ultimate goal: closing fast and getting back to work.<\/p><p><strong>Notes<\/strong><\/p><p><b id=\"footnote1\">1.<\/b> Some great investors still send longer term sheets, but this has more to do with their preference for going a bit deeper into the details at this stage, rather than deferring this until the definitive documents. The definitive documents are derived from the term sheet and are the much longer (100+ pages) binding contracts that everyone signs and closes on. It\u2019s common to negotiate a few additional points at this stage, though deviation from anything explicitly addressed in the term sheet is definitely re-trading. Also, in a few places, this term sheet refers to certain terms as being \u201cstandard.\u201d That may seem vague and circular, but term sheets frequently do describe certain terms that way. What that really means is that there\u2019s an accepted practice of what appears in the docs for these terms among the lawyers who specialize in startups and venture deals, so make sure your lawyer (and the investor\u2019s lawyer) fit that description.<a href=\"#footnoteid1\">\u21a9<\/a><\/p><p><b id=\"footnote2\">2.<\/b> The two most impactful investor vetoes in this section are the veto on a financing, which is covered by clauses (ii) and (iii), and the veto on a sale of the company, which is in clause (vii). We point these out because the concrete implications of these clauses aren\u2019t facially obvious, and because most term sheets use similar technical jargon for these vetoes.<a href=\"#footnoteid2\">\u21a9<\/a><\/p><p><b id=\"footnote3\">3.<\/b> The founders implicitly control those 2 seats because they\u2019re designated by a majority of common, and founders generally control a majority of common for a long time. In even more founder-friendly term sheets, those 2 seats may be designated by the founders themselves (as individuals).<a href=\"#footnoteid3\">\u21a9<\/a><\/p><p><b id=\"footnote4\">4.<\/b> Whether being fired from the company as an employee also triggers the removal of the founder from the board is a separate question and depends on what was negotiated in the financing documents. Sometimes a founder\u2019s right to vote her shares to appoint a director will be conditioned on the founder being currently employed by the company. Whenever conditions are attached to your rights to vote on anything, make sure to ask your lawyer to walk you through the various scenarios in which those conditions matter and how they can hurt you.<a href=\"#footnoteid4\">\u21a9<\/a><\/p><p><em>This is not legal advice.<\/em><\/p><p><em>Thanks to Carolynn Levy, Jon Levy, and Nicole Cadman for their comments on this.<\/em><\/p>","created_at":"2019-01-28T18:41:04+0000","updated_at":"2019-01-28T18:41:04+0000","published_by":["Jason Kwon"],"annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"blog.ycombinator.com","preview_picture":"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2019\/01\/A-Standard-and-Clean-Series-A-Term-Sheet.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["gamedev","physics","simulation"],"is_public":false,"id":794,"title":"liquidfun","url":"http:\/\/google.github.io\/liquidfun\/","created_at":"2019-01-28T17:16:04+0000","updated_at":"2019-01-28T17:16:11+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"google.github.io"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":793,"title":"The worst career advice I ever received","url":"https:\/\/hackernoon.com\/the-worst-career-advice-i-ever-received-54aaf2a50c93?gi=5057607421ee","content":"<p id=\"12f4\" class=\"graf graf--p graf--leading\">The year was 2011, I was a line-level engineering manager and had been working in the same product group at Microsoft for 9 years since the start of my career. I had a friend, Paul, who\u2019d also been working with me as a peer for several years.<\/p><p id=\"b196\" class=\"graf graf--p graf-after--p\">Things weren\u2019t looking good in our organization. Almost every week someone left, product strategy changed bi weekly, and no one was certain which way was up or down. Yet we floated on. I was tasked with several questionable projects and generally kept myself and my reports shielded, but underneath a thin veneer I was afraid. Later, I\u2019d learn to understand what was going on (upper-management, politics, etc). But at the time I was an uninformed junior manager and it simply felt like chaos.<\/p><p id=\"c43b\" class=\"graf graf--p graf-after--p\">For the first time in a long time, I began exploring employment options elsewhere. I managed to line up a risky but interesting management role on a new project team in a totally different part of Microsoft. Having worked in one team for near a decade I wasn\u2019t sure if I should make the switch. I knew all the ropes, my co-workers felt like family, and my manager loved me.<\/p><p id=\"3a4b\" class=\"graf graf--p graf-after--p\">I went to Paul and asked for his advice. He told me, \u201c<strong class=\"markup--strong markup--p-strong\">Stay! When all of these other people leave there will be a power vacuum, open positions, and you\u2019ll be able get promoted quickly.<\/strong>\u201d<\/p><p id=\"37f8\" class=\"graf graf--p graf-after--p\">His opinion felt logical and staying would certainly be less risky. I didn\u2019t immediately reject the thought. At the same time, was that really how I wanted to rise up the ranks? After grappling with the decision for a week, I took the leap of faith and left the team I\u2019d been with for almost a decade.<\/p><p id=\"d927\" class=\"graf graf--p graf-after--p\">The new project was immensely difficult, and I worked as hard as ever to grow the team and bring the product MVP to life. Within 6 months I was promoted to a position that gave me 30 engineers to manage. The product caught fire and 2 years later I was leading a team of 150. A year later I switched roles began work as an executive leader.<\/p><p id=\"c01d\" class=\"graf graf--p graf-after--p\">Paul stayed behind. To this day he is still in that team in the exact same role. We do keep in touch and whenever we talk the conversation eventually makes its way around to him wanting to switch and do something new. Paul doesn\u2019t like the product he works on or the teammates he works with all that much. \u201cThey aren\u2019t innovative.\u201d He doesn\u2019t have other obligations that would hold him back. I ask him why he stays, and he talks about needing the right opportunity. \u201cYou got lucky when you left,\u201d he says. So we talk a lot about finding opportunity, the future, and risk, and luck. I love those talks and always come away feeling personally inspired.<\/p><p id=\"4b36\" class=\"graf graf--p graf-after--p graf--trailing\">Yet, I think Paul will stay a while longer. And I sometimes wonder what my career would have been like if I\u2019d taken his advice. It\u2019s better to look forward.<\/p>","created_at":"2019-01-28T00:16:39+0000","updated_at":"2019-01-28T00:16:39+0000","published_at":"2019-01-20T02:50:29+0000","published_by":["Nick Caldwell"],"annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"hackernoon.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*Me436pcm4idloBhP633EWQ.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":791,"title":"ProseMirror","url":"http:\/\/prosemirror.net\/","content":"<div id=\"content\"><p>Like this one!<\/p><p>Try it out by typing in here or see more <a href=\"http:\/\/prosemirror.net\/examples\">examples<\/a>.<\/p><\/div><p>An ideal content editor produces structured, semantically\nmeaningful documents, but does so in a way that is easy for users to\nunderstand. ProseMirror tries to bridge the gap\nbetween <a href=\"http:\/\/commonmark.org\/\">Markdown<\/a> text editing and\nclassical <a href=\"https:\/\/en.wikipedia.org\/wiki\/WYSIWYG\">WYSIWYG<\/a>\neditors.<\/p><p>It does this by implementing a WYSIWYG-style editing interface for\ndocuments more constrained and structured than plain HTML. You can\ncustomize the shape and structure of the documents your editor\ncreates, and tailor them to your application's needs.<\/p><ul class=\"grid-list\"><li>\n    <h3>Collaborative editing<\/h3>\n    <p>ProseMirror has built-in, ground-up, rock solid support for\n    collaborative editing, where multiple people work on the same\n    document in real time.<\/p>\n  <\/li>\n  <li>\n    <h3>Extensible schemas<\/h3>\n    <p>Document schemas allow editing documents with a custom\n    structure without writing your own editor from scratch.<\/p>\n  <\/li>\n  <li>\n    <h3>Modular<\/h3>\n    <p>A modular architecture makes sure you only load the code you\n    need, and can replace parts of the system as needed.<\/p>\n  <\/li>\n  <li>\n    <h3>Pluggable<\/h3>\n    <p>A plugin system allows you to easily enable additional\n    functionality, and package your own extensions in a convenient\n    format.<\/p>\n  <\/li>\n  <li>\n    <h3>Functional<\/h3>\n    <p>A largely functional and immutable architecture makes it easy\n    to integrate ProseMirror into modern web apps, and confidently\n    implement complex behavior.<\/p>\n  <\/li>\n  <li>\n    <h3>Unopinionated<\/h3>\n    <p>The core library is small and generic, providing a base onto\n    which very different types of editors can be built.<\/p>\n  <\/li>\n<\/ul><p>ProseMirror\nis <a href=\"https:\/\/github.com\/ProseMirror\/prosemirror\/blob\/master\/LICENSE\">open\nsource<\/a>, and you are legally free to use it commercially. Yet,\nwriting, maintaining, supporting, and setting up infrastructure for\nsuch a project takes a lot of work and energy. Therefore...<\/p><p class=\"callout\">If you are using ProseMirror to make profit, there is\na <u>social<\/u> expectation that you help\nfund its maintenance. \u00a0\n<a class=\"fund-button\" href=\"http:\/\/marijnhaverbeke.nl\/fund\/\">Start here.<\/a>\n<\/p><p>ProseMirror is being developed\non <a href=\"https:\/\/github.com\/prosemirror\/\">GitHub<\/a>. Contributions\nare <a href=\"https:\/\/github.com\/ProseMirror\/prosemirror\/blob\/master\/CONTRIBUTING.md\">welcome<\/a>.<\/p><p>The easiest way to install ProseMirror is\nwith <a href=\"https:\/\/www.npmjs.com\/search?q=prosemirror-\">npm<\/a>.\nTake a look at the <a href=\"http:\/\/prosemirror.net\/examples\/basic\/\">examples<\/a> to get\nstarted.<\/p><p>The library works on recent versions of Firefox, Chrome, Safari,\nand Edge, as well as on Internet Explorer 11.<\/p><p>Discussing the project, or asking questions, is best done on\nthe <a href=\"https:\/\/discuss.prosemirror.net\/\">forum<\/a>.\nBugs should be reported through\nthe <a href=\"https:\/\/github.com\/ProseMirror\/prosemirror\/issues\">issue\ntracker<\/a>. We aim to be an inclusive, welcoming community. To make\nthat explicit, we have\na <a href=\"http:\/\/contributor-covenant.org\/version\/1\/1\/0\/\">code of\nconduct<\/a> that applies to communication around the project.<\/p><p>Initial development of the software\nwas <a href=\"https:\/\/www.indiegogo.com\/projects\/prosemirror\/\">crowd-funded<\/a>\nby these 414 excellent <a href=\"http:\/\/prosemirror.net\/backers.html\">individuals and\norganizations<\/a>.<\/p><p><img src=\"http:\/\/prosemirror.net\/logos\/shuttleworth.jpg\" alt=\"Shuttleworth Funded\" \/>In June 2018 the Shuttleworth Foundation supported ProseMirror maintenance with a <a href=\"https:\/\/www.shuttleworthfoundation.org\/fellows\/flash-grants\/\">flash grant<\/a>.<\/p>","created_at":"2019-01-28T00:16:37+0000","updated_at":"2019-01-28T00:16:37+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"prosemirror.net","preview_picture":"http:\/\/prosemirror.net\/img\/picture.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":792,"title":"Object Capabilities for Linked Data v0.3","url":"https:\/\/w3c-ccg.github.io\/ocap-ld\/","content":"Object Capabilities for Linked Data v0.3\n    <section id=\"abstract\"><p>\n        Object Capabilities for Linked Data (<code>OCAP-LD<\/code> for short)\n        provides a\n        secure way for linked data systems to grant and express authority\n        utilizing the\n        <a href=\"https:\/\/en.wikipedia.org\/wiki\/Object-capability_model\">\n          object capability model<\/a>.\n        Capabilities are represented as linked data objects which are signed with\n        <a href=\"https:\/\/w3c-dvcg.github.io\/ld-signatures\/\">\n          Linked Data Signatures<\/a>.\n        OCAP-LD supports delegating authority to other entities\n        on the network by chaining together capability documents.\n        \"Caveats\" may be attached to capability documents which may be used to\n        restrict the scope of their use, for example to restrict the actions\n        which may be used or providing a mechanism by which the capability may\n        be later revoked.\n      <\/p><\/section><section id=\"sotd\"><p>\n        This specification is currently being drafted to be considered as a\n        work item for the\n        <a href=\"https:\/\/w3c-ccg.github.io\/\">Community Credentials Group<\/a>.\n      <\/p><\/section><section id=\"introduction\"><h2>Introduction<\/h2><p><i>\n        This document is being based off of a\n        <a href=\"https:\/\/github.com\/WebOfTrustInfo\/rebooting-the-web-of-trust-fall2017\/blob\/master\/draft-documents\/lds-ocap\/lds-ocap.md\">paper from Rebooting Web of Trust<\/a>.\n        In the early stages of this document some ideas may be better\n        described in that paper than yet in this specification.\n      <\/i><\/p><p>\n        This document does not cover a specific method for delivering OCAP-LD\n        invocations to an object, though something like the inbox property and\n        associated delivery mechanisms from Linked Data Notifications and\n        ActivityPub is one possible system.\n        However there is nothing about OCAP-LD that is specific to HTTP; for\n        example, both capabilities and invocations could be stored on a\n        blockchain or a distributed hash table.\n      <\/p><section><h2>Capabilities vs. Access Control Lists<\/h2><p>\n          One of the first questions that is usually asked about using\n          Object Capabilities for authorization is: \"How are Capabilities\n          different from Access Control Lists?\". Fundamentally, Access Control\n          Lists are about <em>authority by identity<\/em> whereas\n          Object Capabilities are about <em>authority by possession<\/em>.\n        <\/p><p>\n          Authority by identity is the process of giving access to a resource\n          to a specific entity based on their identity. These processes\n          typically ask the question: \"Who are you?\"\n        <\/p><p>\n          Authority by possession is the process of giving access to a resource\n          to any entity that possesses something, like a key. These processes\n          typically ask the question: \"Do you have a key that fits this lock?\"\n        <\/p><p>\n          This document doesn't explain why Access Control Lists lead to\n          a variety of security issues or why Object Capabilities provide\n          stronger security guarantees. For those that would like to\n          learn more about these topics,\n          <a href=\"http:\/\/srl.cs.jhu.edu\/pubs\/SRL2003-02.pdf\">Capability Myths Demolished<\/a>\n          and\n          <a href=\"http:\/\/waterken.sourceforge.net\/aclsdont\/current.pdf\">ACLs Don't<\/a>\n          provide a deeper exploration into the benefits of Object Capabilities.\n        <\/p><\/section><section id=\"ocap-by-example\"><h2>OCAP-LD by Example<\/h2><p>\n          Much of modern computing security infrastructure relies on \"who is\n          doing something\" using access control lists.\n          For example, our friend Alyssa P. Hacker has a car, and she would\n          like to drive it\n          In the access control list world, the car itself may scan Alyssa's\n          face, determine that Alyssa is the driver, and say \"Welcome Alyssa,\n          you are now free to drive.\"\n          Talking cars are appealing and fun, but Alyssa may run into\n          challenges when she would like to allow others to drive her car.\n        <\/p><p>\n          Fortunately, there is another paradigm that is even more familiar to\n          the car scenario, but less familiar in the context of computing,\n          despite providing some superior properties and a greater degree\n          of safety: object capabilities.\n          Object capabilities focus not on \"who\" is performing an action, but\n          upon \"what\" source of authority permits an action to occur.\n          As it turns out, Alyssa does not have a talking and face-scanning\n          car, she has a car that accepts a car key.\n          The car may have no memory of Alyssa whatsoever, but as long as\n          Alyssa holds the kind of key that enables her to drive the car, she\n          can drive the car.\n          Capabilities are very similar to this car metaphor, and we can encode\n          this same idea in OCAP-LD.\n          The following document delegates authority from the car (who always\n          has authority over itself) to Alyssa so that she may drive:\n        <\/p><pre class=\"example highlight javascript\">\n    {\"@context\": [\"https:\/\/example.org\/ocap\/v1\",\n                  \"https:\/\/autopower.example\/\"],\n     \"id\": \"https:\/\/whatacar.example\/a-fancy-car\/proc\/7a397d7b\",\n     \/\/ Since this is the first delegated capability, the parentCapability\n     \/\/ points to the target this capability will operate against\n     \/\/ (in this case, Alyssa's Car)\n     \"parentCapability\": \"https:\/\/whatacar.example\/a-fancy-car\",\n     \/\/ We are granting authority specifically to one of Alyssa's\n     \/\/ cryptographic keys (not to be confused with the car\n     \/\/ key metaphor!)\n     \"invoker\": \"https:\/\/social.example\/alyssa\/#key-for-car\",\n     \/\/ Finally we sign this object with cryptographic material from\n     \/\/ Alyssa's Car's capabilityDelegation field, and using the\n     \/\/ capabilityDelegation proofPurpose.\n     \"proof\": {\n        \"type\": \"RsaSignature2016\",\n        \"proofPurpose\": \"capabilityDelegation\",\n        \"created\": \"2016-02-08T16:02:20Z\",\n        \"creator\": \"https:\/\/whatacar.example\/a-fancy-car#key-1\",\n        \"signatureValue\": \"IOmA4R7TfhkYTYW8...CBMq2\/gi25s=\"}}\n        <\/pre><p>\n          Turning on this car and driving into the sunset is as easy as\n          invoking the capability to do so:\n        <\/p><pre class=\"example highlight javascript\">\n    {\"@context\": [\"https:\/\/example.org\/ocap\/v1\",\n                  \"https:\/\/autopower.example\/\"],\n     \"id\": \"urn:uuid:ad86cb2c-e9db-434a-beae-71b82120a8a4\",\n     \"action\": \"Drive\",\n     \"proof\": {\n        \"type\": \"RsaSignature2016\",\n        \/\/ A linked data document can be an invocation if it has a\n        \/\/ proofPurpose of capabilityInvocation and links to the capability\n        \/\/ chain it is invoking\n        \"proofPurpose\": \"capabilityInvocation\",\n        \"capability\": \"https:\/\/whatacar.example\/a-fancy-car\/proc\/7a397d7b\",\n        \"created\": \"2016-02-08T17:13:48Z\",\n        \"creator\": \"https:\/\/social.example\/alyssa\/#key-for-car\",\n        \"signatureValue\": \"...\"}}\n        <\/pre><p>\n          Alyssa lives with her roommate and long-time friend Ben Bitdiddle.\n          While they're living together, she'd like that he be able to drive\n          her car too.\n          Capabilities support delegation, so she could delegate a new\n          capability to Ben giving him full authority, not unlike going to the\n          hardware store and having them copy the car key.\n          But Alyssa knows that she and Ben will probably not be roommates\n          forever, and she feels no need for him to be able to drive her car\n          once they are not, so she would like to have the option to revoke\n          his authority.\n          In the world of object capabilities, this is entirely possible by\n          adding a \"caveat\" that permits future revocation.\n          We can imagine this to be like a new car key that has a remote\n          destruction mechanism: Alyssa can press a button that she has,\n          a wire burns out inside the car key, and the car key will not be usable\n          anymore.\n          Alyssa generates a new capability document that points at the\n          capability she has, adding the caveat:\n        <\/p><pre class=\"example highlight javascript\">\n    {\"@context\": [\"https:\/\/example.org\/ocap\/v1\",\n                  \"https:\/\/autopower.example\/\"],\n     \"id\": \"https:\/\/social.example\/alyssa\/caps#79795d78\",\n     \/\/ Pointing up the chain at the capability from which Alyssa was\n     \/\/ initially gained authority\n     \"parentCapability\": \"https:\/\/whatacar.example\/a-fancy-car\/proc\/7a397d7b\",\n     \/\/ Alyssa grants authority specifically to one of Ben's\n     \/\/ cryptographic keys\n     \"invoker\": \"https:\/\/chatty.example\/ben\/#key-33\",\n     \/\/ Alyssa adds a caveat: Ben can drive her car, unless she flips\n     \/\/ the bit at this url\n     \"caveat\": [\n       {\"type\": \"ValidWhileTrue\",\n        \"uri\": \"https:\/\/social.example\/alyssa\/ben-can-still-drive\"}],\n     \/\/ Finally Alyssa signs this object with the key she was granted\n     \/\/ authority with\n     \"proof\": {\n        \"type\": \"RsaSignature2016\",\n        \"proofPurpose\": \"capabilityDelegation\",\n        \"created\": \"2017-03-28T06:01:25Z\",\n        \"creator\": \"https:\/\/social.example\/alyssa\/#key-for-car\",\n        \"signatureValue\": \"...\"}}\n        <\/pre><p>\n          Alyssa hands the car key to Ben, and Ben marks the car key in such\n          a way that reminds him that it's specifically to drive Alyssa's car.\n          Several months pass and Ben receives a message from Alyssa asking if\n          he might enjoy joining her at an award ceremony her university is\n          holding.\n          Alyssa is already at the university and is getting a ride to the\n          venue, so she suggests Ben drive her car and meet her there.\n          Ben invokes his capability, the car performs all relevant checks on\n          the capability (including that the proof\/signature is valid and applicable,\n          and that the caveat is still valid), and off Ben drives.\n        <\/p><p>\n          Ben arrives at the fancy award venue, where there is valet parking.\n          The Valet approaches wearing the nametag \"Lem E. Driveit\" and asks\n          for a key to park the car.\n          Ben remembers he has heard some stories about valets going on\n          joyrides and isn't sure if they're true, but decides there's no need\n          to take the risk: he can use his capability to create a new, further\n          restricted capability which he can delegate to Lem.\n          This new car key has the caveat that it can be used to drive up\n          to 5 kilometers, but no more:\n        <\/p><pre class=\"example highlight javascript\">\n    {\"@context\": [\"https:\/\/example.org\/ocap\/v1\",\n                  \"https:\/\/autopower.example\/\"],\n     \"id\": \"https:\/\/chatty.example\/ben\/caps#2cdea8c1\",\n     \"parentCapability\": \"https:\/\/social.example\/alyssa\/caps#79795d78\",\n     \"invoker\": \"https:\/\/lem.example\/#key-bf36\",\n     \/\/ Ben adds this caveat: this capability can be used to drive the\n     \/\/ car, but not for more than 5 kilometers\n     \"caveat\": [\n       {\"type\": \"DriveNoMoreThan\",\n        \/\/ Alyssa's gauge currently says 123854 kilometers driven,\n        \/\/ so this is only 5 km more than the current value\n        \"kilometers\": 123859}],\n     \/\/ Finally Ben signs this object with the key he was granted\n     \/\/ authority with\n     \"proof\": {\n        \"type\": \"RsaSignature2016\",\n        \"proofPurpose\": \"capabilityDelegation\",\n        \"created\": \"2017-06-13T19:15:03Z\",\n        \"creator\": \"https:\/\/chatty.example\/ben\/#key-33\",\n        \"signatureValue\": \"...\"}}\n        <\/pre><p>\n          (It probably doesn't matter to this scenario, but since it is derived\n          from the car key that Ben holds, it also has the same caveat that it\n          will also be remote destructed by Alyssa along with Ben's car key\n          should she press the relevant button.)\n        <\/p><\/section><section id=\"capabilities-are-safer\"><h2>Capabilities Are Safer<\/h2><p>\n          We can see from the above example that by using an object capability\n          system such as OCAP-LD can give us some additional\n          power around delegating and restricting the scope of capabilities.\n          But object capabilities provide improved safety characteristics\n          over Access Control Lists which may make them frequently better\n          choices (for more details, see the paper\n          <a href=\"http:\/\/waterken.sourceforge.net\/aclsdont\/current.pdf\">\n            ACLs Don't<\/a>).\n          Object capability systems are in general more robust against:\n        <\/p><ul><li>\n            <b>Ambient authority problems:<\/b>\n            For example, in an access control list environment,\n            if an attacker gets you to open a message in your email client\n            which contains an image which triggers an arbitrary code execution\n            vulnerability in the image library it is using, an attacker can\n            use this to do everything that you can: read or destroy your\n            documents, they can run programs as you, and they can in general\n            do anything that you can... all from a pin-prick sized vulnerability.\n            Properly using an object capability environment, the worst an\n            attacker could do is mess with your email (or if the program is\n            itself capability aware, they may not be able to do even\n            that... they could just display some nasty imagery on your screen,\n            which they could do already anyway).\n          <\/li>\n          <li>\n            <b>Confused deputy attacks:<\/b>\n            For example, browsers and both backend and frontend software for\n            the web have frequently struggled with Cross-Site Request Forgery\n            attack issues.\n            Your browser may both allow you websites to access resources on\n            other websites and also allow you to log into various websites,\n            say a social networking site or a bank.\n            A malicious website could use this to its advantage to trick\n            your browser into performing actions as you on those other sites...\n            say to read from or spam to your social timeline, or to withdraw\n            money from your bank account.\n            While various techniques to can be used to restrict the scope\n            of these attacks, they are notoriously hard to prevent.\n          <\/li>\n        <\/ul><p>\n          Object capabilities are less vulnerable to these kinds of attacks\n          because capabilities are an encoding of \"the principle of least\n          authority\" in software development practice.\n          OCAP-LD helps bring the power of capabilities to the web, providing\n          specific and directed grants of authority.\n        <\/p><\/section><\/section><section id=\"terminology\"><h2>Terminology<\/h2><dl><dt>capability<\/dt>\n        <dd>\n          Authority which may be invoked to perform some operation upon a\n          (linked data) object.\n        <\/dd>\n        <dt>target<\/dt>\n        <dd>\n          The entity that will be acted upon when the capability is invoked.\n        <\/dd>\n        <dt>capability chain<\/dt>\n        <dd>\n          A chain of capability documents which may be used to delegate\n          authority to other entities in the system.\n          Each capability document in the chain inherits the caveats of\n          previous capabilities in the chain and may only be granted by\n          entities which have existing authority within the chain.\n        <\/dd>\n        <dt>caveat<\/dt>\n        <dd>\n          Also known as \"attenuation\" within object capability literature,\n          a \"caveat\" may be attached to a capability as a restriction on\n          how that capability may be used.\n          Delegated capabilities will preserve caveats and may be themselves\n          may be further restricted by adding more caveats.\n        <\/dd>\n        <dt>invocation<\/dt>\n        <dd>\n          The \"activation\" of a capability, represented as a linked data\n          document which is signed in such a way that matches authority\n          granted through a prior capability.\n          \n          An invocation may have arguments, similar to how a procedure call\n          may have arguments.\n        <\/dd>\n        <dt>action<\/dt>\n        <dd>\n          An argument of an invocation which directs what particular\n          functionality of the object is being used.\n          For example, an invocation against a file storage service\n          may specify either a write action or a read action.\n          (Actions are the same as what are frequently called \"methods\" in\n          computer programming.)\n        <\/dd>\n        <dt>parentCapability<\/dt>\n        <dd>\n          The previous capability document on the chain, which is granting\n          authority to this invocation document.\n          In the case of the first delegated capability document, this points\n          at the target.\n        <\/dd>\n        <dt>capabilityDelegation<\/dt>\n        <dd>\n          This term serves two purposes:\n          <ul><li>\n              As a property from which the target supplies the \"initial source\n              of authority\" on the chain.\n              This is holds cryptographic material which it may use to initially\n              delegate on the chain (and in some cases which it may use to\n              invoke itself as a capability).\n            <\/li>\n            <li>\n              As the value of a proof's <code>proofPurpose<\/code> to indicate\n              that this proof is intended to grant authority to the <code>invoker<\/code>\n              entities on the capability.\n            <\/li>\n          <\/ul><\/dd>\n      <\/dl><\/section><section id=\"capabilities\"><h2>Capabilities<\/h2><p>\n        OCAP-LD capabilities are encoded through a chain of linked data\n        documents, granting authority to a target, possibly restricted through\n        \"caveats\".\n        Authority starts with the target (which always has authority to invoke\n        itself) and extends to further <code>invoker<\/code> entities along\n        the chain.\n        Each capability document granting authority must be signed off with a\n        proof by an entity which has previously been granted authority on the\n        chain.\n        Any caveat applied by a parent in the chain applies to its descendants.\n      <\/p><div class=\"note\"><p>\n          The target's json-ld document is a kind of \"special\" capability document\n          implicitly granting authority to itself.\n          But why?\n          Surely in most protocols, such as when objects are sending\n          invocations over HTTP POST requests, any target that wants to\n          self-modify could do so internally without an explicit invocation\n          process.\n          But in some systems such as blockchains, there is no \"internal\" state,\n          so explicitly doing an invocation to change behavior is still useful.\n        <\/p><p>\n          Another reason is that it simplifies the invocation algorithm: as we\n          delegate authority to the first non-target entity, we do so using the\n          target's cryptographic authority.\n          If this authority can be used to grant capabilities, it may as well\n          be able to be used to invoke them as well.\n        <\/p><\/div><p>\n        Every capability document, except for the target, MUST have:\n      <\/p><ul><li>an associated <code>id<\/code><\/li>\n        <li>\n          <code>parentCapability<\/code>, which links to the target if this is\n          the first delegated capability on the chain, or otherwise links to\n          another capability document\n        <\/li>\n        <li>\n          \n          a <code>proof<\/code> field, which MUST sign the document with\n          <a href=\"https:\/\/w3c-dvcg.github.io\/ld-signatures\/\">\n            Linked Data Signatures<\/a>\n          by cryptographic material which has already been granted authority\n          on the chain (either being cryptographic material from the\n          \n          <code>capabilityDelegation<\/code> property of the target, or from\n          an <code>invoker<\/code> previously granted authority on the chain.\n          \n        <\/li>\n      <\/ul><p>\n        Every capability document, except for the target, MAY have:\n      <\/p><ul><li>\n          <code>invoker<\/code>, which links to one or more instances of\n          cryptographic material (such as public keys) being granted authority\n          to use this capability\n        <\/li>\n      <\/ul><div class=\"note\"><p>\n          Why don't all child capabilities require a invoker property?\n          The reason is that it is perfectly reasonable for an entity which has\n          authority to a capability to give an attenuated capability to\n          <em>itself<\/em> to be used in a specific context.\n        <\/p><p>\n          Capabilities are not structured around \"who has access\" to something\n          but rather holding onto a capability for a particular use, and in\n          order to avoid confused deputy and ambient authority problems,\n          in general objects should not hold on to a \"bucket\" of capabilities\n          which grant them authority, but hold on to them within specific\n          contexts for specific purposes.\n          In general, when an object is granted a capability, it is important\n          that it know for what specific purpose it has been granted that\n          capability so that it does not become confused and use it in the\n          wrong scenario.\n          This is not unlike how in computer programs a procedure called with\n          specific arguments understands for what the meaning of those arguments\n          are and for what they may be used.\n        <\/p><\/div><section id=\"delegation\"><h2>Delegation through Capability Chains<\/h2><p>\n          A capability document that is not the target MUST also have a\n          <code>parentCapability<\/code> property which either points at the\n          target or another capability document.\n          A series of capability chained together in this way is called a\n          \"capability chain\" and is how delegation of capabilities are\n          handled in OCAP-LD.\n          New keys MAY be granted authority to use this capability through\n          the associated <code>invoker<\/code> property.\n        <\/p><\/section><section id=\"caveats\"><h2>Caveats<\/h2><p>\n          Every capability document MAY add restrictions on the way the\n          capability may be used by adding to the <code>caveat<\/code> property.\n          Capabilities inherit the restrictions from all <code>caveat<\/code>\n          properties of their parents, and MAY add new caveats in addition to\n          those of their parents.\n        <\/p><p>\n          The meaning of caveats are determined by their <code>type<\/code>\n          and whatever other properties they have.\n          Due to the way they are interpreted at invocation type by the target,\n          some mutual understanding of terminology must be understood between\n          the entity adding a caveat and the target evaluating (or any other\n          parties observing) the invocation.\n        <\/p><\/section><\/section><section id=\"invocation\"><h2>Invocation<\/h2><p>\n        A capability may be enacted through the process of invocation.\n        In the context of OCAP-LD, an invocation consists of a linked data\n        object which MUST have a <code>proof<\/code> property with a value\n        containing:\n      <\/p><ul><li>\n          a <code>proofPurpose<\/code> of <code>capabilityInvocation<\/code>\n        <\/li>\n        <li>\n          a <code>capability<\/code> property which links to the\n          capability document that grants authority to invoke this capability\n        <\/li>\n        <li>\n          the proof must validate against cryptographic material\n          granted authority by the <code>invoker<\/code> field from the\n          capability chain.\n        <\/li>\n      <\/ul><p>\n        An invocation SHOULD have an <code>id<\/code> (which may also serve as a\n        nonce).\n        Any other properties are considered arguments to the invocation.\n      <\/p><p>\n        Each capability document accrues granted authority in the chain, which\n        is used to both authorize further capability delegation and may be used\n        to invoke the capability chain.\n        The capability document that is invoked's chain is recursively traversed\n        up along the <code>previousCapability<\/code> document until the target is\n        found (the root document that has no <code>previousCapability<\/code>\n        property).\n        The target's <code>capabilityDelegation<\/code> cryptographic material is\n        marked in the initial set of authority.\n        The capability chain is then traversed from this root target back down\n        through each delegated capability all the way to the invoked capability\n        document leaf, validating and accruing authority:\n      <\/p>\n        <ul>\n                    \n               \n               \n                  \n        <\/ul><ul>\n                   \n        <\/ul><ul>\n                  \n              \n        <\/ul><ul>\n                    \n                   \n        <\/ul><p>\n        At this point, the invocation is considered valid and any relevant\n        action may be performed.\n        (This does not guarantee that a specific result will occur, merely that\n        the invocation itself is valid; a invocationTarget could in evaluating\n        the invocation encounter invalid input and return or raise an error,\n        for instance.)\n      <\/p><div class=\"note\"><p>\n          Since invocations are themselves linked data documents, it is\n          possible for an invocation to refer to an invocation as an argument.\n          For this reason it is critical that mechanisms for accepting\n          invocations be sure which invocation is being performed.\n          For example, a mechanism that accepts OCAP-LD capabilities as\n          JSON-LD documents will likely have no trouble telling which\n          invocation is being performed since that will be the \"top\"\n          of the framed document.\n          However, a mechanism which accepted N-Quads would need additional\n          information supplied perhaps in the headers in order to be able\n          to determine the \"root\" of the invocation.\n        <\/p><\/div><section id=\"actions\"><h2>Actions<\/h2><p>\n          Actions are a common way to direct behavior of an invocation.\n          Targets are free to choose their own mechanisms for directing\n          behavior, but MAY support the <code>action<\/code> property on\n          invocations as one common behavioral direction technique.\n          The <code>action<\/code> property points to a URI as a form of\n          vocabulary to determine which action is being performed.\n          For example, a capability to a file storage system may allow\n          for both reading and writing files, and a user could choose\n          to set the <code>action<\/code> to\n          <code>https:\/\/datastore.example\/WriteFile<\/code>\n          as an argument on their invocation.\n        <\/p><\/section><\/section><section class=\"appendix informative\" id=\"relationship-to-vc\"><h2>Relationship to Verifiable Credentials<\/h2><p>\n        The\n        <a href=\"https:\/\/www.w3.org\/TR\/verifiable-claims-data-model\/\">\n          Verifiable Credentials Data Model<\/a>\n        provides a powerful vocabulary for correlating information and\n        presenting correlated information in a secure way.\n        This can be used to make claims about some subject, to present\n        credential certifying you are qualified to do some thing, or so on.\n        This is useful for gathering and presenting information that can\n        be reasoned about.\n        Indeed, this maps nicely to \"the real world\", as humans tend to be\n        \"correlation\/reasoning machines\".\n        For example, Eva Lu Ator is hoping to hire a system administrator and\n        may check for a college diploma, past work history, or recommendations\n        from known entities considered qualified to judge competency to fulfill\n        the role, and all of these can be modeled nicely with the Verifiable\n        Credentials Data Model.\n      <\/p><p>\n        Thus it may be tempting to design a system in which a credential\n        is presented in order to permit or forbid access to some resource.\n        Unfortunately, associating \"who is authorized to do what\" leads us back\n        to all the problems of access control lists,\n        <a href=\"https:\/\/w3c-ccg.github.io\/ocap-ld\/capabilities-are-safer\">which we would like to avoid<\/a>.\n        When returning to our hiring a system administrator example it is\n        easy to imagine how, upon hiring Alice as as a system administrator,\n        Eva would hand Alice a capability to begin administrating her systems.\n        But in considering the previous step where Eva contemplates\n        <i>whether or not to hand Alice that capability<\/i>\n        seems to pull us right back to correlating information about Alice's\n        identity.\n        (What is Alice's educational background?  What is Alice's prior work\n        experience?)\n      <\/p><p>\n        We seem to be in a conundrum.\n        Claims and credentials are forms of correlation that allow us to reason\n        about an entity in our squishy human world, but are unsafe when used as\n        mechanisms to authorize some event to occur within a system.\n        Capabilities are a safe mechanism to model the flow of authority through\n        a system, but there are times when capabilities have not been granted\n        and we need to make a \"judgement call\" by correlating information about\n        that entity.\n        What should we do?\n      <\/p><p>\n        To pose the question is to see the answer: the right approach is to use\n        each system for what it does best.\n        Use correlation (Verifiable Credentials) in a reasoning system (most\n        commonly human reasoning) as a path to make judgements about whether to\n        hand an entity a specific set of initial capabilities.\n        Use capabilities (OCAP-LD) as the mechanism to grant and exercise\n        authority through computing systems.\n        To return to our system administrator example, when Alice applies for the\n        job, she submits a series of credentials about her prior work history\n        and degree, and Eva is able to verify that it is Alice's former\n        employers and university which have made these claims.\n        Deciding that Alice is fit for the job, Eva hands Alice her initial\n        capability which grants her authority to administrate the systems in\n        question (with a caveat that allows Eva to revoke that authority at a\n        future date, if appropriate).\n        Alice uses that capability as the initial entry point into\n        administrating the system.\n      <\/p><\/section><section class=\"appendix informative\" id=\"handling-abuse\"><h2>Handling abuse<\/h2><p>\n        The previous section discussed cases in which correlation (particularly\n        through Verifiable Credentials) is used to grant an initial set of\n        capabilities.\n        This section discusses how correlation may be used to revoke\n        capabilities under certain circumstances, and steps that should be\n        taken to ensure this is done safely.\n      <\/p><p>\n        Capability systems focus on what authority grants that an operation may\n        be performed, not on who is performing the action.\n        In some object capability systems (such as many object capability\n        programming languages) it is not even possible to see what entity is\n        invoking a capability.\n        In the case of OCAP-LD, it is always possible to inspect an invocation\n        to check which authentication material authorized to invoke the\n        capability is being used in the invocation, and this may permit\n        checking a certain amount of \"who\" is performing the action, but to\n        focus on the \"who\" rather than the source of authority would\n        <a href=\"https:\/\/w3c-ccg.github.io\/ocap-ld\/capabilities-are-safer\">\n          bring us back to the dangers of Access Control Lists<\/a>.\n      <\/p><p>\n        Nonetheless, there comes a time when correlating information about\n        usage of a resource becomes critical, and this is in mitigating abuse.\n        Even traditional capability systems, in developing solutions to this\n        problem, develop techniques for correlating information about \"who\"\n        has performed an action.\n        (For example,\n        <a href=\"http:\/\/www.erights.org\/elib\/capability\/horton\/\">Horton<\/a>\n        is one such system for E, though as previously stated,\n        such information can be cleaned already by looking at the invocation\n        and capability chain documents in OCAP-LD.)\n        The key to permit detecting and mitigating abuse safely in a capability\n        based environment is to treat correlation of who is using a service\n        not as something that is used per-invocation, but as something to be\n        reasoned about to possibly revoke authority.\n        (Astute readers may observe that between the previous section and this\n        section we have defined safe ways to use correlation in conjunction\n        with capabilities: as the entry point and possible termination point,\n        with capabilities powering the machine in-between.)\n      <\/p><p>\n        For example, Alice has granted capabilities (with caveats that permit\n        revocation) to a number of users access to run programs on the systems\n        she is administrating, but someone is using their capability to abuse\n        the system.\n        The system itself performs the invocations as the users call them,\n        but Alice, or a program Alice runs, is able to analyze a log of past\n        invocations called.\n        Having finished the analysis, Alice realizes that the abuse is coming\n        from the invocation of a capability granted to some of Mallet's\n        authentication material, and Alyssa revokes that capability.\n      <\/p><\/section>","created_at":"2019-01-28T00:16:37+0000","updated_at":"2019-01-28T00:16:37+0000","annotations":[],"mimetype":"text\/html","reading_time":20,"domain_name":"w3c-ccg.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["gnu","linux","cli","dotfiles"],"is_public":false,"id":790,"title":"Stow\n- GNU Project - Free Software Foundation","url":"https:\/\/www.gnu.org\/software\/stow\/","content":"<hr \/><p> GNU Stow is a symlink farm manager which takes distinct packages\nof software and\/or data located in separate directories on the\nfilesystem, and makes them appear to be installed in the same place.\nFor example, <code>\/usr\/local\/bin<\/code> could contain symlinks to\nfiles within <code>\/usr\/local\/stow\/emacs\/bin<\/code>,\n<code>\/usr\/local\/stow\/perl\/bin<\/code> etc., and likewise recursively\nfor any other subdirectories such as <code>...\/share<\/code>,\n<code>...\/man<\/code>, and so on. <\/p><p> This is particularly useful for keeping track of system-wide and\nper-user installations of software built from source, but\n<a href=\"http:\/\/brandon.invergo.net\/news\/2012-05-26-using-gnu-stow-to-manage-your-dotfiles.html\">can\nalso facilitate a more controlled approach to management of configuration\nfiles in the user's home directory<\/a>, especially when \n<a href=\"http:\/\/lists.gnu.org\/archive\/html\/info-stow\/2011-12\/msg00000.html\">coupled\nwith version control systems<\/a>. <\/p><p> Stow is implemented as a combination of a <a href=\"http:\/\/www.perl.org\/\">Perl<\/a> script providing a CLI interface,\nand a <a href=\"http:\/\/search.cpan.org\/dist\/Stow\/\">backend Perl\nmodule<\/a> which does most of the work.  Stow is <a href=\"http:\/\/www.gnu.org\/philosophy\/free-sw.html\">Free Software<\/a>,\nlicensed under the <a href=\"http:\/\/www.gnu.org\/copyleft\/gpl.html\">GNU\nGeneral Public License<\/a>. <\/p><h3 id=\"news\">Latest news<\/h3><dl><dt><em>Mon 9 Nov 2015<\/em><\/dt>\n  <dd>\n    <a href=\"http:\/\/lists.gnu.org\/archive\/html\/info-stow\/2015-11\/msg00000.html\">Stow\n    2.2.2 has been released<\/a>!  After a long wait, this release contains a\n    number of bug fixes and minor cleanups.\n    <a href=\"http:\/\/git.savannah.gnu.org\/cgit\/stow.git\/tree\/NEWS\">Read\n    details of what's new.<\/a>\n    \n  <\/dd>\n  <dt><em>Sat 18 Feb 2012<\/em><\/dt>\n  <dd>\n    <a href=\"http:\/\/lists.gnu.org\/archive\/html\/info-stow\/2012-01\/msg00000.html\">Stow\n    2.2.0 has been released<\/a>!  This release introduces a long-wanted\n    <code>--no-folding<\/code> option which disables the (re)folding which\n    in some usage scenarios can be confusing or undesirable.  It also\n    drops the abbreviated <code>-a<\/code> version of the <code>--adopt<\/code>\n    option for safety reasons, and makes various improvements\n    to the documentation, test suite, and command-line output.\n    <a href=\"http:\/\/git.savannah.gnu.org\/cgit\/stow.git\/tree\/NEWS\">Read\n    details of what's new.<\/a>\n    \n  <\/dd>\n  <dt><em>Tue 10 Jan 2012<\/em><\/dt>\n  <dd>\n    <a href=\"http:\/\/lists.gnu.org\/archive\/html\/info-stow\/2012-01\/msg00000.html\">Stow\n    2.1.3 has been released<\/a>!  This has several minor improvements\n    to 2.1.0, including a much improved installation procedure,\n    documentation fixes, and a new <code>--adopt<\/code> option.\n    <a href=\"http:\/\/git.savannah.gnu.org\/cgit\/stow.git\/tree\/NEWS\">Read\n    details of what's new.<\/a>\n    \n  <\/dd>\n  <dt><em>Sat 03 Dec 2011<\/em><\/dt>\n  <dd>\n    <a href=\"http:\/\/lists.gnu.org\/archive\/html\/info-stow\/2011-12\/msg00001.html\">Stow\n    2.1.0 has been released<\/a>!  This is the first official release\n    since 2002.  <a href=\"http:\/\/git.savannah.gnu.org\/cgit\/stow.git\/tree\/NEWS\">Read\n    details of what's new.<\/a>\n    \n  <\/dd>\n<\/dl><h3 id=\"download\">Downloading\n  Stow<\/h3><p>Stow\ncan be found on the main GNU ftp server:\n<a href=\"http:\/\/ftp.gnu.org\/gnu\/stow\/\">http:\/\/ftp.gnu.org\/gnu\/stow\/<\/a>\n(via HTTP) and\n<a href=\"ftp:\/\/ftp.gnu.org\/gnu\/stow\/\">ftp:\/\/ftp.gnu.org\/gnu\/stow\/<\/a>\n(via FTP).  It can also be found\non the <a href=\"https:\/\/www.gnu.org\/prep\/ftp.html\">GNU mirrors<\/a>;\nplease\n<a href=\"http:\/\/ftpmirror.gnu.org\/stow\/\">use\na mirror<\/a> if possible.<\/p><p> There is also a <a href=\"https:\/\/savannah.gnu.org\/git\/?group=stow\">git repository<\/a>\ncontaining the latest development code.\n<\/p><h3 id=\"documentation\">Documentation<\/h3><p><a href=\"https:\/\/www.gnu.org\/software\/stow\/manual\/\">Documentation for\nStow<\/a>\nis available online, as\nis <a href=\"https:\/\/www.gnu.org\/manual\/\">documentation for most GNU software<\/a>.  You may\nalso find more information about\nStow\nby running\n<em>info\u00a0stow<\/em>\nor\n<em>man\u00a0stow<\/em>,\nor by looking at\n<em>\/usr\/share\/doc\/stow\/<\/em>,\n<em>\/usr\/local\/doc\/stow\/<\/em>,\nor similar directories on your system.  A brief summary is available by\nrunning <em>stow\u00a0--help<\/em>.<\/p><h3 id=\"mail\">Mailing lists<\/h3><p>Stow\nhas the following mailing lists:<\/p><ul><li><a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/help-stow\">help-stow<\/a> is\nfor general user help and discussion.<\/li>\n<li><a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/stow-devel\">stow-devel<\/a>\nis used to discuss most aspects of\nStow,\nincluding development and enhancement requests.<\/li>\n<li><a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/bug-stow\">bug-stow<\/a>\nis for bug reports.<\/li>\n<\/ul><p>Announcements about\nStow are posted to\n<a href=\"http:\/\/lists.gnu.org\/mailman\/listinfo\/info-stow\">info-stow<\/a>\nand also, as with most other GNU software, to\n<a href=\"http:\/\/lists.gnu.org\/mailman\/listinfo\/info-gnu\">info-gnu<\/a>\n(<a href=\"http:\/\/lists.gnu.org\/archive\/html\/info-gnu\/\">archive<\/a>).<\/p><p>Security reports that should not be made immediately public can be\nsent directly to the maintainer.  If there is no response to an urgent\nissue, you can escalate to the general\n<a href=\"http:\/\/lists.gnu.org\/mailman\/listinfo\/security\">security<\/a>\nmailing list for advice.<\/p><p>\nThe Savannah project also has a\n<a href=\"https:\/\/savannah.gnu.org\/mail\/?group=stow\">mailing lists<\/a> page.\n<\/p><h3 id=\"contribute\">Getting involved<\/h3><p>Development of\nStow,\nand GNU in general, is a volunteer effort, and you can contribute.  For\ninformation, please read <a href=\"https:\/\/www.gnu.org\/help\/\">How to help GNU<\/a>.  If you'd\nlike to get involved, it's a good idea to join the <a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/stow-devel\">stow-devel<\/a> mailing\nlist.<\/p><dl><dt>Bug reporting<\/dt>\n<dd>Please send bug reports to the\n<a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/bug-stow\">bug-stow<\/a>\nmailing list (see <a href=\"#mail\">Mailing lists<\/a> above).<\/dd>\n<dt>Development<\/dt>\n<dd>For <a href=\"https:\/\/savannah.gnu.org\/git\/?group=stow\">development sources<\/a> and other\ninformation, please see the\n<a href=\"http:\/\/savannah.gnu.org\/projects\/stow\/\">Stow\nproject page<\/a>\nat <a href=\"http:\/\/savannah.gnu.org\">savannah.gnu.org<\/a>.\nThere is also a <a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/stow-devel\">stow-devel<\/a>\nmailing list (see <a href=\"#mail\">Mailing lists<\/a> above).<\/dd>\n<dt>Translating\n  Stow<\/dt>\n<dd>Stow is not currently multi-lingual, but patches would be\ngratefully accepted.  Please e-mail <a href=\"https:\/\/lists.gnu.org\/mailman\/listinfo\/stow-devel\">stow-devel<\/a>\nif you intend to work on this.<\/dd>\n<dt>Maintainers<\/dt>\n<dd>Stow\nis currently being maintained by\nAdam Spiers and Troy Mill.\nPlease use the mailing lists for contact.<\/dd>\n<\/dl><h3 id=\"license\">Licensing<\/h3><p>Stow\nis free software; you can redistribute it and\/or modify it under the\nterms of the <a href=\"http:\/\/www.gnu.org\/licenses\/gpl.html\" rel=\"license\">GNU General Public License<\/a> as published by the Free\nSoftware Foundation; either version\u00a03 of the License, or (at your\noption) any later version.<\/p>","created_at":"2019-01-27T14:09:47+0000","updated_at":"2019-01-27T14:09:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"www.gnu.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","caching"],"is_public":false,"id":789,"title":"latest","url":"http:\/\/www.php-cache.com\/en\/latest\/","created_at":"2019-01-23T13:41:37+0000","updated_at":"2019-01-23T13:41:50+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.php-cache.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":788,"title":"Pharo - Welcome to Pharo!","url":"https:\/\/pharo.org\/web","content":"<a class=\"btn download\" href=\"https:\/\/pharo.org\/web\/download\">Download latest version<\/a><div class=\"container\"><div><p>\nRead more about <a href=\"https:\/\/pharo.org\/news\/pharo7.0-released\">here<\/a>\n<\/p><p>\n3000 people registered and follow the Pharo Mooc. You can find it <a href=\"http:\/\/mooc.pharo.org\">http:\/\/mooc.pharo.org<\/a> \n<br \/>Watch the <a href=\"https:\/\/www.dailymotion.com\/video\/x3qedvo_fun-mooc-programmation-objet-immersive-en-pharo_school\">Teaser<\/a>\n<br \/>Follow us on Twitter: <a href=\"https:\/\/twitter.com\/pharoproject\">@pharoproject<\/a>\n<\/p><\/div><div class=\"topic\"><h3>A glimpse of Pharo<\/h3><pre>$ curl https:\/\/get.pharo.org | bash \n$ .\/pharo Pharo.image eval \"42 factorial\"<\/pre><h3>Teaser<\/h3><ul><li><strong>Simple &amp; powerful language:<\/strong> No constructors, no types declaration, no interfaces, no primitive types.  Yet a powerful and elegant language with a full syntax fitting in one postcard! Pharo is objects and messages <em>all the way down<\/em>. <\/li>\n<\/ul><ul><li><strong>Live, immersive environment:<\/strong> Immediate feedback at any moment of your development: Developing, testing, debugging. Even in production environments, you will never be stuck in compiling and deploying steps again!<\/li>\n<\/ul><ul><li><strong>Amazing debugging experience:<\/strong> The Pharo environment includes a debugger unlike anything you've seen before. It allows you to step through code, restart the execution of methods, create methods on the fly, and much more!<\/li>\n<\/ul><ul><li><strong>Pharo is yours:<\/strong> Pharo is made by an incredible community, with more than 100 contributors for the last revision of the platform and hundreds of people contributing constantly with frameworks and libraries.<\/li>\n<\/ul><\/div><\/div>","created_at":"2019-01-22T17:23:42+0000","updated_at":"2019-01-22T17:23:42+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"pharo.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":787,"title":"High Scalability -","url":"http:\/\/highscalability.com\/","content":"<div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Saturday, January 19, 2019 at 1:48PM <\/div> \n    \n  \n      \n      <div class=\"body\">\n  \n        \n        \n        \n          <p><img src=\"http:\/\/farm3.static.flickr.com\/2363\/5781449767_54902da8a5_o.jpg\" alt=\"\" style=\"float: right;\" \/><\/p>\n<h2>Who's Hiring?\u00a0<\/h2>\n<ul><li><strong>Triplebyte<\/strong> lets exceptional software engineers skip screening steps at hundreds of top tech companies like <strong>Apple, Dropbox, Mixpanel, and Instacart<\/strong>. Make your job search <em>O<\/em>(1), not <em>O<\/em>(<em>n<\/em>). <a href=\"https:\/\/triplebyte.com\/?ref=hs_apr18_c2\">Apply here.<\/a> <\/li>\n<li><strong>Need excellent people? Advertise your job here!<\/strong>\u00a0<\/li>\n<\/ul><h2>Fun and Informative Events<\/h2>\n<ul><li><strong>Advertise your event here!<\/strong><\/li>\n<\/ul><h2>Cool Products and Services<\/h2>\n<ul><li><a href=\"https:\/\/www.gkogan.co\/sbc\/\"> Shape the future of software in your industry<\/a>. The Software Buyers Council is a panel of engineers and managers who want to share expert knowledge, contribute to improvement of software, and help startups in their industry. Receive occasional invitations to chat with for 30 minutes about your area of expertise and software usage. No obligations, no marketing emails or sales calls. Upcoming topics include infrastructure and application monitoring, AI\/ML platforms, and more. <a href=\"https:\/\/www.gkogan.co\/sbc\/\">Learn more and join today<\/a>. <\/li>\n<li><strong>InMemory.Net<\/strong>\u00a0provides a <strong>Dot Net native in memory database<\/strong> for analysing large amounts of data. It runs natively on .Net, and provides a native .Net, COM &amp; ODBC apis for integration. It also has an easy to use language for importing data, and supports standard SQL for querying data.\u00a0<a href=\"http:\/\/inmemory.net\/\">http:\/\/InMemory.Net<\/a><\/li>\n<\/ul><ul><li><strong>For heads of IT\/Engineering responsible for building an analytics infrastructure<\/strong>, <a href=\"https:\/\/etleap.com\/?utm_source=high-scalability&amp;utm_medium=sponsored-link&amp;utm_campaign=general\">Etleap is an ETL solution<\/a> for creating perfect data pipelines from day one. Unlike older enterprise solutions, Etleap doesn\u2019t require extensive engineering work to set up, maintain, and scale. It automates most ETL setup and maintenance work, and simplifies the rest into 10-minute tasks that analysts can own. <a href=\"https:\/\/etleap.com\/customers\/?utm_source=high-scalability&amp;utm_medium=sponsored-link&amp;utm_campaign=general\">Read stories from customers like Okta and PagerDuty<\/a>, or <a href=\"https:\/\/etleap.com\/?utm_source=high-scalability&amp;utm_medium=sponsored-link&amp;utm_campaign=general\">try Etleap yourself<\/a>.<\/li>\n<\/ul><ul><li>Build, scale and personalize your\u00a0<a href=\"https:\/\/getstream.io\/try-the-api\/?utm_source=hs&amp;utm_campaign=hs&amp;utm_medium=hs\">news feeds and activity streams with getstream.io<\/a>. Try the API now in this\u00a0<a href=\"https:\/\/getstream.io\/try-the-api\/?utm_source=hs&amp;utm_campaign=hs&amp;utm_medium=hs\">5 minute interactive tutorial<\/a>.\u00a0<strong>Stream\u00a0<\/strong>is free up to 3 million feed updates so it's easy to get started. Client libraries are available for Node, Ruby, Python, PHP, Go, Java and .NET. Stream is currently also hiring Devops and Python\/Go developers in Amsterdam. More than 400 companies rely on Stream for their production feed infrastructure, this includes apps with 30 million users. With your help we'd like to ad a few zeros to that number. Check out the job opening on\u00a0<a href=\"https:\/\/angel.co\/stream\">AngelList<\/a>.<\/li>\n<\/ul><ul><li><a href=\"https:\/\/www.scalyr.com\/\">Scalyr<\/a> is a lightning-fast <a href=\"https:\/\/www.scalyr.com\/product\/centralized-log-management\">log management<\/a> and operational data platform. \u00a0It's a tool (actually, multiple tools) that your entire team will love. \u00a0Get visibility into your production issues without juggling multiple tabs and different services -- all of your logs, server metrics and alerts are in your browser and at your fingertips. . \u00a0Loved and used by teams at Codecademy, ReturnPath, Grab, and InsideSales. Learn more today or see why Scalyr is a great <a href=\"https:\/\/www.scalyr.com\/product\/splunk-alternative\">alternative to Splunk<\/a>.<\/li>\n<li><strong>Advertise your product or service here!<\/strong><\/li>\n<\/ul><p><strong><em> If you are interested in a sponsored post for an event, job, or product, please <a href=\"http:\/\/highscalability.com\/contact\">contact<\/a> us for more information.<\/em><\/strong><\/p>\n<hr \/><h2>Why engineers are joining the Software Buyers Council:<\/h2>\n<p>1) Make a big impact. All studies are requested by startup founders who are committed to making big changes based on the feedback they receive.<\/p>\n<p>2) Be heard. Have one-on-one conversations with me, not some market research corporation. Informal conversation about topics that interest you. No judgement, no sales pitch, no cross-examination.<\/p>\n<p>3) Minimal effort and no commitment. Accept, reject, or ignore invitations as you wish. When you accept an invitation, you can choose the date and time, and the chats are only 30 minutes. Nothing to prepare before or after.<\/p>\n<p>4) Help the startup community and your peers. Your input will help software startups (not mega corporations) and your peers who use or might use that software in the future.<\/p>\n<p><a href=\"https:\/\/www.gkogan.co\/sbc\/\">Learn more and join today<\/a>. Upcoming topics include infrastructure and application monitoring, AI\/ML platforms, and more.<\/p>\n<hr \/><h2>Make Your Job Search <em>O<\/em>(1) \u2014 not <em>O<\/em>(<em>n<\/em>)<\/h2>\n<p><a href=\"https:\/\/triplebyte.com\/?ref=hs_apr18_c3\">Triplebyte<\/a> is unique because they're a team of engineers running their own centralized technical assessment. Companies like <strong>Apple, Dropbox, Mixpanel, and Instacart<\/strong> now let Triplebyte-recommended engineers skip their own screening steps.<\/p>\n<p>We found that <strong>High Scalability readers<\/strong> are about 80% more likely to be in the top bracket of engineering skill.<\/p>\n<p>Take <a href=\"https:\/\/triplebyte.com\/?ref=hs_apr18_c3\">Triplebyte's multiple-choice quiz<\/a> (system design and coding questions) to see if they can help you scale your career faster.<\/p>\n<hr \/><h2>The Solution to Your Operational Diagnostics Woes<\/h2>\n<p><strong><a href=\"https:\/\/www.scalyr.com\/\">Scalyr<\/a><\/strong> gives you instant visibility of your production systems, helping you turn chaotic logs and system metrics into actionable data at interactive speeds. Don't be limited by the slow and narrow capabilities of traditional <a href=\"https:\/\/www.scalyr.com\/\">log monitoring<\/a> tools. View and analyze all your logs and system metrics from multiple sources in one place. Get enterprise-grade functionality with sane pricing and insane performance. <a href=\"https:\/\/www.scalyr.com\/\">Learn more today<\/a>.\u00a0<\/p>\n<hr \/><p><em> If you are interested in a sponsored post for an event, job, or product, please <\/em><em><a href=\"http:\/\/highscalability.com\/contact\">contact<\/a> us for more information.<\/em><\/p>\n        \n  \n          \n  \n         \n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Friday, January 18, 2019 at 8:48AM <\/div> \n    \n  \n      \n      <div class=\"body\">\n  \n        \n        \n        \n          <p>Sorry, Stuff The Internet Says On Scalability has been called on the account of wind, rain, power outages and general mayhem. We're all safe, but it's hard to write a post using stone knives and bear skins. See you next week.<\/p>\n<blockquote class=\"twitter-tweet\">\n<p dir=\"ltr\" lang=\"en\" xml:lang=\"en\">Lots of rain fell over the last 48 hours, but how much?  Here is a preliminary look.  For a more detailed list go here: <a href=\"https:\/\/t.co\/4FDJ80jb8q\">https:\/\/t.co\/4FDJ80jb8q<\/a> <a href=\"https:\/\/twitter.com\/hashtag\/cawx?src=hash&amp;ref_src=twsrc%5Etfw\">#cawx<\/a> <a href=\"https:\/\/twitter.com\/hashtag\/AtmosphericRiver?src=hash&amp;ref_src=twsrc%5Etfw\">#AtmosphericRiver<\/a> <a href=\"https:\/\/t.co\/DjCIH1E9fn\">pic.twitter.com\/DjCIH1E9fn<\/a><\/p><div>\n\u2014 NWS Bay Area (@NWSBayArea) <\/div><a href=\"https:\/\/twitter.com\/NWSBayArea\/status\/1085986398816088064?ref_src=twsrc%5Etfw\">January 17, 2019<\/a><\/blockquote>\n<blockquote class=\"twitter-tweet\">\n<p dir=\"ltr\" lang=\"en\" xml:lang=\"en\">Mudslide closes Hwy. 17 southbound in Santa Cruz Mountains <a href=\"https:\/\/t.co\/dpSfZ6xrpC\">https:\/\/t.co\/dpSfZ6xrpC<\/a> <a href=\"https:\/\/t.co\/k0JZMnOUs9\">pic.twitter.com\/k0JZMnOUs9<\/a><\/p><div>\n\u2014 SFGate (@SFGate) <\/div><a href=\"https:\/\/twitter.com\/SFGate\/status\/1085709933654421504?ref_src=twsrc%5Etfw\">January 17, 2019<\/a><\/blockquote>\n<blockquote class=\"twitter-tweet\">\n<p dir=\"ltr\" lang=\"en\" xml:lang=\"en\">Residents along Hwy 35 in the Santa Cruz mountains say this storm has created major problems - lots of people stranded. The latest at 4 &amp; 6pm. <a href=\"https:\/\/t.co\/daqxt7VwUp\">https:\/\/t.co\/daqxt7VwUp<\/a> <a href=\"https:\/\/twitter.com\/hashtag\/abc7now?src=hash&amp;ref_src=twsrc%5Etfw\">#abc7now<\/a> <a href=\"https:\/\/t.co\/YpCUyGoKYz\">pic.twitter.com\/YpCUyGoKYz<\/a><\/p><div>\n\u2014 David Louie (@abc7david) <\/div><a href=\"https:\/\/twitter.com\/abc7david\/status\/1085989896399970304?ref_src=twsrc%5Etfw\">January 17, 2019<\/a><\/blockquote>\n<blockquote class=\"twitter-tweet\">\n<p dir=\"ltr\" lang=\"en\" xml:lang=\"en\"><a href=\"https:\/\/twitter.com\/RobMayeda?ref_src=twsrc%5Etfw\">@RobMayeda<\/a> :  Santa Cruz Mountains are getting hammered.  Rain rates of 2.13\u201d per hour just now.  3.46\u201d since midnight. <a href=\"https:\/\/t.co\/ACeqH8eZAe\">pic.twitter.com\/ACeqH8eZAe<\/a><\/p><div>\n\u2014 Ricardo Cortes (@RicardoDCortes) <\/div><a href=\"https:\/\/twitter.com\/RicardoDCortes\/status\/1085773214309961733?ref_src=twsrc%5Etfw\">January 17, 2019<\/a><\/blockquote>\n        \n  \n          \n  \n         \n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Friday, January 11, 2019 at 9:11AM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>Wake up! It's HighScalability time:<\/p>&#13;\n<div style=\"margin: auto;\"><iframe width=\"500\" height=\"300\" src=\"https:\/\/www.youtube.com\/embed\/zDAYZU4A3w0\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>&#13;\n&#13;\n<p>The modern day inner sanctum revealed for all to experience. Nausea no extra charge.<\/p>&#13;\n<\/div>&#13;\n&#13;\n<p>Do you like this sort of <em>Stuff? <\/em>Pleasesupport me\u00a0<a href=\"https:\/\/www.patreon.com\/toddhoff\">on Patreon<\/a>. Need cloud? Consume\u00a0<a href=\"https:\/\/www.amazon.com\/dp\/B0765C4SNR\">Explain the Cloud Like I'm 10<\/a>\u00a0(35 nearly 5 star reviews).<\/p>&#13;\n&#13;\n<ul><li><a href=\"https:\/\/v8.dev\/blog\/fast-async\">8x<\/a>: V8 Promiss.all parallel performance improvement; \u00a0<a href=\"https:\/\/www.publishersweekly.com\/pw\/by-topic\/industry-news\/financial-reporting\/article\/78929-print-unit-sales-increased-1-3-in-2018.html\">1.3%<\/a>: print sales increase; <a href=\"https:\/\/www.theverge.com\/2019\/1\/9\/18174631\/old-people-fake-news-facebook-share-nyu-princeton\">11%<\/a>: over 65 shared a hoax;<a href=\"https:\/\/www.techrepublic.com\/article\/40-of-companies-are-adding-jobs-after-deploying-ai-not-killing-them\/\"> 40%<\/a>: add jobs after deploying AI; <a href=\"https:\/\/blog.geteventbot.com\/2019\/01\/06\/how-we-support-open-source-software.html\">3%<\/a>: Eventbot's revenue pledged to open source; <a href=\"https:\/\/twitter.com\/etherchain_org\/status\/1082329360948969472\">51%<\/a>: successful Ethereum attack;\u00a0<a href=\"https:\/\/www.crypto51.app\/\">$308,620<\/a>: cost of a Bitcoin 51% attack; <a href=\"https:\/\/twitter.com\/asymco\/status\/1082296258797232128\">$30<\/a>: Apple services revenue per device per year; <a href=\"https:\/\/www.wired.com\/story\/i-sold-my-data-for-crypto\">.3 cents<\/a>: earnings from selling private data;\u00a0<a href=\"https:\/\/twitter.com\/tom_antonov\/status\/1081503245942558720\">2,000<\/a>:\u00a0baguettes a day produced on French aircraft carrier;\u00a0<a href=\"https:\/\/spectrum.ieee.org\/computing\/hardware\/lasers-vs-microwaves-the-billiondollar-bet-on-the-future-of-magnetic-storage\">5.6 nm<\/a>: future\u00a0smallest grains on a magnetic disk; <a href=\"http:\/\/www.beyondvc.com\/2019\/01\/boldstart-2018-recap-and-whats-hot-in-enterprise-2019.html\">11,000<\/a>: free books from 1923;\u00a0<\/li>\n<li>Quotable Quotes:                                            &#13;\n<ul><li><a href=\"https:\/\/twitter.com\/mekkaokereke\/status\/1081628836729503744\">@mekkaokereke<\/a>: He joined SpaceX as a \"founding employee.\" He designed the Merlin engine. He's CTO of Propulsion. His name is Tom Mueller. Everyone knows Elon Musk. No one knows Tom Mueller, even though Tom is the one currently designing a rocket that will put humans on Mars. \ud83e\udd37\ud83c\udfff\u200d\u2642\ufe0f<\/li>\n<li><a href=\"https:\/\/hackernoon.com\/interview-with-data-scientist-at-kaggle-dr-rachael-tatman-8bc61f9efdb9\">Dr. Rachael Tatman<\/a>: My universal advice is to not get a Ph.D. I even wrote a blog post about it a while ago. The blog\u2019s about linguistics specifically, but most of it applies to machine learning as well. I think that having a Ph.D. can be an advantage when you\u2019re looking for data science jobs, but unless you really want to 1) do research or 2) be a professor there\u2019s no really no benefit to getting a Ph.D. that you can can\u2019t get more quickly doing something else.<\/li>\n<li>&#13;\n<div id=\"_mcePaste\"><a href=\"https:\/\/twitter.com\/unclebobmartin\/status\/1081707790366920710?s=12\">@unclebobmartin<\/a>:\u00a0Unit tests are written by programmers for programmers. Acceptance tests are specified by customers for customers. Integration tests are written by architects for architects.<\/div>&#13;\n<\/li>\n<li><a href=\"https:\/\/twitter.com\/wrathofgnon\/status\/1083203432125612032\">@wrathofgnon<\/a>: A city can be seen as an attempt to solve the predicament that is human existence. No building can solve all problems, and as soon as we come up with what we think is a solution, we are faced with a new (often far worse) set of problems. Historically cities have been compromises.<\/li>\n<li><a href=\"http:\/\/cdixon.org\/2019\/01\/08\/strong-and-weak-technologies\/\">cdixon<\/a>:\u00a0Smartphones are a good example of a broader historical pattern: technologies usually arrive in pairs, a strong form and a weak form...But it\u2019s strong technologies that end up defining new eras...Weak technologies adapt to the world as it currently exists. Strong technologies adapt the world to themselves. Progress depends on strong technologies.<\/li>\n<li>This way to more quotes...&#13;\n<\/li><\/ul><\/li><\/ul><b>Don't miss all that the Internet has to say on Scalability, click below and become eventually consistent with all scalability knowledge<\/b> (which means this post has many more items to read so please keep on reading)...\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2019\/1\/11\/stuff-the-internet-says-on-scalability-for-january-11th-2019.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Tuesday, January 8, 2019 at 11:31AM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>In Part I, we introduced a High Availability (HA) framework for MySQL hosting and discussed various components and their functionality. Now in Part II, we will discuss the details of MySQL semisynchronous replication and the related configuration settings that help us ensure redundancy and consistency of the data in our HA setup. Make sure to check back in for Part III where we will review various failure scenarios that could arise and the way the framework responds and recovers from these conditions.&#13;\n&#13;\nWhat is MySQL Semisynchronous Replication?&#13;\nSimply put, in a MySQL semisynchronous replication configuration, the master commits transactions to the storage engine only after receiving acknowledgement from at least one of the slaves. The slaves would provide acknowledgement only after the events are received and copied to the relay logs and also flushed to the disk. This guarantees that for all transactions committed and returned to the client, the data exists on at least 2 nodes. The term \u2018semi\u2019 in semisynchronous (replication) is due to the fact that the master commits the transactions once the events are received and flushed to relay log, but not necessarily committed to the data files on the slave. This is in contrast to fully synchronous replication, where the transaction would have been committed on both the slave and the master before the session returns to the client.<\/p>\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2019\/1\/8\/mysql-high-availability-framework-explained-part-ii.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Tuesday, January 8, 2019 at 10:44AM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>Have you been experiencing slow MySQL startup times in GTID mode? We recently ran into this issue on one of our\u00a0<a title=\"MySQL Hosting\" href=\"https:\/\/scalegrid.io\/mysql\/products.html\" target=\"_blank\">MySQL hosting<\/a>\u00a0deployments and set out to solve the problem. In this blog, we\u00a0break down the issue that could be slowing down your MySQL restart times, how to debug for your deployment, and what you can do to decrease your start time and improve your understanding of\u00a0<a href=\"https:\/\/dev.mysql.com\/doc\/refman\/5.7\/en\/replication-gtids.html\" target=\"_blank\">GTID-based replication<\/a>.<\/p>&#13;\n<h2>How We Found The Problem<\/h2>\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2019\/1\/8\/slow-mysql-start-time-in-gtid-mode-binary-log-file-size-may.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Friday, January 4, 2019 at 2:48PM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>This\u00a0post\u00a0was written by Wendy Dessler of\u00a0<a title=\"The Blog Frog\" href=\"http:\/\/www.theblogfrog.com\/\" target=\"_blank\">The Blog Frog<\/a>.<\/p>&#13;\n<p><a href=\"https:\/\/scalegrid.io\/\" target=\"_blank\">Database-as-a-Service<\/a>\u00a0(DBaaS) is quickly gaining in popularity across the tech world. These software platform solutions helps users easily manage their database operations without having to really understand any of the abstractions. This allows developers, DBA\u2019s and DevOps engineers to quickly automate their backups, create new SQL and NoSQL clusters, and monitor the performance of their databases for their application without requiring any internal database expertise.<\/p>&#13;\n<p>DBaaS falls under the umbrella of\u00a0<a href=\"https:\/\/scalegrid.io\/blog\/the-future-of-the-application-stack\/\" target=\"_blank\">Platform-as-a-Service<\/a>\u00a0(PaaS) where the platform itself is actually a database or several databases. This is a great choice for DevOps in particular because it allows for more developer agility, productivity, and also security.<\/p>&#13;\n<p>Flexibility and scalability are becoming more important in the world of DevOps and technology in general, and we all know how fast this world moves. Businesses need new ways to keep up with the competition, and developers are looking for an easy, self-service model for managing their databases in order to optimize their app development. Let\u2019s break down the individual benefits so you can <a href=\"https:\/\/scalegrid.io\/blog\/how-devops-should-use-dbaas-database-as-a-service-to-optimize-their-application-development%E2%80%8B\/\">decide if DBaaS is right for your DevOps team<\/a>.<\/p>&#13;\n<h2>1. Outsourced Security and Administration<\/h2>\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2019\/1\/4\/how-devops-should-use-dbaas-database-as-a-service-to-optimiz.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Friday, January 4, 2019 at 9:03AM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>Wake up! It's HighScalability time:<\/p>&#13;\n<div style=\"margin: auto;\"><img src=\"https:\/\/c1.staticflickr.com\/5\/4827\/46549266762_c41629319b.jpg\" alt=\"\" \/><p>Solar system? Nope, the beauty is in your head\u2014<a href=\"https:\/\/www.forbes.com\/sites\/fionamcmillan\/2018\/12\/30\/where-to-find-some-of-the-most-stunning-neuroscience-images-of-2018\">neural art<\/a>.<\/p>&#13;\n<\/div>&#13;\n&#13;\n<p>Do you like this sort of <em>Stuff? <\/em>Pleasesupport me\u00a0<a href=\"https:\/\/www.patreon.com\/toddhoff\">on Patreon<\/a>. Need cloud? <a href=\"https:\/\/www.amazon.com\/dp\/B0765C4SNR\">Explain the Cloud Like I'm 10<\/a>\u00a0(34 almost 5 star reviews).<\/p>&#13;\n&#13;\n<ul><li><a href=\"http:\/\/cidrdb.org\/cidr2019\/papers\/p117-kraska-cidr19.pdf\">45%<\/a>: learned scheduler improves average job completion time;\u00a0<a href=\"https:\/\/twitter.com\/f_kaltheuner\/status\/1079147147361247237\">61%<\/a>: apps share data with Facebook; <a href=\"https:\/\/twitter.com\/netflixfilm\/status\/1078735051406204928\">45,037,125<\/a>: people who watched Bird Box on Netflix in first week; <a href=\"https:\/\/dominikschmidt.xyz\/mars32k\/\">32,368<\/a>:\u00a0color images collected by the Curiosity rover on Mars between August 2012 and November 2018;\u00a0<a href=\"https:\/\/www.beckershospitalreview.com\/artificial-intelligence\/ai-in-healthcare-market-to-grow-to-36-1b-by-2025.html\">$36.1B<\/a>:\u00a0AI \u00a0healthcare market by 2025; <a href=\"https:\/\/www.datanami.com\/2019\/01\/02\/study-pours-cold-water-on-ai-driving-algorithms\/\">20%<\/a>:\u00a0object recognition failure in light rain;\u00a0<a href=\"https:\/\/thenewstack.io\/donald-knuths-christmas-tree-lecture-on-dancing-links-and-organ-music\/\">350<\/a>: pages in Donald Knuth's new book;\u00a0\u00a0<a href=\"https:\/\/arstechnica.com\/science\/2018\/12\/economists-calculate-the-true-value-of-facebook-to-its-users-in-new-study\/\">$1,000<\/a>: price needed to deactivate Facebook account; <a href=\"https:\/\/techcrunch.com\/2018\/12\/27\/epic-fortnite-3-billion-profit\/\">$3B<\/a>: Epic games profit; <a href=\"https:\/\/motherboard.vice.com\/en_us\/article\/vby7ny\/bitcoin-body-heat-mining\">1<\/a>: bitcoin mined from body heat of 44,000 people;\u00a0<\/li>\n<li>Quotable Quotes:                                                    &#13;\n<ul><li><a href=\"https:\/\/twitter.com\/Hannah_Chutzpah\/status\/1074391327629041670\">@Hannah_Chutzpah<\/a>: What are the technical terms, in your field, for 'dunno'? In medicine there's 'idoeopathic' In archeology\/anthropology there's 'ritual purposes' How do you professionally term 'we haven't got a clue'?                                                    &#13;\n<ul><li><a href=\"https:\/\/twitter.com\/peterseibel\/status\/1076875719370653696\">@peterseibel<\/a>: In programming we don't have a special term for it, just like fish don't have a special term for 'water'.<\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/twitter.com\/robn\/status\/1080370433499623425\">@robn<\/a>:\u00a0everyone seems to want stateless services and stateless protocols but I have three kids that can't remember shit I told them ten seconds ago and believe me it's the worst and I don't know why you'd want that<\/li>\n<li><a href=\"https:\/\/twitter.com\/stranter\/status\/1080499048065523713\">@stranter<\/a>: \"A machine learning agent intended to transform aerial images into street maps and back was found to be cheating by hiding information it would need later in \u201ca nearly imperceptible, high-frequency signal\" || \u00a0Sounds like the algo needs some parenting<\/li>\n<li><a href=\"https:\/\/twitter.com\/pmddomingos\/status\/1077291822647525378\">@pmddomingos<\/a>: Overheard at Stanford: trying to run a university in Silicon Valley is like trying to run a monastery in Gomorrah.<\/li>\n<li><a href=\"https:\/\/twitter.com\/scalzi\/status\/1077255285721440257\">@scalzi<\/a>: So, things I've learned in the last few days: 1. Sometimes 15GB+ updates to video games are in fact necessary and unavoidable 2. Devs mostly hate 'em too but right now there's not much to be done about it 3. I still hate it but UGH OKAY FINE I guess<\/li>\n<li><a href=\"https:\/\/twitter.com\/copyconstruct\/status\/1079269975196463105\">@copyconstruct<\/a>:\u00a0\"GPUs will increase 1000\u00d7 in performance by 2025, whereas Moore\u2019s law for CPUs essentially is dead. By replacing branch-heavy algorithms with neural networks, the DBMS can profit from these hardware trends.\"<\/li>\n<li><a href=\"https:\/\/twitter.com\/hotjamz\/status\/1077540553049411585\">@hotjamz<\/a>: he's makin a list \/ it's kept in plain text \/ an elf just got phished \/ you know what is next \/ santa claus has leaked 2.2 billion user's names, addresses, naughty\/nice statuses, and the times they are asleep in their homes<\/li>\n<li>So many more quotes...&#13;\n<\/li><\/ul><\/li><\/ul><b>Don't miss all that the Internet has to say on Scalability, click below and become eventually consistent with all scalability knowledge<\/b> (which means this post has many more items to read so please keep on reading)...\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2019\/1\/4\/stuff-the-internet-says-on-scalability-for-january-4th-2019.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Friday, December 21, 2018 at 8:58AM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>Wake up! It's HighScalability time:<\/p>&#13;\n<div style=\"margin: auto;\"><img src=\"https:\/\/c1.staticflickr.com\/5\/4826\/31458477807_aca616a4e5.jpg\" alt=\"\" \/><p>Have a <a href=\"https:\/\/medium.com\/square-corner-blog\/scaling-christmas-an-illustrated-adventure-a2ea739f5451\">very scalable<\/a> Xmas everyone! See you in the New Year.<\/p>&#13;\n<\/div>&#13;\n&#13;\n<p>Do you like this sort of <em>Stuff? <\/em>Pleasesupport me\u00a0<a href=\"https:\/\/www.patreon.com\/toddhoff\">on Patreon<\/a>. I'd really appreciate it. Still looking for that perfect xmas gift? What could be better than a book on the cloud? <a href=\"https:\/\/www.amazon.com\/dp\/B0765C4SNR\">Explain the Cloud Like I'm 10<\/a>. And if you know someone with hearing problems they might find <a href=\"http:\/\/livecc.info\">Live CC<\/a>\u00a0useful.<\/p>&#13;\n&#13;\n<ul><li><a href=\"https:\/\/www.pornhub.com\/insights\/2018-year-in-review\">33.5 billion<\/a>: Pornhub visits;\u00a0<a href=\"https:\/\/www.popsci.com\/science\/article\/2012-12\/fyi-how-long-would-it-take-santa-deliver-presents-every-kid-earth\">122 million<\/a>: miles traveled by Santa;\u00a0<a href=\"https:\/\/www.cnet.com\/news\/apple-saw-more-government-requests-for-user-data-in-the-first-half-of-2018-changes-transparency-report\/#ftag=CAD-09-10aai5b\">32,342<\/a>: government requests to Apple for user data; <a href=\"http:\/\/www.impactlab.net\/2018\/12\/12\/bell-says-latest-helicopter-was-designed-10-times-faster-with-vr\/\">10x<\/a>: faster helicopter design using VR instead of\u00a0physical models and mockups;\u00a0<a href=\"https:\/\/www.pornhub.com\/insights\/2018-year-in-review\">4403<\/a>: petabytes transferred by Pornhub;<a href=\"https:\/\/kinsta.com\/blog\/disable-google-amp\/\"> 59%<\/a>: dropped leads on Google AMP;\u00a0<a href=\"https:\/\/www.indiewire.com\/2018\/12\/streaming-shows-outnumber-broadcast-cable-1202027973\/\">160<\/a>:\u00a0streaming shows now outnumber their traditional-TV counterparts; <a href=\"https:\/\/medium.com\/@aaronedell\/80-of-all-machine-learning-engineers-work-at-google-or-facebook-6a8e315436e8\">80%<\/a>:\u00a0machine learning engineers work at Google or Facebook; <a href=\"https:\/\/www.studyfinds.org\/survey-quarter-checks-phones-less-than-minute-after-waking\/\">25%<\/a>: adults check phone immediately on waking;\u00a0<a href=\"https:\/\/www.businessinsider.com\/apple-app-store-164-companies-had-first-million-year-2018-2018-12\">164<\/a>: iPhone apps made $1 million through\u00a0in-app subscriptions; <a href=\"https:\/\/www.backblaze.com\/blog\/2018-in-the-rear-view-mirror\/\">750 petabytes<\/a>: Backblaze storage;\u00a0<\/li>\n<li>Quotable Quotes:  &#13;\n<ul><li><a href=\"https:\/\/twitter.com\/flightradar24\/status\/1013088775973556224\">@ flight radar24<\/a>: Yesterday was the busiest day of the year in the skies so far and our busiest day ever. 202,157 flights tracked! The first time we've tracked more than 200,000 flights in a single day on\u00a0<\/li>\n<li><a href=\"https:\/\/twitter.com\/swardley\/status\/1073933737598046209\">@swardley: X<\/a> : What's going to happen in cloud in 2019? Me : Nothing special. 1) Enterprise data centres will continue to close. 2) Cloud will decentralise in terms of provision not power i.e. Amazon will \"invade\" more of those holdouts with AWS Outpost. 3) Serverless will rocket.<\/li>\n<li><a href=\"https:\/\/twitter.com\/odrotbohm\/status\/1073658331703795713\">@odrotbohm<\/a>:\u00a0I\u2019ve seen microservice based systems more deranged after 2 years than any 1,5 decades old monolith could ever have been.<\/li>\n<li><a href=\"https:\/\/www.wired.com\/story\/microsoft-undersea-data-server-scotland\/\">Jason Kehe<\/a>: Savings of even a few degrees Celsius can significantly extend the lifespan of electronic components; Microsoft reports that, on the ocean floor 117 feet down, its racks stay 10 degrees cooler than their land-based counterparts. Half a year after deployment, \u201cthe equipment is happy,\u201d says Ben Cutler, the project\u2019s manager.<\/li>\n<li><a href=\"https:\/\/twitter.com\/kellabyte\/status\/1073619884309790721\">@kellabyte<\/a>:\u00a0\u201cOpen source\u201d infrastructure companies are a giant shit show right now. Whether it\u2019s database or message queues it\u2019s a really weird combo of licenses and features for hostage. Why? Because nobody knows how to make money. Support contracts aren\u2019t enough.<\/li>\n<li><a href=\"https:\/\/www.tbray.org\/ongoing\/When\/201x\/2018\/12\/14\/SF-4\">Tim Bray<\/a>: How to talk about [Serverless Latency] \u00b7 To start with, don\u2019t just say \u201cI need 120ms.\u201d Try some\u00adthing more like \u201cThis has to be in Python, the data\u2019s in Cas\u00adsan\u00addra, and I need the P50 down un\u00adder a fifth of a sec\u00adond, ex\u00adcept I can tol\u00ader\u00adate 5-second la\u00adten\u00adcy if it doesn\u2019t hap\u00adpen more than once an hour.\u201d And in most main\u00adstream ap\u00adpli\u00adca\u00adtion\u00ads, you should be able to get there with server\u00adless. If you plan for it.\u00a0<\/li>\n<li><a href=\"https:\/\/www.jeremydaly.com\/newsletter-issue-16\/\">Jeremy Daly<\/a>:\u00a0Compared to building and maintaining your own systems, cloud computing is ridiculously inexpensive, especially when you\u2019re starting out and haven\u2019t achieved significant scale. Don\u2019t waste your developers\u2019 time trying to shave off nickels and dimes from your bill. Focus on creating more value by delivering and iterating on features faster and worry about cost optimizations later.<\/li>\n<li><a href=\"https:\/\/techcrunch.com\/2018\/12\/15\/the-business-case-for-serverless\/\">Zack Kanter<\/a>: One day, complexity will grow past a breaking point and development velocity will begin to decline irreversibly, and so the ultimate job of the founder is to push that day off as long as humanly possible. The best way to do that is to keep your ball of mud to the minimum possible size\u2014 serverless is the most powerful tool ever developed to do exactly that.<\/li>\n<li> These are the last quotes of the year. Don't be late...&#13;\n<\/li><\/ul><\/li><\/ul><b>Don't miss all that the Internet has to say on Scalability, click below and become eventually consistent with all scalability knowledge<\/b> (which means this post has many more items to read so please keep on reading)...\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2018\/12\/21\/stuff-the-internet-says-on-scalability-for-december-21st-201.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Tuesday, December 18, 2018 at 1:35PM <\/div> \n    \n  \n      \n      <div class=\"body\">\n  \n        \n        \n        \n          <p><img src=\"http:\/\/farm3.static.flickr.com\/2363\/5781449767_54902da8a5_o.jpg\" alt=\"\" style=\"float: right;\" \/><\/p>\n<h2>Who's Hiring?<\/h2>\n<ul><li>\n<p>pMD is a<a href=\"https:\/\/www.pmd.com\/news\/pmd-ranked-among-the-fastest-growing-private-companies-in-the-bay-area-by-the-san-francisco-business-times\">\u00a0fast\u00ad growing<\/a>,\u00a0<a href=\"https:\/\/www.glassdoor.com\/Reviews\/pMD-Reviews-E498449.htm\">highly rated<\/a>\u00a0health care technology company that has been recognized as a Best Place to Work by SF Business Times, Modern Healthcare, and Inc.\u00a0<strong>Senior DevOps Engineer<\/strong>: Your engineering work will focus on using your deep knowledge of the web stack including firewalls, web applications, caches and data stores \u00ad to create innovative infrastructure architectures that are resilient, scalable, and blazingly fast. You will bring best practices for automating the provisioning and patching of servers, replicating and versioning configuration changes, monitoring and alerting, deploying application and data store updates, and high availability \/ disaster recovery. <a href=\"https:\/\/www.pmd.com\/careers?gh_jid=1267819\">Please apply here.<\/a><\/p>\n<\/li>\n<li><strong>Triplebyte<\/strong> lets exceptional software engineers skip screening steps at hundreds of top tech companies like <strong>Apple, Dropbox, Mixpanel, and Instacart<\/strong>. Make your job search <em>O<\/em>(1), not <em>O<\/em>(<em>n<\/em>). <a href=\"https:\/\/triplebyte.com\/?ref=hs_apr18_c2\">Apply here.<\/a> <\/li>\n<li><strong>Need excellent people? Advertise your job here!<\/strong>\u00a0<\/li>\n<\/ul><h2>Fun and Informative Events<\/h2>\n<ul><li><strong>Advertise your event here!<\/strong><\/li>\n<\/ul><h2>Cool Products and Services<\/h2>\n<ul><li><strong>InMemory.Net<\/strong>\u00a0provides a <strong>Dot Net native in memory database<\/strong> for analysing large amounts of data. It runs natively on .Net, and provides a native .Net, COM &amp; ODBC apis for integration. It also has an easy to use language for importing data, and supports standard SQL for querying data.\u00a0<a href=\"http:\/\/inmemory.net\/\">http:\/\/InMemory.Net<\/a><\/li>\n<\/ul><ul><li><strong>For heads of IT\/Engineering responsible for building an analytics infrastructure<\/strong>, <a href=\"https:\/\/etleap.com\/?utm_source=high-scalability&amp;utm_medium=sponsored-link&amp;utm_campaign=general\">Etleap is an ETL solution<\/a> for creating perfect data pipelines from day one. Unlike older enterprise solutions, Etleap doesn\u2019t require extensive engineering work to set up, maintain, and scale. It automates most ETL setup and maintenance work, and simplifies the rest into 10-minute tasks that analysts can own. <a href=\"https:\/\/etleap.com\/customers\/?utm_source=high-scalability&amp;utm_medium=sponsored-link&amp;utm_campaign=general\">Read stories from customers like Okta and PagerDuty<\/a>, or <a href=\"https:\/\/etleap.com\/?utm_source=high-scalability&amp;utm_medium=sponsored-link&amp;utm_campaign=general\">try Etleap yourself<\/a>.<\/li>\n<\/ul><ul><li>Build, scale and personalize your\u00a0<a href=\"https:\/\/getstream.io\/try-the-api\/?utm_source=hs&amp;utm_campaign=hs&amp;utm_medium=hs\">news feeds and activity streams with getstream.io<\/a>. Try the API now in this\u00a0<a href=\"https:\/\/getstream.io\/try-the-api\/?utm_source=hs&amp;utm_campaign=hs&amp;utm_medium=hs\">5 minute interactive tutorial<\/a>.\u00a0<strong>Stream\u00a0<\/strong>is free up to 3 million feed updates so it's easy to get started. Client libraries are available for Node, Ruby, Python, PHP, Go, Java and .NET. Stream is currently also hiring Devops and Python\/Go developers in Amsterdam. More than 400 companies rely on Stream for their production feed infrastructure, this includes apps with 30 million users. With your help we'd like to ad a few zeros to that number. Check out the job opening on\u00a0<a href=\"https:\/\/angel.co\/stream\">AngelList<\/a>.<\/li>\n<\/ul><ul><li><a href=\"https:\/\/www.scalyr.com\/\">Scalyr<\/a> is a lightning-fast <a href=\"https:\/\/www.scalyr.com\/product\/centralized-log-management\">log management<\/a> and operational data platform. \u00a0It's a tool (actually, multiple tools) that your entire team will love. \u00a0Get visibility into your production issues without juggling multiple tabs and different services -- all of your logs, server metrics and alerts are in your browser and at your fingertips. . \u00a0Loved and used by teams at Codecademy, ReturnPath, Grab, and InsideSales. Learn more today or see why Scalyr is a great <a href=\"https:\/\/www.scalyr.com\/product\/splunk-alternative\">alternative to Splunk<\/a>.<\/li>\n<li><strong>Advertise your product or service here!<\/strong><\/li>\n<\/ul><p><strong><em> If you are interested in a sponsored post for an event, job, or product, please <a href=\"http:\/\/highscalability.com\/contact\">contact<\/a> us for more information.<\/em><\/strong><\/p>\n<hr \/><h2>pMd is Hiring<\/h2>\n<p>pMD is a<a href=\"https:\/\/www.pmd.com\/news\/pmd-ranked-among-the-fastest-growing-private-companies-in-the-bay-area-by-the-san-francisco-business-times\"> fast\u00ad growing<\/a>, <a href=\"https:\/\/www.glassdoor.com\/Reviews\/pMD-Reviews-E498449.htm\">highly rated<\/a> health care technology company that has been recognized as a Best Place to Work by SF Business Times, Modern Healthcare, and Inc. We're profitable, have extremely happy customers, and make up a team of people as talented and passionate as you are. We love what we do and care about doing good in the world.<\/p>\n<p><strong> Senior DevOps Engineer<\/strong>: Your engineering work will focus on using your deep knowledge of the web stack \u00ad including firewalls, web applications, caches and data stores \u00ad to create innovative infrastructure architectures that are resilient, scalable, and blazingly fast. You will bring best practices for automating the provisioning and patching of servers, replicating and versioning configuration changes, monitoring and alerting, deploying application and data store updates, and high availability \/ disaster recovery. You have a vision for how a best\u00adin\u00adthe\u00adworld web application stack should run, and you're excited to execute on that vision.<\/p>\n<p>At pMD you\u2019ll wear many hats, have a lot of responsibility, and be part of a fun, smart, and creative team. Our team is small and collaborative, so you\u2019ll experience rapid career growth through real mentorship. You'll be a name (or a nickname) on the team and at social events, rather than Engineer #300. You'll get opportunities to be involved in sales, recruiting, customer success, and other aspects of the business. As the company continues to grow, you're ready to start mentoring others and eventually build and lead a team. At pMD you can grow as quickly as you want to.<\/p>\n<p>Where else would you have a chance to architect a new replication topology in the morning, meet with clients for lunch, build a bare metal server in the afternoon, and drink beers with the team after work? Plus, you'll be able to see the impact of your work immediately and you'll hear from our users how your efforts have improved their lives.<\/p>\n<p><a href=\"https:\/\/www.pmd.com\/careers?gh_jid=1267819\"> Please apply here. <\/a><\/p>\n<hr \/><h2>Make Your Job Search <em>O<\/em>(1) \u2014 not <em>O<\/em>(<em>n<\/em>)<\/h2>\n<p><a href=\"https:\/\/triplebyte.com\/?ref=hs_apr18_c3\">Triplebyte<\/a> is unique because they're a team of engineers running their own centralized technical assessment. Companies like <strong>Apple, Dropbox, Mixpanel, and Instacart<\/strong> now let Triplebyte-recommended engineers skip their own screening steps.<\/p>\n<p>We found that <strong>High Scalability readers<\/strong> are about 80% more likely to be in the top bracket of engineering skill.<\/p>\n<p>Take <a href=\"https:\/\/triplebyte.com\/?ref=hs_apr18_c3\">Triplebyte's multiple-choice quiz<\/a> (system design and coding questions) to see if they can help you scale your career faster.<\/p>\n<hr \/><h2>The Solution to Your Operational Diagnostics Woes<\/h2>\n<p><strong><a href=\"https:\/\/www.scalyr.com\/\">Scalyr<\/a><\/strong> gives you instant visibility of your production systems, helping you turn chaotic logs and system metrics into actionable data at interactive speeds. Don't be limited by the slow and narrow capabilities of traditional <a href=\"https:\/\/www.scalyr.com\/\">log monitoring<\/a> tools. View and analyze all your logs and system metrics from multiple sources in one place. Get enterprise-grade functionality with sane pricing and insane performance. <a href=\"https:\/\/www.scalyr.com\/\">Learn more today<\/a>.\u00a0<\/p>\n<hr \/><p><em> If you are interested in a sponsored post for an event, job, or product, please <\/em><em><a href=\"http:\/\/highscalability.com\/contact\">contact<\/a> us for more information.<\/em><\/p>\n        \n  \n          \n  \n         \n  \n      <\/div><\/div><div class=\"journal-entry-text\"><div class=\"journal-entry-tag journal-entry-tag-post-title\">   <img title=\"date\" alt=\"date\" class=\"inline-icon date-icon\" src=\"http:\/\/highscalability.com\/universal\/images\/transparent.png\" \/>Friday, December 14, 2018 at 9:11AM <\/div> \n    \n  \n     \n      <div class=\"body\">\n  \n        \n  \n         \n  \n        <p>Wake up! It's HighScalability time:<\/p>&#13;\n<iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/JQ8ZiT1sn88\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe>&#13;\n<p>We've come a long way in 50 years. Or have we?<\/p>&#13;\n<p><a href=\"https:\/\/www.amazon.com\/Valley-Genius-Uncensored-History-Founders-ebook\/dp\/B0769XXGXX\/\">Alan Kay<\/a>:\u00a0I believe ARPA spent $ 175,000 of 1968 money for that one demo. That\u2019s probably like a million bucks today.<\/p>&#13;\n<p><a href=\"https:\/\/www.amazon.com\/Valley-Genius-Uncensored-History-Founders-ebook\/dp\/B0769XXGXX\/\">Bill English<\/a>:\u00a0What we did was lease two video circuits from the phone company. They set up a microwave link: two transmitters on the top of the building at SRI, receiver\/ transmitters up on Skyline Boulevard on a truck, and two receivers at the Civic Center. Cables of course going down into the room at both ends. That was our video link. Going back we had two dedicated 1,200-baud lines: high-speed lines at the time. Homemade modems.<\/p>&#13;\n<p><a href=\"https:\/\/www.amazon.com\/Valley-Genius-Uncensored-History-Founders-ebook\/dp\/B0769XXGXX\/\">Doug Engelbart<\/a>:\u00a0It was the very first time the world had ever seen a mouse, seen outline processing, seen hypertext, seen mixed text and graphics, seen real-time videoconferencing.<\/p>&#13;\n<p><a href=\"https:\/\/www.amazon.com\/Valley-Genius-Uncensored-History-Founders-ebook\/dp\/B0769XXGXX\/\">Alan Kay<\/a>:\u00a0We could actually see that ideas could be organized in a different way, that they could be filtered in a different way, that what we were looking at was not something that was trying to automate current modes of thought, but that there should be an amplification relationship between us and this new technology.<\/p>&#13;\n&#13;\n<p>Do you like this sort of <em>Stuff? <\/em>Pleasesupport me\u00a0<a href=\"https:\/\/www.patreon.com\/toddhoff\">on Patreon<\/a>. I'd really appreciate it. Know anyone looking for a simple book explaining the cloud? Then please recommend my well reviewed (32 reviews on Amazon and 77 on Goodreads!)\u00a0book:\u00a0<a href=\"https:\/\/www.amazon.com\/dp\/B0765C4SNR\">Explain the Cloud Like I'm 10<\/a>.\u00a0They'll love it and you'll be their hero forever. And if you know someone with hearing problems they might find <a href=\"http:\/\/livecc.info\">Live CC<\/a> very useful.<\/p>&#13;\n&#13;\n<ul><li><a href=\"https:\/\/www.masterclass.com\/classes\/jimmy-chin-teaches-adventure-photography\">50,000<\/a>: images in a National Geographic shoot; <a href=\"https:\/\/twitter.com\/i\/moments\/1072360261376401408\">11 billion<\/a>: Voyager 2 miles traveled; <a href=\"http:\/\/cdn.aiindex.org\/2018\/AI%20Index%202018%20Annual%20Report.pdf\">83%<\/a>: AI papers originate outside the US; <a href=\"https:\/\/www.usenix.org\/system\/files\/osdi18-alquraan.pdf\">80%<\/a>: network partitions lead to catastrophic failures; <a href=\"https:\/\/www.youtube.com\/watch?v=oA04URWQXnM\">97%<\/a>: large AWS customers use auto scaling;\u00a0<a href=\"https:\/\/brex.com\/blog\/state-of-spend\">82%<\/a>: startup failures due to cashflow problems;\u00a0<a href=\"https:\/\/www.ecnmag.com\/news\/2018\/12\/robots-and-drones-spending-total-1157-billion-2019-says-report\">$115.7 billion<\/a>:\u00a0robot and drone spending in 2019;\u00a0<\/li>\n<li>Quotable Quotes:      &#13;\n<ul><li><a href=\"https:\/\/twitter.com\/jmarhee\/status\/1072210874344058881\">@jmarhee<\/a>:\u00a0\"I went to Kubecon and all I got were these 300 service meshes built by companies that went under by time I got home\"<\/li>\n<li><a href=\"https:\/\/twitter.com\/ReformedBroker\/status\/1071852999004352513\">@ReformedBroker<\/a>: \u201cRetention is the new growth.\u201d - CEO of Adobe ($ADBE), which has gained 793% since moving Photoshop etc to a subscription-based service from a one-time software sale in 2011.<\/li>\n<li><a href=\"https:\/\/twitter.com\/mattray\/status\/1072651885159571456\">@mattray<\/a>:\u00a0OH: \"Cloud native is pretty simple. You just need to know Kubernetes, Prometheus, Fluentd, Jaeger, Envoy, Core DNS, Linkerd, Rook, Vitess, Etcd and Raft.\"<\/li>\n<li><a href=\"https:\/\/twitter.com\/davidfrum\/status\/1071105030340648960\">@davidfrum<\/a>: In order to lose 80% of its value, the bitcoin network expended more electricity than 150 of the world's 195 countries<\/li>\n<li><a href=\"https:\/\/twitter.com\/chappelltracker\/status\/1071099239353249792\">@ChappellTracker<\/a>: So here's something. Users are testing the limits of Tumblr's new algorithm that flags adult content (aka \"censorbot\"). This one found that a man's chest was flagged, but a man's chest next a 50% scale owl went unnoticed.\u00a0<\/li>\n<li><a href=\"https:\/\/www.masterclass.com\/classes\/jimmy-chin-teaches-adventure-photography\">Jimmy Chin<\/a>: Great editors are brutal.<\/li>\n<li><a href=\"https:\/\/twitter.com\/suzatweet\/status\/1070653439468802050\">@danbri<\/a>:\u00a0intro slide: \"Main message is that biology has been computing long before brains evolved. Somatic decision-making and memory are mediated by ancient pre-neural bioelectric networks across all cells. Exploring non-neural Cognition is [an untapped frontier for AI...]\"<\/li>\n<li><a href=\"https:\/\/twitter.com\/swardley\/status\/1072444871024218112\">@swardley<\/a>: X : Large companies are leaving cloud and building internally. If they use it, then it will be hybrid. Serverless is for startups who can't afford their own.\u00a0&#13;\n<div>Me : Hmmm. The only words of comfort I have are ... if you don't like change, you're going to hate irrelevance.<\/div>&#13;\n<\/li><li>Quotes, get'em while they're hot...&#13;\n<\/li><\/ul><\/li><\/ul><b>Don't miss all that the Internet has to say on Scalability, click below and become eventually consistent with all scalability knowledge<\/b> (which means this post has many more items to read so please keep on reading)...\n  \n        <p class=\"journal-read-more-tag\"><a href=\"http:\/\/highscalability.com\/blog\/2018\/12\/14\/stuff-the-internet-says-on-scalability-for-december-14th-201.html\">Click to read more ...<\/a><\/p>\n  \n      <\/div><\/div>","created_at":"2019-01-21T23:46:31+0000","updated_at":"2019-01-21T23:46:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":23,"domain_name":"highscalability.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["keyboard"],"is_public":false,"id":786,"title":"Typewriter Inspired Mechanical Keyboard\u2122","url":"https:\/\/www.qwerkywriter.com\/","content":"<p><strong>Bluetooth Wireless Connectivity Up To 3 Devices<\/strong> - The Qwerkywriter\u00ae S can be connected to multiple wireless devices at the same time. \u00a0So now you can have a mobile phone, a tablet, and a laptop all connected wirelessly to the Qwerkywriter\u00ae S.<\/p><p><strong>Wired USB mode with NKRO support<\/strong> - Made for professionals and type-warriors, the Qwerkywriter\u00ae S\u00a0 can perform at the highest levels without dropping letters or strokes.<\/p><p><strong>Dual Scroll Knob Encoders<\/strong> - The Qwerkywriter\u00ae S features dual function scroll knobs that control Volume on the left knob and Mouse Scroll on the right knob.*<\/p><p><strong>German Engineered Mechanical Switches<\/strong> - At the heart of the new Qwerkywriter\u00ae S is the\u00a0German engineered Cherry MX \"clicky\" switches for unrivaled industry leading\u00a0tactile feedback and performance.<\/p><p>*Vertical Scrolling on iOS is not supported at this time.<\/p>","created_at":"2019-01-21T21:47:41+0000","updated_at":"2019-01-21T21:47:42+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.qwerkywriter.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":785,"title":"Make Something That Makes Something","url":"http:\/\/www.procjam.com\/seeds\/issues\/3\/","content":"<div class=\"container\"><br \/><div class=\"jumbotron\"><p>\n    Welcome to Issue 3 of Seeds! If you've not seen it already, make sure you go and download the PDF version of the zine, with beautiful layout and design from Jupiter and Dann, as well as some bonus content! You can <a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/issue3.pdf\">download it direct from us<\/a> or <a href=\"https:\/\/issuu.com\/jupiterhadley\/docs\/seeds_issue_3__2_\">read it online over on Issuu<\/a>. Our new digital edition, which you're reading now, is great for linking to specific articles or finding stuff on the go, but we recommend reading the zine itself to get the full Seeds experience! We hope you enjoy this new issue!\n    <\/p><p>Background image from Martin O'Leary's article, <a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/mewo.txt.html\">A Year of Daily Generative Sketches<\/a>.<\/p><h3>Issue 3 (2018)<\/h3><ul><li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/ai.txt.html\">Writing Interesting AI<\/a> by <i>Frank Mitchell<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/amin.txt.html\">Intelligent Middle-Level Game Control<\/a> by <i>Amin Babadi<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/ant.txt.html\">Deciphering Generated Symbols as a Game Mechanic: The Design of \"Alien Transmission\"<\/a> by <i>Antonios Liapis<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/apop.txt.html\">Apop: Seeing Patterns Everywhere<\/a> by <i>Isaac Schankler<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/avoyd.txt.html\">Boxes in Space<\/a> by <i>Juliette Foucaut and Doug Binks<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/bugs.txt.html\">Bugs &amp; Beetles<\/a> by <i>Sabine Wieluch<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/cox.txt.html\">Procedural Generation in Twine<\/a> by <i>Dan Cox<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/data.txt.html\">Real-world Data as a Seed<\/a> by <i>Antonios Liapis<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/davide.txt.html\">Procedural Generation, Variety and Reproducibility<\/a> by <i>Davide Ciacco<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/gorm.txt.html\">On the Responsibility of Makers<\/a> by <i>Gorm Lai<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/guppy.txt.html\">Guppy - Procgen as Antidote to Development Boredom<\/a> by <i>Christiaan Moleman<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/hache.txt.html\">Letting Go of Realism<\/a> by <i>Guerric Hach\u00e9<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/hacker.txt.html\">From Hacker Poets to Cybernetic Poets<\/a> by <i>Terry Trowbridge and Joseph Alexander Brown<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/isaac.txt.html\">The Locus of a Generator<\/a> by <i>Issac Karth<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/isl.txt.html\">Islands Are Just Mountains Up To Their Necks in Ocean<\/a> by <i>Scott Turner<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/jewel.txt.html\">Yet Another Contemporary Jewellery Exhibition<\/a> by <i>Enrica Prazzoli<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/koz.txt.html\">High End Procedural Systems (Procedural Trip Report)<\/a> by <i>Denis Kozlov<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/kth.txt.html\">Carving the Infinite Plane into Discrete Regions<\/a> by <i>Keith Evans<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/lac.txt.html\">Mere Juxtaposition - A Lacunic Fiction of Permutations<\/a> by <i>Jasmine Otto<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/mark.txt.html\">Ultima Ratio Regum<\/a> by <i>Mark Johnson<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/mewo.txt.html\">A Year of Daily Generative Sketches<\/a> by <i>Martin O'Leary<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/minus.txt.html\">Minus World: Generative Game Reviews from a Parallel Universe<\/a> by <i>Max Kreminski<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/moo.txt.html\">Superorganism Art<\/a> by <i>Genetic Moo<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/mount.txt.html\">Mountscapes<\/a> by <i>Pol Clarissou<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/name.txt.html\">What's your NAME?<\/a> by <i>Yevhen Loza (EugeneLoza)<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/news.txt.html\">Making a Fantasy Newsroom Bot<\/a> by <i>Damien Crawford<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/nos.txt.html\">Nos Falaises<\/a> by <i>Guillaume Pelletier-Auger<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/not.txt.html\">not_constantinople<\/a> by <i>Seth Alter<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/npc.txt.html\">Towards Generating AI for Generative Worlds: Evolving Behavior Trees in Unreal Engine 4<\/a> by <i>Nathan Partlan and Magy Seif El-Nasr<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/odd.txt.html\">Oddifier: RPG Character Sheets<\/a> by <i>Paul McCann<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/pech.txt.html\">Experiments in Generative Geometry: Building Meshes in Unity, Vertex by Vertex<\/a> by <i>Alexander Pech<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/pgod.txt.html\">But What Do You *Do*: Mechanical Ideas for Turning a Cool Generator Into a Compelling Game<\/a> by <i>Prophet Goddess<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/rand.txt.html\">Game Randomizers: Procedural Variations on Your Nostalgia<\/a> by <i>Jo Mazeika<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/red.txt.html\">Story Generation by Algorithms: A First Attempt at a Digital Storyteller<\/a> by <i>Adam Riddle<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/rende.txt.html\">A Web-Based Node Image Editor<\/a> by <i>Johan Rende<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/rot.txt.html\">Robots on Typewriters<\/a> by <i>Justin Edwards and Allison Perrone<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/samim.txt.html\">Thoughts on the Generative, Creative Economy<\/a> by <i>Samim<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/spell.txt.html\">Procedurally Generated Spellbook Sprites<\/a> by <i>Spencer Egart<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/strata.txt.html\">Designing Strata: A 2D Level Generator for Non-Programmers<\/a> by <i>Matt Schell<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/til.txt.html\">Alien Growth - A 2D Plant Generator<\/a> by <i>Tilman Schmidt<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/tiny.txt.html\">All Dinosaurs Great and Small<\/a> by <i>Elle Sullivan<\/i><\/li>\n<li><a href=\"http:\/\/www.procjam.com\/seeds\/issues\/3\/writing.txt.html\">Representing Writing<\/a> by <i>Mark Rickerby<\/i><\/li>\n<\/ul><\/div><br \/><\/div>\n    \n    \n    <footer class=\"py-5\"><\/footer>","created_at":"2019-01-21T17:23:07+0000","updated_at":"2019-01-21T17:23:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.procjam.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":784,"title":"ryanoasis\/nerd-fonts","url":"https:\/\/github.com\/ryanoasis\/nerd-fonts","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><h2 style=\"text-align: center;\"><a id=\"user-content----1\" class=\"anchor\" aria-hidden=\"true\" href=\"#---1\"><\/a>\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/images\/project-subtitle-phrase.svg\"><img alt=\"Iconic font aggregator, collection, and patcher\" src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/raw\/master\/images\/project-subtitle-phrase.svg?sanitize=true\" \/><\/a>\n<\/h2>\n<p><strong>Nerd Fonts<\/strong> is a project that patches developer targeted fonts with a high number of glyphs (icons). Specifically to add a high number of extra glyphs from popular 'iconic fonts' such as <a href=\"https:\/\/github.com\/FortAwesome\/Font-Awesome\">Font Awesome \u27b6<\/a>, <a href=\"https:\/\/vorillaz.github.io\/devicons\/\" rel=\"nofollow\">Devicons \u27b6<\/a>, <a href=\"https:\/\/github.com\/primer\/octicons\">Octicons \u27b6<\/a>, and <a href=\"#glyph-sets\">others<\/a>.<\/p>\n<p>The following Sankey flow diagram shows the current glyph sets included:<\/p>\n<p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/images\/sankey-glyphs-combined-diagram.svg\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/raw\/master\/images\/sankey-glyphs-combined-diagram.svg?sanitize=true\" alt=\"@SankeyMATIC Diagram\" \/><\/a>\n<\/p>\n<sub><i>Diagram created using <a href=\"http:\/\/sankeymatic.com\/\" title=\"SankeyMATIC (BETA): A Sankey diagram builder for everyone\" rel=\"nofollow\">@SankeyMATIC<\/a><\/i><\/sub><h2><a id=\"user-content-important-notices\" class=\"anchor\" aria-hidden=\"true\" href=\"#important-notices\"><\/a>Important Notices<\/h2>\n<ul><li><code>master<\/code> branch file paths are <strong>not<\/strong> considered stable. <a href=\"#unstable-file-paths\">Verify your repository URI references<\/a><\/li>\n<li>cloning this repository is <strong>not<\/strong> recommended unless you are going to be <a href=\"#contributing\">contributing to development<\/a><\/li>\n<\/ul><h2><a id=\"user-content-table-of-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-contents\"><\/a>Table of Contents<\/h2>\n<p><a href=\"#tldr\"><strong>TL;DR<\/strong><\/a><\/p>\n<p><a href=\"#font-installation\"><strong>Installation Options<\/strong><\/a><\/p>\n<ul><li><a href=\"#option-1-download-and-install-manually\"><strong>1 - Manual<\/strong><\/a><\/li>\n<li><a href=\"#option-2-release-archive-download\"><strong>2 - Release Archive Download<\/strong><\/a><\/li>\n<li><a href=\"#option-3-install-script\"><strong>3 - Install Script<\/strong><\/a><\/li>\n<li><a href=\"#option-4-homebrew-fonts\"><strong>4 - Homebrew Fonts (macOS (OS X))<\/strong><\/a><\/li>\n<li><a href=\"#option-5-clone-the-repo\"><strong>5 - Clone Repo<\/strong><\/a><\/li>\n<li><a href=\"#option-6-ad-hoc-curl-download\"><strong>6 - Ad Hoc Curl Download<\/strong><\/a><\/li>\n<li><a href=\"#option-7-unofficial-arch-user-repository-aur\"><strong>7 - Arch User Repository (AUR) (Arch Linux)<\/strong><\/a><\/li>\n<li><a href=\"#option-8-patch-your-own-font\"><strong>8 - Patch Your Own Font<\/strong><\/a><\/li>\n<\/ul><p><a href=\"#features\"><strong>Features<\/strong><\/a><\/p>\n<ul><li><a href=\"#glyph-sets\"><strong>Glyph\/Icon sets<\/strong><\/a><\/li>\n<li><a href=\"#patched-fonts\"><strong>Patched Fonts<\/strong><\/a><\/li>\n<li><a href=\"#combinations\"><strong>Combinations<\/strong><\/a><\/li>\n<li><a href=\"#font-patcher\"><strong>Font Patcher<\/strong><\/a><\/li>\n<\/ul><p><a href=\"#font-patcher\"><strong>Developer \/ Contributor<\/strong><\/a><\/p>\n<ul><li><a href=\"#font-patcher\"><strong>Font Patcher<\/strong><\/a><\/li>\n<li><a href=\"#gotta-patch-em-all\"><strong>Gotta Patch 'em All Font Patcher!<\/strong><\/a><\/li>\n<li><a href=\"#other-good-fonts-to-patch\"><strong>Other Good Fonts to Patch<\/strong><\/a><\/li>\n<li><a href=\"#contributing\"><strong>Contributing<\/strong><\/a><\/li>\n<\/ul><p><a href=\"#project-motivation\"><strong>Project Motivation<\/strong><\/a><\/p>\n<p><strong>Additional Info<\/strong><\/p>\n<ul><li><a href=\"#unstable-file-paths\"><strong>Unstable file paths on master<\/strong><\/a><\/li>\n<li><a href=\"#changelog\"><strong>Changelog<\/strong><\/a><\/li>\n<li><a href=\"#license\"><strong>License<\/strong><\/a><\/li>\n<\/ul><h2><a id=\"user-content-tldr\" class=\"anchor\" aria-hidden=\"true\" href=\"#tldr\"><\/a>TL;DR<\/h2>\n<p>Nerd Fonts takes popular programming fonts and adds a bunch of Glyphs.\nThere is also a <a href=\"#font-patcher\">font patcher<\/a> available if your desired font isn't already patched.\nFor more high level information see the <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\">wiki<\/a>. If you are looking for the Vim plugin see <a href=\"https:\/\/github.com\/ryanoasis\/vim-devicons\" title=\"VimDevIcons Vim Plugin (external link) \u27b6\">VimDevIcons \u27b6<\/a>.<\/p>\n<h3><a id=\"user-content-various-download-options-for-fonts\" class=\"anchor\" aria-hidden=\"true\" href=\"#various-download-options-for-fonts\"><\/a>Various Download Options for Fonts<\/h3>\n<p><em>If you...<\/em><\/p>\n<ul><li><code>Option 1.<\/code> want to <strong>quickly<\/strong> grab an <strong>individual font<\/strong> download from the <a href=\"#patched-fonts\"><code>patched-fonts\/<\/code> directory<\/a><\/li>\n<li><code>Option 2.<\/code> want to download a <strong>font family<\/strong> package of variations (bold, italic, etc.) see <a href=\"#option-2-release-archive-download\">download an archive<\/a><\/li>\n<li><code>Option 3.<\/code> want to <strong>automate<\/strong> installing or use in <strong>scripts<\/strong> see the <a href=\"#option-3-install-script\">Install Script<\/a><\/li>\n<li><code>Option 4.<\/code> are on <strong>macOS<\/strong> and want to use <strong>Homebrew<\/strong> see <a href=\"#option-4-homebrew-fonts\">Homebrew Fonts<\/a><\/li>\n<li><code>Option 5.<\/code> want <strong>complete control<\/strong> then see <a href=\"#option-5-clone-the-repo\">cloning the repo<\/a><\/li>\n<li><code>Option 6.<\/code> want to use the <strong><code>curl<\/code> command<\/strong> or use in <strong>scripts<\/strong> see <a href=\"#option-6-ad-hoc-curl-download\">Ad Hoc Curl Download<\/a><\/li>\n<li><code>Option 7.<\/code> are on <strong>Arch Linux<\/strong> and want to use <strong>AUR packages<\/strong> see <a href=\"#option-7-unofficial-arch-user-repository-aur\">Unofficial Arch User Repositories<\/a><\/li>\n<li><code>Option 8.<\/code> want to patch your own font see the <a href=\"#option-8-patch-your-own-font\">Font Patcher<\/a><\/li>\n<\/ul><h2><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<ul><li>A <a href=\"#font-patcher\">FontForge Python script<\/a> to patch any font\n<ul><li>Includes an option to create <strong>Monospaced (fixed-pitch, fixed-width)<\/strong> <em>or<\/em> <strong>double-width (non-monospaced)<\/strong> glyphs<\/li>\n<li>For more details see the <a href=\"#font-patcher\"><strong>Font Patcher<\/strong><\/a> section<\/li>\n<\/ul><\/li>\n<li><strong><code>44<\/code><\/strong> already <a href=\"#patched-fonts\">patched font families<\/a><\/li>\n<li>Over <strong><code>1,200,000<\/code><\/strong> unique combinations\/variations of patched fonts <a href=\"#combinations\">(more details)<\/a><\/li>\n<li>Over <strong><code>2,400<\/code><\/strong> glyphs\/icons combined <a href=\"#combinations\">(more details)<\/a>\n<ul><li>Current glyph sets include: <a href=\"https:\/\/github.com\/ryanoasis\/powerline-extra-symbols\">Powerline with Extra Symbols<\/a>, <a href=\"https:\/\/github.com\/FortAwesome\/Font-Awesome\">Font Awesome<\/a>, <a href=\"https:\/\/github.com\/Templarian\/MaterialDesign\">Material Design Icons<\/a>, <a href=\"https:\/\/github.com\/erikflowers\/weather-icons\">Weather<\/a>, <a href=\"https:\/\/vorillaz.github.io\/devicons\/\" rel=\"nofollow\">Devicons<\/a>, <a href=\"https:\/\/github.com\/primer\/octicons\">Octicons<\/a>, <a href=\"https:\/\/github.com\/Lukas-W\/font-logos\">Font Logos<\/a> (Formerly <a href=\"https:\/\/github.com\/Lukas-W\/font-logos\">Font Linux<\/a>), <a href=\"https:\/\/github.com\/gabrielelana\/pomicons\">Pomicons<\/a><\/li>\n<\/ul><\/li>\n<li><strong>Monospaced (fixed-pitch, fixed-width)<\/strong> <em>or<\/em> <strong>double-width (non-monospaced)<\/strong> glyphs version of each font\n<ul><li>This refers to the Nerd Font glyphs themselves not necessarily the Font as a whole<\/li>\n<\/ul><\/li>\n<li>A Developer\/Contributor provided <a href=\"#gotta-patch-em-all\">bash script<\/a> to re-patch all the fonts<\/li>\n<\/ul><h2><a id=\"user-content-glyph-sets\" class=\"anchor\" aria-hidden=\"true\" href=\"#glyph-sets\"><\/a>Glyph Sets<\/h2>\n<p>\ud83d\udd0d\ud83d\udd0d You can now search for glyphs easily on <a href=\"https:\/\/nerdfonts.com\/#cheat-sheet\" rel=\"nofollow\">NerdFonts.com<\/a> via the <a href=\"https:\/\/nerdfonts.com\/#cheat-sheet\" rel=\"nofollow\">Cheat Sheet<\/a><\/p>\n<h3><a id=\"user-content-seti-ui--custom\" class=\"anchor\" aria-hidden=\"true\" href=\"#seti-ui--custom\"><\/a>Seti-UI + Custom<\/h3>\n<blockquote>\n<p>Based off of <a href=\"https:\/\/atom.io\/themes\/seti-ui\" rel=\"nofollow\">Seti-UI<\/a> Icomoon with extra custom glyphs.<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-custom-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-devicons\" class=\"anchor\" aria-hidden=\"true\" href=\"#devicons\"><\/a><a href=\"https:\/\/vorillaz.github.io\/devicons\/\" rel=\"nofollow\">Devicons<\/a><\/h3>\n<blockquote>\n<p>An iconic font made for developers, code jedis, ninjas, HTTPsters, evangelists and nerds(sic). \/ (<a href=\"https:\/\/vorillaz.github.io\/devicons\/\" rel=\"nofollow\">repo<\/a>) \/ (<a href=\"https:\/\/vorillaz.github.io\/devicons\/\" rel=\"nofollow\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-1\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-dev-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-font-awesome\" class=\"anchor\" aria-hidden=\"true\" href=\"#font-awesome\"><\/a><a href=\"https:\/\/github.com\/FortAwesome\/Font-Awesome\">Font Awesome<\/a><\/h3>\n<blockquote>\n<p>The iconic font and CSS toolkit. \/ (<a href=\"https:\/\/github.com\/FortAwesome\/Font-Awesome\">repo<\/a>) \/ (<a href=\"http:\/\/fontawesome.io\/\" rel=\"nofollow\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-2\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-fa-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-font-awesome-extension\" class=\"anchor\" aria-hidden=\"true\" href=\"#font-awesome-extension\"><\/a><a href=\"https:\/\/github.com\/AndreLZGava\/font-awesome-extension\">Font Awesome Extension<\/a><\/h3>\n<blockquote>\n<p>Simple Font Awesome Extension. \/ (<a href=\"https:\/\/github.com\/AndreLZGava\/font-awesome-extension\">repo<\/a>) \/ (<a href=\"https:\/\/andrelzgava.github.io\/font-awesome-extension\/\" rel=\"nofollow\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-3\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-fae-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-material-design-icons\" class=\"anchor\" aria-hidden=\"true\" href=\"#material-design-icons\"><\/a><a href=\"https:\/\/github.com\/Templarian\/MaterialDesign\">Material Design Icons<\/a><\/h3>\n<blockquote>\n<p>2000+ Material Design Icons from the Community. \/ (<a href=\"https:\/\/github.com\/Templarian\/MaterialDesign\">repo<\/a>) \/ (<a href=\"https:\/\/materialdesignicons.com\/\" rel=\"nofollow\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-4\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-mdi-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-weather\" class=\"anchor\" aria-hidden=\"true\" href=\"#weather\"><\/a><a href=\"https:\/\/github.com\/erikflowers\/weather-icons\">Weather<\/a><\/h3>\n<blockquote>\n<p>215 Weather Themed Icons and CSS. \/ (<a href=\"https:\/\/github.com\/erikflowers\/weather-icons\">repo<\/a>) \/ (<a href=\"http:\/\/weathericons.io\/\" rel=\"nofollow\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-5\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-weather-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-octicons\" class=\"anchor\" aria-hidden=\"true\" href=\"#octicons\"><\/a><a href=\"https:\/\/github.com\/primer\/octicons\">Octicons<\/a><\/h3>\n<blockquote>\n<p>GitHub's icons. \/ (<a href=\"https:\/\/github.com\/primer\/octicons\">repo<\/a>) \/ (<a href=\"https:\/\/octicons.github.com\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-6\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-6\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-oct-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a>*<\/h4>\n<p><sub>* To avoid conflicts between icon sets, these were modified in the 'Complete' fonts. To view the original codepoints see the <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/Codepoint-Conflicts\">Wiki<\/a>.<\/sub><\/p>\n<h3><a id=\"user-content-powerline-extra-symbols\" class=\"anchor\" aria-hidden=\"true\" href=\"#powerline-extra-symbols\"><\/a><a href=\"https:\/\/github.com\/ryanoasis\/powerline-extra-symbols\">Powerline Extra Symbols<\/a><\/h3>\n<blockquote>\n<p>Provides extra Powerline separator glyphs and a column number glyph (CN). \/ (<a href=\"https:\/\/github.com\/ryanoasis\/powerline-extra-symbols\">repo<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-7\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-7\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-ple-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-iec-power-symbols\" class=\"anchor\" aria-hidden=\"true\" href=\"#iec-power-symbols\"><\/a><a href=\"https:\/\/unicodepowersymbol.com\/\" rel=\"nofollow\">IEC Power Symbols<\/a><\/h3>\n<blockquote>\n<p>Adding new characters into Unicode \/ (<a href=\"https:\/\/unicodepowersymbol.com\/\" rel=\"nofollow\">website<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-8\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-8\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-iec-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-font-logos-formerly-font-linux\" class=\"anchor\" aria-hidden=\"true\" href=\"#font-logos-formerly-font-linux\"><\/a><a href=\"https:\/\/github.com\/Lukas-W\/font-logos\">Font Logos<\/a> (Formerly <a href=\"https:\/\/github.com\/Lukas-W\/font-logos\">Font Linux<\/a>)<\/h3>\n<blockquote>\n<p>Font-logos is an icon font containing logos of popular linux distributions and other open source software for inclusion in websites. \/ (<a href=\"https:\/\/github.com\/Lukas-W\/font-logos\">repo<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-9\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-9\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-linux-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-pomicons\" class=\"anchor\" aria-hidden=\"true\" href=\"#pomicons\"><\/a><a href=\"https:\/\/github.com\/gabrielelana\/pomicons\">Pomicons<\/a><\/h3>\n<blockquote>\n<p>8 symbols for the <a href=\"https:\/\/cirillocompany.de\/pages\/pomodoro-technique\" rel=\"nofollow\">\"Pomodoro Technique\"\u00ae<\/a>. \/ (<a href=\"https:\/\/github.com\/gabrielelana\/pomicons\">repo<\/a>)<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#mag-bookmark_tabs-see-set-on-the-nerdfontscom-website-10\"><\/a><a href=\"http:\/\/nerdfonts.com\/?set=nf-pom-#cheat-sheet\" rel=\"nofollow\">\ud83d\udd0d\ud83d\udcd1 See set on the NerdFonts.com Website<\/a><\/h4>\n<h3><a id=\"user-content-icon-names-in-shell\" class=\"anchor\" aria-hidden=\"true\" href=\"#icon-names-in-shell\"><\/a>Icon names in shell<\/h3>\n<p>Download provided <code>.sh<\/code> files from <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/bin\/scripts\/lib\">bin\/scripts\/lib\/<\/a> directory somewhere, recommended locations are <code>~\/.local\/share\/fonts\/<\/code> or <code>~\/bin\/<\/code>.<\/p>\n<ul><li><code>i_all.sh<\/code> - helper to load all files you've downloaded at once<\/li>\n<li><code>i_dev.sh<\/code> - Devicons (198 icons, 8 does not have an established name)<\/li>\n<li><code>i_fa.sh<\/code> - Font Awesome (675 icons, 111 aliases)<\/li>\n<li><code>i_fae.sh<\/code> - Font Awesome Extension (170 icons)<\/li>\n<li><code>i_iec.sh<\/code> - IEC Power Symbols (5 icons)<\/li>\n<li><code>i_linux.sh<\/code> - Font Linux (20 icons)<\/li>\n<li><code>i_oct.sh<\/code> - Octicons (172 icons)<\/li>\n<li><code>i_ple.sh<\/code> - Powerline Extra Symbols (37 icons, 2 aliases, 16 does not have an established name)<\/li>\n<li><code>i_pom.sh<\/code> - Pomicons (11 icons)<\/li>\n<li><code>i_seti.sh<\/code> - Seti-UI + Custom (50 icons, 2 aliases, 5 does not have an established name)<\/li>\n<\/ul><p>Then <code>source<\/code> the required file(s) and output <code>$i_*<\/code> variables to see icons:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>source ~\/.local\/share\/fonts\/i_oct.sh\necho $i_oct_heart\n# Output:\n# \u2665<\/pre><\/div>\n<p><strong>NOTE:<\/strong> You have to use one of the Nerd Fonts to see correct icons for some icon sets (Devicons, Font Awesome Extension, Font Linux), but other sets should work with their standard fonts too.<\/p>\n<h2><a id=\"user-content-patched-fonts\" class=\"anchor\" aria-hidden=\"true\" href=\"#patched-fonts\"><\/a>Patched Fonts<\/h2>\n<table><thead><tr><th align=\"left\">Font Name<\/th>\n<th align=\"left\">Font Name and Repository<\/th>\n<th align=\"left\">*RFN<\/th>\n<th align=\"left\">EM Size<\/th>\n<th align=\"left\">Status<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/3270\">3270 Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/rbanffy\/3270font\">3270<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/AnonymousPro\">Anonymice Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/www.marksimonson.com\/fonts\/view\/anonymous-pro\" rel=\"nofollow\">Anonymous Pro<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Arimo\">Arimo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/google\/fonts\/tree\/master\/apache\/arimo\">Arimo<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/AurulentSansMono\">Aurulent Sans Mono Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/BigBlueTerminal\">BigBlueTerminal<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1200<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/BitstreamVeraSansMono\">Bitstream Vera Sans Mono Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Blex\">Blex*<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/IBM\/plex\">IBM Plex<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/CodeNewRoman\">Code New Roman Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Cousine\">Cousine Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/fonts.google.com\/specimen\/Cousine\" rel=\"nofollow\">Cousine<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/DejaVuSansMono\">DejaVu Sans Mono Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/DroidSansMono\">Droid Sans Mono Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/FantasqueSansMono\">Fantasque Sans Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/belluzj\/fantasque-sans\">Fantasque Sans<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/FiraCode\">Fura Code Nerd Font*<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/tonsky\/FiraCode\">Fira<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/FiraMono\">Fura Mono Nerd Font*<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/mozilla\/Fira\">Fira<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Go-Mono\">Go Mono Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/go.googlesource.com\/image\/+\/master\/font\/gofont\/ttfs\/\" rel=\"nofollow\">Go-Mono<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Gohu\">Gohu Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/koemaeda\/gohufont-ttf\">Gohu TTF<\/a>,<a href=\"http:\/\/font.gohu.org\/\" rel=\"nofollow\">Gohu<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Hack\">Hack Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/chrissimpkins\/Hack\">Hack<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Hasklig\">Hasklug Nerd Font*<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/i-tu\/Hasklig\">Hasklig<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/HeavyData\">Heavy Data Mono Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Hermit\">Hermut Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Inconsolata\">Inconsolata Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/InconsolataGo\">Inconsolata Go Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/InconsolataLGC\">Inconsolata LGC Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Iosevka\">Iosevka Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/be5invis\/Iosevka\">Iosevka<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/issues\/83\">#83<\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Lekton\">Lekton Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/LiberationMono\">Literation Mono Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/pagure.io\/liberation-fonts\" rel=\"nofollow\">Liberation<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Meslo\">Meslo Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Monofur\">Monofur Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2400<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Monoid\">Monoid Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1536<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Mononoki\">Mononoki Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/madmalik.github.io\/mononoki\/\" rel=\"nofollow\">Mononoki<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">1024<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/MPlus\">M+ (MPlus) Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Noto\">Noto<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/OpenDyslexic\">OpenDyslexic<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Overpass\">Overpass<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/ProFont\">ProFont (Windows tweaked) Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1200<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/ProFont\">ProFont (x11) Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/ProggyClean\">ProggyClean Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\">Imperfect<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/RobotoMono\">Roboto Mono<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/SourceCodePro\">Sauce Code Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/github.com\/adobe-fonts\/source-code-pro\">Source<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/ShareTechMono\">Shure Tech Mono Nerd Font*<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/fonts.google.com\/specimen\/Share+Tech+Mono\" rel=\"nofollow\">Share Tech Mono<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/SpaceMono\">Space Mono Nerd Font<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/fonts.google.com\/specimen\/Space+Mono\" rel=\"nofollow\">Space Mono<\/a><\/td>\n<td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Terminus\">Terminess Nerd Font*<\/a><\/td>\n<td align=\"left\"><a href=\"http:\/\/terminus-font.sourceforge.net\" rel=\"nofollow\">Terminus Font<\/a><\/td>\n<td align=\"left\">YES<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Tinos\">Tinos<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">YES<\/td>\n<td align=\"left\">2048<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/Ubuntu\">Ubuntu Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/patched-fonts\/UbuntuMono\">Ubuntu Mono Nerd Font<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">NO<\/td>\n<td align=\"left\">1000<\/td>\n<td align=\"left\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/windows-pass-sm.png\" alt=\"w\" title=\"Windows status is working \u263a\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/mac-unknown-sm.png\" alt=\"m2\" title=\"macOS (OSX) status is Unknown\/Un-tested\" \/><\/a> <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/screenshots\/v1.0.x\/linux-pass-sm.png\" alt=\"l\" title=\"Linux status is working \u263a\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><p><sub><em>*RFN = Reserved Font Name<\/em><\/sub><\/p>\n<h2><a id=\"user-content-combinations\" class=\"anchor\" aria-hidden=\"true\" href=\"#combinations\"><\/a>Combinations<\/h2>\n<ul><li>Over <strong><code>1,200,000<\/code><\/strong> unique variations\/combinations (Power Set) of patched fonts:\n<ul><li><strong><code>44<\/code><\/strong> patched font typefaces<\/li>\n<li><strong><code>612<\/code><\/strong> patched font families<\/li>\n<li><strong><code>2,448<\/code><\/strong> 'complete' variations\/combinations<\/li>\n<li><strong><code>1,264,440<\/code><\/strong> <em>possible<\/em> variations\/combinations<\/li>\n<li><strong><code>1,266,888<\/code><\/strong> total calculated combinations (2,448 + 1,264,440)<\/li>\n<\/ul><\/li>\n<li>Combinations for each font are any combination of <a href=\"#variations\">Variations<\/a><\/li>\n<\/ul><h3><a id=\"user-content-variations\" class=\"anchor\" aria-hidden=\"true\" href=\"#variations\"><\/a>Variations<\/h3>\n<h2><a id=\"user-content-font-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#font-installation\"><\/a>Font Installation<\/h2>\n<h3><a id=\"user-content-option-1-download-and-install-manually\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-1-download-and-install-manually\"><\/a><code>Option 1: Download and Install Manually<\/code><\/h3>\n<blockquote>\n<p>Best option for <strong>quickly<\/strong> getting a specific <strong>individual font<\/strong>.<\/p>\n<\/blockquote>\n<p>Download the specific <a href=\"#patched-fonts\">patched font<\/a> of your choice<\/p>\n<h3><a id=\"user-content-option-2-release-archive-download\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-2-release-archive-download\"><\/a><code>Option 2: Release Archive Download<\/code><\/h3>\n<blockquote>\n<p>Best option if you want an <strong>archive<\/strong> or complete <strong>font family<\/strong> of variations (Bold, Italic, etc.).<\/p>\n<\/blockquote>\n<p>Fonts are available for download as packages in the <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/releases\/latest\">latest release<\/a><\/p>\n<h3><a id=\"user-content-option-3-install-script\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-3-install-script\"><\/a><code>Option 3: Install Script<\/code><\/h3>\n<blockquote>\n<p>Best option if you want to <strong>automate<\/strong> installing or for use in <strong>scripts<\/strong>.<\/p>\n<\/blockquote>\n<p><em>Note<\/em>: Only for Linux &amp; macOS (OS X)\n<em>Note<\/em>: <strong>Requires cloning<\/strong> the repo as of now<\/p>\n<h4><a id=\"user-content-all-fonts\" class=\"anchor\" aria-hidden=\"true\" href=\"#all-fonts\"><\/a>All fonts:<\/h4>\n<ul><li>Installs all the patched Fonts (<em>Warning: This is a lot of Fonts adding up to a large size<\/em>)<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>.\/install.sh<\/pre><\/div>\n<h4><a id=\"user-content-single-font\" class=\"anchor\" aria-hidden=\"true\" href=\"#single-font\"><\/a>Single font:<\/h4>\n<ul><li>Installs a single Font of your choice<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>.\/install.sh &lt;FontName&gt;\n.\/install.sh Hack\n.\/install.sh HeavyData<\/pre><\/div>\n<h3><a id=\"user-content-option-4-homebrew-fonts\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-4-homebrew-fonts\"><\/a><code>Option 4: Homebrew Fonts<\/code><\/h3>\n<blockquote>\n<p>Best option if on <strong>macOS<\/strong> and want to use <strong>Homebrew<\/strong>.<\/p>\n<\/blockquote>\n<p>All fonts are available via <a href=\"https:\/\/github.com\/caskroom\/homebrew-fonts\">Homebrew Fonts<\/a> on macOS (OS X)<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>brew tap caskroom\/fonts\nbrew cask install font-hack-nerd-font<\/pre><\/div>\n<h3><a id=\"user-content-option-5-clone-the-repo\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-5-clone-the-repo\"><\/a><code>Option 5: Clone the Repo<\/code><\/h3>\n<blockquote>\n<p>Best option for <strong>full control<\/strong>, <strong>all<\/strong> or <strong>most<\/strong> of the fonts, or <strong>contributing<\/strong> to development.<\/p>\n<\/blockquote>\n<p>Cloning of this repository is <strong>not<\/strong> required nor efficient if you are simply only interested in a limited set of fonts.<\/p>\n<p>However if you do want to clone the repo be sure to <em>shallow<\/em> clone:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>git clone --depth 1<\/pre><\/div>\n<h3><a id=\"user-content-option-6-ad-hoc-curl-download\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-6-ad-hoc-curl-download\"><\/a><code>Option 6: Ad Hoc Curl Download<\/code><\/h3>\n<blockquote>\n<p>Option if you want to use the <strong><code>curl<\/code> command<\/strong> or for use in <strong>scripts<\/strong>.<\/p>\n<\/blockquote>\n<h4><a id=\"user-content-linux\" class=\"anchor\" aria-hidden=\"true\" href=\"#linux\"><\/a>Linux<\/h4>\n<div class=\"highlight highlight-source-shell\"><pre>mkdir -p ~\/.local\/share\/fonts\ncd ~\/.local\/share\/fonts &amp;&amp; curl -fLo \"Droid Sans Mono for Powerline Nerd Font Complete.otf\" https:\/\/github.com\/ryanoasis\/nerd-fonts\/raw\/master\/patched-fonts\/DroidSansMono\/complete\/Droid%20Sans%20Mono%20Nerd%20Font%20Complete.otf<\/pre><\/div>\n<p><em>Note:<\/em> deprecated alternative paths: <code>~\/.fonts<\/code><\/p>\n<h4><a id=\"user-content-macos-os-x\" class=\"anchor\" aria-hidden=\"true\" href=\"#macos-os-x\"><\/a>macOS (OS X)<\/h4>\n<div class=\"highlight highlight-source-shell\"><pre>cd ~\/Library\/Fonts &amp;&amp; curl -fLo \"Droid Sans Mono for Powerline Nerd Font Complete.otf\" https:\/\/github.com\/ryanoasis\/nerd-fonts\/raw\/master\/patched-fonts\/DroidSansMono\/complete\/Droid%20Sans%20Mono%20Nerd%20Font%20Complete.otf<\/pre><\/div>\n<h3><a id=\"user-content-option-7-unofficial-arch-user-repository-aur\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-7-unofficial-arch-user-repository-aur\"><\/a><code>Option 7: Unofficial Arch User Repository (AUR)<\/code><\/h3>\n<blockquote>\n<p>Option for <strong>Arch Linux<\/strong> and wanting to use <strong>AUR packages<\/strong>.<\/p>\n<\/blockquote>\n<p>The following fonts are available via <a href=\"https:\/\/aur.archlinux.org\/\" rel=\"nofollow\">AUR packages<\/a> on Arch Linux:<\/p>\n<ul><li><a href=\"https:\/\/aur.archlinux.org\/packages\/nerd-fonts-complete\/\" rel=\"nofollow\">Nerd Fonts Complete (double-width)<\/a><\/li>\n<li><a href=\"https:\/\/aur.archlinux.org\/packages\/nerd-fonts-complete-mono-glyphs\/\" rel=\"nofollow\">Nerd Fonts Complete (single-width)<\/a><\/li>\n<li><a href=\"https:\/\/aur.archlinux.org\/packages\/nerd-fonts-dejavu-complete\/\" rel=\"nofollow\">Nerd Fonts DejaVu Complete<\/a><\/li>\n<li><a href=\"https:\/\/aur.archlinux.org\/packages\/nerd-fonts-source-code-pro\/\" rel=\"nofollow\">Nerd Fonts Source Code Pro Complete<\/a><\/li>\n<li><a href=\"https:\/\/aur.archlinux.org\/packages\/nerd-fonts-git\/\" rel=\"nofollow\">Nerd Fonts Git (out of date)<\/a><\/li>\n<\/ul><h3><a id=\"user-content-option-8-patch-your-own-font\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-8-patch-your-own-font\"><\/a><code>Option 8: Patch Your Own Font<\/code><\/h3>\n<blockquote>\n<p>The option for <strong>patching<\/strong> your <strong>own font<\/strong> or fully <strong>customizing<\/strong> the patched font.<\/p>\n<\/blockquote>\n<p>Use the provided Python command line script to generate a patched font from your own font to get the extra new glyphs<\/p>\n<p>See: <a href=\"#font-patcher\">Font Patcher<\/a> for usage<\/p>\n<ul><li>use this option if you do <strong>not<\/strong> want to use one of the <a href=\"#patched-fonts\">fonts provided<\/a><\/li>\n<li>you will still need to copy the generated font to the correct font directory on your system<\/li>\n<\/ul><h2 id=\"user-content-font-patcher\" style=\"text-align: center;\"><a id=\"\" class=\"anchor\" aria-hidden=\"true\" href=\"#\"><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/images\/nerd-fonts-patcher-logo.png\"><img src=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/raw\/master\/images\/nerd-fonts-patcher-logo.png\" alt=\"Nerd Fonts Patcher\" \/><\/a>\n<\/h2>\n<p>Patching the font of your own choosing for use with the <a href=\"https:\/\/github.com\/ryanoasis\/vim-devicons\" title=\"VimDevIcons Vim Plugin (external link) \u27b6\">VimDevIcons \u27b6<\/a>:<\/p>\n<ul><li>requires: Python 2 (or Python 3), <code>python-fontforge<\/code> package (version <code>20141231<\/code> or later, see\nthe <a href=\"http:\/\/designwithfontforge.com\/en-US\/Installing_Fontforge.html\" rel=\"nofollow\">install instructions<\/a>)<\/li>\n<li>alternative install method on OSX: <code>brew install fontforge<\/code><\/li>\n<li>Usage:<\/li>\n<\/ul><pre>.\/font-patcher PATH_TO_FONT\n<\/pre>\n<ul><li>Alternative usage: Execute the patcher with the FontForge binary using the script flag:<\/li>\n<\/ul><pre>.\/fontforge -script font-patcher PATH_TO_FONT\n<\/pre>\n<pre>usage: font-patcher [-h] [-v] [-s] [-l] [-q] [-w] [-c] [--fontawesome]\n                    [--fontawesomeextension] [--fontlinux] [--octicons]\n                    [--powersymbols] [--pomicons] [--powerline]\n                    [--powerlineextra] [--material] [--weather]\n                    [--custom [CUSTOM]] [--postprocess [POSTPROCESS]]\n                    [--removeligs] [--configfile [CONFIGFILE]]\n                    [--progressbars | --no-progressbars] [--careful]\n                    [-ext [EXTENSION]] [-out [OUTPUTDIR]]\n                    font\nNerd Fonts Font Patcher: patches a given font with programming and development related glyphs\n* Website: https:\/\/www.nerdfonts.com\n* Version: 2.0.0\n* Development Website: https:\/\/github.com\/ryanoasis\/nerd-fonts\n* Changelog: https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/changelog.md\npositional arguments:\n  font                  The path to the font to patch (e.g., Inconsolata.otf)\noptional arguments:\n  -h, --help            show this help message and exit\n  -v, --version         show program's version number and exit\n  -s, --mono, --use-single-width-glyphs\n                        Whether to generate the glyphs as single-width not double-width (default is double-width)\n  -l, --adjust-line-height\n                        Whether to adjust line heights (attempt to center powerline separators more evenly)\n  -q, --quiet, --shutup\n                        Do not generate verbose output\n  -w, --windows         Limit the internal font name to 31 characters (for Windows compatibility)\n  -c, --complete        Add all available Glyphs\n  --fontawesome         Add Font Awesome Glyphs (http:\/\/fontawesome.io\/)\n  --fontawesomeextension\n                        Add Font Awesome Extension Glyphs (https:\/\/andrelzgava.github.io\/font-awesome-extension\/)\n  --fontlinux, --fontlogos\n                        Add Font Linux and other open source Glyphs (https:\/\/github.com\/Lukas-W\/font-logos)\n  --octicons            Add Octicons Glyphs (https:\/\/octicons.github.com)\n  --powersymbols        Add IEC Power Symbols (https:\/\/unicodepowersymbol.com\/)\n  --pomicons            Add Pomicon Glyphs (https:\/\/github.com\/gabrielelana\/pomicons)\n  --powerline           Add Powerline Glyphs\n  --powerlineextra      Add Powerline Glyphs (https:\/\/github.com\/ryanoasis\/powerline-extra-symbols)\n  --material, --materialdesignicons, --mdi\n                        Add Material Design Icons (https:\/\/github.com\/templarian\/MaterialDesign)\n  --weather, --weathericons\n                        Add Weather Icons (https:\/\/github.com\/erikflowers\/weather-icons)\n  --custom [CUSTOM]     Specify a custom symbol font. All new glyphs will be copied, with no scaling applied.\n  --postprocess [POSTPROCESS]\n                        Specify a Script for Post Processing\n  --removeligs, --removeligatures\n                        Removes ligatures specificed in JSON configuration file\n  --configfile [CONFIGFILE]\n                        Specify a file path for JSON configuration file (see sample: src\/config.sample.json)\n  --progressbars        Show percentage completion progress bars per Glyph Set\n  --no-progressbars     Don't show percentage completion progress bars per Glyph Set\n  --careful             Do not overwrite existing glyphs if detected\n  -ext [EXTENSION], --extension [EXTENSION]\n                        Change font file type to create (e.g., ttf, otf)\n  -out [OUTPUTDIR], --outputdir [OUTPUTDIR]\n                        The directory to output the patched font file to\n<\/pre>\n<h4><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h4>\n<pre>.\/font-patcher Droid\\ Sans\\ Mono\\ for\\ Powerline.otf\n.\/font-patcher Droid\\ Sans\\ Mono\\ for\\ Powerline.otf -s -q\n.\/font-patcher Droid\\ Sans\\ Mono\\ for\\ Powerline.otf --use-single-width-glyphs --quiet\n.\/font-patcher Droid\\ Sans\\ Mono\\ for\\ Powerline.otf -w\n.\/font-patcher Droid\\ Sans\\ Mono\\ for\\ Powerline.otf --windows --quiet\n.\/font-patcher Droid\\ Sans\\ Mono\\ for\\ Powerline.otf --windows --pomicons --quiet\n.\/font-patcher Inconsolata.otf --fontawesome\n.\/font-patcher Inconsolata.otf --fontawesome --octicons --pomicons\n.\/font-patcher Inconsolata.otf\n<\/pre>\n<h2><a id=\"user-content-gotta-patch-em-all-font-patcher\" class=\"anchor\" aria-hidden=\"true\" href=\"#gotta-patch-em-all-font-patcher\"><\/a>Gotta Patch 'em All Font Patcher!<\/h2>\n<ul><li>\n<p>for Contributor or Developer use<\/p>\n<\/li>\n<li>\n<p>re-patches <strong>all<\/strong> fonts in the unpatched directory:<\/p>\n<\/li>\n<\/ul><pre>.\/gotta-patch-em-all-font-patcher\\!.sh\n<\/pre>\n<ul><li>can optionally limit to specific font name pattern:<\/li>\n<\/ul><pre>.\/gotta-patch-em-all-font-patcher\\!.sh Hermit\n<\/pre>\n<h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>See <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/contributing.md\">contributing.md<\/a><\/p>\n<h2><a id=\"user-content-unstable-file-paths\" class=\"anchor\" aria-hidden=\"true\" href=\"#unstable-file-paths\"><\/a>Unstable File Paths<\/h2>\n<p>\u26a0\ufe0f Warning: File paths may change based on releases (especially <strong>major<\/strong> version bumps)<\/p>\n<p>Reference the <strong>release<\/strong> branch and <em>not<\/em> the <del>master<\/del> branch because paths are subject to change for each release<\/p>\n<ul><li>For example:\n<ul><li>\u2705 Use: <code>https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/<b>0.9.0<\/b>\/patched-fonts\/Hermit\/Medium\/complete\/Hurmit%20Medium%20Nerd%20Font%20Complete.otf<\/code><\/li>\n<li>\u274c Instead of: <code>https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/<del>master<\/del>\/patched-fonts\/Hermit\/Medium\/complete\/Hurmit%20Medium%20Nerd%20Font%20Complete.otf<\/code><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-other-good-fonts-to-patch\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-good-fonts-to-patch\"><\/a>Other Good Fonts to Patch<\/h2>\n<ul><li>a list of additional good fonts to patch that cannot be provided or shared due to the license:<\/li>\n<li><a href=\"http:\/\/input.fontbureau.com\/download\/\" rel=\"nofollow\">Input Mono<\/a> (license restriction)\n<ul><li>Possibly coming with external hosting :)<\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/www.fsd.it\/shop\/fonts\/pragmatapro\/\" rel=\"nofollow\">PragmataPro<\/a> (not free)<\/li>\n<li><a href=\"https:\/\/www.microsoft.com\/typography\/fonts\/family.aspx?FID=300\" rel=\"nofollow\">Consolas<\/a> (proprietary)<\/li>\n<li><a href=\"https:\/\/www.typography.com\/fonts\/operator\/\" rel=\"nofollow\">Operator Mono<\/a> (not free)<\/li>\n<li><a href=\"https:\/\/dank.sh\/\" rel=\"nofollow\">Dank Mono<\/a> (not free)<\/li>\n<\/ul><h2><a id=\"user-content-project-motivation\" class=\"anchor\" aria-hidden=\"true\" href=\"#project-motivation\"><\/a>Project Motivation<\/h2>\n<p>See <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/wiki\/Project-Purpose\">Wiki: Project Purpose<\/a><\/p>\n<h2><a id=\"user-content-changelog\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelog\"><\/a>Changelog<\/h2>\n<p>See <a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/changelog.md\">changelog.md<\/a><\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p><a href=\"https:\/\/github.com\/ryanoasis\/nerd-fonts\/blob\/master\/LICENSE\">MIT<\/a> \u00a9 Ryan L McIntyre<\/p>\n<\/article>","created_at":"2019-01-21T14:13:19+0000","updated_at":"2019-01-21T14:13:19+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/8083459?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":783,"title":"Programming Fonts - Test Drive","url":"http:\/\/app.programmingfonts.org\/","content":"<section class=\"select-list\"><section class=\"config\"><p>\n\t\t\t\t\t<label for=\"size\">Size<\/label>\n\t\t\t\t\t<\/p><p>\n\t\t\t\t\t<label for=\"spacing\">Spacing<\/label>\n\t\t\t\t\t<\/p><p>\n\t\t\t\t\t<label for=\"aliasing\">AA<\/label>\n\t\t\t\t\t<\/p><\/section><section class=\"config\"><p class=\"select\">\n\t\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\t<label for=\"select-theme\">Theme<\/label>\n\t\t\t\t\t<\/p><p class=\"select\">\n\t\t\t\t\t<label for=\"select-language\">Language<\/label>\n\t\t\t\t\t<\/p><\/section><footer>\n\t\t\t\t<p>Test drive over 50 free programming fonts!<\/p>\n\t\t\t\t\n\t\t\t\t<p>\n\t\t\t\t\t<a href=\"http:\/\/programmingfonts.org\">Blog<\/a> -\n\t\t\t\t\t<a href=\"http:\/\/atom.io\/packages\/fonts\" rel=\"external\">Atom package<\/a> -\n\t\t\t\t\t<a href=\"https:\/\/www.sublimetext.com\/docs\/3\/font.html\" rel=\"external\">Sublime Text docs<\/a>\n\t\t\t\t<\/p>\n\t\t\t<\/footer><\/section>","created_at":"2019-01-21T14:03:47+0000","updated_at":"2019-01-21T14:03:47+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"app.programmingfonts.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":782,"title":"RE:DOM \u2013\u00a0Tiny turboboosted JavaScript library for creating user interfaces","url":"https:\/\/redom.js.org\/","content":"RE:DOM \u2013\u00a0Tiny turboboosted JavaScript library for creating user interfaces<header><section class=\"center\"><p>Develop web apps with 100 % JavaScript and web standards.<\/p><p class=\"violator solid\">\n        <a href=\"https:\/\/redom.js.org\/documentation\">Documentation<\/a>\n      <\/p><p class=\"violator\">\n        <a href=\"https:\/\/github.com\/redom\/redom\">GitHub<\/a>\n      <\/p><div><p>RE:DOM is a tiny (2 KB) DOM library by <a href=\"https:\/\/pakastin.fi\" target=\"_blank\">Juha Lindstedt<\/a> and <a href=\"https:\/\/github.com\/redom\/redom\/graphs\/contributors\">contributors<\/a>, which adds useful helpers to create DOM elements and keeping them in sync with the data.<\/p><p>Because RE:DOM is so close to the metal and <strong>doesn't use virtual dom<\/strong>, it's actually <strong>faster<\/strong> and uses <strong>less memory<\/strong> than almost all virtual dom based libraries, including React (<a href=\"http:\/\/www.stefankrause.net\/js-frameworks-benchmark7\/table.html\" target=\"_blank\">benchmark<\/a>).<\/p><p>It's also easy to create <strong>reusable components<\/strong> with RE:DOM.<\/p><p>Another benefit is, that you can use just <strong>pure JavaScript<\/strong>, so no complicated templating languages to learn and hassle with. Plus RE:DOM <strong>plays nicely with others<\/strong>. No need to write wrappers for things like Google Maps.<\/p><\/div><p class=\"violator\">\n        <a href=\"https:\/\/redom.js.org\/documentation\">Learn more<\/a>\n      <\/p><h3>Create HTML elements easily<\/h3><pre class=\"language-js\">import { el } from 'redom';\nel('h1', 'Hello RE:DOM!');\n<\/pre><pre class=\"language-markup\">&lt;h1&gt;Hello RE:DOM!&lt;\/h1&gt;<\/pre><h3>No need for bundlers<\/h3><pre class=\"language-html\">&lt;script type=\"module\"&gt;\n  import { el } from 'https:\/\/redom.js.org\/redom.es.min.js';\n  el('h1', 'Hello modules!');\n&lt;\/script&gt;<\/pre><pre class=\"language-markup\">&lt;h1&gt;Hello modules!&lt;\/h1&gt;<\/pre><h3>You can even use plain old ES5<\/h3><pre class=\"language-html\">&lt;script src='https:\/\/redom.js.org\/redom.min.js'&gt;&lt;\/script&gt;\n&lt;script&amp;gt\n    var element = redom.el('h1', 'Hello ES5!');\n    redom.mount(document.body, element);\n&lt;\/script&gt;<\/pre><pre class=\"language-markup\">&lt;h1&gt;Hello ES5!&lt;\/h1&gt;<\/pre><h3>Add attributes<\/h3><pre class=\"language-js\">import { el, mount } from 'redom';\n\/\/ create HTML element\nconst logo = el('img#redom-logo.logo', {\n  src: 'img\/logo.svg'\n});\n\/\/ mount to DOM\nmount(document.body, logo);<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;img id=\"redom-logo\" class=\"logo\" src=\"img\/logo.svg\"&gt;\n&lt;\/body&gt;<\/pre><h3>Listen to events<\/h3><pre class=\"language-js\">import { el } from 'redom';\n\/\/ define Login component\nexport class Login {\n  constructor () {\n    this.el = el('form#login',\n      this.email = el('input.email', { type: 'email' }),\n      this.pass = el('input.pass', { type: 'password' }),\n      this.submit = el('button', { type: 'submit' },\n        'Sign in'\n      )\n    );\n    this.el.onsubmit = e =&gt; {\n      e.preventDefault();\n      console.log({\n        email: this.email.value,\n        pass: this.pass.value\n      });\n    }\n  }\n}<\/pre><pre class=\"language-js\">import { mount } from 'redom';\nimport { Login } from '.\/login.js';\n\/\/ create login\nconst login = new Login();\n\/\/ mount to DOM\nmount(document.body, login);<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;form id=\"login\"&gt;\n    &lt;input class=\"email\" type=\"email\"&gt;\n    &lt;input class=\"pass\" type=\"password\"&gt;\n    &lt;button type=\"submit\"&gt;Sign in&lt;\/button&gt;\n  &lt;\/form&gt;\n&lt;\/body&gt;<\/pre><h3>Sync lists<\/h3><pre class=\"language-js\">import { el } from 'redom';\n\/\/ define &lt;li&gt; component\nexport class Li {\n  constructor () {\n    this.el = el('li');\n  }\n  update (i) {\n    this.el.textContent = `Item ${i}`;\n  }\n}<\/pre><pre class=\"language-js\">import { list, mount } from 'redom';\nimport { Li } from '.\/li.js';\n\/\/ create list\nconst ul = list('ul', Li);\n\/\/ update with data\nul.update([1, 2, 3]);\n\/\/ mount to DOM\nmount(document.body, ul);\n\/\/ schedule another update\nsetTimeout(() =&gt; {\n  ul.update([2, 3, 4, 5]);\n}, 1000);<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;ul&gt;\n    &lt;li&gt;Item 1&lt;\/li&gt;\n    &lt;li&gt;Item 2&lt;\/li&gt;\n    &lt;li&gt;Item 3&lt;\/li&gt;\n  &lt;\/ul&gt;\n&lt;\/body&gt;<\/pre><h3>Sync tables<\/h3><pre class=\"language-js\">import { el, list } from 'redom';\n\/\/ define &lt;td&gt; component\nclass Td {\n  constructor () {\n    this.el = el('td');\n  }\n  update (i) {\n    this.el.textContent = i;\n  }\n}\n\/\/ define &lt;tr&gt; component by extending list\nexport const Tr = list.extend('tr', Td);<\/pre><pre class=\"language-js\">import { list, mount } from 'redom';\nimport { Tr } from '.\/tr';\n\/\/ create table\nconst table = list('table', Tr);\n\/\/ update with data\ntable.update([\n  [ 1, 2, 3 ],\n  [ 4, 5, 6 ],\n  [ 7, 8, 9 ]\n]);\n\/\/ mount to DOM\nmount(document.body, table);\n\/\/ schedule another update\nsetTimeout(() =&gt; {\n  table.update([\n    [ 3, 2, 1 ],\n    [ 6, 5, 4 ],\n    [ 9, 8, 7 ]\n  ]);\n}, 1000);<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;table&gt;\n    &lt;tr&gt;\n      &lt;td&gt;1&lt;\/td&gt;&lt;td&gt;2&lt;\/td&gt;&lt;td&gt;3&lt;\/td&gt;\n    &lt;\/tr&gt;\n    &lt;tr&gt;\n      &lt;td&gt;4&lt;\/td&gt;&lt;td&gt;5&lt;\/td&gt;&lt;td&gt;6&lt;\/td&gt;\n    &lt;\/tr&gt;\n    &lt;tr&gt;\n      &lt;td&gt;7&lt;\/td&gt;&lt;td&gt;8&lt;\/td&gt;&lt;td&gt;9&lt;\/td&gt;\n    &lt;\/tr&gt;\n  &lt;\/table&gt;\n&lt;\/body&gt;<\/pre><h3>Create and update components<\/h3><pre class=\"language-javascript\">import { el, text } from 'redom';\n\/\/ define component\nexport class Hello {\n  \/\/ ..how to create it\n  constructor () {\n    this.el = el('h1',\n      'Hello ',\n      this.target = text('world'),\n      '!'\n    );\n  }\n  \/\/ ..and how to update it\n  update (target) {\n    this.target.textContent = target;\n  }\n}<\/pre><pre class=\"language-javascript\">import { mount } from 'redom';\nimport { Hello } from '.\/hello.js';\nconst hello = new Hello();\nhello.update('RE:DOM');\nmount(document.body, hello);<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;h1&gt;Hello RE:DOM!&lt;\/h1&gt;\n&lt;\/body&gt;<\/pre><h3>Combine list + component<\/h3><pre class=\"language-javascript\">import { el, text } from 'redom';\n\/\/ define component\nexport class Hello {\n  \/\/ ..how to create it\n  constructor () {\n    this.el = el('h1',\n      'Hello ',\n      this.target = text(''),\n      '!'\n    );\n  }\n  \/\/ ..and how to update it\n  update ({ id, target }) {\n    this.target.textContent = target;\n  }\n}<\/pre><pre class=\"language-javascript\">import { list, mount } from 'redom';\nimport { Hello } from '.\/hello.js';\n\/\/ create list of Hello's\nconst hellos = list('section', Hello, 'id');\n\/\/ init with some data\nconst data = [{\n  id: 1,\n  target: 'RE:DOM'\n}, {\n  id: 2,\n  target: 'world'\n}];\nhellos.update(data);\n\/\/ mount to DOM\nmount(document.body, hellos);\n\/\/ schedule an update\nsetTimeout(() =&gt; {\n  const data = [{\n    id: 2,\n    target: 'world'\n  }, {\n    id: 1,\n    target: 'RE:DOM'\n  }];\n  \/\/ reuses DOM elements and only reorders in this case!\n  hellos.update (data);\n}, 1000);\n<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;section&gt;\n    &lt;h1&gt;Hello RE:DOM!&lt;\/h1&gt;\n    &lt;h1&gt;Hello world!&lt;\/h1&gt;\n  &lt;\/section&gt;\n&lt;\/body&gt;<\/pre><h3>Switch sections<\/h3><pre class=\"language-javascript\">const { el } = require('redom');\n\/\/ define A component\nexport class A {\n  constructor () {\n    this.el = el('.a')\n  }\n  update ({ val }) {\n    this.el.textContent = val;\n  }\n}<\/pre><pre class=\"language-javascript\">const { el } = require('redom');\n\/\/ define B component\nexport class B {\n  constructor () {\n    this.el = el('.b')\n  }\n  update ({ val }) {\n    this.el.textContent = val;\n  }\n}<\/pre><pre class=\"language-javascript\">const { el, router } = require('redom');\nimport { A } from '.\/a.js';\nimport { B } from '.\/b.js';\nconst routes = {\n  a: A,\n  b: B\n};\n\/\/ define main app component\nexport class App {\n  constructor () {\n    this.el = el('.app',\n      this.sections = router('.content', routes)\n    );\n  }\n  update ({ section, data }) {\n    this.sections.update(section, data[section]);\n  }\n}<\/pre><pre class=\"language-javascript\">const { mount } = require('redom');\nimport { App } from '.\/app.js';\n\/\/ create app\nconst app = new App();\n\/\/ init app\napp.update({\n  section: 'a',\n  data: {\n    a: 1,\n    b: 2\n  }\n});\n\/\/ mount to DOM\nmount(document.body, app);\n\/\/ update app after 1 second\nsetTimeout(() =&gt; {\n  app.update({\n    section: 'b',\n    data: {\n      a: 1,\n      b: 2\n    }\n  });\n}, 1000);\n<\/pre><pre class=\"language-markup\">&lt;body&gt;\n  &lt;div class=\"app\"&gt;\n    &lt;div class=\"content\"&gt;\n      &lt;div class=\"a\"&gt;1&lt;\/div&gt;\n    &lt;\/div&gt;\n  &lt;\/div&gt;\n&lt;\/body&gt;<\/pre><h3>Works on server-side as well!<\/h3><pre class=\"language-javascript\">const { Document, render } = require('nodom');\n\/\/ create a fake DOM before requiring RE:DOM\nconst document = new Document();\nconst { el, mount } = require('redom');\nmount(document.body, el('h1', 'Hello world!'));\nconst html = render(document.body);\nconsole.log(html); \/\/ &lt;body&gt;&lt;h1&gt;Hello world!&lt;\/h1&gt;&lt;\/body&gt;<\/pre><h3>Looks fun!<\/h3><p class=\"violator solid\">\n        <a href=\"https:\/\/redom.js.org\/documentation\">Documentation<\/a>\n      <\/p><p class=\"violator\">\n        <a href=\"https:\/\/github.com\/redom\/redom\">GitHub<\/a>\n      <\/p><div><img src=\"https:\/\/redom.js.org\/meme.jpg\" alt=\"image\" \/><\/div><\/section><\/header>","created_at":"2019-01-20T20:36:49+0000","updated_at":"2019-01-20T20:36:49+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"redom.js.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":781,"title":"The Art of Unix Programming","url":"https:\/\/arp242.net\/the-art-of-unix-programming\/","content":"The Art of Unix Programming\n<div class=\"legal\"><p>This book and its on-line version are distributed under the terms of the Creative Commons\n\tAttribution-NoDerivs 1.0 license, with the additional proviso that the right to publish it on paper for sale or\n\tother for-profit use is reserved to Pearson Education, Inc. A reference copy of this license may be found at\n\t<a href=\"http:\/\/creativecommons.org\/licenses\/by-nd\/1.0\/legalcode\">\n\t\thttp:\/\/creativecommons.org\/licenses\/by-nd\/1.0\/legalcode<\/a>.\n\t<\/p><p>AIX, AS\/400, DB\/2, OS\/2, System\/360, MVS, VM\/CMS, and IBM PC are trademarks of IBM. Alpha, DEC, VAX, HP-UX,\n\tPDP, TOPS-10, TOPS-20, VMS, and VT-100 are trademarks of Compaq. Amiga and AmigaOS are trademarks of Amiga, Inc.\n\tApple, Macintosh, MacOS, Newton, OpenDoc, and OpenStep are trademarks of Apple Computers, Inc. ClearCase is a\n\ttrademark of Rational Software, Inc. Ethernet is a trademark of 3COM, Inc. Excel, MS-DOS, Microsoft Windows and\n\tPowerPoint are trademarks of Microsoft, Inc. Java. J2EE, JavaScript, NeWS, and Solaris are trademarks of Sun\n\tMicrosystems. SPARC is a trademark of SPARC international. Informix is a trademark of Informix software. Itanium\n\tis a trademark of Intel. Linux is a trademark of Linus Torvalds. Netscape is a trademark of AOL. PDF and\n\tPostScript are trademarks of Adobe, Inc. UNIX is a trademark of The Open Group.<\/p><p>The photograph of Ken and Dennis in <a href=\"#historychapter\" title=\"Chapter 2. History\">Chapter 2<\/a> appears courtesy of Bell Labs\/Lucent\n\t\tTechnologies.\n\t<\/p><p>The epigraph on the Portability chapter is from the Bell System Technical Journal, v57 #6 part 2 (July-Aug.\n\t\t1978) pp. 2021-2048 and is reproduced with the permission of Bell Labs\/Lucent Technologies.<\/p><\/div>\n<div class=\"dedication\"><strong>Dedication<\/strong><p>To Ken Thompson and Dennis Ritchie, because you inspired me.<\/p><\/div>\n<div id=\"toc\">\u00bb<h2>Table of Contents<\/h2><ol type=\"I\" start=\"0\"><li><a href=\"#preface\">Preface<\/a><ol type=\"a\"><li><a href=\"#pr01s01\">Who Should Read This Book<\/a><\/li>\n\t\t\t<li><a href=\"#pr01s02\">How to Use This Book<\/a><\/li>\n\t\t\t<li><a href=\"#pr01s03\">Related References<\/a><\/li>\n\t\t\t<li><a href=\"#pr01s04\">Conventions Used in This Book<\/a><\/li>\n\t\t\t<li><a href=\"#pr01s05\">Our Case Studies<\/a><\/li>\n\t\t\t<li><a href=\"#pr01s06\">Author's Acknowledgements<\/a><\/li>\n\t\t<\/ol><\/li>\n\t\t<li><a href=\"#context\">Context<\/a><ol start=\"1\"><li><a href=\"#philosophychapter\">Philosophy<\/a><ol type=\"a\"><li><a href=\"#ch01s01\">Culture? What Culture?<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch01s02\">The Durability of Unix<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch01s03\">The Case against Learning Unix Culture<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch01s04\">What Unix Gets Wrong<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch01s05\">What Unix Gets Right<\/a><ol type=\"a\"><li><a href=\"#id2808846\">Open-Source Software<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2872776\">Cross-Platform Portability and Open Standards<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2872827\">The Internet and the World Wide Web<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2872945\">The Open-Source Community<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2873031\">Flexibility All the Way Down<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2873078\">Unix Is Fun to Hack<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2873180\">The Lessons of Unix Can Be Applied Elsewhere<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch01s06\">Basics of the Unix Philosophy<\/a><ol type=\"a\"><li><a href=\"#id2877537\">Rule of Modularity: Write simple parts connected by clean interfaces.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2877610\">Rule of Clarity: Clarity is better than cleverness.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2877684\">Rule of Composition: Design programs to be connected with other programs.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2877777\">Rule of Separation: Separate policy from mechanism; separate interfaces from engines.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2877917\">Rule of Simplicity: Design for simplicity; add complexity only where you must.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878022\">Rule of Parsimony: Write a big program only when it is clear by demonstration that nothing else will do.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878054\">Rule of Transparency: Design for visibility to make inspection and debugging easier.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878145\">Rule of Robustness: Robustness is the child of transparency and simplicity.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878263\">Rule of Representation: Fold knowledge into data, so program logic can be stupid and robust.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878339\">Rule of Least Surprise: In interface design, always do the least surprising thing.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878450\">Rule of Silence: When a program has nothing surprising to say, it should say nothing.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878538\">Rule of Repair: Repair what you can \u2014 but when you must fail, fail noisily and as soon as possible.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878666\">Rule of Economy: Programmer time is expensive; conserve it in preference to machine time.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2878742\">Rule of Generation: Avoid hand-hacking; write programs to write programs when you can.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#rule_of_optimization\">Rule of Optimization: Prototype before polishing. Get it working before you optimize it.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2879078\">Rule of Diversity: Distrust all claims for one true way.<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2879112\">Rule of Extensibility: Design for the future, because it will be here sooner than you think.<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch01s07\">The Unix Philosophy in One Lesson<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch01s08\">Applying the Unix Philosophy<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch01s09\">Attitude Matters Too<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#historychapter\">History<\/a><ol type=\"a\"><li><a href=\"#ch02s01\">Origins and History of Unix, 1969-1995<\/a><ol type=\"a\"><li><a href=\"#genesis\">Genesis: 1969\u20131971<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2879627\">Exodus: 1971\u20131980<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2880014\">TCP\/IP and the Unix Wars: 1980-1990<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2886020\">Blows against the Empire: 1991-1995<\/a><\/li>\n\t\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#hackers\">Origins and History of the Hackers, 1961-1995<\/a><ol type=\"a\"><li><a href=\"#id2886621\">At Play in the Groves of Academe: 1961-1980<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2886828\">Internet Fusion and the Free Software Movement: 1981-1991<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#linux_reaction\">Linux and the Pragmatist Reaction: 1991-1998<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch02s03\">The Open-Source Movement: 1998 and Onward<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch02s04\">The Lessons of Unix History<\/a><\/li>\n\t\t\t<\/ol><\/li><li><a href=\"#contrastchapter\">Contrasts<\/a><ol type=\"a\"><li><a href=\"#ch03s01\">The Elements of Operating-System Style<\/a><ol type=\"a\"><li><a href=\"#id2892028\">What Is the Operating System's Unifying Idea?<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2892085\">Multitasking Capability<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2892171\">Cooperating Processes<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2888136\">Internal Boundaries<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2888217\">File Attributes and Record Structures<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2888298\">Binary File Formats<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2888363\">Preferred User Interface Style<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2888484\">Intended Audience<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2888581\">Entry Barriers to Development<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch03s02\">Operating-System Comparisons<\/a><ol type=\"a\"><li><a href=\"#vms\">VMS<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#mac_os_contrast\">MacOS<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#os_2\">OS\/2<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#nt_contrast\">Windows NT<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#beos\">BeOS<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#mvs\">MVS<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2893777\">VM\/CMS<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#linux\">Linux<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch03s03\">What Goes Around, Comes Around<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t<\/ol><\/li>\n\t\t<li><a href=\"#design\">Design<\/a><ol start=\"4\"><li><a href=\"#modularitychapter\">Modularity<\/a><ol type=\"a\"><li><a href=\"#ch04s01\">Encapsulation and Optimal Module Size<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch04s02\">Compactness and Orthogonality<\/a><ol type=\"a\"><li><a href=\"#compactness\">Compactness<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#orthogonality\">Orthogonality<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#spot_rule\">The SPOT Rule<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2895445\">Compactness and the Strong Single Center<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2899405\">The Value of Detachment<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch04s03\">Software Is a Many-Layered Thing<\/a><ol type=\"a\"><li><a href=\"#id2899552\">Top-Down versus Bottom-Up<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2899777\">Glue Layers<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#c_thin_glue\">Case Study: C Considered as Thin Glue<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch04s04\">Libraries<\/a><ol type=\"a\"><li><a href=\"#gimp_plugins\">Case Study: GIMP Plugins<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#unix_and_oo\">Unix and Object-Oriented Languages<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch04s06\">Coding for Modularity<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#textualitychapter\">Textuality<\/a><ol type=\"a\"><li><a href=\"#ch05s01\">The Importance of Being Textual<\/a><ol type=\"a\"><li><a href=\"#id2901332\">Case Study: Unix Password File Format<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2901494\">Case Study: .newsrc Format<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#png\">Case Study: The PNG Graphics File Format<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch05s02\">Data File Metaformats<\/a><ol type=\"a\"><li><a href=\"#id2901882\">DSV Style<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2902039\">RFC 822 Format<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2902164\">Cookie-Jar Format<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2906931\">Record-Jar Format<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2907018\">XML<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2907263\">Windows INI Format<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2907428\">Unix Textual File Format Conventions<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2907903\">The Pros and Cons of File Compression<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch05s03\">Application Protocol Design<\/a><ol type=\"a\"><li><a href=\"#id2908194\">Case Study: SMTP, the Simple Mail Transfer Protocol<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2908383\">Case Study: POP3, the Post Office Protocol<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2908582\">Case Study: IMAP, the Internet Message Access Protocol<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch05s04\">Application Protocol Metaformats<\/a><ol type=\"a\"><li><a href=\"#id2908835\">The Classical Internet Application Metaprotocol<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2908915\">HTTP as a Universal Application Protocol<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2909217\">BEEP: Blocks Extensible Exchange Protocol<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2909294\">XML-RPC, SOAP, and Jabber<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#transparencychapter\">Transparency<\/a><ol type=\"a\"><li><a href=\"#ch06s01\">Studying Cases<\/a><ol type=\"a\"><li><a href=\"#audacity\">Case Study: audacity<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#fetchmail_v\">Case Study: fetchmail's -v option<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2909841\">Case Study: GCC<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2909954\">Case Study: kmail<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2910193\">Case Study: SNG<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2910334\">Case Study: The Terminfo Database<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2914115\">Case Study: Freeciv Data Files<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch06s02\">Designing for Transparency and Discoverability<\/a><ol type=\"a\"><li><a href=\"#zen_of_transparency\">The Zen of Transparency<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2914509\">Coding for Transparency and Discoverability<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2914680\">Transparency and Avoiding Overprotectiveness<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2914758\">Transparency and Editable Representations<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2915107\">Transparency, Fault Diagnosis, and Fault Recovery<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch06s03\">Designing for Maintainability<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#multiprogramchapter\">Multiprogramming<\/a><ol type=\"a\"><li><a href=\"#ch07s01\">Separating Complexity Control from Performance Tuning<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch07s02\">Taxonomy of Unix IPC Methods<\/a><ol type=\"a\"><li><a href=\"#id2915475\">Handing off Tasks to Specialist Programs<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#plumbing\">Pipes, Redirection, and Filters<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2921506\">Wrappers<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2921634\">Security Wrappers and Bernstein Chaining<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2922002\">Slave Processes<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2922148\">Peer-to-Peer Inter-Process Communication<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch07s03\">Problems and Methods to Avoid<\/a><ol type=\"a\"><li><a href=\"#id2923376\">Obsolescent Unix IPC Methods<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2923675\">Remote Procedure Calls<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2923889\">Threads \u2014 Threat or Menace?<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch07s04\">Process Partitioning at the Design Level<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#minilanguageschapter\">Minilanguages<\/a><ol type=\"a\"><li><a href=\"#ch08s01\">Understanding the Taxonomy of Languages<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch08s02\">Applying Minilanguages<\/a><ol type=\"a\"><li><a href=\"#id2924747\">Case Study: sng<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#regexps\">Case Study: Regular Expressions<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2933450\">Case Study: Glade<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2933775\">Case Study: m4<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2934034\">Case Study: XSLT<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2934197\">Case Study: The Documenter's Workbench Tools<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#fetchmailrc\">Case Study: fetchmail Run-Control Syntax<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#awk\">Case Study: awk<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2935613\">Case Study: PostScript<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2935779\">Case Study: bc and dc<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#emacs_lisp_study\">Case Study: Emacs Lisp<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#javascript\">Case Study: JavaScript<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch08s03\">Designing Minilanguages<\/a><ol type=\"a\"><li><a href=\"#id2936413\">Choosing the Right Complexity Level<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2936650\">Extending and Embedding Languages<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2936912\">Writing a Custom Grammar<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#macroexpansion\">Macros \u2014 Beware!<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2937424\">Language or Application Protocol?<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#generationchapter\">Generation<\/a><ol type=\"a\"><li><a href=\"#ch09s01\">Data-Driven Programming<\/a><ol type=\"a\"><li><a href=\"#id2939746\">Case Study: ascii<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#bayes_spam\">Case Study: Statistical Spam Filtering<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#fetchmailconf\">Case Study: Metaclass Hacking in fetchmailconf<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch09s02\">Ad-hoc Code Generation<\/a><ol type=\"a\"><li><a href=\"#id2938615\">Case Study: Generating Code for the ascii Displays<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#htmlgen\">Case Study: Generating HTML Code for a Tabular List<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#configurationchapter\">Configuration<\/a><ol type=\"a\"><li><a href=\"#ch10s01\">What Should Be Configurable?<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch10s02\">Where Configurations Live<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch10s03\">Run-Control Files<\/a><ol type=\"a\"><li><a href=\"#id2942210\">Case Study: The .netrc File<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2942358\">Portability to Other Operating Systems<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch10s04\">Environment Variables<\/a><ol type=\"a\"><li><a href=\"#id2942463\">System Environment Variables<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2942749\">User Environment Variables<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2942882\">When to Use Environment Variables<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2947980\">Portability to Other Operating Systems<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch10s05\">Command-Line Options<\/a><ol type=\"a\"><li><a href=\"#id2948149\">The -a to -z of Command-Line Options<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2950162\">Portability to Other Operating Systems<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch10s06\">How to Choose among the Methods<\/a><ol type=\"a\"><li><a href=\"#fetchmail_environment\">Case Study: fetchmail<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2950578\">Case Study: The XFree86 Server<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch10s07\">On Breaking These Rules<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#interfacechapter\">Interfaces<\/a><ol type=\"a\"><li><a href=\"#ch11s01\">Applying the Rule of Least Surprise<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch11s02\">History of Interface Design on Unix<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch11s03\">Evaluating Interface Designs<\/a><ol type=\"a\"><li><a href=\"#ch11s04\">Tradeoffs between CLI and Visual Interfaces<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2951734\">Case Study: Two Ways to Write a Calculator Program<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch11s05\">Transparency, Expressiveness, and Configurability<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch11s06\">Unix Interface Design Patterns<\/a><ol type=\"a\"><li><a href=\"#id2957637\">The Filter Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2957916\">The Cantrip Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2958032\">The Source Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2958116\">The Sink Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2958199\">The Compiler Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2958336\">The ed pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2958491\">The Roguelike Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2958899\">The \u2018Separated Engine and Interface\u2019 Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2959821\">The CLI Server Pattern<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2959928\">Language-Based Interface Patterns<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch11s07\">Applying Unix Interface-Design Patterns<\/a><ol type=\"a\"><li><a href=\"#id2960228\">The Polyvalent-Program Pattern<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch11s08\">The Web Browser as a Universal Front End<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch11s09\">Silence Is Golden<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#optimizationchapter\">Optimization<\/a><ol type=\"a\"><li><a href=\"#ch12s01\">Don't Just Do Something, Stand There!<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch12s02\">Measure before Optimizing<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch12s03\">Nonlocality Considered Harmful<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch12s04\">Throughput vs. Latency<\/a><ol type=\"a\"><li><a href=\"#id2961306\">Batching Operations<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2961372\">Overlapping Operations<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#binary_caches\">Caching Operation Results<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#complexitychapter\">Complexity<\/a><ol type=\"a\"><li><a href=\"#ch13s01\">Speaking of Complexity<\/a><ol type=\"a\"><li><a href=\"#id2964646\">The Three Sources of Complexity<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2964843\">Tradeoffs between Interface and Implementation Complexity<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2961759\">Essential, Optional, and Accidental Complexity<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2961870\">Mapping Complexity<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2963237\">When Simplicity Is Not Enough<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch13s02\">A Tale of Five Editors<\/a><ol type=\"a\"><li><a href=\"#id2963445\">ed<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#vi_complexity\">vi<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2963798\">Sam<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#emacs_editing\">Emacs<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2967110\">Wily<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch13s03\">The Right Size for an Editor<\/a><ol type=\"a\"><li><a href=\"#id2967267\">Identifying the Complexity Problems<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2967642\">Compromise Doesn't Work<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2967765\">Is Emacs an Argument against the Unix Tradition?<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch13s04\">The Right Size of Software<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t<\/ol><\/li>\n\t\t<li><a href=\"#implementation\">Implementation<\/a><ol start=\"14\"><li><a href=\"#languageschapter\">Languages<\/a><ol type=\"a\"><li><a href=\"#ch14s01\">Unix's Cornucopia of Languages<\/a><\/li>\n\t\t\t\t<li><a href=\"#why_not_c\">Why Not C?<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch14s03\">Interpreted Languages and Mixed Strategies<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch14s04\">Language Evaluations<\/a><ol type=\"a\"><li><a href=\"#c_language\">C<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#cc_language\">C++<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#sh\">Shell<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#perl\">Perl<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#tcl\">Tcl<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#python_language\">Python<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#java\">Java<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#emacs_lisp_language\">Emacs Lisp<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch14s05\">Trends for the Future<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch14s06\">Choosing an X Toolkit<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#toolschapter\">Tools<\/a><ol type=\"a\"><li><a href=\"#ch15s01\">A Developer-Friendly Operating System<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch15s02\">Choosing an Editor<\/a><ol type=\"a\"><li><a href=\"#vi_literacy\">Useful Things to Know about vi<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2979540\">Useful Things to Know about Emacs<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2979678\">The Antireligious Choice: Using Both<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch15s03\">Special-Purpose Code Generators<\/a><ol type=\"a\"><li><a href=\"#id2979797\">yacc and lex<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2986324\">Case Study: Glade<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch15s04\">make: Automating Your Recipes<\/a><ol type=\"a\"><li><a href=\"#id2986550\">Basic Theory of make<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2986902\">make in Non-C\/C++ Development<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2987148\">Utility Productions<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2987644\">Generating Makefiles<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch15s05\">Version-Control Systems<\/a><ol type=\"a\"><li><a href=\"#id2988357\">Why Version Control?<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2988464\">Version Control by Hand<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2988503\">Automated Version Control<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2988625\">Unix Tools for Version Control<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch15s06\">Runtime Debugging<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch15s07\">Profiling<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch15s08\">Combining Tools with Emacs<\/a><ol type=\"a\"><li><a href=\"#id2989559\">Emacs and make<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2989677\">Emacs and Runtime Debugging<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2989781\">Emacs and Version Control<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2989869\">Emacs and Profiling<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2989926\">Like an IDE, Only Better<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#reusechapter\">Reuse<\/a><ol type=\"a\"><li><a href=\"#ch16s01\">The Tale of J. Random Newbie<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch16s02\">Transparency as the Key to Reuse<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch16s03\">From Reuse to Open Source<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch16s04\">The Best Things in Life Are Open<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch16s05\">Where to Look?<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch16s06\">Issues in Using Open-Source Software<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch16s07\">Licensing Issues<\/a><ol type=\"a\"><li><a href=\"#id2991059\">What Qualifies as Open Source<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2993547\">Standard Open-Source Licenses<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2993874\">When You Need a Lawyer<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t<\/ol><\/li>\n\t\t<li><a href=\"#community\">Community<\/a><ol start=\"17\"><li><a href=\"#portabilitychapter\">Portability<\/a><ol type=\"a\"><li><a href=\"#c_evolution\">Evolution of C<\/a><ol type=\"a\"><li><a href=\"#id2998213\">Early History of C<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2994334\">C Standards<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch17s02\">Unix Standards<\/a><ol type=\"a\"><li><a href=\"#id2994594\">Standards and the Unix Wars<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2999310\">The Ghost at the Victory Banquet<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id2999366\">Unix Standards in the Open-Source World<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ietf_process\">IETF and the RFC Standards Process<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch17s04\">Specifications as DNA, Code as RNA<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch17s05\">Programming for Portability<\/a><ol type=\"a\"><li><a href=\"#id3000303\">Portability and Choice of Language<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3000882\">Avoiding System Dependencies<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3000984\">Tools for Portability<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch17s06\">Internationalization<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch17s07\">Portability, Open Standards, and Open Source<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#documentationchapter\">Documentation<\/a><ol type=\"a\"><li><a href=\"#ch18s01\">Documentation Concepts<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch18s02\">The Unix Style<\/a><ol type=\"a\"><li><a href=\"#id3001392\">The Large-Document Bias<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3001522\">Cultural Style<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch18s03\">The Zoo of Unix Documentation Formats<\/a><ol type=\"a\"><li><a href=\"#id3001604\">troff and the Documenter's Workbench Tools<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3001898\">TeX<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3002206\">Texinfo<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3002268\">POD<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3002291\">HTML<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3002310\">DocBook<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch18s04\">The Present Chaos and a Possible Way Out<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch18s05\">DocBook<\/a><ol type=\"a\"><li><a href=\"#id3006015\">Document Type Definitions<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3006182\">Other DTDs<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#db_toolchain\">The DocBook Toolchain<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3006742\">Migration Tools<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3007022\">Editing Tools<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3007143\">Related Standards and Practices<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3007192\">SGML<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3007297\">XML-DocBook References<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch18s06\">Best Practices for Writing Unix Documentation<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#opensourcechapter\">Open Source<\/a><ol type=\"a\"><li><a href=\"#ch19s01\">Unix and Open Source<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch19s02\">Best Practices for Working with Open-Source Developers<\/a><ol type=\"a\"><li><a href=\"#patching\">Good Patching Practice<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#naming\">Good Project- and Archive-Naming Practice<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#develpractice\">Good Development Practice<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#distpractice\">Good Distribution-Making Practice<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#communication\">Good Communication Practice<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch19s03\">The Logic of Licenses: How to Pick One<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch19s04\">Why You Should Use a Standard License<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch19s05\">Varieties of Open-Source Licensing<\/a><ol type=\"a\"><li><a href=\"#id3014860\">MIT or X Consortium License<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3014890\">BSD Classic License<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3014963\">Artistic License<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3015011\">General Public License<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3015063\">Mozilla Public License<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t<\/ol><\/li>\n\t\t\t<li><a href=\"#futurechapter\">Futures<\/a><ol type=\"a\"><li><a href=\"#ch20s01\">Essence and Accident in Unix Tradition<\/a><\/li>\n\t\t\t\t<li><a href=\"#plan9\">Plan 9: The Way the Future Was<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch20s03\">Problems in the Design of Unix<\/a><ol type=\"a\"><li><a href=\"#id3015538\">A Unix File Is Just a Big Bag of Bytes<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3015724\">Unix Support for GUIs Is Weak<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3015838\">File Deletion Is Forever<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3015895\">Unix Assumes a Static File System<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3015943\">The Design of Job Control Was Badly Botched<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3016077\">The Unix API Doesn't Use Exceptions<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3016155\">ioctl2 and fcntl2 Are an Embarrassment<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3019078\">The Unix Security Model May Be Too Primitive<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3019113\">Unix Has Too Many Different Kinds of Names<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3019140\">File Systems Might Be Considered Harmful<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#id3019193\">Towards a Global Internet Address Space<\/a><\/li>\n\t\t\t\t<\/ol><\/li>\n\t\t\t\t<li><a href=\"#ch20s04\">Problems in the Environment of Unix<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch20s05\">Problems in the Culture of Unix<\/a><\/li>\n\t\t\t\t<li><a href=\"#ch20s06\">Reasons to Believe<\/a>\n\t\t\t<\/li><\/ol><\/li>\n\t\t<\/ol><\/li>\n\t\t<li>Appendix<ol type=\"A\"><li><a href=\"#apa\">Glossary of Abbreviations<\/a><\/li>\n\t\t\t<li><a href=\"#apb\">References<\/a><\/li>\n\t\t\t<li><a href=\"#apc\">Contributors<\/a><\/li>\n\t\t\t<li><a href=\"#revhistory\">Revision History<\/a><\/li>\n\t\t\t<li><a href=\"#unix_koans\">Rootless Root<\/a><ol type=\"a\"><li><a href=\"#introduction\">Editor's Introduction<\/a><\/li>\n\t\t\t\t<li><a href=\"#ten-thousand\">Master Foo and the Ten Thousand Lines<\/a><\/li>\n\t\t\t\t<li><a href=\"#script-kiddie\">Master Foo and the Script Kiddie<\/a><\/li>\n\t\t\t\t<li><a href=\"#two_paths\">Master Foo Discourses on the Two Paths<\/a><\/li>\n\t\t\t\t<li><a href=\"#methodology-consultant\">Master Foo and the Methodologist<\/a><\/li>\n\t\t\t\t<li><a href=\"#gui-programmer\">Master Foo Discourses on the Graphical User Interface<\/a><\/li>\n\t\t\t\t<li><a href=\"#zealot\">Master Foo and the Unix Zealot<\/a><\/li>\n\t\t\t\t<li><a href=\"#unix-nature\">Master Foo Discourses on the Unix-Nature<\/a><\/li>\n\t\t\t\t<li><a href=\"#end-user\">Master Foo and the End User<\/a><\/li>\n\t\t\t<\/ol><\/li>\n\t\t<\/ol><\/li>\n\t<\/ol><\/div>\n\n <div class=\"preface\"><h3 id=\"preface\"><a href=\"#preface\">\u00a7<\/a>Preface<\/h3><blockquote class=\"epigraph\">\n   <p>\n    Unix is not so much an operating system as an oral history.\n   <\/p><div>\n   \n    \u2013 Neal Stephenson\n   \n  <\/div><\/blockquote><p>\n   There is a vast difference between knowledge and expertise.\nKnowledge lets you deduce the right thing to do; expertise makes the\nright thing a reflex, hardly requiring conscious thought at\nall.\n  <\/p><p>\n   This book has a lot of knowledge in it, but it is mainly about\nexpertise. It is going to try to teach you the things about Unix\ndevelopment that Unix experts know, but aren't aware that they\nknow. It is therefore less about technicalia and more about\n   <em>shared culture<\/em>\n   than most Unix books \u2014 both\nexplicit and implicit culture, both conscious and unconscious\ntraditions.  It is not a \u2018how-to\u2019 book, it is a\n\u2018why-to\u2019 book.\n  <\/p><p>\n   The why-to has great practical importance, because far too much\nsoftware is poorly designed.  Much of it suffers from bloat, is\nexceedingly hard to maintain, and is too difficult to port to new\nplatforms or extend in ways the original programmers didn't\nanticipate.  These problems are symptoms of bad design.  We hope that\nreaders of this book will learn something of what Unix has to teach\nabout good design.\n  <\/p><p>\n   This book is divided into four parts: Context, Design, Tools,\nand Community.  The first part (Context) is philosophy and history, to\nhelp provide foundation and motivation for what follows.  The second\npart (Design) unfolds the principles of the Unix philosophy into more\nspecific advice about design and implementation.  The third part\n(Tools) focuses on the software Unix provides for helping you solve\nproblems.  The fourth part (Community) is about the human-to-human\ntransactions and agreements that make the Unix culture so effective at\nwhat it does.\n  <\/p><p>\n   Because this is a book about shared culture, I never planned to\nwrite it alone.  You will notice that the text includes guest\nappearances by prominent Unix developers, the shapers of the Unix\ntradition.  The book went through an extended public review process\nduring which I invited these luminaries to comment on and argue with\nthe text.  Rather than submerging the results of that review process\nin the final version, these guests were encouraged to speak with their\nown voices, amplifying and developing and even disagreeing with the\nmain line of the text.\n  <\/p><p>\n   In this book, when I use the editorial \u2018we\u2019 it is\nnot to pretend omniscience but to reflect the fact that it attempts to\narticulate the expertise of an entire community.\n  <\/p><p>\n   Because this book is aimed at transmitting culture, it includes\nmuch more in the way of history and folklore and asides than is normal\nfor a technical book. Enjoy; these things, too, are part of your\neducation as a Unix programmer.  No single one of the historical\ndetails is vital, but the gestalt of them all is important.  We think\nit makes a more interesting story this way.  More\nimportantly, understanding where Unix came from and how it got the\nway it is will help you develop an intuitive feel for the Unix\nstyle.\n  <\/p><p>\n   For the same reason, we refuse to write as if history is over.\nYou will find an unusually large number of references to the time of\nwriting in this book.  We do not wish to pretend that current practice\nreflects some sort of timeless and perfectly logical outcome of\npreordained destiny. References to time of writing are meant as an\nalert to the reader two or three or five years hence that the\nassociated statements of fact may have become dated and should be\ndouble-checked.\n  <\/p><p>\n   Other things this book is not is neither a C tutorial, nor a\nguide to the Unix commands and API. It is not a reference for\n   <em>sed<\/em> or <em>yacc<\/em> or\nPerl or Python. It's not a network programming primer, nor an\nexhaustive guide to the mysteries of X.  It's not a tour of Unix's\ninternals and architecture, either. Other books cover these specifics\nbetter, and this book points you at them as appropriate.\n  <\/p><p>\n   Beyond all these technical specifics, the Unix culture has an\nunwritten engineering tradition that has developed over literally\nmillions of man-years<sup><a href=\"#ftn.id2871746\" id=\"id2871746\">[1]<\/a><\/sup>\n   of skilled effort. This\nbook is written in the belief that understanding that tradition, and\nadding its design patterns to your toolkit, will help you become a\nbetter programmer and designer.\n  <\/p><p>\n   Cultures consist of people, and the traditional way to learn\nUnix culture is from other people and through the folklore, by\nosmosis.  This book is not a substitute for person-to-person\nacculturation, but it can help accelerate the process by allowing you\nto tap the experience of others.\n  <\/p><\/div>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2871746\" id=\"ftn.id2871746\">[1]<\/a><\/sup>\n  The three and a half decades\nbetween 1969 and 2003 is a long time.  Going by the historical\ntrend curve in number of Unix sites during that period, probably\nsomewhere upwards of fifty million man-years have been plowed into\nUnix development worldwide.\n<\/p><\/div>\n\n<h3 id=\"pr01s01\"><a href=\"#pr01s01\">\u00a7<\/a>Who Should Read This Book<\/h3>\n<p>\n  You should read this book if you are an experienced Unix\nprogrammer who is often in the position of either educating novice\nprogrammers or debating partisans of other operating systems, and you\nfind it hard to articulate the benefits of the Unix approach.\n<\/p>\n<p>\n  You should read this book if you are a C, C++, or Java\nprogrammer with experience on other operating systems and you are\nabout to start a Unix-based project.\n<\/p>\n<p>\n  You should read this book if you are a Unix user with\nnovice-level up to middle-level skills in the operating system, but\nlittle development experience, and want to learn how to design\nsoftware effectively under Unix.\n<\/p>\n<p>\n  You should read this book if you are a non-Unix programmer who\nhas figured out that the Unix tradition might have something to\nteach you. We believe you're right, and that the Unix philosophy can\nbe exported to other operating systems. So we will pay more\nattention to non-Unix environments (especially Microsoft operating\nsystems) than is usual in a Unix book; and when tools and case\nstudies are portable, we say so.\n<\/p>\n<p>\n  You should read this book if you are an application architect\nconsidering platforms or implementation strategies for a major\ngeneral-market or vertical application.  It will help you\nunderstand the strengths of Unix as a development platform, and\nof the Unix tradition of open source as a development method.\n<\/p>\n<p>\n  You should <em>not<\/em> read this book if what you\nare looking for is the details of C coding or how to use the Unix\nkernel API.  There are many good books on these topics;\n  <em>Advanced Programming in the Unix Environment<\/em> <sup><a href=\"#Stevens92\" title=\"[Stevens92]\">[Stevens92]<\/a><\/sup>\n  is classic among explorations of the Unix\nAPI, and\n  <em>The Practice of Programming<\/em> <sup><a href=\"#Kernighan-Pike99\" title=\"[Kernighan-Pike99]\">[Kernighan-Pike99]<\/a><\/sup>\n  is recommended reading for all C programmers (indeed for all programmers in any language).\n<\/p>\n\n<h3 id=\"pr01s02\"><a href=\"#pr01s02\">\u00a7<\/a>How to Use This Book<\/h3>\n<p>\n  This book is both practical and philosophical. Some parts are\naphoristic and general, others will examine specific case studies in\nUnix development. We will precede or follow general principles and\naphorisms with examples that illustrate them: examples drawn not from\ntoy demonstration programs but rather from real working code that is\nin use every day.\n<\/p>\n<p>\n  We have deliberately avoided filling the book with lots of code\nor specification-file examples, even though in many places this\nmight have made it easier to write (and in some places perhaps\neasier to read!). Most books about programming give too many\nlow-level details and examples, but fail at giving the reader a\nhigh-level feel for what is really going on. In this book, we\nprefer to err in the opposite direction.\n<\/p>\n<p>\n  Therefore, while you will often be invited to read code and\nspecification files, relatively few are actually included in the book.\nInstead, we point you at examples on the Web.\n<\/p>\n<p>\n  Absorbing these examples will help solidify the principles you\nlearn into semi-instinctive working knowledge. Ideally, you should\nread this book near the console of a running Unix system, with a Web\nbrowser handy. Any Unix will do, but the software case studies are\nmore likely to be preinstalled and immediately available for\ninspection on a Linux system. The pointers in the book are invitations\nto browse and experiment. Introduction of these pointers is paced so\nthat wandering off to explore for a while won't break up exposition\nthat has to be continuous.\n<\/p>\n<p>\n  Note: While we have made every effort to cite URLs that should\nremain stable and usable, there is no way we can guarantee this. If\nyou find that a cited link has gone stale, use common sense and do\na phrase search with your favorite Web search engine. Where possible\nwe suggest ways to do this near the URLs we cite.\n<\/p>\n<p>\n  Most abbreviations used in this book are expanded at first use.\nFor convenience, we have also provided a glossary in an appendix.\n<\/p>\n<p>\n  References are usually by author name.  Numbered footnotes are\nfor URLs that would intrude on the text or that we suspect might be\nperishable; also for asides, war stories, and jokes.<sup><a href=\"#ftn.id2807786\" id=\"id2807786\">[2]<\/a><\/sup><\/p>\n<p>\n  To make this book more accessible to less technical readers, we\ninvited some non-programmers to read it and identify terms that seemed\nboth obscure and necessary to the flow of exposition.  We also use\nfootnotes for definitions of elementary terms that an experienced\nprogrammer is unlikely to need.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2807786\" id=\"ftn.id2807786\">[2]<\/a><\/sup>\n  This particular footnote is dedicated to Terry Pratchett, whose\nuse of footnotes is quite...inspiring.\n<\/p><\/div>\n\n<h3 id=\"pr01s03\"><a href=\"#pr01s03\">\u00a7<\/a>Related References<\/h3>\n<p>\n  Some famous papers and a few books by Unix's early developers\nhave mined this territory before. Kernighan and Pike's\n  <em>The Unix Programming Environment<\/em> <sup><a href=\"#Kernighan-Pike84\" title=\"[Kernighan-Pike84]\">[Kernighan-Pike84]<\/a><\/sup> stands out\namong these and is rightly considered a classic. But today it shows\nits age a bit; it doesn't cover the Internet, and the World Wide Web\nor the new wave of interpreted languages like Perl, Tcl, and\nPython.\n<\/p>\n<p>\n  About halfway into the composition of this book, we learned of\nMike Gancarz's\n  <em>The Unix Philosophy<\/em> <sup><a href=\"#Gancarz\" title=\"[Gancarz]\">[Gancarz]<\/a><\/sup>.  This book is excellent within its\nrange, but did not attempt to cover the full spectrum of topics we\nfelt needed to be addressed.  Nevertheless we are grateful to the\nauthor for the reminder that the very simplest Unix design patterns\nhave been the most persistent and successful ones.\n<\/p>\n<p>\n  <em>The Pragmatic Programmer<\/em> <sup><a href=\"#Hunt-Thomas\" title=\"[Hunt-Thomas]\">[Hunt-Thomas]<\/a><\/sup> is a witty and wise\ndisquisition on good design practice pitched at a slightly\ndifferent level of the software-design craft (more about coding, less\nabout higher-level partitioning of problems) than this book. The\nauthors' philosophy is an outgrowth of Unix experience, and it is an\nexcellent complement to this book.\n<\/p>\n<p>\n  <em>The Practice of Programming<\/em> <sup><a href=\"#Kernighan-Pike99\" title=\"[Kernighan-Pike99]\">[Kernighan-Pike99]<\/a><\/sup> covers some of\nthe same ground as\n  <em>The Pragmatic Programmer<\/em> from a position deep within the Unix tradition.\n<\/p>\n<p>\n  Finally (and with admitted intent to provoke) we recommend\n  <em>Zen Flesh, Zen Bones<\/em> <sup><a href=\"#Reps-Senzaki\" title=\"[Reps-Senzaki]\">[Reps-Senzaki]<\/a><\/sup>, an important collection of\nZen Buddhist primary sources.  References to Zen are scattered throughout\nthis book.  They are included because Zen provides a vocabulary for\naddressing some ideas that turn out to be very important for software\ndesign but are otherwise very difficult to hold in the mind.  Readers\nwith religious attachments are invited to consider Zen not as a\nreligion but as a therapeutic form of mental discipline \u2014 which,\nin its purest non-theistic forms, is exactly what Zen is.\n<\/p>\n\n<h3 id=\"pr01s04\"><a href=\"#pr01s04\">\u00a7<\/a>Conventions Used in This Book<\/h3>\n<p>\n  The term \u201cUNIX\u201d is technically and legally a\ntrademark of The Open Group, and should formally be used only for\noperating systems which are certified to have passed The Open Group's\nelaborate standards-conformance tests. In this book we use\n\u201cUnix\u201d in the looser sense widely current among\nprogrammers, to refer to any operating system (whether formally\nUnix-branded or not) that is either genetically descended from Bell\nLabs's ancestral Unix code or written in close imitation of its\ndescendants. In particular, Linux (from which we draw most of our\nexamples) is a Unix under this definition.\n<\/p>\n<p>\n  This book employs the Unix manual page convention of tagging\nUnix facilities with a following manual section in parentheses,\nusually on first introduction when we want to emphasize that this is a\nUnix command.  Thus, for example, read \u201cmunger(1)\u201d as\n\u201cthe \u2018munger\u2019 program, which will be documented in\nsection 1 (user tools) of the Unix manual pages, if it's present on\nyour system\u201d. Section 2 is C system calls, section 3 is C\nlibrary calls, section 5 is file formats and protocols, section 8 is\nsystem administration tools. Other sections vary among Unixes but\nare not cited in this book.  For more, type\n  <code>man 1 man<\/code> at your Unix shell prompt (older System V Unixes may\nrequire <code>man -s 1 man <\/code>).\n<\/p>\n<p>\n  Sometimes we mention a Unix application (such as\n  <em>Emacs<\/em>), without a manual-section suffix and\ncapitalized.  This is a clue that the name actually represents a\nwell-established family of Unix programs with essentially the same\nfunction, and we are discussing generic properties of all of them.\n  <em>Emacs<\/em>, for example, includes\n  <em>xemacs<\/em>.\n<\/p>\n<p>\n  At various points later in this book we refer to\n\u2018old school\u2019 and \u2018new school\u2019 methods.  As\nwith rap music, new-school starts about 1990.  In this context, it's\nassociated with the rise of scripting\nlanguages, GUIs, open-source Unixes, and the\nWeb.  Old-school refers to the pre-1990 (and especially pre-1985)\nworld of expensive (shared) computers, proprietary Unixes, scripting\nin shell, and C everywhere.  This difference is worth pointing out\nbecause cheaper and less memory-constrained machines have wrought some\nsignificant changes on the Unix programming style.\n<\/p>\n\n<h3 id=\"pr01s05\"><a href=\"#pr01s05\">\u00a7<\/a>Our Case Studies<\/h3>\n<p>\n  A lot of books on programming rely on toy examples constructed\nspecifically to prove a point. This one won't. Our case studies\nwill be real, pre-existing pieces of software that are in\nproduction use every day. Here are some of the major ones:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    <em>cdrtools<\/em> \/ <em>xcdroast<\/em>\n   <\/dt>\n   <dd>\n    <p>\n     These two separate projects are usually used together.\nThe <em>cdrtools<\/em> package is a set of CLI tools for writing CD-ROMs; Web search for \u201ccdrtools\u201d.\nThe <em>xcdroast<\/em> application is a GUI front end for cdrtools; see the\n     <a href=\"http:\/\/www.xcdroast.org\/\">\n      xcdroast project site<\/a>.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <em>fetchmail<\/em> \n   <\/dt>\n   <dd>\n    <p>\n     The\n     <em>fetchmail<\/em> program\nretrieves mail from remote-mail servers using the POP3 or IMAP\npost-office protocols. See the\n     <a href=\"http:\/\/www.catb.org\/~esr\/fetchmail\">\n      fetchmail home page<\/a>\n     (or search for\n\u201cfetchmail\u201d on the Web).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <em>GIMP<\/em> \n   <\/dt>\n   <dd>\n    <p>\n     The\n     <em>GIMP<\/em> (GNU Image\nManipulation Program) is a full-featured paint, draw, and\nimage-manipulation program that can edit a huge variety of graphical\nformats in sophisticated ways. Sources are available from the\n     <a href=\"http:\/\/www.gimp.org\/\">\n      GIMP home page<\/a>\n     (or search for\n\"GIMP\" on the Web).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <em>mutt<\/em> \n   <\/dt>\n   <dd>\n    <p>\n     The\n     <em>mutt<\/em> mail user agent is the current\nbest-of-breed among text-based Unix electronic mail agents, with notably good\nsupport for MIME\n     (Multipurpose Internet Mail Extensions) and the use of privacy aids\nsuch as PGP (Pretty Good Privacy) and GPG (GNU Privacy Guard).  Source\ncode and executable binaries are available at the\n     <a href=\"http:\/\/www.mutt.org\">\n      Mutt project site<\/a>.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <em>xmlto<\/em> \n   <\/dt>\n   <dd>\n    <p>\n     The\n     <em>xmlto<\/em> command renders\nDocBook and other XML documents in various output formats, including\nHTML and text and PostScript.  For sources and documentation, see the\n     <a href=\"http:\/\/cyberelk.net\/tim\/xmlto\/\">\n      xmlto project site<\/a>.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  To minimize the amount of code the user needs to read to\nunderstand the examples, we have tried to choose case studies that can\nbe used more than once, ideally to illustrate several different design\nprinciples and practices.  For this same reason, many of the examples\nare from my projects.  No claim that these are the best possible ones\nis implied, merely that I find them sufficiently familiar to be useful\nfor multiple expository purposes.\n<\/p>\n\n<h3 id=\"pr01s06\"><a href=\"#pr01s06\">\u00a7<\/a>Author's Acknowledgements<\/h3>\n<p>\n  The guest contributors (Ken Arnold, Steven M. Bellovin, Stuart\nFeldman, Jim Gettys, Steve Johnson, Brian Kernighan, David Korn, Mike\nLesk, Doug McIlroy, Marshall Kirk McKusick, Keith Packard, Henry\nSpencer, and Ken Thompson) added a great deal of value to this\nbook. Doug McIlroy, in particular, went far beyond the call of duty in\nthe thoroughness of his critique and the depth of his contributions,\ndisplaying the same care and dedication to excellence which he brought\nto managing the original Unix research group thirty years ago.\n<\/p>\n<p>\n  Special thanks go to Rob Landley and to my wife Catherine\nRaymond, both of whom delivered intensive line-by-line critiques of\nmanuscript drafts.  Rob's insightful and attentive commentary actually\ninspired more than one entire chapter in the final manuscript, and he\nhad a lot to do with its present organization and range; if he had\nwritten all the text he pushed me to improve, I would have to call him\na co-author. Cathy was my test audience representing non-technical\nreaders; to the extent this book is accessible to people who aren't\nalready programmers, that's largely her doing.\n<\/p>\n<p>\n  This book benefited from discussions with many other people over\nthe five years it took me to write it. Mark M. Miller helped me\nachieve enlightenment about threads. John Cowan supplied some insights\nabout interface design patterns and drafted the case studies of\n  <em>wily<\/em> and VM\/CMS, and Jef Raskin showed me\nwhere the Rule of Least Surprise comes from. The UIUC System\nArchitecture Group contributed useful feedback on early chapters. The\nsections on\n  <em>What Unix Gets Wrong<\/em> and\n  <em>Flexibility in Depth<\/em> were directly inspired by\ntheir review. Russell J. Nelson contributed the material on Bernstein\nchaining in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>.\nJay Maynard contributed most of the material in the MVS case study in\n  <a href=\"#contrastchapter\" title=\"Chapter 3. Contrasts\">\n   Chapter 3<\/a>.  Les Hatton provided many helpful\ncomments on the <em>Languages<\/em> chapter and motivated\nthe portion of <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">Chapter 4<\/a> on <em>Optimal Module Size<\/em>.\n  David A. Wheeler contributed many perceptive criticisms and some case-study material, especially in the Design part.\n  Russ Cox helped develop the survey of Plan 9. Dennis Ritchie corrected me on some historical points about C.\n<\/p>\n<p>\n  Hundreds of Unix programmers, far too many to list here,\ncontributed advice and comments during the book's public review period\nbetween January and June of 2003.  As always, I found the process of\nopen peer review over the Web both intensely challenging and intensely\nrewarding.  Also as always, responsibility for any errors in the\nresulting work remains my own.\n<\/p>\n<p>\n  The expository style and some of the concerns of this book have\nbeen influenced by the design patterns\n  movement; indeed, I flirted with the\nidea of titling the book\n  <em>Unix Design Patterns<\/em>. I didn't, because I disagree with some of the\nimplicit central dogmas of the movement and don't feel the need to use\nall its formal apparatus or accept its cultural baggage. Nevertheless,\nmy approach has certainly been influenced by Christopher Alexander's\nwork<sup><a href=\"#ftn.id2808454\" id=\"id2808454\">[3]<\/a><\/sup>\n  (especially\n  <em>The Timeless Way of Building<\/em> and\n  <em>A Pattern Language<\/em>), and I owe the Gang of Four and other\nmembers of their school a large debt of gratitude for showing me how\nit is possible to use Alexander's insights to talk about software\ndesign at a high level without merely uttering vague and useless\ngeneralities. Interested readers should see\n  <em>Design Patterns: Elements of Reusable Object-Oriented Software<\/em><sup><a href=\"#GangOfFour\" title=\"[GangOfFour]\">[GangOfFour]<\/a><\/sup> for an introduction to design\npatterns.\n<\/p>\n<p>\n  The title of this book is, of course, a reference to Donald\nKnuth's\n  <em>The Art of Computer Programming<\/em>.  While not\nspecifically associated with the Unix tradition, Knuth has been an\ninfluence on us all.\n<\/p>\n<p>\n  Editors with vision and imagination aren't as common as they\nshould be. Mark Taub is one; he saw merit in a stalled project and\nskillfully nudged me into finishing it. Copy editors with a good ear\nfor prose style and enough ability to improve writing that isn't like\ntheirs are even less common, but Mary Lou Nohr makes that grade. Jerry\nVotta seized on my concept for the cover and made it look better than I\nhad imagined.  The whole crew at Addison-Wesley gets high marks for\nmaking the editorial and production process as painless as possible,\nand for cheerfully accommodating my control-freak tendencies not just\nover the text but deep into the details of the book's visual design,\nart, and marketing.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2808454\" id=\"ftn.id2808454\">[3]<\/a><\/sup>\n  An appreciation of Alexander's work, with links to\non-line versions of significant portions, may be found at\n  <a href=\"http:\/\/www.math.utsa.edu\/sphere\/salingar\/Chris.text.html\">\n   Some Notes on Christopher Alexander<\/a>.\n<\/p><\/div>\n\n<h2 id=\"context\"><a href=\"#context\">\u00a7<\/a>Context<\/h2>\n\n<h3 id=\"philosophychapter\"><a href=\"#philosophychapter\">\u00a7<\/a>Chapter 1. Philosophy<\/h3>\n <blockquote class=\"epigraph\">\n  <p>Those who do not understand Unix are condemned to reinvent it, poorly.<\/p><div>\n  \u2013 Henry Spencer, Usenet signature, November 1987\n <\/div><\/blockquote>\n\n<h3 id=\"ch01s01\"><a href=\"#ch01s01\">\u00a7<\/a>Culture? What Culture?<\/h3>\n<p>\n  This is a book about Unix programming, but in it we're going to\ntoss around the words \u2018culture\u2019, \u2018art\u2019, and\n\u2018philosophy\u2019 a lot. If you are not a programmer, or you\nare a programmer who has had little contact with the Unix world, this\nmay seem strange. But Unix has a culture; it has a distinctive art of\nprogramming; and it carries with it a powerful design\nphilosophy. Understanding these traditions will help you build better\nsoftware, even if you're developing for a non-Unix platform.\n<\/p>\n<p>\n  Every branch of engineering and design has technical cultures.\nIn most kinds of engineering, the unwritten traditions of the field\nare parts of a working practitioner's education as important as (and,\nas experience grows, often more important than) the official handbooks\nand textbooks. Senior engineers develop huge bodies of implicit\nknowledge, which they pass to their juniors by (as\nZen\n  Buddhists put it)\n\u201ca special transmission, outside the scriptures\u201d.\n<\/p>\n<p>\n  Software engineering is generally an exception to this rule;\ntechnology has changed so rapidly, software environments have come and\ngone so quickly, that technical cultures have been weak and\nephemeral. There are, however, exceptions to this exception. A very\nfew software technologies have proved durable enough to evolve strong\ntechnical cultures, distinctive arts, and an associated design\nphilosophy transmitted across generations of engineers.\n<\/p>\n<p>\n  The Unix culture is one of these. The Internet culture is\nanother \u2014 or, in the twenty-first century, arguably the same\none. The two have grown increasingly difficult to separate since the\nearly 1980s, and in this book we won't try particularly hard.\n<\/p>\n\n<h3 id=\"ch01s02\"><a href=\"#ch01s02\">\u00a7<\/a>The Durability of Unix<\/h3>\n<p>\n  Unix was born in 1969 and has been in continuous production use\never since. That's several geologic eras by computer-industry\nstandards \u2014 older than the PC or workstations or microprocessors\nor even video display terminals, and contemporaneous with the first\nsemiconductor memories. Of all production timesharing systems today,\nonly IBM's VM\/CMS can claim to have existed longer, and Unix machines \nhave provided hundreds of thousands of times more service hours;\nindeed, Unix has probably supported more computing than all other\ntimesharing systems put together.\n<\/p>\n<p>\n  Unix has found use on a wider variety of machines than any other\noperating system can claim. From supercomputers to handhelds and\nembedded networking hardware, through workstations and servers and PCs\nand minicomputers, Unix has probably seen more architectures and more\nodd hardware than any three other operating systems combined.\n<\/p>\n<p>\n  Unix has supported a mind-bogglingly wide spectrum of uses. No\nother operating system has shone simultaneously as a research vehicle,\na friendly host for technical custom applications, a platform for\ncommercial-off-the-shelf business software, and a vital component\ntechnology of the Internet.\n<\/p>\n<p>\n  Confident predictions that Unix would wither away, or be crowded\nout by other operating systems, have been made yearly since its\ninfancy. And yet Unix, in its present-day avatars as\nLinux\n  and BSD and Solaris and MacOS X and half a dozen other variants,\nseems stronger than ever today.\n<\/p>\n <blockquote>\n  <p>\n   Robert Metcalf [the inventor of Ethernet] says that if something\ncomes along to replace Ethernet, it will be called\n\u201cEthernet\u201d, so therefore Ethernet will never\ndie.<sup><a href=\"#ftn.id2876928\" id=\"id2876928\">[4]<\/a><\/sup>\n   Unix has already undergone\nseveral such transformations.\n  <\/p><div>\n  \n   \u2013 Ken Thompson\n  \n <\/div><\/blockquote>\n<p>\n  At least one of Unix's central technologies \u2014 the C\nlanguage \u2014 has been widely naturalized elsewhere. Indeed it is\nnow hard to imagine doing software engineering without\nC\n  as a ubiquitous\ncommon language of systems programming.  Unix also introduced both the\nnow-ubiquitous tree-shaped file namespace with directory nodes and\nthe pipeline for connecting programs.\n<\/p>\n<p>\n  Unix's durability and adaptability have been nothing short of\nastonishing. Other technologies have come and gone like mayflies.\nMachines have increased a thousandfold in power, languages have\nmutated, industry practice has gone through multiple revolutions \u2014\nand Unix hangs in there, still producing, still paying the bills,\nand still commanding loyalty from many of the best and brightest\nsoftware technologists on the planet.\n<\/p>\n<p>\n  One of the many consequences of the exponential\npower-versus-time curve in computing, and the corresponding pace of\nsoftware development, is that 50% of what one knows becomes obsolete\nover every 18 months.  Unix does not abolish this phenomenon, but does\ndo a good job of containing it.  There's a bedrock of unchanging\nbasics \u2014 languages, system calls, and tool invocations \u2014\nthat one can actually keep using for years, even decades.  Elsewhere\nit is impossible to predict what will be stable; even entire operating\nsystems cycle out of use.  Under Unix, there is a fairly sharp\ndistinction between transient knowledge and lasting knowledge, and one\ncan know ahead of time (with about 90% certainty) which category\nsomething is likely to fall in when one learns it.  Thus the loyalty\nUnix commands.\n<\/p>\n<p>\n  Much of Unix's stability and success has to be attributed to its\ninherent strengths, to design decisions Ken\nThompson,\nDennis Ritchie, Brian\nKernighan, Doug McIlroy, Rob Pike\n  and other early Unix developers made back at\nthe beginning; decisions that have been proven sound over and\nover. But just as much is due to the design philosophy, art of\nprogramming, and technical culture that grew up around Unix in the\nearly days. This tradition has continuously and successfully\npropagated itself in symbiosis with Unix ever since.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2876928\" id=\"ftn.id2876928\">[4]<\/a><\/sup>\n  In fact, Ethernet has already been replaced by a\ndifferent technology with the same name \u2014 twice.  Once when coax\nwas replaced with twisted pair, and a second time when gigabit\nEthernet came in.\n<\/p><\/div>\n\n<h3 id=\"ch01s03\"><a href=\"#ch01s03\">\u00a7<\/a>The Case against Learning Unix Culture<\/h3>\n<p>\n  Unix's durability and its technical culture are certainly of\ninterest to people who already like Unix, and perhaps to historians of\ntechnology. But Unix's original application as a general-purpose\ntimesharing system for mid-sized and larger computers is rapidly\nreceding into the mists of history, killed off by personal\nworkstations. And there is certainly room for doubt that it will ever\nachieve success in the mainstream business-desktop market now\ndominated by\nMicrosoft.\n<\/p>\n<p>\n  Outsiders have frequently dismissed Unix as an academic toy or a\nhacker's sandbox.  One well-known polemic, the\n  <em>Unix Hater's Handbook<\/em><sup><a href=\"#Garfinkel\" title=\"[Garfinkel]\">[Garfinkel]<\/a><\/sup>, follows an\nantagonistic line nearly as old as Unix itself in writing its devotees\noff as a cult religion of freaks and losers. Certainly the colossal\nand repeated blunders of\nAT&amp;T,\nSun, Novell,\nand other commercial vendors and standards consortia in mispositioning\nand mismarketing Unix have become legendary.\n<\/p>\n<p>\n  Even from within the Unix world, Unix has seemed to be teetering\non the brink of universality for so long as to raise the suspicion\nthat it will never actually get there. A skeptical outside observer's\nconclusion might be that Unix is too useful to die but too awkward to\nbreak out of the back room; a perpetual niche operating system.\n<\/p>\n<p>\n  What confounds the skeptics' case is, more than anything else,\nthe rise of Linux\n  and other open-source Unixes (such as the\nmodern BSD variants).  Unix's culture proved too vital to be smothered\neven by a decade of vendor mismanagement.  Today the Unix community\nitself has taken control of the technology and marketing, and is\nrapidly and visibly solving Unix's problems (in ways we'll examine in\nmore detail in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">\n   Chapter 20<\/a>).\n<\/p>\n\n<h3 id=\"ch01s04\"><a href=\"#ch01s04\">\u00a7<\/a>What Unix Gets Wrong<\/h3>\n<p>\n  For a design that dates from 1969, it is remarkably difficult\nto identify design choices in Unix that are unequivocally wrong.\nThere are several popular candidates, but each is still a subject of\nspirited debate not merely among Unix fans but across the wider\ncommunity of people who think about and design operating\nsystems.\n<\/p>\n<p>\n  Unix files have no structure above byte level. File deletion is\nirrevocable.  The Unix security model is arguably too primitive.  Job\ncontrol is botched.  There are too many different kinds of names for\nthings.  Having a file system at all may have been the wrong choice.\nWe will discuss these technical issues in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">Chapter 20<\/a>.\n<\/p>\n<p>\n  But perhaps the most enduring objections to Unix are consequences of\na feature of its philosophy first made explicit by the designers of\nthe X windowing system. X strives to provide \u201cmechanism,\nnot policy\u201d, supporting an extremely general set of graphics\noperations and deferring decisions about toolkits and interface\nlook-and-feel (the policy) up to application level. Unix's other\nsystem-level services display similar tendencies; final choices about\nbehavior are pushed as far toward the user as possible. Unix users\ncan choose among multiple shells. Unix programs normally provide many\nbehavior options and sport elaborate preference facilities.\n<\/p>\n<p>\n  This tendency reflects Unix's heritage as an operating system\ndesigned primarily for technical users, and a consequent belief that\nusers know better than operating-system designers what their own needs\nare.\n<\/p>\n <blockquote>\n  <p>\n   This tenet was firmly established at Bell Labs by Dick\nHamming<sup><a href=\"#ftn.id2808745\" id=\"id2808745\">[5]<\/a><\/sup>\n   who insisted in\nthe 1950s when computers were rare and expensive, that open-shop\ncomputing, where customers wrote their own programs, was imperative,\nbecause \u201cit is better to solve the right problem the wrong way\nthan the wrong problem the right way\u201d.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  But the cost of the mechanism-not-policy approach is that when\nthe user <em>can<\/em> set policy, the user <em>must<\/em>\n  set policy. Nontechnical end-users\nfrequently find Unix's profusion of options and interface styles\noverwhelming and retreat to systems that at least pretend to offer\nthem simplicity.\n<\/p>\n<p>\n  In the short term, Unix's laissez-faire approach may lose it a\ngood many nontechnical users. In the long term, however, it may turn\nout that this \u2018mistake\u2019 confers a critical advantage\n\u2014 because policy tends to have a short lifetime, mechanism a\nlong one. Today's fashion in interface look-and-feel too often becomes\ntomorrow's evolutionary dead end (as people using obsolete X toolkits\nwill tell you with some feeling!). So the flip side of the flip side\nis that the \u201cmechanism, not policy\u201d philosophy may enable\nUnix to renew its relevance long after competitors more tied to one\nset of policy or interface choices have faded from\nview.<sup><a href=\"#ftn.id2808799\" id=\"id2808799\">[6]<\/a><\/sup><\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2808745\" id=\"ftn.id2808745\">[5]<\/a><\/sup>\n  Yes, the Hamming of \u2018Hamming distance\u2019\nand \u2018Hamming code\u2019.\n<\/p><p><sup><a href=\"#id2808799\" id=\"ftn.id2808799\">[6]<\/a><\/sup>\n  Jim Gettys, one of the architects of X (and a\ncontributor to this book), has meditated in depth on how X's\nlaissez-faire style might be productively carried forward in\n  <em>The Two-Edged Sword<\/em> <sup><a href=\"#Gettys\" title=\"[Gettys]\">[Gettys]<\/a><\/sup>.  This essay is well worth reading,\nboth for its specific proposals and for its expression of the Unix\nmindset.\n<\/p><\/div>\n\n<h3 id=\"ch01s05\"><a href=\"#ch01s05\">\u00a7<\/a>What Unix Gets Right<\/h3>\n<p>\n  The explosive recent growth of Linux, and the increasing importance\nof the Internet, give us good reasons to suppose that the skeptics'\ncase is wrong. But even supposing the skeptical assessment is true,\nUnix culture is worth learning because there are some things that Unix\nand its surrounding culture clearly do better than any\ncompetitors.\n<\/p>\n<h4 id=\"id2808846\"><a href=\"#id2808846\">\u00a7<\/a>Open-Source Software<\/h4>\n<p>\n  Though the term \u201copen\nsource\u201d\n  and the Open Source Definition were not invented until 1998,\npeer-review-intensive development of freely shared source code was a key\nfeature of the Unix culture from its beginnings.\n<\/p>\n<p>\n  For its first ten years\nAT&amp;T's\n  original\nUnix, and its primary variant Berkeley Unix, were normally distributed\nwith source code.  This enabled most of the other good things that\nfollow here.\n<\/p>\n<h4 id=\"id2872776\"><a href=\"#id2872776\">\u00a7<\/a>Cross-Platform Portability and Open Standards<\/h4>\n<p>\n  Unix is still the only operating system that can present a\nconsistent, documented application programming interface (API) across\na heterogeneous mix of computers, vendors, and special-purpose\nhardware. It is the only operating system that can scale from embedded\nchips and handhelds, up through desktop machines, through servers, and\nall the way to special-purpose number-crunching behemoths and database\nback ends.\n<\/p>\n<p>\n  The Unix API is the closest thing to a hardware-independent\nstandard for writing truly portable software that exists. It is no\naccident that what the IEEE originally called the\n  <em>Portable Operating System Standard<\/em>\n  quickly got a suffix added to\nits acronym and became\nPOSIX.  A\nUnix-equivalent API was the only credible model for such a\nstandard.\n<\/p>\n<p>\n  Binary-only applications for other operating systems die with\ntheir birth environments, but Unix sources are forever. Forever, at\nleast, given a Unix technical culture that polishes and maintains\nthem across decades.\n<\/p>\n<h4 id=\"id2872827\"><a href=\"#id2872827\">\u00a7<\/a>The Internet and the World Wide Web<\/h4>\n<p>\n  The Defense Department's contract for the first production\nTCP\/IP\n  stack went to a\nUnix development group because the Unix in question was largely open\nsource.  Besides\nTCP\/IP, Unix has\nbecome the one indispensable core technology of the Internet Service\nProvider industry. Ever since the demise of the TOPS family of operating\nsystems\n  in the\nmid-1980s, most Internet server machines (and effectively all above\nthe PC level) have relied on Unix.\n<\/p>\n<p>\n  Not even\nMicrosoft's\n  awesome\nmarketing clout has been able to dent Unix's lock on the\nInternet. While the\nTCP\/IP\n  standards (on\nwhich the Internet is based) evolved under\nTOPS-10\n  and\nare theoretically separable from Unix, attempts to make them work on\nother operating systems have been bedeviled by incompatibilities,\ninstabilities, and bugs. The theory and specifications are available\nto anyone, but the engineering tradition to make them into a solid and\nworking reality exists only in the Unix world.<sup><a href=\"#ftn.id2872907\" id=\"id2872907\">[7]<\/a><\/sup><\/p>\n<p>\n  The Internet technical culture\n  and the Unix culture began to\nmerge in the early 1980s, and are now inseparably symbiotic.  The\ndesign of the World Wide Web, the modern face of the Internet, owes as\nmuch to Unix as it does to the ancestral ARPANET.  In particular, the\nconcept of the Uniform Resource Locator (URL) so central to the Web is\na generalization of the Unix idea of one uniform file namespace\neverywhere. To function effectively as an Internet expert, an\nunderstanding of Unix and its culture are indispensable.\n<\/p>\n<h4 id=\"id2872945\"><a href=\"#id2872945\">\u00a7<\/a>The Open-Source Community<\/h4>\n<p>\n  The community that originally formed around the early Unix\nsource distributions never went away \u2014 after the great Internet\nexplosion of the early 1990s, it recruited an entire new generation of\neager hackers\n  on home\nmachines.\n<\/p>\n<p>\n  Today, that community is a powerful support group for all kinds\nof software development.  High-quality open-source development tools\nabound in the Unix world (we'll examine many in this\nbook). Open-source Unix applications are usually equal to, and are\noften superior to, their proprietary equivalents <sup><a href=\"#Fuzz\" title=\"[Fuzz]\">[Fuzz]<\/a><\/sup>. Entire Unix operating systems, with complete\ntoolkits and basic applications suites, are available for free over\nthe Internet. Why code from scratch when you can adapt, reuse,\nrecycle, and save yourself 90% of the work?\n<\/p>\n<p>\n  This tradition of code-sharing depends heavily on hard-won\nexpertise about how to make programs cooperative and reusable. And\nnot by abstract theory, but through a lot of engineering practice\n\u2014 unobvious design rules that allow programs to function not just\nas isolated one-shot solutions but as synergistic parts of a\ntoolkit.  A major purpose of this book is to elucidate those rules.\n<\/p>\n<p>\n  Today, a burgeoning open-source movement is bringing new\nvitality, new technical approaches, and an entire generation of bright\nyoung programmers into the Unix tradition. Open-source projects\nincluding the Linux\n  operating system and symbionts such as\nApache\n  and Mozilla\nhave brought the Unix tradition an unprecedented level of mainstream\nvisibility and success. The open-source movement seems on the verge of\nwinning its bid to define the computing infrastructure of tomorrow\n\u2014 and the core of that infrastructure will be Unix machines\nrunning on the Internet.\n<\/p>\n<h4 id=\"id2873031\"><a href=\"#id2873031\">\u00a7<\/a>Flexibility All the Way Down<\/h4>\n<p>\n  Many operating systems touted as more \u2018modern\u2019 or\n\u2018user friendly\u2019 than Unix achieve their surface glossiness by\nlocking users and developers into one interface policy, and offer\nan application-programming interface that for all its elaborateness\nis rather narrow and rigid. On such systems, tasks the designers\nhave anticipated are very easy \u2014 but tasks they have not\nanticipated are often impossible or at best extremely painful.\n<\/p>\n<p>\n  Unix, on the other hand, has flexibility in depth. The many ways\nUnix provides to glue together programs mean that components of its\nbasic toolkit can be combined to produce useful effects that the\ndesigners of the individual toolkit parts never anticipated.\n<\/p>\n<p>\n  Unix's support of multiple styles of program interface (often\nseen as a weakness because it increases the perceived complexity of\nthe system to end users) also contributes to flexibility; no\nprogram that wants to be a simple piece of data plumbing is forced\nto carry the complexity overhead of an elaborate GUI.\n<\/p>\n<p>\n  Unix tradition lays heavy emphasis on keeping programming\ninterfaces relatively small, clean, and orthogonal \u2014 another trait\nthat produces flexibility in depth. Throughout a Unix system, easy\nthings are easy and hard things are at least possible.\n<\/p>\n<h4 id=\"id2873078\"><a href=\"#id2873078\">\u00a7<\/a>Unix Is Fun to Hack<\/h4>\n<p>\n  People who pontificate about Unix's technical superiority often\ndon't mention what may ultimately be its most important strength,\nthe one that underlies all its successes. Unix is fun to hack.\n<\/p>\n<p>\n  Unix boosters seem almost ashamed to acknowledge this sometimes,\nas though admitting they're having fun might damage their\nlegitimacy somehow. But it's true; Unix is fun to play with and\ndevelop for, and always has been.\n<\/p>\n<p>\n  There are not many operating systems that anyone has ever\ndescribed as \u2018fun\u2019. Indeed, the friction and labor of\ndevelopment under most other environments has been aptly compared to\nkicking a dead whale down the beach.<sup><a href=\"#ftn.id2873107\" id=\"id2873107\">[8]<\/a><\/sup>\n  The kindest\nadjectives one normally hears are on the order of\n\u201ctolerable\u201d or \u201cnot too painful\u201d. In the Unix\nworld, by contrast, the operating system rewards effort rather than\nfrustrating it.  People programming under Unix usually come to see\nit not as an adversary to be clubbed into doing one's bidding by main\neffort but rather as an actual positive help.\n<\/p>\n<p>\n  This has real economic significance. The fun factor started a\nvirtuous circle early in Unix's history. People liked Unix, so they\nbuilt more programs for it that made it nicer to use. Today people\nbuild entire, production-quality open-source Unix systems as a\nhobby. To understand how remarkable this is, ask yourself when you\nlast heard of anybody cloning OS\/360 or VAX VMS\n  or Microsoft\nWindows\n  for fun.\n<\/p>\n<p>\n  The \u2018fun\u2019 factor is not trivial from a design point\nof view, either. The kind of people who become programmers and\ndevelopers have \u2018fun\u2019 when the effort they have to put out\nto do a task challenges them, but is just within their\ncapabilities. \u2018Fun\u2019 is therefore a sign of peak\nefficiency. Painful development environments waste labor and\ncreativity; they extract huge hidden costs in time, money, and\nopportunity.\n<\/p>\n<p>\n  If Unix were a failure in every other way, the Unix engineering\nculture would be worth studying for the ways it keeps the fun\nin development \u2014 because that fun is a sign that it makes\ndevelopers efficient, effective, and productive.\n<\/p>\n<h4 id=\"id2873180\"><a href=\"#id2873180\">\u00a7<\/a>The Lessons of Unix Can Be Applied Elsewhere<\/h4>\n<p>\n  Unix programmers have accumulated decades of experience\nwhile pioneering operating-system features we now take for granted.\nEven non-Unix programmers can benefit from studying that Unix\nexperience. Because Unix makes it relatively easy to apply good design\nprinciples and development methods, it is an excellent place to\nlearn them.\n<\/p>\n<p>\n  Other operating systems generally make good practice rather\nmore difficult, but even so some of the Unix culture's lessons can\ntransfer. Much Unix code (including all its filters, its major\nscripting languages, and many of its code generators) will\nport directly to any operating system supporting ANSI\nC\n  (for the\nexcellent reason that C\n  itself was a Unix invention and the\nANSI C library embodies a substantial chunk of Unix's\nservices!).\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2872907\" id=\"ftn.id2872907\">[7]<\/a><\/sup>\n  Other\noperating systems have generally copied or cloned Unix TCP\/IP\nimplementations. It is their loss that they have not generally adopted\nthe robust tradition of peer review that goes with it, exemplified by\ndocuments like RFC 1025 (\n  <em>TCP and IP Bake Off<\/em>).\n<\/p><p><sup><a href=\"#id2873107\" id=\"ftn.id2873107\">[8]<\/a><\/sup>\n  This was originally\nsaid of the IBM MVS TSO facility by Stephen C. Johnson, perhaps better\nknown as the author of\n  <em>yacc<\/em>.\n<\/p><\/div>\n\n<h3 id=\"ch01s06\"><a href=\"#ch01s06\">\u00a7<\/a>Basics of the Unix Philosophy<\/h3>\n<p>\n  The \u2018Unix philosophy\u2019 originated with Ken\nThompson's\n  early meditations on how to design a small but capable operating\nsystem with a clean service interface. It grew as the Unix culture\nlearned things about how to get maximum leverage out of Thompson's\ndesign.  It absorbed lessons from many sources along the way.\n<\/p>\n<p>\n  The Unix philosophy is not a formal design method. It wasn't\nhanded down from the high fastnesses of theoretical computer\nscience as a way to produce theoretically perfect software. Nor is\nit that perennial executive's mirage, some way to magically extract\ninnovative but reliable software on too short a deadline from\nunmotivated, badly managed, and underpaid programmers.\n<\/p>\n<p>\n  The Unix philosophy (like successful folk traditions in other\nengineering disciplines) is bottom-up, not top-down. It is pragmatic\nand grounded in experience. It is not to be found in official methods\nand standards, but rather in the implicit half-reflexive knowledge, the\n  <em>expertise<\/em>\n  that the Unix culture transmits. It\nencourages a sense of proportion and skepticism \u2014 and shows both\nby having a sense of (often subversive) humor.\n<\/p>\n<p>\n  Doug McIlroy,\n  the inventor of\nUnix pipes\n  and one of the\nfounders of the Unix tradition, had this to say at the time <sup><a href=\"#McIlroy78\" title=\"[McIlroy78]\">[McIlroy78]<\/a><\/sup>:\n<\/p>\n <blockquote>\n  <p>\n   (i) Make each program do one thing well. To do a new job, build\nafresh rather than complicate old programs by adding new\nfeatures.\n  <\/p>\n  <p>\n   (ii) Expect the output of every program to become the input to\nanother, as yet unknown, program. Don't clutter output with\nextraneous information. Avoid stringently columnar or binary input\nformats. Don't insist on interactive input.\n  <\/p>\n  <p>\n   (iii) Design and build software, even operating systems, to be\ntried early, ideally within weeks. Don't hesitate to throw away the\nclumsy parts and rebuild them.\n  <\/p>\n  <p>\n   (iv) Use tools in preference to unskilled help to lighten a\nprogramming task, even if you have to detour to build the tools and\nexpect to throw some of them out after you've finished using\nthem.\n  <\/p>\n <\/blockquote>\n<p>\n  He later summarized it this way (quoted in\n  <em>A Quarter Century of Unix<\/em><sup><a href=\"#Salus\" title=\"[Salus]\">[Salus]<\/a><\/sup>):\n<\/p>\n <blockquote>\n  <p>\n   This is the Unix philosophy: Write programs that do one\nthing and do it well. Write programs to work together. Write\nprograms to handle text streams, because that is a universal\ninterface.\n  <\/p>\n <\/blockquote>\n<p>\n  Rob Pike, who\nbecame one of the great masters of C, offers a slightly different angle\nin\n  <em>Notes on C Programming<\/em> <sup><a href=\"#Pike\" title=\"[Pike]\">[Pike]<\/a><\/sup>:\n<\/p>\n <blockquote>\n  <p>\n   Rule 1. You can't tell where a program is going to spend its\ntime. Bottlenecks occur in surprising places, so don't try to second\nguess and put in a speed hack until you've proven that's where the\nbottleneck\nis.\n  <\/p>\n  <p>\n   Rule 2. Measure. Don't tune for speed until you've measured, and\neven then don't unless one part of the code overwhelms the\nrest.\n  <\/p>\n  <p>\n   Rule 3. Fancy algorithms are slow when\n   <code>n<\/code> is small, and\n   <code>n<\/code> is usually small. Fancy algorithms have big\nconstants. Until you know that\n   <code>n<\/code> is\nfrequently going to be big, don't get fancy. (Even if\n   <code>n<\/code> does get big, use Rule 2 first.)\n  <\/p>\n  <p>\n   Rule 4. Fancy algorithms are buggier than simple ones, and\nthey're much harder to implement. Use simple algorithms as well as\nsimple data structures.\n  <\/p>\n  <p id=\"rule5\">\n   Rule 5. Data dominates. If you've chosen the right data\nstructures and organized things well, the algorithms will almost\nalways be self-evident. Data structures, not algorithms, are central\nto programming.<sup><a href=\"#ftn.id2873492\" id=\"id2873492\">[9]<\/a><\/sup><\/p>\n  <p>\n   Rule 6. There is no Rule 6.\n  <\/p>\n <\/blockquote>\n<p>\n  Ken Thompson, the man who designed and implemented the\nfirst Unix, reinforced Pike's rule 4 with a gnomic maxim worthy of a\nZen\n  patriarch:\n<\/p>\n <blockquote>\n  <p>\n   When in doubt, use brute force.\n  <\/p>\n <\/blockquote>\n<p>\n  More of the Unix philosophy was implied not by what these elders\nsaid but by what they did and the example Unix itself set. Looking\nat the whole, we can abstract the following ideas:\n<\/p>\n <div class=\"orderedlist\"><ol id=\"rules\"><li>Rule of Modularity: Write simple parts connected by clean interfaces.<\/li>\n   <li>\n     Rule of Clarity: Clarity is better than cleverness.\n   <\/li>\n   <li>Rule of Composition: Design programs to be connected to other programs.<\/li>\n   <li>Rule of Separation: Separate policy from mechanism; separate interfaces from engines.<\/li>\n   <li>Rule of Simplicity: Design for simplicity; add complexity only where you must.<\/li>\n   <li>Rule of Parsimony: Write a big program only when it is clear by demonstration that nothing else will do.<\/li>\n   <li> Rule of Transparency: Design for visibility to make inspection and debugging easier.<\/li>\n   <li>Rule of Robustness: Robustness is the child of transparency and simplicity.<\/li>\n   <li>Rule of Representation: Fold knowledge into data so program logic can be stupid and robust.<\/li>\n   <li>Rule of Least Surprise: In interface design, always do the least surprising thing.<\/li>\n   <li>Rule of Silence: When a program has nothing surprising to say, it should say nothing.<\/li>\n   <li>Rule of Repair: When you must fail, fail noisily and as soon as possible.<\/li>\n   <li>Rule of Economy: Programmer time is expensive; conserve it in preference to machine time.<\/li>\n   <li>Rule of Generation: Avoid hand-hacking; write programs to write programs when you can.<\/li>\n   <li>Rule of Optimization: Prototype before polishing. Get it working before you optimize it.<\/li>\n   <li>Rule of Diversity: Distrust all claims for \u201cone true way\u201d.<\/li>\n   <li>Rule of Extensibility: Design for the future, because it will be here sooner than you think.<\/li>\n  <\/ol><\/div>\n<p>\n  If you're new to Unix, these principles are worth some\nmeditation. Software-engineering texts recommend most of them; but\nmost other operating systems lack the right tools and traditions to\nturn them into practice, so most programmers can't apply them with\nany consistency. They come to accept blunt tools, bad designs,\noverwork, and bloated code as normal \u2014 and then wonder what Unix\nfans are so annoyed about.\n<\/p>\n<h4 id=\"id2877537\"><a href=\"#id2877537\">\u00a7<\/a>Rule of Modularity: Write simple parts connected by clean interfaces.<\/h4>\n<p>\n  As Brian Kernighan once observed, \u201cControlling complexity\nis the essence of computer programming\u201d <sup><a href=\"#Kernighan-Plauger\" title=\"[Kernighan-Plauger]\">[Kernighan-Plauger]<\/a><\/sup>. Debugging dominates\ndevelopment time, and getting a working system out the door is\nusually less a result of brilliant design than it is of managing not\nto trip over your own feet too many times.\n<\/p>\n<p>\n  Assemblers, compilers, flowcharting, procedural programming,\nstructured programming, \u201cartificial intelligence\u201d,\nfourth-generation languages, object\norientation, and software-development\nmethodologies without number have been touted and sold as a cure for\nthis problem. All have failed as cures, if only because they\n\u2018succeeded\u2019 by escalating the normal level of program\ncomplexity to the point where (once again) human brains could barely\ncope. As Fred Brooks\n  famously observed <sup><a href=\"#Brooks\" title=\"[Brooks]\">[Brooks]<\/a><\/sup>, there is no silver bullet.\n<\/p>\n<p>\n  The only way to write complex software that won't fall on its\nface is to hold its global complexity down \u2014 to build it out of\nsimple parts connected by well-defined interfaces, so that most\nproblems are local and you can have some hope of upgrading\na part without breaking the whole.\n<\/p>\n<h4 id=\"id2877610\"><a href=\"#id2877610\">\u00a7<\/a>Rule of Clarity: Clarity is better than cleverness.<\/h4>\n<p>\n  Because maintenance is so important and so expensive, write\nprograms as if the most important communication they do is not to the\ncomputer that executes them but to the human beings who will read and\nmaintain the source code in the future (including yourself).\n<\/p>\n<p>\n  In the Unix tradition, the implications of this advice go beyond\njust commenting your code.  Good Unix practice also embraces choosing\nyour algorithms and implementations for future maintainability.  Buying a\nsmall increase in performance with a large increase in the complexity \nand obscurity of your technique is a bad trade \u2014 not merely because\ncomplex code is more likely to harbor bugs, but also because complex\ncode will be harder to read for future maintainers.\n<\/p>\n<p>\n  Code that is graceful and clear, on the other hand, is less\nlikely to break \u2014 and more likely to be instantly comprehended\nby the next person to have to change it.  This is important,\nespecially when that next person might be yourself some years down the\nroad.\n<\/p>\n <blockquote>\n  <p>\n   Never struggle to decipher subtle code three times.  Once might\nbe a one-shot fluke, but if you find yourself having to figure it out\na second time \u2014 because the first was too long ago and you've\nforgotten details \u2014 it is time to comment the code so that the\nthird time will be relatively painless.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<h4 id=\"id2877684\"><a href=\"#id2877684\">\u00a7<\/a>Rule of Composition: Design programs to be connected with other programs.<\/h4>\n<p>\n  It's hard to avoid programming overcomplicated monoliths if none\nof your programs can talk to each other.\n<\/p>\n<p>\n  Unix tradition strongly encourages writing programs that\nread and write simple, textual, stream-oriented, device-independent\nformats.  Under classic Unix, as many programs as possible are written\nas simple\n  <em>filters<\/em>, which take a simple text\nstream on input and process it into another simple text stream on\noutput.\n<\/p>\n<p>\n  Despite popular mythology, this practice is favored not because\nUnix programmers hate graphical user interfaces. It's because if you\ndon't write programs that accept and emit simple text streams, it's\nmuch more difficult to hook the programs together.\n<\/p>\n<p>\n  Text streams are to Unix tools as messages are to objects in an\nobject-oriented setting. The simplicity of the text-stream interface\nenforces the encapsulation of the tools.  More elaborate forms of\ninter-process communication, such as remote procedure calls, show a\ntendency to involve programs with each others' internals too\nmuch.\n<\/p>\n<p>\n  To make programs composable, make them independent.  A program\non one end of a text stream should care as little as possible about\nthe program on the other end.  It should be made easy to replace one\nend with a completely different implementation without disturbing the\nother.\n<\/p>\n<p>\n  GUIs can be a very good thing. Complex binary data formats are\nsometimes unavoidable by any reasonable means. But before writing a\nGUI, it's wise to ask if the tricky interactive parts of your\nprogram can be segregated into one piece and the workhorse\nalgorithms into another, with a simple command stream or\napplication protocol connecting the two. Before devising a tricky\nbinary format to pass data around, it's worth experimenting to see\nif you can make a simple textual format work and accept a little\nparsing overhead in return for being able to hack the data stream\nwith general-purpose tools.\n<\/p>\n<p>\n  When a serialized, protocol-like interface is not natural for the\napplication, proper Unix design is to at least organize as many of the\napplication primitives as possible into a library with a well-defined\nAPI.  This opens up the possibility that the application can be called\nby linkage, or that multiple interfaces can be glued on it for different tasks.\n<\/p>\n<p>\n  (We discuss these issues in detail in <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>.)\n<\/p>\n<h4 id=\"id2877777\"><a href=\"#id2877777\">\u00a7<\/a>Rule of Separation: Separate policy from mechanism; separate interfaces from engines.<\/h4>\n<p>\n  In our discussion of what Unix gets wrong, we observed that the\ndesigners of X\n  made a basic decision to implement\n\u201cmechanism, not policy\u201d\n  \u2014to make X a generic\ngraphics engine and leave decisions about user-interface style to\ntoolkits and other levels of the system.  We justified this by\npointing out that policy and mechanism tend to mutate on different\ntimescales, with policy changing much faster than mechanism. Fashions\nin the look and feel of GUI toolkits may come and go, but raster\noperations and compositing are forever.\n<\/p>\n<p>\n  Thus, hardwiring policy and mechanism together has two bad\neffects: It makes policy rigid and harder to change in response to user\nrequirements, and it means that trying to change policy has a strong\ntendency to destabilize the mechanisms.\n<\/p>\n<p>\n  On the other hand, by separating the two we make it possible to\nexperiment with new policy without breaking mechanisms.  We also make\nit much easier to write good tests for the mechanism (policy,\nbecause it ages so quickly, often does not justify the investment).\n<\/p>\n<p>\n  This design rule has wide application outside the GUI\ncontext.  In general, it implies that we should look for ways to\nseparate interfaces from engines.\n<\/p>\n<p>\n  One way to effect that separation is, for example, to write\nyour application as a library of C\n  service routines that are driven by an\nembedded scripting language, with the application flow of control\nwritten in the scripting language rather than C. A classic example of\nthis pattern is the\n  <em>Emacs<\/em> editor, which uses\nan embedded Lisp\n  interpreter to control editing\nprimitives written in C.  We discuss this style of design in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">Chapter 11<\/a>.\n<\/p>\n<p>\n  Another way is to separate your application into cooperating front-end and back-end processes communicating through a\n  specialized application protocol over sockets; we discuss this kind of design in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">Chapter 5<\/a>\n  and\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>.  The front end implements policy; the\nback end, mechanism.  The global complexity of the pair will often be\nfar lower than that of a single-process monolith implementing the same\nfunctions, reducing your vulnerability to bugs and lowering life-cycle\ncosts.\n<\/p>\n<h4 id=\"id2877917\"><a href=\"#id2877917\">\u00a7<\/a>Rule of Simplicity: Design for simplicity; add complexity only where you must.<\/h4>\n<p>\n  Many pressures tend to make programs more complicated (and\ntherefore more expensive and buggy). One such pressure is technical\nmachismo. Programmers are bright people who are (often justly) proud\nof their ability to handle complexity and juggle abstractions. Often\nthey compete with their peers to see who can build the most intricate\nand beautiful complexities. Just as often, their ability to design\noutstrips their ability to implement and debug, and the result is\nexpensive failure.\n<\/p>\n <blockquote>\n  <p>\n   The notion of \u201cintricate and beautiful complexities\u201d is\nalmost an oxymoron.  Unix programmers vie with each other\nfor \u201csimple and beautiful\u201d honors \u2014 a point that's\nimplicit in these rules, but is well worth making overt.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Even more often (at least in the commercial software world)\nexcessive complexity comes from project requirements that are based on\nthe marketing fad of the month rather than the reality of what\ncustomers want or software can actually deliver. Many a good design\nhas been smothered under marketing's pile of \u201cchecklist\nfeatures\u201d \u2014 features that, often, no customer will ever\nuse. And a vicious circle operates; the competition thinks it has to\ncompete with chrome by adding more chrome. Pretty soon, massive bloat\nis the industry standard and everyone is using huge, buggy programs\nnot even their developers can love.\n<\/p>\n<p>\n  Either way, everybody loses in the end.\n<\/p>\n<p>\n  The only way to avoid these traps is to encourage a software\nculture that knows that small is beautiful, that actively resists\nbloat and complexity: an engineering tradition that puts a high\nvalue on simple solutions, that looks for ways to break program systems up\ninto small cooperating pieces, and that reflexively fights attempts to\ngussy up programs with a lot of chrome (or, even worse, to design\nprograms\n  <em>around<\/em> the chrome).\n<\/p>\n<p>\n  That would be a culture a lot like Unix's.\n<\/p>\n<h4 id=\"id2878022\"><a href=\"#id2878022\">\u00a7<\/a>Rule of Parsimony: Write a big program only when it is clear by demonstration that nothing else will do.<\/h4>\n<p>\n  \u2018Big\u2019 here has the sense both of large in volume of\ncode and of internal complexity.  Allowing programs to get large\nhurts maintainability.  Because people are reluctant to throw\naway the visible product of lots of work, large programs invite\noverinvestment in approaches that are failed or suboptimal.\n<\/p>\n<p>\n  (We'll examine the issue of the right size of software in more\ndetail in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">Chapter 13<\/a>.)\n<\/p>\n<h4 id=\"id2878054\"><a href=\"#id2878054\">\u00a7<\/a>Rule of Transparency: Design for visibility to make inspection and debugging easier.<\/h4>\n<p>\n  Because debugging often occupies three-quarters or more of\ndevelopment time, work done early to ease debugging can be a very\ngood investment. A particularly effective way to ease debugging is to\ndesign for\n  <em>transparency<\/em> and\n  <em>discoverability<\/em>.\n<\/p>\n<p>\n  A software system is\n  <em>transparent<\/em>\n  when you\ncan look at it and immediately understand what it is doing and how. It\nis\n  <em>discoverable<\/em>\n  when it has facilities for\nmonitoring and display of internal state so that your program not only\nfunctions well but can be\n  <em>seen<\/em>\n  to function\nwell.\n<\/p>\n<p>\n  Designing for these qualities will have implications throughout\na project. At minimum, it implies that debugging options should not be\nminimal afterthoughts. Rather, they should be designed in from the\nbeginning \u2014 from the point of view that the program should be\nable to both demonstrate its own correctness and communicate to future\ndevelopers the original developer's mental model of the problem it\nsolves.\n<\/p>\n<p>\n  For a program to demonstrate its own correctness, it needs to be\nusing input and output formats sufficiently simple so that the proper\nrelationship between valid input and correct output is easy to\ncheck.\n<\/p>\n<p>\n  The objective of designing for transparency and discoverability\nshould also encourage simple interfaces that can easily be manipulated\nby other programs \u2014 in particular, test and monitoring harnesses\nand debugging scripts.\n<\/p>\n<h4 id=\"id2878145\"><a href=\"#id2878145\">\u00a7<\/a>Rule of Robustness: Robustness is the child of transparency and simplicity.<\/h4>\n<p>\n  Software is said to be\n  <em>robust<\/em> when\nit performs well under unexpected conditions which stress the designer's\nassumptions, as well as under normal conditions.\n<\/p>\n<p>\n  Most software is fragile and buggy because most programs are too\ncomplicated for a human brain to understand all at once. When you\ncan't reason correctly about the guts of a program, you can't be sure\nit's correct, and you can't fix it if it's broken.\n<\/p>\n<p>\n  It follows that the way to make robust programs is to make their\ninternals easy for human beings to reason about. There are two main\nways to do that:\ntransparency\n  and\nsimplicity.\n<\/p>\n <blockquote>\n  <p>\n   For robustness, designing in tolerance for unusual or extremely\nbulky inputs is also important.  Bearing in mind the Rule of\nComposition helps; input generated by other programs is notorious for\nstress-testing software (e.g., the original Unix C compiler reportedly\nneeded small upgrades to cope well with Yacc output).  The forms\ninvolved often seem useless to humans.  For example, accepting empty\nlists\/strings\/etc., even in places where a human would seldom or never\nsupply an empty string, avoids having to special-case such situations when\ngenerating the input mechanically.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  One very important tactic for being robust under odd inputs is\nto avoid having special cases in your code.  Bugs often lurk in the\ncode for handling special cases, and in the interactions among parts\nof the code intended to handle different special cases.\n<\/p>\n<p>\n  We observed above that software is\n  <em>transparent<\/em> when you can look at it and\nimmediately see what is going on. It is\n  <em>simple<\/em>\n  when what is going on is uncomplicated enough for a human brain to\nreason about all the potential cases without strain.  The more\nyour programs have both of these qualities, the more robust they will\nbe.\n<\/p>\n<p>\n  Modularity (simple parts, clean interfaces) is a way to organize\nprograms to make them simpler. There are other ways to fight for\nsimplicity. Here's another one.\n<\/p>\n<h4 id=\"id2878263\"><a href=\"#id2878263\">\u00a7<\/a>Rule of Representation: Fold knowledge into data, so program logic can be stupid and robust.<\/h4>\n<p>\n  Even the simplest procedural logic is hard for humans to verify,\nbut quite complex data structures are fairly easy to model and reason\nabout. To see this, compare the expressiveness and explanatory power\nof a diagram of (say) a fifty-node pointer tree with a flowchart of a\nfifty-line program.  Or, compare an array initializer expressing a conversion\ntable with an equivalent switch statement. The difference in\ntransparency\n  and\nclarity is dramatic. See Rob Pike's\n  <a href=\"#rule5\">Rule 5<\/a>.\n<\/p>\n<p>\n  Data is more tractable than program logic. It follows that where\nyou see a choice between complexity in data structures and\ncomplexity in code, choose the former. More: in evolving a design,\nyou should actively seek ways to shift complexity from code to\ndata.\n<\/p>\n<p>\n  The Unix community did not originate this insight, but a lot of\nUnix code displays its influence.  The\nC\n  language's facility at\nmanipulating pointers, in particular, has encouraged the use of\ndynamically-modified reference structures at all levels of coding from\nthe kernel upward. Simple pointer chases in such structures frequently\ndo duties that implementations in other languages would instead have\nto embody in more elaborate procedures.\n<\/p>\n<p>\n  (We also cover these techniques in\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">Chapter 9<\/a>.)\n<\/p>\n<h4 id=\"id2878339\"><a href=\"#id2878339\">\u00a7<\/a>Rule of Least Surprise: In interface design, always do the least surprising thing.<\/h4>\n<p>\n  (This is also widely known as the Principle of Least\nAstonishment.)\n<\/p>\n<p>\n  The easiest programs to use are those that demand the least new\nlearning from the user \u2014 or, to put it another way, the easiest\nprograms to use are those that most effectively connect to the user's\npre-existing knowledge.\n<\/p>\n<p>\n  Therefore, avoid gratuitous novelty and excessive cleverness in\ninterface design.  If you're writing a calculator program,\n\u2018+\u2019 should always mean addition! When designing an\ninterface, model it on the interfaces of functionally similar or\nanalogous programs with which your users are likely to be\nfamiliar.\n<\/p>\n<p>\n  Pay attention to your expected audience.  They may be end\nusers, they may be other programmers, or they may be system\nadministrators.  What is least surprising can differ among these\ngroups.\n<\/p>\n<p>\n  Pay attention to tradition. The Unix world has rather well-developed\nconventions about things like the format of configuration and\nrun-control files, command-line switches, and the like. These\ntraditions exist for a good reason: to tame the learning curve.\nLearn and use them.\n<\/p>\n<p>\n  (We'll cover many of these traditions in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">Chapter 5<\/a>\n  and\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">Chapter 10<\/a>.)\n<\/p>\n <blockquote>\n  <p>\n   The flip side of the Rule of Least Surprise is to avoid\nmaking things superficially similar but really a little bit different.\nThis is extremely treacherous because the seeming familiarity raises false\nexpectations.  It's often better to make things distinctly different than\nto make them\n   <em>almost<\/em>\n   the same.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<h4 id=\"id2878450\"><a href=\"#id2878450\">\u00a7<\/a>Rule of Silence: When a program has nothing surprising to say, it should say nothing.<\/h4>\n<p>\n  One of Unix's oldest and most persistent design rules is that when a\nprogram has nothing interesting or surprising to say, it should\n  <em>shut up<\/em>.  Well-behaved Unix programs do their jobs\nunobtrusively, with a minimum of fuss and bother.  Silence is\ngolden.\n<\/p>\n<p>\n  This \u201csilence is golden\u201d rule evolved originally\nbecause Unix predates video displays.  On the slow printing terminals\nof 1969, each line of unnecessary output was a serious drain on the\nuser's time.  That constraint is gone, but excellent reasons for\nterseness remain.\n<\/p>\n <blockquote>\n  <p>\n   I think that the terseness of Unix programs is a central feature\nof the style.  When your program's output becomes another's input, it \nshould be easy to pick out the needed bits.  And for people it is\na human-factors necessity \u2014 important information should not\nbe mixed in with verbosity about internal program behavior.  If all\ndisplayed information is important, important information is easy to\nfind.\n  <\/p><div>\n  \n   \u2013 Ken Arnold\n  \n <\/div><\/blockquote>\n<p>\n  Well-designed programs treat the user's attention and\nconcentration as a precious and limited resource, only to be claimed\nwhen necessary.\n<\/p>\n<p>\n  (We'll discuss the Rule of Silence and the reasons for it in\nmore detail at the end of\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">Chapter 11<\/a>.)\n<\/p>\n<h4 id=\"id2878538\"><a href=\"#id2878538\">\u00a7<\/a>Rule of Repair: Repair what you can \u2014 but when you must fail, fail noisily and as soon as possible.<\/h4>\n<p>\n  Software should be\ntransparent\n  in the way that it fails, as well\nas in normal operation.  It's best when software can cope with\nunexpected conditions by adapting to them, but the worst kinds of bugs\nare those in which the repair doesn't succeed and the problem quietly\ncauses corruption that doesn't show up until much later.\n<\/p>\n<p>\n  Therefore, write your software to cope with incorrect inputs and\nits own execution errors as gracefully as possible.  But when it\ncannot, make it fail in a way that makes diagnosis of the problem as\neasy as possible.\n<\/p>\n<p>\n  Consider also Postel's Prescription:<sup><a href=\"#ftn.id2878578\" id=\"id2878578\">[10]<\/a><\/sup>\n  \u201cBe liberal in what you accept, and conservative in what you\nsend\u201d. Postel was speaking of network service programs, but the\nunderlying idea is more general.  Well-designed programs cooperate\nwith other programs by making as much sense as they can from\nill-formed inputs; they either fail noisily or pass strictly clean and\ncorrect data to the next program in the chain.\n<\/p>\n<p>\n  However, heed also this warning:\n<\/p>\n <blockquote>\n  <p>\n   The original HTML documents recommended \u201cbe generous in\nwhat you accept\u201d, and it has bedeviled us ever since because\neach browser accepts a different superset of the specifications.  It\nis the\n   <em>specifications<\/em>\n   that should be generous,\nnot their interpretation.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  McIlroy adjures us to\n  <em>design<\/em>\n  for\ngenerosity rather than compensating for inadequate standards with \npermissive implementations.  Otherwise, as he rightly points\nout, it's all too easy to end up in tag soup.\n<\/p>\n<h4 id=\"id2878666\"><a href=\"#id2878666\">\u00a7<\/a>Rule of Economy: Programmer time is expensive; conserve it in preference to machine time.<\/h4>\n<p>\n  In the early minicomputer days of Unix, this was still a fairly\nradical idea (machines were a great deal slower and more expensive\nthen). Nowadays, with every development shop and most users (apart\nfrom the few modeling nuclear explosions or doing 3D movie\nanimation) awash in cheap machine cycles, it may seem too obvious\nto need saying.\n<\/p>\n<p>\n  Somehow, though, practice doesn't seem to have quite caught up\nwith reality. If we took this maxim really seriously throughout\nsoftware development, most applications would be written in\nhigher-level languages like Perl, Tcl, Python, Java,\nLisp and even shell \u2014 languages that ease the programmer's burden by\ndoing their own memory management (see <sup><a href=\"#Ravenbrook\" title=\"[Ravenbrook]\">[Ravenbrook]<\/a><\/sup>).\n<\/p>\n<p>\n  And indeed this is happening within the Unix world, though\noutside it most applications shops still seem stuck with the\nold-school Unix strategy of coding in C (or C++). Later in this book\nwe'll discuss this strategy and its tradeoffs in detail.\n<\/p>\n<p>\n  One other obvious way to conserve programmer time is to teach\nmachines how to do more of the low-level work of programming. This\nleads to...\n<\/p>\n<h4 id=\"id2878742\"><a href=\"#id2878742\">\u00a7<\/a>Rule of Generation: Avoid hand-hacking; write programs to write programs when you can.<\/h4>\n<p>\n  Human beings are notoriously bad at sweating the details.\nAccordingly, any kind of hand-hacking of programs is a rich source of\ndelays and errors. The simpler and more abstracted your program\nspecification can be, the more likely it is that the human designer\nwill have gotten it right. Generated code (at\n  <em>every<\/em>\n  level) is almost always cheaper and more\nreliable than hand-hacked.\n<\/p>\n<p>\n  We all know this is true (it's why we have compilers and\ninterpreters, after all) but we often don't think about the\nimplications. High-level-language code that's repetitive and\nmind-numbing for humans to write is just as productive a target for a\ncode generator as machine code. It pays to use code generators when\nthey can raise the level of abstraction \u2014 that is, when the\nspecification language for the generator is simpler than the generated\ncode, and the code doesn't have to be hand-hacked afterwards.\n<\/p>\n<p>\n  In the Unix tradition, code generators are heavily used to\nautomate error-prone detail work. Parser\/lexer generators are the\nclassic examples; makefile generators and GUI interface builders are\nnewer ones.\n<\/p>\n<p>\n  (We cover these techniques in\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">Chapter 9<\/a>.)\n<\/p>\n<h4 id=\"rule_of_optimization\"><a href=\"#rule_of_optimization\">\u00a7<\/a>\n  Rule of Optimization:\nPrototype before polishing. Get it working before you optimize\nit.\n <\/h4>\n<p>\n  The most basic argument for prototyping first is Kernighan &amp;\nPlauger's;\n  \u201c90% of the functionality delivered now\nis better than 100% of it delivered never\u201d. Prototyping first\nmay help keep you from investing far too much time for marginal\ngains.\n<\/p>\n<p>\n  For slightly different reasons, Donald\nKnuth\n  (author of\n  <em>The Art Of Computer Programming<\/em>, one\nof the field's few true classics) popularized the observation that\n\u201cPremature optimization is the root of all\nevil\u201d.<sup><a href=\"#ftn.id2878872\" id=\"id2878872\">[11]<\/a><\/sup>\n  And he was right.\n<\/p>\n<p>\n  Rushing to optimize before the bottlenecks are known may be the\nonly error to have ruined more designs than feature creep. From\ntortured code to incomprehensible data layouts, the results of\nobsessing about speed or memory or disk usage at the expense of\ntransparency and simplicity are everywhere. They spawn innumerable\nbugs and cost millions of man-hours \u2014 often, just to get marginal\ngains in the use of some resource much less expensive than\ndebugging time.\n<\/p>\n<p>\n  Disturbingly often, premature local optimization actually\nhinders global optimization (and hence reduces overall performance).\nA prematurely optimized portion of a design frequently interferes with\nchanges that would have much higher payoffs across the whole design,\nso you end up with both inferior performance and excessively complex\ncode.\n<\/p>\n<p>\n  In the Unix world there is a long-established and very explicit\ntradition (exemplified by Rob Pike's\n  comments above and Ken\nThompson's\n  maxim about brute force) that says:\n  <em>\n   Prototype, then\npolish. Get it working before you optimize it\n<\/em>. Or: Make it\nwork first, then make it work fast. \u2018Extreme programming' guru\nKent Beck,\noperating in a different culture, has usefully amplified this to:\n\u201cMake it run, then make it right, then make it\nfast\u201d.\n<\/p>\n<p>\n  The thrust of all these quotes is the same: get your design\nright with an un-optimized, slow, memory-intensive implementation\nbefore you try to tune. Then, tune systematically, looking for the\nplaces where you can buy big performance wins with the smallest\npossible increases in local complexity.\n<\/p>\n <blockquote>\n  <p>\n   Prototyping is important for system design as well as\noptimization \u2014 it is much easier to judge whether a prototype\ndoes what you want than it is to read a long specification.  I\nremember one development manager at Bellcore who fought against the\n\u201crequirements\u201d culture years before anybody talked about\n\u201crapid prototyping\u201d or \u201cagile development\u201d. He\nwouldn't issue long specifications; he'd lash together some combination of\nshell scripts and awk code that did roughly what was needed, tell the\ncustomers to send him some clerks for a few days, and then have the\ncustomers come in and look at their clerks using the prototype and\ntell him whether or not they liked it.  If they did, he would say\n\u201cyou can have it industrial strength so-many-months from now at\nsuch-and-such cost\u201d.  His estimates tended to be accurate, but\nhe lost out in the culture to managers who believed that requirements \nwriters should be in control of everything.\n  <\/p><div>\n  \n   \u2013 Mike Lesk\n  \n <\/div><\/blockquote>\n<p>\n  Using prototyping to learn which features you don't have to\nimplement helps optimization for performance; you don't have to\noptimize what you don't write. The most powerful optimization tool in\nexistence may be the delete key.\n<\/p>\n <blockquote>\n  <p>One of my most productive days was throwing away 1000 lines of code.<\/p><div>\n  \u2013 Ken Thompson\n <\/div><\/blockquote>\n<p>\n  (We'll go into a bit more depth about related ideas in\n  <a href=\"#optimizationchapter\" title=\"Chapter 12. Optimization\">Chapter 12<\/a>.)\n<\/p>\n<h4 id=\"id2879078\"><a href=\"#id2879078\">\u00a7<\/a>Rule of Diversity: Distrust all claims for \u201cone true way\u201d.<\/h4>\n<p>\n  Even the best software tools tend to be limited by the\nimaginations of their designers. Nobody is smart enough to optimize\nfor everything, nor to anticipate all the uses to which their software\nmight be put. Designing rigid, closed software that won't talk\nto the rest of the world is an unhealthy form of arrogance.\n<\/p>\n<p>\n  Therefore, the Unix tradition includes a healthy\nmistrust of \u201cone true way\u201d approaches to software design or\nimplementation. It embraces multiple languages, open extensible\nsystems, and customization hooks everywhere.\n<\/p>\n<h4 id=\"id2879112\"><a href=\"#id2879112\">\u00a7<\/a>Rule of Extensibility: Design for the future, because it will be here sooner than you think.<\/h4>\n<p>\n  If it is unwise to trust other people's claims for \u201cone\ntrue way\u201d, it's even more foolish to believe them about your\nown designs.  Never assume you have the final answer.  Therefore,\nleave room for your data formats and code to grow; otherwise, you\nwill often find that you are locked into unwise early choices because\nyou cannot change them while maintaining backward compatibility.\n<\/p>\n<p>\n  When you design protocols or file formats, make them sufficiently\nself-describing to be extensible.  Always,\n  <em>always<\/em>\n  either include a version number, or compose the format from\nself-contained, self-describing clauses in such a way that new clauses\ncan be readily added and old ones dropped without confusing \nformat-reading code.  Unix experience tells us that the marginal\nextra overhead of making data layouts self-describing is paid back \na thousandfold by the ability to evolve them forward without \nbreaking things.\n<\/p>\n<p>\n  When you design code, organize it so future developers\nwill be able to plug new functions into the architecture without\nhaving to scrap and rebuild the\narchitecture.\nThis rule is not a license to add features you don't yet need; it's\nadvice to write your code so that adding features later when you\n  <em>do<\/em>\n  need them is easy.\nMake the joints flexible, and put \u201cIf you ever need\nto...\u201d comments in your code.  You owe this grace to people who\nwill use and maintain your code after you.\n<\/p>\n<p>\n  You'll be there in the future too, maintaining code you may have\nhalf forgotten under the press of more recent projects.  When you\ndesign for the future, the sanity you save may be your own.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2873492\" id=\"ftn.id2873492\">[9]<\/a><\/sup>\n  Pike's original adds \u201c(See\nBrooks p. 102.)\u201d here. The reference is to an early edition of\n  <em>The Mythical Man-Month<\/em> <sup><a href=\"#Brooks\" title=\"[Brooks]\">[Brooks]<\/a><\/sup>; the quote is \u201cShow me your\nflow charts and conceal your tables and I shall continue to be\nmystified, show me your tables and I won't usually need your flow\ncharts; they'll be obvious\u201d.\n<\/p><p><sup><a href=\"#id2878578\" id=\"ftn.id2878578\">[10]<\/a><\/sup>\n  Jonathan\nPostel was the first editor of the Internet RFC series of standards,\nand one of the principal architects of the Internet. A tribute\n  <a href=\"http:\/\/www.postel.org\/postel.html\">\n   page<\/a>\n  is maintained\nby the Postel Center for Experimental Networking.\n<\/p><p><sup><a href=\"#id2878872\" id=\"ftn.id2878872\">[11]<\/a><\/sup>\n  In full: \u201cWe should forget about\nsmall efficiencies, say about 97% of the time: premature optimization\nis the root of all evil\u201d. Knuth himself attributes the remark\nto C. A. R. Hoare.\n<\/p><\/div>\n\n<h3 id=\"ch01s07\"><a href=\"#ch01s07\">\u00a7<\/a>The Unix Philosophy in One Lesson<\/h3>\n<p>\n  All the philosophy really boils down to one iron law, the\nhallowed \u2018KISS\n  principle\u2019 of master engineers everywhere:\n<\/p>\n <div class=\"center\"><img alt=\"KISS: Keep It Simple Stupid\" src=\"denied:data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAdQAAADbCAAAAADxbT\/RAAAAAnRSTlMA\/1uRIrUAAAACYktHRAD\/h4\/MvwAAAAlwSFlzAAAA+gAAAPoAxzgVegAAER1JREFUeNrtnTsOvbgVximSNgpVUqQYIUVKNQ3NSCkyBcVsACnlv4jQ1GlcpWcLbIEtsAW2wBbYws01D+MnGPv4XK7lr7pPf8Y\/MLaxj7OXWUNLkp6qfjKDy7SfTh35Nc+SHq0\/\/Ey6yRbq1Jafzm+Srcp2soA6VJ\/OZ9I9VcMF1DEh\/UJV4wnUufl09pLc1MwmqGNqG32t8lEPtft0xpJ81Omgkk\/nKslPRIWabqdfr0aGmq7TCEREqP2n85MEoZ6Hmtq9cWhrA69Q07hgJCoPqOmGGo3IDnVOlW80yucNarpQIxJZoU6fzkcSpKYFqulCLSpC+lfSA9UTUhUGbGSBqr2j5mTy904KqYnoyVGog+6bOiH9Ak21jt3whqqrfYm\/YRKG9PCyl6ZubvzdknCkeQ5TvDJN27f6dE6T7KWZfzRlmqH8wd8qCUuaJlGfkXShfrfUS\/WHBmrnb5SEJ3UW0i\/Zv5TPZn+jJDzNCsC\/ZH9XPvt0LpPuSQH4x+ynBPXL9We1qfRLgvrl+ilBjU8qwey3BPXL9UOFShLULxdJUONTghqhEtQIlaBGqAQ1QiWoESpBjVAJaoRKUCNUghqhEtQIlaBGqAQ1QiWoESpBjVAJaoRKUCNUgipr6kldVVVNulH5rqmAFiuENUlQBU1EWPJXEaHMZ5iyCG7iAdUiunN17iT5WmbZKqq003raQbNkt+CKnEBARTBJUJn0q7BpkW9R6ZcocJ4LAFFMPKCS6iIsWlXxRdu97yFmHkVV2S\/DGoaOnKDNqzuJ7erPYkcVNSErDr\/yRjHxvadOfaPNZt0a8jWpqeeNprVwrbnL7xhfiV+AXZJ2GIZ3W0bj0PqUNooJRENpVn9\/GvSjlU7PznlN3azUFKVrtJGZu\/AbLpFROWc9gl+gmFBBtH7lP1wEiIDL\/igzdU7pOD3k82KWwth4ZBjFhAqkSyNG\/sgvLj2egku9y0lqdThHBTqqRc0JOQm3b\/c+JIrJIhCo4j+uIrkcv6x9VzOLVXntn0x98b1HeaOYrAKBKob+6Gx\/7B\/HR4xZ4dq8mFjdZ6pkhqN2zJ9ssgkEqli2F81PVs8ARJIY7xgbddSLxixxgcqfbLIJZphQ+PV5ndpeHtsdCcaOt9QjblRh\/tFR4G4uKCa7AkA9\/SXLN0zEl3vZvCqCs0Yny7hbfYBiopqhQN27ljlQhFkIqEfb\/TRTeyAbt\/JGMdmFDHVn6tmVuW9sFndfPq\/0yPWV9lkTOREkqA0wUwiorXUCpXt5o5gwoUIFZwoBtbFOYL3cnDpiKCZMmFDhmUJArewTWA7AaWAAxYQJEWoAphBQuVHXq+bJ0i9xGmBGMWHCgxqCKQRULoHLIana1QbFhAkL6v7YCZYpMNTLSq8HgBrOhAkJ6gzdl7E1vpfE5UAO7W66HAKKCRMO1DEQU2iol41OWjYuAwMoJkICwaHuo1\/gTCGgCtMnrsqSVo0ug2EoJkwYUIdgTIG7NNcP+Gkz1mVgAMWECQHqPp4ZgCkEVPFwy4sCr93KG8XE4BYCakimEFClzT3CbAKBYsIUHOo+QBaEKQRUOeL81WX0XBOm0FDDMgV59CbPmXeeZ\/pxk11hobKZroGYgkBVNufJA+ythGKyKyhUNts6FFMQqJrFOQG2K0Qx2RQS6hicKQxUzT5avhOSP2SyKSBUNuEmHFMYqC\/dSjTw6wjFZFU4qAMCUyCos27j5wL4podisioYVLbhWEimQFD5GbecgDd3RjFZFAoqDlMoqC\/N9qJLUYB2J1FMqAJBZXNywjIFg6rZyXDNvudK0Q+YvEJBPeZZfQtUfiWLoAJohjKeSRio\/OLaL6l+3xp1DRmqCvAIUEyCrKURF3h\/Q0Np0WwMIwF410MxCQBVXrR\/+fjQQ5BQlcgFhyCrRwSTAFAVBXwkAQv1GANT5L0+GtMEA2pAqsBQ3+VhiokDOUwQ2gQFajiq4FDN0asgx\/QCmwSBmjdyCs7hGO4YA6U5mNoygFVwWJMQUMtR7Wb7x3e4NgZLtTd0PEArnIAmAaA2NFdIVMNAfTdR9Xc92O5ZMJPwA\/pBqYaCqovjBk81lEm4R28KVZgoD9fGQBq1d70c+MlNEJOAD8kxqAaEaqgeoRvyIUxCTmdBoBoU6mvSDRNAN+QDmASdeKZQBX\/OHxaqPvI0+KkJbhJ2iqhM9Ymr3s6leVgGP5YNbRJ4MndoqsGh6kZq4dvxwCahZ+gHphoeqiZUtGeQufAmwdfSyFQL0MoLAeprVgZqA8zsBDXBW\/W2C7RPgAFVnV1f+CcZ1ARxfWoIqjhQlcoxxGxdQBOMleQBqeJAfc1S8zTIzHo4E5SYDzJVuOajJ9TKNriCtBLm1qJhFBNBONFZglH1hEqsi058QnwrIBmKieyIADUYVU+odBKYXddhulki2CaCsCKeyVSB7kqehTDYn1+1sxWKiSC02ITyzMhHhFtfLg27q6h1tkIxEYQecBKWqm8hZNZX0eBuhWLCCz80LChV30JY+oZWI5dCfJV7bRgUE16IUJXeNcAwsC\/UZRjHrm3KO93rbaCY8MKMzC2vpQYY3PeFuh6rVTdSKKLnmaiGOFCVtdT+VH2hri0Tq4eXvNO9ETwUE164u10oD+J8Bwx9oW4tE4vZI3wb5uZWbCgmvJD3pZHT9h0G9oW6d\/ev60b+fLw5dIJiwgt7WzC5sXROdWwJOa2FbI2NCe2XxeWNgG+7GzqdnzXh9LG93iyobpNiz9Zt2hmfJFRcZ2MR39loHmnCCQSquOOl5a6Ml9XMwd\/co7UyPkuIPZq+KHC+UKZHmhjScYbqtn\/qBVX+8aLpxLQyPk3oONjTAp+4NMgzTbR+HlDFq+9ipEgN0qenyt9fTFM7rLZYPk2IO9iTAufHTcqHmkBDFQcAr5ptyuJV\/T+Em6\/hvLTaDP00IaETYaxjuCwbemGfNwGGKs3DuOp8agJ\/aVbaihe0YXRFHKLS9+zOExKbA\/oIKfwlZGjBPsAEGKr8h6vxLdVSUyuJR6AfB7V6SHuRkJhCrilxfgmTqVfyBBNQqGoMmXOqU6aRUitZnJbqagXdXfUiIXk9d94I9dpE+B8Yg6Q\/wQQO6tzpVlgWrfmh8GCIONNIf7mi1TcWiVwnpMt+9e7fv9WTRqTRmG8sDzABgdoRUhlDAmV51UhDH+\/fE1IXxn9kdcf\/nmcm3KZpOmbjkrSDeNUbE+IKoCGN+VA2nfb4H2ACArXKLkVu\/p7\/i7FTZpUO53zeuyPs41l77bNCPA8z9wCTb4BqHD65C\/V8HEaYFjZ3hgBHZXf15OHzJiBQg2taT+rKe5HDWUKjdNbPPZHryIrYPPf9uMmhJ0N9H0NPepB1gycJve\/QyjcDbS9Q1e879NeYHG6PhprkpAQ1QiWoESpBjVAJaoRKUCNUghqhEtQIlaBGqAQ1QiWoESpBjVAJaoRKUCNUghqhEtQIlaBGqAQ1QiWoESpBjVAJaoRKUCNUghqhEtQIlaBGqKihjpVbqNIxyMYHdvmbO\/8dAsNB3VeLFHTJM6mWpT155R4PlmWrMKyhnttl6VRB+mZbjFQ4BIObuya3P2LV846U\/E1LclvSlXPQs5BX6rLEvN3e9O8DaB0T2jTUfIKyxjxrRnqVZWzpfO6Q93lo7f+l8bwjJX9TX7OP7EPvKwpa\/dL\/zszIP2bozCcof1fsyxffJb0V8Fg4xVe1PmKd5x1p8jczc4rBcb1fcKh7\/kG27uMS7KXiIAfudwn7ulgXnsaz97ojf8WVuh8zyHaMR4JTLhVdwWWTIEHVeio5czQf3AseByoQUw5qKVdcGbe6evbbDNP6iLWepd\/2LDxU15C\/KFA7IKYH1EYJOZPz8Ydqr8Xn1kes82yu40hZmrtH0ceA2smhgea2pGFUZv3bvsleMyne\/YTZkCDbbJT7iq6bV4IxzV25p1mvreeaQqAvytFodyQs5UyW6slyVu\/Z4zOqWLH80VKqaNeIM8+dt5tHgNrJ8czGolvCKW1nufB2ILSdv8V0U+KgHZlRroc1BCcfhGgNmLEwzOlZT8g8vV3muelo4S8xa7R2R0tAzKgi1fPI2chuPc2WC9mK5W\/JbEXD2w1leXxUOV\/y4aG2CtP15J62XY\/Ft+PSU6MXR5er9c8J1C2mHdcXnpaoU7TAaEuyrMb1R83yYlw7vFq7oyUgZFQjxZPLGcvq1ktRrFj+Xktkwm4jyYq7eS7URom1V2xVTrkevvR2KYOllGjzb1ATNEGlvT6q4vjPzP183QWHhtDrt0+ql8Fu\/5ecM41UTxUqe6VY8Z3SvWt9\/I0473cRHOpyKvNUu72EyBLWVnrLl0atRHs9hfq+Xa1hhgj\/e\/OL\/GWw2z5Rc6aRztMEVbXaP5iOZvTxt+65UJdbCl991WLKtWLEDDulMM+h0iJekqmUJo\/Ni046q2q7IlA9r6HuVvsH7bGd1vG3wbncERpKlCp3W5VGb9XBXGY4KtZXUN\/\/WRoitZyUzQtml+3tVsuxatnzGurILNYPquNAvwSqRFXCodLRlIaU4MsMdYvK2JmK8+TFS4Fq3VARPa+hviSouQ7qy3kHP5TBh4Xq9tkoxQ0Eh7q4Se1YF6jjnVEEwfM+1EwL1XkHP5xhQnrMDfuwFH9Tqv8yfWeGSriqsrjFUrVjJX5xqZg8raAK99TsGLvnitt5ZAoH6tLlbtiHbHxp7uS3\/GFNmWaA1wC159piuRvUScKh5kySyfMa6m61f1Ac8Sg9ipsJ6SkNR7XiOir1IL\/lD6u\/0U8dxY5MaSrOkxe91E9VcybJ5HkNdbfiarJG\/rGH0J6n5jvV7mhT0AFZ6S1\/WM2dEaWjU0AHGNpLhOqLRroTKznrcumpmsGTQR12giar\/YOBRdo+xiNejhOsAkOd+f8uqS7DpMtF+y6Nvio1b4\/RnzHLJ2OCOT1DZu76KdjQen3cDLUIZ+UTyW7\/Ss5ZnknBk7Wee87eH+UdnepSFpynaMUyUW1doqVGm\/ZzwXUnUrQ5SuseF2XP7cCy9telt8uR5h0dIS1GXYJr4Nwmy8e54qDTQdPmXcRjfWxzQ4fj+v3FsBbndmZNGVfZinbsx3LO6Eth4yadJ8vZvjdFM9PHL63OiuVvsarH11QuhV\/vZOCms\/wOA3WfTVjSmqarCi4xOmcur9lJL72lR1qoT7zEJKYqy4XHI6Tr141AimYriqFZRvDqbnuRDRMb02uXnzadajdt\/yJqzohcFqonnzOaakYr7KruZs2Rsfxth0ethldWbjuFTIXzbEKVYPYL2D3VTciGd+zmqbD+7VuDtO+GjRVEIPLfEtQ7dgPQrIaw+ocK9acE1aSxcmy64B7ZX1WoPyeoBvV+TNGOTAH4p+yfCapeA\/Fj+jmof9N0aa63vQfUjAsV0Q7LSt3L9FcNVNemtcuBLwtmPBfdPNIOz0rdeIxknfKZ68iGg\/Y9v5AWEyLaoVnNyg6kb6Sj8hnmpZrkK80OgWP2Uknj3lWTfKTZHTx\/ZS\/NpoFAyyaSgmvUXJL1G6p6U80y58dASagysMvWJ1yKqnSxPl6jfm\/Z+SXPxT1Ukv\/95wdJeqR+\/Pu\/vxs2x65fFOqQJUWkYYFqudN00leIPs9elt19OiNJcBo2qKa7atL3aZkbs0Cdcv\/Ukp6gdYrb+iCh808u6QlaxxfY1NSkCLQvi9g6sqkFHIH22fAstETpn2bSZ8UWkLKH84nqt4tbFMxGEufUsflq1dzyfW6AmPinnPQp8VMshLlRfeqvfqlyYV6DOOFtDymT9FXKpdms8izGKfVYv06NvCJHnZo6takd\/EXal82dQ124Vqka\/gLlVatdNmeeRD4MXTsMeHOAk6w1D0PbDeY1yv8HFBcVunK60PsAAAA8dEVYdGNvbW1lbnQAIEltYWdlIGdlbmVyYXRlZCBieSBHTlUgR2hvc3RzY3JpcHQgKGRldmljZT1wbm1yYXcpCvqLFvMAAAAASUVORK5CYII=\" \/><\/div>\n<p>\n  Unix gives you an excellent base for applying the KISS principle.\nThe remainder of this book will help you learn how.\n<\/p>\n\n<h3 id=\"ch01s08\"><a href=\"#ch01s08\">\u00a7<\/a>Applying the Unix Philosophy<\/h3>\n<p>\n  These philosophical principles aren't just vague generalities.\nIn the Unix world they come straight from experience and lead to\nspecific prescriptions, some of which we've already developed\nabove. Here's a by no means exhaustive list:\n<\/p>\n <div><ul><li>\n    <p>\n     Everything that can be a source- and\ndestination-independent filter\n     <em>should<\/em>\n     be\none.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Data streams should if at all possible be textual (so they can\nbe viewed and filtered with standard tools).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Database layouts and application protocols should if\nat all possible be textual (human-readable and\nhuman-editable).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Complex front ends (user interfaces) should be cleanly\nseparated from complex back ends.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Whenever possible, prototype in an interpreted language before\ncoding C.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Mixing languages is better than writing everything in\none, if and only if using only that one is likely to overcomplicate\nthe program.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Be generous in what you accept, rigorous in what you\nemit.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     When filtering, never throw away information you don't\nneed to.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Small is beautiful.  Write programs that do as little\nas is consistent with getting the job done.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  We'll see the Unix design rules, and the prescriptions that\nderive from them, applied over and over again in the remainder of this\nbook.  Unsurprisingly, they tend to converge with the very best\npractices from software engineering in other\ntraditions.<sup><a href=\"#ftn.id2879339\" id=\"id2879339\">[12]<\/a><\/sup><\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2879339\" id=\"ftn.id2879339\">[12]<\/a><\/sup>\n  One notable example is Butler Lampson's\n  <em>Hints for Computer System Design<\/em> <sup><a href=\"#Lampson\" title=\"[Lampson]\">[Lampson]<\/a><\/sup>, which I discovered late in the preparation of\nthis book.  It not only expresses a number of Unix dicta in forms that\nwere clearly discovered independently, but uses many of the same tag\nlines to illustrate them.\n<\/p><\/div>\n\n<h3 id=\"ch01s09\"><a href=\"#ch01s09\">\u00a7<\/a>Attitude Matters Too<\/h3>\n<p>\n  When you see the right thing, do it \u2014 this may look like\nmore work in the short term, but it's the path of least effort in the\nlong run. If you don't know what the right thing is, do the minimum\nnecessary to get the job done, at least until you figure out what the\nright thing is.\n<\/p>\n<p>\n  To do the Unix philosophy right, you have to be loyal to\nexcellence. You have to believe that software design is a craft worth\nall the intelligence, creativity, and passion you can\nmuster. Otherwise you won't look past the easy, stereotyped ways of\napproaching design and implementation; you'll rush into coding when\nyou should be thinking. You'll carelessly complicate when you should\nbe relentlessly simplifying \u2014 and then you'll wonder why your\ncode bloats and debugging is so hard.\n<\/p>\n<p>\n  To do the Unix philosophy right, you have to value your own time\nenough never to waste it. If someone has already solved a problem\nonce, don't let pride or politics suck you into solving it a\nsecond time rather than re-using. And never work harder than you\nhave to; work smarter instead, and save the extra effort for when\nyou need it. Lean on your tools and automate everything you\ncan.\n<\/p>\n<p>\n  Software design and implementation should be a joyous art, a\nkind of high-level play. If this attitude seems preposterous or\nvaguely embarrassing to you, stop and think; ask yourself what\nyou've forgotten. Why do you design software instead of doing\nsomething else to make money or pass the time? You must have\nthought software was worthy of your passion once....\n<\/p>\n<p>\n  To do the Unix philosophy right, you need to have (or recover)\nthat attitude. You need to\n  <em>care<\/em>. You need to\n  <em>play<\/em>. You need to be willing to\n  <em>explore<\/em>.\n<\/p>\n<p>\n  We hope you'll bring this attitude to the rest of this book. Or,\nat least, that this book will help you rediscover it.\n<\/p>\n\n<h3 id=\"historychapter\"><a href=\"#historychapter\">\u00a7<\/a>Chapter 2. History<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Those who cannot remember the past are condemned to repeat it.\n  <\/p><div>\n  \n   \u2013 George Santayana\nThe Life of Reason (1905)\n  \n <\/div><\/blockquote>\n<p>\n  The past informs practice.  Unix has a long and colorful history,\nmuch of which is still live as folklore, assumptions, and (too often)\nbattle scars in the collective memory of Unix programmers.  In this\nchapter we'll survey the history of Unix, with an eye to explaining\nwhy, in 2003, today's Unix culture looks the way it does.\n<\/p>\n\n<h3 id=\"ch02s01\"><a href=\"#ch02s01\">\u00a7<\/a>Origins and History of Unix, 1969-1995<\/h3>\n<p>\n  A notorious \u2018second-system effect\u2018 often afflicts\nthe successors of small experimental prototypes.  The urge to add\neverything that was left out the first time around all too frequently\nleads to huge and overcomplicated design.  Less well known, because\nless common, is the \u2018third-system effect\u2019; sometimes,\nafter the second system has collapsed of its own weight, there is a\nchance to go back to simplicity and get it really right.\n<\/p>\n<p>\n  The original Unix was a third system.  Its grandfather was the\nsmall and simple Compatible Time-Sharing System\n(CTSS), either the first\nor second timesharing system ever deployed (depending on some\ndefinitional questions we are going to determinedly ignore).  Its\nfather was the pioneering\nMultics\n  project, an\nattempt to create a feature-packed \u2018information utility\u2019\nthat would gracefully support interactive timesharing of mainframe\ncomputers by large communities of users. Multics, alas, did\ncollapse of its own weight.  But Unix was born from that collapse.\n<\/p>\n<h4 id=\"genesis\"><a href=\"#genesis\">\u00a7<\/a>Genesis: 1969\u20131971<\/h4>\n<p>\n  Unix was born in 1969 out of the mind of a computer scientist at\nBell Laboratories, Ken Thompson.  Thompson had been a\nresearcher on the Multics project, an experience which spoiled him for\nthe primitive batch computing that was the rule almost everywhere\nelse.  But the concept of timesharing was still a novel one in the late\n1960s; the first speculations on it had been uttered barely ten years\nearlier by computer scientist John McCarthy (also the inventor of the\nLisp\n  language),\nthe first actual deployment had been in 1962, seven years earlier, and\ntimesharing operating systems were still experimental and\ntemperamental beasts.\n<\/p>\n<p>\n  Computer hardware was at that time more primitive than even\npeople who were there to see it can now easily recall.  The most\npowerful machines of the day had less computing power and internal\nmemory than a typical cellphone of today.<sup><a href=\"#ftn.id2881371\" id=\"id2881371\">[13]<\/a><\/sup>\n  Video display terminals were in their infancy and would not be widely\ndeployed for another six years.  The standard interactive device on\nthe earliest timesharing systems was the ASR-33 teletype \u2014 a\nslow, noisy device that printed upper-case-only on big rolls of yellow\npaper.  The ASR-33 was the natural parent of the Unix tradition of\nterse commands and sparse responses.\n<\/p>\n<p>\n  When Bell Labs withdrew from the Multics research consortium,\nKen Thompson was left with some Multics-inspired ideas about how to\nbuild a file system.  He was also left without a machine on which to\nplay a game he had written called Space Travel, a science-fiction\nsimulation that involved navigating a rocket through the solar\nsystem. Unix began its life on a scavenged\nPDP-7\n  minicomputer<sup><a href=\"#ftn.id2881412\" id=\"id2881412\">[14]<\/a><\/sup>\n  like the one shown in Figure 2.1, as a platform for the Space Travel game and a testbed for Thompson's ideas about\n  operating system design.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 2.1. The PDP-7.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"The PDP-7.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  The full origin story is told in <sup><a href=\"#Ritchie79\" title=\"[Ritchie79]\">[Ritchie79]<\/a><\/sup> from the point of view of\nThompson's first collaborator, Dennis Ritchie, the man who would\nbecome known as the co-inventor of Unix and the inventor of the\nC\n  language.\nDennis Ritchie, Doug McIlroy, and a few colleagues had become used to\ninteractive computing under\nMultics\n  and did not\nwant to lose that capability.  Thompson's PDP-7 operating system\noffered them a lifeline.\n<\/p>\n<p>\n  Ritchie observes: \u201cWhat we wanted to preserve was not just\na good environment in which to do programming, but a system around\nwhich a fellowship could form. We knew from experience that the\nessence of communal computing, as supplied by remote-access,\ntime-shared machines, is not just to type programs into a terminal\ninstead of a keypunch, but to encourage close communication\u201d.\nThe theme of computers being viewed not merely as logic devices but as\nthe nuclei of communities was in the air; 1969 was also the year the\nARPANET (the direct ancestor of today's Internet) was invented. The theme\nof \u201cfellowship\u201d would resonate all through Unix's\nsubsequent history.\n<\/p>\n<p>\n  Thompson and Ritchie's Space Travel implementation attracted\nnotice.  At first, the PDP-7's software had to be cross-compiled on a\nGE mainframe.  The utility programs that Thompson and Ritchie wrote to\nsupport hosting game development on the PDP-7 itself became the core\nof Unix \u2014 though the name did not attach itself until 1970.  The\noriginal spelling was \u201cUNICS\u201d (UNiplexed Information\nand Computing Service), which Ritchie later described as \u201ca\nsomewhat treacherous pun on Multics\u201d, which stood for\nMULTiplexed Information and Computing\nService.\n<\/p>\n<p>\n  Even at its earliest stages, PDP-7 Unix bore a strong\nresemblance to today's Unixes and provided a rather more\npleasant programming environment than was available anywhere else in\nthose days of card-fed batch mainframes.  Unix was very close to being\nthe first system under which a programmer could sit down directly at a\nmachine and compose programs on the fly, exploring possibilities and\ntesting while composing.  All through its lifetime Unix has had a pattern\nof growing more capabilities by attracting highly skilled volunteer\nefforts from programmers impatient with the limitations of other\noperating systems.  This pattern was set early, within Bell Labs\nitself.\n<\/p>\n<p>\n  The Unix tradition of lightweight development and informal\nmethods also began at its beginning.  Where Multics had been a large\nproject with thousands of pages of technical specifications written\nbefore the hardware arrived, the first running Unix code was\nbrainstormed by three people and implemented by Ken\nThompson\n  in two\ndays \u2014 on an obsolete machine that had been designed to be a\ngraphics terminal for a \u2018real\u2019 computer.\n<\/p>\n<p>\n  Unix's first real job, in 1971, was to support what would now be\ncalled word processing for the Bell Labs patent department; the first\nUnix application was the ancestor of the\nnroff(1)\ntext formatter.  This project justified the purchase of a\nPDP-11, a much\nmore capable minicomputer.  Management remained blissfully unaware\nthat the word-processing system that Thompson and colleagues were\nbuilding was incubating an operating system.  Operating systems were\nnot in the Bell Labs plan \u2014\nAT&amp;T\n  had joined\nthe Multics consortium precisely to avoid doing an operating\nsystem on its own.  Nevertheless, the completed system was a rousing\nsuccess.  It established Unix as a permanent and valued part of the\ncomputing ecology at Bell Labs, and began another theme in Unix's\nhistory \u2014 a close association with document-formatting,\ntypesetting, and communications tools.  The 1972 manual claimed 10\ninstallations.\n<\/p>\n<p>\n  Later, Doug McIlroy\n  would write of this period <sup><a href=\"#McIlroy91\" title=\"[McIlroy91]\">[McIlroy91]<\/a><\/sup>: \u201cPeer pressure and simple pride in\nworkmanship caused gobs of code to be rewritten or discarded as better\nor more basic ideas emerged. Professional rivalry and protection of\nturf were practically unknown: so many good things were happening that\nnobody needed to be proprietary about innovations\u201d. But it\nwould take another quarter century for all the implications of that\nobservation to come home.\n<\/p>\n<h4 id=\"id2879627\"><a href=\"#id2879627\">\u00a7<\/a>Exodus: 1971\u20131980<\/h4>\n<p>\n  The original Unix operating system was written in assembler, and\nthe applications in a mix of assembler and an interpreted language\ncalled B, which had the virtue that it was small enough to run on the\nPDP-7.  But B\nwas not powerful enough for systems programming, so Dennis Ritchie\nadded data types and structures to it.  The resulting C\nlanguage\n  evolved\nfrom B beginning in 1971; in 1973 Thompson and Ritchie finally\nsucceeded in rewriting Unix in their new language.  This was quite an\naudacious move; at the time, system programming was done in assembler\nin order to extract maximum performance from the hardware, and the\nvery concept of a portable operating system was barely a gleam in\nanyone's eye. As late as 1979, Ritchie\n  could write: \u201cIt seems certain that\nmuch of the success of Unix follows from the readability,\nmodifiability, and portability of its software that in turn follows\nfrom its expression in high-level languages\u201d, in the knowledge\nthat this was a point that still needed making.\n<\/p>\n <div class=\"center\"><img alt=\"\" src=\"denied:data:image\/png;base64,\/9j\/4AAQSkZJRgABAQED6APoAAD\/2wBDAAMCAgMCAgMDAwMEAwMEBQgFBQQEBQoHBwYIDAoMDAsKCwsNDhIQDQ4RDgsLEBYQERMUFRUVDA8XGBYUGBIUFRT\/wAALCAIKAowBAREA\/8QAHQAAAAcBAQEAAAAAAAAAAAAAAQIDBAUGBwgACf\/EAFUQAAECBAQDBgIHAwgGBwgCAwECAwAEBREGEiExBxNBCCJRYXGBFJEVIzJCobHBUnLRCSQzYoKSwvAWQ1OisuEXJXOTo9LxGDREVGODs8NFZJQmVf\/aAAgBAQAAPwDtyan1q+8fnERNzayDqTEJOPqUftWN4jFJW4co1MIOvhgFCFXJ+0rx8vSGExOWBF79bGIeYcKiTexhk5bc23hLmK1ssi3jrBA8fvBKulxoYTdWhQvqkee0KttFICl3CTsB1\/hBgb7Cw6AQo2nN189oesj+ESEugEaxJSraU2vFjp0uh1IBSFHoTEgUhpYQwrnOdU7pHqekSMuySq7zatNijUD2iUlpFt42Q4kn9k6GJFmirTqRElLU\/LYFMS0rIdbWh8JYITDF99KFFDY5rngNh6npANtFagp1WfwQPsj+MS0u50vp4GJNhxPpD1tV\/ODx6CFwXsNT4CBsTv8AKDR6EJudYkWi4+6lpA6qNogprEcxOApkGeW3\/wDMPiw9k7mIxUu2XOdMuKnHxrmc1A9BsI8lx+cWUS7anVDcI6ep2ESclhdbqkrnXLDflNn81fwidlZJiSbyMNJaT1CRa8LFQTuQIZz1RTKKCN1q2SNSYbOMztQllJUEy9\/slWqvlC0hRWJKyiOa9\/tF6n28IeuNIeSUuIC0nooXERszREBJVLr5J\/ZOqf8AlFBr\/E2Qwu65Kur+KmkG2RhQKfdWw\/OMh4gYtVjpTAn5SWSywsOM2TdxBB\/b3+VrxU5mbKrgH2vEY+6Te536QyfcQjcknyiOmZxRBtoD0AiJmFlXX2EQ0ywlSysXbX+2g2MIOTkw0bOJD6bfaGiv4GEC6zNXyKsr9hQsR7QzebAvpvHpWiT1WWW5WWW8rxSNB+kWWncJJh8BdRmEsJP+ra1V84tVOwZSKKAWJNC3AP6V3vKiReJykDQeJ0iJqM+xKNhx95DaCcoUtQSCfAE7nyhw1hquztKmKk3TDI05htTpm6s58E0oAXsCtJXr0OSx8Y5gonGnEX\/tft4SVNI\/0YnmEqYlXGklbYVKB4WctcnNpHO\/GLHGJuEvaLxpMYcqj1PUqfLq2km7ToUkK76DorfqI2fhb21aHiPk0\/GUsmhVA2SJ5m6pZZ8VDUo\/EeYjoNmYlqrJtzUm+1NyzqczTzKgtKweoI0MMn29ze0R0w2dYjJkJAVY6+URcw5ZOg\/CI19C1Xv8hDB9iwN\/xho60NfzvDF1tO5Ta+vlDF9IGv6xGTJtcJIsfxgZZJdp1USB9lCFfJVorz4VcgnYwwWm57xVfyj7BzMwbG508YjJh+\/WI19SlqN9E7lR6RHTM6EJLaNE9VdTEY8+NNdYYvErJNzDF1dvntDZxzTQQ3TnWrQHWFiwUpGa4J0AtqYXlaBNTiwoy7irbNpFx7xKDClTdA\/mSwk\/tG0HRhCooBPI+ahB\/oKeZGsmTb9kiF2qe8kd+VcQR4ph7LSaL2vr4HSH6JdDJAJus7IAuTEzT5Zx0gOK5TZ+4g6n1MWyRpqENpCEgADS0SbNPCrAiHzdMQrRSQoeBh\/LSKkEBCloHkbj5GJFpl5A1Shzz+yYdGdalk3eSpo7AEXufKCL5k4nMVcpn9lJ7xHmekN1IDQyoASnwEJgkesOWnre0O2p3LD5md84UdrTLPdvnX0SmFmw\/MpStaw2g6lAFj84coCECybQfMIQmp5iSaLjzqW0jqTENMV6ZmwRJtclv\/bvD8k7xGKaa5nOeWqaf\/bdN7eg2EGbMxUFlEugueJGiR6mJSVwug5VTTpcPVtGifnuYmmWG5dsIaQltA2SkWEGWsNpKlGwENE1Rtx4NoSpZvY5Re3rCIkZiacJfcLbd9EIOpHrD1qVaZtlQLgWCjqbesLQjNTjEiyp2YeQw0nUrcUEge5ig4j40UmlhbdPSqovjQKHdaB9Tv7Rk+J+JdaxCVJmJxTEuf8AUS\/cTbz6n3ilTE7l8LRGOrUu5R9WPG+nyhkuZU0o5k8zwI0\/CGj00h06LAJ2SdCIYvki\/WI59e\/haGbmib6Xhk51OwhH4RcwoIaQXFHZKBrvExJcMahVilUwgSqN7ufa9hFrpfDKmUzV9Tk+sHTnHQe38YsKJdqVbyMNIaQOiRYQ0ecSLgm9tbDpGdYo41YXoCnGGZs1acSSPhqWkPqBHQruG0n95QPlGV4i4y4orwUmnsy+HWL2C9JqZt6qAbT6ZVesQ+BcaV3AuLXMRSL30lXFy65ZEzVLzK0BRBujMe6dBYDTyiYxKcZcQZlM3iisPqRfOn6QdKUp\/cZH6JjGcUtIoHbX4evIc5iJmVkE8y2XPdos3t\/ZjNe2nThI9obEAA\/94al3gbb3aSD+UYS6gJ3A+cXHh5xhxXwumguh1NxuWvmckX++w543SdvUWPnHV3DXtcYZxoGZOvpThuqKsnM4q8s4fJf3fRXzjZVoEw0HW1BaFC6VINwQeoI3iPmJU5SVaesR78rra1j4RHvN5RtYeMR8wgJNhc9dYjJjKkbC8Rsw4CCD3hEe\/fXXSIyYaJV+F\/OJLD1JenpPECmigolaeqYdC1gHKFoGlzqe8NBrvFYeYSBpqDre0MDL5ibDS9o+tExMWvqDDFxeYFRVYA6q8IjZyaz90aIBuB4+cRrq7nUQyfUB\/CGLxFzqR6QgslRN9R5wiZRxavqwbnoYtGFcGzVTSFKQEt5rFw7D+MXyUwjR6CgzExkccA1fmSLD0GwiDrvFzBuHUqD9VYJT91nX8RpGYYl7ZfDjDqv5zUWkE7cxxIJ\/OIGldtrBOI1OClXnUt\/bU3mIHvaEFduDh0xNuS81U5aWeaNnEqc+z66RbMN9qHh9ihaUSNdkX1q2S3MIv8rxodMxPQ60gFt9pwHxsfxixSNKk3E3YAF+qDe8S8vRVWBHeHimJ6kyxQQlQ9LxYWaeVagXh81TdiUw+akkoTdWgA6x4JU5oyLD\/aKGnt4wZEkls5lXccOmdWp\/5Qi9JJUb5cp8U6Q2XKOjZWYeChCK21IGqD7awTN0B2htMVNuVBF86h0EIszk5UTZJ5DXVR3iWkpqVpyfq0853qtUSUvVnJhiZUTcpSFadNYSFWyIzLUEIH3lG1\/SEl1ibmRaWRym\/wDavDX2T\/GGpQ02vmurVMvD\/WOm9vQbCFWUTNTVZhBUjq4dEj36+0SsnhptBzTK+erfINED26+8TCEJbSEpASkaACCuvoZF1G0NXZ1TzaTKp5tzYkaAR5qRccbWmZdKwoaoToB77w6ZYbYQEtoCEjoBCkRlZxLTMPs8yfnG5fS4STdSvRI1MZniPjirvtUeVCRt8RM7+oSP1+UZjW8VT9de5k9OOTCr3AWbJT6DYRATU9lH2reEMC8twnp67wi4tKfNQ8YZvTBiPfdF7gwwfsoHMAbncw1UVpuELI8laiG61knvoN\/FOoh7T8Mz9XsWJc5D\/rFaD5xZ6dwzZbAXPPl031Q3oPnFmkqTJUxsIlpdDXmBqfeDuuhJN1WvFbxHjSkYaumfnmZZ0i4ZUSp1Xo2kFR+UUOf4jYixC61L4Vw6twPBSm52prCGwBe55aTfoftLSfKMkxg7XavOzMjiGpPzimHVNOSiVBuXCkmxAbRZKhcbqufOG9OwbOTbCTLynIltg65ZtoehNh7CJJjCdOlRebmnJ5f+ylRkRfzWoX+SfeHrLnwN00+XZp6f2mE98+qzdX4w0caK1XUSpROpvvGG9oNtdD7Q3B+pLSW8zMie9obJnHR+REV3t9yRleNzL40EzSmVbdQVp\/SOaXGiBmV3hba\/WEXbqUc2h2sNoRWDm1FiY0Phrx5xZwxcQ1ITpnKYCCqnTl1tEf1eqPb8Y6s4c9pjCfEMNys06KBV12Hw02oBtavBDmx9DY+saVNIAuoag7ekQ8zc7C8Rkw0bnreIx9o7jT3hi5Lb+PrDJ5ka6fKI92WUomw06AQ1ekOagBaAdcyb66wiuUsDrc3sBDJ2XCV2CgB62j6duTt9SRk\/E+kNH5tLgAPdA2Ahk8+Dsq\/rDNx82Iho44TeGyyT423vBcpG4sBrEjTAl1xIIs2dydz6eEW+vY7aw1JMU6mNIVOKbvrolpNtzHGPHrjjiicrDdFo6Z6s1acKkyzEugqSkDTOq2iUxgtX4VYykFOVXHmO5eitL7\/w6QqYcT5BIskfMxRKhKcJ0zhmp+rYhxG8D3j9WwlfoLE2iYXxoweKb9F0zAlUlKeE5QZecKCoedkXiuU9vh3V3nUu4KxDIAgZnEVDN1\/rI1i60OU4M4R5L7k7WJGacH+vaS+oeoTYgRN06l4sqte+kOHXFKmzYQbs0tMyZZy3QKQ5ooxqWFu2fxG4L1VmQ4p4RnFU+4T9JyjRSd976oX7ER35wY4x0LirhyVrmH6g1VKa9pzGtFNq6pWndJHgY2ymyLU4lK0gE+kTLVMCbfdh2hrLoEhZgUSfMOZ3vHon7ohcS48IH4YEwBlAemkJqkQb6bxHzzbcuklWp8BvEBMpdnjlQ2QD0A194RbofJ7y+8rw6CAdbWn0HQQ3XdAzFQbR4q0h1JuPJSsMDIhYyqcdG48k\/wAYWShlhXMWovOj\/WOa29B09ody0pOVMAtN5Uf7RzRPt4xNSeHGGbKfPxLn9Yd0e38YlkpCEgJAAGwEJLm2WzZbiU+pho5MTTz5QwjKlJ+2rb\/nCpp6HnEuPnmKAtbZN\/G0O0pCQAAAB0EeJtFYxHxIoeGwpD00JiYH+ol7KVfz6D3jLcS8aapUwtuQSKawdLpOZwj16e0Z5N1J6bdW666t1xWpW4okn3MNFuKV43PWE1LQlNl94+ERsy6kOZk6enWGTkw5exssD2MIqmArc5T5w3dVm2Nx5Q1c1JAgGabMTzmSXYU6o+Cdon6fw8mXRmnHksJ3yo1VFkkMK02mWUhgOL\/bc7xiRJS2nSyU+toYVKrytNllTE0+1LSydC8+4G0A\/vGKNO8VGZxSG6DTpqtFzOBNZTLyqSkX7zihmI1GqUEG+8UfHFZxKmYRL1CspkEqSrmStGBaSSFqFucfrDoBqCkeUMcD0OWWmedapomlAJURy85Wb63J6n9oxclUacl22Jt6aaoK05xncUl1YQr7qdhtcaBVgBFNnUysvUZl6XabmH3HCszswOY44SblVlAAeyRDCcW9NrzuureVa2ZZJ9obtUt6YZcfCMku2LrfWQltCfFSjYD3jNsZ9oLhvggLbm8QJrU6jT4OgoExr5u3DY9lH0jCsY9tmszRWzhKgyVBa1CZuctOTPqMwDaf7h9Yzalr4i8WsTsYlmWa3ik0xaZh+eU2t1Eu0hWY962VCRYmwsPKN8\/lDaekY7wpPJunn0xSb9DlcJ\/xRya40ULyqGdQAIvpe8N3ASBYW10NtLXhBQGXumxhMtnbeCqsDpeNP4cdobFHD7lyqn\/pekp0+CnFE5R\/UXun01HlHT+AeNuFeJCW2ZWb+AqS\/tSE2Qld\/wCqdle2vlFxmZa17i2m0MHZJQscuw6wydkr3JuDbWGjsmADp7GGDzAAsQPTrDN1g5bFIF\/KGbrBsARvDX4c3PS24Aj6HOzec3NhbYdB6Q2dmr6Aw0dmCQfCG6nFG9vwgpWTp8oEaJuSQPODJly+Ulf2f2f4w9al1psU628Ir+K6POzbzjqFqSXABc7Dpr5RWKbO0vAFQMy7SW5+ecshQSBdxXQZvC\/SK1xC4W4fxBJTWL8ZTPxAbvyKJLrythR8baqjlKuTOKKbJKnaPhigcP6Qp0pYm6nKtNqdSDa4Ll1H2BiovY\/ri1BD3FRhpwDvJp0ktQPpZCRDhniLXZdCko4oTqtf\/iaMpaf+Ew7kMWYhrz2s9grGBIsW6pIJlnVDwBKUH8YetYHwtU3lOYvwJVsBpQM303QZsOyqf62VV7D0MdXdn\/hHiSWMpJvYpl8ecPKiyeS5UWszjV9gpKr3Sdrg6R0FwR7N1I4S8SZisYTdVTKfURy6jQ215pVSvuutj7p8trR1XhxsNzjzadUtLyEfjFlWxzTroPAQoGwkWAAg2UR60DAE2hmJszDrjaUqTkNr+MIzFMS82S6bJGpCYWalEJZTZIGnQQ0m2UNtrWrRIFyYp8zVHZtwiWZ5LYNua7ufQQm2G2lZ1EuOftr\/AE8IlpGkT1RAUlHJaP33NLjyG8WGRw7KyllLT8Q6NczgvY+Q6RKaAeUJLmm03AWCoD7N9TDViYm5ly4bDbPivc+ghRimNNLLirurJvdetvQQ8hN+YalWlOvOJabSLlazYD3ihYj4z0eklbUilVTmBpdvRsH97r7XjLMScTa3iEqS7MmWlz\/qJclCbeZ3MU5yYI2NyekIrcIFyQOnnCZmkC9t\/wCtCTkwVC94aOOm5vrDR1RUPeGjp0tfXpaEUhxxwJQCpXgNTEtTsHT89YqT8Kg\/eXoflFlpuCJGTsqYJml+YsPlE42y3LoCGkIbSNLJFhBFuAnKm6j4JinYv4jUvCk6iQmFuzFTcb5yZCTb5j2QkgKUdEoBIIBURexttEWuexRiWnmZlksYfkXC2lp5VpmaXmBOgPcRa1tleRhB7ANPpplqpOvzFZqEy6Vsvz556mlZcyQ3e+QX00CdREzT1zqGppK5duVlZnK085NqCChKWkoOUDQ3yki5Bio1OnUF2aRMVAO1ucZCglxICWlG99RoPwVBJvE84lGSSbapzAFktsJF0jyPT2tEA5KzVWeUoJdmXN1LN1EeZMZnjrjfw74dlxFZxTKzE6jQ0+kWnH7+BynIk\/vKEc+427dj6uYxgvDDEgnUJn6yv4l71DYshJ8jnjGaliPihx7nCmYmqziVLZzFlsKErLjxyps22POwEOabwVlZQ5sQ4llGFp+1I0VP0g+D4FaSGUn\/AO5fyiy0+nULDZSaLhiXL19J\/EChPPA+IaASyn0UldvGNM4R1atVPHsuzUKzOzSZmSnJQS5dKWQhcu4AlLSbISL20AHSJzt7sGdoHDWppGbmyzqSq\/iltVvzjjh1JSSRfT9nrCKgUq1sPDTrCKkBOYX1vpcbQlZWosSesEUnJt84TOu0eQtTa0qSopUk3CkmxBjZeG\/abxFg8NStXH+kFKQQLPq+vbH9VfX0Vf1EdM4K4mYa4jy3Mo88lUyBmck3u4+j1T1HmLiJx2XKifLpDF2WJvfc6RGuyoSoDUDXrDRyUFhb3hs9LXO34Q1ckbq0TcR3KuZBGv46QkXwRa9hCS1Ai9x6QTU\/8oMF5CNMyjsIXbbIXmVqr8vSHjTXMVrfXwh\/LyRuLHrEg3IqdTlUMyTobxB1\/hi3WFJmZRIbmUG+VWxjPq1giYkq7LzNRlnClrVLaxdObofCIXG\/BXBfGyZkJvEdOAfpQWpDr8ypIe0+yoDcXAsBaMOq3ZwxnRWJypNTUpS6Nzi3I07C9JaL7iel1ubG2+piEf4TYwlAguTOLmCrRKHZ6SCv7hNva8It8NsSVOaEiESk7PrF0SGK6QiXL48ETDOhPvE5gngZXV\/ET2HXJzCtZkHP+ssJ1R74mSmm+vLzdCL2\/OOw+F899BYOpMjR5RluQeVZVPCdZVRPeyHom99NhG9YGwvMTFTVNS90hwAEdEkH7Ua3QaE1RJQtpWp5xay446vUqUdzEpHo9AQGYx6148lAQLCCzCgllZJtoYTS6C0ggjaIDFNREpTyCQM6gmIKk0CfqmVQQZdg68xwbjyHWLfTcOSdOsrLznhrzHNSD5DpEmpQQm50AhuqfaKF8tYWtIvkG\/yhNkzUylXNSGm1DT9qFZaQaltQnMvqtWphzDCrV2n0KXL0\/NtSrY6uKsT6Dc+0ZniPjs02FtUWU5qtufM6J9k7n3tGW13FtTxE6V1CcdfF7hsmyE+iRpEQp1S\/si9vlCClgElSs3kIQXM6aWSPAQg48fGGzq9PPrCBUBfKSmCl1wjooeBGsOpSiTtRUOVLLyn7y9ExOyeBUAJVNvFZ\/Yb0Hzifk6VKU9NmGUN+YGvzhdSwk2BBJ6DeK\/X8dUjDqy1OTzTcza4lWruvnz5aLq9yAIfU1KqvRJesztRksO0qZTnacn182ZWASO6yg2G3VRPiI+enbw4y4w4Y8eZCZwPjGr0+mLpjDzLaHSlpwkrSpSmT3NS3sR7RrfBKs1rivh+mYrrE209UqlR5Vc1NryozrS9Mtnup8kDQCOhJR2Sl8Oy8pN1NTzKUBIl2EZV5kquDm+1bw1G5hvO4tdCA1IMJl2wnLmc7yvlsPxirzzM\/XJ+XWpx2YyE9zU\/IRSMf8Z8A8MgtGIsVyEpMo3p8qr4qav4ctu+U\/vlIjm7Hv8oLIs8yXwVhMvqFwmoV9y\/uGGzYe6z6RhFe4q8XeP0yuQVP1asS6jc02ltFqVQP6zbYCLDxV84ZyHBNuRUFYmxJI05Y+1I0v\/rGbB8CGzykH95wHyi1Umg4boTifojCzc28Np7Ej3xK\/VMujK2n0UXIlpmaqNYCWKtPPTMq2bsyKbNyyR\/VYQEtp9kwoiSukBCQ2kbjawhIyyGXCtxRXawA\/WLrwiUhjiXhnIEkOzqGivyUcp\/AxbO2TLmZ7P8Aw1njophaGVEC5F2LfmiOLl3UlN76C5Pj4Qk60UkXuSQDt0\/yYScbKiT11Op0hopFz0BJPtBVoGa1tbXPSElAgWgloC0LyE9MU2cZmZV5yXmGlBSHWlFKknxBEfR+mEz9Ikpkkr50uhzMTvdIP6wV2UsSLdd4j35PW2\/nDZUn9rTTeEVyQte1+l4aqlSFG2YegjrgOZ\/SAvtaPW1uY8lw7It+9CzajcDzh4yvYEXvD9gA2sCIlJcX6AmJSWSQdQTeJJlxV8jQHM6lWgEPEUpubSUzAEyFb5xeI2b4S0qpXLKVyqlb5dU\/KGiuDNRabSmVnG3m0m6Urum0JL4Y19xKmpumylRYP3VlP4RGVHgZXKsG2m6U0xLNqC20c0XSRtY9IlqT2d65N1UzU07LS7i0hsuBRUUp\/UxrGCOBdKwyykPOGbVe5AGUH9Y06Uk2ZNoNMtJabAsEpFocR6PQEeEeAtBXHEtIKlqCUjck2iuz2OZGXcyMkzNlWWps6D+MR01jhbjTwblAdCEFSt\/WGTvEmSkJdtD7KjMEWS02b5jBJaqT9XfYnX3GJZhslTcslAWr+0Tt7RNjGfINnWUqA6oNoZVbFzD7JclXVtvj7hFifeF6HO1KfbUJ99Al1DQJGZY9xFjlkSsk1mbAA6qvcn1MO0PIcHdUDEBiHH1Fw0FJmpxK3wP6Bk5l\/Lp7xl2JeNtSnkqapbKaeydOarvuH9B+MZrP1OYqD6npqYcmHlbrdUVE\/OGinTBCsWuSSfMQk46Tl6adIQW4rofmIRK9dQLeUEUvS98vrAsScxOKsw2p07d0aRMSWCXnSFTTgYT+ynUxPyGGpGnAKQ1zFj7zmsSKlpbOptbSwhjUqtKUqWVMTsw1Jy4\/1sy4G0\/jvFPqPE9hd0UmRfqXg89eWY+ZBWr2TY+MVapVetVpJE9VFsMK3labeXRbwK7lav7wHlCVMwuW2v5pJIlWFalwgISo+JJ+0fnEizQZVpaviH1zJ3CGhlQPc6n5COD\/AOUopqJfiBhSZbQGkLoqGwhNyBlff6k+Yjaexkpyc4S4XISV3pi2hYXuUT03\/wCcfONcx5xIwdwwl+birElPoiwMwlnnc8yr0ZTdZ9bRzJxB\/lD6FTVOy+CsNP1h0XCZ+tK5LN\/EMoOYj1Un0jn\/ABDx74z8e5tymS0\/VJiXc3pWH2FMshPgpLYuoeayfWIin8A35ZwKxRiCRo6796Rkz9ITnoUNEoQfJbiYuNGwRhiirSKVho1F9Ovx+JnuYL+KZZopQPRalxOTq6jUpVEnPT63ZEHu09hKZeUR6MthKPwv5w1NJQykNpZ5Tfg2m0NzLJaeSpCCopFwSbmHC7FbbgABI0EGQj9oFdr+QHlaG76lqKwUkiwNiL+8SeBXFyGNqBNi6AxPy7t1G2gcBjZ+1hT0v9mCnkpuKdWi0bDYJceb\/SOFVN5gqxNuvlCKm87gA3Nr26b6Qi42BsCB43hB1GUg6n0O0JONgJsNVb7QitsAlIv5adYTULWsYKR4HrpBdo+kvDMif4bYXmdCXKawT5\/VgGJtyUGY2HW28R0xJ2Vewt5w2MqRuNbdYRdZBZyEd2+a9tYZLkjmNhmHjHTtgUkWtrHiAlIJVYDxgmVTg1JSk7jqYWQNRYWEOG0jTWxh6w1rcaRKyUsVnaJ6UkiALj3MSstILXbTKnxtrEzKUlsgXSQfGJuRpKEkWv7iLBJU4C12gfSJhmntZdUlJ8CIK8w2wN7noBuYUl2VOWzd1P7I394lZdpIAA0t0h+0m0Kx6PA3j0DBHHUNJKlqCQNyYr1Rxmy0VNyTZnHRpdOiB6mKtUJqZqZKp+ZK09Jds5UD18Yh3ZlIVZCAEpNglO0FM4Dobp9RaI56RafnW5nNcpFtDvEmhVrFJykbWhvOTi2xc3N76xEPT6LG6io+Ahp9NPya88s8tg9C2oiJWU4qVSSGSYS1OJtuoZVfMfwiCrePavWbhcyqXa6NMHKPnuYq7j\/Mv1PUmG6gQd8g8jDZbmQ6G48xCS3VX218oKXha23rBVOpIsDf0hO6ibA+kPpLD87P6pa5SP2ndIsMjg+WYTeYUZhfhsmJhLLUq3lbSlpPgkWgvNGUlKbgC5UdAPeK5N49pbT3w8q6urTeYo5NOAWAoC+VThIQk+RN4qmIsYYiM0uVbEvQ0BIKuUA\/MDTbOoZR7JPrBaJhyVemG5x1L9SqbqLl+aUp9QBHifs69LgRL4lpxnpmVM22mQDTIbDEuASoDqToB+MNWmGZWwl5dCFDZbnfX8zoPYQqW3XyMxUtZ6kkkxWMccQcJcM2FTOK8SU2gpKbhqafHOWP6rSbrV7CPnX21uPOFuOmJ8OowgJ6Zl6ZLLlnJqZY5fPWXCoctNyq2vUA+UVvhTxC4n1ikM8LKFi6YwxKsKeMvItJLLzry1jM1nQnm3J+6PlDFXZ4q8vVZkYwxHT6Q+hwh1lDqp+eWb6nlN3KSf8A6qkRc6Fw1wfQRmk8Pv1uYTtOYhd7l\/FMsyQB6LcX6RaJmZnJmRElMTikSH\/\/ADpRCJWVT\/8AZaCUH3BMNk08NtctlpKEDSwAFocMSCW0hNzlG\/nBFo1ISnVOxOwhpMBxzVSrJ3sN4bNSyNLruLbdTCEyzkSAgDXu31jwYDRHMspO9r7GGsy+VKHLSnLm1BMMjOPNTcs41cBp5KrBW1jHUvaNlU1DsyYyQBmMpWnHQB0Bm835OR8+8veVbW4trtCFiO6jTp+EJ2JAAFiO9YQiRlXe9iOttRCLoBIuenyhJ1AbNj4C1obkGxG9jsTBSLpFhCavDwj6N9nuYE\/wSwk7vlleUf7Kyn9IvDzFzsLwxfYuT+cMnGRc2hq431I94aql7nb8Y6PcTyzr3lE7CPGXJVdepHyEKIl1KJ6w4blCbg3FoctSpFjYw\/lpLNra0TkhIKQQVKKR4DcxYZBk3BsFeRG0WSRlgbZmzfyMT0nItLtrb1ET0pS02BGU+8J1HEtCw+kqqFUlZTLuFuDSIul8XsHYinPgqZiamPTN8ob+ISFqPgAd4sC0ho6ak7qOpMC1NZPOHjc8kQ7bqLaRdSgkecG+l21nK33jDhpS1DMvujzhTnoHWCuTjbaSVKCQNyTEDP4vbBLck2ZpwaZxoge8V6enHZ05p2Y5g\/2SDZA\/jASslOVNIElLkt3tnPdQPf8AhE0xgFtxKVzk0445uUN6I9PEw7nKW1JMBBZYW0BYAoAtFGrUkpUwfg5J5y52l0lVjBadg2dmRmnHEyl9QhSDm94WdwfPMGzDqHgdrKt+cV7Eks\/ISqkzILToN0gi2bxt4xTnZ7qTrDRybCibbwkt1Vrk2t0MNlud6528ITVMJTYAFIhst0K2NwIRUSVaa6QmpVhC0tT5meUAw0pfnbQRNSeCwqypt7+w3\/GJ2UpUpIgcllII++dSfeHBeTm7t1HbTWDpzO9UpFtdbm0Z3T+JL2MJ9Mph2RQltwKLc\/UiSlQB1UlpJuRa5BKh00iRmMIpqRmn61UpmsJlir+aKsGVDNa3JSLG2U2uCdd4lJFCpWRW0JZmUYU0GeY9ZFk6m4A6i4FjbaId9ijrmlTT6F1CbP2r91snxA\/TWFF1uYWhLTITLNgWCWE5be8V7FmLKDgenqqOKK5IUGW3DtRmEtFX7oJuo+QBMc0cSP5RTh7hbmy+FafPYxnE3AfI+DlL+OZQK1f3R6xzJjTtq8Y+LU4aZRJpyhMPnK3TsMMKQ6seHMGZ0n0UB5RWZPs1YwqTqqljWpymE2nPrHXa3MFycVfryE5nL\/v5R5xqvBXgvwkrFampFh+fxdUJeWeUJ5c0JZth9LS1tqSy3muLo+84fSMzwtNronbFpLyFlonE7QKkm2jjoB\/BUaDQqP8ADpnG1Jylt90KNrlSgsgknziUSz3rBJsd\/KElsFougI0ve5hVb2W3XTciEnJnKu6EnXz3huc6lEuAg9AncQmppAKs31h2A8ISdaAOcDUaXI2hssNlaA2oqPj0hB4BJBBBJG4Gu0RyruukC4QPveMNppsI7iftEXA2jrXiC39M9nLiQ19tRaZm7dO9Kyzv5kx88bHTTVWu17CEnQQqwGW2+mtoRU2dSdAbWFoTUn6vx0uRtDdy6hYWzEm4tvCCjcpuNvH1hFKbp1FiT1PzgFIsDqB10H+fGEHB3jc6+Yj6C9kZ8z3AejpJvyZiYa0\/7Qn9Y1h1m5P6QymGrbnUiGS2bD8oausFQNhpvDYywJJIA9o6Mblsh0BKj1MLCXJ1hwiUub7mHrEmT00h\/L0+52+cSkpTTpZNonpClBShdJi1U2iIKU2AvFilKGLDu2iNx1i+i8L8Pu1WrvBAAIZYB77yvAD9ekca4s7YKapMhp2sONpWtSskiglm19EE2NiBa9z46RBVri3S6nJlL1VZaS4nRKnMyl+iRtHIHGLEMiirOTsoxVJpxC7NTcqgpynz3P4x1F2C+3ZNv4jkOG2Paq\/NMzhDFKqFRBDrbp2ZWs\/aB0AJ1vb2+kb7LjRNxp4xFzFSLd8m37StoQZfemzmW5y2\/wBtX6CJtVVZpMq0uWQFOOD7a\/4Q1GInn13Kio\/hAivur0au6ra4PdHqYZzcwqY7029zR\/swbIHt1heTpc\/VLJl2eSz\/ALRwZU28vGLFTsISsrZcz\/O3f647o9B\/GJrmssjIFIRYaJvaGKakubWtlphZtpzPuiCt0db6FJnXi8D\/AKtJsn+MP5aValGg2ygNoHQRH13EtLw9LlyozbTAtcIUbqV6J3MZNifi81UCpqiSSmCD\/wC9L0P90afOM7qNamZ6ZU5MTC5lz9tSrgQxVMArItp4wUzIGg0894SU6pWoIMJLUb66wmVdLwUNqWQkDOSbADWJGUwzNTBuofDpPVZ1+UTclhyUlCC4C+u2hXt8oksyGgEiyR0SkQXO4oHKLDqVdIrtUxxSJBSm0PLqkynQsyQzgHwKrhI9zeAkMb01VN+JqTzspMFSgKdIpzu2GxU6oZRfyHvHzy7S2MqxQO23h+bplWqFNYm52m52mZpaQW1BkFCrEXBBN\/HWO0+HdOkaJT6Y+49MvLQ0UFTP1eXcaa3v7iLY7iNaC4JKXblUqVcr+0tR8SfGISuVmVokiup16qS1MkUaqm6lMJZbH9pZA9hHPvETt8cLcDJdl6O9OY0qCdAinI5MtfzecGo80pVHMeO+3\/xTx9MKp+FWZfCku8cjbNIZL02q\/TmqBVf9wJikNdnriVjiaNaxnO\/QKH++ufxVOKEwseIaOZ5X923nD1eGuDPDNIXU6hUeIFRSSOUx\/NJPMLXFkErVv1Wg+UR9U7UVXkZNynYMpVPwZTFDKW6YyGlrH9dQ7y\/7alRQaNI4x4yYmYpEh8ZXapMkqSwFd1IAupZ2SlIGpUbACOiewtQ36LxPrclPSomEOS2Rp1KyEKVmUgLQfvDvG3Qxk2LJj6I7S0nOXycuqyMxfw\/olGN3qba5HFGIZYrOVqqTSABuAHlDeBAvlsRcjW8MHUKW+tKiAkdb7wDuRtkn7bh2H5Qwm3VpW0Doq99OsIy6jMIKl5r7DNvCxOQEEgfdBEN3Guc7bUi2pvpCqZS1ikFNr77mE3pRABKlgDfa594ZOBvLbUDxttEY60HljJoUq8Y65pDf01wVxpLjUzOGZZ4A+PwXL\/NmPnOlN0C+1t\/OC8s2KTa+p1MJn626dxuCdz5Q3cSe9rbTQWgjqbrBJuQb7aj\/ADaEHh38xGhsCP4QktBVcXACRvbpBHGwlFwrNa19dbQ0WCL2tfqI7u7Dk18VwfnmOsvVHR6XQg\/xje3mgL3AtDKYYFgbXvDFxmyr26Q2W1YnTbYCG62A4oqIuT1PWOiGm9jsYdts3PiYfMSpURYRJy0ncgJFzExJ0w6FSQT5RNydMvbu\/OLFTqVqNIt1Lo+gOWIPjHxFp3Bjh5VMTTiQ8uWbIl5bMAX3T9lA9THznx9xixNjfBU5i7ELqm6hOKEsy87JqXJ045jmYGW41ISMwN7jU9I5pqeHpgzDVUrc1IKl3SVKdk2Dy1i+t1FRsR6XiekuLuAXKImlybUrRFsXAVLMJXMziuinFqFrD9kaDzjMMU8Va2hbrku\/KzdNBKCGmEtlA6fZFvmBFPl+JYqr7X0hLJLjLocamkHKtBBuPO40sRqDt4R90uynxbZ498FqHVETCpmpSrKZSeUvQlxIACz45hY+t4144YF87p5iugGwhnO0lLSSpZCUjx0AiCnZ1l5ptmXQXVNk969kj3huAhIHOXn\/APpp0T\/ziZp9DqFUQMjYlWOinBbTyEWem4VkpDKtafiHh99zW3oNok3phDDSlkghOpAMMjOTFQaSuTSWxf7TosLeUKClNuuJcmDznAOosn5Q+ACRpoIq2JuJdAwrmRMziXpkf\/Dsd9d\/PoPeMrxLxrrdWC26a0ikSp05q+84R69PYe8Z9MTqpt1Tj7zk6+o3Ljyjr+phs9MLc0Uu6eiRoPlCRUFC2a0evbzPlCZVe14Aq89fCFZWSmJ02aYWoftbCJ2Rwpaypp2\/9RH8Yl2JGWkQOU0hB\/aO8KqdKiAlObz6QxqNTlaWgKnZtuXzfZbvda\/RI1PsIh\/9Ip2oOculU7lpNgJqoEpSbnogan3KYrVekqnPT70jPTDtSU2rKpprusbfsJsCP3rwm3Q0sBKXnUMIAsGmQFEeWmg+cLhMrJKzMy6Qv\/au99f8B8o+ffbgnEUztRYXqcw4G2kfAzC3VnQBJbuT5DLHQeMO2Xwq4XSapJNbViyosqWBL0BHMbPeJF3lZUW2+zm9I5vx3\/KG8QMXzBkMEUaVwy06cjS22\/jZ1V9rKUMoP7qL+cUCZ4I8XeKk4KzjiffpiHBm+OxbOqQ5l\/qMnM6R+6i0GVgbg3w3Rmr2IZ7G9SRr8HTh8LLX8CoZnFD15ZhvMdqN\/Dcq5J4Bw3TMHy6hlL0mzlfUPNy5cP8AaWR5RlVaxfiTG0y4Z+oTdQWs5lN5jlv45RpDiVwDX10dubmaRPS1NmHCmXnnWFBpbiR3kpUdDpvbXQR0lRP5PrEEtxLwJhqr1BE0qtyzFRnkSTSrSbCld5KlH7wSDr4kR2Zw97OGGMB1viq9QadT8LycxJ\/QkseapbqGQg53ASSczh1vfoI5T4GzT0n2qJ6m81KZduWAZbSdEhMwgC49Pwjn3j6yqjcZXF3sU\/DuG3ikBP8AhjoTGCUtcScYJQnNeqvLFzsFHPp\/ehqCEJSSSnu2NtzDcuAKUDe2+p1ho+rmkiwNvsiEVpN05we7oehgqU5F5UkJSohRub6woltKrXTmIve5sBHgdCEISdTt6Qm86SEjNY2+1be8JLALaftAW8IbON505hcJA3\/5RE5SlallI0VceMde8Fz9MYKdlyL\/ABWF0tEeJQ9Nt\/kRHzlKMi1J0uCUkHobwk4CCb2KdLga2ghQoAm4BPUdIIoFNgCo3G23vCTicpBCTa9731\/ztCLrYI0IsbnfWEi0oElIuPz33hEpBIsLncgekNJhvKCctuusdo9gKb5mEsWyd78qcadt+8gj\/DHUD7FiYZPtgC9oZOItpYb62hm4356npCIBEdES7AVbSJOXlhm136CJaVkSu3QROSVPtYBMTknIZiNLGJ+Rpo7ukWCSkEospQsInZUKS3e\/LbAuVK8I+cv8pBx6pmK5\/CmCKHUVqlHaghEzPtJIRzCtKLIWdFFIJuRoLxn\/ABTqEjUV1PB9Ops0ZKSl+TNTnLu2pKU3CRr3UpNrW3Op3jhTFSmJOqLYlmHGXbloqWsZVG9gQOh9Y3LgN2Z5idSisYgl1Jv3mWV6i3ifMxMcbOy7MTLhqmGkiSm1IyvspP1bw8x0MYnI8IpuhyUxLu01T9Xe+rDjibpaB3KfPpfzj6KfyT2DsVcPK\/i2m1d4vUioyDU2wgKKktLQ4UAHwJSq+\/SPoFijFYoqxLssc+ZUm+pslI84os7PzVUeCpx4vEnRlGiB7dYmKThSoVFKVLSJJj+uO8R5D+MW+l4YkaVZaW+a8P8AWum5\/wCUP5mcblWi4e8kaWTqYQLk3MrQpmzTJGvNT3vlB5elMsOqdsVuH7yje3pDlxxDCCtakoQkXKlGwEUPE\/Geh0MLalFGqTQ0ysHuA+atvleMkxRxXr+Jbt\/E\/R8qf9TK3Tceatz+EUps8twuISnP+2d4MXM18xKl+JN4DmW2teE1Kve8EKh5wk46UbBS1HZKASTANUPEdXeSGks0uV6rmBncPokfrFzpuHpeSaQXP5y8N3FiwJ8bRKcxKbBI9ABBjnKCSQ2kC5JO0RIr0tMKUKehdVcTuWbcseqzp16XgFS1SnCBPTaaa2rZqUIzEebh12N9Ak6QyncLNrdT8ChKEJvzJl4kFfgoqOp0tCzSZKntJQt5c2oJylDYypv433iJm3HJgqQhvI0o3DSP86xQeInFnBHCtkuYrxPT6O6BcSinOZMq9GUXX+Fo5Y4kfyklIky5K4Fwu9VHRcJqFcVy27+IZQbn3WPSMAxRhzi12kJ2dx7iukVAUeUlFE1JEgJeWabQk8tCBYXBNhcXOupiZ7OmB8J4h4X1qsYkpi6w5J1NMq3KoUlpOVTKnCpSwCv7hGVJT6xEzvaqVhhl2S4f4XpmEWDdPPlmcr6h5uXLh\/tLPpFISviZxlrVOp7TVVq81VFq+GaQgobfKftKB0SQnqo7dTEJifhXinC9KFZqdMfTRXJlyUaqqRnln3UXzpbc2XYgi400MW7sp8Ev\/aA44Yfwi\/zkUx5an591ggLbl0C6iCdidB7x9AMLdjjCfCvh\/wAXZl+nyU45zw1TXprI+9KshellW7qiCNRaHfGJ7BmHOAvCpTLKKpJyrpmpVuYAKHShQW8lSRpe9x84qPGPty5uJuHqxghp2WYk5P4VZnmAgOKcA0tfVIIRYmMamuNOJcaSlVTU6quRqU458QoINgpSSbpAG3dO3lEbwbnpar8dJOvyMy2+8JUtTaELFwoLQb+5H4+UZX2vpH6P40T6RpYuDT+q+6P0jZsUul\/HdUmDqJpiTmRbqVyrSj+cISjSnApel790HpB3EIDgUtVhrf8AhDSYS2FkthRUDe8MhM81xYW3lXfS53EeDzbtk3F0i9ukNVTi3lZQCLGxubWEKSynHGlAkg308T7wVQQ0tObMpe402gyllANhqnS6jtDR10lJIJG4iTouAMS4nb5lKoc8+1sqY5JS2B4lZ7o9zHUPA+iO4WcoNEnpliYn1UaaLyZZwOIb\/nKVJTnHdJAc1sTa8fOzEUiqRxBV5U7sTbze9tlnSI9SUlN8uYnaxhNaTcAaBOhgjibBIF7C3l0hu4BmAzDzA+cILRfS\/kdYTN02t3iRoLb\/AOdYSULhWpJt03\/ztDZ9KiFBQJsbWHWOsf5PecIqmNZMm2ZiXdA9FLB\/OOw5hPTSGTyAU+A+cRzycpv02hm4g32sfCG6x3tb+xjouUWhRsnQeN7xNyKEEghwFXmIsUizmsbpPoYscjL3A7ukT0nLISMx\/GJmTl89so08TE2w02yjmOHup1JPSIzH+JKbh\/A1aqM3NtNSzUk6srzjWyDt4x8GOKtaxNxY4n4Yl2W3PipBTUtLS7ae4gIVmCkgE+BUonqTHX3G7iBT2sFT023LM06qGVDL7bTYTnuQopJ3Ucw8esNcIcCcNYu4ZS9Qn6SyupzkuHPiVJGdGlxlPS0VjAvGJnh+zPUPFRW99GOZEzLKblTRJCTY6m2x3MW6S4pYXxq24aNVkPuIF1MOdxweqTGc4mnUMzylhKVLvpl12jvDsVVmWo\/Cr42oMqln5qbcUlwp3bSAke1wqNhTKu4pqrs3zQ1JLV3XTqSnyEWyl0+jURIWgp5oGrzuqv8AlD13EMklorZeRM23S0sEj2g7apucUhxJ5DChcoWAVQvK09mUKlNo7ytSo6mDTc9LyDRcmHkMoHVRtFbqOOGC2tEk6wh4\/ZVNLCR\/duCfwjH8eNYtqLZmJ+b+LkjsiWOVsD90b+94oFiPtEadOphNTgJ+z6QQuA6aj2gqlA2N4C+lwdfCHUpS5qaPdbsk\/eVoImZPCzaDeYdzHqlG0SzMtLSgHLaS2fG2sKjMs9xJHmYjKliOmUpRbmJoOPj\/AOHYBcX7gbeptFfnMbVCYzJkpRunt\/7WYs44fRI7o9yY5H4q8SMVyPbFw7h9vEE+ikVSUSy\/I\/EqQw5mllbtiydTrtHZ+F2E0qmONLmkIC0oulpOY6DSx2h7MVRS3rtMjMkAB1zVRA8ukVrFuL6PhGQVU8U16TosoB\/S1CYS0D5JB3PkAY5h4l\/yjfD7CpclMI06dxlOi6UvWMrKg+OZQK1eyR6xhtU4x9pbtIJW3h2TmMNUB05Sulo+BYSn+tNLOY+ff9oPhXsFJuup8QMYKeUEl59ilDMbDVRVMO2HqUpV6xrWGsH8L+FlBaquEsNSFTmDq1NOFM1NLAF8wcduE7XulCfKLvVq1MYnwJi8vFaUTFDezNlxSwopdbsdSdQDHKfY4VLTfDnHMnNNJeSioybiAom6FKamU5hY6nS1jpGLcE+Fq+LPHOiYRSCmXmahaZNwClhCipy1zvlBj6a1LAsjhvGfETECqtTqPT5KkooFKkQ2CqTkg0lSg1YgJUsk3IuYyTtnVDBE32XuG9Hw07KzTCHFqYbacSVtJvdZUgai50uR4xyr2fcZ1bs6cacO4rlZVD6XkusfDOL5aXUqT9kqsctzbXyjZme0jiipymMGanXGKBLVyZM5NyCrPAN5j9UhahcW7ouLGMMdx3T3357ly05U5kFYklyylKbaCiSRbYC5N\/WKfVMRTT0kmVmZyTkEABKrEuPKAOn2L2I03I2iLnsYNulKnHZyqPIN0rmF8lsHxyINzt+0IuvZ4xdOf9LlClkhmWlHluJUxLtJQlR5aiLnc6gbkxZO27KljjHOr0st1+3oXSr\/ABRps5MJcnKFNG1pvDlIdJ6n+ZtpP4oMI805hY3F+m+8ecKJl8nNdSTbvGwhRKSAdLJIvmJ0hhMIKFJOlr2vbeI1xCws3SpBO194BUupKXtQEqAuba+cOMM4dq2I3ly1Kpk3UXc1sss0py3y9o0CicAMRVF0Nzr0pTFDVTKnC++n1aZC1D+0BErPYK4eYHCjibEiZh5NszL0w3L\/APhtc53+8lHtFOrHan4a4GUW8M4fRUHhtMNy6Gv\/ABHues+yW\/aM0xZ22MXYhmkJkmmKXLhVgtKee8keTj5cKT+5ljq7sz4xquKqfRXqvU5uqvy87UJND048p1wJW2w6E5lG9roV6RxTxYkzTuKmL5c93l1aZA8vrFH9Yqigb+atxCRJJNj8ukEWARYjMdddoQIGoPdHgOt4TUCq2u3W+kJkXGU21FwbbQllJ1HW+oMNXUFQVlAI8L7GOj+wLO8jijX5Uq1fpSlW8Sl1H8Y7keBMR7wNrb9IYOd6\/jDNwZVb366w3KQDqR7xvMrUmgnVsi37JiYkqjK7h1aPAWvE7JVRtNrPIPqbROSmIig2TqfG+kWCTxCCE51ZjFhksTtEC6omZTFkmt9uW5zZfc+y1mGZXoIz\/jPwJkOJEsHFVWeo5Da0BiSCOU4pXVaFJN\/a0fPvD3BHEvCLi1iiexKwAhiWWZCqZEhqYbI+6fu2sLp0N45v4g49TVKxU5SWbXMS\/PzlwuXzK6kdB0js\/hHiGWf4X0VCptLSkNBshSgb6eMc8dpPBEnWcUSckmfbEi4hyaASsILrwyhLebpupVri9owOr4fqmEJludzKlJhkBTaGnCtbmv3NSdtxqNItmDuKNXxbX5SQpFMcqsy4oJCFLCM2upUo6JHmSI+qvB3ElJxfwkoa6cyuSaaY+GelV\/bYeR3XEKt1zAm\/W4PWLtIPJlGw004QEjoq0PV1maZbV\/OFqRbYm+kRUlij6EW4qWS2pazdS1jU+8Wmh8Q6lOKSj4JaWzqX8+VAHjrC1f4tJoyeSmaQ\/MEDupA7t9BrGYVTG1SxfVQtEwtsIB75NkpT1VEBOKVbmDO42r7LrhsV+YEHpmI5ylufVOqDZ3bUbpPtDyrSzM9KCpSmVNzZxkKF0n08Ig81xfUQLTa31ZW0FZ8olJXDrrti8Q0PDcxLS1HlZQhQQFq\/aVD3mbBIzH00hKbnWaewp+dmWpRlO63FhI+Ziqz\/ABLkmlKTTJVc+sf6527TXzIzH2FvOIR2rVvEilJXMO8o7sSYLSB6kHMfc28oeU+gIlmQl1aGgD\/Ry4ufntEtLtMs\/wBEwlJ3zr7yvx0\/COHO0zONUXtx8O5yemW5OXV8LzJiZWENoSq6CpSjoB5mNr4i9uzhRw0Y+Gl6o5i6rNICDK0RIW0FAbF42Ra\/7JV6RgNS7XfHrjw45KcM8Jqw3S1EpM7LM81xIPVUy6AhHsB6wxonYpxHjWoprPFHHMxPzjhClsyrqpx4+ILzhCB\/Zzeka3hPhFw04fyj68OYflJioyoF56qNmcduCBu4AgHW\/dR7w+xnxHxHSsdUKnylKFTpSyUzDilkFItoEn7m99APCKTiCVxXifiC5LyT9QTRnae6JiXZfS2hwFK86CpRBUbBNk318IZTXCmoVfhOzRRPKwe6soUWZhAccCAokg2OcaeHXyjQcFzck7hiq0mTnXaqluhTYVOqSAl42QbixP7HW28cw9jh7kSXEWWIvkdp7vyecQf\/AMkYhTsVVDh3xgmKvTJj4Kdkqi8EOj7t1KSfwJjWpjjFN1GcnEYhra5pczdaphleitLWIA8IoqcUsCWd5dP+JHMsiad7jdgbjVVh7RF45xpK15yTHxLbKJTvIblElZzfvGw6dLxVprE3OcW6mW5swo3584svK+R7v4Rb+GHCnGXHB+qimOpYpVHlVTs\/Pzai3KSjY2uEg94kWCUgkmL3SOxVi6o8Mqfi9+bYl11SomSp9MCFKemEJvzHidAlItsbknwi6UfsLVqbr9NpcpT5uoOvhTjrr1mkIQmwza2sCVDxil4m4fucF+1bScPPJlWlSj7AIkzdAC0a69TqbxI9upkf9IzL6Ro4Au\/jmYZX\/ii0SalTmHsAPkgJdwvLJJ\/7Nx5v\/BDqzbaTcX8VEaGPNOh9JAISRoADrC8hITtTfSxLSz80+T3W2GypR8rDWLVJcFcWTzyEvyTNIz7Cpvpbd9mtXD7JMTM9wfwrhMIfxjjJqUWnvFjMiUv6c0h0\/wBloxW6lx+4JYDWtFOkzXplAuFpl1P3P77+VI\/7gxQcX9vSpTTKpSgUCWlZXZInVl9I9GhlZ\/8ACjGMW9o\/iBjNlcvP4hmUyatpWXVymU+jabJHyjO5ubm5shyYdddzbKcJN4J8M8E3LSwCMwJSRp4wkN4+kXZdr\/x2HqGwltpCJSpNEFDKELUHpSYJK1AAqN0jVRJsLRzJ2k5L4DjzjFs90GeLuUdcyUq\/WM6caGUnb01hBYKAVAEm+nn5wBAXqNLeIhFbepsbJAtY6wkWtCfPcwk4k+hJtYwksXAyDw28YbvoVbKLgW1O\/wDnaNp7Es6ZLjzLtE2+Jp8w1r1sAr\/DH0DfvmJhk7e0RryhmNydoZPEBQ10vDYqBOl7RqaZ4I66nbWFUVNSSbXMLNVVVwVuH0iRYrq0kWcHhoYl5CtzrwJaUohP2jm2ESC8YzUsFNtLUp0DVZF0jSMrx5iKuyqvpGQacVPy5LjUw06ruq\/bIB2HpE3wq\/lAmqo2KRjilTAfliGfpeSRcO9My2jYp9U3ESvag4g0TF3ALE0\/R59ipshpJQ6woFxg3F7jdJt0MfK2brktLyy1sKaDillNlJ72o1N\/ONNwFxLapNDXLrmihpDZKAV6ZrRCT2JBiuVmDM1hs2TfItGunQExsXDHEXDBjDM3NSiUKqIYLATOpu6klNrBVtRGw9kbgPjQ4Fq+IuHtLlKTNVkKkm6tUnMrjbYV3uXa\/dJ621y9I6y7PXZBY4L4VfkJzEU5WJ+emFTk64okNc1W6WkE2QkeWp3MazMcNaZytEBNh9rMRFGrNGk2FOsU9Tj7oOQr3aH9rr7RV6nK0jB6G1zr3x86RmDV7IT5nxijYx4qzT86JWUWcqLgNt7DUAafvH8IhpGVqE3WVfFOlrV5V19cibD8yfeHtOm0yku8yomzywFW6oHSMU7V3arnOAlLpztKkWpuu1JSksGZSFNMNp8vlFd7M\/bMc491B\/DuJqdK07ErTZdl5mTTkbmEjdKk9CI6SkZ8NKIdUUsrBSrLE7I0eXKEuqJcChcDpEm0GmO62kD0EKpKlb90eJiBxnjal4BwrUMQ1MvO0+QSFPKl0ZyLqCfS9z4xUZXi4rG0s3M4bUGaa8y263MON5nl5gTYJ2FveE\/9GZ2oTAmZ91alb82cWVLH7qdx7ACJJiiyUmlNmzNLB0U7okf2R+ph+XFplluOrSxKti6lKIbbQPEnQCMW4kdtLhPwyDrC69\/pJUm7j4KhAP6+BduGx7KJ8o56q\/bi4u8YJt2mcKsEmlMnT4ptgzswkeKnFANo\/u6eMZ\/jvs2cQ6ng3FPEXiViNiq1SQkw4JRyeVMzKSVAJuUgoCRr3QrTyjYuzlw7wJS+EGGKw\/hyTq1YqMk5UX5ioNB\/KoPrbCW0qGVI06hUX6pcUX14PfmpVn4adlA5y5ZVlMpSkd0C2iemiUiK1hvEFfxNi+mTdSedbqM1SQn4RptTLKkhZAWTfUq8ett4d4FlcV02Vrhxg5JturcU3KMtuhS+WCnKABraydL6w14vY4l8GYWnMXyaZpdQkUJ5AUsJQHFWQDbW4ub3PyiL4i8SUDifw4kKDNcul1WireqDQZyH4zIFqIVbNoTbfqYp\/FTiFXeHlXp70s5zaXPZ05HACULTqRfexBuL7WML9k7GTNbl5ymWyOopFQRbxHLWoflGZ9ldfw2KeKMmfuyKXbeaJ9n\/AMxjFOJyqcniTiVSJd9LPx7pS0VgHfW5t436bQxoEjWa\/VJGnUGjKenp90S0qhhkuOPLJ+ykqvrr0tGo4V7I3EzHmM8QUCZalqfM0CWXNVKbqE2FsSwSkEoK0Zrq1tlHUHwjX+HnYlkjwUpVarrC53FWLZ8SVLQlS+VIS6VWcmClJGZW9r90CxsdY68wT2b8LU7jdKUyXoVPbwtgqj3kZZ9lKviZlaLKmXQR9YonN3lXtlFoh8FUvCGGeDOOqZMVKVRO1OeW9MtSyUoUpxC+6jQ\/ZJF9vvRW8XdrDC1Aw\/gim0KlFUzh\/kuzAWnupC7Zsp162JvGQcZu2TiisldVkGHKLVGJr4NpxrYN6nW\/iRuI5mVieexjxvw1XanPfSNUm59j4hSlZjcLAFz4Wt8o1Pt5pbmcSUqdbabZS9Lyqw20LISDKtiw8u5Fw4VcPK3j\/hBw4qlL+DUwzT5qnOOPzbbZbcROPKAIUb\/ZWLWBi\/zHBqk4Sl0vYyxdKU1OhLSMrN\/7T6m7\/wBhKorVV428C+HgUiXQvEk0n\/ZtuTVz5lfIQP7qx6xn2Jf5QGoIZclcKYWlqZLnRKpt7T\/umEsoPooLjGsUdp7iTipt1l3E0zT5RzRUrSgmTaI8CloJCve8Zm5NTc++c7rr7rqtiSorJ\/MxeuG\/AnF\/EziZSsDSVLfkK3PkKyVBpTIZay5y6sEXCQnW9tdLbxNt9nqcqnG+V4e0OpJq\/OqYpoqSWS2lRSrK64lJJORJzWN9QPON1qXYmokr2tcO8M6QupTtDbDb9Tm51aSpxtCQ48RlSkJBuEDexvrGhY+4IN8Ru25hzCs5JSVKwrTmUOyNOQEts\/BsmxCUp0JcWCddSCPCKT\/KWyspSeOFKal3kKl3ZFMuZNtnlol0IAygdCCVqOnjHFE7LmVmnGzplOnpHdHZAqZXSUpuO45SnvfOpk\/\/AJYzXtgyRlOP+IdLB5uXdvbxaSD+UY8SFe+toTsNuo8oKsEpyAgC3yP+bQkog6dNoRWSoCxAAP8AkQnyitRAso2uQTvATLTaWEFS\/rFHvJT0iPWCDoQFHQxpnZUnPo\/tA4VJskOuPMnS32mVj84+j8ysZjpcxGvO3TES89v4RHvzHXUiEeeNbkDyJjfHZKXeAzsoUBra0IuyEsRblC\/lDVylMEbFPoYSVS5eWaU+8+US7YuoqNvaKFj3i\/LYYphXLOIbCE5i0hYKlfLe8YhN8Z6xiJxIZmphHMTqy0bJBN7Anx089oj3+0PJYck2abUXZmangC2halhxKL7A+I6RX8X12QxE3LuVRS6FOrWDLz4SeSoAaXP8fn0ik4zmqnOtTOSorkK4EqbK0OK5U+1bZQvY3G2n\/LE3Pi32y64AgJUU22I8dIsOEXZWZmAzPOFLSu7YnQ+USeKUSco6hulPFtz77StQY23sF9mme4\/cWVfSbDzGFqeUvzr9jZ6yh9SlXiep6CPuFT5CmYUokvJyrbFNpkm0G2m0gIbaQkWAHQACKbM8ZqVOVRVMoI+lplAJdfScsuyPFS+voIrWIceNvP5Jqd+KIN+U33Wh5W6+8ZjW+IUzWa\/K01lzlsZ7rSjTQdBEHVW0Vibl5ufmw2284k5QdUpLlv0EZ1iTGFKwy7OLklofUlC1J6qJbdB39IhK3xSm6vKuVSXdS1LSVQU3M945uW8kBCvS5tF3oFTTUqLLTAUL5dSlVzcaRhnad4FznHCbw2iVfQJyWmkIyuDRbZIz\/hr7RccFdnbGuB+JX0uirYbcw8zmEtJsUdluYaTayUczJe1tzfWNZmpOtEqS7KMKSQe8wnKfzi2U+YbkqTKmcdDCggApWdb+m8ITWJcqD8JL3T\/tH+6PZO597RETnx9YTZ1xxxo7i\/La+XX3vFB7SS5tzs94yl3Z5TzcvTTymGxZtGVQPuYrfY\/nnX+A+FloA5qpRCStKe8bFQ336RZ+JPaJ4dcJ23P9IsUSjU8ka0+UV8RNE+GRF8v9q0cy4m\/lAcR40qCqRwmwFMTsyvuonJ9pUy76hlvup\/tKUPKKlOcDeN3HKaS9xNxuaRJq7\/0c4\/zVJHlLNENo\/tFMXui9mnhLwfpKa1WpZyvBr7U9V1cxq97aMosncjQlcXqY4myMth3EYkJdumUqlJAlXkoS20tJSCFpZyhI3sBlvDfHtdbxp2asZzrMwubbmKMXEuK0JstB26b7RVez3NMucEcBJfdLSXKZOS4UElWonFnpEhgSbwxR6ZxJojLhmZ\/D0oaow7PrCDNTLiSQ3YaqAAToD4whUOKtZPCymYolwlTapFt96SubIbIBVZQ17tybHoIzXDXGZmrY+dEyr4ZublWkIbC7thwKVrY7XBHyiQ451pFe4U1ySpctPVdJKQt+Qk3HmWVIcSpSVOJSUpsAdzGR8JV4\/wC0dxgwpQaG1KSdTkm1stvKZVy5ZjJZx1w3ubDX1sOsdqVbsv8ABriRUv8AQFPFWpVLiPJtLWhRcT8PzgmyrNBASQNQQFZt9Yx3h1wNnuztxsptDqE6Ki3VKfUES822mw5iZdxLrSh+0lXzBB6xj\/Z4JleMvEyUScpcpU9a\/imYaX\/hjOa\/gp3GvaGew0w4llyp1dMuHFmwTnULm\/oTH1dp+BcH8POK+GwHZKVo2EqF8NSJdNsqHVpyrfJ6qJOp1MZTQuMeBuFGDeIUjUS5OVevTDkuy8EZufcHVZuNLkn3is1btWTMxhLDTdNlG5VyjoQtt8gBOZDdlJ0toQVdN4yniJ2j8R1TED9SmK+5hyempUyUyuQmLhSAARfKRoSCP7UYwOKk\/NVF5FC+Pq8w6lKAiVZKi4QdbpsbX02H3YvlI4H8ZuKLCHJDDqqJS3U2ceqq+QlNzfrt62jRZbsKMS0qif4l8SUsS5SkONShS22q3\/1XCEHbprD+n1Dsx8FplDNKYGI6vfIh8qU8vMdPtqKEpv4oUqM57dFRlcT0\/DVfkpdcpLz0owpMqrLZjIXmigZQNByx56xB8OsW1qg9nOiM0KddpfMn6sJlUmS0t9aG5dTYcWmylJAWqySbaxzXU6\/Uau+t2cnHphxZupS1HUw8ncC4ip2HZKvzdDqEtRJ5Zblag9LLSw+obhCyLK9o3Ko9iDGElSuGjbMy3NYnxyFPS1FDRR8IwEhQW66TYEgg5baWOukS\/CjsU1uqdpukcNsSBLiZVwzVY+EUShEujVQSrrm+zcePlHTMn2d6XW+3\/T5X6KouHqDh2TaqclS5NtttLrTaylq6UDvLK0lZza6iL5STg3Cfa+4n4mq2I35erolGWg4hLSUoaWLKSCu5uOWgXAjnjBvErCfDvttPVmnMNu4aalpj4XMvPc8sHMFEbk5tfOJLiX2pZlztC0zFVLbSWpmmzMk0zcqCDmCwLgAm5A3EZRifj5iY8aKFjJMwmlVViXXKNKCcqU3uU3B6XPURnHHnE9Wx3VpOu1+sN1qpzWbM6hdyLgZdLC1sttoy2sgTDLE0ndQyr8j\/AJvHW3ZDmlNSjyVJKD9Hy7wBO4bn2FX+V4Q7cUiZfjamZtb4imMHQdQpQv8AhGADoLWJuDBVAgEHU7WAhMm2h22tvCK0gqAHU6wmEkG9tIIhRSAoDX1hFSQdSLq8zvCDydrAXAzRauC039GcZsGTBGiamykk\/wBZQT+sfTKbdAUbnziKmJlI6+kQk5NhKla6RFTU9b\/nDNVQIOwUTveOnraW+cEWkk6aGEi0SbRmPGrHLFAokxLhSVJlwOckE3uo6a+UcS4yxY9Wa4XBNuplQrKsJOgtbr4esQD+IZhcw4y28p0KGRRQL5iNvb084FyT+Oab+Lk+T3QELKQm5\/ag4nKpR6W4y+TWqQs2XKPrvywOqT90+YhpT8SspIlagFTVHmFXQ+qxck3LaWI6fmIrWNKWZCqvKStDzF7hbf2VAjRQt4xW23xLqTkJGuY6XjTuAdDk8dcW8KSVbyoo7lQYROrUq12ysXGny94+zLPEbDfC+mLpmA6HLoATkS6Ww0yn0QkC+vpGVYtxxifF8wVVaddmGt0soOVpPokaRPYfnjQ8BFLZDU5PPKLi72ISnRIiKl6iwXmi9NIdU4sXsdAcwiqVLFbMnxJGRAbbFkeGpFifnGfY6xrNSrvL5q0hoqZy+BCsyYyPEmK3FVBbzOZbaSXikC90LFlj2iMoOOJOgCekKxNJTTqgwJaZVnvmZJ+rcSPFMadwo4z0tSDS1PN86SJYe5abJeQNEPoHUEbx0DgGWZq06ut50rk2QW5dfRSjuoflD7EPFSgYexIzh+bedNYflvi2ZdDR+sRcp0UdNxCi6xUKg5kYRyEndLIzr+Z0\/CFJakKSouPuZVnc35jh9TEoxJobIDbAJGy3O8f4RX8d4\/wvw6kjN4sxJIURkC6BNvgLX+6j7SvQCOSuNXbPw9xIwzXMC8OsM1nF89VpZUp8algttoCvvJRYrV7hIjD+HNI4v4xn3eEkzi1\/CdIw7Jc2blkK5Cmm1KByKyJC3DdwDKSRGq4R7JfDDCTs3N1yfmcXTskA7MpmXQ0y35qbQcx2O6\/aLzKcVcPUilzrGFafLSGHmKeJpuZlGBLtFWoKSgJGbLb71733iwcNcZS+LZWiVZL6pyXc5mriLE2tcZRttsIq1co8pN4LxI\/iWsSSpViqJbm103Pkly6sKZbupIObUXNrecLzVZwRg6en5R9pzLVEpSJaqq5rRQlIR3dNvHW+sFYxPL404KY+bl25eWZapMyyiVlQAhCUOpSCAANwAYovAWqy8p2f8EzMzMsyzbLtQZK3lAD+nCra2170Y7jji63hTiJjJyQYTOJq7TLSVOrUgIs3lJta53MdO8A+CdUxZwHpE1j\/ABRTuHlBqMoJOmtvhKpmaby5Q4c6khIVuBYm2txFfp\/8ntiai8bKJJTVVlqphpM00XJuWOR1UnlK+blPiUlBsTYkHYgxu9Q7QuJMK4zRLYaodJkeG9Gc+EMkJdSpiaTfLdGUWSSdbk7W0VeLVR6DI0Di3xj+jZCXarkpQmJmRnZdGR1MtNBRU2bfayONKKSdQFW2AjnjhcKLV+K+EqbSKdMymJZbECTMN8uyw225mUVqSemRRsVEfa01tE3xoxtLVjtI4alWXErti6flmCnqgSQad\/8AEJ9xHK\/BGXLvagxXIIW22qckKo0lTqsqASypYJPQabxnHE+oP4O47ztQDxRNSUy2+lyVWFALShJBSoaEXG46GNIr\/aAVWqyl0VacnHXmi2RLXU4Eq1sCR422h\/QMM8SOJVNUjD2Cqgt8Ect2otWFv28yra2i\/wBB7D3ESsUnLjLGNPw3SQc7zLKuaU7nU6BPXeH3\/RT2YuEbpexLit3GFQQAFNIf5wUR\/wBldB9FEQhUO3jw9wBKKkuHfDyXYSjRDzzaGAfMpTmPuFiMcx328uKGMHFiVqTVDZVsKe3lWB5OG7n+9GVyLeO+MU7V5luYn689T5J2ozzzz+blMNi6lqUo\/huTtFQk5SZed5rTS18qy1KA2F946z7V0rLK4O4IflytSRLrSStQPeDxUbWA0u7oPOIXgXWXmuz5OSIdKZV6uTcu8gAHOlcqytI\/vM308I56wPQJfEmOqHRpuZ+DlZ2fZlXZgpvy0qWElVvK8fVLtf4IwTJo4R4cW+XKZh9+XzSLakhpUqLBV7G4JKE6\/wAYt3\/tRYFq\/GvAjFNl5JDTS10tMy4kFbOZBCQFnQC4tb8Y5d4ndpGv4R7Z2LKzRJhp2eEs7SJJxlQdbdbSVLTrqL6j01jI6vxfxO9xblq\/MVcUipz0g7IrmmXtWyDnAukkgkkfOKdV63LLxA\/Pz0\/M1P4tj4eZIVldWoKzJUSq+l7biKzP4uYl65J1CSZ+HXKHRh5fMJTaygb+RhhUcaVjEymEtcx+ZlXC4yZWXCMvjYJHW34RDTErPhxZmCiWGbN\/OXkhaT+6Tfr4QgsyuRKZmpqcCSSluVZJt5XVlt+MFRWZSWlyyzIc5Ga\/86dKvwTljpPsnVV2fnJ0ryt3pM8hKECyRkTzQB6ZIu3bzlAMf4ZmyDlfphTe19UuE\/4o5mO5GoHzhM2IIVvawt4QmpIAtc3J6nxhLRIJAOvlBD3bWsfGElJGt9ddLDQwnkuTc3PT+EN3QSqwPXU7dYcYWnDTcYUCa25NQl3L+FnEmPp5OTQuSVWv0iHfnU3OvtEJPToBUNSR1iGmZ6+y7QycqBCzrbyEdh8ggXteCcsnW0Np+aRTZGZm3AeXLtqcNh0AvHz440cRm8Qzs46h1ZDr+ZbZVe5tv+JjIHVOuc9bas6VXvmGgA2v5wmhawgFhH29FjUb72MSHMD3J57zkwsAgJUVEJP8ImZGoJl25eXQoSjWpJUDlUb9R\/GK9XZaZw1PHnyoek5nvEIAKFJO+o+flCVQlpdUm202S4hKC5L33LZPeSfEp3+cQshQJEz7bVQmlSsq+CUTQScv7qh09Y33gIKZwmxWxU6xTl1iluWQmYlzzPhrnRxNtCPEbx3rT6jJ1+nsT9MmG52RfTnbfZVmSofx8ocTTDEshsPOJQ6s2SknUnwAhzUalJSdIZlKhUJdrlg2Te603308Yzl2s0FL7\/w78yoIUFXzAAm99P4RD4lfk8dVNuaos40Jtod5l05M9juD6xSuLtPnTSxVXZRwTKUhEyy2CSlY2X6HxjBWqtN1uptyrziqQgnRwNlRBIttvlJh5IcAMRzk6n4xbMpK3umYnF7A\/soGqkn06xsmCuBWHqFMS72SYq062QpsukobQeoShJuQfBR9o3Kj0qpSy5Ul8STUuQplrQJb80tjSMQx\/Ovr7a\/D9E3MOzyZmluJcU8ft2Kzt4aR1tXKtTMM05ycqk\/KUinN\/bemnUstpHmSQI51xz28uHOGZpdNwpLVDiBWr5W5ekskMlX\/AGhGo\/dSqMnxVxj7QHE\/M27N0vhDRHRo0kkz6kegCnb+YCIochwXwhK1H6RxA\/Vse1ZZzOTdZfUyypXjy0qLih+8sekXhuprpkkJKmtS9JpmqVSVMZTLNkeYTYq\/tExBYPdydpfiEP8Ab4fl3r33sqXP8Yv83h2kSVfxrUpoFiabpi519aFFZ+GSbLVlF7bj1iDkavgnAiJB34ctNTdPSw0qYUt1h9OqlHLsFHNc6jTpBMNcQJfE9PeRKIYlmJF9yXbl5RIQhKAbBQAAtcaxiuJeIkrQ8O8Q8PTVVW27UK4zNolAlSg7yyghRsNLAGNmxrwsxpxf4cU\/FGGsI1Sdpssr4lE4tAb5rOUhRbbUQtQOh0TrbrFB7P7NWpTXEGj1GWmJNqYw\/MTCGphJSoglKgqx6Ea3jYexxO0XBnZTaxtWaMzWajSq1Ny9HYnE3aDqw2S4Qf2bHX1tGvY+qlP4x8K67VMV4coT9dw3PSbUyppiyX5N9SEqQSq621JzqsQdFISoaG0QPapdpeH+J1JRUnyxTpCRMpTZYAlDAS0kpWACBcqVbW4PrYjScD43bwy9wIp1XfEs7OYcmGn1vKKSltRZRL5sxuCSQADrcxQcS9mDGC+MbLLDi\/8AQWorRNVWbmHAW5ZDJuUjMT9oAWygWzEn7MZRN9tql8P+1ViTGSJNddwvXXW8Pplpexccp8ukpMwgE2N3lEpvukK1G8a3jXtE8G8B\/TGN8CUt+Zx\/UZP4VuZflHQZdJG4SRYnQXt9ogXMYX2S+DlW48Y7PEasVSYoVLwhMBUnJzLN1zbiwtbji1kgAqVcqOu4HSMe4TSyHO2g3T1OFpFRVNyfNQgry82TcSFWT0uofjG6Yr4PcA6LVPpzH+KZeqVVttLbsvTnVKbcKBYE5M1jYAd620QUx2wOCvCxlUvgDh2zNPIuEzD7TbV9N72Xm\/3YzbGn8olxEr7S5ejokcPSx0SmUYCiB6rKrHzTaMFxbxaxjj+aDtbr9Qqbh250wty3oCTb2iWonAPH2LMFM4tk6O6\/SpuoIpcopawHZyZUbZGkE5l2O5AsLHXSNN4C9inFHETjCMKYql3cPSVOUpyr6gvNtoSFKSm1xmN0i\/TN1tGu8J+wlQeIVS4kYhn5Gp0rCtFS6ijybb\/fecSDYLcUDfYZrW1PQRtvD\/sy4VoPZQU2p2RpLmJ55C52pJTncVLAjKznGoFwCU3tqYzH+UmoGHMCztFlcOSkvIPMUppgsyzKW0JaBISbAC98ytfERkfGpZqvZjwlMmxU26+kka6FEqv9TFX4AzA\/6H6+kpz\/AA2IZNVr7cyWmU\/4BGCPTDlLxE4+0rlvS80VoKehSu4\/KN34jcaGalSEqpU+8qfXkeUVyyEggi5SSbqO\/jFNxHxfYng0qn0qUk3UuJmA4jMtanPvA5tOp\/CK3MVLEOJagxOTK5j6lwFucfs0lsfvGw\/GGTzbSVkzNUlZcheezJU+sG\/QgZfmqG\/x9LZLh\/n8+tQtda0sJ90jMfDqIGVrByPKkqOwkoTmLiGi8Uf1iV5re1oCZGI56kpn3m6gulLc5KZgoWJfPqcgP2b6HQeBgj+Dq1LNtuTEg6w2tGdK3BYEXtf5xr2OeyHibhrhDC1axBPSTL+IZf4pimMqWqYZbITl5gy2CiVpFgd\/SND7UvZUw52e+FuCZiVbqE3iSoSoeqM84oqZ5ywmzSE2ASB3z1JtvFY7IpCcTSzJXczDU8wEn+vKOD8417tws\/GU7h3U0i4elHU366pbV+pjlfLdNiL\/AKmCFJSL220FoSPiAQPCErZRYb7mw1t0gFEadDfU2hJQUBa3etrCS0lQO1+toSdRdOW3nfaGLq1NPMuovdCgoHxsbx9J2p4zcjLPA3DrSF733AMMZh83IOpvtEROPXUb6dd4iH12vfQHw0hoXTeO5g0PWCqYGthYRT+LUyKVw6rbwUEOKZLaFWuAo6a+UfMbG55s25cILyVqB5ex13v1ity7qsjg1sq1xfeHLbLvL+sWVADS24hxI5w8lXNLZtYZhcX6A+sSRQ1PhCXj9fte9ki529rRY6ZNqn6Uqi1VpqZp4VkQ4ohDjV9Db0MQk\/hCdwjUG5eYaL0shXPkphOqHU\/eSD5jp5QM3RpDEFNqFPlUqz2MzLBP+rIFwnzuLj5RHcHa3iNrFEtR6O27OOzKw2uUSjOCOtweg8Y794c0lPCOkKW\/Pq+JmUXck0atBZ2IGuvS4iCqnF+m0XEKnZmYcm5whWUKVdKEgXJt0tGJYr49sVczU2+6szClqLLCNLHpmt4eW8VdHEWeVRWEsTTnNSu6lX++q5AA8hc+8PKbxRckKrKS7byUrlwVvqBGgHeIJ8c1o1\/hz2g28RvzEnMNtvANL5S3wAHHPuoN\/E2ET0rTp2tzzs5T6DLUMrIBcQkZk6C45ivO+1ot9HwfLS6g5PTK517dSUaA+qjqYuDczK0OQXMOKlqXJNput91YbQkea1H9Yzyp9pTDAmXpXC8pUMeVFvQporP82Qf68yuzYHmCY5v4\/wBZ4pN4hpPFxpuiYcTSJVTEkKdNpnVt5lKSQskZVK7x2FhpCdUwHJV+dZnsf4gr\/EGtKbbfWibmzLyTedAXlSkXUQM1tMgi00qbaoMmmUoclJ4elVaFulMhkqH9ZzVavdRhNDgK3LAqzHXzPnCjtiyCTbW3pCUwhV0k9Ovjp4RGYTIX2ksR9PiMIhQudyA1\/CKdxcx7TcP41rpmZlwmdw+ZFDUv3ruFau6egHXWNMpHAbiHxs4G0Ws4fwwtEvKtNOsOTjyWXZlKEWUWkHVQOtibX6Xhp2YOCTVOpWLMc8TEVSm4PkHEpZpssVNPVF437gIsco2NiNeosY1Kr9lDg1xGrOAcXYfpNVo1MxBVUyU1IvPqUWH0guZHG1lRAWlCgbHS6T1IjSXOKNdZ48MfCYgflaNTJ5mmM4bllhMuWFq5YC28pJNrKBFrW8IpXG96nPcTqgzKpQ3MyTVZpByCxUzyEPpT5hClKA8LxFdjShu8VOxnUcO0hDa6zR6rPMJbdNgVuIzIUd9s4+UMeLTdb4OdmioUHFs03J46xzPsN\/CpcSpUlJS6kqW4op0ASEnUaXWOt4tmEe1XwQ7QeBqZMcSKPMNV2lrSlTa5ZxxD60XAWhaBspO6VW3I1ABjmLtIcVMT9oPiDPSeGKPPCXf5MpLplmVWYlWlZm2kWHVXeJtuABoI2yUXxlq3DtrCmLcXrkKY\/LCWcRXH20zDqNNCLc5foQb9bxR6R2WcFYRqArGJa7P1OYtdAfDdOYAFrBKn7LI8MjR8ouYxfgfDqeXR6RLzTyRopiTVNLB83prIgeqWVRGVPi7VK6+w0umSxlG3EqDVSeXOCwPRsctkf92Yx2qSTFF7f6JeWaRKyrlWLLbbQypQlxooAA6CyrRjHHOjzNKxclh9QJ+FQQG1ZkdRp8osVB7IOPazhPC2IlyzEpIYlcLdNbW5d9wAgZyjcJJIsetxHY2Cf5PPB9K454Ww9OSUxVabRqWZ+vTE48eXPzNhZCUaZWwojbfW5ix4J7MuA6Jww4nYvnaVSDNzE683J8hsLbkEIeKSlobDwB9I0ej4w4XYbmuElFZqEs\/K0p5QZWt5KW2nXE5g8u2guQdSfvRUGu0XRKVxbxtO0ISapWeJZ+NbWChXLUoLWFJJzBRA2Ot4yXCPahrOHsNcSsLMyAqKapPqdZf+J5XwiXrqKwCklYvbTSMXnuIeJqlhlGHX8RTX0HTXXkopynTywVg5V5fHU28IqHE7ipN8RVJfxRiCYqs2uWblUGZygsoAsEJCQNASdTr1i8YyljMdlGXaKgtUjOhBUOuaWQf\/ANUUrs4u5uHvEBnQlqapU0ATYaOPI\/8A2CMmxamknF9WUkzLMtziQhIStRV97W4sL3todIi\/pGQaty6cXyNAZt4qHyTliTp8vieepM9VqVSppNLk7CZnpCSIbYubALdSnu3uNz1hObwVik0uWrM7SKk3TpzMpioTTDgaetvlcIsr2Ma7irslVLA3ATD2Pa3UlS9axFMBFNoCWBm5H+1cWVaXumyQk\/bTrvbRuNHYtp3Cjh3gCVYXPz2P68GnJ1IVdlKnCAhhtIG4ubkknudLxfeOHCeo8JOzvhTh3I4bbw4zXK2xKz1afmU86pPKJILgAGVsW0BP3RHRHGTglhqVp\/B7AczVGadQ6Vy2jJsy5cDzik95wna5KFDU3+sJjNf5QteG5PiDgRqaU6pSJiWYmpZS0hluUCydEjUXudfAROdrTj9hEP0GZordOrDFOqEjmmUy4cytod77YUrSwuOhEc\/9svtLvcd6LISLUgabTqVNIUlJcJzkBScxA0+94RmnZJl5ia4l4caZ5NkToW6X3kNgoN0qsVEXPe0SLk9BG2drVsznBfhtN9WyGSfVkf8AljlDKdB7iCOEADy94SUnONLp6GEVJuNdBsIHIFXvqPCE1oSUgECET5iwv0hFxBIVp3QbkwxnWc7C7Aiw9r+EfQHCE\/8AG4IoExf+lkGFE+JLYhaYeFzrrtETOzI3FhfW8Q8zOAi97esMDNhOmh8zH0ByW6Wgqki+0ZB2pp0SnCadZS7ynnlAo7+UnLr7x82qlPIffWoXSoqIIVEVmLawo6pPy1h42bDMkAEfe8oXW4hKTdKkOAgFQOlvWHMkSslPcU+o91ZV9keUWemMyDjaVTjrSXUkrWoHW\/Q29YuiFTqmG5N2UTVqUr7S5d3M+3uQsJ6EfjtGZzDy8KYrZQpaXA2staJ\/pQCChQ9iN\/COseE+FqTheUfqVOp4kZipfXvTK9XTfXKP2U38IrHFHiiuSnJhKXkc8NktMGxISPvEeZ8Y5xXieeqc3OTbzpcdUhSM5Nkgm4sPLeIqUllz9RUy2blsi7hG6j1t0AHTpE5P1VmkyhbkyHXWEEKmSdG7nW3io\/gBFacmDSKWqZdKufNWAR1CSbi\/7x\/AecS9Erj1Ll2kFeTlq5qlDqu+l\/b847elsb02k8L8J4jmHp2p\/SsshAladKLmHTMBPeQANBte6iOvhFYmeKOMa5NNyVIkafhAOkpQqfP0hUlaX0Yb7jZt+2qKvUsN06ooRUsSVR7EtTz\/AFbWIJgzRBvbuSjRDTfkFExcFIkZKQZlpxEujKqyJecOdYHTJKMjIn3EZ9xtvMdnCbNiA268BdGS1nQfs9PSGEw8FqpUxuX6RIuAerCAfygzbxy6kBINwIVbWEuBOwG5hy04HCArUJOl4I46ChVrWvoYf8EsJyuLO1NMrqiVGgs4PdeqSkkpPJFgRca3OgjZ1YM4RccsMVB6a4X0+iv0Z1hLKpY8tx6UedDYcKkgHMDrrfbfUxNY\/qc5QePWGZaVqM3R5elTUrJUinMuqQw\/LghtxAQk97Q63GxB6RasRUh7G2A8d0\/DS2nJuh4nnilhaM6A4UhzJl8CVqHkTGP4u4kzfBThzwuma3JimVWfxScRzNNCcq5eQZbKXCUmxBIItfqq0aZiCq8G6lXqTxllcY06psU6VXMy1KlnEF2ZeN1N503zJKCoixAt12jmLhqcT8cMf17GUkwmYw3RZWqPVSeLgCVzc00tRS2k6qCRkSD4CJHsFSSn+GGPZmTxc5hSqUmsOPMFEy2yl8rYACHM+hTmQOnSHnEPgdNcV5tNc4g40dmHC2lp2YlEDKtA1CS+5y2rC\/RRHlCdJpHCjhvIplqbKt1ZxvQKcLk+on0Rymf\/ABFe8O5njFMsyypej0sybB0CHXhLtkebMqG\/kpxUVd7GdemXFoZqZkEOfaRTEJlAfVTYC1f2lGItEqwy6pyynHF6qcVqpR8STvCxbbQVOlITpoOkNVFaFFWviEkxU+Jz\/wAD258NzeXKl+qUx6\/iFlu\/5xm\/aJYyYzQFqzqS2tuyumVxQtHduKe0thmZoPDBvCIaW5RKdKzC2SyUBgpCfq9hcemmkZ3iftJ4pm8fJrqqhN0Z6fllSzzjKOUVNhYUQgX0BF9fKM+pvGWsSWH8Y4SbxAiWo87O\/GTDbzSXFTCCb91ZBKe8Lm0ZfUcc0pujNByYK6jLOJ5QKipC0pOgsPIb+sMf9P5hM0fo2lLkZIpIur6lux3GYmI2ax46iSeE1OysjMqV3XJMF5wp0BBIFjsCO90ioT9fkl3UuZnaitZutRUlgHwvbNEWnErks5eSkZNpQN86mQ8oe7mbX0tHQ1DqD1d7K+IfiFqcebnpZaibdWppP6CKf2cptX0FxHl1Eqy0eWmUpP8AUnmL\/wDGYyrGbYTimphNxd0kaeMdXYC7BE5M9mZ3irW6i9Sp9SRNSdOUlKEuS5IstThV3QUlR0F9o6Z4scJKBw77LODcGS1ap0ohyfl5mq\/B\/W\/FTDywM5UNwm5Iv+ynwgvbylMI4Yp3CzDb5fclJGZk5UyKHkttGTvZaiBdVzy0i9or3ai49YUqp4bCk06UckMPVyVF3GCsBhNwpAUrQg2Tpl3SDeI\/jv2kneKFYw7Uk0yYpstQZ5uflc10IUpNwSCN9CetoxDtQcfK1xJk6YJt5La6RUUTTQbUMwKbgHxil4x42VCr1hucXV3H35aZam2VpvdJSQoG+mtjELxk4zP8V64zUqrUZ2o1ltsNrmpoICClKiUBKUAWHeJ26xUK9j5NTl5hh4WcfF1ZE6BWh3uTuIQTO4mxajlS8jOTrDgssNMHLfzVbx8TGw9nHAlfp2NcPtrpzj08qqSzoYY+sLSEuAqUopuE6eJ6Rufagkku9nqjWt\/MKypn2St5v9I49AB0184IpAJv+JghNtD02hEp7xubA7kwAFyCbH0gikixF9PHqYTW0B3SLgaQkpIKRcDWGUy1mQQTt4R2ZwsqnP4XYZOe4RJIbuD1Tp+kS0zOm29vGx1iLnZ8XtfUDr4RCTU4pZsCE28DE\/T8AVepybU0hxhtDguEuE3t7CO9iQQTmEFIAuNI5c7cVcFNotCZyKVm5qwBtukf59Y4SnwzNPqcIKVK1uTvDBaChBBGpItCkq6SLak3GnlDzvIUShRsTchQ2tC6ihCkAvZrJN7a2\/hDkFJcANk6A3Pe\/HpFiw\/XVUmpNTgaU8hlWdTSTYovoTY7xtOFaxS8VS5aq9Gk5qYzXbeEunM3tY5t7xZcZY7k+HWGZp4uBxwIyo018gI5AreKZuoqmahMr\/nk+srWOoT91I8odty6ZGlyYUgkhHNWkjdxQuAfIC3uYbfGmSQUNXEzMHO64TbKDsPf8hFpoOH5acp8vMTRtJBdmmSO9Mu+J\/qj8vaG09gyUxezXJ6brSaXTqIQEKbZ5ip2aWbJQkXFhpa\/QDaISr0lumS0zLNL57pcskjXwtHWfAqdeVwVqdEmmfiHZZhUxLS7hUkKUnXKcpBsfDrFeki0iZkXZwtysi\/MB1xh0ciXbJQb6DvEfn7wxw7RZSbqNUdnJRLyZZ6yFOzYlZZA3BJV3jpsBFsqGL6FISqUs1oLKBkMtQpcS4UfAvKu4r1EVbicS\/2b6lzmXJTnvTC2W5gkrIzJO51VvvEAw8XsP4MdCv6agSdz6JKf0hZD4N0lQFt7wu26EK2zDz6wsl7ukm2Y6a9ISUpSsxJIAsdI0TsvVeQV2iZ\/D00pLb+IMKLlZdStlLbdzlPunMfaLpxKo1d4F8M8XCvVOUTVsRVBmVpLbB0Yk23ue44odAkXHrbxh832teDWPqdS+IJlp1zHlJl3fhKXMMuBpEwU2zlQBSoXAIO\/leOfOC\/FHjZS8fz9UwPJTr9NnlOPVEzsmVS04+4srW8q9gg3VYG4skARZuJnBfGHHrEz9cxbimTlJhxtDT7cglU0pDaTcIAbuhtN9bFQF9TEejCHBrhTT0sVmst1FxoWLT00lRJ8OVL57einExMcPe0zg+oYmk+H2EqS\/TaVWnuTNfDZZRL10lN1BOdajbxc9owvgDNzVAHFp2nvmSn2HpVKJtsAPNoMytKwhdrpvcXIsYuzueqvmYnn3ahMHd6ZWp1Z\/tKJMKNy+RRNgkAAW84PzSUWc26W0gybFaQgDNbe8KOzOVIF9CNhDYvKWsgnNqdD0jylHlrAFgds0UXj7MGQ7S2Cakq4+po8xm22DRiD7TtJfRjiacDZW2ioTcsMo1Cg6ogW32vFSk8dVuk0OTklJl2Ey4ypW+sJWUa6EXv18IbVniIupONmYqq1hk3bDDROXyClWitzGIpZbpWGXZtxehMy6VX9k2\/WHMnN1yeKvo2ScbCdzLNBFh6gX\/GLXSuAPEjFCG5j6Hmm2nl5Q7M93fY67jWNBwX2IcT4nrqZCcmSgpSVlMsjmqIG9rbdB7iKZxQ4LIwRjt3DUm064uWeTLuuPiyi7lCiCm+lgRGyTXZUpmCOy\/N4+nqhSnKtNS4cRKgLceSFLCcuuiSBrFV4c5Zrs9Y4ljfOwJR6w6DnLR+TgEUDs1uA1PG8sonLMYYm9v8A6bjTv\/64oOP0oOK5xbeja8qk2\/dEdpTfaDYxF2DqNgszRYr7C0thtOUJXLNkje977aW6RiGMuKs5XqJKGexOubUZZkok8qjZaAAbk2AIKD84rON8eyWKahf4ufm6jMOtrExOvZkpJIuASbgC5EQWMMbTc7VJuUmHUuFDgKVpBJCgb6E67w5UrFmI+cGpOqzSXBZSwheVJ0OitgLxaKN2dOJ2N1l1miLbYWAFuuqKx6koCreOtottN7GdWmJhLVWxRTpWYGhlpd5DjnkMiFKc\/wDDi5tdlPAeE2EOYjqs0pQ1Px1pNHsZlyW\/4TC6a9wMwERyHaZMzCdLsKXNOf8AgsJSf\/8AI94iap2sMFUsn6Gwu9UXU\/ZWuTYlxp5vmaV8ikxFSPbMxLWamy0xSZGWkGlpUqWm3Zh9LqLgFJTzEtjTwQI3rtEUyVHZ3r7MpLNy0rJ1lTjTDKQlLaVTJIAHQWcjiMXIsNuukFUBt+EFUmx0G3XwhJWxGlgbjygpvZVrQUtg9bH5QUINs+UgbX6el4TcQkhWpudIavpuyRkvfXzjpfg1NlXDGkpJsW+a3Y6f6xUWSZnQhRN73GpiKnJoqOu1tLmIp6ZVci4SY6O4U1Sn1LBsst6eYZdbUW1NrCQRa3iff3joijY1wpiOZRL0yryU8+rUNsPBSiPQGHdQV8LOhpthawU3uDHLfbulZl3BtAm2mHEoEwthaj0zJuP+GOGkFa08tR1SM1z1hs6gk394SSlSVhQvmGsSLai+2FjS28AXl5iAbEbWEKS75SsAgG26lDYfrEpTaquTUqYbTznkkHMR08LGNB4f4raTNlDbQbCQHFZSUhROgT4f+kUjjLiyZr1aRIc0usy45i9bjNFNYlHZxaXVJ+pH1aCf8+kTs\/Ol6ecBUA02cmm\/tEFUJjNMBoq1Wq5N+n\/oLQSaxTM8uzTqkBI5bQSdUpv+pizU+mPy2BpaeQ444pM4Q+Avum6Tkuk+BvqIuPBjhzMYxxSh6duJOXsVDbMTHT9FpKafUnJWRRyiwhXdScpUm3Q+MZjPN4grEo8pVEpmFaU6ChU5VFhLi0+Odw5if3REC8\/g6lqzVCsT2J5o7tU9BZZOlrF1epHomDSnFQUabYNDoFPpMslaSspb50wtIOo5jl7X8QBER2nuKspxYZkXpChpoUtIoUyAZhTrjubqtR9NB5wNNdBwDgBzNZRo3LuPFDzghVpalKIN7W1tDvmkFCUXGnyhZp1NrBeYmHEuHHzy0pUpXRKBcn2hHCjDGE+1Hw8cxQmbpEhOyj7AcCFodIWhaElGUZrlRABEajxcwpg6tVefdxljqozdJS8r4eUn53krQ1fRBzlbp9kRni+OvBzhcx8Phiion3mxouXlQtR\/+9MZvwaEUTGHbWxDWG1IpFIlpVtJsh2eKpxafQL+rHsgRT5KZ4y8eXZNEs5WaxKTc38C0q5RKIeylRSDohNki5tsBDbhb2dMQ8WeMErgluZQlZdcM7U2butNMtklxwH72gsPEkR0472XMN8B69w8xDTUVBVTmqrzw9UHgSiXzZG0FsADMoArJtpoIxHhix8Hi3jbJm\/cb5lv3J9H\/mi3suKKUJTpdN9YUKlp1BvqL6x5YKjZJsNgY84oIsblNtTYa\/OElvpBsk5j1JMJqmMveOvtHkrWoFQBykakiKL2qT8NjnAk9qFKokivMddUWT\/hhv2tGuVxFqDgJCVVCYtY6d5QV+sYRMoK05UglajYAa3Jjprhr2IKtiU0VFVYm2HqgpOgGUai+UEjoLknyjorhv2FsPUrFlXZmvg2TJyyFBM29zSFLKrHu9bJPuYu+H8F8OMOYBqD6JppqqtuvciSZbQkOFCsoJv3iCQT7xZl8WeG9CpkkqnU5qbq9OZS5OtuZlqUMlzoqw1JjN2+12ZXFNRxNh2mplmnqcllqTQA3YpcFz3fIxy9WcUO4644TU5NpWJuamHam4VbHMyAB43B0it44xTU5mmz1LfxOv6MYcU2mlknKNb2Ou1zpFt4SYaXK8JcfGYLipiaozMw337JSgTbKtU21OoN76a6axmPZkcdY4nzkuwAX36LVmUJy5rqEm8pOnXVIgmI+F2KuIVf+lKczLzy5wZnil1tkocubgtkgjysLa6RcMOdjriLN09Pxk0miSOa4W8lxpP95wNpI\/tRe6H2EpZ4B2sYmdm2kak09tTiUnr3mkOpHuoRIv8AAngfgg2rVfl3nm76TVSYQr0KEuur\/wDBv5QZ3i5wJwg4FUyRYqUwj\/WSki\/MqVbxWoyg6fsqiFqnbbp8gs\/QGEnkq2S686xLeX+pZ5n\/AIxPnFAxF2wcb10lLcrSZMH7KnpZU+seiptbtva0UOucccfYiZVLz+MasqWVcfCsTSmGLf8AZt5U\/hFKzrmXlKWtx5w6qUSVH1MW3h1wmxdxSqE7KYVoT9WXJMKmZpeZLTcu0N1rWshKR7xCCVDYVm0KTZQBuPDeJPBgYZxKwuaZVMS2X6xkKKeYkEEpuNRcDcR3\/wAVAitdnzHSmU5UZJaaQm+awLMs6NT+9HCrV8qRcnT0gcp38uhgih3hb3gqk2GhFj56wRWnhr1EEKbjxtAFPXXfeCkZe8kC+2o29Ibut5r\/AHvAW2jdeCsypGBiznH1M24i97gXsf1i1TTxKSCTcna0Rk0\/3QL3AG3hEc+6ep7p8YIzV5iTRkbmFNpvewMa\/wBm50DjZQ0BplsCXc\/o0W+7HXmNa9P0WsSyZSRM2hxBzEKtYxm\/HCkVDilwtq9KcpDrcy2j4qXc3yrRr+IuPePnDVZH4R9RbUVpvbNawhmpJTfvG29yYSWASSLnrBW1lld90HeHewSUGxVrtvBUoKQpSjv4+MeSvlhRCTc9b7GJnDlUTJOlaioIKhcX89TEeujO4grqGmSS7NPJbWu18utrn5iLoMCPyUlO011m03SipLoGhsVXSv3umMsn31NPLSq6VZ+9eGimXpyYyt98na0TNOwfMvoUl5gqUoWQNN+hMaDgrAC3Ociovrs19YJYfZOmkb1gFpvD8iF2CHFqH2RaKnxOx\/U6Vi6eEo45LIclGnWphokEEmx\/EGMonqlN1ibVMzs09NzCjcuPuFZPlcx5DoSm4+1BkvqJHQ2iNxQnm0VzT7wuBF5w64lzhRgJ7dSGpxm1\/wBl8m3+9Fqw9gTEmKyE0mhzs6FG2dtkhAH7x0i4v8Ff9GmEzGNMV0DB7P2imcnErft5IB1ivVDipwEwIohM7XeINQQdG5Fn4aXJ8Lmxt6Xg1N7RPEjFc3KU3hfwupeDWZt1LLVTqDOdYKjYEuOADr4GIHi\/J4zonHXhLLY+qkvVcUSU+mXmZyV+wQp9Kkgd1OwcA26Rj1W4W1TiNx6Tguhuibm5+eLDToCuWhN7rWQdbJAJPpGtYG7G1Ar\/AGhZDA5rkw\/SkLcemS9Zp52XZA5ihbbMdB1AMX3tTcIsC8G8CUun4epEu8hthc68hCitxSnyrlqW4TchICUgepjYeGfErA2BeG2DKfSqaXxJ0FZeWtVkomZjLzVG2pJ1F\/CKNhFGH+CdXqlbwtUefUp5lbKmDMoNgo5tABfpGb1zjDP4xxlTUV+cLxlHW2ZQrdJIyruBb0MULCDSpbjTxwlwPtU+cX\/dnGVRPSSkhptW9wLqJtC+bVJOgv0G0CQcxAFiIRzjmlJt52O0AkJbXcp+cEdTnR9X3eloSdJ5dgbgC1j0ikdqttTsjw5mwLZqGUX21Q+4P0hPtUkTWKecD9ssvEg755dC7\/jGI0+o\/RtSkZ1KAsy76HrLOhyqBA9NI+hVd7UlZxLN099bKJdMkELlVpUUgpWixNzfYGK5SeMBxBTHXKnWG6fVS+UPqdmrqWlKSUXJte119Iy6v8WadRZpqcdm3JlLEwopS19lwE638NQfnCNd4mKRjhuqUCkPz0pVGQpTjCS4Aki1rW3Gx9Ij6FhnidimozLVHw7MmXBIadZRsCLWIFza3lFswB2e8X0ivuTuJ52msTKmlNIQ7OIQ8nMQb5CQvx2TFkHYwo89UnZ+sVabeL7inlpEu4y1cm+heDAI9FmLkxgfCOGZKt4Op1ZlW8RVaiOsyMnPzCkgNN\/XK5SW0LQpR5XV2OS+zJOLpPaUwq8iYdlSiouoK5dKFLKS24CkBaSnUHLqDvGs4n7d+J5BamKFhum0dDtynMt0qAudw0ppsn1RGR4h7WXE6pzK3W6+ilKWT3qVJMSyxr\/tEIC\/mqM3xHj\/ABLi57mVqv1OrrJ3npxx4\/7xMRstTZ2oTGSVlXphSSLhpBVb5RZ8N8NMVYiXMpkKDUJtEqLzDjMs44hgEXuspByjrc2EJVqhOUMoD+TMoZrp2I9Y6awJ2RqfM9k3EfFCvSVQn6\/MFKKNKMuqbbYauAX1gaq0Kjrp3R4xq2CuyfL8OexZiKv1zDkqcX1pJedmqiy2p+RktO62VfZKkg6p1748BF04GcD6PhfsTYhmVTErKzWJ0uTMxONtFxaJcAfU3AuO6hQt\/XMPuz1hjBWAOx\/W6gKqlc\/ienqVNBKkNPruFJDaApWoAURePmu7RJ2WQ4p0oQokry5tRcwnSUlqqy5SCbhQPS2kfQSU\/wCuuzvilBTmL2G5R4AnciRbTf5tGOFWvsgm+vWDgi40JEFIzEDUE7aQVSQPGx2gikFJItY9ba\/52j37IB18oLlCW7lXeva0ChhczdLSFuKPRCSfyhdWHKqUXVIusp\/bf+qT7lVhF\/4cYgpODqBOy9YqsnKrVMlxKG3OebZQP9XmttDyocZsHs5gienJzr\/NpT9VqTEBPcd6NYiUo09MWH2n30tfMBKvzhGSx3M41cZlpZlNFtNNNLebVzFKQsL0ssWvdAtoPtbRGCuFy6uRX3hcgOGbaRmsbXADe0dY9m2Ubl+L9McDyXiM6Rb92O0cQSofnWFHomFm2FfR00lCcyiyoAeJymPlPiiQQiuTrXfQpbirNn7hubgxU5ynOtrIyEJ3v4wwWysE3SoW3jzUuVKKjcJT+cGllLcXywkLvqkdRCuoVbc9QesIrIIJJub7QpS5ecrFXlKfTJZyeqE06lpiWaSVLWsnQAR9JOAfYfkcF4dl6vjJpFQxTNZH3JcG7UoQQpKEkbqFhcxcsc9nyg1yszNVXKhidflzLLcb7uZN9j4xxzxc7GFdmavMzVECAhpBUpCtA4BsR5+MUnBnZ4qVIfdeqbwJQnWXU2bg+IVFlVhVqQDiQm5TpYiHVPp6lzJWgXKwAq+lhFvUltiUCkpJ0EYfxPnlzmMnmnNpZpDftbN\/iirqcN73NtoWQ5dQHTaFAvUH5XhpWzekugnqPzjcOE+IsX4c4C4ecwRgekYlelvjJuaqlWyrEkC8R3QSkW7t9+kUPFvGziNidCm8ScTm6PKm4NOw8Nh+z9SAPmqMymaphCVeU+qm1TEc2dVTFXm+UhR8ciLk\/wB6Ev8ApUqkqAzRJOnULXQUySTzP+8VmWfnE3gOk44rnETC1WqLdRdl2KrKumZqjpQiwdSdC4QDt0joTtrviX7TuEHNi1U5VwnyJZ\/hGUU\/Hs7wi7SkzXKc4WH2npmXziwICytPUGJxzH9ToeP6djyk1F6WdelnJF+bZIWtlSrhV\/W\/WK25xHk6xNzlKqcxNTsnOJMut4JKi1Y3SrTaxiKaxpMYVSJNiYknZTKWHHVOWUpH3Tlte4MViRxbIylbXNTSlTry7nNLtqQAeigb3vp4QFMnpteIJGeTQ5xQdeQ8iZmuYruZrZ0mwSR56xqtBSUdpLimwASJukVG2u\/dQ5\/hiTp2USgB6CxBMLB1Vsydr6Qm++pCRqTfcn\/0hq0pSTcgEX6wX4lS15R009oMHy2NRYAk3vvBQ6p3oQepis9p1ku4C4azFjcSk2wbdbTCz\/ih5x1w1UMaSVAVSZBE1PGkUt8obslx1KpRsZtT3iDp5CKZQOyDxSxM0l1rD5lWSP6SYXZI9xeNjw12KsdvS8mK1iqUkWpdARklgqZsi2gOS9tPKLRI9jDAVNfXMVfEk7UHXLKcQh1ttJPXQrCx\/dMWiQ4ZcHcGNNsIpkrNKBBBmC7NEn0sz\/xGLNLYpwjRVNpk5JuXFgQqXlGGwNOhUhxfychCe4tUxakhyTNQRsETTzkwkeHddUpI9gIYTPFyalmeTIsiRaP3WAGkWP8AVTYfhEDM4inptwrWopBuL3vmHnaMkxNVyjtTcNG8xIBalz5B1a0Ef70ZHwaV8B2k8J5zl\/8A9iYaIO9lOhJ\/OGfF6mU6UxKhFKZeYlEhSeXMPB5YWLBeoSnQquQLaA2ud43XsIdj6ndoeq1nEOL2FOYMpCFsFtmaDCnpoBKgCQbpSEquT+MdB8Fexnw\/oPDniJiqvSFGqC0ImpWQQ+TOt08oKgF7HMb2sddr9YYdlTgXwqq3ZequN8S05dVqTky8t15SwgyxSvIgI71tiD3h1OkXng1iXCGG+xXi3EdMpbSXp0PN1DmOKcLrpVkSFWy2AziwBsLx81XmjX8MOMqOaZlblCralJ1H6j5R2Li\/tY1io8AqXhFmUal5BFJZbee5oQ84A2LpCQRcb6W+cUPGHaWxPi3hPI4anJuWTRnaeG2W+7nuMwtbxuOnpFBp\/HLEY4d07DD1ZqS6RLBxsyDa1cryJF9dFdfCK9SOI66RhunU1bTjiJQLY\/pLJAzE7f2oqdWrYnZhQTLZFAWuVZr269LRFSSimrS5INyvp5gx9A+Eo+luDszL\/a+Jwo2m3iUqm2\/8AjhGW1bAVcqGkOCq6j+UEUrcA2MF5motAEhQ+1brCmdLDCHA0h1SlqBz3sLAeBHjBBV5lF8iWGQdgllBPzIv+MJP1iovIyrn5habWy8w5fltEcWXnCVBK1E7nUiGr8itXdyXNzv+cJCluIFytKbaDU+9oFyRmm0ZiVkakZVRYsCzLjU64VOG6HpV\/U\/szCB+SzFmlqrMyCpmWYmVM8p91K08ltd1ZzqCrXa2kdA9muYdZ4z0NBWoIUtQIPXumO+6zpMtE+Bh1S1pcUAI+cnbHwAvh9xhnXZfOiRqQE8wSk5UlX2wD5KBjEmKspJ76yojoRrCjk+h4qDjaUjoUjeG7ozdwDKRfSGUlMmmVNmZKArlrzWPWHWIJuTmJxyYk0qbad7\/ACyLZFdQPK8aHwJ7MWN+0DXpZql09+Qw8pVpmuzLZTLtpG+Un7aulh7x9QOD\/ZJ4f8ElSk1QaZLrrTLeU1abbL0wokd4hSjZP9kCNVdkZ1ajafRbwMuD+sRlWoU640S2tiaPVK0ls+xF\/wAozDFL6pNTjLjKmnADdpwWVbxHQjzEYJj2abQl3IAk2N7+EY9OKQ4+pX3iYQkCE5u91uD+kPqzVJen0tTsw8hlhsXU4tYSke50GsYbjus06u4rfm6apK2ltNhxSFBQ5gSAdRoeg0iESAU2\/CDpsRbaBcWQbXtCNTVmpzoNibA\/jHUnZdSaj2YcVyIKb\/A1Boa6fav\/AIo5OofD1NaeShl+cnVHTlU2TW4fmqwEaxhnsxz82yHzhdbSLXMxXJzKgDxKEWt7mLKjBeFMGJIrXEWh0QjeUw4wlx8eWZAUv5xYMM0bCr9XkZjDfDnGOParzEuMTtTbUy0pQNwoFRJI0v8AZ6RB9vJ5cvx1w5MOoLLwTJPqbP3TcAj2y2jDePpZZ4pVzO0SUTTp0VbN9Yo+28EwvgfH1Ywy9WqPhmofQS1pBnkoUGlEmwOu+ulwINh3hDjDH3EWkYOaZyVOovpaspfcaT95xQHQC5MbjxP7F1A4WYDVXX6nVKo67PrSy4rIylMq33VuFGpJUq5TrokXixdmvgphiSwvTK8\/X5FiQxbN\/DqQ42FziJRtSsyEqt9XnyEKI1ItGu9q6pYeluCji6ItCn6jVUsMhltKeTLsWShpI3CANfMqJjkqiKH\/ALUmIG\/\/AJ6izg06lUgVX\/CHNIdvKN6EqKdbjeHBWQUjJp4A3jzhCwCpWvQX1hLPzFWQFOKOgSkEm8TlI4YYvxGEKkMNVJ1tWod5CkI\/vEARc5Psz42LCn6kqmUOUtdTs\/NpASPPLmtBv+jjh1hsH6f4oSMw6nVUvRWjNG\/hdGa3umMz7XKcMzHDfh87hF6ZmKQy\/PSxdnEKS8XMyFKzAgad4W0jbOC3E6YVwYwGqWlmnKkxTRKrnJhkcwJbUpCUpUm3dCQLXvD2scSa\/MOHm1Nee\/dLQShQ6boAP4xEz9XmZqVC5mZXNKUN5halqHubxAzdXmnkKS2AGxpa36RFmnzpb562HA0LlTqwQke5ipVnGdFp71na5KpINylLoUUm3gIiJfjLhGlWLj8zOqGuVhon87Q2e7RuGZUuLaolRnFXulLrqGx72vDpfGqurl5aZksOUqntzLQfa+LddeUUHY2SEj8YrDzs7WOIuBsa1OaaM4jEUjIrlpeX5TSG84Wkg5ib6Kv7RT5AihdounLvl+FxKwrzGWYReFuMUkmUxlVmFAgsVKaaJ20zn+Ear2VO0NM8HMEcQ6KwzLumsNBtC5l5TfIK0lBcTlQrMRdOmnTWHFO7TleoHDGuYJbm2JymVqZWZt15J5iEEg3bWVgJza3uk7RKUHtJynDThdifAlOobVVp9fbQpstzPLTJEp+1lCTm1ANrjbzjJJjH+I8OYZThqmV6aaw1NK+KfkOblZeJUFd5OxsRb+yIp0lOiRmk2ILJUQRsCDB6zNIZmGkhwlTScmUG+lyR+cJyiJ5thCkU2ZeZ7wCi0QkA+draHX5wR+WmkoAW5KsZgCoLm21e+VKifwhsqXaGYKqbKgbH6ltxR9rpA\/GPFVOJAUqcmT1ulDX6q\/KDNzss28kS9PCFZgOa68pax6Wsn8DHefZid+MwJSGlah2jTDNt\/sTr+nydEcMuNmXmn2TcctxYsN9CRB7jQ6jz84A6baiCEd4+fTzgCgWAvYeXSDpGaVBN7Jc09x\/yhJSe\/qBc+ME5I33HW8PW3UfCtpQk9b+UIrVmUbWABtYfjDctjld7x\/WCOoBbUm5AsQPlC1AHJ+NVbX4VSvXIpK\/8MXecpjdTmnZxuj1NDcyovJS5MyrZAV3vsqN7a9Y3jgYtDXGHDatiqYt+Bjv6qo+uZPrDqnNBNjYXjPO1HwqlOJ3Cao5koRUqWgzkq+U3IKRdSfQi\/wCEfKqbYW2+v+ooj3hNJ\/a32v0hyHMoC8wuPKEHsriLEX843bssdjrEXHyuMVSptu0fArCwX51abLm7HVtkHe+xVsPWPrRhzD9NwnQ5GjUiVbk6bJMpYYYbFkoQkWAiSKha+0JKdSm97GCF5Nt9IiK\/h+QxHKFmbRYjVDqftIPiDHJvaB4aVjB8q5U+R8VSyCFzMuLoT4FQ+7f5ecc0zEwXwch3GljDJU8uTcTYHMTr5CM44yYlmalLNSCF2lkLBWlJ+2el\/SKZh9ATKGxsAdImUG5Gma\/jCgTqdAD5QVQAUb2T4wjUFXp74At3dfOOl+yDWH6RwIxlMSzLUw\/KInlIZfBKFnK2qxt0MSmNlVimcDsO4lw\/i6XbrtWLZ+hKDT2k8lJBzi5zOXSRa+kYXVMC42xFZ\/Es7MtMnUu4iqQZQB5JcUPwEMk0DBeF0tu1XHdOStBzBijSjk4okHbNZKP96Nh4k\/yij+KHqeuRYrC5uQaDTb8q+iltqtcXUlrOv2ziOeuLnGis8cK1L4grjMtLTMqWZdtEvmtkSSRmKiSpWpuTD7jxJS7vGiYamV5JWZnG+arwQpSb\/gY7rx7xFw1SuFcvhjD6ENzjQcc5StA2ltBSgDpvmFo4g4BcTq5w64rpxJLPuvz7rEwha19+xUnU63AtaNFxhxlmuJcomgVqqoW0uWclJUOO3KFK1A8N\/wA4pWBcbO4FlWaZUpeYKZMONIQhvMWyoaKQfG\/h0iSw9i7GNUrMq7T8LzU\/yllJVNIJU42dxcjU9dIkpOTelu2bSZSpFEhMz0mmWWJn6sIW7TyhIV4XJA940eU7PM9Q5cnEWJ6FQEj\/AG8zrp4Xsn8YFWHOE1AV\/wBZY3nK26N2aVKGxPksJWk\/3oIviNwvoYtSOHs9WXB9l2szgbHrZJWP90Q3c7Stfkxkw7hvDeHGzoFS8nzHR\/aukH3TFdrXG3H9fJ+LxbUmkk6tyakyv4tBJt7xSZ95dVdS9OuvTzt9XZpxTqz7qJMEOVpKkNt5Bbe28IcXCJngNQVlNyxiCbRr\/WZYP6GJbA3GfDWEuH9Cp8xNuqmmZUJUyy0okG5JFzpuT1iNn+0lT2VkydHmJpXRT6wgH5XiPb7Q+IcRzq2pWQplPAbulb6VuAD0vvrEXOY6xnNpdcdxC7LS+cIUunyqWkBRH2c9r3tfSK3UVuVFC\/pKpVGpLIskzM2tSb+Nr2ilNyQuNL+sOUU8PKJNgkQd6mtIbUTdR3t5RqdIaZm6JhFcw265LGlnmKbcQgpCH3kjVX7vQGIt+tszWE3JuWIDNPxRTnEJK8y8uV\/U6Dy1tEDxRX9CcdsQvDT4euuLFumV8kflFk7Q0gRxRxohBTlZqzzllKAJClm1gd9xtGfyVNnZBTigWpcKTlUl55DZIv4EjwhnPWTmRMVKSZUSNElbh9bpSR+MJsztOl0D\/rCbddIsQ2wGwfLMVk\/hD4z8jykWk3ngCdZmZKh8khJG3jBxPqSRyJKUYG2jXM1\/tlUAusVAhZEy80SbnlHlj5JtDJ8LmFFTxU6o7lZJJgUyyCbKTbXQDXWDoYunXS3WDqSVkFQ12EF5YQAbXCVX22jujsmTYXh\/Dyf6tUYI9HJVY\/4zHH+KJT4LGFflbWDU\/MIFxtZxUR99Bvb8oKEk2009IMlO2l\/SCrRax29IMi3wrwIt9Yg3PoqEVpAKb6ix2G0HyZtdgTrCzNgBcaX\/AM\/lCWQJTmte4JtCSmwUjTTzEFUiwUmxIhSiNBU6tsg2cYfR82lQ4xglRnJBwAnmU+WVoSNeWkH8o6c4KOhrith8rTdXxYAJ6R9CKrbMyfOF5Iga7RQe09jWZwNwQr1RlWi6taUy6iPupWcpJ+dvePlNNVEzLrrqwkcxRUUgeMN1OII1G24vtBQ6kkISSpRPdG5J9Osdq9lLsHzeK1SWL+JLK5Kjd12UoagUuzI3CneqUf1dz5R9DpGVlaTIsSUjLtSclLoDbTDKQhCEjYACFHZ5LaCoagHWGU5V8iQAd4jXKwpSb3vaCpqemZa8qEi5JNoYOY4liSiUQqdUNM6dED+119oQGJKg6wtpTUo8y7dKpd5BUgjqneOUe0RwJUtl7EOEWpahvjV+nZ80q6b7pJALRvpY93b7PXhio46xAmsT0hUS5TpqVcLSmOVlKSDqDfWEcTOImpALQsLVe5sYaUIfzY2\/a8ImEABWm3rtBiQCbG46wB7oFuvzhGcTeTdHUJMXfhzxMrnDfgbXZigrl23n60qUf+IZDqS24yCdDp92M8qHFvGFSbLSq9MyjJ\/1MhaVb9LNhMX\/AIEdkfE\/aTL1RlMSSEjJIcU2pc64t2YWQE5iGxrYZgLkiOlKH\/JQUVlKVVnHdSmlfeTJyrbQ+aiqEOIn8mJRJGRlVYUxS5Tn28xmHq66lTZTpawSlNo5Y458DU8EaewycXUfE7026cyKUbljKPvane8WLjVw1xPxHx6w5hvDj80FScvzFS6VHOVMoVnVe9r36aaRf6J2euMFaRJTlUckqItnIXVzj6SFlItmUNwSALjqdesWBHCTAuEqnPT9Zx7QJGcmWy281Tz8StGb7YSgG4v6RXJekdnrC687UjX8WzKVZgvlltsm9\/vlBiXd494fpix\/o\/w1pzStkv1N\/nK\/u5SfkqI6pdo\/H0xLlMlNyNCZ25VOkk6e7mf8IzurVqeqPaZwBWp18vT807THHXyAlS1XSgqNrC+kM0yolq1UQburQ+tJUrU\/bO53MO1FS7KUoeAHSA5lk3SddLiDNODKbpN779I8ty4Kbak9ILqsC1jbUZjBVoJRe9idvKC8RGzMdn4nfkYiQT5Z5c\/+SKpiHDyJ3gBguvJShSpScm6etWyrF1bgH+9FIptNRNXUsG1vsgaExLsUsS4KmRkcOmY66entEnN8QsVS\/DWbwDzpY4dm6gmpuZGbOl1KcoGa\/wBnQG1txvETQZcJWWZ2lzk6oqRlLMzyglH3ge4q5ItY3AHgYjqhKyvxr\/wjS5dkrJS04vmKbFzZJVYXI0BIA1grLPLzabwWYQSyrUXKdhGr4WmKLLYDwm5WsPy9SYSmYlzNPuOZkH4h1QQlCXEAnvi+p0MMcdTUrN8PsWs02lU2mSMjNSUwlEm0gOA8wpAWq5UftnQmwimce05uLmK3Bb62aW8Lf1jnH5xZ+0eyV8T668E6TCmJgKG5zsoVf8YyUoCk5rXVaL5wG7P1b7R3ENnC9FdalSG1PTE5MZuWygA72BOuw0jbeE\/8nvPcQeI9Rw3NVSakqZTFOomazLSanGCpNsqE5invG4OvQxpnZ07AtOruO66cWyU3OYUpL0xLNlx0MLmXEKsnMEqunTU2tHIWOJOTkMVVWUkZZuXlZeadbbbbc5iUpCyAM33h0v5RXVtBB1Gm9zCZTe5AynXUmPIbskKuSR5aQK0ENk62Gu8KFhXLLnS14MJFxyWU6EqIAJJA0\/zpHYPZMnQaPQ0k\/Zq0+wPILlWF\/m2Y5x4tSiZHi3jBnLlCarMG3qskfnFYKQbbCPW1OVWYeceQCEkeOkeWCE90a\/OCtIztvjp3SNfAgfrBV6eN\/wAIMlPc1A2vtCrTeYXtm1tcwHeCSBa4vAlghCLk66jqDALllEJFwQdBrrGhcPeB+IMTOydQlvhkMF0N2edsSToRtvY7RMV7s143U1SUqpOZbckholLyLd1Sh4+AEXvhw4tHEfCswuwUqeSm4G4vH0PrA+rYV5wEm7e0OK3h2mYzw9PUOsSyZumzrRZeaVpcH8iPGPkzxpwnQsD8Tq9Q6A7MP0+QmCwlU2QVgj7WwGgOgihukpvYAC2pMdtdgTstS9YW1xPxbJJek21EUaSfTdLigdX1A7gHRPz8I+gpfKjcmEJmYypJ2AiDNULiJxsHvIvpfyiLXVPiWm1jYje8FRNtMsKU8vI2gXJtFenp1+qTIbeKhI7tsHTmf1leXlDyVCGmyRttppaFwobH7Khobw3qUomsSD8q4QlxSShV05krBHUbEHYiOEO19wLXRGG8YUWRX\/Mwliqy9yooRoG3Ao6qSLgXOuVSb\/ZMcwErfpz61M8saWv6w7oX9AsXJsrpEok6gae8HvdPibx4JypOsJvtpcbW2VBsK0KlXskeOmsTGFKQ\/WuEmK6dK2dcarcq4lSUkg3bcGmml7RJtdnzGGLWJc0+kVJ8g94rki0kC3RRsI6F4O4S4w8B8IuUqiP4Pw0y66uYcrFec5kyjOE3QEpVlAGXreFK7xXn5BZXjbtPFZBuZHC8s0hN\/C7YUYzjEnaI4OCYUZhvGfESb6LqU6tLSj6Zgf8AdjOuNHERGMMCykqxwuTgCQVMCak5xSFcybQElKhmUgZh3gdI0GucaMcyUlQKTI4jmZGmChSC225cJSrWXTclds29+sUapVSo15wuVWozdRWfvTcwt3X+0TDZEuhK8ugt4QdKEA6fI6QohIuSRcnYCFc4S3a3zhjiN34bitw2nABomQXfxyzSk\/pEhXkfB4vrzd7cuefHd6fWKhIq5iz1O4MKpSltKdwbawVyZAJ0Ay9CIFLgcVmNhY7WtrHg5ZfibXHr4Qi66AhR6DbTaD4rBmOAuJ02JDNZkHNT0Lcwk\/pEVLXmOyORa\/wuJnBr5ttn\/FGeUZ4IvfTQRJfSZYBXy0uAfdVCasSTCk2bZlmvNDV\/+K8JqrtRLmYTrqD\/APSOT\/htEOohLyt736wZtwgFOgFoFJQXkBxeVHUgXPkPnaLRJSBreE6Uw4iadZaqU2nJLtqWQcrKr2Gx751hVyh1GRwTjxU3RZ+mySpBoS7s2yWkqUJthVrkDMogE6ecRPGpPPx49Ma\/ziQkpi4655RpRP4xaOPiviK9TZr\/AOaoNKfuTveVauYydeUJTrdJN9tQY7H\/AJNXi5S+G2IscytXn26dJTUkh9L7qdOYg5Ui4BOuY9YvnB3tet4D4j4uqWKnZxykzgTMLRKZn3CsoCU93NY\/ZAv57xQOEna4\/wCiibxrVS0mdYnJtxyVk3U5Fr5tilRI8Ep2v7xzOpaZqbm1uXVzlKWF+Nyf4w1m5ItsNFViFJO2ulyNflDJTaQgkmwA3EOQlj4B1RzF0gBPrDx5yWZoBBau84kJzWUNfU6bR4z7ErQeQQ2VrRlJBSVC5+e0DKzcy\/RTJyDb806pJHLZZKr38x5GOpOzVXJ2YodCkpqYW41S66iXZbJultLspMFQHkVC\/qYw\/tDS3wnHPF7e2aaS5t+02hX6xQV3uB1gR4gWH5wJIBHT1gqwMtx6WjzKcweGgJb6+RB\/SEz4aXvCyU93peFGU5mlHU5VWgFApUpJ18IDmKWhKFKUUJvlSToIIVrS6mx1B67RdMEcWsRYKn2WadPJUwiZSopWnOlQBPjraxPhG80jtHVl2mtJdpkrOLbW62XDodHV6RFYLdZTjTDCUuXcbqCOnS4j6FVs\/wA3YJ8R+UN5d4Jt4xLSj9j5x8pu0hMmc4+Y6eUQu9RWkFO1gAP0gnZ84SL42cWaPhtWdNOK\/iZ9xP3WEG6h6nRPvH2Bp9Pk6HS5Sm0+XRKSEo0lhhlsWShCRYACHAV+URNXn8osCQNtIpjtUVLV0JWqzcwnISdNekMcPVLnc6WcILku6ppQv4H+EKzJdFTdaWbBtHMCT1Udvl+sFlJj4t1alKzBPdIVuDDtC9QnWwFr+MK5wUFJUCLW0jyZlJtf7Q0Ivv8A84guI8ozO4Qn31MIm20S60vtLTcOslJzJI6\/84+Y\/EqhNUB6dkWSS004S0vqts2Ug\/3SIqNBADbgESqbdbb9IOCLi+w8BHjsDYWhGZTmYcubd0xp\/Zam8Zpk8byOB35OXrcwqUUlc+kKaSkczMbG4v52iJxZxg4hTk7NSVf4oOy7jC1MusUlCk2UDYjuJSPxjNapUqBPO86pz9fr753U++GwfclRhh\/pPQpG\/wADhGSUobLnnXHyfa4H4QdviPiXRqktS1NB2TTpFts\/MJv+MdCdqOZnKrwD4RTs+pxycVILS8p37ZVlbvfz0il1wlctg91QAK8PSWp8klP6Q0AFhb\/JjwV977V4MAO9dQKv0gyVAK8Rt6R55ZBv1AtEfjlwsVfh\/O2N22Uf7s24YsePAZPiTitsp7v0k\/pbxWT+sMW0ggWFrjpBnnEjKRa3ptCbgTkKkkEG5H\/rCRXYkAjUXMClwhRBTrqbwRZKmyAq4Guoh5OJ+I4LY+aJuUOSD4A6WWtP+OIfCa2pjsp4sYdeQyWK4HUcxVsxLTWg8zlMZXITN8iQCrMnS2t4nvoydmGSpEq6EW+0pFh89oQao7wtzHZVk9c8yi\/yBvCvwEtlsupMadGkLUfxAH4wg9LUoOZvjJpwEahMulOvqV\/pHnn6OxblSM0q2mZ6bTr7JQPziLqM2l+ZlizLolUJIIDZUcxBvc3JiVo2OK\/h1mekqbV5uSl3nua42w5lzKKQLnz0iDxJW6lWJYGfqEzOAKzAPvKXrY66xaOKaebVqFMWuXsP0xWvX+ZtJP8AwxbuNJkX8KYFnUrmDUnsMyKXEKSnk8ttHLBBvfNmSQelrRjSVApJKjlA2MOqDUzT33lpSpYeQUZdbEXvrbXTSJkprNWa7tInnE2yHJLrKSnprboQPlDmUw1WVEpVKJYKrJKJh9ts+WilAxZaNwyxJUQpDEoXVAd1Muy7MX9OWhQifc7P+MH0NrmqZUZRsJuHHZJTCB7vlsCE3eC8rIC1UxLQ5A21EzX5FpQPmltx1X4Q+kOFGGZt1AGK5GoAbM0+XqVQJ9OTLJB\/vRY2eClCZCHHKXiyYQnXO1hRTCD\/AG5yZSPmIeowjhOkJBVRZplSdc89iajU9X91pLqxDaarGBpbMiaVhhKT\/qp\/EdUnz\/dYYbSfnF34YYmpmIBIS2GKxhqXkqVWZScm6XSaRMtuuhxRYDnPefWSBzLWy9R4xkPaulPhePFbIGUPsy7t9NbtgfpGVDVdyNPIWg1hexGhgcuZO+8ecbTlsNxa94LLo+tV5trGv7p\/hCa02SPw84Xyga6J8ATEhIU2anmnfhmHXhor6tBMOprCNZk2ObMSKpVtCgjM+QjU6gWJveHqcBzgYEw9MSbSMtwrmE6ewtERUKMzKS63UzfO5dvstkJOvQ3iIaXy15077x0Nwywk7XKLOzQUyhtU88EBV72uD+ZMBhF5H+mlCdRopM21e\/hmEfRytKBprC766flEdLvAAa3h4ioBpJUdEpBJJ8BHyYx7UW6rjjEc8lZUiYqD7oUo3uCsmOvf5NTDCFPYzxMpIJSWpBpRG2hWv\/DHcYeSpV73HlBJudDEupV8p6RT56sTMuVrTLpn5c\/bbRbOPSKvWX5KssFVPmC1MNnMZWY7jg9PGKzLVZVPrb80ruomWwpY8HU6fiLfIxaF1BLyGX0uhwuINz57wnT6klQWpJGbx84k0zeVAHUjaCqnAm\/j4CEnJ1OpIy23uYcNrD0u7LPpzy8y2psk7KBBvHzb47yZp+JapT1JsqRCZVR8cqbX+VozChA5XABrpEroDre2u0CFAnwTA5rakaWgjyrtG3VMbN2HHg1xAxA0Tothk6\/vKH6xkOPMPvzPELETEvT5t4pqD4+qaJB+sPWBpXBvEdUWOVh2YUg6Avqyi34RdpLs\/ViVaC5sUajt2vnmlpBHuv8AjD1XDPDsq2UVTijIS4GqmaY7zD6WazflFl7SppUx2deHyqNOvVCnycw9KNzT6FIW5ZIBJCgD0PSM3qjubDuBXOiqC2m\/7rrqf0hs04MoFyE\/OBub3KbwZvMelz1EGCSCFaanpCoAUbEjMBERxPPLo2D3xoEtvpFj+y\/f\/FFy4qJA4rYlyiwXNlz+8Af1iIaUkKGpB6eEKO2dUFE3A8BCDrmwAsEm2nWEVqKtQPaCpubrGgOwgFghNjod9DEk06l3hvxElyQSaaw6Lf1Zpr\/zGMQpeJpiXpC6ey+38Ip4PqYW2laSsDLmsRvbSHknU61VVciQXPTJOzUkhR\/BAi10zgbxTxMnmynD\/Fc62LqL6qW+EWtuVqTa3neEa3wfxzhaRXPVPC0\/LSbYKnHUJS6Gx1KshNgPExVW5pJsoWsRCTs4L32v5w2XN3NxqLxovAPEOBsL49l8Q48kqnWafSxz5SkU6WS78Y\/90OFSgAhP2iOpsNrxXMVImsY43r9YoeHKnKU6oTrkwxKLaU6ppKlEgFQSBfXw087Xhqzw0xVWXUybVK5Ti7WTMvNskeZzqEWzjJSJWjT2GpBmpSVWek6JKSs29Tn0vNoeQkhSM40JToI13CuDsP454VYDrFQla\/KTlPkn6cp7lSAkpptD7miTMvthVgqx0MOV4c4fUNIAlaa3f7Rmq9TGfmJaXfUPYw2mMa8P6ShSG38JND9kztXnj8kpl0GIiY4y4IYcUpmbpvMG3wGCWXD7Lmpl0+9oaudpGQkAUyM1iw5hYCQNOpaR6cmUuNuiogqr2gHKokpdo1XqottWsUVB8e4bdbT+EVx3io6VlyWwVhSXWfvPU74xXzmC4YUleLmPHl5KY9LU07BNIpTEsQfAcpAMLTeIOLNcQpE7WcTvNdUuuvBIHvYRBO4MxHPuAzT6io63fn2wfe67wKeGzy0Hnz8k2QbFK1rcO\/8AVSR+MKN8PJFCsrtSQT4y7ClC\/wDaKY0fs6SQw9xExTTQoLMvLNnmWtmLU\/Lqvb0Bh32y5bk8ZWntRzqYyrTrZSx+gjFGxYX1uYUsdtb+EALJ6a+P4wZSQbXAB6kDWAlrCZbSSbk5dvEWiaw3gWsYr5X0fKksrcDXxDhyt366nU2GukaZR+FdFojahUVqq76gVIcAytoQm2YhN+8SdBeEZjHD89Ig0aSRKSyXhzXZlIS2lKfsJB2t5DW9optSqzBmJjM+uqOur5hzEhlJ9N1W26QzdfcmrF9V0j7KE6JSPACEZ5aXKfMJAA7kVxuXdmX0My7S3n3FBKG2klSlHwAGpMdg8KcH1WQwdLJmW25R1wlwtOupChcAm46G99DrpGdYWbUziumKBByzDZsf3hH0grZ\/6il1Xton8ogmXSoJsdIRxRNKZwnWnEKyLRJPKSQNQQgx8mb3C1ElSiSST1j6WdhqhJwx2fqZMFvLMVeaenFnxGbIk\/JIjoP4vu9faIOs1dQTkKhpFYnJ0uFSkuKYe6ONkXvFaqVbcaJRV5ZE6z0m20ZHUeZIiuVOotSDiUvu8+WdF2pkDcf1h4xFsYuXIVKXk1KBQp4BJKtLKBt+MWun1QJtnJQgnTXr5RLy1VDiQq4uenlDhE4XSQFakaE+MJGZzXsRqdPA+UeptdTTqmadNHLLzHelnFnRK\/2T\/nrHFna5p4keJ9bPJ5XxLLL+U+aLH8Rb2jAKLdJcudDaJVWvlHgrw66Qa4N7G8JuXyE+Ua12MX+TxRqqRpmlWlWP\/bJH6xrclKSVXwbxBq1Yxa7TsUU+bmW5GlU2XZQt02u2QkpK13J1sekc7PUbiliVsc5GIXm1dHnFtI\/Gwhg5wcrhVzKpO0inHcrqNWZBHmRmJ\/CN+4X8asAcH+F83hWpymA6lU5plxpdaZaem5nvFVjo1a6QbDvDYRj\/ABZ4rYcxNwskMKUmamKjMydQdnzNqlBLNZVJIypTmJ67xHTeV7AXD13qKW82f7M07\/GGqAAnbvHUQokXB28oUSo6m2oHSA5gSPU6Ax4O5VBWh8hDHimrPgTCrw2TMTzV\/Tkq\/wAUW7igsK4jTrxF+exKu23+0w2YhlOJy23FgN9ILzwdhYesEU7k+1oTqel4c0OkVLFNZk6NSJF6oVOdcDMvLMDMtxZ2Fv18I7WwP\/JlvTtJl5jGOMFyM64kKckaVLpWGv6vNWdT6Jt6xpdH\/k3+FdPCfjZ3EFWI3D84htJ9kIB\/GL7Q+xjwbw\/KTcuzg1mZam2uRMJnZl58OozJVlIUsi10pO3SLPh\/s68KsKFP0Rw5wvILTstqksZ\/7xTeL1I06SpjQbk5OXlUDQIYaSgfICMa47Yvfnl\/QUo8WpRsZ5nIbcxXRJ8hvb0jnqamghTiDbQ2IMZXIdjfC3FDGVQfkJ6TpC3EiYclHGXVpKj9ooCHGwBe1x4nTSNDpX8nlg2TA+Nq6VEb\/DSCRf8A75To\/CLpSexfwxpK\/rXJp7wBMuxf3aaSfxiwNcC+ENLmmJJyUVMTKx3GX6lMukj93Pa3taDGhcB8H1BuQXT8ISE+o3SibYaK7+ZWPzMW2jY44ds1Bil0qs4dRPunIzKSTzIUpX7ISnr5R8meOJMxi6afNruPuK+a1GLmujS9a7I1PnFi03S3n2kOA6hJnCVJ9CHIouB+Cz+O5qkSsvVJSXmKlfktrQskWKhqbAfdPWNVl+xMWUpdn8TNtyxcLIcZlQRmH2hqvpcHbx8I9iXsmyWGZGWXJJq2JJxahnYkyAlCSkkE2bJuCLEdCDE\/hvs00bIV1TDlQDQaUsLU8pKuZlulJ7yRlvoTFTqfDaVpk28qWaodNkhYINQyrVtrcqChDdNNpNNUM+MKBJqSm5+jS1mH93LDaZruEEJSJnHU3OXN+U1LOkD37wiMcxZw8l1Oa16c8C2EAE\/2gkwg5xIwhL5QxhebnCNSZiZLd\/7qyPwhsvjDIsoWZPBsiySLDnvc780Q2d40VgFBlaPRZLLspmWUD8woRJ8EazM1DihWJ6aUkzU9S591wpFhmDZd0HqiLj225e2PcNzQF+dTVJv+6sn\/ABRgTabe41tAnQgiwPjeBTcK0IuYBV+8b6+MXzgjgGncQMXvU+qTMxLsS0uZoCXsCshSRa5Gg1vHTtC4cUShyjUnJyriJdlRKO+RqdyfM6xJDBtIZJW5T2yoAgZrq\/OOYOOzRp3EJySCg3KIYaW0w2AlCLjWyRp0ihh1OlvxhRta3VAISpQ2FheHgoNVn2liXps28FAjuMKP5CH+GMCYxbdcMng+rzgW2W1LRIO5k9UqQq10kHYiNXkOzLxVxRKoqNRmqhTpl5KSWnGnVOEZRYrKdM3j10iDpswhrFMn8OrOwmaRlJ3IzCPpNWDnwxLq8UoP4RXGXcoA6wu+2mclH5dzvIdbUhQ8QQQY+VeIJBuk1KpS1wQxMONpHWwURH1L4DyaafwiwZLtpADdKYNh4qQFE\/jF9em8rRIvfwEUvEU4pCjdWS\/h0isrr4lyRNJW9LnTMkC6fOGlVTeUVNy63J6nn7Smj32\/VMUGsPDkLSyp5csde83YpP5RRKxXFI+HcUfrJRxJKh95F9D7RpEpVubLF1JJBFwo6gp3Biw0+oLU0wryzWH7PnExIzOV4X1QbXF7kecPJpQZeOYHIo3Ch0MIV+mJqdNzWspBv3fwMcq9s6Qnqi5hOtD6tPw7lMmlFNyVoOZsn1SpX92Ob5GXTLPFIVn0vcw9JIUeukFKhmH6dIMDY36x57RB0G3jeJ7hBV5yi1LGE3TphUpPN0J59h1vdCkOtqBHpaKvVeK2OakFvTWJastC\/tLQ8psH+7aJ3hFwGx12k56daodalgZRSUuu1WeWkkqCiAlNlKUbA9I3ylfyV2LZgIVU8b0uXVpcS8q67+KimDYw\/kxMQUSjJew9idivVTPl+EdlRLoy21OYrPlGK8QuyZxJ4TYdna7iOnSjNLZRlccl5hLhTmIA0HmYbNOhfC7ACyq+Vqdb08plR\/WCNuggW1PgNYnKVhHEFdUPoyh1OoEnQSsm45\/wgxbJTs+cTZ1vOnA1ZYSfvzUuWB83MsMqpwQx\/RlJEzhSoqKrkFhAeGniUEge8N2eEWL1lKHKSJVR1Am5lpk\/JagYh+M2BazQuHuGZJcsZyfcqE44linH4khCm2QCSi4uSk9ekb5L9krFHFFig4jk1vyTc5SJLmsTUg8h1txLCEKBBAH3fGLbS\/5POrPgfGVqYSDuBLNo\/FTh\/KLfSv5OukgJ+Nqs66etplCR8ktn84t1M7BOAJGxmk\/EkHd2ZdV+RSI1bhXwIwHwiqq6tRaPKJqhbLSJvld9tJ+0ASSdfGNPcxAnxQPe5\/KIyrYrdk6e\/MIIBQNCB\/GKKeME8knb+6II5xknkpJ7v90Q2c40VJOwQR+6P4RS6riClzzrs1O09BURmWtT60jQbnWK5TZLDuLJj\/qzC5mpYq79STMOIZB8llXf\/s5vO0WSl4QpdBm0PScslucbvldQpQIBvpck9NPOH707lmC3lU8\/vym+8o+Z6AeZtGb8SOO2HcAhxqp1H4ieAsKbTlhTg8lubJ9rHzMcv467TeJMSpelaOEYcpzm6JT+mc\/fc3JjG5p1yafU6+4XXFG5W4bk+8W7grJSkzxlwa882kvtVFlwOlWpyquPyin8YBnqSXADqpW+nX\/nF6wYkTnZKxC0D\/Qzs3oPIS6\/1MUXCXGjEuCaTJyVKTIN\/ClZamHZYOO2UoqIJUbEXUdLQ9qHaX4nz5ynFK5Zu9yiWlWWhf2ReKzPcWcc1FxSpnGFaczDKcs64jMPDQjSIWcqU9VFZpypTc2bbvzCnL\/Mw4pmEqjW2HHZCmPzjLRstxloqSjS9iYkZrhliKQaQ67Q5spXe3Kb5hHqE3IHmd4NJ8N8RT7qGWKFOKUoaJU0UH8bRKI4OYrQqztMblvFUxOMoA+a4gaJh9urzz0u\/U5CnchQDjk2+E3GaxyftkeF4sEvw8pXOdRM4skNCpCUSjTkw5mG2ZKU6fO8TNB4Qy9TdebLGIqilDtkO02hTKkuIsDcAo0ubjytfUb2\/hhwCx3I8UWn5DBWJU0Uy82ymaqFNWwQHJVxAzXFvtKAv18o1DtM9nfiBxFXhJ+j0Dnuyku43MiYmmWOUTkIvnWPBW3hGQs9jniEw3eoPYapY2vO1+WTb+6owunsi1FpQ+N4i4Dkr7j6VU6R7IbMOU9mLDMqkfSPGXDTZ8JOTmXyP90XgU8CeFMrYTvF6amFDQiSw85r7qWItHD2S4McJq+7VZTEmK6\/MLl1S6kfAMMtlKikndV73EXd\/tIcMpa4YoGI5u2n1swy2D8gYjJvtW4PbB+H4ezD4PWaqqtfUJQIqdb7RmFKtVDOTPCHD81OJbS3zZ5x2YVYXsNSBpDRztPMyyB9H8NcGSShsRSkrI\/vEwmrta4waJ+BpuH6cLaCXpLAt\/uwzf7WvE9zNyq8mWBG0vLtIH4JiGne1BxRmklK8Y1BCT0bcy\/lFdf448QJh0rdxfVyvx+LXr+MPZB1LVRle8dXUm\/hrH0ynFBeCZVd7\/Utm\/sIrDK82oMOmXFEgbx8wOK1Ndo\/EHFEm8ktuNT7+h8CokfgRH1F4REI4U4SVbelS23\/AGaYmKrPnl9xWXySYqdVeUVZx3gdxfpEFOUVc0wqYk1\/ENfebA7yfaKoZqcoU0t+TcU2PvsqV3VeREMqsuVxE25MSHNlZwauyqbAnzT0IjLMQMPJeWG3EPbhxpYyrI8CIs+A6wXcMLYUS49LkMAK3ynVJP5e0aDR1XSyFHXLY2iSM98I8hNibK0PlE+t4TkqFA3ToCPAGG0rVfh31SMydHNEnp5Rl\/FSiyeMMO1fDiUuLcmm+fIrcH2JhGqRr4m6fRUcQoKkzi0kZVgWKeoMLnbTaE1EA6bwKHNTpAurGTQxMcF5WYq+L65TZRlyYmJygzzSGmk5lKOS9gOp02hyxwMxLN0goVSqq25Y2L7IZaGu5zkdI0fgfXsV9mqXqTtITg5c\/PuoWXcQ1hpBl8iVAFKUOXN853jRFdrLjJiB4pax5geRuf6Oi0mZqCx6ZUqvFuwbW+PPFKsSlLlcXV1rmjv1BvCKpKXSOqi6+ED2Fz4CN5pvYglsQyb0vxR4kYhxsw8BzKaJn4SVvvYhPeVr5iNGw32TeCmFaTIU2VwdSnZSRzmXRPOKmMmc5lf0ijudYvNJwdgHCyAKZQKDT8mwlZNlB\/AQzq+LZ90qYp7DMjLjQK5rYWr8dIqdel5ueZSjnNL0zLUqYTqr5xTVUGouOONONNZCCM3MCkqHtf5RnMl2Y5\/EHFSWqy8aVKVww2sTMxQJV1baFZf9WlaVJyoJtpa9rgHa3U7LglWUNMpSy0gZUoQAAkQC5lahqpR94IXL7kx5Lqkg2UfYwBcgnMA8IKXB08IjcQLC6NND+peMVem8qiLgHzhquo5dL3J0AEM51OIZqofR9No7i5iwUqYnCWZdoHa51Kj5JB8yIsdJ4UytmpvEb5rk0mx5C05JRCvJq5B16rKj6RbpiaTKtoJSlpgd1IA+QAG58hFQx\/jSj4ApiqjiWoiksKF2pRJCpuY8AE\/dH4+kcicTO1HXcWNvU7DTZw7RVEglpX84e1sStcYkUuPuLWoqeWo3Utetz5wRaSkqChY+cIOCxFh7wDMy7JzTMxLuKl5hpYWh1BsUqBuCD4wlxEmE1KSk5oWu4DmHgdL\/AIxoXCFBnezljyUSf6N+aXa3QyzX\/kjCU\/0aSnqInKDWaLSmHE1HDqKxMlzMl5c440EptonKkWPU384vtGla9OyaHqPwnpbzLgC235yXcdBSoXTqtYBBGt41aj4cwjLUmVRXayKPUikLflKFhOXBYXbVHNcfOaxJ1AsbROpq3DaTUoM1zHi0KCbpl0SEuk2Fv2FEXOvl6Q2dxLw3S2tBp2NanmAuJrE6m0qtrqGm07xHP424ZsMobRwvROrAIBqddnZgb31HMF4RPFPC0igGQ4R4IYVfPnmZBUwb2376zrADtC1SSWDSsM4TpIT92ToEsk+xKCR84JM9qTiY4oiXraJFs9JSVaa\/4UiIyd4\/8SJ1B5+L6tbwTMrT+RiFnMf4rqiM01iKpPhX+0mlqv8AMxBzM\/OTCjzJp5ZvqVKJhuouqJ7y7epgGkkPtrJum4uD0hotlSCdNSfSEi2VKGtxe8JqTqBvCa2x+wQfKETLqB1Ta9obrZs4okjUddOsIu8tu93Eg+sIOTbCEgl1seioSVUJUDR5APS5hv8AScqL5nUm0EVUJdKiFOZj4p2jRQGRPMkA2Didb7ax9MJZTTmAZBbqVKb+GbJCdzoIZSyKYUAiSmVf2YdoMiFDLTZg+0cC9ujCaaDxSTV2JVctKViUS5ZY3cR3VfhlMdtcOngjhXhDUgfRUtonp9WmHk6+XEiwKR0vEO+gq0V9YDsDpENMyL8vMCYkXly8wnUpz5QryIiJqdYkJ5Smq3KuUyZOiZ1olTSj5+EVPEmFJyVaE1LLL7B7zc1KqzD8NRGeV+rPvoyVKTE7lFviG+67b13MRWCq7K0jEwR8ar4ObIbcamE2UhV+6QfXT3jY6dN8iZS2TdwEknw1h9PPufEpuq\/lFgokyM6W1KshactjDiepP0lLd26Jlo2SfP1iqcSac\/O4YYrcqOXOyKhzQjU3G9\/z+ccV8T6amQ4gzbzCAhifQJ1tKdhn+2B6KCora72HW8Iqc7thteC80JJvcjpbrEzgzCc9xExdSMN0sfz2pzKZZsnZFzqo+QFyfSPoVhP+T44VYb+FmJkVup1VtGVyb+k3JfOojvWDRSQD4XOkX+Q7JfB+RIUrA0jPuDXPUnXZsn15q1Xi5UXhLgLDuU0vBOHqepOypelsIUPcJvFrlg1KIyMNtspGgS0kJH4Q4S+Sb5jETiOdcXO6rNgkRDKm1XtmPzhpM1BuWSVuuhtO11G0NEuVSpn+YyfJbP8A8TO3Qn1CPtK98vrDxGH5CQCnqzNuVeYQM3w5sllHmpA0A\/fJ9YqGK+MVJo8ut9lwVN0OhvkSahlSfNdrW6d0H1ifweqrYmwrXZxTaJdU0MsjLsLUwpuySbKd1UTmULn+roItOE6RNUpqZXOlfxDjhy5pxyYGSwt9qwSdDoPGJ7mabwUuG5t+MeLkCV2O8ASCNrwRR08Ijq6u1KmNvsmMIlpKZrM4tDIKUA95wg2H8TFxomGESSgppCi71fd+37DZPtr5xcJCmolk2Asdyowk878Q48iWShwsf077isrEuP66\/H+qNfSOduMnaxo2BHnqbhFSK9iFIKHKq6n6mXPUNp\/yfHWOR6ziGpY4qr1Urc89Up90lSlvLvb0HQQ2RT27mxCRvb+ECCiWJyuJV0KSL39ojptTDiiQNU+GmkMHWOcTkzG3jDBdybCGFcez0sIJNkruPK4\/5RrfZxbE9ww4jSWa9wmw\/fYeH+CMUoGF63iVkqpNJnakG7BxUqwpYR4AkDS8WRngrjqYQhf+jE6hKti9lbubXFsxESbfAXiXPoS0aZMIaACQhycSQkdAAFHT2i907DlQw7SpSl1JoIm5VrlOoSrNZXTX5Qg\/LpaIC1JSAb94\/wAYauTkoyq6pthHjmdSP1hhM1qlCxXVJMW0\/pkmGL+JqIjVVRl\/LIq\/5CGMxi\/D7QVlny75JaV\/CG6seURIAvMLt1DP8YbvcQ6Ye6iWnFAeKQL\/AIwgriZLN3CKbMK6AqWkCGj\/ABMcJ+qpet9M73\/KG6+Is+u5RT2EA6XUtREIKx5VXBdLEui+wyk\/mYSVjOtOKtmZHmGgYQdxLW3bfzv2CEiCfSlVcBvOuA26aQQztScFlTr58wuE1fFOC6pp426FZgnw6lg5nFqBFrlW0EVJBSvvEjrHkSabHQkEaiDfCJRYWv6QZMmi+wJ136QdEsnKNAPWNNmQfjEgEgBY\/OPp3hIl7h7R1XzH4Nq\/90RNSSRyU6AaQ5QOhEc69uPhDUuJnDGWn6JJqnapRXlP8hsd9bKk2WEjqRYG3lF\/4UyVQqXDXCSGGFpKaVLJXmTsQ2NIn6jhqqtqsqXWtI+9a4ivVFD0nmRMSireVxEE\/VJIgp5qmXDoEvaj8IiKk2++yvJLsuIULEpQVpMUeZplZpLy3qGudknDqWm2yWlf2TFUxJiCorCkYgoSM4\/+Jl2i2r1MZ\/Vm6TVEqLc+WlDZLzdyCPMRqeAMRorFOYUX0zE3LDkPqT97TuqPqB8xGgvo5hZWNlpt7wvLPcp1IG6db22i0ScwHFtqKrpWLKG2vjArpzaX5uUeGaTnmyhwW0vbQjzjhztDYbfwtiakykyjK6hDzQUNloCgpCh4ghRjMFL7um0NHV5SRcAeUbJwl7KGMuKlPbq7gYw7h9xJLdQqF7ufuNjvKF+ug843Ts6dl3EXC7ihOYgmZ+mVVMjIPCnOSSsyzMKskZml5SO6Vdba7x2fRXp9FHaeqVxNL1KFJSkpPhoSIazmJkyaS+pxLbCDY5hqv08PWHtBxNTsRsqcp80h\/KBnQD3k+oiWC9dDtB0K1iv4hmLzpA1VYAWiOakZmZAJPJHiRdXyh4xTpSQcStSS9M2uM3eXbqR4D5CKVjXjPSsNqclmnxOzo0+GlF6JP9dz9E\/OMu5+MOME4qWlimWpiFfWEHlSzd\/E\/ePzMaNhfhdh7Bvw638QS5mkLDji3JlmylDYBJJKRvtrrGiVPE1HwRQH5l19pbEujnFqUspS7690aXOsZvN9qqhNIbWxQqs6lxBUC4hCLKvYIN1Gyj0ERM12smEpUWMOKuWxl588hH1xI+pItorX8IiKh2sarkUqSo1MSkpS0kvTZXlmDl+rITba+8RU72r8QrEz8K3SWkhKJdBDDy1Jmjluk62KdTr6RG1HtJ42eXNpameXkZRLBEpTwpQmTlKlozDVH2vS\/lCtF4pcScR1wSsrPzkwwxIrYfVy2mg5MZO642NCTcHQXtrpG38Hl4uGDpaXxK+p2rNrcLinLOOFJUSnMoG17G1vKL65Qp2Zl1cx5TiVC3Lyp\/GwiuP0VckrKGciB0AsBE3RcPrdknpx1SJeUZF3H3VBKUj1MVvHuJ6Xg6kOVPENQ+haMlN0MBWSbm\/bdpJ8ftnwEcU8Ze1JUuJKF0ShpNDwy3dDcpLHIpY8VW8fmYxH4WVQNVZdep1gAqTacuFuqNt0f84RfqSc\/dQbdOarMYSVVHT+yAetoaKnXAc2gPiEiCKqD\/RwjptDRb5Fwbb+AhlU1KfknEk5tLi\/lGv9kbLMyuPJO+rjUqq1uln0n\/iEQnZ7mizh\/EsonlKdQ+weW+5kSpJQ8lVyCDba9tYsmLOIEvI0GTVPIU3LCYS21LUuqG7aEosFFPMz2NrAK2vqTGMDGmIJ+ozEwzWpiny7rhUllU9mKQdLXUq508YAVx5hqbcm54VOZUyW2kvOhaEEm2YgXuQNoY0Wel6SJp1wS9TfellsNImOYUslQsV2KLEgXtqLHWIsUoWAyXtrtCyKWlKfs2PpBX5QNgHIE+2kG+CTtl1HXzgTLAAnKL9Y8Ja6QLajYCPJlsySMm21hvAfCKQr7I112gUyl1A5CR8oH4ReoIsPOB+FUnXL6wYSZy66+cGEspIuNvCDFglXQfxgfhiFddPKCqlLdYAsC1wLHqIMiWJuBrbpBPhio6Jv+kHRKeIsRaHP0ctAAIUCdbAGO12uxHV1zKXJityoAVchKY6poFF+h8OSlLU8lapdlLWcdbC0SLLPJGUm9vCFkoJOgJiIxLjrDuB2kqr9XlKWFi6UzDgBI9IxSudtPCWH0TkvR67SXChZSwg5wi3icqYzWb7cWK1uKW1iPCiGuiPrif8AghlMduzECk5X3MLT4H2gQ8m\/zTEfPdtGTn5Yh7B9GfmDuUTuVJ\/AGGUj2iqNWWs30bWMNKJ1mKXNc9lPqCIlm8YVmsd2mY+DqdwidaLTg9Yb1Kaxu0336vL1BsjWxCrxn1ddVNOH6SpzQWd1toyn5iG3D+ry2FcWIcRMrRJTaeQ624q4BJ7p9jHS0s6XqYm\/22xeFW3AXkqH3tDbxiXpk8CpbKjlI\/PpFnlkGoyqVJut5uygB1jBO25geTPDrDmKWniqZkp4ygChYlt1JOU+NlJFvUxxoha5lbbLKFOuuEJShAupROwA6x2L2fux7J0uXl8V8SZcvWAfl6MVWQkbhT3j+78\/CNzqvGycrbr0jgWhGYblhy1TziAiVYA0+2dNPAQ44XO45UipTeMqrITranU\/AopwukN7lZJAIPS3kT1i6zM9OukqamVJXf750A8ohq1O06svludS\/MsS6SVMk5UOq8VKBvbyjnvEna9GD8Xv4VwPQpeoVFskO\/DNtstsJ3IUvTMRvFvwnibirj9tT0lXlJbaIQ6CttsgqFxYIB2jpShCqilSiZybvMpbSHFJbHeUBqdfGJRyXSq7zpSFAauK0Foq+K8bUzCUiZmfmkyzdu5pd13yQn9T8jHOuO+ONUxRzpSm5qXS1nvBCjzXR4rVufSKlh\/D9XrylKkZJyZS2Mzi8wAA8SSRp5xpuHsBVmcm6Nh+VpE5Pz80pb03MpKAxIthVu8dbnrqR5AxZ+LEzhngcmSl6X9H1TGinG0\/BVZLpYbSUEhxISBck21J9BHOM\/VF4zq77zoYXOVgquy18QpCJlBKiEgm4QO9b2hy7h6aqDhKKbNNLnEhxCvo9ZDTzd7rOZXW2g84JPUqZbQ7MOyrtPYm\/rsz0oy2ZZxF+9qdVKsLQ2lpk1ZTaWhOS6KoFuqDTiU\/D8vN3k2Ruqw19IK07PVJ9sKlKihmpJUmZAdX9U0i9nB3dL2F1HSNA4UcJ5zia\/W55yWdknAyJORLzayiwt9aFE66Aja2sadw57LT2D69h+orq7r\/ANELUtKFMpRzFKzZr6nu94jx846hpMjJqcsk\/DBW4bsB6xN\/Qkgl0cyZUL63uBDHEb+H6BKpfnC5MqUcjMugAuPK\/ZSBv+XjHOPHbtPUbhdKJXOhl+uJ70hRZZYW3LHopVtFL\/rHQfdHU\/PbiTxVxHxbrrtSrk646FKJblwo5G\/bx84rDbfLRmSLnraPKeSkdAIKp9KbkC+lgPGEDMlRGwO0FQ4oqIvYekApxRB11hErNxobGG6yobEgQRSb3BO\/S28bD2LvgzjvEchO1KUpjL1PSvnTjyW0dx1I3URrZRhnWOyS8mszyU8TOHjcmXlllyYriUqUjNcEpCTY26XMMT2YqHJ2+L418OWCNw3VFOkewSIBzgDw\/kwn4rj1hRA\/\/rSzz35EQKeEnByWI+J49yqiBtK4cmV\/I5tYBOCeAMmr63jHV5sg7S2GX039zGezDUsiamBKuqmJUOENPLTlK0a5VEdCRbSCqRmJASPKG1Tbyy4067+MKBpJCFWG0e5aEm1gT5QCm2x925ghyg6geo6QAGtwMx8oME3sNL+ceyga6GDWBuLQAaSCQUkXNtBAAIzH8I8pKRse76bx4oFzvf8AKA7p1sR43hMhKb9RbeBsM8DkGW41ML06b+CmUOllDpRY5HNjHUWFe0XhCXoMmzN01uXfaQEKbVLIXqN7HTS94k5XFWL5wjn1udVpr9YRHQXAluoTlNmJioTxcSpYCFPLzHzjXJubZpUvzlB2aSDYoYRmJ9ojf+klDKj8PhmrPkdRK2B+Zj579sypVPEXF+emZ+nvyLPKaDLE2NUoyixAvbU3jnxyQAuQlKQPBIgUS5Fx+kClpViQojygXWLixSlVvFIhsplSG1loqaURry1EAjwIjRMD8X52Qk5ejVWiSVdl2U5GXVnkzKU9BzB9r3BjR6dxGpCCMjVTo6yP6GaaDrd\/3kkH8IcT2IkT6c8uilTwtfVxTaj\/AHk2\/GKbiJ5bsu4HcMKTmBs7LEOgHx7pvGw8E8doxPQ1SLxWmfkwG3UuAhRHRVjrF5UvkOjzhZM8WKkNNF9Ys1JriqROIcCu4dd4v2IcA4e454cFLqtPaqVPRZ7IVFJQ4NAQUkEGxPzjNcNdmXAXA\/F85iCUl35spKfhxPHmokjbvZDbU36nURM4\/wCMeHJWiPuzkyZtlI0l0KCEq8j1IjnuW4r4nx9jOm4XotDnJLDNRmEJWqXkloZaa3UcwFhcX163jraXpr0gwhpDSktoSEpQBpYbWhqHpiYUoBYZ8yNhGW8bMYf6I4WnZSgyz9WxDOpLTEvLIK1rcIsDYdBv6CKbwO7JUlSWpKvYoSlurPNDmtB3LmJHeKjuom+pjrfAVBwRw4pSxJvSUu49q6t+ZSSbbbnpEhM8RKAtSm6dNS1UmE\/6mScS4R5qI0SPM2jHOIvaEZphclqZyqhUEH7adZZg+X7ZHjt4Rl+FcC414816ZmJOXfnspKn5ty\/JZG+UE9fBIjoLh52XRTWUfTsk3UHAAG2pldksm+uugJPUagW3O8aPTuF1KlviES9OaEtmPNTKCyFkAAiw0JsAIvNGorNNluW0wmUQdmGUgAeaj1MZxxA7OOHsdYpFenGzMzikpbKX3CEt2FgpIHX1iArXZSosvQp0UGVkmKqlspk3VlQyk7qJubKO14yc9lzGL7v170uCnTZxX6Qu12R6hNzEsapUUhhp1DxbaaSMxSb2JUbW9o3+UwFhqQk0fEvyLLxSkOAKbAygfYv4QSebwZJpXzKzT2wo6hU22LD9n0hgMeYEoLjroxPRmFKGU559vQeAF4p87xvwzLuKDnEzDCQL2DSAogezphmO0RgxKrJ4mSGYH\/USgVr7pVCg4\/4VecSf+keaeUo7tU8H\/wDVGLdoLtWyeD3Zim4ann63iN5vIahNEESzahcAAAAX3sB4XjiypTU\/iKemKpU5pycnXVZlvOm5JMJIbIA08oErNiNdoTWAAD4dY8pNvS8Ey3Ve34QCk62HdG8A4ALgmEFhI236CCbHS94TUQkjTXa0VatNIU+4FpFgq4hvLSjf2iCQR5XELuSzWa6AQkbEixPnbpBTKIGt9LdTCfwSBa5v53gfg2yTYBVtxvFxkm80lLjLb6sdIWDSk2Ive3WGdV0lj0sd4MjVhNwPs9esK8sBI2zeXhCSmx0BNvEQkvLYhOoOkAUlSBfUGAy22F\/C8HSFAEmBO4FttdtxAAg2Bvf84Hu3+zmt1MeKUkfs9N9oDl5j5eIO0EOifP8ACCC9\/TcQYJG34CC6gq39oVJQUG471tDaGyyoEZbkW6R1w5xIkmNBKMpI0OZcWHDnaTGEpAy8vKSC05isKWoki8PZrtr1lk\/UCnt+iCYZSXbcxAubIfm5TlEaBLOxjJuMHE9fGKtTMxOLbdmpdgBtSWsl0g3t52vGKuN2Va2o8YSKbecFCR08dIBTet7DWCKaBG2vnDZ2Xt3rajXSN+4aIwDWqPLfH4jmKdNhI5krMquArrYnpE1iVOAKD3ZFE5V3LfblxdHzEUyexjJgZZCkzbA\/rJ1\/OIqlYwnMPYgaq8tRptx4HU2KUqSdwdbGOmadUGcQUiWqLGYNvIC8p3SeoPpCi0ArQs6KESzK+ewADdVrW\/WAoUjimk1lU9SKo5Jy6Ekvd7u5evkY2HC+IXZjCyn55xLhdu8VOgEEdbg9NoinsGYNncSSdUdw\/Tp0rUCpTrCVKYURdK06WI12IjUmKuy1LlP1YbQnuZEgJIHS3SICt4ilpNkhCkEqVlRc3yqIJHtfS3nHHnFPtStyzM83JNEVJtwt5E3AuN7j2jJcCdrfEOD5mZmvoKTqFRmlArmpp9y9hfKEpGiQLnaL+32v8Y4gWl36Hw02pH2S9LLdUkeRKos1I7T+PFyiyhrDqWkqP2aYFC\/uYeVjjPijGtBlpabelZNCgrninshkPHMftAdALaRVZSXS9OyrazmQt1KVDxBOsdF0ziNXqDTmqdTJ1NPkmRZDMuwhCUj2TEZizi\/jWSw3VJqVr8wJlmXWtGYJKbgE6ptrGeM8cOKE42CnG080m2iWWGUgegyQRfFPiROOcpzH1bR1OQtpJ+SYD\/SvGk04efj3EZ12E5lv8hCzTtenB9djHEjl9wam4PyMLow87NgiYrtdmBb\/AFlTeP8AihccO6U+AHnqg+L\/AOsnnVf4ocN8KcMuWz07m\/8AaPLV+sPmeFOFBb\/qKUVpfvpJ\/OJGV4cYZYupFAkAf+wT\/CJRjCtIk0FSKXItJA+0GEgD3tASEzS11FUgyuSTM\/aQ2m2ZYtr06eETrNN0AKEJ18I4e7QlLRTeMtfRlALhbd8tUCKGFgJVoIKVbnprvBVWAFifSCne97ka32gihm6G8eVdJ3JJ6QRd06DUwi4v9o2v5wgp5sXHMTcdAdYRVNtAWvr5AwiZxBOyvW1oneGYkZrE1Ufm6RKVgS1InJliTqCStkuobzJUoAi9rE2vB2OKc8psBvDOEGhawth6XV\/xAmFm+LFaRfl0zCzA6ZMNSX6tGDL4s4pFuUuiMHb6nDtPH\/6IAcYschNkVxMuB1l6fLNf8LYhBfGHiB0xlVmul2Hsn\/CBB5Sen8QtJqFRm36hPO35kxMuFbiyNNVHUnaFSw5a5TsOoiLrjRMgqybG8IMSylsN5dRlEHXKKUq+toBcqtzvkEk9SekeNPWAdLCwMEcpxSCLbaCCJlylIT1F9QdYASytdYMWVAqJAt+cF5NzsEk7QJZNjsYMWQknUXjxata1r3vcbwQICFHNZQ6X6QUspAPrpBSzsTbTSADSSnX2EBykWtc66wg4S2QATteNqoeF0vLqBdPMUmZWm6jfrD1eE2hc5EWHlDZzCzZ2Q3b0hOTw4zKzaXFIQcpFhaEcSNNSWICsJCUqISqwt3SNYz+sSfwk882eitIj1I8IDKQekFDYUQdR6R4t31EEUi6drR6lycs7UmG5x1MtLuKyreUjOEDxtHQmCOH2ATKpW5iOmVEqTcs5FIJ8tTEo9hHDsu6o0uQp5WD3VKcKh+cNZ3h5MVoBU5XafIy6fstoVYD2EW7AcozhqXcpxxDIVMOKuyywo506a7xPTSlNLNri\/j1gZWoBvuEk3HQw8dxQwmS+jZt5bEm8sB15gd7KNT7QeUxMcfYvZbp7jkrhKjtWdeV3QtIFzf1tE9hziJIVKpTNU+KZakWysobUqysgBCdD5ARCzXHP\/R+lTErML+IVOkmXUTbKk7mK9U+JnOlZV2amcqStVQdSlf2EJGVsH1Osc\/4ppdMxHXKxUpfMyqemFTCS4oEpzG5FtupihVPD4lH7Ky+AKToYmsHy0maxLicnmqcwQR8QtvmIQoJ7pUnqLgA+sWKr8VEVSqcqWkWpGUtkWhpISgkbKAFrRa8LVmVmaXYTTV0LVe6xoDY\/xiUl69TmalKhdRlUq5qLJ5ybnvDpeNzBKtbXJ3hjX2ubQKkgjRcs4Dp\/VMZtSFpep8sEEpzJSoqB9ImpeVWtxtalEFJvYGwVoRY\/OH6GxmQSR3b6aaxIy7TPxDa82VSL2SFWBv4iJpkS61IcKkBQBA7\/AIxItKk87alvtZ06i7gH6w+bdprjrbjkxLlTZJSS6O6SPWHhmqS6pouTcrdpWdN3gLGxF9\/AmHSKvTBcfHyY06vp\/jCU7UaBUpZcrOVKVMs4QHENTiW1KTcEgKBuD5xzZxRwrW2OLTVcpVZl28OIeCmGpeohx5IH7Z0I3tpfTrHR+HcaUV6kyqpqsyLb4QAoOzSc1\/O5vHJnaoVKTfFVc9IzDU3KuyLWZ5hYWkKGYEXHtGQmaZQn7d\/QHaEVz7Q+ylSj6QVU6b91lR8yYTM48q4DSR1B3ghmJv8AZCdOg3hNfxRBKnLDx0AEMZmdQ2e8+onwCiYaOT7RIslRPXzgom1LHcYJHvB2viHlhKwG0nqoRMMS9LTSZhLpmFVXnILLgI5CW7KzhQ3zXy26bxKcKmijHhls2YTFNqDdx1vJum34RUpNP1KdRpbpDsi6BqAr5QUnbprvB7WA1ufPpCa0aa5T+sWPDRKqWbKsEPG3uBEspd1C2o8ojsQs5qU8diBe3lDWltlNOZ6KKBqT5bQ5LFwnvoAOpObWCcsH\/WIHkVCE7oQlV3kHS2qhCWdNtHUKsLnvQULacIHPbGu6lCCOuoS4AHm1AfeChHgptQUovtkJ374EE5rBzXeQB0uqCc9kac5BH7wgfiWAdHUehMF+LZFgHUg7WB3gi35cgfWpUfaCfEs781H96PJmGlAgLC\/HWPcxANyq2sEXMtFSsyiCNT4w1eeC1ApVmFrXGsa6\/UJ6ncY26TzyinTFUDTrY6gqsY6zq\/BbDrKxy0P6j\/aGM+4kcO6ZQcKz83Jl5EwyjMklZIEYrQZl1\/luOuKUSnW50h1iGXbnMQS7Twu06E31ivcQKQJGeacbR9U4gAHzEVApsB0Hh1ggTlJvvHrWHnAEX0gpF\/OE3GgtJJEXDh8g1F1ckqZpbBTayak0khQ8lRqzXC2lTEqpcy7TFOW0RSplQUfxtFJrvD5TRUGBWWU62U4zzU\/NJitS1GqeG6xK1CTqzaJllYUlLocT7EEdRHReHcSnE1HamXGskwkAO5Aoov5Egbws86EJJvaK5X6yDLraUuwIsbaG3hFZrOPphugJoVMUZWSVq8UGxcPW5isTldfdl0shRSygWCR1iHmq0lLiXZp0qQgaIJvt0it1zF89VluBKlJbcIzeYGw9BFddecCSNjC30NWRLMzBpk5yHhmbdLCsqx4pNtRHkyk62oIXKOggahTZuIOXVtkhSADa1inaHNOrXwE1mLLK0rSpspWmwFxa\/tvB5iroad7hazJOi27nUdRpFjovFaYlWS1PTs08sEkPc5zW\/Q69IslK4t\/EOouFT7bWVKkgFWYeY21iSrdZpeIlJdpVMn5cNJAUlgKQlPrEI3VpNA7zE8u3Uv8A8TCqKxJJsfo2YWd7qmB\/5oXTVZVae7RnDrqTMDf+9DtNdbcXdyiJcVoMyn0gmwt+1CorsqP\/AOAYJ65phH\/mhRFflr2NBlB\/99H\/AJoXRW5WxP0JJDXYzDf\/AJoMiuttqu1SJJo2PeTMtD\/FBE11oKOelU5RO384aH+KB+n23ElKafIN+Ymmv4wH0+3mAMpT1dP\/AHlv+MMcUVxmaw7NspYk0pOU\/UzCCvfoAdYzO7ZIsyq1+qhpByllB7qcx84HmpsQGxfcCBzEAq0T1FhBLFablRsNIRdlwsnMFKGxsTDGbpcqzKlxDSUquL31hsCAjKkBA3ItpeDpFh1PjAlAJgBYdRfqYsHC9ZRxSoSb\/wBKpxkj99pxH+KKlLHJcbWJFveF030\/WDFZUdb6wawzCxIMeyFQJIJiewyMkrNJ7ybLSQCPEGJYlOW5NiN4ZVcpdp0xltbLtGfqYK\/G3Q3hViXsnTUX1vB+XYJGX8dIMpgZRlTbyjp7gUGX5OhBxltfOw64g3SDctzrw+dlCOX3WENTsy0QLturSPZREeDKQNRAlgDpp4AQBYFiAARfwgW2km\/Q+YjwZudYPyU9E7He9ukELdrBKRp5R4N5rE\/asYc01OV31SdYkmmnJleRIOaLcOAmMKpJs1GSpyZlp1GZKUuJCiPGxMVR\/A9Wpry5abkXZV9s2U262q4+QjUsftCR42PmwBaqiVD+\/HclQWHW0Lve6AfwjNOKzXNwbVk2v9Sr8o5cw+pIl2yTpYw+xA4EVanODqlOvvDvG0oJ+glYHfZOa\/l1jLXEnwhvbp18oUIsNdz0ggsSRbXeAAzeUAkW0veCNuiUm2ny2l4NqCuWv7KgDqDHSWCsZcNq5TG0qp7lCmCAlTgBKArqM0WCoYPlZtoO0KvGZ0uEpeEUmq1bEWHJxLa0vAAlRW6yFJt6kGK3McS8d4vqDFGpbq1rX3+WlIQ00kffX4ARenp6capAfWsTaW0BD0y2jKlSxuQP2bxmFRxhJ1FxRbm21DrdVog5nEMm1mJfCieiNYhZzE63QpLDZQP2lRFmZceXdffPmYMEFTRVcC3S8A\/INuNlQdKje23lGhUvi7M06nYflUyqnJakS6pYy6pgpRME3yrUBY6X6HpEFVscVKr1B2bS1LMOOG+RDWZJ\/vXP4xU331PPuOOJstaionbWGqr3uQD1vBL5U62ghSkg6W8dYTDaSbJuPGAEvudT6GDpl0gWABhdqXSogWEXSmt8mjN33Dg\/Ef8AKDPLTbpYxX5nKJt42sSQdD0teGoN3Bc6EWuTC7LCCkEAmF\/hbja48IH4UEa7jbyj3wqbEhN7m8CmTSBbTzhRMuhOwsPKFktC2unpB0M8xwBNsxNrqNhCxlA2pacwWEn7Q2PmIJy+Z3RsPGAcRplBuPzgrTahcJ0B6QSelgZGYJ2AzfjEEEjTYjwg6QBrrfyMATprbTpBVXvubxMcO3uTxUwionQ1KXQT+84E\/rEDMtFiozjR3S+4nTyURBkHp1gc9gNb21HlBgAqxOt+vWDJsEnxv0iXw04EuTKbHVAV66j+MSjqwpdtQOkEd5aqc8hQVnVsrNYAWO4+XyihgAFV7k3MKMnunXr1hTQWsNRqTHrBXXU+EdEdn6ctK4ZSRoJapS+99n2XP8Zjn7ELXwuKq4zb+jnn0Af\/AHDDYruAL6bWP5QCSFJB69bwJsTpoB18YMnU3GpPhCuqyToT6QXJdSSTv4QC29bWseg1hPW+pFr6wtKrHxKTpc322OkOxMLZWFIUUrGyk6ERrvDvtG1XCjzUpVnVzMipss84fbQnTe24\/GOj6LjCk12mMTks+zNMuJ7rhAUfQ3N9PCOeOOjX0dxmqarWIm0r\/EGOzmHRM0mTc3CmUH8BFL4kM83CtUSOrCvyjk+igCWQPC8OMTKI+jnNQLW\/GLSlCJqUU0oXQtOUxkdSk1Sk282rdKiIYWAO1\/O0DYK18IKUJCoDKOo1hPKLjxJjy2wpJtprsYk8KYwmcGT61oZZnJN0gPSs0nM2rz8j5iOg8L0mgYhYTOBctSXHUhTSKdNLVcnrZSbARPVEf6O0x8zVYYelEJJUiZsb\/wDOMxmcaP1GiqRT5BmkUZxeV99hGR2cPRtJ3IPU7ARLYyrbOHuGT7a15Xy2GkoT951fQDwSPyjm9toAX8D1hXIFKKrBNzcAHQQZDYI0GkKJaydDChGguNDADawG56R4WSoEgE9IBRyJFvG1oQfTqVeIveGitdL2gik6k3N+loFSdOgI1EECDfYfKDlNgR+QgQgE2tqNYdSqLuJ00uIuCUhFCVpqlaSPx\/jEa4\/mSNdIi5sfXk7BSR09YbFIzpHmdYkpBvMyTbrDsN2KdgRBg0LXHXWADGoGkDybj2g6GLGxGbrCnJ123hRqVBUNBvB1s8xQSkEJGkHWzlQEjc6m0I\/DXVsAYOmWAGog65Tm0+csLjlKtfy1in3BCRbXpaABCrdI8bHUJNvCCqJJ9Yc4emDJ4xw9M3y8qoMOfJ1JguJ5X4XF9dZ2CJ98f+IYY20sn0vCiUX0FgI9kAGp9IPZKLnqPeH1DUGpwg2N21fZiaUcyTYj3EIPnMlWmhFiIppAzqBOl\/CFEt2TYjz1H4wbLcgaXgpCQL22Ebj2f5sJbw8gkZk1GoNbftMML\/wGMi4htCW4kYoa2SKg6fS6r\/rESpACb6G494BNjew8tYMUgKB3sNoP9qyraGD5reQgc19iQCNAekApVlgZvKCkJJA0KtOkCwAl9J2F7ekOFKGYgW9YbTneSk2BUFddoMioTLNw264lJN7IUQL+gjfu1DL\/AAvF2oqGmfIuOssKvfFYPozo1zSjZ\/3REJjdvm4eqSdyWVflHIVIIShSTc2UR+MOMS96mSS\/2SRFoprgMq0rbuiKZj2nBueE0gdx0Xt5iKc8VXF7ADpBN9CflHrjoNdoKb3ubfnHri9uphNYOttYQmUhV9L3\/GLVw+4pzuAnhLTMuzVKSo35EwnMWz4oO49I05\/H8hXv53K0+RbXa6HXgTl9AokX9orlUqSZmYbnZuauGiVcxZskekZ1izF7+KZplFyiRlz9W2ep6qPmYh0AC5tvA5bAE6g+UKhFze2ohTZQBGh3jx18iIC4B9ukFUQBtfrCbhSE79YTUv7p6+MNnNcx89oA7EaQU3OhNrwYBQG4PUwJSTofWPZCCOoMSVPRmcHhFrW3ahzPiEg29xFdKxbeGs0SrKR4EQgDZQ1ESlK7yXBvYg2Pn\/6RIpbuPLwg4bJGm1ukGSzoQB6mDpYt6CHKy48hptaiUMpKUA\/dBJJ\/EwUS9zpbTzgymjYjpe94O22RawvHuQdbmx6mDJl+7\/m0ELQX3QDrD6Ulc0u83b7SFD8IzQ3SbHW3SDA3Juc2vWATbIT0EEOpJtp4wk6+WJqUdFwW3EqzehBiw8SWORxLxUjYCpPEHxuon9YhEZQDe+gg1wkbbQQqBuc1x5x4LsNdBuRDqlrKZ1s5rXuLe0TyHDkFrCAWTtpta+8Up5eSadT4KPlpDnnuPH6xxThCQkKUSbACwHoBCWayrJ1No9munLvGt8CpkMim3Vl5dfCdT0clHR\/gijcXW\/h+LWJkAbvpWB+8hJitp7tr6WgUi1wdNbbwbME6jU2vB0qtY3tAFQBuTvbQR5QsE5jYwJuVDTQHQkQBUFXIsDcwQrNgOvl6w7BCrEa31hN9P1dwSdekJXy6FREdL9r2X5XE9blrZ5dJjovhY\/8AGcM8Pub\/AM0SPkIHFTRVSJweLSh+EcHTVemqZVJhlASWkuK0PTWLfUHPi8MSzxH3rmIyX4ktSTYZclyot924O9odu4qlMVUpxttstutKBsrrFSmBZZCtx4wM1KBhLRDzbvMQFnIdU+R84alJTcQGUBIG9oAnQaWgpGnUCE1AXvDZ1oKEElJuYp6ithQCvBQ0+UenanOVMj4qYW8lOyCbJHtCKQkHTxhdO51hQm+vQbCPJJve9yIOCCdNTvHrFR8oKk3J1jyvw62hBwjPe3tCY1F7W6g3hIgn31gFJ7ot42JjwToDodNIMhIC9FadTAkAG+p6ecHCCD5xJ0wWWDFuDWakzSBc\/UqPrYRTSoWuTCLxKk+UJ5etolaKbvOpHgD\/AJ+cTbbVxfTeFQxffx+cLBjYAbnWDhjUWhQSvWPBmxIt1g\/K8QPeDcnXT8IMGLlIOgHWAU3rYDTc+EDycliBe3SH9NZBV4\/51jJ5xvkzcy2B3kuKTb3MI30G5A849YBOu9\/aBCR3tRr7Q1qItLjxB\/SLVxYcH\/SVXFn\/AFy0Pi\/9dtKv1isJfv8AZIuI9zkgHoekeccAOuhgpcBBuR84cyLoRNsaaZwPGLEnuA5dfeBtptqfGKbUAUTzoPj0gZdYBJJym34wZXiAfHfSPJVmSDa9o0vg89y23Ttyq5Tlj+0mYR\/iEQHHVrk8Xqx05iGV38fq0j9IqDf2QfHQwok6gaEDqYH7PQWPWBFwkAgHWBSkpN7WHtA28bm\/nAWJJJzG2m0FBUkjbS4N4Kq6iSTYK9oWa\/o0gG9tLj8oO6Pqjpr5QgBcaqt6xvvaGxq1j7FjVRaYMugNBvKo3vEthLjnX6Fg2RpcolpDcukoStQubXiHrvGXFNTZWhc+UJUCCEi0Y9VlLdccWq5Wo3Ji\/wAmvnYIZJP2SCYzWojLOvafeJiUww5y0zRB1ABA94laiyCQ6NQvWI5QBMF3OnWC2yjU29Y90trbpBSoX8oQUNbHWCKtsQbwi4Aem8NighRAgALLGaHCVA6D1vA59QN0wZJsSnXxtBwbba9LQYnugXt4wAFra+hgFHQi1yfCG7ija1iTeEwoq06iAtY7QRW+5tHgSSTcgbQI1Ol\/WBJ7w07oMOkJCj1vEvLshpaDa4PhFrkW+ZKuNjTMgj8IoRBt6QRZKlAbkmwtBChSSbix8Ik6GT9I5dsyCB\/n2i0sNWCjbrDlLPUAkDyhRDPTwhwzK3O1\/wAoOtATZKRuNzBeSlIus6mEiD+zYCPFOxIhS4GgvqPYwCGhbUe0AsAanp8okKeO8CRYjwjKcSp5OIqg2Rb65RA231hhYg+XkYEAa208yYCwTe+8N59VpdXXX5RuWLOG+GcXNUKtP193D1WnaTJrmJGamJBXMsykJcSFTSFJCgAbEX9IgU8BqZMqyy2MEuE7DkSy\/wD8c2qDns2zrn\/u9fW8D4UeaWD\/AN2lcJu9mXEKQVIqbRFvv0ipJ0\/\/AMaI57s9YkYBP0hTbD\/aiZZ\/\/IwmI5zhFXZRwqNSoCy2b5RV2EKNugClDWG6Ffa1HsbwJFza\/S8U+sW+k3jvc7eMOaJRalXn3WqVTpupuoRnW1Jy6nVJT4kJBsIeO4OxKyPrcO1Zr9+QdA\/4YYvU6eljlekJpkgH7cupP5iL3wmS4mSrjnLcS2zM011SikhIImQAL+NlH5Q07RbfL4o8zfmybRNj5qH6RRmyCB+YhZKeth5AwoL5BpoTsRtAlNwFAWttHgbaEW08bwUju638LQRRKU6XF9LwC1X879CIUKs6EpypIAvoLE+pgZUHIQfG0S1PoM\/WOYmRknZpQQpRS2gqNgNdIh3Wy2spUkoUNCkjaNNxocz7SvKEaS+Po6x6EwR9LjwshpSr6aCGFYwo+1T\/AIhJUtagDkCdRE9S23GcHOtOoKXE2JSYps3hyfnHFzLDBWyo6EQpR6XMynxHMZcF0WPdiXDa3qcQUlKmjbUdDtEUQQqE1A9DqIITca\/jBc17X0PlBLkg6wRf2rW21hNQsq5OkJqPev8AlCC7q1O0edsT3AQLfeOo8YM0CpJAvprHs41ST0gWlG+vTYQoFE9APMmDm+a17CDJ26\/pBFG\/QDWEXbWOntCI3IsSPKDGxFrWHQQRNikggwdLYA01N+uggqiUkEgDyG0F02IvbqIetWzbXixyTXMYZVra9oslOSEgae0Z7NANTDzdvsrUPxhArykEDUG8WKew+zMpQ61WactRH2Q7Yj5wlT6aiWnm1qn5ZbtyAwhV1KOosItDLK037hHkYeIasRpYeFoWbbGhJNrQZTtgAjf0ggsD4+cFVmVuSYAjTX8OseDYI1A2gC0LWIPzgUggW8vlHtBbS4PWH0k3rt5+UZjj1nkYqmxa2cJWD6pEQqT0y2Hl0gSSbEa+keOu14bTususbePnErxElkvVWivZLl2hyBKiL3swlP8Ahis\/BNBXebB9RHvo9skd2yvG20OpZtyWOVt91rxKFkfkYkGazWJQjkViotW6tTjifyMSDWO8XM6IxdXkJPQVN63pbNEg3meQ24tV1KTdRO5Jg4SN7AmKlXGy3Ujbe29oWw\/iB3Dc4uaYkpWdUpHLKJzmZBqDcFtaCDp4xaGeMM+yDloUo1f\/AOVqlTYt6Wm7fhDhnjjU5cp\/m9ZbtpaXxTUEj\/fcXE5S+IY4jUmtyU5N4ok5iTYbqDKHq6ZuXfLbzd0uJU2FC17ghXSIjtKNj\/TemP20ckQL+i1fxjOZdYSN7AbWMPBcqubXg4sAdNLa9YEXJ+17QYnUHW3UQVSSonx8hCeW5VqfaPEAm5Gm2ggE2vcJ062hWUsc5SnKAY6C7M+OaXRK29JVRCUPPJ5bTpT4nVJHnp8o3ut9lXC\/EOpO1tpDcp8RYqQqZQySd7lOa2oI1F\/WObOKuGhQmJNaTfPcEmHfCnDjVbo8wtYBUh3LqIvScAsJSO8U28ojKrhJthvIHiQRsN4p1XkDIMzsvmKgWwoXguCKE9U6Wp1DiwkOFNhsIveHcIJQ+rnrNim3fTpFW4j0WUw65JJbcJ+LWUG4sNtPxjMZ+WLLygd4ZHbwhNSTbXYawmnU3I0gFG\/p08YTUdTAKSSDCSk5k6aGEVgqTf8AGCE5beG0ezXTc6wIKSfU3gwXtY9d7QYLuLbAmFAo+J1MKJObfWCL10Ihu6AEk38ocypCgnbvC0NnE9\/XeClJSLJtqdTeDBIAAG5gr7ZZdyq12N7wk5r4bw9Y7wGYdIuFEa5rAsk6axYZVFjb5gRntca5NYnUbfWqPz1hioXtfWEg2n9npDqk\/V1iRKRoH0X8hcRs6aXe\/cF99DBxTdhkFvEQY0\/ryzr0tAGmINrpVePCltdUnS3SPKpKCRYdd7bQmaUkHum1tDCSqWNDmFvOCmklWxhJdMXsnfY+UEVTnLjr6Q8lae4nQg+MZrxUl1MYlZUdObLpVfzBIiqpta+u9txBidNCL+EDe6bC9obzNuQq2p8LRM44JcZwo6NSqiMp\/urWj\/DFfTqkkixOogcx6i\/kIXaF7bi+nlCwSAAARc\/OC2te4I9IsUorNJsq\/q2NoMlRB6b9YreIE2nrj9m+nWGCUkXFvPSFLpN\/LxghbzJtbTyG0WfACctQrKUj+ko04LeNmir\/AAxY+0SgTE5hmZFvrJNevj9k\/rGWMjKm214ftuDInUi+wHrC6VA3NzeDkJSm43uDboIBK9dNSSdRAFQJUQdB0EEIF9NTfpBQLEKOv6wYkZRYW0gyRlHdPUabXhzKzcxLzAeQ6rmpOYKB1vHQGBO1AugYdZkamyuamGiQlz+rpYRM9oijiToEi9c6O5bQ07OcuifptURfvIcB\/CNj+hmyhRcVa17RU6mhhl5aFC9utozTHIQqoO5U5UrYIgvCB5DtKnJdea6Xc10xpEgyA4qzwIy310jKuPbiSqkKadzlGcmx2OkUWopE3Ky82ixS6kE+R6xDrAJ9DCZG+phEgpJVvBFG5PhAW13EAry94RcOhEIHTUa3hJR2F9PWCqVlOkFz7G+l4PnsQCNBA5yrZJIhYAgXV3fxMLpdA03JgMw16QktNr218j1g0iqygNhm0PhAzKMj7ibAgncQllJUBpttArAsRbU62vCJ7x1JNoKnW\/Q3h7K6oTbW24MXrCiboF9B4RNMAhw28Yo2L2S3iKa\/rBKv90RCKT4biC3udRCYdLLqHBoUKCh8wY6HRZaEq6K1FoWSDfxPpCgPX3MKoTmPU9YcoYzaZR7iBW0i1igE+YhJMk2rXlpMEckmbn6sAwiqnskEWIgi6c0TuR4wRNKSDYKPkbQu1TiLd69\/wjLOOMiqVnqQ\/wBFNrR8iD+sZuFED2vCiSNIE7n9ITmE3Zc\/dMS2KdaLg9YH\/wDGLTc\/1Zl4RBJFh0APSB2sbga7dIOi422tqBCoWbawUqKlDfTzidpjtpBF9LEg\/OHJIvfptaK5iFIEy2SbEg3vEcgkKAhbS1730g10gA6DXeLBgI8zEZbSbc2RnG9et5ZyLRxrAmsM4JmstwqXKb+qEGMwQjOR1vofSDgd63zB6QYqynUm3+dIOhRToSSeoMKd0nQ6X8I8m5tvePdSQnU6Wv0gVXy31J6ekCEi5B6dTvCgIURsB0tAJNwfE+EFUcqiCnNY76R2l2m6OZrArRlEF5xLw0Gpio9mSi1GnJqpmpRTSHVJSnOLe8bXPy65ZS2lIzEG2YbRUqnRnHC4sZV36DeMo4ny65KelS4MuZoiK9wlU88ao1LJzLuD6Rdnk1FM0ptSVoBFrgaRmXFaXflpyTRManKSLeEV6iOCbp8xKHRTR5jd\/A7wxcbyq1HtBFAEQivYQisFKvPyjxT5GCkQk4LHWwMNVKsNRa3nvCKie8ekIuvJAsYK2h55GZLTikftBJtDybkZilIaM9LuMc5GdsLFiRDdE60DlBsPSFg6laRlXcjoPCFm1XTe2loUCr621gDcDpFowJxMqOBm52Ul5Km1CTmyFuMVCUQ8LgWukkXGh6eEVuZXzJpxegClaAbD0hK+n6QRxRJvrCZFz4W8oEAAdRbaF5SyVWOxPyi+YUWOWTvYRMyqgog236RUsct5Ky2u187CTp5EiK2sdAflCJIBhN\/Vs31IjoLD7gmqFTn+rku2o+6REmGVK6Qs3Lk23h0zK5hoBpCqsqU2FoTyEnMU6eceUVX7oIGxhPIU+vn1j2TKBcW9I8GrjfaPcu9tvUwshoaC1htGa8f5W1GpL4A7j6mz7pv+kYyggW3OkKjY6Wj17iw006GCq+wryHSJXEF14TwYu2glplu\/pMuH\/FEICCARtAXubWH+doMCBtsB9oQOcHXoIFS7KOYkk+cTNEVmk3Bc6Kt+sPb5STbbqTFfxATzG1HTTw2iMQoBQ2062hVKzYXG5gb2BNosGAiE4ypQ\/bcU3t+0lSf8UWvienn8LMGvEE5ChFwf\/pn\/AMsZi0vQEXB8IUWvbMbG\/QQKbEi9yQbi8HCABc6aeEGNjv4ddoAHa\/pYGDZhfQ7nUWhQEXA6+seBKQAU+d4MF2zJtYDy\/CPZiVE\/iP4wmtV1GwBHiY+geNpUTcgGVKCBmve0R2D2ZqVL4LoU3awtE6vOq6nbnzB3iIdmWm3FgJO9tYzfjHh2ZqkvJTcs2e4opOh1vFN4J0ueoVaqy5qQeKF2SkhBsfONylEsTCbOy5BJ0uNY557RjqE44ZYQCkNSycyegvcxm1Im\/gag06o\/V3yrH9UxJ1aXDTuZIuhWqVdCIjDvvCarDW8FOouCCYDKdtNoIEm2nzhFYK15UpKyrokXJhlNFTK1JWgoWn7pFj8omsNcOcS4zcSmmUt1TSv9e93Gx53MbRgfstqpM6xP4ifaqWQ5hJMn6u\/9YneNenKHJLYRKIokmwwBYJQlIAt5RW8ZcNcP4qEsqdlmnnkI5YLV0qSPaKBM9mijTC1BE6+woquDuAIg6h2XpyWUtyn1lh4C5Sh5BST4axkKmXZZ5xh1JbeaWULSehBsRApXZNiLDYQKlJvtqISJKHEqte0KFXMXmIAB1vB0gFQHTe8JL0J8vKCbAWVv4wY6kE+wHWBQrK4gjSxsRF1wxMWbWdNAdIsUgbpAOumwivY\/btNSTni2pPyP\/OKkRc7WPQwmQR5XhJ5KchjeeHjgmsFUlVgSGi2T+6oj9ItTLKRa8OENlVrCwHQwKu\/t3R5QGQ9RtAqQBe0AEWF+kBYAi4GusFyjS\/4x4Cxt46QbLYecLNp8uutxFD46yvNwMl22rM02o+9x+sYC2bC\/4WhZIJsPDwjwOviTBTYi4108IlqzZWAsJuC\/cenWvktCv8UQI1uBoel4UIBJ0Bt0vHtr2OoN94Ak5k+enjeCkWBNze0StCVy230nQgg2PSJK6ktAKIUo3N7RA4hSU5CVX6RFJuLWtfwhVKiQABt7XgUG+h384m8CrAxrh+9yTPsC99rrAi64\/wC\/wbox1vLzQQfKxWmMsZPcBNvOFSSCD9q0HQbg301uSNLQYqUtdresCPDYnS\/hHkr71iQNLXgyd7dbXEHHh77bR5ICje5t5mDpPe16CPLJTqkaEXvCJ7+ua3kVER9IK\/Ll2WPNQL300iPoVIU2tZHeuIlDIutC6mgR6xDVCVusnJlJOwEHakFvNFK0l1I2zQmzT5Zp5SeZyzuRD5pMu2RbK6OoA3jkPj9Pia4pVU5MiUBDYT5BIjO8wubbjxh\/KVVxtvlODmsfsndPpHlKYc1bdCf6qxYwiGsysrag4o9BrCRKE21v5QC3k6gDTzhPnk6CyYTTNvSbvNZeW0sfeQbGPUerLRiylTc0rnK+LbK1O97MMw3vHZczNTK0hUuEBIAAQ2iwt7Qi6\/UkjO4lTsvuUpXYiGkzV25l1tl1p9hA1Ckq1vD5idp7KEpMwXFpNwXU2UD6wqXv5xzOai18wAF9YXXVfigA6lT6irTKkAfhHNvHjCSqFipNWZllMSNR1OmgdH2h76H5xmmYKSe6fCAIAT5mCqtfWDE3AIO0Chw3OlhALtfWAAKjlJFukGSL2vrbaCZrq8TvFmw3NDlK9IuNPIUhIB08SYh+IKbyki4dLLUkH2H8IpdwTaEyrLvrBXblPQAdI2vhA9z8FMoGpZfcR7Xv\/ii8oVZNz0hYLuL+VgIEqvsdIG9xe+kALXubQOyb9RBTbbr0gLW9BtBDZOw18R0gQoBWW4Jte0OUE2HSKvxalvieHtW0+wlCwB5KF45qb12hVJ11gxVlIINjfoILe48T4RKT93OHFCVfRuqTrevm2wf1iCzlJsdNNIMFDQBRPtA7k+WpNoKVlOgG3WPXJsq9j6RI0JQLj6NypN\/xiV1G5sPOIfEN\/h2zoDe+sQgISRY2tCgUb5joOkGz2SNB7RIYcfEtiOkPD\/VzjK7+ixGk42bV\/wBE1QZIsZWqOpt4WmFD9YyBknLcaHeHITa2oFvPWBIJSQk+Y16x4que7fUb+MKNkEEeMeGVVyBm8DBk2Njb9YFKgpINh84FRuLE2PkOkAlZV3jqRoCYEquncgeBhArAJtt6x9Sq5Tw9L5kIAt9q5iKpPJC1pC+9bbwh85OoacOgyjQqMQ03OMB5WRpZCjoSYScmwqyG1C\/UWhq3JFTi1Ftta\/EmDsJcbfTlbbSiOMuOUxzuKleXdNw6B3dthFHCxpC1tbAg6QX28oLLr5U0lQNiOo9I8pVvDSCKPdNvtbmEivu3va52grhspWlul4jJl3lOtLvYpWFfiI76oUk65R5Gaabs28yhYUo7gpBh+1K\/fmSlpQNtE62j09R6euX79ltq72ZO8Q7+GmBzC2rKFJugqtcRCs91PJfSFKzZe6m4Pyhd1pUpNpQ0AgjX7ERPFHD0zjfA9Ul1HmzjSPimE5bEKQL6eouPeOQg6FDrfrfpAlQJFzcdYKtdhr1gUK7gte28KJWFJ8\/PrB1kHQ7+kEKSkCw1gdgLbwTLdQ84kqK+EOlH7XX3jRacRkRqdukMMdtBVBbWBcpeSfwIihFNtOsEINhbaCKTc338I1vgdMcyh1Ng\/cmQux\/rJH\/ljSs17ePpA3sdb67GDpVqYG4Go1Ee9BqNdYMTcA6D0Eev3bWgCMp8oAJtp+ECEjfc9DCzZtrt5xGY0lvi8HVpm180o5a\/iE3\/AEjlNrRI8ocJJT4QPl+MFNiBofWJWZzHhZKugENsV15KlAaDMw1bXpfKYromUkaEE73g\/MTvcWO1+sCl7KCAbJO+kFUR7CBvcnytEhQ1fz4AWAKCNOukTCxc3MRNdSFSe\/WIBKrK11EKhd0jUeYgVKNzl0EHlnSxMMug6IcCr28CI2LGCQ5gPGzV9Gas8of96lX6xisuSLHfS3rDlJvqBfTqN4NcAaAkgwdYGTQXUfDpHkAqaNrq8SBtA\/ZAyki949mykWN7jXSPJVun5+UCi5AA28fCFHHCtISNh5aQVSirW+h00husWIHLKrDePqbUi+ZQtkXR+1FTlXFIm3AlsoAO\/jEmGlPDbNc9TCUxI\/XBtWiQLhQ2iPRL5VnlOhYJ71hrD9uTkX7XU6FDexhdtlpm4bCiAepEcD8W3y7xLxAr\/wDtKEVZK7qGutocpcITtr4wVa4TzWWnoekGUu+l\/aElOWIJ0hNawTcC0Gd7ylAXvvaIuojM2T4b2jv3hziRqu8OsOFhK1JMk0FKuBaybH8REuqVfebLiXipoGwKyLfOI8S04w4VpSqZZV0Qq8LNpQXEBfdWPuOpt+MMX6T8MHFMEpBOYlJB+XhEpITMmqTDTgS64jXvWKoZTBSZlwoRlaUnu+XiDHHHGLBxwVjeaZQ2WZCaUZhhQGhSTqB6G4+UUjmf+sJqftfzgyXiAm5FjCzbg8bWhYuZjY9No8FX1B0tBgu9hp6wFwCD7DWH1KazzTeVWpO0adTZF0IRoFC3SG+MZdSsNzIUNUlKh\/eEZuE6ba23hNWtx4dIIoaeg0jSeBs0UzlYYJ1U225YDaxI\/WNYChfX5wdJJSSd4NmAtffrAgm+sGSD1NvAQJJIFtY9e17knyEDcHyueseuTfTWPA3NrWMKoXlAFoLOtCZp80191xpSSPG4IjkRKChSkbEEj5QoNra6eECpRNjuIKVEja\/vEvhfiNXMGyFQp0i9LOUyddDkxJzsgzNMuKGyilwGxHlEj\/0mU2ZsKjgPDU0kjvKl2HZZR\/7p1IHygFYl4fTqQJjBU7IeK6dWVi3ol1tz84DkcNJwJCahiilLO4dYl5tKf95o\/hHjgvCk2oCQ4hSySfuVKlzDNvVTYcECOFU88f8Aq\/EOGKmVbNt1VDKj7PBBgjnDXFGHUfSM5SiZBv8ApJqWfafbTfS5U2pVhe0I8wlVxudvGI2snPJHu3ANz\/6xWgQo31AvYQrm19Da8HKiCTqoQRSzkJubWvY+MbbVh8ThPHyLghTpmP7zba4xGWuUJ18tYcoXruAbW0hQ2AGhtbWBsbAg6jWBJsAdbEaWj1rDMPUx46G5F\/baAN0k3ta97woXAE2SNLfOCZhmtuPSBSvKoDXb1hFSluG6QQI+pk2p\/wCFyBaVE9BraI2SBZeKFIzk9THqg0Q6OXdN98sNHVOOy7gIUDskW\/GGiZFxlfcOY2+yIUkHgl1xBbcR4lJiRkJZiYdypSouE2IJ1tHA3GSTVTeKGI5daSkpm1G3rFPSTca2hdCrp3vaBUo6EjTrCSngoi2ggzirEm+8I8w7eMJrWL+8LOKsoEGxsPyhjOatr8BHZHZuZmKvwdpS0sqUllTrPMHQBR\/jGmCTElIloNLWDqc57vtDH4l6TDaQQ0yndCdzDtdWQ4wG0slVhe5Tcn3iMqE6llgp+CzhX2sq7GEJOqSrriGUtGUURYhe5hWzRmi1LvNOLtcJLhtGa8dsCPYywqpbUvkqshd+XQDcuI+8kfK\/tHIxWUmxBBHQ7gwip++g1UeghRWZsWWMvkYUbeykaQ4Dt8otB8xO50geaAb62j17kgD3iSpjoRNtm9tdo1umvhcqkjYi94SxCyX8OT4INi0SPbWMnQoZB8oKVXMJqPvrF04OTPJxY+3mtzpVQt6FJ\/jG17EdL9OkLJ0Gt\/GDg6bW8IEaiw18oE62O0DfU9NI8CTfSAuU9LDygwjw311EGQcuusLJVmF9PSOSqzL\/AAddqLJuA1MuJ+SjDZK9Dc+8ezG\/Tbxj3Ujwhi4LvOAnrcXgSjOQQDtr4R4JSR42O1tIBSAb6X84TDSVKINt4KlvKoZVqR6G0L09biJ5i61FOaxzG+hi1loA76DS\/jEfV03klnwHTwiqNuFJCsoUAb2Ox8ovC8YYJmSfisDpl1EaqkKq+z7gLDogefw3m1KAGJqWsmws7LTSB8w0YR+gMGTpWJTHZl1H7IqdKcb+ZbU4Iv8AT5yRqdPxzLU2eTUZZuUaSmZbSUpdIlgkkA2O6TGJS6khAIvmtDpCgD5bXvtChVY6\/ZB1gQsLvbToRaPFQtlA08jAF0EaXvaxgvO7iu97R5Cj3tQL7QPMVYm5HWDFwlWwsBtaPIStw2aSVqP3fExoFJ4E4mq0iiZCUNBWyVpUTbfoCOsfQEtmlpXlKnlq6Wg1PbE6u7ysj19kxJilKQPtpvf1gr9NaaTd1Qyq6xGTUmhkAyykozHUqNzaI9KHGHDco73QdYey4KlpWyrI4DHBXHt8PcXsRnNmPP1PibCKClX+RCza8noY8pSgNdoTJHtALXY2384SW4lJvCbalPEhCVOK8Ei5hw6q2UK7psLg7w1WeZmSkg32vGx8IOMVT4V4dal5KtMTUpzFPTNHmkFNio2+rV1Nhcx0fhzi7QMW0P6QdrTMilJAXKumykHw8xE8phirsh+Sn2H2QLHlqCrwlKsuyAWtxouNnZaTfTwhnPkPrW40hTSlnRKE7esRq1O\/FNc9KXFXygqAuB6w4XUJRDy22FpZINitKQT56w9RMyE06gFCR\/WUbmOXuL3BKcbx9MzdNSEUScAfLwAshR+2APXX3itP4LlaG19UgrWN3nN4oWJUmWqCkEWukERFsv5TcnbWHzMzn2sSekOksOrSpQbIQN1HQQBIRoVXsYUSsACHUo5Z1B312MapR3c0m0bi1olZlJep8w3p3mlJt11BjFQvS3hHs9zqfeClzMN\/SLHwwmeRjqn62Dgcbt49w\/wjoWXbLitbDSFVDKoa6x4EEa6+do8FBWxtBkiwOvXSBBFrW9YFO1hr7QUkqNvOPG5AJO\/iY9e9xeFEnu6a\/pB20kqNtRHLvEBj4bHVabtYfEKV89f1iERqk3I8YG9jrp5QF79IavJVzFKCCUmwBtvDmj0moV+eRI06Ufnpxd8rDDZWojqbDpCdYlX6LUHpGal3ZSZbNltPJKVJPW4MNS4FgWN9r+ceC9xbU+EAbaa38IM0sIeSc1rKB0MW5w5gDuB0hjUgVSbgsE6dOsVJAvp4dR0guXvdLQZKBm7oBVrYEXgiQpHc5qgBsL6fKNN4JozSmKpcWAclgdBvosbe8ZqwpQQE6DzELhXeHlraFM5SAojUHcQCXsx0HvbaALljYnW8AF5gFFQGtheA5mYaDfcQbmgJ2AvpYQYXSQcxI2ywKVBYttbeLrwip8rV8bSctMpS42tWoJt\/npHaoovwrTSWkrU2UBQGUHL5Xi7OtOzxs24pIH3lGHcvLNSbac84CsjU26xIyM\/Lo+pLp0Gqz1hKoTbZbV3sxv3R4wwmMr6czSsviIZ\/E\/AvAuNhxZ2PhD2WdW732095XXYR89eOGdHFnEecd8TJ222ilhzvDwIvcQulzMBp6R5Ttr+UJOTASLJsTe0aDMcCsWS1GRVJluUlJFbYdS89MJSCki4imITTKasfElVRdB\/o2lZWx79YncM4jqU5PoZkpaVlZVP20tsi1vAne8IYs+Fkniky6DNLuSRoBFWLy1E2snyAhRicDSgHWUzLZ3SrQ+xiWlXJCaumVnHJB5X+qfPdPoofrDldcxNQFBLM4+hNrJKV2CvfYxpXBDinWGcWCQrlXU9JTCDmbccvyyPvXPytHTjVPYnmEzLDyeQBcLJ1N\/SI9VOafWXGXsywfvIOkQU\/T5tqYUQGyk7BvQmGKn5iUClLacAvbupzEGEJyrJn5FxlzMtYFgVDYRlFfl0oecyBTiidLjQe0WLhrwKw9xPwrPTVSS98e1MltLzSrWFgbee8Gm+ybSKeslCJmZTuApwiK5UuHUnhbMGKQljLs88NvcxlGJlOIqDjaQopOpUDYX9IgSsk97UgwZK7bX94fynecSPPSNOoBPwjYPTwizsIC2Tp3rRhkyC0+6jqhah73hG+YeAMFzHKB4eUSeD5n4XGFId6CZQn5m36x0kh5SDubbwuhara39YXRpbS48I8ADcecGBIG+kCL65jHs1vSPKO999oKdT0FhHtRew19IPur+EKI1NwLjpr0jnLjGx8PxAn1W\/pENrHge6B+kU\/7u9gYEa9BfygNLeA8I1Lh1MLewm6hFEp1QyPqCnZ2aWjTQ\/ZSOnjfrFnpEhPU+stuyNMw1TpsJP10rJrdUgHoVLX+YhHiBg3E2Lkp+k8RU+ZT90M0plsp8BmTrGQ13BDlCmm5P48Ts6+bMstICeuqla7Q0mcE1uSLmaW5qEi2dChb8bRHzNJqEulanJN8C\/2gkkD3GkMHFqSO9orwVpFubUrlBSTfMkEDbpCT6ry6r9R4X6RUAQCQCd9YLzDptuYFLlrEDJ4Wg5zIcLah3hrY7xpnAlwLrlYZ\/blNQfJQH6xmi7NvuoJ2WQB6GAD3duBr1MDnvoTm22gzajYpIJSdLDqINYZjfpbrePBJI08fSABNjpoT7wFyeu51g+YAk7nWASom5sbbxY8BVY0TFVPm\/spS4Be2mu0fQnC1akqvQpV91sIVltYi\/n09YlUrcaAzBIRfUkwvmTMaoUj1hIMzDWYLUhKCdD4wqmWedCRfMEnVUHcaSwgWbK3Sb5r6QTncxtTLjeZ5X2fKDy7ypNCs9k9No+f\/HGnVJ3ihX5oU2a5DjxUlYbJBHjeM+utKrLSpCh91QsYUbe6H5wq2lc25kaQVL2AEK\/CS8kQqZXznR\/qGzt6mH2I8bVrFXJbqE865LsJCGpcKIQhIFgLRDS7K5uZal2U53XFBCEJ3UToI29\/Cktw9w8lDikqdab5ky54uEap9toxWo1FdUnnZlw\/bOgvsOghsFi+pgbi28FIBMPafUpyVs0yovtk\/wBA4M6T7fwh5Jz8iZxEy2RTZ5CrgOArZPl4iN3wfx0+Bk5eUrMsqRSmwTOSp5kuoedto16mVtqtsIflJ5l1gpBC23Ab3\/KJCXliSVqur03vCT8xLSzllmx2Ui2YRFrRLJmnXHAlaVWKQpH+ekYvjWVm5qqTzMurkS7ayCpAJKh\/kxu\/Z\/S3SOFLbc0044wZl0odYSA5v1HXWLJUKrLltQl3ays\/sBkD8TGL4+cWh5RXJuqeVoDMuc1wew0TGB4qpEwuqrKwpLihcpA18hDKXwRUZ5pa5YNPLCC5yi4A4odSAd4ry2VtKzEaecLyc+2w4FOggDwF40rDlapy2Wk\/FNpWbWStWU\/jF6kkkpFtU7RhteZLNfqbYGiX1nX1iOUbEi8EzGwg8m8ZaflXtuW6hYt5KBjqxtlOVJJFiL79IULqEDT0gAbi8Hv08I8CAQDB+hN\/mY9c2sPDwgSLb6dIINDrtAlWmh0g6SVEW9PODpUCv8IwXjzLlrGMq8NObKp\/BREZ5\/m8HtYbj0MFNtdPxjVeD0+WaLOo5xQfiNElQTe6RsSPKLm6TT5cTM28GmUC6nHHrAeuoH+dIpVd4jqrDhkMKsLm5omy5tVwy15i+59YHDlFlKQtU9OvzE\/V3NZh9abFPkL7CJ1WKpJIKEMBy6hopBJV7jSGTs1LTpJ+FUxpYcogpPqDEb8EyuoOJmEc6TULozAEDTbLFImSlmcfaSAUoWpKcpsAAdIbzCg6wQlVrdDFPU9ZakgXNzrE9L4Aqs7IsTUslp5t1AUO9YiHVBw6abUlqr1LdelOUSLKWUhWlrlvXxiGxE9ILqDiqe0402TYIUsqA8gbA\/OGslV5+kv86Tmn5N8ixWy4UkjcjTptDa61qU44SpZJJJ3vfeDtjvCx0Oh0hZOhvkKladIVtnsLHawuY8psgAkgW6GAUnYkn22gijmJ3Ch+MFUcuxJHrHuZvcEnzgSs6Ap226wqZpaUaW062F\/GOhuE\/HqSo2DmJKqqKpplakgqNjl0t19Y7ILKHArupufKG7cujm5bZVdB0MPHKe8m11XT0Frw2IMu6VOlZBOgG0OHZ5KyA22L+cJKdSpIccZCEIN1OEwd2ZbqDIdBGRI0AFhDBp2UmQpL0mw4hWlloCj+McOdo6lFXE6oLkmEoZFkJbaRltYeAjOUU1MqlLk86G76hpGqj\/CFHKjmaLcsgSzO3d+0fUwwUkJym9yekHRJzD4zIZUR420jZeA\/Cx5ybViuphIlJNWSWaOud3x9BvEXx9xckzaaIwoFQVzJhQ8eiYqfDrhbVeITqnJf+bU9s9+YWND5DxjUGuA1HkkAO86aV1USU39oVR2fKVPq7q5iXSdAUKuB84jZzsxPrUoU6tIXbo+m34iKDijh\/WOHMyh5+clQtNyhbTozHpoDFMWFTJ7qFOLUSe6LkwMjVJykPkyrq2D95J1Sr1SdDFqw\/wAQ\/oyZS4C7R5k2vMSZu2r95s\/pGy4R46zyWgipoTPyij\/77Id7L5qRuI0lnEtHEgquzdQpsjS7aOzj31ij+4NbxlmOO01SJUuM4flF1R3\/AG7yeU0PQbmHlLxL9L4YlalMlpmYmWwtwJ0Fz4RvnDHJL4HprBLjazmcWptN73JP8Isk\/VKPTWFqmZ11JA2WnKIxLiBjykS7ylU5pVUqCiQ2022ShB\/aUesUSkYPmZ8OzlSaIfcUV2UO8SdbwhWsPqp0tLplpBtqYSsp+PCSpxCCSSANra7xR61w2D06USClvFepU2LpB6+kCngbVX5b4iUezLA+wsbGKnV8MVSjzBlZhJM0PtMrT0HU+UXbhFi12dcfo0zdTjaCtkk62G6faKvjhr4XF1R0tmUFW9QIgFWHe0MEUtNhbXyGsJrStSSUp0Gupjp2kTapqlyTpJKXGUKBv4gQ\/Qq41OvnCiVAJ39hCgVe29\/OFE946jS0LBsjcWG0AVIR0+cJKcz6ga2j2Y662gArw2\/KFEEkX8esKhQsfLaMZ7QTFp2jTG5U24jw2IP6xlAVpbU23g17Ea6HxMFWLDpr4xN0DFcxhqnzgalkPlxaTmdcslCrHXKNTFio0p\/pq8iarNQXWUgaSaFFppnytca\/nGh0pyk02XTLy9KLI6IcSUpHsBCi3aQ5cvSz4UoE8tF7DTcWgDJyrlMWiXdWoJGdLWibq89N7dYg1SriUnKGUOKXbKASRbc3\/WHb7Tcu0XOYVueDYOX8DeMtxGkM12bv3c689jcbgHr6xFpWVhSTbXS8VR0lDzg2AOwjbMDzT3+jFObCAPq9PPUjrFglqg0y8UlAuU2KVH57xGzQp\/xK+YDfXuhQy36aaRCVOhSMwUDlNNLOxKb\/AI6\/jFbqWFJdTwstSCkkagEfpEa9hnlqCW3Wyrpc2hqun8kArUE3OpEIllpsXKlAbg5dDCS0ZlkjRB1AJ6QkmwISc4FtxHslsx2se6SDrBQ2lR8ttYHlpG2g23gqUjLY\/MwV4pSg3uNd\/OGma5Ns2\/RVo+u05IMgJLa7EeUM1MBkpKmwVgXCrQBqz\/NCEy+YeJgk2+46gcxkAHaI6aWtxSClGUDcpEIv01+bSfrLtfsRBYjxbSMJ5WqlVZamgovkdVYqHlGXYn7UODsMtq+CfVVppN8qWR3b+scw8QOLlQx3W5idblm5APHZsXUfeK5J0meqDgytLKla53NLxPy2C55tsuutkNjfLreHEtQ5tLiENUtxy+6imLHRcGV6qz0tTpeUQ29NuhptC1akk+EdE4\/VJcIeHrUutSVN01nljLpzXyO8fPX8o5PwDgqo8YcYOuvLLcutwuTMws6AX2EdkUDCklheWZkpNTTUoy0AANibQ7mZOTmW+68kOA2ASAQoREuU56USlTRCknZF738oeok3Wcin2kMJUNAo2JjjjjLJVOTx9P8A0oouNrUVy6hfIEX0AhPgzS3qljph1kKUmWBWUIGYrOwFusdGVng7JVeScm8SStOoUqrvCanlpZUB5AamMI4k4Y4b0Rjl4YxDPVepJ0UEs\/ze\/ko6mM2Zfep74dlX3Jd4bKaVYxNDG87MNobqUtL1NKdLupyr+Yh4zX8NOpV8Zhx0D\/8ArzFjf3EaZwaXhrE6GJCZlnHJ4uK5MqtxSkIbB08rx3BhJMtKybcm0hvloSAlKkgi1olJ6gUya0mZQcs65k6j5RX61gKlNMKck2GgVX7yUi8UheFUNOFx7KlI0iv4gk5CaqLMqXksy7aeZMOEgaX7qb+dj7CFJSp4VlAG+ah0i+VmWRnJ\/ugxGVmpzjjgRS5EUxtxQSl6dQAQOqgi9\/naMj4q12j02mPS8j\/Pam8TnnXNXFn\/AD7CMgweiaolcZqbjolUNkkqWL3B0OnWDYqraq5W3ZoBNiAm6ARmt112iLS2FJBvdV9j0g4ISbDTyEP5ajTUybcrloI+2sWBjfcJXThqnIUbqQylBI620itVfjNh6i1Rci4p99basq3WUZkpPhvr7RM0fiZharlIZq8u2s\/cfu0f960XCVUw+0lxD7bqT1QoKH4Qdc0hGiE+UIKmFOnck+EAFlfSPcwet+sBzO8fDyjwXt0hRKynx0O8K8wm2mmwMZdx\/Z5lEpj4F8kwUel0\/wDKMWQoFOmnWFL2Gm\/SClQI8hDSbIFlE2HjEtSZhqmMSFRp044JkuKbnJJ24Fhay0kaFJva24I6gxdGcVBaElRUSBqvNYD0h21i5QWlS3nRa4SVKJh4rErSZY5XylsDvqUM8BJYlW4s5Zhh9vYXGUgRJtTrc6OWhBadTskgJBPgDcfhGb41S6MQL+ISpLi0JOVR9v0iHQvKRa+m\/nFanTkm3gB13HSN44bz7b+CJBJmWkrQhScitMhzncw4rLssAUuLQSFDVtwHc9fARCvol\/rFtpCVJ2KNbecRK5uZU8tHNuPI5cp9NLwadyrQDzOY5ayhpr57REz9NQ8guBBBGlidQfHS0NHGHGEHO0FJIFgRe3mBDNyZZbA7tyAQTY2+UMXXmwghts5VDS40EIhSFJJBAF7AW3jzUwWTmCgB1BtrCThUpSiSE3OukIlQB6aHYQNwkm2id7H13hOZUFMK+z4QvTMLVWuMKfkpGYmmUqyZ2mioAgA2uPWPre7XafLyIdDOYDTe8MDiBNQ7yGEggaDaI1uslxS0LSlK0+B1gjlYUopQqwA6wT6Ra5nLS5zVn7toUTVuXdC2sqU726xyH210CZrlCm0iyCypA+cUDhNwAnuIcsmfmZhNOphNg4r7S\/SNxpXZ4w9h4oclmvj3Ebre1v7RbWOFdJnqU+0\/KoS4pNkhpIunzinJ4Fysq6sNzU4pNjbvaCIeZ4UPS6yE1aYSlO9jr6RtXZv4B\/RL87jqqTT0yiVQpiQS\/wDZ5hGq\/YRzv2gcRTHF3im1hCjO55CScIedSe6VX7yj6RZ8P0OQwC0imyyxyEJGZ1IuVq6kxZXqu0yy0S8XkrOgSd\/WGn+kT\/xBbYKWbdFAEe0LMYmdlJkuc\/Ltds628SIa1fH6Kg242nOh091K1b38YybHVBOIHmvjZhx5ANw4ekN8N0x\/Abv0lhzmKqABBcQoBQHleKNV5DFOL6i+\/PrmnypZP86dK8sLyXDCZcSkzLxQDuE6RKSmEKfSiFLKHlDa+phScpUjOpCSw2R0IFiPeK1X8JzFMYTMNJU6wshOguUk9DEzwvlp6nYqoTranKe6uayLU4gpGQ73v0jtunVSrUxA5rMu4T9laHgkK9jCtU4kVmUaBVIsptpczAtFGr3GWsS7ZMu7JybuYBWdWca9bRVJjEtQxA6n6Qrc7MBR\/o5RsNo+ceZZlJl9x1ynOLlgrIFTz5IKR4g79TDWr8WKNhAqYYcSXLf+709CQfS4jMsUca6pVQ6iXbRTGnElGcq5r5B\/ARnpqjylFxFytRuXXe8s\/wAIaOOLdKlOrUv97eHslRZipKvLtkotqtRyget4lmcJ8sBT8wFX+6z194dIl5WXWEtyyQE7r1KjBEzTi1JS1zLk6W1MaHJYjFBwY8ieZeYcS25keCRlub5et9z4RzyprmOLXcqKvHrBVsIIOmo6iFZSanKYQuTm35VXiy6U\/rFoo3FTE0g6lCqkZhrwmEBX47xeKZxvm0JAnqay7lNiphwpPrY3izyHGSgzHdfD8mo6fWN5h803iwyOLaJVCBLVOXdX+zzAFfI6xJhYWklKr36jaDIOn6wtzUmwSMot4394Ok6W673ihcbZfnYLLltWphC9fDUfrGEtt505r2B0vHk7AW948oJFhffcQVUuHkm4vbW0EZlEsrDlykAEH06w9kgXWCQ4SANjtDxthdhmG\/gekPZSnLWtADzaEE2OdWg9YfSKAhxSFMsnN0IPzvAGXnpdfcXlRfRKiSPaIbEJccn21uG68mvzMRZcIVoCfO0QFUN51w7XFvWNBwPMKeoTDDLv1oUruKAtvtE+xPBWdtRbZcbOQmwVf8fzhtPzXPUENOF0nQBCiL\/j5QgptsLJnBMNLtYLTdX5mI2fmQynLLvKcN+8HNDaIj6ZmmlFOdwZuhBIMNzMTDSFOmZQsbkFeo8rXhqupqd0XY66KJ1giXrmxCkgbgHQQ3c+0SlSgdxqTpCVtCdR7wmom4Ic0PntHkXJueg3MGJC1XtaJfDGFZrGNdlKRJW50w4BnOyB1UfQR23g\/AspgnDsnSpFLaW2kXWpVrrWd1H1jXfo9DTaGeZsdjrDmXZQwpbmq0jQg6CEKpS0TqUuS7AaV1UDvBUUlbDeZQ5nd28YZy9NnJlRcbQWsp0KhrBvgJoLus3BNiTGTcdsEU7F8mW56QeVMSLZWy4hdkqv7xHcHqJWsL4GnZvEEq7TqXLqSiWUG7py\/tRMYKxtIYnxHUKXLTCZtqXQHEzLQsFDwIi8qKZfKWnQ2tWlz0EeefS1LLCXEzClbgCwEMaHg1zFNUlKdKFKpiYcAURqAOpPpGldq3iPTeBnBluj0xxCp5TPwco22dVukWUu0cRcMMPKwpJLnprMurT3ffWRci+trxY6mylhxp1t0KcSq9lb69DDeflPh5YTT74bBOiUHa8VarTS0FRaMzfTKq+kMxXXA63zXFIsm3mYkkTYbe5jl9RolXnCc1VHi2pSRzGx0MQYqi+alxtK0rF7I2ESqKy6iUWlKAXVkbDWI+admy6FqWrbVKj+HlCJccK0JDQKjuMt4M49KMkjl5l7FAB36xY8A4lYoOIGH3VOtSt7Op\/I6g7R1Hwl7NOFuLc\/MY4rVWmHJWxbbcS8Fpv4+XpGJ8SMeL4X8UJ+ktvP1ShB5cu1LqSFqFhoUHcekMkcSqbXmx8Ph2dWbkFTtkJJiPq9SeZZC\/gJJjOe4hH1q7+d9IgZvFExKslMxMTEwvUctghtKfDURTJ6rTE6TzSooB+wlwqPuTEC44EZ7ZWR4g6n1MM5aTdnCAwwtz+tayfmYnJLCTrneecskfdaFz8zEuzhlqVDZ+Cy5tea7clXpAzkg7nKG86UJTchZ0MNlNPS5AKgA4SAkG8PpHDb0w38RMOolJZOqnVKyj5mEZnGFEwzYUiXFQm0apeXcNpV466qii1yu1HEUwXp6ZLgB0bHdQkeSRpEelBNwlOvgIU+Bdc0SNrDeAXSJrXK0VAdBDJcu6hwpUhVxvpAZ1ZdLgjeFUzTqbHMT6wo1UVpSMyAfMaGJORxjUaVb4Sbflxf7KFmx9totFM41V6T0dcYnUDo63Y\/MWiz0zjxLrIE\/TFtm\/eWw4FfgbfnGk0PEMliSQTNSD6X2jof2knwI6GIjihL\/F4JqadylCVj2UDHPDbhS2Uk6EgnTrArWAQYISbbwhNuIKDmOZd7DKdQYTyc0pRddra3OkTUjJpXlCVqTpYXGkTUow\/kUhFlG26B3gYOzL3yhQWp7Y50fpbSH6WZZSEBZsoaXSN\/KHy6SsNJU042oFQGVKrkadbRXsX08sqlphIUEG4UfPTSKxlS1exsLAaxAVNV5xdt\/SL5gJpt\/DrhKsqm3FWO5T5gdYeOqQ2ylTiUuErJ5ibAH1\/9Yj5lbSgpSAptZ3WgAEH0EMm6tNpKEv8A1jQNio7nz1hWanmnAoryC\/2boGvuIjAsLcul0JCdk6m8IzLrC8yFiy+htp7w1WkNJ\/o9CNVIJtBFvKQyQkJIOx6wVL6HBdbhB8oBTiVElCrpGuo1hECxJIuI8rLYd\/1Gv4QBUb6736R1J2YuHCGqa7iKcSC46eXL21Kk\/e6+Nvx9tAxrxtwzgyuuUuenLTLaElSGgpQTfYG3W1j7xsxn3LlWW1+p6QlOz7gSktKKvHwgyatMFASgk+MPGam+HWwpSkjqCInZV1Uy6my8iQL69YWmWpe13XgkefWKTxMLTGDKvMSrfPmG5ZSmza+ojn2e7W85iXhfJYYmGkNONKUh+4+7aKB2fKmZPG9Zn23OXJlspAUdDrG9KrZqDKjzkoB21iMM9NLWWmJpQGykEb+kWbC\/E2S4VU+fqNQfUiZUkgKULZB4DzMY83Wqn2gccOYqqxWKNIKKJCWXqCfG0WiosuHM38IoZd1ARX3JGYYcUosKUnopUIzFMnq0UoaCCE6IRsbwwmpGbpyeVZD7o0XfXLEDOsFCipxLTTxOiSNYaokJh9alPKKlKNgQYkPoeYQ0A02cx3JOsEbofMKS8VhKTsE7xIMScjLA5mHXV30zaAQ6dcYUQoyIWmw7oFvxhBSpRtzOZZxhV9ARcGHaDSO88ZWzpPROhhnVRQCuyEuKeXY6jup8dBFJw7xjxDwwqNZo8rU5g4fnXw8phCjluDcKT4eYgV4qa4j8U6dNtLcQwwCourTZTiz4j8PaNhkaa0kTxTNMPOyzaluMKOQpQADcA76HpFDxBiqTl2y3JpzqBJKl91A\/UxUubUa+tRl5Z6dSBf6tGVtPqdolKRg6am3EpqE01TGj0QMx91HQRPKwhhTD7ja3+ZMur1CivmE\/oItNQwXTKdTmZ1TTi1LsUyxctpa9yB6xBOVhMqBLttiRR4hHSIp2fmZkqDU4slByggaGDfATBkFrn3mmpc\/acf7vyO5PpFZm8SU2lrUmQl\/jpgD+neFmx6Dc+8VSt1udrHenZlTiR9hoaIT5ADQQylabNTywlhlRA6kWET0pgjKAZt4A7lCDtEpL4flZfLlKWm7aqKTqPWHD2GJKcACFKLltMhuT4RE\/6MJbUTzXQq9u8uwPvaGExRZuWfUeUlwp0BGtvxhN2QS9q5LoPmdDDZzDrEwn6sKbc6gC8R7lBmGkFSO+m37O8MHJN1om7Rt4jWElkJJuCD4WjyFamxA0+cWTAGMX8IYil5nmEyTqg3MNX0KCd\/Ubx0Zi2WTOYVqYQQpLkq4QRr0JjmNtRULbesJrmEhWVIKlXtlAuYBKHF\/0h5af2U7wD0u2lAygJF\/DyiwYcwu\/VnAUD7WgWbiJeUwi+4442BZ1KrBCh9r39t4dN02bk1kLS80pBy6JOh6wq464vKHJhJCTstFzD2WddUQkMNP5j91Q99CIkWkoffShyUBufstA5h7gxXOIEqiTpaEDMDzwSnNcWsfkYoqwQ33Sk2GpG8QdVH84Kt9NrxbMDPpbpD4UALOHwPQeMScw8hTFh3Rax\/8ASI9KilOVOZWUaJvvDJTmclLjZJJ1A1vBJhtAcbTcN6a5zcCGymFEnlqChc6AX+UNJpMw0jMoFXTvJHzhm2\/zVb2A3hyUrcO5Ub2NtPaE1soZAzpIcJ1T194QUpKzoooGuggwutQ71k9QTABSUqOpKbbmJ7B1HXiXEUlINtKcU44lJSnQ2uI7Hx9iaQ4I8Nmn2eWHktfDyUqNMznp1A3P\/OOG6pVJisVCYnZp5cxMPrLjji9SonePrEZZgKSkgHMbHyhw7QW3FDkBPL3JgjUk3KXTygVX+1BHpK7nMDZKYXEiqyVlwI628BArQ1zAtTgcy7i2kRWIp9TrK22GEFgoIWCNCLRw7jThOali2fflQJBhxZIQBpE5gPhWaG\/zHKkCN8idLxrsk5ISkiQ40nPtqbxGVavqapk2JSVSHeWotvDdJA3jlyerVbxzWGJWozzs2tx3IEqOg1jpvCdNYw5TZWQl05W2EAb6E9TFhffQsBDt09D5xFVinOutFLTayrZJGotFamKdUaUgzamylIO+xhJszE2oOhoqSo\/WEDeHcvhVirPLcmEtDW1zqQIXneH7LZbXKzTTpO4UbARFVKhPyU2WwpZYAAUpANyqJul0WzAUpCdNkq3iZRSaenKHpdOdWoUIRepCFvWbZS00VA5nNNPSJSYwwy0wFANraWdFEXtEbMYVl2SOamXcQdCSPygUYSpM9JrbfkWikqsFoGVQHrFCxZwjw\/UFpblpZ9xxOwvqfaLFw+4aU3B04FIog5yrXWs5iIkcR8OX6\/XnpuXnnqZMuM8tbUsgKSU7X1iPb4FUankLmGlT84ixUuYdz3\/sjSJnEFAbl2WA2tMlLoascqLXHQWitNUiTnhlUgrbSbqceNgo+kSFSoNPqzLLbEuzKFhQcW40jvEDpeEpyr0mQSlcxMh11OiWVDO4r2iCqFVbnXVTT7DVOllf\/MJ7x9ExW5\/FTFOb5dKkgtX\/AM3ND8k7D3inTaKhiSaOZb9RmSdEIBVb+EPJLhRXZ1CnJlgyjdr3tmV8hEjKcOpemBK3kh90dXNb+g8YkHafkSOW2gKHdyjS+nhCcvLsuP2dQoHRICtgIcvScqtDgYmA2qwBQrUEecRCKI0ylLmYofKu4tpREA\/TlIRd9S3HQq4DpufwhF6j\/ETdypJz\/aS6cpT8vaDNYbW1nUhlbyVHQgggekRc7Ry22oiXW0ojMVIUDEZ8LONKCUHKlI00y3hJmQupx5bYIOpSLm5t4Q1dpUs4ta0pLSALqSpMRq6SxMKuwM2hFkg7wvTsBVCtOITLZUNqOri9kiNkrWKXME4VRT56fkqm\/wDDhpDOVSHlDLbcEj3IEYIhp14ZnVctG+RB\/WHKW0MgBAAG3nCzLD1QcCGW1LVcC1toteHMHN\/Go+O+2PsoTqkH+t1jRkyiuS22VM2RtkRc39ekeblnmElYKFFKdSpvL8yN4I+pUy7ZJRc3JKRax+UR8y2qalVMTcqjO5\/rkIufXyhinDkzLqzNgqbBJGxF+hiPnVzktMufaQE2UCBp+MOK283inDQkTkFVLgKDyso08SPERTZrAlZkEKWWEO63Ibc1+RsYqlepc3JuoD8s40ojdYsD7w\/oVRTT6Y42E5ipea9\/LwgXqwp5e9wP2hb8o82ovL1cSUDe5It7wslxAR31JCiSRcE39YSfnWUWBZKgetrj8doal0MrUE2yfdJ6eUFeUZpC7fWFA1NtoYplm5i5Ucnibbw+anHZSUWxLKShSiCpakAr9idh6RHOIUtBCgb7gkwTIReybm+94FTWYDQjy6QRTawNhv0jYOztLylLxC5V6grI0w2pVztew8eusVzjhxNmeI2J1uFxQpcuSiUZ\/ZGlz6m0ZwpBJOuYDQG+8fWZE8ytwuLH2doe0ua5zd0LATfQEwM2+tUyAFC14l5VoFpIXokG9z1gzraXr6Ap20iMmqe6jusqGU7gxA1tDrEs4hxzKFCwtFIewahmRVOupDq1mwSREDN0iRp0sp19kIUq4FohgzIJCFqBUtWySdIS+inaoHWZJtK3VoUENhW+kYFL4DruDcTs1Or0p2Vk0PnVZAO+lo3anSj6VcxoOKLoBTnGgEJzTE38QtLjuYpN7AxKyy30MFTk0GidjfrEZV36k4zo+h5tKtlfe9IZyVPmZh4trXyJZ37Q2ynwEJTkq7TkONtuqUQqyUD73rDynVGoTKUp+CuSBqjSxi20hl2ZWGFt5VmxUSnX5xK\/6LNOu8pCLLBJ5t7ACGLtMXTXFKcKXFINgUd42h1VmZapsyzqVhvmAIUgDUGFpWiGli6zz0lIypUq6TDl+ekxL5DLJbVvdKLa+RhilUs4+6ouuZQLpTm39IsuHqF8dmfLCGQUnK45YaQeqUGX+ISgvMOFGpKV2tAyMrLSL3NLZcCgRmvpDWaobUq+qaC1pU6klDCVXynoo\/witGgqKnSS2prdSnDdR+cMp7D78w0j4WWbZStVkqX1T4kQAw+qlZZycWy9Lt6OskENqB0Nze\/vFA4gSlHo8\/LDDcuiU5red0c3nLzk6hJNyB5RXZDAuIsQOFwSLraSbGYnboHqAdTvF3oPBimqQpNYmlTbpuOW2ChpPgdNTFspeCmKLKrbkJRqXbB1UkXuPGFU0goBU0ChtzvKWOvT2iNmaQwlpV2mX175XQL69bxVK9ghpvI+Epa1+0hZJ\/zpDBNHXJy3NTKl4WvcptEbN0P4pC3Uy6gUm6tL2iDmZKZZUotLNr5e6doZTSpixDq3UG9uYBbNCTMow5MqfemH1qUdwdSfOJ5x5qWY5zALKlp2Sbm\/j5Q1mZhDjaHFlKnHTqgD5nTaGM5SJOYCVNczm\/aUlC7gekBMUmVXlCFvMkEApWbG3l0hm9QFuu8pZGiSqy9CodIMMLy0myZqecRKy+55hAH4b+giDquMfhk8ijNclG3xTie8f3U9PUxUpgF1anXVlx0m6nFEkk+ZgJaRdnHQxLtLedVslvX\/ANImJTBcyhQcmlJQgWJQhVzbzizSDDDLfLYbQUoHdSUZVA+e0SUgHFPLdRJAu274AJ023EPxOzEw4pbrb7CkJA5dgnbawtr6w9M7MCTfstaUDZRTufH0gzKJqbQpSFpmEI6oTZftpDudkUuSKCqRXm2VkOUp1Ot4bSlLbm30oUqYZKUZwdSL+t7e0SE3SpmdZKnJtTzYFi2o2HppaIJygo5hdLRZULKCki+2w8oesTky6y3crdTqkFOo\/WM24vNpRJSDgTle5ikqTfXpvGaspW7cJF\/SJWVpanNClSdN94kGcOqWDkWFDe2oJg6KG60pIKLkmyVZTB\/oKd5ymu4gHqtX8dITmqA7LvJMx3FWtcbEQkmUWwRy1HveWn5wM2htS8xaSGwACpwWNz6aQgimJU2oIcUFZtNjpAOUV8C4Wh0nZANiYajIhYbfZKV7FB7to85ISjrmqnGidgRcCC\/Qa02U3MJcBPpEo8\/PS1PS1zCGyLEJGnT+EVyqDIEjRPXTqIZAG32M\/mRePqk7INvH6pasid4RdmiytCGATbTTrE7T2SWEl0KDyzp5RJhtzIEcwhXS8Ly0wpDpbcToN1Qut9pQKiNBtaKpimXS61zAo5Ui4EVtmoO\/CltdlEbDeIJ6nP1ubLU7LKRKJ1zCJbC3DV7GlbbkJRhttpvdQ1VljWKhwCwvg2gzFXeeVkYRdxTisigryEcWYpmGsfYpel23lqlGHrpKlkgAHSNUkJcPyjTfNHdAAuOkM6lTULcKUGxBuVAbxDz1HmpsFppYJOuXa4g9IkG5OWW24hQWDcqOsC8DlJzEkKvmho68+htxy4sLk3G0RlLxG5MpUrmKbUk2SLWuIstIxDUHcyVZcgsQTvE9K1gS7i86uYHCMyfA+UO5Z9h95QSS0Bc97aHkxSUJabWhaHirvAA6CEG6bPLebHxKUMDW3QQlV1usTZSh3MlKcoaCQb+ZiMwvTHH515c2QttJulsC2l9hFtfm2pbMsulKU2AZ2NvKPc1E6pLrbeTSwWdT6wq3JyUtZb\/MUgm5uog3hWaflZlP1csvME7k629YZMUZKlpmCFOADutqIAhkTNNTTq1lOQapSTewgyqlzJfMGklSjl1TdJ84bGh0ySdD8rJSrE\/cLW8lAC1Eb69PaF3cQuPuoK5NTgOgUNQoesSDryEMJU82Su5\/o9BaI2bKktF5lbqZdRBWSm2vgIj2EGdmwlM3zW89glkAKAPj4Wh67Q1Tfef+vSm9jfXeIU4WUQMpWRe9lm4MLOUp9LJaLRUncZVbeRgknJS7zSkOU9tt1JIUpatT6xAVPA8tOla22l80\/ZSCMoPjeKtU+HVUkQpxDedsKvZatD5CK7O0KZl3S3MShJvclvXKOh0habwwthlKJgLl89ilsquT4RGO05TZ+tSsHYFSbwMozyV5UuZniCElCRZV+lostKw7UphCnJl1DbKQLpcRa1vOKtiTG1NoTy2JNIqlQHdvf6pv1PX0EZ1P1eaqk0JioPl5WuVB0QgeATsIZPPB58NNguKVshI1iwUPh3OVHIua7jSjfIm9h6qt+UXSTwk9TGi0xItBq+6FamHjchKuzCkrRZKGwVIU3dRPUA\/OCql5ZDCFsK5KXRnDiU2UpPkf+ceLTYVzGpvknqEm2f8AeHWClptZUHW1TCSLpcSbW023\/SGKUDKskvMlOxURluPKFJZlycliuXL026Valuwy\/jD9qTn0Nkl1Z1sEEm6fH\/N48tU9KBIWs2J2cF0n1JMOi6+20VoSyRYkpQq1z6dYBM67mCHc6SrXvgWHyhNhFFC32ksKYOXuOsqW0c3UkJNiIp+O8PyU\/wApLk3MzLaTfRxJJ9dLiIim0anyCeW1LIsNczp0t5xIlEm0ruFleUXskajxgVyiHFBXKbF9swtc9ISU29ldU01ZxGlt0\/jAyyZhMofiWG81ySsDRMKKRKTwVzUJWsJsAoG1\/wAog5uiOB9QQCWSbhSAPlCBphbCmnSpROllDT5RFrpbTSgfh0Gx3Hd\/L+EOTJtrQjMFBZslI136R6clGg4QpxMxk0OfUk9esMJqUDaHDyQpO9kX0hsmTbVrzFIFvYQ5TzpR4IQ4h1tX2dQqIDEr4mHm1WGbYi22sRbdgnXNvsFWj6yzQCZI2FtekIyaEhto5Re\/hEug\/XJ9YGZJ5253hFxarnvH5w8Y1aF9YYVdIMsu4B0ippQlKyQkA36CFVKPMULm3hBuFM5MS\/F2mBp9xsKWAQhZFxFo7YU9MtyFTQmYdSggXSFm20cbcHhzK5UQrvDKDY6xudPSnKNB8obThsvTTeKop9wVEWcWNf2jBKk+4ltdnFjvHZRhzh\/vySirvHXUwyke\/MTebvd07w7kJdr4lB5aL5N8ohfIlL2iQNOggtNJVPG+uit4nZEXdX7Q8klET6UgkJzDTpFvsFSzdxfXrDZ5tBbUSlN77284hW+7UjbSzotaAqqQp1u4Buo7xOqSESyAkAAAWt00gsx3mVhWot1gZMA0ZF9dVQwZec7wzqsDoLwE8AJsaD7Kfyh6yy2mkJIQkEL3AhpU0pAuAAcitbQ6oiQaQ3cA6Hf1hadSPgSbC4GkP5BIfkMjg5iEtaJVqBtFKZbS1PzORIRc2OUW6wpRXll3llai3Y92+n2h0ix5QAwAABl2HqYIEjngWFivUWhrPoSHNh9nwivT6QErAAA30giJl34Sb+tX3UXHeOkU\/GCEopko8lIDynQFOAd46jcwo8kPU9xTgC1DNYq1OxivqaQXPsJ+XrCuHpdo1EKLSCoEWOUX3iI48zj8vh9pLTzjSVrspKFEBQ8DGFMdfb8oSqJsy4eo6xb8JstiQlFhtOc6lVtTqYvcq84UobK1FvN9i+nyhalTDrk3NBTi1ANJsCom2sJKAU08CLjMND6wNSSDN5CAUDKAk7AXgH2W0jRtI06DzENkHKUAaDKdB7wjO91MzbS6ht6Q9wWtRnENFRLZJOS+l\/SLVjmXal8N0ZxppDTjjLqlrQkAqIcUASepEQa1F6UQHCVjINFawvLoSG5UAAC40tDaqEpWixtca2j2IGGmXGg22lAKdQlIF4rM6kFomwvbwho6w2STy0XtvlEKTKEppyQEgC9rAf1YLLsthTVm0i7ZJ0hYtIbUcqEp1OwtEVJtIIJKE3J3tAKUQ8QCQMp0ibZaQ1SZRaEJSskd5IsdTrFSSolqcUSSRksfeEJhIM60kgFJzXHTaAq6EpZbIABGxAiMmgC0i4vqPyhrLk+O5P5w6p6EuuZVpC05hooXHSIifQlM2\/YAWUbWG2hiDqgBS3fXeItKiCqxO8f\/2Q==\" \/><p><strong>Ken (seated) and Dennis (standing) at a PDP-11 in 1972.<\/strong><\/p><\/div>\n<p>\n  A 1974 paper in\n  <em>Communications of the ACM<\/em> <sup><a href=\"#Ritchie-Thompson\" title=\"[Ritchie-Thompson]\">[Ritchie-Thompson]<\/a><\/sup> gave Unix its first public\nexposure.  In that paper, its authors described the unprecedentedly\nsimple design of Unix, and reported over 600 Unix installations. All\nwere on machines underpowered even by the standards of that day, but\n(as Ritchie and Thompson wrote) \u201cconstraint has encouraged not\nonly economy, but also a certain elegance of design\u201d.\n<\/p>\n<p>\n  After the CACM paper, research labs and universities all over\nthe world clamored for the chance to try out Unix themselves.  Under a\n1958 consent decree in settlement of an antitrust case, AT&amp;T\n  (the parent organization of Bell Labs) had\nbeen forbidden from entering the computer business.  Unix could not,\ntherefore, be turned into a product; indeed, under the terms of\nthe consent decree, Bell Labs was required to license its nontelephone\ntechnology to anyone who asked.  Ken\nThompson\n  quietly began answering requests by shipping out tapes and disk packs\n\u2014 each, according to legend, with a note signed \u201clove,\nken\u201d.\n<\/p>\n<p>\n  This was years before personal computers. Not only was the\nhardware needed to run Unix too expensive to be within an individual's\nreach, but nobody imagined that would change in the foreseeable\nfuture.  So Unix machines were only available by the grace of big\norganizations with big budgets: corporations, universities,\ngovernment agencies. But use of these minicomputers was less regulated\nthan the even-bigger mainframes, and Unix development rapidly took on\na countercultural air.  It was the early 1970s; the pioneering Unix\nprogrammers were shaggy hippies and hippie-wannabes.  They delighted\nin playing with an operating system that not only offered them\nfascinating challenges at the leading edge of computer science, but also\nsubverted all the technical assumptions and business practices that\nwent with Big Computing.  Card punches, COBOL, business suits, and\nbatch IBM\n  mainframes were\nthe despised old wave; Unix\nhackers\n  reveled in\nthe sense that they were simultaneously building the future and\nflipping a finger at the system.\n<\/p>\n<p>\n  The excitement of those days is captured in this quote from\nDouglas Comer: \u201cMany universities contributed to UNIX. At the\nUniversity of Toronto, the department acquired a 200-dot-per-inch\nprinter\/plotter and built software that used the printer to simulate a\nphototypesetter. At Yale University, students and computer scientists\nmodified the UNIX shell. At Purdue University, the Electrical\nEngineering Department made major improvements in performance,\nproducing a version of UNIX that supported a larger number of\nusers. Purdue also developed one of the first UNIX computer\nnetworks. At the University of California at Berkeley, students\ndeveloped a new shell and dozens of smaller utilities. By the late\n1970s, when Bell Labs released Version 7 UNIX, it was clear that the\nsystem solved the computing problems of many departments, and that it\nincorporated many of the ideas that had arisen in universities. The\nend result was a strengthened system. A tide of ideas had started a\nnew cycle, flowing from academia to an industrial laboratory, back to\nacademia, and finally moving on to a growing number of commercial\nsites\u201d <sup><a href=\"#Comer\" title=\"[Comer]\">[Comer]<\/a><\/sup>.\n<\/p>\n<p>\n  The first Unix of which it can be said that essentially all of\nit would be recognizable to a modern Unix programmer was the Version 7\nrelease in 1979.<sup><a href=\"#ftn.id2879903\" id=\"id2879903\">[15]<\/a><\/sup>\n  The first Unix user group had formed the previous year. By this time\nUnix was in use for operations support all through the Bell System<sup><a href=\"#Hauben\" title=\"[Hauben]\">[Hauben]<\/a><\/sup>, and had spread to universities as far away\nas Australia, where John Lions's 1976 notes <sup><a href=\"#Lions\" title=\"[Lions]\">[Lions]<\/a><\/sup> on\nthe Version 6 source code became the first serious documentation of\nthe Unix kernel internals.  Many senior Unix hackers still treasure a\ncopy.\n<\/p>\n <blockquote>\n  <p>\n   The Lions book was a samizdat publishing sensation.  Because\nof copyright infringement or some such it couldn't be published in the\nU.S., so copies of copies seeped everywhere.  I still have my copy,\nwhich was at least 6th generation.  Back then you couldn't be a kernel\nhacker without a Lions.\n  <\/p><div>\n  \n   \u2013 Ken Arnold\n  \n <\/div><\/blockquote>\n<p>\n  The beginnings of a Unix industry were coalescing as well.  The\nfirst Unix company (the Santa Cruz Operation, SCO) began operations in\n1978, and the first commercial C\n  compiler (Whitesmiths) sold that same\nyear.  By 1980 an obscure software company in Seattle was also getting\ninto the Unix game, shipping a port of the\nAT&amp;T\n  version for microcomputers called XENIX.\nBut Microsoft's\n  affection for Unix as a product was not to last very long (though Unix\nwould continue to be used for most internal development work at the\ncompany until after 1990).\n<\/p>\n<h4 id=\"id2880014\"><a href=\"#id2880014\">\u00a7<\/a>TCP\/IP and the Unix Wars: 1980-1990<\/h4>\n<p>\n  The Berkeley\n  campus of the University of California emerged early as the single\nmost important academic hot-spot in Unix development.  Unix research\nhad begun there in 1974, and was given a substantial impetus when Ken\nThompson\n  taught\nat the University during a 1975-76 sabbatical. The first BSD\nrelease\n  had been in\n1977 from a lab run by a then-unknown grad student named Bill\nJoy.  By 1980\nBerkeley was the hub of a sub-network of universities actively\ncontributing to their variant of Unix.  Ideas and code from Berkeley\nUnix (including the\nvi(1)\neditor)\n  were feeding back\nfrom Berkeley to Bell Labs.\n<\/p>\n<p>\n  Then, in 1980, the Defense Advanced Research Projects Agency\nneeded a team to implement its brand-new\nTCP\/IP\n  protocol stack\non the VAX under Unix.  The PDP-10s\n  that\npowered the ARPANET at that time were aging, and indications that DEC\nmight be forced to cancel the 10 in order to support the VAX were\nalready in the air.  DARPA considered contracting DEC to implement\nTCP\/IP, but rejected\nthat idea because they were concerned that DEC might not be responsive\nto requests for changes in their proprietary\nVAX\/VMS\n  operating system <sup><a href=\"#Libes-Ressler\" title=\"[Libes-Ressler]\">[Libes-Ressler]<\/a><\/sup>. Instead, DARPA\nchose Berkeley Unix as a platform \u2014 explicitly because its\nsource code was available and unencumbered <sup><a href=\"#Leonard\" title=\"[Leonard]\">[Leonard]<\/a><\/sup>.\n<\/p>\n<p>\n  Berkeley's\n  Computer Science Research Group was in the right place at the right\ntime with the strongest development tools; the result became arguably\nthe most critical turning point in Unix's history since its\ninvention.\n<\/p>\n<p>\n  Until the TCP\/IP\n  implementation was released with Berkeley\n  4.2 in 1983, Unix\nhad had only the weakest networking support.  Early experiments with\nEthernet were unsatisfactory.  An ugly but serviceable facility called\nUUCP (Unix to Unix Copy Program) had been developed at Bell Labs for\ndistributing software over conventional telephone lines via modem.<sup><a href=\"#ftn.id2880193\" id=\"id2880193\">[16]<\/a><\/sup>\n  UUCP could forward Unix mail\nbetween widely separated machines, and (after Usenet was invented in\n1981) supported\nUsenet, a distributed\nbulletin-board facility that allowed users to broadcast text messages\nto anywhere that had phone lines and Unix systems.\n<\/p>\n<p>\n  Still, the few Unix users aware of the bright lights of the\nARPANET felt like they were stuck in a backwater.  No FTP, no telnet,\nonly the most restricted remote job execution, and painfully slow\nlinks.  Before TCP\/IP,\nthe Internet and Unix cultures did not mix.  Dennis\nRitchie's\n  vision of computers as a way to \u201cencourage close\ncommunication\u201d was one of collegial communities clustered\naround individual timesharing machines or in the same computing\ncenter; it didn't extend to the continent-wide distributed\n\u2018network nation\u2019 that ARPA users had started to form in\nthe mid-1970s.  Early ARPANETters, for their part, considered Unix a\ncrude makeshift limping along on risibly weak hardware.\n<\/p>\n<p>\n  After TCP\/IP,\neverything changed. The ARPANET and Unix cultures began to merge at\nthe edges, a development that would eventually save both from\ndestruction.  But there would be hell to pay first as the result of\ntwo unrelated disasters; the rise of\nMicrosoft\n  and the\nAT&amp;T\n  divestiture.\n<\/p>\n<p>\n  In 1981, Microsoft made its historic deal with\nIBM\n  over the new IBM\nPC. Bill Gates bought QDOS (Quick and Dirty Operating System), a clone\nof CP\/M that its programmer Tim Paterson had thrown together in six\nweeks, from Paterson's employer Seattle Computer Products. Gates,\nconcealing the IBM deal from Paterson and SCP, bought the rights for\n$50,000.  He then talked IBM into allowing Microsoft to market MS-DOS\nseparately from the PC hardware.  Over the next decade, leveraging\ncode he didn't write made Bill Gates a multibillionaire, and\nbusiness tactics even sharper than the original deal gained Microsoft\na monopoly lock on desktop computing.  XENIX as a product was rapidly\ndeep-sixed, and eventually sold to SCO.\n<\/p>\n<p>\n  It was not apparent at the time how successful\n(or how destructive)\nMicrosoft\n  was going\nto be.  Since the IBM PC-1 didn't have the hardware capacity to run\nUnix, Unix people barely noticed it at all (though, ironically enough,\nDOS 2.0 eclipsed CP\/M largely because Microsoft's co-founder Paul Allen\nmerged in Unix features including subdirectories and\npipes).  There were\nthings that seemed much more interesting going on \u2014 like the\n1982 launching of Sun Microsystems.\n<\/p>\n<p>\n  Sun Microsystems\n  founders Bill\nJoy, Andreas\nBechtolsheim, and Vinod Khosla set out to build a dream Unix machine\nwith built-in networking capability.  They combined hardware designed\nat Stanford with the Unix developed at Berkeley to produce a smashing\nsuccess, and founded the workstation industry.  At the time, nobody\nmuch minded watching source-code access to one branch of the Unix tree\ngradually dry up as Sun began to behave less like a freewheeling\nstartup and more like a conventional\nfirm. Berkeley\n  was\nstill distributing BSD\n  with source code. Officially, System\nIII\n  source\nlicenses cost $40,000 each; but Bell Labs was turning a blind eye to\nthe number of bootleg Bell Labs Unix tapes in circulation, the\nuniversities were still swapping code with Bell Labs, and it looked\nlike Sun's\n  commercialization of Unix might just be the best thing to happen to it\nyet.\n<\/p>\n<p>\n  1982 was also the year that C\n  first showed signs of establishing\nitself outside the Unix world as the systems-programming language of\nchoice.  It would only take about five years for\nC\n  to drive machine\nassemblers almost completely out of use.  By the early 1990s C and\nC++\n  would dominate not\nonly systems but application programming; by the late 1990s all other\nconventional compiled languages would be effectively obsolete.\n<\/p>\n<p>\n  When DEC canceled development on the\nPDP-10's\n  successor\nmachine (Jupiter) in 1983, VAXes running Unix began to take over as\nthe dominant Internet machines, a position they would hold until being\ndisplaced by Sun workstations.  By 1985, about 25% of all VAXes\nwould be running Unix despite DEC's stiff opposition.  But the\nlongest-term effect of the Jupiter cancellation was a less obvious\none; the death of the MIT AI Lab's PDP-10-centered hacker culture\nmotivated a programmer named Richard\nStallman\n  to begin writing GNU, a complete free clone of Unix.\n<\/p>\n<p>\n  By 1983 there were no fewer than six Unix-workalike\noperating systems for the IBM-PC: uNETix, Venix, Coherent, QNX, Idris,\nand the port hosted on the Sritek PC daughtercard.  There was still no\nport of Unix in either the\nSystem V\n  or\nBSD\n  versions;\nboth groups considered the 8086 microprocessor woefully\nunderpowered and wouldn't go near it.  None of the Unix-workalikes\nwere significant as commercial successes, but they indicated a\nsignificant demand for Unix on cheap hardware that the major vendors\nwere not supplying.  No individual could afford to meet it, either,\nnot with the $40,000 price-tag on a source-code license.\n<\/p>\n<p>\n  Sun\n  was already a success (with imitators!) when, in 1983, the\nU.S. Department of Justice won its second antitrust case against\nAT&amp;T\n  and broke\nup the Bell System.  This relieved AT&amp;T from the 1958 consent\ndecree that had prevented them from turning Unix into a product.\nAT&amp;T promptly rushed to commercialize Unix System V\u2014a move\nthat nearly killed Unix.\n<\/p>\n <blockquote>\n  <p>\n   So true.  But their marketing did spread Unix internationally.\n  <\/p><div>\n  \n   \u2013 Ken Thompson\n  \n <\/div><\/blockquote>\n<p>\n  Most Unix boosters thought that the\ndivestiture\n  was great news. We thought we saw in\nthe post-divestiture AT&amp;T, Sun Microsystems, and Sun's smaller imitators the\nnucleus of a healthy Unix industry \u2014 one that, using inexpensive\n68000-based workstations, would challenge and eventually break the\noppressive monopoly that then loomed over the computer industry\n\u2014 IBM's.\n<\/p>\n<p>\n  What none of us realized at the time was that the productization\nof Unix would destroy the free exchanges of source code that had nurtured\nso much of the system's early vitality.  Knowing no other model than\nsecrecy for collecting profits from software and no other model than\ncentralized control for developing a commercial product, AT&amp;T\nclamped down hard on source-code distribution.  Bootleg Unix tapes became\nfar less interesting in the knowledge that the threat of lawsuit might\ncome with them.  Contributions from universities began to dry up.\n<\/p>\n<p>\n  To make matters worse, the big new players in the Unix market\npromptly committed major strategic blunders.  One was to seek\nadvantage by product differentiation \u2014 a tactic which resulted\nin the interfaces of different Unixes diverging.  This threw away\ncross-platform compatibility and fragmented the Unix market.\n<\/p>\n<p>\n  The other, subtler error was to behave as if personal computers\nand Microsoft\n  were\nirrelevant to Unix's prospects.  Sun\nMicrosystems\n  failed to see that commoditized PCs would inevitably become an attack\non its workstation market from below.\nAT&amp;T, fixated on\nminicomputers and mainframes, tried several different strategies to\nbecome a major player in computers, and badly botched all of them.  A\ndozen small companies formed to support Unix on PCs; all were\nunderfunded, focused on selling to developers and engineers, and never\naimed at the business and home market that Microsoft was\ntargeting.\n<\/p>\n<p>\n  In fact, for years after divestiture the Unix community was\npreoccupied with the first phase of the Unix\nwars\n  \u2014 an\ninternal dispute, the rivalry between System\nV Unix\n  and BSD\nUnix. The dispute\nhad several levels, some technical (sockets vs. streams,\nBSD\n  tty vs. System V\ntermio) and some cultural.  The divide was roughly between\nlonghairs and shorthairs; programmers and technical people tended to\nline up with Berkeley and BSD, more business-oriented types with AT&amp;T\nand System V.\nThe longhairs, repeating a theme from Unix's early days ten years\nbefore, liked to see themselves as rebels against a corporate empire;\none of the small companies put out a poster showing an X-wing-like\nspace fighter marked \u201cBSD\u201d speeding away from a huge\nAT&amp;T \u2018death star\u2019 logo left broken and in flames. Thus\nwe fiddled while Rome burned.\n<\/p>\n<p>\n  But something else happened in the year of the\nAT&amp;T divestiture that would have more long-term importance for\nUnix.  A programmer\/linguist named Larry Wall quietly invented the\npatch(1)\nutility.  The\n  <em>patch<\/em> program, a simple tool\nthat applies changebars generated by\ndiff(1)\nto a base file, meant that Unix developers could cooperate by passing\naround patch sets \u2014 incremental changes to code \u2014 rather\nthan entire code files.  This was important not only because patches\nare less bulky than full files, but because patches would often apply\ncleanly even if much of the base file had changed since the\npatch-sender fetched his copy.  With this tool, streams of development\non a common source-code base could diverge, run in parallel, and\nre-converge.  The\n  <em>patch<\/em> program did more\nthan any other single tool to enable collaborative development over\nthe Internet \u2014 a method that would revitalize Unix after\n1990.\n<\/p>\n<p>\n  In 1985 Intel shipped the first 386 chip, capable of addressing 4\ngigabytes of memory with a flat address space.  The clumsy segment\naddressing of the 8086 and 286 became immediately obsolete.  This was\nbig news, because it meant that for the first time, a microprocessor in\nthe dominant Intel family had the capability to run Unix without\npainful compromises.  The handwriting was on the wall for\nSun\n  and the\nother workstation makers.  They failed to see it.\n<\/p>\n<p>\n  1985 was also the year that Richard\nStallman\n  issued the GNU manifesto <sup><a href=\"#Stallman\" title=\"[Stallman]\">[Stallman]<\/a><\/sup> and launched the\nFree Software Foundation.  Very few people took him or his\nGNU\n  project\nseriously, a judgment that turned out to be seriously mistaken.  In\nan unrelated development of the same year, the originators of the X\nwindow system released it as source code without royalties,\nrestrictions, or license code.  As a direct result of this decision,\nit became a safe neutral area for collaboration between Unix vendors,\nand defeated proprietary contenders to become Unix's graphics\nengine.\n<\/p>\n<p>\n  Serious standardization efforts aimed at reconciling the System\nV\n  and\nBerkeley\n  APIs also\nbegan in 1983 with the \/usr\/group standard. This was followed in 1985\nby the POSIX\n  standards, an\neffort backed by the IEEE.  These described the intersection set of\nthe BSD\n  and SVR3\n(System V Release 3)\n  calls, with the superior Berkeley signal\nhandling and job control but with SVR3 terminal control.  All later\nUnix standards would incorporate POSIX at their core, and later Unixes\nwould adhere to it closely. The only major addition to the modern Unix\nkernel API to come afterwards was BSD sockets.\n<\/p>\n<p>\n  In 1986 Larry Wall, previously the inventor of\npatch(1),\nbegan work on Perl, which would become the first and most\nwidely used of the open-source scripting languages. In early 1987 the\nfirst version of the GNU C compiler\n  appeared, and by the end of 1987 the\ncore of the GNU toolset was falling into place\n  : editor, compiler, debugger, and\nother basic development tools.  Meanwhile, the X windowing\nsystem\n  was\nbeginning to show up on relatively inexpensive workstations. Together,\nthese would provide the armature for the open-source Unix developments\nof the 1990s.\n<\/p>\n<p>\n  1986 was also the year that PC technology broke free of IBM's\ngrip.  IBM, still trying\nto preserve a price-vs.-power curve across its product line that would\nfavor its high-margin mainframe business, rejected the 386 for most of\nits new line of PS\/2 computers in favor of the weaker 286.  The PS\/2\nseries, designed around a proprietary bus architecture to lock out\nclonemakers, became a colossally expensive failure.<sup><a href=\"#ftn.id2885728\" id=\"id2885728\">[17]<\/a><\/sup>\n  Compaq, the most aggressive of the\nclonemakers, trumped IBM's move by releasing the first 386 machine.\nEven with a clock speed of a mere 16 MHz, the 386 made a tolerable Unix\nmachine.  It was the first PC of which that could be said.\n<\/p>\n<p>\n  It was beginning to be possible to imagine that\nStallman's\n  GNU project\n  might mate with 386 machines to produce\nUnix workstations almost an order of magnitude less costly than anyone\nwas offering.  Curiously, no one seems to have actually got this far\nin their thinking.  Most Unix programmers, coming from the\nminicomputer and workstation worlds, continued to disdain cheap 80x86\nmachines in favor of more elegant 68000-based designs.  And, though a\nlot of programmers contributed to the GNU\nproject, among\nUnix people it tended to be considered a quixotic gesture that was\nunlikely to have near-term practical consequences.\n<\/p>\n<p>\n  The Unix community had never lost its rebel streak.  But in\nretrospect, we were nearly as blind to the future bearing down on us\nas IBM\n  or\nAT&amp;T.  Not even\nRichard Stallman, who had declared a moral crusade against\nproprietary software a few years before, really understood how badly\nthe productization of Unix had damaged the community around it; his\nconcerns were with more abstract and long-term issues.  The rest of us\nkept hoping that some clever variation on the corporate formula would\nsolve the problems of fragmentation, wretched marketing, and strategic\ndrift, and redeem Unix's pre-divestiture\npromise.  But worse was still to\ncome.\n<\/p>\n<p>\n  1988 was the year Ken Olsen (CEO of DEC) famously described Unix\nas \u201csnake oil\u201d.  DEC had been shipping its own variant of\nUnix on PDP-11s\n  since\n1982, but really wanted the business to go to its proprietary\nVMS\n  operating system.  DEC and the minicomputer industry were in deep\ntrouble, swamped by waves of powerful low-cost machines coming out of\nSun Microsystems\n  and the rest of the workstation\nvendors. Most of those workstations ran Unix.\n<\/p>\n<p>\n  But the Unix industry's own problems were growing more severe.\nIn 1988 AT&amp;T took a 20% stake in Sun Microsystems.  These two\ncompanies, the leaders in the Unix market, were beginning to wake up\nto the threat posed by PCs, IBM, and Microsoft, and to realize that\nthe preceding five years of bloodletting had gained them little.  The\nAT&amp;T\/Sun alliance and the development of technical standards around\nPOSIX\n  eventually healed\nthe breach between the System V\n  and BSD Unix\n  lines.  But the second phase of the Unix\nwars\n  began when the\nsecond-tier vendors (IBM,\nDEC, Hewlett-Packard, and others) formed the Open Software Foundation\nand lined up against the AT&amp;T\/Sun axis (represented by Unix\nInternational).  More rounds of Unix fighting Unix ensued.\n<\/p>\n<p>\n  Meanwhile,\nMicrosoft\n  was\nmaking billions in the home and small-business markets that the\nwarring Unix factions had never found the will to address. The 1990\nrelease of Windows 3.0 \u2014 the first successful graphical\noperating system from Redmond \u2014 cemented Microsoft's dominance,\nand created the conditions that would allow them to flatten and\nmonopolize the market for desktop applications in the 1990s.\n<\/p>\n<p>\n  The years from 1989 to 1993 were the darkest in Unix's history.\nIt appeared then that all the Unix community's dreams had failed.\nInternecine warfare had reduced the proprietary Unix industry to a\nsquabbling shambles that never summoned either the determination or\nthe capability to challenge Microsoft.  The elegant Motorola chips\nfavored by most Unix programmers had lost out to Intel's ugly but\ninexpensive processors.  The GNU project\n  failed to produce the free Unix kernel\nit had been promising since 1985, and after years of excuses\nits credibility was beginning to wear thin.  PC technology was being\nrelentlessly corporatized.  The pioneering Unix\nhackers\n  of the 1970s\nwere hitting middle age and slowing down.  Hardware was getting\ncheaper, but Unix was still too expensive.  We were belatedly becoming\naware that the old monopoly of\nIBM\n  had yielded to a\nnewer monopoly of\nMicrosoft, and\nMicrosoft's mal-engineered software was rising around us like a tide\nof sewage.\n<\/p>\n<h4 id=\"id2886020\"><a href=\"#id2886020\">\u00a7<\/a>Blows against the Empire: 1991-1995<\/h4>\n<p>\n  The first glimmer of light in the darkness was the 1990 effort\nby William Jolitz to port BSD\n  onto a 386 box, publicized by a series of\nmagazine articles beginning in 1991.  The 386BSD port was possible because,\npartly influenced by Stallman,\nBerkeley\n  hacker\nKeith Bostic\n  had begun an effort to clean AT&amp;T proprietary code out of the BSD\nsources in 1988.  But the 386BSD project took a severe blow when, near\nthe end of 1991, Jolitz walked away from it and destroyed his own\nwork.  There are conflicting explanations, but a common thread in all\nis that Jolitz wanted his code to be released as unencumbered source\nand was upset when the corporate sponsors of the project opted for a\nmore proprietary licensing model.\n<\/p>\n<p>\n  In August 1991 Linus Torvalds, then an unknown university student from\nFinland, announced the Linux\n  project.  Torvalds is on record that one\nof his main motivations was the high cost of Sun's\n  Unix at his university.  Torvalds\nhas also said that he would have joined the BSD\n  effort had he known of it, rather than\nfounding his own.  But 386BSD was not shipped until early 1992, some\nmonths after the first Linux release.\n<\/p>\n<p>\n  The importance of both these projects became clear only in\nretrospect. At the time, they attracted little notice even within the\nInternet hacker culture \u2014 let alone in the wider Unix community,\nwhich was still fixated on more capable machines than PCs, and on\ntrying to reconcile the special properties of Unix with the\nconventional proprietary model of a software business.\n<\/p>\n<p>\n  It would take another two years and the great Internet explosion\nof 1993\u20131994 before the true importance of\nLinux\n  and the\nopen-source\n  BSD\n  distributions\nbecame evident to the rest of the Unix world.  Unfortunately for the\nBSDers, an AT&amp;T\n  lawsuit against BSDI (the startup company that had backed the Jolitz\nport) consumed much of that time and motivated some key\nBerkeley\n  developers\nto switch to Linux.\n<\/p>\n <blockquote>\n  <p>\n   Code copying and theft of trade secrets was alleged. The actual\ninfringing code was not identified for nearly two years. The lawsuit\ncould have dragged on for much longer but for the fact that Novell\nbought USL from AT&amp;T and sought a settlement. In the end, three files\nwere removed from the 18,000 that made up the distribution, and a\nnumber of minor changes were made to other files. In addition, the\nUniversity agreed to add USL copyrights to about 70 files, with the\nstipulation that those files continued to be freely redistributed.\n  <\/p><div>\n  \n   \u2013 Marshall Kirk McKusick\n  \n <\/div><\/blockquote>\n<p>\n  The settlement set an important precedent by freeing an entire working Unix from proprietary control, but its effects\n  on BSD itself were dire.  Matters were not helped when, in 1992\u20131994, the Computer Science Research Group at Berkeley\n  shut down; afterwards, factional warfare within the BSD community split it into three competing development efforts.\n  As a result, the BSD lineage lagged behind Linux at a crucial time and lost to it the lead position in the Unix\n  community.\n<\/p>\n<p>\n  The Linux\n  and BSD\n  development efforts were native to the\nInternet in a way previous Unixes had not been.  They relied on\ndistributed development and Larry Wall's\npatch(1)\ntool, and recruited developers via email and through\nUsenet\n  newsgroups.\nAccordingly, they got a tremendous boost when Internet Service\nProvider businesses began to proliferate in 1993, enabled by changes\nin telecomm technology and the privatization of the Internet backbone\nthat are outside the scope of this history.  The demand for cheap\nInternet was created by something else: the 1991 invention of\nthe World Wide Web.  The Web was the \u201ckiller app\u201d of the\nInternet, the graphical user interface technology that made it\nirresistible to a huge population of nontechnical end users.\n<\/p>\n<p>\n  The mass-marketing of the Internet both increased the pool of\npotential developers and lowered the transaction costs of distributed\ndevelopment.  The results were reflected in efforts like XFree86,\nwhich used the Internet-centric model to build a more effective\ndevelopment organization than that of the official X\nConsortium.  The\nfirst XFree86 in 1992 gave Linux\n  and the BSDs\n  the graphical-user-interface engine they\nhad been missing.  Over the next decade XFree86 would lead in X\ndevelopment, and an increasing portion of the X Consortium's activity\nwould come to consist of funneling innovations originated in the\nXFree86 community back to the Consortium's industrial sponsors.\n<\/p>\n<p>\n  By late 1993, Linux\n  had both Internet capability and\nX. The entire\nGNU toolkit had been hosted on it from the beginning, providing\nhigh-quality development tools.  Beyond GNU tools,\nLinux\n  acted as a basin of attraction,\ncollecting and concentrating twenty years of open-source software that\nhad previously been scattered across a dozen different proprietary\nUnix platforms.  Though the Linux kernel was still officially in beta\n(at 0.99 level), it was remarkably crash-free.  The breadth and\nquality of the software in Linux\n  distributions was already that of a\nproduction-ready operating system.\n<\/p>\n<p>\n  A few of the more flexible-minded among old-school Unix\ndevelopers began to notice that the long-awaited dream of a cheap Unix\nsystem for everybody had snuck up on them from an unexpected direction.\nIt didn't come from\nAT&amp;T\n  or\nSun\n  or any\nof the traditional vendors. Nor did it rise out of an organized effort\nin academia.  It was a bricolage that bubbled up out of the Internet\nby what seemed like spontaneous generation, appropriating and\nrecombining elements of the Unix tradition in surprising ways.\n<\/p>\n<p>\n  Elsewhere, corporate maneuvering continued.\nAT&amp;T\n  divested\nits interest in Sun in 1992; then sold its Unix Systems Laboratories\nto Novell in 1993; Novell handed off the Unix trademark to the X\/Open\nstandards group in 1994; AT&amp;T and Novell joined OSF in 1994,\nfinally ending the Unix wars.  In 1995 SCO bought UnixWare (and the\nrights to the original Unix sources) from Novell.  In 1996, X\/Open and\nOSF merged, creating one big Unix standards group.\n<\/p>\n<p>\n  But the conventional Unix vendors and the wreckage of their wars\ncame to seem steadily less and less relevant.  The action and energy\nin the Unix community were shifting to Linux\n  and BSD\n  and the open-source\n  developers.  By the time\nIBM, Intel, and SCO\nannounced the Monterey project in 1998 \u2014 a last-gasp attempt to\nmerge One Big System out of all the proprietary Unixes left standing\n\u2014 developers and the trade press reacted with amusement, and the\nproject was abruptly canceled in 2001 after three years of going\nnowhere.\n<\/p>\n<p>\n  The industry transition could not be said to have completed\nuntil 2000, when SCO sold UnixWare and the original Unix source-code\nbase to Caldera \u2014 a Linux distributor. But after 1995, the story\nof Unix became the story of the open-source movement.  There's another\nside to that story; to tell it, we'll need to return to 1961 and the \norigins of the Internet hacker culture.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2881371\" id=\"ftn.id2881371\">[13]<\/a><\/sup>\n  Ken Thompson\nreminded me that today's cellphones have more RAM than the PDP-7 had\nRAM and disk storage combined; a\n  \n   <em>large<\/em>\n  \n  disk, in\nthose days, was less than a megabyte of storage.\n<\/p><p><sup><a href=\"#id2881412\" id=\"ftn.id2881412\">[14]<\/a><\/sup>\n  There is a Web FAQ on the\n  <a href=\"http:\/\/www.faqs.org\/faqs\/dec-faq\/pdp8\/\">\n   PDP computers<\/a>\n  that explains the otherwise extremely obscure PDP-7's place in\nhistory.\n<\/p><p><sup><a href=\"#id2879903\" id=\"ftn.id2879903\">[15]<\/a><\/sup>\n  The version 7 manuals can be browsed\non-line at\n  <a href=\"http:\/\/plan9.bell-labs.com\/7thEdMan\/index.html\">\n   http:\/\/plan9.bell-labs.com\/7thEdMan\/index.html<\/a>.\n<\/p><p><sup><a href=\"#id2880193\" id=\"ftn.id2880193\">[16]<\/a><\/sup>\n  UUCP was hot stuff when a\n  \n   <em>fast<\/em>\n  \n  modem was 300 baud.\n<\/p><p><sup><a href=\"#id2885728\" id=\"ftn.id2885728\">[17]<\/a><\/sup>\n  The\nPS\/2 did, however, leave one mark on later PCs \u2014 they made the\nmouse a standard peripheral, which is why the mouse connector on the\nback of your chassis is called a \u201cPS\/2\nport\u201d.\n<\/p><\/div>\n\n<h3 id=\"hackers\"><a href=\"#hackers\">\u00a7<\/a>Origins and History of the Hackers, 1961-1995<\/h3>\n<p>\n  The Unix tradition is an implicit culture that has always\ncarried with it more than just a bag of technical tricks. It transmits\na set of values about beauty and good design; it has legends and folk\nheroes.  Intertwined with the history of the Unix tradition is another\nimplicit culture that is more difficult to label neatly.  It has its\nown values and legends and folk heroes, partly overlapping with those\nof the Unix tradition and partly derived from other sources.  It has\nmost often been called the \u201chacker culture\u201d, and since\n1998 has largely coincided with what the computer trade press calls\n\u201cthe open source movement\u201d.\n<\/p>\n<p>\n  The relationships between the Unix tradition, the hacker\nculture, and the open-source movement are subtle and complex.  They\nare not simplified by the fact that all three implicit cultures have\nfrequently been expressed in the behaviors of the same human beings.\nBut since 1990 the story of Unix is largely the story of how the\nopen-source hackers changed the rules and seized the initiative from\nthe old-line proprietary Unix vendors.  Therefore, the other half of\nthe history behind today's Unix is the history of the hackers.\n<\/p>\n<h4 id=\"id2886621\"><a href=\"#id2886621\">\u00a7<\/a>At Play in the Groves of Academe: 1961-1980<\/h4>\n<p>\n  The roots of the hacker culture can be traced back to 1961, the\nyear MIT took delivery of its first PDP-1\n  minicomputer. The PDP-1 was one of the earliest interactive computers,\nand (unlike other machines) of the day was inexpensive enough that time\non it did not have to be rigidly scheduled.  It attracted a group of\ncurious students from the Tech Model Railroad Club who experimented\nwith it in a spirit of fun.\n  <em>Hackers: Heroes of the Computer Revolution<\/em><sup><a href=\"#Levy\" title=\"[Levy]\">[Levy]<\/a><\/sup>\nentertainingly describes the early days of the club.  Their most\nfamous achievement was SPACEWAR, a game of dueling rocketships loosely\ninspired by the\n  <em>Lensman<\/em> space operas of\nE.E. \u201cDoc\u201d Smith.<sup><a href=\"#ftn.id2886667\" id=\"id2886667\">[18]<\/a><\/sup><\/p>\n<p>\n  Several of the TMRC experimenters later went on to become core\nmembers of the MIT Artificial Intelligence Lab, which in the 1960s and\n1970s became one of the world centers of cutting-edge computer\nscience.  They took some of TMRC's slang and in-jokes with them,\nincluding a tradition of elaborate (but harmless) pranks called\n\u201chacks\u201d.  The AI Lab programmers appear to have been\nthe first to describe themselves as \u201chackers\u201d.\n<\/p>\n<p>\n  After 1969 the MIT AI Lab was connected, via the early ARPANET, to\nother leading computer science research laboratories at Stanford, Bolt\nBeranek &amp; Newman, Carnegie-Mellon University and elsewhere.\nResearchers and students got the first foretaste of the way fast\nnetwork access abolishes geography, often making it easier to\ncollaborate and form friendships with distant people on the net than\nit would be to do likewise with colleagues closer-by but less\nconnected.\n<\/p>\n<p>\n  Software, ideas, slang, and a good deal of humor flowed over the\nexperimental ARPANET links.  Something like a shared culture began to\nform.  One of its earliest and most enduring artifacts was the Jargon\nFile, a list of shared slang terms that originated at Stanford in 1973\nand went through several revisions at MIT after 1976.  Along the way\nit accumulated slang from CMU, Yale, and other ARPANET sites.\n<\/p>\n<p>\n  Technically, the early hacker culture was largely hosted on\nPDP-10 minicomputers.  They\nused a variety of operating systems that have since passed into\nhistory: TOPS-10, TOPS-20, Multics, ITS, SAIL.  They programmed in\nassembler and dialects of Lisp. PDP-10 hackers took over running the ARPANET\nitself because nobody else wanted the job. Later, they became the\nfounding cadre of the Internet Engineering Task Force\n(IETF)\n  and originated\nthe tradition of standardization through Requests For Comment\n(RFCs).\n<\/p>\n<p>\n  Socially, they were young, exceptionally bright, almost entirely\nmale, dedicated to programming to the point of addiction, and tended\nto have streaks of stubborn nonconformism \u2014 what years later\nwould be called \u2018geeks\u2019.  They, too, tended to be shaggy\nhippies and hippie-wannabes.  They, too, had a vision of computers\nas community-building devices.  They read Robert Heinlein and\nJ. R. R. Tolkien, played in the Society for Creative Anachronism, and\ntended to have a weakness for puns.  Despite their quirks (or perhaps\nbecause of them!) many of them were among the brightest programmers in\nthe world.\n<\/p>\n<p>\n  They were\n  <em>not<\/em>\n  Unix programmers.  The early\nUnix community was drawn largely from the same pool of geeks in\nacademia and government or commercial research laboratories, but the\ntwo cultures differed in important ways.  One that we've already \ntouched on is the weak networking of early Unix.  There was\neffectively no Unix-based ARPANET access until after 1980, and it\nwas uncommon for any individual to have a foot in both camps.\n<\/p>\n<p>\n  Collaborative development and the sharing of source code was a\nvalued tactic for Unix programmers.  To the early ARPANET hackers, on\nthe other hand, it was more than a tactic: it was something\nrather closer to a shared religion, partly arising from the academic\n\u201cpublish or perish\u201d imperative and (in its more extreme\nversions) developing into an almost Chardinist idealism about\nnetworked communities of minds.  The most famous of these hackers,\nRichard M. Stallman, became the ascetic saint of that\nreligion.\n<\/p>\n<h4 id=\"id2886828\"><a href=\"#id2886828\">\u00a7<\/a>Internet Fusion and the Free Software Movement: 1981-1991<\/h4>\n<p>\n  After 1983 and the BSD\n  port of\nTCP\/IP, the Unix and\nARPANET cultures began to fuse together.  This was a natural\ndevelopment once the communication links were in place, since both\ncultures were composed of the same kind of people (indeed, in a few\nbut significant cases the\n  <em>same<\/em>\n  people).  ARPANET\nhackers learned C\n  and began to speak the jargon of pipes, filters, and shells; Unix\nprogrammers learned\nTCP\/IP\n  and started to\ncall each other \u201chackers\u201d.  The process of fusion was\naccelerated after the Project Jupiter cancellation in 1983 killed the\nPDP-10's\n  future.\nBy 1987 the two cultures had merged so completely that most hackers\nprogrammed in C and casually used slang terms that went back to the\nTech Model Railroad Club of twenty-five years earlier.\n<\/p>\n<p>\n  (In 1979 I was unusual in having strong ties to both the Unix\nand ARPANET cultures.  In 1985 that was no longer unusual. By the time\nI expanded the old ARPANET Jargon File into the\n  <em>New Hacker's Dictionary<\/em><sup><a href=\"#Raymond96\" title=\"[Raymond96]\">[Raymond96]<\/a><\/sup> in 1991,\nthe two cultures had effectively fused.  The Jargon File, born on the\nARPANET but revised on\nUsenet, aptly\nsymbolized the merger.)\n<\/p>\n<p>\n  But TCP\/IP networking and slang were not the only things the\npost-1980 hacker culture inherited from its ARPANET roots.  It also\ngot Richard Stallman, and Stallman's moral crusade.\n<\/p>\n<p>\n  Richard M. Stallman\n  (generally known by his login name, RMS) had\nalready proved by the late 1970s that he was one of the most able\nprogrammers alive.  Among his many inventions was the Emacs editor.\nFor RMS, the Jupiter cancellation in 1983 only finished off a\ndisintegration of the MIT AI Lab culture that had begun a few years\nearlier as many of its best went off to help run competing\nLisp-machine companies.  RMS felt ejected from a hacker Eden, and\ndecided that proprietary software was to blame.\n<\/p>\n<p>\n  In 1983 Stallman founded the GNU project, aimed at writing an entire free\noperating system.  Though Stallman was not and had never been a Unix\nprogrammer, under post-1980 conditions implementing a Unix-like\noperating system became the obvious strategy to pursue.  Most of RMS's\nearly contributors were old-time ARPANET hackers newly decanted into\nUnix-land, in whom the ethos of code-sharing ran rather stronger than\nit did among those with a more Unix-centered background.\n<\/p>\n<p>\n  In 1985, RMS published the GNU Manifesto. In it he consciously\ncreated an ideology out of the values of the pre-1980 ARPANET hackers\n\u2014 complete with a novel ethico-political claim, a self-contained\nand characteristic discourse, and an activist plan for change. RMS\naimed to knit the diffuse post-1980 community of hackers into a\ncoherent social machine for achieving a single revolutionary\npurpose. His behavior and rhetoric half-consciously echoed Karl Marx's\nattempts to mobilize the industrial proletariat against the alienation\nof their work.\n<\/p>\n<p>\n  RMS's manifesto ignited a debate that is still live in the\nhacker culture today.  His program went way beyond maintaining a\ncodebase, and essentially implied the abolition of\nintellectual-property rights in software. In pursuit of this goal, RMS\npopularized the term \u201cfree software\u201d, which was the first\nattempt to label the product of the entire hacker culture. He wrote\nthe General Public License (GPL), which was to become both a rallying\npoint and a focus of great controversy, for reasons we will examine in\n  <a href=\"#reusechapter\" title=\"Chapter 16. Reuse\">Chapter 16<\/a>. You can learn more\nabout RMS's position and the Free Software Foundation at the <a href=\"http:\/\/www.gnu.org\">GNU website<\/a>.\n<\/p>\n<p>\n  The term \u201cfree software\u201d was partly a description\nand partly an attempt to define a cultural identity for hackers. On\none level, it was quite successful. Before RMS, people in the hacker\nculture recognized each other as fellow-travelers and used the same\nslang, but nobody bothered arguing about what a \u2018hacker\u2019\nis or should be. After him, the hacker culture became much more\nself-conscious; value disputes (often framed in RMS's language even by\nthose who opposed his conclusions) became a normal feature of\ndebate. RMS, a charismatic and polarizing figure, himself became so\nmuch a culture hero that by the year 2000 he could hardly be\ndistinguished from his legend.\n  <em>Free as in Freedom<\/em><sup><a href=\"#Williams\" title=\"[Williams]\">[Williams]<\/a><\/sup> gives\nus an excellent portrait.\n<\/p>\n<p>\n  RMS's arguments influenced the behavior even of many hackers who\nremained skeptical of his theories.  In 1987, he persuaded the\ncaretakers of BSD Unix\n  that cleaning out AT&amp;T's proprietary\ncode so they could release an unencumbered version would be a good\nidea.  However, despite his determined efforts over more than fifteen\nyears, the post-1980 hacker culture never unified around his\nideological vision.\n<\/p>\n<p>\n  Other hackers were rediscovering open, collaborative development\nwithout secrets for more pragmatic, less ideological reasons.  A few\nbuildings away from Richard Stallman's 9th-floor office at MIT, the X\ndevelopment\n  team thrived during the late 1980s.  It was funded by Unix vendors who\nhad argued each other to a draw over the control and\nintellectual-property-rights issues surrounding the X windowing\nsystem, and saw no better alternative than to leave it free to\neveryone. In 1987\u20131988 the X development prefigured the really huge\ndistributed communities that would redefine the leading edge of Unix\nfive years later.\n<\/p>\n <blockquote>\n  <p>\n   X was one of the first large-scale open-source projects to be\ndeveloped by a disparate team of individuals working for different\norganizations spread across the globe.  E-mail allowed ideas to move\nrapidly among the group so that issues could be resolved as quickly as\nnecessary, and each individual could contribute in whatever capacity\nsuited them best.  Software updates could be distributed in a matter\nof hours, enabling every site to act in a concerted manner during\ndevelopment.  The net changed the way software could be\ndeveloped.\n  <\/p><div>\n  \n   \u2013 Keith Packard\n  \n <\/div><\/blockquote>\n<p>\n  The X developers were no partisans of the GNU master\nplan, but they\nweren't actively opposed to it, either.  Before 1995 the most serious\nopposition to the GNU plan\n  came from the BSD\n  developers. The BSD\npeople, who\nremembered that they had been writing freely redistributable and\nmodifiable software years before RMS's\nmanifesto, rejected GNU's claim to historical and ideological\nprimacy. They specifically objected to the infectious or\n\u201cviral\u201d property of the GPL, holding out the BSD license\nas being \u201cmore free\u201d because it placed fewer restrictions\non the reuse of code.\n<\/p>\n<p>\n  It did not help RMS's case that, although his Free Software\nFoundation had produced most of the rest of a full software toolkit,\nit failed to deliver the central piece.  Ten years after the founding\nof the GNU project, there was still no GNU kernel.  While\nindividual tools like Emacs and GCC proved tremendously useful, GNU\nwithout a kernel neither threatened the hegemony of proprietary Unixes\nnor offered an effective counter to the rising problem of the\nMicrosoft\n  monopoly.\n<\/p>\n<p>\n  After 1995 the debate over RMS's ideology took a somewhat\ndifferent turn.  Opposition to it became closely associated with both\nLinus Torvalds\n  and the author of this\nbook.\n<\/p>\n<h4 id=\"linux_reaction\"><a href=\"#linux_reaction\">\u00a7<\/a>Linux and the Pragmatist Reaction: 1991-1998<\/h4>\n<p>\n  Even as the HURD (the GNU kernel) effort was stalling, new\npossibilities were opening up.  In the early 1990s the combination of\ncheap, powerful PCs with easy Internet access proved a powerful lure\nfor a new generation of young programmers looking for challenges to\ntest their mettle.  The user-space toolkit written by the Free\nSoftware Foundation suggested a way forward that was free of the high\ncost of proprietary software development tools.  Ideology followed\neconomics rather than leading the charge; some of the newbies signed\nup with RMS's crusade and adopted the GPL as their banner, and others\nidentified more with the Unix tradition as a whole and joined the\nanti-GPL camp, but most dismissed the whole dispute as a\ndistraction and just wrote code.\n<\/p>\n<p>\n  Linus Torvalds\n  neatly straddled the GPL\/anti-GPL divide\nby using the GNU toolkit to surround the Linux kernel he had invented\nand the GPL's infectious properties to protect it, but rejecting the\nideological program that went with RMS's license. Torvalds affirmed\nthat he thought free software better in general but occasionally used\nproprietary programs. His refusal to be a zealot even in his own cause\nmade him tremendously attractive to the majority of hackers who had\nbeen uncomfortable with RMS's rhetoric, but had lacked any focus or\nconvincing spokesperson for their skepticism.\n<\/p>\n<p>\n  Torvalds's cheerful pragmatism and adept but low-key style\ncatalyzed an astonishing string of victories for the hacker culture in\nthe years 1993\u20131997, including not merely technical successes\nbut the solid beginnings of a distribution, service, and support\nindustry around the Linux operating system.  As a result his prestige\nand influence skyrocketed.  Torvalds became a hero on Internet time;\nby 1995, he had achieved in just four years the kind of culture-wide\neminence that RMS had required fifteen years to earn \u2014 and far\nexceeded Stallman's\n  record at selling \u201cfree\nsoftware\u201d to the outside world.  By contrast with Torvalds,\nRMS's rhetoric began to seem both strident and unsuccessful.\n<\/p>\n<p>\n  Between 1991 and 1995 Linux\n  went from a proof-of-concept surrounding\nan 0.1 prototype kernel to an operating system that could compete on\nfeatures and performance with proprietary Unixes, and beat most of\nthem on important statistics like continuous uptime.  In 1995, Linux\nfound its killer app: Apache, the\nopen-source webserver.  Like Linux, Apache proved remarkably stable\nand efficient.  Linux machines running Apache quickly became the platform\nof choice for ISPs worldwide; Apache captured about 60% of\nwebsites,<sup><a href=\"#ftn.id2887399\" id=\"id2887399\">[19]<\/a><\/sup>\n  handily beating out both of its\nmajor proprietary competitors.\n<\/p>\n<p>\n  The one thing Torvalds did not offer was a new ideology \u2014\na new rationale or generative myth of hacking, and a positive\ndiscourse to replace RMS's hostility to intellectual property with a\nprogram more attractive to people both within and outside the hacker\nculture.  I\ninadvertently supplied this lack in 1997 as a result of trying to\nunderstand why Linux's development had not collapsed in confusion\nyears before. The technical conclusions of my published papers <sup><a href=\"#Raymond01\" title=\"[Raymond01]\">[Raymond01]<\/a><\/sup>\n  will be summarized in\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">Chapter 19<\/a>. For this historical sketch, it will be\nsufficient to note the impact of the first one's central formula: \u201cGiven a sufficiently large number of eyeballs, all\nbugs are shallow\u201d.\n<\/p>\n<p>\n  This observation implied something nobody in the hacker culture had\ndared to really believe in the preceding quarter-century: that its\nmethods could reliably produce software that was not just more elegant\nbut more reliable and\n  <em>better<\/em>\n  than our proprietary\ncompetitors' code. This consequence, quite unexpectedly, turned out to\npresent exactly the direct challenge to the discourse of \u201cfree\nsoftware\u201d that Torvalds himself had never been interested in\nmounting. For most hackers and almost all nonhackers, \u201cFree\nsoftware because it works better\u201d easily trumped \u201cFree\nsoftware because all software should be free\u201d.\n<\/p>\n<p>\n  The paper's contrast between \u2018cathedral\u2019\n(centralized, closed, controlled, secretive) and \u2018bazaar\u2019\n(decentralized, open, peer-review-intensive) modes of development\nbecame a central metaphor in the new thinking.  In an important sense\nthis was merely a return to Unix's pre-divestiture roots \u2014 it is\ncontinuous with McIlroy's\n  1991 observations about the positive\neffects of peer pressure on Unix development in the early 1970s and\nDennis Ritchie's\n  1979 reflections on fellowship,\ncross-fertilized with the early ARPANET's academic tradition of peer\nreview and with its idealism about distributed communities of mind.\n<\/p>\n<p>\n  In early 1998, the new thinking helped motivate Netscape\nCommunications to release the source code of its Mozilla\nbrowser. The press attention surrounding that event took Linux to Wall\nStreet, helped drive the technology-stock boom of 1999\u20132001, and\nproved to be a turning point in both the history of the hacker culture\nand of Unix.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2886667\" id=\"ftn.id2886667\">[18]<\/a><\/sup>\n  SPACEWAR was not related to Ken\nThompson's Space Travel game, other than by the fact that both\nappealed to science-fiction fans.\n<\/p><p><sup><a href=\"#id2887399\" id=\"ftn.id2887399\">[19]<\/a><\/sup>\n  Current and historical webserver share figures\nare available at the monthly\n  <a href=\"http:\/\/www.netcraft.com\/survey\/\">\n   Netcraft Web Server\nSurvey<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch02s03\"><a href=\"#ch02s03\">\u00a7<\/a>The Open-Source Movement: 1998 and Onward<\/h3>\n<p>\n  By the time of the Mozilla release in 1998, the hacker community\ncould best be analyzed as a loose collection of factions or tribes\nthat included Richard Stallman's\n  Free Software Movement, the Linux community,\nthe Perl\n  community, the\nApache\n  community, the\nBSD community, the X\ndevelopers, the Internet Engineering Task Force\n(IETF), and at least a\ndozen others.  These factions overlap, and an individual developer\nwould be quite likely to be affiliated with two or more.\n<\/p>\n<p>\n  A tribe might be grouped around a particular codebase that they\nmaintain, or around one or more charismatic influence leaders, or\naround a language or development tool, or around a particular software\nlicense, or around a technical standard, or around a caretaker\norganization for some part of the infrastructure.  Prestige tends to\ncorrelate with longevity and historical contribution as well as more\nobvious drivers like current market-share and mind-share; thus, perhaps the\nmost universally respected of the tribes is the\nIETF, which can claim\ncontinuity back to the beginnings of the ARPANET in 1969. The\nBSD\n  community, with\ncontinuous traditions back to the late 1970s, commands considerable\nprestige despite having a much lower installation count than\nLinux. Stallman's\n  Free Software Movement, dating back to the\nearly 1980s, ranks among the senior tribes both on historical\ncontribution and as the maintainer of several of the software tools in\nheaviest day-to-day use.\n<\/p>\n<p>\n  After 1995 Linux\n  acquired a special role as both the\nunifying platform for most of the community's other software and the\nhackers' most publicly recognizable brand name.  The Linux community\nshowed a corresponding tendency to absorb other sub-tribes \u2014\nand, for that matter, to co-opt and absorb the\nhacker\n  factions\nassociated with proprietary Unixes.  The hacker culture as a whole\nbegan to draw together around a common mission: push Linux and\nthe bazaar development model as far as it could go.\n<\/p>\n<p>\n  Because the post-1980 hacker\nculture\n  had become so\ndeeply rooted in Unix, the new mission was implicitly a brief for the\ntriumph of the Unix tradition.  Many of the hacker community's senior\nleaders were also Unix old-timers, still bearing scars from the\npost-divestiture civil wars of the 1980s and getting behind\nLinux\n  as the last, best hope to fulfill\nthe rebel dreams of the early Unix days.\n<\/p>\n<p>\n  The Mozilla release helped further concentrate opinions.  In\nMarch of 1998 an unprecedented summit meeting of community influence\nleaders representing almost all of the major tribes convened to\nconsider common goals and tactics.  That meeting adopted a new label\nfor the common development method of all the factions: open\nsource.\n<\/p>\n<p>\n  Within six months almost all the tribes in the hacker community\nwould accept \u201copen source\u201d as its new banner.  Older\ngroups like IETF\n  and the\nBSD developers would begin to apply it retrospectively to what they\nhad been doing all along.  In fact, by 2000 the rhetoric of open\nsource would not just unify the hacker culture's present practice and\nplans for the future, but re-color its view of its own past.\n<\/p>\n<p>\n  The galvanizing effect of the Netscape announcement, and of the\nnew prominence of Linux, reached well beyond the Unix community and\nthe hacker culture.  Beginning in 1995, developers from various\nplatforms in the path of Microsoft's Windows juggernaut (MacOS; Amiga;\nOS\/2; DOS; CP\/M; the weaker proprietary Unixes; various mainframe,\nminicomputer, and obsolete microcomputer operating systems) had banded\ntogether around Sun Microsystems's Java\nlanguage.  Many\ndisgruntled Windows developers joined them in hopes of maintaining at\nleast some nominal independence from Microsoft.  But Sun's handling of\nJava was (as we discuss in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">Chapter 14<\/a>) clumsy and alienating on several levels.\nMany Java developers liked what they saw in the nascent open-source\nmovement, and followed Netscape's lead into Linux and open source just\nas they had previously followed Netscape into Java.\n<\/p>\n<p>\n  Open-source activists welcomed the surge of immigrants from\neverywhere.  The old Unix hands began to share the new immigrants'\ndreams of not merely passively out-enduring the Microsoft monopoly,\nbut actually reclaiming key markets from it.  The open-source\ncommunity as a whole prepared a major push for mainstream\nrespectability, and began to welcome alliances with major corporations\nthat increasingly feared losing control of their own businesses as\nMicrosoft's lock-in tactics grew ever bolder.\n<\/p>\n<p>\n  There was one exception: Richard\nStallman\n  and the Free Software Movement.  \u201cOpen source\u201d was\nexplicitly intended to replace Stallman's preferred \u201cfree\nsoftware\u201d with a public label that was ideologically neutral,\nacceptable both to historically opposed groups like the\nBSD\n  hackers and\nthose who did not wish to take a position in the GPL\/anti-GPL debate.\nStallman flirted with adopting the term, then rejected it on the\ngrounds that it failed to represent the moral position that was\ncentral to his thinking.  The Free Software Movement has since\ninsisted on its separateness from \u201copen source\u201d,\ncreating perhaps the most significant political\nfissure in the hacker culture of 2003.\n<\/p>\n<p>\n  The other (and more important) intention behind \u201copen\nsource\u201d was to present the hacker community's methods to the\nrest of the world (especially the business mainstream) in a more\nmarket-friendly, less confrontational way.  In this role, fortunately,\nit proved an unqualified success \u2014 and led to a revival of\ninterest in the Unix tradition from which it sprang.\n<\/p>\n\n<h3 id=\"ch02s04\"><a href=\"#ch02s04\">\u00a7<\/a>The Lessons of Unix History<\/h3>\n<p>\n  The largest-scale pattern in the history of Unix is this: when\nand where Unix has adhered most closely to\nopen-source\n  practices, it has prospered.  Attempts to proprietarize it have\ninvariably resulted in stagnation and decline.\n<\/p>\n<p>\n  In retrospect, this should probably have become obvious much\nsooner than it did.  We lost ten years after 1984 learning our lesson,\nand it would probably serve us very ill to ever again forget\nit.\n<\/p>\n<p>\n  Being smarter than anyone else about important but narrow issues\nof software design didn't prevent us from being almost completely\nblind about the consequences of interactions between technology and\neconomics that were happening right under our noses.  Even the most\nperceptive and forward-looking thinkers in the Unix community were at\nbest half-sighted.  The lesson for the future is that over-committing\nto any one technology or business model would be a mistake \u2014 and\nmaintaining the adaptive flexibility of our software and the design\ntradition that goes with it is correspondingly imperative.\n<\/p>\n<p>\n  Another lesson is this: Never bet against the cheap plastic\nsolution. Or, equivalently, the low-end\/high-volume hardware\ntechnology almost always ends up climbing the power curve and winning.\nThe economist Clayton Christensen calls this\n  <em>disruptive technology<\/em>\n  and showed in\n  <em>The Innovator's Dilemma<\/em><sup><a href=\"#Christensen\" title=\"[Christensen]\">[Christensen]<\/a><\/sup> how this happened\nwith disk drives, steam shovels, and motorcycles.  We saw it happen as\nminicomputers displaced mainframes, workstations and servers replaced\nminis, and commodity Intel machines replaced workstations and servers.\nThe open-source movement is winning by commoditizing software.  To\nprosper, Unix needs to maintain the knack of co-opting the cheap\nplastic solution rather than trying to fight it.\n<\/p>\n<p>\n  Finally, the old-school Unix community failed in its efforts to\nbe \u201cprofessional\u201d by welcoming in all the command\nmachinery of conventional corporate organization, finance, and\nmarketing.  We had to be rescued from our folly by a rebel alliance of\nobsessive geeks and creative misfits\n  \u2014who then\nproceeded to show us that professionalism and dedication really meant\nwhat we had been doing\n  <em>before<\/em>\n  we succumbed to the\nmundane persuasions of \u201csound business practices\u201d.\n<\/p>\n<p>\n  The application of these lessons with respect to software \ntechnologies other than Unix is left as an easy exercise for the \nreader.\n<\/p>\n\n<h3 id=\"contrastchapter\"><a href=\"#contrastchapter\">\u00a7<\/a>Chapter 3. Contrasts<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   If you have any trouble sounding condescending, find a Unix user\nto show you how it's done.\n  <\/p><div>\n  \n   \u2013 Scott Adams\nDilbert newsletter 3.0, 1994\n  \n <\/div><\/blockquote>\n<p>\n  The design of operating systems conditions the style of software\ndevelopment under them in many ways both obvious and subtle.  Much of\nthis book traces connections between the design of the Unix operating\nsystem and the philosophy of program design that has evolved around\nit.  For contrast, it will therefore be instructive to compare the\nclassic Unix way with the styles of design and programming native to\nother major operating systems.\n<\/p>\n\n<h3 id=\"ch03s01\"><a href=\"#ch03s01\">\u00a7<\/a>The Elements of Operating-System Style<\/h3>\n<p>\n  Before we can start discussing specific operating systems, we'll\nneed an organizing framework for the ways that operating-system\ndesign can affect programming style for good or ill.\n<\/p>\n<p>\n  Overall, the design and programming styles associated with\ndifferent operating systems seem to derive from three different\nsources: (a) the intentions of the operating-system designers, (b)\nuniformities forced on designs by costs and limitations in the\nprogramming environment, and (c) random cultural drift, early\npractices becoming traditional simply because they were there\nfirst.\n<\/p>\n<p>\n  Even if we take it as given that there is some random cultural\ndrift in every operating-system community, considering the intentions\nof the designers and the costs and limitations of the results does\nreveal some interesting patterns that can help us understand the Unix\nstyle better by contrast.  We can make the patterns explicit by\nanalyzing some of the most important ways that operating systems\ndiffer.\n<\/p>\n<h4 id=\"id2892028\"><a href=\"#id2892028\">\u00a7<\/a>What Is the Operating System's Unifying Idea?<\/h4>\n<p>\n  Unix has a couple of unifying ideas or metaphors that shape its\nAPIs and the development style that proceeds from them.  The most\nimportant of these are probably the \u201ceverything is a\nfile\u201d model and the pipe metaphor<sup><a href=\"#ftn.id2892049\" id=\"id2892049\">[20]<\/a><\/sup>\n  built on top of it.\nIn general, development style under any given operating system is\nstrongly conditioned by the unifying ideas baked into the system by\nits designers \u2014 they percolate upwards into applications\nprogramming from the models provided by system tools and APIs.\n<\/p>\n<p>\n  Accordingly, the most basic question to ask in contrasting Unix\nwith another operating system is: Does it have unifying ideas that\nshape its development, and if so how do they differ from\nUnix's?\n<\/p>\n<p>\n  To design the perfect anti-Unix, have no unifying idea at all,\njust an incoherent pile of ad-hoc features.\n<\/p>\n<h4 id=\"id2892085\"><a href=\"#id2892085\">\u00a7<\/a>Multitasking Capability<\/h4>\n<p>\n  One of the most basic ways operating systems can differ is in\nthe extent to which they can support multiple concurrent processes.  At\nthe lowest end (such as DOS or CP\/M) the operating system is basically\na sequential program loader with no capacity to multitask at all.\nOperating systems of this kind are no longer competitive on\ngeneral-purpose computers.\n<\/p>\n<p>\n  At the next level up, an operating system may have\n  <em>cooperative multitasking<\/em>.  Such systems can\nsupport multiple processes, but a process has to voluntarily give up\nits hold on the processor before the next one can run (thus, simple\nprogramming errors can readily freeze the machine).  This style of\noperating system was a transient adaptation to hardware that was\npowerful enough for concurrency but lacked either a periodic clock\ninterrupt<sup><a href=\"#ftn.id2892115\" id=\"id2892115\">[21]<\/a><\/sup>\n  or a memory-management unit or both; it,\ntoo, is obsolete and no longer competitive.\n<\/p>\n<p>\n  Unix has\n  <em>preemptive multitasking<\/em>, in\nwhich timeslices are allocated by a scheduler which routinely\ninterrupts or pre-empts the running process in order to hand control\nto the next one.  Almost all modern operating systems support\npreemption.\n<\/p>\n<p>\n  Note that \u201cmultitasking\u201d is not the same as\n\u201cmultiuser\u201d.  An operating system can be multitasking\nbut single-user, in which case the facility is used to support a\nsingle console and multiple background processes.  True multiuser\nsupport requires multiple user privilege domains, a feature we'll\ncover in the discussion of internal boundaries a bit further on.\n<\/p>\n<p>\n  To design the perfect anti-Unix, don't support multitasking at\nall \u2014 or, support multitasking but cripple it by surrounding\nprocess management with a lot of restrictions, limitations, and\nspecial cases that mean it's quite difficult to get any actual use out\nof multitasking.\n<\/p>\n<h4 id=\"id2892171\"><a href=\"#id2892171\">\u00a7<\/a>Cooperating Processes<\/h4>\n<p>\n  In the Unix experience, inexpensive process-spawning and easy\ninter-process communication (IPC) makes a whole ecology of small\ntools, pipes, and\nfilters possible.  We'll explore this ecology in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>; here, we need to point out some\nconsequences of expensive process-spawning and IPC.\n<\/p>\n <blockquote>\n  <p>\n   The pipe was technically trivial, but profound in its effect.\nHowever, it would not have been trivial without the fundamental\nunifying notion of the process as an autonomous unit of computation,\nwith process control being programmable.  As in Multics, a shell was\njust another process; process control did not come from God inscribed\nin JCL.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  If an operating system makes spawning new processes expensive\nand\/or process control is difficult and inflexible, you'll usually see\nall of the following consequences:\n<\/p>\n <div><ul><li>\n    <p>\n     Monster monoliths become a more natural way of programming.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Lots of policy has to be expressed within those monoliths. This\nencourages C++\n     and\nelaborately layered internal code organization, rather than C and\nrelatively flat internal hierarchies.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     When processes can't avoid a need to communicate, they do so\nthrough mechanisms that are either clumsy, inefficient, and insecure\n(such as temporary files) or by knowing far too much about each others'\nimplementations.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Multithreading is extensively used for tasks that Unix would\nhandle with multiple communicating lightweight processes.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Learning and using asynchronous I\/O is a must.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  These are examples of common stylistic traits (even in\napplications programming) being driven by a limitation in the OS\nenvironment.\n<\/p>\n<p>\n  A subtle but important property of pipes and the other classic\nUnix IPC methods is that they require communication between programs\nto be held down to a level of simplicity that encourages separation of\nfunction.  Conversely, the result of having no equivalent of the pipe\nis that programs can only be designed to cooperate by building in full\nknowledge of each others' internals.\n<\/p>\n<p>\n  In operating systems without flexible IPC and a strong tradition\nof using it, programs communicate by sharing elaborate data\nstructures.  Because the communication problem has to be solved anew\nfor all programs every time another is added to the set, the\ncomplexity of this solution rises as the square of the number of\ncooperating programs.  Worse than that, any change in one of the\nexposed data structures can induce subtle bugs in an arbitrarily \nlarge number of other programs.\n<\/p>\n <blockquote>\n  <p>\n   Word and Excel and PowerPoint and other Microsoft programs have\nintimate \u2014 one might say promiscuous \u2014 knowledge of each\nothers' internals.  In Unix, one tries to design programs to operate\nnot specifically with each other, but with programs as yet unthought\nof.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  We'll return to this theme in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>.\n<\/p>\n<p>\n  To design the perfect anti-Unix, make process-spawning very\nexpensive, make process control difficult and inflexible, and\nleave IPC as an unsupported or half-supported afterthought.\n<\/p>\n<h4 id=\"id2888136\"><a href=\"#id2888136\">\u00a7<\/a>Internal Boundaries<\/h4>\n<p>\n  Unix has wired into it an assumption that the programmer knows\nbest.  It doesn't stop you or request confirmation when you do\ndangerous things with your own data, like issuing\n  <code>rm -rf *<\/code>.  On the other hand, Unix is rather careful about not\nletting you step on other people's data.  In fact, Unix encourages you\nto have multiple accounts, each with its own attached and possibly\ndiffering privileges, to help you protect yourself from misbehaving\nprograms.<sup><a href=\"#ftn.id2888159\" id=\"id2888159\">[22]<\/a><\/sup>\n  System\nprograms often have their own pseudo-user accounts to confer access to\nspecial system files without requiring unlimited\n(or\n  <em>superuser<\/em>) access.\n<\/p>\n<p>\n  Unix has at least three levels of internal boundaries that guard\nagainst malicious users or buggy programs. One is memory management;\nUnix uses its hardware's memory management unit (MMU) to ensure that\nseparate processes are prevented from intruding on the others'\nmemory-address spaces.  A second is the presence of true privilege\ngroups for multiple users \u2014 an ordinary (nonroot) user's\nprocesses cannot alter or read another user's files without\npermission.  A third is the confinement of security-critical functions\nto the smallest possible pieces of trusted code.  Under Unix, even the\nshell (the system command interpreter) is not a privileged\nprogram.\n<\/p>\n<p>\n  The strength of an operating system's internal boundaries\nis not merely an abstract issue of design: It has important\npractical consequences for the security of the system.\n<\/p>\n<p>\n  To design the perfect anti-Unix, discard or bypass memory\nmanagement so that a runaway process can crash, subvert, or corrupt\nany running program.  Have weak or nonexistent privilege groups, so\nusers can readily alter each others' files and the system's critical\ndata (e.g., a macro virus, having seized control of your word\nprocessor, can format your hard drive).  And trust large volumes of\ncode, like the entire shell and GUI, so that any bug or successful\nattack on that code becomes a threat to the entire system.\n<\/p>\n<h4 id=\"id2888217\"><a href=\"#id2888217\">\u00a7<\/a>File Attributes and Record Structures<\/h4>\n<p>\n  Unix files have neither record structure nor attributes.  In\nsome operating systems, files have an associated record structure; the\noperating system (or its service libraries) knows about files with a \nfixed record length, or about text line termination and whether CR\/LF\nis to be read as a single logical character.\n<\/p>\n<p>\n  In other operating systems, files and directories can have\nname\/attribute pairs associated with them \u2014 out-of-band data\nused (for example) to associate a document file with an application\nthat understands it.  (The classic Unix way to handle these associations is\nto have applications recognize \u2018magic numbers\u2019, or other\ntype data within the file itself.)\n<\/p>\n<p>\n  OS-level record structures are generally an optimization\nhack, and do\nlittle more than complicate APIs and programmers' lives.  They\nencourage the use of opaque record-oriented file formats that generic\ntools like text editors cannot read properly.\n<\/p>\n<p>\n  File attributes can be useful, but (as we will see in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">Chapter 20<\/a>) can raise some awkward semantic issues in a\nworld of byte-stream-oriented tools and\npipes.  When file\nattributes are supported at the operating-system level, they\npredispose programmers to use opaque formats and lean on the file\nattributes to tie them to the specific applications that interpret\nthem.\n<\/p>\n<p>\n  To design the perfect anti-Unix, have a cumbersome set of record \nstructures that make it a hit-or-miss proposition whether any given\ntool will be able to even read a file as the writer intended it.  Add\nfile attributes and have the system depend on them heavily, so that \nthe semantics of a file will not be determinable by looking at the\ndata within it.\n<\/p>\n<h4 id=\"id2888298\"><a href=\"#id2888298\">\u00a7<\/a>Binary File Formats<\/h4>\n<p>\n  If your operating system uses binary formats for critical data\n(such as user-account records) it is likely that no tradition of\nreadable textual formats for applications will develop.  We explain in\nmore detail why this is a problem in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">Chapter 5<\/a>.  For now it's sufficient to note the\nfollowing consequences:\n<\/p>\n <div><ul><li>\n    <p>\n     Even if a command-line interface, scripting, and\npipes\n     are supported,\nvery few filters will evolve.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Data files will be accessible only through dedicated tools.\nDevelopers will think of the tools rather than the data files as central.\nThus, different versions of file formats will tend to be incompatible.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  To design the perfect anti-Unix, make all file formats binary \nand opaque, and require heavyweight tools to read and edit them.\n<\/p>\n<h4 id=\"id2888363\"><a href=\"#id2888363\">\u00a7<\/a>Preferred User Interface Style<\/h4>\n<p>\n  In\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">Chapter 11<\/a>\n  we will develop in some detail the consequences of the differences between\n  <em>command-line interfaces<\/em> (CLIs) and\n  <em>graphical user interfaces<\/em> (GUIs).  Which kind\nan operating system's designers choose as its normal mode of\npresentation will affect many aspects of the design, from process\nscheduling and memory management on up to the\n  <em>application programming interfaces<\/em>\n  (APIs) presented for applications\nto use.\n<\/p>\n<p>\n  It has been enough years since the first Macintosh that very few\npeople need to be convinced that weak GUI facilities in an operating\nsystem are a problem.  The Unix lesson is the opposite: that weak\nCLI facilities are a less obvious but equally severe deficit.\n<\/p>\n<p>\n  If the CLI facilities of an operating system are weak or\nnonexistent, you'll also see the following consequences:\n<\/p>\n <div><ul><li>\n    <p>\n     Programs will not be designed to cooperate with each other in\nunexpected ways \u2014 because they\n     <em>can't<\/em>\n     be.  Outputs\naren't usable as inputs.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Remote system administration will be sparsely supported, more\ndifficult to use, and more network-intensive.<sup><a href=\"#ftn.id2888440\" id=\"id2888440\">[23]<\/a><\/sup><\/p>\n   <\/li>\n   <li>\n    <p>\n     Even simple noninteractive programs will incur the overhead \nof a GUI or elaborate scripting interface.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Servers, daemons, and background processes will probably be\nimpossible or at least rather difficult, to program in any graceful\nway.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  To design the perfect anti-Unix, have no CLI and no \ncapability to script programs \u2014 or, important facilities that\nthe CLI cannot drive.\n<\/p>\n<h4 id=\"id2888484\"><a href=\"#id2888484\">\u00a7<\/a>Intended Audience<\/h4>\n<p>\n  The design of operating systems varies in response to the \nexpected audience for the system.  Some operating systems are\nintended for back rooms, some for desktops.  Some are designed\nfor technical users, others for end users.  Some are intended to\nwork standalone in real-time control applications, others for an \nenvironment of timesharing and pervasive networking.\n<\/p>\n<p>\n  One important distinction is client\nvs. server. \u2018Client\u2019 translates as: being lightweight,\nsuppporting only a single user, able to run on small machines,\ndesigned to be switched on when needed and off when the user is done,\nlacking pre-emptive multitasking, optimized for low latency, and\nputting a lot of its resources into fancy user interfaces.\n\u2018Server\u2019 translates as: being heavyweight, capable of\nrunning continuously, optimized for throughput, fully pre-emptively\nmultitasking to handle multiple sessions.  In origin all operating\nsystems were server operating systems; the concept of a client\noperating system only emerged in the late 1970s with inexpensive but\nunderpowered PC hardware.  Client operating systems are more focused\non a visually attractive user experience than on 24\/7 uptime.\n<\/p>\n<p>\n  All these variables have an effect on development style. One of\nthe most obvious is the level of interface complexity the target\naudience will tolerate, and how it tends to weight perceived\ncomplexity against other variables like cost and capability. Unix is\noften said to have been written by programmers for programmers \u2014\nan audience that is notoriously tolerant of interface\ncomplexity.\n<\/p>\n <blockquote>\n  <p>\n   This is a consequence rather than a goal.  I abhor a system designed\nfor the \u201cuser\u201d, if that word is a coded pejorative\nmeaning \u201cstupid and unsophisticated\u201d.\n  <\/p><div>\n  \n   \u2013 Ken Thompson\n  \n <\/div><\/blockquote>\n<p>\n  To design the perfect anti-Unix, write an operating system that\nthinks it knows what you're doing better than you do.  And then adds\ninjury to insult by getting it wrong.\n<\/p>\n<h4 id=\"id2888581\"><a href=\"#id2888581\">\u00a7<\/a>Entry Barriers to Development<\/h4>\n<p>\n  Another important dimension along which operating systems differ\nis the height of the barrier that separates mere users from becoming\ndevelopers.  There are two important cost drivers here.  One is the\nmonetary cost of development tools, the other is the time cost of\ngaining proficiency as a developer.  Some development cultures evolve\nsocial barriers to entry, but these are usually an effect of the\nunderlying technology costs, not a primary cause.\n<\/p>\n<p>\n  Expensive development tools and complex, opaque APIs produce\nsmall and elitist programming cultures.  In those cultures,\nprogramming projects are large, serious endeavors \u2014 they have to\nbe in order to offer a payoff that justifies the cost of both hard and\nsoft (human) capital invested.  Large, serious projects tend to\nproduce large, serious programs (and, far too often, large expensive\nfailures).\n<\/p>\n<p>\n  Inexpensive tools and simple interfaces support casual\nprogramming, hobbyist cultures, and exploration.  Programming projects\ncan be small (often, formal project structure is plain unnecessary),\nand failure is not a catastrophe.  This changes the style in which\npeople develop code; among other things, they show less tendency to\nover-commit to failed approaches.\n<\/p>\n<p>\n  Casual programming tends to produce lots of small programs and a\nself-reinforcing, expanding community of knowledge.  In a world of\ncheap hardware, the presence or absence of such a community is an\nincreasingly important factor in whether an operating system is\nlong-term viable at all.\n<\/p>\n<p>\n  Unix pioneered casual programming. One of the things Unix was\nfirst at doing was shipping with a compiler and scripting tools as\npart of the default installation available to all users, supporting a\nhobbyist software-development culture that spanned multiple\ninstallations.  Many people who write code under Unix do not think of\nit as writing code \u2014 they think of it as writing scripts to\nautomate common tasks, or as customizing their environment.\n<\/p>\n<p>\n  To design the perfect anti-Unix, make casual programming\nimpossible.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2892049\" id=\"ftn.id2892049\">[20]<\/a><\/sup>\n  For\nreaders without Unix experience, a pipe is a way of connecting the\noutput of one program to the input of another.  We'll explore the ways\nthis idea can be used to help programs cooperate in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>.\n<\/p><p><sup><a href=\"#id2892115\" id=\"ftn.id2892115\">[21]<\/a><\/sup>\n  A periodic clock interrupt from the hardware\nis useful as a sort of heartbeat for a timesharing system; each time\nit fires, it tells the system that it may be time to switch to another\ntask, defining the size of the unit timeslice.  In 2003 Unixes\nusually set the heartbeat to either 60 or 100 times a\nsecond.\n<\/p><p><sup><a href=\"#id2888159\" id=\"ftn.id2888159\">[22]<\/a><\/sup>\n  The modern buzzword for this is\n  <em>role-based security<\/em>.\n<\/p><p><sup><a href=\"#id2888440\" id=\"ftn.id2888440\">[23]<\/a><\/sup>\n  This\nproblem was considered quite serious by Microsoft itself during their\nrebuild of Hotmail.  See <sup><a href=\"#BrooksD\" title=\"[BrooksD]\">[BrooksD]<\/a><\/sup>.\n<\/p><\/div>\n\n<h3 id=\"ch03s02\"><a href=\"#ch03s02\">\u00a7<\/a>Operating-System Comparisons<\/h3>\n<p>\n  The logic of Unix's design choice stands out more clearly when\nwe contrast it with other operating systems.  Here we will attempt\nonly a design overview; for detailed discussion of the technical\nfeatures of different operating systems.<sup><a href=\"#ftn.id2888666\" id=\"id2888666\">[24]<\/a><\/sup><\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 3.1. Schematic history of timesharing.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Schematic history of timesharing.\" src=\"denied:data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAsoAAAL4CAAAAABHT446AAAAAnRSTlMA\/1uRIrUAAAACYktHRAD\/h4\/MvwAAAAlwSFlzAAAAZAAAAGQAD5bF3QAARZtJREFUeNrtnU+o60p64D+SId3zMuMXEW7SA57MPQfSfQ0hxM+Qk2xCDIYeaDJzz4HTvM0Q4TeYgceBYN6LFmcxYJhBNIEevDB4dRuPF4+zaYwhWXjhhZmdoBfeNZgs1GtvtAhZeapKfyzbsqWyqlQl+ftxr6UjyXK59HP5U\/1xwW632SXi7BCkRMDGBrp0LOg40WJ3D9BTnTQE4QHYv92utdv17qPFeKE6XQjCSagyedjCIlx0WtYm75kRpFAilbfk\/yJcLKx7A2NlpFSEKj\/buwVxOFiQsvm5pTppCMJDqLJjGBYpkoMFYQOqk4YgPIQqE3pWfLEzVCcNQXjYqzxubcPFlqxZluqkIQgPsLUBiLUby6Im+4uxcd9Ck5FyERTJ20V8gSDlA2\/ukIqAKiMVAVVGKgKqjFQEVBmpCKgyUhFQZaQioMpIRUCVkYqAKiMVoVQqb74tAM2Hz2AenKNUKi\/mnnTmmndDwTw4B6pcssuIeXAOVLlklxHz4ByocskuI+bBOVDlkl1GzINzoMolu4yYB+dAlUt2GTEPzoEql+wyYh6cA1Uu2WXEPDgHqlyyy5g3D9YVyINkKqPyynj02gBNcrGac8+IXbtmG+5cstKHNTmAMS\/vZUxXmbxJA4zHVSxv+gbcdUkeDAySQ+mfBd3zIJnKqOx178g1gInnud2D7YOBB0aTroHXZ49ev9Iqey5Qe+\/2Jj\/Sx0fy5uknfD5AldVz8TKuaFl7R67akCxdN\/oqba49WBldJrHrq8yWJb2MWQIM9hb3X0yPE\/bGyMeZZUP6CXTPg2Sqo7J31\/VWTXBp0TskF3MCQwPIpja9tEMY+Fc4fCztZcyo8rxL\/X0kscY6jLaA5ks7w9O1z4NkKqTywHD7a2OwDqII+n9FYoqhH1OQ0ON2VO73m90VNZfQD98xXUzuoO+iyuq5fBnXMOySiHmw3qtM\/tM7PrrWNVa3ozL9YJNYuTv0\/2TyBhFHF\/qosnpSLuMjKXlX0A4vJvvPbnlY7Ni8uymV6UOXWut6bRYrTx69Of2U9yH1+brnQTJVUnlIy527wYHKrGCCFSuVIFZGlfYyZlV5eEczZO6tJ96A1d88rsidH3nyqo0qqyflMrq0EGLxxRwG6zkM1wP4gctCRrpnRU1f9S\/WKut\/GdNVpm\/RaLJ65T4Y9J0Pm2S5ot9MBkDXTTuB9nmQTJVUFoLulxHz4ByocskuI+bBOVDlkl1GzINzoMolu4yYB+dAlUt2GTEPzoEqH\/F\/\/mir+m0qzwNUWTpFXMbv4I9Uv03leYAqSyf1MtIOcZc7l69TzuDNv4D\/pHO5nJAH6e\/6Um4kPBFVlk6ayqsJuTID1qi36kN7FS36AEEnZjett+58VAOdp6o\/zYMM79rrG9Bc0eY\/aLvRwn9SQo6gytJJa+3zLxy7qE2\/L76\/WM+9ddjzwE3pTTNffAD49z9X\/Vaz50GWdz2ZeOuu4Xl3fa\/fjxbBk05zBFWWTorK\/cn+ogIdTDEPF563HyLVvXyS+eLnAPA9U\/V7zZwHWd71OtgJa2\/djBZh16qTHEGVpZOiMhvBF15Usk4vqr\/wvEE0SmrymKLy7h1x+XdbmgbMJ3mQ7V0TnbtRLyvw9uon5QiqLJ0UlWNdzL3HgTcnVzNYeMO+EX6NxkYKnVHZpD3Wa5r+xvBJHmR716Tsdc+pfJIjqLJ0OFReGUafFE7Bgl4eegkBvLQOy+Qybgz4t7\/7z6rfbNY8yPiuJyvvnMonOYIqS4dDZS\/oeB4t9sFyqsq7vzLMRU\/1m82aB9neNTPZu1uzO8FggSqrI0Xl9sGFGTbd2MKbhF+1q9RYeef8crezxqrfbcY8yPSuh+RZ7sRr970+ud8LFuGTVhgrF06KygO62x2wjvbrPhuO6S\/mxt1dNDozuOO\/pDLjWc\/K5ZM8yPKuJ+yHbFw6SvVu5YWL4EmnOYIqSyfLKJJgdR5fHBzUvXiO\/WXcdrSswzitV87yri9m20mOoMrSydLal3bd0sbO7y+jnuFycmtfDhJyBFWWTmofjHXqhVulHRC7jFqGywl5kP6uOXOk8iqrnzGu6F5hCeHyF18TvvsN4etoLbaauDFt\/\/HGL3XKg\/LAobL6TCw6BQnhsunIx9IpD8oDqnwpBafhMqqsLajyxRSchMuosragypdTcBwuo8ragipfTsFxuDwqQGXd8qAkoMopKTgKlwsw+XIro\/qroCuocloKDsNlVFlbUOXUFByEy6iytqDKqSk4CJdRZW0pl8pffcPJlz\/ifcZXpymIh8tXmLl8FaqymjwoAUJVdgfhbxcbaYcWVB58\/KGIs8TC5XP6TWsd5wGg4Tizxqh2aDKIVFlVHuiP4FI5fXKAYlX+8LmQ0+zD5bP1yk91xxmBTdR9Ot7Fq7KVJ6nS8kB7JKkcmzdPpcrbd98X0ul4Hy6fVXkKI8epdxzndeQsl6RwTlN5uixGZVF5oD1yVI7Nm+f6c461jbb3mC63YJU\/AXwScqIoXD7f2ld\/cqYNWDqmQyMKG15rQMrnlzrUyJ\/TOjSmzozo\/vTgOObMeaqBWYzKwvJAdySVyvt588JV1+hneL5glT8CfBRzpjBcPq\/yS21pzmovM9Mvhqm\/4Jjklm9GVhtL5nDHdGggTQ9ZzuxiVBaXB5ojV+X4SPVh8NsjRar8AeC9oFMF4fJ5lWfw+kQi5pdZpDL5XwsCDPJvScvqumM2XqfEb7N2vl5DrMoC80BvZKjsrpNU7rfT5z4UrPL2HYAhKFAMwuULfTA65KZvCg9OXGUaQAcqs\/\/1kfnaMf1C\/OncibTNA72RoXLXPVB5xX5YZDJfQ\/oJxKr8iY4z\/iToZH64fN5k55UWwfWXA5U7jRErjp9mzojcEjpmfbasUZVnNPw4g8gsEJsHWiNUZTpjHKEZmzdvTueOm9\/N58Z8fgepsx+KVfkjTY2wQJGFyxdUXlJFaXwxgpfZCF5nLzBadgBMovfyAR5ojcXsiWlNi\/DGayEqi80DnSmgtW\/NcaxYlRvv33\/+viHsdDRcdgpAZBbQPHgvMA80plQN11cg8haKhsulU1lwHmgMqswBCZdRZW1BlbnONlY+IEp5HmgLqszF89+gyrqCKnOx\/TGqrCuoMh9lDDBK2fuYH5Eqn58\/bn20rJLKS1ZfnKTkLM3ZM8fcSCkqGoEqn50\/LtjoZZk1T3eVTxqupzbx8YW13E1NeJhGi2AjxazRfnHOa421kwQLf+Py5VTlGylFRSNO5bPzx3mxHvnps+YJfn+iVT4pk\/2eFMzaht8BP1js+ynbtjN7qjlO3XRMM1yEG5en5bzgJN8K4lS+NH\/cfnBJ9\/JJSqeyae9VZt3fRuFir\/Is2AkzZ9YIF+FG52mEKgtBnMqX5o\/bq5w+a55YZKtcX8ZUJutUZX9xOHqEdr4IehlBtItudOyObJVvJPYWp\/KF+eNiKqfPmicW2SpDbNF5cUbE4WBxqPLTMkFlutFZ1lBlIchR+Xj+OL+ozjZrnliKVHlaq5mkSA4WYVENdGFPnVOV2caEoX+Ck4wqn5Bj1jw4OaysKpsXVKblrBlb7B31pa3P2OCoYBGafKoyNpFchTiVL80ft\/c3w6x5QpGt8sOBjq+NZWwROfpKwo2l7TyYjklu+4JFsNGZdlBlIYhTeUB3J80fF80PR8k6a54oZKv8QoPi5QvQ8dMz02TNJf4i2EiLXzYgYenYdahPnWARbgzrQCSqfCP11ALrlYXOmicK2SrHqoWXo\/giK6e\/AIOtfdchuLUvBY5Z8wQhveF6amcy9qzJ5ukvu6DKV0FU3mQ8tOBZ88QgveE6U0+L80wTtt1IQCAa2NisYHYs6DjRYmzA6Txf2DNuV8KxfbcDBDFGa7fr3UeLe2tnnUiAKu9QZY0JVSYPW1hEi81u0zo+FFXelVPlG4m9I5VJPEEdDha7hBtCVHmHKmtMqPKzvVsQh4MFqnwOVFlbQpUdw7BIkRwsUOVzlHFA1I2pTOhZ0eKexMqochKosrbsVR63ttGiY+0svO1Loowq30g9NWxtAJJ3G8uiJgeLt3u4PwnZUOVdMSr\/nayLXW2CInm7iC+SQZV3Rag8\/x34peBE3wY8fTDKOGOcaJV\/0iP84teEXrQWWz3dOP3x5f0nG78A+FPBib4NOFTmR4MZ45Tf8nDnAe39eSPRrVikqvzh91W\/PfUqc8+atyAq3wudcUF5HhSDTJW37z5TPgmG6svIP2seVRmEJlt1HhSETJV1mDFO9WXkzwOmMohs8lOdBwUhU2UdZoxTfRmvyAOmcovzSTrnQUHIVPkDwAfV70\/1Zbxi1jymMtjVyYOCkKgynTHunepgWfFlvGbWPF9lI+vgnnRupD5EospazBinWOVr8qDlu9zjexYiUWUtZoxTrPI1edDxVb5Xm\/LyIVFlwbPmXYdila\/JA1\/l+7HalJcPqU0kqkUqZwosgH\/3ZwLv+m4FVFm\/FCxu5D5NMKiynikQWfOjPg8KAVXWMgUdkU3X6vOgEFBlLVNgwHOV8qAQUGUtU\/BmCeyEoT4PCgFV1jUF4qJl9XlQCKiypikQGC3fSIUIqqxpCoRGyzcBqqxpCoRGyzcBqlz9FNwIqLK+KVDdQbZkoMrapkDYjZ\/6PCgEVFnbFAi78VOfB4WAKmubAmE3furzoBBQZUxBRUCVNU6BoPs+9XlQCKiyvikQdd+HrX0CKLNI6lOADX5coMr6pkDXBr+Hbwn\/9K+Eb6O12GrixrT9xxu\/5E8Wqlz9FIimn\/93tFO5ItdQZa1ToGWDH6qshHKnQMyNn+g8QJWVUO4UiLnxQ5U1zMRbS4GYGz\/ReTAvQOUrKhBRZb1TICJYFp0HBZjsXZEsVFnrFAgJllFlAZRcJPUpEBIsi27tQ5WVUPIUaNlKgiorAVMgHn4x16gypkDHVpJsMq76Btx1Xc8bGADNOfG56ZFNzRWqfKMpEPrrcWLIVK+8eqSPj+RYgyznA6L0YDLx1l0jk8pYr1y9FGjYPS6Tyo8T+jhvEpW7dMXzmmsWZwCqfKMpEHDjp6S1Lyh8ibhDaM9pMd1mG9ZdVBlToEsKMqkM+8XkDvqu1x+yDTR6RpVvNQV5b\/zUqMyUdf3CuQt9745tmKwymYwqVzEFWwNytnEo6YPRZrHy5NGbr8myD\/5tYFaTsQ9GFVPg5J5ZVUnD9aBJHx9X5M5vTgPlLo0vhmTVnWR5+hXJQpW1T4Ft5Yww1DRcD5tg9Ekh7DYNgK5L44sJm8bNRZUxBZqQMUbIxRXJQpVLkALNGvxQZSVUIAX57\/sEgyoroQIpyH\/fJxgcEKWEKqQg730fju0TQBVE0iIFuVxGlQVQEZFUpyBntIwqC6ASIqlPQc5oWb7Krutd078+IOmJqHJFU5AvWpbecL2aEB8HrAPRqg\/tFetBZAy8daxJJOh1PzSg7UYLf6M7OFUZG64xBQnIbu1z\/X6bTOWm53XvWBfPNXhDlzb2sZ1hr\/u7vtfvh4two3tazl+RLFS5HCnQqZXkWLv+ZK8yeXBh7pEiee73Tx4O9kEE2QlrOjAqWIQbve4cVb6RFOjVSnKsnd99M1CZrBOVJ9Dv+1sf95Ew7XUPgdHh8X5X\/MkjqnwjKdCrleRYu1gve++RFMdEZe\/RaDOVg\/iCQXvdn6jMuuK7Bqp8KynI3TtOJJdUXhlGn5TM\/cm62Y3iC3Lv5wV9lY9VDjowA6qMKVBA\/4LKtJzts\/U1+7sdxRe+tHdrtiNYRF3xT1TGyrgKp+D6Yll2vXL7QMdhk4QMxsQPGtZRfDGcs1737b7XJ9uCRbAx+KUBVPkmUpDnxk+2yoM5DXcHAGTH2r\/dm9zBHd06CKuMw173dMfKCxZRV\/z+BFW+mRTkufGTrXKsWtide1fgnv6gAKpc3RTkuPGT3nC9yjRc77zJQb0dqowpKDgFpyXvOpfKq4Rt2HBd5RRcXSrj7ysLoEIiqU6BRg1+qLISKpMCjRr8UGUlVCcF+jT4ocpKqFQKNHEZVVZChVJwdbSMA6IEUCGR1Kfg6mgZVRZAhUTSIAXXRsuosgCqJBKmIARVVkLFUnBVsYwqC6BiIqlOwZU3fqKbVn76DeEf\/4XwTbRGV\/\/6JwkbT1djG\/\/6T5LP9M03P+NPFqpcohRo1EySxPf+L+cTtr89EvjyqHKZUqBPM0kC\/4P7K+PL3xL5hYEqVz8FxbD9PV6Vf1n7XOQnE1UuVwr0LZV\/Apwqb38An4tMAKpcqhRcNU1wIXmwMXhV\/hLgD0WmAFUuVQqumia4kDxoAKfKDnH\/L0SmAFUuVQqumia4iDz45fcBfsCl8nvi\/nuRSUCVy5YC\/mi5iDz4Aa+YJh1s3RCZBKLyJnlP7lnCd1UUSXkKroiWC8iDT5\/xqvyn78gzvhSZBtjYrGB2LOgQed\/uwaC18PcAPQFnr5xI6lNgwI++\/fbbf\/pXwrffhmux1YSN\/5Sy\/3Qjv2LO337+H2tcz9j8Wf0zsSoHMUZrt+vdk4wiL0H+Hguquq6cSOpT8PbnKy17QIzHvA3knd1GaNViqDJ52JI7UFIkLzrkVVrWJu+ZKZUTSYMUaNqZp8er5ZvoNvhIZZISovIbWLRtdGHdGxgr65kCTVXu8D6hJ6SwjBGq\/EyKY1ov+Gx02Mdr+9wScPbqiaQ+BXqq7HA\/hdv9NEKVHcOwSMlsvW1a\/v3eRkQ1XfVEUp+CeQEq898qWbxPER5fRCoTepZfNxf8bQg4e\/VEUp+CAky+YrwzdyOk8PgipvK4RSIL4223NXZbsmaJuATVE0l9CrRUecutsvD4gtzu2UAr3Tfsdo\/WK98vdmPjviXkClRPJPUp0FJl7nBBfHwRFslbOb9HVj2R1KdAS5W5w4Vn8b1VsQ9G2VIgwNS1cJV5wwX+gCQdVLlsKbjoYBuAeeoasDrcs+obcEdnFhsYAM25WJU3vG9yPBaf0ahy2VJwuV4ZDDYZwqB9NOeSP3HNY5\/N2evNBxdPwl2vbL9xPkFCfIEqly4FKSr3WbF8Nz9S+ZHNsTBvEpWp6ymlMm+aec2UEV+gyqVLQYrKLlV12KZTj\/WbQNQlgUVz4gXTlZKtQ2iniMyvMm+oLCO+QJVLl4IUlT06mWlzTaVlU0rTecQm8\/g0kZM7SJjHJo\/KC96aNRnxBapcuhSkqewa\/bk\/L7o7HEJUHDN5g7mku5DSkYMzzbyjtKTEF6hy6VJwOTqghbFBZ+MFb910ySP4x7dZrDx59OZkHym5L6vM3fOYDynxBapcuhRclhBW3hq6fqRx57pAxJ57q6E3YDP0Pq7InR9xe9W+fBa+FG14BxxJiS9Q5dKl4KKDQzo7b3\/tzfswWRlG32j+vy4YNJwYNslyRWKMpgHQdUWqzFvIyokvUOXSpcArAL4U8Q4gkRNfoMqlS4F+KvOGynLiC1S5dCnQTmXeASTcoXVGUOWypUC7AVG8A0i4W7kzgiqXLQXaqcx7Eye+070Pqly2FOimMm99hKz4AlUuXQp0U5l3PIis+AJVLl0KTlV2XS9Dd\/pzJD2RJ828A0hkxReoculSMD\/2bjUhPg5YU\/SqD+0V6zBksB7JbjM4hnaPIzuGBrTdaNEH2i7oJvRd5rmR41RTWnyBKpcuBSdlMutr73d9I+Z271jv+jX7exL0tZhMvHWXbL3re\/1+uFjP\/aPc03KeIzm8A0ikxReoculScKxdf7JXmXaHg7lHiuQ57WUxcSEWRJB1WHvrZrTw\/CElXneeQ2VeNaXFF6hy6VJwrN2dG1OZrBOVJ9CnPZJXUTdlprPf85MZHR4\/YCX65DGHypxNd\/LiC1S5dCk41i7Wqd57JMUx7dX5aJBYmAYOMZVpF6IjlYd91tMo6MV8ncqcpay8+AJVLl0KLqlMO8ORkrk\/WTe7tIOcLy2wsSMr77RU9uYQO8U1KvMOIJEXX6DKpUtB\/4LKtPD1y2JyQ9ems32E40WYyd7dmt3pBYsoWD5ROXuaOQeQ8P\/gZ3ZQ5bKl4Fjl9oHKwyaJI4xJGDSEkg5J1OFOvHbf65P7vWBBDadnWz1erzJnKXvNBFdZQZXLloJjlQdzGu4OWPm7Zrd7tF75bh5XecIKaJftWHnBYm7c3bHDgzqQa1TmLWUlxheoculScKxyrFrYnXtXEFRMX6WyRvEFqly6FJw0aKwmmYw9a3LCLwlkTjPn1Agy4wtUuXQpOC1517lUXiVsy9pwzduVSGZ8gSqXLgW5vM1I1rRwdiWSGl+gyqVLgUYq83ZVlhpfoMqlS4FGKvMOnZYaX6DKpUuBRipzqik3vkCVS5cCfVTm7RokN75AlUuXAn0GRPEOtZYbX6DKpUuBPipzqrmQfC1Q5bKlQBuVedUUP+nkIahy2VKgjcrajE8NQJXLlgJdVOatVJYw6eQhqHLZUvDTbwj\/+C+Eb6K12GrixrT9xxt\/liEdvJXKsuMLVPkGUiCHDufvccqOL1DlG0iBFHgrlaXHF6jyDaRACrzjTaXHF6jyDaRACrzxgvT4AlW+gRTIgLdSWX58gSrfQAoksOW96ZMfX6DKN5ACGW+Ks\/tFAfEFqnwDKRDPgveX7guIL1DlG0iBcLYt3jmeJE0KdQCqzMcXXxO++w3h62gttpqw8buU\/acbv1SdaamZyvvDb5ImnTwEVebDdOSjPtcuwx1eyJp08hBUmQ9Umb\/Jupj4AlXmBFXmDy+KiS9QZU5QZYffy0LiC1SZk1EBKvPW2RYLf3hRTHyBKnNSgMlyxyXnzlD+GuJi4gtUmZNbV9m5otmumPgCVebk1lXuXNGXopj4AlXm5MZVtq9ogC4ovkCVOcmq4yzHAarz7EJuXjNVWUHxBarMyTn9HgCYnMsaTJ2XGkDDr+tYvkANah2brk\/NGtSflmQtdkCJVN52rpKyoPgCVebkbL0y1J7o4uUBHKdGVkYv4Q4qcY38Ne3QPzv0DAcHnKA+1xJxOld9yCROOnkIqszHeZVNVizXR1RlavUoprLz2iASs6J51HCODiiJyuMrKpQpEiedPARV5uO8ykvq5+sDVfcVHvaeMpWXQVkd\/H1wQDlU7l2bqgI63fugynycV9kxYek0ZkxVuw7mMq4yfYTY3\/EDyqDypnNt2VpYfIEqc3JB5WXNHD2Fxj4BORIAYqUyMHeXQeHMDiiLyovrwmRKYfEFqszJ6LzKzlPtYUZXRjRoNuE4Vn5gsbJNbv4ODzhBuz4Ydo5KiMLiC1SZk7MBLkydGTwxdUe0nm36EFN5VCMav9AbPqczdY4OOEF1nh2xfc4xMK+4+AJV5uScfq9AAgZz5oxMsJeNGsCTHwpPTRJl1J+mftEMNZOuxQ\/QXOWN\/dzJ8y1RXHyBKnPiFIDqPAvZvlkd6y1fA0dx8QWqzMnNqEyC+mc79++wFBhfoMqc3ITKzrj3zD\/sKYkC4wtUmZPKDohyFgvbsp47z+SfNRb1s1gFxheoMieVUvlAX8teLER\/IRT6qUSV+ShC5b+wpGN+IU3fg9wvMlhClfmoVKksnSLjC1SZk1OVl7R+OLEjfVr3+9mZY9TnmiCKfSeoMh8nDddTm\/j4whqhpyY8TFlXIdo9OdhIMWvQIDtea\/CwjBb+\/mVCr2XtGq6vpdD4AlXm5KRMZh3u\/Y4WDcd5qrPOnDOINhJs25k9ka110zHNaBF2NDot51XnmTAKjS9QZU6OtTPtvcpAe8CNHFIkjx7iKs+CP2DmzBrRIupDN6qqygVHSqgyH8fa1Zcxlck6UdkG0++LHO\/6Ngu6f0K0CPfbnaqqXGx8gSpzcqxdvD99hxTHRGWnU3s4UZl2HkpWOezAXD2Vi40vUGVOLqk8rdXoSBLTnjWe9kU1sGElU+ecyg5UVOWia2JQZT7MCyrTwtdk64e3fYHJTn3GdgSL8yqrzzUhFBxfoMqcHKv8cKDja4PEETU7CBoiR19J1LG0nQfTMeloEn8R7p92KqpywfEFqszJscovtP5h+UI73jsz\/3bPrkN9FG2kG4CyZDv8eme6CPcHdSCVU5l3ksrcoMp8HKscqxZeXvXby0HFdPVULmDSyUNQZT5OGjSmdiZjz5qc8BsC6nNNBEXHF1RliZ8e9RdF\/ojr1B86vMg0YVslGq6LmHTyENjYrGB2LKC\/dfB2DwZJw9iAK39W6ZDqqZzL24yozjMhFB5f0FKZqdwiL36\/2xmkkCZ\/31s7S4QEqPLNqlx4fBGpTB62sNiRInnRoWX1btMScHZU+VZVLj6+2KtM4gmi8htY1tbfJOKGEFW+VZWLjy8ilZ9JcUxU3j0bHVT5AqhyNoqPLyKVHcOwSMlsvW1aPVT5PDggKhMK4otI5Z3\/C7oQ3PZt2CI36i8KqqwEBfFFTOVxi0QWxttua5CvB2tn4W1fEqhyJhTEF+R2zwYgebdht3u0Xvl+4S9EhGzqLwqqrAIV8UVYJG\/ltDCpvyiosgqKmhTqAOyDwcdPeoRf\/JrQi9Ziq6cbpz++vD9h489UZ1peipp08hBUWTIff6g6BcVT1KSTh6DKkvnwueoUFI+S+AJVlsz23feVXFel71lJfIEqS+YTwCfVeVA0auILVFkyHwE+qs6DolETX6DKkvkA8F51HhSMovgCVZbL9h2AcWPBsqL4AlWWyyc61vqT6kwoFkXxBaosl49U5dsKllXFF+VSefMtNz\/lfobQPl2N9+8\/f9+Qmse6oSq+KJfKi7knnbno3ijqv5mKRVV8gSqjymIpctLJQ1BlVFkoRU46eQiqjCoLRUWnex9UGVUWibr4AlVGlYWiLr5AlVFloaiLL1BlVFkkCuMLVBlVFonC+OKWVF6jytJRGF9UT+U2gEuEhKZL1pi9rgErb2AANLN8ElDlHKiML6qnstcn4nrecOh5YHTphkEbPM+glg5QZbmojC8qqLJ3RwV+JP+hz4rluzlVmW7EUlkyKuOLKqo8JAKv+1Rllwo8bHtAN7YzPRlVzoHaH6OpoMqe0ff6tDgGEmy4XnNNVfYmd9B3UWWpFD3p5CFVVLlvuDS+IAa7Rn\/e9ZjKnteFPqosFaXxRSVVdo320FfZ6xrtNV2Z01K6D6iyTBT\/2F0VVfa6d2wBK28NXeb0nFbErdqoskzUxhfVVJnd9JE7PWBB87wPE7dpAHRdVFkmauOLaqqcB1T5WlT\/mC6qjCqLeqOK5wNClVFlQSiOL1BlVFkQquMLVBlVFvU+Vc83iCqjymJQHV+gyqiyGBbK32e1VHZdL6WL\/TrlDKjylaiYdPKQSqm8mhBXB6x5etWHNm3so7i0L5Hhd1Z20\/oso8rXoTy+qJTKLutp73cearLm6yEppd0m63i\/DjpguCl9ilDlq1Ay6eQhVVK5P9mrTB5cYIcPSTlMiuR52AGje\/kkqPJVqI8vKqXynRtTmaz7Kj+uPW8C\/aiz8uQRVRaP+viiUipDbPFIymGmMo0vyJ9GO1TZNVBl4WgQX1RW5ZVh0CEkfnxBQo91k3bBB9gfhioLRIP4orIqE7rsBo92vaeb1nB0GKosEA3ii0qp3D5QddikhfKaxRfGZB9XrDBWFo4O8UWlVB7Q3e6A9rj31sF93oDVI0\/u4C58blDPgSoLRIf4olIqx6qM3bNHBpXPqLJIdIgvKqUya+1LwU37BQFUmR8t4otqqZzhFw5XaQegyvwomxTqgGqpLIAUlb\/4mvDdbwhfR2ux1YSN36XsP934pdRrIh5lk04egirzqWw68ilbOa5s0slDUGVUOS96xBeoMqqcF03iC1QZVc6LJvEFqsyp8qgAlUXXoUhGk\/gCVeZUuQCTVY9c5kSX+AJVRpVzokt8gSqjyjnRJb5AlVHlfGgTX5RM5a++4eTLH\/E+4yshKs9uRmVt4otyqczPxx8KPmG6iEuzAQAPprN8Icu6SbxuPAAwuZc1mFZMZW3ii6qr\/OFzwSdMr1d+qr2SxxE9EKjOjvPy4kDtie57eYAMpbLyXONA6aSTh1Rb5e277wsuNNJVZiYTl5nK7H9j5oDJiuX6qGoqK5108pBqq\/wJ4JPYM6arDAerLzVn+kDWlrRYfn1wqqayFp3ufaqt8keAj2LPyKdyDRq2Y77SjSYsWfFcLZU1ii8qrvIHgPdiz5hB5eWR1fUli5pr5ujJqZrKGsUX1VZ5+w7AEBssp\/fBaLweqjzt+GtPtYdZNpVL1AdDo\/ii2ip\/or\/j+UnoKdNFNOu0vm1mhyo\/UbVh6szgycmmstpM40Gn+KLaKn+kKosNljOYSOuVadE8NQGC+OIVwHTMmTMywa6SyjrFF9VWufGe0BB6ygwq50ZtpvGgU3xRbZUlpABVjqFVfIEqc4Iqx9AqvkCVOcEBUTG0ii9QZU5Q5T2aJRRV5gNVjuWtXqFQ1VUW3dyAKu\/RK76ovMqiOVV5SRuqk7vaT+2wEftsV\/ykHSXJNd3SiSrzMTrVlfj44jdRm\/BAW\/UoxOHpQ9gccrDfsetQewkXy5dTlUvScK1ZfIEqc3JSJrMu9X57dMNxnurOK7F4SVZn9diAkdh+p0b2QbRYnpbzqt9jRjSLL1BlTo61M+29qlRMYMX2Kylrn+KOxveTsnj04IQL52lUTpV1iy8qr7LsJpL6MqYq7c3JxOyQELjegX25HN9vg2mStWDh2J1yqqxbfIEqc3KsHcQWHVLOMpVpfEE3mQ8Hh4X7O7UH6nCwWNbKqbJu8QWqzMkllae1msk63tP4wg8nyALgaL9pzxpPTrhI6PepOs+yZYT6a3sEqsyHeUFlJwyQaR97B+Id7eP7wb\/fCxYJKqvPtSwZq90nDlXm41jlhwNVXxuskrnB9jyxAaqn+2s2CyqChT\/KpHwqaxdfVF5l2a19Lyw2fqFd652Zfx9Hf\/eCMKpHt32H+22yYxQtwjqQkqm80C+RVVdZNMcqx6qFlynj\/pL3BxXTZVNZi0knD0GV+Thp0JhmGOF0yW9zebKtDLmmX3xBVT7z+RIR1pfhovBxWrLm+qVDJ+kn5ErQcK3HpJOHwMZmBbNjQcfZbVj\/ge1udw8gYrBL9VTO5W1GVL\/HDGgYX9BSmancIsm7342JxVuyOhZUMKhXGQdESUHD+CJSmTxsgQk8Jt8dnZYl5GOHKldTZR3ji73KpDz2VX7e0KqWe6MasTKqLAMd44tI5Wd7t2Aq0\/iCLp5bAs6OKldTZR3ji0hlxzAsWjKz+IKyEVFNVz2VcUDUTtP4IlKZ0GNZ2Am\/OwwBZ1d\/UXBsnwS0jC9iKo9btFDe0LhiS9YsERmq\/UXhBlXeaRpfkNs9G4Dk3cay2K+32vS7Y2zct4Tkp\/YXhRtUWdf4IiySt3JamHS\/KPygyjpNCnUA9sHg4yc9wi9+TehFa7HVxI1p+483\/kz1m7yMPpNOHlJ1ldWnoHLoM+nkIagywomm8QWqjHCia3yBKiOc6BpfoMoIJ7rGF5VXuQTd2MuFtvFF5VVGBKNtfIEqI3xoG1+gyggX+sYXqDLChb7xBaqMcKFvfFF5ldWnoFLoNenkIagywoFek04egiojHOjZ6d4HVUayo3N8gSojHOgcX1ReZWy4FonO8UXlVUYEonV8gSoj2dE6vkCVkexoHV+gykhm9I4vKq+y+hRUB73jC1QZyYze8QWqjGRF95+aQZWRjOg36eQhqDKSEc3ji8qrjK19otA9vqi8yogodI8vUGUkI7rHF6jybbL5lpv\/wv2MTY5X40fIjCPnQZU1ZTH3pDNfFPtqVVdZfQq0BFXmRb1I6lOgJagyL+pFUp8CLUGVeVEvkvoUaAmqzIt6kdSnQEtQZV7Ui4StfYmgyryoVxlJBFXmBVXWFFSZF1RZU66Ua51xm1CV15mOQpVvlDNytQFcYgU0XbLGFHINWFGdmm6\/CQDtvnewZ2AANOfn5Ep5teOXjk67T8bFF7gxldWnQEvOydVn4g6HngdGl24YtIEtBl1jSIXpH+4x6LaBCJUPTrtPxqUXQJWRC3LdUZ8eqVl9VkrezZnKzTUzmRhzuIfpJ6RUPnzBKBmXXgBVRi7INSQ+rfvULJdaNGx71KyVvwici+0ZQnt+Xq7UVztQOf6CUTIuvQCqjFySy+h7fVo6AvmWd0lpzMzqD+Mqx\/dM7qDvilE5ftooGZdeAFVGLsnVN9xH3yzX6M+7HjPrzvUg8ulgj+d1oS9I5fhpw2RceoEbUxlb+xI5L5drtIe+WV7XaK\/ZyopY1RzunYv2zGnB2YdzcqW\/2oHK+xeMknHxBW5MZSSRC3J173yzVt4auoFixKr+3Yr8sZ4c7JnTejISSAtROfaCUTIuvgCqjFyUi91tkdstYNHqvA8TGl8Ql5sAtGiO73GbBkDXFaHywQtGybj4Aqgygg3X\/KDKmoIq86JeZfUp0BJUmRf1IqlPgZagyryoF0l9CrQEVeZFvUjqU6AlqDIv6kVSnwItQZV5US8StvYlkiqX63opnd7XKWfgUVnIq1VdZSSRNLlWE2LPgLW7rfrQpg1xFNfrG9Bc+fql9SLOrrKYV0OVb5IUuVzWC95vQm6yRuShS5vevMnEW3eN4KCUXj6ZVRb0aqjyTZIiV3+yl4s8uMAOHw787\/mwd0\/38kkyqyzo1VDlmyRFLtbpIpSLrPtyPTK1SEEZHDV5vHiSzCoLejVU+SZJkQtii8eBN2dykW98v3h0g6NcQ4zKgl6t6iqrT4GWcMi1Mow+63c\/9O+8Jiu6B2B\/mEiVc7waqnyTcMhFS0Z2y0W7xQduHR4mUuUcr4Yq3yQpcrUP5Bk2aTG5Zt\/4Q\/JE179NY4NLRKgs6NVQ5ZskRa4B3e0OaG94b933h4kO6Df+JKjwZQQ1D7lVFvRqqPJNklbTu6\/Edc8e6XYvnoOjXlnMq1VdZWy4TiRL+1sKbtqYfr7WvvyvVnWVkURSe0WsU+VapR3A0QdDyKuhyjcJ9ozjBVXWFFSZF1RZU1BlXtSrrD4FWoIq86JeJPUp0BJUmRf1IqlPgZagyryoF0l9CrQEVeZFvUjqU6AlqDIv6kXC1r5EUGVe1KuMJIIq84Iqa8riq2+S+fPHMzv+23\/4hpOvFqmvdpa\/\/hP+V0OVkRj\/G351Zo8D\/1BcMra\/PeJ\/EqqM7Pn5935wdt8ffOYUlo6\/\/K0r7nGqrrL6FJSI5e\/Cx7M7\/xh+sC0oHZ9+5\/MrnoUqIyHLGsDb2b1fAjSKScfWgN+74mmoMhLgEJP\/8PzuNwDoFJKQvwT4syuehiojPk6duPqn5\/dv3wF8ZhaQkE\/fB3h\/xfNQZYTBTIa\/v3DEe7L\/ez+XnhASXgD88RVPrLrK2NqXkV9RU39vc+GIv6KufyY9IX9JX+bLK55IVN5c8bSMqFcZycwP\/uDy9\/rfE5H\/Rn4lRsu4VuWNzQpmx4KOs9uwHx7Y7sYGdE4SvfmWm59yP0Pi5wq5iDXefnkxFHZq\/+ufnwtIyOb9+8\/+\/ornQRBjtHa73v1uTPzdktV7a2edlKjFttsjxbLJZKlVwPXZ9Habawr\/UGXysAWWzrFNy+rdpnV8KKpcZTqZvg+3BRTL135cIpXJ58BX+Xnjbzq5ISylyhitZ+MtY0YVUCxfW3kdqvxs7xZMZRpfoMq3xraT8StdfrHsXHvJQpUdw7BoycziC1T51she2Eovlq1rey3B3tkeu+wsZronsTKqfDs42cta6cXy1Y3je5XHLVoo+3d7HWtnVeO2D1XOwjNHHajkYvnNvvaZsLUByPXeWBaLlmx2prd7uD8p50upMtaIZMDm0Udysdy7umUhKJK3GS55KVVG0tnyfadLLZZzfFA4+mCgyhWlx5fpUovl8fjqp4pW2R2wqaoGsAq3GBmehSorZMF7PyGzWH6+vpOH+FIZYo9YKpeAFq89EovlPKeWqPI6KKddVFlnrDf+p0i7SjZ\/YiLkqTyBoQFdb0j+bhvtcJ7XolXGyrgUFleUg\/KK5TwjriSWyuTfyl94rtHP9mRUuWAyt1gfIKtY3vRyPFmuyuHCG95lCzNQ5YJ5virDZRXLuT4iklR224cq99v9TE9GlYsla4e4YyQVy7lGdItX+ZFOwtZfHag8ma8h07NR5ULZXOuOnGL56k5xDPEqr9sATaLzHIbrAcznMFjPjfn8DtoqVMYakUs8X\/3bWVKK5as7xTGwte+Gsa7uuiOnWM73izGo8u3C0bXzFAnF8vWd4hio8s2y5W7mO3i2+GL5+k5xDFT5ZuHsRXSM8GI574ej6ipjDcY5rq2HCxFeLOfoFMdAlW+Uq+vhIkQXyzk6xTFQ5Rvl+nq4EMHFcu7TiVSZ9YFbH25bTdzYX+uUM6DKRZGjHm5\/DqEXK0+nOIZAlVcT4uqAtVOv+tBe0WV7EvuLdsxHlbUgVz1ciNhiOXfAI05lt8sWTOWm53XviNh3q9hf7KCUrhjY2lcI+erhIkQWy7k6xTHEqdyf7FUmDy7MvW7fi\/3F6F4+CVbGFcIV3e2TEFks5\/9YiFM56Mbpq0zWibx3j0DL5eAvxuQRVVZO3nq4CIHFcv5pTsSpDLHF48CbE3mB9u6M\/vIjjMujVlHlAriuu33imYQVy\/k6xTHkqLwyjD4pi1lksf8LwEsbv4oqF0BH3FySworlfJ3iGHJUJtA4Gdbh393+0WFFqYw1GCdw\/RZRCsKKZQHTqIlTuX2g6rDpkk1db9WO\/mKsCo6VUeVjrhmXeh5BxXLOTnEMcSoP6G53AEAK4HW\/T92d37HbvuAvRlDPgSqrYiMsUGYIKpZzdopjCKxX3lcZuwdHxv8KKp9RZVVsBQbKDCHFspAPhODWvhTcffGMKishZ8\/OU4RYmLdTHENkH4x1qsqrtAOwtU8uIrpeHJ9SQBbn7RTHqHrPOCTOW+7W4VMEFMtiAm5U+YZwBFR5nZK\/WM7dKY6BKt8OgjoRnZw2d5kq5hOGKt8OoisvQvIWy\/k7xTGqrjLWYET0RFQTJJG3WBbUzIIq3wpjeVmR00VBETyqfCPIueXzyVcsC+gUx0CVbwPB7dVH5CqWBXSKY6DKN4Hw9uqj0+cplkV9XVRdZawRYfTEjIA6S45iWUSnOAaPyl99w8mXP+J9xleongwktFcfkqNYFtEpjsGhMj8ffyjz7EhWZLRXH3F1sSxuTJVUlT\/8vsyzIxlxpN7y+VxtpJBOcQyZKm\/ffSY\/D5E0xI1KvcS1xbKQTnEMmSp\/Avgk8fRINuRWXoRcWSwL\/CkNmSp\/BPgo8fSZwMo4ae3VR1xXLIvpFMeQqfIHgA8ST5+Jm1dZeuVFyHXlq8A2SIkqb98BvFMdLN+6ymP5lRch1xTLgjrFMSSqTEJl9cHyjatcoMlXFcsif0BRosofqcqqg+XbVrlIk6\/yUmQfJ4kqN94TGvLOn4mbbj1cyOkN98XXhO9+Q\/g6WqOr\/\/2\/Jmw8XY1v\/M\/JZ\/r66y\/5kyW1iQRRiqymEdORzxXfpqhyZZHWyHeTKt92pKoWec3VqDJSJNuWqB5nJ6DKSlCfAjXI7Gs\/KkDlK27XUeVKInXUSAEmX5N6VLmKyB3\/dJMqq6\/VvU2V5XYhukmV1XOTKkvuDIcqK0H990LxyO7WyWHk7GTL1F6iykg2LNnfRCkWTs0a1J+IsC81gAat7pg1TIAntu\/BJo+vNXhYospICvK7EF2uV5526GOHHFQjy9ELeXj525EzA+p0fUp31k3HTKmc1q5eGSmcAjrDXbawQ8tdZ9QgKtOCmJbKDRpnULGf\/KfCjBTUqDJykSK6dV5WueYvSCH8Cg8jP6qgJTP1ut4BWi6Dv79UKquvP1CfgkIppIPyZZVhv7DrYJKY2HwlWps1k200H1DlsqagSIrpap+iMruhW\/qF8xOQg+tsywiYv0tAlcuagiLfbDGDRi73wWB1FI7dcUY0QjYhuA9kkQcJkqnD9Zl\/F3gB7IOhYQqKo1fQm70s4Qu7oetMyZ3fiMXJT69su03K54cnFjg\/mI6Zctt3RbJQ5cpQ1A9epNUrvzagZpKbu2WjBvC0pPHFqFav06h5VGe3fba\/UKjyuTFd5zd+l7JfyJiui9xOa19hJleh4VrTLtfIjo7cL8xkVBlVlsi2I\/m3wOOgyqiyNCTPz3CEph6gyhWgWJN19YBHZU3HdF3kFj4aTqtQk6ugcgHvQPg1uQGVnWLLZFQZVZZEEdMzHHKq8pI2TR91sw\/62O+72s\/OXfOkHajyCZVXuXiTTwPNqU18fGFN0VMTHmjrh9\/HPlw4R\/tn9JcxYemYNWhMneXLqQeSG65RZf0YF2\/yiQdLNkDE7yDUcJynetTHfhZv1Ivtf13SxkDHtp3ZU42smiI8KFzl2cFCusoVb+0b9xT8FvvxNTPtvaqs79so7GP\/FHc0vp\/w+uJLQDc\/jQR4kFflBwCWoGUN6DcFS+ULTIO1B\/D5APTPpdkg6w\/mftCXfJWrTbG\/nxxyfM38TpyBqmSdXOugj32wcE72Ezp+aTajRbrdKVjlxDtXYMNenJcHiJK7749q+ivmaMk+ezXaR2pk7gd9nVL1gEAothKTT1SOdbZ3Oi\/OiKi672NPF6f7WXzhF9xLJ+rdrFplkxXL9VGSystgZckea35vv9F+0BeqnINtr6gZc464pPK0VjNhGetjzxYAR\/tZfEGxp\/FTKFZ5ScV8fXCSVD6zKRz0hSrnoNBuFwccewAHCxYgB33sg8XpfhJ7zmImn6qsouEaHPopa8x4VA4HfRWgcmU\/GkU38cU49uDhQMfXxnLfx95fnO4PRly\/0jso24kGmqhWeVkzR09OFpVj9rJBX6jy1YzFTajLzfGVe2Gx7wvQSzozWRkV9LEf7fvYH+53Xmh8YQfVy2EdSHEqJwYFQG\/nHmZxlZcP0dosrnLDj5WjQV9JYB+MbG9LzQ2fz7HKsWrhZUo\/neT9QcV0gSonJg5o281T6C77PQ86Gsb\/k96e+mUx++yxT+jMDgd9JSE62yupcpH97BM4+T6d2k4elgnB5nUqH\/3Qv\/Pmf3WdWpWUjFf6tWHOnJEJ9P3MHgAaNmufpDTYyog9kgyg9cqkaA4HfaHK17HpqG34OS1Zz3avyETSML9rGq43NiuYHQtYByvHvy++Bzj9CsuV3oxcTm0ZRxcKZ9GRNstINtR7kAgEMUZrt+vdkw\/8PTvJOPFDof4taNq9sFBU3vD5qPcgkVBl8rCFRfhTCp2WlfDJV\/8WUOXCfuziAuo9SCRSmXzUicr3z0DL5YV1b5yeTb1I6lOgGMU3fD4FXIQ8Kj\/buwVRmaxbHT\/PWifHqhdJfQoUS6SuXSSeigK4Ilmhyo5hWKRkZnEG27E5raZTL1IZRxcK5E1B5+QE1HuQCOyrlmkUBpvob0PDt1BAAnQo9s5gq+icnIB6DxLZqzxukYzq9HZOZ7fdJs5owTWma19bmFzpSA7jH9NVQB5qq7KyjnAn6Kry1gYgz9tYFv3IL+7pbd\/YuG8lnItjTFc9mEbleP\/BYfxjum5YZXUd4U7QVeUgo7JEiMcvd35M12vc+vj+w8O4x3TdrsoLLW74fPRWORPHL3d+TNdDw9xHD\/H9h4dxj+m6LltmiaulUtnSJExm\/KRH+MWvCb1oLbaauDFt\/\/HGn\/EnK4\/K58d0jcx6LXFM1+Fh3GO6zukXjiEcRZPG0T6FUDedg3GEQbBzedo41aIkvW19gguNyaPy+TFdpOztNA4OC\/YfHcY7puvcV1s0hnA\/aRztSE3NPRxHyNJyedo4\/eqV7WfFnS7KgSiVD8d00RIwcUzX8WGcY7rOCRiNIdxPGhf2\/z8cR+i\/6MVp43RTefusW4o0Jc8okvNjuii1o8Piv40AB131s4t06YYDYi8aaPxSOx5HCLG95VB5obhHZ3nIo\/LZMV20unn\/FX4wputg6Bf3mK5UlWOfLqixntOH4whLp7KlvB9cacij8tkxXa+1eiM8+HD\/4dAv7jFd6SrvJ43by8rGEcaCnfKovOlo0HuoLORRmX9M19Fh3GO6Lr0KkzOaNC6U9WgcoX\/bd3HaOI2+z8eqO9mXilwjrgsf03XpbKxAjiaNi1Q+HEfo+35x2jjVVyRiq0HX5DKRb8R10WO6LpyLjSHcTxpH\/2Yfl\/g4wiDYuTxtnOorEr3XlkbfD2WgXD9Kqz4FhYH3e7ygylqqvHnWpRtceUCVdVT5rejpRapAuVTWtE+WYLZ+h1uED1RZO5UXLew8dA3lmoJSfQqks+3p1J+zTKDKeqk8xiq4aym7ypdGF0bInjFOGJuO8vCmvJRLZY7RhdFkF0f77TrUXmhbNh18KGjGOFFY2FCdg3KpfHz0+dGF9Heoo8Pi+2usK\/Vs5HfDEDNjnBgWLew7lAcelbnHdE1\/LHhM17F250cXOvbySOVgPymSR36XDNa3WciMcSLA27288KjMzccfCj7hsXbnRxdOR86xyv5+G8IpBFiJLmTGOAHg7V5upKr84fcFn\/BYu7OjC2ngcKRyOGNcp8bGqL6aNVp+C5kxLjd4uycAmSpv330m+DvzksoHwwbNYMTV6ehC0541\/Ag7PtOgWpXxdk8EMlX+BPBJ7Bkzjy70f03AdBL2O1G\/+1qiyoWXj3i7JwaZKn8E+Cj2jNlnjHNOAoxgf80Ogwqbnk3IjHF5wNs9UchU+QPAB7FnzD5jXEzlw\/203\/3IGdXqdfankBnjcoC3e8KQqPL2HcA7sSWOpjPGXQ3e7glEosqfaLj6Segp1Y8uFMkWb\/dEIlHlj1RlscGy+tGF4thazxhbiESiyo337z9\/3xB6ylzeZkRyhoeM8ScuBCO1iUT4XKaVUXmMFXDCQZUVqLx4xhFP4kGVC1fZ6aDIMkCVC1Z50+thtYUUyqWy+h7TOdla+LMAskCVC1QZ699kgioXpzJWW0il6iovuZ8hS2UUWTKVVHk2GtkmodP6N\/APeqiM9W\/SKZfK2UYXDiHGjwuYMS6VBda\/yadcKmejEzP5+79SkoQDxs8W1r\/Jp4oq\/8+9yeq7nm3HWCIXQ\/VU3u5mocj3yn9HcGvhzV5RVEplUvptDVjs6r7JykvDTQ8n3SuOKqncAWvnANg7i8UWqpvVFr2e6iTcFFVRmRbABjzvdjYpizcAhurv9bdn7GtRLFVQeeuHFbs3KywFF2PFscW4pTy6uTnKrzLVmIUV2oD3ekooucqk6GMa2\/oUghvrWXnFyU3Co\/IXXxO++w3h62gttpqw8buU\/acbv8ycGuouvdHTSWPyweph5zdFlOv3lff40TG70dOH7Rjv9dRRSpXDsCJ2o6cBi15H9d3mTVM6lbc6hhUkQrY72NFCLWVTmWqsWVixox2G8FZPOTwqj\/Kbmsqle6awGUSrsILEOlbH1uob4kYpyWyqYe8K3TTebe0ONk\/rQTlUjnpXaMZb7xkbQ3ShBCpvY70rdGJDAgu809MHvVUO4wrdwgq\/ChnbQrRCgsqzM39n+v3Y+Osxi3WMK1hgoddXBCJY5aXZAIAHM5jWBmD0UgNojBwnXGZWOWoG0S2uIOUxBhY6IrZe+an2Sh5HpkMPpTPojegMNqMXfzam0UvqCVi9sp69Kxgbu4PlsaaIVZmZTJxlv6TC5pCuPbG\/nXCZxt\/p2buC4VidHraEaItYleHkj1d4GMWXKQFKLex7rN2N3luvo1uSkANkq0wnFwsnGTNTf\/Vqqlnf4wAMj8uAYJXjskZzmAazmj5B+gle\/kI3jTE8Lgti+2A0Xo9UHtEaOBOiZSqa1dWS8NjC8LgciK2MM9lEpjM7KqJHtAJu+hAtU1GdH\/G3az9jeFwiBDeR0Hplv2iemgAjZ9moATwto2VZVF5YnY61wLCiTOjdcK2C7ZvVebY0C3SQdFDlOJtx77lna\/B5QvhBlaN3Zz8\/W2OscistZRsQJYeFRTR+w9C41Ny6ypuF1XvuYGhcAW5X5e3CphLbWE9REW5RZSKx1XnuYWVbtcin8pLWFB91qZ\/aRxsT95NlYl98uSpHEuPdXfXI1XA9tYmoL6w5emrCA23pmz7Qpr5go+Oc2e\/UAZ6cZUL\/ZUlB62axsFDiapOnMm75xBZM1YbjPNWJuKzl2jnoI5ew\/9X\/VCxPy3mhb44J3Ov0Oh3LGqPEFSePyqa9V5U8LGHkPJlOssqH+x8aJosuno4L+hGRzqLeEa7V+lBgrJu4FfKoXF\/GVCXrRNV6B\/xy90Tlg\/0js16jh9md01J5QzV+Y0Z3CETKju+3TXfEzXTYhuBY\/2gU+HbJozLEFp0XZ0RUpeP5HpJUPt6\/7DToYy1jgLFl0jJjnzs+gcC+4Ausjbh5RKk8rdVMUvKyOGJfFAOc2+\/M4Mh4CbEyckvkqYyDgwWLg2G2j42PDjvc74\/BPlVZTcM1UgHyqPxwoOprg4TDD09BB\/tTlWP7aXW0Sc827aDKiCDyqPxC6x+WL0DH7M1MNgh1VGe3dcFG58z+11q9wXYHdSCoMpKfPCrHqoWXKeP+kvcHFdOoMiKAXA3XU9vJwzLh1wRQZeRKiMpHrWDO2X67pyVrpl80PMs0YRtWCCNXAhubFcyO5c9v7nToWPmxAZ1ToXN5mxHV+YGUFghijNZu17snJfQ9k+ne2lmnX\/WoMqIxocrkYQuLXc8XGDa7TevkWFQZ0ZhIZRJPEJXvn4GWy7BLuiFElRGNCVV+tncLojJZtzqoMlJCQpUdw7BIyczijHMqV2VAFFJJYO8sjZNJkEz\/viexMqqMlIq9yuMWnT66t3NIgNGxdtbpbR+qjGgMbG2g035sLPYD3Yt7dtv35i+OQJURjQmK5G2WVjZUGdEYnj4YP+kRfvFrQi9ai60mbkzbf7zxZ6ozBCkrPCojiMagykhFQJWRioAqIxUBVUYqAqqMVARUGakIqDJSEVBlpCKgykhFQJWRioAqIxUBVUYqAqqMVARUGakIqDJSEVBlpCKgykhFQJWRioAqIxUBVUYqAqqMVARUGakIqDJSEVBlpCKgykhFQJWRioAqIxUBVUYqAqqMVARUGakIqDJSEVBlpCKgykhFQJWRioAqIxUBVUYqAqqMVIQbUnlT0deqWuquRbTKWxvobJY2RJOxGgW\/I5ICA4znN7ZGscia1SIrnULmt4xei738PZ2pttUBYPpsDXDynv8i51\/IsQy4723JpTEAWllmHM2dkgV9oNfC2RWRAeJLZZB0Xr4UOIYdrG2J1D1jTNYWhagcey2gOhN57B0YPbrP7sjOl3Mv5DzTx2fLL1oWtvx82ELw4Fj3uyIyQKLKwdfYdist8ZdSMG4FazQZzC5yBYt49dhrgf+\/tdmBxUql+4V0lc+8EP2WIokimcKcKiInIHzYGrsiMkCeym8wNqC3G5O\/O0Zn91xcgMZSsA082vakvMu09x+u2sbO6ZC1LTVo3JGekvgLkUiHqEsCi9ZbGOaRrWPoFPKRjlRe9N4KyQCJpTL9boHwc2kVk32HKdjRULkj5V2mvbq\/yiyyxnSjBVtWPEt\/9f0L9difxKO3Rfi6dPF2D1YR35S+ypbV6jmFZIBclcPFbnxfYJhxUCrv6Pc9FPnq+9fys4K+dWCf5p78D1X8hbZj8pUYFscsVVv\/rx4UcdcQBRj2fSEZIEnlbedQZauY2oNYCqJYmQbrrXFxrx57Lfby7IaLlpFGZ1OEytELbVpb8gj+12GHxcpvz7sFDfSsIr6mIP5QQAaIPzO7wbCcA5XfFhsoLsKgr7gw3oK1bW+7s+5pFdDmrYhXj70Wy9we+1pwdhsoImqPvZB1vyVfTT1jsXPGO5t+snfPDrnzIxeChu\/SYV8E7HvxvpAMEH\/mTQdohLhbwHhjw2IB9mZhLBb3UET27WjdD63OpfEZWTPIOr2ItK63qKI5fC2akJ0fX4xp7ba12S0skPtxir+QYxiW0fpVDwz6jThukSXJlG2L5ElPfsBF371FH4wW+QAVkQH\/HzSoJGJpDSwkAAAAPHRFWHRjb21tZW50ACBJbWFnZSBnZW5lcmF0ZWQgYnkgR05VIEdob3N0c2NyaXB0IChkZXZpY2U9cG5tcmF3KQr6ixbzAAAAAElFTkSuQmCC\" \/><\/div><\/div>\n<p>\n  Figure 3.1 indicates the genetic relationships among the timesharing operating systems we'll survey.  A few other\n  operating systems (marked in gray, and not necessarily timesharing) are included for context.  Sytems in solid boxes\n  are still live.  The \u2018birth\u2019 are dates of first shipment;<sup><a href=\"#ftn.id2888745\" id=\"id2888745\">[25]<\/a><\/sup>\n  the \u2018death\u2019 dates are generally when the system was end-of-lifed by its vendor.\n<\/p>\n<p>\n  Solid arrows indicate a genetic relationship or very strong\ndesign influence (e.g., a later system with an API deliberately\nreverse-engineered to match an earlier one). Dashed lines indicate\nsignificant design influence.  Dotted lines indicate weak design\ninfluence. Not all the genetic relationships are acknowledged by the\ndevelopers; indeed, some have been officially denied for legal or\ncorporate-strategy reasons but are open secrets in the\nindustry.\n<\/p>\n<p>\n  The \u2018Unix\u2019 box includes all proprietary Unixes,\nincluding both AT&amp;T and early Berkeley versions.  The\n\u2018Linux\u2019 box includes the open-source Unixes, all of which\nlaunched in 1991.  They have genetic inheritance from early Unix\nthrough code that was freed from AT&amp;T proprietary control by the\nsettlement of a 1993 lawsuit.<sup><a href=\"#ftn.id2888778\" id=\"id2888778\">[26]<\/a><\/sup><\/p>\n<h4 id=\"vms\"><a href=\"#vms\">\u00a7<\/a>VMS<\/h4>\n<p>\n  VMS is the proprietary operating system originally developed for\nthe VAX minicomputer from Digital Equipment Corporation.  It was first\nreleased in 1978, was an important production operating system in the\n1980s and early 1990s, and continued to be maintained when DEC was\nacquired by Compaq and Compaq was acquired by Hewlett-Packard.  It is\nstill sold and supported in mid-2003, though little new development\ngoes on in it today.<sup><a href=\"#ftn.id2888820\" id=\"id2888820\">[27]<\/a><\/sup>\n  VMS is surveyed here to show the contrast between\nUnix and other CLI-oriented operating systems from the minicomputer\nera.\n<\/p>\n<p>\n  VMS has full preemptive multitasking, but makes\nprocess-spawning very expensive.  The VMS file system has an elaborate\nnotion of record types (though not attributes).  These traits have all\nthe consequences we outlined earlier on, especially (in VMS's case)\nthe tendency for programs to be huge, clunky monoliths.\n<\/p>\n<p>\n  VMS features long, readable COBOL-like system commands and\ncommand options. It has very comprehensive on-line help (not for APIs,\nbut for the executable programs and command-line syntax). In fact, the\nVMS CLI and its help system are the organizing metaphor of VMS.\nThough X windows\n  has been retrofitted onto the system, the\nverbose CLI remains the most important stylistic influence on program\ndesign.  This has the following major implications:\n<\/p>\n <div><ul><li>\n    <p>\n     The frequency with which people use command-line functions \u2014 the more voluminously you have to type, the less you want to \ndo it.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The size of programs \u2014 people want to type less, so they\nwant to use fewer programs, and write larger ones with more bundled\nfunctions.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The number and types of options your program\naccepts \u2014 they must conform to the syntactic\nconstraints imposed by the help system.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The ease of using the help system \u2014 it's very complete,\nbut search and discovery tools for it are absent and it has poor\nindexing.  This makes acquiring broad knowledge difficult, encourages\nspecialization, and discourages casual programming.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  VMS has a respectable system of internal boundaries. It was\ndesigned for true multiuser operation and fully employs the hardware\nMMU to protect processes from each other.  The system command\ninterpreter is privileged, but the encapsulation of critical functions\nis otherwise reasonably good.  Security cracks against VMS have been\nrare.\n<\/p>\n<p>\n  VMS tools were initially expensive, and its interfaces are\ncomplex.  Enormous volumes of VMS programmer documentation are only\navailable in paper form, so looking up anything is a time-consuming,\nhigh-overhead operation.  This has tended to discourage exploratory\nprogramming and learning a large toolkit. Only since being nearly\nabandoned by its vendor has VMS developed casual programming and a\nhobbyist culture, and that culture is not particularly strong.\n<\/p>\n<p>\n  Like Unix, VMS predated the client\/server distinction.  It was\nsuccessful in its day as a general-purpose timesharing operating\nsystem.  The intended audience was primarily technical users and\nsoftware-intensive businesses, implying a moderate tolerance for\ncomplexity.\n<\/p>\n<h4 id=\"mac_os_contrast\"><a href=\"#mac_os_contrast\">\u00a7<\/a>MacOS<\/h4>\n<p>\n  The Macintosh operating system was designed at Apple in the\nearly 1980s, inspired by pioneering work on GUIs done earlier at\nXerox's Palo Alto Research Center.  It saw its debut with the\nMacintosh in 1984.  MacOS has gone through two significant design\ntransitions since, and is undergoing a third. The first transition was\nthe shift from supporting only a single application at a time to being\nable to cooperatively multitask multiple applications (MultiFinder);\nthe second was the shift from 68000 to PowerPC processors, which both\npreserved backward binary compatibility with 68K applications and\nbrought in an advanced shared library management system for PowerPC\napplications, replacing the original 68K trap instruction-based\ncode-sharing system.  The third was the merger of MacOS design ideas\nwith a Unix-derived infrastructure in MacOS X.  Except where\nspecifically noted, the discussion here applies to pre-OS-X\nversions.\n<\/p>\n<p>\n  MacOS has a very strong unifying idea that is very different\nfrom Unix's: the Mac Interface Guidelines.  These specify in\ngreat detail what an application GUI should look like and how it\nshould behave.  The consistency of the Guidelines influenced the\nculture of Mac users in significant ways.  Not infrequently,\nsimple-minded ports of DOS or Unix programs that did not follow the\nGuidelines have been summarily rejected by the Mac user base and\nfailed in the marketplace.\n<\/p>\n<p>\n  One key idea of the Guidelines is that things stay where you\nput them. Documents, directories, and other objects have\npersistent locations on the desktop that the system doesn't mess with,\nand the desktop context persists through reboots.\n<\/p>\n<p>\n  The Macintosh's unifying idea is so strong that most of the\nother design choices we discussed above are either forced by it or\ninvisible.  All programs have GUIs.  There is no CLI at all.\nScripting facilities are present but much less commonly used than\nunder Unix; many Mac programmers never learn them.  MacOS's\ncaptive-interface GUI metaphor (organized around a single main event\nloop) leads to a weak scheduler without preemption.  The weak\nscheduler, and the fact that all MultiFinder applications run in a\nsingle large address space, implies that it is not practical to use\nseparated processes or even threads rather than polling.\n<\/p>\n<p>\n  MacOS applications are not, however, invariably monster\nmonoliths.  The system's GUI support code, which is partly implemented\nin a ROM shipped with the hardware and partly implemented in shared\nlibraries, communicates with MacOS programs through an event interface\nthat has been quite stable since its beginnings.  Thus, the design of\nthe operating system encourages a relatively clean separation between\napplication engine and GUI interface.\n<\/p>\n<p>\n  MacOS also has strong support for isolating application metadata\nlike menu structures from the engine code.  MacOS files have both a\n\u2018data fork\u2019 (a Unix-style bag of bytes that\ncontains a document or program code) and a \u2018resource fork\u2019\n(a set of user-definable file attributes).  Mac applications tend\nto be designed so that (for example) the images and sound used in them\nare stored in the resource fork and can be modified separately from\nthe application code.\n<\/p>\n<p>\n  The MacOS system of internal boundaries is very weak.  There is\na wired-in assumption that there is but a single user, so there are no\nper-user privilege groups.  Multitasking is cooperative, not\npre-emptive. All MultiFinder applications run in the same address\nspace, so bad code in any application can corrupt anything outside the\noperating system's low-level kernel. Security cracks against MacOS\nmachines are very easy to write; the OS has been spared an epidemic\nmainly because very few people are motivated to crack it.\n<\/p>\n<p>\n  Mac programmers tend to design in the opposite direction from\nUnix programmers; that is, they work from the interface inward, rather\nthan from the engine outward (we'll discuss some of the implications\nof this choice in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">Chapter 20<\/a>).\nEverything in the design of the MacOS conspires to encourage\nthis.\n<\/p>\n<p>\n  The intended role for the Macintosh was as a client operating\nsystem for nontechnical end users, implying a very low tolerance for\ninterface complexity.  Developers in the Macintosh culture became\nvery, very good at designing simple interfaces.\n<\/p>\n<p>\n  The incremental cost of becoming a developer, assuming you have\na Macintosh already, has never been high.  Thus, despite rather\ncomplex interfaces, the Mac developed a strong hobbyist culture early\non.  There is a vigorous tradition of small tools, shareware, and\nuser-supported software.\n<\/p>\n<p>\n  Classic MacOS has been end-of-lifed.  Most of its facilities\nhave been imported into MacOS X, which mates them to a Unix\ninfrastructure derived from the\nBerkeley\n  tradition.<sup><a href=\"#ftn.id2889123\" id=\"id2889123\">[28]<\/a><\/sup>\n  At\nthe same time, leading-edge Unixes such as\nLinux\n  are beginning to borrow ideas like file attributes (a generalization\nof the resource fork) from MacOS.\n<\/p>\n<h4 id=\"os_2\"><a href=\"#os_2\">\u00a7<\/a>OS\/2<\/h4>\n<p>\n  OS\/2 began life as an\nIBM\n  development project\ncalled ADOS (\u2018Advanced DOS\u2019), one of three competitors to\nbecome DOS 4. At that time, IBM and\nMicrosoft\n  were\nformally collaborating to develop a next-generation operating system\nfor the PC.  OS\/2 1.0 was first released in 1987 for the 286, but was\nunsuccessful.  The 2.0 version for the 386 came out in 1992, but by\nthat time the\nIBM\/Microsoft\n  alliance had already fractured.\nMicrosoft went in a different (and more lucrative) direction with\nWindows 3.0. OS\/2 attracted a loyal minority following, but never\nattracted a critical mass of developers and users.  It remained third\nin the desktop market, behind the Macintosh, until being subsumed into\nIBM's Java\n  initiative after 1996.  The last released version was 4.0 in\n1996. Early versions found their way into embedded systems and still,\nas of mid-2003, run inside many of the world's automated teller\nmachines.\n<\/p>\n<p>\n  Like Unix, OS\/2 was built to be preemptively multitasking and\nwould not run on a machine without an MMU (early versions simulated an\nMMU using the 286's memory segmentation).  Unlike Unix, OS\/2 was never\nbuilt to be a multiuser system.  Process-spawning was relatively\ncheap, but IPC was difficult and brittle.  Networking was initially\nfocused on LAN protocols, but a TCP\/IP stack was added in later\nversions.  There were no programs analogous to Unix service daemons,\nso OS\/2 never handled multi-function networking very well.\n<\/p>\n<p>\n  OS\/2 had both a CLI and GUI.  Most of the positive legendry\naround OS\/2 was about the Workplace Shell (WPS), the OS\/2 desktop.\nSome of this technology was licensed from the developers of the\nAmigaOS Workbench,<sup><a href=\"#ftn.id2892797\" id=\"id2892797\">[29]<\/a><\/sup>\n  a pioneering GUI desktop that still as of 2003 has a loyal fan\nbase in Europe.  This is the one area of the design in which OS\/2\nachieved a level of capability which Unix arguably has not yet\nmatched.  The WPS was a clean, powerful, object-oriented design with\nunderstandable behavior and good extensibility.  Years later it would\nbecome a model for Linux's GNOME project.\n<\/p>\n<p>\n  The class-hierarchy design of WPS was one of OS\/2's unifying\nideas.  The other was multithreading.  OS\/2 programmers used threading\nheavily as a partial substitute for IPC between peer processes.\nNo tradition of cooperating program toolkits developed.\n<\/p>\n<p>\n  OS\/2 had the internal boundaries one would expect in a\nsingle-user OS.  Running processes were protected from each other, and\nkernel space was protected from user space, but there were no per-user\nprivilege groups. This meant the file system had no protection against\nmalicious code.  Another consequence was that there was no analog of a\nhome directory; application data tended to be scattered all over the\nsystem.\n<\/p>\n<p>\n  A further consequence of the lack of multiuser capability was\nthat there could be no privilege distinctions in userspace.  Thus,\ndevelopers tended to trust only kernel code. Many system tasks\nthat in Unix would be handled by user-space daemons were jammed\ninto the kernel or the WPS.  Both bloated as a result.\n<\/p>\n<p>\n  OS\/2 had a text vs. binary mode (that is, a mode in which CR\/LF\nwas read as a single end-of-line, versus one in which no such\ninterpretation was performed), but no other file record structure. It\nsupported file attributes, which were used for desktop persistence\nafter the manner of the Macintosh.  System databases were mostly in\nbinary formats.\n<\/p>\n<p>\n  The preferred UI style was through the WPS. User interfaces\ntended to be ergonomically better than Windows, though not up to\nMacintosh standards (OS\/2's most active period was relatively early in\nthe history of MacOS Classic).  Like Unix and Windows, OS\/2's user\ninterface was themed around multiple, independent per-task groups of\nwindows, rather than capturing the desktop for the running\napplication.\n<\/p>\n<p>\n  The intended audience for OS\/2 was business and nontechnical\nend users, implying a low tolerance for interface complexity.  It was\nused both as a client operating system and as a file and print\nserver.\n<\/p>\n<p>\n  In the early 1990s, developers in the OS\/2 community began to\nmigrate to a Unix-inspired environment called EMX that emulated\nPOSIX\n  interfaces.\nPorts of Unix software started routinely showing up under OS\/2 in the\nlatter half of the 1990s.\n<\/p>\n<p>\n  Anyone could download EMX, which included the GNU Compiler\nCollection and other open-source development tools.\nIBM\n  intermittently gave\naway copies of the system documentation in the OS\/2 developer's\ntoolkit, which was posted on many BBSs and FTP sites.  Because of\nthis, the \u201cHobbes\u201d FTP archive of user-developed OS\/2\nsoftware had already grown to over a gigabyte in size by 1995.  A very\nvigorous tradition of small tools, exploratory programming, and\nshareware developed and retained a loyal following for some years\nafter OS\/2 itself was clearly headed for the dustbin of\nhistory.\n<\/p>\n<p>\n  After the release of Windows 95 the OS\/2 community, feeling\nbeleaguered by Microsoft and encouraged by\nIBM, became increasingly\ninterested in Java.  After the Netscape source code release in early\n1998, the direction of migration changed (rather suddenly), toward\nLinux.\n<\/p>\n<p>\n  OS\/2 is interesting as a case study in how far a multitasking\nbut single-user operating-system design can be pushed.  Most of the\nobservations in this case study would apply well to other operating\nsystems of the same general type, notably\nAmigaOS<sup><a href=\"#ftn.id2892968\" id=\"id2892968\">[30]<\/a><\/sup>\n  and GEM.<sup><a href=\"#ftn.id2892980\" id=\"id2892980\">[31]<\/a><\/sup>\n  A wealth of OS\/2\nmaterial is still available on the Web in 2003, including some\ngood histories.<sup><a href=\"#ftn.id2892994\" id=\"id2892994\">[32]<\/a><\/sup><\/p>\n<h4 id=\"nt_contrast\"><a href=\"#nt_contrast\">\u00a7<\/a>Windows NT<\/h4>\n<p>\n  Windows NT (New Technology) is Microsoft's operating system for\nhigh-end personal and server use; it is shipped in several variants\nthat can all be considered the same for our purposes.  All of\nMicrosoft's operating systems since the demise of Windows ME in 2000\nhave been NT-based; Windows 2000 was NT 5, and Windows XP (current in\n2003) is NT 5.1.  NT is genetically descended from VMS, with which\nit shares some important characteristics.\n<\/p>\n<p>\n  NT has grown by accretion, and lacks a unifying metaphor\ncorresponding to Unix's \u201ceverything is a file\u201d or the\nMacOS\n  desktop.<sup><a href=\"#ftn.id2893067\" id=\"id2893067\">[33]<\/a><\/sup>\n  Because core\ntechnologies are not anchored in a small set of persistent central\nmetaphors, they become obsolete every few years.  Each of the technology\ngenerations \u2014 DOS (1981), Windows 3.1 (1992), Windows 95 (1995),\nWindows NT 4 (1996), Windows 2000 (2000), Windows XP (2002), and\nWindows Server 2003 (2003) \u2014 has\nrequired that developers relearn fundamental things in a different\nway, with the old way declared obsolete and no longer well\nsupported.\n<\/p>\n<p>\n  There are other consequences as well:\n<\/p>\n <div><ul><li>\n    <p>\n     The GUI facilities coexist uneasily with the weak, \nremnant command-line interface inherited from DOS and VMS.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Socket programming has no unifying data object analogous to the\nUnix everything-is-a-file-handle, so multiprogramming and network\napplications that are simple in Unix require several more fundamental\nconcepts in NT.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  NT has file attributes in some of its file system types.  They\nare used in a restricted way, to implement access-control lists on\nsome file systems, and don't affect development style very much.  It\nalso has a record-type distinction, between text and binary files,\nthat produces occasional annoyances (both NT and OS\/2 inherited this\nmisfeature from DOS).\n<\/p>\n<p>\n  Though pre-emptive multitasking is supported, process-spawning\nis expensive \u2014 not as expensive as in VMS, but (at about 0.1\nseconds per spawn) up to an order of magnitude more so than on a modern\nUnix.  Scripting facilities are weak, and the OS makes extensive use\nof binary file formats. In addition to the expected consequences we\noutlined earlier are these:\n<\/p>\n <div><ul><li>\n    <p>\n     Most programs cannot be scripted at all.  Programs rely on\ncomplex, fragile\n     <em>remote procedure call<\/em> (RPC)\nmethods to communicate with each other, a rich source of bugs.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     There are no generic tools at all.  Documents and databases\ncan't be read or edited without special-purpose programs.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Over time, the CLI has become more and more neglected because the\nenvironment there is so sparse.  The problems associated with a weak\nCLI have gotten progressively worse rather than better. (Windows\nServer 2003 attempts to reverse this trend somewhat.)\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  System and user configuration data are centralized in a central\nproperties registry rather than being scattered through numerous\ndotfiles and system data files as in Unix.  This also has consequences\nthroughout the design:\n<\/p>\n <div><ul><li>\n    <p>\n     The registry makes the system completely non-orthogonal.\nSingle-point failures in applications can corrupt the registry,\nfrequently making the entire operating system unusable and requiring a\nreinstall.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The\n     <em>registry creep<\/em> phenomenon: as the\nregistry grows, rising access costs slow down all programs.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  NT systems on the Internet are notoriously vulnerable to worms,\nviruses, defacements, and cracks of all kinds.  There are many reasons\nfor this, some more fundamental than others.  The most fundamental is\nthat NT's internal boundaries are extremely porous.\n<\/p>\n<p>\n  NT has access control lists that can be used to implement\nper-user privilege groups, but a great deal of legacy code\nignores them, and the operating system permits this in order not to\nbreak backward compatibility.  There are no security controls on\nmessage traffic between GUI clients, either,<sup><a href=\"#ftn.id2893243\" id=\"id2893243\">[34]<\/a><\/sup>\n  and adding them would also break backward compatibility.\n<\/p>\n<p>\n  While NT will use an MMU, NT versions after 3.5 have the system\nGUI wired into the same address space as the privileged kernel for\nperformance reasons.  Recent versions even wire the webserver into\nkernel space in an attempt to match the speed of Unix-based\nwebservers.\n<\/p>\n<p>\n  These holes in the boundaries have the synergistic effect of\nmaking actual security on NT systems effectively\nimpossible.<sup><a href=\"#ftn.id2893270\" id=\"id2893270\">[35]<\/a><\/sup>\n  If an intruder can get code run as any user at all (e.g., through the\nOutlook email-macro feature), that code can forge messages through the\nwindow system to any other running application.  And any buffer\noverrun or crack in the GUI or webserver can be exploited to take\ncontrol of the entire system.\n<\/p>\n <p id=\"dll_hell\">\n  Because Windows does not handle library versioning\nproperly, it suffers from a chronic configuration problem called\n\u201cDLL hell\u201d, in which installing new programs can randomly\nupgrade (or even downgrade!) the libraries on which existing programs\ndepend.  This applies to the vendor-supplied system libraries as well\nas to application-specific ones: it is not uncommon for an application\nto ship with specific versions of system libraries, and break silently\nwhen it does not have them.<sup><a href=\"#ftn.id2893313\" id=\"id2893313\">[36]<\/a><\/sup><\/p>\n<p>\n  On the bright side, NT provides sufficient facilities to host\nCygwin, which is a compatibility layer implementing Unix at both the\nutilities and the API level, with remarkably few\ncompromises.<sup><a href=\"#ftn.id2893328\" id=\"id2893328\">[37]<\/a><\/sup>\n  Cygwin\npermits C programs to make use of both the Unix and the native APIs,\nand is the first thing many Unix hackers install on such Windows\nsystems as they are compelled by circumstances to make use of.\n<\/p>\n<p>\n  The intended audience for the NT operating systems is primarily \nnontechnical end users, implying a very low tolerance for interface\ncomplexity.  It is used in both client and server roles.\n<\/p>\n<p>\n  Early in its history Microsoft relied on third-party development\nto supply applications.  They originally published full documentation\nfor the Windows APIs, and kept the price of development tools low.\nBut over time, and as competitors collapsed, Microsoft's strategy\nshifted to favor in-house development, they began hiding APIs from\nthe outside world, and development tools grew more expensive.  As\nearly as Windows 95, Microsoft was requiring nondisclosure agreements\nas a condition for purchasing professional-quality development\ntools.\n<\/p>\n<p>\n  The hobbyist and casual-developer culture that had grown up\naround DOS and earlier Windows versions was large enough to be\nself-sustaining even in the face of increasing efforts by Microsoft to\nlock them out (including such measures as certification programs\ndesigned to delegitimize amateurs).  Shareware never went away, and\nMicrosoft's policy began to reverse somewhat after 2000 under market\npressure from open-source operating systems and Java.  However,\nWindows interfaces for \u2018professional\u2019 programming\ncontinued to grow more complex over time, presenting an increasing\nbarrier to casual (or serious!) coding.\n<\/p>\n<p>\n  The result of this history is a sharp dichotomy between the\ndesign styles practiced by amateur and professional NT developers\n\u2014 the two groups barely communicate.  While the hobbyist culture\nof small tools and shareware is very much alive, professional NT\nprojects tend to produce monster monoliths even bulkier than those\ncharacteristic of \u2018elitist\u2019 operating systems like\nVMS.\n<\/p>\n<p>\n  Unix-like shell facilities, command sets, and library APIs \nare available under Windows through third-party libraries including\nUWIN, Interix, and the open-source Cygwin.\n<\/p>\n<h4 id=\"beos\"><a href=\"#beos\">\u00a7<\/a>BeOS<\/h4>\n<p>\n  Be, Inc. was founded in 1989 as a hardware vendor, building\npioneering multiprocessing machines around the PowerPC chip.  BeOS was\nBe's attempt to add value to the hardware by inventing a new,\nnetwork-ready operating system model incorporating the lessons of both\nUnix and the MacOS family, without being either.  The result was a\ntasteful, clean, and exciting design with excellent performance in its\nchosen role as a multimedia platform.\n<\/p>\n<p>\n  BeOS's unifying ideas were \u2018pervasive threading\u2019,\nmultimedia flows, and the file system as database.  BeOS was designed\nto minimize latency in the kernel, making it well-suited for\nprocessing large volumes of data such as audio and video streams \nin real time.  BeOS \u2018threads\u2019 were actually lightweight\nprocesses in Unix terminology, since they supported thread-local\nstorage and therefore did not necessarily share all address spaces.\nIPC via shared memory was fast and efficient.\n<\/p>\n<p>\n  BeOS followed the Unix model in having no file structure above\nthe byte level. Like the MacOS, it supported and used file\nattributes.  In fact, the BeOS file system was actually a database\nthat could be indexed by any attribute.\n<\/p>\n<p>\n  One of the things BeOS took from Unix was intelligent design of\ninternal boundaries.  It made full use of an MMU, and sealed running\nprocesses off from each other effectively.  While it presented as a\nsingle-user operating system (no login), it supported Unix-like\nprivilege groups in the file system and elsewhere in the OS internals.\nThese were used to protect system-critical files from being touched by\nuntrusted code; in Unix terms, the user was logged in as an anonymous\nguest at boot time, with the only other \u2018user\u2019 being root.\nFull multiuser operation would have been a small change to the upper\nlevels of the system, and there was in fact a BeLogin utility.\n<\/p>\n<p>\n  BeOS tended to use binary file formats and the native database\nbuilt into the file system, rather than Unix-like textual\nformats.\n<\/p>\n<p>\n  The preferred UI style of BeOS was GUI, and it leaned heavily on\nMacOS experience in interface design.  CLI and scripting were,\nhowever, also fully supported.  The command-line shell of BeOS was a\nport of\nbash(1),\nthe dominant open-source Unix shell, running through a\nPOSIX\n  compatibility\nlibrary.  Porting of Unix CLI software was, by design, trivially easy.\nInfrastructure to support the full panoply of scripting, filters, and\nservice daemons that goes with the Unix model was in place.\n<\/p>\n<p>\n  BeOS's intended role was as a client operating system\nspecialized for near-real-time multimedia processing (especially sound\nand video manipulation).  Its intended audience included technical and\nbusiness end users, implying a moderate tolerance for interface\ncomplexity.\n<\/p>\n<p>\n  Entry barriers to BeOS development were low; though the\noperating system was proprietary, development tools were inexpensive\nand full documentation was readily available.  The BeOS effort began\nas part of one of the efforts to unseat Intel's hardware with RISC\ntechnology, and was continued as a software-only effort after the\nInternet explosion.  Its strategists were paying attention during\nLinux's formative period in the early 1990s, and were fully aware of\nthe value of a large casual-developer base.  In fact they succeeded in\nattracting an intensely loyal following; as of 2003 no fewer\nthan five separate projects are attempting to resurrect BeOS in open\nsource.\n<\/p>\n<p>\n  Unfortunately, the business strategy surrounding BeOS was not as\nastute as the technical design.  The BeOS software was originally\nbundled with dedicated hardware, and marketed with only vague hints\nabout intended applications. Later (1998) BeOS was ported to generic PCs\nand more closely focused on multimedia applications, but never\nattracted a critical mass of applications or users.  BeOS finally\nsuccumbed in 2001 to a combination of anticompetitive maneuvering by\nMicrosoft (lawsuit in progress as of 2003) and competition\nfrom variants of Linux\n  that had been adapted for multimedia\nhandling.\n<\/p>\n<h4 id=\"mvs\"><a href=\"#mvs\">\u00a7<\/a>MVS<\/h4>\n<p>\n  MVS (Multiple Virtual Storage) is IBM's flagship operating\nsystem for its mainframe computers. Its roots stretch back to OS\/360,\nwhich began life in the mid-1960s as the operating system IBM wanted\nits customers to use on the then-new System\/360 computer systems.\nDescendants of this code remain at the heart of today's IBM mainframe\noperating systems.  Though the code has been almost entirely\nrewritten, the basic design is largely untouched; backward\ncompatibility has been religiously maintained, to the point that\napplications written for OS\/360 run unmodified on the MVS of 64-bit\nz\/Series mainframe computers three architectural generations\nlater.\n<\/p>\n<p>\n  Of all the operating systems surveyed here, MVS is the only one\nthat could be considered older than Unix (the ambiguity stems from the\ndegree to which it has evolved over time).  It is also the least\ninfluenced by Unix concepts and technology, and represents the\nstrongest design contrast with Unix.  The unifying idea of MVS is that\nall work is batch; the system is designed to make the most efficient\npossible use of the machine for batch processing of huge amounts of\ndata, with minimal concessions to interaction with human users.\n<\/p>\n<p>\n  Native MVS terminals (the 3270 series) operate only in block\nmode.  The user is presented with a screen that he fills in,\nmodifying local storage in the terminal.  No interrupt is presented to\nthe mainframe until the user presses the send key.  Character-level\ninteraction, in the manner of Unix's raw mode, is impossible.\n<\/p>\n<p>\n  TSO, the closest equivalent to the Unix interactive environment,\nis limited in native capabilities.  Each TSO user is represented to\nthe rest of the system as a simulated batch job. The facility is\nexpensive \u2014 so much so that its use is typically limited to\nprogrammers and support staff.  Ordinary users who need to merely run\napplications from a terminal almost never use TSO. Instead, they work\nthrough transaction monitors, a kind of multiuser application server\nthat does cooperative multitasking and supports asynchronous I\/O. In\neffect, each kind of transaction monitor is a specialized timesharing\nplugin (almost, but not entirely unlike a webserver running CGI).\n<\/p>\n<p>\n  Another consequence of the batch-oriented architecture is that\nprocess spawning is a slow operation. The I\/O system deliberately\ntrades high setup cost (and associated latency) for better\nthroughput. These choices are a good match for batch operation, but\ndeadly to interactive response. A predictable result is that TSO users\nnowadays spend almost all their time inside a dialog-driven interactive\nenvironment, ISPF. It is rare for a programmer to do anything inside\nnative TSO except start up an instance of ISPF. This does away with\nprocess-spawn overhead, at the cost of introducing a very large\nprogram that does everything but start the machine room\ncoffeepot.\n<\/p>\n<p>\n  MVS uses the machine MMU; processes have separate address\nspaces. Interprocess communication is supported only through shared\nmemory. There are facilities for threading (which MVS calls\n\u201csubtasking\u201d), but they are lightly used, mainly because\nthe facility is only easily accessible from programs written in\nassembler. Instead, the typical batch application is a short series of\nheavyweight program invocations glued together by JCL (Job Control\nLanguage) which provides scripting, though in a notoriously difficult\nand inflexible way. Programs in a job communicate through temporary\nfiles; filters and the like are nearly impossible to do in a usable\nmanner.\n<\/p>\n<p>\n  Every file has a record format, sometimes implied (inline input\nfiles in JCL are implied to have an 80-byte fixed-length record format\ninherited from punched cards, for example), but more often explicitly\nspecified. Many system configuration files are in text format, but\napplication files are usually in binary formats specific to the\napplication.  Some general tools for examining files have evolved out\nof sheer necessity, but it is still not an easy problem to solve.\n<\/p>\n<p>\n  File system security was an afterthought in the original\ndesign. However, when security was found to be necessary, IBM added it\nin an inspired fashion: They defined a generic security API, then made\nall file access requests pass by that interface before being\nprocessed. As a result, there are at least three competing security\npackages with differing design philosophies \u2014 and all of them\nare quite good, with no known cracks against them between 1980 and\nmid-2003. This variety allows an installation to select the package\nthat best suits local security policy.\n<\/p>\n<p>\n  Networking facilities are another afterthought. There is no\nconcept of one interface for both network connections and local files;\ntheir programming interfaces are separate and quite different. This\ndid allow TCP\/IP to supplant IBM's native SNA (Systems Network\nArchitecture) as the network protocol of choice fairly seamlessly. It\nis still common in 2003 to see both in use at a given\ninstallation, but SNA is dying out.\n<\/p>\n<p>\n  Casual programming for MVS is almost nonexistent except within\nthe community of large enterprises that run MVS. This is not due so\nmuch to the cost of the tools themselves as it is to the cost of the\nenvironment \u2014 when one must spend several million dollars on the\ncomputer system, a few hundred dollars a month for a compiler is\nalmost incidental.  Within that community, however, there is a\nthriving culture of freely available software, mainly programming and\nsystem-administration tools. The first computer user's group, SHARE,\nwas founded in 1955 by IBM users, and is still going strong\ntoday.\n<\/p>\n<p>\n  Considering the vast architectural differences, it is a\nremarkable fact that MVS was the first non-System-V operating system\nto meet the Single Unix Specification (there is less to this than\nmeets the eye, however, as ports of Unix software from elsewhere have\na strong tendency to founder on ASCII-vs.-EBCDIC character-set\nissues). It's possible to start a Unix shell from TSO; Unix\nfile systems are specially formatted MVS data sets.  The MVS Unix\ncharacter set is a special EBCDIC codepage with newline and linefeed\nswapped (so that what appears as linefeed to Unix appears like newline\nto MVS), but the system calls are real system calls implemented in the\nMVS kernel.\n<\/p>\n<p>\n  As the cost of the environment drops into the hobbyist range,\nthere is a small but growing group of users of the last public-domain\nversion of MVS (3.8, dating from 1979). This system, as well as\ndevelopment tools and the emulator to run them, are all available for\nthe cost of a CD.<sup><a href=\"#ftn.id2893754\" id=\"id2893754\">[38]<\/a><\/sup><\/p>\n<p>\n  The intended role of MVS has always been in the back\noffice. Like VMS and Unix itself, MVS predates the server\/client\ndistinction. Interface complexity for back-office users is not only\ntolerated but expected, in the name of making the computer spend\nfewer expensive resources on interfaces and more on the work it's\nthere to get done.\n<\/p>\n<h4 id=\"id2893777\"><a href=\"#id2893777\">\u00a7<\/a>VM\/CMS<\/h4>\n<p>\n  VM\/CMS is IBM's\n  <em>other<\/em>\n  mainframe operating\nsystem.  Historically speaking, it is Unix's uncle: the common\nancestor is the CTSS\n  system, developed at MIT around 1963 and running on the IBM 7094\nmainframe.  The group that developed CTSS then went on to write\nMultics, the immediate ancestor of Unix.  IBM established a group in\nCambridge to write a timesharing system for the IBM 360\/40, a modified\n360 with (for the first time on an IBM system) a paging\nMMU.<sup><a href=\"#ftn.id2893806\" id=\"id2893806\">[39]<\/a><\/sup>\n  The MIT\nand IBM programmers continued to interact for many years thereafter,\nand the new system got a user interface that was very CTSS-like,\ncomplete with a shell named EXEC and a large supply of utilities\nanalogous to those used on Multics and later on Unix.\n<\/p>\n<p>\n  In another sense, VM\/CMS and Unix are funhouse mirror images of\none another.  The unifying idea of the system, provided by the VM\ncomponent, is virtual machines, each of which looks exactly like the\nunderlying physical machine.  They are preemptively multitasked, and\nrun either the single-user operating system CMS or a complete\nmultitasking operating system (typically MVS, Linux, or another\ninstance of VM itself).  Virtual machines correspond to Unix\nprocesses, daemons, and emulators, and communication between them is\naccomplished by connecting the virtual card punch of one machine to\nthe virtual card reader of another.  In addition, a layered tools\nenvironment called CMS Pipelines is provided within CMS, directly\nmodeled on Unix's pipes but architecturally extended to support\nmultiple inputs and outputs.\n<\/p>\n<p>\n  When communication between them has not been explicitly set up,\nvirtual machines are completely isolated from each other.  The\noperating system has the same high reliability, scalability, and\nsecurity as MVS, and has far greater flexibility and is much easier to\nuse.  In addition, the kernel-like portions of CMS do not need to be\ntrusted by the VM component, which is maintained completely\nseparately.\n<\/p>\n<p>\n  Although CMS is record-oriented, the records are essentially\nequivalent to the lines that Unix textual tools use.  Databases are\nmuch better integrated into CMS Pipelines than is typically the case\non Unix, where most databases are quite separate from the operating\nsystem.  In recent years, CMS has been augmented to fully support the\nSingle Unix Specification.\n<\/p>\n<p>\n  The UI style of CMS is interactive and conversational, very\nunlike MVS but like VMS and Unix.  A full-screen editor called XEDIT\nis heavily used.\n<\/p>\n<p>\n  VM\/CMS predates the client\/server distinction, and is nowadays\nused almost entirely as a server operating system with emulated IBM\nterminals.  Before Windows came to dominate the desktop so completely,\nVM\/CMS provided word-processing services and email both internally to\nIBM and between mainframe customer sites \u2014 indeed, many VM systems\nwere installed exclusively to run those applications because of VM's\nready scalability to tens of thousands of users.\n<\/p>\n<p>\n  A scripting language called Rexx supports programming in a style\nnot unlike shell, awk, Perl or Python.  Consequently, casual\nprogramming (especially by system administrators) is very important on\nVM\/CMS.  Free cycles permitting, admins often prefer to run production\nMVS in a virtual machine rather than directly on the bare iron, so\nthat CMS is also available and its flexibility can be taken advantage\nof.  (There are CMS tools that permit access to MVS\nfile systems.)\n<\/p>\n<p>\n  There are even striking parallels between the history of VM\/CMS\nwithin IBM and Unix within Digital Equipment Corporation (which made\nthe hardware that Unix first ran on).  It took IBM years to understand\nthe strategic importance of its unofficial timesharing system, and\nduring that time a community of VM\/CMS programmers arose that was\nclosely analogous in behavior to the early Unix community.  They\nshared ideas, shared discoveries about the system, and above all\nshared source code for utilities.  No matter how often IBM tried to\ndeclare VM\/CMS dead, the community \u2014 which included IBM's own\nMVS system developers! \u2014 insisted on keeping it alive.  VM\/CMS\neven went through the same cycle of de facto open source to closed\nsource back to open source, though not as thoroughly as Unix\ndid.\n<\/p>\n<p>\n  What VM\/CMS lacks, however, is any real analog to C.  Both VM\nand CMS were written in assembler and have remained so implemented.\nThe nearest equivalent to C was various cut-down versions of PL\/I that\nIBM used for systems programming, but did not share with its\ncustomers.  Therefore, the operating system remains trapped on its\noriginal architectural line, though it has grown and expanded as the\n360 architecture became the 370 series, the XA series, and finally the\ncurrent z\/Series.\n<\/p>\n<p>\n  Since the year 2000, IBM has been promoting VM\/CMS on mainframes\nto an unprecedented degree \u2014 as ways to host thousands of\nvirtual Linux machines at once.\n<\/p>\n<h4 id=\"linux\"><a href=\"#linux\">\u00a7<\/a>Linux<\/h4>\n<p>\n  Linux, originated by Linus Torvalds in 1991, leads the pack of\nnew-school open-source Unixes that have emerged since 1990 (also\nincluding FreeBSD, NetBSD, OpenBSD, and Darwin), and is representative of the design\ndirection being taken by the group as a whole.  The trends in it can\nbe taken as typical for this entire group.\n<\/p>\n<p>\n  Linux does not include any code from the original Unix source\ntree, but it was designed from Unix standards to behave like a Unix.\nIn the rest of this book, we emphasize the continuity between Unix and\nLinux.  That continuity is extremely strong, both in terms of\ntechnology and key developers \u2014 but here we emphasize some\ndirections Linux is taking that mark a departure from\n\u2018classical\u2019 Unix tradition.\n<\/p>\n<p>\n  Many developers and activists in the Linux community have\nambitions to win a substantial share of end-user desktops.  This makes\nLinux's intended audience quite a bit broader than was ever the case\nfor the old-school Unixes, which have primarily aimed at the server\nand technical-workstation markets.  This has implications for the\nway Linux hackers design software.\n<\/p>\n<p>\n  The most obvious change is a shift in preferred interface\nstyles.  Unix was originally designed for use on teletypes and slow\nprinting terminals.  Through much of its lifetime it was strongly\nassociated with character-cell video-display terminals lacking either\ngraphics or color capabilities.  Most Unix programmers stayed firmly\nwedded to the command line long after large end-user applications had\nmigrated to X-based GUIs, and the design of both Unix\noperating systems and their applications have continued to reflect\nthis fact.\n<\/p>\n<p>\n  Linux users and developers, on the other hand, have been\nadapting themselves to address the nontechnical user's fear of CLIs.\nThey have moved to building GUIs and GUI tools much more intensively\nthan was the case in old-school Unix, or even in contemporary\nproprietary Unixes.  To a lesser but significant extent, this is true\nof the other open-source Unixes as well.\n<\/p>\n<p>\n  The desire to reach end users has also made Linux developers\nmuch more concerned with smoothness of installation and software\ndistribution issues than is typically the case under proprietary Unix\nsystems.  One consequence is that Linux features binary-package\nsystems far more sophisticated than any analogs in proprietary\nUnixes, with interfaces designed (as of 2003, with only mixed\nsuccess) to be palatable to nontechnical end users.\n<\/p>\n<p>\n  The Linux community wants, more than the old-school Unixes ever\ndid, to turn their software into a sort of universal pipefitting for\nconnecting together other environments.  Thus, Linux features support\nfor reading and (often) writing the file system formats and networking\nmethods native to other operating systems.  It also supports\nmultiple-booting with them on the same hardware, and simulating them\nin software inside Linux itself.  The long-term goal is subsumption;\nLinux emulates so it can absorb.<sup><a href=\"#ftn.id2894051\" id=\"id2894051\">[40]<\/a><\/sup><\/p>\n<p>\n  The goal of subsuming the competition, combined with the drive\nto reach the end-user, has motivated Linux developers to adopt design\nideas from non-Unix operating systems to a degree that makes\ntraditional Unixes look rather insular.  Linux applications using\nWindows .INI format files for configuration is a minor example we'll\ncover in\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\"> Chapter 10<\/a>;\nLinux 2.5's incorporation of extended file attributes, which among\nother things can be used to emulate the semantics of the Macintosh\nresource fork, is a recent major one at time of writing.\n<\/p>\n <blockquote>\n  <p>\n   But the day Linux gives the Mac diagnostic that you can't open a\nfile because you don't have the application is the day Linux becomes\nnon-Unix.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  The remaining proprietary Unixes (such as Solaris, HP-UX, AIX,\netc.) are designed to be big products for big IT budgets.  Their\neconomic niche encourages designs optimized for maximum power on\nhigh-end, leading-edge hardware.  Because Linux has part of its roots\namong PC hobbyists, it emphasizes doing more with less.  Where\nproprietary Unixes are tuned for multiprocessor and server-cluster\noperation at the expense of performance on low-end hardware, core\nLinux developers have explicitly chosen not to accept more complexity\nand overhead on low-end machines for marginal performance gains on\nhigh-end hardware.\n<\/p>\n<p>\n  Indeed, a substantial fraction of the Linux user community is\nunderstood to be wringing usefulness out of hardware as technically\nobsolete today as Ken Thompson's PDP-7 was in 1969. As a consequence,\nLinux applications are under pressure to stay lean and mean that their\ncounterparts under proprietary Unix do not experience.\n<\/p>\n<p>\n  These trends have implications for the future of Unix as a\nwhole, a topic we'll return to in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">Chapter 20<\/a>.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2888666\" id=\"ftn.id2888666\">[24]<\/a><\/sup>\n  See the\n  <a href=\"http:\/\/www.osdata.com\/\">\n   OSData website<\/a>.\n<\/p><p><sup><a href=\"#id2888745\" id=\"ftn.id2888745\">[25]<\/a><\/sup>\n  Except\nfor Multics which exerted most of its influence between the time its\nspecifications were published in 1965 and when it actually shipped in\n1969.\n<\/p><p><sup><a href=\"#id2888778\" id=\"ftn.id2888778\">[26]<\/a><\/sup>\n  For details on the\nlawsuit, see Marshall Kirk McKusick's paper in <sup><a href=\"#OpenSources\" title=\"[OpenSources]\">[OpenSources]<\/a><\/sup>.\n<\/p><\/div>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2888820\" id=\"ftn.id2888820\">[27]<\/a><\/sup>\n  More information is available at the\n  <a href=\"http:\/\/www.openvms.org\">\n   OpenVMS.org site<\/a>.\n<\/p><p><sup><a href=\"#id2889123\" id=\"ftn.id2889123\">[28]<\/a><\/sup>\n  MacOS X actually consists of two\nproprietary layers (ports of the OpenStep and Classic Mac GUIs)\nlayered over an open-source Unix core (Darwin).\n<\/p><\/div>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2892797\" id=\"ftn.id2892797\">[29]<\/a><\/sup>\n  In return for some Amiga technology,\nIBM gave Commodore a license for its REXX scripting language. The deal\nis described at\n  <a href=\"http:\/\/www.os2bbs.com\/os2news\/OS2Warp.html\">\n   http:\/\/www.os2bbs.com\/os2news\/OS2Warp.html<\/a>.\n<\/p><p><sup><a href=\"#id2892968\" id=\"ftn.id2892968\">[30]<\/a><\/sup><a href=\"http:\/\/os.amiga.com\/\">\n   AmigaOS\nPortal<\/a>.\n<\/p><p><sup><a href=\"#id2892980\" id=\"ftn.id2892980\">[31]<\/a><\/sup><a href=\"http:\/\/www.geocities.com\/SiliconValley\/Vista\/6148\/gem.html\">\n   The\nGEM Operating System<\/a>.\n<\/p><p><sup><a href=\"#id2892994\" id=\"ftn.id2892994\">[32]<\/a><\/sup>\n  See, for example, the\n  <a href=\"http:\/\/www.os2voice.org\/\">\n   OS Voice<\/a>\n  and\n  <a href=\"http:\/\/www.os2bbs.com\/\">\n   OS\/2 BBS.COM<\/a>\n  sites.\n<\/p><p><sup><a href=\"#id2893067\" id=\"ftn.id2893067\">[33]<\/a><\/sup>\n  Perhaps. It has been argued that the unifying\nmetaphor of all Microsoft operating systems is \u201cthe customer\nmust be locked in\u201d.\n<\/p><p><sup><a href=\"#id2893243\" id=\"ftn.id2893243\">[34]<\/a><\/sup><a href=\"http:\/\/security.tombom.co.uk\/shatter.html\">\n   http:\/\/security.tombom.co.uk\/shatter.html<\/a>\n<\/p><p><sup><a href=\"#id2893270\" id=\"ftn.id2893270\">[35]<\/a><\/sup>\n  Microsoft actually admitted publicly that NT\nsecurity is impossible in March 2003. See\n  <a href=\"http:\/\/www.microsoft.com\/technet\/treeview\/default.asp?url=\/technet\/security\/bulletin\/MS03-010.asp\">\n   http:\/\/www.microsoft.com\/technet\/treeview\/default.asp?url=\/technet\/security\/bulletin\/MS03-010.asp<\/a>.\n<\/p><p><sup><a href=\"#id2893313\" id=\"ftn.id2893313\">[36]<\/a><\/sup>\n  The DLL hell problem is\nsomewhat mitigated by the .NET development framework, which handles\nlibrary versioning \u2014 but as of 2003 .NET only ships on the\nhighest-end server versions of NT.\n<\/p><p><sup><a href=\"#id2893328\" id=\"ftn.id2893328\">[37]<\/a><\/sup>\n  Cygwin is largely compliant with the\nSingle Unix Specification, but programs requiring direct hardware\naccess run into limitations in the Windows kernel that hosts it.\nEthernet cards are notoriously problematic.\n<\/p><p><sup><a href=\"#id2893754\" id=\"ftn.id2893754\">[38]<\/a><\/sup><a href=\"http:\/\/www.cbttape.org\/cdrom.htm\">\n   http:\/\/www.cbttape.org\/cdrom.htm<\/a>\n<\/p><p><sup><a href=\"#id2893806\" id=\"ftn.id2893806\">[39]<\/a><\/sup>\n  The development machine and initial target was a 40\nwith customized microcode, but it proved insufficiently powerful;\nproduction deployment was on the 360\/67.\n<\/p><p><sup><a href=\"#id2894051\" id=\"ftn.id2894051\">[40]<\/a><\/sup>\n  The results of Linux's\nemulate-and-subsume strategy differ noticeably from the\nembrace-and-extend practiced by some of its competitors. For\nstarters, Linux does not break compatibility with what it is\nemulating in order to lock customers into the \u201cextended\u201d\nversion.\n<\/p><\/div>\n\n<h3 id=\"ch03s03\"><a href=\"#ch03s03\">\u00a7<\/a>What Goes Around, Comes Around<\/h3>\n<p>\n  We attempted to select for comparison timesharing systems that\neither are now or have in the past been competitive with Unix.  The\nfield of plausible candidates is not wide.  Most\n(Multics, ITS, DTSS,\nTOPS-10, TOPS-20, MTS, GCOS, MPE and perhaps a\ndozen others) are so long dead that they are fading from the\ncollective memory of the computing field.  Of those we surveyed, VMS\nand OS\/2 are moribund, and MacOS has been subsumed by a Unix\nderivative. MVS and VM\/CMS were limited to a single proprietary\nmainframe line.  Only Microsoft Windows remains as a viable competitor\nindependent of the Unix tradition.\n<\/p>\n<p>\n  We pointed out Unix's strengths in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">Chapter 1<\/a>, and they are certainly part of the\nexplanation.  But it's actually more instructive to look at the\nobverse of that answer and ask which weaknesses in Unix's competitors\ndid them in.\n<\/p>\n<p>\n  The most obvious shared problem is nonportability.  Most of\nUnix's pre-1980 competitors were tied to a single hardware platform,\nand died with that platform.  One reason VMS survived long enough to\nmerit inclusion here as a case study is that it was successfully\nported from its original VAX hardware to the Alpha processor (and in\n2003 is being ported from Alpha to Itanium).  MacOS successfully\nmade the jump from the Motorola 68000 to PowerPC chips in the late\n1980s.  Microsoft Windows\n  escaped this problem by being in the\nright place when commoditization flattened the market for\ngeneral-purpose computers into a PC monoculture.\n<\/p>\n<p>\n  From 1980 on, another particular weakness continually reemerges\nas a theme in different systems that Unix either steamrollered or\noutlasted: an inability to support networking gracefully.\n<\/p>\n<p>\n  In a world of pervasive networking, even an operating system\ndesigned for single-user use needs multiuser capability (multiple\nprivilege groups) \u2014 because without that, any network\ntransaction that can trick a user into running malicious code will\nsubvert the entire system (Windows macro viruses are only the tip of\nthis iceberg).  Without strong multitasking, the ability of an\noperating system to handle network traffic and run user programs at\nthe same time will be impaired.  The operating system also needs\nefficient IPC so that its network programs can communicate with each\nother and with the user's foreground applications.\n<\/p>\n<p>\n  Windows gets away with having severe deficiencies in these areas\nonly by virtue of having developed a monopoly position before\nnetworking became really important, and by having a user population\nthat has been conditioned to accept a shocking frequency of crashes\nand security breaches as normal.  This is not a stable situation, and\nit is one that partisans of Linux have successfully (in 2003)\nexploited to make major inroads in the server-operating-system\nmarket.\n<\/p>\n<p>\n  Around 1980, during the early heyday of personal computers,\noperating-system designers dismissed Unix and traditional timesharing\nas heavyweight, cumbersome, and inappropriate for the brave new world\nof single-user personal machines \u2014 despite the fact that GUI\ninterfaces tended to demand the reinvention of multitasking to cope\nwith threads of execution bound to different windows and widgets.  The\ntrend toward client operating systems was so intense that server\noperating systems were at times dismissed as steam-powered relics of a\nbygone age.\n<\/p>\n<p>\n  But as the designers of BeOS noticed, the requirements of\npervasive networking cannot be met without implementing something very\nclose to general-purpose timesharing.  Single-user client operating\nsystems cannot thrive in an Internetted world.\n<\/p>\n<p>\n  This problem drove the reconvergence of client and server\noperating systems.  The first, pre-Internet attempts at peer-to-peer\nnetworking over LANs, in the late 1980s, began to expose the\ninadequacy of the client-OS design model.  Data on a network has to\nhave rendezvous points in order to be shared; thus, we can't do\nwithout servers.  At the same time, experience with the Macintosh and\nWindows client operating systems raised the bar on the minimum quality\nof user experience customers would tolerate.\n<\/p>\n<p>\n  With non-Unix models for timesharing effectively dead by 1990,\nthere were not many possible responses client operating-system\ndesigners could mount to this challenge. They could co-opt Unix (as\nMacOS X has done), re-invent roughly equivalent features a patch at a\ntime (as Windows has done), or attempt to reinvent the entire world\n(as BeOS tried and failed to do).  But meanwhile, open-source Unixes\nwere growing client-like capabilities to use GUIs and run on\ninexpensive personal machines.\n<\/p>\n<p>\n  These pressures turned out, however, not to be as symmetrically\nbalanced as the above description might imply.  Retrofitting\nserver-operating-system features like multiple privilege classes and\nfull multitasking onto a client operating system is very difficult,\nquite likely to break compatibility with older versions of the client,\nand generally produces a fragile and unsatisfactory result rife with\nstability and security problems.  Retrofitting a GUI onto a server\noperating system, on the other hand, raises problems that can largely\nbe finessed by a combination of cleverness and throwing\never-more-inexpensive hardware resources at the problem.  As with\nbuildings, it's easier to repair superstructure on top of a solid \nfoundation than it is to replace the foundations without \ntrashing the superstructure.\n<\/p>\n<p>\n  Besides having the native architectural strengths of a server\noperating system,  Unix was always agnostic about its intended\naudience.  Its designers and implementers never assumed they knew all\npotential uses the system would be put to.\n<\/p>\n<p>\n  Thus, the Unix design proved more capable of reinventing itself as\na client than any of its client-operating-system competitors were of\nreinventing themselves as servers.  While many other factors of\ntechnology and economics contributed to the Unix resurgence of\nthe 1990s, this is one that really foregrounds itself in any\ndiscussion of operating-system design style.\n<\/p>\n\n\n<h3 id=\"modularitychapter\"><a href=\"#modularitychapter\">\u00a7<\/a>Chapter 4. Modularity<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   There are two ways of constructing a software design.  One is to make\nit so simple that there are obviously no deficiencies; the other is to\nmake it so complicated that there are no obvious deficiencies. The\nfirst method is far more difficult.\n  <\/p><div>\n  \n   \u2013 C. A. R. Hoare\nThe Emperor's Old Clothes, CACM February 1981\n  \n <\/div><\/blockquote>\n<p>\n  There is a natural hierarchy of code-partitioning methods that\nhas evolved as programmers have had to manage ever-increasing levels\nof complexity.  In the beginning, everything was one big lump of\nmachine code.  The earliest procedural languages brought in the notion\nof partition by subroutine.  Then we invented service libraries to\nshare common utility functions among multiple programs. Next, we\ninvented separated address spaces and communicating processes.  Today we\nroutinely distribute program systems across multiple hosts separated\nby thousands of miles of network cable.\n<\/p>\n<p>\n  The early developers of Unix were among the pioneers in software\nmodularity.  Before them, the Rule of Modularity was computer-science\ntheory but not engineering practice.  In\n  <em>Design Rules<\/em><sup><a href=\"#Baldwin-Clark\" title=\"[Baldwin-Clark]\">[Baldwin-Clark]<\/a><\/sup>, a path-breaking\nstudy of the economics of modularity in engineering design, the\nauthors use the development of the computer industry as a case study \nand argue that the Unix community was in fact the first to\nsystematically apply modular decomposition to production software, as\nopposed to hardware.  Modularity of hardware has of course been one\nof the foundations of engineering since the adoption of standard \nscrew threads in the late 1800s.\n<\/p>\n<p>\n  The Rule of Modularity bears amplification here: The only way to\nwrite complex software that won't fall on its face is to build it out\nof simple modules connected by well-defined interfaces, so that most\nproblems are local and you can have some hope of fixing or optimizing\na part without breaking the whole.\n<\/p>\n<p>\n  The tradition of being careful about modularity and of paying\nclose attention to issues like\northogonality\n  and compactness are still much deeper in the bone among Unix\nprogrammers than elsewhere.\n<\/p>\n <blockquote>\n  <p>\n   Early Unix programmers became good at modularity because they had to\nbe.  An OS is one of the most complicated pieces of code around.  If\nit is not well structured, it will fall apart.  There were a couple of\nearly failures at building Unix that were scrapped.  One can blame the\nearly (structureless) C for this, but basically it was because the OS\nwas too complicated to write.  We needed both refinements in tools (like C \nstructures) and good practice in using them (like Rob Pike's rules for \nprogramming) before we could tame that complexity.\n  <\/p><div>\n  \n   \u2013 Ken Thompson\n  \n <\/div><\/blockquote>\n<p>\n  Early Unix hackers struggled with this in many ways.  In the\nlanguages of 1970 function calls were expensive, either because call\nsemantics were complicated (PL\/1.  Algol) or because the compiler was\noptimizing for other things like fast inner loops at the expense of\ncall time.  Thus, code tended to be written in big lumps.  Ken and\nseveral of the other early Unix developers knew modularity was a good\nidea, but they remembered PL\/1 and were reluctant to write small\nfunctions lest performance go to hell.\n<\/p>\n <blockquote>\n  <p>\n   Dennis Ritchie encouraged modularity by telling all and sundry that\nfunction calls were really, really cheap in C.  Everybody started\nwriting small functions and modularizing.  Years later we found out\nthat function calls were still expensive on the PDP-11, and VAX code \nwas often spending 50% of its time in the CALLS instruction.  Dennis\nhad lied to us!  But it was too late; we were all hooked...\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  All programmers today, Unix natives or not, are taught to\nmodularize at the subroutine level within programs.  Some learn the\nart of doing this at the module or abstract-data-type level and call\nthat \u2018good design\u2019.  The\ndesign-patterns\n  movement is making a noble effort to\npush up a level from there and discover successful design abstractions\nthat can be applied to organize the large-scale structure of\nprograms.\n<\/p>\n<p>\n  Getting better at all these kinds of problem partitioning is a\nworthy goal, and many excellent treatments of them are available\nelsewhere.  We shall not attempt to cover all the issues relating to\nmodularity within programs in too much detail: first, because that is\na subject for an entire volume (or several volumes) in itself; and\nsecond, because this is a book about the art of\n  <em>Unix<\/em>\n  programming.\n<\/p>\n<p>\n  What we will do here is examine more specifically what the Unix\ntradition teaches us about how to follow the Rule of Modularity.  In\nthis chapter, our examples will live within process units.  Later, in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>, we'll\nexamine the circumstances under which partitioning programs into\nmultiple cooperating processes is a good idea, and discuss more\nspecific techniques for doing that partitioning.\n<\/p>\n\n<h3 id=\"ch04s01\"><a href=\"#ch04s01\">\u00a7<\/a>Encapsulation and Optimal Module Size<\/h3>\n<p>\n  The first and most important quality of modular code is\n  <em>encapsulation<\/em>.  Well-encapsulated modules don't\nexpose their internals to each other.  They don't call into the middle\nof each others' implementations, and they don't promiscuously share\nglobal data.  They communicate using application programming\ninterfaces (APIs) \u2014 narrow, well-defined sets of procedure calls\nand data structures.  This is what the Rule of Modularity is\nabout.\n<\/p>\n<p>\n  The APIs between modules have a dual role. On the implementation\nlevel, they function as choke points between the modules,\npreventing the internals of each from leaking into its neighbors.  On\nthe design level, it is the APIs (not the bits of implementation\nbetween them) that really define your architecture.\n<\/p>\n<p>\n  One good test for whether an API is well designed is this one:\nif you try to write a description of it in purely human language\n(with no source-code extracts allowed), does it make sense?  It is a\nvery good idea to get into the habit of writing informal descriptions\nof your APIs before you code them.  Indeed, some of the most able\ndevelopers start by defining their interfaces, writing brief comments\nto describe them, and then writing the code \u2014 since the process\nof writing the comment clarifies what the code must do.  Such\ndescriptions help you organize your thoughts, they make useful module\ncomments, and eventually you might want to turn them into a roadmap\ndocument for future readers of the code.\n<\/p>\n<p>\n  As you push module decomposition harder, the pieces get smaller and the\ndefinition of the APIs gets more important. Global complexity,\nand consequent vulnerability to bugs, decreases.  It has been received\nwisdom in computer science since the 1970s (exemplified in papers such\nas <sup><a href=\"#Parnas\" title=\"[Parnas]\">[Parnas]<\/a><\/sup>) that you ought to design\nyour software systems as hierarchies of nested modules, with the grain\nsize of the modules at each level held to a minimum.\n<\/p>\n<p>\n  It is possible, however, to push this kind of decomposition too hard and make your modules too small.  There is\n  evidence <sup><a href=\"#Hatton97\" title=\"[Hatton97]\">[Hatton97]<\/a><\/sup> that when one plots defect density versus\n  module size, the curve is U-shaped and concave upwards (see Figure 4.1).  Very small and very large modules are\n  associated with more bugs than those of intermediate size. A different way of viewing the same data is to plot lines\n  of code per module versus total bugs.  The curve looks roughly logarithmic up to a \u2018sweet spot\u2019 where it flattens\n  (corresponding to the minimum in the defect density curve), after which it goes up as the square of the number of the\n  lines of code (which is what one might intuitively expect for the whole curve, following Brooks's\nLaw<sup><a href=\"#ftn.id2894437\" id=\"id2894437\">[41]<\/a><\/sup>).\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 4.1. Qualitative plot of defect count and\ndensity vs. module size.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Qualitative plot of defect count and density vs. module size.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  This unexpectedly increasing incidence of bugs at small module\nsizes holds across a wide variety of systems implemented in different\nlanguages.  Hatton has proposed a model relating this nonlinearity to\nthe chunk size of human short-term memory.<sup><a href=\"#ftn.id2894501\" id=\"id2894501\">[42]<\/a><\/sup>\n  Another way to interpret\nthe nonlinearity is that at small module grain sizes, the increasing\ncomplexity of the interfaces becomes the dominating term; it's\ndifficult to read the code because you have to understand everything\nbefore you can understand anything.  In\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">Chapter 7<\/a>\n  we'll examine more advanced forms of\nprogram partitioning; there, too, the complexity of interface\nprotocols comes to dominate the total complexity of the system as the\ncomponent processes get smaller.\n<\/p>\n<p>\n  In nonmathematical terms, Hatton's empirical results imply a sweet spot\nbetween 200 and 400 logical lines of code that minimizes probable\ndefect density, all other factors (such as programmer skill) being\nequal.  This size is independent of the language being used \u2014 an\nobservation which strongly reinforces the advice given elsewhere in\nthis book to program with the most powerful languages and tools you\ncan.  Beware of taking these numbers too literally however. Methods\nfor counting lines of code vary considerably according to what the\nanalyst considers a logical line, and other biases (such as whether\ncomments are stripped).  Hatton himself suggests as a rule of thumb a\n2x conversion between logical and physical lines, suggesting an\noptimal range of 400\u2013800 physical lines.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2894437\" id=\"ftn.id2894437\">[41]<\/a><\/sup>\n  Brooks's Law predicts that adding programmers to a\nlate project makes it later. More generally, it predicts that costs\nand error rates rise as the square of the number of programmers on a\nproject.\n<\/p><p><sup><a href=\"#id2894501\" id=\"ftn.id2894501\">[42]<\/a><\/sup>\n  In Hatton's\nmodel, small differences in the maximum chunk size a programmer can\nhold in short-term memory have a large multiplicative effect on the\nprogrammer's efficiency.  This might be a major contributor to the\norder-of-magnitude (or larger) variations in effectiveness observed by\nFred Brooks and others.\n<\/p><\/div>\n\n<h3 id=\"ch04s02\"><a href=\"#ch04s02\">\u00a7<\/a>Compactness and Orthogonality<\/h3>\n<p>\n  Code is not the only sort of thing with an optimal chunk size.\nLanguages and APIs (such as sets of library or system calls) run up\nagainst the same sorts of human cognitive constraints that produce\nHatton's U-curve.\n<\/p>\n<p>\n  Accordingly, Unix programmers have learned to think very hard\nabout two other properties when designing APIs, command sets,\nprotocols, and other ways to make computers do tricks:\n  <em>compactness<\/em> and\n  <em>orthogonality<\/em>.\n<\/p>\n<h4 id=\"compactness\"><a href=\"#compactness\">\u00a7<\/a>Compactness<\/h4>\n<p>\n  Compactness\n  is the property that a design can fit inside a human being's head.  A\ngood practical test for compactness is this: Does an experienced user\nnormally need a manual?  If not, then the design (or at least the\nsubset of it that covers normal use) is compact.\n<\/p>\n<p>\n  Compact software tools have all the virtues of physical tools\nthat fit well in the hand.  They feel pleasant to use, they don't obtrude\nthemselves between your mind and your work, they make you more\nproductive \u2014 and they are much less likely than unwieldy tools\nto turn in your hand and injure you.\n<\/p>\n<p>\n  Compact is not equivalent to \u2018weak\u2019.  A design can\nhave a great deal of power and flexibility and still be compact if it\nis built on abstractions that are easy to think about and fit together\nwell.  Nor is compact equivalent to \u2018easily learned\u2019; some\ncompact designs are quite difficult to understand until you have\nmastered an underlying conceptual model that is tricky, at which point\nyour view of the world changes and compact\n  <em>becomes<\/em>\n  simple.  For a lot of people, the Lisp\nlanguage is a classic example of this.\n<\/p>\n <blockquote>\n  <p>\n   Nor does compact mean \u2018small\u2019.  If a well-designed \nsystem is predictable and \u2018obvious\u2019 to the experienced\nuser, it might have quite a few pieces.\n  <\/p><div>\n  \n   \u2013 Ken Arnold\n  \n <\/div><\/blockquote>\n<p>\n  Very few software designs are compact in an absolute sense, but\nmany are compact in a slightly looser sense of the term.  They have a\ncompact working set, a subset of capabilities that suffices for 80% or\nmore of what expert users normally do with them.  Practically\nspeaking, such designs normally need a reference card or cheat sheet\nbut not a manual.  We'll call such designs\n  <em>semi-compact<\/em>, as opposed to\n  <em>strictly compact<\/em>.\n<\/p>\n<p>\n  The concept is perhaps best illustrated by examples.  The Unix\nsystem call API is semi-compact, but the standard\nC\n  library is not\ncompact in any sense.  While Unix programmers easily keep a subset of\nthe system calls sufficient for most applications programming (file\nsystem operations, signals, and process control) in their heads, the C\nlibrary on modern Unixes includes many hundreds of entry points,\ne.g., mathematical functions, that won't all fit inside a single\nprogrammer's cranium.\n<\/p>\n<p>\n  <em>The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information<\/em><sup><a href=\"#Miller\" title=\"[Miller]\">[Miller]<\/a><\/sup> is one of the foundation papers in cognitive\npsychology (and, incidentally, the specific reason that U.S.\nlocal telephone numbers have seven digits).  It showed that the number\nof discrete items of information human beings can hold in\nshort-term memory is seven, plus or minus two. This gives us a good\nrule of thumb for evaluating the compactness of APIs: Does a\nprogrammer have to remember more than seven entry points?  Anything\nlarger than this is unlikely to be strictly compact.\n<\/p>\n<p>\n  Among Unix tools, make(1) is compact; autoconf(1) and automake(1) are not.  Among markup languages, HTML is\n  semi-compact, but DocBook (a documentation markup language we shall discuss in\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">Chapter 18<\/a>) is not.  The\nman(7) macros are compact, but troff(1) markup is not.\n<\/p>\n<p>\n  Among general-purpose programming languages, C and Python are semi-compact; Perl, Java, Emacs Lisp, and shell are not\n  (especially since serious shell programming requires you to know half-a-dozen other tools like sed(1) and awk(1)). C++\n  is anti-compact \u2014 the language's designer has admitted that he doesn't expect any one programmer to ever understand it\n  all.\n<\/p>\n<p>\n  Some designs that are not compact have enough internal\nredundancy of features that individual programmers end up carving out\ncompact dialects sufficient for that 80% of common tasks by choosing a\nworking subset of the language.  Perl has this kind of\npseudo-compactness, for example.  Such designs have a built-in trap;\nwhen two programmers try to communicate about a project, they may find\nthat differences in their working subsets are a significant barrier to\nunderstanding and modifying the code.\n<\/p>\n<p>\n  Noncompact designs are not automatically doomed or bad,\nhowever.  Some problem domains are simply too complex for a compact\ndesign to span them.  Sometimes it's necessary to trade away\ncompactness for some other virtue, like raw power and range.  Troff\nmarkup is a good example of this.  So is the\nBSD\n  sockets\nAPI. The purpose of emphasizing compactness as a virtue is not to\ncondition you to treat compactness as an absolute requirement, but to\nteach you to do what Unix programmers do: value compactness\nproperly, design for it whenever possible, and not throw it away\ncasually.\n<\/p>\n<h4 id=\"orthogonality\"><a href=\"#orthogonality\">\u00a7<\/a>Orthogonality<\/h4>\n<p>\n  Orthogonality\n  is one of the most important properties that can help make even\ncomplex designs\ncompact.  In a purely\northogonal design, operations do not have side effects; each action\n(whether it's an API call, a macro invocation, or a language\noperation) changes just one thing without affecting others.  There is\none and only one way to change each property of whatever system you\nare controlling.\n<\/p>\n<p>\n  Your monitor has orthogonal controls.  You can change the\nbrightness independently of the contrast level, and (if the monitor\nhas one) the color balance control will be independent of both.\nImagine how much more difficult it would be to adjust a monitor on\nwhich the brightness knob affected the color balance: you'd\nhave to compensate by tweaking the color balance every time after you\nchanged the brightness.  Worse, imagine if the contrast control also\naffected the color balance; then, you'd have to adjust both knobs\nsimultaneously in exactly the right way to change either contrast or\ncolor balance alone while holding the other constant.\n<\/p>\n<p>\n  Far too many software designs are non-orthogonal.  One common\nclass of design mistake, for example, occurs in code that reads and\nparses data from one (source) format to another (target) format.  A\ndesigner who thinks of the source format as always being stored in a\ndisk file may write the conversion function to open and read from a\nnamed file.  Usually the input could just as well have been any file\nhandle.  If the conversion routine were designed orthogonally,\ne.g., without the side effect of opening a file, it could save work\nlater when the conversion has to be done on a data stream supplied\nfrom standard input, a network socket, or any other source.\n<\/p>\n<p>\n  Doug McIlroy's\n  advice to \u201cDo one thing well\u201d\nis usually interpreted as being about simplicity.  But it's also,\nimplicitly and at least as importantly, about orthogonality.\n<\/p>\n<p>\n  It's not a problem for a program to do one thing well and\nother things as side effects, provided supporting those other things\ndoesn't raise the complexity of the program and its vulnerability\nto bugs.  In\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>\n  we'll\nexamine a program called\n  <em>ascii<\/em> that prints\nsynonyms for the names of ASCII characters, including hex, octal, and\nbinary values; as a side effect, it can serve as a quick base\nconverter for numbers in the range 0\u2013255.  This second use is not an\northogonality violation because the features that support it are all\nnecessary to the primary function; they do not make the program\nmore difficult to document or maintain.\n<\/p>\n<p>\n  The problems with non-orthogonality arise when side effects\ncomplicate a programmer's or user's mental model, and beg to be\nforgotten, with results ranging from inconvenient to dire.  Even when\nyou do not forget the side effects, you're often forced to do extra\nwork to suppress them or work around them.\n<\/p>\n<p>\n  There is an excellent discussion of orthogonality and how to\nachieve it in\n  <em>The Pragmatic Programmer<\/em> <sup><a href=\"#Hunt-Thomas\" title=\"[Hunt-Thomas]\">[Hunt-Thomas]<\/a><\/sup>.  As they point out, orthogonality reduces\ntest and development time, because it's easier to verify code that\nneither causes side effects nor depends on side effects from other\ncode \u2014 there are fewer combinations to test. If it breaks,\northogonal code is more easily replaced without disturbance to the\nrest of the system.  Finally, orthogonal code is easier to document\nand reuse.\n<\/p>\n<p>\n  The concept of\n  <em>refactoring<\/em>, which first\nemerged as an explicit idea from the \u2018Extreme Programming\u2019\nschool, is closely related to orthogonality.  To refactor code is to\nchange its structure and organization without changing its observable\nbehavior.  Software engineers have been doing this since the birth of\nthe field, of course, but naming the practice and identifying a stock\nset of refactoring techniques has helped concentrate peoples'\nthinking in useful ways.  Because these fit so well with the central\nconcerns of the Unix design tradition, Unix developers have quickly\ncoopted the terminology and ideas of refactoring.<sup><a href=\"#ftn.id2895110\" id=\"id2895110\">[43]<\/a><\/sup><\/p>\n<p>\n  The basic Unix APIs were designed for orthogonality with\nimperfect but considerable success.  We take for granted being able to\nopen a file for write access without exclusive-locking it for write,\nfor example; not all operating systems are so graceful.  Old-style\n(System III)\nsignals were non-orthogonal, because signal receipt had the\nside-effect of resetting the signal handler to the default\ndie-on-receipt.  There are large non-orthogonal patches like the BSD\nsockets\n  API and\n  <em>very<\/em>\n  large ones like the X windowing system's\ndrawing libraries.\n<\/p>\n<p>\n  But on the whole the Unix API is a good example:\nOtherwise it not only would not but\n  <em>could<\/em>\n  not\nbe so widely imitated by C\n  libraries on other operating systems.  This is also a reason that the\nUnix API repays study even if you are not a Unix programmer; it has\nlessons about orthogonality to teach.\n<\/p>\n<h4 id=\"spot_rule\"><a href=\"#spot_rule\">\u00a7<\/a>The SPOT Rule<\/h4>\n<p>\n  <em>The Pragmatic Programmer<\/em> articulates a\nrule for one particular kind of orthogonality that is especially\nimportant.  Their \u201cDon't Repeat Yourself\u201d rule is: every\npiece of knowledge must have a\n  <em>single<\/em>,\nunambiguous, authoritative representation within a system.  In this\nbook we prefer, following a suggestion by Brian Kernighan, to call\nthis the Single Point Of Truth or SPOT rule.\n<\/p>\n<p>\n  Repetition leads to inconsistency and code that is subtly\nbroken, because you changed only some repetitions when you needed to\nchange\n  <em>all<\/em>\n  of them.  Often, it also means\nthat you haven't properly thought through the organization \nof your code.\n<\/p>\n<p>\n  Constants, tables, and metadata should be declared and\ninitialized\n  <em>once<\/em>\n  and imported elsewhere.  Any\ntime you see duplicate code, that's a danger sign.  Complexity\nis a cost; don't pay it twice.\n<\/p>\n<p>\n  Often it's possible to remove code duplication by\n  <em>refactoring<\/em>; that is, changing the organization\nof your code without changing the core algorithms.  Data duplication\nsometimes appears to be forced on you.  But when you see it, here are\nsome valuable questions to ask:\n<\/p>\n <div><ul><li>\n    <p>\n     If you have duplicated data in your code because it has to\nhave two different representations in two different places, can you\nwrite a function, tool or code generator to make one representation from the\nother, or both from a common source?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     If your documentation duplicates knowledge in your code, can you\ngenerate parts of the documentation from parts of the code, or\nvice-versa, or both from a common higher-level representation?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     If your header files and interface declarations duplicate\nknowledge in your implementation code, is there a way you can generate\nthe header files and interface declarations from the code?\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  There is an analog of the SPOT rule for data structures:\n\u201cNo junk, no confusion\u201d.  \u201cNo junk\u201d says\nthat the data structure (the model) should be minimal, e.g., not made\nso general that it can represent situations which cannot exist.\n\u201cNo confusion\u201d says that states which must be kept\ndistinct in the real-world problem must be kept distinct in the model.\nIn short, the SPOT rule advocates seeking a data structure whose\nstates have a one-to-one correspondence with the states of the\nreal-world system to be modeled.\n<\/p>\n<p>\n  From deeper within the Unix tradition, we can add some of our\nown corollaries of the SPOT rule:\n<\/p>\n <div><ul><li>\n    <p>\n\t Are you duplicating data because you're caching intermediate results of some computation or lookup?  Consider\n\t carefully whether this is premature optimization; stale caches (and the layers of code needed to keep caches\n\t synchronized) are a fertile source of bugs,<sup><a href=\"#ftn.id2895368\" id=\"id2895368\">[44]<\/a><\/sup> and can even\n\t slow down overall performance if (as often happens) the cache-management overhead is higher than you expected.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     If you see lots of duplicative boilerplate code, can you\ngenerate all of it from a single higher-level representation,\ntwiddling a few knobs to generate the different cases?\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  The reader should begin to see a pattern emerging here.\n<\/p>\n<p>\n  In the Unix world, the SPOT Rule as a unifying idea has seldom\nbeen explicit \u2014 but heavy use of code generators to implement\nparticular\n  <em>kinds<\/em>\n  of SPOT are very much part of the\ntradition.  We'll survey these techniques in\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>.\n<\/p>\n<h4 id=\"id2895445\"><a href=\"#id2895445\">\u00a7<\/a>Compactness and the Strong Single Center<\/h4>\n<p>\n  One subtle but powerful way to promote compactness in a\ndesign is to organize it around a strong core algorithm addressing a\nclear formal definition of the problem, avoiding heuristics and\nfudging.\n<\/p>\n <blockquote>\n  <p>\n   Formalization often clarifies a task spectacularly.  It is not\nenough for a programmer to recognize that bits of his task fall within\nstandard computer-science categories \u2014 a little depth-first\nsearch here and a quicksort there.  The best results occur when the\nnub of the task can be formalized, and a clear model of the job at\nhand can be constructed.  It is not necessary that ultimate users\ncomprehend the model.  The very existence of a unifying core will\nprovide a comfortable feel, unencumbered with the\nwhy-in-hell-did-they-do-that moments that are so prevalent in using\nSwiss-army-knife programs.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  This is an often-overlooked strength of the Unix\ntradition.  Many of its most effective tools are thin wrappers around\na direct translation of some single powerful algorithm.\n<\/p>\n<p>\n  Perhaps the clearest example of this is\ndiff(1),\nthe Unix tool for reporting differences between related files.  This\ntool and its dual,\npatch(1),\nhave become central to the network-distributed development style of\nmodern Unix.  A valuable property of diff is that it seldom surprises\nanyone.  It doesn't have special cases or painful edge conditions,\nbecause it uses a simple, mathematically sound method of sequence\ncomparison.  This has consequences:\n<\/p>\n <blockquote>\n  <p>\n   By virtue of a mathematical model and a solid\nalgorithm, Unix diff contrasts markedly with its\nimitators.  First, the central engine is solid, small, and\nhas never needed one line of maintenance.  Second, the\nresults are clear and consistent, unmarred by surprises\nwhere heuristics fail.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Thus, people who use diff can develop an intuitive feel for what\nit will do in any given situation without necessarily understanding\nthe central algorithm perfectly.  Other well-known examples of\nthis special kind of clarity achieved through a strong central\nalgorithm abound in Unix:\n<\/p>\n <div><ul><li>\n    <p>\n     The\ngrep(1)\nutility for selecting lines out of files by pattern matching is a\nsimple wrapper around a formal algebra of regular-expression patterns\n(see\n     <a href=\"#regexps\" title=\"Case Study: Regular Expressions\">\n      the section called \u201cCase Study: Regular Expressions\u201d<\/a>\n     for discussion).  If it had\nlacked this consistent mathematical model, it would probably look like\nthe design of the original\nglob(1)\nfacility in the oldest Unixes, a handful of ad-hoc wildcards that\ncan't be combined.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The yacc(1) utility for\ngenerating language parsers is a thin wrapper around the formal theory\nof LR(1) grammars. Its partner, the lexical analyzer generator\nlex(1),\nis a similarly thin wrapper around the theory of nondeterministic\nfinite-state automata.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  All three of these programs are so bug-free that their correct\nfunctioning is taken utterly for granted, and compact enough to fit\neasily in a programmer's hand. Only a part of these good qualities are due\nto the polishing that comes with a long service life and frequent use;\nmost of it is that, having been constructed around a strong and\nprovably correct algorithmic core, they never needed much polishing\nin the first place.\n<\/p>\n<p>\n  The opposite of a formal approach is using\n  <em>heuristics<\/em> \u2014rules of thumb leading\ntoward a solution that is probabilistically, but not certainly,\ncorrect.  Sometimes we use heuristics because a deterministically\ncorrect solution is impossible. Think of spam filtering, for example;\nan algorithmically perfect spam filter would need a full solution to\nthe problem of understanding natural language as a module.  Other\ntimes, we use heuristics because known formally correct methods are\nimpossibly expensive.  Virtual-memory management is an example of\nthis; there are near-perfect solutions, but they require so much\nruntime instrumentation that their overhead would swamp any\ntheoretical gain over heuristics.\n<\/p>\n<p>\n  The trouble with heuristics is that they proliferate special\ncases and edge cases.  If nothing else, you usually have to backstop \na heuristic with some sort of recovery mechanism when it fails.  All\nthe usual problems with escalating complexity follow.  To manage the\nresulting tradeoffs, you have to start by being aware of them.  Always\nask if a heuristic actually pays off in performance what it costs \nin code complexity \u2014 and don't guess at the performance\ndifference, actually measure it before making a decision.\n<\/p>\n<h4 id=\"id2899405\"><a href=\"#id2899405\">\u00a7<\/a>The Value of Detachment<\/h4>\n<p>\n  We began this book with a reference to Zen: \u201ca special\ntransmission, outside the scriptures\u201d.  This was not mere\nexoticism for stylistic effect; the core concepts of Unix have always\nhad a spare, Zen-like simplicity that continues to shine through the\nlayers of historical accidents that have accreted around\nthem.  This\nquality is reflected in the cornerstone documents of Unix, like\n  <em>The C Programming Language<\/em> <sup><a href=\"#Kernighan-Ritchie\" title=\"[Kernighan-Ritchie]\">[Kernighan-Ritchie]<\/a><\/sup> and the 1974 CACM paper that introduced\nUnix to the world; one of the famous quotes from that paper observes\n\u201c...constraint has encouraged not only economy, but also a\ncertain elegance of design\u201d.  That simplicity came from trying\nto think not about how much a language or operating system could do,\nbut of how\n  <em>little<\/em>\n  it could do \u2014 not by\ncarrying assumptions but by starting from zero (what in Zen is called\n\u201cbeginner's mind\u201d or \u201cempty mind\u201d).\n<\/p>\n<p>\n  To design for\ncompactness\n  and\northogonality,\nstart from zero.  Zen teaches that attachment leads to suffering;\nexperience with software design teaches that attachment to unnoticed\nassumptions leads to non-orthogonality, noncompact designs, and\nprojects that fail or become maintenance nightmares.\n<\/p>\n<p>\n  To achieve enlightenment and surcease from suffering, Zen\nteaches detachment.  The Unix tradition teaches the value of\ndetachment from the particular, accidental conditions under which a\ndesign problem was posed.  Abstract.  Simplify.  Generalize.  Because\nwe write software to solve problems, we cannot completely detach from\nthe problems \u2014 but it is well worth the mental effort to see how\nmany preconceptions you can throw away, and whether the design becomes\nmore compact\n  and\northogonal as you do that.  Possibilities for code reuse often\nresult.\n<\/p>\n<p>\n  Jokes about the relationship between Unix and Zen are a live\npart of the Unix tradition as well.<sup><a href=\"#ftn.id2899522\" id=\"id2899522\">[45]<\/a><\/sup>\n  This is not an accident.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2895110\" id=\"ftn.id2895110\">[43]<\/a><\/sup>\n  In\nthe foundation text on this topic,\n  <em>Refactoring<\/em> <sup><a href=\"#Fowler\" title=\"[Fowler]\">[Fowler]<\/a><\/sup>, the author comes very close to stating that\nthe principal goal of refactoring is to improve orthogonality. But\nlacking the concept, he can only approximate this idea from several\ndifferent directions: eliminating code duplication and various other\n\u201cbad smells\u201d many of which are some sort of orthogonality\nviolation.\n<\/p><p><sup><a href=\"#id2895368\" id=\"ftn.id2895368\">[44]<\/a><\/sup>\n  An archetypal example of bad caching is the <code>rehash<\/code> directive in csh(1); type <code>man 1 csh<\/code> for\n  details.  See\n  <a href=\"#binary_caches\" title=\"Caching Operation Results\">\n   the section called \u201cCaching Operation Results\u201d<\/a> for another example.\n<\/p><p><sup><a href=\"#id2899522\" id=\"ftn.id2899522\">[45]<\/a><\/sup>\n  For a recent example\nof Unix\/Zen crossover, see\n  <a href=\"https:\/\/arp242.net\/the-art-of-unix-programming\/unix_koans.html\" title=\"Appendix D. Rootless Root\">\n   Appendix D<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch04s03\"><a href=\"#ch04s03\">\u00a7<\/a>Software Is a Many-Layered Thing<\/h3>\n<p>\n  Broadly speaking, there are two directions one can go in\ndesigning a hierarchy of functions or objects.  Which direction you\nchoose, and when, has a profound effect on the layering of your code.\n<\/p>\n<h4 id=\"id2899552\"><a href=\"#id2899552\">\u00a7<\/a>Top-Down versus Bottom-Up<\/h4>\n<p>\n  One direction is bottom-up, from concrete to abstract \u2014\nworking up from the specific operations in the problem domain that you\nknow you will need to perform.  For example, if one is designing\nfirmware for a disk drive, some of the bottom-level primitives might\nbe \u2018seek head to physical block\u2019, \u2018read physical\nblock\u2019, \u2018write physical block\u2019, \u2018toggle drive\nLED\u2019, etc.\n<\/p>\n<p>\n  The other direction is top-down, abstract to concrete \u2014 from the\nhighest-level specification describing the project as a whole, or the\napplication logic, downwards to individual operations.  Thus, if one is\ndesigning software for a mass-storage controller that might drive\nseveral different sorts of media, one might start with abstract\noperations like \u2018seek logical block\u2019, \u2018read logical\nblock\u2019, \u2018write logical block\u2019, \u2018toggle\nactivity indication\u2019.  These would differ from the similarly named\nhardware-level operations above in that they're intended to be \ngeneric across different kinds of physical devices.\n<\/p>\n<p>\n  These two examples could be two ways of approaching design for\nthe same collection of hardware.  Your choice, in cases like this, is\none of these: either abstract the hardware (so the objects encapsulate the real\nthings out there and the program is merely a list of manipulations on\nthose things), or organize around some behavioral model (and\nthen embed the actual hardware manipulations that carry it out in the\nflow of the behavioral logic).\n<\/p>\n<p>\n  An analogous choice shows up in a lot of different contexts.\nSuppose you're writing MIDI sequencer software.  You could organize\nthat code around its top level (sequencing tracks) or around its\nbottom level (switching patches or samples and driving wave\ngenerators).\n<\/p>\n<p>\n  A very concrete way to think about this difference is to ask\nwhether the design is organized around its main event loop (which\ntends to have the high-level application logic close to it) or around\na service library of all the operations that the main loop can invoke.  A\ndesigner working from the top down will start by thinking about the\nprogram's main event loop, and plug in specific events later.  A\ndesigner working from the bottom up will start by thinking about\nencapsulating specific tasks and glue them together into some kind of\ncoherent order later on.\n<\/p>\n<p>\n  For a larger example, consider the design of a Web browser.  The\ntop-level design of a Web browser is a specification of the expected\nbehavior of the browser: what types of URL (like\n  <code>http:<\/code> or\n  <code>ftp:<\/code> or\n  <code>file:<\/code>) it interprets, what kinds of images it is expected to be\nable to render, whether and with what limitations it will accept\nJava\n  or\nJavaScript, etc.  The layer of the implementation\nthat corresponds to this top-level view is its main event loop; each\ntime around, the loop waits for, collects, and dispatches on a user\naction (such as clicking a Web link or typing a character into a\nfield).\n<\/p>\n<p>\n  But the Web browser has to call a large set of domain\nprimitives to do its job.  One group of these is concerned with\nestablishing network connections, sending data over them, and\nreceiving responses.  Another set is the operations of the GUI \ntoolkit the browser will use.  Yet a third set might be concerned\nwith the mechanics of parsing retrieved HTML from text into a \ndocument object tree.\n<\/p>\n<p>\n  Which end of the stack you start with matters a lot, because the\nlayer at the other end is quite likely to be constrained by your\ninitial choices.  In particular, if you program purely from the top\ndown, you may find yourself in the uncomfortable position that the\ndomain primitives your application logic wants don't match the ones\nyou can actually implement.  On the other hand, if you program purely\nfrom the bottom up, you may find yourself doing a lot of work that is\nirrelevant to the application logic \u2014 or merely designing a pile\nof bricks when you were trying to build a house.\n<\/p>\n<p>\n  Ever since the structured-programming controversies of the\n1960s, novice programmers have generally been taught that the correct\napproach is the top-down one: stepwise refinement, where you specify\nwhat your program is to do at an abstract level and gradually fill in\nthe blanks of implementation until you have concrete working code.\nTop-down tends to be good practice when three preconditions are true:\n(a) you can specify in advance precisely what the program is to do,\n(b) the specification is unlikely to change significantly during\nimplementation, and (c) you have a lot of freedom in choosing, at\na low level, how the program is to get that job done.\n<\/p>\n<p>\n  These conditions tend to be fulfilled most often in programs\nrelatively close to the user and high in the software stack \u2014\napplications programming.  But even there those preconditions often\nfail.  You can't count on knowing what the \u2018right\u2019 way for\na word processor or a drawing program to behave is until the user\ninterface has had end-user testing.  Purely top-down programming often\nhas the effect of overinvesting effort in code that has to be scrapped\nand rebuilt because the interface doesn't pass a reality check.\n<\/p>\n<p>\n  In self-defense against this, programmers try to do both things\n\u2014 express the abstract specification as top-down application\nlogic,\n  <em>and<\/em>\n  capture a lot of low-level domain\nprimitives in functions or libraries, so they can be reused when the\nhigh-level design changes.\n<\/p>\n<p>\n  Unix programmers inherit a tradition that is centered in systems\nprogramming, where the low-level primitives are hardware-level\noperations that are fixed in character and extremely important.  They\ntherefore lean, by learned instinct, more toward bottom-up\nprogramming.\n<\/p>\n<p>\n  Whether you're a systems programmer or not, bottom-up can also\nlook more attractive when you are programming in an exploratory way,\ntrying to get a grasp on hardware or software or real-world phenomena\nyou don't yet completely understand.  Bottom-up programming gives you\ntime and room to refine a vague specification.  Bottom-up also appeals to\nprogrammers' natural human laziness \u2014 when you have to scrap and\nrebuild code, you tend to have to throw away larger pieces if you're\nworking top-down than you do if you're working bottom-up.\n<\/p>\n<p>\n  Real code, therefore tends to be programmed both top-down and\nbottom-up.  Often, top-down and bottom-up code will be part of the\nsame project.  That's where \u2018glue\u2019 enters the\npicture.\n<\/p>\n<h4 id=\"id2899777\"><a href=\"#id2899777\">\u00a7<\/a>Glue Layers<\/h4>\n<p>\n  When the top-down and bottom-up drives collide, the result\nis often a mess.  The top layer of application logic and the bottom\nlayer of domain primitives have to be impedance-matched by a layer of\nglue logic.\n<\/p>\n<p>\n  One of the lessons Unix programmers have learned over decades is\nthat glue is nasty stuff and that it is vitally important to keep glue\nlayers as thin as possible.  Glue should stick things together, but\nshould not be used to hide cracks and unevenness in the layers.\n<\/p>\n<p>\n  In the Web-browser example, the glue would include the rendering\ncode that maps a document object parsed from incoming HTML into a\nflattened visual representation as a bitmap in a display buffer, using\nGUI domain primitives to do the painting.  This rendering code is\nnotoriously the most bug-prone part of a browser.  It attracts into\nitself kluges to address problems that originate both in the HTML\nparsing (because there is a lot of ill-formed markup out there) and\nthe GUI toolkit (which may not have quite the primitives that are\nreally needed).\n<\/p>\n<p>\n  A Web browser's glue layer has to mediate not merely between\nspecification and domain primitives, but between several different\nexternal specifications: the network behavior standardized in\nHTTP, HTML document structure, and various graphics and multimedia\nformats as well as the users' behavioral expectations from the\nGUI.\n<\/p>\n<p>\n  And one single bug-prone glue layer is not the worst fate that can\nbefall a design.  A designer who is aware that the glue layer exists,\nand tries to organize it into a middle layer around its own set of\ndata structures or objects, can end up with\n  <em>two<\/em>\n  layers of glue \u2014 one above the midlayer and one below.\nProgrammers who are bright but unseasoned are particularly apt to fall\ninto this trap; they'll get each fundamental set of classes\n(application logic, midlayer, and domain primitives) right and make\nthem look like the textbook examples, only to flounder as the multiple\nlayers of glue needed to integrate all that pretty code get thicker\nand thicker.\n<\/p>\n<p>\n  The thin-glue principle can be viewed as a refinement of the\nRule of Separation.  Policy (the application logic) should be cleanly\nseparated from mechanism (the domain primitives), but if there is a\nlot of code that is neither policy nor mechanism, chances are that it\nis accomplishing very little besides adding global complexity to the\nsystem.\n<\/p>\n<h4 id=\"c_thin_glue\"><a href=\"#c_thin_glue\">\u00a7<\/a>Case Study: C Considered as Thin Glue<\/h4>\n<p>\n  The C language itself is a good example of the effectiveness \nof thin glue.\n<\/p>\n<p>\n  In the late 1990s, Gerrit Blaauw and Fred Brooks observed in\n  <em>Computer Architecture: Concepts and Evolution<\/em> <sup><a href=\"#BlaauwBrooks\" title=\"[BlaauwBrooks]\">[BlaauwBrooks]<\/a><\/sup> that the\narchitectures in every generation of computers, from early mainframes\nthrough minicomputers through workstations through PCs, had tended to\nconverge.  The later a design was in its technology generation, the\nmore closely it approximated what Blaauw &amp; Brooks called the\n\u201cclassical architecture\u201d: binary representation, flat\naddress space, a distinction between memory and working store\n(registers), general-purpose registers, address resolution to\nfixed-length bytes, two-address instructions,\nbig-endianness,<sup><a href=\"#ftn.id2899932\" id=\"id2899932\">[46]<\/a><\/sup>\n  and data types\na consistent set with sizes a multiple of either 4 or 6 bits (the\n6-bit families are now extinct).\n<\/p>\n<p>\n  Thompson\n  and Ritchie\n  designed C to be a sort of structured assembler for an idealized\nprocessor and memory architecture that they expected could be\nefficiently modeled on most conventional computers.  By happy\naccident, their model for the idealized processor was the\nPDP-11, a\nparticularly mature and elegant minicomputer design that closely\napproximated Blaauw &amp; Brooks's classical architecture.  By good\njudgment, Thompson and Ritchie declined to wire into their language\nmost of the few traits (such as little-endian byte order) where the\nPDP-11 didn't match it.<sup><a href=\"#ftn.id2900000\" id=\"id2900000\">[47]<\/a><\/sup><\/p>\n<p>\n  The PDP-11 became an important model for the following\ngenerations of microprocessor architectures.  The basic abstractions\nof C turned out to capture the classical architecture rather neatly.\nThus, C started out as a good fit for microprocessors and, rather than\nbecoming irrelevant as its assumptions fell out of date, actually\nbecame a\n  <em>better<\/em>\n  fit as hardware converged more\nclosely on the classical architecture. One notable example of this\nconvergence was when Intel's 386, with its large flat memory-address\nspace, replaced the 286's awkward segmented-memory addressing after\n1985; pure C was actually a better fit for the 386 than it had been\nfor the 286.\n<\/p>\n<p>\n  It is not a coincidence that the experimental era in computer\narchitectures ended in the mid-1980s at the same time that C (and its\nclose descendant C++) were sweeping all before them as\ngeneral-purpose programming languages.  C, designed as a thin but\nflexible layer over the classical architecture, looks with two\ndecades' additional perspective like almost the best possible design\nfor the structured-assembler niche it was intended to fill.  In\naddition to\ncompactness,\northogonality,\nand detachment (from the machine architecture on which it was\noriginally designed), it also has the important quality of\ntransparency\n  that we will discuss in\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>. The few language designs since that\nare arguably better have needed to make large changes (like\nintroducing garbage collection) in order to get enough functional\ndistance from C not to be swamped by it.\n<\/p>\n<p>\n  This history is worth recalling and understanding because C\nshows us how powerful a clean, minimalist design can be.  If\nThompson\n  and\nRitchie\n  had\nbeen less wise, they would have designed a language that did much\nmore, relied on stronger assumptions, never ported satisfactorily off\nits original hardware platform, and withered away as the world changed\nout from under it.  Instead, C has flourished \u2014 and the example\nThompson and Ritchie set has influenced the style of Unix development\never since.  As the writer, adventurer, artist, and aeronautical\nengineer Antoine de Saint-Exup\u00e9ry once put\nit, writing about the\ndesign of airplanes:\n  \n   <em>\u00abLa perfection est atteinte non quand il ne reste rien \u00e0 ajouter, mais quand il ne reste rien \u00e0 enlever\u00bb.<\/em>\n  \n  (\u201cPerfection is\nattained not when there is nothing more to add, but when there is\nnothing more to remove\u201d.)\n<\/p>\n<p>\n  Ritchie and Thompson lived by this maxim.  Long after the\nresource constraints on early Unix software had eased, they worked at\nkeeping C as thin a layer over the hardware as possible.\n<\/p>\n <blockquote>\n  <p>\n   Dennis used to say to me, when I would ask for some\nparticularly extravagant feature in C, \u201cIf you want PL\/1, you know\nwhere to get it\u201d.  He didn't have to deal with some marketer saying\n\u201cBut we need a check in the box on the sales viewgraph!\u201d\n  <\/p><div>\n  \n   \u2013 Mike Lesk\n  \n <\/div><\/blockquote>\n<p>\n  The history of C is also a lesson in the value of having a\nworking reference implementation\n  <em>before<\/em>\n  you\nstandardize.  We'll return to this point in\n  <a href=\"#portabilitychapter\" title=\"Chapter 17. Portability\">\n   Chapter 17<\/a>\n  when we discuss the evolution of C and\nUnix standards.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2899932\" id=\"ftn.id2899932\">[46]<\/a><\/sup>\n  The terms\n  <em>big-endian<\/em> and\n  <em>little-endian<\/em> refer to architectural choices about\nthe order in which bits are interpreted within a machine word.  Though\nit has no canonical location, a Web search for\n  <em>On Holy Wars and a Plea for Peace<\/em>\n  should turn up a classic and\nentertaining paper on this subject.\n<\/p><p><sup><a href=\"#id2900000\" id=\"ftn.id2900000\">[47]<\/a><\/sup>\n  The widespread belief that the\nautoincrement and autodecrement features entered C because they\nrepresented PDP-11 machine instructions is a myth.  According to\nDennis Ritchie, these operations were present in the ancestral B\nlanguage before the PDP-11 existed.\n<\/p><\/div>\n\n<h3 id=\"ch04s04\"><a href=\"#ch04s04\">\u00a7<\/a>Libraries<\/h3>\n<p>\n  One consequence of the emphasis that the Unix programming style\nput on modularity and well-defined APIs is a strong tendency to\nfactor programs into bits of glue connecting collections of\nlibraries, especially shared libraries (the equivalents of what are\ncalled dynamically-linked libraries or DLLs under Windows and\nother operating systems).\n<\/p>\n<p>\n  If you are careful and clever about design, it is often possible\nto partition a program so that it consists of a\nuser-interface-handling main section (policy) and a collection of\nservice routines (mechanism) with effectively no glue at all.  This\napproach is especially appropriate when the program has to do a lot of\nvery specific manipulations of data structures like graphic images,\nnetwork-protocol packets, or control blocks for a hardware interface.\nSome good general architectural advice from within the Unix tradition,\nparticularly applicable to the resource-management challenges of this\nsort of library is collected in\n  <em>The Discipline and Method Architecture for Reusable Libraries<\/em><sup><a href=\"#Vo\" title=\"[Vo]\">[Vo]<\/a><\/sup>.\n<\/p>\n<p>\n  Under Unix, it is normal practice to make this layering\nexplicit, with the service routines collected in a library that is\nseparately documented.  In such programs, the front end gets to\nspecialize in user-interface considerations and high-level protocol.\nWith a little more care in design, it may be possible to detach the\noriginal front end and replace it with others adapted for different\npurposes.  Some other advantages should become evident from our case\nstudy.\n<\/p>\n<p>\n  There is a flip side to this. In the Unix world, libraries which\nare delivered\n  <em>as libraries<\/em>\n  should come with\nexerciser programs.\n<\/p>\n <blockquote>\n  <p>\n   APIs should come with programs, and vice versa.  An API that you\nmust write C code to use, which cannot be invoked easily from the\ncommand line, is harder to learn and use.  And contrariwise, it's a\nroyal pain to have interfaces whose\n   <em>only<\/em>\n   open,\ndocumented form is a program, so you cannot invoke them easily from a\nC program \u2014 for example,\nroute(1)\nin older Linuxes.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  Besides easing the learning curve, library exercisers often make\nexcellent test frameworks.  Experienced Unix programmers therefore see\nthem not just as a form of thoughtfulness to the library's users but\nas an indication that the code has probably been well tested.\n<\/p>\n<p>\n  An important form of library layering is the\n  <em>plugin<\/em>, a library with a set of known entry\npoints that is dynamically loaded after startup time to perform a\nspecialized task.  For plugins to work, the calling program has to be\norganized largely as a documented service library that the plugin can\ncall back into.\n<\/p>\n<h4 id=\"gimp_plugins\"><a href=\"#gimp_plugins\">\u00a7<\/a>Case Study: GIMP Plugins<\/h4>\n<p>\n  The GIMP (GNU Image Manipulation\nprogram)\n  is a graphics\neditor designed to be driven through an interactive GUI.  But GIMP is\nbuilt as a library of image-manipulation and housekeeping routines\ncalled by a relatively thin layer of control code.  The driver code\nknows about the GUI, but not directly about image formats; the library\nroutines reverse this by knowing about image formats and operations\nbut not about the GUI.\n<\/p>\n<p>\n  The library layer is documented (and, in fact shipped as\n\u201clibgimp\u201d for use by other programs).  This means that C\nprograms called \u201cplugins\u201d can be dynamically loaded by\nGIMP and call the library to do image manipulation, effectively taking\nover control at the same level as the GUI (see Figure 4.2).\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 4.2. Caller\/callee relationships in GIMP with a plugin loaded.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Caller\/callee relationships in GIMP with a plugin loaded.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  Plugins are used to perform lots of special-purpose\ntransformations such as colormap hacking, blurring and despeckling;\nalso for reading and writing file formats not native to the GIMP core;\nfor extensions like editing animations and window manager themes; and\nfor lots of other sorts of image-hacking that can be automated by\nscripting the image-hacking logic in the GIMP core.  A registry of\nGIMP plugins is available on the World Wide Web.\n<\/p>\n<p>\n  Though most GIMP plugins are small, simple C programs, it is also possible to write a plugin that exposes the library\n  API to a scripting language; we'll discuss this possibility in <a href=\"#interfacechapter\" title=\"Chapter 11.&#10;  Interfaces\"> Chapter 11<\/a> when we examine the \u2018polyvalent program\u2019 pattern.\n<\/p>\n\n<h3 id=\"unix_and_oo\"><a href=\"#unix_and_oo\">\u00a7<\/a>Unix and Object-Oriented Languages<\/h3>\n<p>\n  Since the mid-1980s most new language designs have included\nnative support for\n  <em>object-oriented programming<\/em> (OO).  Recall that in object-oriented programming, the functions that\nact on a particular data structure are encapsulated with the data in\nan object that can be treated as a unit.  By contrast, modules in\nnon-OO languages make the association between data and the functions\nthat act on it rather accidental, and modules frequently leak data or\nbits of their internals into each other.\n<\/p>\n<p>\n  The OO design concept initially proved valuable in the design of\ngraphics systems, graphical user interfaces, and certain kinds of\nsimulation. To the surprise and gradual disillusionment of many, it\nhas proven difficult to demonstrate significant benefits of OO outside\nthose areas.  It's worth trying to understand why.\n<\/p>\n<p>\n  There is some tension and conflict between the Unix tradition of\nmodularity and the usage patterns that have developed around OO\nlanguages.  Unix programmers have always tended to be a bit more\nskeptical about OO than their counterparts elsewhere. Part of this is\nbecause of the Rule of Diversity; OO has far too often been promoted\nas the One True Solution to the software-complexity problem.  But\nthere is something else behind it as well, an issue which is worth\nexploring as background before we evaluate specific OO\n(object-oriented) languages in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>.  It will also help throw some\ncharacteristics of the Unix style of non-OO programming into sharper\nrelief.\n<\/p>\n<p>\n  We observed above that the Unix tradition of modularity is one\nof thin glue, a minimalist approach with few layers of abstraction\nbetween the hardware and the top-level objects of a program.  Part of\nthis is the influence of C.  It takes serious effort to simulate true\nobjects in C.  Because that's so, piling up abstraction layers is an\nexhausting thing to do.  Thus, object hierarchies in C tend to be\nrelatively flat and\ntransparent. Even when Unix programmers use other\nlanguages, they tend to want to carry over the\nthin-glue\/shallow-layering style that Unix models have taught\nthem.\n<\/p>\n<p>\n  OO languages make abstraction easy \u2014 perhaps too easy.\nThey encourage architectures with thick glue and elaborate layers.\nThis can be good when the problem domain is truly complex and demands\na lot of abstraction, but it can backfire badly if coders end up doing\nsimple things in complex ways just because they can.\n<\/p>\n<p>\n  All OO languages show some tendency to suck programmers into the\ntrap of excessive layering.  Object frameworks and object browsers are\nnot a substitute for good design or documentation, but they often get\ntreated as one.  Too many layers destroy\ntransparency\n  : It becomes too difficult\nto see down through them and mentally model what the code is actually\ndoing.  The Rules of Simplicity, Clarity, and Transparency get\nviolated wholesale, and the result is code full of obscure bugs and\ncontinuing maintenance problems.\n<\/p>\n<p>\n  This tendency is probably exacerbated because a lot of\nprogramming courses teach thick layering as a way to satisfy the Rule\nof Representation.  In this view, having lots of classes is\nequated with embedding knowledge in your data.  The problem with this\nis that too often, the \u2018smart data\u2019 in the glue layers is\nnot actually about any natural entity in whatever the program is\nmanipulating \u2014 it's just about being glue. (One sure sign of\nthis is a proliferation of abstract subclasses or\n\u2018mixins\u2019.)\n<\/p>\n<p>\n  Another side effect of OO abstraction is that opportunities for\noptimization\n  tend to disappear.  For example,\n  <em>a<\/em>\n  +\n  <em>a<\/em>\n  +\n  <em>a<\/em>\n  +\n  <em>a<\/em>\n  can become\n  <em>a<\/em>\n  * 4 and even\n  <em>a<\/em>\n  &lt;&lt; 2 if a is an integer.  But if one creates a class with\noperators, there is nothing to indicate if they are commutative,\ndistributive, or associative.  Since one isn't supposed to look inside\nthe object, it's not possible to know which of two equivalent\nexpressions is more efficient.  This isn't in itself a good reason to\navoid using OO techniques on new projects; that would be premature\noptimization.\nBut it is reason to think twice before transforming non-OO code into a\nclass hierarchy.\n<\/p>\n<p>\n  Unix programmers tend to share an instinctive sense of these\nproblems. This tendency appears to be one of the reasons that, under\nUnix, OO languages have failed to displace non-OO workhorses like\nC,\nPerl\n  (which\nactually has OO facilities, but they're not heavily used), and\nshell.\nThere is more vocal criticism of OO in the Unix world than orthodoxy\npermits elsewhere; Unix programmers know when\n  <em>not<\/em>\n  to use OO; and when they do use OO languages, they spend more\neffort on trying to keep their object designs uncluttered.  As the\nauthor of\n  <em>The Elements of Networking Style<\/em> once\nobserved in a slightly different context <sup><a href=\"#Padlipsky\" title=\"[Padlipsky]\">[Padlipsky]<\/a><\/sup>:\n\u201cIf you know what you're doing, three layers is enough; if you\ndon't, even seventeen levels won't help\u201d.\n<\/p>\n<p>\n  One reason that OO has succeeded most where it has (GUIs,\nsimulation, graphics) may be because it's relatively difficult to get\nthe ontology of types wrong in those domains.  In GUIs and graphics, for\nexample, there is generally a rather natural mapping between\nmanipulable visual objects and classes.  If you find yourself\nproliferating classes that have no obvious mapping to what goes\non in the display, it is correspondingly easy to notice that the glue\nhas gotten too thick.\n<\/p>\n<p>\n  One of the central challenges of design in the Unix style is how\nto combine the virtue of detachment (simplifying and generalizing\nproblems from their original context) with the virtue of thin glue and\nshallow, flat, transparent hierarchies of code and\ndesign.\n<\/p>\n<p>\n  We'll return to some of these points and apply them when we\ndiscuss object-oriented languages in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>.\n<\/p>\n\n<h3 id=\"ch04s06\"><a href=\"#ch04s06\">\u00a7<\/a>Coding for Modularity<\/h3>\n<p>\n  Modularity is expressed in good code, but it primarily comes\nfrom good design.  Here are some questions to ask about any code you \nwork on that might help you improve its modularity:\n<\/p>\n <div><ul><li>\n    <p>\n     How many global variables does it have?  Global variables are\nmodularity poison, an easy way for components to leak information to\neach other in careless and promiscuous ways.<sup><a href=\"#ftn.id2900865\" id=\"id2900865\">[48]<\/a><\/sup><\/p>\n   <\/li>\n   <li>\n    <p>\n     Is the size of your individual modules in Hatton's sweet spot?  If your\nanswer is \u201cNo, many are larger\u201d, you may have a long-term\nmaintenance problem.  Do you know what your own sweet spot is?  Do\nyou know what it is for other programmers you are cooperating with?\nIf not, best be conservative and stick to sizes near the low end of\nHatton's range.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Are the individual functions in your modules too large?  This\nis not so much a matter of line count as it is of internal complexity.  If you\ncan't informally describe a function's contract with its callers in\none line, the function is probably too large.<sup><a href=\"#ftn.id2900903\" id=\"id2900903\">[49]<\/a><\/sup><\/p>\n    <blockquote>\n     <p>\n      Personally I tend to break up a subprogram when there are too many\nlocal variables.  Another clue is [too many] levels of indentation.\nI rarely look at length.\n     <\/p><div>\n     \n      \u2013 Ken Thompson\n     \n    <\/div><\/blockquote>\n   <\/li>\n   <li>\n    <p>\n     Does your code have internal APIs \u2014 that is, collections\nof function calls and data structures that you can describe to others\nas units, each sealing off some layer of function from the rest of the\ncode?  A good API makes sense and is understandable without looking at\nthe implementation behind it.  The classic test is this: Try to\ndescribe it to another programmer over the phone.  If you fail, it\nis very probably too complex, and poorly designed.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Do any of your APIs have more than seven entry points?  Do any\nof your classes have more than seven methods each?  Do your data\nstructures have more than seven members?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     What is the distribution of the number of entry points per\nmodule across the project?<sup><a href=\"#ftn.id2900990\" id=\"id2900990\">[50]<\/a><\/sup>\n     Does it seem uneven?  Do the modules with lots of entry points really\nneed that many? Module complexity tends to rise as the square of the\nnumber of entry points, too \u2014 yet another reason simple APIs are\nbetter than complicated ones.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  You might find it instructive to compare these with our\nchecklist of questions about\ntransparency,\nand\ndiscoverability\n  in\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2900865\" id=\"ftn.id2900865\">[48]<\/a><\/sup>\n  Globals\nalso mean your code cannot be reentrant; that is, multiple instances in the\nsame process are likely to step on each\nother.\n<\/p><p><sup><a href=\"#id2900903\" id=\"ftn.id2900903\">[49]<\/a><\/sup>\n  Many\nyears ago, I learned from Kernighan &amp; Plauger's\n  <em>The Elements of Programming Style<\/em> a useful\nrule.  Write that one-line comment immediately after the prototype of\nyour function.  For\n  \n   <em>every<\/em>\n  \n  function, without\nexception.\n<\/p><p><sup><a href=\"#id2900990\" id=\"ftn.id2900990\">[50]<\/a><\/sup>\n  A cheap way to collect this\ninformation is to analyze the tags files generated by a utility like\netags(1)\nor\nctags(1).\n<\/p><\/div>\n\n<h3 id=\"textualitychapter\"><a href=\"#textualitychapter\">\u00a7<\/a>Chapter 5. Textuality<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   It's a well-known fact that computing devices such as the abacus\nwere invented thousands of years ago. But it's not well known that the\nfirst use of a common computer protocol occurred in the Old\nTestament. This, of course, was when Moses aborted the Egyptians'\nprocess with a control-sea.\n  <\/p><div>\n  \n   \u2013 Tom Galloway\nrec.arts.comics, February 1992\n  \n <\/div><\/blockquote>\n<p>\n  In this chapter, we'll look at what the Unix tradition has to\ntell us about two different kinds of design that are closely related:\nthe design of file formats for retaining application data in permanent\nstorage, and the design of application protocols for passing data\nand commands between cooperating programs, possibly over a\nnetwork.\n<\/p>\n<p>\n  What unifies these two kinds of design is that they both involve\nthe serialization of in-memory data structures.  For the internal\noperation of computer programs, the most convenient representation of\na complex data structure is one in which all fields have the machine's\nnative data format (e.g. two's-complement binary for integers) and all\npointers are actual memory addresses (as opposed, say, to being named\nreferences).  But these representations are not well suited to storage\nand transmission; memory addresses in the data structure lose their\nmeaning outside memory, and emitting raw native data formats causes\ninteroperability problems passing data between machines with different\nconventions (big- vs. little-endian, say, or 32-bit vs. 64-bit).\n<\/p>\n<p>\n  For transmission and storage, the traversable, quasi-spatial\nlayout of data structures like linked lists needs to be flattened or\nserialized into a byte-stream representation from which the structure\ncan later be recovered.  The serialization (save) operation is\nsometimes called\n  <em>marshaling<\/em> and its inverse\n(load) operation\n  <em>unmarshaling<\/em>.  These terms\nare usually applied with respect to objects in an\nOO\n  language like C++\n  or Python\n  or Java, but could be used with equal justice\nof operations like loading a graphics file into the internal storage\nof a graphics editor and saving it out after modifications.\n<\/p>\n<p>\n  A significant percentage of what C\n  and C++\n  programmers maintain is ad-hoc code for\nmarshaling and unmarshaling operations \u2014 even when the\nserialized representation chosen is as simple as a binary structure\ndump (a common technique under non-Unix environments).  Modern\nlanguages like Python\n  and Java\n  tend to have built-in unmarshal and\nmarshal functions that can be applied to any object or byte-stream\nrepresenting an object, and that reduce this labor substantially.\n<\/p>\n<p>\n  But these na\u00efve methods are often unsatisfactory for\nvarious reasons, including both the machine-interoperability problems\nwe mentioned above and the negative trait of being opaque to other\ntools.  When the application is a network protocol, economy may demand\nthat an internal data structure (such as, say, a message with source\nand destination addresses) be serialized not into a single blob of\ndata but into a series of attempted transactions or messages which the\nreceiving machine may reject (so that, for example, a large message can\nbe rejected if the destination address is invalid).\n<\/p>\n<p>\n  Interoperability,\ntransparency,\nextensibility,\nand storage or transaction economy: these are the important themes in\ndesigning file formats and application protocols.  Interoperability\nand transparency demand that we focus such designs on clean data\nrepresentations, rather than putting convenience of implementation or\nhighest possible performance first.  Extensibility also favors textual\nprotocols, since binary ones are often harder to extend or subset\ncleanly.  Transaction economy sometimes pushes in the opposite\ndirection \u2014 but we shall see that putting that criterion first\nis a form of premature\noptimization\n  that it is often wise to resist.\n<\/p>\n<p>\n  Finally, we must note a difference between data file formats and\nthe run-control files that are often used to set the startup options\nof Unix programs.  The most basic difference is that (with sporadic\nexceptions like GNU Emacs's configuration interface) programs don't\nnormally modify their own run-control files \u2014 the information\nflow is one-way, from file read at startup time to application\nsettings.  Data-file formats, on the other hand, associate properties\nwith named resources and are both read and written by their\napplications. Configuration files are generally hand-edited and small,\nwhereas data files are program-generated and can become arbitrarily\nlarge.\n<\/p>\n<p>\n  Historically, Unix has related but different sets of conventions for these two kinds of representation.  The\n  conventions for run control files are surveyed in <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">\n  Chapter 10<\/a>; only conventions for data files are examined in this chapter.\n<\/p>\n\n<h3 id=\"ch05s01\"><a href=\"#ch05s01\">\u00a7<\/a>The Importance of Being Textual<\/h3>\n<p>\n  Pipes and sockets will pass binary data as well as text. But\nthere are good reasons the examples we'll see in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  are textual: reasons that hark back\nto Doug McIlroy's\n  advice quoted in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>. Text streams are a valuable universal\nformat because they're easy for human beings to read, write, and edit\nwithout specialized tools.  These formats are (or can be designed to\nbe) transparent.\n<\/p>\n<p>\n  Also, the very limitations of text streams help enforce\nencapsulation. By discouraging elaborate representations with rich,\ndensely encoded structure, text streams also discourage programs from\nbeing promiscuous with each other about their internal states and help\nenforce encapsulation.  We'll return to this point at the end of\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  when we discuss RPC.\n<\/p>\n<p>\n  When you feel the urge to design a complex binary file format,\nor a complex binary application protocol, it is generally wise to\nlie down until the feeling passes. If performance is what you're\nworried about, implementing compression on the text protocol stream\neither at some level below or above the application protocol will\ngive you a cleaner and perhaps better-performing design than a \nbinary protocol (text compresses well, and quickly).\n<\/p>\n <blockquote>\n  <p>\n   A bad example of binary formats in Unix history was the way\ndevice-independent\n   <em>troff<\/em> read a binary file\ncontaining device information, supposedly for speed.  The initial\nimplementation generated that binary file from a text description in a\nsomewhat unportable way.  Faced with a need to port\n   <em>ditroff<\/em> <em>quickly<\/em>\n   to a new\nmachine, rather than reinvent the binary goo, I ripped it out and just\nhad\n   <em>ditroff<\/em> read the text file.  With carefully\ncrafted file-reading code, the speed penalty was negligible.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  Designing a textual protocol tends to future-proof your system.\nOne specific reason is that ranges on numeric fields aren't implied by\nthe format itself. Binary formats usually specify the number of bits\nallocated to a given value, and extending them is difficult.  For\nexample, IPv4 only allows 32 bits for an address.  To extend address\nsize to 128 bits (as done by IPv6) requires a major\nrevamping.<sup><a href=\"#ftn.id2901165\" id=\"id2901165\">[51]<\/a><\/sup>\n  In contrast, if you need a larger value\nin a text format, just write it.  It may be that a given program can't\nreceive values in that range, but it's usually easier to modify the\nprogram than to modify all the data stored in that format.\n<\/p>\n<p>\n  The only good justification for a binary protocol is if you're\ngoing to be manipulating large enough data sets that you're genuinely\nworried about getting the most bit-density out of your media, or if\nyou're very concerned about the time or instruction budget required to\ninterpret the data into an in-core structure.  Formats for large\nimages and multimedia are sometimes an example of the former, and\nnetwork protocols with hard latency requirements sometimes an \nexample of the latter.\n<\/p>\n <blockquote>\n  <p>\n   The reciprocal problem with SMTP or HTTP-like text protocols is\nthat they tend to be expensive in bandwidth and slow to parse.  The\nsmallest X request is 4 bytes: the smallest HTTP request is about 100\nbytes.  X requests, including amortized overhead of transport, can be\nexecuted in the order of 100 instructions; at one point, an\nApache\n   [web server]\ndeveloper proudly indicated they were down to 7000 instructions.  For\ngraphics, bandwidth becomes everything on output; hardware is designed\nsuch that these days the graphics-card bus is\n   <em>the<\/em>\n   bottleneck for small operations, so any protocol had better be very\ntight if it is not to be a worse bottleneck.  This is the extreme\ncase.\n  <\/p><div>\n  \n   \u2013 Jim Gettys\n  \n <\/div><\/blockquote>\n<p>\n  These concerns are valid in other extreme cases as well as in X\n\u2014 for example, in the design of graphics file formats intended\nto hold very large images. But they are usually just another case of\npremature-optimization\n  fever.  Textual formats don't necessarily have much lower bit density\nthan binary ones; they do after all use seven out of eight bits per\nbyte. And what you gain by not having to parse text, you generally\nlose the first time you need to generate a test load, or to eyeball a\nprogram-generated example of your format and figure out what's in\nthere.\n<\/p>\n<p>\n  In addition, the kind of thinking that goes into designing tight\nbinary formats tends to fall down on making them cleanly extensible.\nThe X designers experienced this:\n<\/p>\n <blockquote>\n  <p>\n   Against the current X framework is the fact we didn't design enough of\na structure to make it easier to ignore trivial extensions to the protocol;\nwe can do this some of the time, but a bit better framework would have been\ngood.\n  <\/p><div>\n  \n   \u2013 Jim Gettys\n  \n <\/div><\/blockquote>\n<p>\n  When you think you have an extreme case that justifies a binary\nfile format or protocol, you need to think very carefully about\nextensibility\n  and leaving room in the design for growth.\n<\/p>\n<h4 id=\"id2901332\"><a href=\"#id2901332\">\u00a7<\/a>Case Study: Unix Password File Format<\/h4>\n<p>\n  On many operating systems, the per-user data required to\nvalidate logins and start a user's session is an opaque binary\ndatabase. Under Unix, by contrast, it's a text file with records one\nper line and colon-separated fields.\n<\/p>\n<p>\n  Example 5.1 consists of some randomly-chosen example lines:\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.1. Password file example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\ngames:*:12:100:games:\/usr\/games:\ngopher:*:13:30:gopher:\/usr\/lib\/gopher-data:\nftp:*:14:50:FTP User:\/home\/ftp:\nesr:0SmFuPnH5JlNs:23:23:Eric S. Raymond:\/home\/esr:\nnobody:*:99:99:Nobody:\/:\n<\/pre><\/div>\n<p>\n  Without even knowing anything about the semantics of the fields,\nwe can notice that it would be hard to pack the data much tighter in\na binary format. The colon sentinel characters would have to have\nfunctional equivalents taking at least as much space (usually\neither count bytes or NULs). The per-user records would either have\nto have terminators (which could hardly be shorter than a single\nnewline) or else be wastefully padded out to a fixed length.\n<\/p>\n<p>\n  Actually the prospects for saving space through binary encoding\npretty much vanish if you know the actual semantics of the data.  The\nnumeric user ID (3rd) and group ID (4th) fields are integers, thus on\nmost machines a binary representation would be at least 4 bytes, and\nlonger than the text for values up to 999.  But let's agree to ignore\nthis for now and suppose the best case that the numeric fields have a\n0-255 range.\n<\/p>\n<p>\n  We could tighten up the numeric fields (3rd and 4th) by\ncollapsing the numerics to single bytes, and the password strings\n(2nd) to an 8-bit encoding. On this example, that would give about an\n8% size decrease.\n<\/p>\n<p>\n  That 8% of putative inefficiency buys us a lot. It avoids\nputting an arbitrary limit on the range of the numeric fields. It\ngives us the ability to modify the password file with any old text\neditor of our choice, rather than having to build a specialized tool\nto edit a binary format (though in the case of the password file\nitself, we have to be extra careful about concurrent edits). And it\ngives us the ability to do ad-hoc searches and filters and reports on\nthe user account information with text-stream tools such as\ngrep(1).\n<\/p>\n<p>\n  We do have to be a bit careful about not embedding a colon in\nany of the textual fields.  Good practice is to tell the file write\ncode to precede embedded colons with an escape character, and then to\ntell the file read code to interpret it. Unix tradition favors\nbackslash for this use.\n<\/p>\n<p>\n  The fact that structural information is conveyed by field\nposition rather than an explicit tag makes this format faster \nto read and write, but a bit rigid.  If the set of properties\nassociated with a key is expected to change with any frequency, \none of the tagged formats described below might be a better choice.\n<\/p>\n<p>\n  Economy is not a major issue with password files to begin with,\nas they're normally read seldom<sup><a href=\"#ftn.id2901458\" id=\"id2901458\">[52]<\/a><\/sup>\n  and infrequently modified.  Interoperability\nis not an issue, since various data in the file (notably user and\ngroup numbers) are not portable off the originating machine.  For\npassword files, it's therefore quite clear that going where the\ntransparency\ncriterion\n  leads was the right thing.\n<\/p>\n<h4 id=\"id2901494\"><a href=\"#id2901494\">\u00a7<\/a>Case Study: .newsrc Format<\/h4>\n<p>\n  Usenet\n  news is\na worldwide distributed bulletin-board system that anticipated today's\nP2P networking by two decades.  It uses a message format very similar\nto that of RFC 822 electronic-mail messages, except that instead of\nbeing directed to personal recipients messages are sent to topic\ngroups.  Articles posted at any participating site are broadcast to\neach site that it has registered as a neighbor, and eventually\nflood-fill to all news sites.\n<\/p>\n<p>\n  Almost all Usenet news readers understand the\n  <code>.newsrc<\/code> file, which records which Usenet messages\nhave been seen by the calling user.  Though it is named like a\nrun-control file, it is not only read at startup but typically updated\nat the end of the newsreader run.  The\n  <code>.newsrc<\/code> format has been fixed since the first newsreaders around 1980.\n   Example 5.2 is a representative section from a <code>.newsrc<\/code> file.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.2. A\n    <code>.newsrc<\/code> example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nrec.arts.sf.misc! 1-14774,14786,14789\nrec.arts.sf.reviews! 1-2534\nrec.arts.sf.written: 1-876513\nnews.answers! 1-199359,213516,215735\nnews.announce.newusers! 1-4399\nnews.newusers.questions! 1-645661\nnews.groups.questions! 1-32676\nnews.software.readers! 1-95504,137265,137274,140059,140091,140117\nalt.test! 1-1441498\n<\/pre><\/div>\n<p>\n  Each line sets properties for the newsgroup named in the first\nfield. The name is immediately followed by a character that indicates \nwhether the owning user is currently subscribed to the group or not; a colon\nindicates subscription, and an exclamation mark indicates\nnonsubscription. The remainder of the line is a sequence of\ncomma-separated article numbers or ranges of article numbers,\nindicating which articles the user has seen.\n<\/p>\n<p>\n  Non-Unix programmers might have automatically tried to design a\nfast binary format in which each newsgroup status was described by\neither a long but fixed-length binary record, or a sequence of\nself-describing binary packets with internal length fields.  The main\npoint of such a binary representation would be to express ranges with\nbinary data in paired word-length fields, in order to avoid the\noverhead of parsing all the range expressions at startup.\n<\/p>\n<p>\n  Such a layout could be read and written faster than a textual\nformat, but it would have other problems.  A na\u00efve implementation\nin fixed-length records would have placed artificial length limits on\nnewsgroup names and (more seriously) on the maximum number of ranges\nof seen-article numbers.  A more sophisticated binary-packet format\nwould avoid the length limits, but could not be edited with the user's\neyeballs and fingers \u2014 a capability that can be quite useful\nwhen you want to reset just some of the read bits in an individual\nnewsgroup.  Also, it would not necessarily be portable to different\nmachine types.\n<\/p>\n<p>\n  The designers of the original newsreader chose\ntransparency\n  and interoperability over\neconomy.  The case for going in the other direction was not completely\nridiculous;\n  <code>.newsrc<\/code> files can get very large, and\none modern reader (GNOME's Pan) uses a speed-optimized private format\nto avoid startup lag.  But to other implementers, textual\nrepresentation looked like a good tradeoff in 1980, and has looked\nbetter as machines increased in speed and storage dropped in\nprice.\n<\/p>\n<h4 id=\"png\"><a href=\"#png\">\u00a7<\/a>Case Study: The PNG Graphics File Format<\/h4>\n<p>\n  PNG (Portable Network Graphics) is a file format for bitmap\ngraphics.\n  It is like GIF,\nand unlike JPEG, in that it uses lossless compression and is optimized\nfor applications such as line art and icons rather than photographic\nimages.  Documentation and open-source reference libraries of high\nquality are available at the\n  <a href=\"http:\/\/www.libpng.org\/pub\/png\/\">\n   Portable Network Graphics\nwebsite<\/a>.\n<\/p>\n<p>\n  PNG is an excellent example of a thoughtfully designed binary\nformat.  A binary format is appropriate since graphics files may\ncontain very large amounts of data, such that storage size and\nInternet download time would go up significantly if the pixel data\nwere stored textually.  Transaction economy was the prime\nconsideration, with\ntransparency\n  sacrificed.<sup><a href=\"#ftn.id2901736\" id=\"id2901736\">[53]<\/a><\/sup>\n  The designers were, however, careful about\ninteroperability; PNG specifies byte orders, integer word lengths,\nendianness, and (lack of) padding between fields.\n<\/p>\n<p>\n  A PNG file consists of a sequence of chunks, each in a\nself-describing format beginning with the chunk type name and the\nchunk length.  Because of this organization, PNG does not need a\nrelease number.  New chunk types can be added at any time; the case of\nthe first letter in the chunk\ntype name informs PNG-using software whether or not each chunk can be\nsafely ignored.\n<\/p>\n<p>\n  The PNG file header also repays study.  It has been \ncleverly designed to make various common kinds of file corruption\n(e.g., by 7-bit transmission links, or mangling of CR and LF\ncharacters) easy to detect.\n<\/p>\n<p>\n  The PNG standard is precise, comprehensive, and well written. It\ncould serve as a model for how to write file format standards.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2901165\" id=\"ftn.id2901165\">[51]<\/a><\/sup>\n  There is a legend that some early airline\nreservation systems allocated exactly one byte for a plane's passenger\ncount.  Supposedly they became very confused by the arrival of the\nBoeing 747, the first plane that could carry more than 255\npassengers.\n<\/p><p><sup><a href=\"#id2901458\" id=\"ftn.id2901458\">[52]<\/a><\/sup>\n  Password files are\nnormally read once per user session at login time, and after that\noccasionally by file-system utilities like\nls(1)\nthat must map from numeric user and group IDs to\nnames.\n<\/p><p><sup><a href=\"#id2901736\" id=\"ftn.id2901736\">[53]<\/a><\/sup>\n  Confusingly, PNG supports a different kind\nof transparency \u2014 transparent pixels in the\nPNG image.\n<\/p><\/div>\n\n<h3 id=\"ch05s02\"><a href=\"#ch05s02\">\u00a7<\/a>Data File Metaformats<\/h3>\n<p>\n  A data file metaformat is a set of syntactic and lexical\nconventions that is either formally standardized or sufficiently well\nestablished by practice that there are standard service libraries to\nhandle marshaling and unmarshaling it.\n<\/p>\n<p>\n  Unix has evolved or adopted metaformats suitable for a wide range\nof applications. It is good practice to use one of these (rather than\nan idiosyncratic custom format) wherever possible.  The benefits begin\nwith the amount of custom parsing and generation code that you may be\nable to avoid writing by using a service library.  But the most\nimportant benefit is that developers and even many users will instantly \nrecognize these formats and feel comfortable with them, which reduces\nthe friction costs of learning new programs.\n<\/p>\n<p>\n  In the following discussion, when we refer to \u201ctraditional\nUnix tools\u201d we are intending the combination of\ngrep(1),\nsed(1),\nawk(1),\ntr(1),\nand\ncut(1)\nfor doing text searches and transformations.\nPerl\n  and other\nscripting languages\n  tend to have good native support for\nparsing the line-oriented formats that these tools encourage.\n<\/p>\n<p>\n  Here, then, are the standard formats that can serve you as models.\n<\/p>\n<h4 id=\"id2901882\"><a href=\"#id2901882\">\u00a7<\/a>DSV Style<\/h4>\n<p>\n  DSV stands for\n  <em>Delimiter-Separated Values<\/em>.  Our first case study in textual metaformats was\nthe\n  <code>\/etc\/passwd<\/code> file, which is a DSV format with\ncolon as the value separator.  Under Unix, colon is the default\nseparator for DSV formats in which the field values may contain\nwhitespace.\n<\/p>\n<p>\n  <code>\/etc\/passwd<\/code> format (one record \nper line, colon-separated fields) is very traditional under Unix\nand frequently used for tabular data.  Other classic examples \ninclude the\n  <code>\/etc\/group<\/code> file describing security \ngroups and the\n  <code>\/etc\/inittab<\/code> file used to control\nstartup and shutdown of Unix service programs at different run levels\nof the operating system.\n<\/p>\n<p>\n  Data files in this style are expected to support inclusion of\ncolons in the data fields by backslash escaping.  More generally,\ncode that reads them is expected to support record continuation by\nignoring backslash-escaped newlines, and to allow embedding\nnonprintable character data by C-style backslash escapes.\n<\/p>\n<p>\n  This format is most appropriate when the data is tabular,\nkeyed by a name (in the first field), and records are typically\nshort (less than 80 characters long).  It works well with \ntraditional Unix tools.\n<\/p>\n<p>\n  One occasionally sees field separators other than the colon,\nsuch as the pipe character | or even an ASCII NUL.  Old-school Unix\npractice used to favor tabs, a preference reflected in the defaults\nfor\ncut(1)\nand\npaste(1);\nbut this has gradually changed as format designers became aware of the\nmany small irritations that ensue from the fact that tabs and spaces\nare not visually distinguishable.\n<\/p>\n<p>\n  This format is to Unix what CSV (comma-separated value) format \nis under Microsoft Windows and elsewhere outside the Unix world.\nCSV (fields separated by commas, double quotes used to escape \ncommas, no continuation lines) is rarely found under Unix.\n<\/p>\n<p>\n  In fact, the Microsoft version of CSV is a textbook example of\nhow\n  <em>not<\/em>\n  to design a textual file format. Its\nproblems begin with the case in which the separator character (in this\ncase, a comma) is found inside a field. The Unix way would be to\nsimply escape the separator with a backslash, and have a double escape\nrepresent a literal backslash.  This design gives us a single special case\n(the escape character) to check for when parsing the file, and only a\nsingle action when the escape is found (treat the following character\nas a literal). The latter conveniently not only handles the separator\ncharacter, but gives us a way to handle the escape character and\nnewlines for free. CSV, on the other hand, encloses the entire field\nin double quotes if it contains the separator.  If the field contains\ndouble quotes, it must also be enclosed in double quotes, and the\nindividual double quotes in the field must themselves be repeated\ntwice to indicate that they don't end the field.\n<\/p>\n<p>\n  The bad results of proliferating special cases are twofold.\nFirst, the complexity of the parser (and its vulnerability to bugs) is\nincreased.  Second, because the format rules are complex and\nunderspecified, different implementations diverge in their handling of\nedge cases.  Sometimes continuation lines\n  <em>are<\/em>\n  supported, by starting the last field of the line with an unterminated\ndouble quote \u2014 but only in some products!  Microsoft has\nincompatible versions of CSV files between its own applications, and\nin some cases between different versions of the same application\n(Excel being the obvious example here).\n<\/p>\n<h4 id=\"id2902039\"><a href=\"#id2902039\">\u00a7<\/a>RFC 822 Format<\/h4>\n<p>\n  The RFC 822 metaformat derives from the textual format of\nInternet electronic mail messages; RFC 822 is the principal Internet RFC\ndescribing this format (since superseded by RFC 2822).  MIME\n(Multipurpose Internet Media Extension)\n  provides\na way to embed typed binary data within RFC-822-format messages.  (Web\nsearches on either of these names will turn up the relevant\nstandards.)\n<\/p>\n<p>\n  In this metaformat, record attributes are stored one per line,\nnamed by tokens resembling mail header-field names and terminated with\na colon followed by whitespace.  Field names do not contain\nwhitespace; conventionally a dash is substituted instead. The\nattribute value is the entire remainder of the line, exclusive of\ntrailing whitespace and newline.  A physical line that begins with tab\nor whitespace is interpreted as a continuation of the current logical\nline. A blank line may be interpreted either as a record terminator or\nas an indication that unstructured text follows.\n<\/p>\n<p>\n  Under Unix, this is the traditional and preferred textual\nmetaformat for attributed messages or anything that can be closely\nanalogized to electronic mail.  More generally, it's appropriate for\nrecords with a varying set of fields in which the hierarchy of data is\nflat (no recursion or tree structure).\n<\/p>\n<p>\n  Usenet news\n  uses\nit; so do the HTTP 1.1 (and later) formats used by the World Wide Web.\nIt is very convenient for editing by humans.  Traditional Unix search\ntools are still good for attribute searches, though finding record\nboundaries will be a little more work than in a record-per-line\nformat.\n<\/p>\n<p>\n  One weakness of RFC 822 format is that when more than one RFC\n822 message or record is put in a file, the record boundaries may not\nbe obvious \u2014 how is a poor literal-minded computer to know where\nthe unstructured text body of a message ends and the next header\nbegins?  Historically, there have been several different conventions\nfor delimiting messages in mailboxes.  The oldest and most widely\nsupported, leading each message with a line that begins with the\nstring\n  <code>\"From \"<\/code> and sender information, is not\nappropriate for other kinds of records; it also requires that lines in\nmessage text beginning with\n  <code>\"From \"<\/code> be escaped\n(typically with\n  <code>&gt;<\/code>) \u2014 a practice which not infrequently leads\nto confusion.\n<\/p>\n<p>\n  Some mail systems use delimiter lines consisting of control\ncharacters unlikely to appear in messages, such as several ASCII 01\n(control-A) characters in succession. The MIME standard gets around\nthe problem by including an explicit message length in the header, but\nthis is a fragile solution which is very likely to break if messages\nare ever manually edited.  For a somewhat better solution, see the\nrecord-jar style described later in this chapter.\n<\/p>\n<p>\n  For examples of RFC 822 format, look in your mailbox.\n<\/p>\n<h4 id=\"id2902164\"><a href=\"#id2902164\">\u00a7<\/a>Cookie-Jar Format<\/h4>\n<p>\n  Cookie-jar format is used by the\nfortune(1)\nprogram for its database of random quotes.  It is appropriate for\nrecords that are just bags of unstructured text.  It simply uses newline\nfollowed by\n  <code>%%<\/code> (or sometimes newline followed by <code>%<\/code>) as a record separator. Example 5.3\n  is an example section from a file of email signature quotes:\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.3. A fortune file example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n\"Among the many misdeeds of British rule in India, history will look\nupon the Act depriving a whole nation of arms as the blackest.\"\n        -- Mohandas Gandhi, \"An Autobiography\", pg 446\n%\nThe people of the various provinces are strictly forbidden to have \nin their possession any swords, short swords, bows, spears, firearms,\nor other types of arms. The possession of unnecessary implements \nmakes difficult the collection of taxes and dues and tends to foment \nuprisings.\n        -- Toyotomi Hideyoshi, dictator of Japan, August 1588\n%\n\"One of the ordinary modes, by which tyrants accomplish their \npurposes without resistance, is, by disarming the people, and making \nit an offense to keep arms.\"\n        -- Supreme Court Justice Joseph Story, 1840\n<\/pre><\/div>\n<p>\n  It is good practice to accept whitespace after\n  <code>%<\/code> when looking\nfor record delimiters.  This helps cope with human editing mistakes.\nIt's even better practice to use\n  <code>%%<\/code>, and ignore all text from\n  <code>%%<\/code> to\nend-of-line.\n<\/p>\n <blockquote>\n  <p>\n   The cookie-jar separator was originally\n   <code>%%\\n<\/code>.  I wanted\nsomething a bit more visible than\n   <code>%<\/code> would have been.  In fact, any\nstuff after the\n   <code>%%<\/code> is treated as a comment (or at least that's how I\nwrote it).\n  <\/p><div>\n  \n   \u2013 Ken Arnold\n  \n <\/div><\/blockquote>\n<p>\n  Simple cookie-jar format is appropriate for pieces of\ntext that have no natural ordering, distinguishable structure above\nword level, or search keys other than their text context.\n<\/p>\n<h4 id=\"id2906931\"><a href=\"#id2906931\">\u00a7<\/a>Record-Jar Format<\/h4>\n<p>\n  Cookie-jar record separators combine well with the RFC 822\nmetaformat for records, yielding a format we'll call\n\u2018record-jar\u2019. If you need a textual format that will\nsupport multiple records with a variable repertoire of explicit\nfieldnames, one of the least surprising and human-friendliest ways to\ndo it would look like Example 5.4.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.4. Basic data for three planets in a record-jar format.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nPlanet: Mercury\nOrbital-Radius: 57,910,000 km\nDiameter: 4,880 km\nMass: 3.30e23 kg\n%%\nPlanet: Venus\nOrbital-Radius: 108,200,000 km\nDiameter: 12,103.6 km\nMass: 4.869e24 kg\n%%\nPlanet: Earth\nOrbital-Radius: 149,600,000 km\nDiameter: 12,756.3 km\nMass: 5.972e24 kg\nMoons: Luna\n<\/pre><\/div>\n<p>\n  Of course, the record delimiter could be a blank line, but a\nline consisting of \"\n  <code>%%\\n<\/code> \" is more explicit and less likely to be\nintroduced by accident during editing (two printable characters are\nbetter than one because it can't be generated by a single-character\ntypo).  In a format like this it is good practice to simply ignore\nblank lines.\n<\/p>\n<p>\n  If your records have an unstructured text part, your record-jar\nformat is closely approaching a mailbox format. In this case, it's\nimportant that you have a well-defined way to escape the record\ndelimiter so it can appear in text; otherwise, your record reader is\ngoing to choke on an ill-formed text part someday.  Some technique\nanalogous to byte-stuffing (described later in this chapter) is\nindicated.\n<\/p>\n<p>\n  Record-jar format is appropriate for sets of field-attribute \nassociations that are like DSV files, but have a variable repertoire\nof fields, and possibly unstructured text associated with them.\n<\/p>\n<h4 id=\"id2907018\"><a href=\"#id2907018\">\u00a7<\/a>XML<\/h4>\n<p>\n  XML is a very simple syntax resembling HTML \u2014\nangle-bracketed tags and ampersand-led literal sequences.  It is about\nas simple as a plain-text markup can be and yet express recursively\nnested data structures.  XML is just a low-level syntax; it requires\na document type definition (such as XHTML) and associated application\nlogic to give it semantics.\n<\/p>\n<p>\n  XML is well suited for complex data formats (the sort of things\nfor which the old-school Unix tradition would use an RFC-822-like stanza\nformat) though overkill for simpler ones.  It is especially\nappropriate for formats that have a complex nested or recursive\nstructure of the sort that the RFC 822 metaformat does not handle well.\nFor a good introduction to the format, see\n  <em>XML in a Nutshell<\/em><sup><a href=\"#Harold-Means\" title=\"[Harold-Means]\">[Harold-Means]<\/a><\/sup>.\n<\/p>\n <blockquote>\n  <p>\n   Among the hardest things to get right in designing any text file\nformat are issues of quoting, whitespace and other low-level syntax\ndetails.  Custom file formats often suffer from slightly broken syntax\nthat doesn't quite match other similar formats.  Using a standard format\nsuch as XML, which is verifiable and parsed by a standard library,\neliminates most of these issues.\n  <\/p><div>\n  \n   \u2013 Keith Packard\n  \n <\/div><\/blockquote>\n<p>\n   Example 5.5 is a simple example of an XML-based configuration file.  It is part of the <em>kdeprint<\/em> tool shipped\n   with the open-source KDE office suite hosted under Linux.  It describes options for an image-to-PostScript filtering\n   operation, and how to map them into arguments for a filter command.  For another instructive example, see the\n   discussion of\n  <em>Glade<\/em> in <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">Chapter 8<\/a>. <\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.5. An XML example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n&lt;?xml version=\"1.0\"?&gt;\n&lt;kprintfilter name=\"imagetops\"&gt;\n    &lt;filtercommand \n           data=\"imagetops %filterargs %filterinput %filteroutput\" \/&gt;\n    &lt;filterargs&gt;\n        &lt;filterarg name=\"center\" \n                   description=\"Image centering\" \n                   format=\"-nocenter\" type=\"bool\" default=\"true\"&gt;\n            &lt;value name=\"true\" description=\"Yes\" \/&gt;\n            &lt;value name=\"false\" description=\"No\" \/&gt;\n        &lt;\/filterarg&gt;\n        &lt;filterarg name=\"turn\" \n                   description=\"Image rotation\" \n                   format=\"-%value\" type=\"list\" default=\"auto\"&gt;\n            &lt;value name=\"auto\" description=\"Automatic\" \/&gt;\n            &lt;value name=\"noturn\" description=\"None\" \/&gt;\n            &lt;value name=\"turn\" description=\"90 deg\" \/&gt;\n        &lt;\/filterarg&gt;\n        &lt;filterarg name=\"scale\" \n                   description=\"Image scale\" \n                   format=\"-scale %value\" \n                   type=\"float\" \n                        min=\"0.0\" max=\"1.0\" default=\"1.000\" \/&gt;\n        &lt;filterarg name=\"dpi\" \n                   description=\"Image resolution\" \n                   format=\"-dpi %value\" \n                   type=\"int\" min=\"72\" max=\"1200\" default=\"300\" \/&gt;\n    &lt;\/filterargs&gt;\n    &lt;filterinput&gt;\n        &lt;filterarg name=\"file\" format=\"%in\" \/&gt;\n        &lt;filterarg name=\"pipe\" format=\"\" \/&gt;\n    &lt;\/filterinput&gt;\n    &lt;filteroutput&gt;\n        &lt;filterarg name=\"file\" format=\"&gt; %out\" \/&gt;\n        &lt;filterarg name=\"pipe\" format=\"\" \/&gt;\n    &lt;\/filteroutput&gt;\n&lt;\/kprintfilter&gt;\n<\/pre><\/div>\n<p>\n  One advantage of XML is that it is often possible to\ndetect ill-formed, corrupted, or incorrectly generated data through a\nsyntax check, without knowing the semantics of the data.\n<\/p>\n<p>\n  The most serious problem with XML is that it doesn't play well\nwith traditional Unix tools.  Software that wants to read an XML\nformat needs an XML parser; this means bulky, complicated programs.\nAlso, XML is itself rather bulky; it can be difficult to see the\ndata amidst all the markup.\n<\/p>\n<p>\n  One application area in which XML is clearly winning is in\nmarkup formats for document files (we'll have more to say about this\nin\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>).\nTagging in such documents tends to be relatively sparse among large\nblocks of plain text; thus, traditional Unix tools still work fairly\nwell for simple text searches and transformations.\n<\/p>\n<p>\n  One interesting bridge between these worlds is PYX format\n\u2014 a line-oriented translation of XML that can be hacked with\ntraditional line-oriented Unix text tools and then losslessly\ntranslated back to XML.  A Web search for \u201cPyxie\u201d will\nturn up resources.  The xmltk toolkit takes the opposite tack,\nproviding stream-oriented tools analogous to\ngrep(1)\nand\nsort(1)\nfor filtering XML documents; Web search for\n\u201cxmltk\u201d to find it.\n<\/p>\n<p>\n  XML can be a simplifying choice or a complicating one.  There is\na lot of hype surrounding it, but don't become a fashion victim by either\nadopting or rejecting it uncritically. Choose carefully and bear the KISS\nprinciple in mind.\n<\/p>\n<h4 id=\"id2907263\"><a href=\"#id2907263\">\u00a7<\/a>Windows INI Format<\/h4>\n<p>\n  Many Microsoft Windows programs use a textual data format that looks like Example 5.6. This example associates\n  optional resources named\n  <code>account<\/code>, <code>directory<\/code>, <code>numeric_id<\/code>, and <code>developer<\/code> with named projects\n  <code>python<\/code>, <code>sng<\/code>, <code>fetchmail<\/code>, and <code>py-howto<\/code>.  The DEFAULT entry supplies\n  values that will be used when a named entry fails to supply them.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.6. A\n    <code>.INI<\/code> file example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n[DEFAULT]\naccount = esr\n[python]\ndirectory = \/home\/esr\/cvs\/python\/\ndeveloper = 1\n[sng]\ndirectory = \/home\/esr\/WWW\/sng\/\nnumeric_id = 1012\ndeveloper = 1\n[fetchmail]\nnumeric_id = 18364\n[py-howto]\naccount = eric\ndirectory = \/home\/esr\/cvs\/py-howto\/\ndeveloper = 1\n<\/pre><\/div>\n<p>\n  This style of data-file format is not native to Unix, but some\nLinux programs (notably Samba, the suite of tools for accessing\nWindows file shares from Linux) support it under Windows's influence.\nThis format is readable and not badly designed, but like XML it\ndoesn't play well with\ngrep(1)\nor conventional Unix scripting tools.\n<\/p>\n<p>\n  The .INI format is appropriate if your data naturally falls into\nits two-level organization of name-attribute pairs clustered under\nnamed records or sections.  It's not good for data with a fully\nrecursive treelike structure (XML is more appropriate for that), \nand it would be overkill for a simple list of name-value\nassociations (use DSV format for that).\n<\/p>\n<h4 id=\"id2907428\"><a href=\"#id2907428\">\u00a7<\/a>Unix Textual File Format Conventions<\/h4>\n<p>\n  There are long-standing Unix traditions about how textual data\nformats ought to look.  Most of these derive from one or more of the\nstandard Unix metaformats we've just described.  It is wise to follow these\nconventions unless you have strong and specific reasons to do otherwise.\n<\/p>\n<p>\n  In\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">\n   Chapter 10<\/a>\n  we\nwill discuss a different set of conventions used for program\nrun-control files, but you should notice that it will share some of\nthese same rules (especially about the lexical level, the rules by\nwhich characters are assembled into tokens).\n<\/p>\n <div><ul><li>\n    <p>\n     <em>\n      One record per newline-terminated line, if\npossible.\n<\/em>\n     This makes it easy to extract records with\ntext-stream tools.  For data interchange with other operating systems, \nit's wise to make your file-format parser indifferent to whether \nthe line ending is LF or CR-LF. It's also conventional to ignore\ntrailing whitespace in such formats; this protects against common\neditor bobbles.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Less than 80 characters per line, if possible.<\/em>\n     This\nmakes the format browseable in an ordinary-sized terminal window.\nIf many records must be longer than 80 characters, consider a\nstanza format (see below).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Use\n      <code>#<\/code> as an introducer for comments.\n<\/em>\n     It is good to have a way to embed annotations and comments in\ndata files.  It's best if they're actually part of the file\nstructure, and so will be preserved by tools that know its format.\nFor comments that are not preserved during parsing,\n     <code>#<\/code> is the \nconventional start character.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Support the backslash convention.<\/em>\n     The least\nsurprising way to support embedding nonprintable control characters\nis by parsing C-like backslash escapes \u2014\n     <code>\\n<\/code> for a newline,\n     <code>\\r<\/code> for a carriage return,\n     <code>\\t<\/code> for a tab,\n     <code>\\b<\/code> for backspace,\n     <code>\\f<\/code> for formfeed,\n     <code>\\e<\/code> for ASCII escape (27),\n     <code>\\nnn<\/code> or\n     <code>\\onnn<\/code> or\n     <code>\\0nnn<\/code> for the character with octal\nvalue\n     <code>nnn<\/code>,\n     <code>\\xnn<\/code> for the character with hexadecimal value\n     <code>nn<\/code>,\n     <code>\\dnnn<\/code> for the character with decimal value\n     <code>nnn<\/code>,\n     <code>\\\\<\/code> for a literal backslash.  A newer convention, but one worth following,\nis the use of\n     <code>\\unnnn<\/code> for a hexadecimal\nUnicode literal.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      In one-record-per-line formats, use colon or any run\nof whitespace as a field separator.\n<\/em>\n     The colon convention\nseems to have originated with the Unix password file. If your fields\nmust contain instances of the separator(s), use a backslash as the\nprefix to escape them.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Do not allow the distinction between tab and\nwhitespace to be significant.\n<\/em>\n     This is a recipe for serious\nheadaches when the tab settings on your users' editors are different;\nmore generally, it's confusing to the eye.  Using tab alone as a field\nseparator is especially likely to cause problems; allowing any run of\ntabs and spaces to be a field separator, on the other hand, works\nwell.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Favor hex over octal.<\/em>\n     Hex-digit pairs and\nquads are easier to eyeball-map into bytes and today's 32- and 64-bit\nwords than octal digits of three bits each; also marginally more\nefficient.  This rule needs emphasizing because some older Unix tools\nsuch as\nod(1)\nviolate it; that's a legacy from the instruction field sizes in the\nmachine languages of older PDP minicomputers.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      For complex records, use a \u2018stanza\u2019 format:\nmultiple lines per record, with a record separator line of\n      <code>%%\\n<\/code> or\n      <code>%\\n<\/code>.\n<\/em>\n     The separators make useful visual boundaries for human\nbeings eyeballing the file.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      In stanza formats, either have one record field per\nline or use a record format resembling RFC 822 electronic-mail headers,\nwith colon-terminated field-name keywords leading fields.\n<\/em>\n     The second choice is appropriate when fields are often either absent\nor longer than 80 characters, or when records are sparse (e.g., often\nwith empty fields).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>In stanza formats, support line continuation.<\/em>\n     When interpreting the file, either discard backslash followed by\nwhitespace or interpret newline followed by whitespace equivalently to a single\nspace, so that a long logical line can be folded into short (easily\neditable!) physical lines. It's also conventional to ignore\ntrailing whitespace in these formats; this convention protects against common\neditor bobbles.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Either include a version number or design the format\nas self-describing chunks independent of each other.\n<\/em>\n     If\nthere is even the faintest possibility that the format will have to be\nchanged or extended, include a version number so your code can\nconditionally do the right thing on all versions.  Alternatively,\ndesign the format as self-describing chunks so that you can add new\nchunk types without instantly breaking old code.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Beware of floating-point round-off problems.<\/em>\n     Conversion of floating-point numbers from binary to text format and\nback can lose precision, depending on the quality of the conversion\nlibrary you are using.  If the structure you are\nmarshaling\/unmarshaling contains floating point, you should test the\nconversion in both directions.  If it looks like conversion in\neither direction is subject to roundoff errors, be prepared to dump\nthe floating-point field as raw binary instead, or a string encoding\nthereof.  If you're coding in C or some language that has access to\nC printf\/scanf, the C99\n     <code>%a<\/code> specifier may solve this problem.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Don't bother compressing or binary-encoding just part of\nthe file.\n<\/em>\n     See below...\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<h4 id=\"id2907903\"><a href=\"#id2907903\">\u00a7<\/a>The Pros and Cons of File Compression<\/h4>\n<p>\n  Many modern Unix projects, such as OpenOffice.org and AbiWord,\nnow use XML compressed with\nzip(1)\nor\ngzip(1)\nas a data file format.  Compressed XML combines space economy with some of the\nadvantages of a textual format \u2014 notably, it avoids the problem\nthat binary formats must often allocate space for information that may\nnot be used in particular cases (e.g., for unusual options or large\nranges).  But there is some dispute about this, dispute which turns on some\nof the central tradeoffs discussed in this chapter.\n<\/p>\n<p>\n  On the one hand, experiments have shown that documents in a\ncompressed XML file are usually significantly smaller than the\nMicrosoft Word's native file format, a binary format that one might\nimagine would take less space.  The reason relates to a fundamental of\nthe Unix philosophy: Do one thing well.  Creating a single tool\nto do the compression job well is more effective than ad-hoc\ncompression on parts of the file, because the tool can look across all\nthe data and exploit\n  <em>all<\/em>\n  repetition in the\ninformation.\n<\/p>\n<p>\n  Also, by separating the representation design from the\nparticular compression method used, you leave open the possibility of\nusing different compression methods in the future with no more than\nminimal changes to the actual file parsing \u2014 perhaps, with no\nchanges at all.\n<\/p>\n<p>\n  On the other hand, compression does some damage to transparency.\nWhile a human being can estimate from context whether uncompressing\nthe file is likely to show him anything useful, tools such as\nfile(1)\ncannot as of mid-2003 see through the wrapping.\n<\/p>\n<p>\n  Some would advocate a less structured compression format \u2014\nstraight\ngzip(1)-compressed\nXML data, say, without the internal structure and self-identifying\nheader chunk provided by\nzip(1). While\nusing a format similar to that of \nzip(1)\nsolves the identification problem, it means that decoding such\nfiles will be tricky for programs written in the simpler scripting\nlanguages.\n<\/p>\n<p>\n  Any of these solutions (straight text, straight binary, or\ncompressed text) may be optimal depending on the relative weight you\ngive to storage economy, discoverability, or making browsing tools\nas simple as possible to write.  The point of the preceding discussion\nis not to advocate any one of these approaches over the others, but\nrather to suggest how you can think about the options and design\ntradeoffs clearly.\n<\/p>\n<p>\n  This having been said, the truly Unixy solution would probably\nbe to fix\nfile(1)\nto see file prefixes through the compression \u2014 and, failing\nthat, to write a shellscript wrapper around\nfile(1)\nthat would interpret compression as a direction to apply\ngunzip(1)\nand take a second look.\n<\/p>\n\n<h3 id=\"ch05s03\"><a href=\"#ch05s03\">\u00a7<\/a>Application Protocol Design<\/h3>\n<p>\n  In\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>,\nwe'll discuss the advantages of breaking complicated applications up\ninto cooperating processes speaking an application-specific command\nset or protocol with each other.  All the good reasons for data file\nformats to be textual apply to these application-specific protocols as\nwell.\n<\/p>\n<p>\n  When your application protocol is textual and easily parsed\nby eyeball, many good things become easier.  Transaction dumps become\nmuch easier to interpret. Test loads become easier to write.\n<\/p>\n<p>\n  Server processes are often invoked by harness programs such as\ninetd(8)\nin such a way that the server sees commands on standard input and\nships responses to standard output.  We describe this \u201cCLI\nserver\u201d pattern in more detail in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>.\n<\/p>\n<p>\n  A CLI server with a command set that is designed for simplicity has\nthe valuable property that a human tester will be able to type\ncommands direct to the server process to probe the software's\nbehavior.\n<\/p>\n<p>\n  Another issue to bear in mind is the end-to-end design\nprinciple.  Every protocol designer should read the classic\n  <em>End-to-End Arguments in System Design<\/em> <sup><a href=\"#Saltzer\" title=\"[Saltzer]\">[Saltzer]<\/a><\/sup>. There are often serious questions about which\nlevel of the protocol stack should handle features like security and\nauthentication; this paper provides some good conceptual tools for\nthinking about them. Yet a third issue is designing application\nprotocols for good performance.  We'll cover that issue in more detail\nin\n  <a href=\"#optimizationchapter\" title=\"Chapter 12. Optimization\">\n   Chapter 12<\/a>.\n<\/p>\n<p>\n  The traditions of Internet application protocol design evolved\nseparately from Unix before 1980.<sup><a href=\"#ftn.id2908167\" id=\"id2908167\">[54]<\/a><\/sup>\n  But since the 1980s these traditions have become thoroughly\nnaturalized into Unix practice.\n<\/p>\n<p>\n  We'll illustrate the Internet style by looking at three\napplication protocols that are both among the most heavily used, and\nare widely regarded among Internet hackers as paradigmatic: SMTP,\nPOP3, and IMAP.  All three address different aspects of mail transport\n(one of the net's two most important applications, along with the\nWorld Wide Web), but the problems they address (passing messages,\nsetting remote state, indicating error conditions) are generic to\nnon-email application protocols as well and are normally addressed\nusing similar techniques.\n<\/p>\n<h4 id=\"id2908194\"><a href=\"#id2908194\">\u00a7<\/a>Case Study: SMTP, the Simple Mail Transfer Protocol<\/h4>\n<p>\n   Example 5.7 is an example transaction in SMTP (Simple Mail Transfer Protocol), which is described by RFC 2821. In the\n   example, <em>C:<\/em> lines are sent by a mail transport agent (MTA) sending mail, and <em>S:<\/em> lines are returned\n   by the MTA receiving it. Text <em> emphasized like this <\/em> is comments, not part of the actual transaction.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.7. An SMTP session example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n<em>C: &lt;client connects to service port 25&gt;<\/em>\n<em>C:<\/em> HELO snark.thyrsus.com               <em>sending host identifies self<\/em>\n<em>S:<\/em> 250 OK Hello snark, glad to meet you  <em>receiver acknowledges<\/em>\n<em>C:<\/em> MAIL FROM: &lt;esr@thyrsus.com&gt;         <em>identify sending user<\/em>\n<em>S:<\/em> 250 &lt;esr@thyrsus.com&gt;... Sender ok   <em>receiver acknowledges<\/em>\n<em>C:<\/em> RCPT TO: cor@cpmy.com                <em>identify target user<\/em>\n<em>S:<\/em> 250 root... Recipient ok             <em>receiver acknowledges<\/em>\n<em>C:<\/em> DATA\n<em>S:<\/em> 354 Enter mail, end with \".\" on a line by itself\n<em>C:<\/em> Scratch called.  He wants to share\n<em>C:<\/em> a room with us at Balticon.\n<em>C:<\/em>.                                    <em>end of multiline send<\/em>\n<em>S:<\/em> 250 WAA01865 Message accepted for delivery\n<em>C:<\/em> QUIT                                 <em>sender signs off<\/em>\n<em>S:<\/em> 221 cpmy.com closing connection      <em>receiver disconnects<\/em>\n<em>C: &lt;client hangs up&gt;<\/em>\n<\/pre><\/div>\n<p>\n  This is how mail is passed among Internet machines.  Note the\nfollowing features: command-argument format of the requests, responses\nconsisting of a status code followed by an informational message, the\nfact that the payload of the DATA command is terminated by a line\nconsisting of a single dot.\n<\/p>\n<p>\n  SMTP is one of the two or three oldest application protocols \nstill in use on the Internet.   It is simple, effective, and has\nwithstood the test of time.  The traits we have called out here are \ntropes that recur frequently in other Internet protocols.  If there is\nany single archetype of what a well-designed Internet application \nprotocol looks like, SMTP is it.\n<\/p>\n<h4 id=\"id2908383\"><a href=\"#id2908383\">\u00a7<\/a>Case Study: POP3, the Post Office Protocol<\/h4>\n<p>\n  Another one of the classic Internet protocols is POP3, the Post\nOffice Protocol.  It is also used for mail transport, but where SMTP\nis a \u2018push\u2019 protocol with transactions initiated by the\nmail sender, POP3 is a \u2018pull\u2019 protocol with transactions \ninitiated by the mail receiver.  Internet users with intermittent\naccess (like dial-up connections) can let their mail pile up on \na mail-drop machine, then use a POP3 connection to pull mail\nup the wire to their personal machines.\n<\/p>\n<p>\n   Example 5.8 is an example POP3 session. In the example, <em>C:<\/em> lines are sent by the client, and <em>S:<\/em>\n   lines by the mail server. Observe the many similarities with SMTP.  This protocol is also textual and line-oriented,\n   sends payload message sections terminated by a line consisting of a single dot followed by line terminator, and even\n   uses the same exit command, QUIT. Like SMTP, each client operation is acknowledged by a reply line that begins with a\n   status code and includes an informational message meant for human eyes.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.8. A POP3 example session.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n<em>C: &lt;client connects to service port 110&gt;<\/em>\n<em>S:<\/em> +OK POP3 server ready &lt;1896.6971@mailgate.dobbs.org&gt;\n<em>C:<\/em> USER bob\n<em>S:<\/em> +OK bob\n<em>C:<\/em> PASS redqueen\n<em>S:<\/em> +OK bob's maildrop has 2 messages (320 octets)\n<em>C:<\/em> STAT\n<em>S:<\/em> +OK 2 320\n<em>C:<\/em> LIST\n<em>S:<\/em> +OK 2 messages (320 octets)\n<em>S:<\/em> 1 120\n<em>S:<\/em> 2 200\n<em>S:<\/em>.\n<em>C:<\/em> RETR 1\n<em>S:<\/em> +OK 120 octets\n<em>S: &lt;the POP3 server sends the text of message 1&gt;<\/em>\n<em>S:<\/em>.\n<em>C:<\/em> DELE 1\n<em>S:<\/em> +OK message 1 deleted\n<em>C:<\/em> RETR 2\n<em>S:<\/em> +OK 200 octets\n<em>S: &lt;the POP3 server sends the text of message 2&gt;<\/em>\n<em>S:<\/em>.\n<em>C:<\/em> DELE 2\n<em>S:<\/em> +OK message 2 deleted\n<em>C:<\/em> QUIT\n<em>S:<\/em> +OK dewey POP3 server signing off (maildrop empty)\n<em>C: &lt;client hangs up&gt;<\/em>\n<\/pre><\/div>\n<p>\n  There are a few differences.  The most obvious one is that POP3\nuses status tokens rather than SMTP's 3-digit status codes.  Of course\nthe requests have different semantics.  But the family resemblance\n(one we'll have more to say about when we discuss the generic Internet\nmetaprotocol later in this chapter) is clear.\n<\/p>\n<h4 id=\"id2908582\"><a href=\"#id2908582\">\u00a7<\/a>Case Study: IMAP, the Internet Message Access Protocol<\/h4>\n<p>\n  To complete our triptych of Internet application protocol examples, we'll look at IMAP, another post office protocol\n  designed in a slightly different style.  See Example 5.9; as before, <em>C:<\/em> lines are sent by the client, and\n  <em>S:<\/em> lines by the mail server.  Text <em>emphasized like this<\/em> is comments, not part of the actual\n  transaction.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 5.9. An IMAP session example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n<em>C: &lt;client connects to service port 143&gt;<\/em>\n<em>S:<\/em> * OK example.com IMAP4rev1 v12.264 server ready\n<em>C:<\/em> A0001 USER \"frobozz\" \"xyzzy\"\n<em>S:<\/em> * OK User frobozz authenticated\n<em>C:<\/em> A0002 SELECT INBOX\n<em>S:<\/em> * 1 EXISTS\n<em>S:<\/em> * 1 RECENT\n<em>S:<\/em> * FLAGS (\\Answered \\Flagged \\Deleted \\Draft \\Seen)\n<em>S:<\/em> * OK [UNSEEN 1] first unseen message in \/var\/spool\/mail\/esr\n<em>S:<\/em> A0002 OK [READ-WRITE] SELECT completed\n<em>C:<\/em> A0003 FETCH 1 RFC822.SIZE                    <em>Get message sizes<\/em>\n<em>S:<\/em> * 1 FETCH (RFC822.SIZE 2545)\n<em>S:<\/em> A0003 OK FETCH completed\n<em>C:<\/em> A0004 FETCH 1 BODY[HEADER]                   <em>Get first message header<\/em>\n<em>S:<\/em> * 1 FETCH (RFC822.HEADER {1425}\n<em>&lt;server sends 1425 octets of message payload&gt;<\/em>\n<em>S:<\/em>)\n<em>S:<\/em> A0004 OK FETCH completed\n<em>C:<\/em> A0005 FETCH 1 BODY[TEXT]                     <em>Get first message body<\/em>\n<em>S:<\/em> * 1 FETCH (BODY[TEXT] {1120}\n<em>&lt;server sends 1120 octets of message payload&gt;<\/em>\n<em>S:<\/em>)\n<em>S:<\/em> * 1 FETCH (FLAGS (\\Recent \\Seen))\n<em>S:<\/em> A0005 OK FETCH completed\n<em>C:<\/em> A0006 LOGOUT\n<em>S:<\/em> * BYE example.com IMAP4rev1 server terminating connection\n<em>S:<\/em> A0006 OK LOGOUT completed\n<em>C: &lt;client hangs up&gt;<\/em>\n<\/pre><\/div>\n<p>\n  IMAP delimits payloads in a slightly different way.  Instead of\nending the payload with a dot, the payload length is sent just before\nit.  This increases the burden on the server a little bit (messages\nhave to be composed ahead of time, they can't just be streamed up\nafter the send initiation) but makes life easier for the client, which\ncan tell in advance how much storage it will need to allocate to\nbuffer the message for processing as a whole.\n<\/p>\n<p>\n  Also, notice that each response is tagged with a sequence label\nsupplied by the request; in this example they have the form A000n, \nbut the client could have generated any token into that slot.  This\nfeature makes it possible for IMAP commands to be streamed to the\nserver without waiting for the responses; a state machine in the \nclient can then simply interpret the responses and payloads as\nthey come back.  This technique cuts down on latency.\n<\/p>\n<p>\n  IMAP (which was designed to replace POP3) is an excellent\nexample of a mature and powerful Internet application protocol design,\none well worth study and emulation.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2908167\" id=\"ftn.id2908167\">[54]<\/a><\/sup>\n  One relic of this pre-Unix history is that Internet\nprotocols normally use CR-LF as a line terminator rather than\nUnix's bare LF.\n<\/p><\/div>\n\n<h3 id=\"ch05s04\"><a href=\"#ch05s04\">\u00a7<\/a>Application Protocol Metaformats<\/h3>\n<p>\n  Just as data file metaformats have evolved to simplify serialization\nfor storage, application protocol metaformats have evolved to simplify\nserialization for transactions across networks.  The tradeoffs are a\nlittle different in this case; because network bandwidth is more expensive\nthan storage, there is more of a premium on transaction economy.\nStill, the transparency and interoperability benefits of textual\nformats are sufficiently strong that most designers have resisted\nthe temptation to optimize for performance at the cost of readability.\n<\/p>\n<h4 id=\"id2908835\"><a href=\"#id2908835\">\u00a7<\/a>The Classical Internet Application Metaprotocol<\/h4>\n<p>\n  Marshall Rose's RFC 3117,\n  <em>On the Design of Application Protocols<\/em>,<sup><a href=\"#ftn.id2908849\" id=\"id2908849\">[55]<\/a><\/sup>\n  provides an excellent overview of the\ndesign issues in Internet application protocols.  It makes explicit\nseveral of the tropes in classical Internet application protocols that\nwe observed in our examination of SMTP, POP, and IMAP, and provides an\ninstructive taxonomy of such protocols.  It is recommended\nreading.\n<\/p>\n<p>\n  The classical Internet metaprotocol is\ntextual.  It uses single-line requests and responses, except for\npayloads which may be multiline.  Payloads are shipped either with a\npreceding length in octets or with a terminator that is the line\n  <code>\".\\r\\n\"<\/code>. In the latter case the payload is\n  <em id=\"byte_stuffing\">byte-stuffed<\/em>; all lines that start with a period get\nanother period prepended, and the receiver side is responsible for\nboth recognizing the termination and stripping away the stuffing.\nResponse lines consist of a status code followed by a human-readable\nmessage.\n<\/p>\n<p>\n  One final advantage of this classical style is that it is\nreadily extensible.  The parsing and state-machine framework doesn't\nneed to change much to accommodate new requests, and it is easy to\ncode implementations so that they can parse unknown requests and\nreturn an error or simply ignore them.  SMTP, POP3, and IMAP have all\nbeen extended in minor ways fairly often during their lifetimes, with\nminimal interoperability problems.  Na\u00efvely designed binary\nprotocols are, by contrast, notoriously brittle.\n<\/p>\n<h4 id=\"id2908915\"><a href=\"#id2908915\">\u00a7<\/a>HTTP as a Universal Application Protocol<\/h4>\n<p>\n  Ever since the World Wide Web reached critical mass around 1993, \napplication protocol designers have shown an increasing tendency to\nlayer their special-purpose protocols on top of HTTP, using web servers\nas generic service platforms.\n<\/p>\n<p>\n  This is a viable option because, at the transaction layer, HTTP is very simple and general.  An HTTP request is a\n  message in an RFC-822\/MIME-like format; typically, the headers contain identification and authentication information,\n  and the first line is a method call on some resource specified by a Universal Resource Indicator (URI).  The most\n  important methods are GET (fetch the resource), PUT (modify the resource) and POST (ship data to a form or back-end\n  process). The most important form of URI is a URL or Uniform Resource Locator, which identifies the resource by\n  service type, host name, and a location on the host.  An HTTP response is simply an RFC-822\/MIME message and can\n  contain arbitrary content to be interpreted by the client.\n<\/p>\n<p>\n  Web servers handle the transport and request-multiplexing layers\nof HTTP, as well as standard service types like http and ftp.  It is\nrelatively easy to write web server plugins that will handle custom\nservice types, and to dispatch on other elements of the URI format.\n<\/p>\n<p>\n  Besides avoiding a lot of lower-level details, this method means\nthe application protocol will tunnel through the standard HTTP service\nport and not need a\nTCP\/IP\n  service port of\nits own.  This can be a distinct advantage; most firewalls leave port 80\nopen, but trying to punch another hole through can be fraught with\nboth technical and political difficulties.\n<\/p>\n<p>\n  With this advantage comes a risk.  It means that your web server \nand its plugins grow more complex, and cracks in any of that code can\nhave large security implications.  It may become more difficult to\nisolate and shut down problem services.  The usual tradeoffs between \nsecurity and convenience apply.\n<\/p>\n<p>\n  RFC 3205,\n  <em>On the Use of HTTP As a Substrate<\/em>,<sup><a href=\"#ftn.id2909002\" id=\"id2909002\">[56]<\/a><\/sup>\n  has good design advice for anyone\nconsidering using HTTP as the underlayer of an application\nprotocol, including a summary of the tradeoffs and problems involved.\n<\/p>\n<h5 id=\"id2909019\"><a href=\"#id2909019\">\u00a7<\/a>Case Study: The CDDB\/freedb.org Database<\/h5>\n<p>\n  Audio CDs consist of a sequence of music tracks in a digital\nformat called CDDA-WAV.  They were designed to be played by very simple\nconsumer-electronics devices a few years before general-purpose\ncomputers developed enough raw speed and sound capability to decode\nthem on the fly.  Because of this, there is no provision in the \nformat for even simple metainformation such as the album and track\ntitles.  But modern computer-hosted CD players want this information so\nthe user can assemble and edit play lists.\n<\/p>\n<p>\n  Enter the Internet.  There are (at least two) repositories that\nprovide a mapping between a hash code computed from the track-length\ntable on a CD and artist\/album-title\/track-title records.  The\noriginal was\n  <code>cddb.org<\/code>, but\nanother site called\n  <code>freedb.org<\/code> which is probably now more\ncomplete and widely used.  Both sites rely on their users for the\nenormous task of keeping the database current as new CDs come out;\n  <code>freedb.org<\/code> arose from a\ndeveloper revolt after CDDB elected to take all that user-contributed\ninformation proprietary .\n<\/p>\n<p>\n  Queries to these services could have been implemented as a\ncustom application protocol on top of TCP\/IP, but that would have\nrequired steps such as getting a new TCP\/IP port number assigned and\nfighting to get a hole for it punched through thousands of firewalls.\nInstead, the service is implemented over HTTP as a simple CGI query\n(as if the CD's hash code had been supplied by a user filling in a Web\nform).\n<\/p>\n<p>\n  This choice makes all the existing infrastructure of HTTP and\nWeb-access libraries in various programming languages available to\nsupport programs for querying and updating this database.  As a\nresult, adding such support to a software CD player is nearly trivial,\nand effectively every software CD player knows how to use them.\n<\/p>\n<h5 id=\"id2909113\"><a href=\"#id2909113\">\u00a7<\/a>Case Study: Internet Printing Protocol<\/h5>\n<p>\n  Internet Printing Protocol (IPP) is a successful, widely implemented\nstandard for the control of network-accessible printers.  Pointers to\nRFCs, implementations, and much other related material are available\nat the IETF's\n  <a href=\"http:\/\/www.pwg.org\/ipp\/\">\n   Printer Working\nGroup<\/a>\n  site.\n<\/p>\n<p>\n  IPP uses HTTP 1.1 as a transport layer.  All IPP requests are\npassed via an HTTP POST method call; responses are ordinary HTTP\nresponses. (Section 4.2 of RFC 2568,\n  <em>Rationale for the Structure of the Model and Protocol for the Internet Printing Protocol<\/em>, does an excellent job of explaining this\nchoice; it repays study by anyone considering writing a new application\nprotocol.)\n<\/p>\n<p>\n  From the software side, HTTP 1.1 is widely deployed.  It already\nsolves many of the transport-level problems that would otherwise\ndistract protocol developers and implementers from concentrating on\nthe domain semantics of printing.  It is cleanly extensible, so there\nis room for IPP to grow.  The CGI programming model for handling the\nPOST requests is well understood and development tools are widely\navailable.\n<\/p>\n<p>\n  Most network-aware printers already embed a web server, because\nthat's the natural way to make the status of the printer remotely \nqueryable by human beings.  Thus, the incremental cost of adding IPP\nservice to the printer firmware is not large.  (This is an argument\nthat could be applied to a remarkably wide range of other\nnetwork-aware hardware, including vending machines and coffee makers<sup><a href=\"#ftn.id2909179\" id=\"id2909179\">[57]<\/a><\/sup>\n  and hot tubs!)\n<\/p>\n<p>\n  About the only serious drawback of layering IPP over HTTP is\nthat the protocol is completely driven by client requests.  Thus there\nis no space in the model for printers to ship asynchronous alert\nmessages back to clients.  (However, smarter clients could run a\ntrivial HTTP server to receive such alerts formatted as HTTP requests\nfrom the printer.)\n<\/p>\n<h4 id=\"id2909217\"><a href=\"#id2909217\">\u00a7<\/a>BEEP: Blocks Extensible Exchange Protocol<\/h4>\n<p>\n  BEEP (formerly BXXP) is a generic protocol machine that competes\nwith HTTP for the role of universal underlayer for application\nprotocols.  There is a niche open because there is not as yet\nany other more established metaprotocol that is appropriate for truly\npeer-to-peer applications, as opposed to the client-server\napplications that HTTP handles well.  A\n  <a href=\"http:\/\/www.beepcore.org\/beepcore\/docs\/sl-beep.jsp\">\n   project\nwebsite<\/a>\n  provides access to standards and open-source\nimplementations in several languages.\n<\/p>\n<p>\n  BEEP has features to support both client-server and peer-to-peer\nmodes.  The authors designed the BEEP protocol and support library so\nthat picking the right options abstracts away messy issues like data\nencoding, flow control, congestion-handling, support of end-to-end\nencryption, and assembling a large response composed of multiple\ntransmissions,\n<\/p>\n<p>\n  Internally, BEEP peers exchange sequences of self-describing\nbinary packets not unlike chunk types in\nPNG.  The design is tuned\nmore for economy and less for\ntransparency\n  than the classical Internet protocols or HTTP, and might be a better\nchoice when data volumes are large.  BEEP also avoids the HTTP problem\nthat all requests have to be client-initiated; it would be better in\nsituations in which a server needs to send asynchronous status messages\nback to the client.\n<\/p>\n<p>\n  BEEP is still new technology in mid-2003, and has only a few\ndemonstration projects.  But the BEEP papers are good analytical\nsurveys of best practice in protocol design; even if BEEP itself fails\nto gain widespread adoption, the papers will retain considerable tutorial\nvalue.\n<\/p>\n<h4 id=\"id2909294\"><a href=\"#id2909294\">\u00a7<\/a>XML-RPC, SOAP, and Jabber<\/h4>\n<p>\n  There is a developing trend in application protocol design\ntoward using XML within\nMIME\n  to structure\nrequests and payloads.  BEEP peers use this format for channel\nnegotiations.  Three major protocols are going the XML route\nthroughout: XML-RPC and SOAP (Simple Object Access Protocol) for\nremote procedure calls, and Jabber for instant messaging and presence.\nAll three are XML document types.\n<\/p>\n<p>\n  XML-RPC is very much in the Unix spirit (its author observes\nthat he learned how to program in the 1970s by reading the original\nsource code for Unix). It's deliberately minimalist but nevertheless\nquite powerful, offering a way for the vast majority of RPC\napplications that can get by on passing around scalar\nboolean\/integer\/float\/string datatypes to do their thing in a way that\nis lightweight and easy to understand and monitor. XML-RPC's type\nontology is richer than that of a text stream, but still simple and\nportable enough to act as a valuable check on interface\ncomplexity. Open-source implementations are available. An excellent\n  <a href=\"http:\/\/www.xmlrpc.com\/\">\n   XML-RPC home page<\/a>\n  points to\nspecifications and multiple open-source implementations.\n<\/p>\n<p>\n  SOAP is a more heavyweight RPC protocol with a richer type\nontology that includes arrays and C-like structs.  It was inspired by\nXML-RPC, but has been plausibly accused of being an overdesigned\nvictim of the second-system effect.  As of mid-2003 the SOAP\nstandard is still a work in progress, but a trial implementation in\nApache\n  is tracking the\ndrafts.  Open-source client modules in Perl, Python, Tcl, and Java\n  are readily discoverable by a Web\nsearch.  The W3C draft specification is\n  <a href=\"http:\/\/www.w3.org\/TR\/SOAP\/\">\n   available on the Web<\/a>.\n<\/p>\n<p>\n  XML-RPC and SOAP, considered as remote procedure call methods,\nhave some associated risks that we discuss at the end of\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>.\n<\/p>\n<p>\n  Jabber is a peer-to-peer protocol designed to support instant\nmessaging and presence.  What makes it interesting as an application\nprotocol is that it supports passing around XML forms and live\ndocuments. Specifications, documentation, and open-source\nimplementations are available at the\n  <a href=\"http:\/\/www.jabber.org\/about\/overview.html\">\n   Jabber\nSoftware Foundation<\/a>\n  site.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2908849\" id=\"ftn.id2908849\">[55]<\/a><\/sup>\n  See\n  <a href=\"ftp:\/\/ftp.rfc-editor.org\/in-notes\/rfc3117.txt\">\n   RFC\n3117<\/a>.\n<\/p><p><sup><a href=\"#id2909002\" id=\"ftn.id2909002\">[56]<\/a><\/sup>\n  See\n  <a href=\"http:\/\/www.faqs.org\/rfcs\/rfc3205.html\">\n   RFC\n3205<\/a>.\n<\/p><p><sup><a href=\"#id2909179\" id=\"ftn.id2909179\">[57]<\/a><\/sup>\n  See\n  <a href=\"http:\/\/www.ietf.org\/rfc\/rfc2324.txt\">\n   RFC 2324<\/a>\n  and\n  <a href=\"http:\/\/www.ietf.org\/rfc\/rfc2325.txt\">\n   RFC 2325<\/a>.\n<\/p><\/div>\n\n<h3 id=\"transparencychapter\"><a href=\"#transparencychapter\">\u00a7<\/a>Chapter 6. Transparency<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Beauty is more important in\ncomputing than anywhere else in technology because software is so\ncomplicated. Beauty is the ultimate defense against complexity.\n  <\/p><div>\n  \n   \u2013 David Gelernter\nMachine Beauty: Elegance and the Heart of Technology (1998)\n  \n <\/div><\/blockquote>\n<p>\n  In the previous chapter we discussed the importance of\ntextual data formats and application protocols, representations that\nare easy for human beings to examine and interact with.  These promote\nqualities in design that are much valued in the Unix tradition but\nseldom if ever talked about explicitly:\n  <em>transparency<\/em> and\n  <em>discoverability<\/em>.\n<\/p>\n<p>\n  Software systems are transparent when they don't have murky\ncorners or hidden depths.  Transparency is a passive quality.  A\nprogram is transparent when it is possible to form a simple mental\nmodel of its behavior that is actually predictive for all or most\ncases, because you can see through the machinery to what is actually\ngoing on.\n<\/p>\n<p>\n  Software systems are discoverable when they include features\nthat are designed to help you build in your mind a correct mental\nmodel of what they do and how they work.  So, for example, good\ndocumentation helps discoverability to a user.  Good choice of\nvariable and function names helps discoverability to a\nprogrammer. Discoverability\n  is an active quality. To achieve it in your software you cannot merely\nfail to be obscure, you have to go out of your way to be helpful.<sup><a href=\"#ftn.id2911508\" id=\"id2911508\">[58]<\/a><\/sup><\/p>\n<p>\n  Transparency and discoverability are important for both users\nand software developers. But they're important in different ways.\nUsers like these properties in a UI because they mean an easier\nlearning curve.  UI transparency and discoverability are a large part\nof what people mean when they say a UI is \u2018intuitive\u2019;\nmost of the rest is the Rule of Least Surprise.  We'll examine the\nproperties that make user interfaces pleasant and effective in more\ndepth in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>.\n<\/p>\n<p>\n  Software developers like these qualities in the code itself (the\npart users don't see) because they so often need to understand it well\nenough to modify and debug it.  Also, a program designed so that its\ninternal data flows are readily comprehensible is more likely to be\none that does not fail because of bad interactions that the designer\ndidn't notice, and more likely to be able to evolve forward gracefully\n(including accommodating change when new maintainers pick up the\nbaton).\n<\/p>\n<p>\n  Transparency is a major component of what David\nGelernter\n  refers to as \u201cbeauty\u201d in this chapter's epigraph.  Unix\nprogrammers, borrowing from mathematicians, often use the more\nspecific term \u201celegance\u201d for the quality Gelernter speaks\nof.  Elegance is a combination of power and simplicity.  Elegant code\ndoes much with little.  Elegant code is not only correct but visibly,\n  <em>transparently<\/em>\n  correct.  It does not merely\ncommunicate an algorithm to a computer, but also conveys insight and\nassurance to the mind of a human that reads it.  By seeking elegance\nin our code, we build better code.  Learning to write transparent code\nis a first, long step toward learning how to write elegant code\n\u2014 and taking care to make code discoverable helps us learn how\nto make it transparent.  Elegant code is both transparent and\ndiscoverable.\n<\/p>\n<p>\n  It may be easier to appreciate the difference between\ntransparency and discoverability with a pair of extreme examples.  The\nLinux\n  kernel source is remarkably transparent (given the intrinsic\ncomplexity of what it does) but not at all discoverable \u2014\nacquiring the minimum knowledge needed to live in the code and\nunderstand the idiom of the developers is difficult, but once you do\nthe whole makes sense.<sup><a href=\"#ftn.id2911617\" id=\"id2911617\">[59]<\/a><\/sup>\n  On the other hand, the\nEmacs Lisp libraries are discoverable but not transparent.  It's easy\nto acquire enough knowledge to tweak just one thing, but quite\ndifficult to comprehend the whole system.\n<\/p>\n<p>\n  In this chapter, we'll examine features of Unix designs that\npromote transparency and discoverability not just in UIs but in the\nparts users don't normally see.  We'll develop some useful rules you\ncan apply to your coding and development practice.  Later on, in\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>\n  we'll see how good release-engineering\npractices (like having a\n  <code>README<\/code> file with\nappropriate content) can make your source code as discoverable as your\ndesign.\n<\/p>\n<p>\n  If you need a practical reminder why these qualities are\nimportant, remember that the sanity you save by writing transparent,\ndiscoverable systems may well be that of your own future self.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2911508\" id=\"ftn.id2911508\">[58]<\/a><\/sup>\n  An economically-minded friend comments: \u201cDiscoverability is\nabout reducing barriers to entry; transparency is about reducing the \ncost of living in the code\u201d.\n<\/p><p><sup><a href=\"#id2911617\" id=\"ftn.id2911617\">[59]<\/a><\/sup>\n  The Linux kernel makes a number\nof attempts at discoverability, including the Documentation\nsubdirectory in the Linux kernel source tarball and quite a number of\ntutorial websites and books.  These attempts are frustrated by the\nspeed at which the kernel changes; the documentation has a chronic\ntendency to fall behind.\n<\/p><\/div>\n\n<h3 id=\"ch06s01\"><a href=\"#ch06s01\">\u00a7<\/a>Studying Cases<\/h3>\n<p>\n  Normal practice in this book has been to intersperse case studies with\nphilosophy.  But in this chapter we'll begin by looking at several\nUnix designs that exhibit transparency and discoverability, and\nattempt to draw lessons from them only after all have been\npresented.  Each major point of the analysis in the latter half of this\nchapter draws on several of these, and the arrangement avoids forward\nreferences to case studies the reader hasn't seen yet.\n<\/p>\n<h4 id=\"audacity\"><a href=\"#audacity\">\u00a7<\/a>Case Study: audacity<\/h4>\n<p>\n  First, we'll look at an example of transparency in UI design.  It\nis\n  <em>audacity<\/em>, an open-source editor for\nsound files that runs on Unix systems, Mac OS X, and Windows.\nSources, downloadable binaries, documentation, and screen shots are\navailable at the\n  <a href=\"http:\/\/audacity.sourceforge.net\/\">\n   project site<\/a>.\n<\/p>\n<p>\n  This program supports cutting, pasting, and editing of audio\nsamples. It supports multitrack editing and mixing.  The UI is\nsuperbly simple; the sound waveforms are shown in the\n  <em>audacity<\/em> window. The image of the\nwaveform can be cut and pasted; operations on that image are directly\nreflected in the audio sample as soon as they are performed.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 6.1. Screen shot of\n    <em>audacity<\/em>.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Screen shot of audacity.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  Multitrack editing is supported in the simplest possible way;\nthe screen splits into multiple per-track displays in a spatial\nrelationship that conveys their concurrency and makes it easy to \nmatch features by inspection.  Tracks can be dragged right or\nleft with the mouse to change their relative timing.\n<\/p>\n<p>\n  Several features of this UI are subtly excellent and worthy of\nemulation: the large, easily visible and clickable operation \nbuttons with distinguishing colors, the presence of an undo command\nthat removes most of the risk from experimentation, the volume slider\nthat makes softness\/loudness visually obvious in its shape.\n<\/p>\n<p>\n  But these are details.  The central virtue of this program is\nthat it has a superbly transparent and natural user interface, one that\nerects as few barriers between the user and the sound file as\npossible.\n<\/p>\n<h4 id=\"fetchmail_v\"><a href=\"#fetchmail_v\">\u00a7<\/a>Case Study: fetchmail's -v option<\/h4>\n<p>\n  <em>fetchmail<\/em> is a network gateway program. Its main purpose is to translate between\nPOP3 or IMAP remote-mail protocols and the Internet's native SMTP\nprotocol for email exchange. It is in extremely widespread use on Unix\nmachines that use intermittent SLIP or PPP connections to Internet\nservice providers, and as such probably touches an appreciable\nfraction of the Internet's mail traffic.\n<\/p>\n<p>\n  <em>fetchmail<\/em> has no fewer than 60\ncommand-line options (which, as we'll establish later in this book, is\nprobably too many), and a number of other options that are settable\nfrom the run-control file but not from the command line.  Of all\nthese, the most important \u2014 by far \u2014 is\n  <code>-v<\/code>, the verbose option.\n<\/p>\n<p>\n  When\n  <code>-v<\/code> is on,\n  <em>fetchmail<\/em> dumps each one of its POP, IMAP, and SMTP transactions to standard output as they happen.  A\n  developer can actually see the code doing protocol with remote mailservers and the mail transport program it forwards\n  to, in real time.  Users can send session transcripts with their bug reports. Example 6.1 shows a representative\n  session transcript.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 6.1. An example\n    <em>fetchmail<\/em> <code>-v<\/code> transcript.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nfetchmail: 6.1.0 querying hurkle.thyrsus.com (protocol IMAP) \n              at Mon, 09 Dec 2002 08:41:37 -0500 (EST): poll started\nfetchmail: running ssh %h \/usr\/sbin\/imapd \n              (host hurkle.thyrsus.com service imap)\nfetchmail: IMAP&lt; * PREAUTH [42.42.1.0] IMAP4rev1 v12.264 server ready\nfetchmail: IMAP&gt; A0001 CAPABILITY\nfetchmail: IMAP&lt; * CAPABILITY IMAP4 IMAP4REV1 NAMESPACE IDLE SCAN \n              SORT MAILBOX-REFERRALS LOGIN-REFERRALS AUTH=LOGIN \n              THREAD=ORDEREDSUBJECT\nfetchmail: IMAP&lt; A0001 OK CAPABILITY completed\nfetchmail: IMAP&gt; A0002 SELECT \"INBOX\"\nfetchmail: IMAP&lt; * 2 EXISTS\nfetchmail: IMAP&lt; * 1 RECENT\nfetchmail: IMAP&lt; * OK [UIDVALIDITY 1039260713] UID validity status\nfetchmail: IMAP&lt; * OK [UIDNEXT 23982] Predicted next UID\nfetchmail: IMAP&lt; * FLAGS (\\Answered \\Flagged \\Deleted \\Draft \\Seen)\nfetchmail: IMAP&lt; * OK [PERMANENTFLAGS \n               (\\* \\Answered \\Flagged \\Deleted \\Draft \\Seen)] \n               Permanent flags\nfetchmail: IMAP&lt; * OK [UNSEEN 2] first unseen in \/var\/spool\/mail\/esr\nfetchmail: IMAP&lt; A0002 OK [READ-WRITE] SELECT completed\nfetchmail: IMAP&gt; A0003 EXPUNGE\nfetchmail: IMAP&lt; A0003 OK Mailbox checkpointed, no messages expunged\nfetchmail: IMAP&gt; A0004 SEARCH UNSEEN\nfetchmail: IMAP&lt; * SEARCH 2\nfetchmail: IMAP&lt; A0004 OK SEARCH completed\n2 messages (1 seen) for esr at hurkle.thyrsus.com.\nfetchmail: IMAP&gt; A0005 FETCH 1:2 RFC822.SIZE\nfetchmail: IMAP&lt; * 1 FETCH (RFC822.SIZE 2545)\nfetchmail: IMAP&lt; * 2 FETCH (RFC822.SIZE 8328)\nfetchmail: IMAP&lt; A0005 OK FETCH completed\nskipping message esr@hurkle.thyrsus.com:1 (2545 octets) not flushed\nfetchmail: IMAP&gt; A0006 FETCH 2 RFC822.HEADER\nfetchmail: IMAP&lt; * 2 FETCH (RFC822.HEADER {1586}\nreading message esr@hurkle.thyrsus.com:2 of 2 (1586 header octets)\nfetchmail: SMTP&lt; 220 snark.thyrsus.com ESMTP Sendmail 8.12.5\/8.12.5; \n               Mon, 9 Dec\n2002 08:41:41 -0500<\/pre><pre class=\"programlisting\">fetchmail: SMTP&gt; EHLO localhost\nfetchmail: SMTP&lt; 250-snark.thyrsus.com \n               Hello localhost [127.0.0.1], pleased to meet you\nfetchmail: SMTP&lt; 250-ENHANCEDSTATUSCODES\nfetchmail: SMTP&lt; 250-8BITMIME\nfetchmail: SMTP&lt; 250-SIZE\nfetchmail: SMTP&gt; MAIL FROM:&lt;mutt-dev-owner@mutt.org&gt; SIZE=8328\nfetchmail: SMTP&lt; 250 2.1.0 &lt;mutt-dev-owner@mutt.org&gt;... Sender ok\nfetchmail: SMTP&gt; RCPT TO:&lt;esr@localhost&gt;\nfetchmail: SMTP&lt; 250 2.1.5 &lt;esr@localhost&gt;... Recipient ok\nfetchmail: SMTP&gt; DATA\nfetchmail: SMTP&lt; 354 Enter mail, end with \".\" on a line by itself\n#\nfetchmail: IMAP&lt; )\nfetchmail: IMAP&lt; A0006 OK FETCH completed\nfetchmail: IMAP&gt; A0007 FETCH 2 BODY.PEEK[TEXT]\nfetchmail: IMAP&lt; * 2 FETCH (BODY[TEXT] {6742}\n (6742 body octets) *********************.**************************.\n********************************.************************.***********\n**********.***********************.***************\nfetchmail: IMAP&lt; )\nfetchmail: IMAP&lt; A0007 OK FETCH completed\nfetchmail: SMTP&gt;. (EOM)\nfetchmail: SMTP&lt; 250 2.0.0 gB9ffWo08245 Message accepted for delivery\n flushed\nfetchmail: IMAP&gt; A0008 STORE 2 +FLAGS (\\Seen \\Deleted)\nfetchmail: IMAP&lt; * 2 FETCH (FLAGS (\\Recent \\Seen \\Deleted))\nfetchmail: IMAP&lt; A0008 OK STORE completed\nfetchmail: IMAP&gt; A0009 EXPUNGE\nfetchmail: IMAP&lt; * 2 EXPUNGE\nfetchmail: IMAP&lt; * 1 EXISTS\nfetchmail: IMAP&lt; * 0 RECENT\nfetchmail: IMAP&lt; A0009 OK Expunged 1 messages\nfetchmail: IMAP&gt; A0010 LOGOUT\nfetchmail: IMAP&lt; * BYE hurkle IMAP4rev1 server terminating connection\nfetchmail: IMAP&lt; A0010 OK LOGOUT completed\nfetchmail: 6.1.0 querying hurkle.thyrsus.com (protocol IMAP) \n               at Mon, 09 Dec 2002 08:41:42 -0500: poll completed\nfetchmail: SMTP&gt; QUIT\nfetchmail: SMTP&lt; 221 2.0.0 snark.thyrsus.com closing connection\nfetchmail: normal termination, status 0\n<\/pre><\/div>\n<p>\n  The\n  <code>-v<\/code> option makes what\n  <em>fetchmail<\/em> is doing discoverable (by letting you\nsee the protocol exchanges).  This is\n  <em>immensely<\/em>\n  useful.  I considered it so important that I wrote special code to\nmask account passwords out of\n  <code>-v<\/code> transaction dumps so\nthat they could be passed around and posted without anyone having to\nremember to edit sensitive information out of them.\n<\/p>\n<p>\n  This turned out to be a good call.  At least eight out of ten\nproblems reported get diagnosed within seconds of a knowledgeable\nperson's eyes seeing a session transcript.  There are several\nknowledgeable people on the fetchmail mailing list \u2014 in fact,\nbecause most bugs are easy to diagnose, I seldom have to\nhandle them myself.\n<\/p>\n<p>\n  Over the years,\n  <em>fetchmail<\/em> has\nacquired a reputation as a rather bulletproof program.  It can be\nmisconfigured, but it very seldom outright breaks.  Betting that this\nhas nothing to do with the fact that the exact circumstances of eight\nout of ten bugs are rapidly discoverable would not be smart.\n<\/p>\n<p>\n  We can learn from this example.  The lesson is this:  Don't\nlet your debugging tools be mere afterthoughts or treat them as\nthrowaways.  They are your windows into the code; don't just knock\ncrude holes in the walls, finish and glaze them.  If you plan to \nkeep the code maintained, you're always going to need to let light\ninto it.\n<\/p>\n<h4 id=\"id2909841\"><a href=\"#id2909841\">\u00a7<\/a>Case Study: GCC<\/h4>\n<p>\n  GCC, the GNU C compiler used on most modern Unixes, is perhaps\nan even better example of engineering for transparency.  GCC is\norganized as a sequence of processing stages knit together by a \ndriver program.  The stages are: preprocessor, parser, code \ngenerator, assembler, and linker.\n<\/p>\n<p>\n  Each of the first three stages takes in a readable textual\nformat and emits a readable textual format (the assembler has to emit\nand the linker to accept binary formats, pretty much by definition).\nWith various command-line options of the\ngcc(1)\ndriver, you can see not just the results after C preprocessing, after\nassembly generation, and after object code generation \u2014 but you\ncan also monitor the results of many intermediate steps in parsing and\ncode generation.\n<\/p>\n <blockquote>\n  <p>\n   This is exactly the structure of cc, the first (PDP-11) C\ncompiler.\n  <\/p><div>\n  \n   \u2013 Ken Thompson\n  \n <\/div><\/blockquote>\n<p>\n  There are many benefits of this organization.  One that is\nparticularly important for GCC is regression\ntesting.<sup><a href=\"#ftn.id2909913\" id=\"id2909913\">[60]<\/a><\/sup>\n  Because most\nof the various intermediate formats are textual, deviations from\nexpected results in a regression test are easily spotted and analyzed\nusing simple textual diff operations on the intermediate results;\nthere is no need for specialist dump-analysis tools that may well\nharbor their own bugs, and in any case would represent an additional\nmaintenance burden.\n<\/p>\n<p>\n  The design pattern to extract from this example is that the\ndriver program has monitoring switches that merely (but sufficiently)\nexpose the textual data flows among the components.  As with\n  <em>fetchmail's<\/em> <code>-v<\/code> option,\nthese options are not afterthoughts; they are designed in for\ndiscoverability.\n<\/p>\n<h4 id=\"id2909954\"><a href=\"#id2909954\">\u00a7<\/a>Case Study: kmail<\/h4>\n<p>\n  <em>kmail<\/em> is the GUI mailreader\ndistributed with the KDE environment.  The\n  <em>kmail<\/em> UI is tastefully and well designed,\nwith many good features including automatic display of enclosed images\nin a MIME\n  multipart and\nsupport for PGP key encryption\/decryption.  It is friendly to\nend-users \u2014 my beloved but nontechie wife uses and\nenjoys it.\n<\/p>\n<p>\n  Many mail user agents make one gesture in the direction of\ndiscoverability\n  by having a command that toggles display of all the mail headers,\nas opposed to a select few like From and Subject.  The UI of\n  <em>kmail<\/em> takes this a long step\nfurther.\n<\/p>\n<p>\n  A running\n  <em>kmail<\/em> displays status\nnotifications in a one-line subwindow at the bottom of its window, in\nsmall type over a steel-gray background clearly modeled on the\nNetscape\/Mozilla status bar.  When you open a mailbox, for example,\nthe status bar displays counts of total and unread messages.  The\nvisual presentation is unobtrusive; it is easy to ignore the\nnotifications, but also easy to focus on them if you want to.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 6.2. Screen shot of\n    <em>kmail<\/em>.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Screen shot of kmail.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  The\n  <em>kmail<\/em> GUI is good user-interface\ndesign. It's informative, but not distracting; it gets around the\nreason we adduce in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>\n  that the best policy for Unix tools\noperating normally is usually silence.  The authors showed excellent\ntaste in borrowing the look and feel of the browser status bar.\n<\/p>\n<p>\n  But the extent of the\n  <em>kmail<\/em> developers' tastefulness will not become clear until you have to\ntroubleshoot an installation that is having trouble sending mail.  If\nyou watch closely during the send, you will observe that each line of\nthe SMTP transaction with the remote mail transport is echoed into the\n  <em>kmail<\/em> status bar as it happens.\n<\/p>\n<p>\n  The\n  <em>kmail<\/em> developers neatly avoid a\ntrap that often makes GUI programs like\n  <em>kmail<\/em> a terrible pain in a troubleshooter's\nfundament.  Most design teams with kmail's objectives would have\nsuppressed those messages entirely, fearing that they would give Aunt\nTillie a touch of the vapors that would drive her back to the\nmeretricious pseudo-simplicity of a Windows box.\n<\/p>\n<p>\n  Instead, they designed for transparency \u2014 they made the\ntransaction messages show, but also made them visually easy to ignore.\nBy getting the presentation right, they managed to please both Aunt\nTillie and her geeky nephew Melvin who fixes her computer problems.\nThis was brilliant; it's a technique other GUI interfaces could and\nshould emulate.\n<\/p>\n<p>\n  Ultimately, of course, the visibility of those messages is good\nfor Aunt Tillie, because they mean Melvin is far less likely to throw\nup his hands in frustration while trying to solve her email\nproblems.\n<\/p>\n<p>\n  The lesson here is clear.  Dumbing down your UI is only the\nhalf-smart thing to do.  The really smart thing is to find a way \nto leave the details accessible, but make them unobtrusive.\n<\/p>\n<h4 id=\"id2910193\"><a href=\"#id2910193\">\u00a7<\/a>Case Study: SNG<\/h4>\n<p>\n  The program\n  <em>sng<\/em> translates between\nPNG\n  format and an\nall-text representation of it (SNG or Scriptable Network Graphics\nformat) that can be examined and modified with an ordinary text\neditor.  Run on a PNG file, it produces an SNG file; run on an SNG\nfile, it recovers the equivalent PNG.  The transformation is 100%\nfaithful and lossless in both directions.\n<\/p>\n<p>\n  In syntactic style, SNG resembles CSS (Cascading Style Sheets),\nanother language for controlling presentation of graphics; this makes\nat least a gesture in the direction of the Rule of Least\nSurprise. Here is a test example:\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 6.2. An SNG Example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n#SNG: This is a synthetic SNG test file\n# Our first test is a paletted (type 3) image.\nIHDR: {\n        width: 16;\n        height: 19;\n        bitdepth: 8;\n        using color: palette; \n        with interlace;\n}\n# Sample bit depth chunk\nsBIT: {\n  red: 8;\n  green: 8;\n  blue: 8;\n}\n# An example palette: three colors, one of which \n# we will render transparent \nPLTE: {\n   (0,     0, 255)\n   (255,   0,   0)\n   \"dark slate gray\",\n}<\/pre><pre class=\"programlisting\"># Suggested palette\nsPLT {\n   name: \"A random suggested palette\";\n   depth: 8;\n   (0,     0, 255), 255, 7;\n   (255,   0,   0), 255, 5;\n   ( 70,  70,  70), 255, 3;\n}\n# The viewer will actually use this...\nIMAGE: {\n   pixels base64\n2222222222222222\n2222222222222222\n0000001111100000\n0000011111110000\n0000111001111000\n0001110000111100\n0001110000111100\n0000110001111000\n0000000011110000\n0000000111100000\n0000001111000000\n0000001111000000\n0000000000000000\n0000000110000000\n0000001111000000\n0000001111000000\n0000000110000000\n2222222222222222\n2222222222222222\n}\ntEXt: {                                # Ordinary text chunk\n  keyword: \"Title\";\n  text: \"Sample SNG script\";\n}\n# Test file ends here\n<\/pre><\/div>\n<p>\n  The point of this tool is to enable users to edit various\nobscure PNG\n  chunk types\nthat are not necessarily supported by conventional graphics editors.\nRather than writing special-purpose code to grovel through the PNG\nbinary format, the user can simply flip an image into an all-text\nrepresentation, edit that, and massage it back.  Another potential\napplication is in making images amenable to version control; under\nmost version-control systems, text files are much easier to manage\nthan binary blobs, and diff operations on SNG representations actually\nhave some possibility of yielding useful information.\n<\/p>\n<p>\n  The gains here go beyond the time not spent writing\nspecial-purpose code for manipulating binary PNGs, however. The code\nof the\n  <em>sng<\/em> program itself is not especially\ntransparent, but it promotes transparency in larger systems of\nprograms by making the entire contents of PNGs\ndiscoverable.\n<\/p>\n<h4 id=\"id2910334\"><a href=\"#id2910334\">\u00a7<\/a>Case Study: The Terminfo Database<\/h4>\n<p>\n  The terminfo database is a collection of descriptions of\nvideo-display terminals.  Each entry describes the escape sequences \nthat perform various manipulations on the terminal screen, such as\ninserting or deleting lines, erasing from the cursor position to end\nof line or screen, or beginning and ending screen highlights such\nas reverse video, underline, or blink.\n<\/p>\n<p>\n  The terminfo database is primarily used by the\ncurses(3)\nlibraries.  These underlie the \u201croguelike\u201d interface\nstyle we discuss in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>, and some very widely used programs such\nas\nmutt(1),\nlynx(1),\nand\nslrn(1).\nThough the terminal emulators such as\nxterm(1)\nthat run on today's bitmapped displays all have capabilities that are\nminor variations on those of the ANSI X3.64 standard and the venerable\nVT100 terminal, there is still enough variation that hardwiring ANSI\ncapabilities into applications would be a bad idea.  Terminfo is also\nworth studying because problems that are logically similar to the one\nit addressed arise constantly in managing other kinds of peripheral\nhardware that doesn't have a standard way to report their own\ncapabilities.\n<\/p>\n<p>\n  The design of terminfo benefits from experience with an earlier capability format called termcap.  The database of\n  termcap descriptions lived in a textual format in one big file, <code>\/etc\/termcap<\/code>; though this format is now\n  obsolete, your Unix system almost certainly includes a copy.\n<\/p>\n<p>\n  Normally, the key used to look up your terminal type entry is\nthe environment variable\n  <code>TERM<\/code>, which for purposes of\nthis case study is set by magic.<sup><a href=\"#ftn.id2910448\" id=\"id2910448\">[61]<\/a><\/sup>\n  Applications that\nuse terminfo (or termcap) pay a small penalty in startup lag; when the\ncurses(3)\nlibrary initializes itself, it has to look up the entry corresponding\nto\n  <code>TERM<\/code> and load the entry into memory.\n<\/p>\n<p>\n  Experience with termcap showed that the startup penalty was\ndominated by the time required to parse the textual representation of\ncapabilities. Accordingly, terminfo entries are binary structure dumps\nthat can be marshaled and unmarshaled more quickly.  There is a master\ntextual format for the entire database, the terminfo capability file.\nThat file (or individual entries) can be compiled to binary form with\nthe terminfo compiler\ntic(1);\nbinary entries can be decompiled to the editable text format by\ninfocmp(1).\n<\/p>\n<p>\n  The design superficially contradicts the advice we gave in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  against binary\ncaches, but this is actually the extreme case in which that's a good\ntactic.  Edits to the text masters are very rare \u2014 in fact,\nUnixes normally ship with the terminfo database precompiled and the\ntext master serving primarily as documentation.  Thus, the\nsynchronization and inconsistency problems that would normally\nmilitate against this approach almost never arise.\n<\/p>\n<p>\n  The designers of terminfo could have optimized for speed\nin a second way.  The entire database of binary entries could have\nbeen put in some kind of big opaque database file.  What they actually did\ninstead was more clever and more in the Unix spirit.  Terminfo entries\nlive in a directory hierarchy, usually on modern Unixes under\n  <code>\/usr\/share\/terminfo<\/code>.  Consult the\nterminfo(5)\nman page to find the location on your system.\n<\/p>\n<p>\n  If you look in the terminfo directory, you'll see subdirectories\nnamed by single printable characters.  Under each of these are the\nentries for each terminal type that has a name beginning with that\nletter.  The goal of this organization was to avoid having to do a\nlinear search of a very large directory; under more modern Unix\nfile systems, which represent directories with B-trees or other\nstructures optimized for fast lookup, the subdirectories won't be\nnecessary.\n<\/p>\n <blockquote>\n  <p>\n   I found that even on a fairly modern Unix, splitting a big directory up\ninto subdirectories can improve performance substantially.  It was tens\nof thousands of files, an authorized-user database for a big educational\ninstitution, on a late-model DEC Alpha running DEC's Unix.  (Subdirectories\nnamed by first and last letter of name \u2014 e.g., \"johnson\" would be in\ndirectory \"j_n\" \u2014 worked best of the schemes we tested.  Using the first\ntwo letters wasn't nearly as good, because there were a lot of\nsystematically-generated names which differed only toward the end.)\nThis may just say that sophisticated directory indexing is still not as\ncommon as it should be... but even so, that makes an organization which\nworks well without it more portable than one which requires it.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  Thus, the cost of opening a terminfo entry is two file system\nlookups and a file open.  But since mining the same entry from one\nbig database would have required a lookup and open for the database,\nthe incremental cost for terminfo's organization is at most one\nfile system lookup.  Actually, it's less than that; it's the cost\ndifference between one file system lookup and whatever retrieval method\nthe one big database would have used.  This is probably marginal, and\nquite tolerable once per application at startup time.\n<\/p>\n<p>\n  Terminfo uses the file system itself as a simple hierarchical\ndatabase.  This is a superb bit of constructive laziness, obeying the\nRule of Economy and the Rule of Transparency.  It means that all the\nordinary tools for navigating, examining and modifying the file system\ncan be used to navigate, examine, and modify the terminfo database; no\nspecial ones (other than\ntic(1)\nand\ninfocmp(1)\nfor packing and unpacking the individual records) need to be written\nand debugged.  It also means that work on speeding up database access\nwould be work on speeding up the file system itself, tuning that would\nbenefit many more applications than just users of\ncurses(3).\n<\/p>\n<p>\n  There is one additional advantage of this organization that\ndoesn't come up in the terminfo case; you get to use Unix's\npermissions mechanism rather than having to invent your own\naccess-control layer with its own bugs.  This falls out as a\nconsequence of adopting the \u201ceverything is a file\u201d\nphilosophy of Unix rather than trying to fight it.\n<\/p>\n<p>\n  The terminfo directory layout is rather space-inefficient\non most Unix file systems.  The entries are usually between 400 and\n1400 bytes long, but file systems normally allocate a minimum of 4K \nfor every nonempty disk file. The designers accepted this cost\nfor the same reason they chose a packed binary format, to\ncut the startup latency of terminfo-using programs to a minimum.  Disk \ncapacity for constant price has exploded over a thousandfold since,\ntending to vindicate that decision.\n<\/p>\n<p>\n  The contrast with the formats used by the Microsoft Windows\n  registry files is instructive.  Registries are property databases used\nby both Windows itself and applications.  Each registry lives in one\nbig file.  Registries contain a mix of text and binary data that\nrequires specialized editing tools.  The one-big-file approach leads,\namong other things, to the notorious \u2018registry creep\u2019\nphenomenon; average access time rises without bound as new entries are\nadded.  Because there is no standard API for editing the registry\nprovided by the system, applications use ad-hoc code to edit it\nthemselves, making it notoriously subject to corruption that can lock\nup the entire system.\n<\/p>\n<p>\n  Using the Unix file system as a database is a tactic other\napplications with simple database requirements might do well to emulate.\nGood reasons not to do it are more likely to have to do with the\ndatabase keys not naturally looking like filenames than they are with\nany performance problems.  In any case, it's the sort of good fast\nhack that can be very useful in prototyping.\n<\/p>\n<h4 id=\"id2914115\"><a href=\"#id2914115\">\u00a7<\/a>Case Study: Freeciv Data Files<\/h4>\n<p>\n  Freeciv is an open-source strategy game inspired by Sid Meier's\nclassic\n  <em>Civilization II<\/em>.  In it, each player\nbegins with a wandering band of neolithic nomads and builds a\ncivilization. Player civilizations may explore and colonize the world,\nfight wars, engage in trade, and research technological advances.\nSome players may actually be artificial intelligences; solitaire play\nagainst these can be challenging.  One wins either by conquering the\nworld or by being the first player to reach a technology level sufficient\nto get a starship to Alpha Centauri.  Sources and documentation are\navailable at the\n  <a href=\"http:\/\/www.freeciv.org\/\">\n   project\nsite<\/a>.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 6.3. Main window of a Freeciv game.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Main window of a Freeciv game.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  In\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  we'll exhibit the Freeciv strategy game as an example of client-server\npartitioning, with the server maintaining shared state and the client\nconcentrating on GUI presentation.  But this game has another notable\narchitectural feature; much of the game's fixed data, rather than\nbeing wired into the server code, is expressed in a property registry\nread in by the game server at startup time.\n<\/p>\n<p>\n  The game's registry files are written in a textual data-file\nformat that assembles text strings (with associated text and numeric\nproperties) into various internal lists of important data (such as\nnations and unit types) in the game server.  The minilanguage has an\ninclude directive, so game data can be broken up into semantic units\n(different files) that are each separately editable.  This design\nchoice has been carried through to such an extent that it's possible\nto define new nations and new unit types simply by creating new\ndeclarations in the data files, without touching the server code at\nall.\n<\/p>\n<p>\n  The Freeciv server's startup parsing has an interesting feature\nthat creates something of a conflict between two of Unix's design\nrules, and is therefore worth closer examination.  The server ignores\nproperty names it doesn't know how to use. This makes it possible to\ndeclare properties that the server doesn't yet use without breaking\nthe startup parsing. It means that development of the game data\n(policy) and the server engine (mechanism) can be cleanly separated.\nOn the other hand, it also means startup parsing won't catch simple\nmisspellings of attribute names.  This quiet failure seems to violate\nthe Rule of Repair.\n<\/p>\n<p>\n  To resolve this conflict, notice that it's the server's job to\n  <em>use<\/em>\n  the registry data, but the task of carefully\nerror-checking that data could be handed off to another program to be run\nby human editors each time the registry is modified.  One Unix\nsolution would be a separate auditing program that analyzes either\na machine-readable specification of the ruleset format or the source\nof the server code to determine the set of properties it uses, parses\nthe Freeciv registry to determine the set of properties it provides,\nand prepares a difference report.<sup><a href=\"#ftn.id2914267\" id=\"id2914267\">[62]<\/a><\/sup><\/p>\n<p>\n  The aggregate of all Freeciv data files is functionally similar\nto a Windows registry, and even uses a syntax resembling the textual\nportions of registries.  But the creep and corruption problems we\nnoted with the Windows registry don't crop up here because no program\n(either within or outside the Freeciv suite)\n  <em>writes<\/em>\n  to these files.  It's a read-only\nregistry edited only by the game's maintainers.\n<\/p>\n<p>\n  The performance impact of data-file parsing is minimized because\nfor each file the operation is performed only once, at either \nclient or server startup time.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2909913\" id=\"ftn.id2909913\">[60]<\/a><\/sup>\n  Regression testing is a method for detecting\nbugs introduced as software is modified.  It consists of periodically\nchecking the output of the changing software for some fixed test input\nagainst a snapshot of output captured at an earlier stage of the process\nand known (or assumed) to be correct.\n<\/p><p><sup><a href=\"#id2910448\" id=\"ftn.id2910448\">[61]<\/a><\/sup>\n  Actually,\n  <code>TERM<\/code> is set by the system at login time.  For actual\nterminals on serial lines, the mapping from tty lines to\n  <code>TERM<\/code> values is set from a system configuration file at\nboot time; the details vary among Unixes.  Terminal emulators like\nxterm(1)\nset this variable themselves.\n<\/p><p><sup><a href=\"#id2914267\" id=\"ftn.id2914267\">[62]<\/a><\/sup>\n  The ur-ancestor of \nsuch validator programs under Unix was\n  <em>lint<\/em>, a validator for C code separate from\nthe C compiler. Though GCC has absorbed its functions, old Unix hands\nare still apt to refer to the process of running a validator as\n\u2018linting\u2019, and the name survives in utilities such as\n  <em>xmllint<\/em>.\n<\/p><\/div>\n\n<h3 id=\"ch06s02\"><a href=\"#ch06s02\">\u00a7<\/a>Designing for Transparency and Discoverability<\/h3>\n<p>\n  To design for transparency and discoverability, you need to\napply every tactic for keeping your code simple, and also concentrate\non the ways in which your code is a communication to other human\nbeings.  The first questions to ask, after \u201cWill this design\nwork?\u201d are \u201cWill it be readable to other people?  Is it\nelegant?\u201d We hope it is clear by now that these questions are\nnot fluff and that elegance is not a luxury.  These qualities in the\nhuman reaction to software are essential for reducing its bugginess\nand increasing its long-term maintainability.\n<\/p>\n<h4 id=\"zen_of_transparency\"><a href=\"#zen_of_transparency\">\u00a7<\/a>The Zen of Transparency<\/h4>\n<p>\n  One\npattern that emerges from the examples we've examined so far in this\nchapter is this: If you want transparent code, the most effective\nroute is simply not to layer too much abstraction over what you\nare manipulating with the code.\n<\/p>\n<p>\n  In\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  's\nsection on the value of detachment, our advice was to abstract and\nsimplify and generalize, to try and detach from the particular,\naccidental conditions under which a design problem was posed.  The\nadvice to abstract does not actually contradict the advice against\nexcessive abstractions we're developing here, because there is a\ndifference between getting free of assumptions and forgetting the\nproblem you're trying to solve.  This is part of what we were driving\nat when we developed the idea that glue layers need to be kept\nthin.\n<\/p>\n<p>\n  One of the main lessons of Zen is that we ordinarily see the\nworld through a haze of preconceptions and fixed ideas that proceed\nfrom our desires.  To achieve enlightenment, we must follow the Zen\nteaching not merely to let go of desire and attachment, but to\nexperience reality exactly as it is \u2014 without the preconceptions\nand the fixed ideas getting in the way.\n<\/p>\n<p>\n  This is excellent pragmatic advice for software designers.  It's\npart of what's implicit in the classic Unix advice to be minimalist.\nSoftware designers are clever people who form ideas (abstractions)\nabout the application domains they deal with.  They organize the\nsoftware they write around those ideas.  Then, when debugging, they\noften find they have great trouble seeing through those ideas to what\nis actually going on.\n<\/p>\n<p>\n  Any Zen master would recognize this problem instantly, yell\n\u201cThree pounds of flax!\u201d, and probably clout the student a\ngood one.<sup><a href=\"#ftn.id2914429\" id=\"id2914429\">[63]<\/a><\/sup>\n  Consciously\ndesigning for transparency is a slightly less mystical way of\naddressing it.\n<\/p>\n<p>\n  In\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  we\ncriticized object-oriented programming in terms likely to prove a bit\nshocking to programmers who were raised on the 1990s gospel of\nOO. Object-oriented design doesn't have\nto be over-complicated design, but we've observed that too often it\nis.  Too many OO designs are spaghetti-like tangles of is-a and has-a\nrelationships, or feature thick layers of glue in which many of the\nobjects seem to exist simply to hold places in a steep-sided pyramid\nof abstractions.  Such designs are the opposite of transparent; they\nare (notoriously) opaque and difficult to debug.\n<\/p>\n<p>\n  As we've previously noted, Unix programmers are the original\nzealots about modularity, but tend to go about it in a quieter way.\nKeeping glue layers thin is part of it; more generally, our tradition\nteaches us to build lower, hugging the ground with algorithms and\nstructures that are designed to be simple and transparent.\n<\/p>\n<p>\n  As with Zen art, the simplicity of good Unix code depends on\nexacting self-discipline and a high level of craft, neither of which\nare necessarily apparent on casual inspection.  Transparency is hard\nwork, but worth the effort for more than merely artistic reasons.\nUnlike Zen art, software requires debugging \u2014 and usually needs\ncontinuing maintenance, forward-porting, and adaptation throughout\nits lifetime.  Transparency is therefore more than an esthetic\ntriumph; it is a victory that will be reflected in lower costs\nthroughout the software's life cycle.\n<\/p>\n<h4 id=\"id2914509\"><a href=\"#id2914509\">\u00a7<\/a>Coding for Transparency and Discoverability<\/h4>\n<p>\n  Transparency and\ndiscoverability,\nlike modularity, are primarily properties of designs, not code.  It is\nnot sufficient to get right the low-level elements of style, such as\nindenting code in a clear and consistent way or having good\nvariable-naming conventions.  These qualities have much more to do\nwith code properties that are less obvious to inspection.  Here are a\nfew to think about:\n<\/p>\n <div><ul><li>\n    <p>\n     What is the maximum static depth of your procedure-call\nhierarchy?  That is, leaving out recursions, how many levels of call\nmight a human have to model mentally to understand the operation of\nthe code?  Hint: If it's more than four, beware.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Does the code have invariant properties<sup><a href=\"#ftn.id2914557\" id=\"id2914557\">[64]<\/a><\/sup>\n     that are both strong and visible?  Invariant\nproperties help human beings reason about code and detect problem\ncases.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Are the function calls in your APIs individually orthogonal, or\ndo they have too many magic flags and mode bits that have a single call doing\nmultiple tasks?  Avoiding mode flags entirely can lead to a cluttered\nAPI with too many nigh-identical functions, but the obverse error (lots of\neasily-forgotten and confusable mode flags) is even more common.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Are there a handful of prominent data structures or a single\nglobal scoreboard that captures the high-level state of the system?\nIs this state easy to visualize and inspect, or is it diffused among\nmany individual global variables or objects that are hard to\nfind?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Is there a clean, one-to-one mapping between data structures\nor classes in your program and the entities in the world that they \nrepresent?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Is it easy to find the portion of the code responsible for any\ngiven function? How much attention have you paid to the readability\nnot just of individual functions and modules but of the whole codebase?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Does the code proliferate special cases or avoid them?  Every\nspecial case could interact with every other special case; all those\npotential collisions are bugs waiting to happen. But even more\nimportantly, special cases make the code harder to understand.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     How many magic numbers (unexplained constants) does the code have in\nit?  Is it easy to discover the implementation's limits (such as\ncritical buffer sizes) by inspection?\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  It's best for code to be simple.  But if it answers these sorts\nof questions well, it can be very complex without putting an\nimpossible cognitive burden on a human maintainer.\n<\/p>\n<p>\n  The reader might find it instructive to compare these with our\nchecklist questions about modularity in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>.\n<\/p>\n<h4 id=\"id2914680\"><a href=\"#id2914680\">\u00a7<\/a>Transparency and Avoiding Overprotectiveness<\/h4>\n<p>\n  Close kin to the programmer tendency to build overelaborate\ncastles of abstractions is a tendency to overprotect others from the\nlow-level details.  While it's not bad practice to hide those details\nin the program's normal mode of operation\n(\n  <em>fetchmail<\/em> 's\n  <code>-v<\/code> switch is\noff by default), they should be discoverable.  There's an important\ndifference between hiding them and making them inaccessible.\n<\/p>\n<p>\n  Programs that\n  <em>cannot<\/em>\n  reveal what they are\ndoing make troubleshooting far more difficult. Thus, experienced Unix\nusers actually take the presence of debugging and instrumentation\nswitches as a good sign, and their absence as possibly a bad one.\nAbsence suggests an inexperienced or careless developer; presence\nsuggests one with enough wisdom to follow the Rule of\nTransparency.\n<\/p>\n<p>\n  The temptation to overprotect is especially strong in GUI\napplications targeted for end users, like mail readers.  One reason \nUnix developers have been cool toward GUI interfaces is that, in\ntheir designers' haste to make them \u2018user-friendly\u2019 each one\noften becomes frustratingly opaque to anyone who has to solve user problems\n\u2014 or, indeed, interact with it anywhere outside the narrow\nrange predicted by the user-interface designer.\n<\/p>\n<p>\n  Worse, programs that are opaque about what they are doing tend\nto have a lot of assumptions baked into them, and to be frustrating or\nbrittle or both in any use case not anticipated by the designer.\nTools that look glossy but shatter under stress are not good long-term\nvalue.\n<\/p>\n<p>\n  Unix tradition pushes for programs that are flexible for a \nbroader range of uses and troubleshooting situations, including the\nability to present as much state and activity information to the user\nas the user indicates he is willing to handle.  This is good for\ntroubleshooting; it is also good for growing smarter, more\nself-reliant users.\n<\/p>\n<h4 id=\"id2914758\"><a href=\"#id2914758\">\u00a7<\/a>Transparency and Editable Representations<\/h4>\n<p>\n  Another theme that emerges from these examples is the value of\nprograms that flip a problem out of a domain in which transparency is\nhard into one in which it is easy.  Audacity,\nsng(1)\nand the\ntic(1)\/infocmp(1)\npair all have this property.  The objects they manipulate are not\nreadily conformable to the hand and eye; audio files are not visual\nobjects, and although images expressed in\nPNG\n  format are visual,\nthe complexities of PNG annotation chunks are not.  All three\napplications turn manipulation of their binary file formats into a\nproblem to which human beings can more readily apply intuition and\ncompetences gained from everyday experience.\n<\/p>\n<p>\n  A rule all these examples follow is that they degrade the\nrepresentation as little as possible \u2014 in fact, they translate\nit reversibly and losslessly.  This property is very important, and\nworth implementing even if there is no obvious application demand for\nthat kind of 100% fidelity.  It gives potential users confidence that\nthey can experiment without degrading their data.\n<\/p>\n<p>\n  All the advantages of textual data-file formats that we\ndiscussed in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  also apply to the textual formats that\nsng(1),\ninfocmp(1)\nand their kin generate.  One important application for\nsng(1)\nis robotic generation of\nPNG\n  image annotations by\nscripts \u2014 because\nsng(1)\nexists, such scripts are easier to write.\n<\/p>\n<p>\n  Whenever you face a design problem that involves editing some\nkind of complex binary object, the Unix tradition encourages asking\nfirst off whether you can write a tool analogous to\nsng(1)\nor the\ntic(1)\/infocmp(1)\npair that can do a lossless mapping to an editable textual format and\nback. There is no established term for programs of this kind, but\nwe'll call them\n  <em>textualizers<\/em>.\n<\/p>\n<p>\n  If the binary object is dynamically generated or very large,\nthen it may not be practical or possible to capture all the state with\na textualizer.  In that case, the equivalent task is to write a\nbrowser. The paradigm example is\nfsdb(1),\nthe file-system debugger supported under various Unixes; there is a\nLinux equivalent called\ndebugfs(1).\nThe\npsql(1)\nused to browse PostgreSQL databases, and the\nsmbclient(1)\nprogram that can be used to query Windows file shares on a\nSAMBA-equipped Linux machine, are two more.  All five are simple CLI\nprograms that could be driven by scripts and test harnesses.\n<\/p>\n<p>\n  Writing a textualizer or browser is a valuable exercise for at\nleast four reasons:\n<\/p>\n <div><ul><li>\n    <p>\n     <em>You gain an excellent learning experience.<\/em>\n     There may be other ways that are as good to learn about the structure\nof the object, but none that are obviously better.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      You gain the ability to dump the contents of the structure for\ninspection and debugging.\n<\/em>\n     Because such a tool makes dumping\neasy, you'll do it more. You'll get more information, probably leading\nto more insight.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      You gain the ability to easily generate test loads and unusual \ncases.\n<\/em>\n     This means you are more likely to probe the\nodd corners of the object's state space \u2014 and to break the\nassociated software, so you can fix it before your users break it.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>You gain code you may be able to reuse.<\/em>\n     If you're\ncareful about how you write the browser\/textualizer and keep the CLI\ninterpreter properly separated from the marshaling\/unmarshaling library,\nyou may find you have code that can be reused for your actual\napplication.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  After you've done this, you may well discover that it's possible\nto apply the \u201cseparated engine and interface\u201d pattern\n(see\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>) using\nyour textualizer\/debugger as the engine.  All the usual benefits of\nthis pattern will apply.\n<\/p>\n <blockquote>\n  <p>\n   It is desirable, although often difficult, for a textualizer to\nbe able to read and write even a damaged binary object.  For one\nthing, it lets you generate damaged test cases to stress-test\nsoftware; for another, it can make emergency repairs a whole lot\neasier.  It may be hard to handle cases in which the\n   <em>structure<\/em>\n   of the object is messed up, but at\nleast you should handle cases in which the\n   <em>content<\/em>\n   of the structure is nonsense, e.g., by showing nonsense values in hex\nand converting the hex back to the values.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<h4 id=\"id2915107\"><a href=\"#id2915107\">\u00a7<\/a>Transparency, Fault Diagnosis, and Fault Recovery<\/h4>\n<p>\n  Yet another benefit of transparency, related to ease of\ndebugging, is that transparent systems are easier to perform recovery\nactions on after a bug bites \u2014 and, often, more resistant to\ndamage from bugs in the first place.\n<\/p>\n<p>\n  In comparing the terminfo database with Windows registries\nwe noted that registries are notoriously subject to being corrupted\nby buggy application code.  This can make the entire system unusable.  \nEven if it doesn't, recovery can be difficult if the corruption\nconfuses the specialized registry-editing tools.\n<\/p>\n<p>\n  Our Unix case studies illustrate ways that designing for\ntransparency can prevent this class of problem.  Because the terminfo\ndatabase is not one big file, botching one terminfo entry does not\nmake the whole terminfo data set unusable.  Fully textual one-big-file\nformats like termcap are usually parsed with methods which (unlike\nblock reads of binary structure dumps) can recover from single-point\nerrors.  Syntax errors in an SNG file can be corrected by hand without\nrequiring specialized editors that might refuse to load a damaged PNG\nimage.\n<\/p>\n<p>\n  Going back to the\n  <em>kmail<\/em> case study,\nthat program makes fault diagnosis easier because it obeys the Rule of\nRepair: SMTP failures are noisy, usefully so.\nYou don't have to decode a layer of obfuscatory messages\ngenerated by\n  <em>kmail<\/em> itself to see\nwhat the interaction with the SMTP server looks like.  All you have to\ndo is look in the right place, because\n  <em>kmail<\/em> is being transparent and not throwing\naway information about the error state.  (It helps that SMTP itself is\ntextual and includes human-readable status messages in its\ntransactions.)\n<\/p>\n<p>\n  Discoverability tools like textualizers and browsers also make\nfault diagnosis easier.  We've already touched on one reason: they\nmake inspecting the state of the system easier.  But there is another\neffect at work as well; textualized versions of data tend to have\nuseful redundancies (such as using whitespace for visual separation as\nwell as explicit delimiters for parsing). These are present to make\nthem easier to read for humans, but also have the effect of making them\nmore resistant to being irreparably trashed by point failures.  A\ncorrupted chunk in a PNG file is seldom recoverable, but the human\ncapacity for pattern recognition and reasoning from context might be\nable to repair the equivalent SNG form.\n<\/p>\n<p>\n  Over and over again, the Rule of Robustness is clear.\nSimplicity plus transparency lowers costs, reduces everybody's stress,\nand frees people to concentrate on new problems rather than cleaning\nup after old mistakes.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2914429\" id=\"ftn.id2914429\">[63]<\/a><\/sup>\n  See the koan called\n  <em>Tozan's Three Pounds<\/em>\n  in the\n  <em>Gateless Gate<\/em> <sup><a href=\"#Mumon\" title=\"[Mumon]\">[Mumon]<\/a><\/sup>.\n<\/p><p><sup><a href=\"#id2914557\" id=\"ftn.id2914557\">[64]<\/a><\/sup>\n  An\ninvariant is a property of a software design that is preserved by\nevery operation in it.  For example, in most databases it is an\ninvariant that no two records may have the same key. In a\nC\n  \n  program that\ncorrectly manipulates strings, every string buffer must contain a\nterminating NUL byte on exit from each string function.  In an\ninventory system, no parts count can hold a number less than\nzero.\n<\/p><\/div>\n\n<h3 id=\"ch06s03\"><a href=\"#ch06s03\">\u00a7<\/a>Designing for Maintainability<\/h3>\n<p>\n  Software is maintainable to the extent that people who are not\nits author can successfully understand and modify it. Maintainability\ndemands more than code that works; it demands code that follows\nthe Rule of Clarity and communicates successfully to human beings\nas well as the computer.\n<\/p>\n<p>\n  Unix programmers have a lot of implicit knowledge available to\nthem about what makes for maintainable code, because Unix hosts source\ncode that goes back decades.  For reasons we'll discuss in\n  <a href=\"#portabilitychapter\" title=\"Chapter 17. Portability\">\n   Chapter 17<\/a>, Unix programmers learn a tendency to\nscrap and rebuild rather than patching grubby code (see Rob Pike's\nmeditation on this subject in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>).  Thus, any sources that have survived\nmore than a decade of evolutionary pressure have been selected for\nmaintainability. These old, successful, well-established projects with\nmaintainable code are the community's models for practice.\n<\/p>\n<p>\n  A question Unix programmers \u2014 and especially Unix\nprogrammers in the open-source world \u2014 learn to ask about tools\nthey are evaluating for use is: \u201cIs this code live, dormant, or\ndead?\u201d  Live code has an active developer community attached\nto it.  Dormant code has often become dormant because the pain of\nmaintaining it exceeded its utility to its originators.  Dead code has\nbeen dormant for so long that it would be easier to reimplement an\nequivalent from scratch.  If you want your code to live, investing\neffort to make it maintainable (and therefore attractive to future\nmaintainers) will be one of the most effective ways you can spend your\ntime.\n<\/p>\n<p>\n  Code that is designed to be both transparent and\ndiscoverable\n  has gone a long way toward being maintainable.  But there are other\npractices we can observe in the model projects in this chapter that\nare worth emulating.\n<\/p>\n<p>\n  One very important practice is an application of the Rule of\nClarity: choosing simple algorithms.  In\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>\n  we quoted Ken Thompson: \u201cWhen in\ndoubt, use brute force\u201d.  Thompson\n  understood the full cost of complicated\nalgorithms \u2014 not just that they're more bug-prone when initially\nimplemented, but that they're harder for maintainers down the line to\nunderstand.\n<\/p>\n<p>\n  Another important practice is the inclusion of hacker's guides.\nIt has always been highly approved behavior for source code\ndistributions to include guide documents informally describing the key\ndata structures and algorithms in the code. In fact, Unix programmers\nhave often been better about producing hacker's guides than they are\nabout writing end-user documentation.\n<\/p>\n<p>\n  The open-source community has seized on and elaborated this\ncustom.  Besides being advice to future maintainers, hacker's guides\nfor open-source projects are also designed to make it easy for casual\ncontributors to add features or fix bugs.  The Design Notes file\nshipped with\n  <em>fetchmail<\/em> is representative.  The Linux\n  kernel sources include literally dozens\nof these.\n<\/p>\n<p>\n  In\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>\n  we'll\ndescribe conventions that Unix developers have evolved for making\nsource code distributions easy to examine and easy to build running\ncode from. These practices, too, promote maintainability.\n<\/p>\n\n<h3 id=\"multiprogramchapter\"><a href=\"#multiprogramchapter\">\u00a7<\/a>Chapter 7. Multiprogramming<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   If we believe in data structures, we must believe in independent\n(hence simultaneous) processing. For why else would we collect items\nwithin a structure?  Why do we tolerate languages that give us the one\nwithout the other?\n  <\/p><div>\n  \n   \u2013 Alan Perlis\nEpigrams in Programming, in ACM SIGPLAN (Vol 17 #9, 1982)\n  \n <\/div><\/blockquote>\n<p>\n  The most characteristic program-modularization technique of Unix\nis splitting large programs into multiple cooperating processes.  This\nhas usually been called \u2018multiprocessing\u2019 in the Unix world,\nbut in this book we revive the older term\n\u2018multiprogramming\u2019 to avoid confusion with\nmultiprocessor hardware implementations.\n<\/p>\n<p>\n  Multiprogramming is a particularly murky area of design, one in\nwhich there are few guidelines to good practice.  Many programmers\nwith excellent judgment about how to break up code into subroutines\nnevertheless wind up writing whole applications as monster\nsingle-process monoliths that founder on their own internal\ncomplexity.\n<\/p>\n<p>\n  The Unix style of design applies the do-one-thing-well approach\nat the level of cooperating programs as well as cooperating routines\nwithin a program, emphasizing small programs connected by well-defined\ninterprocess communication or by shared files. Accordingly, the Unix operating\nsystem encourages us to break our programs into simpler subprocesses,\nand to concentrate on the interfaces between these subprocesses. It\ndoes this in at least three fundamental ways:\n<\/p>\n <div><ul><li>\n    <p>\n     by making process-spawning cheap;\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     by providing methods (shellouts, I\/O redirection, pipes,\nmessage-passing, and sockets) that make it relatively easy for\nprocesses to communicate;\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     by encouraging the use of simple, transparent, textual\ndata formats\n     that can be passed through\npipes and sockets.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  Inexpensive process-spawning and easy process control are\ncritical enablers for the Unix style of programming. On an operating\nsystem such as VAX VMS, where starting processes is expensive\nand slow and requires special privileges, one must build monster\nmonoliths because one has no choice. Fortunately the trend in the Unix\nfamily has been toward lower\nfork(2)\noverhead rather than higher. Linux, in particular, is famously efficient\nthis way, with a process-spawn faster than thread-spawning on many\nother operating systems.<sup><a href=\"#ftn.id2918458\" id=\"id2918458\">[65]<\/a><\/sup><\/p>\n<p>\n  Historically, many Unix programmers have been encouraged to\nthink in terms of multiple cooperating processes by experience with\nshell programming.  Shell makes it relatively easy to set up groups of\nmultiple processes connected by\npipes, running either\nin background or foreground or a mix of the two.\n<\/p>\n<p>\n  In the remainder of this chapter, we'll look at the\nimplications of cheap process-spawning and discuss how and when to\napply pipes, sockets, and other interprocess communication (IPC)\nmethods to partition your design into cooperating processes. (In the\nnext chapter, we'll apply the same separation-of-functions philosophy\nto interface design.)\n<\/p>\n<p>\n  While the benefit of breaking programs up into cooperating\nprocesses is a reduction in global complexity, the cost is that we\nhave to pay more attention to the design of the protocols which are\nused to pass information and commands between processes.  (In software\nsystems of all kinds, bugs collect at interfaces.)\n<\/p>\n<p>\n  In\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  we\nlooked at the lower level of this design problem \u2014 how to lay\nout application protocols that are transparent, flexible and\nextensible.  But there is a second, higher level to the problem which\nwe blithely ignored.  That is the problem of designing state machines\nfor each side of the communication.\n<\/p>\n<p>\n  It is not hard to apply good style to the syntax of application\nprotocols, given models like SMTP or BEEP or XML-RPC.  The real\nchallenge is not protocol syntax but protocol\n  <em>logic<\/em>\n  \u2014designing a protocol that is both\nsufficiently expressive and deadlock-free. Almost as importantly, the\nprotocol has to be\n  <em>seen<\/em>\n  to be expressive and\ndeadlock-free; human beings attempting to model the behavior of the\ncommunicating programs in their heads and verify its correctness must\nbe able to do so.\n<\/p>\n<p>\n  In our discussion, therefore, we will focus on the kinds of\nprotocol logic one naturally uses with each kind of interprocess\ncommunication.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2918458\" id=\"ftn.id2918458\">[65]<\/a><\/sup>\n  See, for example, the results\nquoted in\n  <em>Improving Context Switching Performance of Idle Tasks under Linux<\/em><sup><a href=\"#Appleton\" title=\"[Appleton]\">[Appleton]<\/a><\/sup>.\n<\/p><\/div>\n\n<h3 id=\"ch07s01\"><a href=\"#ch07s01\">\u00a7<\/a>Separating Complexity Control from Performance Tuning<\/h3>\n<p>\n  First, though, we need to dispose of a few red herrings. Our\ndiscussion is\n  <em>not<\/em>\n  going to be about using\nconcurrency to improve performance. Putting that concern before\ndeveloping a clean architecture that minimizes global complexity is\npremature optimization,\nthe root of all evil (see\n  <a href=\"#optimizationchapter\" title=\"Chapter 12. Optimization\">\n   Chapter 12<\/a>\n  for further discussion).\n<\/p>\n<p>\n  A closely related red herring is threads (that is, multiple\nconcurrent processes sharing the same memory-address space). Threading is a\nperformance hack.  To avoid a long diversion here, we'll\nexamine threads in more detail at the end of this chapter; the\nsummary is that they do not reduce global complexity but rather\n  <em>increase<\/em>\n  it, and should therefore be avoided \nsave under dire necessity.\n<\/p>\n<p>\n  Respecting the Rule of Modularity, on the other hand, is\n  <em>not<\/em>\n  a red herring; doing so can make your\nprograms \u2014 and your life \u2014 simpler. All the reasons for\nprocess partitioning are continuous with the reasons for module\npartitioning that we developed in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>.\n<\/p>\n<p>\n  Another important reason for breaking up programs into\ncooperating processes is for better security.  Under Unix, programs\nthat must be run by ordinary users, but must have write access to\nsecurity-critical system resources, get that access through a feature\ncalled the\n  <em>setuid bit<\/em>.<sup><a href=\"#ftn.id2915417\" id=\"id2915417\">[66]<\/a><\/sup>\n  Executable files are the smallest unit\nof code that can hold a setuid bit; thus, every line of code in a\nsetuid executable must be trusted.  (Well-written setuid programs,\nhowever, take all necessary privileged actions first and then drop\ntheir privileges back to user level for the remainder of their\nexistence.)\n<\/p>\n<p>\n  Usually a setuid program only needs its privileges for one or a\nsmall handful of operations.  It is often possible to break up such a\nprogram into cooperating processes, a smaller one that needs setuid\nand a larger one that does not.  When we can do this, only the code in\nthe smaller program has to be trusted.  It is in significant part\nbecause this kind of partitioning and delegation is possible that Unix\nhas a better security track record<sup><a href=\"#ftn.id2915447\" id=\"id2915447\">[67]<\/a><\/sup>\n  than its competitors.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2915417\" id=\"ftn.id2915417\">[66]<\/a><\/sup>\n  A setuid\nprogram runs not with the privileges of the user calling it, but with\nthe privileges of the owner of the executable. This feature can be\nused to give restricted, program-controlled access to things like the\npassword file that nonadministrators should not be allowed to modify\ndirectly.\n<\/p><p><sup><a href=\"#id2915447\" id=\"ftn.id2915447\">[67]<\/a><\/sup>\n  That is, a better\nrecord measured in security breaches per total machine hours of\nInternet exposure.\n<\/p><\/div>\n\n<h3 id=\"ch07s02\"><a href=\"#ch07s02\">\u00a7<\/a>Taxonomy of Unix IPC Methods<\/h3>\n<p>\n  As in single-process program architectures, the simplest\norganization is the best.  The remainder of this chapter will present\nIPC techniques roughly in order of escalating complexity of\nprogramming them.  Before using a later, more complex technique, you\nshould prove by demonstration \u2014 with prototypes and benchmark\nresults \u2014 that no earlier and simpler technique will do.  Often\nyou will surprise yourself.\n<\/p>\n<h4 id=\"id2915475\"><a href=\"#id2915475\">\u00a7<\/a>Handing off Tasks to Specialist Programs<\/h4>\n<p>\n  In the simplest form of interprogram cooperation enabled by\ninexpensive process spawning, a program runs another to accomplish a\nspecialized task. Because the called program is often specified as a\nUnix shell command through the\nsystem(3)\ncall, this is often called\n  <em>shelling out<\/em>\n  to the\ncalled program. The called program inherits the user's keyboard and\ndisplay and runs to completion. When it exits, the calling program\nresumes control of the keyboard and display and resumes\nexecution.<sup><a href=\"#ftn.id2915505\" id=\"id2915505\">[68]<\/a><\/sup>\n  Because the calling program does not\ncommunicate with the called program during the callee's execution,\nprotocol design is not an issue in this kind of cooperation,\nexcept in the trivial sense that the caller may pass command-line\narguments to the callee to change its behavior.\n<\/p>\n<p>\n  The classic Unix case of shelling out is calling an editor from\nwithin a mail or news program. In the Unix tradition one does\n  <em>not<\/em>\n  bundle purpose-built editors into programs\nthat require general text-edited input. Instead, one allows the user\nto specify an editor of his or her choice to be called when editing\nneeds to be done.\n<\/p>\n<p>\n  The specialist program usually communicates with its parent through the\nfile system, by reading or modifying file(s) with specified\nlocation(s); this is how editor or mailer shellouts work.\n<\/p>\n<p>\n  In a common variant of this pattern, the specialist program may\naccept input on its standard input, and be called with the\nC\n  library entry\npoint\n  <code>popen(..., \"w\")<\/code> or as part of a\nshellscript.  Or it may send output to its standard output, and be\ncalled with\n  <code>popen(..., \"r\")<\/code> or as part of a\nshellscript.  (If it both reads from standard input and writes to\nstandard output, it does so in a batch mode, completing all reads\nbefore doing any writes.)  This kind of child process is not usually\nreferred to as a shellout; there is no standard jargon for it, but it\nmight well be called a \u2018bolt-on\u2019.\n<\/p>\n<p>\n  They key point about all these cases is that the specialist\nprograms don't handshake with the parent while they are running.  They\nhave an associated protocol only in the trivial sense that whichever\nprogram (master or slave) is accepting input from the other has to be\nable to parse it.\n<\/p>\n<h5 id=\"id2915594\"><a href=\"#id2915594\">\u00a7<\/a>Case Study: The mutt Mail User Agent<\/h5>\n<p>\n  The\n  <em>mutt<\/em> mail user agent is the modern\nrepresentative of the most important design tradition in Unix email\nprograms. It has a simple screen-oriented interface with\nsingle-keystroke commands for browsing and reading mail.\n<\/p>\n<p>\n  When you use\n  <em>mutt<\/em> as a mail composer (either by calling it\nwith an address as a command-line argument or by using one of the\nreply commands), it examines the process environment variable\n  <code>EDITOR<\/code>, and then generates a temporary file name. The\nvalue of the\n  <code>EDITOR<\/code> variable is called as a command\nwith the tempfile name as an argument.<sup><a href=\"#ftn.id2915659\" id=\"id2915659\">[69]<\/a><\/sup>\n  When that command terminates,\n  <em>mutt<\/em> resumes on the assumption that the\ntemporary file contains the desired mail text.\n<\/p>\n<p>\n  Almost all Unix mail- and netnews-composition programs observe\nthe same convention. Because they do, composer implementers don't\nneed to write a hundred inevitably diverging editors, and users\ndon't need to learn a hundred divergent interfaces.  Instead, users\ncan carry their chosen editors with them.\n<\/p>\n<p>\n  An important variant of this strategy shells out to a small proxy\nprogram that passes the specialist job to an already-running instance\nof a big program, like an editor or a Web browser.  Thus, developers\nwho normally have an instance of\n  <em>emacs<\/em> running on their X display can set\n  <code>EDITOR=emacsclient<\/code>, and have a buffer pop open in\ntheir\n  <em>emacs<\/em> when they request editing in\n  <em>mutt<\/em>.  The point of this is not really to\nsave memory or other resources, it's to enable the user to unify all\nediting in a single\n  <em>emacs<\/em> process (so that,\nfor example, cut and paste among buffers can carry along internal\n  <em>emacs<\/em> state information like font\nhighlighting).\n<\/p>\n<h4 id=\"plumbing\"><a href=\"#plumbing\">\u00a7<\/a>Pipes, Redirection, and Filters<\/h4>\n<p>\n  After Ken Thompson\n  and Dennis\nRitchie, the\nsingle most important formative figure of early Unix was probably Doug\nMcIlroy. His\ninvention of the\n  <em>pipe<\/em>\n  construct reverberated\nthrough the design of Unix, encouraging its nascent do-one-thing-well\nphilosophy and inspiring most of the later forms of IPC in the Unix\ndesign (in particular, the socket abstraction used for\nnetworking).\n<\/p>\n<p>\n  Pipes depend on the convention that every program has initially\navailable to it (at least) two I\/O data streams: standard input and\nstandard output (numeric file descriptors 0 and 1 respectively).\nMany programs can be written as\n  <em>filters<\/em>, which read\nsequentially from standard input and write only to standard\noutput.\n<\/p>\n<p>\n  Normally these streams are connected to the user's keyboard and\ndisplay, respectively. But Unix shells universally support\n  <em>redirection<\/em>\n  operations which connect these\nstandard input and output streams to files. Thus, typing\n<\/p>\n <pre class=\"programlisting\">\nls &gt;foo\n<\/pre>\n<p>\n  sends the output of the directory lister\nls(1)\nto a file named \u2018foo\u2019. On the other hand, typing:\n<\/p>\n <pre class=\"programlisting\">\nwc &lt;foo\n<\/pre>\n<p>\n  causes the word-count utility\nwc(1)\nto take its standard input from the file \u2018foo\u2019, and\ndeliver a character\/word\/line count to standard output.\n<\/p>\n<p>\n  The pipe operation connects the standard output of one program\nto the standard input of another. A chain of programs connected in\nthis way is called a\n  <em>pipeline<\/em>. If we write\n<\/p>\n <pre class=\"programlisting\">\nls | wc\n<\/pre>\n<p>\n  we'll see a character\/word\/line count for the current directory\nlisting. (In this case, only the line count is really likely to be\nuseful.)\n<\/p>\n <blockquote>\n  <p>\n   One favorite pipeline was \u201c<code>bc | speak<\/code>\u201d\u2014a \ntalking desk calculator.  It knew number names up to a\nvigintillion.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  It's important to note that all the stages in a pipeline run\nconcurrently.  Each stage waits for input on the output of the\nprevious one, but no stage has to exit before the next can run.  This\nproperty will be important later on when we look at interactive uses\nof pipelines, like sending the lengthy output of a command to\nmore(1).\n<\/p>\n<p>\n  It's easy to underestimate the power of combining pipes and\nredirection.  As an instructive example,\n  <em>The Unix Shell As a 4GL<\/em><sup><a href=\"#Schaffer-Wolf\" title=\"[Schaffer-Wolf]\">[Schaffer-Wolf]<\/a><\/sup> shows that with these\nfacilities as a framework, a handful of simple utilities can be\ncombined to support creating and manipulating relational databases\nexpressed as simple textual tables.\n<\/p>\n<p>\n  The major weakness of pipes is that they are unidirectional.\nIt's not possible for a pipeline component to pass control information\nback up the pipe other than by terminating (in which case the previous\nstage will get a\n  <code>SIGPIPE<\/code> signal on the\nnext write).  Accordingly, the protocol for passing data is simply the\nreceiver's input format.\n<\/p>\n<p>\n  So far, we have discussed anonymous pipes\ncreated by the shell. There is a variant called a\n  <em id=\"named_pipes\">\n   named\npipe\n<\/em>\n  which is a special kind of file. If two programs open the file, one for reading and the other for writing, a named\n  pipe acts like a pipe-fitting between them. Named pipes are a bit of a historical relic; they have been largely\n  displaced from use by named <em>sockets<\/em>, which we'll discuss below. (For more on the history of this relic, see\n  the discussion of <a href=\"#messages\" title=\"System V IPC\">System V IPC<\/a> below.)\n<\/p>\n<h5 id=\"id2916058\"><a href=\"#id2916058\">\u00a7<\/a>Case Study: Piping to a Pager<\/h5>\n<p>\n  Pipelines have many uses. For one example, Unix's process lister\nps(1)\nlists processes to standard output without caring that a long listing\nmight scroll off the top of the user's display too quickly for the\nuser to see it.  Unix has another program,\nmore(1),\nwhich displays its standard input in screen-sized chunks, prompting\nfor a user keystroke after displaying each screenful.\n<\/p>\n<p>\n  Thus, if the user types \u201c<code>ps | more <\/code>\u201d, piping the output of\nps(1)\nto the input of\nmore(1),\nsuccessive page-sized pieces of the list of processes will be\ndisplayed after each keystroke.\n<\/p>\n<p>\n  The ability to combine programs like this can be extremely\nuseful. But the real win here is not cute combinations; it's that\nbecause both pipes and\nmore(1)\nexist,\n  <em>other programs can be simpler<\/em>. Pipes mean\nthat programs like\nls(1)\n(and other programs that write to standard out) don't have to grow\ntheir own pagers \u2014 and we're saved from a world of a thousand\nbuilt-in pagers (each, naturally, with its own divergent look and\nfeel). Code bloat is avoided and global complexity reduced.\n<\/p>\n<p>\n  As a bonus, if anyone needs to customize pager behavior, it can\nbe done in\n  <em>one<\/em>\n  place, by changing\n  <em>one<\/em>\n  program.  Indeed, multiple pagers can exist,\nand will all be useful with every application that writes to standard\noutput.\n<\/p>\n<p>\n  In fact, this has actually happened. On modern Unixes,\nmore(1)\nhas been largely replaced by\nless(1),\nwhich adds the capability to scroll back in the displayed file rather\nthan just forward.<sup><a href=\"#ftn.id2916197\" id=\"id2916197\">[70]<\/a><\/sup>\n  Because\nless(1)\nis decoupled from the programs that use it, it's possible to simply\nalias \u2018more\u2019 to \u2018less\u2019 in your shell, set the\nenvironment variable\n  <code>PAGER<\/code> to \u2018less\u2019 (see\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">\n   Chapter 10<\/a>), and get\nall the benefits of a better pager with all properly-written Unix\nprograms.\n<\/p>\n<h5 id=\"id2916247\"><a href=\"#id2916247\">\u00a7<\/a>Case Study: Making Word Lists<\/h5>\n<p>\n  A more interesting example is one in which pipelined programs\ncooperate to do some kind of data transformation for which, in less\nflexible environments, one would have to write custom code.\n<\/p>\n<p>\n  Consider the pipeline\n<\/p>\n <pre class=\"programlisting\">\ntr -c '[:alnum:]' '[\\n*]' | sort -iu | grep -v '^[0-9]*$'\n<\/pre>\n<p>\n  The first command translates non-alphanumerics on standard input\nto newlines on standard output. The second sorts lines on standard\ninput and writes the sorted data to standard output, discarding all\nbut one copy of spans of adjacent identical lines.  The third\ndiscards all lines consisting solely of digits.  Together, these\ngenerate a sorted wordlist to standard output from text on standard\ninput.\n<\/p>\n<h5 id=\"id2916289\"><a href=\"#id2916289\">\u00a7<\/a>Case Study: pic2graph<\/h5>\n<p>\n  Shell source code for the program pic2graph(1) ships with the <em>groff<\/em> suite of text-formatting tools from the\n  Free Software Foundation.  It translates diagrams written in the PIC language to bitmap images. Example 7.1 shows the\n  pipeline at the heart of this code.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 7.1. The\n    <em>pic2graph<\/em> pipeline.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n(echo \".EQ\"; echo $eqndelim; echo \".EN\"; echo \".PS\";cat;echo \".PE\")|\\\n     groff -e -p $groffpic_opts -Tps &gt;${tmp}.ps \\\n     &amp;&amp; convert -crop 0x0 $convert_opts ${tmp}.ps ${tmp}.${format} \\\n     &amp;&amp; cat ${tmp}.${format}\n<\/pre><\/div>\n<p>\n  The\npic2graph(1)\nimplementation illustrates how much one pipeline can do purely by\ncalling preexisting tools.  It starts by massaging its input into an\nappropriate form, continues by feeding it through\ngroff(1)\nto produce PostScript, and finishes by converting the PostScript to a\nbitmap.  All these details are hidden from the user, who simply sees\nPIC source go in one end and a bitmap ready for inclusion in a Web\npage come out the other.\n<\/p>\n<p>\n  This is an interesting example because it illustrates how\npipes\n  and filtering can\nadapt programs to unexpected uses.  The program that interprets PIC,\npic(1),\nwas originally designed only to be used for embedding diagrams in\ntypeset documents.  Most of the other programs in the toolchain it was\npart of are now semiobsolescent.  But PIC remains handy for new uses,\nsuch as describing diagrams to be embedded in HTML.  It gets a renewed\nlease on life because tools like\npic2graph(1)\ncan bundle together all the machinery needed to convert the output of\npic(1)\ninto a more modern format.\n<\/p>\n<p>\n  We'll examine pic(1) more closely, as a minilanguage design, in <a href=\"#minilanguageschapter\" title=\"Chapter 8.&#10;&#9;  Minilanguages\">Chapter 8<\/a>.\n<\/p>\n<h5 id=\"id2916464\"><a href=\"#id2916464\">\u00a7<\/a>Case Study: bc(1) and dc(1)<\/h5>\n<p>\n  Part of the classic Unix toolkit dating back to Version 7 is a\npair of calculator programs. The\ndc(1)\nprogram is a simple calculator that accepts text lines consisting of\nreverse-Polish notation (RPN) on standard input and emits calculated\nanswers to standard output. The\nbc(1)\nprogram accepts a more elaborate infix syntax resembling conventional\nalgebraic notation; it includes as well the ability to set and read\nvariables and define functions for elaborate formulas.\n<\/p>\n<p>\n  While the modern GNU implementation of\nbc(1)\nis standalone, the classic version passed commands to\ndc(1)\nover a pipe. In this division of labor,\nbc(1)\ndoes variable substitution and function expansion and translates infix\nnotation into reverse-Polish \u2014 but doesn't actually do\ncalculation itself, instead passing RPN translations of input\nexpressions to\ndc(1)\nfor evaluation.\n<\/p>\n<p>\n  There are clear advantages to this separation of function. It\nmeans that users get to choose their preferred notation, but the logic\nfor arbitrary-precision numeric calculation (which is moderately\ntricky) does not have to be duplicated. Each of the pair of programs\ncan be less complex than one calculator with a choice of notations\nwould be. The two components can be debugged and mentally modeled\nindependently of each other.\n<\/p>\n<p>\n  In\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  we\nwill reexamine these programs from a slightly different example, as\nexamples of domain-specific minilanguages.\n<\/p>\n<h5 id=\"id2921380\"><a href=\"#id2921380\">\u00a7<\/a>Anti-Case Study: Why Isn't fetchmail a Pipeline?<\/h5>\n<p>\n  In Unix terms,\n  <em>fetchmail<\/em> is an\nuncomfortably large program that bristles with options.  Thinking\nabout the way mail transport works, one might think it would be\npossible to decompose it into a pipeline. Suppose for a moment it were\nbroken up into several programs: a couple of fetch programs to get\nmail from POP3 and IMAP sites, and a local SMTP injector.  The\npipeline could pass Unix mailbox format.  The present elaborate\n  <em>fetchmail<\/em> configuration could be replaced\nby a shellscript containing command lines.  One could even insert\nfilters in the pipeline to block spam.\n<\/p>\n <pre class=\"programlisting\">\n#!\/bin\/sh\nimap jrandom@imap.ccil.org | spamblocker | smtp jrandom\nimap jrandom@imap.netaxs.com | smtp jrandom\n# pop ed@pop.tems.com | smtp jrandom\n<\/pre>\n<p>\n  This would be very elegant and Unixy.  Unfortunately, it can't\nwork.  We touched on the reason earlier; pipelines are\nunidirectional.\n<\/p>\n<p>\n  One of the things the fetcher program\n(\n  <em>imap<\/em> or\n  <em>pop<\/em>)\nwould have to do is decide whether to send a delete request for each\nmessage it fetches.  In\n  <em>fetchmail<\/em> 's present\norganization, it can delay sending that request to the POP or IMAP\nserver until it knows that the local SMTP listener has accepted\nresponsibility for the message.  The pipelined, small-component\nversion would lose that property.\n<\/p>\n<p>\n  Consider, for example, what would happen if the\n  <em>smtp<\/em> injector fails because the SMTP\nlistener reports a disk-full condition. If the fetcher has already\ndeleted the mail, we lose.  This means the fetcher cannot delete mail\nuntil it is notified to do so by the\n  <em>smtp<\/em> injector.  This in turn raises a host of questions. How would they\ncommunicate?  What message, exactly, would the injector pass back?\nThe global complexity of the resulting system, and its vulnerability\nto subtle bugs, would almost certainly be higher than that of a\nmonolithic program.\n<\/p>\n<p>\n  Pipelines are a marvelous tool, but not a universal one.\n<\/p>\n<h4 id=\"id2921506\"><a href=\"#id2921506\">\u00a7<\/a>Wrappers<\/h4>\n<p>\n  The opposite of a shellout is a\n  <em>wrapper<\/em>. A\nwrapper creates a new interface for a called program, or specializes\nit.  Often, wrappers are used to hide the details of elaborate shell\npipelines.  We'll\ndiscuss interface wrappers in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>. Most specialization wrappers are quite\nsimple, but nevertheless very useful.\n<\/p>\n<p>\n  As with shellouts, there is no associated protocol because the\nprograms do not communicate during the execution of the callee; but\nthe wrapper usually exists to specify arguments that modify the\ncallee's behavior.\n<\/p>\n<h5 id=\"id2921550\"><a href=\"#id2921550\">\u00a7<\/a>Case Study: Backup Scripts<\/h5>\n<p>\n  Specialization wrappers are a classic use of the Unix shell and\nother scripting languages. One kind of specialization wrapper\nthat is both common and representative is a backup script. It may be a\none-liner as simple as this:\n<\/p>\n <pre class=\"programlisting\">\ntar -czvf \/dev\/st0 \"$@\"\n<\/pre>\n<p>\n  This is a wrapper for the\ntar(1)\ntape archiver utility which simply supplies one fixed argument (the\ntape device\n  <code>\/dev\/st0<\/code>) and passes to tar all the other arguments\nsupplied by the user (\u201c\n  <code>$@<\/code> \u201d).<sup><a href=\"#ftn.id2921610\" id=\"id2921610\">[71]<\/a><\/sup><\/p>\n<h4 id=\"id2921634\"><a href=\"#id2921634\">\u00a7<\/a>Security Wrappers and Bernstein Chaining<\/h4>\n<p>\n  One common use of wrapper scripts is as\n  <em>security wrappers<\/em>.  A security script may call a gatekeeper program\nto check some sort of credential, then conditionally execute another\nbased on the status value returned by the gatekeeper.\n<\/p>\n<p>\n  Bernstein chaining is a specialized security-wrapper technique\nfirst invented by Daniel J. Bernstein, who has employed it in a number\nof his packages. (A similar pattern appears in commands like\nnohup(1)\nand\nsu(1),\nbut the conditionality is absent.)  Conceptually, a Bernstein chain is\nlike a pipeline, but\neach successive stage replaces the previous one rather than running\nconcurrently with it.\n<\/p>\n<p>\n  The usual application is to confine security-privileged\napplications to some sort of gatekeeper program, which can then hand\nstate to a less privileged one.  The technique pastes several programs\ntogether using execs, or possibly a combination of forks and\nexecs. The programs are all named on one command line. Each program\nperforms some function and (if successful) runs\nexec(2)\non the rest of its command line.\n<\/p>\n<p>\n  Bernstein's\n  <em>rblsmtpd<\/em> package is a\nprototypical example. It serves to look up a host in the antispam DNS\nzone of the Mail Abuse Prevention System. It does this by doing a DNS\nquery on the IP address passed into it in the\n  <code>TCPREMOTEIP<\/code> environment variable.  If the query is\nsuccessful, then\n  <em>rblsmtpd<\/em> runs its own SMTP\nthat discards the mail. Otherwise the remaining command-line arguments\nare presumed to constitute a mail transport agent that knows the SMTP\nprotocol, and are handed to\nexec(2)\nto be run.\n<\/p>\n<p>\n  Another example can be found in Bernstein's\n  <em>qmail<\/em> package. It contains a program called\n  <em>condredirect<\/em>. The first parameter is an\nemail address, and the remainder a gatekeeper program and\narguments.\n  <em>condredirect<\/em> forks and execs the\ngatekeeper with its arguments. If the gatekeeper exits successfully,\n  <em>condredirect<\/em> forwards the email\npending on stdin to the specified email address.  In this case,\nopposite to that of\n  <em>rblsmtpd<\/em>, the security\ndecision is made by the child; this case is a bit more like a\nclassical shellout.\n<\/p>\n<p>\n  A more elaborate example is the\n  <em>qmail<\/em> POP3 server. It consists of three programs,\n  <em>qmail-popup<\/em>,\n  <em>checkpassword<\/em>, and\n  <em>qmail-pop3d<\/em>.\n  <em>Checkpassword<\/em> comes from a separate package\ncleverly called\n  <em>checkpassword<\/em>, and\nunsurprisingly it checks the password. The POP3 protocol has an\nauthentication phase and mailbox phase; once you enter the mailbox\nphase you cannot go back to the authentication phase. This is a\nperfect application for Bernstein chaining.\n<\/p>\n<p>\n  The first parameter of\n  <em>qmail-popup<\/em> is\nthe hostname to use in the POP3 prompts. The rest of its parameters\nare forked and passed to\nexec(2),\nafter the POP3 username and password have been fetched. If the program\nreturns failure, the password must be wrong, so\n  <em>qmail-popup<\/em> reports that and waits for a\ndifferent password. Otherwise, the program is presumed to have\nfinished the POP3 conversation, so\n  <em>qmail-popup<\/em> exits.\n<\/p>\n<p>\n  The program named on\n  <em>qmail-popup<\/em> 's\ncommand line is expected to read three null-terminated strings from\nfile descriptor 3.<sup><a href=\"#ftn.id2921897\" id=\"id2921897\">[72]<\/a><\/sup>\n  These are the username, password, and response to\na cryptographic challenge, if any.  This time it's\n  <em>checkpassword<\/em> which accepts as parameters\nthe name of\n  <em>qmail-pop3d<\/em> and its\nparameters. The\n  <em>checkpassword<\/em> program exits\nwith failure if the password does not match; otherwise it changes to\nthe user's uid, gid, and home directory, and executes the rest of its\ncommand line on behalf of that user.\n<\/p>\n<p>\n  Bernstein chaining is useful for situations in which the\napplication needs setuid or setgid privileges to initialize a\nconnection, or to acquire some credential, and then drop those\nprivileges so that following code does not have to be trusted.\nFollowing the exec, the child program cannot set its real user ID back\nto root. It's also more flexible than a single process, because you\ncan modify the behavior of the system by inserting another program\ninto the chain.\n<\/p>\n<p>\n  For example,\n  <em>rblsmtpd<\/em> (mentioned\nabove) can be inserted into a Bernstein chain, in between tcpserver\n(from the\n  <em>ucspi-tcp<\/em> package) and the real\nSMTP server, typically\n  <em>qmail-smtpd<\/em>. However, it works with\ninetd(8)\nand <code>sendmail -bs<\/code> as well.\n<\/p>\n<h4 id=\"id2922002\"><a href=\"#id2922002\">\u00a7<\/a>Slave Processes<\/h4>\n<p>\n  Occasionally, child programs both accept data from and return\ndata to their callers through\npipes\n  connected to\nstandard input and output, interactively. Unlike simple shellouts and\nwhat we have called \u2018bolt-ons\u2019 above, both master and\nslave processes need to have internal state machines to handle a\nprotocol between them without deadlocking or racing.  This is a\ndrastically more complex and more difficult-to-debug organization than\na simple shellout.\n<\/p>\n<p>\n  Unix's\npopen(3)\ncall can set up either an input pipe or an output pipe for a shellout,\nbut not both for a slave process \u2014 this seems intended to\nencourage simpler programming. And, in fact, interactive master-slave\ncommunication is tricky enough that it is normally only used when\neither (a) the implied protocol is utterly trivial, or (b) the slave\nprocess has been designed to speak an application protocol along the\nlines we discussed in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>.  We'll return to this issue, and ways\nto cope with it, in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>.\n<\/p>\n<p>\n  When writing a master\/slave pair, it is good practice for the\nmaster to support a command-line switch or environment variable that\nallows callers to set their own slave command.  Among other things,\nthis is useful for debugging; you will often find it handy during \ndevelopment to invoke the real slave process from within a harness\nthat monitors and logs transactions between slave and master.\n<\/p>\n<p>\n  If you find that master\/slave interactions in your program are\nbecoming nontrivial, it may be time to think about going the rest of\nthe way to a more peer-to-peer organization, using techniques like\nsockets or shared memory.\n<\/p>\n<h5 id=\"id2922085\"><a href=\"#id2922085\">\u00a7<\/a>Case Study: scp and ssh<\/h5>\n<p>\n  One common case in which the implied protocol really is trivial is\nprogress meters.  The\nscp(1)\nsecure-copy command calls\nssh(1)\nas a slave process, intercepting enough information from ssh's\nstandard output to reformat the reports as an ASCII animation of a\nprogress bar.<sup><a href=\"#ftn.id2922127\" id=\"id2922127\">[73]<\/a><\/sup><\/p>\n<h4 id=\"id2922148\"><a href=\"#id2922148\">\u00a7<\/a>Peer-to-Peer Inter-Process Communication<\/h4>\n<p>\n  All the communication methods we've discussed so far have a sort\nof implicit hierarchy about them, with one program effectively\ncontrolling or driving another and zero or limited feedback passing in\nthe opposite direction. In communications and networking we frequently\nneed channels that are\n  <em>peer-to-peer<\/em>, usually (but\nnot necessarily) with data flowing freely in both directions. We'll\nsurvey peer-to-peer communications methods under Unix here, and\ndevelop some case studies in later chapters.\n<\/p>\n<h5 id=\"id2922172\"><a href=\"#id2922172\">\u00a7<\/a>Tempfiles<\/h5>\n<p>\n  The use of tempfiles as communications drops between\ncooperating programs is the oldest IPC technique there is.  Despite\ndrawbacks, it's still useful in shellscripts, and in one-off programs\nwhere a more elaborate and coordinated method of communication would\nbe overkill.\n<\/p>\n<p>\n  The most obvious problem with using tempfiles as an IPC\ntechnique is that it tends to leave garbage lying around if processing\nis interrupted before the tempfile can be deleted. A less obvious risk\nis that of collisions between multiple instances of a program using\nthe same name for a tempfile. This is why it is conventional for\nshellscripts that make tempfiles to include $$ in their names; this\nshell variable expands to the process-ID of the enclosing shell and\neffectively guarantees that the filename will be unique (the same \ntrick is supported in Perl).\n<\/p>\n<p>\n  Finally, if an attacker knows the location to which a tempfile\nwill be written, it can overwrite on that name and possibly either read\nthe producer's data or spoof the consumer process by inserting\nmodified or spurious data into the file.<sup><a href=\"#ftn.id2922208\" id=\"id2922208\">[74]<\/a><\/sup>\n  This is a security risk.  If the processes involved have root\nprivileges, this is a very serious risk.  It can be mitigated by setting\nthe permissions on the tempfile directory carefully, but such\narrangements are notoriously likely to spring leaks.\n<\/p>\n<p>\n  All these problems aside, tempfiles still have a niche because\nthey're easy to set up, they're flexible, and they're less vulnerable\nto deadlocks or race conditions than more elaborate methods. And\nsometimes, nothing else will do.  The calling conventions of your\nchild process may require that it be handed a file to operate on.  Our\nfirst example of a shellout to an editor demonstrates this\nperfectly.\n<\/p>\n<h5 id=\"id2922239\"><a href=\"#id2922239\">\u00a7<\/a>Signals<\/h5>\n<p>\n  The simplest and crudest way for two processes on the same\nmachine to communicate with each other is for one to send the other a\n  <em>signal<\/em>. Unix signals are a form of soft\ninterrupt; each one has a default effect on the receiving process\n(usually to kill it). A process can declare a\n  <em>\n   signal\nhandler\n<\/em>\n  that overrides the default action for the signal; the\nhandler is a function that is executed asynchronously when the signal\nis received.\n<\/p>\n<p>\n  Signals were originally designed into Unix as a way for the\noperating system to notify programs of certain errors and critical\nevents, not as an IPC facility. The\n  <code>SIGHUP<\/code> signal,\nfor example, is sent to every program started from a given terminal\nsession when that session is terminated. The\n  <code>SIGINT<\/code> signal is sent to whatever process is currently attached to the\nkeyboard when the user enters the currently-defined interrupt\ncharacter (often control-C). Nevertheless, signals can be useful for\nsome IPC situations (and the POSIX-standard signal set includes two\nsignals,\n  <code>SIGUSR1<\/code> and\n  <code>SIGUSR2<\/code>,\nintended for this\nuse). They are often\nemployed as a control channel for\n  <em>daemons<\/em>\n  (programs that run constantly, invisibly, in background), a way for an\noperator or another program to tell a daemon that it needs to either\nreinitialize itself, wake up to do work, or write\ninternal-state\/debugging information to a known location.\n<\/p>\n <blockquote>\n  <p>\n   I insisted\n   <code>SIGUSR1<\/code> and\n   <code>SIGUSR2<\/code> be invented for BSD. People\nwere grabbing system signals to mean what they needed them to mean for IPC,\nso that (for example) some programs that segfaulted would not coredump\nbecause\n   <code>SIGSEGV<\/code> had been hijacked.\n  <\/p>\n  <p>\n   This is a general principle \u2014 people will want to hijack\nany tools you build, so you have to design them to either be\nun-hijackable or to be hijacked cleanly.  Those are your only choices.\nExcept, of course, for being ignored\u2014a highly reliable way to\nremain unsullied, but less satisfying than might at first appear.\n  <\/p><div>\n  \n   \u2013 Ken Arnold\n  \n <\/div><\/blockquote>\n<p>\n  A technique often used with signal IPC is the so-called\n  <em>pidfile<\/em>. Programs that will need to be signaled\nwill write a small file to a known location (often in\n  <code>\/var\/run<\/code> or the invoking user's home directory)\ncontaining their process ID or PID. Other programs can read that file\nto discover that PID. The pidfile may also function as an implicit\n  <em>lock file<\/em>\n  in cases where no more than one\ninstance of the daemon should be running simultaneously.\n<\/p>\n<p>\n  There are actually two different flavors of signals.  In the\nolder implementations (notably V7, System\nIII, and early\nSystem V), the\nhandler for a given signal is reset to the default for that signal\nwhenever the handler fires.  The result of sending two of the same\nsignal in quick succession is therefore usually to kill the process,\nno matter what handler was set.\n<\/p>\n<p>\n  The BSD 4.x\n  versions of Unix changed to\n\u201creliable\u201d signals, which do not reset unless the user\nexplicitly requests it.  They also introduced primitives to block or\ntemporarily suspend processing of a given set of signals.  Modern\nUnixes support both styles.  You should use the BSD-style\nnonresetting entry points for new code, but program defensively in\ncase your code is ever ported to an implementation that does not\nsupport them.\n<\/p>\n<p>\n  Receiving N signals does not necessarily invoke the signal\nhandler N times.  Under the older System V signal model, two or more\nsignals spaced very closely together (that is, within a single\ntimeslice of the target process) can result in various race\nconditions<sup><a href=\"#ftn.id2922479\" id=\"id2922479\">[75]<\/a><\/sup>\n  or anomalies. Depending on what\nvariant of signals semantics the system supports, the second and later\ninstances may be ignored, may cause an unexpected process kill, or may\nhave their delivery delayed until earlier instances have been\nprocessed (on modern Unixes the last is most likely).\n<\/p>\n<p>\n  The modern signals API is portable across all recent Unix\nversions, but not to Windows or classic (pre-OS X) MacOS.\n<\/p>\n<h5 id=\"id2922508\"><a href=\"#id2922508\">\u00a7<\/a>System Daemons and Conventional Signals<\/h5>\n<p>\n  Many well-known system daemons accept\n  <code>SIGHUP<\/code> (originally the signal sent to programs on a serial-line drop, such as\nwas produced by hanging up a modem connection) as a signal to\nreinitialize (that is, reload their configuration files); examples\ninclude Apache\n  and the\nLinux implementations of\nbootpd(8),\ngated(8),\ninetd(8),\nmountd(8),\nnamed(8),\nnfsd(8),\nand\nypbind(8). In\na few cases,\n  <code>SIGHUP<\/code> is accepted in its original sense\nof a session-shutdown signal (notably in Linux\npppd(8)),\nbut that role nowadays generally goes to\n  <code>SIGTERM<\/code>.\n<\/p>\n<p>\n  <code>SIGTERM<\/code> (\u2018terminate\u2019) is\noften accepted as a graceful-shutdown signal (this is as distinct from\n  <code>SIGKILL<\/code>, which does an immediate process\nkill and cannot be blocked or handled).\n  <code>SIGTERM<\/code> actions often involve cleaning up\ntempfiles, flushing final updates out to databases, and the\nlike.\n<\/p>\n<p>\n  When writing daemons, follow the Rule of Least Surprise: use\nthese conventions, and read the manual pages to look for existing\nmodels.\n<\/p>\n<h5 id=\"fetchmail_signals\"><a href=\"#fetchmail_signals\">\u00a7<\/a>Case Study: fetchmail's Use of Signals<\/h5>\n<p>\n  The\n  <em>fetchmail<\/em> utility is normally set\nup to run as a daemon in background, periodically collecting mail from\nall remote sites defined in its run-control file and passing the mail\nto the local SMTP listener on port 25 without user\nintervention.\n  <em>fetchmail<\/em> sleeps for a\nuser-defined interval (defaulting to 15 minutes) between collection\nattempts, so as to avoid constantly loading the network.\n<\/p>\n<p>\n  When you invoke <code>fetchmail<\/code> with no arguments,\nit checks to see if you have a\n  <em>fetchmail<\/em> daemon already running (it does this by looking for a pidfile). If no\ndaemon is running,\n  <em>fetchmail<\/em> starts up\nnormally using whatever control information has been specified in its\nrun-control file. If a daemon is running, on the other hand, the new\n  <em>fetchmail<\/em> instance just signals the old one\nto wake up and collect mail immediately; then the new instance\nterminates. In addition, <code>fetchmail -q<\/code> sends a\ntermination signal to any running\n  <em>fetchmail<\/em> daemon.\n<\/p>\n<p>\n  Thus, typing <code>fetchmail<\/code> means, in effect,\n\u201cpoll now and leave a daemon running to poll later; don't bother me\nwith the detail of whether a daemon was already running or not\u201d.\nObserve that the detail of which particular signals are used for\nwakeup and termination is something the user doesn't have to know.\n<\/p>\n<h5 id=\"sockets\"><a href=\"#sockets\">\u00a7<\/a>Sockets<\/h5>\n<p>\n  Sockets\n  were developed in the\nBSD\n  lineage of Unix\nas a way to encapsulate access to data networks. Two programs\ncommunicating over a socket typically see a bidirectional byte stream\n(there are other socket modes and transmission methods, but they are\nof only minor importance). The byte stream is both sequenced (that is,\neven single bytes will be received in the same order sent) and\nreliable (socket users are guaranteed that the underlying network will\ndo error detection and retry to ensure delivery). Socket descriptors,\nonce obtained, behave essentially like file descriptors.\n<\/p>\n <blockquote>\n  <p>\n   Sockets differ from read\/write in one important case.  If\nthe bytes you send arrive, but the receiving machine fails to ACK, the\nsending machine's TCP\/IP stack will time out.  So getting an error does\n   <em>not<\/em>\n   necessarily mean that the bytes didn't\narrive; the receiver may be using them.  This problem has profound \nconsequences for the design of reliable protocols, because you have to be \nable to work properly when you don't know what was received in the \npast.  Local I\/O is \u2018yes\/no\u2019.  Socket I\/O is \u2018yes\/no\/maybe\u2019.\nAnd nothing can ensure delivery \u2014 the remote machine\nmight have been destroyed by a comet.\n  <\/p><div>\n  \n   \u2013 Ken Arnold\n  \n <\/div><\/blockquote>\n<p>\n  At the time a socket is created, you specify a\n  <em>protocol family<\/em>\n  which tells the network layer how\nthe name of the socket is interpreted. Sockets are usually thought of\nin connection with the Internet, as a way of passing data between\nprograms running on different hosts; this is the AF_INET socket\nfamily, in which addresses are interpreted as host-address and\nservice-number pairs. However, the AF_UNIX (aka AF_LOCAL) protocol\nfamily supports the same socket abstraction for communication between\ntwo processes on the same machine (names are interpreted as the\nlocations of special files analogous to bidirectional named pipes). As\nan example, client programs and servers using the X windowing\nsystem\n  typically use AF_LOCAL sockets to communicate.\n<\/p>\n<p>\n  All modern Unixes support BSD-style\n  sockets, and as a\nmatter of design they are usually the right thing to use for\nbidirectional IPC no matter where your cooperating processes are\nlocated.  Performance pressure may push you to use shared memory or\ntempfiles or other techniques that make stronger locality assumptions,\nbut under modern conditions it is best to assume that your code will\nneed to be scaled up to distributed operation.  More importantly,\nthose locality assumptions may mean that portions of your system get\nchummier with each others' internals than ought to be the case in a\ngood design.  The separation of address spaces that sockets enforce is\na feature, not a bug.\n<\/p>\n<p>\n  To use sockets gracefully, in the Unix tradition, start by\ndesigning an\n  <em>application protocol<\/em>\n  for use between\nthem \u2014 a set of requests and responses which expresses the\nsemantics of what your programs will be communicating about in a\nsuccinct way.  We've already discussed the some major issues in the\ndesign of application protocols in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>.\n<\/p>\n<p>\n  Sockets are supported in all recent Unixes, under Windows, and\nunder classic MacOS as well.\n<\/p>\n <h6 id=\"id2922950\">\n  Case Study: PostgreSQL\n <\/h6>\n<p>\n  PostgreSQL is an open-source database program.  Had it been\nimplemented as a monster monolith, it would be a single program with\nan interactive interface that manipulates database files on disk\ndirectly.  Interface would be welded together with implementation, and\ntwo instances of the program attempting to manipulate the same\ndatabase at the same time would have serious contention and locking\nissues.\n<\/p>\n<p>\n  Instead, the PostgreSQL suite includes a server called\n  <em>postmaster<\/em> and at least three client\napplications.  One\n  <em>postmaster<\/em> server\nprocess per machine runs in background and has exclusive access to the\ndatabase files.  It accepts requests in the SQL query minilanguage through\nTCP\/IP\n  sockets,\nand returns answers in a textual format as well.  When the user runs a\nPostgreSQL client, that client opens a session to\n  <em>postmaster<\/em> and does SQL transactions with\nit.  The server can handle several client sessions at once, and\nsequences requests so that they don't interfere with each other.\n<\/p>\n<p>\n  Because the front end and back end are separate, the server\ndoesn't need to know anything except how to interpret SQL requests\nfrom a client and send SQL reports back to it.  The clients, on the \nother hand, don't need to know anything about how the database is \nstored.  Clients can be specialized for different needs and have\ndifferent user interfaces.\n<\/p>\n<p>\n  This organization is quite typical for Unix databases \u2014 so\nmuch so that it is often possible to mix and match SQL clients and SQL\nservers. The interoperability issues are the SQL server's\nTCP\/IP\n  port number,\nand whether client and server support the same dialect of SQL.\n<\/p>\n <h6 id=\"id2923040\">\n  Case Study: Freeciv\n <\/h6>\n<p>\n  In\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>, we\nintroduced Freeciv as an example of transparent data formats.  But\nmore critical to the way it supports multiplayer gaming is the\nclient\/server partitioning of the code.  This is a representative\nexample of a program in which the application needs to be distributed\nover a wide-area network and handles communication through TCP\/IP\nsockets.\n<\/p>\n<p>\n  The state of a running Freeciv game is maintained by a server\nprocess, the game engine.  Players run GUI clients which exchange\ninformation and commands with the server through a packet protocol.\nAll game logic is handled in the server.  The details of GUI are\nhandled in the client; different clients support different interface\nstyles.\n<\/p>\n<p>\n  This is a very typical organization for a multiplayer online\ngame. The packet protocol uses\nTCP\/IP\n  as a transport,\nso one server can handle clients running on different Internet hosts.\nOther games that are more like real-time simulations (notably\nfirst-person shooters) use raw Internet datagram protocol (UDP) and\ntrade lower latency for some uncertainty about whether any given\npacket will be delivered. In such games, users tend to be issuing\ncontrol actions continuously, so sporadic dropouts are tolerable, but\nlag is fatal.\n<\/p>\n<h5 id=\"id2923100\"><a href=\"#id2923100\">\u00a7<\/a>Shared Memory<\/h5>\n<p>\n  Whereas two processes using sockets to communicate may live on\ndifferent machines (and, in fact, be separated by an Internet\nconnection spanning half the globe), shared memory requires producers\nand consumers to be co-resident on the same hardware.  But, if your\ncommunicating processes can get access to the same physical memory,\nshared memory will be the fastest way to pass information between\nthem.\n<\/p>\n<p>\n  Shared memory may be disguised under different APIs, but on\nmodern Unixes the implementation normally depends on the use of\nmmap(2)\nto map files into memory that can be shared between processes.\nPOSIX\n  defines a\nshm_open(3)\nfacility with an API that supports using files as shared memory; this\nis mostly a hint to the operating system that it need not flush the\npseudofile data to disk.\n<\/p>\n<p>\n  Because access to shared memory is not automatically serialized\nby a discipline resembling read and write calls, programs doing the\nsharing must handle contention and deadlock issues themselves,\ntypically by using semaphore variables located in the shared segment.\nThe issues here resemble those in multithreading (see the end of this\nchapter for discussion) but are more manageable because default is\n  <em>not<\/em>\n  to share memory. Thus, problems are better\ncontained.\n<\/p>\n<p>\n  On systems where it is available and reliable, the\nApache\n  web server's\nscoreboard facility uses shared memory for communication between an\nApache master process and the load-sharing pool of Apache images that it\nmanages. Modern X implementations also use shared memory, to pass\nlarge images between client and server when they are resident on the\nsame machine,  to avoid the overhead of socket communication.\nBoth uses are performance hacks justified by experience and testing,\nrather than being architectural choices.\n<\/p>\n<p>\n  The\nmmap(2)\ncall is supported under all modern Unixes, including\nLinux\n  and the open-source BSD\n  versions; this is described in the Single Unix Specification.  It will\nnot normally be available under Windows, MacOS classic, and other\noperating systems.\n<\/p>\n<p>\n  Before purpose-built\nmmap(2)\nwas available, a common way for two processes to communicate was for\nthem to open the same file, and then delete that file.  The file\nwouldn't go away until all open filehandles were closed, but some old\nUnixes took the link count falling to zero as a hint that they could\nstop updating the on-disk copy of the file.  The downside was that\nyour backing store was the file system rather than a swap device,\nthe file system the deleted file lived on couldn't be unmounted until the\nprograms using it closed, and attaching new processes to an existing\nshared memory segment faked up in this way was tricky at best.\n<\/p>\n<p>\n  After Version 7 and the split between the BSD and System V lineages,\nthe evolution of Unix interprocess communication took two different\ndirections. The BSD direction led to\n  <a href=\"#sockets\" title=\"Sockets\">sockets<\/a>. The AT&amp;T lineage, on the other\nhand, developed named pipes (as\n  <a href=\"#named_pipes\">previously discussed<\/a>) and an IPC facility, specifically designed for\npassing binary data and based on shared-memory bidirectional message queues.  This is called \u2018System V IPC\u2019\u2014or, among\nold timers, \u2018Indian Hill\u2019 IPC after the AT&amp;T facility where it was first written.\n<\/p>\n<p>\n  The upper, message-passing layer of System V IPC has largely\nfallen out of use.  The lower layer, which consists of shared memory\nand semaphores, still has significant applications under circumstances\nin which one needs to do mutual-exclusion locking and some global data\nsharing among processes running on the same machine.  These System V\nshared memory facilities evolved into the POSIX shared-memory API,\nsupported under Linux, the BSDs, MacOS X and Windows, but not classic\nMacOS.\n<\/p>\n<p>\n  By using these shared-memory and semaphore facilities\n(shmget(2),\nsemget(2),\nand friends) one can avoid the overhead of copying data through the\nnetwork stack.  Large commercial databases (including Oracle, DB2, Sybase, and\nInformix) use this technique heavily.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2915505\" id=\"ftn.id2915505\">[68]<\/a><\/sup>\n  A common error in programming shellouts is to\nforget to block signals in the parent while the subprocess runs.\nWithout this precaution, an interrupt typed to the subprocess can have\nunwanted side effects on the parent\nprocess.\n<\/p><p><sup><a href=\"#id2915659\" id=\"ftn.id2915659\">[69]<\/a><\/sup>\n  Actually, the above\nis a slight oversimplification.  See the discussion of\n  <code>EDITOR<\/code> and\n  <code>VISUAL<\/code> in\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">\n   Chapter 10<\/a>\n  for the rest of the\nstory.\n<\/p><p><sup><a href=\"#id2916197\" id=\"ftn.id2916197\">[70]<\/a><\/sup>\n  The\nless(1)\nman page explains the name by observing \u201cLess is\nmore\u201d.\n<\/p><p><sup><a href=\"#id2921610\" id=\"ftn.id2921610\">[71]<\/a><\/sup>\n  A common error is to use\n  <code>$*<\/code> rather than \u201c\n  <code>$@<\/code> \u201d. This does bad things when handed a filename with\nembedded spaces.\n<\/p><p><sup><a href=\"#id2921897\" id=\"ftn.id2921897\">[72]<\/a><\/sup><em>qmail-popup<\/em> 's standard\ninput and standard output are the socket, and standard error (which\nwill be file descriptor 2) goes to a log file.  File descriptor 3 is\nguaranteed to be the next to be allocated. As an infamous kernel comment\nonce observed: \u201cYou are not expected to understand this\u201d.\n<\/p><p><sup><a href=\"#id2922127\" id=\"ftn.id2922127\">[73]<\/a><\/sup>\n  The friend who suggested this case study comments: \u201cYes,\nyou can get away with this technique...if there are just a few\neasily-recognizable nuggets of information coming back from the slave\nprocess, and you have tongs and a radiation suit\u201d.\n<\/p><p><sup><a href=\"#id2922208\" id=\"ftn.id2922208\">[74]<\/a><\/sup>\n  A particularly nasty\nvariant of this attack is to drop in a named Unix-domain socket where\nthe producer and consumer programs are expecting the tempfile to\nbe.\n<\/p><p><sup><a href=\"#id2922479\" id=\"ftn.id2922479\">[75]<\/a><\/sup>\n  A \u2018race condition\u2019 is a class\nof problem in which correct behavior of the system relies on two\nindependent events happening in the right order, but there is no\nmechanism for ensuring that they actually will.  Race conditions\nproduce intermittent, timing-dependent problems that can be devilishly\ndifficult to debug.\n<\/p><\/div>\n\n<h3 id=\"ch07s03\"><a href=\"#ch07s03\">\u00a7<\/a>Problems and Methods to Avoid<\/h3>\n<p>\n  While BSD-style sockets\n  over\nTCP\/IP\n  have become the\ndominant IPC method under Unix, there are still live controversies\nover the right way to partition by multiprogramming.  Some obsolete\nmethods have not yet completely died, and some techniques of\nquestionable utility have been imported from other operating systems\n(often in association with graphics or GUI programming).  We'll be\ntouring some dangerous swamps here; beware the crocodiles.\n<\/p>\n<h4 id=\"id2923376\"><a href=\"#id2923376\">\u00a7<\/a>Obsolescent Unix IPC Methods<\/h4>\n<p>\n  Unix (born 1969) long predates\nTCP\/IP\n  (born 1980) and\nthe ubiquitous networking of the 1990s and later.  Anonymous pipes,\nredirection, and shellout have been in Unix since very early days, but\nthe history of Unix is littered with the corpses of APIs tied to\nobsolescent IPC and networking models, beginning with the\n  <code>mx()<\/code> facility that appeared in Version 6 (1976)\nand was dropped before Version 7 (1979).\n<\/p>\n<p>\n  Eventually BSD sockets\n  won out as IPC was unified with\nnetworking.  But this didn't happen until after fifteen years of\nexperimentation that left a number of relics behind.  It's useful to\nknow about these because there are likely to be references to them in\nyour Unix documentation that might give the misleading impression that\nthey're still in use. These obsolete methods are described in more detail\nin\n  <em>Unix Network Programming<\/em> <sup><a href=\"#Stevens90\" title=\"[Stevens90]\">[Stevens90]<\/a><\/sup>.\n<\/p>\n <blockquote>\n  <p>\n   The real explanation for all the dead IPC facilities in old AT&amp;T\nUnixes was politics.  The Unix Support Group was headed by a low-level\nmanager, while some projects that used Unix were headed by vice\npresidents.  They had ways to make irresistible requests, and would\nnot brook the objection that most IPC mechanisms are\ninterchangeable.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<h5 id=\"messages\"><a href=\"#messages\">\u00a7<\/a>System V IPC<\/h5>\n<p>\n  The System V IPC facilities are message-passing facilities based\non the System V shared memory facility we described earlier.\n<\/p>\n<p>\n  Programs that cooperate using System V IPC usually define\nshared protocols based on exchanging short (up to 8K) binary messages.\nThe relevant manual pages are\nmsgctl(2)\nand friends.  As this style has been largely superseded by text\nprotocols passed between sockets, we do not give an example\nhere.\n<\/p>\n<p>\n  The System V IPC facilities are present in\nLinux\n  and other modern Unixes. However, as they are a legacy feature, they\nare not exercised very often. The Linux version is still known to have\nbugs as of mid-2003.  Nobody seems to care enough to fix them.\n<\/p>\n<h5 id=\"id2923528\"><a href=\"#id2923528\">\u00a7<\/a>Streams<\/h5>\n<p>\n  Streams networking was invented for Unix Version 8 (1985) by\nDennis Ritchie. A re-implementation called STREAMS (yes,\nit is all-capitals in the documentation) first became available in the\n3.0 release of System V Unix (1986).  The STREAMS facility provided a\nfull-duplex interface (functionally not unlike a BSD\nsocket, and like\nsockets, accessible through normal\nread(2)\nand\nwrite(2)\noperations after initial setup) between a user process and a specified\ndevice driver in the kernel.  The device driver might be hardware such\nas a serial or network card, or it might be a software-only\npseudodevice set up to pass data between user processes.\n<\/p>\n<p>\n  An interesting feature of both streams and\nSTREAMS<sup><a href=\"#ftn.id2923595\" id=\"id2923595\">[76]<\/a><\/sup>\n  is that it is possible to push\nprotocol-translation modules into the kernel's processing path, so\nthat the device the user process \u2018sees\u2019 through the\nfull-duplex channel is actually filtered.  This capability could be\nused, for example, to implement a line-editing protocol for a terminal\ndevice.  Or one could implement protocols such as IP or TCP without\nwiring them directly into the kernel.\n<\/p>\n<p>\n  Streams originated as an attempt to clean up a messy feature of the\nkernel called \u2018line disciplines\u2019 \u2014 alternative modes\nof processing character streams coming from serial terminals and early\nlocal-area networks.  But as serial terminals faded from view,\nEthernet LANs became ubiquitous, and TCP\/IP drove out other protocol\nstacks and migrated into Unix kernels, the extra flexibility provided\nby STREAMS had less and less utility.  In 2003, System V\nUnix\n  still\nsupports STREAMS, as do some System V\/BSD\n  hybrids such as Digital Unix and\nSun Microsystems' Solaris.\n<\/p>\n<p>\n  Linux\n  and other open-source Unixes have\neffectively discarded STREAMS. Linux kernel modules and libraries are\navailable from the\n  <a href=\"http:\/\/www.gcom.com\/home\/linux\/lis\/\">\n   LiS<\/a>\n  project, but (as\nof mid-2003) are not integrated into the stock Linux kernel.  They\nwill not be supported under non-Unix operating systems.\n<\/p>\n<h4 id=\"id2923675\"><a href=\"#id2923675\">\u00a7<\/a>Remote Procedure Calls<\/h4>\n<p>\n  Despite occasional exceptions such as NFS (Network File System)\nand the GNOME project, attempts to import CORBA, ASN.1, and other\nforms of remote-procedure-call interface have largely failed \u2014\nthese technologies have not been naturalized into the Unix\nculture.\n<\/p>\n<p>\n  There seem to be several underlying reasons for this.  One is that RPC interfaces are not readily discoverable; that\n  is, it is difficult to query these interfaces for their capabilities, and difficult to monitor them in action without\n  building single-use tools as complex as the programs being monitored (we examined some of the reasons for this in <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\"> Chapter 6<\/a>). They have the same version skew problems\n  as libraries, but those problems are harder to track because they're distributed and not generally obvious at link\n  time.\n<\/p>\n<p>\n  As a related issue, interfaces that have richer type signatures\nalso tend to be more complex, therefore more brittle. Over time, they\ntend to succumb to ontology creep as the inventory of types that get\npassed across interfaces grows steadily larger and the individual types\nmore elaborate.  Ontology creep is a problem because structs are more\nlikely to mismatch than strings; if the ontologies of the programs on\neach side don't exactly match, it can be very hard to teach them to\ncommunicate at all, and fiendishly difficult to resolve bugs.  The\nmost successful RPC applications, such as the Network File System, \nare those in which the application domain naturally has only a\nfew simple data types.\n<\/p>\n<p>\n  The usual argument for RPC is that it permits\n\u201cricher\u201d interfaces than methods like text streams\n\u2014 that is, interfaces with a more elaborate and\napplication-specific ontology of data types. But the Rule of\nSimplicity applies!  We observed in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  that one of the functions of interfaces\nis as choke points that prevent the implementation details of modules\nfrom leaking into each other.  Therefore, the main argument in favor\nof RPC is also an argument that it increases global complexity rather\nthan minimizing it.\n<\/p>\n<p>\n  With classical RPC, it's too easy to do things in a complicated\nand obscure way instead of keeping them simple.  RPC seems to\nencourage the production of large, baroque, over-engineered systems with\nobfuscated interfaces, high global complexity, and serious\nversion-skew and reliability problems \u2014 a perfect example of\nthick glue layers run amok.\n<\/p>\n<p>\n  Windows COM and DCOM are perhaps the archetypal examples of how\nbad this can get, but there are plenty of others.  Apple abandoned\nOpenDoc, and both CORBA and the once wildly hyped Java\nRMI\n  have\nreceded from view in the Unix world as people have gained field\nexperience with them.  This may well be because these methods don't\nactually solve more problems than they cause.\n<\/p>\n<p>\n  Andrew S. Tanenbaum and Robbert van Renesse have given us a\ndetailed analysis of the general problem in\n  <em>A Critique of the Remote Procedure Call Paradigm<\/em><sup><a href=\"#Tanenbaum-VanRenesse\" title=\"[Tanenbaum-VanRenesse]\">[Tanenbaum-VanRenesse]<\/a><\/sup>, a paper which should serve as a\nstrong cautionary note to anyone considering an architecture based on\nRPC.\n<\/p>\n<p>\n  All these problems may predict long-term difficulties for the\nrelatively few Unix projects that use RPC.  Of these projects, perhaps\nthe best known is the GNOME desktop effort.<sup><a href=\"#ftn.id2923816\" id=\"id2923816\">[77]<\/a><\/sup>\n  These problems also contribute to the\nnotorious security vulnerabilities of exposing NFS servers.\n<\/p>\n<p>\n  Unix tradition, on the other hand, strongly favors\ntransparent\n  and\ndiscoverable\n  interfaces. This is one of the forces behind the Unix culture's\ncontinuing attachment to IPC through textual protocols.  It is often\nargued that the parsing overhead of textual protocols is a performance\nproblem relative to binary RPCs \u2014 but RPC interfaces tend to\nhave latency problems that are far worse, because (a) you can't\nreadily anticipate how much data marshaling and unmarshaling a given\ncall will involve, and (b) the RPC model tends to encourage\nprogrammers to treat network transactions as cost-free.  Adding even\none additional round trip to a transaction interface tends to add\nenough network latency to swamp any overhead from parsing or\nmarshaling.\n<\/p>\n<p>\n  Even if text streams were less efficient than RPC, the\nperformance loss would be marginal and linear, the kind better\naddressed by upgrading your hardware than by expending development\ntime or adding architectural complexity. Anything you might lose in\nperformance by using text streams, you gain back in the ability to\ndesign systems that are simpler \u2014 easier to monitor, to model,\nand to understand.\n<\/p>\n<p>\n  Today, RPC and the Unix attachment to text streams are\nconverging in an interesting way, through protocols like XML-RPC and\nSOAP. These, being textual and transparent, are more palatable to Unix\nprogrammers than the ugly and heavyweight binary serialization formats\nthey replace.  While they don't solve all the more general\nproblems pointed out by Tanenbaum and van Renesse, they do in some\nways combine the advantages of both text-stream and RPC worlds.\n<\/p>\n<h4 id=\"id2923889\"><a href=\"#id2923889\">\u00a7<\/a>Threads \u2014 Threat or Menace?<\/h4>\n<p>\n  Though Unix developers have long been comfortable with\ncomputation by multiple cooperating processes, they do not have a\nnative tradition of using threads (processes that share their entire\naddress spaces).  These are a recent import from elsewhere, and the\nfact that Unix programmers generally dislike them is not merely\naccident or historical contingency.\n<\/p>\n<p>\n  From a complexity-control point of view, threads are a bad\nsubstitute for lightweight processes with their own address spaces;\nthe idea of threads is native to operating systems with expensive\nprocess-spawning and weak IPC facilities.\n<\/p>\n<p>\n  By definition, though daughter threads of a process typically\nhave separate local-variable stacks, they share the same global\nmemory. The task of managing contentions and critical regions in this\nshared address space is quite difficult and a fertile source of global\ncomplexity and bugs. It can be done, but as the complexity of one's\nlocking regime rises, the chance of races and deadlocks due to\nunanticipated interactions rises correspondingly.\n<\/p>\n<p>\n  Threads are a fertile source of bugs because they can too easily\nknow too much about each others' internal states.  There is no\nautomatic encapsulation, as there would be between processes with\nseparate address spaces that must do explicit IPC to communicate.\nThus, threaded programs suffer from not just ordinary contention\nproblems, but from entire new categories of timing-dependent bugs that\nare excruciatingly difficult to even reproduce, let alone fix.\n<\/p>\n<p>\n  Thread developers have been waking up to this problem.  Recent\nthread implementations and standards show an increasing concern with\nproviding thread-local storage, which is intended to limit problems\narising from the shared global address space.  As threading APIs move\nin this direction, thread programming starts to look more and more\nlike a controlled use of shared memory.\n<\/p>\n <blockquote>\n  <p>\n   Threads often prevent abstraction.  In order to prevent\ndeadlock, you often need to know how and if the library you are using\nuses threads in order to avoid deadlock problems.  Similarly, the use\nof threads in a library could be affected by the use of threads at the\napplication layer.\n  <\/p><div>\n  \n   \u2013 David Korn\n  \n <\/div><\/blockquote>\n<p>\n  To add insult to injury, threading has performance costs that\nerode its advantages over conventional process partitioning.  While\nthreading can get rid of some of the overhead of rapidly switching\nprocess contexts, locking shared data structures so threads won't step\non each other can be just as expensive.\n<\/p>\n <blockquote>\n  <p>\n   The X server, able to execute literally millions of ops\/second,\nis\n   <em>not<\/em>\n   threaded; it uses a poll\/select loop.\nVarious efforts to make a multithreaded implementation have come to no\ngood result.  The costs of locking and unlocking get too high for\nsomething as performance-sensitive as graphics servers.\n  <\/p><div>\n  \n   \u2013 Jim Gettys\n  \n <\/div><\/blockquote>\n<p>\n  This problem is fundamental, and has also been a continuing issue\nin the design of Unix kernels for symmetric multiprocessing.\nAs your resource-locking gets finer-grained, latency due to locking\noverhead can increase fast enough to swamp the gains from locking\nless core memory.\n<\/p>\n<p>\n  One final difficulty with threads is that threading standards\nstill tend to be weak and underspecified as of\nmid-2003. Theoretically conforming libraries for Unix standards\nsuch as POSIX\n  threads\n(1003.1c) can nevertheless exhibit alarming differences in behavior\nacross platforms, especially with respect to signals, interactions\nwith other IPC methods, and resource cleanup times.  Windows and\nclassic MacOS have native threading models and interrupt facilities\nquite different from those of Unix and will often require considerable\nporting effort even for simple threading cases.  The upshot is that\nyou cannot count on threaded programs to be portable.\n<\/p>\n<p>\n  For more discussion and a lucid contrast with event-driven\nprogramming, see\n  <em>Why Threads Are a Bad Idea<\/em> <sup><a href=\"#Ousterhout96\" title=\"[Ousterhout96]\">[Ousterhout96]<\/a><\/sup>.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2923595\" id=\"ftn.id2923595\">[76]<\/a><\/sup>\n  STREAMS was much more complex. Dennis Ritchie\nis reputed to have said \u201cStreams means something different when\nshouted\u201d.\n<\/p><p><sup><a href=\"#id2923816\" id=\"ftn.id2923816\">[77]<\/a><\/sup>\n  GNOME's main\ncompetitor, KDE, started with CORBA but abandoned it in their 2.0\nrelease. They have been on a quest for lighter-weight IPC methods ever\nsince.\n<\/p><\/div>\n\n<h3 id=\"ch07s04\"><a href=\"#ch07s04\">\u00a7<\/a>Process Partitioning at the Design Level<\/h3>\n<p>\n  Now that we have all these methods, what should we do with them?\n<\/p>\n<p>\n  The first thing to notice is that tempfiles, the more\ninteractive sort of master\/slave process relationship, sockets, RPC,\nand all other methods of bidirectional IPC are at some level\nequivalent \u2014 they're all just ways for programs to exchange data\nduring their lifetimes.  Much of what we do in a sophisticated way\nusing sockets or shared memory we could do in a primitive way using\ntempfiles as mailboxes and signals for notification.  The differences\nare at the edges, in how programs establish communication, where and\nwhen one does the marshalling and unmarshalling of messages, in what\nsorts of buffering problems you may have, and atomicity guarantees you\nget on the messages (that is, to what extent you can know that the\nresult of a single send action from one side will show up as a single\nreceive event on the other).\n<\/p>\n<p>\n  We've seen from the PostgreSQL study that one effective way to\nhold down complexity is to break an application into a client\/server\npair. The PostgreSQL client and server communicate through an application\nprotocol over sockets, but very little about the design pattern would\nchange if they used any other bidirectional IPC method.\n<\/p>\n<p>\n  This kind of partitioning is particularly effective in\nsituations where multiple instances of the application must manage\naccess to resources that are shared among all. A single server process\nmay manage all resource contention, or cooperating peers may each take\nresponsibility for some critical resource.\n<\/p>\n<p>\n  Client-server partitioning can also help distribute cycle-hungry\napplications across multiple hosts.  Or it may make them suitable for\ndistributed computing across the Internet (as with Freeciv). We'll\ndiscuss the related\n  <em>CLI server<\/em> pattern in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>.\n<\/p>\n<p>\n  Because all these peer-to-peer IPC techniques are alike at some\nlevel, we should evaluate them mainly on the amount of\nprogram-complexity overhead they incur, and how much opacity they\nintroduce into our designs.  This, ultimately, is why BSD sockets have\nwon over other Unix IPC methods, and why RPC has generally failed to\nget much traction.\n<\/p>\n<p>\n  Threads are fundamentally different.  Rather than supporting\ncommunication among different programs, they support a sort of\ntimesharing within an instance of a single program.  Rather\nthan being a way to partition a big program into smaller ones \nwith simpler behavior, threading is strictly a performance hack. It\nhas all the problems normally associated with performance hacks, and\na few special ones of its own.\n<\/p>\n<p>\n  Accordingly, while we should seek ways to break up large\nprograms into simpler cooperating processes, the use of threads within\nprocesses should be a last resort rather than a first.  Often, you may\nfind you can avoid them.  If you can use limited shared memory and\nsemaphores, asynchronous I\/O using\n  <code>SIGIO<\/code>, or\npoll(2)\/select(2)\nrather than threading, do it that way.  Keep it simple; use techniques\nearlier on this list and lower on the complexity scale in preference\nto later ones.\n<\/p>\n<p>\n  The combination of threads, remote-procedure-call interfaces, and\nheavyweight object-oriented design is especially dangerous.  Used\nsparingly and tastefully, any of these techniques can be valuable \u2014\nbut if you are ever invited onto a project that is supposed to feature\nall three, fleeing in terror might well be an appropriate\nreaction.\n<\/p>\n<p>\n  We have previously observed that programming in the real world\nis all about managing complexity.  Tools to manage complexity are good\nthings.  But when the effect of those tools is to proliferate\ncomplexity rather than to control it, we would be better off throwing them\naway and starting from zero.  An important part of the Unix wisdom \nis to never forget this.\n<\/p>\n\n<h3 id=\"minilanguageschapter\"><a href=\"#minilanguageschapter\">\u00a7<\/a>Chapter 8. Minilanguages<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   A good notation has a subtlety and suggestiveness which at times \nmakes it almost seem like a live teacher.\n  <\/p><div>\n  \n   \u2013 Bertrand Russell\nThe World of Mathematics (1956)\n  \n <\/div><\/blockquote>\n<p>\n  One of the most consistent results from large-scale studies of\nerror patterns in software is that programmer error rates in defects per \nhundreds of lines are largely independent of the language in which\nthe programmers are coding.<sup><a href=\"#ftn.id2930951\" id=\"id2930951\">[78]<\/a><\/sup>\n  Higher-level languages, which allow\nyou to get more done in fewer lines, mean fewer bugs as well.\n<\/p>\n<p>\n  Unix has a long tradition of hosting little languages\nspecialized for a particular application domain, languages that can\nenable you to drastically reduce the line count of your programs.\nDomain-specific language examples include the numerous Unix\ntypesetting languages (\n  <em>troff<\/em>,\n  <em>eqn<\/em>,\n  <em>tbl<\/em>,\n  <em>pic<\/em>,\n  <em>grap<\/em>),\nshell utilities (\n  <em>awk<\/em>,\n  <em>sed<\/em>,\n  <em>dc<\/em>,\n  <em>bc<\/em>), and software development tools\n(\n  <em>make<\/em>,\n  <em>yacc<\/em>,\n  <em>lex<\/em>).  There is a fuzzy boundary between\ndomain-specific languages and the more flexible sort of application\nrun-control file (\n  <em>sendmail<\/em>, BIND, X);\nanother with data-file formats; and another with scripting\nlanguages\n  (which we'll survey in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>).\n<\/p>\n<p>\n  Historically, domain-specific languages of this kind have been\ncalled \u2018little languages\u2019 or \u2018minilanguages\u2019\nin the Unix world, because early examples were small and low in\ncomplexity relative to general-purpose languages (all three terms for\nthe category are in common use). But if the application domain is\ncomplex (in that it has lots of different primitive operations or\ninvolves manipulation of intricate data structures), an application\nlanguage for it may have to be rather more complex than some\ngeneral-purpose languages.  We'll keep the traditional term\n\u2018minilanguage\u2019 to emphasize that the wise course\nis usually to keep these designs as small and simple as\npossible.\n<\/p>\n<p>\n  The domain-specific little language is an extremely powerful\ndesign idea.  It allows you to define your own higher-level language\nto specify the appropriate methods, rules, and algorithms for the task\nat hand, reducing global complexity relative to a design that uses\nhardwired lower-level code for the same ends.  You can get to a\nminilanguage design in at least three ways, two of them good and one\nof them dangerous.\n<\/p>\n<p>\n  One right way to get there is to realize up front that you can\nuse a minilanguage design to push a given specification of a programming\nproblem up a level, into a notation that is more compact and\nexpressive than you could support in a general-purpose language.  As\nwith code generation and data-driven programming, a minilanguage lets\nyou take practical advantage of the fact that the defect rate in your\nsoftware will be largely independent of the level of the language you\nare using; more expressive languages mean shorter programs and fewer\nbugs.\n<\/p>\n<p>\n  The second right way to get to a minilanguage design is to\nnotice that one of your specification file formats is looking more and\nmore like a minilanguage \u2014 that is, it is developing complex\nstructures and implying actions in the application you are\ncontrolling.  Is it trying to describe control flow as well as data\nlayouts?  If so, it may be time to promote that control flow from\nbeing implicit to being explicit in your specification\nlanguage.\n<\/p>\n<p>\n  The wrong way to get to a minilanguage design is to extend your\nway to it, one patch and crufty added feature at a time. On this path,\nyour specification file keeps sprouting more implied control flow and\nmore tangled special-purpose structures until it has become an ad-hoc\nlanguage without your noticing it.  Some legendary nightmares have\nbeen spawned this way; the example every Unix guru will think of (and\nshudder over) is the\n  <code>sendmail.cf<\/code> configuration\nfile associated with the\n  <em>sendmail<\/em> mail\ntransport.\n<\/p>\n<p>\n  Sadly, most people do their first minilanguage the wrong way,\nand only realize later what a mess it is.  Then the question is: how\nto clean it up?  Sometimes the answer implies rethinking the entire\napplication design. Another notorious example of language-by-feature\ncreep was the editor\n  <em>TECO<\/em>, which grew first\nmacros and then loops and conditionals as programmers wanted to use it\nto package increasingly complex editing routines. The resulting\nugliness was eventually fixed by a redesign of the entire editor to be\nbased on an intentional language; this is how Emacs\nLisp\n  (which we'll survey below) evolved.\n<\/p>\n<p>\n  All sufficiently complicated specification files aspire to the\ncondition of minilanguages.  Therefore, it will often be the case that\nyour only defense against designing a bad minilanguage is knowing how\nto design a good one. This need not be a huge step or involve knowing\na lot of formal language theory; with modern tools, learning a few\nrelatively simple techniques and bearing good examples in mind as you\ndesign should be sufficient.\n<\/p>\n<p>\n  In this chapter we'll examine all the kinds of minilanguages\nnormally supported under Unix, and try to identify the kinds of\nsituation in which each of them represents an effective design\nsolution.  This chapter is not meant to be an exhaustive catalog of\nUnix languages, but rather to bring out the design principles involved\nin structuring an application around a minilanguage.  We'll have much\nmore to say about languages for general-purpose programming in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>.\n<\/p>\n<p>\n  We'll need to start by doing a little taxonomy, so we'll\nknow what we're talking about later on.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2930951\" id=\"ftn.id2930951\">[78]<\/a><\/sup>\n  Les Hatton reports by email on the analysis in his book in\npreparation,\n  <em>Software Failure<\/em> : \u201cProvided\nyou use executable line counts for the density measure, the injected\ndefect densities vary less between languages than they do between\nengineers by about a factor of 10\u201d.\n<\/p><\/div>\n\n<h3 id=\"ch08s01\"><a href=\"#ch08s01\">\u00a7<\/a>Understanding the Taxonomy of Languages<\/h3>\n<p>\n  All the languages in Figure 8.1 are described in case studies, either in this chapter or elsewhere in this book.  For\n  the general-purpose interpreters near the right-hand side, see\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">Chapter 14<\/a>.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 8.1. Taxonomy of languages.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Taxonomy of languages.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  In\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  we\nlooked at Unix conventions for data files.  There's a spectrum of\ncomplexity in these.  At the low end are files that make simple\nassociations between names and properties; the\n  <code>\/etc\/passwd<\/code> and\n  <code>.newsrc<\/code> formats are good examples.  Further up the\nscale we start to get formats that marshal or serialize data\nstructures; the PNG\n  and\nSNG formats are (equivalent) good examples of this.\n<\/p>\n<p>\n  A structured data-file format starts to border on being a\nminilanguage when it expresses not just structure but actions\nperformed on some interpretive context (that is, memory that is\noutside the data file itself).  XML markups tend to straddle this\nborder; the example we'll look at here is\n  <em>Glade<\/em>, a code generator for building GUI\ninterfaces. Formats that are both designed to be read and written by\nhumans (rather than just programs) and are used to generate code, are\nfirmly in the realm of minilanguages.\n  <em>yacc<\/em> and\n  <em>lex<\/em> are the classic examples.  We'll\ndiscuss\n  <em>glade<\/em>,\n  <em>yacc<\/em> and\n  <em>lex<\/em> in\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>.\n<\/p>\n<p>\n  The Unix macro processor,\n  <em>m4<\/em>, is\nanother very simple declarative minilanguage (that is, one in which\nthe program is expressed as a set of desired relationships or\nconstraints rather than explicit actions).  It has often been used as a\npreprocessing stage for other minilanguages.\n<\/p>\n<p>\n  Unix makefiles, which are designed to automate build processes,\nexpress dependency relationships between source and derived files<sup><a href=\"#ftn.id2924480\" id=\"id2924480\">[79]<\/a><\/sup>\n  and the commands required to make each derived\nfile from its sources.\nWhen you run make, it uses those declarations to walk the implied tree\nof dependencies, doing the least work necessary to bring your build up\nto date.  Like\n  <em>yacc<\/em> and\n  <em>lex<\/em> specifications, makefiles are a\ndeclarative minilanguage; they set up constraints that imply actions\nperformed on an interpretive context (in this case, the portion of the\nfile system where the source and generated files live).  We'll return\nto makefiles in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>.\n<\/p>\n<p>\n  XSLT, the language used to describe transformations of XML,\nis at the high end of complexity for declarative minilanguages.  It's\ncomplex enough that it's not normally thought of as a minilanguage \nat all, but it shares some important characteristic of such languages\nwhich we'll examine when we look at it in more detail below.\n<\/p>\n<p>\n  The spectrum of minilanguages ranges from declarative (with\nimplicit actions) to imperative (with explicit actions).  The\nrun-control syntax of\nfetchmail(1)\n  can be viewed as either a very weak imperative language or a\ndeclarative language with implied control flow.  The troff and\nPostScript typesetting languages are imperative languages with a lot\nof special-purpose domain expertise baked into them.\n<\/p>\n<p>\n  Some task-specific imperative minilanguages start to border on\nbeing general-purpose interpreters.  They reach this level when they\nare explicitly\n  <em>Turing-complete<\/em> \u2014that is,\nthey can do both conditionals and loops (or recursion)<sup><a href=\"#ftn.id2924589\" id=\"id2924589\">[80]<\/a><\/sup>\n  with features that are\ndesigned to be used as control structures.  Some languages, by\ncontrast, are only accidentally Turing-complete \u2014 they have\nfeatures that can be used to implement control structures as a \nsort of side effect of what they are actually designed to do.\n<\/p>\n<p>\n  The\nbc(1)\nand\ndc(1)\ninterpreters we looked at in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  are good examples of specialized\nimperative minilanguages that are explicitly Turing-complete.\n<\/p>\n<p>\n  We are over the border into general-purpose interpreters when we\nreach languages like Emacs Lisp\n  and\nJavaScript\n  that are designed to be full\nprogramming languages run in specialized contexts.  We'll have more to\nsay about these when we discuss embedded scripting\nlanguages\n  later\non.\n<\/p>\n<p>\n  The spectrum in interpreters is one of increasing generality;\nthe flip side of this is that a more general-purpose interpreter\nembodies fewer assumptions about the context in which it runs.  With\nincreasing generality there usually comes a richer ontology of data\ntypes.  Shell and Tcl\n  have relatively simple ontologies;\nPerl,\nPython, and\nJava\n  more\ncomplex ones.  We'll return to these general-purpose languages in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2924480\" id=\"ftn.id2924480\">[79]<\/a><\/sup>\n  For less technical readers: the compiled form of a\nC\n  \n  program is\nderived from its C source form by compilation and linkage.  The\nPostScript version of a troff document is derived from the troff\nsource; the command to make the former from the latter is a\n  <em>troff<\/em> invocation.  There are many other\nkinds of derivation; makefiles can express almost all of\nthem.\n<\/p><p><sup><a href=\"#id2924589\" id=\"ftn.id2924589\">[80]<\/a><\/sup>\n  Any Turing-complete language could theoretically be used for\ngeneral-purpose programming, and is theoretically exactly as powerful\nas any other Turing-complete language.  In practice, some\nTuring-complete languages would be far too painful to use for anything\noutside a specified and narrow problem domain.\n<\/p><\/div>\n\n<h3 id=\"ch08s02\"><a href=\"#ch08s02\">\u00a7<\/a>Applying Minilanguages<\/h3>\n<p>\n  Designing with minilanguages involves two distinct challenges.\nOne is having existing minilanguages handy in your toolkit, and\nrecognizing when they can be applied as-is.  The other is knowing when\nit is appropriate to design a custom minilanguage for an application.\nTo help you develop both aspects of your design sense, about half of \nthis chapter will consist of case studies.\n<\/p>\n<h4 id=\"id2924747\"><a href=\"#id2924747\">\u00a7<\/a>Case Study: sng<\/h4>\n<p>\n  In\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>\n  we\nlooked at\nsng(1),\nwhich translates between\nPNG\n  and an editable\nall-text representation of the same bits.  The SNG data-file format is\nworth reexamining for contrast here because it is not quite a\ndomain-specific minilanguage.  It describes a data layout, but doesn't\nassociate any implied sequence of actions with the data.\n<\/p>\n<p>\n  SNG does, however, share one important characteristic with\ndomain-specific minilanguages that binary structured data formats like\nPNG do not \u2014\ntransparency.  Structured data files make\nit possible for editing, conversion, and generation tools to cooperate\nwithout knowing about each others' design assumptions other than through\nthe medium of the minilanguage.  What SNG adds is that, like a\ndomain-specific minilanguage, it's designed to be easy to parse by\neyeball and edit with general-purpose tools.\n<\/p>\n<h4 id=\"regexps\"><a href=\"#regexps\">\u00a7<\/a>Case Study: Regular Expressions<\/h4>\n<p>\n  A kind of specification that turns up repeatedly in Unix tools\nand scripting languages is the\n  <em>regular expression<\/em>\n  (\u2018regexp\u2019 for short).  We consider\nit here as a declarative minilanguage for describing text patterns; it\nis often embedded in other minilanguages.  Regexps are so ubiquitous\nthat the are hardly thought of as a minilanguage, but they replace\nwhat would otherwise be huge volumes of code implementing different\n(and incompatible) search capabilities.\n<\/p>\n<p>\n  This introduction skates over some details like POSIX\n  extensions,\nback-references, and internationalization features; for a more\ncomplete treatment, see\n  <em>Mastering Regular Expressions<\/em><sup><a href=\"#Friedl\" title=\"[Friedl]\">[Friedl]<\/a><\/sup>.\n<\/p>\n<p>\n  Regular expressions describe patterns that may either match or fail to match against strings.  The simplest\n  regular-expression tool is grep(1), a filter that passes through to its output every line in its input matching a\n  specified regexp. Regexp notation is summarized in Table 8.1.\n<\/p>\n  <table class=\"border\"><caption>Table 8.1. Regular-expression examples.<\/caption>\n   <thead><tr><th>\n      Regexp\n     <\/th>\n     <th>\n      Matches\n     <\/th>\n    <\/tr><\/thead><tbody><tr><td>\n      <code>\"x.y\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> followed by any character followed by\n      <code>y<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"x\\.y\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> followed by a literal period followed by\n      <code>y<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"xz?y\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> followed by at most one\n      <code>z<\/code> followed by\n      <code>y<\/code>; thus,\n      <code>\"xy\"<\/code> or\n      <code>\"xzy\"<\/code> but not\n      <code>\"xz\"<\/code> or\n      <code>\"xdy\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"xz*y\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> followed by any number of instances of\n      <code>z<\/code>, followed by\n      <code>y<\/code>;\n   thus,\n      <code>\"xy\"<\/code> or\n      <code>\"xzy\"<\/code> or\n      <code>\"xzzzy\"<\/code> but not\n      <code>\"xz\"<\/code> or\n      <code>\"xdy\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"xz+y\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> followed by one or more instances of\n      <code>z<\/code>, followed by\n      <code>y<\/code>;\n   thus,\n      <code>\"xzy\"<\/code> or\n      <code>\"xzzy\"<\/code> but not\n      <code>\"xy\"<\/code> or\n      <code>\"xz\"<\/code> or\n      <code>\"xdy\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"s[xyz]t\"<\/code> <\/td>\n     <td>\n      <code>s<\/code> followed by any of the characters\n      <code>x<\/code> or\n      <code>y<\/code> or\n      <code>z<\/code>, followed by\n      <code>t<\/code>;\n   thus,\n      <code>\"sxt\"<\/code> or\n      <code>\"syt\"<\/code> or\n      <code>\"szt\"<\/code> but not\n      <code>\"st\"<\/code> or\n      <code>\"sat\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"a[x0-9]b\"<\/code> <\/td>\n     <td>\n      <code>a<\/code> followed by either\n      <code>x<\/code> or characters in the range\n      <code>0<\/code> \u2013\n      <code>9<\/code>, followed by\n      <code>b<\/code>;\n   thus,\n      <code>\"axb\"<\/code> or\n      <code>\"a0b\"<\/code> or\n      <code>\"a4b\"<\/code> but not\n      <code>\"ab\"<\/code> or\n      <code>\"aab\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"s[^xyz]t\"<\/code> <\/td>\n     <td>\n      <code>s<\/code> followed by any character that is not\n      <code>x<\/code> or\n      <code>y<\/code> or\n      <code>z<\/code>, followed by\n      <code>t<\/code>;\n   thus,\n      <code>\"sdt\"<\/code> or\n      <code>\"set\"<\/code> but not\n      <code>\"sxt\"<\/code> or\n      <code>\"syt\"<\/code> or\n      <code>\"szt\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"s[^x0-9]t\"<\/code> <\/td>\n     <td>\n      <code>s<\/code> followed by any character that is not\n      <code>x<\/code> or in the range\n      <code>0<\/code> \u2013\n      <code>9<\/code>, followed by\n      <code>t<\/code>; thus,\n      <code>\"slt\"<\/code> or\n      <code>\"smt\"<\/code> but not\n      <code>\"sxt\"<\/code> or\n      <code>\"s0t\"<\/code> or\n      <code>\"s4t\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"^x\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> at the beginning of a string; \n  thus,\n      <code>\"xzy\"<\/code> or\n      <code>\"xzzy\"<\/code> but not\n      <code>\"yzy\"<\/code> or\n      <code>\"yxy\"<\/code>.\n     <\/td>\n    <\/tr><tr><td>\n      <code>\"x$\"<\/code> <\/td>\n     <td>\n      <code>x<\/code> at the end of a string; \n  thus,\n      <code>\"yzx\"<\/code> or\n      <code>\"yx\"<\/code> but not\n      <code>\"yxz\"<\/code> or\n      <code>\"zxy\"<\/code>.\n     <\/td>\n    <\/tr><\/tbody><\/table><p>\n  There are a number of minor variants of regexp notation:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     <em>Glob expressions.<\/em>\n     This is the limited set\nof wildcard conventions used by early Unix shells for filename\nmatching. There are only three wildcards:\n     <code>*<\/code>, which matches any\nsequence of characters (like\n     <code>.*<\/code> in the other variants);\n     <code>?<\/code>, which\nmatches any single character (like\n     <code>.<\/code> in the other variants); and\n     <code>[...]<\/code>, which matches a character class just as in the other variants.\nSome shells (\n     <em>csh<\/em>,\n     <em>bash<\/em>,\n     <em>zsh<\/em>) later\nadded\n     <code>{}<\/code> for alternation.  Thus,\n     <code>x{a,b}c<\/code> matches\n     <code>xac<\/code> or\n     <code>xbc<\/code> but not\n     <code>xc<\/code>. Some shells further extend globs in the direction of extended\nregular expressions.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Basic regular expressions.<\/em>\n     This is the\nnotation accepted by the original\ngrep(1)\nutility for extracting lines matching a given regexp from a file.  The\nline editor\ned(1),\nthe stream editor\nsed(1),\nalso use these.  Old Unix hands think of these as the basic or\n\u2018vanilla\u2019 flavor of regexp; people first exposed to the\nmore modern tools tend to assume the extended form described next.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Extended regular expressions.<\/em>\n     This is the\nnotation accepted by the extended grep utility\negrep(1)\nfor extracting lines matching a given regexp from a file.  Regular\nexpressions in Lex and the\n     <em>Emacs<\/em> editor are\nvery close to the\n     <em>egrep<\/em> flavor.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Perl regular expressions.<\/em>\n     This is the notation\naccepted by Perl\n     and Python\n     regexp functions.  These are quite a bit more\npowerful than the\n     <em>egrep<\/em> flavor.\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n<p>\n  Now that we've looked at some motivating examples, Table 8.2 is a summary of the standard regular-expression\n  wildcards.  Note: we're not including the glob variant in this table, so a value of \u201cAll\u201d implies only all three of\n  the basic, extended\/Emacs, and Perl\/Python variants.<sup><a href=\"#ftn.id2932978\" id=\"id2932978\">[81]<\/a><\/sup><\/p>\n  <table class=\"border\"><caption>Table 8.2. Introduction to regular-expression operations.<\/caption>\n   <thead><tr><th>\n      Wildcard\n     <\/th>\n     <th>\n      Supported in\n     <\/th>\n     <th>\n      Matches\n     <\/th>\n    <\/tr><\/thead><tbody><tr><td>\n      <code>\\<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      Escape next character.  Toggles whether following punctuation\n  is treated as a wildcard or not. Following letters or digits are interpreted\n  in various different ways depending on the program.\n     <\/td>\n    <\/tr><tr><td>\n      <code>.<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      Any character.\n     <\/td>\n    <\/tr><tr><td>\n      <code>^<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      Beginning of line\n     <\/td>\n    <\/tr><tr><td>\n      <code>$<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      End of line\n     <\/td>\n    <\/tr><tr><td>\n      <code>[...]<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      Any of the characters between the brackets\n     <\/td>\n    <\/tr><tr><td>\n      <code>[^...]<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      Any characters\n      <em>except those<\/em>\n      between the brackets.\n     <\/td>\n    <\/tr><tr><td>\n      <code>*<\/code> <\/td>\n     <td>\n      All\n     <\/td>\n     <td>\n      Accept any number of instances of the previous element.\n     <\/td>\n    <\/tr><tr><td>\n      <code>?<\/code> <\/td>\n     <td>\n      egrep\/Emacs, Perl\/Python\n     <\/td>\n     <td>\n      Accept zero or one instances of the previous element.\n     <\/td>\n    <\/tr><tr><td>\n      <code>+<\/code> <\/td>\n     <td>\n      egrep\/Emacs, Perl\/Python\n     <\/td>\n     <td>\n      Accept one or more instances of the previous element.\n     <\/td>\n    <\/tr><tr><td>\n      <code>{n}<\/code> <\/td>\n     <td>\n      egrep, Perl\/Python; as\n      <code>\\{n\\}<\/code> in Emacs\n     <\/td>\n     <td>\n      Accept exactly\n      <code>n<\/code> repetitions of the previous element.\n  Not supported by some older regexp engines.\n     <\/td>\n    <\/tr><tr><td>\n      <code>{n,}<\/code> <\/td>\n     <td>\n      egrep, Perl\/Python; as\n      <code>\\{n,\\}<\/code> in Emacs\n     <\/td>\n     <td>\n      Accept\n      <code>n<\/code> or more repetitions of the previous element.\n  Not supported by some older regexp engines.\n     <\/td>\n    <\/tr><tr><td>\n      <code>{m,n}<\/code> <\/td>\n     <td>\n      egrep, Perl\/Python; as\n      <code>\\{m,n\\}<\/code> in Emacs\n     <\/td>\n     <td>\n      Accept at least\n      <code>m<\/code> and at most\n      <code>n<\/code> repetitions of the previous \n  element. Not supported by some older regexp engines.\n     <\/td>\n    <\/tr><tr><td>\n      <code>|<\/code> <\/td>\n     <td>\n      egrep, Perl\/Python; as\n      <code>\\|<\/code> in Emacs\n     <\/td>\n     <td>\n      Accept the element to the left or the element to the right.\n  This is usually used with some form of pattern-grouping delimiters.\n     <\/td>\n    <\/tr><tr><td>\n      <code>(...)<\/code> <\/td>\n     <td>\n      Perl\/Python; as\n      <code>\\(...\\)<\/code> in older versions.\n     <\/td>\n     <td>\n      Treat this pattern as a group (in newer regexp engines like\n  Perl and Python's). Older regexp engines such as those in Emacs and grep\n  require\n      <code>\\(...\\)<\/code>.\n     <\/td>\n    <\/tr><\/tbody><\/table><p>\n  Design practice in new languages with regexp support has\nstabilized on the Perl\/Python variant. It is more transparent than the\nothers, notably because backlash before a non-alphanumeric character\nalways means that character as a literal, so there is much\nless confusion about how to quote elements of regexps.\n<\/p>\n<p>\n  Regular expressions are an extreme example of how concise a \nminilanguage can be.  Simple regular expressions express recognition\nbehavior that would otherwise have to be implenented with hundreds of\nlines of fussy, bug-prone code.\n<\/p>\n<h4 id=\"id2933450\"><a href=\"#id2933450\">\u00a7<\/a>Case Study: Glade<\/h4>\n<p>\n  <em>Glade<\/em> is an interface\nbuilder for the open-source GTK toolkit library for\nX.<sup><a href=\"#ftn.id2933471\" id=\"id2933471\">[82]<\/a><\/sup><em>Glade<\/em> allows you to develop a\nGUI interface by interactively picking, placing, and modifying widgets\non an interface panel.  The GUI editor produces an XML file describing\nthe interface; this, in turn, can be fed to one of several code\ngenerators that will actually grind out C,\nC++,\nPython\n  or\nPerl\n  code for\nthe interface.  The generated code then calls functions you write to\nsupply behavior to the interface.\n<\/p>\n<p>\n  <em>Glade<\/em> 's XML format for describing GUIs is a good example of a simple domain-specific minilanguage.  See\n  Example 8.1 for a \u201cHello, world!\u201d GUI in Glade format.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 8.1. Glade \u201cHello, World\u201d.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n&lt;?xml version=\"1.0\"?&gt;\n&lt;GTK-Interface&gt;\n&lt;widget&gt;\n  &lt;class&gt;GtkWindow&lt;\/class&gt;\n  &lt;name&gt;HelloWindow&lt;\/name&gt;\n  &lt;border_width&gt;5&lt;\/border_width&gt;\n  &lt;Signal&gt;\n    &lt;name&gt;destroy&lt;\/name&gt;\n    &lt;handler&gt;gtk_main_quit&lt;\/handler&gt;\n  &lt;\/Signal&gt;\n  &lt;title&gt;Hello&lt;\/title&gt;\n  &lt;type&gt;GTK_WINDOW_TOPLEVEL&lt;\/type&gt;\n  &lt;position&gt;GTK_WIN_POS_NONE&lt;\/position&gt;\n  &lt;allow_shrink&gt;True&lt;\/allow_shrink&gt;\n  &lt;allow_grow&gt;True&lt;\/allow_grow&gt;\n  &lt;auto_shrink&gt;False&lt;\/auto_shrink&gt;\n  &lt;widget&gt;\n    &lt;class&gt;GtkButton&lt;\/class&gt;\n    &lt;name&gt;Hello World&lt;\/name&gt;\n    &lt;can_focus&gt;True&lt;\/can_focus&gt;\n    &lt;Signal&gt;\n      &lt;name&gt;clicked&lt;\/name&gt;\n      &lt;handler&gt;gtk_widget_destroy&lt;\/handler&gt;\n      &lt;object&gt;HelloWindow&lt;\/object&gt;\n    &lt;\/Signal&gt;\n    &lt;label&gt;Hello World&lt;\/label&gt;\n  &lt;\/widget&gt;\n&lt;\/widget&gt;\n&lt;\/GTK-Interface&gt;\n<\/pre><\/div>\n<p>\n  A valid specification in\n  <em>Glade<\/em> format implies a repertoire of actions\nby the GUI in response to user behavior.  The\n  <em>Glade<\/em> GUI treats these specifications as\nstructured data files;\n  <em>Glade<\/em> code\ngenerators, on the other hand, use them to write programs implementing\na GUI.  For some languages (including Python), there are runtime libraries that\nallow you to skip the code-generation step and simply instantiate the\nGUI directly at runtime from the XML specification (interpreting Glade\nmarkup, rather than compiling it to the host language).  Thus, you get\nthe choice of trading space efficiency for startup speed or\nvice versa.\n<\/p>\n<p>\n  Once you get past the verbosity of XML,\n  <em>Glade<\/em> markup is a fairly simple language.\nIt does just two things: declare GUI-widget hierarchies and associate\nproperties with widgets.  You don't actually have to know a lot about\nhow\n  <em>glade<\/em> works to read the specification\nabove.  In fact, if you have any experience programming in GUI\ntoolkits, reading it will immediately give you a fairly good\nvisualization of what\n  <em>glade<\/em> does with the\nspecification.  (Hands up everyone who predicted that this particular\nspecification will give you a single button widget in a window\nframe.)\n<\/p>\n<p>\n  This kind of transparency and simplicity is the mark of a good\nminilanguage\ndesign.  The mapping between the notation and\ndomain objects is very clear.  The relationships between objects are\nexpressed directly, rather than through name references or some other\nsort of indirection that you have to think to follow.\n<\/p>\n<p>\n  The ultimate functional test of a minilanguage like this one is\nsimple: can I hack it without reading the manual?  For a significant\nrange of cases, the\n  <em>Glade<\/em> answer\nis yes.  For example, if you know the C-level constants that GTK uses\nto describe window-positioning hints, you'll recognize\n  <code>GTK_WIN_POS_NONE<\/code> as one and instantly be able to\nchange the positioning hint associated with this GUI.\n<\/p>\n<p>\n  The advantage of using\n  <em>Glade<\/em> should be clear.  It specializes in\ncode generation so you don't have to.  That's one less routine task\nyou have to hand-code, and one fewer source of hand-coded bugs.\n<\/p>\n<p>\n  More information, including source code and documentation and\nlinks to sample applications, is available at the\n  <a href=\"http:\/\/glade.gnome.org\/\">\n   Glade project page<\/a>.\n  <em>Glade<\/em> has been ported to\nWindows.\n<\/p>\n<h4 id=\"id2933775\"><a href=\"#id2933775\">\u00a7<\/a>Case Study: m4<\/h4>\n<p>\n  The\nm4(1)\nmacro processor interprets a declarative minilanguage for describing\ntransformations of text.  An\n  <em>m4<\/em> program is\na set of macros that specifies ways to expand text strings into other\nstrings.  Applying those declarations to an input text with\n  <em>m4<\/em> performs macro expansion and yields an output text.  (The\nC\n  preprocessor performs similar services for C compilers,\nthough in a rather different style.)\n<\/p>\n<p>\n   Example 8.2 shows an <em>m4<\/em> macro that directs <em>m4<\/em> to expand each occurrence of the string \"OS\" in its\n   input into the string \"operating system\" on output.  This is a trivial example; <em>m4<\/em> supports macros with\n   arguments that can be used to do more than transform one fixed string into another.  Typing <code>info m4<\/code> at\n   your shell prompt will probably display on-line documentation for this language.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 8.2. A sample\n    <em>m4<\/em> macro.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\ndefine(`OS', `operating system')\n<\/pre><\/div>\n<p>\n  The\n  <em>m4<\/em> macro language supports\nconditionals and recursion.  The combination can be used to implement\nloops, and this was intended;\n  <em>m4<\/em> is\ndeliberately Turing-complete.  But actually trying to use\n  <em>m4<\/em> as a general-purpose language would be\ndeeply perverse.\n<\/p>\n<p>\n  The\n  <em>m4<\/em> macro processor is usually\nemployed as a preprocessor for minilanguages that lack a built-in\nnotion of named procedures or a built-in file-inclusion feature.  It's\nan easy way to extend the syntax of the base language so the\ncombination with\n  <em>m4<\/em> supports both these\nfeatures.\n<\/p>\n<p>\n  One well-known use of\n  <em>m4<\/em> has been to\nclean up (or at least effectively hide) another minilanguage design\nthat was called out as a bad example earlier in this chapter.  Most\nsystem administrators now generate their\n  <code>sendmail.cf<\/code> configuration files using an\n  <em>m4<\/em> macro package supplied with the\n  <em>sendmail<\/em> distribution.  The macros\nstart from feature names (or name\/value pairs) and generate the\ncorresponding (much uglier) strings in the\n  <em>sendmail<\/em> configuration language.\n<\/p>\n<p>\n  Use\n  <em>m4<\/em> with caution, however.  Unix\nexperience has taught minilanguage designers to be wary of macro\nexpansion,<sup><a href=\"#ftn.id2934007\" id=\"id2934007\">[83]<\/a><\/sup>\n  for reasons we'll discuss\n  <a href=\"#macroexpansion\" title=\"Macros \u2014 Beware!\">later in the chapter<\/a>.\n<\/p>\n<h4 id=\"id2934034\"><a href=\"#id2934034\">\u00a7<\/a>Case Study: XSLT<\/h4>\n<p>\n  XSLT, like\n  <em>m4<\/em> macros, is a\nlanguage for describing transformations of a text stream.  But it does\nmuch more than simple macro substitution; it describes transformations\nof XML data, including query and report generation.  It is the\nlanguage used to write XML stylesheets.  For practical applications,\nsee the description of XML document processing in\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>.  XSLT is described by a World Wide\nWeb Consortium standard and has several open-source\nimplementations.\n<\/p>\n<p>\n  XSLT and\n  <em>m4<\/em> macros are both purely\ndeclarative and Turing-complete, but XSLT supports only recursions and\nnot loops. It is quite complex, certainly the most difficult language\nto master of any in this chapter's case studies \u2014 and probably\nthe most difficult of any language mentioned in this\nbook.<sup><a href=\"#ftn.id2934079\" id=\"id2934079\">[84]<\/a><\/sup><\/p>\n<p>\n  Despite its complexity, XSLT really is a minilanguage.  It\nshares important (though not universal) characteristics of the breed:\n<\/p>\n <div><ul><li>\n    <p>\n     A restricted ontology of types, with (in particular) no analog \nof record structures or arrays.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Restricted interface to the rest of the world.  XSLT processors are\ndesigned to filter standard input to standard output, with a\nlimited ability to read and write files.  They can't open sockets\nor run subcommands.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 8.3. A sample XSLT program.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n&lt;?xml version=\"1.0\"?&gt;\n&lt;xsl:stylesheet xmlns:xsl=\"http:\/\/www.w3.org\/1999\/XSL\/Transform\" \n                version=\"1.0\"&gt;\n  &lt;xsl:output method=\"xml\"\/&gt;\n  &lt;xsl:template match=\"*\"&gt;\n    &lt;xsl:element name=\"{name()}\"&gt;\n      &lt;xsl:for-each select=\"@*\"&gt;\n        &lt;xsl:element name=\"{name()}\"&gt;\n          &lt;xsl:value-of select=\".\"\/&gt;\n        &lt;\/xsl:element&gt;\n      &lt;\/xsl:for-each&gt;\n      &lt;xsl:apply-templates select=\"*|text()\"\/&gt;\n    &lt;\/xsl:element&gt;\n  &lt;\/xsl:template&gt; \n&lt;\/xsl:stylesheet&gt;\n<\/pre><\/div>\n<p>\n  The program in Example 8.3 transforms an XML document so that each attribute of every element is transformed into a\n  new tag pair directly enclosed by that element, with the attribute value as the tag pair's content.\n<\/p>\n<p>\n  We've included a glance at XSLT here partly to illustrate the point that \u2018declarative\u2019 does not imply either \u2018simple\u2019\n  or \u2018weak\u2019, and mostly because if you have to work with XML documents, you will someday have to face the challenge that\n  is XSLT.\n<\/p>\n<p>\n  <em>XSLT: Mastering XML Transformations<\/em> <sup><a href=\"#Tidwell\" title=\"[Tidwell]\">[Tidwell]<\/a><\/sup> is a good introduction to the\nlanguage.  A brief tutorial with examples is available on the\nWeb.<sup><a href=\"#ftn.id2934184\" id=\"id2934184\">[85]<\/a><\/sup><\/p>\n<h4 id=\"id2934197\"><a href=\"#id2934197\">\u00a7<\/a>Case Study: The Documenter's Workbench Tools<\/h4>\n<p>\n  The\ntroff(1)\ntypesetting formatter was, as we noted in\n  <a href=\"#historychapter\" title=\"Chapter 2. History\">\n   Chapter 2<\/a>, Unix's original killer application.\n  <em>troff<\/em> is the center of a suite of\nformatting tools (collectively called Documenter's Workbench or DWB),\nall of which are domain-specific minilanguages of various kinds.  Most\nare either preprocessors or postprocessors for\ntroff markup.  Open-source Unixes host an\nenhanced implementation of Documenter's Workbench called\ngroff(1),\nfrom the Free Software Foundation.\n<\/p>\n<p>\n  We'll examine <em>troff<\/em> in more detail in <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n\t  Chapter 18<\/a>; for now, it's sufficient to note that it is a good example of an imperative minilanguage that\n  borders on being a full-fledged interpreter (it has conditionals and recursion but not loops; it is accidentally\n  Turing-complete).\n<\/p>\n<p>\n  The postprocessors (\u2018drivers\u2019 in DWB terminology)\nare normally not visible to\n  <em>troff<\/em> users.\nThe original troff emitted codes for the\nparticular typesetter the Unix development group had available in\n1970; later in the 1970s these were cleaned up into a\ndevice-independent minilanguage for placing text and simple graphics\non a page.  The postprocessors translate this language (called\n\u201cditroff\u201d for \u201cdevice-independent troff\u201d)\ninto something modern imaging printers can actually accept \u2014 the\nmost important of these (and the modern default) is PostScript.\n<\/p>\n<p>\n  The preprocessors are more interesting, because they actually\nadd capabilities to the troff language.\nThere are three common ones:\ntbl(1)\nfor making tables,\neqn(1)\nfor typesetting mathematical equations, and\npic(1)\nfor drawing diagrams.  Less used, but still live, are\ngrn(1)\nfor graphics, and\nrefer(1)\nand\nbib(1)\nfor formatting bibliographies. Open-source equivalents of all of these\nship with\n  <em>groff<\/em>.  The\ngrap(1)\npreprocessor provided a rather versatile plotting facility; there is\nan open-source implementation separate from\n  <em>groff<\/em>.\n<\/p>\n<p>\n  Some other preprocessors have no open-source implementation and\nare no longer in common use. Best known of these was\nideal(1),\nfor graphics.  A younger sibling of the family,\nchem(1),\ndraws chemical structural formulas; it is available as part of Bell\nLabs's netlib code.<sup><a href=\"#ftn.id2934417\" id=\"id2934417\">[86]<\/a><\/sup><\/p>\n<p>\n  Each of these preprocessors is a little program that accepts a\nminilanguage and compiles it into troff requests.  Each one recognizes\nthe markup it is supposed to interpret by looking for a unique start\nand end request, and passes through unaltered any markup outside those\n(\n  <em>tbl<\/em> looks for\n.TS\/.TE,\n  <em>pic<\/em> looks for\n.PS\/.PE, etc.).  Thus, most of the\npreprocessors can normally be run in any order without stepping on\neach other.  There are some exceptions: in particular,\n  <em>chem<\/em> and\n  <em>grap<\/em> both issue\n  <em>pic<\/em> commands, and so must come\nbefore it in the pipeline.\n<\/p>\n <pre class=\"programlisting\">\ncat thesis.ms | chem | tbl | refer | grap | pic | eqn \\\n                                             | groff -Tps &gt;thesis.ps\n<\/pre>\n<p>\n  The preceding is a full-Monty example of a Documenter's Workbench\nprocessing pipeline,\nfor a hypothetical thesis incorporating chemical formulas,\nmathematical equations, tables, bibliographies, plots, and diagrams.\n(The\ncat(1)\ncommand simply copies its input or a file argument to its output; we\nuse it here to emphasize the order of operations.)  In practice modern\ntroff implementations tend to support\ncommand-line options that can invoke at least\ntbl(1),\neqn(1)\nand\npic(1),\nso it isn't necessary to write such an elaborate pipeline.  Even if it\nwere, these sorts of build recipes are normally composed just once and\nstashed away in a makefile or shellscript wrapper for repeated use.\n<\/p>\n<p>\n  The document markup of Documenter's Workbench is in some ways\nobsolete, but the range of problems these preprocessors address gives\nsome indication of the power of the minilanguage model \u2014 it\nwould be extremely difficult to embed equivalent knowledge into a\nWYSIWYG word processor.  There are some ways in which modern XML-based\ndocument markups and toolchains are still, in 2003, playing\ncatch-up with capabilities that Documenter's Workbench had in 1979.\nWe'll discuss these issues in more detail in\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>.\n<\/p>\n<p>\n  The design themes that gave Documenter's Workbench so much power\nshould by now be familiar ones; all the tools share a common\ntext-stream representation of documents, and the formatting system is\nbroken up into independent components that can be debugged and\nimproved separately.  The pipeline architecture supports plugging in\nnew, experimental preprocessors and postprocessors without disturbing\nold ones.  It is modular and extensible.\n<\/p>\n<p>\n  The architecture of Documenter's Workbench as a whole teaches us\nsome things about how to fit multiple specialist minilanguages into a\ncooperating system. One preprocessor can build on another.  Indeed,\nthe Documenter's Workbench tools were an early exemplar of the power\nof pipes, filtering, and minilanguages that influenced a lot of later\nUnix design by example.  The design of the individual preprocessors\nhas more lessons to teach about what effective minilanguage designs\nlook like.\n<\/p>\n<p>\n  One of these lessons is negative. Sometimes users writing\ndescriptions in the minilanguages do unclean things with low-level\ntroff markup inserted by hand.  This can\nproduce interactions and bugs that are hard to diagnose, because the\ngenerated troff coming out of the pipeline\nis not visible \u2014 and would not be readable if it were.  This is\nanalogous to the sorts of bugs that happen in code that mixes C with\nsnippets of in-line assembler.  It might have been better to separate\nthe language layers more completely, if that were possible.\nMinilanguage designers should take note of this.\n<\/p>\n<p>\n  All the preprocessor languages (though not troff markup itself)\nhave relatively clean, shell-like syntaxes that follow many of the\nconventions we described in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  for the design of data-file formats.\nThere are a few embarrassing exceptions; notably,\ntbl(1)\ndefaults to using a tab as a field separator between table columns,\nreplicating an infamous botch in the design of\nmake(1)\nand causing annoying bugs when editors or other tools invisibly change\nthe composition of whitespace.\n<\/p>\n<p>\n  While troff itself is a specialized\nimperative language, one theme that runs through at least three of the\nDocumenter's Workbench minilanguages is declarative semantics: doing\nlayout from constraints.  This is an idea that shows up in modern GUI\ntoolkits as well \u2014 that, instead of giving pixel coordinates for\ngraphical objects, what you really want to do is declare spatial\nrelationships among them (\u201cwidget A is above widget B, which is\nto the left of widget C\u201d) and have your software compute a\nbest-fit layout for A, B, and C according to those constraints.\n<\/p>\n<p>\n  The pic(1) program uses this approach to lay out elements for diagrams.  The language taxonomy diagram at Figure 8.1\n  was produced with\nthe\n  <em>pic<\/em> source code in Example 8.4<sup><a href=\"#ftn.id2934708\" id=\"id2934708\">[87]<\/a><\/sup> run through\n  <em>pic2graph<\/em>, one of our case studies in <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n  Chapter 7<\/a>.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 8.4. Taxonomy of languages \u2014 the\n    <em>pic<\/em> source.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n# Minilanguage taxonomy\n#\n# Base ellipses\ndefine smallellipse {ellipse width 3.0 height 1.5}\nM: ellipse width 3.0 height 1.8 fill 0.2\nline from M.n to M.s dashed\nD: smallellipse() with .e at M.w + (0.8, 0)\nline from D.n to D.s dashed\nI: smallellipse() with .w at M.e - (0.8, 0)\n#\n# Captions\n\"\" \"Data formats\" at D.s\n\"\" \"Minilanguages\" at M.s\n\"\" \"Interpreters\" at I.s\n#\n# Heads\narrow from D.w + (0.4, 0.8) to D.e + (-0.4, 0.8)\n\"flat to structured\" \"\" at last arrow.c\narrow from M.w + (0.4, 1.0) to M.e + (-0.4, 1.0)\n\"declarative to imperative\" \"\" at last arrow.c\narrow from I.w + (0.4, 0.8) to I.e + (-0.4, 0.8)\n\"less to more general\" \"\" at last arrow.c\n#\n# The arrow of loopiness\narrow from D.w + (0, 1.2) to I.e + (0, 1.2)\n\"increasing loopiness\" \"\" at last arrow.c\n#\n# Flat data files\n\"\/etc\/passwd\" \".newsrc\" at 0.5 between D.c and D.w\n# Structured data files\n\"SNG\" at 0.5 between D.c and M.w\n# Datafile\/minilanguage borderline cases\n\"regexps\" \"Glade\" at 0.5 between M.w and D.e\n# Declarative minilanguages\n\"m4\" \"Yacc\" \"Lex\" \"make\" \"XSLT\" \"pic\" \"tbl\" \"eqn\" \\\n\t\t\tat 0.5 between M.c and D.e\n# Imperative minilanguages\n\"fetchmail\" \"awk\" \"troff\" \"Postscript\" at 0.5 between M.c and I.w\n# Minilanguage\/interpreter borderline cases\n\"dc\" \"bc\" at 0.5 between I.w and M.e\n# Interpreters\n\"Emacs Lisp\" \"JavaScript\" at 0.25 between M.e and I.e\n\"sh\" \"tcl\" at 0.55 between M.e and I.e\n\"Perl\" \"Python\" \"Java\" at 0.8 between M.e and I.e\n<\/pre><\/div>\n<p>\n  This is a very typical Unix minilanguage design, and as such has\nsome points of interest even on the purely syntactic level.  Notice\nhow much it looks like a shell program: # leads comments, and\nthe syntax is obviously token-oriented with the simplest possible\nconvention for strings.  The designer of\npic(1)\nknew that Unix programmers expect minilanguage syntaxes to look like\nthis unless there is a strong and specific reason they should not.\nThe Rule of Least Surprise is in full operation here.\n<\/p>\n<p>\n  It probably doesn't take a lot of effort to discern that the\nfirst line of code is a macro definition; the later references to\n  <code>smallellipse()<\/code> encapsulate a repeated design\nelement of the diagram.  Nor will it take much scrutiny to deduce that\n  <code>box invis<\/code> declares a box with invisible borders,\nactually just a frame for text to be stacked inside.  The\n  <code>arrow<\/code> command is equally obvious.\n<\/p>\n<p>\n  With these as clues and one eye on the actual diagram, the\nmeaning of the remaining pieces of the syntax (position references\nlike\n  <code>M.s<\/code> and constructions like\n  <code>last arrow<\/code> or\n  <code>\n   at\n0.25 between M.e and I.e\n  <\/code>\n  or the addition of vector offsets\nto a location) should become rapidly apparent.  As with\nGlade markup and\n  <em>m4<\/em>, an\nexample like this one can teach a good bit of the language without any\nreference to a manual (a compactness property\ntroff(1)\nmarkup, unfortunately, does\n  <em>not<\/em>\n  have).\n<\/p>\n<p>\n  The example of\npic(1)\nreflects a common design theme in minilanguages, which we also\nsaw reflected in Glade \u2014 the use of a\nminilanguage interpreter to encapsulate some form of constraint-based\nreasoning and turn it into actions.  We could actually choose to view\npic(1)\nas an imperative language rather than a declarative one; it has\nelements of both, and the dispute would quickly grow\ntheological.\n<\/p>\n<p>\n  The combination of macros with constraint-based layout gives\npic(1)\nthe ability to express the structure of diagrams in a way that more\nmodern vector-based markups like SVG cannot.  It is therefore\nfortunate that one effect of the Documenter's Workbench design is to\nmake it relatively easy to keep\npic(1)\nuseful outside the DWB context.  The\n  <em>pic2graph<\/em> script we used as a case study in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  was an\nad-hoc way to accomplish this, using the retrofitted PostScript\ncapability of\ngroff(1)\nas a half-way step to a modern bitmap format.\n<\/p>\n<p>\n  A cleaner solution is the\npic2plot(1)\nutility distributed with the GNU plotutils\npackage, which\nexploited the internal modularity of the GNU\npic(1)\ncode.  The code was split into a parsing front end and a back end that\ngenerated troff markup, the two communicating through a layer of\ndrawing primitives.  Because this design obeyed the Rule of\nModularity,\n  <em>pic2plot(1)<\/em> implementers were able to split off the GNU\n  <em>pic<\/em> parsing stage and reimplement the\ndrawing primitives using a modern plotting library.  Their solution\nhas the disadvantage, however, that text in the output is generated\nwith fonts built into\n  <em>pic2plot<\/em> that won't\nmatch those of troff.\n<\/p>\n<h4 id=\"fetchmailrc\"><a href=\"#fetchmailrc\">\u00a7<\/a>\n  Case Study:\nfetchmail Run-Control Syntax\n <\/h4>\n<p>\n  See Example 8.5 for an example.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 8.5. Synthetic example of a\n    <code>fetchmailrc<\/code>.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n# Poll this site first each cycle.\npoll pop.provider.net proto pop3\n    user \"jsmith\" with pass \"secret1\" is \"smith\" here\n    user jones with pass \"secret2\" is \"jjones\" here with options keep\n# Poll this site second, unless Lord Voldemort zaps us first.\npoll billywig.hogwarts.com with proto imap:\n    user harry_potter with pass \"floo\" is harry_potter here\n# Poll this site third in the cycle.  \n# Password will be fetched from ~\/.netrc\npoll mailhost.net with proto imap:\n    user esr is esr here\n<\/pre><\/div>\n<p>\n  This run-control file can be viewed as an imperative\nminilanguage.  There is an implied flow of execution: cycle through\nthe list of poll commands repeatedly (sleeping for a while at the end\nof each cycle), and for each site entry collect mail for each\nassociated user in sequence.  It is far from being general-purpose; \nall it can do is sequence the program's polling behavior.\n<\/p>\n<p>\n  As with\npic(1),\none could choose to view this minilanguage as either declarations or a\nvery weak imperative language, and argue endlessly over the\ndistinction.  On the one hand, it has neither conditionals nor\nrecursion nor loops; in fact, it has no explicit control structures at\nall.  On the other hand, it does describe actions rather than just\nrelationships, which distinguishes it from a purely declarative syntax\nlike Glade GUI descriptions.\n<\/p>\n<p>\n  Run-control minilanguages for complex programs often straddle\nthis border.  We're making a point of this fact because not having\nexplicit control structures in an imperative minilanguage can be a\ntremendous simplification if the problem domain lets you get away\nwith it.\n<\/p>\n<p>\n  One notable feature of\n  <code>.fetchmailrc<\/code> syntax\nis the use of optional noise keywords that are supported simply in\norder to make the specifications read a bit more like English.  The\n\u2018with\u2019 keywords and single occurrence of\n\u2018options\u2019 in the example aren't actually necessary, but\nthey help make the declarations easier to read at a glance.\n<\/p>\n<p>\n  The traditional term for this sort of thing is\n  <em>syntactic sugar<\/em>; the maxim that goes with this\nis a famous quip that \u201csyntactic sugar causes cancer of the\nsemicolon\u201d.<sup><a href=\"#ftn.id2935218\" id=\"id2935218\">[88]<\/a><\/sup>\n  Indeed, syntactic sugar needs to be used\nsparingly lest it obscure more than help.\n<\/p>\n<p>\n  In\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>\n  we'll see how data-driven programming helps provide an elegant\nsolution to the problem of editing\n  <em>fetchmail<\/em> run-control files through a\nGUI.\n<\/p>\n<h4 id=\"awk\"><a href=\"#awk\">\u00a7<\/a>Case Study: awk<\/h4>\n<p>\n  The\n  <em>awk<\/em> minilanguage is an old-school\nUnix tool, formerly much used in shellscripts.  Like\n  <em>m4<\/em>, it's intended for writing small but\nexpressive programs to transform textual input into textual output.\nVersions ship with all Unixes, several in open source; the command\n  <code>info gawk<\/code> at your Unix shell prompt is quite likely\nto take you to on-line documentation.\n<\/p>\n<p>\n  Programs in\n  <em>awk<\/em> consist of\npattern\/action pairs.  Each pattern is a\n  <em>regular expression<\/em>, a concept we'll describe in detail in\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>.  When an\n  <em>awk<\/em> program is run, it steps through each line of the input file.  Each\nline is checked against every pattern\/action pair in order.  If the\npattern matches the line, the associated action is performed.\n<\/p>\n<p>\n  Each action is coded in a language resembling a subset of C, with\nvariables and conditionals and loops and an ontology of types\nincluding integers, strings, and (unlike C)\ndictionaries.<sup><a href=\"#ftn.id2935337\" id=\"id2935337\">[89]<\/a><\/sup><\/p>\n<p>\n  The\n  <em>awk<\/em> action language is\nTuring-complete, and can read and write files. In some versions it can\nopen and use network sockets.  But\n  <em>awk<\/em> has\nprimarily seen use as a report generator, especially for interpreting\nand reducing tabular data.  It is seldom used standalone, but rather\nembedded in scripts.  There is an example\n  <em>awk<\/em> program in the\n  <a href=\"#htmlgen\" title=\"Case Study: Generating HTML Code for a Tabular List\">case study on HTML generation<\/a>\n  included in <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">Chapter 9<\/a>.\n<\/p>\n<p>\n  A case study of\n  <em>awk<\/em> is included to\npoint out that it is\n  <em>not<\/em>\n  a model for emulation;\nin fact, since 1990 it has largely fallen out of use. It has been\nsuperseded by new-school scripting\nlanguages\n  \u2014notably \nPerl, which was explicitly designed to be an\n  <em>awk<\/em> killer.  The reasons are worthy of\nexamination, because they constitute a bit of a cautionary tale for\nminilanguage designers.\n<\/p>\n<p>\n  The\n  <em>awk<\/em> language was originally\ndesigned to be a small, expressive special-purpose language for report\ngeneration.  Unfortunately, it turns out to have been designed at a\nbad spot on the complexity-vs.-power curve.  The action language is\nnoncompact, but\nthe pattern-driven framework it sits inside keeps it from being\ngenerally applicable \u2014 that's the worst of both worlds.  And the\nnew-school scripting languages can do anything\n  <em>awk<\/em> can; their equivalent programs are\nusually just as readable, if not more so.\n<\/p>\n <blockquote>\n  <p>\n   Awk has also fallen out of use because more modern shells have\nfloating point arithmetic, associative arrays, RE pattern matching,\nand substring capabilities, so that equivalents of small awk scripts\ncan be done without the overhead of process creation.\n  <\/p><div>\n  \n   \u2013 David Korn\n  \n <\/div><\/blockquote>\n<p>\n  For a few years after the release of Perl in 1987,\n  <em>awk<\/em> remained competitive simply because it\nhad a smaller, faster implementation.  But as the cost of compute\ncycles and memory dropped, the economic reasons for favoring a\nspecial-purpose language that was relatively thrifty with both lost\ntheir force. Programmers increasingly chose to do awklike things with\nPerl or (later) Python, rather than keep two different scripting\nlanguages\n  in their heads.<sup><a href=\"#ftn.id2935542\" id=\"id2935542\">[90]<\/a><\/sup>\n  By the year 2000\n  <em>awk<\/em> had become little more than a memory\nfor most old-school Unix\nhackers, and not a\nparticularly nostalgic one.\n<\/p>\n<p>\n  Falling costs have changed the tradeoffs in minilanguage design.\nRestricting your design's capabilities to buy\ncompactness\n  may\nstill be a good idea, but doing so to economize on machine resources\nis a bad one.  Machine resources get cheaper over time, but space in\nprogrammers' heads only gets more expensive.  Modern minilanguages can\neither be general but noncompact, or specialized but very\ncompact; specialized\nbut noncompact simply won't compete.\n<\/p>\n<h4 id=\"id2935613\"><a href=\"#id2935613\">\u00a7<\/a>Case Study: PostScript<\/h4>\n<p>\n  PostScript is a minilanguage specialized for describing typeset\ntext and graphics to imaging devices. It is an import into Unix,\nbased on design work done at the legendary Xerox Palo Alto\nResearch Center\n  along with the earliest laser printers.  For years after its first\ncommercial release in 1984, it was available only as a proprietary\nproduct from Adobe, Inc., and was primarily associated with Apple\ncomputers.  It was cloned under license terms very close to\nopen-source in 1988, and has since become the de-facto standard for\nprinter control under Unix.  A fully open-source version is shipped\nwith most most modern Unixes.<sup><a href=\"#ftn.id2935641\" id=\"id2935641\">[91]<\/a><\/sup>\n  A good technical introduction to PostScript is also\navailable.<sup><a href=\"#ftn.id2935658\" id=\"id2935658\">[92]<\/a><\/sup><\/p>\n<p>\n  PostScript bears some functional resemblance to troff markup;\nboth are intended to control printers and other imaging devices, and\nboth are normally generated by programs or macro packages rather than\nbeing hand-written by humans.  But where troff requests are a\njumped-up set of format-control codes with some language features\ntacked on as an afterthought, PostScript was designed from the ground\nup as a language and is far more expressive and powerful.  The main\nthing that makes Postscript useful is that algorithmic descriptions of\nimages written in it are far smaller than the bitmaps they render to,\nand so take up less storage and communication bandwidth.\n<\/p>\n<p>\n  PostScript is explicitly Turing-complete, supporting\nconditionals and loops and recursion and named procedures.  The\nontology of types includes integers, reals, strings, and arrays (each\nelement of an array may be of any type) but no equivalent of\nstructures.  Technically, PostScript is a stack-based language;\narguments of PostScript's primitive procedures (operators) are normally\npopped off a push-down stack of arguments, and the result(s) are pushed\nback onto it.\n<\/p>\n<p>\n  There are about 40 basic operators out of a total of around 400.\nThe one that does most of the work is <code>show<\/code>, which\ndraws a string onto the page.  Others set the current font, change the\ngray level or color, draw lines or arcs or Bezier curves, fill closed\nregions, set clipping regions, etc.  A PostScript interpreter is\nsupposed to be able to interpret these commands into bitmaps to be\nthrown on a display or print medium.\n<\/p>\n<p>\n  Other PostScript operators implement arithmetic, control\nstructures, and procedures. These allow repetitive or stereotyped\nimages (such as text, which is composed of repeated letterforms) to be\nexpressed as programs that combine images.  Part of the utility of\nPostScript comes from the fact that PostScript programs to print text or\nsimple vector graphics are much less bulky than the bitmaps the text\nor vectors render to, are device-resolution independent, and travel more\nquickly over a network cable or serial line.\n<\/p>\n<p>\n  Historically, PostScript's stack-based interpretation resembles\na language called FORTH, originally designed to control telescope\nmotors in real time, which was briefly popular in the 1980s.\nStack-based languages are famous for supporting extremely tight,\neconomical coding and infamous for being difficult to read.\nPostScript shares both traits.\n<\/p>\n<p>\n  PostScript is often implemented as firmware built into a\nprinter.  The open-source implementation Ghostscript can translate\nPostScript to various graphics formats and (weaker) printer-control\nlanguages.  Most other software treats PostScript as a final output\nformat, meant to be handed to a PostScript-capable imaging device but\nnot edited or eyeballed.\n<\/p>\n<p>\n  PostScript (either in the original or the trivial variant\nEPSF, with a bounding box\ndeclared around it so it can be embedded in other graphics) is a very\nwell designed example of a special-purpose control language and\ndeserves careful study as a model. It is a component of other\nstandards such as PDF, the Portable Document Format.\n<\/p>\n<h4 id=\"id2935779\"><a href=\"#id2935779\">\u00a7<\/a>Case Study: bc and dc<\/h4>\n<p>\n  We first examined bc(1) and dc(1) in <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\"> Chapter 7<\/a>\n  as a case study in shellouts.  They are examples of domain-specific minilanguages of the imperative type.\n<\/p>\n <blockquote>\n  <p>\n   <em>dc<\/em> is the oldest language on Unix; it\nwas written on the PDP-7 and ported to the PDP-11 before Unix [itself]\nwas ported.\n  <\/p><div>\n  \n   \u2013 Ken Thompson\n  \n <\/div><\/blockquote>\n<p>\n  The domain of these two languages is unlimited-precision\narithmetic.  Other programs can use them to do such calculations\nwithout having to worry about the special techniques needed to do\nthose calculations.\n<\/p>\n <blockquote>\n  <p>\n   In fact, the original motivation for dc had nothing to do with\nproviding a general-purpose interactive calculator, which could have\nbeen done with a simple floating-point program.  The motivation was\nBell Labs' long interest in numerical analysis: calculating constants\nfor numerical algorithms,\n   <em>accurately<\/em>\n   is greatly\naided by being able to work to much higher precision than the\nalgorithm itself will use.  Hence dc's arbitrary-precision\narithmetic.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  Like SNG and\n  <em>Glade<\/em> markup, one of the\nstrengths of both of these languages is their simplicity.  Once you\nknow that\ndc(1)\nis a reverse-Polish-notation calculator and\nbc(1)\nan algebraic-notation calculator, very little about interactive use of\neither of these languages is going to be novel.  We'll return to the\nimportance of the Rule of Least Surprise in interfaces in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>.\n<\/p>\n<p>\n  These minilanguages have both conditionals and loops; they are\nTuring-complete, but have a very restricted ontology of types\nincluding only unlimited-precision integers and strings.  This puts\nthem in the borderland between interpreted minilanguages and full\nscripting languages.  The programming features have been\ndesigned not to intrude on the common use as a calculator; indeed,\nmost\n  <em>dc<\/em> \/\n  <em>bc<\/em> users\nare probably unaware of them.\n<\/p>\n<p>\n  Normally,\n  <em>dc<\/em> \/\n  <em>bc<\/em> are used\nconversationally, but their capacity to support libraries of\nuser-defined procedures gives them an additional kind of utility\n\u2014 programmability. This is actually the most important advantage\nof imperative minilanguages, one that we observed in the case study\nof the Documenter's Workbench tools to be very powerful whether or not\na program's normal mode is conversational; you can use them to write\nhigh-level programs that embody task-specific intelligence.\n<\/p>\n<p>\n  Because the interface of\n  <em>dc<\/em> \/\n  <em>bc<\/em> is so\nsimple (send a line containing an expression, get back a line\ncontaining a value) other programs and scripts can easily get access\nto all these capabilities by calling these programs as slave\nprocesses.\n   Example 8.6 is one famous example, an implementation of the Rivest-Shamir-Adelman public-key cipher in Perl that was\n   widely published in signature blocks and on T-shirts as a protest against U.S. export retrictions on cryptography, c.\n   1995; it shells out to <em>dc<\/em> to do the unlimited-precision arithmetic required.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 8.6. RSA implementation using\n    <em>dc<\/em>.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nprint pack\"C*\",split\/\\D+\/,`echo \"16iII*o\\U@{$\/=$z;[(pop,pop,unpack\n\"H*\",&lt;&gt;)]}\\EsMsKsN0[lN*1lK[d2%Sa2\/d0&lt;X+d*lMLa^*lN%0]dsXx++\\\nlMlN\/dsM0&lt;J]dsJxp\"|dc`\n<\/pre><\/div>\n<h4 id=\"emacs_lisp_study\"><a href=\"#emacs_lisp_study\">\u00a7<\/a>Case Study: Emacs Lisp<\/h4>\n<p>\n  Rather than merely being run as a slave process to accomplish\nspecific tasks, a special-purpose interpreted language can become the\ncore of an entire architecture; we'll consider the advantages and\ndisadvantages of this approach in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">\n   Chapter 13<\/a>.  troff requests were an early example;\ntoday, the\n  <em>Emacs<\/em> editor is one of the\nbest-known and most powerful modern ones.  It's built around a dialect\nof Lisp\n  with\nprimitives for both describing actions on editing buffers and\ncontrolling slave processes.\n<\/p>\n<p>\n  The fact that Emacs is built around a powerful language for\ndescribing editing actions or front ends for other programs means that\nit can be used for many other things besides ordinary editing.  We'll\nexamine the applications of Emacs's task-specific intelligence for\nday-to-day program development (compilation, debugging, version\ncontrol) in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>. Emacs \u2018modes\u2019\nare user-defined libraries \u2014 programs written in Emacs Lisp that\nspecialize the editor for a particular job \u2014 usually, but not\nnecessarily, one related to editing.\n<\/p>\n<p>\n  Thus there are specialized modes that know the syntax of a large\nnumber of programming languages, and of markup languages like\nSGML, XML, and HTML. But\nmany people also use Emacs modes to send\nand receive email (these use Unix system mail utilities as slaves) or\nUsenet news. Emacs can browse the web, or act as a front-end for\nvarious chat programs. There is also a calendaring package, Emacs's\nown calculator program, and even a fairly wide selection of games\nwritten as Emacs Lisp modes (including a descendant of the famous\nELIZA program that simulates a Rogersian psychiatrist).<sup><a href=\"#ftn.id2936188\" id=\"id2936188\">[93]<\/a><\/sup><\/p>\n<h4 id=\"javascript\"><a href=\"#javascript\">\u00a7<\/a>Case Study: JavaScript<\/h4>\n<p>\n  JavaScript is an open-source language designed to be embedded in\nC\n  programs. Though\nit is also embedded in web servers, its original and best-known\nmanifestation is client-side JavaScript, which allows you to embed\nexecutable code in Web pages to be run by any JavaScript-capable\nbrowser.  That is the version we will survey here.\n<\/p>\n<p>\n  JavaScript is a fully Turing-complete interpreted language with\nintegers, real numbers, booleans, strings, and lightweight\ndictionary-based objects resembling those of\nPython.\nValues are typed, but variables can hold any type; conversions between\ntypes are automatic in many contexts.  Syntactically JavaScript resembles\nJava\n  with some\ninfluence from Perl, and features Perl-like regular\nexpressions.\n<\/p>\n<p>\n  Despite all these features, client-side JavaScript is not quite a\ngeneral-purpose language.  Its capabilities are severely restricted\nto prevent attacks on the browser user through Web pages\ncontaining JavaScript code.  It can accept input from the user and\ngenerate or modify Web pages, but it cannot directly alter the\ncontents of disk files and cannot open its own network\nconnections.\n<\/p>\n<p>\n  Over time, the JavaScript language has become more general and\nless bound to its client-side environment.  This is something that can\nbe expected to happen to any successful specialized language as its\npossibilities unfold in the minds of developers and users.  Client\nJavaScript now interacts with its environment by reading and writing\nvalues in a single special object called the browser DOM (Document\nObject Model).  The language still has some legacy APIs to the browser\nthat don't go through the DOM, but these are deprecated, not present\nin the ECMA-262 standard for JavaScript, and may not be supported in\nfuture versions.\n<\/p>\n<p>\n  The standard reference for JavaScript is\n  <em>JavaScript: The Definitive Guide<\/em><sup><a href=\"#FlanaganJavaScript\" title=\"[FlanaganJavaScript]\">[FlanaganJavaScript]<\/a><\/sup>.  Source code\nis downloadable.<sup><a href=\"#ftn.id2936325\" id=\"id2936325\">[94]<\/a><\/sup>\n  JavaScript makes an interesting study for two reasons. First, it's\nabout as close to being a general-purpose language as one can get\nwithout actually being there.  Second, the binding between client-side\nJavaScript and its browser environment via a single DOM object is well\ndesigned, and could serve as a model for other embedding\nsituations.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2932978\" id=\"ftn.id2932978\">[81]<\/a><\/sup>\n  The POSIX\nstandard for regular expressions introduces some symbolic ranges like\n  <code>[[:lower;;]]<\/code> and\n  <code>[[:digit:]]<\/code>, and some specific tools have extra\nwildcards not covered here, but these will suffice to interpret most\nregexps.\n<\/p><p><sup><a href=\"#id2933471\" id=\"ftn.id2933471\">[82]<\/a><\/sup>\n  For non-Unix programmers, an X toolkit is a graphics\nlibrary that supplies GUI widgets (like labels, buttons, and pull-down\nmenus) to the programs that link to it.  Under most other graphical\noperating systems, the OS supplies one toolkit that everyone uses.\nUnix and X support multiple toolkits; this is part of the separation\nof policy from mechanism that we called out as a design goal of X in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>. GTK and Qt\nare the two most popular open-source X toolkits.\n<\/p><p><sup><a href=\"#id2934007\" id=\"ftn.id2934007\">[83]<\/a><\/sup>\n  Whether or not \u201cmacro expansion\u201d\nshould be spelled \u201cmacroexpansion\u201d is a matter for some\ndispute.  The latter is found mainly among Lisp\nprogrammers.\n<\/p><p><sup><a href=\"#id2934079\" id=\"ftn.id2934079\">[84]<\/a><\/sup>\n  It is not clear that XSLT could be any simpler\nand still do its job, however, so we cannot characterize it as a bad\ndesign.\n<\/p><p><sup><a href=\"#id2934184\" id=\"ftn.id2934184\">[85]<\/a><\/sup><a href=\"http:\/\/nwalsh.com\/docs\/tutorials\/xsl\/xsl\/slides.html\">\n   XSL\nConcepts and Practical Use<\/a>.\n<\/p><p><sup><a href=\"#id2934417\" id=\"ftn.id2934417\">[86]<\/a><\/sup><a href=\"http:\/\/www.netlib.org\/\">\n   http:\/\/www.netlib.org\/<\/a>\n<\/p><p><sup><a href=\"#id2934708\" id=\"ftn.id2934708\">[87]<\/a><\/sup>\n  It is also quite traditional\nfor Unix books that describe\npic(1)\nto include their own illustrations as coding examples.\n<\/p><p><sup><a href=\"#id2935218\" id=\"ftn.id2935218\">[88]<\/a><\/sup>\n  The line is owed to Alan Perlis, who did\nsome of the pioneering work in software modularity around 1970. The\nsemicolon in question was the statement separator or terminator\nin various Algol-descended languages, including Pascal and\nC.\n<\/p><p><sup><a href=\"#id2935337\" id=\"ftn.id2935337\">[89]<\/a><\/sup>\n  For those who have never programmed in a\nmodern scripting language, a dictionary is a lookup table of\nkey-to-value associations, often implemented through a hash table.\nC\n  \n  programmers spend a lot\nof their coding time implementing dictionaries in various elaborate\nways.\n<\/p><p><sup><a href=\"#id2935542\" id=\"ftn.id2935542\">[90]<\/a><\/sup>\n  I was at one time an\n  <em>awk<\/em> wizard, but I had to be reminded by someone\nelse that the language was applicable to the HTML-generation problem\nwhere this book's only\n  <em>awk<\/em> example\noccurs.\n<\/p><p><sup><a href=\"#id2935641\" id=\"ftn.id2935641\">[91]<\/a><\/sup>\n  There is a\n  <a href=\"http:\/\/www.cs.wisc.edu\/~ghost\/\">\n   GhostScript Project\nsite<\/a>.\n<\/p><p><sup><a href=\"#id2935658\" id=\"ftn.id2935658\">[92]<\/a><\/sup><a href=\"http:\/\/www.cs.indiana.edu\/docproject\/programming\/postscript\/postscript.html\">\n   A First Guide To PostScript<\/a>.\n<\/p><p><sup><a href=\"#id2936188\" id=\"ftn.id2936188\">[93]<\/a><\/sup>\n  One of the silliest things you can do with a modern Unix machine\nis run the Eliza mode of Emacs against\nrandom quotes from Zippy the Pinhead.\n  <code>M-x psychoanalyze-pinhead<\/code>; type control-G when you've had\nenough.\n<\/p><p><sup><a href=\"#id2936325\" id=\"ftn.id2936325\">[94]<\/a><\/sup>\n  Open-source JavaScript implementations\nin C and Java are\n  <a href=\"http:\/\/www.mozilla.org\/js\/\">\n   available<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch08s03\"><a href=\"#ch08s03\">\u00a7<\/a>Designing Minilanguages<\/h3>\n<p>\n  When is designing a minilanguage appropriate?  We've observed\nthat minilanguages offer a way to push problem specifications to a\nhigher level, and seen how this operates in several case studies.  The\nflip side of this observation is that a minilanguage is likely to be a\ngood approach whenever the domain primitives in your application area\nare simple and stereotyped, but the ways in which users are likely to\nwant to apply them are fluid and varying.\n<\/p>\n<p>\n  For some related ideas, find a description of the\n  <a href=\"http:\/\/www.c2.com\/cgi\/wiki?AlternateHardAndSoftLayers\">\n   Alternate\nHard And Soft Layers<\/a>\n  and\n  <a href=\"http:\/\/www.doc.ic.ac.uk\/~np2\/patterns\/scripting\/scripting.html\">\n   Scripted\nComponents<\/a>\n  design patterns.\n<\/p>\n<p>\n  An interesting survey of design styles and techniques in\nminilanguages is\n  <em>Notable Design Patterns for Domain-Specific Languages<\/em><sup><a href=\"#Spinellis\" title=\"[Spinellis]\">[Spinellis]<\/a><\/sup>.\n<\/p>\n<h4 id=\"id2936413\"><a href=\"#id2936413\">\u00a7<\/a>Choosing the Right Complexity Level<\/h4>\n<p>\n  The first important thing to bear in mind when designing a\nminilanguage is, as usual, to keep it as simple as possible.  The\ntaxonomy diagram we used to organize the case studies implies a\nhierarchy of complexity; you want to keep your design as far toward the\nleft-hand edge as possible.  If you can get away with designing a\nstructured data file rather than a minilanguage that is going to\nmodify external data when it's interpreted, by all means do\nso.\n<\/p>\n<p>\n  One very pragmatic reason to stick with structured data rather\nthan a minilanguage is that in a networked world, embedded\nminilanguage facilities are subject to abuses that can be inconvenient\nor even dangerous.  JavaScript\n  is a prime example in the\n\u2018inconvenient\u2019 category; its designers didn't anticipate\nthat it would be used for pop-up advertisements so obnoxious as to\ncreate a demand for browser features that suppress JavaScript\ninterpretation.\n<\/p>\n<p>\n  Microsoft Word\n  macro viruses show how this sort of thing\ncan become actively dangerous, a security hole that costs billions of\ndollars in downtime and lost productivity annually.  It is instructive\nto note that despite the existence of at least twenty million Unix\nusers worldwide<sup><a href=\"#ftn.id2936473\" id=\"id2936473\">[95]<\/a><\/sup>\n  there has never been any Unix equivalent\nof Windows's frequent macro-virus outbreaks.  There are a number of\nreasons for this, including the fundamentally better security design\nof Unix; but at least one is the fact that Unix mail agents do\n  <em>not<\/em>\n  default to executing live content in any\ndocument that the user\nviews.<sup><a href=\"#ftn.id2936490\" id=\"id2936490\">[96]<\/a><\/sup><\/p>\n<p>\n  If there is any way that your application's users might end up\nrunning programs from untrusted sources, risky features of your\napplication minilanguage might end up having to be suppressed.\nLanguages like Java and JavaScript are explicitly\n  <em>sandboxed<\/em> \u2014that is, they have limited\naccess to their environment not merely to simplify their design but to\ntry to prevent potentially destructive operations by buggy or\nmalicious code.\n<\/p>\n<p>\n  On the other hand, a lot of bad designs have been\nbotched by designers who failed to face up to the fact that they\nreally needed a minilanguage rather than a data-file format.  Too\noften, language-like features get pasted on as an afterthought.\nThe two most common symptoms of this problem are weak, ad-hoc control\nstructures and poor or nonexistent facilities for declaring\nprocedures.\n<\/p>\n<p>\n  It's risky to design minilanguages that are only accidentally\nTuring-complete.  If you do this the odds are good that, sometime in\nthe future, some clever fellow is going to think he needs to press \nyour language into doing loops and conditionals for him.  Because\nthese are only available in an obfuscated way, he'll produce\nobfuscated code. The results may be serviceable in the short term, \nbut are likely to be a nightmare for those who come after him.\n<\/p>\n<p>\n  Minilanguage design is both powerful and esthetically rewarding,\nbut it's also full of similar traps.  There are kinds of design in\nwhich it is appropriate to take the bottom-up approach of pasting\ntogether a bunch of low-level services and worrying about their\norganization after you have explored the problem domain for a\nwhile.  One of the virtues of minilanguages is that they can help you\nget a good design out of bottom-up programming by allowing you to\ndefer some top-down decisions into the control flow of programs in\nyour minilanguage.  But if you take a bottom-up approach to the minilanguage\ndesign\n  <em>itself<\/em>, you are likely to end up with an\nugly syntax reflecting a weak language and a poorly-thought-out\nimplementation.\n<\/p>\n<p>\n  There are many places in a minilanguage design where small\nchoices make a large difference in the useability and ease of the\ntool:\n<\/p>\n <blockquote>\n  <p>\n   As a language designer, it is a good principle to consider the\nalternatives to giving an error message.  When there is true ambiguity\nin the intent of the programmer an error message is appropriate, but\nin many cases the intent is clear, and making the language silently do\nthe right thing is a great boon.  A good example is C accommodating an\nextra comma at the end of an array initializer list, which makes both\nediting and machine generation of array initializers much easier.\nAnti-examples are the pickiness of various HTML readers, especially\ntheir habit of silently discarding parts of your document because of\ntrivial nesting errors.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  On this issue, as elsewhere, there is no substitute for good\ntaste and engineering judgment.  If you're going to design a\nminilanguage, don't do it halfway.  Declarative minilanguages should\nhave a clear, consistent language-like syntax designed to be readable\nby humans.  Imperative ones should add a full range of control\nstructures adapted from language models you can expect your users to\nbe familiar with.  Think about the language\n  <em>as<\/em>\n  a\nlanguage; ask yourself esthetic questions like \u201cWill this be\ncomfortable to program in?\u201d and even \u201cWill it be pleasant\nto look at?\u201d Here, as elsewhere in software design, David\nGelernter's maxim is apt: beauty is the ultimate defense against\ncomplexity.\n<\/p>\n<h4 id=\"id2936650\"><a href=\"#id2936650\">\u00a7<\/a>Extending and Embedding Languages<\/h4>\n<p>\n  One fundamentally important question is whether you can\nimplement your minilanguage by extending or embedding an existing\nscripting language.  This is often the right way to go\nfor an imperative minilanguage, but much less appropriate for a\ndeclarative one.\n<\/p>\n<p>\n  Sometimes it's possible to write your imperative language simply\nby coding service functions in an interpreted language, which we'll\ncall the \u2018host\u2019 language for purposes of this\ndiscussion. Your minilanguage programs are then just scripts that load\nyour service library and use the host language's control structures\nand other facilities as a framework.  Every facility the host language\nsupplies is one you don't have to write.\n<\/p>\n<p>\n  This is the easiest way to write a minilanguage.  Old-school\nLisp\n  programmers (including me) love this technique and use it heavily.  It\nunderlies the design of the\n  <em>Emacs<\/em> editor,\nand has been rediscovered in the new-school scripting languages like\nTcl,\nPython, and\nPerl. There are\ndrawbacks to it, however.\n<\/p>\n<p>\n  Your host language may be unable to interface to a code library\nthat you need.  Or, internally, its ontology of data types may be\ninadequate for the kind of computation you need to do.  Or, after\nmeasuring the performance of a prototype, you discover that it's too\nslow.  When any of these things happen, your solution is usually\ngoing to involve coding in C\n  (or\nC++) and integrating the\nresults into your minilanguage.\n<\/p>\n<p>\n  The option of extending a scripting\nlanguage\n  with C code, or of embedding a scripting language in a C program,\nrelies on the existence of scripting languages designed for it.  You\nextend a scripting language by telling it to dynamically load a C\nlibrary or module in such a way that the C entry points become visible\nas functions in the extended language.  You embed a scripting language\nin a C program by sending commands to an instance of the interpreter\nand receiving the results back as values in C.\n<\/p>\n<p>\n  Both techniques also rely on the ability to move data between\nthe type ontology of C and the type ontology of your scripting\nlanguage.  Some scripting languages are designed from the ground up to\nsupport this.  One such is Tcl, which we'll cover in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>.  Another is\n  <em>Guile<\/em>, an open-source dialect of the\nLisp\n  variant\nScheme.\n  <em>Guile<\/em> is shipped as a library and\nspecifically designed to be embedded in C\n  programs.\n<\/p>\n<p>\n  It is possible (though in 2003 still rather painful and\ndifficult) to extend or embed Perl.  It is very easy to extend\nPython\n  and\nonly slightly more difficult to embed it; C extension is especially\nheavily used in the Python world.  Java\n  has an interface to call \u2018native\nmethods\u2019 in C, though the practice is explicitly discouraged\nbecause it tends to break portability.  Most versions of shell are not\ndesigned for embeddability and extension, but the Korn shell (ksh93\nand later versions) is a notable exception.\n<\/p>\n<p>\n  There are lots of bad reasons not to piggyback your imperative\nminilanguage on an existing scripting\nlanguage.\nOne of the few good ones is that you actually want to implement your own\ncustom grammar for error checking.  If that's the case, then see the\nadvice about\n  <em>yacc<\/em> and\n  <em>lex<\/em> below.\n<\/p>\n<h4 id=\"id2936912\"><a href=\"#id2936912\">\u00a7<\/a>Writing a Custom Grammar<\/h4>\n<p>\n  For declarative minilanguages, one major question is whether or\nnot you should use XML as a base syntax and specify your grammar as an\nXML document type.  This may well be the right thing for elaborately\nstructured declarative minilanguages, but the same caveats we noted in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  about the\ndesign of data-file formats apply \u2014 XML might be overkill.  If\nyou don't use XML, follow the Rule of Least Surprise by supporting the\nUnix conventions we described for data files (simple token-oriented\nsyntax, supporting C\n  backslash conventions, etc.).\n<\/p>\n<p>\n  If you do need a custom grammar,\n  <em>yacc<\/em> and\n  <em>lex<\/em> (or their local equivalent in the\nlanguage you're using) should probably be your best friends, unless\nthe grammar of your language is so simple that hand-coding a\nrecursive-descent parser is trivial. Even then,\n  <em>yacc<\/em> may give you better error recovery,\nand a\n  <em>yacc<\/em> specification will be easier to\nmodify as the language syntax evolves.  See\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>\n  for a look at the\n  <em>yacc<\/em> - and\n  <em>lex<\/em> -derived tools available in different\nimplementation languages.\n<\/p>\n<p>\n  Even if you decide you must implement your own syntax, consider\nwhat mileage you can get from reusing existing tools.  If you need a\nmacro facility, consider whether preprocessing with\nm4(1)\nmight be the right answer \u2014 but consider the cautions in the\nnext section first.\n<\/p>\n<h4 id=\"macroexpansion\"><a href=\"#macroexpansion\">\u00a7<\/a>Macros \u2014 Beware!<\/h4>\n<p>\n  Macro expansion facilities were a favored tactic for language\ndesigners in early Unix; the C\n  language has one, of course, and we have seen them show up in some of\nthe more complex special-purpose minilanguages like\npic(1).\nThe\n  <em>m4<\/em> preprocessor provides a generic tool for implementing\nmacro-expanding preprocessors.\n<\/p>\n<p>\n  Macro expansion is easy to specify and implement, and you can do\na lot of cute tricks with it.  Those early designers appear to have been\ninfluenced by experience with assemblers, in which macro facilities\nwere often the only device available for structuring programs.\n<\/p>\n<p>\n  The strength of macro expansion is that it knows nothing about\nthe underlying syntax of the base language, and can be used to extend\nthat syntax.  Unfortunately, this power is very easily abused to\nproduce code that is opaque, surprising, and a fertile source of\nhard-to-characterize bugs.\n<\/p>\n<p>\n  In C, the classic example of this sort of problem is a macro\nsuch as this:\n<\/p>\n <pre class=\"programlisting\">\n#define max(x, y)\tx &gt; y ? x : y\n<\/pre>\n<p>\n  There are at least two problems with this macro.  One is that it\ncan produce surprising results if either of the arguments is an\nexpression including an operator of lower precedence than\n  <code>&gt;<\/code> or\n  <code>?:<\/code>.\nConsider the expression\n  <code>\n   max(a = b,\n++c)\n  <\/code>.  If the programmer has forgotten that\n  <code>max<\/code> is a macro, he will be expecting the\nassignment\n  <code>a = b<\/code> and the preincrement\noperation on\n  <code>c<\/code> to be executed before\nthe resulting values are passed as arguments to\n  <code>max<\/code>.\n<\/p>\n<p>\n  But that's not what will happen.  Instead, the preprocessor will\nexpand this expression to\n  <code>\n   a = b &gt; ++c ? a = b :\n++c\n  <\/code>, which the C compiler's precedence rules make it\ninterpret as\n  <code>\n   a = (b &gt; ++c ? a = b :\n++c)\n  <\/code>.  The effect will be to assign to\n  <code>a<\/code> !\n<\/p>\n<p>\n  This sort of bad interaction can be headed off by coding the\nmacro definition more defensively.\n<\/p>\n <pre class=\"programlisting\">\n#define max(x, y)\t((x) &gt; (y) ? (x) : (y))\n<\/pre>\n<p>\n  With this definition, the expansion would be\n  <code>((a = b) &gt; (++c) ? (a = b) : (++c))<\/code>.  This\nsolves one problem \u2014 but notice that\n  <code>c<\/code> may be incremented twice!  There are subtler\nversions of this trap, such as passing the macro a function-call with\nside effects.\n<\/p>\n<p>\n  In general, interactions between macros and expressions with\nside effects can lead to unfortunate results that are hard to\ndiagnose.  C's macro processor is a deliberately lightweight and\nsimple one; more powerful ones can actually get you in worse\ntrouble.\n<\/p>\n<p>\n  The\n  <em>TeX<\/em> formatting\nlanguage (see\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>) well illustrates the general\nproblem.\n  <em>TeX<\/em> is intentionally\nTuring-complete (it has conditionals, loops, and recursion), but while\nit can be made to do amazing things,\n  <em>TeX<\/em> code tends to be unreadable and\npainful to debug.  The sources for\n  <em>LaTeX<\/em>, the the most widely used\n  <em>TeX<\/em> macro package, are an instructive\nexample: they're in very good\n  <em>TeX<\/em> style, but even so are extremely\ndifficult to follow.\n<\/p>\n<p>\n  A minor problem, compared to this one, is that macro expansion\ntends to screw up error diagnostics.  The base language processor\ngenerates its error reports relative to the macro expanded text, not\nthe original the programmer is looking at.  If the relationship\nbetween the two has been obfuscated by macro expansion, the emitted\ndiagnostic can be very difficult to associate with the actual\nlocation of the error.\n<\/p>\n<p>\n  This is especially a problem with preprocessors and macros\nthat can have multiline expansions, conditionally include or exclude\ntext, or otherwise change line numbers in the expanded text.\n<\/p>\n<p>\n  Macro expansion stages that are built into a language can do\ntheir own compensation, fiddling line numbers to refer back to the\npreexpanded text.  The macro facility in\npic(1)\ndoes this, for example.  This problem is more difficult to solve when\nthe macro expansion is done by a preprocessor.\n<\/p>\n<p>\n  The C preprocessor\n  addresses this problem by emitting\n  <code>#line<\/code> directives whenever it does an inclusion\nor multiline expansion. The C compiler is expected to interpret these\nand adjust the line numbers in its error reports accordingly.\nUnfortunately,\n  <em>m4<\/em> has no such\nfacility.\n<\/p>\n<p>\n  These are reasons to use macro expansion with extreme caution.  \nOne of the long-term lessons of the Unix experience is that macros\ntend to create more problems than they solve.  Modern language and\nminilanguage designs have moved away from them.\n<\/p>\n<h4 id=\"id2937424\"><a href=\"#id2937424\">\u00a7<\/a>Language or Application Protocol?<\/h4>\n<p>\n  Another important question you need to ask is whether your\nminilanguage engine will be called interactively by other\nprograms, as a slave process.  If so, your design should probably look\nless like a conversational language for human interaction and more\nlike the kind of application protocols we looked at in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>.\n<\/p>\n<p>\n  The main difference is how carefully marked the boundaries\nof transactions are.  Human beings are good at spotting where\nconversational output from a CLI ends, and where the prompt for the\nnext input is.  They can use context to tell what's significant and\nwhat should be ignored.  Computer programs have much more trouble with \nthis.  Without either unambiguous end markers on output or advance\nknowledge of the length of the output, they can't tell when to stop\nreading.\n<\/p>\n <blockquote>\n  <p>\n   Even worse is when a program's input is buffered (often\ninadvertently, as by stdio).  A program that stops overtly reading at\nthe right place can nonetheless eat past it.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Programs in which master processes are trying to do interactive\nthings with slaved minilanguages that are not carefully designed\naround this problem are prone to deadlock as the master and slave fall\nout of synchronization (a problem we first noted in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>).\n<\/p>\n<p>\n  There are workarounds for driving minilanguages that are not so\ncarefully designed.  The prototype for most of them is the\nTcl\n  <em>expect<\/em> package.  This package \nassists conversation with CLIs.  It's built around the following\noperation: read from slave until either a given regular-expression\npattern is matched or a specified timeout elapses.  With this (and, of\ncourse, a send-to-slave operation) it's often possible to construct\nmaster programs to do reliable dialogues with slave processes even\nwhen the latter have not been tailored for the role.\n<\/p>\n<p>\n  Workalikes of the\n  <em>expect<\/em> package in\nother languages are available; a Web search for the name of your\nfavorite language with the added keywords \u201cTcl expect\u201d is\nquite likely to turn up something useful.  As a minilanguage designer,\nhowever, you would be unwise to assume that all your users will be\n  <em>expect<\/em> gurus. Even if they are, this is an\nextra glue layer and a place for things to go wrong.\n<\/p>\n<p>\n  Be aware of this issue when designing your minilanguage.  It may\nbe a good idea to add an option that changes its conversational\nbehavior to make it respond more like an application protocol, with\nunambiguous end-of-output delimiters and an analog of\n  <a href=\"#byte_stuffing\">byte stuffing<\/a>.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2936473\" id=\"ftn.id2936473\">[95]<\/a><\/sup>\n  20M is a conservative estimate based on\nmid-2003 figures from the Linux Counter and\nelsewhere.\n<\/p><p><sup><a href=\"#id2936490\" id=\"ftn.id2936490\">[96]<\/a><\/sup><em>Kmail<\/em>, which we looked\nat in\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>, won't\neven chase off-site links in HTML for this\nreason.\n<\/p><\/div>\n\n<h3 id=\"generationchapter\"><a href=\"#generationchapter\">\u00a7<\/a>Chapter 9. Generation<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   The programmer at wit's end ... can often do best by\ndisentangling himself from his code, rearing back, and\ncontemplating his data. Representation is the essence of\nprogramming.\n  <\/p><div>\n  \n   \u2013 Fred Brooks\nThe Mythical Man-Month, Anniversary Edition (1975-1995), p. 103\n  \n <\/div><\/blockquote>\n<p>\n  In\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>\n  we\nobserved that human beings are better at visualizing data than they\nare at reasoning about control flow.  We recapitulate: To see this,\ncompare the expressiveness and explanatory power of a diagram of a\nfifty-node pointer tree with a flowchart of a fifty-line program. Or\n(better) of an array initializer expressing a conversion table with an\nequivalent switch statement. The difference in\ntransparency\n  and clarity is\ndramatic.<sup><a href=\"#ftn.id2939543\" id=\"id2939543\">[97]<\/a><\/sup><\/p>\n<p>\n  Data is more tractable than program logic. That's\ntrue whether the data is an ordinary table, a declarative markup\nlanguage, a templating system, or a set of macros that will expand to\nprogram logic.  It's good practice to move as much of the complexity\nin your design as possible away from procedural code and into\ndata, and good practice to pick data representations that are\nconvenient for humans to maintain and manipulate.  Translating\nthose representations into forms that are convenient for machines\nto process is another job for machines, not for humans.\n<\/p>\n <blockquote>\n  <p>\n   Another important advantage of higher-level, more declarative\nnotations is that they lend themselves better to compile-time checking.\nProcedural notations inherently have complex runtime behavior which is\ndifficult to analyze at compile time.  Declarative notations give the\nimplementation much more leverage for finding mistakes, by permitting much\nmore thorough understanding of the intended behavior.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  These insights ground in theory a set of practices that have\nalways been an important part of the Unix programmer's toolkit \u2014\nvery high-level languages, data-driven programming, code generators,\nand domain-specific minilanguages.  What unifies these is that they\nare all ways of lifting the generation of code up some levels, so that\nspecifications can be smaller. We've previously noted that defect\ndensities tend to be nearly constant across programming languages; all\nthese practices mean that whatever malign forces generate our bugs\nwill get fewer lines to wreak their havoc on.\n<\/p>\n<p>\n  In\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  we\ndiscussed the uses of domain-specific minilanguages. In\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  we'll make the argument for\nvery-high-level languages.  In this chapter we'll look at some\ndesign studies in data-driven programming and a few examples of ad-hoc\ncode generation; we'll look at some code-generation tools in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>.  As with minilanguages, these methods can\nenable you to drastically cut the line count of your programs, and\ncorrespondingly lower debugging time and maintenance costs.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2939543\" id=\"ftn.id2939543\">[97]<\/a><\/sup>\n  For further development of this point see<sup><a href=\"#Bentley\" title=\"[Bentley]\">[Bentley]<\/a><\/sup>.\n<\/p><\/div>\n\n<h3 id=\"ch09s01\"><a href=\"#ch09s01\">\u00a7<\/a>Data-Driven Programming<\/h3>\n<p>\n  When doing\n  <em>data-driven programming<\/em>, one\nclearly distinguishes code from the data structures on which it acts,\nand designs both so that one can make changes to the logic of the program \nby editing not the code but the data structure.\n<\/p>\n<p>\n  Data-driven programming is sometimes confused with object\norientation, another style in which data\norganization is supposed to be central.  There are at least two\ndifferences.  One is that in data-driven programming, the data is not\nmerely the state of some object, but actually defines the control flow\nof the program.  Where the primary concern in\nOO\n  is\nencapsulation, the primary concern in data-driven programming is\nwriting as little fixed code as possible.  Unix has a stronger\ntradition of data-driven programming than of OO.\n<\/p>\n<p>\n  Programming data-driven style is also sometimes confused with\nwriting state machines.  It is in fact possible to express the logic\nof a state machine as a table or data structure, but hand-coded state\nmachines are usually rigid blocks of code that are far harder to\nmodify than a table.\n<\/p>\n<p>\n  An important rule when doing any kind of code generation or\ndata-driven programming is this: always push problems upstream.  Don't\nhack the generated code or any intermediate representations by hand\n\u2014 instead, think of a way to improve or replace your translation\ntool.  Otherwise you're likely to find that hand-patching bits which\nshould have been generated correctly by machine will have turned into\nan infinite time sink.\n<\/p>\n<p>\n  At the upper end of its complexity scale, data-driven\nprogramming merges into writing interpreters for p-code or simple\nminilanguages of the kind we surveyed in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>.  At other edges, it merges into code\ngeneration and state-machine programming.  The distinctions are not\nactually that important; the important part is moving program logic\naway from hardwired control structures and into data.\n<\/p>\n<h4 id=\"id2939746\"><a href=\"#id2939746\">\u00a7<\/a>Case Study: ascii<\/h4>\n<p>\n  I maintain a program called\n  <em>ascii<\/em>, a\nvery simple little utility that tries to interpret its command-line\narguments as names of ASCII (American Standard Code for Information\nInterchange) characters and report all the equivalent names.  Code and\ndocumentation for the tool are available from the\n  <a href=\"http:\/\/www.catb.org\/~esr\/ascii\">\n   project page<\/a>. Here is an illustrative\nscreenshot:\n<\/p>\n <pre class=\"screen\">\nesr@snark:~\/WWW\/writings\/taoup$ ascii 10\nASCII 1\/0 is decimal 016, hex 10, octal 020, bits 00010000: called ^P, DLE\nOfficial name: Data Link Escape\nASCII 0\/10 is decimal 010, hex 0a, octal 012, bits 00001010: called ^J, LF, NL\nOfficial name: Line Feed\nC escape: '\\n'\nOther names: Newline\nASCII 0\/8 is decimal 008, hex 08, octal 010, bits 00001000: called ^H, BS\nOfficial name: Backspace\nC escape: '\\b'\nOther names:\nASCII 0\/2 is decimal 002, hex 02, octal 002, bits 00000010: called ^B, STX\nOfficial name: Start of Text\n<\/pre>\n<p>\n  One indication that this program was a good idea is the fact\nthat it has an unexpected use \u2014 as a quick CLI aid to converting\nbetween decimal, hex, octal, and binary representations of bytes.\n<\/p>\n<p>\n  The main logic of this program could have been coded as a \n128-branch case statement.  This would, however, have made the code\nbulky and difficult to maintain.  It would also have tangled parts\nthat change relatively rapidly (like the list of slang names for\ncharacters) with parts that change slowly or not at all (like\nthe official names), putting them both in the same legend string and\nmaking errors during editing much more likely to touch data that\nought to be stable.\n<\/p>\n<p>\n  Instead, we apply data-driven programming.  All the character\nname strings live in a table structure that is quite a bit larger than\nany of the functions in the code (indeed, counted in lines it is\nlarger than any\n  <em>three<\/em>\n  of the functions in the\nprogram).  The code merely navigates the table and does low-level\ntasks like radix conversions.  The initializer actually lives in the\nfile\n  <code>nametable.h<\/code>, which is generated in a way\nwe'll describe later in this chapter.\n<\/p>\n<p>\n  This organization makes it easy to add new character names,\nchange existing ones, or delete old names by simply editing the table,\nwithout disturbing the code.\n<\/p>\n<p>\n  (The way the program is built is good Unix style, but the\noutput format is questionable.  It's hard to see how the output could\nusefully become the input of any other program, so it does not play\nwell with others.)\n<\/p>\n<h4 id=\"bayes_spam\"><a href=\"#bayes_spam\">\u00a7<\/a>Case Study: Statistical Spam Filtering<\/h4>\n<p>\n  One interesting case of data-driven programming is statistical\nlearning algorithms for detecting spam (unsolicited bulk email).\nA whole class of mail filter programs (those easily findable\nby Web search include\n  <em>popfile<\/em>,\n  <em>spambayes<\/em>, and\n  <em>bogofilter<\/em>) use a database of word\ncorrelations to replace the elaborate pattern-matching conditional\nlogic of pattern-matching spam filters.\n<\/p>\n<p>\n  Programs like these became common on the Internet very rapidly\nfollowing Paul Graham's landmark paper\n  <em>A Plan for Spam<\/em><sup><a href=\"#Graham\" title=\"[Graham]\">[Graham]<\/a><\/sup> in 2002.  While the\nexplosion was triggered by the increasing cost of the pattern-matching\narms race, the statistical-filtering idea was adopted first and\nfastest by Unix shops.  In part, this was certainly because almost all\nthe Internet service providers (who are most burdened by spam, and\nthus had most incentive to adopt effective new techniques) are Unix\nshops \u2014 but undoubtedly the harmony with some traditional themes\nin Unix software design helped as well.\n<\/p>\n<p>\n  Conventional spam filters require that a system administrator,\nor some other responsible party, maintain information on patterns of\ntext found in spam \u2014 names of sites that emit nothing but spam,\ncome-on phrases often used by pornography sites or Internet scam\nartists, and the like.  In his paper, Graham noted accurately that\ncomputer programmers like the idea of pattern-matching filters, and\nsometimes have difficulty seeing past that approach, because it offers\nthem so many opportunities to be clever.\n<\/p>\n<p>\n  Statistical spam filters, on the other hand, work by collecting\nfeedback about what the user judges to be spam versus nonspam.  That\nfeedback is processed into databases of statistical correlation\ncoefficients or\n  <em>weights<\/em> connecting words or phrases\nto the user's spam\/nonspam classification.  The most popular\nalgorithms use minor variants of Bayes's Theorem on conditional\nprobabilities, but other techniques (including various sorts of\npolynomial hashing) are also employed.\n<\/p>\n<p>\n  In all these programs, the correlation check is a relatively\ntrivial mathematical formula.  The weights fed into the formula along\nwith the message being checked serve as implicit control\nstructure for the filtering algorithm.\n<\/p>\n<p>\n  The problem with conventional pattern-matching spam filters is \nthat they are brittle.  Spammers are constantly gaming against the\nfilter-rule databases, forcing the filter maintainers to constantly\nreprogram their filters to stay ahead in the arms race.  Statistical\nspam filters generate their own filter rules from the user feedback.\n<\/p>\n<p>\n  In fact, experience with statistical filters seems to show that the \nparticular learning algorithm used is far less important than the\nquality of the spam and nonspam data sets from which the learning algorithm\ncomputes its weights.  So the results of statistical filters really\nare driven more by the shape of the data than by the algorithm.\n<\/p>\n<p>\n  <em>A Plan for Spam<\/em> was something of a\nbombshell because its author argued convincingly that a simple, even\ncrude, statistical approach gave a lower rate of nonspam being\nerroneously classified as spam than either elaborate pattern-matching\ntechniques or the human eyeball could manage.  For Unix programmers,\nseeing past the lure of clever pattern-matching was far easier than in\nother programming cultures without as strong an attachment to\n\u201cKeep It Simple, Stupid!\u201d\n<\/p>\n<h4 id=\"fetchmailconf\"><a href=\"#fetchmailconf\">\u00a7<\/a>Case Study: Metaclass Hacking in fetchmailconf<\/h4>\n<p>\n  The\nfetchmailconf(1)\ndotfile configurator shipped with\nfetchmail(1)\n  contains an instructive example of advanced data-driven programming in\na very high-level, object-oriented\n  language.\n<\/p>\n<p>\n  In October 1997 a series of questions on the fetchmail-friends mailing list made it clear that end-users were having\n  increasing troubles generating configuration files for <em>fetchmail<\/em>.  The file uses a simple, classically-Unixy\n  free-format syntax, but can become forbiddingly complicated when a user has POP3 and IMAP accounts at multiple sites.\n  See Example 9.1 for a somewhat simplified version of the <em>fetchmail<\/em> author's configuration file.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.1. Example of fetchmailrc syntax.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nset postmaster \"esr\"\nset daemon 300\npoll imap.ccil.org with proto IMAP and options no dns\n    aka snark.thyrsus.com locke.ccil.org ccil.org\n       user esr there is esr here \n            options fetchall dropstatus warnings 3600\npoll imap.netaxs.com with proto IMAP\n       user \"esr\" there is esr here options dropstatus warnings 3600\n<\/pre><\/div>\n<p>\n  The design objective of\n  <em>fetchmailconf<\/em> was to completely hide the control file syntax behind a fashionable,\nergonomically-correct GUI replete with selection buttons,\nslider bars and fill-out forms.  But the beta design had a problem: it\ncould easily generate configuration files from the user's GUI actions,\nbut could not read and edit existing ones.\n<\/p>\n<p>\n  The parser for\n  <em>fetchmail<\/em> 's\n  configuration file syntax is rather elaborate.  It's actually written\nin\n  <em>yacc<\/em> and\n  <em>lex<\/em>,\nthe two classic Unix tools for generating language-parsing code in C.\nFor\n  <em>fetchmailconf<\/em> to be able to edit\nexisting configuration files, it at first appeared that it would be\nnecessary to replicate that elaborate parser in fetchmailconf's\nimplementation language \u2014 Python.\n<\/p>\n<p>\n  This tactic seemed doomed.  Even leaving aside the amount of\nduplicative work implied, it is notoriously hard to be certain that\ntwo parsers in two different languages accept the same grammar.\nKeeping them synchronized as the configuration language evolved bid\nfair to be a maintenance nightmare.  It would have violated the SPOT\nrule\n  we discussed\nin\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  wholesale.\n<\/p>\n<p>\n  This problem stumped me for a while.  The insight that cracked\nit was that\n  <em>fetchmailconf<\/em> could use\n  <em>fetchmail<\/em> 's own parser as a filter!  I\nadded a\n  <code>--configdump<\/code> option to\n  <em>fetchmail<\/em> that would parse\n  <code>.fetchmailrc<\/code> and dump the result to standard\noutput in the format of a Python initializer.  For the file above, the result would look roughly like Example 9.2 (to\nsave space, some data not relevant to the example is omitted).\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.2. Python structure dump of a\n    <em>fetchmail<\/em> configuration.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nfetchmailrc = {\n    'poll_interval':300,\n    \"logfile\":None,\n    \"postmaster\":\"esr\",\n    'bouncemail':TRUE,\n    \"properties\":None,\n    'invisible':FALSE,\n    'syslog':FALSE,\n    # List of server entries begins here\n    'servers': [\n        # Entry for site `imap.ccil.org' begins:\n        {\n            \"pollname\":\"imap.ccil.org\",\n            'active':TRUE,\n            \"via\":None,\n            \"protocol\":\"IMAP\",\n            'port':0,\n            'timeout':300,\n            'dns':FALSE,\n            \"aka\":[\"snark.thyrsus.com\",\"locke.ccil.org\",\"ccil.org\"],\n            'users': [\n                {\n                    \"remote\":\"esr\",\n                    \"password\":\"masked_one\",\n                    'localnames':[\"esr\"],\n                    'fetchall':TRUE,\n                    'keep':FALSE,\n                    'flush':FALSE,\n                    \"mda\":None,\n                    'limit':0,\n                    'warnings':3600,\n                },            ]\n        },\n        # Entry for site `imap.netaxs.com' begins:\n        {\n            \"pollname\":\"imap.netaxs.com\",\n            'active':TRUE,\n            \"via\":None,\n            \"protocol\":\"IMAP\",\n            'port':0,\n            'timeout':300,\n            'dns':TRUE,\n            \"aka\":None,\n            'users': [\n                {\n                    \"remote\":\"esr\",\n                    \"password\":\"masked_two\",\n                    'localnames':[\"esr\"],\n                    'fetchall':FALSE,\n                    'keep':FALSE,\n                    'flush':FALSE,\n                    \"mda\":None,\n                    'limit':0,\n                    'warnings':3600,\n                },            ]\n        },\n    ]\n}\n<\/pre><\/div>\n<p>\n  The major hurdle had been leapt.  The Python interpreter could\nthen evaluate the\n  <em>fetchmail<\/em> <code>--configdump<\/code> output and read the configuration\navailable to\n  <em>fetchmailconf<\/em> as the value of\nthe variable \u2018fetchmail\u2019.\n<\/p>\n<p>\n  But this wasn't quite the last obstacle in the race.  What was\nreally needed wasn't just for\n  <em>fetchmailconf<\/em> to have the existing configuration, but to turn it into a linked tree\nof live objects.  There would be three kinds of objects in this tree:\n  <code>Configuration<\/code> (the top-level object representing the\nentire configuration),\n  <code>Site<\/code> (representing one of the\nservers to be polled), and\n  <code>User<\/code> (representing user data\nattached to a site).  The example file describes three site objects, each\nwith one user object attached to it.\n<\/p>\n<p>\n  The three object classes already existed in\n  <em>fetchmailconf<\/em>.  Each had a method that\ncaused it to pop up a GUI edit panel to modify its instance data.  The\nlast remaining problem was to somehow transform the static data in this\nPython initializer into live objects.\n<\/p>\n<p>\n  I considered writing a glue layer that would explicitly\nknow about the structure of all three classes and use that knowledge\nto grovel through the initializer creating matching objects, but\nrejected that idea because new class members were likely to be added\nover time as the configuration language grew new features.  If the\nobject-creation code were written in the obvious way, it would once\nagain be fragile and tend to fall out of synchronization when either\nthe class definitions or the initializer structure dumped by the\n  <code>--configdump<\/code> report generator changed.  Again, a\nrecipe for endless bugs.\n<\/p>\n<p>\n  The better way would be data-driven programming \u2014 code\nthat would analyze the shape and members of the initializer, query the\nclass definitions themselves about their members, and then\nimpedance-match the two sets.\n<\/p>\n<p>\n  Lisp\n  and\nJava\n  programmers call this\n  <em>introspection<\/em>; in\nsome other object-oriented languages it's called\n  <em>\n   metaclass\nhacking\n<\/em>\n  and is generally considered fearsomely esoteric,\ndeep black magic.  Most object-oriented languages don't support it at\nall; in those that do (Perl\n  and Java among them), it tends to be a\ncomplicated and fragile undertaking.  But Python's facilities for\nintrospection and metaclass hacking are unusually accessible.\n<\/p>\n<p>\n  See Example 9.3 for the solution code, from near line 1895 of the 1.43 version.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.3.\n    <code>copy_instance<\/code> metaclass code.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\ndef copy_instance(toclass, fromdict):\n# Make a class object of given type from a conformant dictionary.\n    class_sig = toclass.__dict__.keys(); class_sig.sort()\n    dict_keys = fromdict.keys(); dict_keys.sort()\n    common = set_intersection(class_sig, dict_keys)\n    if 'typemap' in class_sig: \n        class_sig.remove('typemap')\n    if tuple(class_sig) != tuple(dict_keys):\n        print \"Conformability error\"\n#        print \"Class signature: \" + `class_sig`\n#        print \"Dictionary keys: \" + `dict_keys`\n        print \"Not matched in class signature: \"+ \\\n                                        `set_diff(class_sig, common)`\n        print \"Not matched in dictionary keys: \"+ \\\n                                        `set_diff(dict_keys, common)`\n        sys.exit(1)\n    else:\n        for x in dict_keys:\n            setattr(toclass, x, fromdict[x])\n<\/pre><\/div>\n<p>\n  Most of this code is error-checking against the possibility that\nthe class members and\n  <code>--configdump<\/code> report generation\nhave drifted out of synchronization.  It ensures that if the code\nbreaks, the breakage will be detected early \u2014 an implementation\nof the Rule of Repair.  The heart of this function is the last\ntwo lines, which set attributes in the class from corresponding\nmembers in the dictionary.  They're equivalent to this:\n<\/p>\n <pre class=\"programlisting\">\ndef copy_instance(toclass, fromdict):\n        for x in fromdict.keys():\n                setattr(toclass, x, fromdict[x])\n<\/pre>\n<p>\n  When your code is this simple, it is far more likely to be right. See Example 9.4 for the code that calls it.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.4. Calling context for\n    <code>copy_instance<\/code>.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n    # The tricky part - initializing objects from the `configuration' \n    # global.  `Configuration' is the top level of the object tree \n    # we're going to mung \n    Configuration = Controls()\n    copy_instance(Configuration, configuration)\n    Configuration.servers = [];\n    for server in configuration['servers']:\n        Newsite = Server()\n        copy_instance(Newsite, server)\n        Configuration.servers.append(Newsite)\n        Newsite.users = [];\n        for user in server['users']:\n            Newuser = User()\n            copy_instance(Newuser, user)\n            Newsite.users.append(Newuser)\n<\/pre><\/div>\n<p>\n  The key point to extract from this code is that it traverses the\nthree levels of the initializer (configuration\/server\/user),\ninstantiating the correct objects at each level into lists contained\nin the next object up.  Because\n  <code>copy_instance<\/code> is\ndata-driven and completely generic, it can be used on all three levels\nfor three different object types.\n<\/p>\n<p>\n  This is a new-school sort of example; Python was not even\ninvented until 1990.  But it reflects themes that go back to 1969 in\nthe Unix tradition.  If meditating on Unix programming as practiced by\nhis predecessors had not taught me constructive laziness\n\u2014 insisting on reuse, and refusing to write duplicative glue\ncode in accordance with the SPOT rule\n  \u2014I might have rushed into coding a\nparser in Python.  The first key insight that\n  <em>fetchmail<\/em> itself could be made into\n  <em>fetchmailconf<\/em> 's configuration parser might\nnever have happened.\n<\/p>\n<p>\n  The second insight (that\n  <code>copy_instance<\/code> could be generic) proceeded from\nthe Unix tradition of looking assiduously for ways to avoid\nhand-hacking. But more specifically, Unix programmers are very used to\nwriting parser specifications to generate parsers for processing\nlanguage-like markups; from there it was a short step to believing\nthat the rest of the job could be done by some kind of generic\ntree-walk of the configuration structure. Two separate stages of\ndata-driven programming, one building on the other, were needed to\nsolve the design problem cleanly.\n<\/p>\n<p>\n  Insights like this can be extraordinarily powerful.  The code we\nhave been looking at was written in about ninety minutes, worked the\nfirst time it was run, and has been stable in the years since (the\nonly time it has ever broken is when it threw an exception in the\npresence of genuine version skew).  It's less than forty lines and\nbeautifully simple.  There is no way that the na\u00efve approach of\nbuilding an entire second parser could possibly have produced this\nkind of maintainability, reliability or\ncompactness.\nReuse, simplification, generalization,\northogonality\n  :\nthis is the\n  Zen of Unix\nin action.\n<\/p>\n<p>\n  In\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">\n   Chapter 10<\/a>,\nwe'll examine the run-control syntax of\n  <em>fetchmail<\/em> as an example of the standard shell-like metaformat for run-control\nfiles.  In\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  we'll use\n  <em>fetchmailconf<\/em> as an example of\nPython's strength in rapidly building GUIs.\n<\/p>\n\n<h3 id=\"ch09s02\"><a href=\"#ch09s02\">\u00a7<\/a>Ad-hoc Code Generation<\/h3>\n<p>\n  Unix comes equipped with some powerful special-purpose code\ngenerators for purposes like building lexical analyzers (tokenizers)\nand parsers; we'll survey these in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>.  But there are much simpler, lighter-weight\nsorts of code generation we can use to make life easier without having\nto know any compiler theory or write (error-prone) procedural\nlogic.\n<\/p>\n<p>\n  Here are a couple of simple case studies to illustrate this\npoint:\n<\/p>\n<h4 id=\"id2938615\"><a href=\"#id2938615\">\u00a7<\/a>Case Study: Generating Code for the ascii Displays<\/h4>\n<p>\n  Called without arguments, <em>ascii<\/em> generates a usage screen that looks like Example 9.5.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.5.\n    <em>ascii<\/em> usage screen.\n   <\/b>\n  <\/p><pre class=\"screen\">\nUsage: ascii [-dxohv] [-t] [char-alias...]\n   -t = one-line output  -d = Decimal table  -o = octal table  -x = hex table\n   -h = This help screen -v = version information\nPrints all aliases of an ASCII character. Args may be chars, C \\-escapes,\nEnglish names, ^-escapes, ASCII mnemonics, or numerics in decimal\/octal\/hex.\nDec Hex    Dec Hex    Dec Hex  Dec Hex  Dec Hex  Dec Hex   Dec Hex   Dec Hex\n  0 00 NUL  16 10 DLE  32 20    48 30 0  64 40 @  80 50 P   96 60 `  112 70 p\n  1 01 SOH  17 11 DC1  33 21 !  49 31 1  65 41 A  81 51 Q   97 61 a  113 71 q\n  2 02 STX  18 12 DC2  34 22 \"  50 32 2  66 42 B  82 52 R   98 62 b  114 72 r\n  3 03 ETX  19 13 DC3  35 23 #  51 33 3  67 43 C  83 53 S   99 63 c  115 73 s\n  4 04 EOT  20 14 DC4  36 24 $  52 34 4  68 44 D  84 54 T  100 64 d  116 74 t\n  5 05 ENQ  21 15 NAK  37 25 %  53 35 5  69 45 E  85 55 U  101 65 e  117 75 u\n  6 06 ACK  22 16 SYN  38 26 &amp;  54 36 6  70 46 F  86 56 V  102 66 f  118 76 v\n  7 07 BEL  23 17 ETB  39 27 '  55 37 7  71 47 G  87 57 W  103 67 g  119 77 w\n  8 08 BS   24 18 CAN  40 28 (  56 38 8  72 48 H  88 58 X  104 68 h  120 78 x\n  9 09 HT   25 19 EM   41 29 )  57 39 9  73 49 I  89 59 Y  105 69 i  121 79 y\n 10 0A LF   26 1A SUB  42 2A *  58 3A :  74 4A J  90 5A Z  106 6A j  122 7A z\n 11 0B VT   27 1B ESC  43 2B +  59 3B ;  75 4B K  91 5B [  107 6B k  123 7B {\n 12 0C FF   28 1C FS   44 2C ,  60 3C &lt;  76 4C L  92 5C \\  108 6C l  124 7C |\n 13 0D CR   29 1D GS   45 2D -  61 3D =  77 4D M  93 5D ]  109 6D m  125 7D }\n 14 0E SO   30 1E RS   46 2E .  62 3E &gt;  78 4E N  94 5E ^  110 6E n  126 7E ~\n 15 0F SI   31 1F US   47 2F \/  63 3F ?  79 4F O  95 5F _  111 6F o  127 7F DEL\n<\/pre><\/div>\n<p>\n  This screen is carefully designed to fit in 23 rows and 79\ncolumns, so that it will fit in a 24\u00d780 terminal window.\n<\/p>\n<p>\n  This table could be generated at runtime, on the fly.  Grinding\nout the decimal and hex columns would be easy enough.  But between\nwrapping the table at the right places and knowing when to print\nmnemonics like NUL rather than characters, there would have been enough\nodd corner cases to make the code distinctly unpleasant.  Furthermore,\nthe columns had to be unevenly spaced to make the table fit in 79\ncolumns.  But any Unix programmer would reflexively express it as a block\nof data before finding out these things.\n<\/p>\n<p>\n  The most na\u00efve way to generate the usage screen would have been\nto put each line into a C\n  initializer in the\n  <code>ascii.c<\/code> source code, and then\nhave all lines be written out by code that steps through the\ninitializer.  The problem with this method is that the extra data in\nthe C initializer format (trailing newline, string quotes, comma) would\nmake the lines longer than 79 characters, causing them to wrap and\nmaking it rather difficult to map the appearance of the code to the\nappearance of the output.  This, in turn, would make the display\ndifficult to edit, which was annoying when I was tinkering it\nto fit in 24\u00d780 screen cells.\n<\/p>\n<p>\n  A more sophisticated method using the string-pasting behavior of\nthe ANSI C\n  preprocessor collided with a variant of the same problem.\nEssentially, any way of inlining the usage screen explicitly would\ninvolve punctuation at start and end of line that there's no room\nfor.<sup><a href=\"#ftn.id2938770\" id=\"id2938770\">[98]<\/a><\/sup>\n  And copying the table to the screen from a\nfile at runtime seemed like a fragile expedient; after all, the file\ncould get lost.\n<\/p>\n<p>\n  Here's the solution.  The source distribution contains a file\nthat just contains the usage screen, exactly as listed above and named\n  <code>splashscreen<\/code>. The C\n  source contains the following\nfunction:\n<\/p>\n <pre class=\"programlisting\">\nvoid \nshowHelp(FILE *out, char *progname) \n{\n  fprintf(out,\"Usage: %s [-dxohv] [-t] [char-alias...]\\n\", progname);\n#include \"splashscreen.h\"\n  exit(0);\n}\n<\/pre>\n<p>\n  And\n  <code>splashscreen.h<\/code> is generated by a makefile production:\n<\/p>\n <pre class=\"programlisting\">\nsplashscreen.h: splashscreen\n        sed &lt;splashscreen &gt;splashscreen.h \\\n            -e 's\/\\\\\/\\\\\\\\\/g' -e 's\/\"\/\\\\\"\/' -e 's\/.*\/puts(\"&amp;\");\/' \n<\/pre>\n<p>\n  So when the program is built, the\n  <code>splashscreen<\/code> file is automatically massaged into\na series of output function calls, which are then included by the C\npreprocessor in the right function.\n<\/p>\n<p>\n  By generating the code from data, we get to keep the editable\nversion of the usage screen identical to its display appearance.  This\npromotes\ntransparency.  Furthermore, we could modify the usage\nscreen at will without touching the C code at all, and the right thing\nwould automatically happen on the next build.\n<\/p>\n<p>\n  For similar reasons, the initializer that holds the name synonym\nstrings is also generated via a\n  <em>sed<\/em> script\nin the makefile, from a file called\n  <code>nametable<\/code> in\nthe\n  <em>ascii<\/em> source distribution.  Most of\n  <code>nametable<\/code> is simply copied into the C\ninitializer.  But the generation process would make it easy to adapt\nthis tool for other 8-bit character sets such as the ISO-8859 series\n(Latin-1 and friends).\n<\/p>\n<p>\n  This is an almost trivial example, but it nevertheless\nillustrates the advantages of even simple and ad-hoc code\ngeneration.  Similar techniques could be applied to larger\nprograms with correspondingly greater benefits.\n<\/p>\n<h4 id=\"htmlgen\"><a href=\"#htmlgen\">\u00a7<\/a>Case Study: Generating HTML Code for a Tabular List<\/h4>\n<p>\n  Let's suppose that we want to put a page of tabular data on a Web page.  We want the first few lines to look like\n  Example 9.6.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.6. Desired output format for the star table.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nAalat         David Weber             The Armageddon Inheritance\nAelmos        Alan Dean Foster        The Man who Used the Universe \nAedryr        Steve Miller\/Sharon Lee Scout's Progress \nAergistal     Gerard Klein            The Overlords of War \nAfdiar        L. Neil Smith           Tom Paine Maru \nAgandar       Donald Kingsbury        Psychohistorical Crisis \nAghirnamirr   Jo Clayton              Shadowkill \n<\/pre><\/div>\n<p>\n  The thick-as-a-plank way to handle this would be to hand-write\nHTML table code for the desired appearance.  Then, each time we want\nto add a name, we'd have to hand-write another set of\n&lt;tr&gt; and &lt;td&gt; tags for\nthe entry.  This would get very tedious very quickly.  But what's\nworse, changing the format of the list would require hand-hacking\nevery entry.\n<\/p>\n<p>\n  The superficially clever way to handle this would be to make\nthis data a three-column relation in a database, then use some fancy\nCGI<sup><a href=\"#ftn.id2941426\" id=\"id2941426\">[99]<\/a><\/sup>\n  technique or a database-capable templating\nengine like PHP to generate the page on the fly.  But suppose we know\nthat the list will not change very often, don't want to run a database\nserver just to be able to display this list, and don't want to load\nthe server with unnecessary CGI traffic?\n<\/p>\n<p>\n  There's a better solution.  We put the data in a tabular flat-file format like Example 9.7.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 9.7. Master form of the star table.\n   <\/b>\n  <\/p><pre class=\"screen\">\nAalat         :David Weber                 :The Armageddon Inheritance\nAelmos        :Alan Dean Foster            :The Man who Used the Universe \nAedryr        :Steve Miller\/Sharon Lee     :Scout's Progress \nAergistal     :Gerard Klein                :The Overlords of War \nAfdiar        :L. Neil Smith               :Tom Paine Maru \nAgandar       :Donald Kingsbury            :Psychohistorical Crisis \nAghirnamirr   :Jo Clayton                  :Shadowkill \n<\/pre><\/div>\n<p>\n  We could in a pinch have done without the explicit colon field\ndelimiters, using the pattern consisting of two or more spaces as a\ndelimiter, but the explicit delimiter protects us in case we press\nspacebar twice while editing a field value and fail to notice\nit.\n<\/p>\n<p>\n  We then write a script in shell, Perl, Python, or Tcl\n  that massages\nthis file into an HTML table, and run that each time we add an entry.\nThe old-school Unix way would revolve around the following\nnigh-unreadable sed(1) invocation\n<\/p>\n <pre class=\"programlisting\">\nsed -e 's,^,&lt;tr&gt;&lt;td&gt;,' -e 's,$,&lt;\/td&gt;&lt;\/tr&gt;,' -e 's,:,&lt;\/td&gt;&lt;td&gt;,g'\n<\/pre>\n<p>\n  or this perhaps slightly more scrutable awk(1) program:\n<\/p>\n <pre class=\"programlisting\">\nawk -F: '{printf(\"&lt;tr&gt;&lt;td&gt;%s&lt;\/td&gt;&lt;td&gt;%s&lt;\/td&gt;&lt;td&gt;%s&lt;\/td&gt;&lt;\/tr&gt;\\n\", \\\n                 $1, $2, $3)}'\n<\/pre>\n<p>\n  (If either of these examples interests but mystifies, read the\ndocumentation for\nsed(1)\nor\nawk(1).\nWe\n  <a href=\"#awk\" title=\"Case Study: awk\">\n   explained<\/a>\n  in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  that the latter has largely fallen\nout of use.  The former is still an important Unix tool that we\nhaven't examined in detail because (a) Unix programmers already know\nit, and (b) it's easy for non-Unix programmers to pick up from the\nmanual page once they grasp the basic ideas about pipelines and\nredirection.)\n<\/p>\n<p>\n  A new-school solution might center on this Python code, or\non equivalent Perl:\n<\/p>\n <pre class=\"programlisting\">\nfor row in map(lambda x:x.rstrip().split(':'),sys.stdin.readlines()):\n    print \"&lt;tr&gt;&lt;td&gt;\" + \"&lt;\/td&gt;&lt;td&gt;\".join(row) + \"&lt;\/td&gt;&lt;\/tr&gt;\"\n<\/pre>\n<p>\n  These scripts took about five minutes each to write and debug,\ncertainly less time than would have been required to either hand-hack\nthe initial HTML or create and verify the database.  The combination\nof the table and this code will be much simpler to maintain than\neither the under-engineered hand-hacked HTML or the over-engineered\ndatabase.\n<\/p>\n<p>\n  A further advantage of this way of solving the problem is that\nthe master file stays easy to search and modify with an ordinary text\neditor.  Another is that we can experiment with different\ntable-to-HTML transformations by tweaking the generator script, or\neasily make a subset of the report by putting a\ngrep(1)\nfilter before it.\n<\/p>\n<p>\n  I actually use this technique to maintain the Web page that\nlists\n  <em>fetchmail<\/em> test sites; the example above is science-fictional only because\npublishing the real data would reveal account usernames and\npasswords.\n<\/p>\n<p>\n  This was a somewhat less trivial example than the previous\none. What we've actually designed here is a separation between content\nand formatting, with the generator script acting as a stylesheet.\n(This is yet another mechanism-vs.-policy separation.)\n<\/p>\n<p>\n  The lesson in all these cases is the same.  Do as little work\nas possible.  Let the data shape the code. Lean on your tools.\nSeparate mechanism from policy.  Expert Unix programmers learn to see\npossibilities like these quickly and automatically.  Constructive\nlaziness is one of the cardinal virtues of the master programmer.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2938770\" id=\"ftn.id2938770\">[98]<\/a><\/sup>\n  Scripting languages tend to solve this problem more\nelegantly than C does.  Investigate the shell's\n  <em>here documents<\/em>\n  and Python's triple-quote construct to find out\nhow.\n<\/p><p><sup><a href=\"#id2941426\" id=\"ftn.id2941426\">[99]<\/a><\/sup>\n  Here, CGI refers not to Computer Graphic Inagery\nbut to the Common Gateway Interface used for live Web\ncontent.\n<\/p><\/div>\n\n<h3 id=\"configurationchapter\"><a href=\"#configurationchapter\">\u00a7<\/a>Chapter 10. Configuration<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Let us watch well our beginnings, and results will manage themselves.\n  <\/p><div>\n  \n   \u2013 Alexander Clark\n  \n <\/div><\/blockquote>\n<p>\n  Under Unix, programs can communicate with their environment in a\nrich variety of ways. It's convenient to divide these into (a)\nstartup-environment queries and (b) interactive channels. In this\nchapter, we'll focus primarily on startup-environment queries. The\nnext chapter will discuss interactive channels.\n<\/p>\n\n<h3 id=\"ch10s01\"><a href=\"#ch10s01\">\u00a7<\/a>What Should Be Configurable?<\/h3>\n<p>\n  Before plunging into the details of different kinds of program\nconfiguration, we should ask a high-level question: What things should \nbe configurable?\n<\/p>\n<p>\n  The gut-level Unix answer is \u201ceverything\u201d.  The\nRule of Separation that we discussed in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">Chapter 1<\/a>\n  encourages Unix programmers to build\nmechanism and defer policy decisions outward toward the user wherever\npossible.  While this tends to produce programs that are powerful and\nrewarding for expert users, it also tends to produce interfaces that\noverwhelm novices and casual users with a surfeit of choices, and with\nconfiguration files sprouting like weeds.\n<\/p>\n<p>\n  Unix programmers aren't going to be cured of their tendency to\ndesign for their peers and the most sophisticated users any time soon\n(we'll grapple a bit with the question of whether such a change would\nactually be desirable in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">Chapter 20<\/a>).  So it's perhaps more useful to invert the\nquestion and ask \u201cWhat things should\n  <em>not<\/em>\n  be\nconfigurable?\u201d Unix practice does offer some guidelines on\nthis.\n<\/p>\n<p>\n  First,\n  <em>\n   don't provide configuration switches for what\nyou can reliably detect automatically\n<\/em>. This is a\nsurprisingly common mistake.  Instead, look for ways to eliminate\nconfiguration switches by autodetection, or by trying alternative\nmethods at runtime until one succeeds. If this strikes you as\ninelegant or too expensive, ask yourself if you haven't fallen into\npremature optimization.\n<\/p>\n <blockquote>\n  <p>\n   One of the nicest examples of autodetection I experienced was\nwhen Dennis Ritchie and I were porting Unix to the Interdata 8\/32.\nThis was a big-endian machine, and we had to generate data for that\nmachine on a PDP-11, write a magnetic tape, and then load the magnetic\ntape on the Interdata.  A common error was to forget to twiddle the\nbyte order; a checksum error showed you that you had to unmount,\nremount again on the PDP-11, regenerate the tape, unmount, and\nremount.  Then one day Dennis hacked the Interdata tape reader program\nso that if it got a checksum error it rewound the tape, toggled\n\u2018byte flip\u2019 switch and reread it.  A second checksum error\nwould kill the load, but 99% of the time it just read the tape and did\nthe right thing.  Our productivity shot up, and we pretty much ignored\ntape byte order from that point on.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  A good rule of thumb is this: Be adaptive unless doing so costs\nyou 0.7 seconds or more of latency.  0.7 seconds is a magic number\nbecause, as Jef Raskin discovered while designing the Canon Cat,\nhumans are almost incapable of noticing startup latency shorter than\nthat; it gets lost in the mental overhead of changing the focus of\nattention.\n<\/p>\n<p>\n  Second,\n  <em>\n   users should not see optimization\nswitches\n<\/em>.  As a designer, it's\n  <em>your<\/em>\n  job to make the program run economically, not the user's.  The\nmarginal gains in performance that a user might collect from\noptimization switches are usually not worth the interface-complexity\ncost.\n<\/p>\n <blockquote>\n  <p>\n   File-format nonsense (record length, blocking factor, etc)\nwas blessedly eschewed by Unix, but the same kind of\nthing has roared back in excess configuration goo.  KISS became\nMICAHI: make it complicated and hide it.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Finally,\n  <em>\n   don't do with a configuration switch what can\nbe done with a script wrapper or a trivial pipeline\n<\/em>.  Don't\nput complexity inside your program when you can easily enlist other\nprograms to help get the work done. (Recall our discussion in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  of why\nls(1)\ndoes not have a built-in pager, or an option to invoke it).\n<\/p>\n<p>\n  Here are some more general questions to consider whenever you\nfind yourself thinking about adding a configuration option:\n<\/p>\n <div><ul><li>\n    <p>\n     Can I leave this feature out?  Why am I\nfattening the manual and burdening the user?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Could the program's normal behavior be changed in an\ninnocuous way that would make the option unnecessary?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Is this option merely cosmetic?  Should I be thinking\nless about how to make the user interface configurable and more \nabout how to make it right?\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Should the behavior enabled by this option be\na separate program instead?\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  Proliferating unnecessary options has many bad effects.  One of\nthe subtlest but most serious is what it will do to your test\ncoverage.\n<\/p>\n <blockquote>\n  <p>\n   Unless it is done very carefully, the addition of an on\/off\nconfiguration option can lead to a need to double the amount of\ntesting.  Since in practice one never does double the amount of\ntesting, the practical effect is to reduce the amount of testing that\nany given configuration receives.  Ten options leads to 1024 times as\nmuch testing, and pretty soon you are talking real reliability\nproblems.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n\n<h3 id=\"ch10s02\"><a href=\"#ch10s02\">\u00a7<\/a>Where Configurations Live<\/h3>\n<p>\n  Classically, a Unix program can look for control information in\nfive places in its startup-time environment:\n<\/p>\n <div><ul><li>\n    <p>\n     Run-control files under\n     <code>\/etc<\/code> (or at fixed location elsewhere in\nsystemland).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     System-set environment variables.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Run-control files (or \u2018dotfiles\u2019) in the\nuser's home directory. (See\n     <a href=\"#contrastchapter\" title=\"Chapter 3. Contrasts\">\n      Chapter 3<\/a>\n     for a discussion of this important\nconcept, if it is unfamiliar.)\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     User-set environment variables.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Switches and arguments passed to the program on the command\nline that invoked it.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  These queries are usually done in the order listed above.  That\nway, later (more local) settings override earlier (more global)\nones. Settings found earlier can help the program compute locations\nfor later retrievals of configuration data.\n<\/p>\n<p>\n  When thinking about which mechanism to use to pass configuration\ndata to a program, bear in mind that good Unix practice demands using\nwhichever one most closely matches the expected lifetime of the\npreference.  Thus: for preferences which are very likely to change\nbetween invocations, use command-line switches.  For preferences which\nchange seldom, but that should be under individual user control, use\na run-control file in the user's home directory.  For preference\ninformation that needs to be set site-wide by a system administrator\nand\n  <em>not<\/em>\n  changed by users, use a run-control file\nin system space.\n<\/p>\n<p>\n  We'll discuss each of these places in more detail, then examine\nsome case studies.\n<\/p>\n\n<h3 id=\"ch10s03\"><a href=\"#ch10s03\">\u00a7<\/a>Run-Control Files<\/h3>\n<p>\n  A run-control file is a file of declarations or commands\nassociated with a program that it interprets on startup. If a program\nhas site-specific configuration shared by all users at a site, it will\noften have a run-control file under the\n  <code>\/etc<\/code> directory. (Some Unixes have an\n  <code>\/etc\/conf<\/code> subdirectory that collects such data.)\n<\/p>\n<p>\n  User-specific configuration information is often carried in a\nhidden run-control file in the user's home directory. Such files are\noften called \u2018dotfiles\u2019 because they exploit the Unix\nconvention that a filename beginning with a dot is normally invisible\nto directory-listing tools.<sup><a href=\"#ftn.id2941866\" id=\"id2941866\">[100]<\/a><\/sup><\/p>\n<p>\n  Programs can also have run-control or dot directories. These\ngroup together several configuration files that are related to the\nprogram, but that are most conveniently treated separately (perhaps\nbecause they relate to different subsystems of the program, or have\ndiffering syntaxes).\n<\/p>\n<p>\n  Whether file or directory, convention now dictates that the\nlocation of the run-control information has the same basename as the\nexecutable that reads it. An older convention still common among\nsystem programs uses the executable's name with the suffix\n\u2018rc\u2019 for \u2018run control\u2019.<sup><a href=\"#ftn.id2941902\" id=\"id2941902\">[101]<\/a><\/sup>\n  Thus, if you write a\nprogram called \u2018seekstuff\u2019 that has both site-wide and\nuser-specific configuration, an experienced Unix user would expect to\nfind the former at\n  <code>\/etc\/seekstuff<\/code> and the latter\nat\n  <code>.seekstuff<\/code> in the user's home directory; but\nit would be unsurprising if the locations were\n  <code>\/etc\/seekstuffrc<\/code> and\n  <code>.seekstuffrc<\/code>, especially if seekstuff were a\nsystem utility of some sort.\n<\/p>\n<p>\n  In\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  we\ndescribed a somewhat different set of design rules for textual data-file \nformats, and discussed how to optimize for different weightings\nof interoperability,\ntransparency\n  and transaction economy.\nRun-control files are typically only read once at program startup and\nnot written; economy is therefore usually not a major concern.\nInteroperability and transparency both push us toward textual formats\ndesigned to be read by human beings and modified with an ordinary text\neditor.\n<\/p>\n<p>\n  While the semantics of run-control files are of course\ncompletely program dependent, there are some design rules about\nrun-control syntax that are widely observed. We'll describe\nthose next; but first we'll describe an important exception.\n<\/p>\n<p>\n  If the program is an interpreter for a language, then it is\nexpected to be simply a file of commands in the syntax of that\nlanguage, to be executed at startup. This is an important rule,\nbecause Unix tradition strongly encourages the design of all kinds of\nprograms as special-purpose languages and minilanguages. Well-known\nexamples with dotfiles of this kind include the various Unix command\nshells and the\n  <em>Emacs<\/em> programmable editor.\n<\/p>\n<p>\n  (One reason for this design rule is the belief that special\ncases are bad news \u2014 thus, that any switch that changes the\nbehavior of a language should be settable from within the language.\nIf as a language designer you find that you\n  <em>cannot<\/em>\n  express all the startup settings of a\nlanguage in the the language itself, a Unix programmer would say you\nhave a design problem \u2014 which is what you should be fixing,\nrather than devising a special-case run-control syntax.)\n<\/p>\n<p>\n  This exception aside, here are the normal style rules for \nrun-control syntaxes. Historically, they are patterned on the syntax of\nUnix shells:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     <em>\n      Support explanatory comments, and lead them with\n      <code>#<\/code>.\n<\/em>\n     The syntax should also ignore whitespace before\n     <code>#<\/code>, so\nthat comments on the same line as configuration directives are\nsupported.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Don't make insidious whitespace\ndistinctions.\n<\/em>\n     That is, treat runs of spaces and tabs,\nsyntactically the same as a single space.  If your directive format is\nline-oriented, it is good form to ignore trailing spaces and tabs on\nlines.  The metarule is that the interpretation of the file should\nnot depend on distinctions a human eye can't\nsee.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Treat multiple blank lines and comment lines\nas a single blank line\n<\/em>.  If the input format uses blank\nlines as separators between records, you probably want to ensure that\na comment line does not end a record.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Lexically treat the file as a simple\nsequence of whitespace-separated tokens, or lines of tokens.\n<\/em>\n     Complicated lexical rules are hard to learn, hard to remember, and\nhard for humans to parse.  Avoid them.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      But, support a string syntax for tokens with\nembedded whitespace.\n<\/em>\n     Use single quote or double quote as\nbalanced delimiters. If you support both, beware of giving them different\nsemantics as they have in shell; this is a well-known source\nof confusion.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      Support a backslash syntax for embedding\nunprintable and special characters in strings\n<\/em>. The standard\npattern for this is the backslash-escape syntax supported by C\ncompilers.  Thus, for example, it would be quite surprising if the\nstring\n     <code>\"a\\tb\"<\/code> were not interpreted as a character\n\u2018a\u2019, followed by a tab, followed by the character\n\u2018b\u2019.\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n<p>\n  Some aspects of shell syntax, on the other hand, should\n  <em>not<\/em>\n  be emulated in run-control syntaxes \u2014\nat least not without a good and specific reason. The shell's baroque\nquoting and bracketing rules, and its special metacharacters for\nwildcards and variable substitution, both fall into this\ncategory.\n<\/p>\n<p>\n  It bears repeating that the point of these conventions is to\nreduce the amount of novelty that users have to cope with when they\nread and edit the run-control file for a program they have never seen\nbefore. Therefore, if you have to break the conventions, try to do so\nin a way that makes it visually obvious that you have done so,\ndocument your syntax with particular care, and (most importantly)\ndesign it so it's easy to pick up by example.\n<\/p>\n<p>\n  These standard style rules only describe conventions about\ntokenizing and comments. The names of run-control files, their\nhigher-level syntax, and the semantic interpretation of the syntax are\nusually application-specific. There are a very few exceptions to\nthis rule, however; one is dotfiles which have become\n\u2018well-known\u2019 in the sense that they routinely carry\ninformation used by a whole class of applications. Sharing\nrun-control file formats in this way reduces the amount of novelty\nusers have to cope with.\n<\/p>\n<p>\n  Of these, probably the best established is the\n  <code>.netrc<\/code> file.  Internet client programs that must\ntrack host\/password pairs for a user can usually get them from\nthe\n  <code>.netrc<\/code> file, if it exists.\n<\/p>\n<h4 id=\"id2942210\"><a href=\"#id2942210\">\u00a7<\/a>Case Study: The .netrc File<\/h4>\n<p>\n  The <code>.netrc<\/code> file is a good example of the standard rules in action.  An example, with the passwords\n  changed to protect the innocent, is in Example 10.1.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 10.1. A\n    <code>.netrc<\/code> example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n# FTP access to my Web host\nmachine unix1.netaxs.com\n        login esr\n        password joesatriani\n# My main mailserver at Netaxs\nmachine imap.netaxs.com\n        login esr\n        password jeffbeck\n# Auxiliary IMAP maildrop at CCIL\nmachine imap.ccil.org\n    login esr\n    password marcbonilla\n# Auxiliary POP maildrop at CCIL\nmachine pop3.ccil.org\n    login esr\n    password ericjohnson\n# Shell account at CCIL\nmachine locke.ccil.org\n    login esr\n    password stevemorse\n<\/pre><\/div>\n<p>\n  Observe that this format is easy to parse by eyeball even if\nyou've never seen it before; it's a set of machine\/login\/password\ntriples, each of which describes an account on a remote host. This\nkind of\ntransparency\n  is important \u2014 much more\nimportant, actually, than the time economy of faster interpretation or\nthe space economy of a more compact and cryptic file format. It\neconomizes the far more valuable resource that is\n  <em>human<\/em>\n  time, by making it likely that a human\nbeing will be able to read and modify the format without having to\nread a manual or use a tool less familiar than a plain old text\neditor.\n<\/p>\n<p>\n  Observe also that this format is used to supply information for\nmultiple services \u2014 an advantage, because it means sensitive\npassword information need only be stored in one place. The\n  <code>.netrc<\/code> format was designed for the original Unix\nFTP client program. It's used by all FTP clients, and also understood\nby some telnet clients, and by the\nsmbclient(1)\ncommand-line tool, and by the\n  <em>fetchmail<\/em> program. If you are writing an Internet client that must do password\nauthentication through remote logins, the Rule of Least Surprise\ndemands that it use the contents of\n  <code>.netrc<\/code> as\ndefaults.\n<\/p>\n<h4 id=\"id2942358\"><a href=\"#id2942358\">\u00a7<\/a>Portability to Other Operating Systems<\/h4>\n<p>\n  Systemwide run-control files are a design tactic that can be\nused on almost any operating system, but dotfiles are rather more\ndifficult to map to a non-Unix environment. The critical thing\nmissing from most non-Unix operating systems is true multiuser\ncapability and the notion of a per-user home directory. DOS and\nWindows versions up to ME (including 95 and 98), for example,\ncompletely lack any such notion; all configuration information has to\nbe stored either in systemwide run-control files at a fixed location,\nthe Windows registry, or configuration files in the same directory a\nprogram is run from.  Windows NT has some notion of per-user home\ndirectories (which made its way into Windows 2000 and XP), but it is\nonly poorly supported by the system tools.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2941866\" id=\"ftn.id2941866\">[100]<\/a><\/sup>\n  To make dotfiles visible,\nuse the\n  <code>-a<\/code> option of\nls(1).\n<\/p><p><sup><a href=\"#id2941902\" id=\"ftn.id2941902\">[101]<\/a><\/sup>\n  The \u2018rc\u2019\nsuffix goes back to Unix's grandparent,\nCTSS\n  .  It had a\ncommand-script feature called \"runcom\".  Early Unixes used\n\u2018rc\u2019 for the name of the operating system's boot script,\nas a tribute to CTSS runcom.\n<\/p><\/div>\n\n<h3 id=\"ch10s04\"><a href=\"#ch10s04\">\u00a7<\/a>Environment Variables<\/h3>\n<p>\n  When a Unix program starts up, the environment accessible to it\nincludes a set of name to value associations (names and values are\nboth strings). Some of these are set manually by the user; others are\nset by the system at login time, or by your shell or terminal\nemulator (if you're running one).  Under Unix, environment variables\ntend to carry information about file search paths, system defaults,\nthe current user ID and process number, and other key bits of\ninformation about the runtime einvironment of programs.  At a shell\nprompt, typing <code>set<\/code> followed by a newline will list\nall currently defined shell variables.\n<\/p>\n<p>\n  In C\n  and\nC++\n  these values\ncan be queried with the library function\ngetenv(3). Perl\n  and Python\n  initialize environment-dictionary\nobjects at startup. Other languages generally follow one of these two\nmodels.\n<\/p>\n<h4 id=\"id2942463\"><a href=\"#id2942463\">\u00a7<\/a>System Environment Variables<\/h4>\n<p>\n  There are a number of well-known environment variables you can\nexpect to find defined on startup of a program from the Unix\nshell.  These (especially\n  <code>HOME<\/code>) will often need to be evaluated\n  <em>before<\/em>\n  you read a local dotfile.\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>USER<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Login name of the account under which this session is logged in\n(BSD convention).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>LOGNAME<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Login name of the\naccount under which this session is logged in (System V\n     convention).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>HOME<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Home directory of the user running this\nsession.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>COLUMNS<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The number of character-cell columns on the controlling\nterminal or terminal-emulator window.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>LINES<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The number of character-cell rows on the controlling terminal\nor terminal-emulator window.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>SHELL<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The name of the user's command shell (often used by shellout\ncommands).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>PATH<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The list of directories that the shell searches when looking\nfor executable commands to match a name.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>TERM<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Name of the terminal type of the session console or\nterminal emulator window (see the terminfo case study in\n     <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n      Chapter 6<\/a>\n     for background).  TERM is special in\nthat programs to create remote sessions over the network (such as\n     <em>telnet<\/em> and\n     <em>ssh<\/em>)\nare expected to pass it through and set it in the remote\nsession.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  (This list is representative, but not exhaustive.)\n<\/p>\n<p>\n  The\n  <code>HOME<\/code> variable is especially important,\nbecause many programs use it to find the calling user's dotfiles\n(others call some functions in the C\n  runtime library to get the calling\nuser's home directory).\n<\/p>\n<p>\n  Note that some or all of these system environment variables may\n  <em>not<\/em>\n  be set when a program is started by some\nother method than a shell spawn. In particular, daemon listeners on a\nTCP\/IP\n  socket often\ndon't have these variables set \u2014 and if they do, the values are\nunlikely to be useful.\n<\/p>\n<p>\n  Finally, note that there is a tradition (exemplified by the\n  <code>PATH<\/code> variable) of using a colon as a separator when an\nenvironment variable must contain multiple fields, especially when the\nfields can be interpreted as a search path of some sort. Note that\nsome shells (notably\n  <em>bash<\/em> and\n  <em>ksh<\/em>)\n  <em>always<\/em>\n  interpret\ncolon-separated fields in an environment variable as filenames, which\nmeans in particular that they expand\n  <code>~<\/code> in these fields to the user's\nhome directory.\n<\/p>\n<h4 id=\"id2942749\"><a href=\"#id2942749\">\u00a7<\/a>User Environment Variables<\/h4>\n<p>\n  Although applications are free to interpret environment\nvariables outside the system-defined set, it is nowadays fairly\nunusual to actually do so. Environment values are not really suitable\nfor passing structured information into a program (though it can in\nprinciple be done via parsing of the values). Instead, modern Unix\napplications tend to use run-control files and dotfiles.\n<\/p>\n<p>\n  There are, however, some design patterns in which user-defined\nenvironment variables can be useful:\n<\/p>\n<p>\n  <em>\n   Application-independent preferences that need to be\nshared by a large number of different programs.\n<\/em>\n  This set of\n\u2018standard\u2019 preferences changes only slowly, because lots\nof different programs need to recognize each one before it becomes\nuseful.<sup><a href=\"#ftn.id2942782\" id=\"id2942782\">[102]<\/a><\/sup>\n  Here are the standard ones:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>EDITOR<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The name of the user's preferred editor (often used by shellout\ncommands).<sup><a href=\"#ftn.id2942810\" id=\"id2942810\">[103]<\/a><\/sup><\/p>\n   <\/dd>\n   <dt>\n    \n     <code>MAILER<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The name of the user's preferred mail user agent (often used by\nshellout commands).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>PAGER<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The name of the\nuser's preferred program for browsing plaintext.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>BROWSER<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The name of the\nuser's preferred program for browsing Web URLs.  This one, as of\n2003, is still very new and not yet widely\nimplemented.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<h4 id=\"id2942882\"><a href=\"#id2942882\">\u00a7<\/a>When to Use Environment Variables<\/h4>\n<p>\n  What both user and system environment variables have in common\nis that it would be annoying to have to replicate the information they\ncontain in a large number of application run-control files, and\nextremely annoying to have to change that information everywhere when\nyour preference changes.  Typically, the user sets these variables in\nhis or her shell session startup file.\n<\/p>\n<p>\n  <em>\n   A value varies across several contexts that share\ndotfiles, or a parent needs to pass information to multiple child\nprocesses.\n<\/em>\n  Some pieces of start-up information are expected\nto vary across several contexts in which the calling user would share\ncommon run-control files and dotfiles. For a system-level example,\nconsider several shell sessions open through terminal emulator windows\non an X desktop. They will all see the same dotfiles, but might have\ndifferent values of\n  <code>COLUMNS<\/code>,\n  <code>LINES<\/code>, and\n  <code>TERM<\/code>. (Old-school shell programming used this method\nextensively; makefiles still do.)\n<\/p>\n<p>\n  <em>\n   A value varies too often for dotfiles, but doesn't\nchange on every startup.\n<\/em>\n  A user-defined environment\nvariable may (for example) be used to pass a file system or Internet\nlocation that is the root of a tree of files that the program should\nplay with. The CVS version-control system interprets the variable\n  <code>CVSROOT<\/code> this way, for example. Several newsreader\nclients that fetch news from servers using the NNTP protocol interpret\nthe variable\n  <code>NNTPSERVER<\/code> as the location of the server\nto query.\n<\/p>\n<p>\n  <em>\n   A process-unique override needs to be expressed in a\nway that doesn't require the command-line invocation to be\nchanged.\n<\/em>\n  A user-defined environment variable can be useful\nfor situations in which, for whatever reason, it would be inconvenient to\nhave to change an application dotfile or supply command-line options\n(perhaps it is expected that the application will normally be used\ninside a shell wrapper or within a makefile).  A particularly\nimportant context for this sort of use is debugging.  Under Linux, for\nexample, manipulating the variable\n  <code>LD_LIBRARY_PATH<\/code> associated with the ld(1) linking loader enables you to change where\nlibraries are loaded from \u2014 perhaps to pick up versions that do\nbuffer-overflow checking or profiling.\n<\/p>\n<p>\n  In general, a user-defined environment variable can be an\neffective design choice when the value changes often enough to make\nediting a dotfile each time inconvenient, but not necessarily every\ntime (so always setting the location with a command-line option\nwould also be inconvenient).  Such variables should typically be\nevaluated\n  <em>after<\/em>\n  a local dotfile and be permitted \nto override settings in it.\n<\/p>\n<p>\n  There is one traditional Unix design pattern that we do not\nrecommend for new programs. Sometimes, user-set environment variables\nare used as a lightweight substitute for expressing a program\npreference in a run-control file.  The venerable\nnethack(1)\ndungeon-crawling game, for example, reads a\n  <code>NETHACKOPTIONS<\/code> environment variable for user\npreferences.  This is an old-school technique; modern practice would\nlean toward parsing them from a\n  <code>.nethack<\/code> or\n  <code>.nethackrc<\/code> run-control file.\n<\/p>\n<p>\n  The problem with the older style is that it makes tracking where\nyour preference information lives more difficult than it would be if\nyou knew the program had a run-control file under your home directory.\nEnvironment variables can be set anywhere in several different shell\nrun-control files \u2014 under Linux these are likely to include\n  <code>.profile<\/code>,\n  <code>.bash_profile<\/code>, and\n  <code>.bashrc<\/code> at least.  These files are cluttered and\nfragile things, so as the code overhead of having an option-parser has\ncome to seem less significant preference information has tended to\nmigrate out of environment variables into dotfiles.\n<\/p>\n<h4 id=\"id2947980\"><a href=\"#id2947980\">\u00a7<\/a>Portability to Other Operating Systems<\/h4>\n<p>\n  Environment variables have only very limited portability off\nUnix. Microsoft operating systems\n  have an environment-variable feature\nmodeled on that of Unix, and use a\n  <code>PATH<\/code> variable as\nUnix does to set the binary search path, but most of other variables\nthat Unix shell programmers take for granted (such as process ID or\ncurrent working directory) are not supported. Other operating systems\n(including classic MacOS) generally do not have any local equivalent\nof environment variables.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2942782\" id=\"ftn.id2942782\">[102]<\/a><\/sup>\n  Nobody knows a really graceful way to represent\nthis sort of distributed preference data; environment variables\nprobably are not it, but all the known alternatives have equally nasty\nproblems.\n<\/p><p><sup><a href=\"#id2942810\" id=\"ftn.id2942810\">[103]<\/a><\/sup>\n  Actually, most Unix programs first check\n  <code>VISUAL<\/code>, and only if that's not set will they consult\n  <code>EDITOR<\/code>. That's a relic from the days when people had\ndifferent preferences for line-oriented editors and visual\neditors.\n<\/p><\/div>\n\n<h3 id=\"ch10s05\"><a href=\"#ch10s05\">\u00a7<\/a>Command-Line Options<\/h3>\n<p>\n  Unix tradition encourages the use of command-line switches to\ncontrol programs, so that options can be specified from scripts. This\nis especially important for programs that function as pipes or\nfilters. Three conventions for how to distinguish command-line options\nfrom ordinary arguments exist; the original Unix style, the GNU style,\nand the X toolkit style.\n<\/p>\n<p>\n  In the original Unix tradition, command-line options are single\nletters preceded by a single hyphen. Mode-flag options that do not\ntake following arguments can be ganged together; thus, if\n  <code>-a<\/code> and\n  <code>-b<\/code> are mode options,\n  <code>-ab<\/code> or\n  <code>-ba<\/code> is also correct and\nenables both. The argument to an option, if any, follows it\n(optionally separated by whitespace). In this style, lowercase\noptions are preferred to uppercase. When you use uppercase options,\nit's good form for them to be special variants of the lowercase\noption.\n<\/p>\n<p>\n  The original Unix style evolved on slow ASR-33 teletypes that made\nterseness a virtue; thus the single-letter options. Holding down the\nshift key required actual effort; thus the preference for lower case,\nand the use of \u201c-\u201d (rather than the perhaps more\nlogical \u201c+\u201d) to enable options.\n<\/p>\n<p>\n  The GNU style uses option keywords (rather than keyword letters)\npreceded by two hyphens. It evolved years later when some of the\nrather elaborate GNU utilities began to run out of single-letter\noption keys (this constituted a patch for the symptom, not a cure for\nthe underlying disease). It remains popular because GNU options are\neasier to read than the alphabet soup of older styles.  GNU-style\noptions cannot be ganged together without separating whitespace. An\noption argument (if any) can be separated by either whitespace or a\nsingle \u201c=\u201d (equal sign) character.\n<\/p>\n<p>\n  The GNU double-hyphen option leader was chosen so that traditional\nsingle-letter options and GNU-style keyword options could be\nunambiguously mixed on the same command line. Thus, if your initial\ndesign has few and simple options, you can use the Unix style without\nworrying about causing an incompatible \u2018flag day\u2019 if you\nneed to switch to GNU style later on. On the other hand, if you are\nusing the GNU style, it is good practice to support single-letter\nequivalents for at least the most common options.\n<\/p>\n<p>\n  The X toolkit style, confusingly, uses a single hyphen and\nkeyword options. It is interpreted by X toolkits that filter out and\nprocess certain options (such as\n  <code>-geometry<\/code> and\n  <code>-display<\/code>) before handing the filtered command line to\nthe application logic for interpretation. The X toolkit style is not\nproperly compatible with either the classic Unix or GNU styles, and\nshould not be used in new programs unless the value of being\ncompatible with older X conventions seems very high.\n<\/p>\n<p>\n  Many tools accept a bare hyphen, not associated with any option\nletter, as a pseudo-filename directing the application to read from\nstandard input.  It is also conventional to recognize a double hyphen\nas a signal to stop option interpretation and treat all following\narguments literally.\n<\/p>\n<p>\n  Most Unix programming languages offer libraries that will parse\na command line for you in either classic-Unix or GNU style\n(interpreting the double-hyphen convention as well).\n<\/p>\n<h4 id=\"id2948149\"><a href=\"#id2948149\">\u00a7<\/a>The -a to -z of Command-Line Options<\/h4>\n<p>\n  Over time, frequently-used options in well-known Unix programs\nhave established a loose sort of semantic standard for what various flags\nmight be expected to mean. The following is a list of options and\nmeanings that should prove usefully unsurprising to an experienced\nUnix user:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>-a<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     All (without argument). If there is a GNU-style\n     <code>--all<\/code> option, for\n     <code>-a<\/code> to be anything but a synonym for it\nwould be quite surprising. Examples:\nfuser(1),\nfetchmail(1).\n    <\/p>\n    <p>\n     Append, as in\ntar(1).\nThis is often paired with\n     <code>-d<\/code> for delete.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-b<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Buffer or block size (with argument). Set a critical buffer\nsize, or (in a program having to do with archiving or managing storage\nmedia) set a block size. Examples:\ndu(1),\ndf(1),\ntar(1).\n    <\/p>\n    <p>\n     Batch. If the program is naturally interactive,\n     <code>-b<\/code> may be used to suppress prompts or set other\noptions appropriate to accepting input from a file rather than a human\noperator. Example:\nflex(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-c<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Command (with argument). If the program is an interpreter that\nnormally takes commands from standard input, it is expected that the\noption of a\n     <code>-c<\/code> argument will be passed to it as a\nsingle line of input. This convention is particularly strong for\nshells and shell-like interpreters. Examples:\nsh(1),\nash(1),\nbsh(1),\nksh(1),\npython(1).\nCompare\n     <code>-e<\/code> below.\n    <\/p>\n    <p>\n     Check (without argument). Check the correctness of the file\nargument(s) to the command, but don't actually perform normal\nprocessing. Frequently used as a syntax-check option by programs that\ndo interpretation of command files. Examples:\ngetty(1),\nperl(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-d<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Debug (with or without argument). Set the level of debugging\nmessages. This one is very common.\n    <\/p>\n    <p>\n     Occasionally\n     <code>-d<\/code> has the sense of\n\u2018delete\u2019 or \u2018directory\u2019.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-D<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Define (with argument). Set the value of some symbol in an\ninterpreter, compiler, or (especially) macro-processor-like\napplication. The model is the use of\n     <code>-D<\/code> by the C\n     compiler's macro preprocessor.  This is a strong association \nfor most Unix programmers; don't try to fight it.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-e<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Execute (with argument). Programs that are wrappers, or that can\nbe used as wrappers, often allow\n     <code>-e<\/code> to set the program they hand off\ncontrol to. Examples:\nxterm(1),\nperl(1).\n    <\/p>\n    <p>\n     Edit. A program that can open a resource in either a read-only\nor editable mode may allow\n     <code>-e<\/code> to specify opening in the editable\nmode. Examples:\ncrontab(1),\nand the\nget(1)\nutility of the SCCS version-control system.\n    <\/p>\n    <p>\n     Occasionally\n     <code>-e<\/code> has the sense of \u2018exclude\u2019 or\n\u2018expression\u2019.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-f<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     File (with argument). Very often used with an argument to\nspecify an input (or, less frequently, output) file for programs that\nneed to randomly access their input or output (so that redirection via\n&lt; or &gt; won't suffice). The classic example is\ntar(1);\nothers abound.  It is also used to indicate that arguments normally\ntaken from the command line should be taken from a file instead;  see\nawk(1)\nand egrep(1)\nfor classic examples.\nCompare\n     <code>-o<\/code> below; often,\n     <code>-f<\/code> is the input-side analog of\n     <code>-o<\/code>.\n    <\/p>\n    <p>\n     Force (typically without argument). Force some operation (such\nas a file lock or unlock) that is normally performed\nconditionally. This is less common.\n    <\/p>\n    <p>\n     Daemons often use\n     <code>-f<\/code> in a way that combines\nthese two meanings, to force processing of a configuration file from a\nnondefault location. Examples:\nssh(1),\nhttpd(1),\nand many other daemons.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-h<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Headers (typically without argument). Enable, suppress, or\nmodify headers on a tabular report generated by the program.\nExamples:\npr(1),\nps(1).\n    <\/p>\n    <p>\n     Help. This is actually less common than one might expect offhand\n\u2014 for much of Unix's early history developers tended to think of\non-line help as memory-footprint overhead they couldn't\nafford. Instead they wrote manual pages (this shaped the man-page\nstyle in ways we'll discuss in\n     <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n      Chapter 18<\/a>).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-i<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Initialize (usually without argument). Set some critical\nresource or database associated with the program to an initial or\nempty state.  Example:\nci(1)\nin RCS.\n    <\/p>\n    <p>\n     Interactive (usually without argument). Force a program that\ndoes not normally query for confirmation to do so. There are classical\nexamples\n(rm(1),\nmv(1))\nbut this use is not common.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-I<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Include (with argument). Add a file or directory name to those\nsearched for resources by the application. All Unix compilers with any\nequivalent of source-file inclusion in their languages use\n     <code>-I<\/code> in this sense. It would be extremely surprising to\nsee this option letter used in any other way.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-k<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Keep (without argument). Suppress the normal deletion of\nsome file, message, or resource. Examples:\npasswd(1),\nbzip(1),\nand\nfetchmail(1).\n    <\/p>\n    <p>\n     Occasionally\n     <code>-k<\/code> has the sense of\n\u2018kill\u2019.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-l<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     List (without argument). If the program is an archiver or\ninterpreter\/player for some kind of directory or archive format, it\nwould be quite surprising for\n     <code>-l<\/code> to do anything but request an item\nlisting.  Examples:\narc(1),\nbinhex(1),\nunzip(1). (However,\ntar(1)\nand\ncpio(1)\nare exceptions.)\n    <\/p>\n    <p>\n     In programs that are already report generators,\n     <code>-l<\/code> almost invariably means \u201clong\u201d and\ntriggers some kind of long-format display revealing more detail than\nthe default mode.  Examples:\nls(1),\nps(1).\n    <\/p>\n    <p>\n     Load (with argument). If the program is a linker or a language\ninterpreter,\n     <code>-l<\/code> invariably loads a library, in some\nappropriate sense. Examples:\ngcc(1),\nf77(1),\nemacs(1).\n    <\/p>\n    <p>\n     Login.  In programs such as\nrlogin(1)\nand\nssh(1)\nthat need to specify a network identity,\n     <code>-l<\/code> is how you do it.\n    <\/p>\n    <p>\n     Occasionally\n     <code>-l<\/code> has the sense of \u2018length\u2019 or\n\u2018lock\u2019.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-m<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Message (with argument). Used with an argument,\n     <code>-m<\/code> passes it\nin as a message string for some logging or announcement purpose.\nExamples:\nci(1),\ncvs(1).\n    <\/p>\n    <p>\n     Occasionally\n     <code>-m<\/code> has the sense of \u2018mail\u2019,\n\u2018mode\u2019, or \u2018modification-time\u2019.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-n<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Number (with argument). Used, for example, for page number\nranges in programs such as\nhead(1),\ntail(1),\nnroff(1),\nand\ntroff(1). Some\nnetworking tools that normally display DNS names accept\n     <code>-n<\/code> as an option that causes them to display the raw\nIP addresses instead;\nifconfig(1)\nand\ntcpdump(1)\nare the archetypal examples.\n    <\/p>\n    <p>\n     Not (without argument). Used to suppress normal actions in\nprograms such as make(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-o<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Output (with argument). When a program needs to specify an\noutput file or device by name on the command line, the\n     <code>-o<\/code> option does it. Examples:\nas(1),\ncc(1),\nsort(1). On\nanything with a compiler-like interface, it would be extremely\nsurprising to see this option used in any other way.  Programs that\nsupport\n     <code>-o<\/code> often (like gcc) have logic that allows it\nto be recognized after ordinary arguments as well as before.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-p<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Port (with argument). Especially used for options that specify\nTCP\/IP\n     port numbers.\nExamples:\ncvs(1),\nthe PostgreSQL tools, the\nsmbclient(1),\nsnmpd(1),\nssh(1).\n    <\/p>\n    <p>\n     Protocol (with argument). Examples:\nfetchmail(1),\nsnmpnetstat(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-q<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Quiet (usually without argument). Suppress normal result or\ndiagnostic output. This is very common.  Examples: ci(1), co(1), \nmake(1).  See also the \u2018silent\u2019 sense of\n     <code>-s<\/code>.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-r<\/code> (also\n     <code>-R<\/code>)\n    \n   <\/dt>\n   <dd>\n    <p>\n     Recurse (without argument). If the program operates on a\ndirectory, then this option might tell it to recurse on all\nsubdirectories. Any other use in a utility that operated on\ndirectories would be quite surprising. The classic example is, of\ncourse, cp(1).\n    <\/p>\n    <p>\n     Reverse (without argument). Examples:\nls(1),\nsort(1). A\nfilter might use this to reverse its normal translation action\n(compare\n     <code>-d<\/code>).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-s<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Silent (without argument). Suppress normal diagnostic or result\noutput (similar to\n     <code>-q<\/code>; when both are supported, q\nmeans \u2018quiet\u2019 but\n     <code>-s<\/code> means \u2018utterly\nsilent\u2019). Examples:\ncsplit(1),\nex(1),\nfetchmail(1).\n    <\/p>\n    <p>\n     Subject (with argument).\n     <em>Always<\/em>\n     used with\nthis meaning on commands that send or manipulate mail or news\nmessages.  It is extremely important to support this, as programs\nthat send mail expect it. Examples: mail(1), elm(1), mutt(1).\n    <\/p>\n    <p>\n     Occasionally\n     <code>-s<\/code> has the sense of\n\u2018size\u2019.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-t<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Tag (with argument). Name a location or give a string for a\nprogram to use as a retrieval key. Especially used with text editors\nand viewers. Examples:\ncvs(1),\nex(1),\nless(1),\nvi(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-u<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     User (with argument). Specify a user, by name or numeric UID.\nExamples:\ncrontab(1),\nemacs(1),\nfetchmail(1),\nfuser(1),\nps(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-v<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Verbose (with or without argument). Used to enable\ntransaction-monitoring, more voluminous listings, or debugging\noutput. Examples:\ncat(1),\ncp(1),\nflex(1),\ntar(1),\nmany others.\n    <\/p>\n    <p>\n     Version (without argument). Display program's version on\nstandard output and exit. Examples:\ncvs(1),\nchattr(1),\npatch(1),\nuucp(1). More\nusually this action is invoked by\n     <code>-V<\/code>.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-V<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Version (without argument). Display program's version on\nstandard output and exit (often also prints compiled-in configuration\ndetails as well). Examples: gcc(1), flex(1), hostname(1),\nmany others. It would be quite surprising for this switch to be\nused in any other way.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-w<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Width (with argument). Especially used for specifying widths in\noutput formats. Examples:\nfaces(1),\ngrops(1),\nod(1),\npr(1),\nshar(1).\n    <\/p>\n    <p>\n     Warning (without argument). Enable warning diagnostics, or\nsuppress them. Examples:\nfetchmail(1),\nflex(1),\nnsgmls(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-x<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Enable debugging (with or without argument). Like\n     <code>-d<\/code>. Examples:\nsh(1),\nuucp(1).\n    <\/p>\n    <p>\n     Extract (with argument). List files to be extracted from an\narchive or working set. Examples:\ntar(1),\nzip(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-y<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Yes (without argument). Authorize potentially destructive\nactions for which the program would normally require confirmation.\nExamples:\nfsck(1),\nrz(1).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>-z<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Enable compression (without argument). Archiving and backup\nprograms often use this. Examples:\nbzip(1),\nGNU\ntar(1),\nzcat(1),\nzip(1),\ncvs(1).\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  The preceding examples are taken from the\nLinux\n  toolset, but should be good on most modern Unixes.\n<\/p>\n<p>\n  When you're choosing command-line option letters for your\nprogram, look at the manual pages for similar tools.  Try to use the\nsame option letters they use for the analogous functions of your\nprogram.  Note that some particular application areas that have\nparticularly strong conventions about command-line switches which you\nviolate at your peril \u2014 compilers, mailers, text filters,\nnetwork utilities and X software are all notable for this.  Anybody\nwho wrote a mail agent that used\n  <code>-s<\/code> as anything but a\nSubject switch, for example, would have scorn rightly heaped upon the\nchoice.\n<\/p>\n<p>\n  The GNU project\n  recommends conventional meanings for\na few double-dash options in the GNU coding\nstandards.<sup><a href=\"#ftn.id2950141\" id=\"id2950141\">[104]<\/a><\/sup>\n  It also lists long options\nwhich, though not standardized, are used in many GNU programs.\nIf you are using GNU-style options, and some option you need has\na function similar to one of those listed, by all means \nobey the Rule of Least Surprise and reuse the name.\n<\/p>\n<h4 id=\"id2950162\"><a href=\"#id2950162\">\u00a7<\/a>Portability to Other Operating Systems<\/h4>\n<p>\n  To have command-line options, you have to have a command line.\nThe MS-DOS family does, of course, though in Windows it's hidden by a\nGUI and its use is discouraged; the fact that the option character is\nnormally \u2018\n  <code>\/<\/code> \u2019 rather than \u2018\n  <code>-<\/code> \u2019 is merely a\ndetail.  MacOS classic and other pure GUI environments have no close\nequivalent of command-line options.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2950141\" id=\"ftn.id2950141\">[104]<\/a><\/sup>\n  See the\n  <a href=\"http:\/\/www.gnu.org\/prep\/standards.html\">\n   Gnu Coding\nStandards<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch10s06\"><a href=\"#ch10s06\">\u00a7<\/a>How to Choose among the Methods<\/h3>\n<p>\n  We've looked in turn at system and user run-control files, at\nenvironment variables, and at command-line arguments. Observe the\nprogression from least easily changed to most easily changed. There is\na strong convention that well-behaved Unix programs that use more than\none of these places should look at them in the order given, allowing\nlater settings to override earlier ones (there are specific\nexceptions, such as command-line options that specify where a dotfile\nshould be found).\n<\/p>\n<p>\n  In particular, environment settings usually override dotfile\nsettings, but can be overridden by command-line options. It is good\npractice to provide a command-line option like the\n  <code>-e<\/code> of\nmake(1) \nthat can override environment settings or declarations in run-control files;\nthat way the program can be scripted with well-defined behavior\nregardless of the way the run-control files look or environment\nvariables are set.\n<\/p>\n<p>\n  Which of these places you choose to look at depends on how much\npersistent configuration state your program needs to keep around\nbetween invocations. Programs designed mainly to be used in a batch\nmode (as generators or filters in pipelines, for example) are usually\ncompletely configured with command-line options. Good examples of this\npattern include\nls(1),\ngrep(1)\nand\nsort(1). At\nthe other extreme, large programs with complicated interactive\nbehavior may rely entirely on run-control files and environment\nvariables, and normal use involves few command-line options or none at\nall. Most X window managers are a good example of this pattern.\n<\/p>\n<p>\n  (Unix has the capability for the same file to have multiple\nnames or \u2018links\u2019.  At startup time, every program has\navailable to it the filename through which it was called.\nOne other way to signal to a program that has several modes of\noperation which one it should come up in is to give it a link for each\nmode, have it find out which link it was called through, and change \nits behavior accordingly.  But this technique is generally considered\nunclean and seldom used.)\n<\/p>\n<p>\n  Let's look at a couple of programs that gather configuration\ndata from all three places. It will be instructive to consider why,\nfor each given piece of configuration data, it is collected as it\nis.\n<\/p>\n<h4 id=\"fetchmail_environment\"><a href=\"#fetchmail_environment\">\u00a7<\/a>\n  Case Study:\nfetchmail\n <\/h4>\n<p>\n  The\n  <em>fetchmail<\/em> program uses only two\nenvironment variables,\n  <code>USER<\/code> and\n  <code>HOME<\/code>. These variables are in the predefined set\ninitialized by the system; many programs use them.\n<\/p>\n<p>\n  The value of\n  <code>HOME<\/code> is used to find the dotfile\n  <code>.fetchmailrc<\/code>, which contains configuration\ninformation in a fairly elaborate syntax obeying the shell-like\nlexical rules described above. This is appropriate because, once it\nhas been initially set up, Fetchmail's\nconfiguration will change only infrequently.\n<\/p>\n<p>\n  There is neither an\n  <code>\/etc\/fetchmailrc<\/code> nor\nany other systemwide file specific to fetchmail. Normally such files\nhold configuration that's not specific to an individual user.\nfetchmail does use a small set of properties with this kind of scope\n\u2014 specifically, the name of the local postmaster, and a few\nswitches and values describing the local mail transport setup (such as\nthe port number of the local SMTP listener). In practice, however,\nthese are seldom changed from their compiled-in default values.  When\nthey are changed, they tend to be modified in user-specific\nways. Thus, there has been no demand for a systemwide fetchmail \nrun-control file.\n<\/p>\n<p>\n  Fetchmail can retrieve host\/login\/password triples from a\n  <code>.netrc<\/code> file. Thus, it gets authenticator\ninformation in the least surprising way.\n<\/p>\n<p>\n  Fetchmail has an elaborate set of command-line options, which\nnearly but do not entirely replicate what the\n  <code>.fetchmailrc<\/code> can express. The set was not\noriginally large, but grew over time as new constructs were added to\nthe\n  <code>.fetchmailrc<\/code> minilanguage and parallel\ncommand-line options for them were added more or less\nreflexively.\n<\/p>\n<p>\n  The intent of supporting all these options was to make\nfetchmail easier to script by allowing\nusers to override bits of its run control from the command line. But\nit turns out that outside of a few options like\n  <code>--fetchall<\/code> and\n  <code>--verbose<\/code> there is\nlittle demand for this \u2014 and none that can't be satisfied with a\nshellscript that creates a temporary run-control file on the fly and\nthen feeds it to fetchmail using the\n  <code>-f<\/code> option.\n<\/p>\n<p>\n  Thus, most of the command-line options are never used, and in\nretrospect including them was probably a mistake; they bulk up the\nfetchmail code a bit without accomplishing\nanything very useful.\n<\/p>\n <blockquote>\n  <p>\n   If bulking up the code were the only problem, nobody would care,\nexcept for a couple of maintainers.  However, options increase the\nchances of error in code, particularly due to unforeseen interactions\namong rarely used options.  Worse, they bulk up the manual, which is a\nburden on everybody.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  There is a lesson here; had I thought carefully enough about\nfetchmail's usage pattern and been a little less ad-hoc about adding\nfeatures, the extra complexity might have been avoided.\n<\/p>\n <blockquote>\n  <p>\n   An alternative way of dealing with such situations, which\ndoesn't clutter up either the code or the manual much, is to have a\n\u201cset option variable\u201d option, such as the\n   <code>-O<\/code> option of\n   <em>sendmail<\/em>,\nwhich lets you specify an option name and value, and sets that name to\nthat value as if such a setting had been given in a configuration\nfile.  A more powerful variant of this is what\n   <em>ssh<\/em> does with its\n   <code>-o<\/code> option: the argument to\n   <code>-o<\/code> is treated as if it were a\nline appended to the configuration file, with the full config-file\nsyntax available.  Either of these approaches gives people with\nunusual requirements a way to override configuration from the command\nline, without requiring you to provide a separate option for each bit\nof configuration that might be overridden.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<h4 id=\"id2950578\"><a href=\"#id2950578\">\u00a7<\/a>Case Study: The XFree86 Server<\/h4>\n<p>\n  The X windowing system is the engine that supports bitmapped \ndisplays on Unix machines. Unix applications running through a client machine\nwith a bitmapped display get their input events through X and \nsend screen-painting requests to it.  Confusingly, X \u2018servers\u2019\nactually run on the client machine \u2014 they exist to serve requests to\ninteract with the client machine's display device.  The applications\nsending those requests to the X server are called \u2018X\nclients\u2019, even though they may be running on a server machine.\nAnd no, there is no way to explain this inverted terminology that is\nnot confusing.\n<\/p>\n<p>\n  X servers have a forbiddingly complex interface to their\nenvironment. This is not surprising, as they have to deal with a wide\nrange of complex hardware and user preferences. The environment\nqueries common to all X servers, documented on the\nX(1)\nand\nXserver(1)\npages, therefore make a useful example for study.  The implementation\nwe examine here is XFree86, the X implementation used under\nLinux\n  and several other open-source Unixes.\n<\/p>\n<p>\n  At startup, the XFree86 server examines a systemwide run-control file; the exact pathname varies between X builds on\n  different platforms, but the basename is XF86Config. The XF86Config file has a shell-like syntax as described above.\n  Example 10.2 is a sample section of an XF86Config file.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 10.2. X configuration example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n# The 16-color VGA server\nSection \"Screen\"\n    Driver      \"vga16\"\n    Device      \"Generic VGA\"\n    Monitor     \"LCD Panel 1024x768\"\n    Subsection  \"Display\"\n        Modes       \"640x480\" \"800x600\"\n        ViewPort    0 0\n    EndSubsection\nEndSection\n<\/pre><\/div>\n<p>\n  The XF86Config file describes the host machine's display\nhardware (graphics card, monitor), keyboard, and pointing device\n(mouse\/trackball\/glidepad). It's appropriate for this information to\nlive in a systemwide run-control file, because it applies to all\nusers of the machine.\n<\/p>\n<p>\n  Once X has acquired its hardware configuration from the run\ncontrol file, it uses the value of the environment variable\n  <code>HOME<\/code> to find two dotfiles in the calling user's home\ndirectory. These files are\n  <code>.Xdefaults<\/code> and\n  <code>.xinitrc<\/code>.<sup><a href=\"#ftn.id2950710\" id=\"id2950710\">[105]<\/a><\/sup><\/p>\n<p>\n  The\n  <code>.Xdefaults<\/code> file specifies per-user,\napplication-specific resources relevant to X (trivial examples of\nthese might include font and foreground\/background colors for a\nterminal emulator). The phrase \u2018relevant to X\u2019 indicates a\ndesign problem, however.  Collecting all these resource declarations\nin one place is convenient for inspecting and editing them, but it is\nnot always clear what should be declared in\n  <code>.Xdefaults<\/code> and what belongs in an\napplication-specific dotfile. The\n  <code>.xinitrc<\/code> file\nspecifies the commands that should be run to initialize the user's X\ndesktop just after server startup. These programs will almost always\ninclude a window or session manager.\n<\/p>\n<p>\n  X servers have a large set of command-line options. Some of\nthese, such as the\n  <code>-fp<\/code> (font path) option, override\nthe XF86Config.  Some are intended to help track server bugs, such as\nthe\n  <code>-audit<\/code> option; if these are used at all, they are\nlikely to vary quite frequently between test runs and are therefore\npoor candidates to be included in a run-control file. A very important\noption is the one that sets the server's display number. Multiple\nservers may run on a host provided each has a unique display number,\nbut all instances share the same run-control file(s); thus, the\ndisplay number cannot be derived solely from those files.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2950710\" id=\"ftn.id2950710\">[105]<\/a><\/sup>\n  The\n  <code>.xinitrc<\/code> is analogous to a Startup folder on\nWindows and other operating systems.\n<\/p><\/div>\n\n<h3 id=\"ch10s07\"><a href=\"#ch10s07\">\u00a7<\/a>On Breaking These Rules<\/h3>\n<p>\n  The conventions described in this chapter are not absolute, but\nviolating them will increase friction costs for users and\ndevelopers in the future. Break them if you must \u2014 but be sure you\nknow exactly why you are doing so before you do it. And if you do\nbreak them, make sure that attempts to do things in conventional ways\nbreak noisily, giving proper error feedback in accordance with the\nRule of Repair.\n<\/p>\n\n<h3 id=\"interfacechapter\"><a href=\"#interfacechapter\">\u00a7<\/a>Chapter 11. Interfaces<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   All our knowledge has its origins in our perceptions.\n  <\/p><div>\n  \n   \u2013 Leonardo Da Vinci\n  \n <\/div><\/blockquote>\n<p>\n  The interface of a program is the sum of all the ways that it\ncommunicates with human users and other programs.  In\n  <a href=\"#configurationchapter\" title=\"Chapter 10. Configuration\">\n   Chapter 10<\/a>, we discussed the use of environment\nvariables, switches, run-control files and other parts of\nstart-up-time interfaces. In this chapter, we'll untangle the history\nand explain the pragmatics of Unix interfaces after startup time.\nBecause user-interface code normally consumes 40% or more of development\ntime, knowing good design patterns is especially important here in\norder to avoid a lot of false starts and time-intensive\nrewrites.\n<\/p>\n<p>\n  In the Unix tradition of interface design, we encounter two\nthemes over and over again. One is anticipatory design for\ncommunication with other programs; the other is the Rule of Least\nSurprise.\n<\/p>\n<p>\n  Unix programs can give you extra power from being used in\nsynergistic combinations; we discussed various methods for hooking\ntogether such combinations in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>. The \u2018other programs\u2019 part\nof Unix interface design is not an afterthought or a marginal case as\nit is under many other operating systems. Rather, it is a central\nchallenge that has to be balanced and integrated carefully with the\ndemands of interface design for human users.\n<\/p>\n<p>\n  Much of Unix-community tradition about program interface design\nmay seem odd and arbitrary \u2014 or even, in the age of the GUI,\ndownright regressive \u2014 when you encounter that tradition for the\nfirst time.  But in spite of various blemishes and irregularities,\nthat tradition has an inner logic to it which is worth learning and\nunderstanding. It reflects heuristics accumulated over Unix's long\nhistory about ways to do effective communication both with human\nbeings and with other programs. And it includes a set of conventions\nwhich create commonalities between programs \u2014 it defines\n\u2018least surprising\u2019 alternatives for a wide range of common\ninterface-design problems.\n<\/p>\n<p>\n  After startup, programs normally get input or commands from the\nfollowing sources:\n<\/p>\n <div><ul><li>\n    <p>\n     Data and commands presented on the program's standard\ninput.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Inputs passed through IPC, such as X server events and\nnetwork messages.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Files and devices in known locations (such as a data file name\npassed to or computed by the program).\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  Programs can emit results in all the same ways (with output\ngoing to standard output).\n<\/p>\n<p>\n  Some Unix programs are graphical, some have screen-oriented\ncharacter interfaces, and some use a starkly simple text-filter\ndesign unchanged from the days of mechanical teletypes. To the\nuninitiated, it is often far from obvious why any given program\nuses the style it does \u2014 or, indeed, why Unix supports such a\nplethora of interface styles at all.\n<\/p>\n<p>\n  Unix has several competing interface styles.  All are still\nalive for a reason; they're optimized for different situations. By\nunderstanding the fit between task and interface style, you will learn\nhow to choose the right styles for the jobs you need to do.\n<\/p>\n\n<h3 id=\"ch11s01\"><a href=\"#ch11s01\">\u00a7<\/a>Applying the Rule of Least Surprise<\/h3>\n<p>\n  The Rule of Least Surprise is a general principle in the design\nof all kinds of interfaces, not just software: \u201cDo the least\nsurprising thing\u201d. It's a consequence of the fact that human\nbeings can only pay attention to one thing at one time (see\n  <em>The Humane Interface<\/em> <sup><a href=\"#Raskin\" title=\"[Raskin]\">[Raskin]<\/a><\/sup>).  Surprises in the interface focus\nthat single locus of attention on the interface, rather than on the\ntask where it belongs.\n<\/p>\n<p>\n  Thus, to design usable interfaces, it's best when possible not\nto design an entire new interface model. Novelty is a barrier to\nentry; it puts a learning burden on the user, so minimize it. Instead,\nthink carefully about the experience and knowledge of your user\nbase. Try to find functional similarities between your program and\nprograms they are likely to already know about. Then mimic the\nrelevant parts of the existing interfaces.\n<\/p>\n<p>\n  The Rule of Least Surprise should not be interpreted as a call\nfor mechanical conservatism in design. Novelty raises the cost of a\nuser's first few interactions with an interface, but poor design\nwill make the interface needlessly painful forever. As in other\nsorts of design, rules are not a substitute for good taste and\nengineering judgment. Consider your tradeoffs carefully \u2014 and\nconsider them from the\n  <em>user's<\/em>\n  point of view. The bias\nimplied by the Rule of Least Surprise is a good one to hold\nconsciously, mainly because interface designers (like other\nprogrammers) have an unconscious tendency to be too clever for the\nuser's good.\n<\/p>\n<p>\n  One implication of the Rule of Least Surprise is this: Wherever\npossible, allow the user to delegate interface functions to a familiar\nprogram. We already observed in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  that, if your program requires the\nuser to edit significant amounts of text, you should write it to call\nan editor (specifiable by the user) rather than building in your own\nintegrated editor.  This will enable the\n  <em>users<\/em>,\nwho know their preferences better than you, to choose the least\nsurprising alternative.\n<\/p>\n<p>\n  Elsewhere in this book we have advocated symbiosis and\ndelegation as tactics for promoting code reuse and minimizing\ncomplexity. The point here is that when users can intercept the\ndelegation, and direct it to an agent of their own choice,\nthese techniques become not merely economical for the developer but\nactively empowering to users.\n<\/p>\n<p>\n  Further: When you can't delegate, emulate. The purpose of the\nRule of Least Surprise is to reduce the amount of complexity a user\nmust absorb to use an interface. Continuing the editor example, this\nmeans that if you must implement an embedded editor, it's best if the\neditor commands are a subset of those for a well-known general-purpose\neditor. (Or more than one.  Both\n  <em>bash<\/em> and\n  <em>ksh<\/em> have command-line editors that allow\nthe user to choose between\n  <em>vi<\/em> and\n  <em>Emacs<\/em> editing styles.)\n<\/p>\n<p>\n  Under the Unix versions of the Netscape and Mozilla Web\nbrowsers, for example, fill-in fields in forms recognize a subset of\nthe default bindings for the\n  <em>Emacs<\/em> editor. Control-A goes to start of line, Control-D deletes the next\ncharacter, and so forth. This choice helps people who know\n  <em>Emacs<\/em>, and leaves others no worse off than\nan arbitrary, idiosyncratic command set would have. The only way it\ncould have been bettered was by choosing key bindings associated with\nsome editor significantly more widely used than\n  <em>Emacs<\/em>; and among Netscape's original user\npopulation there was no such animal.\n<\/p>\n<p>\n  These principles can be applied in many other areas of interface\ndesign. They suggest, for example, that it is deeply foolish to create\nnovel document formats for an on-line help system when users are\ncomfortable with an HTML Web browser. Or even that if you are\ndesigning an arcade-style game, it is wise to look at the gesture sets\nof previous games to see if you can give new users a feeling of\ncomfort by allowing them to transfer joystick skills learned in other\ngames.\n<\/p>\n\n<h3 id=\"ch11s02\"><a href=\"#ch11s02\">\u00a7<\/a>History of Interface Design on Unix<\/h3>\n<p>\n  Unix predates the modern graphics-intensive style of software\ninterface design. For over a decade after the first Unix in 1969,\ncommand-line interfaces (CLIs) on teletypes and dumb text-mode\nterminals were the norm. Most of the basic Unix toolset (programs like\nls(1),\ncat(1),\nand\ngrep(1))\nstill reflect this heritage.\n<\/p>\n<p>\n  Gradually, after 1980, Unix evolved support for screen-painting\non character-cell terminals. Programs began to mix command-line and\nvisual interfaces, with common commands often bound to keystrokes that\nwould not be echoed to the screen. Some of the early programs written\nin this style (often called \u2018curses\u2019 programs, after the\nscreen-painting cursor-control library normally used to implement\nthem, or \u2018roguelike\u2019 after the first application to use\ncurses) are still used today; notable examples include the\ndungeon-crawling game\nrogue(1),\nthe\nvi(1)\n  text editor, and (from a few years later) the\nelm(1)\nmailer and its modern descendant\nmutt(1).\n<\/p>\n<p>\n  A few years later in the mid-1980s, the computing world as a\nwhole began to assimilate the results of the pioneering work on\ngraphical user interfaces (GUIs) that had been going on at Xerox's\nPalo Alto Research Center since the early 1970s. On personal\ncomputers, the Xerox PARC\n  work inspired the Apple Macintosh interface\nand through that the design of Microsoft\nWindows. Unix's adaptation of these ideas took a\nrather more complicated path.\n<\/p>\n<p>\n  Around 1987 the X windowing system\n  outcompeted several early contenders and\nprototype efforts to become the standard graphical-interface facility\nfor Unix. Whether this was a good or a bad thing has remained a topic\nof debate ever since; some of the other contenders (notably Sun's\nNetwork Window System or NeWS)\n  were arguably rather more powerful\nand elegant. X, however, had one overriding virtue; it was open\nsource. The code had been developed at MIT by a research group more\ninterested in exploring the problem space than in creating a product,\nand it remained freely redistributable and modifiable. It was thus\nable to attract support from a wide range of developers and sponsoring\ncorporations who would have been reluctant to line up behind a single\nvendor's closed product. (This, of course, prefigured an important\ntheme in the breakout of the Linux operating\nsystem\n  ten years later.)\n<\/p>\n<p>\n  The designers of X decided early on that X would support\n\u201cmechanism, not policy\u201d. Their objective was to make X as\nflexible and portable across platforms as possible, while putting as\nfew constraints on the look and feel of X programs as they could\nmanage. Look and feel, they decided, would be handled by\n\u2018toolkits\u2019 \u2014 libraries calling X services linked to\nuser programs. X would also be designed to support multiple window\nmanagers,<sup><a href=\"#ftn.id2951032\" id=\"id2951032\">[106]<\/a><\/sup>\n  and would not require a\nwindow manager to have any special privileges or uniquely close\nintegration with X's machinery.\n<\/p>\n<p>\n  This approach was the polar opposite of that taken by the\nMacintosh and Windows commercial products, which enforced\nparticular look-and-feel policies by designing them right into the\nsystem. The difference in approach ensured that X would have a\nlong-run evolutionary advantage by remaining adaptable as new\ndiscoveries were made about the human factors in interface design\n\u2014 but it also ensured that the X world would be divided by\nmultiple toolkits, a profusion of window managers, and many\nexperiments in look and feel.\n<\/p>\n<p>\n  Since the mid-1990s X has become ubiquitous even on the\nlowest-end personal Unix machines. Use of Unix from text-mode\nterminals, as opposed to graphics-capable computer consoles, has\nsharply declined and seems headed for extinction. Accordingly, the\nuse of curses-style interfaces for new applications is also in\ndecline; most new applications that would formerly have been\ndesigned in that style now use an X toolkit. It is instructive to\nnote that Unix's older CLI design tradition is still\nquite vigorous and successfully competes with X in many areas.\n<\/p>\n<p>\n  It is also instructive to note that there are a few\nspecific application areas in which curses-style (or\n\u2018roguelike\u2019) character-cell interfaces remain the norm\n\u2014 especially text editors and interactive communications\nprograms such as mailers, newsreaders, and chat clients.\n<\/p>\n<p>\n  For historical reasons, then, there is a wide range of\ninterface styles in Unix programs. Line-oriented, character-cell\nscreen-oriented, and X-based \u2014 with the X-based world somewhat\nbalkanized by the competition between multiple X toolkits and\nwindow managers (though this is less an issue in 2003 than\nwas the case five or even three years ago).\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2951032\" id=\"ftn.id2951032\">[106]<\/a><\/sup>\n  A window manager handles associations between\nwindows on the screen and running tasks. Window managers handle\nbehaviors like title bars, placement, minimizing, maximizing, moving,\nresizing, and shading windows.\n<\/p><\/div>\n\n<h3 id=\"ch11s03\"><a href=\"#ch11s03\">\u00a7<\/a>Evaluating Interface Designs<\/h3>\n<p>\n  All these interface styles survive because they are adapted for\ndifferent jobs. When making design decisions about a project, it's\nimportant to know how to pick a style (or combine styles) that will be\nappropriate to your application and your user population.\n<\/p>\n<p>\n  We will use five basic metrics to categorize interface\nstyles:\n  <em>concision<\/em>,\n  <em>expressiveness<\/em>,\n  <em>ease<\/em>,\n  <em>transparency<\/em>, and\n  <em>scriptability<\/em>.  We've already used some of\nthese terms earlier in this book in ways that were preparation for\ndefining them here.  They are comparatives, not absolutes; they have\nto be evaluated with respect to a particular problem domain and with\nsome knowledge of the users' skill base. Nevertheless, they will help\norganize our thinking in useful ways.\n<\/p>\n<p>\n  A program interface is \u2018concise\u2019 when the length and\ncomplexity of actions required to do a transaction with it has a low\nupper bound (the measurement might be in keystrokes, gestures, or\nseconds of attention required). Concise interfaces pack a lot of\nleverage into a relatively few bits or state changes.\n<\/p>\n<p>\n  Interfaces are \u2018expressive\u2019 when they can readily be\nused to command a wide variety of actions. The\n  <em>most<\/em>\n  expressive interfaces can command\ncombinations of actions not anticipated by the designer of the\nprogram, but which nevertheless give the user useful and consistent\nresults.\n<\/p>\n<p>\n  The difference between concision and expressiveness is an\nimportant one. Consider two different ways of entering text: from a\nkeyboard, or by picking characters from a screen display with mouse\nclicks. These have equal expressiveness, but the keyboard is more\nconcise (as we can easily verify by comparing average text-entry\nspeeds). On the other hand, consider two dialects of the same\nprogramming language, one with a complex-number type and one not.\nWithin the problem domain they have in common, their concision will\nbe identical; but for a mathematician or electrical engineer, the\ndialect with complex numbers will be much more expressive.\n<\/p>\n<p>\n  The \u2018ease\u2019 of an interface is inversely proportional\nto the mnemonic load it puts on the user \u2014 how many things\n(commands, gestures, primitive concepts) the user has to remember\nspecifically to support using that interface. Programming languages\nhave a high mnemonic load and low ease; menus and well-labeled\non-screen buttons are simpler.\n<\/p>\n<p>\n  Recall that we devoted an entire earlier chapter to\n\u2018transparency\u2019.  In that chapter we touched on the idea of\ninterface transparency, and gave the\n  <a href=\"#audacity\" title=\"Case Study: audacity\"><em>audacity audio editor<\/em><\/a>\n  as one superb example of it.  But we were\nthen much more interested in transparency of a different kind, one\nthat relates to the structure of code rather than of user interfaces.\nWe therefore described UI transparency in terms of its effect (nothing\nobtrudes between the user and the problem domain) rather than the\nspecific features of design that produce it.  Now it's time to zero in\non these.\n<\/p>\n<p>\n  The \u2018transparency\u2019 of an interface is how few things\nthe user has to remember about the state of his problem, his data, or\nhis program while\n  <em>using<\/em>\n  the interface.  An interface has high\ntransparency when it naturally presents intermediate results, useful\nfeedback, and error notifications on the effects of a user's\nactions. So-called WYSIWYG (What You See Is What You Get) interfaces\nare intended to maximize transparency, but sometimes backfire \u2014\nespecially by presenting an over-simplified view of the\ndomain.\n<\/p>\n<p>\n  The related concept of discoverability applies to interface\ndesign, as well.  A discoverable interface provides the user with\nassistance in learning it, such as a greeting message pointing to\ncontext-sensitive help, or explanatory balloon popups.  Though\ndiscoverability has to be implemented in rather different ways\nfor each of the interface styles we shall consider, the degree to\nwhich it is achievable is largely independent of interface style.\nThus, we shall not use it as a metric in this discussion.\n<\/p>\n<p>\n  Note that transparency of code and design does not automatically imply \ntransparency of interface, or vice versa!  It is all too easy to point\nto code that has one but not the other.\n<\/p>\n<p>\n  The \u2018scriptability\u2019 of an interface is the ease with\nwhich it can be manipulated by other programs (e.g., through the IPC\nmechanisms discussed in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>). Scriptable programs are readily\nusable as components by other programs, reducing the need for costly\ncustom coding and making it relatively easy to automate repetitive\ntasks.\n<\/p>\n<p>\n  That last point \u2014 automating repetitive tasks \u2014\ndeserves more attention than it usually gets.  Unix programmers,\nadministrators, and users develop a habit of thinking through the\nroutine procedures they use, then packaging them so they no longer\nhave to manually execute or even think about them any more.  This\nhabit depends on scriptable interfaces.  It is a quiet but tremendous\nproductivity booster not available in most other software\nenvironments.\n<\/p>\n<p>\n  It will be useful to bear in mind that humans and computer\nprograms have very different cost functions with respect to these\nmetrics. So do novice and expert human users in a particular problem\ndomain. We'll explore how the tradeoffs between them change for\ndifferent user populations.\n<\/p>\n\n<h3 id=\"ch11s04\"><a href=\"#ch11s04\">\u00a7<\/a>Tradeoffs between CLI and Visual Interfaces<\/h3>\n<p>\n  The CLI style of early Unix has retained its utility long after\nthe demise of teletypes for two reasons. One is that command-line and\ncommand-language interfaces are more expressive than visual\ninterfaces, especially for complex tasks. The other is that CLI\ninterfaces are highly scriptable \u2014 they readily support the\ncombining of programs, as we discussed in detail in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>.  Usually (though not always) CLIs\nhave an advantage in concision as well.\n<\/p>\n<p>\n  The disadvantage of the CLI style, of course, is that it almost\nalways has high mnemonic load (low ease), and usually has low\ntransparency. Most people (especially\nnon-technical end users) find such interfaces relatively cryptic and\ndifficult to learn.\n<\/p>\n <blockquote>\n  <p>\n   On the other hand, the \u2018user-friendly\u2019 GUIs of other\noperating systems have their own problems.  Finding the right buttons\nto push is like playing Adventure: the interfaces are just as\nburdensome as any Unix command line interface, save that one can in\ntheory find the treasure by sufficient exploration.  In Unix, one\nneeds the manual.\n  <\/p><div>\n  \n   \u2013 Brian Kernighan\n  \n <\/div><\/blockquote>\n<p>\n  Database queries are a good example of the kind of interface for\nwhich pushing buttons is not just burdensome but extremely limiting.\nNeither keystroke commands to a full-screen character interface nor\nGUI gestures on a graphic display can express typical actions in\nthe problem domain as expressively or concisely as typing SQL\ndirect to a server. And it is certainly easier to make a client\nprogram utter SQL queries than it would be to have it simulate a\nuser clicking a GUI!\n<\/p>\n<p>\n  On the other hand, many non-technical database users are so\nresistant to having to remember SQL syntax that they prefer a less\nconcise and less expressive full-screen or GUI interface.\n<\/p>\n<p>\n  SQL is a good example for illustrating another point.  The most\npowerful CLIs are not ad-hoc collections of commands, but imperative\nminilanguages designed along the lines we described in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>. These minilanguages are the\nhighest-power, highest-complexity end of the CLI spectrum; they\nmaximize expressiveness, but minimize ease.  They are difficult to use\nand generally need to be discreetly veiled from ordinary end-users,\nbut unbeatable when the capability and flexibility of the interface is\nthe most important thing.  When properly designed, they also score\nhigh on scriptability.\n<\/p>\n<p>\n  Some applications, unlike database queries, are naturally\nvisual. Paint programs, Web browsers, and presentation software make\nthree excellent examples. What these application domains have in common\nis that (a) transparency is extremely valuable, and (b) the primitive\nactions in the problem domain are themselves visual: \u201cdraw\nthis\u201d, \u201cshow me what I'm pointing at\u201d, \u201cput\nthis here\u201d.\n<\/p>\n<p>\n  The flip side of paint programs is that it is\ndifficult to capture relationships within the pictures they are \nmanipulating. It takes careful, thoughtful design to give the user any\nhandle on the structure of images with repeated elements, for\nexample. This is a general design problem with visual interfaces.\n<\/p>\n<p>\n  In\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>\n  we\nlooked at the Audacity sound file editor.  Its interface design\nsucceeds because it does a particularly clean job of mapping its audio\napplication domain onto a simple set of visual representations\n(borrowed from equalizer displays on stereos).  It does this by\nthoroughly following through the consequences of a single translation:\nsounds to waveform images.  The visual operations are not a mere\ngrab-bag of low-level tweaks; they are all tied to that\ntranslation.\n<\/p>\n<p>\n  In applications that are\n  <em>not<\/em>\n  naturally\nvisual, however, visual interfaces are most appropriate for simple\none-shot or infrequent tasks performed by novice users (a point the\ndatabase example illustrates).\n<\/p>\n<p>\n  Resistance to CLI interfaces tends to decrease as users become\nmore expert. In many problem domains, users (especially\n  <em>frequent<\/em>\n  users) reach a crossover point at which\nthe concision and expressiveness of CLI becomes more valuable than\navoiding its mnemonic load. Thus, for example, computing novices\nprefer the ease of GUI desktops, but experienced users often gradually\ndiscover that they prefer typing commands to a shell.\n<\/p>\n<p>\n  CLIs also tend to gain utility as problems scale up and involve\nmore in the way of canned, procedural and repetitive actions. Thus,\nfor example, a WYSIWYG desktop-publishing program is usually the\neasiest route to composing relatively small and unstructured documents\nsuch as business letters. But for complex book-sized documents that\nare assembled from sections and may require many global format changes\nor structural manipulation during composition, a minilanguage\nformatter such as\n  <em>troff<\/em>,\n  <em>Tex<\/em>, or some XML-markup\nprocessor\n  is usually a\nmore effective choice (see\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>\n  for more discussion of this\ntradeoff).\n<\/p>\n<p>\n  Even in domains that are naturally visual, scaling up the\nproblem size tends to tilt the tradeoff toward a CLI. If you need to\nfetch and save one Web page from a given URL, point and click (or type\nand click) is fine.  But for Web forms, you're going to use a\nkeyboard.  And if you need to fetch and save the pages corresponding\nto a given list of fifty URLs, a CLI client that can read URLs from\nstandard input or the command line can save you a lot of unnecessary\nmotion.\n<\/p>\n<p>\n  As another example, consider modifying the color table in a\ngraphic image. If you want to change one color (say, to lighten it by\nan amount you will only know is right when you see it) a visual\ndialogue with a color-picker widget is almost mandatory. But suppose\nyou need to replace the entire table with a set of specified RGB\nvalues, or to create and index large numbers of thumbnails. These are\noperations that GUIs usually lack the expressive power to\nspecify. Even when they do, invoking a properly designed CLI or filter\nprogram will do the job far more concisely.\n<\/p>\n<p>\n  Finally (as we observed earlier on) CLIs are important in \nfacilitating using programs from other programs. A GUI graphics\neditor that\n  <em>can<\/em>\n  handle making a batch of\nthumbnails for a list of files probably does it with a plugin written\nin a scripting language, calling an internal CLI of the graphics\neditor (as in the GIMP's\n  script-fu facility). Unix environments bring the value of CLIs into\nsharper relief precisely because their IPC facilities are rich, have\nlow overhead, and are easily accessible from user programs.\n<\/p>\n<p>\n  The explosion of interest in GUIs since 1984 has had the\nunfortunate effect of obscuring the virtues of CLIs. The design of\nconsumer software, in particular, has become heavily skewed toward\nGUIs. While this is a good choice for the novice and casual users that\nconstitute most of the consumer market, it also exacts hidden costs on\nmore expert users as they run up against the expressiveness limits of\nGUIs \u2014 costs which steadily increase as the users take on more\ndemanding problems. Most of these costs derive from the fact that GUIs\nare simply not scriptable at all \u2014\n  <em>every<\/em>\n  interaction with them has to be human-driven.\n<\/p>\n<p>\n  Gentner &amp; Nielsen sum up the tradeoff very well in\n  <em>The Anti-Mac Interface<\/em> <sup><a href=\"#Gentner-Nielsen\" title=\"[Gentner-Nielsen]\">[Gentner-Nielsen]<\/a><\/sup>: \u201c[Visual interfaces] work well for\nsimple actions with a small number of objects, but as the number of\nactions or objects increases, direct manipulation quickly becomes\nrepetitive drudgery.  The dark side of a direct manipulation interface\nis that you have to manipulate everything. Instead of an executive who\ngives high-level instructions, the user is reduced to an assembly-line\nworker who must carry out the same task over and over\u201d. Noted\nscience-fiction writer Neal Stephenson made the same point, less\ndirectly but more entertainingly, in his brilliant and discursive\nessay\n  <em>In the Beginning Was the Command Line<\/em> <sup><a href=\"#Stephenson\" title=\"[Stephenson]\">[Stephenson]<\/a><\/sup>.\n<\/p>\n<p>\n  A typical Unix old hand's take on this problem is rather less\ntheoretical:\n<\/p>\n <blockquote>\n  <p>\n   The commercial world generally goes for the novice mode because \n(a) purchase decisions are often made on the basis of 30 seconds trial, \nand (b) it minimizes the demands on customer support to have only a \ndumbed-down GUI.  I find many non-Unix systems very frustrating because,\nfor example, they will provide no way to do something on a hundred or a\nthousand files; I want to write a script, and there's no support for it.  The\nbasic problem is that they've assumed all users are novices all the\ntime, and then they bash Unix because it doesn't cater to that model.\n  <\/p><div>\n  \n   \u2013 Mike Lesk\n  \n <\/div><\/blockquote>\n<p>\n  For the long haul, then \u2014 for serving both casual and expert\nusers, for cooperating with other computer programs, and whether\nthe problem domain is naturally visual or not \u2014 support for\n  <em>both<\/em>\n  CLI and visual interfaces is important. Unix's\nhistory positions it well to meet both sets of needs. After\npresenting an indicative case study, we will examine the\ncharacteristic design patterns that the Unix tradition has evolved\nto meet them.\n<\/p>\n<h4 id=\"id2951734\"><a href=\"#id2951734\">\u00a7<\/a>Case Study: Two Ways to Write a Calculator Program<\/h4>\n<p>\n  To be more concrete, let us contrast how the GUI and CLI styles\ncan be usefully applied to the design of a simple interactive program:\na desk calculator. Our examples for contrast are\ndc(1)\/bc(1)\nand\nxcalc(1).\n<\/p>\n<p>\n  The original Unix desk calculator program, first distributed\nwith Version 7, was\ndc(1)\u2014a reverse-Polish-notation calculator that could handle\nunlimited-precision arithmetic. Later, an algebraic (infix notation)\ncalculator language,\nbc(1),\nwas implemented on top of dc (we used the relationship between these\nprograms as a case study in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>, and again in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>). Both of these programs use a\nCLI. You type an expression on standard input, you press enter, and\nthe value of the expression is printed on standard output.\n<\/p>\n<p>\n  The\nxcalc(1)\nprogram, on the other hand, visually simulates a simple calculator,\nwith clickable buttons and a calculator-style display.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 11.1. The xcalc GUI.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"The xcalc GUI.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  The\nxcalc(1)\napproach is simpler to describe because it mimics an interface with\nwhich novice users will be familiar; the man page says, in fact,\n\u201cThe numbered keys, the +\/\u2212 key, and the +, \u2212, *, \/, and = keys\nall do exactly what you would expect them to\u201d. All the\ncapabilities of the program are conveyed by the visible button\nlabels. This is the Rule of Least Surprise in its strongest form, and\na real advantage for infrequent and novice users who will never have\nto read a man page to use the program.\n<\/p>\n<p>\n  However,\n  <em>xcalc<\/em> also inherits the\nalmost complete\nnon-transparency\n  of a calculator; when\nevaluating a complex expression, you don't get to see and sanity-check\nyour keystrokes \u2014 which can be a problem if, say, you misplace a\ndecimal point in an expression like (2.51 + 4.6) * 0.3. There's no\nhistory, so you can't check. You'll get a result, but it won't be the\nresult of the calculation you intended.\n<\/p>\n<p>\n  With the\ndc(1)\nand\nbc(1)\nprograms, on the other hand, you can edit mistakes out of the\nexpression as you build it. Their interface is more transparent,\nbecause you can see the calculation that is being performed at every\nstage. It is more expressive because the\n  <em>dc<\/em> \/\n  <em>bc<\/em> interpreter, not being limited to what fits on a reasonably-sized\nvisual mockup of a calculator, can include a much larger repertoire of\nfunctions (and facilities such as if\/then\/else, stored variables, and\niteration). It also incurs, of course, a higher mnemonic load.\n<\/p>\n<p>\n  Concision is more of a toss-up; good typists will find the CLI\nmore concise, while poor ones may find it faster to point and\nclick. Scriptability is not a toss-up;\n  <em>dc<\/em> \/\n  <em>bc<\/em> can easily\nbe used as a filter, but xcalc can't be scripted at all.\n<\/p>\n<p>\n  The tradeoff between ease for novices and utility for expert\nusers is very clear here. For casual use in situations where a\nmental-arithmetic error check is not hard,\n  <em>xcalc<\/em> wins. For more complex calculations\nwhere the steps must not only be correct but must be\n  <em>seen<\/em>\n  to be correct, or in which they are most\nconveniently generated by another program,\n  <em>dc<\/em> \/\n  <em>bc<\/em> wins.\n<\/p>\n\n<h3 id=\"ch11s05\"><a href=\"#ch11s05\">\u00a7<\/a>Transparency, Expressiveness, and Configurability<\/h3>\n<p>\n  Unix programmers inherit a strong bias toward making interfaces\nexpressive and configurable.  Like programmers from other traditions,\nthey think about how to match their interfaces to the target audience\n\u2014 but they differ in how they deal with uncertainty about\nthat target audience.  Software developers whose experience is\nprimarily with client operating systems default toward making\ninterfaces simple; they are willing to sacrifice expressiveness to\ngain ease.  Unix programmers default toward making interfaces\nexpressive and transparent, and are more willing to sacrifice ease to\nget these qualities.\n<\/p>\n<p>\n  The results of this attitude have often been described as\ninterfaces written \u201cby programmers, for programmers\u201d.\nBut this oversimplifies the matter in an important way.  When a Unix\nprogrammer opts for configurability and expressiveness over ease, he\nis not necessarily thinking of his audience as consisting solely of\nother programmers; rather, he is often acting on a gut-level instinct\nthat in the absence of knowledge about end-users' intentions it is\nbest not to patronize or second-guess them.\n<\/p>\n <blockquote>\n  <p>\n   The downside of this attitude (which is a close cousin to\n\u201cmechanism, not policy\u201d) is a tendency to assume that\nwhen the highly configurable and expressive interface is done, the job\nis finished... even if the result is almost impossible for anyone else\nto use without lengthy study.  The flip side of configurability is an\nurgent need for good defaults and an easy way to set everything to the\ndefault.  The flip side of expressivity is a need for guidance \u2014 be\nit in the program or the documentation \u2014 on where to get started\nand how to achieve the most commonly-desired results.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  The Rule of Transparency also has an influence.  When a Unix\nprogrammer is writing to meet an RFC or other standard that defines a\nset of control options, he tends to assume that his job is to provide\na complete and transparent interface to all of those options; whether\nor not he thinks any given one will actually be used is secondary.\nHis job is mechanism; policy belongs to the user.\n<\/p>\n<p>\n  This mindset leads to a much stricter attitude about what\nconstitutes standards conformance, one in which incomplete support is\nmuch less tolerable.  In cases where a Macintosh or Windows developer\nwould say \u201cWe don't need to support that feature of the\nstandard; most users won't care, and it's too complicated for\nthem\u201d, a Unix developer is likely to say \u201cWe don't know\nthat nobody will ever want this feature or option, therefore we must\nsupport it\u201d.\n<\/p>\n<p>\n  These attitudes can lead to clashes when a Unix programmer is\nworking with others, who are likely to interpret his design choices as\na blithe willingness to burden users with technical details that are\nobscure, pointless, and even frightening.  Mac or Windows programmers\nfear scaring away the many  to serve the advanced needs of the\nfew.\n<\/p>\n<p>\n  The Unix programmer, on the other hand, is likely to see\ndefaulting away from expressiveness as a sort of cop-out or even\nbetrayal of future users, who will know their own requirements better\nthan the present implementer.  Ironically, though the Unix attitude is\noften construed as a sort of programmer arrogance, it is actually a\nform of humility \u2014 one often acquired along with years of battle\nscars.\n<\/p>\n<p>\n  The extent to which the Unix attitude is appropriate varies.\nWhichever side of this divide you the reader are on, it is wise to\nlearn to listen to the other, and wise to understand the premises\nbehind the opposing point of view.  Rather than falling into the trap\nof either intimidating users or condescending to them, it may be\npossible to build transparent interfaces in which the advanced\nfeatures are present but inconspicuous.  The\n  <em>audacity<\/em> and\n  <em>kmail<\/em> case studies in\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>\n  are good examples to follow.\n<\/p>\n<p>\n  Finally, a note about user-interface design for nontechnical\nend-users.  This is a demanding art, and Unix programmers don't have a\ntradition of being very good at it.  But with the ideas we've\ndeveloped from examining the Unix tradition, it is possible to make\none strong and useful statement about it.  That is: when people say a\nuser interface is\n  <em>intuitive<\/em>, what they mean is\nthat it (a) is discoverable, (b) is transparent in use, and (c) obeys\nthe Rule of Least Surprise.<sup><a href=\"#ftn.id2957556\" id=\"id2957556\">[107]<\/a><\/sup>\n  Of these three rules, Least\nSurprise is the least binding; initial surprises can be coped with \nif discoverability and transparency make longer-term use rewarding.\n<\/p>\n<p>\n  The user interfaces of today's cellphones (for example) have\nrelatively high mnemonic load in that you have to maintain at least a\nrough mental map of the interface menus to use them rapidly without\nconstantly having to spend attention on checking where you are in the\nhierarchy.  But the better-designed ones rapidly become\n\u2018intuitive\u2018 for their users anyway, because they have\nthese three qualities.\n<\/p>\n<p>\n  Intuitiveness is not quite the same quality as ease, because (as\nthe cellphone example shows) people can develop what they think of as\n\u2018intuitions\u2018 about transparent interfaces that have fairly\nhigh mnemonic load, as long as simple operations are easy and there is\na discovery path that allows them to assimilate the interface's more\ndifficult corners one step at a time.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2957556\" id=\"ftn.id2957556\">[107]<\/a><\/sup>\n  This insight comes to us\nfrom a nontechnical end-user who just happens to be the author's wife\nCatherine Raymond.\n<\/p><\/div>\n\n<h3 id=\"ch11s06\"><a href=\"#ch11s06\">\u00a7<\/a>Unix Interface Design Patterns<\/h3>\n<p>\n  In the Unix tradition, the tradeoffs we described above are met\nby well-established interface design patterns. Here is a bestiary\nof these patterns, with analyses and examples. We'll follow it with\na discussion of how to apply them.\n<\/p>\n<p>\n  Note that this bestiary does not include GUI design patterns\n(though it includes a design pattern that can use a GUI as a\ncomponent). There are no design patterns in graphical user interfaces\nthemselves that are specifically native to Unix. A promising beginning\nof a discussion of GUI design patterns in general can be found at\n  <em>Experiences \u2014 A Pattern Language for User Interface Design<\/em><sup><a href=\"#Coram-Lee\" title=\"[Coram-Lee]\">[Coram-Lee]<\/a><\/sup>.\n<\/p>\n<p>\n  Also note that programs may have modes that fit more than one\ninterface pattern. A program that has a compiler-like interface,\nfor example, may behave as a filter when no file arguments are\nspecified on the command line (many format converters behave like\nthis).\n<\/p>\n<h4 id=\"id2957637\"><a href=\"#id2957637\">\u00a7<\/a>The Filter Pattern<\/h4>\n<p>\n  The interface-design pattern most classically associated with\nUnix is the\n  <em>filter<\/em>. A filter program takes data on\nstandard input, transforms it in some fashion, and sends the result\nto standard output. Filters are not interactive; they may query\ntheir startup environment, and are typically controlled by\ncommand-line options, but they do not require feedback or commands\nfrom the user in their input stream.\n<\/p>\n<p>\n  Two classic examples of filters are\ntr(1)\nand\ngrep(1). The\ntr(1)\nprogram is a utility that translates data on standard input to\nresults on standard output using a translation specification given on\nthe command line. The\ngrep(1)\nprogram selects lines from standard input according to a match\nexpression specified on the command line; the resulting selected lines\ngo to standard output. A third is the\nsort(1)\nutility, which sorts lines in input according to criteria specified on\nthe command line and issues the sorted result to standard\noutput.\n<\/p>\n<p>\n  Both\ngrep(1)\nand\nsort(1)\n(but not\ntr(1))\ncan alternatively take data input from a file (or files) named on the\ncommand line, in which case they do not read standard input but act\ninstead as though that input were the catenation of the named files\nread in the order they appear. (In this case it is also expected that\nspecifying \u201c-\u201d as a filename on the command line will\ndirect the program explicitly to read from standard input.) The\narchetype of such \u2018catlike\u2019 filters is\ncat(1),\nand filters are expected to behave this way unless there are\napplication-specific reasons to treat files named on the command line\ndifferently.\n<\/p>\n<p>\n  When designing filters, it is well to bear in mind some\nadditional rules, partly developed in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>\n  :\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     <em>\n      Remember Postel's Prescription: Be generous in what\nyou accept, rigorous in what you emit.\n<\/em>\n     That is, try to\naccept as loose and sloppy an input format as you can and emit as\nwell-structured and tight an output format as you can.  Doing the\nformer reduces the odds that the filter will be brittle in the face of\nunexpected inputs, and break in someone's hand (or in the middle of\nsomeone's toolchain).  Doing the latter increases the odds that your\nfilter will someday be useful as an input to other programs.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>\n      When filtering, never throw away information you don't\nneed to.\n<\/em>\n     This, too, increases the odds that your filter\nwill someday be useful as an input to other programs. Information you\ndiscard is information that no later stage in a pipeline can\nuse.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>When filtering, never add noise.<\/em>\n     Avoid\nadding nonessential information, and avoid reformatting in ways that\nmight make the output more difficult for downstream programs to parse.\nThe most common offenders are cosmetic touches like headers, footers,\nblank\/ruler lines, summaries and conversions like adding aligned\ncolumns, or writing a factor of \"1.5\" as \"150%\".  Times and dates are\na particular bother because they're hard for downstream programs to\nparse. Any such additions should be optional and controlled by\nswitches.  If your program emits dates, it's good practice to have a\nswitch that can force them into ISO8601 YYYY-MM-DD and hh:mm:ss\nformats \u2014 or, better yet, use those by default.\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n<p>\n  The term \u201cfilter\u201d for this pattern is\nlong-established Unix jargon.\n<\/p>\n <blockquote>\n  <p>\n   \u201cFilter\u201d is indeed long-established.  It came into\nuse on day one of pipes.  The term was a natural transferral from\nelectrical-engineering usage: data flowed from source through filters\nto sink.  Source or sink could be either process or file.  The\ncollective EE term, \u201ccircuit\u201d, was never considered,\nsince the plumbing metaphor for data flow was already well\nestablished.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Some programs have interface design patterns like the filter,\nbut even simpler (and, importantly, even easier to script). They\nare cantrips, sources, and sinks.\n<\/p>\n<h4 id=\"id2957916\"><a href=\"#id2957916\">\u00a7<\/a>The Cantrip Pattern<\/h4>\n<p>\n  The cantrip interface design pattern is the simplest of all. No\ninput, no output, just an invocation and a numeric exit status. A cantrip's\nbehavior is controlled only by startup conditions. Programs don't\nget any more scriptable than this.\n<\/p>\n<p>\n  Thus, the cantrip design pattern is an excellent default when\nthe program doesn't require any runtime interaction with the user\nother than fairly simple setup of initial conditions or control\ninformation.\n<\/p>\n<p>\n  Indeed, because scriptability is important, Unix designers learn\nto resist the temptation to write more interactive programs when\ncantrips will do. A collection of cantrips can always be driven from\nan interactive wrapper or shell program, but interactive programs are harder\nto script. Good style therefore demands that you try to find a cantrip\ndesign for your tool before giving in to the temptation to write an\ninteractive interface that will be harder to script. And when\ninteractivity seems necessary, remember the characteristic Unix design\npattern of separating the engine from the interface; often, the right\nthing is an interactive wrapper written in some scripting\nlanguage\n  that calls a cantrip to do the real work.\n<\/p>\n<p>\n  The console utility\nclear(1),\nwhich simply clears your screen, is the purest possible cantrip; it\ndoesn't even take command-line options. Other classic simple examples\nare\nrm(1)\nand\ntouch(1).\nThe\nstartx(1)\nprogram used to launch X is a complex example, typical of a whole\nclass of daemon-summoning cantrips.\n<\/p>\n<p>\n  This interface design pattern, though fairly common, has not\ntraditionally been named; the term \u201ccantrip\u201d is my\ninvention. (In origin, it's a Scots-dialect word for a magic spell,\nwhich has been picked up by a popular fantasy-role-playing game to tag\na spell that can be cast instantly, with minimal or no\npreparation.)\n<\/p>\n<h4 id=\"id2958032\"><a href=\"#id2958032\">\u00a7<\/a>The Source Pattern<\/h4>\n<p>\n  A\n  <em>source<\/em>\n  is a filter-like program that\nrequires no input; its output is controlled only by startup\nconditions. The paradigmatic example would be\nls(1),\nthe Unix directory lister.  Other classic examples include\nwho(1)\nand\nps(1).\n<\/p>\n<p>\n  Under Unix, report generators like\nls(1),\nps(1),\nand\nwho(1)\ntend strongly to obey the source pattern, so their output can be\nfiltered with standard tools.\n<\/p>\n<p>\n  The term \u2018source\u2019 is, as Doug McIlroy noted, very\ntraditional. It is less common than it might be because\n\u2018source\u2019 has other important meanings.\n<\/p>\n<h4 id=\"id2958116\"><a href=\"#id2958116\">\u00a7<\/a>The Sink Pattern<\/h4>\n<p>\n  A\n  <em>sink<\/em>\n  is a filter-like program that\nconsumes standard input but emits nothing to standard output. Again,\nits actions on the input data are controlled only by startup\nconditions.\n<\/p>\n<p>\n  This interface pattern is unusual, and there are few well-known\nexamples. One is\nlpr(1),\nthe Unix print spooler. It will queue text passed to it on standard\ninput for printing. Like many sink programs, it will also process\nfiles named to it on the command line.  Another example is\nmail(1)\nin its mail-sending mode.\n<\/p>\n<p>\n  Many programs that might appear at first glance to be sinks take\ncontrol information as well as data on standard input and are actually\ninstances of something like the\n  <em>ed<\/em> pattern\n(see below).\n<\/p>\n<p>\n  The term\n  <em>sponge<\/em> is sometimes applied\nspecifically to sink programs like\nsort(1)\nthat must read their entire input before they can process any of\nit.\n<\/p>\n<p>\n  The term \u2018sink\u2019 is traditional and common.\n<\/p>\n<h4 id=\"id2958199\"><a href=\"#id2958199\">\u00a7<\/a>The Compiler Pattern<\/h4>\n<p>\n  Compiler-like programs use neither standard output nor standard\ninput; they may issue error messages to standard error, however.\nInstead, a compiler-like program takes file or resource names from\nthe command line, transforms the names of those resources in some\nway, and emits output under the transformed names. Like cantrips,\ncompiler-like programs do not require user interaction after\nstartup time.\n<\/p>\n<p>\n  This pattern is so named because its paradigm is the\nC\n  compiler,\ncc(1)\n(or, under Linux\n  and many other modern Unixes,\ngcc(1)). But\nit is also widely used for programs that do (for example) graphics\nfile conversions or compression\/decompression.\n<\/p>\n<p>\n  A good example of the former is the\ngif2png(1)\nprogram used to convert GIF (Graphic Interchange Format) to PNG\n(Portable Network\nGraphics).<sup><a href=\"#ftn.id2958281\" id=\"id2958281\">[108]<\/a><\/sup>\n  Good examples of the latter are\nthe\ngzip(1)\nand\ngunzip(1)\nGNU compression utilities, almost certainly shipped with your Unix\nsystem.\n<\/p>\n<p>\n  In general, the compiler interface design pattern is a good\nmodel when your program often needs to operate on multiple named\nresources and can be written to have low interactivity (with its\ncontrol information supplied at startup time). Compiler-like\nprograms are readily scriptable.\n<\/p>\n<p>\n  The term \u201ccompiler-like interface\u201d for this pattern\nis well-understood in the Unix community.\n<\/p>\n<h4 id=\"id2958336\"><a href=\"#id2958336\">\u00a7<\/a>The ed pattern<\/h4>\n<p>\n  All the previous patterns have very low interactivity; they use\nonly control information passed in at startup time, and separate\nfrom the data. Many programs, of course, need to be driven by a\ncontinuing dialog with the user after startup time.\n<\/p>\n<p>\n  In the Unix tradition, the simplest interactive design pattern\nis exemplified by\ned(1),\nthe Unix line editor. Other classic examples of this pattern include\nftp(1)\nand\nsh(1),\nthe Unix shell.  The\ned(1)\nprogram takes a filename argument; it modifies that file.  On its\ninput, it accepts command lines. Some of the commands result in output\nto standard output, which is intended to be seen immediately by the\nuser as part of the dialog with the program.\n<\/p>\n<p>\n  An actual sample\ned(1)\nsession will be included in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">\n   Chapter 13<\/a>.\n<\/p>\n<p>\n  Many browserlike and editorlike programs under Unix obey this\npattern, even when the named resource they edit is something other\nthan a text file. Consider\ngdb(1),\nthe GNU symbolic debugger, as an example.\n<\/p>\n<p>\n  Programs obeying the\n  <em>ed<\/em> interface\ndesign pattern are not quite so scriptable as would be the simpler\ninterface types resembling filters. You can feed them commands on\nstandard input, but it is trickier to generate sequences of commands\n(and interpret any output they might ship back) than it is to just set\nenvironment variables and command-line options.  If the action of the\ncommands is not so predictable that they can be run blind (e.g., with a\nhere-document as input and ignoring output), driving\n  <em>ed<\/em> -like programs requires a protocol, and a\ncorresponding state machine in the calling process.  This raises the\nproblems we noted in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  during the discussion of slave process\ncontrol.\n<\/p>\n<p>\n  Nevertheless, this is the simplest and most scriptable\npattern that supports fully interactive programs. Accordingly, it\nis still quite useful as a component of the \u201cseparated engine and\ninterface\u201d pattern we'll describe below.\n<\/p>\n<h4 id=\"id2958491\"><a href=\"#id2958491\">\u00a7<\/a>The Roguelike Pattern<\/h4>\n<p>\n  The roguelike pattern is so named because its first example was the dungeon-crawling game rogue(1) (see Figure 11.2)\n  under BSD; the adjective \u201croguelike\u201d for this pattern is widely recognized in Unix tradition. Roguelike programs are\n  designed to be run on a system console, an X terminal emulator, or a video display terminal. They use the full screen\n  and support a visual interface style, but with character-cell display rather than graphics and a mouse.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 11.2. Screen shot of the original Rogue game.\n   <\/b>\n  <\/p><pre class=\"screen\">\n                                                a) some food\n                                                b) +1 ring mail [4] being worn\n-----------------------              ########## c) a +1,+2 mace in hand \n|                     +###############          d) a +1,+0 short bow\n|                     |                         e) 28 +0,+0 arrows\n---------------+-------                         f) a short bow\n               #                                i) a magnesium wand\n               #                                g) a magnesium wand\n             ###               ---------------- j) a potion of detect things\n     --------+----------       |                l) a scroll of teleportation\n     |                 |      #+                --press space to continue--\n     |                 |      #|                 |             #\n     |                 +#######|                 |            ##\n     |                 |       |                 +##############\n     --------+----------       -------------------             #\n        ######                                                 #\n  ------+----------                                            ######\n  |...........@..!|                                                 #\n  |...........%...|                 ----------------                #\n  |...............|                #+              |          #######\n  |...............+#################|              |          #\n  |...............|                 |              +###########\n  -----------------                 ----------------\nLevel: 3  Gold: 73     Hp: 36(36)   Str: 14(16) Arm: 4  Exp: 4\/78\n<\/pre><\/div>\n<p>\n  Commands are typically single keystrokes not echoed to the user\n(as opposed to the command lines of the\n  <em>ed<\/em> pattern), though some will open a command window (often, though not\nalways, the last line of the screen) on which more elaborate\ninvocations can be typed.  The command architecture often makes heavy\nuse of the arrow keys to select screen locations or lines on which to\noperate.\n<\/p>\n<p>\n  Programs written in this pattern tend to model themselves on either\nvi(1)\n  or\nemacs(1)\nand (obeying the Rule of Least Surprise) use their command sequences\nfor common operations such as getting help or terminating the\nprogram. Thus, for example, one can expect one of the commands\n\u2018x\u2019, \u2018q\u2019, or \u2018C-x C-c\u2019 to\nterminate a program written to this pattern.\n<\/p>\n<p>\n  Some other interface tropes associated with this pattern\ninclude: (a) the use of one-item-per-line menus, with the\ncurrently-selected item indicated by bold or reverse-video\nhighlighting, and (b) \u2018mode lines\u2019 \u2014 program status\nsummaries carried on a highlighted screen line, often near the bottom\nor at the top of the screen.\n<\/p>\n<p>\n  The roguelike pattern evolved in a world of video display\nterminals; many of these didn't have arrow or function keys.  In a\nworld of graphics-capable personal computers, with character-cell\nterminals a fading memory, it's easy to forget what an influence this\npattern exerted on design; but the early exemplars of the roguelike\npattern were designed a few years before IBM standardized the PC\nkeyboard in 1981.  As a result, a traditional but now archaic part of\nthe roguelike pattern is the use of the h, j, k, and l as cursor keys\nwhenever they are not being interpreted as self-inserting characters\nin an edit window; invariably k is up, j is down, h is left, and l is\nright.  This history also explains why older Unix programs tend not to\nuse the ALT keys and to use function keys in a limited way if at\nall.\n<\/p>\n<p>\n  Programs obeying this pattern are legion: The\nvi(1)\ntext editor in all its variants, and the\nemacs(1)\neditor;\nelm(1),\npine(1),\nmutt(1), and most other Unix mail readers;\ntin(1),\nslrn(1),\nand other Usenet\n  newsreaders; the\nlynx(1)\nWeb browser; and many others.  Most Unix programmers spend most of\ntheir time driving programs with interfaces like these.\n<\/p>\n<p>\n  The roguelike pattern is hard to script; indeed scripting it is\nseldom even attempted. Among other things, this pattern uses raw-mode\ncharacter-by-character input, which is inconvenient for scripting.\nIt's also quite hard to interpret the output programmatically, because\nit usually consists of sequences of incremental screen-painting\nactions.\n<\/p>\n<p>\n  Nor does this pattern have the visual slickness of a mouse-driven\nfull GUI. While the point of using the full screen interface is to\nsupport simple kinds of direct-manipulation and menu interfaces,\nroguelike programs still require users to learn a command\nrepertoire. Indeed, interfaces built on the roguelike pattern show a\ntendency to degenerate into a sort of cluttered wilderness of modes\nand meta-shift-cokebottle commands that only hard-core\nhackers\n  can love. It\nwould seem that this pattern has the worst of both worlds, being\nneither scriptable nor conforming to recent fashions in design for\nend-users.\n<\/p>\n<p>\n  But there must be some value in this pattern. Roguelike mailers,\nnewsreaders, editors, and other programs remain extremely popular even\namong people who invariably run them through terminal emulators on an\nX display that supports GUI competitors. Moreover, the roguelike\npattern is so pervasive that under Unix even GUI programs often\nemulate it, adding mouse and graphics support to a command and display\ninterface that still looks rather roguelike.  The X mode of\nemacs(1),\nand the\nxchat(1)\nclient are good examples of such adaptation. What accounts for the\npattern's continuing popularity?\n<\/p>\n<p>\n  Efficiency, and perceived efficiency, seem to be important\nfactors. Roguelike programs tend to be fast and lightweight\nrelative to their nearest GUI competitors. For startup and runtime\nspeed, running a roguelike program in an Xterm may be preferable to\ninvoking a GUI that will chew up substantial resources setting up\nits displays and respond more slowly afterwards. Also, programs\nwith a roguelike design pattern can be used over telnet links or\nlow-speed dialup lines for which X is not an option.\n<\/p>\n<p>\n  Touch-typists often prefer roguelike programs because they can\navoid taking their hands off the keyboard to move a mouse.  Given a\nchoice, touch-typists will prefer interfaces that minimize keystrokes\nfar off the home row; this may account for a significant percentage of\nvi(1)'s\npopularity.\n<\/p>\n<p>\n  Perhaps more importantly, roguelike interfaces are predictable\nand sparing in their use of screen real estate on an X display;\nthey do not clutter the display with multiple windows, frame\nwidgets, dialog boxes, or other GUI impedimenta. This makes the\npattern well suited for use in programs that must frequently share\nthe user's attention with other programs (as is especially the case\nwith editors, mailers, newsreaders, chat clients, and other\ncommunication programs).\n<\/p>\n<p>\n  Finally (and probably most importantly) the roguelike pattern\ntends to appeal more than GUIs to people who value the concision and\nexpressiveness of a command set enough to tolerate the added mnemonic\nload. We saw above that there are good reasons for this preference to\nbecome more common as task complexity, use frequency, and user\nexperience rise. The roguelike pattern meets this preference while\nalso supporting GUI-like elements of direct manipulation as an\n  <em>ed<\/em> -pattern program cannot. Thus, far from\nhaving only the worst of both worlds, the roguelike interface design\npattern can capture some of the best.\n<\/p>\n<h4 id=\"id2958899\"><a href=\"#id2958899\">\u00a7<\/a>The \u2018Separated Engine and Interface\u2019 Pattern<\/h4>\n<p>\n  In\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  we\nargued against building monster single-process monoliths, and that it\nis often possible to lower the global complexity of programs by\nsplitting them into communicating pieces.  In the Unix world, this\ntactic is frequently applied by separating the \u2018engine\u2019\npart of the program (core algorithms and logic specific to its\napplication domain) from the \u2018interface\u2019 part (which\naccepts user commands, displays results, and may provide services such\nas interactive help or command history).  In fact, this\nseparated-engine-and-interface pattern is probably the one most\ncharacteristic interface design pattern of Unix.\n<\/p>\n<p>\n  (The other, more obvious candidate for that distinction would be\nfilters.  But filters are more often found in non-Unix environments\nthan engine\/interface pairs with bidirectional traffic between them.\nSimulating pipelines is easy; the more sophisticated IPC mechanisms \nrequired for engine\/interface pairs are hard.)\n<\/p>\n<p>\n  Owen Taylor, maintainer of the GTK+ library widely used for writing user interfaces under X, beautifully brings out\n  the engineering benefits of this kind of partitioning at the end of his note <a href=\"http:\/\/www.gtk.org\/setuid.html\">\n  Why GTK_MODULES is not a security hole<\/a>; he finishes by writing \"[T]he secure setuid program is a 500 line program\n  that does only what it needs to, rather than a 500,000 line library whose essential task is user interfaces\".\n<\/p>\n<p>\n  This is not a new idea. Xerox PARC's\n  early research into graphical user\ninterfaces led them to propose the\n\u201cmodel-view-controller\u201d\n  pattern as an archetype for\nGUIs.\n<\/p>\n <div><ul><li>\n    <p>\n     The \u201cmodel\u201d is what in the Unix world is usually\ncalled an \u201cengine\u201d.  The model contains the\ndomain-specific data structures and logic for your application.\nDatabase servers are archetypal examples of models.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The \u201cview\u201d part is what renders your domain objects\ninto a visible form.  In a really well-separated model\/view\/controller\napplication, the view component is notified of updates to the model\nand responds on its own, rather than being driven synchronously by\nthe controller or by explicit requests for a refresh.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The \u201ccontroller\u201d processes user requests and passes\nthem as commands to the model.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  In practice, the view and controller parts tend to be more\nclosely bound together than either is to the model.  Most GUIs, for\nexample, combine view and controller behavior.  They tend to be\nseparated only when the application demands multiple views of the\nmodel.\n<\/p>\n<p>\n  Under Unix, application of the model\/view\/controller pattern is\nfar more common than elsewhere precisely because there is a strong\n\u201cdo one thing well\u201d tradition, and IPC methods are both\neasy and flexible.\n<\/p>\n<p>\n  An especially powerful form of this technique couples a policy\ninterface (often a GUI combining view and controller functions) with\nan engine (model) that contains an interpreter for a domain-specific\nminilanguage.  We examined this pattern in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>, focusing on minilanguage design; now\nit's time to look at the different ways that such engines can form\ncomponents of larger systems of code.\n<\/p>\n<p>\n  There are several major variants of this pattern.\n<\/p>\n<h5 id=\"id2959084\"><a href=\"#id2959084\">\u00a7<\/a>Configurator\/Actor Pair<\/h5>\n<p>\n  In a configurator\/actor pair, the interface part controls the\nstartup environment of a filter or daemon-like program which then runs\nwithout requiring user commands.\n<\/p>\n<p>\n  The programs\nfetchmail(1)\n  and\nfetchmailconf(1)\n(which we've already used as case studies in\ndiscoverability\n  and data-driven programming and will encounter again as language case\nstudies in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>) are\na good example of a configurator\/actor pair.\n  <em>fetchmailconf<\/em> is the interactive dotfile\nconfigurator that ships with\nfetchmail.\n  <em>fetchmailconf<\/em> can also serve as a GUI\nwrapper that runs fetchmail in either foreground or background\nmode.\n<\/p>\n<p>\n  This design pattern enables both\n  <em>fetchmail<\/em> and\n  <em>fetchmailconf<\/em> to specialize in what they do\nwell, and indeed to be written in different languages appropriate to\ntheir task domains. Fetchmail, which usually runs in background as a\ndaemon, need not be bloated with GUI code. Conversely,\n  <em>fetchmailconf<\/em> can specialize in elaborate\nGUIness without exacting size and complexity costs from\nfetchmail. Finally, because the information channels between them are\nnarrow and well-defined, it remains possible to drive\n  <em>fetchmail<\/em> from the command line and from\nscripts other than\n  <em>fetchmailconf<\/em>.\n<\/p>\n<p>\n  The term \u201cconfigurator\/actor\u201d is my invention.\n<\/p>\n<h5 id=\"id2959228\"><a href=\"#id2959228\">\u00a7<\/a>Spooler\/Daemon Pair<\/h5>\n<p>\n  A slight variant of the configurator\/actor pair can be useful in\nsituations that require serialized access to a shared resource in a\nbatch mode; that is, when a well-defined job stream or sequence\nof requests requires some shared resource, but no individual\njob requires user interaction.\n<\/p>\n<p>\n  In this spooler\/daemon pattern, the spooler or front end simply\ndrops job requests and data in a spool area. The job requests and data\nare simply files; the spool area is typically just a directory. The\nlocation of the directory and the format of the job requests are agreed\non by the spooler and daemon.\n<\/p>\n<p>\n  The daemon runs forever in background, polling the spool\ndirectory, looking there for work to do. When it finds a job request,\nit tries to process the associated data. If it succeeds, the job\nrequest and data are deleted out of the spool area.\n<\/p>\n<p>\n  The classic example of this pattern is the Unix print spooler\nsystem,\nlpr(1)\/lpd(1). The\nfront end is\nlpr(1);\nit simply drops files to be printed in a spool area periodically\nscanned by\n  <em>lpd<\/em>.\n  <em>lpd<\/em> 's job\nis simply to serialize access to the printer devices.\n<\/p>\n<p>\n  Another classic example is the pair\nat(1)\/atd(1),\nwhich schedules commands for execution at specified times. A\nthird example, historically important though no longer in wide use,\nwas UUCP \u2014 the Unix-to-Unix Copy Program commonly used as a mail\ntransport over dial-up lines before the Internet explosion of the\nearly 1990s.\n<\/p>\n<p>\n  The spooler\/daemon pattern remains important in mail-transport\nprograms (which are batchy by nature). The front ends of mail\ntransports such as\nsendmail(1)\nand\nqmail(1)\nusually make one try at delivering mail immediately, through SMTP over an\noutbound Internet connection. If that attempt fails, the mail will\nfall into a spool area; a daemon version or mode of the mail transport\nwill retry the delivery later.\n<\/p>\n<p>\n  Typically, a spooler\/daemon system has four parts: a job\nlauncher, a queue lister, a job-cancellation utility, and a\nspooling daemon, In fact, the presence of the first three parts is\na sure clue that there is a spooler daemon behind them\nsomewhere.\n<\/p>\n<p>\n  The terms \u201cspooler\u201d and \u201cdaemon\u201d are\nwell-established Unix jargon. (\u2018Spooler\u2019 actually\ndates back to early mainframe days.)\n<\/p>\n<h5 id=\"id2959389\"><a href=\"#id2959389\">\u00a7<\/a>Driver\/Engine Pair<\/h5>\n<p>\n  In this pattern, unlike a configurator\/actor or spooler\/server\npair, the interface part supplies commands to and interprets output\nfrom an engine after startup; the engine has a simpler interface\npattern. The IPC method used is an implementation detail; the engine\nmay be a slave process of the driver (in the sense we discussed in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>) or the\nengine and driver may communicate through sockets, or shared memory,\nor any other IPC method. The key points are (a) the interactivity of\nthe pair, and (b) the ability of the engine to run standalone with its\nown interface.\n<\/p>\n<p>\n  Such pairs are trickier to write than configurator\/actor pairs\nbecause they are more tightly and intricately coupled; the driver\nmust have knowledge not merely about the engine's expected startup\nenvironment but about its command set and response formats as well.\n<\/p>\n<p>\n  When the engine has been designed for scriptability, however, it\nis not uncommon for the driver part to be written by someone other\nthan the engine author, or for more than one driver to front-end a\ngiven engine. An excellent example of both is provided by the programs\ngv(1)\nand\nghostview(1),\nwhich are drivers for\ngs(1),\nthe Ghostscript interpreter. GhostScript renders PostScript to various\ngraphics formats and lower-level printer-control languages. The gv and\nghostview programs provide GUI wrappers for GhostScript's rather\nidiosyncratic invocation switches and command syntax.\n<\/p>\n<p>\n  Another excellent example of this pattern is the\n  <em>xcdroast<\/em> \/\n  <em>cdrtools<\/em> combination.  The cdrtools distribution provides a program\ncdrecord(1)\nwith a command-line interface.  The\n  <em>cdrecord<\/em> code specializes in knowing\neverything about talking to CD-ROM hardware.\n  <em>xcdroast<\/em> is a GUI; it specializes in\nproviding a pleasant user experience.  The\nxcdroast(1)\n  program calls\ncdrecord(1)\nto do most of its work.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 11.3. The Xcdroast GUI.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"The Xcdroast GUI.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  <em>xcdroast<\/em> also calls other CLI tools:\ncdda2wav(1)\n(a sound file converter) and\nmkisofs(1)\n(a tool for creating ISO-9660 CD-ROM file system images from a list of\nfiles).  The details of how these tools are invoked are hidden from\nthe user, who can think in terms centered on the task of making CDs\nrather than having to know directly about the arcana of sound-file\nconversion or file-system structure.  Equally important, the\nimplementers of each of these tools can concentrate on their\ndomain-specific expertise without having to be user-interface\nexperts.\n<\/p>\n <blockquote>\n  <p>\n   A key pitfall of driver\/engine organization is that frequently\nthe driver must understand the state of the engine in order to reflect\nit to the user.  If the engine action is practically instantaneous,\nit's not a problem, but if the engine can take a long time (e.g., when\naccessing many URLs) the lack of feedback can be a significant issue.\nA similar problem is responding to errors.  For example, the\ntraditional (although not very Unix-like) confirmation question about\nwhether it's OK to overwrite a file that already exists is kind of\npainful to write in the driver\/engine world; the engine, which detects\nthe problem, has to ask the driver to do the confirmation prompting.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  It's important to design the engine so that it not only does the\nright thing, but also notifies the driver about what it's doing so\nthe driver can present a graceful interface with appropriate feedback.\n<\/p>\n<p>\n  The terms \u201cdriver\u201d and \u201cengine\u201d are\nuncommon but established in the Unix community.\n<\/p>\n<h5 id=\"id2959688\"><a href=\"#id2959688\">\u00a7<\/a>Client\/Server Pair<\/h5>\n<p>\n  A client\/server pair is like a driver\/engine pair, except that\nthe engine part is a daemon running in background which is not\nexpected to be run interactively, and does not have its own user\ninterface. Usually, the daemon is designed to mediate access to\nsome sort of shared resource \u2014 a database, or a transaction\nstream, or specialized shared hardware such as a sound device.\nAnother reason for such a daemon may be to avoid performing \nexpensive startup actions each time the program is invoked.\n<\/p>\n<p>\n  Yesterday's paradigmatic example was the\nftp(1)\/ftpd(1)\npair that implements FTP, the File Transfer Protocol; or perhaps two\ninstances of\nsendmail(1),\nsender in foreground and listener in background, passing Internet\nemail. Today's would have to be any browser\/web server pair.\n<\/p>\n<p>\n  However, this pattern is not limited to communication programs;\nanother important case is in databases, such as the\npsql(1)\/postmaster(1)\npair. In this one,\n  <em>psql<\/em> serializes access\nto a shared database managed by the postgres daemon, passing it SQL\nrequests and presenting data sent back as responses.\n<\/p>\n<p>\n  These examples illustrate an important property of such pairs,\nwhich is that the cleanliness of the protocol that serializes\ncommunication between them is all-important. If it is well-defined\nand described by an open standard, it can become a tremendous\nopportunity for leverage by insulating client programs from the\ndetails of how the server's resource is managed, and allowing\nclients and servers to evolve semi-independently. All\nseparated-engine-and-interface programs potentially get this kind\nof benefit from clean separation of function, but in the\nclient\/server case the payoffs for getting it right tend to be\nparticularly high exactly because managing shared resources is\nintrinsically difficult.\n<\/p>\n<p>\n  Message queues and pairs of named\npipes\n  can be and have\nbeen used for front-end\/back-end communication, but the benefits of\nbeing able to run the server on a different machine from the client\nare so great that nowadays almost all modern client-server pairs use\nTCP\/IP\n  sockets.\n<\/p>\n<h4 id=\"id2959821\"><a href=\"#id2959821\">\u00a7<\/a>The CLI Server Pattern<\/h4>\n<p>\n  It's normal in the Unix world for server processes to be invoked\nby harness programs<sup><a href=\"#ftn.id2959831\" id=\"id2959831\">[109]<\/a><\/sup>\n  such as\ninetd(8)\nin such a way that the server sees commands on standard input and\nships responses to standard output; the harness program then takes\ncare of ensuring that the server's stdin and stdout are connected to a\nspecified TCP\/IP\n  service port.  One benefit of this division of labor is that the\nharness program can act as a single security gatekeeper for all of the\nservers it launches.\n<\/p>\n<p>\n  One of the classic interface patterns is therefore a CLI server.\nThis is a program which, when invoked in a foreground mode, has a\nsimple CLI interface reading from standard input and writing to\nstandard output.  When backgrounded, the server detects this and\nconnects its standard input and standard output to a specified\nTCP\/IP\n  service\nport.\n<\/p>\n<p>\n  In some variants of this pattern, the server backgrounds itself\nby default, and has to be told with a command-line switch when it\nshould stay in foreground.  This is a detail; the essential point is\nthat most of the code neither knows nor cares whether it is running in\nforeground or a TCP\/IP\n  harness.\n<\/p>\n<p>\n  POP, IMAP, SMTP, and HTTP servers normally obey this pattern.\nIt can be combined with any of the server\/client patterns described\nearlier in this chapter.  An HTTP server can also act as a harness\nprogram; the CGI scripts that supply most live content on the Web run\nin a special environment provided by the server where they can take\ninput (form arguments) from standard input, and write the generated\nHTML that is their result to standard output.\n<\/p>\n<p>\n  Though this pattern is quite traditional, the term \u201cCLI\nserver\u201d is my invention.\n<\/p>\n<h4 id=\"id2959928\"><a href=\"#id2959928\">\u00a7<\/a>Language-Based Interface Patterns<\/h4>\n<p>\n  In\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  we\nexamined domain-specific minilanguages as a means of pushing program\nspecification up a level, gaining flexibility, and minimizing bugs.\nThese virtues make the language-based CLI an important style of Unix\ninterface \u2014 one exemplified by the Unix shell itself.\n<\/p>\n<p>\n  The strengths of this pattern are well illustrated by the case\nstudy earlier in the chapter comparing\ndc(1)\/bc(1)\nwith\nxcalc(1).\nThe advantages that we observed earlier (the gain in expressiveness\nand scriptability) are typical of minilanguages; they generalize to\nother situations in which you routinely have to sequence complex\noperations in a specialized problem domain. Often, unlike the\ncalculator case, minilanguages also have a clear advantage in\nconcision.\n<\/p>\n<p>\n  One of the most potent Unix design patterns is the combination\nof a GUI front end with a CLI minilanguage back end.  Well-designed\nexamples of this type are necessarily rather complex, but often a\ngreat deal simpler and more flexible than the amount of ad-hoc code\nthat would be necessary to cover even a fraction of what the\nminilanguage can do.\n<\/p>\n<p>\n  This general pattern is not, of course, unique to Unix. Modern\ndatabase suites everywhere normally consist of one or more GUI\nfront ends and report generators, all of which talk to a common\nback-end using a query language such as SQL. But this pattern\nmainly evolved under Unix and is still much better understood and\nmore widely applied there than elsewhere.\n<\/p>\n<p>\n  When the front and back ends of a system fulfilling this design\npattern are combined in a single program, that program is often said\nto have an \u2018embedded scripting\nlanguage\u2019.  In the Unix world,\n  <em>Emacs<\/em> is one of the best-known exemplars of\nthis pattern; refer to our discussion of it in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  for some advantages.\n<\/p>\n<p>\n  The script-fu facility of\nGIMP\n  is another good\nexample. GIMP is a powerful open-source graphics editor.  It has a GUI\nresembling that of Adobe Photoshop.  Script-fu allows GIMP to be\nscripted using Scheme (a dialect of Lisp); scripting through\nTcl, or\nPerl\n  or\nPython\n  is\nalso available.  Programs written in any of these languages can call\nGIMP internals through its plugin interface.  The demonstration\napplication for this facility is a Web page<sup><a href=\"#ftn.id2960094\" id=\"id2960094\">[110]<\/a><\/sup>\n  which allows people to\nconstruct simple logos and graphic buttons through a CGI interface\nthat passes a generated Scheme program to an instance of GIMP, and\nreturns a finished image.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2958281\" id=\"ftn.id2958281\">[108]<\/a><\/sup>\n  Sources for this program, and other converters with\nsimilar interfaces, are available at the\n  <a href=\"http:\/\/www.cdrom.com\/pub\/png\/\">\n   PNG\nwebsite<\/a>.\n<\/p><p><sup><a href=\"#id2959831\" id=\"ftn.id2959831\">[109]<\/a><\/sup>\n  A harness program is a wrapper\nwhose job it is to make some special sort of resource available to the\nprogram(s) it calls.  The term is most often used for test harnesses,\nwhich make available test loads and (often) examples of correct output\nfor the actual output to be checked against.\n<\/p><p><sup><a href=\"#id2960094\" id=\"ftn.id2960094\">[110]<\/a><\/sup><a href=\"http:\/\/www.xcf.berkeley.edu\/~gimp\/script-fu\/script-fu.html\">\n   Script-Fu\npage<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch11s07\"><a href=\"#ch11s07\">\u00a7<\/a>Applying Unix Interface-Design Patterns<\/h3>\n<p>\n  To facilitate scripting and pipelining (see\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>) it is wise to choose the\nsimplest interface pattern possible \u2014 that is, the pattern with the\nfewest channels to the environment and the least\ninteractivity.\n<\/p>\n<p>\n  In many of the single-component patterns described above, it is\nemphasized that the pattern does not require user interaction after\nstartup time. When the \u2018user\u2019 is often expected to be another\nprogram (and thus to lack the range and flexibility of a human\nbrain) this is a very valuable feature, maximizing\nscriptability.\n<\/p>\n<p>\n  We've seen that different interface design patterns optimize for\ntraits valuable in differing circumstances. In particular, there is\na strong and inherent tension between the GUIs and design patterns\nappropriate for novice and nontechnical end-users (on the one\nhand) and those which serve expert users and maximize scriptability\n(on the other).\n<\/p>\n<p>\n  One way around this dilemma is to make programs with modes that\nexhibit more than one pattern. An excellent example is the Web browser\nlynx(1). It\nnormally has a roguelike interface for interactive use, but can be\ncalled with a\n  <code>-dump<\/code> option that makes it into a\nsource, formatting a specified Web page to text dumped on standard\noutput.\n<\/p>\n<p>\n  Such dual-mode interfaces, however, are not normally attempted\nwhen the program has to have a true GUI. The reasons for this are\npartly historical, but mostly have to do with controlling global\ncomplexity. GUIs tend to require complex startup configurations and\nlarge volumes of specialized code; these features coexist uneasily\nwith the simpler patterns. In the worst case, a dual-mode GUI\/non-GUI\nprogram could require two separate command-interpreter loops, with\nall that implies in the way of code bloat and potential\ninconsistencies.\n<\/p>\n<p>\n  Thus, when \u201cchoose the simplest pattern\u201d conflicts\nwith a requirement to produce a GUI, the Unix way is to split the\nprogram in two, applying the \u2018separated engine and\ninterface\u2019 design pattern.\n<\/p>\n<p>\n  In fact, by combining a theme from\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  with this idea, we can perhaps name a\nnew design pattern emerging under Linux\n  and other modern, open-source\nUnixes where GUIs are not merely a reluctant add-on but an active\nfocus of lots of development effort.\n<\/p>\n<h4 id=\"id2960228\"><a href=\"#id2960228\">\u00a7<\/a>The Polyvalent-Program Pattern<\/h4>\n<p>\n  A polyvalent program has the following traits:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     The program's application-domain logic lives in a library with a\ndocumented API, which can be linked to other programs.  The program's\ninterface logic to the rest of the world is a thin layer over the\nlibrary.  Or perhaps there are several layers with different UI\nstyles, any of which the library can be linked to.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     One UI mode is a cantrip, compiler-like or CLI pattern that executes its\ninteractive commands in batch mode.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     One UI mode is a GUI, either linked directly to the\ncore library or acting as as a separate process driving the CLI\ninterface.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     One UI mode is a scripting interface using a modern\ngeneral-purpose scripting language like \nPerl,\nPython, or\nTcl.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Optional extra: One UI mode is a roguelike interface\nusing\ncurses(3).\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 11.4. Caller\/callee relationships in a polyvalent program.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Caller\/callee relationships in a polyvalent program.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  Notably, the GIMP\n  actually fulfills this pattern.\n<\/p>\n\n<h3 id=\"ch11s08\"><a href=\"#ch11s08\">\u00a7<\/a>The Web Browser as a Universal Front End<\/h3>\n<p>\n  Separating your CLI back end from a GUI interface has become an even\nmore attractive strategy since the transformation of computing by\nthe World Wide Web in the mid-1990s. For a large class of\napplications, it makes increasing sense not to write a custom GUI\nfront end at all, but rather to press Web browsers into service in\nthat role.\n<\/p>\n<p>\n  This approach has many advantages. The most obvious is that you\ndon't have to write procedural GUI code \u2014 instead, you can\ndescribe the GUI you want in languages (HTML and JavaScript) that are\nspecialized for it. This avoids a lot of expensive and complex\nsingle-purpose coding and often more than halves the total project\neffort. Another is that it makes your application instantly\nInternet-ready; the front end may be on the same host as the back end,\nor may be a thousand miles away. Yet another is that all the minor\npresentation details of the application (such as fonts and color) are\nno longer your back end's problem, and indeed can be customized by\nusers to their own tastes through mechanisms like browser preferences\nand cascading style sheets. Finally, the uniform elements of the Web\ninterface substantially ease the user's learning task.\n<\/p>\n<p>\n  There are disadvantages. The two most important are (a) the\nbatch style of interaction that the Web enforces, and (b) the\ndifficulties of managing persistent sessions using a stateless\nprotocol. Though these are not exclusively Unix issues, we'll discuss\nthem here \u2014 because it's very important to think clearly on the\n  <em>design<\/em>\n  level about when it's worthwhile to accept\nor work around these constraints.\n<\/p>\n<p>\n  CGI, the Common Gateway Interface through which a browser can\ninvoke a program on the server host, does not support fine-grained\ninteractivity well.  Nor do the templating systems, application\nservers, and embedded server scripts that are gradually replacing it\n(in a mild abuse of language, we will use CGI for all of these in\nthis section).\n<\/p>\n<p>\n  You can't do character-by-character or\nGUI-gesture-by-GUI-gesture I\/O through a CGI gateway; instead, you\nhave to fill out an HTML form and click a submit button that sends the\nform contents to a CGI script. The CGI script then runs and the server\nhands you back a page of HTML that it generated (which may itself be\nanother CGI form).\n<\/p>\n<p>\n  This is essentially a batch style of interaction, not that far\nremoved in concept from dropping punched cards in an input hopper and\ngetting back a printout. It can be made more palatable by using\nJavaScript to interact with the user, batching up transactions into\nmessages to be shipped to the server.\n<\/p>\n<p>\n  Java\n  applets can open up their own character-stream connections back to the\nserver to support smoother interactivivity. But Java has technical\nproblems (it can only use a fixed display area on the page, and can't\nchange the portion of the display outside that rectangle) and much\nworse political ones (proprietary licensing from Sun has stalled Java\ndeployment and made others reluctant to commit to it; you can't count\non every user's browser to support it).\n<\/p>\n<p>\n  Both Java\n  and JavaScript\n  can run into browser incompatibilities,\nas well.\nMicrosoft's\n  resistance\nto implementing JDK 1.2 and Swing on Internet Explorer is a serious\nproblem for Java applets, and differing Javascript version levels can\nalso break your application (though Javascript bugs are easier to\nfix).  Nevertheless, it is frequently less effort to work around these\nproblems than it would be to write and deploy a custom front end.\nA problem harder to work around is that a growing number of sophisticated\nusers routinely disable Java and even JavaScript in their browsers because\nof security problems and interface abuses.\n<\/p>\n<p>\n  As an independent issue, it is tricky to maintain session\ninformation across multiple CGI forms. The server doesn't keep any\nstate about client sessions between CGI transactions, so you can't\nrely on it to connect later form submissions with earlier ones by the\nsame user. There are two standard dodges around this: chained forms\nand browser cookies.\n<\/p>\n<p>\n  When you chain forms, you arrange for the CGI for the first form\nto generate a unique ID in an invisible field of the second form, and\nfor the second and all subsequent forms to pass that ID to their\nsuccessors. Cookies give a similar effect in a less direct way\nanalogous to environment variables (see any of the hundreds of books\non CGI design for details).  In either case, your CGI has to use the\nID as a session index (or cookies to cache state directly) and to\nhandle multiplexing the sessions explicitly.\n<\/p>\n<p>\n  It is often possible to live with these restrictions. Many\nnontrivial applications can fit into a single form and response,\nevading both problems. Even when this isn't true and the\napplication requires multiple forms, the complexity and cost\nsavings from not having to build and distribute a specialized front\nend are so large that they can easily pay for the effort required\nto write CGIs smart enough to do their own session tracking.\n<\/p>\n<p>\n  The session management problem can be addressed with application\nservers like Zope or Enhydra which provide a session abstraction, and\nservices like user authentication to programs embedded inside\nthem. The drawback of these programs is identical to their advantage:\nthe fact that they make it easier to keep per-user state on the\nserver.  That per-user state can be a problem; it eats resources, and\nit has to be timed out, because between transactions there is no way to\nknow that the user is still on the other end of the wire.\n<\/p>\n<p>\n  As usual, the best advice is to choose the simplest pattern\npossible.  Resist the temptation to do a heavyweight design relying on\nJava or an application server when simple CGIs and cookies will\ndo the job.\n<\/p>\n<p>\n  One problem with the browser-as-universal-front-end approach is that \nCGI back ends aren't readily separable from the browser environment, so\nit can be hard to script or automate transactions to the back end.\nThe Unix answer  is a three-tier architecture \u2014 Web forms\ncalling CGIs which call commands.  The automation interface is the\ncommands.\n<\/p>\n<p>\n  The way that browsers decouple front and back ends has larger\nimplications. On the Web, locking in consumers to closed, proprietary\nprotocols and APIs has become more difficult and less attractive as\nthis trend has advanced. The economics of software development are\ntherefore tilting toward HTML, XML, and other open, text-based\nInternet standards.  This trend synergizes in interesting ways with\nthe evolution of the open-source development model, which we'll survey\nin\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>. In the\nworld that the Web is creating, Unix's design tradition \u2014\nincluding the approaches to interface design we've surveyed in this\nchapter \u2014 looks more at home than ever before.\n<\/p>\n\n<h3 id=\"ch11s09\"><a href=\"#ch11s09\">\u00a7<\/a>Silence Is Golden<\/h3>\n<p>\n  We cannot leave the subject of interactive user interfaces\nwithout exploring one of the oldest and most persistent design tropes\nof Unix, the Rule of Silence.  We observed in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>\n  that well-designed\nUnix programs with nothing interesting or surprising to say should\nshut up, and suggested there are good reasons for this that have long\noutlasted the slow teletypes on which Unix was born.\n<\/p>\n<p>\n  Here's one: Programs that babble don't tend to play well with\nother programs.  If your CLI program emits status messages to standard\noutput, then programs that try to interpret that output will be put to\nthe trouble of interpreting or discarding those messages (even if\nnothing went wrong!).  Better to send only real errors to standard\nerror and not to emit unrequested data at all.\n<\/p>\n<p>\n  Here's another: The user's vertical screen space is precious.\nEvery line of junk your program emits is one less line of context\nstill available on the user's display.\n<\/p>\n<p>\n  Here's a third: Junk messages are a careless waste of the human\nuser's bandwidth.  They're one more source of distracting motion on a\nscreen display that may be mediating for more important foreground\ntasks, such as communication with other humans.\n<\/p>\n<p>\n  Go ahead and give your GUIs progress bars for long operations.\nThat's good style \u2014 it helps the user time-share his brain\nefficiently by cuing him that he can go off and read mail or do other\nthings while waiting for completion.  But don't clutter GUI interfaces\nwith confirmation popups except when you have to guard operations\nthat might lose or trash data \u2014 and even then, hide them when\nthe parent window is minimized, and bury them unless the parent window\nhas focus.<sup><a href=\"#ftn.id2960697\" id=\"id2960697\">[111]<\/a><\/sup>\n  Your job as an\ninterface designer is to assist the user, not to gratuitously get in\nhis face.\n<\/p>\n<p>\n  In general, it's bad style to tell the user things he already\nknows (\"Program &lt;foo&gt; is starting up...\", or \"Program\n&lt;foo&gt; is exiting\" are two classic offenders).  Your interface\ndesign as a whole should obey the Rule of Least Surprise, but the\ncontent of messages should obey a Rule of\n  <em>Most<\/em>\n  Surprise \u2014 be chatty only about things that deviate from\nwhat's normally expected.\n<\/p>\n<p>\n  This rule has even greater force for confirmation prompts.\nConstantly asking for confirmation where the answer is almost always\n\u201cyes\u201d conditions the user to press \u201cyes\u201d without\nthinking about it, a habit that can have very unfortunate\nconsequences.  Programs should request confirmation only when there is\ngood reason to suspect that the answer might be \u201cno no\nno!\u201d  A confirmation request that is not a surprise is a\nstrong hint of bad design.  Any confirmation prompts at all may be a\nsign that what your interface really needs is an undo command.\n<\/p>\n<p>\n  If you want chatty progress messages for debugging purposes,\ndisable them by default with a verbosity switch.  Before releasing for\nproduction, relegate as many of the normal messages as possible to\nbeing displayed only when the verbosity switch is on.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2960697\" id=\"ftn.id2960697\">[111]<\/a><\/sup>\n  If your windowing system supports translucent\npopups that intrude less between the user and the application,\n  \n   <em>use them<\/em>\n  .\n<\/p><\/div>\n\n<h3 id=\"optimizationchapter\"><a href=\"#optimizationchapter\">\u00a7<\/a>Chapter 12. Optimization<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Premature optimization is the root of all evil.\n  <\/p><div>\n  \n   \u2013 C. A. R. Hoare\n  \n <\/div><\/blockquote>\n<p>\n  This is going to be a very short chapter, because the main thing\nUnix experience teaches us about optimizing for performance is how to\nknow when not to do it. A secondary lesson is that the most effective\noptimization tactics are usually things we do for other reasons, such as\ncleanness of design.\n<\/p>\n\n<h3 id=\"ch12s01\"><a href=\"#ch12s01\">\u00a7<\/a>Don't Just Do Something, Stand There!<\/h3>\n<p>\n  The most powerful optimization technique in any programmer's\ntoolbox is to do nothing.\n<\/p>\n<p>\n  This very Zen advice is true for several reasons.  One is the\nexponential effect of Moore's Law \u2014 the smartest, cheapest, and\noften\n  <em>fastest<\/em>\n  way to collect performance gains is\nto wait a few months for your target hardware to become more capable.\nGiven the cost ratio between hardware and programmer time, there are\nalmost always better things to do with your time than to optimize a\nworking system.\n<\/p>\n<p>\n  We can get mathematically specific about this.  It is almost\nnever worth doing optimizations that reduce resource use by merely a\nconstant factor; it's smarter to concentrate effort on cases in which\nyou can reduce average-case running time or space use from\nO(\n  <em>n<\/em><sup>\n   2<\/sup>) to\nO(\n  <em>n<\/em>) or O(\n  <em>n<\/em>\n  log\n  <em>n<\/em>),<sup><a href=\"#ftn.id2962140\" id=\"id2962140\">[112]<\/a><\/sup>\n  or similarly reduce from a higher\norder. Linear performance gains tend to be rapidly swamped by Moore's\nLaw.<sup><a href=\"#ftn.id2960862\" id=\"id2960862\">[113]<\/a><\/sup><\/p>\n<p>\n  Another very constructive form of doing nothing is to not write\ncode.  The program can't be slowed down by code that isn't there.  It\ncan be slowed down by code that\n  <em>is<\/em>\n  there but\nless efficient than it could be \u2014 but that's a different\nmatter.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2962140\" id=\"ftn.id2962140\">[112]<\/a><\/sup>\n  For readers unfamiliar with O\nnotation, it is a way of indicating how the average running time of an\nalgorithm changes with the size of its inputs. An O(1) algorithm runs\nin constant time.  An O(\n  \n   <em>n<\/em>\n  ) algorithm runs in a\ntime that is predicted by\n  <code>\n   A\n   \n    <em>n<\/em>\n   \n   + C\n  <\/code>, where\n  <code>A<\/code> is some unknown\nconstant of proportionality and\n  <code>C<\/code> is an\nunknown constant representing setup time.  Linear search of a list for\na specified value is O(\n  \n   <em>n<\/em>\n  ).  An\nO(\n  \n   <em>n<\/em>\n  <sup>\n   2<\/sup>) algorithm runs\nin time\n  <code>\n   A\n   \n    <em>n<\/em>\n   <sup>\n    2<\/sup><\/code>\n  plus\nlower-order terms (which might be linear, or logarithmic, of any other\nfunction lower than a quadratic).  Checking a list for duplicate values\n(by the na\u00efve method, not sorting it) is\nO(\n  \n   <em>n<\/em>\n  <sup>\n   2<\/sup>).  Similarly,\nO(\n  \n   <em>n<\/em>\n  <sup>\n   3<\/sup>) algorithms have\nan average run time predicted by the cube of problem size; these tend\nto be too slow for practical use.  O(log\n  \n   <em>n<\/em>\n  ) is typical of tree searches.  Intelligent\nchoice of algorithm can often reduce running time from\nO(\n  \n   <em>n<\/em>\n  <sup>\n   2<\/sup>) to\nO(log\n  \n   <em>n<\/em>\n  ).  Sometimes when\nwe are interested in predicting an algorithm's memory utilization, we\nmay notice that it varies as O(1) or O(\n  \n   <em>n<\/em>\n  ) or\nO(\n  \n   <em>n<\/em>\n  <sup>\n   2<\/sup>); in general,\nalgorithms with O(\n  \n   <em>n<\/em>\n  <sup>\n   2<\/sup>)\nor higher memory utilization are not practical\neither.\n<\/p><\/div>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2960862\" id=\"ftn.id2960862\">[113]<\/a><\/sup>\n  The eighteen-month doubling time usually quoted for\nMoore's Law implies that you can collect a 26% performance gain just\nby buying new hardware in six months.\n<\/p><\/div>\n\n<h3 id=\"ch12s02\"><a href=\"#ch12s02\">\u00a7<\/a>Measure before Optimizing<\/h3>\n<p>\n  When you have real-world evidence that your application is too\nslow, then (and\n  <em>only<\/em>\n  then) is the time to think\nabout optimizing the code.  But before you do more than think about\noptimizing, measure.\n<\/p>\n<p>\n  Recall Rob Pike's six rules in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>.  One of the lessons that the original\nUnix programmers learned early is that intuition is a poor guide to\nwhere the bottlenecks are, even for one who knows the code in question\nintimately. Unixes, unlike most other operating systems, usually come with\nprofilers; use them.\n<\/p>\n<p>\n  Reading profiler results is something of an art.  There are a\ncouple of recurring problems: one is instrumentation noise, another is\nthe effect of imposed external latencies, and a third is\noverweighting of upper nodes in the call graph.\n<\/p>\n<p>\n  The instrumentation-noise problem is fundamental.  Profilers\nwork by inserting instructions that report execution time at the entry\nand exit points of subroutines, also at fixed intervals within the inline\ncode of routines.  These instructions themselves take time to execute.\nThe effect is to reduce the dispersion of call times: very short\nsubroutines tend to look more expensive than they are, with a \nlot of noise in their comparative call times, while for longer\nones the instrumentation overhead is invisible.\n<\/p>\n<p>\n  Bearing instrumentation noise in mind, it's wise to assume\nthat the times listed for the fastest, shortest subroutines are\ngoing to have a lot of froth and air in them.  They can still be\neating a lot of time if they are called very frequently, however, \nso pay particular attention to their call-count statistics.\n<\/p>\n<p>\n  The external-latency problem is also fundamental.  There are \nvarious sorts of delay and distortion that can happen behind the \nprofiler's back. The simplest is overhead from operations with\nunpredictable latency \u2014 disk and network accesses, cache fills,\nprocess-context switches, and the like.  The problem is not so much\nthat these overheads happen \u2014 they may actually be what you're\ntrying to measure, especially if you're focusing on whole-system\nperformance rather than just tuning a critical inner loop.  The\nproblem is that they have a random component that means the\nresults from any individual profiling run may not be very useful.\n<\/p>\n<p>\n  One way to minimize the effects of these noise sources, and get a\nbetter picture of where the time is going in the average case, is to\nadd together the results from a lot of profiling runs.  There are a lot\nof good reasons to build test harnesses and test loads for your\nprograms before you get to optimizing; the most important reason, usually\nfar more important than performance tuning, is so you can\nregression-test your program for correctness as you change it.  Once\nyou've done this, being able to profile repeated tests under load is a\nnice side effect that will often give you better information than\na few runs by hand.\n<\/p>\n<p>\n  Various effects tend to allocate time spent to calling routines\nrather than callees, overweighting upper modes in the call graph.\nFunction-call overhead, for example, is often charged to the calling\nroutine (whether or not this is true depends partly on your machine\narchitecture and where the profiler is allowed to insert probes).\nMacros and inline functions, if your compiler supports them, won't\nshow up in the profiling report at all; every bit of their time gets\ncharged to the calling function.\n<\/p>\n<p>\n  More importantly, many time-reporting tools give a display in\nwhich time spent in subroutines is charged to the caller. (The\ngprof(1)\nprofiler distributed with open-source Unixes has this trait.)\nNa\u00efvely subtracting callee time from caller time won't give you a\nuseful result if the same routine can have more than one caller\n\u2014 the effect would be to artificially deflate both callers'\ntimes. Especially nasty is the common case of a utility function with\nmultiple call sites, some of which make lots of trivial calls and\nothers of which make a few complicated ones.\n<\/p>\n<p>\n  To get more transparent results, factor your code so that\nupper-level routines consist as much as possible of calls to\nlower-level routines, rather than in-line code.  If you keep\nthe overhead of upper-level control logic to a minimum, the\ncall structure of the code will tend to organize the profile report\nin a way that is relatively easy to read.\n<\/p>\n<p>\n  You'll get more insight from using profilers if you think of them less\nas ways to collect individual performance numbers, and more as ways to\nlearn how performance varies as a function of interesting parameters\n(e.g., problem size, CPU speed, disc speed, memory size, compiler\noptimization, or whatever else is relevant).  Try fitting those\nnumbers to a model, using open-source software like R or a good-quality\nproprietary tool like MATLAB.\n<\/p>\n <blockquote>\n  <p>\n   The natural smoothing of the data that results from model\nfitting tends to focus on the big effects and cover up the small,\nnoisy ones.  For example, by fitting a cubic to the matrix inversion\nroutine in MATLAB on random matrices from 10 \u00d7 10 to 1000 \u00d7 1000, it is\nclear that we actually have three cubics, with clearly defined\nboundaries, that correspond roughly to \u201cin cache\u201d,\n\u201cin memory but out of cache\u201d, and \u201cout of\nmemory\u201d.  The data shows us this effect even if weren't looking\nfor it, just by looking at the deviations from the best fit.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n\n<h3 id=\"ch12s03\"><a href=\"#ch12s03\">\u00a7<\/a>Nonlocality Considered Harmful<\/h3>\n<p>\n  The most effective way to optimize your code is to keep it\nsmall and simple.  We've been through lots of good reasons to\nkeep it small and simple earlier in this book.  Here's a new \none: you want the central data structures and the time-critical loops\nin your code never to fall out of cache.\n<\/p>\n<p>\n  Consider your target machine as a hierarchy of memory types\narranged by distance from the processor.  There are the processor's\nown registers; its instruction pipeline; the level-one (L1) cache; the\nlevel-two (L2) cache; possibly a level-three (L3) cache; main memory\n(what Unix old hands still quaintly call \u2018core\u2019); and the disk\ndrives where swap space lives.  Technologies like SMP,\nshared-memory clusters, and non-uniform memory access (NUMA) add \nmore layers to the picture but only widen the overall spread.\n<\/p>\n<p>\n  Every kind of access to that stack is getting faster. Processor\ncycles are almost free, outside of a few demanding applications like\nmodeling nuclear explosions or real-time video compression.  But\nwhat's also happening is that the speed ratios between layers in the\nstorage hierarchy are all increasing as processor speeds go up.  Thus,\nthe relative cost of a cache miss is increasing.\n<\/p>\n<p>\n  So we have an interesting paradox.  As machine resources plummet, the\nexpected cost of large data structures falls \u2014 but because the cost\nspread between adjacent cache levels is also going up, the\nperformance impact of being just large enough to break a cache\nboundary is also rising.\n<\/p>\n<p>\n  \u201cSmall is beautiful\u201d is therefore better advice\nthan ever, particularly with regard to central data structures that\nmust live in the fastest possible cache.  The advice applies to code\nas well; the average instruction spends more time being loaded than\nit does executing.\n<\/p>\n<p>\n  This turns some traditional advice on its head.  Compiler\noptimizations like loop unrolling, which get rid of relatively\nexpensive machine instructions in return for an increase in total code\nsize, may no longer be worth doing.  Another example is precomputing\nsmall tables \u2014 for example, a table of sin(x) by degree for\noptimizing rotations in a 3D graphics engine will take 365 \u00d7 \n4 bytes on a modern machine.  Before processors got enough faster than\nmemory to demand caching, this was an obvious speed optimization.\nNowadays it may be faster to recompute each time rather than pay for\nthe percentage of additional cache misses caused by the table.\n<\/p>\n<p>\n  But in the future, this might turn around again as caches grow\nlarger.  More generally, many optimizations are temporary and can\neasily turn into pessimizations as cost ratios change.  The only way\nto know is to measure and see.\n<\/p>\n\n<h3 id=\"ch12s04\"><a href=\"#ch12s04\">\u00a7<\/a>Throughput vs. Latency<\/h3>\n<p>\n  Another effect of fast processors is that performance is \nusually bounded by the cost of I\/O and \u2014 especially with \nprograms that use the Internet \u2014 network transactions.  It's\ntherefore valuable to know how to design network protocols for\ngood performance.\n<\/p>\n<p>\n  The most important issue is avoiding protocol round trips as\nmuch as possible.  Every protocol transaction that requires a\nhandshake turns any latency in the connection into a potentially\nserious slowdown.  Avoiding such handshakes is not specifically a\nUnix-tradition practice, but it's one that needs mention here because\nso many protocol designs lose huge amounts of performance to\nthem.\n<\/p>\n <blockquote>\n  <p>\n   I cannot say enough about latency.  X11 went well beyond\nX10 in avoiding round trip requests: the Render extension goes even\nfurther.  X (and these days, HTTP\/1.1) is a streaming protocol.  For\nexample, on my laptop, I can execute over 4\n   <em>million<\/em>\n   1\u00d71 rectangle requests (8 million no-op\nrequests) per second.  But round trips are hundreds or thousands of\ntimes more expensive. Anytime you can get a client to do something\nwithout having to contact the server, you have a tremendous\nwin.\n  <\/p><div>\n  \n   \u2013 Jim Gettys\n  \n <\/div><\/blockquote>\n<p>\n  In fact, a good rule of thumb is to design for the lowest\npossible latency and ignore bandwidth costs until your profiling tells\nyou otherwise.  Bandwidth problems can be solved later in development\nby tricks like compressing a protocol stream on the fly; but getting\nrid of high latency baked into an existing design is much, much harder\n(often, effectively impossible).\n<\/p>\n<p>\n  While this effect shows up most clearly in network protocol\ndesign, throughput vs. latency tradeoffs are a much more general\nphenomenon.  In writing applications, you will sometimes face a choice\nbetween doing an expensive computation once in anticipation that it\nwill be used several times, or computing only when actually needed\n(even if that means you will often recompute results).  In most cases\nwhere you face a tradeoff like this, the right thing to do is bias\ntoward low latency.  That is, don't try to precompute expensive\noperations unless you have a throughput requirement and know by actual\nmeasurement that the throughput you are getting is too low.\nPrecomputation may seem efficient because it minimizes total use of\nprocessor cycles, but processor cycles are cheap.  Unless you are\ndoing one of a handful of monstrously compute-intensive applications\nlike data mining, animation rendering, or the aforementioned bomb\nsimulations, it is usually better to opt for short startup times and\nquick response.\n<\/p>\n<p>\n  In Unix's early days this advice might have been considered\nheretical.  Processors were much slower and cost ratios were very\ndifferent then; also, the pattern of Unix use was tilted rather more\nstrongly toward server operations.  The point about the value of low\nlatency needs to be made partly because even newer Unix developers\nsometimes inherit an old-time cultural prejudice toward optimizing for\nthroughput.  But times have changed.\n<\/p>\n<p>\n  Three general strategies for reducing latency are (a) batching\ntransactions that can share startup costs, (b) allowing transactions\nto overlap, and (c) caching.\n<\/p>\n<h4 id=\"id2961306\"><a href=\"#id2961306\">\u00a7<\/a>Batching Operations<\/h4>\n<p>\n  Graphics APIs are frequently written on the assumption that the\nfixed setup cost for a physical screen update is large.  Consequently, the\nwrite operations actually modify an internal buffer.  It is up to\nthe programmer to decide when enough of these updates have been batched \nand to issue the call that turns them into a physical screen update.\nPicking the right spacing of physical updates can make a great deal\nof difference to the feel of the graphics client.  Both the \nX server and the\ncurses(3)\nlibrary used by roguelike programs are organized in this way.\n<\/p>\n<p>\n  Persistent service daemons are a more Unix-specific example of\nbatching.  There are two reasons, one obvious and one subtle, to write\npersistent daemons (as opposed to CLI servers that are started up\nfresh for each session).  The obvious reason is to manage updates to a\nshared resource.  The less obvious reason, which obtains even for daemons\nthat don't handle updates, is to amortize the cost of reading in the\ndaemon's database across multiple requests.  A perfect example of this\nis the DNS service daemon\nnamed(8),\nwhich must sometimes handle thousands of requests per second, each one\nof which may actually be blocking a user's Web page load.  One of the\ntactics that makes\nnamed(8)\nfast is that it replaces parses of expensive on-disk text files\ndescribing DNS zones with accesses to a cache held in memory.\n<\/p>\n<h4 id=\"id2961372\"><a href=\"#id2961372\">\u00a7<\/a>Overlapping Operations<\/h4>\n<p>\n  In\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>\n  we\ncompared the POP3 and IMAP protocols for querying remote-mail\nservers.  We noted that IMAP requests (unlike POP3 requests) are\ntagged with a request identifier generated by the client; the server,\nwhen it ships back a response, includes the tag of the request it\npertains to.\n<\/p>\n<p>\n  POP3 requests have to be processed in lockstep by both client\nand server; the client sends a request, waits for the response to that\nrequest, and only then can prepare and ship the next one.  IMAP\nrequests, on the other hand, are are tagged so they can be overlapped.\nIf an IMAP client knows that it wants to fetch multiple messages, it\ncan stream several fetch requests (each with a different tag) to the\nIMAP server, without waiting for responses between them.  Responses,\neach tagged, will come back when the server is ready; responses to\nearly requests may come in while the client is still shipping later\nones.\n<\/p>\n<p>\n  This strategy is general to more areas than network protocols.\nIf you want to cut latency, blocking or waiting on intermediate\nresults is deadly.\n<\/p>\n<h4 id=\"binary_caches\"><a href=\"#binary_caches\">\u00a7<\/a>Caching Operation Results<\/h4>\n<p>\n  Sometimes you can get the best of both worlds (low latency and\ngood throughput) by computing expensive results as needed and caching\nthem for later use.  Earlier we mentioned that\n  <em>named<\/em> reduces latency by batching; it\nalso reduces latency by caching the results of previous network\ntransactions with other DNS servers.\n<\/p>\n<p>\n  Caching has its own problems and tradeoffs, which are well\nillustrated by one application: the use of binary caches to eliminate\nparsing overhead associated with text database files.  Some variants\nof Unix have used this technique to speed up access to their password\ninformation (the usual motivation was to cut latency on logins at very\nlarge sites).\n<\/p>\n<p>\n  To make this work, all code that looks at the binary cache has\nto know that it should check the timestamps on both files and\nregenerate the cache if the text master is newer.  Alternatively, all\nchanges to the textual master must be made through a wrapper that will\nupdate the binary format.\n<\/p>\n<p>\n  While this approach can be made to work, it has all the\ndisadvantages that the SPOT\n  rule would lead us to expect.  The\nduplication of data means that it doesn't yield any economy of storage\n\u2014 it's purely a speed\noptimization.\nBut the real problem with it is that the code to ensure coherency\nbetween cache and master is notoriously leaky and bug-prone.  Very\nfrequently updated cache files can lead to subtle race conditions\nsimply because of the 1-second resolution of timestamps.\n<\/p>\n<p>\n  Coherency can be guaranteed in simple cases.  One such is the\nPython interpreter, which compiles and deposits on disk a\np-code file with extension\n  <code>.pyc<\/code> when a\nPython library file is first imported. On subsequent runs the cached\ncopy of the p-code is loaded unless the source has since changed (this\navoids reparsing the library source code on every run).  Emacs\nLisp uses a similar technique with\n  <code>.el<\/code> and\n  <code>.elc<\/code> files.  This technique works because both\nread and write accesses to the cache go through a single\nprogram.\n<\/p>\n<p>\n  When the update pattern of the master is more complex, however,\nthe synchronization code tends to spring leaks. The Unix variants that\nused this technique to speed up access to critical system databases\nwere infamous for spawning system-administrator horror stories that\nreflected this.\n<\/p>\n<p>\n  In general, binary cache files are a brittle technique and\nprobably best avoided.  The work that went into implementing a\nspecial-purpose hack to reduce latency in this one case would have\nbeen better spent improving the application design so it doesn't have\na bottleneck there \u2014 or even on tuning to improve the speed of\nthe file system or the virtual-memory implementation.\n<\/p>\n<p>\n  When you think you are in a situation that demands caching, it\nis wise to look one level deeper and ask why the caching is necessary.\nIt may well be no more difficult to solve that problem than it would\nbe to get all the edge cases in the caching software right.\n<\/p>\n\n<h3 id=\"complexitychapter\"><a href=\"#complexitychapter\">\u00a7<\/a>Chapter 13. Complexity<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Everything should be made as simple as possible, but no simpler.\n  <\/p><div>\n  \n   \u2013 Albert Einstein\n  \n <\/div><\/blockquote>\n<p>\n  At the end of\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>, we summarized the Unix philosophy as\n\u201cKeep It Simple, Stupid!\u201d  Throughout the Design\nsection, one of the continuing themes has been the importance of\nkeeping designs and implementations as simple as possible.  But what\n  <em>is<\/em>\n  \u201cas simple as possible\u201d?  How do\nyou tell?\n<\/p>\n<p>\n  We've held off on addressing this question until now because\nunderstanding simplicity is complicated.  It needs some of the ideas\nwe developed earlier in the Design section, especially in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  and\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>, as background.\n<\/p>\n<p>\n  The large questions in this chapter are central preoccupations\nof the Unix tradition, some of them motivating holy wars that have\nsimmered for decades.  This chapter starts from established Unix\npractice and vocabulary, then goes a bit further beyond it than we do in\nthe rest of the book.  We don't try to develop simple answers to\nthese questions, because there aren't any \u2014 but we can hope that\nyou will walk away with better conceptual tools for developing\nyour own answers.\n<\/p>\n\n<h3 id=\"ch13s01\"><a href=\"#ch13s01\">\u00a7<\/a>Speaking of Complexity<\/h3>\n<p>\n  As with previous issues about modularity and interface design,\nUnix programmers react to a set of distinctions they have often \nlearned from experience without knowing how to articulate.  Therefore\nwe'll need to start by developing some terminology.\n<\/p>\n<p>\n  We will start by defining what software complexity is.  We will\nmake some horizontal distinctions between different flavors of\ncomplexity, which sometimes have to be traded off against each other.\nWe will finish by making some even more important vertical\ndistinctions, between the kinds of complexity we must live with and\nthe kinds we have the option to eliminate.\n<\/p>\n<h4 id=\"id2964646\"><a href=\"#id2964646\">\u00a7<\/a>The Three Sources of Complexity<\/h4>\n<p>\n  Questions about simplicity, complexity, and the right size of\nsoftware arouse a lot of passion in the Unix world.  Unix programmers\nhave learned a view of the world in which simplicity is beauty is\nelegance is good, and in which complexity is ugliness is\ngrotesquery is evil.\n<\/p>\n<p>\n  Underlying the Unix programmer's passion for simplicity is a\npragmatic fact: complexity costs.  Complex software is harder to think\nabout, harder to test, harder to debug, and harder to maintain \u2014\nand above all, harder to learn and use.  The costs of complexity,\nrough as they are during development, bite hardest after\ndeployment. Complexity creates places for bugs to nest, from which\nthey will emerge to trouble the world through the entire lifetime of\ntheir software.\n<\/p>\n<p>\n  All kinds of pressures tend to drag programmers into a swamp of\ncomplexity nevertheless.  We've examined a rogue's gallery of these in\nearlier chapters; feature creep and premature optimization are the two\nmost notorious.  Traditionally, Unix programmers push back against\nthese tendencies by proclaiming with religious fervor a rhetoric that\ncondemns all complexity as bad.\n<\/p>\n<p>\n  So what exactly do we mean by \u2018complexity\u2019?  This\npoint is worth pinning down, because it varies by observer.\n<\/p>\n<p>\n  Unix programmers (like other programmers) tend to focus on\n  <em>implementation complexity<\/em> \u2014basically,\nthe degree of difficulty a programmer will experience in attempting to\nunderstand a program so he or she can mentally model or debug\nit.\n<\/p>\n<p>\n  Customers and users, on the other hand, tend to see complexity\nin terms of the program's\n  <em>interface complexity<\/em>.\nIn\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>\n  we discussed\nthe quality of ease and its inverse, mnemonic load.  To a user,\ncomplexity correlates closely with mnemonic load.  Poor expressiveness\nand concision can matter too, if a weak interface forces the user to\nperform lots of error-prone or merely tedious low-level operations\nrather than a few high-level ones.\n<\/p>\n<p>\n  Driven by both of these is a third measure that is much simpler:\nthe total number of lines of code in the system, its\n  <em>codebase size<\/em>.  In terms of life-cycle costs,\nthis is usually the most important measure.  The reasons go back to\nperhaps the most important empirical result in software engineering,\none we've cited before: the defect density of code, bugs per hundred\nlines, tends to be a constant independent of implementation language.\nMore lines of code means more bugs, and debugging is the most\nexpensive and time-consuming part of development.\n<\/p>\n<p>\n  Codebase size, interface complexity and implementation\ncomplexity may all rise together.  That is the usual result of feature\ncreep, and why programmers especially dread it.  Premature\noptimization doesn't tend to raise interface complexity, but it has\nbad effects (often severely bad) on implementation complexity and\ncodebase size.  But those sorts of arguments against complexity are\nrelatively easy to win; the difficult ones begin when these three\nmeasures have to be traded off against each other.\n<\/p>\n<p>\n  We've already mentioned one situation in which two measures vary in\nopposite directions: a user interface that has been designed primarily\nto preserve implementation simplicity, or keep codebase size down, may\nsimply dump low-level tasks on the user.  (A crude example of this,\nbarely imaginable to a Unix programmer but all too common elsewhere,\nmight be an editor that lacked a global-replace feature.)  Though this\nsort of design failure is all too common, it does not traditionally\nhave a name. We'll call it a\n  <em>manularity trap<\/em>.\n<\/p>\n<p>\n  Pressure to keep the codebase size down by using extremely dense\nand complicated implementation techniques can cause a cascade of\nimplementation complexity in the system, leading to an un-debuggable\nmess.  This used to happen frequently when fitting programs onto very\nsmall systems demanded assembler programming or tricks like\nself-modifying code; nowadays it is uncommon except in embedded systems,\nand rapidly becoming rare even there.  This kind of design failure doesn't\nhave a traditional name, but one might call it a\n  <em>blivet trap<\/em>, after an old Army term for the results of attempting\nto stuff ten pounds of horse manure into a five-pound bag.\n<\/p>\n<p>\n  The blivet trap won't appear in our case studies, but we've\ndefined it for contrast with its opposite.  It can happen that the\ndesigners of a project are so wary of implementation complexity that\nthey reject a complex but unified way to solve a whole class of\nproblems in favor of lots of duplicative, ad-hoc code that solves each\nindividual one in turn. The result is bloat in the size of the\ncodebase, and maintainability problems more severe than if the unified\nmethod had been accepted.  For example, a Web project that really\nneeds a centralized relational database behind its pages might instead\nspawn several different keyed data files containing information that\nhas to be reintegrated at page generation time.  This sort of failure\nis all too common.  It doesn't have a traditional name; we'll call it\nan\n  <em>adhocity trap<\/em>.\n<\/p>\n<p>\n  These are the three faces of complexity, and some of the traps\ndesigners fall into in attempts to avoid them.<sup><a href=\"#ftn.id2964826\" id=\"id2964826\">[114]<\/a><\/sup>\n  We'll see more examples\nwhen we get to the case studies later in the chapter.\n<\/p>\n<h4 id=\"id2964843\"><a href=\"#id2964843\">\u00a7<\/a>Tradeoffs between Interface and Implementation Complexity<\/h4>\n<p>\n  One of the most perceptive observations ever made about the Unix\ntradition by someone standing outside it was contained in Richard\nGabriel's paper called\n  <em>Lisp: Good News, Bad News, and How to Win Big<\/em><sup><a href=\"#Gabriel\" title=\"[Gabriel]\">[Gabriel]<\/a><\/sup>.\nGabriel is a long-time leader of the Lisp community, and the paper was\nprimarily an argument for a particular style of\nLisp\n  design,\nbut the author himself acknowledges that it is now remembered\nprimarily for the section called \u2018The Rise of\n  <em>\n   Worse Is\nBetter\n<\/em>\n  \u2019.\n<\/p>\n<p>\n  The paper argued that Unix and C have the characteristics of\nviruses, and that in the evolutionary struggle among software designs\ntraits like implementation simplicity and portability which lead to\nrapid propagation (infectiousness) are more effective than correctness\nand completeness of the design.  Gabriel came so close to anticipating\nthe \u2018many-eyeballs\u2019 effect on open-source software that\nthe open-source community retrospectively adopted him as one of its\ntheorists after 1997.\n<\/p>\n<p>\n  Less remembered is that the Gabriel's central argument was about\na very specific tradeoff between implementation and interface\ncomplexity, one which rather exactly fits the categories we have\nexamined in this chapter.  Gabriel contrasts an \u2018MIT\u2019\nphilosophy most valuing interface simplicity with a \u2018New\nJersey\u2019 philosophy most valuing implementation simplicity.\nHe then proposes that although the MIT philosophy leads to software\nthat is better in the abstract, the (worse) New Jersey model has better\npropagation characteristics.  Over time, people pay more attention \nto software written in the New Jersey style, so it improves faster.\nWorse becomes better.\n<\/p>\n<p>\n  In fact, the MIT and New Jersey philosophies have analogs as\nconflicting tendencies within the Unix design tradition itself. One\nstrain of Unix thinking emphasizes small sharp tools, starting designs\nfrom zero, and interfaces that are simple and consistent.  This point\nof view has been most famously championed by Doug McIlroy. Another\nstrain emphasizes doing simple implementations that work, and that\nship quickly, even if the methods are brute-force and some edge cases\nhave to be punted.  Ken Thompson's code and his maxims about\nprogramming have often seemed to lean in this direction.\n<\/p>\n<p>\n  The tension between these approaches arises precisely because\none can sometimes get a simpler interface if one is willing to pay \nimplementation complexity for it, or vice versa.  Gabriel's original\nexample, about how system calls that do long operations handle\ninterrupts they cannot hold or mask, is still one of the best.  Under\nthe MIT philosophy, the right thing to do would be to back out of\nthe system call and automatically resume it once the interrupt has\nbeen handled; this is harder to implement but leads to a simpler\ninterface.  Under the New Jersey philosophy, the system call would\nreturn an error indicating that it has been interrupted and the\nuser must re-execute; this can be implemented far more simply, but\nleads to a programming interface that is more difficult to use.\n<\/p>\n<p>\n  Both approaches have been tried.  Old Unix hands will instantly\nthink of System-V-style vs. BSD-style handling of software signals;\nthe latter follows the MIT philosophy, while the former hails from New\nJersey. Underlying the choice between them is a pressing question\nthat has nothing directly to do with the software's infectiousness:\nif your goal is to hold down total global complexity, where are you\nmost willing to pay to do that?  Where\n  <em>should<\/em>\n  you\nbe most willing to pay?\n<\/p>\n<p>\n  One epochal example not mentioned in Gabriel's paper is from\ndistributed hypertext systems.  Early distributed-hypertext projects\nsuch as NLS and Xanadu were severely constrained by the MIT-philosophy\nassumption that dangling links were an unacceptable breakdown in the\nuser interface; this constrained the systems to either browsing only a\ncontrolled, closed set of documents (such as on a single CD-ROM) or\nimplementing various increasingly elaborate replication, caching, and\nindexing methods in an attempt to prevent documents from randomly\ndisappearing.  Tim Berners-Lee cut through this Gordian knot by\npunting the problem in classic New Jersey style.  The simplicity of\nimplementation he bought by allowing \u201c404: Not Found\u201d as\na response was what made the World Wide Web lightweight enough to\npropagate and succeed.\n<\/p>\n<p>\n  Gabriel himself, while sticking with the observation that\n\u2018worse\u2019 is more infectious and tends to win in the end,\nhas publicly changed his mind several times about the underlying\ncomplexity-related question of whether or not this is actually a good\nthing.  His uncertainty mirrors a lot of ongoing design \ndebates within the Unix community.\n<\/p>\n<p>\n  We cannot offer a one-size-fits-all answer.  As with most of the\nlarge questions in this chapter, good taste and engineering judgement\nwill demand different answers in different situations.  The important\nthing is to develop the habit of thinking carefully about this issue\non each and every one of your designs.  As we have observed before in\ndiscussing software modularity, complexity is a cost you must budget\nvery carefully.\n<\/p>\n<h4 id=\"id2961759\"><a href=\"#id2961759\">\u00a7<\/a>Essential, Optional, and Accidental Complexity<\/h4>\n<p>\n  In an ideal world, Unix programmers would craft only small,\nperfect gems of software, each minimal, each elegant, each\nperfect. But one of the unfortunate things about reality is that it\noften poses complex problems that demand complex solutions.  You can't\ncontrol a jetliner with an elegant ten-line procedure.  There are too\nmany pieces of equipment, too many channels and interfaces, too many\ndifferent processors \u2014 too many different subsystems defined by\nindependently operating human beings who often don't agree even on\nfundamental conventions.  Even if you are successful at making all the\nindividual software parts of an avionics system elegant, integration\nis likely to produce a large, complex, and grubby body of code with\n(one hopes) the single virtue that it will actually\n  <em>work<\/em>.\n<\/p>\n<p>\n  Jetliners have\n  <em>essential complexity<\/em>.  \nThere is a rather sharp point past which it's not possible to trade\naway features for simplicity, because the plane has to stay in the air.  \nBecause of that very fact, avionics control systems do not tend to\nspawn religious wars about complexity \u2014 and Unix programmers\ntend to stay away from them.\n<\/p>\n<p>\n  Jetliners are certainly not immune from system failures due to\novercomplexity.  But the design issues are easier to discern and think\nabout in software for which the requirements are more flexible, in\nwhich it is easy to trade off between anticipated features and\ncomplexity. (Here, and in the rest of this chapter, we will use\n\u2018feature\u2019 in a very general sense that includes things\nlike performance gains or overall degree of interface polish.)\n<\/p>\n<p>\n  To sharpen our vision, we need to begin by noticing a difference\nbetween\n  <em>accidental complexity<\/em> and\n  <em>optional complexity<\/em>.<sup><a href=\"#ftn.id2961828\" id=\"id2961828\">[115]<\/a><\/sup>\n  Accidental complexity\nhappens because someone didn't find the simplest way to implement a\nspecified set of features.  Accidental complexity can be eliminated by\ngood design, or good redesign.  Optional complexity, on the other\nhand, is tied to some desirable feature.  Optional complexity can be\neliminated only by changing the project's objectives.\n<\/p>\n<p>\n  When we fail to distinguish between optional and accidental\ncomplexity, design debates become seriously confused.  Questions about\nwhat a project's objectives are get confused with questions about the\naesthetics of simplicity, and whether people have been sufficiently\nclever.\n<\/p>\n<h4 id=\"id2961870\"><a href=\"#id2961870\">\u00a7<\/a>Mapping Complexity<\/h4>\n<p>\n  So far, we've developed two different scales for thinking about complexity.  These scales are actually orthogonal to\n  each other. Figure 13.1 may help clarify the relationships.  Each of the nine boxes of the figure lists a common\n  source of a particular kind of complexity.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 13.1. Sources and kinds of complexity.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Sources and kinds of complexity.\" src=\"denied:data:image\/png;base64,iVBORw0KGgoAAAANSUhEUgAAAr4AAADtCAAAAAB8aSnZAAAAAnRSTlMA\/1uRIrUAAAACYktHRAD\/h4\/MvwAAAAlwSFlzAAAAZAAAAGQAD5bF3QAAHAlJREFUeNrtXTuuxDpyZeDAiTEmjMkfFBgz6VNowJEChw8wGvAKFE+mFTxAKzCgaILJlMwCGHgDAmYF2gK3QLP4p\/7dLUrNvnWA2x9KIg9LpymqildFxMn4v\/8+u8Y5xiYbjEg3JV1ytrR++\/f08mUsfRsnMkW6yeieLt8\/\/eEK2unbOJEp0k1G92z58j\/+M7+AdvImzmSKdJPRPVu+fyXkrxfQTt7EmUyRbjK6Z8v3N0J+u4B28ibOZIp0k9E9W75\/IuSXC2gnb+JMpkg3Gd2T5cv\/SAhNPvnNycBINyXdk+Urp74XTH5zMjDSTUn3ZPn+BvJNPvnNycBINyXdk+X7519++cMvf37yIN6tbhpWaCe2y2nITg+Z0T09bCGalfJuLlKqXnm3SmJY3pKTgZFuSrrXybesVg+ZkhhWt1jaqQ1zFrLTQ2Z0L5NvP9CZWbjxUTgSo2bVuj1Qvkh3i+5l8m1Eo4bfhpKy169q0tAWhMr3ByF07ElHSS0okQVNSUqUL9LdpnuVfAcmOJWDa9PLgZjpVxBnI2fEIxEdeCwaKJDTXQZV1Eq7KF+ku0X3KvnWoM\/a3q3pVxAn1e91ZwuEkS\/vOpQv0t2he5F8RyjtyCiIso1+VcMrU++1Umwg37HkOPoi3T26F8lXLeMXtBG1vIEbOv0K4nyUclYhJw+yYOytfCsumoJzlC\/S3aF7jXwZ3J2JB6WM10SKWL8y0o5c3rM1RSsaXdCNrRyPS9IOlDa0\/JvcY5H2MSrDsd36hbLx2KF77WSnh\/Pp9kd3PHi2IrrX+X1PxIKBK0K43EBKTr094s7RtdpmNmgpISXjrZyuF9CdoaGkqHlFNKatD+tGXNeDr\/0U0M2vm\/0\/RDdCRQpweTbkwI\/8sMKUFQ9wjOh+i3ylLeHXG4f2fOdWf9mDWJihgBHB+UxAZHKfBxQ+mkbv28xaf0W+rvYNHByPDux2sKajoy+h5Xa\/g2YPK4wEBx2l+zXyFUUtXx4+FhIYRAdCYIMbLngQIZnLF6pi1hEiHur6x0pu9lUNhAPPq\/KFv6CecbLSNAjfLE9mgj6ozvnjOZ\/X5MtWB83D8h2UiXy\/F2ocV6zLR7F8GNEcw17v4JvkC36NsREqFjIUpBy0QVT4AwIhsMHERWQhpQSIqggJadQXFTkxVVXMGLSloZvPn4yO9HXDtZ9aFfqjZ0zX5dtSVY8+WN4KVKTSXBxr00Swm3D960hNaec6Fxwvv7JCTk6GSf\/1cdYK4im6MfWOqKvTUo2WXF\/\/c2Bdc0qM5ePDXIgKOKpe00J2YHfd+DfJF\/wa4OCAHsmBsq70x9p5ks1YqsQwil5d\/vQGrgtV5EShL0jD4RoJEUIr2Ui+cJCrUhneHz1luiZfVTvU44M2o\/lBzFgHuwX946Itot1G94OSlujbaf+9XZbn68flKydr\/UqNnlxgXVNqLR8f5kNUxJySuhZ8X0hfJd+G8odwVrBuNxX+CM8wnE0pAq5C2MGGOpo216ATq1s1DHATYTFwYtbt1MuLPvcmD+YnFgZtNPsp62C3uH+zzrndBzUcRjVFxz1Jd0q9psOsRkKiokkHuLd8fJgPUdmDRjJ2+3Pgr5Ivp1UnpidRhz+mZ7gnpFKaDOWrIydQCBfoxnudKzX37R\/2xAn3IZCvO3rOdF++YdBGDZ0z1sFucSFdky9XPYxrItGbeIrulDovi+UafdG0A97yUWEQorIHyeH3gI4+RL7PevxWDFwX1nr1KNhDX+NU+IMFJlOlVmjBBhM5gcJS1j9U7oS0ap7xUCyJPVLLd1COIZg8uKPnTHflGwVt4C9k7ZoI2rD94\/CjmnbO\/qAY+Fzj\/nu7vClfsAWnyzX6opCZKTWWjw4LQlT2IDHSA3dwnyHf2Txsz\/u3YmAVmYZYiBx54DcOH3X4YyxJC99MXARuHCrtTzAbesJ05ATAS3n7UPOhIYZYV8pNcMagSLfdq1moPIQVTDXpjp4zXaRranf1QNCGMPmV92TwrE0TwW6KYaXHsIJUo+tDcLz8+ju4lWsR99\/aRUeHnqEbodNT8IEKsVCjLlKMA+vqUmv56DAfolKWhF7LS96Bf\/j9DPkujgRbI\/K7caGBccb651zky9jz258Qxtpo4vSTlzxIeNzyR2R0i3wD96Q5NTPX4I73710D0x6iXt17lRxCWj3kJ9+Dlh9Ze4TEVfLVa9S1sy9yTyrHpfebVrQSj3HX+\/eugbuCkMfzIfYXkFQP\/ZKr7nPpAg5avj3WsYvka1ane2efc08qx2XgN+W0Yfvev5+9BgbperqXyNeHrUKvY+R2sT7NruD73r+cDIx0U9K9RL5+dbqYy5dGftOmava9fzkZGOmmpHuJfM3q9ImPkFjHZeA37dlI2K73LycDI92UdC+Rr16d7px9zj1pHJfeb8ooYwX5dc\/7l5OBkW5Kuvf6fXdbX64sJwMj3ZR0P1m+q96\/nAyMdFPSvVW+O47LVe9fTgZGuinpfkTQ+Hna6ds4kSnSTUYX5ZueKdJNRhflm54p0k1GF+WbninSTUYX5ZueKdJNRhflm54p0k1GF+WbninSTUYX5ZueKdJNRhflm54p0k1GF+WbninSTUY3T\/nWTSaoGdJNSTdL+SIQGihfRMZA+SIyBsoXkTGylO949z3DcYxINyXdLOWbk2sH6aaki\/JNzxTpJqOL8k3PFOkmo4vyTc8U6Saji\/JNzxTpJqOL8k3PFOkmo4vyTc8U6Saji\/JNzxTpJqOL8k3PFOkmo7svX770KPeOlmupM06Sb8rcFtfhA\/RwIG32J9F9BpF8O3gMbwdJocLUGbxbUPhIxnJNX1vyNanZo7LlPB3D1u8qJwM\/RVelmSekqQgk4GKk5LYENlFCayE39WrT6BLcc5XUh8lyWcTCdO9DTQl5PFLRvRvx6AtPP6\/ng+qCkLbEtSFfm5rdF6zv+yPlaxL+9VJ8YJqu8yU6B7L6g3ygDy6CBPcuR6J6er1L9z7qhyT\/GPkO+nm6Ue51bZshlvVr8rWp2We50DVm+cXXad9tuKN4Sb5ghwIyuj7CEvdY74G2ApLzBAnunXzVM79duvda50k83nze8mV6QhvkXm8LonInUvOwR50cCC5bq33ckK\/LcTHJhb6SX3zbylngFflyEB\/krlZpFl2JGmmVBTvKHn539+bk69K90wOp\/d6jey8m8i1JlBt6IE2nswARbhKemuRAL46+3thxLvS1\/OLrtO823FE8L18JlSycNqIZoxLSFLW+QHXEzrqIP4p4+dp076qEPcjhDHZ5y1fUelB12SdsEvWGds5cfDUjrsaWfF1qdl3HcqJzn198nfbdhjuKlyYPytYN5Y+4hIjSunvixOCz0delex\/jvc+ney9mt26y2yKQr1IzfGrVdCrO\/LyCDfn61OxxLvS1\/OLrtO823FG8JF91O8Bp1cUlkKexDvYKPszkq9O9l0289\/l078VMvqKiHfcifZRMjYuj9TWY5EAvytenZo9zoa\/lF1+nfbfhjuI1z4Ny\/9TFpISofOOuTH\/g\/jsxBT7de0PU\/ONHyHdsIOU4uBl97vW\/P4iccbXiQUo9FJi84C1ZN8mW39elZo9zoa\/kF9+28lH004IgDvNKYth+sZQvzy+f1MPQyDtlQtSvXN+4uRL5oQOvWSsNZfPamwT3sEneS5B6gIL\/8uneRSdvsmnVHm3+GN1n4iBb2Ah9HaZ7U9D4vWaXDSxPlbKsHHgCWc56GMRhhhdoLFtsMbgjMhzOtunypoQbySOneDdr\/Fbo6zDdL5KvIHpq2FZen4tNkYVPx0GeKf4y+dbqFp6dsi5gNnY8r+W75PtmLvQ1+apAqShcp1RY5LB8+f5VcWMm+SPkaxxQfocFk5milWhUgInFWCsWjxjWZxjfteKMqDvzrgLDyEk7HSEsIkWtYi46NGLjMOabCmLpj01JKfkLLQQrtL1shKaFCK2qTnSkrxtTny7RIRZT6RrT75Gv72QYaTI20iYxwSc1mQoLhDRb2RuzCaFDX6ER5cni2vOvzGz2dAGzZbpfJV95q81FOaq4E6xz0Y45rq5TOjRi4zA+UGI+QpSrL0Vdy\/mdrs3sIv94o6vTN\/q6PlOiQiy20jWm3yjfKNKkbGRNYoJPkKc6LGjkLW\/PrCF1ZaER1ZTEp7p2B9uA2TLd75Ivpw1TxqjVVc76lX1oxMZhJoESIu\/IGsEruJ\/oBneqlM9Qloy1j9qYP1OiQiy20jWmXyRfdyGPDKhsZE3i7R0V6Fu5OnDQxEaM5Ov3dAGzZbrfJV95rQF\/nDQMsOAT+co\/G4eJCtWfnI6DF7quG1+btmPTmOoi+aoSHWJxwZ0Vpt8j39JpKTYg2MgaKZBvWKBdfdaQYmbEmXzNnjZgtkx3V75P3w++IN9n21iT7yDHAb3YqqNMjL0Ni\/jQiI3DBIES87HRRh2pvYOwERpWdKa6UL66RIdYbKVrTFPJ9ywH7HG6TQGnCiwRRZqUjayRvASjglp+GTprSDEzoophyTPIjHzNnuOWd3NfvrODd\/15z8v36TaWDdzBXKkZwasPC2Zp48Iivfzp69AIN3EYGyhpRxuJkbciFSzorOz1UZdLwHoZXR04TGx9qkSHWP5hKl1jeoJ8w1XrurZW3urAz8ZuGMibTtRDdMHvq4bgKNKkbaRNYoqUacMCXisTakPCSYfQV2DEEU6W\/M4Kpv1Sek8XMFumuzv6Lu2waagXRt9n20gwnA2MM9bTs+c+542+4ap1APzEnWNQbhjJCa3cPtd57oIyk+\/IpzXNglY89NAtYPf8v99GAgPTHoa431l7btXnytetWgfCMEtiwQbyZgT2XLqXAOSr\/XHajyeHfHlRUI5Ns6JcOP9dRSvxGOEeEdypPPCQxliS78ltJDBwVxDykBe0k288T5WvW7WuCJOKRRuKj6J7CaR8tT8u8OON1nGnVpQH\/jvwSjkPXeghjbFQeHYbORn4LLp+1bop6AuibjfthvrlqlPQvQRSvu4feEJHaOxbss66ruBeWoGHNMaC3s5uIycDn0U3WLVuUZMm2HB8XdkVdC+BlC9xlyAxlxYN\/XeiqRovrcBDGmOh9Ow2cjLwWXSDVevw4W8w3W1IsOEMN1p+8tX+uIkfz0TuzP9FGBdcz+Du1njoQg9pjAXBnd1GTgY+i264al1a9O\/wPI6hCjd8FN1LIOVr\/HHWj0fA68Z7MpgV5YH\/jjJWkF\/he6kW\/VfLJluQ79lt5GTgc+iGq9YBJS\/ldylat2G8xO\/7Wdjy++56hJunis9sIycDI92UdF+V77jqIT1Pvqtt5GRgpJuS7qqAdlaUr3tIn5Dvq23kZGCkm5LuN604+0Bkp4fM6KJ80zNFusnoonzTM0W6yeiifNMzRbrJ6KJ80zNFusnoonzTM0W6yeiifNMzRbrJ6KJ80zNFusnoonzTM0W6yeiifNMzRbrJ6OYp37rJBDVDuinpZilfBEID5YvIGChfRMZA+SIyRpbyHe++ZziOEemmpJulfHNy7SDdlHRRvumZIt1kdFG+6Zki3WR0Ub7pmSLdZHRRvumZIt1kdFG+6Zki3WR0Ub7pmSLdZHRRvumZIt1kdFG+6Zki3WR0P1e+G0+cy8nASDcl3Qvl65LgbMElFdrK9p6TgZFuSrqXjr4uO84yhtnOq7RTG+YsZKeHzOheLl\/1EHCdJdzlCtcPBlfJhHiYdnSd9sV2ehnZ6SEzulfLF5Lg6Czh6jXMOQ7JhFQicp1hHOWLdHfpXitfnQRHZwmH1yjnuMtAXLu8F6u07zbcUWSnh8zoXj556GyWcHiNco47+eoM4yhfpLtL93L5wuxWZwmXr1HOcStfk2Ec5Yt0d+le73mouc4SDq9RznGb\/91kGEf5It1duhfK1yXB0VnC1WuYc9zmf9cZxv8G6YVWab\/F8IQUPEerOKyHzaxsvNuv4BScLd8TLL1N93Ojbpu09\/bgrfyp0Meij3l4v8+HqziiB96UcEu7bjfenX+WnqK7G3Ciy8W7ZqLiLXytfPXcY1jOani8z0P09kIVR+RbqztZtmU319zCcLY8wr027q3S3Qg4bbV0jrpWW\/hy+YquXN2kYa7Z4zT7YRBIcW\/zKobdXJZH5Ev11MDvNJ9IWMYzJgNXRbMjgh3HeP\/X6JJVgpNgU9wcWeuQARwXbJyehugsLFD\/dvnCLSBERv4DUnQ+uI6S6IG5IOUgOlJTqR45467k1VtFU6aBFAn1pg+wb8o9TcluX4\/I158CS6qvG80U0BZyHmRiOYqJjuoIw0AHe\/r6L9p5LmRB2cOOXBeo2kzHdxlvyZfXwlRprKObcsEmTd5tC0JPIQXZG3mA4ybC3ganYXYWFql\/u3wFmAhmbXUlp5jCR0lEyUWtIijSoNpdZy06DaSY67o5wB4nbPTlANPD8vWkOjfXbDrNrXa+xdodAgy0u5Fb53kjKfVM7ejzRrte7TFel68JOAkbURpsU85b7y0zkEnoyVNoWj2mhNnZfW\/daVgIZy1R\/3b5SkvpyMhIxm6MoyQzM6q\/eSBFOIe0OcAe19B9h8Ah+YbLPEzltaua2g0Qy1FMdFRHGAZOPvrP3AyF8hWuul3Gm5OHdmIuGhlnYtEo9OQp2N7Edre9dd2Yn4VF6t8uXzn3NZGRuoFlFmGUxNmPRvKdBVLiMxS8mejLPtMduqU7Mb5yQwO+M1WoYznAxER1hGEwkS9hjnIoX1PdHuNN+U6NoJuatG\/e4tCTp2B6M5WvoeflOz8LS9Q\/RL7P3iYflS+jvTCREVZIkQRRklpOtR5q5OsfofnngRSh38wB9jih4y6HmO7QbQroP1wbfeWGhsSjZDCq6VgOMDFRHWEY2GCPIIOcPMo5IhNDp3bUBTpSr6vbZbwrX1+lbcq1H1iGTEJPnkIlZQ3zg6gi11t3GuZnYZH6Z8h3RmzPHbgv36GRk6eiBl2oyIgQlbCfITQiuLxJALsXpIJ7BsJ6ebPTS6tOAylwGLyZA\/SbqkJHXw4w3aMLfl9Vla5cZXo2rGWZvItpilbHckbJxHyCbYqBLLJHsILxWh8oS3WBzhutq9tlvEIXrCmM\/WSVxjqmKRNscpZR26LQU0BhqEjREF9ROwa9DU7D7CwsUv8M+S66B7dG5PPiQqnjAVdGYcfdgvfpPl\/lAtyy73dxi3zHmZNwSoLzJV9rSPssst8k32zonmb0q+SrnZHaMxh494zbVfVHefoqWonHCPes4OzjtJBz1gVH+2kG7vf+9e5doHxnUPOBs+heIl\/jjPSeQeNkDdyuxtPHacPUd+3hrMFbu0T7dNaJgPJNS\/cS+ZpbsdgzGPtPrOuvK7iXL3hrl+bAORkY6aake4l8zdViFiMI3K7W9ddUjZev0OvZF2jfbbijyE4PmdG9RL7GGRl5Bm1As48cnT0bpdaJ8bWKcXnJWE4GRrop6V4iX+MhtJ5B72S1btd2NP5WylhBfoXv2uVaLa6QysnASDcl3Xv9vrutL1eWk4GRbkq6nyzfkbXLlszJwEg3Jd1b5bvjdm3XtuZkYKSbku5HBI2fp52+jROZIt1kdFG+6Zki3WR0Ub7pmSLdZHRRvumZIt1kdFG+6Zki3WR0Ub7pmSLdZHRRvumZIt1kdFG+6Zki3WR0Ub7pmSLdZHRRvumZIt1kdM+X7wWdZ3WTCWqGdFPSverRmwhEAqB8ERkD5YvIGChfRMZA+SIyBsoXkTFQvoiMgfJFZAyULyJjoHwzQOLsfpeho+VuLqbngPL9JFRk6emLT+dRfDPb3\/k90BjJWJ78S0T5fhT44vnYP0nDwqf7ezDhckI60ylQvp8F9eRxd4WdPb57O3+i\/bSWJfC6HkxYmQKU77cDnk7ont89e3z3Zv5El7hwIUvgtT0IWSmaKvegnFcQptMV6seV+9SL7zSG+CCYR2+a53fPHt+9lT\/RJy6cZwm8vgeGlUuIaDbVPndit53m+3BjiM8Bif5mj+\/eyp\/oExfOswTe0gPgEtD06Qr1rWW9n9XxUGOIz8FcvtHjuzfyJ\/rEhQvPr7+lByzOBOnTFWrnhE+9+F5jiM8B4ZF8p4\/v3sif6BMXzrME3tMD4BLQ9OkK9ePKferFNxq7tGs3YnKPwN2V600DngrI\/9cEz++ePb57I3+iT1w4zxJ4eQ98PkFFUz1KdGgJ0ST\/oR9X7lIvvo6fIF8OaStJFZqK26zW0rJ7h98YBPgJZ+ct\/AQD1SobOYt\/6WT24QPxydw+Aj\/BQFRPFOJL6Jp8Bz79yk+O0x9H8qyJ2eMnyNf3UfvyRVsQaj3qgsgpWG+c\/nKgloLx7vRaxQ6k8gs5+xjO8LMjTsXPkq\/25Tedcv57j3pTGH+6\/DL2gTsdvqpQgRwJW3GGnx1xKn6EfN3VX\/vyqf7oPerc+tNFIycagTu9sSFbMdTn+NkRp+InyLd0sjPeVCYmjn9q\/OlyXkHr0J0uv6piDu7UM\/zsiFPxE+TbFLByDzzk2pf\/KBkMuN6j3tfGny5nCQMJ3OmjWnggP8DTk\/oz\/Own4u6lkZ9A9yfIF27LiBqCtS+fy1uwpjAedbgfe3BhnP4P2M+70+ErBAFg1lvUZ\/jZz8NsJc6N\/unb6P4I+X4nlk7dB4\/ISeiifDPBOFuCPj11nIcLxH8GXZTv50Iv6ta+6mCh+mwRe0Ur8RjBcQILxDktBCtuuMO8gy7K92OhF3UH687NQvX5InZOG6b\/UUMtW68Fv2GKfgtdlO\/HwtzbTNadk\/DNeqK7gns9jGTsbpgD30IX5fuxMOscJ+vOzcdoEbtoYDmd1YPQnusfQRfl+7HQi7qn68711ThexN6zUYoH9ADRFTHSO+7gbqGL8v1YcL2o26479wvVZ4vYKWMF+RW+wwJxIao7QoO30EX55ofdc\/Y5sZXUdFG++WH7nI2s\/azMVinponyzw84i9vbD1nQmpYvyRWQMlC8iY6B8ERkD5YvIGChfRMZA+SIyBsoXkTFQvoiMgfJFZAyULyJjoHwRGQPli8gYKF9ExkD5IjIGyheRMVC+iIyB8kVkDJQvIjcE\/9mJ8kVkBvZvf\/rtr0bCKF9EZmDwqKk\/agnnKd8G8XPxP8Tgn\/7lf\/OUL+IHQ42+5F+L\/\/x9zHT0RfxgMCNdAMoXkRnG331uPZTvvXgly2FHywNPBfvgPAHnAeV7I1pKSPn0I5JGMpb72pylQvnw1C2vAeV7I0BQz+d3GI6ds8W9vm1IRvneCFqLaabwxenEJE340jnbT4XyWalbTgLK90Z0pALx6mwmXKeC6EhfNybNiU4BrtKEC7dfRYh+kPlzqVCEyYVyW+aWJED53om+IA0PspmYh5GbNCcuZ7hJ5Wn2M+fs2VQo9mn8N2VuSQOU772oQZ9RNhOTCkK4TCYqL7gQLuuJOWfPpkKx8r0pc0saoHzvA4NZakPmsjNpTmwmk1bNkSOJun2Op0IJMlF9zeCL8r0RDJxmQ+WymcAUVp8RnebEZDIZra\/B7GfO2bOpUFwulHsyt6QByvc+8FLeStXcZDOBvCWsYOph5CbNic0ZXupJgN5vaIk6aU+mQqmYy4VyS+aWNED5fg5eicAtILPMLWn7isgNmWVuSdhXRH7ILHPLe\/h\/+nLzg0xNYIAAAAA8dEVYdGNvbW1lbnQAIEltYWdlIGdlbmVyYXRlZCBieSBHTlUgR2hvc3RzY3JpcHQgKGRldmljZT1wbm1yYXcpCvqLFvMAAAAASUVORK5CYII=\" \/><\/div><\/div>\n<p>\n  We've touched on some of these varieties of complexity earlier\nin this book, especially the accidental ones. In\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  we saw that accidental interface\ncomplexity often comes from non-orthogonality in the interface design\n\u2014 that is, failing to carefully factor the interface operations\nso that each does exactly one thing.  Accidental code complexity\n(making code more complicated than it needs to be to get the job done)\noften results from premature optimization. Accidental codebase bloat\noften results from violating the SPOT rule, duplicating code or\norganizing it poorly so that opportunities for reuse aren't\nrecognized.\n<\/p>\n<p>\n  Essential interface complexity usually can't be cut without\ntrimming the basic functional requirements for the software (a theme\nwe'll develop further in this chapter's case studies).  Essential\ncodebase size is related to choice of development tools because, if\nthe feature list is held constant, the most important factor in\ncodebase size is probably the choice of implementation language (as we\nimplied in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>).\n<\/p>\n<p>\n  Sources of optional complexity are the most difficult to make\nuseful generalizations about, because they so often depend on\ndelicate judgments about which features it is worth paying the\ncomplexity cost for.  Optional interface complexity often comes from\nadding convenience features that make life easier for users but aren't\nessential to the function of the program.  Optional increases in\ncodebase size (supposing the user-visible features and the algorithms\nused are held constant) can often come from various sorts of practices\nintended to make it more maintainable \u2014 adding mode comments,\nusing long variable names, and so forth. Optional implementation\ncomplexity tends to be driven by\n  <em>everything<\/em>\n  that\ntouches a project.\n<\/p>\n<p>\n  The sources of complexity have to be grappled with in different\nways.  Codebase size can be attacked with better tools.\nImplementation complexity can be addressed with better choice of\nalgorithms.  Interface complexity has to be addressed with better\ninteraction design, a skill involving considerations of ergonomics and\nuser psychology.  This skill is less common (and possibly more\ndifficult) than writing code.\n<\/p>\n<p>\n  Attacking the kinds of complexity, on the other hand, has to be\ndone more with insight than with methods.  You cut accidental\ncomplexity by noticing that there is a simpler way to do things.\nYou cut optional complexity by making context-dependent judgments\nabout what features are worthwhile.  You can only cut essential \ncomplexity by having an epiphany, fundamentally redefining the\nproblem you are addressing.\n<\/p>\n<h4 id=\"id2963237\"><a href=\"#id2963237\">\u00a7<\/a>When Simplicity Is Not Enough<\/h4>\n<p>\n  The failure mode that goes with the Unix tradition's insistence\non simplicity is that Unix programmers often talk (and sometimes even\nbehave) as though all optional complexity is accidental.  More than\nthis, there is a strong bias in the Unix tradition toward removing\nfeatures rather than accepting optional complexity.\n<\/p>\n<p>\n  The case for this attitude is easy to make (indeed, we spend\nmuch of this book making it).  Clean minimalism makes us feel virtuous\non many levels, and designing for it is a valuable counter to\nthe natural tendency of software systems to develop\never-more-elaborate encrustations of ill-considered features.  But\ncomputing resources and human thinking time, like wealth, find their\njustification not in being hoarded but in being spent.  As with other\nforms of asceticism, one has to ask when design minimalism stops being\na valuable form of self-discipline and starts being a mere hair shirt\n\u2014 a way to indulge those feelings of virtue at the expense of\nactually\n  <em>using<\/em>\n  that wealth to get work\ndone.\n<\/p>\n<p>\n  This is a perilous question, all too easily turned into an\nargument for abandoning good design discipline altogether.  Unix old\nhands often shy away from it, fearing that failing to hold the hardest\npossible line against complexity and bloat will lead us inexorably to\ndamnation.  But it's also a\n  <em>necessary<\/em>\n  question.\nWe'll tackle it directly when analyzing this chapter's case\nstudies.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2964826\" id=\"ftn.id2964826\">[114]<\/a><\/sup>\n  The terms\nwe have invented for these design traps, unlikely as they may sound,\ncome from established hacker jargon described in <sup><a href=\"#Raymond96\" title=\"[Raymond96]\">[Raymond96]<\/a><\/sup>.\n<\/p><p><sup><a href=\"#id2961828\" id=\"ftn.id2961828\">[115]<\/a><\/sup>\n  The\ndistinction between accidental and optional complexity means that the\ncategories we're discussing here are\n  \n   <em>not<\/em>\n  \n  the same\nas essence and accident in Fred Brooks's essay\n  <em>No Silver Bullet<\/em><sup><a href=\"#Brooks\" title=\"[Brooks]\">[Brooks]<\/a><\/sup>, but they have common\nancestry in philosophy.\n<\/p><\/div>\n\n<h3 id=\"ch13s02\"><a href=\"#ch13s02\">\u00a7<\/a>A Tale of Five Editors<\/h3>\n<p>\n  Now we're going to use five different Unix editors as case \nstudies.  It will be helpful to bear in mind a set of benchmark\ntasks as we examine these designs:\n<\/p>\n <div><ul><li>\n    <p>\n     <em>Plain-text editing.<\/em>\n     Manipulating\nplain ASCII (or, in this internationalized age, perhaps Unicode) files\nwith no structure known to the editor above byte level, or perhaps\nline level.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Rich-text editing.<\/em>\n     Editing of text\nwith attributes; these might include font changes, color, or other sorts of\nproperties of text spans (such as being a hyperlink).  Editors that can\ndo this have to be able to translate between some presentation of the\nattributes in the user interface and some on-disk representation of\nthe data (such as HTML, XML, or other rich-text formats.)\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Syntax awareness.<\/em>\n     An editor that\nis syntax-aware knows that input events have a grammar, and does things like\nautomatically changing the indent level when it recognizes the\nbeginning or end of a block scope in a programming language.  Editors\nthat are syntax-aware also commonly highlight syntax with colors\nor distinguished fonts.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Output parsing<\/em>\n     of batch command\noutput. The commonest case of this in the Unix world is running a C\ncompilation from inside the editor, trapping the error messages, and\nthen being able to step through the error locations without leaving the\neditor.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Interaction<\/em>\n     with helper\nsubprocesses that persist and maintain state between editor commands.\nThis capability, when present, has powerful consequences:\n    <\/p>\n    <div><ul><li>\n       <p>\n        It's possible to drive a version-control system\n    from inside the editor, performing file checkins and checkouts\n    without dropping out to a shell window or separate\n    utility.\n       <\/p>\n      <\/li>\n      <li>\n       <p>\n        It's possible to front-end a symbolic debugger\n    inside the editor, such that (for example) when the run stops on a\n    breakpoint the appropriate file and line is automatically\n    visited.\n       <\/p>\n      <\/li>\n      <li>\n       <p>\n        It's possible to edit remote files within the\n    editor, by having it recognize when a filename refers to another\n    host (recognizing some syntax like\n        <code>\/user@host:\/path\/to-file<\/code>).  Provided you have\n    the right access, such an editor can automatically run a\n    utility like\n    scp(1)\n    or\n    ftp(1)\n    to fetch a local copy, then automatically copy the edited version\n    back to the remote location at file-save time.\n       <\/p>\n      <\/li>\n     <\/ul><\/div>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  All our case studies can edit plain text.  (The reader should\nnot take this capability for granted \u2014 there are many things\ncalled editors, such as \u2018word processors\u2019 that are too\nspecialized to do this!) We begin seeing variable degrees of optional\ncomplexity in how they handle the more complex tasks.\n<\/p>\n<h4 id=\"id2963445\"><a href=\"#id2963445\">\u00a7<\/a>ed<\/h4>\n<p>\n  ed(1)\nis the truly Unix-minimalist way of plain-text editing.  It dates from\nthe days of teletypes.<sup><a href=\"#ftn.id2963466\" id=\"id2963466\">[116]<\/a><\/sup>\n  It has a simple, austere CLI, and there is\nno screen display.  In the following listing, computer output is\n  <em>emphasized<\/em>.\n<\/p>\n <pre class=\"programlisting\">\ned sample.txt\n<em>sample.txt: No such file or directory<\/em>\n# This is a comment line, not a command.\n# The message above warns that the sample.txt file is newly created.\na\n<em>the quick brown fox\njumped over the lazy dog<\/em>\n.\n# That was an append command, which added text to the file.  \n# The dot on a line by itself terminated the append.\n1s\/f[a-z]x\/dragon\/\n# On line 1, replace the first substring matching an f followed by a\n# lowercase alphabetic followed by x with \u2018dragon\u2019.  The\n# substitute command accepts basic regular expressions.\n1,$p\nthe quick brown dragon\njumped over the lazy dog\n# Print all lines from 1 to the last.\nw\n<em>51<\/em>\n# That wrote the file to disk. The \u2018q\u2019 command ends the\n# editing session.\nq\n<\/pre>\n<p>\n  Unbelievable as it may seem to a modern reader, most of Unix's\noriginal code was written with this editor.  The reader with DOS\nexperience may recognize here the original on which\n  <em>EDLIN<\/em> was (crudely) modeled.\n<\/p>\n<p>\n  If one defines the job of an editor simply as enabling the user\nto create and modify plain text files,\ned(1)\nis entirely sufficient for the job.  Importantly to the Unix view of\ndesign correctness, it does nothing else.  Many old-school Unix\nprogrammers half-seriously maintain that all editors with more\nfeatures than ed has are simply bloated \u2014 and a few still who\nseriously believe this.\n<\/p>\n<p>\n  Appropriately,\n  <em>ed<\/em> was Ken Thompson's\ndeliberate simplification of the earlier\n  <em>qed<\/em> <sup><a href=\"#RitchieQED\" title=\"[RitchieQED]\">[RitchieQED]<\/a><\/sup> editor\n\u2014 which was very similar (and the first editor to use regular\nexpressions in the characteristic Unix way) but had multiple-buffer\ncapability that Ken deliberately discarded.  He judged it not worth\nthe additional complexity.\n<\/p>\n<p>\n  A notable characteristic of ed(1) and all its descendants is the\nobject-operation format of its commands (the session example shows an\nexplicit range on the \u2018p\u2019 command).  There is a relatively\npowerful syntax for specifying line ranges, either numerically, or by\nregular-expression pattern match, or by special shorthands for the\ncurrent and last line.  Most editor operations can be applied to any\nrange.  This is a good example of orthogonality.\n<\/p>\n<p>\n  Nowadays,\ned(1)\nis primarily used as a program-driven editing tool in scripts \u2014\na role to which editors with more elaborate modes of interactivity are\nunsuited.  There is a close variant called\nex(1)\nwhich adds a few useful interactivity features such as command\nprompts; it is occasionally useful in rare cases when editing must be\ndone over a slow serial line, or in certain unusual crash-recovery\nsituations where the library support needed to run other editors is\nnot accessible.  For these reasons, every Unix includes an\n  <em>ed<\/em> implementation and most include\n  <em>ex<\/em> as well.\n<\/p>\n<p>\n  The\nsed(1)\nstream editor mentioned in\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>\n  is also closely related to ed; many\nof the basic commands are the same, though designed to be invoked\nthrough command-line switches rather than from standard input.\n<\/p>\n<p>\n  Almost all Unix programmers have strayed from the path of\nausterity and minimalist virtue enough to normally use editors that at\nleast present a roguelike, screen-oriented interface.  However, the\nfact that the religion of ed\npersists<sup><a href=\"#ftn.id2963675\" id=\"id2963675\">[117]<\/a><\/sup>\n  says a great deal that is worth\nnoting about the Unix mindset.\n<\/p>\n<h4 id=\"vi_complexity\"><a href=\"#vi_complexity\">\u00a7<\/a>vi<\/h4>\n<p>\n  The original\nvi(1)\neditor was the first attempt to bolt a visual, roguelike interface\nonto the command set of\ned(1).\nLike ed, its commands are generally single\nkeystrokes, and it is particularly well suited to use by\ntouch-typists.\n<\/p>\n<p>\n  The original vi didn't have mouse\nsupport, editing menus, macros, assignable key bindings, or any form\nof user customization. In line with the religion of ed, vi's partisans\nconsidered the lack of these features a virtue.  On this view, one of\nvi's most important virtues is that you can start editing immediately\non a new Unix system without having to carry along your customizations\nor worrying that the default command bindings will be dangerously\ndifferent from what you're used to.\n<\/p>\n<p>\n  One characteristic of vi that\nbeginners tend to find frustrating is a result of its terse\nsingle-keystroke commands. It has a\n  <em>moded<\/em> interface \u2014 you are either in command mode or in text-insertion\nmode. In text-insertion mode, the only commands that work are the\nESC key for mode exit and (on newer versions) the cursor-movement\nkeys. In command mode, typing text will be interpreted as commands and\ndo odd (and probably destructive) things to your content.\n<\/p>\n<p>\n  On the other hand, one property of the command set that\nvi fans particularly tout is the\nobject-operation format it inherited from\ned. Most of the extended commands also\noperate in a natural way on any line range.\n<\/p>\n<p>\n  Over the years, vi has bulked up\nconsiderably.  Modern versions add mouse support, editing menus,\nunlimited undo (the original vi could only\nundo the last command), multiple files in separate buffers, and\ncustomization with a run-control file.  However, the use of\nrun-control files is still unusual, and in contrast to\nEmacs, the use of embedded general-purpose\nscripting has never caught on.  Instead, vi\nimplementations have grown individual capabilities to do things, like\nsyntax awareness of C code and output parsing of C compiler error\nmessages, by adding C code to vi itself.\nSubprocess interaction is not supported.\n<\/p>\n<h4 id=\"id2963798\"><a href=\"#id2963798\">\u00a7<\/a>Sam<\/h4>\n<p>\n  The\n  <em>Sam<\/em> editor<sup><a href=\"#ftn.id2963818\" id=\"id2963818\">[118]<\/a><\/sup>\n  was written by Rob Pike at Bell Labs in the mid-1980s.  Sam was\ndesigned for the Plan 9 operating system, which we'll survey in\n  <a href=\"#futurechapter\" title=\"Chapter 20. Futures\">\n   Chapter 20<\/a>.  While the Sam editor is not widely known\noutside the Labs, it's favored by many of the original Unix developers\nwho went on to work on Plan 9, including Ken Thompson himself.\n<\/p>\n<p>\n  Sam is a fairly straightforward descendant of ed, remaining much\ncloser to its parent than vi.  Sam incorporates only two new concepts:\na curses-style text display and text selection with the mouse.\n<\/p>\n<p>\n  Each Sam session has exactly one command window, and one or more\ntext windows.  Text windows edit text, and command windows accept\ned-style editing commands.  The mouse is used to move between windows,\nand to select text regions within text windows.  This is a clean,\northogonal, modeless design that discards most of the interface\ncomplexity of vi.\n<\/p>\n<p>\n  Most commands operate by default on a select region that can be\npainted with a mouse drag operation.  The select region for a command\ncan also be set by specifying a line range in the fashion of ed, but\nSam gains considerable power from the fact that the user can select at\nfiner granularity than a line range.  Because the mouse is available\nto do selections and rapidly change focus between buffers (including\nthe command buffer), Sam needs no equivalent of the default (command)\nmode of vi.  The hundreds of extended vi commands are unnecessary and,\ntherefore, omitted. Overall, Sam adds only about a dozen commands to\nthe seventeen or so in the ed set, for a total of about thirty.\n<\/p>\n<p>\n  Four of the new commands in Sam join two inherited from\ned(1)\nand\nvi(1),\nas ways to apply regular expressions to the task of selecting files\nand file regions to operate on.  These provide limited but effective\nloop and conditional facilities to the command language.  There is,\nhowever, no way to name or parameterize command-language\nprocedures. Nor can the language do interactive control of a\nsubprocess.\n<\/p>\n<p>\n  An interesting feature of Sam is that it's split into two parts.\nseparating a back end that manipulates files and does searches from a\nfront end that handles the screen interface.  This instance of the \n\u201cseparated engine and interface\u201d chapter has the \nimmediate practical benefit that, though the program has a GUI, it\ncan run easily over a low-bandwidth connection to edit files on a\nremote server.  Also, the front and back ends can be retargeted\nrelatively easily.\n<\/p>\n<p>\n  Sam, like recent versions of vi, has infinite undo.  By design,\nit supports neither rich-text editing, nor output parsing, nor\nsubprocess interaction.\n<\/p>\n<h4 id=\"emacs_editing\"><a href=\"#emacs_editing\">\u00a7<\/a>Emacs<\/h4>\n<p>\n  <em>Emacs<\/em> is undoubtedly the most\npowerful programmer's editor in existence. It's a big, feature-laden\nprogram with a great deal of flexibility and customizability. As we\nobserved in the\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  section on Emacs Lisp, Emacs has an entire programming language inside\nit that can be used to write arbitrarily powerful editor\nfunctions.\n<\/p>\n<p>\n  Unlike\nvi,\nEmacs doesn't have interface modes;\ninstead, commands are normally control characters or prefixed with an\nESC. However, in Emacs it is possible to\nbind just about any key sequence to any command, and commands can be\nstock or customized Lisp programs.\n<\/p>\n<p>\n  Emacs can edit multiple files, each\nin a separate buffer, and supports moving text among the buffers.\nVersions running under X have native mouse support.\n<\/p>\n<p>\n  The Lisp programs bound to Emacs keystrokes can perform\narbitrary text transformations on a buffer.  This capability is\nheavily used, among other things to define syntax-aware and rich-text\nediting modes for dozens of different languages and markup formats\n(beginning with support and color highlighting of C code as in\nvi, but going way beyond that).  Each mode\nis simply a library file of Lisp code that is loaded on demand.\n<\/p>\n<p>\n  Emacs Lisp programs can also interactively control arbitrary\nsubprocesses.  Some notable consequences of this capability were\nlisted earlier, including the ability to serve as a front end for\nversion-control systems, debuggers, and the like.\n<\/p>\n<p>\n  The designers of Emacs<sup><a href=\"#ftn.id2967056\" id=\"id2967056\">[119]<\/a><\/sup>\n  built a programmable editor\nthat could have task-related intelligence customized into it for\nhundreds of different specialized editing jobs.  They then gave it the\nability to drive other tools. As a result, Emacs supports dealing with\nall things textual in one shared context \u2014 files, mail, news,\ndebugger symbols.  It can serve as a customizable front end to any\ncommand with an interactive textual interface.\n<\/p>\n<p>\n  It is a common joke, both among fans and detractors of\nEmacs, to describe it as an operating\nsystem masquerading as an editor.  That overstates the case, but\nEmacs certainly does fulfill the role\noccupied by integrated development environments (IDEs) under non-Unix\noperating systems (a theme to which we shall return in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>).\n<\/p>\n<p>\n  This power comes at a price in complexity. To use a customized\nEmacs you have to carry around the Lisp\nfiles that define your personal Emacs\npreferences. Learning how to customize\nEmacs is an entire art in\nitself. Emacs is correspondingly harder to\nlearn than vi.\n<\/p>\n<h4 id=\"id2967110\"><a href=\"#id2967110\">\u00a7<\/a>Wily<\/h4>\n<p>\n  The\n  <em>wily<\/em> editor<sup><a href=\"#ftn.id2967125\" id=\"id2967125\">[120]<\/a><\/sup>\n  is a clone of the Plan 9\n  editor\n  <em>acme<\/em>.<sup><a href=\"#ftn.id2967150\" id=\"id2967150\">[121]<\/a><\/sup>\n  It shares some facilities with Sam, but is intended to provide a\nfundamentally different user experience.  Although Wily probably sees\nthe least widespread use of any of these editors, it is interesting\nbecause it illustrates a different and arguably more Unixy way of\nimplementing an Emacs-like programmable editor.\n<\/p>\n<p>\n  Wily could be described as a minimalist IDE, an implementation\nof Emacs-style extensibility without the decades of accompanying\ncruft.  In Wily, even global search and replace, that\n  \n   <em>sine qua non<\/em> \n  of Unix editors, is\nsupplied by an external program.  The built-in commands relate almost\nexclusively to windowing operations.  Wily is designed from the ground\nup to use the mouse as much, and as well, as possible.\n<\/p>\n<p>\n  Wily attempts to replace not only conventional editors but\nconventional terminal windows such as\nxterm(1)\nas well.  In Wily, any piece of text within the main window (which\ncontains multiple non-overlapping Wily windows) can be an action or a\nsearch expression.  The left mouse button is used to select text, the\nmiddle button to execute text as a command (either built-in or\nexternal), and the right button to search either Wily's buffers or the\nfile system for text.  No permanent or popup menus are\nrequired.\n<\/p>\n<p>\n  In Wily, the keyboard is used\n  <em>only<\/em>\n  to enter\ntext.  Shortcuts are achieved not by special use of the keyboard, but\nby holding down more than one mouse button at the same time.  These\nshortcuts are always equivalent to using the middle button on some\nbuilt-in command.\n<\/p>\n<p>\n  Wily can also be used as the front\nend for C, Python, or Perl programs, reporting to them whenever a\nwindow is changed or an execute or search command is performed with\nthe mouse.  These plugins function analogously to Emacs modes, but\ndon't run in the same address space with Wily; instead, they\ncommunicate with it via a very simple set of remote procedure calls.\nWily comes packaged with an\n  <em>xterm<\/em> analog\nand a mail tool which uses it as the editing front end.\n<\/p>\n<p>\n  Because Wily depends on the mouse so heavily, it cannot be used\non a character-cell-only console display; nor can it be used over a\nremote link without X forwarding.  As an editor, Wily is designed for\nediting plain text; it has only two fonts (one proportional and one\nfixed-width) and has no mechanism that could support rich-text editing\nor syntax awareness.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2963466\" id=\"ftn.id2963466\">[116]<\/a><\/sup>\n  Younger readers may not be aware\nthat terminals used to print.  On paper.  Very\nslowly.\n<\/p><p><sup><a href=\"#id2963675\" id=\"ftn.id2963675\">[117]<\/a><\/sup>\n  The religion of ed\nis exemplified by a famous Usenet posting which the reader may be able\nto find with a Web search for \u201cEd is the standard\neditor\u201d.  While it is clearly intended as parody, it is by no\nmeans clear that the author was entirely joking.  Most Unix hackers\nwould read it as an example of \u201cHa ha, only\nserious\u201d.\n<\/p><p><sup><a href=\"#id2963818\" id=\"ftn.id2963818\">[118]<\/a><\/sup><a href=\"http:\/\/plan9.bell-labs.com\/sys\/doc\/sam\/sam.html\">\n   http:\/\/plan9.bell-labs.com\/sys\/doc\/sam\/sam.html<\/a>\n<\/p><p><sup><a href=\"#id2967056\" id=\"ftn.id2967056\">[119]<\/a><\/sup>\n  The designers of Emacs were\nRichard M. Stallman, Bernie Greenberg, and Richard M. Stallman.  The\noriginal Emacs was Stallman's invention, the first version with an\nembedded Lisp was Greenberg's, and the now-definitive version is\nStallman's derived from Greenberg's.  No complete account of the\ndesign history has been written in 2003, but Greenberg's\n  <em>Multics Emacs: The History, Design, and Implementation<\/em>\n  is illuminating and readily discoverable\nvia keyword search on the Web.\n<\/p><\/div>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2967125\" id=\"ftn.id2967125\">[120]<\/a><\/sup><a href=\"http:\/\/www.cs.yorku.ca\/~oz\/wily\">\n   http:\/\/www.cs.yorku.ca\/~oz\/wily<\/a>\n<\/p><p><sup><a href=\"#id2967150\" id=\"ftn.id2967150\">[121]<\/a><\/sup><a href=\"http:\/\/plan9.bell-labs.com\/sys\/doc\/acme\/acme.html\">\n   http:\/\/plan9.bell-labs.com\/sys\/doc\/acme\/acme.html<\/a>\n<\/p><\/div>\n\n<h3 id=\"ch13s03\"><a href=\"#ch13s03\">\u00a7<\/a>The Right Size for an Editor<\/h3>\n<p>\n  Now let us examine our case studies using the complexity\ncategories we developed at the beginning of this chapter.\n<\/p>\n<h4 id=\"id2967267\"><a href=\"#id2967267\">\u00a7<\/a>Identifying the Complexity Problems<\/h4>\n<p>\n  Every text editor has a certain amount of essential complexity.\nAt minimum, it has to maintain an internal buffer copy of the file or\nfiles the user is editing. Functions to import and export file data\nare a minimum requirement (usually from and to disk, though the stream\neditor\nsed(1)\nis an interesting exception).  Some way to modify the buffer must be\nsupported, though we cannot specify what way without describing\nspecific features that are optional.  Our four examples show widely\nvarying levels of optional and accidental complexity beyond\nthis.\n<\/p>\n<p>\n  Of all of these,\ned(1)\nhas the least complexity.  Almost the only non-orthogonal feature in\nits command set is the fact that many of its commands can take a\n\u2018p\u2019 or \u2018l\u2019 suffix to print or list command\nresults.  Even after three decades of feature additions there are\nfewer than thirty editing commands, and the normal working set for\nmost users will be less than a dozen.  There is not much in the way of\noptional complexity that could be removed here, and it's hard to\nidentify any accidental complexity at all.  The user interface of\ned is strictly compact.\n<\/p>\n<p>\n  On the flip side, the ed interface is\nnot really suitable for editing tasks even as basic as rapidly\nflipping through a text file.  One has to limit one's objectives\npretty sharply for ed to become an\nacceptable solution for interactive editing.\n<\/p>\n<p>\n  Suppose, then, that we add \u201csupport visual browsing and\nediting of multiple files\u201d as an objective?  Then\nSam seems not very far from being the\nminimal ed extension that could achieve\nthis.  The fact that the designers did not change the semantics of the\ninherited ed commands is notable; they kept\nan existing, orthogonal set and added a relatively small set of\ncapabilities that are themselves orthogonal.\n<\/p>\n<p>\n  One large increase in optional (implementation) complexity is\nSam's infinite-undo capability.  Another\nsignificant one is the new regular-expression-based loop and iteration\nfacility in the command language.  These, and the fact that the mouse\ncan be used as a selection device, are about all that distinguish\nSam from a hypothetical\ned with a mouse-and-windows\ninterface.\n<\/p>\n<p>\n  Without a thorough code audit it's difficult to be sure, but at\nthe design level it's hard to identify any accidental complexity in\nSam.  The interface is at least semi-compact and arguably strictly compact.\nThis editor lives up to the very highest standards of Unix design\n\u2014 unsurprisingly, given its provenance.\n<\/p>\n<p>\n  By contrast, vi looks rather bloated\nand flabby.  There are hundreds of commands, many of them\nduplicative. These are at best optional complexity, and perhaps\naccidental. At a guess, most users don't know more than 5% of the\ncommand set.  With the example of Sam\nbefore us, it's fair to wonder why the interface complexity of\nvi is so high.\n<\/p>\n<p>\n  In\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>\n  we\ndescribed the effect of the absence of standard arrow keys on early\nroguelike programs; vi was one of these.  When vi was built, its\nauthor knew that many of his users would need to be able to use the\ncursor motion keys traditional on Unix glass teletypes.  This made a\nmodal interface inevitable.  Once the hjkl keys had mode-dependent\nmeanings in an edit buffer, it was all too easy to fall into the habit\nof adding new commands in an ad-hoc way.\n<\/p>\n<p>\n  Sam, designed as it is to depend on a\nbitmapped display with both arrow keys and a mouse, can be much\ncleaner.  And it is.\n<\/p>\n<p>\n  But the clutter of vi commands is a relatively superficial\nproblem.  It's interface complexity, yes, but of a kind most users can\nand do ignore (the interface is semi-compact in the sense we developed\nin\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>).  The deeper problem is an\nadhocity trap. Over the years, vi has had progressively more and more\nspecial-purpose C code bolted onto it to perform tasks that Sam\nrefuses to do and that Emacs would attack with Lisp code modules and\nsubprocess control. The extensions are not, as in Emacs, libraries\nloaded as needed; users pay the overhead for the resulting code bloat\nall the time.  As a result, the size difference between a modern vi\nand a modern Emacs is not nearly as great as one might expect; in\nmid-2003 on an Intel-architecture machine, it's 1500KB for GNU Emacs\nversus 900KB for vim.  There is a whole lot of both optional and\naccidental complexity in that 900KB.\n<\/p>\n<p>\n  For vi partisans, not having an embedded scripting language\n\u2014 not being Emacs \u2014 has become an identity issue, a central\npart of the shared myth that vi is a lightweight editor.  While vi\nfans like to talk about filtering buffers with external programs and\nscripts to do what Emacs's embedded scripting does, the reality is\nthat vi's \u201c!\u201d command cannot filter regions of an edit\nbuffer selected at finer granularity than a range of lines (Sam and\nWily, though they have no more subprocess management than vi does, can\nat least filter arbitrary text ranges, not just line ranges).  All\nknowledge of file formats and syntaxes that vary at a finer\ngranularity (and most do) has to be built in to C code if vi is going\nto have it available at all.  There is thus little prospect that the\ncodebase-size ratio between Emacs and vi will improve in favor of vi;\nindeed, it seems likely to get worse.\n<\/p>\n<p>\n  Emacs is sufficiently large, and has\na sufficiently tangled history, to make separating its optional from\nits accidental complexity quite a challenge.  We can at least begin by\ntrying to separate the dispensable accidents of the\nEmacs design from its indispensable\nessentials.\n<\/p>\n<p>\n  Perhaps the most conspicuously dispensable part of the\nEmacs design is Emacs\nLisp.  It\nis essential to what Emacs does that it\nfeatures what we nowadays call an embedded scripting language, but\nEmacs would be little different in\ncapability if that language had been Python or Java or Perl.  At the\ntime Emacs was designed in the 1970s,\nhowever, Lisp was about the only language that had the characteristics\n(including unlimited-extent types and garbage collection) to fit it to\nthe job.\n<\/p>\n<p>\n  Much in the particulars of the way\n  <em>emacs<\/em> handles event processing and drives a\nbitmapped display (including the support for internationalization) is\naccidental as well.  The one great schism in its history (the GNU\nEmacs\/XEmacs fork) was over these issues, and demonstrates that\nnothing in the rest of the design prefers or requires any one event\nmodel.\n<\/p>\n<p>\n  On the other hand, the ability to bind arbitrary event\nsequences to arbitrary built-in or user-defined functions is\nindispensable.  The scripting language could change and the event model\ncould change, but without the anything-goes polymorphism in the way\nthey are connected, the Emacs design would\nbe both unrecognizable and crippled.  Extension modes would have\nto fight each other for ownership of a limited event set, and\nactivating multiple cooperating modes on the same buffer would\nbe difficult or impossible.\n<\/p>\n<p>\n  The huge library of extension modes shipped with\nEmacs is accidental as well.  The\n  <em>ability<\/em>\n  to construct such extensions may be\nessential, but the particular set we have is a product of history and\nchance. They could all be different or replaced; the result would\nstill, recognizably, be Emacs.\n<\/p>\n<p>\n  But subprocess interaction is indispensable.  Without it,\nEmacs modes could not perform the expected\nIDE-like integration and front-ending of many different tools.\n<\/p>\n<p>\n  Experience with small editors that clone the default keybindings\nand appearance of Emacs without emulating\nits extensibility is instructive.  There have been several such\nclones, of which the best known are probably\n  <em>MicroEmacs<\/em> and\n  <em>pico<\/em>, but none have ever acquired\nsignificant mindshare.\n<\/p>\n<p>\n  Having identified accident and essence in the\nEmacs design helps us get a handle on which\nof its complexity is optional and which accidental. But, more\nimportantly, they help us see past the superficial differences between\nEmacs and the previous three editors we\nhave considered, to the really critical difference: the fact that the\nobjectives of the Emacs design are far more\nbroad.  Emacs wants to be a unified\ninterface to all tools that operate on text.\n<\/p>\n<p>\n  Wily makes an interesting contrast with Emacs.  As with Sam, the\namount of optional complexity is low; the Wily user interface can be\nsuccinctly but effectively described in a single page.\n<\/p>\n<p>\n  But this elegance comes with a price; it is not possible to bind\nfunctions to any keystrokes or input gestures other than a restricted\nset of mouse chords.  Instead, every editor function other than very\nbasic text insertion and deletion has to be implemented with a program\noutboard of the editor, either a standalone script or a specialized\nsymbiont process listening to Wily input events.  (The former\ntechnique relies on outboard program startups being fast enough not to\nproduce noticeable interface lag, something which was emphatically not\nthe case in either Emacs's natal environment or under the Unixes it\nwas first ported to.)\n<\/p>\n<p>\n  Optional complexity which Emacs would\nimplement in Lisp extension modes is instead distributed through\nspecialized symbionts; each has to know the special\nWily messaging interface. An advantage of\nthis approach is that such symbionts can be written in any language\nthe user chooses. In addition, the symbionts (because they run\noutboard) cannot adversely affect each other or the\nWily core (which is not true of\nEmacs modes). A disadvantage is that\nWily itself cannot directly do subprocess\ninteraction with ordinary Unix tools at all.\n<\/p>\n<p>\n  In this and other ways,\n  <em>wily's<\/em> distributed scripting is not as powerful as the embedded scripting of\nEmacs.  The scope of Wily's objectives is correspondingly narrower;\nthe authors disclaim any interest in syntax-aware editing, or rich\ntext, for example, and neither Wily nor its Plan 9 ancestor\n  <em>acme<\/em> can do these things.\n<\/p>\n<p>\n  This brings us to another, and sharper way of posing the central\nquestion of this chapter: When do large objectives justify a large\nprogram?\n<\/p>\n<h4 id=\"id2967642\"><a href=\"#id2967642\">\u00a7<\/a>Compromise Doesn't Work<\/h4>\n<p>\n  The comparison between Sam and vi suggests strongly that, at\nleast where editors are concerned, attempts to compromise between the\nminimalism of ed and the all-singing-all-dancing comprehensiveness of\nEmacs don't work very well; vi attempts this, and ends up with neither\nvirtue.  Instead, it falls into an adhocity trap. Wily avoids the\nadhocity trap, but cannot match the power of Emacs and must demand a\ncustom process interface from each of its interactive symbionts in\norder to come anywhere close.\n<\/p>\n<p>\n  Evidently something about editors tends to push them in the\ndirection of increasing complexity. In the case of\nvi, that something is not hard to identify;\nit's the desire for convenience.  While ed\nmay be theoretically adequate, very few people (other than perhaps Ken\nThompson himself) would forgo screen-oriented editing to make a\nstatement about software bloat.\n<\/p>\n<p>\n  More generally, programs that mediate between the user and the\nrest of the universe notoriously attract features.  This includes not\njust editors but Web browsers, mail and newsgroup readers, and other\ncommunications programs.  All tend to evolve in accordance with the\nLaw of Software Envelopment, aka Zawinski's Law: \u201cEvery program\nattempts to expand until it can read mail. Those programs which cannot\nso expand are replaced by ones which can\u201d.\n<\/p>\n<p>\n  Jamie Zawinski, inventor of the Law (and one of the principal\nauthors of the Netscape and Mozilla Web browsers), maintains more\ngenerally that all really useful programs tend to turn into Swiss Army\nknives. The commercial success of large, integrated application suites\noutside the Unix world tends to confirm this, and directly challenges\nthe Unix philosophy of minimalism.\n<\/p>\n<p>\n  To the extent Zawinski's Law is correct, it suggests that some\nthings want to be small and some want to be large, but the middle\nground is unstable.  The superficial problems with\nvi can be put down to history, but the\ndeeper ones trace back to the combination of steady pressure to add\nfeatures with refusal to embed the scripting and subprocess-control\nfeatures that vi partisans associate with\nexcessive size.  On a different level, accepting that there would \nbe two modes in the interface (insertion versus character-motion)\nopened a can of worms \u2014 it became far too easy to add new\ncommands without thinking about their complexity impact on the\noverall design.\n<\/p>\n<p>\n  The examples of Emacs and\nWily further suggest\n  <em>why<\/em>\n  some things want to be large: so that several\nrelated tasks can share context.  Editing and version control (or\nediting and mail, editing and symbolic debugging, etc.) are separate\ntasks from the point of view of the implementers \u2014 but users\nwould often prefer to have one big environment that lets them point at\npieces of text, rather than spend time and attention ping-ponging\nbetween several programs that each have to have the same filename or\nthe contents of some cut buffer handed to them.\n<\/p>\n<p>\n  More generally, let's suppose we view the entire Unix\nenvironment as a single work of design by community.  Then the religion\nof \u201csmall, sharp tools\u201d, the pressure to keep interface\ncomplexity and codebase size down, may lead right to a manularity trap\n\u2014 the user has to maintain all the shared context himself,\nbecause the tools won't do it for him.\n<\/p>\n<p>\n  Returning to the specific context of editors, Sam shows us that\nvi is the wrong thing.  Wily is a valiant\neffort to avoid the vastness of Emacs that falls short because it\ncan't be syntax-aware.  But Wily, or some\nrealization of the Emacs design ideas cleaned up and stripped of\nhistorical baggage, might be the right thing.  The value of optional\ncomplexity depends on the objectives you choose, and the ability to\nshare context among all the text-oriented tools related to a task is\nvaluable.\n<\/p>\n<h4 id=\"id2967765\"><a href=\"#id2967765\">\u00a7<\/a>Is Emacs an Argument against the Unix Tradition?<\/h4>\n<p>\n  The traditional Unix view of the world, however, is so attached\nto minimalism that it isn't very good at distinguishing between the\nadhocity-trap problems of vi and the\noptional complexity of Emacs.\n<\/p>\n <blockquote>\n  <p>\n   The reason that vi and emacs never caught on among old-school\nUnix programmers is that they are\n   <em>ugly<\/em>.\nThis complaint may be \u201cold Unix\u201d speaking, but had it not\nbeen for the singular taste of old Unix, \u201cnew Unix\u201d would\nnot exist.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Attacks on Emacs by vi users \u2014 along with attacks on vi by\nthe hard-core old-school types still attached to ed \u2014 are\nepisodes in a larger argument, a contest between the exuberance of\nwealth and the virtues of austerity.  This argument correlates with\nthe tension between the old-school and new-school styles of\nUnix.\n<\/p>\n<p>\n  The \u201csingular taste of old Unix\u201d was partly a\nconsequence of poverty in exactly the same way that Japanese\nminimalism was \u2014 one learns to do more with less most effectively\nwhen having more is not an option.  But Emacs (and new-school Unix,\nreinvented on powerful PCs and fast networks) is a child of wealth.\n<\/p>\n <blockquote>\n  <p>\n   As, in a different way, was old-school Unix.  Bell Labs had enough\nresources so that Ken was not confined by demands to have a\nproduct yesterday.  Recall Pascal's apology for writing a long\nletter because he didn't have enough time to write a short one.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Ever since, Unix programmers have maintained a tradition that\nexalts the elegant over the excessive.\n<\/p>\n<p>\n  The vastness of Emacs, on the other\nhand, did not originate under Unix, but was invented by Richard\nM. Stallman\n  within a very different culture that\nflourished at the MIT Artificial Intelligence Lab in the 1970s.  The\nMIT AI lab was one of the wealthiest corners of computer-science\nacademia; people learned to treat computing resources as cheap,\nanticipating an attitude that would not be viable elsewhere until\nfifteen years later. Stallman was unconcerned with minimalism;\nhe sought the maximum power and scope for his code.\n<\/p>\n<p>\n  The central tension in the Unix tradition has always been\nbetween doing more with less and doing more with more.  It recurs in a\nlot of different contexts, often as a struggle between designs that\nhave the quality of clean minimalism and others that choose expressive\nrange and power even at the cost of high complexity.  For both sides,\nthe arguments for or against Emacs have\nexemplified this tension since it was first ported to Unix in the\nearly 1980s.\n<\/p>\n<p>\n  Programs that are both as useful and as large as\nEmacs make Unix programmers uncomfortable\nprecisely because they force us to face the tension.  They suggest\nthat old-school Unix minimalism is valuable as a discipline, but\nthat we may have fallen into the error of dogmatism.\n<\/p>\n<p>\n  There are two ways Unix programmers can address this problem.\nOne is to deny that large is actually large. The other is to develop\na way of thinking about complexity that is not a dogma.\n<\/p>\n<p>\n  Our thought experiment with replacing Lisp and the extension\nlibraries gives us a new perspective on the oft-heard charge that\nEmacs is bloated because its extension\nlibrary is so large.  Perhaps this is as unfair as charging that\n  <em>\/bin\/sh<\/em> is bloated because the collection\nof all shellscripts on a system is large.\nEmacs could be considered a virtual machine\nor framework around a collection of small, sharp tools (the modes)\nthat happen to be written in Lisp.\n<\/p>\n<p>\n  On this view, the main difference between the shell and\nEmacs is that Unix distributors don't ship\nall the world's shellscripts along with the shell.  Objecting to\nEmacs because having a general-purpose\nlanguage in it feels like bloat is approximately as silly as refusing\nto use shellscripts because shell has conditionals and for loops.\nJust as one doesn't have to learn shell to use shellscripts, one\ndoesn't have to learn Lisp to use Emacs.  If\nEmacs has a design problem, it's not so\nmuch the Lisp interpreter (the framework part) as the fact that the\nmode library is an untidy heap of historical accretions \u2014 but\nthat's a source of complexity users can ignore, because they won't be\naffected by what they don't use.\n<\/p>\n<p>\n  This mode of argument is very comforting.  It can be applied to\nother tool-integration frameworks, such as the (uncomfortably large)\nGNOME and KDE desktop projects.  There is some force to it.\nAnd yet, we should be suspicious of any \u2018perspective\u2019 that\noffers to resolve all our doubts so neatly; it might be a\nrationalization, not a rationale.\n<\/p>\n<p>\n  Therefore, let's avoid the possibility of falling into denial\nand accept that Emacs is both useful\nand large \u2014 that it\n  <em>is<\/em>\n  an argument against\nUnix minimalism.  What does our analysis of the kinds of complexity in\nit, and the motives for it, suggest beyond that?  And is there reason\nto believe that those lessons generalize?\n<\/p>\n\n<h3 id=\"ch13s04\"><a href=\"#ch13s04\">\u00a7<\/a>The Right Size of Software<\/h3>\n<p>\n  There is a hidden dual of the Unix gospel of small, sharp tools;\na background so implicit that many Unix practitioners do not notice \nit, any more than fish notice the water they swim in.  It is the\npresence of frameworks.\n<\/p>\n<p>\n  Small, sharp tools in the Unix style have trouble sharing data,\nunless they live inside a framework that makes communication among\nthem easy. Emacs is such a framework, and\n  <em>unified management of shared context<\/em>\n  is what the\noptional complexity of Emacs is buying.\nThe practical impact of unified management of shared context is that\nthe user is not burdened with low-level naming and resource-management\nissues.\n<\/p>\n<p>\n  In old-school Unix, the only framework was pipelines,\nredirection, and the shell; the integration was done with scripts, and\nthe shared context was (essentially) the file system itself.  But that\nwas not the end of evolution.\n<\/p>\n<p>\n  Emacs unifies the file system with a world of text buffers and\nhelper processes, largely leaving the shell framework behind.  Wily is\nalso about buffers and helpers, but incorporates the shell framework\ninto itself. Modern desktop environments provide a communication\nframework for GUIs, also leaving the shell framework behind.  Each\nframework has strengths and weaknesses of its own.  Frameworks become\nhomes to ecologies of tools \u2014 the shell to shellscripts, Emacs\nto Lisp modes, and desktop environments to flocks of GUIs\ncommunicating both via drag and drop and by more esoteric means such\nas object brokers.\n<\/p>\n<p>\n  This suggests a Rule of Minimality:\n  <em>\n   Choose the shared\ncontext you want to manage, and build your programs as small as those\nboundaries will allow.\n<\/em>\n  This is \u201cas simple as\npossible, but no simpler\u201d, but it focuses attention on the\nchoice of shared context.  It applies not just to frameworks, but\nto applications and program systems.\n<\/p>\n<p>\n  It is, however, all too easy to get sloppy about how large your\nshared context needs to be. The pressure behind Zawinski's Law is the\ntendency of applications to want to share context for convenience.\nIt's easy to end up carrying around too much weight, too many\nassumptions, and to write programs that are over-complex, bloated, and\nhuge.  The paradigmatic example in the 1990s was the way that the\nmailto: URL induced the growth of huge mail clients\nembedded in Web browsers.\n<\/p>\n<p>\n  The corrective to this tendency comes straight from the\nold-school Unix hymnbook.  It is the Rule of Parsimony:\n  <em>\n   Write a big program only when it is clear by demonstration\nthat nothing else will do\n<\/em>\n  \u2014that is, when attempts to\npartition the problem have been made and failed. This maxim implies an\nastringent skepticism about large programs, and a strategy for\navoiding them: look for the small-program solution first. If a single\nsmall program won't do the job, try building a toolkit of cooperating\nsmall programs within an existing framework to attack it. Only if both\napproaches fail are you free (in the Unix tradition) to build a large\nprogram (or a new framework) without feeling you have failed the\ndesign challenge.\n<\/p>\n<p>\n  When you do write a framework, remember the Rule of Separation.\nFrameworks should be mechanism, and have as little policy as possible.\nIn most cases, that is no policy at all. Factor as much behavior as\npossible into modules that use the framework.  One of the benefits of\nwriting or reusing a framework is that it can help you separate what\nwould otherwise be big lumps of policy into separate modules, modes,\nor tools \u2014 pieces that can be usefully recombined with\nothers.\n<\/p>\n<p>\n  These rules are valuable heuristics, but the tension at the\nheart of the Unix tradition does not resolve neatly into a set of\n  \n   <em>a-priori<\/em> \n  prescriptions for optimal size\nof any given project.  Circumstances alter cases, and exercising good\njudgment and good taste is what software designers are for.  As in\nSoto Zen, the journey\n  <em>is<\/em>\n  the destination;\nenlightenment has to be rediscovered in every day of practice.\n<\/p>\n\n<h2 id=\"implementation\"><a href=\"#implementation\">\u00a7<\/a>Implementation<\/h2>\n\n<h3 id=\"languageschapter\"><a href=\"#languageschapter\">\u00a7<\/a>Chapter 14. Languages<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   The limits of my language are the limits of my world.\n  <\/p><div>\n  \n   \u2013 Ludwig Wittgenstein\nTractatus Logico-Philosophicus 5.6, 1918\n  \n <\/div><\/blockquote>\n\n<h3 id=\"ch14s01\"><a href=\"#ch14s01\">\u00a7<\/a>Unix's Cornucopia of Languages<\/h3>\n<p>\n  Unix supports a wider variety of application languages than does\nany other single operating system; indeed, it may well have hosted\nmore different languages than every other operating system in the\nhistory of computing combined.<sup><a href=\"#ftn.id2970798\" id=\"id2970798\">[122]<\/a><\/sup><\/p>\n<p>\n  There are at least two excellent reasons for this huge\ndiversity. One is the wide use of Unix as a research and teaching\nplatform. The other (far more relevant for working programmers) is the\nfact that matching your application design with the proper\nimplementation language(s) can make an immense difference in your\nproductivity. Therefore the Unix tradition encourages the design of\ndomain-specific languages (as we mentioned in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  and\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>) and what are now generally\ncalled\n  <em>scripting languages<\/em> \u2014those designed specifically to\nglue together other applications and tools.\n<\/p>\n <blockquote>\n  <p>\n   The term \u201cscripting language\u201d probably derives from\nthe term \u201cscript\u201d that was applied to a potted input for\na normally interactive program, in particular sh or ed \u2014 a much\nmore felicitous term than the \u201cruncom\u201d we inherited from\nUnix's ancestor CTSS.\n\u201cScript\u201d appears in the V7 manual (1979).  I don't recall\nwho coined the name.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  In truth, the term \u2018scripting language\u2019 is a\nsomewhat awkward one.  Many of the the major languages usually so\ndescribed (Perl, Tcl, Python, etc.) have outgrown the group's\nscripting origins and are now standalone general-purpose programming\nlanguages of considerable power.  The term tends to obscure strong\nsimilarities in style with other languages that are not usually lumped\nin with this group, notably Lisp and Java.  The only argument for\ncontinuing to use it is that nobody has yet invented a better\nterm.\n<\/p>\n<p>\n  Part of the reason all these can be lumped together under the\nrubric of \u2018scripting language\u2019 is that they all have\npretty much the same ontogeny.  Having a runtime to do interpretation\nalso makes it relatively easy to automate dynamic storage management.\nAutomating dynamic storage management almost requires using references\n(opaque memory addresses that you can't do arithmetic on) rather than\npassing value copies or explicit pointers around.  Using references\nmakes runtime polymorphism and OO an easy next step.\n  \n   <em>Voila:<\/em> \n  the modern scripting\nlanguage!\n<\/p>\n<p>\n  To apply the Unix philosophy effectively, you'll need to have\nmore than just C\n  in your\ntoolkit. You'll need to learn how to use some of Unix's other\nlanguages (especially the scripting languages), and how to be comfortable\nmixing multiple languages in specialist roles within large program\nsystems.\n<\/p>\n<p>\n  In this chapter we'll survey C and its most important\nalternatives, discussing their strengths and weaknesses and the\nsorts of tasks to which they are best matched. The languages\ncovered will be C, C++, shell, Perl, Tcl, Python, Java, and Emacs\nLisp. Each survey section will include case studies on applications\nwritten using these languages, and references to other examples and\ntutorial material. High-quality implementations of all these\nlanguages are available in open source on the Internet.\n<\/p>\n<p>\n  Warning: Choice of application language is one of the archetypal\nreligious issues in the Internet\/Unix world. People get very attached\nto these tools and will sometimes defend them past all reason. If this\nchapter achieves its aim, zealots of all stripes may be offended by\nthis chapter, but everyone else will learn from it.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2970798\" id=\"ftn.id2970798\">[122]<\/a><\/sup>\n  See the\n  <a href=\"ftp:\/\/ftp.idiom.com\/pub\/compilers-list\/free-compilers\">\n   Free\nCompiler and Interpreter List<\/a>\n  for\ndetails.\n<\/p><\/div>\n\n<h3 id=\"why_not_c\"><a href=\"#why_not_c\">\u00a7<\/a>Why Not C?<\/h3>\n<p>\n  C is the native language of Unix. Since the early 1980s it has\ncome to dominate systems programming almost everywhere in the computer\nindustry. Outside of Fortran's dwindling niche in scientific and\nengineering computing, and excluding the vast invisible dark mass of\nCOBOL financial applications at banks and insurance companies, C and\nits offspring C++ have now (in 2003) dominated applications\nprogramming almost completely for more than a decade.\n<\/p>\n<p>\n  It may therefore seem perverse to assert that C and C++ are\nnowadays almost always the wrong vehicle for beginning new\napplications development. But it's true; C and C++ optimize for\nmachine efficiency at the expense of increased implementation and\n(especially) debugging time. While it still makes sense to write\nsystem programs and time-critical kernels of applications in C or C++,\nthe world has changed a great deal since these languages came to\nprominence in the 1980s.  In 2003, processors are a thousand\ntimes faster, memories are a thousand times larger, and disks are a\nfactor of\n  <em>ten<\/em>\n  thousand larger, for roughly\nconstant dollars.<sup><a href=\"#ftn.id2971041\" id=\"id2971041\">[123]<\/a><\/sup><\/p>\n<p>\n  These plunging costs change the economics of programming in a\nfundamental way.  Under most circumstances it no longer makes sense\nto try to be as sparing of machine resources as C permits.  Instead,\nthe economically optimal choice is to minimize debugging time and\nmaximize the long-term maintainability of the code by human beings.\nMost sorts of implementation (including application prototyping) are\ntherefore better served by the newer generation of interpreted and\nscripting languages.  This transition exactly parallels\nthe conditions that, last time around the wheel, led to the rise of\nC\/C++ and the eclipse of assembler programming.\n<\/p>\n<p>\n  The central problem of C and C++ is that they require\nprogrammers to do their own memory management \u2014 to declare\nvariables, to explicitly manage pointer-chained lists, to dimension buffers,\nto detect or prevent buffer overruns, and to allocate and deallocate\ndynamic storage.  Some of this task can be automated away by\nunnatural acts like retrofitting C with a garbage collector such as\nthe Boehm-Weiser implementation, but the design of C is such that this\ncannot be a complete solution.\n<\/p>\n<p>\n  C memory management is an enormous source of complication and\nerror. One study (cited in <sup><a href=\"#Boehm\" title=\"[Boehm]\">[Boehm]<\/a><\/sup>)\nestimates that 30% or 40% of development time is devoted to storage\nmanagement for programs that manipulate complex data structures.  This\ndid not even include the impact on debugging cost.  While hard figures\nare lacking, many experienced programmers believe that\nmemory-management bugs are the single largest source of persistent\nerrors in real-world code.<sup><a href=\"#ftn.id2968237\" id=\"id2968237\">[124]<\/a><\/sup>\n  Buffer overruns are a\ncommon cause of crashes and security holes.  Dynamic-memory management\nis particularly notorious for spawning insidious and hard-to-track\nbugs, such as memory leaks and stale-pointer problems.\n<\/p>\n<p>\n  Not so long ago, manual memory management made sense anyway.\nBut there are no \u2018small systems\u2019 any more, not in\nmainstream applications programming. Under today's conditions, an\nimplementation language that automates away memory management (and\nbuys an order of magnitude decrease in bugs at the expense of using a\nbit more cycles and core) makes a lot more sense.\n<\/p>\n<p>\n  A recent paper <sup><a href=\"#Prechelt\" title=\"[Prechelt]\">[Prechelt]<\/a><\/sup>\nmusters an impressive array of statistical evidence for a claim that\nprogrammers with experience in both worlds will find very plausible:\nprogrammers are just about twice as productive in scripting\nlanguages\n  as they are in C or C++. This accords well with the 30%\u201340% penalty\nestimate noted earlier, plus debugging overhead.  The performance\npenalty of using a scripting language\n  is very often insignificant for\nreal-world programs, because real-world programs tend to be limited by\nwaits for I\/O events, network latency, and cache-line fills rather\nthan by the efficiency with which they use the CPU itself.\n<\/p>\n<p>\n  The Unix world has been slowly coming around to this point of\nview in practice, especially since 1990 or so, as is shown by the\nincreasing popularity of Perl and other scripting languages. But the\nevolution of practice has not yet (as of mid-2003) led to a\nwholesale change in conscious attitudes; many\nUnix programmers are still absorbing the lesson\nPerl\n  and\nPython\n  have\nbeen teaching.\n<\/p>\n<p>\n  We can see the same trend happening, albeit more slowly, outside\nthe Unix world \u2014 for example, in the continuing shift from C++\nto Visual Basic evident in applications development under Microsoft\nWindows and NT, and the move toward\nJava\n  in the\nmainframe world.\n<\/p>\n<p>\n  The arguments against C and C++ apply with equal force to other\nconventional compiled languages such as Pascal, Algol, PL\/I, FORTRAN,\nand compiled Basic dialects. Despite occasional heroic efforts such as\nAda, the differences between conventional languages remain superficial\nwhen set against their basic design decision to leave memory\nmanagement to the programmer.  Though high-quality open-source\nimplementations of most languages ever written are available under\nUnix, no other conventional languages remain in widespread use in the\nUnix or Windows worlds; they have been abandoned in favor of C and\nC++.  Accordingly we will not survey them here.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2971041\" id=\"ftn.id2971041\">[123]<\/a><\/sup>\n  Outside the Unix world, this\nthree-orders-of-magnitude improvement in hardware performance has been\nmasked to a significant extent by a corresponding drop in software\nperformance.\n<\/p><p><sup><a href=\"#id2968237\" id=\"ftn.id2968237\">[124]<\/a><\/sup>\n  The severity of this problem\nis attested to by the rich slang Unix programmers have developed for\ndescribing different varieties: \u2018aliasing bug\u2019,\n\u2018arena corruption\u2019, \u2018memory leak\u2019,\n\u2018buffer overflow\u2019, \u2018stack smash\u2019,\n\u2018fandango on core\u2019, \u2018stale pointer\u2019,\n\u2018heap trashing\u2019, and the rightly dreaded \u2018secondary\ndamage\u2019.  See the\n  <a href=\"http:\/\/www.catb.org\/~esr\/jargon\">\n   Jargon\nFile<\/a>\n  for elucidation.\n<\/p><\/div>\n\n<h3 id=\"ch14s03\"><a href=\"#ch14s03\">\u00a7<\/a>Interpreted Languages and Mixed Strategies<\/h3>\n<p>\n  Languages that avoid manual memory management do it by having a\nmemory manager built into their runtime executable somewhere.\nTypically, runtime environments in these languages are split into a\nprogram part (the running script itself) and the interpreter part,\nwith the interpreter managing dynamic storage. On Unixes (and other\nmodern operating systems) the interpreter core can be shared by\nmultiple program parts, reducing the effective overhead for each\none.\n<\/p>\n<p>\n  Scripting is nowhere near a new idea\nin the Unix world. As far back as the mid-1970s, in an era of far\nsmaller machines, the Unix shell (the interpreter for commands typed\nto a Unix console) was designed as a full interpreted programming\nlanguage. It was common even then to write programs entirely in\nshell, or\nto use the shell to write glue logic that knit together canned\nutilities and custom programs in C into wholes greater than the sum of\ntheir parts. Classical introductions to the Unix environment (such as\n  <em>The Unix Programming Environment<\/em> <sup><a href=\"#Kernighan-Pike84\" title=\"[Kernighan-Pike84]\">[Kernighan-Pike84]<\/a><\/sup>) have dwelt\nheavily on this tactic, and with good reason: it was one of Unix's\nmost important innovations.\n<\/p>\n<p>\n  Advanced shell programming mixes languages freely, employing\nboth binaries and interpreted elements from half a dozen or more\nother languages for subtasks. Each language does what it does best,\neach component is a module with narrow interfaces to the others,\nand the global complexity of the whole is much lower than it would\nbe had it been coded as a single monster monolith in a general-purpose\nlanguage.\n<\/p>\n\n<h3 id=\"ch14s04\"><a href=\"#ch14s04\">\u00a7<\/a>Language Evaluations<\/h3>\n<p>\n  Mixing languages is a knowledge-intensive (rather than coding-intensive)\nstyle of programming. To make it work, you have to have both\nworking knowledge of a suitable variety of languages and expertise\nabout what they're best at and how to fit them together. In this\nsection, we will try to point you at references to help you with the\nfirst and an overview to convey the second. For each language\nsurveyed we will include case studies of successful programs that\nexemplify its strengths.\n<\/p>\n<h4 id=\"c_language\"><a href=\"#c_language\">\u00a7<\/a>C<\/h4>\n<p>\n  Despite the memory-management problem, there are some\napplication niches for which C is still king. Programs that require maximum speed, have\nreal-time requirements, or are tightly coupled to the OS kernel are\ngood candidates for C.\n<\/p>\n<p>\n  Programs that must be portable across multiple operating systems\nmay also be good candidates for C. Some of the alternatives to C that\nwe shall discuss below are, however, increasingly penetrating major\nnon-Unix operating systems; in the near future, portability may be less a\ndistinguishing advantage of C.\n<\/p>\n<p>\n  Sometimes the leverage to be gained from existing programs like\nparser generators or GUI builders that generate C code is so great\nthat it justifies C coding of the rest of a small application.\n<\/p>\n<p>\n  And, of course, C proved indispensable to the developers of all\nits alternatives. Dig down through enough implementation layers\nunder any of the other languages surveyed here and you will find a\ncore implemented in pure, portable C. These languages inherit \nmany of the advantages of C.\n<\/p>\n<p>\n  Under modern conditions, it's perhaps best to think of C as a\nhigh-level assembler for the Unix virtual machine (recall the\ndiscussion of the success of C as a case study in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>).  C standards have exported many of the\nfacilities of this virtual machine, such as the standard I\/O library,\nto other operating systems.  C is where you go when you want to get as\nclose as possible to the bare metal but stay portable.\n<\/p>\n<p>\n  One good reason to learn C, even if your programming needs are\nsatisfied by a higher-level language, is that it can help you learn to\nthink at hardware-architecture level.  The best reference and tutorial\non C for people who are already programmers is still\n  <em>The C Programming Language<\/em><sup><a href=\"#Kernighan-Ritchie\" title=\"[Kernighan-Ritchie]\">[Kernighan-Ritchie]<\/a><\/sup>.\n<\/p>\n<p>\n  Porting C code between Unix variants is almost always possible\nand usually easy, but specific areas of variation (like\nsignals and process control) can be tricky to get right.  We\nhighlight some of these issues in\n  <a href=\"#portabilitychapter\" title=\"Chapter 17. Portability\">\n   Chapter 17<\/a>.  Differing C bindings on other\noperating systems can of course cause C portability problems, although\nWindows NT at least theoretically supports an ANSI\/POSIX-compliant C\nAPI.\n<\/p>\n<p>\n  High-quality C compilers are available as open-source software\nover the Internet; the best-known and most widely used is the Free\nSoftware Foundation's GNU C compiler (part of GCC, the GNU Compiler\nCollection), which has become the native C of\nall open-source Unix systems and many even in the closed-source\nworld. GCC ports are even available for\nMicrosoft's\n  family of operating systems. GCC sources\nare available at the FSF's\n  <a href=\"ftp:\/\/ftp.gnu.org\/pub\/gnu\">\n   FTP\nsite<\/a>.\n<\/p>\n<p>\n  Summing up: C's best side is resource efficiency and\ncloseness to the machine.  Its worst side is that programming \nin it is a resource-management hell.\n<\/p>\n<h5 id=\"fetchmail\"><a href=\"#fetchmail\">\u00a7<\/a>\n  C Case Study:\nfetchmail\n <\/h5>\n<p>\n  The best case study for C is the Unix kernel itself, for which a\nlanguage that naturally supports hardware-level operations is\nactually a strong advantage. But\n  <em>fetchmail<\/em> is a good example of the kind of\nuser-land utility that is still best coded in C.\n<\/p>\n<p>\n  <em>fetchmail<\/em> does only the simplest kind\nof dynamic-memory management; its only complex data structure is a\nsingly-linked list of per-mailserver control blocks built just once,\nat startup time, and changed only in fairly trivial ways\nafterwards. This substantially erodes the case against using C by\nsidestepping C's greatest weakness.\n<\/p>\n<p>\n  On the other hand, these control blocks are fairly complex\n(including all of string, flag, and numeric data) and would be\ndifficult to handle as coherent fast-access objects in an\nimplementation language without an equivalent of the C struct\nfeature. Most of the alternatives to C are weaker than C in this\nrespect (Python\n  and Java\n  being notable exceptions).\n<\/p>\n<p>\n  Finally,\n  <em>fetchmail<\/em> requires the\nability to parse a fairly complex specification syntax for\nper-mail-server control information. In the Unix world this sort of\nthing is classically handled by using C code generators that grind out\nsource code for a tokenizer and grammar parser from declarative\nspecifications. The existence of\n  <em>yacc<\/em> and\n  <em>lex<\/em> was a point in favor of C.\n<\/p>\n<p>\n  <em>fetchmail<\/em> might reasonably have been coded in\nPython,\nalbeit with possibly significant loss of performance. Its size and\ndata-structure complexity would have excluded shell and\nTcl\n  right off\nand strongly counterindicated Perl, and the application domain is outside\nthe natural scope of Emacs Lisp. A Java\n  implementation wouldn't have been an\nunreasonable path, but Java's object-oriented style and garbage\ncollection would have offered little purchase on\n  <em>fetchmail<\/em> 's specific problems over what\nC\n  already\nyields. Nor could C++\n  have done much to simplify the\nrelatively simple internal logic of\n  <em>fetchmail<\/em>.\n<\/p>\n<p>\n  However, the real reason\n  <em>fetchmail<\/em> is\na C program is that it evolved by gradual mutation from an ancestor\nalready written in C.  The existing implementation has been\nextensively tested on many different platforms and against many odd\nand quirky servers.  Carrying all that implicit knowledge through to a\nre-implementation in a different language would be messy and\ndifficult.  Furthermore,\n  <em>fetchmail<\/em> depends\non imported code for functions (like NTLM authentication) that don't\nseem to be available above C level.\n<\/p>\n<p>\n  <em>fetchmail<\/em> 's interactive configurator,\nwhich did not have a C legacy problem, is written in Python; we'll\ndiscuss that case along with that language.\n<\/p>\n<h4 id=\"cc_language\"><a href=\"#cc_language\">\u00a7<\/a>C++<\/h4>\n<p>\n  When C++ was first released to the world in the mid-1980s\nobject-oriented (OO)\n  languages were being widely touted\nas the silver bullet for the software-complexity problem. C++'s OO\nfeatures appeared to be an overwhelming advantage over the ancestral\nC, and partisans expected that it would rapidly make the older\nlanguage obsolete.\n<\/p>\n<p>\n  This has not happened. Part of the fault can be laid to problems\nin C++ itself; the requirement that it be backward-compatible with C\nforced a great many compromises on the design.  Among other things,\nthat requirement prevented C++ from going to fully automatic\ndynamic-memory management and addressing C's most serious problem.\nLater, feature arms races between different compiler implementers,\nunconstrained by a weak and premature standardization effort, pushed\nC++ to become rather baroque and excessively complicated.\n<\/p>\n<p>\n  Another part of the fault must be laid to the failure of\nOO\n  itself\nto live up to expectations.  We examined this problem in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>, observing the tendency of OO methods to\nlead to thick glue layers and maintenance problems.  Today\n(2003), inspection of open-source archives (in which choice of\nlanguage reflects developers' judgments rather than corporate\nmandates) reveals that C++ usage is still heavily concentrated in GUIs,\nmultimedia toolkits and games (the major success areas for OO\ndesign), and little used elsewhere.\n<\/p>\n<p>\n  It may be that C++'s realization of OO\n  is particularly problem-prone. There\nis some evidence that C++ programs have higher life-cycle costs than\nequivalents in C, FORTRAN, or Ada. Whether this is a problem with\nOO\n  or\nspecifically with C++ or both remains unclear, though there is reason\nto suspect both are implicated <sup><a href=\"#Hatton98\" title=\"[Hatton98]\">[Hatton98]<\/a><\/sup>.\n<\/p>\n<p>\n  In recent years, C++ has incorporated some important non-OO ideas.  It has exceptions similar to those in Lisp; that\n  is, it is possible to throw an object or value up the call stack until it is caught by a handler.  STL (Standard\n  Template Library) provides generic programming; that is, it is possible to code algorithms that are independent of the\n  type signature of their data and have them compiled to do the right thing at runtime. (Only languages that do\n  compile-time static type-checking need this; more dynamic languages simply pass around typeless references and support\n  type identification at runtime.)\n<\/p>\n<p>\n  Efficient compiled language; upward-compatible with C;\nobject-oriented platform; vehicle for cutting-edge techniques like STL\nand generics \u2014 C++ tries to be all things to all people, but the cost\nis more complexity than the mind of any individual programmer can\nhandle. As we noted in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>, the language's principal designer has\nconceded that he doesn't expect any one programmer to grasp it\nall. Unix hackers do not react well to this; one anonymous but famous\ncharacterization is \u201cC++: an octopus made by nailing extra legs\nonto a dog\u201d.\n<\/p>\n<p>\n  When all is said and done, however, C++'s most fundamental\nproblem is that it is basically just another conventional language. It\nconfines the memory-management problem better than it did before the\ninvention of the Standard Template Library, and a lot better than C\ndoes, but the confinement is brittle; it breaks unless your code\nuses objects and only objects. For many types of application its\nOO\n  features are not significant, and simply add complexity to C without\nyielding much advantage. Open-source C++ compilers are available; if\nC++ were unequivocally superior to C it would now dominate.\n<\/p>\n<p>\n  Summing up: C++'s best side is its combination of compiled\nefficiency with facilities for OO\n  and generic programming.  Its worst\nside is that it is baroque and complex, and tends to encourage\nover-complex designs.\n<\/p>\n<p>\n  Consider using C++ if an existing C++ toolkit or service library\noffers powerful leverage for your application, or if you're in one of\nthe application areas mentioned above for which an\nOO\n  language is known to be a large win.\n<\/p>\n<p>\n  The classic C++ reference is Stroustrup's\n  <em>The C++ Programming Language<\/em><sup><a href=\"#Stroustrup\" title=\"[Stroustrup]\">[Stroustrup]<\/a><\/sup>. You will find an excellent\nbeginner's tutorial on C++ and basic OO methods in\n  <em>C++: A Dialog<\/em><sup><a href=\"#Heller\" title=\"[Heller]\">[Heller]<\/a><\/sup>.\n  <em>C++ Annotations<\/em> <sup><a href=\"#Brokken\" title=\"[Brokken]\">[Brokken]<\/a><\/sup> is a condensed introduction to C++\nfor expert C programmers.\n<\/p>\n<p>\n  The Gnu Compiler Collection\n  includes a C++ compiler. The language\nis therefore universally available on Unix and on\nMicrosoft\n  operating systems; comments made under C\nabove also apply here.  Strong collections of open-source\n  <a href=\"http:\/\/www.boost.org\/\">\n   support libraries<\/a>\n  are available.\nHowever, portability is compromised by the fact that (as of\nmid-2003) actual C++ implementations implement widely varying subsets of the\ndraft ISO standard now in preparation.<sup><a href=\"#ftn.id2969184\" id=\"id2969184\">[125]<\/a><\/sup><\/p>\n<h5 id=\"id2969192\"><a href=\"#id2969192\">\u00a7<\/a>C++ Case Study: The Qt Toolkit<\/h5>\n<p>\n  The Qt interface toolkit is one of the notable C++ success\nstories in today's open-source world. It provides a widget set and API\nfor writing graphical user interfaces under X, one deliberately (and\nrather effectively) designed to emulate the visual look and feel of\nMotif, MacOS Platinum, or the Microsoft\nWindows\n  interface.  Qt actually provides more\nthan just GUI services; it also provides a portable application layer,\nwith classes for XML, file access, sockets, threads, timers, time\/date\nhandling, database access, various abstract data types, and\nUnicode.\n<\/p>\n<p>\n  The Qt toolkit is a critical and visible component of the KDE\nproject, the senior of the open-source world's two efforts to\nproduce a competitive GUI and integrated set of desktop\nproductivity tools.\n<\/p>\n<p>\n  Qt's C++ implementation exhibits the strengths of an\nOO\n  language for encapsulating user-interface components. In a language\nsupporting objects, a visual hierarchy of interface widgets can be\ncleanly expressed in the code by a hierarchy of class instances. While\nthis sort of thing can be simulated in C\n  with explicit indirection through\nhand-rolled method tables, such code is much cleaner in C++.\nComparison with the notoriously baroque C API of Motif is\ninstructive.\n<\/p>\n<p>\n  The Qt source code and reference documentation are available at the\n  <a href=\"http:\/\/www.trolltech.com\/\">\n   Trolltech site<\/a>.\n<\/p>\n<h4 id=\"sh\"><a href=\"#sh\">\u00a7<\/a>Shell<\/h4>\n<p>\n  The \u2018Bourne shell\u2019 (\n  <em>sh<\/em>)\nof Version 7 Unix was Unix's first (and for many years its only)\nportable interpreted language. Today the ancestral Bourne shell has\nlargely been displaced by variants of the upward-compatible Korn\nShell\n  (\n  <em>ksh<\/em>); the single most important of these\nis the Bourne Again Shell, bash.\n<\/p>\n<p>\n  A few other shells exist and are used interactively, but are not\nsignificant as programming languages; of these, the best known is\nprobably the C shell\n  <em>csh<\/em>, which is\nnotoriously<sup><a href=\"#ftn.id2969353\" id=\"id2969353\">[126]<\/a><\/sup>\n  unsuitable for writing\nscripts.\n<\/p>\n<p>\n  Simple shell programs are extremely easy and natural to write.\nThe Unix tradition of rapid prototyping in interpretive languages\nbegan with shell.\n<\/p>\n <blockquote>\n  <p>\n   I wrote the very first version of netnews as a 150-line\nshellscript.  It had multiple newsgroups and cross-posting;\nnewsgroups were directories and cross-posting was implemented\nas multiple links to the article.  It was far too slow to\nuse for production, but the flexibility permitted endless\nexperimentation with the protocol design.\n  <\/p><div>\n  \n   \u2013 Steven M. Bellovin\n  \n <\/div><\/blockquote>\n<p>\n  As program size gets larger, however, they tend to become rather\nad-hoc. Some parts of shell syntax (notably its quoting and\nstatement-syntax rules) can be very confusing. These drawbacks\ngenerally relate to compromises in the programming-language part of\nthe shell's design made to preserve its utility as an interactive\ncommand-line interpreter.\n<\/p>\n<p>\n  Programs are described as being \u2018in shell\u2019 even when\nthey are not pure shell but include heavy use of\nC\n  filters like\nsort(1)\nand of standard text-processing minilanguages like\nsed(1)\nor\nawk(1). This\nsort of programming has been in decline for some years, however;\nnowadays such elaborate glue logic is generally written in\nPerl\n  or\nPython, with\nshell being reserved for the simplest kinds of wrappers (for which\nthese languages would be overkill) and system boot-time initialization\nscripts (which cannot assume they are available).\n<\/p>\n<p>\n  Such basic shell programming should be adequately covered in any\nintroductory Unix book.\n  <em>The Unix Programming Environment<\/em><sup><a href=\"#Kernighan-Pike84\" title=\"[Kernighan-Pike84]\">[Kernighan-Pike84]<\/a><\/sup> remains one of\nthe best sources on intermediate and advanced shell programming.  Korn\nshell implementations or clones are present on every Unix.\n<\/p>\n<p>\n  Complex shellscripts often have portability problems, not so\nmuch because of the shell itself but because they make assumptions\nabout what other programs are available as components.  While Bourne\nand Korn-shell clones have been sporadically available on non-Unix\noperating systems, shell programs are not (practically speaking) at\nall portable off Unix.\n<\/p>\n<p>\n  Summing up: shell's best side is that it is very natural and\nquick for small scripts.  Its worst side is that large shellscripts\ndepend on lots of auxiliary commands that aren't necessarily\nidentically behaved nor even present on all target machines.  Nor\nis it easy to analyze the dependencies in a large shellscript.\n<\/p>\n<p>\n  It is almost never necessary to build or install a shell, since all\nUnix systems and Unix emulators come equipped with them.  The standard\nshell on Linux\n  and other leading-edge Unix variants is\nnow\n  <em>bash<\/em>.\n<\/p>\n<h5 id=\"xmlto\"><a href=\"#xmlto\">\u00a7<\/a>Case Study: xmlto<\/h5>\n<p>\n  <em>xmlto<\/em> is a driver script\nthat calls all the commands needed to render an XML-DocBook document\nas HTML, PostScript, plain text, or in any one of several other\nformats (we'll take a closer look at DocBook in\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>).  It is written\nin bash.\n<\/p>\n<p>\n  <em>xmlto<\/em> handles the details\nof calling an XSLT engine with appropriate stylesheet, then handing\noff the result to a postprocessor.  For HTML and XHTML the XSLT\ntransformation does the entire job.  For plain text, the XML is also\nprocessed into HTML, but then handed to a postprocessor \u2014\nlynx(1)\nin its\n  <code>-dump<\/code> mode, which renders HTML to flat text.\nFor PostScript, the XML is transformed to XML FO (formatting\nobjects)\n  which a\npostprocessor then massages into\n  <em>TeX<\/em> macros, to DVI\n  format via\ntex(1),\nand then finally to PostScript via the well-known\ndvi2ps(1)\ntool.\n<\/p>\n<p>\n  <em>xmlto<\/em> consists of a\nsingle front-end shellscript.  It calls any one of several script\nplugins, each named after the target format.  Each plugin is a\nshellscript.  Depending on how it's called, it either supplies a\nstylesheet for the front end to apply, or calls the appropriate\npostprocessor(s) with various canned arguments.\n<\/p>\n<p>\n  This architecture means that all the information about a given\noutput format lives in one place (the corresponding script plugin), so\nadding new output types can be done without disturbing the front-end\ncode at all.\n<\/p>\n<p>\n  <em>xmlto<\/em> is a good example of a\nmedium-sized shell application.  Neither C\n  nor C++ would have made sense\nbecause they are awkward for scripting.  Any of the other scripting\nlanguages in this chapter could have been used for this job; but it's\nall simple command dispatching, with no internal data structures or\ncomplex logic, so shell is good enough. Shell has the significant\nadvantage of being ubiquitous on the intended target systems.\n<\/p>\n<p>\n  In theory this script could run on any system supporting\n  <em>bash<\/em>.  The real constraint is the\nrequirement for one of the XSLT engines and all the postprocessors\nneeded to be present on the system.  In practice, this script is not\nlikely to run anywhere but under one of the modern open-source\nUnixes.\n<\/p>\n<h5 id=\"id2975671\"><a href=\"#id2975671\">\u00a7<\/a>Case Study: Sorcery Linux<\/h5>\n<p>\n  Sorcerer GNU\/Linux is a Linux\n  distribution that you install as a small,\nbootable foothold system just powerful enough to run\nbash(1)\nand a couple of download utilities.  With this code in place, you can\ninvoke Sorcery, the Sorcerer package system.\n<\/p>\n<p>\n  Sorcery handles installing, removing, and integrity-checking\nsoftware packages.  When you \u201ccast spells\u201d, Sorcery\ndownloads the source code, compiles it, installs it, and saves a list\nof files that were installed (along with a build log and checksums for\nall the files).  Installed packages can be \u201cdispelled\u201d or\nremoved. Package listing and integrity checks are also available.\nMore details are available at the\n  <a href=\"http:\/\/sorcerer.wox.org\">\n   Sorcery project site<\/a>.\n<\/p>\n<p>\n  The Sorcery system is written entirely in shell.  Program\ninstallation procedures tend to be small, simple programs for which\nshell is appropriate.  In this particular application, the main\ndrawback of shell is neutralized because Sorcery's authors can\nguarantee that the helper programs they need will be present in the\nfoothold system.\n<\/p>\n<h4 id=\"perl\"><a href=\"#perl\">\u00a7<\/a>Perl<\/h4>\n<p>\n  Perl is shell on steroids. It was specifically designed to\nreplace\nawk(1),\nand expanded to replace shell as the \u2018glue\u2019 for\nmixed-language script programming.\n  It was first released in 1987.\n<\/p>\n<p>\n  Perl's strongest point is its extremely powerful built-in\nfacilities for pattern-directed processing of textual, line-oriented\ndata formats; it is unsurpassed at this. It also includes far stronger\ndata structures than shell, including dynamic arrays of mixed element\ntypes and a \u2018hash\u2019 or \u2018dictionary\u2019 type that\nsupports convenient and fast lookup of name-value pairs.\n<\/p>\n<p>\n  Additionally, Perl includes a rather complete and\nwell-thought-out internal binding of virtually the entire Unix API,\ndrastically reducing the need for C\n  and making it suitable for jobs like\nsimple TCP\/IP\n  clients\nand even servers. Another strong advantage of Perl is that a large and\nvigorous open-source community has grown up around it. Its home on\nthe net is the\n  <a href=\"http:\/\/www.cpan.org\">\n   Comprehensive Perl\nArchive Network<\/a>. Dedicated Perl hackers have written hundreds\nof freely reusable Perl modules for many different programming\ntasks. These include everything from structure-walking of directory\ntrees through X toolkits for GUI building, through excellent canned\nfacilities for supporting HTTP robots and CGI programming.\n<\/p>\n<p>\n  Perl's main drawback is that parts of it are irredeemably ugly,\ncomplicated, and must be used with caution and in stereotyped ways\nlest they bite (its argument-passing conventions for functions are a\ngood example of all three problems). It is harder to get started in\nPerl than it is in shell. Though small programs in Perl can\nbe extremely powerful, careful discipline is required to maintain\nmodularity and keep a design under control as program size\nincreases. Because some limiting design decisions early in Perl's\nhistory could not be reversed, many of the more advanced features have\na fragile, klugey feel about them.\n<\/p>\n<p>\n  The definitive reference on Perl is\n  <em>Programming Perl<\/em><sup><a href=\"#Wall2000\" title=\"[Wall2000]\">[Wall2000]<\/a><\/sup>. This book has nearly\neverything you will ever need to know in it, but is notoriously badly\norganized; you will have to dig to find what you want. A more\nintroductory and narrative treatment is available in\n  <em>Learning Perl<\/em> <sup><a href=\"#Schwartz-Christiansen\" title=\"[Schwartz-Christiansen]\">[Schwartz-Christiansen]<\/a><\/sup>.\n<\/p>\n<p>\n  Perl is universal on Unix systems.  Perl scripts at the same\nmajor release level tend to be readily portable between Unixes\n(provided they don't use extension modules). Perl implementations are\navailable (and even well documented) for the\nMicrosoft\n  family of operating systems and on\nMacOS\n  as well. Perl\/Tk provides cross-platform GUI capability.\n<\/p>\n<p>\n  Summing up: Perl's best side is as a power tool for small glue\nscripts involving a lot of regular-expression grinding.  Its worst\nside is that it is ugly, spiky, and nigh-unmaintainable in large\nvolumes.\n<\/p>\n<h5 id=\"id2975926\"><a href=\"#id2975926\">\u00a7<\/a>A Small Perl Case Study: blq<\/h5>\n<p>\n  The\n  <em>blq<\/em> script is a tool\nfor querying block lists (lists of Internet sites that have been\nidentified as habitual sources of unsolicited bulk email, aka spam).\nYou can find current sources at the\n  <a href=\"http:\/\/www.unicom.com\/sw\/blq\/\">\n   <em>blq<\/em> project page<\/a>.\n<\/p>\n<p>\n  <em>blq<\/em> is a good example of\na small Perl script, illustrating both the strengths and weaknesses of\nthe language. It makes intensive use of regular-expression matching.\nOn the other hand, the Net::DNS Perl extension module it uses has to\nbe conditionally included, because it is not guaranteed to be present\nin any given Perl installation.\n<\/p>\n<p>\n  <em>blq<\/em> is exceptionally\nclean and disciplined as Perl code goes, and I recommend it as an\nexample of good style (the other Perl tools referenced from the\n  <em>blq<\/em> project page are good examples as\nwell).  But parts of the code are unreadable unless you are familiar\nwith very specific Perl idioms \u2014 the very first line of code,\n  <code>$0 =~ s!.*\/!!;<\/code>, is an example.  While all languages\nhave some of this kind of opacity, Perl has it worse than most.\n<\/p>\n<p>\n  Tcl\n  and\nPython\n  are\nboth good for small scripts of this type, but both lack the Perl\nconvenience features for regular-expression matching that\n  <em>blq<\/em> uses heavily; an implementation in\neither would have been reasonable, but probably less\ncompact\n  and\nexpressive.  An Emacs Lisp\n  implementation would have been even\nfaster to write and more compact than the Perl one, but probably\npainfully slow to use.\n<\/p>\n<h5 id=\"id2976082\"><a href=\"#id2976082\">\u00a7<\/a>A Large Perl Case Study: keeper<\/h5>\n<p>\n  <em>keeper<\/em> is the tool used to file\nincoming packages and maintain both FTP and WWW index files for the\nhuge Linux\n  free-software archives at ibiblio. You\ncan find sources and documentation in the\n  <a href=\"http:\/\/www.ibiblio.org\">\n   search tools subdirectory of the ibiblio\narchive<\/a>.\n<\/p>\n<p>\n  <em>keeper<\/em> is a good example of a\nmedium-to-large interactive Perl application. The command-line\ninterface is line-oriented and patterned after a specialized shell or\ndirectory editor; note the embedded help facilities. The working parts\nmake heavy use of file and directory handling, pattern matching, and\npattern-directed editing. Note the ease with which\n  <em>keeper<\/em> generates Web pages and\nelectronic-mail notifications from programmatic templates. Note also\nthe use of a canned Perl module to automate walking various functions\nover directory trees.\n<\/p>\n<p>\n  At about 3300 lines, this application is probably pushing the\nsize and complexity limit of what one should attempt in a single Perl\nprogram. Nevertheless, most of it was written in a period of six\ndays. In C,\nC++,\n  or\nJava\n  it would\nhave taken a minimum of six weeks and been extremely difficult to\ndebug or modify after the fact. It is way too large for pure\nTcl. A\nPython\n  version would probably be structurally cleaner, more readable, and\nmore maintainable \u2014 but also more verbose (especially near the\npattern-matching parts).  An Emacs Lisp\n  mode could readily do the job, but\nEmacs is not well suited for use over a\ntelnet link that is often slowed to a crawl by server\ncongestion.\n<\/p>\n<h4 id=\"tcl\"><a href=\"#tcl\">\u00a7<\/a>Tcl<\/h4>\n<p>\n  Tcl (Tool Command Language) is a small language interpreter\ndesigned to link with compiled\nC\n  libraries, providing\nscripted control of C code (\n  <em>\n   extended\nscripts\n<\/em>).  Its original application was to\ncontrol libraries for electronic simulators (SPICE-like applications).\nTcl is also suitable for\n  <em>embedded scripts<\/em>\n  \u2014that \nis, scripts called from within C programs and returning values to\nthose programs.  Tcl had its first general public release in\n1990.\n<\/p>\n<p>\n  Some facilities built on top of Tcl have achieved wide use outside\nthe Tcl community itself. The two most important of these are:\n<\/p>\n <div><ul><li>\n    <p>\n     The Tk toolkit, a kinder and gentler X interface that makes it\neasy to rapidly build buttons, dialog boxes, menu trees, and\nscrolling text widgets and collect input from them.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Expect, a language that makes it relatively easy to script\nfully interactive programs with widely variable responses.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  The Tk toolkit is so important that the language is often\nreferred to as Tcl\/Tk. Tk is also frequently used with\nPerl\n  and\nPython.\n<\/p>\n<p>\n  The main advantage of Tcl itself is that it is extremely\nflexible and radically simple. The syntax is very odd (based on a\npositional parser) but totally consistent. There are no reserved\nwords, and there is no syntactic distinction between a function call\nand \u2018built in\u2019 syntax; thus the Tcl language interpreter\nitself can be effectively redefined from within Tcl (which is what\nmakes projects like\n  <em>Expect<\/em> reasonable).\n<\/p>\n<p>\n  The main drawback of Tcl is that the pure language has only weak\nfacilities for namespace control and modularity, and two of them\n(<code>upvar<\/code> and <code>uplevel<\/code>) are rather\ndangerous if not used with great caution. Also, there are no data\nstructures other than association lists.  Tcl therefore scales up very\npoorly \u2014 it is difficult to organize and debug pure Tcl programs of\neven moderate size (more than a few hundred lines) without tripping\nover your own feet. In practice, almost all large Tcl programs use one\nof several OO\n  extensions to the language.\n<\/p>\n<p>\n  The oddities of the syntax can at first be a problem as well; the\ndistinction between string quotes and braces will probably give you\nheadaches for a while, and the rules for when things need to be quoted\nor braced are a bit tricky.\n<\/p>\n<p>\n  Pure Tcl only provides access to a relatively small\nand commonly used part of the Unix API (essentially just file handling,\nprocess-spawning, and sockets). Indeed, Tcl has the flavor of an\nexperiment in seeing how small a scripting\nlanguage\n  can get and still be useful.  Tcl extensions (similar to Perl modules)\nprovide a richer set of capabilities, but are (like CPAN modules) not\nguaranteed to be installed everywhere.\n<\/p>\n<p>\n  The original Tcl reference is\n  <em>Tcl and the Tk Toolkit<\/em><sup><a href=\"#Ousterhout94\" title=\"[Ousterhout94]\">[Ousterhout94]<\/a><\/sup>, but that book has\nbeen largely superseded by\n  <em>Practical Programming in Tcl and Tk<\/em><sup><a href=\"#Welch\" title=\"[Welch]\">[Welch]<\/a><\/sup>.  Brian Kernighan has written a\ndescription of a real-world Tcl project <sup><a href=\"#Kernighan95\" title=\"[Kernighan95]\">[Kernighan95]<\/a><\/sup> that summarizes Tcl's strengths and weaknesses\nas a rapid-prototyping and production tool; his contrast with\nMicrosoft Visual Basic is particularly balanced and\ninstructive.\n<\/p>\n<p>\n  The Tcl world doesn't have one central repository run by a core\ngroup analogous to Perl's or Python's, but several excellent\nwebsites both point to each other and cover most Tcl tool and\nextension development. Look at the\n  <a href=\"http:\/\/www.tcltk.com\">\n   Tcl Developer Xchange<\/a>\n  first; among\nother things, it offers Tcl sources of an interactive Tcl tutorial.\nThere is also a\n  <a href=\"http:\/\/sourceforge.net\/foundry\/tcl-foundry\/\">\n   Tcl foundry at\nSourceForge<\/a>.\n<\/p>\n<p>\n  Tcl scripts have portability problems similar to those of shell\nscripts; the language itself is highly portable, but the components it\ncalls may not be.  Tcl implementations exist for the Microsoft family\nof operating systems, MacOS, and many other platforms. Tcl\/Tk scripts\nwill run on any platform with GUI capabilities.\n<\/p>\n<p>\n  Summing up: Tcl's best side is its spare,\ncompact\n  design\nand the\nextensibility\n  of the Tcl interpreter.  Its worst side is the odd positional parser\nand the weakness of its data structures and namespace controls; the\nlatter defect makes it scale poorly for large projects.\n<\/p>\n<h5 id=\"id2976538\"><a href=\"#id2976538\">\u00a7<\/a>Case Study: TkMan<\/h5>\n<p>\n  <em>TkMan<\/em> is a browser for Unix man pages\nand Texinfo documents.  At roughly 1200 lines, it is quite large to be\nwritten in pure Tcl, but the code is unusually well-modularized and\nmature.  It uses Tk to provide a GUI interface quite a bit nicer than\neither the stock\nman(1)\nor\nxman(1)\nutilities support.\n<\/p>\n<p>\n  TkMan makes a good case study because it exhibits almost the\nfull gamut of Tcl techniques.  Highlights include Tk integration, \nscripted control of other Unix applications (such as the Glimpse \nsearch engine), and the use of Tcl to parse Texinfo markup.\n<\/p>\n<p>\n  Any of the other languages would have made for a less direct\ninterface to the Tk GUI that constitutes most of this code.\n<\/p>\n<p>\n  A Web search for \u201cTkMan\u201d should turn up sources\nand documentation.\n<\/p>\n<h5 id=\"id2976607\"><a href=\"#id2976607\">\u00a7<\/a>Moodss: A Large Tcl Case Study<\/h5>\n<p>\n  The Moodss system is a graphical monitoring application for\nsystem administrators.  It can watch logs and gather statistics for\nMySQL, Linux, SNMP networks, and\nApache, and presents a\ndigested view of them through spreadsheet-like GUI panels called\n\u2018dashboards\u2019. Monitoring modules can be written in\nPython\n  or\nPerl\n  as well as\nTcl.  The code is polished, mature, and considered an exemplar in the\nTcl community.  There is a\n  <a href=\"http:\/\/jfontain.free.fr\/moodss\/\">\n   project website<\/a>.\n<\/p>\n<p>\n  The Moodss core consists of about 18,000 lines of Tcl.  It uses\nseveral Tcl extensions including a custom object system; the Moodss author\nadmits that without these \u201cwriting such a big application would \nnot have been possible\u201d.\n<\/p>\n<p>\n  Again, any of the other languages would have made for a less direct\ninterface to the Tk GUI that constitutes most of this code.\n<\/p>\n<h4 id=\"python_language\"><a href=\"#python_language\">\u00a7<\/a>Python<\/h4>\n<p>\n  Python is a scripting language\n  designed for close integration with\nC. It can both\nimport data from and export data to dynamically loaded C libraries,\nand can be called as an embedded scripting language from C. Its syntax\nis rather like a cross between that of C and the Modula family, but\nhas the unusual feature that block structure is actually controlled by\nindentation (there is no analog of explicit begin\/end or C curly\nbrackets).  Python was first publicly released in 1991.\n<\/p>\n<p>\n  The Python language is a very clean, elegant design with\nexcellent modularity features. It offers designers the option to write\nin an object-oriented style but does not force that choice (it can be\ncoded in a more classically procedural C-like way). It has a type\nsystem comparable in expressive power to\nPerl's,\nincluding dynamic container objects and association lists, but less\nidiosyncratic (actually, it is a matter of record that\nPerl's\n  object\nsystem was built in imitation of Python's). It even pleases\nLisp\n  hackers\nwith anonymous lambdas (function-valued objects that can be passed\naround and used by iterators). Python ships with the Tk toolkit, which\ncan be used to easily build GUI interfaces.\n<\/p>\n<p>\n  The standard Python distribution includes client classes for\nmost of the important Internet protocols (SMTP, FTP, POP3, IMAP, HTTP)\nand generator classes for HTML. It is therefore very well suited to\nbuilding protocol robots and network administrative plumbing. It is\nalso excellent for Web CGI work, and competes successfully with\nPerl\n  at the\nhigh-complexity end of that application area.\n<\/p>\n<p>\n  Of all the interpreted languages we describe, Python and\nJava\n  are the\ntwo most clearly suited for scaling up to large complex projects\nwith many cooperating developers. In many ways Python is simpler than\nJava, and its friendliness to rapid prototyping may give it an edge\nover Java for standalone use in applications that are neither hugely\ncomplex nor speed critical. An implementation of Python in Java,\ndesigned to facilitate mixed use of these two languages, is available\nand in production use; it is called Jython.\n<\/p>\n<p>\n  Python cannot compete with C\n  or C++\n  on raw execution speed (though using a\nmixed-language strategy on today's fast processors probably makes that\nrelatively unimportant). In fact it's generally thought to be the\nleast efficient and slowest of the major scripting languages, a price\nit pays for runtime type polymorphism.  Beware of rejecting Python on\nthese grounds, however; most applications do not actually need better\nperformance than Python offers, and even those that appear to are\ngenerally limited by external latencies such as network or disk waits\nthat entirely swamp the effects of Python's interpretive\noverhead. Also, by way of compensation, Python is exceptionally easy\nto combine with C, so performance-critical Python modules can be readily\ntranslated into that language for substantial speed gains.\n<\/p>\n<p>\n  Python loses in expressiveness to\nPerl\n  for small\nprojects and glue scripts heavily dependent on regular-expression\ncapability. It would be overkill for tiny projects, to which\nshell\n  or\nTcl\n  might be\nbetter suited.\n<\/p>\n<p>\n  Like Perl, Python has a well-established development community\nwith a\n  <a href=\"http:\/\/www.python.org\">\n   central website<\/a>\n  carrying a great many useful Python implementations, tools and\nextension modules.\n<\/p>\n<p>\n  The definitive Python reference is\n  <em>Programming Python<\/em><sup><a href=\"#Lutz\" title=\"[Lutz]\">[Lutz]<\/a><\/sup>. Extensive\non-line documentation on Python extensions is also available at the\nPython website.\n<\/p>\n<p>\n  Python programs tend to be quite portable between Unixes and even\nacross other operating systems; the standard library is powerful\nenough to significantly cut the use of nonportable helper programs.\nPython implementations are available for\nMicrosoft\n  operating systems and for\nMacOS. Cross-platform GUI development is\npossible with either Tk or two other toolkits. Python\/C applications\ncan be \u2018frozen\u2019, quasi-compiled into pure C sources that\nshould be portable to systems with no Python installed.\n<\/p>\n<p>\n  Summing up: Python's best side is that it encourages clean,\nreadable code and combines accessibility with scaling up well to large\nprojects.  Its worst side is inefficiency and slowness, not just\nrelative to compiled languages but relative to other scripting\nlanguages as well.\n<\/p>\n<h5 id=\"id2976960\"><a href=\"#id2976960\">\u00a7<\/a>A Small Python Case Study: imgsizer<\/h5>\n<p>\n  Imgsizer is a utility that rewrites WWW pages so that\nimage-inclusion tags get the right image size parameters\nautomatically plugged in (this speeds up page loading on many\nbrowsers). You can find sources and documentation in the URL\n  <a href=\"http:\/\/www.ibiblio.org\">\n   WWW tools\nsubdirectory of the ibiblio archive<\/a>.\n<\/p>\n<p>\n  <em>imgsizer<\/em> was originally written in Perl, and was a nearly ideal example of the sort of small, pattern-driven\n  text-processing tool at which Perl excels.  It was later translated to Python to take advantage of Python's library\n  support for HTTP fetching; this eliminated a dependency on an external page-fetching utility.  Observe the use of\n  file(1) and ImageMagick identify(1) as specialist tools for extracting the pixel sizes of images.\n<\/p>\n<p>\n  The dynamic string handling and sophisticated regular-expression matching required would have made <em>imgsizer<\/em>\n  quite painful to write in C or C++; that version would also have been much larger and harder to read. Java would have\n  solved the implicit memory-management problem, but is hardly more expressive than C or C++ at text pattern matching.\n<\/p>\n<h5 id=\"id2977085\"><a href=\"#id2977085\">\u00a7<\/a>A Medium-Sized Python Case Study: fetchmailconf<\/h5>\n<p>\n  In <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\"> Chapter 11<\/a> we examined the\n  <em>fetchmail<\/em> \/ <em>fetchmailconf<\/em> pair as an example of one way to separate implementation from interface.\n  Python's strengths are well illustrated by <em>fetchmailconf<\/em>. <\/p> <p> <em>fetchmailconf<\/em> uses the Tk toolkit\n  to implement a multi-panel GUI configuration editor (Python bindings also exist for GTK+ and other toolkits, but Tk\n  bindings ship with every Python interpreter).\n<\/p>\n<p>\n  In expert mode, the GUI supports editing of about sixty\nattributes divided among three panel levels. Attribute widgets include\na mix of checkboxes, radio buttons, text fields, and scrolling\nlistboxes. Despite this complexity, the first fully-functional version\nof the configurator took me less than a week to design and code,\ncounting the four days it took to learn Python and Tk.\n<\/p>\n<p>\n  Python excels at rapid prototyping of GUI interfaces, and (as\n  <em>fetchmailconf<\/em> illustrates) such prototypes are often deliverable.\nPerl\n  and\nTcl\n  have similar\nstrengths in this area (including the Tk toolkit, which was written\nfor Tcl) but are hard to control at the complexity level\n(approximately 1400 lines) of\n  <em>fetchmailconf<\/em>. Emacs\nLisp\n  is\nnot suited for GUI programming. Choosing Java\n  would have increased the complexity\noverhead of this programming task without delivering significant\nbenefits for this nonspeed-intensive application.\n<\/p>\n<h5 id=\"id2977218\"><a href=\"#id2977218\">\u00a7<\/a>A Large Python Case Study: PIL<\/h5>\n<p>\n  PIL, the Python Imaging Library, supports the manipulation of\nbitmap graphics. It supports many popular formats, including\nPNG, JPEG, BMP, TIFF,\nPPM, XBM, and GIF. Python programs can use it to convert and transform\nimages; supported transformations include cropping, rotation, scaling,\nand shearing. Pixel editing, image convolution, and color-space\nconversions are also supported. The PIL distribution includes Python\nprograms that make these library facilities available from the\ncommand line. Thus PIL can be used either for batch-mode image\ntransformation or as a strong toolkit over which to implement\nprogram-driven image processing of bitmaps.\n<\/p>\n<p>\n  The implementation of PIL illustrates the way Python can be\nreadily augmented with loadable object-code extensions to the Python\ninterpreter. The library core, implementing fundamental operations on\nbitmap objects, is written in\nC\n  for speed. The upper\nlevels and sequencing logic are in Python, slower but much easier to\nread and modify and extend.\n<\/p>\n<p>\n  The analogous toolkit would be difficult or impossible to write\nin Emacs Lisp\n  or shell, which don't have or don't document\na C extension interface at all.  Tcl\n  has a good C extension facility, but\nPIL would be an uncomfortably large project in Tcl.\nPerl\n  has such\nfacilities (Perl XS), but they are ad-hoc, poorly documented, complex,\nand unstable by comparison to Python's and use of them is\nrare. Java's\n  Native Method Interface appears to provide a facility roughly\ncomparable to Python's; PIL would probably have made a reasonable Java\nproject.\n<\/p>\n<p>\n  The PIL code and documentation is available at the\n  <a href=\"http:\/\/www.pythonware.com\/products\/pil\/\">\n   project\nwebsite<\/a>.\n<\/p>\n<h4 id=\"java\"><a href=\"#java\">\u00a7<\/a>Java<\/h4>\n<p>\n  The Java programming language was designed to be \u201cwrite\nonce, run anywhere\u201d and to support embedding interactive\nprograms (or\n  <em>applets<\/em>) in Web pages that would\nbe runnable from any browser. Thanks to a series of technical and\nstrategic blunders by its owner, Sun\nMicrosystems, it has failed in both its original\nobjectives.  But it is still sufficiently strong at both systems and\napplications programming to be seriously challenging\nC\n  and C++.  Java was announced in 1995.\n<\/p>\n<p>\n  Java is cleverly designed to capture the huge benefit of\nautomatic memory management and the lesser but not insignificant\nbenefit of supporting OO\n  design, while being far smaller and\nsimpler than C++. It retains a broadly C-like syntax\nthat most programmers will find comfortable. It includes support for\ncallouts to dynamically-loaded C and calling Java as an embedded\nlanguage from C.  Nor is it trivial that Sun\n  has done an excellent job of making\ngood Java documentation available on the Web.\n<\/p>\n<p>\n  Against Java, we can say that (compared to, say,\nPython) some\nparts of it appear over-complex and others deficient. Java's\nclass-visibility and implicit-scoping rules are baroque. The interface\nfacility avoids complex problems with multiple inheritance at the cost\nof being only slightly less difficult to understand and use in itself.\nFeatures like inner and anonymous classes can lead to very confusing\ncode. The absence of reliable destructor methods means that it is\ndifficult to ensure proper management of resources other than memory,\nsuch as mutexes and file locks.  Significant parts of the Unix\noperating-system facilities are not accessible from stock Java,\nincluding signals, poll, and select.  While Java's I\/O facilities are\nvery powerful, simple reading of text files is not simple.\n<\/p>\n<p>\n  There is a particularly invidious problem, resembling Windows\nDLL hell, with libraries.  Java has no method to manage\ndifferent library versions. This can create huge problems in\nenvironments like application servers, where the server might come\nequipped with one version of (say) an XML library, but the application\nships with a different (usually newer) version.  The only handle on\nsuch problems is the\n  <code>CLASSPATH<\/code> environment variable,\na source of chronic deployment problems.\n<\/p>\n<p>\n  Furthermore, Sun's handling of the Java language has been both\npolitically and technically obtuse. Java's first GUI toolkit, AWT, was\na mess that had to be essentially replaced. Withdrawing the language\nfrom ECMA\/ISO standardization further nettled many developers already\nupset by features of the Sun Community Source License (SCSL).\nRestrictions in the SCSL continue to hamper open-source\nimplementations of Java 1.2 and their J2EE (Java 2 Enterprise Edition)\nspecification.  This compromises Java's original objective of\nuniversal portability.\n<\/p>\n<p>\n  Sadly, browser applets are dead.\nMicrosoft's\n  decision not to support Java 1.2 in Internet Explorer effectively\nkilled them.  However, Java seems to have found a secure niche in the\ncomputing ecology, for \u2018servlets\u2019 running within Web\napplication servers.  It has also become commonly used for a lot of\nin-house corporate programming not directly tied to databases or\nwebservers.  It has become major competition for both Microsoft's\nASP\/COM platform and Perl CGIs.  Finally, it is in widespread and\nincreasing use as a language for teaching introductory programming (a\nrole for which it is extremely well suited).\n<\/p>\n<p>\n  Overall, we can fairly judge Java to be superior to\nC++\n  (which is\nboth far more complex and does less to attack the memory-management\nproblem) for all but systems programming and the most speed-critical\napplications.  Experience seems to show that Java programmers are\nsomewhat less likely to fall into the trap of excessive\nOO\n  layering than are C++ programmers, though this remains a significant\nproblem.\n<\/p>\n<p>\n  How Java will fare in equilibrium with the other languages\nwe describe here is unclear as yet, and may depend largely on project\nscale. We may expect its proper niche to resemble\nPython's. Like Python, it cannot compete with\nC\n  or\nC++\n  on raw\nexecution speed, nor against Perl\n  on small projects that use\npattern-driven editing heavily. It is (more definitely than Python)\noverkill for small projects. We may guess that Python will have an\nedge in smaller projects and Java in larger ones, but the verdict of\nexperience is not yet in.\n<\/p>\n<p>\n  The best single reference on paper is probably\n  <em>Java In A Nutshell<\/em><sup><a href=\"#FlanaganJava\" title=\"[FlanaganJava]\">[FlanaganJava]<\/a><\/sup>, but this is\nnot the best tutorial introduction; that would probably be\n  <em>Thinking in Java<\/em> <sup><a href=\"#Eckel\" title=\"[Eckel]\">[Eckel]<\/a><\/sup>. Trails to all the world's Java websites begin at\n  <a href=\"http:\/\/java.sun.com\">\n   Sun's Java site<\/a>, which also\nhas complete HTML documentation available for download for free. The\n  <a href=\"http:\/\/dmoz.org\/Computers\/Programming\/Languages\/Java\/\">\n   Open\nDirectory Java Page<\/a>\n  also collects useful Java links.\n<\/p>\n<p>\n  Java implementations are available for all Unixes, for\nMicrosoft operating systems, MacOS, and many other platforms.\n<\/p>\n<p>\n  Sources for Kaffe, an open-source Java implementation with class\nlibraries conforming to most of JDK 1.1 and portions of JDK 1.2,\nare available at the\n  <a href=\"http:\/\/www.kaffe.org\/\">\n   Kaffe project site<\/a>.\n<\/p>\n<p>\n  There is a Java front end for GCC. GCJ can compile Java code to\neither Java bytecode or native code, and can compile Java bytecode to\nnative code as well. It comes packaged with open-source class\nlibraries that implement most of JDK 1.2, and a Java bytecode\ninterpreter called\n  <em>gij<\/em>. Details are at the\n  <a href=\"http:\/\/gcc.gnu.org\/java\/\">\n   GCJ project page<\/a>.\n<\/p>\n<p>\n  There is a Java IDE for Emacs at the\n  <a href=\"http:\/\/jdee.sunsite.dk\/\">\n   JDEE project site<\/a>.\n<\/p>\n<p>\n  Java portability is excellent at the language level.  Incomplete\nlibrary implementations (especially older JDK 1.1 versions that don't\nsupport the newer JDK 1.2) can be an issue.\n<\/p>\n<p>\n  Java's best side is that it comes close enough to\nachieving write-once-run-anywhere to be useful as an OS-independent\nenvironment of its own.  Its worst side is that the Java 1\/Java 2\nsplit compromises that goal in deeply frustrating ways.\n<\/p>\n<h5 id=\"id2977737\"><a href=\"#id2977737\">\u00a7<\/a>Case Study: FreeNet<\/h5>\n<p>\n  Freenet is a peer-to-peer networking project that is intended to\nmake censorship and content suppression\nimpossible.<sup><a href=\"#ftn.id2977747\" id=\"id2977747\">[127]<\/a><\/sup>\n  Freenet developers envision the\nfollowing applications:\n<\/p>\n <div><ul><li>\n    <p>\n     Uncensorable dissemination of controversial information: Freenet\nprotects freedom of speech by enabling anonymous and uncensorable\npublication of material ranging from grassroots alternative journalism\nto banned expos\u00e9s.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Efficient distribution of high-bandwidth content: Freenet's\nadaptive caching and mirroring is being used to distribute Debian\nLinux software updates.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Universal personal publishing: Freenet enables anyone to have a\nwebsite, without space restrictions or compulsory advertising, even if\nthe would-be webmaster doesn't own a computer.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  Freenet addresses these goals by providing a virtual space in\nwhich to publish documents that is not tied to any specific machine.\nPublished information and Freenet's own internal data indexes are\nreplicated and distributed across the network in such a way that even\nFreenet administrators don't know at any given time where all the physical\ncopies are located.  Privacy for people browsing or submitting to\nFreenet is protected by strong cryptography.\n<\/p>\n<p>\n  Java was a good choice for this project for at least two reasons.\nFirst: the goals of the project put a heavy premium on having\ncompatible implementations on the widest possible variety of machines,\nso Java's high portability is a dominating advantage.  Second: the\nnature of the project is such that the network API is important,\nand Java has a strong one built in.\n<\/p>\n<p>\n  C is traditional for infrastructure projects of this kind that\nhave high performance demands, but the lack of a standardized network\nAPI would have made porting a significant difficulty.  C++ would have\nhad the same difficulty.  Tcl, Perl, or Python\n  might have reduced the porting burden,\nbut at a greater cost in performance.  Emacs Lisp\n  would\nhave been painfully slow and totally inappropriate.\n<\/p>\n<h4 id=\"emacs_lisp_language\"><a href=\"#emacs_lisp_language\">\u00a7<\/a>Emacs Lisp<\/h4>\n<p>\n  Emacs Lisp is a scripting language\n  used to program the behavior of the\nEmacs text editor.  Its first public release was in 1984.\n<\/p>\n<p>\n  Emacs Lisp is not a general-purpose language in quite the same\nway as the others surveyed in this chapter; while it is powerful\nenough to theoretically be used as such, it is traditionally employed\nonly to write control programs for the\nEmacs editor itself and does not\ncommunicate as fluently with other software as would a modern\nscripting language.\n<\/p>\n<p>\n  Nevertheless, there is a significant range of applications in\nwhich Emacs Lisp is more effective than anything else. Many of these\nhave to do with providing a front-end for development tools such as\nthe C compiler and linker,\nmake(1),\nversion-control systems, and symbolic debuggers; we'll discuss these\nin\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>.\n<\/p>\n<p>\n  More generally, Emacs is to pattern-\nor syntax-directed\n  <em>interactive<\/em>\n  editing what\nPerl\n  is to\npattern-directed\n  <em>batch<\/em>\n  editing. Any application\nthat involves interactively hacking a special file format or text\ndatabase is an excellent candidate to be prototyped (and possibly\ndelivered) as an Emacs mode (an Emacs Lisp\nprogram that specializes the editor's behavior).\n<\/p>\n<p>\n  Emacs Lisp is also valuable for building applications that have\nto be closely integrated with a text editor, or that function\nprimarily as text browsers with some editing capability. User agents\nfor email and Usenet\n  news fall in this category. So do certain kinds of database front\nends.\n<\/p>\n<p>\n  Emacs Lisp is a Lisp. It follows as the night the day that it\nmanages memory automatically and is far more elegant and powerful than\nmost conventional languages, or indeed most\n  <em>unconventional<\/em>\n  languages; it can compete with\nJava\n  or\nPython\n  on\nthis level and laugh at C\n  or C++, Perl, shell or Tcl. Lisp's perennial problem of lacking a\nstandardized OS binding for portability is solved by the Emacs core,\nwhich in effect\n  <em>is<\/em>\n  its OS binding.\n<\/p>\n<p>\n  Lisp's other perennial problem \u2014 being a resource hog\n\u2014 is no longer a real issue on modern machines. Parody\nexpansions like \u2018Emacs Makes A Computer Slow\u2019 and\n\u2018Eventually Munches All Computer Storage\u2019 used to be\ncommon (in fact the Emacs distribution itself includes a list of\nthem). But many other commonly used categories of programs (such as\nWeb browsers) have nowadays grown larger and more complex than Emacs,\nwhich has come to appear rather moderate by comparison.\n<\/p>\n<p>\n  The definitive Emacs Lisp reference is\n  <em>\n   The GNU Emacs\nLisp Reference Manual\n<\/em>, which may be browseable through your\nEmacs's \u2018info\u2019 help system. If not, it can be downloaded from the\n  <a href=\"ftp:\/\/ftp.gnu.org\/pub\/gnu\">\n   FSF FTP\n   site<\/a>. If you\nfind that impenetrable,\n  <em>Writing GNU Emacs Extensions<\/em><sup><a href=\"#Glickstein\" title=\"[Glickstein]\">[Glickstein]<\/a><\/sup> may help.\n<\/p>\n<p>\n  Portability of Emacs Lisp programs is excellent.\nEmacs implementations are available for all\nUnixes, the Microsoft operating systems, and Mac\nOS.\n<\/p>\n<p>\n  Summing up: Emacs Lisp's best point is that it combines an\nexcellent base language, Lisp, with powerful domain primitives for\ntext manipulation.  Its worst point is poor performance and\ndifficulties using it in communication with other programs.\n<\/p>\n<p>\n  For more information, see the discussion of\n  <em>Emacs<\/em> under editors in the next\nchapter.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2969184\" id=\"ftn.id2969184\">[125]<\/a><\/sup>\n  The last C++ standard,\ndating from 1998, was widely implemented but weak, especially in the \narea of libraries.\n<\/p><p><sup><a href=\"#id2969353\" id=\"ftn.id2969353\">[126]<\/a><\/sup>\n  See Tom Christiansen's essay\n  <em>Csh Programming Considered Harmful<\/em>, which should be readily\nfindable via Web search.\n<\/p><p><sup><a href=\"#id2977747\" id=\"ftn.id2977747\">[127]<\/a><\/sup>\n  There is a\n  <a href=\"http:\/\/freenetproject.org\">\n   Freenet project\nwebsite<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch14s05\"><a href=\"#ch14s05\">\u00a7<\/a>Trends for the Future<\/h3>\n<p>\n   Table 14.1 gives a rough indication of today's distribution of language usage.  We give figures from both\n   SourceForge<sup><a href=\"#ftn.id2978184\" id=\"id2978184\">[128]<\/a><\/sup> and Freshmeat,<sup><a href=\"#ftn.id2978197\" id=\"id2978197\">[129]<\/a><\/sup> the two most important new-release sites, as of March 2003.\n<\/p>\n<p>\n  The SourceForge figures are soft in several ways: Notably,\nSourceForge's query interface doesn't permit filtering on OS and\nlanguage simultaneously, so some of these numbers represent MacOS and\nWindows projects.  The effect is probably to exaggerate\nC++\n  and\nJava's\n  share\nconsiderably. However, Unix-based projects dominate sufficiently (by\nabout a 3:1 ratio) so that the effect on the figures for languages other\nthan these is probably not too distorting.\n<\/p>\n<p>\n  The Freshmeat sample is smaller, but the site hosts only\nUnix-based releases \u2014 and it counts actual releases, not the\nhuge clutter of failed and inactive SourceForge projects.  It is thus\ninteresting that the population figures track SourceForge's by about a\n1:2 ratio except in precisely the cases (C++ and Java) where we would\nexpect them to be out of proportion because of the absence of Windows\nprojects.\n<\/p>\n <table class=\"border\"><caption>Table 14.1. Language choices.<\/caption>\n   <thead><tr><th>\n      Language\n     <\/th>\n     <th>\n      SourceForge\n     <\/th>\n     <th>\n      Freshmeat\n     <\/th>\n    <\/tr><\/thead><tbody><tr><td>\n      C\n     <\/td>\n     <td>\n      10296\n     <\/td>\n     <td>\n      4845\n     <\/td>\n    <\/tr><tr><td>\n      C++\n     <\/td>\n     <td>\n      9880\n     <\/td>\n     <td>\n      2098\n     <\/td>\n    <\/tr><tr><td>\n      Shell\n     <\/td>\n     <td>\n      1058\n     <\/td>\n     <td>\n      487\n     <\/td>\n    <\/tr><tr><td>\n      Perl\n     <\/td>\n     <td>\n      4394\n     <\/td>\n     <td>\n      2508\n     <\/td>\n    <\/tr><tr><td>\n      Tcl\n     <\/td>\n     <td>\n      649\n     <\/td>\n     <td>\n      328\n     <\/td>\n    <\/tr><tr><td>\n      Python\n     <\/td>\n     <td>\n      2222\n     <\/td>\n     <td>\n      948\n     <\/td>\n    <\/tr><tr><td>\n      Java\n     <\/td>\n     <td>\n      8032\n     <\/td>\n     <td>\n      1900\n     <\/td>\n    <\/tr><tr><td>\n      Emacs Lisp\n     <\/td>\n     <td>\n      ?\n     <\/td>\n     <td>\n      31\n     <\/td>\n    <\/tr><\/tbody><\/table><p>\n  This chapter was first drafted in 1997; at time of writing\nit is mid-2003. That is a long enough time base that the relative\npositions of the languages we surveyed above have changed somewhat\nsince first writing, indicating adoption trends that may suggest\nwhat their futures will be like. (Community size is an important\npredictor of the quality and amount of work that will go into\nimproving the most-used open-source implementations of these\nlanguages; both growth and decline tend to be\nself-reinforcing.)\n<\/p>\n<p>\n  Broadly speaking, C\n  and C++\n  and Emacs Lisp\n  have remained stable across the\n1997-2003 time period, appealing to much the same constituencies in\n2003 as they did in 1997. C has gained slowly at the expense of older\nconventional languages such as FORTRAN; C++, on the other hand, has\nlost some ground to Java.\n<\/p>\n<p>\n  Perl\n  usage has grown respectably, but the language itself has been stagnant\nfor some time. Perl's internals are notoriously grubby; it's been\nunderstood for years that the language's implementation needs to be\nrewritten from scratch, but an attempt in 1999 failed and another\nseems presently stalled in mid-2003.  Nevertheless, Perl is still\nthe 800-pound gorilla of scripting\nlanguages, and dominates Web scripting and\nCGI.\n<\/p>\n<p>\n  Tcl\n  has\nbeen in a period of relative decline, or at least of diminishing\nvisibility.  In 1996 a widely-reported and plausible estimate of\ncommunity sizes held that for every Python\n  hacker there were five\nTcl\n  hackers and\ntwelve Perl\n  hackers. Today the SourceForge figures suggest those ratios are about\n3:1:7.  However, Tcl is reported to be very widely used for scripting\nof specialized components in several industries, including electronic\ndesign automation, radio and television broadcasting, and the film\nindustry.\n<\/p>\n<p>\n  Python\n  has risen in popularity as rapidly as Tcl\n  has fallen. Though the\nPerl\n  community\nis still twice the size of Python's, a visible tendency of the\nbrightest Perl hackers to migrate to Python has been rather ominous\nfor the former language \u2014 especially as there is no migration at\nall in the opposite direction.\n<\/p>\n<p>\n  Java\n  has\nbecome widely used at sites already invested in Sun\nMicrosystems\n  technology and is in increasing deployment as an instructional\nlanguage in undergraduate computer science curricula. Elsewhere,\nhowever, it is only marginally more popular than it was in 1997. Sun's\ndetermination to stick to a proprietary licensing model has prevented\nthe major breakout many observers then predicted; under\nLinux\n  and in the wider open-source community Java has not made the headway\nagainst C that it has elsewhere.\n<\/p>\n<p>\n  No new general-purpose language has emerged to seriously\nchallenge those we've surveyed here. PHP is making inroads in Web\ndevelopment, challenging Perl CGIs (as well as ASP and server-side\nJava) but is\nalmost never used for standalone programming.  Non-Emacs\nLisp\n  dialects,\na once-promising area that seemed headed for a renaissance in the\nmid-1990s, have continued to fade. Recent efforts such as Ruby (a sort\nof Python-Perl-Smalltalk cross developed in Japan) and Squeak (an\nopen-source Smalltalk port) look promising, but have so far neither\nattracted hackers far outside their development groups nor\ndemonstrated staying power.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2978184\" id=\"ftn.id2978184\">[128]<\/a><\/sup><a href=\"http:\/\/sourceforge.net\/softwaremap\/trove_list.php?form_cat=160\">\n   Query\nfor statistics<\/a>.\n<\/p><p><sup><a href=\"#id2978197\" id=\"ftn.id2978197\">[129]<\/a><\/sup><a href=\"http:\/\/freshmeat.net\/browse\/160\/?topic_id=160\">\n   Query for\nstatistics<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch14s06\"><a href=\"#ch14s06\">\u00a7<\/a>Choosing an X Toolkit<\/h3>\n<p>\n  An issue related to choice of language is choice of X toolkit\nfor GUI programming.  Recall the discussion in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>\n  of how X separates mechanism from\npolicy.  Each possible choice of toolkit will give you a slightly\ndifferent look and feel.\n<\/p>\n<p>\n  Your choice of X toolkit will be connected to your choice of\napplication language in two ways: first, because some languages ship\nwith a binding to a preferred toolkit, and second because some\ntoolkits only have bindings to a limited set of languages.\n<\/p>\n<p>\n  Java, of\ncourse, has its own cross-platform toolkits built in, so your choice\nwill be between AWT (universally deployed) and Swing (more capable,\nmore complex, slower, and only in JDK 1.2\/Java 2). The remainder of\nthis section focuses on the other languages we have\nsurveyed. Similarly, if you're using Tcl, Tk comes bundled.  There\nprobably is not a lot of point in evaluating alternatives.\n<\/p>\n<p>\n  The once-ubiquitous Motif toolkit is effectively dead. It\ncouldn't keep up with the newer toolkits distributed without license fees\nor restrictions.  These attracted more developer effort until they\nsurged past closed-source toolkits in capability and features;\nnowadays, the competition is all in open source.\n<\/p>\n<p>\n  The four toolkits to consider seriously in 2003 are Tk,\nGTK, Qt, and wxWindows, with GTK and Qt being the clear front runners.\nAll four have ports on MacOS and Windows, so any choice will give you\nthe capability to do cross-platform development.\n<\/p>\n<p>\n  The Tk toolkit is the oldest of the four and has the advantage\nof incumbency; it's native in Tcl and bindings to it are shipped with\nthe stock version of Python.  Libraries to provide language\nbindings to Tk are generally available for C\n  and C++.  Unfortunately, Tk also shows its age\nin that its standard widget set is both limited and rather ugly.  On\nthe other hand, the Tk Canvas widget has capabilities that other\ntoolkits still match only with difficulty.\n<\/p>\n<p>\n  GTK began life as a replacement for Motif, and was invented to\nsupport the GIMP.  It is\nnow the preferred toolkit of the GNOME project and is used by hundreds\nof GNOME applications. The native API is C; bindings are available for\nC++,\nPerl, and\nPython, but\ndo not ship with the stock language distributions.  It's the only one\nof these four with a native C binding.\n<\/p>\n<p>\n  Qt is a toolkit associated with the KDE project.  It is natively\na C++\n  library;\nbindings are available for Python\n  and Perl\n  but do not ship with the stock\ninterpreters. Qt has a reputation for having the best-designed and\nmost expressive API of these four, but adoption was initially hindered\nby controversy over early versions of the Qt license and was further\nslowed down by the fact that a C binding was slow in coming.\n<\/p>\n<p>\n  wxWindows is also natively C++\n  with bindings available in\nPerl\n  and\nPython.  The\nwxWindows developers emphasize their support for cross-platform\ndevelopment heavily and appear to regard it as the main selling point\nof the toolkit.  Another selling point is that wxWindows is actually a\nwrapper around the native (GTK, Windows, and MacOS 9) widgets on each\nplatform, so applications written using it retain a native look and\nfeel.\n<\/p>\n<p>\n  As of mid-2003 few detailed comparisons have been written, but a Web search for \u201cX toolkit comparison\u201d may turn up\n  some useful hits. Table 14.2 summarizes the state of play.\n<\/p>\n  <table class=\"border\"><caption>Table 14.2. Summary of X Toolkits.<\/caption>\n   <thead><tr><th rowspan=\"2\">\n      Toolkit\n     <\/th>\n     <th rowspan=\"2\">\n      Native language\n     <\/th>\n     <th rowspan=\"2\">\n      Shipped with\n     <\/th>\n     <th colspan=\"5\">\n      Bindings\n     <\/th>\n    <\/tr><tr><th>\n      C\n     <\/th>\n     <th>\n      C++\n     <\/th>\n     <th>\n      Perl\n     <\/th>\n     <th>\n      Tcl\n     <\/th>\n     <th>\n      Python\n     <\/th>\n    <\/tr><\/thead><tbody><tr><td>\n      Tk\n     <\/td>\n     <td>\n      Tcl\n     <\/td>\n     <td>\n      Tcl, Python\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n    <\/tr><tr><td>\n      GTK\n     <\/td>\n     <td>\n      C\n     <\/td>\n     <td>\n      Gnome\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n    <\/tr><tr><td>\n      Qt\n     <\/td>\n     <td>\n      C++\n     <\/td>\n     <td>\n      KDE\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n    <\/tr><tr><td>\n      wxWindows\n     <\/td>\n     <td>\n      C++\n     <\/td>\n     <td>\n      \u2014\n     <\/td>\n     <td>\n      \u2014\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n     <td>\n      Y\n     <\/td>\n    <\/tr><\/tbody><\/table><p>\n  Architecturally, these libraries are all written at about the\nsame abstraction level.  GTK and Qt use a slot-and-signal apparatus\nfor event-handling so similar that ports between them have been\nreported to be almost trivial.  Your choice among them will probably\nbe conditioned more by the availability of bindings to your chosen\ndevelopment language than anything else.\n<\/p>\n\n<h3 id=\"toolschapter\"><a href=\"#toolschapter\">\u00a7<\/a>Chapter 15. Tools<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Unix is user-friendly \u2014 it's just choosy about who its\nfriends are.\n  <\/p><div>\n  \n   \u2013 Anonymous\n  \n <\/div><\/blockquote>\n\n<h3 id=\"ch15s01\"><a href=\"#ch15s01\">\u00a7<\/a>A Developer-Friendly Operating System<\/h3>\n<p>\n  Unix has a long-established reputation as a good environment to\ndevelop under. It's well equipped with tools written by programmers\nfor programmers. These automate away many of the grubby little\ntasks that would otherwise distract you from concentrating on the\nmost important (and most enjoyable!) aspect of development\u2014 your design.\n<\/p>\n<p>\n  While all the tools you'll need are there and individually well\ndocumented, they're not knit together by an integrated development\nenvironment (IDE). Finding and assembling them into a kit that\nsuits your needs has traditionally taken considerable effort.\n<\/p>\n<p>\n  If you're used to a good IDE \u2014 the kind of GUI-driven\ncombination of editor, configuration-manager, compiler, and\ndebugger now common on Macintosh and Windows systems \u2014 the Unix\napproach may seem casual, murky, and primitive. But there's\nactually method in it.\n<\/p>\n<p>\n  IDEs make a lot of sense for single-language programming in a\ntool-poor environment. If what you're doing is confined to grinding\nout C\n  or\nC++\n  code by hand\nand the yard, they're quite appropriate.  Under Unix, however, your\nlanguages and implementation options are a lot more varied. It's\ncommon to use multiple code generators, custom configurators, and many\nother standard and custom tools.\n<\/p>\n<p>\n  IDEs do exist under Unix (there are several good open-source\nones, including emulations of the major Macintosh and Windows\nIDEs). But it's difficult to control an open-ended variety of\nprogramming tools with them, and they're not much used. Unix\nencourages a more flexible style, one less exclusively centered on\nthe edit\/compile\/debug loop.\n<\/p>\n<p>\n  In this chapter we introduce you to the tactics of\ndevelopment under Unix \u2014 building code, managing code\nconfigurations, profiling, debugging, and automating away a lot\nof the drudgery associated with these tasks so you can concentrate on\nthe fun parts. As usual, the exposition focuses more on the\narchitectural picture than the how-to details. When you\n  <em>want<\/em>\n  how-to details, most of the tools in this\nchapter are well described in\n  <em>Programming with GNU Software<\/em><sup><a href=\"#Loukides-Oram\" title=\"[Loukides-Oram]\">[Loukides-Oram]<\/a><\/sup>.\n<\/p>\n<p>\n  Many of these tools automate things that you could do yourself\nby hand, albeit more slowly and with a higher error rate.  The\none-time cost of climbing the learning curve should be more than\npaid off by the ability to write programs more efficiently, and\nspend less attention on low-level details and more on design.\n<\/p>\n<p>\n  Unix programmers traditionally learn how to use these tools by\nosmosis from other programmers, and by exploration over a period of\nyears. If you're a novice, pay careful attention; we're going to try\nto jump you over a big section of the Unix learning curve by showing\nyou what is possible right at the outset. If you are an experienced\nUnix programmer in a hurry, you can skip this chapter \u2014 but\nmaybe you shouldn't. There might just be some bit of useful lore here\nthat even you don't know.\n<\/p>\n\n<h3 id=\"ch15s02\"><a href=\"#ch15s02\">\u00a7<\/a>Choosing an Editor<\/h3>\n<p>\n  The first and most basic tool of development is a text editor\nsuitable for modifying and writing programs.\n<\/p>\n<p>\n  Literally dozens of text editors are available under Unix;\nwriting one seems to be one of the standard finger exercises for\nbudding open-source\nhackers. Most of\nthese are ephemera, not suitable for extended use by anyone other than\ntheir authors. A few are emulations of non-Unix editors, useful as\ntransition aids for programmers used to other operating systems. You\ncan browse through a wide variety at SourceForge or ibiblio or any\nother major open-source archive.\n<\/p>\n<p>\n  For serious editing work, two editors completely dominate the\nUnix programming scene. Each is available in a couple of minor variant\nimplementations, but has a standard version you can rely on finding on\nany modern Unix system. These two editors are\n  <em>vi<\/em> and\n  <em>Emacs<\/em>.\nWe discussed them in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">\n   Chapter 13<\/a>\n  as part of our discussion of the right\nsize of software.\n<\/p>\n<p>\n  As we noted in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">\n   Chapter 13<\/a>, these two editors express sharply\ncontrasting design philosophies, but both are extremely popular and\ncommand great loyalty from identifiable core user populations. Surveys\nof Unix programmers consistently indicate about a 50\/50 split between\nthem, with all other editors barely registering.\n<\/p>\n<p>\n  In our earlier examinations of\n  <em>vi<\/em> and\n  <em>Emacs<\/em>, we were primarily concerned with\ntheir optional complexity and the surrounding design-philosophy issues.\nMany other things are worth knowing about these editors, both as a\nmatter of practicality and of Unix cultural literacy.\n<\/p>\n<h4 id=\"vi_literacy\"><a href=\"#vi_literacy\">\u00a7<\/a>Useful Things to Know about vi<\/h4>\n<p>\n  The name of\n  <em>vi<\/em> is an abbreviation for\n\u201cvisual editor\u201d and is pronounced\n  \n   \/vee eye\/\n  \n  (not\n  \n   \/vie\/\n  \n  and\n  <em>definitely<\/em>\n  not\n  \n   \/siks\/\n  \n  !).\n<\/p>\n<p>\n  <em>vi<\/em> was not quite the earliest\nscreen-oriented editor; that palm goes to the Rand editor,\n  <em>re<\/em>, that ran on Version 6 Unix in the 1970s. But\n  <em>vi<\/em> is the longest-lived screen-oriented\neditor built for Unix that is still in use, and is a hallowed part of\nUnix tradition.\n<\/p>\n<p>\n  The original\n  <em>vi<\/em> was the version\npresent in the earliest BSD software distributions\n  beginning in 1976; it is now obsolete. Its\nreplacement was \u2018new vi\u2019 which shipped with 4.4BSD and is\nfound on modern 4.4BSD variants such as BSD\/OS, FreeBSD, and NetBSD\nsystems. There are several variants with extended features, notably\n  <em>vim<\/em>,\n  <em>vile<\/em>,\n  <em>elvis<\/em>, and\n  <em>xvi<\/em>;\nof these\n  <em>vim<\/em> is probably the most popular\nand is found on many Linux\n  systems. All the variants are rather\nsimilar and share a core command set unchanged from the original\n  <em>vi<\/em>.\n<\/p>\n<p>\n  Ports of\n  <em>vi<\/em> are available for the\nWindows operating systems\n  and MacOS.\n<\/p>\n<p>\n  Most introductory Unix books include a chapter describing basic <em>vi<\/em> usage. One place a <em>vi<\/em> FAQ is\n  available is the <a href=\"http:\/\/www.faqs.org\/faqs\/editor-faq\/vi\/\"> Editor FAQ\/vi<\/a>; you can find many other copies\n  with a WWW keyword search for page titles including \u201cvi\u201d and \u201cFAQ\u201d.\n<\/p>\n<h4 id=\"id2979540\"><a href=\"#id2979540\">\u00a7<\/a>Useful Things to Know about Emacs<\/h4>\n<p>\n  <em>Emacs<\/em> stands for \u2018EDiting\nMACroS\u2019 (pronounce it\n  \n   \/ee\u00b4\u00b7maks\/\n  ).  It was originally\nwritten in the late 1970s as a set of macros in an editor called TECO,\nthen reimplemented several times in different ways.  In an amusing\ntwist, modern Emacs implementations include a TECO emulation\nmode.\n<\/p>\n<p>\n  In our earlier discussion of editors and optional complexity, we\nnoted that many people consider Emacs\nexcessively heavyweight.  However, investing the time to learn it can\nyield rich rewards in productivity.  Emacs\nsupports many powerful editing modes that offer help with the syntax\nof various programming languages and markups.  We'll see later in this\nchapter how Emacs can be used in\ncombination with other development tools to give capabilities\ncomparable to (and in many ways surpassing) those of conventional\nIDEs.\n<\/p>\n<p>\n  The standard Emacs, universally available on modern Unixes, is <em>GNU Emacs<\/em>; this is what generally runs if you\n  type <code>emacs<\/code> to a Unix shell prompt. GNU Emacs sources and documentation are available at the <a href=\"ftp:\/\/gnu.org\/pub\/gnu\"> Free Software Foundation archive site<\/a>.\n<\/p>\n<p>\n  The only major variant is called <em>XEmacs<\/em>; it has a better X interface but otherwise quite similar capabilities\n  (it forked from Emacs 19). <em>XEmacs<\/em> has a <a href=\"http:\/\/www.xemacs.org\"> home page<\/a>. Emacs (and Emacs\n  Lisp) is universally available under modern Unixes. It has been ported to MS-DOS (where it works poorly) and Windows\n  95 and NT (where it is said to work reasonably well).\n<\/p>\n<p>\n  Emacs includes its own interactive\ntutorial and very complete on-line documentation; you'll find\ninstructions on how to invoke both on the default\nEmacs startup screen. A good introduction\non paper is\n  <em>Learning GNU Emacs<\/em> <sup><a href=\"#Cameron\" title=\"[Cameron]\">[Cameron]<\/a><\/sup>.\n<\/p>\n<p>\n  The keystroke commands used in the Unix ports of\nNetscape\/Mozilla and Internet Explorer text windows (in forms and the\nmailer) are copied from the stock Emacs\nbindings for basic text editing.  These bindings are the closest thing\nto a cross-platform standard for editor keystrokes.\n<\/p>\n<h4 id=\"id2979678\"><a href=\"#id2979678\">\u00a7<\/a>The Antireligious Choice: Using Both<\/h4>\n<p>\n  Many people who regularly use both\n  <em>vi<\/em> and\nEmacs tend to use them for different\nthings, and find it valuable to know both.\n<\/p>\n<p>\n  In general,\n  <em>vi<\/em> is best for small jobs\n\u2014 quick replies to mail, simple tweaks to system configuration,\nand the like. It is especially useful when you're using a new system\n(or a remote one over a network) and don't have your\nEmacs customization files handy.\n<\/p>\n<p>\n  Emacs comes into its own for extended\nediting sessions in which you have to handle complex tasks, modify\nmultiple files, and use results from other programs during the\nsession. For programmers using X on their console (which is typical on\nmodern Unixes), it's normal to start up\nEmacs shortly after login time in a large\nwindow and leave it running forever, possibly visiting dozens of files\nand even running programs in multiple Emacs\nsubwindows.\n<\/p>\n\n<h3 id=\"ch15s03\"><a href=\"#ch15s03\">\u00a7<\/a>Special-Purpose Code Generators<\/h3>\n<p>\n  Unix has a long-standing tradition of hosting tools that are\nspecifically designed to generate code for various special purposes.\nThe venerable monuments of this tradition, which go back to Version 7\nand earlier days, and were actually used to write the original\nPortable C Compiler back in the 1970s, are\nlex(1)\nand\nyacc(1).\nTheir modern, upward-compatible successors are\nflex(1)\nand\nbison(1),\npart of the GNU\n  toolkit and still heavily used today.  These programs have set an\nexample that is carried forward in projects like GNOME's\n  <em>Glade<\/em> interface builder.\n<\/p>\n<h4 id=\"id2979797\"><a href=\"#id2979797\">\u00a7<\/a>yacc and lex<\/h4>\n<p>\n  <em>yacc<\/em> and\n  <em>lex<\/em> are tools for generating language\nparsers.  We observed in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  that your first minilanguage is all\ntoo likely to be an accident rather than a design.  That accident is\nlikely to have a hand-coded parser that costs you far too much\nmaintenance and debugging time \u2014 especially if you have not\nrealized it is a parser, and have thus failed to properly separate it\nfrom the remainder of your application code.  Parser generators are\ntools for doing better than an accidental, ad-hoc implementation; they\ndon't just let you express your grammar specification at a higher\nlevel, they also wall off all the parser's implementation complexity\nfrom the rest of your code.\n<\/p>\n<p>\n  If you reach a point where you are planning to implement a\nminilanguage from scratch, rather than by extending or embedding an\nexisting scripting language\n  or parsing XML,\n  <em>yacc<\/em> and\n  <em>lex<\/em> will probably be your most important tools after your\nC\n  compiler.\n<\/p>\n<p>\n  <em>lex<\/em> and\n  <em>yacc<\/em> each generate code for a single\nfunction \u2014 respectively, \u201cget a token from the input\nstream\u201d and \u201cparse a sequence of tokens to see if it\nmatches a grammar\u201d.  Usually, the\n  <em>yacc<\/em> -generated parser function calls a\nLex-generated tokenizer function each time it wants to get another\ntoken.  If there are no user-written C callbacks at all in the\n  <em>yacc<\/em> -generated parser, all it will do is a\nsyntax check; the value returned will tell the caller if the input\nmatched the grammar it was expecting.\n<\/p>\n<p>\n  More usually, the user's C code, embedded in the generated\nparser, populates some runtime data structures as a side-effect of\nparsing the input.  If the minilanguage is declarative, your\napplication can use these runtime data structures directly.  If your\ndesign was an imperative minilanguage, the data structures might\ninclude a parse tree which is immediately fed to some kind of\nevaluation function.\n<\/p>\n<p>\n  <em>yacc<\/em> has a rather ugly interface,\nthrough exported global variables with the name prefix\n  <code>yy_<\/code>.  This is because it predates structs in C; in\nfact,\n  <em>yacc<\/em> predates C itself; the first\nimplementation was written in C's predecessor B.  The crude\nthough effective algorithm\n  <em>yacc<\/em> -generated\nparsers use to try to recover from parse errors (pop tokens until an\nexplicit error production is matched) can also lead to\nproblems, including memory leaks.\n<\/p>\n <blockquote>\n  <p>\n   If you are building parse trees, using malloc to make nodes, and\nyou start popping things off the stack in error recovery, you don't\nget to recover (free) the storage.  In general, Yacc can't do it,\nsince it doesn't know enough about what's on the stack.  If the yacc\nparser were in C++, it could assume that the values were classes and\n\u201cdestruct\u201d them.  In \u201creal\u201d compilers, parse\ntree nodes are generated using an arena-based allocator, so the nodes\ndon't leak, but there is a logical leak anyway that needs to be\nthought about to make industrial-strength error recovery.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  <em>lex<\/em> is a lexical analyzer generator.\nIt's a member of the same functional family as\ngrep(1)\nand\nawk(1),\nbut more powerful because it enables you to arrange for arbitrary C\ncode to be executed on each match.  It accepts a declarative\nminilanguage and emits skeleton C code.\n<\/p>\n<p>\n  A crude but useful way to think about what a\n  <em>lex<\/em> -generated tokenizer does is as a sort\nof inverse\ngrep(1).\nWhere\ngrep(1)\ntakes a single regular expression and returns a list of matches in the\nincoming data stream, each call to a\n  <em>lex<\/em> -generated tokenizer takes a list of\nregular expressions and indicates which expression occurs next in the\ndatastream.\n<\/p>\n <blockquote>\n  <p>\n   Splitting input analysis into tokenizing input and parsing\nthe token stream is a useful tactic even if you're not using Yacc and\nLex and your \u201ctokens\u201d are nothing like the usual ones in\na compiler.  More than once I've found that splitting input handling\ninto two levels made the code much simpler and easier to understand,\ndespite the complexity added by the split itself.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  <em>lex<\/em> was written to automate the task\nof generating lexical analyzers (tokenizers) for compilers.  It turned\nout to have a surprisingly wide range of uses for other kinds of\npattern recognition, and has since been described as \u201cthe\nSwiss-army knife of Unix programming\u201d.<sup><a href=\"#ftn.id2980161\" id=\"id2980161\">[130]<\/a><\/sup><\/p>\n<p>\n  If you are attacking any kind of pattern-recognition or\nstate-machine problem in which all the possible input stimuli will fit\nin a byte,\n  <em>lex<\/em> may enable you to generate\ncode that will be more efficient and reliable than a hand-crafted\nstate machine.\n<\/p>\n <blockquote>\n  <p>\n   John Jarvis at Holmdel [an AT&amp;T laboratory] used\n   <em>lex<\/em> to find faults in circuit boards, by\nscanning the board, using a chain-encoding technique to represent the\nedges of areas on the board, and then using Lex to define patterns\nthat would catch common fabrication errors.\n  <\/p><div>\n  \n   \u2013 Mike Lesk\n  \n <\/div><\/blockquote>\n<p>\n  Most importantly, the\n  <em>lex<\/em> specification minilanguage is much higher-level and more\ncompact\n  than\nequivalent handcrafted C.  Modules are available to use\n  <em>flex<\/em>, the open-source version, with\nPerl\n  (find them\nwith a Web search for \u201clex perl\u201d), and a work-alike\nimplementation is part of\n  <em>PLY<\/em> in\nPython.\n<\/p>\n<p>\n  <em>lex<\/em> generates parsers that are up\nto an order of magnitude slower than hand-coded parsers.  This is\nnot a good reason to hand-code, however; it's an argument for\nprototyping with\n  <em>lex<\/em> and hand-hacking\nonly if prototyping reveals an actual bottleneck.\n<\/p>\n<p>\n  <em>yacc<\/em> is a parser generator.  It, too,\nwas written to automate part of the job of writing compilers.  It\ntakes as input a grammar specification in a declarative minilanguage\nresembling BNF (Backus-Naur Form) with C code associated with each\nelement of the grammar.  It generates code for a parser function\nthat, when called, accepts text matching the grammar from an input\nstream.  As each grammar element is recognized, the parser function\nruns the associated C code.\n<\/p>\n<p>\n  The combination of\n  <em>lex<\/em> and\n  <em>yacc<\/em> is very effective for writing language\ninterpreters of all kinds.  Though most Unix programmers never get to\ndo the kind of general-purpose compiler-building that these tools were\nmeant to assist, they're extremely useful for writing parsers for\nrun-control file syntaxes and domain-specific minilanguages.\n<\/p>\n<p>\n  <em>lex<\/em> -generated tokenizers are very\nfast at recognizing low-level patterns in input streams, but the\nregular-expression minilanguage that\n  <em>lex<\/em> knows is not good at counting things, or recognizing recursively\nnested structures.  For parsing those, you want\n  <em>yacc<\/em>.  On the other hand, while you\ntheoretically could write a\n  <em>yacc<\/em> grammar to\ndo its own token-gathering, the grammar to specify that would be\nhugely bloated and the parser extremely slow.  For tokenizing input,\nyou want\n  <em>lex<\/em>.  Thus, these tools are\nsymbiotic.\n<\/p>\n<p>\n  If you can implement your parser in a higher-level language than\nC (which we recommend you do; see\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  for discussion), then look for equivalent\nfacilities like Python's PLY (which covers both\n  <em>lex<\/em> and\n  <em>yacc<\/em>)<sup><a href=\"#ftn.id2985990\" id=\"id2985990\">[131]<\/a><\/sup>\n  or Perl's\n  PY\nand Parse::Yapp modules, or Java's\n  CUP,<sup><a href=\"#ftn.id2986024\" id=\"id2986024\">[132]<\/a><\/sup>\n  Jack,<sup><a href=\"#ftn.id2986040\" id=\"id2986040\">[133]<\/a><\/sup>\n  or Yacc\/M<sup><a href=\"#ftn.id2986056\" id=\"id2986056\">[134]<\/a><\/sup>\n  packages.\n<\/p>\n<p>\n  As with macro processors, one of the problems with code\ngenerators and preprocessors is that compile-time errors in the\ngenerated code may carry line numbers that are relative to the\ngenerated code (which you don't want to edit) rather than the\ngenerator input (which is where you need to make corrections).\n  <em>yacc<\/em> and\n  <em>lex<\/em> address this by generating the same\n  <code>#line<\/code> constructs that the C\n  preprocessor does; these set the current line number for error\nreporting so the numbers will come out right. Any program that\ngenerates C\n  or\nC++\n  should do\nlikewise.\n<\/p>\n<p>\n  More generally, well-designed procedural-code generators\nshould never require the user to hand-alter or even look at the\ngenerated parts.  Getting those right is the code generator's\njob.\n<\/p>\n<h5 id=\"id2986148\"><a href=\"#id2986148\">\u00a7<\/a>Case Study: The fetchmailrc Grammar<\/h5>\n<p>\n  The canonical demonstration example that seems to have appeared\nin every\n  <em>lex<\/em> and\n  <em>yacc<\/em> tutorial ever written is a toy\ninteractive calculator program that parses and evaluates arithmetic\nexpressions entered by the user.  We will spare you yet another\nrepetition of this cliche; if you are interested, consult the source\ncode of the\nbc(1)\nand\ndc(1)\ncalculator implementations from the GNU project, or the paradigm example\n\u2018hoc\u2019<sup><a href=\"#ftn.id2986209\" id=\"id2986209\">[135]<\/a><\/sup>\n  from <sup><a href=\"#Kernighan-Pike84\" title=\"[Kernighan-Pike84]\">[Kernighan-Pike84]<\/a><\/sup>.\n<\/p>\n<p>\n  Instead, the grammar of\n  <em>fetchmail<\/em> 's\n  run-control-file parser provides a good medium-sized case study in\n  <em>lex<\/em> and\n  <em>yacc<\/em> usage.  There are a couple of points of interest here.\n<\/p>\n<p>\n  The\n  <em>lex<\/em> specification, in\n  <code>rcfile_l.l<\/code>, is a very typical implementation of a\nshell-like syntax.  Note how two complementary rules support either\nsingle or double-quoted strings; this is a good idea in general.  The\nrules for accepting (possibly signed) integer literals and discarding\ncomments are also pretty generic.\n<\/p>\n<p>\n  The\n  <em>yacc<\/em> specification, in\n  <code>rcfile_y.y<\/code>, is long but straightforward. It does\nnot perform any\n  <em>fetchmail<\/em> actions, just\nsets bits in a list of internal control blocks. After startup,\n  <em>fetchmail<\/em> 's normal mode of operation is\njust to repeatedly walk that list, using each record to drive a\nretrieval session with a remote site.\n<\/p>\n<h4 id=\"id2986324\"><a href=\"#id2986324\">\u00a7<\/a>Case Study: Glade<\/h4>\n<p>\n  We looked at\n  <em>Glade<\/em> in\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  as a good example of a\ndeclarative minilanguage. We also noted that its back end produces a\nresult by generating code in any one of several languages.\n<\/p>\n<p>\n  <em>Glade<\/em> is a good modern example of an\napplication-code generator.  What makes it Unixy in spirit are the\nfollowing features, which most GUI builders (especially most\nproprietary GUI builders) don't have:\n<\/p>\n <div><ul><li>\n    <p>\n     Rather than being glued together as one monster\nmonolith, the\n     <em>Glade<\/em> GUI and\n     <em>Glade<\/em> code generator obey the Rule of\nSeparation (following the \u201cseparated engine and\ninterface\u201d design pattern).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The GUI and code generator are connected by an (XML-based) textual\ndata file format that can be read and modified by other tools.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Multiple target languages (as opposed to just\nC\n     or\nC++) are supported.  More\ncould easily be added.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  The design implies that it should also be possible to\nreplace the\n  <em>Glade<\/em> GUI editor component,\nshould that ever become desirable.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2980161\" id=\"ftn.id2980161\">[130]<\/a><\/sup>\n  The common latter-day description of Perl as a \n\u201cSwiss-army chainsaw\u201d is derivative.\n<\/p><p><sup><a href=\"#id2985990\" id=\"ftn.id2985990\">[131]<\/a><\/sup>\n  PLY is\n  <a href=\"http:\/\/systems.cs.uchicago.edu\/ply\/\">\n   downloadable<\/a>.\n<\/p><p><sup><a href=\"#id2986024\" id=\"ftn.id2986024\">[132]<\/a><\/sup>\n  CUP is\n  <a href=\"http:\/\/www.cs.princeton.edu\/~appel\/modern\/java\/CUP\/\">\n   downloadable<\/a>.\n<\/p><p><sup><a href=\"#id2986040\" id=\"ftn.id2986040\">[133]<\/a><\/sup>\n  Jack is\n  <a href=\"http:\/\/www.javaworld.com\/javaworld\/jw-12-1996\/jw-12-jack.html\">\n   downloadable<\/a>.\n<\/p><p><sup><a href=\"#id2986056\" id=\"ftn.id2986056\">[134]<\/a><\/sup>\n  Yacc\/M is\n  <a href=\"http:\/\/david.tribble.com\/yaccm.html\">\n   downloadable<\/a>.\n<\/p><p><sup><a href=\"#id2986209\" id=\"ftn.id2986209\">[135]<\/a><\/sup><a href=\"http:\/\/cm.bell-labs.com\/cm\/cs\/upe\/\">\n   http:\/\/cm.bell-labs.com\/cm\/cs\/upe\/<\/a>\n<\/p><\/div>\n\n<h3 id=\"ch15s04\"><a href=\"#ch15s04\">\u00a7<\/a>make: Automating Your Recipes<\/h3>\n<p>\n  Program sources by themselves don't make an application. The way\nyou put them together and package them for distribution matters,\ntoo. Unix provides a tool for semi-automating these processes;\nmake(1).\n  <em>Make<\/em> is covered in most introductory Unix\nbooks. For a really thorough reference, you can consult\n  <em>Managing Projects with Make<\/em> <sup><a href=\"#Oram-Talbot\" title=\"[Oram-Talbot]\">[Oram-Talbot]<\/a><\/sup>. If you're using\n  <em>GNU make<\/em> (the most advanced make, and the\none normally shipped with open-source Unixes) the treatment in\n  <em>Programming with GNU Software<\/em> <sup><a href=\"#Loukides-Oram\" title=\"[Loukides-Oram]\">[Loukides-Oram]<\/a><\/sup> may be better in some\nrespects. Most Unixes that carry\n  <em>GNU make<\/em> will also support GNU Emacs; if yours does you will probably find a\ncomplete make manual on-line through Emacs's\n  <em>info<\/em> documentation system.\n<\/p>\n<p>\n  Ports of\n  <em>GNU make<\/em> to DOS and Windows\nare available from the FSF.\n<\/p>\n<h4 id=\"id2986550\"><a href=\"#id2986550\">\u00a7<\/a>Basic Theory of make<\/h4>\n<p>\n  If you're developing in C\n  or C++, an important part of the recipe for\nbuilding your application will be the collection of compilation and\nlinkage commands needed to get from your sources to working\nbinaries. Entering these commands is a lot of tedious detail work, and\nmost modern development environments include a way to put them in\ncommand files or databases that can automatically be re-executed to\nbuild your application.\n<\/p>\n<p>\n  Unix's make(1) program, the original of all these facilities, was designed specifically to help C programmers manage\n  these recipes. It lets you write down the dependencies between files in a project in one or more \u2018makefiles\u2019. Each\n  makefile consists of a series of <em>productions<\/em>; each one tells <em>make<\/em> that some given target file\n  depends on some set of source files, and says what to do if any of the sources are newer than the target. You don't\n  actually have to write down all dependencies, as the <em>make<\/em> program can deduce a lot of the obvious ones from\n  filenames and extensions.\n<\/p>\n<p>\n  For example: You might put in a makefile that the binary\n  <code>myprog<\/code> depends on three object files\n  <code>myprog.o<\/code>,\n  <code>helper.o<\/code>, and\n  <code>stuff.o<\/code>. If you have source files\n  <code>myprog.c<\/code>,\n  <code>helper.c<\/code>, and\n  <code>stuff.c<\/code>,\n  <em>make<\/em> will\nknow without being told that each\n  <code>.o<\/code> file depends\non the corresponding\n  <code>.c<\/code> file, and supply its own\nstandard recipe for building a\n  <code>.o<\/code> file from a\n  <code>.c<\/code> file.\n<\/p>\n <blockquote>\n  <p>\n   Make originated with a visit from Steve Johnson (author of\n   <em>yacc<\/em>, etc.), storming into my office, cursing\nthe Fates that had caused him to waste a morning debugging a correct\nprogram (bug had been fixed, file hadn't been compiled,\n   <code>cc *.o<\/code> was therefore unaffected).  As I had spent a part of the\nprevious evening coping with the same disaster on a project I was\nworking on, the idea of a tool to solve it came up.  It began with an\nelaborate idea of a dependency analyzer, boiled down to something much\nsimpler, and turned into Make that weekend.  Use of tools that were\nstill wet was part of the culture. Makefiles were text files, not\nmagically encoded binaries, because that was the Unix ethos:\nprintable, debuggable, understandable stuff.\n  <\/p><div>\n  \n   \u2013 Stuart Feldman\n  \n <\/div><\/blockquote>\n<p>\n  When you run <code>make<\/code> in a project directory, the\n  <em>make<\/em> program looks at all productions and\ntimestamps and does the minimum amount of work necessary to make sure\nderived files are up to date.\n<\/p>\n<p>\n  You can read a good example of a moderately complex makefile in\nthe sources for\n  <em>fetchmail<\/em>.\nIn the subsections below we'll refer to it again.\n<\/p>\n<p>\n  Very complex makefiles, especially when they call subsidiary\nmakefiles, can become a source of complications rather than\nsimplifying the build process.  A now-classic warning is issued\nin\n  <em>Recursive Make Considered Harmful<\/em>.<sup><a href=\"#ftn.id2986827\" id=\"id2986827\">[136]<\/a><\/sup>\n  The argument in this paper has\nbecome widely accepted since it was written in 1997, and has come\nnear to reversing previous community practice.\n<\/p>\n<p>\n  No discussion of\nmake(1)\nwould be complete without an acknowledgement that it includes one of\nthe worst design botches in the history of Unix. The use of tab\ncharacters as a required leader for command lines associated with a\nproduction means that the interpretation of a makefile can change\ndrastically on the basis of invisible differences in whitespace.\n<\/p>\n <blockquote>\n  <p>\n   Why the tab in column 1? Yacc was new, Lex was brand new. I\nhadn't tried either, so I figured this would be a good excuse to\nlearn. After getting myself snarled up with my first stab at Lex, I\njust did something simple with the pattern newline-tab. It\nworked, it stayed. And then a few weeks later I had a user population\nof about a dozen, most of them friends, and I didn't want to screw up\nmy embedded base. The rest, sadly, is history.\n  <\/p><div>\n  \n   \u2013 Stuart Feldman\n  \n <\/div><\/blockquote>\n<h4 id=\"id2986902\"><a href=\"#id2986902\">\u00a7<\/a>make in Non-C\/C++ Development<\/h4>\n<p>\n  <em>make<\/em> is not just useful for C\/C++\nrecipes, however.  Scripting languages like those we described in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  may not require\nconventional compilation and link steps, but there are often other\nkinds of dependencies that\nmake(1)\ncan help you with.\n<\/p>\n<p>\n  Suppose, for example, that you actually generate part of your\ncode from a specification file, using one of the techniques from\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>. You can use\n  <em>make<\/em> to tie the spec file and the generated\nsource together. This will ensure that whenever you change the spec\nand remake, the generated code will automatically be rebuilt.\n<\/p>\n<p>\n  It's quite common to use makefile productions to express recipes\nfor making documentation as well as code. You'll often see this\napproach used to automatically generate PostScript or other derived\ndocumentation from masters written in some markup language (like HTML\nor one of the Unix document-macro languages we'll survey in\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>). In fact, this sort of use is so\ncommon that it's worth illustrating with a case study.\n<\/p>\n<h5 id=\"id2986982\"><a href=\"#id2986982\">\u00a7<\/a>Case Study: make for Document-File Translation<\/h5>\n<p>\n  In the\nfetchmail\n  makefile,\nfor example, you'll see three productions that relate files named\n  <code>FAQ<\/code>,\n  <code>FEATURES<\/code>, and\n  <code>NOTES<\/code> to HTML sources\n  <code>fetchmail-FAQ.html<\/code>,\n  <code>fetchmail-features.html<\/code>, and\n  <code>design-notes.html<\/code>.\n<\/p>\n<p>\n  The HTML files are meant to be accessible on the fetchmail Web\npage, but all the HTML markup makes them uncomfortable to look at\nunless you're using a browser. So the\n  <code>FAQ<\/code>,\n  <code>FEATURES<\/code>, and\n  <code>NOTES<\/code> are\nflat-text files meant to be flipped through quickly with an editor or\npager program by someone reading the\n  <em>fetchmail<\/em> sources themselves (or, perhaps, distributed to FTP sites that don't\nsupport Web access).\n<\/p>\n<p>\n  The flat-text forms can be made from their HTML masters by using\nthe common open-source program\nlynx(1).\n  <em>lynx<\/em> is a Web browser for text-only\ndisplays; but when invoked with the\n  <code>-dump<\/code> option it\nfunctions reasonably well as an HTML-to-ASCII formatter.\n<\/p>\n<p>\n  With the productions in place, the developer can edit the HTML\nmasters without having to remember to manually rebuild the flat-text\nforms afterwards, secure in the knowledge that\n  <code>FAQ<\/code>,\n  <code>FEATURES<\/code>, and\n  <code>NOTES<\/code> will be properly rebuilt whenever they are\nneeded.\n<\/p>\n<h4 id=\"id2987148\"><a href=\"#id2987148\">\u00a7<\/a>Utility Productions<\/h4>\n<p>\n  Some of the most heavily used productions in typical makefiles\ndon't express file dependencies at all. They're ways to bundle up\nlittle procedures that a developer wants to mechanize, like making\na distribution package or removing all object files in order to do\na build from scratch.\n<\/p>\n <blockquote>\n  <p>\n   Non-file productions were intentional and in there from day\none. \u2018Make all\u2019 and \u2018clean\u2019 were my own\nconventions from earliest days.  One of the older Unix jokes is\n\u201cMake love\u201d which results in \u201cDon't know how to\nmake love\u201d.\n  <\/p><div>\n  \n   \u2013 Stuart Feldman\n  \n <\/div><\/blockquote>\n<p>\n  There is a well-developed set of conventions about what utility\nproductions should be present and how they should be named.\nFollowing these will make your makefile much easier to understand\nand use.\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>all<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Your\n     <code>all<\/code> production should make every\nexecutable of your project. Usually the\n     <code>all<\/code> production doesn't have an explicit rule; instead it refers to all of\nyour project's top-level targets (and, not accidentally, documents\nwhat those are). Conventionally, this should be the first production in\nyour makefile, so it will be the one executed when the developer types\n     <code>make<\/code> with no argument.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>test<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Run the program's automated test suite, typically consisting of\na set of unit tests<sup><a href=\"#ftn.id2987288\" id=\"id2987288\">[137]<\/a><\/sup>\n     to find regressions, bugs, or other deviations\nfrom expected behavior during the development process. The\n\u2018test\u2019 production can also be used by end-users of the\nsoftware to ensure that their installation is functioning correctly.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>clean<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Remove all files (such as binary executables and object files)\nthat are normally created when you\n     <code>make all<\/code>.  A <code>make clean<\/code>\n     should reset the process of building the software to a good initial state.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>dist<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Make a source archive (usually with the\ntar(1)\nprogram) that can be shipped as a unit and used to rebuild the program\non another machine. This target should do the equivalent of depending\non <code>all<\/code> so that a <code>make dist<\/code>\n     automatically rebuilds the whole project before making the\ndistribution archive \u2014 this is a good way to avoid last-minute\nembarrassments, like not shipping derived files that are actually\nneeded (like the flat-text\n     <code>README<\/code>\n     in <em>fetchmail<\/em>, which is actually generated from an HTML source).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>distclean<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Throw away everything but what you would include if you were\nbundling up the source with\n     <code>make dist<\/code>. This may be the the same as <code>make clean<\/code>\n     but should be included\nas a production of its own anyway, to document what's going on. When\nit's different, it usually differs by throwing away local\nconfiguration files that aren't part of the normal\n     <code>make all<\/code> build sequence (such as those generated by autoconf(1);\nwe'll talk about autoconf(1) in\n     <a href=\"#portabilitychapter\" title=\"Chapter 17. Portability\">\n      Chapter 17<\/a>).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>realclean<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Throw away everything you can rebuild using the makefile. This\nmay be the same as\n     <code>make distclean<\/code>, but should be\nincluded as a production of its own anyway, to document what's going\non. When it's different, it usually differs by throwing away files\nthat are derived but (for whatever reason) shipped with the project\nsources anyway.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>install<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Install the project's executables and documentation in system\ndirectories so they will be accessible to general users (this\ntypically requires root privileges). Initialize or update any\ndatabases or libraries that the executables require in order to\nfunction.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>uninstall<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Remove files installed in system directories by\n     <code>make install<\/code>\n     (this typically requires root privileges). This should completely and perfectly reverse a\n     <code>make install<\/code>.  The presence of an uninstall production implies a\nkind of humility that experienced Unix hands look for as a sign of\nthoughtful design; conversely, not having an uninstall production is\nat best careless, and (when, for example, an installation creates\nlarge database files) can be quite rude and thoughtless.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  Working examples of all the standard targets are available for\ninspection in the\n  <em>fetchmail<\/em> makefile. By studying all of them together you will see a pattern\nemerge, and (not incidentally) learn much about the\n  <em>fetchmail<\/em> package's structure. One of the\nbenefits of using these standard productions is that they form an\nimplicit roadmap of their project.\n<\/p>\n<p>\n  But you need not limit yourself to these utility productions.\nOnce you master\n  <em>make<\/em>, you'll find yourself\nmore and more often using the makefile machinery to automate little\ntasks that depend on your project file state. Your makefile is a\nconvenient central place to put these; using it makes them readily\navailable for inspection and avoids cluttering up your workspace with\ntrivial little scripts.\n<\/p>\n<h4 id=\"id2987644\"><a href=\"#id2987644\">\u00a7<\/a>Generating Makefiles<\/h4>\n<p>\n  One of the subtle advantages of Unix\n  <em>make<\/em> over the dependency databases built\ninto many IDEs is that makefiles are simple text files \u2014 files\nthat can be generated by programs.\n<\/p>\n<p>\n  In the mid-1980s it was fairly common for large Unix program\ndistributions to include elaborate custom shellscripts that would\nprobe their environment and use the information they gathered to\nconstruct custom makefiles.  These custom configurators reached absurd\nsizes.  I wrote one once that was 3000 lines of shell, about twice as\nlarge as any single module in the program it was configuring \u2014\nand this was not unusual.\n<\/p>\n<p>\n  The community eventually said \u201cEnough!\u201d and various\npeople set out to write tools that would automate away part or all of\nthe process of maintaining makefiles. These tools generally tried to\naddress two issues:\n<\/p>\n<p>\n  One issue is\n  <em>portability<\/em>. Makefile\ngenerators are commonly built to run on many different hardware\nplatforms and Unix variants. They generally try to deduce things about\nthe local system (including everything from machine word size up to\nwhich tools, languages, service libraries, and even document\nformatters it has available). They then try to use those deductions to\nwrite makefiles that exploit the local system's facilities and\ncompensate for its quirks.\n<\/p>\n<p>\n  The other issue is\n  <em>\n   dependency\nderivation\n<\/em>. It's possible to deduce a great deal about the\ndependencies of a collection of C sources by analyzing the sources\nthemselves (especially by looking at what include files they use and\nshare). Many makefile generators do this in order to mechanically\ngenerate\n  <em>make<\/em> dependencies.\n<\/p>\n<p>\n  Each different makefile generator tackles these objectives in a\nslightly different way. Probably a dozen or more generators have\nbeen attempted, but most proved inadequate or too difficult to drive\nor both, and only a few are still in live use. We'll survey the major\nones here. All are available as open-source software on the\nInternet.\n<\/p>\n<h5 id=\"id2987736\"><a href=\"#id2987736\">\u00a7<\/a>makedepend<\/h5>\n<p>\n  Several small tools have tackled the rule automation part of the\nproblem exclusively. This one, distributed along with the X windowing\nsystem from MIT, is the fastest and most useful and comes\npreinstalled under all modern Unixes, including all\nLinuxes.\n<\/p>\n<p>\n  <em>makedepend<\/em> takes a collection of\nC\n  sources and\ngenerates dependencies for the corresponding\n  <code>.o<\/code> files from their\n  <code>#include<\/code> directives. These can be\nappended directly to a makefile, and in fact makedepend is defined to\ndo exactly that.\n<\/p>\n<p>\n  <em>makedepend<\/em> is useless for anything\nbut C projects. It doesn't try to solve more than one piece of the\nmakefile-generation problem. But what it does it does quite\nwell.\n<\/p>\n<p>\n  <em>makedepend<\/em> is sufficiently documented by its manual page. If you type\n  <code>man makedepend<\/code> at a terminal window you will quickly learn what you need to know about\ninvoking it.\n<\/p>\n<h5 id=\"id2987839\"><a href=\"#id2987839\">\u00a7<\/a>Imake<\/h5>\n<p>\n  <em>Imake<\/em> was written in an attempt to\nmechanize makefile generation for the X window\nsystem. It\nbuilds on\n  <em>makedepend<\/em> to tackle both the\ndependency-derivation and portability problems.\n<\/p>\n<p>\n  <em>Imake<\/em> system effectively replaces\nconventional makefiles with Imakefiles. These are written in a more\ncompact and powerful notation which is (effectively) compiled into\nmakefiles. The compilation uses a rules file which is system-specific\nand includes a lot of information about the local environment.\n<\/p>\n<p>\n  <em>Imake<\/em> is well suited to X's\nparticular portability and configuration challenges and universally\nused in projects that are part of the X distribution. However, it has\nnot achieved much popularity outside the X developer community. It's\nhard to learn, hard to use, hard to extend, and produces generated\nmakefiles of mind-numbing size and complexity.\n<\/p>\n<p>\n  The\n  <em>Imake<\/em> tools will be available on\nany Unix that supports X, including Linux. There has been one heroic\neffort <sup><a href=\"#DuBois\" title=\"[DuBois]\">[DuBois]<\/a><\/sup> to make the mysteries of\n  <em>Imake<\/em> comprehensible to non-X-programming\nmortals. These are worth learning if you are going to do X\nprogramming.\n<\/p>\n<h5 id=\"id2987944\"><a href=\"#id2987944\">\u00a7<\/a>autoconf<\/h5>\n<p>\n  <em>autoconf<\/em> was written by people who\nhad seen and rejected the\n  <em>Imake<\/em> approach. It generates per-project\n  <code>configure<\/code> shellscripts that are like the old-fashioned custom script\nconfigurators. These\n  <code>configure<\/code> scripts can\ngenerate makefiles (among other things).\n<\/p>\n<p>\n  <em>Autoconf<\/em> is focused on portability\nand does no built-in dependency derivation at all. Although it is\nprobably as complex as\n  <em>Imake<\/em>, it is much\nmore flexible and easier to extend. Rather than relying on a\nper-system database of rules, it generates\n  <code>configure<\/code> shell code that goes out and searches\nyour system for things.\n<\/p>\n<p>\n  Each\n  <code>configure<\/code> shellscript is built from a\nper-project template that you have to write, called\n  <code>configure.in<\/code>. Once generated, though, the\n  <code>configure<\/code> script will be self-contained and can\nconfigure your project on systems that don't carry\nautoconf(1)\nitself.\n<\/p>\n<p>\n  The\n  <em>autoconf<\/em> approach to makefile\ngeneration is like imake's in that you start by writing a makefile\ntemplate for your project. But\n  <em>autoconf<\/em> 's\n  <code>Makefile.in<\/code> files are basically just makefiles\nwith placeholders in them for simple text substitution; there's no\nsecond notation to learn. If you want dependency derivation, you must take\nexplicit steps to call\nmakedepend(1)\nor some similar tool \u2014 or use\nautomake(1).\n<\/p>\n<p>\n  <em>autoconf<\/em> is documented by an on-line\nmanual in the GNU\n  <em>info<\/em> format. The source scripts\nof autoconf are available from the FSF\n  archive site, but are\nalso preinstalled on many Unix and Linux\n  versions. You should be able to\nbrowse this manual through your Emacs's help system.\n<\/p>\n<p>\n  Despite its lack of direct support for dependency derivation,\nand despite its generally ad-hoc approach, in mid-2003\n  <em>autoconf<\/em> is clearly the most popular of the\nmakefile generators, and has been for some years. It has eclipsed\n  <em>Imake<\/em> and driven at least one major\ncompetitor (\n  <em>metaconfig<\/em>) out of use.\n<\/p>\n<p>\n  A reference,\n  <em>GNU Autoconf, Automake and Libtool<\/em>\n  is available <sup><a href=\"#Vaughan\" title=\"[Vaughan]\">[Vaughan]<\/a><\/sup>. We'll have\nmore to say about\n  <em>autoconf<\/em>, from a slightly\ndifferent angle, in\n  <a href=\"#portabilitychapter\" title=\"Chapter 17. Portability\">\n   Chapter 17<\/a>.\n<\/p>\n<h5 id=\"id2988205\"><a href=\"#id2988205\">\u00a7<\/a>automake<\/h5>\n<p>\n  <em>automake<\/em> is an attempt to add\n  <em>Imake<\/em> -like dependency derivation as a layer\non top of\nautoconf(1). You\nwrite\n  <code>Makefile.am<\/code> templates in a broadly\n  <em>Imake<\/em> -like notation;\nautomake(1)\ncompiles them to\n  <code>Makefile.in<\/code> files, which\n  <em>autoconf<\/em> 's\n  <code>configure<\/code> scripts then operate\non.\n<\/p>\n<p>\n  <em>automake<\/em> is still relatively new\ntechnology in mid-2003. It is used in several\nFSF\n  projects but has not yet been widely adopted elsewhere.  While its\ngeneral approach looks promising, it is as yet rather brittle \u2014\nit works when used in stereotyped ways but tends to break badly if you\ntry to do anything unusual with it.\n<\/p>\n<p>\n  Complete on-line documentation is shipped with\n  <em>automake<\/em>, which can be downloaded from the\nFSF\n  archive site.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2986827\" id=\"ftn.id2986827\">[136]<\/a><\/sup><a href=\"http:\/\/www.tip.net.au\/~millerp\/rmch\/recu-make-cons-harm.html\">\n   Available\non the Web<\/a>.\n<\/p><p><sup><a href=\"#id2987288\" id=\"ftn.id2987288\">[137]<\/a><\/sup>\n  A unit test is test code attached to a\nmodule to verify correct performance.  Use of the term \u2018unit\ntest\u2019 suggests that the test is written concurrently\nwith the code by the developer of the code, and implies a discipline\nin which module releases aren't considered complete until they have\nattached test code. The term and the concept originated in the\n\u201cExtreme Programming\u201d methodology popularized by Kent\nBeck, but has gained wide acceptance among Unix programmers since\nabout 2001.\n<\/p><\/div>\n\n<h3 id=\"ch15s05\"><a href=\"#ch15s05\">\u00a7<\/a>Version-Control Systems<\/h3>\n<p>\n  Code evolves. As a project moves from first-cut prototype to\ndeliverable, it goes through multiple cycles in which you explore\nnew ground, debug, and then stabilize what you've accomplished. And\nthis evolution doesn't stop when you first deliver for production.\nMost projects will need to be maintained and enhanced past the 1.0\nstage, and will be released multiple times.  Tracking all that \ndetail is just the sort of thing computers are good at and \nhumans are not.\n<\/p>\n<h4 id=\"id2988357\"><a href=\"#id2988357\">\u00a7<\/a>Why Version Control?<\/h4>\n<p>\n  Code evolution raises several practical problems that can be\nmajor sources of friction and drudgery \u2014 thus a serious drain on\nproductivity. Every moment spent on these problems is a moment not\nspent on getting the design and function of your project right.\n<\/p>\n<p>\n  Perhaps the most important problem is\n  <em>reversion<\/em>.  If you make a change, and discover\nit's not viable, how can you revert to a code version that is known\ngood? If reversion is difficult or unreliable, it's hard to risk\nmaking changes at all (you could trash the whole project, or make many\nhours of painful work for yourself).\n<\/p>\n<p>\n  Almost as important is\n  <em>change tracking<\/em>. You\nknow your code has changed; do you know why? It's easy to forget the\nreasons for changes and step on them later. If you have collaborators\non a project, how do you know what they have changed while you weren't\nlooking, and who was responsible for each change?\n<\/p>\n <blockquote>\n  <p>\n   Amazingly often, it is useful to ask what\n   <em>you<\/em>\n   have changed since the last known-good\nversion, even if you have no collaborators.  This often uncovers\nunwanted changes, such as forgotten debugging code.  I now do this\nroutinely before checking in a set of changes.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  Another issue is\n  <em>bug tracking<\/em>. It's quite common\nto get new bug reports for a particular version after the code has\nmutated away from it considerably. Sometimes you can recognize\nimmediately that the bug has already been stomped, but often you\ncan't. Suppose it doesn't reproduce under the new version. How do you\nget back the state of the code for the old version in order to\nreproduce and understand it?\n<\/p>\n<p>\n  To address these problems, you need procedures for keeping a\nhistory of your project, and annotating it with comments that\nexplain the history. If your project has more than one developer,\nyou also need mechanisms for making sure developers don't overwrite\neach others' versions.\n<\/p>\n<h4 id=\"id2988464\"><a href=\"#id2988464\">\u00a7<\/a>Version Control by Hand<\/h4>\n<p>\n  The most primitive (but still very common) method is all\nhand-hacking. You snapshot the project periodically by manually\ncopying everything in it to a backup. You include history comments\nin source files. You make verbal or email arrangements with other\ndevelopers to keep their hands off certain files while you hack\nthem.\n<\/p>\n<p>\n  The hidden costs of this hand-hacking method are high,\nespecially when (as frequently happens) it breaks down. The\nprocedures take time and concentration; they're prone to error, and\ntend to get slipped under pressure or when the project is in\ntrouble \u2014 that is, exactly when they are most needed.\n<\/p>\n<p>\n  As with most hand-hacking, this method does not scale well.  It\nrestricts the granularity of change tracking, and tends to lose\nmetadata details such as the order of changes, who did them, and why.\nReverting just a part of a large change can be tedious and time\nconsuming, and often developers are forced to back up farther than\nthey'd like after trying something that doesn't work.\n<\/p>\n<h4 id=\"id2988503\"><a href=\"#id2988503\">\u00a7<\/a>Automated Version Control<\/h4>\n<p>\n  To avoid these problems, you can use a\n  <em>\n   version-control\nsystem\n<\/em>\n  (VCS), a suite of programs that automates away most of\nthe drudgery involved in keeping an annotated history of your\nproject and avoiding modification conflicts.\n<\/p>\n<p>\n  Most VCSs share the same basic logic. To use one, you start by\n  <em>registering<\/em>\n  a collection of source files \u2014\nthat is, telling your VCS to start archive files describing their\nchange histories. Thereafter, when you want to edit one of these\nfiles, you have to\n  <em>check out<\/em>\n  the file \u2014\nassert an exclusive lock on it. When you're done, you\n  <em>\n   check\nin\n<\/em>\n  the file, adding your changes to the archive, releasing\nthe lock, and entering a change comment explaining what you\ndid.\n<\/p>\n<p>\n  The history of the project is not necessarily linear. All VCSs\nin common use actually allow you to maintain a tree of variant\nversions (for ports to different machines, say) with tools for merging\nbranches back into the main \u201ctrunk\u201d version.  This\nfeature becomes important as the size and dispersion of the \ndevelopment group increases.  It needs to be used with care, \nhowever; multiple active variants of the code base can be \nvery confusing (just associated bug reports to the right version\nare not necessarily easy), and automated merging of branches does\nnot guaranteed that the combined code works.\n<\/p>\n<p>\n  Most of the rest of what a VCS does is convenience: labeling,\nand reporting features surrounding these basic operations, and tools\nwhich allow you to view differences between versions, or to group a\ngiven set of versions of files as a named\n  <em>release<\/em>\n  that can be examined or reverted to at any time without losing later\nchanges.\n<\/p>\n<p>\n  VCSs have their problems. The biggest one is that using a VCS\ninvolves extra steps every time you want to edit a file, steps\nthat developers in a hurry tend to want to skip if they have to be\ndone by hand. Near the end of this chapter we'll discuss a way to\nsolve this problem.\n<\/p>\n<p>\n  Another problem is that some kinds of natural operations tend to\nconfuse VCSs. Renaming files is a notorious trouble spot; it's not\neasy to automatically ensure that a file's version history will be\ncarried along with it when it is renamed.  Renaming problems are\nparticularly difficult to resolve when the VCS supports\nbranching.\n<\/p>\n<p>\n  Despite these difficulties, VCSs are a huge boon to productivity\nand code quality in many ways, even for small single-developer\nprojects. They automate away many procedures that are just tedious\nwork. They help a lot in recovering from mistakes. Perhaps most\nimportantly, they free programmers to experiment by guaranteeing\nthat reversion to a known-good state will always be easy.\n<\/p>\n<p>\n  (VCSs, by the way, are not merely good for program code; the\nmanuscript of this book was maintained as a collection of files\nunder\n  <em>RCS<\/em> while it was being written.)\n<\/p>\n<h4 id=\"id2988625\"><a href=\"#id2988625\">\u00a7<\/a>Unix Tools for Version Control<\/h4>\n<p>\n  Historically, three VCSs have been of major significance in the\nUnix world, and we'll survey them here. For an extended introduction\nand tutorial, consult\n  <em>Applying RCS and SCCS<\/em> <sup><a href=\"#Bolinger-Bronson\" title=\"[Bolinger-Bronson]\">[Bolinger-Bronson]<\/a><\/sup>.\n<\/p>\n<h5 id=\"id2988648\"><a href=\"#id2988648\">\u00a7<\/a>Source Code Control System (SCCS)<\/h5>\n<p>\n  The first was\n  <em>SCCS<\/em>, the original\nSource Code Control System developed by Bell Labs around 1980 and\nfeatured in System III\n  Unix.\n  <em>SCCS<\/em> seems to have been the first serious attempt at a unified source-code\nmanagement system; concepts that it pioneered are still found at some\nlevel in all later ones, including commercial Unix and Windows\nproducts such as ClearCase.\n<\/p>\n<p>\n  <em>SCCS<\/em> itself is, however, now\nobsolete; it was proprietary Bell Labs software.  Superior open-source\nalternatives have since been developed, and most of the Unix world has\nconverted to those.\n  <em>SCCS<\/em> is still in use to\nmanage old projects at some commercial vendors, but can no longer be\nrecommended for new projects.\n<\/p>\n<p>\n  No complete open-source implementation of\n  <em>SCCS<\/em> exists. A clone called CSSC\n(Compatibly Stupid Source Control) is in development under the\nsponsorship of the FSF.\n<\/p>\n<h5 id=\"id2988732\"><a href=\"#id2988732\">\u00a7<\/a>Revision Control System (RCS)<\/h5>\n<p>\n  The superior open-source alternatives began with RCS (Revision\nControl System), born at Purdue University a few years after\n  <em>SCCS<\/em> and originally distributed with 4.3BSD\nUnix. It is\nlogically similar to\n  <em>SCCS<\/em> but has a cleaner\ncommand interface, and good facilities for grouping together entire\nproject releases under symbolic names.\n<\/p>\n<p>\n  <em>RCS<\/em> is currently the most widely\nused version control system in the Unix world. Some other Unix\nversion-control systems use it as a back end or underlayer. It is well\nsuited for single-developer or small-group projects hosted at a single\ndevelopment shop.\n<\/p>\n<p>\n  The\n  <em>RCS<\/em> sources are maintained and\ndistributed by the FSF. Free ports are available for\nMicrosoft operating systems\n  and VAX VMS.\n<\/p>\n<h5 id=\"id2988824\"><a href=\"#id2988824\">\u00a7<\/a>Concurrent Version System (CVS)<\/h5>\n<p>\n  CVS (Concurrent Version System) began life as a front end to\n  <em>RCS<\/em> developed in the early 1990s, but the\nmodel of version control it uses was different enough that it\nimmediately qualified as a new design. Modern implementations don't\nrely on\n  <em>RCS<\/em>.\n<\/p>\n<p>\n  Unlike\n  <em>RCS<\/em> and\n  <em>SCCS<\/em>,\n  <em>CVS<\/em> doesn't exclusively lock files when they're checked out. Instead, it\ntries to reconcile nonconflicting changes mechanically when they're\nchecked back in, and requests human help on conflicts. The design\nworks because patch conflicts are much less common than one might\nintuitively think.\n<\/p>\n<p>\n  The interface of\n  <em>CVS<\/em> is significantly\nmore complex than that of\n  <em>RCS<\/em>, and it needs\na lot more disk space. These properties make it a poor choice for small\nprojects. On the other hand,\n  <em>CVS<\/em> is well\nsuited to large multideveloper efforts distributed across several\ndevelopment sites connected by the\nInternet.\n  <em>CVS<\/em> tools on a client machine can\neasily be told to direct their operations to a repository located on a\ndifferent host.\n<\/p>\n<p>\n  The open-source community makes heavy use of\n  <em>CVS<\/em> for projects such as GNOME and\nMozilla. Typically, such\n  <em>CVS<\/em> repositories\nallow anyone to check out sources remotely. Anyone can, therefore,\nmake a local copy of a project, modify it, and mail change patches to\nthe project maintainers. Actual write access to the repository is more\nlimited and has to be explicitly granted by the project maintainers. A\ndeveloper who has such access can perform a commit option from his\nmodified local copy, which will cause the local changes to get made\ndirectly to the remote repository.\n<\/p>\n<p>\n  You can see an example of a well-run\n  <em>CVS<\/em> repository, accessible over the\nInternet, at the\n  <a href=\"http:\/\/cvs.gnome.org\">\n   GNOME CVS\nsite<\/a>. This site illustrates the use of\n  <em>CVS<\/em> -aware browsing tools such as Bonsai,\nwhich are useful in helping a large and decentralized group of\ndevelopers coordinate their work.\n<\/p>\n<p>\n  The social machinery and philosophy accompanying the use of\n  <em>CVS<\/em> is as important as the details of the\ntools. The assumption is that projects\n  <em>will<\/em>\n  be\nopen and decentralized, with code subject to peer review and\ninspection even by developers who are not officially members of the\nproject group.\n<\/p>\n<p>\n  Just as importantly,\n  <em>CVS<\/em> 's\nnonlocking philosophy means that projects can't be blocked by a lock\nif a programmer disappears in the middle of making some changes.\n  <em>CVS<\/em> thus allows developers to avoid the\n\u201csingle person point of failure\u201d problem; in turn, this\nmeans that project boundaries can be fluid, casual contributions are\nrelatively easy, and projects are not required to have an elaborate\nhierarchy of control.\n<\/p>\n<p>\n  The\n  <em>CVS<\/em> sources are maintained and\ndistributed by the FSF.\n<\/p>\n<p>\n  <em>CVS<\/em> has significant problems. Some\nare merely implementation bugs, but one basic problem is that your\nproject's file namespace is not versioned in the same way changes to\nfiles themselves are. Thus,\n  <em>CVS<\/em> is easily\nconfused by file renamings, deletions, and additions.  Also,\n  <em>CVS<\/em> records changes on a per-file basis,\nrather than as\n  <em>sets<\/em>\n  of changes made to files.  This\nmakes it harder to back out to specific versions, and harder to handle\npartial check-ins.  Fortunately, none of these problems are intrinsic\nto the nonlocking style, and they have been successfully addressed by\nnewer version-control systems.\n<\/p>\n<h5 id=\"id2989087\"><a href=\"#id2989087\">\u00a7<\/a>Other Version-Control Systems<\/h5>\n<p>\n  <em>CVS<\/em> 's design problems are sufficient\nto have created demand for a better open-source VCS. Several such\nefforts are under way as of 2003. The most notable of these are\n  <em>Aegis<\/em> and\n  <em>Subversion<\/em>.\n<\/p>\n<p>\n  <a href=\"http:\/\/www.pcug.org.au\/~millerp\/aegis\/aegis.html\">\n   Aegis<\/a>\n  has the longest history of any of these alternatives, has hosted\nits own development since 1991, and is a mature production system.\nIt features a heavy emphasis on regression-testing and\nvalidation.\n<\/p>\n<p>\n  <a href=\"http:\/\/subversion.tigris.org\/\">\n   Subversion<\/a>\n  is\npositioned as \u201cCVS done right\u201d, with the known design\nproblems fully addressed, and in 2003 probably has the best\nnear-term prospect of replacing\n  <em>CVS<\/em>.\n<\/p>\n<p>\n  The\n  <a href=\"http:\/\/www.bitkeeper.com\">\n   BitKeeper<\/a>\n  project\nexplores some interesting design ideas related to change-sets and\nmultiple distributed code repositories. Linus Torvalds uses Bitkeeper \nfor the Linux kernel sources.  Its non-open-source license\nis, however, controversial, and has significantly retarded the\nacceptance of the product.\n<\/p>\n\n<h3 id=\"ch15s06\"><a href=\"#ch15s06\">\u00a7<\/a>Runtime Debugging<\/h3>\n<p>\n  Anyone who has been programming longer than a week knows that\ngetting the syntax of your programming language right is the\n  <em>easy<\/em>\n  part of debugging. The hard part comes after that,\nwhen you need to understand why your syntactically correct program\ndoesn't behave as you expect.\n<\/p>\n<p>\n  The Unix tradition encourages developers to anticipate this\nproblem by designing for transparency \u2014 in particular, designing\nprograms in such a way that their internal data flows are readily\nmonitored with the naked eye and simple tools, and readily mentally\nmodeled. This is a topic we covered in detail in\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>.  Design for\ntransparency is valuable both for preventing bugs and for easing the\nruntime-debugging task.\n<\/p>\n<p>\n  Design for transparency is not, however, sufficient in itself.\nWhen you are debugging a program at runtime, it's extremely useful to\nbe able to examine the state of your program at runtime, set\nbreakpoints, and execute pieces of it down to the single-statement\nlevel in a controlled way. Unix has a long tradition of hosting\nprograms to help you with this. Open-source Unixes feature a powerful\none called\n  <em>gdb<\/em> (yet another\nFSF\n  project) that supports C\n  and C++\n  debugging.\n<\/p>\n<p>\n  Perl,\nPython,\nJava, and Emacs\nLisp\n  all\nsupport standard packages or programs (included with their base\ndistributions) that allow you to set breakpoints, control execution,\nand do general runtime-debugger things. Tcl, designed as a small language for small\nprojects, has no such facility (though it does have a trace facility\nthat can be used to watch variables at runtime).\n<\/p>\n<p>\n  Remember the Unix philosophy. Spend your time on design quality, not\nthe low-level details, and automate away everything you can \u2014\nincluding the detail work of runtime debugging.\n<\/p>\n\n<h3 id=\"ch15s07\"><a href=\"#ch15s07\">\u00a7<\/a>Profiling<\/h3>\n<p>\n  As a general rule, 90% of the execution time of your program\nwill be spent in 10% of its code. Profilers are tools that help you\nidentify the 10% of hot spots that constrain the speed of your\nprogram. This is a good thing for making it faster.\n<\/p>\n<p>\n  But in the Unix tradition, profilers have a far more important\nfunction. They enable you\n  <em>not<\/em>\n  to optimize the other 90%!\nThis is good, and not just because it saves you work. The\n  <em>really<\/em>\n  valuable effect is that not optimizing that 90%\nholds down global complexity and reduces bugs.\n<\/p>\n<p>\n  You may recall that we quoted Donald\nKnuth\n  observing \u201cPremature optimization is the root of all\nevil\u201d\n  in\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">\n   Chapter 1<\/a>, and that\nRob Pike\n  and Ken\nThompson\n  had a\nfew pungent observations on the topic as well. These were the voices\nof experience. Do good design.  Think about what's\n  <em>right<\/em>\n  first. Tune for efficiency later.\n<\/p>\n<p>\n  Profilers help you do this. If you get in the good habit of\nusing them, you can get rid of the bad habit of premature\noptimization. Profilers don't just change the way you work; they\nchange how you think.\n<\/p>\n<p>\n  Profilers for compiled languages rely on instrumenting object\ncode, so they are even more platform-dependent than compilers. On the\nother hand, a compiled-language profiler doesn't care about the source\nlanguage of the programs it instruments. Under Unix, the single\nprofiler\ngprof(1)\nhandles C,\nC++, and all\nother compiled languages.\n<\/p>\n<p>\n  Perl,\nPython, and\nEmacs Lisp\n  have their own profilers included in\ntheir basic distributions; these are portable across all platforms on\nwhich the host languages themselves run. Java\n  has built-in profiling.\nTcl\n  has no\nprofiling support as yet.\n<\/p>\n\n<h3 id=\"ch15s08\"><a href=\"#ch15s08\">\u00a7<\/a>Combining Tools with Emacs<\/h3>\n<p>\n  One of the things the Emacs editor is\nvery good at is acting as a front end for other development tools (we\ndiscussed this from a philosophical angle in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">\n   Chapter 13<\/a>).  In fact, nearly every tool we've\ndiscussed in this chapter can be driven from within an\nEmacs editor session through front ends\nthat give them greater utility than they would have running\nstandalone.\n<\/p>\n<p>\n  To illustrate this, we'll walk you through the use of these\ntools with Emacs in a typical\nbuild\/test\/debug cycle. For details on them, see\nEmacs's own on-line help system; this\nsection just gives you an overview, to motivate you to learn\nmore.\n<\/p>\n<p>\n  Read and learn \u2014 not just about\nEmacs, but about the mental habit of\nlooking for synergies between programs, and creating them. Try to read\nthis section as instruction in philosophy, not just technique.\n<\/p>\n<h4 id=\"id2989559\"><a href=\"#id2989559\">\u00a7<\/a>Emacs and make<\/h4>\n<p>\n  <em>Make<\/em>, for example, can be started with the Emacs command\n  <code>ESC-x compile<\/code> followed by an Enter. This command will run make(1)\nin the current directory, capturing the output in an Emacs buffer.\n<\/p>\n<p>\n  This by itself wouldn't be very useful. But\nEmacs's\n  <em>make<\/em> mode knows about the error message format (featuring a source file and\nline number) emitted by Unix C\n  compilers and many other tools.\n<\/p>\n<p>\n  If anything run by\n  <em>make<\/em> issues error\nmessages, the command <code>Ctl-X `<\/code> (control-X-backquote)\nwill try to parse them and take you to each error location in turn,\npopping open a window on the appropriate file and taking the cursor to\nthe error line.<sup><a href=\"#ftn.id2989640\" id=\"id2989640\">[138]<\/a><\/sup><\/p>\n<p>\n  This makes it extremely easy to step through an entire build,\nfixing any syntax that has been broken since the last compile.\n<\/p>\n<h4 id=\"id2989677\"><a href=\"#id2989677\">\u00a7<\/a>Emacs and Runtime Debugging<\/h4>\n<p>\n  For catching runtime errors, Emacs\noffers similar integration with your symbolic debugger \u2014 that\nis, you can use an Emacs mode to set\nbreakpoints in your programs and examine their runtime state. You run\nthe debugger by sending it commands through an\nEmacs window. Whenever the debugger stops\non a breakpoint, the message the debugger ships back about the source\nlocation is parsed and used to pop up a window on the source around the\nbreakpoint.\n<\/p>\n<p>\n  Emacs's Grand Unified Debugger mode\nsupports all the major C debuggers:\ngdb(1),\nsdb(1),\ndbx(1),\nand\nxdb(1). It\nalso supports Perl\n  symbolic debugging using the perldb\nmodule, and the standard debuggers for both\nJava\n  and\nPython. Facilities built into Emacs Lisp\nitself support interactive debugging of Emacs Lisp\ncode.\n<\/p>\n<p>\n  At time of writing (mid-2003) there is not yet support for\nTcl\n  debugging\nfrom within Emacs. The design of Tcl is\nsuch that it seems unlikely to be added.\n<\/p>\n<h4 id=\"id2989781\"><a href=\"#id2989781\">\u00a7<\/a>Emacs and Version Control<\/h4>\n<p>\n  Once you've corrected your program's syntax and fixed its\nruntime bugs, you may want to save the changes into a\nversion-controlled archive. If you've only tried running\nversion-control tools from the shell, it's hard to blame you for\nsloughing off this important step. Who wants to have to remember to\nrun checkout\/checkin commands around every edit operation?\n<\/p>\n<p>\n  Fortunately, Emacs offers help here\ntoo. Code built into Emacs implements a\nsimple-to-use front end for\n  <em>SCCS<\/em>,\n  <em>RCS<\/em>,\n  <em>CVS<\/em>, or\nSubversion. The single command <code>Ctl-x v v<\/code> tries to\ndeduce the next logical version-control operation to do on the file\nyou are visiting. The operations this includes are registering a file,\nchecking out and locking it, and checking it back in (accepting a\nchange comment in a pop-up buffer).<sup><a href=\"#ftn.id2989834\" id=\"id2989834\">[139]<\/a><\/sup><\/p>\n<p>\n  Emacs also helps you view the change\nhistory of version-controlled files, and helps you back out changes\nyou don't want. It makes it easy to apply version-control operations\nto whole sets or project directory trees of files. In general, it does\na pretty good job of making version-control operations painless.\n<\/p>\n<p>\n  The implications of these features are larger than you might\nguess before you've gotten used to it. You'll find, once you get used\nto fast and easy version control, that it's extremely\nliberating. Because you know you can always revert to a known-good\nstate, you'll find you feel more free to develop in a fluid and\nexploratory way, trying lots of changes out to see their\neffects.\n<\/p>\n<h4 id=\"id2989869\"><a href=\"#id2989869\">\u00a7<\/a>Emacs and Profiling<\/h4>\n<p>\n  Surprise...this is perhaps the only phase of the development\ncycle in which Emacs front-ending does\n  <em>not<\/em>\n  offer substantial help. Profiling is an\nintrinsically batchy operation \u2014 instrument your program, run\nit, view the statistics, speed-tune the code with an editor,\nrepeat. There isn't much room for Emacs\nleverage in the profiling-specific parts of this cycle.\n<\/p>\n<p>\n  Nevertheless, there's a good tutorial reason for us to think\nabout Emacs and profiling. If you found\nyourself analyzing a\n  <em>lot<\/em>\n  of profiling reports, it\nmight pay you to write a mode in which a mouse click or keystroke on a\nprofile report line visited the source of the relevant function. This\nactually would be fairly easy to do using the\nEmacs \u2018tags\u2019 code. In fact, by\nthe time you read this, some other reader may already have written\nsuch a mode and contributed it to the public\nEmacs code base.\n<\/p>\n<p>\n  The real point here is again a philosophical one. Don't drudge\n\u2014 drudging wastes your time and productivity! If you find yourself\nspending a lot of time on the low-level mechanical parts of\ndevelopment, step back. Apply the Unix philosophy. Use your toolkit\nto automate or semi-automate the task.\n<\/p>\n<p>\n  Then give back something in return for all you've inherited, by\nposting your solution as open-source software to the Internet. Help\nliberate your fellow programmers from drudgery, too.\n<\/p>\n<h4 id=\"id2989926\"><a href=\"#id2989926\">\u00a7<\/a>Like an IDE, Only Better<\/h4>\n<p>\n  Earlier in this chapter we asserted that\nEmacs can give you capabilities resembling\nthose of a conventional integrated development environment, only\nbetter. By now you should have enough facts in hand to see how that\ncan be true. You can run entire development projects from inside\nEmacs, driving the low-level mechanics with\na few keystrokes and saving yourself the mental effort and disruption\nof constantly switching contexts.\n<\/p>\n<p>\n  The Emacs-enabled development style\ntrades away some capabilities of advanced IDEs, like graphical views\nof program structure. But those are frills. What\nEmacs gives you in return is flexibility\nand control. You're not limited by the imagination of the IDE designer:\nyou can tweak, customize, and add task-related intelligence\nusing Emacs Lisp. Also, Emacs is better at\nsupporting mixed-language development than conventional IDEs.\n<\/p>\n<p>\n  Finally, you're not limited to accepting what one small group of\nIDE developers sees fit to support. By keeping an eye on the\nopen-source community, you can benefit from the work of thousands of your\npeers, Emacs-using developers facing\nchallenges much like yours. This is much more effective \u2014 and\nmuch more fun.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2989640\" id=\"ftn.id2989640\">[138]<\/a><\/sup>\n  Look at\np+processes-&gt;compile\nunder the Emacs help menu for more\ninformation on these and related compilation-control\ncommands.\n<\/p><p><sup><a href=\"#id2989834\" id=\"ftn.id2989834\">[139]<\/a><\/sup>\n  See the subsection\nof the Emacs on-line documentation titled\n  <em>Version Control<\/em>\n  for more details on these and related\ncommands.\n<\/p><\/div>\n\n<h3 id=\"reusechapter\"><a href=\"#reusechapter\">\u00a7<\/a>Chapter 16. Reuse<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   When the superior man refrains from acting, his force is felt\nfor a thousand miles.\n  <\/p><div>\n  \n   \u2013 Tao Te Ching (as popularly mistranslated)\n  \n <\/div><\/blockquote>\n<p>\n  Reluctance to do unnecessary work is a great virtue in\nprogrammers. If the Chinese sage Lao-Tze were alive today and still\nteaching the way of the Tao, he would probably be mistranslated as:\nWhen the superior programmer refrains from coding, his force is felt\nfor a thousand miles. In fact, recent translators have suggested that\nthe Chinese term\n  \n   <em>wu-wei<\/em> \n  that has\ntraditionally been rendered as \u201cinaction\u201d or\n\u201crefraining from action\u201d should probably be read as\n\u201cleast action\u201d or \u201cmost efficient action\u201d or\n\u201caction in accordance with natural law\u201d, which is an even\nbetter description of good engineering practice!\n<\/p>\n<p>\n  Remember the Rule of Economy.  Re-inventing fire and the wheel\nfor every new project is terribly wasteful.  Thinking time is precious\nand very valuable relative to all the other inputs that go into\nsoftware development; accordingly, it should be spent solving new\nproblems rather than rehashing old ones for which known solutions\nalready exist. This attitude gives the best return both in the\n\u201csoft\u201d terms of developing human capital and in the\n\u201chard\u201d terms of economic return on development\ninvestment.\n<\/p>\n <blockquote>\n  <p>\n   Reinventing the wheel is bad not only because it\nwastes time, but because reinvented wheels are often square.  There is an\nalmost irresistible temptation to economize on reinvention time by\ntaking a shortcut to a crude and poorly-thought-out version, which\nin the long run often turns out to be false economy.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  The most effective way to avoid reinventing the wheel is to\nborrow someone else's design and implementation of it. In other\nwords, to reuse code.\n<\/p>\n<p>\n  Unix supports reuse at every level from individual library\nmodules up to entire programs, which Unix helps you script and\nrecombine.  Systematic reuse is one of the most important\ndistinguishing behaviors of Unix programmers, and the experience of\nusing Unix should teach you a habit of trying to prototype solutions\nby combining existing components with a minimum of new invention,\nrather than rushing to write standalone code that will only be used\nonce.\n<\/p>\n<p>\n  The virtuousness of code reuse is one of the great\napple-pie-and-motherhood verities of software development.  But many\ndevelopers entering the Unix community from a basis of experience in\nother operating systems have never learned (or have unlearned) the\nhabit of systematic reuse.  Waste and duplicative work is rife, even\nthough it seems to be against the interests both of those who pay for\ncode and those who produce it.  Understanding why such dysfunctional\nbehavior persists is the first step toward changing it.\n<\/p>\n\n<h3 id=\"ch16s01\"><a href=\"#ch16s01\">\u00a7<\/a>The Tale of J. Random Newbie<\/h3>\n<p>\n  Why do programmers reinvent wheels?  There are many reasons,\nreaching all the way from the narrowly technical to the psychology of\nprogrammers and the economics of the software production system.  The\ndamage from the endemic waste of programming time reaches all these\nlevels as well.\n<\/p>\n<p>\n  Consider the first, formative job experience of J. Random\nNewbie, a programmer fresh out of college.  Let us assume that he (or\nshe) has been taught the value of code reuse and is brimming with\nyouthful zeal to apply it.\n<\/p>\n<p>\n  Newbie's first project puts him on a team building some large\napplication.  Let's say for the sake of example that it's a GUI\nintended to help end users intelligently construct queries for and\nnavigate through a large database.  The project managers have\nassembled what they deem to be a suitable collection of tools and\ncomponents, including not merely a development language but many\nlibraries as well.\n<\/p>\n<p>\n  The libraries are crucial to the project.  They package many\nservices \u2014 from windowing widgets and network connections on up\nto entire subsystems like interactive help \u2014 that would\notherwise require immense quantities of additional coding, with a\nsevere impact on the project's budget and its ship date.\n<\/p>\n<p>\n  Newbie is a little worried about that ship date.  He may lack\nexperience, but he's read\n  <em>Dilbert<\/em> and heard a\nfew war stories from experienced programmers.  He knows management has\na tendency to what one might euphemistically call\n\u201caggressive\u201d schedules.  Perhaps he has read Ed Yourdon's\n  <em>Death March<\/em> <sup><a href=\"#Yourdon\" title=\"[Yourdon]\">[Yourdon]<\/a><\/sup>, which as long ago as 1996 noted\nthat a majority of projects are on a time and resource budget at least\n50% too tight, and that the trend is for that squeeze to get worse.\n<\/p>\n<p>\n  But Newbie is bright and energetic.  He figures his best chance\nof succeeding is to learn to use the tools and libraries that have\nbeen handed to him as intelligently as possible.  He limbers up his\ntyping fingers, hurls himself at the challenge...and enters hell.\n<\/p>\n<p>\n  Everything takes longer and is more painful than he expects.\nBeneath the surface gloss of their demo applications, the components\nhe is re-using seem to have edge cases in which they behave\nunpredictably or destructively \u2014 edge cases his code tickles\ndaily.  He often finds himself wondering what the library programmers\nwere thinking.  He can't tell, because the components are inadequately\ndocumented \u2014 often by technical writers who aren't programmers\nand don't think like programmers.  And he can't read the source code\nto learn what it is actually doing, because the libraries are opaque\nblocks of object code under proprietary licenses.\n<\/p>\n<p>\n  Newbie has to code increasingly elaborate workarounds for\ncomponent problems, to the point where the net gain from using the\nlibraries starts to look marginal.  The workarounds make his code\nprogressively grubbier.  He probably hits a few places where a library\nsimply cannot be made to do something crucially important that is\ntheoretically within its specifications.  Sometimes he is sure there\nis some way to actually make the black box perform, but he can't\nfigure out what it is.\n<\/p>\n<p>\n  Newbie finds that as he puts more strain on the libraries, his\ndebugging time rises exponentially.  His code is bedeviled with\ncrashes and memory leaks that have trace paths leading into the\nlibraries, into code he can't see or modify.  He knows most of those\ntrace paths probably lead back out to his code, but without source it\nis very difficult to trace through the bits he didn't write.\n<\/p>\n<p>\n  Newbie is growing horribly frustrated.  He had heard in college\nthat in industry, a hundred lines of finished code a week is\nconsidered good performance. He had laughed then, because he was many\ntimes more productive than that on his class projects and the code he\nwrote for fun.  Now it's not funny any more.  He is wrestling not\nmerely with his own inexperience but with a cascade of problems\ncreated by the carelessness or incompetence of others \u2014 problems\nhe can't fix, but can only work around.\n<\/p>\n<p>\n  The project schedule is slipping.  Newbie, who dreamed of being\nan architect, finds himself a bricklayer trying to build with bricks\nthat won't stack properly and that crumble under load-bearing\npressure.  But his managers don't want to hear excuses from a novice\nprogrammer; complaining too loudly about the poor quality of the\ncomponents is likely to get him in political trouble with the senior\npeople and managers who selected them.  And even if he could win that\nbattle, changing components would be a complicated proposition\ninvolving batteries of lawyers peering narrowly at licensing\nterms.\n<\/p>\n<p>\n  Unless Newbie is very, very lucky, he is not going to be able to\nget library bugs fixed within the lifetime of his project.  In his\nsaner moments, he may realize that the working code in the libraries \ndoesn't draw his attention the way the bugs and omissions do. He'd love\nto sit down for a clarifying chat with the component developers; he\nsuspects they can't be the idiots their code sometimes suggests, just\nprogrammers like him working within a system that frustrates their\nattempts to do the right thing.  But he can't even find out who they are\n\u2014 and if he could, the software vendor they work for probably\nwouldn't let them talk to him.\n<\/p>\n<p>\n  In desperation, Newbie starts making his own bricks \u2014\nsimulating less stable library services with more stable ones and\nwriting his own implementations from scratch.  His replacement code,\nbecause he has a complete mental model of it that he can refresh by\nrereading, tends to work relatively well and be easier to debug than\nthe combination of opaque components and workarounds it\nreplaces.\n<\/p>\n<p>\n  Newbie is learning a lesson; the less he relies on other\npeoples' code, the more lines of code he can get written.  This lesson\nfeeds his ego.  Like all young programmers, deep down he thinks he is\nsmarter than anyone else.  His experience seems, superficially, to be\nconfirming this.  He begins building his own personal toolkit, one\nbetter fitted to his hand.\n<\/p>\n<p>\n  Unfortunately, the roll-your-own reflexes Newbie is acquiring\nare a short-term local optimization that will cause long-term\nproblems.  He may get more lines of code written, but the actual value\nof what he produces is likely to drop substantially relative to what\nit would have been if he were doing successful reuse.  More code does\nnot equal better code, not when it's written at a lower level and\nlargely devoted to reinventing wheels.\n<\/p>\n<p>\n  Newbie has at least one more demoralizing experience in store,\nwhen he changes jobs.  He is likely to discover that he can't take his\ntoolkit with him.  If he walks out of the building with code he wrote\non company time, his old employers could well regard this as\nintellectual-property theft.  His new employers, knowing this, are not\nlikely to react well if he admits to reusing any of his old\ncode.\n<\/p>\n<p>\n  Newbie could well find his toolkit is useless even if he can\nsneak it into the building at his new job.  His new employers may use\na different set of proprietary tools, languages, and libraries. It is\nlikely he will have to learn a somewhat new set of techniques and\nreinvent a new set of wheels each time he changes projects.\n<\/p>\n<p>\n  Thus do programmers have reuse (and other good practices that\ngo with it, like modularity and transparency) systematically\nconditioned out of them by a combination of technical problems,\nintellectual-property barriers, politics, and personal ego needs.\nMultiply J. Random Newbie by a hundred thousand, age him by decades,\nand have him grow more cynical and more used to the system year by\nyear.  There you have the state of much of the software industry, a\nrecipe for enormous waste of time and capital and human skill \u2014\neven\n  <em>before<\/em>\n  you factor in vendors' market-control\ntactics, incompetent management, impossible deadlines, and all the\nother pressures that make doing good work difficult.\n<\/p>\n<p>\n  The professional culture that springs from J. Random Newbie's\nexperiences will reflect them in the large.  Programming shops will\nhave a ferocious Not Invented Here complex.  They will be poisonously\nambivalent about code reuse, pushing inadequate but heavily marketed\nvendor components on their programmers in order to meet schedule\ncrunches, while simultaneously rejecting reuse of the programmers' own\ntested code.  They will churn out huge volumes of ad-hoc, duplicative\nsoftware produced by programmers who know the results will be garbage but\nare glumly resigned to never being able to fix anything but their own\nindividual pieces.\n<\/p>\n<p>\n  The closest equivalent of code reuse to emerge in such a\nculture will be a dogma that code once paid for can never be thrown\naway, but must instead be patched and kluged even when all parties\nknow that it would be better to scrap and start anew.  The\nproducts of this culture will become progressively more bloated and\nbuggy over time even when every individual involved is trying his or\nher hardest to do good work.\n<\/p>\n\n<h3 id=\"ch16s02\"><a href=\"#ch16s02\">\u00a7<\/a>Transparency as the Key to Reuse<\/h3>\n<p>\n  We field-tested the tale of J. Random Newbie on a number of\nexperienced programmers.  If you the reader are one yourself, we\nexpect you responded to it much as they did: with groans of\nrecognition.  If you are not a programmer but you manage programmers,\nwe sincerely hope you found it enlightening.  The tale is intended to\nillustrate the ways in which different levels of pressure against reuse\nreinforce each other to create a magnitude of problem not linearly\npredictable from any individual cause.\n<\/p>\n<p>\n  So accustomed are most of us to the background assumptions of\nthe software industry that it can take considerable mental effort\nbefore the primary causes of this problem can be separated from the\naccidents of narrative.  But they are not, in the end, very\ncomplex.\n<\/p>\n<p>\n  At the bottom of most of J. Random Newbie's troubles (and the\nlarge-scale quality problems they imply) is\ntransparency\n  \u2014 or, rather, the lack of it.  You can't fix what you can't see\ninside.  In fact, for any software with a nontrivial API, you can't\neven properly\n  <em>use<\/em>\n  what you can't see inside.\nDocumentation is inadequate not merely in practice but in principle;\nit cannot convey all the nuances that the code embodies.\n<\/p>\n<p>\n  In\n  <a href=\"#transparencychapter\" title=\"Chapter 6. Transparency\">\n   Chapter 6<\/a>, we\nobserved how central transparency is to good software.\nObject-code-only components destroy the transparency of a software\nsystem, On the other hand, the frustrations of code reuse are far less\nlikely to bite when the code you are attempting to reuse is available\nfor reading and modification. Well-commented source code is its own\ndocumentation. Bugs in source code can be fixed. Source can be\ninstrumented and compiled for debugging to make probing its behavior\nin obscure cases easier. And if you need to change its behavior, you\ncan do that.\n<\/p>\n<p>\n  There is another vital reason to demand source code.  A lesson\nUnix programmers have learned through decades of constant change is\nthat source code lasts, object code doesn't.  Hardware platforms\nchange, service components like support libraries change, the\noperating system grows new APIs and deprecates old ones.  Everything\nchanges \u2014 but opaque binary executables cannot adapt to change.\nThey are brittle, cannot be reliably forward-ported, and have to be\nsupported with increasingly thick and error-prone layers of emulation\ncode.  They lock users into the assumptions of the people who built\nthem.  You need source because, even if you have neither the\nintention nor the need to change the software, you will have to\nrebuild it in new environments to keep it running.\n<\/p>\n<p>\n  The importance of transparency and the code-legacy problem are\nreasons that you should require the code you reuse to be open to\ninspection and modification.<sup><a href=\"#ftn.id2990172\" id=\"id2990172\">[140]<\/a><\/sup>\n  It is not a complete argument for what\nis now called \u2018open source\u2019; because \u2018open\nsource\u2019 has rather stronger implications than simply requiring\ncode to be transparent and visible.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2990172\" id=\"ftn.id2990172\">[140]<\/a><\/sup>\n  NASA, which consciously\nbuilds software intended to have a service life of decades, has learned\nto insist on source-code availability for all space avionics\nsoftware.\n<\/p><\/div>\n\n<h3 id=\"ch16s03\"><a href=\"#ch16s03\">\u00a7<\/a>From Reuse to Open Source<\/h3>\n<p>\n  In the early days of Unix, components of the operating system,\nits libraries, and its associated utilities were passed around as\nsource code; this openness was a vital part of the Unix culture.  We\ndescribed in\n  <a href=\"#historychapter\" title=\"Chapter 2. History\">\n   Chapter 2<\/a>\n  how,\nwhen this tradition was disrupted after 1984, Unix lost its initial\nmomentum. We have also described how, a decade later, the rise of the\nGNU\n  toolkit and\nLinux\n  prompted a rediscovery of the value of open-source code.\n<\/p>\n<p>\n  Today, open-source code is again one of the most powerful tools\nin any Unix programmer's kit. Accordingly, though the explicit concept\nof \u201copen source\u201d and the most widely used\nopen-source\n  licenses are decades younger than Unix itself, it's important to\nunderstand both to do leading-edge development in today's\nUnix culture.\n<\/p>\n<p>\n  Open source\n  relates to code reuse in much the way romantic love relates to sexual\nreproduction \u2014 it's possible to explain the former in terms of\nthe latter, but to do so is to risk overlooking much of what makes the\nformer interesting.  Open source does not reduce to merely being a\ntactic for supporting reuse in software development.  It is an\nemergent phenomenon, a social contract among developers and users that\ntries to secure several advantages related to transparency.\nAs such, there are several different ways to approaching an\nunderstanding of it.\n<\/p>\n<p>\n  Our historical description earlier in this book chose one angle\nby focusing on causal and cultural relationships between Unix and open\nsource.  We'll discuss the institutions and tactics of open-source\ndevelopment in\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>. In discussing the theory and practice\nof code reuse, it's useful to think of open source more specifically,\nas a direct response to the problems we dramatized in the tale of\nJ. Random Newbie.\n<\/p>\n<p>\n  Software developers want the code they use to be transparent.\nFurthermore, they don't want to lose their toolkits and their\nexpertise when they change jobs.   They get tired of being victims, fed\nup with being frustrated by blunt tools and intellectual-property\nfences and having to repeatedly re-invent the wheel.\n<\/p>\n<p>\n  These are the motives for open source that flow from J. Random\nNewbie's painful initiatory experience with reuse.  Ego needs play a\npart here, too; they give pervasive emotional force to what would\notherwise be a bloodless argument about engineering best practices.\nSoftware developers are like every other kind of craftsman and\nartificer; they want, not so secretly, to be artists.  They have the\ndrives and needs of artists, including the desire to have an audience.\nThey not only want to reuse code, they want their code to be reused.\nThere is an imperative here that goes beyond and overrides short-term\neconomic goal-seeking and that cannot be satisfied by closed-source\nsoftware production.\n<\/p>\n<p>\n  Open source is a kind of ideological preemptive strike on all\nthese problems.  If the root of most of J. Random Newbie's problems\nwith reuse is the opacity of closed-source code, then the\ninstitutional assumptions that produce closed-source code must be\nsmashed.  If corporate territoriality is a problem, it must be\nattacked or bypassed until the corporations have caught on to how\nself-destructive their territorial reflexes are.  Open source is what\nhappens when code reuse gets a flag and an army.\n<\/p>\n<p>\n  Accordingly, since the late 1990s, it no longer makes any\nsense to try to recommend strategies and tactics for code reuse\nwithout talking about open source, open-source practices, open-source\nlicensing, and the open-source community.  Even if those issues could\nbe separated elsewhere, they have become inextricably bound together\nin the Unix world.\n<\/p>\n<p>\n  In the remainder of this chapter, we'll survey various issues\nassociated with re-using open-source code: evaluation, documentation,\nand licensing. In\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>\n  we'll discuss the open-source\ndevelopment model more generally, and examine the conventions you\nshould follow when you are releasing code for others to use.\n<\/p>\n\n<h3 id=\"ch16s04\"><a href=\"#ch16s04\">\u00a7<\/a>The Best Things in Life Are Open<\/h3>\n<p>\n  On the Internet, literally terabytes of Unix sources for systems\nand applications software, service libraries, GUI toolkits and\nhardware drivers are available for the taking. You can have most built\nand running in minutes with standard tools.  The mantra is\n  <code>.\/configure; make; make install<\/code>; usually you have to be root to do the install part.\n<\/p>\n<p>\n  People from outside the Unix world (especially non-technical\npeople) are prone to think open-source\n  (or \u2018free\u2019) software is\nnecessarily inferior to the commercial kind, that it's shoddily made\nand unreliable and will cause more headaches than it saves. They miss\nan important point: in general, open-source software is written by\npeople who care about it, need it, use it themselves, and are putting\ntheir individual reputations among their peers on the line by\npublishing it.  They also tend to have less of their time consumed by\nmeetings, retroactive design changes, and bureaucratic overhead. They\nare therefore both more strongly motivated and better positioned to do\nexcellent work than wage slaves toiling Dilbert-like to meet\nimpossible deadlines in the cubicles of proprietary software\nhouses.\n<\/p>\n<p>\n  Furthermore, the open-source user community (those peers) is not\nshy about nailing bugs, and its standards are high. Authors who put\nout substandard work experience a lot of social pressure to fix\ntheir code or withdraw it, and can get a lot of skilled help fixing\nit if they choose. As a result, mature open-source packages are\ngenerally of high quality and often functionally superior to any\nproprietary equivalent. They may lack polish and have documentation\nthat assumes much, but the vital parts will usually work quite\nwell.\n<\/p>\n<p>\n  Besides the peer-review effect, another reason to expect better\nquality is this: in the open-source world developers are never forced\nby a deadline to close their eyes, hold their noses, and ship.  A\nmajor consequent difference between open-source practice and elsewhere\nis that a release level of 1.0 actually means the software is ready to\nuse.  In fact, a version number of 0.90 or above is a fairly reliable\nsignal that the code is production-ready, but the developers are not\nquite ready to bet their reputations on it.\n<\/p>\n<p>\n  If you are a programmer from outside the Unix world, you may\nfind this claim difficult to believe. If so, consider this: on modern\nUnixes, the C\n  compiler itself is almost invariably open source. The Free Software\nFoundation's GNU Compiler Collection\n  (GCC) is so powerful, so well\ndocumented, and so reliable that there is effectively no proprietary\nUnix compiler market left, and it has become normal for Unix vendors\nto port GCC to their platforms rather than do in-house compiler\ndevelopment.\n<\/p>\n<p>\n  The way to evaluate an open-source\n  package is to read its documentation and\nskim some of its code. If what you see appears to be competently\nwritten and documented with care, be encouraged. If there also is\nevidence that the package has been around for a while and has\nincorporated substantial user feedback, you may bet that it is quite\nreliable (but test anyway).\n<\/p>\n<p>\n  A good gauge of maturity and the volume of user feedback is the\nnumber of people besides the original author mentioned in the\n  <code>README<\/code> and project news or history files in the\nsource distribution. Credits to lots of people for sending in fixes\nand patches are signs both of a significant user base keeping the\nauthors on their toes, and of a conscientious maintainer who is\nresponsive to feedback and will take corrections.  It is also \nan indication that, if early code tends to be a minefield of bugs, \nthere has since been a thundering herd run through it without\ntoo many recent explosions.\n<\/p>\n<p>\n  It's also a good omen when the software has its own Web page,\non-line FAQ (Frequently Asked Questions) list, and an associated\nmailing list or Usenet\n  newsgroup. These are all signs that a live and substantial community\nof interest has grown up around the software. On Web pages, recent\nupdates and an extensive mirror list are reliable signs of a project\nwith a vigorous user community. Packages that are duds just don't get\nthis kind of continuing investment, because they can't reward\nit.\n<\/p>\n<p>\n  Ports to multiple platforms are also a valuable indication of a\ndiversified user base.  Project pages tend to advertise new ports\nprecisely because they signal credibility.\n<\/p>\n<p>\n  Here are some examples of what Web pages associated with\nhigh-quality open-source software look like:\n<\/p>\n <div><ul><li>\n    <p>\n     <a href=\"http:\/\/www.gimp.org\/\">\n      GIMP<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/www.gnome.org\">\n      GNOME<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/www.kde.org\">\n      KDE<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/www.python.org\">\n      Python<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/www.kernel.org\">\n      The Linux kernel<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/www.postgresql.org\">\n      PostgreSQL<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/xfree86.org\">\n      XFree86<\/a>\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <a href=\"http:\/\/www.info-zip.org\/pub\/infozip\/\">\n      InfoZip<\/a>\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  Looking at Linux distributions is another good way to find quality.\nDistribution-makers for Linux and other open-source Unixes carry\na lot of specialist expertise about which projects are best-of-breed\n\u2014 that's a large part of the value they add when they integrate\na release.  If you are already using an open-source Unix, something\nelse to check is whether the package you are evaluating is already\ncarried by your distribution.\n<\/p>\n\n<h3 id=\"ch16s05\"><a href=\"#ch16s05\">\u00a7<\/a>Where to Look?<\/h3>\n<p>\n  Because so much open source is available in the Unix\nworld, skill at finding code to reuse can have an enormous payoff\n\u2014 much greater than is the case for other operating systems.\nSuch code comes in many forms: individual code snippets and\nexamples, code libraries, utilities to be reused in scripts.  Under\nUnix most code reuse is not a matter of actual cut-and-paste into\nyour program \u2014 in fact, if you find yourself doing that, there\nis almost certainly a more graceful mode of reuse that you are\nmissing. Accordingly, one of the most useful skills to cultivate under\nUnix is a good grasp of all the different ways to glue together code,\nso you can use the Rule of Composition.\n<\/p>\n<p>\n  To find re-usable code, start by looking under your nose. Unixes\nhave always featured a rich toolkit of re-usable utilities and \nlibraries; modern ones, such as any current Linux system, include\nthousands of programs, scripts, and libraries that may be re-usable.  \nA simple\n  <code>man -k<\/code> search with a few keywords often yields useful results.\n<\/p>\n<p>\n  To begin to grasp something of the amazing wealth of resources\nout there, surf to SourceForge, ibiblio, and Freshmeat.net.  Other\nsites as important as these three may exist by the time you read this\nbook, but all three of these have shown continuing value and\npopularity over a period of years, and seem likely to endure.\n<\/p>\n<p>\n  <a href=\"http:\/\/www.sourceforge.net\">\n   SourceForge<\/a>\n  is a\ndemonstration site for software specifically designed to support\ncollaborative development, complete with associated project-management\nservices. It is not merely an archive but a free development-hosting\nservice, and in mid-2003 is undoubtedly the largest single hub of\nopen-source activity in the world.\n<\/p>\n<p>\n  The Linux archives at\n  <a href=\"http:\/\/www.ibiblio.org\">\n   ibiblio<\/a>\n  were the largest in the\nworld before SourceForge.  The ibiblio archives are passive, simply a\nplace to publish packages.  They do, however, have a better interface to\nthe World Wide Web than most passive sites (the program that creates\nits Web look and feel was one of our case studies in the discussion of\nPerl\n  in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>). It's also the home site of the Linux\nDocumentation Project, which maintains many documents that are\nexcellent resources for Unix users and developers.\n<\/p>\n<p>\n  <a href=\"http:\/\/www.freshmeat.net\">\n   Freshmeat<\/a>\n  is a\nsystem dedicated to providing release announcements of new software,\nand new releases of old software.  It lets users and third parties\nattach reviews to releases.\n<\/p>\n<p>\n  These three general-purpose sites contain code in many\nlanguages, but most of their content is C\n  or C++. There are also sites specialized\naround some of the interpreted languages as discussed in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>.\n<\/p>\n<p>\n  The CPAN archive is the central repository for useful free code\nin Perl. It is easily reached from the\n  <a href=\"http:\/\/www.perl.com\/perl\">\n   Perl home\npage<\/a>.\n<\/p>\n<p>\n  The Python Software Activity makes an archive of\nPython\n  software and documentation available at the\n  <a href=\"http:\/\/www.python.org\">\n   Python Home Page<\/a>.\n<\/p>\n<p>\n  Many Java\n  applets and pointers to other sites featuring free Java software are\nmade available at the\n  <a href=\"http:\/\/java.sun.com\/applets\/\">\n   Java\nApplets page<\/a>.\n<\/p>\n<p>\n  One of the most valuable ways you can invest your time as a Unix\ndeveloper is to spend time wandering around these sites learning\nwhat is available for you to use. The coding time you save may be\nyour own!\n<\/p>\n<p>\n  Browsing the package metadata is a good idea, but don't stop\nthere. Sample the code, too.  You'll get a better grasp on what the code\nis doing, and be able to use it more effectively.\n<\/p>\n<p>\n  More generally, reading code is an investment in the future.\nYou'll learn from it \u2014 new techniques, new ways to partition\nproblems, different styles and approaches.  Both using the code and\nlearning from it are valuable rewards.  Even if you don't use the\ntechniques in the code you study, the improved definition of the\nproblem you get from looking at other peoples' solutions may well help\nyou invent a better one of your own.\n<\/p>\n<p>\n  Read before you write; develop the habit of reading code.  There\nare seldom any completely new problems, so it is almost always\npossible to discover code that is close enough to what you need to be\na good starting point.  Even when your problem is genuinely novel, it\nis likely to be genetically related to a problem someone else has\nsolved before, so the solution you need to develop is likely to be\nrelated to some pre-existing one as well.\n<\/p>\n\n<h3 id=\"ch16s06\"><a href=\"#ch16s06\">\u00a7<\/a>Issues in Using Open-Source Software<\/h3>\n<p>\n  There are three major issues in using or re-using\nopen-source\n  software; quality, documentation, and licensing terms. We've seen\nabove that if you exercise a little judgment in picking through your\nalternatives, you will generally find one or more of quite respectable\nquality.\n<\/p>\n<p>\n  Documentation is often a more serious issue. Many high-quality\nopen-source packages are less useful than they technically ought to be\nbecause they are poorly documentated. Unix tradition encourages a\nrather hieratic style of documentation, one which (while it may\ntechnically capture all of a package's features) assumes that the\nreader is intimately familiar with the application domain and reading\nvery carefully. There are good reasons for this, which we'll discuss\nin\n  <a href=\"#documentationchapter\" title=\"Chapter 18. Documentation\">\n   Chapter 18<\/a>, but the\nstyle can present a bit of a barrier.  Fortunately, extracting value\nfrom it is a learnable skill.\n<\/p>\n<p>\n  It is worth doing a Web search for phrases including the \nsoftware package, or topic keywords, and the string \u201cHOWTO\u201d or\n\u201cFAQ\u201d. These queries will often turn up documentation\nmore useful to novices than the man page.\n<\/p>\n<p>\n  The most serious issue in reusing open-source software\n(especially in any kind of commercial product) is understanding what\nobligations, if any, the package's license puts upon you. In the next\ntwo sections we'll discuss this issue in detail.\n<\/p>\n\n<h3 id=\"ch16s07\"><a href=\"#ch16s07\">\u00a7<\/a>Licensing Issues<\/h3>\n<p>\n  Anything that is not public domain has a copyright, possibly\nmore than one. Under U.S. federal law, the authors of a work hold\ncopyright even if there is no copyright notice.\n<\/p>\n<p>\n  Who counts as an author under copyright law can be complicated,\nespecially for software that has been worked on by many hands.  This\nis why licenses are important.  They can authorize uses of code in\nways that would be otherwise impermissible under copyright law and,\ndrafted appropriately, can protect users from arbitrary actions by the\ncopyright holders.\n<\/p>\n<p>\n  In the proprietary software world, the license terms are designed to\nprotect the copyright. They're a way of granting a few rights to\nusers while reserving as much legal territory as possible for the\nowner (the copyright holder). The copyright holder is very\nimportant, and the license logic so restrictive that the exact\ntechnicalities of the license terms are usually unimportant.\n<\/p>\n<p>\n  As will be seen below, the copyright holder typically uses the\ncopyright to protect the license, which makes the code freely\navailable under terms he intends to perpetuate\nindefinitely. Otherwise, only a few rights are reserved and most\nchoices pass to the user. In particular, the copyright holder cannot\nchange the terms on a copy you already have. Therefore, in open-source\nsoftware the copyright holder is almost irrelevant \u2014 but the\nlicense terms are very important.\n<\/p>\n<p>\n  Normally the copyright holder of a project is the current\nproject leader or sponsoring organization. Transfer of the project to\na new leader is often signaled by changing the copyright holder.\nHowever, this is not a hard and fast rule; many open-source projects\nhave multiple copyright holders, and there is no instance on record of\nthis leading to legal problems. Some projects choose to assign\ncopyright to the Free Software Foundation, on the theory that it has\nan interest in defending open source and lawyers available to do\nit.\n<\/p>\n<h4 id=\"id2991059\"><a href=\"#id2991059\">\u00a7<\/a>What Qualifies as Open Source<\/h4>\n<p>\n  For licensing purposes, we can distinguish several different\nkinds of rights that a license may convey. There are rights to copy and\nredistribute, rights to use, rights to modify for personal use, and\nrights to redistribute modified copies. A license may restrict or\nattach conditions to any of these rights.\n<\/p>\n<p>\n  The\n  <a href=\"http:\/\/www.opensource.org\/osd.html\">\n   Open Source\nDefinition<\/a>\n  is the result of a great deal of thought about what\nmakes software \u201copen source\u201d or (in older terminology)\n\u201cfree\u201d.  It is widely accepted in the open-source\ncommunity as an articulation of the social contract among open-source\ndevelopers. Its constraints on licensing impose the following\nrequirements:\n<\/p>\n <div><ul><li>\n    <p>\n     An unlimited right to copy be granted.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     An unlimited right to redistribute in unmodified form be granted.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     An unlimited right to modify for personal use be granted.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  The guidelines prohibit restrictions on redistribution of\nmodified binaries; this meets the needs of software distributors, who\nneed to be able to ship working code without encumbrance. It allows\nauthors to require that modified sources be redistributed as pristine\nsources plus patches, thus establishing the author's intentions and an\n\u201caudit trail\u201d of any changes by others.\n<\/p>\n<p>\n  The OSD is the legal definition of the \u201cOSI Certified Open\nSource\u201d certification mark, and as good a definition of\n\u201cfree software\u201d as anyone has ever come up with. All of\nthe standard licenses (MIT,\nBSD, Artistic, GPL\/LGPL,\nand MPL) meet it (though some, like GPL, have other restrictions which\nyou should understand before choosing it).\n<\/p>\n<p>\n  Note that licenses that allow only noncommercial use do not\nqualify as open-source licenses, even if they are based on GPL or some\nother standard license. Such licenses discriminate against particular\noccupations, persons, and groups, a practice which the OSD's Clause 5\nexplicitly forbids.\n<\/p>\n<p>\n  Clause 5 was written after years of painful experience.\nNo-commercial-use licenses turn out to have the problem that there is\nno bright-line legal test for what sort of redistribution qualifies as\n\u2018commercial\u2019.  Selling the software as a product\nqualifies, certainly.  But what if it were distributed at a nominal\nprice of zero in conjunction with other software or data, and a price\nis charged for the whole collection?  Would it make a difference\nwhether the software were essential to the function of the whole\ncollection?\n<\/p>\n<p>\n  Nobody knows.  The very fact that no-commercial-use licenses\ncreate uncertainty about a redistributor's legal exposure is a serious\nstrike against them. One of the objectives of the OSD is to ensure\nthat people in the distribution chain of OSD-conforming software do\nnot need to consult with intellectual-property lawyers\nto know what their rights are.  OSD forbids complicated restrictions \nagainst persons, groups, and occupations partly so that people\ndealing with collections of software will not face a combinatorial\nexplosion of slightly differing (and perhaps conflicting) restrictions\non what they can do with it.\n<\/p>\n<p>\n  This concern is not hypothetical, either.  One important part of\nthe open-source distribution chain is CD-ROM distributors who\naggregate it in useful collections ranging from simple anthology\nCDs up to bootable operating systems.  Restrictions that would make life\nprohibitively complicated for CD-ROM distributors, or others trying to\nspread open-source software commercially, have to be forbidden.\n<\/p>\n<p>\n  On the other hand, the OSD has nothing to say about the laws of\nyour jurisdiction.  Some countries have laws against exporting certain\nrestricted technologies to named \u2018rogue states\u2019. The OSD\ncannot negate those, it only says that licensors may not add\nrestrictions of their own.\n<\/p>\n<h4 id=\"id2993547\"><a href=\"#id2993547\">\u00a7<\/a>Standard Open-Source Licenses<\/h4>\n<p>\n  Here are the standard open-source license terms you are likely\nto encounter. The abbreviations listed here are in general use.\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <a href=\"http:\/\/www.opensource.org\/licenses\/mit-license.html\">\n      MIT<\/a>\n    \n   <\/dt>\n   <dd>\n    <p>\n     MIT X Consortium license (like BSD's but with no advertising\nrequirement)\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <a href=\"http:\/\/www.opensource.org\/licenses\/bsd-license.html\">\n      BSD<\/a>\n    \n   <\/dt>\n   <dd>\n    <p>\n     University of California at Berkeley Regents copyright (used on\nBSD\n     code)\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <a href=\"http:\/\/www.opensource.org\/licenses\/artistic-license.html\">\n      Artistic License<\/a>\n    \n   <\/dt>\n   <dd>\n    <p>\n     Same terms as Perl Artistic License\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <a href=\"http:\/\/www.gnu.org\/copyleft.html\">\n      GPL<\/a>\n    \n   <\/dt>\n   <dd>\n    <p>\n     GNU General Public License\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <a href=\"http:\/\/www.gnu.org\/copyleft.html\">\n      LGPL<\/a>\n    \n   <\/dt>\n   <dd>\n    <p>\n     Library (or \u2018Lesser\u2019) GPL\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <a href=\"http:\/\/www.opensource.org\/licenses\/MPL-1.1.html\">\n      MPL<\/a>\n    \n   <\/dt>\n   <dd>\n    <p>\n     Mozilla Public License\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  We'll discuss these licenses in more detail, from a developer's\npoint of view, in\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>.  For the purposes of this chapter, the\nonly important distinction among them is whether they are infectious\nor not.  A license is\n  <em>infectious<\/em> if it requires\nthat any derivative work of the licensed software also be placed under\nits terms.\n<\/p>\n<p>\n  Under these licenses, the only kind of open-source use you\nshould really worry about is actual incorporation of the\nfree-software code into a proprietary product (as opposed, say, to\nmerely using open-source development tools to make your product).\nIf you're prepared to include proper license acknowledgements and\npointers to the source code you're using in your product\ndocumentation, even direct incorporation should be safe provided\nthe license is not infectious.\n<\/p>\n<p>\n  The GPL is both the most widely used and the most controversial\ninfectious license.  And it is clause 2(b), requiring\nthat any derivative work of a GPLed program itself be GPLed, that\ncauses the controversy. (Clause 3(b) requiring licensors to make\nsource available on physical media on demand used to cause some,\nbut the Internet explosion has made publishing source code archives\nas required by 3(a) so cheap that nobody worries about the source-publication\nrequirement any more.)\n<\/p>\n<p>\n  Nobody is quite certain what the \u201ccontains or is derived\nfrom\u201d in clause 2(b) means, nor what kinds of use are protected\nby the \u201cmere aggregation\u201d language a few paragraphs later.\nContentious issues include library linking and inclusion of\nGPL-licensed header files. Part of the problem is that the U.S. copyright \nstatutes do not define what derivation is; it has been left to the\ncourts to hammer out definitions in case law, and computer software is\nan area in which this process (as of mid-2003) has barely begun.\n<\/p>\n<p>\n  At one end, the \u201cmere aggregation\u201d certainly makes\nit safe to ship GPLed software on the same media with your proprietary\ncode, provided they do not link to or call each other.  They may even\nbe tools operating on the same file formats or on-disk structures;\nthat situation, under copyright law, would not make one a derivative\nof the other.\n<\/p>\n<p>\n  At the other end, splicing GPLed code into your proprietary\ncode, or linking GPLed object code to yours, certainly does make your\ncode a derivative work and requires it to be GPLed.\n<\/p>\n<p>\n  It is generally believed that one program may execute a second\nprogram as a subprocess without either program becoming thereby a\nderivative work of the other.\n<\/p>\n<p>\n  The case that causes dispute is dynamic linking of shared\nlibraries.  The Free Software Foundation's\n  position is that if a\nprogram calls another program as a shared library, then that program\nis a derivative work of the library.  Some programmers think this\nclaim is overreaching.  There are technical, legal, and political\narguments on both sides that we won't rehash here.  Since the Free\nSoftware Foundation\n  wrote and owns the license, it would\nbe prudent to behave as if the FSF's position is correct until a court\nrules otherwise.\n<\/p>\n<p>\n  Some people think the 2(b) language is deliberately designed to\ninfect every part of any commercial program that uses even a snippet\nof GPLed code; such people refer to it as the GPV, or \u201cGeneral\nPublic Virus\u201d. Others think the \u201cmere aggregation\u201d\nlanguage covers everything short of mixing GPL and non-GPL code in the\nsame compilation or linkage unit.\n<\/p>\n<p>\n  This uncertainty has caused enough agitation in the open-source\ncommunity that the FSF had to develop the special, slightly more\nrelaxed \u201cLibrary GPL\u201d (which they have since renamed the\n\u201cLesser GPL\u201d) to reassure people they could continue to\nuse runtime libraries that came with the FSF's GNU compiler\ncollection.\n<\/p>\n<p>\n  You'll have to choose your own interpretation of clause 2(b);\nmost lawyers will not understand the technical issues involved, and there is no\ncase law. As a matter of empirical fact, the\nFSF\n  has never (from its founding in 1984 to mid-2003, at least) sued\nanyone under the GPL but it has enforced the GPL by threatening\nlawsuit, in all known cases successfully.  And, as another\nempirical fact, Netscape includes the source and object of a GPLed\nprogram with the commercial distribution of its Netscape Navigator\nbrowser.\n<\/p>\n<p>\n  The MPL and LGPL are infectious in a more limited way than GPL.\nThey explicitly allow linking with proprietary code without turning\nthat code into a derivative work, provided all traffic between the\nGPLed and non-GPLed code goes through a library API or other\nwell-defined interface.\n<\/p>\n<h4 id=\"id2993874\"><a href=\"#id2993874\">\u00a7<\/a>When You Need a Lawyer<\/h4>\n<p>\n  This section is directed to commercial developers considering\nincorporating software that falls under one of these standard licenses\ninto closed-source products.\n<\/p>\n<p>\n  Having gone through all this legal verbiage, the expected thing\nfor us to do at this point is to utter a somber disclaimer to the\neffect that we are not lawyers, and that if you have any\ndoubts about the legality of something you want to do with open-source\nsoftware, you should immediately consult a lawyer.\n<\/p>\n<p>\n  With all due respect to the legal profession, this would be\nfearful nonsense. The language of these licenses is as clear as\nlegalese gets \u2014 they were written to be clear \u2014 and should\nnot be at all hard to understand if you read it carefully. The lawyers\nand courts are actually more confused than you are. The law of\nsoftware rights is murky, and case law on open-source licenses is\n(as of mid-2003) nonexistent; no one has ever been sued under\nthem.\n<\/p>\n<p>\n  This means a lawyer is unlikely to have a significantly better\ninsight than a careful lay reader. But lawyers are professionally\nparanoid about anything they don't understand. So if you ask one, he\nis rather likely to tell you that you shouldn't go anywhere near\nopen-source software, despite the fact that he probably doesn't\nunderstand the technical aspects or the author's intentions anywhere\nnear as well as you do.\n<\/p>\n<p>\n  Finally, the people who put their work under open-source\nlicenses\n  are\ngenerally not mega-corporations attended by schools of lawyers looking\nfor blood in the water; they're individuals or volunteer groups who\nmainly want to give their software away.  The few exceptions (that is,\nlarge companies both issuing under open-source licenses and with\nmoney to hire lawyers) have a stake in open source and don't want to\nantagonize the developer community that produces it by stirring up\nlegal trouble.  Therefore, your odds of getting hauled into court on\nan innocent technical violation are probably lower than your chances\nof being struck by lightning in the next week.\n<\/p>\n<p>\n  This isn't to say you should treat these licenses as jokes. That\nwould be disrespectful of the creativity and sweat that went into the\nsoftware, and you wouldn't enjoy being the first litigation target of\nan enraged author no matter how the lawsuit came out. But in the\nabsence of definitive case law, a visible good-faith effort to meet\nthe author's intentions is 99% of what you can do; the additional 1%\nof protection you might (or might not) get by consulting a lawyer is\nunlikely to make a difference.\n<\/p>\n\n\n<h3 id=\"portabilitychapter\"><a href=\"#portabilitychapter\">\u00a7<\/a>Chapter 17. Portability<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   The realization that the operating systems of the target machines\nwere as great an obstacle to portability as their hardware architecture\nled us to a seemingly radical suggestion: to evade that part of the\nproblem altogether by moving the operating system itself.\n  <\/p><div>\n  \n   \u2013 &lt;authorgroup&gt;Steve JohnsonDennis Ritchie&lt;\/authorgroup&gt;\nPortability of C Programs and the UNIX System (1978)\n  \n <\/div><\/blockquote>\n<p>\n  Unix was the first production operating system to be ported\nbetween differing processor families (Version 6 Unix, 1976-77). Today,\nUnix is routinely ported to every new machine powerful enough to sport\na memory-management unit.  Unix applications are routinely moved\nbetween Unixes running on wildly differing hardware; in fact, it is\nunheard of for a port to fail.\n<\/p>\n<p>\n  Portability has always been one of Unix's principal advantages.\nUnix programmers tend to write on the assumption that hardware is\nevanescent and only the Unix API is stable, making as few assumptions\nas possible about machine specifics such as word length, endianness or\nmemory architecture.  In fact, code that is hardware-dependent in any\nway that goes beyond the abstract machine model of\nC\n  is considered\nbad form in Unix circles, and only really tolerated in very\nspecial cases like operating system kernels.\n<\/p>\n<p>\n  Unix programmers have learned that it is easy to be wrong when\nanticipating that a software project will have a short\nlifetime.<sup><a href=\"#ftn.id2998082\" id=\"id2998082\">[141]<\/a><\/sup>\n  Thus, they\ntend to avoid making software dependent on specific and perishable\ntechnologies, and to lean heavily on open standards.  These habits of\nwriting for portability are so ingrained in the Unix tradition that\nthey are applied even to small single-use projects that are thought of as\nthrowaway code. They have had secondary effects all through the design\nof the Unix development toolkit, and on programming languages like\nPerl and Python and Tcl that were developed under Unix.\n<\/p>\n<p>\n  The direct benefit of portability is that it is normal for Unix\nsoftware to outlive its original hardware platform, so tools and\napplications don't have to be reinvented every few years.  Today,\napplications originally written for Version 7 Unix (1979) are\nroutinely used not merely on Unixes genetically descended from V7, but\non variants like Linux\n  in which the operating system API was\nwritten from a Unix specification and shares no code with the Bell\nLabs source tree.\n<\/p>\n<p>\n  The indirect benefits are less obvious but may be more important.\nThe discipline of portability tends to exert a simplifying influence on\narchitectures, interfaces, and implementations.  This both increases\nthe odds of project success and reduces life-cycle maintenance\ncosts.\n<\/p>\n<p>\n  In this chapter, we'll survey the scope and history of Unix\nstandards.  We'll discuss which ones are still relevant today and\ndescribe the areas of greater and lesser variance in the Unix API.\nWe'll examine the tools and practices that Unix developers use to keep\ncode portable, and develop some guides to good practice.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2998082\" id=\"ftn.id2998082\">[141]<\/a><\/sup>\n  PDP-7 Unix and Linux were both examples of\nunexpected persistence.  Unix originated as a research toy hacked\ntogether by some researchers between projects, half to play with\nfile-system ideas and half to host a game.  The second was summed up by\nits creator as \u201cMy terminal emulator grew legs\u201d <sup><a href=\"#Torvalds\" title=\"[Torvalds]\">[Torvalds]<\/a><\/sup>.\n<\/p><\/div>\n\n<h3 id=\"c_evolution\"><a href=\"#c_evolution\">\u00a7<\/a>Evolution of C<\/h3>\n<p>\n  The central fact of the Unix programming experience has always\nbeen the stability of the C language and the handful of service\ninterfaces that always travel with it (notably, the standard I\/O\nlibrary and friends).  The fact that a language originated in 1973 has\nrequired as little change as this one has in thirty years of heavy use\nis truly remarkable, and without parallels anywhere else in computer\nscience or engineering.\n<\/p>\n<p>\n  In\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>, we\nargued that C has been successful because it acts as a layer of thin\nglue over computer hardware approximating the \u201cstandard\narchitecture\u201d of <sup><a href=\"#BlaauwBrooks\" title=\"[BlaauwBrooks]\">[BlaauwBrooks]<\/a><\/sup>.  There is, of course, more to the story than\nthat. To understand the rest of the story, we'll need to take a brief\nlook at the history of C.\n<\/p>\n<h4 id=\"id2998213\"><a href=\"#id2998213\">\u00a7<\/a>Early History of C<\/h4>\n<p>\n  C began life in 1971 as a systems-programming language for the\nPDP-11\n  port of\nUnix, based on Ken Thompson's\n  earlier B interpreter which had in turn been\nmodeled on BCPL, the Basic Common Programming Language designed at\nCambridge University in 1966-67.<sup><a href=\"#ftn.id2998245\" id=\"id2998245\">[142]<\/a><\/sup><\/p>\n<p>\n  Dennis Ritchie's\n  original C compiler (often called the\n\u201cDMR\u201d compiler after his initials) served the rapidly\ngrowing community around Unix versions 5, 6, and 7.  Version 6 C\nspawned Whitesmiths C, a reimplementation that became the first\ncommercial C compiler and the nucleus of IDRIS, the first Unix\nworkalike. But most modern C implementations are patterned on Steven\nC. Johnson's \u201cportable C compiler\u201d (PCC) which debuted in\nVersion 7 and replaced the DMR compiler entirely in both System\nV\n  and the BSD\n4.x\n  releases.\n<\/p>\n<p>\n  In 1976, Version 6 C introduced the\n  <code>typedef<\/code>,\n  <code>union<\/code>, and\n  <code>\n   unsigned\nint\n  <\/code>\n  declarations.  The approved syntax for variable\ninitializations and some compound operators also changed.\n<\/p>\n<p>\n  The original description of C was Brian Kernighan and Dennis\nM. Ritchie's\n  original\n  <em>The C Programming Language<\/em> aka\n\u201cthe White Book\u201d <sup><a href=\"#Kernighan-Ritchie\" title=\"[Kernighan-Ritchie]\">[Kernighan-Ritchie]<\/a><\/sup>.  It\nwas published in 1978, the same year the Whitemiths C compiler\nbecame available.\n<\/p>\n<p>\n  The White Book described enhanced Version 6 C, with one\nsignificant exception involving the handling of public storage.\nRitchie's original intention had been to model C's rules on FORTRAN\nCOMMON declarations, on the theory that any machine that could handle\nFORTRAN would be ready for C.  In the common-block model, a public\nvariable may be declared multiple times; identical declarations are\nmerged by the linker.  But two early C ports (to Honeywell and IBM 360\nmainframes) happened to be to machines with very limited common\nstorage or a primitive linker or both.  Thus, the Version 6 C compiler\nwas moved to the stricter definition-reference model (requiring at\nmost one definition of any given public variable and the\n  <code>extern<\/code>\n  keyword tagging references to it) described\nin <sup><a href=\"#Kernighan-Ritchie\" title=\"[Kernighan-Ritchie]\">[Kernighan-Ritchie]<\/a><\/sup>.\n<\/p>\n<p>\n  This decision was reversed in the C compiler that shipped with\nVersion 7 after it developed that a great deal of existing source\ndepended on the looser rules.  Pressure for backward-compatibility\nwould foil yet another attempt to switch (in 1983's System V\nRelease 1)\nbefore the ANSI Draft Standard finally settled on definition-reference\nrules in 1988.  Common-block public storage is still admitted as an\nacceptable variation by the standard.\n<\/p>\n<p>\n  V7 C introduced\n  <code>enum<\/code> and\ntreated\n  <code>struct<\/code> and\n  <code>union<\/code> values as first-class objects that\ncould be assigned, passed as arguments, and returned from functions\n(rather than being passed around by address).\n<\/p>\n <blockquote>\n  <p>\n   Another major change in V7 was that Unix data structure declarations\nwere now documented on header files, and included.  Previous Unixes\nhad actually printed the data structures (e.g., for directories) in\nthe manual, from which people would copy it into their code.  Needless\nto say, this was a major portability problem.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  The System III\n  C version of the PCC compiler (which also\nshipped with BSD 4.1c) changed the handling of struct\ndeclarations so that members with the same names in different structs\nwould not clash.  It also introduced\n  <code>void<\/code> and\n  <code>unsigned char<\/code> declarations.  The scope of\n  <code>extern<\/code> declarations local to a function was\nrestricted to the function, and no longer included all code following\nit.\n<\/p>\n<p>\n  The ANSI C Draft Proposed Standard added\n  <code>const<\/code> (for read-only storage) and\n  <code>volatile<\/code> (for locations such as memory-mapped\nI\/O registers that might be modified asynchronously from the thread of\nprogram control). The\n  <code>unsigned<\/code> type\nmodifier was generalized to apply to any type, and a symmetrical\n  <code>signed<\/code> was added.  Initialization\nsyntax for\n  <code>auto<\/code> array and structure\ninitializers and\n  <code>union<\/code> types was\nadded. Most importantly, function prototypes were added.\n<\/p>\n<p>\n  The most important changes in early C were the switch to\ndefinition-reference and the introduction of function prototypes in\nthe Draft Proposed ANSI C Standard. The language has been essentially\nstable since copies of the X3J11 committee's working papers on the\nDraft Proposed Standard signaled the committee's intentions to\ncompiler implementers in 1985-1986.\n<\/p>\n<p>\n  A more detailed history of early C, written by its designer, can\nbe found at <sup><a href=\"#Ritchie93\" title=\"[Ritchie93]\">[Ritchie93]<\/a><\/sup>.\n<\/p>\n<h4 id=\"id2994334\"><a href=\"#id2994334\">\u00a7<\/a>C Standards<\/h4>\n<p>\n  C standards development has been a conservative process with\ngreat care taken to preserve the spirit of the original C language,\nand an emphasis on ratifying experiments in existing compilers rather\nthan inventing new features. The C9X charter<sup><a href=\"#ftn.id2994347\" id=\"id2994347\">[143]<\/a><\/sup>\n  document is an excellent expression\nof this mission.\n<\/p>\n<p>\n  Work on the first official C standard began in 1983 under the\nauspices of the X3J11 ANSI committee. The major functional additions\nto the language were settled by the end of 1986, at which point it\nbecame common for programmers to distinguish between \u201cK&amp;R\nC\u201d and \u201cANSI C\u201d.\n<\/p>\n <blockquote>\n  <p>\n   Many people don't realize how\n   <em>unusual<\/em>\n   the C\nstandardization effort, especially the original ANSI C work, was in\nits insistence on standardizing only tested features.  Most language\nstandard committees spend much of their time inventing new features,\noften with little consideration of how they might be implemented.\nIndeed, the few ANSI C features that\n   <em>were<\/em>\n   invented from scratch \u2014 e.g., the notorious\n\u201ctrigraphs\u201d\u2014were the most disliked and least\nsuccessful features of C89.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n <blockquote>\n  <p>\n   Void pointers were invented as part of the standards effort, and have\nbeen a winner.  But Henry's point is still well taken.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  While the core of ANSI C was settled early, arguments over the\ncontents of the standard libraries dragged on for years. The formal\nstandard was not issued until the end of 1989, well after most\ncompilers had implemented the 1985 recommendations.  The standard was\noriginally known as ANSI X3.159, but was redesignated ISO\/IEC\n9899:1990 when the International Standards Organization (ISO) took\nover sponsorship in 1990.  The language variant it describes is\ngenerally known as C89 or C90.\n<\/p>\n<p>\n  The first book on C and Unix portability practice,\n  <em>Portable C and Unix Systems Programming<\/em> <sup><a href=\"#Lapin\" title=\"[Lapin]\">[Lapin]<\/a><\/sup>, was published in 1987 (I wrote it under a corporate\npseudonym forced on me by my employers at the time).  The Second\nEdition of <sup><a href=\"#Kernighan-Ritchie\" title=\"[Kernighan-Ritchie]\">[Kernighan-Ritchie]<\/a><\/sup> came out in 1988.\n<\/p>\n<p>\n  A very minor revision of C89, known as Amendment 1, AM1, or C93,\nwas floated in 1993.  It added more support for wide characters and\nUnicode. This became ISO\/IEC 9899-1:1994.\n<\/p>\n<p>\n  Revision of the C89 standard began in 1993.  In 1999, ISO\/IEC\n9899 (generally known as C99) was adopted by ISO.  It incorporated\nAmendment 1, and added a great many minor features.  Perhaps the most\nsignificant one for most programmers is the C++-like ability to\ndeclare variables at any point in a block, rather than just at the\nbeginning.  Macros with a variable number of arguments were also\nadded.\n<\/p>\n<p>\n  The C9X working group has a\n  <a href=\"http:\/\/anubis.dkuug.dk\/JTC1\/SC22\/WG14\/www\/projects\">\n   Web page<\/a>, but no third standards effort is planned as\nof mid-2003.  They are developing an addendum on C for \nembedded systems.\n<\/p>\n<p>\n  Standardization of C has been greatly aided by the fact that\nworking and largely compatible implementations were running on a wide\nvariety of systems before standards work was begun.  This made it\nharder to argue about what features should be in the standard.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2998245\" id=\"ftn.id2998245\">[142]<\/a><\/sup>\n  The \u2018C\u2019 in\nC therefore stands for Common \u2014 or, perhaps, for\n\u2018Christopher\u2019.  BCPL originally stood for \u201cBootstrap\nCPL\u201d\u2014a much simplified version of CPL, the very\ninteresting but overly ambitious and never implemented Common\nProgramming Language of Oxford and Cambridge, also known\naffectionately as \u201cChristopher's Programming Language\u201d\nafter its prime advocate, computer-science pioneer Christopher\nStrachey.\n<\/p><p><sup><a href=\"#id2994347\" id=\"ftn.id2994347\">[143]<\/a><\/sup><a href=\"http:\/\/anubis.dkuug.dk\/JTC1\/SC22\/WG14\/www\/charter\">\n   Available on\nthe Web<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch17s02\"><a href=\"#ch17s02\">\u00a7<\/a>Unix Standards<\/h3>\n<p>\n  The 1973 rewrite of Unix in C\n  made it unprecedentedly easy to port\nand modify.  As a result, the ancestral Unix diverged into a family of\noperating systems early on.  Unix standards originally developed to\nreconcile the APIs of the different branches of the family\ntree.\n<\/p>\n<p>\n  The Unix standards that evolved after 1985 were quite successful\nat this \u2014 so much so that they serve as valuable documentation\nof the API of modern Unix implementations.  In fact, real-world Unixes\nfollow published standards so closely that developers can (and\nfrequently do) lean more on documents like the POSIX specification\nthan on the official manual pages for the Unix variant they happen to\nbe using.\n<\/p>\n<p>\n  In fact, on the newer open-source Unixes (such as\nLinux), it is common for operating-system\nfeatures to have been engineered using published standards as the\nspecification.  We'll return to this point when we examine the RFC\nstandards process later in this chapter.\n<\/p>\n<h4 id=\"id2994594\"><a href=\"#id2994594\">\u00a7<\/a>Standards and the Unix Wars<\/h4>\n<p>\n  The original motivation for the development of Unix standards\nwas the split between the\nAT&amp;T\n  and\nBerkeley\n  lines of\ndevelopment that we examined in\n  <a href=\"#historychapter\" title=\"Chapter 2. History\">\n   Chapter 2<\/a>.\n<\/p>\n<p>\n  The 4.x BSD Unixes were descended from the 1979 Version 7.\nAfter the release of 4.1BSD in 1980 the BSD\n  line quickly developed a reputation as the\ncutting edge of Unix.  Important additions included the\n  <em>vi<\/em> visual editor, job control facilities for managing multiple foreground\nand background tasks from a single console, and improvements in\nsignals (see\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>). By far the most important addition\nwas to be TCP\/IP\n  networking, but though Berkeley got the contract to do it in 1980,\nTCP\/IP was not to ship in an external release for three years.\n<\/p>\n<p>\n  But another version, 1981's System III, became the basis of AT&amp;T's later\ndevelopment. System III reworked the Version 7 terminals interface\ninto a cleaner and more elegant form that was completely incompatible\nwith the Berkeley enhancements.  It retained the older (non-resetting)\nsemantics of signals (again, see\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>\n  for discussion of this point).  The\nJanuary 1983 release of System V Release 1\n  incorporated some\nBSD\n  utilities (such\nas\nvi(1)).\n<\/p>\n<p>\n  The first attempt to bridge the gap came in February 1983 from\nUniForum, an influential Unix user group.  Their Uniforum 1983 Draft\nStandard (UDS 83) described a \u201ccore Unix System\u201d consisting of a\nsubset of the System III\n  kernel and libraries plus a file-locking\nprimitive. AT&amp;T\n  declared support for UDS 83, but the standard was an inadequate subset\nof evolving practice based on 4.1BSD.  The problem was exacerbated by the July\n1983 release of 4.2BSD, which added many new features (including\nTCP\/IP\n  networking) and\nintroduced some subtle incompatibilities with the ancestral Version\n7.\n<\/p>\n<p>\n  The 1984 divestiture of the Bell operating companies and the\nbeginnings of the Unix wars\n  (see\n  <a href=\"#historychapter\" title=\"Chapter 2. History\">\n   Chapter 2<\/a>) significantly complicated matters.  Sun\nMicrosystems\n  was leading the workstation industry in a BSD direction; AT&amp;T was\ntrying to get into the computer business and use control of Unix as a\nstrategic weapon even as it continued to license the operating system\nto competitors like Sun.  All the vendors were making business\ndecisions to differentiate their versions of Unix for competitive\nadvantage.\n<\/p>\n<p>\n  During the Unix wars, technical standardization became something\nthat cooperating technical people pushed for and most product managers\naccepted grudgingly or actively resisted.  The one large and important\nexception was\nAT&amp;T,\nwhich declared its intention to cooperate with user groups in setting\nstandards when it announced System V Release\n2 (SVr2)\n  in January\n1984.  The second revision of the UniForum Draft Standard, in 1984,\nboth tracked and influenced the API of SVr2.  Later Unix standards\nalso tended to track System V except in areas where\nBSD\n  facilities were\nclearly functionally superior (thus, for example, modern Unix\nstandards describe the System V\n  terminal controls rather than the BSD\ninterface to the same facilities).\n<\/p>\n<p>\n  In 1985,\nAT&amp;T\n  released\nthe\n  <em>System V Interface Definition<\/em> (SVID).  SVID\nprovided a more formal description of the SVr2 API, incorporating UDS\n84.  Later revisions SVID2 and SVID3 tracked the interfaces of System V\nreleases 3 and 4.  SVID became the basis for the POSIX standards, which\nultimately tipped most of the Berkeley\/AT&amp;T disputes over system\nand C\n  library calls in\nAT&amp;T's favor.\n<\/p>\n<p>\n  But this would not become obvious for a few years yet;\nmeanwhile, the Unix wars\n  raged on.  For example, 1985 saw the\nrelease of two competing API standards for file system sharing over\nnetworks: Sun's\n  Network File System (NFS) and\nAT&amp;T's\n  Remote\nFile System (RFS).  Sun's NFS prevailed because Sun was willing to\nshare not merely specifications but open-source code with\nothers.\n<\/p>\n<p>\n  The lesson of this success should have been all the more pointed\nbecause on purely logical grounds RFS was the superior model. It\nsupported better file-locking semantics and better mapping among user\nidentities on different systems, and generally made an effort to get\nthe finer details of Unix file system semantics precisely right,\nunlike NFS. The lesson was ignored, however, even when it was repeated\nin 1987 by the open-source X windowing system's\n  victory over Sun's proprietary\nNetworked Window System (NeWS).\n<\/p>\n<p>\n  After 1985 the main thrust of Unix standardization passed to the\nInstitute of Electrical and Electronic Engineers (IEEE).  The IEEE's\n1003 committee developed a series of standards generally known as\nPOSIX.<sup><a href=\"#ftn.id2995003\" id=\"id2995003\">[144]<\/a><\/sup>\n  These went beyond describing\nmerely systems calls and C\n  library facilities; they specified\ndetailed semantics of a shell and a minimum command set, and also\ndetailed bindings for various non-C programming languages.  The first\nrelease in 1990 was followed by a second edition in 1996. The\nInternational Standards Organization adopted them as ISO\/IEC\n9945.\n<\/p>\n<p>\n  Key POSIX standards include the following:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     1003.1 (released 1990)\n    \n   <\/dt>\n   <dd>\n    <p>\n     Library procedures.  Described the C system call API, much like\nVersion 7 except for signals and the terminal-control interface.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     1003.2 (released 1992)\n    \n   <\/dt>\n   <dd>\n    <p>\n     Standard shell and utilities. Shell semantics strongly resemble those \nof the System V Bourne shell.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     1003.4 (released 1993)\n    \n   <\/dt>\n   <dd>\n    <p>\n     Real-time Unix.  Binary semaphores, process memory locking,\nmemory-mapped files, shared memory, priority scheduling, real-time\nsignals, clocks and timers, IPC message passing, synchronized I\/O,\nasynchronous I\/O, real-time files.\n    <\/p>\n    <p>\n     In the 1996 Second Edition, 1003.4 was split into 1003.1b\n(real-time) and 1003.1c (threads).\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  Despite being underspecified in a couple of key areas such as\nsignal-handling semantics and omitting BSD\n  sockets, the original POSIX standards\nbecame the basis of all later Unix standardization work.  They are\nstill cited as an authority, albeit indirectly through references like\n  <em>POSIX Programmer's Guide<\/em> <sup><a href=\"#Lewine\" title=\"[Lewine]\">[Lewine]<\/a><\/sup>.  The de facto Unix API standard is still\n\u201cPOSIX plus sockets\u201d, with later standards mainly adding\nfeatures and specifying conformance in unusual edge cases more\nclosely.\n<\/p>\n<p>\n  The next player on the scene was X\/Open (later renamed the Open\nGroup), a consortium of Unix vendors formed in 1984.  Their X\/Open\nPortability Guides (XPGs) initially developed in parallel with the\nPOSIX drafts, then after 1990 the XPGs incorporated and extended\nPOSIX.  Unlike POSIX, which attempted to capture a safe subset of all\nUnixes, the XPGs were oriented more toward common practice at the\nleading edge; even XPG1 in 1985, spanning SVr2 and 4.2BSD,\n  included\nsockets.\n<\/p>\n<p>\n  XPG2 in 1987 added a terminal-handling API that was essentially\nSystem V\ncurses(3).\nXPG3 in 1990 merged in the X11 API.  XPG4 in 1992 mandated full\ncompliance with the 1989 ANSI C standard.  XPG2, 3, and 4 were heavily concerned\nwith support of internationalization and described an elaborate API\nfor handling codesets and message catalogs.\n<\/p>\n<p>\n  In reading about Unix standards you might come across references\nto \u201cSpec 1170\u201d (from 1993), \u201cUnix 95\u201d (from\n1995) and \u201cUnix 98\u201d (from 1998).  These were\ncertification marks based on the X\/Open standards; they are now of\nhistorical interest only.  But the work done on XPG4 turned into \nSpec 1170, which turned into the first version of the Single Unix\nSpecification (SUS).\n<\/p>\n<p>\n  In 1993 seventy-five systems and software vendors including\nevery major Unix company put a final end to the Unix\nwars\n  when they\ndeclared backing for X\/Open to develop a common definition of Unix.\nAs part of the arrangement, X\/Open acquired the rights to the Unix\ntrademark.  The merged standard became Single Unix Standard version 1.\nIt was followed in 1997 by a version 2.  In 1999 X\/Open absorbed the\nPOSIX activity.\n<\/p>\n<p>\n  In 2001, X\/Open (now The Open Group) issued the\n  <a href=\"http:\/\/www.unix.org\/version3\/\">\n   Single Unix Standard version 3<\/a>.  All the threads of Unix API standardization were finally\ngathered into one bundle.  This reflected facts on the ground; the\ndifferent varieties of Unix had re-converged on a common API.  And, at\nleast among old-timers who remembered the turbulence of the 1980s,\nthere was much rejoicing.\n<\/p>\n<h4 id=\"id2999310\"><a href=\"#id2999310\">\u00a7<\/a>The Ghost at the Victory Banquet<\/h4>\n<p>\n  There was, unfortunately, an awkward detail \u2014 the\nold-school Unix vendors who had backed the effort were under severe\npressure from the new school of open-source Unixes, and were in some\ncases in the process of abandoning (in favor of\nLinux)\nthe proprietary Unixes for which they had gone to so much effort to\nsecure conformance.\n<\/p>\n<p>\n  The conformance testing needed to verify Single Unix\nSpecification conformance is an expensive proposition.  It would need\nto be done on a per-distribution basis, but is well out of the reach\nof most distributors of open-source operating systems.  In any case,\nLinux changes so fast that any given release of a distribution would\nprobably be obsolete by the time it could get\ncertified.<sup><a href=\"#ftn.id2999345\" id=\"id2999345\">[145]<\/a><\/sup><\/p>\n<p>\n  Standards like the Single Unix Specification have not entirely lost their\nrelevance.  They're still valuable guides for Unix implementers.  But\nhow The Open Group and other institutions of the old-school Unix\nstandardization process will adapt to the rapid tempo of open-source\nreleases (and to the low- or zero-budget operation of open-source\ndevelopment groups!) remains to be seen.\n<\/p>\n<h4 id=\"id2999366\"><a href=\"#id2999366\">\u00a7<\/a>Unix Standards in the Open-Source World<\/h4>\n<p>\n  In the mid-1990s, the open-source community began\nstandardization efforts of its own.  These efforts built on the\nsource-code-level compatibility secured by\nPOSIX\n  and its\ndescendants.  Linux, in particular, had been written from\nscratch in a way that depended on the availability of Unix API\nstandards like POSIX.<sup><a href=\"#ftn.id2999397\" id=\"id2999397\">[146]<\/a><\/sup><\/p>\n<p>\n  In 1998 Oracle ported its market-leading database product to\nLinux, in a move that was rightly seen as a major breakthrough in\nLinux's mainstream acceptance.  The engineer in charge of the port\nprovided a definitive demonstration that API standards had done their\njob when he was asked by a reporter what technical challenges Oracle\nhad had to surmount.  The engineer's reply was \u201cWe typed\n\u2018make\u2019.\u201d\n<\/p>\n<p>\n  The problem for the new-school Unixes, therefore, was not API\ncompatibility at the source-code level.  Everybody took for granted\nthe ability to move source code between different\nLinux,\nBSD, and\nproprietary-Unix distributions without more than a trivial amount of\nporting labor.  The new problem was not source compatibility but\nbinary compatibility.  For the ground under Unix had shifted in a\nsubtle way as a consequence of the triumph of commodity PC\nhardware.\n<\/p>\n<p>\n  In the old days, each Unix had run on what was effectively its\nown hardware platform.  There was enough variety in processor\ninstruction sets and machine architectures that applications had to be\nported at source level to move at all.  On the other hand, there were\na relatively few major Unix releases, each with relatively long\nservice lifetimes.  Application vendors like Oracle could afford the\ncost of building and shipping separate binary distributions for each\nof three or four hardware\/software combinations, because they could\namortize the low cost of source-code porting over large customer\npopulations and a long enough product life cycle.\n<\/p>\n<p>\n  But then the minicomputer and workstation vendors were swamped by\ninexpensive 386-based supermicros, and open-source Unixes changed the \nrules.  Vendors found they no longer had a stable platform to ship\ntheir binaries to.\n<\/p>\n<p>\n  The superficial problem, at first, was the large number of Unix\ndistributors \u2014 but as the Linux\n  distribution market\nconsolidated, it became clear that the real issue was the rate of\nchange over time.  APIs were stable, but the expected locations of\nsystem administrative files, utility programs, and things like the\nprefix of the paths to user mailbox names and system log files kept\nchanging.\n<\/p>\n<p>\n  The first standards effort to develop within the new-school\nLinux\n  and BSD\n  community\nitself (beginning in 1993) was the Filesystem Hierarchy Standard (FHS).\nThis was incorporated into the Linux Standards Base (LSB), which also\nstandardized an expected set of service libraries and helper\napplications.  Both standards became activities of the\n  <a href=\"http:\/\/www.freestandards.org\/\">\n   Free Standards Group<\/a>,\nwhich by 2001 developed a role similar to X\/Open's position amidst the\nold-school Unix vendors.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2995003\" id=\"ftn.id2995003\">[144]<\/a><\/sup>\n  The\noriginal 1986 trial-use standard was called IEEE-IX.  The name\n\u2018POSIX\u2019 was suggested by Richard Stallman. The\nintroduction to POSIX.1 says: \u201cIt is expected to be pronounced\npahz-icks as in positive, not poh-six, or other variations. The\npronounciation has been published in an attempt to promulgate a\nstandardized way of referring to a standard operating system\ninterface\u201d.\n<\/p><p><sup><a href=\"#id2999345\" id=\"ftn.id2999345\">[145]<\/a><\/sup>\n  One Linux distributor, Lasermoon in Great\nBritain, did achieve POSIX.1 FIPS 151-2 certification \u2014 and went\nout of business, because potential customers didn't\ncare.\n<\/p><p><sup><a href=\"#id2999397\" id=\"ftn.id2999397\">[146]<\/a><\/sup>\n  See\n  <em>Just for Fun<\/em><sup><a href=\"#Torvalds\" title=\"[Torvalds]\">[Torvalds]<\/a><\/sup> for\ndiscussion.\n<\/p><\/div>\n\n<h3 id=\"ietf_process\"><a href=\"#ietf_process\">\u00a7<\/a>IETF and the RFC Standards Process<\/h3>\n<p>\n  When the Unix community merged with the culture of Internet\nengineers, it also inherited a mindset formed by the RFC standards\nprocess of the Internet Engineering Task Force (IETF).  In IETF tradition,\nstandards have to arise from experience with a working prototype\nimplementation \u2014 but once they\n  <em>become<\/em>\n  standards, code that does not conform to them is considered broken and\nmercilessly scrapped.\n<\/p>\n<p>\n  This is not, sadly, the way standards are normally developed.  The\nhistory of computing is full of instances in which technical standards\nwere derived by a process that combined the worst features of\nphilosophical axe-grinding with murky back-room politics \u2014\nproducing specifications that failed to resemble anything ever\nimplemented. Worse, many were either so demanding that they could\nnot be practically implemented or so underspecified that they caused\nmore confusion than they resolved.  Then they were promulgated\nto vendors who ignored them wherever they were inconvenient.\n<\/p>\n<p>\n  One of the more notorious examples of standards nonsense was the\nOpen Systems Interconnect networking protocols that briefly contended\nwith TCP\/IP\n  in the\n1980s \u2014 its 7-layer model looked elegant from a distance but\nproved overcomplicated and unimplementable in\npractice.<sup><a href=\"#ftn.id2999607\" id=\"id2999607\">[147]<\/a><\/sup>\n  The\nANSI X3.64 standard for video-display terminal capabilities is another\nfamous horror story; bedeviled by subtle incompatibilities between\nlegally conformant implementations.  Even after character-cell\nterminals have been largely displaced by bitmapped displays these\ncontinue to cause problems (in particular, this is why the function\nand special keys in your\nxterm(1)\nwill occasionally break).  The RS232 standard for serial\ncommunications was so underspecified that it sometimes seemed that no\ntwo serial cables were alike.  Standards horror stories of similar\nkind could fill a book the size of this one.\n<\/p>\n<p>\n  The IETF's philosophy has been famously summarized as \u201cWe\nreject kings, presidents, and voting.  We believe in rough consensus\nand running code\u201d.<sup><a href=\"#ftn.id2999646\" id=\"id2999646\">[148]<\/a><\/sup>\n  That demand for a working implementation\nfirst has saved it from the worst category of blunders.  In fact its\ncriterion is stronger:\n<\/p>\n <blockquote>\n  <p>\n   [A] candidate specification must be implemented\nand tested for correct operation and interoperability by multiple\nindependent parties and utilized in increasingly demanding\nenvironments, before it can be adopted as an Internet\nStandard.\n  <\/p>\n <\/blockquote>\n<p>\n  All IETF standards pass through a stage as RFCs (Requests for\nComment).  The submission process for RFCs is deliberately informal.\nRFCs may propose standards, survey results, suggest philosophical\nbases for subsequent RFCs, or even make jokes.  The appearance of the\nannual April 1st RFC is the closest equivalent of a high holy day\nobservance among Internet\nhackers, and has\nproduced such gems as\n  <em>A Standard for the Transmission of IP Datagrams on Avian Carriers (RFC 1149)<\/em><sup><a href=\"#ftn.id2999703\" id=\"id2999703\">[149]<\/a><\/sup>\n  the\n  <em>The Hyper Text Coffee Pot Control Protocol (RFC 2324)<\/em>,<sup><a href=\"#ftn.id2999728\" id=\"id2999728\">[150]<\/a><\/sup>\n  and\n  <em>The Security Flag in the IPv4 Header (RFC 3514)<\/em>.<sup><a href=\"#ftn.id2999745\" id=\"id2999745\">[151]<\/a><\/sup><\/p>\n<p>\n  But joke RFCs are about the\n  <em>only<\/em>\n  sort of\nsubmission that instantly becomes an RFC.  Serious proposals actually\nstart as \u201cInternet-Drafts\u201d floated for public comment via\nIETF directories on several well-known hosts.  Individual Internet-Drafts\nhave no formal status and can be changed or dropped by their\noriginators at any time.  If they are neither withdrawn nor promoted to\nRFC status, they are removed after six months.\n<\/p>\n<p>\n  Internet-Drafts are not specifications, and software\nimplementers and vendors are specifically barred from claiming\ncompliance with them as if they were specifications. Internet-Drafts\nare focal points for discussion, usually in a working group connected\nthrough an electronic mailing list.  When the working group leadership\ndeems fit, the Internet-Draft is submitted to the RFC editor for\nassignment of an RFC number.\n<\/p>\n<p>\n  Once an Internet-Draft has been published with an RFC number,\nit is a specification to which implementers may claim conformance. It\nis expected that the authors of the RFC and the community at large \nwill begin correcting the specification with field experience.\n<\/p>\n<p>\n  Some RFCs go no further.  A specification that fails to attract\nuse and survive field testing can be quietly forgotten, and eventually\nmarked \u201cNot recommended\u201d or \u201cSuperseded\u201d by\nthe RFC editor.  Failed proposals are accepted as one of the overheads\nof the process, and no stigma is attached to being associated\nwith one.\n<\/p>\n<p>\n  The steering committee of the IETF (IESG, or Internet\nEngineering Steering Group) is responsible for putting successful RFCs\non the standards track.  They do this by designating the RFC a\n\u2018Proposed Standard\u2019. For the RFC to qualify, the\nspecification must be stable, peer-reviewed, and have attracted\nsignificant interest from the Internet community.  Implementation\nexperience is not absolutely required before an RFC is given Proposed\nStandard designation, but it is considered highly desirable, and the IESG\nmay require it if the RFC touches the Internet core protocols or\nmight be otherwise destabilizing.\n<\/p>\n<p>\n  Proposed Standards are still subject to revision, and may even\nbe withdrawn if the IESG and IETF identify a better solution. They\nare not recommended for use in \u201cdisruption-sensitive\nenvironments\u201d\u2014don't put them in your\nair-traffic-control systems or on intensive-care equipment.\n<\/p>\n<p>\n  When there are at least two working, complete, independently\noriginated, and interoperable implementations of a Proposed Standard,\nthe IESG may elevate it to Draft Standard status.  RFC 2026 says: \n\u201cElevation to Draft Standard is a major advance in status, indicating\na strong belief that the specification is mature and will be\nuseful\u201d.\n<\/p>\n<p>\n  Once an RFC has reached Draft Standard status, it will be\nchanged only to address bugs in the logic of the specification. Draft\nStandards are expected to be ready for deployment in\ndisruption-sensitive environments.\n<\/p>\n<p>\n  When a Draft Standard has passed the test of widespread\nimplementation and reached general acceptance, it may be blessed as an\nInternet Standard.  Internet Standards keep their RFC numbers, but also\nget a number in the STD series.  At time of writing there are over\n3000 RFCs but only 60 STDs.\n<\/p>\n<p>\n  RFCs not on standards track may be labeled Experimental,\nInformational (the joke RFCs get this label), or Historic.  The\nHistoric label is applied to obsolete standards. RFC 2026 notes:\n\u201c(Purists have suggested that the word should be\n\u2018Historical\u2019; however, at this point, the use of\n\u2018Historic\u2019 is historical.)\u201d\n<\/p>\n<p>\n  The IETF standards process is designed to encourage\nstandardization driven by practice rather than theory, and to \nensure that standard protocols have undergone rigorous peer\nreview and testing.  The success of this model is evident in its \nresults \u2014 the worldwide Internet.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id2999607\" id=\"ftn.id2999607\">[147]<\/a><\/sup>\n  A Web search is likely to turn up a popular\npage comparing the OSI 7-layer model with the Taco Bell 7-layer\nburrito \u2014 unfavorably to the former.\n<\/p><p><sup><a href=\"#id2999646\" id=\"ftn.id2999646\">[148]<\/a><\/sup>\n  This line was first uttered\nby senior IETF cadre Dave Clark at the tumultuous 1992 meeting during\nwhich the IETF rejected the Open Systems Interconnect\nprotocol.\n<\/p><p><sup><a href=\"#id2999703\" id=\"ftn.id2999703\">[149]<\/a><\/sup><a href=\"http:\/\/www.ietf.org\/rfc\/rfc1149.txt\">\n   RFC\n1149 is available on the Web<\/a>. Not only that, it has been\n  <a href=\"http:\/\/www.blug.linux.no\/rfc1149\/writeup.html\">\n   implemented<\/a>.\n<\/p><p><sup><a href=\"#id2999728\" id=\"ftn.id2999728\">[150]<\/a><\/sup><a href=\"http:\/\/www.ietf.org\/rfc\/rfc2324.txt\">\n   RFC 2324 is available on the\nWeb<\/a>.\n<\/p><p><sup><a href=\"#id2999745\" id=\"ftn.id2999745\">[151]<\/a><\/sup><a href=\"http:\/\/www.ietf.org\/rfc\/rfc3514.txt\">\n   RFC 3514 is available on the\nWeb<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch17s04\"><a href=\"#ch17s04\">\u00a7<\/a>Specifications as DNA, Code as RNA<\/h3>\n<p>\n  Even in the paleolithic period of the PDP-7, Unix programmers\nhad always been more prone than their counterparts elsewhere to treat\nold code as disposable.  This was doubtless a product of the Unix\ntradition's emphasis on modularity, which makes it easier to discard\nand replace small pieces of systems without losing everything.  Unix\nprogrammers have learned by experience that trying to salvage bad code\nor a bad design is often more work than rebooting the project. Where\nin other programming cultures the instinct would be to patch the\nmonster monolith because you have so much work invested in it, the\nUnix instinct is usually to scrap and rebuild.\n<\/p>\n<p>\n  The IETF\n  tradition\nreinforced this by teaching us to think of code as secondary to\nstandards. Standards are what enable programs to cooperate; they knit\nour technologies into wholes that are more than the sum of the parts.\nThe IETF showed us that careful standardization, aimed at capturing\nthe best of existing practice, is a powerful form of humility that\nachieves more than grandiose attempts to remake the world around a\nnever-implemented ideal.\n<\/p>\n<p>\n  After 1980, the impact of that lesson was increasingly widely\nfelt in the Unix community.  Thus, while the ANSI\/ISO C standard from\n1989 is not completely without flaws, it is exceptionally clean and\npractical for a standard of its size and importance.  The Single Unix\nSpecification contains fossils from three decades of experimentation\nand false starts in a more complicated domain, and is therefore\nmessier than ANSI C.  But the component standards it was composed from\nare quite good; strong evidence for this is the fact that Linus\nTorvalds successfully built a Unix from scratch by reading them. The\nIETF's\n  quiet but\npowerful example created one of the critical pieces of context that\nmade Linus Torvalds's feat possible.\n<\/p>\n<p>\n  Respect for published standards and the\nIETF\n  process has become\ndeeply ingrained in the Unix culture; deliberately violating Internet\nSTDs is simply Not Done.  This can sometimes create chasms of mutual\nincomprehension between people with a Unix background and others prone\nto assume that the most popular or widely deployed implementation of a\nprotocol is by definition correct \u2014 even if it breaks the\nstandard so severely that it will not interoperate with properly\nconforming software.\n<\/p>\n<p>\n  The Unix programmer's respect for published standards is more\ninteresting because he is likely to be rather hostile to a-priori\nspecifications of other kinds.  By the time the \u2018waterfall\nmodel\u2019\n  (specify exhaustively first, then implement, then debug, with no\nreverse motion at any stage) fell out of favor in the\nsoftware-engineering literature, it had been an object of derision\namong Unix programmers for years.  Experience, and a strong tradition\nof collaborative development, had already taught them that prototyping\nand repeated cycles of test and evolution are a better\nway.\n<\/p>\n<p>\n  The Unix tradition clearly recognizes that there can be great\nvalue in good specifications, but it demands that they be treated as\nprovisional and subject to revision through field experience in the\nway that Internet-Drafts and Proposed Standards are.  In best Unix\npractice, the documentation of the program is used as a specification\nsubject to revision analogously to an Internet Proposed\nStandard.\n<\/p>\n <blockquote>\n  <p>\n   Unlike other environments, in Unix development the documentation is often\nwritten before the program, or at least in conjunction with it.  For X11, the\ncore X standards were finished before the first release of X and have remained\nessentially unchanged since that date. Compatibility among different X systems\nis improved further by rigorous specification-driven testing.\n  <\/p>\n  <p>\n   The existence of a well-written specification made the\ndevelopment of the X test suite much easier.  Each statement in the X\nspecification was translated into code to test the implementation, a\nfew minor inconsistencies were uncovered in the specification during\nthis process, but the result is a test suite that covers a significant\nfraction of the code paths within the sample X library and server, and\nall without referring to the source code of that implementation.\n  <\/p><div>\n  \n   \u2013 Keith Packard\n  \n <\/div><\/blockquote>\n<p>\n  Semiautomation of the test-suite generation proved to be a\nmajor advantage.  While field experience and advances in the state\nof the graphics art led many to criticize X on design grounds, and\nvarious portions of it (such as the security and user-resource models) came\nto seem clumsy and over-engineered, the X implementation achieved a\nremarkable level of stability and cross-vendor interoperation.\n<\/p>\n<p>\n  In\n  <a href=\"#generationchapter\" title=\"Chapter 9. Generation\">\n   Chapter 9<\/a>\n  we discussed the value of pushing coding up to the highest possible\nlevel to minimize the effects of constant defect density.  Implicit in\nKeith Packard's account is the idea that the X documentation\nconstituted no mere wish-list but a form of high-level code.  Another\nkey X developer confirms this:\n<\/p>\n <blockquote>\n  <p>\n   In X, the specification has always ruled.  Sometimes specs have bugs that\nneed to be fixed too, but code is usually buggier than specs (for any\nspec worth its ink, anyway).\n  <\/p><div>\n  \n   \u2013 Jim Gettys\n  \n <\/div><\/blockquote>\n<p>\n  Jim goes on to observe that X's process is actually quite\nsimilar to the IETF's.\nNor is its utility limited to constructing good test suites; it means\nthat arguments about the system's behavior can be conducted at a\nfunctional level with respect to the specification, avoiding too much\nentanglement in implementation issues.\n<\/p>\n <blockquote>\n  <p>\n   Having a well-considered specification driving development allows for\nlittle argument about bug vs. feature; a system which incorrectly implements the\nspecification is broken and should be fixed.\n  <\/p>\n  <p>\n   I suspect this is so ingrained into most of us that we lose\nsight of its power.\n  <\/p>\n  <p>\n   A friend of mine who worked for a small software firm east of\nBellevue wondered how Linux applications developers could get OS\nchanges synchronized with application releases.  In that company,\nmajor system-level APIs change frequently to accommodate application\nwhims and so essential OS functionality must often be released along\nwith each application.\n  <\/p>\n  <p>\n   I described the power held by the specifications and how the\nimplementation was subservient to them, and then went on to assert\nthat an application which got an unexpected result from a documented\ninterface was either broken or had discovered a bug.  He found this\nconcept startling.\n  <\/p>\n  <p>\n   Discerning such bugs is a simple matter of verifying the\nimplementation of the interface against the specification.  Of course,\nhaving source for the implementation makes that a bit easier.\n  <\/p><div>\n  \n   \u2013 Keith Packard\n  \n <\/div><\/blockquote>\n<p>\n  This standards-come-first attitude has benefits for end users as\nwell.  While that no-longer-small company east of Bellevue has trouble\nkeeping its office suite compatible with its own previous releases,\nGUI applications written for X11 in 1988 still run without change on\ntoday's X implementations.  In the Unix world, this sort of longevity\nis normal \u2014 and the standards-as-DNA attitude is the reason\nwhy.\n<\/p>\n<p>\n  Thus, experience shows that the standards-respecting,\nscrap-and-rebuild culture of Unix tends to yield better\ninteroperability over extended time than perpetual patching of a code\nbase without a standard to provide guidance and continuity.  This\nmay, indeed, be one of the most important Unix lessons.\n<\/p>\n<p>\n  Keith's last comment brings us directly to an issue that the\nsuccess of open-source Unixes has brought to the forefront \u2014 the\nrelationship between open standards and open source.   We'll address\nthis at the end of the chapter \u2014 but before doing that, it's\ntime to address the practical question of how Unix programmers \ncan actually\n  <em>use<\/em>\n  the tremendous body of\naccumulated standards and lore to achieve software portability.\n<\/p>\n\n<h3 id=\"ch17s05\"><a href=\"#ch17s05\">\u00a7<\/a>Programming for Portability<\/h3>\n<p>\n  Software portability is usually thought of in quasi-spatial\nterms: can this code be moved sideways to existing hardware and\nsoftware platforms other than the one it was built for?  But Unix\nexperience over decades tells us that durability down through time is\njust as important, if not more so.  If we could predict the future of\nsoftware in detail it would probably be the present \u2014\nnevertheless, in programming for portability we should try to think\nabout making choices that will base the software on the features of\nits environment that are likeliest to persist, and avoid technologies\nthat seem likely to face end-of-life in the foreseeable future.\n<\/p>\n<p>\n  Under Unix, two decades of attention to the issues of\nspecifying portable APIs has largely solved that problem.  Facilities\ndescribed in the Single Unix Specification are likely to be present\non all modern Unix platforms today and rather unlikely to go \nunsupported in the future.\n<\/p>\n<p>\n  But not all platform dependencies have to do with the system\nor library APIs.  Your implementation language can matter; file-system\nlayout and configuration differences between the source and target\nsystem can be a problem as well.  But Unix practice has evolved ways \nto cope.\n<\/p>\n<h4 id=\"id3000303\"><a href=\"#id3000303\">\u00a7<\/a>Portability and Choice of Language<\/h4>\n<p>\n  The first issue in programming for portability is your choice of\nimplementation language.  All the major languages we surveyed in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  are highly portable\nin the sense that compatible implementations are available across all\nmodern Unixes; for most, implementations under Windows and MacOS are\navailable as well.  Portability problems tend to arise not in the core\nlanguages but in support libraries and degree of integration with the\nlocal environment (especially IPC and concurrent-process management,\nincluding the infrastructure for GUIs).\n<\/p>\n<h5 id=\"c_portability\"><a href=\"#c_portability\">\u00a7<\/a>C Portability<\/h5>\n<p>\n  The core C language is extremely portable.  The standard Unix\nimplementation is the GNU C compiler, which is ubiquitous not only in\nopen-source Unixes but modern proprietary Unixes as well.  GNU C has\nbeen ported to Windows and classic MacOS, but is not widely used in\neither environment because it lacks portable bindings to the native\nGUI.\n<\/p>\n<p>\n  The standard I\/O library, mathematics routines, and\ninternationalization support are portable across all C\nimplementations.  File I\/O, signals, and process control are portable\nacross Unixes provided one takes care to use only the modern APIs\ndescribed in the Single Unix Specification; older C code often has\nthickets of preprocessor conditionals for portability, but those\nhandle legacy pre-POSIX interfaces from older proprietary Unixes that \nare obsolete or close to it in 2003.\n<\/p>\n<p>\n  C portability starts to be a more serious problem near IPC,\nthreads, and GUI interfaces.  We discussed IPC and threads portability\nissues in\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>. The real practical problem is GUI\ntoolkits.  A number of open-source GUI toolkits are universally\nportable across modern Unixes and to Windows and classic MacOS as well\n\u2014 Tk, wxWindows, GTK, and Qt are four well-known ones with source\ncode and documentation readily discoverable by Web search.  But none\nof them is shipped with all platforms, and (for reasons more legal\nthan technical) none of these offers the native-GUI look and feel on\nall platforms.  We gave some guidelines for coping in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>.\n<\/p>\n<p>\n  Volumes have been written on the subject of how to write \nportable C code.  This book is not going to be one of them.   Instead,\nwe recommend a careful reading of\n  <em>Recommended C Style and Coding Standards<\/em><sup><a href=\"#Cannon\" title=\"[Cannon]\">[Cannon]<\/a><\/sup>\nand the chapter on portability in\n  <em>The Practice of Programming<\/em> <sup><a href=\"#Kernighan-Pike99\" title=\"[Kernighan-Pike99]\">[Kernighan-Pike99]<\/a><\/sup>.\n<\/p>\n<h5 id=\"cc_portability\"><a href=\"#cc_portability\">\u00a7<\/a>C++ Portability<\/h5>\n<p>\n  C++ has all the same operating-system-level portability issues as C,\nand some of its own.\nAn additional one is that the open-source GNU compiler for C++ has\nlagged substantially behind the proprietary implementations for most\nof its existence; thus, there is not yet as of mid-2003 a\nuniversally deployed equivalent of GNU C on which to base a de-facto\nstandard.  Furthermore, no C++ compiler yet implements the full C++99\nISO standard for the language, though GNU C++ comes closest.\n<\/p>\n<h5 id=\"shell_portability\"><a href=\"#shell_portability\">\u00a7<\/a>Shell Portability<\/h5>\n<p>\n  Shell-script portability is, unfortunately, poor.  The problem\nis not shell itself;\nbash(1)\n(the open-source Bourne Again shell) has become sufficiently\nubiquitous that pure shellscripts can run almost anywhere.  The\nproblem is that most shellscripts make heavy use of other commands and\nfilters that are much less portable, and by no means guaranteed to be\nin the toolkit in any given target machine.\n<\/p>\n<p>\n  This problem can be overcome by dint of heroic effort, as in the\nautoconf(1)\ntools.  But it is sufficiently severe that most of the heavier sort of\nprogramming that used to be done in shell has moved to\nsecond-generation scripting languages\n  like Perl, Python, and Tcl.\n<\/p>\n<h5 id=\"perl_portability\"><a href=\"#perl_portability\">\u00a7<\/a>Perl Portability<\/h5>\n<p>\n  Perl has good portability. Stock Perl even offers a portable set\nof bindings to the Tk toolkit that supports portable GUIs across\nUnix, MacOS and Windows. One issue dogs it, however.  Perl scripts often\nrequire add-on libraries from CPAN (the Comprehensive Perl Archive\nNetwork) which are not guaranteed to be present with every Perl\nimplementation.\n<\/p>\n<h5 id=\"python_portability\"><a href=\"#python_portability\">\u00a7<\/a>Python Portability<\/h5>\n<p>\n  Python has excellent portability.  Like Perl, stock Python even\noffers a portable set of bindings to the Tk toolkit that supports\nportable GUIs across Unix, MacOS, and Windows.\n<\/p>\n<p>\n  Stock Python has a much richer standard library than does\nPerl\n  and no\nequivalent of CPAN for programmers to rely on; instead, important\nextension modules are routinely incorporated into the stock Python\ndistribution during minor releases. This trades a spatial problem for\na temporal one, making Python much less subject to the missing-module\neffect at the cost of making Python minor version numbers somewhat\nmore important than Perl release levels are.  In practice, the\ntradeoff seems to favor Python.\n<\/p>\n<h5 id=\"tcl_portability\"><a href=\"#tcl_portability\">\u00a7<\/a>Tcl Portability<\/h5>\n<p>\n  Tcl portability is good, overall, but varies sharply by project\ncomplexity.  The Tk toolkit for cross-platform GUI programming is\nnative to Tcl.  As with Python, evolution of the core language has\nbeen relatively smooth, with few version-skew problems.\nUnfortunately, Tcl relies even more heavily than Perl on extension\nfacilities that are not guaranteed to ship with every implementation\n\u2014 and there is no equivalent of CPAN to centrally distribute\nthem.\n<\/p>\n<p>\n  For smaller projects not reliant on extensions, therefore, Tcl\nportability is excellent.  But larger projects tend to depend heavily\non both extensions and (as with shell programming) calling external\ncommands that may or may not be present on the target machine; their\nportability tends to be poor.\n<\/p>\n<p>\n  Tcl may have suffered, ironically, from\nthe ease of adding extensions to it.  By the time a particular extension\nstarted to look interesting as part of the standard distribution, there\ntypically were several different versions of it in existence.  At the\n1995 Tcl\/Tk Workshop, John Ousterhout explained why there was no OO\nsupport in the standard Tcl distribution:\n<\/p>\n <blockquote>\n  <p>\n   Think of five mullahs sitting around in a circle, all saying\n\u201cKill him, he's a heathen\u201d.  If I put a specific OO\nscheme into the core, then one of them will say \u201cBless you, my\nson, you may kiss my ring\u201d, and the other four will say\n\u201cKill him, he's a heathen\u201d.\n  <\/p>\n <\/blockquote>\n<p>\n  The lot of a language designer is not necessarily a happy one.\n<\/p>\n<h5 id=\"java_portability\"><a href=\"#java_portability\">\u00a7<\/a>Java Portability<\/h5>\n<p>\n  Java portability is excellent \u2014 it was, after all,\ndesigned with \u201cwrite once, run everywhere\u201d as a primary\ngoal.  Portability fails, however, to be perfect.  The difficulties are\nmostly version-skew problems between JDK 1.1 and the older AWT GUI\ntoolkit (on the one hand) and JDK 1.2 with the newer Swing GUI toolkit.\nThere are several important reasons for these:\n<\/p>\n <div><ul><li>\n    <p>\n     Sun's AWT design was so deficient that it had to be replaced \nwith Swing.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Microsoft's\n     refusal to support Java development on Windows and attempt to replace\nit with C#.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Microsoft's decision to hold Internet Explorer's applet support \nat the JDK 1.1 level.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Sun\n     licensing terms that make open-source implementations of JDK 1.2\nimpossible, retarding its deployment (especially in the Linux\nworld).\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  For programs that involve GUIs, Java developers seeking\nportability will, for the foreseeable future, face a choice: Stay in\nJDK 1.1\/AWT with a poorly designed toolkit for maximum portability\n(including to Microsoft Windows), or get the better toolkit and\ncapabilities of JDK 1.2 at the cost of sacrificing some\nportability.\n<\/p>\n<p>\n  Finally, as we noted previously, the Java thread support has\nportability problems.  The Java API, unlike less ambitious\noperating-system bindings for other languages, bravely tried to bridge\nthe gaps between the diverging process models offered by different\noperating systems.  It does not quite manage the trick.\n<\/p>\n<h5 id=\"emacs_lisp_portability\"><a href=\"#emacs_lisp_portability\">\u00a7<\/a>Emacs Lisp Portability<\/h5>\n<p>\n  Emacs Lisp portability is excellent. Emacs installations tend to\nbe upgraded frequently, so seriously out-of-date environments are\nrare.  The same extension Lisp\n  is supported everywhere and effectively\nall extensions are shipped with Emacs itself.\n<\/p>\n<p>\n  Then, too, the primitive set of Emacs is quite stable.  It\nachieved completeness for the things an editor has to do (manipulating\nbuffers, bashing text) years ago.  Only the introduction of X has\ndisturbed this picture at all, and very few Emacs modes need to be\naware of X.  Portability problems are usually manifestations of quirks\nin the C-level bindings of operating-system facilities; control of\nsubordinate processes in modes like mail agents is about the only\nissue where such problems manifest with any frequency.\n<\/p>\n<h4 id=\"id3000882\"><a href=\"#id3000882\">\u00a7<\/a>Avoiding System Dependencies<\/h4>\n<p>\n  Once your language and support libraries are chosen, the next\nportability issue is usually the location of key system files and\ndirectories: mail spools, logfile directories and the like.\nThe archetype of this sort of problem is whether the mail spool\ndirectory is\n  <code>\/var\/spool\/mail<\/code> or\n  <code>\/var\/mail<\/code>.\n<\/p>\n<p>\n  Often, you can avoid this sort of dependency by stepping back\nand reframing the problem.  Why are you opening a file in the mail\nspool directory, anyway?  If you're writing to it, wouldn't it be\nbetter to simply invoke the local mail transport agent to do it for\nyou so the file-locking gets done right?  If you're reading from it, \nmight it be better to query it through a POP3 or IMAP server?\n<\/p>\n<p>\n  The same sort of question applies elsewhere.  If you find\nyourself opening logfiles manually, shouldn't you be using\nsyslog(3)\ninstead?  Function-call interfaces through the C library are better\nstandardized than system file locations.  Use that fact!\n<\/p>\n<p>\n  If you must have system file locations in your code, your best\nalternative depends on whether you will be distributing in source code\nor binary form.  If you are distributing in source, the\n  <em>autoconf<\/em> tools we discuss in the next\nsection will help you.  If you're distributing in binary, then it's\ngood practice to have your program poke around at runtime and see if\nit can automatically adapt itself to local conditions \u2014 say, by\nactually checking for the existence of\n  <code>\/var\/mail<\/code> and\n  <code>\/var\/spool\/mail<\/code>.\n<\/p>\n<h4 id=\"id3000984\"><a href=\"#id3000984\">\u00a7<\/a>Tools for Portability<\/h4>\n<p>\n  You can often use the open-source GNU\nautoconf(1)\nwe surveyed in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>\n  to\nhandle portability issues, do system-configuration probes, and tailor\nyour makefiles.  People building from sources today expect to be able\nto type\n  <code>configure; make; make install<\/code>\n  and get a clean build.  There is a good\n  <a href=\"http:\/\/seul.org\/docs\/autotut\/\">\n   tutorial on these tools<\/a>. Even if you're distributing in binary, the\nautoconf(1)\ntools can help automate away the problem of conditionalizing your code\nfor different platforms.\n<\/p>\n<p>\n  Other tools that address this problem; two of the better known\nare the\nImake(1)\ntool associated with the X windowing system and the\n  <em>Configure<\/em> tool built by Larry Wall (later\nthe inventor of Perl) and adapted for many different\nprojects.  All are at least as complicated as the autoconf suite, and\nno longer as often used.  They don't cover as wide a range of target\nsystems.\n<\/p>\n\n<h3 id=\"ch17s06\"><a href=\"#ch17s06\">\u00a7<\/a>Internationalization<\/h3>\n<p>\n  An in-depth discussion of code internationalization \u2014\ndesigning software so the interface readily incorporates multiple\nlanguages and the vagaries of different character sets \u2014 would be\nout of scope for this book.  However, a few lessons for good practice \ndo stand out from Unix experience.\n<\/p>\n<p>\n  First,\n  <em>\n   separate the message base from the\ncode\n<\/em>.  Good Unix practice is to separate the message\nstrings a program uses from its code. so that message dictionaries\nin other languages can be plugged in without modifying the\ncode.\n<\/p>\n<p>\n  The best-known tool for this job is GNU\n  <em>gettext<\/em>, which requires that you wrap \nnative-language strings that need to be internationalized in\na special macro.  The macro uses each string as a key into per-language\ndictionaries which can be supplied as separate files.  If no such\ndictionaries are available (or if they are but the string lookup does\nnot return a match), the macro simply returns its argument, implicitly\nfalling back on the native language in the code.\n<\/p>\n<p>\n  While\n  <em>gettext<\/em> itself is messy and\nfragile as of mid-2003, its general philosophy is sound.  For many\nprojects, it is possible to craft a lighter-weight version of this\nidea with good results.\n<\/p>\n<p>\n  Second, there is a clear trend in modern Unixes to scrap all the\nhistorical cruft associated with multiple character sets and make\napplications natively speak UTF-8, the 8-bit shift encoding of the\nUnicode character set (as opposed to, say, making them natively speak\n16-bit wide characters).  The low 128 characters of UTF-8 are ASCII,\nand the low 256 are Latin-1, which means this choice is\nbackward-compatible with the two most widely used character sets.  The\nfact that XML and Java have made this choice helps, but the momentum\nis present even where XML and Java are not.\n<\/p>\n<p>\n  Third, beware of character ranges in regular expressions.  The\nelement\n  <code>[a-z]<\/code>\n  will not necessarily catch all\nlower-case letters if the script or program it's in is applied to\n(say) German, where the sharp-s or \u00df character is considered\nlower-case but does not fall in that range; similar problems arise\nwith French accented letters. Its safer to use\n  <code>[[:lower:]]<\/code>. and other symbolic ranges described in\nthe POSIX standard.\n<\/p>\n\n<h3 id=\"ch17s07\"><a href=\"#ch17s07\">\u00a7<\/a>Portability, Open Standards, and Open Source<\/h3>\n<p>\n  Portability requires standards.  Open-source reference\nimplementations are the most effective method known for both\npromulgating a standard and for pressuring proprietary vendors into\nconforming.  If you are a developer, open-source implementations of a\npublished standard can both tremendously reduce your coding workload\nand allow your product to benefit (in ways both expected and\nunexpected) from the labor of others.\n<\/p>\n<p>\n  Let's suppose, for example, you are designing image-capture\nsoftware for a digital camera.  Why write your own format for saving\nimage bits or buy proprietary code when (as we noted in\n  <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n   Chapter 5<\/a>) there is a well-tested, full-featured\nlibrary for writing PNGs in open source?\n<\/p>\n<p>\n  The (re)invention of open source has had a significant impact on\nthe standards process as well.  Though it is not formally a\nrequirement, the IETF\n  has since around 1997 grown increasingly resistant to\nstandard-tracking RFCs that do not have at least one open-source\nreference implementation. In the future, it seems likely that\nconformance to any given standard will increasingly be measured by\nconformance to (or outright use of!) open-source implementations that\nhave been blessed by the standard's authors.\n<\/p>\n <blockquote>\n  <p>\n   The flip side of this is that often the best way to\n   <em>make<\/em>\n   something a standard is to distribute a\nhigh-quality open-source implementation of it.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  In the end, the most effective step you can take to ensure the\nportability of your code is to not rely on proprietary technology.\nYou never know when the closed-source library or tool or code\ngenerator or network protocol you are depending on will be\nend-of-lifed, or when the interface will be changed in some\nbackwards-incompatible way that breaks your project.  With open-source\ncode, you have a path forward even if the leading-edge version changes\nin a way that breaks your project; because you have access to source\ncode, you can forward-port it to new platforms if you need to.\n<\/p>\n<p>\n  Until the late 1990s this advice would have been \nimpractical.  The few alternatives to relying on proprietary operating\nsystems and development tools were noble experiments, academic\nproofs-of-concept, or toys.  But the Internet changed everything; in\nmid-2003 Linux\n  and the other open-source Unixes exist\nand have proven their mettle as platforms for delivering\nproduction-quality software.  Developers have a better option now than\nbeing dependent on short-term business decisions designed to protect\nsomeone else's monopoly.  Practice defensive design \u2014 build on\nopen source and don't get stranded!\n<\/p>\n\n<h3 id=\"documentationchapter\"><a href=\"#documentationchapter\">\u00a7<\/a>Chapter 18. Documentation<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   I've never met a human being who would want to read 17,000 pages of\ndocumentation, and if there was, I'd kill him to get him out of the\ngene pool.\n  <\/p><div>\n  \n   \u2013 Joseph Costello\n  \n <\/div><\/blockquote>\n<p>\n  Unix's first application, in 1971, was as a platform for\ndocument preparation \u2014 Bell Labs used it to prepare patent\ndocuments for filing.  Computer-driven phototypesetting was still a\nnovel idea then, and for years after it debuted in 1973 Joe Ossana's\ntroff(1)\nformatter defined the state of the art.\n<\/p>\n<p>\n  Ever since, sophisticated document formatters, typesetting\nsoftware, and page-layout programs of various sorts have been an\nimportant theme in the Unix tradition.  While\ntroff(1)\nhas proven surprisingly durable, Unix has also hosted a lot of\ngroundbreaking work in this application area.  Today, Unix developers\nand Unix tools are at the cutting edge of far-reaching changes in\ndocumentation practice triggered by the advent of the World Wide\nWeb.\n<\/p>\n<p>\n  At the user-presentation level, Unix-community practice has been\nmoving rapidly toward \u2018everything is HTML, all references are\nURLs\u2019 since the mid-1990s.  Increasingly, modern Unix help\nbrowsers are simply Web browsers that know how to parse certain\nspecialized kinds of URLs (for example, \u2018man:ls(1)\u2019\ninterprets the ls(1) man page into HTML).  This relieves the problems\ncreated by having lots of different formats for documentation masters,\nbut does not entirely solve them.  Documentation composers still have\nto grapple with issues about which master format best meets their\nparticular needs.\n<\/p>\n<p>\n  In this chapter, we'll survey the rather unfortunate surfeit \nof different documentation formats and tools left behind by decades of\nexperimentation, and we'll develop guidelines for good practice\nand good style.\n<\/p>\n\n<h3 id=\"ch18s01\"><a href=\"#ch18s01\">\u00a7<\/a>Documentation Concepts<\/h3>\n<p>\n  Our first distinction is between \u201cWhat You See Is What You\nGet\u201d (WYSIWYG) documentation programs and\n  <em>markup-centered tools<\/em>. Most desktop-publishing\nprograms and word processors are in the former category; they have\nGUIs in which what one types is inserted directly into an on-screen\npresentation of the document intended to resemble the final printed\nversion as closely as possible.  In a markup-centered system, by\ncontrast, the master document is normally flat text containing\nexplicit, visible control tags and not at all resembling the intended\noutput.  The marked-up source can be modified with an ordinary text\neditor, but has to be fed to a formatter program to produce rendered\noutput for printing or display.\n<\/p>\n<p>\n  The visual-interface, WYSIWYG style was too expensive for early\ncomputer hardware, and remained rare until the advent of the Macintosh\npersonal computer in 1984. It is completely dominant on non-Unix\noperating systems today, Native Unix document tools, on the other\nhand, are almost all markup-centered. The Unix\ntroff(1)\nof 1971 was a markup formatter, and is probably the oldest such\nprogram still in use.\n<\/p>\n<p>\n  Markup-centered tools still have a role because actual implementations of WYSIWYG tend to be broken in various ways \u2014\n  some superficial, some deep.  WYSIWYG document processors have the general problem with GUIs that we discussed in <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\"> Chapter 11<\/a>; the fact that you <em>can<\/em> visually\n  manipulate anything tends to mean you <em>must<\/em> visually manipulate everything.  That would remain a problem even\n  if the WYSIWIG correspondence between screen and printer output were perfect \u2014 but it almost never is.\n<\/p>\n<p>\n  In truth, WYSIWYG document processors aren't exactly\nWYSIWIG. Most have interfaces that obscure the differences between\nscreen presentation and printer output without actually eliminating\nthem. Thus they violate the Rule of Least Surprise: the visual aspect\nof the interface encourages you to use the program like a\ntypewriter even though it is not, and your input will occasionally\nproduce an unexpected and undesired result.\n<\/p>\n<p>\n  In further truth, WYSIWIG systems actually rely on markup codes\nbut expend great effort on keeping them invisible in normal use.  Thus they\nbreak the Rule of Transparency: you can't see all of the markup, so it\nis difficult to fix documents that break because of misplaced markup\ncodes.\n<\/p>\n<p>\n  Despite its problems, WYSIWYG document processors can be very\nnice if what you want is to slide a picture three ems to the right on\nthe cover of a four-page brochure.  But they tend to be constricting\nany time you need to make a global change to the layout of a 300-page\nmanuscript. WYSIWYG users faced with that kind of challenge must give\nit up or suffer the death of a thousand mouse clicks; in situations\nlike that, there is really no substitute for being able to edit\nexplicit markup, and Unix's markup-centered document tools offer\nbetter solutions.\n<\/p>\n<p>\n  Today, in a world influenced by the example of the Web and XML,\nit has become common to make a distinction between\n  <em>presentation<\/em> and\n  <em>structural<\/em> markup in documents \u2014 the\nformer being instructions about how a document should look, the latter\nbeing instructions about how it's organized and what it means.\nThis distinction wasn't clearly understood or followed through in\nearly Unix tools, but it's important for understanding the design\npressures that led to today's descendants of them.\n<\/p>\n<p>\n  Presentation-level markup carries all the formatting information\n(e.g., about desired whitespace layout and font changes) in the\ndocument itself.  In a structural-markup system, the document has to\nbe combined with a\n  <em>stylesheet<\/em> that tells  the\nformatter how to translate the structure markup in the document to a\nphysical layout.  Both kinds of markup ultimately control the physical\nappearance of a printed or browsed document, but structural markup\ndoes it through one more level of indirection that turns out to be\nnecessary if you want to produce good results for both printing \nand the Web.\n<\/p>\n<p>\n  Most markup-centered documentation systems support a macro\nfacility.  Macros are user-defined commands that are expanded by text\nsubstitution into sequences of built-in markup requests.  Usually,\nthese macros add structural features (like the ability to declare\nsection headings) to the markup language.\n<\/p>\n <blockquote>\n  <p>\n   The troff macro sets (\n   <em>mm<\/em>,\n   <em>me<\/em>, and my\n   <em>ms<\/em> package) were actually designed to push\npeople away from format-oriented editing and toward content-oriented\nediting.  The idea was to label the semantic parts and then have\ndifferent style packages that would know whether in this style the\ntitle should be boldfaced or not, centered or not, and so on.  Thus\nthere was at one point a set of macros that tried to imitate ACM\nstyle, and another that imitated Physical Review style, but used the\nbasic\n   <code>-ms<\/code> markup.  All of the macros lost out to\npeople who were focused on producing one document, and controlling its\nappearance, just as Web pages get bogged down in the dispute over\nwhether the reader or author should control the appearance.  I\nfrequently found secretaries who were using the .AU\n(author name) command just to produce italics, noticing that it did\nthat, and then getting into trouble with its other effects.\n  <\/p><div>\n  \n   \u2013 Mike Lesk\n  \n <\/div><\/blockquote>\n<p>\n  Finally, we note that there are significant differences between\nthe sorts of things composers want to do with small documents\n(business and personal letters, brochures, newsletters) and the things\nthey want to do with large ones (books, long articles, technical\npapers, and manuals).  Large documents tend to have more structure, to\nbe pieced together from parts that may have to be changed separately,\nand to need automatically-generated features like tables of contents;\nthese are both traits that favor markup-centered tools.\n<\/p>\n\n<h3 id=\"ch18s02\"><a href=\"#ch18s02\">\u00a7<\/a>The Unix Style<\/h3>\n<p>\n  The Unix style of documentation (and documentation tools) has\nseveral technical and cultural traits that set it apart from the way\ndocumentation is done elsewhere.  Understanding these signature traits\nfirst will create context for you to understand why the programs\nand the practice look the way they do, and why the documentation reads\nthe way it does.\n<\/p>\n<h4 id=\"id3001392\"><a href=\"#id3001392\">\u00a7<\/a>The Large-Document Bias<\/h4>\n<p>\n  Unix documentation tools have always been designed primarily for\nthe challenges involved in composing large and complex documents.\nOriginally it was patent applications and paperwork; later it was\nscientific and technical papers, technical documentation of all sorts.\nConsequently, most Unix developers learned to love markup-centered\ndocumentation tools.  Unlike the PC users of the time, the Unix\nculture was unimpressed with WYSIWYG word processors when they became\ngenerally available in the late 1980s and early 1990s \u2014 and even\namong today's younger Unix hackers it is still unusual to find anyone\nwho really prefers them.\n<\/p>\n<p>\n  Dislike of opaque binary document formats \u2014 and especially\nof opaque\n  <em>proprietary<\/em>\n  binary formats \u2014 also\nplayed a part in the rejection of WYSIWYG tools.  On the other hand,\nUnix programmers seized on PostScript (the now-standard language for\ncontrolling imaging printers) with enthusiasm as soon as the language\ndocumentation became available; it fits neatly in the Unix tradition\nof domain-specific languages.  Modern open-source Unix systems have\nexcellent PostScript and Portable Document Format\n(PDF)\n  tools.\n<\/p>\n<p>\n  Another consequence of this history is that Unix documentation\ntools have tended to have relatively weak support for including\nimages, but strong support for diagrams, tables, graphing, and\nmathematical typesetting \u2014 the sorts of things often needed in\ntechnical papers.\n<\/p>\n<p>\n  The Unix attachment to markup-centered systems has often been\ncaricatured as a prejudice or a troglodyte trait, but it is not really\nanything of the kind.  Just as the putatively \u2018primitive\u2019\nCLI style of Unix is in many ways better adapted to the needs of power\nusers than GUIs, the markup-centered design of tools like\ntroff(1)\nis a better fit for the needs of power documenters than are WYSIWYG\nprograms.\n<\/p>\n<p>\n  The large-document bias in Unix tradition did not just keep Unix\ndevelopers attached to markup-based formatters like\n  <em>troff<\/em>, it also made them interested in\nstructural markup.  The history of Unix document tools is one of\nlurching, muddled, and erratic movement in a general direction away\nfrom presentation markup and toward structural markup.  In mid-2003\nthis journey is not yet over, but the end is distantly in\nsight.\n<\/p>\n<p>\n  The development of the World Wide Web meant that the ability to\nrender documents in multiple media (or, at least, for both print and\nHTML display) became the central challenge for documentation tools\nafter about 1993.  At the same time, even ordinary users were, under\nthe influence of HTML, becoming more comfortable with markup-centered\nsystems.  This led directly to an explosion of interest in structural\nmarkup and the invention of XML after 1996.  Suddenly the old-time\nUnix attachment to markup-centered systems started looking prescient\nrather than reactionary.\n<\/p>\n<p>\n  Today, in mid-2003, most of the leading-edge development of\nXML-based documentation tools using structural markup is taking place\nunder Unix.  But, at the same time, the Unix culture has yet to let go\nof its older tradition of presentation-level markup systems. The\ncreaking, clanking, armor-plated dinosaur that is\n  <em>troff<\/em> has only partly been displaced by\nHTML and XML.\n<\/p>\n<h4 id=\"id3001522\"><a href=\"#id3001522\">\u00a7<\/a>Cultural Style<\/h4>\n<p>\n  Most software documentation is written by technical writers for\nthe least-common-denominator ignorant \u2014 the knowledgeable\nwriting for the knowledgeless.  The documentation that ships with Unix\nsystems has traditionally been written by programmers for their peers.\nEven when it is not peer-to-peer documentation, it tends to be\ninfluenced in style and format by the enormous mass of\nprogrammer-to-programmer documentation that ships with Unix\nsystems.\n<\/p>\n<p>\n  The difference this makes can be summed up in one observation:\nUnix manual pages traditionally have a section called BUGS.  In other\ncultures, technical writers try to make the product look good by\nomitting and skating over known bugs.  In the Unix culture, peers\ndescribe the known shortcomings of their software to each other in\nunsparing detail, and users consider a short but informative BUGS\nsection to be an encouraging sign of quality work.  Commercial Unix\ndistributions that have broken this convention, either by suppressing\nthe BUGS section or euphemizing it to a softer tag like LIMITATIONS or\nISSUES or APPLICATION USAGE, have invariably fallen into\ndecline.\n<\/p>\n<p>\n  Where most other software documentation tends to to oscillate\nbetween incomprehensibility and oversimplifying condescension, classic\nUnix documentation is written to be telegraphic but complete.  It does\nnot hold you by the hand, but it usually points in the right\ndirection.  The style assumes an active reader, one who is able\nto deduce obvious unsaid consequences of what is said, and who has\nthe self-confidence to trust those deductions.\n<\/p>\n<p>\n  Unix programmers tend to be good at writing references,\nand most Unix documentation has the flavor of a reference or\n  \n   <em>aide memoire<\/em> \n  for someone who thinks\nlike the document-writer but is not yet an expert at his or her\nsoftware.  The results often look much more cryptic and sparse than\nthey actually are.  Read every word carefully, because whatever you\nwant to know will probably be there, or deducible from what's there.\nRead every word carefully, because you will seldom be told anything\ntwice.\n<\/p>\n\n<h3 id=\"ch18s03\"><a href=\"#ch18s03\">\u00a7<\/a>The Zoo of Unix Documentation Formats<\/h3>\n<p>\n  All the major Unix documentation formats except the very newest\none are presentation-level markups assisted by macro packages.  We\nexamine them here from oldest to newest.\n<\/p>\n<h4 id=\"id3001604\"><a href=\"#id3001604\">\u00a7<\/a>troff and the Documenter's Workbench Tools<\/h4>\n<p>\n  We discussed the Documenter's Workbench architecture and tools\nin\n  <a href=\"#minilanguageschapter\" title=\"Chapter 8. Minilanguages\">\n   Chapter 8<\/a>\n  as an\nexample of how to integrate a system of multiple minilanguages.  Now\nwe return to these tools in their functional role as a typesetting\nsystem.\n<\/p>\n<p>\n  The <em>troff<\/em> formatter interprets a presentation-level markup language.  Recent implementations like the GNU\n  project's groff(1) emit PostScript by default, though it is possible to get other forms of output by selecting a\n  suitable driver.  See Example 18.1 for several of the <em>troff<\/em> codes you might encounter in document sources.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 18.1. groff(1) markup example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nThis is running text.\n.\\\" Comments begin with a backslash and double quote.\n.ft B\nThis text will be in bold font.\n.ft R\nThis text will be back in the default (Roman) font.\nThese lines, going back to \"This is running text\", will \nbe formatted as a filled paragraph.\n.bp\nThe bp request forces a new page and a paragraph break.\nThis line will be part of the second filled paragraph.\n.sp 3\nThe .sp request emits the number of blank lines given as argument\n.nf\nThe nf request switches off paragraph filling.\nUntil the fi request switches it back on\nwhitespace and layout will be preserved.\nOne word in this line will be in \\fBbold\\fR font.\n.fi\nParagraph filling is back on.\n<\/pre><\/div>\n<p>\n  troff(1) has many other requests, but you are unlikely to see most of them directly.  Very few documents are written\n  in bare <em>troff<\/em>.  It supports a macro facility, and half a dozen macro packages are in more or less general\n  use.  Of these, the overwhelmingly most common is the man(7) macro package used to write Unix manual pages.  See\n  Example 18.2 for a sample.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 18.2.\n    <em>man<\/em> markup example.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\n.SH SAMPLE SECTION\nThe SH macro starts a section, boldfacing the section title.  \n.P\nThe P request starts a new paragraph.  The I request sets its \nargument in\n.I italics.\n.IP *\nThis starts an indented paragraph with an asterisk label.\nMore text for the first bulleted paragraph.\n.TP\nThis first line will become a paragraph label\nThis will be the first line in the paragraph, further indented\nrelative to the label.\nThe blank line just above this is treated almost exactly like a\nparagraph break (actually, like the troff-level request .sp 1).\n.SS A subsection\nThis is subsection text.  \n<\/pre><\/div>\n<p>\n  Two of the other half-dozen historical\n  <em>troff<\/em> macro libraries, \nms(7)\nand\nmm(7)\nare still in use.  BSD Unix has its own elaborate\nextended macro set,\nmdoc(7).\nAll these are designed for writing technical manuals and long-form\ndocumentation.  They are similar in style but more elaborate than man\nmacros, and oriented toward producing typeset output.\n<\/p>\n<p>\n  A minor variant of\ntroff(1)\ncalled\nnroff(1)\nproduces output for devices that can only support constant-width\nfonts, like line printers and character-cell terminals.  When you view\na Unix manual page within a terminal window, it is nroff that has\nrendered it for you.\n<\/p>\n<p>\n  The Documenter's Workbench tools do the technical-documentation\njobs they were designed for quite well, which is why they have\nremained in continuous use for more than thirty years while computers\nincreased a thousandfold in capacity. They produce typeset text of\nreasonable quality on imaging printers, and can throw a tolerable\napproximation of a formatted manual page on your screen.\n<\/p>\n<p>\n  They fall down badly in a couple of areas, however.  Their stock\nselection of available fonts is limited.  They don't handle images\nwell. It's hard to get precise control of the positioning of text or\nimages or diagrams within a page.  Support for multilingual documents\nis nonexistent.  There are numerous other problems, some chronic but\nminor and some absolute showstoppers for specific uses.  But the most\nserious problem is that because so much of the markup is presentation\nlevel, it's difficult to make good Web pages out of unmodified\n  <em>troff<\/em> sources.\n<\/p>\n<p>\n  Nevertheless, at time of writing man pages remain the single \nmost important form of Unix documentation.\n<\/p>\n<h4 id=\"id3001898\"><a href=\"#id3001898\">\u00a7<\/a>TeX<\/h4>\n<p>\n  <em>TeX<\/em> (pronounced\n  \n   \/teH\/\n  \n  with a rough h as though you are\ngargling) is a very capable typesetting program that, like the\nEmacs editor, originated outside the Unix\nculture but is now naturalized in it.  It was created by noted\ncomputer scientist Donald Knuth\n  when he became impatient with the quality of\ntypography, and especially mathematical typesetting, that was\navailable to him in the late 1970s.\n<\/p>\n<p>\n  <em>TeX<\/em>, like\ntroff(1),\nis a markup-centered system.\n  <em>TeX<\/em> 's request\nlanguage is rather more powerful than\n  <em>troff<\/em> 's; among other things, it is better\nat handling images, page-positioning content precisely, and\ninternationalization.\n  <em>TeX<\/em> is particularly\ngood at mathematical typesetting, and unsurpassed at basic typesetting\ntasks like kerning, line filling, and hyphenating.\n  <em>TeX<\/em> has become the standard submission\nformat for most mathematical journals.  It is actually now maintained\nas open source by a working group of the the American Mathematical\nSociety.  It is also commonly used for scientific papers.\n<\/p>\n<p>\n  As with\ntroff(1),\nhuman beings usually do not write large volumes of raw\n  <em>TeX<\/em> macros by hand; they use macro packages\nand various auxiliary programs instead.  One particular macro package,\n  <em>LaTeX<\/em>, is almost universal, and most people who say they're composing\nin\n  <em>TeX<\/em> almost always actually mean they're\nwriting\n  <em>LaTeX<\/em>.  Like\n  <em>troff<\/em> 's macro\npackages, a lot of its requests are semi-structural.\n<\/p>\n<p>\n  One important use of\n  <em>TeX<\/em> that is\nnormally hidden from the user is that other document-processing tools\noften generate\n  <em>LaTeX<\/em> to be turned into PostScript, rather than\nattempting the much more difficult job of generating PostScript\nthemselves.  The\nxmlto(1)\nfront end that we discussed as a shell-programming case study in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  uses this tactic; so does the XML-DocBook\ntoolchain we'll examine later in this chapter.\n<\/p>\n<p>\n  <em>TeX<\/em> has a wider application range\nthan\ntroff(1)\nand is in most ways a better design.  It has the same fundamental\nproblems as\n  <em>troff<\/em> in an increasingly\nWeb-centric world; its markup has strong ties to the presentation\nlevel, and automatically generating good Web pages from\n  <em>TeX<\/em> sources is difficult and\nfault-prone.\n<\/p>\n<p>\n  <em>TeX<\/em> is never used for Unix system\ndocumentation and only rarely used for application documentation;\nfor those purposes,\n  <em>troff<\/em> is sufficient.\nBut some software packages that originated in academia outside the\nUnix community have imported the use of\n  <em>TeX<\/em> as a documentation master format; the Python\nlanguage\n  is\none example.  As we noted above, it is also heavily used for\nmathematical and scientific papers, and will probably dominate that\nniche for some years yet.\n<\/p>\n<h4 id=\"id3002206\"><a href=\"#id3002206\">\u00a7<\/a>Texinfo<\/h4>\n<p>\n  Texinfo is a documentation markup invented by the Free Software\nFoundation and used mainly for GNU\n  project documentation \u2014 including\nthe documentation for such essential tools as\nEmacs and the GNU Compiler\nCollection.\n<\/p>\n<p>\n  Texinfo was the first markup system specifically designed to\nsupport both typeset output on paper and hypertext output for\nbrowsing.  The hypertext format was not, however, HTML; it was a more\nprimitive variety called \u2018info\u2019, originally designed to be\nbrowsed from within Emacs. On the print\nside, Texinfo turns into\n  <em>TeX<\/em> macros and can\ngo from there to PostScript.\n<\/p>\n<p>\n  The Texinfo tools can now generate HTML.  But they don't do a\nvery good or complete job, and because a lot of Texinfo's markup\nis at presentation level it is doubtful that they ever will.  As of\nmid-2003, the Free Software Foundation is working on heuristic Texinfo \nto DocBook translation.  Texinfo will probably remain a live format\nfor some time.\n<\/p>\n<h4 id=\"id3002268\"><a href=\"#id3002268\">\u00a7<\/a>POD<\/h4>\n<p>\n  Plain Old Documentation is the markup system used by the\nmaintainers of Perl.  It generates manual pages, and has\nall the familiar problems of presentation-level markups, including\ntrouble generating good HTML.\n<\/p>\n<h4 id=\"id3002291\"><a href=\"#id3002291\">\u00a7<\/a>HTML<\/h4>\n<p>\n  Since the World Wide Web entered the mainstream in the early\n1990s, a small but increasing percentage of Unix projects have been\nwriting their documentation directly in HTML.  The problem with this\napproach is that it is difficult to generate high-quality typeset\noutput from HTML.  There are particular problems with indexing as\nwell; the information needed to generate indexes is not present in\nHTML.\n<\/p>\n<h4 id=\"id3002310\"><a href=\"#id3002310\">\u00a7<\/a>DocBook<\/h4>\n<p>\n  DocBook is an SGML and XML document type definition designed for\nlarge, complex technical documents.  It is alone among the markup\nformats used in the Unix community in being purely structural.  The\nxmlto(1)\ntool discussed in\n  <a href=\"#languageschapter\" title=\"Chapter 14. Languages\">\n   Chapter 14<\/a>\n  supports rendering to HTML, XHTML,\nPostScript, PDF, Windows\nHelp markup, and several less important formats.\n<\/p>\n<p>\n  Several major open-source projects (including the Linux\nDocumentation Project, FreeBSD,\nApache, Samba, GNOME,\nand KDE) already use DocBook as a master format.  This book was\nwritten in XML-DocBook.\n<\/p>\n<p>\n  DocBook is a large topic. We'll return to it after summing up\nthe problems with the current state of Unix documentation.\n<\/p>\n\n<h3 id=\"ch18s04\"><a href=\"#ch18s04\">\u00a7<\/a>The Present Chaos and a Possible Way Out<\/h3>\n<p>\n  Unix documentation is, at present, a mess.\n<\/p>\n<p>\n  Between\n  <em>man<\/em>,\n  <em>ms<\/em>,\n  <em>mm<\/em>,\n  <em>TeX<\/em>, Texinfo, POD, HTML, and DocBook, the\ndocumentation master files on modern Unix systems are scattered across\neight different markup formats.  There is no uniform way to view all\nthe rendered versions.  They aren't Web-accessible, and they aren't\ncross-indexed.\n<\/p>\n<p>\n  Many people in the Unix community are aware that this is a\nproblem.  At time of writing most of the effort toward solving it has\ncome from open-source developers, who are more actively interested in\ncompeting for acceptance by nontechnical end users than developers\nfor proprietary Unixes have been.  Since 2000, practice has been\nmoving toward use of XML-DocBook as a documentation interchange\nformat.\n<\/p>\n<p>\n  The goal, which is within sight but will take a lot of effort to\nachieve, is to equip every Unix system with software that will act as\na systemwide document registry. When system administrators install\npackages, one step will be to enter the package's XML-DocBook \ndocumentation into the registry.  It will then be rendered into a\ncommon HTML document tree and cross-linked to the documentation\nalready present.\n<\/p>\n<p>\n  Early versions of the document-registry software are already\nworking.  The problem of forward-converting documentation from the\nother formats into XML-DocBook is a large and messy one, but the\nconversion tools are falling into place.  Other political and\ntechnical problems remain to be attacked, but are probably\nsolvable. While there is not as of mid-2003 a communitywide\nconsensus that the older formats have to be phased out, that seems the\nlikeliest working out of events.\n<\/p>\n<p>\n  Accordingly, we'll next take a very detailed look at DocBook and\nits toolchain. This description should be read as an introduction to\nXML under Unix, a pragmatic guide to practice and as a major case\nstudy. It's a good example of how, in the context of the Unix\ncommunity, cooperation between different project groups develops\naround shared standards.\n<\/p>\n\n<h3 id=\"ch18s05\"><a href=\"#ch18s05\">\u00a7<\/a>DocBook<\/h3>\n<p>\n  A great many major open-source projects are converging on\nDocBook as a standard format for their documentation.  The advocates\nof XML-based markup seem to have won the theoretical argument against\npresentation-level and for structural-level markup, and an effective\nXML-DocBook toolchain is available in open source.\n<\/p>\n<p>\n  Nevertheless, a lot of confusion still surrounds DocBook and the\nprograms that support it.  Its devotees speak an argot that is dense\nand forbidding even by computer-science standards, slinging around\nacronyms that have no obvious relationship to the things you need to\ndo to write markup and make HTML or PostScript from it.  XML standards\nand technical papers are notoriously obscure. In the rest of this\nsection, we'll try to dispel the fog of jargon.\n<\/p>\n<h4 id=\"id3006015\"><a href=\"#id3006015\">\u00a7<\/a>Document Type Definitions<\/h4>\n<p>\n  (Note: to keep the explanation simple, most of this section\ntells some lies, mainly by omitting a lot of history.  Truthfulness\nwill be fully restored in a following section.)\n<\/p>\n<p>\n  DocBook is a structural-level markup language.  Specifically, it\nis a dialect of XML.  A DocBook document is a piece of XML that uses\nXML tags for structural markup.\n<\/p>\n<p>\n  For a document formatter to apply a stylesheet to your document\nand make it look good, it needs to know things about the overall\nstructure of your document.  For example, in order to physically\nformat chapter headers properly, it needs to know that a book\nmanuscript normally consists of front matter, a sequence of chapters,\nand back matter.  In order for it to know this sort of thing, you\nneed to give it a\n  <em>Document Type Definition<\/em>\n  or DTD. The\nDTD tells your formatter what sorts of elements can be in the document\nstructure, and in what order they can appear.\n<\/p>\n<p>\n  What we mean by calling DocBook a \u2018dialect\u2019 of XML\nis actually that DocBook is a DTD \u2014 a rather large DTD, with\nsomewhere around 400 tags in it.<sup><a href=\"#ftn.id3006067\" id=\"id3006067\">[152]<\/a><\/sup><\/p>\n<p>\n  Lurking behind DocBook is a kind of program called a\n  <em>validating parser<\/em>. When you format a DocBook document, the\nfirst step is to pass it through a validating parser (the front end of\nthe DocBook formatter).  This program checks your document against the\nDocBook DTD to make sure you aren't breaking any of the DTD's\nstructural rules (otherwise the back end of the formatter, the part\nthat applies your stylesheet, might become quite confused).\n<\/p>\n<p>\n  The validating parser will either throw an error, giving you messages\nabout places where the document structure is broken, or translate the\ndocument into a stream of XML elements and text that the parser back\nend combines with the information in your stylesheet to produce\nformatted output.\n<\/p>\n<p>\n   Figure 18.1 diagrams the whole process.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 18.1. Processing structural documents.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Processing structural documents.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  The part of the diagram inside the dotted box is your formatting\nsoftware, or\n  <em>toolchain<\/em>. Besides the obvious and\nvisible input to the formatter (the document source) you'll need to\nkeep the two hidden inputs of the formatter (DTD and stylesheet) in\nmind to understand what follows.\n<\/p>\n<h4 id=\"id3006182\"><a href=\"#id3006182\">\u00a7<\/a>Other DTDs<\/h4>\n<p>\n  A brief digression into other DTDs may help clarify what parts of\nthe previous section are specific to DocBook and what parts are general to\nall structural-markup languages.\n<\/p>\n<p>\n  <a href=\"http:\/\/www.tei-c.org\/\">\n   TEI<\/a>\n  (Text Encoding\nInitiative) is a large, elaborate DTD used primarily in academia for\ncomputer transcription of literary texts.  TEI's Unix-based toolchains\nuse many of the same tools that are involved with DocBook, but with\ndifferent stylesheets and (of course) a different DTD.\n<\/p>\n<p>\n  XHTML, the latest version of HTML, is also an XML application\ndescribed by a DTD, which explains the family resemblance between\nXHTML and DocBook tags. The XHTML toolchain consists of Web browsers\nthat can format HTML as flat ASCII, together with any of a number of\nad-hoc HTML-to-print utilities.\n<\/p>\n<p>\n  Many other XML DTDs are maintained to help people exchange\nstructured information in fields as diverse as bioinformatics and\nbanking.  You can look at a\n  <a href=\"http:\/\/www.xml.com\/pub\/rg\/DTD_Repositories\">\n   list of\nrepositories<\/a>\n  to get some idea of the variety available.\n<\/p>\n<h4 id=\"db_toolchain\"><a href=\"#db_toolchain\">\u00a7<\/a>The DocBook Toolchain<\/h4>\n<p>\n  Normally, what you'll do to make XHTML from your DocBook sources\nis use the\nxmlto(1)\nfront end.  Your commands will look like this:\n<\/p>\n <pre class=\"screen\">\nbash$ xmlto xhtml foo.xml\nbash$ ls *.html\nar01s02.html ar01s03.html ar01s04.html index.html\n<\/pre>\n<p>\n  In this example, you converted an XML-DocBook  document named\n  <code>foo.xml<\/code> with three top-level sections into an\nindex page and two parts.  Making one big page is just as easy:\n<\/p>\n <pre class=\"screen\">\nbash$ xmlto xhtml-nochunks foo.xml\nbash$ ls *.html\nfoo.html\n<\/pre>\n<p>\n  Finally, here is how you make PostScript for printing:\n<\/p>\n <pre class=\"screen\">\nbash$ xmlto ps foo.xml       # To make PostScript\nbash$ ls *.ps\nfoo.ps\n<\/pre>\n<p>\n  To turn your documents into HTML or PostScript, you need an engine that can apply the combination of DocBook DTD and a\n  suitable stylesheet to your document. Figure 18.2 illustrates how the open-source tools for doing this fit together.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 18.2. Present-day XML-DocBook toolchain.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Present-day XML-DocBook toolchain.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  Parsing your document and applying the stylesheet transformation\nwill be handled by one of three programs.  The most likely one is\n  <em>xsltproc<\/em>,\nthe parser that ships with Red Hat Linux.  The other possibilities are two\nJava\n  programs,\n  <em>Saxon<\/em> and\n  <em>Xalan<\/em>.\n<\/p>\n<p>\n  It is relatively easy to generate high-quality XHTML from either\nDocBook; the fact that XHTML is simply another XML DTD helps a lot.\nTranslation to HTML is done by applying a rather simple stylesheet,\nand that's the end of the story.  RTF is also simple to generate in\nthis way, and from XHTML or RTF it's easy to generate a flat ASCII\ntext approximation in a pinch.\n<\/p>\n<p>\n  The awkward case is print.  Generating high-quality printed\noutput \u2014 which means, in practice, Adobe's\nPDF\n  (Portable Document Format) \u2014 is difficult.  Doing it right requires\nalgorithmically duplicating the delicate judgments of a human\ntypesetter moving from content to presentation level.\n<\/p>\n<p>\n  So, first, a stylesheet translates DocBook's structural markup\ninto another dialect of XML \u2014 FO\n  (Formatting Objects).  FO\nmarkup is very much presentation-level; you can think of it as a sort\nof XML functional equivalent of\n  <em>troff<\/em>.  It\nhas to be translated to PostScript for packaging in a PDF.\n<\/p>\n<p>\n  In the toolchain shipped with Red Hat\nLinux, this job\nis handled by a\n  <em>TeX<\/em> macro package called\n  <em>PassiveTeX<\/em>. It\ntranslates the formatting objects generated by\n  <code>xsltproc<\/code> into Donald Knuth's <em>TeX<\/em> language. <em>TeX<\/em> 's output, known as\nDVI (DeVice Independent) format, is then massaged into PDF.\n<\/p>\n<p>\n  If you think this bucket chain of XML to\n  <em>TeX<\/em> macros to DVI to\nPDF sounds like an awkward kludge, you're right.  It clanks, it\nwheezes, and it has ugly warts.  Fonts are a significant problem,\nsince XML and\n  <em>TeX<\/em> and PDF have very\ndifferent models of how fonts work; also, handling\ninternationalization and localization is a nightmare. About the only\nthing this code path has going for it is that it works.\n<\/p>\n<p>\n  The elegant way will be\nFOP, a direct\nFO-to-PostScript translator being developed by the\nApache\n  project.  With\nFOP, the internationalization problem is, if not solved, at least well\nconfined; XML tools handle Unicode all the way through to FOP.  The\nmapping from Unicode glyphs to Postscript font is also strictly FOP's\nproblem.  The only trouble with this approach is that it doesn't work\n\u2014 yet.  As of mid-2003, FOP is in an unfinished alpha state\n\u2014 usable, but with rough edges and missing features.\n<\/p>\n<p>\n   Figure 18.3 illustrates what the FOP toolchain looks like.\n<\/p>\n <div class=\"figure\"><p class=\"title\">\n   <b>\n    Figure 18.3. Future XML-DocBook toolchain with FOP.\n   <\/b>\n  <\/p><div class=\"center\"><img alt=\"Future XML-DocBook toolchain with FOP.\" src=\"denied:data:image\/png;base64,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\" \/><\/div><\/div>\n<p>\n  FOP has competition.  Another project called\n  <em>xsl-fo-proc<\/em> aims to do the same\nthings as FOP, but in C++\n  (and therefore both faster than\nJava\n  and not\nrelying on the Java environment).  As of mid-2003,\n  <em>xsl-fo-proc<\/em> is in an unfinished alpha\nstate, not as far along as FOP.\n<\/p>\n<h4 id=\"id3006742\"><a href=\"#id3006742\">\u00a7<\/a>Migration Tools<\/h4>\n<p>\n  The second biggest problem with DocBook is the effort needed to\nconvert old-style presentation markup to DocBook markup.  Human beings\ncan usually parse the presentation of a document into logical\nstructure automatically, because (for example) they can tell from \ncontext when an italic font means \u2018emphasis\u2019 and when it means\nsomething else such as \u2018this is a foreign phrase\u2019.\n<\/p>\n<p>\n  Somehow, in converting documents to DocBook, those\nsorts of distinctions need to be made explicit.  Sometimes\nthey're present in the old markup; often they are not, and the\nmissing  structural information has to be either deduced by \nclever heuristics or added by a human.\n<\/p>\n<p>\n  Here is a summary of the state of conversion tools from\nvarious other formats.  None of these do a completely perfect job;\ninspection and perhaps a bit of hand-editing by a human being\nwill be needed after conversion.\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     GNU Texinfo\n    \n   <\/dt>\n   <dd>\n    <p>\n     The Free Software Foundation\n     intends to support DocBook as an\ninterchange format.  Texinfo has enough structure to make reasonably\ngood automatic conversion possible (human editing is still needed\nafterwards, but not much of it), and the 4.x versions of\n     <em>makeinfo<\/em> feature a\n     <code>--docbook<\/code> switch that generates DocBook.  More at the\n     <a href=\"http:\/\/www.gnu.org\/directory\/texinfo.html\">\n      <em>makeinfo<\/em> project page<\/a>.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     POD\n    \n   <\/dt>\n   <dd>\n    <p>\n     A\n     <a href=\"http:\/\/www.cpan.org\/modules\/by-module\/Pod\/\">\n      POD::DocBook<\/a>\n     module translates Plain Old Documentation markup to DocBook.  It\nclaims to translate every POD tag except the L&lt;&gt; italic tag.\nThe man page also says \u201cNested =over\/=back lists are not supported\nwithin DocBook\u201d, but notes that the module has been heavily\ntested.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <em>LaTeX<\/em> \n   <\/dt>\n   <dd>\n    <p>\n     A project called\n     <a href=\"http:\/\/www.lrz-muenchen.de\/services\/software\/sonstiges\/tex4ht\/mn.html\">\n      TeX4ht<\/a>\n     can, according to the author of Passive\n     <em>TeX<\/em>, \ngenerate DocBook from\n     <em>LaTeX<\/em>.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     man pages and other\n     <em>troff<\/em> -based markups\n    \n   <\/dt>\n   <dd>\n    <p>\n     These are generally considered the biggest and nastiest conversion\nproblems.  And indeed, the basic\ntroff(1) markup is at too low a presentation\nlevel for automatic conversion tools to do much of any good.  However,\nthe gloom in the picture lightens significantly if we consider\ntranslation from sources of documents written in macro packages like\nman(7).  These have enough structural\nfeatures for automatic translation to get some traction.\n    <\/p>\n    <p>\n     I wrote a tool to do troff-to-DocBook\nmyself, because I couldn't find anything else that did a tolerable\njob of it.  It's called\n     <a href=\"http:\/\/www.catb.org\/~esr\/doclifter\/\">\n      doclifter<\/a>.  It\nwill translate to either SGML or XML DocBook from\nman(7),\nmdoc(7),\nms(7), or\nme(7) macros.  See the documentation\nfor details.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<h4 id=\"id3007022\"><a href=\"#id3007022\">\u00a7<\/a>Editing Tools<\/h4>\n<p>\n  One thing we do not have in mid-2003 is a good\nopen-source structure editor for SGML\/XML documents.\n<\/p>\n<p>\n  <a href=\"http:\/\/www.lyx.org\/\">\n   LyX<\/a>\n  is a GUI word processor\nthat uses\n  <em>LaTeX<\/em> for printing and supports structural editing of\n  <em>LaTeX<\/em> markup.  There is a\n  <em>LaTeX<\/em> package that generates DocBook, and a\n  <a href=\"http:\/\/bgu.chez.tiscali.fr\/doc\/db4lyx\/\">\n   how-to document<\/a>\n  describing how to write SGML and XML in the LyX GUI.\n<\/p>\n<p>\n  <a href=\"http:\/\/www.math.u-psud.fr\/~anh\/TeXmacs\/TeXmacs.html\">\n   GNU\n   <em>TeX<\/em> Macs<\/a>\n  is a project aimed at producing an editor that is good\nfor technical and mathematical material, including displayed formulas.\n1.0 was released in April 2002.  The developers plan XML support in\nthe future, but it's not there yet.\n<\/p>\n<p>\n  Most people still hack DocBook tags by hand using either\n  <em>vi<\/em> or\n  <em>emacs<\/em>.\n<\/p>\n<h4 id=\"id3007143\"><a href=\"#id3007143\">\u00a7<\/a>Related Standards and Practices<\/h4>\n<p>\n  The tools are coming together, if slowly, to edit and format\nDocBook markup. But DocBook itself is a means, not an end.  We'll need\nother standards besides DocBook itself to accomplish the\nsearchable-documentation-database objective. There are two big issues:\ndocument cataloging and metadata.\n<\/p>\n<p>\n  The\n  <a href=\"http:\/\/scrollkeeper.sourceforge.net\/\">\n   ScrollKeeper<\/a>\n  project aims directly to meet this need. It provides a simple set of\nscript hooks that can be used by package install and uninstall\nproductions to register and unregister their documentation.\n<\/p>\n<p>\n  ScrollKeeper uses the\n  <a href=\"http:\/\/www.ibiblio.org\/osrt\/omf\/\">\n   Open Metadata Format<\/a>.\nThis is a standard for indexing open-source documentation analogous to\na library card-catalog system.  The idea is to support rich search\nfacilities that use the card-catalog metadata as well as the source \ntext of the documentation itself.\n<\/p>\n<h4 id=\"id3007192\"><a href=\"#id3007192\">\u00a7<\/a>SGML<\/h4>\n<p>\n  In previous sections, we have deliberately omitted a lot of\nDocBook's history.  XML has an older brother,\nStandard Generalized Markup Language (SGML).\n<\/p>\n<p>\n  Until mid-2002, no discussion of DocBook would have been\ncomplete without a long excursion into SGML, the differences between\nSGML and XML, and detailed descriptions of the SGML DocBook toolchain.\nLife can be simpler now; an XML DocBook toolchain is available in open\nsource, works as well as the SGML toolchain ever did, and is easier to\nuse.\n<\/p>\n<h4 id=\"id3007297\"><a href=\"#id3007297\">\u00a7<\/a>XML-DocBook References<\/h4>\n<p>\n  One of the things that makes learning DocBook difficult is that\nthe sites related to it tend to overwhelm the newbie with long lists\nof W3C standards, massive exercises in SGML theology, and dense\nthickets of abstract terminology.  See\n  <em>XML in a Nutshell<\/em><sup><a href=\"#Harold-Means\" title=\"[Harold-Means]\">[Harold-Means]<\/a><\/sup> for a good\nbook-length general introduction.\n<\/p>\n<p>\n  Norman Walsh's\n  <em>DocBook: The Definitive Guide<\/em>\n  is available\n  <a href=\"http:\/\/www.oreilly.com\/catalog\/docbook\/\">\n   in print<\/a>\n  and\n  <a href=\"http:\/\/www.docbook.org\/tdg\/en\/html\/docbook.html\">\n   on the\nWeb<\/a>.  This is indeed the definitive reference, but as an\nintroduction or tutorial it's a disaster.  Instead, read this:\n<\/p>\n<p>\n  <a href=\"http:\/\/xml.web.cern.ch\/XML\/goossens\/dbatcern\/\">\n   Writing\nDocuments Using DocBook<\/a>.  This is an\nexcellent tutorial.\n<\/p>\n<p>\n  There is an equally excellent\n  <a href=\"http:\/\/www.dpawson.co.uk\/docbook\/\">\n   DocBook FAQ<\/a>\n  with a lot\nof material on styling HTML output.  There is also a DocBook\n  <a href=\"http:\/\/docbook.org\/wiki\/moin.cgi\">\n   wiki<\/a>.\n<\/p>\n<p>\n  Finally, the\n  <a href=\"http:\/\/xml.coverpages.org\/\">\n   The XML\nCover Pages<\/a>\n  will take you into the jungle of XML standards\nif you really want to go there.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3006067\" id=\"ftn.id3006067\">[152]<\/a><\/sup>\n  In XML-speak, what we\nhave been calling a \u2018dialect\u2019 is called an\n\u2018application\u2019; we've avoided that usage, since it collides\nwith another more common sense of the word.\n<\/p><\/div>\n\n<h3 id=\"ch18s06\"><a href=\"#ch18s06\">\u00a7<\/a>Best Practices for Writing Unix Documentation<\/h3>\n<p>\n  The advice we gave earlier in the chapter about reading Unix\ndocumentation can be turned around.  When you write documentation for\npeople within the Unix culture,\n  <em>\n   don't dumb it\ndown\n<\/em>.  If you write as if for idiots, you will be\nwritten off as an idiot yourself.  Dumbing documentation down\nis very different from making it accessible; the former is\nlazy and omits important things, whereas the latter requires\ncareful thought and ruthless editing.\n<\/p>\n<p>\n  Don't think for a moment that volume will be mistaken for\nquality.  And especially, never\n  <em>ever<\/em>\n  omit\nfunctional details because you fear they might be confusing, nor\nwarnings about problems because you don't want to look bad.  It is\n  <em>unanticipated<\/em>\n  problems that will cost you\ncredibility and users, not the problems you were honest about.\n<\/p>\n<p>\n  Try to hit a happy medium in information density. Too low\nis as bad as too high.  Use screen shots sparingly; they tend\nto convey little information beyond the style and feel of the\ninterface.  They are never a good substitute for clear textual\ndescription.\n<\/p>\n<p>\n  If your project is of any significant size, you should probably\nbe shipping three different kinds of documentation: man pages as\nreference material, a tutorial manual, and a FAQ (Frequently Asked\nQuestions) list.  You should have a website as well, to serve as a\ncentral point of distribution (see the guidelines on communication \nin\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>).\n<\/p>\n<p>\n  Huge man pages are viewed with some disfavor; navigation within\nthem can be difficult. If yours are getting large, consider writing a\nreference manual, with the man page(s) giving a quick summary,\npointers into the reference manual, and details of how the program(s)\nare invoked.\n<\/p>\n<p>\n  In your source code, include the standard metainformation files\ndescribed in the\n  <a href=\"#opensourcechapter\" title=\"Chapter 19. Open Source\">\n   Chapter 19<\/a>\n  section on open-source release\npractices, such as\n  <code>README<\/code>.  Even if your code is\ngoing to be proprietary, these are Unix conventions and future\nmaintainers coming from a Unix background will come up to speed faster\nif the conventions are followed.\n<\/p>\n<p>\n  Your man pages should be command references in the traditional\nUnix style for the traditional Unix audience.  The tutorial manual\nshould be long-form documentation for nontechnical users.  And the\nFAQ should be an evolving resource that grows as your software support\ngroup learns what the frequent questions are and how to answer\nthem.\n<\/p>\n<p>\n  There are more specific habits you should adopt if\nyou want to get a little ahead of mid-2003's practice:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     Maintain your document masters in XML-DocBook.  Even your man pages\ncan be DocBook\n     <code class=\"sgmltag-element\">\n      RefEntry\n     <\/code>\n     documents.  There is a very good\n     <a href=\"http:\/\/www.tldp.org\/HOWTO\/mini\/Man-Page.html\">\n      HOWTO<\/a>\n     on writing manual pages\nthat explains the sections and organization your users will expect to\nsee.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Ship the XML masters.  Also, in case your users' systems don't\nhave\nxmlto(1)\nship the\n     <em>troff<\/em> sources that you get by running\n     <code>xmlto man<\/code>\n     on your masters. Your software distribution's\ninstallation procedure should install those in the normal way, but\ndirect people to the XML files if they want to write or edit\ndocumentation.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Make your project's installation package ScrollKeeper-ready.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Generate XHTML from your masters (with <code>xmlto xhtml<\/code>)\n     and make it available from your project's Web page.\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n<p>\n  Whether or not you're using XML-DocBook as a master format,\nyou'll want to find a way to convert your documentation to HTML.\nWhether your software is open-source or proprietary, users are\nincreasingly likely to find it via the Web.  Putting your\ndocumentation on-line has the direct effect of making it easier for\npotential users and customers who know your software exists to read it\nand learn about it.  It has the indirect effect that your software\nwill become more likely to turn up in a Web search.\n<\/p>\n\n<h3 id=\"opensourcechapter\"><a href=\"#opensourcechapter\">\u00a7<\/a>Chapter 19. Open Source<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   Software is like sex \u2014 it's better when it's free.\n  <\/p><div>\n  \n   \u2013 Linus Torvalds\n  \n <\/div><\/blockquote>\n<p>\n  We concluded\n  <a href=\"#historychapter\" title=\"Chapter 2. History\">\n   Chapter 2<\/a>\n  by observing the largest-scale pattern in\nUnix's history; it flourished when its practices most closely\napproximated open source, and stagnated when they did not.  We then\nasserted in\n  <a href=\"#reusechapter\" title=\"Chapter 16. Reuse\">\n   Chapter 16<\/a>\n  that\nopen-source development tools tend to be of high quality.  We'll begin\nthis chapter by sketching an explanation of how and why open-source\ndevelopment works.  Most of its behaviors are simply intensifications\nof long-established Unix-tradition practices.\n<\/p>\n<p>\n  We'll then descend from realm of abstraction and describe some\nof the most important folk customs that Unix has picked up from the\nopen-source community \u2014 in particular, the community-evolved\nguidelines for what a good source-code release looks like.  Many of\nthese customs could be profitably adopted by developers on other modern\noperating systems as well.\n<\/p>\n<p>\n  We'll describe these customs on the assumption that you are\ndeveloping open source; most are still good ideas even if you are\nwriting proprietary software.  The open-source assumption is also\nhistorically appropriate, because many of these customs found their\nway back into proprietary Unix shops via ubiquitous open-source tools\nlike\npatch(1),\n  <em>Emacs<\/em>, and GCC.\n<\/p>\n\n<h3 id=\"ch19s01\"><a href=\"#ch19s01\">\u00a7<\/a>Unix and Open Source<\/h3>\n<p>\n  Open-source development exploits the fact that characterizing\nand fixing bugs \u2014 unlike, say, implementing a particular\nalgorithm \u2014 is a task that lends itself well to being split into\nmultiple parallel subtasks.  Exploration of the neighborhood of\npossibilities near a prototype design also parallelizes well.  With\nthe right technological and social machinery in place, development\nteams that are loosely networked and very large can do astoundingly\ngood work.\n<\/p>\n<p>\n  Astoundingly, that is, if you are carrying around the mental\nhabits developed by people who treat process secrecy and proprietary\ncontrol as a given. From\n  <em>The Mythical Man-Month<\/em> <sup><a href=\"#Brooks\" title=\"[Brooks]\">[Brooks]<\/a><\/sup> until the rise of\nLinux,\nthe orthodoxy in software engineering was all about small, closely\nmanaged teams within heavyweight organizations like corporations and\ngovernment. The practice was of\n  <em>large<\/em>\n  teams\nclosely managed.\n<\/p>\n<p>\n  The early Unix community, before the\nAT&amp;T\n  divestiture, was a paradigmatic\nexample of open source in action.  While the pre-divestiture Unix code\nwas technically and legally proprietary, it was treated as a commons\nwithin its user\/developer community.  Volunteer efforts were\nself-directed by the people most strongly motivated to solve problems.\nFrom these choices many good things flowed.  Indeed, the technique of\nopen-source development evolved as an unconscious folk practice in the\nUnix community for more than a quarter century, many years before it was\nanalyzed and labeled in the late 1990s (See\n  <em>The Cathedral and the Bazaar<\/em><sup><a href=\"#Raymond01\" title=\"[Raymond01]\">[Raymond01]<\/a><\/sup> and\n  <em>Understanding Open Source Software Development<\/em><sup><a href=\"#Feller-Fitzgerald\" title=\"[Feller-Fitzgerald]\">[Feller-Fitzgerald]<\/a><\/sup>.\n<\/p>\n<p>\n  In retrospect, it is rather startling how oblivious we all were\nto the implications of our own behavior. Several people came very\nclose to understanding the phenomenon; Richard Gabriel in his\n\u201cWorse Is Better\u201d\n  paper from 1990 <sup><a href=\"#Gabriel\" title=\"[Gabriel]\">[Gabriel]<\/a><\/sup>\nis the best known, but one can find prefigurations in Brooks <sup><a href=\"#Brooks\" title=\"[Brooks]\">[Brooks]<\/a><\/sup> (1975) and as far back as Vyssotsky and\nCorbat\u00f3's meditations on the Multics design (1965). I failed to\nget it over more than twenty years of observing software development,\nbefore being awakened by Linux in the mid-1990s.  This experience\nshould make any thoughtful and humble person wonder what other\nimportant unifying concepts are still implicit in our behavior and\nlurking right under our collective noses, hidden not by their\ncomplexity but by their very simplicity.\n<\/p>\n<p>\n  The rules of open-source development are simple:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     <em>Let the source be open.<\/em>\n     Have no secrets.  Make the code and the process that produces it\npublic. Encourage third-party peer review.  Make sure that others \ncan modify and redistribute the code freely.  Grow the co-developer\ncommunity as big as you can.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Release early, release often.<\/em>\n     A rapid\nrelease tempo means quick and effective feedback.  When each\nincremental release is small, changing course in response to\nreal-world feedback is easier.\n    <\/p>\n    <p>\n     Just make sure your first release builds, runs, and demonstrates\npromise.  Usually, an initial version of an open-source program\ndemonstrates promise by doing at least some portion of its final job,\nsufficient to show that the initiator can actually continue the\nproject.  For example, an initial version of a word processor might\nsupport typing in text and displaying it on the screen.\n    <\/p>\n    <p>\n     A first release that cannot be compiled or run can kill a\nproject (as, famously, almost happened to the Mozilla browser).\nReleases that cannot compile suggest that the project developers will\nbe unable to complete the project, Also, non-working programs are\ndifficult for other developers to contribute to, because they cannot\neasily determine if any change they made improved the program or not.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     <em>Reward contribution with praise.<\/em>\n     If you can't give your co-developers material rewards, \ngive psychological ones.  Even if you can, remember that people will\noften work harder for reputation than they would for gold.\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n <blockquote>\n  <p>\n   A corollary of rule 2 is that individual releases should not be\nmomentous events, with many promises attached and much preparation.\nIt's important to ruthlessly streamline your release process, so that\nyou\n   <em>can<\/em>\n   do frequent releases painlessly.  A setup\nwhere all other work must stop during release preparation is a\nterrible mistake.  (Notably, if you're using CVS or something similar,\nreleases in preparation should be branches off the main line of\ndevelopment, so that they don't block main-line progress.)  To sum up,\ndon't treat releases as big special events; make them part of normal\nroutine.\n  <\/p><div>\n  \n   \u2013 Henry Spencer\n  \n <\/div><\/blockquote>\n<p>\n  Remember that the reason for frequent releases is to shorten and\nspeed the feedback loop connecting your user population to your\ndevelopers.  Therefore, resist thinking of the next release as a\npolished jewel that cannot ship until everything is perfect.  Don't\nmake long wish lists.  Make progress incrementally, admit and advertise\ncurrent bugs, and have confidence that perfection will come with\ntime.  Accept that you will go through dozens of point releases on the\nway, and don't get upset as the version numbers mount.\n<\/p>\n<p>\n  Open-source development uses large teams of programmers\ndistributed over the Internet and communicating primarily through\nemail and Web documents.  Typically, most contributors to any given\nproject are volunteers contributing in order to be rewarded by the\nincreased usefulness of the software to them, and by reputation\nincentives.  A central individual or core group steers the project;\nother contributors may drop in and drop out sporadically.  To\nencourage casual contributors, it is important to avoid erecting\nsocial barriers between them and the core team.  Minimize the core\nteam's privileged status, and work hard to keep the boundaries\ninconspicuous.\n<\/p>\n<p>\n  Open-source projects follow the Unix-tradition advice of\nautomating wherever possible.  They use the\npatch(1)\ntool to pass around incremental changes.  Many projects (and all large\nones) have network-accessible code repositories using version-control\nsystems like CVS (recall the discussion in\n  <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n   Chapter 15<\/a>). Use of automated bug- and patch-tracking\nsystems is also common.\n<\/p>\n<p>\n  In 1997, almost nobody outside the hacker culture understood\nthat it was even possible to run a large project this way, let alone\nget high-quality results.  In 2003 this is no longer news;\nprojects like Linux,\nApache, and Mozilla\nhave achieved both success and high public visibility.\n<\/p>\n<p>\n  Abandoning the habit of secrecy in favor of process\ntransparency\n  and peer review was\nthe crucial step by which alchemy became chemistry.  In the same way,\nit is beginning to appear that open-source development may signal the\nlong-awaited maturation of software development as a\ndiscipline.\n<\/p>\n\n<h3 id=\"ch19s02\"><a href=\"#ch19s02\">\u00a7<\/a>Best Practices for Working with Open-Source Developers<\/h3>\n<p>\n  Much of what constitutes best practice in the open-source\ncommunity is a natural adaptation to distributed development; you'll\nread a lot in the rest of this chapter about behaviors that maintain\ngood communication with other developers.  Where Unix conventions are\narbitrary (such as the standard names of files that convey\nmetainformation about a source distribution) they often trace back\neither to Usenet\n  in\nthe early 1980s, or to the conventions and standards of the GNU\nproject.\n<\/p>\n<h4 id=\"patching\"><a href=\"#patching\">\u00a7<\/a>Good Patching Practice<\/h4>\n<p>\n  Most people become involved in open-source software by writing patches\nfor other people's software before releasing projects of their own.\nSuppose you've written a set of source-code changes for someone else's\nbaseline code. Now put yourself in that person's shoes. How is he\nto judge whether to include the patch?\n<\/p>\n<p>\n  It is very difficult to judge the quality of code, so\ndevelopers tend to evaluate patches by the quality of the submission.\nThey look for clues in the submitter's style and communications\nbehavior instead \u2014 indications that the person has been in their\nshoes and understands what it's like to have to evaluate and merge an\nincoming patch.\n<\/p>\n<p>\n  This is actually a rather reliable proxy for code quality.  In\nmany years of dealing with patches from many hundreds of strangers,\nI have only seldom seen a patch that was thoughtfully\npresented and respectful of my time but technically bogus.  On the\nother hand, experience teaches that patches which look careless or\nare packaged in a lazy and inconsiderate way are very likely to\nactually\n  <em>be<\/em>\n  bogus.\n<\/p>\n<p>\n  Here are some tips on how to get your patch accepted:\n<\/p>\n<h5 id=\"id3007894\"><a href=\"#id3007894\">\u00a7<\/a>Do send patches, don't send whole archives or files.<\/h5>\n<p>\n  If your change includes a new file that doesn't exist in the code,\nthen of course you have to send the whole file.  But if you're modifying\nan already-existing file, don't send the whole file.  Send a diff instead;\nspecifically, send the output of the \ndiff(1)\ncommand run to compare the baseline distributed version against your\nmodified version.\n<\/p>\n<p>\n  The\ndiff(1)\ncommand and its dual,\npatch(1),\nare the most basic tools of open-source development.  Diffs are better\nthan whole files because the developer you're sending a patch to may\nhave changed the baseline version since you got your copy.  By sending\nhim a diff you save him the effort of separating your changes from\nhis; you show respect for his time.\n<\/p>\n<h5 id=\"id3007953\"><a href=\"#id3007953\">\u00a7<\/a>Send patches against the current version of the code.<\/h5>\n<p>\n  It is both counterproductive and rude to send a maintainer\npatches against the code as it existed several releases ago, and expect\nhim to do all the work of determining which changes duplicate things he\nhas since done, versus which things are actually novel in your patch.\n<\/p>\n<p>\n  As a patch submitter, it is\n  <em>your<\/em>\n  responsibility\nto track the state of the source and send the maintainer a minimal patch\nthat expresses what you want done to the main-line codebase. That means\nsending a patch against the current version.\n<\/p>\n<h5 id=\"id3007988\"><a href=\"#id3007988\">\u00a7<\/a>Don't include patches for generated files.<\/h5>\n<p>\n  Before you send your patch, walk through it and delete any patch\nbands for files in it that are going to be automatically regenerated\nonce the maintainer applies the patch and remakes.  The classic\nexamples of this error are C\n  files\ngenerated by\n  <em>Bison<\/em> or\n  <em>Flex<\/em>.\n<\/p>\n<p>\n  These days the most common mistake of this kind is sending a\ndiff with a huge band that is nothing but changebars between your\n  <code>configure<\/code> script and the maintainer's.  This file is generated by <code>autoconf<\/code>.\n<\/p>\n<p>\n  This is inconsiderate.  It means your recipient is put to the trouble\nof separating the real content of the patch from a lot of bulky noise.\nIt's a minor error, not as important as some of the things we'll get\nto further on \u2014 but it will count against you.\n<\/p>\n<h5 id=\"id3008059\"><a href=\"#id3008059\">\u00a7<\/a>Don't send patch bands that just tweak RCS or SCCS $-symbols.<\/h5>\n<p>\n  Some people put special tokens in their source files that are\nexpanded by the version-control system when the file is checked in:\nthe $Id$ construct used by RCS and CVS,\nfor example.\n<\/p>\n<p>\n  If you're using a local version-control system yourself, your\nchanges may alter these tokens.  This isn't really harmful, because\nwhen your recipient checks his code back in after applying your patch\nthe tokens will be re-expanded in accordance with the\n  <em>maintainer's<\/em>\n  version-control status.  But those\nextra patch bands are noise.  They're distracting.  It's more\nconsiderate not to send them.\n<\/p>\n<p>\n  This is another minor error.  You'll be forgiven for it if you\nget the big things right.  But you want to avoid it anyway.\n<\/p>\n<h5 id=\"id3008100\"><a href=\"#id3008100\">\u00a7<\/a>Do use -c or -u format, don't use the default (-e) format.<\/h5>\n<p>\n  The default (\n  <code>-e<\/code>) format of\ndiff(1)\nis very brittle.  It doesn't include any context, so the patch tool\ncan't cope if any lines have been inserted or deleted in the baseline \ncode since you took the copy you modified.\n<\/p>\n<p>\n  Getting an\n  <code>-e<\/code> diff is annoying, and suggests\nthat the sender is either an extreme newbie, careless, or clueless.\nMost such patches get tossed out without a second thought.\n<\/p>\n<h5 id=\"id3008156\"><a href=\"#id3008156\">\u00a7<\/a>Do include documentation with your patch.<\/h5>\n<p>\n  This is very important.  If your patch makes a user-visible addition\nor change to the software's features,\n  <em>\n   include changes to the\nappropriate man pages and other documentation files in your patch\n<\/em>.\nDo\n  <em>not<\/em>\n  assume that the recipient will be happy to\ndocument your code for you, or to have undocumented features lurking\nin the code.\n<\/p>\n<p>\n  Documenting your changes well demonstrates some good things.  First,\nit's considerate to the person you are trying to persuade.  Second, it\nshows that you understand the ramifications of your change well enough\nto explain it to somebody who can't see the code.  Third, it demonstrates\nthat you care about the people who will ultimately use the software.\n<\/p>\n<p>\n  Good documentation is usually the most visible sign of what separates\na solid contribution from a quick and dirty hack.  If you take the time\nand care necessary to produce it, you'll find you're already 85% of the\nway to having your patch accepted by most developers.\n<\/p>\n<h5 id=\"id3008201\"><a href=\"#id3008201\">\u00a7<\/a>Do include an explanation with your patch.<\/h5>\n<p>\n  Your patch should include cover notes explaining why you think the\npatch is necessary or useful.  This is explanation directed not to the\nusers of the software but to the maintainer to whom you are sending the\npatch.\n<\/p>\n<p>\n  The note can be short \u2014 in fact, some of the most\neffective cover notes I've ever seen just said \u201cSee the\ndocumentation updates in this patch\u201d.  But it should show the\nright attitude.\n<\/p>\n<p>\n  The right attitude is helpful, respectful of the maintainer's time,\nquietly confident but unassuming.  It's good to display understanding of\nthe code you're patching.  It's good to show that you can identify with the\nmaintainer's problems. It's also good to be up front about any risks you\nperceive in applying the patch.  Here are some examples of the sorts of\nexplanatory comments that experienced developers send:\n<\/p>\n<p>\n  \u201cI've seen two problems with this code, X and Y.  I fixed\nproblem X, but I didn't try addressing problem Y because I don't think\nI understand the part of the code that I believe is\ninvolved\u201d.\n<\/p>\n<p>\n  \u201cFixed a core dump that can happen when one of the foo\ninputs is too long.  While I was at it, I went looking for similar\noverflows elsewhere.  I found a possible one in blarg.c, near line\n666.  Are you sure the sender can't generate more than 80 characters\nper transmission?\u201d\n<\/p>\n<p>\n  \u201cHave you considered using the Foonly algorithm for this\nproblem?  There is a good implementation at\n&lt;http:\/\/www.example.com\/~jsmith\/foonly.html&gt;\u201d.\n<\/p>\n<p>\n  \u201cThis patch solves the immediate problem, but I realize it\ncomplicates the memory allocation in an unpleasant way.  Works for me,\nbut you should probably test it under heavy load before\nshipping\u201d.\n<\/p>\n<p>\n  \u201cThis may be featuritis, but I'm sending it anyway.\nMaybe you'll know a cleaner way to implement the\nfeature\u201d.\n<\/p>\n<h5 id=\"id3008290\"><a href=\"#id3008290\">\u00a7<\/a>Do include useful comments in your code.<\/h5>\n<p>\n  A maintainer will want to have strong confidence that he\nunderstands your changes before merging them in.  This isn't an\ninvariable rule; if you have a track record of good work with the\nmaintainer, he may just run a casual eye over the changes before\nchecking them in semiautomatically.  But everything you can do to\nhelp him understand your code and decrease his uncertainty increases\nyour chances that your patch will be accepted.\n<\/p>\n<p>\n  Good comments in your code help the maintainer understand it.\nBad comments don't.\n<\/p>\n<p>\n  Here's an example of a bad comment:\n<\/p>\n <pre class=\"programlisting\">\n\/* norman newbie fixed this 13 Aug 2001 *\/\n<\/pre>\n<p>\n  This conveys no information.  It's nothing but a muddy\nterritorial bootprint you're planting in the middle of the\nmaintainer's code.  If he takes your patch (which you've made less\nlikely) he will almost certainly strip out this comment.  If you want\na credit, include a patch band for the project\n  <code>NEWS<\/code> or\n  <code>HISTORY<\/code> file.  He's\nmore likely to take that.\n<\/p>\n<p>\n  Here's an example of a good comment:\n<\/p>\n <pre class=\"programlisting\">\n\/*\n * This conditional needs to be guarded so that crunch_data() never\n * gets passed a NULL pointer.  &lt;norman_newbie@foosite.com&gt;\n *\/\n<\/pre>\n<p>\n  This comment shows that you understand not only the maintainer's\ncode but the kind of information that he needs to have confidence in\nyour changes.  This kind of comment\n  <em>gives<\/em>\n  him\nconfidence in your changes.\n<\/p>\n<h5 id=\"id3008383\"><a href=\"#id3008383\">\u00a7<\/a>Don't take it personally if your patch is rejected<\/h5>\n<p>\n  There are lots of reasons a patch can be rejected that don't\nreflect on you.  Remember that most maintainers are under heavy time\npressure, and have to be conservative in what they accept lest the\nproject code get broken.  Sometime resubmitting with improvements\nwill help.  Sometimes it won't.  Life is hard.\n<\/p>\n<h4 id=\"naming\"><a href=\"#naming\">\u00a7<\/a>Good Project- and Archive-Naming Practice<\/h4>\n<p>\n  As the load on maintainers of archives like ibiblio, SourceForge,\nand CPAN increases, there is an increasing trend for submissions\nto be processed partly or wholly by programs (rather than entirely by\na human).\n<\/p>\n<p>\n  This makes it more important for project and archive-file names\nto fit regular patterns that computer programs can parse and\nunderstand.\n<\/p>\n<h5 id=\"id3008426\"><a href=\"#id3008426\">\u00a7<\/a>Use GNU-style names with a stem and major.minor.patch numbering.<\/h5>\n<p>\n  It's helpful to everybody if your archive files all have GNU-like\nnames \u2014 all-lower-case alphanumeric stem prefix, followed by a hyphen,\nfollowed by a version number, extension, and other suffixes.\n<\/p>\n<p>\n  A good general form of name has these parts in order:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     project prefix\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     dash\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     version number\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     dot\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     \u201csrc\u201d or \u201cbin\u201d (optional)\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     dot or dash (dot preferred)\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     binary type and options (optional)\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     archiving and compression extensions\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n<p>\n  Name stems in this style can contain hyphen or underscores to\nseparate syllables; dashes are actually preferred.  It is good\npractice to group related projects by giving the stems a common \nhyphen-terminated prefix.\n<\/p>\n<p>\n  Let's suppose you have a project you call \u2018foobar\u2019\nat major version 1, minor version or release 2, patchlevel 3.  If it's\ngot just one archive part (presumably the sources), here's what its\nnames should look like like:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>foobar-1.2.3.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The source archive.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar.lsm<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The LSM file (assuming you're submitting to ibiblio).\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  Please\n  <em>don't<\/em>\n  use names like these:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>foobar123.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     This looks to many programs like an archive\nfor a project called \u201cfoobar123\u201d with no version number.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar1.2.3.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     This looks to many programs like an archive\nfor a project called \u201cfoobar1\u201d at version 2.3.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar-v1.2.3.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Many programs think this goes with a\nproject called \u201cfoobar-v1\u201d.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foo_bar-1.2.3.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     The underscore is hard for people to speak,\ntype, and remember.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>FooBar-1.2.3.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Unless you\n     <em>like<\/em>\n     looking like a\nmarketing weenie.  This is also hard for people to speak, type, and \nremember.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  If you have to differentiate between source and binary archives, or\nbetween different kinds of binary, or express some kind of build\noption in the file name, please treat that as a file extension to go\n  <em>after<\/em>\n  the version number. That is, please do this:\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     <code>foobar-1.2.3.src.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Sources.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar-1.2.3.bin.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     Binaries, type not specified.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar-1.2.3.bin.i386.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     i386 binaries.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar-1.2.3.bin.i386.static.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     i386 binaries statically linked.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     <code>foobar-1.2.3.bin.SPARC.tar.gz<\/code> \n   <\/dt>\n   <dd>\n    <p>\n     SPARC binaries.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  Please\n  <em>don't<\/em>\n  use names like\n\u2018foobar-i386-1.2.3.tar.gz\u2019, because programs have a hard\ntime telling type infixes (like \u2018-i386\u2019) from the\nstem.\n<\/p>\n<p>\n  The convention for distinguishing major from minor release is\nsimple: you increment the patch level for fixes or minor features, the \nminor version number for compatible new features, and the major\nversion number when you make incompatible changes.\n<\/p>\n<h5 id=\"local\"><a href=\"#local\">\u00a7<\/a>But respect local conventions where appropriate.<\/h5>\n<p>\n  Some projects and communities have well-defined conventions for\nnames and version numbers that aren't necessarily compatible with the\nabove advice.  For instance,\nApache\n  modules are\ngenerally named like mod_foo, and have both their own version number\nand the version of Apache with which they work.  Likewise,\nPerl\n  modules\nhave version numbers that can be treated as floating point numbers\n(e.g., you might see 1.303 rather than 1.3.3), and the distributions\nare generally named Foo-Bar-1.303.tar.gz for version 1.303 of module\nFoo::Bar. (Perl itself, on the other hand, switched to using the\nconventions described here in late 1999.)\n<\/p>\n<p>\n  Look for and respect the conventions of specialized\ncommunities and developers; for general use, follow the above\nguidelines.\n<\/p>\n<h5 id=\"prefix\"><a href=\"#prefix\">\u00a7<\/a>Try hard to choose a name prefix that is unique and easy to type.<\/h5>\n<p>\n  The stem prefix should be common to all of a project's files, and it\nshould be easy to read, type, and remember.  So please don't use\nunderscores.  And don't capitalize or BiCapitalize without extremely\ngood reason \u2014 it messes up the natural human-eyeball search order and\nlooks like some marketing weenie trying to be clever.\n<\/p>\n<p>\n  It confuses people when two different projects have the same stem name.  So try to check for collisions before your\n  first release. Two good places to check are the <a href=\"http:\/\/metalab.unc.edu\/pub\/Linux\"> index file of ibiblio<\/a>\n  and the application index at <a href=\"http:\/\/www.freshmeat.net\"> Freshmeat<\/a>.  Another good place to check is\n  <a href=\"http:\/\/www.sourceforge.net\"> SourceForge<\/a>; do a name search there.\n<\/p>\n<h4 id=\"develpractice\"><a href=\"#develpractice\">\u00a7<\/a>Good Development Practice<\/h4>\n<p>\n  Here are some of the behaviors that can make the difference \nbetween a successful project with lots of contributors and one that\nstalls out after attracting no interest:\n<\/p>\n<h5 id=\"proprietary\"><a href=\"#proprietary\">\u00a7<\/a>Don't rely on proprietary code.<\/h5>\n<p>\n  Don't rely on proprietary languages, libraries, or other code.\nDoing so is risky business at the best of times; in the open-source\ncommunity, it is considered downright rude. Open-source developers\ndon't trust code for which they can't review the source.\n<\/p>\n<h5 id=\"autoconf\"><a href=\"#autoconf\">\u00a7<\/a>Use GNU Autotools.<\/h5>\n<p>\n  Configuration choices should be made at compile time. A\nsignificant advantage of open-source distributions is that they allow\nthe package to adapt at compile-time to the environment it finds. This\nis critical because it allows the package to run on platforms its\ndevelopers have never seen, and it allows the software's community of\nusers to do their own ports. Only the largest of development teams can\nafford to buy all the hardware and hire enough employees to support\neven a limited number of platforms.\n<\/p>\n<p>\n  Therefore: Use the GNU autotools to handle portability issues,\ndo system-configuration probes, and tailor your makefiles.  People\nbuilding from sources today expect to be able to type\n  <code>configure; make; make install<\/code>\n  and get a clean build \u2014 and rightly so.  There is a\n  <a href=\"http:\/\/seul.org\/docs\/autotut\/\">\n   good tutorial on these tools <\/a>.\n<\/p>\n<p>\n  <em>autoconf<\/em> and\n  <em>autoheader<\/em> are mature.\n  <em>automake<\/em>, as we've previously noted, is\nstill buggy and brittle as of mid-2003; you may have to maintain\nyour own\n  <code>Makefile.in<\/code>.  Fortunately\nit's the least important of the autotools.\n<\/p>\n<p>\n  Regardless of your approach to configuration, do not ask the user\nfor system information at compile-time. The user installing the package\ndoes not know the answers to your questions, and this approach is doomed\nfrom the start. The software must be able to determine for itself any\ninformation that it may need at compile- or install-time.\n<\/p>\n<p>\n  But\n  <em>autoconf<\/em> should not be regarded\nas a license for knob-ridden designs.  If at all possible, program to\nstandards like POSIX and refrain also from asking the system for\nconfiguration information.  Keep ifdefs to a minimum \u2014 or,\nbetter yet, have none at all.\n<\/p>\n<h5 id=\"testcode\"><a href=\"#testcode\">\u00a7<\/a>Test your code before release.<\/h5>\n<p>\n  A good test suite allows the team to easily run regression tests\nbefore releases. Create a strong, usable test framework so that you\ncan incrementally add tests to your software without having to train\ndevelopers in the specialized intricacies of the test suite.\n<\/p>\n<p>\n  Distributing the test suite allows the community of users to test\ntheir ports before contributing them back to the group.\n<\/p>\n<p>\n  Encourage your developers to use a wide variety of platforms as\ntheir desktop and test machines, so that code is continuously being\ntested for portability flaws as part of normal development.\n<\/p>\n<p>\n  It is good practice, and encourages confidence in your code, when\nit ships with the test suite you use, and that test suite can be run\nwith\n  <code>make test<\/code>.\n<\/p>\n<h5 id=\"sanitycode\"><a href=\"#sanitycode\">\u00a7<\/a>Sanity-check your code before release.<\/h5>\n<p>\n  By \u201csanity check\u201d we mean: use every tool available\nthat has a reasonable chance of catching errors a human would be prone\nto overlook. The more of these you catch with tools, the fewer your\nusers and you will have to contend with.\n<\/p>\n<p>\n  If you're writing C\n  \/C++\n  using GCC, test-compile with -Wall and\nclean up all warning messages before each release. Compile your code\nwith every compiler you can find \u2014 different compilers often\nfind different problems. Specifically, compile your software on a true\n64-bit machine. Underlying datatypes can change on 64-bit machines,\nand you will often find new problems there. Find a Unix vendor's\nsystem and run the lint utility over your software.\n<\/p>\n<p>\n  Run tools that look for memory leaks and other runtime errors;\nElectric Fence and Valgrind are two good ones available in open source.\n<\/p>\n<p>\n  For Python\n  projects, the\n  <a href=\"http:\/\/sourceforge.net\/projects\/pychecker\">\n   PyChecker<\/a>\n  program can be a useful check.  It often catches nontrivial errors.\n<\/p>\n<p>\n  If you're writing Perl, check your code with perl\n  <code>-c<\/code> (and maybe\n  <code>-T<\/code>, if\napplicable). Use perl\n  <code>-w<\/code> and 'use strict'\nreligiously. (See the Perl documentation for further\ndiscussion.)\n<\/p>\n<h5 id=\"sanitydocs\"><a href=\"#sanitydocs\">\u00a7<\/a>Spell-check your documentation and READMEs before release.<\/h5>\n<p>\n  Spell-check your documentation, README files and error messages\nin your software. Sloppy code, code that produces warning messages when\ncompiled, and spelling errors in README files or error messages, all lead\nusers to believe the engineering behind it is also haphazard and\nsloppy.\n<\/p>\n<h5 id=\"cport\"><a href=\"#cport\">\u00a7<\/a>Recommended C\/C++ Portability Practices<\/h5>\n<p>\n  If you are writing C, feel free to use the full ANSI\nfeatures.  Specifically, do use function prototypes, which will help\nyou spot cross-module inconsistencies.  The old-style K&amp;R\ncompilers are ancient history.\n<\/p>\n<p>\n  Do not assume compiler-specific features such as the GCC\n  <code>-pipe<\/code> option or nested functions are available. These\nwill come around and bite you the second somebody ports to a\nnon-Linux, non-GCC system.\n<\/p>\n<p>\n  Code required for portability should be isolated to a single\narea and a single set of source files (for example, an\n  <code>os<\/code> subdirectory).  Compiler, library and\noperating system interfaces with portability issues should be\nabstracted to files in this directory.\n<\/p>\n<p>\n  A portability layer is a library (or perhaps just a set of\nmacros in header files) that abstracts away just the parts of an\noperating system's API your program is interested in.  Portability\nlayers make it easier to do new software ports. Often, no member of\nthe development team knows the porting platform (for example, there\nare literally hundreds of different embedded operating systems, and\nnobody knows any significant fraction of them). By creating a separate\nportability layer, it becomes possible for a specialist who knows a\nplatform to port your software without having to understand anything\noutside the portability layer.\n<\/p>\n<p>\n  Portability layers also simplify applications. Software rarely\nneeds the full functionality of more complex system calls such as\nmmap(2)\nor\nstat(2),\nand programmers commonly configure such complex interfaces\nincorrectly. A portability layer with abstracted interfaces\n(say, something named\n  <code>__file_exists<\/code> instead of a call to\nstat(2))\nallows you to import only the limited, necessary functionality from\nthe system, simplifying the code in your application.\n<\/p>\n<p>\n  Always write your portability layer to select based on a\nfeature, never based on a platform. Trying to create a separate\nportability layer for each supported platform results in a multiple\nupdate problem maintenance nightmare. A \u201cplatform\u201d is\nalways selected on at least two axes: the compiler and the\nlibrary\/operating system release. In some cases there are three axes,\nas when Linux vendors select a C\n  library independently of the operating\nsystem release. With\n  <em>M<\/em>\n  vendors,\n  <em>N<\/em>\n  compilers, and\n  <em>O<\/em>\n  operating system\nreleases, the number of platforms quickly scales out of reach of any\nbut the largest development teams.  On the other hand, by using\nlanguage and systems standards such as ANSI and POSIX 1003.1, the set\nof features is relatively\nconstrained.\n<\/p>\n<p>\n  Portability choices can be made along either lines of code or\ncompiled files. It doesn't make a difference if you select alternate\nlines of code on a platform, or one of a few different files. A rule\nof thumb is to move portability code for different platforms into\nseparate files when the implementations diverge significantly (shared\nmemory mapping on Unix vs. Windows), and leave portability code in a\nsingle file when the differences are minimal (for example, whether\nyou're using\n  <code>gettimeofday<\/code>,\n  <code>clock_gettime<\/code>,\n  <code>ftime<\/code> or\n  <code>time<\/code> to\nfind out the current time-of-day).\n<\/p>\n<p>\n  For anywhere outside a portability layer, heed this advice:\n<\/p>\n <blockquote>\n  <p>\n   <code>#ifdef<\/code> and\n   <code>#if<\/code> are last resorts, usually a sign of failure\nof imagination, excessive product differentiation, gratuitous\n\u201coptimization\u201d or accumulated trash. In the middle of\ncode they are anathema.\n   <code>\/usr\/include\/stdio.h<\/code> from GNU is an archetypical horror.\n  <\/p><div>\n  \n   \u2013 Doug McIlroy\n  \n <\/div><\/blockquote>\n<p>\n  Use of\n  <code>#ifdef<\/code> and\n  <code>#if<\/code> is permissible (if well controlled) within a\nportability layer. Outside it, try hard to confine these to\nconditionalizing\n  <code>#includes<\/code> based on feature\nsymbols.\n<\/p>\n<p>\n  Never intrude on the namespace of any other part of the system,\nincluding filenames, error return values and function names.  Where\nthe namespace is shared, document the portion of the namespace that you\nuse.\n<\/p>\n<p>\n  Choose a coding standard. The debate over the choice of standard\ncan go on forever \u2014 regardless, it is too difficult and\nexpensive to maintain software built using multiple coding standards,\nand so some common style must be chosen. Enforce your coding standard\nruthlessly, as consistency and cleanliness of the code are of the\nhighest priority; the details of the coding standard itself are a\ndistant second.\n<\/p>\n<h4 id=\"distpractice\"><a href=\"#distpractice\">\u00a7<\/a>Good Distribution-Making Practice<\/h4>\n<p>\n  These guidelines describe how your distribution should look when\nsomeone downloads, retrieves and unpacks it.\n<\/p>\n<h5 id=\"tarballs\"><a href=\"#tarballs\">\u00a7<\/a>Make sure tarballs always unpack into a single new directory.<\/h5>\n<p>\n  The single most annoying mistake fledgling contributors make is to build\ntarballs that unpack the files and directories in the distribution into\nthe current directory, potentially overwriting files already located there.\n  <em>Never do this!<\/em>\n<\/p>\n<p>\n  Instead, make sure your archive files all have a common directory part\nnamed after the project, so they will unpack into a single top-level \ndirectory directly\n  <em>beneath<\/em>\n  the current one.\nConventionally, the name of the directory should be the same as the\nstem of the tarball's name.  So, for example, a tarball named\n  <code>foo-0.23.tar.gz<\/code> is expected to unpack into\na subdirectory named\n  <code>foo-0.23<\/code>.\n<\/p>\n<p>\n   Example 19.1 shows a makefile trick that, assuming your distribution directory is named \u201cfoobar\u201d and SRC contains a\n   list of your distribution files, accomplishes this.\n<\/p>\n <div class=\"example\"><p class=\"title\">\n   <b>\n    Example 19.1.\n    <em>tar<\/em> archive maker production.\n   <\/b>\n  <\/p><pre class=\"programlisting\">\nfoobar-$(VERS).tar.gz:\n\t@ls $(SRC) | sed s:^:foobar-$(VERS)\/: &gt;MANIFEST\n\t@(cd ..; ln -s foobar foobar-$(VERS))\n\t(cd ..; tar -czvf foobar\/foobar-$(VERS).tar.gz `cat foobar\/MANIFEST`)\n\t@(cd ..; rm foobar-$(VERS))\n<\/pre><\/div>\n<h5 id=\"readme\"><a href=\"#readme\">\u00a7<\/a>Include a README.<\/h5>\n<p>\n  Include a file called\n  <code>README<\/code> that is a roadmap\nof your source distribution.  By ancient convention (originating with\nDennis Ritchie himself before 1980, and promulgated on\nUsenet\n  in the early\n1980s), this is the first file intrepid explorers will read after\nunpacking the source.\n<\/p>\n<p>\n  <code>README<\/code> files should be short and easy to\nread.  Make yours an introduction, not an epic. Good things to have in\nthe\n  <code>README<\/code> include the following:\n<\/p>\n <div class=\"orderedlist\"><ol type=\"1\"><li>\n    <p>\n     A brief description of the project.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     A pointer to the project website (if it has \none).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Notes on the developer's build environment and\npotential portability problems.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     A roadmap describing important files and subdirectories.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Either build\/installation instructions or a pointer to a file\ncontaining same (usually\n     <code>INSTALL<\/code>).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Either a maintainers\/credits list or a pointer to a\nfile containing same (usually\n     <code>CREDITS<\/code>).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Either recent project news or a pointer to a file\ncontaining same (usually\n     <code>NEWS<\/code>).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Project mailing list addresses.\n    <\/p>\n   <\/li>\n  <\/ol><\/div>\n<p>\n  At one time this file was commonly\n  <code>READ.ME<\/code>,\nbut this interacts badly with browsers, who are all too likely to\nassume that the\n  <code>.ME<\/code> suffix means it's not textual and can only be\ndownloaded rather than browsed.  This usage is deprecated.\n<\/p>\n<h5 id=\"filenames\"><a href=\"#filenames\">\u00a7<\/a>Respect and follow standard file-naming practices.<\/h5>\n<p>\n  Before even looking at the\n  <code>README<\/code>, your\nintrepid explorer will have scanned the filenames in the top-level\ndirectory of your unpacked distribution.  Those names can themselves\nconvey information.  By adhering to certain standard naming practices,\nyou can give the explorer valuable clues about where to look\nnext.\n<\/p>\n<p>\n  Here are some standard top-level file names and what they mean.  Not\nevery distribution needs all of these.\n<\/p>\n <div class=\"variablelist\"><dl><dt>\n    \n     README\n    \n   <\/dt>\n   <dd>\n    <p>\n     The roadmap file, to be read first.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     INSTALL\n    \n   <\/dt>\n   <dd>\n    <p>\n     Configuration, build, and installation instructions.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     AUTHORS\n    \n   <\/dt>\n   <dd>\n    <p>\n     List of project contributors (GNU convention).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     NEWS\n    \n   <\/dt>\n   <dd>\n    <p>\n     Recent project news.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     HISTORY\n    \n   <\/dt>\n   <dd>\n    <p>\n     Project history.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     CHANGES\n    \n   <\/dt>\n   <dd>\n    <p>\n     Log of significant changes between revisions.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     COPYING\n    \n   <\/dt>\n   <dd>\n    <p>\n     Project license terms (GNU convention).\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     LICENSE\n    \n   <\/dt>\n   <dd>\n    <p>\n     Project license terms.\n    <\/p>\n   <\/dd>\n   <dt>\n    \n     FAQ\n    \n   <\/dt>\n   <dd>\n    <p>\n     Plain-text Frequently-Asked-Questions document for\nthe project.\n    <\/p>\n   <\/dd>\n  <\/dl><\/div>\n<p>\n  Note the overall convention that filenames with all-caps names\nare human-readable metainformation about the package, rather than\nbuild components. This elaboration of the\n  <code>README<\/code> was developed early on at the Free Software\nFoundation.\n<\/p>\n<p>\n  Having a\n  <code>FAQ<\/code> file can save you a lot of\ngrief.  When a question about the project comes up often, put it in\nthe FAQ; then direct users to read the FAQ before sending questions or\nbug reports.  A well-nurtured FAQ can decrease the support burden on\nthe project maintainers by an order of magnitude or more.\n<\/p>\n<p>\n  Having a\n  <code>HISTORY<\/code> or\n  <code>NEWS<\/code> file with timestamps in it for each release\nis valuable.  Among other things, it may help establish prior art if\nyou are ever hit with a patent-infringement lawsuit (this hasn't\nhappened to anyone yet, but best to be prepared).\n<\/p>\n<h5 id=\"upgradability\"><a href=\"#upgradability\">\u00a7<\/a>Design for upgradability.<\/h5>\n<p>\n  Your software will change over time as you put out new releases.  Some\nof these changes will not be backward-compatible.  Accordingly, you\nshould give serious thought to designing your installation layouts so\nthat multiple installed versions of your code can coexist on the same\nsystem.  This is especially important for libraries \u2014 you can't \ncount on all your client programs to upgrade in lockstep with your\nAPI changes.\n<\/p>\n<p>\n  The Emacs, Python, and Qt projects have a good convention for\nhandling this: version-numbered directories (another practice that\nseems to have been made routine by the FSF).  Here's how an installed\nQt library hierarchy looks (\n  <code>${ver}<\/code> is the version\nnumber):\n<\/p>\n <pre class=\"screen\">\n\/usr\/lib\/qt\n\/usr\/lib\/qt-${ver}\n\/usr\/lib\/qt-${ver}\/bin          # Where you find moc\n\/usr\/lib\/qt-${ver}\/lib          # Where you find .so\n\/usr\/lib\/qt-${ver}\/include      # Where you find header files\n<\/pre>\n<p>\n  With this organization, multiple versions can coexist.  Client programs have to specify the library version they want,\n  but that's a small price to pay for not having the interfaces break on them.  This good practice avoids the notorious\n  <a href=\"#dll_hell\"> \u201cDLL Hell\u201d<\/a> failure mode of Windows.\n<\/p>\n<h5 id=\"rpms\"><a href=\"#rpms\">\u00a7<\/a>Under Linux, provide RPMs.<\/h5>\n<p>\n  The de facto standard format for installable binary packages\nunder Linux that used by the Red Hat Package manager, RPM.  It's\nfeatured in the most popular Linux\n  distribution, and supported by\neffectively all other Linux distributions (except Debian and\nSlackware; and Debian can install from RPMs). Accordingly, it's a good\nidea for your project site to provide installable RPMs as well as\nsource tarballs.\n<\/p>\n<p>\n  It's also a good idea for you to include in your source tarball\nthe RPM spec file, with a production that makes RPMs from it in your\n  <code>makefile<\/code>.  The spec file should have the\nextension\n  <code>.spec<\/code>; that's how the\n  <em>rpm<\/em> <code>-t<\/code> option finds it in a\ntarball.\n<\/p>\n<p>\n  For extra style points, generate your spec file with a\nshellscript that automatically plugs in the correct version number by\nanalyzing the project\n  <code>makefile<\/code> or a\n  <code>version.h<\/code>.\n<\/p>\n<p>\n  Note: If you supply source RPMs, use BuildRoot to make the\nprogram be built in\n  <code>\/tmp<\/code> or\n  <code>\/var\/tmp<\/code>.  If you don't, during the course of\nrunning the make install part of your build, the install will install\nthe files in the real final places. This will happen even if there are\nfile collisions, and even if you didn't want to install the package at\nall. When you're done, the files will have been installed and your\nsystem's RPM database will not know about it. Such badly behaved SRPMs\nare a minefield and should be eschewed.\n<\/p>\n<h5 id=\"checksums\"><a href=\"#checksums\">\u00a7<\/a>Provide checksums.<\/h5>\n<p>\n  Provide checksums with your binaries (tarballs, RPMs, etc.).  This\nwill allow people to verify that they haven't been corrupted or had\nTrojan-horse code inserted in them.\n<\/p>\n<p>\n  While there are several commands you can use for this purpose\n(such as <code>sum<\/code> and <code>cksum<\/code>) it is\nbest to use a cryptographically-secure hash function.  The GPG package\nprovides this capability via the\n  <code>--detach-sign<\/code> option; so does the GNU command <code>md5sum<\/code>.\n<\/p>\n<p>\n  For each binary you ship, your project Web page should list\nthe checksum and the command you used to generate it.\n<\/p>\n<h4 id=\"communication\"><a href=\"#communication\">\u00a7<\/a>Good Communication Practice<\/h4>\n<p>\n  Your software and documentation won't do the world much good if\nnobody but you knows they exist.  Also, developing a visible presence for\nthe project on the Internet will assist you in recruiting users and\nco-developers.  Here are the standard ways to do that.\n<\/p>\n<h5 id=\"announce\"><a href=\"#announce\">\u00a7<\/a>Announce to Freshmeat.<\/h5>\n<p>\n  Announce to\n  <a href=\"http:\/\/www.freshmeat.net\">\n   Freshmeat<\/a>.\nBesides being widely read itself, this group is a major feeder for \nWeb-based technical news channels.\n<\/p>\n<p>\n  Never assume the audience has been reading your release\nannouncements since the beginning of time.  Always include at least a\none-line description of what the software does.  Bad example:\n\u201cAnnouncing the latest release of FooEditor, now with themes\nand ten times faster\u201d.  Good example: \u201cAnnouncing the\nlatest release of FooEditor, the scriptable editor for touch-typists, \nnow with themes and ten times faster\u201d.\n<\/p>\n<h5 id=\"newsgroup\"><a href=\"#newsgroup\">\u00a7<\/a>Announce to a relevant topic newsgroup.<\/h5>\n<p>\n  Find a Usenet topic group directly relevant to your application,\nand announce there as well.  Post only where the\n  <em>function<\/em>\n  of the code is relevant, and exercise\nrestraint.\n<\/p>\n<p>\n  If (for example) you are releasing a program written in Perl\nthat queries IMAP servers, you should certainly post to\n  <code>comp.mail.imap<\/code>.  But you should probably\nnot post to\n  <code>comp.lang.perl<\/code> unless the program is also an instructive example of cutting-edge Perl\ntechniques.\n<\/p>\n<p>\n  Your announcement should include the URL of a project website.\n<\/p>\n<h5 id=\"website\"><a href=\"#website\">\u00a7<\/a>Have a website.<\/h5>\n<p>\n  If you intend trying to build any substantial user or developer community\naround your project, it should have a website.  Standard things to have\non the website include:\n<\/p>\n <div><ul><li>\n    <p>\n     The project charter (why it exists, who the audience is, etc.).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Download links for the project sources.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Instructions on how to join the project mailing list(s).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     A FAQ (Frequently Asked Questions) list.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     HTMLized versions of the project documentation.\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     Links to related and\/or competing projects.\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n<p>\n  Refer to the website examples in\n  <a href=\"#reusechapter\" title=\"Chapter 16. Reuse\">\n   Chapter 16<\/a>\n  for examples of what a well-educated project\nwebsite looks like.\n<\/p>\n<p>\n  An easy way to have a website is to put your project on one of\nthe sites that specializes in providing free hosting. In 2003\nthe two most important of these are SourceForge (which is a\ndemonstration and test site for proprietary collaboration tools) or\nSavannah (which hosts open-source projects as an ideological\nstatement).\n<\/p>\n<h5 id=\"lists\"><a href=\"#lists\">\u00a7<\/a>Host project mailing lists.<\/h5>\n<p>\n  It's standard practice to have a private development list\nthrough which project collaborators can communicate and exchange\npatches.  You may also want to have an announcements list for people\nwho want to be kept informed of the project's progress.\n<\/p>\n<p>\n  If you are running a project named \u2018foo\u2019, your\ndeveloper list might be\n  <code>\n   &lt;\n   <a href=\"mailto:foo-dev\">\n    foo-dev<\/a>\n   &gt;\n  <\/code>\n  or\n  <code>\n   &lt;\n   <a href=\"mailto:foo-friends\">\n    foo-friends<\/a>\n   &gt;\n  <\/code>; your announcement list might be\n  <code>\n   &lt;\n   <a href=\"mailto:foo-announce\">\n    foo-announce<\/a>\n   &gt;\n  <\/code>.\n<\/p>\n<p>\n  An important decision is just how private the\n\u201cprivate\u201d development list is.  Wider participation in\ndesign discussions is often a good thing, but if the list is\nrelatively open, sooner or later you\n  <em>will<\/em>\n  get\npeople asking new-user questions on it.  Opinions vary on how best to\nsolve this problem.  Just having the documentation tell the new users\nnot to ask elementary questions on the development list is not a\nsolution; such a request must be enforced somehow.\n<\/p>\n<p>\n  An announcements list needs to be tightly controlled.  Traffic should be\nat most a few messages a month; the whole purpose of such a list is to\naccommodate people who want to know when something important happens, but\ndon't want to hear about day-to-day details.  Most such people will\nquickly unsubscribe if the list starts generating significant clutter in\ntheir mailboxes.\n<\/p>\n<p>\n  See the section\n  <em>Where Should I Look?<\/em> in\n  <a href=\"#reusechapter\" title=\"Chapter 16. Reuse\">\n   Chapter 16<\/a>\n  for specifics on\nthe major open-source archive sites.  You should release your package\nto these.\n<\/p>\n<p>\n  Other important locations include:\n<\/p>\n <div><ul><li>\n    <p>\n     The\n     <a href=\"http:\/\/www.python.org\">\n      Python Software Activity<\/a>\n     site (for software written in Python).\n    <\/p>\n   <\/li>\n   <li>\n    <p>\n     The\n     <a href=\"http:\/\/language.perl.com\/CPAN\">\n      CPAN<\/a>, the\nComprehensive Perl Archive Network (for software written in Perl).\n    <\/p>\n   <\/li>\n  <\/ul><\/div>\n\n<h3 id=\"ch19s03\"><a href=\"#ch19s03\">\u00a7<\/a>The Logic of Licenses: How to Pick One<\/h3>\n<p>\n  The choice of license terms involves decisions about what, if\nany restrictions the author wants to put on what people do with the\nsoftware.\n<\/p>\n<p>\n  If you want to make no restrictions at all, you should put your\nsoftware in the public domain. An appropriate way to do this would\nbe to include something like the following text at the head of each\nfile:\n<\/p>\n <pre class=\"programlisting\">\nPlaced in public domain by J. Random Hacker, 2003.  Share and enjoy!\n<\/pre>\n<p>\n  If you do this, you are surrendering your copyright. Anyone can\ndo anything they like with any part of the text. It doesn't get any\nfreer than this.\n<\/p>\n<p>\n  But very little open-source software is actually placed in the\npublic domain.  Some open-source developers want to use\ntheir ownership of the code to ensure that it stays open (these tend\nto adopt the GPL).  Others simply want to control their legal exposure;\none of the things\n  <em>all<\/em>\n  open-source licenses have\nin common is a disclaimer of warranty.\n<\/p>\n\n<h3 id=\"ch19s04\"><a href=\"#ch19s04\">\u00a7<\/a>Why You Should Use a Standard License<\/h3>\n<p>\n  The widely known licenses conforming to the Open Source\nDefinition have well-established interpretive traditions. Developers\n(and, to the extent they care, users) know what they imply, and have a\nreasonable take on the risks and tradeoffs they involve. Therefore,\nuse one of the standard licenses carried on the OSI site if at all\npossible.\n<\/p>\n<p>\n  If you must write your own license, be sure to have it certified\nby OSI. This will avoid a lot of argument and overhead. Unless you've\nbeen through it, you have no idea how nasty a licensing flamewar can\nget; people become passionate because the licenses are regarded as\nalmost-sacred covenants touching the core values of the open-source\ncommunity.\n<\/p>\n<p>\n  Furthermore, the presence of an established interpretive\ntradition may prove important if your license is ever tested in\ncourt. At time of writing (mid-2003) there is no case law either\nsupporting or invalidating any open-source license. However, it is\na legal doctrine (at least in the United States, and probably in other\ncommon-law countries such as England and the rest of the British\nCommonwealth) that courts are supposed to interpret licenses and\ncontracts according to the expectations and practices of the\ncommunity in which they originated. There is thus good reason to\nhope that open-source community practice will be determinative when\nthe court system finally has to cope.\n<\/p>\n\n<h3 id=\"ch19s05\"><a href=\"#ch19s05\">\u00a7<\/a>Varieties of Open-Source Licensing<\/h3>\n<h4 id=\"id3014860\"><a href=\"#id3014860\">\u00a7<\/a>MIT or X Consortium License<\/h4>\n<p>\n  The loosest kind of free-software license is one that grants\nunrestricted rights to copy, use, modify, and redistribute modified\ncopies as long as a copy of the copyright and license terms is\nretained in all modified versions.  But when you accept this\nlicense you do give up the right to sue the maintainers.\n<\/p>\n<p>\n  You can find a template for the standard X Consortium license at\nthe\n  <a href=\"http:\/\/www.opensource.org\/licenses\/mit-license.html\">\n   OSI\nsite<\/a>.\n<\/p>\n<h4 id=\"id3014890\"><a href=\"#id3014890\">\u00a7<\/a>BSD Classic License<\/h4>\n<p>\n  The next least restrictive kind of license grants unrestricted\nrights to copy, use, modify, and redistribute modified copies as long\nas a copy of the copyright and license terms is retained in all\nmodified versions, and an acknowledgment is made in advertising or\ndocumentation associated with the package.  Grantee has to give up\nthe right to sue the maintainers.\n<\/p>\n<p>\n  The original BSD license is the best-known license of this kind.\nAmong parts of the free-software culture that trace their lineages\nback to BSD Unix,\nthis license is used even on a lot of free software that was written\nthousands of miles from\nBerkeley.\n<\/p>\n<p>\n  It is also not uncommon to find minor variants of the BSD\nlicense that change the copyright holder and omit the advertising\nrequirement (making it effectively equivalent to the MIT license).\nNote that in mid-1999 the Office of Technology Transfer of the\nUniversity of California rescinded the advertising clause in the BSD\nlicense. So the license on the BSD software has been relaxed in\nexactly this way.  Should you choose the BSD approach, we strongly\nrecommend that you use the new license (without advertising clause)\nrather than the old.  That requirement was dropped because it led to\nsignificant legal and procedural complications over what constituted\nadvertising.\n<\/p>\n<p>\n  You can find a BSD license template at the\n  <a href=\"http:\/\/www.opensource.org\/licenses\/bsd-license.html\">\n   OSI\nsite<\/a>.\n<\/p>\n<h4 id=\"id3014963\"><a href=\"#id3014963\">\u00a7<\/a>Artistic License<\/h4>\n<p>\n  The next most restrictive kind of license grants unrestricted\nrights to copy, use, and locally modify. It allows redistribution of\nmodified binaries, but restricts redistribution of modified sources in\nways intended to protect the interests of the authors and the\nfree-software community.\n<\/p>\n<p>\n  The Artistic License, devised for Perl\n  and widely used in the Perl developer\ncommunity, is of this kind. It requires modified files to contain\n\u201cprominent notice\u201d that they have been altered. It also\nrequires people who redistribute changes to make them freely available\nand make efforts to propagate them back to the free-software\ncommunity.\n<\/p>\n<p>\n  You can find a copy of the Artistic License at the\n  <a href=\"http:\/\/www.opensource.org\/licenses\/artistic-license.html\">\n   OSI\nsite<\/a>.\n<\/p>\n<h4 id=\"id3015011\"><a href=\"#id3015011\">\u00a7<\/a>General Public License<\/h4>\n<p>\n  The GNU General Public License (and its derivative, the Library\nor \u201cLesser\u201d GPL) is the single most widely used\nfree-software license. Like the Artistic License, it allows\nredistribution of modified sources provided the modified files bear\n\u201cprominent notice\u201d.\n<\/p>\n<p>\n  The GPL requires that any program containing parts that are under\nGPL be wholly GPLed. (The exact circumstances that trigger this\nrequirement are not perfectly clear to everybody.)\n<\/p>\n<p>\n  These extra requirements actually make the GPL more restrictive\nthan any of the other commonly used licenses. (Larry Wall developed\nthe Artistic License to avoid them while serving many of the same\nobjectives.)\n<\/p>\n<p>\n  You can find a pointer to the GPL, and instructions about how to\napply it, at\n  <a href=\"http:\/\/www.gnu.org\/copyleft.html\">\n   FSF\ncopyleft site<\/a>.\n<\/p>\n<h4 id=\"id3015063\"><a href=\"#id3015063\">\u00a7<\/a>Mozilla Public License<\/h4>\n<p>\n  The Mozilla Public License supports software that is open\nsource, but may be linked with closed-source modules or extensions. It\nrequires that the distributed software (\"Covered Code\")\nremain open, but permits add-ons called through a defined API to\nremain closed.\n<\/p>\n<p>\n  You can find a template for the MPL at the\n  <a href=\"http:\/\/www.opensource.org\/licenses\/MPL-1.1.html\">\n   OSI\nsite<\/a>.\n<\/p>\n\n<h3 id=\"futurechapter\"><a href=\"#futurechapter\">\u00a7<\/a>Chapter 20. Futures<\/h3>\n <blockquote class=\"epigraph\">\n  <p>\n   The best way to predict the future is to invent it.\n  <\/p><div>\n  \n   \u2013 Alan Kay\nUttered during a 1971 meeting at XEROX PARC\n  \n <\/div><\/blockquote>\n<p>\n  History is not over.  Unix will continue to grow and change.\nThe community and the tradition around Unix will continue to evolve.\nTrying to forecast the future is a chancy business, but we can perhaps\nanticipate it in two ways: first, by looking at how Unix has coped with\ndesign challenges in the past; second, by identifying problems that are\nlooking for solutions and opportunities waiting to be\nexploited.\n<\/p>\n\n<h3 id=\"ch20s01\"><a href=\"#ch20s01\">\u00a7<\/a>Essence and Accident in Unix Tradition<\/h3>\n<p>\n  To understand how Unix's design might change in the future, we can\nstart by looking at how Unix programming style has changed over time\nin the past.  This effort leads us directly to one of the challenges\nof understanding the Unix style \u2014 distinguishing between\naccident and essence. That is, recognizing traits that arise from\ntransient technical circumstances versus those that are deeply tied to\nthe central Unix design challenge \u2014 how to do modularity and\nabstraction right while also keeping systems transparent and\nsimple.\n<\/p>\n<p>\n  This distinction can be difficult, because traits that arose as\naccidents have sometimes turned out to have essential utility.\nConsider as an example the \u2018Silence is golden\u2019 rule of\nUnix interface design we examined in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>; it began as an adaptation to slow\nteletypes, but continued because programs with terse output could be\ncombined in scripts more easily.  Today, in an environment where\nhaving many programs running visibly through a GUI is normal, it has a\nthird kind of utility: silent programs don't distract or waste the\nuser's attention.\n<\/p>\n<p>\n  On the other hand, some traits that once seemed essential to\nUnix turned out to be accidents tied to a particular set of cost\nratios.  For example, old-school Unix favored program designs (and\nminilanguages like awk(1)) that did line-at-a-time processing of an\ninput stream or record-at-a-time processing of binary files, with any\ncontext that needed to be maintained between pieces carried by\nelaborate state-machine code.  New-school Unix design, on the other\nhand, is generally happy with the assumption that a program can read\nits entire input into memory and thereafter randomly access it at will.\nIndeed, modern Unixes supply an mmap(2) call that allows the\nprogrammer to map an entire file into virtual memory and completely\nhides the serialization of I\/O to and from disk space.\n<\/p>\n<p>\n  This change trades away storage economy to get simpler and more\ntransparent code.  It's an adaptation to the plunging cost of memory\nrelative to programmer time.  Many of the differences between\nold-school Unix designs in the 1970s and 1980s and those of the new\npost-1990 school can be traced to the huge shift in relative costs\nthat today makes all machine resources several orders of magnitude\ncheaper relative to programmer time than they were in 1969.\n<\/p>\n<p>\n  Looking back, we can identify three specific technology changes\nthat have driven significant changes in Unix design style:\ninternetworking, bitmapped graphics displays, and the personal\ncomputer.  In each case, the Unix tradition has adapted to the\nchallenge by discarding accidents that were no longer adaptive and\nfinding new applications for its essential ideas.  Biological\nevolution works this way too.  Evolutionary biologists have a rule:\n\u201cDon't assume that historical origin specifies current utility,\nor vice versa\u201d. A brief look at how Unix adapted in each of\nthese cases may provide some clues to how Unix might adapt itself to\nfuture technology shifts that we cannot yet anticipate.\n<\/p>\n<p>\n  <a href=\"#historychapter\" title=\"Chapter 2. History\">\n   Chapter 2<\/a>\n  described\nthe first of these changes: the rise of internetworking, from the\nangle of cultural history, telling how\nTCP\/IP\n  brought the\noriginal Unix and ARPANET cultures together after 1980. In\n  <a href=\"#multiprogramchapter\" title=\"Chapter 7. Multiprogramming\">\n   Chapter 7<\/a>, the material on obsolescent IPC and\nnetworking methods such as System V\n  STREAMS hints at the many false starts,\nmissteps, and dead ends that preoccupied Unix developers through much\nof the following decade.  There was a good deal of confusion about\nprotocols,<sup><a href=\"#ftn.id3017690\" id=\"id3017690\">[153]<\/a><\/sup>\n  and about the\nrelationship between intermachine networking and interprocess\ncommunication among processes on the same machine.\n<\/p>\n<p>\n  Eventually the confusion was cleared up when TCP\/IP won and BSD\nsockets\n  reasserted Unix's essential\neverything-is-a-byte-stream\n  metaphor.  It became normal to use BSD\nsockets for both IPC and networking, older methods for both largely\nfell out of use, and Unix software grew increasingly indifferent to\nwhether communicating components were hosted on the same or different\nmachines.  The invention of the World Wide Web in 1990-1991 was the\nlogical result.\n<\/p>\n<p>\n  When bitmapped graphics and the example of the Macintosh\narrived in 1984 a few years after TCP\/IP, they posed a rather more\ndifficult challenge.  The original GUIs from Xerox\nPARC\n  and Apple\nwere beautiful, but wired together far too many levels of the system\nfor Unix programmers to feel comfortable with their design.  The\nprompt response of Unix programmers was to make separation of policy\nfrom mechanism an explicit principle; the X windowing system\nestablished it by 1988.  By splitting X widget sets away from the\ndisplay manager that was doing low-level graphics, they created an\narchitecture that was modular and clean in Unix terms, and one that\ncould easily evolve better policy over time.\n<\/p>\n<p>\n  But that was the easy part of the problem.  The hard part was\ndeciding whether Unix ought to have a unified interface policy at all,\nand if so what it ought to be.  Several different attempts to\nestablish one through proprietary toolkits (like Motif) failed.  Today,\nin 2003, GTK and Qt contend with each other for the role.  While the\ndebate on this question is not over in 2003, the persistence of\ndifferent UI styles that we noted in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>\n  seems telling. New-school Unix design has\nkept the command line, and dealt with the tension between GUI and CLI\napproaches by writing lots of CLI-engine\/GUI-interface pairs that can\nbe used in both styles.\n<\/p>\n<p>\n  The personal computer posed few major design challenges as a\ntechnology in itself.  The 386 and later chips were powerful enough to\ngive the systems designed around them cost ratios similar to those of\nthe minicomputers, workstations, and servers on which Unix matured.\nThe true challenge was a change in the potential market for Unix; the\nmuch lower overall price of the hardware made personal computers\nattractive to a vastly broader, less technically sophisticated user\npopulation.\n<\/p>\n<p>\n  The proprietary-Unix vendors, accustomed to the fatter margins from\nselling more powerful systems to sophisticated buyers, were never\ninterested in this wider market.  The first serious initiatives\ntoward the end-user desktop came out of the open-source community and\nwere mounted for essentially ideological reasons.  As of mid-2003,\nmarket surveys indicate that\nLinux\n  has reached about\n4%\u20135% share there, closely comparable to the Apple Macintosh's.\n<\/p>\n<p>\n  Whether or not Linux ever does substantially better than this,\nthe nature of the Unix community's response is already clear. We\nexamined it in the study of Linux in\n  <a href=\"#contrastchapter\" title=\"Chapter 3. Contrasts\">\n   Chapter 3<\/a>.  It includes adopting a few technologies\n(such as XML) from elsewhere, and putting a lot of effort into\nnaturalizing GUIs into the Unix world.  But underneath the themed GUIs\nand the installation packaging, the main emphasis is still on\nmodularity and clean code \u2014 on getting the infrastructure for\nserious, high-reliability computing and communications right.\n<\/p>\n<p>\n  The history of the large-scale desktop-focused developments like\nMozilla and OpenOffice.org that were launched in the late 1990s\nillustrates this emphasis well.  In both these cases, the most\nimportant theme in community feedback wasn't demand for new features\nor pressure to make a ship date \u2014 it was distaste for monster\nmonoliths, and a general sense that these huge programs would have to\nbe slimmed down, refactored, and carved into modules before they would\nbe other than embarrassments.\n<\/p>\n<p>\n  Despite being accompanied by a great deal of innovation, the\nresponses to all three technologies were conservative with regard to\nthe fundamental Unix design rules \u2014 modularity,\ntransparency, separation of policy from mechanism, and the other qualities\nwe've tried to characterize earlier in this book.  The learned\nresponse of Unix programmers, reinforced over thirty years, was to go\nback to first principles \u2014 to try to get more leverage out of\nUnix's basic abstractions of streams, namespaces, and processes in\npreference to layering on new ones.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3017690\" id=\"ftn.id3017690\">[153]<\/a><\/sup>\n  For a few years it looked like ISO's 7-layer\nnetworking standard might compete successfully with TCP\/IP.  It was\npromoted by a European standards committee politically horrified at\nthe thought of adopting any technology birthed in the bowels of the\nPentagon.  Alas, their indignation exceeded their technical acuity.\nThe result proved overcomplicated and unhelpful; see <sup><a href=\"#Padlipsky\" title=\"[Padlipsky]\">[Padlipsky]<\/a><\/sup> for details.\n<\/p><\/div>\n\n<h3 id=\"plan9\"><a href=\"#plan9\">\u00a7<\/a>Plan 9: The Way the Future Was<\/h3>\n<p>\n  We know what Unix's future used to look like.  It was designed\nby the research group at Bell Labs that built Unix and called\n\u2018Plan 9 from Bell Labs\u2019.<sup><a href=\"#ftn.id3015199\" id=\"id3015199\">[154]<\/a><\/sup>\n  Plan 9 was an attempt to do Unix over\nagain, better.\n<\/p>\n<p>\n  The central design challenge the designers attempted to meet in\nPlan 9 was integrating graphics and ubiquitous networking into a\ncomfortable Unix-like framework. They kept the Unix choice to mediate\naccess to as many system services as possible through a single big\nfile-hierarchy name space.  In fact, they improved on it; many\nfacilities that under Unix are accessed through various ad-hoc\ninterfaces like BSD sockets,\nfcntl(2),\nand\nioctl(2)\nare in Plan 9 accessed through ordinary read and write operations on\nspecial files analogous to device files.  For portability and ease of\naccess, almost all device interfaces are textual rather than binary.\nMost system services (including, for example, the window system) are\n  <em>file servers<\/em> containing special files or\ndirectory trees representing the served resources.  By representing\nall resources as files, Plan 9 turns the problem of accessing\nresources on different servers into the problem of accessing files on\ndifferent servers.\n<\/p>\n<p>\n  Plan 9 combined this more-Unix-than-Unix file model with a new\nconcept: private name spaces.  Every user (in fact, every\nprocess) can have its own view of the system's services by creating\nits own tree of file-server mounts.  Some of the file server mounts\nwill have been manually set up by the user, and others automatically\nset up at login time.  So (as the\n  <em>Plan 9 from Bell Labs<\/em>\n  survey paper points out)\n\u201c<code>\/dev\/cons<\/code>\n  always refers to your terminal\ndevice and <code>\/bin\/date<\/code>\n  to the correct version of\nthe date command to run, but which files those names represent depends\non circumstances such as the architecture of the machine executing\n  <code>date<\/code>\u201d.\n<\/p>\n<p>\n  The single most important feature of Plan 9 is that all mounted\nfile servers export the same file-system-like interface, regardless of\nthe implementation behind them.  Some might correspond to local file\nsystems, some to remote file systems accessed over a network, some to\ninstances of system servers running in user space (like the window\nsystem or an alternate network stack), and some to kernel interfaces.\nTo users and client programs, all these cases look alike.\n<\/p>\n<p>\n  One of the examples from the Plan 9 survey paper is the way FTP\naccess to remote sites is implemented.  There is no\nftp(1)\ncommand under Plan 9.  Instead there is an\n  <em>ftpfs<\/em> fileserver, and each FTP connection\nlooks like a file system mount.\n  <em>ftpfs<\/em> automatically translates open, read, and write commands on files and\ndirectories under the mount point into FTP protocol transactions.\nThus, all ordinary file-handling tools such as\nls(1),\nmv(1)\nand\ncp(1)\nsimply work, both underneath the FTP mount point and across the\nboundaries with the rest of the user's view of the namespace.  The only\ndifference the user (or his scripts and programs) will notice is\nretrieval speed.\n<\/p>\n<p>\n  Plan 9 has much else to recommend it, including the reinvention\nof some of the more problematic areas of the Unix system-call\ninterface, the elimination of superuser, and many other interesting\nrethinkings. Its pedigree is impeccable, its design elegant, and it\nexposes some significant errors in the design of Unix.  Unlike most\nefforts at a second system, it produced an architecture that was in\nmany ways simpler and more elegant than its predecessor. Why didn't it\ntake over the world?\n<\/p>\n<p>\n  One could argue for a lot of specific reasons \u2014 lack of\nany serious effort to market it, scanty documentation, much confusion\nand stumbling over fees and licensing.  For those unfamiliar with Plan\n9, it seemed to function mainly as a device for generating interesting\npapers on operating-systems research.  But Unix itself had previously\nsurmounted all these sorts of obstacles to attract a dedicated\nfollowing that spread it worldwide.  Why didn't Plan 9?\n<\/p>\n<p>\n  The long view of history may tell a different story, but in\n2003 it looks like Plan 9 failed simply because it fell short of\nbeing a compelling enough improvement on Unix to displace its\nancestor.  Compared to Plan 9, Unix creaks and clanks and has obvious\nrust spots, but it gets the job done well enough to hold its position.\nThere is a lesson here for ambitious system architects: the most\ndangerous enemy of a better solution is an existing codebase that is\njust good enough.\n<\/p>\n<p>\n  Some Plan 9 ideas have been absorbed into modern Unixes,\nparticularly the more innovative open-source versions.  FreeBSD has a\n  <code>\/proc<\/code> file system modeled exactly on\nthat of Plan 9 that can be used to query or control running processes.\nFreeBSD's\nrfork(2)\nand Linux's\nclone(2)\nsystem calls are modeled on Plan 9's\nrfork(2).\nLinux's\n  <code>\/proc<\/code> file system, in addition\nto presenting process information, holds a variety of synthesized Plan\n9-like device files used to query and control kernel internals using\npredominantly textual interfaces.  Experimental 2003 versions of\nLinux are implementing per-process mount points, a long step toward\nPlan 9's private namespaces.  The various open-source Unixes are all\nmoving toward systemwide support for UTF-8, an encoding actually\ninvented for Plan 9.<sup><a href=\"#ftn.id3015486\" id=\"id3015486\">[155]<\/a><\/sup><\/p>\n<p>\n  It may well be that over time, much more of Plan 9 will work its\nway into Unix as various portions of Unix's architecture slide into\nsenescence.  This is one possible line of development for Unix's \nfuture.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3015199\" id=\"ftn.id3015199\">[154]<\/a><\/sup>\n  The name is a\ntribute to the 1958 movie that has passed into legend as \u201cthe\nworst ever made\u201d,\n  <em>Plan 9 from Outer Space<\/em>.  (The legend is, unfortunately, incorrect, as the\nfew who have seen an even worse stinkeroo from 1966 called\n  <em>Manos: The Hands of Fate<\/em> can\nattest.) Documentation, including a survey paper\ndescribing the architecture, along with complete source code and a\ndistribution that installs on PCs, can be readily found with a Web\nsearch for the phrase \u2018Plan 9 from Bell\nLabs\u2019.\n<\/p><p><sup><a href=\"#id3015486\" id=\"ftn.id3015486\">[155]<\/a><\/sup>\n  The tale of how UTF-8 was born\ninvolves\n  <a href=\"http:\/\/www.cl.cam.ac.uk\/~mgk25\/ucs\/utf-8-history.txt\">\n   Ken\nThompson, Rob Pike, a new Jersey diner, and a frenzied overnight\nhack<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ch20s03\"><a href=\"#ch20s03\">\u00a7<\/a>Problems in the Design of Unix<\/h3>\n<p>\n  Plan 9 cleans up Unix, but only really adds one new concept\n(private namespaces) to its basic set of design ideas.  But are there\nserious problems with those basic design ideas? In\n  <a href=\"#philosophychapter\" title=\"Chapter 1. Philosophy\">Chapter 1<\/a>\n  we touched on\nseveral issues that Unix arguably got wrong.  Now that the open-source\nmovement has put the design future of Unix back in the hands of\nprogrammers and technical people, these are no longer decisions we\nhave to live with forever.  We'll reexamine them in order to get a\nbetter handle on how Unix might evolve in the future.\n<\/p>\n<h4 id=\"id3015538\"><a href=\"#id3015538\">\u00a7<\/a>A Unix File Is Just a Big Bag of Bytes<\/h4>\n<p>\n  A Unix file is just a big bag of bytes, with no other \nattributes.  In particular, there is no capability to store\ninformation about the file type or a pointer to an associated\napplication program outside the file's actual data.\n<\/p>\n<p>\n  More generally, everything is a byte stream; even hardware devices are byte streams.\nThis metaphor was a tremendous success of early Unix, and a real\nadvance over a world in which (for example) compiled programs could\nnot produce output that could be fed back to the compiler.\nPipes\n  and shell\nprogramming\n  sprang from this metaphor.\n<\/p>\n<p>\n  But Unix's byte-stream metaphor is\n  <em>so<\/em>\n  central that Unix has trouble integrating software objects with\noperations that don't fit neatly into the byte stream or file\nrepertoire of operations (create, open, read, write, delete). This is\nespecially a problem for GUI objects such as icons, windows, and\n\u2018live\u2019 documents.  Within a classical Unix model of the\nworld, the only way to extend the everything-is-a-byte-stream metaphor\nis through\n  <code>ioctl<\/code> calls, a notoriously ugly\ncollection of back doors into kernel space.\n<\/p>\n<p>\n  Fans of the Macintosh\n  family of operating systems tend to be\nvociferous about this.  They advocate a model in which a single\nfilename may have both data and resource \u2018forks\u2019, the data\nfork corresponding to the Unix byte stream and the resource fork being\na collection of name\/value pairs. Unix partisans prefer approaches\nthat make file data self-describing so that effectively the same sort\nof metadata is stored within the file.\n<\/p>\n<p>\n  The problem with the Unix approach is that every program\nthat writes the file has to know about it.  Thus, for example, if we\nwant the file to carry type information inside it, every tool that\ntouches it has to take care to either preserve the type field\nunaltered or interpret and then rewrite it.  While this would be\ntheoretically possible to arrange, in practice it would be far too\nfragile.\n<\/p>\n<p>\n  On the other hand, supporting file attributes raises awkward\nquestions about which file operations should preserve them.  It's\nclear that a copy of a named file to another name should copy\nthe source file's attributes as well as its data \u2014 but suppose\nwe\ncat(1)\nthe file, redirecting the output of\ncat(1)\nto a new name?\n<\/p>\n<p>\n  The answer to this question depends on whether the attributes\nare actually properties of filenames or are in some magical way\nbundled with the file's data as a sort of invisible preamble or\npostamble.  Then the question becomes: Which operations make the\nproperties visible?\n<\/p>\n<p>\n  Xerox PARC\n  file-system designs grappled with this problem as far back as the\n1970s. They had an \u2018open serialized\u2019 call that returned a\nbyte stream containing both attributes and content.  If applied to a\ndirectory, it returned a serialization of the directory's attributes\nplus the serialization of all the files in it.  It is not clear that\nthis approach has ever been bettered.\n<\/p>\n<p>\n  Linux 2.5\n  already supports attaching arbitrary\nname\/value pairs as properties of a filename, but at time of writing\n this capability is not yet much used by\napplications.  Recent versions of Solaris have a roughly equivalent\nfeature.\n<\/p>\n<h4 id=\"id3015724\"><a href=\"#id3015724\">\u00a7<\/a>Unix Support for GUIs Is Weak<\/h4>\n<p>\n  The Unix experience proves that using a handful of metaphors as\nthe basis for a framework is a powerful strategy (recall the\ndiscussion of frameworks and shared context in\n  <a href=\"#complexitychapter\" title=\"Chapter 13. Complexity\">Chapter 13<\/a>).  The visual metaphor at the heart of\nmodern GUIs (files represented by icons, and opened by clicking which\ninvokes some designated handler program, typically able to create and\nedit these files) has proven both successful and long-lived, exerting\na strong hold on users and interface designers ever since Xerox\nPARC pioneered it in the 1970s.\n<\/p>\n<p>\n  Despite considerable recent effort, in 2003 Unix still\nsupports this metaphor only poorly and grudgingly \u2014 there are\nlots of layers, few conventions, and only weak construction utilities.\nA typical reaction from a Unix old hand is to suspect that this\nreflects deeper problems with the GUI metaphor itself.\n<\/p>\n <blockquote>\n  <p>\n   I think part of the problem is that we still don't have the\nmetaphor right.  For example, on the Mac I drag a file to the\ntrashcan to delete it, but when I drag it to a disc it gets copied,\nand can't drag it to a printer icon to print it because that's done\nthrough the menus.  I could go on and on.  It's like files were in\nOS\/360, before Unix came along with its simple (but not too simple),\nfile idea.\n  <\/p><div>\n  \n   \u2013 Steve Johnson\n  \n <\/div><\/blockquote>\n<p>\n  We quoted Brian Kernighan and Mike Lesk to similar effect in\n  <a href=\"#interfacechapter\" title=\"Chapter 11. Interfaces\">\n   Chapter 11<\/a>.  But the inquiry can't stop with\nindicting the GUI, because with all its flaws there is tremendous\ndemand for GUIs from end users.  Supposing we could get the metaphor\nright at the level of the design of user interactions, would Unix be\ncapable of supporting it gracefully?\n<\/p>\n<p>\n  The answer is: probably not.  We touched on this problem in\nconsidering whether the bag-of-bytes model is adequate.\nMacintosh-style file attributes may help provide the mechanism for\nricher support of GUIs, but it seems very unlikely that they are the\nwhole answer.  Unix's object model doesn't include the right\nfundamental constructs.  We need to think through what a really strong\nframework for GUIs would be like \u2014 and, just as importantly,\nhow it can be integrated with the existing frameworks of Unix.  This\nis a hard problem, demanding fundamental insights that have yet to\nemerge from the noise and confusion of ordinary software engineering\nor academic research.\n<\/p>\n<h4 id=\"id3015838\"><a href=\"#id3015838\">\u00a7<\/a>File Deletion Is Forever<\/h4>\n<p>\n  People with VMS\n  experience, or who remember\nTOPS-20\n  often\nmiss these systems' file-versioning facilities. Opening an existing file\nfor write or deleting it actually renamed it in a predictable way\nincluding a version number; only an explicit removal operation on a\nversion file actually erased data.\n<\/p>\n<p>\n  Unix does without this, at a not inconsiderable cost in user\nirritation when the wrong files get deleted through a typo or\nunexpected effects of shell wildcarding.\n<\/p>\n<p>\n  There does not seem to be any foreseeable prospect that this will\nchange at the operating system level.  Unix developers like clear, simple\noperations that do what the user tells them to do, even if the user's\ninstructions could amount to commanding \u201cshoot me in the\nfoot\u201d.  Their instinct is to say that protecting the user\nfrom himself should be done at the GUI or application level, not\nin the operating system.\n<\/p>\n<h4 id=\"id3015895\"><a href=\"#id3015895\">\u00a7<\/a>Unix Assumes a Static File System<\/h4>\n<p>\n  Unix has, in one sense, a very static model of the world.\nPrograms are implicitly assumed to run only briefly, so the\nbackground of files and directories can be assumed static during their\nexecution.  There is no standard, well-established way to ask the\nsystem to notify an application if and when a specified file or\ndirectory changes.  This becomes a significant issue when writing\nlong-lived user-interface software which wants to know about changes\nto the background.\n<\/p>\n<p>\n  Linux has file- and directory-change notification\nfeatures,<sup><a href=\"#ftn.id3015918\" id=\"id3015918\">[156]<\/a><\/sup>\n  and some versions of BSD have copied them, but\nthese are not yet portable to other Unixes.\n<\/p>\n<h4 id=\"id3015943\"><a href=\"#id3015943\">\u00a7<\/a>The Design of Job Control Was Badly Botched<\/h4>\n<p>\n  Apart from the ability to suspend processes (in itself a\ntrivial addition to the scheduler which could be made fairly\ninoffensive) what job control is about is switching a\nterminal among multiple processes.  Unfortunately, it does the easiest\npart \u2014 deciding where keystrokes go \u2014 and punts all the\nhard parts, like saving and restoring the state of the screen, to the\napplication.\n<\/p>\n<p>\n  A really good implementation of such a facility would be\ncompletely invisible to user processes: no dedicated signals, no need\nto save and restore terminal modes, no need for the applications to\nredraw the screen at random times.  The model ought to be a virtual\nkeyboard that is sometimes connected to the real one (and blocks you\nif you ask for input when it isn't connected) and a virtual screen\nwhich is sometimes visible on the real one (and might or might not\nblock on output when it's not), with the system doing the multiplexing\nin the same way it multiplexes access to the disk, the processor,\netc... and no impact on user programs at all.<sup><a href=\"#ftn.id3015976\" id=\"id3015976\">[157]<\/a><\/sup><\/p>\n<p>\n  Doing it right would have required the Unix tty driver to track\nthe entire current screen state rather than just maintaining a line\nbuffer, and to know about terminal types at kernel level (possibly\nwith help from a daemon process) so it could do restores properly when\na suspended process is foregrounded again.  A consequence of doing it\nwrong is that the Unix kernel can't detach a session, such as an\n  <em>xterm<\/em> or\n  <em>Emacs<\/em> job, from one terminal and re-attach it to another (which could be of\na different type).\n<\/p>\n<p>\n  As Unix usage has shifted to X displays and terminal emulators,\njob control has become relatively less important, and this issue does\nnot have quite the force it once did.  It is still annoying that there\nis no suspend\/attach\/detach, however; this feature could be useful for\nsaving the state of terminal sessions between logins.\n<\/p>\n<p>\n  A common open-source program called\nscreen(1)\nsolves several of these problems.<sup><a href=\"#ftn.id3016047\" id=\"id3016047\">[158]<\/a><\/sup>\n  However, since it has to be called explicitly by the user, its\nfacilities are not guaranteed to be present in every terminal session;\nalso, the kernel-level code that overlaps with it in function has not been\nremoved.\n<\/p>\n<h4 id=\"id3016077\"><a href=\"#id3016077\">\u00a7<\/a>The Unix API Doesn't Use Exceptions<\/h4>\n<p>\n  C lacks a facility for throwing named exceptions with attached\ndata.<sup><a href=\"#ftn.id3016087\" id=\"id3016087\">[159]<\/a><\/sup>\n  Thus, the C functions in the Unix API\nindicate errors by returning a distinguished value (usually \u22121 or a\nNULL character pointer) and setting a global errno variable.\n<\/p>\n<p>\n  In retrospect, this is the source of many subtle errors.\nProgrammers in a hurry often neglect to check return values.  Because\nno exception is thrown, the Rule of Repair is violated; program flow\ncontinues until the error condition manifests as some kind of failure\nor data corruption later in execution.\n<\/p>\n<p>\n  The absence of exceptions also means that some tasks which ought \nto be simple idioms \u2014 like aborting from a signal handler on a\nversion with Berkeley-style signals \u2014 have to be performed with\ncode that is complex, subject to portability glitches, and bug-prone.\n<\/p>\n<p>\n  This problem can be (and normally is) hidden by bindings of the\nUnix API in languages such as Python or Java that have exceptions.\n<\/p>\n<p>\n  The lack of exceptions is actually an indicator of a problem\nwith larger immediate implications; C's weak type ontology\nmakes communication between higher-level languages implemented in\nit problematic.  Most of the more modern languages, for example, have \nlists and dictionaries as primary data types \u2014 but, because\nthese don't have any canonical representation in the universe of C, \nattempting to pass lists between (say) Perl and Python is an unnatural\nact requiring a lot of glue.\n<\/p>\n<p>\n  There are technologies that address the larger problem, such as CORBA,\nbut they tend to involve a lot of runtime translation and be\nunpleasantly heavyweight.\n<\/p>\n<h4 id=\"id3016155\"><a href=\"#id3016155\">\u00a7<\/a>ioctl(2) and fcntl(2) Are an Embarrassment<\/h4>\n<p>\n  The ioctl(2) and fcntl(2) mechanisms provide a way to write\nhooks into a device driver.  The original, historical use of ioctl(2)\nwas to set parameters like baud rate and number of framing bits in a\nserial-communications driver, thus the name (for \u2018I\/O\ncontrol\u2019).  Later, ioctl calls were added for other driver\nfunctions, and fcntl(2) was added as a hook into the\nfile system.\n<\/p>\n<p>\n  Over the years,\n  <code>ioctl<\/code> and\n  <code>fcntl<\/code> calls have proliferated.  They are\noften poorly documented, and often a source of portability problems as\nwell.  With each one comes a grubby pile of macro definitions\ndescribing operation types and special argument values.\n<\/p>\n<p>\n  The underlying problem is the same as \u2018big bag of\nbytes\u2019; Unix's object model is weak, leaving no natural places\nto put many auxiliary operations.  Designers have an untidy choice\namong unsatisfactory alternatives;\n  <code>fcntl<\/code> \/\n  <code>ioctl<\/code> going\nthrough devices in\n  <code>\/dev<\/code>, new\nspecial-purpose system calls, or hooks through special-purpose virtual\nfile systems that hook into the kernel (e.g.\n  <code>\/proc<\/code> under Linux and elsewhere).\n<\/p>\n<p>\n  It is not clear whether or how Unix's object model will be\nenriched in the future.  If MacOS-like file attributes become a common\nfeature of Unix, tweaking magic named attributes on device drivers may\ntake over the role\n  <code>ioctl<\/code> \/\n  <code>fcntl<\/code> now have (this would at least have the\nmerit of not requiring piles of macro definitions before the interface\ncould be used).  We've already seen that Plan 9, which uses the named\nfile server or file system as its basic object, rather than the\nfile\/bytestream, presents another possible path.\n<\/p>\n<h4 id=\"id3019078\"><a href=\"#id3019078\">\u00a7<\/a>The Unix Security Model May Be Too Primitive<\/h4>\n<p>\n  Perhaps root is too powerful, and Unix should have finer-grained\ncapabilities or ACLs (Access Control Lists) for system-administration\nfunctions, rather than one superuser that can do anything.  People who\ntake this position argue that too many system programs have permanent\nroot privileges through the set-user-ID mechanism; if even one can be\ncompromised, intrusions everywhere will follow.\n<\/p>\n<p>\n  This argument is weak, however.  Modern Unixes allow any given\nuser account to belong to multiple security groups.  Through use of\nthe execute-permission and set-group-ID bits on program executables,\neach group can in effect function as an ACL for files or programs.\n<\/p>\n<p>\n  This theoretical possibility is very little used, however,\nsuggesting that the demand for ACLs is much less in practice than it\nis in theory.\n<\/p>\n<h4 id=\"id3019113\"><a href=\"#id3019113\">\u00a7<\/a>Unix Has Too Many Different Kinds of Names<\/h4>\n<p>\n  Unix unified files and local devices \u2014 they're all just\nbyte streams.  But network devices accessed through sockets have\ndifferent semantics in a different namespace.  Plan 9\n  demonstrates that\nfiles can be smoothly unified with both local and remote (network)\ndevices, and all of these things can be managed through a namespace\nthat is dynamically adjustable per-user and even per-program.\n<\/p>\n<h4 id=\"id3019140\"><a href=\"#id3019140\">\u00a7<\/a>File Systems Might Be Considered Harmful<\/h4>\n<p>\n  Was having a file system at all the wrong thing? Since the late\n1970s there has been an intriguing history of research into persistent\nobject stores and operating systems that don't have a shared global\nfile system at all, but rather treat disk storage as a huge swap area\nand do everything through virtualized object pointers.\n<\/p>\n<p>\n  Modern efforts in this line (such as\nEROS<sup><a href=\"#ftn.id3019167\" id=\"id3019167\">[160]<\/a><\/sup>) hint that such designs can offer large benefits including\nboth provable conformance to a security policy and higher\nperformance.  It must be noted, however, that if this is a failure of Unix,\nit is equally a failure of all of its competitors; no major\nproduction operating system has yet followed EROS's\nlead.<sup><a href=\"#ftn.id3019184\" id=\"id3019184\">[161]<\/a><\/sup><\/p>\n<h4 id=\"id3019193\"><a href=\"#id3019193\">\u00a7<\/a>Towards a Global Internet Address Space<\/h4>\n<p>\n  Perhaps URLs don't go far enough.  We'll leave the last word\non possible future directions of Unix to Unix's inventor:\n<\/p>\n <blockquote>\n  <p>\n   My ideal for the future is to develop a file system remote interface\n(a la Plan 9) and then have it implemented across the Internet as the\nstandard rather than HTML.  That would be ultimate cool.\n  <\/p><div>\n  \u2013 Ken Thompson\n <\/div><\/blockquote>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3015918\" id=\"ftn.id3015918\">[156]<\/a><\/sup>\n  Look for\n  <code>F_NOTIFY<\/code> under\nfcntl(2).\n<\/p><p><sup><a href=\"#id3015976\" id=\"ftn.id3015976\">[157]<\/a><\/sup>\n  This\nparagraph is based on a 1984 analysis by Henry Spencer.  He went on to\nnote that job control was necessary and appropriate for POSIX.1 and\nlater Unix standards to consider precisely\n  \n   <em>because<\/em>\n  \n  it oozes its way into every program, and\nhence has to be thought about in any application-to-system interface.\nHence, POSIX's endorsement of a mis-design, while proper solutions were\n\u201cout of scope\u201d and hence were not even\nconsidered.\n<\/p><p><sup><a href=\"#id3016047\" id=\"ftn.id3016047\">[158]<\/a><\/sup>\n  There is a\nproject site for\nscreen(1)\nat\n  <a href=\"http:\/\/www.math.fu-berlin.de\/~guckes\/screen\/\">\n   http:\/\/www.math.fu-berlin.de\/~guckes\/screen\/<\/a>.\n<\/p><p><sup><a href=\"#id3016087\" id=\"ftn.id3016087\">[159]<\/a><\/sup>\n  For nonprogrammers,\n  <em>throwing an exception<\/em>\n  is a way for a program to bail out in the middle\nof a procedure. It's not quite an exit because the throw can be\nintercepted by catcher code in an enclosing procedure.  Exceptions are\nnormally used to signal errors or unexpected conditions that mean it\nwould be pointless to try to continue normal\nprocessing.\n<\/p><p><sup><a href=\"#id3019167\" id=\"ftn.id3019167\">[160]<\/a><\/sup><a href=\"http:\/\/www.eros-os.org\/\">\n   http:\/\/www.eros-os.org\/<\/a>\n<\/p><p><sup><a href=\"#id3019184\" id=\"ftn.id3019184\">[161]<\/a><\/sup>\n  The operating systems of the Apple Newton, the\nAS\/400 minicomputer and the Palm handheld could be considered\nexceptions.\n<\/p><\/div>\n\n<h3 id=\"ch20s04\"><a href=\"#ch20s04\">\u00a7<\/a>Problems in the Environment of Unix<\/h3>\n<p>\n  The old-time Unix culture has largely reinvented itself in the \nopen-source movement.  Doing so saved us from extinction, but it also\nmeans that the problems of open source are now ours as well.\n<\/p>\n<p>\n  One of these is how to make open-source development economically\nsustainable.  We have reconnected with our roots in the collaborative,\nopen process of Unix's early days.  We have largely won the technical\nargument for abandoning secrecy and proprietary control.  We have\nthought of ways to cooperate with markets and managers on more equal\nterms than we ever could in the 1970s and 1980s, and in many ways our\nexperiments have succeeded.  In 2003 the open-source Unixes, and their\ncore development groups, have achieved a degree of mainstream\nrespectability and authority that would have been unimaginable as\nrecently as the mid-1990s.\n<\/p>\n<p>\n  We have come a long way.  But we have a long way to go yet.  We\nknow what business models might work in theory, and now we can even\npoint at a sporadic handful of successes that demonstrate that they\nwork in practice; now we have to show that they can be made to work\nreliably over a longer term.\n<\/p>\n<p>\n  It's not necessarily going to be an easy transition.  Open\nsource turns software into a service industry.  Service-provider firms\n(think of medical and legal practices) can't be scaled up by injecting\nmore capital into them; those that try only scale up their fixed\ncosts, overshoot their revenue base, and starve to death.  The choices\ncome down to singing for your supper (getting paid through tips and\ndonations), running a corner shop (a small, low-overhead service\nbusiness), or finding a wealthy patron (some large firm that needs to\nuse and modify open-source software for its business purposes).\n<\/p>\n<p>\n  In total, the amount of money spent to hire software developers\ncan be expected to rise, for the same reasons that mechanics' hourly\nwages go up as the price of automobiles drops.<sup><a href=\"#ftn.id3019301\" id=\"id3019301\">[162]<\/a><\/sup>\n  But it is going to become\nmore difficult for any one individual or firm to capture a lot of that\nspending.  There will be many programmers who are well off, but\nfewer millionaires. This is actually a sign of progress, of\ninefficiencies being competed out of the system.  But it will\nrepresent a big change in climate, and probably means that investors\nwill lose what little interest they have left in funding software\nstartups.\n<\/p>\n<p>\n  One important subproblem related to the increasing difficulty\nof sustaining really large software businesses is how to organize\nend-user testing.  Historically, the Unix culture's concentration on\ninfrastructure has meant that we have not tended to build programs\nthat depended for their value on providing a comfortable interface for\nend-users.  Now, especially in the open-source Unixes that aim to\ncompete directly with\nMicrosoft\n  and\nApple, that is changing.  But end-user interfaces need to be\nsystematically tested with real end users \u2014 and therein lie some\nchallenges.\n<\/p>\n<p>\n  Real end-user testing demands facilities, specialists, and a\nlevel of monitoring that are difficult for the distributed volunteer\ngroups characteristic of open-source development to arrange.  It may\nbe, therefore, that open-source word processors, spreadsheets, and\nother \u2018productivity\u2019 applications have to be left in the\nhands of large corporate-sponsored efforts like OpenOffice.org that\ncan afford the overhead.  Open-source developers consider single\ncorporations to be single points of failure and worry about such\ndependencies, but no better solution has yet evolved.\n<\/p>\n<p>\n  These are economic problems.  We have other problems of a more\npolitical nature, because success makes enemies.\n<\/p>\n<p>\n  Some are familiar.\nMicrosoft's\n  ambition for an unchallengeable monopoly lock on computing made the\ndefeat of Unix a strategic goal for the company in the mid-1980s, five\nyears before we knew we were in a fight.  In mid-2003, despite\nhaving had several growth markets it was counting on largely usurped\nby Linux, Microsoft is still the wealthiest and\nmost powerful software company in the world.  Microsoft knows very\nwell that it must defeat the new-school Unixes of the open-source\nmovement to survive.  To defeat them, it must destroy or discredit the\nculture that produced them.\n<\/p>\n<p>\n  Unix's comeback in the hands of the open-source community, and\nits association with the freewheeling culture of the Internet, has\nmade it newer enemies as well.  Hollywood and Big Media feel deeply\nthreatened by the Internet and have launched a multipronged attack on\nuncontrolled software development.  Existing legislation like the\nDigital Millennium Copyright Act has already been used to prosecute\nsoftware developers who were doing things the media moguls disliked\n(the most notorious cases, of course, involve the DeCSS software that\nenables playing of encrypted DVDs).  Contemplated schemes like the\nso-called Trusted Computing Platform Alliance and Palladium\nthreaten<sup><a href=\"#ftn.id3019418\" id=\"id3019418\">[163]<\/a><\/sup>\n  to make open-source development\neffectively illegal \u2014 and if open source goes down, Unix is very\nlikely to go down with it.\n<\/p>\n<p>\n  Unix and the\nhackers\n  and the\nInternet against\nMicrosoft\n  and\nHollywood and Big Media.  It's a struggle we need to win for all our\ntraditional reasons of professionalism, allegiance to our craft, and\nmutual tribal loyalty.  But there are larger reasons this struggle is\nimportant.  The possibilities of politics are increasingly shaped by\ncommunication technology \u2014 who can use it, who can censor it,\nwho can control it.  Government and corporate control of the content\nof the nets, and of what people can do with their computers, is a\nsevere long-term threat to political freedom.  The nightmare scenario\nis one in which corporate monopolism and statist power-seeking, always\nnatural allies, feed back into each other and create rationales for\nincreasing regulation, repression, and criminalization of digital\nspeech. In opposing this, we are the warriors of liberty \u2014\nnot merely our own liberty, but everyone else's as well.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3019301\" id=\"ftn.id3019301\">[162]<\/a><\/sup>\n  For a\nmore complete discussion of this effect, see\n  <em>The Magic Cauldron<\/em>\n  in <sup><a href=\"#Raymond01\" title=\"[Raymond01]\">[Raymond01]<\/a><\/sup>.\n<\/p><p><sup><a href=\"#id3019418\" id=\"ftn.id3019418\">[163]<\/a><\/sup>\n  See the\n  <a href=\"http:\/\/www.cl.cam.ac.uk\/~rja14\/tcpa-faq.html\">\n   TCPA FAQ<\/a>\n  for a rather hair-raising summary of the possibilities by a noted\nsecurity specialist.\n<\/p><\/div>\n\n<h3 id=\"ch20s05\"><a href=\"#ch20s05\">\u00a7<\/a>Problems in the Culture of Unix<\/h3>\n<p>\n  Just as important as the technical problems with Unix itself and\nthe challenges consequent on its success are the cultural problems of\nthe community around it.  There are at least two serious ones: a\nlesser challenge of internal transition, and a greater one of\novercoming our historical elitism.\n<\/p>\n<p>\n  The lesser challenge is that of friction between the old-school\nUnix gurus and the new-school open-source crowd.  The success of\nLinux, in particular, is not an\nentirely comfortable phenomenon for a lot of older Unix programmers.\nThis is partly a generational problem.  The raucous energy,\nna\u00efvete and gleeful zealotry of the Linux kids sometimes grates\non elders who have been around since the 1970s and (often rightly)\nconsider themselves wiser.  It's only exacerbated by the fact that the\nkids are succeeding where the elders failed.\n<\/p>\n<p>\n  The greater problem of psychology only became clear to me\nafter spending three days at a Macintosh developer conference in 2000.\nIt was a very enlightening experience to be immersed in a programming\nculture with assumptions diametrically opposed to those of the Unix\nworld.\n<\/p>\n<p>\n  Macintosh programmers are all about the user experience.\nThey're architects and decorators.  They design from the outside in,\nasking first \u201cWhat kind of interaction do we want to\nsupport?\u201d and then building the application logic behind it to\nmeet the demands of the user-interface design.  This leads to programs\nthat are very pretty and infrastructure that is weak and rickety.  In\none notorious example, as late as Release 9 the MacOS memory manager\nsometimes required the user to manually deallocate memory by manually\nchucking out exited but still-resident programs.  Unix people are\nviscerally revolted by this kind of mal-design; they don't understand\nhow Macintosh people could live with it.\n<\/p>\n<p>\n  By contrast, Unix people are all about infrastructure.  We are\nplumbers and stonemasons.  We design from the inside out, building\nmighty engines to solve abstractly defined problems (like \u201cHow\ndo we get reliable packet-stream delivery from point A to point B over\nunreliable hardware and links?\u201d).  We then wrap thin and often\nprofoundly ugly interfaces around the engines.  The commands\ndate(1),\nfind(1),\nand\ned(1)\nare notorious examples, but there are hundreds of others.  Macintosh\npeople are viscerally revolted by this kind of mal-design; they don't\nunderstand how Unix people can live with it.\n<\/p>\n<p>\n  Both design philosophies have some validity, but the two camps\nhave a great deal of difficulty seeing each other's points. The\ntypical Unix developer's reflex is to dismiss Macintosh software as\ngaudy fluff, eye-candy for the ignorant, and to continue\nbuilding software that appeals to other Unix developers. If end-users\ndon't like it, so much the worse for the end users; they will come\naround when they get a clue.\n<\/p>\n<p>\n  In many ways this kind of parochialism has served us well.  We\nare the keepers of the Internet and the World Wide Web.  Our software\nand our traditions dominate serious computing, the applications where\n24\/7 reliability and minimal downtime is a must.  We really are\nextremely good at building solid infrastructure; not perfect by any\nmeans, but there is no other software technical culture that has\nanywhere close to our track record, and it is one to be proud\nof.\n<\/p>\n<p>\n  The problem is that we increasingly face challenges that demand\na more inclusive view.  Most of the computers in the world don't live\nin server rooms, but rather in the hands of those end users.  In early\nUnix days, before personal computers, our culture defined itself\npartly as a revolt against the priesthood of the mainframes, the\nkeepers of the big iron.  Later, we absorbed the power-to-the-people\nidealism of the early microcomputer enthusiasts.  But today\n  <em>we<\/em>\n  are the priesthood;\n  <em>we<\/em>\n  are the people who run the networks and the big iron.  And our\nimplicit demand is that if you want to use our software, you must\nlearn to think like us.\n<\/p>\n<p>\n  In 2003, there is a deep ambivalence in our attitude \u2014 a\ntension between elitism and missionary populism.  We want to reach and\nconvert the 92% of the world for whom computing means games and\nmultimedia and glossy GUI interfaces and (at their most technical)\nlight email and word processing and spreadsheets.  We are spending\nmajor effort on projects like GNOME and KDE designed to give Unix a\npretty face.  But we are still elitists at heart, deeply reluctant and\nin many cases unable to identify with or listen to the needs of the\nAunt Tillies of the world.\n<\/p>\n<p>\n  To non-technical end users, the software we build tends to be\neither bewildering and incomprehensible, or clumsy and condescending,\nor both at the same time.  Even when we try to do the\nuser-friendliness thing as earnestly as possible, we're woefully\ninconsistent at it.  Many of the attitudes and reflexes we've\ninherited from old-school Unix are just wrong for the job.  Even when\nwe want to listen to and help Aunt Tillie, we don't know how \u2014\nwe project our categories and our concerns onto her and give her\n\u2018solutions\u2019 that she finds as daunting as her\nproblems.\n<\/p>\n<p>\n  Our greatest challenge as a culture is whether we can outgrow\nthe assumptions that have served us so well \u2014 whether we can\nacknowledge, not merely intellectually but in the sinew of daily\npractice, that the Macintosh people have a point.  Their point is made\nin more general, less Mac-specific way in\n  <em>The Inmates Are Running the Asylum<\/em><sup><a href=\"#Cooper\" title=\"[Cooper]\">[Cooper]<\/a><\/sup>, an insightful and argumentative\nbook about what its author calls\n  <em>interaction design<\/em>\n  that (despite occasional crotchets) contains a good\ndeal of hard truth that every Unix programmer ought to know.\n<\/p>\n<p>\n  We can turn aside from this; we can remain a priesthood\nappealing to a select minority of the best and brightest, a geek\nmeritocracy focused on our historical role as the keepers of the\nsoftware infrastructure and the networks.  But if we do this, we will\nvery likely go into decline and eventually lose the dynamism that has\nsustained us through decades.  Someone else will serve the people;\nsomeone else will put themselves where the power and the money are, and\nown the future of 92% of all software. The odds are, whether that\nsomeone else is\nMicrosoft\n  or not,\nthat they will do it using practices and software we don't much\nlike.\n<\/p>\n<p>\n  Or we can truly accept the challenge.  The open-source movement\nis trying hard to do so.  But the kind of sustained work and\nintelligence we have brought to other problems in the past will not\nalone suffice.  Our attitudes must change in a fundamental and\ndifficult way.\n<\/p>\n<p>\n  In\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  we\ndiscussed the importance of throwing away limiting assumptions and\ndiscarding the past in solving technical problems, suggesting a\nparallel with the Zen ideas of detachment and \u2018beginner's\nmind\u2019.  We have a larger kind of detachment to work on now.  We\nmust learn humility before Aunt Tillie, and relinquish some of the\nlong-held prejudices that have made us so successful in the\npast.\n<\/p>\n<p>\n  Tellingly, the Macintosh culture has begun to converge with ours\n\u2014 MacOS X has Unix underneath, and in 2003 Mac developers\nare (albeit with a struggle in some cases) making the mental \nadjustment to learn the infrastructure-focused virtues of Unix.\nOur challenge will be, reciprocally, to embrace the user-centered\nvirtues of the Macintosh.\n<\/p>\n<p>\n  There are other signs that the Unix culture is shedding its\ninsularity as well.  One is the convergence that seems to be going on\nbetween the Unix\/open-source community and the movement called\n\u201cagile programming\u201d.<sup><a href=\"#ftn.id3019774\" id=\"id3019774\">[164]<\/a><\/sup>\n  We noted in\n  <a href=\"#modularitychapter\" title=\"Chapter 4. Modularity\">\n   Chapter 4<\/a>\n  that Unix programmers have seized happily\non the concept of refactoring, one of the preoccupations of the\nagile-programming thinkers.  Refactoring, and other agile concepts\nlike unit-testing and design around stories, seem to articulate and\nsharpen practices that have heretofore been widespread but only\nimplicit in the Unix tradition. The Unix tradition, on the other hand,\ncan bring groundedness and the lessons of long experience to the\nagile-programming party.  As open-source software gains market share\nit is even conceivable that these cultures will fuse, much as the\nold-time Internet and early Unix cultures did after 1980.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3019774\" id=\"ftn.id3019774\">[164]<\/a><\/sup>\n  For an introduction\nto agile programming, see the\n  <a href=\"http:\/\/agilemanifesto.org\/\">\n   Agile\nManifesto<\/a>\n<\/p><\/div>\n\n<h3 id=\"ch20s06\"><a href=\"#ch20s06\">\u00a7<\/a>Reasons to Believe<\/h3>\n<p>\n  The future of Unix is full of difficult problems.  Would we\ntruly want it any other way?\n<\/p>\n<p>\n  For more than thirty years we have thrived on challenges.  We\npioneered the best practices of software engineering. We created\ntoday's Internet and Web.  We have built the largest, most complex,\nand most reliable software systems ever to exist.  We outlasted the\nIBM\n  monopoly and we're\nmaking a run against the\nMicrosoft\n  monopoly\nthat is good enough to deeply frighten it.\n<\/p>\n<p>\n  Not that everything has been triumph by any means.  In the 1980s\nwe nearly destroyed ourselves by acceding to the proprietary capture\nof Unix.  We neglected the low end, the nontechnical end users, for\nfar too long and thereby left\nMicrosoft\n  an\nopening to grossly lower the quality standards of software.\nIntelligent observers have pronounced our technology, our community,\nand our values to be dead any number of times.\n<\/p>\n<p>\n  But always we have come storming back.  We make mistakes, but we\nlearn from our mistakes.  We have transmitted our culture across\ngenerations; we have absorbed much of what was best from the early\nacademic hackers\n  and\nthe ARPANET experimenters and the microcomputer enthusiasts and a\nnumber of other cultures.  The open-source movement has resurrected\nthe vigor and idealism of our early years, and today we are stronger\nand more numerous than we have ever been.\n<\/p>\n<p>\n  So far, betting against the Unix hackers has always been\nshort-term smart but long-term stupid.  We can prevail \u2014 if we\nchoose to.\n<\/p>\n\n<h3 id=\"apa\"><a href=\"#apa\">\u00a7<\/a>Appendix A. Glossary of Abbreviations<\/h3>\n<p>The most important abbreviations and acronyms used in the main text are defined here.<\/p>\n<dl><dt>\n     API\n    <\/dt>\n    <dd>\n     <p>\n      <em>Application Programming Interface.<\/em>\n      The\n   set of procedure calls that communicates with a linkable procedure\n   library or an operating-system kernel or the combination of\n   both.\n     <\/p>\n    <\/dd>\n    <dt>\n     BSD\n    <\/dt>\n    <dd>\n     <p>\n\t  <em>Berkeley System Distribution<\/em>; also <em>Berkeley Software Distribution<\/em>; sources are ambiguous. The\n\t  generic name of the Unix distributions issued by the Computer Science Research Group at the University of\n\t  California at Berkeley between 1976 and 1994, and of the open-source Unixes genetically descended from them.\n     <\/p>\n    <\/dd>\n    <dt>\n     CLI\n    <\/dt>\n    <dd>\n     <p>\n      <em>Command Line Interface.<\/em>\n      Considered\n   archaic by some, but still very useful in the Unix world.\n     <\/p>\n    <\/dd>\n    <dt>\n     CPAN\n    <\/dt>\n    <dd>\n     <p>\n      <em>\n       Comprehensive Perl\n       Archive Network.\n<\/em>\n      The\n   central\n      <a href=\"http:\/\/cpan.org\/\">\n       Web repository<\/a>\n      for\n   Perl modules and extensions.\n     <\/p>\n    <\/dd>\n    <dt>\n     GNU\n    <\/dt>\n    <dd>\n     <p>\n      <em>\n       GNU's Not\n   Unix!\n<\/em>\n      The recursive acronym for the Free Software\n   Foundation's project to produce an entire free-software clone of\n   Unix.  This effort didn't completely succeed, but did produce many of\n   the essential tools of modern Unix development including Emacs and\n   the GNU Compiler Collection.\n     <\/p>\n    <\/dd>\n    <dt>\n     GUI\n    <\/dt>\n    <dd>\n     <p>\n      <em>Graphical User Interface.<\/em>\n      The modern\n   style of application interface using mice, windows, and icons\n   invented at Xerox PARC during the 1970s, as opposed to the older\n   CLI or roguelike styles.\n     <\/p>\n    <\/dd>\n    <dt>\n     IDE\n    <\/dt>\n    <dd>\n     <p>\n      <em>Integrated Development Environment.<\/em>\n      A GUI workbench for developing code, featuring facilities like\n   symbolic debugging, version control, and data-structure browsing.\n   These are not commonly used under Unix, for reasons discussed in\n      <a href=\"#toolschapter\" title=\"Chapter 15. Tools\">\n       Chapter 15<\/a>.\n     <\/p>\n    <\/dd>\n    <dt>\n     IETF\n    <\/dt>\n    <dd>\n     <p>\n      <em>Internet Engineering Task Force.<\/em>\n      The\n   entity responsible for defining Internet protocols such as\n   TCP\/IP. A loose, collegial organization mainly of technical\n   people.\n     <\/p>\n    <\/dd>\n    <dt>\n     IPC\n    <\/dt>\n    <dd>\n     <p>\n      <em>Inter-Process Communication.<\/em>\n      Any\n   method of passing data between processes running in separate\n   address spaces.\n     <\/p>\n    <\/dd>\n    <dt>\n     MIME\n    <\/dt>\n    <dd>\n     <p>\n      <em>Multipurpose Internet Mail Extensions.<\/em>\n      A\n   series of RFCs that describe standards for embedding binary and\n   multipart messages within RFC-822 mail.  Besides being used for\n   mail transport, MIME is used as an underlevel by important\n   application protocols including HTTP and BEEP.\n     <\/p>\n    <\/dd>\n    <dt>\n     OO\n    <\/dt>\n    <dd>\n     <p>\n      <em>Object Oriented.<\/em>\n      A style of programming\n   that tries to encapsulate data to be manipulated and the code that\n   manipulates it in (theoretically) sealed containers called objects.\n   By contrast, non-object-oriented programming is more casual about\n   exposing the internals of the data structure and code.\n     <\/p>\n    <\/dd>\n    <dt>\n     OS\n    <\/dt>\n    <dd>\n     <p>\n      <em>Operating System.<\/em>\n      The foundation software of a\n   machine; that which schedules tasks, allocates storage, and presents a\n   default interface to the user between applications.  The facilities an\n   operating system provides and its general design philosophy exert an\n   extremely strong influence on programming style and on the technical\n   cultures that grow up around its host machines.\n     <\/p>\n    <\/dd>\n    <dt>\n     PDF\n    <\/dt>\n    <dd>\n     <p>\n      <em>Portable Document Format.<\/em>\n      The\n   PostScript language for control of printers and other imaging\n   devices is designed to be streamed to printers.  PDF is a sequence\n   of PostScript pages, packaged with annotations so it can\n   conveniently be used as a display format.\n     <\/p>\n    <\/dd>\n    <dt>\n     PDP-11\n    <\/dt>\n    <dd>\n     <p>\n      <em>Programmable Data Processor 11.<\/em>\n      Possibly\n   the single most successful minicomputer design in history; first\n   shipped in 1970, last shipped in 1990, and the immediate ancestor\n   of the VAX.  The PDP-11 was the first major Unix platform.\n     <\/p>\n    <\/dd>\n    <dt>\n     PNG\n    <\/dt>\n    <dd>\n     <p>\n      <em>Portable Network Graphics.<\/em>\n      The World\n   Wide Web Consortium's standard and recommended format for bitmap\n   graphics images. An elegantly designed binary graphics format\n   described in\n      <a href=\"#textualitychapter\" title=\"Chapter 5. Textuality\">\n       Chapter 5<\/a>.\n     <\/p>\n    <\/dd>\n    <dt>\n     RFC\n    <\/dt>\n    <dd>\n     <p>\n      <em>Request For Comment.<\/em>\n      An Internet\n   standard.  The name arose at a time when the documents were\n   regarded as proposals to be submitted to a then-nonexistent \n   but anticipated formal approval process of some sort.  The\n   formal approval process never materialized.\n     <\/p>\n    <\/dd>\n    <dt>\n     RPC\n    <\/dt>\n    <dd>\n     <p>\n      <em>Remote Procedure Call.<\/em>\n      Use of\n   IPC methods that attempt to create the illusion that the \n   processes exchanging them are running in the same address \n   space, so they can cheaply (a) share complex structures, and (b)\n   call each other like function libraries, ignoring latency and other\n   performance considerations.  This illusion is notoriously\n   difficult to sustain.\n     <\/p>\n    <\/dd>\n    <dt>\n     TCP\/IP\n    <\/dt>\n    <dd>\n     <p>\n      <em>\n       Transmission Control Protocol\/Internet\n   Protocol.\n<\/em>\n      The basic protocol of the Internet since the\n   conversion from NCP (Network Control Protocol) in 1983. Provides \n   reliable transport of data streams.\n     <\/p>\n    <\/dd>\n    <dt>\n     UDP\/IP\n    <\/dt>\n    <dd>\n     <p>\n      <em>\n       Universal Datagram Protocol\/Internet\n   Protocol.\n<\/em>\n      Provides unreliable but low-latency transport\n   for small data packets.\n     <\/p>\n    <\/dd>\n    <dt>\n     UI\n    <\/dt>\n    <dd>\n     <p>\n      <em>User Interface.<\/em>\n     <\/p>\n    <\/dd>\n    <dt>\n     VAX\n    <\/dt>\n    <dd>\n     <p>\n      Formally,\n      <em>Virtual Address Extension<\/em>\n      :\n   the name of a classic minicomputer design developed by Digital\n   Equipment Corporation (later merged with Compaq, later merged with\n   Hewlett-Packard) from the PDP-11. The\n   first VAX shipped in 1977.  For ten years after 1980 VAXen were\n   among the most important Unix platforms.  Microprocessor\n   reimplementations are still shipping today.\n     <\/p>\n    <\/dd>\n   <\/dl><h3 id=\"apb\"><a href=\"#apb\">\u00a7<\/a>Appendix B. References<\/h3>\n<p>\n\tEvent timelines of the\n\t<a href=\"http:\/\/snap.nlc.dcccd.edu\/learn\/drkelly\/hst-hand.htm\">Unix Industry<\/a>\n\tand of <a href=\"http:\/\/www.robotwisdom.com\/linux\/timeline.html\">GNU\/Linux and Unix<\/a>\n\tare available on the Web.  A timeline tree of <a href=\"http:\/\/www.levenez.com\/unix\/\">Unix releases<\/a> is also available.\n<\/p>\n<h3 id=\"id3020425\"><a href=\"#id3020425\">\u00a7<\/a>Bibliography<\/h3>\n<ul><li id=\"Appleton\">\n\t\t<sup>[Appleton]<\/sup>\n\t\tRandy Appleton.\n\t\t<em class=\"title\">Improving Context Switching Performance of Idle Tasks under Linux<\/em>.\n\t\t2001.\n\t\t<a href=\"http:\/\/cs.nmu.edu\/~randy\/Research\/Papers\/Scheduler\/\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Baldwin-Clark\">\n\t\t<sup>[Baldwin-Clark]<\/sup>\n\t\tCarliss Baldwin and Kim Clark.\n\t\t<em class=\"title\">Design Rules, Vol 1: The Power of Modularity.<\/em>\n\t\t2000.\n\t\tMIT Press.\n\t\tISBN 0-262-024667.\n\t<\/li>\n\t<li id=\"Bentley\">\n\t\t<sup>[Bentley]<\/sup>\n\t\tJon Bentley.\n\t\t<em class=\"title\">Programming Pearls.<\/em>\n\t\t2nd Edition.\n\t\t2000.\n\t\tAddison-Wesley.\n\t\tISBN 0-201-65788-0.\n\t\tThe third essay in this book, \u201cData Structures Programs\u201d, argues a case similar to that of\n\t\t<a href=\"#generationchapter\" title=\"Chapter 9. Generation\"> Chapter 9<\/a> with\n\t\tBentley's characteristic eloquence.  Some of the book is\n\t\t<a href=\"http:\/\/www.cs.bell-labs.com\/cm\/cs\/pearls\/\"> available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"BlaauwBrooks\">\n\t\t<sup>[BlaauwBrooks]<\/sup>\n\t\tGerrit A. Blaauw and Frederick P. Brooks.\n\t\t<em class=\"title\">Computer Architecture: Concepts and Evolution.<\/em>\n\t\t1997.\n\t\tISBN 0-201-10557-8.\n\t\tAddison-Wesley.\n\t<\/li>\n\t<li id=\"Bolinger-Bronson\">\n\t\t<sup>[Bolinger-Bronson]<\/sup>\n\t\tDan Bolinger and Tan Bronson.\n\t\t<em class=\"title\">Applying RCS and SCCS.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1995.\n\t\tISBN 1-56592-117-8.\n\t\tNot just a cookbook, this also surveys the design issues in version-control systems.\n\t<\/li>\n\t<li id=\"Brokken\">\n\t\t<sup>[Brokken]<\/sup>\n\t\tFrank Brokken.\n\t\t<em class=\"title\">C++ Annotations Version.<\/em>\n\t\t2002.\n\t\t<a href=\"http:\/\/www.icce.rug.nl\/documents\/cplusplus\/cplusplus.html\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"BrooksD\">\n\t\t<sup>[BrooksD]<\/sup>\n\t\tDavid Brooks.\n\t\t<em class=\"title\">Converting a UNIX .COM Site to Windows.<\/em>\n\t\t2000.\n\t\t<a href=\"http:\/\/www.securityoffice.net\/mssecrets\/hotmail.html#_Toc491601819\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Brooks\">\n\t\t<sup>[Brooks]<\/sup>\n\t\tFrederick P. Brooks.\n\t\t<em class=\"title\">The Mythical Man-Month.<\/em>\n\t\t20th Anniversary Edition.\n\t\tAddison-Wesley.\n\t\t1995.\n\t\tISBN 0-201-83595-9.\n\t<\/li>\n\t<li id=\"Boehm\">\n\t\t<sup>[Boehm]<\/sup>\n\t\tHans Boehm.\n\t\t<em class=\"title\">Advantages and Disadvantages of Conservative Garbage Collection.<\/em>\n\t\tThorough discussion of tradeoffs between garbage-collected and non-garbage-collected environments.\n\t\t<a href=\"http:\/\/www.hpl.hp.com\/personal\/Hans_Boehm\/gc\/issues.html\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Cameron\">\n\t\t<sup>[Cameron]<\/sup>\n\t\tDebra Cameron, Bill Rosenblatt, and Eric Raymond.\n\t\t<em class=\"title\">Learning GNU Emacs.<\/em>\n\t\t2nd Edition.\n\t\tO'Reilly &amp; Associates.\n\t\t1996.\n\t\tISBN 1-56592-152-6.\n\t<\/li>\n\t<li id=\"Cannon\">\n\t\t<sup>[Cannon]<\/sup>\n\t\tL. W. Cannon, R. A. Elliot, L. W. Kirchhoff, J. A. Miller, J. M. Milner, R. W. Mitzw, E. P. Schan, N. O. Whittington, Henry Spencer, David Keppel, and Mark Brader.\n\t\t<em class=\"title\">Recommended C Style and Coding Standards.<\/em>\n\t\t1990.\n\t\tAn updated version of the <em>Indian Hill C Style and Coding Standards<\/em> paper, with modifications by the last\n\t\tthree authors. It describes a recommended coding standard for C programs.\n\t\t<a href=\"http:\/\/www.apocalypse.org\/pub\/u\/paul\/docs\/cstyle\/cstyle.htm\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Christensen\">\n\t\t<sup>[Christensen]<\/sup>\n\t\tClayton Christensen.\n\t\t<em class=\"title\">The Innovator's Dilemma.<\/em>\n\t\tHarperBusiness.\n\t\t2000.\n\t\tISBN 0-066-62069-4.\n\t\tThe book that introduced the term \u201cdisruptive technology\u201d.  A fascinating and lucid examination of how and why\n\t\ttechnology companies doing everything right get mugged by upstarts.  A business book technical people should read.\n\t<\/li>\n\t<li id=\"Comer\">\n\t\t<sup>[Comer]<\/sup><em>Unix Review<\/em>.\n\t\t\n\t\t\n\t\tDouglas Comer.\n\t\t\u201cPervasive Unix: Cause for Celebration\u201d.\n\t\tOctober 1985.\n\t\t\n\t\tp. 42.\n\t\t\n\t\t\n\t<\/li>\n\t<li id=\"Cooper\">\n\t\t<sup>[Cooper]<\/sup>\n\t\tAlan Cooper.\n\t\t<em class=\"title\">The Inmates Are Running the Asylum.<\/em>\n\t\tSams.\n\t\t1999.\n\t\tISBN 0-672-31649-8.\n\t\tDespite some occasional quirks and crotchets, this \n\t\tbook is a trenchant and brilliant analysis of what's wrong with \n\t\tsoftware interface designs, and how to put it right.\n\t<\/li>\n\t<li id=\"Coram-Lee\">\n\t\t<sup>[Coram-Lee]<\/sup>\n\t\tTod Coram and Ji Lee.\n\t\t<em class=\"title\">Experiences \u2014 A Pattern Language for User Interface Design.<\/em>\n\t\t1996.\n\t\t<a href=\"http:\/\/www.maplefish.com\/todd\/papers\/Experiences.html\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"DuBois\">\n\t\t<sup>[DuBois]<\/sup>\n\t\tPaul DuBois.\n\t\t<em class=\"title\">Software Portability with Imake.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1993.\n\t\tISBN 1-56592-055-4.\n\t<\/li>\n\t<li id=\"Eckel\">\n\t\t<sup>[Eckel]<\/sup>\n\t\tBruce Eckel.\n\t\t<em class=\"title\">Thinking in Java.<\/em>\n\t\t3rd Edition.\n\t\tPrentice-Hall.\n\t\t2003.\n\t\tISBN 0-13-100287-2.\n\t\t<a href=\"http:\/\/www.mindview.net\/Books\/TIJ\/\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Feller-Fitzgerald\">\n\t\t<sup>[Feller-Fitzgerald]<\/sup>\n\t\tJoseph Feller and Brian Fitzgerald.\n\t\t<em class=\"title\">Understanding Open Source Software.<\/em>\n\t\t2002.\n\t\tISBN 0-201-73496-6.\n\t\tAddison-Wesley.\n\t<\/li>\n\t<li id=\"FlanaganJava\">\n\t\t<sup>[FlanaganJava]<\/sup>\n\t\tDavid Flanagan.\n\t\t<em class=\"title\">Java in a Nutshell.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1997.\n\t\tISBN 1-56592-262-X.\n\t<\/li>\n\t<li id=\"FlanaganJavaScript\">\n\t\t<sup>[FlanaganJavaScript]<\/sup>\n\t\tDavid Flanagan.\n\t\t<em class=\"title\">JavaScript: The Definitive Guide.<\/em>\n\t\t4th Edition.\n\t\tO'Reilly &amp; Associates.\n\t\t2002.\n\t\tISBN 1-596-00048-0.\n\t<\/li>\n\t<li id=\"Fowler\">\n\t\t<sup>[Fowler]<\/sup>\n\t\tMartin Fowler.\n\t\t<em class=\"title\">Refactoring.<\/em>\n\t\tAddison-Wesley.\n\t\t1999.\n\t\tISBN 0-201-48567-2.\n\t<\/li>\n\t<li id=\"Friedl\">\n\t\t<sup>[Friedl]<\/sup>\n\t\tJeffrey Friedl.\n\t\t<em class=\"title\">Mastering Regular Expressions.<\/em>\n\t\t2nd Edition.\n\t\t2002.\n\t\tISBN  0-596-00289-0.\n\t\tO'Reilly &amp; Associates.\n\t\t\n\t\t484pp..\n\t\t\n\t<\/li>\n\t<li id=\"Fuzz\">\n\t\t<sup>[Fuzz]<\/sup>\n\t\tBarton Miller, David Koski, Cjin Pheow Lee, Vivekananda Maganty, Ravi Murthy, Ajitkumar Natarajan, and Jeff Steidl.\n\t\t<em class=\"title\">Fuzz Revisited: A Re-examination of the Reliability of Unix Utilities and Services.<\/em>\n\t\t2000.\n\t\t<a href=\"http:\/\/www.opensource.org\/advocacy\/fuzz-revisited.pdf\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Gabriel\">\n\t\t<sup>[Gabriel]<\/sup>\n\t\tRichard Gabriel.\n\t\t<em class=\"title\">Good News, Bad News, and How to Win Big.<\/em>\n\t\t1990.\n\t\t<a href=\"http:\/\/www.dreamsongs.com\/WorseIsBetter.html\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Gancarz\">\n\t\t<sup>[Gancarz]<\/sup>\n\t\tMike Gancarz.\n\t\t<em class=\"title\">The Unix Philosophy.<\/em>\n\t\tDigital Press.\n\t\t1995.\n\t\tISBN 1-55558-123-4.\n\t<\/li>\n\t<li id=\"GangOfFour\">\n\t\t<sup>[GangOfFour]<\/sup>\n\t\tErich Gamma, Richard Helm, Ralph Johnson, and John Vlissides.\n\t\t<em class=\"title\">Design Patterns: Elements of Reusable Object-Oriented Software.<\/em>\n\t\tAddison-Wesley.\n\t\t1997.\n\t\tISBN 0-201-63361-2.\n\t<\/li>\n\t<li id=\"Garfinkel\">\n\t\t<sup>[Garfinkel]<\/sup>\n\t\tSimson Garfinkel, Daniel Weise, and Steve Strassman.\n\t\t<em class=\"title\">The Unix Hater's Handbook.<\/em>\n\t\tIDG Books.\n\t\t1994.\n\t\tISBN 1-56884-203-1.\n\t\t<a href=\"http:\/\/research.microsoft.com\/~daniel\/unix-haters.html\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Gentner-Nielsen\">\n\t\t<sup>[Gentner-Nielsen]<\/sup><em>Communications of the ACM<\/em>.\n\t\tAssociation for Computing Machinery.\n\t\t\n\t\t\n\t\tDon Gentner and Jacob Nielsen.\n\t\t\u201cThe Anti-Mac Interface\u201d.\n\t\tAugust 1996.\n\t\t<a href=\"http:\/\/www.acm.org\/cacm\/AUG96\/antimac.htm\">Available on the Web<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Gettys\">\n\t\t<sup>[Gettys]<\/sup>\n\t\tJim Gettys.\n\t\t<em class=\"title\">The Two-Edged Sword.<\/em>\n\t\t1998.\n\t\t<a href=\"http:\/\/freshmeat.net\/articles\/view\/122\/\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Glickstein\">\n\t\t<sup>[Glickstein]<\/sup>\n\t\tBob Glickstein.\n\t\t<em class=\"title\">Writing GNU Emacs Extensions.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1997.\n\t\tISBN 1-56592-261-1.\n\t<\/li>\n\t<li id=\"Graham\">\n\t\t<sup>[Graham]<\/sup>\n\t\tPaul Graham.\n\t\t<em class=\"title\">A Plan for Spam.<\/em>\n\t\t<a href=\"http:\/\/www.paulgraham.com\/spam.html\">Available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Harold-Means\">\n\t\t<sup>[Harold-Means]<\/sup>\n\t\tElliotte Rusty Harold and W. Scott Means.\n\t\t<em class=\"title\">XML in a Nutshell.<\/em>\n\t\t2nd Edition.\n\t\tO'Reilly &amp; Associates.\n\t\t2002.\n\t\tISBN 0-596-00292-0.\n\t<\/li>\n\t<li id=\"Hatton97\">\n\t\t<sup>[Hatton97]<\/sup><em>IEEE Software<\/em>.\n\t\t\n\t\t\n\t\tLes Hatton.\n\t\t\u201cRe-examining the Defect-Density versus Component Size Distribution\u201d.\n\t\tMarch\/April 1997.\n\t\t<a href=\"http:\/\/www.cs.ukc.ac.uk\/people\/staff\/lh8\/pubs\/pubis697\/Ubend_IS697.pdf.gz\">\n\t\t\tAvailable\n\t\ton the Web<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Hatton98\">\n\t\t<sup>[Hatton98]<\/sup><em>IEEE Software<\/em>.\n\t\t\n\t\t\n\t\tLes Hatton.\n\t\t\u201cDoes OO Sync with the Way We Think?\u201d.\n\t\t\n\t\t15.\n\t\t\n\t\t\n\t\t(3).\n\t\t\n\t\t<a href=\"http:\/\/www.cs.ukc.ac.uk\/people\/staff\/lh8\/pubs\/pubis698\/OO_IS698.pdf.gz\">\n\t\t\tAvailable\n\t\ton the Web<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Hauben\">\n\t\t<sup>[Hauben]<\/sup>\n\t\tRonda Hauben.\n\t\t<em class=\"title\">History of UNIX.<\/em>\n\t\t<a href=\"http:\/\/www.dei.isep.ipp.pt\/docs\/unix.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Heller\">\n\t\t<sup>[Heller]<\/sup>\n\t\tSteve Heller.\n\t\t<em class=\"title\">C++: A Dialog.<\/em>\n\t\t\n\t\tProgramming with the C++ Standard Library.\n\t\t\n\t\tPrentice-Hall.\n\t\t2003.\n\t\tISBN 0-13-009402-1.\n\t<\/li>\n\t<li id=\"Hunt-Thomas\">\n\t\t<sup>[Hunt-Thomas]<\/sup>\n\t\tAndrew Hunt and David Thomas.\n\t\t<em class=\"title\">The Pragmatic Programmer: From Journeyman to Master.<\/em>\n\t\tAddison-Wesley.\n\t\t2000.\n\t\tISBN 0-201-61622-X.\n\t<\/li>\n\t<li id=\"Kernighan95\">\n\t\t<sup>[Kernighan95]<\/sup>\n\t\tBrian Kernighan.\n\t\t<em class=\"title\">Experience with Tcl\/Tk for Scientific and Engineering Visualization.<\/em>\n\t\tUSENIX Association Tcl\/Tk Workshop Proceedings.\n\t\t1995.\n\t\t<a href=\"http:\/\/www.usenix.org\/publications\/library\/proceedings\/tcl95\/full_papers\/kernighan.txt\">\n\t\tAvailable\n\t\ton the Web<\/a>.\n\t<\/li>\n\t<li id=\"Kernighan-Pike84\">\n\t\t<sup>[Kernighan-Pike84]<\/sup>\n\t\tBrian Kernighan and Rob Pike.\n\t\t<em class=\"title\">The Unix Programming Environment.<\/em>\n\t\tPrentice-Hall.\n\t\t1984.\n\t\tISBN 0-13-937681-X.\n\t<\/li>\n\t<li id=\"Kernighan-Pike99\">\n\t\t<sup>[Kernighan-Pike99]<\/sup>\n\t\tBrian Kernighan and Rob Pike.\n\t\t<em class=\"title\">The Practice of Programming.<\/em>\n\t\t1999.\n\t\tISBN 0-201-61586-X.\n\t\tAddison-Wesley.\n\t\tAn excellent treatise on writing high-quality programs,\n\t\tsurely destined to become a classic of the field.\n\t<\/li>\n\t<li id=\"Kernighan-Plauger\">\n\t\t<sup>[Kernighan-Plauger]<\/sup>\n\t\tBrian Kernighan and P. J. Plauger.\n\t\t<em class=\"title\">Software Tools.<\/em>\n\t\tAddison-Wesley.\n\t\t1976.\n\t\tISBN 201-03669-X.\n\t<\/li>\n\t<li id=\"Kernighan-Ritchie\">\n\t\t<sup>[Kernighan-Ritchie]<\/sup>\n\t\tBrian Kernighan and Dennis Ritchie.\n\t\t<em class=\"title\">The C Programming Language.<\/em>\n\t\t2nd Edition.\n\t\tPrentice-Hall Software Series.\n\t\t1988.\n\t\tISBN 0-13-110362-8.\n\t<\/li>\n\t<li id=\"Lampson\">\n\t\t<sup>[Lampson]<\/sup><em>ACM Operating Systems Review<\/em>.\n\t\tAssociation for Computing Machinery.\n\t\t\n\t\t\n\t\tButler Lampson.\n\t\t\u201cHints for Computer System Design\u201d.\n\t\tOctober 1983.\n\t\t<a href=\"http:\/\/research.microsoft.com\/~lampson\/33-Hints\/WebPage.html\">\n\t\t\tAvailable\n\t\ton the Web<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Lapin\">\n\t\t<sup>[Lapin]<\/sup>\n\t\tJ. E. Lapin.\n\t\t<em class=\"title\">Portable C and Unix Systems Programming.<\/em>\n\t\tPrentice-Hall.\n\t\t1987.\n\t\tISBN 0-13-686494-5.\n\t<\/li>\n\t<li id=\"Leonard\">\n\t\t<sup>[Leonard]<\/sup>\n\t\tAndrew Leonard.\n\t\t<em class=\"title\">BSD Unix: Power to the People, from the Code.<\/em>\n\t\t2000.\n\t\t<a href=\"http:\/\/dir.salon.com\/tech\/fsp\/2000\/05\/16\/chapter_2_part_one\/index.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Levy\">\n\t\t<sup>[Levy]<\/sup>\n\t\tSteven Levy.\n\t\t<em class=\"title\">Hackers: Heroes of the Computer Revolution.<\/em>\n\t\tAnchor\/Doubleday.\n\t\t1984.\n\t\tISBN 0-385-19195-2.\n\t\t<a href=\"http:\/\/www.stanford.edu\/group\/mmdd\/SiliconValley\/Levy\/Hackers.1984.book\/contents.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Lewine\">\n\t\t<sup>[Lewine]<\/sup>\n\t\tDonald Lewine.\n\t\t<em class=\"title\">POSIX Programmer's Guide: Writing Portable Unix Programs.<\/em>\n\t\t1992.\n\t\tO'Reilly &amp; Associates.\n\t\tISBN 0-937175-73-0.\n\t\t\n\t\t607pp..\n\t\t\n\t<\/li>\n\t<li id=\"Libes-Ressler\">\n\t\t<sup>[Libes-Ressler]<\/sup>\n\t\tDon Libes and Sandy Ressler.\n\t\t<em class=\"title\">Life with Unix.<\/em>\n\t\t1989.\n\t\tISBN 0-13-536657-7.\n\t\tPrentice-Hall.\n\t\tThis book gives a more detailed version of Unix's\n\t\tearly history. It's particularly strong for the period\n\t\t1979\u20131986.\n\t<\/li>\n\t<li id=\"Lions\">\n\t\t<sup>[Lions]<\/sup>\n\t\tJohn Lions.\n\t\t<em class=\"title\">Lions's Commentary on Unix 6th Edition.<\/em>\n\t\t1996.\n\t\t1-57398-013-7.\n\t\tPeer-To-Peer Communications.\n\t\tPostScript rendering of Lions's original floats around \n\t\tthe Web.\n\t\t<a href=\"http:\/\/www.upl.cs.wisc.edu\/~epaulson\/lionc.ps\">\n\t\tThis URL may\n\t\tbe unstable<\/a>.\n\t<\/li>\n\t<li id=\"Loukides-Oram\">\n\t\t<sup>[Loukides-Oram]<\/sup>\n\t\tMike Loukides and Andy Oram.\n\t\t<em class=\"title\">Programming with GNU Software.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1996.\n\t\tISBN 1-56592-112-7.\n\t<\/li>\n\t<li id=\"Lutz\">\n\t\t<sup>[Lutz]<\/sup>\n\t\tMark Lutz.\n\t\t<em class=\"title\">Programming Python.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1996.\n\t\tISBN 1-56592-197-6.\n\t<\/li>\n\t<li id=\"McIlroy78\">\n\t\t<sup>[McIlroy78]<\/sup><em>The Bell System Technical Journal<\/em>.\n\t\tBell Laboratories.\n\t\t\n\t\t\n\t\tM. D. McIlroy, E. N. Pinson, and B. A. Tague.\n\t\t\u201cUnix Time-Sharing System Forward\u201d.\n\t\t1978.\n\t\t\n\t\t57 (6, part 2).\n\t\t\n\t\t\n\t\tp. 1902.\n\t\t\n\t\t\n\t<\/li>\n\t<li id=\"McIlroy91\">\n\t\t<sup>[McIlroy91]<\/sup><em>Proc. Virginia Computer Users Conference<\/em>.\n\t\t\n\t\tVolume 21.\n\t\t\n\t\t\n\t\t\n\t\tM. D. McIlroy.\n\t\t\u201cUnix on My Mind\u201d.\n\t\t\n\t\tp. 1-6.\n\t\t\n\t\t\n\t<\/li>\n\t<li id=\"Miller\">\n\t\t<sup>[Miller]<\/sup><em>The Psychological Review<\/em>.\n\t\t\n\t\t\n\t\tGeorge Miller.\n\t\t\u201cThe Magical Number Seven, Plus or Minus Two\u201d.\n\t\t\n\t\tSome limits on our capacity for processing information.\n\t\t\n\t\t1956.\n\t\t\n\t\t63.\n\t\t\n\t\t\n\t\tpp. 81-97.\n\t\t\n\t\t<a href=\"http:\/\/www.well.com\/user\/smalin\/miller.html\">\n\t\t\tAvailable on the\n\t\tWeb<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Mumon\">\n\t\t<sup>[Mumon]<\/sup>\n\t\tMumon.\n\t\t<em class=\"title\">The Gateless Gate.<\/em>\n\t\t<a href=\"http:\/\/www.ibiblio.org\/zen\/cgi-bin\/koan-index.pl\">\n\t\tA good modern translation is available on the Web<\/a>.\n\t<\/li>\n\t<li id=\"OpenSources\">\n\t\t<sup>[OpenSources]<\/sup>\n\t\tSam Ockman and Chris DiBona.\n\t\t<em class=\"title\">Open Sources: Voices from the Open Source Revolution.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1999.\n\t\tISBN 1-56592-582-3.\n\t\t\n\t\t280pp..\n\t\t\n\t\t<a href=\"http:\/\/www.oreilly.com\/catalog\/opensources\/book\/toc.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Oram-Talbot\">\n\t\t<sup>[Oram-Talbot]<\/sup>\n\t\tAndrew Oram and Steve Talbot.\n\t\t<em class=\"title\">Managing Projects with Make.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t1991.\n\t\tISBN 0-937175-90-0.\n\t<\/li>\n\t<li id=\"Ousterhout94\">\n\t\t<sup>[Ousterhout94]<\/sup>\n\t\tJohn Ousterhout.\n\t\t<em class=\"title\">Tcl and the Tk Toolkit.<\/em>\n\t\tAddison-Wesley.\n\t\t1994.\n\t\tISBN 0-201-63337-X.\n\t<\/li>\n\t<li id=\"Ousterhout96\">\n\t\t<sup>[Ousterhout96]<\/sup>\n\t\tJohn Ousterhout.\n\t\t<em class=\"title\">Why Threads Are a Bad Idea (for most purposes).<\/em>\n\t\t1996.\n\t\tAn invited talk at USENIX 1996. There is no written paper\n\t\tthat corresponds to it, but the slide presentation is\n\t\t<a href=\"http:\/\/home.pacbell.net\/ouster\/threads.pdf\">\n\t\tavailable on\n\t\tthe Web<\/a>.\n\t<\/li>\n\t<li id=\"Padlipsky\">\n\t\t<sup>[Padlipsky]<\/sup>\n\t\tMichael Padlipsky.\n\t\t<em class=\"title\">The Elements of Networking Style.<\/em>\n\t\tiUniverse.com.\n\t\t2000.\n\t\tISBN 0-595-08879-1.\n\t<\/li>\n\t<li id=\"Parnas\">\n\t\t<sup>[Parnas]<\/sup><em>Communications of the ACM<\/em>.\n\t\t\n\t\t\n\t\tParnas L. David.\n\t\t\u201cOn the Criteria to Be Used in Decomposing Systems into Modules\u201d.\n\t\tAvailable on the Web at the\n\t\t<a href=\"http:\/\/www.acm.org\/classics\/may96\/\">\n\t\t\tACM Classics\n\t\tpage<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Pike\">\n\t\t<sup>[Pike]<\/sup>\n\t\tRob Pike.\n\t\t<em class=\"title\">Notes on Programming in C.<\/em>\n\t\tThis\n\t\tdocument is popular on the Web; a title search is sure to find\n\t\tseveral copies.\n\t\t<a href=\"http:\/\/www.lysator.liu.se\/c\/pikestyle.html\">\n\t\tHere is one<\/a>.\n\t<\/li>\n\t<li id=\"Prechelt\">\n\t\t<sup>[Prechelt]<\/sup>\n\t\tLutz Prechelt.\n\t\t<em class=\"title\"><a href=\"http:\/\/www.ubka.uni-karlsruhe.de\/cgi-bin\/psview?document=ira\/2000\/5\"> An Empirical Comparison of C, C++, Java, Perl, Python, Rexx, and Tcl for a Search\/String-Processing Program <\/a>.<\/em>\n\t<\/li>\n\t<li id=\"Raskin\">\n\t\t<sup>[Raskin]<\/sup>\n\t\tJef Raskin.\n\t\t<em class=\"title\">The Humane Interface.<\/em>\n\t\tAddison-Wesley.\n\t\t2000.\n\t\tISBN 0-201-37937-6.\n\t\tA summary is\n\t\t<a href=\"http:\/\/humane.sourceforge.net\/humane_interface\/summary_of_thi.html\">\n\t\tavailable\n\t\ton the Web<\/a>.\n\t<\/li>\n\t<li id=\"Ravenbrook\">\n\t\t<sup>[Ravenbrook]<\/sup><em class=\"title\">The Memory Management Reference.<\/em>\n\t\t<a href=\"http:\/\/www.memorymanagement.org\/\">\n\t\tAvailable on\n\t\tthe Web<\/a>.\n\t<\/li>\n\t<li id=\"Raymond96\">\n\t\t<sup>[Raymond96]<\/sup>\n\t\t\n\t\tEric S. Raymond.\n\t\t\n\t\t<em class=\"title\">The New Hacker's Dictionary.<\/em>\n\t\t3rd Edition.\n\t\t1996.\n\t\tISBN 0-262-68092-0.\n\t\tMIT Press.\n\t\t\n\t\t547pp..\n\t\t\n\t\tAvailable on the Web at\n\t\t<a href=\"http:\/\/www.catb.org\/~esr\/jargon\">\n\t\tJargon File\n\t\tResource Page<\/a>.\n\t<\/li>\n\t<li id=\"Raymond01\">\n\t\t<sup>[Raymond01]<\/sup>\n\t\tEric S. Raymond.\n\t\t<em class=\"title\">The Cathedral and the Bazaar.<\/em>\n\t\t2nd Edition.\n\t\t1999.\n\t\tISBN 0-596-00131-2.\n\t\tO'Reilly &amp; Associates.\n\t\t\n\t\t240pp..\n\t\t\n\t<\/li>\n\t<li id=\"Reps-Senzaki\">\n\t\t<sup>[Reps-Senzaki]<\/sup>\n\t\tPaul Reps and Nyogen Senzaki.\n\t\t<em class=\"title\">Zen Flesh, Zen Bones.<\/em>\n\t\t1994.\n\t\tShambhala Publications.\n\t\tISBN 1-570-62063-6.\n\t\t\n\t\t285pp..\n\t\t\n\t\tA superb anthology of Zen primary sources, presented just as\n\t\tthey are.\n\t<\/li>\n\t<li id=\"Ritchie79\">\n\t\t<sup>[Ritchie79]<\/sup>\n\t\tDennis M. Ritchie.\n\t\t<em class=\"title\">The Evolution of the Unix Time-Sharing System.<\/em>\n\t\t1979.\n\t\t<a href=\"http:\/\/cm.bell-labs.com\/cm\/cs\/who\/dmr\/hist.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Ritchie93\">\n\t\t<sup>[Ritchie93]<\/sup>\n\t\tDennis M. Ritchie.\n\t\t<em class=\"title\">The Development of the C Language.<\/em>\n\t\t1993.\n\t\t<a href=\"http:\/\/cm.bell-labs.com\/cm\/cs\/who\/dmr\/chist.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"RitchieQED\">\n\t\t<sup>[RitchieQED]<\/sup>\n\t\tDennis M. Ritchie.\n\t\t<em class=\"title\">An Incomplete History of the QED Text Editor.<\/em>\n\t\t2003.\n\t\t<a href=\"http:\/\/cm.bell-labs.com\/cm\/cs\/who\/dmr\/qed.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Ritchie-Thompson\">\n\t\t<sup>[Ritchie-Thompson]<\/sup><em class=\"title\">The Unix Time-Sharing System.<\/em>\n\t\tDennis M. Ritchie and Ken Thompson.\n\t\t<a href=\"http:\/\/cm.bell-labs.com\/cm\/cs\/who\/dmr\/cacm.html\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Saltzer\">\n\t\t<sup>[Saltzer]<\/sup><em>ACM Transactions on Computer Systems<\/em>.\n\t\tAssociation for Computing Machinery.\n\t\t\n\t\t\n\t\tJames. H. Saltzer, David P. Reed, and David D. Clark.\n\t\t\u201cEnd-to-End Arguments in System Design\u201d.\n\t\tNovember 1984.\n\t\t<a href=\"http:\/\/web.mit.edu\/Saltzer\/www\/publications\/endtoend\/endtoend.pdf\">\n\t\t\tAvailable on the Web<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Salus\">\n\t\t<sup>[Salus]<\/sup>\n\t\tPeter H. Salus.\n\t\t<em class=\"title\">A Quarter-Century of Unix.<\/em>\n\t\tAddison-Wesley.\n\t\t1994.\n\t\tISBN 0-201-54777-5.\n\t\tAn excellent overview of Unix history, explaining many of the\n\t\tdesign decisions in the words of the people who made them.\n\t<\/li>\n\t<li id=\"Schaffer-Wolf\">\n\t\t<sup>[Schaffer-Wolf]<\/sup>\n\t\tEvan Schaffer and Mike Wolf.\n\t\t<em class=\"title\">The Unix Shell as a Fourth-Generation Language.<\/em>\n\t\t1991.\n\t\t<a href=\"http:\/\/www.rdb.com\/lib\/4gl.pdf\">\n\t\tAvailable on the Web<\/a>.  An open-source implementation,\n\t\t<em>NoSQL<\/em>, is available and readily turned up\n\tby a Web search.\n\t<\/li>\n\t<li id=\"Schwartz-Christiansen\">\n\t\t<sup>[Schwartz-Christiansen]<\/sup>\n\t\tRandal Schwartz and Tom Phoenix.\n\t\t<em class=\"title\">Learning Perl.<\/em>\n\t\t3rd Edition.\n\t\tO'Reilly &amp; Associates.\n\t\t2001.\n\t\tISBN 0-596-00132-0.\n\t<\/li>\n\t<li id=\"Spinellis\">\n\t\t<sup>[Spinellis]<\/sup><em>Journal of Systems and Software<\/em>.\n\t\t\n\t\t\n\t\tDiomidis Spinellis.\n\t\t\u201cNotable Design Patterns for Domain-Specific Languages\u201d.\n\t\t\n\t\t56.\n\t\t\n\t\t\n\t\t(1).\n\t\t\n\t\tFebruary 2001.\n\t\t\n\t\tp. 91-99.\n\t\t\n\t\t\n\t<\/li>\n\t<li id=\"Stallman\">\n\t\t<sup>[Stallman]<\/sup>\n\t\tRichard M. Stallman.\n\t\t<em class=\"title\">The GNU Manifesto.<\/em>\n\t\t<a href=\"http:\/\/www.gnu.org\/gnu\/manifesto.html\">\n\t\tAvailable on\n\t\tthe Web<\/a>.\n\t<\/li>\n\t<li id=\"Stephenson\">\n\t\t<sup>[Stephenson]<\/sup>\n\t\tNeal Stephenson.\n\t\t<em class=\"title\">In the Beginning Was the Command Line.<\/em>\n\t\t1999.\n\t\t<a href=\"http:\/\/www.cryptonomicon.com\/beginning.html\">\n\t\tAvailable on the Web<\/a>,\n\t\tand also as a trade paperback from Avon Books.\n\t<\/li>\n\t<li id=\"Stevens90\">\n\t\t<sup>[Stevens90]<\/sup>\n\t\tW. Richard Stevens.\n\t\t<em class=\"title\">Unix Network Programming.<\/em>\n\t\tPrentice-Hall.\n\t\t1990.\n\t\tISBN 0-13-949876-1.\n\t\tThe classic on this topic. Note: Some later editions of this\n\t\tbook omit coverage of the Version 6 networking facilities like\n\t\t<em>mx()<\/em>.\n\t<\/li>\n\t<li id=\"Stevens92\">\n\t\t<sup>[Stevens92]<\/sup>\n\t\tW. Richard Stevens.\n\t\t<em class=\"title\">Advanced Programming in the Unix Environment.<\/em>\n\t\t1992.\n\t\tISBN 0-201-56317-7.\n\t\tAddison-Wesley.\n\t\tStevens's comprehensive guide to the Unix API.  A feast for\n\t\tthe experienced programmer or the bright novice, and a worthy \n\t\tcompanion to\n\t\t<em>Unix Network Programming<\/em>.\n\t<\/li>\n\t<li id=\"Stroustrup\">\n\t\t<sup>[Stroustrup]<\/sup>\n\t\tBjarne Stroustrup.\n\t\t<em class=\"title\">The C++ Programming Language.<\/em>\n\t\tAddison-Wesley.\n\t\t1991.\n\t\tISBN 0-201-53992-6.\n\t<\/li>\n\t<li id=\"Tanenbaum-VanRenesse\">\n\t\t<sup>[Tanenbaum-VanRenesse]<\/sup>\n\t\tAndrew S. Tanenbaum and Robbert van Renesse.\n\t\t<em class=\"title\">A Critique of the Remote Procedure Call Paradigm.<\/em>\n\t\tEUTECO'88 Proceedings, Participants Edition.\n\t\t1988.\n\t\t\n\t\tpp. 775-783.\n\t\t\n\t<\/li>\n\t<li id=\"Tidwell\">\n\t\t<sup>[Tidwell]<\/sup>\n\t\tDoug Tidwell.\n\t\t<em class=\"title\">XSLT: Mastering XML Transformations.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t2001.\n\t\tISBN 1-596-00053-7.\n\t<\/li>\n\t<li id=\"Torvalds\">\n\t\t<sup>[Torvalds]<\/sup>\n\t\tLinus Torvalds and David Diamond.\n\t\t<em class=\"title\">Just for Fun.<\/em>\n\t\t\n\t\tThe Story of an Accidental Revolutionary.\n\t\t\n\t\tHarperBusiness.\n\t\t2001.\n\t\tISBN 0-06-662072-4.\n\t<\/li>\n\t<li id=\"Vaughan\">\n\t\t<sup>[Vaughan]<\/sup>\n\t\tGary V. Vaughan, Tom Tromey, and Ian Lance Taylor.\n\t\t<em class=\"title\">GNU Autoconf, Automake, and Libtool.<\/em>\n\t\tNew Riders Publishing.\n\t\t2000.\n\t\t\n\t\t390 p..\n\t\t\n\t\tISBN 1-578-70190-2.\n\t\tA user's guide to the GNU autoconfiguration tools.\n\t\t<a href=\"http:\/\/sources.redhat.com\/autobook\/\">\n\t\tAvailable on the Web<\/a>.\n\t<\/li>\n\t<li id=\"Vo\">\n\t\t<sup>[Vo]<\/sup><em>Software Practice &amp; Experience<\/em>.\n\t\t\n\t\t\n\t\tKiem-Phong Vo.\n\t\t\u201cThe Discipline and Method Architecture for Reusable Libraries\u201d.\n\t\t2000.\n\t\t\n\t\t30.\n\t\t\n\t\t\n\t\tp. 107-128.\n\t\t\n\t\t<a href=\"http:\/\/www.research.att.com\/sw\/tools\/vcodex\/dm-spe.ps\">\n\t\t\tAvailable\n\t\ton the Web<\/a>.\n\t\t\n\t<\/li>\n\t<li id=\"Wall2000\">\n\t\t<sup>[Wall2000]<\/sup>\n\t\tLarry Wall, Tom Christiansen, and Jon Orwant.\n\t\t<em class=\"title\">Programming Perl.<\/em>\n\t\t3rd Edition.\n\t\tO'Reilly &amp; Associates.\n\t\t2000.\n\t\tISBN  0-596-00027-8.\n\t<\/li>\n\t<li id=\"Welch\">\n\t\t<sup>[Welch]<\/sup>\n\t\tBrent Welch.\n\t\t<em class=\"title\">Practical Programming in Tcl and Tk.<\/em>\n\t\tPrentice-Hall.\n\t\t1999.\n\t\tISBN 0-13-022028-0.\n\t<\/li>\n\t<li id=\"Williams\">\n\t\t<sup>[Williams]<\/sup>\n\t\tSam Williams.\n\t\t<em class=\"title\">Free as in Freedom.<\/em>\n\t\tO'Reilly &amp; Associates.\n\t\t2002.\n\t\tISBN 0-596-00287-4.\n\t\t<a href=\"http:\/\/www.oreilly.com\/openbook\/freedom\/index.html\">\n\t\tAvailable\n\t\ton the Web<\/a>.\n\t<\/li>\n\t<li id=\"Yourdon\">\n\t\t<sup>[Yourdon]<\/sup>\n\t\tEdward Yourdon.\n\t\t<em class=\"title\">Death March.<\/em>\n\t\t\n\t\tThe Complete Software Developer's Guide to Surviving \u201cMission Impossible\u201d Projects.\n\t\t\n\t\tPrentice-Hall.\n\t\t1997.\n\t\tISBN 0-137-48310-4.\n\t<\/li>\n<\/ul><h3 id=\"apc\"><a href=\"#apc\">\u00a7<\/a>Appendix C. Contributors<\/h3>\n<blockquote class=\"epigraph\">\n\t<p>Anyone who has attended a USENIX conference in a fancy hotel can tell you that a sentence like \u201cYou're one of\n\tthose computer people, aren't you?\u201d is roughly equivalent to \u201cLook, another amazingly mobile form of slime mold!\u201d in\n\tthe mouth of a hotel cocktail waitress.<\/p><div>\n\t\u2013 Elizabeth Zwicky\n<\/div><\/blockquote>\n  <p>\n   <em>Ken Arnold<\/em>\n   was part of the group that created the 4BSD Unix releases.  He wrote the original curses(3) library and was one of\n   the authors of the original rogue(6) game. He is a co-author of the <em>Java Reference Manual<\/em>, and one of the\n   leading experts on Java and OO techniques.\n  <\/p>\n  <p>\n   <em>Steven M. Bellovin<\/em> created Usenet (with Tom Truscott and Jim Ellis) while at University of North Carolina in\n   1979. In 1982 he joined AT&amp;T Bell Laboratories, where he has done pioneering research in security, cryptography,\n   and networking on Unix systems and elsewhere.  He is an active member of the Internet Engineering Task Force, and a\n   member of the National Academy of Engineering.\n  <\/p>\n  <p>\n   <em>Stuart Feldman<\/em> was a member of the Bell Labs Unix development group.  He wrote make(1) and f77(1). He is now\n   the vice-president in charge of computing research at IBM.\n  <\/p>\n  <p>\n   <em>Jim Gettys<\/em> was, with Bob Scheifler and Keith Packard, one of the principal architects of the X windowing\n   system in the late 1980s.  He wrote much of the X library, the X license, and articulated the \u201cmechanism, not policy\u201d\n   central credo of the X design.\n  <\/p>\n  <p>\n   <em>Steve Johnson<\/em> wrote yacc(1) and then used it to write the Portable C Compiler, which replaced the original DMR C and became the ancestor of most later Unix C compilers.\n  <\/p>\n  <p>\n   <em>Brian Kernighan<\/em> has been the Unix community's single most articulate exponent of good style.  He has\n   authored or coauthored several books that are indispensable classics of the tradition, including <em>The Practice of\n   Programming<\/em>, <em>The C Programming Language<\/em>, <em>The Unix Programming Environment<\/em>.  While at Bell\n   Labs, he coauthored the awk(1) language and had a major hand in the development of the troff family of tools,\n   including eqn(1) (co-written with Lorinda Cherry), pic(1), and grap(1) (Jon Bentley).\n  <\/p>\n  <p>\n   <em>David Korn<\/em> wrote the Korn shell, the stylistic ancestor of almost all modern Unix shell designs.  He created\n   UWIN, a UNIX emulator for those that are forced to use Windows.  David has also done research in the design of file\n   systems and tools for source-code portability.\n  <\/p>\n  <p>\n   <em>Mike Lesk<\/em> was part of the original Unix crew at Bell Labs.  Among other contributions, he wrote the\n   <em>ms<\/em> macro package, the tbl(1) and refer(1) tools for word processing, the lex(1) lexical-analyzer generator,\n   and <em>UUCP<\/em> (Unix-to-Unix copy program).\n  <\/p>\n  <p>\n   <em>Doug McIlroy<\/em> headed the research group at Bell Labs where Unix was born and invented the Unix pipe.  He\n   wrote spell(1), diff(1), sort(1), join(1), tr(1), and other classic Unix tools, and helped define the traditional\n   style of Unix documentation.  He has also done pioneering work in storage-allocation algorithms, computer security,\n   and theorem-proving.\n  <\/p>\n  <p>\n   <em>Marshall Kirk McKusick<\/em> implemented the 4.2BSD fast file system and was the Research Computer Scientist at the Berkeley Computer Systems Research Group (CSRG) overseeing the development and release of 4.3BSD and 4.4BSD.\n  <\/p>\n  <p>\n   <em>Keith Packard<\/em> was a major contributor to the original X11 code.  During a second phase of involvement\n   beginning in 1999, Keith rewrote the X rendering code, producing a more powerful but dramatically smaller\n   implementation suitable for handheld computers and PDAs.\n  <\/p>\n  <p>\n   <em>Eric S. Raymond<\/em> has been writing Unix software since 1982. In 1991 he edited <em>The New Hacker's\n   Dictionary<\/em>, and has since been studying the Unix community and the Internet hacker culture from a historical and\n   anthropological perspective.  In 1997 that study produced <em>The Cathedral and the Bazaar<\/em>, that helped\n   (re)define and energize the open-source movement.  He presently maintains more than thirty open-source software\n   projects and about a dozen key FAQ documents.\n  <\/p>\n  <p>\n   <em>Henry Spencer<\/em> was a leader among the first wave of programmers to embrace Unix when it escaped from Bell\n   Labs in the mid-1970s.  His contributions included the public-domain getopt(3), the first open-source string library,\n   and an open-source regular-expression engine which found use in 4.4BSD and as the reference for the POSIX standard.\n   He is also a noted expert on the arcana of C. He was a coauthor of C News, and has for many years been a voice of\n   reason on Usenet and one of its most respected regulars.\n  <\/p>\n  <p>\n   <em>Ken Thompson<\/em> invented Unix.\n  <\/p>\n\n<h2 id=\"revhistory\"><a href=\"#revhistory\">\u00a7<\/a>Revision History<\/h2>\n<strong>Revision 1.1<\/strong> \u2013 15 January 2019 \u2013 Martin Tournoij &lt;martin@arp242.net&gt;<br \/>Reformat as single HTML docment with more modern styling.<br \/><strong>Revision 1.0<\/strong> \u2013 19 September 2003 \u2013 esr<br \/>This is the content that went to Addison-Wesley's printers.<br \/><strong>Revision 0.4<\/strong> \u2013 5 February 2003 \u2013 esr<br \/>Release for public review.<br \/><strong>Revision 0.3<\/strong> \u2013 22 January 2003 \u2013 esr<br \/>First eighteen-chapter draft. Manuscript walkthrough at Chapter 12. Limited release for early reviewers.<br \/><strong>Revision 0.2<\/strong> \u2013 2 January 2003 \u2013 esr<br \/>First manuscript walkthrough at Chapter 7. Released to Dmitry Kirsanov at AW production.<br \/><strong>Revision 0.1<\/strong> \u2013 16 November 2002 \u2013 esr<br \/>First DocBook draft, fifteen chapters. Languages rewritten to incorporate lots of feedback. Transparency, Shipped to Mark Taub at AW.<br \/><strong>Revision 0.0<\/strong> \u2013 1999 \u2013 esr<br \/>Public HTML draft, first four chapters only.<br \/><h3 id=\"unix_koans\"><a href=\"#unix_koans\">\u00a7<\/a>Appendix D. Rootless Root<\/h3>\n<div class=\"center\"><img alt=\"\" src=\"denied:data:image\/png;base64,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\" \/><\/div>\n\n<p>\n  The discovery of the collection of koans known as the <em>Rootless Root<\/em>, apparently preserved for decades in the\n  dry upper air of the Western Mountains, has ignited great controversy in scholarly circles.  Are these authentic\n  documents shedding new light on the teaching of the early Unix patriarchs?  Or are they clever pastiches from a later\n  age, commanding the authority of semi-mythical figures such as the Patriarchs Thompson, Ritchie and McIlroy for\n  doctrines which evolved closer to our own era?\n<\/p>\n<p>\n  It is impossible to say for certain.  All sides in the dispute\nhave made much of an alleged similarity to that venerable classic,\n  <em>The Tao of Programming<\/em>.<sup><a href=\"#ftn.id3033994\" id=\"id3033994\">[165]<\/a><\/sup>\n  But\n  <em>Rootless Root<\/em> is\nquite different in tone and style from the loose, poetic anecdotes of\nthe James translation, focused as it is on the remarkable and\nenigmatic figure of Master Foo.\n<\/p>\n<p>\n  It would be more apposite to seek parallels in the\n  <em>AI Koans<\/em>;<sup><a href=\"#ftn.id3034026\" id=\"id3034026\">[166]<\/a><\/sup>\n  indeed,\nthere are textual clues that the author of the\n  <em>Rootless Root<\/em>\n  may have redacted certain versions of the\n  <em>AI Koans<\/em>. We are also on much firmer ground in\nseeking connections with the Loginataka;<sup><a href=\"#ftn.id3034054\" id=\"id3034054\">[167]<\/a><\/sup>\n  indeed, it is\nquite possible that the unknown authors of\n  <em>Rootless Root<\/em>\n  and of\n  <em>Loginataka<\/em> were one and\nthe same person, possibly a student of Master Foo himself.\n<\/p>\n<p>\n  Mention should also be made of the\n  <em>Tales of Zen Master Greg<\/em>,<sup><a href=\"#ftn.id3034091\" id=\"id3034091\">[168]<\/a><\/sup>\n  though\nthe Nine Inch Nails references have cast some doubt on their antiquity\nand it is thus unlikely that they influenced\n  <em>Rootless Root<\/em>.\n<\/p>\n<p>\n  That the title of the work was intended as a reference to the\nZen classic\n  <em>Gateless Gate<\/em> <sup><a href=\"#ftn.id3034123\" id=\"id3034123\">[169]<\/a><\/sup>\n  of Mumon, we can say with fair confidence.  There are echoes of Mumon\nin several of the koans.\n<\/p>\n<p>\n  There is considerable dispute over whether Master Foo should \nbe assigned to the Eastern (New Jersey) School, or the Western\nSchool that grew out of the Patriarch Thompson's epochal early journey\nto Berkeley.  If this question has not been settled, it is perhaps\nbecause we cannot even establish that Master Foo ever existed!  He\nmight merely be a composite of a group of teachers, or of an entire dharma \nlineage.\n<\/p>\n<p>\n  Even supposing the legend of Master Foo accreted around the\nteaching of some a single person, what of his favored student\nNubi?  Nubi has all the earmarks of a stock figure, the perfect\ndisciple.  One is reminded of the tales surrounding the Buddha's \nfavorite follower Ananda.  It seems likely that there was a\nhistorical Ananda, but no trace of his actual personality has\nsurvived the euhemerizing process by which the life of the Buddha\nwas polished into timeless myth.\n<\/p>\n<p>\n  In the end, all we can do is take these teaching stories on\ntheir own terms, and extract what kernels of wisdom may be found\nthere.\n<\/p>\n<p>\n  The redaction of the\n  <em>Rootless Root<\/em> is a\nwork in progress, as the source materials present many difficulties \nin reconstruction and interpretation. Future versions may include \nmore stories as these difficulties are overcome.\n<\/p>\n<div class=\"footnotes\"><strong>Footnotes<\/strong><p><sup><a href=\"#id3033994\" id=\"ftn.id3033994\">[165]<\/a><\/sup><em>The Tao of Programming<\/em>\n  is available\n  <a href=\"http:\/\/www.canonical.org\/~kragen\/tao-of-programming.html\">\n   on the Web<\/a>.\n<\/p><p><sup><a href=\"#id3034026\" id=\"ftn.id3034026\">[166]<\/a><\/sup>\n  The\n  <em>AI Koans<\/em> are available\n  <a href=\"http:\/\/www.catb.org\/~esr\/jargon\/html\/Some-AI-Koans.html\">\n   on the Web<\/a>.\n<\/p><p><sup><a href=\"#id3034054\" id=\"ftn.id3034054\">[167]<\/a><\/sup>\n  The\n  <em>Loginataka<\/em> is available\n  <a href=\"http:\/\/www.catb.org\/~esr\/faqs\/loginataka.html\">\n   on the Web<\/a>.\n<\/p><p><sup><a href=\"#id3034091\" id=\"ftn.id3034091\">[168]<\/a><\/sup>\n  The\n  <em>Tales of Zen Master Greg<\/em>\n  are available\n  <a href=\"http:\/\/www.gu.uwa.edu.au\/users\/greg\/\">\n   on the Web<\/a>.\n<\/p><p><sup><a href=\"#id3034123\" id=\"ftn.id3034123\">[169]<\/a><\/sup>\n  The\n  <em>Gateless Gate<\/em> is available\n  <a href=\"http:\/\/www.ibiblio.org\/zen\/cgi-bin\/koan-index.pl\">\n   on the Web<\/a>.\n<\/p><\/div>\n\n<h3 id=\"ten-thousand\"><a href=\"#ten-thousand\">\u00a7<\/a>Master Foo and the Ten Thousand Lines<\/h3>\n<p>\n  Master Foo once said to a visiting programmer: \u201cThere is\nmore Unix-nature in one line of shell script than there is in ten thousand\nlines of C\u201d.\n<\/p>\n<p>\n  The programmer, who was very proud of his mastery of C, said:\n\u201cHow can this be?  C is the language in which the very kernel of\nUnix is implemented!\u201d\n<\/p>\n<p>\n  Master Foo replied: \u201cThat is so. Nevertheless, there is\nmore Unix-nature in one line of shell script than there is in ten thousand\nlines of C\u201d.\n<\/p>\n<p>\n  The programmer grew distressed.  \u201cBut through the C\nlanguage we experience the enlightenment of the Patriarch Ritchie!  We\nbecome as one with the operating system and the machine, reaping\nmatchless performance!\u201d\n<\/p>\n<p>\n  Master Foo replied: \u201cAll that you say is true. But there\nis still more Unix-nature in one line of shell script than there is in ten\nthousand lines of C\u201d.\n<\/p>\n<p>\n  The programmer scoffed at Master Foo and rose to depart.  But\nMaster Foo nodded to his student Nubi, who wrote a line of shell\nscript on a nearby whiteboard, and said: \u201cMaster programmer,\nconsider this pipeline.  Implemented in pure C, would it not span ten\nthousand lines?\u201d\n<\/p>\n<p>\n  The programmer muttered through his beard, contemplating what\nNubi had written.  Finally he agreed that it was so.\n<\/p>\n<p>\n  \u201cAnd how many hours would you require to implement and\ndebug that C program?\u201d asked Nubi.\n<\/p>\n<p>\n  \u201cMany\u201d, admitted the visiting programmer.\n\u201cBut only a fool would spend the time to do that when so many\nmore worthy tasks await him\u201d.\n<\/p>\n<p>\n  \u201cAnd who better understands the Unix-nature?\u201d\nMaster Foo asked. \u201cIs it he who writes the ten thousand\nlines, or he who, perceiving the emptiness of the task, gains merit by\nnot coding?\u201d\n<\/p>\n<p>\n  Upon hearing this, the programmer was enlightened.\n<\/p>\n\n<h3 id=\"script-kiddie\"><a href=\"#script-kiddie\">\u00a7<\/a>Master Foo and the Script Kiddie<\/h3>\n<p>\n  A stranger from the land of Woot came to Master Foo as he was\neating the morning meal with his students.\n<\/p>\n<p>\n  \u201cI hear y00 are very l33t\u201d, he said.\n\u201cPl33z teach m3 all y00 know\u201d.\n<\/p>\n<p>\n  Master Foo's students looked at each other, confused by the\nstranger's barbarous language. Master Foo just smiled and replied:\n\u201cYou wish to learn the Way of Unix?\u201d\n<\/p>\n<p>\n  \u201cI want to b3 a wizard hax0r\u201d, the stranger\nreplied, \u201cand 0wn ever3one's b0xen\u201d.\n<\/p>\n<p>\n  \u201cI do not teach that Way\u201d, replied Master Foo.\n<\/p>\n<p>\n  The stranger grew agitated.  \u201cD00d, y00 r nothing but a\np0ser\u201d, he said.  \u201cIf y00 n00 anything, y00 wud t33ch\nm3\u201d.\n<\/p>\n<p>\n  \u201cThere is a path\u201d, said Master Foo, \u201cthat\nmight bring you to wisdom\u201d.  The master scribbled an IP address on\na piece of paper.  \u201cCracking this box should pose you little\ndifficulty, as its guardians are incompetent.  Return and tell me \nwhat you find\u201d.\n<\/p>\n<p>\n  The stranger bowed and left. Master Foo finished his meal.\n<\/p>\n<p>\n  Days passed, then months.  The stranger was forgotten.\n<\/p>\n<p>\n  Years later, the stranger from the land of Woot returned.\n<\/p>\n<p>\n  \u201cDamn you!\u201d he said, \u201cI cracked that box, and\nit was easy like you said.  But I got busted by the FBI and thrown\nin jail\u201d.\n<\/p>\n<p>\n  \u201cGood\u201d, said Master Foo.  \u201cYou are ready for\nthe next lesson\u201d.  He scribbled an IP address on another piece\nof paper and handed it to the stranger.\n<\/p>\n<p>\n  \u201cAre you\n  <em>crazy<\/em>\n  ?\u201d the stranger\nyelled. \u201cAfter what I've been through, I'm never going to break\ninto a computer again!\u201d\n<\/p>\n<p>\n  Master Foo smiled.  \u201cHere\u201d, he said, \u201cis the\nbeginning of wisdom\u201d.\n<\/p>\n<p>\n  On hearing this, the stranger was enlightened.\n<\/p>\n\n<h3 id=\"two_paths\"><a href=\"#two_paths\">\u00a7<\/a>Master Foo Discourses on the Two Paths<\/h3>\n<p>\n  Master Foo instructed his students:\n<\/p>\n<p>\n  \u201cThere is a line of dharma teaching, exemplified by the\nPatriarch McIlroy's mantra \u2018Do one thing well\u2019, which\nemphasizes that software partakes of the Unix way when it has simple\nand consistent behavior, with properties that can be readily modeled by\nthe mind of the user and used by other programs\u201d.\n<\/p>\n<p>\n  \u201cBut there is another line of dharma teaching, exemplified by the\nPatriarch Thompson's great mantra \u2018When in doubt, use brute\nforce\u2019, and various sutras on the value of getting 90% of cases\nright\n  <em>now<\/em>, rather than 100%\n  <em>later<\/em>, which emphasizes robustness and simplicity of\nimplementation\u201d.\n<\/p>\n<p>\n  \u201cNow tell me: which programs have the Unix nature?\u201d\n<\/p>\n<p>\n  After a silence, Nubi observed:\n<\/p>\n<p>\n  \u201cMaster, these teachings may conflict\u201d.\n<\/p>\n<p>\n  \u201cA simple implementation is likely to lack logic for edge\ncases, such as resource exhaustion, or failure to close a race window,\nor a timeout during an uncompleted transaction\u201d.\n<\/p>\n<p>\n  \u201cWhen such edge cases occur, the behavior of the software\nwill become irregular and difficult.  Surely this is not the Way of\nUnix?\u201d\n<\/p>\n<p>\n  Master Foo nodded in agreement.\n<\/p>\n<p>\n  \u201cOn the other hand, it is well known that fancy algorithms\nare brittle.  Further, each attempt to cover an edge case tends to\ninteract with both the program's central algorithms and the code\ncovering other edge cases\u201d.\n<\/p>\n<p>\n  \u201cThus, attempts to cover all edge cases in advance, guaranteeing\n\u2018simplicity of description\u2019, may in fact produce code that is\novercomplicated and brittle or which, plagued by bugs, never ships at all.\nSurely this is not the Way of Unix?\u201d\n<\/p>\n<p>\n  Master Foo nodded in agreement.\n<\/p>\n<p>\n  \u201cWhat, then, is the proper dharma path?\u201d asked Nubi.\n<\/p>\n<p>\n  The master spoke:\n<\/p>\n<p>\n  \u201cWhen the eagle flies, does it forget that its feet have\ntouched the ground?  When the tiger lands upon its prey, does it\nforget its moment in the air? Three pounds of VAX!\u201d\n<\/p>\n<p>\n  On hearing this, Nubi was enlightened.\n<\/p>\n\n<h3 id=\"methodology-consultant\"><a href=\"#methodology-consultant\">\u00a7<\/a>Master Foo and theMethodologist<\/h3>\n<p>\n  When Master Foo and his student Nubi journeyed among the sacred\nsites, it was the Master's custom in the evenings to offer public\ninstruction to Unix neophytes of the towns and villages in which they\nstopped for the night.\n<\/p>\n<p>\n  On one such occasion, a methodologist was among those\nwho gathered to listen.\n<\/p>\n<p>\n  \u201cIf you do not repeatedly profile your code for hot spots\nwhile tuning, you will be like a fisherman who casts his net in an\nempty lake\u201d, said Master Foo.\n<\/p>\n<p>\n  \u201cIs it not, then, also true\u201d, said the methodology\nconsultant, \u201cthat if you do not continually measure your\nproductivity while managing resources, you will be like a fisherman who\ncasts his net in an empty lake?\u201d\n<\/p>\n<p>\n  \u201cI once came upon a fisherman who just at that moment let\nhis net fall in the lake on which his boat was floating\u201d, said\nMaster Foo.  \u201cHe scrabbled around in the bottom of his boat for\nquite a while looking for it\u201d.\n<\/p>\n<p>\n  \u201cBut\u201d, said the methodologist, \u201cif\nhe had dropped his net in the lake, why was he looking in the\nboat?\u201d\n<\/p>\n<p>\n  \u201cBecause he could not swim\u201d, replied Master Foo.\n<\/p>\n<p>\n  Upon hearing this, the methodologist was enlightened.\n<\/p>\n\n<h3 id=\"gui-programmer\"><a href=\"#gui-programmer\">\u00a7<\/a>Master Foo Discourses on theGraphical User Interface<\/h3>\n<p>\n  One evening, Master Foo and Nubi attended a gathering of\nprogrammers who had met to learn from each other.  One of the\nprogrammers asked Nubi to what school he and his master belonged. Upon\nbeing told they were followers of the Great Way of Unix, the\nprogrammer grew scornful.\n<\/p>\n<p>\n  \u201cThe command-line tools of Unix are crude and\nbackward\u201d, he scoffed.  \u201cModern, properly designed\noperating systems do everything through a graphical user\ninterface\u201d.\n<\/p>\n<p>\n  Master Foo said nothing, but pointed at the moon.  A nearby\ndog began to bark at the master's hand.\n<\/p>\n<p>\n  \u201cI don't understand you!\u201d said the programmer.\n<\/p>\n<p>\n  Master Foo remained silent, and pointed at an image of the Buddha.\nThen he pointed at a window.\n<\/p>\n<p>\n  \u201cWhat are you trying to tell me?\u201d asked the\nprogrammer.\n<\/p>\n<p>\n  Master Foo pointed at the programmer's head.\nThen he pointed at a rock.\n<\/p>\n<p>\n  \u201cWhy can't you make yourself clear?\u201d demanded the\nprogrammer.\n<\/p>\n<p>\n  Master Foo frowned thoughtfully, tapped the the programmer\ntwice on the nose, and dropped him in a nearby trashcan.\n<\/p>\n<p>\n  As the programmer was attempting to extricate himself from\nthe garbage, the dog wandered over and piddled on him.\n<\/p>\n<p>\n  At that moment, the programmer achieved enlightenment.\n<\/p>\n\n<h3 id=\"zealot\"><a href=\"#zealot\">\u00a7<\/a>Master Foo and the Unix Zealot<\/h3>\n<p>\n  A Unix zealot, having heard that Master Foo was wise in the\nGreat Way, came to him for instruction.  Master Foo said to him:\n<\/p>\n<p>\n  \u201cWhen the Patriarch Thompson invented Unix, he did not\nunderstand it.  Then he gained in understanding, and no longer invented\nit\u201d.\n<\/p>\n<p>\n  \u201cWhen the Patriarch McIlroy invented the pipe, he knew\nthat it would transform software, but did not know that it would \ntransform mind\u201d.\n<\/p>\n<p>\n  \u201cWhen the Patriarch Ritchie invented C, he condemned \nprogrammers to a thousand hells of buffer overruns, heap corruption,\nand stale-pointer bugs\u201d.\n<\/p>\n<p>\n  \u201cTruly, the Patriarchs were blind and foolish!\u201d\n<\/p>\n<p>\n  The zealot was greatly angered by the Master's words.\n<\/p>\n<p>\n  \u201cThese enlightened ones\u201d, he protested. \u201cgave\nus the Great Way of Unix.  Surely, if we mock them we will lose merit\nand be reborn as beasts or MCSEs\u201d.\n<\/p>\n<p>\n  \u201cIs your code ever completely without stain and\nflaw?\u201d demanded Master Foo.\n<\/p>\n<p>\n  \u201cNo\u201d, admitted the zealot, \u201cno man's\nis\u201d.\n<\/p>\n<p>\n  \u201cThe wisdom of the Patriarchs\u201d, said Master Foo,\n\u201cwas that they\n  <em>knew<\/em>\n  they were\nfools\u201d.\n<\/p>\n<p>\n  Upon hearing this, the zealot was enlightened.\n<\/p>\n\n<h3 id=\"unix-nature\"><a href=\"#unix-nature\">\u00a7<\/a>Master Foo Discourses on the Unix-Nature<\/h3>\n<p>\n  A student said to Master Foo: \u201cWe are told that the firm\ncalled SCO holds true dominion over Unix\u201d.\n<\/p>\n<p>\n  Master Foo nodded.\n<\/p>\n<p>\n  The student continued, \u201cYet we are also told that the firm\ncalled OpenGroup also holds true dominion over Unix\u201d.\n<\/p>\n<p>\n  Master Foo nodded.\n<\/p>\n<p>\n  \u201cHow can this be?\u201d asked the student.\n<\/p>\n<p>\n  Master Foo replied:\n<\/p>\n<p>\n  \u201cSCO indeed has dominion over the code of Unix, but the\ncode of Unix is not Unix.  OpenGroup indeed has dominion over the name\nof Unix, but the name of Unix is not Unix\u201d.\n<\/p>\n<p>\n  \u201cWhat, then, is the Unix-nature?\u201d asked the student.\n<\/p>\n<p>\n  Master Foo replied:\n<\/p>\n<p>\n  \u201cNot code.  Not name.  Not mind.  Not things.  Always changing,\nyet never changing\u201d.\n<\/p>\n<p>\n  \u201cThe Unix-nature is simple and empty.  Because it is\nsimple and empty, it is more powerful than a typhoon\u201d.\n<\/p>\n<p>\n  \u201cMoving in accordance with the law of nature, it unfolds\ninexorably in the minds of programmers, assimilating designs to its\nown nature. All software that would compete with it must become\nlike to it; empty, empty, profoundly empty, perfectly void,\nhail!\u201d\n<\/p>\n<p>\n  Upon hearing this, the student was enlightened.\n<\/p>\n\n<h3 id=\"end-user\"><a href=\"#end-user\">\u00a7<\/a>Master Foo and the End User<\/h3>\n<p>\n  On another occasion when Master Foo gave public instruction, an\nend user, having heard tales of the Master's wisdom, came to him for\nguidance.\n<\/p>\n<p>\n  He bowed three times to Master Foo. \u201cI wish to learn the\nGreat Way of Unix\u201d, he said \u201cbut the command line\nconfuses me\u201d.\n<\/p>\n<p>\n  Some of the onlooking neophytes began to mock the end user,\ncalling him \u201cclueless\u201d and saying that the Way of\nUnix is only for those of discipline and intelligence.\n<\/p>\n<p>\n  The Master held up a hand for silence, and called the most\nobstreperous of the neophytes who had mocked forward, to where\nhe and the end user sat.\n<\/p>\n<p>\n  \u201cTell me\u201d, he asked the neophyte, \u201cof the\ncode you have written and the works of design you have\nuttered\u201d.\n<\/p>\n<p>\n  The neophyte began to stammer out a reply, but fell silent.\n<\/p>\n<p>\n  Master Foo turned to the end-user. \u201cTell me\u201d, he\ninquired, \u201cwhy do you seek the Way?\u201d\n<\/p>\n<p>\n  \u201cI am discontent with the software I see around\nme\u201d, the end user replied. \u201cIt neither performs reliably\nnor pleases the eye and hand.  Having heard that the Unix way, though\ndifficult, is superior, I seek to cast aside all snares and\ndelusions\u201d.\n<\/p>\n<p>\n  \u201cAnd what do you do in the world\u201d, asked Master Foo,\n\u201cthat you must strive with software?\u201d\n<\/p>\n<p>\n  \u201cI am a builder\u201d, the end user replied, \u201cMany\nof the houses of this town were made under my chop\u201d.\n<\/p>\n<p>\n  Master Foo turned back to the neophyte.  \u201cThe housecat may\nmock the tiger\u201d, said the master, \u201cbut doing so will not\nmake his purr into a roar\u201d.\n<\/p>\n<p>\n  Upon hearing this, the neophyte was enlightened.\n<\/p>","created_at":"2019-01-19T14:52:28+0000","updated_at":"2019-01-19T14:52:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":963,"domain_name":"arp242.net","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":780,"title":"The Route of a Text Message","url":"https:\/\/scottbot.net\/the-route-of-a-text-message\/","content":"<p>This is the third post in my <a href=\"https:\/\/scottbot.net\/tag\/full-stack-dev\/\">full-stack dev<\/a> (f-s d) series on the secret life of data. This installment is about a single text message: how it was typed, stored, sent, received, and displayed.  I sprinkle in some history and context to break up the alphabet soup of protocols, but though the piece gets technical, it should all be easily understood.<\/p>\n\n\n\n<p>The first two installments of this series are <a href=\"https:\/\/scottbot.net\/cetus\/\">Cetus<\/a>, about the propagation of errors in a 17th century spreadsheet, and <a href=\"https:\/\/scottbot.net\/down-the-rabbit-hole\/\">Down the Rabbit Hole<\/a>, about the insane lengths one occasionally needs to go through to track down the source of a dataset.<\/p>\n\n\n\n<h2>A Love Story<\/h2>\n\n\n\n<p>My leg involuntarily twitches with vibration\u2014was it my phone, or just a phantom feeling?\u2014and a quick inspection reveals a blinking blue notification. \u201cI love you\u201d, my wife texted me. I walk downstairs to wish her goodnight, because I know the difference between the message and <em>the message<\/em>, you know? <br \/><\/p>\n\n\n\n<p>It\u2019s a bit like encryption, or maybe steganography: anyone can see the text, but only I can decode the hidden data.<br \/><\/p>\n\n\n\n<p>My translation, if we\u2019re being honest, is just one extra link in a remarkably long chain of data events, all to send a message (\u201ccome downstairs and say goodnight\u201d) in under five seconds across about 40 feet. <br \/><\/p>\n\n\n\n<p>The message presumably began somewhere in my wife\u2019s brain and somehow ended up in her thumbs, but that\u2019s a signal for a different story. Ours begins as her thumb taps a translucent screen, one letter at a time, and ends as light strikes my retinas. <\/p>\n\n\n\n<h1>Through the Looking Glass<\/h1>\n\n\n\n<p>With each tap, a small electrical current passes from the screen to her hand. Because electricity flows easily through human bodies, sensors on the phone register a change in voltage wherever her thumb presses against the screen. But the world is messy, and the phone senses random fluctuations in voltage across the rest of the screen, too, so an algorithm determines the biggest, thumbiest-looking voltage fluctuations and assumes that\u2019s where she intended to press.<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh3.googleusercontent.com\/NlsLmOO0Q1v8hLg6yRk-ted8dNx1BN7ZcHjDyaDdR0mHBb-XVL0zUwISiIrXXulr882BXgZBWBEXJuW1x7yFzU3yQMTepoNicbU7VdiNVNQ6yoZtkV-GnFv798-F5dilBYAO3oKR\" alt=\"\" \/><figcaption><br \/><em>Figure 0. Capacitive touch.<\/em> <\/figcaption><\/figure><p>So she starts tap-tap-tapping on the keyboard, one letter at a time.<\/p>\n\n\n\n<p>I-<em>spacebar<\/em>-l-o-v-e-<em>spacebar<\/em>-y-o-u. <br \/><\/p>\n\n\n\n<p>She\u2019s not a keyboard swiper (I am, but somehow she still types faster than me). The phone reliably records the (x,y) coordinates of each thumbprint and aligns it with the coordinates of each key on the screen. It\u2019s harder than you think; sometimes her thumb slips, yet somehow the phone realizes she\u2019s not trying to swipe, that it was just a messy press.<\/p>\n\n\n\n<p>Deep in the metal guts of the device, an algorithm tests whether each thumb-shaped voltage disruption moves more that a certain number of pixels, called <em>touch slop<\/em>. If the movement is sufficiently small, the phone registers a keypress rather than a swipe.<br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img data-attachment-id=\"42110\" data-permalink=\"https:\/\/scottbot.net\/?attachment_id=42110\" data-orig-file=\"https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?fit=798%2C310&amp;ssl=1\" data-orig-size=\"798,310\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"google android code\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?fit=300%2C117&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?fit=798%2C310&amp;ssl=1\" src=\"https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?w=840\" alt=\"\" class=\"wp-image-42110\" srcset=\"https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?w=798&amp;ssl=1 798w, https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?resize=300%2C117&amp;ssl=1 300w, https:\/\/i1.wp.com\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png?resize=768%2C298&amp;ssl=1 768w\" data-recalc-dims=\"1\" \/><figcaption>Fig 1. Android\u2019s code for detecting \u2018touch slop\u2019. Notice the developers had my wife\u2019s gender in mind.<\/figcaption><\/figure><p>She finishes her message, a measly 10 characters of her allotted 160. <\/p>\n\n\n\n<p>The allotment of 160 characters is a carefully chosen number, if you believe the legend: In 1984, German telephone engineer Friedhelm Hillebrand sat at his typewriter and wrote as many random sentences as came to his mind. His team then looked at postcards and telex messages, and noticed most fell below 160 characters. \u201cEureka!\u201d, they presumably yelled in German, before setting the character limit of text messages in stone for the next three-plus decades.<br \/><\/p>\n\n\n\n<h1>Character Limits &amp; Legends<\/h1>\n\n\n\n<p>Legends rarely tell the whole story, and the legend of SMS is no exception. Hillebrand and his team hoped to relay messages over a secondary channel that phones were already using to exchange basic information with their local stations. <br \/><\/p>\n\n\n\n<p>Signalling System no. 7 (SS7) are a set of protocols used by cell phones to stay in constant contact with their local tower; they need this continuous connection to know when to ring, to get basic location tracking, to check for voicemail, and communicate other non-internet reliant messages. Since the protocol\u2019s creation in 1980, it had a hard limit of 279 bytes of information. If Hillebrand wanted text messages to piggyback on the SS7 protocol, he had to deal with this pretty severe limit.<br \/><\/p>\n\n\n\n<p>Normally, 279 bytes equals 279 characters. A byte is eight bits (each bit is a 0 or 1), and in common encodings, a single letter is equivalent to eight 0s and 1s in a row. <\/p>\n\n\n\n<p>\u2018A\u2019 is<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">0100 0001<\/pre>\n\n\n\n<p>\u2018B\u2019 is<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">0100 0010<\/pre>\n\n\n\n<p>\u2018C\u2019 is<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">0100 0011<\/pre>\n\n\n\n<p>and so on. <br \/><\/p>\n\n\n\n<p>Unfortunately, getting messages across the SS7 protocol isn\u2019t a simple matter of sending 2,232 (that\u2019s 279 bytes at 8 bits each) 0s or 1s through radio signals from my phone to yours. Part of that 279-byte signal needs to contain your phone number, and part of it needs to contain my phone number. Part of it needs to let the cell tower know \u201chey, this is a message, not a call, don\u2019t ring the phone!\u201d. <br \/><\/p>\n\n\n\n<p>By the time Hillebrand and his team finished cramming all the necessary contextual bits into the 279-byte signal, they were left with only enough space for 140 characters at 1 byte (8 bits) a piece, or 1,120 bits. <br \/><\/p>\n\n\n\n<p>But what if they could encode a character in only 7 bits? At 7 bits per character, they could squeeze 160 (1,140 \/ 7 = 160) characters into each SMS, but those extra twenty characters demanded a sacrifice: fewer possible letters. <br \/><\/p>\n\n\n\n<p>An 8-bit encoding allows 256 possible characters: lowercase \u2018a\u2019 takes up one possible space, uppercase \u2018A\u2019 another space, a period takes up a third space, an \u2018@\u2019 symbol takes up a fourth space, a line break takes up a fifth space, and so on up to 256. To squeeze an alphabet down to 7 bits, you need to remove some possible characters: the 1\/2 symbol (\u00bd), the degree symbol (\u00b0), the pi symbol (\u03c0), and so on. But assuming people will never use those symbols in text messages (a poor assumption, to be sure), this allowed Hillebrand and his colleagues to stuff 160 characters into a 140-byte space, which in turn fit neatly into a 279-byte SS7 signal: exactly the number of characters they claim to have discovered was the perfect length of a message. (A bit like the miracle of Hanukkah, if you ask me.)<br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh4.googleusercontent.com\/9aJOO5Qoa2Hg531TGIiX1b3-DFtCIf5gJBTOZ4Vw6cErngpblcDc9IKrQFFjTthm0z8rz4THEqguMSwaNLcrbKlVjdYX2owAWgGyrqqXyZJB7D_ctl_jLm6jL-rIzBu4rjSfm_lG\" alt=\"\" \/><\/figure><figure class=\"wp-block-image\"><img src=\"https:\/\/lh6.googleusercontent.com\/EYhYMzHKdsNheP0KVn9uBSgrHThR3MWtFqo8wvUq8lJqv0-GNnGpkmgQycMXV9qVqugkgiy92w6Cm13dz_WwgrwqG8RZ6-lZGXyC6jIR97oYHmD3QBAzrVp-Rq7MkGGJL7X8NxtS\" alt=\"\" \/><figcaption><em>Fig 2. The GSM-7 character set.<\/em> <\/figcaption><\/figure><p>So there my wife is, typing \u201cI love you\u201d into a text message, all the while the phone converts those letters into this 7-bit encoding scheme, called GSM-7. <\/p>\n\n\n\n<p>\u201cI\u201d (notice it\u2019s at the intersection of <em>4x<\/em> and <em>x9<\/em> above) =<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">49 <\/pre>\n\n\n\n<p>Spacebar (notice it\u2019s at the intersection of <em>2x<\/em> and <em>x0<\/em> above) =<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">20 <\/pre>\n\n\n\n<p>\u201cl\u201d =<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">6C<\/pre>\n\n\n\n<p>\u201co\u201d = <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">6F<\/pre>\n\n\n\n<p>and so on down the line. <br \/><\/p>\n\n\n\n<p>In all, her slim message becomes: <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">49 20 6C 6F 76 65 20 79 6F 75 <\/pre>\n\n\n\n<p>(10 bytes combined). Each two-character code, called a <em>hex code<\/em>, is one 8-bit chunk, and together it spells \u201cI love you\u201d.<\/p>\n\n\n\n<p>But this is actually not how the message is stored on her phone. It has to convert the 8-bit text to 7-bit hex codes, which it does by essentially borrowing the remaining bit at the end of every byte. The math is a bit more <a href=\"http:\/\/web.archive.org\/web\/20081017024616\/http:\/\/www.dreamfabric.com:80\/sms\/hello.html\">complicated<\/a> than is worth getting into here, but the resulting message appears as <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">49 10 FB 6D 2F 83 F2 EF 3A <\/pre>\n\n\n\n<p>(9 bytes in all) in her phone.<br \/><\/p>\n\n\n\n<p>When my wife finally finishes her message (it takes only a few seconds), she presses \u2018send\u2019 and a host of tiny angels retrieve the encoded message, flutter their invisible wings the 40 feet up to the office, and place it gently into my phone. The process isn\u2019t entirely frictionless, which is why my phone vibrates lightly upon delivery.<br \/><\/p>\n\n\n\n<p>The so-called \u201ctelecommunication engineers\u201d will tell you a different story, and for the sake of completeness I\u2019ll relay it to you, but I wouldn\u2019t trust them if I were you.<\/p>\n\n\n\n<h1>SIM-to-Send<\/h1>\n\n\n\n<p>The engineers would say that, when the phone senses voltage fluctuations over the \u2018send\u2019 button, it sends the encoded message to the SIM card (that tiny card your cell provider puts in your phone so it knows what your phone number is), and in the process it wraps it in all sorts of useful contextual data. By the time it reaches my wife\u2019s SIM, it goes from a 140-byte message (just the text) to a 176-byte message (text + context). <br \/><\/p>\n\n\n\n<p>The extra 36 bytes are used to encode all sorts of information, seen below. <\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh5.googleusercontent.com\/sMclPt0j-mW8sVUshmo1tQF4OFB7TzawmLxdxphQh1srtfJ6LVZH9BxoSp8RCIf0dmMnoVnvlcl6zpf7pkiSwkDYLccpz8KwKt7Gz-UpGi0xGNM-HThLOziWNPGVtr3_Hp82OnUf\" alt=\"\" \/><figcaption><em>Fig 3. Here, bytes are called octets (8 bits). Counting all possible bytes yields 174 (10+1+1+12+1+1+7+1+140). The other two bytes are reserved for some SIM card bookkeeping.<\/em> <\/figcaption><\/figure><p>The first ten bytes are reserved for the telephone number (service center address, or SCA) of the SMS service center (SMSC), tasked with receiving, storing, forwarding, and delivering text messages. It\u2019s essentially a switchboard: my wife\u2019s phone sends out a signal to the local cell tower and gives it the number of the SMSC, which forwards her text message from the tower to the SMSC. The SMSC, which in our case is operated by AT&amp;T, routes the text to the mobile station nearest to my phone. Because I\u2019m sitting three rooms away from my wife, the text just bounces back to the same mobile station, and then to my phone.<br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh3.googleusercontent.com\/U-8uwEgJzETkO5-v0moohHGtYNba-PN-Kz4CsTRdt1S0RJvs-aTh4Qqh5lf048oNIMboekQaEBd5k_Edxp5BzvbB5zx6hIuTcqDf9rjPd-B3D0U1K2CxhXzovUe_1gfp6kxh-Gd7\" alt=\"\" \/><figcaption><em>Fig 4. SMS cellular network<\/em> <\/figcaption><\/figure><p>The next byte (PDU-type) encodes some basic housekeeping on how the phone should interpret the message, including whether it was sent successfully, whether the carrier requests a status report, and (importantly) whether this is a single text or part of a string of connected messages. <br \/><\/p>\n\n\n\n<p>The byte after the PDU-Type is the message reference (MR). It\u2019s a number between 1 and 255, and is essentially used as a short-term ID number to let the phone and the carrier know which text message it\u2019s dealing with. In my wife\u2019s case the number is set to 0, because her phone has its own message ID system independent of this particular file.<br \/><\/p>\n\n\n\n<p>The next twelve bytes or so are reserved for the recipient\u2019s phone number, called the destination address (DA). With the exception of the 7-bit letter character encoding I mentioned earlier, that helps us stuff 160 letters into a 140-character space, the phone number encoding is the stupidest, most confusing bits you\u2019ll encounter in this SMS. It\u2019s called reverse nibble notation, and it reverses every other digit in a large number. (Get it? Part of a byte is a nibble, hahahahaha, nobody\u2019s laughing, engineers.) <br \/><\/p>\n\n\n\n<p>My number, which is usually 1-352-537-8376, is logged in my wife\u2019s phone as: <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">3125358773f6<\/pre>\n\n\n\n<p>The 1-3 is represented by <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">31<\/pre>\n\n\n\n<p>The 52 is represented by<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">25<\/pre>\n\n\n\n<p>The 53 is represented by<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">35<\/pre>\n\n\n\n<p>The 7-8 is represented by <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">87<\/pre>\n\n\n\n<p>The 37 is represented by <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">73<\/pre>\n\n\n\n<p>And the 6 is represented by\u2026<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">f6<\/pre>\n\n\n\n<p>Where the fuck did the \u2018f\u2019 come from? It means it\u2019s the end of the phone number, but for some awful reason (again, reverse nibble notation) it\u2019s one character before the final digit. <br \/><\/p>\n\n\n\n<p>It\u2019s like pig latin for numbers.<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">tIs'l ki eip galit nof runbmre.s<br \/><\/pre>\n\n\n\n<p>But I\u2019m not bitter.<\/p>\n\n\n\n<p>[Edit: <a href=\"https:\/\/twitter.com\/seangies\/status\/1085302529917763585\">Sean Gies points out<\/a> that reverse nibble notation is an inevitable artifact of representing 4-bit <a href=\"https:\/\/en.wikipedia.org\/wiki\/Endianness\">little-endian<\/a> numbers in 8-bit chunks. That doesn\u2019t invalidate the above description, but it does add some context for those who know what it means, and makes the decision seem more sensible.]<\/p>\n\n\n\n<p>The Protocol Identifier (PID) byte is honestly, at this point, mostly wasted space. It takes about 40 possible values, and it tells the service provider how to route the message. A value of <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">22 <\/pre>\n\n\n\n<p>means my wife is sending \u201cI love you\u201d to a fax machine; a value of <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">24 <\/pre>\n\n\n\n<p>means she\u2019s sending it to a voice line, somehow. Since she\u2019s sending it as an SMS to my phone, which receives texts, the PID is set to<\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">0<\/pre>\n\n\n\n<p>(Like every other text sent in the modern world.)<br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh6.googleusercontent.com\/GzrQ7QoYr4YOne9hwSA_Ua0GuF-uvCeH54sUH5zUmJwpDq1zyVIw4kjKRZEtR-BrctmhUrEj99MOH9AhWy4-jUgl2i8284_7mWYUFncB3cV0WXpR5P2AnzedxGJAqRftSLPRuxsG\" alt=\"\" \/><figcaption><em>Fig 5. Possible PID Values<\/em> <\/figcaption><\/figure><p>The next byte is the Data Coding Scheme (DCS, <a href=\"https:\/\/www.etsi.org\/deliver\/etsi_ts\/123000_123099\/123038\/10.00.00_60\/ts_123038v100000p.pdf\">see this doc for details<\/a>), which tells the carrier and the receiving phone which character encoding scheme was used. My wife used GSM-7, the 7-bit alphabet I mentioned above that allows her to stuff 160 letters into a 140-character space, but you can easily imagine someone wanting to text in Chinese, or someone texting a complex math equation (ok, maybe you can\u2019t easily imagine that, but a guy can dream, right?). <br \/><\/p>\n\n\n\n<p>In my wife\u2019s text, the DCS byte was set to <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">0<\/pre>\n\n\n\n<p>meaning she used a 7-bit alphabet, but she may have changed that value to use an 8- or 16-bit alphabet, which would allow her many more possible letters, but a much smaller space to fit them. Incidentally, this is why when you text emoji to your friend, you have fewer characters to work with.<br \/><\/p>\n\n\n\n<p>There\u2019s also a little flag in the DCS byte that tells the phone whether to self-destruct the message after sending it, <em>Mission Impossible<\/em> style, so that\u2019s neat.<br \/><\/p>\n\n\n\n<p>The validity period (VP) space can take up to seven bytes, and sends us into another aspect of how text messages actually work. Take another look at Figure 4, above. It\u2019s okay, I\u2019ll wait.<br \/><\/p>\n\n\n\n<p>When my wife finally hits \u2018send\u2019, the text gets sent to the SMS Service Center (SMSC), which then routes the message to me. I\u2019m upstairs and my phone is on, so I receive the text in a handful of seconds, but what if my phone were off? Surely my phone can\u2019t accept a message when it\u2019s not receiving any power, so the SMSC has to do <em>something<\/em> with the text.<br \/><\/p>\n\n\n\n<p>If the SMSC can\u2019t find my phone, my wife\u2019s message will just bounce around in its system until the moment my phone reconnects, at which point it sends the text out immediately. I like to think of the SMSC continuously checking every online phone to see if its mine like a puppy waiting for its human by the door: is that smell my human? No. Is <em>that<\/em> smell my human? No. Is <em>this<\/em> smell my human? YESYESJUMPNOW.<br \/><\/p>\n\n\n\n<p>The validity period (VP) bytes tell the carrier how long the puppy will wait before it gets bored and finds a new home. It\u2019s either a timestamp or a duration, and it basically says \u201cif you don\u2019t see the recipient phone pop online in the next however-many days, just don\u2019t bother sending it.\u201d The default validity period for a text is 10,080 minutes, which means if it takes me more than seven days to turn my phone back on, I\u2019ll never receive her text.<br \/><\/p>\n\n\n\n<p>Because there\u2019s often a lot of empty space in an SMS, a few bits here or there are dedicated to letting the phone and carrier know exactly which bytes are unused. If my wife\u2019s SIM card expects a 176-byte SMS, but because she wrote an exceptionally short message it only receives a 45-byte SMS, it may get confused and assume something broke along the way. The user data length (UDL) byte solves this problem: it relays exactly how many bytes the text in the text message actually take up. <br \/><\/p>\n\n\n\n<p>In the case of \u201cI love you\u201d, the UDL claims the subsequent message is 9 bytes. You\u2019d expect it to be 10 bytes, one for each of the 10 characters in <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">I-spacebar-l-o-v-e-spacebar-y-o-u<\/pre>\n\n\n\n<p>but because each character is 7 bits rather than 8 bits (a full byte), we\u2019re able to shave an extra byte off in the translation. That\u2019s because 7 bits * 10 characters = 70 bits, divided by 8 (the number of bits in a byte) = 8.75 bytes, rounded up to 9 bytes. <br \/><\/p>\n\n\n\n<p>Which brings us to the end of every SMS: the message itself, or the UD (User Data). The message can take up to 140 bytes, though as I just mentioned, \u201cI love you\u201d will pack into a measly 9. Amazing how much is packed into those 9 bytes\u2014not just the message (my wife\u2019s presumed love for me, which is already difficult enough to compress into 0s and 1s), but also the <em>message<\/em> (I need to come downstairs and wish her goodnight). Those bytes are: <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">49 10 FB 6D 2F 83 F2 EF 3A.<br \/><\/pre>\n\n\n\n<p>In all, then, this is the text message stored on my wife\u2019s SIM card:<br \/><\/p>\n\n\n\n<p>SCA[1-10]-PDU[1]-MR[1]-DA[1-12]-DCS[1]-VP[0, 1, or 7]-UDL[1]-UD[0-140]<br \/><\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">00 - 11 - 00 - 07 31 25 35 87 73 F6 - ?? 00 ?? - ?? - 09 - 49 10 FB 6D 2F 83 F2 EF 3A<br \/><\/pre>\n\n\n\n<p>(Note: to get the full message, I need to do some more digging. Alas, you only see <em>most<\/em> of the message here, hence the ??s.)<\/p>\n\n\n\n<h1>Waves in the \u00c6ther<\/h1>\n\n\n\n<p>Somehow [he says in David Attenborough\u2019s voice], the SMS must now begin its arduous journey from the SIM card to the nearest base station. \u00a0To do that, my wife\u2019s phone must convert a string of 176 bytes to the 279 bytes readable by the SS7 protocol, convert those digital bytes to an analog radio signal, and then send those signals out into the \u00e6ther at a frequency of somewhere between 800 and 2000 megahertz. That means each wave is between 6 and 14 inches from one peak to the next. <br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh5.googleusercontent.com\/RNm3grWjf8Ar88wrfuRcnuj1K_T2Q9DMOl3kBaOWZqCDcZ_ByF_no_WL1NUFoBUppuVcv-bjX4X78KW2iGDRXqfehRpdU1rnlyP3HT6MxKDhn2udLeV4ednGofn1ozZ-OCBMvaMl\" alt=\"\" \/><figcaption><em>Fig 6. Wavelength<\/em> <\/figcaption><\/figure><p>In order to efficiently send and receive signals, antennas should be no smaller than half the size of the radio waves they\u2019re dealing with. If cell waves are 6 to 14 inches, their antennas need to be 3-7 inches. Now stop and think about the average height of a mobile phone, and why they never seem to get much smaller.<br \/><\/p>\n\n\n\n<p>Through some digital gymnastics that would take entirely too long to explain, suddenly my wife\u2019s phone shoots a 279-byte information packet containing \u201cI love you\u201d at the speed of light in every direction, eventually fizzling into nothing after about 30 miles. <br \/><\/p>\n\n\n\n<p>Well before getting that far, her signal strikes the AT&amp;T HSPA Base Station ID199694204 LAC21767. This base transceiver station (BTS) is about 5 blocks from my favorite bakery in Hazelwood, La Gourmandine, and though I was able to find its general location using an android app called OpenSignal, the antenna is camouflaged beyond my ability to find it.<\/p>\n\n\n\n<p>The really fascinating bit here is that it reaches the base transceiver station at all, given everything else going on. Not only is my wife texting me \u201cI love you\u201d in the 1000ish mhz band of the electromagnetic spectrum; tens of thousands of other people are likely talking on the phone or texting within the 30 mile radius around my house, beyond which cell signals disintegrate. On top of that, a slew of radio and TV signals are jostling for attention in our immediate airspace, alongside visible light bouncing this way and that, to name a few of the many electromagnetic waves that seem like they ought to be getting in the way.<br \/><\/p>\n\n\n\n<p>As <a href=\"https:\/\/www.youtube.com\/watch?v=1qQQXTMih1A\">Richard Feynman eloquently put it in 1983<\/a>, it\u2019s a bit like the cell tower is a little blind bug resting gently atop the water on one end of a pool, and based only on the frequency and direction of waves that cause it to bounce up and down, it\u2019s able to reconstruct who\u2019s swimming and where. <br \/><\/p>\n\n\n\n<figure class=\"wp-block-embed-youtube wp-block-embed is-type-video is-provider-youtube wp-embed-aspect-4-3 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe class=\"youtube-player\" width=\"840\" height=\"473\" src=\"https:\/\/www.youtube.com\/embed\/1qQQXTMih1A?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>\n<\/div><figcaption>Feynman discussing waves.<\/figcaption><\/figure><p>In part due to the complexity of competing signals, each base transceiver station generally can\u2019t handle more than 200 active users (using voice or data) at a time. So \u201cI love you\u201d pings my local base transceiver station, about a half a mile away, and then shouts itself into the void in every direction until it fades into the noise of everyone else.<\/p>\n\n\n\n<h1>Switching<\/h1>\n\n\n\n<p>I\u2019m pretty lucky, all things considered. Were my wife and I on different cell providers, or were we in different cities, the route of her message to me would be a good deal more circuitous. <br \/><\/p>\n\n\n\n<p>My wife\u2019s message is massaged into the 279-byte SS7 channel, and sent along to the local base transceiver station (BTS) near the bakery. From there, it gets routed to the base station controller (BSC), which is the brain of not just our antenna, but several other local antennas besides. The BSC flings the text to AT&amp;T Pittsburgh\u2019s mobile switching center (MSC), which relies on the text message\u2019s SCA (remember the service center address embedded within every SMS? That\u2019s where this comes in) to get it to the appropriate short message service center (SMSC). <br \/><\/p>\n\n\n\n<p>This alphabet soup is easier to understand with the diagram from figure 7; I just described steps 1 and 3. If my wife were on a different carrier, we\u2019d continue through steps 4-7, because that\u2019s where the mobile carriers all talk to each other. The SMS has to go from the SMSC to a global switchboard and then potentially bounce around the world before finding its way to my phone. <br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh3.googleusercontent.com\/5PSN844ciuDUjzZnVHDSooXHrkKhvX9rjG14k_oUUffT7Xm97KOlT0dIYKPje2wY0LEr6MlXCKDkU4OaHhQlieI52s6njAZDagiIodE3eElX-UcpDCRoBgyDIHh15KgnZBVBl34y\" alt=\"\" \/><figcaption><em>Fig 7. SMS routed through a GSM network.<\/em> <\/figcaption><\/figure><p>But she\u2019s on AT&amp;T and I\u2019m on AT&amp;T, and our phones are connected to the same tower, so after step 3 the 279-byte packet of love just does an about-face and returns through the same mobile service center, through the same base station, and now to my phone instead of hers. A trip of a few dozen miles in the blink of an eye.<\/p>\n\n\n\n<h1>Sent-to-SIM<\/h1>\n\n\n\n<p><em>Buzzzzz<\/em>. My pocket vibrates. A notification lets me know an SMS has arrived through my nano-SIM card, a circuit board about the size of my pinky nail. Like Bilbo Baggins or any good adventurer, it changed a bit in its trip there and back again.<br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh6.googleusercontent.com\/NbA9m8nW2fHDCSlmcittyXIGl9-tDh85sAFQixng2j-aE7Hbsv9gaDk-BAtGwkWqt1CZOgsyp5C0q0CNopbREswegN3YUWQgEOm8ckis89UNfp8Cvf0Z9omyafYUKyJ7kwM3nCOH\" alt=\"\" \/><figcaption><em>Fig 8. Received message, as opposed to sent message (figure 3).<\/em> <\/figcaption><\/figure><p>Figure 8 shows the structure of the message \u201cI love you\u201d now stored on my phone. Comparing figures 3 and 8, we see a few differences. The SCA (phone number of the short message service center), the PDU (some mechanical housekeeping), the PID (phone-to-phone rather than, say, phone-to-fax), the DCS (character encoding scheme), the UDL (length of message), and the UD (the message itself) are all mostly the same, but the VP (the text\u2019s expiration date), the MR (the text\u2019s ID number), and the DA (my phone number) are missing.<br \/><\/p>\n\n\n\n<p>Instead, on my phone, there are two new pieces of information: the OA (originating address, or my wife\u2019s phone number), and the SCTS (service center time stamp, or when my wife sent the message). <br \/><\/p>\n\n\n\n<p>My wife\u2019s phone number is stored in the same annoying reverse nibble notation (like dyslexia but for computers) that my phone number was stored in on her phone, and the timestamp is stored in the same format as the expiration date was stored in on on her phone. <br \/><\/p>\n\n\n\n<p>These two information inversions make perfect contextual sense. Her phone needed to reach me by a certain time at a certain address, and I now need to know who sent the message and when. Without the home address, so to speak, I wouldn\u2019t know whether the \u201cI love you\u201d came from my wife or my mother, and the difference would change my interpretation of the message fairly significantly.<\/p>\n\n\n\n<h1>Through a Glass Brightly<\/h1>\n\n\n\n<p>In much the same way that any computer translates a stream of bytes into a series of (x,y) coordinates with specific color assignments, my phone\u2019s screen gets the signal to render <\/p>\n\n\n\n<pre class=\"wp-block-preformatted\">49 10 FB 6D 2F 83 F2 EF 3A<\/pre>\n\n\n\n<p>on the screen in front of me as \u201cI love you\u201d in backlit black-and-white. It\u2019s an interesting process, but as it\u2019s not particularly unique to smartphones, you\u2019ll have to look it up elsewhere. Let\u2019s instead focus on how those instructions become points of light.<br \/><\/p>\n\n\n\n<p>The friendly marketers at Samsung call my screen a Super AMOLED (Active Matrix Organic Light-Emitting Diode) display, which is somehow both redundant and not particularly informative, so we\u2019ll ignore unpacking the acronym as yet another distraction, and dive right into the technology.<br \/><\/p>\n\n\n\n<p>There are about 330,000 tiny sources of light, or pixels, crammed inside each of my phone screen\u2019s 13 square inches. For that many pixels, each needs to be about 45\u00b5m (micrometers) wide: thinner than a human hair. There\u2019s 4 million of \u2018em in all packed into the palm of my hand.<br \/><\/p>\n\n\n\n<p>But you already know how screens work. You know that every point of light, like the Christian God or Musketeers (minus d\u2019Artagnan), is always a three-for-one sort of deal. Red, green, and blue combine to form white light in a single pixel. Fiddle with the luminosity of each channel, and you get every color in the rainbow. And since 4 x 3 = 12, that\u2019s 12 million tiny sources of light sitting innocently dormant behind my black mirror, waiting for me to press the power button to read my wife\u2019s text.<\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh5.googleusercontent.com\/lyM52NvkH6b7UpTojk2SL6fo4U0sb06_R7vhKRPQ-hPL8mRurOmX1XcemN-NPYDLh8O0HCs_g8j3xDl5SUh_aq8U2Qm7J4bbZN0A5ZPv262Hs5aujULUzsYpjVlO5VBNfRI9FRxi\" alt=\"\" \/><figcaption><em>Fig 9. The subpixel array of a Samsung OLED display.<\/em> <\/figcaption><\/figure><p>Each pixel, as the acronym suggests, is an organic light-emitting diode. That\u2019s fancy talk for an electricity sandwich: <br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh6.googleusercontent.com\/PVA7sQBXXtUC6EJUsxV0jLZnVd5VdErlPFfGSIwC5T8XsQZ8nv8y8ni1M5iqh1M870FH0nbx25mti-j_yCy8OQ8jlrXnh17ISmaK11xJ6WkbkfPC_0j8P1r76Q9YxVdxERxthBAC\" alt=\"\" \/><figcaption><em>Fig 10. An electricity sandwich.<\/em><\/figcaption><\/figure><p>The layers aren\u2019t too important, beyond the fact that it\u2019s a cathode plate (negatively charged), below a layer of organic molecules (remember back to highschool: it\u2019s just some atoms strung together with carbon), below an anode plate (positively charged). <br \/><\/p>\n\n\n\n<p>When the phone wants the screen on, it sends electrons from the cathode plate to the anode plate. The sandwiched molecules intercept the energy, and in response they start emitting visible light, photons, up through the transparent anode, up through the screen, and into my waiting eyes.<br \/><\/p>\n\n\n\n<p>Since each pixel is three points of light (red, green, and blue), there\u2019s actually three of these sandwiches per pixel. They\u2019re all essentially the same, except the organic molecule is switched out: poly(p-phenylene) for blue light, polythiophene for red light, and poly(p-phenylene vinylene) for green light. Because each is slightly different, they shine different colors when electrified. <br \/><\/p>\n\n\n\n<p>(Fun side fact: blue subpixels burn out much faster, due to a process called \u201cexciton-polaron annihilation\u201d, which sounds really exciting, doesn\u2019t it?)<br \/><\/p>\n\n\n\n<p>All 4 million pixels are laid out on an indexed matrix. An index works in a computer much the same way it works in a book: when my phone wants a specific pixel to light a certain color, it looks that pixel up in the index, and then sends a signal to the address it finds. <em>Let there be light, and there was light<\/em>.<br \/><\/p>\n\n\n\n<p>(Fun side fact: now you know what \u201cActive Matrix Light-Emitting Diode\u201d means, and you didn\u2019t even try.)<br \/><\/p>\n\n\n\n<p>My phone\u2019s operating system interprets my wife\u2019s text message, figures out the shape of each letter, and maps those shapes to the indexed matrix. It sends just the right electric pulses through the Super AMOLED screen to render those three little words that have launched ships and vanquished curses.<br \/><\/p>\n\n\n\n<p>The great strangeness here is that my eyes never see \u201cI love you\u201d in bright OLED lights; it appear on the screen black-on-white. The phone creates the illusion of text through negative space, washing the screen white by setting every red, green, &amp; blue to maximum brightness, then turning off the bits where letters should be. Its complexity is offensively mundane.<br \/><\/p>\n\n\n\n<figure class=\"wp-block-image\"><img src=\"https:\/\/lh3.googleusercontent.com\/ET7St4Ie1cKCmH9Qw-pcqM_VEiPqjKSWhsxJqcQL4NBLabfceSr3OSI9hMjsqLYK1UsojdX35bMINeFrScfgXQeqnDCAAvAwLseTltg7l-JWKq97lUQ_r5Nd8H_JJE-6kYYFB0IX\" alt=\"\" \/><figcaption><em>Fig 11. Negative space.<\/em><\/figcaption><\/figure><p>In displaying everything <em>but<\/em> my wife\u2019s text message, and letting me read it in the gaps, my phone succinctly betrays the lie at the heart of the information age: that communication is <em>simple<\/em>. Speed and ease hide a mountain of mediation. <br \/><\/p>\n\n\n\n<p>And that mediation isn\u2019t just technical. My wife\u2019s text wouldn\u2019t have reached me had I not paid the phone bill on time, had there not been a small army of workers handling financial systems behind the scenes. Technicians keep the phone towers in working order, which they reach via a network of roads partially subsidized by federal taxes collected from hundreds of millions of Americans across 50 states. Because so many transactions still occur via mail, if the U.S. postal system collapsed tomorrow, my phone service would falter. Exploited factory workers in South America and Asia assembled parts in both our phones, and exhausted programmers renting expensive Silicon Valley closets are as-you-read-this pushing out code ensuring our phones communicate without interruption. <br \/><\/p>\n\n\n\n<p>All of this underneath a 10-character text. A text which, let\u2019s be honest, means much more than it says. My brain subconsciously peels back years of interactions with my wife to decode the message appearing on my phone, but between her and me there\u2019s still a thicket of sociotechnical mediation, a stew of people and history and parts, that can never be untangled.<\/p>\n\n\n\n<h1>The Aftermath<\/h1>\n\n\n\n<p>So here I am, in the office late one Sunday night. \u201cI love you,\u201d my wife texted from the bedroom downstairs, before the message traversed 40 or so feet to my phone in a handful of seconds. I realize what it means: it\u2019s time to wish her goodnight, and perhaps wrap up this essay. I tap away the last few words, now slightly more cognizant of the complex layering of miles, signals, years of history, and human sweat it took to keep my wife from having to shout upstairs that it\u2019s about damn time I get some rest.<\/p>\n\n\n\n<p><em>Thanks to Christopher Warren, Vika Zafrin, and Nechama Weingart for comments on earlier drafts.<\/em><\/p>\n\n\n<p><\/p>\n\n<div id=\"jp-relatedposts\" class=\"jp-relatedposts\">\n\t<h3 class=\"jp-relatedposts-headline\"><em>Related<\/em><\/h3>\n<\/div>","created_at":"2019-01-16T17:49:03+0000","updated_at":"2019-01-16T17:49:03+0000","published_at":"2019-01-15T16:57:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":25,"domain_name":"scottbot.net","preview_picture":"https:\/\/scottbot.net\/wp-content\/uploads\/2019\/01\/image1.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":779,"title":"Some Lisp books (and then some)","url":"http:\/\/blog.fogus.me\/2012\/07\/25\/some-lisp-books-and-then-some\/","content":"Jul 25, 2012<h2>Some Lisp books<\/h2><p><em>caveat: this is not a best-of nor a comprehensive list of Lisp books; it is merely a selection of Lisp books you may not have heard of or that special to me in some way.<\/em><\/p><p>My first Lisp book:<\/p><p><a href=\"http:\/\/www.cs.cmu.edu\/~dst\/LispBook\/\"> \n<img src=\"http:\/\/farm9.staticflickr.com\/8147\/7643955638_dbf848ebf6_n_d.jpg\" alt=\"symbolic-comp\" \/><\/a><\/p><p>My first Lisp:<\/p><p><a href=\"http:\/\/www.amazon.com\/LISP-STAT-Object-Oriented-Environment-Statistical-Probability\/dp\/0471509167\/?tag=fogus-20\">\n<img src=\"http:\/\/farm8.staticflickr.com\/7266\/7643885134_eeb7dcca8f_n_d.jpg\" alt=\"lisp-stat\" \/><\/a><\/p><p>My favorite undergrad Lisp\/Prolog book:<\/p><p><a href=\"http:\/\/www.amazon.com\/Symbolic-Computing-PROLOG-Robert-Mueller\/dp\/0471607711\/?tag=fogus-20\"> \n<img src=\"http:\/\/farm9.staticflickr.com\/8283\/7643812708_d796319b72_n_d.jpg\" alt=\"lisp-prolog\" \/><\/a><\/p><p>My favorite Lisp:<\/p><p><a href=\"http:\/\/www.amazon.com\/Programming-Language-Dialect-Lisp\/dp\/013881905X\/?tag=fogus-20\"> \n<img src=\"http:\/\/farm9.staticflickr.com\/8292\/7643701892_acf82b18e5_n_d.jpg\" alt=\"T\" \/><\/a><\/p><p>The best Lisp book I\u2019ve ever read and could ever possibly read<sup id=\"fnref:1\"><a href=\"#fn:1\" rel=\"footnote\">1<\/a><\/sup>:<\/p><p><a href=\"http:\/\/www.amazon.com\/Architecture-Computers-Mcgraw-Hill-Supercomputing-Processing\/dp\/0070355967\/?tag=fogus-20\">\n<img src=\"http:\/\/farm8.staticflickr.com\/7252\/7644006104_e55c00e35c_m_d.jpg\" alt=\"arch\" \/><\/a><\/p><p><a href=\"http:\/\/n01se.net\/chouser\/\">Chouser\u2019s book<\/a>:<\/p><p><a href=\"http:\/\/www.joyofclojure.com\">\n<img src=\"http:\/\/www.joyofclojure.com\/images\/joy-of-clojure-cover-239x300.jpg\" alt=\"joy\" \/><\/a><\/p><p>Mind-blowing and bombastic:<\/p><p><a href=\"http:\/\/www.amazon.com\/Let-Over-Lambda-Doug-Hoyte\/dp\/1435712757\/?tag=fogus-20\">\n<img src=\"http:\/\/ecx.images-amazon.com\/images\/I\/41rpu2m9vkL.jpg\" alt=\"letoverlambda\" \/><\/a>\n<em>excerpts at <a href=\"http:\/\/letoverlambda.com\/textmode.cl\/guest\/toc\">the author\u2019s site<\/a><\/em><\/p><p>The best Clojure book for beginners:<\/p><p><a href=\"http:\/\/www.amazon.com\/Clojure-Programming-Chas-Emerick\/dp\/1449394701\/?tag=fogus-20\">\n<img src=\"http:\/\/akamaicovers.oreilly.com\/images\/0636920013754\/cat.gif\" alt=\"clojure-book\" \/><\/a><\/p><p>Language implementation 101:<\/p><p><a href=\"http:\/\/www.amazon.com\/Essentials-Programming-Languages-Daniel-Friedman\/dp\/0262062798\/?tag=fogus-20\">\n<img src=\"http:\/\/www.eopl3.com\/cover.jpg\" alt=\"lang-impl-101\" \/><\/a><\/p><h2>Some books are worth any price<\/h2><h4>On Lisp by Paul Graham<\/h4><p><a href=\"http:\/\/www.paulgraham.com\/onlisptext.html\">\n<img src=\"http:\/\/ecx.images-amazon.com\/images\/I\/412T69YKAXL.jpg\" alt=\"on-lisp\" \/><\/a><\/p><h4>Compiling with Continuations by Andrew Appel<sup id=\"fnref:2\"><a href=\"#fn:2\" rel=\"footnote\">2<\/a><\/sup><\/h4><p><a href=\"http:\/\/www.amazon.com\/Compiling-Continuations-Andrew-W-Appel\/dp\/052103311X\/?tag=fogus-20\"> \n<img src=\"http:\/\/ecx.images-amazon.com\/images\/I\/41TXGKT68AL.jpg\" alt=\"k\" \/><\/a><\/p><h4>Lisp in Small Pieces by Christian Queinnec<\/h4><p><a href=\"http:\/\/www.amazon.com\/Lisp-Small-Pieces-Christian-Queinnec\/dp\/0521545668\/?tag=fogus-20\">\n<img src=\"http:\/\/photo.goodreads.com\/books\/1181590037l\/1168500.jpg\" alt=\"LiSP\" \/><\/a><\/p><h4>LISP 1.5 Programmer\u2019s Manual by John McCarthy<\/h4><p><a href=\"http:\/\/www.softwarepreservation.org\/projects\/LISP\/book\/LISP%201.5%20Programmers%20Manual.pdf\">\n<img src=\"http:\/\/i43.tower.com\/images\/mm101111984\/lisp-1-5-programmers-manual-john-mccarthy-paperback-cover-art.jpg\" alt=\"lisp15\" \/><\/a><\/p><h4>Applicative Higher-order Programming by Stefan Sokolowski<\/h4><p><a href=\"http:\/\/www.amazon.com\/dp\/B0041V599Q?tag=fogus-20\">\n<img src=\"http:\/\/img51.imageshack.us\/img51\/5781\/001d0115medium.jpg\" alt=\"applicative\" \/><\/a><\/p><h4>PAIP<\/h4><p><a href=\"http:\/\/www.amazon.com\/Paradigms-Artificial-Intelligence-Programming-Studies\/dp\/1558601910\/?tag=fogus-20\">\n<img src=\"http:\/\/redditorschoice.com\/img\/paradigms-artificial-intelligence-programming-case-studies-common-lisp_3618_500.jpg\" alt=\"paip\" \/><\/a><\/p><p>:F<\/p><div class=\"footnotes\"><hr \/><ol><li id=\"fn:1\">\n<p>This is one of those \u201cI wish I had written this\u201d books.\u00a0<a href=\"#fnref:1\" rev=\"footnote\">\u21a9<\/a><\/p>\n<\/li>\n<li id=\"fn:2\">\n<p>Appel also wrote a series of <a href=\"http:\/\/www.amazon.com\/Andrew-W.-Appel\/e\/B000AP5X0W\/?tag=fogus-20\">compiler implementation books in various languages<\/a> that I\u2019ve yet to read.  Any thoughts on the series?\u00a0<a href=\"#fnref:2\" rev=\"footnote\">\u21a9<\/a><\/p>\n<\/li>\n<\/ol><\/div>","created_at":"2019-01-15T16:23:09+0000","updated_at":"2019-01-15T16:23:09+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"blog.fogus.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":778,"title":"The most advanced operating system","url":"https:\/\/ambrevar.xyz\/guix-advance\/index.html","content":"<p>\n<a href=\"https:\/\/www.gentoo.org\/\">Gentoo<\/a> together with all distributions using the <a href=\"https:\/\/wiki.gentoo.org\/wiki\/Project:Portage\">Portage<\/a> package manager are\nsource-based distributions with an iconic feature: the <a href=\"https:\/\/wiki.gentoo.org\/wiki\/Project:Portage\">USE flags<\/a>, which lets the\nuser enable features across the whole system (for instance disable sound, enable\ngraphical interface support, etc.).\n<\/p><p>\nUSE flags make it trivial to toggle features that were exposed by the packagers\n(and they have the benefit of being tested).  On the other hand, Portage does\nnot let you configure features that were not thought of in advance by the\npackager.  For instance, if a package has optional audio but the packager didn\u2019t\nexpose the corresponding USE flag, the user can\u2019t do anything about it (beside\ncreating a new package definition).\n<\/p><p>\nConversely, Guix gives you full customization for everything, albeit with a bit\nmore Scheme code. Roughly, in pseudo-code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-lisp\">(loop-over (TARGET-PACKAGES)\n  (package\n    (inherit TARGET)\n    (changes-here... including patches, build options, etc.))\n<\/pre><\/div><p>\nwill batch-define <code>TARGET-PACKAGES<\/code> with your changes.  None of the changes have\nto be carved into the package definition.  The user has 100% control over the\nchanges to make to the package at any time.\n<\/p><p>\nI\u2019ve loved Gentoo while I was using it, but after moving to Guix it became\napparent that Portage\u2019s approach to package customization is more limited.\n<\/p><ul class=\"org-ul\"><li>The USE flag system does not let you customize unplanned, arbitrary features.<\/li>\n<li>USE flags add a whole class of complexity (see the rather complex <a href=\"https:\/\/dev.gentoo.org\/~zmedico\/portage\/doc\/man\/ebuild.5.html\">atom\nsemantic<\/a>) to describe and manager feature relationships across packages.  Guix\ncompletely nullifies this complexity layer by using Guile Scheme to program\nthe relationships between packages.<\/li>\n<\/ul><p>\nMoreover, Portage suffer from the same aforementioned issue with the lack of\nproper support for multiple versions.  Indeed, USE flags make it an order of\nmagnitude worse (a frequent complaint about Portage): when incompatible USE\nflags propagate to some dependencies, the user must find a resolution manually.\nSometimes, it means that the desired feature won\u2019t be applicable (at least not\nwithout significant work on the package definitions).  A source of bad\nheadaches\u2026\n<\/p><p>\nOn the practical level, Guix provides pre-compiled packages which can be a huge\ntime-saver compared to Gentoo which does not (but Portage support binary package\ndistribution).\n<\/p><p>\nThe *BSD systems (e.g. FreeBSD) suffer from similar issues with the <code>make\nconfig<\/code> customization interface.\n<\/p>","created_at":"2019-01-15T04:33:13+0000","updated_at":"2019-01-15T04:33:13+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"ambrevar.xyz","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":777,"title":"ActivityPub - one protocol to rule them all?","url":"https:\/\/schub.io\/blog\/2018\/02\/01\/activitypub-one-protocol-to-rule-them-all.html","content":"<p>AcivityPub, a protocol that aims to unify federation and client interactions within decentralized social networks, reached its W3C Recommendation state recently. Along with their state change, they got a bit of press coverage, and obviously, the specification is known within my bubble.<\/p><p>Naturally, as I am one of the people behind diaspora*, I get asked what I think about ActivityPub and whether diaspora* will change its implementation to be compatible with it. I could be cheap and just link to <a href=\"https:\/\/xkcd.com\/927\/\" target=\"_blank\" rel=\"noopener\">xkcd-927<\/a> and be done with it, but that would neither be useful to anyone nor respectful towards all the people who worked on ActivityPub, so here we are.<\/p><h2 id=\"preface\">Preface<\/h2><p>Distributed social networking is hard. Not only are we facing a lot of very tricky challenges whenever we implement a feature, we also have to deal with a lot of different beliefs and perspectives. In itself, social networking is a very deep subject with a lot of different approaches. Just compare Facebook and Twitter! Both can be considered social networks, but their approach could not be more different. One is designed to focus on person-to-person communication and sharing items with a \u201cfriends only\u201d group, while the other is designed to have a public, more open, communication approach. It is very hard, if not impossible, to find a \u201cperfect\u201d approach to social networks. In my opinion, all different approaches have their place and their use, so I usually do not discuss this.<\/p><p>When talking about distributed social networking, you not only have to figure out how your network should look and feel like, you also need to work on the deep, technical issues. How to get content distributed. How to ensure privacy. How to ensure security. How to maintain good usability, even when the technical aspects are way more complex than they would be in centralized system. So, different approaches to this issue came up with their own solutions. As for diaspora*, we <a href=\"https:\/\/diaspora.github.io\/diaspora_federation\/\" target=\"_blank\" rel=\"noopener\">came up with a protocol<\/a> based on some already existing specifications (like, just to name one example, WebFinger for profile discovery), with some custom extensions to fit our use case. Other networks did their own work on that.<\/p><p><a href=\"https:\/\/www.w3.org\/TR\/activitypub\/\" target=\"_blank\" rel=\"noopener\">ActivityPub<\/a> is a protocol built by the W3C Social Working Group<sup id=\"fnref:1\"><a href=\"#fn:1\" class=\"footnote\">1<\/a><\/sup> that aims to unify communication in federation social networks, with the ultimate goal to have <em>all<\/em> distributed social networks exchanging content with each others. ActivityPub itself outlines some generics, mainly how objects should be exchanged, but it by itself is not the full story. ActivityPub is based on <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/\" target=\"_blank\" rel=\"noopener\">Activity Streams 2.0<\/a>, which is a specification defining the representation of actions and objects based on a JSON format, as well as the <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/\" target=\"_blank\" rel=\"noopener\">Activity Vocabulary<\/a>, which defines the actors, actions, and objects that can be used within ActivityStreams, and thus, within ActivityPub.<\/p><p>ActivityStreams is not a new invention. In fact, ActivityStreams 1.0 <a href=\"http:\/\/activitystrea.ms\/specs\/json\/1.0\/\" target=\"_blank\" rel=\"noopener\">is a published specification since 2011<\/a>. It got some implementors, but since it was only used as a presentation format for data, it never took off. ActivityStreams 2 introduced some changes to that, while keeping the main idea.<\/p><h2 id=\"generic-idea\">Generic idea<\/h2><p>The goal to define something that works for all implementations of distributed social networking is a hard one. In fact, I might even say it is impossible, but nonetheless, ActivityPub tried. If you combine ActivityPub and ActivityStreams, you will realize that these specifications offer you <em>a lot<\/em> of flexibility. With the amount of different actors, activities, objects, and the high amount of customizability defined within ActivityStreams, it should be possible to represent any kind of interaction possible. However, in my opinion, this creates more issues than it is trying to solve: the high number of possibilities makes it <em>very hard<\/em> to write server implementations and user interfaces able to handle this. I will go into details at a later point in this post.<\/p><p>It is also notable that ActivityPub is not only defining server-to-server communications. It also covers the communication between servers and the client. This is a very interesting approach, and in theory, it would solve a lot of issues. Clients could be compatible with multiple server implementations, and extensive libraries to build software could be build. In reality, however, I expect a lot of ActivityPub implementations to use some kind of customized payloads, so clients might not be interchangeable at all.<\/p><p>Also, as I outlined earlier, there are a lot of different approaches to social networking. There are a lot of design considerations around how a network should look and feel, and the clients for these networks are designed around their core concepts. Based on ActivityPub, you can build applications that share concepts with Facebook, as well as applications more like Twitter. One could assume that all compatible clients work just fine with all ActivityPub implementers, allowing the user to choose their preferred style of network. However, I think ActivityPub offers too many ambiguous ways to implement certain aspects, so I doubt this would work.<\/p><p>Ultimately, I feel like server-to-client communication should not be a part of a specification that works on server-to-server communication. There are a lot of different requirements conflicting here, and while ActivityPub\u2019s approach is nice, a more complex server that provides a <em>simple<\/em> REST API for entities to clients would be the better approach, since building a client based on ActivityPub actually makes the job harder.<\/p><h2 id=\"profile-discovery\">Profile discovery<\/h2><p>Undeniably, a really important part of social networking is being able to find and get in touch with people. It is important to ensure that finding your contacts is not too hard. This task is easy for centralized systems: they already have all the profile data. You are easily able to search for names, locations, even phone numbers. In distributed systems\u2026 not so much. Since data is distributed over a lot of nodes, there is a high chance your node never received the profile of the person you want to contact.<\/p><p>This means contacts have to exchange some kind of identifier. For diaspora* (and, in fact, a lot of current systems), we use an account identifier that looks like an email address: <code class=\"highlighter-rouge\">alice@example.com<\/code>, for example. This is a simple and obvious decision, since it includes all the relevant information for us: the username and the host. From there, we use WebFinger and well-known URLs to discover details and fetch information.<\/p><p>With ActivityPub, this is not the case. Instead of having a fixed-format account identifier, users are actually much more like JSON documents floating around on some webserver. There is no fixed URL you can query for user information if you know the username, you have to know the full URL. This can be something simple like <code class=\"highlighter-rouge\">https:\/\/example.com\/alice\/<\/code>, but obviously, this entirely depends on the server implementation, and can be more complex.<\/p><p>This is a <a href=\"https:\/\/github.com\/w3c\/activitypub\/issues\/20\" target=\"_blank\" rel=\"noopener\">design decision<\/a> made by the people in the Social Working Group. Parts of the folks there <em>really<\/em> dislike non-dereferencable<sup id=\"fnref:2\"><a href=\"#fn:2\" class=\"footnote\">2<\/a><\/sup> identifiers, and their reasoning is something along the lines of \u201cthe entire web works with full URLs, so social networking should do, too\u201d. Regardless of the discussion about non-dereferencable identifiers, this was a bad decision. It does make the protocol easier, since developers do not have to write anything about discovering accounts, but it deeply impacts the user experience.<\/p><p>I am confident that, for most non-technical users, something that just looks like an email address is <em>way<\/em> easier to remember than remembering an arbitrary URL.<\/p><h2 id=\"user-profiles-aka-actors\">User profiles aka Actors<\/h2><p>There is no such thing as a <em>profile<\/em>, really. But there is something called \u201cactor\u201d. An actor is an object that can, well, act on other objects. Users are actors. There are <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#actor-types\" target=\"_blank\" rel=\"noopener\">several actor types<\/a>: besides the to-be-expected <code class=\"highlighter-rouge\">Person<\/code>, there is also <code class=\"highlighter-rouge\">Application<\/code>, <code class=\"highlighter-rouge\">Group<\/code>, <code class=\"highlighter-rouge\">Organization<\/code>, and <code class=\"highlighter-rouge\">Service<\/code>. I really like this approach, since it makes it easy to distinguish between persons, organization profiles, and bots.<\/p><p>All of these actors inherit their attributes from the generic ActivityStream <code class=\"highlighter-rouge\">object<\/code>, so they have a lot of attributes. Besides some endpoints like <code class=\"highlighter-rouge\">inbox<\/code> and <code class=\"highlighter-rouge\">outbox<\/code>, which I will cover later, you can provide a name, a profile description, and some other details.<\/p><p>However, this list is not exhaustive. I immediately spot that a <code class=\"highlighter-rouge\">birthday<\/code> field is missing, and here comes the first catch with ActivityStreams. <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/#actors\" target=\"_blank\" rel=\"noopener\">Quoting the ActivityStreams specification<\/a> here:<\/p><blockquote>\n  <p>This specification intentionally defines Actors in only the most generalized way, stopping short of defining semantically specific properties for each. All Actor objects are specializations of Object and inherit all of the core properties common to all Objects. External vocabularies can be used to express additional detail not covered by the Activity Vocabulary.<\/p>\n<\/blockquote><p>To be fair, in this case, they go on and mention implementations <em>should<\/em><sup id=\"fnref:3\"><a href=\"#fn:3\" class=\"footnote\">3<\/a><\/sup> provide more profile details in a <a href=\"https:\/\/www.w3.org\/TR\/vcard-rdf\/\" target=\"_blank\" rel=\"noopener\">vCard<\/a>. However, even vCard will not cover all imaginable use-cases, but because of the very open and extensible nature of ActivityStreams, you are free to supply any format you want. The handling of unknown attributes is not specified, but I assume implementations would simply ignore them, which could work fine in most cases. However, this is just one of the many sources of confusion. Just as an example, if you have a public profile description and a private profile description, there is no way to map that to the existing fields, so you end up defining custom fields.<\/p><p>Ultimately, you will end up with a lot of different custom fields for different use cases, and everyone is <em>kind of<\/em> cooking their own soup anyway. But because exchanging objects still works just fine, some missing fields can cause a lot of confusion. In my example, people would very easily get confused why their private profile description is visible on some servers but not on others.<\/p><h2 id=\"collections-objects-and-activities\">Collections, Objects, and Activities<\/h2><p>ActivityPub defines <code class=\"highlighter-rouge\">Collections<\/code>, which are just a collection of objects. As such, there are two special collections for federating contents. There is the <code class=\"highlighter-rouge\">outbox<\/code> and the <code class=\"highlighter-rouge\">inbox<\/code>. If you want to send something to someone, you push an <code class=\"highlighter-rouge\">Activity<\/code> to your <code class=\"highlighter-rouge\">outbox<\/code> and the server takes care of sending it out. Likewise, all incoming activities will land in your <code class=\"highlighter-rouge\">inbox<\/code>.<\/p><p>It should be no surprise that the URLs for these collections are not static, either. Instead, they are defined in the <code class=\"highlighter-rouge\">Actor<\/code> object. This means you have to store the URLs for those in your database and use lots of different URLs, but that is not a big issue. Interestingly, this allows you to <code class=\"highlighter-rouge\">GET<\/code> another actors <code class=\"highlighter-rouge\">outbox<\/code>, effectively receiving a list of their past activities.<\/p><p>Just to make everything clear, I will explain two more important words. <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#object-types\" target=\"_blank\" rel=\"noopener\"><code class=\"highlighter-rouge\">Object<\/code>s<\/a> are items you can pass around, like <code class=\"highlighter-rouge\">Image<\/code>, <code class=\"highlighter-rouge\">Note<\/code>, or <code class=\"highlighter-rouge\">Profile<\/code>. <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#activity-types\" target=\"_blank\" rel=\"noopener\"><code class=\"highlighter-rouge\">Activities<\/code><\/a> on the other hand are the <em>actions<\/em> an <code class=\"highlighter-rouge\">Actor<\/code> does to an <code class=\"highlighter-rouge\">Object<\/code>, like <code class=\"highlighter-rouge\">Create<\/code>, <code class=\"highlighter-rouge\">Follow<\/code>, <code class=\"highlighter-rouge\">Like<\/code> and so on. The lists of available types are <em>really<\/em> long, so that is good. I am pretty sure that you can model just about any interaction you can have.<\/p><p>Servers never send <code class=\"highlighter-rouge\">Objects<\/code> alone, they are always encapsulated in an <code class=\"highlighter-rouge\">Activity<\/code>. This makes sense: a note does not just appear out of nowhere, someone has to <code class=\"highlighter-rouge\">Create<\/code> it. Overall, I really like this concept.<\/p><p>There is an option to send public activities, which is a bit like broadcasting your post into the world. Since all activities require a recipient, you have to use <a href=\"https:\/\/www.w3.org\/TR\/activitypub\/#public-addressing\" target=\"_blank\" rel=\"noopener\">a special address<\/a> to mark a post as public.<\/p><p>To avoid having to send an object into everyones inbox, there is an optional inbox called the <a href=\"https:\/\/www.w3.org\/TR\/activitypub\/#shared-inbox-delivery\" target=\"_blank\" rel=\"noopener\"><code class=\"highlighter-rouge\">sharedInbox<\/code><\/a>. The shared inbox is a bit comparable to a global receive URL per server\u2026 except it is not. Like with the other endpoints, there is no guaranteed URL, and there could be more than one shared inbox. Instead of having one inbox per server, each actor may specify a <code class=\"highlighter-rouge\">sharedInbox<\/code> attribute. When delivering activities, you only have to send it once per <code class=\"highlighter-rouge\">sharedInbox<\/code>, which may or may not be once per host. It is the receiving servers\u2019 responsibility to then route the item to individual inboxes. Also, the spec allows to simply hit all shared inboxes if an activity is marked as public.<\/p><p>Pulling\/retrieving an activity is easy, since, as expected, their IDs are just URLs <a href=\"https:\/\/www.w3.org\/TR\/activitypub\/#retrieving-objects\" target=\"_blank\" rel=\"noopener\">you can pull<\/a>. Please note that, according to the specification, it is required that a server responds to these queries. However, requiring authentication for some or all interactions is <em>optional<\/em>. I will write a bit more about this later, but keep in mind that, in theory, a server would be perfectly spec-compliant with publicly distributing non-public activities.<\/p><p>With all the flexibility in both ActivityPub and AcitivtyStreams, I was really surprised to see a really fixed relationship model between actors. Users follow each other. Here, this has several implications. When Bob is following Alice, Alice will send all public activities (or all activities sent to all followers) to Bob, but Bob has no obligation to return something. While this model works fine for applications like Twitter, I do not think it is a good generic solution.<\/p><p>In diaspora*, we do the opposite. Alice can <em>start sharing<\/em> with Bob. From that point on, Bob\u2019s host will receive public posts from Alice. Alice has the option to now target Bob in <em>limited<\/em> (i.e. non-public) posts. Also, Alice will send her private profile to Bob, and Alice\u2019s server will fetch Bob\u2019s public profile<sup id=\"fnref:4\"><a href=\"#fn:4\" class=\"footnote\">4<\/a><\/sup>. In my opinion, the \u201cspecifying to whom I push content to\u201d model is more fitting than the \u201cspecifying who should push me content\u201d model, at least for distributed social networking. In any case, adopting ActivityPub in diaspora* would cause a complete rework of that logic, as well as potentially confusing all users.<\/p><p>I have to talk about collections again. Each <code class=\"highlighter-rouge\">Actor<\/code> has two \u201ccontact collections\u201d, <code class=\"highlighter-rouge\">following<\/code> and <code class=\"highlighter-rouge\">followers<\/code>, again, with no fixed URLs, but fixed names. These can be used as a recipient collection for activities, to easily allow sending something to all of your followers.<\/p><p>I am not sure if a server can hold more collections. In diaspora*, we allow users to group contacts into what we call \u201caspects\u201d, and those aspects get used as a target for publishing content. In theory, this could be handled with having multiple collections. However, since the collection remains in the <code class=\"highlighter-rouge\">to<\/code> field<sup id=\"fnref:5\"><a href=\"#fn:5\" class=\"footnote\">5<\/a><\/sup>, this would be a bad solution for people who call their collections \u201cannoying people\u201d or \u201csecret crush\u201d. This can probably solved by using <code class=\"highlighter-rouge\">bto<\/code> or <code class=\"highlighter-rouge\">bcc<\/code>, which <a href=\"https:\/\/www.w3.org\/TR\/activitypub\/#remove-bto-bcc-before-delivery\" target=\"_blank\" rel=\"noopener\">have to be removed<\/a>, leaving an empty <code class=\"highlighter-rouge\">to<\/code> behind. However, even this is probably not a working solution, since replies are delivered based on the visible recipient fields. Directly setting everyone as <code class=\"highlighter-rouge\">to<\/code>\/<code class=\"highlighter-rouge\">cc<\/code> would not solve this either, since the contacts would be visible again and an author might not want to show everyone who is also receiving the content.<\/p><p>All in all, I am surprised by the limited possibilities here, given that everything else in ActivityPub\/ActivityStreams is designed to build literally everything.<\/p><h2 id=\"private-messaging\">Private messaging<\/h2><p>Talking about missing things I am surprised about: there is no support for private communication. Most large implementation of distributed social networking have some form of private conversation, and given the usage I see on diaspora*, it seems like this is a required feature.<\/p><p>While private conversations are technically just objects, they get exposed in a different form of user interface. While regular content is usually shown in a \u201cstream\u201d-like fashion with comments underneath it, private conversations are strictly separated and show messages in a more linear, grouped, fashion.<\/p><p>One could technically implement private conversations by using activities with only one recipient, but distinguishing between regular content and private conversations, and building a different UI for that, would be tricky. The other solution would be to use custom activities with custom objects, but that would make interoperability a matter of luck, and nothing you can rely on.<\/p><h2 id=\"general-ambiguity-in-objects-and-activities\">General ambiguity in Objects and Activities<\/h2><p>As mentioned earlier, there are <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#object-types\" target=\"_blank\" rel=\"noopener\">a lot of objects<\/a> and even more <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#activity-types\" target=\"_blank\" rel=\"noopener\">activities<\/a>. While I am fine with the activities, some objects confuse me. There is <code class=\"highlighter-rouge\">Article<\/code>, <code class=\"highlighter-rouge\">Document<\/code>, and <code class=\"highlighter-rouge\">Note<\/code>. Let\u2019s check their descriptions:<\/p><blockquote>\n  <p>Article - Represents any kind of multi-paragraph written work.<br \/>Document - Represents a document of any kind.<br \/>Note - Represents a short written work typically less than a single paragraph in length.<\/p>\n<\/blockquote><p>So\u2026 what should diaspora* use? If I have a look at the example, <code class=\"highlighter-rouge\">Document<\/code> seems to be the wrong choice, since it does not include contents, but only a URL. However, all these objects inherit all their attributes from the generic <code class=\"highlighter-rouge\">Object<\/code> type, so everything contains <code class=\"highlighter-rouge\">content<\/code>, <code class=\"highlighter-rouge\">url<\/code>, or both. Generally speaking, <em>most<\/em> posts on diaspora* are less than a single paragraph, but they might as well get really, really long.<\/p><p>Even worse: sometimes, people attach images to their posts. While every object also has an <code class=\"highlighter-rouge\">attachment<\/code> attribute, which is an array of objects to attach to, what if an user wants to be able to embed an image, a video, or a location at a very specific position?<\/p><p>If all networks were to agree on simply using <code class=\"highlighter-rouge\">Note<\/code> for post-like activities, contents would still not be exchangeable without problems. The <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-content\" target=\"_blank\" rel=\"noopener\"><code class=\"highlighter-rouge\">content<\/code><\/a> attribute is one of those. While ActivityPub seems to assume the content is always HTML and <a href=\"https:\/\/www.w3.org\/TR\/activitypub\/#source-property\" target=\"_blank\" rel=\"noopener\">thus added the <code class=\"highlighter-rouge\">source<\/code> attribute for markup languages<\/a>, this is not defined. In fact, the ActivityStreams spec <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-content\" target=\"_blank\" rel=\"noopener\">is clear that it is <em>not<\/em> always HTML<\/a>:<\/p><blockquote>\n  <p>By default, the value of content is HTML. The mediaType property can be used in the object to indicate a different content type.<\/p>\n<\/blockquote><p>So, it can be HTML, but it can also be something very different. diaspora* exclusively submits markdown, Friendica sends their BBCode by default. And since the ActivityPub specification is not limiting to HTML, this is another huge issue for interop.<\/p><p>ActivityStreams <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/#extensibility\" target=\"_blank\" rel=\"noopener\">extensibility<\/a> is another cause of concern. While it is <em>nice<\/em> to have a protocol that you can use for everything, having a specification that allows to send what basically is arbitrary JSON, is a huge danger for the compatibility between implementations. While the specification clearly warns that \u201cimplementations that rely too heavily on the use of extensions may experience reduced interoperability with other implementations\u201d, having such possibilities in the first place is maybe not a good idea if you <em>want<\/em> something that is interoperable.<\/p><h2 id=\"replies-and-interactions\">Replies and Interactions<\/h2><p>A reoccurring issue with distributed social networking is a high sensitivity to inconsistencies in the delivery of interactions (likes, comments, etc.). ActivityPub <a href=\"https:\/\/www.w3.org\/TR\/activitypub\/#inbox-forwarding\" target=\"_blank\" rel=\"noopener\">acknowledges this issue<\/a> and writes the following:<\/p><blockquote>\n  <p>The following section is to mitigate the \u201cghost replies\u201d problem which occasionally causes problems on federated networks. This problem is best demonstrated with an example.<\/p>\n  <p>Alyssa makes a post about her having successfully presented a paper at a conference and sends it to her followers collection, which includes her friend Ben. Ben replies to Alyssa\u2019s message congratulating her and includes her followers collection on the recipients. However, Ben has no access to see the members of Alyssa\u2019s followers collection, so his server does not forward his messages to their inbox. Without the following mechanism, if Alyssa were then to reply to Ben, her followers would see Alyssa replying to Ben without having ever seen Ben interacting. This would be very confusing!<\/p>\n<\/blockquote><p>Yes, indeed, that would be very confusing.<\/p><p>Let us have a look at the implementation in the diaspora* protocol first. We have a pretty easy rule: Whenever Bob interacts to something Alice shared, that interaction will be sent to the Alice\u2019s host and the Alice\u2019s host alone. Since interactions are designed to be relayable without loosing the option to validate them, the Alice\u2019s host can forward these interactions to everyone who received the shareable in the first place. Alice\u2019s host is the one who delivered the shareable, so it feels somewhat natural to also ask Alice\u2019s host to distribute the interactions.<\/p><p>That is, however, not the case with ActivityPub. In the example (and, in fact, everywhere else), it is clear that Bob has to address individual recipients in his interaction. We have seen this implementation before and <em>it is built to fail<\/em>. Just imagine this very simple case: Alice posts an activity to Bob, but has Charlie on <code class=\"highlighter-rouge\">bcc<\/code>. Bob posts a response. Bob sends his response back to Alice.<\/p><p>Since the <code class=\"highlighter-rouge\">to<\/code> field of that response does not include a collection owned by Alice\u2019s host, the ActivityPub specification clearly says Alice\u2019s server <em>must not<\/em> forward this interaction. Because of that, Charlie will never receive Bob\u2019s response. However, when Alice then responds to Bob\u2019s comment, Charlie will get Alice\u2019s response, since Alice\u2019s server knows that Charlie is on <code class=\"highlighter-rouge\">bcc<\/code>. The way of sending interactions outlined by ActivityPub does <em>not<\/em> solve the ghost reply issue. If anything, it creates more complex and confusing edge-cases, since some interactions will be forwarded, while others will not.<\/p><p>As per the ActivityStreams spec, all objects <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-replies\" target=\"_blank\" rel=\"noopener\">have a <code class=\"highlighter-rouge\">replies<\/code><\/a> property. So, a more sensible, reliable, and even more AcitivityStream\u2019y way of handling replies would probably be adding the interaction to the replies collection and sending an <code class=\"highlighter-rouge\">update<\/code>.<\/p><p>I saved this point for the end, as it is both the most easy and the most difficult point in case of ActivityPub. Technically, all of the points in this sub-headline are not part of the specification, and thus, I could say \u201cthere is no such thing in ActivityPub\u201d. In reality, though, this is not so easy.<\/p><p>While it is true that these issues are not part of the official specification, there is <a href=\"https:\/\/www.w3.org\/wiki\/SocialCG\/ActivityPub\/Authentication_Authorization\" target=\"_blank\" rel=\"noopener\">a wiki page<\/a> talking about authentication and authorization. These notes are, as I said, not part of any specification and are mere best practice reports. It is <em>not<\/em> safe to assume all implementation will follow those, and it is <em>not<\/em> safe to assume these are reviewed, or even stable. That being said, I still want to share my thoughts on them.<\/p><p>For client-to-server authentication, they suggest using OAuth 2.0. For working around the client registration on multiple endpoints, they suggest using OpenIDs Dynamic Client Registration. There is nothing bad to say about this; I really like it.<\/p><p>Now for the tricky part. As I mentioned earlier, some endpoints may allow accessing potentially private data. A good example for that is retrieving an activity from its URL. When you do not put any authentication in place, everyone can get access to any activity as long as the ID is known. This is not good.<\/p><p>In the wiki page, they suggest <a href=\"https:\/\/tools.ietf.org\/html\/draft-cavage-http-signatures-08\" target=\"_blank\" rel=\"noopener\">HTTP signatures<\/a> for authorizing HTTP requests. While this is surely not the prettiest method, it is probably the most appropriate in this case, since ActivityPub highly depends on having everything accessible via plain HTTP anyway. Basically, if Alice wanted to fetch an activity from Bob, Alice\u2019s server would sign the HTTP request with Alice\u2019s private key, so Bob can validate that with the cached version of her public key.<\/p><p>As for actually signing the contents of an activity, not just the transport request, the Working Group suggests using <a href=\"https:\/\/w3c-dvcg.github.io\/ld-signatures\/\" target=\"_blank\" rel=\"noopener\">Linked Data Signatures<\/a>. This would work just fine and indeed ensure that contents cannot be modified. But there is a catch. The working group claims that \u201cfor most forms of delivery, HTTP Signatures is enough\u201d. I disagree.<\/p><p>In fact, I would love to see signatures being <em>required<\/em> everywhere. When distributing contents, for example as a payload in something like a <em>reshare<\/em>, it would be way too easy to alter the contents without a signature. If they, for whatever reason, do not want to have signatures, they could just stop sending payloads entirely, and just distribute IDs which the receiving hosts could then fetch.<\/p><p>Another important aspect, which is unfortunately completely missing, is encryption. Communication within ActivityPub <em>cannot be encrypted<\/em>. If you consider that the specs do not even require <code class=\"highlighter-rouge\">https<\/code>, this is really bad. Although they <em>prefer<\/em> using <code class=\"highlighter-rouge\">https<\/code>, the specification lists <code class=\"highlighter-rouge\">http<\/code> as a valid protocol as well. So you might end up distributing potentially private contents unsigned and unencrypted, and you might not even be aware of it.<\/p><p>Distributed social networks are distributed for a reason, and privacy and data ownership are two of them, for sure. Signatures for all contents and encryption for private contents should be mandatory. They are mandatory in diaspora* since its beginning, and I cannot come up with a reason against that.<\/p><h2 id=\"summary\">Summary<\/h2><p>Okay, I\u2019ll make this <em>short<\/em> after this huge post, I promise.<\/p><p>ActivityPub tries to work for everything and everyone. And because of that, they introduced a lot of flexibility and, sadly, a lot of ambiguity. Even though they tried, I found some reasons as for why we, as diaspora* developers, would not be able to build upon this new protocol without using heavily customized objects and activities.<\/p><p>Due to my past personal and professional<sup id=\"fnref:6\"><a href=\"#fn:6\" class=\"footnote\">6<\/a><\/sup> life, I do not like specifications and protocols that allow implementations to <em>do their own thing<\/em>. If you want your specification to be interoperable, you have to take care of that. Even the slightest wiggle room will cause issues sooner or later, and when you are dealing with peoples\u2019 social networks, this causes unsatisfied and confused users very quickly. This is not the way to grow.<\/p><p>In my personal opinion, a <em>very strict<\/em> specification would be the only way for me to be comfortable. And I am not even talking about limiting everyone to one set of fields and objects that everyone has to use. Instead, in my opinion, a specification should define all possible objects\u2019 types and all possible fields, define which attributes are optional and which are required, and rule out any room for interpretation or extension outside this specification. If an implementation needs additional fields to build the features they want, the specification maintainers should adapt the spec accordingly after some discussion and consideration.<\/p><p>I am aware that the W3C is not the right organization for such an approach. I do not expect to see a <em>Living Standards<\/em> in the W3C and maybe it is not fair to judge ActivityPub by its framework. They did a good job, and for sure, they are not able to frequently update the list of allowed fields, so I do not expect that. Still, I believe that a \u201cLiving Federation Protocol\u201d would be the only way to build something that works for <em>most<\/em> folks, can be considered interoperable by default, and can quickly adapt to new requirements.<\/p><p>As much as I hope that ActivityPub and the related documents get traction and find widespread use, we, as the diaspora* project, and I personally, am neither comfortable nor <em>able<\/em> to support it. For diaspora*, it would bring a lot of downsides with no real upside. Even if we would migrate our federation, we would still be incompatible with a lot of stuff, and probably have a more inconsistent behavior altogether. And as for me, personally, I do not feel comfortable suggesting ActivityPub as a specification that ensures interoperability. Some of the core design decisions make everything way too undefined for me to be comfortable claiming so.<\/p><p><strong>Update 2019-01-13<\/strong>: I published a <em>second part<\/em> of this post, <a href=\"https:\/\/schub.io\/blog\/2019\/01\/13\/activitypub-final-thoughts-one-year-later.html\"><em>\u201cActivityPub - Final thoughts, one year later.\u201d<\/em><\/a>, with more thoughts on the general development, specifications, and other related topics.<\/p><div class=\"footnotes\"><ol><li id=\"fn:1\">\n      <p>One thing I should add: I did not contribute to the SocialWGs efforts, and thus to the specifications I am writing about, in any meaningful manner. While I did read the IRC channels, meeting minutes, and attended some meetings, at the point I was able to contribute, it was pretty clear that the goal of this working group was to build something based upon an improved version of ActivityStreams. Since I disagreed with this decision in the first place, me getting involved in the working groups efforts would actually have hampered their work. However, because of my lack of involvement, you are free to think my opinion is a bit hypocritical, and that is fine.\u00a0<a href=\"#fnref:1\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:2\">\n      <p>I actually think that, although the <code class=\"highlighter-rouge\">[user]@[host]<\/code> address does violate some Linked Data concepts, it is still pretty easy to look up. However, this is not a fight I want to be a part of.\u00a0<a href=\"#fnref:2\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:3\">\n      <p>Note that, as usual with specs, this is an <a href=\"https:\/\/tools.ietf.org\/html\/rfc2119\" target=\"_blank\" rel=\"noopener\">RFC 2119<\/a> SHOULD, which means, it is a recommendation, but not binding. You would be perfectly spec compliant with ignoring it. In fact, Mastodon, for example, does not provide a vCard.\u00a0<a href=\"#fnref:3\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:4\">\n      <p>This profile exchange is mainly done for cryptographic reasons, since we use that to exchange the public keys.\u00a0<a href=\"#fnref:4\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:5\">\n      <p>At least, this is how I understand the document. It\u2019s not written otherwise, so I assume so.\u00a0<a href=\"#fnref:5\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:6\">\n      <p>My work includes figuring out why web things work in some browsers but not in others, and you would be surprised if I told you how creatively people can misunderstand specifications, even if there is no obvious room for that.\u00a0<a href=\"#fnref:6\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>","created_at":"2019-01-14T15:35:51+0000","updated_at":"2019-01-14T15:35:51+0000","annotations":[],"mimetype":"text\/html","reading_time":23,"domain_name":"schub.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":776,"title":"Example: Deploying WordPress and MySQL with Persistent Volumes","url":"https:\/\/kubernetes.io\/docs\/tutorials\/stateful-application\/mysql-wordpress-persistent-volume\/","content":"<p>\n  <a href=\"https:\/\/github.com\/kubernetes\/website\/edit\/master\/content\/en\/docs\/tutorials\/stateful-application\/mysql-wordpress-persistent-volume.md\" id=\"editPageButton\" target=\"_blank\">\n    Edit This Page\n  <\/a>\n<\/p><p>This tutorial shows you how to deploy a WordPress site and a MySQL database using Minikube. Both applications use PersistentVolumes and PersistentVolumeClaims to store data.<\/p><p>A <a href=\"https:\/\/kubernetes.io\/docs\/concepts\/storage\/persistent-volumes\/\">PersistentVolume<\/a> (PV) is a piece of storage in the cluster that has been manually provisioned by an administrator, or dynamically provisioned by Kubernetes using a <a href=\"https:\/\/kubernetes.io\/docs\/concepts\/storage\/storage-classes\">StorageClass<\/a>.  A <a href=\"https:\/\/kubernetes.io\/docs\/concepts\/storage\/persistent-volumes\/#persistentvolumeclaims\">PersistentVolumeClaim<\/a> (PVC) is a request for storage by a user that can be fulfilled by a PV. PersistentVolumes and PersistentVolumeClaims are independent from Pod lifecycles and preserve data through restarting, rescheduling, and even deleting Pods.<\/p><blockquote class=\"warning\">\n  <p><strong>Warning:<\/strong> This deployment is not suitable for production use cases, as it uses single instance WordPress and MySQL Pods. Consider using <a href=\"https:\/\/github.com\/kubernetes\/charts\/tree\/master\/stable\/wordpress\" target=\"_blank\">WordPress Helm Chart<\/a> to deploy WordPress in production.<\/p>\n<\/blockquote><blockquote class=\"note\">\n  <p><strong>Note:<\/strong> The files provided in this tutorial are using GA Deployment APIs and are specific to kubernetes version 1.9 and later. If you wish to use this tutorial with an earlier version of Kubernetes, please update the API version appropriately, or reference earlier versions of this tutorial.<\/p>\n<\/blockquote><ul id=\"markdown-toc\"><li><a href=\"#objectives\">Objectives<\/a><\/li>\n<li><a href=\"#before-you-begin\">Before you begin<\/a><\/li>\n<li><a href=\"#create-persistentvolumeclaims-and-persistentvolumes\">Create PersistentVolumeClaims and PersistentVolumes<\/a><\/li>\n<li><a href=\"#create-a-secret-for-mysql-password\">Create a Secret for MySQL Password<\/a><\/li>\n<li><a href=\"#deploy-mysql\">Deploy MySQL<\/a><\/li>\n<li><a href=\"#deploy-wordpress\">Deploy WordPress<\/a><\/li>\n<li><a href=\"#cleaning-up\">Cleaning up<\/a><\/li>\n<li><a href=\"#what-s-next\">What's next<\/a><\/li>\n<\/ul><h2 id=\"objectives\">Objectives<\/h2><ul><li>Create PersistentVolumeClaims and PersistentVolumes<\/li>\n<li>Create a Secret<\/li>\n<li>Deploy MySQL<\/li>\n<li>Deploy WordPress<\/li>\n<li>Clean up<\/li>\n<\/ul><h2 id=\"before-you-begin\">Before you begin<\/h2><p>You need to have a Kubernetes cluster, and the kubectl command-line tool must\nbe configured to communicate with your cluster. If you do not already have a\ncluster, you can create one by using\n<a href=\"https:\/\/kubernetes.io\/docs\/setup\/minikube\">Minikube<\/a>,\nor you can use one of these Kubernetes playgrounds:<\/p><ul><li><a href=\"https:\/\/www.katacoda.com\/courses\/kubernetes\/playground\" target=\"_blank\">Katacoda<\/a><\/li>\n<li><a href=\"http:\/\/labs.play-with-k8s.com\/\" target=\"_blank\">Play with Kubernetes<\/a><\/li>\n<\/ul><p>To check the version, enter <code>kubectl version<\/code>.<\/p><p>Download the following configuration files:<\/p><ol><li><p><a href=\"https:\/\/kubernetes.io\/examples\/application\/wordpress\/mysql-deployment.yaml\">mysql-deployment.yaml<\/a><\/p><\/li>\n<li><p><a href=\"https:\/\/kubernetes.io\/examples\/application\/wordpress\/wordpress-deployment.yaml\">wordpress-deployment.yaml<\/a><\/p><\/li>\n<\/ol><h2 id=\"create-persistentvolumeclaims-and-persistentvolumes\">Create PersistentVolumeClaims and PersistentVolumes<\/h2><p>MySQL and Wordpress each require a PersistentVolume to store data.  Their PersistentVolumeClaims will be created at the deployment step.<\/p><p>Many cluster environments have a default StorageClass installed.  When a StorageClass is not specified in the PersistentVolumeClaim, the cluster\u2019s default StorageClass is used instead.<\/p><p>When a PersistentVolumeClaim is created, a PersistentVolume is dynamically provisioned based on the StorageClass configuration.<\/p><blockquote class=\"warning\">\n  <div><strong>Warning:<\/strong>In local clusters, the default StorageClass uses the<code>hostPath<\/code>provisioner.<code>hostPath<\/code>volumes are only suitable for development and testing. With<code>hostPath<\/code>volumes, your data lives in<code>\/tmp<\/code>on the node the Pod is scheduled onto and does not move between nodes. If a Pod dies and gets scheduled to another node in the cluster, or the node is rebooted, the data is lost.<\/div>\n<\/blockquote><blockquote class=\"note\">\n  <div><strong>Note:<\/strong>If you are bringing up a cluster that needs to use the<code>hostPath<\/code>provisioner, the<code>--enable-hostpath-provisioner<\/code>flag must be set in the<code>controller-manager<\/code>component.<\/div>\n<\/blockquote><blockquote class=\"note\">\n  <p><strong>Note:<\/strong> If you have a Kubernetes cluster running on Google Kubernetes Engine, please follow <a href=\"https:\/\/cloud.google.com\/kubernetes-engine\/docs\/tutorials\/persistent-disk\" target=\"_blank\">this guide<\/a>.<\/p>\n<\/blockquote><h2 id=\"create-a-secret-for-mysql-password\">Create a Secret for MySQL Password<\/h2><p>A <a href=\"https:\/\/kubernetes.io\/docs\/concepts\/configuration\/secret\/\">Secret<\/a> is an object that stores a piece of sensitive data like a password or key. The manifest files are already configured to use a Secret, but you have to create your own Secret.<\/p><ol><li><p>Create the Secret object from the following command. You will need to replace\n<code>YOUR_PASSWORD<\/code> with the password you want to use.<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl create secret generic mysql-pass --from-literal=password=YOUR_PASSWORD<\/pre><\/div><\/li>\n<li><p>Verify that the Secret exists by running the following command:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl get secrets<\/pre><\/div>\n<p>The response should be like this:<\/p>\n<pre>  NAME                  TYPE                    DATA      AGE\n  mysql-pass            Opaque                  1         42s\n<\/pre><\/li>\n<\/ol><blockquote class=\"note\">\n  <div><strong>Note:<\/strong>To protect the Secret from exposure, neither<code>get<\/code>nor<code>describe<\/code>show its contents.<\/div>\n<\/blockquote><h2 id=\"deploy-mysql\">Deploy MySQL<\/h2><p>The following manifest describes a single-instance MySQL Deployment. The MySQL container mounts the PersistentVolume at \/var\/lib\/mysql. The <code>MYSQL_ROOT_PASSWORD<\/code> environment variable sets the database password from the Secret.<\/p><table class=\"includecode\" id=\"application-wordpress-mysql-deployment-yaml\"><thead><tr><th>\n                <a href=\"https:\/\/raw.githubusercontent.com\/kubernetes\/website\/master\/content\/en\/examples\/application\/wordpress\/mysql-deployment.yaml\" download=\"application\/wordpress\/mysql-deployment.yaml\">\n                    <code>application\/wordpress\/mysql-deployment.yaml<\/code>\n                <\/a>\n                <img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/951ae1fcc65e28202164b32c13fa7ae04fab4a0b\/b77dc\/images\/copycode.svg\" title=\"Copy application\/wordpress\/mysql-deployment.yaml to clipboard\" alt=\"image\" \/><\/th>\n        <\/tr><\/thead><tbody><tr><td><div class=\"highlight\"><pre class=\"language-yaml\" data-lang=\"yaml\">apiVersion: v1\nkind: Service\nmetadata:\n  name: wordpress-mysql\n  labels:\n    app: wordpress\nspec:\n  ports:\n    - port: 3306\n  selector:\n    app: wordpress\n    tier: mysql\n  clusterIP: None\n---\napiVersion: v1\nkind: PersistentVolumeClaim\nmetadata:\n  name: mysql-pv-claim\n  labels:\n    app: wordpress\nspec:\n  accessModes:\n    - ReadWriteOnce\n  resources:\n    requests:\n      storage: 20Gi\n---\napiVersion: apps\/v1 # for versions before 1.9.0 use apps\/v1beta2\nkind: Deployment\nmetadata:\n  name: wordpress-mysql\n  labels:\n    app: wordpress\nspec:\n  selector:\n    matchLabels:\n      app: wordpress\n      tier: mysql\n  strategy:\n    type: Recreate\n  template:\n    metadata:\n      labels:\n        app: wordpress\n        tier: mysql\n    spec:\n      containers:\n      - image: mysql:5.6\n        name: mysql\n        env:\n        - name: MYSQL_ROOT_PASSWORD\n          valueFrom:\n            secretKeyRef:\n              name: mysql-pass\n              key: password\n        ports:\n        - containerPort: 3306\n          name: mysql\n        volumeMounts:\n        - name: mysql-persistent-storage\n          mountPath: \/var\/lib\/mysql\n      volumes:\n      - name: mysql-persistent-storage\n        persistentVolumeClaim:\n          claimName: mysql-pv-claim\n<\/pre><\/div>  <\/td>\n        <\/tr><\/tbody><\/table><ol><li><p>Deploy MySQL from the <code>mysql-deployment.yaml<\/code> file:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl create -f https:\/\/k8s.io\/examples\/application\/wordpress\/mysql-deployment.yaml<\/pre><\/div><\/li>\n<li><p>Verify that a PersistentVolume got dynamically provisioned. Note that it can\nIt can take up to a few minutes for the PVs to be provisioned and bound.<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl get pvc<\/pre><\/div>\n<p>The response should be like this:<\/p>\n<pre>  NAME             STATUS    VOLUME                                     CAPACITY ACCESS MODES   STORAGECLASS   AGE\n  mysql-pv-claim   Bound     pvc-91e44fbf-d477-11e7-ac6a-42010a800002   20Gi     RWO            standard       29s\n<\/pre><\/li>\n<li><p>Verify that the Pod is running by running the following command:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl get pods<\/pre><\/div>\n<blockquote class=\"note\">\n<div><strong>Note:<\/strong><pre>  It can take up to a few minutes for the Pod\u2019s Status to be <code>RUNNING<\/code>.\n<\/pre><\/div>\n<\/blockquote>\n<p>The response should be like this:<\/p>\n<pre>  NAME                               READY     STATUS    RESTARTS   AGE\n  wordpress-mysql-1894417608-x5dzt   1\/1       Running   0          40s\n<\/pre><\/li>\n<\/ol><h2 id=\"deploy-wordpress\">Deploy WordPress<\/h2><p>The following manifest describes a single-instance WordPress Deployment and Service. It uses many of the same features like a PVC for persistent storage and a Secret for the password. But it also uses a different setting: <code>type: LoadBalancer<\/code>. This setting exposes WordPress to traffic from outside of the cluster.<\/p><table class=\"includecode\" id=\"application-wordpress-wordpress-deployment-yaml\"><thead><tr><th>\n                <a href=\"https:\/\/raw.githubusercontent.com\/kubernetes\/website\/master\/content\/en\/examples\/application\/wordpress\/wordpress-deployment.yaml\" download=\"application\/wordpress\/wordpress-deployment.yaml\">\n                    <code>application\/wordpress\/wordpress-deployment.yaml<\/code>\n                <\/a>\n                <img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/951ae1fcc65e28202164b32c13fa7ae04fab4a0b\/b77dc\/images\/copycode.svg\" title=\"Copy application\/wordpress\/wordpress-deployment.yaml to clipboard\" alt=\"image\" \/><\/th>\n        <\/tr><\/thead><tbody><tr><td><div class=\"highlight\"><pre class=\"language-yaml\" data-lang=\"yaml\">apiVersion: v1\nkind: Service\nmetadata:\n  name: wordpress\n  labels:\n    app: wordpress\nspec:\n  ports:\n    - port: 80\n  selector:\n    app: wordpress\n    tier: frontend\n  type: LoadBalancer\n---\napiVersion: v1\nkind: PersistentVolumeClaim\nmetadata:\n  name: wp-pv-claim\n  labels:\n    app: wordpress\nspec:\n  accessModes:\n    - ReadWriteOnce\n  resources:\n    requests:\n      storage: 20Gi\n---\napiVersion: apps\/v1 # for versions before 1.9.0 use apps\/v1beta2\nkind: Deployment\nmetadata:\n  name: wordpress\n  labels:\n    app: wordpress\nspec:\n  selector:\n    matchLabels:\n      app: wordpress\n      tier: frontend\n  strategy:\n    type: Recreate\n  template:\n    metadata:\n      labels:\n        app: wordpress\n        tier: frontend\n    spec:\n      containers:\n      - image: wordpress:4.8-apache\n        name: wordpress\n        env:\n        - name: WORDPRESS_DB_HOST\n          value: wordpress-mysql\n        - name: WORDPRESS_DB_PASSWORD\n          valueFrom:\n            secretKeyRef:\n              name: mysql-pass\n              key: password\n        ports:\n        - containerPort: 80\n          name: wordpress\n        volumeMounts:\n        - name: wordpress-persistent-storage\n          mountPath: \/var\/www\/html\n      volumes:\n      - name: wordpress-persistent-storage\n        persistentVolumeClaim:\n          claimName: wp-pv-claim\n<\/pre><\/div>  <\/td>\n        <\/tr><\/tbody><\/table><ol><li><p>Create a WordPress Service and Deployment from the <code>wordpress-deployment.yaml<\/code> file:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl create -f https:\/\/k8s.io\/examples\/application\/wordpress\/wordpress-deployment.yaml<\/pre><\/div><\/li>\n<li><p>Verify that a PersistentVolume got dynamically provisioned:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl get pvc<\/pre><\/div>\n<blockquote class=\"note\">\n<div><strong>Note:<\/strong><pre>  It can take up to a few minutes for the PVs to be provisioned and bound.\n<\/pre><\/div>\n<\/blockquote>\n<p>The response should be like this:<\/p>\n<pre>  NAME             STATUS    VOLUME                                     CAPACITY ACCESS MODES   STORAGECLASS   AGE\n  wp-pv-claim      Bound     pvc-e69d834d-d477-11e7-ac6a-42010a800002   20Gi     RWO            standard       7s\n<\/pre><\/li>\n<li><p>Verify that the Service is running by running the following command:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl get services wordpress<\/pre><\/div>\n<p>The response should be like this:<\/p>\n<pre>  NAME        TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)        AGE\n  wordpress   ClusterIP   10.0.0.89    &lt;pending&gt;     80:32406\/TCP   4m\n<\/pre>\n<blockquote class=\"note\">\n<div><strong>Note:<\/strong><pre>  Minikube can only expose Services through <code>NodePort<\/code>. The EXTERNAL-IP is always pending.\n<\/pre><\/div>\n<\/blockquote><\/li>\n<li><p>Run the following command to get the IP Address for the WordPress Service:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  minikube service wordpress --url<\/pre><\/div>\n<p>The response should be like this:<\/p>\n<pre>  http:\/\/1.2.3.4:32406\n<\/pre><\/li>\n<li><p>Copy the IP address, and load the page in your browser to view your site.<\/p><\/li>\n<\/ol><p>You should see the WordPress set up page similar to the following screenshot.<\/p><p><img src=\"https:\/\/raw.githubusercontent.com\/kubernetes\/examples\/master\/mysql-wordpress-pd\/WordPress.png\" alt=\"wordpress-init\" \/><\/p><blockquote class=\"warning\">\n  <div><strong>Warning:<\/strong>Do not leave your WordPress installation on this page. If another user finds it, they can set up a website on your instance and use it to serve malicious content.<p>Either install WordPress by creating a username and password or delete your instance.<\/p><\/div>\n<\/blockquote><h2 id=\"cleaning-up\">Cleaning up<\/h2><ol><li><p>Run the following command to delete your Secret:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl delete secret mysql-pass<\/pre><\/div><\/li>\n<li><p>Run the following commands to delete all Deployments and Services:<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl delete deployment -l app=wordpress\n  kubectl delete service -l app=wordpress<\/pre><\/div><\/li>\n<li><p>Run the following commands to delete the PersistentVolumeClaims.  The dynamically provisioned PersistentVolumes will be automatically deleted.<\/p>\n<div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">  kubectl delete pvc -l app=wordpress<\/pre><\/div><\/li>\n<\/ol><h2 id=\"what-s-next\">What's next<\/h2><ul><li>Learn more about <a href=\"https:\/\/kubernetes.io\/docs\/tasks\/debug-application-cluster\/debug-application-introspection\/\">Introspection and Debugging<\/a><\/li>\n<li>Learn more about <a href=\"https:\/\/kubernetes.io\/docs\/concepts\/workloads\/controllers\/jobs-run-to-completion\/\">Jobs<\/a><\/li>\n<li>Learn more about <a href=\"https:\/\/kubernetes.io\/docs\/tasks\/access-application-cluster\/port-forward-access-application-cluster\/\">Port Forwarding<\/a><\/li>\n<li>Learn how to <a href=\"https:\/\/kubernetes.io\/docs\/tasks\/debug-application-cluster\/get-shell-running-container\/\">Get a Shell to a Container<\/a><\/li>\n<\/ul><h2>Feedback<\/h2><p class=\"feedback--prompt\">Was this page helpful?<\/p>","created_at":"2019-01-14T11:15:40+0000","updated_at":"2019-01-14T11:15:40+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"kubernetes.io","preview_picture":"https:\/\/kubernetes.io\/images\/kubernetes-horizontal-color.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":775,"title":"Running WordPress in a Kubernetes Cluster","url":"https:\/\/deliciousbrains.com\/running-wordpress-kubernetes-cluster\/","content":"<img width=\"1540\" height=\"748\" src=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/10131149\/db-WPKubernetesCluster-1540x748.jpg\" class=\"featured wp-post-image\" alt=\"\" srcset=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/10131149\/db-WPKubernetesCluster-1540x748.jpg 1540w, https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/10131149\/db-WPKubernetesCluster-385x187.jpg 385w, https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/10131149\/db-WPKubernetesCluster-770x374.jpg 770w\" \/><p>As a developer I try to keep my eye on the progression of technologies that I might not use every day, but are important to understand as they might indirectly affect my work. For example the recent rise of containerization, <a href=\"http:\/\/www.zdnet.com\/article\/what-is-docker-and-why-is-it-so-darn-popular\/\">popularized by Docker<\/a>, used for hosting web apps at scale. I\u2019m not technically a devops person but as I build web apps on a daily basis it\u2019s good for me to keep my eye on how these technologies are progressing.<\/p><p>A good example of this progression is the rapid development of container orchestration platforms that allow you to easily deploy, scale and manage containerized applications. The main players at the moment seem to be <a href=\"https:\/\/kubernetes.io\/\">Kubernetes (by Google)<\/a>, <a href=\"https:\/\/docs.docker.com\/engine\/swarm\/\">Docker Swarm<\/a> and <a href=\"http:\/\/mesos.apache.org\/\">Apache Mesos<\/a>. If you want a good intro to each of these technologies and their differences I recommend giving <a href=\"https:\/\/mesosphere.com\/blog\/docker-vs-kubernetes-vs-apache-mesos\/\">this article<\/a> a read.<\/p><p>In this article, we\u2019re going to start simple and take a look at the Kubernetes platform and how you can set up a WordPress site on a single node cluster on your local machine.<\/p><h2 id=\"installing-kubernetes\">Installing Kubernetes<\/h2><p>The <a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/\">Kubernetes docs<\/a> have a great interactive tutorial that covers a lot of this stuff but for the purpose of this article I\u2019m just going to cover installation and usage on macOS.<\/p><p>The first thing we need to do is install Kubernetes on your local machine. We\u2019re going to use a tool called <a href=\"https:\/\/kubernetes.io\/docs\/getting-started-guides\/minikube\/\">Minikube<\/a> which is specifically designed to make it easy to set up a Kubernetes cluster on your local machine for testing.<\/p><p>As per the Minikube docs, there are a few prerequisites before we get going. Make sure you have a Hypervisor installed (\u2018m going to use Virtualbox). Next we need to <a href=\"https:\/\/kubernetes.io\/docs\/tasks\/tools\/install-kubectl\/\">install the Kubernetes command-line tool<\/a> (known as <code>kubectl<\/code>). If you use Homebrew this is as simple as running:<\/p><pre>$ brew install kubectl\n<\/pre><p>Now we can actually <a href=\"https:\/\/github.com\/kubernetes\/minikube\/releases\">install Minikube<\/a>:<\/p><pre>$ curl -Lo minikube https:\/\/storage.googleapis.com\/minikube\/releases\/v0.21.0\/minikube-darwin-amd64 &amp;&amp; chmod +x minikube &amp;&amp; sudo mv minikube \/usr\/local\/bin\/\n<\/pre><p>Finally we want to <a href=\"https:\/\/kubernetes.io\/docs\/getting-started-guides\/minikube\/#quickstart\">start Minikube<\/a> which will create a virtual machine which will act as our single-node Kubernetes cluster. At this point I should state that, although we\u2019re running things locally in this article, most of the following concepts will apply when running a full Kubernetes cluster on <a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/\">real servers<\/a>. On a multi-node cluster a \u201cmaster\u201d node would be responsible for managing the other worker nodes (VM\u2019s or physical servers) and Kubernetes would automate the distribution and scheduling of application containers across the cluster.<\/p><pre>$ minikube start --vm-driver=virtualbox\n<\/pre><h2 id=\"installing-helm\">Installing Helm<\/h2><p>At this point we should now have a (single node) Kubernetes cluster running on our local machine. We can now interact with Kubernetes in any way we want. I found <a href=\"http:\/\/kubernetesbyexample.com\/\">kubernetesbyexample.com<\/a> to be a good introduction to Kubernetes concepts and terms if you want to start playing around.<\/p><p>While we could set things up manually, we\u2019re actually going to use a separate tool to install our WordPress application to our Kubernetes cluster. <a href=\"https:\/\/docs.helm.sh\/\">Helm<\/a> is labelled as a \u201cpackage manager for Kubernetes\u201d and works by allowing you to easily deploy pre-built software packages to your cluster, known as \u201cCharts\u201d. You can think of a Chart as a group of container definitions and configs that are designed for a specific application (such as WordPress). First let\u2019s install Helm on our local machine:<\/p><pre>$ brew install kubernetes-helm\n<\/pre><p>Next we need to install Helm on our cluster. Thankfully this is as simple as running:<\/p><pre>$ helm init\n<\/pre><h2 id=\"installing-wordpress\">Installing WordPress<\/h2><p>Now that Helm is running on our cluster we can install the <a href=\"https:\/\/kubeapps.com\/charts\/stable\/wordpress\">WordPress chart<\/a> by running:<\/p><pre>$ helm install --namespace wordpress --name wordpress --set serviceType=NodePort stable\/wordpress  \n<\/pre><p>The will install and run WordPress in a container and MariaDB in a container for the database. This is known as a \u201cPod\u201d in Kubernetes. A <a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/explore-intro\/\">Pod<\/a> is basically an abstraction that represents a group of one or more application containers and some shared resources for those containers (e.g. storage volumes, networking etc.).<\/p><p>We give the release a namespace and a name to keep things organized and make them easy to find. We also set the <code>serviceType<\/code> to <code>NodePort<\/code>. This is important because, by default, the service type will be set to <code>LoadBalancer<\/code> and, as we currently don\u2019t have a load balancer for our cluster, we wouldn\u2019t be able to access our WordPress site from outside the cluster.<\/p><p>In the last part of the output from this command you will notice some helpful instructions on how to access your WordPress site. Run these commands to get the external IP address and port for our WordPress site:<\/p><pre>$ export NODE_PORT=$(kubectl get --namespace wordpress -o jsonpath=\"{.spec.ports[0].nodePort}\" services wordpress-wordpress)\n$ export NODE_IP=$(kubectl get nodes --namespace wordpress -o jsonpath=\"{.items[0].status.addresses[0].address}\")\n$ echo http:\/\/$NODE_IP:$NODE_PORT\/admin\n<\/pre><p>You should now be able to visit the resulting URL (ignoring the <code>\/admin<\/code> bit) and see WordPress running on your very own Kubernetes cluster!<\/p><h2 id=\"scaling-wordpress\">Scaling WordPress<\/h2><p>One of the great things about container orchestration platforms such as Kubernetes is that it makes scaling and managing your application really simple. Let\u2019s check the status of our deployments:<\/p><pre>$ kubectl get deployments --namespace=wordpress\n<\/pre><p><a href=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image4.png\"><img src=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image4.png\" alt=\"kubectl get deployments\" class=\"aligncenter size-full wp-image-30814\" srcset=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image4.png 653w, https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image4-385x70.png 385w\" \/><\/a><\/p><p>We should see that we have 2 deployments, one for the Mariadb database and one for WordPress itself. Now let\u2019s say your WordPress site is starting to see a lot of traffic and we want to split the load over multiple instances. We can scale our <code>wordpress-wordpress<\/code> deployment by running a simple command:<\/p><pre>$ kubectl scale --replicas 2 deployments wordpress-wordpress --namespace=wordpress\n<\/pre><p>If we run the <code>kubectl get deployments<\/code> command again we should now see something like this:<\/p><p><a href=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120710\/image2.png\"><img src=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120710\/image2.png\" alt=\"kubectl get deployments\" class=\"aligncenter size-full wp-image-30812\" srcset=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120710\/image2.png 652w, https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120710\/image2-385x72.png 385w\" \/><\/a><\/p><p>You\u2019ve just scaled up your WordPress site! Easy peasy, right? There are now multiple WordPress containers that traffic can be load-balanced across. For more info on Kubernetes scaling check out <a href=\"https:\/\/kubernetes.io\/docs\/tutorials\/kubernetes-basics\/scale-intro\/\">this tutorial<\/a>.<\/p><h2 id=\"high-availability\">High Availability<\/h2><p>Another great feature of platforms such as Kubernetes is the ability to not only scale easily, but to provide high availability by implementing self-healing components. Say one of your WordPress deployments fails for some reason. Kubernetes will automatically replace the deployment instantly. We can simulate this by deleting one of the pods running in our WordPress deployment.<\/p><p>First get a list of pods by running:<\/p><pre>$ kubectl get pods --namespace=wordpress\n<\/pre><p><a href=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image3.png\"><img src=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image3.png\" alt=\"kubectl get pods\" class=\"aligncenter size-full wp-image-30813\" srcset=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image3.png 678w, https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120711\/image3-385x79.png 385w\" \/><\/a><\/p><p>Then delete one of the pods:<\/p><pre>$ kubectl delete pod {POD-ID} --namespace=wordpress\n<\/pre><p>If you run the <code>kubectl get pods<\/code> command again you should see Kubernetes spinning up the replacement pod straight away.<\/p><p><a href=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120709\/image1.png\"><img src=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120709\/image1.png\" alt=\"kubectl get pods\" class=\"aligncenter size-full wp-image-30811\" srcset=\"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120709\/image1.png 710w, https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/07120709\/image1-385x163.png 385w\" \/><\/a><\/p><h2 id=\"going-further\">Going Further<\/h2><p>We\u2019ve only really scratched the surface of what Kubernetes can do. If you want to delve a bit deeper, I would recommend having a look at some of the following features:<\/p><ul><li><a href=\"https:\/\/kubernetes.io\/docs\/tasks\/run-application\/horizontal-pod-autoscale\/\">Horizontal scaling<\/a><\/li>\n<li><a href=\"https:\/\/kubernetes.io\/docs\/concepts\/workloads\/controllers\/replicationcontroller\/#what-is-a-replicationcontroller\">Self healing<\/a><\/li>\n<li><a href=\"https:\/\/kubernetes.io\/docs\/concepts\/workloads\/controllers\/deployment\/#what-is-a-deployment\">Automated rollouts and rollbacks<\/a><\/li>\n<li><a href=\"https:\/\/kubernetes.io\/docs\/concepts\/configuration\/secret\/\">Secret management<\/a><\/li>\n<\/ul><p>Have you ever run WordPress on a container platform? Have you ever used Kubernetes (or another container orchestration platform) and got any good tips? How do you normally scale your WordPress sites? Let us know in the comments.<\/p>","created_at":"2019-01-14T11:01:21+0000","updated_at":"2019-01-14T11:01:21+0000","published_at":"2017-08-22T13:00:39+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":5,"domain_name":"deliciousbrains.com","preview_picture":"https:\/\/cdn.deliciousbrains.com\/content\/uploads\/2017\/08\/10131149\/db-WPKubernetesCluster-1540x748.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":774,"title":"12 Things I Learned from Chris Dixon about Startups","url":"https:\/\/a16z.com\/2015\/01\/18\/12-things-learned-from-chris-dixon-about-startups\/","content":"<div class=\"entry-content__primary\">\n\n\t\t\t<div class=\"entry-content__post-body wp-content\">\n\t\t\t\t<p><strong>#1 \u201cIf everyone loves your idea, I might be worried that it\u2019s not forward thinking enough.\u201d\u00a0<\/strong><\/p>\n<p>Anyone thinking about starting a business should be searching for mispriced opportunities. \u00a0While markets are mostly efficient in eliminating opportunities for extraordinary profit, there are always areas of an economy in which there are significant uncertainty. These areas are excellent places for a startup to look for opportunities.<\/p>\n<p>The best entrepreneurs have learned that large businesses are investing huge amounts of capital in areas in which extensive information is already available, and that it\u2019s most advantageous to create new businesses where information is not available. In other words, the best opportunities for startups tend to be in areas that are overlooked and less well-known by others. What Chris Dixon is saying is that if everyone loves your idea, this may be a \u201ctell\u201d that there will not be opportunities for extraordinary profit.\u00a0The venture investor Peter Thiel has said: \u201cThe best startups are good ideas that look like bad ideas. Good ideas that look like good ideas are already being worked on by big companies.\u201d<\/p>\n<p>Dixon\u00a0also says in\u00a0<a href=\"http:\/\/cdixon.org\/2009\/08\/22\/why-you-shouldnt-keep-your-startup-idea-secret\/\" target=\"_blank\" rel=\"noopener\">this insightful post\u00a0<\/a>that <strong>\u201cyou shouldn\u2019t keep your startup idea\u00a0secret.\u201d<\/strong> He identifies a range of positive benefits that flow from sharing the idea and getting feedback, while pointing out <strong>\u201cthere are at best a handful of people in the world who might actually drop everything and copy your idea.\u201d<\/strong><\/p>\n<p>While most businesses do not require a result that generates extraordinary profit to be a success, this is not the case for a startup that seeks to raise venture capital. Venture capitalists invest in a portfolio of startups, knowing that only one to three in every fund could likely be a massive success. Chris Dixon\u2019s partner Marc Andreessen describes the approach of a venture capitalist more technically as \u2018buying a portfolio of <em>long<\/em><strong>\u2013<\/strong><em>dated<\/em>, deeply<strong>\u2013<\/strong><em>out-of-the-money call<\/em> options\u2019. Entrepreneurs are in the business of creating those options and selling some of them to investors to fund the business.<\/p>\n<p>Like a startup or any other investor, a venture capitalist is seeking a mispriced opportunity. \u00a0All intelligent investors seek mispriced assets and if you want to explore this topic you can do no better than <a href=\"http:\/\/www.oaktreecapital.com\/MemoTree\/It's%20All%20a%20Big%20Mistake_06_20_12.pdf\" target=\"_blank\" rel=\"noopener\">reading Howard Marks<\/a>. \u00a0Why do large businesses and others leave this opportunity in areas with significant uncertainty available for startups? Howard Marks traces the source to bias and closed mindedness, capital rigidity, psychological success, and herd behavior. Markets are not fully efficient. Private, emerging and obscure markets are especially inefficient.<\/p>\n<p><strong>#2 \u201c<\/strong><strong>How do you develop a good idea that looks like a bad idea? You need to know a secret \u2014 in the <\/strong><strong>Peter Thiel sense: something you believe that most other people don\u2019t believe. How do you develop a secret?<\/strong> <strong>(a) know the tools better than anyone else; (b) know the problems better than anyone else; and\/or (c) draw from unique life experience.\u201d<\/strong><\/p>\n<p><strong>\u201cFounders have to choose a market long before they have any idea whether they will reach product\/market fit. In my opinion, the best predictor of success is whether there is what David Lee calls \u2018founder\/market fit.\u2019 Founder\/market fit means the founders have a deep understanding of the market they are entering, and are people who \u2018personify their product, business, and ultimately their company.\u2019\u201d \u00a0<\/strong><\/p>\n<p>Chris Dixon is saying that the people most likely to know the \u201ca secret\u201d about a business opportunity are people who have deep domain expertise. In other words, it is not nearly as likely that someone without deep domain expertise will be successful without understanding the technology, the best methods to create the product, the best ways to bring products to market or the needs of the customer. Another way to think about this point is in terms of a moat or sustainable competitive advantage. Founders and employees of the startup are themselves contributors to the moat of a company, both directly and indirectly.<\/p>\n<p>People who work for the startup who have deep domain knowledge are a likely source of what is called \u201coptionality\u201d which I have explained in <a href=\"http:\/\/25iq.com\/2014\/04\/05\/the-best-venture-capitalists-harvest-optionality-dealing-with-risk-uncertainty-and-ignorance\/\" target=\"_blank\" rel=\"noopener\">a previous blog post<\/a>. When a team of people in a startup\u00a0have\u00a0what Dixon called \u201csecrets\u201d as a result of deep domain expertise, their ability to adapt and innovate gives the startup and the investors optionality. Teams\u00a0that\u00a0do not have this optionality usually can\u2019t adapt to changing environments, and fail more often.<\/p>\n<p>The other skill that people with deep understanding have is the ability to see a phenomenon that is emerging\u00a0within\u00a0a complex adaptive system. When something bigger than the sum of its parts is emerging in an economy, some people with deep domain expertise are going to see the potential (the secret) before other people.<\/p>\n<p><strong>#3 \u201c[The] business of seed investing, and frankly, early-stage entrepreneurship, is so much about getting good information. And almost all of that information, unfortunately, is not published.\u201d<\/strong><\/p>\n<p>The fact that information about a business is hard to get is actually a great thing for a startup, since it can help create the mispriced opportunity they seek. Uncertainty in the early stages of a startup is the friend of the entrepreneur. As the team pushes forward to reduce technology risk, find product\/market fit and discover methods to scale the business, uncertainty is retired and value is created. The job of a great venture capitalist is in no small part to provide entrepreneurs with an entry into networks that allows them to quickly and cost-effectively find this private information. The same principle applies to the venture capitalists themselves. Great venture capitalists are always trying to find good sources of information, particularly information that is not published.<\/p>\n<p><strong>#4 \u201cIdeas \u2026matter, just not in the narrow sense in which startup ideas are popularly defined. Good startup ideas are well developed, multi-year plans that contemplate many possible paths according to how the world changes.\u201d<\/strong><\/p>\n<p><strong>\u201cCharacteristics of the best ideas: (a) powerful people dismiss them as toys; (b) they unbundle functions done by others; (c) they often start off as hobbies and\/or (d) they often challenge social norms.\u201d\u00a0<\/strong><\/p>\n<p><strong>\u201cThe best ideas come through direct experience. \u2026When you differentiate your direct experience from conventional wisdom, that\u2019s where the best startup ideas come from.\u201d<\/strong><\/p>\n<p>My deepest exposure\u00a0to what Chris Dixon is talking about immediately above (direct experience) came during the time I worked for Craig McCaw, who without question is a savant when it comes to ideas that can be developed into great businesses. \u00a0In the very early days, the \u201ccell phone\u201d only offered\u00a0enough value to be a commercial success to a very small number of users. In the beginning, the device was so big it required a suitcase or a car installation to be useful. I remember real estate agents and construction sites as the biggest users. On the infrastructure side, a city like Seattle could be served by radios on only three very tall towers.<\/p>\n<p>Eventually mobile phones appeared, but they were very expensive, analog, heavy and large. During that time period, McKinsey famously predicted\u00a0that no one would ever use a mobile phone if a land line phone was available. McKinsey placed little or no value on what Craig McCaw called the ability of people to be \u201cnomadic.\u201d The mobile phone had each of the attributes Chris Dixon noted above. Some people thought of the mobile phone as a toy. Since my first mobile phone cost more than $4,000 dollars I actually felt awkward using it in some social settings. Talking into a mobile phone in some pubic settings would cause people to frown at you.<\/p>\n<p>To continue my example, Craig McCaw\u00a0was also\u00a0an<em> enthusiastic personal user<\/em> of what we then called \u201ccellular\u201d phones. Craig McCaw loved working\u00a0out of the \u2018mobile office\u2019 \u2013 meaning cars, planes, boats and ships. Which meant he was a natural enthusiast for the product. When the time came to sell his cable TV business in order to double down on the mobile phone business, the choice was made easy by his love of the mobile phone.<\/p>\n<p>#<strong>5 \u201cThere is a widespread myth that the most important part of building a great company is coming up with a great idea.\u201d \u201c<\/strong><strong>What you should really be focused on when pitching your early stage startup is pitching yourself and your team. Of course a great way to show you can build stuff is to build a prototype of the product you are raising money for. This is why so many VCs tell entrepreneurs to \u2018come back when you have a demo.\u2019 They aren\u2019t wondering whether your product can be built \u2013 they are wondering whether you can build it.\u201d<\/strong><\/p>\n<p>Great ideas matter, as the previous quotations noted. But the ability of a team to execute is a more important consideration than a clever idea. Most everyone has said more than once \u201cI thought of that idea first\u201d when they see a new business being formed. The best idea in the world without a team to make it happen, won\u2019t amount to a hill of beans. Chris Dixon is saying that the best evidence that a team can actually execute, is actually executing on something like a demo. The best evidence that you can do something like create software, is actually creating software. As the old proverb points out, <em>the proof of the pudding is in the eating.<\/em><\/p>\n<p><strong>#6 \u201cWhat the smartest people do on the weekend is what everyone else will do during the week in ten\u00a0years.\u201d \u201cH<\/strong><strong>obbies are what the smartest people spend their time on when they aren\u2019t constrained by near-term financial goals.<\/strong><strong>\u201d\u00a0<\/strong><\/p>\n<p>Chris Dixon is not referring to smart people who play ping pong in their garage or pay fantasy baseball in their dorm rooms on weekends. He is referencing\u00a0the smartest people that are building things like the first PCs, drones, or a better search engine. \u00a0The Homebrew Computer Club, an early hobbyist group which had it first meeting on March 5, 1975, is just one example. The critical element here is an advanced technology that is useful to very smart hobbyists but does not yet have obvious financial returns associated with its use. Inevitably, technologies driven by phenomena like Moore\u2019s law drive costs to lower levels and performance to high enough levels so that what was once a hobby becomes a thriving business.<\/p>\n<p>As an aside, it is easier to understand Moore\u2019s law now. But I will say that although I am probably in the top few percent in the US who understand its power, I underestimate that power every single year.<\/p>\n<p><strong>#7 \u201cThis era of technology, it seems to be the core theme is about moving beyond bits to atoms. Meaning technology that affects real word, and transportation and housing and healthcare and all these other things, as opposed to just moving bits around. And those areas tend to be more heavily regulated and, this issue is only beginning to be significant and will probably the defining issue of the next decade in technology.\u201d\u00a0<\/strong><\/p>\n<p>I include this quotation since it is an example of something that I value, which is an ability of a person to make a genuine non-consensus prediction about the future rather than predicting the present. Too many venture capitalists are camp followers. They are moving into a trend when others have long since moved on to other opportunities. To outperform, a venture capitalist must think like Howard Marks described here: \u201cTo achieve superior investment results, your insight into value has to be superior. Thus you must learn things others don\u2019t, see things differently or do a better job of analyzing them \u2013 ideally all three.\u201d Deviating from a consensus view for its own sake is suicidal, but doing it occasionally and being very right is what makes a great venture capitalist like Dixon.<\/p>\n<p><strong>#8 \u201cAnyone who has pitched VCs knows they are obsessed with market size.\u201d\u00a0\u201cIf you can\u2019t make the case that you\u2019re addressing a possible billion dollar market, you\u2019ll have difficulty getting VCs to invest. (Smaller, venture-style investors like angels and seed funds also prioritize market size but are usually more flexible \u2013 they\u2019ll often invest when the market is \u201conly\u201d ~$100M).\u00a0 This is perfectly rational since VC returns tend to be driven by a few big hits in big markets.\u201d\u00a0<\/strong><\/p>\n<p><strong>\u201cIf you are arguing market size with a VC using a spreadsheet, you\u2019ve already lost the debate.\u201d For early-stage companies, you should never rely on quantitative analysis to estimate market size. Venture-style startups are bets on broad, secular trends. Good VCs understand this.\u201d\u00a0\u201cStartups that fill white spaces [areas where there is latent demand without supply] aren\u2019t usually world-changing companies, but they often have solid exits. They force incumbents to see a demand they had missed, and those incumbents often respond with an acquisition.\u201d<\/strong><\/p>\n<p>It is not possible to make silk purse out of a pig\u2019s ear. \u00a0For a startup to generate the necessary financial return for a VC, the potential market being addressed must be massive. The entrepreneur who pulls out a spreadsheet and tries to make the case that the market is large enough\u00a0based on fake quantitative assumptions does little but destroy his or her credibility. Venture capitalists hate to see hockey stick shaped distribution curves based on unrealistic assumptions that don\u2019t map to reality. Chris Dixon is saying that what they do want to see for markets that can\u2019t be defined with much certainty is a strongly argued narrative which explains that the market has or will attain the required size. Yes, they want to see as many related facts as possible that support the narrative. No, they don\u2019t want to hear wild guesses presented as facts.<\/p>\n<p><strong>#9 \u201cThere are two kinds of investors: Ron Conways who try to create value by finding good people and helping them create something great, and others, who want a piece of someone else\u2019s things. The builders and the extractors. Avoid the extractors.\u201d\u00a0<\/strong><\/p>\n<p><strong>\u201cFounders too often view raising capital as a transaction, when it is actually a very deep relationship. They think of money as money, when there is actually smart money, dumb money, high-integrity money, and low-integrity money.\u201d<\/strong><\/p>\n<p>Particularly in the United States, money is not the scarce resource in venture capital.\u00a0 The scarce resource is fundable startups. The outcome for any startup will increasingly be determined by access to networks of people and resources. \u00a0If a startup has a choice between (1) just money and (2) money plus access to these networks, is it wise to choose the latter.\u00a0 Because startups most compete in an Extremistan environment (i.e., winner-take-all or winner-take-most-all) even the smallest advantage can end up topping the balance of success and cumulative advantage to one company. \u00a0Perhaps some founders are cheered up by a venture capitalist who is mostly a cheerleader, but the smart entrepreneurs want someone who can directly help with tasks like recruiting and problems like pricing and distribution.<\/p>\n<p><strong>#10 \u201cVCs have a portfolio, and they want to have big wins. They\u2019d rather have a few more lottery tickets.. while for the entrepreneurs, it\u2019s their whole life, and let\u2019s say you raised five million bucks, and you have a fifty million dollar offer, and the entrepreneurs are like, \u201cLook, I make whatever millions of dollars. I\u2019ll be able to start another company.\u201d And the VCs are like, \u2018Wait!\u00a0We invested billions of dollars.\u2019 That is usually where tension comes.\u201d<\/strong><\/p>\n<p>Chris Dixon has been both a founder and a VC. He has empathy for both venture capitalists and founders on this set of issues. He is most certainly correct that this type of situation\u00a0creates tension. The wave of discussion about this topic is proof of that. The question is: what is the best way to resolve the tension in ways that are mutually beneficial? First, a brief note about what is at stake. The economist Harry Markowitz called diversification the <em>only free lunch investing. <\/em>Warren Buffett discussed the issues involved as follows:<\/p>\n<blockquote><p>\u201cOf course, some investment strategies require wide diversification. If significant risk exists in a single transaction, overall risk should be reduced by making that purchase one of many mutually-independent commitments. Thus, you may consciously purchase a risky investment \u2013 one that indeed has a significant possibility of causing loss or injury \u2013 if you believe that your gain, weighted for probabilities, considerably exceeds your loss, comparably weighted, and if you can commit to a number of similar, but unrelated opportunities.<\/p>\n<p>Most venture capitalists employ this strategy. Should you choose to pursue this course, you should adopt the outlook of the casino that owns a roulette wheel, which will want to see lots of action because it is favored by probabilities, but will refuse to accept a single, huge bet.<\/p>\n<p>Another situation requiring wide diversification occurs when an investor who does not understand the economics of specific businesses nevertheless believes it in his interest to be a long-term owner of American industry. That investor should both own a large number of equities and space out his purchases. By periodically investing in an index fund, for example, the know-nothing investor can actually out-perform most investment professionals.\u00a0 Paradoxically, when \u2018dumb\u2019 money acknowledges its limitations, it ceases to be dumb.<\/p>\n<p>On the other hand, if you are a <em>know-something<\/em> investor, able to understand business economics and to find five to ten sensibly priced companies that possess important long-term competitive advantages, conventional diversification makes no sense for you. It is apt simply to hurt your results and increase your risk. I cannot understand why an investor of that sort elects to put money into a business that is his 20th favorite rather than simply adding that money to his top choices \u2013 the businesses he understands best and that present the least risk, along with the greatest profit potential.\u00a0 In the words of the prophet Mae West: \u00a0\u2018Too much of a good thing can be wonderful.\u2019\u201d<\/p><\/blockquote>\n<p>The wisest outcome on founder and employee liquidity issues will depend on the facts and circumstances of each case. There is no connect-the-dots-formula that is right in all cases. Fred Wilson has written a <a href=\"http:\/\/avc.com\/2014\/12\/some-thoughts-on-founder-liquidity\/\" target=\"_blank\" rel=\"noopener\">very thoughtful post<\/a>\u00a0on this issue.\u00a0He points out that even from the view of the VC there is an incentive to create some liquidity:<em> \u201c<\/em>providing some founder liquidity, at the appropriate time, will incentivize the founders to have a longer term focus and\u00a0that will result in exits at much larger valuations because, contrary to popular belief, founders drive the timing of exit way more than VCs do.\u201d<\/p>\n<p>In addition to what Fred Wilson notes, it is one thing to concentrate your investments if you have a net worth measured in millions and quite another if you have little financial cushion if the business fails. The important point that Chis Dixon raises is that there is an issue here, and it can create tension if not dealt with intelligently. Founders are smarter and better informed than ever before and they want a venture capitalist who is empathetic and thoughtful.<\/p>\n<p><strong>#11 \u201cIf you aren\u2019t getting rejected on a daily basis, your goals are not ambitious enough. The most valuable lesson I had starting out in my career was when I was trying to break in the tech world and I applied to jobs at big companies and at startups, at VC firms. I got rejected everywhere. I had sort of an unusual background. I was a philosophy major, a self-taught programmer. It turned out to be the most valuable experience of my career because I eventually developed such thick skin that I just didn\u2019t care anymore about getting rejected. And, in fact, I kind of turned it around and started embracing it. I eventually \u2014 that sort of emboldened me. Through those sort of bolder tactics, eventually landed a job that got my first startup funded. So every day to this day I try to make sure I get rejected.\u201d<\/strong><\/p>\n<p>The ability to handle rejection in a sales process is something that has always fascinated me. \u00a0Why can some people knock on door after door and suffer rejection and after rejection and still maintain a positive attitude long enough to generate the eventual sale? For some people a single rejection turns them into a nervous wreck, while others power through to close a sale. It seem to me to be explained by a combination of innate personality and a learned skill.<\/p>\n<p>In any event, starting a business and even building a successful career involves way more selling than people who have never done it before imagine. Entrepreneurs are constantly selling themselves, their business and its products to potential employees, suppliers, distributors, investors and customers. If you can\u2019t sell, starting a business is probably unwise.<\/p>\n<p><strong>#12 \u201cBefore I started my first company, an experienced entrepreneur I know said, \u2018Get ready to feel sick to your stomach for the next five years.\u2019 And I was, \u2018Eh, whatever.\u2019 Then later, I was, \u2018Shoot, I should listen to the guy.\u2019\u201d\u00a0<\/strong><\/p>\n<p><strong>\u201cYou\u2019ve either started a company or you haven\u2019t. \u2018Started\u2019 doesn\u2019t mean joining as an early employee, or investing or advising or helping out. It means starting with no money, no help, no one who believes in you (except perhaps your closest friends and family), and building an organization from a borrowed cubicle with credit card debt and nowhere to sleep except the office. It almost invariably means being dismissed by arrogant investors who show up a half hour late, totally unprepared and then instead of saying \u2018no\u2019 give you non-committal rejections like \u2018we invest in later stage companies.\u2019 It means looking prospective employees in the eyes and convincing them to leave safe jobs, quit everything and throw their lot in with you. It means having pundits in the press and blogs who\u2019ve never built anything criticize you and armchair quarterback your every mistake. It means lying awake at night worrying about running out of cash and having a constant knot in your stomach during the day fearing you\u2019ll disappoint the few people who believed in you and validate your smug doubters<\/strong>.\u201d<\/p>\n<p>I was the third employee of a company founded by Craig McCaw, and although I wasn\u2019t a founder it was nevertheless a life changing experience. \u00a0It was a particularly notable day since I received two job offers the very same day. One job was a very safe position with an established company. The other was with the startup. What tipped the decision was that I wanted to have the life experience of being part of a startup. I wanted the experience more than the immediate monetary rewards that the other position offered.<\/p>\n<p>From a post by Chris Dixon on <a href=\"http:\/\/cdixon.org\/2009\/09\/19\/climbing-the-wrong-hill\/%20\" target=\"_blank\" rel=\"noopener\">climbing the wrong hill in your career<\/a>:\u00a0<strong>\u201cPeople tend to systematically overvalue near-term over long-term rewards. \u00a0This effect seems to be even stronger in more ambitious people. \u00a0Their ambition seems to make it hard for them to forgo the nearby upward step.\u201d<\/strong><\/p>\n<p>There were lots of times I thought that I had made a mistake in not taking the safer and better-paying job. I experienced all the things Chris Dixon talks about in that post, including attacks from armchair critics. I worried often about other employees and how my family would cope with failure if that happened. The tale of this startup is actually one of the great untold stores in business history. It resulted in a 4X return for early investors which was not terrible, but not great either. That\u2019s a story for another time.<\/p>\n<hr \/><p><em>Posted here by permission of the author (who blogs at <a href=\"https:\/\/25iq.com\/\" target=\"_blank\" rel=\"noopener\">25iq.com<\/a>).\u00a0<\/em><em>About the author:\u00a0<\/em><em><a href=\"https:\/\/twitter.com\/trengriffin\" target=\"_blank\" rel=\"noopener\">Tren Griffin<\/a>\u2019s professional background has primarily involved areas where business meets technologies like software and mobile communications. He currently works at Microsoft. Previously, he was a partner at private equity firm\u00a0Eagle River (established by Craig McCaw) and before that, a consultant in Asia. Griffin\u2019s latest book, <a href=\"http:\/\/www.amazon.com\/Charlie-Munger-Complete-Investor-Publishing-ebook\/dp\/B010EB3EUM\/\" target=\"_blank\" rel=\"noopener\">Charlie Munger: The Complete Investor<\/a> is about the legendary Berkshire Hathaway vice chairman, and how he\u00a0invokes a set of interdisciplinary \u201cmental models\u201d involving economics, business, psychology, ethics, and management to keep emotions out of his investments and avoid the common pitfalls of bad judgment.\u00a0<\/em><\/p>\n<p><strong>sources:<\/strong><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2009\/08\/22\/why-you-shouldnt-keep-your-startup-idea-secret\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>Why you shouldn\u2019t keep you startup idea secret\u00a0<\/em><\/a><\/p>\n<p><a href=\"https:\/\/github.com\/charlesfeng\/startup-school-2013\/blob\/master\/ChrisDixon.md\" target=\"_blank\" rel=\"noopener\">Github \u2013 Startup School 2013<\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2013\/08\/04\/the-idea-maze\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 The<em> idea maze<\/em><\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2013\/03\/03\/what-the-smartest-people-do-on-the-weekend-is-what-everyone-else-will-do-during-the-week-in-ten-years\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>What the smartest people do on the weekend<\/em><\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2011\/06\/20\/foundermarket-fit\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>Founder Market Fit<\/em>\u00a0<\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2010\/04\/03\/size-markets-using-narratives-not-numbers\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>Size markets using narratives, not\u00a0numbers<\/em><\/a><\/p>\n<p><a href=\"http:\/\/thisweekinstartups.com\/chris-dixon-general-partner-at-andreessen-horowitz-twist-e336\/\" target=\"_blank\" rel=\"noopener\">This Week in Startups \u2013 Chris Dixon, General Partner at a16z (video)<\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2010\/04\/03\/size-markets-using-narratives-not-numbers\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>Size markets using narratives, not\u00a0numbers<\/em><\/a><\/p>\n<p><a href=\"http:\/\/www.logomaker.com\/blog\/2013\/12\/16\/9-inspirational-quotes-startups-small-business-chris-dixon\/\" target=\"_blank\" rel=\"noopener\">9 inspirational startup quotes from Chris Dixon<\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2009\/11\/14\/pitch-yourself-not-your-idea\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>Pitch yourself, not your idea<\/em><\/a><\/p>\n<p><a href=\"https:\/\/gigaom.com\/2010\/05\/28\/chris-dixon\/\" target=\"_blank\" rel=\"noopener\">Gigaom \u2013\u00a0Why the VC Model is Broken<\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2009\/09\/19\/climbing-the-wrong-hill\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>Climbing the Wrong Hill<\/em><\/a><\/p>\n<p><a href=\"http:\/\/cdixon.org\/2010\/09\/12\/if-you-arent-getting-rejected-on-%20%20a-daily-basis-your-goals-arent-ambitious-enough\/\" target=\"_blank\" rel=\"noopener\">Chris Dixon \u2013 <em>If you are not getting rejected<\/em><\/a><\/p>\n\t\t\t<\/div>\n\n\t\t\t<div class=\"entry-content__post-date\">\n\t\t\t\tJanuary 18, 2015\t\t\t<\/div>\n\n\t\t\t\n<div class=\"entry-content__related-posts related-posts related-posts--entry-content\">\n\t<div class=\"related-stories\">\n\t\t<ul class=\"related-stories__list post-list--related\">\n\t\t      \t\n\n\t\n\n\t\n\t\t\t\n\t\n\t\t\n\t\t\t \t\t\n<\/ul><\/div>\n\n<ul class=\"post-terms post-terms--post-list\"><li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/apparently-all-our-listicles-are-16\/\">\n\t\tapparently all our listicles are 16 (or multiples of 16!)\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/end-of-year-roundups\/\">\n\t\tend of year roundups\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/holiday-specials\/\">\n\t\tholiday specials\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/listicles\/\">\n\t\tlisticles\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/what-were-reading\/\">\n\t\twhat we're reading\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/articles\/\">\n\t\tarticles\t\t<\/a>\n\t<\/li>\n\n<\/ul><article class=\"post-list__item post-list__item--post\"><h3 class=\"post-list__item-title\">\n\t\t<a class=\"post-list__item-title-link\" href=\"https:\/\/a16z.com\/2017\/02\/18\/12-things-about-product-market-fit\/\">12 Things about Product-Market Fit<\/a>\t<\/h3>\n\n\t\n<div class=\"byline byline--post-list\">\n\t\n\t\tby\t\n\t\n\t\t\n\t\t\t<a href=\"https:\/\/a16z.com\/author\/tren-griffin\/\" target=\"\" class=\"author\">Tren Griffin<\/a>\t\t\n<\/div>\n\n<ul class=\"post-terms post-terms--post-list\"><li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/advice-and-how-to\/growth\/\">\n\t\tgrowth\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/advice-and-how-to\/how-innovation-happens\/\">\n\t\thow innovation happens\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/company-building-101\/\">\n\t\tCompany Building 101\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/listicles\/\">\n\t\tlisticles\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/of-moats-and-margins\/\">\n\t\tof moats and margins\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/product-market-fit\/\">\n\t\tproduct-market fit\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/articles\/\">\n\t\tarticles\t\t<\/a>\n\t<\/li>\n\n<\/ul><\/article><article class=\"post-list__item post-list__item--post\"><h3 class=\"post-list__item-title\">\n\t\t<a class=\"post-list__item-title-link\" href=\"https:\/\/a16z.com\/2018\/12\/27\/endless-chasm-enterprise-startups\/\">The Unending Chasm, and How to Survive It<\/a>\t<\/h3>\n\n\t\n<div class=\"byline byline--post-list\">\n\t\n\t\tby\t\n\t\n\t\t\n\t\t\t<a href=\"https:\/\/a16z.com\/author\/martin-casado\/\" target=\"\" class=\"author\">Martin Casado<\/a>\t\t\n<\/div>\n\n<ul class=\"post-terms post-terms--post-list\"><li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/advice-and-how-to\/go-to-market\/\">\n\t\tgo to market\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/enterprise-startups\/\">\n\t\tenterprise startups\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/pre-chasm\/\">\n\t\tpre-chasm\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/sales\/\">\n\t\tsales\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/articles\/\">\n\t\tarticles\t\t<\/a>\n\t<\/li>\n\n<\/ul><\/article><\/div>\n<\/div>\n\n\t\t\n\n\t\t\n<aside class=\"entry-sidebar sidebar sidebar--single\" role=\"complementary\"><div class=\"sidebar__widgets widget-area\">\n\n\t\t<section class=\"widget widget--post-terms\"><ul class=\"post-terms post-terms--single\"><li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/advice-and-how-to\/career\/\">\n\t\tcareer &amp; self development\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/tag\/listicles\/\">\n\t\tlisticles\t\t<\/a>\n\t<\/li>\n\n\n\t<li>\n\t\t<a rel=\"tag\" href=\"https:\/\/a16z.com\/articles\/\">\n\t\tarticles\t\t<\/a>\n\t<\/li>\n\n<\/ul><\/section><section class=\"widget widget--post-social\"><ul class=\"social-sharing\"><li class=\"social-sharing__item\">\n\t\t<a class=\"social-sharing__link\" href=\"https:\/\/www.facebook.com\/sharer\/sharer.php?u=https:\/\/a16z.com\/2015\/01\/18\/12-things-learned-from-chris-dixon-about-startups\/\" target=\"_blank\">\n\t\t\tFacebook\n\t\t<\/a>\n\t<\/li>\n\t<li class=\"social-sharing__item\">\n\t\t<a class=\"social-sharing__link\" href=\"https:\/\/www.linkedin.com\/shareArticle?mini=true&amp;url=https:\/\/a16z.com\/2015\/01\/18\/12-things-learned-from-chris-dixon-about-startups\/\" target=\"_blank\">\n\t\t\tLinkedIn\n\t\t<\/a>\n\t<\/li>\n\t<li class=\"social-sharing__item\">\n\t\t<a class=\"social-sharing__link\" href=\"https:\/\/twitter.com\/intent\/tweet\/?text=12+Things+I+Learned+from+Chris+Dixon+about+Startups&amp;url=https:\/\/a16z.com\/2015\/01\/18\/12-things-learned-from-chris-dixon-about-startups\/&amp;via=a16z\">\n\t\t\tTwitter\n\t\t<\/a>\n\t<\/li>\n<\/ul><\/section><\/div>\n\n<\/aside>","created_at":"2019-01-13T19:19:56+0000","updated_at":"2019-01-13T19:19:56+0000","published_at":"2015-01-18T22:34:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":21,"domain_name":"a16z.com","preview_picture":"https:\/\/s2.wp.com\/wp-content\/themes\/vip\/a16z-2015\/static\/images\/a16z-logo.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":773,"title":"how-to-choose-a-startup-to-work-for","url":"https:\/\/triplebyte.com\/blog\/how-to-choose-a-startup-to-work-for","created_at":"2019-01-13T14:29:16+0000","updated_at":"2019-01-13T14:29:16+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"triplebyte.com"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":772,"title":"Tilt Up","url":"https:\/\/tilt.build\/","content":"<div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-dce9541ea932d1b88cb5\"><div class=\"sqs-block-content\"><h3>Tilt makes it possible to develop all your microservices locally in Kubernetes while collaborating with your team.<\/h3><p>Write a Tiltfile script that describes how your services fit together. Share it with your team so that any engineer can hack on any server. See a complete view of your system, from building to deploying to logging to crashing.<\/p><\/div><\/div><div class=\"sqs-block button-block sqs-block-button\" data-block-type=\"53\" id=\"block-a5a3b8102e7da07fd146\"><div class=\"sqs-block-content\"><p><a href=\"https:\/\/github.com\/windmilleng\/tilt\" class=\"sqs-block-button-element--medium sqs-block-button-element\">Get Tilt<\/a><\/p><\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-2 span-2\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-aa1d1c108da1264c2e24\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0efed76d2a73aee7ba576d\/1544486366272\/local.png\" alt=\"local.png\" \/><\/noscript><img class=\"thumb-image\" data-image=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0efed76d2a73aee7ba576d\/1544486366272\/local.png\" data-image-dimensions=\"800x800\" data-image-focal-point=\"0.5,0.5\" alt=\"local.png\" data-load=\"false\" data-image-id=\"5c0efed76d2a73aee7ba576d\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0efed76d2a73aee7ba576d\/1544486366272\/local.png\" \/><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"col sqs-col-10 span-10\"><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-a711b2e836e65b45b310\"><div class=\"sqs-block-content\"><h2>Keep your local development tools<\/h2><p>Edit in your IDE, save to your filesystem. No commit or push required. Tilt watches the filesystem and updates your servers in seconds. Tilt can update your servers on Docker For Mac, Minikube, or your favorite Kubernetes cluster.<\/p><\/div><\/div><\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-2 span-2\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-24716b0781658ce419ab\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0f286389858336ab286c16\/1544497256628\/collab.png\" alt=\"collab.png\" \/><\/noscript><img class=\"thumb-image\" data-image=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0f286389858336ab286c16\/1544497256628\/collab.png\" data-image-dimensions=\"800x800\" data-image-focal-point=\"0.5,0.5\" alt=\"collab.png\" data-load=\"false\" data-image-id=\"5c0f286389858336ab286c16\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0f286389858336ab286c16\/1544497256628\/collab.png\" \/><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"col sqs-col-10 span-10\"><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-a7f23d73723691dd674d\"><div class=\"sqs-block-content\"><h2>Find your way in a maze of microservices<\/h2><p>Stop playing 20 questions with kubectl each time your app misbehaves. Tilt collects problems from across tools and services into one UI. One place to see build breakages, yaml typos, crash loops, and request exceptions.<\/p><\/div><\/div><\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-2 span-2\"><div class=\"sqs-block image-block sqs-block-image\" data-aspect-ratio=\"100\" data-block-type=\"5\" id=\"block-c64b2e414099652e2c07\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0efed76d2a73aee7ba5771\/1544486376529\/insight.png\" alt=\"insight.png\" \/><\/noscript><img class=\"thumb-image\" data-image=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0efed76d2a73aee7ba5771\/1544486376529\/insight.png\" data-image-dimensions=\"800x800\" data-image-focal-point=\"0.5,0.5\" alt=\"insight.png\" data-load=\"false\" data-image-id=\"5c0efed76d2a73aee7ba5771\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c0efed76d2a73aee7ba5771\/1544486376529\/insight.png\" \/><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"col sqs-col-10 span-10\"><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-ffb82851b73f0257d1fe\"><div class=\"sqs-block-content\"><h2>Share the workflow that keeps you in-flow<\/h2><p>Any engineer just runs \u2018tilt up\u2019 to see the app live at http:\/\/localhost. No more \u201cit worked on my machine\u201d \u2013 everything runs in containers so the right dependencies are always there. Tilt updates with container optimization tricks &amp; best practices, so that even complex projects update in seconds.<\/p><\/div><\/div><\/div><\/div><div class=\"sqs-block button-block sqs-block-button\" data-block-type=\"53\" id=\"block-924347fd547ed28a8a64\"><div class=\"sqs-block-content\"><p><a href=\"https:\/\/github.com\/windmilleng\/tilt\" class=\"sqs-block-button-element--medium sqs-block-button-element\">Try Tilt Now<\/a><\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1544487663311_58147\"><div class=\"sqs-block-content\"><p>Already have a Dockerfile and a Kubernetes config? You\u2019ll be able to setup Tilt quickly and start getting things done. <a href=\"https:\/\/docs.tilt.build\">Let\u2019s Go!<\/a><\/p><p>Having trouble developing your servers in Kubernetes? We\u2019d love to chat.<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1544487663311_22046\"><div class=\"sqs-block-content\"><p>Built with \ud83d\udc96 at Windmill Engineering in NYC<\/p><\/div><\/div>","created_at":"2019-01-13T13:57:18+0000","updated_at":"2019-01-13T13:57:18+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"tilt.build","preview_picture":"http:\/\/static1.squarespace.com\/static\/58d96727be659476d0a45a13\/t\/5c1030c703ce642769552fa8\/1544564936514\/Tilt-Beta.png?format=1500w","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":771,"title":"Re-decentralizing the Web, for good this time","url":"https:\/\/ruben.verborgh.org\/articles\/redecentralizing-the-web\/","content":"<p id=\"summary\">Originally designed as a decentralized network, the Web has undergone a significant centralization in recent years. In order to regain freedom and control over the digital aspects of our lives, we should understand how we arrived at this point and how we can get back on track. This chapter explains the history of decentralization in a Web context, and details Tim Berners-Lee\u2019s role in the continued battle for a free and open Web. The challenges and solutions are not purely technical in nature, but rather fit into a larger socio-economic puzzle, to which all of us are invited to contribute. Let us take back the Web for good, and leverage its full potential as envisioned by its creator.<\/p><h2 id=\"power-to-the-people\">Power to the people<\/h2><p id=\"power-to-the-people-p-1\">As an inventor, you might envision a purpose and destiny for your creation\u2014<wbr \/>yet ultimately, people decide how they put it to use. John Pemberton aimed to cure morphine addicts when he started brewing the potion now known as <em>Coca-Cola<\/em>, Noah McVicker\u2019s <em>Play-Doh<\/em> served as a wall-cleaner before it became a children\u2019s toy, and Alfred Nobel declared yearly prizes so he would not be remembered for dynamite\u2019s military purposes. Admirably, Tim Berners-Lee never even intended to control his own invention: his former employer <abbr class=\"caps\">CERN<\/abbr> released the World Wide Web software openly, and the Web itself is designed in a <em>decentralized<\/em> way so that no one can decide who can say what. This unprecedented openness has led to large-scale <em>permissionless innovation<\/em> and unbounded creativity, provides a voice to more than half of the world\u2019s population, and has revolutionized communication, education, and business. However, a consequence of this freedom is also that anyone can create things that go against the spirit of the Web, such as illegal materials and\u2014<wbr \/>ironically\u2014<wbr \/>platforms whose primary goal is <em>centralization<\/em>.<\/p><p id=\"power-to-the-people-p-2\">The concept of centralization does not pose a problem in and of itself: there are good reasons for bringing people and things together. The situation becomes problematic when we are robbed of our choice, deceived into thinking there is only one access gate to a space that, in reality, we collectively own. Some time ago, it seemed unimaginable that a fundamentally open platform like the Web would become the foundation for closed spaces, where we pay with our personal data for a fraction of the freedoms that are actually already ours. Yet a majority of Web users today find themselves confined to the boundaries of a handful of influential social networks for their daily interactions. Such networks gather opinions from all over the world, only to condense that richness into one space, where they simultaneously act as the director and judge of the resulting stream they present to us.<\/p><p id=\"power-to-the-people-p-3\">Because this change happened so suddenly, perhaps we need a reminder that the Web landscape looked quite different not even that long ago. In 2008, Iranian blogger Hossein Derakhshan was sentenced to 20 years of jail, primarily because of blog posts he had written. He and many others were able to state their critical opinions because they had the Web as an open platform, so they did not depend on anyone\u2019s permission to publish their words. Crucially, the Web\u2019s hyperlinking mechanism lets blogs point to each other, again without requiring any form of permission. This allows for a decentralized value network between equals, where readers remain in active and conscious control of their next move. When Derakhshan was eventually released in 2014, he <a href=\"https:\/\/medium.com\/matter\/the-web-we-have-to-save-2eb1fe15a426\" target=\"_blank\">came back to an entirely different Web<\/a> <a href=\"#ref-1\" class=\"reference\">[1]<\/a>: critical readers had transformed into passive viewers, as if watching television. While Web technology had of course evolved, its core foundations had not\u2014<wbr \/>it was the way people were using the Web that had become unrecognizable in a mere 6 years.<\/p><p id=\"power-to-the-people-p-4\">Of course, social media are not our enemies here: they should be credited with lowering the barrier for the online publication of short texts and photos by anyone. Yet they operate under a <em>winner-takes-all<\/em> strategy, each striving to become the dominant portal instead of mutually interoperating like the rest of the Web. In contrast to blogs, we typically cannot interact with posts in one network from within another: we would need to either move the people or the data. This famous <a href=\"https:\/\/www.economist.com\/node\/10880516\" target=\"_blank\"><q>walled gardens<\/q> problem of social media<\/a> <a href=\"#ref-2\" class=\"reference\">[2]<\/a> has significantly worsened since 2008, because some gardens have grown huge while their walls remain in place. A major problem is that access to the dominant networks invariably means giving up control over our personal data: we can enter through the door in the wall if we pay with our digital belongings. That personal data can then be leveraged to unwittingly influence us through excessively personalized advertising for brands, products, and even political agendas. Furthermore, once there, people tend to form small conversational circles within each garden\u2014<wbr \/>an effect that is further amplified by the inward focus of social media platforms and their algorithms that favor maximizing engagement over diversity. The resulting <em>filter bubble<\/em> <a href=\"#ref-3\" class=\"reference\">[3]<\/a> isolates us into our own echo chambers, whereas the Web\u2019s purpose\u2014<wbr \/>and social media\u2019s claim\u2014<wbr \/>has always been to <em>connect<\/em>.<\/p><p id=\"power-to-the-people-p-5\">Unsurprisingly, these problems are reflected in <a href=\"https:\/\/webfoundation.org\/2017\/03\/web-turns-28-letter\/\" target=\"_blank\">three challenges for the Web<\/a> <a href=\"#ref-4\" class=\"reference\">[4]<\/a> that Tim Berners-Lee put forward in 2017:<\/p><ul><li>taking back control of our personal data;<\/li> <li>preventing the spread of misinformation;<\/li> <li>realizing transparency for political advertising.<\/li> <\/ul><p id=\"power-to-the-people-p-6\">Clearly, it is undesirable to tackle these challenges through centralized solutions, for instance by appointing an authority for personal data, news, and advertising. This would create yet another single point of failure, which\u2014<wbr \/>even assuming the best of intentions\u2014<wbr \/>would always be more vulnerable to abuse. The core issue in this situation is ultimately not one with the individual social networks, but with their hyper-centralization of data and people, and therefore power. We want control, but we want to put that control in the hands of every person, as a right they can choose to exercise over the data they create.<\/p><p id=\"power-to-the-people-p-7\">From the above, it is clear that <a href=\"https:\/\/blog.dshr.org\/2018\/01\/it-isnt-about-technology.html\" target=\"_blank\">our primary obstacles are not technological<\/a> <a href=\"#ref-5\" class=\"reference\">[5]<\/a>; hence <a href=\"https:\/\/webfoundation.org\/2018\/03\/web-birthday-29\/\" target=\"_blank\">Tim Berners-Lee\u2019s call<\/a> <a href=\"#ref-6\" class=\"reference\">[6]<\/a> to <q>assemble the brightest minds from business, technology, government, civil society, the arts, and academia to tackle the threats to the Web\u2019s future<\/q>. Yet at the same time, computer scientists and engineers need to deliver the <em>technological burden of proof<\/em> that decentralized personal data networks can scale globally and that they can provide people with an experience similar to that of centralized platforms.<\/p><p id=\"power-to-the-people-p-8\">In this chapter, we will therefore start with a technological perspective on decentralization, highlighting Tim Berners-Lee\u2019s role in the continuing fight to keep the Web open and decentralized. After a historical overview of power struggles on the Web, we will zoom in on the changes that decentralization requires, and examine what a more healthy ecosystem would look like. As a concrete implementation of these principles, we will study the Solid project. We will end with a discussion of open challenges and an outlook on the future.<\/p><h2 id=\"ashort-history-of-de-centralization-and-the-web\">A short history of (de-)centralization and the Web<\/h2><p id=\"ashort-history-of-de-centralization-and-the-web-p-1\">The arrows of the decentralization movement have not always been aimed at social media\u2014<wbr \/>and they likely will not be anymore at some point in the future. The forces causing centralization have instead been a moving target: every time a threat had been addressed, an even bigger one superseded it. Understanding these threats will provide us with insights into the different facets of decentralization and their importance.<\/p><h3 id=\"decentralization-as-the-unspoken-assumption\">Decentralization as the unspoken assumption<\/h3><p id=\"decentralization-as-the-unspoken-assumption-p-1\">Decentralized systems, which do not require a central mediator to function, were already around at the time the Web was invented. Most notably, the Internet was increasingly gaining popularity as a large-scale decentralized network. Email was even more decentralized than the traditional postal mail service it mimicked, since different mail servers would directly exchange messages with each other. Long forgotten protocols such as the Network News Transfer Protocol (<abbr class=\"caps\">NNTP<\/abbr>) allowed for the decentralized exchange of news articles. In short, decentralization was not some crazy new idea, but rather the spirit of the time.<\/p><p id=\"decentralization-as-the-unspoken-assumption-p-2\">Therefore, when Tim Berners-Lee set out to design a new hypertext system in 1989, it was presumed to be decentralized, in contrast to documentation systems of the time, but in alignment with many others. The main selling point of the Web was its <a href=\"https:\/\/www.w3.org\/2005\/Talks\/0323-yorkshire-tbl\/slide5-2.html\" target=\"_blank\"><em>universality<\/em><\/a> <a href=\"#ref-7\" class=\"reference\">[7]<\/a>, its independence of, among others, hardware and software; decentralization was simply the unspoken assumption. This is reflected in the original article introducing the Web <a href=\"#ref-8\" class=\"reference\">[8]<\/a>, which emphasizes universal readability across operating systems, but does not mention the term <q>decentralization<\/q> at all.<\/p><p id=\"decentralization-as-the-unspoken-assumption-p-3\">The only component with centralized roots in the Web\u2019s architectural design is the Domain Name System (<abbr class=\"caps\">DNS<\/abbr>), which resolves the domain name part of a Web address (such as <em>example.org<\/em>) to a physical machine on the Internet. This was not as much of an issue back in the days when the number of domains was relatively small and domain ownership would be stationary. Nowadays, millions of domain names frequently change hands, thereby breaking existing links in possibly malicious ways. By manipulating <abbr class=\"caps\">DNS<\/abbr>, governments can block or alter access to existing websites. Tim Berners-Lee has indicated that, in hindsight, a more decentralized naming system might have been preferred. Apart from that, the Web contained all ingredients to thrive in a decentralized way.<\/p><h3 id=\"the-race-for-our-desktop\">The race for our desktop<\/h3><p id=\"the-race-for-our-desktop-p-1\">A first wave of centralization resulted as collateral damage from the <em>browser war<\/em> of the nineties, in which companies competed to become the sole vendor of the software through which we access the Web. The Web\u2019s design principle of universality demanded readability on any platform, so the emergence of multiple browsers was a blessing\u2014<wbr \/>except that they strived for market domination rather than mutually beneficial co-existence. The Netscape browser and Microsoft\u2019s Internet Explorer tried to convert each other\u2019s users through new features, with Internet Explorer reaching over 90% of connected desktops at its peek.<\/p><p id=\"the-race-for-our-desktop-p-2\">While competition through innovation is fine, these features came at the cost of incompatibility across browsers and therefore directly endangered the Web\u2019s universality. Websites would carry badges such as <q>best viewed in Internet Explorer<\/q>, since a consistent experience across platforms could not be guaranteed. Those who did not want to use a particular browser\u2014<wbr \/>or who could not install it because no compatible version for their system existed\u2014<wbr \/>would be unable to access such websites fully or at all. The resulting de-facto browser monopoly infringed on people\u2019s preference for browser or operating system, centralizing the Web\u2019s decision process in one company that thereby determined the rate of innovation.<\/p><p id=\"the-race-for-our-desktop-p-3\">The World Wide Web Consortium (<abbr class=\"caps\">W3C<\/abbr>) was founded by Tim Berners-Lee with a mission of compatibility, enabling cross-browser consistency through recommendations that specify the correct workings of Web technologies. While <abbr class=\"caps\">W3C<\/abbr> standardization is administratively centralized, it incorporates feedback from a decentralized network of members through a consensus-driven process. A problem by the early 2000s was that Internet Explorer deviated from <abbr class=\"caps\">W3C<\/abbr> recommendations at crucial points, forcing developers to follow either the actual standards or their incorrect implementation in the most popular browser.<\/p><p id=\"the-race-for-our-desktop-p-4\">Fortunately, pressure from Firefox and Safari during a second browser war eventually <a href=\"https:\/\/alistapart.com\/article\/beyonddoctype\" target=\"_blank\">forced Microsoft onto a more standards-oriented course<\/a> <a href=\"#ref-9\" class=\"reference\">[9]<\/a>. Since 2010, no single browser has gained more than two thirds of global market share anymore, meaning that standards compatibility is now in the interest of browser vendors and Web developers alike. The balkanization of the Web through centralized browser development has thereby largely been averted.<\/p><h3 id=\"the-race-for-our-personal-data-and-identity\">The race for our personal data and identity<\/h3><p id=\"the-race-for-our-personal-data-and-identity-p-1\">While the hegemony of Google still continues, social media have found an even more powerful way of collecting and marketing our personal data. The social Web revolution of the 2000s encouraged people to be present online, which drove many of us to various platforms to share blog posts, bookmarks, photos, videos, and more. Some year later, social media companies created centralized platforms to take over many of these features, which until then were spread out across multiple providers. These platforms store our personal data and request far-reaching usage rights in exchange for their services, all of which operate within their own walled garden.<\/p><p id=\"the-race-for-our-personal-data-and-identity-p-2\">Like search engines, the main service of social networks consists of a linear list of content, ranked by factors and algorithms we can only minimally influence. In contrast to search, a social feed is generated without any input terms from our side, like a television that no longer requires a remote. The ensuing show is meticulously personalized based on data we consciously left on social network platforms, combined with traces from our browsing history picked up\u2014<a href=\"https:\/\/berjon.com\/advertising-war-on-consent\/\" target=\"_blank\">without our concious consent<\/a> <a href=\"#ref-10\" class=\"reference\">[10]<\/a>\u2014by social trackers on third-party websites. In his 2018 Dertouzos distinguished lecture, Tim Berners-Lee mentioned that <a href=\"https:\/\/www.csail.mit.edu\/news\/utopia-dystopia-29-short-years\" target=\"_blank\">political advertising has been banned from television in the <abbr class=\"caps\">UK<\/abbr><\/a> <a href=\"#ref-11\" class=\"reference\">[11]<\/a> because of concerns about the impact of such a direct medium. Yet by that logic, he continued, we should be much more concerned about the heavily personalized political advertising that current social media platforms enable and allow. Even if we refrain from explicitly sharing certain sensitive traits, seemingly insignificant pieces of other data can be combined into reliable predictors of highly personal information <a href=\"#ref-12\" class=\"reference\">[12]<\/a> such as sexual orientation, ethnicity, and religious or political views, which are subsequently used to target us.<\/p><p id=\"the-race-for-our-personal-data-and-identity-p-3\">As in the previous two centralization races, a subtle force is exerted upon us: we feel pressured to be part of the large networks, because not joining means missing out on the volatile virtual traces of our friends\u2019 and family members\u2019 lives. Often the easiest way for grandparents to see their grandchildren\u2019s latest pictures is to create a Facebook or Instagram account. This is how the digital memory of a large part of today\u2019s generation ends up in one space, often beyond control of those that are part of the memories. The centralization of our online activities has become so extreme that <a href=\"http:\/\/lirneasia.net\/2014\/08\/more-facebook-users-than-internet-users-in-south-east-asia\/\" target=\"_blank\">some Facebook users have become unaware of their ability to access the Internet<\/a> <a href=\"#ref-13\" class=\"reference\">[13]<\/a>. This paradox has sadly become a reality in many countries, where Facebook\u2019s <em>Internet.org<\/em> initiative provides a severely constrained version of the Web that further reduces people\u2019s options, in blatant violation of Net Neutrality.<\/p><p id=\"the-race-for-our-personal-data-and-identity-p-4\">Meanwhile, another race is happening in the background, namely the battle to become our <em>identity provider<\/em>. An increasing number of websites are gradually replacing their own login systems with authentication tied to large platforms such as Google or Facebook. For people with an existing account, the <q>Log in with Facebook<\/q> buttons are a convenience. For those without, they form additional pressure to join. And in both cases, such buttons are yet another way of tracking our online activities. This centralization of identity takes away our freedom to assume the persona we want\u2014<wbr \/>be it anonymous, pseudonymous, our just ourselves\u2014<wbr \/>without needing to expose data we consider our own.<\/p><h2 id=\"data-ownership-by-decoupling-storage-from-service\">Data ownership by decoupling storage from service<\/h2><p id=\"data-ownership-by-decoupling-storage-from-service-p-1\">A recurring theme in the above centralization races is the lack of choice: a choice of browser and operating system, of entry point to the Web, of storage for our personal data. Decentralization is fundamentally about <em>enabling choice<\/em>, by breaking up artificially coupled decisions into individual options that can be combined at will. Just as we are free to choose any combination of device, operating system, and browser to access the Web, we should be able to interact with websites and other people without commitment to a single social or other platform.<\/p><p id=\"data-ownership-by-decoupling-storage-from-service-p-2\">Taking back control of our personal data, as envisioned by Tim Berners-Lee, is realized by <em>decoupling data storage from services<\/em>. This means people can store their data wherever they choose, while still enjoying the services they want. We can pick any provider to store our texts, photos, and videos\u2014<wbr \/>or simply store them on our own Web server\u2014<wbr \/>and rely on any third-party service to interact with them, regardless of storage location. The crucial service of <em>identity<\/em> can, but does not need to, be provided by the data storage.<\/p><p id=\"data-ownership-by-decoupling-storage-from-service-p-3\">This mindset gives rise to the concept of a <em>personal data pod<\/em>, in which we can store every single piece of information we produce. As shown in <a href=\"#social-media-post\">the figure below<\/a> this statement can be taken quite literally: even a seemingly trivial piece of data, such as simple <q>like<\/q> we gave a certain webpage or thing, can be stored in our own pod. While such a degree of decentralization might seem extreme, recall that even supposedly trivial likes can reveal much deeper personal information <a href=\"#ref-12\" class=\"reference\">[12]<\/a>, so it makes sense to give people control over them. Furthermore, since we do not depend on anyone\u2019s permission to publish data in our own pod, we can place likes, annotations, and comments on anything we want, without fear of them being censored or deleted.<\/p><p id=\"data-ownership-by-decoupling-storage-from-service-p-4\">This total data ownership enables highly granular <em>access control<\/em>: people can selectively give permission to friends or applications to read or write specific parts of their data pod. For instance, they can decide whether or not they make their profile picture and full name public, who can see which of their likes and comments, and what applications can edit their pictures or posts on their behalf. These permissions can be changed or revoked at any time. People can have multiple data pods for different purposes, for instance, a pod for personal and family pictures at home, a pod governed by retention policies for professional data at the workplace, and a university pod with study materials and grades. Upon creation, they can decide which data is stored in which pod.<\/p><p id=\"data-ownership-by-decoupling-storage-from-service-p-5\">By choosing the storage location of our own data, we prevent unauthorized access and exploitation. We are no longer obliged to pay with our data in order to access a certain service. Moreover, we can protect the most sensitive parts of our data by keeping them to ourselves, and limit sharing to people and services that really require it\u2014<wbr \/>but only for as long as they need it.<\/p><h2 id=\"independent-innovation-on-separate-data-and-service-markets\">Independent innovation on separate data and service markets<\/h2><p id=\"independent-innovation-on-separate-data-and-service-markets-p-1\">When people store their own data, privacy-unfriendly business models centered around data ownership will not be viable anymore. Such an economic change can be accelerated through legislation, like the <abbr class=\"caps\">EU<\/abbr>\u2019s General Data Protection Regulation (<abbr class=\"caps\">GDPR<\/abbr>), as well as growing awareness among the general population about the dangers of centralization, given recent data scandals at companies such as Equifax and Facebook. Consequently, new business models for applications become necessary.<\/p><p id=\"independent-innovation-on-separate-data-and-service-markets-p-2\">Decentralization requires the nature of applications to evolve from silos to shared views. As shown in <a href=\"#apps-as-views\">the figure below<\/a>, current Web apps combine <em>data<\/em> and <em>service<\/em>. Because of this coupling, our LinkedIn contacts cannot comment on our Facebook pictures, and an <abbr class=\"caps\">RSVP<\/abbr> on a Facebook event will not be reflected in our Doodle calendar\u2019s availability. Decentralized applications, on the other hand, act as <em>views<\/em> on top of our data pod and those of others. When granted specific access rights, photos uploaded into our data pod by a photo gallery application can be accessed by a social feed app. Events in my personal calendar that have public visibility can show up in the same feed. Our friends can view the parts of our data to which we grant them access through whatever application they wish to use.<\/p><figure id=\"apps-as-views\"> <img src=\"https:\/\/ruben.verborgh.org\/images\/blog\/apps-as-views.svg\" alt=\"\" class=\"white\" \/><figcaption> Centralized Web applications act as silos that do not share data with each other. Decentralized Web applications act as shared views on top of personal data pods. <\/figcaption><\/figure><p id=\"independent-innovation-on-separate-data-and-service-markets-p-3\">Because the choice of data and service provider becomes decoupled, <em>separate markets<\/em> for data and services emerge. <a href=\"#separate-markets\">The figure below<\/a> shows that centralized applications compete in a single market based on data ownership, because usage of a service is coupled with usage of its storage. As such, people cannot easily switch to a better application experience, as migrating their data\u2014<wbr \/>if possible\u2014<wbr \/>is technically challenging. Furthermore, new applications that could offer a better experience have trouble joining the market, since they do not own sufficient data yet. With decentralized Web applications, people select their storage and service providers separately, which allows an independent competition on the level of storage and on the level of services. On both levels, the competition is solely based on service quality and features versus cost.<\/p><p id=\"independent-innovation-on-separate-data-and-service-markets-p-4\">This independence means we can freely switch data and service providers, without requiring our friends to choose the same ones. This brings down the walls in between the gardens, because we gain the ability to reuse and move our data, and can interact with anyone in the entire landscape. Data and service providers can evolve without dependency on each other, which enables a faster and more creative innovation cycle. Anyone can enter either market and attract customers by providing a better experience than others, without asking for control of our data.<\/p><figure id=\"separate-markets\"> <img src=\"https:\/\/ruben.verborgh.org\/images\/blog\/data-and-app-markets.svg\" alt=\"\" class=\"white\" \/><figcaption> Centralized applications compete in a single market, based on data ownership. On a decentralized Web, data and service providers compete in different markets. <\/figcaption><\/figure><h2 id=\"the-solid-project\">The Solid project<\/h2><p id=\"the-solid-project-p-1\">In order to realize this vision of data ownership and data\/<wbr \/>service independence, Tim Berners-Lee started the <a href=\"https:\/\/solid.mit.edu\/\" target=\"_blank\">Solid project<\/a> <a href=\"#ref-14\" class=\"reference\">[14]<\/a>. Solid consists of <em>specifications<\/em> for interoperability, <em>implementations<\/em> of servers, clients, and applications, and a <em>community<\/em> of people who build new things. In the next sections, we will discuss some of Solid\u2019s unique aspects.<\/p><h3 id=\"personal-data-linking-and-integration\">Personal data linking and integration<\/h3><p id=\"personal-data-linking-and-integration-p-1\">The goal of Solid is empowering people through <em>personal data management<\/em>, as a counterpart to enterprise data management. We can consider a Solid server or data pod as the equivalent of a hard disk for the Web, on which we can store arbitrary documents. Then Solid apps are like our desktop applications, except that they open documents from Solid servers on the Web. In contrast to actual hard disks, Solid servers are typically public to the entire world, so detailed access control settings allow us to specify who can view or edit which of our documents. Tim Berners-Lee has been leading by example, by managing his personal and professional life with Solid for several years already.<\/p><p id=\"personal-data-linking-and-integration-p-2\">In order for such data management to work at Web scale, data in different pods need to link to each other, similar to how hypertext documents allow us to jump from one website to another. Solid uses <a href=\"https:\/\/www.w3.org\/DesignIssues\/LinkedData.html\" target=\"_blank\"><em>Linked Data<\/em><\/a> <a href=\"#ref-15\" class=\"reference\">[15]<\/a> to achieve this: every piece of data can link to any other. This is how, for example, a comment in your data pod can be attached to a photo in someone else\u2019s pod, while both of you can remain owners of your data. At runtime, Solid applications integrate data from multiple sources and blend them together into a single experience.<\/p><p id=\"personal-data-linking-and-integration-p-3\">Solid pods can offer people a decentralized means of identification. People can pick a so-called <em>WebID<\/em>, which is a Web address that identifies them. This Web address leads to their public profile, and people can log on to any pod with their own WebID, instead of requiring a new login on every website or resorting to a centralized identity provider.<\/p><h3 id=\"readwrite-web\">Read\u2013Write Web<\/h3><p id=\"readwrite-web-p-1\">One of the crucial aspects of Solid is that it provides a Read\u2013Write platform, as was Tim Berners-Lee\u2019s original intention for the Web <a href=\"#ref-16\" class=\"reference\">[16]<\/a>. While writing has always been possible, in the sense that anyone could start their own website, the Web 2.0 and social media revolutions should be credited with making writing considerably easier. This explains part of the success of these platforms, as anyone can now be a content producer at any time, especially through their mobile devices.<\/p><p id=\"readwrite-web-p-2\">Solid should make authoring content similarly easy, the difference being of course that we would always write to our own data pods instead of to the application through which we create. In doing so, we guarantee that everyone can express themselves without risking censorship. To maximize interoperability, our Linked Data should be stored using Semantic Web technologies <a href=\"#ref-17\" class=\"reference\">[17]<\/a>, which interweave a piece of data with its meaning. That way, application can make sense of (parts of) each other\u2019s data, without having to agree upfront exactly what our data should look like. When storing data in our own pods, we need a mechanism to inform others when things have been created or modified\u2014<wbr \/>especially if these are comments on their data. This is enabled through <a href=\"https:\/\/www.w3.org\/TR\/ldn\/\" target=\"_blank\"><em>Linked Data Notifications<\/em><\/a> <a href=\"#ref-18\" class=\"reference\">[18]<\/a>, small automated messages similar to email, which different data pods can send to each other. By combining these technologies, Solid aims to realize the Read\u2013Write Linked Data vision <a href=\"#ref-16\" class=\"reference\">[16]<\/a>, in order to ensure that everyone can participate in the Web of Data.<\/p><h3 id=\"potential-for-disruption\">Potential for disruption<\/h3><p id=\"potential-for-disruption-p-1\">By transforming data ownership and the role of applications in a decentralized ecosystem, Solid is able to disrupt many interactions that happen on the Web. Many processes that currently depend on centralization can be revolutionized in a decentralized way, by cutting out the middlemen that control these processes. This can stimulate innovation in areas that are embracing the current status quo and resisting change.<\/p><p id=\"potential-for-disruption-p-2\">A first obvious target are social interactions between people. Sharing multimedia with friends, colleagues, and family members without privacy concerns becomes possible through Solid. Other examples include collaborating on various kinds of documents under transparent access control, and organizing meetings and events\u2014<wbr \/>again with full data ownership, choice of application and storage, and synchronization between different apps.<\/p><p id=\"potential-for-disruption-p-3\">Moreover, Solid has the technological potential to disrupt entire industries, such as for instance <em>scholarly publishing<\/em>. The current scholarly publication process assumes that an author uploads a scientific manuscript to a centralized platform, where a closed group of reviewers evaluates it. After acceptance, the manuscript is published as an article and then becomes accessible to the public, possibly at a fee. This process is rather slow, as the wider scientific community can only read the article at the end\u2014<wbr \/>if accepted. It is also non-transparent because valuable artefacts of the process, such as reviews and revisions, remain hidden. Further participation is typically only possible through a reply that has to undergo a similar slow process. A decentralized authoring application such as <a href=\"http:\/\/csarven.ca\/dokieli-rww\" target=\"_blank\"><em>dokieli<\/em><\/a> <a href=\"#ref-19\" class=\"reference\">[19]<\/a> instead allows researchers to self-publish their manuscripts online in their own Solid pod. Their peers can annotate these manuscripts with comments and reviews, which are stored in their own pods, guaranteeing freedom of expression to any researcher who wants to participate. All outcomes of this process are online, and the scientific community can continuously provide feedback, even after publication on the Web.<\/p><h2 id=\"adecentralized-web-for-all\">A decentralized Web for all<\/h2><p id=\"adecentralized-web-for-all-p-1\">Re-decentralizing the Web along the lines of the Solid vision can help us tackle <a href=\"https:\/\/webfoundation.org\/2017\/03\/web-turns-28-letter\/\" target=\"_blank\">Tim Berners-Lee\u2019s three challenges<\/a> <a href=\"#ref-4\" class=\"reference\">[4]<\/a>. We can take back control of our personal data by storing data in our own data pods. The spread of misinformation can be halted, because a free choice of applications allows us to influence our news feed\u2014<wbr \/>and all information in there can be traced back to its source. Political advertising becomes more transparent, as we can decide which parts of our data we expose to whom; moreover, the separate data and services markets allow us to consider other options that are not based on advertising in the first place. While this does not fully address all aspects of the challenges, data ownership and choice are major factors.<\/p><p id=\"adecentralized-web-for-all-p-2\">Freedom of course always comes at a cost: what constitutes a victory for personal rights and freedom of speech also facilitates the spread of illegal messages, since decentralized networks make it harder to control what information is exchanged. Legality is of course a tricky matter, as some countries instate laws that prevent their citizens from voicing opinions that would be legal elsewhere. An intriguing case is the <a href=\"http:\/\/www.ethanzuckerman.com\/blog\/2017\/08\/18\/mastodon-is-big-in-japan-the-reason-why-is-uncomfortable\/\" target=\"_blank\">increased popularity of the decentralized social network Mastodon in Japan<\/a> <a href=\"#ref-20\" class=\"reference\">[20]<\/a>: as Twitter started removing images that were deemed questionable under <abbr class=\"caps\">US<\/abbr> norms, Japanese users began publishing them on platforms with less censorship. We will have to accept this trade-off between freedom and control\u2014<wbr \/>and in absence of a globally accepted set of norms, centralized filtering of questionable or illegal content can never yield an adequate solution.<\/p><p id=\"adecentralized-web-for-all-p-3\">This brings us to another aspect of decentralization, which is the tension between freedom and universality. The Paradox of Freedom states that we can only be free if we subject ourselves to certain rules. Simply said, we can take our bike and ride anywhere\u2014<wbr \/>if only we stay on the right side of the road (which in several countries is actually left). Without such rules, we would not be able to get anywhere without causing accidents. Given that <a href=\"https:\/\/www.w3.org\/2005\/Talks\/0323-yorkshire-tbl\/slide5-2.html\" target=\"_blank\">universality has always been a main goal of the Web<\/a> <a href=\"#ref-7\" class=\"reference\">[7]<\/a>, decentralized communities can only flourish if they agree on some basic framework on <em>how<\/em> to decentralize. As with the universality of browsers, there is a major role for the <abbr class=\"caps\">W3C<\/abbr> in creating the standards that will allow decentralized data pods and apps to interoperate. Fortunately, we do not have to agree on everything. Linked Data enables layered agreements, in which a few rules are adopted by many, and sets of additional rules are agreed upon by smaller groups as needed.<\/p><p id=\"adecentralized-web-for-all-p-4\">Importantly, the arrows of decentralization and Solid are not aimed at specific companies such as Google, Facebook, or Twitter. Instead, they point at centralization in general, since many of the problems and challenges faced by these companies are inherent to centralization and the business model of data ownership. We have come to the point where companies possess so much data that they themselves are <a href=\"https:\/\/berjon.com\/advertising-war-on-consent\/\" target=\"_blank\">unable to predict the long-term effects that such a centralization might have<\/a> <a href=\"#ref-10\" class=\"reference\">[10]<\/a>. Therefore, it is unreasonable to use <q>informed consent<\/q> as an excuse, since no individual can reasonably understand what giving up control over small or large pieces of their data will eventually lead to. Storing our data in a trusted place of our choice, combined with a granular permission model, is therefore the only safe bet.<\/p><p id=\"adecentralized-web-for-all-p-5\">Note that none of us are dreaming of a Web without large players. Quite the contrary: Tim Berners-Lee insists that the Web should always remain scale-free <a href=\"#ref-21\" class=\"reference\">[21]<\/a>, with room for the very large and the very small. The problem is that the very large are currently trying to make the rest obsolete, which endangers the online freedoms we have enjoyed for so many years. As argued above, decentralization is foremost about choice, so people should be free to join large or small communities. And while there are several technical issues ahead of us for decentralized applications, notably guaranteeing a similar user experience as centralized platforms in terms of usability and speed, the first technological proof has been delivered with Solid. Now, it is up to all of us to anchor this technological progress in today\u2019s and tomorrow\u2019s socio-economic reality in order to re-decentralize the Web for good. Only when we succeed in taking back control and choice over our most precious digital assets, we are able to truly say: <em>this is for everyone<\/em>.<\/p><h2 id=\"references\">References<\/h2><dl class=\"references\"><dt id=\"ref-1\">[1]<\/dt> <dd>Derakhshan, H. (2015), \u201c<cite>The Web We Have to Save<\/cite>\u201d, 14 July, available at: <a href=\"https:\/\/medium.com\/matter\/the-web-we-have-to-save-2eb1fe15a426\" target=\"_blank\">https:\/\/medium.com\/<wbr \/>matter\/<wbr \/>the-web-we-have-to-save-2eb1fe15a426<\/a>.<\/dd> <dt id=\"ref-2\">[2]<\/dt> <dd>\u201c<cite>Break down these walls<\/cite>\u201d. (2008), <i>The Economist<\/i>, available at: <a href=\"https:\/\/www.economist.com\/node\/10880516\" target=\"_blank\">https:\/\/www.economist.com\/<wbr \/>node\/<wbr \/>10880516<\/a>.<\/dd> <dt id=\"ref-3\">[3]<\/dt> <dd>Pariser, E. (2011), <cite class=\"italic\">The Filter Bubble<\/cite>, Penguin Books.<\/dd> <dt id=\"ref-4\">[4]<\/dt> <dd>Berners-Lee, T. (2017), \u201c<cite>Three challenges for the Web, according to its inventor<\/cite>\u201d, Web Foundation, 12 March, available at: <a href=\"https:\/\/webfoundation.org\/2017\/03\/web-turns-28-letter\/\" target=\"_blank\">https:\/\/webfoundation.org\/<wbr \/>2017\/<wbr \/>03\/<wbr \/>web-turns-28-letter\/<\/a>.<\/dd> <dt id=\"ref-5\">[5]<\/dt> <dd>Rosenthal, D. (2018), \u201c<cite>It Isn\u2019t About The Technology<\/cite>\u201d, 11 January, available at: <a href=\"https:\/\/blog.dshr.org\/2018\/01\/it-isnt-about-technology.html\" target=\"_blank\">https:\/\/blog.dshr.org\/<wbr \/>2018\/<wbr \/>01\/<wbr \/>it-isnt-about-technology.html<\/a>.<\/dd> <dt id=\"ref-6\">[6]<\/dt> <dd>Berners-Lee, T. (2018), \u201c<cite>The Web is under threat. Join us and fight for it\u201d.<\/cite>, Web Foundation, 12 March, available at: <a href=\"https:\/\/webfoundation.org\/2018\/03\/web-birthday-29\/\" target=\"_blank\">https:\/\/webfoundation.org\/<wbr \/>2018\/<wbr \/>03\/<wbr \/>web-birthday-29\/<\/a>.<\/dd> <dt id=\"ref-7\">[7]<\/dt> <dd>Berners-Lee, T. (2005), \u201c<cite>Universality of the Web<\/cite>\u201d, 23 March, available at: <a href=\"https:\/\/www.w3.org\/2005\/Talks\/0323-yorkshire-tbl\/slide5-2.html\" target=\"_blank\">https:\/\/www.w3.org\/<wbr \/>2005\/<wbr \/>Talks\/<wbr \/>0323-yorkshire-tbl\/<wbr \/>slide5-2.html<\/a>.<\/dd> <dt id=\"ref-8\">[8]<\/dt> <dd>Berners\u2010Lee, T., Cailliau, R., Groff, J.F. and Pollermann, B. (1992), \u201c<cite>World\u2010wide web: the information universe<\/cite>\u201d, <i>Electronic Networking<\/i>, Vol. 2 No. 1.<\/dd> <dt id=\"ref-9\">[9]<\/dt> <dd>Gustafson, A. (2008), \u201c<cite>Beyond <abbr class=\"caps\">DOCTYPE<\/abbr>: Web Standards, Forward Compatibility, and <abbr class=\"caps\">IE8<\/abbr><\/cite>\u201d, 21 January, available at: <a href=\"https:\/\/alistapart.com\/article\/beyonddoctype\" target=\"_blank\">https:\/\/alistapart.com\/<wbr \/>article\/<wbr \/>beyonddoctype<\/a>.<\/dd> <dt id=\"ref-10\">[10]<\/dt> <dd>Berjon, R. (2018), \u201c<cite>Advertising\u2019s War on Consent<\/cite>\u201d, 19 March, available at: <a href=\"https:\/\/berjon.com\/advertising-war-on-consent\/\" target=\"_blank\">https:\/\/berjon.com\/<wbr \/>advertising-war-on-consent\/<\/a>.<\/dd> <dt id=\"ref-11\">[11]<\/dt> <dd>Berners-Lee, T. (2018), \u201c<cite>From Utopia to Dystopia in 29 Short Years<\/cite>\u201d, 18 May, available at: <a href=\"https:\/\/www.csail.mit.edu\/news\/utopia-dystopia-29-short-years\" target=\"_blank\">https:\/\/www.csail.mit.edu\/<wbr \/>news\/<wbr \/>utopia-dystopia-29-short-years<\/a>.<\/dd> <dt id=\"ref-12\">[12]<\/dt> <dd>Kosinski, M., Stillwell, D. and Graepel, T. (2013), \u201c<cite>Private traits and attributes are predictable from digital records of human behavior<\/cite>\u201d, <i>Proceedings of the National Academy of Sciences<\/i>, National Academy of Sciences, Vol. 110 No. 15, pp. 5802\u20135805.<\/dd> <dt id=\"ref-13\">[13]<\/dt> <dd>Samarajiva, R. (2014), \u201c<cite>More Facebook users than Internet users in South East Asia?<\/cite>\u201d, 30 August, available at: <a href=\"http:\/\/lirneasia.net\/2014\/08\/more-facebook-users-than-internet-users-in-south-east-asia\/\" target=\"_blank\">http:\/\/lirneasia.net\/<wbr \/>2014\/<wbr \/>08\/<wbr \/>more-facebook-users-than-internet-users-in-south-east-asia\/<\/a>.<\/dd> <dt id=\"ref-14\">[14]<\/dt> <dd>\u201c<cite>Solid<\/cite>\u201d. (n.d.). , available at: <a href=\"https:\/\/solid.mit.edu\/\" target=\"_blank\">https:\/\/solid.mit.edu\/<\/a>.<\/dd> <dt id=\"ref-15\">[15]<\/dt> <dd>Berners-Lee, T. (2006), \u201c<cite>Linked Data<\/cite>\u201d, 27 July, available at: <a href=\"https:\/\/www.w3.org\/DesignIssues\/LinkedData.html\" target=\"_blank\">https:\/\/www.w3.org\/<wbr \/>DesignIssues\/<wbr \/>LinkedData.html<\/a>.<\/dd> <dt id=\"ref-16\">[16]<\/dt> <dd>Berners-Lee, T. and O\u2019Hara, K. (2013), \u201c<cite>The read\u2013write Linked Data Web<\/cite>\u201d, <i>Philosophical Transactions of the Royal Society A<\/i>, Vol. 371 No. 1987.<\/dd> <dt id=\"ref-17\">[17]<\/dt> <dd>Berners-Lee, T., Hendler, J. and Lassila, O. (2001), \u201c<cite>The Semantic Web<\/cite>\u201d, <i>Scientific American<\/i>, Vol. 284 No. 5, pp. 34\u201343.<\/dd> <dt id=\"ref-18\">[18]<\/dt> <dd>Capadisli, S. and Guy, A. (Eds.). (2017), <cite class=\"italic\">Linked Data Notifications<\/cite>, Recommendation, World Wide Web Consortium, available at: <a href=\"https:\/\/www.w3.org\/TR\/ldn\/\" target=\"_blank\">https:\/\/www.w3.org\/<wbr \/><abbr class=\"caps\">TR<\/abbr>\/<wbr \/>ldn\/<\/a>.<\/dd> <dt id=\"ref-19\">[19]<\/dt> <dd>Capadisli, S., Guy, A., Verborgh, R., Lange, C., Auer, S. and Berners-Lee, T. (2017), \u201c<cite>Decentralised Authoring, Annotations and Notifications for a Read\u2013Write Web with dokieli<\/cite>\u201d, in <i>Proceedings of the 17<sup>th<\/sup> International Conference on Web Engineering<\/i>, pp. 469\u2013481, available at: <a href=\"http:\/\/csarven.ca\/dokieli-rww\" target=\"_blank\">http:\/\/csarven.ca\/<wbr \/>dokieli-rww<\/a>.<\/dd> <dt id=\"ref-20\">[20]<\/dt> <dd>Zuckerman, E. (2017), \u201c<cite>Mastodon is big in Japan. The reason why is\u2026 uncomfortable<\/cite>\u201d, 18 August, available at: <a href=\"http:\/\/www.ethanzuckerman.com\/blog\/2017\/08\/18\/mastodon-is-big-in-japan-the-reason-why-is-uncomfortable\/\" target=\"_blank\">http:\/\/www.ethanzuckerman.com\/<wbr \/>blog\/<wbr \/>2017\/<wbr \/>08\/<wbr \/>18\/<wbr \/>mastodon-is-big-in-japan-the-reason-why-is-uncomfortable\/<\/a>.<\/dd> <dt id=\"ref-21\">[21]<\/dt> <dd>Barab\u00e1si, A.-L. and Albert, R. (1999), \u201c<cite>Emergence of Scaling in Random Networks<\/cite>\u201d, <i>Science<\/i>, Vol. 286, pp. 509\u2013512.<\/dd> <\/dl>","created_at":"2019-01-13T13:57:17+0000","updated_at":"2019-01-13T13:57:17+0000","published_at":"2019-01-11T16:30:00+0000","published_by":["LinkedIn"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":28,"domain_name":"ruben.verborgh.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":770,"title":"You Weren't Meant to Have a Boss","url":"http:\/\/www.paulgraham.com\/boss.html","content":"<table style=\"width: 100%; border-spacing: 0px;\"><tr><td style=\"background-color: #ff9922;\"><img src=\"http:\/\/www.virtumundo.com\/images\/spacer.gif\" height=\"15\" width=\"1\" alt=\"image\" \/><span>\n<b>Want to start a startup?<\/b>  Get funded by\n<a href=\"http:\/\/ycombinator.com\/apply.html\">Y Combinator<\/a>.\n<\/span>\n<br \/><img src=\"http:\/\/www.virtumundo.com\/images\/spacer.gif\" height=\"5\" width=\"1\" alt=\"image\" \/><\/td><\/tr><\/table><p>\nMarch 2008, rev. June 2008<\/p><p>Technology tends to separate normal from natural.  Our bodies\nweren't designed to eat the foods that people in rich countries eat, or\nto get so little exercise.  \nThere may be a similar problem with the way we work: \na normal job may be as bad for us intellectually as white flour\nor sugar is for us physically.<\/p><p>I began to suspect this after spending several years working \nwith startup founders.  I've now worked with over 200 of them, and I've\nnoticed a definite difference between programmers working on their\nown startups and those working for large organizations.\nI wouldn't say founders seem happier, necessarily;\nstarting a startup can be very stressful. Maybe the best way to put\nit is to say that they're happier in the sense that your body is\nhappier during a long run than sitting on a sofa eating\ndoughnuts.<\/p><p>Though they're statistically abnormal, startup founders seem to be\nworking in a way that's more natural for humans.<\/p><p>I was in Africa last year and saw a lot of animals in the wild that\nI'd only seen in zoos before. It was remarkable how different they\nseemed. Particularly lions. Lions in the wild seem about ten times\nmore alive. They're like different animals. I suspect that working\nfor oneself feels better to humans in much the same way that living\nin the wild must feel better to a wide-ranging predator like a lion.\nLife in a zoo is easier, but it isn't the life they were designed\nfor.<\/p><p><b>Trees<\/b><\/p><p>What's so unnatural about working for a big company?  The root of\nthe problem is that humans weren't meant to work in such large\ngroups.<\/p><p>Another thing you notice when you see animals in the wild is that\neach species thrives in groups of a certain size.  A herd of impalas\nmight have 100 adults; baboons maybe 20; lions rarely 10.  Humans\nalso seem designed to work in groups, and what I've read about\nhunter-gatherers accords with research on organizations and my own\nexperience to suggest roughly what the ideal size is: groups of 8\nwork well; by 20 they're getting hard to manage; and a group of 50\nis really unwieldy.\n<span>[<a href=\"#f1n\"><span>1<\/span><\/a>]<\/span><\/p><p>Whatever the upper limit is, we are clearly not meant to work in\ngroups of several hundred.  And yet\u2014for reasons having more\nto do with technology than human nature\u2014a great many people\nwork for companies with hundreds or thousands of employees.<\/p><p>Companies know groups that large wouldn't work, so they divide\nthemselves into units small enough to work together.  But to\ncoordinate these they have to introduce something new: bosses.<\/p><p>These smaller groups are always arranged in a tree structure.  Your\nboss is the point where your group attaches to the tree.  But when\nyou use this trick for dividing a large group into smaller ones,\nsomething strange happens that I've never heard anyone mention\nexplicitly.  In the group one level up from yours, your boss\nrepresents your entire group.  A group of 10 managers is not merely\na group of 10 people working together in the usual way.  It's really\na group of groups.  Which means for a group of 10 managers to work\ntogether as if they were simply a group of 10 individuals, the group\nworking for each manager would have to work as if they were a single\nperson\u2014the workers and manager would each share only one\nperson's worth of freedom between them.<\/p><p>In practice a group of people are never able to act as if they were\none person.  But in a large organization divided into groups in\nthis way, the pressure is always in that direction.  Each group\ntries its best to work as if it were the small group of individuals\nthat humans were designed to work in.  That was the point of creating\nit.  And when you propagate that constraint, the result is that\neach person gets freedom of action in inverse proportion to the\nsize of the entire tree.\n<span>[<a href=\"#f2n\"><span>2<\/span><\/a>]<\/span><\/p><p>Anyone who's worked for a large organization has felt this.  You\ncan feel the difference between working for a company with 100\nemployees and one with 10,000, even if your group has only 10 people.<\/p><p><b>Corn Syrup<\/b><\/p><p>A group of 10 people within a large organization is a kind of fake\ntribe.  The number of people you interact with is about right.  But\nsomething is missing: individual initiative.  Tribes of hunter-gatherers\nhave much more freedom.  The leaders have a little more power than other\nmembers of the tribe, but they don't generally tell them what to\ndo and when the way a boss can.<\/p><p>It's not your boss's fault.  The real problem is that in the group\nabove you in the hierarchy, your entire group is one virtual person.\nYour boss is just the way that constraint is imparted to you.<\/p><p>So working in a group of 10 people within a large organization feels\nboth right and wrong at the same time.   On the surface it feels\nlike the kind of group you're meant to work in, but something major\nis missing.  A job at a big company is like high fructose corn\nsyrup: it has some of the qualities of things you're meant to like,\nbut is disastrously lacking in others.<\/p><p>Indeed, food is an excellent metaphor to explain what's wrong with\nthe usual sort of job.<\/p><p>For example, working for a big company is the default thing to do,\nat least for programmers.  How bad could it be?  Well, food shows\nthat pretty clearly.  If you were dropped at a random point in\nAmerica today, nearly all the food around you would be bad for you.\nHumans were not designed to eat white flour, refined sugar, high\nfructose corn syrup, and hydrogenated vegetable oil.  And yet if\nyou analyzed the contents of the average grocery store you'd probably\nfind these four ingredients accounted for most of the calories.\n\"Normal\" food is terribly bad for you.  The only people who eat\nwhat humans were actually designed to eat are a few Birkenstock-wearing\nweirdos in Berkeley.<\/p><p>If \"normal\" food is so bad for us, why is it so common?  There are\ntwo main reasons. One is that it has more immediate appeal.  You\nmay feel lousy an hour after eating that pizza, but eating the first\ncouple bites feels great.  The other is economies of scale.\nProducing junk food scales; producing fresh vegetables doesn't.\nWhich means (a) junk food can be very cheap, and (b) it's worth\nspending a lot to market it.<\/p><p>If people have to choose between something that's cheap, heavily\nmarketed, and appealing in the short term, and something that's\nexpensive, obscure, and appealing in the long term, which do you\nthink most will choose?<\/p><p>It's the same with work.  The average MIT graduate wants to work\nat Google or Microsoft, because it's a recognized brand, it's safe,\nand they'll get paid a good salary right away.  It's the job\nequivalent of the pizza they had for lunch.  The drawbacks will\nonly become apparent later, and then only in a vague sense of\nmalaise.<\/p><p>And founders and early employees of startups, meanwhile, are like\nthe Birkenstock-wearing weirdos of Berkeley:  though a tiny minority\nof the population, they're the ones living as humans are meant to.\nIn an artificial world, only extremists live naturally.<\/p><p><b>Programmers<\/b><\/p><p>The restrictiveness of big company jobs is particularly hard on\nprogrammers, because the essence of programming is to build new\nthings.  Sales people make much the same pitches every day; support\npeople answer much the same questions; but once you've written a\npiece of code you don't need to write it again.  So a programmer\nworking as programmers are meant to is always making new things.\nAnd when you're part of an organization whose structure gives each\nperson freedom in inverse proportion to the size of the tree, you're\ngoing to face resistance when you do something new.<\/p><p>This seems an inevitable consequence of bigness.  It's true even\nin the smartest companies.  I was talking recently to a founder who\nconsidered starting a startup right out of college, but went to\nwork for Google instead because he thought he'd learn more there.\nHe didn't learn as much as he expected.  Programmers learn by doing,\nand most of the things he wanted to do, he couldn't\u2014sometimes\nbecause the company wouldn't let him, but often because the company's\ncode wouldn't let him.  Between the drag of legacy code, the overhead\nof doing development in such a large organization, and the restrictions\nimposed by interfaces owned by other groups, he could only try a\nfraction of the things he would have liked to.  He said he has\nlearned much more in his own startup, despite the fact that he has\nto do all the company's errands as well as programming, because at\nleast when he's programming he can do whatever he wants.<\/p><p>An obstacle downstream propagates upstream.  If you're not allowed\nto implement new ideas, you stop having them.  And vice versa: when\nyou can do whatever you want, you have more ideas about what to do.\nSo working for yourself makes your brain more powerful in the same\nway a low-restriction exhaust system makes an engine more powerful.<\/p><p>Working for yourself doesn't have to mean starting a startup, of\ncourse.  But a programmer deciding between a regular job at a big\ncompany and their own startup is probably going to learn more doing\nthe startup.<\/p><p>You can adjust the amount of freedom you get by scaling the size\nof company you work for.  If you start the company, you'll have the\nmost freedom.  If you become one of the first 10 employees you'll\nhave almost as much freedom as the founders.  Even a company with\n100 people will feel different from one with 1000.<\/p><p>Working for a small company doesn't ensure freedom.  The tree\nstructure of large organizations sets an upper bound on freedom,\nnot a lower bound.  The head of a small company may still choose\nto be a tyrant.  The point is that a large organization is compelled\nby its structure to be one.<\/p><p><b>Consequences<\/b><\/p><p>That has real consequences for both organizations and individuals.\nOne is that companies will inevitably slow down as they grow larger,\nno matter how hard they try to keep their startup mojo.  It's a\nconsequence of the tree structure that every large organization is\nforced to adopt.<\/p><p>Or rather, a large organization could only avoid slowing down if\nthey avoided tree structure.  And since human nature limits the\nsize of group that can work together, the only way I can imagine\nfor larger groups to avoid tree structure would be to have no\nstructure: to have each group actually be independent, and to work\ntogether the way components of a market economy do.<\/p><p>That might be worth exploring.  I suspect there are already some\nhighly partitionable businesses that lean this way.  But I don't\nknow any technology companies that have done it.<\/p><p>There is one thing companies can do short of structuring themselves\nas sponges:  they can stay small.  If I'm right, then it really\npays to keep a company as small as it can be at every stage.\nParticularly a technology company.  Which means it's doubly important\nto hire the best people.  Mediocre hires hurt you twice: they get\nless done, but they also make you big, because you need more of\nthem to solve a given problem.<\/p><p>For individuals the upshot is the same: aim small.  It will always\nsuck to work for large organizations, and the larger the organization,\nthe more it will suck.<\/p><p>In an <a href=\"http:\/\/www.paulgraham.com\/start.html\">essay<\/a> I wrote a couple years ago \nI advised graduating seniors\nto work for a couple years for another company before starting their\nown.  I'd modify that now.  Work for another company if you want\nto, but only for a small one, and if you want to start your own\nstartup, go ahead.<\/p><p>The reason I suggested college graduates not start startups immediately\nwas that I felt most would fail.  And they will.  But ambitious\nprogrammers are better off doing their own thing and failing than\ngoing to work at a big company.  Certainly they'll learn more.  They\nmight even be better off financially.  A lot of people in their\nearly twenties get into debt, because their expenses grow even\nfaster than the salary that seemed so high when they left school.\nAt least if you start a startup and fail your net worth will be\nzero rather than negative.  \n<span>[<a href=\"#f3n\"><span>3<\/span><\/a>]<\/span><\/p><p>We've now funded so many different types of founders that we have\nenough data to see patterns, and there seems to be no benefit from\nworking for a big company.  The people who've worked for a few years\ndo seem better than the ones straight out of college, but only\nbecause they're that much older.<\/p><p>The people who come to us from big companies often seem kind of\nconservative.  It's hard to say how much is because big companies\nmade them that way, and how much is the natural conservatism that\nmade them work for the big companies in the first place.  But\ncertainly a large part of it is learned.  I know because I've seen\nit burn off.<\/p><p>Having seen that happen so many times is one of the things that\nconvinces me that working for oneself, or at least for a small\ngroup, is the natural way for programmers to live.  Founders arriving\nat Y Combinator often have the downtrodden air of refugees.  Three\nmonths later they're transformed: they have so much more \n<a href=\"http:\/\/paulmckellar.com\/photos\/03l\">confidence<\/a>\nthat they seem as if they've grown several inches taller. \n<span>[<a href=\"#f4n\"><span>4<\/span><\/a>]<\/span>\nStrange as this sounds, they seem both more worried and happier at the same\ntime.  Which is exactly how I'd describe the way lions seem in the\nwild.<\/p><p>Watching employees get transformed into founders makes it clear\nthat the difference between the two is due mostly to environment\u2014and\nin particular that the environment in big companies is toxic to\nprogrammers.   In the first couple weeks of working on their own\nstartup they seem to come to life, because finally they're working\nthe way people are meant to.<\/p><p><b>Notes<\/b><\/p><p>[<a name=\"f1n\" id=\"f1n\"><span>1<\/span><\/a>]\nWhen I talk about humans being meant or designed to live a\ncertain way, I mean by evolution.<\/p><p>[<a name=\"f2n\" id=\"f2n\"><span>2<\/span><\/a>]\nIt's not only the leaves who suffer.  The constraint propagates\nup as well as down.  So managers are constrained too; instead of\njust doing things, they have to act through subordinates.<\/p><p>[<a name=\"f3n\" id=\"f3n\"><span>3<\/span><\/a>]\nDo not finance your startup with credit cards.  Financing a\nstartup with debt is usually a stupid move, and credit card debt\nstupidest of all.  Credit card debt is a bad idea, period.  It is\na trap set by evil companies for the desperate and the foolish.<\/p><p>[<a name=\"f4n\" id=\"f4n\"><span>4<\/span><\/a>]\nThe founders we fund used to be younger (initially we encouraged\nundergrads to apply), and the first couple times I saw this I used\nto wonder if they were actually getting physically taller.<\/p><p><b>Thanks<\/b> to Trevor Blackwell, Ross Boucher, Aaron Iba, Abby\nKirigin, Ivan Kirigin, Jessica Livingston, and Robert Morris for\nreading drafts of this.<\/p>","created_at":"2019-01-11T14:57:32+0000","updated_at":"2019-01-11T14:57:32+0000","annotations":[],"mimetype":"text\/html","reading_time":12,"domain_name":"www.paulgraham.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":769,"title":"The worst ANSI renderer, except for all the others \u2013 Et tu, Cthulhu","url":"https:\/\/hpjansson.org\/blag\/2019\/01\/07\/the-worst-ansi-renderer-except-for-all-the-others\/","content":"<p style=\"text-align: center;\"><img class=\"alignnone size-medium\" src=\"https:\/\/i.imgur.com\/2RLD6MO.gif\" width=\"480\" height=\"504\" alt=\"image\" \/><\/p>\n<p><a href=\"https:\/\/hpjansson.org\/chafa\/\">Chafa<\/a> (<a href=\"https:\/\/github.com\/hpjansson\/chafa\">github<\/a>) started out as a small piece of supporting code for an obscure personal project I may announce at some indefinite point in the future. Then I <a href=\"https:\/\/hpjansson.org\/blag\/2018\/04\/24\/introducing-chafa\/\">decided<\/a> to release it as a tongue-in-cheek thing for the VT100 anniversary last year, and, well\u2026 it gathered a bit of steam.<\/p>\n<h2>Chafa 1.0<\/h2>\n<p>Since I\u2019m not one to leave well enough alone, I packaged it up over the holidays for the 1.0 <a href=\"https:\/\/github.com\/hpjansson\/chafa\/releases\">release<\/a>. It brings a pile of improvements, e.g. new symbol ranges like ASCII and Braille, better image preprocessing and a new <code>--fill<\/code> option for halftone\/dithering. It\u2019s also, like, real fast now, and the build is much less brittle.<\/p>\n<p>Big thanks to everyone who contributed to this release: <a href=\"https:\/\/github.com\/kilobyte\">Adam Borowski<\/a>, <a href=\"https:\/\/github.com\/felixonmars\">Felix Yan<\/a>, <a href=\"https:\/\/github.com\/lalyos\">Lajos Papp<\/a>, <a href=\"https:\/\/github.com\/cdluminate\">Mo Zhou<\/a>, <a href=\"https:\/\/github.com\/rarguello\">Ricardo Arguello<\/a>, <a href=\"https:\/\/github.com\/eclipseo\">Robert-Andr\u00e9 Mauchin<\/a>, <a href=\"https:\/\/github.com\/dcb314\">@dcb314<\/a> and <a href=\"https:\/\/github.com\/medusacle\">@medusacle<\/a>.<\/p>\n<p>You\u2019ll find <a href=\"https:\/\/pkgs.org\/download\/chafa\">packages<\/a> out there for Arch, Debian, Fedora, <a href=\"https:\/\/packages.gentoo.org\/packages\/media-gfx\/chafa\">Gentoo<\/a> and Ubuntu now; check your repositories. Extra big thanks to the package maintainers.<\/p>\n<p>As the post title implies, I think Chafa is now the least bad tool in this tiny but tradition-rich niche. So what\u2019s so not-quite-terrible about it?<\/p>\n<h2>Only the best for your aixterm<\/h2>\n<p>If you\u2019ve been around text terminals for a while, you\u2019ll know what this means:<\/p>\n<p style=\"text-align: center;\"><a href=\"https:\/\/hpjansson.org\/blag\/wp-content\/uploads\/2019\/01\/terminal-palette-tango.png\"><img class=\"aligncenter size-full wp-image-310\" src=\"https:\/\/hpjansson.org\/blag\/wp-content\/uploads\/2019\/01\/terminal-palette-tango.png\" alt=\"\" width=\"329\" height=\"83\" \/><\/a><\/p>\n<p>Up until fairly recently, the most colorful terminal applications would operate within the confines of some variant of the above palette, and many still do. It\u2019s a nice and distinct (not to mention cheerful) set of colors, but it\u2019s not a great match for most image material you\u2019ll come across, which makes legible Chafa output a challenge. And that is precisely why I had to try (here with a <a href=\"https:\/\/web.archive.org\/web\/20180328052246if_\/http:\/\/baaad.com.au\/resources\/P4048.JPG?timestamp=1396074690936\">nice woodblock print<\/a>):<\/p>\n<p style=\"text-align: center;\"><a href=\"https:\/\/hpjansson.org\/blag\/wp-content\/uploads\/2019\/01\/chafa-examples-preproc.png\"><img class=\"aligncenter size-full wp-image-306\" src=\"https:\/\/hpjansson.org\/blag\/wp-content\/uploads\/2019\/01\/chafa-examples-preproc.png\" alt=\"\" width=\"766\" height=\"550\" \/><\/a><\/p>\n<p>The top left image is the best reproduction you can get with Chafa using a modern 24-bit color terminal emulator (in this case, GNOME Terminal) at 60 characters wide. At top right is the 16-color mapping you get without applying any tricks; this is pretty close to the mark given the muted moonlight colors of the input image, but it doesn\u2019t make good use of our palette, nor does it convey the essence of the scene very well. Blue sky, rolling hills, green grass. A shady waterfront pavilion. Given our limitations, the output will look nothing like the original anyway, so we\u2019re better off trying to capture the gist of it.<\/p>\n<p>We do this simply by cranking up the contrast and saturation to levels where our cheerful old palette can do a decent job (bottom left). Chafa no longer relies on ImageMagick for this, so it\u2019s available in the <a href=\"https:\/\/hpjansson.org\/chafa\/ref\/\">library API<\/a>, and integer-only math makes it performant enough to run in real time on animations and video (more on this in a later post, perhaps).<\/p>\n<p>It gets even better if you do color assignment in DIN99d space (bottom right), but that\u2019s way slow, so you have to explicitly enable it.<\/p>\n<h2>No ANSI? No problem<\/h2>\n<p style=\"text-align: center;\"><a href=\"https:\/\/hpjansson.org\/blag\/wp-content\/uploads\/2019\/01\/chafa-examples-noansi.png\"><img class=\"aligncenter size-full wp-image-311\" src=\"https:\/\/hpjansson.org\/blag\/wp-content\/uploads\/2019\/01\/chafa-examples-noansi.png\" alt=\"\" width=\"762\" height=\"274\" \/><\/a><\/p>\n<p>You can generate varied output without any color codes at all. The above examples also demonstrate use of the <code>--fg<\/code> option, which, along with <code>--bg<\/code>, exists to tell Chafa what your terminal\u2019s default colors look like so it can target those, but is equally useful for tweaking monochrome output thresholds.<\/p>\n<p>So if for some reason your terminal can\u2019t handle ANSI or you still have one of those warm &amp; fuzzy monochrome tubes sitting around, we\u2019ve got you covered. Or break out your civil rights-era dot-matrix printer and make some precision banners to hang in your lab!<\/p>\n<p style=\"text-align: center;\"><img class=\"alignnone size-medium\" src=\"https:\/\/hpjansson.org\/chafa\/img\/example-3.gif\" width=\"481\" height=\"289\" alt=\"image\" \/><\/p>","created_at":"2019-01-08T15:07:27+0000","updated_at":"2019-01-08T15:07:27+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"hpjansson.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":768,"title":"r\/lisp - How innovative is Clojure as a Lisp dialect from Lisp expert's perspective?","url":"https:\/\/www.reddit.com\/r\/lisp\/comments\/acid7a\/how_innovative_is_clojure_as_a_lisp_dialect_from\/ed89dkq\/","content":"<header class=\"_2GyPfdsi-MbQFyHRECo9GO cx1ohrUAq6ARaXTX2u8YN  lwbhnq-0 cXhqKZ s1wmmuoa-0 iFqXhZ\" data-redditstyle=\"true\"><div class=\"_2vkeRJojnV7cb9pMlPHy7d\"><div class=\"_3dnbqz69WJTFCss8wl7Wlk\">Press J to jump to the feed. Press question mark to learn the rest of the keyboard shortcuts<a aria-label=\"Home\" class=\"_30BbATRhFv3V83DHNDjJAO\" href=\"https:\/\/www.reddit.com\/\"><\/a><\/div><\/div><\/header><div class=\"_1nxEQl5D2Bx2jxDILRHemb  lwbhnq-1 jnERZH\" aria-hidden=\"false\"><div><div class=\"ml0npc-0 iawzai\"><div class=\"ml0npc-1 elAttH\"><div class=\"ml0npc-5 dtrVxj\"><div class=\"s18m1py3-0 gQkYwg\"><div class=\"s18hdlw1-2 dHeuJR\"><div class=\"s69d4o1-10 hvSwby _2rszc84L136gWQrkwH6IaM Post t3_acid7a  s1ukwo15-0 RqhAo\" id=\"t3_acid7a\" tabindex=\"-1\"><div data-test-id=\"post-content\"><div class=\"_3MC4c3Q_Y41YKtl1TcvyMt\"><div class=\"_3UMN4RCVY5288m_fOZlkcg\"><div class=\"s1ssr92a-0 hKePuf\"><div class=\"cZPZhMe-UCZ8htPodMyJ5\"><div class=\"_3AStxql1mQsrZuUIFP9xSg nU4Je7n-eSXStTBAPMYt8\">Posted by<div class=\"xvda30-0 camSYk\"><a class=\"_2tbHP6ZydRpjI44J3syuqC s1461iz-1 gWXVVu\" href=\"https:\/\/www.reddit.com\/user\/lyssphacker\">u\/lyssphacker<\/a><\/div><\/div><\/div><\/div><div class=\"_1rcejqgj_laTEewtp2DbWG s69d4o1-0 jkTHH\"><div class=\"_1hLrLjnE1G_RBCNcN9MVQf\"><img src=\"https:\/\/www.redditstatic.com\/desktop2x\/img\/renderTimingPixel.png\" alt=\"image\" \/><\/div><\/div><div><div class=\"_3-miAEojrCvx_4FQ8x3P-s s69d4o1-4 fFSuw\"><a rel=\"nofollow\" data-click-id=\"comments\" data-test-id=\"comments-page-link-num-comments\" class=\"_1UoeAeSRhOKSNdY_h3iS1O _1Hw7tY9pMr-T1F4P1C-xNU _2qww3J5KKzsD7e5DO0BvvU\" href=\"https:\/\/www.reddit.com\/r\/lisp\/comments\/acid7a\/how_innovative_is_clojure_as_a_lisp_dialect_from\/\">138 comments<\/a><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"s15ogqrm-1 hOSIHi\"><p>Sort by<\/p><div class=\"_21kEJV3xUcv3UfNXceidq4\">Single comment thread.\u00a0<a class=\"NlqxHck26RgynrMKMT-9H\" href=\"https:\/\/www.reddit.com\/r\/lisp\/comments\/acid7a\/how_innovative_is_clojure_as_a_lisp_dialect_from\/\">View all comments\u00a0<\/a><\/div><\/div><\/div><\/div><div class=\"ml0npc-6 garQss\"><div class=\"RffBqG1xCvuy1r49wYhLr dicowf-2 OIJll\"><div class=\"izgy8a-0 dQDarC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Community Details<\/p><\/div><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1wbv0ui-1 hDRsxi\"><a class=\"s1wbv0ui-6 iRWtcD\" href=\"https:\/\/www.reddit.com\/r\/lisp\/\">r\/lisp<\/a><\/div><div class=\"s1wbv0ui-10 JYeEH\"><div class=\"s1wbv0ui-11 caYJOw\"><p class=\"s1wbv0ui-12 ffQWxW\">162<\/p><p class=\"s1wbv0ui-13 egStHf\" id=\"IdCard--CurrentlyViewing--undefined--t5_2qh35\">Lispers online<\/p><\/div><\/div><p class=\"s1wbv0ui-15 emnQcM\">A subreddit for the Lisp family of programming languages.<\/p><p><a class=\"s1wbv0ui-18 jZqmse s1w1mqsg-2 ifvzlp\" href=\"https:\/\/www.reddit.com\/r\/lisp\/submit\">Create Post<\/a><\/p><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Reddit Friends<\/p><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Language References<\/p><\/div><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><p><strong>Common Lisp<\/strong><\/p>\n<ul><li><a href=\"http:\/\/www.lispworks.com\/documentation\/HyperSpec\/Front\/Contents.htm\">CL Hyperspec<\/a><\/li>\n<li><a href=\"http:\/\/phoe.tymoon.eu\/clus\">CL UltraSpec<\/a> (work-in-progress)<\/li>\n<li><a href=\"http:\/\/cvberry.com\/tech_writings\/notes\/common_lisp_standard_draft.html\">CL Spec (as searchable pdf)<\/a><\/li>\n<li>CL Meta-Object Protocol: <a href=\"http:\/\/mop.lisp.se\/www.alu.org\/mop\/\">alu-original<\/a>, <a href=\"http:\/\/metamodular.com\/CLOS-MOP\/\">vanilla-html<\/a>, <a href=\"https:\/\/clos-mop.hexstreamsoft.com\/\">HTML5-mobile<\/a><\/li>\n<li><a href=\"http:\/\/clqr.boundp.org\">CL Quick Reference<\/a><\/li>\n<li><a href=\"https:\/\/docs.google.com\/document\/pub?id=1SYSr0KE8iyv61srv6AMm-YRr-O6zM7Oi6CLSJgZc9yc&amp;pli=1\">CL Type Hierarchy<\/a> <a href=\"http:\/\/sellout.github.io\/media\/CL-type-hierarchy.png\">(alt w\/mop)<\/a><\/li>\n<\/ul><\/li>\n<li><p><strong>Scheme<\/strong><\/p>\n<ul><li><a href=\"http:\/\/www.r7rs.org\/\">r<sup>7<\/sup> rs<\/a><\/li>\n<li><a href=\"http:\/\/www.r6rs.org\/\">r<sup>6<\/sup> rs<\/a><\/li>\n<li><a href=\"http:\/\/www.schemers.org\/Documents\/Standards\/R5RS\/\">r<sup>5<\/sup> rs<\/a> <a href=\"https:\/\/docs.racket-lang.org\/r5rs\/index.html\">mirror<\/a><\/li>\n<li><a href=\"https:\/\/docs.racket-lang.org\/reference\/\">Racket<\/a><\/li>\n<\/ul><\/li>\n<li><p><strong>Clojure<\/strong><\/p>\n<ul><li><a href=\"https:\/\/clojure.org\/reference\/reader\">Reference<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><a href=\"http:\/\/www.quicklisp.org\/\">Quicklisp<\/a> (CL Library manager)<\/li>\n<li><a href=\"https:\/\/portacle.github.io\/\">Portacle<\/a> (Portable CL environment)<\/li>\n<li><a href=\"https:\/\/github.com\/roswell\/roswell\/\">Roswell<\/a> (CL Environment Setup Utility)<\/li>\n<li>IDEs for Common Lisp\n<ul><li><a href=\"http:\/\/common-lisp.net\/project\/slime\/\">SLIME<\/a> (GNU Emacs)<\/li>\n<li><a href=\"http:\/\/www.vim.org\/scripts\/script.php?script_id=2531\">SLIMV<\/a> (Vim)<\/li>\n<li><a href=\"https:\/\/github.com\/l04m33\/vlime\">Vlime<\/a> (Vim)<\/li>\n<li><a href=\"https:\/\/atom.io\/packages\/atom-slime\">Atom-Slime<\/a> (Atom)<\/li>\n<\/ul><\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Tutorials\/FAQs<\/p><\/div><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><a href=\"http:\/\/articulate-lisp.com\">Articulate Common Lisp, how to write Common Lisp today<\/a><\/li>\n<li><a href=\"http:\/\/lisp-lang.org\/\">(Common) Lisp the Language <\/a> (Advocacy, getting started, tutorials, wiki)<\/li>\n<li><a href=\"http:\/\/xach.livejournal.com\/325276.html\">Zach Beane: Where to get help with Common Lisp<\/a><\/li>\n<li><a href=\"http:\/\/www.cs.northwestern.edu\/academics\/courses\/325\/readings\/luv-slides.pdf\">Lisp Style Guide Norvig\/Pitman<\/a> (PDF)<\/li>\n<li><a href=\"http:\/\/random-state.net\/files\/nikodemus-cl-faq.html\">Nikodemus' CL FAQ<\/a><\/li>\n<li><a href=\"https:\/\/google.github.io\/styleguide\/lispguide.xml\">Google Common Lisp Style Guide<\/a> (2014)<\/li>\n<li><a href=\"https:\/\/awesome-cl.com\/\">Awesome CL<\/a> (library recommendations)<\/li>\n<li><a href=\"http:\/\/stevelosh.com\/blog\/2018\/08\/a-road-to-common-lisp\/\"><em>A<\/em> Road to Common Lisp (2018)<\/a> (noob recommendations)<\/li>\n<li><a href=\"https:\/\/lispmethods.com\">Lisp Methods<\/a> (noobs)<\/li>\n<li><a href=\"http:\/\/borretti.me\/article\/common-lisp-sotu-2015\">State of the CL Ecosystem 2015<\/a> (library recommendations)<\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><a href=\"http:\/\/planet.lisp.org\/\">Planet Lisp<\/a> (meta-blog)<\/li>\n<li><a href=\"http:\/\/scheme.dk\/planet\/\">Planet Scheme<\/a> (meta-blog)<\/li>\n<li><a href=\"http:\/\/www.lispforum.com\/\">Lisp Forum<\/a><\/li>\n<li><a href=\"http:\/\/www.cliki.net\/index\">Cliki<\/a> (library reference)<\/li>\n<li><a href=\"http:\/\/common-lisp.net\/\">Common-Lisp.net<\/a> (project hosting, mailing lists)<\/li>\n<li><a href=\"http:\/\/groups.google.com\/forum\/#!forum\/comp.lang.lisp\">comp.lang.lisp<\/a> (usenet group)<\/li>\n<li><a href=\"https:\/\/mailman.common-lisp.net\/listinfo\/pro\">Common Lisp Professionals Mailing List<\/a><\/li>\n<li><a href=\"http:\/\/lispjobs.wordpress.com\">Lisp Jobs<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Common_Lisp\">Common Lisp, Wikipedia<\/a><\/li>\n<li><a href=\"https:\/\/web.archive.org\/web\/20160827235937\/http:\/\/alu.org\/alu\/home\">Association of Lisp Users<\/a> [<strong>archive<\/strong>]<\/li>\n<li><a href=\"http:\/\/stackoverflow.com\/questions\/tagged\/lisp\">Stackoverflow Lisp questions<\/a>, <a href=\"http:\/\/stackoverflow.com\/questions\/tagged\/common-lisp\">Common Lisp only<\/a>, <a href=\"https:\/\/stackoverflow.com\/questions\/tagged\/scheme\">Scheme only<\/a><\/li>\n<li><a href=\"https:\/\/codereview.stackexchange.com\/questions\/tagged\/lisp+or+scheme+or+common-lisp\">Code Review<\/a> (Lisp, CL, Scheme)<\/li>\n<li><a href=\"http:\/\/rosettacode.org\/wiki\/Category:Common_Lisp\">Rosetta Code, Common Lisp<\/a><\/li>\n<li><a href=\"https:\/\/mailman.common-lisp.net\/listinfo\">Mailing Lists<\/a>, <a href=\"https:\/\/common-lisp.net\/independent-lists\/\">more<\/a><\/li>\n<li><a href=\"https:\/\/twitter.com\/search?q=%23lisp\">Twitter #lisp<\/a><\/li>\n<li><a href=\"https:\/\/common-lisp.net\/project\/ecl\/tag\/quarterly.html\">ECL Quarterly<\/a><\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><p><em>Free, Complete, On-line, Authorized<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.cs.cmu.edu\/%7Edst\/LispBook\/\">Common Lisp: A Gentle Introduction to Symbolic Computation<\/a> (Touretzsky, 1990)<\/li>\n<li><a href=\"https:\/\/www.cse.buffalo.edu\/%7Eshapiro\/Commonlisp\/\">Common Lisp: An Interactive Approach<\/a> (Shapiro, 1992)<\/li>\n<li><a href=\"https:\/\/lispcookbook.github.io\/cl-cookbook\/\">The Common Lisp Cookbook 2018<\/a> \/ <a href=\"http:\/\/cl-cookbook.sourceforge.net\">Original 2007<\/a><\/li>\n<li><a href=\"http:\/\/www.cs.cmu.edu\/Groups\/AI\/html\/cltl\/cltl2.html\">Common Lisp The Language, 2nd Edition<\/a> [Pre ANSI] (Steele, 1990)<\/li>\n<li><a href=\"http:\/\/nicklevine.org\/lisp-book\/\">Lisp Outside the Box<\/a> (unfinished, Levine, 2011)<\/li>\n<li><a href=\"http:\/\/www.paulgraham.com\/onlisp.html\">On Lisp<\/a> (Graham, 1993)<\/li>\n<li><a href=\"https:\/\/github.com\/norvig\/paip-lisp\/\">Paradigms of Artificial Intelligence Programming<\/a> (Norvig, 1992)<\/li>\n<li><a href=\"http:\/\/www.gigamonkeys.com\/book\/\">Practical Common Lisp<\/a> (Seibel, 2005)<\/li>\n<li><a href=\"http:\/\/mitpress.mit.edu\/sicp\/full-text\/book\/book.html\">Structure and Interpretation of Computer Programs<\/a>, <a href=\"https:\/\/sarabander.github.io\/sicp\/\">[HTML5\/EPUB3]<\/a> [Scheme] (Abelson\/Sussman, 1996)<\/li>\n<\/ul><\/li>\n<li><p><em>Other Books<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.paulgraham.com\/acl.html\">ANSI Common Lisp<\/a> (Graham, 1995)<\/li>\n<li><a href=\"http:\/\/weitz.de\/cl-recipes\">Common Lisp Recipes<\/a> (Weitz, 2016)<\/li>\n<li><a href=\"http:\/\/landoflisp.com\">Land of Lisp<\/a> (Barski, 2010)<\/li>\n<li><a href=\"http:\/\/letoverlambda.com\">Let over Lambda<\/a> (Hoyte, 2008)<\/li>\n<li><a href=\"http:\/\/people.csail.mit.edu\/phw\/Books\/\">Lisp, 3rd Edition<\/a> (Winston\/Horn, 1989)<\/li>\n<li><a href=\"https:\/\/www.pearsonhighered.com\/program\/Keene-Object-Oriented-Programming-in-Common-Lisp-A-Programmer-s-Guide-to-CLOS\/PGM262224.html\">Object-Oriented Programming in Common Lisp: A Programmer's Guide to CLOS<\/a> (Keene, 1989)<\/li>\n<li><a href=\"https:\/\/mitpress.mit.edu\/books\/art-metaobject-protocol\">The Art of the Metaobject Protocol<\/a> (Kiczales\/des Rivi\u00e8res\/Bobrow, 1991)<\/li>\n<\/ul><\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Food for Thought<\/p><\/div><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><a href=\"http:\/\/lambda-the-ultimate.org\/\">Lambda the Ultimate<\/a><\/li>\n<li><a href=\"http:\/\/c2.com\/cgi\/fullSearch?search=CategoryLisp\">C2 Wiki<\/a><\/li>\n<li><a href=\"https:\/\/www.xach.com\/naggum\/articles\/\">Erik Naggum comp.lang.lisp archive<\/a><\/li>\n<li><a href=\"https:\/\/www.xach.com\/rpw3\/articles\/\">Rob Warnock Lisp usenet archive<\/a><\/li>\n<li><a href=\"http:\/\/www.p-cos.net\/lisp\/guide.html\">Pascal Costanza's Highly Opinionated Guide to Lisp<\/a><\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"izgy8a-0 dQDarC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Implementations<\/p><\/div><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1i4gs3u-0 hNcetJ\"><div class=\"s1i4gs3u-1 YfRFq ser2k5-0 fiwlJt\"><div class=\"md\"><ul><li><em>CL Open Source<\/em>\n<ul><li><a href=\"http:\/\/common-lisp.net\/project\/armedbear\/\">ABCL<\/a> (for JVM)<\/li>\n<li><a href=\"http:\/\/ccl.clozure.com\/\">CCL<\/a><\/li>\n<li><a href=\"http:\/\/www.clisp.org\/\">CLISP<\/a><\/li>\n<li><a href=\"http:\/\/cmucl.org\/\">CMUCL<\/a>, <a href=\"https:\/\/gitlab.common-lisp.net\/cmucl\/cmucl\/wikis\/home\">mirror<\/a> (*nix)<\/li>\n<li><a href=\"https:\/\/common-lisp.net\/project\/ecl\/\">ECL<\/a><\/li>\n<li><a href=\"http:\/\/www.gnu.org\/software\/gcl\/\">GCL<\/a><\/li>\n<li><a href=\"http:\/\/www.sbcl.org\/\">SBCL<\/a><\/li>\n<\/ul><\/li>\n<li><em>CL Commercial<\/em>\n<ul><li><a href=\"http:\/\/www.franz.com\/products\/allegro-common-lisp\/\">Allegro CL<\/a><\/li>\n<li><a href=\"http:\/\/www.lispworks.com\/\">LispWorks<\/a><\/li>\n<li><a href=\"https:\/\/wukix.com\/mocl\">mocl<\/a> for OSX, iOS and Android<\/li>\n<li><a href=\"http:\/\/www.symbolics-dks.com\/\">Open Genera<\/a><\/li>\n<li><a href=\"http:\/\/www.scieneer.com\/scl\/\">Scieneer CL<\/a><\/li>\n<\/ul><\/li>\n<li><em>CL Developmental<\/em>\n<ul><li><a href=\"https:\/\/github.com\/drmeister\/clasp\">Clasp<\/a> (llvm)<\/li>\n<li><a href=\"https:\/\/github.com\/jscl-project\/jscl\">JSCL<\/a> (javascript)<\/li>\n<li><a href=\"https:\/\/github.com\/froggey\/Mezzano\">Mezzano<\/a> (Lisp OS)<\/li>\n<li><a href=\"http:\/\/common-lisp.net\/project\/mkcl\/\">MKCL<\/a> (fork of ECL)<\/li>\n<li><a href=\"https:\/\/github.com\/robert-strandh\/SICL\">SICL<\/a> (modular)<\/li>\n<\/ul><\/li>\n<li><em>CL Historical<\/em>\n<ul><li><a href=\"https:\/\/www.informatik.uni-kiel.de\/\/%7Ewg\/clicc.html\">CLiCC<\/a>  (CL\u2192C)<\/li>\n<li><a href=\"https:\/\/github.com\/sharplispers\/cormanlisp\">Corman Lisp<\/a> (MS-Windows)<\/li>\n<li><a href=\"https:\/\/github.com\/blakemcbride\/eclipse-lisp\">Eclipse Common Lisp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/vsedach\/Thinlisp-1.1\">ThinLisp<\/a> (CL\u2192C)<\/li>\n<li><a href=\"http:\/\/pgc.com\/commonlisp\/\">WCL<\/a> (embeddable)<\/li>\n<\/ul><\/li>\n<li><em>Scheme<\/em>\n<ul><li><a href=\"http:\/\/www-sop.inria.fr\/mimosa\/fp\/Bigloo\/\">Bigloo<\/a><\/li>\n<li><a href=\"https:\/\/www.call-cc.org\/\">Chicken<\/a><\/li>\n<li><a href=\"http:\/\/gambitscheme.org\/wiki\/index.php\/Main_Page\">Gambit<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/vyzo\/gerbil\/wiki\/Gerbil-Scheme\">Gerbil<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/guile\/\">Guile<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/kawa\/\">Kawa<\/a><\/li>\n<li><a href=\"http:\/\/www.racket-lang.org\/\">Racket<\/a><\/li>\n<\/ul><\/li>\n<li><em>Mature Dialects<\/em>\n<ul><li><a href=\"https:\/\/clojure.org\/\">Clojure<\/a><\/li>\n<li><a href=\"https:\/\/opendylan.org\/\">Dylan<\/a><\/li>\n<li><a href=\"http:\/\/www.islisp.org\/\">ISLisp<\/a><\/li>\n<li><a href=\"http:\/\/www.newlisp.org\/\">NewLisp<\/a><\/li>\n<li><a href=\"http:\/\/www.picolisp.com\/\">PicoLisp<\/a><\/li>\n<li><a href=\"http:\/\/shenlanguage.org\/\">Shen<\/a><\/li>\n<\/ul><\/li>\n<li><em>Developing Dialects<\/em>\n<ul><li><a href=\"http:\/\/www.arclanguage.org\/\">Arc<\/a>\/<a href=\"https:\/\/github.com\/arclanguage\/anarki\">Anarki<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/carp-lang\/carp\">Carp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/LuxLang\/lux\">Lux<\/a><\/li>\n<li><a href=\"http:\/\/pixielang.org\/\">Pixie<\/a><\/li>\n<li><a href=\"http:\/\/www.toccata.io\/\">Toccata<\/a><\/li>\n<li><a href=\"http:\/\/www.nongnu.org\/txr\/txr-lisp.html\">TXR<\/a><\/li>\n<li><a href=\"http:\/\/www.ulisp.com\/\">uLisp<\/a> (Arduino)<\/li>\n<\/ul><\/li>\n<li><em>Exo-Lisps<\/em>\n<ul><li><a href=\"https:\/\/clojurescript.org\/\">ClojureScript<\/a> (Clojure\u2192Javascript)<\/li>\n<li><a href=\"https:\/\/github.com\/bakpakin\/Fennel\">Fennel<\/a> (Lua)<\/li>\n<li><a href=\"http:\/\/docs.hylang.org\/en\/stable\/\">Hy<\/a> (Python)<\/li>\n<li><a href=\"https:\/\/github.com\/haskell-lisp\/liskell\">Liskell<\/a> (Haskell)<\/li>\n<li><a href=\"https:\/\/github.com\/rvirding\/lfe\">Lisp Flavored Erlang<\/a><\/li>\n<li><a href=\"http:\/\/lush.sourceforge.net\/\">Lush<\/a> (numerical\/graphical)<\/li>\n<li><a href=\"https:\/\/common-lisp.net\/project\/parenscript\/\">Parenscript<\/a> (CL\u2192Javascript)<\/li>\n<li><a href=\"http:\/\/urn-lang.com\/\">Urn<\/a> (Lua)<\/li>\n<\/ul><\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2019-01-07T04:13:16+0000","updated_at":"2019-01-07T04:13:16+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.reddit.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":767,"title":"Hacker's guide to Neural Networks","url":"http:\/\/karpathy.github.io\/neuralnets\/","content":"<p>Hi there, I\u2019m a <a href=\"http:\/\/cs.stanford.edu\/people\/karpathy\/\">CS PhD student at Stanford<\/a>. I\u2019ve worked on Deep Learning for a few years as part of my research and among several of my related pet projects is <a href=\"http:\/\/convnetjs.com\">ConvNetJS<\/a> - a Javascript library for training Neural Networks. Javascript allows one to nicely visualize what\u2019s going on and to play around with the various hyperparameter settings, but I still regularly hear from people who ask for a more thorough treatment of the topic. This article (which I plan to slowly expand out to lengths of a few book chapters) is my humble attempt. It\u2019s on web instead of PDF because all books should be, and eventually it will hopefully include animations\/demos etc.<\/p><p>My personal experience with Neural Networks is that everything became much clearer when I started ignoring full-page, dense derivations of backpropagation equations and just started writing code. Thus, this tutorial will contain <strong>very little math<\/strong> (I don\u2019t believe it is necessary and it can sometimes even obfuscate simple concepts). Since my background is in Computer Science and Physics, I will instead develop the topic from what I refer to as <strong>hackers\u2019s perspective<\/strong>. My exposition will center around code and physical intuitions instead of mathematical derivations. Basically, I will strive to present the algorithms in a way that I wish I had come across when I was starting out.<\/p><blockquote>\n  <p>\u201c\u2026everything became much clearer when I started writing code.\u201d<\/p>\n<\/blockquote><p>You might be eager to jump right in and learn about Neural Networks, backpropagation, how they can be applied to datasets in practice, etc. But before we get there, I\u2019d like us to first forget about all that. Let\u2019s take a step back and understand what is really going on at the core. Lets first talk about real-valued circuits.<\/p><p><em>Update note<\/em>: I suspended my work on this guide a while ago and redirected a lot of my energy to teaching CS231n (Convolutional Neural Networks) class at Stanford. The notes are on <a href=\"http:\/\/cs231n.github.io\">cs231.github.io<\/a> and the course slides can be found <a href=\"http:\/\/cs231n.stanford.edu\/syllabus.html\">here<\/a>. These materials are highly related to material here, but more comprehensive and sometimes more polished.<\/p><h2 id=\"chapter-1-real-valued-circuits\">Chapter 1: Real-valued Circuits<\/h2><p>In my opinion, the best way to think of Neural Networks is as real-valued circuits, where real values (instead of boolean values <code class=\"highlighter-rouge\">{0,1}<\/code>) \u201cflow\u201d along edges and interact in gates. However, instead of gates such as <code class=\"highlighter-rouge\">AND<\/code>, <code class=\"highlighter-rouge\">OR<\/code>, <code class=\"highlighter-rouge\">NOT<\/code>, etc, we have binary gates such as <code class=\"highlighter-rouge\">*<\/code> (multiply), <code class=\"highlighter-rouge\">+<\/code> (add), <code class=\"highlighter-rouge\">max<\/code> or unary gates such as <code class=\"highlighter-rouge\">exp<\/code>, etc. Unlike ordinary boolean circuits, however, we will eventually also have <strong>gradients<\/strong> flowing on the same edges of the circuit, but in the opposite direction. But we\u2019re getting ahead of ourselves. Let\u2019s focus and start out simple.<\/p><h3 id=\"base-case-single-gate-in-the-circuit\">Base Case: Single Gate in the Circuit<\/h3><p>Lets first consider a single, simple circuit with one gate. Here\u2019s an example:<\/p><p>xy*<\/p><p>The circuit takes two real-valued inputs <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> and computes <code class=\"highlighter-rouge\">x * y<\/code> with the <code class=\"highlighter-rouge\">*<\/code> gate. Javascript version of this would very simply look something like this:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var forwardMultiplyGate = function(x, y) {\n  return x * y;\n};\nforwardMultiplyGate(-2, 3); \/\/ returns -6. Exciting.\n<\/pre><\/div><\/div><p>And in math form we can think of this gate as implementing the real-valued function:<\/p><p>As with this example, all of our gates will take one or two inputs and produce a <strong>single<\/strong> output value.<\/p><h4 id=\"the-goal\">The Goal<\/h4><p>The problem we are interested in studying looks as follows:<\/p><ol><li>We provide a given circuit some specific input values (e.g. <code class=\"highlighter-rouge\">x = -2<\/code>, <code class=\"highlighter-rouge\">y = 3<\/code>)<\/li>\n  <li>The circuit computes an output value (e.g. <code class=\"highlighter-rouge\">-6<\/code>)<\/li>\n  <li>The core question then becomes: <em>How should one tweak the input slightly to increase the output?<\/em><\/li>\n<\/ol><p>In this case, in what direction should we change <code class=\"highlighter-rouge\">x,y<\/code> to get a number larger than <code class=\"highlighter-rouge\">-6<\/code>? Note that, for example, <code class=\"highlighter-rouge\">x = -1.99<\/code> and <code class=\"highlighter-rouge\">y = 2.99<\/code> gives <code class=\"highlighter-rouge\">x * y = -5.95<\/code>, which is higher than <code class=\"highlighter-rouge\">-6.0<\/code>. Don\u2019t get confused by this: <code class=\"highlighter-rouge\">-5.95<\/code> is better (higher) than <code class=\"highlighter-rouge\">-6.0<\/code>. It\u2019s an improvement of <code class=\"highlighter-rouge\">0.05<\/code>, even though the <em>magnitude<\/em> of <code class=\"highlighter-rouge\">-5.95<\/code> (the distance from zero) happens to be lower.<\/p><p>Okay. So wait, we have a circuit, we have some inputs and we just want to tweak them slightly to increase the output value? Why is this hard? We can easily \u201cforward\u201d the circuit to compute the output for any given <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>. So isn\u2019t this trivial? Why don\u2019t we tweak <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> randomly and keep track of the tweak that works best:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ circuit with single gate for now\nvar forwardMultiplyGate = function(x, y) { return x * y; };\nvar x = -2, y = 3; \/\/ some input values\n\/\/ try changing x,y randomly small amounts and keep track of what works best\nvar tweak_amount = 0.01;\nvar best_out = -Infinity;\nvar best_x = x, best_y = y;\nfor(var k = 0; k &lt; 100; k++) {\n  var x_try = x + tweak_amount * (Math.random() * 2 - 1); \/\/ tweak x a bit\n  var y_try = y + tweak_amount * (Math.random() * 2 - 1); \/\/ tweak y a bit\n  var out = forwardMultiplyGate(x_try, y_try);\n  if(out &gt; best_out) {\n    \/\/ best improvement yet! Keep track of the x and y\n    best_out = out; \n    best_x = x_try, best_y = y_try;\n  }\n}\n<\/pre><\/div><\/div><p>When I run this, I get <code class=\"highlighter-rouge\">best_x = -1.9928<\/code>, <code class=\"highlighter-rouge\">best_y = 2.9901<\/code>, and <code class=\"highlighter-rouge\">best_out = -5.9588<\/code>. Again, <code class=\"highlighter-rouge\">-5.9588<\/code> is higher than <code class=\"highlighter-rouge\">-6.0<\/code>. So, we\u2019re done, right? Not quite: This is a perfectly fine strategy for tiny problems with a few gates if you can afford the compute time, but it won\u2019t do if we want to eventually consider huge circuits with millions of inputs. It turns out that we can do much better.<\/p><h4 id=\"stategy-2-numerical-gradient\">Stategy #2: Numerical Gradient<\/h4><p>Here\u2019s a better way. Remember again that in our setup we are given a circuit (e.g. our circuit with a single <code class=\"highlighter-rouge\">*<\/code> gate) and some particular input (e.g. <code class=\"highlighter-rouge\">x = -2, y = 3<\/code>). The gate computes the output (<code class=\"highlighter-rouge\">-6<\/code>) and now we\u2019d like to tweak <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> to make the output higher.<\/p><p>A nice intuition for what we\u2019re about to do is as follows: Imagine taking the output value that comes out from the circuit and tugging on it in the positive direction. This positive tension will in turn translate through the gate and induce forces on the inputs <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>. Forces that tell us how <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> should change to increase the output value.<\/p><p>What might those forces look like in our specific example? Thinking through it, we can intuit that the force on <code class=\"highlighter-rouge\">x<\/code> should also be positive, because making <code class=\"highlighter-rouge\">x<\/code> slightly larger improves the circuit\u2019s output. For example, increasing <code class=\"highlighter-rouge\">x<\/code> from <code class=\"highlighter-rouge\">x = -2<\/code> to <code class=\"highlighter-rouge\">x = -1<\/code> would give us output <code class=\"highlighter-rouge\">-3<\/code> - much larger than <code class=\"highlighter-rouge\">-6<\/code>. On the other hand, we\u2019d expect a negative force induced on <code class=\"highlighter-rouge\">y<\/code> that pushes it to become lower (since a lower <code class=\"highlighter-rouge\">y<\/code>, such as <code class=\"highlighter-rouge\">y = 2<\/code>, down from the original <code class=\"highlighter-rouge\">y = 3<\/code> would make output higher: <code class=\"highlighter-rouge\">2 x -2 = -4<\/code>, again, larger than <code class=\"highlighter-rouge\">-6<\/code>). That\u2019s the intuition to keep in mind, anyway. As we go through this, it will turn out that forces I\u2019m describing will in fact turn out to be the <strong>derivative<\/strong> of the output value with respect to its inputs (<code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>). You may have heard this term before.<\/p><blockquote>\n  <p>The derivative can be thought of as a force on each input as we pull on the output to become higher.<\/p>\n<\/blockquote><p>So how do we exactly evaluate this force (derivative)? It turns out that there is a very simple procedure for this. We will work backwards: Instead of pulling on the circuit\u2019s output, we\u2019ll iterate over every input one by one, increase it very slightly and look at what happens to the output value. The amount the output changes in response is the derivative. Enough intuitions for now. Lets look at the mathematical definition. We can write down the derivative for our function with respect to the inputs. For example, the derivative with respect to <code class=\"highlighter-rouge\">x<\/code> can be computed as:<\/p><p>$$\n\\frac{\\partial f(x,y)}{\\partial x} = \\frac{f(x+h,y) - f(x,y)}{h}\n$$<\/p><p>Where \\( h \\) is small - it\u2019s the tweak amount. Also, if you\u2019re not very familiar with calculus it is important to note that in the left-hand side of the equation above, the horizontal line does <em>not<\/em> indicate division. The entire symbol \\( \\frac{\\partial f(x,y)}{\\partial x} \\) is a single thing: the derivative of the function \\( f(x,y) \\) with respect to \\( x \\). The horizontal line on the right <em>is<\/em> division. I know it\u2019s confusing but it\u2019s standard notation. Anyway, I hope it doesn\u2019t look too scary because it isn\u2019t: The circuit was giving some initial output \\( f(x,y) \\), and then we changed one of the inputs by a tiny amount \\(h \\) and read the new output \\( f(x+h, y) \\). Subtracting those two quantities tells us the change, and the division by \\(h \\) just normalizes this change by the (arbitrary) tweak amount we used. In other words it\u2019s expressing exactly what I described above and translates directly to this code:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = -2, y = 3;\nvar out = forwardMultiplyGate(x, y); \/\/ -6\nvar h = 0.0001;\n\/\/ compute derivative with respect to x\nvar xph = x + h; \/\/ -1.9999\nvar out2 = forwardMultiplyGate(xph, y); \/\/ -5.9997\nvar x_derivative = (out2 - out) \/ h; \/\/ 3.0\n\/\/ compute derivative with respect to y\nvar yph = y + h; \/\/ 3.0001\nvar out3 = forwardMultiplyGate(x, yph); \/\/ -6.0002\nvar y_derivative = (out3 - out) \/ h; \/\/ -2.0\n<\/pre><\/div><\/div><p>Lets walk through <code class=\"highlighter-rouge\">x<\/code> for example. We turned the knob from <code class=\"highlighter-rouge\">x<\/code> to <code class=\"highlighter-rouge\">x + h<\/code> and the circuit responded by giving a higher value (note again that yes, <code class=\"highlighter-rouge\">-5.9997<\/code> is <em>higher<\/em> than <code class=\"highlighter-rouge\">-6<\/code>: <code class=\"highlighter-rouge\">-5.9997 &gt; -6<\/code>). The division by <code class=\"highlighter-rouge\">h<\/code> is there to normalize the circuit\u2019s response by the (arbitrary) value of <code class=\"highlighter-rouge\">h<\/code> we chose to use here. Technically, you want the value of <code class=\"highlighter-rouge\">h<\/code> to be infinitesimal (the precise mathematical definition of the gradient is defined as the limit of the expression as <code class=\"highlighter-rouge\">h<\/code> goes to zero), but in practice <code class=\"highlighter-rouge\">h=0.00001<\/code> or so works fine in most cases to get a good approximation. Now, we see that the derivative w.r.t. <code class=\"highlighter-rouge\">x<\/code> is <code class=\"highlighter-rouge\">+3<\/code>. I\u2019m making the positive sign explicit, because it indicates that the circuit is tugging on x to become higher. The actual value, <code class=\"highlighter-rouge\">3<\/code> can be interpreted as the <em>force<\/em> of that tug.<\/p><blockquote>\n  <p>The derivative with respect to some input can be computed by tweaking that input by a small amount and observing the change on the output value.<\/p>\n<\/blockquote><p>By the way, we usually  talk about the <em>derivative<\/em> with respect to a single input, or about a <strong>gradient<\/strong> with respect to all the inputs. The gradient is just made up of the derivatives of all the inputs concatenated in a vector (i.e. a list). Crucially, notice that if we let the inputs respond to the tug by following the gradient a tiny amount (i.e. we just add the derivative on top of every input), we can see that the value increases, as expected:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var step_size = 0.01;\nvar out = forwardMultiplyGate(x, y); \/\/ before: -6\nx = x + step_size * x_derivative; \/\/ x becomes -1.97\ny = y + step_size * y_derivative; \/\/ y becomes 2.98\nvar out_new = forwardMultiplyGate(x, y); \/\/ -5.87! exciting.\n<\/pre><\/div><\/div><p>As expected, we changed the inputs by the gradient and the circuit now gives a slightly higher value (<code class=\"highlighter-rouge\">-5.87 &gt; -6.0<\/code>). That was much simpler than trying random changes to <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>, right? A fact to appreciate here is that if you take calculus you can prove that the gradient is, in fact, the direction of the steepest increase of the function. There is no need to monkey around trying out random pertubations as done in Strategy #1. Evaluating the gradient requires just three evaluations of the forward pass of our circuit instead of hundreds, and gives the best tug you can hope for (locally) if you are interested in increasing the value of the output.<\/p><p><strong>Bigger step is not always better.<\/strong> Let me clarify on this point a bit. It is important to note that in this very simple example, using a bigger <code class=\"highlighter-rouge\">step_size<\/code> than 0.01  will always work better. For example, <code class=\"highlighter-rouge\">step_size = 1.0<\/code> gives output <code class=\"highlighter-rouge\">-1<\/code> (higer, better!), and indeed infinite step size would give infinitely good results. The crucial thing to realize is that once our circuits get much more complex (e.g. entire neural networks), the function from inputs to the output value will be more chaotic and wiggly. The gradient guarantees that if you have a very small (indeed, infinitesimally small) step size, then you will definitely get a higher number when you follow its direction, and for that infinitesimally small step size there is no other direction that would have worked better. But if you use a bigger step size (e.g. <code class=\"highlighter-rouge\">step_size = 0.01<\/code>) all bets are off. The reason we can get away with a larger step size than infinitesimally small is that our functions are usually relatively smooth. But really, we\u2019re crossing our fingers and hoping for the best.<\/p><p><strong>Hill-climbing analogy.<\/strong> One analogy I\u2019ve heard before is that the output value of our circut is like the height of a hill, and we are blindfolded and trying to climb upwards. We can sense the steepness of the hill at our feet (the gradient), so when we shuffle our feet a bit we will go upwards. But if we took a big, overconfident step, we could have stepped right into a hole.<\/p><p>Great, I hope I\u2019ve convinced you that the numerical gradient is indeed a very useful thing to evaluate, and that it is cheap. But. It turns out that we can do <em>even<\/em> better.<\/p><h4 id=\"strategy-3-analytic-gradient\">Strategy #3: Analytic Gradient<\/h4><p>In the previous section we evaluated the gradient by probing the circuit\u2019s output value, independently for every input. This procedure gives you what we call a <strong>numerical gradient<\/strong>. This approach, however, is <em>still<\/em> expensive because we need to compute the circuit\u2019s output as we tweak every input value independently a small amount. So the complexity of evaluating the gradient is linear in number of inputs. But in practice we will have hundreds, thousands or (for neural networks) even tens to hundreds of millions of inputs, and the circuits aren\u2019t just one multiply gate but huge expressions that can be expensive to compute. We need something better.<\/p><p>Luckily, there is an easier and <em>much<\/em> faster way to compute the gradient: we can use calculus to derive a direct expression for it that will be as simple to evaluate as the circuit\u2019s output value. We call this an <strong>analytic gradient<\/strong> and there will be no need for tweaking anything. You may have seen other people who teach Neural Networks derive the gradient in huge and, frankly, scary and confusing mathematical equations (if you\u2019re not well-versed in maths). But it\u2019s unnecessary. I\u2019ve written plenty of Neural Nets code and I rarely have to do mathematical derivation longer than two lines, and 95% of the time it can be done without writing anything at all. That is because we will only ever derive the gradient for very small and simple expressions (think of it as the <strong>base case<\/strong>) and then I will show you how we can compose these very simply with <strong>chain rule<\/strong> to evaluate the full gradient (think inductive\/recursive case).<\/p><blockquote>\n  <p>The analytic derivative requires no tweaking of the inputs. It can be derived using mathematics (calculus).<\/p>\n<\/blockquote><p>If you remember your product rules, power rules, quotient rules, etc. (see e.g. <a href=\"http:\/\/www.mathsisfun.com\/calculus\/derivatives-rules.html\">derivative rules<\/a> or <a href=\"http:\/\/en.wikipedia.org\/wiki\/Differentiation_rules\">wiki page<\/a>), it\u2019s very easy to write down the derivitative with respect to both <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> for a small expression such as <code class=\"highlighter-rouge\">x * y<\/code>. But suppose you don\u2019t remember your calculus rules. We can go back to the definition. For example, here\u2019s the expression for the derivative w.r.t <code class=\"highlighter-rouge\">x<\/code>:<\/p><p>$$\n\\frac{\\partial f(x,y)}{\\partial x} = \\frac{f(x+h,y) - f(x,y)}{h}\n$$<\/p><p>(Technically I\u2019m not writing the limit as <code class=\"highlighter-rouge\">h<\/code> goes to zero, forgive me math people). Okay and lets plug in our function ( \\( f(x,y) = x y \\) ) into the expression. Ready for the hardest piece of math of this entire article? Here we go:<\/p><p>$$\n\\frac{\\partial f(x,y)}{\\partial x} = \\frac{f(x+h,y) - f(x,y)}{h}\n= \\frac{(x+h)y - xy}{h}\n= \\frac{xy + hy - xy}{h}\n= \\frac{hy}{h}\n= y\n$$<\/p><p>That\u2019s interesting. The derivative with respect to <code class=\"highlighter-rouge\">x<\/code> is just equal to <code class=\"highlighter-rouge\">y<\/code>. Did you notice the coincidence in the previous section? We tweaked <code class=\"highlighter-rouge\">x<\/code> to <code class=\"highlighter-rouge\">x+h<\/code> and calculated <code class=\"highlighter-rouge\">x_derivative = 3.0<\/code>, which exactly happens to be the value of <code class=\"highlighter-rouge\">y<\/code> in that example. It turns out that wasn\u2019t a coincidence at all because that\u2019s just what the analytic gradient tells us the <code class=\"highlighter-rouge\">x<\/code> derivative should be for <code class=\"highlighter-rouge\">f(x,y) = x * y<\/code>. The derivative with respect to <code class=\"highlighter-rouge\">y<\/code>, by the way, turns out to be <code class=\"highlighter-rouge\">x<\/code>, unsurprisingly by symmetry. So there is no need for any tweaking! We invoked powerful mathematics and can now transform our derivative calculation into the following code:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = -2, y = 3;\nvar out = forwardMultiplyGate(x, y); \/\/ before: -6\nvar x_gradient = y; \/\/ by our complex mathematical derivation above\nvar y_gradient = x;\nvar step_size = 0.01;\nx += step_size * x_gradient; \/\/ -1.97\ny += step_size * y_gradient; \/\/ 2.98\nvar out_new = forwardMultiplyGate(x, y); \/\/ -5.87. Higher output! Nice.\n<\/pre><\/div><\/div><p>To compute the gradient we went from forwarding the circuit hundreds of times (Strategy #1) to forwarding it only on order of number of times twice the number of inputs (Strategy #2), to forwarding it a single time! And it gets EVEN better, since the more expensive strategies (#1 and #2) only give an approximation of the gradient, while #3 (the fastest one by far) gives you the <em>exact<\/em> gradient. No approximations. The only downside is that you should be comfortable with some calculus 101.<\/p><p>Lets recap what we have learned:<\/p><ul><li>INPUT: We are given a circuit, some inputs and compute an output value.<\/li>\n  <li>OUTPUT: We are then interested finding small changes to each input (independently) that would make the output higher.<\/li>\n  <li>Strategy #1: One silly way is to <strong>randomly search<\/strong> for small pertubations of the inputs and keep track of what gives the highest increase in output.<\/li>\n  <li>Strategy #2: We saw we can do much better by computing the gradient. Regardless of how complicated the circuit is, the <strong>numerical gradient<\/strong> is very simple (but relatively expensive) to compute. We compute it by <em>probing<\/em> the circuit\u2019s output value as we tweak the inputs one at a time.<\/li>\n  <li>Strategy #3: In the end, we saw that we can be even more clever and analytically derive a direct expression to get the <strong>analytic gradient<\/strong>. It is identical to the numerical gradient, it is fastest by far, and there is no need for any tweaking.<\/li>\n<\/ul><p>In practice by the way (and we will get to this once again later), all Neural Network libraries always compute the analytic gradient, but the correctness of the implementation is verified by comparing it to the numerical gradient. That\u2019s because the numerical gradient is very easy to evaluate (but can be a bit expensive to compute), while the analytic gradient can contain bugs at times, but is usually extremely efficient to compute. As we will see, evaluating the gradient (i.e. while doing <em>backprop<\/em>, or <em>backward pass<\/em>) will turn out to cost about as much as evaluating the <em>forward pass<\/em>.<\/p><h3 id=\"recursive-case-circuits-with-multiple-gates\">Recursive Case: Circuits with Multiple Gates<\/h3><p>But hold on, you say: <em>\u201cThe analytic gradient was trivial to derive for your super-simple expression. This is useless. What do I do when the expressions are much larger? Don\u2019t the equations get huge and complex very fast?\u201d<\/em>. Good question. Yes the expressions get much more complex. No, this doesn\u2019t make it much harder. As we will see, every gate will be hanging out by itself, completely unaware of any details of the huge and complex circuit that it could be part of. It will only worry about its inputs and it will compute its local derivatives as seen in the previous section, except now there will be a single extra multiplication it will have to do.<\/p><blockquote>\n  <p>A single extra multiplication will turn a single (useless gate) into a cog in the complex machine that is an entire neural network.<\/p>\n<\/blockquote><p>I should stop hyping it up now. I hope I\u2019ve piqued your interest! Lets drill down into details and get two gates involved with this next example:<\/p><p>xyz+q*f<\/p><p>The expression we are computing now is \\( f(x,y,z) = (x + y) z \\). Lets structure the code as follows to make the gates explicit as functions:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var forwardMultiplyGate = function(a, b) { \n  return a * b;\n};\nvar forwardAddGate = function(a, b) { \n  return a + b;\n};\nvar forwardCircuit = function(x,y,z) { \n  var q = forwardAddGate(x, y);\n  var f = forwardMultiplyGate(q, z);\n  return f;\n};\nvar x = -2, y = 5, z = -4;\nvar f = forwardCircuit(x, y, z); \/\/ output is -12\n<\/pre><\/div><\/div><p>In the above, I am using <code class=\"highlighter-rouge\">a<\/code> and <code class=\"highlighter-rouge\">b<\/code> as the local variables in the gate functions so that we don\u2019t get these confused with our circuit inputs <code class=\"highlighter-rouge\">x,y,z<\/code>. As before, we are interested in finding the derivatives with respect to the three inputs <code class=\"highlighter-rouge\">x,y,z<\/code>. But how do we compute it now that there are multiple gates involved? First, lets pretend that the <code class=\"highlighter-rouge\">+<\/code> gate is not there and that we only have two variables in the circuit: <code class=\"highlighter-rouge\">q,z<\/code> and a single <code class=\"highlighter-rouge\">*<\/code> gate. Note that the <code class=\"highlighter-rouge\">q<\/code> is is output of the <code class=\"highlighter-rouge\">+<\/code> gate. If we don\u2019t worry about <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> but only about <code class=\"highlighter-rouge\">q<\/code> and <code class=\"highlighter-rouge\">z<\/code>, then we are back to having only a single gate, and as far as that single <code class=\"highlighter-rouge\">*<\/code> gate is concerned, we know what the (analytic) derivates are from previous section. We can write them down (except here we\u2019re replacing <code class=\"highlighter-rouge\">x,y<\/code> with <code class=\"highlighter-rouge\">q,z<\/code>):<\/p><p>Simple enough: these are the expressions for the gradient with respect to <code class=\"highlighter-rouge\">q<\/code> and <code class=\"highlighter-rouge\">z<\/code>. But wait, we don\u2019t want gradient with respect to <code class=\"highlighter-rouge\">q<\/code>, but with respect to the inputs: <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>. Luckily, <code class=\"highlighter-rouge\">q<\/code> is computed as a function of <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> (by addition in our example). We can write down the gradient for the addition gate as well, it\u2019s even simpler:<\/p><p>That\u2019s right, the derivatives are just 1, regardless of the actual values of <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>. If you think about it, this makes sense because to make the output of a single addition gate higher, we expect a positive tug on both <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>, regardless of their values.<\/p><h4 id=\"backpropagation\">Backpropagation<\/h4><p>We are finally ready to invoke the <strong>Chain Rule<\/strong>: We know how to compute the gradient of <code class=\"highlighter-rouge\">q<\/code> with respect to <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> (that\u2019s a single gate case with <code class=\"highlighter-rouge\">+<\/code> as the gate). And we know how to compute the gradient of our final output with respect to <code class=\"highlighter-rouge\">q<\/code>. The chain rule tells us how to combine these to get the gradient of the final output with respect to <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code>, which is what we\u2019re ultimately interested in. Best of all, the chain rule very simply states that the right thing to do is to simply multiply the gradients together to chain them. For example, the final derivative for <code class=\"highlighter-rouge\">x<\/code> will be:<\/p><p>There are many symbols there so maybe this is confusing again, but it\u2019s really just two numbers being multiplied together. Here is the code:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ initial conditions\nvar x = -2, y = 5, z = -4;\nvar q = forwardAddGate(x, y); \/\/ q is 3\nvar f = forwardMultiplyGate(q, z); \/\/ output is -12\n\/\/ gradient of the MULTIPLY gate with respect to its inputs\n\/\/ wrt is short for \"with respect to\"\nvar derivative_f_wrt_z = q; \/\/ 3\nvar derivative_f_wrt_q = z; \/\/ -4\n\/\/ derivative of the ADD gate with respect to its inputs\nvar derivative_q_wrt_x = 1.0;\nvar derivative_q_wrt_y = 1.0;\n\/\/ chain rule\nvar derivative_f_wrt_x = derivative_q_wrt_x * derivative_f_wrt_q; \/\/ -4\nvar derivative_f_wrt_y = derivative_q_wrt_y * derivative_f_wrt_q; \/\/ -4\n<\/pre><\/div><\/div><p>That\u2019s it. We computed the gradient (the forces) and now we can let our inputs respond to it by a bit. Lets add the gradients on top of the inputs. The output value of the circuit better increase, up from -12!<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ final gradient, from above: [-4, -4, 3]\nvar gradient_f_wrt_xyz = [derivative_f_wrt_x, derivative_f_wrt_y, derivative_f_wrt_z]\n\/\/ let the inputs respond to the force\/tug:\nvar step_size = 0.01;\nx = x + step_size * derivative_f_wrt_x; \/\/ -2.04\ny = y + step_size * derivative_f_wrt_y; \/\/ 4.96\nz = z + step_size * derivative_f_wrt_z; \/\/ -3.97\n\/\/ Our circuit now better give higher output:\nvar q = forwardAddGate(x, y); \/\/ q becomes 2.92\nvar f = forwardMultiplyGate(q, z); \/\/ output is -11.59, up from -12! Nice!\n<\/pre><\/div><\/div><p>Looks like that worked! Lets now try to interpret intuitively what just happened. The circuit wants to output higher values. The last gate saw inputs <code class=\"highlighter-rouge\">q = 3, z = -4<\/code> and computed output <code class=\"highlighter-rouge\">-12<\/code>. \u201cPulling\u201d upwards on this output value induced a force on both <code class=\"highlighter-rouge\">q<\/code> and <code class=\"highlighter-rouge\">z<\/code>: To increase the output value, the circuit \u201cwants\u201d <code class=\"highlighter-rouge\">z<\/code> to increase, as can be seen by the positive value of the derivative(<code class=\"highlighter-rouge\">derivative_f_wrt_z = +3<\/code>). Again, the size of this derivative can be interpreted as the magnitude of the force. On the other hand, <code class=\"highlighter-rouge\">q<\/code> felt a stronger and downward force, since <code class=\"highlighter-rouge\">derivative_f_wrt_q = -4<\/code>. In other words the circuit wants <code class=\"highlighter-rouge\">q<\/code> to decrease, with a force of <code class=\"highlighter-rouge\">4<\/code>.<\/p><p>Now we get to the second, <code class=\"highlighter-rouge\">+<\/code> gate which outputs <code class=\"highlighter-rouge\">q<\/code>. By default, the <code class=\"highlighter-rouge\">+<\/code> gate computes its derivatives which tells us how to change <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> to make <code class=\"highlighter-rouge\">q<\/code> higher. BUT! Here is the <strong>crucial point<\/strong>: the gradient on <code class=\"highlighter-rouge\">q<\/code> was computed as negative (<code class=\"highlighter-rouge\">derivative_f_wrt_q = -4<\/code>), so the circuit wants <code class=\"highlighter-rouge\">q<\/code> to <em>decrease<\/em>, and with a force of <code class=\"highlighter-rouge\">4<\/code>! So if the <code class=\"highlighter-rouge\">+<\/code> gate wants to contribute to making the final output value larger, it needs to listen to the gradient signal coming from the top. In this particular case, it needs to apply tugs on <code class=\"highlighter-rouge\">x,y<\/code> opposite of what it would normally apply, and with a force of <code class=\"highlighter-rouge\">4<\/code>, so to speak. The multiplication by <code class=\"highlighter-rouge\">-4<\/code> seen in the chain rule achieves exactly this: instead of applying a positive force of <code class=\"highlighter-rouge\">+1<\/code> on both <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> (the local derivative), the full circuit\u2019s gradient on both <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> becomes <code class=\"highlighter-rouge\">1 x -4 = -4<\/code>. This makes sense: the circuit wants both <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> to get smaller because this will make <code class=\"highlighter-rouge\">q<\/code> smaller, which in turn will make <code class=\"highlighter-rouge\">f<\/code> larger.<\/p><blockquote>\n  <p>If this makes sense, you understand backpropagation.<\/p>\n<\/blockquote><p>Lets <strong>recap<\/strong> once again what we learned:<\/p><ul><li>\n    <p>In the previous chapter we saw that in the case of a single gate (or a single expression), we can derive the analytic gradient using simple calculus. We interpreted the gradient as a force, or a tug on the inputs that pulls them in a direction which would make this gate\u2019s output higher.<\/p>\n  <\/li>\n  <li>\n    <p>In case of multiple gates everything stays pretty much the same way: every gate is hanging out by itself completely unaware of the circuit it is embedded in. Some inputs come in and the gate computes its output and the derivate with respect to the inputs. The <em>only<\/em> difference now is that suddenly, something can pull on this gate from above. That\u2019s the gradient of the final circuit output value with respect to the ouput this gate computed. It is the circuit asking the gate to output higher or lower numbers, and with some force. The gate simply takes this force and multiplies it to all the forces it computed for its inputs before (chain rule). This has the desired effect:<\/p>\n  <\/li>\n<\/ul><ol><li>If a gate experiences a strong positive pull from above, it will also pull harder on its own inputs, scaled by the force it is experiencing from above<\/li>\n  <li>And if it experiences a negative tug, this means that circuit wants its value to decrease not increase, so it will flip the force of the pull on its inputs to make its own output value smaller.<\/li>\n<\/ol><blockquote>\n  <p>A nice picture to have in mind is that as we pull on the circuit\u2019s output value at the end, this induces pulls downward through the entire circuit, all the way down to the inputs.<\/p>\n<\/blockquote><p>Isn\u2019t it beautiful? The only difference between the case of a single gate and multiple interacting gates that compute arbitrarily complex expressions is this additional multipy operation that now happens in each gate.<\/p><h4 id=\"patterns-in-the-backward-flow\">Patterns in the \u201cbackward\u201d flow<\/h4><p>Lets look again at our example circuit with the numbers filled in. The first circuit shows the raw values, and the second circuit shows the gradients that flow back to the inputs as discussed. Notice that the gradient always starts off with <code class=\"highlighter-rouge\">+1<\/code> at the end to start off the chain. This is the (default) pull on the circuit to have its value increased.<\/p><p>(Values)-25-4+3*-12(Gradients)-4-43+-4*1<\/p><p>After a while you start to notice patterns in how the gradients flow backward in the circuits. For example, the <code class=\"highlighter-rouge\">+<\/code> gate always takes the gradient on top and simply passes it on to all of its inputs (notice the example with -4 simply passed on to both of the inputs of <code class=\"highlighter-rouge\">+<\/code> gate). This is because its own derivative for the inputs is just <code class=\"highlighter-rouge\">+1<\/code>, regardless of what the actual values of the inputs are, so in the chain rule, the gradient from above is just multiplied by 1 and stays the same. Similar intuitions apply to, for example, a <code class=\"highlighter-rouge\">max(x,y)<\/code> gate. Since the gradient of <code class=\"highlighter-rouge\">max(x,y)<\/code> with respect to its input is <code class=\"highlighter-rouge\">+1<\/code> for whichever one of <code class=\"highlighter-rouge\">x<\/code>, <code class=\"highlighter-rouge\">y<\/code> is larger and <code class=\"highlighter-rouge\">0<\/code> for the other, this gate is during backprop effectively just a gradient \u201cswitch\u201d: it will take the gradient from above and \u201croute\u201d it to the input that had a higher value during the forward pass.<\/p><p><strong>Numerical Gradient Check.<\/strong> Before we finish with this section, lets just make sure that the (analytic) gradient we computed by backprop above is correct as a sanity check. Remember that we can do this simply by computing the numerical gradient and making sure that we get <code class=\"highlighter-rouge\">[-4, -4, 3]<\/code> for <code class=\"highlighter-rouge\">x,y,z<\/code>. Here\u2019s the code:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ initial conditions\nvar x = -2, y = 5, z = -4;\n\/\/ numerical gradient check\nvar h = 0.0001;\nvar x_derivative = (forwardCircuit(x+h,y,z) - forwardCircuit(x,y,z)) \/ h; \/\/ -4\nvar y_derivative = (forwardCircuit(x,y+h,z) - forwardCircuit(x,y,z)) \/ h; \/\/ -4\nvar z_derivative = (forwardCircuit(x,y,z+h) - forwardCircuit(x,y,z)) \/ h; \/\/ 3\n<\/pre><\/div><\/div><p>and we get <code class=\"highlighter-rouge\">[-4, -4, 3]<\/code>, as computed with backprop. phew! :)<\/p><h3 id=\"example-single-neuron\">Example: Single Neuron<\/h3><p>In the previous section you hopefully got the basic intuition behind backpropagation. Lets now look at an even more complicated and borderline practical example. We will consider a 2-dimensional neuron that computes the following function:<\/p><p>In this expression, \\( \\sigma \\) is the <em>sigmoid<\/em> function. Its best thought of as a \u201csquashing function\u201d, because it takes the input and squashes it to be between zero and one: Very negative values are squashed towards zero and positive values get squashed towards one. For example, we have <code class=\"highlighter-rouge\">sig(-5) = 0.006, sig(0) = 0.5, sig(5) = 0.993<\/code>. Sigmoid function is defined as:<\/p><p>The gradient with respect to its single input, as you can check on Wikipedia or derive yourself if you know some calculus is given by this expression:<\/p><p>For example, if the input to the sigmoid gate is <code class=\"highlighter-rouge\">x = 3<\/code>, the gate will compute output <code class=\"highlighter-rouge\">f = 1.0 \/ (1.0 + Math.exp(-x)) = 0.95<\/code>, and then the (local) gradient on its input will simply be <code class=\"highlighter-rouge\">dx = (0.95) * (1 - 0.95) = 0.0475<\/code>.<\/p><p>That\u2019s all we need to use this gate: we know how to take an input and <em>forward<\/em> it through the sigmoid gate, and we also have the expression for the gradient with respect to its input, so we can also <em>backprop<\/em> through it. Another thing to note is that technically, the sigmoid function is made up of an entire series of gates in a line that compute more <em>atomic<\/em> functions: an exponentiation gate, an addition gate and a division gate. Treating it so would work perfectly fine but for this example I chose to collapse all of these gates into a single gate that just computes sigmoid in one shot, because the gradient expression turns out to be simple.<\/p><p>Lets take this opportunity to carefully structure the associated code in a nice and modular way. First, I\u2019d like you to note that every <strong>wire<\/strong> in our diagrams has two numbers associated with it:<\/p><ol><li>the value it carries during the forward pass<\/li>\n  <li>the gradient (i.e the <em>pull<\/em>) that flows back through it in the backward pass<\/li>\n<\/ol><p>Lets create a simple <code class=\"highlighter-rouge\">Unit<\/code> structure that will store these two values on every wire. Our gates will now operate over <code class=\"highlighter-rouge\">Unit<\/code>s: they will take them as inputs and create them as outputs.<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ every Unit corresponds to a wire in the diagrams\nvar Unit = function(value, grad) {\n  \/\/ value computed in the forward pass\n  this.value = value; \n  \/\/ the derivative of circuit output w.r.t this unit, computed in backward pass\n  this.grad = grad; \n}\n<\/pre><\/div><\/div><p>In addition to Units we also need 3 gates: <code class=\"highlighter-rouge\">+<\/code>, <code class=\"highlighter-rouge\">*<\/code> and <code class=\"highlighter-rouge\">sig<\/code> (sigmoid). Lets start out by implementing a multiply gate. I\u2019m using Javascript here which has a funny way of simulating classes using functions. If you\u2019re not a Javascript - familiar person, all that\u2019s going on here is that I\u2019m defining a class that has certain properties (accessed with use of <code class=\"highlighter-rouge\">this<\/code> keyword), and some methods (which in Javascript are placed into the function\u2019s <em>prototype<\/em>). Just think about these as class methods. Also keep in mind that the way we will use these eventually is that we will first <code class=\"highlighter-rouge\">forward<\/code> all the gates one by one, and then <code class=\"highlighter-rouge\">backward<\/code> all the gates in reverse order. Here is the implementation:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\nvar multiplyGate = function(){ };\nmultiplyGate.prototype = {\n  forward: function(u0, u1) {\n    \/\/ store pointers to input Units u0 and u1 and output unit utop\n    this.u0 = u0; \n    this.u1 = u1; \n    this.utop = new Unit(u0.value * u1.value, 0.0);\n    return this.utop;\n  },\n  backward: function() {\n    \/\/ take the gradient in output unit and chain it with the\n    \/\/ local gradients, which we derived for multiply gate before\n    \/\/ then write those gradients to those Units.\n    this.u0.grad += this.u1.value * this.utop.grad;\n    this.u1.grad += this.u0.value * this.utop.grad;\n  }\n}\n<\/pre><\/div><\/div><p>The multiply gate takes two units that each hold a value and creates a unit that stores its output. The gradient is initialized to zero. Then notice that in the <code class=\"highlighter-rouge\">backward<\/code> function call we get the gradient from the output unit we produced during the forward pass (which will by now hopefully have its gradient filled in) and multiply it with the local gradient for this gate (chain rule!). This gate computes multiplication (<code class=\"highlighter-rouge\">u0.value * u1.value<\/code>) during forward pass, so recall that the gradient w.r.t <code class=\"highlighter-rouge\">u0<\/code> is <code class=\"highlighter-rouge\">u1.value<\/code> and w.r.t <code class=\"highlighter-rouge\">u1<\/code> is <code class=\"highlighter-rouge\">u0.value<\/code>. Also note that we are using <code class=\"highlighter-rouge\">+=<\/code> to add onto the gradient in the <code class=\"highlighter-rouge\">backward<\/code> function. This will allow us to possibly use the output of one gate multiple times (think of it as a wire branching out), since it turns out that the gradients from these different branches just add up when computing the final gradient with respect to the circuit output. The other two gates are defined analogously:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var addGate = function(){ };\naddGate.prototype = {\n  forward: function(u0, u1) {\n    this.u0 = u0; \n    this.u1 = u1; \/\/ store pointers to input units\n    this.utop = new Unit(u0.value + u1.value, 0.0);\n    return this.utop;\n  },\n  backward: function() {\n    \/\/ add gate. derivative wrt both inputs is 1\n    this.u0.grad += 1 * this.utop.grad;\n    this.u1.grad += 1 * this.utop.grad;\n  }\n}\n<\/pre><\/div><\/div><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var sigmoidGate = function() { \n  \/\/ helper function\n  this.sig = function(x) { return 1 \/ (1 + Math.exp(-x)); };\n};\nsigmoidGate.prototype = {\n  forward: function(u0) {\n    this.u0 = u0;\n    this.utop = new Unit(this.sig(this.u0.value), 0.0);\n    return this.utop;\n  },\n  backward: function() {\n    var s = this.sig(this.u0.value);\n    this.u0.grad += (s * (1 - s)) * this.utop.grad;\n  }\n}\n<\/pre><\/div><\/div><p>Note that, again, the <code class=\"highlighter-rouge\">backward<\/code> function in all cases just computes the local derivative with respect to its input and then multiplies on the gradient from the unit above (i.e. chain rule). To fully specify everything lets finally write out the forward and backward flow for our 2-dimensional neuron with some example values:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ create input units\nvar a = new Unit(1.0, 0.0);\nvar b = new Unit(2.0, 0.0);\nvar c = new Unit(-3.0, 0.0);\nvar x = new Unit(-1.0, 0.0);\nvar y = new Unit(3.0, 0.0);\n\/\/ create the gates\nvar mulg0 = new multiplyGate();\nvar mulg1 = new multiplyGate();\nvar addg0 = new addGate();\nvar addg1 = new addGate();\nvar sg0 = new sigmoidGate();\n\/\/ do the forward pass\nvar forwardNeuron = function() {\n  ax = mulg0.forward(a, x); \/\/ a*x = -1\n  by = mulg1.forward(b, y); \/\/ b*y = 6\n  axpby = addg0.forward(ax, by); \/\/ a*x + b*y = 5\n  axpbypc = addg1.forward(axpby, c); \/\/ a*x + b*y + c = 2\n  s = sg0.forward(axpbypc); \/\/ sig(a*x + b*y + c) = 0.8808\n};\nforwardNeuron();\nconsole.log('circuit output: ' + s.value); \/\/ prints 0.8808\n<\/pre><\/div><\/div><p>And now lets compute the gradient: Simply iterate in reverse order and call the <code class=\"highlighter-rouge\">backward<\/code> function! Remember that we stored the pointers to the units when we did the forward pass, so every gate has access to its inputs and also the output unit it previously produced.<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>s.grad = 1.0;\nsg0.backward(); \/\/ writes gradient into axpbypc\naddg1.backward(); \/\/ writes gradients into axpby and c\naddg0.backward(); \/\/ writes gradients into ax and by\nmulg1.backward(); \/\/ writes gradients into b and y\nmulg0.backward(); \/\/ writes gradients into a and x\n<\/pre><\/div><\/div><p>Note that the first line sets the gradient at the output (very last unit) to be <code class=\"highlighter-rouge\">1.0<\/code> to start off the gradient chain. This can be interpreted as tugging on the last gate with a force of <code class=\"highlighter-rouge\">+1<\/code>. In other words, we are pulling on the entire circuit to induce the forces that will increase the output value. If we did not set this to 1, all gradients would be computed as zero due to the multiplications in the chain rule. Finally, lets make the inputs respond to the computed gradients and check that the function increased:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var step_size = 0.01;\na.value += step_size * a.grad; \/\/ a.grad is -0.105\nb.value += step_size * b.grad; \/\/ b.grad is 0.315\nc.value += step_size * c.grad; \/\/ c.grad is 0.105\nx.value += step_size * x.grad; \/\/ x.grad is 0.105\ny.value += step_size * y.grad; \/\/ y.grad is 0.210\nforwardNeuron();\nconsole.log('circuit output after one backprop: ' + s.value); \/\/ prints 0.8825\n<\/pre><\/div><\/div><p>Success! <code class=\"highlighter-rouge\">0.8825<\/code> is higher than the previous value, <code class=\"highlighter-rouge\">0.8808<\/code>. Finally, lets verify that we implemented the backpropagation correctly by checking the numerical gradient:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var forwardCircuitFast = function(a,b,c,x,y) { \n  return 1\/(1 + Math.exp( - (a*x + b*y + c))); \n};\nvar a = 1, b = 2, c = -3, x = -1, y = 3;\nvar h = 0.0001;\nvar a_grad = (forwardCircuitFast(a+h,b,c,x,y) - forwardCircuitFast(a,b,c,x,y))\/h;\nvar b_grad = (forwardCircuitFast(a,b+h,c,x,y) - forwardCircuitFast(a,b,c,x,y))\/h;\nvar c_grad = (forwardCircuitFast(a,b,c+h,x,y) - forwardCircuitFast(a,b,c,x,y))\/h;\nvar x_grad = (forwardCircuitFast(a,b,c,x+h,y) - forwardCircuitFast(a,b,c,x,y))\/h;\nvar y_grad = (forwardCircuitFast(a,b,c,x,y+h) - forwardCircuitFast(a,b,c,x,y))\/h;\n<\/pre><\/div><\/div><p>Indeed, these all give the same values as the backpropagated gradients <code class=\"highlighter-rouge\">[-0.105, 0.315, 0.105, 0.105, 0.210]<\/code>. Nice!<\/p><p>I hope it is clear that even though we only looked at an example of a single neuron, the code I gave above generalizes in a very straight-forward way to compute gradients of arbitrary expressions (including very deep expressions #foreshadowing). All you have to do is write small gates that compute local, simple derivatives w.r.t their inputs, wire it up in a graph, do a forward pass to compute the output value and then a backward pass that chains the gradients all the way to the input.<\/p><h3 id=\"becoming-a-backprop-ninja\">Becoming a Backprop Ninja<\/h3><p>Over time you will become much more efficient in writing the backward pass, even for complicated circuits and all at once. Lets practice backprop a bit with a few examples. In what follows, lets not worry about Unit, Circuit classes because they obfuscate things a bit, and lets just use variables such as <code class=\"highlighter-rouge\">a,b,c,x<\/code>, and refer to their gradients as <code class=\"highlighter-rouge\">da,db,dc,dx<\/code> respectively. Again, we think of the variables as the \u201cforward flow\u201d and their gradients as \u201cbackward flow\u201d along every wire. Our first example was the <code class=\"highlighter-rouge\">*<\/code> gate:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = a * b;\n\/\/ and given gradient on x (dx), we saw that in backprop we would compute:\nvar da = b * dx;\nvar db = a * dx;\n<\/pre><\/div><\/div><p>In the code above, I\u2019m assuming that the variable <code class=\"highlighter-rouge\">dx<\/code> is given, coming from somewhere above us in the circuit while we\u2019re doing backprop (or it is +1 by default otherwise). I\u2019m writing it out because I want to explicitly show how the gradients get chained together. Note from the equations that the <code class=\"highlighter-rouge\">*<\/code> gate acts as a <em>switcher<\/em> during backward pass, for lack of better word. It remembers what its inputs were, and the gradients on each one will be the value of the other during the forward pass. And then of course we have to multiply with the gradient from above, which is the chain rule. Here\u2019s the <code class=\"highlighter-rouge\">+<\/code> gate in this condensed form:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = a + b;\n\/\/ -&gt;\nvar da = 1.0 * dx;\nvar db = 1.0 * dx;\n<\/pre><\/div><\/div><p>Where <code class=\"highlighter-rouge\">1.0<\/code> is the local gradient, and the multiplication is our chain rule. What about adding three numbers?:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ lets compute x = a + b + c in two steps:\nvar q = a + b; \/\/ gate 1\nvar x = q + c; \/\/ gate 2\n\/\/ backward pass:\ndc = 1.0 * dx; \/\/ backprop gate 2\ndq = 1.0 * dx; \nda = 1.0 * dq; \/\/ backprop gate 1\ndb = 1.0 * dq;\n<\/pre><\/div><\/div><p>You can see what\u2019s happening, right? If you remember the backward flow diagram, the <code class=\"highlighter-rouge\">+<\/code> gate simply takes the gradient on top and routes it equally to all of its inputs (because its local gradient is always simply <code class=\"highlighter-rouge\">1.0<\/code> for all its inputs, regardless of their actual values). So we can do it much faster:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = a + b + c;\nvar da = 1.0 * dx; var db = 1.0 * dx; var dc = 1.0 * dx;\n<\/pre><\/div><\/div><p>Okay, how about combining gates?:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = a * b + c;\n\/\/ given dx, backprop in-one-sweep would be =&gt;\nda = b * dx;\ndb = a * dx;\ndc = 1.0 * dx;\n<\/pre><\/div><\/div><p>If you don\u2019t see how the above happened, introduce a temporary variable <code class=\"highlighter-rouge\">q = a * b<\/code> and then compute <code class=\"highlighter-rouge\">x = q + c<\/code> to convince yourself. And here is our neuron, lets do it in two steps:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ lets do our neuron in two steps:\nvar q = a*x + b*y + c;\nvar f = sig(q); \/\/ sig is the sigmoid function\n\/\/ and now backward pass, we are given df, and:\nvar df = 1;\nvar dq = (f * (1 - f)) * df;\n\/\/ and now we chain it to the inputs\nvar da = x * dq;\nvar dx = a * dq;\nvar dy = b * dq;\nvar db = y * dq;\nvar dc = 1.0 * dq;\n<\/pre><\/div><\/div><p>I hope this is starting to make a little more sense. Now how about this:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = a * a;\nvar da = \/\/???\n<\/pre><\/div><\/div><p>You can think of this as value <code class=\"highlighter-rouge\">a<\/code> flowing to the <code class=\"highlighter-rouge\">*<\/code> gate, but the wire gets split and becomes both inputs. This is actually simple because the backward flow of gradients always adds up. In other words nothing changes:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var da = a * dx; \/\/ gradient into a from first branch\nda += a * dx; \/\/ and add on the gradient from the second branch\n\/\/ short form instead is:\nvar da = 2 * a * dx;\n<\/pre><\/div><\/div><p>In fact, if you know your power rule from calculus you would also know that if you have \\( f(a) = a^2 \\) then \\( \\frac{\\partial f(a)}{\\partial a} = 2a \\), which is exactly what we get if we think of it as wire splitting up and being two inputs to a gate.<\/p><p>Lets do another one:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = a*a + b*b + c*c;\n\/\/ we get:\nvar da = 2*a*dx;\nvar db = 2*b*dx;\nvar dc = 2*c*dx;\n<\/pre><\/div><\/div><p>Okay now lets start to get more complex:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = Math.pow(((a * b + c) * d), 2); \/\/ pow(x,2) squares the input JS\n<\/pre><\/div><\/div><p>When more complex cases like this come up in practice, I like to split the expression into manageable chunks which are almost always composed of simpler expressions and then I chain them together with chain rule:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x1 = a * b + c;\nvar x2 = x1 * d;\nvar x = x2 * x2; \/\/ this is identical to the above expression for x\n\/\/ and now in backprop we go backwards:\nvar dx2 = 2 * x2 * dx; \/\/ backprop into x2\nvar dd = x1 * dx2; \/\/ backprop into d\nvar dx1 = d * dx2; \/\/ backprop into x1\nvar da = b * dx1;\nvar db = a * dx1;\nvar dc = 1.0 * dx1; \/\/ done!\n<\/pre><\/div><\/div><p>That wasn\u2019t too difficult! Those are the backprop equations for the entire expression, and we\u2019ve done them piece by piece and backpropped to all the variables. Notice again how for every variable during forward pass we have an equivalent variable during backward pass that contains its gradient with respect to the circuit\u2019s final output. Here are a few more useful functions and their local gradients that are useful in practice:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = 1.0\/a; \/\/ division\nvar da = -1.0\/(a*a);\n<\/pre><\/div><\/div><p>Here\u2019s what division might look like in practice then:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = (a + b)\/(c + d);\n\/\/ lets decompose it in steps:\nvar x1 = a + b;\nvar x2 = c + d;\nvar x3 = 1.0 \/ x2;\nvar x = x1 * x3; \/\/ equivalent to above\n\/\/ and now backprop, again in reverse order:\nvar dx1 = x3 * dx;\nvar dx3 = x1 * dx;\nvar dx2 = (-1.0\/(x2*x2)) * dx3; \/\/ local gradient as shown above, and chain rule\nvar da = 1.0 * dx1; \/\/ and finally into the original variables\nvar db = 1.0 * dx1;\nvar dc = 1.0 * dx2;\nvar dd = 1.0 * dx2;\n<\/pre><\/div><\/div><p>Hopefully you see that we are breaking down expressions, doing the forward pass, and then for every variable (such as <code class=\"highlighter-rouge\">a<\/code>) we derive its gradient <code class=\"highlighter-rouge\">da<\/code> as we go backwards, one by one, applying the simple local gradients and chaining them with gradients from above. Here\u2019s another one:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = Math.max(a, b);\nvar da = a === x ? 1.0 * dx : 0.0;\nvar db = b === x ? 1.0 * dx : 0.0;\n<\/pre><\/div><\/div><p>Okay this is making a very simple thing hard to read. The <code class=\"highlighter-rouge\">max<\/code> function passes on the value of the input that was largest and ignores the other ones. In the backward pass then, the max gate will simply take the gradient on top and route it to the input that actually flowed through it during the forward pass. The gate acts as a simple switch based on which input had the highest value during forward pass. The other inputs will have zero gradient. That\u2019s what the <code class=\"highlighter-rouge\">===<\/code> is about, since we are testing for which input was the actual max and only routing the gradient to it.<\/p><p>Finally, lets look at the Rectified Linear Unit non-linearity (or ReLU), which you may have heard of. It is used in Neural Networks in place of the sigmoid function. It is simply thresholding at zero:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var x = Math.max(a, 0)\n\/\/ backprop through this gate will then be:\nvar da = a &gt; 0 ? 1.0 * dx : 0.0;\n<\/pre><\/div><\/div><p>In other words this gate simply passes the value through if it\u2019s larger than 0, or it stops the flow and sets it to zero. In the backward pass, the gate will pass on the gradient from the top if it was activated during the forawrd pass, or if the original input was below zero, it will stop the gradient flow.<\/p><p>I will stop at this point. I hope you got some intuition about how you can compute entire expressions (which are made up of many gates along the way) and how you can compute backprop for every one of them.<\/p><p>Everything we\u2019ve done in this chapter comes down to this: We saw that we can feed some input through arbitrarily complex real-valued circuit, tug at the end of the circuit with some force, and backpropagation distributes that tug through the entire circuit all the way back to the inputs. If the inputs respond slightly along the final direction of their tug, the circuit will \u201cgive\u201d a bit along the original pull direction. Maybe this is not immediately obvious, but this machinery is a powerful <em>hammer<\/em> for Machine Learning.<\/p><blockquote>\n  <p>\u201cMaybe this is not immediately obvious, but this machinery is a powerful <em>hammer<\/em> for Machine Learning.\u201d<\/p>\n<\/blockquote><p>Lets now put this machinery to good use.<\/p><h2 id=\"chapter-2-machine-learning\">Chapter 2: Machine Learning<\/h2><p>In the last chapter we were concerned with real-valued circuits that computed possibly complex expressions of their inputs (the forward pass), and also we could compute the gradients of these expressions on the original inputs (backward pass). In this chapter we will see how useful this extremely simple mechanism is in Machine Learning.<\/p><h3 id=\"binary-classification\">Binary Classification<\/h3><p>As we did before, lets start out simple. The simplest, common and yet very practical problem in Machine Learning is <strong>binary classification<\/strong>. A lot of very interesting and important problems can be reduced to it. The setup is as follows: We are given a dataset of <code class=\"highlighter-rouge\">N<\/code> vectors and every one of them is labeled with a <code class=\"highlighter-rouge\">+1<\/code> or a <code class=\"highlighter-rouge\">-1<\/code>. For example, in two dimensions our dataset could look as simple as:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>vector -&gt; label\n---------------\n[1.2, 0.7] -&gt; +1\n[-0.3, 0.5] -&gt; -1\n[-3, -1] -&gt; +1\n[0.1, 1.0] -&gt; -1\n[3.0, 1.1] -&gt; -1\n[2.1, -3] -&gt; +1\n<\/pre><\/div><\/div><p>Here, we have <code class=\"highlighter-rouge\">N = 6<\/code> <strong>datapoints<\/strong>, where every datapoint has two <strong>features<\/strong> (<code class=\"highlighter-rouge\">D = 2<\/code>). Three of the datapoints have <strong>label<\/strong> <code class=\"highlighter-rouge\">+1<\/code> and the other three label <code class=\"highlighter-rouge\">-1<\/code>. This is a silly toy example, but in practice a +1\/-1 dataset could be very useful things indeed: For example spam\/no spam emails, where the vectors somehow measure various features of the content of the email, such as the number of times certain enhancement drugs are mentioned.<\/p><p><strong>Goal<\/strong>. Our goal in binary classification is to learn a function that takes a 2-dimensional vector and predicts the label. This function is usually parameterized by a certain set of parameters, and we will want to tune the parameters of the function so that its outputs are consistent with the labeling in the provided dataset. In the end we can discard the dataset and use the learned parameters to predict labels for previously unseen vectors.<\/p><h4 id=\"training-protocol\">Training protocol<\/h4><p>We will eventually build up to entire neural networks and complex expressions, but lets start out simple and train a linear classifier very similar to the single neuron we saw at the end of Chapter 1. The only difference is that we\u2019ll get rid of the sigmoid because it makes things unnecessarily complicated (I only used it as an example in Chapter 1 because sigmoid neurons are historically popular but modern Neural Networks rarely, if ever, use sigmoid non-linearities). Anyway, lets use a simple linear function:<\/p><p>In this expression we think of <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> as the inputs (the 2D vectors) and <code class=\"highlighter-rouge\">a,b,c<\/code> as the parameters of the function that we will want to learn. For example, if <code class=\"highlighter-rouge\">a = 1, b = -2, c = -1<\/code>, then the function will take the first datapoint (<code class=\"highlighter-rouge\">[1.2, 0.7]<\/code>) and output <code class=\"highlighter-rouge\">1 * 1.2 + (-2) * 0.7 + (-1) = -1.2<\/code>. Here is how the training will work:<\/p><ol><li>We select a random datapoint and feed it through the circuit<\/li>\n  <li>We will interpret the output of the circuit as a confidence that the datapoint has class <code class=\"highlighter-rouge\">+1<\/code>. (i.e. very high values = circuit is very certain datapoint has class <code class=\"highlighter-rouge\">+1<\/code> and very low values = circuit is certain this datapoint has class <code class=\"highlighter-rouge\">-1<\/code>.)<\/li>\n  <li>We will measure how well the prediction aligns with the provided labels. Intuitively, for example, if a positive example scores very low, we will want to tug in the positive direction on the circuit, demanding that it should output higher value for this datapoint. Note that this is the case for the the first datapoint: it is labeled as <code class=\"highlighter-rouge\">+1<\/code> but our predictor unction only assigns it value <code class=\"highlighter-rouge\">-1.2<\/code>. We will therefore tug on the circuit in positive direction; We want the value to be higher.<\/li>\n  <li>The circuit will take the tug and backpropagate it to compute tugs on the inputs <code class=\"highlighter-rouge\">a,b,c,x,y<\/code><\/li>\n  <li>Since we think of <code class=\"highlighter-rouge\">x,y<\/code> as (fixed) datapoints, we will ignore the pull on <code class=\"highlighter-rouge\">x,y<\/code>. If you\u2019re a fan of my physical analogies, think of these inputs as pegs, fixed in the ground.<\/li>\n  <li>On the other hand, we will take the parameters <code class=\"highlighter-rouge\">a,b,c<\/code> and make them respond to their tug (i.e. we\u2019ll perform what we call a <strong>parameter update<\/strong>). This, of course, will make it so that the circuit will output a slightly higher score on this particular datapoint in the future.<\/li>\n  <li>Iterate! Go back to step 1.<\/li>\n<\/ol><p>The training scheme I described above, is commonly referred as <strong>Stochastic Gradient Descent<\/strong>. The interesting part I\u2019d like to reiterate is that <code class=\"highlighter-rouge\">a,b,c,x,y<\/code> are all made up of the same <em>stuff<\/em> as far as the circuit is concerned: They are inputs to the circuit and the circuit will tug on all of them in some direction. It doesn\u2019t know the difference between parameters and datapoints. However, after the backward pass is complete we ignore all tugs on the datapoints (<code class=\"highlighter-rouge\">x,y<\/code>) and keep swapping them in and out as we iterate over examples in the dataset. On the other hand, we keep the parameters (<code class=\"highlighter-rouge\">a,b,c<\/code>) around and keep tugging on them every time we sample a datapoint. Over time, the pulls on these parameters will tune these values in such a way that the function outputs high scores for positive examples and low scores for negative examples.<\/p><h4 id=\"learning-a-support-vector-machine\">Learning a Support Vector Machine<\/h4><p>As a concrete example, lets learn a <strong>Support Vector Machine<\/strong>. The SVM is a very popular linear classifier; Its functional form is exactly as I\u2019ve described in previous section, \\( f(x,y) = ax + by + c\\). At this point, if you\u2019ve seen an explanation of SVMs you\u2019re probably expecting me to define the SVM loss function and plunge into an explanation of slack variables, geometrical intuitions of large margins, kernels, duality, etc. But here, I\u2019d like to take a different approach. Instead of definining loss functions, I would like to base the explanation on the <em>force specification<\/em> (I just made this term up by the way) of a Support Vector Machine, which I personally find much more intuitive. As we will see, talking about the force specification and the loss function are identical ways of seeing the same problem. Anyway, here it is:<\/p><p><strong>Support Vector Machine \u201cForce Specification\u201d:<\/strong><\/p><ul><li>If we feed a positive datapoint through the SVM circuit and the output value is less than 1, pull on the circuit with force <code class=\"highlighter-rouge\">+1<\/code>. This is a positive example so we want the score to be higher for it.<\/li>\n  <li>Conversely, if we feed a negative datapoint through the SVM and the output is greater than -1, then the circuit is giving this datapoint dangerously high score: Pull on the circuit downwards with force <code class=\"highlighter-rouge\">-1<\/code>.<\/li>\n  <li>In addition to the pulls above, always add a small amount of pull on the parameters <code class=\"highlighter-rouge\">a,b<\/code> (notice, not on <code class=\"highlighter-rouge\">c<\/code>!) that pulls them towards zero. You can think of both <code class=\"highlighter-rouge\">a,b<\/code> as being attached to a physical spring that is attached at zero. Just as with a physical spring, this will make the pull proprotional to the value of each of <code class=\"highlighter-rouge\">a,b<\/code> (Hooke\u2019s law in physics, anyone?). For example, if <code class=\"highlighter-rouge\">a<\/code> becomes very high it will experience a strong pull of magnitude <code class=\"highlighter-rouge\">|a|<\/code> back towards zero. This pull is something we call <strong>regularization<\/strong>, and it ensures that neither of our parameters <code class=\"highlighter-rouge\">a<\/code> or <code class=\"highlighter-rouge\">b<\/code> gets disproportionally large. This would be undesirable because both <code class=\"highlighter-rouge\">a,b<\/code> get multiplied to the input features <code class=\"highlighter-rouge\">x,y<\/code> (remember the equation is <code class=\"highlighter-rouge\">a*x + b*y + c<\/code>), so if either of them is too high, our classifier would be overly sensitive to these features. This isn\u2019t a nice property because features can often be noisy in practice, so we want our classifier to change relatively smoothly if they wiggle around.<\/li>\n<\/ul><p>Lets quickly go through a small but concrete example. Suppose we start out with a random parameter setting, say, <code class=\"highlighter-rouge\">a = 1, b = -2, c = -1<\/code>. Then:<\/p><ul><li>If we feed the point <code class=\"highlighter-rouge\">[1.2, 0.7]<\/code>, the SVM will compute score <code class=\"highlighter-rouge\">1 * 1.2 + (-2) * 0.7 - 1 = -1.2<\/code>. This point is labeled as <code class=\"highlighter-rouge\">+1<\/code> in the training data, so we want the score to be higher than 1. The gradient on top of the circuit will thus be positive: <code class=\"highlighter-rouge\">+1<\/code>, which will backpropagate to <code class=\"highlighter-rouge\">a,b,c<\/code>. Additionally, there will also be a regularization pull on <code class=\"highlighter-rouge\">a<\/code> of <code class=\"highlighter-rouge\">-1<\/code> (to make it smaller) and regularization pull on <code class=\"highlighter-rouge\">b<\/code> of <code class=\"highlighter-rouge\">+2<\/code> to make it larger, toward zero.<\/li>\n  <li>Suppose instead that we fed the datapoint <code class=\"highlighter-rouge\">[-0.3, 0.5]<\/code> to the SVM. It computes <code class=\"highlighter-rouge\">1 * (-0.3) + (-2) * 0.5 - 1 = -2.3<\/code>. The label for this point is <code class=\"highlighter-rouge\">-1<\/code>, and since <code class=\"highlighter-rouge\">-2.3<\/code> is smaller than <code class=\"highlighter-rouge\">-1<\/code>, we see that according to our force specification the SVM should be happy: The computed score is very negative, consistent with the negative label of this example. There will be no pull at the end of the circuit (i.e it\u2019s zero), since there no changes are necessary. However, there will <em>still<\/em> be the regularization pull on <code class=\"highlighter-rouge\">a<\/code> of <code class=\"highlighter-rouge\">-1<\/code> and on <code class=\"highlighter-rouge\">b<\/code> of <code class=\"highlighter-rouge\">+2<\/code>.<\/li>\n<\/ul><p>Okay there\u2019s been too much text. Lets write the SVM code and take advantage of the circuit machinery we have from Chapter 1:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ A circuit: it takes 5 Units (x,y,a,b,c) and outputs a single Unit\n\/\/ It can also compute the gradient w.r.t. its inputs\nvar Circuit = function() {\n  \/\/ create some gates\n  this.mulg0 = new multiplyGate();\n  this.mulg1 = new multiplyGate();\n  this.addg0 = new addGate();\n  this.addg1 = new addGate();\n};\nCircuit.prototype = {\n  forward: function(x,y,a,b,c) {\n    this.ax = this.mulg0.forward(a, x); \/\/ a*x\n    this.by = this.mulg1.forward(b, y); \/\/ b*y\n    this.axpby = this.addg0.forward(this.ax, this.by); \/\/ a*x + b*y\n    this.axpbypc = this.addg1.forward(this.axpby, c); \/\/ a*x + b*y + c\n    return this.axpbypc;\n  },\n  backward: function(gradient_top) { \/\/ takes pull from above\n    this.axpbypc.grad = gradient_top;\n    this.addg1.backward(); \/\/ sets gradient in axpby and c\n    this.addg0.backward(); \/\/ sets gradient in ax and by\n    this.mulg1.backward(); \/\/ sets gradient in b and y\n    this.mulg0.backward(); \/\/ sets gradient in a and x\n  }\n}\n<\/pre><\/div><\/div><p>That\u2019s a circuit that simply computes <code class=\"highlighter-rouge\">a*x + b*y + c<\/code> and can also compute the gradient. It uses the gates code we developed in Chapter 1. Now lets write the SVM, which doesn\u2019t care about the actual circuit. It is only concerned with the values that come out of it, and it pulls on the circuit.<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ SVM class\nvar SVM = function() {\n  \n  \/\/ random initial parameter values\n  this.a = new Unit(1.0, 0.0); \n  this.b = new Unit(-2.0, 0.0);\n  this.c = new Unit(-1.0, 0.0);\n  this.circuit = new Circuit();\n};\nSVM.prototype = {\n  forward: function(x, y) { \/\/ assume x and y are Units\n    this.unit_out = this.circuit.forward(x, y, this.a, this.b, this.c);\n    return this.unit_out;\n  },\n  backward: function(label) { \/\/ label is +1 or -1\n    \/\/ reset pulls on a,b,c\n    this.a.grad = 0.0; \n    this.b.grad = 0.0; \n    this.c.grad = 0.0;\n    \/\/ compute the pull based on what the circuit output was\n    var pull = 0.0;\n    if(label === 1 &amp;&amp; this.unit_out.value &lt; 1) { \n      pull = 1; \/\/ the score was too low: pull up\n    }\n    if(label === -1 &amp;&amp; this.unit_out.value &gt; -1) {\n      pull = -1; \/\/ the score was too high for a positive example, pull down\n    }\n    this.circuit.backward(pull); \/\/ writes gradient into x,y,a,b,c\n    \n    \/\/ add regularization pull for parameters: towards zero and proportional to value\n    this.a.grad += -this.a.value;\n    this.b.grad += -this.b.value;\n  },\n  learnFrom: function(x, y, label) {\n    this.forward(x, y); \/\/ forward pass (set .value in all Units)\n    this.backward(label); \/\/ backward pass (set .grad in all Units)\n    this.parameterUpdate(); \/\/ parameters respond to tug\n  },\n  parameterUpdate: function() {\n    var step_size = 0.01;\n    this.a.value += step_size * this.a.grad;\n    this.b.value += step_size * this.b.grad;\n    this.c.value += step_size * this.c.grad;\n  }\n};\n<\/pre><\/div><\/div><p>Now lets train the SVM with Stochastic Gradient Descent:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var data = []; var labels = [];\ndata.push([1.2, 0.7]); labels.push(1);\ndata.push([-0.3, -0.5]); labels.push(-1);\ndata.push([3.0, 0.1]); labels.push(1);\ndata.push([-0.1, -1.0]); labels.push(-1);\ndata.push([-1.0, 1.1]); labels.push(-1);\ndata.push([2.1, -3]); labels.push(1);\nvar svm = new SVM();\n\/\/ a function that computes the classification accuracy\nvar evalTrainingAccuracy = function() {\n  var num_correct = 0;\n  for(var i = 0; i &lt; data.length; i++) {\n    var x = new Unit(data[i][0], 0.0);\n    var y = new Unit(data[i][1], 0.0);\n    var true_label = labels[i];\n    \/\/ see if the prediction matches the provided label\n    var predicted_label = svm.forward(x, y).value &gt; 0 ? 1 : -1;\n    if(predicted_label === true_label) {\n      num_correct++;\n    }\n  }\n  return num_correct \/ data.length;\n};\n\/\/ the learning loop\nfor(var iter = 0; iter &lt; 400; iter++) {\n  \/\/ pick a random data point\n  var i = Math.floor(Math.random() * data.length);\n  var x = new Unit(data[i][0], 0.0);\n  var y = new Unit(data[i][1], 0.0);\n  var label = labels[i];\n  svm.learnFrom(x, y, label);\n  if(iter % 25 == 0) { \/\/ every 10 iterations... \n    console.log('training accuracy at iter ' + iter + ': ' + evalTrainingAccuracy());\n  }\n}\n<\/pre><\/div><\/div><p>This code prints the following output:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>training accuracy at iteration 0: 0.3333333333333333\ntraining accuracy at iteration 25: 0.3333333333333333\ntraining accuracy at iteration 50: 0.5\ntraining accuracy at iteration 75: 0.5\ntraining accuracy at iteration 100: 0.3333333333333333\ntraining accuracy at iteration 125: 0.5\ntraining accuracy at iteration 150: 0.5\ntraining accuracy at iteration 175: 0.5\ntraining accuracy at iteration 200: 0.5\ntraining accuracy at iteration 225: 0.6666666666666666\ntraining accuracy at iteration 250: 0.6666666666666666\ntraining accuracy at iteration 275: 0.8333333333333334\ntraining accuracy at iteration 300: 1\ntraining accuracy at iteration 325: 1\ntraining accuracy at iteration 350: 1\ntraining accuracy at iteration 375: 1 \n<\/pre><\/div><\/div><p>We see that initially our classifier only had 33% training accuracy, but by the end all training examples are correctly classifier as the parameters <code class=\"highlighter-rouge\">a,b,c<\/code> adjusted their values according to the pulls we exerted. We just trained an SVM! But please don\u2019t use this code anywhere in production :) We will see how we can make things much more efficient once we understand what is going on at the core.<\/p><p><strong>Number of iterations needed<\/strong>. With this example data, with this example initialization, and with the setting of step size we used, it took about 300 iterations to train the SVM. In practice, this could be many more or many less depending on how hard or large the problem is, how you\u2019re initializating, normalizing your data, what step size you\u2019re using, and so on. This is just a toy demonstration, but later we will go over all the best practices for actually training these classifiers in practice. For example, it will turn out that the setting of the step size is very imporant and tricky. Small step size will make your model slow to train. Large step size will train faster, but if it is too large, it will make your classifier chaotically jump around and not converge to a good final result. We will eventually use witheld validation data to properly tune it to be just in the sweet spot for your particular data.<\/p><p>One thing I\u2019d like you to appreciate is that the circuit can be arbitrary expression, not just the linear prediction function we used in this example. For example, it can be an entire neural network.<\/p><p>By the way, I intentionally structured the code in a modular way, but we could have trained an SVM with a much simpler code. Here is really what all of these classes and computations boil down to:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var a = 1, b = -2, c = -1; \/\/ initial parameters\nfor(var iter = 0; iter &lt; 400; iter++) {\n  \/\/ pick a random data point\n  var i = Math.floor(Math.random() * data.length);\n  var x = data[i][0];\n  var y = data[i][1];\n  var label = labels[i];\n  \/\/ compute pull\n  var score = a*x + b*y + c;\n  var pull = 0.0;\n  if(label === 1 &amp;&amp; score &lt; 1) pull = 1;\n  if(label === -1 &amp;&amp; score &gt; -1) pull = -1;\n  \/\/ compute gradient and update parameters\n  var step_size = 0.01;\n  a += step_size * (x * pull - a); \/\/ -a is from the regularization\n  b += step_size * (y * pull - b); \/\/ -b is from the regularization\n  c += step_size * (1 * pull);\n}\n<\/pre><\/div><\/div><p>this code gives an identical result. Perhaps by now you can glance at the code and see how these equations came about.<\/p><p><strong>Variable pull?<\/strong> A quick note to make at this point: You may have noticed that the pull is always 1,0, or -1. You could imagine doing other things, for example making this pull proportional to how bad the mistake was. This leads to a variation on the SVM that some people refer to as <em>squared hinge loss<\/em> SVM, for reasons that will later become clear. Depending on various features of your dataset, that may work better or worse. For example, if you have very bad outliers in your data, e.g. a negative data point that gets a score <code class=\"highlighter-rouge\">+100<\/code>, its influence will be relatively minor on our classifier because we will only pull with force of <code class=\"highlighter-rouge\">-1<\/code> regardless of how bad the mistake was. In practice we refer to this property of a classifier as <strong>robustness<\/strong> to outliers.<\/p><p>Lets <strong>recap<\/strong>. We introduced the <strong>binary classification<\/strong> problem, where we are given N D-dimensional vectors and a label +1\/-1 for each. We saw that we can combine these features with a set of parameters inside a real-valued circuit (such as a <strong>Support Vector Machine<\/strong> circuit in our example). Then, we can repeatedly pass our data through the circuit and each time tweak the parameters so that the circuit\u2019s output value is consistent with the provided labels. The tweaking relied, crucially, on our ability to <strong>backpropagate<\/strong> gradients through the circuit. In the end, the final circuit can be used to predict values for unseen instances!<\/p><h4 id=\"generalizing-the-svm-into-a-neural-network\">Generalizing the SVM into a Neural Network<\/h4><p>Of interest is the fact that an SVM is just a particular type of a very simple circuit (circuit that computes <code class=\"highlighter-rouge\">score = a*x + b*y + c<\/code> where <code class=\"highlighter-rouge\">a,b,c<\/code> are weights and <code class=\"highlighter-rouge\">x,y<\/code> are data points). This can be easily extended to more complicated functions. For example, lets write a 2-layer Neural Network that does the binary classification. The forward pass will look like this:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ assume inputs x,y\nvar n1 = Math.max(0, a1*x + b1*y + c1); \/\/ activation of 1st hidden neuron\nvar n2 = Math.max(0, a2*x + b2*y + c2); \/\/ 2nd neuron\nvar n3 = Math.max(0, a3*x + b3*y + c3); \/\/ 3rd neuron\nvar score = a4*n1 + b4*n2 + c4*n3 + d4; \/\/ the score\n<\/pre><\/div><\/div><p>The specification above is a 2-layer Neural Network with 3 hidden neurons (n1, n2, n3) that uses Rectified Linear Unit (ReLU) non-linearity on each hidden neuron. As you can see, there are now several parameters involved, which means that our classifier is more complex and can represent more intricate decision boundaries than just a simple linear decision rule such as an SVM. Another way to think about it is that every one of the three hidden neurons is a linear classifier and now we\u2019re putting an extra linear classifier on top of that. Now we\u2019re starting to go <em>deeper<\/em> :). Okay, lets train this 2-layer Neural Network. The code looks very similar to the SVM example code above, we just have to change the forward pass and the backward pass:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ random initial parameters\nvar a1 = Math.random() - 0.5; \/\/ a random number between -0.5 and 0.5\n\/\/ ... similarly initialize all other parameters to randoms\nfor(var iter = 0; iter &lt; 400; iter++) {\n  \/\/ pick a random data point\n  var i = Math.floor(Math.random() * data.length);\n  var x = data[i][0];\n  var y = data[i][1];\n  var label = labels[i];\n  \/\/ compute forward pass\n  var n1 = Math.max(0, a1*x + b1*y + c1); \/\/ activation of 1st hidden neuron\n  var n2 = Math.max(0, a2*x + b2*y + c2); \/\/ 2nd neuron\n  var n3 = Math.max(0, a3*x + b3*y + c3); \/\/ 3rd neuron\n  var score = a4*n1 + b4*n2 + c4*n3 + d4; \/\/ the score\n  \/\/ compute the pull on top\n  var pull = 0.0;\n  if(label === 1 &amp;&amp; score &lt; 1) pull = 1; \/\/ we want higher output! Pull up.\n  if(label === -1 &amp;&amp; score &gt; -1) pull = -1; \/\/ we want lower output! Pull down.\n  \/\/ now compute backward pass to all parameters of the model\n  \/\/ backprop through the last \"score\" neuron\n  var dscore = pull;\n  var da4 = n1 * dscore;\n  var dn1 = a4 * dscore;\n  var db4 = n2 * dscore;\n  var dn2 = b4 * dscore;\n  var dc4 = n3 * dscore;\n  var dn3 = c4 * dscore;\n  var dd4 = 1.0 * dscore; \/\/ phew\n  \/\/ backprop the ReLU non-linearities, in place\n  \/\/ i.e. just set gradients to zero if the neurons did not \"fire\"\n  var dn3 = n3 === 0 ? 0 : dn3;\n  var dn2 = n2 === 0 ? 0 : dn2;\n  var dn1 = n1 === 0 ? 0 : dn1;\n  \/\/ backprop to parameters of neuron 1\n  var da1 = x * dn1;\n  var db1 = y * dn1;\n  var dc1 = 1.0 * dn1;\n  \n  \/\/ backprop to parameters of neuron 2\n  var da2 = x * dn2;\n  var db2 = y * dn2;\n  var dc2 = 1.0 * dn2;\n  \/\/ backprop to parameters of neuron 3\n  var da3 = x * dn3;\n  var db3 = y * dn3;\n  var dc3 = 1.0 * dn3;\n  \/\/ phew! End of backprop!\n  \/\/ note we could have also backpropped into x,y\n  \/\/ but we do not need these gradients. We only use the gradients\n  \/\/ on our parameters in the parameter update, and we discard x,y\n  \/\/ add the pulls from the regularization, tugging all multiplicative\n  \/\/ parameters (i.e. not the biases) downward, proportional to their value\n  da1 += -a1; da2 += -a2; da3 += -a3;\n  db1 += -b1; db2 += -b2; db3 += -b3;\n  da4 += -a4; db4 += -b4; dc4 += -c4;\n  \/\/ finally, do the parameter update\n  var step_size = 0.01;\n  a1 += step_size * da1; \n  b1 += step_size * db1; \n  c1 += step_size * dc1;\n  a2 += step_size * da2; \n  b2 += step_size * db2;\n  c2 += step_size * dc2;\n  a3 += step_size * da3; \n  b3 += step_size * db3; \n  c3 += step_size * dc3;\n  a4 += step_size * da4; \n  b4 += step_size * db4; \n  c4 += step_size * dc4; \n  d4 += step_size * dd4;\n  \/\/ wow this is tedious, please use for loops in prod.\n  \/\/ we're done!\n}\n<\/pre><\/div><\/div><p>And that\u2019s how you train a neural network. Obviously, you want to modularize your code nicely but I expended this example for you in the hope that it makes things much more concrete and simpler to understand. Later, we will look at best practices when implementing these networks and we will structure the code much more neatly in a modular and more sensible way.<\/p><p>But for now, I hope your takeaway is that a 2-layer Neural Net is really not such a scary thing: we write a forward pass expression, interpret the value at the end as a score, and then we pull on that value in a positive or negative direction depending on what we want that value to be for our current particular example. The parameter update after backprop will ensure that when we see this particular example in the future, the network will be more likely to give us a value we desire, not the one it gave just before the update.<\/p><p>Now that we understand the basics of how these circuits function with data, lets adopt a more conventional approach that you might see elsewhere on the internet and in other tutorials and books. You won\u2019t see people talking too much about <strong>force specifications<\/strong>. Instead, Machine Learning algorithms are specified in terms of <strong>loss functions<\/strong> (or <strong>cost functions<\/strong>, or <strong>objectives<\/strong>).<\/p><p>As I develop this formalism I would also like to start to be a little more careful with how we name our variables and parameters. I\u2019d like these equations to look similar to what you might see in a book or some other tutorial, so let me use more standard naming conventions.<\/p><h4 id=\"example-2-d-support-vector-machine\">Example: 2-D Support Vector Machine<\/h4><p>Lets start with an example of a 2-dimensional SVM. We are given a dataset of \\( N \\) examples \\( (x_{i0}, x_{i1}) \\) and their corresponding labels \\( y_{i} \\) which are allowed to be either \\( +1\/-1 \\) for positive or negative example respectively. Most importantly, as you recall we have three parameters \\( (w_0, w_1, w_2) \\). The SVM loss function is then defined as follows:<\/p><p>Notice that this expression is always positive, due to the thresholding at zero in the first expression and the squaring in the regularization. The idea is that we will want this expression to be as small as possible. Before we dive into some of its subtleties let me first translate it to code:<\/p><div class=\"language-javascript highlighter-rouge\"><div class=\"highlight\"><pre>var X = [ [1.2, 0.7], [-0.3, 0.5], [3, 2.5] ] \/\/ array of 2-dimensional data\nvar y = [1, -1, 1] \/\/ array of labels\nvar w = [0.1, 0.2, 0.3] \/\/ example: random numbers\nvar alpha = 0.1; \/\/ regularization strength\nfunction cost(X, y, w) {\n  \n  var total_cost = 0.0; \/\/ L, in SVM loss function above\n  N = X.length;\n  for(var i=0;i&lt;N;i++) {\n    \/\/ loop over all data points and compute their score\n    var xi = X[i];\n    var score = w[0] * xi[0] + w[1] * xi[1] + w[2];\n    \n    \/\/ accumulate cost based on how compatible the score is with the label\n    var yi = y[i]; \/\/ label\n    var costi = Math.max(0, - yi * score + 1);\n    console.log('example ' + i + ': xi = (' + xi + ') and label = ' + yi);\n    console.log('  score computed to be ' + score.toFixed(3));\n    console.log('  =&gt; cost computed to be ' + costi.toFixed(3));\n    total_cost += costi;\n  }\n  \/\/ regularization cost: we want small weights\n  reg_cost = alpha * (w[0]*w[0] + w[1]*w[1])\n  console.log('regularization cost for current model is ' + reg_cost.toFixed(3));\n  total_cost += reg_cost;\n  console.log('total cost is ' + total_cost.toFixed(3));\n  return total_cost;\n}\n<\/pre><\/div><\/div><p>And here is the output:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>cost for example 0 is 0.440\ncost for example 1 is 1.370\ncost for example 2 is 0.000\nregularization cost for current model is 0.005\ntotal cost is 1.815 \n<\/pre><\/div><\/div><p>Notice how this expression works: It measures how <em>bad<\/em> our SVM classifier is. Lets step through this explicitly:<\/p><ul><li>The first datapoint <code class=\"highlighter-rouge\">xi = [1.2, 0.7]<\/code> with label <code class=\"highlighter-rouge\">yi = 1<\/code> will give score <code class=\"highlighter-rouge\">0.1*1.2 + 0.2*0.7 + 0.3<\/code>, which is <code class=\"highlighter-rouge\">0.56<\/code>. Notice, this is a positive example so we want to the score to be greater than <code class=\"highlighter-rouge\">+1<\/code>. <code class=\"highlighter-rouge\">0.56<\/code> is not enough. And indeed, the expression for cost for this datapoint will compute: <code class=\"highlighter-rouge\">costi = Math.max(0, -1*0.56 + 1)<\/code>, which is <code class=\"highlighter-rouge\">0.44<\/code>. You can think of the cost as quantifying the SVM\u2019s unhappiness.<\/li>\n  <li>The second datapoint <code class=\"highlighter-rouge\">xi = [-0.3, 0.5]<\/code> with label <code class=\"highlighter-rouge\">yi = -1<\/code> will give score <code class=\"highlighter-rouge\">0.1*(-0.3) + 0.2*0.5 + 0.3<\/code>, which is <code class=\"highlighter-rouge\">0.37<\/code>. This isn\u2019t looking very good: This score is very high for a negative example. It should be less than -1. Indeed, when we compute the cost: <code class=\"highlighter-rouge\">costi = Math.max(0, 1*0.37 + 1)<\/code>, we get <code class=\"highlighter-rouge\">1.37<\/code>. That\u2019s a very high cost from this example, as it is being misclassified.<\/li>\n  <li>The last example <code class=\"highlighter-rouge\">xi = [3, 2.5]<\/code> with label <code class=\"highlighter-rouge\">yi = 1<\/code> gives score <code class=\"highlighter-rouge\">0.1*3 + 0.2*2.5 + 0.3<\/code>, and that is <code class=\"highlighter-rouge\">1.1<\/code>. In this case, the SVM will compute <code class=\"highlighter-rouge\">costi = Math.max(0, -1*1.1 + 1)<\/code>, which is in fact zero. This datapoint is being classified correctly and there is no cost associated with it.<\/li>\n<\/ul><blockquote>\n  <p>A cost function is an expression that measuress how bad your classifier is. When the training set is perfectly classified, the cost (ignoring the regularization) will be zero.<\/p>\n<\/blockquote><p>Notice that the last term in the loss is the regularization cost, which says that our model parameters should be small values. Due to this term the cost will never actually become zero (because this would mean all parameters of the model except the bias are exactly zero), but the closer we get, the better our classifier will become.<\/p><blockquote>\n  <p>The majority of cost functions in Machine Learning consist of two parts: 1. A part that measures how well a model fits the data, and 2: Regularization, which measures some notion of how complex or likely a model is.<\/p>\n<\/blockquote><p>I hope I convinced you then, that to get a very good SVM we really want to make the <strong>cost as small as possible<\/strong>. Sounds familiar? We know exactly what to do: The cost function written above is our circuit. We will forward all examples through the circuit, compute the backward pass and update all parameters such that the circuit will output a <em>smaller<\/em> cost in the future. Specifically, we will compute the <em>gradient<\/em> and then update the parameters in the <em>opposite direction<\/em> of the gradient (since we want to make the cost small, not large).<\/p><blockquote>\n  <p>\u201cWe know exactly what to do: The cost function written above is our circuit.\u201d<\/p>\n<\/blockquote><p>todo: clean up this section and flesh it out a bit\u2026<\/p><h2 id=\"chapter-3-backprop-in-practice\">Chapter 3: Backprop in Practice<\/h2><h3 id=\"building-up-a-library\">Building up a library<\/h3><h3 id=\"example-practical-neural-network-classifier\">Example: Practical Neural Network Classifier<\/h3><ul><li>Multiclass: Structured SVM<\/li>\n  <li>Multiclass: Logistic Regression, Softmax<\/li>\n<\/ul><h3 id=\"example-regression\">Example: Regression<\/h3><p>Tiny changes needed to cost function. L2 regularization.<\/p><h3 id=\"example-structured-prediction\">Example: Structured Prediction<\/h3><p>Basic idea is to train an (unnormalized) energy model<\/p><h3 id=\"vectorized-implementations\">Vectorized Implementations<\/h3><p>Writing a Neural Net classfier in Python with numpy\u2026.<\/p><h3 id=\"backprop-in-practice-tipstricks\">Backprop in practice: Tips\/Tricks<\/h3><ul><li>Monitoring of Cost function<\/li>\n  <li>Monitoring training\/validation performance<\/li>\n  <li>Tweaking initial learning rates, learning rate schedules<\/li>\n  <li>Optimization: Using Momentum<\/li>\n  <li>Optimization: LBFGS, Nesterov accelerated gradient<\/li>\n  <li>Importance of Initialization: weights and biases<\/li>\n  <li>Regularization: L2, L1, Group sparsity, Dropout<\/li>\n  <li>Hyperparameter search, cross-validations<\/li>\n  <li>Common pitfalls: (e.g. dying ReLUs)<\/li>\n  <li>Handling unbalanced datasets<\/li>\n  <li>Approaches to debugging nets when something doesnt work<\/li>\n<\/ul><h2 id=\"chapter-4-networks-in-the-wild\">Chapter 4: Networks in the Wild<\/h2><p>Case studies of models that work well in practice and have been deployed in the wild.<\/p><h3 id=\"case-study-convolutional-neural-networks-for-images\">Case Study: Convolutional Neural Networks for images<\/h3><p>Convolutional layers, pooling, AlexNet, etc.<\/p><h3 id=\"case-study-recurrent-neural-networks-for-speech-and-text\">Case Study: Recurrent Neural Networks for Speech and Text<\/h3><p>Vanilla Recurrent nets, bi-directional recurrent nets. Maybe overview of LSTM<\/p><h3 id=\"case-study-word2vec\">Case Study: Word2Vec<\/h3><p>Training word vector representations in NLP<\/p><h3 id=\"case-study-t-sne\">Case Study: t-SNE<\/h3><p>Training embeddings for visualizing data<\/p><h2 id=\"acknowledgements\">Acknowledgements<\/h2><p>Thanks a lot to the following people who made this guide better: wodenokoto (HN), zackmorris (HN).<\/p><p>This guide is a work in progress and I appreciate feedback, especially regarding parts that were unclear or only made half sense. Thank you!<\/p><p>Some of the Javascript code in this tutorial has been translated to Python by Ajit, find it over on <a href=\"https:\/\/github.com\/urwithajit9\/HG_NeuralNetwork\">Github<\/a>.<\/p>","created_at":"2019-01-07T04:08:03+0000","updated_at":"2019-01-07T04:08:03+0000","annotations":[],"mimetype":"text\/html","reading_time":67,"domain_name":"karpathy.github.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":766,"title":"AWS re:Invent 2018: Amazon DynamoDB Deep Dive: Advanced Design Patterns for DynamoDB (DAT401)","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=HaEPXoXVf2k","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/HaEPXoXVf2k?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2019-01-05T03:36:04+0000","updated_at":"2019-01-05T03:36:04+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/HaEPXoXVf2k\/hqdefault.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":765,"title":"Coconut Programming Language","url":"http:\/\/coconut-lang.org\/","content":"Coconut Programming Language\n<p>\n  \n  <a href=\"https:\/\/github.com\/evhub\/coconut\" class=\"btn\">GitHub<\/a>\n  <a href=\"http:\/\/coconut.readthedocs.io\/en\/master\/HELP.html\" class=\"btn\">Tutorial<\/a>\n  <a href=\"http:\/\/coconut.readthedocs.io\/en\/master\/DOCS.html\" class=\"btn\">Documentation<\/a>\n  <a href=\"http:\/\/coconut.readthedocs.io\/en\/master\/FAQ.html\" class=\"btn\">FAQ<\/a><\/p><section class=\"interpreter\"><section class=\"main-content\"><p>Coconut is a <strong>functional programming language<\/strong> that <strong>compiles to <a href=\"https:\/\/www.python.org\/\">Python<\/a><\/strong>. Since <strong>all valid Python is valid Coconut<\/strong>, using Coconut will only extend and enhance what you're already capable of in Python.<\/p><p>Why use Coconut? Coconut is built to be useful. Coconut enhances the repertoire of Python programmers to include the tools of modern functional programming, in such a way that those tools are <strong>easy<\/strong> to use and immensely <strong>powerful<\/strong>; that is, Coconut does to functional programming what Python did to imperative programming. And Coconut code runs the same on <strong>any Python version<\/strong>, making the Python 2\/3 split a thing of the past.<\/p><p>Installing Coconut is as easy as<\/p><ol><li>installing <a href=\"https:\/\/www.python.org\/downloads\/\">Python<\/a>,<\/li>\n<li>opening a command-line prompt,<\/li>\n<li>and entering:<\/li>\n<\/ol><p>pip install coconut<\/p><p>which will give you access to all the features of Coconut, which adds to Python <strong>built-in, syntactical support<\/strong> for:<\/p><ul><li>pipeline-style programming<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\n\"hello, world!\" |&gt; print\n<\/pre><\/div><\/div><ul class=\"simple\"><li>prettier lambdas<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\nx -&gt; x ** 2\n<\/pre><\/div><\/div><ul class=\"simple\"><li>partial application<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\nrange(10) |&gt; map$(pow$(?, 2)) |&gt; list\n<\/pre><\/div><\/div><ul class=\"simple\"><li>pattern-matching<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\nmatch [head] + tail in [0, 1, 2, 3]:\n    print(head, tail)\n<\/pre><\/div><\/div><ul class=\"simple\"><li>destructuring assignment<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\n{\"list\": [0] + rest} = {\"list\": [0, 1, 2, 3]}\n<\/pre><\/div><\/div><ul class=\"simple\"><li>infix notation<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\n5 `mod` 3 == 2\n<\/pre><\/div><\/div><ul class=\"simple\"><li>operator functions<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\nproduct = reduce$(*)\n<\/pre><\/div><\/div><ul class=\"simple\"><li>function composition<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\n(f..g..h)(x, y, z)\n<\/pre><\/div><\/div><ul class=\"simple\"><li>lazy lists<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\n(| first_elem() |) :: rest_elems()\n<\/pre><\/div><\/div><ul class=\"simple\"><li>parallel programming<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\nrange(100) |&gt; parallel_map$(pow$(2)) |&gt; list\n<\/pre><\/div><\/div><ul class=\"simple\"><li>tail call optimization<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\ndef factorial(n, acc=1):\n    case n:\n        match 0:\n            return acc\n        match _ is int if n &gt; 0:\n            return factorial(n-1, acc*n)\n<\/pre><\/div><\/div><ul class=\"simple\"><li>algebraic data types<\/li>\n<\/ul><div class=\"highlight-coconut\"><div class=\"highlight\"><pre>\ndata Empty()\ndata Leaf(n)\ndata Node(l, r)\ndef size(Empty()) = 0\n@addpattern(size)\ndef size(Leaf(n)) = 1\n@addpattern(size)\ndef size(Node(l, r)) = size(l) + size(r)\n<\/pre><\/div><\/div><p>and much more!<\/p><p>Like what you see? Don't forget to star Coconut on GitHub!<\/p><p><a class=\"github-button\" href=\"https:\/\/github.com\/evhub\/coconut\" data-icon=\"octicon-star\" data-size=\"large\" data-show-count=\"true\" aria-label=\"Star evhub\/coconut on GitHub\">Star<\/a><\/p><p>Ready to get started? Here are some links to help you out:<\/p><ul><li>\n<a href=\"http:\/\/coconut.readthedocs.io\/en\/master\/HELP.html\">Coconut's <strong>tutorial<\/strong><\/a> will guide you through the process of starting to enhance your Python with Coconut in a straightforward, easy-to-follow way.<\/li>\n<li>\n<a href=\"http:\/\/coconut.readthedocs.io\/en\/master\/DOCS.html\">Coconut's <strong>documentation<\/strong><\/a> is an extensive catalog of information on all of Coconut's features for whenever you see something that you want more information about.<\/li>\n<li>\n<a href=\"http:\/\/coconut.readthedocs.io\/en\/master\/FAQ.html\">The Coconut FAQ<\/a> should hopefully answer any questions you might have about who Coconut is built for and whether or not you should use it.<\/li>\n<li>\n<a href=\"https:\/\/github.com\/evhub\/coconut\/issues\/new\">Creating a new issue<\/a> is the best way for you to get help if you're having a problem with Coconut\u2014just detail the problem in the issue and it will be addressed as soon as possible.<\/li>\n<li>\n<a href=\"https:\/\/gitter.im\/evhub\/coconut\">Coconut's chat room<\/a> is a great place if you want to pose any general questions, concerns, or comments you have about Coconut to other Coconut developers.<\/li>\n<\/ul><\/section><\/section>","created_at":"2019-01-03T14:26:02+0000","updated_at":"2019-01-03T14:26:02+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"coconut-lang.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":764,"title":"openage - free strategy game engine","url":"https:\/\/openage.sft.mx\/","content":"<p>You know Age of Empires. You know its bugs. You know the fun playing it.<\/p><p>\n                    <a href=\"https:\/\/openage.sft.mx\/pics\/openage_2.png\">\n                        <img src=\"https:\/\/openage.sft.mx\/pics\/openage_2.png\" alt=\"epic picture here\" class=\"contentpic\" \/><\/a>\n                <\/p><a href=\"https:\/\/blog.openage.sft.mx\">Blog<\/a><a href=\"https:\/\/github.com\/SFTtech\/openage\">Code<\/a><a href=\"https:\/\/github.com\/SFTtech\/openage\/blob\/master\/doc\/status.md\">Feature list<\/a><a href=\"https:\/\/github.com\/SFTtech\/openage\/blob\/master\/doc\/building.md\">Build and install<\/a><p>\n                    We are creating a free (as in freedom) clone of the Age of Empires II engine in C++17 from scratch (GPLv3+).\n                    We use \"modern\" technologies like OpenGL\/GLSL, Python, Qt5, CMake, ...\n                <\/p><p>\n                    We aim to create an exact clone (for now), using the original media files.\n                <\/p><p>\n                    This allows people using <b>GNU\/Linux<\/b>, <b>BSD<\/b>, <b>macOS<\/b> or <b>Windows<\/b> to play the game\n                    natively, minus features, plus additional epic bugs (or was it vice versa?).\n                <\/p><p>\n                    The project is run by volunteers:<br \/>Students and professionals from all over the world.<br \/>Join us and help to create an awesome game!\n                <\/p><ul><li><strong>Matrix<\/strong>: <a href=\"https:\/\/riot.im\/app\/#\/room\/#sfttech:matrix.org\">#sfttech:matrix.org<\/a><\/li>\n                    <li><strong>IRC<\/strong>: <a href=\"https:\/\/webchat.freenode.net\/?channels=sfttech\">#sfttech on freenode.org<\/a><\/li>\n                    <li><strong>Reddit<\/strong>: <a href=\"https:\/\/reddit.com\/r\/openage\">\/r\/openage<\/a><\/li>\n                    <li><strong>Money sink<\/strong>: <a href=\"https:\/\/liberapay.com\/SFTtech\">liberapay\/SFTtech<\/a><\/li>\n                <\/ul>","created_at":"2018-12-29T00:50:23+0000","updated_at":"2018-12-29T00:50:23+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"openage.sft.mx","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":763,"title":"mogenslund\/liquid","url":"https:\/\/github.com\/mogenslund\/liquid\/blob\/master\/README.md","content":"<p><a href=\"https:\/\/clojars.org\/mogenslund\/liquid\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/01a1a13f5c831cb73e4111e1a56fa169812bf325\/68747470733a2f2f696d672e736869656c64732e696f2f636c6f6a6172732f762f6d6f67656e736c756e642f6c69717569642e737667\" alt=\"Clojars Project\" data-canonical-src=\"https:\/\/img.shields.io\/clojars\/v\/mogenslund\/liquid.svg\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/mogenslund\/liquid\/blob\/master\/logo.png\"><img src=\"https:\/\/github.com\/mogenslund\/liquid\/raw\/master\/logo.png\" width=\"281\" height=\"360\" alt=\"image\" style=\"float: left;\" \/><\/a><\/p>\n<ul><li><b>Fluidable<\/b>  -  Embed your code into \u03bbiquid or \u03bbiquid into your code<\/li>\n<li><b>Dependency free<\/b>  -  Everything is pure Clojure<\/li>\n<li><b>Translateable<\/b>  -  Tiny core, simple model, no tricks<\/li>\n<li><b>Destraction Free<\/b>  -  Typeahead and completions on demand... Your demand!<\/li>\n<li><b>Comfortable keybindings<\/b>  -  Most used commands are close at hand<\/li>\n<li><b>Servable<\/b>  -  Run as server, local or remote<\/li>\n<\/ul><h2><a id=\"user-content-website\" class=\"anchor\" aria-hidden=\"true\" href=\"#website\"><\/a>Website<\/h2>\n<p>Find more information on <a href=\"http:\/\/salza.dk\" rel=\"nofollow\">salza.dk<\/a><\/p>\n<h2><a id=\"user-content-who-is-this-for\" class=\"anchor\" aria-hidden=\"true\" href=\"#who-is-this-for\"><\/a>Who is this for<\/h2>\n<p>\u03bbiquid is absolutely not for everyone! It is for you if<\/p>\n<ul><li>You get a kick of doing everything in Clojure<\/li>\n<li>You love to tinker with your editor<\/li>\n<li>Terminals are the best<\/li>\n<li>You love distractionfree power<\/li>\n<li>You love modal editing, like in vim<\/li>\n<\/ul><h2><a id=\"user-content-what-can-you-do-with-it\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-can-you-do-with-it\"><\/a>What can you do with it<\/h2>\n<p>There is no made up language or syntax to extend \u03bbiquid.<\/p>\n<ul><li>Since \u03bbiquid is pure Clojure it can be used to extend any Clojure or Java program, so the editor lives inside the program.. like magic.<\/li>\n<li>Turn it around, you can use any Clojure and any Java library to extend \u03bbiquid.. if the library can do it then \u03bbiquid can do it. Configure keyboard shortcuts and commands to tame the libraries. Use code to integrate.<\/li>\n<\/ul><h2><a id=\"user-content-demo-video\" class=\"anchor\" aria-hidden=\"true\" href=\"#demo-video\"><\/a>Demo video<\/h2>\n<p>Watch these videos to get a sense of what it can and how the editor works:<\/p>\n<p><a href=\"https:\/\/www.youtube.com\/watch?v=B3pYpnYlXgU&amp;feature=youtu.be\" rel=\"nofollow\">Demo<\/a><\/p>\n<p><a href=\"https:\/\/drive.google.com\/file\/d\/0B5Au3PxHCmgBM2MySHl4aUVWTEE\/view?usp=sharing\" rel=\"nofollow\">liquid01<\/a><\/p>\n<h2><a id=\"user-content-community\" class=\"anchor\" aria-hidden=\"true\" href=\"#community\"><\/a>Community<\/h2>\n<p>For discussions, help, tips and support, use<\/p>\n<p><strong><a href=\"https:\/\/clojurians.slack.com\/messages\/liquid\/\" rel=\"nofollow\">#liquid<\/a><\/strong> channel in <a href=\"http:\/\/clojurians.net\/\" rel=\"nofollow\">Clojure Slack<\/a><\/p>\n<h2><a id=\"user-content-some-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#some-extensions\"><\/a>Some extensions<\/h2>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<h3><a id=\"user-content-download-and-execution\" class=\"anchor\" aria-hidden=\"true\" href=\"#download-and-execution\"><\/a>Download and execution<\/h3>\n<h4><a id=\"user-content-try-it-with-clj\" class=\"anchor\" aria-hidden=\"true\" href=\"#try-it-with-clj\"><\/a>Try it with clj<\/h4>\n<p>You can run from command line using the <code>clj<\/code> command:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>clj -Sdeps '{:deps {mogenslund\/liquid {:mvn\/version \"1.0.0\"}}}' -m dk.salza.liq.core<\/pre><\/div>\n<h4><a id=\"user-content-quick-start\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-start\"><\/a>Quick start<\/h4>\n<p>Download the single jar (Uberjar containing both Clojure and Liquid) <a href=\"http:\/\/salza.dk\/liq.jar\" rel=\"nofollow\">liq.jar<\/a><\/p>\n<p><strong>IMPORTANT<\/strong> Please have the <a href=\"http:\/\/salza.dk\/liquid\/cheatsheet.html\" rel=\"nofollow\">Cheatsheet<\/a> nearby when using Liquid for the first time. The keybindings are powerful, but needs to be learned. <strong>You will get lost without it<\/strong><\/p>\n<p>On Windows double click the liq.jar.<br \/>On Linux and Mac execute the following in a terminal:<\/p>\n<pre>java -jar liq.jar\n<\/pre>\n<p>If the terminal causes problems, like hard to see colors or artifacts when navigating, try the JFrame implementation:<\/p>\n<pre>java -jar liq.jar --jframe\n<\/pre>\n<p>My experience is that the terminal \"Terminology\", with dark background works very well.<\/p>\n<h4><a id=\"user-content-starter-kit\" class=\"anchor\" aria-hidden=\"true\" href=\"#starter-kit\"><\/a>Starter kit<\/h4>\n<p>A place to start is <a href=\"https:\/\/github.com\/mogenslund\/liquid-starter-kit\">Liquid Starter Kit<\/a><\/p>\n<p>It sort of corresponds to a sample <code>.emacs<\/code> file. Some code you can extend to create your own version of the editor.<\/p>\n<h4><a id=\"user-content-local-setup-and-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#local-setup-and-extensions\"><\/a>Local setup and extensions<\/h4>\n<p>For a recommended approach to local setup and managing extensions, please read<\/p>\n<p><a href=\"http:\/\/salza.dk\/setupbasic.html\" rel=\"nofollow\">Basic Setup<\/a><\/p>\n<h4><a id=\"user-content-salza-\u03bbiquid-text-editor-is-designed-with-clojure-developers-in-mind\" class=\"anchor\" aria-hidden=\"true\" href=\"#salza-\u03bbiquid-text-editor-is-designed-with-clojure-developers-in-mind\"><\/a>Salza \u03bbiquid text editor is designed with clojure developers in mind<\/h4>\n<p>The editor is written in pure Clojure. That is, there are no other dependencies, not even curses or lanterna!<\/p>\n<p>I have been using Emacs for many years. I have tweaked it a lot and implemented modes, to simulate the Vim way of switching between insert mode and normal mode for navigation. I like being able to use the whole keyboard for navigation. As you will see in the video below, I am very inspired by Emacs and Vim, but I have choosen not to try to copy or reimplement Emacs or Vim. I have only stolen the features that I need or like. I hope others will just create plugins or extensions to make the editor support other features that they like.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/mogenslund\/liquid\/blob\/master\/liquid01.png\"><img src=\"https:\/\/github.com\/mogenslund\/liquid\/raw\/master\/liquid01.png\" alt=\"Liquid Text Editor\" \/><\/a><\/p>\n<p>I could tell a lot about the editor, but I think you should see the video below for a demonstration.<\/p>\n<h2><a id=\"user-content-technical-video-explaining-the-core-datastructure-of-the-editor\" class=\"anchor\" aria-hidden=\"true\" href=\"#technical-video-explaining-the-core-datastructure-of-the-editor\"><\/a>Technical video explaining the core datastructure of the editor<\/h2>\n<p>This video is for technical minded people to get a view of the inside of the construction of the editor. Useful for understanding the idea behind the editor or as inspiration for similar projects or extensions of this: <a href=\"https:\/\/drive.google.com\/file\/d\/0B5Au3PxHCmgBNjJodzNjQ2w4c2M\/view?usp=sharing\" rel=\"nofollow\">slider01<\/a><\/p>\n<h2><a id=\"user-content-use-cases\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-cases\"><\/a>Use cases<\/h2>\n<ul><li>\n<p><b>Tight integration<\/b> with Clojure code. It should be easy to have the same code executed from within the editor as well as outside. Using a real language for extensions provides a great advantage compared to a \"home made\" extension language, that only fits the specific editor.<br \/>Example: As a QA Engineer I do test automation using Selenium WebDriver. I have included the Selenium jar into my local environment, so now I can execute tests, parts of tests or snippets from within \u03bbiquid.<br \/>All other tools that I create in Clojure, which are useful as is, are easy to make accessible from within the editor as well.<\/p>\n<\/li>\n<li>\n<p><b>Embedding<\/b> the editor into your your application. Include the \u03bbiquid to your project and use it as part of your program or for debugging and patching, just like you already do with the REPL, but with some advantages: It is easier to open a file with code and execute parts of it, or jump between snippets, do modifications, and execute again.<br \/>Example: I have a file with snippets like: <code>(patch \"myserver\")<\/code>, <code>(status \"myserver\")<\/code> and <code>(run-test \"name-of-test\")<\/code>. I just modify \"myserver\", if needed, and press \"e\" to execute the command.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>Dependencies<\/h2>\n<p>Right now the only dependency is Clojure itself. I would like to keep it that way, as for core functionality. When starting \u03bbiquid, any resources can just be added to the classpath, and the code will be available from inside \u03bbiquid. So it is not a limitation to \u03bbiquid, it just means that anything requiring other dependencies should be implemented as plugins and be loaded together with \u03bbiquid. Extensions can be loaded through a .liq file in the home folder.<\/p>\n<h2><a id=\"user-content-inspiration-from-emacs-and-vi\" class=\"anchor\" aria-hidden=\"true\" href=\"#inspiration-from-emacs-and-vi\"><\/a>Inspiration from Emacs and Vi<\/h2>\n<p>From Emacs I have been inspired by extensibility. Clojure is also the language for writing extensions. That was in fact a primary motivation for creating the editor. Actually it is possible to include the editor into a project and make the editor part of the program. This thought is very similar to the REPL (read-eval-print-loop), but more like an edit-eval-render-loop.<\/p>\n<p>From Vi I have stolen the concept of normal\/insert mode. I really like the ability to use the whole keyboard to navigate the file and execute commands at one time, while typing text in another. The tab key is used to switch between the two modes. The color of the cursor will indicate which mode is in use. I have resisted the attempt to use the same keybindings as Vi. Some are the same, but as default I have chosen the keys jkli as navigation keys. They are located where the fingers are placed and layed out like the arrow keys. Look at jkli on the keyboard. Different up and down operations are mapped to different combinations with \"i\" and \"k\" (up and down), while different sideway actions are mapped to \"j\" and \"l\". E.g going to the end of the line is done with shift+l and the beginning of the line with shift+j.<\/p>\n<p>S-expressions in clojure can be evaluated with \"e\", while the whole file is evaluated with shift+e.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2018 Mogens Br\u00f8dsgaard Lund<\/p>\n<p>Distributed under the Eclipse Public License either version 1.0 or any later version.<\/p>","created_at":"2018-12-27T23:43:32+0000","updated_at":"2018-12-27T23:43:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/7754224?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":762,"title":"piprate\/json-gold","url":"https:\/\/github.com\/piprate\/json-gold","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>This library is an implementation of the <a href=\"http:\/\/json-ld.org\/\" rel=\"nofollow\">JSON-LD 1.0<\/a> specification in Go.\nIt supports both URDNA2015 and URGNA2012 RDF dataset normalisation algorithms.<\/p>\n<p>The <a href=\"https:\/\/github.com\/kazarena\/json-gold\">original library<\/a> was written by Stan Nazarenko\n(<a href=\"https:\/\/github.com\/kazarena\">@kazarena<\/a>). See the full list of contributors\n<a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/CONTRIBUTORS.md\">here<\/a>.<\/p>\n<h2><a id=\"user-content-testing--compliance\" class=\"anchor\" aria-hidden=\"true\" href=\"#testing--compliance\"><\/a>Testing &amp; Compliance<\/h2>\n<p>As of December 26, 2018:<\/p>\n<ul><li>all JSON-LD 1.0 tests from the <a href=\"https:\/\/github.com\/json-ld\/json-ld.org\/tree\/master\/test-suite\">official JSON-LD test suite<\/a> pass, with one exception: <a href=\"https:\/\/github.com\/melville-wiley\/json-ld.org\/blob\/3461fd0005cd8e338cd3729c4714163e9217e619\/test-suite\/tests\/frame-manifest.jsonld#L530\">framing test #tg010<\/a>. See the discussion <a href=\"https:\/\/github.com\/json-ld\/json-ld.org\/pull\/663\">here<\/a>. It appears that while it's meant to be a 1.0 test, the fix in pyLD was implemented in the 1.1 section of the framing code. This needs further investigation.<\/li>\n<li>all RDF Dataset Normalisation tests from the <a href=\"https:\/\/json-ld.github.io\/normalization\/tests\/index.html\" rel=\"nofollow\">current test suite<\/a> pass<\/li>\n<li>JSON-LD 1.1 spec is not currenty supported<\/li>\n<\/ul><h2><a id=\"user-content-inspiration\" class=\"anchor\" aria-hidden=\"true\" href=\"#inspiration\"><\/a>Inspiration<\/h2>\n<p>This implementation was heavily influenced by <a href=\"https:\/\/github.com\/jsonld-java\/jsonld-java\">JSONLD-Java<\/a> with some techniques borrowed from <a href=\"https:\/\/github.com\/digitalbazaar\/pyld\">PyLD<\/a> and <a href=\"https:\/\/github.com\/linkeddata\/gojsonld\">gojsonld<\/a>. Big thank you to the contributors of the forementioned libraries for figuring out implementation details of the core algorithms.<\/p>\n<h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<h3><a id=\"user-content-expand\" class=\"anchor\" aria-hidden=\"true\" href=\"#expand\"><\/a>Expand<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/expand.go\">examples\/expand.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\n\/\/ expanding remote document\nexpanded, err := proc.Expand(\"http:\/\/json-ld.org\/test-suite\/tests\/expand-0002-in.jsonld\", options)\nif err != nil {\n\tlog.Println(\"Error when expanding JSON-LD document:\", err)\n\treturn\n}\n\/\/ expanding in-memory document\ndoc := map[string]interface{}{\n\t\"@context\": \"http:\/\/schema.org\/\",\n\t\"@type\": \"Person\",\n\t\"name\": \"Jane Doe\",\n\t\"jobTitle\": \"Professor\",\n\t\"telephone\": \"(425) 123-4567\",\n\t\"url\": \"http:\/\/www.janedoe.com\",\n}\nexpanded, err = proc.Expand(doc, options)<\/pre><\/div>\n<h3><a id=\"user-content-compact\" class=\"anchor\" aria-hidden=\"true\" href=\"#compact\"><\/a>Compact<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/compact.go\">examples\/compact.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\ndoc := map[string]interface{}{\n\t\"@id\": \"http:\/\/example.org\/test#book\",\n\t\"http:\/\/example.org\/vocab#contains\": map[string]interface{}{\n\t\t\"@id\": \"http:\/\/example.org\/test#chapter\",\n\t},\n\t\"http:\/\/purl.org\/dc\/elements\/1.1\/title\": \"Title\",\n}\ncontext := map[string]interface{}{\n\t\"@context\": map[string]interface{}{\n\t\t\"dc\": \"http:\/\/purl.org\/dc\/elements\/1.1\/\",\n\t\t\"ex\": \"http:\/\/example.org\/vocab#\",\n\t\t\"ex:contains\": map[string]interface{}{\n\t\t\t\"@type\": \"@id\",\n\t\t},\n\t},\n}\ncompactedDoc, err := proc.Compact(doc, context, options)<\/pre><\/div>\n<h3><a id=\"user-content-flatten\" class=\"anchor\" aria-hidden=\"true\" href=\"#flatten\"><\/a>Flatten<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/flatten.go\">examples\/flatten.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\ndoc := map[string]interface{}{\n\t\"@context\": []interface{}{\n\t\tmap[string]interface{}{\n\t\t\t\"name\": \"http:\/\/xmlns.com\/foaf\/0.1\/name\",\n\t\t\t\"homepage\": map[string]interface{}{\n\t\t\t\t\"@id\": \"http:\/\/xmlns.com\/foaf\/0.1\/homepage\",\n\t\t\t\t\"@type\": \"@id\",\n\t\t\t},\n\t\t},\n\t\tmap[string]interface{}{\n\t\t\t\"ical\": \"http:\/\/www.w3.org\/2002\/12\/cal\/ical#\",\n\t\t},\n\t},\n\t\"@id\": \"http:\/\/example.com\/speakers#Alice\",\n\t\"name\": \"Alice\",\n\t\"homepage\": \"http:\/\/xkcd.com\/177\/\",\n\t\"ical:summary\": \"Alice Talk\",\n\t\"ical:location\": \"Lyon Convention Centre, Lyon, France\",\n}\nflattenedDoc, err := proc.Flatten(doc, nil, options)<\/pre><\/div>\n<h3><a id=\"user-content-frame\" class=\"anchor\" aria-hidden=\"true\" href=\"#frame\"><\/a>Frame<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/frame.go\">examples\/frame.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\ndoc := map[string]interface{}{\n\t\"@context\": map[string]interface{}{\n\t\t\"dc\": \"http:\/\/purl.org\/dc\/elements\/1.1\/\",\n\t\t\"ex\": \"http:\/\/example.org\/vocab#\",\n\t\t\"ex:contains\": map[string]interface{}{\"@type\": \"@id\"},\n\t},\n\t\"@graph\": []interface{}{\n\t\tmap[string]interface{}{\n\t\t\t\"@id\": \"http:\/\/example.org\/test\/#library\",\n\t\t\t\"@type\": \"ex:Library\",\n\t\t\t\"ex:contains\": \"http:\/\/example.org\/test#book\",\n\t\t},\n\t\tmap[string]interface{}{\n\t\t\t\"@id\": \"http:\/\/example.org\/test#book\",\n\t\t\t\"@type\": \"ex:Book\",\n\t\t\t\"dc:contributor\": \"Writer\",\n\t\t\t\"dc:title\": \"My Book\",\n\t\t\t\"ex:contains\": \"http:\/\/example.org\/test#chapter\",\n\t\t},\n\t\tmap[string]interface{}{\n\t\t\t\"@id\": \"http:\/\/example.org\/test#chapter\",\n\t\t\t\"@type\": \"ex:Chapter\",\n\t\t\t\"dc:description\": \"Fun\",\n\t\t\t\"dc:title\": \"Chapter One\",\n\t\t},\n\t},\n}\nframe := map[string]interface{}{\n\t\"@context\": map[string]interface{}{\n\t\t\"dc\": \"http:\/\/purl.org\/dc\/elements\/1.1\/\",\n\t\t\"ex\": \"http:\/\/example.org\/vocab#\",\n\t},\n\t\"@type\": \"ex:Library\",\n\t\"ex:contains\": map[string]interface{}{\n\t\t\"@type\": \"ex:Book\",\n\t\t\"ex:contains\": map[string]interface{}{\n\t\t\t\"@type\": \"ex:Chapter\",\n\t\t},\n\t},\n}\nframedDoc, err := proc.Frame(doc, frame, options)<\/pre><\/div>\n<h3><a id=\"user-content-to-rdf\" class=\"anchor\" aria-hidden=\"true\" href=\"#to-rdf\"><\/a>To RDF<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/to_rdf.go\">examples\/to_rdf.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\noptions.Format = \"application\/nquads\"\n\/\/ this JSON-LD document was taken from http:\/\/json-ld.org\/test-suite\/tests\/toRdf-0028-in.jsonld\ndoc := map[string]interface{}{\n\t\"@context\": map[string]interface{}{\n\t\t\"sec\":        \"http:\/\/purl.org\/security#\",\n\t\t\"xsd\":        \"http:\/\/www.w3.org\/2001\/XMLSchema#\",\n\t\t\"rdf\":        \"http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#\",\n\t\t\"dc\":         \"http:\/\/purl.org\/dc\/terms\/\",\n\t\t\"sec:signer\": map[string]interface{}{\"@type\": \"@id\"},\n\t\t\"dc:created\": map[string]interface{}{\"@type\": \"xsd:dateTime\"},\n\t},\n\t\"@id\":                \"http:\/\/example.org\/sig1\",\n\t\"@type\":              []interface{}{\"rdf:Graph\", \"sec:SignedGraph\"},\n\t\"dc:created\":         \"2011-09-23T20:21:34Z\",\n\t\"sec:signer\":         \"http:\/\/payswarm.example.com\/i\/john\/keys\/5\",\n\t\"sec:signatureValue\": \"OGQzNGVkMzVm4NTIyZTkZDYMmMzQzNmExMgoYzI43Q3ODIyOWM32NjI=\",\n\t\"@graph\": map[string]interface{}{\n\t\t\"@id\":      \"http:\/\/example.org\/fact1\",\n\t\t\"dc:title\": \"Hello World!\",\n\t},\n}\ntriples, err := proc.ToRDF(doc, options)<\/pre><\/div>\n<h3><a id=\"user-content-from-rdf\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-rdf\"><\/a>From RDF<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/from_rdf.go\">examples\/from_rdf.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\ntriples := `\n\t&lt;http:\/\/example.com\/Subj1&gt; &lt;http:\/\/www.w3.org\/1999\/02\/22-rdf-syntax-ns#type&gt; &lt;http:\/\/example.com\/Type&gt; .\n\t&lt;http:\/\/example.com\/Subj1&gt; &lt;http:\/\/example.com\/prop1&gt; &lt;http:\/\/example.com\/Obj1&gt; .\n\t&lt;http:\/\/example.com\/Subj1&gt; &lt;http:\/\/example.com\/prop2&gt; \"Plain\" .\n\t&lt;http:\/\/example.com\/Subj1&gt; &lt;http:\/\/example.com\/prop2&gt; \"2012-05-12\"^^&lt;http:\/\/www.w3.org\/2001\/XMLSchema#date&gt; .\n\t&lt;http:\/\/example.com\/Subj1&gt; &lt;http:\/\/example.com\/prop2&gt; \"English\"@en .\n`\ndoc, err := proc.FromRDF(triples, options)<\/pre><\/div>\n<h3><a id=\"user-content-normalize\" class=\"anchor\" aria-hidden=\"true\" href=\"#normalize\"><\/a>Normalize<\/h3>\n<p>See complete code in <a href=\"https:\/\/github.com\/piprate\/json-gold\/blob\/master\/examples\/normalize.go\">examples\/normalize.go<\/a>.<\/p>\n<div class=\"highlight highlight-source-go\"><pre>proc := ld.NewJsonLdProcessor()\noptions := ld.NewJsonLdOptions(\"\")\noptions.Format = \"application\/nquads\"\noptions.Algorithm = \"URDNA2015\"\ndoc := map[string]interface{}{\n\t\"@context\": map[string]interface{}{\n\t\t\"ex\": \"http:\/\/example.org\/vocab#\",\n\t},\n\t\"@id\": \"http:\/\/example.org\/test#example\",\n\t\"@type\": \"ex:Foo\",\n\t\"ex:embed\": map[string]interface{}{\n\t\t\"@type\": \"ex:Bar\",\n\t},\n}\nnormalizedTriples, err := proc.Normalize(doc, options)<\/pre><\/div>\n<\/article>","created_at":"2018-12-27T04:53:54+0000","updated_at":"2018-12-27T04:53:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/17619914?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":761,"title":"redblobgames\/mapgen4","url":"https:\/\/github.com\/redblobgames\/mapgen4\/blob\/master\/README.org","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/91d65a362d0aa227711f3010a74ff316f17ff18d\/687474703a2f2f756e6d61696e7461696e65642e746563682f62616467652e737667\"><img src=\"https:\/\/camo.githubusercontent.com\/91d65a362d0aa227711f3010a74ff316f17ff18d\/687474703a2f2f756e6d61696e7461696e65642e746563682f62616467652e737667\" alt=\"http:\/\/unmaintained.tech\/badge.svg\" data-canonical-src=\"http:\/\/unmaintained.tech\/badge.svg\" \/><\/a><\/p>\n<p><a href=\"https:\/\/www.redblobgames.com\/maps\/mapgen4\/\" rel=\"nofollow\">Mapgen4<\/a> is a procedural wilderness map generator I wrote during 2017 and 2018. It\u2019s written in JavaScript and designed to:<\/p>\n<ul><li>run <i>fast<\/i> enough to regenerate in real time as you paint terrain<\/li>\n  <li>look <i>pretty<\/i> instead of generating realistic terrain<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/7e352057d8ff35a88f420ba092c581069d1143ea\/68747470733a2f2f7777772e726564626c6f6267616d65732e636f6d2f6d6170732f6d617067656e342f626c6f672f73637265656e73686f7431352d736d616c6c2e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/7e352057d8ff35a88f420ba092c581069d1143ea\/68747470733a2f2f7777772e726564626c6f6267616d65732e636f6d2f6d6170732f6d617067656e342f626c6f672f73637265656e73686f7431352d736d616c6c2e6a7067\" alt=\"https:\/\/www.redblobgames.com\/maps\/mapgen4\/blog\/screenshot15-small.jpg\" data-canonical-src=\"https:\/\/www.redblobgames.com\/maps\/mapgen4\/blog\/screenshot15-small.jpg\" \/><\/a><\/p>\n<p>If you like <code>yarn<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>yarn global add browserify\nyarn install\nmake build\/points-5.data\nyarn build<\/pre><\/div>\n<p>If you like <code>npm<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>npm -g install browserify\nnpm install\nmake build\/points-5.data\nnpm run build<\/pre><\/div>\n<p>Then run <code>python -m SimpleHTTPServer 8000<\/code> and visit http:\/\/localhost:8000\/embed.html in your browser.<\/p>\n<p>I have a <a href=\"http:\/\/simblob.blogspot.com\/search\/label\/mapgen4\" rel=\"nofollow\">series of blog posts<\/a> about how I made these maps:<\/p>\n<ul><li><a href=\"http:\/\/simblob.blogspot.com\/2018\/08\/mapgen4-goals.html\" rel=\"nofollow\">History<\/a> of the project<\/li>\n  <li><a href=\"https:\/\/www.redblobgames.com\/x\/1721-voronoi-alternative\/\" rel=\"nofollow\">Alternative to Voronoi cells<\/a><\/li>\n  <li><a href=\"https:\/\/www.redblobgames.com\/x\/1722-b-rep-triangle-meshes\/\" rel=\"nofollow\">Compact data structure<\/a> for the Delaunay+Voronoi mesh<\/li>\n  <li><a href=\"http:\/\/simblob.blogspot.com\/2018\/08\/mapgen4-elevation.html\" rel=\"nofollow\">Elevation<\/a> to match the desired look instead of tweaking the look to match the elevation<\/li>\n  <li><a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-elevation-painting.html\" rel=\"nofollow\">Distance fields for elevation<\/a><\/li>\n  <li><a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-threads.html\" rel=\"nofollow\">Multithreading<\/a> to make it run acceptably fast<\/li>\n  <li>Fixing the <a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-river-appearance.html\" rel=\"nofollow\">appearance of rivers<\/a><\/li>\n  <li>Revisiting <a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-elevation-painting-revisited.html\" rel=\"nofollow\">distance fields<\/a>, which didn\u2019t work out like I hoped<\/li>\n  <li><a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-rainfall.html\" rel=\"nofollow\">Rainfall<\/a>, biomes, evaporation, wind<\/li>\n  <li>Rendering with an <a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-oblique-projection.html\" rel=\"nofollow\">oblique projection<\/a>, not the standard rotate+translate+scale<\/li>\n  <li>Rendering <a href=\"http:\/\/simblob.blogspot.com\/2018\/10\/mapgen4-outlines.html\" rel=\"nofollow\">outlines<\/a><\/li>\n  <li>Some <a href=\"http:\/\/simblob.blogspot.com\/2018\/09\/mapgen4-bug-fixes.html\" rel=\"nofollow\">bug fixes<\/a><\/li>\n  <li>River <a href=\"https:\/\/www.redblobgames.com\/x\/1723-procedural-river-growing\/\" rel=\"nofollow\">generation<\/a> and <a href=\"http:\/\/simblob.blogspot.com\/2018\/10\/mapgen4-river-representation.html\" rel=\"nofollow\">data structures<\/a><\/li>\n<\/ul><p>It\u2019s a continuation of ideas I developed for <a href=\"https:\/\/github.com\/amitp\/mapgen2\/\">mapgen2<\/a> back in 2010, but at a much larger scale. The underlying code can support 1 million+ Voronoi cells (change <code>spacing<\/code> in config.js to 0.7), including <a href=\"https:\/\/www.redblobgames.com\/maps\/mapgen4\/blog\/3565944-triangles-600kregions.png\" rel=\"nofollow\">a very detailed river network<\/a>, but the rendering code and other parameters are designed to look prettiest around 25k cells.<\/p>\n<p>There\u2019s plenty more that could be done to make it even faster and prettier. There are plenty of features that could be added, such as <a href=\"https:\/\/www.redblobgames.com\/x\/1723-procedural-river-growing\/#draw\" rel=\"nofollow\">drawing your own rivers<\/a>, <a href=\"https:\/\/www.redblobgames.com\/x\/1843-planet-generation\/\" rel=\"nofollow\">sphere output<\/a>, <a href=\"https:\/\/www.redblobgames.com\/x\/1736-resource-placement\/\" rel=\"nofollow\">natural resources<\/a>, towns, forests, names, roads, and nations, but I\u2019m leaving those for a future project.<\/p>\n<p>Mapgen4 and helper libraries I wrote (dual-mesh, prng) are licensed under Apache v2. You can use this code in your own project, including commercial projects.<\/p>\n<p>The map generator uses these libraries:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/mapbox\/delaunator.git\">Delaunator<\/a> from MapBox is licensed under the ISC license.<\/li>\n  <li><a href=\"https:\/\/github.com\/kchapelier\/poisson-disk-sampling\">poisson-disk-sampling<\/a> from Kevin Chapelier is licensed under the MIT license.<\/li>\n  <li><a href=\"https:\/\/github.com\/jwagner\/simplex-noise.js\">simple-noise<\/a> from Jonas Wagner is licensed under the MIT license.<\/li>\n  <li><a href=\"https:\/\/github.com\/mourner\/flatqueue\">flatqueue<\/a> from Vladimir Agafonkin is licensed under the ISC license.<\/li>\n  <li><a href=\"https:\/\/github.com\/mikolalysenko\/hash-int\">hash-int<\/a> from Mikola Lysenko is licensed under the MIT license.<\/li>\n<\/ul><p>The rendering code uses these libraries:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/regl-project\/regl\">regl<\/a> from Mikola Lysenko is licensed under the MIT license.<\/li>\n  <li><a href=\"https:\/\/github.com\/toji\/gl-matrix\">gl-matrix<\/a> from Brandon Jones, Colin MacKenzie IV is licensed under the MIT license.<\/li>\n<\/ul><p>The build step uses:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/browserify\/browserify\">Browserify<\/a> + <a href=\"https:\/\/github.com\/substack\/webworkify\">Webworkify<\/a> are licensed under the MIT license.<\/li>\n  <li><a href=\"https:\/\/github.com\/hughsk\/uglifyify\">Uglifyify<\/a> from Hugh Kennedy is licensed under the MIT license.<\/li>\n<\/ul>","created_at":"2018-12-26T11:49:12+0000","updated_at":"2018-12-26T11:49:12+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/8757369?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":760,"title":"My one-liner Linux Dropbox client","url":"http:\/\/lpan.io\/one-liner-dropbox-client\/","content":"<p>In this blog post, I want to discuss one of my recent attempts to create a\nsimple one-liner Linux Dropbox client using only free and open-source\ncomponents, including <a href=\"https:\/\/rclone.org\/\">rclone<\/a>,\n<a href=\"http:\/\/eradman.com\/entrproject\/\">entr<\/a>, and\n<a href=\"https:\/\/www.freedesktop.org\/wiki\/Software\/systemd\/\">systemd<\/a>.<\/p><p>Recently, the proprietary Dropbox Linux client dropped support for all Linux\nfile systems except <em>unencrypted ext4<\/em>. And, my home directory is\n\u201cunfortunately\u201d encrypted.<\/p>\n<p>In early December, the proprietary Dropbox Linux client I have been using\nstopped working. It logged me out, and prompted me to choose a different sync\nfolder on a \u201csupported file system\u201d.<\/p>\n<p>By the way, I run Ubuntu bionic on a two-year-old Thinkpad t460s.<\/p>\n<p>I am a heavy <a href=\"https:\/\/orgmode.org\/\">Org mode<\/a> user. I take notes in plain text,\nand my use case of Dropbox is simply \u201ccontinuously backing up my notes while I\ntype\u201d.<\/p>\n<p>If you are also into data infrastructure, my use case is very similar to\n\u201casynchronous single-master replication\u201d. All writes go through my Thinkpad, the\nmaster. The remote Dropbox folder is just a read-only follower that I\noccasionally \u201cissue read-only queries\u201d to, or used as a backup to construct a\nnew master when the current master fails or gets stolen.<\/p>\n<p>Nevertheless, this replication setup has saved my life multiple times. I still\nremember, very vividly, that my Thinkpad couldn\u2019t boot during the exam season at\nthe end of my sophomore year. Since I continuously replicated all my notes to\nDropbox, I didn\u2019t lose any data and I was able to view my latest notes on my\nmom\u2019s Macbook. Thanks mom!<\/p>\n<p>When the Dropbox client stopped working, my main focus was to find another\nsimilarly feature-rich remote storage client for Linux. I also wouldn\u2019t mind\nmigrating to another storage back-end, such as Google Drive or AWS S3. Some of\nthe other options I considered include\n<a href=\"https:\/\/www.thefanclub.co.za\/overgrive\">overGrive<\/a> and\n<a href=\"https:\/\/www.insynchq.com\/\">insync<\/a>.<\/p>\n<p>However, I concluded that these solutions are way too feature-rich, and not\n<strong>well-suited for my use case<\/strong>.<\/p>\n<p>For example, these clients are modelled as <em>mounting a remote file system onto\nyour file system<\/em>. They try really hard to abstract the remote file systems away\nby making them look and feel like local file systems. They typically implement\ntwo-way syncing, automatically mapping remote file types to Linux file types,\netc.<\/p>\n<p>I don\u2019t need this level of abstraction. I just need something simple that allows\nme to back up my notes to the cloud continuously while I type. In addition, the\nabstraction also makes it harder to set up and debug. Not to mention that most\nof these \u201cfeature-rich\u201d clients are proprietary.<\/p>\n<p>I ran into <a href=\"https:\/\/rclone.org\/\">rclone<\/a> and realized that it is exactly what I\nwas looking for. <code class=\"highlighter-rouge\">rclone<\/code> is simple but powerful. It is very similar to the\n<code class=\"highlighter-rouge\">rsync<\/code> tool, but for \u201ccloud storage\u201d.<\/p>\n<p>For example, it not only takes care of fault tolerance (integrity checks),\nefficient synchronization algorithms, etc., but also provides <a href=\"https:\/\/github.com\/ncw\/rclone\/blob\/6b1f915ebccdf232cb128540ba67098b754282d6\/fs\/fs.go#L210-L244\">a simple CRUD\ninterface<\/a>\nto interact with popular cloud storage services, including Amazon S3, Google\nDrive, and <a href=\"https:\/\/rclone.org\/dropbox\/\">Dropbox<\/a>.<\/p>\n<p>The following command will make the remote directory <code class=\"highlighter-rouge\">org<\/code> identical with the\nlocal directory <code class=\"highlighter-rouge\">\/home\/lpan\/org<\/code><\/p>\n<div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>ORG_DIR=\/home\/lpan\/org\nREMOTE=dropbox\nrclone sync $ORG_DIR $REMOTE:org\n<\/pre><\/div><\/div>\n<p><a href=\"http:\/\/eradman.com\/entrproject\/\">entr<\/a> is a command runner that leverages the\n<a href=\"http:\/\/man.he.net\/?section=all&amp;topic=inotify\">inotify<\/a> API. It basically allows\nyou to run a command triggered by file changes, without <em>polling<\/em> the file\nsystem.<\/p>\n<p>One of its common usages is to <em>rebuild the project when any of the source file\nchanges<\/em>.<\/p>\n<p><code class=\"highlighter-rouge\">entr<\/code> takes a list of absolute paths from <code class=\"highlighter-rouge\">stdin<\/code>, and then it will execute a\ncommand, passed in as an <code class=\"highlighter-rouge\">ARG<\/code>, when any of the watched file changes.<\/p>\n<div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>WORKDIR=\/path\/to\/myproject\nfind $WORKDIR | grep \"\\.cpp$\" | entr make\n<\/pre><\/div><\/div>\n<p>Now we learned about <code class=\"highlighter-rouge\">rclone<\/code> and <code class=\"highlighter-rouge\">entr<\/code>. The final script turned out to be very\nsimple. To remind you about my use case of Dropbox, all I want to do is to\ncontinuously replicate my local Org file changes to Dropbox. Therefore, we just\nneed to use <code class=\"highlighter-rouge\">entr<\/code> to watch the files that we want to replicate, and use\n<code class=\"highlighter-rouge\">rclone<\/code> to \u201csync\u201d them to the remote storage.<\/p>\n<p>The final script (<code class=\"highlighter-rouge\">\/home\/lpan\/sync_dropbox.sh<\/code>) looks like the following:<\/p>\n<div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>#!\/bin\/bash\nORG_DIR=\/home\/lpan\/org\nREMOTE=dropbox\nfind $ORG_DIR | entr -r rclone sync -v $ORG_DIR $REMOTE:org\n<\/pre><\/div><\/div>\n<p>A <a href=\"https:\/\/en.wikipedia.org\/wiki\/Daemon_(computing)\">Daemon<\/a> is \u201cjust a computer\nprogram that runs as a background process\u201d. We make this script into a\n\u201cbackground process\u201d because we want it to <strong>continuously<\/strong> sync local file\nchanges to the remote file system in the background.<\/p>\n<p><a href=\"https:\/\/www.freedesktop.org\/wiki\/Software\/systemd\/\">systemd<\/a> provides an\ninterface for us to manage daemon processes.<\/p>\n<p>I created a <em>Dropbox Service<\/em> in <code class=\"highlighter-rouge\">~\/.config\/systemd\/user\/dropbox.service<\/code>.<\/p>\n<div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>[Unit]\nDescription=Dropbox Daemon\n[Service]\nExecStart=\/home\/lpan\/sync_dropbox.sh\nRestart=always\n[Install]\nWantedBy=default.target\n<\/pre><\/div><\/div>\n<p>Then, you can manage the daemon with the following commands<\/p>\n<div class=\"highlighter-rouge\"><div class=\"highlight\"><pre># reload the service file\nsystemctl --user daemon-reload\n# start the daemon\nsystemctl --user start dropbox.service\n# start the daemon on login\nsystemctl --user enable dropbox.service\n# inspect the status of the daemon\nsystemctl --user status dropbox.service\n<\/pre><\/div><\/div>\n<p>In this post, I discussed how I apply the UNIX philosophy and use a set of free\nand open-source tools to replace the proprietary &amp; deprecated Dropbox client. We\ntalked about <code class=\"highlighter-rouge\">rclone<\/code> and <code class=\"highlighter-rouge\">entr<\/code>. I also showed you how I make this process a\nDaemon and manage it using <code class=\"highlighter-rouge\">systemd<\/code>.<\/p>\n<p>I want to remind you that the <strong>key idea<\/strong> of this post is <em>simplicity<\/em>. We want\nto solve our simple problems with simple solutions. My use case of Dropbox is\nvery simple. And this is why one simple line of shell script would be a better\nsolution than using a feature-rich, and possibly proprietary remote storage\nclient.<\/p>\n<p>Thanks so much for reading! I really hope you enjoy this post. Got a better way\nto do the same job, or know how to extend the script for another use case? Let\nme know in the comment section below!<\/p>\n<p><strong>Edit 2018-12-24:<\/strong> This article has been discuss on\n<a href=\"https:\/\/www.reddit.com\/r\/linux\/comments\/a92m1u\/my_oneliner_linux_dropbox_client\/\">r\/linux<\/a>.\nThanks @boi_from_the_loo for submitting!<\/p>","created_at":"2018-12-25T14:55:28+0000","updated_at":"2018-12-25T14:55:28+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"lpan.io","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":759,"title":"orbitdb\/orbit-db","url":"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/README.md","content":"<p><a href=\"https:\/\/gitter.im\/orbitdb\/Lobby\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/fed5b5512498193ce4bba599fd94cd12b9f56491\/68747470733a2f2f696d672e736869656c64732e696f2f6769747465722f726f6f6d2f6e776a732f6e772e6a732e737667\" alt=\"Gitter\" data-canonical-src=\"https:\/\/img.shields.io\/gitter\/room\/nwjs\/nw.js.svg\" \/><\/a>\n<a href=\"http:\/\/webchat.freenode.net\/?channels=%23orbitdb\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/579bef4a50bb49be721f0b9ddeb76b543264cf7a\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f667265656e6f64652d2532336f7262697464622d626c75652e7376673f7374796c653d666c61742d737175617265\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/freenode-%23orbitdb-blue.svg?style=flat-square\" \/><\/a>\n<a href=\"https:\/\/circleci.com\/gh\/orbitdb\/orbit-db\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/4a88ae5add694b5aba393c87446427746d107e85\/68747470733a2f2f636972636c6563692e636f6d2f67682f6f7262697464622f6f726269742d64622e7376673f7374796c653d736869656c64\" alt=\"CircleCI Status\" data-canonical-src=\"https:\/\/circleci.com\/gh\/orbitdb\/orbit-db.svg?style=shield\" \/><\/a>\n<a href=\"https:\/\/www.npmjs.com\/package\/orbit-db\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/da7c0bc350c5af63a4e9167603e3b37e997e620d\/68747470733a2f2f62616467652e667572792e696f2f6a732f6f726269742d64622e737667\" alt=\"npm version\" data-canonical-src=\"https:\/\/badge.fury.io\/js\/orbit-db.svg\" \/><\/a>\n<a href=\"https:\/\/www.npmjs.com\/package\/orbit-db\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/03a11f2212fb6d94739543d4376435dce18b9fcd\/68747470733a2f2f696d672e736869656c64732e696f2f6e6f64652f762f6f726269742d64622e737667\" alt=\"node\" data-canonical-src=\"https:\/\/img.shields.io\/node\/v\/orbit-db.svg\" \/><\/a>\n<a href=\"https:\/\/waffle.io\/orbitdb\/orbit-db\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a0c3af88f1805a668ffc7274109b0c1dbc820324\/68747470733a2f2f62616467652e776166666c652e696f2f6f7262697464622f6f726269742d64622e7376673f636f6c756d6e733d496e25323050726f6772657373267469746c653d496e25323050726f6772657373\" alt=\"Project Status\" data-canonical-src=\"https:\/\/badge.waffle.io\/orbitdb\/orbit-db.svg?columns=In%20Progress&amp;title=In%20Progress\" \/><\/a><\/p>\n<blockquote>\n<p>A peer-to-peer database for the decentralized web<\/p>\n<\/blockquote>\n<p>OrbitDB is a serverless, distributed, peer-to-peer database. OrbitDB uses <a href=\"https:\/\/ipfs.io\" rel=\"nofollow\">IPFS<\/a> as its data storage and <a href=\"https:\/\/github.com\/ipfs\/go-ipfs\/blob\/master\/core\/commands\/pubsub.go#L23\">IPFS Pubsub<\/a> to automatically sync databases with peers. It's an eventually consistent database that uses <a href=\"https:\/\/en.wikipedia.org\/wiki\/Conflict-free_replicated_data_type\" rel=\"nofollow\">CRDTs<\/a> for conflict-free database merges making OrbitDB an excellent choice for decentralized apps (dApps), blockchain applications and offline-first web applications.<\/p>\n<p><strong>Test it live at <a href=\"https:\/\/ipfs.io\/ipfs\/QmeESXh9wPib8Xz7hdRzHuYLDuEUgkYTSuujZ2phQfvznQ\/\" rel=\"nofollow\">Live demo 1<\/a>, <a href=\"https:\/\/ipfs.io\/ipfs\/QmasHFRj6unJ3nSmtPn97tWDaQWEZw3W9Eh3gUgZktuZDZ\/\" rel=\"nofollow\">Live demo 2<\/a>, or <a href=\"https:\/\/ipfs.io\/ipfs\/QmTJGHccriUtq3qf3bvAQUcDUHnBbHNJG2x2FYwYUecN43\/\" rel=\"nofollow\">P2P TodoMVC app<\/a><\/strong>!<\/p>\n<p>OrbitDB provides various types of databases for different data models and use cases:<\/p>\n<ul><li><strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/API.md#orbitdblognameaddress\">log<\/a><\/strong>: an immutable (append-only) log with traversable history. Useful for <em>\"latest N\"<\/em> use cases or as a message queue.<\/li>\n<li><strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/API.md#orbitdbfeednameaddress\">feed<\/a><\/strong>: a mutable log with traversable history. Entries can be added and removed. Useful for *\"shopping cart\" type of use cases, or for example as a feed of blog posts or \"tweets\".<\/li>\n<li><strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/API.md#orbitdbkeyvaluenameaddress\">keyvalue<\/a><\/strong>: a key-value database just like your favourite key-value database.<\/li>\n<li><strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/API.md#orbitdbdocsnameaddress-options\">docs<\/a><\/strong>: a document database to which JSON documents can be stored and indexed by a specified key. Useful for building search indices or version controlling documents and data.<\/li>\n<li><strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/API.md#orbitdbcounternameaddress\">counter<\/a><\/strong>: Useful for counting events separate from log\/feed data.<\/li>\n<\/ul><p>All databases are <a href=\"https:\/\/github.com\/orbitdb\/orbit-db-store\">implemented<\/a> on top of <a href=\"https:\/\/github.com\/orbitdb\/ipfs-log\">ipfs-log<\/a>, an immutable, operation-based conflict-free replicated data structure (CRDT) for distributed systems. If none of the OrbitDB database types match your needs and\/or you need case-specific functionality, you can easily <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/GUIDE.md#custom-stores\">implement and use a custom database store<\/a> of your own.<\/p>\n<h4>Project status &amp; support<\/h4>\n<p>Status: <strong>in active development<\/strong><\/p>\n<p><em><strong>NOTE!<\/strong><\/em> <em>OrbitDB is <strong>alpha-stage<\/strong> software. It means OrbitDB hasn't been security audited and programming APIs and data formats can still change. We encourage you to <a href=\"https:\/\/gitter.im\/orbitdb\/Lobby\" rel=\"nofollow\">reach out to the maintainers<\/a> if you plan to use OrbitDB in mission critical systems.<\/em><\/p>\n<p>This is the Javascript implementation and it works both in <strong>Browsers<\/strong> and <strong>Node.js<\/strong> with support for Linux and OS X (Windows is not supported yet). The minimum required version of Node.js is now 8.0.0. To use with older versions of Node.js, we provide an ES5-compatible build through the npm package, located in <code>dist\/es5\/<\/code> when installed through npm.<\/p>\n<h4>Community Calls<\/h4>\n<p>We also have <strong>regular community calls<\/strong>, which we announce in the issues in <a href=\"https:\/\/github.com\/orbitdb\/welcome\/issues\">the @orbitdb welcome repository<\/a>. Join us!<\/p>\n<h2>Table of Contents<\/h2>\n<ul><li><a href=\"#usage\">Usage<\/a>\n<ul><li><a href=\"#cli\">CLI<\/a><\/li>\n<li><a href=\"#module-with-ipfs-instance\">Module with IPFS Instance<\/a><\/li>\n<li><a href=\"#module-with-ipfs-daemon\">Module with IPFS Daemon<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#api\">API<\/a><\/li>\n<li><a href=\"#examples\">Examples<\/a>\n<ul><li><a href=\"#install-dependencies\">Install dependencies<\/a><\/li>\n<li><a href=\"#browser-example\">Browser example<\/a><\/li>\n<li><a href=\"#nodejs-example\">Node.js example<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#packages\">Packages<\/a>\n<ul><li><a href=\"#orbitdb-store-packages\">OrbitDB Store Packages<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#development\">Development<\/a>\n<ul><li><a href=\"#run-tests\">Run Tests<\/a><\/li>\n<li><a href=\"#build\">Build<\/a><\/li>\n<li><a href=\"#benchmark\">Benchmark<\/a><\/li>\n<li><a href=\"#logging\">Logging<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#contributing\">Contributing<\/a><\/li>\n<li><a href=\"#sponsors\">Sponsors<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2>Usage<\/h2>\n<p>Read the <strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/GUIDE.md\">GETTING STARTED<\/a><\/strong> guide for a more in-depth tutorial and to understand how OrbitDB works.<\/p>\n<h3>CLI<\/h3>\n<p>For the CLI tool to manage orbit-db database, see <strong><a href=\"https:\/\/github.com\/orbitdb\/orbit-db-cli\">OrbitDB CLI<\/a><\/strong>.<\/p>\n<p>It can be installed from Npm with:<\/p>\n<pre>npm install orbit-db-cli -g\n<\/pre>\n<h3>Module with IPFS Instance<\/h3>\n<p>If you're using <code>orbitd-db<\/code> to develop <strong>browser<\/strong> or <strong>Node.js<\/strong> applications, use it as a module with the javascript instance of IPFS<\/p>\n<p>Install dependencies:<\/p>\n<pre>npm install orbit-db ipfs\n<\/pre>\n<div class=\"highlight highlight-source-js\"><pre>const IPFS = require('ipfs')\nconst OrbitDB = require('orbit-db')\n\/\/ OrbitDB uses Pubsub which is an experimental feature\n\/\/ and need to be turned on manually.\n\/\/ Note that these options need to be passed to IPFS in\n\/\/ all examples even if not specified so.\nconst ipfsOptions = {\n  EXPERIMENTAL: {\n    pubsub: true\n  }\n}\n\/\/ Create IPFS instance\nconst ipfs = new IPFS(ipfsOptions)\nipfs.on('error', (e) =&gt; console.error(e))\nipfs.on('ready', async () =&gt; {\n  const orbitdb = new OrbitDB(ipfs)\n  \/\/ Create \/ Open a database\n  const db = await orbitdb.log('hello')\n  await db.load()\n  \/\/ Listen for updates from peers\n  db.events.on('replicated', (address) =&gt; {\n    console.log(db.iterator({ limit: -1 }).collect())\n  })\n  \/\/ Add an entry\n  const hash = await db.add('world')\n  console.log(hash)\n  \/\/ Query\n  const result = db.iterator({ limit: -1 }).collect()\n  console.log(JSON.stringify(result, null, 2))\n})<\/pre><\/div>\n<h3>Module with IPFS Daemon<\/h3>\n<p>Alternatively, you can use <a href=\"https:\/\/npmjs.org\/package\/ipfs-api\" rel=\"nofollow\">ipfs-api<\/a> to use <code>orbit-db<\/code> with a locally running IPFS daemon. Use this method if you're using <code>orbitd-db<\/code> to develop <strong>backend<\/strong> or <strong>desktop<\/strong> applications, eg. with <a href=\"https:\/\/electron.atom.io\" rel=\"nofollow\">Electron<\/a>.<\/p>\n<p>Install dependencies:<\/p>\n<pre>npm install orbit-db ipfs-api\n<\/pre>\n<div class=\"highlight highlight-source-js\"><pre>const IpfsApi = require('ipfs-api')\nconst OrbitDB = require('orbit-db')\nconst ipfs = IpfsApi('localhost', '5001')\nconst orbitdb = new OrbitDB(ipfs)\nconst db = await orbitdb.log('hello')\n...<\/pre><\/div>\n<h2>API<\/h2>\n<p>See <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/API.md\">API.md<\/a> for the full documentation.<\/p>\n<h2>Examples<\/h2>\n<h3>Install dependencies<\/h3>\n<pre>git clone https:\/\/github.com\/orbitdb\/orbit-db.git\ncd orbit-db\nnpm install\n<\/pre>\n<p>You'll also need babel and webpack, if you don't have them installed already:<\/p>\n<pre>npm install --global babel-cli\nnpm install --global webpack\n<\/pre>\n<p>Some dependencies depend on native addon modules, so you'll also need to meet <a href=\"https:\/\/github.com\/nodejs\/node-gyp#installation\">node-gyp's<\/a> installation prerequisites. Therefore, Linux users may need to<\/p>\n<pre>make clean &amp;&amp; make\n<\/pre>\n<p>to redo the local package-lock.json with working native dependencies.<\/p>\n<h3>Browser example<\/h3>\n<p>In macOS:<\/p>\n<pre>npm run build\nnpm run examples:browser-macos\n<\/pre>\n<p>In Linux:<\/p>\n<pre>npm run build\nnpm run examples:browser-linux\n<\/pre>\n<p>\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/orbitdb\/orbit-db\/master\/screenshots\/example1.png\"><img src=\"https:\/\/raw.githubusercontent.com\/orbitdb\/orbit-db\/master\/screenshots\/example1.png\" width=\"33%\" alt=\"image\" \/><\/a>\n<\/p>\n<p>Check the code in <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/examples\/browser\/browser.html\">examples\/browser\/browser.html<\/a> and try the <a href=\"https:\/\/ipfs.io\/ipfs\/QmRosp97r8GGUEdj5Wvivrn5nBkuyajhRXFUcWCp5Zubbo\/\" rel=\"nofollow\">live example<\/a>.<\/p>\n<h3>Node.js example<\/h3>\n<pre>npm run examples:node\n<\/pre>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/orbitdb\/orbit-db\/master\/screenshots\/orbit-db-demo3.gif\"><img src=\"https:\/\/raw.githubusercontent.com\/orbitdb\/orbit-db\/master\/screenshots\/orbit-db-demo3.gif\" width=\"66%\" alt=\"image\" \/><\/a><\/p>\n<p><strong>Eventlog<\/strong><\/p>\n<p>See the code in <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/examples\/eventlog.js\">examples\/eventlog.js<\/a> and run it with:<\/p>\n<pre>node examples\/eventlog.js\n<\/pre>\n<p>More examples at <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/tree\/master\/examples\">examples<\/a>.<\/p>\n<h2>Packages<\/h2>\n<p>OrbitDB uses the following modules:<\/p>\n<h3>OrbitDB Store Packages<\/h3>\n<p>To understand a little bit about the architecture, check out a visualization of the data flow at <a href=\"https:\/\/github.com\/haadcode\/proto2\">https:\/\/github.com\/haadcode\/proto2<\/a> or a live demo: <a href=\"http:\/\/celebdil.benet.ai:8080\/ipfs\/Qmezm7g8mBpWyuPk6D84CNcfLKJwU6mpXuEN5GJZNkX3XK\/\" rel=\"nofollow\">http:\/\/celebdil.benet.ai:8080\/ipfs\/Qmezm7g8mBpWyuPk6D84CNcfLKJwU6mpXuEN5GJZNkX3XK\/<\/a>.<\/p>\n<h2>Development<\/h2>\n<h3>Run Tests<\/h3>\n<pre>npm test\n<\/pre>\n<h3>Build<\/h3>\n<pre>npm run build\n<\/pre>\n<h3>Benchmark<\/h3>\n<pre>node benchmarks\/benchmark-add.js\n<\/pre>\n<p>See <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/tree\/master\/benchmarks\">benchmarks\/<\/a> for more benchmarks.<\/p>\n<h3>Logging<\/h3>\n<p>To enable OrbitDB's logging output, set a global ENV variable called <code>LOG<\/code> to <code>debug<\/code>,<code>warn<\/code> or <code>error<\/code>:<\/p>\n<pre>LOG=debug node &lt;file&gt;\n<\/pre>\n<h2>Contributing<\/h2>\n<p>We would be happy to accept PRs! If you want to work on something, it'd be good to talk beforehand to make sure nobody else is working on it. You can reach us on IRC <a href=\"http:\/\/webchat.freenode.net\/?channels=%23orbitdb\" rel=\"nofollow\">#orbitdb<\/a> on Freenode, or in the comments of the <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/issues\">issues section<\/a>.<\/p>\n<p>We also have <strong>regular community calls<\/strong>, which we announce in the issues in <a href=\"https:\/\/github.com\/orbitdb\/welcome\/issues\">the @orbitdb welcome repository<\/a>. Join us!<\/p>\n<p>If you want to code but don't know where to start, check out the issues labelled <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/issues?q=is%3Aopen+is%3Aissue+label%3A%22help+wanted%22+sort%3Areactions-%2B1-desc\">\"help wanted\"<\/a> or the project's <a href=\"https:\/\/waffle.io\/orbitdb\/orbit-db\" rel=\"nofollow\">status board<\/a>.<\/p>\n<p>Please note that we have a <a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/CODE_OF_CONDUCT.md\">Code of Conduct<\/a>, and that all activity in the <a href=\"https:\/\/github.com\/orbitdb\">@orbitdb<\/a> organization falls under it. Read it when you get the chance, as being part of this community means that you agree to abide by it. Thanks.<\/p>\n<h2>Sponsors<\/h2>\n<p>The development of OrbitDB has been sponsored by:<\/p>\n<ul><li><a href=\"https:\/\/haja.io\" rel=\"nofollow\">Haja Networks<\/a><\/li>\n<li><a href=\"https:\/\/protocol.ai\/\" rel=\"nofollow\">Protocol Labs<\/a><\/li>\n<li><a href=\"https:\/\/maintainer.io\" rel=\"nofollow\">Maintainer Mountaineer<\/a><\/li>\n<\/ul><p>If you want to sponsor developers to work on OrbitDB, please reach out to @haadcode.<\/p>\n<h2>License<\/h2>\n<p><a href=\"https:\/\/github.com\/orbitdb\/orbit-db\/blob\/master\/LICENSE\">MIT<\/a> \u00a9 2015-2018 Protocol Labs Inc., Haja Networks Oy<\/p>","created_at":"2018-12-25T04:24:12+0000","updated_at":"2018-12-25T04:24:16+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/25079463?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":758,"title":"SQLite As An Application File Format","url":"https:\/\/www.sqlite.org\/appfileformat.html","content":"SQLite As An Application File Format<h2>Executive Summary<\/h2><p>An SQLite database file with a defined schema\noften makes an excellent application file format.\nHere are a dozen reasons why this is so:\n<\/p><ol><li> Simplified Application Development\n<\/li><li> Single-File Documents\n<\/li><li> High-Level Query Language\n<\/li><li> Accessible Content\n<\/li><li> Cross-Platform\n<\/li><li> Atomic Transactions\n<\/li><li> Incremental And Continuous Updates\n<\/li><li> Easily Extensible\n<\/li><li> Performance\n<\/li><li> Concurrent Use By Multiple Processes\n<\/li><li> Multiple Programming Languages\n<\/li><li> Better Applications\n<\/li><\/ol><p>Each of these points will be described in more detail below,\nafter first considering more closely the meaning of\n\"application file format\".  See also the <a href=\"https:\/\/www.sqlite.org\/aff_short.html\">short version<\/a> of this\nwhitepaper.\n<\/p><h2>What Is An Application File Format?<\/h2><p>\nAn \"application file format\" is the file format\nused to persist application state to disk or to exchange\ninformation between programs.\nThere are thousands of application file formats in use today.\nHere are just a few examples:\n<\/p><ul><li>DOC - Word Perfect and Microsoft Office documents\n<\/li><li>DWG - AutoCAD drawings\n<\/li><li>PDF - Portable Document Format from Adobe\n<\/li><li>XLS - Microsoft Excel Spreadsheet\n<\/li><li>GIT - Git source code repository\n<\/li><li>EPUB - The Electronic Publication format used by non-Kindle eBooks\n<\/li><li>ODT - The Open Document format used by OpenOffice and others\n<\/li><li>PPT - Microsoft PowerPoint presentations\n<\/li><li>ODP - The Open Document presentation format used by OpenOffice and others\n<\/li><\/ul><p>We make a distinction between a \"file format\" and an \"application format\".\nA file format is used to store a single object.  So, for example, a GIF or\nJPEG file stores a single image, and an XHTML file stores text,\nso those are \"file formats\" and not \"application formats\".  An EPUB file, \nin contrast, stores both text and images (as contained XHTML and GIF\/JPEG\nfiles) and so it is considered an \"application format\".  This article is\nabout \"application formats\".\n<\/p><p>The boundary between a file format and an application format is fuzzy.\nThis article calls JPEG a file format, but for an image editor, JPEG \nmight be considered the application format.  Much depends on context.\nFor this article, let us say that a file format stores a single object\nand an application format stores many different objects and their relationships\nto one another.\n<\/p><p>Most application formats fit into one of these three categories:\n<\/p><ol><li><p><b>Fully Custom Formats.<\/b>\nCustom formats are specifically designed for a single application.\nDOC, DWG, PDF, XLS, and PPT are examples of custom formats.  Custom\nformats are usually contained within a single file, for ease of transport.\nThey are also usually binary, though the DWG format is a notable exception.\nCustom file formats require specialized application code\nto read and write and are not normally accessible from commonly\navailable tools such as unix command-line programs and text editors.\nIn other words, custom formats are usually \"opaque blobs\".\nTo access the content of a custom application file format, one needs\na tool specifically engineered to read and\/or write that format.\n<\/p><\/li><li><p><b>Pile-of-Files Formats.<\/b>\nSometimes the application state is stored as a hierarchy of\nfiles.  Git is a prime example of this, though the phenomenon occurs\nfrequently in one-off and bespoke applications.  A pile-of-files format\nessentially uses the filesystem as a key\/value database, storing small\nchunks of information into separate files.  This gives the\nadvantage of making the content more accessible to common utility\nprograms such as text editors or \"awk\" or \"grep\".  But even if many \nof the files in a pile-of-files format\nare easily readable, there are usually some files that have their\nown custom format (example: Git \"Packfiles\") and are hence\n\"opaque blobs\" that are not readable\nor writable without specialized tools.  It is also much less convenient\nto move a pile-of-files from one place or machine to another, than\nit is to move a single file.  And it is hard to make a pile-of-files\ndocument into an email attachment, for example.  Finally, a pile-of-files\nformat breaks the \"document metaphor\":\nthere is no one file that a user can point to\nthat is \"the document\".\n<\/p><\/li><li><p><b>Wrapped Pile-of-Files Formats.<\/b>\nSome applications use a Pile-of-Files that is then encapsulated into\nsome kind of single-file container, usually a ZIP archive.  \nEPUB, ODT,and ODP are examples of this approach.\nAn EPUB book is really just a ZIP archive that contains various\nXHTML files for the text of book chapters, GIF and JPEG images for\nthe artwork, and a specialized catalog file that tells the eBook\nreader how all the XML and image files fit together.  OpenOffice\ndocuments (ODT and ODP) are also ZIP archives containing XML and\nimages that represent their content as well as \"catalog\" files that\nshow the interrelationships between the component parts.\n<\/p><p>A wrapped pile-of-files format is a compromise between a full\ncustom file format and a pure pile-of-files format.\nA wrapped pile-of-files format is not an opaque blob in the same sense\nas a custom format, since the component parts can still be accessed\nusing any common ZIP archiver, but the format is not quite as accessible\nas a pure pile-of-files format because one does still need the ZIP \narchiver, and one cannot normally use command-line tools like \"find\"\non the file hierarchy without first un-zipping it.  On the other\nhand, a wrapped pile-of-files format does preserve the document\nmetaphor by putting all content into a single disk file.  And\nbecause it is compressed, the wrapped pile-of-files format tends to\nbe more compact.\n<\/p><p>As with custom file formats, and unlike pure pile-of-file formats,\na wrapped pile-of-files format is not as easy to edit, since\nusually the entire file must be rewritten in order to change any\ncomponent part.\n<\/p><\/li><\/ol><p>The purpose of this document is to argue in favor of a fourth\nnew category of application file format: An SQLite database file.\n<\/p><h2>SQLite As The Application File Format<\/h2><p>\nAny application state that can be recorded in a pile-of-files can\nalso be recorded in an SQLite database with a simple key\/value schema\nlike this:\n<\/p><blockquote><pre>\nCREATE TABLE files(filename TEXT PRIMARY KEY, content BLOB);\n<\/pre><\/blockquote>\nIf the content is compressed, then such an <a href=\"https:\/\/www.sqlite.org\/sqlar.html\">SQLite Archive<\/a> database is\n<a href=\"https:\/\/www.sqlite.org\/affcase1.html#smaller\">the same size<\/a> (\u00b11%)\nas an equivalent ZIP archive, and it has the advantage\nof being able to update individual \"files\" without rewriting\nthe entire document.\n<p>\nBut an SQLite database is not limited to a simple key\/value structure\nlike a pile-of-files database.  An SQLite database can have dozens\nor hundreds or thousands of different tables, with dozens or\nhundreds or thousands of fields per table, each with different datatypes\nand constraints and particular meanings, all cross-referencing each other,\nappropriately and automatically indexed for rapid retrieval,\nand all stored efficiently and compactly in a single disk file.\nAnd all of this structure is succinctly documented for humans\nby the SQL schema.\n<\/p><p>In other words, an SQLite database can do everything that a \npile-of-files or wrapped pile-of-files format can do, plus much more,\nand with greater lucidity.\nAn SQLite database is a more versatile container than key\/value\nfilesystem or a ZIP archive.  (For a detailed example, see the\n<a href=\"https:\/\/www.sqlite.org\/affcase1.html\">OpenOffice case study<\/a> essay.)\n<\/p><p>The power of an SQLite database could, in theory, be achieved using\na custom file format.  But any custom file format that is as expressive\nas a relational database would likely require an enormous design specification \nand many tens or hundreds of thousands of lines of code to \nimplement.  And the end result would be an \"opaque blob\" that is\ninaccessible without specialized tools.\n<\/p><p>\nHence, in comparison to other approaches, the use of\nan SQLite database as an application file format has\ncompelling advantages.  Here are a few of these advantages,\nenumerated and expounded:\n<\/p><ol><li><p><b>Simplified Application Development.<\/b>\nNo new code is needed for reading or writing the application file.\nOne has merely to link against the SQLite library, or include the \n<a href=\"https:\/\/www.sqlite.org\/amalgamation.html\">single \"sqlite3.c\" source file<\/a> with the rest of the\napplication C code, and SQLite will take care of all of the application\nfile I\/O.  This can reduce application code size by many thousands of\nlines, with corresponding saving in development and maintenance costs.\n<\/p><p>SQLite is one of the\n<a href=\"https:\/\/www.sqlite.org\/mostdeployed.html\">most used<\/a> software libraries in the world.\nThere are literally tens of billions of SQLite database files in use \ndaily, on smartphones and gadgets and in desktop applications.\nSQLite is <a href=\"https:\/\/www.sqlite.org\/testing.html\">carefully tested<\/a> and proven reliable.  It is not\na component that needs much tuning or debugging, allowing developers\nto stay focused on application logic.\n<\/p><\/li><li><p><b>Single-File Documents.<\/b>\nAn SQLite database is contained in a single file, which is easily\ncopied or moved or attached.  The \"document\" metaphor is preserved.\n<\/p><p>SQLite does not have any file naming requirements\nand so the application can use any custom file suffix that it wants\nto help identify the file as \"belonging\" to the application.\nSQLite database files contain a 4-byte <a href=\"https:\/\/www.sqlite.org\/fileformat2.html#appid\">Application ID<\/a> in\ntheir headers that can be set to an application-defined value\nand then used to identify the \"type\" of the document for utility\nprograms such as <a href=\"http:\/\/linux.die.net\/man\/1\/file\">file(1)<\/a>, further\nenhancing the document metaphor.\n<\/p><\/li><li><p><b>High-Level Query Language.<\/b>\nSQLite is a complete relational database engine, which means that the\napplication can access content using high-level queries.  Application\ndevelopers need not spend time thinking about \"how\" to retrieve the\ninformation they need from a document.  Developers write SQL that\nexpresses \"what\" information they want and let the database engine\nto figure out how to best retrieve that content.  This helps developers\noperate \"heads up\" and remain focused on solving the user's problem,\nand avoid time spent \"heads down\" fiddling with low-level file\nformatting details.\n<\/p><p>A pile-of-files format can be viewed as a key\/value database.  \nA key\/value database is better than no database at all.\nBut without transactions or indices or a high-level query language or\na proper schema,\nit is much harder and more error prone to use a key\/value database than\na relational database.\n<\/p><\/li><li><p><b>Accessible Content.<\/b>\nInformation held in an SQLite database file is accessible using \ncommonly available open-source command-line tools - tools that \nare installed by default on Mac and Linux systems and that are \nfreely available as a self-contained EXE file on Windows.\nUnlike custom file formats, application-specific programs are\nnot required to read or write content in an SQLite database.\nAn SQLite database file is not an opaque blob.  It is true\nthat command-line tools such as text editors or \"grep\" or \"awk\" are\nnot useful on an SQLite database, but the SQL query language is a much\nmore powerful and convenient way for examining the content, so the\ninability to use \"grep\" and \"awk\" and the like is not seen as a loss.\n<\/p><p>An SQLite database is a <a href=\"https:\/\/www.sqlite.org\/fileformat2.html\">well-defined and well-documented<\/a>\nfile format that is in widespread use by literally millions of applications\nand is backwards compatible to its inception in 2004 and which promises\nto continue to be compatible in decades to come.  The longevity of\nSQLite database files is particularly important to bespoke applications,\nsince it allows the document content to be accessed far in the\nfuture, long after all traces of the original application have been lost.\nData lives longer than code.\nSQLite databases are <a href=\"https:\/\/www.sqlite.org\/locrsf.html\">recommended by the US Library of Congress<\/a>\nas a storage format for long-term preservation of digital content.\n<\/p><\/li><li><p><b>Cross-Platform.<\/b>\nSQLite database files are portable between 32-bit and 64-bit machines and\nbetween big-endian and little-endian architectures and between any of the\nvarious flavors of Windows and Unix-like operating systems.\nThe application using an SQLite application file format can store\nbinary numeric data without having to worry about the byte-order of\nintegers or floating point numbers.\nText content can be read or written as UTF-8, UTF-16LE, or UTF-16BE and \nSQLite will automatically perform any necessary translations on-the-fly.\n<\/p><\/li><li><p><b>Atomic Transactions.<\/b>\nWrites to an SQLite database are <a href=\"https:\/\/www.sqlite.org\/atomiccommit.html\">atomic<\/a>.  \nThey either happen completely\nor not at all, even during system crashes or power failures.  So\nthere is no danger of corrupting a document just because the power happened\nto go out at the same instant that a change was being written to disk.\n<\/p><p>SQLite is transactional, meaning that multiple changes can be grouped\ntogether such that either all or none of them occur, and so that the\nchanges can be rolled back if a problem is found prior to commit.\nThis allows an application to make a change incrementally, then run\nvarious sanity and consistency checks on the resulting data prior to\ncommitting the changes to disk.  The\n<a href=\"http:\/\/www.fossil-scm.org\/\">Fossil<\/a> DVCS \n<a href=\"http:\/\/www.fossil-scm.org\/fossil\/doc\/tip\/www\/selfcheck.wiki\">uses this technique<\/a>\nto verify that no repository history has been lost prior to each change.\n<\/p><\/li><li><p><b>Incremental And Continuous Updates.<\/b>\nWhen writing to an SQLite database file, only those parts of the file that\nactually change are written out to disk.  This makes the writing happen faster\nand saves wear on SSDs.  This is an enormous advantage over custom\nand wrapped pile-of-files formats, both of which usually require a\nrewrite of the entire document in order to change a single byte.  \nPure pile-of-files formats can also\ndo incremental updates to some extent, though the granularity of writes is \nusually larger with pile-of-file formats (a single file) than with SQLite\n(a single page).\n<\/p><p>SQLite also supports continuous update.\nInstead of collecting changes in memory and then writing\nthem to disk only on a File\/Save action, changes can be written back to\nthe disk as they occur.  This avoids loss of work on a system crash or\npower failure.  An <a href=\"https:\/\/www.sqlite.org\/undoredo.html\">automated undo\/redo stack<\/a>, managed using triggers,\ncan be kept in the on-disk database, meaning that undo\/redo can occur \nacross session boundaries.\n<\/p><\/li><li><p><b>Easily Extensible.<\/b>\nAs an application grows, new features can be added to an\nSQLite application file format simply by adding new tables to the schema\nor by adding new columns to existing tables.  Adding columns or tables\ndoes not change the meaning of prior queries, so with a \nmodicum of care to ensuring that the meaning of legacy columns and\ntables are preserved, backwards compatibility is maintained.\n<\/p><p>It is possible to extend custom or pile-of-files formats too, of course,\nbut doing is often much harder.  If indices are added, then all application\ncode that changes the corresponding tables must be located and modified to\nkeep those indices up-to-date.  If columns are added, then all application\ncode that accesses the corresponding table must be located and modified to \ntake into account the new columns.\n<\/p><\/li><li><p><b>Performance.<\/b>\nIn many cases, an SQLite application file format will be \n<a href=\"https:\/\/www.sqlite.org\/fasterthanfs.html\">faster than a pile-of-files format<\/a> or\na custom format.  In addition to being faster for raw read and\nwrites, SQLite can often dramatically improves start-up times because \ninstead of having to\nread and parse the entire document into memory, the application can\ndo queries to extract only the information needed for the initial screen.\nAs the application progresses, it only needs to load as much material as\nis needed to draw the next screen, and can discard information from\nprior screens that is no longer in use.  This helps keep the memory\nfootprint of the application under control.\n<\/p><p>A pile-of-files format can be read incrementally just like SQLite.\nBut many developers are surprised to learn that SQLite can read and \nwrite smaller BLOBs (less than about 100KB in size) from its database \nfaster than those same blobs can be read or written as separate files \nfrom the filesystem.  (See\n<a href=\"https:\/\/www.sqlite.org\/fasterthanfs.html\">35% Faster Than The Filesystem<\/a> and\n<a href=\"https:\/\/www.sqlite.org\/intern-v-extern-blob.html\">Internal Versus External BLOBs<\/a> for further information.)\nThere is overhead associated with operating a relational\ndatabase engine, however one should not assume that direct file I\/O\nis faster than SQLite database I\/O, as often it is not.\n<\/p><p>In either case, if performance problems do arise in an SQLite application \nthose problems can often be resolved by adding one or two <a href=\"https:\/\/www.sqlite.org\/lang_createindex.html\">CREATE INDEX<\/a>\nstatements to the schema or perhaps running <a href=\"https:\/\/www.sqlite.org\/lang_analyze.html\">ANALYZE<\/a> one time\nand without having to touch a single line of\napplication code.  But if a performance problem comes up in a custom or \npile-of-files format, the fix will often require extensive changes\nto application code to add and maintain new indices or to extract \ninformation using different algorithms.\n<\/p><\/li><li><p><b>Concurrent Use By Multiple Processes.<\/b>\nSQLite automatically coordinates concurrent access to the same\ndocument from multiple threads and\/or processes.  Two or more\napplications can connect and read from the same document at the\nsame time.  Writes are serialized, but as writes normally only\ntake milliseconds, applications simply take turns writing.\nSQLite automatically ensures that the low-level format of the\ndocument is uncorrupted.  Accomplishing the same with a custom\nor pile-of-files format, in contrast, requires extensive support\nin the application.  And the application logic needed to support\nconcurrency is a notorious bug-magnet.\n<\/p><\/li><li><p><b>Multiple Programming Languages.<\/b>\nThough SQLite is itself written in ANSI-C, interfaces exist for\njust about every other programming language you can think of:\nC++, C#, Objective-C, Java, Tcl, Perl, Python, Ruby, Erlang,\nJavaScript, and so forth.  So programmers can develop in whatever\nlanguage they are most comfortable with and which best matches\nthe needs of the project.\n<\/p><p>An SQLite application file format is a great \nchoice in cases where there is a collection or \"federation\" of\nseparate programs, often written in different languages and by\ndifferent development teams.\nThis comes up commonly in research or laboratory\nenvironments where one team is responsible for data acquisition\nand other teams are responsible for various stages of analysis.\nEach team can use whatever hardware, operating system,\nprogramming language and development methodology that they are \nmost comfortable with, and as long as all programs use an SQLite \ndatabase with a common schema, they can all interoperate.\n<\/p><\/li><li><p><b>Better Applications.<\/b>\nIf the application file format is an SQLite database, the complete\ndocumentation for that file format consists of the database schema,\nwith perhaps a few extra words about what each table and column\nrepresents.  The description of a custom file format,\non the other hand, typically runs on for hundreds of \npages.  A pile-of-files format, while much simpler and easier to\ndescribe than a fully custom format, still tends to be much larger\nand more complex than an SQL schema dump, since the names and format\nfor the individual files must still be described.\n<\/p><p>This is not a trivial point.  A clear, concise, and easy to understand\nfile format is a crucial part of any application design.\nFred Brooks, in his all-time best-selling computer science text,\n<i>The Mythical Man-Month<\/i> says:\n<\/p><blockquote><div><i>Representation is the\nessence of computer programming.<br \/>...<br \/>Show me your flowcharts and conceal your tables, and I shall \ncontinue to be mystified.  Show me your tables, and I won't usually\nneed your flowcharts; they'll be obvious.<\/i><\/div><\/blockquote>\n<p>Rob Pike, in his\n<i>Rules of Programming<\/i> expresses the same idea this way:\n<\/p><blockquote>\n<div><i>Data dominates.  If you've chosen the right data structures\nand organized things well, the algorithms will almost always\nbe self-evident.  Data structures, not algorithms, are central\nto programming.<\/i><\/div><\/blockquote>\n<p> Linus Torvalds used different words to say\nmuch the same thing on the Git mailing list on 2006-06-27: \n<\/p><blockquote>\n<div><i>Bad programmers worry about the code.  Good programmers worry\nabout data structures and their relationships.<\/i>\n<\/div><\/blockquote>\n<p>The point is this: an SQL database schema almost always does \na far better job of defining and organizing the tables and \ndata structures and their relationships.\nAnd having clear, concise, and well-defined representation\nalmost always results in an application that performs better,\nhas fewer problems, and is easier to develop and maintain.\n<\/p><\/li><\/ol><h2>Conclusion<\/h2><p>\nSQLite is not the perfect application file format for every situation.\nBut in many cases, SQLite is a far better choice than either a custom\nfile format, a pile-of-files, or a wrapped pile-of-files.\nSQLite is a high-level, stable, reliable, cross-platform, widely-deployed,\nextensible, performant, accessible, concurrent file format.  It deserves\nyour consideration as the standard file format on your next application\ndesign.\n<\/p>","created_at":"2018-12-24T18:59:08+0000","updated_at":"2018-12-24T18:59:08+0000","annotations":[],"mimetype":"text\/html","reading_time":16,"domain_name":"www.sqlite.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["algorithms","regex"],"is_public":false,"id":757,"title":"Regular Expression Matching Can Be Simple And Fast","url":"https:\/\/swtch.com\/~rsc\/regexp\/regexp1.html","content":"Regular Expression Matching Can Be Simple And Fast\n<h2 class=\"au\">\n<a href=\"http:\/\/swtch.com\/~rsc\/\">Russ Cox<\/a>\n<br \/><i>rsc@swtch.com<\/i>\n<br \/>January 2007\n<\/h2>\n<h2 class=\"sh\">Introduction<\/h2>\n<p class=\"pp\">\nThis is a tale of two approaches to regular expression matching.\nOne of them is in widespread use in the\nstandard interpreters for many languages, including Perl.\nThe other is used only in a few places, notably most implementations\nof awk and grep.\nThe two approaches have wildly different\nperformance characteristics:\n<\/p>\n<div class=\"fig\"><div style=\"text-align: center;\">\n<table cellpadding=\"0\" border=\"0\" style=\"border-spacing: 0px;\"><tr><td valign=\"bottom\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/grep3p.png\" alt=\"Perl graph\" width=\"301\" height=\"148\" \/><\/td><td style=\"width: 20px;\"><\/td><td valign=\"bottom\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/grep4p.png\" alt=\"Thompson NFA graph\" width=\"301\" height=\"148\" \/><\/td><\/tr><tr><td style=\"height: 10px;\">\n<\/td><\/tr><tr><td colspan=\"3\" align=\"center\">\nTime to match <code>a?<\/code><sup><i>n<\/i><\/sup><code>a<\/code><sup><i>n<\/i><\/sup> against <code>a<\/code><sup><i>n<\/i><\/sup><\/td><\/tr><\/table><\/div><\/div>\n<p class=\"lp\">\nLet's use superscripts to denote string repetition,\nso that \n<code>a?<sup>3<\/sup>a<sup>3<\/sup><\/code>\nis shorthand for\n<code>a?a?a?aaa<\/code>.\nThe two graphs plot the time required by each approach\nto match the regular expression \n<code>a?<\/code><sup><i>n<\/i><\/sup><code>a<\/code><sup><i>n<\/i><\/sup>\nagainst the string <code>a<\/code><sup><i>n<\/i><\/sup>.\n<\/p>\n<p class=\"pp\">\nNotice that Perl requires over sixty seconds to match\na 29-character string.\nThe other approach, labeled Thompson NFA for\nreasons that will be explained later,\nrequires twenty <i>microseconds<\/i> to match the string.\nThat's not a typo.  The Perl graph plots time in seconds,\nwhile the Thompson NFA graph plots time in microseconds:\nthe Thompson NFA implementation\nis a million times faster than Perl\nwhen running on a miniscule 29-character string.\nThe trends shown in the graph continue: the\nThompson NFA handles a 100-character string in under 200 microseconds,\nwhile Perl would require over 10<sup>15<\/sup> years.\n(Perl is only the most conspicuous example of a large\nnumber of popular programs that use the same algorithm;\nthe above graph could have been Python, or PHP, or Ruby,\nor many other languages.  A more detailed\ngraph later in this article presents data for other implementations.)\n<\/p>\n<p class=\"pp\">\nIt may be hard to believe the graphs: perhaps you've used Perl,\nand it never seemed like regular expression matching was\nparticularly slow.\nMost of the time, in fact, regular expression matching in Perl\nis fast enough.  \nAs the graph shows, though, it is possible\nto write so-called \u201cpathological\u201d regular expressions that\nPerl matches very <i>very<\/i> slowly.\nIn contrast, there are no regular expressions that are \npathological for the Thompson NFA implementation.\nSeeing the two graphs side by side prompts the question, \n\u201cwhy doesn't Perl use the Thompson NFA approach?\u201d\nIt can, it should, and that's what the rest of this article is about.\n<\/p>\n<p class=\"pp\">\nHistorically, regular expressions are one of computer science's\nshining examples of how using good theory leads to good programs.\nThey were originally developed by theorists as a\nsimple computational model,\nbut Ken Thompson introduced them to\nprogrammers in his implementation of the text editor QED\nfor CTSS.\nDennis Ritchie followed suit in his own implementation\nof QED, for GE-TSS.\nThompson and Ritchie would go on to create Unix,\nand they brought regular expressions with them.\nBy the late 1970s, regular expressions were a key\nfeature of the Unix landscape, in tools such as\ned, sed, grep, egrep, awk, and lex.\n<\/p>\n<p class=\"pp\">\nToday, regular expressions have also become a shining\nexample of how ignoring good theory leads to bad programs.\nThe regular expression implementations used by\ntoday's popular tools are significantly slower\nthan the ones used in many of those thirty-year-old Unix tools.\n<\/p>\n<p class=\"pp\">\nThis article reviews the good theory: \nregular expressions, finite automata, \nand a regular expression search algorithm\ninvented by Ken Thompson in the mid-1960s.\nIt also puts the theory into practice, describing \na simple implementation of Thompson's algorithm.\nThat implementation, less than 400 lines of C,\nis the one that went head to head with Perl above.\nIt outperforms the more complex real-world\nimplementations used by\nPerl, Python, PCRE, and others.\nThe article concludes with a discussion of how \ntheory might yet be converted into practice\nin the real-world implementations.\n<\/p>\n<h2 class=\"sh\">\nRegular Expressions\n<\/h2>\n<p class=\"pp\">\nRegular expressions are a notation for\ndescribing sets of character strings.\nWhen a particular string is in the set\ndescribed by a regular expression,\nwe often say that the regular expression\n<i>matches<\/i>\nthe string.\n<\/p>\n<p class=\"pp\">\nThe simplest regular expression is a single literal character.\nExcept for the special metacharacters \n<code>*+?()|<\/code>,\ncharacters match themselves.\nTo match a metacharacter, escape it with\na backslash:\n<code>\\+<\/code>\nmatches a literal plus character.\n<\/p>\n<p class=\"pp\">\nTwo regular expressions can be alternated or concatenated to form a new\nregular expression:\nif <i>e<\/i><sub>1<\/sub> matches\n<i>s<\/i>\nand <i>e<\/i><sub>2<\/sub> matches\n<i>t<\/i>,\nthen <i>e<\/i><sub>1<\/sub><code>|<\/code><i>e<\/i><sub>2<\/sub> matches\n<i>s<\/i>\nor\n<i>t<\/i>,\nand\n<i>e<\/i><sub>1<\/sub><i>e<\/i><sub>2<\/sub>\nmatches \n<i>st<\/i>.\n<\/p>\n<p class=\"pp\">\nThe metacharacters\n<code>*<\/code>,\n<code>+<\/code>,\nand\n<code>?<\/code>\nare repetition operators:\n<i>e<\/i><sub>1<\/sub><code>*<\/code>\nmatches a sequence of zero or more (possibly different)\nstrings, each of which match <i>e<\/i><sub>1<\/sub>;\n<i>e<\/i><sub>1<\/sub><code>+<\/code>\nmatches one or more; \n<i>e<\/i><sub>1<\/sub><code>?<\/code>\nmatches zero or one.\n<\/p>\n<p class=\"pp\">\nThe operator precedence, from weakest to strongest binding, is\nfirst alternation, then concatenation, and finally the\nrepetition operators.\nExplicit parentheses can be used to force different meanings,\njust as in arithmetic expressions.\nSome examples:\n<code>ab|cd<\/code>\nis equivalent to\n<code>(ab)|(cd)<\/code>;\n<code>ab*<\/code>\nis equivalent to\n<code>a(b*)<\/code>.\n<\/p>\n<p class=\"pp\">\nThe syntax described so far is a subset of the traditional Unix\negrep\nregular expression syntax.\nThis subset suffices to describe all regular\nlanguages: loosely speaking, a regular language is a set\nof strings that can be matched in a single pass through\nthe text using only a fixed amount of memory.\nNewer regular expression facilities (notably Perl and\nthose that have copied it) have added \n<a href=\"http:\/\/www.perl.com\/doc\/manual\/html\/pod\/perlre.html\">many new operators\nand escape sequences<\/a>.  These additions make the regular\nexpressions more concise, and sometimes more cryptic, but usually\nnot more powerful:\nthese fancy new regular expressions almost always have longer\nequivalents using the traditional syntax.\n<\/p>\n<p class=\"pp\">\nOne common regular expression extension that \ndoes provide additional power is called\n<i>backreferences<\/i>.\nA backreference like\n<code>\\1<\/code>\nor\n<code>\\2<\/code>\nmatches the string matched\nby a previous parenthesized expression, and only that string:\n<code>(cat|dog)\\1<\/code>\nmatches\n<code>catcat<\/code>\nand\n<code>dogdog<\/code>\nbut not\n<code>catdog<\/code>\nnor\n<code>dogcat<\/code>.\nAs far as the theoretical term is concerned,\nregular expressions with backreferences\nare not regular expressions.\nThe power that backreferences add comes at great cost:\nin the worst case, the best known implementations require\nexponential search algorithms,\nlike the one Perl uses.\nPerl (and the other languages)\ncould not now remove backreference support,\nof course, but they could employ much faster algorithms\nwhen presented with regular expressions that don't have\nbackreferences, like the ones considered above.\nThis article is about those faster algorithms.\n<\/p>\n<h2 class=\"sh\">\nFinite Automata\n<\/h2>\n<p class=\"pp\">\nAnother way to describe sets of character strings is with\nfinite automata.\nFinite automata are also known as state machines,\nand we will use \u201cautomaton\u201d and \u201cmachine\u201d interchangeably.\n<\/p>\n<p class=\"pp\">\nAs a simple example, here is a machine recognizing\nthe set of strings matched by the regular expression\n<code>a(bb)+a<\/code>:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig0.png\" alt=\"DFA for a(bb)+a\" width=\"278\" height=\"54\" \/><\/p>\n<p class=\"pp\">\nA finite automaton is always in one of its states,\nrepresented in the diagram by circles.\n(The numbers inside the circles are labels to make this\ndiscussion easier; they are not part of the machine's operation.)\nAs it reads the string, it switches from state to state.\nThis machine has two special states: the start state <i>s<\/i><sub>0<\/sub>\nand the matching state <i>s<\/i><sub>4<\/sub>.\nStart states are depicted with lone arrowheads pointing at them,\nand matching states are drawn as a double circle.\n<\/p>\n<p class=\"pp\">\nThe machine reads an input string one character at a time,\nfollowing arrows corresponding to the input to move from\nstate to state.\nSuppose the input string is\n<code>abbbba<\/code>.\nWhen the machine reads the first letter of the string, the\n<code>a<\/code>,\nit is in the start state <i>s<\/i><sub>0<\/sub>.  It follows the\n<code>a<\/code>\narrow to state <i>s<\/i><sub>1<\/sub>.\nThis process repeats as the machine reads the rest of the string:\n<code>b<\/code>\nto\n<code><i>s<\/i><sub>2<\/sub><\/code>,\n<code>b<\/code>\nto\n<code><i>s<\/i><sub>3<\/sub><\/code>,\n<code>b<\/code>\nto\n<code><i>s<\/i><sub>2<\/sub><\/code>,\n<code>b<\/code>\nto\n<code><i>s<\/i><sub>3<\/sub><\/code>,\nand finally\n<code>a<\/code>\nto\n<code><i>s<\/i><sub>4<\/sub><\/code>.\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig1.png\" alt=\"DFA execution on abbbba\" width=\"357\" height=\"426\" \/><\/p>\n<p class=\"lp\">\nThe machine ends in <i>s<\/i><sub>4<\/sub>, a matching state, so it\nmatches the string.\nIf the machine ends in a non-matching state, it does not \nmatch the string.\nIf, at any point during the machine's execution, there is no\narrow for it to follow corresponding to the current\ninput character, the machine stops executing early.\n<\/p>\n<p class=\"pp\">\nThe machine we have been considering is called a\n<i>deterministic<\/i>\nfinite automaton (DFA),\nbecause in any state, each possible input letter\nleads to at most one new state.\nWe can also create machines\nthat must choose between multiple possible next states.\nFor example, this machine is equivalent to the previous\none but is not deterministic:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig2.png\" alt=\"NFA for a(bb)+a\" width=\"278\" height=\"54\" \/><\/p>\n<p class=\"lp\">\nThe machine is not deterministic because if it reads a\n<code>b<\/code>\nin state <i>s<\/i><sub>2<\/sub>, it has multiple choices for the next state:\nit can go back to <i>s<\/i><sub>1<\/sub> in hopes of seeing another\n<code>bb<\/code>,\nor it can go on to <i>s<\/i><sub>3<\/sub> in hopes of seeing the final\n<code>a<\/code>.\nSince the machine cannot peek ahead to see the rest of\nthe string, it has no way to know which is the correct decision.\nIn this situation, it turns out to be interesting to\nlet the machine\n<i>always guess correctly<\/i>.\nSuch machines are called non-deterministic finite automata\n(NFAs or NDFAs).\nAn NFA matches an input string if there is some way \nit can read the string and follow arrows to a matching state.\n<\/p>\n<p class=\"pp\">\nSometimes it is convenient to let NFAs have arrows with no\ncorresponding input character.  We will leave these arrows unlabeled.\nAn NFA can, at any time, choose to follow an unlabeled arrow\nwithout reading any input.\nThis NFA is equivalent to the previous two, but the unlabeled arrow\nmakes the correspondence with\n<code>a(bb)+a<\/code>\nclearest:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig3.png\" alt=\"Another NFA for a(bb)+a\" width=\"278\" height=\"39\" \/><\/p>\n<h2 class=\"sh\">\nConverting Regular Expressions to NFAs\n<\/h2>\n<p class=\"pp\">\nRegular expressions and NFAs turn out to be exactly\nequivalent in power: every regular expression has an\nequivalent NFA (they match the same strings) and vice versa.\n(It turns out that DFAs are also equivalent in power \nto NFAs and regular expressions; we will see this later.)\nThere are multiple ways to translate regular expressions into NFAs.\nThe method described here was first described by Thompson\nin his 1968 CACM paper.\n<\/p>\n<p class=\"pp\">\nThe NFA for a regular expression is built up from partial NFAs\nfor each subexpression, with a different construction for\neach operator.  The partial NFAs have\nno matching states: instead they have one or more dangling arrows,\npointing to nothing.  The construction process will finish by\nconnecting these arrows to a matching state.\n<\/p>\n<p class=\"pp\">\nThe NFAs for matching single characters look like:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig4.png\" alt=\"Single-character NFA\" width=\"113\" height=\"21\" \/><\/p>\n<p class=\"lp\">\nThe NFA for the concatenation <i>e<\/i><sub>1<\/sub><i>e<\/i><sub>2<\/sub>\nconnects the final arrow of the <i>e<\/i><sub>1<\/sub> \nmachine to the start of the <i>e<\/i><sub>2<\/sub> machine:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig5.png\" alt=\"Concatenation NFA\" width=\"242\" height=\"20\" \/><\/p>\n<p class=\"lp\">\nThe NFA for the alternation <i>e<\/i><sub>1<\/sub><code>|<\/code><i>e<\/i><sub>2<\/sub>\nadds a new start state with a choice of either the\n<i>e<\/i><sub>1<\/sub> machine or the <i>e<\/i><sub>2<\/sub> machine.\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig6.png\" alt=\"Alternation NFA\" width=\"202\" height=\"62\" \/><\/p>\n<p class=\"lp\">\nThe NFA for <i>e<\/i><code>?<\/code> alternates the <i>e<\/i> machine with an empty path:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig7.png\" alt=\"Zero or one NFA\" width=\"184\" height=\"56\" \/><\/p>\n<p class=\"lp\">\nThe NFA for <i>e<\/i><code>*<\/code> uses the same alternation but loops a \nmatching <i>e<\/i> machine back to the start:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig8.png\" alt=\"Zero or more NFA\" width=\"184\" height=\"56\" \/><\/p>\n<p class=\"lp\">\nThe NFA for <i>e<\/i><code>+<\/code> also creates a loop, but one that\nrequires passing through <i>e<\/i> at least once:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig9.png\" alt=\"One or more NFA\" width=\"190\" height=\"41\" \/><\/p>\n<p class=\"pp\">\nCounting the new states in the diagrams above, we can see\nthat this technique creates exactly one state per character\nor metacharacter in the regular expression,\nexcluding parentheses.\nTherefore the number of states in the final NFA is at most\nequal to the length of the original regular expression.\n<\/p>\n<p class=\"pp\">\nJust as with the example NFA discussed earlier, it is always possible\nto remove the unlabeled arrows, and it is also always possible to generate\nthe NFA without the unlabeled arrows in the first place.\nHaving the unlabeled arrows makes the NFA easier for us to read\nand understand, and they also make the C representation\nsimpler, so we will keep them.\n<\/p>\n<h2 class=\"sh\">\nRegular Expression Search Algorithms\n<\/h2>\n<p class=\"pp\">\nNow we have a way to test whether a regular expression\nmatches a string: convert the regular expression to an NFA\nand then run the NFA using the string as input.\nRemember that NFAs are endowed with the ability to guess\nperfectly when faced with a choice of next state:\nto run the NFA using an ordinary computer, we must find\na way to simulate this guessing.\n<\/p>\n<p class=\"pp\">\nOne way to simulate perfect guessing is to guess\none option, and if that doesn't work, try the other.\nFor example, consider the NFA for\n<code>abab|abbb<\/code>\nrun on the string\n<code>abbb<\/code>:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig10.png\" alt=\"NFA for abab|abbb\" width=\"364\" height=\"79\" \/><\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig11.png\" alt=\"Backtracking execution on abbb\" width=\"729\" height=\"619\" \/><\/p>\n<p class=\"lp\">\nAt step 0, the NFA must make a choice: try to match\n<code>abab<\/code>\nor\ntry to match\n<code>abbb<\/code>?\nIn the diagram, the NFA tries\n<code>abab<\/code>,\nbut that fails after step 3.\nThe NFA then tries the other choice, leading to step 4 and eventually a match.\nThis backtracking approach\nhas a simple recursive implementation\nbut can read the input string many times\nbefore succeeding.\nIf the string does not match,\nthe machine must try\n<i>all<\/i>\npossible execution paths before\ngiving up.\nThe NFA tried only two different paths in the example,\nbut in the worst case, there can be exponentially\nmany possible execution paths, leading to very slow run times.\n<\/p>\n<p class=\"pp\">\nA more efficient but more complicated way to simulate perfect\nguessing is to guess both options simultaneously. \nIn this approach, the simulation allows the machine\nto be in multiple states at once.  To process each letter,\nit advances all the states along all the arrows that\nmatch the letter.\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig12.png\" alt=\"Parallel execution on abbb\" width=\"329\" height=\"511\" \/><\/p>\n<p class=\"lp\">\nThe machine starts in the start state and all the states\nreachable from the start state by unlabeled arrows.\nIn steps 1 and 2, the NFA is in two states simultaneously.\nOnly at step 3 does the state set narrow down to a single state.\nThis multi-state approach tries both paths at the same time,\nreading the input only once.\nIn the worst case, the NFA might be in\n<i>every<\/i>\nstate at each step, but this results in at worst a constant amount\nof work independent of the length of the string,\nso arbitrarily\nlarge input strings can be processed in linear time.\nThis is a dramatic improvement over the exponential time\nrequired by the backtracking approach.\nThe efficiency comes from tracking the set of reachable\nstates but\n<i>not<\/i>\nwhich paths were used to reach them.\nIn an NFA with \n<i>n<\/i>\nnodes, there can only be \n<i>n<\/i>\nreachable states at any step, but there might be\n2<sup><i>n<\/i><\/sup> paths through the NFA.\n<\/p>\n<h2 class=\"sh\">\nImplementation\n<\/h2>\n<p class=\"pp\">\nThompson introduced the multiple-state simulation approach\nin his 1968 paper.\nIn his formulation, the states of the NFA were represented\nby small machine-code sequences, and the list of possible states\nwas just a sequence of function call instructions.\nIn essence, Thompson compiled the regular expression into clever\nmachine code.\nForty years later, computers are much faster and the \nmachine code approach is not as necessary.\nThe following sections\npresent an implementation written in portable ANSI C.\nThe full source code (under 400 lines)\nand the benchmarking scripts are \n<a href=\"http:\/\/swtch.com\/~rsc\/regexp\/\">available online<\/a>.\n(Readers who are unfamiliar or uncomfortable with C or pointers should\nfeel free to read the descriptions and skip over the actual code.)\n<\/p>\n<h2 class=\"sh\" id=\"compiling\">\nImplementation: Compiling to NFA\n<\/h2>\n<p class=\"pp\">\nThe first step is to compile the regular expression\ninto an equivalent NFA.\nIn our C program, we will represent an NFA as a\nlinked collection of \n<code>State<\/code>\nstructures:\n<\/p>\n<pre class=\"p1\">\nstruct State\n{\n\tint c;\n\tState *out;\n\tState *out1;\n\tint lastlist;\n};\n<\/pre><p class=\"lp\">\nEach\n<code>State<\/code>\nrepresents one of the following three NFA fragments,\ndepending on the value of\n<code>c<\/code>.\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig13.png\" alt=\"Possible per-State NFA fragments\" width=\"340\" height=\"109\" \/><\/p>\n<p class=\"lp\">\n(<code>Lastlist<\/code>\nis used during execution and is explained in the next section.)\n<\/p>\n<p class=\"pp\">\nFollowing Thompson's paper,\nthe compiler builds an NFA from a regular expression in\n<i>postfix<\/i>\nnotation with dot\n(<code>.<\/code>) added\nas an explicit concatenation operator.\nA separate function\n<code>re2post<\/code>\nrewrites infix regular expressions like\n\u201c<code>a(bb)+a<\/code>\u201d\ninto equivalent postfix expressions like\n\u201c<code>abb.+.a.<\/code>\u201d.\n(A \u201creal\u201d implementation would certainly\nneed to use dot as the \u201cany character\u201d metacharacter\nrather than as a concatenation operator.\nA real implementation would also probably build the \nNFA during parsing rather than build an explicit postfix expression.\nHowever, the postfix version is convenient and follows \nThompson's paper more closely.)\n<\/p>\n<p class=\"pp\">\nAs the compiler scans the postfix expression, it maintains\na stack of computed NFA fragments.\nLiterals push new NFA fragments onto the stack, while\noperators pop fragments off the stack and then\npush a new fragment.\nFor example, \nafter compiling the\n<code>abb<\/code> in <code>abb.+.a.<\/code>,\nthe stack contains NFA fragments for\n<code>a<\/code>,\n<code>b<\/code>,\nand\n<code>b<\/code>.\nThe compilation of the\n<code>.<\/code>\nthat follows pops the two\n<code>b<\/code>\nNFA fragment from the stack and pushes an NFA fragment for the\nconcatenation\n<code>bb.<\/code>.\nEach NFA fragment is defined by its start state and its\noutgoing arrows:\n<\/p><pre class=\"p1\">\nstruct Frag\n{\n\tState *start;\n\tPtrlist *out;\n};\n<\/pre><p class=\"lp\">\n<code>Start<\/code>\npoints at the start state for the fragment,\nand\n<code>out<\/code>\nis a list of pointers to \n<code>State*<\/code>\npointers that are not yet connected to anything.\nThese are the dangling arrows in the NFA fragment.\n<\/p>\n<p class=\"pp\">\nSome helper functions manipulate pointer lists:\n<\/p><pre class=\"p1\">\nPtrlist *list1(State **outp);\nPtrlist *append(Ptrlist *l1, Ptrlist *l2);\nvoid patch(Ptrlist *l, State *s);\n<\/pre><p class=\"lp\">\n<code>List1<\/code>\ncreates a new pointer list containing the single pointer\n<code>outp<\/code>.\n<code>Append<\/code>\nconcatenates two pointer lists, returning the result.\n<code>Patch<\/code>\nconnects the dangling arrows in the pointer list\n<code>l<\/code>\nto the state\n<code>s<\/code>:\nit sets\n<code>*outp<\/code>\n<code>=<\/code>\n<code>s<\/code>\nfor each pointer\n<code>outp<\/code>\nin\n<code>l<\/code>.\n<\/p>\n<p class=\"pp\">\nGiven these primitives and a fragment stack,\nthe compiler is a simple loop over the postfix expression.\nAt the end, there is a single fragment left:\npatching in a matching state completes the NFA.\n<\/p><pre class=\"p1\">\nState*\npost2nfa(char *postfix)\n{\n\tchar *p;\n\tFrag stack[1000], *stackp, e1, e2, e;\n\tState *s;\n\t#define push(s) *stackp++ = s\n\t#define pop()   *--stackp\n\tstackp = stack;\n\tfor(p=postfix; *p; p++){\n\t\tswitch(*p){\n\t\t\/* <i>compilation cases, described below<\/i> *\/\n\t\t}\n\t}\n\t\n\te = pop();\n\tpatch(e.out, matchstate);\n\treturn e.start;\n}\n<\/pre><p class=\"lp\">\nThe specific compilation cases mimic the translation \nsteps described earlier.\n<\/p>\n<table cellpadding=\"0\" border=\"0\" style=\"border-spacing: 0px;\"><tr><td><p class=\"tlp\">\nLiteral characters:\n<\/p><pre class=\"p1\">\ndefault:\n\ts = state(*p, NULL, NULL);\n\tpush(frag(s, list1(&amp;s-&gt;out));\n\tbreak;\n<\/pre>\n<\/td><td><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig14.png\" alt=\"\" width=\"61\" height=\"24\" \/><\/td><\/tr><tr><td><p class=\"tlp\">\nCatenation:\n<\/p><pre class=\"p1\">\ncase '.':\n\te2 = pop();\n\te1 = pop();\n\tpatch(e1.out, e2.start);\n\tpush(frag(e1.start, e2.out));\n\tbreak;\n<\/pre>\n<\/td><td><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig15.png\" alt=\"\" width=\"182\" height=\"20\" \/><\/td><\/tr><tr><td><p class=\"tlp\">\nAlternation:\n<\/p><pre class=\"p1\">\ncase '|':\n\te2 = pop();\n\te1 = pop();\n\ts = state(Split, e1.start, e2.start);\n\tpush(frag(s, append(e1.out, e2.out)));\n\tbreak;\n<\/pre>\n<\/td><td><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig16.png\" alt=\"\" width=\"140\" height=\"62\" \/><\/td><\/tr><tr><td><p class=\"tlp\">\nZero or one:\n<\/p><pre class=\"p1\">\ncase '?':\n\te = pop();\n\ts = state(Split, e.start, NULL);\n\tpush(frag(s, append(e.out, list1(&amp;s-&gt;out1))));\n\tbreak;\n<\/pre>\n<\/td><td><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig17.png\" alt=\"\" width=\"140\" height=\"68\" \/><\/td><\/tr><tr><td><p class=\"tlp\">\nZero or more:\n<\/p><pre class=\"p1\">\ncase '*':\n\te = pop();\n\ts = state(Split, e.start, NULL);\n\tpatch(e.out, s);\n\tpush(frag(s, list1(&amp;s-&gt;out1)));\n\tbreak;\n<\/pre>\n<\/td><td><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig18.png\" alt=\"\" width=\"131\" height=\"68\" \/><\/td><\/tr><tr><td><p class=\"tlp\">\nOne or more:\n<\/p><pre class=\"p1\">\ncase '+':\n\te = pop();\n\ts = state(Split, e.start, NULL);\n\tpatch(e.out, s);\n\tpush(frag(e.start, list1(&amp;s-&gt;out1)));\n\tbreak;\n<\/pre>\n<\/td><td><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig19.png\" alt=\"\" width=\"140\" height=\"53\" \/><\/td><\/tr><\/table><h2 class=\"sh\">\nImplementation: Simulating the NFA\n<\/h2>\n<p class=\"pp\">\nNow that the NFA has been built, we need to simulate it.\nThe simulation requires tracking \n<code>State<\/code>\nsets, which are stored as a simple array list:\n<\/p><pre class=\"p1\">\nstruct List\n{\n\tState **s;\n\tint n;\n};\n<\/pre><p class=\"lp\">\nThe simulation uses two lists:\n<code>clist<\/code>\nis the current set of states that the NFA is in,\nand\n<code>nlist<\/code>\nis the next set of states that the NFA will be in,\nafter processing the current character.\nThe execution loop initializes\n<code>clist<\/code>\nto contain just the start state and then\nruns the machine one step at a time.\n<\/p><pre class=\"p1\">\nint\nmatch(State *start, char *s)\n{\n\tList *clist, *nlist, *t;\n\t\/* l1 and l2 are preallocated globals *\/\n\tclist = startlist(start, &amp;l1);\n\tnlist = &amp;l2;\n\tfor(; *s; s++){\n\t\tstep(clist, *s, nlist);\n\t\tt = clist; clist = nlist; nlist = t;\t\/* swap clist, nlist *\/\n\t}\n\treturn ismatch(clist);\n}\n<\/pre><p class=\"lp\">\nTo avoid allocating on every iteration of the loop,\n<code>match<\/code>\nuses two preallocated lists\n<code>l1<\/code>\nand\n<code>l2<\/code>\nas\n<code>clist<\/code>\nand\n<code>nlist<\/code>,\nswapping the two after each step.\n<\/p>\n<p class=\"pp\">\nIf the final state list contains the matching state,\nthen the string matches.\n<\/p><pre class=\"p1\">\nint\nismatch(List *l)\n{\n\tint i;\n\tfor(i=0; i&lt;l-&gt;n; i++)\n\t\tif(l-&gt;s[i] == matchstate)\n\t\t\treturn 1;\n\treturn 0;\n}\n<\/pre>\n<p class=\"pp\">\n<code>Addstate<\/code>\nadds a state to the list,\nbut not if it is already on the list.\nScanning the entire list for each add would be inefficient;\ninstead the variable\n<code>listid<\/code>\nacts as a list generation number.\nWhen\n<code>addstate<\/code>\nadds\n<code>s<\/code>\nto a list,\nit records\n<code>listid<\/code>\nin\n<code>s-&gt;lastlist<\/code>.\nIf the two are already equal,\nthen \n<code>s<\/code>\nis already on the list being built.\n<code>Addstate<\/code>\nalso follows unlabeled arrows:\nif \n<code>s<\/code>\nis a\n<code>Split<\/code>\nstate with two unlabeled arrows to new states,\n<code>addstate<\/code>\nadds those states to the list instead of\n<code>s<\/code>.\n<\/p><pre class=\"p1\">\nvoid\naddstate(List *l, State *s)\n{\n\tif(s == NULL || s-&gt;lastlist == listid)\n\t\treturn;\n\ts-&gt;lastlist = listid;\n\tif(s-&gt;c == Split){\n\t\t\/* follow unlabeled arrows *\/\n\t\taddstate(l, s-&gt;out);\n\t\taddstate(l, s-&gt;out1);\n\t\treturn;\n\t}\n\tl-&gt;s[l-&gt;n++] = s;\n}\n<\/pre>\n<p class=\"pp\">\n<code>Startlist<\/code>\ncreates an initial state list by adding just the start state:\n<\/p><pre class=\"p1\">\nList*\nstartlist(State *s, List *l)\n{\n\tlistid++;\n\tl-&gt;n = 0;\n\taddstate(l, s);\n\treturn l;\n}\n<\/pre>\n<p class=\"pp\">\nFinally,\n<code>step<\/code>\nadvances the NFA past a single character, using\nthe current list\n<code>clist<\/code>\nto compute the next list\n<code>nlist<\/code>.\n<\/p><pre class=\"p1\">\nvoid\nstep(List *clist, int c, List *nlist)\n{\n\tint i;\n\tState *s;\n\tlistid++;\n\tnlist-&gt;n = 0;\n\tfor(i=0; i&lt;clist-&gt;n; i++){\n\t\ts = clist-&gt;s[i];\n\t\tif(s-&gt;c == c)\n\t\t\taddstate(nlist, s-&gt;out);\n\t}\n}\n<\/pre>\n<h2 class=\"sh\">\nPerformance\n<\/h2>\n<p class=\"pp\">\nThe C implementation just described was not written with performance in mind.\nEven so, a slow implementation of a linear-time algorithm\ncan easily outperform a fast implementation of an \nexponential-time algorithm once the exponent is large enough.\nTesting a variety of popular regular expression engines on \na so-called pathological regular expression demonstrates this nicely.\n<\/p>\n<p class=\"pp\">\nConsider the regular expression\n<code>a?<sup><i>n<\/i><\/sup>a<sup><i>n<\/i><\/sup><\/code>.\nIt matches the string\n<code>a<sup><i>n<\/i><\/sup><\/code>\nwhen the\n<code>a?<\/code>\nare chosen not to match any letters,\nleaving the entire string to be matched by the\n<code>a<sup><i>n<\/i><\/sup><\/code>.\nBacktracking regular expression implementations\nimplement the zero-or-one\n<code>?<\/code>\nby first trying one and then zero.\nThere are\n<i>n<\/i>\nsuch choices to make, a total of\n2<sup><i>n<\/i><\/sup> possibilities.\nOnly the very last\npossibility\u2014choosing zero for all the <code>?<\/code>\u2014will lead to a match.\nThe backtracking approach thus requires\n<i>O<\/i>(2<sup><i>n<\/i><\/sup>) time, so it will not scale much beyond <i>n<\/i>=25.\n<\/p>\n<p class=\"pp\">\nIn contrast, Thompson's algorithm maintains state lists of length\napproximately <i>n<\/i> and processes the string, also of length <i>n<\/i>,\nfor a total of <i>O<\/i>(<i>n<\/i><sup>2<\/sup>) time.\n(The run time is superlinear,\nbecause we are not keeping the regular expression constant\nas the input grows.\nFor a regular expression of length <i>m<\/i> run on text of length <i>n<\/i>,\nthe Thompson NFA requires <i>O<\/i>(<i>mn<\/i>) time.)\n<\/p>\n<p class=\"pp\">\nThe following graph plots time required to check whether\n<code>a?<sup><i>n<\/i><\/sup>a<sup><i>n<\/i><\/sup><\/code>\nmatches\n<code>a<sup><i>n<\/i><\/sup><\/code>:\n<\/p>\n<div class=\"fig\"><div style=\"text-align: center;\">\n<table cellpadding=\"0\" border=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n<div class=\"box\"><div style=\"text-align: center;\">\n<img src=\"https:\/\/swtch.com\/~rsc\/regexp\/grep1p.png\" alt=\"Performance graph\" width=\"779\" height=\"388\" \/><br \/>regular expression and text size <i>n<\/i>\n<br \/><code>a?<\/code><sup><i>n<\/i><\/sup><code>a<\/code><sup><i>n<\/i><\/sup>\nmatching \n<code>a<\/code><sup><i>n<\/i><\/sup><\/div><\/div>\n<\/td><\/tr><\/table><\/div><\/div>\n<p class=\"lp\">\nNotice that the graph's <i>y<\/i>-axis has a logarithmic scale,\nin order to be able to see a wide variety of times on a single graph.\n<\/p>\n<p class=\"pp\">\nFrom the graph it is clear that Perl, PCRE, Python, and Ruby are\nall using recursive backtracking.\nPCRE stops getting the right answer at \n<i>n<\/i>=23,\nbecause it aborts the recursive backtracking after a maximum number\nof steps.\nAs of Perl 5.6, Perl's regular expression engine is\n<a href=\"http:\/\/perlmonks.org\/index.pl?node_id=502408\">said to memoize<\/a>\nthe recursive backtracking search, which should, at some memory cost,\nkeep the search from taking exponential amounts of time \nunless backreferences are being used.\nAs the performance graph shows, the memoization is not complete:\nPerl's run time grows exponentially even though there\nare no backreferences\nin the expression.\nAlthough not benchmarked here, Java uses a backtracking\nimplementation too.\nIn fact, the\n<code>java.util.regex<\/code>\ninterface requires a backtracking\nimplementation, because arbitrary Java code\ncan be substituted into the matching path.\nPHP uses the PCRE library.\n<\/p>\n<p class=\"pp\">\nThe thick blue line is the C implementation of Thompson's algorithm given above.\nAwk, Tcl, GNU grep, and GNU awk \nbuild DFAs, either precomputing them or using the on-the-fly\nconstruction described in the next section.\n<\/p>\n<p class=\"pp\">\nSome might argue that this test is unfair to\nthe backtracking implementations, since it focuses on an\nuncommon corner case.\nThis argument misses the point:\ngiven a choice between an implementation\nwith a predictable, consistent, fast running time on all inputs\nor one that usually runs quickly but can take\nyears of CPU time (or more) on some inputs,\nthe decision should be easy.\nAlso, while examples as dramatic as this one\nrarely occur in practice, less dramatic ones do occur.\nExamples include using\n<code>(.*)<\/code>\n<code>(.*)<\/code>\n<code>(.*)<\/code>\n<code>(.*)<\/code>\n<code>(.*)<\/code>\nto split five space-separated fields, or using\nalternations where the common cases\nare not listed first.\nAs a result, programmers often learn which constructs are\nexpensive and avoid them, or they turn to so-called\n<a href=\"http:\/\/search.cpan.org\/~dankogai\/Regexp-Optimizer-0.15\/lib\/Regexp\/Optimizer.pm\">optimizers<\/a>.\nUsing Thompson's NFA simulation does not require such adaptation:\nthere are no expensive regular expressions.\n<\/p>\n<h2 class=\"sh\">\nCaching the NFA to build a DFA\n<\/h2>\n<p class=\"pp\">\nRecall that DFAs are more efficient to execute than NFAs,\nbecause DFAs are only ever in one state at a time: they never\nhave a choice of multiple next states.\nAny NFA can be converted into an equivalent DFA\nin which each DFA state corresponds to a\nlist of NFA states.\n<\/p>\n<p class=\"pp\">\nFor example, here is the NFA we used earlier for\n<code>abab|abbb<\/code>,\nwith state numbers added:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig20.png\" alt=\"NFA for abab|abbb\" width=\"424\" height=\"91\" \/><\/p>\n<p class=\"lp\">\nThe equivalent DFA would be:\n<\/p>\n<p class=\"fig\"><img src=\"https:\/\/swtch.com\/~rsc\/regexp\/fig21.png\" alt=\"DFA for abab|abbb\" width=\"496\" height=\"170\" \/><\/p>\n<p class=\"lp\">\nEach state in the DFA corresponds to a list of \nstates from the NFA.\n<\/p>\n<p class=\"pp\">\nIn a sense, Thompson's NFA simulation is\nexecuting the equivalent DFA: each\n<code>List<\/code>\ncorresponds to some DFA state,\nand the \n<code>step<\/code>\nfunction is computing, given a list and a next character,\nthe next DFA state to enter.\nThompson's algorithm simulates the DFA by \nreconstructing each DFA state as it is needed.\nRather than throw away this work after each step,\nwe could cache the\n<code>Lists<\/code>\nin spare memory, avoiding the cost of repeating the computation\nin the future\nand essentially computing the equivalent DFA as it is needed.\nThis section presents the implementation of such an approach.\nStarting with the NFA implementation from the previous section,\nwe need to add less than 100 lines to build a DFA implementation.\n<\/p>\n<p class=\"pp\">\nTo implement the cache, we first introduce a new data type\nthat represents a DFA state:\n<\/p><pre class=\"p1\">\nstruct DState\n{\n\tList l;\n\tDState *next[256];\n\tDState *left;\n\tDState *right;\n};\n<\/pre><p class=\"lp\">\nA\n<code>DState<\/code>\nis the cached copy of the list\n<code>l<\/code>.\nThe array\n<code>next<\/code>\ncontains pointers to the next state for each\npossible input character:\nif the current state is\n<code>d<\/code>\nand the next input character is\n<code>c<\/code>,\nthen\n<code>d-&gt;next[c]<\/code>\nis the next state.\nIf\n<code>d-&gt;next[c]<\/code>\nis null, then the next state has not been computed yet.\n<code>Nextstate<\/code>\ncomputes, records, and returns the next state\nfor a given state and character.\n<\/p>\n<p class=\"pp\">\nThe regular expression match follows\n<code>d-&gt;next[c]<\/code>\nrepeatedly, calling\n<code>nextstate<\/code>\nto compute new states as needed.\n<\/p><pre class=\"p1\">\nint\nmatch(DState *start, char *s)\n{\n\tint c;\n\tDState *d, *next;\n\t\n\td = start;\n\tfor(; *s; s++){\n\t\tc = *s &amp; 0xFF;\n\t\tif((next = d-&gt;next[c]) == NULL)\n\t\t\tnext = nextstate(d, c);\n\t\td = next;\n\t}\n\treturn ismatch(&amp;d-&gt;l);\n}\n<\/pre>\n<p class=\"pp\">\nAll the\n<code>DStates<\/code>\nthat have been computed need to be saved in a \nstructure that lets us look up a\n<code>DState<\/code>\nby its\n<code>List<\/code>.\nTo do this, we arrange them \nin a binary tree\nusing the sorted\n<code>List<\/code>\nas the key.\nThe\n<code>dstate<\/code>\nfunction returns the\n<code>DState<\/code>\nfor a given\n<code>List<\/code>,\nallocating one if necessary:\n<\/p><pre class=\"p1\">\nDState*\ndstate(List *l)\n{\n\tint i;\n\tDState **dp, *d;\n\tstatic DState *alldstates;\n\tqsort(l-&gt;s, l-&gt;n, sizeof l-&gt;s[0], ptrcmp);\n\t\/* look in tree for existing DState *\/\n\tdp = &amp;alldstates;\n\twhile((d = *dp) != NULL){\n\t\ti = listcmp(l, &amp;d-&gt;l);\n\t\tif(i &lt; 0)\n\t\t\tdp = &amp;d-&gt;left;\n\t\telse if(i &gt; 0)\n\t\t\tdp = &amp;d-&gt;right;\n\t\telse\n\t\t\treturn d;\n\t}\n\t\n\t\/* allocate, initialize new DState *\/\n\td = malloc(sizeof *d + l-&gt;n*sizeof l-&gt;s[0]);\n\tmemset(d, 0, sizeof *d);\n\td-&gt;l.s = (State**)(d+1);\n\tmemmove(d-&gt;l.s, l-&gt;s, l-&gt;n*sizeof l-&gt;s[0]);\n\td-&gt;l.n = l-&gt;n;\n\t\/* insert in tree *\/\n\t*dp = d;\n\treturn d;\n}\n<\/pre><p class=\"lp\">\nNextstate runs the NFA\n<code>step<\/code>\nand returns the corresponding\n<code>DState<\/code>:\n<\/p><pre class=\"p1\">\nDState*\nnextstate(DState *d, int c)\n{\n\tstep(&amp;d-&gt;l, c, &amp;l1);\n\treturn d-&gt;next[c] = dstate(&amp;l1);\n}\n<\/pre><p class=\"lp\">\nFinally, the DFA's start state is the\n<code>DState<\/code>\ncorresponding to the NFA's start list:\n<\/p><pre class=\"p1\">\nDState*\nstartdstate(State *start)\n{\n\treturn dstate(startlist(start, &amp;l1));\n}\n<\/pre><p class=\"lp\">\n(As in the NFA simulation,\n<code>l1<\/code>\nis a preallocated\n<code>List<\/code>.)\n<\/p>\n<p class=\"pp\">\nThe\n<code>DStates<\/code>\ncorrespond to DFA states, but the DFA is only built as needed:\nif a DFA state has not been encountered during the search,\nit does not yet exist in the cache.\nAn alternative would be to compute the entire DFA at once.\nDoing so would make\n<code>match<\/code>\na little faster by removing the conditional branch,\nbut at the cost of increased startup time and\nmemory use.\n<\/p>\n<p class=\"pp\">\nOne might also worry about bounding the amount of\nmemory used by the on-the-fly DFA construction.\nSince the\n<code>DStates<\/code>\nare only a cache of the \n<code>step<\/code>\nfunction, the implementation of\n<code>dstate<\/code>\ncould choose to throw away the entire DFA so far\nif the cache grew too large.\nThis cache replacement policy \nonly requires a few extra lines of code in \n<code>dstate<\/code>\nand in\n<code>nextstate<\/code>,\nplus around 50 lines of code for memory management.\nAn implementation is\n<a href=\"http:\/\/swtch.com\/~rsc\/regexp\/\">available online<\/a>.\n(<a href=\"http:\/\/cm.bell-labs.com\/cm\/cs\/awkbook\/\">Awk<\/a>\nuses a similar limited-size cache strategy,\nwith a fixed limit of 32 cached states; this explains the discontinuity\nin its performance at <i>n<\/i>=28 in the graph above.)\n<\/p>\n<p class=\"pp\">\nNFAs derived from regular expressions\ntend to exhibit good locality: they visit the same states\nand follow the same transition arrows over and over\nwhen run on most texts.\nThis makes the caching worthwhile: the first time an arrow\nis followed, the next state must be computed as in the NFA\nsimulation, but future traversals of the arrow are just\na single memory access.\nReal DFA-based implementations can make use\nof additional optimizations to run even faster.\nA companion article (not yet written) will explore\nDFA-based regular expression implementations in more detail.\n<\/p>\n<h2 class=\"sh\">\nReal world regular expressions\n<\/h2>\n<p class=\"pp\">\nRegular expression usage in real programs\nis somewhat more complicated than what the regular expression\nimplementations described above can handle.\nThis section briefly describes the common complications;\nfull treatment of any of these is beyond the scope of this\nintroductory article.\n<\/p>\n<p class=\"pp\">\n<i>Character classes<\/i>.\nA character class, whether \n<code>[0-9]<\/code>\nor\n<code>\\w<\/code>\nor\n<code>.<\/code> (dot),\nis just a concise representation of an alternation.\nCharacter classes can be expanded into alternations\nduring compilation, though it is more efficient to add\na new kind of NFA node to represent them explicitly.\n<a href=\"http:\/\/www.opengroup.org\/onlinepubs\/009695399\/basedefs\/xbd_chap09.html\">POSIX<\/a>\ndefines special character classes\nlike <code>[[:upper:]]<\/code> that change meaning\ndepending on the current locale, but the hard part of\naccommodating these is determining their meaning,\nnot encoding that meaning into an NFA.\n<\/p>\n<p class=\"pp\">\n<i>Escape sequences<\/i>.\nReal regular expression syntaxes need to handle\nescape sequences, both as a way to match metacharacters\n(<code>\\(<\/code>,\n<code>\\)<\/code>,\n<code>\\\\<\/code>,\netc.)\nand to specify otherwise difficult-to-type characters such as\n<code>\\n<\/code>.\n<\/p>\n<p class=\"pp\">\n<i>Counted repetition<\/i>.\nMany regular expression implementations provide a counted\nrepetition operator\n<code>{<i>n<\/i>}<\/code>\nto match exactly \n<i>n<\/i>\nstrings matching a pattern;\n<code>{<\/code><i>n<\/i><code>,<\/code><i>m<\/i><code>}<\/code>\nto match at least \n<i>n<\/i>\nbut no more than\n<i>m<\/i>;\nand\n<code>{<\/code><i>n<\/i><code>,}<\/code>\nto match\n<i>n<\/i>\nor more.\nA recursive backtracking implementation can implement\ncounted repetition using a loop; an NFA or DFA-based\nimplementation must expand the repetition:\n<i>e<\/i><code>{3}<\/code>\nexpands to\n<i>eee<\/i>;\n<i>e<\/i><code>{3,5}<\/code>\nexpands to\n<i>eeee<\/i><code>?<\/code><i>e<\/i><code>?<\/code>,\nand\n<i>e<\/i><code>{3,}<\/code>\nexpands to\n<i>eee<\/i><code>+<\/code>.\n<\/p>\n<p class=\"pp\">\n<i>Submatch extraction<\/i>.\nWhen regular expressions are used for splitting or parsing strings,\nit is useful to be able to find out which sections of the input string\nwere matched by each subexpression.\nAfter a regular expression like\n<code>([0-9]+-[0-9]+-[0-9]+)<\/code>\n<code>([0-9]+:[0-9]+)<\/code>\nmatches a string (say a date and time),\nmany regular expression engines make the\ntext matched by each parenthesized expression\navailable.\nFor example, one might write in Perl:\n<\/p><pre class=\"p1\">\nif(\/([0-9]+-[0-9]+-[0-9]+) ([0-9]+:[0-9]+)\/){\n\tprint \"date: $1, time: $2\\n\";\n}\n<\/pre><p class=\"lp\">\nThe extraction of submatch boundaries has been mostly ignored\nby computer science theorists, and it is perhaps the most\ncompelling argument for using recursive backtracking.\nHowever, Thompson-style algorithms can be adapted to\ntrack submatch boundaries without giving up efficient performance.\nThe Eighth Edition Unix\n<i>regexp<\/i>(3)\nlibrary implemented such an algorithm as early as 1985,\nthough as explained below,\nit was not very widely used or even noticed.\n<\/p>\n<p class=\"pp\">\n<i>Unanchored matches<\/i>.\nThis article has assumed that regular expressions\nare matched against an entire input string.\nIn practice, one often wishes to find a substring\nof the input that matches the regular expression.\nUnix tools traditionally return the longest matching substring\nthat starts at the leftmost possible point in the input.\nAn unanchored search for \n<i>e<\/i>\nis a special case\nof submatch extraction: it is like searching for\n<code>.*(<i>e<\/i>).*<\/code>\nwhere the first\n<code>.*<\/code>\nis constrained to match as short a string as possible.\n<\/p>\n<p class=\"pp\">\n<i>Non-greedy operators<\/i>.\nIn traditional Unix regular expressions, the repetition operators\n<code>?<\/code>,\n<code>*<\/code>,\nand\n<code>+<\/code>\nare defined to match as much of the string as possible while\nstill allowing the entire regular expression to match:\nwhen matching\n<code>(.+)(.+)<\/code>\nagainst\n<code>abcd<\/code>,\nthe first\n<code>(.+)<\/code>\nwill match\n<code>abc<\/code>,\nand the second\nwill match\n<code>d<\/code>.\nThese operators are now called\n<i>greedy<\/i>.\nPerl introduced\n<code>??<\/code>,\n<code>*?<\/code>,\nand\n<code>+?<\/code>\nas non-greedy versions, which match as little of the string\nas possible while preserving the overall match:\nwhen matching\n<code>(.+?)(.+?)<\/code>\nagainst\n<code>abcd<\/code>,\nthe first\n<code>(.+?)<\/code>\nwill match only\n<code>a<\/code>,\nand the second\nwill match\n<code>bcd.<\/code>\nBy definition, whether an operator is greedy\ncannot affect whether a regular expression matches a\nparticular string as a whole; it only affects the\nchoice of submatch boundaries.\nThe backtracking algorithm admits a simple implementation\nof non-greedy operators:\ntry the shorter match before the longer one.\nFor example, in a standard backtracking implementation,\n<code><i>e<\/i>?<\/code>\nfirst tries using\n<i>e<\/i>\nand then tries not using it;\n<code><i>e<\/i>??<\/code>\nuses the other order.\nThe submatch-tracking variants of Thompson's algorithm\ncan be adapted to accommodate non-greedy operators.\n<\/p>\n<p class=\"pp\">\n<i>Assertions<\/i>.\nThe traditional regular expression metacharacters\n<code>^<\/code>\nand\n<code>$<\/code>\ncan be viewed as\n<i>assertions<\/i>\nabout the text around them:\n<code>^<\/code>\nasserts that the previous character\nis a newline (or the beginning of the string),\nwhile\n<code>$<\/code>\nasserts that the next character is a newline\n(or the end of the string).\nPerl added more assertions, like\nthe word boundary\n<code>\\b<\/code>,\nwhich asserts that \nthe previous character is alphanumeric but the next\nis not, or vice versa.\nPerl also generalized the idea to arbitrary\nconditions called lookahead assertions:\n<code>(?=<\/code><i>re<\/i><code>)<\/code>\nasserts that the text after the current input position matches\n<i>re<\/i>,\nbut does not actually advance the input position;\n<code>(?!<\/code><i>re<\/i><code>)<\/code>\nis similar but \nasserts that the text does not match\n<i>re<\/i>.\nThe lookbehind assertions\n<code>(?&lt;=<\/code><i>re<\/i><code>)<\/code>\nand\n<code>(?&lt;!<\/code><i>re<\/i><code>)<\/code>\nare similar but make assertions about the text\nbefore the current input position.\nSimple assertions like\n<code>^<\/code>,\n<code>$<\/code>,\nand\n<code>\\b<\/code>\nare easy to accommodate in an NFA,\ndelaying the match one byte for forward assertions.\nThe generalized assertions\nare harder to accommodate but in principle could\nbe encoded in the NFA.\n<\/p>\n<p class=\"pp\">\n<i>Backreferences<\/i>.\nAs mentioned earlier, no one knows how to \nimplement regular expressions with backreferences efficiently,\nthough no one can prove that it's impossible either.\n(Specifically, the \n<a href=\"http:\/\/perl.plover.com\/NPC\/NPC-3SAT.html\">problem is NP-complete<\/a>, meaning that if\nsomeone did find an efficient implementation, that would\nbe <i>major<\/i> news to computer scientists and would\nwin a <a href=\"http:\/\/www.claymath.org\/Popular_Lectures\/Minesweeper\/\">million dollar prize<\/a>.)\nThe simplest, most effective strategy for backreferences,\ntaken by the original awk and egrep, is not to implement them.\nThis strategy is no longer practical: users have come to\nrely on backreferences for at least occasional use,\nand backreferences are part of\nthe\n<a href=\"http:\/\/www.opengroup.org\/onlinepubs\/009695399\/basedefs\/xbd_chap09.html\">POSIX standard for regular expressions<\/a>.\nEven so, it would be reasonable to use Thompson's NFA simulation\nfor most regular expressions, and only bring out\nbacktracking when it is needed.\nA particularly clever implementation could combine the two,\nresorting to backtracking only to accommodate the backreferences.\n<\/p>\n<p class=\"pp\">\n<i>Backtracking with memoization<\/i>.\nPerl's approach of using memoization to avoid exponential blowup\nduring backtracking\nwhen possible is a good one.  At least in theory, it should make\nPerl's regular expressions behave more like an NFA and\nless like backtracking.  \nMemoization does not completely solve the problem, though:\nthe memoization itself requires a memory footprint roughly \nequal to the size of the text times the size of the regular expression.\nMemoization also does not address the issue of the stack space used\nby backtracking, which is linear in the size of the text:\nmatching long strings typically causes a backtracking\nimplementation to run out of stack space:\n<\/p><pre class=\"p1\">\n$ perl -e '(\"a\" x 100000) =~ \/^(ab?)*$\/;'\nSegmentation fault (core dumped)\n$\n<\/pre>\n<p class=\"pp\">\n<i>Character sets<\/i>.\nModern regular expression implementations must deal with \nlarge non-ASCII character sets such as Unicode.\nThe \n<a href=\"http:\/\/swtch.com\/plan9port\/unix\/\">Plan 9 regular expression library<\/a>\nincorporates Unicode by running an NFA with a\nsingle Unicode character as the input character for each step.\nThat library separates the running of the NFA from decoding\nthe input, so that the same regular expression matching code\nis used for both \n<a href=\"http:\/\/plan9.bell-labs.com\/sys\/doc\/utf.html\">UTF-8<\/a>\nand wide-character inputs.\n<\/p>\n<h2 class=\"sh\" id=\"History\">\nHistory and References\n<\/h2>\n<p class=\"pp\">\nMichael Rabin and Dana Scott\nintroduced non-deterministic finite automata\nand the concept of non-determinism in 1959\n[<a href=\"#rabin-scott\">7<\/a>],\nshowing that NFAs can be simulated by\n(potentially much larger) DFAs in which \neach DFA state corresponds to a set of NFA states.\n(They won the Turing Award in 1976 for the introduction\nof the concept of non-determinism in that paper.)\n<\/p>\n<p class=\"pp\">\n<a name=\"mcnaughton-yamada-b\">R. McNaughton and H. Yamada\n<\/a>[<a href=\"#mcnaughton-yamada\">4<\/a>]\nand \nKen Thompson\n[<a href=\"#thompson\">9<\/a>]\nare commonly credited with giving the first constructions\nto convert regular expressions into NFAs,\neven though neither paper mentions the\nthen-nascent concept of an NFA.\nMcNaughton and Yamada's construction\ncreates a DFA,\nand Thompson's construction creates IBM 7094 machine code,\nbut reading between the lines one can\nsee latent NFA constructions underlying both.\nRegular expression to NFA constructions differ only in how they encode \nthe choices that the NFA must make.\nThe approach used above, mimicking Thompson,\nencodes the choices with explicit choice\nnodes\n(the\n<code>Split<\/code>\nnodes above)\nand unlabeled arrows.\nAn alternative approach,\nthe one most commonly credited to McNaughton and Yamada,\nis to avoid unlabeled arrows, instead allowing NFA states to\nhave multiple outgoing arrows with the same label.\nMcIlroy\n[<a href=\"#mcilroy\">3<\/a>]\ngives a particularly elegant implementation of this approach\nin Haskell.\n<\/p>\n<p class=\"pp\">\nThompson's regular expression implementation\nwas for his QED editor running on the CTSS \n[<a href=\"#vanvleck\">10<\/a>]\noperating\nsystem on the IBM 7094.\nA copy of the editor can be found in archived CTSS sources\n[<a href=\"#pierce\">5<\/a>].\nL. Peter Deutsch and Butler Lampson\n[<a href=\"#deutsch-lampson\">1<\/a>]\ndeveloped the first QED, but\nThompson's reimplementation was the first to use\nregular expressions.\nDennis Ritchie, author of yet another QED implementation,\nhas documented the early history of the QED editor\n[<a href=\"#ritchie\">8<\/a>]\n(Thompson, Ritchie, and Lampson later won\nTuring awards for work unrelated to QED or finite automata.)\n<\/p>\n<p class=\"pp\">\nThompson's paper marked the \nbeginning of a long line of regular expression implementations.\nThompson chose not to use his algorithm when \nimplementing the text editor ed, which appeared in \nFirst Edition Unix (1971), or in its descendant grep,\nwhich first appeared in the Fourth Edition (1973).\nInstead, these venerable Unix tools used\nrecursive backtracking!\nBacktracking was justifiable because the\nregular expression syntax was quite limited:\nit omitted grouping parentheses and the\n<code>|<\/code>,\n<code>?<\/code>,\nand\n<code>+<\/code>\noperators.\nAl Aho's egrep,\nwhich first appeared in the Seventh Edition (1979),\nwas the first Unix tool to provide\nthe full regular expression syntax, using a\nprecomputed DFA.\nBy the Eighth Edition (1985), egrep computed the DFA on the fly,\nlike the implementation given above.\n<\/p>\n<p class=\"pp\">\nWhile writing the text editor sam \n[<a href=\"#pike\">6<\/a>]\nin the early 1980s,\nRob Pike wrote a new regular expression implementation,\nwhich Dave Presotto extracted into a library that \nappeared in the Eighth Edition.\nPike's implementation\nincorporated submatch tracking into an efficient NFA simulation\nbut, like the rest of the Eighth Edition source, was not widely\ndistributed.\nPike himself did not realize that his technique was anything new.\nHenry Spencer reimplemented the Eighth Edition library\ninterface from scratch, but using backtracking,\nand\n<a href=\"http:\/\/arglist.com\/regex\/\">released his implementation<\/a>\ninto the public domain.\nIt became very widely used, eventually serving as the basis\nfor the slow regular expression implementations\nmentioned earlier: Perl, PCRE, Python, and so on.\n(In his defense,\nSpencer knew the routines could be slow,\nand he didn't know that a more efficient algorithm existed.\nHe even warned in the documentation,\n\u201cMany users have found the speed perfectly adequate,\nalthough replacing the insides of egrep with this code\nwould be a mistake.\u201d)\nPike's regular expression implementation, extended to\nsupport Unicode, was made freely available\nwith sam in \n<a href=\"http:\/\/groups.google.com\/group\/comp.os.research\/msg\/f1783504a2d18051\">late 1992<\/a>,\nbut the particularly efficient\nregular expression search algorithm went unnoticed.\nThe code is now available in many forms: as \n<a href=\"http:\/\/plan9.bell-labs.com\/sources\/plan9\/sys\/src\/cmd\/sam\/\">part of sam<\/a>,\nas \n<a href=\"http:\/\/plan9.bell-labs.com\/sources\/plan9\/sys\/src\/libregexp\/\">Plan\u00a09's regular expression library<\/a>,\nor\n<a href=\"http:\/\/swtch.com\/plan9port\/unix\/\">packaged separately for Unix<\/a>.\nVille Laurikari independently discovered Pike's algorithm\nin 1999, developing a theoretical foundation as well\n[<a href=\"#laurikari\">2<\/a>].\n<\/p>\n<p class=\"pp\">\nFinally, any discussion of regular expressions\nwould be incomplete without mentioning \nJeffrey Friedl's book\n<i>Mastering Regular Expressions<\/i>,\nperhaps the most popular reference among today's programmers.\nFriedl's book teaches programmers how best to use today's\nregular expression implementations, but not how best to implement them.\nWhat little text it devotes to implementation\nissues perpetuates the widespread belief that recursive backtracking\nis the only way to simulate an NFA.\nFriedl makes it clear that he \n<a href=\"http:\/\/regex.info\/blog\/2006-09-15\/248\">neither understands nor respects<\/a>\nthe underlying theory.\n<\/p>\n<h2 class=\"sh\">\nSummary\n<\/h2>\n<p class=\"pp\">\nRegular expression matching can be simple and fast, using\nfinite automata-based techniques that have been known for decades.\nIn contrast, Perl, PCRE, Python, Ruby, Java,\nand many other languages\nhave regular expression implementations based on \nrecursive backtracking that are simple but can be\nexcruciatingly slow.\nWith the exception of backreferences, the features\nprovided by the slow backtracking implementations\ncan be provided by the automata-based implementations\nat dramatically faster, more consistent speeds.\n<\/p>\n<p class=\"pp\">\nThe next article in this series,\n\u201c<a href=\"https:\/\/swtch.com\/~rsc\/regexp\/regexp2.html\">Regular Expression Matching: the Virtual Machine Approach<\/a>,\u201d discusses NFA-based submatch extraction.\nThe third article, \u201c<a href=\"https:\/\/swtch.com\/~rsc\/regexp\/regexp3.html\">Regular Expression Matching in the Wild<\/a>,\u201d examines a production implementation.\nThe fourth article, \u201c<a href=\"https:\/\/swtch.com\/~rsc\/regexp\/regexp4.html\">Regular Expression Matching with a Trigram Index<\/a>,\u201d explains how Google Code Search was implemented.\n<\/p>\n<h2 class=\"sh\">\nAcknowledgements\n<\/h2>\n<p class=\"pp\">\nLee Feigenbaum,\nJames Grimmelmann,\nAlex Healy,\nWilliam Josephson,\nand\nArnold Robbins\nread drafts of this article and made many helpful suggestions.\nRob Pike clarified some of the history surrounding his\nregular expression implementation.\nThanks to all.\n<\/p>\n<h2 class=\"sh\">\nReferences\n<\/h2>\n<p class=\"lp-left\">\n[<a href=\"#deutsch-lampson-b\">1<\/a>]\nL. Peter Deutsch and Butler Lampson,\n\u201cAn online editor,\u201d\nCommunications of the ACM 10(12) (December 1967), pp.\u00a0793\u2013799.\n<a href=\"http:\/\/doi.acm.org\/10.1145\/363848.363863\"><i>http:\/\/doi.acm.org\/10.1145\/363848.363863<\/i><\/a>\n<\/p><p class=\"lp-left\">\n[<a href=\"#laurikari-b\">2<\/a>]\nVille Laurikari,\n\u201cNFAs with Tagged Transitions,\ntheir Conversion to Deterministic Automata\nand\nApplication to Regular Expressions,\u201d\nin Proceedings of the Symposium on String Processing and\nInformation Retrieval, September 2000.\n<a href=\"http:\/\/laurikari.net\/ville\/spire2000-tnfa.ps\"><i>http:\/\/laurikari.net\/ville\/spire2000-tnfa.ps<\/i><\/a>\n<\/p><p class=\"lp-left\">\n[<a href=\"#mcilroy-b\">3<\/a>]\nM. Douglas McIlroy,\n\u201cEnumerating the strings of regular languages,\u201d\nJournal of Functional Programming 14 (2004), pp.\u00a0503\u2013518.\n<a href=\"http:\/\/www.cs.dartmouth.edu\/~doug\/nfa.ps.gz\"><i>http:\/\/www.cs.dartmouth.edu\/~doug\/nfa.ps.gz<\/i><\/a> (preprint)\n<\/p><p class=\"lp-left\">\n[<a href=\"#mcnaughton-yamada-b\">4<\/a>]\nR. McNaughton and H. Yamada,\n\u201cRegular expressions and state graphs for automata,\u201d\nIRE Transactions on Electronic Computers EC-9(1) (March 1960), pp.\u00a039\u201347.\n<\/p><p class=\"lp-left\">\n[<a href=\"#pierce-b\">5<\/a>]\nPaul Pierce,\n\u201cCTSS source listings.\u201d\n<a href=\"http:\/\/www.piercefuller.com\/library\/ctss.html\"><i>http:\/\/www.piercefuller.com\/library\/ctss.html<\/i><\/a> \n(Thompson's QED is in the file\n<code>com5<\/code>\nin the source listings archive and is marked as\n<code>0QED<\/code>)\n<\/p><p class=\"lp-left\">\n[<a href=\"#pike-b\">6<\/a>]\nRob Pike,\n\u201cThe text editor sam,\u201d\nSoftware\u2014Practice &amp; Experience 17(11) (November 1987), pp.\u00a0813\u2013845.\n<a href=\"http:\/\/plan9.bell-labs.com\/sys\/doc\/sam\/sam.html\"><i>http:\/\/plan9.bell-labs.com\/sys\/doc\/sam\/sam.html<\/i><\/a>\n<\/p><p class=\"lp-left\">\n[<a href=\"#rabin-scott-b\">7<\/a>]\nMichael Rabin and Dana Scott,\n\u201cFinite automata and their decision problems,\u201d\nIBM Journal of Research and Development 3 (1959), pp.\u00a0114\u2013125.\n<a href=\"http:\/\/www.research.ibm.com\/journal\/rd\/032\/ibmrd0302C.pdf\"><i>http:\/\/www.research.ibm.com\/journal\/rd\/032\/ibmrd0302C.pdf<\/i><\/a>\n<\/p><p class=\"lp-left\">\n[<a href=\"#ritchie-b\">8<\/a>]\nDennis Ritchie,\n\u201cAn incomplete history of the QED text editor.\u201d\n<a href=\"http:\/\/plan9.bell-labs.com\/~dmr\/qed.html\"><i>http:\/\/plan9.bell-labs.com\/~dmr\/qed.html<\/i><\/a>\n<\/p><p class=\"lp-left\">\n[<a href=\"#thompson-b\">9<\/a>]\nKen Thompson,\n\u201cRegular expression search algorithm,\u201d\nCommunications of the ACM 11(6) (June 1968), pp.\u00a0419\u2013422.\n<a href=\"http:\/\/doi.acm.org\/10.1145\/363347.363387\"><i>http:\/\/doi.acm.org\/10.1145\/363347.363387<\/i><\/a>\n(<span><a href=\"http:\/\/www.cs.chalmers.se\/~coquand\/AUTOMATA\/thompson.pdf\">PDF<\/a><\/span>)\n<\/p><p class=\"lp-left\">\n[<a href=\"#vanvleck-b\">10<\/a>]\nTom Van Vleck,\n\u201cThe IBM 7094 and CTSS.\u201d\n<a href=\"http:\/\/www.multicians.org\/thvv\/7094.html\"><i>http:\/\/www.multicians.org\/thvv\/7094.html<\/i><\/a>\n<\/p>\n<p class=\"lp-left\">\nDiscussion on <a href=\"http:\/\/programming.reddit.com\/info\/10c60\/comments\">reddit<\/a> and <a href=\"http:\/\/perlmonks.org\/?node_id=597262\">perlmonks<\/a> and\n<a href=\"http:\/\/lambda-the-ultimate.org\/node\/2064\">LtU<\/a>\n<\/p>\n<div style=\"text-align: center;\">\n<p class=\"copy\">\nCopyright \u00a9 2007 Russ Cox.  All Rights Reserved.\n<br \/><a href=\"http:\/\/swtch.com\/~rsc\/regexp\/\">http:\/\/swtch.com\/~rsc\/regexp\/<\/a>\n<\/p>\n<\/div>","created_at":"2018-12-24T03:08:25+0000","updated_at":"2018-12-24T03:08:31+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":35,"domain_name":"swtch.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":756,"title":"12","url":"https:\/\/tour.golang.org\/basics\/12","created_at":"2018-12-23T13:49:20+0000","updated_at":"2018-12-23T13:49:20+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"tour.golang.org"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":755,"title":"PDF","url":"https:\/\/www.doc.ic.ac.uk\/~susan\/475\/HowToBeAProgrammer.pdf","content":"HowTo Be aProgrammer: AShort,<br \/>\nComprehensiv e,and Personal Summary<br \/>\nRob ertL.Read <br \/>\nDecem ber 16, 2002<br \/>\nDedicated tothe Programmers ofHire.com<br \/>\nCon tents<br \/>\n1 Intro duction 1<br \/>\nI Beginner 2<br \/>\n2 Personal Skills 2<br \/>\n2.1 Learn toDebug .... .. .. ... .. .. ... .. ... .. .. . 2<br \/>\n2.2 Howto Debug by Splitting theProblem Space.. ... .. .. . 3<br \/>\n2.3 Howto Debug UsingaLog .... .. .. ... .. ... .. .. . 4<br \/>\n2.4 Howto Understand Performance Problems.. .. ... .. .. . 5<br \/>\n2.5 Howto Fix Performance Problems.. .. ... .. ... .. .. . 5<br \/>\n2.6 Howto Optimize Loops ... ... .. .. ... .. ... .. .. . 6<br \/>\n2.7 Howto Deal withI\/OExpense .. .. .. ... .. ... .. .. . 7<br \/>\n2.8 Howto Manage Memory ... ... .. .. ... .. ... .. .. . 7<br \/>\n2.9 Howto Deal withIntermitten tBugs ... ... .. ... .. .. . 8<br \/>\n2.10 Howto Learn Design Skills.... .. .. ... .. ... .. .. . 9<br \/>\n3 Team Skills 9<br \/>\n3.1 Whyestimation isimp ortan t... .. .. ... .. ... .. .. . 9<br \/>\n3.2 Howto Estimate Programming Time... ... .. ... .. .. . 10<br \/>\n3.3 Howto Find Information .. ... .. .. ... .. ... .. .. . 11<br \/>\n3.4 Howto Utilize People asInformation Sources... ... .. .. . 12<br \/>\n3.5 Howto Do cumen tWisely .. ... .. .. ... .. ... .. .. . 13<br \/>\n3.6 Howto Work with Poor Co de ... .. .. ... .. ... .. .. . 14<br \/>\n3.7 Howto Use Source Code Con trol ... .. ... .. ... .. .. . 14<br \/>\n3.8 Howto Unit Test ... .. .. ... .. .. ... .. ... .. .. . 15<br \/>\n3.9 Howto Stress Test .. .. .. ... .. .. ... .. ... .. .. . 15<br \/>\n3.10 HowTo Recognize WhenTo Break orGo Home .... .. .. . 16<br \/>\n3.11 Howto Deal withDi cult People ... .. ... .. ... .. .. . 17<br \/>\n1<br \/>\n<br \/>\nII Intermediate 18<br \/>\n4 Personal Skills 18<br \/>\n4.1 Howto Sta yMotiv ated ... ... .. .. ... .. ... .. .. . 18<br \/>\n4.2 Howto Tradeo  Timeversus Space .. .. ... .. ... .. .. . 18<br \/>\n4.3 Howto Balance Brevityand Abstraction .... .. ... .. .. . 19<br \/>\n4.4 Howto Learn NewSkills ... ... .. .. ... .. ... .. .. . 20<br \/>\n4.5 Howto do Integration Testing ... .. .. ... .. ... .. .. . 20<br \/>\n5 Team Skills 20<br \/>\n5.1 Howto Manage Developmen tTime .. .. ... .. ... .. .. . 20<br \/>\n5.2 Howto Manage ThirdPart ySoft ware Risks .. .. ... .. .. . 22<br \/>\n5.3 Howto Manage Consultan ts.... .. .. ... .. ... .. .. . 22<br \/>\n5.4 Howto Comm unicate theRigh tAmoun t. ... .. ... .. .. . 23<br \/>\n5.5 Howto Disagree Honestly andGetAway with It.... .. .. . 23<br \/>\n6 Judgmen t 24<br \/>\n6.1 Howto Tradeo  QualityAgainst Developmen tTime .. .. .. . 24<br \/>\n6.2 Howto Manage Software System Dependence ... ... .. .. . 24<br \/>\n6.3 Howto Decide IfSoft ware IsToo Immature .. .. ... .. .. . 25<br \/>\n6.4 Howto Mak ea Buy vs.Build Decision .. ... .. ... .. .. . 25<br \/>\n6.5 Howto Gro wProfessionally .... .. .. ... .. ... .. .. . 26<br \/>\n6.6 Howto Evaluate Interview ees... .. .. ... .. ... .. .. . 26<br \/>\n6.7 Howto Kno wWhen toApply Fancy Computer Science.. .. . 27<br \/>\n6.8 Howto Talk toNon-Engineers .. .. .. ... .. ... .. .. . 28<br \/>\nI II Adv anced 28<br \/>\n7 Tec hnological Judgment 29<br \/>\n7.1 Howto Tell the Hard From theImp ossible ... .. ... .. .. . 29<br \/>\n7.2 Howto Utilize Embedded Languages ... ... .. ... .. .. . 29<br \/>\n7.2.1 Choosing Languages .... .. .. ... .. ... .. .. . 30<br \/>\n8 Compromising Wisely 30<br \/>\n8.1 Howto Figh tSc hedule Pressure .. .. .. ... .. ... .. .. . 30<br \/>\n8.2 Howto Understand theUser ... .. .. ... .. ... .. .. . 31<br \/>\n8.3 Howto Get aPromotion ... ... .. .. ... .. ... .. .. . 31<br \/>\n9 Serving Your Team 32<br \/>\n9.1 Howto Dev elop Talen t. . .. ... .. .. ... .. ... .. .. . 32<br \/>\n9.2 Howto Cho oseWhat toWork On .. .. ... .. ... .. .. . 33<br \/>\n9.3 Howto Get theMost From Your Teammates ... ... .. .. . 33<br \/>\n9.4 Howto Divide Problems Up.... .. .. ... .. ... .. .. . 33<br \/>\n9.5 Howto Gather SupportforaPro ject ... ... .. ... .. .. . 34<br \/>\n9.6 Howto Gro waSystem .. .. ... .. .. ... .. ... .. .. . 34<br \/>\n9.7 Howto Comm unicate Well .... .. .. ... .. ... .. .. . 35<br \/>\n2<br \/>\n<br \/>\n9.8Howto Tell People Things TheyDon't Wan tto Hear ... .. . 35<br \/>\n9.9 Howto Deal withManagerial Myths... ... .. ... .. .. . 36<br \/>\n9.10 Howto Deal withOrganizational Chaos.... .. ... .. .. . 36<br \/>\n9.11 Request forFeedbac k... .. ... .. .. ... .. ... .. .. . 37<br \/>\nA Glossary 37<br \/>\n1 Intro duction<br \/>\nT o be a go od programmer isdi cult andnoble. Thehardest partofmak-<br \/>\ning real acollectiv evision ofasoft ware project isdealing withone's cowork ers<br \/>\nand customers. Writing computer programsisimp ortan tand takesgreat in-<br \/>\ntelligence andskill. Butitis really child's playcompared toeverything else<br \/>\nthat ago od programmer must dotomak ea soft ware system thatsucceeds<br \/>\nfor both thecustomer andmyriad colleagues forwhom sheispartially re-<br \/>\nsp onsible. Computer programming istaugh tin courses. Theexcellen tbooks<br \/>\nThe Pragmatic Programmer\t<br \/>\n[8], Co de Complete\t[3], Rapid Developmen t\t[2], and<br \/>\nExtreme Programming Explained\t<br \/>\n[4]all teac hcomputer programming andthe<br \/>\nlarger issuesofbeing ago od programmer. TheessaysofPaul Graham[6 ]and<br \/>\nEric Raymond[7 ]should certainly be read before reading thisarticle. Thisessay<br \/>\ndi ers fromthose excellen tworks by emphasizing social problems andcompre-<br \/>\nhensiv elysummarizing theentire setofnecessary skillsasIsee them.<br \/>\nThis isvery subjectiv e.This essayis therefore doomed tobe personal and<br \/>\nsomewhat opinionated. Icon ne myself toproblems thataprogrammer isvery<br \/>\nlik ely tohave to face inher work. Manyof these problems andtheir solutions<br \/>\nare sogeneral tothe human condition thatIwill probably seempreachy. Ihop e<br \/>\nin spite ofthis that thisbook will be useful. Iha ve attempted tosummarize<br \/>\nas concisely aspossible thosethings thatIwish someone hadexplained tome<br \/>\nwhen Iw as tw en ty-one.<br \/>\nIn this book the term boss torefer towhomev ergiv esyou pro jects todo. I<br \/>\nuse thewords business ,comp any,and tribe, synon ymously exceptthatbusiness<br \/>\nconnotes moneymaking, companyconnotes themodern workplace andtribeis<br \/>\ngenerally thepeople you share loyalt ywith.<br \/>\nW elcome tothe tribe.<br \/>\n3<br \/>\n<br \/>\nPart I<br \/>\nBeginner 2 Personal Skills<br \/>\n2.1 Learn toDebug<br \/>\nDebugging isthe cornerstone ofbeing aprogrammer. The rstmeaning ofthe<br \/>\nw ord toremo ve errors, butthemeaning thatmatters isto seeinto theexecution<br \/>\nof aprogr am byexamining it. A programmer thatcannot debuge ectiv elyis<br \/>\nblind. Idealists maythink thatdesign, oranalysis, orcomplexit ytheory ,or what-<br \/>\nnot, ismore fundamen tal,but they arenot working programmers. Theworking<br \/>\nprogrammer does not livein an ideal world. Even ifshe isperfect, sheissur-<br \/>\nrounded by and must interact withcode written by ma jor soft ware companies,<br \/>\norganizations likeGNU, andhercolleagues. Mostofthis code isimp erfect and<br \/>\nimp erfectly documen ted.Without theabilit yto gain visibilit yinto the execu-<br \/>\ntion ofthis code the sligh test bump willthrowher permanen tly.Often this<br \/>\nvisibilit ycan only be gained by exp erimen tation, thatis,debugging.<br \/>\nDebugging isab out therunning ofprograms, notprograms themselves.If<br \/>\ny ou buy something fromama jor software compan y, you usually don'tgetto<br \/>\nsee the program. Butthere willstill arise places wherethecode does not<br \/>\nconform tothe documen tation (crashing your entire machine isacommon and<br \/>\nsp ectacular example), orwhere thedocumen tation ismute. More commonly ,<br \/>\nthe programmer createsanerror, examines thecode she wrote andhasno<br \/>\nclue how the error canbe occurring. Inevitably,this means someassumption<br \/>\nshe ismaking isnot quite correct, orsome condition arisesthatshedidnot<br \/>\nan ticipate. Sometimes themagic trickof staring into the source code works.<br \/>\nWhen itdo esn't, shemust debug.<br \/>\nT o get visibilit yin to the execution ofaprogram onemust be able toexecute<br \/>\nthe code and observ esomething about it.Sometimes thisisvisible, likewhat<br \/>\nis being displa yed on ascreen, orthe dela ybet ween tw o ev en ts. Inman yother<br \/>\ncases, itinvolv esthings thatarenot mean tto be visible, likethe state ofsome<br \/>\nv ariables insidethecode, orwhic hlines ofcode are actually being executed,<br \/>\nor whether certainassertions holdacross acomplicated datastructure. These<br \/>\nhidden thingsmust be rev ealed.<br \/>\nThe common ways oflooking into the innards ofan executing programcan<br \/>\nb e categorized as:<br \/>\n  using adebugging tool,<br \/>\n  prin tlining |making atemp orary modi cation tothe program, typically<br \/>\nadding linesthatprintinformation out,and<br \/>\n  logging |creating apermanen twindo winto the programs execution in<br \/>\nthe form ofalog.<br \/>\n4<br \/>\n<br \/>\nDebuggingtools are wonderful whentheywork, buttheother tw o metho dsare<br \/>\nev en more importan t.Debugging tools often lagbehind language developmen t,<br \/>\nso atanypoin tin time theymaynot be available. Because thedebugging tool<br \/>\nma ychange theway the program executesinsubtle ways itma ynot be practical<br \/>\nin all cases. Finally,there aresome kinds ofdebugging, suchas chec king an<br \/>\nassertion againstalarge datastructure, thatrequire writing code no matter<br \/>\nwhat. Itisgo od to kno who w to use debugging tools when theyarestable, but<br \/>\nit iscritical tobe able toemplo ythe other tw o metho ds.<br \/>\nSome beginners have a sub conscious fearofdebugging inthe sense ofmo di-<br \/>\nfying andexecuting code. This isunderstandable|it isalittle likeexploratory<br \/>\nsurgery .But beginners have to learn topok eat the code tomak eit jump.<br \/>\nThey have to learn toexp erimen ton it,and that nothing theycandotoitwill<br \/>\nmak eit an yworse. Ifyou are ateac herormen torforthese timid people, please<br \/>\nhelp them getov er this fear by gen tly showing them how to do itand holding<br \/>\ntheir hand ifyou have to. Welose alot ofgood programmers atthe delicate<br \/>\nb eginning tothat fear.<br \/>\n2.2 Howto Debug by Splitting theProblem Space<br \/>\nDebugging isfun, because itbegins withamystery .You think itshould do<br \/>\nsomething, butinstead itdo es something else.Itisnot always quite sosimple|<br \/>\nan yexamples Ican givewill be con triv edcompared towhat sometimes happens<br \/>\nin practice. Debugging requirescreativityand ingen uity. If there isasingle key<br \/>\nto debugging isto use the divide andconquer technique onthe mystery .<br \/>\nSupp oseforexample we created aprogram thatshould doabout 10things<br \/>\nin asequence. Whenwe run itand itcrashes. Wedidn't program itto do<br \/>\nthat, sonow we ha ve a mystery|\\It crashes.\"Wecan seethat itdid the rst<br \/>\n#7 by just looking atthe output. Thelastthree would notbe visible from<br \/>\nthe output, sonow our mystery issmaller \\Itcrashed onthing #8,thing #9,<br \/>\nor thing #10\". Canwe design anexp erimen tto see whic hthing itcrashed<br \/>\non? Sure. Wecan useadebugger orwe can add printline statemen ts(or the<br \/>\nequiv alentin whatev erlanguage you are working in)after #8and #9. Then we<br \/>\nrun itagain, ourmystery willbe smaller, suchas \\It crashed onthing #9.\" I<br \/>\n nd that bearing inmind exactly whatthemystery isat anypoin tin time helps<br \/>\nk eep one focused. Whenseveral people areworking together underpressure on<br \/>\na problem itcan getprett yconfusing.<br \/>\nThe key todivide andconquer asadebugging technique isthe same asitis<br \/>\nfor algorithm design.Aslong asyou doago od job splitting themystery inthe<br \/>\nmiddle, you won't have to split ittooman ytimes, andyou will be debugging<br \/>\nquic kly.But what isthe middle ofam ystery? Thereiswhere truecreativit yand<br \/>\nexp erience comesin.To atrue beginner, thespace ofpossible errorslooks like<br \/>\nthe lines inthe source code. She doesn't have the vision shewill later develop<br \/>\nto see the other dimensions ofthe program, suchas the space ofexecuted lines,<br \/>\nthe data structure, thememory managemen t,the interaction withforeign code,<br \/>\nthe code that isrisky andthecode that issimple. Theseotherdimensions let<br \/>\nan exp erienced programmer formanimp erfect butvery useful mentalmodel of<br \/>\n5<br \/>\n<br \/>\nallthe things thatcangowrong. Having thatmentalmodel iswhat helps one<br \/>\n nd themiddle ofthe mystery e ectiv ely.<br \/>\nOnce you have ev enly subdivided thespace ofall that cangowrong, you<br \/>\nm ust trytodecide inwhic hspace theerror lies.Inthe simple casewhere the<br \/>\nm ystery is\\Is the single linethat makesitcrash executed before, orafter this<br \/>\nline isexecuted?\" you can just observ ewhic hlines getexecuted andyou are<br \/>\ndone. Inother casesyour subdivision ofthe mystery willbe more like\\Either<br \/>\nthere isapoin ter inthat graph thatpoin tsto the wrong node, ormy the<br \/>\nalgorithm thatadds upthe variables inthat graph doesn't work.\" Inthat case<br \/>\ny ou mayha ve to write asmall program tochec kthat thepoin ters inthe graph<br \/>\nare allcorrect inorder todecide whichpart ofthe subdivided mystery canbe<br \/>\neliminated. 2.3 Howto Debug UsingaLog<br \/>\nLogging isthe practice ofwriting asystem sothat itpro duces asequence of<br \/>\ninformativ erecords calledalog. The practice ofprin tlining isjust producing a<br \/>\nsimple andusually temporary log.Generally loggingmeansapermanen tand<br \/>\np ossibly con gurable log,that o ers theseadvantages:<br \/>\n  The logmaypro vide useful information about bugs thatarehard torepro-<br \/>\nduce (suchas those thatoccur inthe production environmen tbut cannot<br \/>\nb e repro duced inthe test environmen t.)<br \/>\n  The logcan provide statistical andperformance information.<br \/>\n  If con gurable, thelogmayallo wgeneral information tobe captured to<br \/>\ndebug asp eci c problem thatwould be di cult toreinstate forevery<br \/>\nproblem ifit were done onatemp orary basis.<br \/>\nThe amoun tof information outputinto the logisalw ays acompromise bet ween<br \/>\ninformation andbrevit y. Too muc hinformation makesthe logexpensiv eand<br \/>\nhard touse, toolittle information anditma ynot contain theinformation you<br \/>\nneed. Making theamoun tof information outputcon gurable isvery useful in<br \/>\nthis regard. Typically ,eac hrecord inthe logwill identify itsposition inthe<br \/>\nsource code, the thread thatexecuted itifthat isan issue, theprecise timeof<br \/>\nexecution, andcommonly anadditional usefulpieceofinformation, suchas the<br \/>\nv alue ofsome variable, theamoun tof free memory ,the num ber ofdata objects,<br \/>\netc. These logstatemen tsare sprinkled throughout thesource code but in<br \/>\nparticular atma jor functionalit ypoin ts, and around riskycode. Eachstatemen t<br \/>\ncan be assigned alev eland willonly output arecord ifthe system iscurren tly<br \/>\ncon gured tooutput thatlevel. One should trytoanticipate whereproblems<br \/>\nmigh toccur anddesign thelogstatemen tsto try toaddress thoseproblems.<br \/>\nThe need tomeasure theperformance ofparticular subsystems isusually easy<br \/>\nto anticipate andquite appropriate foralog asitallo wsthe performance data<br \/>\nto be collected inevery environmen t.<br \/>\n6<br \/>\n<br \/>\nIfyou have a permanen tlog, printlining cannowbe done interms ofthe log<br \/>\nrecords, andsome ofthe debugging statementswill probably be permanen tly<br \/>\nadded tothe logging system.<br \/>\n2.4 Howto Understand Performance Problems<br \/>\nUnderstanding how to learn about theperformance ofarunning systemisun-<br \/>\na v oidable forthe same reason thatdebugging is.Even ifthe code you write<br \/>\nconsiders performance perfectly ,the demands placedupon itwill change, the<br \/>\nhardw areituses willchange, andthesoftware systems itinterfaces withwillbe<br \/>\nhidden andperhaps surprising interms oftheir performance. Howev er, inprac-<br \/>\ntice performance problemsarealittle di eren tand alittle easier thandebugging<br \/>\nin general.<br \/>\nSupp osethat you consider asystem orasubsystem tobe to oslo w. Before<br \/>\ny ou try tomak eit faster, you must build amen talmodel ofwh yit is slo w. To do<br \/>\nthis you have to kno wwhere thetime orother resource isreally being spent. A<br \/>\npro ling tool or ago od log isvery useful forthis. There isafamous dictumthat<br \/>\n90% ofthe time willbe sp entin 10% ofthe code. Iwould addtothat howev er<br \/>\nthe importance input\/output expense (I\/O) toperformance issues.Oftenmost<br \/>\nof the time isspentin I\/O inone way or another. Findingtheexpensiv eI\/O<br \/>\nand theexpensiv e10% ofthe code isago od  rst step.<br \/>\nThere areman ydimensions tothe performance ofacomputer system,and<br \/>\nman yresources consumed. The rst resource tomeasure iswall-clo ck time, the<br \/>\ntotal timethatpasses forthe computation. Howev er, this maynot be the whole<br \/>\npicture. Sometimes somethingthattakesalittle longer butdoesn't burn up<br \/>\nso man ypro cessor seconds willbe m uc hbetter incomputing environmen tyou<br \/>\nactually have to deal with. Similarly ,memory ,net work bandwidth, database<br \/>\nor other serveraccesses mayin the end be far more expensiv ethan processor<br \/>\nseconds. Con tention forshared resources thataresync hronized cancause deadlo ck<br \/>\nand starv ation. Deadlo ck is the inabilit yto pro ceed because ofimprop ersyn-<br \/>\nc hronization orresource demands. Starvation isthe failure toschedule acom-<br \/>\np onen tprop erly.If itcan be an ticipated, itis best tohave a way of measuring<br \/>\nthis contention from thestart ofyour project. Even ifthis contention does not<br \/>\no ccur, itis very helpful tobe able toassert thatwith con dence. Loggingwall-<br \/>\nclo ck time isparticularly valuable because itcan inform about unpredictable<br \/>\ncircumstance thatarise insituations whereotherpro ling isimpractical.<br \/>\n2.5 Howto Fix Performance Problems<br \/>\nMost software projects canbe made tentoone-h undred timesfaster thanthey<br \/>\nare atthe code complete date(theearliest datethatallcode iscompletely<br \/>\nfunctional) withrelativ elylittle e ort. Undertime-to-mark et-pressure,itis<br \/>\nb oth wise ande ectiv eto cho ose asolution thatgetsthejobdone simply and<br \/>\nquic kly,but less e cien tlythan some other solution. Howev er, performance is<br \/>\na part ofusabilit y, and often itmust even tually be considered morecarefully .<br \/>\n7<br \/>\n<br \/>\nThekey toimpro vingtheperformance ofavery complicated systemisto<br \/>\nanalyze itwell enough to nd therough spots. These places wheremostofthe<br \/>\nresources areconsumed arethen directly improved. There isnot muc hsense in<br \/>\noptimizing afunction thataccoun tsfor only 1%ofthe computation time.You<br \/>\nshould doaperformance analysis rstto nd outwhere thetime isreally going,<br \/>\nand decide whattoimpro ve from there. Asarule ofthum byou should think<br \/>\ncarefully before doing anything unlessyou think itisgoing tomak ethe system<br \/>\nor asigni can tpart ofitat least twice asfast. There isusually away to do<br \/>\nthis. Work onanything lessthan doubling theperformance onlyafter you'v e<br \/>\nthough tcarefully about thebigwins. Onereason todo this isto consider the<br \/>\ntest and qualit yassurance e ortthatwillhave to be applied toyour change.<br \/>\nEac hchange brings atest burden withit,soitis muc hbetter tohave a few big<br \/>\nc hanges.<br \/>\nAfter you'v emade atw o-fold improvemen tin something, you need toat<br \/>\nleast rethink andperhaps reanalyze todisco ver the next-most-rough spot inthe<br \/>\nsystem, andattac kthat toget another tw o-fold improvemen t.<br \/>\nOften, therough spots inperformance willbe an example ofcoun ting cows<br \/>\nb y coun ting legsand dividing by four, instead ofcoun ting heads. For example,<br \/>\nI'v emade errors suchas failing topro vide arelational databasesystemwith<br \/>\na prop erindex onacolumn Ilo ok up alot, whic hprobably madeittw en ty<br \/>\ntimes slower at least. Other examples includedoingunnecessary I\/Oininner<br \/>\nlo ops, leaving indebugging statementsthat arenolonger needed, unnecessary<br \/>\nmemory allocation, andinparticular inexpertuse oflibraries andother subsys-<br \/>\ntems thatareoften poorly documen tedwith respecttoperformance. Thiskind<br \/>\nof impro vemen tis sometimes calledlow-hanging fruit,meaning thatitcan be<br \/>\neasily picked topro vide some bene t.<br \/>\nWhat doyou dowhen you start torun outoflow-hanging fruit?Well,<br \/>\ny ou can reachhigher, orchop thetree down. You can continue making small<br \/>\nimpro vemen tsor you can seriously redesignasystem orasubsystem. (Thisisa<br \/>\ngreat opportunit yto use your skills asago od programmer, notonly inthe new<br \/>\ndesign butalso incon vincing your boss that thisisago od idea.) Before you<br \/>\nargue forthe redesign ofasubsystem, you should askyourself ifyou can make<br \/>\nit  v eto ten time better.<br \/>\n2.6 Howto Optimize Loops<br \/>\nSometimes oneencoun terloops, orrecursiv efunctions, thattakea long time<br \/>\nand arerough spots inyour product. You can probably makethe loop alittle<br \/>\nfaster, butspend afew minutes considering ifthere isaway to remo ve it en tirely .<br \/>\nW ould adi eren talgorithm doit? Could you compute thatwhile computing<br \/>\nsomething else?Ifyou can't  ndaw ay around itor you cho ose not to,then<br \/>\ny ou can optimize theloop. This issimple; move stu  out.Inthe end, thiswill<br \/>\nrequire notonly ingen uitybut also anunderstanding ofthe expense ofeac hkind<br \/>\nof statemen tand expression. Buthere aresome suggestions:<br \/>\n  Remo ve <br \/>\roating poin top erations.<br \/>\n8<br \/>\n<br \/>\n Inline subroutine calls.<br \/>\n  Don't allocate newmemory blocks unnecessarily .<br \/>\n  Fold constan tstogether.<br \/>\n  Mo ve I\/O into abu er.<br \/>\n  Try not todivide.<br \/>\n  Try not todo exp ensiv ecasts.<br \/>\n  Mo ve a poin ter rather thanrecomputing indices.<br \/>\n2.7 Howto Deal withI\/OExpense<br \/>\nF or alot ofproblems, processors arefast compared tothe cost ofcomm unicating<br \/>\nwith ahardw aredevice. Thiscostisusually abbreviated I\/O,anditcan include<br \/>\nnet work cost, diskI\/O, database queries, leI\/O, andother things thattend to<br \/>\nmak esome hardw arenot very close tothe processor dosomething. Therefore<br \/>\nbuilding afast system isoften moreaquestion ofimpro vingI\/Othan impro ving<br \/>\nthe code insome tightlo op, oreven impro vinganalgorithm.<br \/>\nThere aretw o very fundamen taltechniques toimpro vingI\/O: caching and<br \/>\nrepresen tation.Caching isav oiding I\/O(generally av oiding thereading ofsome<br \/>\nabstract value) by storing acop yof that value locally sono I\/O isperformed<br \/>\nto get the value. Itbrings withitcac hecoherency problems,butcanbe very<br \/>\ne ectiv eand alot has been written about it.Represen tationisthe approac h<br \/>\nof making I\/Ocheap erby represen tingdata more e cien tly.This isoften in<br \/>\ntension withother demands, likehuman readabilit yand portabilit y.<br \/>\nRepresen tationscanoften be impro ved by afactor oftw o or three from<br \/>\ntheir  rstimplemen tation.Tec hniques fordoing thisinclude usingabinary<br \/>\nrepresen tationinstead ofahuman readable one,transmitting adictionary of<br \/>\nsym bols along withthedata sothat long symbols don't have to be enco ded,<br \/>\nand, atthe extreme, thingslikeHu man encoding. Ithink thisoften represen ts<br \/>\na form oflow-hanging fruit.<br \/>\nA third technique thatissometimes possible isto impro ve the localit yof<br \/>\nreference by pushing thecomputation closertothe data. For instance, ifyou<br \/>\nare reading somedatafrom adatabase andcomputing somethingsimplefrom<br \/>\nit, suc has asummation, trytoget the database serverto do itfor you. This<br \/>\nis highly dependen ton the kind ofsystem you're working with,butingeneral<br \/>\nbuilding aserv erto pro duce computation resultsclosetothe data sothat only<br \/>\nsmall amoun tsof data have to be transferred shouldbe explored moreoften.<br \/>\n2.8 Howto Manage Memory<br \/>\nMemory isaprecious resource thatyou can't a ord torun outof.You can<br \/>\noften ignore itfor awhile buteven tually you will have to decide howto manage<br \/>\nmemory .<br \/>\n9<br \/>\n<br \/>\nSpacethatneeds topersists bey ond thescop eof asingle subroutine isoften<br \/>\ncalled heap allocate d. Dep ending onthe system you use, you mayha ve to<br \/>\nexplicitly deallocatesuchspace whenitis ab out tobecome garbage oryou may<br \/>\nrely ongarbage collector. Ach unk ofmemory isgarbage whennothing refersto<br \/>\nit. Agarbage collector noticesgarbage andfrees itsspace without anyaction<br \/>\nrequired by the programmer. Garbagecollection iswonderful. Itlessens errors<br \/>\nand increases code brevit yand concision cheaply .Use itwhen you can. But<br \/>\nev en with garbage collection, you can  llupallmemory withgarbage. Aclassic<br \/>\nmistak eis to use ahash table asacac he and forget toremo ve the references<br \/>\nin the hash table. Sincethereference remains,thereferen tis uncollectable but<br \/>\nuseless. Thisiscalled amemory leak.You should look for and  xmemory<br \/>\nleaks early.If you have long running systemsmemory maynev erbe exhausted<br \/>\nin testing butwill be exhausted by the user.<br \/>\nThe creation ofnew objects ismo derately expensiv eon anysystem. Memory<br \/>\nallo cated directly inthe local variables ofasubroutine isnot because thepolicy<br \/>\nfor freeing itcan be very simple. You should av oid unnecessary object creation<br \/>\nand unnecessary allocation ofmemory outsideoflocal subroutine parameters.<br \/>\nAn imp ortan tcase occurs when you can de ne anupp erbound onthe num-<br \/>\nb er ofobjects you will need atone time. Ifthese objects alltakeup the same<br \/>\namoun tof memory ,you maybe able toallo cate asingle block of memory ,or<br \/>\na bu er, tohold them all.The objects you need canbe allo cated andreleased<br \/>\ninside thisbu er inaset rotation pattern, soitissometimes calledaring bu er.<br \/>\nThis isusually fasterthanheap allocation.<br \/>\nSometimes you have to explicitly freeallocated space soitcan be reallo cated<br \/>\nrather thanrelyongarbage collection. Thenyou must apply careful intelligence<br \/>\nto eac hch unk ofallo cated memory anddesign away for itto be deallo cated at<br \/>\nthe appropriate time.Themetho dma ydi er foreachkind ofob ject you create.<br \/>\nY ou must mak esure that every execution ofamemory allocating operation is<br \/>\nmatc hedby amemory deallocating operation even tually .This isso di cult that<br \/>\np eople oftensimply implemen ta rudimen taryform orgarbage collection, such<br \/>\nas reference counting, todo this forthem.<br \/>\n2.9 Howto Deal withIntermitten tBugs<br \/>\nThe intermitten tbug isacousin ofthe 50-fo ot-invisible-scorpion-from-outer-<br \/>\nspace kindofbug. Thisnightmare occurs sorarely thatitishard toobserv e,<br \/>\ny et often enough thatitcan't be ignored. You can't debug because you can't<br \/>\n nd it.<br \/>\nThe intermitten tbug hastoobey the same laws oflogic everything elsedoes.<br \/>\nWhat makesithard isthat itoccurs onlyunder unknownconditions. Try to<br \/>\nrecord thecircumstances underwhichit do es occur, sothat you can guess at<br \/>\nwhat thevariabilit yreally is.The condition maybe related todata values,<br \/>\nsuc has \\This onlyhapp enswhen we en ter Wyoming asavalue.\" Ifthat isnot<br \/>\nthe source ofvariabilit y, the next suspectshould be improp erlysynchronized<br \/>\nconcurrency .<br \/>\nT ry ,try ,try torepro duce it.Ifyou can't reproduce it,set atrap forit<br \/>\n10<br \/>\n<br \/>\nby building alogging system, asp ecial oneifyou have to, that canlogwhat<br \/>\ny ou guess iswhat you need when itreally occurs. Ifthe bug only occurs in<br \/>\npro duction, thisisalong process. Eachhin tthat you get from thelogabout<br \/>\nwhat theproblem maybe ma ynot provide thesolution butsuggest theneed for<br \/>\nmore logging. Thenew logging maytak ea long time tobe put into pro duction.<br \/>\nThen you have to wait for the bug tooccur toget more information. Thiscycle<br \/>\ncan goon for some time.<br \/>\n2.10 Howto Learn Design Skills<br \/>\nT o learn howto design software, study theaction ofamen torby being physically<br \/>\npresen twhen theyaredesigning. Thenstudy well-written piecesofsoft ware.<br \/>\nAfter that,you can read some books onthe latest design techniques.<br \/>\nThen you must doityourself. Startwithasmall project. When you are<br \/>\n nally done,consider how the design failedorsucceeded andhow you diverged<br \/>\nfrom your original conception. Themove on tolarger projects, hopefully in<br \/>\nconjunction withother people. Design isamatter ofjudgmen tthat takesyears<br \/>\nto acquire. Asmart programmer canlearn thebasics adequately intw o mon ths<br \/>\nand canimpro ve from there.<br \/>\nIt isnatural andhelpful todev elop your own style, butremem ber that design<br \/>\nis an art, notascience. People whowrite books onthe subject have a vested<br \/>\nin terest inmaking itseem scienti c. Don't become dogmatic about particular<br \/>\ndesign styles.<br \/>\n3 Team Skills<br \/>\n3.1 Whyestimation isimp ortan t<br \/>\nT o get aworking software system inactiv euse asquic klyaspossible requires<br \/>\nplanning thedevelopmen t,but also planning thedocumen tation, deploymen t,<br \/>\nmark eting, sales,and nance. Without predictabilit yof the developmen ttime,<br \/>\nit isimp ossible toplan these e ectiv ely.<br \/>\nGo od estimation provides predictabilit y. Managers love it, aswell they<br \/>\nshould. Thefactthat itis imp ossible, both theoretically andpractically ,to<br \/>\npredict accurately howlong itwill taketo dev elop software ifoften lostonman-<br \/>\nagers. Weare askedtodo this impossible thingallthe time, andwe m ust face<br \/>\nup toithonestly .Ho wev er, itwould be dishonest nottoadmit thisimpossibilit y,<br \/>\nand when necessary ,explain it.There isalot ofroom formiscomm unication<br \/>\nab out estimates, aspeople have a startling tendency tothink wishfully thatthe<br \/>\nsen tence:<br \/>\nIestimate itmigh tbe possible ifIreally understand thatproblem<br \/>\nthat itisabout 50% likely tobe completed in5weeks ifno one<br \/>\nb others usinthat time.<br \/>\nreally means:<br \/>\n11<br \/>\n<br \/>\nIpromise tohave it all done 5weeks from now.<br \/>\nTherefore explicitlydiscusswhattheestimate meanswiththeperson you give<br \/>\nit to, asifthey were asimpleton. Restateyour assumptions, nomatter how<br \/>\nob vious theyseem toyou.<br \/>\n3.2 Howto Estimate Programming Time<br \/>\nEstimation takespractice. Italso takeslab or. Ittak esso muc hlab oritma ybe<br \/>\na go od idea toestimate thetime itwill taketo mak ethe estimate, especially if<br \/>\ny ou are askedtoestimate something you consider stupid.<br \/>\nWhen askedtopro vide anestimate ofsomething big,themost honest thing<br \/>\nto do isto stall. Mostengineers areenthusiastic andeager toplease, andstalling<br \/>\ncertainly willdisplease thestalled. Butanon-the-sp otestimate probably can't<br \/>\nb e accurate andhonest.<br \/>\nWhile stalling, itma ybe possible toconsider doingorprotot yping thetask.<br \/>\nIf political pressure permits, thisisthe most accurate way of pro ducing the<br \/>\nestimate, andmakesreal progress.<br \/>\nWhen notpossible, you should  rstestablish themeaning ofthe estimate<br \/>\nv ery clearly .Restate thatmeaning asthe  rst andlastpart ofyour written<br \/>\nestimate. Prepareawritten estimate by decomp osingthetask into progressiv ely<br \/>\nsmaller subtasks until eac hsmall taskisno more thanada y, and ideally halfa<br \/>\nda yat most inlength. Themost importan tthing isnot toleave an ything out.<br \/>\nF or instance, documen tation, testing, timeforplanning, timeforcomm unicating<br \/>\nwith other groups, andvacation timeareallvery imp ortan t.Ifyou spend part<br \/>\nof eac hda ydealing withknuckleheads, putaline item forthat inthe estimate.<br \/>\nThis givesyour boss visibilit yinto what isusing upyour time ataminim um,<br \/>\nand migh tget you more time.<br \/>\nI kno wgood engineers whopadestimates implicitly,but Irecommend that<br \/>\ny ou donot. Oneofthe results ofthis isto decrease trustsomewhat. For<br \/>\ninstance, anengineer mightestimate 3da ys for atask that heorshe truly<br \/>\nthinks willtakeone day. The engineer mayplan tospend 2da ys documen ting<br \/>\nit, ortw o da ys working onsome other useful project. Butitwill be detectable<br \/>\nthat thetask was done inonly oneday(if itturns outthat way), and the<br \/>\napp earance ofslac king orov erestimating willbe created. It'sfarbetter togiv e<br \/>\nprop ervisibilit yinto what oneisactually doing.Ifdo cumen tation takestwice<br \/>\nas long ascoding andtheestimate says so, tremendous advantage isgained by<br \/>\nmaking thisvisible tothe manager.<br \/>\nP ad explicitly instead.Ifatask willprobably takeone daybut migh ttak e<br \/>\nten days ifyour approac hdo esn't work, note thissomeho win the estimate ifyou<br \/>\ncan; ifnot, atleast doanav erage weigh tedby your estimates ofthe probabilities.<br \/>\nAn yrisk factor thatyou can identify and assign anestimate toshould gointo<br \/>\nthe schedule. Oneperson isunlik elytobe sic kin anygiv en week. Butalarge<br \/>\npro ject with manyengineers willhave some sicktime; likewise vacation time.<br \/>\nAnd what isthe probabilit yof amandatory company-widetraining seminar?<br \/>\nIf itcan be estimated, stickit in. There areofcourse, unknownunkno wns,<br \/>\n12<br \/>\n<br \/>\norunk-unks. Unk-unksby de nition cannotbe estimated individually .You<br \/>\ncan trytocreate aglobal lineitem forallunk-unks, orhandle theminsome<br \/>\nother way that you comm unicate toyour boss. You cannot, howev er, letyour<br \/>\nb oss forget thatthey exist, anditisdeucedly easyforanestimate tobecome a<br \/>\nsc hedule without theunk-unks considered.<br \/>\nIn ateam environmen t,you should trytohave the people whowilldothe<br \/>\nw ork dothe estimate, andyou should trytohave team-wide consensuson<br \/>\nestimates. People vary widely inskill, experience, preparedness, andcon dence.<br \/>\nCalamit ystrik esifastrong programmer estimatesforherself andthen weak<br \/>\nprogrammers areheld tothis estimate. Theactofhaving thewhole teamagree<br \/>\non aline-b y-line basistothe estimate clari estheteam understanding, aswell<br \/>\nas allo wing theopportunit yfor tactical reassignmen tof resources (forinstance,<br \/>\nshifting burdenaw ay from weak erteam members tostronger.)<br \/>\nIf there arebig risks thatcannot be ev aluated, itisyour dutyto state so<br \/>\nforcefully enoughthatyour manager does not commit tothem andthen become<br \/>\nem barassed whentherisk occurs. Hopefully insuc hacase whatev erisneeded<br \/>\nwill be done todecrease therisk.<br \/>\nIf you can convince your compan yto use Extreme Programming, you will<br \/>\nonly have to estimate relativelysmall things, andthis isboth more funand<br \/>\nmore productiv e.<br \/>\n3.3 Howto Find Information<br \/>\nThe nature ofthe what you need tokno wdetermines how you should  ndit.<br \/>\nIf you need information about concrete things\t<br \/>\nthatisobjectiv eand easy<br \/>\nto verify ,for example thelatest patchlev elof asoft ware product, askalarge<br \/>\nn um ber ofpeople politely by searc hing theinternet foritor by posting ona<br \/>\ndiscussion group.Anything thatsmacksofeither opinion orsub jectiv einter-<br \/>\npretation shouldnotbe searc hedforonthe internet, asthe ratio todriv elof<br \/>\ntruth istoolow.<br \/>\nIf you need general knowledge about something subjectiv e\t<br \/>\n,the history of<br \/>\nwhat people have though tab out it,go tothe library (thephysical building in<br \/>\nwhic hbooks arestored.) For example, tolearn about math ormushro omsor<br \/>\nm ysticism, gotothe library .<br \/>\nIf you need tokno whow to do something thatisnot trivial\t<br \/>\n,get tw o or<br \/>\nthree books onthe subject and read them. You migh tlearn how to do some-<br \/>\nthing trivial, likeinstall asoft ware package, from theInternet. You can even<br \/>\nlearn importan tthings likego od programming technique, butyou can easily<br \/>\nsp end more timesearching andsorting theresults andattempting todivine the<br \/>\nauthorit yof the results thanitwould taketo read thepertinen tpart ofasolid<br \/>\nb ook.<br \/>\nIf you need information thatnoone else could be exp ected tokno w\t<br \/>\n,for ex-<br \/>\nample, doesthis software that isbrand newwork ongigan ticdata sets,you must<br \/>\nstill searc hthe internet andthelibrary .After those options arecompleted ex-<br \/>\nhausted, you maydesign anexp erimen tto ascertain it,suc has trying it.<br \/>\n13<br \/>\n<br \/>\nIfyou wan tan opinion oravalue judgmen tthat takesinto accoun tsome<br \/>\nunique circumstance, talktoan exp ert. For instance, ifyou wan tto kno wif<br \/>\nit is ago od idea tobuild amo dern database managemen tsystem inLISP ,you<br \/>\nshould talktoaLISP expert and adatabase expert.<br \/>\nIf you wan tto kno whow lik ely itis\t<br \/>\nthat afaster algorithm foraparticular<br \/>\napplication existsthathasnot yet been published, talktosomeone working in<br \/>\nthat  eld.<br \/>\nIf you wan tto mak ea personal decision thatonly you can mak e\t<br \/>\n,suc has<br \/>\nwhether ornot you should startabusiness, thenconsider theFranklin method<br \/>\nor divination. Supposeyou have studied theidea from allangles, have done<br \/>\nall you homew ork,andwork edout allthe consequences andpros andcons in<br \/>\ny our mind, andyet still remain indecisiv e.The multitude ofavailable divination<br \/>\ntec hniques arevery useful fordetermining your own semi-concious desires,as<br \/>\nthe eachpresen ta complete ambiguous andrandom patternthatyour own<br \/>\nsub concious willassign meaning to.<br \/>\n3.4 Howto Utilize People asInformation Sources<br \/>\nResp ectevery person's time,andbalance itagainst your own. Asking someone<br \/>\na question accomplishes farmore thanjustreceiving theansw er.The person<br \/>\nlearns about you, both by enjo ying your presence andhearing theparticular<br \/>\nquestions. You learn about theperson inthe same way, and you maylearn the<br \/>\nansw eryou seek. Thisisusually farmore importan tthan your question.<br \/>\nHo wev er, the value ofthis diminishes themore you do it. You are, after all,<br \/>\nusing themost precious commodityaperson has,their time. Thebene ts of<br \/>\ncomm unication must be weighed against thecosts. Furthermore, theparticular<br \/>\ncosts andbene ts deriveddi er fromperson toperson. Istrongly believ ethat<br \/>\nan executiv eof one hundred people should spend  vemin utes amon thtalking<br \/>\nto eac hperson inher organization, whichwould be ab out  veper cen tof her<br \/>\ntime. Buttenminutes migh tbe to om uc h, and  vemin utes istoom uc hif they<br \/>\nha ve one thousand employees. Theamoun tof time you spend talking toeac h<br \/>\np erson inyour organization depends ontheir role(more thantheir position).<br \/>\nY ou should talktoyour boss more thanyour boss's boss, butyou should talk<br \/>\nto your boss's boss alittle. Itma ybe uncomfortable, butIbeliev eyou have a<br \/>\ndut yto talk alittle bittoall your superiors eachmon thno matter what.<br \/>\nThe basic ruleisthat every one bene ts fromtalking toyou alittle bit,and<br \/>\nthe more theytalktoyou the less bene t theyderive.Itisyour jobtopro vide<br \/>\nthem thisbene t, andtoget the bene t ofcomm unicating withthem, keeping<br \/>\nthe bene t inbalance withthetime spent.<br \/>\nIt isimp ortan tto resp ectyour own time. Iftalking tosomeone, even ifit<br \/>\nwill cost them time, willsave you agreat dealoftime, thenyou should doit<br \/>\nunless you think theirtimeismore valuable thanyours tothe tribeby that<br \/>\nfactor.<br \/>\nAstrange example ofthis isthe summer intern. Asummer intern ina<br \/>\nhighly technical position can'tbe exp ected toaccomplish toomuc h; they can<br \/>\nb e exp ected topester thehell outofeveryb ody there. Sowhyis this tolerated?<br \/>\n14<br \/>\n<br \/>\nBecausethepestered arereceiving something importan tfrom theintern. They<br \/>\nget achance tosho wo  alittle. Theygetachance tohear some newideas,<br \/>\nma ybe;they getachance tosee things fromadi eren tpersp ectiv e.They may<br \/>\nalso be trying torecruit theintern, buteven ifthat isnot tohapp enthey gain<br \/>\na lot ofbene t.<br \/>\nY ou should askpeople foralittle bitoftheir wisdom andjudgmen twhenev er<br \/>\ny ou honestly believ ethey have something tosay. This <br \/>\ratters themandyou will<br \/>\nlearn something andteachthem something. Agood programmer does not often<br \/>\nneed theadvice ofaVice Presiden tof Sales, butifshe ever does, she should be<br \/>\nsure toask forit.<br \/>\n3.5 Howto Do cumen tWisely<br \/>\nLife istoo short towrite crapnobody will read. Ifyou write crap,nobody<br \/>\nwill read it.Therefore alittle good do cumen tation isbest. Bad documen tation<br \/>\nis very bad. Managers oftendon't understand this,because even bad docu-<br \/>\nmen tation givesthem afalse sense ofsecurit ythat they arenot dependen ton<br \/>\ntheir programmers. Ifsomeone absolutely insiststhatyou write trulyuseless<br \/>\ndo cumen tation, sayyes and quietly begin looking forabetter job.<br \/>\nThere's nothing quitease ectiv eas putting anaccurate estimate ofthe<br \/>\namoun tof time itwill taketo pro duce good do cumen tation into an estimate to<br \/>\nslac ken the demand fordocumen tation. Thetruth iscold andhard: documen-<br \/>\ntation, liketesting, cantakeman ytimes longer thandeveloping code.<br \/>\nW riting good do cumen tation is, rst ofall, good writing. Isuggest you<br \/>\n nd books onwriting, studythem, andpractice. Buteven ifyou are alousy<br \/>\nwriter orhave poor command ofthe language inwhic hyou must documen t,<br \/>\nthe Golden Ruleisall you really need: \\Dounto others asyou would have<br \/>\nthem dounto you.\" Tak etime toreally thinkabout who willbe reading your<br \/>\ndo cumen tation, whattheyneed toget out ofit, and howyou can teachthat to<br \/>\nthem. Ifyou dothat, you will be an abov e av erage documen tation writer, and<br \/>\na go od programmer.<br \/>\nWhen itcomes toactually documen tingcode itself, asopp osed topro ducing<br \/>\ndo cumen tsthat canactually be read by non-programmers, thebest program-<br \/>\nmers I'veev er kno wn hold auniv ersal sentimen t:write self-explanatory code<br \/>\nand don't documen tco de except inthe places thatyou cannot makeit clear.<br \/>\nThere aretw o go od reasons forthis. First, anyone who needs tosee code-lev el<br \/>\ndo cumen tation willinmost cases be able toand prefer toread thecode any-<br \/>\nw ay. Admittedly ,this seems easiertothe experienced programmer thantothe<br \/>\nb eginner. Moreimportan tlyhowev er, isthat thecode and thedocumen tation<br \/>\ncannot getour ofsync ifthere isno documen tation. Thesource code can at<br \/>\nw orst be wrong andconfusing. Thedocumen tation, ifnot written perfectly ,can<br \/>\nlie, and that isathousand timesworse.<br \/>\nThis does not mak eit easier onthe resp onsible programmer. Howdo es one<br \/>\nwrite self-explanatory code? What does that even mean? Itmeans:<br \/>\n  writing code kno wing thatsomeone willhave to read it;<br \/>\n15<br \/>\n<br \/>\n applying thegolden rule;<br \/>\n  using whatev errules ofgood writing you mayha ve learned;<br \/>\n  cho osing asolution thatisstraigh tforward,even ifyou could getby with<br \/>\nanother solution faster;<br \/>\n  sacri cing smalloptimizations thatobfusticate thecode;<br \/>\n  thinking about thereader andspending someofyour precious timeto<br \/>\nmak eit easier onher.<br \/>\n3.6 Howto Work with Poor Co de<br \/>\nIt isvery common tohave to work with poor qualit yco de that someone else<br \/>\nwrote. Don'tthinktoo poorly ofthem, howev er, until you have walk ed in<br \/>\ntheir shoes. They mayha ve been askedvery consciously toget something done<br \/>\nquic klytomeet schedule pressure. Butinorder towork with unclear code you<br \/>\nm ust understand it.To understand ittak eslearning time,andthat time will<br \/>\nha ve to come outofsome schedule, somewhere, andyou must insist onit. To<br \/>\nunderstand it,you will have to read thesource code. You will probably have to<br \/>\nexp erimen twith it.<br \/>\nThis isago od time todocumen t,even ifit is only foryourself, because<br \/>\nthe actoftrying todocumen tthe code will force you toconsider anglesyou<br \/>\nmigh tnot have considered, andtheresulting documen tma ybe useful. While<br \/>\ny ou're doing this,consider whatitwould taketo rewrite someorall ofthe code.<br \/>\nW ould itactually save time torewrite someofit? Could you trust itbetter if<br \/>\ny ou rewrote it?Becareful ofarrogance here.Ifyou rewrite it,itwill be easier<br \/>\nfor you todeal with, butwill itreally be easier forthe next person whohasto<br \/>\nread it?Ifyou rewrite it,what willthetest burden be? Will theneed tore-test<br \/>\nit out weigh anybene ts thatmightbe gained?<br \/>\nIn anyestimate thatyou mak efor work against code you didn't write,the<br \/>\nqualit yof that code should a ectyour perception ofthe risk ofproblems and<br \/>\nunk-unks. Itisimp ortan tto remem ber that abstraction andencapsulation, tw o of the<br \/>\nprogrammers best tools, areparticularly applicabletolousy code. You maynot<br \/>\nb e able toredesign alarge block of code, but ifyou can add acertain amountof<br \/>\nabstraction toityou can obtain someofthe bene ts ofago od design without<br \/>\nrew orking thewhole mess.Inparticular, you can trytowall o  the parts that<br \/>\nare particularly badsothat they maybe redesigned independen tly.<br \/>\n3.7 Howto Use Source Code Con trol<br \/>\nSource code con trol systems letyou manage projects e ectiv ely.They're very<br \/>\nuseful forone person, andessen tialforagroup. Thetrackall changes indi eren t<br \/>\nv ersions sothat nocode isever lost and meaning canbe assigned tochanges. I<br \/>\nw as late toappreciate thebene ts ofthem butnowIw ouldn't livewithout one<br \/>\nev en on aone-p erson project.<br \/>\n16<br \/>\n<br \/>\nAgood tec hnique forusing asource code con trol system isto sta ywithin a<br \/>\nfew days ofbeing uptodate atall time. Code that can't be  nished inafew<br \/>\nda ys ischec ked in, but insuc haway that isinactiv eand willnotbe called, and<br \/>\nhence notcreate anyproblem foranyb ody else, until ithas been unit tested.<br \/>\n3.8 Howto Unit Test<br \/>\nUnit testing, thetesting ofan individual pieceofcoded functionalit yby the<br \/>\nteam thatwrote it,isapart ofcoding, notsomething di erentfrom it.Part<br \/>\nof designing thecode isdesigning how itwill be tested. You should writedown<br \/>\na test plan, even ifit isonly onesentence. Sometimes thetest willbe simple:<br \/>\n\\Do esthe button look good?\" Sometimes itwill be complex: \\Didthismatc hing<br \/>\nalgorithm returnprecisely thecorrect matches?\"<br \/>\nUse assertion chec king andtestdriverswhenev erpossible. Thisnotonly<br \/>\ncatc hesbugs early,but isvery useful lateronand letsyou eliminate mysteries<br \/>\nthat you would otherwise have to worry about.<br \/>\nThe Extreme Programming develop ersare writing extensiv elyonunit testing<br \/>\ne ectiv ely;Ican dono better thantorecommend theirwritings.<br \/>\n3.9 Howto Stress Test<br \/>\nUnlik eunit testing, stresstesting isfun. At  rst itapp ears that thepurp ose<br \/>\nof stress testing isto  nd outifthe system works under aload. Inrealit y, it<br \/>\nis common thatthesystem does work under aload butfails towork insome<br \/>\nw ay when theload ishea vyenough. Icall this hitting thewall orbonking 1<br \/>\n.<br \/>\nThere maybe some exceptions, butthere isalmost always awall. The purp ose<br \/>\nof stress testing isto  gure outwhere thewall is,and then  gure outhow to<br \/>\nmo ve the wall further out.<br \/>\nA plan forstress testing shouldbe dev elop edearly inthe project, because<br \/>\nit often helps toclarify exactly whatisexp ected. Istw o seconds foraweb page<br \/>\nrequest amiserable failureorasmashing success?Is vehundred concurren t<br \/>\nusers enough? Thatofcourse depends, butonemust knowthe answ erwhen<br \/>\ndesigning thesystem thatanswersthe request. Thestress testneeds tomo del<br \/>\nrealit ywell enough tobe useful. Itisn't really possible tosim ulate  vehundred<br \/>\nh uman usingasystem concurren tlyvery easily ,but onecanatleast create  ve<br \/>\nh undred simulations andtrytomo del some partofwhat theymightdo.<br \/>\nIn stress testing, startoutwith aligh tload andload thesystem alongsome<br \/>\ndimension suchas input rateorinput sizeuntil you bonk. Ifthe wall istoo<br \/>\nclose tosatisfy your needs,  gureoutwhic hresource isthe bottlenec k(there is<br \/>\nusually adominan tone.) Isitmemory ,pro cessor, I\/O,network bandwidth, or<br \/>\ndata contention? Then gure outhowyou can move the wall. Note thatmoving<br \/>\nthe wall, that is,increase themaxim umload thesystem canhandle, mightnot<br \/>\nhelp ormigh tactually hurt the performance ofaligh tlyloaded system. Usually,<br \/>\n1<br \/>\nThis term hasseveral meanings, derivedfrom \tohit\", butisin particular usedby athletes<br \/>\nto describ erunning outofblo od sugar orsome other basicresource thatmanifests asasudden<br \/>\nrather thangradual degradation ofperformance orspirt.<br \/>\n17<br \/>\n<br \/>\nperformance underheavyload ismore importan tthan performance underaligh t<br \/>\nload. You mayha ve to get visibilit yin to sev eral di eren tdimensions tobuild upa<br \/>\nmen talmodel ofit; no single technique issu cien t.For instance, loggingoften<br \/>\ngiv esago od idea ofthe wall clock time bet ween tw o ev en ts in the system, but<br \/>\nunless carefully constructed, doesn't givevisibilit yinto memory utilization or<br \/>\nev en data structure size.Similarly ,in amo dern system anum ber ofcomputers<br \/>\nand manysoft ware systems maybe co op erating. Particularly whenyou are<br \/>\nb onking (thatis,the performance isnon-linear inthe size ofthe input) these<br \/>\nother software systems maybe a bottlenec k.Visibilit yinto these systems, even<br \/>\nif only measuring theprocessor loadonallparticipating machines canbe very<br \/>\nhelpful. Kno wing where thewall isessen tialnotonly tomo ving thewall, but also to<br \/>\npro viding predictabilit yso that thebusiness canbe managed e ectively.<br \/>\n3.10 HowTo Recognize WhenTo Break orGo Home<br \/>\nComputer programming isan activit y, but itis also aculture. Theunfortunate<br \/>\nfact isthat itisnot aculture thatvalues mentalorphysical health very muc h.<br \/>\nF or both cultural\/historical reasons(theneed towork atnigh ton unloaded<br \/>\ncomputers, forexample) andbecause ofov erwhelming time-to-mark etpressure<br \/>\nand thescarcit yof programmers, computerprogrammers areov erw orked. I<br \/>\ndon't think you can trust allthe stories you hear, butIthink sixtyhours aweek<br \/>\nis common, and ftyis prett ymuc haminim um.This means thatoften muc h<br \/>\nmore thanthatisrequired. Thisisserious problem forthe good programmer,<br \/>\nwho isresp onsible notonly forthemselv esand their teammates. You have to<br \/>\nrecognize whentogo home, andsometimes whentosuggest thatother people<br \/>\ngo home. Therecan'tbe an y xed rulesforsolving thisproblem, anymore than<br \/>\nthere canbe  xed rulesforraising achild, forthe same reason|ev eryhuman<br \/>\nb eing isdi eren t.<br \/>\nBey ond sixtyhours aweek isextraordinary e ortforme, whic hIcan apply<br \/>\nfor short perio dsoftime (about one week), andthat issometimes expected of<br \/>\nme. Idon't knowifit is fair toexp ectsixt yhours ofwork from aperson; Idon't<br \/>\nev en kno wiffort yis fair. Iam sure howev er that itisstupid towork somuc h<br \/>\nthat you are getting littleoutofthat extra houryou work. For me personally ,<br \/>\nthat's anymore thansixtyhours aweek. Ipersonally thinkaprogrammer<br \/>\nshould exercise noblesse obligeandshoulder ahea vy burden. Howev er, itis<br \/>\nnot aprogrammer's dutyto be patsy .The sadfact isprogrammers areoften<br \/>\nask ed tobe patsies inorder toput onasho wfor someb ody ,for example a<br \/>\nmanager tryingtoimpress anexecutiv e.Programmers oftensuccum bto this<br \/>\nb ecause theyareeager toplease andnotvery good at saying no.There arefour<br \/>\ndefenses againstthis:<br \/>\n  comm unicate asmuc has possible withevery one inthe compan yso that<br \/>\nno one canmislead theexecutiv esabout what isgoing on,<br \/>\n18<br \/>\n<br \/>\n learn toestimate andschedule defensiv elyand explicitly andgiveev ery one<br \/>\nvisibilit yinto what theschedule isand where itstands,<br \/>\n  learn tosayno, and sayno asateam when necessary ,and<br \/>\n  quit ifyou have to.<br \/>\nMost programmers aregood programmers, andgood programmers wan tto get<br \/>\na lot done. To do that, theyhave to manage theirtimee ectiv ely.There<br \/>\nis acertain amountof men talinertia associated withgetting warmed uptoa<br \/>\nproblem anddeeply involv ed init. Man yprogrammers  ndthey work best<br \/>\nwhen theyhave long, uninterrupted blocks oftime inwhic hto get warmed up<br \/>\nand concen trate. Howev er, people must sleep andperform otherduties. Each<br \/>\np erson needs to nd away to satisfy both their human rhythm andtheir work<br \/>\nrh ythm. Eachprogrammer needstodo whatev erittak esto pro cure e cien t<br \/>\nw ork perio ds,suchas reserving certaindays inwhic honly most critical meetings<br \/>\nwill be attended.<br \/>\nSince Iha ve children, Itry tospend evenings withthem sometimes. The<br \/>\nrh ythm thatworks best forme isto work avery long day, sleep inthe o ce or<br \/>\nnear theo ce (Ihave a long comm utefrom home towork) then gohome early<br \/>\nenough thenext dayto spend time withmy children before theygotobed. I<br \/>\nam not comfortable withthis,butitis the best compromise Iha ve been able<br \/>\nto work out. Gohome ifyou have a con tagious disease. You should gohome<br \/>\nif you are thinking suicidalthoughts.You should takea break orgo home if<br \/>\ny ou think homicidal thoughtsfor more thanafew seconds. You should send<br \/>\nsomeone homeifthey showserious mentalmalfunctioning orsigns ofmen tal<br \/>\nillness bey ond mild depression. Ifyou are being tempted tobe dishonest or<br \/>\ndeceptiv ein away that you normally arenot due tofatigue, you should take<br \/>\na break. Don'tusecocaine oramphetamines tocom bat fatigue. Don'tabuse<br \/>\nca eine. 3.11 Howto Deal withDi cult People<br \/>\nY ou will probably have to deal with di cult persons. You mayev en be one.<br \/>\nIf you are the kind ofperson whohasalot ofcon<br \/>\ricts withcowork ers and<br \/>\nauthorit y gures, you should cherish theindep endence thisimplies, butwork<br \/>\non your interp ersonal skillswithout sacri cing your intelligence orprinciples.<br \/>\nThis canbe very disturbing tomost programmers whohave no exp erience in<br \/>\nthis sort ofthing, andwhose previous lifeexperience hastaugh tthem patterns of<br \/>\nb eha vior that arenot useful inthe workplace. Di cultpersons areoften inured<br \/>\nto disagreemen t,and they areless a ected by social pressure tocompromise<br \/>\nthan others. Thekey isto resp ectthem therightamoun t,whic his more than<br \/>\ny ou will wan tto but notasmuc has they mightwan t.<br \/>\nProgrammers have to work together asateam. When disagreemen tarise, it<br \/>\nm ust be resolv edsomeho w,itcannot be duc ked for long. Di cult people often<br \/>\nare extremely intelligen tand have something very useful tosay. It iscritical that<br \/>\none listen andunderstand thedi cult person without prejudice causedby the<br \/>\n19<br \/>\n<br \/>\nperson. Afailure tocomm unicate isoften thebasis ofdisagreemen tand itcan<br \/>\nsometimes be remo ved with great patience. Try tokeep this comm unication<br \/>\nco ol and cordial, anddon't accept anybaits forgreater con<br \/>\rict thatmaybe<br \/>\no ered. Afterareasonable perio dof trying tounderstand, makea decision.<br \/>\nDon't letabully forceyou todo something you don't agreewith. Ifyou are<br \/>\nthe leader dowhat you think isbest. Don't makea decision foranypersonal<br \/>\nreasons, andbe prepared toexplain thereasons foryour decision. Ifyou are a<br \/>\nteammate withadi cult person don'tletthe leaders decision have an ypersonal<br \/>\nimpact. Ifitdo esn't goyour way, do itthe other way whole-heartedly .<br \/>\nDi cult people dochange andimpro ve. I'veseen itwith my own eyes, but<br \/>\nit isvery rare. Howev er, every one hastransitory upsand downs.<br \/>\nOne ofthe challenges thatevery programmer butespecially leaders faceis<br \/>\nk eeping thedi cult person fullyengaged. Theyaremore prone toduc kwork<br \/>\nand resist passiv elythan others.<br \/>\nP art II<br \/>\nIn termediate<br \/>\n4 Personal Skills<br \/>\n4.1 Howto Sta yMotiv ated<br \/>\nIt isawonderful andsurprising factthat programmers arehighly motivatedby<br \/>\nthe desire tocreate artifacts thatarebeautiful, useful,ornift y. This desire is<br \/>\nnot unique toprogrammers noruniv ersal butitis so strong andcommon among<br \/>\nprogrammers thatitseparates thethem fromothers inmost companies.<br \/>\nThis haspractical andimportan tconsequences. Ifprogrammers areasked<br \/>\nto do something thatisnot beautiful, useful,ornift y, they willhave lo w morale.<br \/>\nThere's alot ofmoney tobe made doing ugly,stupid, andboring stu ;butin<br \/>\nthe end funwill makethe most money forthe compan y.<br \/>\n4.2 Howto Tradeo  Timeversus Space<br \/>\nY ou can be a go od programmer withoutgoingtocollege, butyou can't be a<br \/>\ngo od intermediate programmer withoutknowing basic computational complex-<br \/>\nit y theory .You migh tbe able tointuit how to tradeo  timeagainst space<br \/>\nwithout it,but you will nothave a  rm basis forcomm unicating withyour<br \/>\ncolleagues withoutit.<br \/>\nTime (processor cycles) andspace (memory) canbe traded o against each<br \/>\nother. Engineering isab out compromise, andthis isa ne example. Itisnot<br \/>\nalw ays systematic. Ingeneral, howev er, one cansave space by enco ding things<br \/>\nmore tightly,at the expense ofmore computation timewhen you have to deco de<br \/>\nthem. Onecansave time by cac hing, thatis,spending spacetostore alocal<br \/>\ncop yof something, atthe expense ofhaving tomain taintheconsistency ofthe<br \/>\n20<br \/>\n<br \/>\ncache. One cansometimes save time by main taining moreinformation inadata<br \/>\nstructure. Thisusually costasmall amoun tof space butmaycomplicate the<br \/>\nalgorithm alot.<br \/>\nImpro vingthespace\/time tradeo canoften change oneorthe other dra-<br \/>\nmatically .Ho wev er, before you work onthis you should askyourself ifwhat<br \/>\ny ou are impro vingisreally thething thatmost needs tobe impro ved inyour<br \/>\nsystem. It'sfuntowork onanalgorithm, butyou can't letthat blind you tothe<br \/>\ncold hard factthat impro vingsomething thatisnot aproblem willnotmake<br \/>\nan ynoticeable di erence, andwillcreate atest burden.<br \/>\nMemory onmo dern computers appears cheap, because unlikepro cessor time<br \/>\ny ou can't seeitbeing used until you bonk. Butwhen you bonk, you bonk hard.<br \/>\nThere arealso other hidden coststousing memory ,suc has your e ect onother<br \/>\nprograms thatmust be residen t,and thetime toallo cate anddeallo cateit.<br \/>\nConsider thiscarefully before you trade aw ay space togain speed.<br \/>\n4.3 Howto Balance Brevityand Abstraction<br \/>\nAbstraction iskey toprogramming. Oneshould cho ose howabstract oneneeds<br \/>\nto be carefully .Beginning programmers intheir enthusiasm oftencreate more<br \/>\nabstraction thanisreally useful. Onesignofthis isifyou create classes that<br \/>\ndon't really contain anyco de and don't really doanything exceptserveto ab-<br \/>\nstract something. Theattraction ofthis isunderstandable butthevalue ofcode<br \/>\nbrevit ymust be measured againstthevalue ofabstraction. Occasionally one<br \/>\nsees agen tlemistak emade by en thusiastic idealists: atthe start ofthe project<br \/>\na lot classes arede ned thatarewonderfully abstractandshould cover every<br \/>\nev en tualit y. As the project progresses andfatigue setsin,the code itself be-<br \/>\ncomes messy.Function bodies become longerthantheyshould be. The empt y<br \/>\nclasses areaburden todocumen tthat isignored whenunder pressure. The<br \/>\n nal result would have been better ifthe energy spenton abstraction hadbeen<br \/>\nsp enton keeping thingsshortandsimple. Istrongly recommend thearticle<br \/>\n\\Succinctness isPow er\" by Paul Graham[6 ].<br \/>\nThere isacertain dogmaassociated withuseful techniques suchas infor-<br \/>\nmation hidingandobject orien tedprogramming thataresometimes takentoo<br \/>\nfar. These techniques letone code abstractly andanticipate change. Iper-<br \/>\nsonally think,howev er, that oneshould notproduce muc h speculativ eco de.<br \/>\nF or example, itisan accepted style tohide aninteger variable onanobject<br \/>\nb ehind mutators andaccessors, sothat thevariable itselfisnot exposed, only<br \/>\nthe little interface toit. This does allo wthe implemen tationofthat variable<br \/>\nto be changed without a ecting thecalling code, and isperhaps appropriate<br \/>\nto alibrary writerwhomust publish avery stable API.ButIdon't think the<br \/>\nb ene t ofthis outweighs thecost ofthe wordiness ofitwhen my team isowns<br \/>\nthe calling code and hence canrecodethe caller aseasily asthe called. Four or<br \/>\n v eextra linesofcode isahea vyprice topayfor this speculativ ebene t.<br \/>\nP ortabilit yposes asimilar problem. Shouldcode be portable toadi eren t<br \/>\ncomputer, compiler,software system orwhatev er,orsimply easilyported? I<br \/>\nthink anon-p ortable, short-and-easily-p ortedpieceofcode isbetter thanalong<br \/>\n21<br \/>\n<br \/>\nportable one.Itisrelativ elyeasy andcertainly ago od idea tocon ne non-<br \/>\np ortable code todesignated areas,suchas aclass thatmakesdatabase queries<br \/>\nthat arespeci c toagiv en DBMS.<br \/>\n4.4 Howto Learn NewSkills<br \/>\nLearning newskills, especially non-technicalones,isthe greatest funofall.<br \/>\nMost companies would have better morale ifthey understo od ho w muc hthis<br \/>\nmotiv ateprogrammers.<br \/>\nHumans learnby doing. Book-reading andclass-taking areuseful. Butcould<br \/>\ny ou have an yresp ectforaprogrammer whohadneverwritten aprogram? To<br \/>\nlearn anyskill, you have to put yourself inaforgiving position whereyou can<br \/>\nexercise thatskill. When learning anew programming language,trytodo a<br \/>\nsmall project itin before you have to do alarge project. When learning to<br \/>\nmanage asoft ware project, trytomanage asmall one rst.<br \/>\nA good men torisno replacemen tfor doing things yourself, butisalot<br \/>\nb etter thanabook. What canyou o er apoten tial mentorinexc hange for<br \/>\ntheir knowledge? At a minim um,you should o ertostudy hardsotheir time<br \/>\nw on't be wasted.<br \/>\nT ry toget your boss tolet you have formal training, butunderstand that<br \/>\nit often nobetter thanthesame amoun tof time spentsimply playing with<br \/>\nthe new skill you wan tto learn. Itishowev er, easier toask fortraining than<br \/>\npla ytime inour imperfect world, even though alot offormal training isjust<br \/>\nsleeping throughlectureswaiting forthe dinner party.<br \/>\nIf you lead people, understand how they learn andhelp them learnby as-<br \/>\nsigning themprojects thatarethe rightsize and that exercise skillstheywan t<br \/>\nto learn. Don'tforget thatthemost importan tskills foraprogrammer arenot<br \/>\nthe technical ones.Giveyour people achance toplayand practice courage,<br \/>\nhonest y, and comm unication.<br \/>\n4.5 Howto do Integration Testing<br \/>\nIn tegration testingisthe testing ofthe integration ofvarious componentsthat<br \/>\nha ve been unit tested. Integration isexp ensiv eand itcomes outinthe testing.<br \/>\nY ou must include timeforthis inyour estimates andyour schedule.<br \/>\nIdeally you should organize apro ject sothat there isnot aphase atthe<br \/>\nend where things have to be integrated. Itisfar better togradually integrate<br \/>\nthings asthey arecompleted ov er the course ofthe project. Ifitisuna voidable<br \/>\nestimate itcarefully .<br \/>\n5 Team Skills<br \/>\n5.1 Howto Manage Developmen tTime<br \/>\nT o manage developmen ttime, maintainaconcise andup-to-date project plan. A<br \/>\npro ject plan isan estimate, asc hedule, aset ofmilestones formarking progress,<br \/>\n22<br \/>\n<br \/>\nanassignmen tof your team oryour own time toeac htask onthe estimate. It<br \/>\nshould alsoinclude otherthings you have to remem ber to do, suchas meeting<br \/>\nwith thequalit yassurance people, preparing documen tation, orordering equip-<br \/>\nmen t.Ifyou are onateam, theproject plan should be a consensual agreement,<br \/>\nb oth atthe start andasyou go.<br \/>\nThe project plan exists tohelp makedecisions, nottosho whow organized<br \/>\ny ou are. Ifthe project plan iseither toolong ornot up-to-date itwill be useless<br \/>\nfor making decisions. Inrealit y, these decisions areabout individual persons.<br \/>\nThe plan andyour judgmen tlet you decide ifyou should shifttasks fromone<br \/>\np erson toanother. Themilestones markyour progress. Ifyou use afancy<br \/>\npro ject planning tool, donot be seduced into creating abig design upfron tfor<br \/>\nthe project, butuseitmain tainconcision andup-to-dateness.<br \/>\nIf you miss amilestone, you should takeimmediate actionsuchas informing<br \/>\ny our boss that thescheduled completion ofthat project hasslipp edby that<br \/>\namoun t.The estimate andschedule couldneverhave been perfect tobegin<br \/>\nwith; thiscreates theillusion thatyou migh tbe able tomak eup the days you<br \/>\nmissed inthe latter partofthe project. You migh t.But itisjust aslikely that<br \/>\ny ou have underestimated thatpartasthat you have ov erestimated it.Therefore<br \/>\nthe scheduled completion ofthe project hasalready slipped,whether you likeit<br \/>\nor not.<br \/>\nMak esure you plan includes timefor:<br \/>\n  internal teammeetings,<br \/>\n  demos,<br \/>\n  do cumen tation,<br \/>\n  scheduled perio dicactivities,<br \/>\n  integration testing,<br \/>\n  dealing withoutsiders,<br \/>\n  sic kness,<br \/>\n  vacations,<br \/>\n  main tenance ofexisting products, and<br \/>\n  main tenance ofthe developmen ten vironmen t.<br \/>\nThe project plan canserveas away to giv eoutsiders oryour boss aview into<br \/>\nwhat you oryour team aredoing. For this reason itshould be short andup-to-<br \/>\ndate.<br \/>\n23<br \/>\n<br \/>\n5.2Howto Manage ThirdPart ySoft ware Risks<br \/>\nA pro ject often depends onsoft ware produced by organizations thatitdo es not<br \/>\ncon trol. There aregreat risksassociated withthird partysoft ware that must be<br \/>\nrecognized by ev ery one involv ed.<br \/>\nNev er,ever, rest anyhop eson vapor. Vap or isanyalleged software that<br \/>\nhas been promised butisnot yet available. Thisisthe surest way to go out of<br \/>\nbusiness. Itisun wise tobe sk eptical ofasoft ware compan y'spromise torelease<br \/>\na certain product withacertain feature atacertain date;itis far wiser toignore<br \/>\nit completely andforget you ever heard it.Nev erlet itbe written down inany<br \/>\ndo cumen tsused by your compan y.<br \/>\nIf third partysoft ware isnot vap or, itis still risky ,but atleast itis arisk<br \/>\nthat canbe tac kled. Ifyou are considering usingthirdpartysoft ware you should<br \/>\ndev ote energy earlytoevaluating it.People mightnot liketo hear thatitwill<br \/>\ntak etw o weeks ortw o mon thstoevaluate eachof three products forsuitabilit y.<br \/>\nBut ithas tobe done asearly aspossible. Thecost ofintegrating cannotbe<br \/>\naccurately estimatedwithoutaprop erevaluation.<br \/>\nUnderstanding thesuitabilit yof existing thirdpartysoft ware foraparticular<br \/>\npurp oseisvery tribal knowledge. Itisvery subjectiv eand generally residesin<br \/>\nexp erts. You can save a lot oftime ifyou can  nd those experts. Often a<br \/>\npro ject willdepend onathird partysoft ware system socompletely thatifthe<br \/>\nin tegration failstheproject willfail. Express riskslikethat clearly inwriting<br \/>\nin the schedule. Try tohave a con tingency plan,suchas another systemthat<br \/>\ncan be used orthe abilit yto write thefunctionalit yyourself ifthe risk can't be<br \/>\nremo ved early .Nev erlet aschedule depend onvap or.<br \/>\n5.3 Howto Manage Consultan ts<br \/>\nUse consultan ts,but don't relyonthem. Theyarewonderful people anddeserv e<br \/>\na great dealofresp ect. They oftenknowmore about speci c technologies and<br \/>\nev en programming techniques thanprogrammers thatwork asemplo yees ofa<br \/>\ncompan y. The best way to use them isas educators in-housethatcanteachby<br \/>\nexample.<br \/>\nHowev er, they cannot usually become partofthe team inthe same sense<br \/>\nthat regular employees are, ifonly because you maynot have enough time<br \/>\nto learn theirstrengths andweaknesses. Their nancial commitmen tis muc h<br \/>\nlo wer. They canmove more easily.They mayha ve less togain ifthe compan y<br \/>\ndo es well. Some willbe go od, some willbe av erage andsome willbe bad, but<br \/>\nusually your selection ofconsultan tswill notbe as careful asyour selection of<br \/>\nemplo yees, soyou will getmore badones.<br \/>\nIf they're goingtowrite code you must review itcarefully asyou goalong.<br \/>\nY ou cannot gettothe end ofthe apro ject with therisk ofalarge block of code<br \/>\nthat hasnot been review ed.This istrue ofall team members, really ,but you<br \/>\nwill usually have more knowledge ofyour closer teammembers.<br \/>\n24<br \/>\n<br \/>\n5.4Howto Comm unicate theRigh tAmoun t<br \/>\nConsider carefullythecost ofameeting. Itcosts itsduration multiplied by<br \/>\nthe num ber ofparticipan ts.Meetings aresometimes necessary,but smaller is<br \/>\nusually better. Thequalit yof comm unication insmall meetings isbetter, and<br \/>\nless time ov erall iswasted. Ifyou are having ameeting andanyone isbored,<br \/>\nthat should be a sign that perhaps you should have organized smallermeetings.<br \/>\nEv erything possible should be done toencourage informalcommunication.<br \/>\nMore useful work done during luncheswith colleagues thanduring anyother<br \/>\ntime. Itisashame thatmore companies tonot recognize thisfact andsupp ort<br \/>\nit. 5.5 Howto Disagree Honestly andGet Away with It<br \/>\nDisagreemen tis agreat opportunit yto mak ea go od decision, butitshould be<br \/>\nhandled delicately .Hop efully you feel that you have expressed your though ts<br \/>\nadequately andbeen listened tobefore thedecision ismade. Inthat casethere<br \/>\nis nothing moretosay, and you should decidewhether you will stand behind<br \/>\nthe decision even though tyou disagree withit.Ifyou can supp ortthis decision<br \/>\nev en though you disagree, sayso. This showshowvaluable you are because you<br \/>\nare indep enden tand notayes-man, butrespectful ofthe decision andateam<br \/>\npla yer.<br \/>\nSometimes adecision willbe made thatyou disagree withwhen thedecision<br \/>\nmak ersdid not have the full bene t ofyou opinion. You should thenevaluate<br \/>\nwhether you should raisetheissue onthe basis ofthe bene t tothe compan yor<br \/>\ntrib e.Ifitisasmall mistak ein your opinion, itma ynot be worth reconsidering.<br \/>\nIf itis alarge mistak ein you opinion, thenofcourse you must presen tan<br \/>\nargumen t.<br \/>\nUsually ,this isnot aproblem. Insome stressful circumstances andwith<br \/>\nsome personalit ytyp es this canlead tothings being taken personally .For<br \/>\ninstance, somevery good programmers lackthe con dence neededtochallenge<br \/>\na decision even when theyhave go od reason tobeliev eit to be wrong. Inthe<br \/>\nw orst ofcircumstances thedecision makerisinsecure andtakesitas apersonal<br \/>\nc hallenge toher authorit y. It isbest toremem ber insuc hcircumstances that<br \/>\np eople reactwiththereptilian partoftheir brains. You should presentyour<br \/>\nargumen tin priv ate, and trytosho whow new knowledge changes thebasis on<br \/>\nwhic hthe decision was made.<br \/>\nWhether thedecision isrev ersed ornot, you must remem ber that probably<br \/>\nno one will ever have a righ tto say\"I told you so!\" since theother decision will<br \/>\nnot have been fully explored.<br \/>\n25<br \/>\n<br \/>\n6Judgmen t<br \/>\n6.1 Howto Tradeo  QualityAgainst Developmen tTime<br \/>\nSoft ware developmen tis alw ays acompromise bet ween what theproject does<br \/>\nand getting theproject done. Butyou maybe ask ed tomak ea tradeo  of<br \/>\nqualit yto speed thedeplo ymen tof apro ject that iso ends your engineering<br \/>\nsensibilities inthe same way that making atoaster thatisdesigned towear out<br \/>\nis o ensiv e.For example, you maybe ask ed todo something thatisapoor<br \/>\nsoft ware engineering practiceandwilllead toalot ofmain tenance problems.<br \/>\nIf this happ ensyour  rst responsibilit yis to inform your team andtoclearly<br \/>\nexplain thecost ofthe decrease inqualit y. After all,your understanding ofit<br \/>\nshould be muc hbetter thanyour boss's understanding. Makeit clear what is<br \/>\nb eing lostand what isbeing gained, andhow at what costthelost ground will<br \/>\nb e regained inthe next cycle. Inthis thevisibilit ypro vided by ago od pro ject<br \/>\nplan should be helpful. Ifthe qualit ytradeo  a ectsthequalit yassurance e ort<br \/>\np oin tthat out(both toyour boss and qualit yassurance people.) Ifthe qualit y<br \/>\ntradeo  willlead tomore bugsbeing reported afterthequalit yassurance perio d,<br \/>\np oin tthat out.<br \/>\nIf she still insists you should trytoisolate theshoddiness into particular<br \/>\ncomp onentsthat you can plan torewrite orimpro ve in the next cycle. Explain<br \/>\nthis toyour team sothat they canplan forit.<br \/>\n6.2 Howto Manage Software System Dependence<br \/>\nMo dern software systems tendtodep end onalarge num ber ofcomp onentsthat<br \/>\nare not directly underyour control. Thisincreases productivit ythrough synergy<br \/>\nand reuse. Howev er, eachcomp onentbrings withitsome problems:<br \/>\n  Ho wwill you  x bugs inthe comp onent?<br \/>\n  Do esthe comp onentrestrict you toparticular hardwareorsoft ware sys-<br \/>\ntems?<br \/>\n  What willyou doifthe comp onentfails completely?<br \/>\nIt isalw ays best toencapsulate thecomp onentin some way so that itisiso-<br \/>\nlated andsothat itcan be sw app edout. Ifthe comp onentpro ves completely<br \/>\nun work able, you maybe able toget adi eren tone, butyou mayha ve to write<br \/>\ny our own. Encapsulation isnot portabilit y, but itmak esporting easierandthat<br \/>\nis almost asgood.<br \/>\nHa ving thesource code for acomp onentdecreases therisk by afactor of<br \/>\nfour. Withit,you can evaluate iteasier, debugiteasier,  ndwork arounds<br \/>\neasier, andmake xes easier. Ifyou mak e xes, you should givethem tothe<br \/>\no wner ofthe comp onentand trytoget them toincorp oratethem into an o cial<br \/>\nrelease; otherwise you will have to main tainanuno cial version uncomfortably .<br \/>\n26<br \/>\n<br \/>\n6.3Howto Decide IfSoft ware IsToo Immature<br \/>\nThe useofsoft ware other people wrote isone ofthe most e ectiv eways tobuild<br \/>\nquic klyasolid system. Itshould notbe discouraged, buttherisks associated<br \/>\nwith itmust be examined. Oneofthe biggest risksisthe perio dof bugginess<br \/>\nand near inoperabilit ythat isoften associated withsoftware before itmatures<br \/>\nthrough usageinto ausable product. Whenconsidering integrating withor<br \/>\nb ecoming dependen tin some way on asoft ware system, whether createdin<br \/>\nhouse orby athird party, it is very importan tto consider ifit isreally mature<br \/>\nenough tobe used. Herearesome questions you should askyourself about it:<br \/>\n  Is itvap or? (Promises arevery immature).<br \/>\n  Is there anaccessible body oflore about thesoftware?<br \/>\n  Are you the  rst user?<br \/>\n  Is there astrong incentivefor continuation?<br \/>\n  Has ithad amain tenance e ort?<br \/>\n  Will itsurviv edefection ofthe curren tmain tainers?<br \/>\n  Is there aseasoned alternativ eat least halfasgood?<br \/>\n  Is itkno wn toyour tribeor compan y?<br \/>\n  Is itdesirable toyour tribeor compan y?<br \/>\n  Can you hire people towork oniteven ifit is bad?<br \/>\n6.4 Howto Mak ea Buy vs.Build Decision<br \/>\nAn entrepreneurial companythat istrying toaccomplish somethingwithsoft-<br \/>\nw are has tomak ebuy versus build decisions constantly.This requires agreat<br \/>\ncom bination ofbusiness, managemen t,and engineering savvy .It should perhaps<br \/>\nb e called abuy andintegrate vs.build andintegrate decision because thecost<br \/>\nof integration must be considered.<br \/>\n  Ho wwell do your needs matchthose forwhic hit was designed?<br \/>\n  What portion ofwhat you buy willyou need?<br \/>\n  What isthe cost ofevaluating theintegration?<br \/>\n  What isthe cost ofintegration?<br \/>\n  Will buying increase ordecrease longterm maintenance costs?<br \/>\n  Will building putyou inabusiness position you don't wan tto be in?<br \/>\n27<br \/>\n<br \/>\nYou should thinktwice before building something thatisbig enough toserv eas<br \/>\nthe basis foranentire other business. Suchideas areoften proposed by brigh t<br \/>\nand optimistic persons thatwillhave a lot tocon tribute toyour team. Iftheir<br \/>\nidea iscomp elling, you ma ywish tochange your business plan;butdonot invest<br \/>\nin asolution biggerthanyour own business withoutconcious thought.<br \/>\nAfter considering thesequestions, you should perhaps prepare tw o draft<br \/>\npro ject plans, oneforbuilding andoneforbuying. Thiswillforce you toconsider<br \/>\nthe integration costs.You should alsoconsider thelong term maintenance costs<br \/>\nof both solutions. To estimate theintegration costs,you will have to be able<br \/>\nto do athorough evaluation ofthe software before you buy it.Ifyou can't<br \/>\nev aluate ityou will assume anunreasonable riskinbuying itand you should<br \/>\ndecide against buyingthatparticular product. Ifthere areseveral buy decisions<br \/>\nunder consideration, someenergy willhave to be sp entev aluating each.<br \/>\n6.5 Howto Gro wProfessionally<br \/>\nAssume responsibilit yin excess ofyour authorit y. Pla ythe role that you de-<br \/>\nsire. Express appreciation forpeople's contribution tothe success ofthe larger<br \/>\norganization, aswell asthings asthat help you personally .<br \/>\nIf you wan tto become ateam leader, instigate theformation ofconsensus.<br \/>\nIf you wan tto become amanager, takeresp onsibilit yfor the schedule. You can<br \/>\nusually dothis comfortably whileworking withaleader oramanager, sincethis<br \/>\nfrees heruptotak egreater responsibilit y. If that istoomuc hto try,do ita<br \/>\nlittle atatime.<br \/>\nEv aluate yourself. Ifyou wan tto become abetter programmer, asksomeone<br \/>\ny ou admire how you can become likeher. You can also askyour boss, who will<br \/>\nkno wless buthave a greater impactonyour career.<br \/>\nPlan ways tolearn newskills, butthetrivial technical kindlikelearning a<br \/>\nnew software system andthehard social kind likewriting well, by integrating<br \/>\nthem into your work.<br \/>\n6.6 Howto Evaluate Interview ees<br \/>\nEv aluating poten tialemplo yees isnot given the energy itdeserv es.Abad hire,<br \/>\nlik ea bad marriage, isterrible. Asigni can tportion ofevery one's energy should<br \/>\nb e dev oted torecruitmen t,but this israrely done.<br \/>\nThere aredi eren tin terviewing styles. Some aretorturous, designedtoput<br \/>\nthe candidate underagreat dealofstress. Thisservesavery valuable purpose<br \/>\nof possibly revealing character <br \/>\raws and weaknesses understress. Candidates<br \/>\nare nomore honest withinterview eesthan theyarewith themselv es,and the<br \/>\nh uman capacit yfor self-deception isastonishing.<br \/>\nY ou should ataminim umgivethe candidate theequiv alentof an oral ex-<br \/>\namination onthe technical skillsfortw o hours. Withpractice, you will be<br \/>\nable toquic klycover what theyknowand quicklyretract fromwhat theydon't<br \/>\nkno wto mark outtheboundary .In terview eeswill respectthis. Iha ve sev eral<br \/>\ntimes heard withmyown ears them saythat thewas one oftheir motiv ations for<br \/>\n28<br \/>\n<br \/>\ncho osing acompan y. Go od ones wan tto be hired fortheir skills, notwhere they<br \/>\nw ork edlast orwhat scho olthey wen tto orsome other inessen tialcharacteristic.<br \/>\nIn doing this,you should alsoevaluate theirabilityto learn, whichis far<br \/>\nmore importan tthan what theyknow. You should alsowatc hfor the whi <br \/>\nof brimstone thatisgiv en o  by di cult people. You can often recognize it<br \/>\nlater, butinthe heat ofthe interview itishard torecognize. Howwell people<br \/>\ncomm unicate andwork with people ismore importan tthan being uponthe<br \/>\nlatest programming language.<br \/>\nFinally ,in terviewing isalso apro cess ofselling. You should be trying tosell<br \/>\ny our compan yto the candidate. Howev er, you are talking toaprogrammer, so<br \/>\nnormal salesmanship does not work. Don't trytocolor thetruth. Starto with<br \/>\nthe bad stu , then nish strong withthegood stu .<br \/>\n6.7 Howto Kno wWhen toApply Fancy Computer Sci-<br \/>\nence<br \/>\nThere isabody ofkno wledge about algorithms, datastructures, mathematics,<br \/>\nand other gee-whiz stu thatmost programmers knowabout but rarely use.<br \/>\nIn practice, thiswonderful stu istoocomplicated andgenerally unnecessary .<br \/>\nThere isno poin tin impro vinganalgorithm whenmostofyour time isspent<br \/>\nmaking ine cien tdatabase calls,forinstance. Anunfortunate amountof pro-<br \/>\ngramming consistsofgetting systems totalk toeac hother andusing very simple<br \/>\ndata structures tobuild anice user interface.<br \/>\nWhen ishigh technology theappropriate technology? Whenshould one<br \/>\ncrac kabook toget something otherthanarun-of-the-mill algorithm?Itis<br \/>\nsometimes usefultodo this butitshould be ev aluated carefully .<br \/>\nThe three mostimport considerations forthe poten tial computer science<br \/>\ntec hnique are:<br \/>\n  Is itwell encapsulated sothat therisk toother systems islow and the<br \/>\no v erall increase incomplexit yand main tenance costislow?<br \/>\n  Is the bene t startling (forexample, afactor oftw o in amature system<br \/>\nor afactor often inanew system)?<br \/>\n  Will you be able totest andevaluate ite ectiv ely?<br \/>\nIf awell-isolated algorithmthatusesasligh tlyfancy algorithm candecrease<br \/>\nhardw arecost orincrease performance by afactor oftw o across anentire sys-<br \/>\ntem, thenitwould be criminal nottoconsider it.One ofthe keys toarguing<br \/>\nfor suchan approac his to sho wthat therisk isreally quitelow, since thepro-<br \/>\np osed technology hasprobably been well studied, theonly issue isthe risk of<br \/>\nin tegration. Hereaprogrammer's experience andjudgmen tcan truly synergize<br \/>\nwith thefancy technology tomak ein tegration easy.<br \/>\nBy carefully designing theencapsulating sothat therisk islow, you should<br \/>\nb e able topro duce anestimate withcon dence thatallowsthe costs andbene ts<br \/>\nof the prop osalbe prop erlyjudged.<br \/>\n29<br \/>\n<br \/>\n6.8Howto Talk toNon-Engineers<br \/>\nEngineers andprogrammers inparticular arerecognized by popular culture as<br \/>\nb eing di eren tfrom other people. Thisimplies thatother people aredi eren t<br \/>\nfrom us.This isworth bearing inmind when communicating withthem. One<br \/>\nshould always understand theaudience.<br \/>\nNon-engineers aresmart, butnotasgrounded increating technical things as<br \/>\nw e are. Wemak ethings. Theysellthings andhandle thingsandcoun tthings<br \/>\nand manage things,butthey arenot experts onmaking things.<br \/>\nThey arenot asgood at working together onteams asengineers are(there<br \/>\nare not doubt exceptions.) Theirsocial skills aregenerally asgood as orbet-<br \/>\nter than engineers innon-team environmen ts,but their work does not always<br \/>\ndemand thatthey practice thekind ofintimate, precisecommunication and<br \/>\ncareful subdivisions oftasks thatwe do. Their teams aremore likegroups.<br \/>\nNon-engineers maybe to oeager toplease andthey maybe in timidated by<br \/>\ny ou. Just likeus, they maysayyes without reallymeaning itto please you or<br \/>\nb ecause theyarealittle scared ofyou, and then notstand behind theirwords.<br \/>\nNon-programmers canunderstand technical things butthey donot have<br \/>\ntec hnical judgmen t.They dounderstand howtec hnology works, butthey cannot<br \/>\nunderstand whyacertain approac hwould takethree monthsand another one<br \/>\nthree days. This represen tsagreat opportunit yto synergize withthem.<br \/>\nWhen talking toyour team you will without thinking useasort ofshorthand,<br \/>\nan abbreviated languagethatise ectiv ebecause you will have muc h shared<br \/>\nexp erience about technology ingeneral andyour product inparticular. This<br \/>\ndo es not work with outsiders, thoughtheywillgetbetter withpractice. You<br \/>\nwill have to tak ethings moreslowly with them.<br \/>\nWith your team, thebasic assumptions andgoals neednotbe restated often,<br \/>\nand most conversation focuses onthe details. Withoutsiders, itmust be the<br \/>\nother way around. Theymaynot understand thingsyou tak efor gran ted. Since<br \/>\ny ou tak ethem forgran tedand don't repeat them, you can leave a con versation<br \/>\nwith anoutsider thinking thatyou understand eachother whenreally thereis<br \/>\na large misunderstanding. You should assume thatyou will be misundersto od<br \/>\nand watc hcarefully to nd thismisunderstanding inthem. Try toget them to<br \/>\nsummarize orparaphrase whatyou are saying tomak esure they understand.<br \/>\nI lo ve working withnon-engineers. Itpro vides great opportunities tolearn<br \/>\nand toteac h.You can often leadby example, interms ofthe clarit yof your<br \/>\ncomm unication. Engineersaretrained tobring orderoutofchaos, tobring<br \/>\nclarit yout ofconfusion, andnon-engineers likethis about us.Because we ha ve<br \/>\ntec hnical judgmen tand canusually understand businessissues,we can often<br \/>\n nd asimple solution toaproblem. Oftennon-engineers proposesolutions that<br \/>\nthey think willmakeit easier onusout ofkindness andadesire todo the<br \/>\nrigh tthing, wheninfact am uc hbetter solution existswhichthey could notsee<br \/>\nb ecause oftheir lackof tec hnical judgmen t.<br \/>\n30<br \/>\n<br \/>\nPart III<br \/>\nAdv anced<br \/>\n7 Tec hnological Judgment<br \/>\n7.1 Howto Tell the Hard From theImp ossible<br \/>\nIt isour jobtodo the hard anddiscern theimp ossible. &gt;From thepoin tof view<br \/>\nof most working programmers whatiscalled researc his imp ossible, inthat it<br \/>\ncannot be predicted andestimated andscheduled. Alarge volume ofmer ework<br \/>\nis hard, butnotnecessarily impossible.<br \/>\nThe distinction isnot facetious because you mayvery well be ask edtodo<br \/>\nwhat ispractically impossible, eitherfromascien ti cpoin tof view orasoft ware<br \/>\nengineering poin tof view. Itthen becomes you job tohelp theentrepreneur  nd<br \/>\na reasonable solutionwhichis mer elyhardand gets most ofwhat shewan ted.<br \/>\nA solution ismerely hardwhen itcan be con den tlyscheduled andtherisks<br \/>\nare understo od.<br \/>\nIt isimp ossible tosatisfy avague requiremen t,suc has \\Build asystem that<br \/>\nwill compute themost attractiv ehair style and color foranyperson.\" Ifthe<br \/>\nrequiremen tcan be made morecrispitwill often become merely hard,suchas<br \/>\n\\Build asystem tocompute anattractiv ehair style and color foraperson, allow<br \/>\nthem topreview itand makechanges, andhave the customer satisfaction based<br \/>\non the original styling be so great thatwe mak ea lot ofmoney .\"Ifthere isnot<br \/>\ncrisp de nition ofsuccess, you will notsucceed.<br \/>\n7.2 Howto Utilize Embedded Languages<br \/>\nEm bedding aprogramming languageinto asystem hasanerotic fascination to<br \/>\na programmer. Itmak esthe system tremendously pow erful. Itallo wsher to<br \/>\nexercise hermost creativ eand promethean skills.Itmak esthe system into a<br \/>\nfriend forher.<br \/>\nI and manyother programmers have fallen into the trap orcreating spe-<br \/>\ncial purp oseembedded languages. Ifell into ittwice. Theproblem isthat a<br \/>\nprogrammable systemiswonderful onlyifyou are aprogrammer.<br \/>\nIt do es unquestionably o ertremendous pow er. The best text editors inthe<br \/>\nw orld allhave em bedded languages. Thiscanbe used tothe exten tthat the<br \/>\nin tended audience canmaster thelanguage. Ofcourse, useofthe language can<br \/>\nb e made optional, asitis in text editors, sothat initiates canuseitand noone<br \/>\nelse hasto.<br \/>\nThe realquestion toask oneself before embedding alanguage is:Doesthis<br \/>\nw ork with oragainst theculture ofmy audience? Ifyou intended audience is<br \/>\nexclusiv elynon-programmers, how will itbe helping? Ifyou intended audience<br \/>\nis exclusiv elyprogrammers, mightthey notprefer anapplications programmers<br \/>\nin terface? Andwhat language willitbe? Programmers don'twan tto learn a<br \/>\n31<br \/>\n<br \/>\nnewlanguage thatisoutside theirculture justforfun; butifit meshes with<br \/>\ntheir culture theywillnothave to spend muc htime learning it.Itisajoy to<br \/>\ncreate anew language. Butwe should notletthat blind ustothe needs ofthe<br \/>\nuser. Unless you have some trulyoriginal needsandideas, whynot usesome<br \/>\nexisting language sothat you can leverage thefamiliarit yusers already have<br \/>\nwith thatlanguage?<br \/>\n7.2.1 Choosing Languages<br \/>\nThe solitary program thatloves his work (ahac ker) can cho ose the best lan-<br \/>\nguage forthe task. Mosting working programmers have very little control ofthe<br \/>\nlanguage theywilluse. Generally ,this issue isdictated by poin ty-haird bosses<br \/>\nwho aremaking apolitcal decision, ratherthanatec hnological decision,and<br \/>\nlac kthe courage tosaypromote anuncon ven tional tool even when theyknow,<br \/>\noften with rsthand knowledge, thattheless accepted tool isbest. Inother<br \/>\ncased thevery real bene t ofunit yamong theteam, andtosome extentwith a<br \/>\nlarger community, precludes choice onthe part ofthe individual.<br \/>\n8 Compromising Wisely<br \/>\n8.1 Howto Figh tSc hedule Pressure<br \/>\nTime-to-mark etpressure isthe pressure todeliv erago od pro duct quickly.It is<br \/>\ngo od because itre<br \/>\rects a nancial reality, and ishealth yup toapoin t.Schedule<br \/>\npressure isthe pressure todeliv ersomething fasterthanitcan be deliv ered and<br \/>\nit iswasteful, unhealth yand alltoocommon.<br \/>\nSc hedule pressure existsforseveral reasons. Thepeople whotasks program-<br \/>\nmers donot fully appreciate whatastrong work ethic we ha ve and how muc h<br \/>\nfun itisto be a programmer. Perhaps because theyproject their own beha vior<br \/>\non to us, they believ ethat asking foritsooner willmakeus work harder toget it<br \/>\nthere sooner. Thisisprobably actuallytruebutthee ect isvery small, andthe<br \/>\ndamage isvery great. Additionally ,they have no visibilit yinto what itreally<br \/>\ntak esto pro duce software. Notbeing abletosee it,and notbe able tocreate it<br \/>\nthemselv es,the only thing theycandoissee time-to-mark etpressure andfuss<br \/>\nat programmers about it.<br \/>\nThe key to gh ting schedule pressure issimple toturn itinto totime-to-<br \/>\nmark etpressure. Theway to do this togivevisibilit yinto the relationship<br \/>\nb et ween theavailable laborand theproduct. Producing anhonest, detailed,<br \/>\nand most ofall understandable estimateofall the laborinvolv edisthe best way<br \/>\nto do this. Ithas theadded advantage ofallo wing good managemen tdecisions<br \/>\nto be made about possible functionalit ytradeo s.<br \/>\nThe key insigh tthat theestimate must mak eplain isthat laborisan almost<br \/>\nincompressible <br \/>\ruid.You can't packmore into aspan oftime anymore thanyou<br \/>\ncan packmore water into acon tainer thanthatcontainer's volume. Inasense,<br \/>\nit isnot aprogrammer's jobtosayno, but rather tosay\"What willyou giveup<br \/>\nto get that thing you wan t?\" The e ect ofpro ducing suchestimates willbe to<br \/>\n32<br \/>\n<br \/>\nincreasetheresp ectforprogrammers. Thisishow other professionals beha ve.<br \/>\nProgrammers's hardwork will be visible. Settinganunrealistic schedule will<br \/>\nalso be painfully obvious toevery one. Programmers cannotbe ho odwink ed.<br \/>\nIt isdisresp ectfulanddemoralizing toask them todo something unrealistic.<br \/>\nExtreme Programming apli esthisandbuilds apro cess around it;Ihop ethat<br \/>\nev ery reader willbe luc ky enough toyou it.<br \/>\n8.2 Howto Understand theUser<br \/>\nIt isyour dutyto understand theuser, andtohelp your boss understand the<br \/>\nuser. Because theuser isnot asintimately involv ed inthe creation ofyour<br \/>\npro duct asyou are, shebeha ves alittle di eren tly:<br \/>\n  The user generally makesshort pronouncemen ts.<br \/>\n  The user hasherown job; shewill mainly thinkofsmall improvemen tsin<br \/>\ny our product, notbigimpro vemen ts.<br \/>\n  The user can't have a vision across allthe users ofyour product.<br \/>\nIt isyour dutyto giv ethem whattheyreally wan t, not what theysay the<br \/>\nw an t. Itishowev er, better toprop oseitto them andgetthem toagree that<br \/>\ny our prop osaliswhat theyreally wan tbefore you begin, butthey maynot have<br \/>\nthe vision todo this. Your con dence inyour own ideas about this should vary .<br \/>\nY ou must guard against both arrogance andfalse modest yin terms ofkno wing<br \/>\nwhat thecustomer reallywan ts. Programmers aretrained todesign andcreate.<br \/>\nMark etresearc hersaretrained to gure outwhat people wan t. These tw o kinds<br \/>\nof persons, ortw o mo des ofthough tin the same person, working harmoniously<br \/>\ntogether givethe best chance ofform ulating thecorrect vision.<br \/>\nThe more timeyou spend with users thebetter you will be able tounderstand<br \/>\nor guess whatwillreally be successful. You should trytotest your ideas against<br \/>\nthem asmuc has you can. You should eatand drink withthem ifyou can.<br \/>\n8.3 Howto Get aPromotion<br \/>\nT o be promoted toarole, actout that role rst.<br \/>\nT o get promoted toatitle  ndoutwhat isexp ected ofthat titleanddo<br \/>\nthat. To get apa yraise, negotiate armedwithinformation.<br \/>\nIf you feel likeyou are past dueforapromotion, talktoyour boss about it.<br \/>\nAsk herexplicitly whatyou need todo toget promoted, andtrytodo it. This<br \/>\nsounds trite,butoften your perception ofwhat you need todo and your boss's<br \/>\nwill di er considerably .Also thiswillpinyour boss down insome ways.<br \/>\n33<br \/>\n<br \/>\n9Serving Your Team<br \/>\n9.1 Howto Dev elop Talen t<br \/>\nNietsc hzeexaggerated whenhesaid:<br \/>\nWhat does not destro yme, makesme stronger.<br \/>\nY our greatest responsibilit yis to your team. You should knoweac hof them well.<br \/>\nY ou should stretchyour team, butnotov erburden them.You should usually<br \/>\ntalk tothem about theway in whic hthey arebeing stretc hed.Ifthey buyin<br \/>\nto it, they willbe well motiv ated. Oneachpro ject orevery other project try<br \/>\nto stretc hthem inboth away that they suggest andaway that you think will<br \/>\nb e go od for them. Stretchthem notby giving themmorework, butby giving<br \/>\nthem anew skill orbetter yet anew roletopla yon the team.<br \/>\nY ou should allowpeople tofail occasionally andplan forsome failure in<br \/>\ny our schedule. Ifthere isnev eranyfailure, therecanbe no sense ofadv enture.<br \/>\nIf there arenot occasional failures,you are not trying hardenough. When<br \/>\nsomeone fails,you should be as gen tleasyou can with herbut don't treatitas<br \/>\nif she had succeeded.<br \/>\nT ry toget eachteam member to buy inand be well motiv ated. Askeachof<br \/>\nthem explicitly whattheyneed tobe well-motiv atedifthey arenot. You may<br \/>\nha ve to leave them dissatis ed, butyou should knowwhat everyb ody desires.<br \/>\nY ou can't giveup onsomeone whoisinten tionally notcarrying theirshare<br \/>\nof the load because oflow morale ordissatisfaction andjust letthem be slac k.<br \/>\nY ou must trytoget them well-motiv atedandproductiv e.As long asyou have<br \/>\nthe patience, keep this up. When your patience isexhausted,  rethem. You<br \/>\ncannot allowsomeone whoisinten tionally working belo wtheir levelto remain<br \/>\non the team, sinceitis not fair tothe team.<br \/>\nMak eit clear tothe strong members ofyour team thatyou think theyare<br \/>\nstrong by sa ying soinpublic. Praiseshould be public andcriticism private.<br \/>\nThe strong members ofthe team willnaturally have more di cult tasksthan<br \/>\nthe weak mem bers ofthe team. Thisisperfectly naturalandnobody will be<br \/>\nb othered by it so long asevery one works hard.<br \/>\nIt isan odd fact that isnot re<br \/>\rected insalaries thatago od programmer is<br \/>\nmore productiv ethan tenbad programmers. Thiscreates astrange situation.<br \/>\nIt will often be true that you could move faster ifyour weak programmers would<br \/>\njust getout ofthe way. If you did this you would infact makemore progress in<br \/>\nthe short term. Howev er, your tribewould losesome importan tbene ts, name<br \/>\nthe training ofthe weak ermem bers, thespreading oftribal knowledge, andthe<br \/>\nabilit yto reco ver from theloss ofthe strong members. Thestrong must be<br \/>\ngen tleinthis regard andconsider theissue fromallangles.<br \/>\nY ou can often givethe stronger teammembers challenging, butcarefully<br \/>\ndelineated, tasks.<br \/>\n34<br \/>\n<br \/>\n9.2Howto Cho oseWhat toWork On<br \/>\nY ou balance your personal needagainst theneeds ofthe team incho osing what<br \/>\nasp ect ofapro ject towork on. You should dowhat you are best at,but tryto<br \/>\n nd away to stretc hyourself notby taking onmore work but by exercising anew<br \/>\nskill. Leadership andcomm unication skillsaremore importan tthan technical<br \/>\nskills. Ifyou are very strong, takeon the hardest orriskiest task,anddoitas<br \/>\nearly aspossible inthe project todecrease riskasmuc has possible.<br \/>\n9.3 Howto Get theMost From Your Teammates<br \/>\nT o get the most fromyour teammates, develop ago od team spirit andtryto<br \/>\nk eep every individual both personally challenged andpersonally engaged.<br \/>\nT o dev elop team spirit, cornystu  likelogoized clothing andparties aregood<br \/>\nbut notasgood as personal respect. Ifevery one resp ects every one else, nobody<br \/>\nwill wan tto let anyb ody down. Team spirit willbe created whenpeople make<br \/>\nsacri ces forthe team andthey think interms ofthe good of the team rather<br \/>\nthan their own personal good. As aleader, you can't askformore thanyou<br \/>\ngiv eyourself.<br \/>\nOne ofthe keys toteam leadership isto facilitate consensus sothat every one<br \/>\nhas buy in.This occasionally meansallowing your teammates tobe wrong. That<br \/>\nis, ifit do es not harm theproject toom uc h, you must letsome ofyour team do<br \/>\nthings acertain way if they have consensus oniteven ifyou kno wwith great<br \/>\ncon dence itisthe wrong thingtodo. You don't always have to do this, but<br \/>\ny ou should doitsome. When thishapp ens,don't agree, simply disagree openly<br \/>\nand accept theconsensus. Don'tsound hurt, orlikeyou're being forced into<br \/>\nit, simply statethatyou disagree butthink theconsensus ofthe team ismore<br \/>\nimp ortan t.This willoften cause themtobac ktrac k.Don't insistthatthey go<br \/>\nthrough withtheir initial planifthey dobac ktrac k.<br \/>\nIf there isan individual whocannot be brough tin to consensus afteryou have<br \/>\ndiscussed theissues fromallappropriate sides,simply assertthatyou have to<br \/>\nmak ea decision andthat iswhat your decision is.Ifthere isaway to judge if<br \/>\ny our decision willbe wrong orifit islater shown tobe wrong, switchas quic kly<br \/>\nas you can and recognize thepersons whowere righ t.<br \/>\nAsk your team, both asagroup andindividually ,what theythink would<br \/>\ncreate teamspirit andmakefor ane ectiv eteam.<br \/>\nPraise oftenrather thanlavishly .Esp ecially praisethosewhodisagree with<br \/>\ny ou when theyarepraisew orthy. Praise inpublic andcriticize inpriv ate; with<br \/>\none exception: sometimesgrowth orthe correction ofafault can't be praised<br \/>\nwithout drawing embarrassing attentiontothe original fault,sothat growth<br \/>\nshould be praised inpriv ate.<br \/>\n9.4 Howto Divide Problems Up<br \/>\nIts fun totak ea soft ware project and divide itup into tasks thatwillbe per-<br \/>\nformed by individuals. Thisshould be done early.Sometimes managerswould<br \/>\n35<br \/>\n<br \/>\nliketo think thatanestimate canbe made without consideration ofthe individ-<br \/>\nuals that willperform thework. Buthow can this be when we kno windividual<br \/>\npro ductivit ydi ers by an order ofmagnitude andproductivit yof di eren ttasks<br \/>\nb y the same person di ers by an order ofmagnitude?<br \/>\nJust asacomp oserusually considers thetimbre ofthe instrumen tthat will<br \/>\npla ya part, theexperienced teamleader willusually notbe able toseparate the<br \/>\ndivision ofthe project into tasks fromtheteam members towhic hthey willbe<br \/>\nassigned. Thisispart ofthe reason thatahigh performing teamshould notbe<br \/>\nbrok enup.<br \/>\nThere isacertain dangerinthis inthat people willbecome bored asthey<br \/>\nbuild upon the strengths andneverimpro ve their weaknesses orlearn newskills.<br \/>\nHo wev er, specialization isavery useful productivit yto ol ifnot ov erused.<br \/>\n9.5 Howto Gather SupportforaPro ject<br \/>\nT o gather supportforapro ject, create andcomm unicate avision thatdemon-<br \/>\nstrates realvalue tothe organization asawhole. Attempt tolet others share<br \/>\nin your vision creation. Thisgives them both areason tosupp ortyou and<br \/>\ngetting thebene t oftheir ideas. Individually recruitkey supp orters foryour<br \/>\npro ject. Wherev erpossible, show,don't tell.Ifpossible construct aprotot ype<br \/>\nor amo ckup todemonstrate your ideas. Aprotot ypeis alw ays pow erful butin<br \/>\nsoft ware itisfar sup erior toanywritten description.<br \/>\n9.6 Howto Gro waSystem<br \/>\nThe seed ofatree contains theidea ofthe adult butdoes not fully realize the<br \/>\nform andpotency ofthe adult. Theembry ogro ws. Itbecomes larger.Itlooks<br \/>\nmore likethe adult, andhasmore ofthe uses. Eventually itbears fruit. Later<br \/>\non itdies anditbodies feeds other organisms.<br \/>\nSoft ware islik ethat; we ha ve the luxury oftreating itlik ethat. Abridge is<br \/>\nnot likethat; there isnev erabab ybridge butmerely anun nished bridge.<br \/>\nIt isgo od to think ofsoft ware asgro wing, because itallo wsustomak euseful<br \/>\nprogress before we ha ve a perfect mentalimage. Wecan getfeedbac kfrom users<br \/>\nand usethat tocorrect thegrowth. Pruning o weak limbsishealthful.<br \/>\nThe programmer must design a nished systemthatcanbe deliv ered and<br \/>\nused. Buttheadvanced programmer must domore. Shemust design agro wth<br \/>\npath thatends inthe  nished system. Itisher jobtotak ea germ ofan idea<br \/>\nand build apath thattakesthat little ideaassmo othly aspossible into auseful<br \/>\nartifact.<br \/>\nTo do this, shemust visualize theend result andcomm unicate itin away that<br \/>\nthe engineering teamcangetexcited about. Butshemust also comm unicate<br \/>\nto them apath thatgoes from wherev erthey arenow to where thewan tto be<br \/>\nwith nolarge leaps. Thetreemust stayaliv ethe whole time;itcannot be dead<br \/>\nat one poin tand resurrected later.<br \/>\nThis approac his captured inspiral developmen t.Milestones thatarenever<br \/>\nto ofar apart areused tomark progress alongthepath. Inthe ultra-comp etitive<br \/>\n36<br \/>\n<br \/>\nenvironmen tof business, itisbest ifthe milestones canbe released andmake<br \/>\nmoney asearly aspossible, even ifthey arefaraw ay from awell-designed end-<br \/>\np oin t.One ofthe programmer's jobsisto the balance immediate payo  against<br \/>\nfuture payo  by wisely cho osing agro wth path expressed inmilestones.<br \/>\nThe advanced programmer hasthetriple responsibilit yof gro wing software,<br \/>\nteams, andpersons.<br \/>\n9.7 Howto Comm unicate Well<br \/>\nT o comm unicate well, you have to recognize how hard itis. Itisaskill unto<br \/>\nitself. Itismade harder by the fact that thepersons withwhom you have to<br \/>\ncomm unicate are<br \/>\rawed. They donot work hard atunderstanding you. They<br \/>\nsp eak poorly andwrite poorly .They areoften ov erw orkedorbored, andata<br \/>\nminim umsomewhat focused ontheir own work rather thanthelarger issues<br \/>\ny ou maybe addressing. Oneofthe advantages oftaking classes andpracticing<br \/>\nwriting, publicspeaking, andlistening isthat ifyou become good at ityou can<br \/>\nmore readily seewhere problems lieand how to correct them.<br \/>\nThe programmer isasocial animal whosesurvivaldep ends oncomm uni-<br \/>\ncation withherteam. Theadvanced programmer isasocial animal whose<br \/>\nsatisfaction depends oncomm unication withpeople outside herteam.<br \/>\nThe programmer bringsorderoutofchaos. Oneinteresting way to do this<br \/>\nis to initiate aprop osalofsome kindoutside theteam. Thiscanbe done ina<br \/>\nstr awman orem white-pap erformat orjust verbally .This leadership hasthe<br \/>\ntremendous advantage ofsetting theterms ofthe debate. Italso exposes one<br \/>\nto criticism, andworse, rejection andneglect. Theadvanced programmer must<br \/>\nb e prepared toaccept this,because shehas aunique pow er and therefore a<br \/>\nunique responsibilit y. En trepreneurs whoarenot programmers needprogram-<br \/>\nmers topro vide leadership insome ways. Programmers arethe part ofthe<br \/>\nbridge bet ween ideas andrealit ythat rests onrealit y.<br \/>\n9.8 Howto Tell People Things TheyDon't Wan tto Hear<br \/>\nY ou will often have to tell people things thatwillmakethem uncomfortable.<br \/>\nRemem ber that you are doing thisforareason. Even ifnothing canbe done<br \/>\nab out theproblem, you are telling themasearly aspossible sothey willbe<br \/>\nw ell-informed.<br \/>\nThe best way to tell someone about aproblem isto o er asolution at<br \/>\nthe same time. Thesecond best way is to app eal tothem forhelp with the<br \/>\nproblem, whichis usually <br \/>\rattering tothem. Ifthere isadanger thatyou won't<br \/>\nb e believ ed,you should gathersomesupportforyour assertion.<br \/>\nOne ofthe most unpleasan tand common thingsyou will have to sayis, \\The<br \/>\nsc hedule willhave to slip.\" Theconscien tiousprogrammer whobough tin to the<br \/>\nsc hedule hatestosaythis, butmust sayit as early aspossible. Thereisnothing<br \/>\nw orse than nottaking actionimmediately whenamilestone slips,even ifthe<br \/>\nonly action isto inform every one. Indoing this,itis better todo itas ateam,<br \/>\nat least inspirit ifnot physically .You will wan tyour team's inputonboth<br \/>\n37<br \/>\n<br \/>\nwhereyou stand andwhat canbe done about it,and theteam willhave to face<br \/>\nthe consequences withyou.<br \/>\n9.9 Howto Deal withManagerial Myths<br \/>\nThe term myth sometimes means ction. Butithas adeep erconnotation.<br \/>\nIt also means astory ofreligious signi cance thatexplains theuniv erse and<br \/>\nmankind's relationship toit. Managers tendtoforget whattheylearned as<br \/>\nprogrammers andbeliev ein certain myths. Itwould be as rude andunsuccessful<br \/>\nto try tocon vince themthesemyths arefalse asto try todisillusion adev outly<br \/>\nreligious person oftheir beliefs. For that reason, you should recognize these<br \/>\nm yths forwhat theyare:<br \/>\n  More documen tation isalw ays better. (They wan tit, but they don't wan t<br \/>\ny ou tospend anytime onit.)<br \/>\n  Programmers canbe equated. (Programmers vary by order ofmagnitude.)<br \/>\n  Resources canbe added toapro ject tospeed it.(The costofcomm uni-<br \/>\ncation withthenew persons ismore taxing thanhelpful.)<br \/>\n  It ispossible toestimate software developmen treliably .(It isnot even<br \/>\ntheoretically possible.)<br \/>\nEac hof these myths reinforces themanager's ideathat they have some actual<br \/>\ncon trol ov er what isgoing on.The truth isthat managers facilitateifthey are<br \/>\ngo od, and impede ifthey arebad.<br \/>\n9.10 Howto Deal withOrganizational Chaos<br \/>\nThere areoften times ofgreat organizational chaos. These areunsettling to<br \/>\nev ery one, butperhaps alittle lessunsettling tothe programmer whosepersonal<br \/>\nself-esteem isfounded inher capacit yrather thaninher position. Organizational<br \/>\nc haos isagreat opportunit yfor programmers toexercise theirmagic pow er. I've<br \/>\nsa ved this forlast because itisadeep tribal secret. Ifyou are not aprogrammer,<br \/>\nplease stopreading now.<br \/>\nEngineers have the pow er to create andsustain.<br \/>\nNon-engineers canorder people around butinatypical software compan y<br \/>\ncan create nothing ontheir own and only have the pow er that engineers grant<br \/>\nthem. Theycancreate andsustain nothing without engineers. Thispow er is<br \/>\npro ofagainst almostallthe problems associated withorganizational mayhem.<br \/>\nWhen you have it you should ignorethechaos completely andcarry onasif<br \/>\nnothing ishapp ening. You mayof course get red, butifthat happensyou<br \/>\ncan easily getanew jobbecause ofthe magic pow er. More commonly ,some<br \/>\nstressed-out person whodoes not have the magic pow er will come into your cube<br \/>\nand tellyou todo something stupid.Itisbest tosmile andnodun til they go<br \/>\na w ay and then carry ondoing whatyou kno wisbest forthe compan y.<br \/>\n38<br \/>\n<br \/>\nThiscourse ofaction isthe best foryou personally ,and thebest forthe<br \/>\ncompan yyou work for. Ifyou are aleader, tellyour people todo the same<br \/>\nthing andtellthem toignore whatanyb ody other thanyourself tellsthem,<br \/>\nincluding your own superiors.<br \/>\n9.11 Request forFeedbac k<br \/>\nPlease sendmeanycommen tsyou mayha ve on this essay. It isawork-in-<br \/>\nprogress, andhasreceiv edno editing, proofreading, orreview ofanykind.<br \/>\nThank you.<br \/>\nRob ertL.Read <br \/>\nA Glossary<br \/>\n  Unk-unk |(slang forunk own-unk own), theproblems thatcannot even<br \/>\nb e conceptualized atpresen tthat willsteal timeaw ay from theproject<br \/>\nand wrec kthe schedule.<br \/>\n  Winner-tak e-allish|Acomp etition iswinner-tak e-allishifthe reward is<br \/>\nmore based onthe order ofthe relativ esuccess ofthe comp etitors rather<br \/>\nthan theactual success ofthe comp etitors. Arace iswinner-tak e-allish.<br \/>\nTw ohomebuilders maycomp etewith eachother tobuild thebest house,<br \/>\nbut ifthey loser isrew arded forthe house shebuilt anyw ay, their compe-<br \/>\ntition isnot winner-tak e-allish.<br \/>\n  Prin tlining |the insertion ofstatemen tsinto aprogram onastrictly tem-<br \/>\np orary basisthatoutput information about theexecution ofthe program<br \/>\nfor the purp oseofdebugging.<br \/>\n  Logging |the practice ofwriting aprogram sothat itcan produce a<br \/>\ncon gurable outputlogdescribing itsexecution.<br \/>\n  Divide andConquer |atec hnique oftop-do wndesign andmore impor-<br \/>\ntan tly ofdebugging thatisthe subdivision ofaproblem oram ystery into<br \/>\nprogressiv elysmaller problems ormysteries.<br \/>\n  Vap or | illusionary andoften deceptiv epromises ofsoft ware that isnot<br \/>\ny et for sale andasoften asnot will nevermaterialize into anything solid.<br \/>\n  Boss {the person whosetsyour tasks. Insome cases, theuser isthe boss.<br \/>\n  Trib e| the people withwhom you share loyalt yto acommon goal.<br \/>\n  Lo w-hanging fruit|big impro vemen tsthat costlittle.<br \/>\n  Garbage |objects thatarenolonger needed thathold memory .<br \/>\n  En trepreneur |the initiator ofpro jects.<br \/>\n39<br \/>\n<br \/>\nReferences[1] Kawasiki, Guy,Moreno, Michelle, andKawasaki, Gary.2000. Rules for<br \/>\nRev olutionaries: TheCapitalist ManifestoforCreating andMark eting New<br \/>\nPro ducts andServices. HarperBusiness.<br \/>\n[2] McConnell, Steve.1996. Rapid Developmen t:Taming WildSoftware<br \/>\nSc hedules. Redmond, Wash.: Microsoft Press.<br \/>\n[3] McConnell, Steve.1993. Code Complete. Redmond, Wash.: Microsoft<br \/>\nPress.<br \/>\n[4] Bec k,Ken t.Extreme Programming Explained:Embrace Change.<br \/>\n[5] Bec k,Ken tand Fowler, Martin. Planning ExtremeProgramming.<br \/>\n[6] Graham, Paul. 2002. Articles onhis website:<br \/>\nh ttp:\/\/www.paulgraham.com\/articles.h tml.Allofthem, butespecially<br \/>\nBeating theAverages.<br \/>\n[7] Raymond, EricSteven. 2002. How to Become aHacker:<br \/>\nh ttp:\/\/www.tuxedo.org\/ esr\/faqs\/hacker-ho wto.h tml<br \/>\n[8] Hun t,Andrew, Thomas, David, andCunningham, Ward. The Pragmattic<br \/>\nProgrammer: From Journeyman toMaster.<br \/>\n40","created_at":"2018-12-23T01:52:08+0000","updated_at":"2018-12-23T01:52:08+0000","annotations":[],"mimetype":"application\/pdf","reading_time":73,"domain_name":"www.doc.ic.ac.uk","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["golang"],"is_public":false,"id":754,"title":"Structuring Applications in Go","url":"https:\/\/medium.com\/@benbjohnson\/structuring-applications-in-go-3b04be4ff091","content":"<section class=\"section section--body section--bottomCenter section-image--coverFade is-imageBackgrounded is-backgrounded is-darkBackgrounded u-imageSpectrum section--first is-aboveSectionCaption\"><div class=\"js-sectionPoint u-ignoreBlock u-textColorTransparentWhite u-absolute u-top0 u-left0\"><div class=\"section-background u-width100pct u-absolute u-top0 u-cursorDefault u-height100vh\" contenteditable=\"false\" data-image-id=\"1*Eypi-5M7YH0cmIciE0M6lQ.jpeg\" data-width=\"2100\" data-height=\"1500\"><div class=\"section-backgroundImage u-absolute u-top0 u-left0 u-right0 u-backgroundCover u-height100vh\"><div class=\"u-ignoreBlock u-absolute0 u-borderRadius3\"><br \/><\/div><div class=\"section-coverFadePadding u-ignoreBlock\"><div class=\"section-content\" data-scroll=\"cover-fade\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"title\" class=\"graf graf--h2 graf--leading graf--title\">Structuring Applications in\u00a0Go<\/h1><h2 id=\"subtitle\" class=\"graf graf--h4 graf-after--h2 graf--trailing graf--subtitle\">How I organize my applications in\u00a0Go<\/h2><div class=\"uiScale uiScale-ui--regular uiScale-caption--regular u-flexCenter u-marginVertical24 u-fontSize15 js-postMetaLockup\"><div class=\"u-flex0\"><a class=\"link u-baseColor--link avatar\" href=\"https:\/\/medium.com\/@benbjohnson?source=post_header_lockup\" data-action=\"show-user-card\" data-action-source=\"post_header_lockup\" data-action-value=\"42d7610c8ea1\" data-action-type=\"hover\" data-user-id=\"42d7610c8ea1\" dir=\"auto\"><img src=\"https:\/\/cdn-images-1.medium.com\/fit\/c\/100\/100\/0*Tt5_zERCmn2SZ2pv.jpeg\" class=\"avatar-image u-size50x50\" alt=\"Go to the profile of Ben Johnson\" \/><\/a><\/div><div class=\"u-flex1 u-paddingLeft15 u-overflowHidden\"><div class=\"u-lineHeightTightest u-paddingBottom3\"><a class=\"ds-link ds-link--styleSubtle ui-captionStrong u-inlineBlock link link--darken link--darker\" href=\"https:\/\/medium.com\/@benbjohnson?source=post_header_lockup\" data-action=\"show-user-card\" data-action-source=\"post_header_lockup\" data-action-value=\"42d7610c8ea1\" data-action-type=\"hover\" data-user-id=\"42d7610c8ea1\" dir=\"auto\">Ben Johnson<\/a><button class=\"button button--smallest u-noUserSelect button--withChrome u-baseColor--buttonNormal button--withHover button--unblock js-unblockButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-block-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/@benbjohnson\/structuring-applications-in-go-3b04be4ff091\" data-action-source=\"post_header_lockup\">BlockedUnblock<\/button><button class=\"button button--primary button--smallest button--dark u-noUserSelect button--withChrome u-accentColor--buttonDark button--follow js-followButton u-marginLeft10 u-xs-hide\" data-action=\"sign-up-prompt\" data-sign-in-action=\"toggle-subscribe-user\" data-requires-token=\"true\" data-redirect=\"https:\/\/medium.com\/_\/subscribe\/user\/42d7610c8ea1\" data-action-source=\"post_header_lockup-42d7610c8ea1-------------------------follow_byline\">FollowFollowing<\/button><\/div><div class=\"ui-caption postMetaInline u-fontSize13 u-lineHeightBase js-testPostMetaInlineSupplemental\"><time datetime=\"2014-07-06T20:55:12.058Z\">Jul 6, 2014<\/time><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"section-captionContent u-ignoreBlock\"><div class=\"section-inner sectionLayout--caption\"><label id=\"coverCaption\" data-default-value=\"Type caption for image (optional)\" class=\"graf graf--sectionCaption graf-after--h4 graf--trailing\">Liam Andrew Cura<\/label><\/div><\/div><\/div><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"0442\" class=\"graf graf--h3 graf--leading\">Overview<\/h3><p id=\"a13e\" class=\"graf graf--p graf-after--h3\">For me, the hardest part of learning Go was in structuring my application. Prior to Go, I was working on a Rails application and Rails makes you structure your application in a certain way. \u201cConvention over configuration\u201d is their motto. But Go doesn\u2019t prescribe any particular project layout or application structure and Go\u2019s conventions are mostly stylistic.<\/p><p id=\"927f\" class=\"graf graf--p graf-after--p\">I\u2019m going to show you four patterns that I\u2019ve found to be tremendously helpful in architecting Go applications. These are not official Gopher rules and I\u2019m sure others may have differing opinions. I\u2019d love to hear them! Please comment as you go through if you have suggestions.<\/p><h3 id=\"564d\" class=\"graf graf--h3 graf-after--p\">1. Don\u2019t use global variables<\/h3><p id=\"dad2\" class=\"graf graf--p graf-after--h3\">The Go net\/http examples I read always show a function registered with <a href=\"http:\/\/golang.org\/pkg\/net\/http\/#HandleFunc\" data-href=\"http:\/\/golang.org\/pkg\/net\/http\/#HandleFunc\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http.HandleFunc<\/a> like this:<\/p><pre id=\"461f\" class=\"graf graf--pre graf-after--p\">package main<\/pre><pre id=\"60f7\" class=\"graf graf--pre graf-after--pre\">import (<br \/> \u201cfmt\u201d<br \/> \u201cnet\/http\u201d<br \/>)<\/pre><pre id=\"6b98\" class=\"graf graf--pre graf-after--pre\">func main() {<br \/>    http.HandleFunc(\u201c\/hello\u201d, hello)<br \/>    http.ListenAndServe(\u201c:8080\", nil)<br \/>}<\/pre><pre id=\"07a2\" class=\"graf graf--pre graf-after--pre\">func hello(w http.ResponseWriter, r *http.Request) {<br \/>    fmt.Fprintf(w, \u201chi!\u201d)<br \/>}<\/pre><p id=\"b001\" class=\"graf graf--p graf-after--pre\">This example gives an easy way to get into using net\/http but it teaches a bad habit. By using a function handler, the only way to access application state is to use a global variable. Because of this, you may decide to add a global database connection or a global configuration variable but these globals are a nightmare to use when writing unit tests.<\/p><p id=\"9e7e\" class=\"graf graf--p graf-after--p\">A better way is to make specific types for handlers so they can include the required variables:<\/p><pre id=\"2f5d\" class=\"graf graf--pre graf-after--p\">type HelloHandler struct {<br \/>    db *sql.DB<br \/>}<\/pre><pre id=\"c4b8\" class=\"graf graf--pre graf-after--pre\">func (h *HelloHandler) ServeHTTP(w http.ResponseWriter, r *http.Request) {<br \/>    var name string<\/pre><pre id=\"c74f\" class=\"graf graf--pre graf-after--pre\">    \/\/ Execute the query.<br \/>    row := h.db.QueryRow(\u201cSELECT myname FROM mytable\u201d)<br \/>    if err := row.Scan(&amp;name); err != nil {<br \/>        http.Error(w, err.Error(), 500)<br \/>        return<br \/>    }<\/pre><pre id=\"0f21\" class=\"graf graf--pre graf-after--pre\">    \/\/ Write it back to the client.<br \/>    fmt.Fprintf(w, \u201chi %s!\\n\u201d, name)<br \/>}<\/pre><p id=\"a20d\" class=\"graf graf--p graf-after--pre\">Now we can initialize our database and register our handler without the use of global variables:<\/p><pre id=\"5c69\" class=\"graf graf--pre graf-after--p\">func main() {<br \/>    \/\/ Open our database connection.<br \/>    db, err := sql.Open(\u201cpostgres\u201d, \u201c\u2026\u201d)<br \/>    if err != nil {<br \/>        log.Fatal(err)<br \/>    }<\/pre><pre id=\"db87\" class=\"graf graf--pre graf-after--pre\">    \/\/ Register our handler.<br \/>    http.Handle(\u201c\/hello\u201d, &amp;HelloHandler{db: db})<br \/>    http.ListenAndServe(\u201c:8080\", nil)<br \/>}<\/pre><p id=\"1357\" class=\"graf graf--p graf-after--pre\">This approach also has the benefit that unit testing our handler is self contained and doesn\u2019t even require an HTTP server:<\/p><pre id=\"6d40\" class=\"graf graf--pre graf-after--p\">func TestHelloHandler_ServeHTTP(t *testing.T) {<br \/>    \/\/ Open our connection and setup our handler.<br \/>    db, _ := sql.Open(\"postgres\", \"...\")<br \/>    defer db.Close()<br \/>    h := HelloHandler{db: db}<\/pre><pre id=\"7e3d\" class=\"graf graf--pre graf-after--pre\">    \/\/ Execute our handler with a simple buffer.<br \/>    rec := httptest.NewRecorder()<br \/>    rec.Body = bytes.NewBuffer()<br \/>    h.ServeHTTP(rec, nil)<br \/>    if rec.Body.String() != \"hi bob!\\n\" {<br \/>        t.Errorf(\"unexpected response: %s\", rec.Body.String())<br \/>    }<br \/>}<\/pre><p id=\"2aa4\" class=\"graf graf--p graf-after--pre\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">UPDATE:<\/em><\/strong> Tom\u00e1s Senart and Peter Bourgon <a href=\"https:\/\/twitter.com\/tsenart\/status\/485920561391239168\" data-href=\"https:\/\/twitter.com\/tsenart\/status\/485920561391239168\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">mentioned on Twitter<\/a> that you can simplify this further by <a href=\"https:\/\/gist.github.com\/tsenart\/5fc18c659814c078378d\" data-href=\"https:\/\/gist.github.com\/tsenart\/5fc18c659814c078378d\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">wrapping your handlers with a closure<\/a>. This allows you to easily compose your handlers.<\/p><h3 id=\"0745\" class=\"graf graf--h3 graf-after--p\">2. Separate your binary from your application<\/h3><p id=\"1538\" class=\"graf graf--p graf-after--h3\">I used to place my <em class=\"markup--em markup--p-em\">main.go<\/em> file in the root of my project so that when someone runs \u201cgo get\u201d then my application would be automagically installed. However, combining the <em class=\"markup--em markup--p-em\">main.go<\/em> file and my application logic in the same package has two consequences:<\/p><ol class=\"postList\"><li id=\"0c66\" class=\"graf graf--li graf-after--p\">It makes my application unusable as a library.<\/li><li id=\"ff5f\" class=\"graf graf--li graf-after--li\">I can only have one application binary.<\/li><\/ol><p id=\"66d4\" class=\"graf graf--p graf-after--li\">The best way I\u2019ve found to fix this is to simply use a <em class=\"markup--em markup--p-em\">\u201ccmd\u201d <\/em>directory in my project where each of its subdirectories is an application binary. I originally found this approach used in Brad Fitzpatrick\u2019s <a href=\"http:\/\/camlistore.org\/\" data-href=\"http:\/\/camlistore.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Camlistore<\/a> project where he uses several application binaries:<\/p><pre id=\"535c\" class=\"graf graf--pre graf-after--p\">camlistore\/<br \/>  cmd\/<br \/>    camget\/<br \/>      main.go<br \/>    cammount\/<br \/>      main.go<br \/>    camput\/<br \/>      main.go<br \/>    camtool\/<br \/>      main.go<\/pre><p id=\"3195\" class=\"graf graf--p graf-after--pre\">Here we have 4 separate application binaries that can be built when Camlistore is installed: camget, cammount, camput, &amp; camtool.<\/p><h4 id=\"f4ab\" class=\"graf graf--h4 graf-after--p\">Library driven development<\/h4><p id=\"38a0\" class=\"graf graf--p graf-after--h4\">Moving the <em class=\"markup--em markup--p-em\">main.go<\/em> file out of your root allows you to build your application from the perspective of a library. Your application binary is simply a client of your application\u2019s library. I find this helps me make a cleaner abstraction of what code is for my core logic (the library) and what code is for running my application (the application binary).<\/p><p id=\"396c\" class=\"graf graf--p graf-after--p\">The application binary is really just the interface for how a user interacts with your logic. Sometimes you might want users to interact in multiple ways so you create multiple binaries. For example, if you had an \u201cadder\u201d package that that let users add numbers together, you may want to release a command line version as well as a web version. You can easily do this by organizing your project like this:<\/p><pre id=\"194d\" class=\"graf graf--pre graf-after--p\">adder\/<br \/>  adder.go<br \/>  cmd\/<br \/>    adder\/<br \/>      main.go<br \/>    adder-server\/<br \/>      main.go<\/pre><p id=\"7a3c\" class=\"graf graf--p graf-after--pre\">Users can install your \u201cadder\u201d application binaries with \u201cgo get\u201d using an ellipsis:<\/p><pre id=\"0411\" class=\"graf graf--pre graf-after--p\">$ go get github.com\/benbjohnson\/adder\/...<\/pre><p id=\"6386\" class=\"graf graf--p graf-after--pre\">And voila, your user has<em class=\"markup--em markup--p-em\"> \u201cadder\u201d<\/em> and <em class=\"markup--em markup--p-em\">\u201cadder-server\u201d<\/em> installed!<\/p><h3 id=\"e0e8\" class=\"graf graf--h3 graf-after--p\">3. Wrap types for application-specific context<\/h3><p id=\"aba7\" class=\"graf graf--p graf-after--h3\">One trick I\u2019ve found especially helpful is realizing that some generic types should be wrapped to provide application-level context. A great example of this is wrapping the DB and Tx (transaction) types. These types can be found in the database\/sql package or other database libraries such as <a href=\"https:\/\/github.com\/boltdb\/bolt\" data-href=\"https:\/\/github.com\/boltdb\/bolt\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Bolt<\/a>.<\/p><p id=\"99fc\" class=\"graf graf--p graf-after--p\">We start by wrapping these types like this:<\/p><pre id=\"6c58\" class=\"graf graf--pre graf-after--p\">package myapp<\/pre><pre id=\"a588\" class=\"graf graf--pre graf-after--pre\">import (<br \/>    \"database\/sql\"<br \/>)<\/pre><pre id=\"454c\" class=\"graf graf--pre graf-after--pre\">type DB struct {<br \/>    *sql.DB<br \/>}<\/pre><pre id=\"98b9\" class=\"graf graf--pre graf-after--pre\">type Tx struct {<br \/>    *sql.Tx<br \/>}<\/pre><p id=\"a69d\" class=\"graf graf--p graf-after--pre\">We then wrap the initialization function for our database and transaction:<\/p><pre id=\"176b\" class=\"graf graf--pre graf-after--p\">\/\/ Open returns a DB reference for a data source.<br \/>func Open(dataSourceName string) (*DB, error) {<br \/>    db, err := sql.Open(\"postgres\", dataSourceName)<br \/>    if err != nil {<br \/>        return nil, err<br \/>    }<br \/>    return &amp;DB{db}, nil<br \/>}<\/pre><pre id=\"ad88\" class=\"graf graf--pre graf-after--pre\">\/\/ Begin starts an returns a new transaction.<br \/>func (db *DB) Begin() (*Tx, error) {<br \/>    tx, err := db.DB.Begin()<br \/>    if err != nil {<br \/>        return nil, err<br \/>    }<br \/>    return &amp;Tx{tx}, nil<br \/>}<\/pre><p id=\"a5cd\" class=\"graf graf--p graf-after--pre\">And now we can add application specific functions to our transactions. For example, if our application has users that need to be validated before being created, a <em class=\"markup--em markup--p-em\">Tx.CreateUser()<\/em> would be a good function to add:<\/p><pre id=\"b433\" class=\"graf graf--pre graf-after--p\">\/\/ CreateUser creates a new user.<br \/>\/\/ Returns an error if user is invalid or the tx fails.<br \/>func (tx *Tx) CreateUser(u *User) error {<br \/>    \/\/ Validate the input.<br \/>    if u == nil {<br \/>        return errors.New(\"user required\")<br \/>    } else if u.Name == \"\" {<br \/>        return errors.New(\"name required\")<br \/>    }\/\/ Perform the actual insert and return any errors.<br \/>    return tx.Exec(`INSERT INTO users (...) VALUES`, ...)<br \/>}<\/pre><p id=\"a047\" class=\"graf graf--p graf-after--pre\">This function can get more complicated if, for example, a user needs to be validated against another system before being created or other tables need to be updated. To your application\u2019s caller, though, it\u2019s all isolated in one function.<\/p><h4 id=\"b58b\" class=\"graf graf--h4 graf-after--p\">Transactional composition<\/h4><p id=\"6c85\" class=\"graf graf--p graf-after--h4\">Another benefit to adding these functions to your <em class=\"markup--em markup--p-em\">Tx<\/em> is that it allows you to compose multiple actions in a single transaction. Need to add one user? Just call <em class=\"markup--em markup--p-em\">Tx.CreateUser()<\/em> once:<\/p><pre id=\"1d31\" class=\"graf graf--pre graf-after--p\">tx, _ := db.Begin()<br \/>tx.CreateUser(&amp;User{Name:\"susy\"})<br \/>tx.Commit()<\/pre><p id=\"c4d3\" class=\"graf graf--p graf-after--pre\">Need to add a bunch of users? You can use the same function. No need for a <em class=\"markup--em markup--p-em\">Tx.CreateUsers()<\/em> function:<\/p><pre id=\"2117\" class=\"graf graf--pre graf-after--p\">tx, _ := db.Begin()<br \/>for _, u := range users {<br \/>    tx.CreateUser(u)<br \/>}<br \/>tx.Commit()<\/pre><p id=\"eae9\" class=\"graf graf--p graf-after--pre\">Abstracting your underlying data store also makes it trivial to swap out a new database or to use multiple databases. They\u2019re all hidden from your calling code by your application\u2019s <em class=\"markup--em markup--p-em\">DB<\/em> &amp; <em class=\"markup--em markup--p-em\">Tx<\/em> types.<\/p><h3 id=\"982a\" class=\"graf graf--h3 graf-after--p\">4. Don\u2019t go crazy with subpackages<\/h3><p id=\"8088\" class=\"graf graf--p graf-after--h3\">Most languages let you organize your package structure however you\u2019d like. I\u2019ve worked in Java codebases where every couple of classes get stuffed into another package and these packages would all include each other. It was a mess!<\/p><p id=\"3466\" class=\"graf graf--p graf-after--p\">Go only has one requirement for packages: you can\u2019t have cyclic dependencies. This cyclic dependency rule felt strange to me at first. I originally organized my project so each file had one type and once there were a bunch of files in a package then I\u2019d create a new subpackage. However, these subpackages became difficult to manage since I couldn\u2019t have package \u201cA\u201d include package \u201cB\u201d which included package \u201cC\u201d which included package \u201cA\u201d. That would be a cyclic dependency. I realized that I had no good reason for separating out packages except for having \u201ctoo many files\u201d.<\/p><p id=\"bf9a\" class=\"graf graf--p graf-after--p\">Recently I\u2019ve found myself going the other direction\u200a\u2014\u200aonly using a single root package. Usually my project\u2019s types are all very related so it fits better from a usability and API standpoint. These types can also take advantage of calling unexported between them which keeps the API small and clear.<\/p><p id=\"4da3\" class=\"graf graf--p graf-after--p\">I found a few things helped me move toward larger packages:<\/p><ol class=\"postList\"><li id=\"aa05\" class=\"graf graf--li graf-after--p\">Group related types and code together in each file. If your types and functions are well organized then I find that files tend to be between 200 and 500 SLOC. This might sound like a lot but I find it easy to navigate. 1000 SLOC is usually my upper limit for a single file.<\/li><li id=\"f71b\" class=\"graf graf--li graf-after--li\">Organize the most important type at the top of the file and add types in decreasing importance towards the bottom of the file.<\/li><li id=\"4067\" class=\"graf graf--li graf-after--li\">Once your application starts getting above 10,000 SLOC you should seriously evaluate whether it can be broken into smaller projects.<\/li><\/ol><p id=\"9039\" class=\"graf graf--p graf-after--li\"><a href=\"https:\/\/github.com\/boltdb\/bolt\" data-href=\"https:\/\/github.com\/boltdb\/bolt\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Bolt<\/a> is a good example of this. Each file is a grouping of types related to a single Bolt construct:<\/p><pre id=\"9d48\" class=\"graf graf--pre graf-after--p\">bucket.go<br \/>cursor.go<br \/>db.go<br \/>freelist.go<br \/>node.go<br \/>page.go<br \/>tx.go<\/pre><h3 id=\"c5a6\" class=\"graf graf--h3 graf-after--pre\">Conclusion<\/h3><p id=\"638a\" class=\"graf graf--p graf-after--h3\">Code organization is one of the hardest parts about writing software and it rarely gets the focus it deserves. Use global variables sparingly, move your application binary code to its own package, wrap some types for application-specific context, and limit your subpackages. These are just a few tricks that can help make Go code easier and more maintainable.<\/p><p id=\"af65\" class=\"graf graf--p graf-after--p graf--trailing\">If you\u2019re writing Go projects the same way you write Ruby, Java, or Node.js projects then you\u2019re probably going to be fighting with the language.<\/p><\/div><\/div><\/section>","created_at":"2018-12-22T15:10:28+0000","updated_at":"2018-12-22T15:10:30+0000","published_at":"2014-07-06T20:55:12+0000","published_by":["Ben Johnson"],"annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*Eypi-5M7YH0cmIciE0M6lQ.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":753,"title":"Golang test driven development to check database queries involved methods","url":"https:\/\/stackoverflow.com\/questions\/24823307\/golang-test-driven-development-to-check-database-queries-involved-methods","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I want to create a database driven application using Golang. I am trying to do it TDD way.\nWhen I try to test methods that make Sql queries, What all are the packages available ?<\/p>\n<ul><li><p>I don't want to connect to the default database that I use for development. I can write code to take up another test database while running a test, but is there any go library that already does it.<\/p><\/li>\n<li><p>Is there any library that does db tests without connecting to database at all ?<\/p><\/li>\n<\/ul><p>What is the standard way to do database test with golang ? <\/p><\/div><div class=\"user-action-time\">&#13;\n        asked Jul 18 '14 at 10:52&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/132347\/sumit-m-asok\">Sumit M Asok<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,52462235&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"2\">&#13;\n                                2 Answers&#13;\n                                2&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I had a similar question not long ago when refactoring some of my own tests, and there's a couple of ways you can do it:<\/p>\n<p>a) Provide an exported type and an <code>Open<\/code> or <code>Connect<\/code> function that returns it - e.g.<\/p>\n<pre>type DB struct {\n    db *sql.DB\n}\n\/\/ Using http:\/\/jmoiron.github.io\/sqlx\/ for this example, but\n\/\/ it has the same interface as database\/sql\nfunc Open(opts *Options) (*DB, error) {\n    db, err := sqlx.Connect(opts.Driver, fmt.Sprintf(\"host=%s user=%s dbname=%s sslmode=%s\", opts.Host, opts.User, opts.Name, opts.SSL))\n    if err != nil {\n        return nil, err\n    }\n    return &amp;DB{db}, nil\n}\n<\/pre>\n<p>... and then <em>each of your<\/em> tests, write setup &amp; teardown functions that return an instance of <code>*DB<\/code> that you define your database functions on (as methods - i.e. <code>func (db *DB) GetUser(user *User) (bool, error)<\/code>):<\/p>\n<pre>\/\/ Setup the test environment.\nfunc setup() (*DB, error) {\n    err := withTestDB()\n    if err != nil {\n        return nil, err\n    }\n    \/\/ testOptions is a global in this case, but you could easily\n    \/\/ create one per-test\n    db, err := Open(testOptions)\n    if err != nil {\n        return nil, err\n    }\n    \/\/ Loads our test schema\n    db.MustLoad()\n    return db, nil\n}\n\/\/ Create our test database.\nfunc withTestDB() error {\n    db, err := open()\n    if err != nil {\n        return err\n    }\n    defer db.Close()\n    _, err = db.Exec(fmt.Sprintf(\"CREATE DATABASE %s;\", testOptions.Name))\n    if err != nil {\n        return err\n    }\n    return nil\n}\n<\/pre>\n<p>Note that this is somewhat \"integration\" testing, but I strongly prefer to test against a \"real\" database since mocking the interface won't help you catch issues with your queries\/query syntax.<\/p>\n<p>b) The alternative, although less extensible on the application side, is to have a global <code>db *sql.DB<\/code> variable that you initialise in <code>init()<\/code> within your tests\u2014since tests have no guaranteed order you'll need to use <code>init()<\/code>\u2014and then run your tests from there. i.e.<\/p>\n<pre>var db *sql.DB\nfunc init() {\n    var err error\n    \/\/ Note the = and *not* the assignment - we don't want to shadow our global\n    db, err = sqlx.Connect(...)\n    if err != nil {\n        ...\n    }\n    err := db.loadTestSchema\n    \/\/ etc.\n}\nfunc TestGetUser(t *testing.T) {\n   user := User{}\n   exists, err := db.GetUser(user)\n   ...\n}\n<\/pre>\n<p>You can find some practical examples in <a href=\"https:\/\/github.com\/drone\/drone\/tree\/master\/pkg\/database\/testing\">drone.io's GitHub repo<\/a>, and I'd also recommend <a href=\"https:\/\/medium.com\/@benbjohnson\/structuring-applications-in-go-3b04be4ff091\">this article on structuring Go applications<\/a> (especially the DB stuff).<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/24824118\/revisions\" title=\"show all edits to this post\">edited Mar 25 '15 at 20:16<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/576932\/ford\">ford<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            5,4162647&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jul 18 '14 at 11:36&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/556573\/elithrar\">elithrar<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            14.3k54681&#13;\n        <\/div>&#13;<\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I use a global variable to store the data source (or connection string) of current database and set to different value in test function. Since there is only one database I need to operate so I choose the easiest way.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jul 18 '14 at 11:13&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3828156\/chendesheng\">chendesheng<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,129611&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-12-22T04:39:43+0000","updated_at":"2018-12-22T04:39:43+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":752,"title":"Database Testing Patterns in Go","url":"https:\/\/www.clypd.com\/database-testing-patterns-in-go\/","content":"<img width=\"250\" height=\"191\" src=\"https:\/\/www.clypd.com\/wp-content\/uploads\/2014\/07\/testing1.png\" class=\"attachment-full size-full wp-post-image\" alt=\"\" title=\"\" \/><p>At clypd, we place a lot of value on testing as a mechanism for ensuring code correctness and our approach to testing constantly evolves. We\u2019ve spent the past few months learning and creating a new approach to mocking and testing functions that need to access data, whether from a database, a file, or over the network.<\/p><h3>Growing pains<\/h3><p>When we wrote our first lines of Go code in January 2014, our tests were exclusively unit tests. We wrote our processes so that we could ingest all the required data at the very beginning, operate on it, and output the result. The inputs and outputs were everything from files to databases to network calls \u2013 but it didn\u2019t matter, because we could test that separately from the actual logic. Most of our top level functions looked something like this:<\/p><p><code>func main() {<br \/>database := Open(\"dbname\")<br \/>data := GetData(database)<br \/>results := DoSomethingWith(data)<br \/>CommitResults(database, result)<br \/>}<\/code><br \/>Eventually the size of our data outgrew this pattern \u2014 we couldn\u2019t hold all the data we needed to operate in memory, and so we needed to gather new data and output results during the execution of the program. We took the simplest approach (which is also often the best) to solve this problem and attempted to deal with subsets of data at a time. \u00a0The function above turned into:<\/p><p><code>func main() {<br \/>database := Open(\"dbname\")<br \/>ids := GetIDs(database)<br \/>for _, id := range IDs {<br \/>data := GetDataFor(database, id)<br \/>result := DoSomethingWith(data)<br \/>CommitResult(database, result)<br \/>}<br \/>}<\/code><br \/>We\u2019d retrieve a slice of all the primary keys we wanted to operate on from the database at the start of execution. We\u2019d then take a subset of those keys and retrieve the associated information, consume it, and repeat. Database calls moved lower and lower into the application and consequently, our test coverage dropped. \u00a0Eventually, we moved to correct this.<\/p><h3>Mocking the database<\/h3><p>Our first approach was very simple: mock out the database. We wrapped our database library, with various methods like Select() and Exec() that accepted queries as arguments. These methods formed an interface for our database. We created an alternate implementation with a map for a backend that would match SQL queries up to responses to help with unit testing.<\/p><p>It solved our immediate problem and allowed us to provide data to our programs. By writing tests for functions that accessed the database, our coverage improved. \u00a0However, after a short time using this approach, we \u00a0grew increasingly dissatisfied. The tests just weren\u2019t very rigorous. Code coverage isn\u2019t necessarily a good measurement for test quality.<\/p><p>Some of the weaknesses of this approach worried us. First, we wanted some degree of input validation. While the ability to provide data to our programs improved, we couldn\u2019t validate the correctness of the output (what was inserted, updated, or deleted in the database). We also had some difficulty exactly replicating the behaviour of our database package.<\/p><p>Our database package uses reflection for much of its API, so our test solution had to make use of reflection in much the same way. We are fairly inexperienced with reflection, and tend to avoid it because of the additional complexity associated with generic solutions. There are exceptions, times when reflection has drastically reduced code duplication by enabling a generic solution.<\/p><p>We had some ideas to improve this implementation, but weren\u2019t very excited about them. There was the possibility of doing input validation with some ugly type assertions. While not inherently evil, type assertions add some burden of knowledge to the programmer and generally reduce the effectiveness of writing in a language with strict typing. To solve the bugs relating to an inexact implementation of the database library, we could limit the power of our database wrapper to the features. But if many on the team were already unhappy with the design, forcing additional features is not the right approach.<\/p><h3>Interfaces<\/h3><p>One of the interesting things about Go is that there is usually an idiomatic solution to a problem that is far cleaner and easier to implement than any other solution. Clearly, if we were already struggling this much, then this solution was not <i>the<\/i> solution. It was time to \u00a0scrap it and try again.<\/p><p>Our second approach relied heavily on Go\u2019s interfaces. We chose an approach based on interfaces because, at the very least, we\u2019d learn more about interfaces and become more comfortable with them. So we picked an isolated area of the codebase, a place that could be a prototype for the pattern, and came up with a structure that encapsulated our database calls. The implementation was a series of one-line functions that passed their data straight through to our database package. The example above becomes:<\/p><p><code>type datasource interface {<br \/>GetIDs()<br \/>GetDataFor(int)<br \/>CommitResult(resultType)<br \/>}<\/code><\/p><p>type db struct {<br \/>database<br \/>}<\/p><p>func (db *db) GetIDs() {<br \/>GetIDs(database)<br \/>}<\/p><p>func (db *db) GetDataFor(id int) {<br \/>GetDataFor(database, id)<br \/>}<\/p><p>func (db *db) CommitResult(result resultType) {<br \/>CommitResult(database, result)<br \/>}<\/p><p>func Run(db *db) {<br \/>ids := db.GetIDs()<br \/>for _, id := range IDs {<br \/>data := db.GetDataFor(id)<br \/>result := DoSomethingWith(data)<br \/>db.CommitResult(result)<br \/>}<br \/>}<\/p><p>func main() {<br \/>database := Open(\"dbname\")<br \/>Run(&amp;db{database})<br \/>}<\/p><p>We can now write a test for the function Run(). The test implementation of datasource is a structure that had two slices for each function in the interface, one to capture inputs, and one to provide outputs.\u00a0Each function definition, however, was only two lines \u2013 which surprised us. \u00a0One line to add the the inputs to a slice and one line to pop the outputs out of a slice. The new design was simpler, with no need to worry about the implementation of the database wrapper, and benefitted from Go\u2019s strict typing. It seemed like a good solution, but how did it scale?<\/p><p>As we rolled out the new pattern, the solution scaled without much pain. Many of the benefits can be attributed to Go\u2019s interfaces. We kept our interfaces function local. Each function that had to access data would have an interface associated with it, and those interfaces are combined as we climbed higher in the code.<\/p><p>The end result was one structure per package that handled all the production logic and several much smaller, more specialized structures for testing purposes. While this new pattern did introduce more test code than the previous approach, the additional boilerplate of setting up test structures was expected. Test code bloat is much less worrying than production code bloat, and the increased clarity is well worth the additional cost. There is some extra production code, but it has the nice side effect of concentrating all of our external calls in one place.<\/p><p>With this new pattern in place, other programmers can easily see, at a glance, all the information that a program will need to run. Also, this approach has the nice side effect of being able to run much of the same code from any data source. We can run the same processes from a CSV and all that\u2019s needed is another implementation of the data interface.<\/p><p>This new pattern suits us, for now. It\u2019s a clean solution that avoids reflection entirely. We don\u2019t have to worry about the implementation details of how we get our data and it allows us to provide and test functions that gather data and output data.<\/p><p>While tests can always be improved, this pattern has now been rolled out to the entire codebase as a means of testing data-access calls and we\u2019ve witnessed the power of Go\u2019s interfaces firsthand.<\/p>","created_at":"2018-12-22T04:36:23+0000","updated_at":"2018-12-22T04:36:23+0000","published_by":["Joey Goode"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"www.clypd.com","preview_picture":"https:\/\/www.clypd.com\/wp-content\/uploads\/2014\/07\/testing1.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["json-ld"],"is_public":false,"id":751,"title":"digitalbazaar\/jsonld-signatures","url":"https:\/\/github.com\/digitalbazaar\/jsonld-signatures","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/digitalbazaar\/jsonld-signatures\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f9131d1778d7b87353c82579435eff26143e2b05\/68747470733a2f2f7472617669732d63692e6f72672f6469676974616c62617a6161722f6a736f6e6c642d7369676e6174757265732e706e673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/digitalbazaar\/jsonld-signatures.png?branch=master\" \/><\/a><\/p>\n<p>An implementation of the Linked Data Signatures specification for JSON-LD.\nThis software works in all modern browsers as well as node.js via <a href=\"https:\/\/www.npmjs.com\/package\/jsonld-signatures\" rel=\"nofollow\">npm<\/a>.<\/p>\n<h2><a id=\"user-content-introduction\" class=\"anchor\" aria-hidden=\"true\" href=\"#introduction\"><\/a>Introduction<\/h2>\n<p>Install with npm:<\/p>\n<pre>npm install jsonld-signatures\n<\/pre>\n<p>In Node.js, include the library like this:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var jsonld = require('jsonld');\nvar jsig = require('jsonld-signatures');\njsig.use('jsonld', jsonld);<\/pre><\/div>\n<p>In a browser environment, include <code>jsonld<\/code>, <code>forge<\/code>, and\n<code>dist\/jsonld-signatures.min.js<\/code> via script tag or other mechanism.<\/p>\n<h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ to generate the next two lines, run the following command:\n\/\/\n\/\/ openssl genrsa -out key.pem; cat key.pem; openssl rsa -in key.pem -pubout -out pubkey.pem; cat pubkey.pem; rm key.pem pubkey.pem\n\/\/\n\/\/ for an example of how to specify these keys, look at [key-example]:\nvar testPublicKeyPem = \"-----BEGIN PUBLIC KEY-----\\r\\n...\";\nvar testPrivateKeyPem = \"-----BEGIN PRIVATE KEY-----\\r\\n...\";\n\/\/ specify the public key object\nvar testPublicKey = {\n  '@context': jsig.SECURITY_CONTEXT_URL,\n  '@id': 'https:\/\/example.com\/i\/alice\/keys\/1',\n  owner: 'https:\/\/example.com\/i\/alice',\n  publicKeyPem: testPublicKeyPem\n};\n\/\/ specify the public key owner object\nvar testPublicKeyOwner = {\n  \"@context\": jsig.SECURITY_CONTEXT_URL,\n  '@id': 'https:\/\/example.com\/i\/alice',\n  publicKey: [testPublicKey]\n};\n\/\/ create the JSON-LD document that should be signed\nvar testDocument = {\n  \"@context\": {\n    schema: 'http:\/\/schema.org\/',\n    name: 'schema:name',\n    homepage: 'schema:url',\n    image: 'schema:image'\n  },\n  name: 'Manu Sporny',\n  homepage: 'https:\/\/manu.sporny.org\/',\n  image: 'https:\/\/manu.sporny.org\/images\/manu.png'\n};\n\/\/ sign the document and then verify the signed document\njsig.sign(testDocument, {\n  privateKeyPem: testPrivateKeyPem,\n  creator: 'https:\/\/example.com\/i\/alice\/keys\/1'\n}, function(err, signedDocument) {\n  if(err) {\n    return console.log('Signing error:', err);\n  }\n  console.log('Signed document:', signedDocument);\n  \/\/ verify the signed document\n  jsig.verify(signedDocument, {\n    publicKey: testPublicKey,\n    publicKeyOwner: testPublicKeyOwner,\n  }, function(err, verified) {\n    if(err) {\n      return console.log('Signature verification error:', err);\n    }\n    console.log('Signature is valid:', verified);\n  });\n});\n\/\/ verification\nvar sign = jsig.promises.sign(testDocument, {\n  privateKeyPem: testPrivateKeyPem,\n  creator: 'https:\/\/example.com\/i\/alice\/keys\/1'\n});\nsign.then(function(signedDocument) {...}, function(err) {...});\nvar verify = jsig.promises.verify(signedDocument, {\n  publicKey: testPublicKey,\n  publicKeyOwner: testPublicKeyOwner\n});\nverify.then(function(verified) {...}, function(err) {...});<\/pre><\/div>\n<h2><a id=\"user-content-commercial-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#commercial-support\"><\/a>Commercial Support<\/h2>\n<p>Commercial support for this library is available upon request from\nDigital Bazaar: <a href=\"mailto:support@digitalbazaar.com\">support@digitalbazaar.com<\/a><\/p>\n<h2><a id=\"user-content-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#source\"><\/a>Source<\/h2>\n<p>The source code for the JavaScript implementation of the JSON-LD Signatures API\nis available at:<\/p>\n<p><a href=\"https:\/\/github.com\/digitalbazaar\/jsonld-signatures\">https:\/\/github.com\/digitalbazaar\/jsonld-signatures<\/a><\/p>\n<h2><a id=\"user-content-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#tests\"><\/a>Tests<\/h2>\n<p>This library includes a sample testing utility which may be used to verify\nthat changes to the processor maintain the correct output.<\/p>\n<p>To run the sample tests you will need to get the test suite files by cloning\nthe <a href=\"https:\/\/github.com\/digitalbazaar\/jsonld-signatures\/\">jsonld-signatures repository<\/a> hosted on GitHub.<\/p>\n<p><a href=\"https:\/\/github.com\/digitalbazaar\/jsonld-signatures\/\">https:\/\/github.com\/digitalbazaar\/jsonld-signatures\/<\/a><\/p>\n<p>Run the Node.js tests using the following command:<\/p>\n<pre>npm run test\n<\/pre>\n<p>Run browser tests using PhantomJS using the following command:<\/p>\n<pre>npm run test-karma\n<\/pre>\n<p>Run browser tests using a selection of browsers using the following command:<\/p>\n<pre>npm run test-karma -- --browsers Firefox,Chrome,PhantomJS\n<\/pre>\n<p>Code coverage of node tests can be generated in <code>coverage\/<\/code>:<\/p>\n<pre>npm run coverage\n<\/pre>\n<\/article>","created_at":"2018-12-21T14:36:04+0000","updated_at":"2018-12-21T14:36:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/167436?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["timeseries"],"is_public":false,"id":750,"title":"What the heck is time-series data (and why do I need a time-series database)?","url":"https:\/\/blog.timescale.com\/what-the-heck-is-time-series-data-and-why-do-i-need-a-time-series-database-dcf3b1b18563?_branch_match_id=604498028704843480&gi=23479871a5ab","content":"<section class=\"section section--body section--first\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><figure id=\"c37b\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*mWcVDjr-cp0cHZYOhqhIXA.png\" data-width=\"1365\" data-height=\"768\" data-is-featured=\"true\" data-action=\"zoom\" data-action-value=\"1*mWcVDjr-cp0cHZYOhqhIXA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*mWcVDjr-cp0cHZYOhqhIXA.png\" alt=\"image\" \/><\/div><\/figure><p id=\"fc1e\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--figure\">Here\u2019s a riddle: what do self-driving Teslas, autonomous Wall Street trading algorithms, smart homes, transportation networks that fulfill lightning-fast same-day deliveries, and an open-data-publishing NYPD have in common?<\/p><p id=\"03a8\" class=\"graf graf--p graf-after--p\">For one, they are signs that our world is changing at warp speed, thanks to our ability to capture and analyze more and more data in faster and faster ways than before.<\/p><p id=\"26cd\" class=\"graf graf--p graf-after--p\">But if you look closely, you\u2019ll notice that each of these applications requires a special kind of data:<\/p><ul class=\"postList\"><li id=\"b73e\" class=\"graf graf--li graf-after--p\">Self-driving cars continuously collect data about how their local environment is changing around them.<\/li><li id=\"07c4\" class=\"graf graf--li graf-after--li\">Autonomous trading algorithms continuously collect data on how the markets are changing.<\/li><li id=\"f754\" class=\"graf graf--li graf-after--li\">Our smart homes monitor what\u2019s going on inside of them to regulate temperature, identify intruders, and respond to our beck-and-call (\u201cAlexa, play some relaxing music\u201d).<\/li><li id=\"a5fb\" class=\"graf graf--li graf-after--li\">Our retail industry monitors how their assets are moving with such precision and efficiency that cheap same-day delivery is a luxury that many of us take for granted.<\/li><li id=\"1bd8\" class=\"graf graf--li graf-after--li\">The NYPD tracks its vehicles to allow us to hold them more accountable (e.g., for analyzing <a href=\"http:\/\/www.nyc.gov\/html\/911reporting\/html\/reports\/end-to-end.shtml\" data-href=\"http:\/\/www.nyc.gov\/html\/911reporting\/html\/reports\/end-to-end.shtml\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">911 response times<\/a>).<\/li><\/ul><p id=\"5773\" class=\"graf graf--p graf-after--li\">These applications rely on a form of data that measures how things change over time. Where time isn\u2019t just a metric, but a primary axis.<\/p><p id=\"9cc0\" class=\"graf graf--p graf-after--p\">This is time-series data. And it\u2019s starting to play a larger role in our world.<\/p><p id=\"1459\" class=\"graf graf--p graf-after--p\">Software developer usage patterns already reflect this. In the past 24 months, time-series databases (TSDBs) have emerged as the fastest growing category of databases:<\/p><figure id=\"38dd\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*l4ZYKn3RxYKCxWOPao_-uQ.png\" data-width=\"1229\" data-height=\"634\" data-action=\"zoom\" data-action-value=\"1*l4ZYKn3RxYKCxWOPao_-uQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*l4ZYKn3RxYKCxWOPao_-uQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><strong class=\"markup--strong markup--figure-strong\"><em class=\"markup--em markup--figure-em\">Source: DB-Engines, June 2017. <\/em><\/strong><a href=\"https:\/\/db-engines.com\/en\/ranking_categories\" data-href=\"https:\/\/db-engines.com\/en\/ranking_categories\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--figure-strong\"><em class=\"markup--em markup--figure-em\">https:\/\/db-engines.com\/en\/ranking_categories<\/em><\/strong><\/a><\/figcaption><\/figure><p id=\"f29f\" class=\"graf graf--p graf-after--figure\">As the developers of a new, open source time-series database, we are often asked about this trend. Typically, we get these three questions:<\/p><ol class=\"postList\"><li id=\"3868\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">What is time-series data?<\/strong><\/li><li id=\"cff0\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">When would I need a time-series database?<\/strong><\/li><li id=\"6dd2\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Why should I use (or not use) TimescaleDB?<\/strong><\/li><\/ol><p id=\"000e\" class=\"graf graf--p graf-after--li\">We initially developed this material for a <a href=\"https:\/\/www.youtube.com\/watch?v=7hxXU9dceaE\" data-href=\"https:\/\/www.youtube.com\/watch?v=7hxXU9dceaE\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">talk we gave at Percona Live in April<\/a>. We\u2019re now publishing it in this post to (try to) thoughtfully answer these questions in detail and share these answers with as many people as possible.<\/p><p id=\"d5c8\" class=\"graf graf--p graf-after--p graf--trailing\">Let\u2019s start.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"829c\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">What is time-series data?<\/strong><\/h3><p id=\"3d28\" class=\"graf graf--p graf-after--h3\">Some think of \u201ctime-series data\u201d as a sequence of data points, measuring the same thing over time, stored in time order. That\u2019s true, but it just scratches the surface.<\/p><p id=\"7523\" class=\"graf graf--p graf-after--p\">Others may think of a series of numeric values, each paired with a timestamp, defined by a name and a set of labeled dimensions (or \u201ctags\u201d). This is perhaps one way to model time-series data, but not a definition of the data itself.<\/p><p id=\"5e01\" class=\"graf graf--p graf-after--p\">Let\u2019s go a little deeper.<\/p><p id=\"6e84\" class=\"graf graf--p graf-after--p\">Here\u2019s a basic illustration. Imagine sensors collecting data from three settings: a city, farm, and factory. In this example, each of these sources periodically sends new readings, creating a series of measurements collected over time.<\/p><figure id=\"2a97\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*GIOYjAyTjAaOK7HgAdAdVA.gif\" data-width=\"562\" data-height=\"481\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*GIOYjAyTjAaOK7HgAdAdVA.gif\" alt=\"image\" \/><\/div><\/figure><p id=\"a366\" class=\"graf graf--p graf-after--figure\">Here\u2019s another example, with real data from the City of New York, showing every taxicab ride for the first few seconds of 2016. As you can see, each row is a \u201cmeasurement\u201d collected at a specific time:<\/p><figure id=\"e23e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*FmoscB-nsG0QvlDX.\" data-width=\"640\" data-height=\"360\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*FmoscB-nsG0QvlDX.\" alt=\"image\" \/><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"8e14\" class=\"graf graf--p graf-after--figure\">There are many other kinds of time-series data. To name a few: <strong class=\"markup--strong markup--p-strong\">DevOps monitoring data, mobile\/web application event streams, industrial machine data, scientific measurements.<\/strong><\/p><p id=\"1c3a\" class=\"graf graf--p graf-after--p\">These datasets primarily have 3 things in common:<\/p><ol class=\"postList\"><li id=\"d6b9\" class=\"graf graf--li graf-after--p\">The data that arrives is almost always recorded as a new entry<\/li><li id=\"7c0c\" class=\"graf graf--li graf-after--li\">The data typically arrives in time order<\/li><li id=\"7f81\" class=\"graf graf--li graf-after--li\">Time is a primary axis (time-intervals can be either regular or irregular)<\/li><\/ol><p id=\"034a\" class=\"graf graf--p graf-after--li\">In other words, time-series data workloads are generally \u201cappend-only.\u201d While they may need to correct erroneous data after the fact, or handle delayed or out-of-order data, these are exceptions, not the norm.<\/p><p id=\"0049\" class=\"graf graf--p graf-after--p\">You may ask:<em class=\"markup--em markup--p-em\"> How is this different than just having a time-field in a dataset?<\/em><\/p><p id=\"115a\" class=\"graf graf--p graf-after--p\">Well, it depends: how does your dataset track changes? By updating the current entry, or by inserting a new one?<\/p><p id=\"ce34\" class=\"graf graf--p graf-after--p\">When you collect a new reading for sensor_x, do you overwrite your previous reading, or do you create a brand new reading in a separate row? While both methods will provide you the current state of the system, only by writing the new reading in a separate row will you be able to track all states of the system over time.<\/p><p id=\"972b\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Simply put: time-series datasets track changes to the overall system as INSERTs, not UPDATEs.<\/strong><\/p><p id=\"b761\" class=\"graf graf--p graf-after--p\">This practice of recording each and every change to the system as a new, different row is what makes time-series data so powerful. It allows us to <em class=\"markup--em markup--p-em\">measure change<\/em>: analyze how something <em class=\"markup--em markup--p-em\">changed<\/em> in the past, monitor how something <em class=\"markup--em markup--p-em\">is changing<\/em> in the present, predict how it <em class=\"markup--em markup--p-em\">may change<\/em> in the future.<\/p><p id=\"5d53\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">And so, here\u2019s how we define time-series data: data that collectively represents how a system\/process\/behavior changes over time.<\/strong><\/p><p id=\"9fb5\" class=\"graf graf--p graf-after--p\">This is more than just an academic distinction: by centering our definition around \u201cchange\u201d, we can start to identify time-series datasets that we aren\u2019t collecting today, but that we <em class=\"markup--em markup--p-em\">should<\/em> be collecting.<\/p><p id=\"8867\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">In fact, what we\u2019ve found is that often people have time-series data but don\u2019t realize it.<\/strong><\/p><figure id=\"4269\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*edtKtsbt403tuoo_1UK3iw.png\" data-width=\"710\" data-height=\"532\" data-action=\"zoom\" data-action-value=\"1*edtKtsbt403tuoo_1UK3iw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*edtKtsbt403tuoo_1UK3iw.png\" alt=\"image\" \/><\/div><\/figure><p id=\"2b30\" class=\"graf graf--p graf-after--figure\">Imagine you maintain a web application. Every time a user logs in, you may just update a \u201clast_login\u201d timestamp for that user in a single row in your \u201cusers\u201d table. But what if you treated each login as a separate event, and collected them over time? Then you could: track historical login activity, see how usage is (in-\/de-)creasing over time, bucket users by how often they access the app, and more.<\/p><p id=\"a53e\" class=\"graf graf--p graf-after--p\">This example illustrates a key point: by preserving the inherent time-series nature of our data, we are able to preserve valuable information on how that data changes over time.<\/p><p id=\"2b2d\" class=\"graf graf--p graf-after--p\">(In fact, this example illustrates another point: event data is also time-series data.)<\/p><p id=\"4c98\" class=\"graf graf--p graf-after--p\">Of course, storing data at this resolution comes with an obvious problem: you end up with a lot of data, rather fast. <strong class=\"markup--strong markup--p-strong\">So that\u2019s the catch: time-series data piles up very quickly.<\/strong><\/p><p id=\"3733\" class=\"graf graf--p graf-after--p\">Having a lot of data creates problems when both recording it and querying it in a performant way.<\/p><p id=\"e1b7\" class=\"graf graf--p graf-after--p graf--trailing\">Which is why people are now turning to time-series databases.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"5565\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">Why do I need a time-series database?<\/strong><\/h3><p id=\"a429\" class=\"graf graf--p graf-after--h3\">You might ask: <em class=\"markup--em markup--p-em\">Why can\u2019t I just use a \u201cnormal\u201d (i.e., non-time-series) database?<\/em><\/p><p id=\"c01e\" class=\"graf graf--p graf-after--p\">The truth is that you can, and some people do:<\/p><figure id=\"9c3a\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*gDkrxs6ikLZuofhg.\" data-width=\"825\" data-height=\"461\" data-action=\"zoom\" data-action-value=\"0*gDkrxs6ikLZuofhg.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*gDkrxs6ikLZuofhg.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><strong class=\"markup--strong markup--figure-strong\"><em class=\"markup--em markup--figure-em\">Source: Percona, February 2017. <\/em><\/strong><a href=\"https:\/\/www.percona.com\/blog\/2017\/02\/10\/percona-blog-poll-database-engine-using-store-time-series-data\/\" data-href=\"https:\/\/www.percona.com\/blog\/2017\/02\/10\/percona-blog-poll-database-engine-using-store-time-series-data\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--figure-strong\"><em class=\"markup--em markup--figure-em\">https:\/\/www.percona.com\/blog\/2017\/02\/10\/percona-blog-poll-database-engine-using-store-time-series-data\/<\/em><\/strong><\/a><\/figcaption><\/figure><p id=\"fd4a\" class=\"graf graf--p graf-after--figure\">Yet why do the majority of respondents to that survey use a time-series database instead of a normal db? And why are TSDBs the fastest growing category of databases today?<\/p><p id=\"6d36\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Two reasons: (1) scale and (2) usability.<\/strong><\/p><p id=\"94ca\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Scale: <\/strong>Time-series data accumulates very quickly. (For example, a single connected car will collect <a href=\"https:\/\/qz.com\/344466\/connected-cars-will-send-25-gigabytes-of-data-to-the-cloud-every-hour\/\" data-href=\"https:\/\/qz.com\/344466\/connected-cars-will-send-25-gigabytes-of-data-to-the-cloud-every-hour\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">25GB of data per hour<\/a>.) And normal databases are not designed to handle that scale: relational databases fare poorly with very large datasets; NoSQL databases fare better at scale, but can still be outperformed by a database fine-tuned for time-series data. In contrast, time-series databases (which can be based on relational or NoSQL databases) handle scale by introducing efficiencies that are only possible when you treat time as a first class citizen. These efficiencies result in performance improvements, including: higher ingest rates, faster queries at scale (although some support more queries than others), and better data compression.<\/p><p id=\"881d\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Usability: <\/strong>TSDBs also typically include functions and operations common to time-series data analysis: data retention policies, continuous queries, flexible time aggregations, etc. Even if scale it not a concern at the moment (e.g., if you are just starting to collect data), these features can still provide a better user experience and make your life easier.<\/p><p id=\"a7d1\" class=\"graf graf--p graf-after--p\">This is why developers are increasingly adopting time-series databases and using them for a variety of use cases:<\/p><ul class=\"postList\"><li id=\"e21e\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Monitoring software systems:<\/strong> Virtual machines, containers, services, applications<\/li><li id=\"5366\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Monitoring physical systems: <\/strong>Equipment, machinery, connected devices, the environment, our homes, our bodies<\/li><li id=\"b724\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Asset tracking applications:<\/strong> Vehicles, trucks, physical containers, pallets<\/li><li id=\"abfa\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Financial trading systems:<\/strong> Classic securities, newer cryptocurrencies<\/li><li id=\"b7e4\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Eventing applications: <\/strong>Tracking user\/customer interaction data<\/li><li id=\"c89f\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Business intelligence tools:<\/strong> Tracking key metrics and the overall health of the business<\/li><li id=\"c685\" class=\"graf graf--li graf-after--li\">(and more)<\/li><\/ul><p id=\"2a9a\" class=\"graf graf--p graf-after--li graf--trailing\">But even then, you\u2019ll need to pick a time-series database that best fits your data model and write\/read patterns.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"9bfd\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">Why should I use (or not use) TimescaleDB?<\/strong><\/h3><p id=\"8cb5\" class=\"graf graf--p graf-after--h3\">If you do need a time-series database, there are already <a href=\"https:\/\/misfra.me\/2016\/04\/09\/tsdb-list\/\" data-href=\"https:\/\/misfra.me\/2016\/04\/09\/tsdb-list\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">quite a lot of options<\/a> out there. You may be happy with one of them. But we weren\u2019t.<\/p><p id=\"3699\" class=\"graf graf--p graf-after--p\">Why weren\u2019t we satisfied with the state-of-the-art? Because we wanted the power of full SQL at scale, which none of the existing options provided.<\/p><p id=\"c25e\" class=\"graf graf--p graf-after--p\">Specifically, we found that existing time-series databases:<\/p><ul class=\"postList\"><li id=\"3758\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Underperformed on many of our queries<\/strong> (read: high latencies)<\/li><li id=\"741b\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Would not even support many other queries<\/strong> (varies by database)<\/li><li id=\"ea76\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Required us to learn a new query language<\/strong> (read: not SQL)<\/li><li id=\"e589\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Would not work with most of our existing tools<\/strong> (read: poor compatibility)<\/li><li id=\"0a68\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Required us to silo our data into two databases<\/strong>: a \u201cnormal\u201d relational one, and a second time-series one (read: operational and development headaches)<\/li><\/ul><p id=\"4053\" class=\"graf graf--p graf-after--li\">So we built TimescaleDB because we needed it. And then other people wanted to use it, <a href=\"https:\/\/blog.timescale.com\/when-boring-is-awesome-building-a-scalable-time-series-database-on-postgresql-2900ea453ee2\" data-href=\"https:\/\/blog.timescale.com\/when-boring-is-awesome-building-a-scalable-time-series-database-on-postgresql-2900ea453ee2\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">so earlier this year we open sourced it<\/a> under an Apache 2 license.<\/p><figure id=\"ef1e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*NQUMsEfQKXj2ScvsVVSoKg.jpeg\" data-width=\"406\" data-height=\"374\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*NQUMsEfQKXj2ScvsVVSoKg.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><strong class=\"markup--strong markup--figure-strong\">TimescaleDB: SQL at scale for time-series data<\/strong><\/figcaption><\/figure><p id=\"d2df\" class=\"graf graf--p graf-after--figure\">When should you consider TimescaleDB? If you want:<\/p><ol class=\"postList\"><li id=\"93f5\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">A normal SQL interface <\/strong>to time-series data, even at scale<\/li><li id=\"02d2\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Operational simplicity<\/strong>: one database for your relational and time-series data<\/li><li id=\"a9f3\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">JOINs<\/strong> across relational and time-series data at query time<\/li><li id=\"584e\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">PostgreSQL!<\/strong> (Timescale looks and acts just like PostgreSQL)<\/li><li id=\"b776\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Query performance<\/strong>, especially for a broad, varied set of queries (via robust secondary index support)<\/li><li id=\"cb13\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Native support for geospatial data<\/strong> (via PostGIS compatibility)<\/li><li id=\"ac55\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Third-party tools: <\/strong>Timescale supports anything that speaks SQL, including BI tools like Tableau<\/li><\/ol><p id=\"22ad\" class=\"graf graf--p graf-after--li\">Then again, if any of the following is true, you might not want to use TimescaleDB:<\/p><ul class=\"postList\"><li id=\"c636\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">If you have only simple query patterns<\/strong> (e.g., key-value lookups, one dimensional rollups). Other database options are more optimized for these types of queries.<\/li><li id=\"3a51\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">If you have sparse and\/or completely unstructured data.<\/strong> While TimescaleDB supports structured and semi-structured data (including JSON), if your data is typically sparse or completely unstructured, there are other better database options.<\/li><\/ul><p id=\"ca7e\" class=\"graf graf--p graf-after--li graf--trailing\">But if TimescaleDB does seem interesting\/useful, we welcome you to <a href=\"http:\/\/www.timescale.com\/\" data-href=\"http:\/\/www.timescale.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">learn more<\/a>, download\/install it, and give it a whirl. (Available on <a href=\"https:\/\/github.com\/timescale\/timescaledb\/\" data-href=\"https:\/\/github.com\/timescale\/timescaledb\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">GitHub<\/a>. Stars appreciated.)<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"a3f2\" class=\"graf graf--h3 graf--leading\"><strong class=\"markup--strong markup--h3-strong\">A parting thought: Is all data time-series data?<\/strong><\/h3><p id=\"a39c\" class=\"graf graf--p graf-after--h3\">Let\u2019s return to the topic of time-series data for one parting thought.<\/p><p id=\"2b0b\" class=\"graf graf--p graf-after--p\">For the past decade or so, we have lived in the era of \u201cBig Data\u201d, collecting massive amounts of information about our world and applying computational resources to make sense of it. Even though this era started with modest computing technology, our ability to capture, store, and analyze data has improved at an exponential pace, thanks to major macro-trends: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Moore%27s_law\" data-href=\"https:\/\/en.wikipedia.org\/wiki\/Moore%27s_law\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Moore\u2019s law<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mark_Kryder\" data-href=\"https:\/\/en.wikipedia.org\/wiki\/Mark_Kryder\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Kryder\u2019s law<\/a>, cloud computing, an entire industry of \u201cbig data\u201d technologies.<\/p><figure id=\"9ccf\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*_Ozk54h8SH0Z7Cu9c1EeXw.png\" data-width=\"1324\" data-height=\"619\" data-action=\"zoom\" data-action-value=\"1*_Ozk54h8SH0Z7Cu9c1EeXw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*_Ozk54h8SH0Z7Cu9c1EeXw.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><strong class=\"markup--strong markup--figure-strong\"><em class=\"markup--em markup--figure-em\">Under Moore\u2019s Law, computational power (transistor density) doubles every 18 months, while Kryder\u2019s Law postulates that storage capacity doubles every 12\u00a0months.<\/em><\/strong><\/figcaption><\/figure><p id=\"4dd0\" class=\"graf graf--p graf-after--figure\">Now we need more. No longer content to just observe the state of the world, we now want to measure how our world changes over time, down to sub-second intervals. Our \u201cbig data\u201d datasets are now being dwarfed by another type of data, one that relies heavily on time to preserve information about the change that is happening.<\/p><p id=\"7553\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">But does all data start off as time-series data?<\/strong> Recall our earlier web application example: we had time-series data but didn\u2019t realize it.<\/p><p id=\"5608\" class=\"graf graf--p graf-after--p\">Or think of any \u201cnormal\u201d dataset. Say, the current accounts and balances at a major retail bank. Or the source code for a software project. Or the text for this blog post.<\/p><p id=\"8eb3\" class=\"graf graf--p graf-after--p\">Typically we choose to store the latest state of the system. But what if we instead stored every change, and computed the latest state at query time? Isn\u2019t a \u201cnormal\u201d dataset just a view on top of an inherently time-series dataset (cached for performance reasons)? Don\u2019t banks have transaction ledgers? (And aren\u2019t blockchains just distributed, immutable time-series logs?) Wouldn\u2019t a software project have version control (e.g., git commits)? Doesn\u2019t this blog post have revision history? (Undo. Redo.)<\/p><p id=\"11f2\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">Put differently: Don\u2019t all databases have logs?<\/em><\/p><p id=\"d232\" class=\"graf graf--p graf-after--p\">We recognize that many applications may never require time-series data (and would be better served by a \u201ccurrent-state view\u201d). But as we continue along the exponential curve of technological progress, it would seem that these \u201ccurrent-state views\u201d become less necessary. <strong class=\"markup--strong markup--p-strong\">And that by storing more and more data in its time-series form, we may be able to understand it better.<\/strong><\/p><p id=\"1f70\" class=\"graf graf--p graf-after--p\">Is all data time-series data? We\u2019ve yet to find a good counter example. If you\u2019ve got one, we\u2019re open to hearing it.<\/p><p id=\"39de\" class=\"graf graf--p graf-after--p graf--trailing\">Regardless, one thing is clear: time-series data already surrounds us. It\u2019s time we put it to use.<\/p><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"a070\" class=\"graf graf--p graf--leading\"><em class=\"markup--em markup--p-em\">Thank you for reading our post. If you found it helpful, we\u2019d appreciate a recommend or share.<\/em><\/p><p id=\"c866\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">If you have any follow up questions or comments, we welcome them via <\/em><a href=\"mailto:support@timescale.com\" data-href=\"mailto:support@timescale.com\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\"><em class=\"markup--em markup--p-em\">email<\/em><\/a><em class=\"markup--em markup--p-em\"> or <\/em><a href=\"https:\/\/twitter.com\/timescaledb\" data-href=\"https:\/\/twitter.com\/timescaledb\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Twitter<\/em><\/a><em class=\"markup--em markup--p-em\">.<\/em><\/p><p id=\"6860\" class=\"graf graf--p graf-after--p graf--trailing\"><em class=\"markup--em markup--p-em\">And if you\u2019d like to learn more about TimescaleDB, please check out our <\/em><a href=\"https:\/\/github.com\/timescale\/timescaledb\" data-href=\"https:\/\/github.com\/timescale\/timescaledb\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">GitHub<\/em><\/a><em class=\"markup--em markup--p-em\">, and let us know how we can help.<\/em><\/p><\/div><\/div><\/section>","created_at":"2018-12-21T03:09:48+0000","updated_at":"2018-12-21T22:15:56+0000","published_at":"2017-06-12T19:05:43+0000","published_by":["Ajay Kulkarni"],"annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"blog.timescale.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*mWcVDjr-cp0cHZYOhqhIXA.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["lisp","cl"],"is_public":false,"id":749,"title":"clasp-developers\/clasp","url":"https:\/\/github.com\/clasp-developers\/clasp","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<h2><a id=\"user-content-note--november-28-2018---this-contains-a-pre-release-of-clasp-09\" class=\"anchor\" aria-hidden=\"true\" href=\"#note--november-28-2018---this-contains-a-pre-release-of-clasp-09\"><\/a>NOTE:  November 28, 2018 - This contains a pre-release of Clasp 0.9.<\/h2>\n<p><a href=\"https:\/\/www.youtube.com\/watch?v=mbdXeRBbgDM&amp;feature=youtu.be\" rel=\"nofollow\">We have a new talk up on youtube!<\/a><\/p>\n<h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2>\n<p>Clasp is a new <a href=\"https:\/\/common-lisp.net\/\" rel=\"nofollow\">Common Lisp<\/a> implementation that seamlessly interoperates with C++ libraries and programs using <a href=\"http:\/\/llvm.org\/\" rel=\"nofollow\">LLVM<\/a> for compilation to native code. This allows Clasp to take advantage of a vast array of preexisting libraries and programs, such as out of the scientific computing ecosystem. Embedding them in a Common Lisp environment allows you to make use of rapid prototyping, incremental development, and other capabilities that make it a powerful language.<\/p>\n<h2><a id=\"user-content-latest-release\" class=\"anchor\" aria-hidden=\"true\" href=\"#latest-release\"><\/a>Latest Release<\/h2>\n<p>The upcoming release is <a href=\"https:\/\/github.com\/clasp-developers\/clasp\/releases\/tag\/0.9.0\">Clasp 0.9<\/a>. <a href=\"https:\/\/github.com\/clasp-developers\/clasp\/milestone\/3\">Changes<\/a><\/p>\n<p>See the <a href=\"https:\/\/github.com\/clasp-developers\/clasp\/blob\/master\/RELEASE-NOTES\">RELEASE-NOTES<\/a><\/p>\n<h3><a id=\"user-content-building-clasp\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-clasp\"><\/a>Building Clasp<\/h3>\n<p>At the moment, Clasp is supported on Linux and Mac OS X. On these systems, you should be able to build it from source.<\/p>\n<p><a href=\"https:\/\/github.com\/clasp-developers\/clasp\/wiki\/Build-Instructions\">Follow the instructions.<\/a><\/p>\n<p>In case things go wrong, the quickest way to get help is to either <a href=\"#reporting-problems\">file an issue<\/a>, or to <a href=\"#irc\">chat with us directly<\/a>.<\/p>\n<p>Building on most systems will take around 8-12GB of memory and ~1-2 hours with a relatively modern processor, so be prepared to watch a movie or do some other useful work until Clasp is all done. You may also want to activate a large enough swap file if your system does not have enough DRAM.<\/p>\n<p>Building can be greately accelerated if you have a multi-core machine - but it requires a specially configured version of the Boehm garbage collector found here: (<a href=\"https:\/\/github.com\/clasp-developers\/forkable-boehm.git\">https:\/\/github.com\/clasp-developers\/forkable-boehm.git<\/a>)<\/p>\n<p>Currently there are no binary releases available.<\/p>\n<p>There is a docker image for a superset of Clasp called <a href=\"https:\/\/hub.docker.com\/r\/drmeister\/cando\/\" rel=\"nofollow\">Cando<\/a><\/p>\n<h3><a id=\"user-content-common-lisp-ecosystem-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#common-lisp-ecosystem-support\"><\/a>Common Lisp Ecosystem Support<\/h3>\n<p>Clasp supports the following major components:<\/p>\n<p>Post on the issues or <a href=\"#irc\">contact us<\/a> if you're interested in changing that.<\/p>\n<h2><a id=\"user-content-contributing-to-clasp\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing-to-clasp\"><\/a>Contributing to Clasp<\/h2>\n<p>We very much welcome any kind of contribution to Clasp, even if it is just bug finding and testing. A lot can be done all around the project, if you want to dive into something large. See the <a href=\"https:\/\/github.com\/clasp-developers\/clasp\/blob\/dev\/CONTRIBUTING.md\">CONTRIBUTING<\/a> file for the few guidelines we've set up around contributions.<\/p>\n<h2><a id=\"user-content-reporting-problems\" class=\"anchor\" aria-hidden=\"true\" href=\"#reporting-problems\"><\/a>Reporting Problems<\/h2>\n<p>Generally you can report problems in two fashions, either by <a href=\"https:\/\/github.com\/clasp-developers\/clasp\/issues\/new\">opening an issue ticket<\/a> or by <a href=\"#irc\">chatting with us directly<\/a>. In both cases, though, you should have the following pieces handy in order for us to be able to help you out as quickly and painlessly as possible:<\/p>\n<ul><li>Your operating system name and version.<\/li>\n<li>The branches that you're using of Clasp and Externals-Clasp.<\/li>\n<li>A paste of the build log or failure point that you reached.<\/li>\n<li>Patience.<\/li>\n<\/ul><h2><a id=\"user-content-irc\" class=\"anchor\" aria-hidden=\"true\" href=\"#irc\"><\/a>IRC<\/h2>\n<p>Clasp has an IRC channel on <a href=\"https:\/\/freenode.net\/\" rel=\"nofollow\">Freenode<\/a> called #clasp. If you have questions, problems, suggestions, or generally would like to just hang out with us devs, come and stop by!<\/p>\n<h2><a id=\"user-content-more-on-clasp\" class=\"anchor\" aria-hidden=\"true\" href=\"#more-on-clasp\"><\/a>More on Clasp<\/h2>\n<p>For more information on Clasp and the discussion around it, see the following sites:<\/p>\n<ul><li><a href=\"https:\/\/drmeister.wordpress.com\" rel=\"nofollow\">Christian Schafmeister's blog<\/a><\/li>\n<li><a href=\"https:\/\/hn.algolia.com\/?query=clasp&amp;sort=byPopularity&amp;prefix&amp;page=0&amp;dateRange=all&amp;type=story\" rel=\"nofollow\">Hackernews<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/lisp\/search?q=clasp&amp;restrict_sr=on\" rel=\"nofollow\">Reddit<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=8X69_42Mj-g\" rel=\"nofollow\">Google Tech Talks<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=5bQhGS8V6dQ\" rel=\"nofollow\">ELS2016<\/a><\/li>\n<\/ul><h2><a id=\"user-content-acknowledgments\" class=\"anchor\" aria-hidden=\"true\" href=\"#acknowledgments\"><\/a>Acknowledgments<\/h2>\n<p>Clasp was supported by:<\/p>\n<ul><li>The Defense Threat Reduction Agency (DOD-DTRA) (HDTRA1-09-1-0009)<\/li>\n<li>The National Institutes of Health (NIH\/NIGMS Grant number: 2R01GM067866-07A2)<\/li>\n<li>The National Science Foundation (Grant number: 1300231)<\/li>\n<\/ul><\/article>","created_at":"2018-12-19T13:54:08+0000","updated_at":"2018-12-19T13:54:15+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/9052708?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":747,"title":"Stripe Atlas: Writing copy for landing pages","url":"https:\/\/stripe.com\/atlas\/guides\/landing-page-copy","content":"<p>In the beginning, there\u2019s the landing\u00a0page.<\/p><p>It\u2019s often a company\u2019s first impression: a digital introduction to an organization, the problem it tackles, and the solutions it offers. The sophistication of landing pages can vary. Some companies use their homepage as a landing page. Others create specific pages that target discrete customer profiles and move them through a marketing\u00a0funnel.<\/p><p>Landing pages also take various forms depending on the growth stage of a company. But that\u2019s one of the common missteps that I see: landing pages prioritize the story of the startup, rather than the journey <em>of the customer<\/em> with the\u00a0startup.<\/p><p>Landing page copy improves the more customers see themselves in it. Rethink and revise your landing page by <a href=\"#the-foundational-elements-of-effective-landing-page-copy\">building it around three elements<\/a>. If that\u2019s too much effort and you\u2019re looking for fast fixes, the following tips will improve your current landing page\u00a0copy.<\/p><p><strong>Focus copy on them.<\/strong> Landing pages often say \u201cwe offer\u201d or \u201cour solution,\u201d which focuses on the wrong thing\u2014your company, not your customers. Go through each sentence in your copy and rewrite it to address your customers. One way to do this is to begin with the word \u201cyou.\u201d Another tip is to start your sentence with a verb. Focusing on them nearly guarantees that your copy will address\u2014and speak to\u2014your visitor. Here\u2019s a good example of what this looks\u00a0like:<\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/modern_fertility.jpg\" alt=\"modern fertility screenshot\" class=\"expandable\" \/><p>The only possible exception to leading with \u201cyou\u201d in your copy is if you\u2019re a service business, where prospects want to see what you do differently as a service provider. That\u2019s often expressed as \u201cwe\u201d language. But it doesn\u2019t have to be. It just very often\u00a0is.<\/p><p><strong>Add pattern, texture, and shine to a block of copy.<\/strong> Accurate, succinct, and grammatically correct copy can still feel flat. Patterns, texture, and shine can add another dimension to your copy, making it more engaging and memorable. Here\u2019s\u00a0how:<\/p><ul><li>\n    <p><strong>Apply a pattern to a sentence.<\/strong> Patterns may be used to subtly reference a logo or map out a theme. Take a logistics company that transports goods by railway. Perhaps it not only wants to deliver a message, but also simulate a train with a pattern of evenly spaced dots in a horizontal line. To mimic that visual, rewrite landing page copy to link only words of similar length, such as three- or four-letter\u00a0words.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Vary sentence length and formatting to create texture.<\/strong> Texture can make copy feel more conversational, natural, and engaging. To create texture, write a smooth, polished sentence and juxtapose it with a more staccato sentence. Throw in some short sentences. More. More. And maybe one more. And then add a sentence that goes a lot longer, using clauses to lure your reader along. Then stop. The result?\u00a0Texture.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Include a glossy word or two for shine.<\/strong> Polish might mean swapping in a few new words. Find a bland word in your copy, and replace it with a more dazzling synonym. Take the <a href=\"https:\/\/www.collectiveretreats.com\/\">Collective Retreats<\/a> landing page copy below. Instead of \u201cFind destinations without losing luxury,\u201d try \u201cExplore extraordinary destinations without sacrificing luxury.\u201d The bright adjective and stronger verbs adds punch to the\u00a0copy.<\/p>\n  <\/li>\n<\/ul><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/collective_retreats.jpg\" alt=\"collective retreats screenshot\" class=\"expandable\" \/><p><strong>Defang objections with an \u201ceven if\u201d clause.<\/strong> If you can anticipate what might keep someone from believing your claim or assertion, undercut that opposition by acknowledging it. It\u2019ll hint that you understand their fear, uncertainty, and doubt\u2014and suggest that your solution takes those considerations into account. The formula is simple: \u201c[Claim] even if [objection].\u201d A very simple example is: \u201cBe creative even if you\u2019re not creative.\u201d Here\u2019s an \u201ceven if\u201d clause in the\u00a0wild:<\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/woorank.jpg\" alt=\"woorank screenshot\" class=\"expandable\" \/><p><strong>Limit each sentence to one idea.<\/strong> Sentences have the capacity to carry a lot of information, but your reader cannot. Your readers depend on periods, question marks, and even exclamation marks to give their brains a short rest\u2014just enough of a reprieve to absorb information before moving on. The more you help readers with information digestion, the more appetite they\u2019ll have to read on. So edit every sentence to have just one thought. Not two. Not three. Apple is skilled at this technique, but even it has\u00a0opportunities:<\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/apple_tv.jpg\" alt=\"apple tv screenshot\" class=\"expandable\" \/><p>Take its copy for Apple TV 4K. Most of the sentences don\u2019t qualify as sentences, but they absolutely follow the rule. Here\u2019s the one that doesn\u2019t: \u201cApple TV 4K lets you watch movies and shows in amazing 4K HDR\u2014and now it completes the picture with immersive sound from Dolby Atmos.\u201d This sentence merges two distinct features. The average reader would be better able to take in the information if that sentence was broken in\u00a0two:<\/p><p class=\"indented-copy\">\n        Apple TV 4K lets you watch movies and shows in amazing 4K HDR. It\u2019s got immersive sound from Dolby Atmos. Streams your favorite channels live. Has great content from apps like Netflix, Amazon Prime Video, and ESPN. And thanks to Siri, you can control it all with just your voice.\n      <\/p><p><strong>Create a landing page that\u2019s <em>not<\/em> your homepage.<\/strong> If you\u2019ve recently created a website for your company, your landing page might be your homepage, but they\u2019re two different creatures. A landing page is designed to convert prospects into customers. It speaks to visitors looking for something specific, features content relevant to that particular item, and contains a call-to-action customized to that precise offering. On the other hand, a homepage serves a broad audience, features widely-relevant content, and may not have an immediate\u00a0CTA.<\/p><p>More sophisticated companies will route visitors who, say, search for products for large companies to a page geared toward enterprises rather than a homepage with general information. If you have a website with more than one page\u2014or a more advanced information architecture and sitemap\u2014consider linking to the page on your site that best addresses a visitor\u2019s\u00a0intent.<\/p><p>First, the spoiler: landing page copy is never done. Like your company, it will\u2014and should\u2014evolve over time. It\u2019s an iterative process, and can always be improved. There are hundreds of copywriting formulas (many of them documented <a href=\"https:\/\/copyhackers.com\/2015\/10\/copywriting-formula\/\">here<\/a>) that can help you craft a headline, draft bullets, or structure a CTA. But if you don\u2019t have the fundamentals down, the improvement will be incremental, not\u00a0game-changing.<\/p><p>Copy for a high-performing landing page has three foundational\u00a0qualities:<\/p><p><strong>It delivers a convincing first impression for the startup.<\/strong> At a fundamental level, this involves direct, specific, and grammatically-correct copy. That level of precision and professionalism sets expectations and raises the bar for future engagements with a company\u2019s product and\u00a0team.<\/p><p><strong>It considers the maturity of the market.<\/strong> Every market\u2014like every company\u2014is in its own stage of development. Companies in highly mature markets can use short copy, because most visitors already get the ins-and-outs of the solution or category (e.g. disposable razors). Their focus should be on product differentiators, the brand story, and who\u2019s using it. If a market is still emerging (e.g. cryptocurrency in 2018) customers likely need more information, because a company is not only educating about its solution, but also helping define the category more\u00a0broadly.<\/p><p><strong>It reflects the customer\u2019s stage of awareness.<\/strong> Effective landing page copy mirrors customer comprehension, which is layered and includes: how well a customer understands the challenge the company addresses (e.g. data privacy), what\u2019s at stake (e.g. personal data and\/or compliance), the changing landscape (e.g. <a href=\"https:\/\/stripe.com\/guides\/general-data-protection-regulation\">GDPR<\/a>), and tools that help (e.g. a specific\u00a0product).<\/p><p>In short, great copy demonstrates that a company grasps how aware customers are of the market, its pain points, and potential solutions. The best copy does all that <em>and<\/em> signals that the company is clued into precisely how aware the customers are of <em>themselves<\/em>. Here are the five stages of awareness for any visitor to your landing\u00a0page:<\/p><ul><li>\n    <p><strong>Most aware.<\/strong> Visitors totally understand your solution and likely believe it\u2019s a top contender for them. They just needs nudging. Purchases happen\u00a0here.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Product-aware.<\/strong> Visitors are learning about your product. Free trials, demos, and purchases happen\u00a0here.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Solution-aware.<\/strong> Visitors are considering solutions to their pain or\u00a0problem.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Problem-aware.<\/strong> Visitors are feeling pain or dealing with a\u00a0problem.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Unaware.<\/strong> Visitors haven\u2019t experienced a need that would drive them to your\u00a0solution.<\/p>\n  <\/li>\n<\/ul><p>Landing page copy reflects the customer if you can answer \u201cyes\u201d to the following\u00a0questions:<\/p><ul><li>\n    <p>Is the language accessible and does it mirror a visitor\u2019s stage of\u00a0awareness?<\/p>\n  <\/li>\n  <li>\n    <p>Does the copy move them from where they are to where they want to\u00a0be?<\/p>\n  <\/li>\n  <li>\n    <p>Are visitors prompted to take action once they become Product-aware or Most\u00a0aware?<\/p>\n  <\/li>\n<\/ul><p>Every landing page should <strong>deliver a convincing first impression, consider the maturity of the market,<\/strong> and <strong>reflect the customer\u2019s stage of awareness<\/strong>. The best way to see how the absence or presence of these elements alter a landing page is to review real-life examples. So, with the three foundational elements in mind, I\u2019ve evaluated the landing pages of nine companies\u2014from member management software to a vegan candy brand to a court date notification service. Here\u00a0goes:<\/p><p>\n        <b>Company<\/b>: <a href=\"https:\/\/www.kapwing.com\">Kapwing<\/a><br \/><b>What it does<\/b>: Modern editor for videos, GIFs, and images<br \/><b>Landing Page Challenge<\/b>: Copy that\u2019s short, but vague<br \/><b>The Fix to Apply<\/b>: Get specific\u2014but, first, scroll down<br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/kapwing.jpg\" alt=\"kapwing screenshot\" class=\"expandable\" \/><p><strong>The\u00a0Analysis<\/strong><\/p><p>The headline\u2014\u201dCreativity made easy\u201d\u2014 raises this question: Are people who\u2019d use this solution <em>not<\/em> creative? The phrase suggests that it\u2019s hard for your audience to be creative. Is that\u00a0true?<\/p><p>The page falls victim to the myth that copy must be kept short at all costs. Many startups write three- or four-word headlines that try to say everything. It reduces the elevator pitch to a hiccup. Along the way, we tell ourselves we\u2019re being succinct because we understand that being succinct is good when it comes to\u00a0communication.<\/p><p>But are you being succinct or are you being vague? Worse, are you sacrificing a clear message for a short one? Your headline is prime real estate. Let it do some real work\u2014it\u2019s made for\u00a0it.<\/p><p>Think back to your last ten conversations with a customer or prospect. (If you haven\u2019t had at least ten, <a href=\"https:\/\/copyhackers.com\/2018\/05\/how-to-listen\/\">do that first<\/a>.) At what point in the conversation did they perk up? That\u2019s the fodder for a good headline\u2014at least one that\u2019s worth testing. That\u2019s what should top your landing\u00a0page.<\/p><p>With headlines, don\u2019t take rich, nuanced feedback from customers and whittle it down to the smallest possible version. Instead listen. Write down what you hear. And put it on the page. You\u2019ll clean it up so it\u2019s on brand and persuasive, but it shouldn\u2019t be boiled\u00a0down.<\/p><p>The concern in boiling it down is that you\u2019ll get too vague. Kapwing\u2019s tools run the gamut. They make memes, add subtitles to video, create montages and more. If you\u2019re the marketing team, the temptation is to look at everything Kapwing does, plot it in a Venn diagram, and promote the overlapping part: the quality that all features, uses cases and outputs have in common. That\u2019s likely how \u201cCreativity made easy\u201d\u00a0emerged.<\/p><p>That\u2019s how mediocre headlines surface\u2014and it happens surprisingly often, even to the most copy-savvy startups. Many founders use more effective headlines in their verbal elevator pitches or elsewhere in copy. That\u2019s the case for Kapwing; the better options for a headline are further down in the\u00a0copy:<\/p><p class=\"indented-copy\">\n        <strong>Tasks that take hours in iMovie take minutes in Kapwing.<\/strong><br \/>No install. No passwords. No technical tutorials. It works on every OS, on your phone, and your computer.\n      <\/p><p>OR<\/p><p class=\"indented-copy\">\n        <strong>Bring on the memes.<\/strong><br \/>Creative jobs that take hours in iMovie take just minutes in Kapwing.\n      <\/p><p>Either copy expresses the ease of being creative in Kapwing in a specific, compelling way\u2014and serve as entry points to its other suite of tools. These messages should lead the page as a headline. Hidden headline copy happens to many businesses. It\u2019s not that companies can\u2019t generate good headlines; they\u2019ve likely just buried\u00a0them.<\/p><p>\n        <b>Company:<\/b> <a href=\"https:\/\/www.webgazer.io\/\">WebGazer<\/a><br \/><b>What it does:<\/b> Free website monitoring<br \/><b>Landing Page Challenge:<\/b> We-focused copy<br \/><b>The Fix to Apply:<\/b> \"The Rule of You\"<br \/><\/p><p><strong>The\u00a0Analysis<\/strong><\/p><p>This headline works, as long as visitors arrive on the page knowing what WebGazer is and does. The copy might suffice if this is a retargeting\u00a0page.<\/p><p>But what if it\u2019s not? Let\u2019s start with the phrase: \u201cWe help you keep your business running.\u201d The very first word (\u201cwe\u201d) is a problem. We-focused copy is an issue\u2014almost without fail. It communicates from the perspective of the company, not the customer. It\u2019s the equivalent of a person introducing themselves and dominating the airspace with their story. Correct \u201cwe-focused\u201d copy by rewriting every sentence to begin with the word \u201cyou\u201d or a verb. Here\u2019s how WebGazer can change three sentences on its landing\u00a0page:<\/p><table style=\"width: 100%;\"><thead><tr><th>We-focused copy<\/th>\n            <th>You-focused copy<\/th>\n          <\/tr><\/thead><tbody><tr><td>We monitor your sites without a rest, see what we have done today yourself.\n            <\/td>\n            <td>You can rest easy. Because we\u2019ll monitor your site without rest.  <\/td>\n          <\/tr><tr><td>WebGazer checks if your website is up as it should be and notifies you if anything goes wrong.<\/td>\n            <td>Keep your online business running without interruption. WebGazer monitors it night and day. And notifies you if anything goes wrong. <\/td>\n          <\/tr><tr><td>We help you keep your business running.<\/td>\n            <td>How would you know if your website was suddenly down?<\/td>\n          <\/tr><\/tbody><\/table><p>\n        <b>Company:<\/b> <a href=\"https:\/\/www.influencekit.com\/\">InfluenceKit<\/a><br \/><b>What it does:<\/b> Automated reporting for sponsored content<br \/><b>Landing Page Challenge:<\/b> A timely call to action<br \/><b>The Fix to Apply:<\/b> <a href=\"https:\/\/copyhackers.com\/2015\/10\/copywriting-formula\/#AIDA)[ framework](https:\/\/copyhackers.com\/2015\/10\/copywriting-formula\/#AIDA\">AIDA<\/a><br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/influence_kit.jpg\" alt=\"influence kit screenshot\" class=\"expandable\" \/><p><strong>The\u00a0Analysis<\/strong><\/p><p>My first impression is how much is <em>right<\/em> with this landing page copy. It reads well and follows some of the strongest copywriting better practices,\u00a0including:<\/p><ul><li>\n    <p>Removing distracting calls to action (e.g. no global\u00a0navigation)<\/p>\n  <\/li>\n  <li>\n    <p>Hooking visitors immediately with a highly desirable benefit (why they should\u00a0care)<\/p>\n  <\/li>\n  <li>\n    <p>Expanding on the benefit with product-specific support (how the product does\u00a0it)<\/p>\n  <\/li>\n  <li>\n    <p>Formatting copy for maximum readability (the three paragraphs of body copy under \u201cYou have a blog?\u201d is a manageable column to read instead of spanning the width of the page, which is extremely fatiguing for the reading\u00a0eye.)<\/p>\n  <\/li>\n  <li>\n    <p>Using a voice that is present but not overwhelming (With few exceptions, they\u2019ve used their voice in crossheads only, minimizing \u2018interruptions\u2019 in how readers process\u00a0information.)<\/p>\n  <\/li>\n  <li>\n    <p>Leading with \u201cyou\u201d wherever\u00a0possible<\/p>\n  <\/li>\n  <li>\n    <p>Making sentences short, and easy to\u00a0consume<\/p>\n  <\/li>\n<\/ul><p>The landing point copy has done so much right\u2014the challenge is that it hasn\u2019t cashed in on it. After making a convincing argument to the prospect, it forgets to end with a CTA. The AIDA framework can help outline this progression and missed\u00a0opportunity.<\/p><p><strong>AIDA: Attention, Interest, Desire,\u00a0Action.<\/strong><\/p><p><strong>Grab their attention<\/strong>. In this case, InfluenceKit starts to build trust and empathy by clearly articulating the problem many bloggers face: demonstrating value from content. The copy\u00a0reads:<\/p><p class=\"indented-copy\">\n        ...it\u2019s really hard proving to brands how much value you\u2019re delivering.\n      <\/p><p><strong>Build up their interest in the subject<\/strong>. Here, the copy deepens that understanding by elaborating on the problem, causing readers to nod along and more fully recognize themselves in the challenge. This shared understanding builds further interest. The copy\u00a0continues:<\/p><p class=\"indented-copy\">Yes, it\u2019s about the content, but it\u2019s also about knowing how many lives were impacted by what you created.<\/p><p class=\"indented-copy\">We all like to know where our money's going (dang you Target!!). Brands are just like us; they want to know what they're getting for their investment. Reporting impressions is the industry standard for internet marketing, and up until now, bloggers have been a black hole of information.<\/p><p class=\"indented-copy\">Your blog is more than just pretty pictures, you have real influence. You know it, and we know it, and together we\u2019re going to prove it.<\/p><p><strong>Turn that casual interest into a direct desire<\/strong> for them to add the solution to their lives. The solution that\u2019s offered is succinct and clear. The copy\u00a0reads:<\/p><p class=\"indented-copy\">Show Your Real Influence with Powerful and Dynamic Reports<\/p><p><strong>Call them to action.<\/strong> There\u2019s a CTA, but it\u2019s in the wrong place. It\u2019s buried beneath a list of integrations, the four-step process to set-up InfluenceKit, an infographic on the two types of bloggers, and a Meet-the-Team section. Finally, there\u2019s a call to\u00a0action:<\/p><p class=\"indented-copy\">Ready to Prove Your Value &amp; Make More Money? Request Access Now. (Secret handshake to follow)<\/p><p>The copy has sufficiently made a case for action by the end of the four-part \u201cHow Does InfluenceKit Work?\u201d section. Here, I\u2019d expect\u2014and encourage\u2014a <em>closer<\/em> section,\u00a0featuring:<\/p><ul><li>\n    <p>A restatement of the value\u00a0prop<\/p>\n  <\/li>\n  <li>\n    <p>A quick bullet list of results generated by the\u00a0solution<\/p>\n  <\/li>\n  <li>\n    <p>A powerful\u00a0screenshot<\/p>\n  <\/li>\n  <li>\n    <p>A\u00a0CTA<\/p>\n  <\/li>\n  <li>\n    <p>A testimonial from a blogger or\u00a0sponsor<\/p>\n  <\/li>\n  <li>\n    <p>A data point about average\u00a0results<\/p>\n  <\/li>\n<\/ul><p>Perhaps InfluenceKit is a newer company, which is why there\u2019s no data yet for how it\u2019s paying off. But there\u2019s power in being new. I\u2019d suggest the founders share their own story and results thanks to InfluenceKit\u2014being transparent all along that they\u2019re the founders\u2014to demonstrate that others can also benefit. Part of making a lasting first impression is creating a window for not only a prospect to relate, but act, and interact. That turns an impression into a\u00a0relationship.<\/p><p>\n        <b>Company:<\/b> <a href=\"https:\/\/nombacandies.com\/\">Nomba<\/a><br \/><b>What it does:<\/b> Digital-first vegan candy brand<br \/><b>Landing Page Challenge:<\/b> The model to try or demo<br \/><b>The Fix to Apply:<\/b> Map your market over your first customer engagement<br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/nomba.jpg\" alt=\"nomba screenshot\" class=\"expandable\" \/><p><strong>The\u00a0Analysis<\/strong><\/p><p>The \u201cTry Now\u201d language at the top of the landing page\u2014and throughout the site\u2014is confusing. But what\u2019s more puzzling is the idea of a trial for this\u00a0product.<\/p><p>For the mainstream consumer, trials are not typical in the world of buying candy. By offering a trial, you\u2019ve effectively taken a super-sophisticated market of buyers\u2014people who\u2019ve bought candy, understand buying candy, and just wanna get vegan candy\u2014and introduced doubt. Suddenly visitors are asking small questions that escalate into bigger\u00a0concerns:<\/p><ul><li>\n    <p>What does a trial mean in the world of\u00a0candy?<\/p>\n  <\/li>\n  <li>\n    <p>Try food that\u2019s shipped to me? Do I ship it\u00a0back?<\/p>\n  <\/li>\n  <li>\n    <p>Will I get a sample or a full-sized\u00a0product?<\/p>\n  <\/li>\n  <li>\n    <p>Wait. Is this a subscription I\u2019ll have to work to get out\u00a0of?<\/p>\n  <\/li>\n  <li>\n    <p>Do I even <em>want<\/em> spicy\u00a0candy?<\/p>\n  <\/li>\n<\/ul><p>And just like that, they\u2019re\u00a0gone.<\/p><p>Your visitors go from wanting to give this spicy vegan candy a shot to hitting the back button and feeling relief that they got out of there as quickly as they\u00a0did.<\/p><p>Nomba\u2019s copy has a new challenge: it needs to deal with the \u201cproblem\u201d of prospective customers not understanding the trial model for a food item. Here\u2019s how it can do\u00a0that:<\/p><ul><li>\n    <p><strong>First, know that you have to address the problem.<\/strong> But you <em>do not have to go into great detail<\/em> to address it. After all, this is a $5 treat or $15 investment in\u00a0sugar.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Second, ask: what\u2019s the easiest, quickest way is to address the problem?<\/strong> Often, it\u2019s in naming. Sometimes product features are better explained by renaming them and pricing tiers are better explained by re-labeling each tier. In this case, the <em>trial<\/em> may be better grokked by renaming it to <em>sample pack<\/em> or\u00a0<em>sampler<\/em>.<\/p>\n  <\/li>\n<\/ul><p>Now the \u201cTry Now\u201d CTA could become any of the\u00a0following:<\/p><ul><li>\n    <p>Get the\u00a0Sampler<\/p>\n  <\/li>\n  <li>\n    <p>Sample\u00a0Nomba<\/p>\n  <\/li>\n  <li>\n    <p>Sample\u00a0It<\/p>\n  <\/li>\n<\/ul><p>Nomba should keep that new language consistent throughout each product page and checkout to avoid further confusion. Any startup that offers demos or trials to win customers should craft their copy with not only their customer, but their existing market in mind. The market part of product\/market fit also matters.\u00a0<\/p><p>\n        <b>Company<\/b>: <a href=\"https:\/\/eve.io\/\">Eve<\/a><br \/><b>What it does<\/b>: Event management software<br \/><b>Landing Page Challenge<\/b>: Overemphasis on the solution, not the problem<br \/><b>The Fix to Apply<\/b>: <a href=\"https:\/\/copyhackers.com\/2015\/10\/copywriting-formula\/#PAS\">PAS framework<\/a><br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/eve.jpg\" alt=\"eve screenshot\" class=\"expandable\" \/><p><strong>The\u00a0Analysis<\/strong><\/p><p>The prospects for event management software may already recognize the need for a solution, but if not, this landing page needs more education on the challenges of planning events. After all, there\u2019s a lot of complexity in running events: multiple stakeholders, shifting timelines, numerous vendors, and changing agendas. The list goes\u00a0on.<\/p><p>The landing page starts out relatively strong with a benefit-driven headline and a product-level value\u00a0proposition:<\/p><p class=\"indented-copy\">\n        The easier way to plan events.<br \/>Eve allows your entire team to plan, execute and review your events, all in one place \u2014 without changing your existing workflow.\n      <\/p><p>Leading with a value proposition is the standard choice for SaaS companies. That\u2019s fine, but a lot of marketers struggle with what should go next on the page. For a product like eve\u2014where it\u2019s easy for the prospect and the whole world to minimize the pain of event planning until you\u2019re two days away from the event and panicking\u2014I\u2019d recommend moving swiftly from discussing the value prop to outlining and fleshing out the problem eve\u00a0solves.<\/p><p>It may seem like that\u2019s accomplished with the first crosshead under the\u00a0hero:<\/p><p class=\"indented-copy\">\n        Easy software to end event stress\n      <\/p><p>But that crosshead leads with the solution (\u201ceasy software\u201d) when it ought to focus entirely on the pain (\u201cevent stress\u201d). The PAS\u2014or Problem, Agitation, Solution\u2014framework helps lay out copy in a way that it zeroes in on the driving force behind choosing a painkiller solution like eve: the\u00a0pain.<\/p><p>My biggest advice is choosing copy that doesn\u2019t glide over that pain. So swap\u00a0this:<\/p><p class=\"indented-copy\">\n        \"How hard could it be to plan and coordinate a live event?\"\n      <\/p><p>with\u00a0this:<\/p><p class=\"indented-copy\">\n        \"How hard could it be to plan and coordinate a two-day conference with 20 speakers from around the globe and 500 attendees?\"\n      <\/p><p>The first option is safe. The second option is\u00a0sticky.<\/p><p>Too often we try to boil everything down to its safest version. And we sacrifice specific, sticky copy that might actually hook and convert actual prospects. It\u2019s not our fault. Everyone we know is writing safe copy. If we go bold, we\u2019re taking a\u00a0risk.<\/p><p>Test it and see, but my encouragement is to go for it. Rare is the business that makes waves today without either 1) a huge marketing budget to overcome their safe messaging or 2) a memorable way of connecting with people. Endeavor to be uncomfortably specific with your prospects. It shows you\u2019re in touch with the precise pain felt by the professionals in your\u00a0market.<\/p><p>\n        <b>Company:<\/b> <a href=\"https:\/\/www.scribendi.com\/\">Scribendi<\/a><br \/><b>What it does:<\/b> Editing and proofreading services for English documents<br \/><b>Landing Page Challenge:<\/b> Understanding scale and results<br \/><b>The Fix to Apply:<\/b> Visualize\u2014don\u2019t just display\u2014data<br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/scribendi.jpg\" alt=\"scribendi screenshot\" class=\"expandable\" \/><p><strong>The\u00a0Analysis<\/strong><\/p><p>Overall, the copy on this page is strong. But copy is not always words, but\u00a0numbers.<\/p><p>Further down the page, there is some impressive proof of impact on clients served, words proofread and editors in\u00a0action:<\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/scribendi_2.jpg\" alt=\"scribendi2 screenshot\" class=\"expandable\" \/><p>These are striking figures, but they also come with\u00a0complications.<\/p><p>First, I\u2019m sure I\u2019m not the only person who misread 1.6B as 1.6M. That sounds like a lot, but how many documents does that mean? Is it 10,000+ ESL papers for academics or 100,000 essays for\u00a0undergraduates?<\/p><p>I suggest keeping the data as is, but adding in a visual that shows the scale of the achievement. Here\u2019s one way to back into a compelling visual via a quick calculation: 1.6B words is actually 6.4M pages. That\u2019s 318,939 millimeters. That\u2019s 1,046 feet. That\u2019s amazing. That\u2019s more than half the size of the CN Tower in Canada\u2014or insert your local landmark relevant to your\u00a0audience.<\/p><p>Help your prospects visualize the work you\u2019ve done\u2014and don\u2019t be afraid to reference something <em>outside<\/em> your market to make the case <em>for<\/em> your market. Numbers are good. But I believe the saying goes: whoever draws a picture\u00a0wins.<\/p><p>\n        <b>Company:<\/b> <a href=\"https:\/\/www.joinit.org\/\">Join It<\/a><br \/><b>What it does:<\/b> Member management software<br \/><b>Landing Page Challenge:<\/b> Assumes visitor is solution-aware<br \/><b>The Fix to Apply:<\/b> Shift to the mindset of a solution-aware prospect<br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/join_it.jpg\" alt=\"join it screenshot\" class=\"expandable\" \/><p><strong>The\u00a0Analysis<\/strong><\/p><p>This headline\u2014Membership Management Software\u2014is great for search engine optimization (SEO), but it\u2019s so technical that it\u2019s jarring, especially as the first three bold words on the\u00a0page.<\/p><p>For a headline like \u201cMembership Management Software\u201d to work, its prospects need to be sophisticated and \u201csolution-aware.\u201d Here\u2019s where that stacks on the five stages of\u00a0awareness:<\/p><ul><li>\n    <p><strong>Most aware.<\/strong> Visitors totally understand your solution and likely believe it\u2019s a top contender for them. They just need nudging. Purchases happen\u00a0here.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Product-aware.<\/strong> Visitors are learning about your product. Free trials, demos and purchases happen\u00a0here.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Solution-aware.<\/strong> Visitors are considering solutions to their pain or\u00a0problem.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Problem-aware.<\/strong> Visitors are feeling pain or dealing with a\u00a0problem.<\/p>\n  <\/li>\n  <li>\n    <p><strong>Unaware.<\/strong> Visitors haven\u2019t experienced a need that would drive them to your\u00a0solution.<\/p>\n  <\/li>\n<\/ul><p>If prospects are solution-aware, they\u2019re actively seeking \u201cmembership management software\u201d and would easily call it that. However, Join It is eliminating a segment of prospects by only addressing solution-aware visitors. By finessing headlines and copy to match the <em>problems or state of mind<\/em> of prospects, Join It can open its aperture to capture more potential customers. Here\u2019s how that might look with tweaks to\u00a0copy:<\/p><table style=\"width: 100%;\"><thead><tr><th>Solution-aware prospect\u2019s state of mind on arrival<\/th>\n            <th>Corresponding headline copy<\/th>\n          <\/tr><\/thead><tbody><tr><td>I\u2019ve heard that membership management software can help me automate all the community stuff I\u2019m struggling with<\/td>\n            <td>Now You Can Automate Community Engagement with Easy Membership Management Software<\/td>\n          <\/tr><tr><td>I need membership management software that does this big list of stuff<\/td>\n            <td>Fast-Growing Organizations Need Membership Management Software That:\n              <ul><li>Brings your community together<\/li>\n                <li>Works intuitively<\/li>\n                <li>Automates your tasks<\/li>\n                <li>Scales as you grow<\/li>\n              <\/ul><\/td>\n          <\/tr><tr><td>I\u2019m thinking of switching from X membership management software, which I don\u2019t like because of Y<\/td>\n            <td>Put an End to Y with Easy Membership Management Software That Scales<\/td>\n          <\/tr><tr><td>Should I be terrified of all the work that\u2019s about to come when I introduce a membership management solution in my organization?<\/td>\n            <td><small>Membership management software<\/small><br \/>How Much Easier Would Growing Your Business Be If You Could\n              <b>Automate Your Community?<\/b>\n            <\/td>\n          <\/tr><\/tbody><\/table><p>A shift from solution-aware language to the state-of-mind of a prospect may seem like you\u2019re downshifting to a problem-aware visitor. This stage of visitor might seem further from a product-aware prospect, where a sale can\u00a0occur.<\/p><p>But given the market maturity for membership management software, Join It will open its top-of-the-funnel to more prospects if it makes its headline more accessible to more visitors. And it\u2019ll still be relatable to solution-aware and product-aware\u00a0prospects.<\/p><p>\n        <b>Company:<\/b> <a href=\"https:\/\/parkpnp.com\/us\/\">Parkpnp<\/a><br \/><b>What it does:<\/b> Marketplace for unused or underutilized parking spaces<br \/><b>Landing Page Challenge:<\/b> Incorrect assumptions about the visitor<br \/><b>The Fix to Apply:<\/b> <a href=\"https:\/\/copyhackers.com\/2015\/10\/copywriting-formula\/#PAS\">The Rule of One<\/a><br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/park_pnp.jpg\" alt=\"park pnp screenshot\" class=\"expandable\" \/><p>A key part of customer awareness is understanding the context and drive for their visit\u2014all of which should be reflected in the copy and experience of a page. That nuance can be the difference between a visitor identifying you as a hero in their lives\u2014or a\u00a0villain.<\/p><p>In the case of Parkpnp, context and drive are pivotal factors for its solution. An app that allows you to find a parking space might be used in advance of an anticipated painful parking experience, such as major sporting events. But it\u2019s more likely to be used in the heat of the moment, when the pain for a prospect is most visceral and distractions\u00a0abound.<\/p><p>In short, most people will use Parkpnp when they are driving their cars looking for a place to park in a busy area. Their pain is high. Their motivation to solve the pain is\u00a0high.<\/p><p>They need a\u00a0hero.<\/p><p>Is your site a hero or a villain in that\u00a0moment?<\/p><p>The current hero section copy and experience sets up the solution to be more the villain than the hero. First, it puts work on the user\u2019s plate by asking for an address\u2014which holds true on both mobile and desktop. But perhaps more glaringly, it makes assumptions about what \u201cperfect\u201d means to the user. In fact, <em>perfect parking<\/em> could\u00a0mean:<\/p><ul><li>\n    <p>It\u2019s near where I am\u00a0now.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s closest to my final\u00a0destination.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s available\u00a0now.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s available regularly \/\u00a0always.<\/p>\n  <\/li>\n  <li>\n    <p>It allows overnight\u00a0parking.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s the cheapest option in the\u00a0vicinity.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s\u00a0free.<\/p>\n  <\/li>\n  <li>\n    <p>It accepts credit card\u00a0payments.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s patrolled by\u00a0security.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s\u00a0covered.<\/p>\n  <\/li>\n  <li>\n    <p>It has extra-wide parking\u00a0spots.<\/p>\n  <\/li>\n  <li>\n    <p>It doesn\u2019t require parallel\u00a0parking.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s got special needs\u00a0access.<\/p>\n  <\/li>\n<\/ul><p>The word \u201cperfect\u201d is trying to cover all manner of desires in the headline for this page. But perfect for User A may not be perfect for User B. This is where the business needs to make a decision: What\u2019s the primary prospect\u2019s top definition of <em>perfect parking<\/em> in the moment they use the\u00a0solution?<\/p><p>This question gets to the heart of what copywriters call the Rule of One, which holds that you are always writing a landing page with the following four points identified\u00a0first:<\/p><ul><li>\n    <p>Your One\u00a0Reader<\/p>\n  <\/li>\n  <li>\n    <p>Your One Big Idea (or culture-shifting\u00a0idea)<\/p>\n  <\/li>\n  <li>\n    <p>Your One Promise (or desirable measurable\u00a0outcome)<\/p>\n  <\/li>\n  <li>\n    <p>Your One\u00a0Offer<\/p>\n  <\/li>\n<\/ul><p>Of all parts of the Rule of One, your One Reader is the most important. You can\u2019t write an effective page without first identifying your One Reader. Once you know your One Reader well, you know what pains drive them, what benefits they\u2019re looking for, what offer they\u2019re most likely to respond best to, what kinds of social proof will move them\u2014the list goes on. You do not have to get incredibly detailed on your One Reader, but you do have to know them well enough to look at the above list of definitions of <em>perfect parking<\/em> and know which one they would choose if they could only choose\u00a0one.<\/p><p>The one they would use is the one you\u00a0use.<\/p><p>In the case of Parkpnp, if the One Reader defines <em>perfect parking<\/em> as <em>parking that\u2019s near where I am now<\/em>, the headline would change and so would the CTA. Here\u2019s\u00a0how:<\/p><table style=\"width: 100%;\"><thead><tr><th>From...<\/th>\n            <th>To...<\/th>\n          <\/tr><\/thead><tbody><tr><td>Headline<\/td>\n            <td>FIND YOUR PERFECT PARKING SPACE<\/td>\n            <td>FIND AN AVAILABLE PARKING SPACE NEARBY NOW<\/td>\n          <\/tr><tr><td>Primary CTA<\/td>\n            <td>Field: Where would you like to park?<br \/>Button: Find Parking<\/td>\n            <td>Button: Find Parking Near Me Now<\/td>\n          <\/tr><tr><td>Subordinate CTA<\/td>\n            <td>Button: Find Parking Nearby<\/td>\n            <td>Link: Find Parking for a Future Date, Event, or Long-Term Need<\/td>\n          <\/tr><\/tbody><\/table><p>That\u2019s how your site goes from being a villain to being a\u00a0hero.<\/p><p>Now, if your One Reader would define <em>perfect parking<\/em> as <em>parking near my work that doesn\u2019t cost half my paycheck<\/em> or <em>safe parking for my daughter who\u2019s away at college<\/em>\u2014that is, if they\u2019re planning for future parking needs\u2014the headline and experience would change accordingly, of course. You don\u2019t have to get that specific. But if you <em>can<\/em> get that specific in your headline, you\u2019re more likely to convert your visitor. The alternative is writing a headline that serves a huge range of audiences. But, be forewarned: a hero to all is a boring character\u2014and more likely to be ineffectual for the\u00a0business.<\/p><p>\n        <b>Company:<\/b> <a href=\"https:\/\/ecourtdate.com\/\">eCourt Date<\/a><br \/><b>What it does:<\/b> Timely, digital court date reminders<br \/><b>Landing Page Challenge:<\/b> Unclear who the user is<br \/><b>The Fix to Apply:<\/b> \"Ideal for\" statement<br \/><\/p><img src=\"https:\/\/d37ugbyn3rpeym.cloudfront.net\/atlas-guides\/landing-page-copy-guide\/e_court_date.jpg\" alt=\"e court date screenshot\" class=\"expandable\" \/><p>Here\u2019s the first\u2014and most important\u2014question about eCourt Date\u2019s landing page: do people land on this page and know if the service is meant for\u00a0them?<\/p><p>One of the cleanest, easiest ways to optimize your copy is to use an \u201cideal for\u201d statement. That statement can appear anywhere on your site, but it\u2019s best that it appears closer to the top of your landing page. That\u2019s especially true for a service like eCourt Date, where the user is unclear. It might be for lawyers, police officers, the court, defendants and plaintiffs\u2014maybe even those serving jury\u00a0duty.<\/p><p>Whichever one it is, put that on the page in an ideal-for statement,\u00a0like:<\/p><p class=\"indented-copy\">\n        Ideal for lawyers with criminal and DUI cases\n      <\/p><p>If it\u2019s for more than one user, that\u2019s fine. Put those names on the\u00a0page:<\/p><p class=\"indented-copy\">\n        Ideal for criminal lawyers, traffic cops, and courthouse administrators\n      <\/p><p>From there, the headline and subhead in the hero have context and communicate the service more clearly. To enhance the copy, take a few extra steps. Answer why it matters and add proof that you\u2019ve done it. This rule is known as \u201cSo What\u201d\/\u201dProve\u00a0it.\u201d<\/p><p>Here\u2019s how the copy might look with those two elements\u00a0addressed:<\/p><p class=\"indented-copy\">\n        Ideal for busy courthouse administrators<br \/><b>Send Court Date Reminders.<\/b><br \/><b>Reduce Failure to Appear Rates.<\/b><br \/>Notify defendants about court dates and other legal events. So you save up to 20 hours per month of court staff time. Reduce court no-shows by 33%. And humanize the defendant experience.\n      <\/p><p>Bolster this copy with testimonials from previously frustrated and now happy court staff or judges. These improvements all stem from articulating the ideal user for the service at the\u00a0onset.<\/p><p>Landing page copy is an underleveraged, powerful tool. Done right, it builds brand, engenders trust, and sells product\u2014to anyone with an internet connection, on their schedule. But it\u2019s not automatic. Landing page copy must deliver a convincing first impression, consider the maturity of the market, and reflect the customer\u2019s stage of awareness. It must meet prospects where they are and get them to where they\u2014and the business\u2014want them to be. Once it does, hand waves became high fives, and high fives turn into handshakes\u2014and conversions can happen without much human\u00a0intervention.<\/p><p>Don\u2019t overthink where you should start. Just get started. A landing page can\u2014and must\u2014always improve. If you have limited time and resources, run through <a href=\"#quick-wins-for-better-landing-page-copy\">the quick wins<\/a> to make tweaks that generate outsized gains. If you have more time, <a href=\"#the-foundational-elements-of-effective-landing-page-copy\">revamp your copy<\/a> to orient around making a lasting impression, the maturity of the market, and customer\u00a0awareness.<\/p><aside class=\"disclaimer\"><p>\n        Landing page copy is only one channel to the customer. Get <a href=\"https:\/\/stripe.com\/atlas\/guides\/email-marketing-ecommerce\">tips on email marketing<\/a>, or hone your cold emails and calls with <a href=\"https:\/\/stripe.com\/atlas\/guides\/starting-sales#how-should-you-communicate-with-customers\">this guide on finding your first customers<\/a>. If there\u2019s a related topic you wish we\u2019d cover, drop us a line at <a href=\"mailto:atlas@stripe.com\">atlas@stripe.com<\/a>.\n      <\/p>\n    <\/aside><a href=\"https:\/\/stripe.com\/atlas\/guides\" class=\"common-Link common-Link--arrowL\">Return to guides<\/a>","created_at":"2018-12-19T12:12:03+0000","updated_at":"2018-12-19T12:12:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":28,"domain_name":"stripe.com","preview_picture":"https:\/\/stripe.com\/img\/v3\/atlas\/guide\/social.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":748,"title":"OpenWeb Icons (a font!)","url":"https:\/\/pfefferle.github.io\/openwebicons\/icons\/","content":"<section id=\"icons\"><div class=\"row\"><p><\/p><h2>Uncategorized<\/h2><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-apml\">\n              <i class=\"openwebicons-apml\">\u00a0openwebicons-apml\n              <small>by\u00a0<a href=\"http:\/\/suda.co.uk\/projects\/apml\/\" target=\"_blank\">Brian Suda<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/apml.areyoupayingattention.com\/\" target=\"_blank\">APML<\/a> (Attention Profile Markup Language) allows users to share their own personal Attention Profile in much the same way that OPML allows the exchange of reading lists between News Readers.<\/p><\/div><div class=\"col-md-4\"><p>The <a href=\"http:\/\/www.shareaholic.com\/openshareicon\" target=\"_blank\">Open Share icon<\/a> conveys the act of sharing by visually representing one hand passing an object to another hand, as in \"pass it on\" or \"sharing\". The icon also represents an \"eye\", as in \"look at this\".<\/p><\/div><div class=\"col-md-4\"><p>An <a href=\"http:\/\/shareicons.com\/\">icon<\/a> to represent the generic action of sharing a web page. This could mean adding the page to a social bookmarking site or memetracker, e-mailing the page to someone, etc.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-feed\">\n              <i class=\"openwebicons-feed\">\u00a0openwebicons-feed\n              <small>by\u00a0<a href=\"http:\/\/feedicons.com\/\" target=\"_blank\">Feedicons.com<\/a><\/small>\n            <\/i><\/h3><p>A <a href=\"http:\/\/en.wikipedia.org\/wiki\/Web_feed\" target=\"_blank\">web feed<\/a> (or news feed) is a data format used for providing users with frequently updated content. Content distributors syndicate a web feed, thereby allowing users to subscribe to it.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-ostatus\">\n              <i class=\"openwebicons-ostatus\">\u00a0openwebicons-ostatus\n              <small>by\u00a0OStatus.org<\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/ostatus.org\/\" target=\"_blank\">OStatus<\/a> is an open standard for distributed status updates. Our goal is a specification that allows different messaging hubs to route status updates between users in near-real-time.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-opml\">\n              <i class=\"openwebicons-opml\">\u00a0openwebicons-opml\n              <small>by\u00a0<a href=\"http:\/\/opmlicons.com\/\" target=\"_blank\">Opmlicons.com<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/dev.opml.org\/\" target=\"_blank\">OPML<\/a> an XML-based format that allows exchange of outline-structured information between applications running on different operating systems and environments.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-activity\">\n              <i class=\"openwebicons-activity\">\u00a0openwebicons-activity\n              <small>by\u00a0<a href=\"http:\/\/wiki.activitystrea.ms\/w\/page\/23953797\/Icon\" target=\"_blank\"><abbr title=\"Activity Streams\">AS<\/abbr>-Wiki<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/activitystrea.ms\/\" target=\"_blank\">Activity Streams<\/a> specification defines a format for describing socially interesting events, or \"Activities\", about the people and things an individual may care about.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-geo\">\n              <i class=\"openwebicons-geo\">\u00a0openwebicons-geo\n              <small>by\u00a0<a href=\"http:\/\/geotagicons.com\/\" target=\"_blank\">GeotagIcons.com<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/geotagicons.com\/\" target=\"_blank\">Geotag Icon<\/a> is intended as a web \u201cstandard\u201d icon for identifying geotagged content to humans.<\/p><\/div><div class=\"col-md-4\"><p><a href=\"http:\/\/www.opensearch.org\/\" target=\"_blank\">OpenSearch<\/a> is a collection of simple formats for the sharing of search results.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-oauth\">\n              <i class=\"openwebicons-oauth\">\u00a0openwebicons-oauth\n            <\/i><\/h3><p>An <a href=\"http:\/\/oauth.net\/\" target=\"_blank\">open protocol<\/a> to allow secure API authorization in a simple and standard method from desktop and web applications.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-openid\">\n              <i class=\"openwebicons-openid\">\u00a0openwebicons-openid\n              <small>by\u00a0<a href=\"http:\/\/openid.net\/foundation\/news\/logos\/\" target=\"_blank\">OpenID<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/openid.net\/\">OpenID<\/a> is a safe, faster, and easier way to log in to web sites.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-rdf\">\n              <i class=\"openwebicons-rdf\">\u00a0openwebicons-rdf\n              <small>by\u00a0<a href=\"http:\/\/www.w3.org\/RDF\/icons\/\" target=\"_blank\">W3C<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.w3.org\/RDF\/\" target=\"_blank\">RDF<\/a> is a standard model for data interchange on the Web. RDF has features that facilitate data merging even if the underlying schemas differ, and it specifically supports the evolution of schemas over time without requiring all the data consumers to be changed.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-json-ld\">\n              <i class=\"openwebicons-json-ld\">\u00a0openwebicons-json-ld\n              <small>by\u00a0<a href=\"http:\/\/json-ld.org\/images\/\" target=\"_blank\">JSON-LD<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/json-ld.org\" target=\"_blank\">JSON-LD<\/a> is a lightweight Linked Data format. It is based on the already successful JSON format and provides a way to help JSON data interoperate at Web-scale.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-semantic-web\">\n              <i class=\"openwebicons-semantic-web\">\u00a0openwebicons-semantic-web\n              <small>by\u00a0<a href=\"http:\/\/www.w3.org\/2007\/10\/sw-logos.html\" target=\"_blank\">W3C<\/a><\/small>\n            <\/i><\/h3><p>In addition to the classic \u201cWeb of documents\u201d W3C is helping to build a technology stack to support a \u201cWeb of data,\u201d the sort of data you find in databases. The ultimate goal of the Web of data is to enable computers to do more useful work and to develop systems that can support trusted interactions over the network. The term \"<a href=\"http:\/\/www.w3.org\/standards\/semanticweb\/\" target=\"_blank\">Semantic Web<\/a>\" refers to W3C\u2019s vision of the Web of linked data.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-dataportability\">\n              <i class=\"openwebicons-dataportability\">\u00a0openwebicons-dataportability\n              <small>a <abbr title=\"Trademark\">TM<\/abbr> of the <a href=\"http:\/\/www.dataportability.org\/\" target=\"_blank\">DataPortability Organisation<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.dataportability.org\/\" target=\"_blank\">Data portability<\/a> is the ability for people to reuse their data across interoperable applications. The DataPortability Project works to advance this vision by identifying, contextualizing and promoting efforts in the space.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-federated\">\n              <i class=\"openwebicons-federated\">\u00a0openwebicons-federated\n              <small>by\u00a0<a href=\"http:\/\/status.net\/2010\/06\/28\/federated-social-web-summit-2010-announced\" target=\"_blank\">Status.net<\/a><\/small>\n            <\/i><\/h3><p>To collaborate and build the future of the Internet together via the <a href=\"http:\/\/www.w3.org\/2005\/Incubator\/federatedsocialweb\/\" target=\"_blank\">federated social web<\/a> with OStatus and other open web technologies.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-open-web\">\n              <i class=\"openwebicons-open-web\">\u00a0openwebicons-open-web\n              <small>by the <a href=\"http:\/\/www.openwebfoundation.org\/\" target=\"_blank\">Open Web Foundation<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/www.openwebfoundation.org\/\" target=\"_blank\">Open Web Foundation<\/a> is an independent non-profit dedicated to the development and protection of open, non-proprietary specifications for web technologies.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-web-intents\">\n              <i class=\"openwebicons-web-intents\">\u00a0openwebicons-web-intents\n              <small>by\u00a0<a href=\"http:\/\/aerotwist.com\/\" target=\"_blank\">Paul Lewis<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/webintents.org\/\" target=\"_blank\">Web Intents<\/a> is a framework for client-side service discovery and inter-application communication. Web Intents puts the user in control of service integrations and makes the developer's life simple.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-xmpp\">\n              <i class=\"openwebicons-xmpp\">\u00a0openwebicons-xmpp\n              <small>by\u00a0XMPP<\/small>\n            <\/i><\/h3><p>The Extensible Messaging and Presence Protocol (<a href=\"http:\/\/xmpp.org\/\" target=\"_blank\">XMPP<\/a>) is an open technology for real-time communication, which powers a wide range of applications including instant messaging, presence, multi-party chat, voice and video calls, collaboration, lightweight middleware, content syndication, and generalized routing of XML data.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-qr\">\n              <i class=\"openwebicons-qr\">\u00a0openwebicons-qr\n            <\/i><\/h3><p>The <a href=\"http:\/\/www.denso-wave.com\/qrcode\/index-e.html\" target=\"_blank\">QR code<\/a> was a kind of two dimensional symbology [...] The QR code is that carry meaningful information in the vertical direction as well as the horizontal, hence the two-dimensional term.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-epub\">\n              <i class=\"openwebicons-epub\">\u00a0openwebicons-epub\n              <small>by\u00a0<a href=\"http:\/\/idpf.org\/epub\" target=\"_blank\">ePub<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/idpf.org\/epub\/30\/spec\/epub30-overview.html\" target=\"_blank\">EPUB\u00ae specification<\/a> is a distribution and interchange format standard for digital publications and documents. EPUB defines a means of representing, packaging and encoding structured and semantically enhanced Web content \u2014 including HTML5, CSS, SVG, images, and other resources \u2014 for distribution in a single-file format.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-opengraph\">\n              <i class=\"openwebicons-opengraph\">\u00a0openwebicons-opengraph\n              <small>by\u00a0<a href=\"http:\/\/ogp.me\/\" target=\"_blank\">Open Graph protocol<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/ogp.me\/\" target=\"_blank\">Open Graph protocol<\/a> enables any web page to become a rich object in a social graph. For instance, this is used on Facebook to allow any web page to have the same functionality as any other object on Facebook.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-foaf\">\n              <i class=\"openwebicons-foaf\">\u00a0openwebicons-foaf\n              <small>by\u00a0<a href=\"http:\/\/www.foaf-project.org\/\" target=\"_blank\">FOAF<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/www.foaf-project.org\/\" target=\"_blank\">Friend of a Friend (FOAF)<\/a> project is creating a Web of machine-readable pages describing people, the links between them and the things they create and do; it is a contribution to the linked information system known as the Web.<\/p><\/div><div class=\"col-md-4\"><p><a href=\"http:\/\/informationcard.net\/\" target=\"_blank\">Information Cards<\/a> are the digital version of the cards you carry in your purse or wallet today. You use them with a new kind of digital wallet called a selector.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-browserid\">\n              <i class=\"openwebicons-browserid\">\u00a0openwebicons-browserid\n              <small>by\u00a0<a href=\"https:\/\/github.com\/mozilla\/browserid\/tree\/dev\/resources\/assets\" target=\"_blank\">Mozilla<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/browserid.org\/\" target=\"_blank\">BrowserID<\/a> is an easy way to sign into websites. Instead of making up a hard-to-remember username and password for each site, BrowserID lets you use your email address and a single password to sign in to any site that supports it.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-persona\">\n              <i class=\"openwebicons-persona\">\u00a0openwebicons-persona\n              <small>by\u00a0<a href=\"http:\/\/people.mozilla.org\/~smartell\/persona\/\" target=\"_blank\">Mozilla<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.mozilla.org\/en-US\/persona\/\" target=\"_blank\">Persona<\/a> allows you to sign in to sites using an email address you choose. So, instead of having to manage multiple usernames and passwords across your favorite sites and devices, you\u2019ll have more time to do the important stuff. We\u2019ll manage the details!<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-remote-storage\">\n              <i class=\"openwebicons-remote-storage\">\u00a0openwebicons-remote-storage\n              <small>by\u00a0<a href=\"http:\/\/unhosted.org\" target=\"_blank\">Unhosted<\/a><\/small>\n            <\/i><\/h3><p>Adding WebFinger, OAuth and Cross-Origin Resource Sharing (CORS) to an online storage makes it usable as per-user storage for web apps. The <a href=\"http:\/\/www.w3.org\/community\/unhosted\/wiki\/RemoteStorage\" target=\"_blank\">remoteStorage specification<\/a> describes a common interface for such a per-user online data storage.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-odata\">\n              <i class=\"openwebicons-odata\">\u00a0openwebicons-odata\n              <small>by\u00a0<a href=\"http:\/\/www.odata.org\/libraries\" target=\"_blank\">OData<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/www.odata.org\/\" target=\"_blank\">Open Data Protocol (OData)<\/a> is a Web protocol for querying and updating data that provides a way to unlock your data and free it from silos that exist in applications today.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-markdown\">\n              <i class=\"openwebicons-markdown\">\u00a0openwebicons-markdown\n              <small>by\u00a0<a href=\"https:\/\/github.com\/dcurtis\/markdown-mark\" target=\"_blank\">Dustin Curtis<\/a><\/small>\n            <\/i><\/h3><p>Use this mark to identify <a href=\"http:\/\/daringfireball.net\/projects\/markdown\/\" target=\"_blank\">Markdown<\/a>. It should be used to identify user input areas which support Markdown-compiled HTML output or to identify general Markdown support.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-tosdr\">\n              <i class=\"openwebicons-tosdr\">\u00a0openwebicons-tosdr\n              <small>by\u00a0<a href=\"https:\/\/github.com\/didnotread\/didnotread.org\/tree\/gh-pages\/img\" target=\"_blank\">Terms of Service; Didn't Read<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/tos-dr.info\" target=\"_blank\">Terms of Service; Didn't read!<\/a> \"I have read and agree to the Terms\" is the biggest lie on the web. We aim to fix that. We are a user rights initiative to rate and label website terms &amp; privacy policies, from very good \"Class A\" to very bad \"Class E\".<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-indieweb\">\n              <i class=\"openwebicons-indieweb\">\u00a0openwebicons-indieweb\n              <small>by\u00a0<a href=\"http:\/\/indiewebcamp.com\/press-kit\" target=\"_blank\">IndieWebCamp<\/a><\/small>\n            <\/i><\/h3><p>Rather than posting content on third-party silos of content, we should all own the content we're creating. Publish on your own domain, and syndicate out to silos. This is the basis of the \"Indie Web\" movement.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-indiewebcamp\">\n              <i class=\"openwebicons-indiewebcamp\">\u00a0openwebicons-indiewebcamp\n              <small>by\u00a0<a href=\"http:\/\/indiewebcamp.com\/press-kit\" target=\"_blank\">IndieWebCamp<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/indiewebcamp.com\/\" target=\"_blank\">IndieWebCamp<\/a> is a regular gathering of web creators building and sharing open web technologies to advance the state of the indie web, and empower users to own their own identities &amp; content on the web.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-git\">\n              <i class=\"openwebicons-git\">\u00a0openwebicons-git\n              <small>by\u00a0<a href=\"http:\/\/git-scm.com\/downloads\/logos\" target=\"_blank\">Git SCM<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/git-scm.com\" target=\"_blank\">Git<\/a> is a free and open source distributed version control system designed to handle everything from small to very large projects with speed and efficiency.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-webhooks\">\n              <i class=\"openwebicons-webhooks\">\u00a0openwebicons-webhooks\n              <small>by\u00a0<a href=\"http:\/\/webhooks.org\/\" target=\"_blank\">webhooks.org<\/a><\/small>\n            <\/i><\/h3><p>A <a href=\"http:\/\/webhooks.org\/\" target=\"_blank\">WebHook<\/a> is an HTTP callback: an HTTP POST that occurs when something happens; a simple event-notification via HTTP POST.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-osi\">\n              <i class=\"openwebicons-osi\">\u00a0openwebicons-osi\n              <small>by\u00a0<a href=\"http:\/\/opensource.org\/node\/442\" target=\"_blank\"><abbr title=\"Open Source Initiative\">OSI<\/abbr><\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"http:\/\/opensource.org\" target=\"_blank\">Open Source Initiative (OSI)<\/a> is a non-profit corporation with global scope formed to educate about, and advocate, for the benefits of open source, and to build bridges among different constituencies in the open source community.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-webfinger\">\n              <i class=\"openwebicons-webfinger\">\u00a0openwebicons-webfinger\n              <small>by\u00a0<a href=\"http:\/\/hueniverse.com\" target=\"_blank\">Eran Hammer<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/webfinger.net\" target=\"_blank\">WebFinger<\/a> is used to discover information about people or other entities on the Internet that are identified by a URI or IRI using standard Hypertext Transfer Protocol (HTTP) methods over a secure transport.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-bitcoin\">\n              <i class=\"openwebicons-bitcoin\">\u00a0openwebicons-bitcoin\n              <small>by\u00a0<a href=\"http:\/\/bitcoin.org\/\" target=\"_blank\">bitcoin<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/bitcoin.org\/\" target=\"_blank\">Bitcoin<\/a> is a digital currency, a protocol, and a software that enables Instant peer-to-peer transactions, worldwide payments, almost no processing fees, and much more.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-pubsubhubbub\">\n              <i class=\"openwebicons-pubsubhubbub\">\u00a0openwebicons-pubsubhubbub\n              <small>by\u00a0<a href=\"https:\/\/code.google.com\/p\/pubsubhubbub\/\" target=\"_blank\">PubSubHubbub<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/code.google.com\/p\/pubsubhubbub\/\" target=\"_blank\">PubSubHubbub<\/a> is a simple, open, server-to-server webhook-based pubsub (publish\/subscribe) protocol for any web accessible resources.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-svg\">\n              <i class=\"openwebicons-svg\">\u00a0openwebicons-svg\n              <small>by\u00a0<a href=\"http:\/\/www.w3.org\/2009\/08\/svg-logos.html\" target=\"_blank\">W3C<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.w3.org\/TR\/SVG\/\" target=\"_blank\">SVG<\/a> is a language for describing two-dimensional graphics in XML. SVG allows for three types of graphic objects: vector graphic shapes (e.g., paths consisting of straight lines and curves), images and text.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-tent\">\n              <i class=\"openwebicons-tent\">\u00a0openwebicons-tent\n              <small>by\u00a0<a href=\"https:\/\/tent.io\/trademark-guidelines\" target=\"_blank\">Tent.io<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/tent.io\/\" target=\"_blank\">Tent<\/a> is a communications protocol. It lets users store any kind of data and send that data to other users in real time. Just like other protocol-based services like email or the web, anyone can host their own server or write apps. Also like email, users can host their own Tent server or use a Tent hosting provider.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-copyleft\">\n              <i class=\"openwebicons-copyleft\">\u00a0openwebicons-copyleft\n              <small>by\u00a0<a href=\"http:\/\/en.wikipedia.org\/wiki\/File:Copyleft.svg\" target=\"_blank\">Wikipedia<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/www.gnu.org\/copyleft\/\" target=\"_blank\">Copyleft<\/a> is a general method for making a program (or other work) free, and requiring all modified and extended versions of the program to be free as well.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-gnu\">\n              <i class=\"openwebicons-gnu\">\u00a0openwebicons-gnu\n              <small>by\u00a0<a href=\"http:\/\/www.gnu.org\/graphics\/gnusvgart.html\" target=\"_blank\">GNU<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/www.gnu.org\" target=\"_blank\">GNU<\/a> is a Unix-like operating system that is free software\u2014it respects your freedom. You can install versions of GNU (more precisely, GNU\/Linux systems) which are entirely free software.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-webmention\">\n              <i class=\"openwebicons-webmention\">\u00a0openwebicons-webmention\n              <small>by\u00a0<a href=\"http:\/\/webmention.io\" target=\"_blank\">Aaron Parecki<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/webmention.org\/\" target=\"_blank\">WebMention<\/a> is a simple way to automatically notify any URL when you link to it on your site. From the receivers perpective, it's a way to request notification when other sites link to it. It\u2019s a modern alternative to Pingback and other forms of Linkback.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-javascript\">\n              <i class=\"openwebicons-javascript\">\u00a0openwebicons-javascript\n              <small>by\u00a0<a href=\"https:\/\/github.com\/voodootikigod\/logo.js\/blob\/master\/registry.md\" target=\"_blank\">logo.js<\/a><\/small>\n            <\/i><\/h3><p>Officially announced at <a href=\"http:\/\/jsconf.com\" target=\"_blank\">JSConf EU 2011<\/a>, but used for almost a year and half prior by the community, we are just going to offer this logo for use with JS projects. If you like it use it, if you don't that's cool too. Just don't complain unless you want to offer a different one, post your alternative to twitter with the tag #jslogo.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-wtfpl\">\n              <i class=\"openwebicons-wtfpl\">\u00a0openwebicons-wtfpl\n              <small>by\u00a0<a href=\"http:\/\/www.wtfpl.net\/download\/\" target=\"_blank\">WTFPL<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.wtfpl.net\/\" target=\"_blank\">WTFPL<\/a> - The <em>Do What The Fuck You Want To Public License<\/em> (WTFPL) is a free software license.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-wordpress\">\n              <i class=\"openwebicons-wordpress\">\u00a0openwebicons-wordpress\n              <small>by\u00a0<a href=\"https:\/\/wordpress.org\/about\/logos\/\" target=\"_blank\">WordPress.org<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/wordpress.org\/\" target=\"_blank\">WordPress<\/a> started in 2003 with a single bit of code to enhance the typography of everyday writing and with fewer users than you can count on your fingers and toes. Since then it has grown to be the largest self-hosted blogging tool in the world, used on millions of sites and seen by tens of millions of people every day.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-owncloud\">\n              <i class=\"openwebicons-owncloud\">\u00a0openwebicons-owncloud\n              <small>by\u00a0<a href=\"http:\/\/owncloud.org\/press\/\" target=\"_blank\">ownCloud<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/owncloud.org\/\" target=\"_blank\">ownCloud<\/a> gives you universal access to your files through a web interface or WebDAV. It also provides a platform to easily view &amp; sync your contacts, calendars and bookmarks across all your devices and enables basic editing right on the web.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-known\">\n              <i class=\"openwebicons-known\">\u00a0openwebicons-known\n              <small>by\u00a0<a href=\"https:\/\/withknown.com\" target=\"_blank\">Known<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/withknown.com\" target=\"_blank\">Known<\/a> is an open publishing and collaboration platform that allows anyone to share their stories using many types of media. Whether you are publishing as a group, or just as an individual, Known allows you to post to a site that you really own, while reaching your audience whever they might be.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-barcamp\">\n              <i class=\"openwebicons-barcamp\">\u00a0openwebicons-barcamp\n              <small>by\u00a0<a href=\"http:\/\/barcamp.org\/w\/page\/405009\/SpreadBarCamp\" target=\"_blank\">BarCamp.org<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.barcamp.org\" target=\"_blank\">BarCamp<\/a> is an international network of user-generated conferences primarily focused around technology and the web. They are open, participatory workshop-events, the content of which is provided by participants.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-indiehosters\">\n              <i class=\"openwebicons-indiehosters\">\u00a0openwebicons-indiehosters\n              <small>by\u00a0<a href=\"https:\/\/indiehosters.net\" target=\"_blank\">IndieHosters.net<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/indiehosters.net\" target=\"_blank\">IndieHosters<\/a> is a new network of independent hosters who offer CMS Software as a Service.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-cccs\">\n              <i class=\"openwebicons-cccs\">\u00a0openwebicons-cccs\n              <small>by\u00a0<a href=\"https:\/\/www.c3s.cc\/en\/press-area\/\" target=\"_blank\">C3S<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/www.c3s.cc\/\" target=\"_blank\">Cultural Commons Collecting Society<\/a> SCE mit beschr\u00e4nkter Haftung (C3S SCE) \u2013 is a European co-operative. It is intended to be founded as a European non-profit collecting society managing musical works which are released under Creative Commons non-commercial (CC NC) and other free licences, including those under no licence at all.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-mozilla\">\n              <i class=\"openwebicons-mozilla\">\u00a0openwebicons-mozilla\n              <small>by\u00a0<a href=\"https:\/\/blog.mozilla.org\/press\/media-library\/\" target=\"_blank\">mozilla<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/www.mozilla.org\/\" target=\"_blank\">The Mozilla Corporation<\/a> was established in August 2005 as a wholly owned taxable subsidiary that serves the non-profit, public benefit goals of its parent, the Mozilla Foundation, and the vast Mozilla community.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-unhosted\">\n              <i class=\"openwebicons-unhosted\">\u00a0openwebicons-unhosted\n              <small>by\u00a0<a href=\"https:\/\/github.com\/unhosted\/design\" target=\"_blank\">unhosted<\/a><\/small>\n            <\/i><\/h3><p>Also known as \"serverless\", \"client-side\", or \"static\" web apps, <a href=\"https:\/\/unhosted.org\" target=\"_blank\">unhosted<\/a> web apps do not send your user data to their server. Either you connect your own server at runtime, or your data stays within the browser.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-indie\">\n              <i class=\"openwebicons-indie\">\u00a0openwebicons-indie\n              <small>by\u00a0<a href=\"https:\/\/ind.ie\/style-guide\/\" target=\"_blank\">Ind.ie<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/ind.ie\/\" target=\"_blank\">Ind.ie<\/a> is a tiny social enterprise working to protect human rights and democracy in the digital age. Our activities include raising awareness of the issues, advising policy makers and educators, and creating free and open independent alternatives to mainstream technologies.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-ghost\">\n              <i class=\"openwebicons-ghost\">\u00a0openwebicons-ghost\n              <small>by\u00a0<a href=\"https:\/\/ghost.org\/about\/logos\/\" target=\"_blank\">Ghost<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/ghost.org\/\" target=\"_blank\">Ghost<\/a> is a simple, powerful publishing platform that allows you to share your stories with the world.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-diaspora\">\n              <i class=\"openwebicons-diaspora\">\u00a0openwebicons-diaspora\n              <small>by\u00a0<a href=\"https:\/\/wiki.diasporafoundation.org\/File:Diaspora_logo_roboto.svg\" target=\"_blank\">Diaspora<\/a><\/small>\n            <\/i><\/h3><p>Use <a href=\"https:\/\/diasporafoundation.org\" target=\"_blank\">diaspora*<\/a> as your home base to post to your profiles on other major social services. This way your friends will still be able to keep in touch with what you\u2019re up to, even if they\u2019re not yet on diaspora*.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-webplattform\">\n              <i class=\"openwebicons-webplattform\">\u00a0openwebicons-webplattform\n              <small>by\u00a0<a href=\"http:\/\/www.webplatform.org\/logo\/\" target=\"_blank\">WebPlattform<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/www.webplatform.org\/\" target=\"_blank\">Web Platform<\/a> is a community-driven site that aims to become a comprehensive and authoritative source for web developer documentation.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-opensourcedesign\">\n              <i class=\"openwebicons-opensourcedesign\">\u00a0openwebicons-opensourcedesign\n              <small>by\u00a0<a href=\"http:\/\/opensourcedesign.net\" target=\"_blank\">Open Source Design<\/a><\/small>\n            <\/i><\/h3><p>We are a community of designers and developers pushing more open design processes and improving the user experience and interface design of open source software.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-opendesign\">\n              <i class=\"openwebicons-opendesign\">\u00a0openwebicons-opendesign\n              <small>by\u00a0<a href=\"http:\/\/opendesign.foundation\" target=\"_blank\">OpenDesign<\/a><\/small>\n            <\/i><\/h3><p>We think the best way to improve the design industry is to work in the open.<\/p><\/div><div class=\"col-md-4\"><p><a href=\"http:\/\/gnu.io\/social\/\" target=\"_blank\">GNU social<\/a> is a continuation of the StatusNet project. It is social communication software for both public and private communications. It is widely supported and has a large userbase.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-mastodon\">\n              <i class=\"openwebicons-mastodon\">\u00a0openwebicons-mastodon\n              <small>by\u00a0<a href=\"https:\/\/joinmastodon.org\" target=\"_blank\">Mastodon<\/a><\/small>\n            <\/i><\/h3><p>With thousands of interconnected communities to choose from, and the tools to make your own, Mastodon is the world\u2019s largest free, open-source, decentralized microblogging network.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-mastodon-simple\">\n              <i class=\"openwebicons-mastodon-simple\">\u00a0openwebicons-mastodon-simple\n              <small>by\u00a0<a href=\"https:\/\/joinmastodon.org\" target=\"_blank\">Mastodon<\/a><\/small>\n            <\/i><\/h3><p>...<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-activitypub\">\n              <i class=\"openwebicons-activitypub\">\u00a0openwebicons-activitypub\n              <small>by\u00a0<a href=\"https:\/\/activitypub.rocks\/news\/new-tutorial-new-logo.html\" target=\"_blank\">ActivityPub.rocks<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/www.w3.org\/TR\/activitypub\/\" target=\"_blank\">ActivityPub<\/a> is a decentralized social networking protocol based on the ActivityStreams 2.0 data format. ActivityPub is an official W3C recommended standard published by the W3C Social Web Working Group.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-wapuu\">\n              <i class=\"openwebicons-wapuu\">\u00a0openwebicons-wapuu\n              <small>by\u00a0<a href=\"https:\/\/ma.tt\/2015\/12\/wapuus-history\/\" target=\"_blank\">Wapuu<\/a><\/small>\n            <\/i><\/h3><p>This cute little Japanese monster had its start in 2009 at the WordCamp Tokyo after-party. Matt Mullenweg asked Japanese users for ideas to promote WordPress in Japan. One suggestion was to make a mascot in Japan for WordPress.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-micro-blog\">\n              <i class=\"openwebicons-micro-blog\">\u00a0openwebicons-micro-blog\n              <small>by\u00a0<a href=\"http:\/\/help.micro.blog\/2018\/resources\/\" target=\"_blank\">Micro.blog<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/micro.blog\" target=\"_blank\">Micro.blog<\/a> encourages publishing at your own domain name, where you can control your own content, but it still integrates posts into a familiar timeline user interface, with centralized replies, favorites, and an open API based on JSON Feed and IndieWeb standards.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-indieauth\">\n              <i class=\"openwebicons-indieauth\">\u00a0openwebicons-indieauth\n              <small>by\u00a0<a href=\"https:\/\/indieauth.com\" target=\"_blank\">IndieAuth<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/www.w3.org\/TR\/indieauth\/\" target=\"_blank\">IndieAuth<\/a> is an identity layer on top of OAuth 2.0, primarily used to obtain an OAuth 2.0 Bearer Token for use by [Micropub] clients. End-Users and Clients are all represented by URLs. IndieAuth enables Clients to verify the identity of an End-User, as well as to obtain an access token that can be used to access resources under the control of the End-User.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-friendica\">\n              <i class=\"openwebicons-friendica\">\u00a0openwebicons-friendica\n              <small>by\u00a0<a href=\"https:\/\/github.com\/friendica\/friendica\" target=\"_blank\">Friendi.ca<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/friendi.ca\/\" target=\"_blank\">Friendica<\/a> is a software to create a distributed social network. It is part of the federation of servers that are running with Friendica, Diaspora and Hubzilla which are compatible to each other.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-jsonfeed\">\n              <i class=\"openwebicons-jsonfeed\">\u00a0openwebicons-jsonfeed\n              <small>by\u00a0<a href=\"https:\/\/github.com\/brentsimmons\/JSONFeed\/tree\/master\/graphics\" target=\"_blank\">JSONFeed<\/a><\/small>\n            <\/i><\/h3><p>The <a href=\"https:\/\/jsonfeed.org\/\" target=\"_blank\">JSON Feed<\/a> format is a pragmatic syndication format, like RSS and Atom, but with one big difference: it\u2019s JSON instead of XML.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-pump-io\">\n              <i class=\"openwebicons-pump-io\">\u00a0openwebicons-pump-io\n              <small>by\u00a0<a href=\"https:\/\/de.wikipedia.org\/wiki\/Datei:Pump.io.svg\" target=\"_blank\">Pump.io<\/a><\/small>\n            <\/i><\/h3><p><a href=\"http:\/\/pump.io\/\" target=\"_blank\">Pump.io<\/a> is a stream server that does most of what people really want from a social network.<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><h3 id=\"openwebicons-hubzilla\">\n              <i class=\"openwebicons-hubzilla\">\u00a0openwebicons-hubzilla\n              <small>by\u00a0<a href=\"https:\/\/treer.gitlab.io\/hubzilla-graphics\/example.html\" target=\"_blank\">Hubzilla<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/hubzilla.org\/\" target=\"_blank\">Hubzilla<\/a> is a powerful platform for creating interconnected websites featuring a decentralized identity, communications, and permissions framework built using common webserver technology.<\/p><\/div><div class=\"col-md-4\"><h3 id=\"openwebicons-fediverse\">\n              <i class=\"openwebicons-fediverse\">\u00a0openwebicons-fediverse\n              <small>by\u00a0<a href=\"https:\/\/commons.wikimedia.org\/wiki\/File:Fediverse_logo_proposal_(mono_version).svg\" target=\"_blank\">Fediverse<\/a><\/small>\n            <\/i><\/h3><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fediverse\" target=\"_blank\">Fediverse<\/a> (a portmanteau of \"federation\" and \"universe\") is the ensemble of federated servers that are used for web publishing (i.e. social networking, microblogging or websites) and file hosting.<\/p><\/div><\/div><\/section><section id=\"html5-icons\"><div class=\"row\"><div class=\"col-md-12\"><h2><i class=\"openwebicons-html5\"> <em>HTML5<\/em> Icons<\/i><\/h2><p>The Web-Font is based on the original SVGs from the <em><a href=\"http:\/\/www.w3.org\/html\/logo\/\" target=\"_blank\">HTML5 Logo Site<\/a><\/em> by the W3C.\n            <\/p><table class=\"table table-bordered table-striped\"><thead><tr><th>Icon<\/th>\n                  <th>class-name<\/th>\n                  <th>About<\/th>\n                <\/tr><\/thead><tbody><tr><td><i class=\"openwebicons-html5\"> <\/i><\/td>\n                  <td>openwebicons-html5<\/td>\n                  <td><a href=\"http:\/\/www.whatwg.org\/html\/\" target=\"_blank\">HTML5<\/a> contains powerful capabilities for Web-based applications with more powerful interaction,\n                    video support, graphics, more styling effects, and a full set of APIs.\n                    HTML5 adapts to any device, whether desktop, mobile, tablet, or television. HTML5 is an open platform developed under royalty free licensing terms. <\/td>\n                <\/tr><tr><td><i class=\"openwebicons-css3\"> <\/i><\/td>\n                  <td>openwebicons-css3<\/td>\n                  <td><a href=\"http:\/\/www.w3.org\/Style\/CSS\/current-work\" target=\"_blank\">CSS Level 3<\/a> builds on CSS Level 2 module by module, using the CSS2.1 specification\n                    as its core. Each module adds functionality and\/or replaces part of the CSS2.1 specification. The CSS Working Group intends that the new CSS modules will not\n                    contradict the CSS2.1 specification: only that they will add functionality and refine definitions.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-connectivity\"> <\/i><\/td>\n                  <td>openwebicons-connectivity<\/td>\n                  <td>More efficient connectivity means more real-time chats, faster games, and better communication. Web Sockets and Server-Sent Events are pushing\n                    (pun intended) data between client and server more efficiently than ever before. This icon can also be used for <a href=\"http:\/\/webintents.org\/\" target=\"_blank\"><em>Web Intents<\/em><\/a>\n                    or <a href=\"http:\/\/xauth.org\/\" target=\"_blank\">XAuth<\/a>.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-semantics\"> <\/i><\/td>\n                  <td>openwebicons-semantics<\/td>\n                  <td>Giving meaning to structure, semantics are front and center with HTML5. A richer set of tags, along with <a href=\"http:\/\/www.w3.org\/TR\/rdfa-core\/\" target=\"_blank\">RDFa<\/a>,\n                    <a href=\"http:\/\/www.w3.org\/TR\/html5\/microdata.html\" target=\"_blank\">microdata<\/a>, and <a href=\"http:\/\/microformats.org\/\" target=\"_blank\">microformats<\/a>, are enabling a more\n                    useful, data driven web for both programs and your users.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-device-access\"> <\/i><\/td>\n                  <td>openwebicons-device-access<\/td>\n                  <td>Beginning with the Geolocation API, Web Applications can present rich, device-aware features and experiences. Incredible device access innovations are being developed and implemented, from audio\/video input access to microphones and cameras, to local data such as contacts &amp; events, and even tilt orientation.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-offline-storage\"> <\/i><\/td>\n                  <td>openwebicons-offline-storage<\/td>\n                  <td>Web Apps can start faster and work even if there is no internet connection, thanks to the HTML5 App Cache, as well as the Local Storage, Indexed DB, and the File API specifications.<\/td>\n                <\/tr><tr><td> <\/td>\n                  <td>openwebicons-multimedia<\/td>\n                  <td>Audio and video are first class citizens in the HTML5 web, living in harmony with your apps and sites. Lights, camera, action!<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-3deffects\"> <\/i><\/td>\n                  <td>openwebicons-3deffects<\/td>\n                  <td>Between SVG, Canvas, WebGL, and CSS3 3D features, you're sure to amaze your users with stunning visuals natively rendered in the browser.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-perfintegration\"> <\/i><\/td>\n                  <td>openwebicons-perfintegration<\/td>\n                  <td>Make your Web Apps and dynamic web content faster with a variety of techniques and technologies such as Web Workers and XMLHttpRequest 2. No user should ever wait on your watch.<\/td>\n                <\/tr><\/tbody><\/table><p>This work is attributed to the W3C. Suggested markup: <code>&lt;span&gt;HTML5 Logo by &lt;a href=\"http:\/\/www.w3.org\/\"&gt;&lt;abbr title=\"World Wide Web Consortium\"&gt;W3C&lt;\/abbr&gt;&lt;\/a&gt;&lt;\/span&gt;<\/code>.<\/p><\/div><\/div><\/section><section id=\"cc-icons\"><div class=\"row\"><div class=\"col-md-12\"><h2><i class=\"openwebicons-cc\"> <em>Creative Commons<\/em> Icons<\/i><\/h2><p>The Web-Font is made using the original SVGs from the <em><a href=\"http:\/\/creativecommons.org\/about\/downloads\" target=\"_blank\">Creative Commons<\/a><\/em> Project. If you need a Font for one of your Text\/Word-Dokuments, there is also an official TrueType-Font: <a href=\"http:\/\/creativecommons.org\/about\/downloads\">http:\/\/creativecommons.org\/about\/downloads<\/a><\/p><table class=\"table table-bordered table-striped\"><thead><tr><th>Icon<\/th>\n                  <th>class-name<\/th>\n                  <th>About<\/th>\n                <\/tr><\/thead><tbody><tr><td><i class=\"openwebicons-cc\"><\/i><\/td>\n                  <td>openwebicons-cc<\/td>\n                  <td><a href=\"http:\/\/creativecommons.org\/\" target=\"_blank\">Creative Commons<\/a> is a nonprofit organization that enables the sharing and use of creativity and knowledge through free legal tools.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-by\"><\/i><\/td>\n                  <td>openwebicons-cc-by<\/td>\n                  <td><strong>Attribution<\/strong> \u2014 You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work).<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-nc\"><\/i><\/td>\n                  <td>openwebicons-cc-nc<\/td>\n                  <td><strong>Noncommercial<\/strong> \u2014 You may not use this work for commercial purposes.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-nc-eu\"><\/i><\/td>\n                  <td>openwebicons-cc-nc-eu<\/td>\n                  \n                <\/tr><tr><td><i class=\"openwebicons-cc-nc-jp\"><\/i><\/td>\n                  <td>openwebicons-cc-nc-jp<\/td>\n                  \n                <\/tr><tr><td><i class=\"openwebicons-cc-sa\"><\/i><\/td>\n                  <td>openwebicons-cc-sa<\/td>\n                  <td><strong>Share Alike<\/strong> \u2014 If you alter, transform, or build upon this work, you may distribute the resulting work only under the same or similar license to this one.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-nd\"><\/i><\/td>\n                  <td>openwebicons-cc-nd<\/td>\n                  <td><strong>No Derivative Works<\/strong> \u2014 You may not alter, transform, or build upon this work.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-public\"><\/i><\/td>\n                  <td>openwebicons-cc-public<\/td>\n                  <td><strong>Public Domain<\/strong> \u2014 Where the work or any of its elements is in the <a href=\"http:\/\/wiki.creativecommons.org\/Public_domain\" target=\"_blank\">public domain<\/a> under applicable law, that status is in no way affected by the license.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-zero\"><\/i><\/td>\n                  <td>openwebicons-cc-zero<\/td>\n                  <td><a href=\"http:\/\/creativecommons.org\/about\/cc0\" target=\"_blank\">CC0<\/a> enables scientists, educators, artists and other creators and owners of copyright- or database-protected content to waive those interests in their works and thereby place them as completely as possible in the public domain, so that others may freely build upon, enhance and reuse the works for any purposes without restriction under copyright or database law.<\/td>\n                <\/tr><tr><td>\n                  <\/td><td>openwebicons-cc-share<\/td>\n                  <td><strong>to Share<\/strong> \u2014 to copy, distribute and transmit the work<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-cc-remix\"><\/i><\/td>\n                  <td>openwebicons-cc-remix<\/td>\n                  <td><strong>to Remix<\/strong> \u2014 to adapt the work<\/td>\n                <\/tr><\/tbody><\/table><p>From the <a href=\"http:\/\/creativecommons.org\/policies\" target=\"_blank\"><em>Creative Commons<\/em> Policy<\/a>:<\/p><blockquote><div><strong>Creative Commons License Buttons:<\/strong> The Creative Commons buttons that describe a key term of our license, such as BY, NC, ND, SA, Sampling, Sampling Plus and Noncommercial Sampling Plus may only be used in the context of pointing to a Creative Commons license on the Creative Commons server that includes that license term or to otherwise describe the Creative Commons license, that includes that license term and that applies to a particular work.<\/div><\/blockquote><\/div><\/div><\/section><section id=\"microformats-icons\"><div class=\"row\"><div class=\"col-md-12\"><h2><i class=\"openwebicons-microformats\"> Microformats Icons<\/i><\/h2><p>The microformats-icons are based on the work of <a href=\"http:\/\/bartelme.at\/journal\/archive\/microformats_icons\" target=\"_blank\">Wolfgang Bartelme<\/a> and Chris Messina.\n            You can find the .png files on the <a href=\"http:\/\/microformats.org\/wiki\/icons\" target=\"_blank\">microformats-wiki<\/a>.<\/p><table class=\"table table-bordered table-striped\"><thead><tr><th>Icon<\/th>\n                  <th>class-name<\/th>\n                  <th>About<\/th>\n                <\/tr><\/thead><tbody><tr><td><i class=\"openwebicons-microformats\"> <\/i><\/td>\n                  <td>openwebicons-microformats<\/td>\n                  <td><a href=\"http:\/\/microformats.org\/\" target=\"_blank\">microformats<\/a> are extensions to HTML for marking up people, events, blog posts, reviews, recipes etc.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-hcard-download\"><\/i><\/td>\n                  <td>openwebicons-hcard-download<\/td>\n                  <td><a href=\"http:\/\/microformats.org\/wiki\/hcard\" target=\"_blank\">hCard<\/a> is a simple, open format for publishing people, companies, organizations on the web, using a 1:1 representation of vCard (<a href=\"http:\/\/microformats.org\/wiki\/rfc-2426\" title=\"rfc-2426\" target=\"_blank\">RFC2426<\/a>) properties and values in HTML. hCard is one of several open <a href=\"https:\/\/pfefferle.github.io\/wiki\/microformats\" title=\"microformats\">microformat<\/a> standards suitable for embedding data in HTML\/HTML5, and Atom\/RSS\/XHTML or other XML.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-hcard-add\"><\/i><\/td>\n                  <td>openwebicons-hcard-add<\/td>\n                  \n                <\/tr><tr><td><i class=\"openwebicons-hresume\"><\/i><\/td>\n                  <td>openwebicons-hresume<\/td>\n                  <td>The <a href=\"http:\/\/microformats.org\/wiki\/hresume\" target=\"_blank\">hResume<\/a> format is based on a set of fields common to numerous resumes published today on the web. Where possible field names have been chosen and reused from preexisting microformats.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-hatom\"><\/i><\/td>\n                  <td>openwebicons-hatom<\/td>\n                  <td><a href=\"http:\/\/microformats.org\/wiki\/hatom\" target=\"_blank\">hAtom<\/a> is a microformat for identifying semantic information in weblog posts and practically any other place Atom may be used, such as news articles.<\/td>\n                <\/tr><\/tbody><\/table><\/div><\/div><\/section><section id=\"openfontlicense-icons\"><div class=\"row\"><div class=\"col-md-12\"><h2><i class=\"openwebicons-ofl\"> SIL Open Font License - Icons<\/i><\/h2><p>The <a href=\"http:\/\/scripts.sil.org\/OFL\" target=\"_blank\">SIL Open Font License (OFL)<\/a> is a free, libre and open source license specifically designed for fonts and related software based on our experience in font design and linguistic software engineering.<\/p><p>The Web-Font is made using the SVGs from the <a href=\"http:\/\/oflb.open-fonts.org\" target=\"_blank\">Open Font Community<\/a>.<\/p><table class=\"table table-bordered table-striped\"><thead><tr><th>Icon<\/th>\n                  <th>class-name<\/th>\n                  <th>About<\/th>\n                <\/tr><\/thead><tbody><tr><td><i class=\"openwebicons-ofl\"> <\/i><\/td>\n                  <td>openwebicons-ofl<\/td>\n                  \n                <\/tr><tr><td><i class=\"openwebicons-ofl-attribution\"> <\/i><\/td>\n                  <td>openwebicons-ofl-attribution<\/td>\n                  <td>The requirement is for proper attribution of the author(s): name(s) and notice(s) must be preserved and abuse of the name(s) and reputation of the author(s) is forbidden.<\/td>\n                <\/tr><tr><td>\n                  <\/td><td>openwebicons-ofl-share<\/td>\n                  <td>The requirement is for derivative works to remain under the same license to encourage fair collaboration and prevent anyone from locking away contributions.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-ofl-embedding\"><\/i><\/td>\n                  <td>openwebicons-ofl-embedding<\/td>\n                  <td>The permission is for fonts to be embedded in any kind of document. This does not affect the licensing status of the document but makes it easier for documents to be used in different environments.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-ofl-renaming\"><\/i><\/td>\n                  <td>openwebicons-ofl-renaming<\/td>\n                  <td>The requirement is for derivative fonts to be renamed to allow branching while retaining artistic integrity.<\/td>\n                <\/tr><tr><td><i class=\"openwebicons-ofl-selling\"><\/i><\/td>\n                  <td>openwebicons-ofl-selling<\/td>\n                  <td>The requirement is for fonts to be bundled with software when they are sold. Fonts cannot be sold on their own. Redistribution without selling is not restricted.<\/td>\n                <\/tr><\/tbody><\/table><p>The spirit and working model of the OFL can be expressed in human-readable Creative Commons-like terminology using the following permits \/ requires elements and visual representations:<\/p><p>Please note that this terminology and visual representation is simply an expression of the working model of the license and has no legal value in itself. It is designed to help you understand and use the Open Font License in a similar way to the OFL FAQ. It is always intended to link back to the full license text of the OFL. Please note that although the terminology and visual representation of the OFL is based on work by Creative Commons, the OFL is not officially affiliated with Creative Commons.<\/p><p>Source: <a href=\"http:\/\/scripts.sil.org\/cms\/scripts\/page.php?site_id=nrsi&amp;id=OFL#9ccf5052\" target=\"_blank\">SIL Open Font License (OFL)<\/a>.<\/p><\/div><\/div><\/section><div class=\"modal fade\" id=\"icon-modal\"><div class=\"modal-dialog\"><div class=\"modal-content\"><p>\n              <\/p><h4 class=\"modal-title\">Modal title<\/h4><\/div><\/div><\/div>","created_at":"2018-12-19T12:12:03+0000","updated_at":"2018-12-19T12:12:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":19,"domain_name":"pfefferle.github.io","preview_picture":"http:\/\/pfefferle.github.io\/openwebicons\/img\/OpenWeb-Icons.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["latex","typesetting"],"is_public":false,"id":746,"title":"LaTeX Documentation","url":"https:\/\/www.latex-project.org\/help\/documentation\/","content":"<hr \/><div class=\"row\"><div class=\"col cell1of2\"><p>This page contains references to core documentation about LaTeX written by the\n    <a href=\"https:\/\/www.latex-project.org\/about\/team\/\">LaTeX team<\/a>. Articles on specific topics, talks,\n    etc. can be found on\n    <a href=\"https:\/\/www.latex-project.org\/publications\/indexbytopic\/\">the publication page<\/a> indexed\n    by topic.<\/p><p>In addition pointer to\n    documentation in other languages (usually developed and maintained by user groups)\n    is given. To find documentation that is\n    available elsewhere on the net, visit the <a href=\"https:\/\/www.latex-project.org\/help\/links\">links\n    page<\/a>. Also, there's a whole page\n    dedicated to <a href=\"https:\/\/www.latex-project.org\/help\/books\">books on LaTeX and related\n    topics<\/a>.<\/p><p>If you know about a good resource of TeX and related documentation not listed here or on\n    the <a href=\"https:\/\/www.latex-project.org\/help\/links\">links page<\/a>,\n    please <a href=\"https:\/\/www.latex-project.org\/contact\">contact us<\/a>.<\/p><\/div><div class=\"col cell1of2\"><ul class=\"toc\"><li><a href=\"#documentation-distributed-with-latex\">Distributed LaTeX documentation<\/a>\n        <ul><li><a href=\"#general-documentation\">General documentation<\/a><\/li>\n\t  <li><a href=\"#typesetting-complex-mathematics\">Typesetting mathematics<\/a><\/li>\n\t  <li><a href=\"#source-code-documentation\">Source code documentation<\/a><\/li>\n\t  <li><a href=\"#japanese-translations\">Japanese translations<\/a><\/li>\n        <\/ul><\/li>\t\n      <li><a href=\"#other-non-english-documentation\">Other non-english documentation resources<\/a>\n        <ul><li><a href=\"#german\">German<\/a><\/li>\n          <li><a href=\"#french\">French<\/a><\/li>\n        <\/ul><\/li>\n    <\/ul><\/div><\/div><hr \/><h2 id=\"documentation-distributed-with-latex\">Documentation distributed with LaTeX<\/h2><h3 id=\"general-documentation\">General documentation<\/h3><p>A short introduction to LaTeX (with some focus on the differences\nbetween the old LaTeX 2.09 and current standard LaTeX2e) is given in\nthe following document:<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/usrguide.pdf\" target=\"_blank\">LaTeX2e for authors<\/a><\/li>\n<\/ul><p>More advanced documentation about LaTeX, which is available via the net includes:<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/clsguide.pdf\" target=\"_blank\">LaTeX2e\n   for class and package writers<\/a><\/li>\n  <li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/fntguide.pdf\" target=\"_blank\">LaTeX2e\n   font selection<\/a><\/li>\n  <li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/encguide.pdf\" target=\"_blank\">LaTeX2e\n   font encodings<\/a><\/li>\n  <li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/cyrguide.pdf\" target=\"_blank\">Cyrillic\n   languages support in LaTeX<\/a><\/li>\n  <li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/cfgguide.pdf\" target=\"_blank\">Configuration\n   options for LaTeX2e<\/a><\/li>\n  <li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/modguide.pdf\" target=\"_blank\">Modifying\n   LaTeX<\/a><\/li>\n  <li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/ltx3info.pdf\" target=\"_blank\">The LaTeX3\n   Project<\/a><\/li>\n<\/ul><h3 id=\"typesetting-complex-mathematics\">Typesetting complex mathematics<\/h3><p>Specifically targeting the typesetting of mathematics is:<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/amsldoc.pdf\" target=\"_blank\">User\u2019s Guide for the amsmath Package (Version 2.1)<\/a><\/li>\n<\/ul><h3 id=\"japanese-translations\">Japanese translations<\/h3><p>Yukitoshi FUJIMURA kindly translated two of the above documents to the Japanese language. These are<\/p><p>He also provided a translation of \u201cShort Math Guide for LaTeX\u201d distributed and maintained\nby the American Mathematical Society (AMS):<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/short-math-guide_jpn.pdf\" target=\"_blank\">\u306f\u3084\u308f\u304b\u308a LaTeX \u3067\u6570\u5f0f\u7d44\u7248 <\/a>(Short Math Guide for LaTeX)\n<a href=\"https:\/\/www.latex-project.org\/help\/documentation\/short-math-guide_jpn.tex\" target=\"_blank\"> [source]<\/a><\/li>\n<\/ul><p>Here are <a href=\"https:\/\/www.latex-project.org\/help\/documentation\/readme_jpn\/\">his comments on the translation<\/a>.<\/p><h3 id=\"quick-summary-of-changes-by-release\">Quick summary of changes by release<\/h3><p>Changes made to the LaTeX kernel or to core packages maintained by the\nLaTeX team are discussed in some detail in the\n<a href=\"https:\/\/www.latex-project.org\/news\/latex2e-news\/\">LaTeX News Newsletters<\/a>\nthat come as part of each release.\nA summary sorted by release is given in the following document:<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/latexchanges.pdf\" target=\"_blank\">Changes to the LaTeX2\u03b5 format since 2014<\/a><\/li>\n<\/ul><h3 id=\"source-code-documentation\">Source code documentation<\/h3><p>The full documentation of the source code with all commands and their\nimplementation can be obtained by processing <code class=\"highlighter-rouge\">source2e.tex<\/code>\ndistributed as part of the LaTeX2e distribution. A compiled version\nwith a list of all major changes and an index of all commands and\ntheir usage within the kernel is<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/source2e.pdf\" target=\"_blank\">The LaTeX2e Sources, release 2018-12-01 (660+ pages)<\/a><\/li>\n<\/ul><p>As a companion document Martin Scharrer compiled a useful reference\nlist with links back into the previous document if both are stored in\nthe same directory:<\/p><ul><li><a href=\"http:\/\/mirrors.ctan.org\/info\/macros2e\/macros2e.pdf\" target=\"_blank\">List of internal LaTeX2e Macros useful to Package Authors<\/a><\/li>\n<\/ul><p>For the implementation of the standard classes <code class=\"highlighter-rouge\">article<\/code>, <code class=\"highlighter-rouge\">report<\/code>,\nand <code class=\"highlighter-rouge\">book<\/code> there also exists a document that contains the complete sources\nwith commentary:<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/classes.pdf\" target=\"_blank\">Standard Document Classes for LaTeX2e<\/a><\/li>\n<\/ul><p>In 2015 we introduced a roll-back\/roll-forward functionality by which\nit becomes possible to reset the kernel code (though not external\npackages at this stage) to the behavior that it had on a particular\ndate. This can be useful when processing older documents. It is\ndocumented here:<\/p><ul><li><a href=\"https:\/\/www.latex-project.org\/help\/documentation\/latexrelease.pdf\" target=\"_blank\">The latexrelease package<\/a><\/li>\n<\/ul><h2 id=\"other-non-english-documentation\">Other non-english documentation<\/h2><p>There is a lot of documentation in languages other than English. Lists of books and other resources in other languages are maintained by TeX user groups in the respective countries. The following are known to us:<\/p><h3 id=\"german\">German<\/h3><ul><li><a href=\"http:\/\/www.dante.de\/\">German LaTeX documentation maintained by German TeX User Group Dante e.V.<\/a><\/li>\n<\/ul><h3 id=\"french\">French<\/h3><ul><li><a href=\"https:\/\/www.gutenberg.eu.org\/-TeXniques-\">French LaTeX documentation maintained by the Francophone TeX Users Group GUTenberg<\/a><\/li>\n  <li><a href=\"http:\/\/lozzone.free.fr\/index.php?vlunch=latex\">Tout ce que vous avez toujours voulu savoir sur LaTeX<\/a><\/li>\n  <li><a href=\"http:\/\/www.framasoft.net\/rubrique266.html\">A collection of french LaTeX documentation<\/a> maintained by <a href=\"http:\/\/www.framasoft.net\/\">Framasoft<\/a><\/li>\n  <li><a href=\"http:\/\/www.babafou.eu.org\/Apprends_LaTeX\/\">Apprends LaTeX!<\/a><\/li>\n<\/ul><hr \/><hr \/>","created_at":"2018-12-19T02:37:45+0000","updated_at":"2018-12-19T02:37:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.latex-project.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":745,"title":"The Observation Deck \u00bb Open source confronts its midlife crisis","url":"http:\/\/dtrace.org\/blogs\/bmc\/2018\/12\/14\/open-source-confronts-its-midlife-crisis\/","content":"<h1>Open source confronts its midlife crisis<\/h1>&#13;\n&#13;\n    \t\t\t<p>Midlife is tough: the idealism of youth has faded, as has inevitably some of its fitness and vigor.  At the same time, the responsibilities of adulthood have grown:  the kids that were such a fresh adventure when they were infants and toddlers are now grappling with their own transition into adulthood \u2014 and you try to remind yourself that the kids that you have sacrificed so much for probably don\u2019t <i>actually<\/i> hate your guts, regardless of that post they just liked on the \u2018gram.  Making things more challenging, while you are navigating the turbulence of teenagers, your own parents are likely entering life\u2019s twilight, needing help in new ways from their adult children.  By midlife, in addition to the singular joys of life, you have also likely experienced its terrible sorrows:  death, heartbreak, betrayal.  Taken together, the fading of youth, the growth in responsibility and the endurance of misfortune can lead to cynicism or (worse) drastic and poorly thought-out choices.  Add in a little fear of mortality and some existential dread, and you have the stuff of which midlife crises are made\u2026\n<\/p><p>I raise this not because of my own adventures at midlife, but because it is clear to me that open source \u2014 now several decades old and fully adult \u2014 is going through its own midlife crisis.  This has long been in the making:  for years, I (and others) have been critical of <a href=\"https:\/\/www.slideshare.net\/bcantrill\/leaping-the-chasm-from-proprietary-to-open-a-survivors-guide\/11\">service providers\u2019 parastic relationship with open source<\/a>, as cloud service providers turn open source software into a service offering without giving back to the communities upon which they implicitly depend.  At the same time, open source has been (rightfully) entirely unsympathetic to the proprietary software models that have been burned to the ground \u2014 but also seemingly oblivious as to the larger economic waves that have buoyed them.\n<\/p><p>So it seemed like only a matter of time before the companies built around open source software would have to confront their own crisis of confidence: open source business models are really tough, selling software-as-a-service is one of the most natural of them, the cloud service providers are really good at it \u2014 and their commercial appetites seem boundless.  And, like a new cherry red two-seater sports car next to a minivan in a suburban driveway, some open source companies are dealing with this crisis exceptionally poorly: they are trying to restrict the way that their open source software can be used.  These companies want it both ways: they want the advantages of open source \u2014 the community, the positivity, the energy, the adoption, the downloads \u2014 but they also want to enjoy the fruits of proprietary software companies in software lock-in and its concomitant monopolistic rents.  If this were entirely transparent (that is, if some bits were merely being made explicitly proprietary), it would be fine:  we could accept these companies as essentially proprietary software companies, albeit with <a href=\"http:\/\/dtrace.org\/blogs\/bmc\/2004\/08\/28\/the-economics-of-software\/\">an open source loss-leader<\/a>.  But instead, these companies are trying to license their way into this self-contradictory world:  continuing to claim to be entirely open source, but perverting the license under which portions of that source are available.  Most gallingly, they are doing this by hijacking open source nomenclature.  Of these, the laughably named <a href=\"https:\/\/commonsclause.com\/\">commons clause<\/a> is the worst offender (it is plainly designed to be confused with the purely virtuous <a href=\"https:\/\/creativecommons.org\/\">creative commons<\/a>), but others (including CockroachDB\u2019s <a href=\"https:\/\/www.cockroachlabs.com\/cockroachdb-community-license\/\">Community License<\/a>, MongoDB\u2019s <a href=\"https:\/\/www.mongodb.com\/licensing\/server-side-public-license\">Server Side Public License<\/a>, and Confluent\u2019s <a href=\"https:\/\/www.confluent.io\/confluent-community-license\">Community License<\/a>) are little better.  And in particular, as it apparently needs to be said: no, \u201ccommunity\u201d is <b>not<\/b> the opposite of \u201copen source\u201d \u2014 please stop sullying its good name by attaching it to licenses that are deliberately not open source!  But even if they were more aptly named (e.g. \u201cthe restricted clause\u201d or \u201cthe controlled use license\u201d or \u2014 perhaps most honest of all \u2014 \u201cthe please-don\u2019t-put-me-out-of-business-during-the-next-reInvent-keynote clause\u201d), these licenses suffer from a serious problem:  <b>they are almost certainly asserting rights that the copyright holder doesn\u2019t in fact have<\/b>.\n<\/p><p>If I sell you a book that I wrote, I can restrict your right to read it aloud for an audience, or sell a translation, or write a sequel; these restrictions are rights afforded the copyright holder.  I cannot, however, tell you that you can\u2019t put the book on the same bookshelf as that of my rival, or that you can\u2019t read the book while flying a particular airline I dislike, or that you aren\u2019t allowed to read the book and also work for a company that competes with mine.  (Lest you think that last example absurd, that\u2019s almost verbatim the language in the new <a href=\"https:\/\/www.confluent.io\/confluent-community-license\">Confluent Community (sic) License<\/a>.) I personally think that none of these licenses would withstand a court challenge, but I also don\u2019t think it will come to that: because the vendors behind these licenses will surely fear that they wouldn\u2019t survive litigation, they will deliberately avoid inviting such challenges.  In some ways, this netherworld is even worse, as the license becomes a vessel for unverifiable fear of arbitrary liability.\n<\/p><p>Legal dubiousness aside, as with that midlife hot rod, the licenses <b>aren\u2019t going to address the underlying problem.<\/b> To be clear, the underlying problem is not the licensing, it\u2019s that these companies don\u2019t know how to make money \u2014 they want open source to be its own business model, and seeing that the cloud service providers have an entirely viable business model, they want a piece of the action.  But as a result of these restrictive riders, one of two things will happen with respect to a cloud services provider that wants to build a service offering around the software:\n<\/p><ol><li>The cloud services provider will build their service not based on the software, but rather on another open source implementation that doesn\u2019t suffer from the complication of a lurking company with brazenly proprietary ambitions.\n<\/li><li>The cloud services provider will build their service on the software, but will use only the truly open source bits, reimplementing (and keeping proprietary) any of the surrounding software that they need.\n<\/li><\/ol><p>In the first case, the victory is strictly pyrrhic:  yes, the cloud services provider has been prevented from monetizing the software \u2014 but the software will now have less of the adoption that is the lifeblood of a thriving community.  In the second case, there is no real advantage over the current state of affairs:  the core software is still being used without the open source company being explicitly paid for it.  Worse, the software and its community have been harmed: where one could previously appeal to the social contract of open source (namely, that cloud service providers have a social responsibility to contribute back to the projects upon which they depend), now there is little to motivate such reciprocity.  Why should the cloud services provider contribute anything back to a company that has declared war on it?  (Or, worse, <a href=\"https:\/\/techcrunch.com\/2018\/11\/29\/the-crusade-against-open-source-abuse\/\">implicitly accused it of malfeasance<\/a>.) Indeed, as long as fights are being picked with them, cloud service providers are likely clutch their bug fixes in the open core as a differentiator, cackling to themselves over the gnarly race conditions that they have fixed of which the community is blissfully unaware.  Is this in any way a desired end state?\n<\/p><p>So those are the two cases, and they are both essentially bad for the open source project.  Now, one may notice that there is a choice missing, and for those open source companies that still harbor magical beliefs, let me put this to you as directly as possible: <b>cloud services providers are emphatically not going to license your proprietary software.<\/b> I mean, you knew that, right?  The whole premise with your proprietary license is that you are finding that there is no way to compete with the operational dominance of the cloud services providers; did you really believe that those same dominant cloud services providers can\u2019t simply reimplement your LDAP integration or whatever?  The cloud services providers are currently reproprietarizing <i>all<\/i> of computing \u2014 <a href=\"https:\/\/www.wired.com\/story\/new-amazon-chips-cloud-computing\/\">they are making their own CPUs<\/a> for crying out loud! \u2014 reimplementing the bits of your software that they need in the name of the service that their customers want (and will pay for!) won\u2019t even move the needle in terms of their effort.\n<\/p><p>Worse than all of this (and the reason why this madness needs to stop): <b>licenses that are vague with respect to permitted use are corporate toxin<\/b>.  Any company that has been through an acquisition can speak of the peril of the due diligence license audit:  the acquiring entity is almost always deep pocketed and (not unrelatedly) risk averse; the last thing that any company wants is for a deal to go sideways because of concern over unbounded liability to some third-party knuckle-head.  So companies that engage in license tomfoolery are doing worse than merely not solving their own problem:  they are potentially poisoning the wellspring of their own community.\n<\/p><p>So what to do?  Those of us who have been around for a while \u2014 who came up in the era of proprietary software and saw the merciless transition to open source software \u2014 know that there\u2019s no way to cross back over the Rubicon.  Open source software companies need to come to grips with that uncomfortable truth: their business model isn\u2019t their community\u2019s problem, and they should please stop trying to make it one.  And while they\u2019re at it, it would be great if they could please stop making outlandish threats about the demise of open source; they sound like shrieking proprietary software companies from the 1990s, warning that open source will be ridden with nefarious backdoors and unspecified legal liabilities.  (Okay, yes, a confession: just as one\u2019s first argument with their teenager is likely to give their own parents uncontrollable fits of smug snickering, those of us who came up in proprietary software may find companies decrying the economically devastating use of their open source software to be amusingly ironic \u2014 but our schadenfreude cups runneth over, so they can definitely stop now.)\n<\/p><p>So yes, these companies have a clear business problem:  they need to find goods and services that people will exchange money for.  There are many business models that are complementary with respect to open source, and some of the best open source software (and certainly the least complicated from a licensing drama perspective!) comes from companies that simply needed the software and open sourced it because they wanted to build a community around it.  (There are many examples of this, but the outstanding <a href=\"https:\/\/www.envoyproxy.io\/\">Envoy<\/a> and <a href=\"https:\/\/www.jaegertracing.io\/\">Jaeger<\/a> both come to mind \u2014 the former from Lyft, the latter from Uber.) In this regard, open source is like a remote-friendly working policy: it\u2019s something that you do because it makes economic and social sense; even as it\u2019s very core to your business, its not a business model in and of itself.\n<\/p><p>That said, it <i>is<\/i> possible to build business models around the open source software that is a company\u2019s expertise and passion!  Even though the VC that led the last round wants to puke into a trashcan whenever they hear it, business models like \u201csupport\u201d, \u201cservices\u201d and \u201ctraining\u201d are entirely viable!  (That\u2019s the good news; the bad news is that they may not deliver the up-and-to-the-right growth that these companies may have promised in their pitch deck \u2014 and they may come at too low a margin to pay for large teams, lavish perks, or outsized exits.)  And of course, making software available as a service is also an entirely viable business model \u2014 but I\u2019m pretty sure they\u2019ve heard about that one <a href=\"https:\/\/twitter.com\/bcantrill\/status\/1067871017006071808\">in the keynote<\/a>.\n<\/p><p>As part of their quest for a business model, these companies should read Adam Jacob\u2019s excellent blog entry on <a href=\"https:\/\/medium.com\/sustainable-free-and-open-source-communities\/we-need-sustainable-free-and-open-source-communities-edf92723d619\">sustainable free and open source communities<\/a>. Adam sees what I see (and <a href=\"https:\/\/redmonk.com\/sogrady\/2018\/09\/10\/tragedy-of-the-commons-clause\/\">Stephen O\u2019Grady sees<\/a> and <a href=\"https:\/\/medium.com\/@rhatr\/is-it-time-for-cloud-native-open-source-db0ad6e695e5\">Roman Shaposhnik sees<\/a>), and he has taken a really positive action by starting the <a href=\"https:\/\/sfosc.org\/\">Sustainable Free and Open Source Communities project<\/a>.  This project has a lot to be said for it: it explicitly focuses on building community; it emphasizes social contracts; it seeks longevity for the open source artifacts; it shows the way to viable business models; it rejects copyright assignment to a corporate entity.  Adam\u2019s efforts can serve to clear our collective head, and to focus on what\u2019s really important:  the health of the communities around open source.  By focusing on longevity, we can plainly see restrictive licensing as the death warrant that it is, shackling the fate of a community to that of a company.  (Viz. after the company behind AGPL-licensed RethinkDB capsized, it took the Linux Foundation buying the assets and relicensing them to <a href=\"https:\/\/www.joyent.com\/blog\/the-liberation-of-rethinkdb\">rescue the community<\/a>.) Best of all, it\u2019s written by someone who has built a business that has open source software at its heart.  Adam has <a href=\"https:\/\/sfosc.org\/sfosc-book\/motivations\/\">endured the challenges of the open core model<\/a>, and is refreshingly frank about its economic and psychic tradeoffs.  And if he doesn\u2019t make it explicit, Adam\u2019s fundamental optimism serves to remind us, too, that any perceived \u201cdanger\u201d to open source is overblown: open source is going to endure, as no company is going to be able to repeal the <a href=\"http:\/\/dtrace.org\/blogs\/bmc\/2004\/12\/16\/the-economics-of-software-redux\/\">economics of software<\/a>.  That said, as we collectively internalize that open source is not a business model on its own, we will likely see fewer VC-funded open source companies (though I\u2019m honestly not sure that that\u2019s a bad thing).\n<\/p><p>I don\u2019t think that it\u2019s an accident that Adam, Stephen, Roman and I see more or less the same thing and are more or less the same age: not only have we collectively experienced many sides of this, but we are at once young enough to still recall our own idealism, yet old enough to know that coercion never endures in the limit.  In short, this too shall pass \u2014 and in the end, open source will survive its midlife questioning just as people in midlife get through theirs:  by returning to its core values and by finding rejuvenation in its communities.  Indeed, we can all find solace in the fact that while life is finite, our values and our communities survive us \u2014 and that our engagement with them is our most important legacy.<br \/><\/p><div id=\"postmeta\">&#13;\n&#13;\n            Posted on December 14, 2018 at 10:50 pm by <a href=\"http:\/\/dtrace.org\/blogs\/bmc\" title=\"Visit bmc\u2019s website\" rel=\"external\">bmc<\/a> \u00b7 <a href=\"http:\/\/dtrace.org\/blogs\/bmc\/2018\/12\/14\/open-source-confronts-its-midlife-crisis\/\" rel=\"bookmark\" title=\"Permanent link to this post\">Permalink<\/a>&#13;\n<br \/>&#13;\nIn:\u00a0<a href=\"http:\/\/dtrace.org\/blogs\/bmc\/category\/uncategorized\/\" title=\"View all posts in Uncategorized\" rel=\"category tag\">Uncategorized<\/a>&#13;\n          <\/div>&#13;\n&#13;\n          &#13;\n&#13;\n&#13;\n  &#13;\n          <br \/><div id=\"navigation\">&#13;\n\t   \t  \t\t<div class=\"left\"><a href=\"http:\/\/dtrace.org\/blogs\/bmc\/2018\/10\/05\/assessing-software-engineering-candidates\/\" rel=\"prev\">\u00ab Previous post<\/a><\/div>&#13;\n\t    \t  \t<div class=\"right\"><\/div>&#13;\n\t\t      <\/div>&#13;\n&#13;\n    \t\t&#13;\n\t\t&#13;","created_at":"2018-12-16T05:03:09+0000","updated_at":"2018-12-16T05:03:09+0000","annotations":[],"mimetype":"text\/html","reading_time":12,"domain_name":"dtrace.org","preview_picture":"http:\/\/dtrace.org\/blogs\/ahl\/files\/2011\/10\/icon.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":744,"title":"Show HN: Stock Trading with Insomnia REST Client and Alpaca API","url":"https:\/\/news.ycombinator.com\/item?id=18675280","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D18675280\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"18675280\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=18675280\" class=\"storylink\">Show HN: Stock Trading with Insomnia REST Client and Alpaca API<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        17 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=shift8\" class=\"hnuser\">shift8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18675280\">2 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=18675280&amp;goto=item%3Fid%3D18675280\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Show%20HN%3A%20Stock%20Trading%20with%20Insomnia%20REST%20Client%20and%20Alpaca%20API&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Show%20HN%3A%20Stock%20Trading%20with%20Insomnia%20REST%20Client%20and%20Alpaca%20API\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=18675280&amp;auth=ae2807d55d2fd46cea43ab9069d5c4903e13eb42\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=18675280\">discuss<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>I've created an Insomnia workspace to make it easier to debug your stock trading code and learn the Alpaca API. You can read about how here: \n<a href=\"https:\/\/medium.com\/automation-generation\/using-alpacas-stock-trading-api-with-insomnia-client-a796f064a5d7\" rel=\"nofollow\">https:\/\/medium.com\/automation-generation\/using-alpacas-stock...<\/a><p>The actual workspace JSON for you to import into Insomnia can be found here: <a href=\"https:\/\/github.com\/alpacahq\/insomnia-workspace\" rel=\"nofollow\">https:\/\/github.com\/alpacahq\/insomnia-workspace<\/a><\/p><\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-12-15T02:40:41+0000","updated_at":"2018-12-15T02:40:41+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":743,"title":"Asking the right question is more important than getting the right answer","url":"https:\/\/lemire.me\/blog\/2018\/12\/06\/asking-the-right-question-is-more-important-than-getting-the-right-answer\/","content":"<p>Schools train us to provide the right answers to predefined questions. Yet anyone with experience from the real world knows that, more often than not, the difficult part is to find the right question.<\/p><p>To make a remarkable contribution, you need to start by asking the right question. I will go further than this: the questions you are asking might define who you are.<\/p><p>What is a good question?<\/p><ul><li>The great questions are tractable and fruitful. They lead you on a path of discovery. It is easy to ask how to cure cancer, but that\u2019s not a good question because it does not help anyone do medical research.<\/li><li>Secret questions are the best: if you are the only one with this question in mind, then you may be holding a gold mine. Questions that everyone is having are proportionally worthless. (E.g., see <a href=\"http:\/\/a-fwd.com\/asin-com=0804139296&amp;com=daniellemires-20&amp;ca=danilemisblog-20&amp;uk=danilemisblog-21&amp;de=danilemisbl05-21&amp;fr=danilemisbl0e-21&amp;it=danilemisbl02-21&amp;jp=danilemisblog-22&amp;cn=danilemisblog-23&amp;fb=com&amp;sc=w\" rel=\"nofollow\">Zero to One<\/a> by Peter Thiel)<\/li><\/ul><p>You may think that by studying hard, by learning all the answers, you will get better at asking great questions. I am not sure it works.<\/p><p>In fact, knowing too much can harm you. I would take a B student who has fresh questions as a Ph.D. student over a typical overeager A+ student who frets about getting everything right. It is a poorly held secret that some of the very best researchers and innovators were average students.<\/p><p>Do the following experiment. Pick a scholarly field, any field, then spend two weeks reading everything about it that you can. Next, write down 5 questions. I can almost guarantee you that these 5 questions will be already covered by sources you read. They will be \u201cknown\u201d questions.<\/p><p>So to find good questions, you have to maintain some distance from the material. This should be uncontroversial if you consider that I define \u201cgood questions\u201d to be \u201csecret\u201d or \u201chighly original\u201d.<\/p><p>Our minds tend to frame everything in terms of the patterns we have learned. Spend two years studying Marxism and every single problem will feel like a Marxist problem to you. It becomes difficult for you to come up with new questions outside of the frame.<\/p><p>Don\u2019t get me wrong: smart people who know more tend to be more creative, everything else being equal\u2026 but there is a difference between being knowledgeable and having been locked into a frame of mind.<\/p><p>Yet here is how many researchers work. They survey the best papers from the last major conference or journal issue in their field. Importantly, they make sure to read what everyone is reading and to make sure to make theirs the frame of minds of the best people. They make sure that they can repeat the most popular questions and answers. They look at the papers, look for holes or possibilities for improvement and work from there. What this ensures that there are a few leaders (people writing about genuine novel ideas) followed by a long and nearly endless stream of \u201cme too\u201d papers that offer minor and inconsequential variations.<\/p><p>It is easier to judge these things in retrospect. In computer science, we had the XML craze at the turn of the century. Dozens of XML papers appeared each year at each of the top database conferences. <a href=\"https:\/\/lemire.me\/blog\/2013\/01\/14\/xml-for-databases-a-dead-idea\/\">I wrote about the untold story of the death of this idea<\/a>. How could so many people get so excited at the same time by what was a dead-end?<\/p><p>I believe that people are happy to be handed out questions and will often rush out to provide highly sophisticated thorough answers\u2026 whether or not the question is the right one.<\/p><p>My claim is that the people leading are not unnaturally smart, knowledgeable or creative. The people who answer other people\u2019s questions are not dumb or unimaginative. The main difference is one of focus. You either focus on asking good questions or you focus on providing good answers.<\/p><p>The world would be better if we had more people asking better questions.<\/p><p>How might we ask better questions?<\/p><ul><li>Pay attention to what is around you and violates your worldview. How did Fleming discover penicillin? He noticed that some mold that had invaded his dirty lab appeared to kill bacteria. He asked the right question at that time.<\/li><li>Be patient. Reportedly, Einstein once stated, \u201cIt\u2019s Not That I\u2019m so Smart, It\u2019s Just That I Stay with Problems Longer.\u201d The longer you work on a problem, the more likely you are to find interesting questions. (See <a href=\"https:\/\/onlinelibrary.wiley.com\/doi\/abs\/10.1002\/jocb.350\">Forthmann et al. 2018<\/a>) The easiest way to miss the great questions is to dismiss the problems as uninteresting and move on too quickly.<\/li><li>Be physically active, go for a walk. Chaining yourself to a desk is likely counterproductive. I used to think that being an all-out intellectual was the best route, but I now believe that I was grossly mistaken. I personally take a walk outside almost every morning on weekdays. (See <a href=\"http:\/\/psycnet.apa.org\/buy\/2014-14435-001\">Oppezzo and Schwartz, 2014<\/a>).<\/li><li>Don\u2019t be too social. Social pressure toward conformity trigger intense instinctive reactions. It is simply hard to go against the herd. Thus you are better off not know too much about where the herd is. In concrete terms, spend entirely days by yourself.<\/li><li>Ask a lot of questions. If you want to become good at providing the right answers, train yourself to answer lots of questions. If you want to become good at asking questions, ask a lot of them.<\/li><li>Always question your own thoughts and work.<\/li><\/ul><blockquote><p>The scientific mind does not so much provide the right answers as ask the right questions. (attributed to Levi-Strauss)<\/p><\/blockquote>","created_at":"2018-12-14T18:49:33+0000","updated_at":"2018-12-14T18:49:33+0000","published_by":["View all posts by Daniel Lemire"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"lemire.me","preview_picture":"https:\/\/lemire.me\/img\/portrait2018facebook.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["algorithms"],"is_public":false,"id":742,"title":"The Convex Hull of a Planar Point Set","url":"http:\/\/geomalgorithms.com\/a10-_hull-1.html","content":"<p>Computing a convex hull (or just \"hull\") is one of the first sophisticated geometry algorithms, and there are many variations of it. The most common form of this algorithm involves determining the smallest convex set (called the \"convex hull\") containing a discrete set of points. This algorithm also applies to a polygon, or just any set of line segments, whose hull is the same as the hull of its vertex point set. There are numerous applications for convex hulls: collision avoidance, hidden object determination, and shape analysis to name a few. And they are a minimal linear bounding container.<\/p>&#13;\n              <p>The most popular hull algorithms are the \"Graham scan\" algorithm [Graham, 1972] and the \"divide-and-conquer\" algorithm [Preparata &amp; Hong, 1977]. Implementations of both these algorithms are readily available (see [O'Rourke, 1998]). Both are <img id=\"Picture1009\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time algorithms, but the Graham has a low runtime constant in 2D and runs very fast there. However, the Graham algorithm does not generalize to 3D and higher dimensions whereas the divide-and-conquer algorithm has a natural extension. We do not consider 3D algorithms here (see [O'Rourke, 1998] for more information).<\/p>&#13;\n              <p>Here is a list of some well-known 2D hull algorithms. Let <i>n<\/i> = # points in the input set, and <i>h<\/i> = # vertices on the output hull. Note that <img id=\"Picture1011\" height=\"21\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/h.le.n.gif\" border=\"0\" alt=\"h.le.n\" title=\"h.le.n\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, so <img id=\"Picture1012\" height=\"24\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/nh.le.n2.gif\" border=\"0\" alt=\"nh.le.n2\" title=\"nh.le.n2\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. The list is ordered by date of first publication.<\/p>&#13;\n              <table border=\"0\" cellpadding=\"0\" style=\"width: 100%; border-spacing: 0px;\"><tr><td align=\"center\">&#13;\n                    <table id=\"Table5\" border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 547px;\"><colgroup><col \/><col \/><col \/><\/colgroup><tr align=\"left\"><th id=\"Text100\" style=\"width: 223px;\">&#13;\n                          <p><b><u>Algorithm<\/u><\/b><\/p>&#13;\n                        <\/th>\n                        <th id=\"Text105\" style=\"width: 112px;\">&#13;\n                          <p><b><u>Speed<\/u><\/b><\/p>&#13;\n                        <\/th>\n                        <th id=\"Text110\" style=\"width: 211px;\">&#13;\n                          <p><b><u>Discovered By<\/u><\/b><\/p>&#13;\n                        <\/th>\n                      <\/tr><tr><td id=\"Text112\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text116\">&#13;\n                          <p><img id=\"Picture1008\" height=\"28\" width=\"50\" src=\"http:\/\/geomalgorithms.com\/O-n4-.gif\" border=\"0\" alt=\"O(n4)\" title=\"O(n4)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text120\">&#13;\n                          <p>[Anon, the dark ages]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text122\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text126\">&#13;\n                          <p><img id=\"Picture1005\" height=\"25\" width=\"52\" src=\"http:\/\/geomalgorithms.com\/O-nh-.gif\" border=\"0\" alt=\"O(nh)\" title=\"O(nh)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text130\">&#13;\n                          <p>[Chand &amp; Kapur, 1970]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text132\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text136\">&#13;\n                          <p><img id=\"Picture1003\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text140\">&#13;\n                          <p>[Graham, 1972]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text142\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text146\">&#13;\n                          <p><img id=\"Picture1006\" height=\"25\" width=\"52\" src=\"http:\/\/geomalgorithms.com\/O-nh-.gif\" border=\"0\" alt=\"O(nh)\" title=\"O(nh)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text150\">&#13;\n                          <p>[Jarvis, 1973]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text152\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text156\">&#13;\n                          <p><img id=\"Picture1007\" height=\"25\" width=\"52\" src=\"http:\/\/geomalgorithms.com\/O-nh-.gif\" border=\"0\" alt=\"O(nh)\" title=\"O(nh)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text160\">&#13;\n                          <p>[Eddy, 1977], [Bykat, 1978]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text162\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text166\">&#13;\n                          <p><img id=\"Picture1000\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text170\">&#13;\n                          <p>[Preparata &amp; Hong, 1977]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text172\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text176\">&#13;\n                          <p><img id=\"Picture1001\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text180\">&#13;\n                          <p>[Andrew, 1979]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text182\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text186\">&#13;\n                          <p><img id=\"Picture1002\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text190\">&#13;\n                          <p>[Kallay, 1984]<\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text192\">&#13;\n                          &#13;\n                        <\/td>\n                        <td id=\"Text196\">&#13;\n                          <p><img id=\"Picture1004\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-h-.gif\" border=\"0\" alt=\"O(nlog-h)\" title=\"O(nlog-h)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n                        <\/td>\n                        <td id=\"Text200\">&#13;\n                          <p>[Kirkpatrick &amp; Seidel, 1986]<\/p>&#13;\n                        <\/td>\n                      <\/tr><\/table><\/td>\n                <\/tr><\/table><h2><br \/>Convex Hulls<\/h2>&#13;\n              <p>The convex hull of a geometric object (such as a point set or a polygon) is the smallest convex set containing that object. There are many equivalent definitions for a convex set <b>S<\/b>. The most basic of these is:<\/p>&#13;\n              <p><b>Def 1. <\/b>A set <b>S<\/b> is <b><i>convex<\/i><\/b> if whenever two points <i>P<\/i> and <i>Q<\/i> are inside <b>S<\/b>, then the whole line segment <i>PQ<\/i> is also in <b>S<\/b>.<\/p>&#13;\n              <p><img id=\"Picture1074\" height=\"160\" width=\"326\" src=\"http:\/\/geomalgorithms.com\/Pic_convex1.gif\" border=\"0\" alt=\"Pic_convex1\" title=\"Pic_convex1\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n              <p>But this definition does not readily lead to algorithms for constructing convex sets. A more useful definition states:<\/p>&#13;\n              <p><b>Def 2.<\/b> A set <b>S<\/b> is <b><i>convex<\/i><\/b> if it is exactly equal to the intersection of all the half planes containing it.<\/p>&#13;\n              <p>It can be shown that these two definitions are equivalent. However, the second one gives us a better computational handle, especially when the set <b>S<\/b> is the intersection of a <b>finite<\/b> number of half planes. In this case, the boundary of <b>S<\/b> is polygon in 2D, and polyhedron in 3D, with which it can be identified.<\/p>&#13;\n              <p><b>Def 3.<\/b> The <b><i>convex hull<\/i><\/b> of a finite point set <b>S<\/b> = {<i>P<\/i>} is the smallest 2D convex polygon <img id=\"Picture994\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 2px; margin-bottom: 2px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> (or polyhedron in 3D) that contains <b>S<\/b>. That is, there is no other convex polygon (or polyhedron) <img id=\"Picture993\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/LAMBDA.gif\" border=\"0\" alt=\"LAMBDA\" title=\"LAMBDA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> with <img id=\"Picture995\" height=\"21\" width=\"88\" src=\"http:\/\/geomalgorithms.com\/S.in.LAMBDA.in.OMEGA.gif\" border=\"0\" alt=\"S.in.LAMBDA.in.OMEGA\" title=\"S.in.LAMBDA.in.OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>.<\/p>&#13;\n              <p>Also, this convex hull has the smallest area and the smallest perimeter of all convex polygons that contain <b>S<\/b>.<\/p>&#13;\n              <h2><br \/>2D Hull Algorithms<\/h2>&#13;\n              <p>For this algorithm we will cover two similar fast 2D hull algorithms: the Graham scan, and Andrew's Monotone Chain scan. They both use a similar idea, and are implemented as a stack. In practice, they are both very fast, but Andrew's algorithm will execute slightly faster since its sort comparisons and rejection tests are more efficient. An implementation of Andrew's algorithm is given below in our <a href=\"#chainHull_2D()\">chainHull_2D()<\/a> routine.<\/p>&#13;\n              <h3><br \/>The \"Graham Scan\" Algorithm<\/h3>&#13;\n              <p>The Graham scan algorithm [Graham, 1972] is often cited ([Preparata &amp; Shamos, 1985], [O'Rourke, 1998]) as the first real \"computational geometry\" algorithm. As the size of the geometric problem (namely, <i>n<\/i> = the number of points in the set) increases, it achieves the optimal asymptotic efficiency of <img id=\"Picture996\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time. This algorithm and its implementation has been covered in great detail by [O'Rourke, 1998, Sect 3.5, 72-86] with downloadable C code available from his web site: <a id=\"Text74Link2\" href=\"http:\/\/cs.smith.edu\/%7Eorourke\/books\/compgeom.html\">Computational Geometry in C<\/a>. We do not repeat that level of detail here, and only give a conceptual overview of the algorithm. The <img id=\"Picture997\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time for the Graham scan is spent doing an initial radial sort of the input set points. After that, the algorithm employs a stack-based method which runs in just <img id=\"Picture998\" height=\"25\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/O-n-.gif\" border=\"0\" alt=\"O(n)\" title=\"O(n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time. In fact, the method performs at most 2<i>n<\/i> simple stack push and pop operations. Thus, it executes very rapidly, bounded only by the speed of sorting.<\/p>&#13;\n              <p>Let <b>S<\/b> = {<i>P<\/i>} be a finite set of points. The algorithm starts by picking a point in <b>S<\/b> known to be a vertex of the convex hull. This can be done in <img id=\"Picture999\" height=\"25\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/O-n-.gif\" border=\"0\" alt=\"O(n)\" title=\"O(n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time by selecting the rightmost lowest point in the set; that is, a point with first a minimum (lowest) <i>y<\/i> coordinate, and second a maximum (rightmost) <i>x<\/i> coordinate. Call this base point <i>P<\/i><sub>0<\/sub>. Then, the algorithm sorts the other points <i>P<\/i> in <b>S<\/b> radially by the increasing counter-clockwise (ccw) angle the line segment <i>P<\/i><sub>0<\/sub><i>P<\/i> makes with the <i>x<\/i>-axis. If there is a tie and two points have the same angle, discard the one that is closest to <i>P<\/i><sub>0<\/sub>. For efficiency, it is important to note that the sort comparison between two points <i>P<\/i><sub>1<\/sub> and <i>P<\/i><sub>2<\/sub> can be made without actually computing their angles. In fact, computing angles would use slow inaccurate trigonometry functions, and doing these computations would be a bad mistake. Instead, one just observes that <i>P<\/i><sub>2<\/sub> would make a greater angle than <i>P<\/i><sub>1<\/sub> if (and only if) <i>P<\/i><sub>2<\/sub> lies on the left side of the directed line segment <i>P<\/i><sub>0<\/sub><i>P<\/i><sub>1<\/sub> as shown in the following diagram. This condition can be tested by a fast accurate computation that uses only 5 additions and 2 multiplications. The code for this test was given in the <a href=\"http:\/\/geomalgorithms.com\/a01-_area.html#isLeft()\">isLeft()<\/a> routine from Algorithm 1 about the <a href=\"http:\/\/geomalgorithms.com\/a01-_area.html\">Area of Triangles and Polygons<\/a>.<\/p>&#13;\n              <p><img id=\"Picture1075\" height=\"220\" width=\"328\" src=\"http:\/\/geomalgorithms.com\/Pic_graham1.gif\" border=\"0\" alt=\"Pic_graham1\" title=\"Pic_graham1\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n              <p>After sorting, let the ccw-radially-ordered point set be <img id=\"Picture1014\" height=\"26\" width=\"144\" src=\"http:\/\/geomalgorithms.com\/S--P0-P1-...-Pn-1-.gif\" border=\"0\" alt=\"S={P0,P1,...,Pn-1}\" title=\"S={P0,P1,...,Pn-1}\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. It looks like a fan with a pivot at the point <i>P<\/i><sub>0<\/sub>.<\/p>&#13;\n              <p><img id=\"Picture1076\" height=\"230\" width=\"328\" src=\"http:\/\/geomalgorithms.com\/Pic_graham2.gif\" border=\"0\" alt=\"Pic_graham2\" title=\"Pic_graham2\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n              <p>We next loop through the points of <b>S<\/b> one-by-one testing for convex hull vertices. The algorithm is an inductive incremental procedure using a stack of points. At each stage, we save (on the stack) the vertex points for the convex hull of all points already processed. This is the induction condition. We start with <i>P<\/i><sub>0<\/sub> and <i>P<\/i><sub>1<\/sub> on the stack. Then at the <i>k<\/i>-th stage, we add the next point <i>P<\/i><i><sub>k<\/sub><\/i>, and compute how it alters the prior convex hull. Because of the way <b>S<\/b> was sorted, <i>P<\/i><i><sub>k<\/sub><\/i> is outside the hull of the prior points <i>P<\/i><i><sub>i<\/sub><\/i> with <i>i<\/i> &lt; <i>k<\/i>, and it must be added as a new hull vertex on the stack. But it's addition may cause previous stack points to no longer be a hull vertices. If this happens, the previous points must be popped off the stack and discarded. One tests for this by checking if the new point <i>P<\/i><i><sub>k<\/sub><\/i> is to the left or the right of the line joining the top two points of the stack. Again, we use the routine <a href=\"http:\/\/geomalgorithms.com\/a01-_area.html#isLeft()\">isLeft()<\/a> to quickly make this test. If <i>P<\/i><i><sub>k<\/sub><\/i> is on the left of the top segment, then prior hull vertices remain intact, and <i>P<\/i><i><sub>k<\/sub><\/i> gets pushed onto the stack. But, if it is on the right side of the top segment, then the prior point at the stack top will get absorbed inside the new hull, and that prior point must be popped off the stack. This test against the line segment at the stack top continues until either <i>P<\/i><i><sub>k<\/sub><\/i> is left of that line or the stack is reduced to the single base point <i>P<\/i><sub>0<\/sub>. In either case, <i>P<\/i><i><sub>k<\/sub><\/i> gets pushed onto the stack, and the algorithm proceeds to the next point <i>P<\/i><sub><i>k<\/i>+1<\/sub> in the set. The different possibilities involved are illustrated in the following diagram. <\/p>&#13;\n              <p><img id=\"Picture1078\" height=\"250\" width=\"716\" src=\"http:\/\/geomalgorithms.com\/Pic_graham3.gif\" border=\"0\" alt=\"Pic_graham3\" title=\"Pic_graham3\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n              <p>It is easy to understand why this works by viewing it as an incremental algorithm. The old stack <img id=\"Picture1015\" height=\"26\" width=\"169\" src=\"http:\/\/geomalgorithms.com\/Sk-1--P0-...-Pt2-Pk-1-.gif\" border=\"0\" alt=\"Sk-1={P0,...,Pt2,Pk-1}\" title=\"Sk-1={P0,...,Pt2,Pk-1}\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, with <i>P<\/i><sub><i>k\u2013<\/i>1<\/sub> at the top, is the convex hull of all points <i>P<\/i><i><sub>i<\/sub><\/i> with<i> i <\/i>&lt; <i>k<\/i>. The next point <i>P<\/i><i><sub>k<\/sub><\/i> is outside this hull since it is left of the line <i>P<\/i><sub>0<\/sub><i>P<\/i><sub><i>k\u2013<\/i>1<\/sub> which is an edge of the <b>S<\/b><sub><i>k\u2013<\/i>1<\/sub> hull. To incrementally extend <b>S<\/b><sub><i>k\u2013<\/i>1<\/sub> to include <i>P<\/i><i><sub>k<\/sub><\/i>, we need to find the two tangents from <i>P<\/i><i><sub>k<\/sub><\/i> to <b>S<\/b><sub><i>k\u2013<\/i>1<\/sub>. One tangent is clearly the line <i>P<sub>k<\/sub>P<\/i><sub>0<\/sub><i>. The other is a line P<sub>k<\/sub>P<sub>t<\/sub><\/i> such that <i>P<sub>k<\/sub><\/i> is left of the segment in <b>S<\/b><sub><i>k\u2013<\/i>1<\/sub> preceding <i>P<sub>t<\/sub><\/i> and is right of the segment following <i>P<sub>t<\/sub><\/i> (when it exists). This uniquely characterizes the second tangent since <b>S<\/b><sub><i>k\u2013<\/i>1<\/sub> is a convex polygon. The way to find <i>P<sub>t<\/sub><\/i><sub>is simply to search from the top of the stack down until the point with the property is found. The points above <i>P<sub>t<\/sub><\/i> in <b>S<\/b><sub><i>k\u2013<\/i>1<\/sub> are easily seen to be contained inside the triangle <img id=\"Picture1017\" height=\"26\" width=\"65\" src=\"http:\/\/geomalgorithms.com\/DELTA-P0PtPk.gif\" border=\"0\" alt=\"DELTA-P0PtPk\" title=\"DELTA-P0PtPk\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, and are thus no longer on the hull extended to include <i>P<\/i><i><sub>k<\/sub><\/i>. So, they can be discarded by popping them off the stack during the search for <i>P<sub>t<\/sub><\/i>. Then, the <i>k<\/i>-th convex hull is the new stack <img id=\"Picture1016\" height=\"26\" width=\"141\" src=\"http:\/\/geomalgorithms.com\/Sk--P0-...-Pt-Pk-.gif\" border=\"0\" alt=\"Sk={P0,...,Pt,Pk}\" title=\"Sk={P0,...,Pt,Pk}\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>.<\/sub><\/p>&#13;\n              <p>At the end, when <i>k <\/i>= <i>n<\/i>-1, the points remaining on the stack are precisely the ordered vertices of the convex hull's polygon boundary. Note that for each point of <b>S<\/b> there is one push and at most one pop operation, giving at most 2<i>n<\/i> stack operations for the whole algorithm.<\/p>&#13;\n              <p>This procedure is summarized by the following pseudo-code.<\/p>&#13;\n              <p><b>Pseudo-Code: Graham Scan Algorithm<\/b><\/p>&#13;\n              <table border=\"0\" cellpadding=\"0\" style=\"width: 100%; border-spacing: 0px;\"><tr><td align=\"center\">&#13;\n                    <table id=\"Table6\" border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 645px;\"><colgroup><col \/><\/colgroup><tr><td id=\"Text206\" style=\"width: 645px;\">&#13;\n                          <p><img id=\"Picture379\" height=\"4\" width=\"500\" src=\"http:\/\/geomalgorithms.com\/trans5x500.gif\" border=\"0\" alt=\"image\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: bottom;\" \/><\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text209\">&#13;\n                          <p>\u00a0\u00a0\u00a0 <b>Input:<\/b> a\u00a0 set of points <b>S<\/b> = {P = (P.x,P.y)}<\/p><p>\u00a0\u00a0\u00a0 Select the rightmost lowest point P<sub>0<\/sub> in <b> S<\/b><br \/>\u00a0\u00a0\u00a0 Sort <b>S<\/b> radially (ccw) about P<sub>0<\/sub> as a center {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Use isLeft() comparisons<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 For ties, discard the closer points<br \/>\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0 Let P[N] be the sorted array of points with P[0]=P<sub>0<\/sub><\/p><p>\u00a0\u00a0\u00a0 Push P[0] and P[1] onto a stack <img id=\"Picture1018\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p><p>\u00a0\u00a0\u00a0 while i &lt; N<br \/>\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Let P<sub>T1<\/sub> = the top point on <img id=\"Picture1019\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 If (P<sub>T1<\/sub> == P[0]) {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Push P[i] onto <img id=\"Picture1020\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 i++\u00a0\u00a0\u00a0\u00a0 \/\/ increment i<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Let P<sub>T2<\/sub> = the second top point on <img id=\"Picture1021\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 If (P[i] is strictly left of the line\u00a0 P<sub>T2<\/sub> to P<sub>T1<\/sub>) { <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Push P[i] onto <img id=\"Picture1022\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 i++\u00a0\u00a0\u00a0\u00a0 \/\/ increment i<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 else<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Pop the top point P<sub>T1<\/sub> off the stack<br \/>\u00a0\u00a0\u00a0 }<\/p><p>\u00a0\u00a0\u00a0 <b>Output: <\/b><img id=\"Picture1023\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> = the convex hull of <b>S<\/b>.<\/p>&#13;\n                        <\/td>\n                      <\/tr><\/table><\/td>\n                <\/tr><\/table><h3><br \/>Andrew's Monotone Chain Algorithm<\/h3>&#13;\n              <p>[Andrew, 1979] discovered an alternative to the Graham scan that uses a linear lexographic sort of the point set by the <i>x<\/i> and <i>y<\/i>-coordinates. This is an advantage if this ordering is already known for a set, which is sometimes the case. But even if sorting is required, this is a faster sort than the angular Graham-scan sort with its more complicated comparison function. The \"Monotone Chain\" algorithm computes the upper and lower hulls of a monotone chain of points, which is why we refer to it as the \"Monotone Chain\" algorithm. Like the Graham scan, it runs in <img id=\"Picture1024\" height=\"25\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/O-nlog-n-.gif\" border=\"0\" alt=\"O(nlog-n)\" title=\"O(nlog-n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time due to the sort time. After that, it only takes <img id=\"Picture1025\" height=\"25\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/O-n-.gif\" border=\"0\" alt=\"O(n)\" title=\"O(n)\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> time to compute the hull. This algorithm also uses a stack in a manner very similar to Graham's algorithm.<\/p>&#13;\n              <p>First the algorithm sorts the point set <img id=\"Picture1026\" height=\"26\" width=\"144\" src=\"http:\/\/geomalgorithms.com\/S--P0-P1-...-Pn-1-.gif\" border=\"0\" alt=\"S={P0,P1,...,Pn-1}\" title=\"S={P0,P1,...,Pn-1}\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> by increasing <i>x<\/i> and then <i>y<\/i> coordinate values. Let the minimum and maximum <i>x<\/i>-coordinates be <i>x<sub>min<\/sub><\/i> and <i>x<sub>max<\/sub><\/i>. Clearly, <img id=\"Picture1034\" height=\"26\" width=\"86\" src=\"http:\/\/geomalgorithms.com\/P0.x-x-min.gif\" border=\"0\" alt=\"P0.x=x-min\" title=\"P0.x=x-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, but there may be other points with this minimum <i>x<\/i>-coordinate. Let <img id=\"Picture1035\" height=\"26\" width=\"57\" src=\"http:\/\/geomalgorithms.com\/Pminmin.gif\" border=\"0\" alt=\"Pminmin\" title=\"Pminmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> be a point in <b>S<\/b> with <img id=\"Picture1036\" height=\"26\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/P.x-x-min.gif\" border=\"0\" alt=\"P.x=x-min\" title=\"P.x=x-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> first and then min <i>y<\/i> among all those points. Also, let <img id=\"Picture1038\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pminmax.gif\" border=\"0\" alt=\"Pminmax\" title=\"Pminmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> be the point with <img id=\"Picture1095\" height=\"26\" width=\"81\" src=\"http:\/\/geomalgorithms.com\/P.x-x-min.gif\" border=\"0\" alt=\"P.x=x-min\" title=\"P.x=x-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> first and then max <i>y<\/i> second. Note that <img id=\"Picture1039\" height=\"26\" width=\"130\" src=\"http:\/\/geomalgorithms.com\/Pminmin-Pminmax.gif\" border=\"0\" alt=\"Pminmin=Pminmax\" title=\"Pminmin=Pminmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> when there is a unique <i>x<\/i>-minimum point. Similarly define <img id=\"Picture1040\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pmaxmin.gif\" border=\"0\" alt=\"Pmaxmin\" title=\"Pmaxmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and <img id=\"Picture1041\" height=\"25\" width=\"61\" src=\"http:\/\/geomalgorithms.com\/Pmaxmax.gif\" border=\"0\" alt=\"Pmaxmax\" title=\"Pmaxmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> as the points with <img id=\"Picture1042\" height=\"25\" width=\"84\" src=\"http:\/\/geomalgorithms.com\/P.x-x-max.gif\" border=\"0\" alt=\"P.x=x-max\" title=\"P.x=x-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> first, and then <i>y<\/i> min or max second. Again note that <img id=\"Picture1043\" height=\"26\" width=\"134\" src=\"http:\/\/geomalgorithms.com\/Pmaxmin-Pmaxmax.gif\" border=\"0\" alt=\"Pmaxmin=Pmaxmax\" title=\"Pmaxmin=Pmaxmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> when there is a unique <i>x<\/i>-maximum point. Next, join the lower two points, <img id=\"Picture1083\" height=\"26\" width=\"57\" src=\"http:\/\/geomalgorithms.com\/Pminmin.gif\" border=\"0\" alt=\"Pminmin\" title=\"Pminmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and <img id=\"Picture1084\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pmaxmin.gif\" border=\"0\" alt=\"Pmaxmin\" title=\"Pmaxmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> to define a lower line <img id=\"Picture1058\" height=\"26\" width=\"40\" src=\"http:\/\/geomalgorithms.com\/L-min.gif\" border=\"0\" alt=\"L-min\" title=\"L-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. Also, join the upper two points, <img id=\"Picture1085\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pminmax.gif\" border=\"0\" alt=\"Pminmax\" title=\"Pminmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and <img id=\"Picture1086\" height=\"25\" width=\"61\" src=\"http:\/\/geomalgorithms.com\/Pmaxmax.gif\" border=\"0\" alt=\"Pmaxmax\" title=\"Pmaxmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> to define an upper line <img id=\"Picture1059\" height=\"25\" width=\"41\" src=\"http:\/\/geomalgorithms.com\/L-max.gif\" border=\"0\" alt=\"L-max\" title=\"L-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. These points and lines are shown in the following example diagram.<\/p>&#13;\n              <p><img id=\"Picture1082\" height=\"275\" width=\"581\" src=\"http:\/\/geomalgorithms.com\/Pic_chain.gif\" border=\"0\" alt=\"Pic_chain\" title=\"Pic_chain\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/><\/p>&#13;\n              <p>The algorithm now proceeds to construct a lower convex vertex chain <img id=\"Picture1066\" height=\"26\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/OMEGA-min.gif\" border=\"0\" alt=\"OMEGA-min\" title=\"OMEGA-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> below <img id=\"Picture1060\" height=\"26\" width=\"40\" src=\"http:\/\/geomalgorithms.com\/L-min.gif\" border=\"0\" alt=\"L-min\" title=\"L-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and joining the two lower points <img id=\"Picture1087\" height=\"26\" width=\"57\" src=\"http:\/\/geomalgorithms.com\/Pminmin.gif\" border=\"0\" alt=\"Pminmin\" title=\"Pminmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and <img id=\"Picture1088\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pmaxmin.gif\" border=\"0\" alt=\"Pmaxmin\" title=\"Pmaxmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>; and also an upper convex vertex chain <img id=\"Picture1067\" height=\"25\" width=\"44\" src=\"http:\/\/geomalgorithms.com\/OMEGA-max.gif\" border=\"0\" alt=\"OMEGA-max\" title=\"OMEGA-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> above <img id=\"Picture1064\" height=\"25\" width=\"41\" src=\"http:\/\/geomalgorithms.com\/L-max.gif\" border=\"0\" alt=\"L-max\" title=\"L-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and joining the two upper points <img id=\"Picture1089\" height=\"25\" width=\"61\" src=\"http:\/\/geomalgorithms.com\/Pmaxmax.gif\" border=\"0\" alt=\"Pmaxmax\" title=\"Pmaxmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and <img id=\"Picture1090\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pminmax.gif\" border=\"0\" alt=\"Pminmax\" title=\"Pminmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. Then the convex hull <img id=\"Picture1056\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> of <b>S<\/b> is constructed by joining <img id=\"Picture1068\" height=\"26\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/OMEGA-min.gif\" border=\"0\" alt=\"OMEGA-min\" title=\"OMEGA-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> and <img id=\"Picture1070\" height=\"25\" width=\"44\" src=\"http:\/\/geomalgorithms.com\/OMEGA-max.gif\" border=\"0\" alt=\"OMEGA-max\" title=\"OMEGA-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> together.<\/p>&#13;\n              <p>The lower or upper convex chain is constructed using a stack algorithm almost identical to the one used for the Graham scan. For the lower chain, start with <img id=\"Picture1091\" height=\"26\" width=\"57\" src=\"http:\/\/geomalgorithms.com\/Pminmin.gif\" border=\"0\" alt=\"Pminmin\" title=\"Pminmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> on the stack. Then process the points of <b>S<\/b> in sequence. For <img id=\"Picture1069\" height=\"26\" width=\"42\" src=\"http:\/\/geomalgorithms.com\/OMEGA-min.gif\" border=\"0\" alt=\"OMEGA-min\" title=\"OMEGA-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, only consider points strictly below the lower line <img id=\"Picture1061\" height=\"26\" width=\"40\" src=\"http:\/\/geomalgorithms.com\/L-min.gif\" border=\"0\" alt=\"L-min\" title=\"L-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. Suppose that at any stage, the points on the stack are the convex hull of points below <img id=\"Picture1062\" height=\"26\" width=\"40\" src=\"http:\/\/geomalgorithms.com\/L-min.gif\" border=\"0\" alt=\"L-min\" title=\"L-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> that have already been processed. Now consider the next point <i>P<\/i><i><sub>k<\/sub><\/i> that is below <img id=\"Picture1063\" height=\"26\" width=\"40\" src=\"http:\/\/geomalgorithms.com\/L-min.gif\" border=\"0\" alt=\"L-min\" title=\"L-min\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. If the stack contains only the one point <img id=\"Picture1092\" height=\"26\" width=\"57\" src=\"http:\/\/geomalgorithms.com\/Pminmin.gif\" border=\"0\" alt=\"Pminmin\" title=\"Pminmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> then put <i>P<\/i><i><sub>k<\/sub><\/i> onto the stack and proceed to the next stage. Otherwise, determine whether P<i><sub>k<\/sub><\/i> is strictly left of the line between the top two points on the stack. If it is, put <i>P<\/i><i><sub>k<\/sub><\/i> onto the stack and proceed. If it is not, pop the top point off the stack, and test <i>P<\/i><i><sub>k<\/sub><\/i> against the stack again. Continue until <i>P<\/i><i><sub>k<\/sub><\/i> gets pushed onto the stack. After this stage, the stack again contains the vertices of the lower hull for the points already considered. The geometric rationale is exactly the same as for the Graham scan. After all points have been processed, push <img id=\"Picture1093\" height=\"26\" width=\"60\" src=\"http:\/\/geomalgorithms.com\/Pmaxmin.gif\" border=\"0\" alt=\"Pmaxmin\" title=\"Pmaxmin\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> onto the stack to complete the lower convex chain.<\/p>&#13;\n              <p>The upper convex chain <img id=\"Picture1071\" height=\"25\" width=\"44\" src=\"http:\/\/geomalgorithms.com\/OMEGA-max.gif\" border=\"0\" alt=\"OMEGA-max\" title=\"OMEGA-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> is constructed in an analogous manner. But, process <b>S<\/b> in decreasing order <img id=\"Picture1072\" height=\"26\" width=\"133\" src=\"http:\/\/geomalgorithms.com\/-Pn-1-Pn-2-...-P0-.gif\" border=\"0\" alt=\"{Pn-1,Pn-2,...,P0}\" title=\"{Pn-1,Pn-2,...,P0}\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, starting at <img id=\"Picture1094\" height=\"25\" width=\"61\" src=\"http:\/\/geomalgorithms.com\/Pmaxmax.gif\" border=\"0\" alt=\"Pmaxmax\" title=\"Pmaxmax\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>, and only considering points above <img id=\"Picture1065\" height=\"25\" width=\"41\" src=\"http:\/\/geomalgorithms.com\/L-max.gif\" border=\"0\" alt=\"L-max\" title=\"L-max\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/>. Once the two hull chains have been found, it is easy to join them together (but be careful to avoid duplicating the endpoints).<\/p>&#13;\n              <p><br \/><b>Pseudo-Code: Andrew's Monotone Chain Algorithm<\/b><\/p>&#13;\n              <table border=\"0\" cellpadding=\"0\" style=\"width: 100%; border-spacing: 0px;\"><tr><td align=\"center\">&#13;\n                    <table id=\"Table7\" border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><colgroup><col \/><\/colgroup><tr><td id=\"Text212\" style=\"width: 774px;\">&#13;\n                          <p><img id=\"Picture385\" height=\"4\" width=\"500\" src=\"http:\/\/geomalgorithms.com\/trans5x500.gif\" border=\"0\" alt=\"image\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: bottom;\" \/><\/p>&#13;\n                        <\/td>\n                      <\/tr><tr><td id=\"Text215\">&#13;\n                          <p>\u00a0\u00a0\u00a0 <b>Input:<\/b> a set <b>S<\/b> = {P = (P.x,P.y)} of N points<\/p><p>\u00a0\u00a0\u00a0 Sort <b>S<\/b> by increasing x and then y-coordinate.<br \/>\u00a0\u00a0\u00a0 Let P[] be the sorted array of N points.<\/p><p>\u00a0\u00a0\u00a0 Get the points with 1st x min or max and 2nd y min or max<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 minmin = index of P with min x first and min y second<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 minmax = index of P with min x first and max y second<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 maxmin = index of P with max x first and min y second<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 maxmax = index of P with max x first and max y second<\/p><p>\u00a0\u00a0\u00a0 Compute the lower hull stack as follows:<br \/>\u00a0\u00a0\u00a0 (1) Let L_min be the lower line joining P[minmin] with\u00a0 P[maxmin].<br \/>\u00a0\u00a0\u00a0 (2) Push P[minmin] onto the stack.<br \/>\u00a0\u00a0\u00a0 (3) for i = minmax+1 to maxmin-1 (the points between xmin and xmax)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (P[i] is above or on L_min)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Ignore it and continue.<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 while (there are at least 2 points on the stack)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Let P<sub>T1<\/sub> = the top point on the stack.<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Let P<sub>T2<\/sub> = the second point on the stack.<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (P[i] is strictly left of the line from P<sub>T2<\/sub> to P<sub>T1<\/sub>)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 break out of this while loop.<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Pop the top point P<sub>T1<\/sub> off the stack.<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Push P[i] onto the stack.<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0 (4) Push P[maxmin] onto the stack.<\/p><p>\u00a0\u00a0\u00a0 Similarly, compute the upper hull stack.<\/p><p>\u00a0\u00a0\u00a0 Let <img id=\"Picture1032\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> = the join of the lower and upper hulls.<\/p><p>\u00a0\u00a0\u00a0 <b>Output: <\/b><img id=\"Picture1033\" height=\"20\" width=\"20\" src=\"http:\/\/geomalgorithms.com\/OMEGA.gif\" border=\"0\" alt=\"OMEGA\" title=\"OMEGA\" style=\"margin-top: 0px; margin-bottom: 0px; margin-left: 0px; margin-right: 0px; text-align: top;\" \/> = the convex hull of <b>S<\/b>.<br \/>\u00a0<\/p>&#13;\n                        <\/td>\n                      <\/tr><\/table><\/td>\n                <\/tr><\/table><h2><br \/>Implementation<\/h2>&#13;\n              <p>Here is a \"C++\" implementation of the Chain Hull algorithm. <\/p>&#13;\n              <p>\/\/ Copyright 2001 softSurfer, 2012 Dan Sunday<br \/>\/\/ This code may be freely used and modified for any purpose<br \/>\/\/ providing that this copyright notice is included with it.<br \/>\/\/ SoftSurfer makes no warranty for this code, and cannot be held<br \/>\/\/ liable for any real or imagined damage resulting from its use.<br \/>\/\/ Users of this code must verify correctness for their application.<br \/>\u00a0<\/p>&#13;\n              <p>\/\/ Assume that a class is already given for the object:<br \/>\/\/\u00a0\u00a0\u00a0 <b>Point<\/b> with coordinates {float x, y;}<br \/>\/\/===================================================================<br \/>\u00a0<\/p>&#13;\n              <p>\/\/ <b>isLeft():<\/b> tests if a point is Left|On|Right of an infinite line.<br \/>\/\/\u00a0\u00a0\u00a0 Input:\u00a0 three points P0, P1, and P2<br \/>\/\/\u00a0\u00a0\u00a0 Return: &gt;0 for P2 left of the line through P0 and P1<br \/>\/\/\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 =0 for P2 on the line<br \/>\/\/\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;0 for P2 right of the line<br \/>\/\/\u00a0\u00a0\u00a0 See: Algorithm 1 on <a href=\"http:\/\/geomalgorithms.com\/a01-_area.html\">Area of Triangles<\/a><br \/>inline float<br \/>isLeft( Point P0, Point P1, Point P2 )<br \/>{<br \/>\u00a0\u00a0\u00a0 return (P1.x - P0.x)*(P2.y - P0.y) - (P2.x - P0.x)*(P1.y - P0.y);<br \/>}<br \/>\/\/===================================================================<\/p>&#13;\n              <p><br \/>\/\/ <b>chainHull_2D():<\/b> Andrew's monotone chain 2D convex hull algorithm<br \/>\/\/\u00a0\u00a0\u00a0\u00a0 Input:\u00a0 P[] = an array of 2D points <br \/>\/\/\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 <b>presorted<\/b> by increasing x and y-coordinates<br \/>\/\/\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 n =\u00a0 the number of points in P[]<br \/>\/\/\u00a0\u00a0\u00a0\u00a0 Output: H[] = an array of the convex hull vertices (max is n)<br \/>\/\/\u00a0\u00a0\u00a0\u00a0 Return: the number of points in H[]<br \/>int<br \/>chainHull_2D( Point* P, int n, Point* H )<br \/>{<br \/>\u00a0\u00a0\u00a0 \/\/ the output array H[] will be used as the stack<br \/>\u00a0\u00a0\u00a0 int\u00a0\u00a0\u00a0 bot=0, top=(-1);\u00a0\u00a0 \/\/ indices for bottom and top of the stack<br \/>\u00a0\u00a0\u00a0 int\u00a0\u00a0\u00a0 i;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ array scan index<\/p><p>\u00a0\u00a0\u00a0 \/\/ Get the indices of points with min x-coord and min|max y-coord<br \/>\u00a0\u00a0\u00a0 int minmin = 0, minmax;<br \/>\u00a0\u00a0\u00a0 float xmin = P[0].x;<br \/>\u00a0\u00a0\u00a0 for (i=1; i&lt;n; i++)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (P[i].x != xmin) break;<br \/>\u00a0\u00a0\u00a0 minmax = i-1;<br \/>\u00a0\u00a0\u00a0 if (minmax == n-1) {\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ degenerate case: all x-coords == xmin<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] = P[minmin];<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (P[minmax].y != P[minmin].y) \/\/ a\u00a0 nontrivial segment<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] =\u00a0 P[minmax];<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] = P[minmin];\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ add polygon endpoint<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 return top+1;<br \/>\u00a0\u00a0\u00a0 }<\/p><p>\u00a0\u00a0\u00a0 \/\/ Get the indices of points with max x-coord and min|max y-coord<br \/>\u00a0\u00a0\u00a0 int maxmin, maxmax = n-1;<br \/>\u00a0\u00a0\u00a0 float xmax = P[n-1].x;<br \/>\u00a0\u00a0\u00a0 for (i=n-2; i&gt;=0; i--)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (P[i].x != xmax) break;<br \/>\u00a0\u00a0\u00a0 maxmin = i+1;<\/p><p>\u00a0\u00a0\u00a0 \/\/ Compute the <b>lower hull<\/b> on the stack H<br \/>\u00a0\u00a0\u00a0 H[++top] = P[minmin];\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ push\u00a0 minmin point onto stack<br \/>\u00a0\u00a0\u00a0 i = minmax;<br \/>\u00a0\u00a0\u00a0 while (++i &lt;= maxmin)<br \/>\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ the lower line joins P[minmin]\u00a0 with P[maxmin]<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (isLeft( P[minmin], P[maxmin], P[i])\u00a0 &gt;= 0 &amp;&amp; i &lt; maxmin)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 continue;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ ignore P[i] above or on the lower line<\/p><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 while (top &gt; 0)\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ there are at least 2 points on the stack<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ test if\u00a0 P[i] is left of the line at the stack top<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (isLeft(\u00a0 H[top-1], H[top], P[i]) &gt; 0)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 break;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ P[i] is a new hull\u00a0 vertex<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 else<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 top--;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ pop top point off\u00a0 stack<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] = P[i];\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ push P[i] onto stack<br \/>\u00a0\u00a0\u00a0 }<\/p><p>\u00a0\u00a0\u00a0 \/\/ Next, compute the <b>upper hull<\/b> on the stack H above\u00a0 the bottom hull<br \/>\u00a0\u00a0\u00a0 if (maxmax != maxmin)\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ if\u00a0 distinct xmax points<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] = P[maxmax];\u00a0 \/\/ push maxmax point onto stack<br \/>\u00a0\u00a0\u00a0 bot = top;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ the bottom point of the upper hull stack<br \/>\u00a0\u00a0\u00a0 i = maxmin;<br \/>\u00a0\u00a0\u00a0 while (--i &gt;= minmax)<br \/>\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ the upper line joins P[maxmax]\u00a0 with P[minmax]<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (isLeft( P[maxmax], P[minmax], P[i])\u00a0 &gt;= 0 &amp;&amp; i &gt; minmax)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 continue;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ ignore P[i] below or on the upper line<\/p><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 while (top &gt; bot)\u00a0\u00a0\u00a0\u00a0 \/\/ at least 2 points on the upper stack<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ test if\u00a0 P[i] is left of the line at the stack top<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (isLeft(\u00a0 H[top-1], H[top], P[i]) &gt; 0)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 break;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ P[i] is a new hull\u00a0 vertex<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 else<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 top--;\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ pop top point off\u00a0 stack<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] = P[i];\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ push P[i] onto stack<br \/>\u00a0\u00a0\u00a0 }<br \/>\u00a0\u00a0\u00a0 if (minmax != minmin)<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 H[++top] = P[minmin];\u00a0 \/\/ push\u00a0 joining endpoint onto stack<\/p><p>\u00a0\u00a0\u00a0 return top+1;<br \/>}<br \/>\u00a0<\/p>&#13;\n              <h2><br \/>References<\/h2>&#13;\n              <p>S.G. Akl &amp; Godfried Toussaint,\u00a0 \"Efficient Convex Hull Algorithms for Pattern Recognition Applications\", Proc.\u00a0 4th Int'l Joint Conf. on Pattern Recognition, Kyoto, Japan, 483-487 (1978)<\/p>&#13;\n              <p>A.M. Andrew, \"Another Efficient\u00a0 Algorithm for Convex Hulls in Two Dimensions\", Info.\u00a0 Proc. Letters 9, 216-219 (1979)<\/p>&#13;\n              <p>A. Bykat, \"Convex Hull of a Finite\u00a0 Set of Points in Two Dimensions\", Info.\u00a0 Proc. Letters 7, 296-298 (1978)<\/p>&#13;\n              <p>W. Eddy, \"A New Convex Hull Algorithm\u00a0 for Planar Sets\", ACM Trans. Math. Software 3(4), 398-403 (1977)<\/p>&#13;\n              <p>Ronald Graham, \"An Efficient\u00a0 Algorithm for Determining the Convex Hull of a Finite Point Set\", Info.\u00a0 Proc. Letters 1, 132-133 (1972)<\/p>&#13;\n              <p>R.A. Jarvis, \"On the Identification\u00a0 of the Convex Hull of of a Finite Set of Points in the Plane\", Info.\u00a0 Proc. Letters 2, 18-21 (1973)<\/p>&#13;\n              <p>M. Kallay, \"The Complexity of\u00a0 Incremental Convex Hull Algorithms in R<sup>d<\/sup>\", Info.\u00a0 Proc. Letters 19, 197 (1984)<\/p>&#13;\n              <p>D.G. Kirkpatrick &amp; R. Seidel, \"The\u00a0 Ultimate Planar Convex Hull Algorithm?\", SIAM Jour. Comput. 15, 287-299 (1986)<\/p>&#13;\n              <p>Joseph O'Rourke, <a id=\"Text74Link7\" href=\"http:\/\/www.amazon.com\/gp\/product\/0521649765\/ref=as_li_ss_tl?ie=UTF8&amp;tag=geometry_algorithms-20&amp;linkCode=as2&amp;camp=1789&amp;creative=390957&amp;creativeASIN=0521649765\">Computational Geometry in C (2nd Edition)<\/a>, Chap. 3 \"Convex Hulls in 2D\"\u00a0 (1998)<\/p>&#13;\n              <p>Franco Preparata &amp; Michael Shamos, <a id=\"Text74Link8\" href=\"http:\/\/www.amazon.com\/gp\/product\/0387961313\/ref=as_li_ss_tl?ie=UTF8&amp;tag=geometry_algorithms-20&amp;linkCode=as2&amp;camp=1789&amp;creative=390957&amp;creativeASIN=0387961313\">Computational Geometry: An Introduction<\/a>,\u00a0 Chap. 3 \"Convex Hulls: Basic Algorithms\" (1985)<\/p>&#13;\n              <p>Franco Preparata &amp; S.J. Hong, \"Convex\u00a0 Hulls of Finite Sets of Points in Two and Three Dimensions\", Comm. ACM 20, 87-93\u00a0 (1977)<\/p>&#13;","created_at":"2018-12-14T00:04:04+0000","updated_at":"2018-12-14T00:04:08+0000","annotations":[],"mimetype":"text\/html","reading_time":17,"domain_name":"geomalgorithms.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":741,"title":"Ask HN: What's the best book on modern UI\/UX?","url":"https:\/\/news.ycombinator.com\/item?id=18662992","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D18662992\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"18662992\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=18662992\" class=\"storylink\">Ask HN: What's the best book on modern UI\/UX?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        123 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=InGodsName\" class=\"hnuser\">InGodsName<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18662992\">3 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=18662992&amp;goto=item%3Fid%3D18662992\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20What's%20the%20best%20book%20on%20modern%20UI%2FUX%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20What's%20the%20best%20book%20on%20modern%20UI%2FUX%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=18662992&amp;auth=7b3866844128ccfbc507a3b197b4675e8da7fb46\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=18662992\">40\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>For designing websites like Stripe, Airbnb, Slack etc...?<p>What tools do they use?<\/p><p>What books they read or have written?<\/p><p>Which school teaches modern UI\/UX?<\/p><\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-12-12T21:04:10+0000","updated_at":"2018-12-12T21:04:10+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","dependencies","closures","functional"],"is_public":false,"id":740,"title":"On PHP 5.3, Lambda Functions, and Closures | Articles","url":"http:\/\/fabien.potencier.org\/on-php-5-3-lambda-functions-and-closures.html","content":"<p>PHP 5.3 will have a lot of exciting new features. One of the most important one\nfor me is the introduction of lambda functions and closures support. I won't\ntalk too much about what lambda functions or closures are, as you can find\nmany good blog posts describing them in great details. To sum up, a lambda\nfunction is an anonymous PHP function that can be stored in a variable and\npassed as an argument to other functions or methods. A closure is a lambda\nfunction that is aware of its surrounding context.<\/p><p>The first obvious use for lambda functions and closures is in conjunction with\nthe <code>array_map()<\/code>, <code>array_reduce()<\/code>, and <code>array_filter()<\/code> native PHP\nfunctions:<\/p><pre class=\"language-php\">$input = array(1, 2, 3, 4, 5);\n$output = array_filter($input, function ($v) { return $v &gt; 2; });\n<\/pre><p>\nThe above example filters the input array by removing all values greater than\n2:<\/p><pre class=\"language-php\">$output == array(2 =&gt; 3, 3 =&gt; 4, 4 =&gt; 5)\n<\/pre><p>\n<code>function ($v) { return $v &gt; 2; }<\/code> is the lambda function definition and it\ncan be stored in a PHP variable to be reusable:<\/p><pre class=\"language-php\">$max_comparator = function ($v) { return $v &gt; 2; };\n$input = array(1, 2, 3, 4, 5);\n$output = array_filter($input, $max_comparator);\n<\/pre><p>\nBut what if I want to change the maximum number allowed in the filtered array?\nI can either create another lambda function or use a closure:<\/p><pre class=\"language-php\">$max_comparator = function ($max)\n{\n  return function ($v) use ($max) { return $v &gt; $max; };\n};\n$input = array(1, 2, 3, 4, 5);\n$output = array_filter($input, $max_comparator(2));\n<\/pre><p>\nNow, the <code>$max_comparator<\/code> function takes the maximum allowed number and\nreturns a function that is different according to this maximum. Even for such\na contrived example, I hope you see the great power it gives you!<\/p><p>Closures also opens up a lot of great possibilities, like the implementation\nof the cryptic\n<a href=\"http:\/\/php100.wordpress.com\/2009\/04\/13\/php-y-combinator\/\">Y-combinator<\/a>, as\ndemonstrated by Stanislav Malyshev in one of his recent blog post:<\/p><pre class=\"language-php\">function Y($F)\n{\n  $func = function ($f) { return $f($f); };\n  return $func(function ($f) use($F)\n  {\n    return $F(function ($x) use($f)\n    {\n      $ff = $f($f);\n      return $ff($x);\n    });\n  });\n}\n<\/pre><p>\nToday, I want to talk about yet great another example on how to use lambda\nfunctions and closures. You will see that it can simplify your code a lot when\nused appropriately.<\/p><p>If you have read my blog recently, you know that I am quite obsessed with\ndependency injection these days. This post will show you how to implement a\nvery simple but still full-featured dependency injection container, thanks to\nthe new features of PHP 5.3.<\/p><p>A dependency injection container must be able to manage two different kind of\ndata: objects and parameters. And objects must be created on-demand the first\ntime they are accessed from the container.<\/p><p>Using a simple class that implements the magic <code>__get()<\/code> and <code>__set()<\/code>\nmethods, we can easily manage both the objects and the parameters:<\/p><pre class=\"language-php\">class DIContainer\n{\n  protected $values = array();\n  function __set($id, $value)\n  {\n    $this-&gt;values[$id] = $value;\n  }\n  function __get($id)\n  {\n    return is_callable($this-&gt;values[$id]) ? $this-&gt;values[$id]($this) : $this-&gt;values[$id];\n  }\n}\n<\/pre><p>\nUsing the container is quite simple:<\/p><pre class=\"language-php\">$container = new DIContainer();\n\/\/ define the parameters\n$container-&gt;cookie_name = 'SESSION_ID';\n$container-&gt;storage_class = 'SessionStorage';\n\/\/ defined the objects\n$container-&gt;storage = function ($c)\n{\n  return new $c-&gt;storage_class($c-&gt;cookie_name);\n};\n$container-&gt;user = function ($c)\n{\n  return new User($c-&gt;storage);\n};\n\/\/ get the user object\n$user = $container-&gt;user;\n\/\/ the above call is roughly equivalent to the following code:\n\/\/ $storage = new SessionStorage('SESSION_ID');\n\/\/ $user = new User($storage);\n<\/pre><blockquote class=\"note\"><p>The examples I use in this article are the same as the one I have used in my\nseries on <a href=\"http:\/\/fabien.potencier.org\/article\/11\/what-is-dependency-injection\">dependency injection<\/a>.\n<\/p><\/blockquote><p>The definition of an object is done by defining a lambda function that returns\nan instance of the object.<\/p><p>The <code>__get()<\/code> method checks if the value associated with a key is a PHP\ncallable (and lambda functions are callables) to make the difference between\nan object definition and a parameter.<\/p><p>Also notice how we call the lambda:<\/p><pre class=\"language-php\">$this-&gt;values[$id]($this)\n<\/pre><p>\nThe trick here is to pass the container as an argument of the lambda function\nso that it can use the container to access parameters and other objects\nmanaged by the container.<\/p><p>We can make the container a bit better by adding error support:<\/p><pre class=\"language-php\">class DIContainer\n{\n  protected $values = array();\n  function __set($id, $value)\n  {\n    $this-&gt;values[$id] = $value;\n  }\n  function __get($id)\n  {\n    if (!isset($this-&gt;values[$id]))\n    {\n      throw new InvalidArgumentException(sprintf('Value \"%s\" is not defined.', $id));\n    }\n    return is_callable($this-&gt;values[$id]) ? $this-&gt;values[$id]($this) : $this-&gt;values[$id];\n  }\n}\n<\/pre><p>\nDefining objects with lambda functions is great because the developer can do\nwhatever he wants to actually create and configure instances. But we still\nneed to implement one important feature of any dependency injection container:\nshared instances. This can be done manually like this:<\/p><pre class=\"language-php\">$container-&gt;user = function ($c)\n{\n  static $object;\n  if (is_null($object))\n  {\n    $object = new User($c-&gt;storage);\n  }\n  return $object;\n};\n<\/pre><p>\nBy declaring a static variable in the lambda function, the first time it is\ncalled, the object is created and returned. For all subsequent calls, the same\ninstance will always be returned.<\/p><p>It works as expected, but this is quite repetitive, tedious, and error prone.\nFor all instances that must be unique, we need to add this boilerplate code.\nInstead, I want to support shared instances as a native feature of the\ncontainer itself. Thanks to closures, that's quite easy to accomplish:<\/p><pre class=\"language-php\">class DIContainer\n{\n  \/\/ ...\n  function asShared($callable)\n  {\n    return function ($c) use ($callable)\n    {\n      static $object;\n      if (is_null($object))\n      {\n        $object = $callable($c);\n      }\n      return $object;\n    };\n  }\n}\n<\/pre><p>\nThe <code>asShared()<\/code> method wraps the given lambda function to add the needed\nlogic we have seen before. Declaring an object as unique for a given container\nis now dead simple:<\/p><pre class=\"language-php\">$c-&gt;user = $c-&gt;asShared(function ($c)\n{\n  return new User($c-&gt;storage);\n});\n<\/pre><p>\nIn less than 30 lines of PHP, we have coded a full-featured dependency\ninjection container. Quite impressive!<\/p><p>If we remove the need for shared instance, and if we define parameters and\nobjects as lambdas, we can even compact the code so that it fits in a tweet:<\/p><pre class=\"language-php\">class Container {\n protected $s=array();\n function __set($k, $c) { $this-&gt;s[$k]=$c; }\n function __get($k) { return $this-&gt;s[$k]($this); }\n}\n<\/pre><p>\nI have called this tweet container <em>twittee<\/em>, and it has its own\n<a href=\"http:\/\/twittee.org\/\">dedicated website<\/a> and\n<a href=\"http:\/\/github.com\/fabpot\/twittee\/tree\/master\">github repository<\/a>.<\/p><p>PHP 5.3 is really a great step forward for the PHP language.<\/p>","created_at":"2018-12-12T15:42:06+0000","updated_at":"2018-12-12T15:42:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"fabien.potencier.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":739,"title":"If You\u2019re So Successful, Why Are You Still Working 70 Hours a Week?","url":"https:\/\/hbr.org\/2018\/02\/if-youre-so-successful-why-are-you-still-working-70-hours-a-week","content":"<div class=\"article-executive-summary is-hidden\"><h4 class=\"text-gray-light mbn mt-large2\">Executive Summary<\/h4><p>In the old days, if you were a white-collar worker, the deal was that you worked as hard as you could at the start of your career to earn the right to be rewarded later on, with security of tenure and a series of increasingly senior positions. This is no longer true. Today, many senior leaders work\u00a0longer and harder than ever. At the heart of it is insecurity, and indeed, elite professional organizations deliberately set out to identify and recruit \u201cinsecure overachievers.\u201d Insecure overachievers are exceptionally capable and fiercely ambitious, yet are driven by a profound sense of their own inadequacy. If this sounds familiar, you should try to work exceptionally long hours when you need to or want to\u00a0\u2014 but do\u00a0it consciously, for specified time periods, and to achieve specific goals. Don\u2019t let it become a habit because you have forgotten how to work or live any other way.<\/p><\/div><div class=\"article article-first-row\"><div class=\"mbn pbn\"><figure><a href=\"https:\/\/hbr.org\/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960.jpg\"><img class=\"alignnone size-full wp-image-193430\" src=\"https:\/\/hbr.org\/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960.jpg\" srcset=\"\/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-300x169.jpg 300w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-1024x576.jpg 1024w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-500x281.jpg 500w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-383x215.jpg 383w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-383x215.jpg 383w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-500x281.jpg 500w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-700x394.jpg 700w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-850x478.jpg 850w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960-1200x675.jpg 1200w, \/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960.jpg 1200w\" alt=\"feb18_1_904950960\" width=\"1200\" height=\"675\" \/><\/a>\n   \n  <figcaption class=\"credit ptn mtn\">\n   Jena Ardell\/Getty Images\n  <\/figcaption><\/figure><\/div><p>\u201cI really became a robot,\u201d a manager at an accounting firm explained. She and her colleagues worked extraordinarily long hours, but, she said, \u201cI thought it was normal. It\u2019s like brainwashing. You are in a kind of mental system where you are under increasing demands, and you say to yourself that it doesn\u2019t matter, that you will rest afterwards, but that moment never comes.\u201d<\/p><p>Through my research, I\u2019ve heard stories like this over and over again from people in accounting firms, law firms, consulting firms, and other white-collar jobs. We all know that chronic overwork is bad for our mental and physical health and can seriously jeopardize the quality of our work. We wish we could change the way we work, but we don\u2019t really know how.<\/p><p>Long hours are most common in managerial and professional occupations. This is something of a recent trend. In the old days, if you were a white-collar worker, the deal was that you worked as hard as you could at the start of your career to earn the right to be rewarded later on, with security of tenure and a series of increasingly senior positions. In professional organizations, such as law firms, accountancy firms, management consultancies, and investment banks, the prize was partnership. The competition was relentless, but once you won the prize, it was yours for keeps. Partners had autonomy to choose how and when to work and what to work on. Of course, some senior partners spent a surprising amount of their \u201cbusiness development time\u201d on the golf course, but that was OK because they had already paid their dues to the organization.<\/p><p>This is no longer true. As a director of HR in a leading accounting firm told me, \u201cThe head of audit is in the office regularly from 5:30\u00a0AM until 10 PM,\u00a0on weekends, too. So is our managing partner. This is not exceptional. The rest of the firm sees the senior people working these hours and emulates them.\u201d<\/p><p>My <a href=\"https:\/\/global.oup.com\/academic\/product\/leading-professionals-9780198744788?cc=us&amp;lang=en&amp;\">research<\/a>, published in my new book about leadership in professional organizations, shows that our tendency to overwork and burn out is framed by a complex combination of factors involving our profession, our organization, and ourselves. At the heart of it is insecurity. As one senior business unit leader in a law firm admitted to me: \u201cI just come in here and work as hard as I can all the time. I feel like I\u2019m doing a good job, but it\u2019s hard to measure. That\u2019s the nature of what we do: It\u2019s feast or famine. And we all tend to be such insecure people that we\u2019re all scared all the time.\u201d<\/p><p>The 500 interviews I conducted for my book showed a pattern:\u00a0A professional\u2019s insecurity is rooted in the inherent intangibility of knowledge work. How do you\u00a0convince your client that you know something worthwhile and justify the high fees you charge? The insecurity caused by this intangibility is exacerbated by the rigorous \u201cup or out\u201d promotion system perpetuated by elite professional organizations, which turns your colleagues into your competitors. How do you convince your boss that you\u2019re worth more than your closest colleague? There is no chance for a professional to rest on their laurels\u00a0\u2014 or even to rest.<\/p><p>Exacerbating this problem, elite professional organizations deliberately set out to identify and recruit \u201cinsecure overachievers\u201d \u2014 some leading professional organizations explicitly use this terminology, though not in public. Insecure overachievers are exceptionally capable and fiercely ambitious, yet driven by a profound sense of their own inadequacy. This typically stems from childhood, and may result from various factors, such as experience of financial or physical deprivation, or a belief that\u00a0their parents\u2019 love was contingent upon their behaving and performing well.<\/p><p>As the recruiters I interviewed explained, these individuals are immensely attractive to elite professional organizations because they are entirely self-motivating and self-disciplining. The firm in effect tells the insecure overachiever, \u201cWe are the best in the business, and because we want you to work for us, that makes you the best, too.\u201d But\u00a0upon joining the firm, insecure overachievers discover that the rigorous up-or-out policy exacerbates their insecurity and their fear of being \u201cexposed\u201d as inadequate \u2014 and ultimately rejected.<\/p><p>In the short term, insecure overachievers respond by delivering exceptional performance. As the chair of a consulting firm told me, \u201cMy theory is that the best client relationship builders in our firm are insecure. They are so hell-bent on making their clients feel good about them that they work overtime. Clients feel their passion and respond to that.\u201d<\/p><p>The tendency to hard work is reinforced by the strong culture of social control created by elite professional organizations. On the one hand, this is comforting. Some professionals I have studied refer to their firms as being like a \u201cfamily,\u201d or something even more intense. As one consultant described it, \u201cWhen I first came here, I thought, This place\u00a0feels like a cult. But now I have been here a while, I think it is great.\u201d Taken to extremes, the insecure overachiever\u2019s sense of commitment can lead to extreme conformity and the normalization of unhealthy behaviors.<\/p><p>Paradoxically, the professionals I studied still believe that they have autonomy and that they are overworking by choice. They do not blame their organizations,\u00a0which after all have invested in work-life balance initiatives and wellness programs. Instead, they blame themselves for being inadequate. Their colleagues seem to be coping, and they take that as further evidence of their own inadequacy. They do not talk honestly to their colleagues about their problems, thus perpetuating the myth of the invincible professional, which encourages their colleagues to feel inadequate in turn. If they suffer burnout, they think it is their fault. Their organization and its leadership are absolved of responsibility, so nothing fundamental changes.<\/p><p>As a result, by the time insecure overachievers become leaders of their organizations, they unconsciously replicate the systems of social control and overwork that helped to create them.<\/p><p>If you are a leader who is wondering \u201cWhy am I working harder than ever?,\u201d take a good look at yourself, the organization that has created you, and the organizational practices you are perpetuating. Working hard can be rewarding and exhilarating. But consider how you are living. Recognize when you are driving yourself and your staff too hard, and learn how to help yourself and your colleagues to step back from the brink.<\/p><p>Your insecurities may have helped to get you where you are today, but are they still working for you? Is it time to acknowledge that you have \u201cmade it\u201d and to start enjoying the experience a little bit more? And if your boss is an insecure overachiever, recognize how they are projecting their insecurity onto you \u2014 how they make you feel insecure for not being able to keep up with them.<\/p><p>Work exceptionally long hours when you need to or want to, but do so consciously, for specified time periods, and to achieve specific goals. Don\u2019t let it become a habit because you have forgotten how to work or live any other way.<\/p><p>And notice how you judge colleagues who are working less hard than you \u2014 they may have discovered something you need to learn.<\/p><p>If you are a leader, you have a responsibility not just to your firm but to the people who work within it. Help your colleagues to achieve their full potential, but do not allow yourself to exacerbate and exploit their insecurities. And remember that your ultimate \u201cduty of care\u201d is to yourself.<\/p><\/div>","created_at":"2018-12-11T18:53:43+0000","updated_at":"2018-12-11T18:53:43+0000","published_at":"2018-02-01T17:00:50+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"hbr.org","preview_picture":"https:\/\/hbr.org\/resources\/images\/article_assets\/2018\/02\/feb18_1_904950960.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["graphs","algorithms"],"is_public":false,"id":738,"title":"Graph traversals","url":"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/lec24-12sp.html","content":"Graph traversals\n<h3>Topics:<\/h3>\n<ul><li>tricolor algorithm<\/li>\n<li>breadth-first search<\/li>\n<li>depth-first search<\/li>\n<li>cycle detection<\/li>\n<li>topological sort<\/li>\n<li>connected components<\/li>\n<\/ul><h2>Graph traversals<\/h2>\n<p>\nWe often want to solve problems that are expressible in terms of a traversal or search over a graph. Examples include:<\/p>\n<ul><li>Finding all reachable nodes (for garbage collection)<\/li>\n<li>Finding the best reachable node (single-player game search) or\nthe minmax best reachable node (two-player game search)<\/li>\n<li>Finding the best path through a graph (for routing and map directions)<\/li>\n<li>Determining whether a graph is a DAG.<\/li>\n<li>Topologically sorting a graph.<\/li>\n<\/ul><p> The goal of a graph traversal, generally, is to find all nodes reachable\nfrom a given set of root nodes. In an undirected graph we follow all edges; in\na directed graph we follow only out-edges.<\/p>\n<h2>Tricolor algorithm<\/h2>\n<p>\nAbstractly, graph traversal can be expressed in terms of the <b>tricolor\nalgorithm<\/b> due to Dijkstra and others. In this algorithm, graph nodes are\nassigned one of three colors that can change over time:\n<\/p>\n<ul><li><b>White<\/b> nodes are undiscovered nodes that have not been seen yet in the current traversal and may even be unreachable.\n<\/li><li><b>Black<\/b> nodes are nodes that are reachable and that the algorithm is done with.\n<\/li><li><b>Gray<\/b> nodes are nodes that have been discovered but that the algorithm is not done with yet. These nodes are on a frontier\nbetween white and black.\n<\/li><\/ul><p>\nThe progress of the algorithm is depicted by the following figure. Initially there are no black nodes and the roots are gray. As the\nalgorithm progresses, white nodes turn into gray nodes and gray nodes turn into black nodes. Eventually there are no gray nodes left\nand the algorithm is done.\n<\/p>\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/tricolor.png\" alt=\"image\" \/><\/div>\n<p>\nThe algorithm maintains a key invariant at all times: there are no edges from\nwhite nodes to black nodes. This is clearly true initially, and because it is\ntrue at the end, we know that any remaining white nodes cannot be reached from\nthe black nodes.\n<\/p>\n<p>\nThe algorithm pseudo-code is as follows:\n<\/p>\n<ol><li>Color all nodes white, except for the root nodes, which are colored gray.\n<\/li><li>While some gray node n exists:\n<ul><li>color some white successors of n gray.\n<\/li><li>if n has no white successors, optionally color n black.\n<\/li><\/ul><\/li><\/ol><p>This algorithm is abstract enough to describe many different graph\ntraversals. It allows the particular implementation to choose the node n from\namong the gray nodes; it allows choosing which and how many white successors to\ncolor gray, and it allows delaying the coloring of gray nodes black. We says\nthat such an algorithm is <i>nondeterministic<\/i> because its behavior is not\nfully defined. However, as long as it does some work on each gray node that it picks,\nany implementation that can be described in terms of this algorithm will finish. Further,\nbecause the black-white invariant is maintained, it must reach all reachable nodes in the graph.\n<\/p>\n<p>One value of defining graph search in terms of the tricolor algorithm is\nthat the tricolor algorithm works even when gray nodes are worked on\nconcurrently, as long as the black-white invariant is maintained. Thinking\nabout this invariant therefore helps us ensure that whatever graph traversal we\nchoose will work when parallelized, which is increasingly important.\n<\/p>\n<h2>Breadth-first search<\/h2>\n<p>\nBreadth-first search (BFS) is a graph traversal algorithm that explores nodes in the\norder of their distance from the roots, where distance is defined as the minimum path length\nfrom a root to the node. Its pseudo-code looks like this:\n<\/p>\n<pre class=\"code\">\n\/\/ let s be the source node\nfrontier = new Queue()\nmark root visited (set root.distance = 0)\nfrontier.push(root)\nwhile frontier not empty {\n    Vertex v = frontier.pop()\n    for each successor v' of v {\n\tif v' unvisited {\n\t    frontier.push(v')\n\t    mark v' visited (v'.distance = v.distance + 1)\n\t}\n    }\n}\n<\/pre>\n<p>\nHere the white nodes are those not marked as visited, the gray nodes are those\nmarked as visited and that are in <code>fronter<\/code>, and the black nodes are\nvisited nodes no longer in <code>frontier<\/code>. Rather than having a visited flag,\nwe can keep track of a node's distance in the field <code>v.distance<\/code>. When\na new node is discovered, its distance is set to be one greater than its predecessor\n<code>v<\/code>.\n<\/p><p>\nWhen <code>frontier<\/code> is a first-in, first-out (FIFO) queue, we get\nbreadth-first search. All the nodes on the queue have a minimum path length\nwithin one of each other. In general, there is a set of nodes to be popped off,\nat some distance <i>k<\/i> from the source, and another set of elements, later\non the queue, at distance <i>k+1<\/i>. Every time a new node is pushed onto the\nqueue, it is at distance <i>k+1<\/i> until all the nodes at distance <i>k<\/i>\nare gone, and <i>k<\/i> then goes up by one.  Therefore newly pushed nodes are\nalways at a distance at least as great as any other gray node.\n<\/p>\n<p>Suppose that we run this algorithm on the following graph, assuming\nthat successors are visited in alphabetic order from any given node:\n:<\/p>\n<div style=\"margin: auto;\"><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/graph1.png\" alt=\"image\" \/><\/div>\n<p>\nIn that case, the following sequence of nodes pass through the queue,\nwhere each node is annotated by its minimum distance from the source node A.\nNote that we're pushing onto the right of the queue and popping from the left.\n<\/p>\n<pre>\nA0  B1  D1  E1  C2\n<\/pre>\n<p>\nClearly, nodes are popped in distance order: A, B, D, E, C. This is very useful\nwhen we are trying to find the shortest path through the graph to something.\nWhen a queue is used in this way, it is known as a <b>worklist<\/b>; it keeps\ntrack of work left to be done.<\/p>\n<h2>Depth-first search<\/h2>\n<p>\nWhat if we were to replace the FIFO queue with a LIFO stack?\nIn that case we get a completely different order of traversal.\nAssuming that successors are pushed onto the stack in <i>reverse<\/i>\nalphabetic order, the successive stack states look like this:\n<\/p>\n<pre>\nA\nB D E\nC D E\nD E\nE\n<\/pre>\n<p>With a stack, the search will proceed from a given node as far as it can\nbefore backtracking and considering other nodes on the stack. For example,\nthe node E had to wait until all nodes reachable from B and D were considered.\nThis is a <b>depth-first search<\/b>.\n<\/p>\n<p>\nA more standard way of writing depth-first search is as a recursive function,\nusing the program stack as the stack above. We start with every node white\nexcept the starting node and apply the function <code>DFS<\/code> to the\nstarting node:\n<\/p>\n<pre>\nDFS(Vertex v) {\n    mark v visited\n    set color of v to gray\n    for each successor v' of v {\n\tif v' not yet visited {\n\t    DFS(v')\n\t}\n    }\n    set color of v to black\n}\n<\/pre>\n<p>\nYou can think of this as a person walking through the graph following arrows\nand never visiting a node twice except when backtracking, when a dead end is\nreached. Running this code on the graph above yields the following graph\ncolorings in sequence, which are reminiscent of but a bit different from what\nwe saw with the stack-based version:\n<\/p>\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph1.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph2.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph3.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph4.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph5.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph6.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph7.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph8.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph9.png\" alt=\"image\" \/><img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/dfsgraph10.png\" alt=\"image\" \/><p>\nNotice that at any given time there is a single path of gray nodes\nleading from the starting node and leading to the current node\n<code>v<\/code>. This path corresponds to the stack in the earlier\nimplementation, although the nodes end up being visited in a different\norder because the recursive algorithm only marks one successor gray at\na time.\n<\/p>\n<p>The sequence of calls to DFS form a tree. This is called the <b>call\ntree<\/b> of the program, and in fact, any program has a call tree. In this case\nthe call tree is a subgraph of the original graph:\n<\/p><div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/call-tree.png\" alt=\"image\" \/><\/div>\n<p>The algorithm maintains an amount of state that is\nproportional to the size of this path from the root. This makes DFS rather different\nfrom BFS, where the amount of state (the queue size) corresponds to the\nsize of the perimeter of nodes at distance <i>k<\/i> from the starting\nnode. In both algorithms the amount of state can be O(|V|). For DFS this\nhappens when searching a linked list. For BFS this happens when\nsearching a graph with a lot of branching, such as a binary tree,\nbecause there are 2<i><sup>k<\/sup><\/i> nodes at distance <i>k<\/i> from\nthe root. On a balanced binary tree, DFS maintains state proportional to\nthe height of the tree, or O(log |V|). Often the graphs that we want to\nsearch are more like trees than linked lists, and so DFS tends to run\nfaster.\n<\/p>\n<p>\nThere can be at most |V| calls to DFS_visit. And the body of the loop\non successors can be executed at most |E| times. So the asymptotic performance of\nDFS is O(|V| + |E|), just like for breadth-first search.\n<\/p>\n<p>\nIf we want to search the whole graph, then a single recursive traversal may not\nsuffice. If we had started a traversal with node C, we would miss all\nthe rest of the nodes in the graph. To do a depth-first search of an entire\ngraph, we call DFS on an arbitrary unvisited node, and repeat until every node\nhas been visited.  For example, consider the original graph expanded with two new\nnodes F and G:\n<\/p>\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/graph2.png\" alt=\"image\" \/><\/div>\n<p>\nDFS starting at A will not search all the nodes. Suppose we next choose F to\nstart from.  Then we will reach all nodes. Instead of constructing just one\ntree that is a subgraph of the original graph, we get a <b>forest<\/b> of two trees:\n<\/p>\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/graph2-forest.png\" alt=\"image\" \/><\/div>\n<h2>Topological sort\n<\/h2><p>\nOne of the most useful algorithms on graphs is topological sort, in\nwhich the nodes of an acyclic graph are placed in an order consistent with the\nedges of the graph. This is useful when you need to order a set of\nelements where some elements have no ordering constraint relative to other\nelements.\n<\/p>\n<p>\nFor example, suppose you have a set of tasks to perform, but some tasks\nhave to be done before other tasks can start. In what order should you\nperform the tasks? This problem can be solved by representing the tasks\nas node in a graph, where there is an edge from task 1 to task 2 if\ntask 1 must be done before task 2. Then a topological sort of the\ngraph will give an ordering in which task 1 precedes task 2. Obviously,\nto topologically sort a graph, it cannot have cycles. For example, if\nyou were making lasagna, you might need to carry out tasks described by\nthe following graph:\n<\/p>\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/lasagna.png\" alt=\"image\" \/><p>There is some flexibility about what order to do things in, but clearly we\nneed to make the sauce before we assemble the lasagna.  A topological sort will\nfind some ordering that obeys this and the other ordering constraints.  Of\ncourse, it is impossible to topologically sort a graph with a cycle in it.\n<\/p>\n<p>The key observation is that a node finishes (is marked black) after all\nof its descendants have been marked black. Therefore, a node that is marked\nblack later must come earlier when topologically sorted. A \na <i>postorder<\/i> traversal generates nodes in the reverse of\na topological sort:\n<\/p>\n<blockquote>\n<div><b>Algorithm:<\/b><p>\nPerform a depth-first search over the entire graph, starting anew with\nan unvisited node if previous starting nodes did not visit every node.\nAs each node is finished (colored black), put it on the head of an\ninitially empty list. This clearly takes time linear in the size of\nthe graph: O(|V| + |E|).\n<\/p>\n<\/div><\/blockquote>\n<p>\nFor example, in the traversal example above, nodes are marked black in the\norder C, E, D, B, A. Reversing this, we get the ordering A, B, D, E, C. This is\na topological sort of the graph. Similarly, in the lasagna example, assuming\nthat we choose successors top-down, nodes are marked black in the order bake,\nassemble lasagna, make sauce, fry sausage, boil pasta, grate cheese. So the\nreverse of this ordering gives us a recipe for successfully making lasagna,\neven though successful cooks are likely to do things more in parallel!\n<\/p>\n\n<h3>Detecting cycles<\/h3>\n<p>\nSince a node finishes after its descendents, a cycle involves a gray node\npointing to one of its gray ancestors that hasn't finished yet. If one of a\nnode's successors is gray, there must be a cycle. To detect cycles in graphs,\ntherefore, we choose an arbitrary white node and run DFS. If that completes and\nthere are still white nodes left over, we choose another white node arbitrarily\nand repeat. Eventually all nodes are colored black. If at any time we follow an\nedge to a gray node, there is a cycle in the graph. Therefore, cycles can be detected in\nO(|V+E|) time.\n<\/p>\n<h3>Edge classifications<\/h3>\n<p>\nWe can classify the various edges of the graph based on the color of the node\nreached when the algorithm follows the edge. Here is the expanded (A\u2013G) graph\nwith the edges colored to show their classification.\n<\/p>\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/graph1-classified.png\" alt=\"image\" \/><\/div>\n<p>\nNote that the classification of\nedges depends on what trees are constructed, and therefore depends on what node\nwe start from and in what order the algorithm happens to select successors to\nvisit.\n<\/p>\n<p>\nWhen the destination node of a followed edge is white, the algorithm performs a\nrecursive call. These edges are called <b>tree edges<\/b>, shown as solid black\narrows. The graph looks different in this picture because the nodes have been\nmoved to make all the tree edges go downward.  We have already seen that tree\nedges show the precise sequence of recursive calls performed during the\ntraversal.\n<\/p>\n<p> When the destination of the followed edge is gray, it is a <b>back\nedge<\/b>, shown in red. Because there is only a single path of gray\nnodes, a back edge is looping back to an earlier gray node, creating a\ncycle. A graph has a cycle if and only if it contains a back edge when\ntraversed from some node.\n<\/p>\n<p> When the destination of the followed edge is colored black, it is a\n<b>forward edge<\/b> or a <b>cross edge<\/b>. It is a cross edge if\nit goes between one tree and another in the forest; otherwise it is a forward edge.\n<\/p>\n<h2>Detecting cycles\n<\/h2><p>It is often useful to know whether a graph has cycles.  To detect\nwhether a graph has cycles, we perform a depth-first search of the\nentire graph.  If a back edge is found during any traversal, the graph\ncontains a cycle. If all nodes have been visited and no back edge has\nbeen found, the graph is acyclic.\n<\/p>\n<h2>Connected components<\/h2>\n<p>Graphs need not be connected, although we have been drawing\nconnected graphs thus far. A graph is connected if there is a\npath between every two nodes. However, it is entirely possible to have a graph\nin which there is no path from one node to another node, even\nfollowing edges backward. For connectedness, we don't care\nwhich direction the edges go in, so we might as well consider an\nundirected graph. A <b>connected component<\/b> is\na subset S such that for every two adjacent vertices\nv and v', either v and v' are both in S or neither one is.\n<\/p>\n<p>\nFor example, the following undirected graph has three\nconnected components:\n<\/p>\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/components.png\" alt=\"image\" \/><p>\nThe connected components problem is to determine how many connected\ncomponents make up a graph, and to make it possible to find, for each\nnode in the graph, which component it belongs to. This can be a useful way\nto solve problems. For example, suppose\nthat different components correspond to different jobs that need to be\ndone, and there is an edge between two components if they need to be\ndone on the same day. Then to find out what is the maximum number of\ndays that can be used to carry all the jobs, we need to count the\ncomponents.\n<\/p>\n<blockquote>\n<div><b>Algorithm:<\/b><p>\nPerform a depth-first search over the graph. As each traversal starts,\ncreate a new component. All nodes reached during the traversal belong\nto that component. The number of traversals done during the depth-first\nsearch is the number of components. Note that if the graph is directed, the\nDFS needs to follow both in- and out-edges.\n<\/p>\n<\/div><\/blockquote>\n<p>For directed graphs, it is usually more useful to define <b>strongly\nconnected components<\/b>. A strongly connected component (SCC)\nis a <i>maximal<\/i> subset of vertices such that every vertex in the set\nis reachable from every other. All cycles in a graph are part of the\nsame strongly connected component, which means that every graph can be\nviewed as a DAG composed of SCCs.\nThere is a simple and  efficient algorithm due to Kosaraju that uses\ndepth-first search twice: <\/p><ol><li>Topologically sort the nodes using DFS. The SCCs will appear in sequence.\n<\/li><li>Now traverse the <i>transposed<\/i> graph, but pick new (white) roots in\n    topological order. Each new subtraversal reaches a distinct SCC.\n<\/li><\/ol><p>\nFor example, consider the following graph, which is clearly not a DAG:\n<\/p><div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/scc-example.png\" alt=\"image\" \/><\/div>\n<p>\nRunning a depth-first traversal in which we  happen to choose children left-to-right,\nand extracting the nodes in reverse postorder, we obtain the ordering 1, 4, 5, 6, 2, 3, 7.\nNotice that the SCCs occur sequentially within this ordering.\nThe job of the second part\nof the algorithm is to identify the boundaries.\n<\/p>\n<p>In the second phase, we start with 1 and find it has no predecessors, so {1} is the first\nSCC. We then start with 4 and find that 5 and 6 are reachable via backward edges, so the second SCC is {4,5,6}.\nStarting from 2, we discover {2,3}, and the final SCC is {7}. The resulting DAG of SCCs is the following:\n<\/p><div style=\"text-align: center;\">\n<img src=\"https:\/\/www.cs.cornell.edu\/courses\/cs2112\/2012sp\/lectures\/lec24\/example-dag.png\" alt=\"image\" \/><\/div>\n<p>\nNotice that the SCCs are also topologically sorted by this algorithm.\n<\/p>\n<p>One inconvenience of this algorithm is that it requires being able to walk\nedges backward. Tarjan's algorithm for identifying strongly connected\ncomponents is only slightly more complicated, yet performs just one depth-first\ntraversal of the graph and only in a forward direction.\n<\/p>\n<p>\nSee Cormen, Leiserson, and Rivest for more details.\n<\/p>\n<hr \/><h2>Further reading<\/h2>\n<ul><li> Carrano. <i>Data Structures and Abstractions with Java<\/i>, Chapter 31.\n<\/li><li> Cormen, Leiserson, and Rivest. <i>Introduction to Algorithms<\/i>, Chapter 23.\n<\/li><\/ul><hr \/><p><small>Notes by Andrew Myers, 4\/19\/12.<\/small><\/p>","created_at":"2018-12-09T22:19:30+0000","updated_at":"2018-12-09T22:19:50+0000","annotations":[],"mimetype":"text\/html","reading_time":14,"domain_name":"www.cs.cornell.edu","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["www","internet"],"is_public":false,"id":737,"title":"Opinion | How to Save the Web","url":"https:\/\/www.nytimes.com\/2018\/12\/06\/opinion\/tim-berners-lee-saving-the-internet.html","content":"<div class=\"css-u5vfum StoryBodyCompanionColumn\"><div class=\"css-4w7y5l\"><p class=\"css-1ygdjhk e2kc3sl0\"><strong class=\"css-8qgvsz euv7paa0\"><em class=\"css-2fg4z9 ehxkw330\">Turning Point:<\/em><\/strong><em class=\"css-2fg4z9 ehxkw330\"> By 2019, 50 percent of the global population will be online.<\/em><\/p><p class=\"css-1ygdjhk e2kc3sl0\">In recent years, it has become clear that the web is not living up to the high hopes we had for it. Built as an open tool for collaboration and empowerment, the web has been hijacked by crooks and trolls who have used it to manipulate people all over the world.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">To preserve a web that serves all of humanity, not just the privileged and the powerful, we will have to fight for it. That\u2019s why I\u2019m asking governments, companies and citizens across the globe to commit to a set of core principles for the web.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">By the end of 2019, 50 percent of the world\u2019s population will be using the internet, according to a <a class=\"css-1g7m0tk\" href=\"https:\/\/www.itu.int\/dms_pub\/itu-s\/opb\/pol\/S-POL-BROADBAND.19-2018-PDF-E.pdf\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">recent report<\/a> from the United Nations-affiliated Broadband Commission for Digital Development. At any other time in the web\u2019s 30-year history, the collective response to this tipping point would likely have been: \u201cGreat! Now let\u2019s get everyone else connected as quickly as possible.\u201d But the world has changed. After years of the web being seen as a potential net force for good, such technological optimism has been eclipsed by fears that the web might be damaging our societies.<\/p><\/div><aside class=\"css-1m2ozyi\"><\/aside><\/div><div class=\"css-u5vfum StoryBodyCompanionColumn\"><div class=\"css-4w7y5l\"><p class=\"css-1ygdjhk e2kc3sl0\">These worries are justified. In recent years, we\u2019ve seen governments engage in <a class=\"css-1g7m0tk\" href=\"https:\/\/www.bloomberg.com\/features\/2018-government-sponsored-cyber-militia-cookbook\/\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">state-sponsored trolling<\/a> to quash dissent and attack opposition. We\u2019ve seen hacking and foreign interference distort politics and <a class=\"css-1g7m0tk\" href=\"https:\/\/www.politico.eu\/article\/russia-probe-us-senate-intelligence-committee-backs-assessment-of-russia-election-meddling\/\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">undermine elections<\/a>. And we\u2019ve seen how the spread of fake news on social media can trigger chaos, confusion and <a class=\"css-1g7m0tk\" href=\"https:\/\/www.wired.com\/story\/how-facebooks-rise-fueled-chaos-and-confusion-in-myanmar\/\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">lethal violence<\/a>.<\/p><\/div><aside class=\"css-1m2ozyi\"><\/aside><\/div><div data-testid=\"photoviewer-wrapper\" class=\"css-79elbk ehw59r11\"><div class=\"css-z3e15g\" data-testid=\"photoviewer-wrapper-hidden\"><div data-testid=\"photoviewer-children\" class=\"css-1a48zt4 ehw59r111\"><figure class=\"css-1lghxe1 e1a8i6eb0\" aria-label=\"media\" role=\"group\" itemprop=\"associatedMedia\" itemscope=\"itemscope\" itemid=\"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE1\/merlin_146370507_2e247e7f-62da-4e6a-b81b-742bc26f7342-articleLarge.jpg?quality=90&amp;auto=webp\" itemtype=\"http:\/\/schema.org\/ImageObject\"><div class=\"css-1xdhyk6 e1vv25i80\">Image<img alt=\"\" class=\"css-1m50asq\" src=\"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE1\/merlin_146370507_2e247e7f-62da-4e6a-b81b-742bc26f7342-articleLarge.jpg?quality=75&amp;auto=webp&amp;disable=upscale\" srcset=\"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE1\/merlin_146370507_2e247e7f-62da-4e6a-b81b-742bc26f7342-articleLarge.jpg?quality=90&amp;auto=webp 600w, https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE1\/merlin_146370507_2e247e7f-62da-4e6a-b81b-742bc26f7342-jumbo.jpg?quality=90&amp;auto=webp 1024w, https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE1\/merlin_146370507_2e247e7f-62da-4e6a-b81b-742bc26f7342-superJumbo.jpg?quality=90&amp;auto=webp 2048w\" itemprop=\"url\" itemid=\"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE1\/merlin_146370507_2e247e7f-62da-4e6a-b81b-742bc26f7342-articleLarge.jpg?quality=75&amp;auto=webp&amp;disable=upscale\" \/><\/div><figcaption itemprop=\"caption description\" class=\"css-1l44abu e3zkro30\">Students learn how to use Google on their laptops in Anuppur, India.CreditAtul Loke for The New York Times<\/figcaption><\/figure><\/div><\/div><div class=\"css-u5vfum StoryBodyCompanionColumn\"><div class=\"css-4w7y5l\"><p class=\"css-1ygdjhk e2kc3sl0\">When we learned last year that Cambridge Analytica had used the personal information of up to <a class=\"css-1g7m0tk\" href=\"https:\/\/www.theguardian.com\/technology\/2018\/apr\/04\/facebook-cambridge-analytica-user-data-latest-more-than-thought\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">87 million Facebook users<\/a> to influence voters in the 2016 American presidential election, we woke up to the fact that we had lost control of our data <em class=\"css-2fg4z9 ehxkw330\">\u2014 <\/em>and that the consequences could change the world.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">But we must not give up on the promise of the web. All technologies come with risks. We drive cars despite the possibility of serious accidents. We take prescription drugs despite the danger of abuse and addiction. We build safeguards into new innovations so we can manage the risks while benefiting from the opportunities.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">The web is a global platform <em class=\"css-2fg4z9 ehxkw330\">\u2014<\/em> its challenges stretch across borders and cultures. Just as the web was built by millions of people collaborating around the world, its future relies on our collective ability to make it a better tool for everyone.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">As we forge the web of tomorrow, we need a set of guiding principles that can define the kind of web we want. Identifying these will not be easy <em class=\"css-2fg4z9 ehxkw330\">\u2014 <\/em>any agreement that covers a diverse group of countries, cultures and interests will never be. But I believe it\u2019s possible to develop a set of basic ideals that we can all agree on, and that will make the web work better for everyone, including the 50 percent of the world\u2019s population that has yet to come online.<\/p><\/div><aside class=\"css-1m2ozyi\"><\/aside><\/div><div class=\"css-u5vfum StoryBodyCompanionColumn\"><div class=\"css-4w7y5l\"><p class=\"css-1ygdjhk e2kc3sl0\">Governments, companies and individuals all have unique roles to play. The <a class=\"css-1g7m0tk\" href=\"https:\/\/webfoundation.org\/\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">World Wide Web Foundation<\/a>, an organization I founded in 2009 to protect the web as a public good, has drawn up a set of <a class=\"css-1g7m0tk\" href=\"https:\/\/fortheweb.webfoundation.org\/\" title=\"\" rel=\"noopener noreferrer\" target=\"_blank\">core principles<\/a> outlining the responsibilities that each party has to protect a web that serves all of humanity. We\u2019re asking everyone to sign on to these principles and join us as we create a formal Contract for the Web in 2019.<\/p><\/div><aside class=\"css-1m2ozyi\"><\/aside><\/div><div data-testid=\"photoviewer-wrapper\" class=\"css-79elbk ehw59r11\"><div class=\"css-z3e15g\" data-testid=\"photoviewer-wrapper-hidden\"><div data-testid=\"photoviewer-children\" class=\"css-1a48zt4 ehw59r111\"><figure class=\"css-1lghxe1 e1a8i6eb0\" aria-label=\"media\" role=\"group\" itemprop=\"associatedMedia\" itemscope=\"itemscope\" itemid=\"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE2\/merlin_146372604_2f8ea0a4-2c4f-4305-8ec2-9a1659167fc0-articleLarge.jpg?quality=90&amp;auto=webp\" itemtype=\"http:\/\/schema.org\/ImageObject\"><div class=\"css-1xdhyk6 e1vv25i80\">Image<div class=\"css-zjzyr8\"><div data-testid=\"lazyimage-container\"><\/div><\/div><figcaption itemprop=\"caption description\" class=\"css-1l44abu e3zkro30\">A copy of the world\u2019s first web page, from 1992.CreditFabrice Coffrini\/Agence France-Presse \u2014 Getty Images<\/figcaption><\/div><\/figure><\/div><\/div><div class=\"css-u5vfum StoryBodyCompanionColumn\"><div class=\"css-4w7y5l\"><p class=\"css-1ygdjhk e2kc3sl0\">The principles specify that governments are responsible for connecting their citizens to an open web that respects their rights. They say that internet companies must play their part in making sure the web is safe, accessible and protects user data. They make it clear that individual citizens have a responsibility to act with compassion and challenge negative behavior they wouldn\u2019t tolerate offline. Most importantly, the principles say that we have to stand up and fight for a web that serves everyone. If we, the billions of people using the web, don\u2019t defend it, who will?<\/p><p class=\"css-1ygdjhk e2kc3sl0\">Based on these fundamental principles, the Contract for the Web will establish a new set of norms to guide the digital policy agendas of governments and the decisions of companies as they build tomorrow\u2019s web technologies. Those who support and help develop the contract will not only show their commitment to the future of the web; they will help shape it. If we want a web that works for us, we must work for the web\u2019s future.<\/p><div data-testid=\"photoviewer-wrapper\" class=\"css-79elbk ehw59r11\"><div class=\"css-z3e15g\" data-testid=\"photoviewer-wrapper-hidden\"><div data-testid=\"photoviewer-children\" class=\"css-1a48zt4 ehw59r111\"><figure class=\"css-1mpi71h e1a8i6eb0\" aria-label=\"media\" role=\"group\" itemprop=\"associatedMedia\" itemscope=\"itemscope\" itemid=\"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/03TP-BERNERS-LEE-author\/03TP-BERNERS-LEE-author-articleLarge.jpg?quality=90&amp;auto=webp\" itemtype=\"http:\/\/schema.org\/ImageObject\"><div class=\"css-zgakxe e1vv25i80\">Image<div class=\"css-zjzyr8\"><div data-testid=\"lazyimage-container\"><\/div><\/div><figcaption itemprop=\"caption description\" class=\"css-198aork e3zkro30\">Tim Berners-LeeCreditHenry Thomas<\/figcaption><\/div><\/figure><\/div><\/div><p class=\"css-1ygdjhk e2kc3sl0\">Ten years ago, when Rosemary Leith and I started the World Wide Web Foundation, fewer than one in four people were online. Our task was to get more people connected and to keep the web open and free so that everyone could benefit from all it had to offer. The web saves lives and creates livelihoods. It puts the world\u2019s information at our fingertips and connects us with friends and family across the globe. It powers social movements and has created countless new industries, fueling widespread innovation.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">As a comparatively young invention, this is just the beginning of what the web has to offer. Imagine what we will be able to accomplish once the next huge tranche of the world\u2019s population is online and contributing to the web\u2019s explosive creativity.<\/p><p class=\"css-1ygdjhk e2kc3sl0\">The good news is, the appetite to take on the web\u2019s challenges has never been greater. Let\u2019s make sure the next billion people connect to a web worth having. Let\u2019s make 2019 the year we push back against the forces subverting the open spirit of the web. We need a free and open web for everyone.<\/p><\/div><aside class=\"css-1m2ozyi\"><\/aside><\/div><\/div><\/div><\/div>","created_at":"2018-12-09T17:18:40+0000","updated_at":"2018-12-09T22:44:23+0000","published_at":"2018-12-06T20:00:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.nytimes.com","preview_picture":"https:\/\/static01.nyt.com\/images\/2018\/12\/03\/opinion\/international-world\/03serners-leeHRjpg-INYT\/03serners-leeHRjpg-INYT-facebookJumbo.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["programming","programming languages"],"is_public":false,"id":736,"title":"The Programming Languages Zoo","url":"http:\/\/plzoo.andrej.com\/","content":"<p>The Programming Languages Zoo is a collection of miniature programming languages which\ndemonstrates various concepts and techniques used in programming language design and\nimplementation. It is a good starting point for those who would like to implement their\nown programming language, or just learn how it is done.<\/p><p>The following features are demonstrated:<\/p><ul><li>functional, declarative, object-oriented, and procedural languages<\/li>\n  <li>source code parsing with a parser generator<\/li>\n  <li>recoding of source code positions<\/li>\n  <li>pretty-printing of values<\/li>\n  <li>interactive shell (REPL) and non-interactive file processing<\/li>\n  <li>untyped, statically and dynamically typed languages<\/li>\n  <li>type checking and type inference<\/li>\n  <li>subtyping, parametric polymorphism, and other kinds of type systems<\/li>\n  <li>eager and lazy evaluation strategies<\/li>\n  <li>recursive definitions<\/li>\n  <li>exceptions<\/li>\n  <li>interpreters and compilers<\/li>\n  <li>abstract machine<\/li>\n<\/ul><h2 id=\"installation\">Installation<\/h2><p>See the <a href=\"http:\/\/plzoo.andrej.com\/install.html\">installation &amp; compilation instructions<\/a>.<\/p><h2 id=\"the-languages\">The languages<\/h2><p>The following languages are on display:<\/p><dl><dt><a href=\"http:\/\/plzoo.andrej.com\/language\/miniml_error.html\">miniml_error<\/a><\/dt>\n  <dd>like miniml that can also abort execution<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/minihaskell.html\">minihaskell<\/a><\/dt>\n  <dd>lazy, functional, integers, booleans, lists, recursion, statically typed<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/miniprolog.html\">miniprolog<\/a><\/dt>\n  <dd>logic programming, Horn clauses, unification<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/levy.html\">levy<\/a><\/dt>\n  <dd>call-by-push value, statically typed<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/comm.html\">comm<\/a><\/dt>\n  <dd>A procedural language with integer arithmetic, local variables, conditional statements, <code class=\"highlighter-rouge\">while<\/code> loops and <code class=\"highlighter-rouge\">print<\/code>, compiled to simple machine code.<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/calc.html\">calc<\/a><\/dt>\n  <dd>integer arithmetic <code class=\"highlighter-rouge\">+<\/code>, <code class=\"highlighter-rouge\">-<\/code>, <code class=\"highlighter-rouge\">*<\/code>, <code class=\"highlighter-rouge\">\/<\/code><\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/sub.html\">sub<\/a><\/dt>\n  <dd>eager, mutable records, statically typed, subtyping<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/boa.html\">boa<\/a><\/dt>\n  <dd>object-oriented, eager, first-class functions, dynamic types, extensible objects<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/lambda.html\">lambda<\/a><\/dt>\n  <dd>untyped \u03bb-calculus, several evaluation strategies<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/calc_var.html\">calc_var<\/a><\/dt>\n  <dd>integer arithmetic <code class=\"highlighter-rouge\">+<\/code>, <code class=\"highlighter-rouge\">-<\/code>, <code class=\"highlighter-rouge\">*<\/code>, <code class=\"highlighter-rouge\">\/<\/code>, variables<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/miniml.html\">miniml<\/a><\/dt>\n  <dd>eager, functional, recursive functions, statically typed, compiler, abstract machine<\/dd>\n  <dt><a href=\"http:\/\/plzoo.andrej.com\/language\/poly.html\">poly<\/a><\/dt>\n  <dd>lazy, functional, statically typed, parametric polymorphism, type inference<\/dd>\n<\/dl><h2 id=\"usage\">Usage<\/h2><p>The languages are not really meant to be used. Rather, you should <a href=\"https:\/\/github.com\/andrejbauer\/plzoo\/tree\/master\/src\">read and study the\nsource code<\/a>, which is decorated\nwith ample comments. Also, each language <code class=\"highlighter-rouge\">lang<\/code> has its own <code class=\"highlighter-rouge\">README.markdown<\/code> and\n<code class=\"highlighter-rouge\">example.lang<\/code> in the subdirectory <code class=\"highlighter-rouge\">src\/lang<\/code>.<\/p><p>Nevertheless, all the language are fully functioning miniature versions of real languages\nand can be executed. For each language <code class=\"highlighter-rouge\">lang<\/code> you can:<\/p><ol><li>\n    <p>see what command-line options are available with<\/p>\n    <div class=\"highlighter-rouge\"><div class=\"highlight\"><pre> .\/lang.native --help\n<\/pre><\/div><\/div>\n  <\/li>\n  <li>\n    <p>run the toplevel with<\/p>\n    <div class=\"highlighter-rouge\"><div class=\"highlight\"><pre> .\/lang.native\n<\/pre><\/div><\/div>\n  <\/li>\n  <li>\n    <p>run files non-interactively with<\/p>\n    <div class=\"highlighter-rouge\"><div class=\"highlight\"><pre> .\/lang.native &lt;file&gt; &lt;file&gt; ...`\n<\/pre><\/div><\/div>\n  <\/li>\n  <li>\n    <p>load files and enter the toplevel<\/p>\n    <div class=\"highlighter-rouge\"><div class=\"highlight\"><pre> .\/lang.native -l &lt;file&gt; -l &lt;file&gt; ...`\n<\/pre><\/div><\/div>\n  <\/li>\n<\/ol><ul><li><a href=\"http:\/\/andrej.com\/\">Andrej Bauer<\/a><\/li>\n  <li><a href=\"http:\/\/matija.pretnar.info\/\">Matija Pretnar<\/a><\/li>\n<\/ul><h2 id=\"license\">License<\/h2><p>The project is open source and released under the permissive <a href=\"http:\/\/plzoo.andrej.com\/license.html\">MIT license<\/a>.<\/p><h2 id=\"contributing\">Contributing<\/h2><p>New contributions are <em>very<\/em> welcome. If you would like to contribute to the project,\nplease contact us through the <a href=\"https:\/\/github.com\/andrejbauer\/plzoo\">GitHub project page<\/a>:<\/p><ul><li>If you discover a problem, <a href=\"https:\/\/github.com\/andrejbauer\/plzoo\/issues\/new\">open an issue<\/a>.<\/li>\n  <li>Even better, fix the problem and <a href=\"https:\/\/github.com\/andrejbauer\/plzoo\/compare\">submit a pull request<\/a>!<\/li>\n  <li>If you would like to help but do not know how, have a look at\n<a href=\"https:\/\/github.com\/andrejbauer\/plzoo\/issues\">open issues<\/a> and volunteer to resolve one.<\/li>\n  <li>If you have an idea for a new language, we will be happy to take it in. Please note that all\nthe languages are purposely kept simple for educational purposes.<\/li>\n<\/ul><p>Before you contibute a new langauge, please read these\n<a href=\"http:\/\/plzoo.andrej.com\/contributions.html\">guidelines for contributing<\/a>.<\/p>","created_at":"2018-12-09T17:16:49+0000","updated_at":"2018-12-09T22:44:55+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"plzoo.andrej.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["programming","math","book"],"is_public":false,"id":735,"title":"A Programmer's Introduction to Mathematics","url":"https:\/\/pimbook.org\/","content":"<p><\/p><table><tr><td class=\"links-left\"> <a href=\"https:\/\/pimbook.org\/pdf\/pim_first_pages.pdf\">First pages<\/a> <\/td>\n                <td> \u2022 <\/td>\n                <td> A preview of the book <\/td>\n              <\/tr><tr><td class=\"links-left\"> <a href=\"https:\/\/amzn.to\/2KzMT5t\">Physical book<\/a> <\/td>\n                <td> \u2022 <\/td>\n                <td> Purchase through Amazon <\/td>\n              <\/tr><tr><td class=\"links-left\"> <a href=\"https:\/\/gum.co\/pim-book\">E-book<\/a> <\/td>\n                <td> \u2022 <\/td>\n                <td> Purchase through Gumroad <\/td>\n              <\/tr><tr><td class=\"links-left\"> <a href=\"https:\/\/github.com\/pim-book\/programmers-introduction-to-mathematics\">Github<\/a> <\/td>\n                <td> \u2022 <\/td>\n                <td> Code from each chapter <\/td>\n              <\/tr><tr><td class=\"links-left\"> <a href=\"https:\/\/pimbook.org\/demo\/index.html\">Demos<\/a> <\/td>\n                <td> \u2022 <\/td>\n                <td> Interactive demos <\/td>\n              <\/tr><tr><td class=\"links-left\"> <a href=\"https:\/\/docs.google.com\/forms\/d\/e\/1FAIpQLScZ0EZfIbsl3l7FeMQ3q1wskHkWXqEWT6b739C2mZEaltrkuw\/viewform?usp=sf_link\">Submit an erratum<\/a> <\/td>\n                <td> \u2022 <\/td>\n                <td> <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1TKNsal6wHkTWsDkjY-T9SGkDN2ax7Ag4EVoRlFadhh4\/edit?usp=sharing\">(Submitted errata)<\/a> <\/td>\n              <\/tr><\/table>","created_at":"2018-12-09T13:33:59+0000","updated_at":"2018-12-09T22:45:11+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"pimbook.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","phpunit"],"is_public":false,"id":734,"title":"Testing methods where DateTime object is used with PhpUnit","url":"http:\/\/www.inanzzz.com\/index.php\/post\/dhg6\/testing-methods-where-datetime-object-is-used-with-phuunit","content":"<p>It can be a bit tricky when it comes to testing methods where <code>DateTime<\/code> object is used. This example shows us how to test such methods.<\/p><div class=\"prettyprint_cover\"><pre class=\"prettyprint linenums lang-rb\">declare(strict_types=1);namespace Api\\ExamBundle\\Helper;class ExamDate<br \/>{<br \/>public function getLastExamDay()<br \/>{<br \/>$date = $this-&gt;getDate('last day of this month');if ($date == $this-&gt;getDate()) {<br \/>$date = $this-&gt;getDate('last day of next month');<br \/>}if ($date-&gt;format('l') === 'Saturday') {<br \/>$date-&gt;modify('-1 day');<br \/>} elseif ($date-&gt;format('l') === 'Sunday') {<br \/>$date-&gt;modify('-2 days');<br \/>}return $date-&gt;format('Y-m-d');<br \/>}protected function getDate(?string $time = null): DateTime<br \/>{<br \/>return new DateTime($time);<br \/>}<br \/>}<br \/><\/pre><\/div><div class=\"prettyprint_cover\"><pre class=\"prettyprint linenums lang-rb\">declare(strict_types=1);namespace tests\\Api\\ExamBundle\\Helper;use Api\\ExamBundle\\Helper;<br \/>use DateTime;<br \/>use PHPUnit_Framework_MockObject_MockObject;<br \/>use PHPUnit_Framework_TestCase;class ExamDateTest extends PHPUnit_Framework_TestCase<br \/>{<br \/>\/**<br \/>* @test<br \/>*<br \/>* @dataProvider getLastExamDayProvider<br \/>*\/<br \/>public function it_returns_last_exam_day($lastDayOfThisMonth, $today, $lastDayOfNextMonth, $expectedDay)<br \/>{<br \/>\/** @var Adaptor|PHPUnit_Framework_MockObject_MockObject $mock *\/<br \/>$mock = $this-&gt;getMockBuilder(ExamDate::class)<br \/>-&gt;disableOriginalConstructor()<br \/>-&gt;setMethods(['getDate'])<br \/>-&gt;getMock();$mock<br \/>-&gt;expects($this-&gt;at(0))<br \/>-&gt;method('getDate')<br \/>-&gt;willReturn(new DateTime($lastDayOfThisMonth));$mock<br \/>-&gt;expects($this-&gt;at(1))<br \/>-&gt;method('getDate')<br \/>-&gt;willReturn(new DateTime($today));if ($lastDayOfNextMonth) {<br \/>$mock<br \/>-&gt;expects($this-&gt;at(2))<br \/>-&gt;method('getDate')<br \/>-&gt;willReturn(new DateTime($lastDayOfNextMonth));<br \/>}$this-&gt;assertEquals($expectedDay, $mock-&gt;getLastExamDay());<br \/>}\/**<br \/>* 1)<br \/>* Last day of this month : 2015-07-31 (Friday)<br \/>* Today                  : 2015-07-31 (Friday)<br \/>* Last day of this month : 2015-08-31 (Monday) *<br \/>* Result                 : 2015-08-31 (Monday) same day of *<br \/>*<br \/>* 2)<br \/>* Last day of this month : 2015-09-30 (Wednesday)<br \/>* Today                  : 2015-09-30 (Wednesday)<br \/>* Last day of this month : 2015-10-31 (Saturday) *<br \/>* Result                 : 2015-10-30 (Friday) -1 day of *<br \/>*<br \/>* 3)<br \/>* Last day of this month : 2015-04-30 (Thursday)<br \/>* Today                  : 2015-04-30 (Thursday)<br \/>* Last day of this month : 2015-05-31 (Sunday) *<br \/>* Result                 : 2015-15-29 (Friday) -2 days of *<br \/>*<br \/>* 4)<br \/>* Last day of this month : 2015-07-31 (Friday) *<br \/>* Today                  : 2015-07-30 (Thursday)<br \/>* Last day of this month :<br \/>* Result                 : 2015-07-31 (Friday) same day of *<br \/>*<br \/>* 5)<br \/>* Last day of this month : 2015-10-31 (Saturday) *<br \/>* Today                  : 2015-10-16 (Friday)<br \/>* Last day of this month :<br \/>* Result                 : 2015-10-30 (Friday) -1 day of *<br \/>*<br \/>* 6)<br \/>* Last day of this month : 2015-05-31 (Sunday) *<br \/>* Today                  : 2015-05-15 (Friday)<br \/>* Last day of this month :<br \/>* Result                 : 2015-05-29 (Friday) -2 days of *<br \/>*\/<br \/>public function getLastExamDayProvider(): array<br \/>{<br \/>return [<br \/>['2015-07-31', '2015-07-31', '2015-08-31', '2015-08-31'],<br \/>['2015-09-30', '2015-09-30', '2015-10-31', '2015-10-30'],<br \/>['2015-04-30', '2015-04-30', '2015-05-31', '2015-05-29'],<br \/>['2015-07-31', '2015-07-30', null, '2015-07-31'],<br \/>['2015-10-31', '2015-10-16', null, '2015-10-30'],<br \/>['2015-05-31', '2015-05-15', null, '2015-05-29'],<br \/>];<br \/>}<br \/>}<br \/><\/pre><\/div>","created_at":"2018-12-08T23:55:06+0000","updated_at":"2018-12-08T23:55:10+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.inanzzz.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":733,"title":"Seminars - Programming Research Laboratory","url":"http:\/\/prl.ccs.neu.edu\/seminars.html","content":"<div class=\"row\"><div class=\"col-md-12\"><p>The PL seminar meets in WVH\n[<a href=\"http:\/\/www.ccs.neu.edu\/home\/wand\/directions.html\">directions<\/a>]. The <a href=\"http:\/\/lists.ccs.neu.edu\/pipermail\/pl-seminar\">mailing\nlist<\/a> is public. An <a href=\"https:\/\/calendar.google.com\/calendar\/embed?src=k4cg1vgb3l2n8r2ph4t01dmtpc@group.calendar.google.com&amp;ctz=America\/New_York\">HTML calendar<\/a>\n  and an <a href=\"https:\/\/calendar.google.com\/calendar\/ical\/k4cg1vgb3l2n8r2ph4t01dmtpc%40group.calendar.google.com\/public\/basic.ics\">ICAL calendar<\/a>\nare available for your convenience.<\/p><\/div><\/div><div class=\"row\"><div id=\"seminar-1\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Better Stories, Better Languages \u2014 Reframing from Programs to Programming<\/a>\n<br \/>Fran\u00e7ois-Ren\u00e9 Rideau\nAlacris\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/fare.tunes.org\">http:\/\/fare.tunes.org<\/a>\n<br \/>30 January, 2019 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Software tools imply a story. New stories can help invent new tools. Explicit stories are a great meta-tool... and provide a systematic way to improve the design of programming language design. To illustrate my point, I will present a series of pairs of stories, each time approaching a same topic with very different consequences. Many of my stories will be familiar to most of you, some other stories less so. Hopefully they will be enlightening as well as entertaining... and so will the meta-story be. (Earlier versions of this talk were given at International Lisp Conference 2009, Lisp NYC 2014-10, and more recently LambdaConf 2017.)<\/p><p>Bio\n<\/p><p>Fran\u00e7ois-Ren\u00e9 Rideau is Co-Founder and Chief Architect at Alacris.io,a company developing a \"Blockchain Layer 2 Operating System\" using formal methods. Once founder of the TUNES Project, he left academic research on Programming Languages and Distributed Systems to build industrial software at ITA Software, then Google and Bridgewater Associates, before he finally became his own Entrepreneur. Long-standing member of the Lisp community, he runs the Boston Lisp Meeting and still co-maintains ASDF (the Common Lisp build system),though he recently jumped ship to Gerbil Scheme, and now uses OCaml at work.<\/p><\/div><\/div><br \/><div id=\"seminar-2\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Whip: Higher-order Contracts for Modern Services<\/a>\n<br \/>Lucas Waye\nFacebook\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/lucaswaye.com\/\">http:\/\/lucaswaye.com\/<\/a>\n<br \/>12 December, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Modern service-oriented applications forgo semantically rich protocols and middleware when composing services. Instead, they embrace the loosely-coupled development and deployment of services that communicate via simple network protocols. Even though these applications do expose interfaces that are higher-order in spirit, the simplicity of the network protocols forces them to rely on brittle low-level encodings. To bridge the apparent semantic gap, programmers introduce ad-hoc and error-prone defensive code. Inspired by Design by Contract, we choose a different route to bridge this gap. We introduce Whip, a contract system for modern services. Whip (i) provides programmers with a higher-order contract language tailored to the needs of modern services; and (ii) monitors services at run time to detect services that do not live up to their advertised interfaces. Contract monitoring is local to a service. Services are treated as black boxes, allowing heterogeneous implementation languages without modification to services' code. Thus, Whip does not disturb the loosely coupled nature of modern services.<\/p><p>Bio\n<\/p><p>I'm currently a Software Engineer at Facebook thinking about privacy in the data warehouse. Previously I helped build TiVo's Targeted Audience Delivery platform. And before that I worked on the backend for Timeful (acquired by Google) and on Hulu's analytics platform. In graduate school I worked on Whip, a contract monitor for microservices. I received a Ph.D. and Master's degree from Harvard under the guidance of Stephen Chong and a Bachelor's degree from Cornell.<\/p><\/div><\/div><br \/><div id=\"seminar-3\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Goblins and Spritely: from the actor model to distributed virtual worlds<\/a>\n<br \/>Christoper Lemmer Webber\nActivityPub\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/dustycloud.org\/\">https:\/\/dustycloud.org\/<\/a>\n<br \/>5 December, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Goblins is an actor model implementation for Racket, inspired by object-capability secure distributed progamming languages such as E with a few new twists of its own thrown in. Goblins is the foundation for a project to build secure, player\/user-programmable virtual worlds named Spritely. Spritely builds on the success of the ideas from the ActivityPub protocol, which has succeeded at connecting together various distributed social network projects, extending them to pull in rich interactive and game aspects. See what is built so far, what's coming next, and how the actor model and object capabilities tie together all these ideas to enable a robust and secure distributed system.<\/p><p>Bio\n<\/p><p>Christopher Lemmer Webber is a user freedom advocate who works on distributed social network technology. They are most well known for their work on the W3C ActivityPub federated protocol, which connects over 1.5 million users across the distributed social web. Chris enjoys programming in lisp\/scheme flavored languages, especially Racket.<\/p><\/div><\/div><br \/><div id=\"seminar-4\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Programming language development and standardization: How they (don\u2019t) do in Haskell<\/a>\n<br \/>Vitaly Bragilevsky\nSouthern Federal University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/sfedu.ru\/www\/stat_pages22.show?p=RR\/per_eng\/D&amp;params=(p_per_id=%3E2733)\">http:\/\/sfedu.ru\/www\/stat_pages22.show?p=RR\/per_eng\/D&amp;params=(p_per_id=%3E2733)<\/a>\n<br \/>28 November, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Programming languages differ a lot in the ways of their development. Who is responsible for including new features and removing obsolete ones? What is a process of doing that? Is there any more or less formal language definition? How respected is that definition by the compilers? How is this changed if there is only one compiler? These questions are answered quite differently for various programming languages. Haskell takes its own way which was crystallized relatively recently. We have GHC Proposals in the form of the GitHub pull requests to suggest new language features. We also have the GHC Steering Committee with its own process to either accept or reject proposals. Finally, there is the Haskell 2020 Language (Haskell Prime) Committee with its own problem of being effectively dead. In this talk, I\u2019d like to describe how this system works for Haskell and why it doesn\u2019t work sometimes. I will discuss the most recent examples of transforming Haskell kinds system towards dependent types (started before the GHC Steering Committee was formed) and introducing linear types in Haskell (subject to the full-blown Committee discussion) among others. I will also characterize the current state and perspectives of the Haskell standardization process.<\/p><p>Bio\n<\/p><p>Vitaly Bragilevsky serves as both the Haskell 2020 Language Committee and the GHC Steering Committee member. He works as a Senior Lecturer at the Southern Federal University in Rostov-on-Don, Russia where he teaches undergraduate students functional programming and theory of computations. He is currently a grantee of the Fulbright Faculty Development Program holding temporary Courtesy Research Assistant position at the University of Oregon under the supervision of Prof. Zena Ariola. Vitaly Bragilevsky translated into Russian and edited translations of the several books on Haskell and the theory of programming languages. He is the author of \u2018Haskell in Depth\u2019 (Manning Publications, available via Manning\u2019s early access program).<\/p><\/div><\/div><br \/><div id=\"seminar-5\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Synthesizing Solutions to the Expression Problem<\/a>\n<br \/>George T. Heineman\nWorcester Polytechnic Institute\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/www.wpi.edu\/people\/faculty\/heineman\">https:\/\/www.wpi.edu\/people\/faculty\/heineman<\/a>\n<br \/>17 October, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The Expression Problem (EP) describes a common software issue that appears regardless of programming language or programming paradigm. As coined by Philip Wadler in 1998, EP is a new name for an old problem. Given a data type defined by cases, consider two independent extensions, namely, adding new cases to the data type and new functions over the datatype. Dozens of papers in the research literature present their solutions to the EP, using a variety of programming languages and approaches. In this talk I present a Scala-based framework that can synthesize full solutions to the EP problem. The first step is to model the domain of interest (such as mathematical expressions or geometric shapes) to ensure that our synthesized code can work with any desired application domain. Second, we demonstrate how to synthesize code that reproduces a dozen EP solutions from the literature, in C++, Java and Haskell. We designed a common API for all approaches, independent of programming language. While we do not introduce any new approach for EP, we explore the challenges that EP solutions face with binary methods and producer methods. An important contribution of this work is the scientific reproduction of known solutions from the literature.\nThis work is based on an ongoing collaboration with with Jan Bessai (TU Dortmund) and Boris D\u00fcdder (Univ. Copenhagen).<\/p><p>Bio\n<\/p><p>George T. Heineman is an Associate Professor of Computer Science at WPI. His research interests are in Software Engineering, specifically synthesizing software systems from composable units. He is co-author of \"Algorithms in a Nutshell (2ed),\" published by O'Reilly Media in 2016. Aside from his professional pursuits, George is an avid puzzler. He invented Sujiken(R), a Sudoku variation played on a right-triangle arrangement of cells in which numbers cannot repeat in a horizontal row, vertical column or diagonal in any direction.<\/p><\/div><\/div><br \/><div id=\"seminar-6\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Rehabilitating the POSIX shell<\/a>\n<br \/>Michael Greenberg\nPomona College\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.cs.pomona.edu\/~michael\/\">http:\/\/www.cs.pomona.edu\/~michael\/<\/a>\n<br \/>10 October, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>We build intricate systems with complex algorithms and invariants, aiming for guarantees of correctness and performance... and then we maintain and deploy these systems with shell scripts! What *are* shell scripts? If the POSIX shell is a programming language, what are its syntax and semantics? Can we apply PL tools to reason about the shell? Why haven't prior PL attempts at understanding the shell redeemed it?<\/p><p>Bio\n<\/p><p>Michael Greenberg is an assistant professor at Pomona College. He received his BA in Computer Science and Egyptology from Brown University (2007) and his PhD in Computer Science from the University of Pennsylvania (2013). His work has ranged from functional-reactive JavaScript (with Shriram Krishnamurthi at Brown) to runtime verification of higher-order programs using contracts (with Benjamin Pierce at Penn) to software-defined networking (with Dave Walker at Princeton) to present activities focused on Kleene algebra with tests and the POSIX shell. He is always looking for new climbing partners.<\/p><\/div><\/div><br \/><div id=\"seminar-7\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Oracle Parfait: The Flavour of Real-World Vulnerability Detection<\/a>\n<br \/>Cristina Cifuentes\nOracle Labs\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/labs.oracle.com\/pls\/apex\/f?p=labs:bio:0:21\">https:\/\/labs.oracle.com\/pls\/apex\/f?p=labs:bio:0:21<\/a>\n<br \/>3 October, 2018 1:15PM \nWVF 010<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The Parfait static code analysis tool focuses on detecting vulnerabilities that matter in C, C++, Java and PL\/SQL languages. Its focus has been on key items expected out of a commercial tool that lives in a commercial organization, namely, precision of results (i.e., high true positive rate), scalability (i.e., being able to run quickly over millions of lines of code), incremental analysis (i.e., being able to run over deltas of the code quickly), and usability (i.e., ease of integration into standard build processes, reporting of traces to the vulnerable location, etc). Today, Parfait is used by thousands of developers at Oracle worldwide on a day-to-day basis. In this presentation we\u2019ll sample a flavour of Parfait \u2014 we explore some real world challenges faced in the creation of a robust vulnerability detection tool, look into two examples of vulnerabilities that severely affected the Java platform in 2012\/2013 and most machines in 2017\/2018, and conclude by recounting what matters to developers for integration into today\u2019s continuous integration and continuous delivery (CI\/CD) pipelines.<\/p><p>Bio\n<\/p><p>Cristina is the Director of Oracle Labs Australia and an Architect at Oracle. Headquartered in Brisbane, the Lab focuses on Program Analysis as it applies to finding vulnerabilities in software and enhancing the productivity of developers worldwide. Prior to founding Oracle Labs Australia, Cristina was the Principal Investigator of the Parfait bug tracking project at Sun Microsystems, then Oracle.<\/p><\/div><\/div><br \/><div id=\"seminar-8\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">On the expressive power of user-defined effects<\/a>\n<br \/>Ohad Kammar\nUniversity of Oxford\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/www.cs.ox.ac.uk\/people\/ohad.kammar\/main.html\">https:\/\/www.cs.ox.ac.uk\/people\/ohad.kammar\/main.html<\/a>\n<br \/>3 October, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Computational effects, such as mutable state, memory allocation, non-determinism, and I\/O interaction, allow programs to implicitly exhibit complex behaviour. I will discuss three abstractions that allow programmers to extend a language with user-defined effects. Control operators (call-cc, shift\/reset, etc.)  are a well-established abstraction for user-defined effects, and I will consider the specific shift\/dollar without answer-type-modification. Since the 90s, functional languages have used monads for user-defined effects, and I will consider Filinski's monadic reflection. In the last decade, the community started looking into Plotkin and Pretnar's handlers for algebraic effects (extensible effects in Haskell, OCaml multicore, the Eff, Frank, Koka, and Links programming languages) as a programming abstraction, and I will consider deep handlers without forwarding. I will compare the relative expressive power of these three abstractions using Felleisen's notion of macro-translations.  This comparison demonstrates the sensitivity of relative expressiveness of user-defined effects to seemingly orthogonal language features. This talk is based on the paper: On the expressive power of user-defined effects: effect handlers,monadic reflection, delimited control. Yannick Forster, Ohad Kammar, Sam Lindley, and Matija Pretnar. Proceedings of the 22nd ACM SIGPLAN International Conference on Functional Programming, PACMPL 1(ICFP): 13:1-13:29 (2017),arXiv:1610.09161, DOI: 10.1145\/3110257.<\/p><p>Bio\n<\/p><p>Ohad Kammar is a postdoctoral research associate at the University of Oxford Department of Computer Science, a Career Development Fellow at Balliol College Oxford, and an incoming Royal Society University Research Fellow at the University of Edinburgh School of Informatics. His fields of interests include programming language modelling and design, logic, and the foundations of computer science.<\/p><\/div><\/div><br \/><div id=\"seminar-9\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">From trash to treasure: Timing-sensitive garbage collection<\/a>\n<br \/>Mathias Pedersen\nAarhus University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/cs.au.dk\/~mvp\/\">http:\/\/cs.au.dk\/~mvp\/<\/a>\n<br \/>19 September, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>We study information flows arising from timing channels in the presence of automatic memory management. We construct a series of example attacks that illustrate how garbage collectors form a shared resource that can be used to reliably leak sensitive information at a rate of up to 1 byte\/sec on a contemporary general-purpose computer. The created channel is also observable across a network connection in a datacenter-like setting. We subsequently present a design of an automatic memory management system and prove that the presented attacks are impossible on garbage collectors satisfying this design. The guarantees provided by the language has been mechanized in the Coq proof assistant.<\/p><p>Bio\n<\/p><p>Mathias is a third year PhD student in the Logic and Semantics group at Aarhus University in Denmark, advised by Aslan Askarov. His work is in the area of language-based security, with a focus on provable mitigation of side channels. In general, anything related to compilers and the semantics of programming languages will be on his list of interests.<\/p><\/div><\/div><br \/><div id=\"seminar-10\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Inferring Type Rules for Syntactic Sugar<\/a>\n<br \/>Justin Pombrio\nBrown University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/justinpombrio.net\">http:\/\/justinpombrio.net<\/a>\n<br \/>7 August, 2018 11:30AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Type systems and syntactic sugar are both valuable to programmers, but sometimes at odds. While sugar is a valuable mechanism for implementing realistic languages, the expansion process obscures program source structure. As a result, type errors can reference terms the programmers did not write (and even constructs they do not know), baffling them. The language developer must also manually construct type rules for the sugars, to give a typed account of the surface language. We address these problems by presenting a process for automatically reconstructing type rules for the surface language using rules for the core. We have implemented this theory, and show several interesting case studies.<\/p><p>Bio\n<\/p><p>Justin Pombrio is a recent PhD graduate from Brown University. His research is mainly in programming languages, with a focus on syntactic sugar, but also includes CS education.<\/p><\/div><\/div><br \/><div id=\"seminar-11\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Debugging Debug Information and Beyond<\/a>\n<br \/>Francesco Zappa Nardelli\nInria Paris \u2013 Rocquencourt\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.di.ens.fr\/~zappa\/\">http:\/\/www.di.ens.fr\/~zappa\/<\/a>\n<br \/>15 June, 2018 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The spectacular results achieved by computer science in the recent years rely on hidden, obscure, and badly specified components that lurk at the very heart of our computing infrastructure. Consider DWARF debug information. Debug information is relied upon by debuggers and plays a key role in the in the implementation of program analysis tools. More surprisingly, debug information can be relied upon by the runtime of high-level programming languages, for instance by C++ to unwind the stack and implement exceptions. The debug information itself can be pervaded by subtle bugs, making the whole infrastructure unreliable. In this talk I will describe techniques and tools to perform validation and synthesis of the DWARF stack unwinding tables. I will also report on adventurous projects that we might build on top of reliable DWARF information.<\/p><p>Bio\n<\/p><p>Francesco Zappa Nardelli is a Researcher at Inria Paris \u2013 Rocquencourt. His research interests focus on concurrent computation on top of relaxed memory models, ranging from hardware models of modern architectures to high-level programming language specification and compilation. He received his Ph.D. from Universit\u00e9 Paris 7 in 2003. Since then, he has worked on language design for distributed computation, type systems for integrating typed and untyped code in scripting languages, and tool support for semantics (including the Ott tool).<\/p><\/div><\/div><br \/><div id=\"seminar-12\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Call-By-Name Gradual Type Theory<\/a>\n<br \/>Max New\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/maxsnew.github.io\/\">http:\/\/maxsnew.github.io\/<\/a>\n<br \/>27 April, 2018 11:30AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>I will present Call-by-name (CBN) Gradual Type Theory [1], an axiomatic semantics of CBN gradual typing that directly axiomatizes the extensionality (eta) principles of types and the \"gradual guarantee\" of [2]. By axiomatizing the \"term precision\" relation of [2] we can provide a specification for casts as meets and joins with respect to precision. Using our type theory, we then show that the classic function contract ([3]) is in fact the unique implementation of function casts that satisfies extensionality and the gradual guarantee. This shows that the combination of these properties is a very strong constraint on the design of a gradually typed language and any implementation that departs must sacrifice one of these properties (typically extensionality).<\/p><p>[1] Call-by-Name Gradual Type Theory, New and Licata, to appear FSCD '18,https:\/\/arxiv.org\/abs\/1802.00061<\/p><p>[2] Refined Criteria for Gradual Typing, Siek, Vitousek, Cimini and Boyland, SNAPL '15, http:\/\/drops.dagstuhl.de\/opus\/volltexte\/2015\/5031\/<\/p><p>[3] Contracts for Higher-Order Functions, Findler and Felleisen, ICFP '02,https:\/\/dl.acm.org\/citation.cfm?id=581484<\/p><p>Bio\n<\/p><p>Max New is a PhD student at Northeastern University working on the semantic foundations of programming languages. He hates proving the same theorem twice and you should too.<\/p><\/div><\/div><br \/><div id=\"seminar-13\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Compilation of Simple Probabilistic Programs to Gibbs Sampling.<\/a>\n<br \/>John Tristan\nOracle Labs\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/jtristan.github.io\/\">https:\/\/jtristan.github.io\/<\/a>\n<br \/>10 April, 2018 11:45AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>One of the several interesting challenges of probabilistic programming is that of compiling probabilistic programs to inference algorithms. One of these inference algorithms, Gibbs sampling, is particularly relevant because it is often statistically efficient, but unfortunately, it is difficult to derive and therefore compile to.<\/p><p>In this talk, after a brief explanation of probabilistic programming and why its relevance to data science, I will explain some of the ideas behind the design of a compiler from (very) simple probabilistic programs to Gibbs sampling. I will also attempt to explain what it would mean for such a compiler to be correct.<\/p><p>Bio\n<\/p><p>Jean-Baptiste Tristan is a researcher in the machine learning group at Oracle Labs. He holds a Ph.D. in Computer Science from the French Institute for Research in Computer Science and Automation (INRIA) and a M.Sc. in Computer Science from the Ecole Normale Superieure of Paris.<\/p><\/div><\/div><br \/><div id=\"seminar-14\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Design by Introspection using the D Language<\/a>\n<br \/>Andrei Alexandrescu\nD Language Foundation\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/erdani.com\">http:\/\/erdani.com<\/a>\n<br \/>6 February, 2018 11:45AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Years ago, D started modestly as an improved offering in the realm of systems programming languages, sharing a good deal of philosophy with C and C++. With time, however, D became a very distinct language with unique features (and, of course, its own mistakes).<\/p><p>One angle of particular interest has been D's ability to perform compile-time introspection. Artifacts in a D program can be \"looked at\" during compilation. Coupled with a full-featured compile-time evaluation engine and with an ability to generate arbitrary code during compilation, this has led to a number of interesting applications.<\/p><p>This talk shares early experience with using these features of the D language. Design by Introspection is a proposed programming paradigm that assembles designs by performing introspection on components during compilation.<\/p><p>Bio\n<\/p><p>Andrei Alexandrescu is a researcher, software engineer, and author. He wrote three best-selling books on programming (Modern C++ Design, C++ Coding Standards, and The D Programming Language) and numerous articles and papers on wide-ranging topics from programming to language design to Machine Learning to Natural Language Processing. Andrei holds a PhD in Computer Science from the University of Washington and a BSc in Electrical Engineering from University \"Politehnica\" Bucharest. He currently works on the D Language Foundation. http:\/\/erdani.com<\/p><\/div><\/div><br \/><div id=\"seminar-15\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Solver-Aided Declarative Programming<\/a>\n<br \/>Molham Aref\nRelational AI\n<a class=\"pn-url\" target=\"_blank\" href=\"#\">#<\/a>\n<br \/>14 December, 2017 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>I will summarize our work on a declarative programming language that offers native language support for model (or instance) finding. This capability can be used to express predictive (e.g. machine learning) and prescriptive (e.g. combinatorial optimization) analytics. The presentation gives an overview of the platform and the language. In particular, it focuses on the important role of integrity constraints,which are used not only for maintaining data integrity, but also, for the formal specification of complex optimization problems and probabilistic programming.<\/p><p>Bio\n<\/p><p>Mr. Molham Aref is the Chief Executive Officer of Relational AI. Mr. Aref has more than 25 years of experience in developing and implementing enterprise-grade analytic, predictive, optimization and simulation solutions for the demand chain, supply chain, and revenue management across various industries. Relational AI combines the latest advances in Artificial Intelligence with a understanding of business processes to develop solutions that shape better decisions, improve agility, and reduce risk. Prior to Relational AI, he was co-founder and CEO of LogicBlox where he led the company from inception through a successful sale to Infor. Previously, he was CEO of Optimi (acquired by Ericsson),a leader in wireless network simulation and optimization, and co-founder of Brickstream (renamed Nomi and then acquired by FLIR), a leading provider of computer-vision-based behavior intelligence solutions.<\/p><\/div><\/div><br \/><div id=\"seminar-16\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Compositional Creativity<\/a>\n<br \/>Chris Martens\nNorth Carolina State University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/sites.google.com\/ncsu.edu\/cmartens\">https:\/\/sites.google.com\/ncsu.edu\/cmartens<\/a>\n<br \/>11 December, 2017 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Creativity is seen as a core cognitive ability for tasks like storytelling, conversation, and aesthetic design. Simulating creative processes with computers enables them to amplify humans' natural creativity and to support applications like interactive narrative, chatbots, and procedural level generators for games. By now, a large number of techniques exist for simulating creativity, such as search, planning, production grammars, logic programming, stochastic modeling from data, and hybrid approaches. However, even the simplest methods are commonly considered programming tasks for \"AI experts.\" High-level programming languages have the potential to broaden participation in authorship of these systems, but only if they are designed with compositionality -- the mathematical principle that's also functional programming's secret weapon -- in mind. This talk will discuss what compositionality has achieved so far for creative computing and what we can do with it next.<\/p><p>Bio\n<\/p><p>Chris Martens is an assistant professor in the Computer Science (CSC) Department at NC State, where she is affiliated with the Digital Games research group. Her interests span artificial intelligence, programming languages, and formal methods.<\/p><\/div><\/div><br \/><div id=\"seminar-17\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Efficient Nominal Gradual Typing<\/a>\n<br \/>Fabian Muehlboeck\nCornell University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.cs.cornell.edu\/~fabianm\/\">http:\/\/www.cs.cornell.edu\/~fabianm\/<\/a>\n<br \/>5 December, 2017 2:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Gradual Typing is the idea that we can support both static and dynamic type checking in different parts of the same program in the same language. Most existing work on gradual typing aims to add gradual typing to an existing language. However, this severely constrains some key decisions in creating gradually typed languages, often leading designers to trade off either soundness or efficiency.<\/p><p>I will talk about designing a nominal object-oriented language with gradual typing as a core component from the start. This affects key design decisions for the other type system features, as well as the design of the runtime. Experiments with a prototype language I implemented suggest that this is a viable approach to achieve sound, yet efficient gradual typing. I will present those results and talk about some key design challenges that have yet to be solved.<\/p><p>Bio\n<\/p><p>Fabian Muehlboeck is a Ph.D. student working with Ross Tate at Cornell University.<\/p><\/div><\/div><br \/><div id=\"seminar-18\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">What does CPS have to do with deep learning?<\/a>\n<br \/>Jeffrey M. Siskind\nPurdue University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/engineering.purdue.edu\/~qobi\/\">https:\/\/engineering.purdue.edu\/~qobi\/<\/a>\n<br \/>11 September, 2017 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Deep learning is formulated around backpropagation, a method for computing gradients of a particular class of data flow graphs to train models to minimize error by gradient descent.  Backpropagation is a special case of a more general method known as reverse mode automatic differentiation (AD) for computing gradients of functions expressed as programs.  Reverse mode AD imposes only a small constant-factor overhead in operation count over the original computation, but has storage requirements that grow, in the worst case, in proportion to the time consumed by the original computation.  This storage blowup can be ameliorated by check pointing, a process that reorders application of classical reverse-mode AD over an execution interval to tradeoff space vs. time.  Application of check pointing in a divide-and-conquer fashion to strategically chosen nested execution intervals can break classical reverse-mode AD into stages which can reduce the worst-case growth in storage from linear to logarithmic.  Doing this has been fully automated only for computations of particularly simple form, with checkpoints spanning execution intervals resulting from a limited set of program constructs.  We show how the technique can be automated for arbitrary computations.  Doing so relies on implementing general purpose mechanisms for counting the number of instructions executed by a program, interrupting the execution after a specified number of steps, and resuming the execution with a nonstandard interpretation.  We implement these general purpose mechanisms with a compiler that converts programs to continuation-passing style (CPS).  The process of efficiently computing gradients with check pointing requires running, and rerunning, little bits of the program out of order.  This is made easier by applying the technique to functional programs. There is a deeper and higher-level message in this talk: machine learning can benefit from advanced techniques from programming language theory.<\/p><p>Bio\n<\/p><p>Jeffrey M. Siskind received the B.A. degree in Computer Science from the Technion, Israel Institute of Technology, Haifa, in 1979, the S.M. degree in Computer Science from the Massachusetts Institute of Technology (M.I.T.), Cambridge, in 1989, and the Ph.D. degree in Computer Science from M.I.T. in 1992.  He did a postdoctoral fellowship at the University of Pennsylvania Institute for Research in Cognitive Science from 1992 to 1993.  He was an assistant professor at the University of Toronto Department of Computer Science from 1993 to 1995, a senior lecturer at the Technion Department of Electrical Engineering in 1996, a visiting assistant professor at the University of Vermont Department of Computer Science and Electrical Engineering from 1996 to 1997, and a research scientist at NEC Research Institute, Inc. from 1997 to 2001.  He joined the Purdue University School of Electrical and Computer Engineering in 2002 where he is currently an associate professor.  His research interests include computer vision, robotics, artificial intelligence, neuroscience, cognitive science, computational linguistics, child language acquisition, automatic differentiation, and programming languages and compilers.<\/p><\/div><\/div><br \/><div id=\"seminar-19\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Relational Parametricity for Polymorphic Blame Calculus<\/a>\n<br \/>Dustin Jamner\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/github.com\/DIJamner\">https:\/\/github.com\/DIJamner<\/a>\n<br \/>23 June, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The polymorphic blame calculus integrates static typing, including universal types, with dynamic typing. The primary challenge with this integration is preserving parametricity: even dynamically-typed code should satisfy it once it has been cast to a universal type. Ahmed et al. (2011) employ runtime type generation in the polymorphic blame calculus to preserve parametricity, but a proof that it does so has been elusive. Matthews and Ahmed (2008) gave a proof of parametricity for a closely related system that combines ML and Scheme, but later found a flaw in their proof. In this work we prove that the polymorphic blame calculus satisfies relational parametricity. The proof relies on a step-indexed Kripke logical relation. The step-indexing is required to make the logical relation well-defined in the case for the dynamic type. The possible worlds include the mapping of generated type names to their concrete types and the mapping of type names to relations. We prove the Fundamental Property of this logical relation and that it is sound with respect to contextual equivalence.<\/p><p>Bio\n<\/p><p>Dustin Jamner is an undergraduate at Northeastern University in Computer Science, working with Amal Ahmed.<\/p><\/div><\/div><br \/><div id=\"seminar-20\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">CoCoSpec: A Mode-aware Contract Language for Reactive Systems<\/a>\n<br \/>Cesare Tinelli\nUniversity of Iowa\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/homepage.cs.uiowa.edu\/~tinelli\/\">http:\/\/homepage.cs.uiowa.edu\/~tinelli\/<\/a>\n<br \/>1 June, 2017 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Contract-based software development is a leading methodology for the construction of safety- and mission-critical embedded systems. Contracts are an effective way to establish boundaries between components and can be used efficiently to verify global properties by using compositional reasoning techniques. A contract specifies the assumptions a component makes on its context and the guarantees it provides. Requirements in the specification of a component are often case-based, with each case describing what the component should do depending on a specific situation (or mode) the component is in.<\/p><p>This talk introduces CoCoSpec, a mode-aware assume-guarantee-based contract language for embedded systems. CoCoSpec is built as an extension of the Lustre language and lets users specify mode behavior directly, instead of encoding it as conditional guarantees, thus preventing a loss of mode-specific information. Mode-aware model checkers supporting CoCoSpec can increase the effectiveness of the compositional analysis techniques found in assume-guarantee frameworks and improve scalability. Such tools can also produce much better feedback during the verification process, as well as valuable qualitative information on the contract itself. I will presents the CoCoSpec language and illustrate the benefits of mode-aware model-checking on a case study involving a flight-critical avionics system. The evaluation uses Kind 2, a collaborative, parallel, SMT-based model checker developed at the University of Iowa that provides full support for CoCoSpec.<\/p><p>Bio\n<\/p><p>Cesare Tinelli received a Ph.D. in Computer Science from the UIUC in 1999 and is currently a professor in Computer Science at the University of Iowa. His research interests include automated reasoning and formal methods. He has done seminal work in automated reasoning, in particular in Satisfiability Modulo Theories (SMT), a field he helped establish through his research and service activities. His work has been funded both by governmental agencies (AFOSR, AFRL, DARPA, NASA, and NSF) and corporations (Intel, General Electric, Rockwell Collins, and United Technologies) and has appeared in more than 70 refereed publications. He is a founder and coordinator of the SMT-LIB initiative, an international effort aimed at standardizing benchmarks and I\/O formats for SMT solvers. He has led the development of the award winning Darwin theorem prover and the Kind model checker. He co-leads the development of CVC4,a widely used and award winning SMT solver, and StarExec, a cross community web-based service for the comparative evaluation of logic solvers.<\/p><\/div><\/div><br \/><div id=\"seminar-21\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Multi-language programming system: a linear experiment<\/a>\n<br \/>5 May, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Programming language research keeps inventing more powerful dynamic and static features for programming language. Ideally, thorough language design should be able to propose a single, unified, powerful yet simple view of them as orthogonal aspects of a single programming language that would remain usable by everyone. In practice, it always takes much more time and effort to find the simple\/orthogonal presentation than to come up with the desirable features, and we become emotionally attached to programming language monoliths (C++, Scala, GHC Haskell, OCaml, Common Lisp...) that suffer from feature creep. An alternative is to design programming systems as multi-language environment instead. This point of view makes it easier to detect flaws in a proposed programming system design: there is no clear objective definition of what \"orthogonal features\" really mean, but we have clear ideas of how multi-language systems should work. The Racket Manifesto says: \"Each language and component must be able to protect its specific invariants\".<\/p><p>In this work we give a formal specification, in terms of full abstraction, for what it means for a new language added to the system to not introduce \"abstraction leaks\" in the other languages, and discuss scenarios where this translates in (informal) usability properties: having a nice teachable language subset, or gracefully introducing an advanced language for experts. We present a multi-language design that exhibits this formal property by adding a language L with linear types and linear state to a typical ML-style programming language U. L allows to write programs with resource usage guarantees (space consumption, typestate-style safety guarantees) that were not possible in U, without introducing abstraction leaks. We will demonstrate several useful examples, some of them that crucially rely on the fine-grained use of language boundaries, at a subterm level rather than whole-module level.<\/p><p>Bio\n<\/p><\/div><\/div><br \/><div id=\"seminar-22\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">From Datalog to Flix: A Declarative Language for Fixed Points on Lattices<\/a>\n<br \/>Magnus Madsen\nUniversity of Waterloo\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/prl.ccs.neu.edu\/plg.uwaterloo.ca\/~mmadsen\/\">plg.uwaterloo.ca\/~mmadsen\/<\/a>\n<br \/>14 April, 2017 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>We present FLIX, a declarative programming language for specifying and solving least fixed point problems, particularly static program analyses. FLIX is inspired by Datalog and extends it with lattices and monotone functions. Using FLIX,implementors of static analyses can express a broader range of analyses than is currently possible in pure Datalog, while retaining its familiar rule-based syntax. We define a model-theoretic semantics of FLIX as a natural extension of the Datalog semantics. This semantics captures the declarative meaning of FLIX programs without imposing any specific evaluation strategy. An efficient strategy is semi-na\u00efve evaluation which we adapt for FLIX. We have implemented a compiler and runtime for FLIX, and used it to express several well-known static analyses, including the IFDS and IDE algorithms. The declarative nature of FLIX clearly exposes the similarity between these two algorithms.<\/p><p>This work has previously been presented at PLDI 2016. The talk will cover some of that material as well as recent developments.<\/p><p>Bio\n<\/p><p>The speaker is a post doctoral fellow at the University of Waterloo.<\/p><\/div><\/div><br \/><div id=\"seminar-23\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Simple Invariants for proving the safety of  distributed protocols and networks<\/a>\n<br \/>Mooly Sagiv\nTel Aviv University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.cs.tau.ac.il\/~msagiv\/\">http:\/\/www.cs.tau.ac.il\/~msagiv\/<\/a>\n<br \/>31 March, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Safety of a distributed protocol means that the protocol never reaches a bad state, e.g., a state where two nodes become leaders in a leader-election protocol. Proving safety is obviously undecidable since such protocols are run by an unbounded number of nodes, and their safety needs to be established for any number of nodes. I will describe a deductive approach for proving safety, based on the concept of universally quantified inductive invariants --- an adaptation of the mathematical concept of induction to the domain of programs. In the deductive approach, the programmer specifies a candidate inductive invariant and the system automatically checks if it is inductive. By restricting the invariants to be universally quantified, this approach can be effectively implemented with a SAT solver.<\/p><p>This is a joint work with Ken McMillan (MSR), Oded Padon (TAU), Aurojit Panda(Berkeley) , and Sharon Shoham(TAU) and was integrated into the IVY system: http:\/\/www.cs.tau.ac.il\/~odedp\/ivy\/. The work is inspired by Shachar Itzhaky's thesis available from http:\/\/people.csail.mit.edu\/shachari\/<\/p><p>Bio\n<\/p><p>Mooly Sagiv is a professor in the School of Computer Sciences at Tel-Aviv University. He is a leading researcher in the area of large scale (inter-procedural) program analysis, and one of the key contributors to shape analysis. His fields of interests include programming languages, compilers, abstract interpretation, profiling, pointer analysis, shape analysis, inter-procedural dataflow analysis, program slicing, and language-based programming environments. Sagiv is a recipient of a 2013 senior ERC research grant for Verifying and Synthesizing Software Composition. Prof. Sagiv served as Member of the Advisory Board of Panaya Inc acquired by Infosys. He received best-paper awards at PLDI'11 and PLDI'12  for his work on composing concurrent data structures and a ACM SIGSOFT Retrospective Impact Paper Award (2011) for program slicing. He is an ACM fellow and a recipient of Microsoft Research Outstanding Collaborator Award 2016.<\/p><\/div><\/div><br \/><div id=\"seminar-24\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Learning to Program and Debug, Automatically<\/a>\n<br \/>Swarat Chaudhuri\nRice University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/www.cs.rice.edu\/~sc40\/\">https:\/\/www.cs.rice.edu\/~sc40\/<\/a>\n<br \/>24 March, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Automating programming and debugging are long-standing goals in computer science. In spite of significant progress in formal methods over the years, we remain very far from achieving these goals. For example, a freshman CS major will typically program circles around today's best program synthesizers. Debugging and verification tools rely on formal specifications, which are hard to provide in many important applications.<\/p><p>Two critical components of the gap between human and machine programmers are that humans learn from experience, i.e., data, and can easily generalize from incomplete problem definitions. In this talk, I will present a general framework for formal methods, based on Bayesian statistical learning, that aims to eliminate these differences. In our framework, descriptions of programming tasks are seen to be \"clues\" towards a hidden (probabilistic) specification that fully defines the task. Large corpora of real-world programs are used to construct a statistical model that correlates specifications with the form and function of their implementations. The framework can be implemented in a variety of ways, but in particular, through a neural architecture called Bayesian variational encoder-decoders. Inferences made using the framework can be used to guide traditional algorithms for program synthesis and bug-finding.<\/p><p>I will show that this data-driven approach can lead to giant leaps in the scope and performance of automated program synthesis and debugging algorithms.  Specifically, I will give a demo of Bayou, a system for Bayesian inductive synthesis of Java programs that goes significantly beyond the state of the art in program synthesis. I will also describe Salento, a debugging system based on our framework that can find subtle violations of API contracts without any kind of specification.<\/p><p>Bio\n<\/p><p>Swarat Chaudhuri is an Associate Professor of Computer Science at Rice University. His research lies at the interface of programming systems and artificial intelligence. Much of his recent work is on program synthesis, the problem of automatically generating computer programs from high-level specifications.<\/p><\/div><\/div><br \/><div id=\"seminar-25\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Templates and Types in C++<\/a>\n<br \/>Jimmy Hartzell\nTower Research\n<br \/>10 March, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Like many modern programming languages, C++ supports both generic\n       programming (templates) and object-oriented programming (inheritance and\n  virtual functions). Unlike many modern programming languages, the generic\n programming mechanism does not fully leverage or interact with the type\n  system. It amounts to compile-time duck typing: C++ function templates,\n  like functions from a non-statically typed programming language, have\n  their full interface implicit in their implementations -- making it hard\n  to discern what that interface actually is.  The concept of Concepts\n  (compile-time typeclasses) was slated for introduction into C++ in\n the C++11 version of the standard, then C++14, then C++17, now C++20... What can\nwe do to work around their absence?<\/p><p>Relatedly, how do we bridge the gap\nbetween static and dynamic polymorphism? C++ provides both, and sometimes we\nwant to mix and mash them. Template tricks can also be done to accomplish this\nas well.<\/p><p>Bio\n<\/p><p>Jimmy Hartzell is a software developer and C++ instructor at Tower Research.<\/p><\/div><\/div><br \/><div id=\"seminar-26\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Overlapping pattern matching for programming with continuous functions<\/a>\n<br \/>Ben Sherman\nMIT\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.ben-sherman.net\/\">http:\/\/www.ben-sherman.net\/<\/a>\n<br \/>24 February, 2017 12:00PM \nWVH 366\n| <a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/github.com\/nuprl\/website\/blob\/master\/seminar-notes\/sherman-patterns.md\">Notes<\/a><\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Topology, when viewed from an unusual perspective (formal topology), describes how to compute with certain objects that are beyond the reach of induction, such as real numbers, probability distributions, streams, and function spaces. Accordingly, one gets a programming language whose types are spaces and whose functions are continuous maps.<\/p><p>Does this language have pattern matching? In functional programming, pattern matching allows definition of a function by partitioning the input and defining the result in each case. We generalize to programming with spaces, where patterns need not represent decidable predicates and also may overlap, potentially allowing nondeterministic behavior in overlapping regions. These overlapping pattern matches are useful for writing a wide array of computer programs on spaces, such as programs that make approximate computations or decisions based on continuous values or that manipulate \"partial\" datatypes.<\/p><p>This talk will introduce topology from a computational perspective, and explore some programs that can be built with this framework using overlapping pattern matching.<\/p><p>Bio\n<\/p><p>Ben Sherman is a second-year PhD student at MIT, advised by Adam Chlipala and Michael Carbin.<\/p><\/div><\/div><br \/><div id=\"seminar-27\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Comparative Study of Generic Programming Features in Object-Oriented Languages<\/a>\n<br \/>Julia Belyakova\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/staff.mmcs.sfedu.ru\/~juliet\/index.en.html\">http:\/\/staff.mmcs.sfedu.ru\/~juliet\/index.en.html<\/a>\n<br \/>3 February, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Earlier comparative studies of language support for generic programming (GP) have shown that mainstream object-oriented (OO) languages such as C# and Java provide weaker support for GP as compared with functional languages such as Haskell or SML.<\/p><p>One reason is that generic mechanisms of C# and Java are based on F-bounded polymorphism, which has certain deficiencies. Some of these deficiencies are eliminated in more recent languages, such as Scala and Kotlin. However, there are peculiarities of object-oriented languages as a whole, which prevent them from being as expressible as Haskell in terms of generic programming.<\/p><p>In this talk we will cover the main problems of language support for generic programming in C#\/Java as well as other modern object-oriented languages, including Scala and Swift. It turns out that all of these languages follow the same approach to constraining type parameters, which leads to inevitable shortcomings. To overcome them, an alternative approach is suggested in several extensions for OO languages, in particular, JavaGenus. The advantages and drawbacks of both approaches will be discussed in the talk.<\/p><p>Here are Julia's <a href=\"http:\/\/staff.mmcs.sfedu.ru\/~juliet\/files\/materials\/seminar-talks\/belyakova-PL-sem-2017-slides.pdf\">Slides<\/a>.<\/p><p>Bio\n<\/p><p>Julia Belyakova graduated in 2014 from Southern Federal University (Rostov-on-Don, Russia) as Master of Science in \u201cComputer Science and Information Technologies\u201d. She is enrolled for PhD program in Southern Federal University and came to Northeastern University as a visiting scientist. She is currently doing research in the area of generic programming for object-oriented languages.<\/p><\/div><\/div><br \/><div id=\"seminar-28\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Making the Browser Reasonable for Sane Programmers<\/a>\n<br \/>John Vilk\nUniversity of Massachusetts Amherst\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/jvilk.com\/\">https:\/\/jvilk.com\/<\/a>\n<br \/>27 January, 2017 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Billions of people use browsers to access the web from a variety of devices, but it remains difficult to write and debug the client-side of web applications. The browser exposes unfamiliar, redundant, and incompatible abstractions for a variety of common tasks, preventing developers from adapting existing code written for other platforms. The browser environment is completely event-driven and pervasively concurrent, resulting in complex control flow and race conditions that are difficult to debug.<\/p><p>In this talk, I will describe how my research makes it easier to write and debug web applications. Doppio is a JavaScript runtime system that lets developers run unaltered code written in general-purpose languages, such as Java and C\/C++, directly inside the browser. Browsix further enhances Doppio with a kernel, processes, and shared operating system resources, letting multiple off-the-shelf programs concurrently interoperate on the same webpage. ReJS is a low-overhead and high-fidelity time-traveling debugger that lets developers instantaneously step forward and backward through a program's execution. ReJS accurately recreates the application's complete behavior, including the GUI and multiple types of race conditions, while producing miniscule and portable application traces. These resources transform the browser into a first-class application platform.<\/p><p>Bio\n<\/p><p>John Vilk is a Ph.D. candidate in the College of Information and Computer Sciences at the University of Massachusetts Amherst. He is a member of the PLASMA lab, and is advised by Emery Berger. John's research aims to improve the browser as an application platform by making it easier to write, debug, and optimize complex web applications. John is a Facebook Fellow (2015), and his work on Doppio: Breaking the Browser Language Barrier is a SIGPLAN Research Highlight (2014) and a PLDI Distinguished Artifact (2014). His research forms the basis of Microsoft ChakraCore's time-traveling debugger and lets millions of people interact with legacy software and games in their browser at the Internet Archive (https:\/\/archive.org\/). John received his MS from the University of Massachusetts Amherst (2013) and his BS from Worcester Polytechnic Institute (2011).<\/p><p>You can learn more about John at his website, https:\/\/jvilk.com\/.<\/p><\/div><\/div><br \/><div id=\"seminar-29\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Retractions and Blame<\/a>\n<br \/>Max New\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/maxsnew.github.io\/\">http:\/\/maxsnew.github.io\/<\/a>\n<br \/>15 December, 2016 12:00PM \nBehrakis 105\n| <a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/github.com\/nuprl\/website\/blob\/master\/seminar-notes\/new-retractionsblame.md\">Notes<\/a><\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>It has long been noted that research on contracts and gradual typing bears a striking resemblance to Dana Scott's work on retracts in domain theory. However, a concrete relationship has been elusive, especially since the notion of *blame* has seemingly no counterpart in domain theory.<\/p><p>We connect these fields by means of gradual type precision, which has been characterized previously using blame in [1]. We propose a definition in terms of section-retraction pairs, allowing us to relate blame and retractions in a precise way. Our definition agrees with [1] on most rules, but differs in a way that suggests a modification to their definition of blame.<\/p><p>[1] Wadler and Findler, Well-typed Programs can't Be Blamed:\nhttp:\/\/homepages.inf.ed.ac.uk\/wadler\/topics\/blame.html#blame-esop<\/p><p>Bio\n<\/p><p>Max New is a PhD student at Northeastern University working on the semantic foundations of programming languages. He hates proving the same theorem twice and you should too.<\/p><\/div><\/div><br \/><div id=\"seminar-30\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Deciding program equivalence with sums and the empty type<\/a>\n<br \/>Gabriel Scherer\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.ccs.neu.edu\/home\/gasche\/\">http:\/\/www.ccs.neu.edu\/home\/gasche\/<\/a>\n<br \/>7 December, 2016 12:00PM \nWVH 110<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The simply-typed \u03bb-calculus, with only function types, is strongly normalizing, and its program equivalence relation is decidable: unlike in more advanced type system with polymorphism or effects, the natural \"syntactic\" equivalence (\u03b2\u03b7-equivalence) corresponds to natural \"semantic\" equivalence (observational or contextual equivalence), and is decidable. Adding product types (pairs) is easy and preserves these properties, but sums (disjoint unions) are, surprisingly, harder. It is only in 1995 that Neil Ghani proved that the equivalence in presence of sums is decidable, and the situation in presence of the empty type (zero-ary sum) was unknown.<\/p><p>We propose an equivalence algorithm for sums and the empty type that takes inspiration from a proof-theoretic technique, named \"focusing\",designed for proof search. The exposition will be introductory; I will present the difficulties caused by sums and the empty type, present some of the existing approaches for sums in the literature, introduce focusing, and give a high-level intuition of our saturation-based algorithm and its termination argument. No previous knowledge of focusing is assumed. I expect some familiarity with typed lambda-calculi, but will recall all concepts used (for example, \u03b2\u03b7-equivalence) during the talk.<\/p><p>Bio\n<\/p><p>Gabriel is interested in theoretical aspects of type systems,\nprogramming language implementation, general programming language\nconcepts, and even some syntactic aspects. He has a preference for the\nformalizable aspects, or formalizable approaches to programming\nlanguage aspects, rather than the often subjective appeal to taste or\nintuition.<\/p><\/div><\/div><br \/><div id=\"seminar-31\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Analysis of Android hybrid applications and other fun with WALA<\/a>\n<br \/>Julian Dolby\nIBM Research\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/researcher.ibm.com\/person\/us-dolby\">http:\/\/researcher.ibm.com\/person\/us-dolby<\/a>\n<br \/>1 December, 2016 12:00PM \nRichards Hall 140<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Hybrid apps help developers build multiple apps for different platforms with less duplicated effort, by providing platform-specific functionality via native code and user interactions via javascript code. however, most hybrid apps are developed in multiple programming languages with different semantics, complicating programming. moreover, untrusted javascript code may access device-specific features via native code, exposing hybrid apps to attacks. Unfortunately, there are no existing tools to detect such vulnerabilities. In this paper, we present HybriDroid, the first static analysis framework for Android hybrid apps. First, we investigate the semantics of interoperation of Android Java and JavaScript. Then, we design and implement a static analysis framework that analyzes inter-communication between Android Java and JavaScript. We demonstrate HybriDroid with a bug detector that identifies programmer errors due to the hybrid semantics, and a taint analyzer that finds information leaks cross language boundaries. Our empirical evaluation shows that the tools are practically usable in that they found previously uncovered bugs in real-world Android hybrid apps and possible information leaks via a widely-used advertising platform.<\/p><p>The bulk of this presentation will focus on ASE 2016 work on analysis of hybrid apps (1), a blend of per-platform native code and portable JavaScript. I will also briefly discuss two other recent projects involving WALA: ASE 2015 work on a practically tunable static analysis framework for large-scale JavaScript applications (2), and ISSTA 2015 work on scalable and precise taint analysis for Android (3).<\/p><p>1: Sungho Lee, Julian Dolby, Sukyoung Ryu: HybriDroid: static analysis framework for Android hybrid applications. ASE 2016: 250-261<\/p><p>2: Yoonseok Ko, Hongki Lee, Julian Dolby, Sukyoung Ryu: Practically Tunable Static Analysis Framework for Large-Scale JavaScript Applications (T). ASE 2015: 541-551<\/p><p>3: Wei Huang, Yao Dong, Ana Milanova, Julian Dolby: Scalable and precise taint analysis for Android. ISSTA 2015: 106-117<\/p><p>Bio\n<\/p><p>Julian Dolby is a Research Staff Member at the IBM Thomas J. Watson Research Center, where he works on a range of topics, including static program analysis, software testing, concurrent programming models and the semantic Web. He is one of the primary authors of the publically available Watson Libraries for Analysis (WALA) program analysis infrastructure, and his recent WALA work has focused on creating the WALA Mobile infrastructure.<\/p><p>His program analysis work has recently focused on scripting languages like JavaScript and on security analysis of Web applications; this work has been included in IBM products, most notably Rational AppScan, Standard Edition and Source Edition. He was educated at the University of Illinois at Urbana-Champaign as a graduate student where he worked with Professor Andrew Chien on programming systems for massively-parallel machines.<\/p><\/div><\/div><br \/><div id=\"seminar-32\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Sequent calculus as a programming language<\/a>\n<br \/>Zena Ariola\nUniversity of Oregon\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/ix.cs.uoregon.edu\/~ariola\/\">http:\/\/ix.cs.uoregon.edu\/~ariola\/<\/a>\n<br \/>28 November, 2016 12:00PM \nWVH 110<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>We will present and demonstrate the usefulness of the sequent\ncalculus as a formal model of computation based on interactions\nbetween producers and consumers of results. This model leads\nto a better understanding of call-by-name evaluation by\nreconciling the conflicting principles of extensionality and\nweak-head evaluation, thus internalizing a known parametricity\nresult. It allows one to explore two dualities of computation:\nthe duality between call-by-name and call-by-value, and the\nduality between construction and deconstruction. This ultimately\nleads to a better linguistic foundation for co-induction as dual\nto induction. From a more practical point of view, the sequent\ncalculus provides a useful inspiration for the design of\nintermediate languages.<\/p><p>Bio\n<\/p><p>TBD<\/p><\/div><\/div><br \/><div id=\"seminar-33\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Contextual Equivalence of Probabilistic Programs<\/a>\n<br \/>Ryan Culpepper\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.ccs.neu.edu\/home\/ryanc\/\">http:\/\/www.ccs.neu.edu\/home\/ryanc\/<\/a>\n<br \/>3 November, 2016 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>In this talk I introduce a probabilistic programming language with continuous random variables and soft constraints, define contextual equivalence for the language, and then present a logical relation that is sound with respect to contextual equivalence.<\/p><p>The question of equivalence for probabilistic programs is interesting and difficult because the meaning of a program is a measure defined by integrating over the space of all of the program's possible evaluations. Thus creating a workable logical relation requires proving integral equalities involving intermediate measures that are ``equal enough'' but still respect contextual equivalence.<\/p><p>Bio\n<\/p><p>Ryan Culpepper is a research scientist at Northeastern University.  He works on probabilistic programming as well as extensible languages and tools for building domain-specific languages. He received his PhD at Northeastern and did a postdoc at the University of Utah.<\/p><\/div><\/div><br \/><div id=\"seminar-34\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Declarative, Convergent Edge Computation<\/a>\n<br \/>Christopher Meiklejohn\nUniversit\u00e9 Catholique\nde Louvain\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/christophermeiklejohn.com\/\">https:\/\/christophermeiklejohn.com\/<\/a>\n<br \/>27 October, 2016 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Consistency is hard and coordination is expensive. As we move into the\nworld of connected 'Internet of Things' style applications, or\nlarge-scale mobile applications, devices have less power, periods of\nlimited connectivity, and operate over unreliable asynchronous\nnetworks. This poses a problem with shared state: how do we handle\nconcurrent operations over shared state, while clients are offline,\nand ensure that values converge to a desirable result without making\nthe system unavailable?<\/p><p>We look at a new programming model, called Lasp. This programming\nmodel combines distributed convergent data structures with a dataflow\nexecution model designed for distribution over large-scale\napplications. This model supports arbitrary placement of processing\nnode: this enables the user to author applications that can be\ndistributed across data centers and pushed to the edge.<\/p><p>Bio\n<\/p><p>Christopher Meiklejohn loves distributed systems and programming\nlanguages. Previously, Christopher worked at Basho Technologies, Inc.\non the distributed key-value store, Riak. Christopher develops a\nprogramming model for distributed computation, called Lasp.\nChristopher is currently a Ph.D. student at the Universit\u00e9 Catholique\nde Louvain in Belgium.<\/p><\/div><\/div><br \/><div id=\"seminar-35\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">The Hack Experiment &amp; HHVM, Then And Now<\/a>\n<br \/>Julien Verlaguet &amp; Brett Simmers\nFacebook\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/hacklang.org\/\">http:\/\/hacklang.org\/<\/a>\n<br \/>25 October, 2016 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The Hack programming language developed at Facebook is an evolution of PHP. With tens of millions  of lines of PHP code in house, the lack of type and poor IDE support were felt to be major threats to Facebook. We designed Hack gradually, adding feature at regular intervals and encouraging developers to adopt them in their code. This talk will touch on the design of the Hack language, its type system, the implementation of the type checker, and the social processes involved in converting 20 millions lines of untyped code to a rich type system.<\/p><p>HHVM is an open source virtual machine for PHP and Hack programs. It was developed by Facebook but is also used by Wikipedia, Baidu, Etsy, and many others. This talk will discuss the history of the project, what goes on behind the scenes while executing your PHP code, and what we're currently working on.<\/p><p>Bio\n<\/p><p>Julien Verlaguet is a OCaml programmer with a Master from Universit\u00e9 Pierre et Marie Curie in Paris. Before joining Facebook in 2011, he worked at Esterel technologies on verified compilation.  At Facebook he designed the Hack programming language, a typed offspring of PHP and managed to convince the majority of developers in the company to add types to their code.<\/p><p>Brett Simmers is a Software Engineer at Facebook, where he's been working on HHVM for the past five years. He primarily works on the internals of the JIT compiler and HHVM's profile-guided optimizations. Before joining Facebook he spent two years at VMware working on the Linux version of VMware Workstation.<\/p><\/div><\/div><br \/><div id=\"seminar-36\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Implementing a Functional Language for Flix<\/a>\n<br \/>Ming-Ho Yee\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/mhyee.com\">http:\/\/mhyee.com<\/a>\n<br \/>20 October, 2016 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Static program analysis is a powerful technique for maintaining software, with applications such as compiler optimizations, code refactoring, and bug finding. Static analyzers are typically implemented in general-purpose programming languages, such as C++ and Java; however, these analyzers are complex and often difficult to understand and maintain. An alternate approach is to use Datalog, a declarative language. Implementors can express analysis constraints declaratively, which makes it easier to understand and ensure correctness of the analysis. Furthermore, the declarative nature of the analysis allows multiple, independent analyses to be easily combined.<\/p><p>Flix is a programming language for static analysis, consisting of a logic language and a functional language. The logic language is inspired by Datalog, but supports user-defined lattices. The functional language allows implementors to write functions, something which is not supported in Datalog. These two extensions, user-defined lattices and functions, allow Flix to support analyses that cannot be expressed by Datalog, such as a constant propagation analysis. Datalog is limited to constraints on relations, and although it can simulate finite lattices, it cannot express lattices over an infinite domain. Finally, another advantage of Flix is that it supports interoperability with existing tools written in general-purpose programming languages.<\/p><p>This talk provides an overview of the implementation and evaluation of the Flix functional language. The implementation involves abstract syntax tree transformations, an interpreter back-end, and a code generator back-end, and must support a number of interesting language features, such as pattern matching, first-class functions, and interoperability. The evaluation compares the interpreter and code generator back-ends in terms of correctness and performance.<\/p><p>Bio\n<\/p><p>Ming-Ho Yee is a Ph.D. student at Northeastern University. He works on programming language design and implementation with Jan Vitek. He received a Bachelor of Software Engineering from the University of Waterloo, and then continued for a Master of Mathematics in Computer Science under the supervision of Ond\u0159ej Lhot\u00e1k.<\/p><\/div><\/div><br \/><div id=\"seminar-37\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Behavioral Non-Portability in Decision-Making Programs<\/a>\n<br \/>Thomas Wahl\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.ccs.neu.edu\/home\/wahl\/\">http:\/\/www.ccs.neu.edu\/home\/wahl\/<\/a>\n<br \/>13 October, 2016 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>The precise semantics of floating-point arithmetic programs depends on the execution platform, including the compiler and the target hardware. Such platform dependencies infringe on the highly desirable goal of software portability (which is in fact promised by heterogeneous computing frameworks like OpenCL): the same program run on the same inputs on different platforms can produce different results. In other words, portability does not guarantee reproducibility, and this is a more or less accepted state of affairs.<\/p><p>Serious doubts on the portability of numeric applications arise when differences in results are behavioral, i.e. when they lead to changes in the control flow of a program. In this talk I will first present an algorithm that takes a numeric procedure and determines whether a given input can lead to different decisions depending merely on how the arithmetic in the procedure is compiled and executed. I will then show how this algorithm can be used in static and dynamic analyses of programs, to estimate their numeric stability. I will illustrate the results on examples characteristic of numeric computing where control flow divergence actually occurs across different execution platforms.<\/p><p>Joint with Yijia Gu, Mahsa Bayati, and Miriam Leeser, Northeastern University, Boston, USA<\/p><p>Bio\n<\/p><p>Thomas Wahl joined the faculty of Northeastern University in 2011. His research concerns the reliability (whatever that means) of complex computing systems. Two domains notorious for their fragility are concurrency and numerical computing. With colleagues, Wahl has investigated how floating-point arithmetic can \"hijack\" a program's computation when run on non-standard architectures, such as heterogeneous and custom-made embedded platforms. You will witness some hijacking attempts in the talk today.<\/p><\/div><\/div><br \/><div id=\"seminar-38\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Concolic Testing: A Decade Later<\/a>\n<br \/>Koushik Sen\nUniversity of California, Berkeley\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/www.cs.berkeley.edu\/~ksen\/\">https:\/\/www.cs.berkeley.edu\/~ksen\/<\/a>\n<br \/>6 October, 2016 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Symbolic execution, which was introduced more than four decades ago, is typically used in software testing to explore as many different program paths as possible in a given amount of time, and for each path to generate a set of concrete input values exercising it, and check for the presence of various kinds of errors including assertion violations, uncaught exceptions, security vulnerabilities, and memory corruption.  A key limitation of classical symbolic execution is that it cannot generate useful test inputs if the program under test uses complex operations such as pointer manipulations and non-linear arithmetic operations.<\/p><p>Our research on Concolic Testing (also known as DART: Directed Automated Random Testing or Dynamic Symbolic Execution) alleviated the limitations of classical symbolic execution by combining concrete execution and symbolic execution.  We demonstrated that concolic testing is an effective technique for generating high-coverage test suites and for finding deep errors in complex software applications. The success of concolic testing in scalably and exhaustively testing real-world software was a major milestone in the ad hoc world of software testing and has inspired the development of several industrial and academic automated testing and security tools.<\/p><p>One of the key challenges of concolic testing is the huge number of programs paths in all but the smallest programs, which is usually exponential in the number of static branches in the code.  In this talk I will describe MultiSE, a new technique for merging states incrementally during symbolic execution, without using auxiliary variables. The key idea of MultiSE is based on an alternative representation of the state, where we map each variable, including the program counter, to a set of guarded symbolic expressions called a value summary. MultiSE has several advantages over conventional symbolic execution and state merging techniques: 1) value summaries enable sharing of symbolic expressions and path constraints along multiple paths, 2) value-summaries avoid redundant execution, 3) MultiSE does not introduce auxiliary symbolic values, which enables it to make progress even when merging values not supported by the constraint solver, such as floating point or function values.  We have implemented MultiSE for JavaScript programs in a publicly available open-source tool. Our evaluation of MultiSE on several programs shows that MultiSE can run significantly faster than traditional symbolic execution.<\/p><p>Bio\n<\/p><p>Koushik Sen is an associate professor in the Department of Electrical Engineering and Computer Sciences at the University of California, Berkeley. His research interest lies in Software Engineering, Programming Languages, and Formal methods. He is interested in developing software tools and methodologies that improve programmer productivity and software quality. He is best known for his work on \u00b3DART: Directed Automated Random Testing\u00b2 and concolic testing. He has received a NSF CAREER Award in 2008, a Haifa Verification Conference (HVC) Award in 2009, a IFIP TC2 Manfred Paul Award for Excellence in Software: Theory and Practice in 2010, a Sloan Foundation Fellowship in 2011, a Professor R. Narasimhan Lecture Award in 2014, and an Okawa Foundation Research Grant in 2015. He has won several ACM SIGSOFT Distinguished Paper Awards. He received the C.L. and Jane W-S. Liu Award in 2004, the C. W. Gear Outstanding Graduate Award in 2005, and the David J. Kuck Outstanding Ph.D. Thesis Award in 2007, and a Distinguished Alumni Educator Award in 2014 from the UIUC Department of Computer Science. He holds a B.Tech from Indian Institute of Technology, Kanpur, and M.S. and Ph.D. in CS from University of Illinois at Urbana-Champaign.<\/p><\/div><\/div><br \/><div id=\"seminar-39\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">LaCasa: Lightweight Affinity and Object Capabilities in Scala<\/a>\n<br \/>Philipp Haller\nKTH Royal Institute of Technology\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.csc.kth.se\/~phaller\/\">http:\/\/www.csc.kth.se\/~phaller\/<\/a>\n<br \/>29 September, 2016 12:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Aliasing is a known source of challenges in the context of imperative object-oriented languages, which have led to important advances in type systems for aliasing control. However, their large-scale adoption has turned out to be a surprisingly difficult challenge. While new language designs show promise, they do not address the need of aliasing control in existing languages.<\/p><p>This paper presents a new approach to isolation and uniqueness in an existing, widely-used language, Scala. The approach is unique in the way it addresses some of the most important obstacles to the adoption of type system extensions for aliasing control. First, adaptation of existing code requires only a minimal set of annotations. Only a single bit of information is required per class. Surprisingly, the paper shows that this information can be provided by the object-capability discipline, widely-used in program security. We formalize our approach as a type system and prove key soundness theorems. The type system is implemented for the full Scala language, providing, for the first time, a sound integration with Scala\u2019s local type inference. Finally, we empirically evaluate the conformity of existing Scala open-source code on a corpus of over 75,000 LOC.<\/p><p>Bio\n<\/p><p>Philipp Haller is an assistant professor in the theoretical computer science group at KTH Royal Institute of Technology, the leading technical university in Sweden. His main research interests are programming languages, type systems, concurrent, and distributed programming. Philipp is co-author of Scala's async\/await extension for asynchronous computations, and one of the lead designers of Scala's futures and promises library. As main author of the book \"Actors in Scala,\" he created Scala's first widely-used actors library. Philipp was co-chair of the 2013 and 2014 editions of the Scala Workshop, and co-chair of the 2015 ACM SIGPLAN Scala Symposium. Previously, he has held positions at Typesafe, Stanford University, and EPFL. He received a PhD in computer science from EPFL in 2010.<\/p><\/div><\/div><br \/><div id=\"seminar-40\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Performance in Julia<\/a>\n<br \/>Jan Vitek\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/janvitek.org\/\">http:\/\/janvitek.org\/<\/a>\n<br \/>15 September, 2016 12:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Julia, like R, is a dynamic language for scientific computing but, unlike R, it was explicitly designed to deliver performance competitive to traditional batch-compiled languages. To achieve this Julia's designers made a number of unusual choices, including the presence of a set of type annotations that are used for dispatching methods and speed up code, but not for type-checking. The result is that many Julia programs are competitive with equivalent programs written in C. This talk gives a brief overview of the key points of Julia's design and considers whether similar ideas could be adopted in R.<\/p><p>Bio\n<\/p><p>Jan Vitek, <a href=\"http:\/\/prl.ccs.neu.edu\/seminars.html\">\"http:\/\/janvitek.org\/\"<\/a>, is a Professor at Northeastern University CCIS. He works on the design and implementation of programming abstractions for areas and in languages that are *terrifying*. For instance, he has worked on a real-time JVM for running Java on an actual embedding real-time system, on understanding JavaScript to make it more secure, and on getting R (yes R) to be scalable for data analysis. Apparently he does these things because they are used in the \"real world\" to \"solve problems\". He also has an excellent sense of humor and didn't give me a bio, so I took some liberties and hope he doesn't mind.<\/p><\/div><\/div><br \/><div id=\"seminar-41\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">A Categorical Perspective on Type Refinement<\/a>\n<br \/>Noam Zeilberger\nInria Ecole Polytechnique\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/noamz.org\/\">http:\/\/noamz.org\/<\/a>\n<br \/>13 July, 2016 11:45AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>A \"type refinement system\" is a type system built on top of\na typed programming language, as an extra layer of typing.  Over the\npast few years Paul-Andr\u00e9 Melli\u00e8s and I have investigated a\ncategorical foundation for type refinement systems, which is based on\nthe idea of looking at a type refinement system \"backwards\", in terms\nof the functor which erases the extra layer of types.  The talk will\ngive an introduction to this framework as well as to some of the\ninsights that it offers, such as:<\/p><p>* The idea of viewing Hoare logic as a type refinement system<\/p><p>* The importance of the bifibrational operations of \"pushing forward\" and\n  \"pulling back\" (which generalize strongest postconditions and weakest preconditions)<\/p><p>* How those operations can be combined with a monoidal closed structure, to\n  model type refinement systems built over pure lambda calculus<\/p><p>I will not assume any prior background in category theory, although as background reading audience\nmembers may be interested in the lecture notes for my recent course at OPLSS\n(<a href=\"http:\/\/noamz.org\/oplss16\/refinements-notes.pdf&quot;\">http:\/\/noamz.org\/oplss16\/refinements-notes.pdf\"<\/a>, especially Ch.3).<\/p><p>Bio\nSince receiving his PhD from Carnegie Mellon in 2009, Noam Zeilberger has been actively pursuing a career as an itinerant postdoc, with gigs in Paris (Laboratoire PPS), Madrid (IMDEA Software), Princeton (IAS), and back in Paris (MSR-Inria).  He is currently a member of Inria Team Parsifal at Ecole Polytechnique.  In his free time, he enjoys enumerating lambda terms, and his favorite combinators are B, C and I.<\/p><\/div><\/div><br \/><div id=\"seminar-42\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Leveraging Fine-grained Data Flows in Web Applications<\/a>\n<br \/>James Mickens\nHarvard University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/mickens.seas.harvard.edu\/\">http:\/\/mickens.seas.harvard.edu\/<\/a>\n<br \/>5 July, 2016 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nA modern web page contains megabytes of HTML, CSS, images, and JavaScript. Loading such a page requires a browser to evaluate a complex dependency graph involving those resources; once the page is loaded, subsequent interactions between those resources and the user can lead to tricky-to-diagnose bugs. In this talk, I'll describe how tracking fine-grained data flows can allow us to reduce page load times by prioritizing the loads of the highest ancestors in the data flow graph. I'll also describe initial work in using data flows to assist with time-travel debugging (in which developers use a logging-and-replay framework to analyze buggy program executions).<\/p><p>Bio\nJames Mickens is an IEEE Knight of the Republic, an ACM Templar for Non-Open Access, and a Royal Proceeding of Her Majesty\u2019s Royal Proceedings. His appreciation for syntactically correct code has led him to be called \u201ca semicolon in human form.\u201d His online shopping habits have too many dimensions to be k-means clustered, so he is only shown ads about dinosaurs and ancient siege machines. This does not bother James Mickens, and explains why he spends his summers attacking France with triceratops horns.<\/p><\/div><\/div><br \/><div id=\"seminar-43\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">A Coq Library For Internal Verification of Running-Times<\/a>\n<br \/>Jay McCarthy\nU. Massachusetts, Lowell\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/jeapostrophe.github.io\/home\/\">https:\/\/jeapostrophe.github.io\/home\/<\/a>\n<br \/>17 May, 2016 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nWe present a Coq library that lifts an abstract yet precise\nnotion of running-time into the type of a function. Our library is\nbased on a monad that counts abstract steps, controlled by one of the\nmonadic operations. The monad's computational content, however, is\nsimply that of the identity monad so programs written in our\nmonad (that recur on the natural structure of their arguments) extract\ninto idiomatic OCaml code. We evaluated the expressiveness of the\nlibrary by proving that red-black tree insertion and search, merge\nsort, insertion sort, Fibonacci, iterated list insertion, BigNum\naddition, and Okasaki's Braun Tree algorithms all have their expected\nrunning times.<\/p><p>Bio\nJay McCarthy is an associate professor at UMass Lowell in the Computer\nScience Department. He is a member of the PLT research group and works\non the Racket programming language. He is passionate about computer\nscience education &amp; diversity, formal verification, programming\nlanguage expressiveness, and his wonderful family.<\/p><\/div><\/div><br \/><div id=\"seminar-44\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">New language ideas for user-defined side-effects: algebraic effect handlers.<\/a>\n<br \/>Gabriel Scherer\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/gallium.inria.fr\/~scherer\/\">http:\/\/gallium.inria.fr\/~scherer\/<\/a>\n<br \/>18 May, 2016 11:45AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nThis talk reports on some cool ideas I learned during a week-long\nseminar in Dagstuhl last month; it will present other people's work,\nnot my own. It is aimed at an audience interested in programming\nlanguage design in general, but not familiar with the theoretical\ntreatment of pure languages and side-effects.\nWe shall start with *demos* first, explaining short code examples in\nthree programming languages: ML with exceptions as a warmup, then the\nnew language 'Eff' of Matija Pretnar and Andrej Bauer, that first\nimplemented so-called effect handlers, and finally the language\n'Frank' of Conor McBride, that integrates these handlers in a more\nuniform style of \"effectful call-by-value programming\".\nOnly second will we discuss some *theory*, in an accessible way:\nmonads and algebraic effects, which are two distinct ways to formalize\nside-effects, and the difference between \"direct\" and \"indirect\" style\nof effectful programming.\nUnderlying this talk are two larger, important questions of\nprogramming language design, that will be touched during the talk and\nwe can discuss further afterwards:\n- Do programming language need a facility for user-defined side-effects?\n- When should we encode new design ideas as libraries\/macros in an\n  expressive language, and when should we design languages afresh for\n  them?<\/p><p>Bio\nGabriel is interested in theoretical aspects of type systems,\nprogramming language implementation, general programming language\nconcepts, and even some syntactic aspects. He has a preference for the\nformalizable aspects, or formalizable approaches to programming\nlanguage aspects, rather than the often subjective appeal to taste or\nintuition.<\/p><\/div><\/div><br \/><div id=\"seminar-45\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Internet of Things: From Small- to Large-Scale Orchestration<\/a>\n<br \/>Charles Consel\nBordeaux Institute of Technology\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/phoenix.inria.fr\/charles-consel\">http:\/\/phoenix.inria.fr\/charles-consel<\/a>\n<br \/>2 May, 2016 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nThe domain of Internet of Things (IoT) is rapidly expanding beyond research and becoming a major industrial market with such stakeholders as major manufacturers of chips and connected objects, and fast-growing operators of low-power wide-area networks. Importantly, this emerging domain is driven by applications that leverage the infrastructure to provide users with innovative, high-value services. Because of this application-centric approach, software development plays a key role to realize the full potential of IoT.\nIn this talk, we argue that there is a continuum between orchestrating connected objects in the small and in the large, fostering a unified approach to application development. We examine the requirements for orchestrating connected objects and address them with domain-specific design concepts. We then show how to map these design concepts into dedicated programming patterns and runtime mechanisms.\nOur work revolves around domain-specific notations integrated into a tool-based design methodology, dedicated to develop IoT applications. We have applied our work across a spectrum of infrastructure sizes; we present examples, ranging from an automated pilot in avionics, to an assisted living platform for the home of seniors, to a parking management system in a smart city.<\/p><p>Bio\nCharles Consel is a professor of Computer Science at Bordeaux Institute of Technology. He served on the faculty of Yale University, Oregon Graduate Institute and the University of Rennes.\nHis research contributions cover programming languages, software engineering, operating systems, pervasive computing, and assistive computing.\nHe leads the Phoenix group at Inria that conducts multi-disciplinary research to design, develop, deploy and assess assistive computing support. This research combines (1) Cognitive Science to study user needs and make a rigorous assessment of assistive services; (2) Sensing and actuating expertise to support the user, based on accurate and rich interactions with the environment; (3) Computer Science to support and guide the development process of the assistive services.<\/p><\/div><\/div><br \/><div id=\"seminar-46\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">STOKE: Search-Based Compiler Optimization<\/a>\n<br \/>Alex Aiken\nStanford University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/theory.stanford.edu\/~aiken\/\">https:\/\/theory.stanford.edu\/~aiken\/<\/a>\n<br \/>15 April, 2016 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nThe optimization component of a compiler takes a program as input and\nproduces another, hopefully faster, program as output.  At a high level,\noptimizers solve a search problem over a space of programs, but the\ntraditional architecture of a compiler's optimizer does no search at\nall.  This talk will present the STOKE project, which is exploring the\nuse of Monte Carlo search methods as the basis of a modern compiler\noptimizer.  We will show that search-based program optimization can\nconsistently improve, sometimes substantially, on current production\ncompilers at their highest levels of optimization and can even compete\nwith expert, hand-written  assembly.  We will also discuss the unique\nand challenging verification problems that arise when a compiler\nproduces code using a random process.<\/p><p>Bio\nAlex Aiken is the Alcatel-Lucent Professor and the current\nTencent Chair of the Computer Science Department at Stanford. Alex\nreceived his Bachelors degree in Computer Science and Music from\nBowling Green State University in 1983 and his Ph.D. from Cornell\nUniversity in 1988. Alex was a Research Staff Member at the IBM\nAlmaden Research Center (1988-1993) and a Professor in the EECS\ndepartment at UC Berkeley (1993-2003) before joining the Stanford\nfaculty in 2003. His research interest is in areas related to\nprogramming languages. He is an ACM Fellow, a recipient of Phi Beta\nKappa's Teaching Award, and a former National Young Investigator.<\/p><\/div><\/div><br \/><div id=\"seminar-47\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">The Secret Life of Mobile Applications<\/a>\n<br \/>Julia Rubin\nMIT\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/people.csail.mit.edu\/mjulia\/\">https:\/\/people.csail.mit.edu\/mjulia\/<\/a>\n<br \/>5 April, 2016 10:30AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nAs software becomes increasingly more complex and yet more pervasive, poor understanding of software behavior compromises the quality and the integrity of software systems that we use. In this talk, I will show that automated analysis techniques can help to identify and reason about software behavior characteristics that matter to humans. After a brief overview of my current research directions, I will focus on techniques for identifying privacy violations in mobile applications, i.e., leakages of sensitive information such as user location and shopping preferences. I will present a set of solutions that rely on contextual, functional and usage-based clues for improving the accuracy of leakage detection and for distinguishing between \u201clegitimate\u201d and \u201cillegitimate\u201d information distribution patterns.<\/p><p>Bio\nJulia Rubin is a Postdoctoral Researcher in the EECS department at MIT. Prior to that, she was a Research Staff Member and, part of the time, a manager at IBM Research in Haifa, Israel. She received her PhD in Computer Science from the University of Toronto, Canada in 2014. Julia\u2019s research interests are in software engineering, program analysis and software security, focusing on improving the quality and the integrity of modern software systems. Her recent work in this area won an ACM Distinguished Paper Award at ASE, two Best Paper Awards, at SPLC and CSMR, and was nominated for Facebook\u2019s Internet Defense Prize at the USENIX Security Symposium.<\/p><\/div><\/div><br \/><div id=\"seminar-48\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Reasoning about Object Capabilities with Logical Relations and Effect Parametricity<\/a>\n<br \/>Dominique Devriese\nKatholieke Universiteit Leuven\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/distrinet.cs.kuleuven.be\/people\/dominiqu\">https:\/\/distrinet.cs.kuleuven.be\/people\/dominiqu<\/a>\n<br \/>25 February, 2016 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nObject capabilities are a technique for fine-grained privilege separation in programming languages and systems,with important applications in security. However, current formal characterisations do not fully capture capability-safety of a programming language and are not sufficient for verifying typical applications. Using state-of-the-art techniques from programming languages research, we define a logical relation for a core calculus of JavaScript that better characterises capability-safety. The relation is powerful enough to reason about typical capability patterns and supports evolvable invariants on shared data structures, capabilities with restricted authority over them and isolated components with restricted communication channels. We use a novel notion of effect parametricity for deriving properties about effects. We demonstrate that our results imply memory access bounds that have previously been used to characterise capability-safety.<\/p><p>Bio\nDominique is a postdoctoral researcher in the research group DistriNet, part of the Computer Science department of the Katholieke Universiteit Leuven. He holds a postdoctoral research fellowship of the Research Foundation - Flanders (FWO). He works on formalising properties of object-oriented and object-capability programming languages---specifically a property called effect parametricity--is the author of the grammar-combinators Haskell parsing library, and has added instance arguments to the programming language\/proof assistant Agda. He is interested in information flow security, secure compilation,full abstraction, and functional and dependently typed programming languages.<\/p><\/div><\/div><br \/><div id=\"seminar-49\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">JavaScript in the Small<\/a>\n<br \/>Satish Chandra\nSamsung\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/sites.google.com\/site\/schandraacmorg\/\">https:\/\/sites.google.com\/site\/schandraacmorg\/<\/a>\n<br \/>22 February, 2016 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nEmerging consumer electronics products are running the same software platforms that power smartphones. This leads to the appealing idea that a uniform programming abstraction can be used for app development for a range of devices, from wearables to smartphones. In practice, however, devices wary in their hardware capabilities and this has an impact on app development. In this talk, I will focus on ways in which JavaScript may be used and\/or run differently to accommodate the different hardware capabilities across devices. Specifically, I will present a subset of JavaScript that strikes a balance between retaining the flavor of JavaScript (e.g. use of prototype inheritance, no explicit type declarations) and the possibility of ahead-of-time compilation to efficient code.<\/p><p>Bio\nSatish Chandra obtained a PhD from the University of Wisconsin-Madison in 1997, and a B.Tech from the Indian Institute of Technology-Kanpur in 1991, both in computer science. From 1997 to 2002, he was a member of technical staff at Bell Laboratories, where his research focused on program analysis, domain-specific languages, and data-communication protocols. From 2002 to 2013, he was a research staff member at IBM Research, where his research focused on bug finding and verification, software synthesis, and test automation. He joined Samsung Electronics in 2013, where he leads the advanced programming tools research team. He is an ACM Distinguished Scientist.<\/p><\/div><\/div><br \/><div id=\"seminar-50\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Journey to Find Bugs in Real-World Web Applications in the Wild<\/a>\n<br \/>Sukyoung Ryu\nKorea Advanced Institute of Science and Technology\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/plrg.kaist.ac.kr\/ryu\">http:\/\/plrg.kaist.ac.kr\/ryu<\/a>\n<br \/>19 February, 2016 11:45AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nAnalyzing real-world web applications is a challenging task. On top of understanding the semantics of JavaScript, it requires modeling of web documents, platform objects, and interactions between them. Most web applications load JavaScript code dynamically, which makes pure static analysis approaches inapplicable. In this talk, we present our journey to analyze JavaScript web applications in the wild using various approaches. From pure JavaScript programs to JavaScript web applications using platform-specific libraries, we explain technical challenges in analyzing each of them and how we built an open-source analysis framework for JavaScript, SAFE, that addresses the challenges incrementally.  Finally, we discuss our ongoing work on analysis of Android hybrid applications.<\/p><p>Bio\n<\/p><\/div><\/div><br \/><div id=\"seminar-51\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Taming release-acquire consistency<\/a>\n<br \/>Ori Lahav\nMax Planck Institute for Software Systems\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/www.mpi-sws.org\/~orilahav\/\">https:\/\/www.mpi-sws.org\/~orilahav\/<\/a>\n<br \/>27 January, 2016 11:45AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Multiprocessors and concurrent programming are now pervasive. Typically, they do not guarantee sequential consistency (a.k.a. interleaving semantics), which is the standard assumption by most work on semantics and verification. Instead, they employ subtle memory models, exposing unexpected relaxed behaviors arising from hardware and compiler optimizations.<\/p><p>In this talk, I will focus on one such model --- the release-acquire fragment of the C\/C++11 memory model. I will describe its merits, and show how it can be further improved, without additional implementation costs, to: (i) forbid dubious behaviors that are not observed in any implementation; (ii) support fence instructions that restore sequential consistency; and (iii) admit an equivalent intuitive operational semantics.<\/p><p>The talk is based on a joint work with Nick Giannarakis and Viktor Vafeiadis, to be presented in POPL'16.<\/p><p>Bio\nOri Lahav is a postdoctoral researcher at MPI-SWS. He obtained his PhD from Tel Aviv University in the area of non-classical logics. His current research interests are in programming languages generally, with specific focus on memory models, concurrency, verification, and logic.<\/p><\/div><\/div><br \/><div id=\"seminar-52\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Secure Compilation to Protected Module Architectures<\/a>\n<br \/>Marco Patrignani\nMax Planck Institute for Software Systems\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.mpi-sws.org\/~marcopat\/\">http:\/\/www.mpi-sws.org\/~marcopat\/<\/a>\n<br \/>25 January, 2016 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nThis talk will informally describe protected module architectures (PMA), a security architecture that provides an assembly-level memory isolation mechanism. \tThen it will describe how to devise a secure (fully-abstract) compiler for an object-oriented language to PMA.  Finally, it will present how to prove the said compiler to be secure and discuss open, related research trajectories.<\/p><p>Bio\nMarco Patrignani did his bachelor and masters study in Bologna, then he obtained a Ph.D. in computer science from the KU Leuven, Belgium. There, with his advisors Dave Clarke and Frank Piessens, he studied secure (fully-abstract) compilation for Intel-SGX like architectures, i.e., untyped assembly languages extended with a memory isolation mechanism. Additionally, he investigated trace-based characterisation of the behaviour of those architectures. He is now a post-doc at MPI-SWS, Germany working on more secure-compilation-related topics with Deepak Garg.<\/p><\/div><\/div><br \/><div id=\"seminar-53\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Program verification under weak memory consistency<\/a>\n<br \/>Viktor Vafeiadis\nMax Planck Institute for Software Systems\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.mpi-sws.org\/~viktor\/\">http:\/\/www.mpi-sws.org\/~viktor\/<\/a>\n<br \/>14 January, 2016 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Weak memory models formalize the inconsistent behaviors that one can observe in multithreaded programs running on modern hardware. In so doing, they complicate the already-difficult task of reasoning about correctness of concurrent code. Worse, they render impotent most formal methods that have been developed to tame concurrency, which almost universally assume a strong (i.e., sequentially consistent) memory model. In response, we have developed a number of alternative reasoning techniques that are sound for programs running weak memory consistency. I will cover both program logics, such as relaxed separation logic, as well as theorems that allow reducing reasoning about well-structured weakly consistent implementations down to sequential consistency, and show how these can be applied to reason about a practical RCU implementation.<\/p><p>Bio\nViktor Vafeiadis is a tenure-track researcher at MPI-SWS. He received his BA (2004) and PhD (2008) from the University of Cambridge. Before joining MPI-SWS in October 2010, he was a postdoc at Microsoft Research and at the University of Cambridge. He is interested in programming languages and verification with a focus program logics for weak memory, program logics for concurrency, compiler verifications, automated verification of concurrent programs, and interactive theorem proving.<\/p><\/div><\/div><br \/><div id=\"seminar-54\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Lightweight Formal Methods for LLVM Verification<\/a>\n<br \/>Santosh Nagarakatte\nRutgers University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.cs.rutgers.edu\/~santosh.nagarakatte\/\">http:\/\/www.cs.rutgers.edu\/~santosh.nagarakatte\/<\/a>\n<br \/>13 January, 2016 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Compilers form an integral component of the software\ndevelopment ecosystem. Compiler writers must understand the\nspecification of source and target languages to design sophisticated\nalgorithms that transform programs while preserving\nsemantics. Unfortunately, compiler bugs in mainstream compilers are\ncommon. Compiler bugs can manifest as crashes during compilation, or,\nmuch worse, result in the silent generation of incorrect\nprograms. Such mis-compilations can introduce subtle errors that are\ndifficult to diagnose and generally puzzling to software\ndevelopers.<\/p><p>The talk will describe the problems in developing peephole\noptimizations that perform local rewriting to improve the efficiency\nof LLVM code. These optimizations are individually difficult to get\nright, particularly in the presence of undefined behavior; taken\ntogether they represent a persistent source of bugs.  The talk will\npresent Alive, a domain-specific language for writing optimizations\nand for automatically either proving them correct or else generating\ncounterexamples.  A transformation in Alive is shown to be correct\nautomatically by encoding the transformation into constraints, which\nare automatically checked for validity using a Satisfiability Modulo\nTheory (SMT) solver. Furthermore, Alive can be automatically\ntranslated into C++ code that is suitable for inclusion in an LLVM\noptimization pass.<\/p><p>Alive is based on an attempt to balance usability and formal\nmethods; for example, it captures\u2014but largely hides\u2014the detailed\nsemantics of three different kinds of undefined behavior in LLVM. We\nhave translated more than 300 LLVM optimizations into Alive and, in\nthe process, found that eight of them were wrong.  I will conclude the\ntalk highlighting the lessons learned and the challenges in\nincorporating lightweight formal methods in the tool-chain of the\ncompiler developer.<\/p><p>Bio\nSantosh Nagarakatte is an Assistant Professor of Computer Science at Rutgers University. He obtained his PhD from the University of Pennsylvania. His research interests are in Hardware-Software Interfaces spanning Programming Languages, Compilers, Software Engineering, and Computer Architecture. His papers have been selected as IEEE MICRO TOP Picks papers of computer architecture conferences in 2010 and 2013. He has received the NSF CAREER Award in 2015, PLDI 2015 Distinguished Paper Award, and the Google Faculty Research Award in 2014 for his research on incorporating lightweight formal methods for LLVM compiler verification.<\/p><\/div><\/div><br \/><div id=\"seminar-55\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Feature Specific Profiling for R<\/a>\n<br \/>Leif Andersen\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/leifandersen.net\">http:\/\/leifandersen.net<\/a>\n<br \/>2 December, 2015 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Programmers use profilers to understand the performance\ncharacteristics of\ntheir programs and to focus on those pieces whose improvement may yield the\nlargest gains. A conventional profiler measures the time that a program\nspends in functions, methods, expressions, and statements. Racket's novel\nfeature-specific profiler supplements this information with timings of\ninstances of linguistic features. This paper reports the results of a\nsuccessful reproducibility effort to adapt feature-specific profiling to\nthe R programming language. Specifically, the paper demonstrates how easy\nand effective it is to add the necessary language support, that the\napproach usefully enhances the information produced by a classical\nprofiler, and that the additional overhead is tolerable.<\/p><p>Bio\nLeif Andersen is a second year Ph.D. student at\nNortheastern University, studying programming language with Matthias\nFelleisen.<\/p><\/div><\/div><br \/><div id=\"seminar-56\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Performance Matters<\/a>\n<br \/>Emery Berger\nU. Massachusetts, Amherst\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/emeryberger.com\/\">http:\/\/emeryberger.com\/<\/a>\n<br \/>20 November, 2015 10:30AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>Performance clearly matters to users. The most common software update\non the AppStore *by far* is \"Bug fixes and performance enhancements.\"\nNow that Moore's Law Free Lunch has ended, programmers have to work\nhard to get high performance for their applications. But why is\nperformance so hard to deliver?<\/p><p> I will first explain why our current approaches to evaluating and\n optimizing performance don't work, especially on modern hardware and\n for modern applications. I will then present two systems that address\n these challenges. Stabilizer is a tool that enables statistically\n sound performance evaluation, making it possible to understand the\n impact of optimizations and conclude things like the fact that the -O2\n and -O3 optimization levels are indistinguishable from noise\n (unfortunately true).<\/p><p> Since compiler optimizations have largely run out of steam, we need\n better profiling support, especially for modern concurrent,\n multi-threaded applications. Coz is a novel \"causal profiler\" that\n lets programmers optimize for throughput or latency, and which\n pinpoints and accurately predicts the impact of optimizations. Coz's\n approach unlocks numerous previously unknown optimization\n opportunities. Guided by Coz, we improved the performance of Memcached\n by 9%, SQLite by 25%, and accelerated six Parsec applications by as\n much as 68%; in most cases, these optimizations involved modifying\n under 10 lines of code.<\/p><p> This talk is based on work with Charlie Curtsinger published at ASPLOS\n 2013 (Stabilizer) and SOSP 2015 (Coz), where it received a Best Paper\n Award.<\/p><p>Bio\n<\/p><p>Emery Berger is a Professor in the College of Information and Computer\nSciences at the University of Massachusetts Amherst, the flagship\ncampus of the UMass system. He graduated with a Ph.D. in Computer\nScience from the University of Texas at Austin in 2002. Professor\nBerger has been a Visiting Scientist at Microsoft Research (7 times)\nand at the Universitat Politecnica de Catalunya (UPC) \/ Barcelona\nSupercomputing Center (BSC).<\/p><p>Professor Berger\u2019s research spans programming languages, runtime\nsystems, and operating systems, with a particular focus on systems\nthat transparently improve reliability, security, and performance. He\nis the creator of a number of influential software systems including\nHoard, a fast and scalable memory manager that accelerates\nmultithreaded applications (used by companies including British\n                                 Telecom, Cisco, Cr\u00e9dit Suisse, Reuters, Royal Bank of Canada, SAP, and\n                                 Tata, and on which the Mac OS X memory manager is based); DieHard, an\nerror-avoiding memory manager that directly influenced the design of\nthe Windows 7 Fault-Tolerant Heap; and DieHarder, a secure memory\nmanager that was an inspiration for hardening changes made to the\nWindows 8 heap.<\/p><p>His honors include a Microsoft Research Fellowship, an NSF CAREER\nAward, a Lilly Teaching Fellowship, a Most Influential Paper Award at\nOOPSLA 2012, a Google Research Award, a Microsoft SEIF Award, a Best\nArtifact Award at PLDI, and Best Paper Awards at FAST, OOPSLA, and\nSOSP; he was named an ACM Senior Member in 2010. Professor Berger is\ncurrently a Member of the SIGPLAN Executive Committee and an Associate\nEditor of the ACM Transactions on Programming Languages and Systems,\nand will serve as Program Chair for PLDI 2016.<\/p><\/div><\/div><br \/><div id=\"seminar-57\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Programming Languages Meets Programmable Networks<\/a>\n<br \/>Arjun Guha\nU. Massachusetts, Amherst\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/people.cs.umass.edu\/~arjun\/home\/\">https:\/\/people.cs.umass.edu\/~arjun\/home\/<\/a>\n<br \/>16 November, 2015 1:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n            <\/p><p>Computer networks do not simply connect machines together, but run several\n         applications on network devices, such as load balancers, intrusion detection\n         systems, authentication portals, and more. Historically, these applications were\n         black-boxes running on proprietary hardware, but software-defined networking\n         (SDN) now allows anyone to write their own programs using open networking\n         protocols (e.g., OpenFlow). So, what are the right abstractions for programming networks? This talk will try\n         to address this question in three ways.<\/p><p>First, we present a syntactic theory of network forwarding called NetKAT, which supports equational reasoning about network-wide behavior. Using NetKAT, programmers can ask and answer questions like, \"Can A communicate with B?\",\n                        \"Does all traffic traverse my intrusion detection system?\", \"Is there a loop in\nmy network?\", and so on.<\/p><p>Second, we present a fast and efficient compiler for NetKAT. Although several\nnetwork compilers already exist, they are unusable on even moderately sized\nnetworks. Using new data structures and compilation algorithms, our new compiler\nis two orders of magnitudes faster than prior work and scales to large\ndatacenter networks.<\/p><p>Finally, we consider the problem of building a reliable runtime system for\nNetKAT. NetKAT abstracts away several low-level details of networking hardware.\nAlthough this is a boon for the network programmer, the burden now shifts to us\nto engineer abstractions correctly. We present a Coq-certified runtime system\nthat is proven correct with respect to a detailed operational model software-\ndefined networks.<\/p><p>Bio\nArjun Guha is an assistant professor of Computer Science at UMass Amherst. He\n      enjoys tackling problems in systems using the tools and principles of\n      programming languages. Apart from network programming, he has worked on Web\n      security and system configuration languages. He received a PhD in Computer\n      Science from Brown University in 2012 and a BA in Computer Science from Grinnell\n      College in 2006.<\/p><\/div><\/div><br \/><div id=\"seminar-58\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Declarative Programming for Eventual Consistency<\/a>\n<br \/>Suresh Jagannathan\nPurdue University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/www.cs.purdue.edu\/homes\/suresh\/\">https:\/\/www.cs.purdue.edu\/homes\/suresh\/<\/a>\n<br \/>13 November, 2015 10:30AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n<\/p><p>In geo-replicated distributed data stores, the need to ensure responsiveness\n   in the face of network partitions and processor failures results in\n   implementations that provide only weak (so-called eventually consistent)\n   guarantees on when data updated by one process becomes visible to another.\n   Applications must be carefully constructed to be aware of unwanted\n   inconsistencies permitted by such implementations (e.g., having negative\n                                                            balances in a bank account, or having an item appear in a shopping cart\n                                                            after it has been removed), but must balance correctness concerns with\n                                                            performance and scalability needs.  Because understanding these tradeoffs\n                                                            requires subtle reasoning and detailed knowledge about the underlying data\n                                                            store, implementing robust distributed applications in such environments is\n                                                            often an error-prone and expensive task.<\/p><p>To overcome these issues, this talk presents a declarative programming model\nfor eventually consistent data stores called Quelea.  The model comprises a\ncontract language, capable of defining fine-grained application-level\nconsistency properties for replicated data types (and transactions over\n                                                      objects of these types), and a contract enforcement system to analyze\n                                                      contracts and automatically generate the appropriate consistency protocol\n                                                      for the method protected by the contract.  By doing so, Quelea enables\n                                                      programmers to reason compositionally about consistency from the perspective\n                                                      of high-level application requirements, not low-level implementation\n                                                      features.<\/p><p>This is joint work with Gowtham Kaki and K.C. Sivaramakrishnan.<\/p><p>Bio\n<\/p><p>Suresh Jagannathan is a Professor of Computer Science at Purdue University\n       where he has been on leave since September 2013, serving as a program\n       manager in the Information Innovation Office at DARPA.  He has also been a\n       visiting faculty at Cambridge University, where he spent a sabbatical year\n       in 2010; and, prior to joining Purdue, was a senior research scientist at\n       the NEC Research Institute in Princeton, N.J.  He received his Ph.D from\n       MIT.<\/p><p>His research interests are in programming languages generally, with specific\nfocus on compilers, functional programming, program verification, and\nconcurrent and distributed systems.  At DARPA, he manages programs on\nprobabilistic programming and machine learning (PPAML), program synthesis\nand repair leveraging predictive analytics over large software corpora\n(MUSE), and self-adaptive software through resource-aware analyses,\nruntimes, and architectures (BRASS).<\/p><\/div><\/div><br \/><div id=\"seminar-59\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Hop.js: multitier programming in JavaScript<\/a>\n<br \/>Manuel Serrano\nINRIA\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www-sop.inria.fr\/members\/Manuel.Serrano\/\">http:\/\/www-sop.inria.fr\/members\/Manuel.Serrano\/<\/a>\n<br \/>3 November, 2015 10:30AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nHop.js is a multitier extension of JavaScript. It allows a single\n       JavaScript program to describe the client-side and the server-side\n       components of a Web application. Its runtime environment ensures a\n       consistent execution of the application on the server and on the\n       client. This talk will shows its minimal set of JavaScript extensions\n       that makes Web programming easier. It will present its runtime\n       environment, with an emphasize on the handling of server-side\n       parallelism.<\/p><p>Bio\nManuel is a researcher at INRIA Sophia Antipolis, he used to work on Scheme.<\/p><\/div><\/div><br \/><div id=\"seminar-60\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Scalloc and Selfie: Fast Memory Allocation and Self-referential Systems Software<\/a>\n<br \/>Christoph Kirsch\nUniversity of Salzburg\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/cs.uni-salzburg.at\/~ck\/\">http:\/\/cs.uni-salzburg.at\/~ck\/<\/a>\n<br \/>2 November, 2015 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n  This talk is about scalloc, a fast, multicore-scalable, low-fragmentation memory allocator and selfie, a 4000-line implementation of a tiny self-compiling C compiler and a tiny self-executing MIPS emulator for teaching systems engineering. Scalloc is a typical example of a very complex, multi-year research effort while selfie is, at least for now, a purely educational, many-year effort in teaching compiler, operating system, and virtual machine design based on a single, highly principled software platform. So far scalloc and selfie only share the passion of their authors and are otherwise two distinct projects. Yet earlier versions of selfie, before they were even identified as such, were instrumental in bringing up the generation of students who did scalloc.\nThe main ideas behind scalloc are: uniform treatment of small and big objects through so-called virtual spans, efficiently and effectively reclaiming free memory through fast and scalable global data structures, and constant-time (modulo synchronization) allocation and deallocation operations that trade off memory reuse and spatial locality without being subject to false sharing. The main ideas behind selfie are: a compiler written in and for a tiny subset of C called C* which uses the dereferencing * operator of C for memory access but lacks data structures and many other features and a MIPS emulator written in C* that can execute itself. Both are combined and extended by students to do very cool stuff.<\/p><p>Bio\n Christoph Kirsch is Professor at the University of Salzburg, Austria. From\n1999 to 2004 he worked as Postdoctoral Researcher at UC, Berkeley. He later\nreturned to Berkeley as Visiting Scholar (2008-2013) and Visiting Professor\n(2014) as part of a collaborative research effort in Cyber-Physical\nSystems. His most recent research interests are in concurrent data\nstructures, memory management, and so-called spatial programming. Dr. Kirsch\nco-invented embedded programming languages and systems such as Giotto, HTL,\nand the Embedded Machine, and more recently co-designed high-performance,\nmulticore-scalable concurrent data structures and memory management\nsystems. He co-founded the International Conference on Embedded Software\n(EMSOFT) and served as ACM SIGBED chair from 2011 until 2013 and ACM TODAES\nassociate editor from 2011 until 2014.<\/p><\/div><\/div><br \/><div id=\"seminar-61\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Coordinated Concurrent Programming in Syndicate<\/a>\n<br \/>Tony Garnock-Jones\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/www.ccs.neu.edu\/home\/tonyg\/\">http:\/\/www.ccs.neu.edu\/home\/tonyg\/<\/a>\n<br \/>28 October, 2015 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nMost programs interact with the world: via graphical\nuser interfaces, networks, etc. This form of interactivity entails\nconcurrency, and concurrent program components must coordinate their\ncomputations. This talk will present Syndicate, a design for a coordinated,\nconcurrent programming language. Each concurrent component in Syndicate is a\nfunctional actor that participates in scoped conversations. The medium of\nconversation arranges for message exchanges and coordinates access to common\nknowledge. As such, Syndicate occupies a point in design space halfway\nbetween actors and threads.<\/p><p>Bio\n Tony is a graduate student working with Matthias.<\/p><\/div><\/div><br \/><div id=\"seminar-62\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Memory corruption: why protection is hard<\/a>\n<br \/>Mathias Payer\nPurdue University\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/nebelwelt.net\">https:\/\/nebelwelt.net<\/a>\n<br \/>23 October, 2015 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n  Memory corruption plagues systems since the dawn of computing. With the\nrise of defense techniques like stack cookies, ASLR, and DEP, attacks\nhave become much more complicated, yet control-flow hijack attacks are\nstill prevalent. Attacks leverage code reuse attacks, often using some\nform of information disclosure. Stronger defense mechanisms have been\nproposed but none have seen wide deployment so far (i) due to the time\nit takes to deploy a security mechanism, (ii) incompatibility with\nspecific features, and (iii) most severely due to performance overhead.\nIn the talk, we evaluate the security benefits and limitations of the\nstatus quo and look into upcoming defense mechanisms (and their attacks).\nControl-Flow Integrity (CFI) and Code-Pointer Integrity (CPI) are two of\nthe hottest upcoming defense mechanisms. CFI guarantees that the runtime\ncontrol flow follows the statically determined control-flow graph. An\nattacker may reuse any of the valid transitions at any control flow\ntransfer. CPI on the other hand is a dynamic property that enforces\nmemory safety guarantees like bounds checks for code pointers by\nseparating code pointers from regular data. We will discuss differences\nand advantages\/disadvantages of both approaches, especially the security\nbenefits they give under novel attacks like Counterfeit Object-Oriented\nProgramming (COOP) and Control-Flow Bending (CFB).  COOP reuses complete\nfunctions as gadgets and CFB bends the control flow along valid but\nunintended paths in the control-flow graph of a program.\n<\/p><p>Bio\nMathias Payer is a security researcher and an assistant professor in\ncomputer science at Purdue university. His interests are related to\nsystem security, binary exploitation, user-space software-based fault\nisolation, binary translation\/recompilation, and (application)\nvirtualization. His research focuses on protecting applications even in\nthe presence of vulnerabilities, with a focus on memory corruption.\nBefore joining Purdue in 2014 he spent two years as PostDoc in Dawn\nSong's BitBlaze group at UC Berkeley. He graduated from ETH Zurich with\na Dr. sc. ETH in 2012.<\/p><\/div><\/div><br \/><div id=\"seminar-63\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Using LLVM as a backend for R<\/a>\n<br \/>Jan Vitek\nNortheastern University\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/janvitek.org\">http:\/\/janvitek.org<\/a>\n<br \/>2 October, 2015 12:00PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nI will provide an update on the status of the Reactor project which aims to use LLVM as just-in-time compiler for the R language.  I will discuss early challenges such as integration in the environment and garbage collection support.<\/p><p>Bio\nJan Vitek is a Professor at Northeastern, he works on programming language design and implementation.<\/p><\/div><\/div><br \/><div id=\"seminar-64\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Completions and Diagnostics in RStudio<\/a>\n<br \/>Kevin Ushey\nRStudio\n<a class=\"pn-url\" target=\"_blank\" href=\"https:\/\/kevinushey.github.io\/\">https:\/\/kevinushey.github.io\/<\/a>\n<br \/>2 October, 2015 11:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\nKevin will discuss the details behind the implementation of completions +\ndiagnostics, as well as some future goals re: enabling user extensibility of\nboth the autocompletion and diagnostics systems.<\/p><p>Bio\nKevin is a software engineer at RStudio. He graduated from the University of British Columbia with an MSc in Statistics, with a thesis focusing on the use of non-linear mixed effects models in the analysis of yeast growth curves in gene knockout studies. Shortly thereafter, he worked doing data analysis at St. Paul's hospital in Vancouver as part of the Daley lab, and later at the Fred Hutchinson Cancer Research Center in Seattle as part of the Gottardo lab. At RStudio, Kevin primarily works on the RStudio IDE, but also maintains the R package packrat.<\/p><\/div><\/div><br \/><div id=\"seminar-65\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Some New Developments for the R Engine<\/a>\n<br \/>Luke Tierney\nUniversity of Iowa\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/homepage.stat.uiowa.edu\/~luke\/\">http:\/\/homepage.stat.uiowa.edu\/~luke\/<\/a>\n<br \/>2 October, 2015 10:00AM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n      R is a dynamic language for statistical\ncomputing and graphics. In recent years R has become a major framework\nfor both statistical practice and research. This talk present a very\nbrief outline of the R language and its evolution and describe some\ncurrent efforts on improvements to the core computational engine,\nincluding work on compilation of R code, efforts to take advantage of\nmultiple processor cores, and modifications to support working with\nlarger data sets.<\/p><p>Bio\nLuke Tierney is a Professor Statistics at the University of Iowa and one of the key contributors to the R project.<\/p><\/div><\/div><br \/><div id=\"seminar-66\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Verification Infrastructure for Permission-based Reasoning<\/a>\n<br \/>Alex Summers\nETH Zu\u0308rich\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/people.inf.ethz.ch\/summersa\/wiki\/\">http:\/\/people.inf.ethz.ch\/summersa\/wiki\/<\/a>\n<br \/>26 June, 2015 1:45PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n      Modern verification techniques are becoming\never-more powerful and sophisticated, and building tools to implement\nthem is a time-consuming and difficult task. Writing a new verifier to\nvalidate each on-paper approach is impractical; for this reason\nintermediate verification languages such as Boogie and Why3 have become\npopular over the last decade. However, verification approaches geared\naround complex program logics (such as separation logic) have typically\nbeen implemented in specialised tools, since the reasoning is hard to\nmap down to first-order automated reasoning. In practice, this means\nthat a rich variety of modern techniques have no corresponding tool\nsupport.  In this talk, I will present the new Silver intermediate\nverification language, which has been designed to facilitate the\nlightweight implementation of a variety of modern methodologies for\nprogram verification. In contrast to lower-level verification languages,\nSilver provides native support for heap reasoning; modes of reasoning\nsuch as concurrent separation logic, dynamic frames and\nrely-guarantee\/invariants can be simply encoded.  Silver has been\ndeveloped as part of the Viper project, which provides two automated\nback-end verifiers for Silver programs. Since releasing our software in\nSeptember last year, it has been used for (internal and external)\nprojects to build tools for Java verification, non-blocking concurrency\nreasoning, flow-sensitive typing and reasoning about GPU and Linux\nkernel code.<\/p><p>Bio\nAlex Summers obtained his PhD from Imperial College\nLondon in 2009, in the area of type systems and\nclassical logics. Since then he has worked in a\nvariety of areas concerning software correctness and\nverification, at Imperial College London and ETH\nZurich. His research interests include developing\nspecification techniques for different (usually\nconcurrent) programming paradigms, and implementing\nthese in automatic verification tools. He was\nrecently awarded the 2015 AITO Dahl-Nygaard Junior\nPrize for his work on type systems and the\nverification of object-oriented programs.<\/p><\/div><\/div><br \/><div id=\"seminar-67\" class=\"col-md-12 pn-seminar\"><p><a class=\"pn-title\">Concurrent Trace Sets for Synchronization Synthesis<\/a>\n<br \/>Roopsha Samanta\nIST Austria\n<a class=\"pn-url\" target=\"_blank\" href=\"http:\/\/pub.ist.ac.at\/~rsamanta\/\">http:\/\/pub.ist.ac.at\/~rsamanta\/<\/a>\n<br \/>26 June, 2015 1:30PM \nWVH 366<\/p><div class=\"pn-abstract-bio\"><p>Abstract\n      In this talk, I will first present a method and\na tool TARA for generating succinct representations of sets of\nconcurrent traces. In our work, we focus on trace sets that contain all\ncorrect or all incorrect permutations of events from a given trace. We\nrepresent such trace sets as Boolean combinations of happens-before\nordering constraints between events.  Our trace set representations can\ndrive diverse verification, fault localization, repair, and synthesis\ntechniques for concurrent programs. In the remainder of the talk, I will\nfocus on the use of our representation for synchronization synthesis.\nThis work appears in POPL 2015 and CAV 2015, and is joint work with\nPavol Cerny, Ed Clarke, Ashutosh Gupta, Tom Henzinger, Arjun\nRadhakrishna, Leonid Ryzhyk and Thorsten Tarrach.<\/p><p>Bio\n<\/p><\/div><\/div><br \/><\/div>","created_at":"2018-12-04T12:50:56+0000","updated_at":"2018-12-04T12:50:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":87,"domain_name":"prl.ccs.neu.edu","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure"],"is_public":false,"id":732,"title":"Clojure's missing\u00a0piece - thoughts from the red planet - thoughts from the red planet","url":"http:\/\/nathanmarz.com\/blog\/clojures-missing-piece.html","content":"<p>After two years of being open source and a lot of development, <a href=\"https:\/\/github.com\/nathanmarz\/specter\">Specter<\/a> has reached 1.0. In this post I'll explain why I call Specter \"Clojure's missing piece\" and why it should be used in most Clojure and ClojureScript programs.<\/p><h3>Clojure's glaring hole<\/h3><p>I've been using Clojure as my primary language for almost 7 years. With it I created <a href=\"http:\/\/storm.apache.org\">Apache Storm<\/a> which became a large, important, international project in the Big Data world. I also used Clojure to do a successful startup called BackType that was <a href=\"https:\/\/techcrunch.com\/2011\/07\/05\/twitter-acquires-social-analytics-startup-backtype\/\">acquired by Twitter<\/a>. I consider Clojure to be by far the best general-purpose language and have long evangelized it.<\/p><p>As much as I enjoyed programming in Clojure, I came to realize that beyond basic use cases Clojure fails to provide an elegant API for immutable programming. This deficiency in Clojure manifested itself in my projects as unreadable code, complexity, lots of bugs, and bad performance.<\/p><p>To demonstrate this deficiency, I'll start with a few examples where Clojure does immutable programming well:<\/p><pre class=\"brush: clojure;\">\n(assoc {:a 1 :b 2} :a 3) ; =&gt; {:a 3, :b 2}\n(dissoc {:a 1 :b 2} :a) ; =&gt; {:b 2}\n(conj [1 2 3] 4) ; =&gt; [1 2 3 4]\n(pop [1 2 3]) ; =&gt; [1 2]\n(conj #{1 2 3} 4) ; =&gt; #{1 2 3 4}\n<\/pre><p>This is how immutable programming should be. You perform an operation on your data and get back a new data structure with your changes represented. The details of how immutability is implemented, like structural sharing, copying, and creating new instances of the same type, are encapsulated.<\/p><p>Unfortunately, you only get this elegance when doing basic operations on basic data structures. The elegance evaporates for a lot of common operations, especially those involving compound data structures. And unfortunately, your problem domain frequently dictates that you must use compound data structures. In my own work, I use a lot of directed acyclic graphs. Nodes can be a variety of record types, with fields ranging the whole gamut of data structures. Some nodes can have embedded within them another DAG. Manipulating a data structure like this with vanilla Clojure was a non-starter and I was forced to come up with Specter to handle it.<\/p><p>As an example of the lack of elegance of vanilla Clojure, consider the task of namespace qualifying the keys of a map. The obvious way to do this with vanilla Clojure is as follows:<\/p><pre class=\"brush: clojure;\">\n(defn ns-qualify-keys [m]\n  (into {}\n    (map (fn [[k v]] [(keyword (str *ns*) (name k)) v]))\n    m\n    ))\n<\/pre><p>This function has two problems: it's slow, and it's wrong. Performance-wise, this function runs upwards of <a href=\"https:\/\/gist.github.com\/nathanmarz\/8e1809bbd3a7632311c2b45ba02eb880\">3x as slow<\/a> as an optimal implementation. It turns out the best way to iterate over the input map and construct the output map completely changes depending on the type of the map. Writing optimal code for this task requires esoteric knowledge of the internals of Clojure<sup class=\"footnote\"><a href=\"#fn1\">1<\/a><\/sup>. For something as fundamental as manipulating program data, the idiomatic way should also be the most performant way. This point alone is enough to warrant a new approach to immutable programming.<\/p><p>However, there's an even more important point: this code is wrong. If you give this function a sorted map, it will return an unsorted map as the result. That's a totally unintended effect to what should just be a transformation on each map key. Whereas the goal of the function is to change some subvalues of the map (the map keys), to write the function correctly you have to reconstruct everything around those subvalues as well. This is a bunch of extra \"stuff\" you have to do and it's easy to make errors in the reconstruction process. So while you can fix the function by replacing <code>{}<\/code> with <code>(empty m)<\/code>, the very fact you have to do so is the problem.<\/p><p>It's not just the map you have to reconstruct in this function. It's also the keywords. The only thing you want to modify is the namespace of each keyword, but to do so you have to construct a whole new keyword with everything else exactly the same. Hence the need to call <code>name<\/code> and <code>keyword<\/code> in the anonymous function. Reconstructing the map and keywords is intertwined with the desire to change the namespaces. This is the very definition of complexity as defined by Rich Hickey in his great talk <a href=\"https:\/\/www.infoq.com\/presentations\/Simple-Made-Easy\">Simple Made Easy<\/a>. I can't even begin to describe how often I wanted to tear my hair out debugging issues that were ultimately caused by a stupid error during reconstruction, like using <code>map<\/code> instead of <code>mapv<\/code>.<\/p><p>The way to improve this code is to factor out reusable functions <code>map-keys<\/code> and <code>update-namespace<\/code> which each concern themselves with doing their one task well. You can then compose them to implement the original function with proper separation of concerns (and optimal performance):<\/p><pre class=\"brush: clojure;\">\n(defn ns-qualify-keys [m]\n  (map-keys\n    (fn [k]\n      (update-namespace k (fn [_] (str *ns*)))\n      )\n    m\n    ))\n<\/pre><p>This code is significantly improved. Gone is the need to reconstruct the map and keywords \u2013 the code concerns itself only with navigating to the namespaces and changing them. That prior intertwinement between updating the namespaces and reconstructing the surrounding data structure has been teased out.<\/p><p>Unfortunately, this isn't quite enough. While having an expanded set of transformation functions like <code>map-keys<\/code> and <code>update-namespace<\/code> is the right start, composing them together using nested anonymous functions just doesn't work well. Code quickly becomes unreadable and hard to maintain. Even this code is hard to read, and it gets far worse with more levels of nesting. Consider this code which increments the even numbers in a map of vector of maps (with a hypothetical <code>map-vals<\/code> function):<\/p><pre class=\"brush: clojure;\">\n(map-vals\n  (fn [v]\n    (mapv\n      (fn [m2]\n        (map-vals\n          (fn [n] (if (even? n) (inc n) n))\n          m2\n          ))\n      v\n      ))\n  m)\n<\/pre><p>This is totally unreadable. What's needed are the transformation functions <strong>and<\/strong> an elegant way to compose them together. Clojure lacks both of these things.<\/p><p>Here's a non-exhaustive list of basic data structure transformations missing from Clojure:<\/p><p>- Transform each key of a map<br \/>- Transform each value of a map<br \/>- Transform each value of a generic sequence<br \/>- Append to a generic sequence<br \/>- Prepend to a generic sequence<br \/>- Add anywhere into a generic sequence (e.g. <code>[1 2 4 5]<\/code> =&gt; <code>[1 2 3 4 5]<\/code>)<br \/>- Remove anywhere from a generic sequence (e.g. <code>'(1 :a 2)<\/code> =&gt; <code>'(1 2)<\/code>)<br \/>- Update the name or namespace of a keyword or symbol<\/p><p><em>(Some of these cannot be implemented \"efficiently\", like prepending to a vector in O(1) time. I'm guessing this is why Clojure core does not provide the functionality. However, in reality the need to do these \"inefficient\" operations sometimes comes up and the efficiency aspect doesn't matter because it's infrequent or not a bottleneck.)<\/em><\/p><p>Some of these are quite difficult to implement optimally. And again, implementing these is not enough. What's also needed is an elegant way to compose them so that a nested data structure is just as easy (and readable) to handle as a top-level data structure.<\/p><h3>Navigation, the missing abstraction<\/h3><p>Specter provides an abstraction called \"navigation\" which completely solves the problem of doing elegant, high-performance immutable transformations on arbitrarily complicated data structures with all the details (e.g. reconstruction) encapsulated. Here's how to namespace qualify the keys of a map with Specter:<\/p><pre class=\"brush: clojure;\">\n(setval [MAP-KEYS NAMESPACE] (str *ns*) any-map)\n<\/pre><p>This code only changes the keys of the maps \u2013\u00a0the type of the map remains identical to what it was before. Additionally, the <a href=\"https:\/\/gist.github.com\/nathanmarz\/8e1809bbd3a7632311c2b45ba02eb880\">performance<\/a> is near-optimal regardless of the type of the map (that is, it's extremely difficult to write faster code).<\/p><p><code>setval<\/code> is like <code>assoc-in<\/code> on steroids. It's given a path, a value, and a piece of data to transform. The path tells it what subvalues to set to the new value. You should think of the path as navigating into the data structure, step by step. In this case, it navigates to each map key, and then for each map key navigates to its namespace. The navigators encapsulate the details of iteration and any necessary reconstruction during the transformation.<\/p><p><code>setval<\/code> is a thin wrapper over the more general operation <code>transform<\/code>. <code>transform<\/code> takes in a function to apply to each navigated subvalue. For example:<\/p><pre class=\"brush: clojure;\">\n(transform [ALL :a even?] inc [{:a 1} {:a 2 :b 1} {:a 4}])\n;; =&gt; [{:a 1} {:a 3, :b 1} {:a 5}]\n<\/pre><p>This transformation navigates to each element of the vector, then to the value of <code>:a<\/code> for every map, stays navigated at a value only if it's even, and then applies the <code>inc<\/code> function.<\/p><p>Specter has two components. The first is the core of Specter which defines the navigator abstraction (<code>defnav<\/code>) and implements a high-performance method of composing navigators together. You can read about how that works <a href=\"https:\/\/github.com\/nathanmarz\/specter\/wiki\/Specter%27s-inline-caching-implementation\">here<\/a>. Implementing this method was the most difficult code I've ever written, and the fact it's even possible is a huge testament to the power of Clojure.<\/p><p>The second component of Specter is a comprehensive set of navigators for Clojure's core data structures. These include implementations of all of Clojure's aforementioned missing transformation functions, as well as many other kinds of navigations. These navigators encapsulate the most efficient means possible of performing traversal and reconstruction, often utilizing esoteric knowledge of Clojure's internals. It's important to note that navigators (like <code>ALL<\/code>, <code>MAP-KEYS<\/code>, <code>NAMESPACE<\/code>) are first-class objects and <strong>not<\/strong> syntax, a common point of confusion to those new to Specter.<\/p><p>As discussed before, these are the two missing pieces of Clojure. By filling these holes, Specter enables all immutable transformations to be done elegantly and with near-optimal performance.<\/p><p>Initially I developed Specter for the kinds of examples already shown: transforming subvalues in a compound data structure. Then I discovered a new category of navigators which shocked me in their expressiveness: substructure navigators. They give an extraordinary amount of control over the manipulation of data and made me believe that manipulating data via the composition of navigators should be a <strong>fundamental skill for all functional programmers<\/strong>. I'll go through a few examples of substructure navigators.<\/p><p><code>srange<\/code> navigates you to a subsequence of a list or vector. That subsequence is then replaced by whatever sequence it's transformed to. Among other tasks, this navigator can be used to splice new elements into a sequence or remove elements. For example:<\/p><pre class=\"brush: clojure;\">\n(setval [:a (srange 2 4)] [] {:a [1 2 3 4 5]})\n;; =&gt; {:a [1 2 5]}\n(setval (srange 2 2) [:A :A] '(1 2 3 4 5))\n;; =&gt; (1 2 :A :A 3 4 5)\n(transform [:a (srange 1 5)] reverse {:a [1 2 3 4 5 6]})\n;; =&gt; {:a [1 5 4 3 2 6]}\n<\/pre><p><code>filterer<\/code> navigates you to a sequence of all elements matching the predicate. Transformations on that sequence are represented back in the original sequence. For example:<\/p><pre class=\"brush: clojure;\">\n(transform (filterer even?) reverse [1 2 3 4 5 6 7 8 9])\n;; =&gt; [1 8 3 6 5 4 7 2 9]\n<\/pre><p><code>subselect<\/code> navigates you to the sequence of elements matched by the given path. Like <code>filterer<\/code>, transformations on that sequence are represented back in the original locations for each value. For example:<\/p><pre class=\"brush: clojure;\">\n(transform (subselect ALL :a even?)\n  reverse\n  [{:a 1} {:a 2 :b 1} {:a 4} {:a 5} {:a 6} {:a 8}])\n;; =&gt; [{:a 1} {:a 8 :b 1} {:a 6} {:a 5} {:a 4} {:a 2}]]\n<\/pre><p>Combining substructure navigators with other navigators enables some truly beautiful code. Take a look at how to increment the last odd number in a sequence:<\/p><pre class=\"brush: clojure;\">\n(transform [(filterer odd?) LAST] inc [1 2 3 4 5 6])\n;; =&gt; [1 2 3 4 6 6]\n<\/pre><p>This is not an operation on a compound data structure, yet navigation proves to be a beautiful abstraction for this task, far more elegant than anything you could write in vanilla Clojure. I cannot emphasize this point enough. Although compound data structures are the most glaring use cases in need of Specter, Specter proves to be incredibly useful for the manipulation of non-compound data structures. Its nature as a <em>composable abstraction<\/em> lets you concisely express very sophisticated transformations. Every newly defined navigator increases the range of use cases <em>combinatorially<\/em>.<\/p><h3>Querying, another natural use case for navigation<\/h3><p>The navigation abstraction has been shown for performing transformations, but the abstraction also lends itself naturally for querying. Here's one example:<\/p><pre class=\"brush: clojure;\">\n(select [ALL :a even?] [{:a 1} {:a 2 :b 1} {:a 4}])\n;; =&gt; [2 4]\n<\/pre><p><code>select<\/code> always returns a vector of results. To select just a single value, use <code>select-any<\/code>:<\/p><pre class=\"brush: clojure;\">\n(select-any [:a :b :c] {:a {:b {:c 1}}})\n;; =&gt; 1\n<\/pre><p>This code looks effectively identical to the equivalent <code>get-in<\/code> usage, except it runs <a href=\"https:\/\/gist.github.com\/nathanmarz\/b7c612b417647db80b9eaab618ff8d83#file-results-clojure-1-7-0-clj-L6\">30% faster<\/a>. All navigators can be used for querying, as the mental model of navigation is identical for both querying and transformation.<\/p><h3>Navigation is generic and extensible<\/h3><p>As explained before, the core of Specter is the <code>defnav<\/code> operator upon which Specter's entire collection of navigators is built. Specter's extensibility is one of its most important features, as you may use data structures other than the ones Clojure provides. As I mentioned, I do a lot of work with graphs, and I just wouldn't be able to work with them in any reasonable way without my internal collection of graph navigators (e.g. subgraph, topological traversal, to a node id, to outgoing nodes, to incoming nodes, etc.).<\/p><p>Specter's navigator interface is completely generic, able to express any navigator. Let's see how it works by looking at the implementation of a simple navigator, <code>NAMESPACE<\/code>:<\/p><pre class=\"brush: clojure;\">\n(defnav ^{:doc \"Navigates to the namespace portion of the keyword or symbol\"}\n  NAMESPACE\n  []\n  (select* [this structure next-fn]\n    (next-fn (namespace structure)))\n  (transform* [this structure next-fn]\n    (let [name (name structure)\n          new-ns (next-fn (namespace structure))]\n      (cond (keyword? structure) (keyword new-ns name)\n            (symbol? structure) (symbol new-ns name)\n            :else (i\/throw-illegal \"NAMESPACE can only be used on symbols or keywords - \" structure)\n            ))))\n<\/pre><p>There are two codepaths, one for querying (<code>select*<\/code>) and one for transforming (<code>transform*<\/code>). Querying in Specter works very similar to how transducers work. It achieves great performance by avoiding the creation of any intermediate data structures. In this case, it navigates to the namespace of the value by calling <code>next-fn<\/code> on it. A navigator that navigates to multiple subvalues (like <code>ALL<\/code>) must call <code>next-fn<\/code> on each subvalue. (As an aside, <code>select*<\/code> used to work exactly like the list monad. However, materializing intermediate lists during navigation killed performance so the method was changed in 0.12.0 to avoid that problem.)<\/p><p><code>transform*<\/code> is expected to transform any subvalues using <code>next-fn<\/code> and perform any necessary operations to reconstruct the surrounding data structure. In this case, it transforms the namespace using <code>next-fn<\/code> and creates a new keyword or symbol depending on the type of the original value.<\/p><p>As you can see, the way navigators are defined is completely generic. Each codepath is a function free to do whatever it needs. It's a simple but very effective interface. Since every navigator composes with all the other navigators, the possibilities are endless. I recommend <a href=\"https:\/\/github.com\/nathanmarz\/specter\/blob\/master\/src\/clj\/com\/rpl\/specter.cljc\">browsing the source<\/a> to see how the built-in navigators are defined.<\/p><p>Besides defining a navigator using <code>defnav<\/code>, you can also define one by combining other navigators together. This is often useful when working with a specific compound data structure for your particular application. For example, suppose you're working on a Blackjack game and storing game information like this:<\/p><pre class=\"brush: clojure;\">\n{:players [{:name \"Hopper\", :funds 100, :games-played 10}\n           {:name \"Eleven\", :funds 6941, :games-played 7}\n           {:name \"Will\", :funds -12, :games-played -8}]\n :bank {:funds 9850}}\n<\/pre><p>Players are indexed by their id (e.g. index 0, index 1) but not by their name. If you wanted to easily manipulate a player by their name, you could define a navigator for it:<\/p><pre class=\"brush: clojure;\">\n(defn player-by-name [name]\n  (path :players\n        ALL\n        #(= (:name %) name)))\n;; Take $1 from Hopper\n(transform [(player-by-name \"Hopper\") :funds] dec data)\n;; Retrieve :games-played for Will\n(select-any [(player-by-name \"Will\") :games-played] data)\n<\/pre><p>You could imagine defining a similar navigator <code>player-by-id<\/code> as well. Abstracting out domain-specific navigators like this significantly reduces the cognitive load of working with your data, as you can write your code in terms of application-specific concepts instead of lower level data structure concepts.<\/p><p>Finally, it's worth mentioning one last concept for building navigators called \"dynamic navs\". These are kind of like macros, but for navigators. Dynamic navs are too much to explain for this post, but you can read about them <a href=\"https:\/\/github.com\/nathanmarz\/specter\/wiki\/Specter%27s-inline-caching-implementation\">here<\/a>. Also, the <a href=\"https:\/\/github.com\/nathanmarz\/specter\/blob\/master\/src\/clj\/com\/rpl\/specter.cljc\">source code<\/a> has many examples of them.<\/p><h3>Recursive navigation<\/h3><p>Unsurprisingly, Specter is fantastic at working with recursive data structures. For a simple example of this, suppose you had a tree represented using vectors:<\/p><pre class=\"brush: clojure;\">\n(def tree [1 [2 [[3]] 4] [[5] 6] [7] 8 [[9]]])\n<\/pre><p>You can define a navigator to all the leaves of the tree like this:<\/p><pre class=\"brush: clojure;\">\n(def TREE-VALUES\n  (recursive-path [] p\n    (if-path vector?\n      [ALL p]\n      STAY)))\n<\/pre><p><code>recursive-path<\/code> allows your path definition to refer to itself, in this case using <code>p<\/code>. This path definition leverages the <code>if-path<\/code> navigator which uses the predicate to determine which path to continue on. If currently navigated to a vector, it recurses navigation at all elements of the vector. Otherwise, it uses <code>STAY<\/code> to stop traversing and finish navigation at the current point. The effect of all this is a depth-first traversal to each leaf node.<\/p><p>You can then compose <code>TREE-VALUES<\/code> with other navigators to do all sorts of manipulations:<\/p><pre class=\"brush: clojure;\">\n;; Increment even leaves\n(transform [TREE-VALUES even?] inc tree)\n;; =&gt; [1 [3 [[3]] 5] [[5] 7] [7] 9 [[9]]]\n;; Get odd leaves\n(select [TREE-VALUES odd?] tree)\n;; =&gt; [1 3 5 7 9]\n;; Reverse order of even leaves (order based on depth-first search)\n(transform (subselect TREE-VALUES even?) reverse tree)\n;; =&gt; [1 [8 [[3]] 6] [[5] 4] [7] 2 [[9]]]\n<\/pre><p>That you can define how to get to the values you care about once and easily reuse that logic for both querying and transformation is invaluable. And, as always, the performance is near-optimal for both querying and transformation.<\/p><h3>New in Specter 1.0<\/h3><p>Specter 1.0 is a really big milestone for the project. My goal with Specter had always been to hit that sweet spot of:<\/p><p>- Provides a set of navigators for Clojure's core data structures that feels complete<br \/>- Can be used completely dynamically (e.g. pass paths as arguments, store and use paths later, etc.)<br \/>- Near-optimal performance for all use cases<\/p><p>It took a few years to figure out how to achieve all this, and after a lot of work Specter is finally there.<\/p><p>The most important part of the Specter 1.0 release is a commitment to backwards compatibility. The 0.11.0, 0.12.0, and 0.13.0 releases had major breaking changes, all of which were required to get Specter to that sweet spot. Now that Specter is there, it will be very stable moving forward.<\/p><p>Specter 1.0 also has some significant new features filling in gaps in the library. For starters, it now can remove elements from maps and sequences with high performance:<\/p><pre class=\"brush: clojure;\">\n;; Remove nil elements form a vector\n(setval [ALL nil?] NONE [1 2 nil 3 nil])\n;; =&gt; [1 2 3]\n;; Remove even values from a map\n(setval [MAP-VALS even?] NONE {:a 1 :b 2 :c 3})\n;; =&gt; {:a 1, :c 3}\n;; Remove a key from a nested map\n(setval [:a :b :c] NONE {:a {:b {:c 1}}})\n;; =&gt; {:a {:b {}}}\n<\/pre><p>Specter 1.0 extends many of the core navigators to operate on strings:<\/p><pre class=\"brush: clojure;\">\n;; Append to a string\n(setval [:a END] \"!!!\" {:a \"Hi\"})\n{:a \"Hi!!!\"}\n;; Remove middle of a string\n(setval (srange 1 3) \"\" \"abcd\")\n;; =&gt; \"ad\"\n<\/pre><p>Another major new feature is integration with Clojure's transducers via the new <code>traverse-all<\/code> operation. A lot of common transducer use cases can be expressed more elegantly with <code>traverse-all<\/code>:<\/p><pre class=\"brush: clojure;\">\n;; Using Vanilla Clojure\n(transduce\n  (comp (map :a) (mapcat identity) (filter odd?))\n  +\n  [{:a [1 2]} {:a [3]} {:a [4 5]}])\n;; =&gt; 9\n;; The same logic expressed with Specter\n(transduce\n  (traverse-all [:a ALL odd?])\n  +\n  [{:a [1 2]} {:a [3]} {:a [4 5]}])\n<\/pre><p>Specter 1.0 also adds some new navigators and has improved performance for a number of use cases. See the full release notes <a href=\"https:\/\/github.com\/nathanmarz\/specter\/blob\/master\/CHANGES.md\">here<\/a>.<\/p><h3>Conclusion<\/h3><p>I think of Specter differently than other libraries. Whereas most libraries provide functionality for a particular task, like interacting with a database, what Specter provides is fundamental to functional programming. Manipulating data structures is one of the most core things you do as a programmer, so even improving that a little bit would have a big effect on your programs. And I would argue that Specter improves that <strong>a lot<\/strong>, letting you work at a significantly higher level of abstraction while simultaneously improving performance a great deal.<\/p><p>If you have a Haskell background, I'm sure you're screaming to yourself \"Lenses! Lenses!\" I actually didn't know about lenses before I made Specter, but they are certainly very similar. I'm not on expert on Haskell, but what I do know is it explicitly distinguishes between navigators that go to one element (Lens) vs. zero or more (Traversal). I fail to see how that complication adds any sort of expressive or performance benefit, but perhaps a Haskeller out there can educate me.<\/p><p>And to answer one of the most common questions I get: no, Specter is not going to become part of Clojure core anytime soon. I'm open to the idea and think it would be a major improvement to Clojure, but the core team has <a href=\"https:\/\/groups.google.com\/forum\/#!topic\/clojure\/qN1UPMVQmaM\">no interest whatsoever<\/a>. You can, however, achieve the same effect by adding Specter as a dependency and writing <code>(use 'com.rpl.specter)<\/code>.<\/p><p>Leveraging Specter to its maximum potential requires you to change how you think about manipulating data. I've noticed many people struggle to grasp all the ways Specter can be applied, no doubt an instance of the <a href=\"http:\/\/paulgraham.com\/avg.html\">blub paradox<\/a>. Substructure navigation, recursive paths, and higher-order navigators are a few of the more advanced concepts that can seem overwhelming. I recommend starting with the most obvious use case, transformations of compound data structures, and letting your brain slowly adapt to the navigation way of thinking. That's the way I started, and the library slowly evolved from there as I saw the different ways these ideas could be leveraged. Eventually it will become part of your basic programming instinct, and you'll wonder how you ever lived without it.<\/p><p class=\"footnote\" id=\"fn1\"><sup>1<\/sup> For a small map (PersistentArrayMap), the optimal transformation uses the IMapIterable interface to iterate over the keys and values. The underlying array is created manually and the PersistentArrayMap\/createAsIfByAssoc static method is used to create the new map. For a larger map (PersistentHashMap), reduce-kv for iteration plus transients for construction prove to be the most efficient method.<\/p>","created_at":"2018-12-03T11:11:44+0000","updated_at":"2018-12-03T14:44:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"nathanmarz.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":731,"title":"Mastodon","url":"https:\/\/mastodon.at\/about","content":"<div class=\"column-0\"><div class=\"brand\"><a href=\"https:\/\/mastodon.at\/\"><img alt=\"Mastodon\" src=\"https:\/\/mastodon.at\/packs\/logo_full-96e7a97fe469f75a23a74852b2478fa3.svg\" \/><\/a><\/div><\/div><div class=\"column-1\"><div class=\"landing-page__forms\"><div class=\"brand\"><a href=\"https:\/\/mastodon.at\/\"><img alt=\"Mastodon\" src=\"https:\/\/mastodon.at\/packs\/logo_full-96e7a97fe469f75a23a74852b2478fa3.svg\" \/><\/a><\/div><form class=\"simple_form new_user\" id=\"new_user\" novalidate=\"novalidate\" action=\"\/auth\" accept-charset=\"UTF-8\" method=\"post\"><div class=\"input with_label string required user_account_username\"><div class=\"label_input\"><div class=\"label_input__wrapper\"><p>@mastodon.at<\/p><\/div><\/div><\/div>\n<p class=\"hint subtle-hint\">By clicking \"Sign up\" below you agree to follow <a href=\"https:\/\/mastodon.at\/about\/more\">the rules of the instance<\/a> and <a href=\"https:\/\/mastodon.at\/terms\">our terms of service<\/a>.<\/p>\n<\/form><p>or<\/p><a class=\"button button-alternative-2 webapp-btn\" href=\"https:\/\/mastodon.at\/auth\/sign_in\">Log in<\/a><\/div><\/div><div class=\"column-2\"><div class=\"landing-page__hero\"><img alt=\"Mastodon\" src=\"https:\/\/mastodon.at\/packs\/preview-9a17d32fc48369e8ccd910a75260e67d.jpg\" \/><\/div><div class=\"landing-page__call-to-action\" dir=\"ltr\"><div class=\"row\"><div class=\"row__mascot\"><div class=\"landing-page__mascot\"><img alt=\"\" src=\"https:\/\/mastodon.at\/packs\/elephant_ui_plane-e3f2d57c12c376e189c274cbe81af8dd.svg\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"column-4 landing-page__information\"><div class=\"landing-page__features\"><div class=\"features-list\"><div><h3>What is Mastodon?<\/h3><p>Mastodon is a social network based on open web protocols and free, open-source software. It is decentralized like e-mail.<\/p><\/div><\/div><div class=\"features-list\"><div class=\"features-list__row\"><p><\/p><h6>Built for real conversation<\/h6>\nWith 500 characters at your disposal and support for granular content and media warnings, you can express yourself the way you want to.<\/div><div class=\"features-list__row\"><p><\/p><h6>You\u2019re a person, not a product<\/h6>\nMastodon is not a commercial network. No advertising, no data mining, no walled gardens. There is no central authority.<\/div><div class=\"features-list__row\"><p><\/p><h6>Always within reach<\/h6>\nMultiple apps for iOS, Android, and other platforms thanks to a developer-friendly API ecosystem allow you to keep up with your friends anywhere.<\/div><div class=\"features-list__row\"><p><\/p><h6>A more humane approach<\/h6>\nLearning from failures of other networks, Mastodon aims to make ethical design choices to combat the misuse of social media.<\/div><\/div><p><a class=\"button button-alternative\" href=\"https:\/\/joinmastodon.org\/\">Learn more<\/a><\/p><\/div><div class=\"landing-page__footer\"><p>\n<a href=\"https:\/\/github.com\/tootsuite\/mastodon\">Source code<\/a>\n (2.6.2)\n<\/p><\/div><\/div>","created_at":"2018-11-30T03:20:26+0000","updated_at":"2018-11-30T03:20:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"mastodon.at","preview_picture":"https:\/\/mastodon.at\/packs\/preview-9a17d32fc48369e8ccd910a75260e67d.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["backpack"],"is_public":false,"id":730,"title":"Staad Laptop BackPack","url":"https:\/\/www.sfbags.com\/products\/staad-laptop-backpack","content":"<div class=\"wrapper\"><div id=\"parentHorizontalTab\" class=\"mt40\"><ul class=\"resp-tabs-list hor_1\"><li id=\"Video\"><i class=\"fa fa-youtube-play\">Video<\/i><\/li>\n      \n      <li id=\"Product_Info\">Product Info<\/li>\n      \n      <li id=\"Specifications\"><i class=\"fa fa-check\">Specifications<\/i><\/li>\n      \n      \n      <li id=\"What_Fits\"><i class=\"fa fa-question-circle\">What Fits<\/i><\/li>\n      \n      \n      \n      \n      <li id=\"Press\"><i class=\"fa fa-newspaper-o\">Press<\/i><\/li>\n      \n      <li id=\"Reviews\"><i class=\"fa fa-star-o\">Reviews<\/i><\/li>\n      \n      <li id=\"FAQs\"><i class=\"fa fa-question-circle\">FAQs<\/i><\/li>\n      \n    <\/ul><div class=\"resp-tabs-container hor_1\"><div><iframe width=\"853\" height=\"480\" src=\"https:\/\/www.youtube.com\/embed\/j_teLj5CIDk?rel=0\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><\/div><div><div class=\"rte\" itemprop=\"description\"><p><strong>Move on up.<\/strong> Upgrade from casual to a more formal work environment with the refined <strong>Staad BackPack<\/strong>. Tapered top creates a slim profile while maximizing capacity. Plushly-cushioned interior pockets for a laptop and tablet keep the main compartment organized. By request, we'll adjust the laptop pocket in the Stout size Staad to comfortably fit your 15\u201d MacBook Pro Touch Bar. Swing the Staad to one side for quick access to keys or phone in angled, zippered side pockets. Choose from tan waxed-canvas or black ballistic and a black, chocolate or grizzly leather for a striking flap and accents. Vintage WWII-era buckle is easy to use and ultra hip. Give your back the pack it deserves.<br \/>Part of the <a href=\"https:\/\/www.sfbags.com\/collections\/staad\">Staad Collection<\/a>.<\/p><p class=\"red-text\"><em>The Stout padded laptop compartment fits up to a 15-inch laptop. If you'd prefer smaller 12\/13-inch compartment, just make the request in the special instructions section when you check out.<\/em><\/p><\/div><\/div><div><div class=\"rte\"><p>All our products are crafted locally in our workshop by the most skilled sewing team in San Francisco. <\/p><p><strong>Material:<\/strong>\n<\/p><ul><li>Waxed Canvas or Ballistic Nylon body for durability<\/li>\n<li>Full-grain premium leather Flap and Handle<\/li>\n<\/ul><p><strong>Dimensions &amp;  Weight:<\/strong>\n<\/p><ul><li><strong>Slim<\/strong>: 15.5\" x 12\" x 1\" - 3\" (tapers); 2.4 lbs.<\/li>\n  \n<li><strong>Stout<\/strong>: 16\" x 14.25\" x 1.5\" - 5.5\" (tapers); 2.8 lbs. <\/li>\n  \n<\/ul><p><strong>Hardware:<\/strong> \n<\/p><ul><li>WWII-era ammunition buckle <\/li>\n<li>YKK Zippers<\/li>\n<li>Splash Proof Zippers<\/li>\n<\/ul><strong>Features:<\/strong><ul><li>Flap is double-sided full-grain leather<\/li>\n<li>Fully-padded built-in laptop\/tablet pocket<\/li>\n<li>Easy-to-see gold liner<\/li>\n<\/ul><\/div><\/div><div><div class=\"rte\">The Staad Laptop Backpack comes in two sizes each with a built-in padded laptop pocket - and  - a built-in padded tablet pocket:<p><strong>Slim<\/strong> Laptop Compartment fits devices up to <strong>13\" x 9\"<\/strong>\n<\/p><ul><li>13\" MacBook Pro with Touch Bar, Retina, Air Retina<\/li>\n<li>12\" MacBook<\/li>\n<li>Pixel Slate<\/li>\n<li>Surface Book 2 (13.5\")<\/li>\n<li>Surface Pro 6, Go<\/li>\n<li>Surface Laptop 2<\/li>\n<li>Dell XPS 13<\/li>\n<\/ul><strong>Slim<\/strong>Tablet Compartment fits devices up to<strong>9.75\" x 7\"<\/strong><ul><li>iPad Pro (11\" &amp; 10.5\")<\/li>\n<li>reMarkable (paper tablet)<\/li>\n<\/ul><br \/><strong>Stout<\/strong>Laptop Compartment fits devices up to<strong>14.25\" x 9.75\"<\/strong><ul><li>15\" MacBook Pro with Touch Bar, Retina<\/li>\n<li>Dell XPS 15<\/li>\n<li>Lenovo X1 Carbon<\/li>\n<li>Surface Book 2 (15\")<\/li>\n<li>Surface Laptop 2<\/li>\n<\/ul><strong>Stout<\/strong>Tablet Compartment fits devices up to<strong>12\" x 8.75\"<\/strong><ul><li>iPad Pro (11\" &amp; 12.9\")<\/li>\n<li>reMarkable (paper tablet)<\/li>\n<li>Surface Pro 6, Go<\/li> \n<li>Pixel Slate<\/li>\n<\/ul><\/div><\/div><div><div id=\"reviews\"><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/www.cultofmac.com\/251889\/waterfields-staad-slim-backpack-is-impeccably-designed-and-here-for-the-long-haul-review\/\" title=\"Waterfield's Staad Slim Backpack Is Impeccably Designed And Here For The Long Haul [Review]&#10;\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/cultofmac.jpg?5267\" alt=\"Waterfield's Staad Slim Backpack Is Impeccably Designed And Here For The Long Haul [Review]&#10;\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/www.cultofmac.com\/251889\/waterfields-staad-slim-backpack-is-impeccably-designed-and-here-for-the-long-haul-review\/\" title=\"Waterfield's Staad Slim Backpack Is Impeccably Designed And Here For The Long Haul [Review]&#10;\" target=\"_blank\">Waterfield's Staad Slim Backpack Is Impeccably Designed And Here For The Long Haul [Review]\n<\/a><\/h3><strong>Cult of Mac<\/strong>- October 30, 2013<p>\"If budget isn't an issue, the Staad backpack is a solid addition to anyone's tech kit. The heavy construction and custom-designed fittings seem like artifacts from an earlier, less obsolescent era while the more modern options promise a long, productive, useful lifespan.\"\n                \n                <\/p><p><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/cultofmac-excellent.png?5268\" alt=\"image\" \/><\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/macsparky.com\/blog\/2016\/5\/the-waterfield-staad-backpack\" title=\"THE WATERFIELD STAAD BACKPACK\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/macsparky-logo.png?13772073425897486857\" alt=\"{ diy['media9_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/macsparky.com\/blog\/2016\/5\/the-waterfield-staad-backpack\" title=\"THE WATERFIELD STAAD BACKPACK\" target=\"_blank\">THE WATERFIELD STAAD BACKPACK<\/a><\/h3><strong>MacSparky<\/strong>- May 31, 2016<p>\u201cI love the way this bag looks and I plan to use the hell out of it over many years into the future.\u201d\n                \n              <\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/www.zdnet.com\/staad-slim-backpack-hands-on-taking-the-gear-bag-to-a-whole-new-level-7000026201\/\" title=\"Review: Staad slim Backpack (hands on): Taking the gear bag to a whole new level\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/zdnet.png?7234\" alt=\"{ diy['media8_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/www.zdnet.com\/staad-slim-backpack-hands-on-taking-the-gear-bag-to-a-whole-new-level-7000026201\/\" title=\"Review: Staad slim Backpack (hands on): Taking the gear bag to a whole new level\" target=\"_blank\">Review: Staad slim Backpack (hands on): Taking the gear bag to a whole new level<\/a><\/h3><strong>ZDNet<\/strong>- February 14, 2014<p>\"There are bags and then there are bags that blend style with function. The Staad Backpack from Waterfield Designs is one of the latter that feels as much at home in the boardroom as it does on the street.\"\n<\/p><p><strong>Rating: 9.5 out of 10<\/strong>\n                \n              <\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/www.zdnet.com\/waterfield-designs-staad-stout-review-high-quality-comfort-and-functionality-are-worth-the-price-7000025874\/\" title=\"Review: WaterField Designs Staad Stout review: High quality, comfort, and functionality are worth the price\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/zdnet.png?7234\" alt=\"{ diy['media7_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/www.zdnet.com\/waterfield-designs-staad-stout-review-high-quality-comfort-and-functionality-are-worth-the-price-7000025874\/\" title=\"Review: WaterField Designs Staad Stout review: High quality, comfort, and functionality are worth the price\" target=\"_blank\">Review: WaterField Designs Staad Stout review: High quality, comfort, and functionality are worth the price<\/a><\/h3><strong>ZDNet<\/strong>- February 3, 2014<p>\"WaterField Designs make some of the finest mobile tech bags in the US and one of their latest models is the fantastic Staad BackPack. The Stout model is worth the $329 price and earns my highest praise.\"\n<\/p><p><strong>Contributor's rating: 9.5 out of 10<\/strong>\n                \n              <\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/www.gottabemobile.com\/2013\/11\/17\/staad-backpack-waterfield-designs-review\/\" title=\"Review: Staad Backpack from WaterField Designs Review\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/gottabemobile.png?7234\" alt=\"{ diy['media6_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/www.gottabemobile.com\/2013\/11\/17\/staad-backpack-waterfield-designs-review\/\" title=\"Review: Staad Backpack from WaterField Designs Review\" target=\"_blank\">Review: Staad Backpack from WaterField Designs Review<\/a><\/h3><strong>Gotta Be Mobile<\/strong>- November 17, 2013<p>\"WaterField Designs crafted a great bag in the Staad Backpack. It's one of the best computer bags we've tested\"\n<br \/>\"It's one of the best computer bags we've tested. Despite the higher price, we recommend it without any reservations.\"\n                \n              <\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/theawesomer.com\/staad-laptop-backpack\/269724\/\" title=\"Review: Staad Laptop BackPack\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/awesomer.jpg?7234\" alt=\"{ diy['media5_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/theawesomer.com\/staad-laptop-backpack\/269724\/\" title=\"Review: Staad Laptop BackPack\" target=\"_blank\">Review: Staad Laptop BackPack<\/a><\/h3><strong>The Awesomer<\/strong><p>\"beautifully hand crafted\"\n                \n              <\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/appadvice.com\/appnn\/2013\/10\/the-new-staad-laptop-backpack-is-perfect-for-the-techie-in-your-life\" title=\"The New Staad Laptop BackPack Is Perfect For The Techie In Your Life\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/appadvice.jpg?5266\" alt=\"{ diy['media4_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/appadvice.com\/appnn\/2013\/10\/the-new-staad-laptop-backpack-is-perfect-for-the-techie-in-your-life\" title=\"The New Staad Laptop BackPack Is Perfect For The Techie In Your Life\" target=\"_blank\">The New Staad Laptop BackPack Is Perfect For The Techie In Your Life<\/a><\/h3><strong>AppAdvice<\/strong>- October 28, 2013<p>\"How Does it Stack Up? I haven't found a bag this nice to look at, which also offers this level of protection.\"\n<br \/>Aesthetic Appeal: \u2605\u2605\u2605\u2605\u2605  <br \/>Wow Factor: \u2605\u2605\u2605\u2605\u2605  <br \/>Build Quality: \u2605\u2605\u2605\u2605\u2605  <br \/>Value: \u2605\u2605\u2605\u2605\u2606<br \/><\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/gadgetmac.com\/reviews\/waterfield-staad-laptop-backpack-review.html\" title=\"Review: WaterField Staad Laptop Backpack Review\" target=\"_blank\"><img src=\"http:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/gadgetmac.png?7234\" alt=\"{ diy['media3_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/gadgetmac.com\/reviews\/waterfield-staad-laptop-backpack-review.html\" title=\"Review: WaterField Staad Laptop Backpack Review\" target=\"_blank\">Review: WaterField Staad Laptop Backpack Review<\/a><\/h3><strong>Gadget Mac<\/strong>- January 23, 2014<p><img src=\"http:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/4ratingGadgetMacRecommend.png?7234\" alt=\"image\" \/><\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/www.imore.com\/review-waterfield-designs-staad-backpack\" title=\"Waterfield Designs Staad Backpack review\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/imore.png?5267\" alt=\"{ diy['media2_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/www.imore.com\/review-waterfield-designs-staad-backpack\" title=\"Waterfield Designs Staad Backpack review\" target=\"_blank\">Waterfield Designs Staad Backpack review<\/a><\/h3><strong>iMore<\/strong>- November 9, 2013<p>\"It's easy to find a laptop-carrying backpack these days; everyone and their brother makes one. It's a different story to find one that looks this good, dresses up nice, and is this well-made. Admittedly, you pay for that quality of construction and appearance. But if you're looking for a laptop backpack that looks as good in the boardroom as it does on the subway, the Staad fits the bill.\"\n                \n              <\/p><\/div><\/div><\/div><hr \/><div class=\"grid\"><div class=\"review\"><div class=\"grid__item large--one-quarter medium--one-quarter text-right\"><a href=\"http:\/\/www.engadget.com\/2013\/11\/01\/waterfield-designs-staad-backpack-carries-your-gear-looks-aweso\/\" title=\"Waterfield Designs Staad BackPack carries your gear, looks awesome doing it\" target=\"_blank\"><img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/files\/engadget.png?5523382429729024448\" alt=\"{ diy['media1_title'] }}\" \/><\/a><\/div><div class=\"grid__item large--three-quarters medium--three-quarters\"><h3><a href=\"http:\/\/www.engadget.com\/2013\/11\/01\/waterfield-designs-staad-backpack-carries-your-gear-looks-aweso\/\" title=\"Waterfield Designs Staad BackPack carries your gear, looks awesome doing it\" target=\"_blank\">Waterfield Designs Staad BackPack carries your gear, looks awesome doing it<\/a><\/h3><strong>Engadget<\/strong>- November 1, 2013<p>\"If you want the hands-free convenience of a backpack, but want something that will be at home in the executive suite, then pull out your wallet and buy the Waterfield Designs Staad BackPack.\"\n                \n              <\/p><\/div><\/div><\/div><hr \/><\/div><\/div><div><div class=\"rte\"><div id=\"faqs\"><h3>Will a 15\" laptop fit in the Slim Staad?<\/h3><div><p>No, but we can fit the Slim version with the larger Stout pocket to accommodate your laptop. Just leave a note in the special instructions and it will be done.<\/p><\/div><h3>Will a water bottle fit in the Slim Staad?<\/h3><div><p>It's going to be snug but it'll work. <\/p><\/div><h3>Is the backpack waterproof?<\/h3><div><p>It's water resistant not waterproof but does hold up well in the rain.<\/p><\/div><\/div><p>Have a question about this product? Contact us at <a href=\"mailto:HeyGary@SFBags.com\">HeyGary@SFBags.com<\/a>.<\/p><\/div><\/div><\/div><\/div><\/div>\n<section id=\"review-scroller-product\"><div class=\"grid\"><div class=\"grid__item large--two-thirds push--large--one-sixth medium-down--one-whole text-center\"><ul class=\"newsticker\"><li>\n    <p><i class=\"icon-quote-left\"> Staad, couldn\u2019t decide so bought two! <i class=\"icon-quote-right\"><\/i><\/i><\/p>\n    <p>- Alex L., Hawaii<\/p>\n  <\/li>\n   \n  <li>\n    <p><i class=\"icon-quote-left\"> Best backpack ever. Worth every penny! <i class=\"icon-quote-right\"><\/i><\/i><\/p>\n    <p>- Brandon H., Idaho<\/p>\n  <\/li>\n   \n  <li>\n    <p><i class=\"icon-quote-left\"> Staad backpack is elegant and functional! <i class=\"icon-quote-right\"><\/i><\/i><\/p>\n    <p>- Brian J., Oregon<\/p>\n  <\/li>\n   \n  <li>\n    <p><i class=\"icon-quote-left\"> Staad: sleek, sturdy, stylish-- worth every$! <i class=\"icon-quote-right\"><\/i><\/i><\/p>\n    <p>- Helena S., Victoria, Australia<\/p>\n  <\/li>\n   \n  <li>\n    <p><i class=\"icon-quote-left\"> My last backpack ever, love it. <i class=\"icon-quote-right\"><\/i><\/i><\/p>\n    <p>- Magnus L., WI<\/p>\n  <\/li>\n   \n  <li>\n    <p><i class=\"icon-quote-left\"> Love, Love, LOVE my Staad Backpack! <i class=\"icon-quote-right\"><\/i><\/i><\/p>\n    <p>- Stephen C., California<\/p>\n  <\/li>\n  \n<\/ul><\/div><\/div><\/section>","created_at":"2018-11-29T14:22:55+0000","updated_at":"2018-11-29T14:23:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.sfbags.com","preview_picture":"http:\/\/cdn.shopify.com\/s\/files\/1\/0281\/2796\/products\/staad-backpack-lifestyle_grande.jpg?v=1529446047","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":729,"title":"wasamasa\/eyebrowse","url":"https:\/\/github.com\/wasamasa\/eyebrowse","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/52c20e6653c2214dd68d674d14c2f75634bbe378\/68747470733a2f2f7261772e6769746875622e636f6d2f776173616d6173612f65796562726f7773652f6d61737465722f696d672f65796562726f77732e676966\"><img alt=\"https:\/\/raw.github.com\/wasamasa\/eyebrowse\/master\/img\/eyebrows.gif\" src=\"https:\/\/camo.githubusercontent.com\/52c20e6653c2214dd68d674d14c2f75634bbe378\/68747470733a2f2f7261772e6769746875622e636f6d2f776173616d6173612f65796562726f7773652f6d61737465722f696d672f65796562726f77732e676966\" data-canonical-src=\"https:\/\/raw.github.com\/wasamasa\/eyebrowse\/master\/img\/eyebrows.gif\" \/><\/a><\/p>\n<h2><a id=\"user-content-about\" class=\"anchor\" aria-hidden=\"true\" href=\"#about\"><\/a>About<\/h2>\n<p><code>eyebrowse<\/code> is a global minor mode for Emacs that allows you to\nmanage your window configurations in a simple manner, just like tiling\nwindow managers like i3wm with their workspaces do.  It displays their\ncurrent state in the modeline by default.  The behaviour is modeled\nafter <a href=\"http:\/\/ranger.nongnu.org\/\" rel=\"nofollow\">ranger<\/a>, a file manager written in\nPython.<\/p>\n<h2><a id=\"user-content-screenshot\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshot\"><\/a>Screenshot<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/0d2219cad4b260efe41bc47a1295595e6da646a9\/68747470733a2f2f7261772e6769746875622e636f6d2f776173616d6173612f65796562726f7773652f6d61737465722f696d672f7363726f742e706e67\"><img alt=\"https:\/\/raw.github.com\/wasamasa\/eyebrowse\/master\/img\/scrot.png\" src=\"https:\/\/camo.githubusercontent.com\/0d2219cad4b260efe41bc47a1295595e6da646a9\/68747470733a2f2f7261772e6769746875622e636f6d2f776173616d6173612f65796562726f7773652f6d61737465722f696d672f7363726f742e706e67\" data-canonical-src=\"https:\/\/raw.github.com\/wasamasa\/eyebrowse\/master\/img\/scrot.png\" \/><\/a><\/p>\n<p>See the lighter and the modeline indicator at the right side of the\nbottom modeline?  That's what you get to see after enabling eyebrowse.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>Install via <code>package.el<\/code> from the <a href=\"https:\/\/marmalade-repo.org\/\" rel=\"nofollow\">Marmalade<\/a> or <a href=\"http:\/\/melpa.org\/\" rel=\"nofollow\">MELPA (stable)<\/a> repository by setting them up if you haven't\nalready and executing <code>M-x package-install RET eyebrowse RET<\/code>.<\/p>\n<h2><a id=\"user-content-quick-tutorial\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-tutorial\"><\/a>Quick Tutorial<\/h2>\n<p>Use <code>M-x eyebrowse-mode<\/code> to enable <code>eyebrowse<\/code> interactively.  If\nyou want to enable it automatically on startup, add <code>(eyebrowse-mode\nt)<\/code> to your init file (either <code>~\/.emacs<\/code> or\n<code>~\/.emacs.d\/init.el<\/code>).<\/p>\n<p>You start with your current window config on slot 1.  Once you hit\n<code>C-c C-w 2<\/code>, you will see the modeline indicator appearing and\nshowing slot 1 and 2 with slot 2 slightly emphasized.  Slot 1 has been\nsaved automatically for you and contains your last window config.  Do\nsomething meaningful like a window split, then hit <code>C-c C-w 1<\/code>.  The\nwindow config on slot 2 is saved and the window config from slot 1 is\nloaded.  Try switching back and forth between them with <code>C-c C-w '<\/code>\nto get a feeling for how subsequent window manipulations are handled.<\/p>\n<p>To make keeping track of workspaces easier, a tagging feature was\nadded.  Use <code>C-c C-w ,<\/code> to set a tag for the current window config,\nit will both appear in the modeline indicator and when using <code>M-x\neyebrowse-switch-to-window-config<\/code>.  Setting the tag to an empty\nvalue will undo this change.<\/p>\n<h2><a id=\"user-content-key-bindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#key-bindings\"><\/a>Key bindings<\/h2>\n<p>The default key bindings are:<\/p>\n<table><thead valign=\"bottom\"><tr><th>Key bind<\/th>\n<th>Function<\/th>\n<\/tr><\/thead><tbody valign=\"top\"><tr><td><code>C-c C-w &lt;<\/code><\/td>\n<td>Switch to previous window config<\/td>\n<\/tr><tr><td><code>C-c C-w &gt;<\/code><\/td>\n<td>Switch to next window config<\/td>\n<\/tr><tr><td><code>C-c C-w '<\/code><\/td>\n<td>Switch to last window config<\/td>\n<\/tr><tr><td><code>C-c C-w \"<\/code><\/td>\n<td>Close current window config<\/td>\n<\/tr><tr><td><code>C-c C-w ,<\/code><\/td>\n<td>Rename current window config<\/td>\n<\/tr><tr><td><code>C-c C-w 0<\/code><\/td>\n<td>Switch to window config <code>0<\/code><\/td>\n<\/tr><tr><td>...<\/td>\n<td>...<\/td>\n<\/tr><tr><td><code>C-c C-w 9<\/code><\/td>\n<td>Switch to window config <code>9<\/code><\/td>\n<\/tr><\/tbody><\/table>\n<h2><a id=\"user-content-further-customization\" class=\"anchor\" aria-hidden=\"true\" href=\"#further-customization\"><\/a>Further Customization<\/h2>\n<p>Use <code>M-x customize-group RET eyebrowse<\/code> for a list of customizable\noptions.  The more interesting ones would be\n<code>eyebrowse-wrap-around<\/code> and <code>eyebrowse-switch-back-and-forth<\/code>\nwhich affect both wrap around and lazier switching.  It is also\npossible to change the behaviour of creation of new workspaces by\ncustomizing <code>eyebrowse-new-workspace<\/code>.  By default the last one is\nsimply cloned, setting it to <code>t<\/code> will start out with as empty of a\nslate as possible (by just displaying a single window with the scratch\nbuffer in it).<\/p>\n<p>The prefix for each binding defaults to <code>C-c C-w<\/code>, but you can change\nit to something else by customizing <code>eyebrowse-keymap-prefix<\/code>.  If\nyou want to change it in your init file, insert the customization\nbefore loading <code>eyebrowse.el<\/code>.  This includes enabling it with an\nauto-loaded function, such as <code>eyebrowse-setup-evil-keys<\/code>,\n<code>eyebrowse-setup-opinionated-keys<\/code> or <code>eyebrowse-mode<\/code>.<\/p>\n<p>If you're not happy with the default keybindings, a riskier set can be\nenabled additionally either by executing <code>M-:\n(eyebrowse-setup-opinionated-keys)<\/code> interactively or inserting\n<code>(eyebrowse-setup-opinionated-keys)<\/code> in your init file.  If the\nfunction detects the <a href=\"https:\/\/bitbucket.org\/lyro\/evil\/wiki\/Home\" rel=\"nofollow\">evil<\/a> package, it\nwill enable extra key bindings for it as well.<\/p>\n<p>The extra key bindings are:<\/p>\n<table><thead valign=\"bottom\"><tr><th>Key bind<\/th>\n<th>Function<\/th>\n<\/tr><\/thead><tbody valign=\"top\"><tr><td><code>C-&lt;<\/code>, <code>gT<\/code><\/td>\n<td>Switch to previous window config<\/td>\n<\/tr><tr><td><code>C-&gt;<\/code>, <code>gt<\/code><\/td>\n<td>Switch to next window config<\/td>\n<\/tr><tr><td><code>C-'<\/code>, <code>zx<\/code><\/td>\n<td>Switch to last window config<\/td>\n<\/tr><tr><td><code>C-\"<\/code>, <code>gc<\/code><\/td>\n<td>Close current window config<\/td>\n<\/tr><tr><td><code>M-0<\/code><\/td>\n<td>Switch to window config <code>0<\/code><\/td>\n<\/tr><tr><td>...<\/td>\n<td>...<\/td>\n<\/tr><tr><td><code>M-9<\/code><\/td>\n<td>Switch to window config <code>9<\/code><\/td>\n<\/tr><\/tbody><\/table>\n<h2><a id=\"user-content-internals\" class=\"anchor\" aria-hidden=\"true\" href=\"#internals\"><\/a>Internals<\/h2>\n<p>This mode basically wraps what <code>C-x r w<\/code> and <code>C-x r j<\/code> would do,\nbut takes care of automatically saving and loading to a separate data\nstructure for you and does it in a slightly different manner (see\n<code>window-state-put<\/code> and <code>window-state-get<\/code> for more details) to\nallow for features like persistency in combination with <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Saving-Emacs-Sessions.html#Saving-Emacs-Sessions\" rel=\"nofollow\">desktop.el<\/a>.<\/p>\n<h2><a id=\"user-content-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#notes\"><\/a>Notes<\/h2>\n<p>The <code>window-state-put<\/code> and <code>window-state-get<\/code> functions do not\nsave all window parameters.  If you use features like side windows\nthat store the window parameters <code>window-side<\/code> and <code>window-slot<\/code>,\nyou will need to customize <code>window-persistent-parameters<\/code> for them\nto be saved as well:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'window-persistent-parameters '(window-side . writable))\n(add-to-list 'window-persistent-parameters '(window-slot . writable))<\/pre><\/div>\n<p>See <a href=\"https:\/\/github.com\/wasamasa\/eyebrowse\/issues\/52\">#52<\/a> for\nfurther discussion.<\/p>\n<h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>If you find bugs, have suggestions or any other problems, feel free to\nreport an issue on the issue tracker or hit me up on IRC, I'm always on\n<code>#emacs<\/code>.  Patches are welcome, too, just fork, work on a separate\nbranch and open a pull request with it.<\/p>\n<h2><a id=\"user-content-alternatives\" class=\"anchor\" aria-hidden=\"true\" href=\"#alternatives\"><\/a>Alternatives<\/h2>\n<p>The two most popular window configuration packages are <a href=\"https:\/\/github.com\/shosti\/elscreen\">elscreen<\/a> and <a href=\"https:\/\/github.com\/emacsmirror\/escreen\">escreen<\/a>.  Both are fairly old and\nhave their share of bugs.  The closest package I've found so far to\neyebrowse with workspace-specific buffers would be <a href=\"https:\/\/github.com\/nex3\/perspective-el\">perspective<\/a>.  <a href=\"https:\/\/github.com\/ilohmar\/wconf\">wconf<\/a> is a minimal alternative with\nhalf the lines of code (and features).  To have fancy features such\nas morphing, try <a href=\"https:\/\/github.com\/tlh\/workgroups.el\">workgroups<\/a>\nor <a href=\"https:\/\/github.com\/pashinin\/workgroups2\">workgroups2<\/a>.<\/p>\n<h2><a id=\"user-content-name\" class=\"anchor\" aria-hidden=\"true\" href=\"#name\"><\/a>Name<\/h2>\n<p>Actually, I wanted to name this mode \"eyebrows\" for no real reason,\nbut then a silly typo happened.  The typo stuck.  So did the new name.<\/p>\n<\/article>","created_at":"2018-11-29T14:09:22+0000","updated_at":"2018-11-29T14:09:22+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/3588994?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","evil-mode"],"is_public":false,"id":728,"title":"noctuid\/evil-guide","url":"https:\/\/github.com\/noctuid\/evil-guide","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<ul><li><a href=\"#emacsevil-for-vim-users\">Emacs\/Evil for Vim Users<\/a><\/li>\n  <li><a href=\"#bare-minimum-emacs-knowledge\">Bare Minimum Emacs Knowledge<\/a>\n    <ul><li><a href=\"#terminology\">Terminology<\/a>\n        <ul><li><a href=\"#cutting-and-pasting\">Cutting and Pasting<\/a><\/li>\n          <li><a href=\"#buffer-window-frame-etc\">Buffer, Window, Frame, etc.<\/a><\/li>\n          <li><a href=\"#point-and-mark\">Point and Mark<\/a><\/li>\n          <li><a href=\"#minibuffer\">Minibuffer<\/a><\/li>\n          <li><a href=\"#modes\">Modes<\/a><\/li>\n          <li><a href=\"#hooks\">Hooks<\/a><\/li>\n          <li><a href=\"#commands-and-functions\">Commands and Functions<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#key-notation\">Key Notation<\/a><\/li>\n      <li><a href=\"#default-keybindings-and-getting-help\">Default Keybindings and Getting Help<\/a><\/li>\n      <li><a href=\"#quoting\">Quoting<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#other-resources\">Other Resources<\/a><\/li>\n  <li><a href=\"#settings-and-hooks\">Settings and Hooks<\/a><\/li>\n  <li><a href=\"#keybindings-and-states\">Keybindings and States<\/a>\n    <ul><li><a href=\"#keybindings-in-emacs\">Keybindings in Emacs<\/a><\/li>\n      <li><a href=\"#keymap-precedence\">Keymap Precedence<\/a><\/li>\n      <li><a href=\"#global-keybindings-and-evil-states\">Global Keybindings and Evil States<\/a><\/li>\n      <li><a href=\"#unbinding-a-key\">Unbinding a Key<\/a><\/li>\n      <li><a href=\"#leader-key\">Leader Key<\/a><\/li>\n      <li><a href=\"#mode-specific-keybindings\">Mode Specific Keybindings<\/a><\/li>\n      <li><a href=\"#buffer-local-keybindings\">Buffer Local Keybindings<\/a><\/li>\n      <li><a href=\"#binding-keys-to-keys-keyboard-macros\">Binding Keys to Keys (Keyboard Macros)<\/a><\/li>\n      <li><a href=\"#mapping-under-keys-that-arent-prefix-keys\">Mapping Under Keys That Aren\u2019t Prefix Keys<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#advice\">Advice<\/a><\/li>\n  <li><a href=\"#further-integrating-evil-and-emacs\">Further Integrating Evil and Emacs<\/a>\n    <ul><li><a href=\"#what-overrides-evil\">What Overrides Evil?<\/a><\/li>\n      <li><a href=\"#evils-tools\">Evil\u2019s Tools<\/a><\/li>\n      <li><a href=\"#preventing-certain-keys-from-being-overridden\">Preventing Certain Keys From Being Overridden<\/a>\n        <ul><li><a href=\"#prevent-text-property-maps-from-overriding-evil\">Prevent Text Property Maps from Overriding Evil<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#use-some-emacs-keybindings\">Use Some Emacs Keybindings<\/a>\n        <ul><li><a href=\"#switching-between-evil-and-emacs\">Switching Between Evil and Emacs<\/a><\/li>\n          <li><a href=\"#using-emacs-keybindings-in-normal-state\">Using Emacs Keybindings in Normal State<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#use-evil-everywhere\">Use Evil Everywhere<\/a>\n        <ul><li><a href=\"#make-evil-normal-state-the-initial-state-always\">Make Evil Normal State the Initial State Always<\/a><\/li>\n          <li><a href=\"#undoprevent-overridingintercept-maps\">Undo\/Prevent Overriding\/Intercept Maps<\/a><\/li>\n          <li><a href=\"#modal-minibuffer\">Modal Minibuffer<\/a><\/li>\n          <li><a href=\"#example-integration-with-pdf-tools\">Example: Integration with Pdf Tools<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#command-properties\">Command Properties<\/a>\n        <ul><li><a href=\"#repeat\">:repeat<\/a><\/li>\n          <li><a href=\"#jump\">:jump<\/a><\/li>\n          <li><a href=\"#type\">:type<\/a><\/li>\n          <li><a href=\"#move-point\">:move-point<\/a><\/li>\n          <li><a href=\"#keep-visual\">:keep-visual<\/a><\/li>\n          <li><a href=\"#suppress-operator\">:suppress-operator<\/a><\/li>\n          <li><a href=\"#motion\">:motion<\/a><\/li>\n          <li><a href=\"#declaring-a-motion\">Declaring a Motion<\/a><\/li>\n        <\/ul><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#other-evil-tips\">Other Evil Tips<\/a>\n    <ul><li><a href=\"#state-indicator\">State Indicator<\/a><\/li>\n      <li><a href=\"#user-created-states\">User Created States<\/a><\/li>\n      <li><a href=\"#using-escape-to-exit-the-minibuffer\">Using Escape to Exit the Minibuffer<\/a><\/li>\n      <li><a href=\"#ex-command-definition\">Ex Command Definition<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#autocommand-equivalents-unfinished\">Autocommand Equivalents (unfinished)<\/a><\/li>\n  <li><a href=\"#other-emacs-tips\">Other Emacs Tips<\/a>\n    <ul><li><a href=\"#packagesplugins\">Packages\/Plugins<\/a><\/li>\n      <li><a href=\"#debugging-your-init-file\">Debugging Your Init File<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#faq\">FAQ<\/a>\n    <ul><li><a href=\"#is-emacs-slow-to-start\">Is emacs slow to start?<\/a><\/li>\n      <li><a href=\"#how-do-i-improve-emacs-performance\">How do I improve emacs\u2019 performance?<\/a><\/li>\n      <li><a href=\"#does-emacs-have-vim-like-tabs-distinct-window-configurations\">Does emacs have vim-like tabs (distinct window configurations)?<\/a><\/li>\n      <li><a href=\"#whats-the-equivalent-of-nnoremap-y-y\">What\u2019s the equivalent of <code>nnoremap Y y$<\/code>?<\/a><\/li>\n      <li><a href=\"#whats-the-equivalent-of-nnoremap-n-nzz\">What\u2019s the equivalent of <code>nnoremap n nzz<\/code>?<\/a><\/li>\n      <li><a href=\"#whats-the-equivalent-of-inoremap-jk-escape\">What\u2019s the equivalent of <code>inoremap jk &lt;escape&gt;<\/code>?<\/a><\/li>\n      <li><a href=\"#whats-the-equivalent-of-nnoremap-cw-ciw\">What\u2019s the equivalent of <code>nnoremap cw ciw<\/code>?<\/a><\/li>\n      <li><a href=\"#whats-the-equivalent-of-vim-unimpaireds-cow\">What\u2019s the equivalent of vim-unimpaired\u2019s <code>cow<\/code>?<\/a><\/li>\n      <li><a href=\"#how-do-i-copy-and-paste-tofrom-the-clipboard-in-terminal-emacs\">How do I copy and paste to\/from the clipboard in terminal emacs?<\/a><\/li>\n      <li><a href=\"#can-i-have-better-completion-in-the-ex-command-line\">Can I have better completion in the ex command line?<\/a><\/li>\n      <li><a href=\"#how-do-i-prevent-parentheses-becoming-unbalanced-in-my-init-file\">How do I prevent parentheses becoming unbalanced in my init file?<\/a><\/li>\n      <li><a href=\"#how-can-i-have-relative-line-numbers\">How can I have relative line numbers?<\/a><\/li>\n      <li><a href=\"#does-emacs-have-support-for-folding\">Does emacs have support for folding?<\/a><\/li>\n      <li><a href=\"#why-dont-keys-defined-with-evil-define-key-work-immediately\">Why don\u2019t keys defined with <code>evil-define-key<\/code> work (immediately)?<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#plan-to-add\">Plan to Add<\/a><\/li>\n<\/ul>\n<p>This is not meant to be a comprehensive introductory guide to emacs or a comparison between vim and emacs. It is meant to focus specifically on evil and address questions vim users might have when first using emacs with evil. Where there are already detailed, informative resources on other related topics, I will include references instead of re-explaining things myself. This aims to be a general guide to teach the reader about how to use and configure evil as opposed to a collection of configuration snippets.<\/p>\n<p>If you have any suggestions for questions or problems it would be useful to discuss, feel free to make an issue or pull request.<\/p>\n<h2><a id=\"user-content-terminology\" class=\"anchor\" aria-hidden=\"true\" href=\"#terminology\"><\/a>Terminology<\/h2>\n<p>For a more complete list of terminology, see the emacs manual\u2019s <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Glossary.html\" rel=\"nofollow\">glossary<\/a><\/p>\n<h3><a id=\"user-content-cutting-and-pasting\" class=\"anchor\" aria-hidden=\"true\" href=\"#cutting-and-pasting\"><\/a>Cutting and Pasting<\/h3>\n<p>In emacs, cutting is called \u201ckilling.\u201d Pasting is called \u201cyanking.\u201d This is probably the most confusing terminology difference between vim and emacs. Just remember that emacs packages that talk about \u201cyanking\u201d are talking about pasting, whereas evil-related packages will use \u201cyanking\u201d to mean the same thing as in vim (copying).<\/p>\n<h3><a id=\"user-content-buffer-window-frame-etc\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer-window-frame-etc\"><\/a>Buffer, Window, Frame, etc.<\/h3>\n<p>Buffers and windows basically mean the same thing in emacs as they do in vim. A window in emacs displays a buffer, and a \u201cframe\u201d can hold multiple windows (some people refer to them as \u201csplits\u201d). An emacs frame is basically a system-level window. Emacs does not have vim-like tabs builtin, but there are various plugins for storing multiple window configurations in the same frame (see <a href=\"#does-emacs-have-vim-like-tabs-distinct-window-configurations\">this part of the faq<\/a>). Also <a href=\"http:\/\/www.braveclojure.com\/basic-emacs\/\" rel=\"nofollow\">this<\/a> excellent article has some screenshots to give you visual explanation.<\/p>\n<h3><a id=\"user-content-point-and-mark\" class=\"anchor\" aria-hidden=\"true\" href=\"#point-and-mark\"><\/a>Point and Mark<\/h3>\n<p>The point refers to the cursor. The mark refers to the other side of a selected region (the \u201cactive region\u201d).<\/p>\n<h3><a id=\"user-content-minibuffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#minibuffer\"><\/a>Minibuffer<\/h3>\n<p>The minibuffer is located at the bottom of emacs. It is used for the evil ex command line among other things. This is also the location of the \u201cecho area\u201d where non-popup messages are be displayed (e.g. <code>(message \"message\")<\/code>).<\/p>\n<p>For more information on the minibuffer, see the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Minibuffer.html\" rel=\"nofollow\">corresponding section<\/a> in the emacs manual.<\/p>\n<h3><a id=\"user-content-modes\" class=\"anchor\" aria-hidden=\"true\" href=\"#modes\"><\/a>Modes<\/h3>\n<p>In emacs, the word \u201cmode\u201d is already taken, so evil refers to vim\u2019s modes as \u201cstates\u201d (e.g. \u201cnormal state\u201d). In emacs, there are major modes and minor modes. Each buffer usually only has one major mode, which is comparable to the filetype in vim. A buffer can have multiple minor modes that can be toggled on and off. An example is <code>flyspell-mode<\/code>, which enables spell checking. Modes have their own keymaps that apply only when the mode is active.<\/p>\n<h3><a id=\"user-content-hooks\" class=\"anchor\" aria-hidden=\"true\" href=\"#hooks\"><\/a>Hooks<\/h3>\n<p>A hook is similar to <code>autocmd<\/code> in vim.<\/p>\n<h3><a id=\"user-content-commands-and-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#commands-and-functions\"><\/a>Commands and Functions<\/h3>\n<p>In emacs, commands are functions that can be bound to a key (interactive) or run with <code>M-x<\/code> (<code>&lt;a-x&gt;<\/code>). Most commands can also be run from evil\u2019s ex command line with <code>:command-name&lt;cr&gt;<\/code>. An exception is commands that have numbers in them such as <code>w3m<\/code> or <code>mu4e<\/code>. I will be referring to commands as <code>command-name<\/code>.<\/p>\n<p>If you want to evaluate a function, you can use <code>M-:<\/code> (<code>&lt;a-:&gt;<\/code>) to evaluate an expression in the minibuffer. You can also run elisp by using the <code>eval-...<\/code> functions (e.g. <code>eval-defun<\/code>) in an emacs lisp mode buffer or by using <code>ielm<\/code> or <code>eshell<\/code>.<\/p>\n<h2><a id=\"user-content-key-notation\" class=\"anchor\" aria-hidden=\"true\" href=\"#key-notation\"><\/a>Key Notation<\/h2>\n<p>In vim, Space followed by Control+a would be written as <code>&lt;space&gt;&lt;c+a&gt;<\/code>. In emacs, it would be written as <code>SPC C-a<\/code>. See the Emacs Wiki <a href=\"https:\/\/www.emacswiki.org\/emacs\/EmacsKeyNotation\" rel=\"nofollow\">entry on key notation<\/a> for more information.<\/p>\n<h2><a id=\"user-content-default-keybindings-and-getting-help\" class=\"anchor\" aria-hidden=\"true\" href=\"#default-keybindings-and-getting-help\"><\/a>Default Keybindings and Getting Help<\/h2>\n<p>Some people prefer to learn emacs\u2019 keybindings first without using evil. I never did, but Sacha Chua\u2019s <a href=\"http:\/\/sachachua.com\/blog\/2013\/05\/how-to-learn-emacs-a-hand-drawn-one-pager-for-beginners\/\" rel=\"nofollow\">visual guide<\/a> for emacs beginners might be a good resource for those who want to. <a href=\"https:\/\/tuhdo.github.io\/\" rel=\"nofollow\">Tuhdo\u2019s guides<\/a> are also very good.<\/p>\n<p>I\u2019ve been able to get by without almost ever using emacs\u2019 default keybindings. The exceptions for me are <code>C-g<\/code> and <code>C-h<\/code>. Even if you don\u2019t plan on learning emacs\u2019 keybindings in full, I recommend learning these when starting out.<\/p>\n<p><code>C-g<\/code> is bound to <code>keyboard-quit<\/code> (or an equivalent) by default. You use it, for example, to exit the minibuffer (e.g. if you type <code>M-:<\/code> which is bound to <code>eval-expression<\/code> and want to exit). You can replace <code>C-g<\/code> with <code>&lt;escape&gt;<\/code> for most cases (see <a href=\"#using-escape-to-exit-the-minibuffer\">Using Escape to Exit the Minibuffer<\/a>), but it is still useful to know about it. If emacs freezes from a long-running command or infinite loop, you can use <code>C-g<\/code> to cancel it.<\/p>\n<p><code>C-h<\/code> is a prefix key for getting help. Since emacs has a different help system, <code>:help<\/code> will not work the same as in vim. If you want to find out about a variable, you can use <code>C-h v<\/code> (<code>describe-variable<\/code>). To find out what a key is bound to, you can use <code>C-h k<\/code> to pop open a help buffer with information about the command. For example, you can find out that <code>C-h k<\/code> is bound to <code>describe-key<\/code> by pressing <code>C-h k C-h k<\/code>. Knowing about <code>C-h k<\/code> can be useful, for example, if you want to find out what a nested key is bound to. By \u201cnested key\u201d, I mean that you can type <code>d C-h k i w<\/code> (<code>d&lt;c-h&gt;kiw<\/code>) to find out that <code>iw<\/code> here is bound to <code>evil-inner-word<\/code>.<\/p>\n<p>To make things more friendly, you can use something like <code>ivy-mode<\/code> from <a href=\"https:\/\/github.com\/abo-abo\/swiper\">ivy<\/a> or <code>helm-mode<\/code> from <a href=\"https:\/\/github.com\/emacs-helm\/helm\">helm<\/a> to allow you to quickly narrow your selection. Helm also provides a <code>helm-apropos<\/code> command that will allow you to search commands, functions, and variables all at once (as well as faces).<\/p>\n<p>Another useful package is <a href=\"https:\/\/github.com\/purcell\/elisp-slime-nav\">elisp-slime-nav<\/a> which provides commands that allow you to jump to the definition or corresponding help page for emacs lisp symbols.<\/p>\n<p>For more information on getting help, see the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Help.html\" rel=\"nofollow\">corresponding section from the emacs manual<\/a>.<\/p>\n<h2><a id=\"user-content-quoting\" class=\"anchor\" aria-hidden=\"true\" href=\"#quoting\"><\/a>Quoting<\/h2>\n<p>Quoting is used to prevent evaluation of a symbol or a list. Quoting is done with <code>(quote ...)<\/code> or by prefixing the symbol or list with a single quote\/apostrophe. When using a function as an argument, you use a sharp quote (equivalent to <code>(function ...)<\/code>). For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(+ 1 3 1)\n;; =&gt; 5\n(apply #'+ '(1 3 1))\n;; =&gt; 5<\/pre><\/div>\n<p>In this example, the <code>+<\/code> function is sharp quoted so that it is not treated as a variable. The list of arguments to pass to the <code>+<\/code> function is quoted so that it is treated as a literal list. Otherwise, <code>(1 3 1)<\/code> would be treated as a function call to <code>1<\/code>. Note that <code>(quote (1 3 1))<\/code> is not the same as <code>(list 1 3 1)<\/code>. Either works in this case, but the latter creates a fresh list.<\/p>\n<p>Here is what will happen if you did not quote the arguments:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(apply + '(1 3 1))\n;; =&gt; Symbol's value as a variable is void: +\n;; if you actually want to store a function name in a variable:\n(setq my-plus-func #'+)\n(apply my-plus-func '(1 3 1))\n;; =&gt; 5\n(apply #'+ (1 3 1))\n;; =&gt; Invalid function: 1\n;; if you wanted to store the argument list in a variable:\n(setq my-arg-list '(1 3 1))\n(apply #'+ my-arg-list)\n;; =&gt; 5<\/pre><\/div>\n<p>This can be confusing to a beginner when setting options or using functions. To simplify things, if you don\u2019t want a function argument to be treated as a variable, you must quote it since functions evaluate their arguments. Note that this applies to <i>symbols<\/i> and not <i>literals<\/i> (i.e. you do not need to quote strings, numbers, etc).<\/p>\n<p>There are some exceptions to this rule. For example, <code>nil<\/code> and <code>t<\/code> do not need to be quoted since they evaluate to themselves. Some macros do not require symbols to be quoted; the most common examples would probably be <code>defun<\/code> and <code>setq<\/code>. For convenience, the name of the function being defined or variable being set does not need to be quoted:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun hello-world ()\n  (message \"Hello world\"))\n(setq my-var t)<\/pre><\/div>\n<p>For more information, see the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Quoting.html\" rel=\"nofollow\">corresponding section<\/a> in the emacs manual.<\/p>\n<p>In addition to the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/\" rel=\"nofollow\">emacs manual<\/a> and <a href=\"https:\/\/tuhdo.github.io\/\" rel=\"nofollow\">Tuhdo\u2019s emacs mini manual<\/a> for general emacs information, there is also the evil manual for specific evil information. It\u2019s very short, and this guide goes into more depth about a lot of things mentioned (e.g. <code>evil-define-key<\/code>). It might be useful for reading about some of the basic settings (though it leaves most settings out). It can be read from emacs with <code>M-x info RET<\/code> or simply <code>C-h i<\/code>, searching for evil, and following the link. If you plan on writing motions, operators, and text objects, you may want to read those sections under \u201cMacros.\u201d<\/p>\n<p>Emacs is configured and extended in emacs lisp, so if you want to learn more about emacs lisp at some point, you may want to read <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/eintr\/\" rel=\"nofollow\">An Introduction to Programming in Emacs Lisp<\/a>. This (and the emacs manual of course) can be read from emacs in info mode as well.<\/p>\n<p>For asking questions, there is the <a href=\"https:\/\/emacs.stackexchange.com\/\" rel=\"nofollow\">emacs stack exchange<\/a> and the <a href=\"https:\/\/www.reddit.com\/r\/emacs\/\" rel=\"nofollow\">emacs subreddit<\/a>.<\/p>\n<p>The basic syntax for emacs settings is <code>(setq &lt;variable&gt; &lt;value&gt; ...)<\/code>. Note that <code>setq<\/code> can be used to set multiple options at once:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq evil-search-wrap t\n      evil-regexp-search t)<\/pre><\/div>\n<p>For settings that have buffer local values by default (the help for the variable will tell you if this is the case), you\u2019ll want to use <code>setq-default<\/code> to set the default value instead:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq-default indent-tabs-mode nil\n              tab-width 4)<\/pre><\/div>\n<p>You can use <code>setq-local<\/code> set the local value of a variable. If the variable is not already buffer local, it will be made buffer local. You could use this with a mode hook, for example, to determine whether indentation is done with tabs or spaces for a specific programming language. Note that the hook should be quoted:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'c-mode-hook\n          (lambda () (setq-local indent-tabs-mode t)))<\/pre><\/div>\n<p>This would be the vim equivalent:<\/p>\n<pre lang=\"vimrc\">au c_settings\n\tau!\n\tau FileType c setlocal noexpandtab\naugroup END\n<\/pre>\n<p>Functions will only be added to hooks once, even if they are anonymous functions (lambdas).<\/p>\n<p>Also note that for variables created by packages, you can set them before the package is loaded without issues. In some cases, you <i>need<\/i> to set them before a package is loaded (e.g. the evil manual gives some of the <code>evil-want-...<\/code> variables as an example). You can also use <code>add-hook<\/code> with a hook that does not yet exist.<\/p>\n<p>Emacs also provides a <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Easy-Customization.html\" rel=\"nofollow\">GUI for customization<\/a>, but this probably won\u2019t be all that interesting to most vim users.<\/p>\n<h2><a id=\"user-content-keybindings-in-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#keybindings-in-emacs\"><\/a>Keybindings in Emacs<\/h2>\n<p>Unlike in vim where keybindings are often made in terms of other keys, in emacs you usually bind keys to named commands. You <i>can<\/i> bind keys to act as other keys, but there is no concept of \u201cdefault\u201d keybindings, so there is no exact equivalent of vim\u2019s <code>noremap<\/code> (though the key translation functions provided by general.el and evil-collection are similar). When possible, you should prefer to bind to named commands and keymaps, but there are some cases where it may be simpler to use keyboard macros (see <a href=\"#binding-keys-to-keys-keyboard-macros\">Binding Keys to Keys (Keyboard Macros)<\/a>).<\/p>\n<p>The main function you\u2019ll use as an evil user for binding keys is <code>evil-define-key<\/code>. Here are some of the other ones provided to you:<\/p>\n<ul><li><code>global-set-key<\/code><\/li>\n  <li><code>evil-global-set-key<\/code><\/li>\n  <li><code>evil-local-set-key<\/code><\/li>\n  <li><code>evil-define-minor-mode-key<\/code><\/li>\n<\/ul><p><code>evil-define-key<\/code> can be used instead of any of these. All of these, including <code>evil-define-key<\/code>, are just wrappers around <code>define-key<\/code>, but they all serve different purposes. I will elaborate on how these functions work and what they can be used for in the upcoming sections. I\u2019d highly recommend looking at <a href=\"https:\/\/github.com\/noctuid\/general.el\">general.el<\/a> for a unified wrapper for all keybinding functions that reduces the verbosity of key definition and provides functions that are more similar to vim\u2019s (such as <code>general-nmap<\/code>) among other things.<\/p>\n<p>As a quick disclaimer, I\u2019m going to be quoting (instead of sharp quoting) commands in example key definitions. Sharp quoting commands (since they are functions) is perfectly valid and, if anything, is more correct. You generally want to sharp quote functions, but for keybindings, you\u2019ll hardly ever see people do it (including in the emacs manual). I think this is mainly for historical reasons, but it may also be a stylistic preference for some.<\/p>\n<h2><a id=\"user-content-keymap-precedence\" class=\"anchor\" aria-hidden=\"true\" href=\"#keymap-precedence\"><\/a>Keymap Precedence<\/h2>\n<p>In emacs, there is a <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Searching-Keymaps.html\" rel=\"nofollow\">hierarchy of keymaps<\/a> that are searched one by one until a definition for a key is found. Evil keymaps are found in <code>emulation-mode-map-alists<\/code> which puts them close to the top in terms of precedence. Here is the order of precedence of evil\u2019s keymaps as explained in <code>evil-core.el<\/code>:<\/p>\n<ul><li>Intercept keymaps   - <code>evil-make-intercept-map<\/code><\/li>\n  <li>Local state keymap  - <code>evil-local-set-key<\/code><\/li>\n  <li>Minor-mode keymaps  - <code>evil-define-minor-mode-key<\/code><\/li>\n  <li>Auxiliary keymaps   - <code>evil-define-key<\/code><\/li>\n  <li>Overriding keymaps  - <code>evil-make-overriding-map<\/code><\/li>\n  <li>Global state keymap - <code>evil-global-set-key<\/code><\/li>\n<\/ul><p>I will be bringing up precedence later on when it is relevant. For more information, see <a href=\"https:\/\/github.com\/syl20bnr\/spacemacs\/wiki\/Keymaps-guide\">spacemacs\u2019 keymap guide<\/a> (though it is missing minor-mode keymaps) and the commentary in <code>evil-core.el<\/code>.<\/p>\n<h2><a id=\"user-content-global-keybindings-and-evil-states\" class=\"anchor\" aria-hidden=\"true\" href=\"#global-keybindings-and-evil-states\"><\/a>Global Keybindings and Evil States<\/h2>\n<p>To make global keybindings in emacs without evil, one would normally use <code>global-set-key<\/code>. <code>global-set-key<\/code> is just a small wrapper function around <code>define-key<\/code> that defines a key in the current global map and signals a error when the key isn\u2019t a string or vector. As an evil user, you won\u2019t often use this function since evil provides several of its own global keymaps corresponding to vim modes. They are as follows:<\/p>\n<ul><li><code>evil-insert-state-map<\/code><\/li>\n  <li><code>evil-emacs-state-map<\/code><\/li>\n  <li><code>evil-normal-state-map<\/code><\/li>\n  <li><code>evil-visual-state-map<\/code><\/li>\n  <li><code>evil-motion-state-map<\/code><\/li>\n  <li><code>evil-operator-state-map<\/code><\/li>\n  <li><code>evil-outer-text-objects-map<\/code><\/li>\n  <li><code>evil-inner-text-objects-map<\/code><\/li>\n  <li><code>evil-replace-state-map<\/code><\/li>\n<\/ul><p>There are also buffer local versions of these (e.g. <code>evil-normal-state-local-map<\/code>).<\/p>\n<p>Most of these should be self-explanatory coming from vim. Emacs state is similar to insert state but uses emacs keybindings (e.g. <code>C-n<\/code> is bound to <code>next-line<\/code> instead of to <code>evil-complete-next<\/code>). For the most part, the keys are the same as if you weren\u2019t using evil at all in emacs state (apart from <code>evil-toggle-key<\/code> which enters\/exits emacs state, <code>C-z<\/code> by default).<\/p>\n<p>Motion state is a bit strange. Keys bound in motion state are inherited in the normal, visual, and operator state keymaps if they are not shadowed. The same inheritance rules apply to normal state, and the main reason motion state exists is for use with read-only modes where insertion keybindings aren\u2019t useful. For example, motion state is the default state for <code>help-mode<\/code>. This means that, by default, only keys bound in motion state will work in <code>help-mode<\/code>.<\/p>\n<p>I personally think that the existence of motion state is a bad idea since it often confuses beginners as there is no vim equivalent and its purpose may not be immediately clear, has a misleading name (it isn\u2019t only used for motions), addresses what I consider a non-issue (e.g. accidentally pressing <code>i<\/code> in a read-only buffer), and addresses this issue poorly. For example, motion state isn\u2019t suitable for all read-only modes (e.g. motions don\u2019t make sense in <code>ediff-mode<\/code>), and remapping insertion commands to be ignored (which is what <a href=\"https:\/\/github.com\/emacs-evil\/evil-collection\">evil-collection<\/a> now does) is a more foolproof and unobtrusive alternative to creating new states. That said, as long as you remember that evil binds motions and some other commands in motion state by default and are aware of <code>evil-motion-state-modes<\/code> and <code>evil-set-initial-state<\/code> (see <a href=\"#make-evil-normal-state-the-initial-state-always\">Make Evil Normal State the Initial State Always<\/a> for information on using normal state instead of motion state in all modes), you shouldn\u2019t encounter any issues.<\/p>\n<p>If you are ever want to know what state a key is bound in, you can check <code>evil-maps.el<\/code> or use <code>lookup-key<\/code>. For example, <code>evil-next-visual-line<\/code> is bound to <code>gj<\/code> in motion state instead of in the normal state keymap (you can check this with <code>(lookup-key evil-normal-state-map \"gj\")<\/code> which will return <code>nil<\/code>). Similarly, if you look up the operator keys such as <code>d<\/code>, you will find that they are only explicitly bound in normal state and not in visual state. Generally, keys are only bound directly in visual state when they have a different behavior from the normal state keys (e.g. <code>u<\/code> and <code>U<\/code> for altering case).<\/p>\n<p>Also note that defining a key in <code>evil-visual-state-map<\/code> is more like <code>xmap<\/code> in vim since there is no \u201cselect\u201d state in evil.<\/p>\n<p>These are the other evil keymaps that might be useful:<\/p>\n<ul><li><code>evil-ex-search-keymap<\/code> (<code>\/<\/code> and <code>?<\/code>)<\/li>\n  <li><code>evil-ex-completion-map<\/code> (<code>:<\/code>)<\/li>\n  <li><code>evil-command-window-mode-map<\/code> (<code>q:<\/code>; you\u2019d use <code>evil-define-key<\/code> for this)<\/li>\n  <li><code>evil-window-map<\/code> (a prefix map for the <code>C-w<\/code> keys)<\/li>\n<\/ul><p>Since <code>define-key<\/code> is the basis for key definition in emacs, I will begin by explaining it. The basic format of <code>define-key<\/code> is <code>(define-key &lt;keymap&gt; &lt;key&gt; &lt;definition&gt;)<\/code>. The specified key can be a string (or something that evaluates to a string) or a vector. You probably won\u2019t want to use a vector of characters instead of a string, but you can use a vector to <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Remapping-Commands.html\" rel=\"nofollow\">remap a command<\/a>, for example. The definition will normally be a command (or something that evaluates to one), but it can also be a keymap or a string. A key bound to a keymap is a prefix key. Binding a key to a string will cause emacs to execute that string as a keyboard macro (see <a href=\"#binding-keys-to-keys-keyboard-macros\">Binding Keys to Keys (Keyboard Macros)<\/a> for examples). See the help text for <code>define-key<\/code> (e.g. <code>C-h f define-key RET<\/code>) for more information on valid definitions.<\/p>\n<p>Here is what a basic <code>nmap<\/code> command equivalent would look like in emacs:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key evil-normal-state-map \"j\" 'evil-next-visual-line)\n(define-key evil-normal-state-map \"k\" 'evil-previous-visual-line)\n;; with `evil-define-key'\n(evil-define-key nil evil-normal-state-map\n  \"j\" 'evil-next-visual-line\n  \"k\" 'evil-previous-visual-line)\n;; with general.el\n(general-nmap\n  \"j\" 'evil-next-visual-line\n  \"k\" 'evil-previous-visual-line)<\/pre><\/div>\n<p>Evil also provides a convenience function called <code>evil-global-set-key<\/code> that allows you to simply specify the name of the state as opposed to the full keymap name:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-global-set-key 'motion \"j\" 'evil-next-visual-line)\n(evil-global-set-key 'motion \"k\" 'evil-previous-visual-line)\n;; `evil-define-key' can also used with \"global\"\n(evil-define-key 'motion 'global\n  \"j\" 'evil-next-visual-line\n  \"k\" 'evil-previous-visual-line)<\/pre><\/div>\n<p>Remember that binding a key in motion state is like binding a key in the normal, visual, and operator states all at once (unless that key is already bound in one of those states).<\/p>\n<p>You can write the key portion as just a string, but often people will use <code>kbd<\/code> to conveniently write keys that have special characters in them like control and space. This follows the format mentioned in <a href=\"#key-notation\">Key Notation<\/a>. These are equivalent:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key evil-normal-state-map \"\\C-j\" 'evil-next-visual-line)\n(define-key evil-normal-state-map (kbd \"C-j\") 'evil-next-visual-line)\n;; general.el implicitily adds a kbd by default\n(general-nmap \"C-j\" 'evil-next-visual-line)<\/pre><\/div>\n<h2><a id=\"user-content-unbinding-a-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#unbinding-a-key\"><\/a>Unbinding a Key<\/h2>\n<p>There is no dedicated alternative to <code>define-key<\/code> for unbinding a key in emacs (though there are wrappers around <code>define-key<\/code> like <code>global-unset-key<\/code>). To unbind a key, you simply bind it to <code>nil<\/code>.<\/p>\n<h2><a id=\"user-content-leader-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#leader-key\"><\/a>Leader Key<\/h2>\n<p>There is no exact equivalent of a \u201cleader\u201d key in evil. You can have named prefix keys with a package like <a href=\"https:\/\/github.com\/noctuid\/general.el\">general.el<\/a> or bind a prefix key to a named keymap. This will allow you to easily change your \u201cleader\u201d\/prefix key later. Here\u2019s an example that doesn\u2019t use any extra packages:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar my-leader-map (make-sparse-keymap)\n  \"Keymap for \\\"leader key\\\" shortcuts.\")\n;; binding \",\" to the keymap\n(define-key evil-normal-state-map \",\" my-leader-map)\n;; binding \",b\"\n(define-key my-leader-map \"b\" 'list-buffers)\n;; change the \"leader\" key to space\n(define-key evil-normal-state-map \",\" 'evil-repeat-find-char-reverse)\n(define-key evil-normal-state-map (kbd \"SPC\") my-leader-map)\n;; general.el can automate the process of prefix map\/command creation\n(general-nmap\n  :prefix \"SPC\"\n  :prefix-map 'my-leader-map\n  \",\" 'list-buffers)<\/pre><\/div>\n<p>This isn\u2019t quite the same as the leader key in vim. In vim, <code>&lt;leader&gt;<\/code> is builtin and sometimes used by plugins to bind keys (despite being considered bad practice). This could potentially be convenient since it gives you some control over what you would like to use as a \u201cmain\u201d prefix key without having to manually make keybindings for it with every plugin. In emacs, evil packages generally do not force the use of some extra package that provides \u201cleader\u201d functionality onto the user, and there is no standard, generic \u201cleader\u201d prefix map provided by evil. This means that \u201cleader\u201d keybindings in emacs will be your personal ones. Note that some packages do provide prefix keymaps that you can then choose a prefix key for though (e.g. <code>projectile-command-map<\/code>).<\/p>\n<p>In terms of functionality, it might be said that named prefixes are actually slightly more convenient in emacs than <code>&lt;leader&gt;<\/code>. You can use as many prefix keymaps as you would like and can bind as many keys to the same prefix keymap as you would like (which may be useful if you want to use a different key to access a prefix keymap in insert state). Note that you can essentially achieve the same functionality (multiple named prefixes) in vim using <code>&lt;Plug&gt;<\/code> mappings.<\/p>\n<p>For an example of a prefix keymap used by evil, see <code>evil-window-map<\/code> which is used for <code>C-w<\/code> commands. From <code>evil-maps.el<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-prefix-command 'evil-window-map)\n(define-key evil-window-map \"b\" 'evil-window-bottom-right)\n(define-key evil-window-map \"c\" 'evil-window-delete)\n...\n(define-key evil-motion-state-map \"\\C-w\" 'evil-window-map)<\/pre><\/div>\n<p>Note the use of <code>define-prefix-command<\/code> instead of <code>defvar<\/code>. Either way works, but <code>define-prefix-command<\/code> is specifically intended for this purpose (see its documentation for more information).<\/p>\n<p>You can check out another alternative for emulating the leader key in the <a href=\"https:\/\/github.com\/noctuid\/evil-guide\/wiki#using-hydra-for-leader-key\">wiki<\/a><\/p>\n<h2><a id=\"user-content-mode-specific-keybindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#mode-specific-keybindings\"><\/a>Mode Specific Keybindings<\/h2>\n<p><code>evil-define-key<\/code> can be used to define keys in specific states for specific modes. The basic format is <code>(evil-define-key &lt;state&gt; &lt;keymap&gt; &lt;key&gt; &lt;definition&gt; ...)<\/code>. Unlike with <code>define-key<\/code>, <code>evil-define-key<\/code> can be used to define multiple keys at once. The state argument can be a single state or a list of states. <code>evil-define-key<\/code> will also defer keybindings if the specified keymap does not exist. This means that you can use it without putting it in an <code>eval-after-load<\/code> for packages that haven\u2019t been loaded yet.<\/p>\n<p>Here is an example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-key 'normal org-mode-map\n  (kbd \"TAB\") 'org-cycle\n  \"&gt;\" 'org-shiftmetaright\n  \"&lt;\" 'org-shiftmetaleft)<\/pre><\/div>\n<p>Coming from vim, this is a lot nicer than using buffer local keybindings with autocommands or ftplugin files in my opinion.<\/p>\n<p>The state can also be nil, so you could also use it like <code>define-key<\/code> except to define multiple keys at once, for example, in <code>evil-normal-state-map<\/code>. I\u2019d recommend using general.el instead if you want this functionality.<\/p>\n<p>If you don\u2019t need keybindings to be deferred and would rather use a function (<code>evil-define-key<\/code> is a macro), <code>evil-define-key*<\/code> was recently added. Also note that <code>evil-declare-key<\/code> is an alias for <code>evil-define-key<\/code>.<\/p>\n<p>There is also a function called <code>evil-define-minor-mode-key<\/code> that is similar to <code>evil-define-key<\/code>. Some differences are that <code>evil-define-minor-mode-key<\/code> only works with minor modes, only allows specifying a single state that cannot be nil, and keys defined with it have a higher precedence than those defined with <code>evil-define-key<\/code>. You probably won\u2019t need to use this function often, but it has a main practical difference that allows it to be used as a workaround for some shortcomings of <code>evil-define-key<\/code> (see <a href=\"#why-dont-keys-defined-with-evil-define-key-work-immediately\">Why don\u2019t keys defined with <code>evil-define-key<\/code> work (immediately)?<\/a>).<\/p>\n<h2><a id=\"user-content-buffer-local-keybindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#buffer-local-keybindings\"><\/a>Buffer Local Keybindings<\/h2>\n<p>Emacs does not have a builtin function for creating buffer local keybindings (that\u2019s not to say there is no such thing as local keymaps; any variable in emacs can be made buffer-local). There is <code>local-set-key<\/code>, but it will bind a key for a mode instead of for a buffer. General.el provides a way to locally bind keys for both evil and non-evil keybindings. Evil also provides <code>evil-local-set-key<\/code> which will work as expected. It is similar to <code>evil-global-set-key<\/code> in that it is a simple wrapper around <code>define-key<\/code> and can only take a single key and definition. For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-local-set-key 'normal key def)\n;; is the same as\n(define-key evil-normal-state-local-map key def)\n;; alternatively with `evil-define-key'\n(evil-define-key 'normal 'local key def)<\/pre><\/div>\n<p>There are good use cases for local keybindings (e.g. maybe you want to bind keys to jump to particular headings in a specific org file), but most are specific to the person and not generally useful. Here\u2019s an example that is a workaround to a deficiency with <code>evil-define-key<\/code> (again, [[<a href=\"#why-dont-keys-defined-with-evil-define-key-work-immediately\">see here<\/a> for a preferable solution). Maybe you want to bind <code>SPC '= to toggle editing an org source block. Keys bound with ~evil-define-key~ in =org-src-mode-map<\/code> won\u2019t take effect immediately, so you can use a hook and local keybindings as one possible workaround:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-key 'normal org-mode-map\n  (kbd \"SPC '\") 'org-edit-special)\n;; you can do this, but the key won't work immediately\n;; (evil-define-key 'normal org-src-mode-map\n;;   (kbd \"SPC '\") 'org-edit-src-exit)\n;; this is a potential workaround\n(defun my-setup-org-edit-src-exit ()\n  (evil-local-set-key 'normal (kbd \"SPC '\") 'org-edit-src-exit))\n(add-hook 'org-src-mode-hook #'my-setup-org-edit-src-exit)<\/pre><\/div>\n<p>This is closer to how you might define local keybindings in vim (with an autocommand and buffer local keybindings). Note that you can replace the <code>#'my-setup...<\/code> with the actual <code>(defun...)<\/code> without problems, but <code>defun<\/code>\u2019s return value is technically undefined, so this may not work in future versions of Emacs.<\/p>\n<h2><a id=\"user-content-binding-keys-to-keys-keyboard-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#binding-keys-to-keys-keyboard-macros\"><\/a>Binding Keys to Keys (Keyboard Macros)<\/h2>\n<p>While you should generally avoid binding keys to keyboard macros when you can bind directly to a command or keymap, binding to a sequence of keys can be simpler than creating new commands:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-key 'normal 'global\n  ;; select the previously pasted text\n  \"gp\" \"`[v`]\"\n  ;; run the macro in the q register\n  \"Q\" \"@q\")\n(evil-define-key 'visual 'global\n  ;; run macro in the q register on all selected lines\n  \"Q\" (kbd \":norm @q RET\")\n  ;; repeat on all selected lines\n  \".\" (kbd \":norm . RET\"))\n;; alternative command version\n(defun my-norm@q ()\n  \"Apply macro in q register on selected lines.\"\n  (interactive)\n  (evil-ex-normal (region-beginning) (region-end) \"@q\"))\n(evil-define-key 'visual 'global \"Q\" #'my-norm@q)<\/pre><\/div>\n<p>These examples are similar to how you might do things in vim. Keyboard macros are fine for simple cases, but note that they do have some limitations. For example:<\/p>\n<ul><li>The prefix argument\/count will apply to the macro (i.e. it will run that many times), not to the next command that runs<\/li>\n  <li>Macro are not suitable for incomplete sequences (e.g. <code>C-c<\/code> or another key bound to a keymap)<\/li>\n  <li>Using <code>C-h k<\/code>, the help buffer will just show the keyboard macro, not the help text for the actual command that will run<\/li>\n<\/ul><p>For a potentially better approach for simulating keys see <a href=\"#using-emacs-keybindings-in-normal-state\">Using Emacs Keybindings in Normal State<\/a>.<\/p>\n<h2><a id=\"user-content-mapping-under-keys-that-arent-prefix-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#mapping-under-keys-that-arent-prefix-keys\"><\/a>Mapping Under Keys That Aren\u2019t Prefix Keys<\/h2>\n<p>In vim, it is somewhat common to bind non-operator functionality under operators (e.g. <code>co&lt;keys&gt;<\/code> to toggling options). It\u2019s is also somewhat common for people to do something like remap <code>cw<\/code> to <code>ciw<\/code>. With evil, it is not possible to bind something like <code>cow<\/code> directly since <code>c<\/code> is not a prefix key (it is already bound to <code>evil-change<\/code>). For this specific case, you can bind under <code>evil-operator-state-map<\/code>. If you want to have different things executed based on the specific operator (<code>d<\/code> vs. <code>c<\/code>) you can check <code>evil-this-operator<\/code>. <a href=\"https:\/\/github.com\/emacs-evil\/evil-collection\/blob\/6ddfc3f7ffc09ae8fcee05e044d3a35aaddacf94\/evil-collection-vdiff.el#L40\">This<\/a> is how evil-collection defines <code>dp<\/code> and <code>do<\/code> for vdiff mode and how I would recommend other packages implement this functionality when possible.<\/p>\n<p>This method won\u2019t work, however, if you wanted to rebind something like <code>ct&lt;key&gt;<\/code> or <code>cw<\/code> (you\u2019d have to redefine <code>evil-find-char-to<\/code> and <code>evil-forward-word-begin<\/code>). For a more general solution that will work for both cases, there is general.el\u2019s <code>general-key-dispatch<\/code> macro. For more information and specific examples see <a href=\"https:\/\/github.com\/noctuid\/general.el#mapping-under-non-prefix-keys\">here<\/a>.<\/p>\n<p>Since this functionality is used in the next section, I\u2019ll go ahead and mention it now. Emacs allows \u201cadvising\u201d a function. This means that you can have certain code execute before, after, or even instead of a function. The examples in this guide are fairly simple, but you can see the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Advising-Functions.html\" rel=\"nofollow\">corresponding section<\/a> of the emacs manual for more information.<\/p>\n<p>There is a common misconception that evil is unable integrate well with certain parts of emacs. What is true is that evil has some default configuration that may be annoying and does not provide default keybindings for all emacs packages. That said, once you know about the tools evil gives you, the process of integration becomes much easier. In the following sections, I will present various techniques for reconciling emacs and evil keybindings.<\/p>\n<p>Some people prefer to use evil only for text editing and use the default emacs keybindings for applications such as dired and mail clients. Evil makes this easy to do by altering initial states or using overriding keymaps for these modes. I personally prefer to use evil everywhere. Some people argue that the lack of default keybindings for applications like dired means that far too much work is required to use evil with them. Some argue that the lack of consistency makes evil not worth using at all. In my experience, making your own keybindings for some application like mu4e takes significantly less time than reading the documentation and can be done simultaneously. Even if you disagree, it\u2019s no longer the case that most modes have no evil support. In many cases there are packages for specific modes that will make evil keybindings for you, such as <a href=\"https:\/\/github.com\/justbur\/evil-magit\">evil-magit<\/a>. I don\u2019t personally use these unless they provide new functionality too, but some people find these packages indispensable. On the other hand, I\u2019d highly recommend looking at <a href=\"https:\/\/github.com\/noctuid\/evil-guide\/blob\/master\/evil-collection\">evil-collection<\/a>. The main difference between this package and others is that it attempts to cover everything as opposed to a single mode. The main benefit of this approach is that <code>evil-collection<\/code> uses a consistent set of rules for what keys are bound to what types of actions. Because of this, I will likely switch my personal configuration for all relevant modes to use <code>evil-collection<\/code> as a base in the future. Even if you don\u2019t agree with the specific key choices, it is easy enough to swap them for your own. On the other hand, most emacs applications use inconsistent keybindings for common actions such as filtering, sorting, marking, etc., so one might even argue that with evil-collection, using evil for such applications is actually easier and more consistent than using them normally.<\/p>\n<h2><a id=\"user-content-what-overrides-evil\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-overrides-evil\"><\/a>What Overrides Evil?<\/h2>\n<p>If you\u2019ve ever entered some buffer and noticed that your normal state keybindings weren\u2019t working, it was probably because of some configuration done by evil (see <code>evil-integration.el<\/code>). There are very few cases where another keymap takes precedence over an evil one.<\/p>\n<p>Referring back to the fact that evil\u2019s keymaps are located in <code>emulation-mode-map-alists<\/code> and the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Searching-Keymaps.html\" rel=\"nofollow\">Searching Keymaps<\/a> section of the emacs manual, you\u2019ll notice that emacs will check in the keymap char property before reaching evil\u2019s keymaps. An example of where this would override evil keybindings is when the point is in a magit diff section in the magit status buffer. See <a href=\"#prevent-text-property-maps-from-overriding-evil\">here<\/a> for information on how to deal with this.<\/p>\n<p>The other main case where evil keybindings will be overridden is by keybindings in <code>overriding-terminal-local-map<\/code>, which has the highest precedence in emacs. Normally it is used by <code>set-transient-map<\/code> to temporarily to elevate a keymap to the highest precedence. Note that generally this will not get in the way of evil keybindings (e.g. this is the mechanism used by <code>hydra<\/code>) For an example of packages that use <code>set-transient-map<\/code>, see <a href=\"http:\/\/oremacs.com\/2014\/12\/31\/keymap-arms-race\/\" rel=\"nofollow\">this article<\/a>.<\/p>\n<p>Finally, it may be possible for other keymaps in <code>emulation-mode-map-alists<\/code> to override evil. For example, when the company popup is active, keys in <code>company-active-map<\/code> will have precedence. If this causes any annoyances, you can unbind the offending key in <code>company-active-map<\/code>.<\/p>\n<h2><a id=\"user-content-evils-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#evils-tools\"><\/a>Evil\u2019s Tools<\/h2>\n<p>Evil provides a way to set the initial state for a mode as well as to allow keybindings in a keymap to override global keybindings for some\/all states. I will be referencing these variables\/functions in later sections, so I will briefly explain them now.<\/p>\n<p><b>Initial States<\/b><\/p>\n<p>Evil has \u201cinitial state\u201d lists containing modes. For example, if you wanted <code>org-mode<\/code> buffers to start in emacs state, you could add <code>org-mode<\/code> to <code>evil-emacs-state-modes<\/code> and remove it from the list it was previously in or just use <code>(evil-set-initial-state 'org-mode 'emacs)<\/code>.<\/p>\n<p><b>Overriding\/Intercept Keymaps<\/b><\/p>\n<p>Evil has two variables called <code>evil-overriding-maps<\/code> and <code>evil-intercept-maps<\/code>. They both have a similar effect. Keybindings made in keymaps listed in <code>evil-override-maps<\/code> will override global evil keybindings. For example, <code>(Info-mode-map . motion)<\/code> is in this list by default, meaning that keys bound in <code>Info-mode-map<\/code> (when it is active) will override keys bound in <code>evil-motion-state-map<\/code>. If no state is specified (e.g. <code>(compilation-mode-map)<\/code>, another default), keybindings in all global keymaps will be overridden. The difference between intercept and overriding keymaps has to do with precedence (refer back to <a href=\"#keymap-precedence\">Keymap Precedence<\/a>). Keys bound in a overriding keymap will not override keys bound with <code>evil-local-set-key<\/code>, <code>evil-define-minor-mode-key<\/code>, or <code>evil-define-key<\/code>, but keys bound in an intercept keymap will.<\/p>\n<p>Note that changing these variables after evil is loaded using <code>setq<\/code> will have no effect. You can use customize, but I recommend using the corresponding functions instead: <code>evil-make-overriding-map<\/code> and <code>evil-make-intercept-map<\/code>.<\/p>\n<p>Evil also has a function called <code>evil-add-hjkl-bindings<\/code> that can be used to add back <code>hjkl<\/code> movement keybindings for a mode after making its keymap an overriding keymap.<\/p>\n<p><b>Evil Command Properties<\/b><\/p>\n<p>Evil has a concept of \u201ccommand properties\u201d that can be added with <code>evil-add-command-properties<\/code>, <code>evil-set-command-property<\/code>, or <code>evil-set-command-properties<\/code> and gotten with <code>evil-get-command-property<\/code> or <code>evil-get-command-properties<\/code>. These can be used to, for example, customize whether or not and how a command will be repeated later with <code>.<\/code> (<code>evil-repeat<\/code>).<\/p>\n<h2><a id=\"user-content-preventing-certain-keys-from-being-overridden\" class=\"anchor\" aria-hidden=\"true\" href=\"#preventing-certain-keys-from-being-overridden\"><\/a>Preventing Certain Keys From Being Overridden<\/h2>\n<p>Regardless of whether you want to sometimes have emacs keys override keys in normal\/motion state, you may want to have certain keys universally available (e.g. prefix keys used for window\/workgroup\/buffer\/file navigation). The suggested method for doing this is to use evil intercept keymaps since they have the highest precedence. This means that no standard method a package could use to define an evil key (<code>evil-define-key<\/code>, <code>evil-define-minor-mode-key<\/code>, <code>evil-local-set-key<\/code>, etc.) can override keys you\u2019ve bound in an intercept keymap. Here\u2019s an example of how to create such a mode\/keymap yourself:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar my-intercept-mode-map (make-sparse-keymap)\n  \"High precedence keymap.\")\n(define-minor-mode my-intercept-mode\n  \"Global minor mode for higher precedence evil keybindings.\"\n  :global t)\n(my-intercept-mode)\n(dolist (state '(normal visual insert))\n  (evil-make-intercept-map\n   ;; NOTE: This requires an evil version from 2018-03-20 or later\n   (evil-get-auxiliary-keymap my-intercept-mode-map state t t)\n   state))\n(evil-define-key 'normal my-intercept-mode-map\n  (kbd \"SPC f\") 'find-file)\n;; ...<\/pre><\/div>\n<p>If you are using <a href=\"https:\/\/github.com\/noctuid\/general.el\">https:\/\/github.com\/noctuid\/general.el<\/a>, this configuration is done automatically, so you can just use the =\u2019override= keymap:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-override-mode)\n(general-def 'normal 'override\n  \"SPC f\" 'find-file)<\/pre><\/div>\n<h3><a id=\"user-content-prevent-text-property-maps-from-overriding-evil\" class=\"anchor\" aria-hidden=\"true\" href=\"#prevent-text-property-maps-from-overriding-evil\"><\/a>Prevent Text Property Maps from Overriding Evil<\/h3>\n<p>Locations in a buffer can have their own keymaps. As these keymaps have a higher precedence than evil, you will have to clear them to prevent them from overriding your keys. As of emacs 25, help pages will tell you where a key is bound, so to find the keymap you could press <code>C-h k &lt;key that is being overriden&gt;<\/code>. A good example of when you might encounter these keymaps is for links (enter and mouse clicks are often remapped) and for magit-status diff sections. To control the keybindings in these locations, you need to clear the keymap (or at least unbind the keys you don\u2019t want) and then define the keys as you like. Note that you should use <code>define-key<\/code> and not <code>evil-define-key<\/code> for this.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq magit-hunk-section-map (make-sparse-keymap))\n(define-key magit-hunk-section-map \"s\" 'magit-stage)<\/pre><\/div>\n<h2><a id=\"user-content-use-some-emacs-keybindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-some-emacs-keybindings\"><\/a>Use Some Emacs Keybindings<\/h2>\n<h3><a id=\"user-content-switching-between-evil-and-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#switching-between-evil-and-emacs\"><\/a>Switching Between Evil and Emacs<\/h3>\n<p>Some people prefer to just use evil for editing and stick to emacs keybindings elsewhere. This method just involves altering the initial state for certain modes or using <code>evil-make-overriding-map<\/code>. For example, if you just wanted to use dired\u2019s keybindings as they are without touching your normal state keybindings in dired-mode, you could do the following:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-set-initial-state 'dired-mode 'emacs)<\/pre><\/div>\n<p>If you wanted to override normal state with dired\u2019s keybindings, you could do this:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-make-overriding-map dired-mode-map 'normal)<\/pre><\/div>\n<p>The latter is what evil does by default (followed by an <code>evil-add-hjkl-bindings<\/code>).<\/p>\n<p>Note that at any time you can use <code>C-z<\/code> (bound to <code>evil-emacs-state<\/code>) to enter emacs state or <code>\\<\/code> (bound to <code>evil-execute-in-emacs-state<\/code>) to execute the next command in emacs state. In emacs state, <code>C-z<\/code> and <code>ESC<\/code> are bound to switch to the previous state. This may not be what you want if you\u2019ve entered emacs state from insert state, so you may want to rebind <code>ESC<\/code> to always enter normal state instead:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key evil-emacs-state-map [escape] 'evil-normal-state)<\/pre><\/div>\n<p>Note that in this case, attempting to rebind <code>(kbd \"ESC\")<\/code> will not work.<\/p>\n<p>If you want to use emacs keybindings instead of the ones that evil makes in insert state, you can change the <code>evil-insert-state-bindings<\/code> variable to your liking or set <code>evil-disable-insert-state-bindings<\/code> to <code>t<\/code> before loading evil (or use customize to set it afterwards). I recommend doing this instead of aliasing or overriding <code>evil-insert-state<\/code> to <code>evil-emacs-state<\/code> because the result is pretty much the same and evil intentionally does not record repeat information in emacs state.<\/p>\n<p>These are the keybindings evil makes in insert state by default:<\/p>\n<table><tbody><tr><th>key<\/th><th>command<\/th><th>emacs default<\/th><\/tr><tr><td><code>C-v<\/code><\/td><td><code>quoted-insert<\/code><\/td><td><code>scroll-up-command<\/code><\/td><\/tr><tr><td><code>C-k<\/code><\/td><td><code>evil-insert-digraph<\/code><\/td><td><code>kill-line<\/code><\/td><\/tr><tr><td><code>C-o<\/code><\/td><td><code>evil-execute-in-normal-state<\/code><\/td><td><code>open-line<\/code><\/td><\/tr><tr><td><code>C-r<\/code><\/td><td><code>evil-paste-from-register<\/code><\/td><td><code>isearch-backward<\/code><\/td><\/tr><tr><td><code>C-y<\/code><\/td><td><code>evil-copy-from-above<\/code><\/td><td><code>yank<\/code><\/td><\/tr><tr><td><code>C-e<\/code><\/td><td><code>evil-copy-from-below<\/code><\/td><td><code>move-end-of-line<\/code><\/td><\/tr><tr><td><code>C-n<\/code><\/td><td><code>evil-complete-next<\/code><\/td><td><code>next-line<\/code><\/td><\/tr><tr><td><code>C-p<\/code><\/td><td><code>evil-complete-previous<\/code><\/td><td><code>previous-line<\/code><\/td><\/tr><tr><td><code>C-x C-n<\/code><\/td><td><code>evil-complete-next-line<\/code><\/td><td><code>set-goal-column<\/code><\/td><\/tr><tr><td><code>C-x C-p<\/code><\/td><td><code>evil-complete-previous-line<\/code><\/td><td><code>mark-page<\/code><\/td><\/tr><tr><td><code>C-t<\/code><\/td><td><code>evil-shift-right-line<\/code><\/td><td><code>transpose-chars<\/code><\/td><\/tr><tr><td><code>C-d<\/code><\/td><td><code>evil-shift-left-line<\/code><\/td><td><code>delete-char<\/code><\/td><\/tr><tr><td><code>C-a<\/code><\/td><td><code>evil-paste-last-insertion<\/code><\/td><td><code>move-beginning-of-line<\/code><\/td><\/tr><tr><td><code>C-w<\/code><\/td><td><code>evil-delete-backward-word<\/code><\/td><td><code>kill-region<\/code><\/td><\/tr><tr><td>or <code>evil-window-map<\/code><\/td><\/tr><tr><td>(see <code>evil-want-C-w-delete<\/code>)<\/td><\/tr><\/tbody><\/table><p>In <code>evil-insert-state-bindings<\/code>, evil also replaces <code>delete-backward-char<\/code> with <code>evil-delete-backward-char-and-join<\/code> and binds <code>&lt;mouse-2&gt;<\/code> to <code>mouse-yank-primary<\/code> (same as the default). Regardless of the value of <code>evil-insert-state-bindings<\/code> or <code>evil-disable-insert-state-bindings<\/code>, evil will bind the following in insert state:<\/p>\n<table><tbody><tr><th>key<\/th><th>command<\/th><th>emacs default<\/th><\/tr><tr><td><code>&lt;delete&gt;<\/code><\/td><td><code>delete-char<\/code><\/td><td><code>delete-forward-char<\/code><\/td><\/tr><tr><td><code>&lt;escape&gt;<\/code><\/td><td><code>evil-normal-state<\/code><\/td><td>acts like meta\/alt<\/td><\/tr><tr><td><code>evil-toggle-key<\/code><\/td><td><code>evil-execute-in-normal-state<\/code><\/td><td>depends<\/td><\/tr><\/tbody><\/table><p>If you don\u2019t like these, you can always unbind or rebind them. <code>evil-toggle-key<\/code> defaults to <code>C-z<\/code> (bound to <code>suspend-frame<\/code> by default).<\/p>\n<h3><a id=\"user-content-using-emacs-keybindings-in-normal-state\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-emacs-keybindings-in-normal-state\"><\/a>Using Emacs Keybindings in Normal State<\/h3>\n<p>For modes that still involve editing text but add extra keybindings, you don\u2019t always have to rely on a package to make keybindings for you in normal state or rebind everything yourself. <code>C-c<\/code> is used as a mode-specific prefix in emacs, and if you are okay with the keys under it for a mode, you can simply change the prefix to something else in normal state. While this won\u2019t always cover all the keybindings made by a mode (e.g. org-mode), it can be helpful.<\/p>\n<p>For example, using <a href=\"https:\/\/github.com\/noctuid\/general.el#simulating-keypresses\">~general-simulate-key~ or ~general-key~<\/a>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key evil-normal-state-map (kbd \"SPC\") (general-simulate-key \"C-c\"))\n;; act as whatever C-n is currently bound to in emacs state (eg. `next-line' or\n;; `dired-next-line')\n(define-key evil-normal-state-map (kbd \"j\") (general-key \"C-n\" :state 'emacs))<\/pre><\/div>\n<p>With the above configuration, you could, for example, press <code>SPC C-e<\/code> in normal state in org mode to bring up the export dispatcher. Emacs allows you to bind keys to keymaps, so the following is also possible:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key evil-normal-state-map (kbd \"SPC h\") help-map)\n(define-key evil-normal-state-map (kbd \"SPC x\") ctl-x-map)<\/pre><\/div>\n<p>There are other ways to simulate keys of course, but the way listed here is well-suited for keybindings. This method allows prefix arguments to work properly for the command that ends up running, whereas a keyboard macro would eat the prefix argument. I personally also prefer this method to the key translation methods mentioned <a href=\"https:\/\/www.emacswiki.org\/emacs\/Evil#toc13\" rel=\"nofollow\">here<\/a>.<\/p>\n<h2><a id=\"user-content-use-evil-everywhere\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-evil-everywhere\"><\/a>Use Evil Everywhere<\/h2>\n<h3><a id=\"user-content-make-evil-normal-state-the-initial-state-always\" class=\"anchor\" aria-hidden=\"true\" href=\"#make-evil-normal-state-the-initial-state-always\"><\/a>Make Evil Normal State the Initial State Always<\/h3>\n<p>You can use the following configuration to have all modes start in normal state:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq evil-emacs-state-modes nil)\n(setq evil-insert-state-modes nil)\n(setq evil-motion-state-modes nil)<\/pre><\/div>\n<p>Since <code>evil-default-state<\/code> defaults to <code>normal<\/code>, you can simply clear the other mode lists. If you want to be more explicit, you can do this before clearing them.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq evil-normal-state-modes\n      (append evil-emacs-state-modes\n              evil-insert-state-modes\n              evil-normal-state-modes\n              evil-motion-state-modes))<\/pre><\/div>\n<p>If you\u2019d rather have REPLs start in insert state, you may want to keep <code>evil-insert-state-modes<\/code> as it is.<\/p>\n<p>Later if you want to change the state a mode starts in, you should use <code>evil-set-initial-state<\/code> as it will automatically remove the mode from any other state list.<\/p>\n<h3><a id=\"user-content-undoprevent-overridingintercept-maps\" class=\"anchor\" aria-hidden=\"true\" href=\"#undoprevent-overridingintercept-maps\"><\/a>Undo\/Prevent Overriding\/Intercept Maps<\/h3>\n<p>As a disclaimer, overriding keymaps will not interfere with keys <a href=\"#preventing-certain-keys-from-being-overridden\">defined in intercept keymaps<\/a>, and the default <code>evil-intercept-maps<\/code> (which contains <code>edebug-mode-map<\/code> at the time of writing) is something you might want to leave as is. If you just never want emacs keys overriding any evil keys (e.g. there are some keys that you don\u2019t want the same everywhere but don\u2019t want overridden by emacs keys either, you want to make all your keybindings for used modes, or you are using <a href=\"https:\/\/github.com\/noctuid\/evil-guide\/blob\/master\/evil-collection\">evil-collection<\/a> which already provides evil keybindings for used modes), the following information may be useful.<\/p>\n<p>Undoing an override or intercept involves unbinding either <code>[override-state]<\/code> or <code>[intercept-state]<\/code> like so:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key keymap [override-state] nil)\n(define-key keymap [intercept-state] nil)<\/pre><\/div>\n<p>As an example, to undo evil\u2019s default overriding of <code>Info-mode-map<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key Info-mode-map [override-state] nil)<\/pre><\/div>\n<p>Instead of specifically undoing all the overrides that evil makes, you may want to instead prevent evil from ever overriding anything using a more generic method. Evil provides variables containing keymaps to elevate. They must be set to nil <i>before<\/i> evil is loaded:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq evil-overriding-maps nil\n      evil-intercept-maps nil)\n;; ...\n(require 'evil)<\/pre><\/div>\n<p>If you don\u2019t want anything to be overriden, this is not enough. In <code>evil-integration.el<\/code>, <code>evil-make-overriding-map<\/code> is used for dired and ibuffer. If you want to prevent <code>evil-integration.el<\/code> from being loaded, you can set <code>evil-want-integration<\/code> to nil <i>before<\/i> loading evil. If you use <a href=\"https:\/\/github.com\/noctuid\/evil-guide\/blob\/master\/evil-collection\">evil-collection<\/a>, which provides the useful functionality from <code>evil-integration.el<\/code> without creating overriding keymaps, you should set this variable to nil.<\/p>\n<p>If you really want to prevent overriding maps from ever being created (e.g. some other evil package could do it), you can advise <code>evil-overiding-map<\/code> to prevent it from ever doing anything:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(advice-add 'evil-make-overriding-map :override #'ignore)<\/pre><\/div>\n<p>You can always remove this advice later:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(advice-remove 'evil-make-overriding-map #'ignore)<\/pre><\/div>\n<h3><a id=\"user-content-modal-minibuffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#modal-minibuffer\"><\/a>Modal Minibuffer<\/h3>\n<p>Normal state does <i>kind of<\/i> work in the minibuffer if you bind a key to <code>evil-normal-state<\/code>. Evil collection  For the ex command line specifically, it\u2019s worth noting that evil provides <code>q:<\/code>.<\/p>\n<p>Missing using normal mode with Unite, I wrote\n  <a href=\"https:\/\/noctuid.github.io\/blog\/2015\/02\/03\/a-more-evil-helm\/\" rel=\"nofollow\">a blog post<\/a> a while back with the idea of using a hydra to implement modality for helm. Since then, people have created improved versions of my hydra <a href=\"https:\/\/github.com\/abo-abo\/hydra\/wiki\/Helm\">for helm<\/a>, and ivy has such a hydra builtin.<\/p>\n<h3><a id=\"user-content-example-integration-with-pdf-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#example-integration-with-pdf-tools\"><\/a>Example: Integration with Pdf Tools<\/h3>\n<p>Configuring a package for evil is not all that different from configuring a package for vanilla emacs. Often the main difference is that you\u2019ll be using <code>evil-define-key<\/code> instead of <code>define-key<\/code> to change keybindings. You start off by reading the documentation for the package to learn how it works and what keybindings it provides.<\/p>\n<p>Pdf-tools has a <a href=\"https:\/\/github.com\/politza\/pdf-tools#some-keybindings\">section in the readme<\/a> that lists its keybindings. If you are happy with them, you could simply let pdf-tool\u2019s keymap override normal state (excluding your \u201cspecial\u201d non-overridable keys). The readme doesn\u2019t tell you the mode\u2019s keymap name specifically, but it is not hard to figure out. After setting up the basics for pdf-tools, you can open a pdf and evaluate <code>major-mode<\/code> to find out that you are in <code>pdf-view-mode<\/code>. You can get a lot more information with <code>C-h m<\/code> (<code>describe-mode<\/code>). Mode\u2019s keymaps generally match their mode\u2019s name, and in this case the main keymap is <code>pdf-view-mode-map<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-make-overriding-map pdf-view-mode-map 'normal)<\/pre><\/div>\n<p>Alternatively, you can find out what keymaps pdf-tools provides by typing <code>pdf map<\/code> after running <code>helm-apropos<\/code>.<\/p>\n<p>Pdf-tools has some other modes, the other main one being the outline mode (<code>pdf-outline-buffer-mode-map<\/code>). For packages that have 2+ main modes for different contexts, you can just repeat this process as necessary and be done with things if you are content with the default keybindings.<\/p>\n<p>If you\u2019re like me though, you\u2019ll prefer to use vim-like keybindings everywhere. You can either change a few keybindings and use the previous configuration (keys bound with <code>evil-define-key<\/code> here won\u2019t be overriden) or bind all the keys you use yourself. You can either look at the keys mentioned in the readme and check what they are bound to with <code>C-h k<\/code> or use <code>C-h m<\/code> to look at all the keys bound. Here are some basic <code>hjkl<\/code> keybindings:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-key 'normal pdf-view-mode-map\n  \"h\" 'pdf-view-previous-page-command\n  \"j\" (lambda () (interactive) (pdf-view-next-line-or-next-page 5))\n  \"k\" (lambda () (interactive) (pdf-view-previous-line-or-previous-page 5))\n  \"l\" 'pdf-view-next-page-command)<\/pre><\/div>\n<p>You could even bind things in terms of <code>general-simulate-keys<\/code> without even looking up the keys if you preferred to:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-evil-define-key 'normal pdf-view-mode-map\n  \"h\" (general-simulate-keys \"p\" t)\n  \"j\" (general-simulate-keys \"C-n\" t)\n  \"k\" (general-simulate-keys \"C-p\" t)\n  ;; alternatively to scroll more\n  \"j\" (general-simulate-keys \"SPC\" t)\n  \"k\" (general-simulate-keys \"DEL\" t)\n  \"l\" (general-simulate-keys \"n\" t))<\/pre><\/div>\n<p>We can go further if we want:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-key 'normal pdf-view-mode-map\n  \"g\" 'pdf-view-first-page\n  \"G\" 'pdf-view-last-page\n  ;; alternatively\n  \"g\" 'image-bob\n  \"G\" 'image-eob\n  (kbd \"C-o\") 'pdf-history-backward\n  (kbd \"C-i\") 'pdf-history-forward\n  \"m\" 'pdf-view-position-to-register\n  \"'\" 'pdf-view-jump-to-register\n  \"\/\" 'pdf-occur\n  \"o\" 'pdf-outline\n  \"f\" 'pdf-links-action-perform\n  \"b\" 'pdf-view-midnight-minor-mode\n  ...)<\/pre><\/div>\n<p>Using the tools mentioned in this section, none of this is difficult. It may be time consuming, but I think reading the documentation for a new mode takes the majority of the time when compared to making 10-20 basic keybindings for it.<\/p>\n<p>As a bonus, here are some functions I wrote to make pdf-tools even more vimmy. Want to have <code>G<\/code> double as a way for jumping to a specific page number? No problem:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun noct:pdf-view-goto-page (count)\n  \"Goto page COUNT.\nIf COUNT is not supplied, go to the last page.\"\n  (interactive \"P\")\n  (if count\n      (pdf-view-goto-page count)\n    (pdf-view-last-page)))\n(evil-define-key 'normal pdf-view-mode-map\n  \"G\" 'noct:pdf-view-goto-page)<\/pre><\/div>\n<p>Want to copy text using vim keys? Pdf-tools displays pdfs using images, but you can open the current page in a text buffer and use vim keys for selection\/copying there:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun noct:pdf-view-page-as-text ()\n  \"Inserts current pdf page into a buffer for keyboard selection.\"\n  (interactive)\n  (pdf-view-mark-whole-page)\n  (pdf-view-kill-ring-save)\n  (switch-to-buffer (make-temp-name \"pdf-page\"))\n  (save-excursion\n    (yank)))\n(evil-define-key 'normal pdf-view-mode-map\n  \"y\" 'noct:pdf-view-page-as-text)<\/pre><\/div>\n<p>Once you\u2019re done, you can delete the buffer (<code>kill-this-buffer<\/code>) and continue reading where you left off.<\/p>\n<h2><a id=\"user-content-command-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#command-properties\"><\/a>Command Properties<\/h2>\n<p>Evil provides the following functions for customizing how it deals with commands:<\/p>\n<ul><li><code>evil-set-command-property<\/code>: Set one property of a command<\/li>\n  <li><code>evil-put-command-property<\/code>: Alias for <code>evil-set-command-property<\/code><\/li>\n  <li><code>evil-set-command-properties<\/code>: Set all the properties of a command<\/li>\n  <li><code>evil-add-command-properties<\/code>: Set one or more command properties of a command<\/li>\n<\/ul><p>Unless you want to remove command properties entirely from a command, you can just use <code>evil-add-command-properties<\/code>.<\/p>\n<p>When creating motions, text-objects, and commands, you can set command properties with keywords (e.g. to control whether an operator will move the point or exit visual state; see the <code>evil\/Macros<\/code> info node). For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-operator some-operator (args)\n  \"Docstring.\"\n  ;; command properties go after the docstring and before the interactive codes\n  :repeat nil\n  (interactive \"...\")\n  ;; ...\n  )<\/pre><\/div>\n<p>Note that operators, commands, motions, and text objects all have default non-nil properties. Not all properties are applicable to all macros (again, see the <code>Macros<\/code> section of the evil info manual).<\/p>\n<p>The default properties for operators are <code>:repeat t :move-point t keep-visual t :supress-operator t<\/code>.<\/p>\n<p>The default properties for commands are <code>:repeat t<\/code>.<\/p>\n<p>The default properties for motions are <code>:repeat 'motion :keep-visual t<\/code>.<\/p>\n<p>The default properties for text objects are <code>:repeat 'motion :extend-selection t :keep-visual t<\/code>.<\/p>\n<h3><a id=\"user-content-repeat\" class=\"anchor\" aria-hidden=\"true\" href=\"#repeat\"><\/a>:repeat<\/h3>\n<p>The <code>:repeat<\/code> property is used to determine how evil records information for repeating later with <code>evil-repeat<\/code>. These symbols are the possible values by default:<\/p>\n<ul><li><code>t<\/code>: record by keystrokes<\/li>\n  <li><code>motion<\/code>: record by keystrokes only in insert state<\/li>\n  <li><code>change<\/code>: record by changes to the buffer<\/li>\n  <li><code>ignore<\/code> or <code>nil<\/code>: don\u2019t record the command<\/li>\n  <li><code>abort<\/code>: immediately abort recording<\/li>\n<\/ul><p>There is also <code>insert-at-point<\/code> which has a less common use case. If a command does not have a <code>:repeat<\/code> property, evil will treat it as if the repeat property was <code>t<\/code>. You can also create your own recording functions and use them by setting the repeat property to that function\u2019s name. You could also use a custom symbol by adding something like <code>(my-repeat-type . my-repeat-function)<\/code> to <code>evil-repeat-types<\/code>.<\/p>\n<p>Evil also provides some wrappers around <code>evil-add-command-properties<\/code> to set the repeat property for a command:<\/p>\n<ul><li><code>evil-declare-repeat<\/code>: set to <code>t<\/code><\/li>\n  <li><code>evil-declare-not-repeat<\/code>: set to <code>nil<\/code><\/li>\n  <li><code>evil-declare-change-repeat<\/code>: set to <code>change<\/code><\/li>\n  <li><code>evil-declare-ignore-repeat<\/code>: set to <code>ignore<\/code><\/li>\n  <li><code>evil-delare-abort-repeat<\/code>: set to <code>abort<\/code><\/li>\n<\/ul><p><code>evil-declare-repeat<\/code> and <code>evil-declare-not-repeat<\/code> are the most commonly useful ones. You\u2019ll use them for configuring whether a command should be repeatable. See the help text and functions in <code>evil-repeat.el<\/code> for more information. For examples of these being used, I\u2019d recommend looking at <code>evil-integration.el<\/code>.<\/p>\n<h3><a id=\"user-content-jump\" class=\"anchor\" aria-hidden=\"true\" href=\"#jump\"><\/a>:jump<\/h3>\n<p>The <code>:jump<\/code> property takes a boolean value. If a command has a non-nil jump property value, the location prior to running the command will be recorded in the jump list to later be navigated to with <code>C-o<\/code> (<code>evil-jump-backward<\/code>) and <code>C-i<\/code> (<code>evil-jump-forward<\/code>). Commands without this command property will not add a position to the jump list. For example, you could use this to have git-gutter\u2019s commands for navigating hunks save the current location before jumping:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-add-command-properties #'git-gutter:next-hunk :jump t)\n(evil-add-command-properties #'git-gutter:previous-hunk :jump t)<\/pre><\/div>\n<h3><a id=\"user-content-type\" class=\"anchor\" aria-hidden=\"true\" href=\"#type\"><\/a>:type<\/h3>\n<p>The <code>:type<\/code> command property determines how commands, motions, and text objects act with operators (e.g. see <code>evil-delete<\/code> as an example of how an operator can be defined to handle different types). The possible values by default are as follows:<\/p>\n<ul><li><code>:line<\/code><\/li>\n  <li><code>:inclusive<\/code><\/li>\n  <li><code>:exclusive<\/code><\/li>\n  <li><code>:block<\/code><\/li>\n<\/ul><p>This property is mainly useful for evil text objects and motions, but it can also be used for non-evil commands. For example, if you bound <code>j<\/code> <code>next-line<\/code> in operator state and set its type to <code>:inclusive<\/code>, <code>dj<\/code> would no longer delete both lines entirely (<code>next-line<\/code> has a type of <code>line<\/code> by default). Evil allows adding new types using <code>evil-define-type<\/code> (see the <code>Macros<\/code> section of the evil manual for more information).<\/p>\n<h3><a id=\"user-content-move-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#move-point\"><\/a>:move-point<\/h3>\n<p>The <code>:move-point<\/code> property applies when defining operators and determines whether evil will move the point to the beginning of the operator range before running the operator code. Note that it defaults to <code>t<\/code>.<\/p>\n<h3><a id=\"user-content-keep-visual\" class=\"anchor\" aria-hidden=\"true\" href=\"#keep-visual\"><\/a>:keep-visual<\/h3>\n<p>TODO<\/p>\n<h3><a id=\"user-content-suppress-operator\" class=\"anchor\" aria-hidden=\"true\" href=\"#suppress-operator\"><\/a>:suppress-operator<\/h3>\n<p>Commands with a non-nil <code>:suppress-operator<\/code> property (e.g. <code>keyboard-quit<\/code>, <code>evil-force-normal-state<\/code>, and operators) will cause the operator (and repeat recording) to be aborted. For example, if you press <code>d ESC<\/code> or <code>d C-g<\/code>, <code>evil-delete<\/code> will quit and not delete anything.<\/p>\n<h3><a id=\"user-content-motion\" class=\"anchor\" aria-hidden=\"true\" href=\"#motion\"><\/a>:motion<\/h3>\n<p>The <code>:motion<\/code> command property is used for operators to automatically use the range given by some motion. This means that the defined operator will not be usable with motions\/text objects, so it is generally not useful. For example, this is how <code>s<\/code> \/ <code>evil-substitute<\/code> is defined:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-operator evil-substitute (beg end type register)\n  \"Change a character.\"\n  :motion evil-forward-char\n  (interactive \"&lt;R&gt;&lt;x&gt;\")\n  (evil-change beg end type register))<\/pre><\/div>\n<h3><a id=\"user-content-declaring-a-motion\" class=\"anchor\" aria-hidden=\"true\" href=\"#declaring-a-motion\"><\/a>Declaring a Motion<\/h3>\n<p>As an example, you don\u2019t always need use <code>evil-define-motion<\/code> to create new motions. If all you want is to do is control the repeating behavior or the behavior in visual state, you can simply change the command properties of a command. To have an emacs command act like a motion, evil provides <code>evil-declare-motion<\/code>, which will set the <code>:repeat<\/code> property to <code>motion<\/code> and the <code>:keep-visual<\/code> property to <code>t<\/code>. Usually the <code>:keep-visual<\/code> property doesn\u2019t matter for emacs commands (they already won\u2019t exit visual state). Setting the <code>repeat<\/code> property will cause a command to only be part of a recorded repeat in insert state (for example, after a <code>C-o<\/code>).<\/p>\n<h2><a id=\"user-content-state-indicator\" class=\"anchor\" aria-hidden=\"true\" href=\"#state-indicator\"><\/a>State Indicator<\/h2>\n<p>I prefer not to have a state indicator on my mode line and instead to just rely on the color and shape of the cursor to determine what state I\u2019m in. For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq evil-mode-line-format nil\n      evil-insert-state-cursor '(bar \"White\")\n      evil-visual-state-cursor '(box \"#F86155\"))<\/pre><\/div>\n<p>See the <code>Settings\/The cursor<\/code> section of the evil info node for more information.<\/p>\n<h2><a id=\"user-content-user-created-states\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-created-states\"><\/a>User Created States<\/h2>\n<p>Evil lets you create new states with <code>evil-define-state<\/code> (see <code>Macros\/States<\/code> under the evil info node). You may never need to use this, and if you\u2019re looking for something like <a href=\"https:\/\/github.com\/kana\/vim-submode\">vim-submode<\/a>, I\u2019d highly recommend looking at <a href=\"https:\/\/github.com\/abo-abo\/hydra\">hydra<\/a> instead.<\/p>\n<h2><a id=\"user-content-using-escape-to-exit-the-minibuffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-escape-to-exit-the-minibuffer\"><\/a>Using Escape to Exit the Minibuffer<\/h2>\n<p>Escape is used as a prefix key in some parts of emacs, so you need to rebind it to <code>keyboard-escape-quit<\/code> in certain minibuffer-related keymaps for it to always act as expected. You could, for example, use <a href=\"https:\/\/github.com\/emacs-evil\/evil-collection\/blob\/master\/evil-collection-minibuffer.el\">evil-collection-minibuffer.el<\/a> to do this.<\/p>\n<h2><a id=\"user-content-ex-command-definition\" class=\"anchor\" aria-hidden=\"true\" href=\"#ex-command-definition\"><\/a>Ex Command Definition<\/h2>\n<p>You can define your own ex commands using <code>evil-ex-define-cmd<\/code>. For example, this is how <code>copy<\/code> is defined:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-ex-define-cmd \"co[py]\" 'evil-copy)\n(evil-ex-define-cmd \"t\" \"copy\")<\/pre><\/div>\n<p>You could, for example, use this to get some emacs commands with numbers in them to work from the command line (this won\u2019t work with \u201cw3m\u201d because of the write command):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-ex-define-cmd \"mu[4e]\" 'mu4e)<\/pre><\/div>\n<p>Here the hooks that are closest to common vim autocommands are listed. This is fairly incomplete, and in many cases there are not direct\/exact equivalents, or common use cases of the vim hooks are unnecessary (e.g. you don\u2019t need to use hooks in emacs to make keybindings for specific filetypes). See <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Standard-Hooks.html\" rel=\"nofollow\">here<\/a> for the standard hooks that are part of emacs.<\/p>\n<table><tbody><tr><th>vim<\/th><th>emacs<\/th><\/tr><tr><td><code>InsertLeave<\/code><\/td><td><code>evil-insert-state-exit-hook<\/code><\/td><\/tr><tr><td><code>FileType python<\/code><\/td><td><code>python-mode-hook<\/code><\/td><\/tr><tr><td><code>BufReadPost<\/code><\/td><td><code>find-file-hook<\/code><\/td><\/tr><tr><td><code>BufWrite(Pre)<\/code><\/td><td><code>before-save-hook<\/code> or <code>write-file-hooks<\/code><\/td><\/tr><tr><td><code>BufWritePost<\/code><\/td><td><code>after-save-hook<\/code><\/td><\/tr><tr><td><code>FocusGained<\/code><\/td><td><code>focus-in-hook<\/code><\/td><\/tr><tr><td><code>FocusLost<\/code><\/td><td><code>focus-out-hook<\/code><\/td><\/tr><tr><td><code>VimLeave(Pre)<\/code><\/td><td><code>kill-emacs-hook<\/code><\/td><\/tr><\/tbody><\/table>\n<h2><a id=\"user-content-packagesplugins\" class=\"anchor\" aria-hidden=\"true\" href=\"#packagesplugins\"><\/a>Packages\/Plugins<\/h2>\n<p>Unlike vim, emacs has a standard way to install plugins (<code>package.el<\/code>). This has some upsides such as allowing a package author to specify dependencies. There are also some differences vim users might consider to be downsides. For example, you normally install packages through a package repository such as <a href=\"https:\/\/melpa.org\/#\/\" rel=\"nofollow\">MELPA<\/a>. If you want to install a package that is not in a package repository immediately (without having to add it yourself), you can use <a href=\"https:\/\/github.com\/raxod502\/straight.el\">straight.el<\/a> or <a href=\"https:\/\/github.com\/dimitri\/el-get\">elget<\/a> or <a href=\"https:\/\/github.com\/quelpa\/quelpa\">quelpa<\/a> to grab it from the source repository like you would with a vim plugin manager. Quelpa and straight.el are also useful if you want the latest version of a package from MELPA (or with your own recipe). MELPA builds packages daily, but sometimes you may want the latest commit for testing a bug fix. You can, of course, always manually clone a repo and put it in your <code>load-path<\/code>. For comparison, <code>package.el<\/code> and other emacs package managers additionally compile all elisp files and generate autoloads from autoload cookies. I personally use and recommend <a href=\"https:\/\/github.com\/raxod502\/straight.el\">straight.el<\/a> which can use MELPA\u2019s recipes but also allows you to specify your own and is trivial to switch to if you are using <a href=\"https:\/\/github.com\/jwiegley\/use-package\">use-package<\/a>.<\/p>\n<p>For basic functionality, you can use <code>package-list-packages<\/code> (or just <code>list-packages<\/code>) to view and install available packages or just  <code>package-install<\/code>.<\/p>\n<h2><a id=\"user-content-debugging-your-init-file\" class=\"anchor\" aria-hidden=\"true\" href=\"#debugging-your-init-file\"><\/a>Debugging Your Init File<\/h2>\n<p>You can start emacs with the <code>--debug-init<\/code> flag when there is some problem in your init to put you in the debugger with a backtrace. You can use <a href=\"https:\/\/emacs.stackexchange.com\/questions\/7852\/show-line-number-on-error?lq=1\" rel=\"nofollow\">wasamasa\u2019s hack<\/a> to have the line number where the error was encountered displayed as well.<\/p>\n<p><a href=\"https:\/\/github.com\/flycheck\/flycheck\">Flycheck<\/a> can help to prevent some errors. If you don\u2019t have flycheck installed, you can also byte-compile your init file to get information about detectable errors and jump to them using <code>byte-compile-file<\/code>; byte-compiling your init file will also give you other nice information such as telling you when you\u2019re using obsolete functions\/variables. The emacs manual <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Init-File.html\" rel=\"nofollow\">does not recommend<\/a> using a byte-compiled init file, so you may want to remove the corresponding \u201celc\u201d file afterwards if you do this.<\/p>\n<h2><a id=\"user-content-is-emacs-slow-to-start\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-emacs-slow-to-start\"><\/a>Is emacs slow to start?<\/h2>\n<p>No, people\u2019s init files are usually the problem. To test emacs\u2019 startup speed, you can start it without loading your init file using <code>emacs -Q<\/code>. Even with hundreds of packages, the startup time shouldn\u2019t be increased very much if you properly defer the loading of your packages when possible.<\/p>\n<p>Deferring a package from loading often just involves not putting a <code>(require 'package)<\/code> in your config. When you install a package through <code>package.el<\/code> (and most alternative package managers), autoloads are automatically generated for functions that have autoload cookies (<code>;;;###autoload<\/code>). This means that if you bind a key to one of these autoloaded commands, the corresponding package will be loaded when you first press that key. Major modes should normally only be loaded when a file of the corresponding type is first loaded. Minor mode activation commands also should be autoloaded. A common way of activating minor modes is by using hooks (e.g. <code>(add-hook 'emacs-lisp-mode #'lispy-mode)<\/code>).<\/p>\n<p>Even if your init is not too optimized, you can you can use emacs\u2019 server functionality so that you only need to start emacs once. You can start the server either putting <code>(server-start)<\/code> in your init file or by using <code>emacsclient<\/code> or <code>emacs<\/code> (with the <code>--daemon<\/code> flag) to create it. You can connect to a server using <code>emacsclient<\/code>. I personally use <code>emacsclient<\/code> as my EDITOR and have a key bound to <code>emacsclient -a \"\" -c<\/code>, which will open a new graphical emacs frame and start the server if it isn\u2019t already running. See the <code>emacsclient<\/code> manpage and the <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Emacs-Server.html\" rel=\"nofollow\">corresponding emacs manual section<\/a> for more information.<\/p>\n<p>For a convenient way to control package loading, see <a href=\"https:\/\/github.com\/jwiegley\/use-package\">use-package<\/a>. For a way to profile your init file(s), see <a href=\"https:\/\/github.com\/dholm\/benchmark-init-el\">benchmark-init<\/a> and <a href=\"https:\/\/github.com\/jschaf\/esup\">esup<\/a>.<\/p>\n<h2><a id=\"user-content-how-do-i-improve-emacs-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-improve-emacs-performance\"><\/a>How do I improve emacs\u2019 performance?<\/h2>\n<p>If you\u2019re encountering lag while using emacs, it\u2019s likely due to part of your configuration. A common culprit for slowdown is <code>linum-mode<\/code>. <code>nlinum<\/code> is a faster alternative, and emacs now has line numbers builtin which should be preferred (see the <code>display-line-numbers<\/code> variable).<\/p>\n<p>I\u2019ve also found that <code>git-gutter<\/code>, for example, can cause major slowdowns in large buffers with a lot of changes. I\u2019ve heard that <code>fic-mode<\/code> can also cause problems. In really large files, you may need to disable some of your minor modes, switch to fundamental mode, or use <a href=\"https:\/\/github.com\/m00natic\/vlfi\">vlf<\/a>. If you\u2019re having trouble quickly finding the culprit of slowdowns, you should try profiling with <code>profiler-start<\/code>.<\/p>\n<h2><a id=\"user-content-does-emacs-have-vim-like-tabs-distinct-window-configurations\" class=\"anchor\" aria-hidden=\"true\" href=\"#does-emacs-have-vim-like-tabs-distinct-window-configurations\"><\/a>Does emacs have vim-like tabs (distinct window configurations)?<\/h2>\n<p>No, but there are plenty of packages that add this feature. Elscreen is often recommended, but it is limited to 10 tabs\/screens, old, and not as good as the alternatives in my opinion. I personally use <a href=\"https:\/\/github.com\/pashinin\/workgroups2\">workgroups2<\/a>. It probably has the most features compared with alternatives, but it is unmaintained, so I\u2019d probably recommend using <a href=\"https:\/\/github.com\/wasamasa\/eyebrowse\">eyebrowse<\/a> instead. There are other alternatives listed in the eyebrowse readme as well.<\/p>\n<h2><a id=\"user-content-whats-the-equivalent-of-nnoremap-y-y\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-the-equivalent-of-nnoremap-y-y\"><\/a>What\u2019s the equivalent of <code>nnoremap Y y$<\/code>?<\/h2>\n<p>You can of course bind <code>Y<\/code> to simulate <code>y$<\/code> or redefine the operator, but evil has an option for this builtin. You can set <code>evil-want-Y-yank-to-eol<\/code> to a non-nil value before loading emacs to make this change.<\/p>\n<p>You might also want to look at the other <code>evil-want<\/code> variables in <code>evil-vars.el<\/code> such as <code>evil-want-change-word-to-end<\/code>.<\/p>\n<h2><a id=\"user-content-whats-the-equivalent-of-nnoremap-n-nzz\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-the-equivalent-of-nnoremap-n-nzz\"><\/a>What\u2019s the equivalent of <code>nnoremap n nzz<\/code>?<\/h2>\n<p>You can advise <code>evil-search-next<\/code> to have the command <code>zz<\/code> is bound to run afterwards.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun my-center-line (&amp;rest _)\n  (evil-scroll-line-to-center nil))\n(advice-add 'evil-search-next :after #'my-center-line)<\/pre><\/div>\n<p>You could advise several commands at once like this using <code>dolist<\/code>.<\/p>\n<h2><a id=\"user-content-whats-the-equivalent-of-inoremap-jk-escape\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-the-equivalent-of-inoremap-jk-escape\"><\/a>What\u2019s the equivalent of <code>inoremap jk &lt;escape&gt;<\/code>?<\/h2>\n<p>As this is not possible by default with emacs\u2019 keybinding system, you have to use one of a few <a href=\"https:\/\/github.com\/noctuid\/general.el#mapping-under-non-prefix-keys\">workarounds<\/a>.<\/p>\n<h2><a id=\"user-content-whats-the-equivalent-of-nnoremap-cw-ciw\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-the-equivalent-of-nnoremap-cw-ciw\"><\/a>What\u2019s the equivalent of <code>nnoremap cw ciw<\/code>?<\/h2>\n<p>This is also not possible by default. See the previous link.<\/p>\n<h2><a id=\"user-content-whats-the-equivalent-of-vim-unimpaireds-cow\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-the-equivalent-of-vim-unimpaireds-cow\"><\/a>What\u2019s the equivalent of vim-unimpaired\u2019s <code>cow<\/code>?<\/h2>\n<p>This is also not possible by default. See the previous link.<\/p>\n<h2><a id=\"user-content-how-do-i-copy-and-paste-tofrom-the-clipboard-in-terminal-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-copy-and-paste-tofrom-the-clipboard-in-terminal-emacs\"><\/a>How do I copy and paste to\/from the clipboard in terminal emacs?<\/h2>\n<p>For osx (with pbcopy) and linux (with xclip), you can install <code>xclip.el<\/code> for this functionality. I have this in my configuration to turn it on when I open emacs in a terminal:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun noct:conditionally-toggle-xclip-mode ()\n  (if (display-graphic-p)\n      (if (bound-and-true-p xclip-mode)\n          (xclip-mode -1))\n    (xclip-mode)))\n(noct:conditionally-toggle-xclip-mode)\n(add-hook 'focus-in-hook\n          #'noct:conditionally-toggle-xclip-mode)<\/pre><\/div>\n<p>If you don\u2019t use emacsclient for terminal instances, you don\u2019t need to use a hook at all. If you do use emacsclient for both graphical and terminal instances, then this should work in theory if your terminal supports <code>focus-in-hook<\/code> (e.g. st and kitty). That said, even though the mode is correctly toggled for me, it doesn\u2019t work in an emacsclient terminal frame unfortunately.<\/p>\n<h2><a id=\"user-content-can-i-have-better-completion-in-the-ex-command-line\" class=\"anchor\" aria-hidden=\"true\" href=\"#can-i-have-better-completion-in-the-ex-command-line\"><\/a>Can I have better completion in the ex command line?<\/h2>\n<p><code>ivy-mode<\/code>, for example, does work in the ex command line, but it must be manually triggered. Because of how completion in the command line works, there is no way as far as I\u2019m aware to have automatic completion popups. Company does work with <code>q:<\/code>, but by default, the completions it suggests may not be too useful.<\/p>\n<h2><a id=\"user-content-how-do-i-prevent-parentheses-becoming-unbalanced-in-my-init-file\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-prevent-parentheses-becoming-unbalanced-in-my-init-file\"><\/a>How do I prevent parentheses becoming unbalanced in my init file?<\/h2>\n<p>The simplest way is to install and use a package like <a href=\"https:\/\/github.com\/luxbock\/evil-cleverparens\">evil-cleverparens<\/a> or <a href=\"https:\/\/github.com\/noctuid\/lispyville\">lispyville<\/a> that will prevent evil\u2019s operators from unbalancing parentheses. Lispyville only remaps evil\u2019s operators by default, so you can ignore its other functionality (and lispy too) if you want.<\/p>\n<h2><a id=\"user-content-how-can-i-have-relative-line-numbers\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-can-i-have-relative-line-numbers\"><\/a>How can I have relative line numbers?<\/h2>\n<p>Using builtin line numbers is now the best solution. Here\u2019s my configuration that mimics <a href=\"https:\/\/github.com\/myusuf3\/numbers.vim\">numbers.vim<\/a>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq-default display-line-numbers 'visual\n              display-line-numbers-widen t\n              ;; this is the default\n              display-line-numbers-current-absolute t)\n(defun noct:relative ()\n  (setq-local display-line-numbers 'visual))\n(defun noct:absolute ()\n  (setq-local display-line-numbers t))\n(add-hook 'evil-insert-state-entry-hook #'noct:absolute)\n(add-hook 'evil-insert-state-exit-hook #'noct:relative)\n;; example of customizing colors\n(custom-set-faces '(line-number-current-line ((t :weight bold\n                                                 :foreground \"goldenrod\"\n                                                 :background \"slate gray\"))))<\/pre><\/div>\n<h2><a id=\"user-content-does-emacs-have-support-for-folding\" class=\"anchor\" aria-hidden=\"true\" href=\"#does-emacs-have-support-for-folding\"><\/a>Does emacs have support for folding?<\/h2>\n<p>Yes, evil has integration with various emacs \u201cfolding\u201d mechanisms builtin (such as origami, hideshow, and outline-mode\/org-mode\/markdown-mode; see <code>evil-fold-list<\/code>). Not all of vim\u2019s <code>z<\/code> keys will work though.<\/p>\n<p>For manual creation of folds from selected regions, there are the <a href=\"https:\/\/github.com\/mrkkrp\/vimish-fold\">vimish-fold<\/a> and <a href=\"https:\/\/github.com\/alexmurray\/evil-vimish-fold\">evil-vimish-fold<\/a> packages.<\/p>\n<h2><a id=\"user-content-why-dont-keys-defined-with-evil-define-key-work-immediately\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-dont-keys-defined-with-evil-define-key-work-immediately\"><\/a>Why don\u2019t keys defined with <code>evil-define-key<\/code> work (immediately)?<\/h2>\n<p>This has been a known problem for a while (see <a href=\"https:\/\/github.com\/emacs-evil\/evil\/issues\/130\">issue 130<\/a> and especially <a href=\"https:\/\/github.com\/emacs-evil\/evil\/issues\/301\">issue 301<\/a>, which explains some of the issues with <code>evil-define-key<\/code>). This doesn\u2019t happen for most modes, but when it does happen, it\u2019s annoying.<\/p>\n<p>There are several possible workarounds. You can use the mode\u2019s hook to either bind the keys locally with <code>evil-local-set-key<\/code> as shown in the <a href=\"#buffer-local-keybindings\">Buffer Local Keybindings<\/a> section. A more direct solution would be to continue to use <code>evil-define-key<\/code> and to use the hook to call <code>evil-normalize-keymaps<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'org-src-mode-hook #'evil-normalize-keymaps)<\/pre><\/div>\n<p>The other way would be to use <code>evil-define-minor-mode-key<\/code> which was introduced specifically as a result of this issue:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-define-minor-mode-key 'normal 'org-src-mode\n  (kbd \"SPC '\") 'org-edit-src-exit)\n;; `evil-define-key' with a quoted symbol instead of a keymap works the same\n(evil-define-key 'normal 'org-src-mode\n  (kbd \"SPC '\") 'org-edit-src-exit)<\/pre><\/div>\n<ul><li>Add a section on configuring undo (e.g. <code>evil-want-fine-undo<\/code>, <code>evil-with-undo<\/code>, <code>evil-with-single-undo<\/code>, etc.)<\/li>\n  <li>Add section explaining evil interactive codes<\/li>\n  <li>Explain all command properties<\/li>\n  <li>Add section on evil\u2019s supported\/missing functionality (e.g. numerical prefixes before operators aren\u2019t repeated and <code>:put<\/code>, <code>:next<\/code>, <code>:rewind<\/code>, etc. are missing)<\/li>\n  <li>Mention <code>evil-ex-map<\/code><\/li>\n  <li>Mention <code>evil-without-repeat<\/code><\/li>\n<\/ul><\/article>","created_at":"2018-11-28T22:44:53+0000","updated_at":"2018-11-28T22:44:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":59,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/4250696?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":727,"title":"Mastering Key Bindings in Emacs","url":"https:\/\/www.masteringemacs.org\/article\/mastering-key-bindings-emacs","content":"<p>By <em>Mickey Petersen<\/em><\/p><img src=\"https:\/\/www.masteringemacs.org\/static\/img\/fleuron2.gif\" class=\"img-responsive center-block\" alt=\"image\" \/><p class=\"text-justify drop-cap\">Altering the key bindings in Emacs should not, on the face of it, be a difficult task. But there\u2019s a reason why the Emacs manual has dedicated 30-odd pages to describing, in great detail, all the subtleties and nuances of how to bind keys.<\/p><p class=\"text-justify\">To save you the time of reading all of that, I\u2019ve written a guide that covers what you need to know to bind keys to even complex commands, and a set of templates you can use in your own code.<\/p><p class=\"text-justify\"><em>This guide will assume that you have little or no knowledge of elisp<\/em><\/p><h2 id=\"what-makes-up-a-key-bind\">What makes up a key bind?<\/h2><p class=\"text-justify\">Emacs is \u201cself-hosting\u201d and, depending on who you ask, close to achieving sentience. Almost all of Emacs is written in Emacs-Lisp (weighing in at roughly 1.2 million lines of code!), and most of the complex elisp functions you use are in turn built with simpler building blocks and so on, right down to the core C source code layer.<\/p><p class=\"text-justify\">Emacs\u2019s key engine is very similar, because Emacs lets you control almost every facet of <em>self-insertion<\/em> and other key binds. For instance, when you press \u201cf\u201d it runs <code>self-insert-command<\/code>, a special command that repeats the last typed key <em>N<\/em> times. So if you type <code>C-u 10 f<\/code> you will see <code>ffffffffff<\/code> printed on your screen. As you can see, every keypress \u2013 even elemental ones like typing a character on the screen \u2013 has an associated key binding that you can modify or even remove.<\/p><h3 id=\"keymaps\">Keymaps<\/h3><p class=\"text-justify\">A <em>keymap<\/em> is an internal data structure used by Emacs to store keys and their associated actions. Keymaps are rarely modified directly, but through a set of commands that manipulate the data structure for you. Most Emacs users will never interact with keymaps aside from indirectly assigning keys to them.<\/p><p class=\"text-justify\">Every buffer and most major and minor modes have a keymap, and that keymap defines what the keys do when key sequences are sent to that buffer. Keys can be divided into three categories: <em>undefined<\/em>, <em>prefix key<\/em>, or <em>complete key<\/em>. <em>Undefined<\/em> is self-explanatory: it does no operation when it is invoked. <em>Prefix keys<\/em> are keys like <code>C-x<\/code> and <code>C-c<\/code>; they are parts of a <em>complete key<\/em>, and each constituent part of a <em>prefix key<\/em> is made up of its own keymap. A <em>complete key<\/em> is a command that, when input, executes its associated command.<\/p><p class=\"text-justify\">It is possible for Emacs to enumerate all the active minor and major mode key bindings in a buffer by typing <code>C-h m<\/code>. This command is <em>very<\/em> useful if you want to learn more about what a major or minor mode can do. Likewise, you can type out parts of a complete key (say <code>M-s<\/code>) and then type <code>C-h<\/code> to get a list of all keys that belong to that prefix.<\/p><h3 id=\"key-bind-commands\">Key Bind Commands<\/h3><p class=\"text-justify\">There are several ways you can define (or undefine) keys, as the table below shows, but in reality there are dozens of ways you can hack the keymaps.<\/p><dl><dt><code>(define-key KEYMAP KEY DEF)<\/code><\/dt><dd><p class=\"text-justify\">Defines a key against a keyboard map. Use this if you want to change a keymap that isn\u2019t the current buffer map.<\/p><\/dd><dt><code>(local-set-key KEY COMMAND)<\/code><\/dt><dd><p class=\"text-justify\">Binds a key to the local keymap used by the active buffer, unlike <code>define-key<\/code> which takes an explicit keymap to bind a key against.<\/p><\/dd><dt><code>(global-set-key KEY COMMAND)<\/code><\/dt><dd><p class=\"text-justify\">Binds a key to the global keymap, making it available in all buffers (with a caveat \u2013 see below.)<\/p><\/dd><dt><code>(global-unset-key KEY)<\/code><\/dt><dd><p class=\"text-justify\">Removes KEY from the global keymap<\/p><\/dd><dt><code>(local-unset-key KEY)<\/code><\/dt><dd><p class=\"text-justify\">Removes KEY from the active, local keymap.<\/p><\/dd><\/dl><h3 id=\"representing-keys-in-code\">Representing Keys in Code<\/h3><p class=\"text-justify\">In order to actually bind a key you must first tell Emacs what key you intend to use. Unfortunately there\u2019s more than one way of representing keys in Emacs: as a <em>string<\/em>, or as a <em>vector<\/em>. We won\u2019t concern ourselves with either, as we will use a handy macro built in to Emacs called <code>kbd<\/code>.<\/p><p class=\"text-justify\">The <code>kbd<\/code> macro translates a human-readable key into a format Emacs can understand.<\/p><p class=\"text-justify\">One important point to note is that you <strong>must<\/strong> surround <em>function<\/em> and <em>navigation<\/em> keys with <code>&lt;<\/code> and <code>&gt;<\/code>. Those keys include F-keys, arrow keys and home row keys, like so: <code>&lt;home&gt;<\/code>, <code>&lt;f8&gt;<\/code> and <code>&lt;down&gt;<\/code>. But if you want represent the key <code>C-c p<\/code> then write <code>(kbd \"C-c p\")<\/code>.<\/p><h4 id=\"remapping-commands\">Remapping Commands<\/h4><p class=\"text-justify\">You can tell Emacs that you want to <em>replace<\/em> all keys pointing to a certain command with one of your own choosing by using the <code>remap<\/code> event; this should be done instead of passing a key to the key bind function you are using. This is arguably the best way of replacing existing commands with your own as Emacs will automagically handle the key reassignment in the background.<\/p><p class=\"text-justify\">Example:<\/p><pre class=\"sourceCode commonlisp\">(define-key (current-global-map) [remap kill-line] 'my-homemade-kill-line)<\/pre><p class=\"text-justify\">Here I globally remap all key binds that point to <code>kill-line<\/code> to <code>my-homemade-kill-line<\/code>.<\/p><p class=\"text-justify\">For some more hands-on examples read my article Fixing the mark commands in transient mark mode.<\/p><h3 id=\"reserved-keys\">Reserved Keys<\/h3><p class=\"text-justify\">You can pick any keyboard combination you desire \u2013 even if that key bind is already taken, so be careful \u2013 but Emacs has set aside certain keys for use by users. Generally, all keys prefixed with <code>C-c ?<\/code> (where <em>?<\/em> is a single character) are reserved for you, and you alone. In practice most third-party packages don\u2019t give a hoot and will gladly stuff their own key binds in there.<\/p><p class=\"text-justify\">The other set of reserved keys are the F-keys from <code>F5<\/code> and onwards. The other two prefix keys reserved to you are <em>hyper<\/em> and <em>super<\/em>. They are remnants from ancient keyboards used in the 80s, but live on today in Emacs. Most PC-compatible keyboards won\u2019t have a <em>super<\/em> or <em>hyper<\/em> key so some people rebind the <em>Windows key<\/em> and the <em>Application Context<\/em> key to be <em>hyper<\/em> and <em>super<\/em> instead.<\/p><p class=\"text-justify\">Anyway, if you want to use <em>hyper<\/em> then use the prefix key <code>H-<\/code> (e.g., <code>H-q<\/code>) and if you want <em>super<\/em> use the prefix key <code>s-<\/code>. <strong>Take note of the lower case \u2018s\u2019!<\/strong><\/p><p class=\"text-justify\">In Windows you can add this to to your .emacs to enable <em>hyper<\/em> and <em>super<\/em>:<\/p><pre class=\"sourceCode commonlisp\">(setq w32-apps-modifier 'hyper)\n(setq w32-lwindow-modifier 'super)\n(setq w32-rwindow-modifier 'hyper)<\/pre><p class=\"text-justify\">In X you\u2019ll have to play around with <em>xmodmap<\/em> or your own tool of choice.<\/p><h4 id=\"keymap-lookup-order\">Keymap Lookup Order<\/h4><p class=\"text-justify\">Emacs will look for keys in a certain order, and that order I have described below. Keep in mind that only <em>active<\/em> keymaps are used, and that the order is top-to-bottom; the first \u201cmatch\u201d is used, subject to criteria that we don\u2019t care about.<\/p><ol><li><code>overriding-terminal-local-map<\/code> for terminal-specific key binds.<\/li><li><code>overriding-local-map<\/code> for keys that should override all other local keymaps. Be VERY careful if you use this!<\/li><li><em>Keymap char property at point<\/em> for keymaps that are local to the character point is at. This is used for stuff like fields in yasnippet and the customize dialog.<\/li><li><code>emulation-mode-map-alists<\/code> for advanced multi-mode keymap management<\/li><li><code>minor-mode-overriding-map-alist<\/code> for overriding the keymaps used by minor modes in major modes.<\/li><li><code>minor-mode-map-alist<\/code> is exactly like the <em>overriding<\/em> version above, but the preferred means of specifying the keymaps for minor modes.<\/li><li><em>Keymap text property at point<\/em> is like the one above for char properties but is for text properties only.<\/li><li><code>current-local-map<\/code> for keymaps defined in the buffers\u2019 current local map<\/li><li><code>current-global-map<\/code> is the last place Emacs will look for key binds and it is for the global ones.<\/li><\/ol><p class=\"text-justify\">All you care about is that <em>minor mode<\/em> keys come before <em>local<\/em> keys, and they in turn are checked before <em>global<\/em> keys.<\/p><h4 id=\"global-vs-local\">Global vs Local<\/h4><p class=\"text-justify\">A global key is functionally identical to that of a local one, except it is declared in a \u201cglobal\u201d keymap, governed by the <code>current-global-map<\/code> function (but usually it points to the default, the <code>global-map<\/code> variable.) Therefore, it is possible for you to define a global key simply by passing the <code>current-global-map<\/code> function to <code>define-key<\/code>. The other \u2013 often better \u2013 alternative is to use the designated function, <code>global-set-key<\/code>.<\/p><p class=\"text-justify\">In a similar vein, to bind a local key you can use the \u201cdesignated\u201d function <code>local-set-key<\/code> or the more general <code>define-key<\/code>. Like the global map, there exists an equivalent function <code>current-local-map<\/code> that returns the keymap local to the buffer. It\u2019s important that you know that although the major mode in a buffer will define the bulk of the key binds used in a buffer, the minor modes often add, remove or change the key binds as they take precedence over the local keymap.<\/p><h2 id=\"defining-your-command\">Defining your Command<\/h2><p class=\"text-justify\">If you thought the mechanics of keymaps and keys were difficult, think again! Deciding on what you want to bind to those keys is even harder \u2013 especially so if you want the command to do very specific things, like switching to a specific buffer name. But I\u2019ve got a few tricks up my sleeve to help you cut out most of the elisp writing by letting Emacs do all the heavy lifting.<\/p><h3 id=\"invoking-a-command\">Invoking a command<\/h3><p class=\"text-justify\">Before I talk about the how and why of Emacs commands it is important that I mention how Emacs handles <em>interactive functions<\/em> \u2013 commonly known as <em>commands<\/em>. An Emacs <em>command<\/em> is one that has the <code>(interactive ...)<\/code> statement at the top of a function body; this magic statement tells Emacs the command is interactive and that it is intended for consumption by users. That is also what determines if you can invoke a command through the <code>M-x<\/code> prompt.<\/p><p class=\"text-justify\">When you bind a function to a key it is important that you keep the above in mind, as you <strong>cannot<\/strong> invoke a non-interactive function through a key binding \u2013 it must be a <em>command<\/em>. Also, for a key definition to function it must invoke a command with <strong>no parameters<\/strong>. If it is your intent \u2013 as it so often is \u2013 to call a function with parameters you must wrap it in a little helper function, like a <code>lambda<\/code> expression or even a standard <code>defun<\/code>. I have prepared templates (see below) that show you how to do this.<\/p><h3 id=\"describing-the-command\">Describing the command<\/h3><p class=\"text-justify\">There are two ways of describing a command in elisp: the <em>manual<\/em> way, and the <em>smart<\/em> way. Let\u2019s start out with the manual way.<\/p><p class=\"text-justify\">The manual way is to use <code>C-h k KEY<\/code> to describe what <code>KEY<\/code> is bound to; the other way is to describe a function, <code>FUNCTION<\/code>, with <code>C-h f FUNCTION<\/code>. Because Emacs is a self-documenting editor all functions, variables, keys, etc. loaded in Emacs can be accessed by way of the <code>describe-xxx<\/code> commands (type <code>C-h C-h<\/code> to see them all!) and that is what most Emacs hackers use.<\/p><p class=\"text-justify\">But there\u2019s another way\u2026 the <em>smarter<\/em> way. It has one teensy-weensy little downside: it\u2019ll only work for <em>interactive<\/em> functions (the ones we call commands, remember?), and then only <em>complex<\/em> commands \u2013 that is, a command that needs user input from the minibuffer.<\/p><p class=\"text-justify\">The smarter way has a name, <code>repeat-complex-command<\/code>, and it is conveniently bound to the key <code>C-x ESC ESC<\/code>. When you type it, Emacs will ask you to <em>redo<\/em> the last typed command, but it will do so showing you the elisp expression to execute; so you can not only change it then-and-there and re-run it, but you can <em>copy it and use it in your own code!<\/em>.<\/p><p class=\"text-justify\">Here\u2019s a practical example: <code>C-M-% foo RET bar RET<\/code> \u2013 which does a <code>query-replace-regexp<\/code> replacing <em>foo<\/em> with <em>bar<\/em>. Now, if you type <code>C-x ESC ESC<\/code> you should see something similar to this (reformatted for clarity):<\/p><pre class=\"sourceCode commonlisp\">(query-replace-regexp \"foo\" \"bar\" nil\n   (if (and transient-mark-mode mark-active)\n       (region-beginning))\n   (if (and transient-mark-mode mark-active)\n       (region-end)))<\/pre><p class=\"text-justify\">Observe how Emacs has kindly filled in all the function arguments to <code>query-replace-regexp<\/code>, including the optional parameters. If you were to run that command again (by say pasting it into the prompt in <code>M-:<\/code>) you will be asked to interactively search and replace with the search term <em>foo<\/em> and its replacement <em>bar<\/em> already filled in!<\/p><p class=\"text-justify\">Here\u2019s another one, this time I type <code>C-x b *scratch*<\/code> to switch to the <code>*scratch*<\/code> buffer:<\/p><pre class=\"sourceCode commonlisp\">(switch-to-buffer \"*scratch*\")<\/pre><p class=\"text-justify\">And there you have it. Obviously the commands seen here are the <em>interactive<\/em> commands. Stuff like replacing strings in a buffer is normally done with specialist elisp functions that don\u2019t require user input, but that may not be a big deal if you are writing your own quick-and-dirty key binds.<\/p><h2 id=\"putting-it-all-together\">Putting it all together<\/h2><p class=\"text-justify\">Yay. You\u2019ve made it this far. I just need to explain a few more important concepts that confuse a lot of newbies: mode hooks and mode-specific keymaps, and then it\u2019s on to the templates.<\/p><p class=\"text-justify\">Most major and minor modes will <em>usually<\/em> set their keys once, when the module is first loaded; that\u2019s good news for us, as we can use <code>define-key<\/code> to add key definitions straight into their <em>mode map<\/em>, the map that holds keys relevant to that mode. Some modes are very advanced, and have several maps \u2013 isearch is a good example \u2013 and for those you will have to mess around with the source code (or possibly the info manual) to find out how to add keys.<\/p><p class=\"text-justify\">It\u2019s the modes that force you to use a <em>mode hook<\/em> that\u2019s the problem: for that you must set your keys when the mode is launched (using its <em>mode hook<\/em>), and that takes a little bit more work.<\/p><p class=\"text-justify\">It is a convention in Emacs that all major mode functions (the ones that activate the mode) end in <code>-mode<\/code> (e.g., <code>python-mode<\/code>), and it is also required that its <em>mode map<\/em> is named <code>xxxx-mode-map<\/code>, and its major mode hook <code>xxxx-mode-hook<\/code>.<\/p><h3 id=\"keymaps-1\">Keymaps<\/h3><p class=\"text-justify\">Modifying a keymap is dirt simple, you use <code>define-key<\/code> as I\u2019ve mentioned before. What\u2019s not so easy is determining what keymap to modify in the first place.<\/p><h4 id=\"listing-all-the-mode-maps\">Listing all the Mode Maps<\/h4><p class=\"text-justify\">If you type this Emacs will give you an apropos buffer with all the known mode maps that follow the major mode naming scheme:<\/p><p class=\"text-justify\"><code>C-u M-x apropos-variable RET -mode-map$ RET<\/code><\/p><p class=\"text-justify\">Amusingly, you can use the repeat-complex-command trick to get the expression for use in a key bind or custom function\u2026<\/p><h4 id=\"quick-keymap-example\">Quick Keymap Example<\/h4><p class=\"text-justify\"><em>This assumes you are using the built-in python mode that comes with GNU Emacs<\/em><\/p><p class=\"text-justify\">Let\u2019s say I want to extend <code>python-mode<\/code> by adding a key, I\u2019ll use <code>F12<\/code>, that switches to the python process buffer. Currently, that\u2019s bound to <code>C-c C-z<\/code>.<\/p><pre class=\"sourceCode commonlisp\">(define-key python-mode-map (kbd \"&lt;f12&gt;\") 'python-switch-to-python)<\/pre><p class=\"text-justify\">I had to use the <em>manual<\/em> way (<code>C-h k C-c C-z<\/code>) to determine the name of the command because the <em>smart<\/em> way wouldn\u2019t work: the command is not complex, as it does not require user input from the minibuffer.<\/p><h3 id=\"hooks\">Hooks<\/h3><p class=\"text-justify\">A <em>mode hook<\/em> has zero-or-more functions that are called when its mode is activated, like say when you open a file that uses that mode or when you change the major mode in a buffer.<\/p><p class=\"text-justify\">You can add a mode hook using the special function <code>add-hook<\/code> that takes the name of a hook (say <code>python-mode-hook<\/code>) and the name of a function to call.<\/p><h4 id=\"listing-all-the-mode-hooks\">Listing all the Mode Hooks<\/h4><p class=\"text-justify\">If you are unsure of the exact name of the mode hook, you can use this handy trick to list the ones Emacs can see:<\/p><p class=\"text-justify\"><code>C-u M-x apropos-variable RET -mode-hook$ RET<\/code>.<\/p><p class=\"text-justify\">That will show <strong>all<\/strong> mode hooks known to Emacs, including their docstring description. If you don\u2019t see your mode it may be because Emacs hasn\u2019t loaded it outright or it lacks the <em>autoload<\/em> keyword.<\/p><h4 id=\"quick-hook-example\">Quick Hook Example<\/h4><p class=\"text-justify\">Let\u2019s add a key (<code>C-c q<\/code> to run <code>M-x shell<\/code>) local to <code>python-mode<\/code> using a hook. For that to work we will need our special hook function, I\u2019ve named it <code>mp-add-python-keys<\/code>, and in it we need <code>local-set-key<\/code>, the function that adds a key to the active buffer\u2019s local map. Observe that the command will be called within <em>the context of the buffer that invoked the major mode<\/em>.<\/p><p class=\"text-justify\">Next is the command that tells Emacs that we want to add a mode hook to Python.<\/p><pre class=\"sourceCode commonlisp\">(defun mp-add-python-keys ()\n  (local-set-key (kbd \"C-c q\") 'shell))\n(add-hook 'python-mode-hook 'mp-add-python-keys)<\/pre><h2 id=\"templates\">Templates<\/h2><p class=\"text-justify\">Here\u2019s a bunch of templates for various use-cases that you can cut\u2019n\u2019paste and use in your own code. I recommend naming things sensibly, and giving them a docstring (in-code documentation that explains what the code does) as well. The best way to avoid accidentally overriding another function with the same name, I would suggest you use a moniker or prefix (I use <code>mp-<\/code>.)<\/p><p class=\"text-justify\">I suggest you read my article, Evaluating Elisp in Emacs, to learn how to evaluate and test the code you write!<\/p><h4 id=\"purpose\">Purpose<\/h4><p class=\"text-justify\">Required if you want to invoke non-interactive functions or functions\/commands with parameters. Use this functions\u2019 name in the command definition argument.<\/p><h4 id=\"definition\">Definition<\/h4><pre class=\"sourceCode commonlisp\">(defun my-function-name-here ()\n  (interactive)\n  ;;; Place your code below this line, but inside the bracket.\n  )<\/pre><h4 id=\"example\">Example<\/h4><p class=\"text-justify\">Displays the message \u201cHello, World\u201d in the echo area when invoked.<\/p><pre class=\"sourceCode commonlisp\">(defun mp-display-message ()\n  (interactive)\n  ;;; Place your code below this line, but inside the bracket.\n  (message \"Hello, World\")\n  )<\/pre><h3 id=\"basic-global-key-bind\">Basic Global Key Bind<\/h3><h4 id=\"purpose-1\">Purpose<\/h4><p class=\"text-justify\">Creates a global key bind available to all buffers.<\/p><h4 id=\"definition-1\">Definition<\/h4><pre class=\"sourceCode commonlisp\">(global-set-key (kbd \"key-bind-here\") 'interactive-command-here)<\/pre><h4 id=\"example-1\">Example<\/h4><p class=\"text-justify\">Binds <code>F1<\/code> to <code>M-x shell<\/code><\/p><pre class=\"sourceCode commonlisp\">(global-set-key (kbd \"&lt;f1&gt;\") 'shell)<\/pre><h3 id=\"add-key-definition-to-a-keymap\">Add key definition to a keymap<\/h3><h4 id=\"purpose-2\">Purpose<\/h4><p class=\"text-justify\">Adds a key and its associated command to a local keymap. Useful for extending major modes with your own custom key binds.<\/p><h4 id=\"definition-2\">Definition<\/h4><pre class=\"sourceCode commonlisp\">(define-key KEYMAP (kbd \"key-bind-here\") 'interactive-command-here)<\/pre><h4 id=\"example-2\">Example<\/h4><p class=\"text-justify\">Binds <code>C-c p<\/code> to <code>python-switch-to-python<\/code><\/p><pre class=\"sourceCode commonlisp\">(define-key python-mode-map (kbd \"C-c p\") 'python-switch-to-python)<\/pre><h3 id=\"complex-command-key-bind\">Complex Command Key Bind<\/h3><h4 id=\"purpose-3\">Purpose<\/h4><p class=\"text-justify\">Creates a global key bind that invokes multiple commands in a row. Use this to create compound keys or invoke commands that take parameters. Uses code from the <em>Function Template<\/em>.<\/p><h4 id=\"definition-3\">Definition<\/h4><pre class=\"sourceCode commonlisp\">(defun name-of-interactive-command-here ()\n  (interactive)\n  ;;; Insert your compound commands below\n  )\n(global-set-key (kbd \"key-bind-here\") 'name-of-interactive-command-here)<\/pre><h4 id=\"example-3\">Example<\/h4><p class=\"text-justify\">Switches to the *scratch* buffer and inserts \u201cHello, World\u201d where point is, and switches back to where it came from.<\/p><pre class=\"sourceCode commonlisp\">(defun switch-to-scratch-and-insert-text ()\n  (interactive)\n  (save-excursion\n    (set-buffer \"*scratch*\")\n    (insert \"Hello, World\")))\n(global-set-key (kbd \"C-c i\") 'switch-to-scratch-and-insert-text)<\/pre><h3 id=\"binding-keys-with-a-mode-hook\">Binding keys with a mode hook<\/h3><h4 id=\"purpose-4\">Purpose<\/h4><p class=\"text-justify\">Use this mode hook template to bind keys that won\u2019t work with a standard define-key template or that require local binding for other reasons. The advantage of a mode hook is that it also gives you the opportunity to set mode-specific settings like indentation, etc.<\/p><h4 id=\"definition-4\">Definition<\/h4><pre class=\"sourceCode commonlisp\">(defun my-hook-function ()\n  ;; add your code here. it will be called every\n  ;; time the major mode is run.\n  )\n(add-hook 'my-mode-hook 'my-hook-function)<\/pre><h4 id=\"example-4\">Example<\/h4><p class=\"text-justify\">Make the return key automatically indent on a newline in emacs-lisp-mode, and enable eldoc mode as well.<\/p><pre class=\"sourceCode commonlisp\">(defun enable-my-elisp-settings ()\n  (turn-on-eldoc-mode)\n  (local-set-key (kbd \"C-m\") 'newline-and-indent))\n(add-hook 'emacs-lisp-mode-hook 'enable-my-elisp-settings)<\/pre><h4 id=\"purpose-5\">Purpose<\/h4><p class=\"text-justify\">This template will remap all keys that <em>point<\/em> to a specific function; say you want to rebind <code>kill-line<\/code> which is bound to <code>C-k<\/code> but you want your code to only override the key binds that <code>kill-line<\/code> is <em>actually<\/em> bound to. Use this template to replace existing commands with those of your own, without worrying about explicitly rebinding each key.<\/p><h4 id=\"definition-5\">Definition<\/h4><p class=\"text-justify\">Note: to replace a global key, you must use <code>global-map<\/code> or call <code>current-global-map<\/code>.<\/p><pre class=\"sourceCode commonlisp\">(define-key keymap [remap original-function] 'my-own-function)<\/pre><h4 id=\"example-5\">Example<\/h4><p class=\"text-justify\">Example here is taken from my article on fixing the mark commands in transient mark mode. I remap the keys that point to <code>exchange-point-and-mark<\/code> to my own function <code>exchange-point-and-mark-no-activate<\/code>.<\/p><pre class=\"sourceCode commonlisp\">(defun exchange-point-and-mark-no-activate ()\n  \"Identical to \\\\[exchange-point-and-mark] but will not activate the region.\"\n  (interactive)\n  (exchange-point-and-mark)\n  (deactivate-mark nil))\n(define-key global-map [remap exchange-point-and-mark] 'exchange-point-and-mark-no-activate)<\/pre><h3 id=\"custom-prefixes\">Custom Prefixes<\/h3><h4 id=\"purpose-6\">Purpose<\/h4><p class=\"text-justify\">Creating a prefix is easy nowadays as you don\u2019t have to explicitly create your own prefix keymaps, provided you use <code>local-set-key<\/code> or <code>global-set-key<\/code>. Use custom prefixes to group or categorize your commands.<\/p><h4 id=\"definition-6\">Definition<\/h4><pre class=\"sourceCode commonlisp\">(global-set-key (kbd \"subkey_1 ... endkey_1\") 'my-command-1)\n(global-set-key (kbd \"subkey_1 ... endkey_2\") 'my-command-2)<\/pre><h4 id=\"example-6\">Example<\/h4><p class=\"text-justify\">Global keys that will insert either the time of the day, or the current date. Type <code>C-c i d<\/code> to insert the date; and <code>C-c i t<\/code> to insert the time. Type <code>C-c i C-h<\/code> to list all bound keys under the <code>C-c i<\/code> prefix.<\/p><pre class=\"sourceCode commonlisp\">(defun mp-insert-date ()\n  (interactive)\n  (insert (format-time-string \"%x\")))\n(defun mp-insert-time ()\n  (interactive)\n  (insert (format-time-string \"%X\")))\n(global-set-key (kbd \"C-c i d\") 'mp-insert-date)\n(global-set-key (kbd \"C-c i t\") 'mp-insert-time)<\/pre><h2 id=\"conclusion\">Conclusion<\/h2><p class=\"text-justify\">I have covered almost every facet of key binding that most Emacs users would care to know about. I know there\u2019s a lot of tutorials out there that essentially capture only what I have in the templates section, but for your edification I decided I wanted to cover <em>everything<\/em> (or near as well anyway) that make up a key definition.<\/p><p class=\"text-justify\">That I can dedicate <em>so much<\/em> to a topic that on the outset appears straightforward highlights the flexibility \u2013 and complexity \u2013 of Emacs. I hope the guide has taught you how to do one of the most frequently-asked things in Emacs.<\/p>","created_at":"2018-11-28T22:44:47+0000","updated_at":"2018-11-28T22:44:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"www.masteringemacs.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs"],"is_public":false,"id":726,"title":"noctuid\/general.el","url":"https:\/\/github.com\/noctuid\/general.el\/tree\/master","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/melpa.org\/#\/general\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/be373599d758041a30afb4c4950c3b079df4e161\/68747470733a2f2f6d656c70612e6f72672f7061636b616765732f67656e6572616c2d62616467652e737667\" alt=\"https:\/\/melpa.org\/packages\/general-badge.svg\" data-canonical-src=\"https:\/\/melpa.org\/packages\/general-badge.svg\" \/><\/a> <a href=\"https:\/\/travis-ci.org\/noctuid\/general.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/96076fda3f94775bd571802af47e2710f26a0ad6\/68747470733a2f2f7472617669732d63692e6f72672f6e6f63747569642f67656e6572616c2e656c2e7376673f6272616e63683d6d6173746572\" alt=\"https:\/\/travis-ci.org\/noctuid\/general.el.svg?branch=master\" data-canonical-src=\"https:\/\/travis-ci.org\/noctuid\/general.el.svg?branch=master\" \/><\/a><\/p>\n<p><a href=\"https:\/\/github.com\/noctuid\/general.el\"><img src=\"https:\/\/camo.githubusercontent.com\/b1298535828025bd6dada5abbb8661b43c7c7785\/687474703a2f2f692e696d6775722e636f6d2f535841363679372e706e67\" alt=\"http:\/\/i.imgur.com\/SXA66y7.png\" data-canonical-src=\"http:\/\/i.imgur.com\/SXA66y7.png\" \/><\/a><\/p>\n<blockquote>\n  <p>A general is a leader. \u2013 onioncheese<\/p>\n<\/blockquote>\n<h2><a id=\"user-content-2018-01-21-general-default--variables-are-obsolete\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-01-21-general-default--variables-are-obsolete\"><\/a>2018-01-21 <code>general-default-...<\/code> variables are obsolete<\/h2>\n<p><code>general-default-prefix<\/code>, <code>general-default-non-normal-prefix<\/code>, <code>general-default-global-prefix<\/code>, <code>general-default-states<\/code>, and <code>general-default-keymaps<\/code> still work. However, they will eventually be removed, so please switch to using <code>general-create-definer<\/code> if you want to use a definer with different defaults.<\/p>\n<h2><a id=\"user-content-2018-01-20-general-create-vim-definer-and-general-create-dual-vim-definer-have-been-removed\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-01-20-general-create-vim-definer-and-general-create-dual-vim-definer-have-been-removed\"><\/a>2018-01-20 <code>general-create-vim-definer<\/code> and <code>general-create-dual-vim-definer<\/code> have been removed<\/h2>\n<p><code>general-create-definer<\/code> should now be used instead as it is now capable of the same functionality (<code>general-evil-setup<\/code> now uses it). Additionally, <code>general-vim-definer-default<\/code> is obsolete and will be removed eventually. The second argument to <code>general-evil-setup<\/code> is no longer used and will also be removed eventually. The vim definers will now always set the default <code>:states<\/code> (and never the default <code>:keymaps<\/code>) because of the change below.<\/p>\n<h2><a id=\"user-content-2018-01-20-states-normal-is-now-the-same-as-keymaps-normal\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-01-20-states-normal-is-now-the-same-as-keymaps-normal\"><\/a>2018-01-20 <code>:states 'normal<\/code> is now the same as <code>:keymaps 'normal<\/code><\/h2>\n<p><code>:keymaps 'global :states 'normal<\/code> will now bind in <code>evil-normal-state-keymap<\/code> as opposed to the normal state auxiliary keymap of <code>(current-global-map)<\/code> (see <a href=\"#note-for-evil-users\">Note for Evil Users<\/a>). It is not recommended to bind in a state and <code>(current-global-map)<\/code>. If you want to prevent certain keys from being overridden, please use evil intercept keymaps instead.<\/p>\n<p>If you update general, please make sure that you are also using a recent version of evil.<\/p>\n<h2><a id=\"user-content-2018-01-20-general-simulate-keys-is-now-obsolete\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-01-20-general-simulate-keys-is-now-obsolete\"><\/a>2018-01-20: <code>general-simulate-keys<\/code> is now obsolete<\/h2>\n<p>Please switch to <code>general-key<\/code> or <code>general-simulate-key<\/code>. Note that keyword arguments have replaced the positional arguments of <code>general-simulate-keys<\/code>. <code>general-simulate-keys<\/code> will likely be removed sometime in the future.<\/p>\n<ul><li><a href=\"#about\">About<\/a><\/li>\n  <li><a href=\"#key-features\">Key Features<\/a><\/li>\n  <li><a href=\"#reading-recommendations\">Reading Recommendations<\/a><\/li>\n  <li><a href=\"#usage-recommendations-and-documentation-clarifications\">Usage Recommendations and Documentation Clarifications<\/a><\/li>\n  <li><a href=\"#basic-examples\">Basic Examples<\/a>\n    <ul><li><a href=\"#general-examples\">General Examples<\/a><\/li>\n      <li><a href=\"#evil-examples\">Evil Examples<\/a><\/li>\n      <li><a href=\"#switching-completely-to-general\">Switching Completely to General<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#general-define-key-details\">~general-define-key~ Details<\/a>\n    <ul><li><a href=\"#definitions\">Definitions<\/a><\/li>\n      <li><a href=\"#keyword-arguments\">Keyword Arguments<\/a>\n        <ul><li><a href=\"#predicates\">Predicates<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#keymapstate-aliases\">Keymap\/State Aliases<\/a><\/li>\n      <li><a href=\"#general-define-key-wrappers\">~general-define-key~ Wrappers<\/a>\n        <ul><li><a href=\"#positional-argument-wrappers\">Positional Argument Wrappers<\/a><\/li>\n          <li><a href=\"#mass-key-unbinding-wrapper\">Mass Key Unbinding Wrapper<\/a><\/li>\n          <li><a href=\"#creating-new-key-definers\">Creating New Key Definers<\/a><\/li>\n          <li><a href=\"#vim-like-definers\">Vim-like Definers<\/a><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#note-for-evil-users\">Note for Evil Users<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#override-keymaps-and-buffer-local-keybindings\">Override Keymaps and Buffer Local Keybindings<\/a><\/li>\n  <li><a href=\"#displaying-keybindings\">Displaying Keybindings<\/a><\/li>\n  <li><a href=\"#functionsmacros-to-aid-key-definition\">Functions\/Macros to Aid Key Definition<\/a>\n    <ul><li><a href=\"#disclaimer\">Disclaimer<\/a><\/li>\n      <li><a href=\"#simulating-keypresses\">Simulating Keypresses<\/a><\/li>\n      <li><a href=\"#mapping-under-non-prefix-keys\">Mapping Under Non-prefix Keys<\/a><\/li>\n      <li><a href=\"#choosing-definitions-based-on-predicates\">Choosing Definitions Based on Predicates<\/a><\/li>\n      <li><a href=\"#key-translation\">Key \u201cTranslation\u201d<\/a><\/li>\n      <li><a href=\"#automatic-key-unbinding\">Automatic Key Unbinding<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#non-keybinding-related-configuration-helpers\">Non-keybinding-related Configuration Helpers<\/a>\n    <ul><li><a href=\"#settings\">Settings<\/a><\/li>\n      <li><a href=\"#hooks-and-advice\">Hooks and Advice<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#integration-with-other-packages\">Integration with Other Packages<\/a>\n    <ul><li><a href=\"#use-package-keywords\">Use-package Keywords<\/a>\n        <ul><li><a href=\"#general-keyword\">:general Keyword<\/a><\/li>\n          <li><a href=\"#hook-keywords\">Hook Keywords<\/a>\n            <ul><li><a href=\"#ghook-keyword\">:ghook Keyword<\/a><\/li>\n              <li><a href=\"#gfhook-keyword\">:gfhook Keyword<\/a><\/li>\n            <\/ul><\/li>\n        <\/ul><\/li>\n      <li><a href=\"#use-with-key-chord\">Use with Key-chord<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#extended-definition-syntax\">Extended Definition Syntax<\/a>\n    <ul><li><a href=\"#autoloaded-keymaps\">\u201cAutoloaded\u201d Keymaps<\/a><\/li>\n      <li><a href=\"#which-key-integration\">Which Key Integration<\/a><\/li>\n      <li><a href=\"#evil-command-properties\">Evil Command Properties<\/a><\/li>\n      <li><a href=\"#user-defined-extended-definition-keywords\">User-defined Extended Definition Keywords<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#user-defined-key-definers\">User-defined Key Definers<\/a>\n    <ul><li><a href=\"#wrapping-evil-define-minor-mode-key\">Wrapping ~evil-define-minor-mode-key~<\/a><\/li>\n      <li><a href=\"#lispy-integration-wrapping-lispy-define-key\">Lispy Integration\/ Wrapping ~lispy-define-key~<\/a><\/li>\n      <li><a href=\"#worf-integration-wrapping-worf-define-key\">Worf Integration\/ Wrapping ~worf-define-key~<\/a><\/li>\n      <li><a href=\"#other-provided-definers\">Other Provided Definers<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#faq\">FAQ<\/a>\n    <ul><li><a href=\"#how-do-i-prevent-key-sequence-starts-with-non-prefix-key-errors\">How do I prevent <code>Key sequence starts with non-prefix key<\/code> errors?<\/a><\/li>\n      <li><a href=\"#why-dont-some-evil-keybindings-work-immediately\">Why don\u2019t some evil keybindings work (immediately)?<\/a><\/li>\n    <\/ul><\/li>\n<\/ul>\n<p><code>general.el<\/code> provides a more convenient method for binding keys in emacs (for both evil and non-evil users). Like <code>use-package<\/code>, which provides a convenient, unified interface for managing packages, <code>general.el<\/code> is intended to provide a convenient, unified interface for key definitions. While this package does implement some completely new functionality (such as the ability to make vim-style keybindings under non-prefix keys with an optional timeout), its primary purpose is to build on existing functionality to make key definition more clear and concise. <code>general-define-key<\/code> is user-extensible and supports defining multiple keys in multiple keymaps at once, implicitly wrapping key strings with <code>(kbd ...)<\/code>, using named prefix key sequences (like the leader key in vim), and much more.<\/p>\n<p>One advantage of using <code>general-define-key<\/code> (or a wrapper for it) even in cases where its extra functionality isn\u2019t necessary and doesn\u2019t significantly improve brevity is that all keybindings are recorded and can be displayed later with <code>general-describe-keybindings<\/code>.<\/p>\n<p>This manual explains the most relevant parts of every <code>general.el<\/code> feature. All user-facing functions, macros, and variables also have docstrings (e.g. accessible with <code>C-h f<\/code> or <code>C-h v<\/code>), so please consult these for further details.<\/p>\n<ul><li>Provides a single function, <code>general-define-key<\/code>, that is usable for all key definition; wrappers are provided as well<\/li>\n  <li>Does not hide important details of key definition (unlike <code>evil-leader.el<\/code>); users should be familiar with <code>define-key<\/code> and other definers (e.g. <code>evil-define-key(*)<\/code> for evil users) before using this package<\/li>\n  <li>Uses a syntax similar to <code>setq<\/code> for key definitions (like <code>evil-define-key<\/code>, <code>bind-map<\/code>, <code>evil-leader.el<\/code>, etc.; unlike <code>bind-key<\/code>)<\/li>\n  <li>Provides tight (and optional) integration with evil (unlike <code>bind-key<\/code>)<\/li>\n  <li><code>general-def<\/code> can act as a drop-in replacement for the following definers (see the documentation below for a minor caveat) (unique):\n    <ul><li><code>general-define-key<\/code> and <code>global-set-key<\/code> (no positional keymap argument)<\/li>\n      <li><code>define-key<\/code> and <code>evil-global-set-key<\/code> (positional argument for keymap)<\/li>\n      <li><code>evil-define-key<\/code> (positional argument for state and keymap)<\/li>\n    <\/ul><\/li>\n  <li>With the <code>:definer<\/code> keyword, <code>general-define-key<\/code> can be extended to use any key definition function (e.g. <code>evil-define-minor-mode-key<\/code>, <code>lispy-define-key<\/code>, etc.) (unique)<\/li>\n  <li>With \u201cextended\u201d definitions, user-created keywords can be added globally (in <code>general-define-key<\/code>) and locally (in an \u201cextended\u201d definition plist) to extend the behavior of <code>general-define-key<\/code> (unique)<\/li>\n  <li>Allows binding keys in multiple keymaps\/states at once (unlike <code>bind-key<\/code>)<\/li>\n  <li>Automatically wraps string keys and definitions with <code>kbd<\/code> (this behavior can be turned off for compatibility with <code>define-key<\/code>)<\/li>\n  <li>Allows using an arbitrary number of prefix keys or \u201cleaders\u201d of any length (but does not require prefix keys like) (unlike <code>evil-leader.el<\/code>)<\/li>\n  <li>Allows for automatically creating prefix commands (but does not require creating them like <code>bind-key<\/code> does)<\/li>\n  <li>Allows for buffer-local keybindings (unlike <code>local-set-key<\/code>)<\/li>\n  <li>Allows deferring keybindings until the specified keymap exists (no need to use <code>(with-)eval-after-load<\/code>) (like <code>evil-define-key<\/code>)<\/li>\n  <li>Allows displaying defined keys (like <code>bind-key.el<\/code>)<\/li>\n  <li>Provides integration with other packages such as <code>key-chord.el<\/code> and <code>which-key.el<\/code> (unique)<\/li>\n  <li>Provides other helpers for keybindings (unique):\n    <ul><li>A method for creating \u201cautoloaded\u201d keymaps (like <code>bind-key.el<\/code>)<\/li>\n      <li>A potentially better way to simulate keypresses (works with prefix args and for incomplete key sequences, i.e. a key bound to a keymap)<\/li>\n      <li>A method for binding under non-prefix keys with an optional timeout (like in vim; e.g. bind <code>jk<\/code> in insert mode without losing <code>j<\/code>)<\/li>\n      <li>A helper to create a menu item to dispatch to different definitions based on predicates<\/li>\n    <\/ul><\/li>\n  <li>Provides other helpers for configuration (e.g. more convenient functions for hooks and advice)<\/li>\n  <li>Is well tested (unlike <code>evil-leader.el<\/code>)<\/li>\n<\/ul>\n<p>Before using <code>general.el<\/code>, you should first be familiar with <code>define-key<\/code>, <code>global-set-key<\/code>, and <a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/emacs\/Key-Bindings.html\" rel=\"nofollow\">emacs\u2019 key binding system<\/a>. I recommend reading <a href=\"https:\/\/www.masteringemacs.org\/article\/mastering-key-bindings-emacs\" rel=\"nofollow\">Mastering Key Bindings in Emacs<\/a> if you are new to emacs. Also see <code>define-key<\/code>\u2019s help text for information on valid keys and definitions.<\/p>\n<p>If you are also using evil, you should first be familiar with how <code>evil-define-key<\/code>, <code>evil-define-minor-mode-key<\/code>, etc. work. If you are a new evil user, I\u2019d recommend looking at my <a href=\"https:\/\/github.com\/noctuid\/evil-guide\">evil guide<\/a>. The <a href=\"https:\/\/github.com\/noctuid\/evil-guide#keybindings-and-states\">Keybindings and States<\/a> section in particular may be useful<\/p>\n<p>A large number of issues opened on this repository are not specifically related to <code>general.el<\/code> (e.g. user usage or syntax errors that would also occur without <code>general-define-key<\/code>). While I don\u2019t mind generic questions about keybinding issues, you may save yourself some time if you first determine whether or not an issue is related to <code>general-define-key<\/code> by, if possible, testing with an equivalent <code>define-key<\/code>, <code>evil-define-key<\/code>, etc. statement.<\/p>\n<p>See also the <a href=\"#faq\">FAQ<\/a> for commonly asked questions.<\/p>\n<p>To facilitate extensibility and easy creation of wrappers, <code>general-define-key<\/code> uses keyword arguments to specify everything besides the key definitions, including for the <code>:states<\/code> and <code>:keymaps<\/code>. Since users will most often specify one or both of these keyword arguments, <code>general-define-key<\/code> is often less concise than <code>define-key<\/code> or <code>evil-define-key<\/code>. It is for this reason that it is recommended that <code>general-define-key<\/code> not be used directly. <code>general.el<\/code> provides wrappers around <code>general-define-key<\/code> that take positional arguments, and it is recommended that you use these instead of <code>general-define-key<\/code>. <code>general-create-definer<\/code> can also be used to create a new definer with certain default settings (e.g. prefix settings). For clarity and consistency, examples in the documentation usually use <code>general-define-key<\/code> unless the example is explicitly for a wrapper. However, <a href=\"#positional-argument-wrappers\">~general-def~<\/a> is recommended over <code>general-define-key<\/code> as it is more flexible and concise. Positional arguments are <i>optional but not required<\/i>, so <code>general-def<\/code> can basically act as a drop-in replacement for many key definers (including <code>general-define-key<\/code>, <code>define-key<\/code>, and <code>evil-define-key<\/code>). Note that <code>general-create-definer<\/code> and the <code>:general<\/code> keyword argument for <code>use-package<\/code> use <code>general-def<\/code>. I personally only use <code>general-def<\/code>.<\/p>\n<p>Since it is more common for commands to not be sharp quoted in key definitions, this package\u2019s examples use single quotes for commands. I personally prefer to always properly sharp quote functions, so commands in the actual non-example code are always sharp quoted.<\/p>\n<p>Although <code>general-define-key<\/code> will automatically defer keybindings until the specified keymaps exist, you can still use it with <code>with-eval-after-load<\/code> or use-package\u2019s <code>:config<\/code> keyword whenever it makes sense to. If you have a lot of keybindings, this could potentially shave some fraction of a second off of your startup time.<\/p>\n<h2><a id=\"user-content-general-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-examples\"><\/a>General Examples<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'general)\n;; * Global Keybindings\n;; `general-define-key' acts like `global-set-key' when :keymaps is not\n;; specified (because \":keymaps 'global\" is the default)\n;; kbd is not necessary and arbitrary amount of key def pairs are allowed\n(general-define-key\n \"M-x\" 'amx                             ; or 'smex\n \"C-s\" 'counsel-grep-or-swiper)\n;; * Mode Keybindings\n;; `general-define-key' is comparable to `define-key' when :keymaps is specified\n(general-define-key\n ;; NOTE: keymaps specified with :keymaps must be quoted\n :keymaps 'org-mode-map\n \"C-c C-q\" 'counsel-org-tag\n ;; ...\n )\n;; `general-def' can be used instead for `define-key'-like syntax\n(general-def org-mode-map\n  \"C-c C-q\" 'counsel-org-tag\n  ;; ...\n  )\n;; * Prefix Keybindings\n;; :prefix can be used to prevent redundant specification of prefix keys\n(general-define-key\n :prefix \"C-c\"\n ;; bind \"C-c a\" to 'org-agenda\n \"a\" 'org-agenda\n \"b\" 'counsel-bookmark\n \"c\" 'org-capture)\n;; for frequently used prefix keys, the user can create a custom definer with a\n;; default :prefix\n;; using a variable is not necessary, but it may be useful if you want to\n;; experiment with different prefix keys and aren't using `general-create-definer'\n(defconst my-leader \"C-c\")\n(general-create-definer my-leader-def\n  ;; :prefix my-leader\n  ;; or without a variable\n  :prefix \"C-c\")\n;; ** Global Keybindings\n(my-leader-def\n  \"a\" 'org-agenda\n  \"b\" 'counsel-bookmark\n  \"c\" 'org-capture)\n;; ** Mode Keybindings\n(my-leader-def\n  :keymaps 'clojure-mode-map\n  ;; bind \"C-c C-l\"\n  \"C-l\" 'cider-load-file\n  \"C-z\" 'cider-switch-to-repl-buffer)\n;; `general-create-definer' creates wrappers around `general-def', so\n;; `define-key'-like syntax is also supported\n(my-leader-def clojure-mode-map\n  \"C-l\" 'cider-load-file\n  \"C-z\" 'cider-switch-to-repl-buffer)\n;; * Settings\n;; change `auto-revert-interval' after autorevert has been loaded (`setq' will\n;; not work)\n(general-setq auto-revert-interval 10)<\/pre><\/div>\n<h2><a id=\"user-content-evil-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#evil-examples\"><\/a>Evil Examples<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'general)\n;; * Global Keybindings\n;; `general-define-key' acts like `evil-define-key' when :states is specified\n(general-define-key\n :states 'motion\n ;; swap ; and :\n \";\" 'evil-ex\n \":\" 'evil-repeat-find-char)\n;; same as\n(general-define-key\n :states 'motion\n \";\" 'evil-ex\n \":\" 'evil-repeat-find-char)\n;; `general-def' can be used instead for `evil-global-set-key'-like syntax\n(general-def 'motion\n  \";\" 'evil-ex\n  \":\" 'evil-repeat-find-char)\n;; alternative using `general-translate-key'\n;; swap ; and : in `evil-motion-state-map'\n(general-swap-key nil 'motion\n  \";\" \":\")\n;; * Mode Keybindings\n(general-define-key\n :states 'normal\n :keymaps 'emacs-lisp-mode-map\n ;; or xref equivalent\n \"K\" 'elisp-slime-nav-describe-elisp-thing-at-point)\n;; `general-def' can be used instead for `evil-define-key'-like syntax\n(general-def 'normal emacs-lisp-mode-map\n  \"K\" 'elisp-slime-nav-describe-elisp-thing-at-point)\n;; * Prefix Keybindings\n;; :prefix can be used to prevent redundant specification of prefix keys\n;; again, variables are not necessary and likely not useful if you are only\n;; using a definer created with `general-create-definer' for the prefixes\n;; (defconst my-leader \"SPC\")\n;; (defconst my-local-leader \"SPC m\")\n(general-create-definer my-leader-def\n  ;; :prefix my-leader\n  :prefix \"SPC\")\n(general-create-definer my-local-leader-def\n  ;; :prefix my-local-leader\n  :prefix \"SPC m\")\n;; ** Global Keybindings\n(my-leader-def\n  :keymaps 'normal\n  ;; bind \"SPC a\"\n  \"a\" 'org-agenda\n  \"b\" 'counsel-bookmark\n  \"c\" 'org-capture)\n;; `general-create-definer' creates wrappers around `general-def', so\n;; `evil-global-set-key'-like syntax is also supported\n(my-leader-def 'normal\n  \"a\" 'org-agenda\n  \"b\" 'counsel-bookmark\n  \"c\" 'org-capture)\n;; to prevent your leader keybindings from ever being overridden (e.g. an evil\n;; package may bind \"SPC\"), use :keymaps 'override\n(my-leader-def\n  :states 'normal\n  :keymaps 'override\n  \"a\" 'org-agenda)\n;; or\n(my-leader-def 'normal 'override\n  \"a\" 'org-agenda)\n;; ** Mode Keybindings\n(my-local-leader-def\n  :states 'normal\n  :keymaps 'org-mode-map\n  \"y\" 'org-store-link\n  \"p\" 'org-insert-link\n  ;; ...\n  )\n;; `general-create-definer' creates wrappers around `general-def', so\n;; `evil-define-key'-like syntax is also supported\n(my-local-leader-def 'normal org-mode-map\n  \"y\" 'org-store-link\n  \"p\" 'org-insert-link\n  ;; ...\n  )\n;; * Setings\n;; change evil's search module after evil has been loaded (`setq' will not work)\n(general-setq evil-search-module 'evil-search)<\/pre><\/div>\n<p>Vim-like definitions:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-evil-setup)\n;; * Global Keybindings\n;; all keywords arguments are still supported\n;; these are just wrappers around `general-def' that set a default :states\n(general-nmap\n  :prefix \"SPC\"\n  \"p\" 'helm-mini)\n;; bind in motion state (inherited by the normal, visual, and operator states)\n(general-mmap\n  \";\" 'evil-ex\n  \":\" 'evil-repeat-find-char)\n;; alternatively, for shorter names\n(general-evil-setup t)\n(mmap\n  \";\" 'evil-ex\n  \":\" 'evil-repeat-find-char)\n;; * Mode Keybindings\n(general-nmap\n  :keymaps 'emacs-lisp-mode-map\n  \"K\" 'elisp-slime-nav-describe-elisp-thing-at-point)\n;; same as\n(general-nmap emacs-lisp-mode-map\n  \"K\" 'elisp-slime-nav-describe-elisp-thing-at-point)\n<\/pre><\/div>\n<h2><a id=\"user-content-switching-completely-to-general\" class=\"anchor\" aria-hidden=\"true\" href=\"#switching-completely-to-general\"><\/a>Switching Completely to General<\/h2>\n<p>It is possible to gradually switch to using general by using it only for new configuration and slowly converting old configuration if desired. If you would like to quickly convert all keybindings in your init file to use general so that they show up with <code>general-describe-keybindings<\/code>, you can potentially use regexp replace. For example, you could use <code>M-&lt; C-M-% \\(global-set-key\\|define-key\\|evil-global-set-key\\|evil-define-key\\) RET general-def RET !<\/code>. The evil equivalent would be <code>:%s\/\\(global-set-key\\|define-key\\|evil-global-set-key\\|evil-define-key\\)\/general-def\/g<\/code>.<\/p>\n<p>There are two caveats. The old key definers all require using <code>kbd<\/code>. This means that you will either have to remove every <code>kbd<\/code> in these key definers (e.g. <code>:%s\/(kbd ?\\(.*?\\))\/\\1\/gc<\/code>; you should likely confirm whether each <code>kbd<\/code> should be removed) or set <code>general-implicit-kbd<\/code> to nil for the old configuration. Furthermore, <code>general-def<\/code> can only correctly replace definer statements where the first specified key is a string or vector. It will not work correctly to replace a definer that uses a variable or function for the first key (e.g. <code>(global-set-key my-key 'command)<\/code> cannot be replaced directly with <code>general-def<\/code>). To use general for definitions like this, you must either use the actual equivalent definer that <code>general-def<\/code> ends up using  (<code>general-define-key<\/code>, <code>general-emacs-define-key<\/code>, or <code>evil-define-key<\/code>) or explicitly separate the positional arguments from the first key with a bogus keyword argument (e.g. <code>(general-def :start-maps my-key 'command)<\/code>).<\/p>\n<p>If you decide to do this, please make sure that your configuration is backed up, and test this out to make sure that there are no errors before permanently changing your configuration.<\/p>\n<p>This package provides one main function, <code>general-define-key<\/code>, for key definitions for both evil and non-evil users. It is recommended you use the provided wrappers around it or create your own with <code>general-create-definer<\/code>, but first you should understand the keyword arguments provided by <code>general-define-key<\/code>.<\/p>\n<h2><a id=\"user-content-definitions\" class=\"anchor\" aria-hidden=\"true\" href=\"#definitions\"><\/a>Definitions<\/h2>\n<p>The only positional arguments for <code>general-define-key<\/code> are any number of key\/definition pairs. General supports all key and definition types supported by <code>define-key<\/code> (see its help text) as well as its own <a href=\"#extended-definition-syntax\">\u201cextended definitions\u201d<\/a>. Here are a few examples of definitions that aren\u2019t standard =\u201dstring key\u201d \u2018command= pairs:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; vector keys, including [t] and [remap] are supported\n(general-define-key\n :keymaps 'org-capture-mode-map\n [remap evil-save-and-close]          'org-capture-finalize\n [remap evil-save-modified-and-close] 'org-capture-finalize\n [remap evil-quit]                    'org-capture-kill)\n(general-define-key\n :states 'normal\n :keymaps 'org-capture-mode-map\n ;; keyboard macro definition\n \"RET\" \"C-c C-c\"\n ;; general.el extended definition\n \"SPC k\" '(org-capture-kill :which-key \"abort capture\"))<\/pre><\/div>\n<p><code>kbd<\/code> will automatically be called on every string key. <code>general-implicit-kbd<\/code> can be set to nil if you want to manually use <code>(kbd \"key\")<\/code>. This option is mainly provided to make it easy to transition to <code>general-define-key<\/code> or <code>general-def<\/code> from other key definers with search and replace and therefore only applies to <code>general-define-key<\/code> (and wrappers). <code>kbd<\/code> will always be called on string keys for other helpers such as <code>general-key<\/code>, <code>general-key-dispatch<\/code>, and <code>general-translate-key<\/code>.<\/p>\n<h2><a id=\"user-content-keyword-arguments\" class=\"anchor\" aria-hidden=\"true\" href=\"#keyword-arguments\"><\/a>Keyword Arguments<\/h2>\n<p><code>:prefix<\/code>, <code>:states<\/code>, and <code>:keymaps<\/code> are the most basic keyword arguments. By default, there is no prefix or state (each is nil), and the keymap is =\u2019global=. Each keymap can either be a quoted keymap, quoted <a href=\"#keymapstate-aliases\">keymap alias<\/a>, =\u2019global=, or =\u2019local=. This is the biggest contrast between <code>general-define-key<\/code> and other definers such as <code>define-key<\/code>, where the keymap is passed in directly. Note that the provided wrappers such as <code>general-def<\/code> do not require quoting keymaps. When the keymap is =\u2019local=, the key will be bound only in the current buffer (see <a href=\"#override-keymaps-and-buffer-local-keybindings\">here<\/a> for more details). When the keymap is =\u2019global=, the key will be bound in <code>(current-global-map)<\/code> (or the corresponding evil global map if <code>:states<\/code> is specified; see <a href=\"#note-for-evil-users\">Note for Evil Users<\/a> for more information).<\/p>\n<p><code>:states<\/code> and <code>:keymaps<\/code> can be lists or a single element, allowing the user to define keys for multiple evil states or keymaps simultaneously. This can be useful in certain situations to prevent redundancy.<\/p>\n<p>Using a different prefix for the insert and emacs states (or any state in <code>general-non-normal-states<\/code>) can be done with <code>:non-normal-prefix<\/code> or <code>:global-prefix<\/code>. By default, <code>:prefix<\/code> will apply to all keys, but if one (or both) of the other prefix keywords is specified, <code>:prefix<\/code> will only apply to evil states not listed in <code>general-non-normal-states<\/code>. This is also the case for the global evil keymaps such as <code>evil-normal-state-map<\/code>. <code>:non-normal-prefix<\/code> will always only apply to the non-normal states. <code>:global-prefix<\/code> will always apply to all keys. For example, this command will bind <code>SPC \/<\/code> to swiper in normal state and <code>M-SPC \/<\/code> to swiper in emacs and insert state:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :keymaps '(normal insert emacs)\n :prefix \"SPC\"\n :non-normal-prefix \"M-SPC\"\n \"\/\" 'swiper)<\/pre><\/div>\n<p>If you would like to create a named prefix keymap for your prefix keys, you can also specify <code>:prefix-command<\/code> and\/or <code>:prefix-map<\/code>. All prefix keys will then be bound to the prefix command or prefix keymap in the correct keymaps. If <code>:prefix-command<\/code> is specified, <code>define-prefix-command<\/code> will be used with <code>prefix-map<\/code> and <code>prefix-name<\/code> passed in as additional arguments to <code>define-prefix-command<\/code>. If only <code>:prefix-map<\/code> is specified, a prefix keymap alone will be created with a menu item\/prompt corresponding to <code>:prefix-name<\/code>. Note that existing prefix commands\/keymaps will not be redefined, so reevaluating a general.el form that uses <code>:prefix-command<\/code> or <code>:prefix-map<\/code> will not clear the previously created keymap.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :keymaps '(normal insert emacs)\n :prefix \"SPC\"\n :non-normal-prefix \"M-SPC\"\n :prefix-command 'my-prefix-command\n :prefix-map 'my-prefix-map\n \"\/\" 'swiper)<\/pre><\/div>\n<p>General is flexible in allowing you to choose how you write things, so if the above would be something you\u2019d use often, you could create a function with the above keyword arguments as defaults using <a href=\"#creating-new-key-definers\">~general-create-definer~<\/a> and write the definition like this:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(my-normal-and-insert-define-key \"\/\" 'swiper)<\/pre><\/div>\n<p>The <code>:infix<\/code> keyword can be used to sandwich keys in between all of the specified prefix keys and the keys in each mapping. This is mainly useful when using multiple prefix keywords and especially when using wrappers. For example, if you wanted to define several keys that were prefixed with <code>SPC g<\/code> in normal state and <code>M-SPC g<\/code> in insert state, you could use the previous wrapper with <code>:infix<\/code> instead of re-specifying both <code>:prefix<\/code> and <code>:non-normal-prefix<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(my-normal-and-insert-define-key :infix \"g\" &lt;maps...&gt;)<\/pre><\/div>\n<p>If you just want to create the prefix keymap and bind keys directly in it without immediately binding a prefix key to the prefix keymap, simply don\u2019t specify <code>:keymaps<\/code> or <code>:prefix<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; bind \"\/\" directly in the newly created my-prefix-map\n(general-define-key :prefix-map 'my-prefix-map \"\/\" 'swiper)<\/pre><\/div>\n<p>There is also a <code>:predicate<\/code> keyword for giving a condition under which a map should be active.<\/p>\n<h3><a id=\"user-content-predicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#predicates\"><\/a>Predicates<\/h3>\n<p>The user can use the <code>:predicate<\/code> keyword to specify a condition under which the map(s) should be active. For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :keymaps 'local\n :predicate '(eobp)\n \"&lt;right&gt;\" 'beginning-of-buffer)<\/pre><\/div>\n<p><code>&lt;right&gt;<\/code> will now behave normally except at the end of the buffer where it will jump to the beginning of the buffer. Note that with <code>:predicate<\/code>, you can still only have a key bound once in a single keymap. If you want to have a key take different actions depending on conditions in a single keymap, see <a href=\"#choosing-definitions-based-on-predicates\">Choosing Definition Based on Predicates<\/a>.<\/p>\n<p>See <a href=\"http:\/\/endlessparentheses.com\/define-context-aware-keys-in-emacs.html\" rel=\"nofollow\">this post<\/a> for more information about how this works.<\/p>\n<h2><a id=\"user-content-keymapstate-aliases\" class=\"anchor\" aria-hidden=\"true\" href=\"#keymapstate-aliases\"><\/a>Keymap\/State Aliases<\/h2>\n<p>To prevent the need to type out long keymap names like <code>evil-inner-text-objects-map<\/code>, general allows the user to specify shorthand names for keymaps by altering <code>general-keymap-aliases<\/code> (and for states by altering <code>general-state-aliases<\/code>). These are alists of either an alias or a list of aliases to the full keymap name:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(push '(help . help-map) general-keymap-aliases)\n;; or\n(push '((h help) . help-map) general-keymap-aliases)\n;; or (emacs 25+)\n(setf (alist-get 'help general-keymap-aliases) 'help-map)\n;; or (emacs 25+)\n(setf (alist-get '(h help) general-keymap-aliases) 'help-map)\n;; now\n(general-define-key :keymaps 'help ...)\n;; is the same as\n(general-define-key :keymaps 'help-map ...)<\/pre><\/div>\n<p>Note that earlier entries in the alist take precedence.<\/p>\n<p>By default, the global evil state and text object keymaps have aliases. This allows for using the same syntax as <code>evil-global-set-key<\/code> and <code>evil-define-key<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key :keymaps 'motion ...)\n;; or\n(general-define-key :keymaps 'm ...)<\/pre><\/div>\n<p>See <code>general-keymap-aliases<\/code> for all default aliases.<\/p>\n<p>All keymap symbols are immediately processed by <code>general--unalias<\/code>. By overriding this function, it would be possible to, for example, automatically append <code>-map<\/code> or <code>-mode-map<\/code> to keymap names that don\u2019t end in <code>-map<\/code> or do something more complicated to create a generic shorthand without having manually specify all aliases. This is not recommended as it could potentially become confusing (and would currently break <code>:definer 'minor-mode<\/code>), but if anyone would find this useful, feel free to make an issue, and I\u2019ll consider adding it as an option.<\/p>\n<h2><a id=\"user-content-general-define-key-wrappers\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-define-key-wrappers\"><\/a><code>general-define-key<\/code> Wrappers<\/h2>\n<h3><a id=\"user-content-positional-argument-wrappers\" class=\"anchor\" aria-hidden=\"true\" href=\"#positional-argument-wrappers\"><\/a>Positional Argument Wrappers<\/h3>\n<p>When defining keys in specific keymaps and states, using positional arguments can be shorter. General has two macros that can basically act as drop-in replacements for <code>define-key<\/code> and <code>evil-define-key<\/code> and another macro that can basically act is a drop-in replacement for both of those and more. They are <code>general-emacs-define-key<\/code>, <code>general-evil-define-key<\/code>, and <code>general-def<\/code> respectively. These are simply wrappers for <code>general-define-key<\/code> that pass the positional arguments to the corresponding keywords. However, for compatibility with <code>define-key<\/code> and <code>evil-define-key<\/code>, it is not necessary to quote keymaps. Both keymaps and states can be left quoted or unquoted (regardless of whether they are lists).<\/p>\n<p>For example, the following are all equivalent:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :keymaps 'org-mode-map\n \"M-n\" 'org-next-visible-heading\n \"M-p\" 'org-previous-visible-heading)\n(general-emacs-define-key org-mode-map\n  \"M-n\" 'org-next-visible-heading\n  \"M-p\" 'org-previous-visible-heading)\n;; rough equivalent with define-key\n(with-eval-after-load 'org-mode\n  (define-key org-mode-map (kbd \"M-n\") 'org-next-visible-heading)\n  (define-key org-mode-map (kbd \"M-p\") 'org-previous-visible-heading))<\/pre><\/div>\n<p>Similarly, the following are all equivalent:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :states '(normal visual)\n :keymaps 'org-mode-map\n \"gj\" 'org-next-visible-heading\n \"gk\" 'org-previous-visible-heading)\n(general-evil-define-key '(normal visual) org-mode-map\n  \"gj\" 'org-next-visible-heading\n  \"gk\" 'org-previous-visible-heading)\n;; equivalent with evil-define-key\n(evil-define-key '(normal visual) org-mode-map\n  \"gj\" 'org-next-visible-heading\n  \"gk\" 'org-previous-visible-heading)<\/pre><\/div>\n<p>The actual behavior of these two macros is the same as <code>general-define-key<\/code>. You can still use <code>general-define-key<\/code>\u2019s keyword arguments after the positional arguments (however, <code>:keymaps<\/code> and <code>:states<\/code> will not override the positional arguments):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; these are both valid\n(general-emacs-define-key 'global\n  :prefix \"C-c\"\n  \"\/\" 'swiper)\n(general-evil-define-key 'normal org-mode-map\n  :prefix \"SPC\"\n  \"g\" 'worf-goto)<\/pre><\/div>\n<p>As for <code>global-set-key<\/code> and <code>evil-global-set-key<\/code>, wrappers are not needed. By default <code>general-define-key<\/code> acts like <code>global-set-key<\/code>, and <code>general-emacs-define-key<\/code> can also act like <code>global-evil-set-key<\/code> using the symbols for evil\u2019s states (see <a href=\"#keymapstate-aliases\">keymap aliases<\/a>).<\/p>\n<p>The third macro, <code>general-def<\/code>, is provided for those who would prefer to use a single, succinctly named definer for all of the previous cases. It will act the same as <code>general-define-key<\/code>, <code>general-emacs-define-key<\/code>, or <code>general-evil-define-key<\/code> depending on the number of positional arguments.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; use `general-define-key' when no \"positional\" arguments\n(general-def\n  \"key\" 'def\n  ...)\n;; example equivalents\n(general-define-key \"key\" 'def)\n(global-set-key (kbd \"key\") 'def)\n;; use `general-emacs-define-key' when one \"positional\" argument\n(general-def org-mode-map\n  \"key\" 'def\n  ...)\n;; example equivalent\n(define-key org-mode-map (kbd \"key\") 'def)\n;; act like `evil-global-set-key'\n(general-def 'normal\n  \"key\" 'def\n  ...)\n;; example equivalents\n(evil-global-set-key 'normal (kbd \"key\") 'def)\n(evil-define-key 'normal 'global (kbd \"key\") 'def)\n;; use `general-evil-define-key' when two \"positional\" arguments\n(general-def 'normal org-mode-map\n  \"key\" 'def\n  ...)\n;; example equivalent\n(evil-define-key 'normal org-mode-map (kbd \"key\") 'def)<\/pre><\/div>\n<p>Note that all leading quoted and unquoted symbols and lists are considered to be positional arguments. This means that if you want to use a variable or function for a key that could be a positional argument, you should either use the definer <code>general-def<\/code> would end up using (<code>general-define-key<\/code>, <code>general-emacs-define-key<\/code>, or <code>evil-define-key<\/code>)  or explicitly separate the positional arguments from the first key with a bogus keyword argument:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-def\n  :start-maps t\n  some-key 'some-command)<\/pre><\/div>\n<h3><a id=\"user-content-mass-key-unbinding-wrapper\" class=\"anchor\" aria-hidden=\"true\" href=\"#mass-key-unbinding-wrapper\"><\/a>Mass Key Unbinding Wrapper<\/h3>\n<p><code>general-unbind<\/code> acts as <code>general-def<\/code>, but the positional arguments should all be keys (instead of pairs of keys and definitions) that should be unbound:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-unbind 'insert\n  \"C-v\"\n  \"C-k\"\n  \"C-y\"\n  \"C-e\")\n;; equivalent to\n(general-def 'insert\n  \"C-v\" nil\n  \"C-k\" nil\n  \"C-y\" nil\n  \"C-e\" nil)<\/pre><\/div>\n<p>This wrapper can also be used, for example, if you want to disable certain commands or keys from working in certain modes by using with <code>:with<\/code> keyword argument (example use case taken from <a href=\"https:\/\/github.com\/emacs-evil\/evil-collection\/blob\/9fc1a19807dfcd0cc2b221832b6e6faad80a291d\/evil-collection-util.el#L32\">evil-collection<\/a>):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-unbind 'normalInfo-mode-map\n  :with 'ignore\n  [remap evil-append]\n  [remap evil-append-line]\n  [remap evil-insert]\n  [remap evil-insert-line])\n;; equivalent to\n(general-def 'normal Info-mode-map\n  [remap evil-append] 'ignore\n  [remap evil-append-line] 'ignore\n  [remap evil-insert] 'ignore\n  [remap evil-insert-line] 'ignore)<\/pre><\/div>\n<p>The reason that this functionality is implemented as a wrapper and not as a keyword argument for <code>general-define-key<\/code> is that <code>cl-defun<\/code> cannot correctly parse keyword arguments when the keyword is in an odd position (e.g. <code>(\"a\" :keyword 'arg)<\/code> instead of <code>(:keyword 'arg \"a\")<\/code>). For example, if this functionality was implemented with an <code>:unbind<\/code> keyword, the <code>:general<\/code> use-package keyword and any definer created with <code>general-create-definer<\/code> would not work if the user specified an odd number of keys to unbind (because the default keyword arguments would be at the end of the arglist, in the wrong positions). As I\u2019d rather not re-implement keyword argument parsing just for this use case, this functionality is provided as a macro. This macro will correctly handle any positioning for keyword arguments.<\/p>\n<h3><a id=\"user-content-creating-new-key-definers\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-new-key-definers\"><\/a>Creating New Key Definers<\/h3>\n<p>The <code>general-create-definer<\/code> macro can create definers that wrap <code>general-def<\/code> but with certain default settings. For example, it can be used to create a definer that will default to a certain prefix (like <code>evil-leader<\/code> does):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; basic example\n(general-create-definer my-leader-def\n  :prefix \"C-c\")\n;; bind \"C-c o\" to `other-window'\n(my-leader-def \"o\" 'other-window)\n;; more complex example\n(general-create-definer tyrant-def\n  :states '(normal insert emacs)\n  :prefix \"SPC\"\n  :non-normal-prefix \"M-SPC\"\n  :prefix-command 'tyrant-prefix-command\n  :prefix-map 'tyrant-prefix-map)\n;; globally bind \"SPC \/\" in normal state and \"M-SPC \/\" in the insert\/emacs\n;; states to `swiper'\n(tyrant-def \"\/\" 'swiper)\n;; for org-mode, bind \"SPC o\" in normal state and \"M-SPC \/\" in the insert\/emacs\n;; states to `counsel-org-goto'\n(tyrant-def org-mode-map \"o\" 'counsel-org-goto)\n;; same as\n(tyrant-def :keymaps 'org-mode-map \"o\" 'counsel-org-goto)<\/pre><\/div>\n<p>It takes an optional <code>:wrapping<\/code> keyword argument that can be specified to use another definer instead of <code>general-def<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-create-definer my-prefix-def\n  :wrapping general-emacs-define-key\n  :prefix \"M-,\")<\/pre><\/div>\n<h3><a id=\"user-content-vim-like-definers\" class=\"anchor\" aria-hidden=\"true\" href=\"#vim-like-definers\"><\/a>Vim-like Definers<\/h3>\n<p><code>general-evil-setup<\/code> can be used to generate key definition functions that are named similarly to vim\u2019s. Currently, the following functions will be created:<\/p>\n<ul><li><code>general-imap<\/code><\/li>\n  <li><code>general-emap<\/code><\/li>\n  <li><code>general-nmap<\/code><\/li>\n  <li><code>general-vmap<\/code><\/li>\n  <li><code>general-omap<\/code><\/li>\n  <li><code>general-mmap<\/code><\/li>\n  <li><code>general-rmap<\/code><\/li>\n  <li><code>general-iemap<\/code><\/li>\n  <li><code>general-nvmap<\/code><\/li>\n  <li><code>general-otomap<\/code><\/li>\n  <li><code>general-itomap<\/code><\/li>\n  <li><code>general-tomap<\/code><\/li>\n<\/ul><p>These are wrappers around <code>general-def<\/code> created with <code>general-create-definer<\/code> that set the default <code>:states<\/code>. You can see the help text for each for a more specific description. <code>general-evil-setup<\/code> can be called with a non-nil argument (i.e. <code>(general-evil-setup t)<\/code>) to create non-prefixed aliases for these definers (e.g. <code>nmap<\/code>).<\/p>\n<p>Here is an example using <code>general-nmap<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-evil-setup)\n;; define in evil-normal-state-map\n(general-nmap \"key\" 'def ...)\n;; define in the normal state auxiliary map for org-mode-map\n(general-nmap org-mode-map \"key\" 'def ...)\n;; same as\n(general-nmap :keymaps 'org-mode-map \"key\" 'def ...)<\/pre><\/div>\n<h2><a id=\"user-content-note-for-evil-users\" class=\"anchor\" aria-hidden=\"true\" href=\"#note-for-evil-users\"><\/a>Note for Evil Users<\/h2>\n<p>When <code>:states<\/code> is specified, <code>general-define-key<\/code> will act as a wrapper around <code>evil-define-key*<\/code>. <code>evil-define-key*<\/code> now directly supports the symbol <code>global<\/code> for the keymap argument, so the following are equivalent:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n ;; (default)\n ;; :keymaps 'global\n :states '(normal visual)\n ...)\n(general-define-key\n :keymaps '(normal visual)\n ...)<\/pre><\/div>\n<p>Note that this previously was not the case and <code>(general-define-key :states 'normal ...)<\/code> would bind in the normal state auxiliary map for <code>(current-global-map)<\/code>. Since auxiliary maps have a higher precedence than evil global and override keymaps, this was previously mentioned as one possible way of preventing certain keybindings from being overridden. However, this is not a reliable method. Keys bound in auxiliary maps can override keys bound in other auxiliary maps, for example, and keys bound in evil local or minor-mode keymaps will always override keys bound in regular auxiliary maps. If you need this functionality, please use evil intercept keymaps instead (see <a href=\"#override-keymaps-and-buffer-local-keybindings\">Override Keymaps<\/a>).<\/p>\n<p>General.el provides the equivalent of <code>bind-key<\/code>\u2019s <code>override-global-map<\/code> as <code>general-override-mode-map<\/code> (keymap alias is =\u2019override=). When <code>general-override-mode<\/code> is enabled, keys bound in <code>general-override-mode-map<\/code> will take precedence over keys bound in any other minor mode keymaps. By default, general.el will automatically enable <code>general-override-mode<\/code> when binding a key in <code>general-override-mode-map<\/code>. If you would prefer to enable it manually (e.g. you wish to toggle it at some point), you can set <code>general-override-auto-enable<\/code> to nil.<\/p>\n<p>General also provides a local equivalent called <code>general-override-local-mode<\/code> which is used to add support for buffer-local keybindings (with higher precedence than mode keybindings) by specifying <code>:keymaps 'local<\/code>. Unlike with the global override mode, <code>:keymaps 'local<\/code> should always be used instead of the actual keymap name since <code>:keymaps 'local<\/code> will cause general.el to automatically turn on the corresponding minor mode and perform some necessary extra setup. Note that this is not the same as using <code>local-set-key<\/code> (which will bind the key for the current buffer\u2019s major mode, affecting other buffers). When <code>:states<\/code> is specified with <code>:keymaps 'local<\/code>, <code>evil-local-set-key<\/code> will be used instead.<\/p>\n<p>Note that binding directly in <code>general-override-mode-map<\/code> (i.e. no <code>:states<\/code> specified) is only useful for non-evil keybindings. Evil keybindings already override almost all normal emacs keybindings using the same method used here (i.e. evil keymaps are in <code>emulation-mode-map-alists<\/code>). The main exceptions where evil keybindings will be overridden by non-evil keybindings are noted <a href=\"https:\/\/github.com\/noctuid\/evil-guide#what-overrides-evil\">here<\/a> with explanations on how to deal with these cases. To understand which evil keybindings override others, review the <a href=\"https:\/\/github.com\/noctuid\/evil-guide#keymap-precedence\">precedence for evil keymaps<\/a>. If you want a global evil keybinding to not be overridden by any other evil keymaps (e.g. overriding keymaps created in <code>evil-integration.el<\/code> or auxiliary keymaps created by some evil package), you can use intercept keymaps. You can make any keymap an intercept keymap, but it may be convenient to just use <code>general-override-mode-map<\/code> for this purpose since the necessary setup (<code>evil-make-intercept-map<\/code>) has already been performed:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; keybindings that should not be overriden\n(general-define-key\n :states 'normal\n :keymaps 'override\n :prefix \"SPC\"\n \"f\" 'find-file)\n;; the above has precedence over the following (excerpt from evil-collection)\n;; \"SPC f\" will still work as `find-file'\n(evil-define-key 'normal transmission-mode-map\n  (kbd \"SPC\") 'scroll-up-command)<\/pre><\/div>\n<p>Note that by default, evil keybindings made with <code>:keymaps 'override<\/code> will override even those made with <code>:keymaps 'local<\/code>.<\/p>\n<p>General keeps track of all your keybindings and allows presenting them as tables in an org buffer using <code>general-describe-keybindings<\/code>. By default, they will be displayed in this order:<\/p>\n<ul><li>Buffer local keybindings (i.e. <code>:keymaps 'local<\/code>)<\/li>\n  <li>Global keybindings (i.e. <code>:keymaps 'global<\/code>)<\/li>\n  <li>Global evil keybindings (e.g. <code>:keymaps 'evil-normal-state-map<\/code>)<\/li>\n  <li>Other keybindings<\/li>\n<\/ul><p>Within these categories keymaps, states, and keybindings will be presented in the order they were created in. For each keybinding created, this command will display the key, the definition, and the previous definition. The previous definition will only be updated when the definition changes by default. To have it only be updated when the key was previously unbound, the user can set <code>general-describe-update-previous-definition<\/code> to <code>nil<\/code>.<\/p>\n<p>The order in which keybindings are displayed is customizable. All keymaps listed in <code>general-describe-priority-keymaps<\/code> will be displayed first. The rest can optionally be sorted by setting <code>general-describe-keymap-sort-function<\/code> (nil by default). The order evil states are displayed in can be altered either by changing <code>general-describe-state-sort-function<\/code> or changing the order of states in <code>general-describe-evil-states<\/code>. Keybindings can also be sorted if the user sets <code>general-describe-keybinding-sort-function<\/code>. Here is an example that will sort everything alphabetically:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq general-describe-priority-keymaps nil\n      general-describe-keymap-sort-function #'general-sort-by-car\n      general-describe-state-sort-function #'general-sort-by-car)\n;; sort keybindings alphabetically by key\n(setq general-describe-keybinding-sort-function #'general-sort-by-car)\n;; sort keybindings alphabetically by definition\n(setq general-describe-keybinding-sort-function #'general-sort-by-cadr)<\/pre><\/div>\n<p>For reference, keybindings are stored in an alist. Here is what is passed to each sorting function:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; `general-keybindings' - an alist of keymap to state alist\n;; passed to `general-describe-keymap-sort-function'\n((keymap-name . state-alist) ...)\n;; a state alist (state name is nil if there is no state)\n;; passed to `general-describe-state-sort-function'\n((state-name . keybindings) ...)\n;; the list of keybindings is passed to `general-describe-keybinding-sort-function'\n((\"key after kbd applied\" 'def 'previous-def) ...)<\/pre><\/div>\n<p>To actually change how the keybinding table is printed, the user could override  <code>general--print-map<\/code>.<\/p>\n<h2><a id=\"user-content-disclaimer\" class=\"anchor\" aria-hidden=\"true\" href=\"#disclaimer\"><\/a>Disclaimer<\/h2>\n<p>Key simulation (for <code>general-simulate-key<\/code> and <code>general-key-dispatch<\/code> but not for <code>general-key<\/code>) can result in duplicate keys being recorded for keyboard macros and evil repeating. To work around this issue, <code>general.el<\/code> will discard these duplicate keys during macro playback (i.e. <code>executing-kbd-macro<\/code> is non-nil). So far, this seems to be a reliable method for getting macros and repeating to work correctly with key simulation. However, it is hard (and maybe impossible) to test some of these cases automatically since it involves simulating keys that in turn simulate keys, and, for example, I haven\u2019t found a way to correctly simulate recording a macro in these cases. Therefore, if you find any issues with macro playback or evil repeating when using <code>general-simulate-key<\/code> or <code>general-key-dispatch<\/code>, please make an issue.<\/p>\n<h2><a id=\"user-content-simulating-keypresses\" class=\"anchor\" aria-hidden=\"true\" href=\"#simulating-keypresses\"><\/a>Simulating Keypresses<\/h2>\n<p>General provides two macros called <code>general-key<\/code> and <code>general-simulate-key<\/code> that can be used to simulate key sequences. In some cases, they can be used similarly to keyboard macros, but they have some advantages. Unlike with a keyboard macro, prefix arguments will work for the command that ends up running. Also, the key simulated does not have to correspond to the full key sequence for a command. See <a href=\"https:\/\/www.emacswiki.org\/emacs\/Evil#toc14\" rel=\"nofollow\">here<\/a> for information on an alternative method of doing some of the things these key simulation helpers can do using <code>key-translation-map<\/code>. I personally prefer general\u2019s helpers as they are simple and more powerful.<\/p>\n<p>Note that when a named prefix keymap\/command exists (e.g. <code>help-command<\/code>), you should generally prefer to bind directly to that. However, this is not possible for a key like <code>C-c<\/code> whose definition varies depending on the buffer. Therefore, you need to use either <code>general-key<\/code> or <code>general-simulate-key<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap \"SPC\" (general-simulate-key \"C-c\"))\n;; or\n(general-nmap \"SPC\" (general-key \"C-c\"))<\/pre><\/div>\n<p>Although both will work correctly, <a href=\"https:\/\/github.com\/justbur\/emacs-which-key\">which-key<\/a> does not currently show all available keys when <code>general-key<\/code> is used, so I would currently recommend using <code>general-simulate-key<\/code> instead for an example like this.<\/p>\n<p>On the other hand, <code>general-key<\/code> should be preferred for simulating a key that corresponds to a single command. Unlike <code>general-simulate-key<\/code>, which creates\/returns a function, <code>general-key<\/code> expands to an extended menu item like <code>general-predicate-dispatch<\/code>. Using an extended menu item is a simpler and more direct approach as emacs will dynamically look up and act as the specified key. This has the advantage of showing the docstring for the exact command with <code>C-h k<\/code>. If the key to act as is unbound, key lookup can continue (like if <code>:predicate<\/code> returns nil), so having a fallback keybinding is possible with <code>general-key<\/code> but not with <code>general-simulate-key<\/code>.<\/p>\n<p><code>general-key<\/code> may be useful when you want to have a key act as another without having to bind it to the exact command in every relevant keymap:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap \"RET\" (general-key \"C-c C-c\"))\n;; a keyboard macro works, but C-h k will not show the command docstring\n(general-nmap \"RET\" \"C-c C-c\")<\/pre><\/div>\n<p><code>general-simulate-key<\/code> and <code>general-key<\/code> also support keyword arguments (<code>:state<\/code> and <code>:keymap<\/code>) to control the context the keys are simulated in. For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap \"j\" (general-simulate-key \"C-n\" :state 'emacs))\n;; `general-key' supports :state only`\n(general-nmap \"j\" (general-key \"C-n\" :state 'emacs))<\/pre><\/div>\n<p>The advantage of <code>general-simulate-key<\/code> over <code>general-key<\/code> is that it can be used to simulate a key sequence corresponding to multiple commands or a command followed by a key sequence. The key argument can be replaced by a list of a command and keys (e.g. <code>(general-simulate-key ('evil-delete \"iw\"))<\/code>). For example, the following is possible with <code>general-simulate-key<\/code> but not with <code>general-key<\/code> or a keyboard macro:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap \"s\" (general-simulate-key ('evil-ex \"s\/\")))<\/pre><\/div>\n<p>See the next section for another reasonable use case for this feature.<\/p>\n<p>When a command is specified for <code>general-simulate-key<\/code>, general will used the remapped version of it if it exists (e.g. if <code>[remap evil-delete] 'lispyville-delete<\/code> is in an active keymap, <code>lispyville-delete<\/code> will be used instead of <code>evil-delete<\/code>). To use the exact command instead, <code>:remap nil<\/code> can be specified<\/p>\n<p><code>general-simulate-key<\/code> creates a named function with a docstring, so which-key and <code>describe-key<\/code> will work properly for keys bound to a command created with it. The automatically generated function name, docstring, and which-key description can be replaced with keyword arguments:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap \"SPC\" (general-simulate-key \"C-c\"\n                      :state 'emacs\n                      :name general-SPC-simulates-C-c\n                      :docstring \"Simulate C-c in emacs state with SPC.\"\n                      :which-key \"Simulate C-c\"))<\/pre><\/div>\n<p>Make sure that you don\u2019t bind a key to simulate itself (e.g. <code>(general-emap \"C-n\" (general-simulate-key \"C-n\" :state 'emacs))<\/code>) as this will cause an infinite loop.<\/p>\n<h2><a id=\"user-content-mapping-under-non-prefix-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#mapping-under-non-prefix-keys\"><\/a>Mapping Under Non-prefix Keys<\/h2>\n<p>This functionality is mainly targeted at evil users, but it could potentially be useful for non-evil users as well. In vim you can bind something like <code>cow<\/code> without a problem. With evil, <code>c<\/code> is bound to <code>evil-change<\/code>, so you can\u2019t bind directly to <code>cow<\/code>. A workaround for this case is to bind a key in <code>evil-operator-state-map<\/code>, but this won\u2019t work when operator state is not used (e.g. you want to bind something like <code>ctb<\/code> or <code>jk<\/code> in insert state). I\u2019ve come up with a more general workaround called <code>general-key-dispatch<\/code>. Consider the following example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap \"c\" (general-key-dispatch 'evil-change\n                    \"ow\" 'toggle-word-wrap\n                    \"tb\" 'some-command\n                    \"c\" 'evil-change-whole-line\n                    ;; alternatively if there was no linewise version:\n                    \"c\" (general-simulate-key ('evil-change \"c\"))))\n;; `evil-change' is not bound in `evil-visual-state-map' by default but\n;; inherited from `evil-normal-state-map'\n;; if you don't want \"c\" to be affected in visual state, you should add this\n(general-vmap \"c\" 'evil-change)<\/pre><\/div>\n<p><code>general-key-dispatch<\/code> is a function-creating macro. In this example, the command created will wait for user input and try to match one of the specified key sequences (e.g. <code>ow<\/code>). If a key sequence is matched, the corresponding command will be executed. Otherwise it will fall back to simulating the fallback command followed by the unmatched keys (using the same mechanism as <code>general-simulate-key<\/code>). For example, <code>ow<\/code> is bound, so <code>cow<\/code> would run <code>toggle-word-wrap<\/code>. On the other hand, <code>b<\/code> is not mapped, so <code>cb<\/code> would act the same as <code>cb<\/code> would by default. Counts and repeating should still work for both the mapped keys and fallback command. Because evil handles <code>cc<\/code> differently (since <code>c<\/code> is not a motion), <code>c<\/code> must be explicitly bound to <code>evil-change-whole-line<\/code> (or to simulate <code>('evil-change \"c\")<\/code>) to keep its behavior. <code>c<\/code> is not actually bound in visual state by default, so to keep <code>c<\/code> working the same in visual state, you should explicitly bind it to <code>evil-change<\/code>.<\/p>\n<p>Like with <code>general-simulate-key<\/code>, general will first check to see if the command to be executed has been remapped (e.g. if <code>[remap evil-delete] 'lispyville-delete<\/code> is in an active keymap, <code>lispyville-delete<\/code> will be used instead of <code>evil-delete<\/code>). To use the exact command instead, <code>:remap nil<\/code> can be specified.<\/p>\n<p>Another thing to note is that you can\u2019t bind a key in the <code>general-key-dispatch<\/code> section to simulate the base key (i.e. the key you bind to the resulting command, in this case <code>c<\/code>). For this example, you couldn\u2019t bind <code>w<\/code> to <code>(general-simulate-key \"ciw\")<\/code>. While this wouldn\u2019t cause an infinite loop, it wouldn\u2019t work either, so you would have to use the command name instead (e.g <code>(general-simulate-key ('evil-change \"iw\"))<\/code>).<\/p>\n<p>Also, if you use a count in the middle (e.g. <code>c2tb<\/code> and <code>2<\/code> is not explicitly bound), the fallback command will be run immediately. If anyone cares about this, feel free to make an issue. I could potentially add an option to allow changing the count in the middle without immediately falling back to the default command.<\/p>\n<p>Another possible use case of <code>general-key-dispatch<\/code> is to emulate vim\u2019s <code>imap<\/code>. For example, you can recreate the common <code>jk<\/code> to <code>&lt;esc&gt;<\/code> keybinding:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-imap \"j\"\n              (general-key-dispatch 'self-insert-command\n                \"k\" 'evil-normal-state))<\/pre><\/div>\n<p>Commands created in this way support an optional timeout, meaning you could still insert <code>jk<\/code> (without <code>C-q<\/code> \/ <code>quoted-insert<\/code>) like with <a href=\"https:\/\/www.emacswiki.org\/emacs\/key-chord.el\" rel=\"nofollow\">key-chord.el<\/a>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-imap \"j\"\n              (general-key-dispatch 'self-insert-command\n                :timeout 0.25\n                \"k\" 'evil-normal-state))<\/pre><\/div>\n<p>If you are using <code>general-key-dispatch<\/code> with a timeout to mirror some prefix keymap in insert state, it may also convenient to use the <code>:inherit-keymap<\/code> keyword. This allows using prefix keybindings without the need to re-specify them in the <code>general-key-dispatch<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-nmap :prefix \",\"\n              :prefix-command 'my-prefix-map\n              \"g\" 'magit-status)\n(general-imap \",\"\n              (general-key-dispatch 'self-insert-command\n                :timeout 0.25\n                :inherit-keymap my-prefix-map))<\/pre><\/div>\n<p>If you bind more keys under your prefix later on in normal state, they will still be available when pressing the prefix in insert state without the need to re-evaluate the <code>general-key-dispatch<\/code>.<\/p>\n<p>By default, <code>general-key-dispatch<\/code> will prevent name clashes by appending a unique number to name of the created command (e.g. <code>general-dispatch-self-insert-command-G402<\/code>). If you would like to reference the created command by name, you can name it yourself using the <code>:name<\/code> keyword argument (e.g. <code>:name general-insert-prefix-dispatch<\/code>).<\/p>\n<p>Like with <code>general-simulate-key<\/code> used with a command name, the behavior of <code>evil-repeat<\/code> will depend on the command that ends up running. Having repeating work correctly requires handling a lot of edge cases, so please make an issue if you find any problems. Note that evil does not support repeating a count that comes before an operator currently, but repeating should work when the count follows the operator key (<code>3cc<\/code> vs <code>c3c<\/code>).<\/p>\n<h2><a id=\"user-content-choosing-definitions-based-on-predicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#choosing-definitions-based-on-predicates\"><\/a>Choosing Definitions Based on Predicates<\/h2>\n<p><code>general-predicate-dispatch<\/code> can be used to generate a <code>menu-item<\/code> that will behave differently based on the provided predicates. It takes a fallback definition as the first argument and then a list of predicates and alternate definitions (which can be commands, keymaps, etc.). Predicates are checked in order. If no predicate is matched and the fallback command is nil, then the mapping will be ignored (the keymap with the next highest precedence, if one exists, will be checked for the pressed key(s)).<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key \"&lt;right&gt;\"\n                    (general-predicate-dispatch 'right-char\n                      ;; pred def ...\n                      (eolp) 'beginning-of-line))<\/pre><\/div>\n<p>The <code>:docstring<\/code> keyword can be specified to add a description to the extended menu item.<\/p>\n<h2><a id=\"user-content-key-translation\" class=\"anchor\" aria-hidden=\"true\" href=\"#key-translation\"><\/a>Key \u201cTranslation\u201d<\/h2>\n<p><code>general-translate-key<\/code> allows binding a key to the definition of another key in the same keymap (comparable to how vim\u2019s keybindings work). Its arguments are the <code>states<\/code> (which can be nil for non-evil keymaps) and <code>keymaps<\/code> (both symbols or lists of symbols like for <code>general-define-key<\/code>) to bind\/look up the key(s) in followed optionally by keyword arguments (currently only <code>:destructive<\/code>) and key\/replacement pairs.<\/p>\n<p><code>evil-collection-translate-key<\/code> allows binding a key to the definition of another key in the same keymap (comparable to how vim\u2019s keybindings work). Its arguments are the <code>states<\/code> and <code>keymaps<\/code> to bind\/look up the key(s) in followed optionally by keyword arguments (currently only <code>:destructive<\/code>) and key\/replacement pairs. <code>states<\/code> can be nil for non-evil keymaps, and both <code>states<\/code> and <code>keymaps<\/code> can be a single symbol or a list of symbols.<\/p>\n<p>This can be particularly useful, for example, when you want make key swaps\/cycles en masse. This use case is similar to one for <code>general-simulate-key<\/code> (i.e. make a key act as another key that has a consistent meaning but different commands for different modes without having to individually bind the key to the exact definition in each mode\u2019s keymap). However, <code>general-simulate-key<\/code> is not always suitable for this purpose. It can be used to, for example, make <code>j<\/code> in normal state act as <code>C-n<\/code> in emacs state (to use the default \u201cdown\u201d navigation key for all modes without needing to individually make keybindings for every mode), but it cannot be used to swap\/cycle keys within a single keymap, as this would cause an infinite loop of simulating the other key(s).<\/p>\n<p>An example use case of <code>general-translate-key<\/code> is for non-QWERTY users who want to retain the hjkl keyboard positions for movement in dired, mu4e, etc. When using a package that already creates hjkl keybindings for the desired mode(s) (e.g. <a href=\"https:\/\/github.com\/jojojames\/evil-collection\">evil-collection<\/a>), it is easily possible to make these cycles in a single statement:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; single invocation example\n(general-translate-key nil 'evil-normal-state-keymap\n  \"n\" \"j\"\n  \"e\" \"k\"\n  ...)\n;; cycling keys en masse\n(dolist (keymap keymaps-with-hjkl-keybindings)\n  (general-translate-key 'normal keymap\n    ;; colemak hnei is qwerty hjkl\n    \"n\" \"j\"\n    \"e\" \"k\"\n    \"i\" \"l\"\n    ;; add back nei\n    \"j\" \"e\"\n    \"k\" \"n\"\n    \"l\" \"i\"))<\/pre><\/div>\n<p>By default, the first invocation of <code>general-translate-key<\/code> will make a backup of the keymap. Each subsequent invocation will look up keys in the backup instead of the original. This means that a call to <code>general-translate-key<\/code> will always have the same behavior even if evaluated multiple times. When <code>:destructive t<\/code> is specified, keys are looked up in the keymap as it is currently. This means that a call to <code>general-translate-key<\/code> that swapped two keys would continue to swap\/unswap them with each call. Therefore when <code>:destructive t<\/code> is used, all cycles\/swaps must be done within a single call to <code>general-translate-key<\/code>. To make a comparison to vim keybindings, <code>:destructive t<\/code> is comparable to vim\u2019s <code>map<\/code>, and <code>:destructive nil<\/code> is comparable to vim\u2019s <code>noremap<\/code> (where the \u201coriginal\u201d keybindings are those that existed in the keymap when <code>general-translate-key<\/code> was first used).<\/p>\n<p>You\u2019ll almost always want to use the default behavior (especially in your init file). The limitation of <code>:destructive nil<\/code> is that you can\u2019t translate a key to another key that was defined after the first <code>evil-collection-translate-key<\/code>, so <code>:destructive t<\/code> may be useful for interactive experimentation.<\/p>\n<p>Note that general state and keymap aliases (as well as <code>local<\/code> and <code>global<\/code>) and <code>general-implicit-kbd<\/code> are supported by <code>general-translate-key<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; normal -&gt; evil-normal-state-keymap\n(general-translate-key nil 'normal\n  ;; kbd not necessary by default\n  \"C-p\" \"C-n\")<\/pre><\/div>\n<p>Keys are bound using <code>general-define-key<\/code>, so they are viewable with <code>general-describe-keybindings<\/code>.<\/p>\n<p><code>general-swap-key<\/code> is provided as a wrapper around <code>general-translate-key<\/code> that allows swapping keys:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-swap-key nil 'normal\n  \";\" \":\"\n  \"a\" \"A\")\n;; equivalent to\n(general-translate-key nil 'normal\n  \";\" \":\"\n  \":\" \";\"\n  \"a\" \"A\"\n  \"A\" \"a\")<\/pre><\/div>\n<h2><a id=\"user-content-automatic-key-unbinding\" class=\"anchor\" aria-hidden=\"true\" href=\"#automatic-key-unbinding\"><\/a>Automatic Key Unbinding<\/h2>\n<p>To automatically prevent <code>Key sequence starts with a non-prefix key<\/code> errors without the need to explicitly unbind non-prefix keys, you can add <code>(general-auto-unbind-keys)<\/code> to your configuration file. This will advise <code>define-key<\/code> to unbind any bound subsequence of the <code>KEY<\/code>. Currently, this will only have an effect for <code>general.el<\/code> key definers. The advice can later be removed with <code>(general-auto-unbind-keys t)<\/code>.<\/p>\n<p>The reason that advice is used is because <code>general-define-key<\/code> does not always define keys in the same manner. Because customer definers are supported with <code>:definer<\/code>, <code>general-define-key<\/code> does not have the necessary information to handle every case itself.<\/p>\n<p>As a final note, if you, for example, bind <code>s<\/code> to a command using <code>general-define-key<\/code> and then later bind <code>s &lt;key&gt;<\/code> to something, <code>s<\/code> will still show up in <code>general-describe-keybindings<\/code> even though it\u2019s no longer bound. Since this is preventable by simply removing the initial unused keybinding, I likely will not try to add a workaround to fix this.<\/p>\n<p>General.el also provides a few helper functions\/macros for other configuration purposes. They are intended to be slightly more convenient versions of functions\/macros provided by default.<\/p>\n<h2><a id=\"user-content-settings\" class=\"anchor\" aria-hidden=\"true\" href=\"#settings\"><\/a>Settings<\/h2>\n<p><code>general-setq<\/code> is a stripped-down <code>customize-set-variable<\/code> that can act as a drop-in replacement for <code>setq<\/code>. The reason you might want to use it instead of <code>setq<\/code> is that <code>setq<\/code> cannot correctly set all variables. Some variables defined with <code>defcustom<\/code> specify a custom setter with <code>:set<\/code> that must be used for changes to take effect (e.g. <code>auto-revert-interval<\/code>). If the corresponding package has already been loaded, using <code>setq<\/code> will generally not work to set these variables. On the other hand, <code>general-setq<\/code> will correctly use the custom setter when necessary. One benefit of <code>general-setq<\/code> over <code>customize-set-variable<\/code> is that it can be used to set multiple variables at once. It does not do everything <code>customize-set-variable<\/code> does (e.g. it cannot be used interactively, does not attempt to load variable dependencies, and does not allow the user to specify comments). From some basic testing, it is 10x to 100x faster because of this, but the speed difference should not really be noticeable if you aren\u2019t setting thousands of variables during emacs initialization.<\/p>\n<p>Here\u2019s an example using variables that have a custom setter:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-setq auto-revert-interval 10\n              evil-want-Y-yank-to-eol t\n              evil-search-module 'evil-search)<\/pre><\/div>\n<p>Note that <code>setq<\/code> will work as expected as long it is used before the corresponding package is loaded, but with <code>customize-set-variable<\/code> or <code>general-setq<\/code>, you do not need to worry about whether or not the package has been loaded. If you decide to use <code>general-setq<\/code>, I\u2019d recommend aliasing it to something shorter like <code>gsetq<\/code>.<\/p>\n<p>One major difference from <code>customize-set-variable<\/code> that you should be aware of is that <code>general-setq<\/code> falls back to using <code>set<\/code> instead of <code>set-default<\/code>. This means that, like <code>setq<\/code>, it will alter the local value of buffer-local variables instead of the default value.<\/p>\n<p><code>general-setq-default<\/code> and <code>general-setq-local<\/code> also exist but do not attempt to call custom setters. The reason for this is that I have never seen any custom setters for variables that make sense to set both globally and locally (custom setters I\u2019ve seen just use <code>set-default<\/code>). <code>setq-default<\/code> is useful when you want to globally change the default for a buffer-local variable. <code>setq-local<\/code> is useful when you want to make a non-buffer-local variable buffer-local and then change its local value (<code>setq<\/code> already preferentially alters the buffer-local value of a variable if there is one). For now, the general.el equivalents are just aliases, but in the future, they will likely record user settings to be displayed in a table later.<\/p>\n<h2><a id=\"user-content-hooks-and-advice\" class=\"anchor\" aria-hidden=\"true\" href=\"#hooks-and-advice\"><\/a>Hooks and Advice<\/h2>\n<p><code>general-add-hook<\/code>, <code>general-remove-hook<\/code>, <code>general-advice-add<\/code>, and <code>general-advice-remove<\/code> all act as drop-in replacements for their corresponding functions but allow lists for some of the arguments. The hook functions allow specifying lists for the hooks and functions, and the advice functions allow specifying lists for the symbols and functions. Because I don\u2019t like the difference in naming for the default advice functions, <code>general-add-advice<\/code> and <code>general-remove-advice<\/code> are also provided as aliases.<\/p>\n<p>For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-add-hook my-lisp-mode-hooks\n                  (list #'lispy-mode #'rainbow-delimiters-mode))\n;; note that setting the :jump command property is recommended instead of this\n(general-add-advice (list #'git-gutter:next-hunk\n                          #'git-gutter:previous-hunk)\n                    :before #'evil-set-jump)<\/pre><\/div>\n<h2><a id=\"user-content-use-package-keywords\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-package-keywords\"><\/a>Use-package Keywords<\/h2>\n<h3><a id=\"user-content-general-keyword\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-keyword\"><\/a>:general Keyword<\/h3>\n<p>General also optionally provides a use-package keyword. <code>:general<\/code> is similar to <code>:bind<\/code> in that it implies <code>:defer t<\/code> whenever there are bound commands that can be autoloaded (e.g. it will not imply <code>:defer t<\/code> if the only bound command is to a lambda, for example). Whenever autoloadable commands are bound, use-package will create autoloads for them (though this is usually not necessary). The keyword is followed by one or more lists containing arguments for <code>general-def<\/code>; there is no difference in syntax:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package org\n  :general\n  (\"C-c c\" 'org-capture)\n  (:keymaps 'org-mode-map\n   \"TAB\" 'org-cycle)\n  ;; uses `general-def' not `general-define-key', so this is fine\n  (org-mode-map\n   \"TAB\" 'org-cycle))<\/pre><\/div>\n<p>The <code>:general<\/code> keyword also supports using any other key definer\/wrapper by manually specifying it:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package org\n  :general\n  (general-nmap \"SPC c\" 'org-capture))<\/pre><\/div>\n<p>One annoyance you may encounter is that the default function for indentation will indent a list starting with a keyword like a function:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(:keymaps 'org-mode-map\n          \"TAB\" 'org-cycle)<\/pre><\/div>\n<p>This is an annoyance you may have using other emacs packages as well and can be fixed by modifying <code>lisp-indent-function<\/code> (see <a href=\"http:\/\/emacs.stackexchange.com\/q\/10230\/5278\" rel=\"nofollow\">this emacs stackexchange question<\/a> and Fuco1\u2019s modified <code>lisp-indent-function<\/code> in one of the answers there).<\/p>\n<h3><a id=\"user-content-hook-keywords\" class=\"anchor\" aria-hidden=\"true\" href=\"#hook-keywords\"><\/a>Hook Keywords<\/h3>\n<p>General provides two alternatives to <code>:hook<\/code> that use <code>general-add-hook<\/code> called <code>:ghook<\/code> and <code>:gfhook<\/code>. Both take any number of arguments of symbols or lists. List arguments work the same for both; they correspond to a list of arguments for <a href=\"#hooks-and-advice\">~general-add-hook~<\/a>. The primary difference between the two is that symbol arguments to <code>:ghook<\/code> are <i>hooks<\/i>, but they are <i>functions<\/i> for <code>:gfhook<\/code> (hence the <code>f<\/code>). Furthermore, <code>:ghook<\/code> usually implies <code>:defer t<\/code>, and <code>:gfhook<\/code> never implies <code>:defer t<\/code>. <code>:ghook<\/code> should be used when the <code>general-add-hook<\/code> is meant to trigger the loading of the package. <code>:gfhook<\/code> should be used when the <code>general-add-hook<\/code> is meant to trigger some function in response to the package\u2019s mode being enabled (or toggled in the case of a minor mode). More simply put, <code>:ghook<\/code> is suited towards enabling minor modes, and <code>:gfhook<\/code> is suited towards performing setup once some mode has loaded. The use case for each is further explained below.<\/p>\n<h4><a id=\"user-content-ghook-keyword\" class=\"anchor\" aria-hidden=\"true\" href=\"#ghook-keyword\"><\/a>:ghook Keyword<\/h4>\n<p><code>:ghook<\/code> is intended to be used to add a package\u2019s minor mode enabling function to a user-specified <i>hook<\/i>, so that when hook is run, the package will be loaded and the mode enabled. This means that <code>:ghook<\/code> will usually imply <code>:defer t<\/code>. While it does not always imply <code>:defer t<\/code>, it will add any non-lambda functions to <code>:commands<\/code> (this is the same behavior as <code>:hook<\/code>). Though this is usually unnecessary (the commands probably already have autoloads), it will in turn imply <code>:defer t<\/code>.<\/p>\n<p>Symbols specified with <code>:ghook<\/code> correspond to hooks, and the function to add to each hook is inferred from the package\u2019s name (i.e. <code>-mode<\/code> is automatically added to the package name unless the package\u2019s name already ends in <code>-mode<\/code>). For example, these are all the same:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package rainbow-delimiters\n  :ghook 'prog-mode-hook)\n(use-package rainbow-delimiters\n  ;; `general-add-hook' arglist: HOOKS FUNCTIONS &amp;optional APPEND LOCAL\n  ;; a missing FUNCTIONS argument will be replaced with inferred minor mode\n  :ghook ('prog-mode-hook))\n(use-package rainbow-delimiters\n  ;; a null or non-symbol placeholder for FUNCTIONS will be replaced with\n  ;; inferred minor mode command; this may be useful if you want to keep the\n  ;; inferred command but also want to set the APPEND and\/or LOCAL arguments\n  ;; afterwards, e.g. ('prog-mode-hook nil t)\n  :ghook ('prog-mode-hook nil))\n(use-package rainbow-delimiters\n  ;; the full arglist for `general-add-hook' can be specified\n  ;; this is necessary if inference is not possible (see below for an example)\n  :ghook ('prog-mode-hook #'rainbow-delimiters-mode))\n(use-package\n  ;; :commands implies :defer t\n  :commands rainbow-delimiters-mode\n  :init (general-add-hook 'prog-mode-hook #'rainbow-delimiters-mode))<\/pre><\/div>\n<p>If you are already familiar with <code>:hook<\/code>, you should note that there are quite a few syntactic differences between <code>:ghook<\/code> and <code>:hook<\/code>. Firstly, quoting the hooks and functions is required. Like <code>:general<\/code> uses the same syntax as <code>general-def(ine-key)<\/code> (unlike <code>:bind<\/code>), <code>:ghook<\/code> uses the same syntax as <code>(general-)add-hook<\/code> for both clarity and convenience. For example, the user may want to use a helper function\/macro to generate the function(s) to add to the hook (see the <a href=\"#gfhook-keyword\">:gfhook section<\/a> for a specific example). The user may also want to specify a variable containing a list of hooks instead of an actual hook name:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defconst my-lisp-mode-hooks\n  '(lisp-mode-hook\n    emacs-lisp-mode-hook\n    clojure-mode-hook\n    scheme-mode-hook\n    ;; ...\n    ))\n(use-package lispy\n  :ghook my-lisp-mode-hooks)\n;; same as\n(use-package lispy\n  :ghook (my-lisp-mode-hooks))\n;; same as\n(use-package lispy\n  ;;  `general-add-hook' can take a list of hooks for the HOOK argument\n  :ghook ('(lisp-mode-hook\n            emacs-lisp-mode-hook\n            clojure-mode-hook\n            scheme-mode-hook\n            ;; ...\n            )))<\/pre><\/div>\n<p>Furthermore, <code>:ghook<\/code> will not automatically add <code>-hook<\/code> to specified hook symbols (i.e. you must specify <code>prog-mode-hook<\/code>; <code>prog-mode<\/code> is not sufficient). This design decision is intended to help prevent confusion since <code>:gfhook<\/code> also exists, and its symbols correspond to functions (not hooks) that could also end in <code>-mode<\/code> (and could potentially not be sharp quoted). I don\u2019t think the loss in conciseness is major, and hopefully this will help always make it immediately clear whether symbols correspond to functions or hooks.<\/p>\n<p>Lastly, <code>:hook<\/code> only takes one argument, whereas <code>:ghook<\/code> can take an arbitrary number of arguments (just like <code>:general<\/code>):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lispy\n  ;; any number of symbols (or lists) is allowed\n  :ghook\n  'lisp-mode-hook\n  'emacs-lisp-mode-hook\n  'clojure-mode-hook\n  'scheme-mode-hook)<\/pre><\/div>\n<p>Note that if the function name cannot be inferred from the package name (i.e. the package name or the package name with <code>-mode<\/code> appended is not correct), you need to specify a full <code>general-add-hook<\/code> arglist:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package yasnippet\n  :ghook ('(text-mode-hook prog-mode-hook) #'yas-minor-mode))<\/pre><\/div>\n<h4><a id=\"user-content-gfhook-keyword\" class=\"anchor\" aria-hidden=\"true\" href=\"#gfhook-keyword\"><\/a>:gfhook Keyword<\/h4>\n<p><code>:gfhook<\/code> is intended to be used to specify <i>functions<\/i> to add to the package\u2019s mode hook. The hook is inferred from the package\u2019s name (by appending either <code>-mode-hook<\/code> or just <code>-hook<\/code> if the package\u2019s name ends in <code>-mode<\/code>). If the hook cannot be inferred from the package name, then the full arglist must be specified just as with <code>:ghook<\/code>. Unlike <code>:ghook<\/code>, <code>:gfhook<\/code> never adds functions to <code>:commands<\/code> and therefore never implies <code>:defer t<\/code>. This is because the functions specified are ones that should be run when turning on (or toggling) the mode(s) the package provides. The specified functions are external to the package, could be called elsewhere, and therefore should not trigger the package to load. The following all have the same effect:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package org\n  ;; for a major-mode package, you might use :mode to imply :defer t (or just\n  ;; use :defer t; or just `use-package-always-defer' which I personally prefer)\n  :gfhook\n  #'visual-line-mode\n  #'my-org-setup\n  ;; ...\n  )\n(use-package org\n  :init\n  (general-add-hook 'org-mode-hook #'visual-line-mode)\n  (general-add-hook 'org-mode-hook #'my-org-setup))\n;; this is also valid but less concise\n(use-package org\n  ;; specify null or non-symbol placeholder for HOOKS to use inferred hook\n  :gfhook (nil (list #'visual-line-mode #'my-org-setup)))\n(use-package org\n  :init\n  (general-add-hook 'org-mode-hook (list #'visual-line-mode #'my-org-setup)))<\/pre><\/div>\n<p>Like with <code>:ghook<\/code>, <code>:gfhook<\/code> still requires quoting, so you can use variables and function\/macro calls to generate the function to add to the hook:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmacro disable (mode)\n  `(lambda () (,mode -1)))\n(use-package proced\n  ;; must be in a `general-add-hook' argument list, so that it itself is not\n  ;; considered one\n  :gfhook (nil (disable visual-line-mode)))<\/pre><\/div>\n<p>Although you could use <code>:gfhook<\/code> to enable minor modes for some major mode (e.g. enable flyspell inside <code>(use-package org)<\/code>), it is probably more logical\/organized to group these hooks along with their minor modes\u2019 use-package declarations (e.g. using <code>:ghook<\/code>). <code>:gfhook<\/code> is more suited for setup functions. Expanding on the proced example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun my-proced-setup ()\n  (visual-line-mode -1)\n  ;; not global; has to be run in buffer\n  (proced-toggle-auto-update t))\n(use-package proced\n  :gfhook #'my-proced-setup)<\/pre><\/div>\n<h2><a id=\"user-content-use-with-key-chord\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-with-key-chord\"><\/a>Use with Key-chord<\/h2>\n<p>General provides a simple function that will rewrite a string into a key-chord vector. This allows you to easily use general to create definitions for <code>key-chord.el<\/code>. The following are equivalent:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(key-chord-define evil-insert-state-map \"jk\" 'evil-normal-state)\n(general-define-key :keymaps 'evil-insert-state-map\n                    (general-chord \"jk\") 'evil-normal-state\n                    (general-chord \"kj\") 'evil-normal-state)<\/pre><\/div>\n<p>Note that the order of the keys does matter unlike with the default <code>key-chord-define<\/code>.<\/p>\n<p>General.el supports some extra per-definition keywords. It has \u201ctype\u201d keywords that give general.el some extra information to use to create definitions (e.g. <code>:prefix-command<\/code> and <code>:keymap<\/code>) and other keywords that will alter or ignore definitions (e.g. <code>:predicate<\/code> and <code>:ignore<\/code>).<\/p>\n<p>The system that allows for the default keywords can also be extended by the user to support more keywords that can either directly alter the definition or just be used for side effects (like <code>:which-key<\/code>). An extended definition keyword can have any number of helper keywords (and can also be used as a helper keyword itself, e.g. <code>:keymap<\/code>). See <a href=\"#user-defined-extended-definition-keywords\">User-defined Extended Definition Keywords<\/a> for more information on creating new keywords.<\/p>\n<p>Here are the keywords available by default (helper keywords are subitems; specific examples are given later):<\/p>\n<ul><li><code>:def<\/code> - Implicit; this is paired with the actual definition (helper keyword; does not trigger any special behavior by itself)<\/li>\n<\/ul><p>\u201cType\u201d specifiers:<\/p>\n<ul><li><code>:keymap<\/code> - For keymaps; if the keymap is not defined, will create an \u201cautoloaded\u201d keymap for <code>:package<\/code>\n    <ul><li><code>:package<\/code> - The package to load (also global)<\/li>\n    <\/ul><\/li>\n  <li><code>:prefix-command<\/code> and\/or <code>:prefix-map<\/code> - These are the same as <code>:def<\/code> and <code>:keymap<\/code> respectively but will create a prefix command and\/or keymap (these behave the same as the global keyword arguments except for any key as opposed to just <code>:prefix<\/code>)\n    <ul><li><code>:prefix-name<\/code> The keymap menu name\/prompt (global value never considered)<\/li>\n    <\/ul><\/li>\n  <li><code>:ignore<\/code> - Do not create a keybinding for the key def pair<\/li>\n<\/ul><p>Note that every bindable definition must have <code>:def<\/code>, but general allows for shorthand where <code>:def<\/code> can be omitted or a \u201ctype\u201d specifier can be used instead:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; shorthand\n'(swiper :wk \"swipe\")\n;; rewritten to\n'(:def swiper :wk \"swipe\")\n;; shorthand\n'(:keymap some-keymap)\n;; rewritten to\n'(:def some-keymap :keymap some-keymap)\n;; same as\n'(:def some-keymap :keymap t)\n;; shorthand\n'(:prefix-command my-prefix-cmd :prefix-map my-prefix-map)\n;; rewritten to\n'(:def my-prefix-cmd :prefix-command my-prefix-cmd :prefix-map my-prefix-map)<\/pre><\/div>\n<p>After the shorthand expansion, the type keywords are handled exactly the same as any other extended definition keyword.<\/p>\n<p>Which-key functionality (see below for more details):<\/p>\n<ul><li><code>:which-key<\/code> or <code>:wk<\/code> - The replacement text (or cons or function)\n    <ul><li><code>:major-modes<\/code> - Major modes to match (optional; also global)<\/li>\n      <li><code>:wk-match-keys<\/code> - Whether to include the keys in the match cons (defaults to <code>t<\/code> globally)<\/li>\n      <li><code>:wk-match-binding<\/code> - Whether to include the binding in the match cons (defaults to <code>t<\/code>; also global)<\/li>\n      <li><code>:wk-full-keys<\/code> - Whether the bound keys correspond to the full sequence to match (defaults to <code>t<\/code>; also global)<\/li>\n      <li><code>:keymap<\/code> - When non-nil, general will not try to match a keymap symbol as if it was a command<\/li>\n    <\/ul><\/li>\n<\/ul><p>Evil command properties (see below for more details):<\/p>\n<ul><li><code>:properties<\/code> - The list of properties to add to the command (also global)<\/li>\n  <li><code>:repeat<\/code> - The repeat property to set for the command (also global)<\/li>\n  <li><code>:jump<\/code> - The jump property to set for the command (also global)<\/li>\n<\/ul><p>Global keywords that can be overridden locally:<\/p>\n<ul><li><code>:predicate<\/code><\/li>\n<\/ul><p>The default value for a keyword is <code>nil<\/code> unless otherwise specified.<\/p>\n<h2><a id=\"user-content-autoloaded-keymaps\" class=\"anchor\" aria-hidden=\"true\" href=\"#autoloaded-keymaps\"><\/a>\u201cAutoloaded\u201d Keymaps<\/h2>\n<p>As the first example, an extended definition can be used to create an \u201cautoload\u201d for a keymap like use-package\u2019s <code>:bind-keymap<\/code> keyword does:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n \"C-c p\" '(:keymap projectile-command-map :package projectile))<\/pre><\/div>\n<p>Using this feature, a key can be bound to a keymap that does not exist yet and still work as expected. Projectile will be loaded when <code>C-c p<\/code> is used for the first time. This is done by using an intermediate function to load the package and rebind the keys.<\/p>\n<p><code>:keymap<\/code> is the primary keyword that triggers this check. It can also be used as a helper keyword (e.g. for <code>which-key<\/code>). If the keymap already exists, general will not try to create an autoloaded keymap, and <code>:package<\/code> is not required.<\/p>\n<p><code>:package<\/code> is a helper keyword that can be specified locally within the extended definition or globally. When using the use-package <code>:general<\/code> keyword, it will automatically be specified.<\/p>\n<h2><a id=\"user-content-which-key-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#which-key-integration\"><\/a>Which Key Integration<\/h2>\n<p>If you are not already familiar with which-key\u2019s replacement system, please see the docstring for <code>which-key-replacement-alist<\/code> if you don\u2019t understand any of the examples or information here.<\/p>\n<p>There are several benefits to using general.el to add which-key replacements. The main benefit is that because the keys and definition are already specified, general.el can automatically assemble the match cons. This reuse of information saves a little space since it is not necessary to make an additional call to <code>which-key-add-key-based-replacements<\/code> with the key information. It is also useful since which-key does not currently provide any convenience function for creating a replacement that matches a binding (you have to manually add to <code>which-key-replacement-alist<\/code>). However, see which-key\u2019s <a href=\"https:\/\/github.com\/justbur\/emacs-which-key#automatic\">which-key-enable-extended-define-key<\/a> which provides another method for automatically creating replacements and binding keys simultaneously.<\/p>\n<p>Another related benefit of using <code>:which-key<\/code> instead of <code>which-key-add-key-based-replacements<\/code> directly even for keys that won\u2019t be bound is that replacements will be added for all prefix combinations (i.e. when <code>:non-normal-prefix<\/code> and\/or <code>:global-prefix<\/code> are also specified).<\/p>\n<p>The argument supplied to <code>:which-key<\/code> or <code>:wk<\/code> is equivalent to the REPLACEMENT argument in <code>which-key-add-key-based-replacements<\/code>. It can be a full replacement cons of <code>(KEY . BINDING)<\/code> or just a string (which will be used as the BINDING and serve as the new description). Additionally it can be a function that will return a replacement cons (see the docstring for <code>which-key-replacements-alist<\/code> or the which-key README).<\/p>\n<p>The <code>:which-key<\/code> keyword can be used with the <code>:major-modes<\/code> keyword (locally or globally) which can be compared to using <code>which-key-add-major-mode-key-based-replacements<\/code>. <code>:major-modes<\/code> can have the following values (see the examples below):<\/p>\n<ul><li><code>t<\/code> - the major mode will be obtained from all keymaps by removing \u201c-map\u201d<\/li>\n  <li>the major mode name (when only one keymap is specified)<\/li>\n  <li>a list of the following values:\n    <ul><li><code>t<\/code> - same behavior as above but only for corresponding index in <code>:keymaps<\/code><\/li>\n      <li>the major mode name for that index<\/li>\n      <li><code>nil<\/code> (or no item at the index) - don\u2019t match the major mode<\/li>\n    <\/ul><\/li>\n<\/ul><p><code>:wk-match-keys<\/code>, <code>:wk-match-binding<\/code>, and <code>:wk-full-keys<\/code> can be used to customize the match cons. Generally these will not need to be adjusted. The binding is only included in the match cons if one is available, and <code>:wk-full-keys<\/code> only needs to be specified as <code>nil<\/code> if you are binding keys in a prefix map.<\/p>\n<p>Here are some examples:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :prefix \"SPC\"\n :keymaps 'normal\n ;; unbind SPC and give it a title for which-key (see echo area)\n \"\" '(nil :which-key \"my lieutenant general prefix\")\n ;; bind nothing but give SPC f a description for which-key\n \"f\" '(:ignore t :which-key \"file prefix\")\n ;; use a cons as a replacement\n \"g\" '(:ignore t :wk (\"g-key\" . \"git prefix\"))\n ;; toggle lispy; use a function as a replacement to show if currently on\n \"l\" '(lispy-mode :wk my-lispy-which-key-display)\n ;; for a keymap, only the keys will be matched;\n ;; :no-match-binding is not necessary\n \"p\" '(:keymap projectile-command-map :wk \"projectile prefix\")\n ...)\n(general-define-key\n :keymaps 'help-map\n ;; allow keys before bound keys in match\n ;; since binding in a prefix map\n :wk-full-keys nil\n ;; make a prefix-command and add description\n \"A\" '(:prefix-command apropos-prefix-map :which-key \"apropos\"))\n;; an equivalent of the above\n(general-define-key :keymaps 'help-map\n  :wk-full-keys nil\n  :prefix \"A\"\n  :prefix-command 'apropos-prefix-map\n  ;; make a prefix-command and add description\n  \"\" '(:ignore t :which-key \"apropos\"))\n;; :major-modes\n(general-define-key\n :keymaps 'emacs-lisp-mode-map\n :major-modes t\n ...)\n(general-define-key\n :keymaps '(no-follow-convention-mode-keymap1\n            org-mode-map)\n :major-modes '(no-follow-convention-mode t)\n ...)<\/pre><\/div>\n<h2><a id=\"user-content-evil-command-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#evil-command-properties\"><\/a>Evil Command Properties<\/h2>\n<p>The <code>:properties<\/code>, <code>:repeat<\/code>, and <code>:jump<\/code> keywords can be used to add evil command properties:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :keymaps 'normal\n :prefix \"SPC\"\n \"gj\" '(git-gutter:next-hunk :properties (:repeat t :jump t))\n \"gk\" '(git-gutter:previous-hunk :repeat t :jump t))\n;; they also work globally\n(general-define-key\n :keymaps 'normal\n :prefix \"SPC\"\n :properties '(:repeat t :jump t)\n ;; or\n :repeat t\n :jump t\n \"gj\" 'git-gutter:next-hunk\n \"gk\" 'git-gutter:previous-hunk)\n<\/pre><\/div>\n<p>Note that the default for commands without a repeat property are treated the same as commands with <code>:repeat t<\/code>, so the above repeat configuration isn\u2019t explicitly necessary in this case.<\/p>\n<p>If you would like for more keywords to be added that correspond to specific properties (like <code>:repeat<\/code>), feel free to make an issue or pull request. For more information on command properties see evil\u2019s documentation and <a href=\"https:\/\/github.com\/noctuid\/evil-guide#command-properties\">here<\/a>.<\/p>\n<h2><a id=\"user-content-user-defined-extended-definition-keywords\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-defined-extended-definition-keywords\"><\/a>User-defined Extended Definition Keywords<\/h2>\n<p>New keywords and functionality can be added by the user by adding a keyword to <code>general-extended-def-keywords<\/code> and creating a corresponding function named <code>general-extended-def-:&lt;keyword&gt;<\/code>.<\/p>\n<p>Whenever this keyword is specified, general calls the corresponding function with the arguments <code>state keymap key edef kargs<\/code>. Generally, you can ignore at least some of these arguments. <code>state<\/code> and <code>keymap<\/code> are the evil state (nil if none) and keymap that the <code>key<\/code> (internal representation; <code>kbd<\/code> already used if necessary) is being bound in. Note that <code>keymap<\/code> will be the symbol for the keymap in case it is needed. To get the actual keymap, using <code>general--get-keymap<\/code> is recommended. <code>edef<\/code> is the extended definition itself, and <code>kargs<\/code> is the plist of all the keyword arguments given to the original <code>general-define-key<\/code>.<\/p>\n<p>Extended definition functions can optionally alter the definitions. Keywords that have this behavior must be added to either <code>general-rewrite-def-keywords<\/code> or <code>general-rewrite-def-after-keywords<\/code> instead of to <code>general-extended-def-keywords<\/code>. The difference between the two is that the former will alter the definition before the functions for the keywords in <code>general-extended-def-keywords<\/code> are called. Functions that alter the definition should return a new extended definition plist with the <code>:def<\/code> entry updated. For a simple example of a function that does not alter the definition, see <code>general-extended-def-:properties<\/code>. For a simple example of a function that does alter the definition, see <code>general-extended-def-:predicate<\/code>.<\/p>\n<p>Extended definition keywords may use any number of helper keywords. These do not need to be added to any variables but should be distinct from any other keywords.<\/p>\n<p>Note that the keywords in <code>general-extended-def-keywords<\/code> and their helper keywords can all be specified both globally and locally. Since globally specifying keywords may not always make sense, it is up to the <code>general-extended-def-:&lt;keyword&gt;<\/code> function to decide how to handle things. When a keyword can be specified both globally and locally, <code>general--getf<\/code> may be useful to get the local value or the global value if there is no local one (e.g. <code>(general--getf edef kargs :predicate)<\/code>). If it does not make sense for your keyword to be specified globally, you can add it to <code>general-extended-def-global-ignore-keywords<\/code>. This will prevent your function from being called unless the keyword is specified locally.<\/p>\n<p>Although <code>general--get-keymap<\/code> and <code>general--getf<\/code> are marked internal, they will continue to exist and keep their current functionality; they are intended to be used as helpers for extended definitions.<\/p>\n<p>You can rely on <code>edef<\/code> being a valid extended definition plist with a <code>:def<\/code> keyword. Even if the user only specifies a keyword globally and does not explicitly write definitions as plists or explicitly specify <code>:def<\/code>, general will automatically rewrite definitions to be valid plists. Consider the following example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :predicate '(eobp)\n \"&lt;right&gt;\" 'beginning-of-buffer)\n;; call `general-extended-def-:predicate' with this as an edef argument:\n'(:def beginning-of-buffer)\n(general-define-key\n \"&lt;right&gt;\" '(beginning-of-buffer :predicate (eobp)))\n;; call `general-extended-def-:predicate' with this as an edef argument:\n'(:def beginning-of-buffer :predicate (eobp))<\/pre><\/div>\n<p>For more information, see the docstring of <code>general-extended-def-keywords<\/code>.<\/p>\n<p>In addition to being able to add new keywords for extended definitions, the user can also create their own key definers. These are generally useful when you want to use some package-specific key definer that has some additional functionality (e.g. <code>lispy-define-key<\/code>).<\/p>\n<p>Alternate definers can be used by specifying the <code>:definer<\/code> keyword (globally or inside an extended definition):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key :definer 'my\n  \"key\" 'def\n  \"key2\" '(def2 :definer 'my-other))<\/pre><\/div>\n<p>The user-created function should be named <code>general-&lt;definer&gt;-define-key<\/code>. It will be passed <code>state keymap key def orig-def kargs<\/code>. These arguments are the same as for extended definition functions except for <code>def<\/code> and <code>orig-def<\/code>. <code>def<\/code> is the transformed definition in its final form (though the definer may also alter it before binding it). On the other hand, <code>orig-def<\/code> is the original definition but always as an extended definition plist (e.g. =\u2019(:def command)= if the user only specified =\u2019command=).<\/p>\n<p>Like extended definitions, custom definers can have any number of helper keyword arguments specified locally in an extended definition or globally in the arguments to <code>general-define-key<\/code>. In cases where a keyword can be both global and local, <code>general--getf<\/code> is a useful helper function. Since the keymap passed in is a symbol, <code>general--get-keymap<\/code> may be useful as well for transforming it to the keymap value. <code>key-description<\/code> will also be useful if the underlying definition function uses <code>kbd<\/code> (since <code>key<\/code> is the internal representation ready to be passed directly to <code>define-key<\/code>; note that <code>key-description<\/code> will work with both strings and vectors, including something like <code>[remap kill-line]<\/code>).<\/p>\n<p>See <code>general-lispy-define-key<\/code> for a basic example.<\/p>\n<h2><a id=\"user-content-wrapping-evil-define-minor-mode-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#wrapping-evil-define-minor-mode-key\"><\/a>Wrapping <code>evil-define-minor-mode-key<\/code><\/h2>\n<p>If you want to use <code>evil-define-minor-mode-key<\/code> instead of <code>evil-define-key*<\/code>, you can use <code>:definer 'minor-mode<\/code>. This will repurpose <code>:keymaps<\/code> to specify minor mode names instead of keymap names:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :definer 'minor-mode\n :states 'normal\n :keymaps 'org-src-mode\n \"RET\" 'org-edit-src-exit)<\/pre><\/div>\n<p>If you are wondering why you might want to use <code>evil-define-minor-mode-key<\/code>, see <a href=\"https:\/\/github.com\/noctuid\/evil-guide#why-dont-keys-defined-with-evil-define-key-work-immediately\">here<\/a>.<\/p>\n<h2><a id=\"user-content-lispy-integration-wrapping-lispy-define-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#lispy-integration-wrapping-lispy-define-key\"><\/a>Lispy Integration\/ Wrapping <code>lispy-define-key<\/code><\/h2>\n<p>To use <code>lispy-define-key<\/code> to make the definitions, <code>:definer 'lispy<\/code> can be specified. <code>:lispy-plist<\/code> can be specified globally or in an extended definition to set the last argument to <code>lispy-define-key<\/code>.<\/p>\n<h2><a id=\"user-content-worf-integration-wrapping-worf-define-key\" class=\"anchor\" aria-hidden=\"true\" href=\"#worf-integration-wrapping-worf-define-key\"><\/a>Worf Integration\/ Wrapping <code>worf-define-key<\/code><\/h2>\n<p>To use <code>worf-define-key<\/code> to make the definitions, <code>:definer 'worf<\/code> can be specified. <code>:worf-plist<\/code> can be specified globally or in an extended definition to set the last argument to <code>worf-define-key<\/code>.<\/p>\n<h2><a id=\"user-content-other-provided-definers\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-provided-definers\"><\/a>Other Provided Definers<\/h2>\n<p>To use <code>lpy-define-key<\/code> to make the definitions, <code>:definer 'lpy<\/code> can be specified.<\/p>\n<h2><a id=\"user-content-how-do-i-prevent-key-sequence-starts-with-non-prefix-key-errors\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-do-i-prevent-key-sequence-starts-with-non-prefix-key-errors\"><\/a>How do I prevent <code>Key sequence starts with non-prefix key<\/code> errors?<\/h2>\n<p>By default, emacs does not support binding a key sequence where a subsequence of the key is already bound in the same keymap (e.g. you cannot bind <code>C-a a<\/code> to a command in a keymap where <code>C-a<\/code> is already bound to a command).<\/p>\n<p>If you want to be able to bind both key sequences and fall back to the shorter key\u2019s command after a timeout or unmatched keypress, see <a href=\"#mapping-under-non-prefix-keys\">general-key-dispatch<\/a>.<\/p>\n<p>Otherwise, you should unbind the non-prefix key. For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(general-define-key\n :keymaps 'normal\n :prefix \"s\"\n ;; prefix keys are prepended to other keys, so \"\" refers to the prefix itself\n \"\" nil\n \"a\" #'def\n ;; ...\n )<\/pre><\/div>\n<p>If you would rather force key definitions to always be made regardless of whether a subsequence of the key is already bound, <a href=\"#automatic-key-unbinding\">general can automatically unbind keys when necessary to prevent this error<\/a>.<\/p>\n<h2><a id=\"user-content-why-dont-some-evil-keybindings-work-immediately\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-dont-some-evil-keybindings-work-immediately\"><\/a>Why don\u2019t some evil keybindings work (immediately)?<\/h2>\n<p>This is a <a href=\"https:\/\/github.com\/emacs-evil\/evil\/issues\/301\">known issue for evil<\/a>. To work around this problem, you can use <a href=\"#wrapping-evil-define-minor-mode-key\">:definer \u2018minor-mode<\/a>. See <a href=\"https:\/\/github.com\/noctuid\/evil-guide#why-dont-keys-defined-with-evil-define-key-work-immediately\">here<\/a> for more information.<\/p>\n<\/article>","created_at":"2018-11-28T22:44:40+0000","updated_at":"2018-11-28T22:44:41+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":67,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/4250696?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["css","webdev","retro"],"is_public":false,"id":725,"title":"NES.css","url":"https:\/\/bcrikko.github.io\/NES.css\/","content":"NES.css - NES-style CSS Framework\n    <header>\n      <p>NES-style CSS Framework.<\/p>\n    <\/header><a class=\"github-link\" href=\"https:\/\/github.com\/BcRikko\/NES.css\" target=\"_blank\">\n      <p class=\"balloon from-right\">Fork me<br \/>on GitHub<\/p>\n      <i class=\"octocat\">\n    <\/i><\/a>\n    <section class=\"container with-title\"><h2 class=\"title\">Buttons<\/h2><\/section><section class=\"container with-title\"><h2 class=\"title\">Containers<\/h2><div class=\"containers\"><div class=\"container with-title is-center\"><label class=\"title\">Container.is-center<\/label><p>Good morning. Thou hast had a good night's sleep I hope.<\/p><\/div><div class=\"container is-dark with-title\"><label class=\"title\">Container.is-dark<\/label><p>Good morning. Thou hast had a good night's sleep I hope.<\/p><\/div><div class=\"container is-rounded\"><p>Good morning. Thou hast had a good night's sleep I hope.<\/p><\/div><div class=\"container is-rounded is-dark\"><p>Good morning. Thou hast had a good night's sleep I hope.<\/p><\/div><\/div><\/section><section class=\"container with-title\"><h2 class=\"title\">Radios<\/h2><p><label>\n          \n          Yes\n        <\/label>\n        <label>\n          \n          No\n        <\/label><\/p><\/section><p><\/p><h2 class=\"title\">Checkboxes<\/h2>\n      <label>\n        \n        Enable\n      <\/label><section class=\"form container with-title\"><h2 class=\"title\">Form<\/h2><p><label>Your name<\/label>\n        <\/p><p><label>.input.is-success<\/label>\n        <\/p><p><label>.input.is-warning<\/label>\n        <\/p><p><label>.input.is-error<\/label>\n        <\/p><\/section><section class=\"balloon container with-title\"><h2 class=\"title\">Balloons<\/h2><div class=\"messages\"><div class=\"message -left\"><i class=\"bcrikko\">Hello NES.css<\/i><\/div><\/div><div class=\"message -right\"><div class=\"balloon from-right\"><p>Good morning. Thou hast had a good night's sleep I hope.<\/p><\/div><i class=\"bcrikko\"><\/i><\/div><\/section><section class=\"container with-title\"><h2 class=\"title\">Icons<\/h2><section class=\"container with-title\"><h3 class=\"title\">Reaction<\/h3><\/section><section class=\"container with-title\"><h3 class=\"title\">SNS<\/h3><\/section><section class=\"container with-title\"><h3 class=\"title\">Others<\/h3><\/section><\/section>","created_at":"2018-11-28T14:52:10+0000","updated_at":"2018-11-28T14:52:14+0000","annotations":[],"mimetype":"text\/html","language":"ja","reading_time":0,"domain_name":"bcrikko.github.io","preview_picture":"https:\/\/user-images.githubusercontent.com\/5305599\/49061716-da649680-f254-11e8-9a89-d95a7407ec6a.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","dgraph"],"is_public":false,"id":724,"title":"jepsen-io\/jepsen","url":"https:\/\/github.com\/jepsen-io\/jepsen\/blob\/master\/dgraph\/src\/jepsen\/dgraph\/client.clj","content":"(ns jepsen.dgraph.client\n      \n          \"Clojure wrapper for the Java DGraph client library. This whole thing is\n      \n          *riddled* with escaping vulnerabilities, so, you know, tread carefully.\"\n      \n          (:require [clojure.string :as str]\n      \n                    [clojure.tools.logging :refer [info warn]]\n      \n                    [dom-top.core :refer [with-retry]]\n      \n                    [wall.hack]\n      \n                    [cheshire.core :as json]\n      \n                    [jepsen.client :as jc]\n      \n                    [jepsen.dgraph.trace :as t])\n      \n          (:import (java.util.concurrent TimeUnit)\n      \n                   (com.google.protobuf ByteString)\n      \n                   (io.grpc ManagedChannel\n      \n                            ManagedChannelBuilder)\n      \n                   (io.dgraph DgraphGrpc\n      \n                              DgraphGrpc$DgraphStub\n      \n                              DgraphClient\n      \n                              DgraphAsyncClient\n      \n                              Transaction\n      \n                              DgraphProto$Assigned\n      \n                              DgraphProto$Mutation\n      \n                              DgraphProto$Response\n      \n                              DgraphProto$Operation\n      \n                              TxnConflictException)))\n      \n        \n\n      \n        (def default-port \"Default dgraph alpha GRPC port\" 9080)\n      \n        \n\n      \n        ;; milliseconds given to the grpc blockingstub as a deadline\n      \n        (def deadline 30000)\n      \n        \n\n      \n        (defn open\n      \n          \"Creates a new DgraphClient for the given node.\"\n      \n          ([node]\n      \n           (t\/with-trace \"client.open!\"\n      \n             (open node default-port)))\n      \n          ([node port]\n      \n           (t\/with-trace \"client.open!\"\n      \n             (let [channel (.. (ManagedChannelBuilder\/forAddress node port)\n      \n                               (usePlaintext true)\n      \n                               (build))\n      \n                   stub  (DgraphGrpc\/newStub channel)\n      \n                   stub  (.withDeadlineAfter stub deadline TimeUnit\/MILLISECONDS)\n      \n                   stubs (into-array DgraphGrpc$DgraphStub [stub])]\n      \n               (DgraphClient. stubs)))))\n      \n        \n\n      \n        (defn close!\n      \n          \"Closes a client. Close is asynchronous; resources may be freed some time\n      \n          after calling (close! client).\"\n      \n          [client]\n      \n          (t\/with-trace \"client.close!\"\n      \n            (let [async-client (wall.hack\/field DgraphClient :asyncClient client)]\n      \n              (doseq [c (wall.hack\/field DgraphAsyncClient :stubs async-client)]\n      \n                (.. c getChannel shutdown)))))\n      \n        \n\n      \n        (defn abort-txn!\n      \n          \"Aborts a transaction object.\"\n      \n          [^Transaction t]\n      \n          (t\/with-trace \"client.abort-txn!\"\n      \n            (try (.discard t)\n      \n               (catch io.grpc.StatusRuntimeException e\n      \n                 (if (re-find #\"ABORTED: Transaction has been aborted\\. Please retry\\.\"\n      \n                              (.getMessage e))\n      \n                   :aborted\n      \n                   (throw e))))))\n      \n        \n\n      \n        (defmacro with-txn\n      \n          \"Takes a vector of a symbol and a client. Opens a transaction on the client,\n      \n          binds it to that symbol, and evaluates body. Calls commit at the end of\n      \n          the body, or discards the transaction if an exception is thrown.\n      \n        \n      \n          If you commit or abort the transaction *within* body (e.g. before with-txn\n      \n          commits it for you), with-txn will attempt to commit, *not* throw, and return\n      \n          the result of `body`.\"\n      \n          [[txn-sym client] &amp; body]\n      \n          `(let [~txn-sym (.newTransaction ^DgraphClient ~client)]\n      \n             (try\n      \n               (let [res# (do ~@body)]\n      \n                 (try (.commit ~txn-sym)\n      \n                      ;; if our trasaction aborts when we try to complete it, we\n      \n                      ;; want to pass the value back up\n      \n                      (catch io.grpc.StatusRuntimeException e#\n      \n                        (if (re-find #\"ABORTED\" (.getMessage e#))\n      \n                          res#\n      \n                          (throw e#)))\n      \n                      ;; If the user manually committed or aborted, that's OK.\n      \n                      (catch io.dgraph.TxnFinishedException e#\n      \n                        nil))\n      \n                 res#)\n      \n               (finally\n      \n                 (.discard ~txn-sym)))))\n      \n        \n\n      \n        (defmacro with-unavailable-backoff\n      \n          \"Wraps an expression returning a completion operation; for selected failure\n      \n          modes, sleeps a random amount of time before returning, so we don't spin our\n      \n          wheels against a down system.\"\n      \n          [&amp; body]\n      \n          `(let [res# (do ~@body)]\n      \n             (when (and (= :fail (:type res#))\n      \n                        (#{:unavailable\n      \n                           :predicate-moving\n      \n                           :unhealthy-connection} (:error res#)))\n      \n               (Thread\/sleep (rand-int 2000)))\n      \n             res#))\n      \n        \n\n      \n        (defmacro with-conflict-as-fail\n      \n          \"Takes an operation and a body. Evaluates body; if a transaction conflict is\n      \n          thrown, returns `op` with :type :fail, :error :conflict.\"\n      \n          [op &amp; body]\n      \n          `(with-unavailable-backoff\n      \n             (try ~@body\n      \n                  (catch io.grpc.StatusRuntimeException e#\n      \n                    (condp re-find (.getMessage e#)\n      \n                      #\"DEADLINE_EXCEEDED:\"\n      \n                      (assoc ~op, :type :info, :error :timeout)\n      \n        \n\n      \n                      #\"context deadline exceeded\"\n      \n                      (assoc ~op, :type :info, :error :timeout)\n      \n        \n\n      \n                      #\"Conflicts with pending transaction. Please abort.\"\n      \n                      (assoc ~op :type :fail, :error :conflict)\n      \n        \n\n      \n                      #\"readTs: \\d+ less than minTs: \\d+ for key:\"\n      \n                      (assoc ~op :type :fail, :error :old-timestamp)\n      \n        \n\n      \n                      #\"Predicate is being moved, please retry later\"\n      \n                      (assoc ~op :type :fail, :error :predicate-moving)\n      \n        \n\n      \n                      #\"Tablet isn't being served by this instance\"\n      \n                      (assoc ~op :type :fail, :error :tablet-not-served-by-instance)\n      \n        \n\n      \n                      #\"Request sent to wrong server\"\n      \n                      (assoc ~op :type :fail, :error :wrong-server)\n      \n        \n\n      \n                      #\"Please retry again, server is not ready to accept requests\"\n      \n                      (assoc ~op :type :fail, :error :not-ready-for-requests)\n      \n        \n\n      \n                      #\"UNAVAILABLE\"\n      \n                      (assoc ~op, :type :fail, :error :unavailable)\n      \n        \n\n      \n                      #\"No connection exists\"\n      \n                      (assoc ~op :type :fail, :error :no-connection)\n      \n        \n\n      \n                      ; Guessssing this means it couldn't even open a conn but not sure\n      \n                      ; This might be a fail???\n      \n                      #\"Unavailable desc = all SubConns are in TransientFailure\"\n      \n                      (assoc ~op :type :info, :error :unavailable-all-subconns-down)\n      \n        \n\n      \n                      #\"rpc error: code = Unavailable desc = transport is closing\"\n      \n                      (assoc ~op :type :info, :error :unavailable-transport-closing)\n      \n        \n\n      \n                      #\"dispatchTaskOverNetwork: while retrieving connection. error: Unhealthy connection\"\n      \n                      (assoc ~op :type :info, :error :unhealthy-connection)\n      \n        \n\n      \n                      #\"Only leader can decide to commit or abort\"\n      \n                      (assoc ~op :type :fail, :error :only-leader-can-commit)\n      \n        \n\n      \n                      #\"This server doesn't serve group id:\"\n      \n                      (assoc ~op :type :fail, :error :server-doesn't-serve-group)\n      \n        \n\n      \n                      ;; FIXME For some reason these don't get caught??\n      \n                      #\"ABORTED\"\n      \n                      (assoc ~op :type :fail, :error :transaction-aborted)\n      \n        \n\n      \n                      (throw e#)))\n      \n        \n\n      \n                  (catch TxnConflictException e#\n      \n                    (assoc ~op :type :fail, :error :conflict)))))\n      \n        \n\n      \n        (defn str-&gt;byte-string\n      \n          \"Converts a string to a protobuf bytestring.\"\n      \n          [s]\n      \n          (ByteString\/copyFromUtf8 s))\n      \n        \n\n      \n        (defn alter-schema!\n      \n          \"Takes a schema string (or any number of strings) and applies that alteration\n      \n          to dgraph. Retries if DEADLINE_EXCEEDED, since dgraph likes to throw this for\n      \n          ??? reasons at the start of the test. Should be idempotent, so... hopefully\n      \n          we can retry, at least in this context?\"\n      \n          [^DgraphClient client &amp; schemata]\n      \n          (t\/with-trace \"client.alter-schema!\"\n      \n            (with-retry [i 0]\n      \n              (.alter client (.. (DgraphProto$Operation\/newBuilder)\n      \n                                 (setSchema (str\/join \"\\n\" schemata))\n      \n                                 build))\n      \n              (catch io.grpc.StatusRuntimeException e\n      \n                (if (and (&lt; i 3)\n      \n                         (re-find #\"DEADLINE_EXCEEDED\" (.getMessage e)))\n      \n                  (do (warn \"Alter schema failed due to DEADLINE_EXCEEDED, retrying...\")\n      \n                      (retry (inc i)))\n      \n                  (throw e))))))\n      \n        \n\n      \n        (defn ^DgraphProto$Assigned mutate!*\n      \n          \"Takes a mutation object and applies it to a transaction. Returns an\n      \n          Assigned.\"\n      \n          [^Transaction txn mut]\n      \n          ;(info \"Mutate:\" mut)\n      \n          (.mutate txn (.. (DgraphProto$Mutation\/newBuilder)\n      \n                           (setSetJson (str-&gt;byte-string (json\/generate-string mut)))\n      \n                           build)))\n      \n        \n\n      \n        (defn mutate!\n      \n          \"Like mutate!*, but returns a map of key names to UID strings.\"\n      \n          [txn mut]\n      \n          (t\/with-trace \"client.mutate\"\n      \n            (.getUidsMap (mutate!* txn mut))))\n      \n        \n\n      \n        (defn ^DgraphProto$Assigned set-nquads!*\n      \n          \"Takes a transaction and an n-quads string, and adds those set mutations to\n      \n          the transaction.\"\n      \n          [^Transaction txn nquads]\n      \n          (.mutate txn (.. (DgraphProto$Mutation\/newBuilder)\n      \n                           (setSetNquads (str-&gt;byte-string nquads))\n      \n                           build)))\n      \n        \n\n      \n        (defn set-nquads!\n      \n          \"Like set-nquads!*, but returns a map of key names to UID strings.\"\n      \n          [txn nquads]\n      \n          (t\/with-trace \"client.set-nquads!\"\n      \n            (.getUidsMap (set-nquads!* txn nquads))))\n      \n        \n\n      \n        (defn check-str-or-map\n      \n          \"If the given value is a string, wraps it in a map with the :uid field.\"\n      \n          [x]\n      \n          (if (string? x)\n      \n            {:uid x}\n      \n            x))\n      \n        \n\n      \n        (defn delete!\n      \n          \"Deletes a record. Can take either a map (treated as a JSON deletion), or a\n      \n          UID string, in which case every outbound edge for the given entity is\n      \n          deleted.\"\n      \n          [^Transaction txn str-or-map]\n      \n          (t\/with-trace \"client.delete!\"\n      \n            (let [target (check-str-or-map str-or-map)]\n      \n              (.mutate txn (.. (DgraphProto$Mutation\/newBuilder)\n      \n                               (setDeleteJson (-&gt; target\n      \n                                                  json\/generate-string\n      \n                                                  str-&gt;byte-string))\n      \n                               build)))))\n      \n        \n\n      \n        (defn graphql-type\n      \n          \"Takes an object and infers a type in the query language, e.g.\n      \n        \n      \n              \\\"4\\\" -&gt; \\\"string\\\",\n      \n              4     -&gt; \\\"int\\\"\"\n      \n          [x]\n      \n          (when-not x\n      \n            (throw (IllegalArgumentException.\n      \n                     \"Can't infer graphql+- type for `nil`; did you mean to pass a non-nil value instead?\")))\n      \n        \n\n      \n          (condp instance? x\n      \n            Long    \"int\"\n      \n            Integer \"int\"\n      \n            String  \"string\"\n      \n            Boolean \"bool\"\n      \n            Double  \"float\"\n      \n            clojure.lang.BigInt \"int\"\n      \n            (throw (IllegalArgumentException.\n      \n                     (str \"Don't know graphql+- type of \" (pr-str x))))))\n      \n        \n\n      \n        (defn query*\n      \n          \"Runs a query given a graphql+- query string, and a map of variables for the\n      \n          query. Variables can be a map of strings, keywords, or symbols to strings,\n      \n          keywords, or symbols; they're all coerced to their string names, and prefixed\n      \n          with $.\n      \n        \n      \n              query(txn \\\"query all($a: string) { all(func: eq(name, $a)) { uid } }\\\"\n      \n                    {:a \\\"cat\\\"})\"\n      \n          ([^Transaction txn query-str]\n      \n           (json\/parse-string (.. txn (query query-str) getJson toStringUtf8)\n      \n                              true))\n      \n          ([^Transaction txn query vars]\n      \n           ;(info \"Query (vars:\" (pr-str vars) \"):\" query)\n      \n           (let [vars (-&gt;&gt; vars\n      \n                           (map (fn [[k v]] [(str \"$\" (name k)) (str v)]))\n      \n                           (into {}))\n      \n                 res (.queryWithVars txn query vars)]\n      \n             (json\/parse-string (.. res getJson toStringUtf8) true))))\n      \n        \n\n      \n        (defn query\n      \n          \"Like query*, but automatically generates the top-level `query` block with\n      \n          query variables inferred from the vars map. Example:\n      \n        \n      \n              query(txn,\n      \n                    \\\"{ all(func: eq(name, $a)) { uid } }\\\"\n      \n                    {:a \\\"cat\\\"})\"\n      \n          ([txn query-str]\n      \n           (t\/with-trace \"client.query\"\n      \n             (query* txn query-str)))\n      \n          ([txn query-str vars]\n      \n           (t\/with-trace \"client.query\"\n      \n             (query* txn\n      \n                     (str \"query all(\"\n      \n                          (-&gt;&gt; vars\n      \n                               (map (fn [[k v]] (str \"$\" (name k) \": \" (graphql-type v))))\n      \n                               (str\/join \", \"))\n      \n                          \") \" query-str)\n      \n                     vars))))\n      \n        \n\n      \n        (defn schema\n      \n          \"Retrieves the current schema as JSON\"\n      \n          [txn]\n      \n          (query txn \"schema {}\"))\n      \n        \n\n      \n        (defn await-ready\n      \n          \"Blocks until the server is up and responding to requests, or throws. Returns\n      \n          client.\"\n      \n          [client]\n      \n          (with-retry [attempts 16]\n      \n            (with-txn [t client]\n      \n              (schema t))\n      \n            (catch io.grpc.StatusRuntimeException e\n      \n              (cond (&lt;= attempts 1)\n      \n                    (throw e)\n      \n        \n\n      \n                    (and (.getCause e)\n      \n                         (instance? java.net.ConnectException\n      \n                                    (.getCause (.getCause e))))\n      \n                    (do (info \"GRPC interface unavailable, retrying in 5 seconds\")\n      \n                        (Thread\/sleep 5000)\n      \n                        (retry (dec attempts)))\n      \n        \n\n      \n                    (re-find #\"server is not ready to accept requests\"\n      \n                             (.getMessage e))\n      \n                    (do (info \"Server not ready, retrying in 5 seconds\")\n      \n                        (Thread\/sleep 5000)\n      \n                        (retry (dec attempts)))\n      \n        \n\n      \n                    :else\n      \n                    (throw e))))\n      \n          client)\n      \n        \n\n      \n        (defn upsert!\n      \n          \"Takes a transaction, a predicate, and a record map. If only one map is\n      \n          provided, it is used as the predicate. If no record exists for the given\n      \n          predicate, inserts the record map.\n      \n        \n      \n          Predicate can be a keyword, which is used as the primary key of the record.\n      \n          TODO: add more complex predicates.\n      \n        \n      \n          Returns nil if upsert did not take place. Returns mutation results otherwise.\"\n      \n          [t pred record]\n      \n          (t\/with-trace \"client.upsert!\"\n      \n            (if-let [pred-value (get record pred)]\n      \n              (let [res (-&gt; (query t (str \"{\\n\"\n      \n                                          \"  all(func: eq(\" (name pred) \", $a)) {\\n\"\n      \n                                          \"    uid\\n\"\n      \n                                          \"  }\\n\"\n      \n                                          \"}\")\n      \n                                   {:a pred-value}))]\n      \n                                                ;(info \"Query results:\" res)\n      \n                (when (empty? (:all res))\n      \n                                                ;(info \"Inserting...\")\n      \n                  (mutate! t record)))\n      \n        \n\n      \n              (throw (IllegalArgumentException.\n      \n                      (str \"Record \" (pr-str record) \" has no value for \"\n      \n                           (pr-str pred)))))))\n      \n        \n\n      \n        (defn gen-pred\n      \n          \"Generates a predicate for a key, given a count of keys, and a prefix.\"\n      \n          [prefix n k]\n      \n          (str prefix \"_\" (mod (hash k) n)))\n      \n        \n\n      \n        (defn gen-preds\n      \n          \"Given a key prefix and a number of keys, generates all predicate names that\n      \n          might be used.\"\n      \n          [prefix n]\n      \n          (-&gt;&gt; (range n)\n      \n               (map (fn [i] (str prefix \"_\" i)))))\n      \n        \n\n      \n        (defrecord TxnClient [opts conn]\n      \n          jc\/Client\n      \n          (open! [this test node]\n      \n            (assoc this :conn (open node)))\n      \n        \n\n      \n          (setup! [this test]\n      \n            (let [keys (-&gt;&gt; (gen-preds \"key\" (:key-predicate-count opts))\n      \n                            (map (fn [pred] (str pred \": int @index(int)\"\n      \n                                                 (when (:upsert-schema test) \" @upsert\")\n      \n                                                 \" .\\n\")))\n      \n                            str\/join)\n      \n                  vals (-&gt;&gt; (gen-preds \"val\" (:value-predicate-count opts))\n      \n                            (map (fn [pred] (str pred \": int .\\n\")))\n      \n                            str\/join)]\n      \n              (alter-schema! conn (str keys vals))))\n      \n        \n\n      \n          (invoke! [this test op]\n      \n            (with-conflict-as-fail op\n      \n              (with-txn [t conn]\n      \n                (-&gt;&gt; (:value op)\n      \n                     (reduce\n      \n                      (fn [txn' [f k v :as micro-op]]\n      \n                        (let [kp (gen-pred \"key\" (:key-predicate-count opts) k)\n      \n                              vp (gen-pred \"val\" (:value-predicate-count opts) k)]\n      \n                          (case f\n      \n                            :r\n      \n                            (let [res (query t (str \"{ q(func: eq(\" kp \", $key)) {\\n\"\n      \n                                                    \"  \" vp \"\\n\"\n      \n                                                    \"}}\")\n      \n                                             {:key k})\n      \n                                  reads (:q res)]\n      \n                              (conj txn' [f k (condp = (count reads)\n      \n                                                ; Not found\n      \n                                                0 nil\n      \n                                                ; Found\n      \n                                                1 (get (first reads)\n      \n                                                       (keyword vp))\n      \n                                                ; Ummm\n      \n                                                (throw (RuntimeException.\n      \n                                                        (str \"Unexpected multiple results for key \"\n      \n                                                             (pr-str k) \": \"\n      \n                                                             (pr-str reads)))))]))\n      \n        \n\n      \n                                                ; TODO: we should be able to optimize this to do pure\n      \n                                                ; inserts and UID-direct writes without the upsert\n      \n                                                ; read-write cycle, at least when we know the state\n      \n                            :w (do (if (:blind-insert-on-write? opts)\n      \n                                     (mutate! t {(keyword kp) k, (keyword vp) v})\n      \n                                     (upsert! t (keyword kp)\n      \n                                              {(keyword kp) k, (keyword vp) v}))\n      \n                                   (conj txn' micro-op)))))\n      \n                      [])\n      \n                     (assoc op :type :ok, :value)))))\n      \n        \n\n      \n          (teardown! [this test])\n      \n        \n\n      \n          (close! [this test]\n      \n            (close! conn)))\n      \n        \n\n      \n        (defn txn-client\n      \n          \"A client which can execute generic transcational workloads over arbitrary\n      \n          integer keys and values.\n      \n        \n      \n          Options:\n      \n        \n      \n            :blind-insert-on-write?   If true, don't do upserts; just insert on every\n      \n                                      write. Only appropriate when you'll never write\n      \n                                      the same thing twice.\n      \n            :value-predicate-count    How many predicates to stripe values over.\n      \n                                      Default: 5.\n      \n            :key-predicate-count      How many predicates to stripe keys over. Default:\n      \n                                      5.\"\n      \n          [opts]\n      \n          (TxnClient.\n      \n            (merge {:blind-insert-on-write? false\n      \n                    :value-predicate-count  5\n      \n                    :key-predicate-count    5}\n      \n                   opts)\n      \n            nil))","created_at":"2018-11-27T20:43:39+0000","updated_at":"2018-11-27T20:43:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/19334403?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["vuc","diaspora","audio","fediverse"],"is_public":false,"id":723,"title":"VUC727 \u2013 Pluspora and Diaspora","url":"https:\/\/www.vuc.me\/2018\/vuc727-pluspora-and-diaspora\/","content":"<div class=\"subtitle\">\n\t\t<h4 class=\"subtitle\">Event: Friday November 23rd, 2018 at 12 Noon EST, 17:00 UTC<\/h4>\n\n<h4 class=\"subtitle\"><\/h4><p>All <a href=\"http:\/\/packetsofeverything.libsyn.com\">audio versions are here<\/a><\/p>\n<br \/><iframe width=\"550\" height=\"320\" src=\"https:\/\/www.youtube.com\/embed\/3Qh2OKxWxPc\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>\n<br \/>\n   IP Communicationsl \n\t\t<\/div>\n\t\n<p>As soon as Google announced the sunset of their failed social network, Google+ denizens began looking at alternatives. There\u2019s plenty of time, the closure is scheduled for August 2019! Still, why invest any further time in a dying network? This week we talk about the many ways to communicate via open source platforms and software, especially the federated ones like Diaspora (and variants). An important step was taken with the creation of <a href=\"https:\/\/pluspora.com\">Pluspora<\/a>, a Diaspora pod. This has already attracted thousands of former (and current) Google+ denizens. We\u2019ll talk about how and why this was done with one or more of the creators.<\/p>\n<div id=\"GSKL2868197\" class=\"envelope\">About the Diaspora federation: <a href=\"https:\/\/the-federation.info\/diaspora\">https:\/\/the-federation.info\/diaspora<\/a><\/div>\n<div id=\"J8313M9B197\" class=\"envelope\">List of Diaspora pods: <a href=\"https:\/\/podupti.me\/\">https:\/\/podupti.me\/<\/a><\/div>\n<div>Pluspora pod <a href=\"https:\/\/pluspora.com\">https:\/\/pluspora.com<\/a><\/div>\n\t\t\t\t\t<p class=\"post_tags\">Tagged as:\n\t\t\t\t\t\t<a href=\"https:\/\/www.vuc.me\/tag\/diaspora\/\" rel=\"tag nofollow\">Diaspora<\/a>, \n\t\t\t\t\t\t<a href=\"https:\/\/www.vuc.me\/tag\/federation\/\" rel=\"tag nofollow\">federation<\/a>, \n\t\t\t\t\t\t<a href=\"https:\/\/www.vuc.me\/tag\/pluspora\/\" rel=\"tag nofollow\">Pluspora<\/a>, \n\t\t\t\t\t\t<a href=\"https:\/\/www.vuc.me\/tag\/social-network\/\" rel=\"tag nofollow\">social network<\/a>\n\t\t\t\t\t<\/p>","created_at":"2018-11-27T13:19:29+0000","updated_at":"2018-11-27T13:19:37+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"www.vuc.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":722,"title":"VUC727 \u2013 Pluspora and Diaspora","url":"https:\/\/www.vuc.me\/","content":"<div class=\"subtitle\">\n\t\t<h4 class=\"subtitle\">Event: Friday November 23rd, 2018 at 12 Noon EST, 17:00 UTC<\/h4>\n\n<h4 class=\"subtitle\"><\/h4><p>All <a href=\"http:\/\/packetsofeverything.libsyn.com\">audio versions are here<\/a><\/p>\n<br \/><iframe width=\"550\" height=\"320\" src=\"https:\/\/www.youtube.com\/embed\/3Qh2OKxWxPc\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>\n<br \/>\n   IP Communicationsl \n\t\t<\/div>\n\t\n<p>As soon as Google announced the sunset of their failed social network, Google+ denizens began looking at alternatives. There\u2019s plenty of time, the closure is scheduled for August 2019! Still, why invest any further time in a dying network? This week we talk about the many ways to communicate via open source platforms and software, especially the federated ones like Diaspora (and variants). An important step was taken with the creation of <a href=\"https:\/\/pluspora.com\">Pluspora<\/a>, a Diaspora pod. This has already attracted thousands of former (and current) Google+ denizens. We\u2019ll talk about how and why this was done with one or more of the creators.<\/p>\n<div id=\"GSKL2868197\" class=\"envelope\">About the Diaspora federation: <a href=\"https:\/\/the-federation.info\/diaspora\">https:\/\/the-federation.info\/diaspora<\/a><\/div>\n<div id=\"J8313M9B197\" class=\"envelope\">List of Diaspora pods: <a href=\"https:\/\/podupti.me\/\">https:\/\/podupti.me\/<\/a><\/div>\n<div>Pluspora pod <a href=\"https:\/\/pluspora.com\">https:\/\/pluspora.com<\/a><\/div>\n<p class=\"to_comments\">{ Comments on this entry are closed }<\/p>","created_at":"2018-11-27T13:19:20+0000","updated_at":"2018-11-27T13:19:20+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"www.vuc.me","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":721,"title":"The Greatest Sales Deck I\u2019ve Ever Seen","url":"https:\/\/medium.com\/the-mission\/the-greatest-sales-deck-ive-ever-seen-4f4ef3391ba0","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"6853\" class=\"graf graf--figure graf--layoutOutsetCenter graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*cWrckQhDDGxKiwpb6YtPtg.jpeg\" data-width=\"4195\" data-height=\"2907\" data-action=\"zoom\" data-action-value=\"1*cWrckQhDDGxKiwpb6YtPtg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*cWrckQhDDGxKiwpb6YtPtg.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Photo by <a href=\"https:\/\/unsplash.com\/@flenjoore\" data-href=\"https:\/\/unsplash.com\/@flenjoore\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Olu\u00a0Eletu<\/a><\/figcaption><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"4330\" class=\"graf graf--h3 graf-after--figure graf--title\">The Greatest Sales Deck I\u2019ve Ever\u00a0Seen<\/h1><h2 id=\"c33b\" class=\"graf graf--h4 graf-after--h3 graf--subtitle\">It\u2019s Zuora\u2019s, and it\u2019s brilliant. Here\u2019s\u00a0why.<\/h2><p id=\"1120\" class=\"graf graf--p graf-after--h4\">A few months ago, my friend Tim took a new sales job at a Series C tech company that had raised over $60 million from A-list investors. He\u2019s one of the best salespeople I know, but soon after starting, he emailed me to say he was struggling.<\/p><p id=\"c9c2\" class=\"graf graf--p graf--startsWithDoubleQuote graf-after--p\">\u201cI\u2019ve landed a few small accounts,\u201d Tim said. \u201cBut my pitch falls flat at big enterprises.\u201d<\/p><p id=\"2614\" class=\"graf graf--p graf-after--p\">As I\u2019ve written before, I love helping teams craft the high-level strategic story that powers sales, marketing, fundraising\u200a\u2014\u200aeverything. So Tim and I met for lunch at the Amber India restaurant off San Francisco\u2019s Market Street to review his deck.<\/p><p id=\"a251\" class=\"graf graf--p graf-after--p\">After loading up on the all-you-can-eat buffet, I asked Tim, \u201cAt what point do prospects tune out?\u201d<\/p><p id=\"ad40\" class=\"graf graf--p graf--startsWithDoubleQuote graf-after--p\">\u201cUsually a few slides in,\u201d he said.<\/p><p id=\"2f22\" class=\"graf graf--p graf-after--p\">Intent on maximizing dining ROI, Tim went back to the buffet for seconds. When he returned, I pulled out my laptop and launched into a Powerpoint presentation.<\/p><p id=\"8289\" class=\"graf graf--p graf--startsWithDoubleQuote graf-after--p\">\u201cWhat\u2019s this?\u201d Tim asked.<\/p><p id=\"6778\" class=\"graf graf--p graf--startsWithDoubleQuote graf-after--p\">\u201cThis,\u201d I said, \u201cis the greatest sales deck I have ever seen.\u201d<\/p><h3 id=\"cf30\" class=\"graf graf--h3 graf-after--p\">The 5 Elements of a Brilliant Sales Narrative<\/h3><p id=\"7da0\" class=\"graf graf--p graf-after--h3\">The sales deck I showed Tim came from <a href=\"https:\/\/medium.com\/@subscribed\" data-href=\"https:\/\/medium.com\/@subscribed\" data-anchor-type=\"2\" data-user-id=\"b5ea6a97d651\" data-action-value=\"b5ea6a97d651\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Zuora<\/a>, the IPO-bound Silicon Valley company that sells a SaaS platform for subscription billing. If you pay for anything on a recurring basis (e.g. enterprise software), there\u2019s a good chance that Zuora facilitates those transactions.<\/p><p id=\"5a3e\" class=\"graf graf--p graf-after--p\">I had received the deck from an ex-Zuora salesperson, who said it helped him close the biggest deals of his career. (I have no connection to Zuora, and no relationship with anyone who currently works there. <em class=\"markup--em markup--p-em\">UPDATE: Some current Zuora employees have connected with me after reading this.<\/em>)<\/p><p id=\"ac0a\" class=\"graf graf--p graf-after--p\">Abandoning his naan in a puddle of curried goat, Tim grabbed pen and paper and took notes as we ran through what made the Zuora deck so effective.<\/p><p id=\"ab62\" class=\"graf graf--p graf-after--p\">Specifically, we noted how brilliantly the deck led prospects through the following five elements, in precisely this order:<\/p><p id=\"5fb0\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">(The ex-Zuora salesperson asked that I not share the Zuora deck publicly, and I will honor that request. However, I found slides on Zuora\u2019s website and SlideShare channel that exhibit nearly the same narrative flow; all of the images below come from those public sources.)<\/em><\/p><h3 id=\"300f\" class=\"graf graf--h3 graf-after--p\">#1. Name a Big, Relevant Change in the\u00a0World<\/h3><p id=\"c055\" class=\"graf graf--p graf-after--h3\">Don\u2019t kick off a sales presentation by talking about your product, your headquarters locations, your investors, your clients, or anything about yourself.<\/p><p id=\"192f\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">Instead, name the undeniable shift in the world that creates both (a) big stakes and (b) huge urgency for your prospect.<\/em><\/p><p id=\"cc4e\" class=\"graf graf--p graf-after--p\">The first slide of virtually every Zuora deck\u200a\u2014\u200asales or otherwise\u200a\u2014\u200ais some version of this:<\/p><figure id=\"c186\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*4EhO9r4Hp5JmDuREvja-uw.png\" data-width=\"1722\" data-height=\"990\" data-action=\"zoom\" data-action-value=\"1*4EhO9r4Hp5JmDuREvja-uw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*4EhO9r4Hp5JmDuREvja-uw.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"c747\" class=\"graf graf--p graf-after--figure\">Zuora came up with the phrase \u201csubscription economy\u201d to name the trend in which buyers increasingly choose recurring service payments over outright purchases. Zuora usually follows that with a slide laying out the history of the change:<\/p><figure id=\"edfd\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*-k3idPAh0ATPmVWBviYWxA.png\" data-width=\"1904\" data-height=\"1072\" data-action=\"zoom\" data-action-value=\"1*-k3idPAh0ATPmVWBviYWxA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*-k3idPAh0ATPmVWBviYWxA.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"5681\" class=\"graf graf--p graf-after--figure\"><em class=\"markup--em markup--p-em\">Note the subtle but important difference from what most pitch advice tells you, which is to start with \u201cthe problem.\u201d <\/em>When you assert that your prospects have a problem, you put them on the defensive. They may be unaware of the problem, or uncomfortable admitting they suffer from it.<\/p><p id=\"a625\" class=\"graf graf--p graf-after--p\">But when you highlight a shift in the world, you get prospects to open up about how that shift affects them, how it scares them, and where they see opportunities. Most importantly, you grab their attention. As Hollywood screenwriting guru Robert McKee says:<\/p><blockquote id=\"283b\" class=\"graf graf--pullquote graf-after--p\"><div>\u2026what attracts human attention is change.\u00a0\u2026if the temperature around you changes, if the phone rings\u200a\u2014\u200athat gets your attention. The way in which a story begins is a starting event that creates a moment of\u00a0change.<\/div><\/blockquote><h3 id=\"30b6\" class=\"graf graf--h3 graf-after--pullquote\">#2. Show There\u2019ll Be Winners and\u00a0Losers<\/h3><p id=\"16db\" class=\"graf graf--p graf-after--h3\">All prospects suffer from what economists call \u201closs aversion.\u201d That is, they tend to avoid a possible loss by sticking to the status quo, rather than risk a possible gain by opting for change.<\/p><p id=\"9c55\" class=\"graf graf--p graf-after--p\">To combat loss aversion, you must demonstrate how the change you cited above will create big winners <em class=\"markup--em markup--p-em\">and<\/em> big losers. In other words, you have to show <em class=\"markup--em markup--p-em\">both<\/em> of the following:<\/p><ol class=\"postList\"><li id=\"40fb\" class=\"graf graf--li graf-after--p\">That adapting to the change you cited will likely result in a highly positive future for the prospect; and<\/li><li id=\"cbff\" class=\"graf graf--li graf-after--li\">That not doing so will likely result in an <em class=\"markup--em markup--li-em\">unacceptably<\/em> negative future for the prospect<\/li><\/ol><p id=\"825e\" class=\"graf graf--p graf-after--li\">Zuora neatly accomplishes this by documenting a \u201cmass extinction\u201d among Fortune 500 companies\u2026<\/p><figure id=\"0b33\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*Ivu9eFe4M3xoZV-hQNEUIw.jpeg\" data-width=\"720\" data-height=\"405\" data-action=\"zoom\" data-action-value=\"1*Ivu9eFe4M3xoZV-hQNEUIw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Ivu9eFe4M3xoZV-hQNEUIw.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"4578\" class=\"graf graf--p graf-after--figure\">\u2026and then showing how the \u201cwinners\u201d have shifted from product ownership to subscription services. Those include upstarts\u2026<\/p><figure id=\"8792\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*0NoObiX6A4mbCCEofUpjZQ.jpeg\" data-width=\"720\" data-height=\"405\" data-action=\"zoom\" data-action-value=\"1*0NoObiX6A4mbCCEofUpjZQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*0NoObiX6A4mbCCEofUpjZQ.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"c42d\" class=\"graf graf--p graf-after--figure\">\u2026as well as rejuvenated incumbents:<\/p><figure id=\"cce2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*9U0xh4v_3df5g-H8AYlOZg.jpeg\" data-width=\"720\" data-height=\"405\" data-action=\"zoom\" data-action-value=\"1*9U0xh4v_3df5g-H8AYlOZg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*9U0xh4v_3df5g-H8AYlOZg.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"5e37\" class=\"graf graf--p graf-after--figure\">To bring the point home, Zuora asks the following:<\/p><figure id=\"a079\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*FnmdADPmftv--yW4VTtkPA.jpeg\" data-width=\"720\" data-height=\"405\" data-action=\"zoom\" data-action-value=\"1*FnmdADPmftv--yW4VTtkPA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*FnmdADPmftv--yW4VTtkPA.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"e980\" class=\"graf graf--p graf-after--figure\">Of course, by this point the common thread is already well established in prospects\u2019 minds: Winners adopt the subscription service models that Zuora supports.<\/p><h3 id=\"1d8a\" class=\"graf graf--h3 graf-after--p\">#3. Tease the Promised\u00a0Land<\/h3><p id=\"d479\" class=\"graf graf--p graf-after--h3\">It\u2019s tempting at this point to jump into the details of your product or service. Resist that urge.<\/p><p id=\"d614\" class=\"graf graf--p graf-after--p\">If you introduce product\/service details too soon, prospects won\u2019t yet have enough context for why those details are important, and they\u2019ll tune out.<\/p><p id=\"7635\" class=\"graf graf--p graf-after--p\">Instead, first present a \u201cteaser\u201d vision of the happily-ever-after that your product\/service will help the prospect achieve\u2014what I call the <em class=\"markup--em markup--p-em\">Promised Land<\/em>.<\/p><p id=\"3f72\" class=\"graf graf--p graf-after--p\">Your Promised Land should be both desirable (obviously) and difficult for the prospect to achieve without outside help. Otherwise, why does your company exist?<\/p><p id=\"ed1e\" class=\"graf graf--p graf-after--p\">After demonstrating that the subscription economy will result in winners and losers, Zuora presents this Promised Land slide, which offers concrete criteria for what it means to win in the subscription economy:<\/p><figure id=\"1f1d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*dmRFjO2Hkg3YpLA_soxauQ.png\" data-width=\"1900\" data-height=\"1064\" data-action=\"zoom\" data-action-value=\"1*dmRFjO2Hkg3YpLA_soxauQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*dmRFjO2Hkg3YpLA_soxauQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"c344\" class=\"graf graf--p graf-after--figure\">Note that the Promised Land is a new future state, not your product or service.<\/p><p id=\"1ee8\" class=\"graf graf--p graf-after--p\">(Over lunch, I asked my friend Tim to articulate his Promised Land, and he said, \u201cYou\u2019ll have the most innovative platform for ____.\u201d Nope: the Promised Land is not <em class=\"markup--em markup--p-em\">having<\/em> your technology, but what life is like <em class=\"markup--em markup--p-em\">thanks<\/em> to having your technology.)<\/p><p id=\"78e0\" class=\"graf graf--p graf-after--p\">Your Promised Land is also crucial for helping prospects pitch your solution to colleagues after your sales meeting ends. In your absence, those colleagues will ask, \u201cWhat do those guys do again?\u201d Armed with a compelling Promised Land, your prospects are more likely to supply an answer that gets others on board.<\/p><h3 id=\"4318\" class=\"graf graf--h3 graf-after--p\">#4. Introduce Features as \u201cMagic Gifts\u201d for Overcoming Obstacles to the Promised\u00a0Land<\/h3><p id=\"d190\" class=\"graf graf--p graf-after--h3\">If it\u2019s not clear by now, successful sales decks follow the same narrative structure as epic films and fairy tales. Your prospect is Luke, and you\u2019re Obi Wan, furnishing a lightsaber to help him defeat the Empire. Your prospect is Frodo, and you\u2019re Gandalf, wielding wizardry to help him destroy the ring. Your prospect is Cinderella, and you\u2019re the fairy godmother, casting spells to get her to the ball.<\/p><p id=\"0f24\" class=\"graf graf--p graf-after--p\">When you introduce your product or service, do so by positioning its capabilities like the lightsaber, wizardry and spells\u2014as \u201cmagic gifts\u201d for helping your main character (prospect) reach that much-desired Promised Land.<\/p><figure id=\"014d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*mKJi6s4U0wCKVwvyxvUbIQ.jpeg\" data-width=\"720\" data-height=\"405\" data-action=\"zoom\" data-action-value=\"1*mKJi6s4U0wCKVwvyxvUbIQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*mKJi6s4U0wCKVwvyxvUbIQ.jpeg\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"3a8c\" class=\"graf graf--p graf-after--figure\">For example, above is the slide where Zuora talks about the structure of its customer record. Out of context, this detail would likely bore even the most technical prospect.<\/p><p id=\"972f\" class=\"graf graf--p graf-after--p\">Positioned in the context of transitioning from an \u201cold world\u201d to a \u201cnew world,\u201d however, it\u2019s the foundation for an engaging conversation with prospects\u2014technical and otherwise\u2014about why it\u2019s so hard to reach the Promised Land with traditional solutions.<\/p><h3 id=\"46d5\" class=\"graf graf--h3 graf-after--p\">#5. Present Evidence that You Can Make the Story Come\u00a0True<\/h3><p id=\"ecf3\" class=\"graf graf--p graf-after--h3\">In telling the sales narrative this way, you\u2019re making a commitment to prospects: If they go with you, you\u2019ll get them to the Promised Land.<\/p><p id=\"6589\" class=\"graf graf--p graf-after--p\">But the road to the Promised Land is, by definition, littered with obstacles, so prospects are rightly skeptical of your ability to deliver. The last piece of the pitch, then, is the best evidence you can offer that you can make the story you\u2019re telling come true.<\/p><p id=\"3a82\" class=\"graf graf--p graf-after--p\">By far, the most effective type of evidence is a success story about how you\u2019ve already helped someone else (who is similar to the prospect) reach the Promised Land. Zuora has a set of customer success stories that sales reps draw on, and while they\u2019re more elaborate in the actual deck, this testimonial captures the essence:<\/p><figure id=\"a0b4\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*Ke6faHWcc4l8sZ7s9sjwug.png\" data-width=\"2002\" data-height=\"910\" data-action=\"zoom\" data-action-value=\"1*Ke6faHWcc4l8sZ7s9sjwug.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Ke6faHWcc4l8sZ7s9sjwug.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"426e\" class=\"graf graf--p graf-after--figure\">I also like this one, from an exec at NCR (a Zuora customer), which speaks more explicitly to Zuora\u2019s stated Promised Land:<\/p><figure id=\"a651\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*wPiGUtu6kyvezEPN_NEwVQ.png\" data-width=\"1990\" data-height=\"910\" data-action=\"zoom\" data-action-value=\"1*wPiGUtu6kyvezEPN_NEwVQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*wPiGUtu6kyvezEPN_NEwVQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"7e99\" class=\"graf graf--p graf-after--figure\">What if you don\u2019t yet have a huge number of successful customers? Product demos are the next most effective evidence, but again, features should always be presented in the context of how they help a prospect reach the Promised Land.<\/p><h3 id=\"0963\" class=\"graf graf--h3 graf-after--p\">A Sales Narrative Works Best When Everyone Tells\u00a0It<\/h3><p id=\"0bd7\" class=\"graf graf--p graf-after--h3\">Of course, successful sales rarely happen <em class=\"markup--em markup--p-em\">solely<\/em> as the result of a great deck. In order for salespeople to be successful, the entire organization must align around the narrative about change, Promised Land, and Magic Gifts.<\/p><p id=\"4b00\" class=\"graf graf--p graf-after--p\">There\u2019s no better example of that than Zuora. If you ever see a Zuora executive speak\u2014including CEO <a href=\"https:\/\/medium.com\/@tientzuo\" data-href=\"https:\/\/medium.com\/@tientzuo\" data-anchor-type=\"2\" data-user-id=\"3d26841cd79c\" data-action-value=\"3d26841cd79c\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Tien Tzuo<\/a>\u2014you\u2019ll almost certainly hear about the subscription economy and the winners and losers it\u2019s creating. In fact, that\u2019s the theme of virtually all the company\u2019s marketing communications and campaigns, as well as its public vision statement:<\/p><figure id=\"708f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*8OfsnIAn34oRWrlBbgHeDg.png\" data-width=\"2102\" data-height=\"988\" data-action=\"zoom\" data-action-value=\"1*8OfsnIAn34oRWrlBbgHeDg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*8OfsnIAn34oRWrlBbgHeDg.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"8ca2\" class=\"graf graf--p graf-after--figure\">According to the ex-Zuora salesperson, this company-wide alignment around the story made him wildly successful:<\/p><blockquote id=\"9532\" class=\"graf graf--pullquote graf-after--p\"><div>The Zuora marketing folks ran campaigns and branding around this shift to the subscription economy, and [CEO] Tien [Tzuo] talked it up all the time. All of that was like air cover for my in-person sales ground attack. By the time I arrived, prospects were already convinced they had to act. It was the closest thing I\u2019ve ever experienced to sales\u00a0nirvana.<\/div><\/blockquote><h3 id=\"24f7\" class=\"graf graf--h3 graf-after--pullquote\">Biggest Deal\u00a0Ever<\/h3><p id=\"1aa9\" class=\"graf graf--p graf-after--h3\">Just three weeks after our lunch, Tim called to say he was seeing promising changes in how prospects at large enterprises were reacting to his new deck, which we drafted together based on Zuora\u2019s template. For one thing, prospects opened up much faster about the challenges they were facing. Also, he said the new pitch was more effective at engaging CFOs and other senior gatekeepers.<\/p><p id=\"f8b3\" class=\"graf graf--p graf-after--p\">Then, a week after that, Tim emailed with even better news: <strong class=\"markup--strong markup--p-strong\">He had just signed the largest deal in his company\u2019s history.<\/strong><\/p><p id=\"f7cc\" class=\"graf graf--p graf-after--p graf--trailing\">Next week, we\u2019re headed back to Amber India to celebrate.<\/p><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"4471\" class=\"graf graf--p graf--leading graf--trailing\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">About Andy Raskin:<\/em><\/strong><em class=\"markup--em markup--p-em\"><br \/>I help leaders align around a strategic story\u200a\u2014\u200ato power sales, marketing, fundraising, product, and recruiting. My clients include teams backed by Andreessen Horowitz, First Round Capital, GV, and other top venture firms. I\u2019ve also led strategic story training at Uber, Yelp, General Assembly, HourlyNerd, Neustar, and Stanford. To learn more or get in touch, visit <\/em><a href=\"http:\/\/andyraskin.com\" data-href=\"http:\/\/andyraskin.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">http:\/\/andyraskin.com<\/em><\/a><em class=\"markup--em markup--p-em\">.<\/em><\/p><\/div><\/div><\/section>","created_at":"2018-11-26T19:26:28+0000","updated_at":"2018-11-26T19:26:28+0000","published_at":"2016-09-15T16:20:59+0000","published_by":["Andy Raskin"],"annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*cWrckQhDDGxKiwpb6YtPtg.jpeg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["clojure","spec"],"is_public":false,"id":720,"title":"Dynamic Languages and Types","url":"http:\/\/danboykis.com\/posts\/dynamic-languages-and-types\/","content":"<p>Many proponents of static typing, including many of the commenters who responded to <a href=\"http:\/\/yogthos.net\/posts\/2016-09-06-Static-Typing-WebSockets.html\">this post<\/a>,\nstrongly believe that its presence in a language directly translates into correctness, even though there is very little\nempirical evidence to support this assertion. More importantly, the assumption is that the safety guarantees provided\nby static typing can\u2019t be provided via any other mechanism.<\/p><p>To me, and many others who work with statically-typed languages, the biggest benefit of types is auto documentation.<\/p>\n<p>Here\u2019s an example. Let\u2019s suppose the developer is encountering this code, in a generic statically-typed language,\nfor the first time.<\/p>\n<pre class=\"language-java\">Account generateAccount(Event e)\n<\/pre>\n<p>It\u2019s not clear what what <code>generateAccount<\/code> actually does, but we know it takes an <code>Event<\/code> and returns an <code>Account<\/code>.\nIn an IDE, it\u2019s easy to click onto Event or Account and see what values they describe to get a sense of what\u2019s happening.\nIf <code>Account<\/code> and <code>Event<\/code> are immutable, we have a pretty good idea of what to expect. The second guarantee is that\n<code>generateAccount<\/code> can\u2019t take <code>Potato<\/code> and return <code>Tomato<\/code> where <code>Potato<\/code> and <code>Tomato<\/code> have no relation to <code>Event<\/code> and\n<code>Account<\/code>. That\u2019s simply not allowed. We pass in <code>Event<\/code> and get back an <code>Account<\/code>, and that\u2019s it. Sure it might throw\nan exception, but the approximate understanding of the function exists, and we can move on to other code.<\/p>\n<p>In a dynamic language, the same function might look something like this:<\/p>\n<pre class=\"language-python\">def generateAccount(e)\n<\/pre>\n<p>It\u2019s not at all obvious what \u2018<code>e<\/code>\u2019 is and what <code>generateAccount<\/code> returns, and we\u2019ll need to get into the internals of\nthe function to figure it out. That\u2019s a lot more effort than simply reading the types by sight in the previous example.\nAnd what if there are 30 of these functions? And if we do finally figure out what it is <code>generateAccount<\/code> does and\ncomment above <code>generateAccount<\/code> describing the types informally, we have to hope that no one changes the comments.\nAll of this function \u201cmetadata\u201d is maintained by the compiler in the statically defined <code>generateAccount<\/code>.<\/p>\n<p>Every developer that I\u2019ve talked to who favors static typing agrees this is the crux of the issue. It\u2019s not monads, or\nlinear types or even type inference that people want. What\u2019s essential is being able to look at a function and figure out\nthe inputs and outputs to get an approximate understanding without having to dig into the internals.<\/p>\n<p>I believe this situation causes some people to prefer statically-typed languages, because the dynamic languages they\nare used to are not powerful enough, and don\u2019t have maintainers who were sufficiently thoughtful to give this feature\nout of the box. My two favorite languages <a href=\"http:\/\/clojure.org\/about\/spec\">Clojure<\/a> and <a href=\"http:\/\/docs.racket-lang.org\/guide\/contract-boundaries.html\">Racket<\/a>,\nboth solve this problem in a powerful and compelling way. I am going to show off a little bit of Clojure\u2019s way of\naddressing this problem. Luckily for me someone thought of a <a href=\"https:\/\/gist.github.com\/honza\/5897359\">decent example<\/a>\nalready. Here\u2019s the challenge recreated here:<\/p>\n<p>The <a href=\"https:\/\/github.com\/honza\/bibles\/blob\/master\/ESV\/ESV.json\">JSON<\/a> data is in the following format<\/p>\n<pre class=\"language-javascript\">{\n    \"Genesis\": {\n        \"1\": {\n            \"1\": \"In the beginning...\" ,\n            \"2\": \"...\"\n        },\n        \"2\": { ... }\n    },\n    \"Exodus\": { ... },\n    ...\n}\n<\/pre>\n<p>In JSON, keys aren\u2019t ordered and must be strings.<\/p>\n<p>The goal is to parse the JSON file, order everything correctly (which means parsing the keys into integers), and\nproduce an array that looks like this:<\/p>\n<pre class=\"language-python\">[\n    (\"Genesis\", 1, 1, \"In the beginning..\")\n    (\"Genesis\", 1, 2, \"\")\n]\n<\/pre>\n<p>The data structure can be a tuple or some new type.<\/p>\n<p>Here\u2019s the Clojure code to do the challenge:<\/p>\n<pre class=\"language-clojure\">(def book-order [\"Genesis\", \"Exodus\", \"Leviticus\", \"Numbers\", \"Deuteronomy\",\n    \"Joshua\", \"Judges\", \"Ruth\", \"1 Samuel\", \"2 Samuel\", \"1 Kings\", \"2 Kings\",\n    \"1 Chronicles\", \"2 Chronicles\", \"Ezra\", \"Nehemiah\", \"Esther\", \"Job\",\n    \"Psalms\", \"Proverbs\", \"Ecclesiastes\", \"Song of Solomon\", \"Isaiah\",\n    \"Jeremiah\", \"Lamentations\", \"Ezekiel\", \"Daniel\", \"Hosea\", \"Joel\", \"Amos\",\n    \"Obadiah\", \"Jonah\", \"Micah\", \"Nahum\", \"Habakkuk\", \"Zephaniah\", \"Haggai\",\n    \"Zechariah\", \"Malachi\", \"Matthew\", \"Mark\", \"Luke\", \"John\", \"Acts\",\n    \"Romans\", \"1 Corinthians\", \"2 Corinthians\", \"Galatians\", \"Ephesians\",\n    \"Philippians\", \"Colossians\", \"1 Thessalonians\", \"2 Thessalonians\",\n    \"1 Timothy\", \"2 Timothy\", \"Titus\", \"Philemon\", \"Hebrews\", \"James\",\n    \"1 Peter\", \"2 Peter\", \"1 John\", \"2 John\", \"3 John\", \"Jude\", \"Revelation\"])\n \n(def numeric-str (s\/and string? st\/numeric? (s\/conformer #(Integer\/parseInt %)) pos-int?))\n(s\/def ::bible   (s\/map-of ::book ::chapter))\n(s\/def ::book    (set book-order))\n(s\/def ::chapter (s\/map-of numeric-str ::verse :conform-keys true))\n(s\/def ::verse   (s\/map-of numeric-str (s\/and string? not-empty) :conform-keys true))\n(s\/def ::sorted-bible (s\/coll-of (s\/tuple ::book pos-int? pos-int? string?)))\n \n(defn get-books [bible]\n  (map (fn [book] [book, (bible book)]) book-order))\n \n(defn get-chapters [books]\n  (mapcat (fn [[book chapter]]\n            (let [i (into (sorted-map) chapter)]\n              (map (fn [[n c]] [book, n, c]) i)))\n          books))\n \n(defn get-verses [verses]\n  (mapcat (fn [[b c d]]\n            (let [i (into (sorted-map) d)]\n              (map (fn [[v t]] [b, c, v, t]) i))) verses))\n \n(defn parse-bible []\n  (let [parse (comp #(s\/conform ::sorted-bible %) get-verses get-chapters get-books)\n        bible (-&gt;&gt; \".\/resources\/ESV.json\" slurp parse-string (s\/conform ::bible))]\n    (if (= ::s\/invalid bible)\n      (throw (ex-info \"not valid bible data\" (s\/explain-data bible)))\n      (parse bible))))\n<\/pre>\n<p>This code is pretty straightforward, but I do want to call attention to the <code>s\/def<\/code> spec section. A <code>::bible<\/code> is a map\nwhose keys must exist in <code>book-order<\/code>. The value of <code>bible<\/code> is chapter which is a map whose keys are numeric strings.\nThe values of <code>chapter<\/code> are maps named verse whose keys are again numeric strings and values are nonempty strings.\nAll of this done declaratively. Spec not only validates data it can also transform it. In this case, we can <code>conform<\/code>\nthe keys to the spec which means that once we validate the bible map, the numeric keys will be converted into integers.<\/p>\n<p>The upside of spec is that now we have robust schema validation. The program will abort early if the data passed to it\nviolates spec, so we don\u2019t have to do error handling in the program\u2019s functions, because the data is in good shape. We\ncan just concentrate on transforming it. An even bigger upside is that now it\u2019s easy to tell what kind of data goes in,\njust follow the definition of <code>::bible<\/code>. What data comes out? Follow <code>::sorted-bible<\/code>.<\/p>\n<p>Best of all, the errors are data! Here\u2019s a corrupt piece of data that violates the spec rule being tested against the\nspec. The name \u201cGe\u201d is not in the set of valid books. Since the error is data, we can interrogate it.<\/p>\n<pre class=\"language-clojure\">=&gt; (s\/explain-data ::bible {\"Ge\" {\"1\" {\"1\" \"In the...\"}}})\n \n#:clojure.spec{:problems ({:path [0], :pred (set book-order), :val \"Ge\", :via [:bible.core\/bible :bible.core\/book], :in [\"Ge\" 0]})}\n \n=&gt; (-&gt; (s\/explain-data ::bible {\"Ge\" {\"1\" {\"1\" \"In the...\"}}}) ::s\/problems first :pred)\n \n(set book-order)\n<\/pre>\n<p>Here are some more cases of bad data.<\/p>\n<pre class=\"language-clojure\">=&gt; (s\/explain-data ::bible {\"Genesis\" {\"1\" {\"1\" \"In the...\"}}}) ;Good data\n \nnil\n \n=&gt; (s\/explain-data ::bible {\"Genesis\" {\"A\" {\"1\" \"In the...\"}}}) ;A is not a number\n \n#:clojure.spec{:problems ({:path [1 0], :pred numeric?, :val \"A\", :via [:bible.core\/bible :bible.core\/chapter], :in [\"Genesis\" 1 \"A\" 0]})}\n \n=&gt; (s\/explain-data ::bible {\"Genesis\" {\"1\" {\"1\" 42}}}) ;42 is not a string\n \n#:clojure.spec{:problems ({:path [1 1 1], :pred string?, :val 42, :via [:bible.core\/bible :bible.core\/chapter :bible.core\/verse], :in [\"Genesis\" 1 \"1\" 1 \"1\" 1]})}\n<\/pre>\n<p>With just a few lines of code, we get all of the following:<\/p>\n<p>1) The ability to verify that data is in the shape and form the program expects cuts down on error checking<\/p>\n<p>2) Some free data parsing, in this case the string to int conversion<\/p>\n<p>3) A robust system for error messages that can be customized, because it\u2019s just data<\/p>\n<p>4) Spec allows powerful predicates like <code>(set book-order)<\/code> and <code>numeric-str<\/code> which are not easy to do in a conventional\ntyped language. Often types provide limited guarantees like \u201cthis key is a String\u201d and not the fact that it\u2019s numeric or\nnonempty. These are exactly the type of invariants we want but rarely get out of types.<\/p>\n<p>5) Specs can be composed and reused in arbitrary ways<\/p>\n<p>6) Using spec to generate test data or just get an example output of what kind of data conforms to the spec.<\/p>\n<p>For example:<\/p>\n<pre class=\"language-clojure\">=&gt; (pprint (ffirst (s\/exercise ::sorted-bible 1)))\n \n[[\"2 Kings\" 1 1 \"Z\"]\n [\"Esther\" 2 1 \"1\"]\n [\"1 Samuel\" 1 1 \"7Tk9\"]\n [\"Haggai\" 2 2 \"8\"]]\n<\/pre>\n<p>7) Since specs live in a registry, we can query the registry for specs that are defined in our program programmatically\nin the REPL. This is another aspect that\u2019s desirable in a large code base, because it enables reuse.<\/p>\n<p>8) We don\u2019t have know how the specs look like beginning development. It\u2019s possible to incrementally build up a program\nin the REPL and once we\u2019re happy with the code, to define the specs. Even this is completely optional. Personally, I\nfind that I like writing specs upfront to conform the input to my code and once I figure out the flow of my code to\nwrite specs again afterwards.<\/p>\n<p>9) This is a subtle point, but it\u2019s important to realize that spec supports the creation of open systems that are\namenable to change and end up less brittle. The following spec<\/p>\n<pre class=\"language-clojure\">(s\/keys :req [::a ::b])\n<\/pre>\n<p>validates that a map contains two keys <code>::a<\/code> and <code>::b<\/code> if there are other keys in the map that don\u2019t have specs defined\nthey are left alone. If there\u2019s no spec for <code>::a<\/code>, the only thing validated is that a key <code>::a<\/code> is present. When\nrequirements change, spec allows for a smooth progression to new requirements. We don\u2019t have to make all of our\ndecisions upfront and our code can still operate.<\/p>\n<p>There are many more uses for specs. I recommend following the excellent tutorial <a href=\"http:\/\/clojure.org\/guides\/spec\">here<\/a>\nto learn more.<\/p>\n<p>I don\u2019t believe that static typing is uniquely qualified to give programmers safety guarantees or better readability.\nThe github link for the challenge provides a Haskell solution along side a Clojure one. Since Haskell is treated as the\ncurrent poster child for statically typed languages, it would be great for somebody to provide a comparable Haskell\nsolution that gives similar guarantees.<\/p>","created_at":"2018-11-26T14:33:50+0000","updated_at":"2018-11-26T14:33:51+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"danboykis.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["podcast","changelog","ai","hardware","intel"],"is_public":false,"id":719,"title":"Practical AI #21: UBER and Intel\u2019s Machine Learning platforms with Cormac Brick and Mike Del Balso","url":"https:\/\/changelog.com\/practicalai\/21","content":"<section class=\"episode-body-section\"><div class=\"richtext richtext--small\"><ul><li>Cormac Brick \u2013 <a href=\"https:\/\/twitter.com\/cormacb\">Twitter<\/a>, <a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\">Website<\/a><\/li>\n              <li>Mike Del Balso \u2013 <a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\">Website<\/a><\/li>\n              <li>Chris Benson \u2013 <a href=\"https:\/\/twitter.com\/chrisbenson\">Twitter<\/a>, <a href=\"https:\/\/github.com\/chrisbenson\">GitHub<\/a>, <a href=\"https:\/\/chrisbenson.com\">Website<\/a><\/li>\n              <li>Daniel Whitenack \u2013 <a href=\"https:\/\/twitter.com\/dwhitena\">Twitter<\/a>, <a href=\"https:\/\/github.com\/dwhitena\">GitHub<\/a>, <a href=\"https:\/\/www.datadan.io\/\">Website<\/a><\/li>\n            <\/ul><\/div><\/section><section class=\"episode-body-section\"><div class=\"richtext richtext--small\"><p>\n                <strong><a href=\"https:\/\/do.co\/changelog\">DigitalOcean<\/a><\/strong> \u2013 DigitalOcean is simplicity at scale. Whether your business is running one virtual machine or ten thousand, DigitalOcean gets out of your way so your team can build, deploy, and scale faster and more efficiently. New accounts get $100 in credit to use in your first 60 days.\n              <\/p><p>\n                <strong><a href=\"https:\/\/www.fastly.com\/?utm_source=changelog&amp;utm_medium=podcast&amp;utm_campaign=changelog-sponsorship\">Fastly<\/a><\/strong> \u2013 <strong>Our bandwidth partner.<\/strong> Fastly powers fast, secure, and scalable digital experiences. Move beyond your content delivery network to their powerful edge cloud platform. Learn more at <a href=\"https:\/\/www.fastly.com\/?utm_source=changelog&amp;utm_medium=podcast&amp;utm_campaign=changelog-sponsorship\">fastly.com<\/a>.\n              <\/p><p>\n                <strong><a href=\"https:\/\/rollbar.com\/changelog\">Rollbar<\/a><\/strong> \u2013 <strong>We catch our errors before our users do because of Rollbar.<\/strong> Resolve errors in minutes, and deploy your code with confidence. Learn more at <a href=\"https:\/\/rollbar.com\/changelog\">rollbar.com\/changelog<\/a>.\n              <\/p><p>\n                <strong><a href=\"https:\/\/linode.com\/changelog\">Linode<\/a><\/strong> \u2013 <strong>Our cloud server of choice.<\/strong> Deploy a fast, efficient, native SSD cloud server for only $5\/month. Get 4 months free using the code <code>changelog2018<\/code>. Start your server - head to <a href=\"https:\/\/linode.com\/changelog\">linode.com\/changelog<\/a>\n              <\/p><\/div><\/section><section class=\"episode-body-section\"><div class=\"richtext richtext--small\"><ul><li><a href=\"https:\/\/www.movidius.com\/\">Intel\u2019s Movidius Group<\/a><\/li>\n<li><a href=\"https:\/\/software.intel.com\/en-us\/openvino-toolkit\">OpenVINO Toolkit<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/NervanaSystems\/ngraph\">ngraph<\/a><\/li>\n<li><a href=\"https:\/\/eng.uber.com\/michelangelo\/\">Uber\u2019s Michelangelo<\/a><\/li>\n<\/ul><\/div><\/section><section class=\"episode-body-section\"><div class=\"transcript\"><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-0\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Cormac, thanks for joining me here at O\u2019Reilly AI. It\u2019s great to have the chance to talk to you. I know you just got out of your talk a little bit earlier\u2026 You talked about \u201cPortability and performance in embedded deep learning - can we have both?\u201d I wanna dig into that a little bit more later, but first I\u2019d like to hear \u2013 I know you work and help the Movidius group at Intel, and I\u2019d love for you to let the audience know what Movidius is if they haven\u2019t heard about it, what you\u2019re doing and what you\u2019re working on now.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-1\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, thanks Daniel. Good to talk. My name is Cormac Brick. I lead VPU architecture at Movidius, part of Intel. VPU for us is a Visual Processing Unit, and that\u2019s the key engine we have in our product line; I kind of lead that architecture.<\/p><p>At Movidius we\u2019re very passionate about machine learning and computer vision at the edge. This is something we\u2019ve beet at for a long time, going back 5-6 years, even before we were part of Intel, and we have multiple products now in the field. We\u2019ve learned a lot as a result of all of that interaction with customers over the years, and the goal of the talk this morning was to really kind of reflect back some of that knowledge, as in what have we learned about tuning neural networks for embedded silicon, and then also tuning embedded silicon for neural networks\u2026 To kind of just reflect back what some of the realities are when you go to take a network to the edge, what\u2019s really required to make that run really well.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-2\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Awesome. Just to dig into that a little bit deeper, when you\u2019re talking about customers that are tuning neural networks for the edge on things like VPUs, which you mention, what are some of the customer use cases around this, and people that have found a lot of value in going down that road?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-3\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, sure. At Movidius we have customers who are engaged heavily in things like digital security and smart city-type use cases, really making more intelligent commerce; that\u2019s one big use case. We\u2019ve also shipped a lot of products on drones, that\u2019s another use case, as well as alot of things around robotics and smart devices and camera devices as well. There are Google Clips products on the market now that uses our Myriad 2 silicon. A lot of the DJI drones have used the Myriad 2 silicon as well, and there are things like \u2013 you can wave at your drone using your hands to control it, and then put out your palm and the drone can land in the palm of your hand, so\u2026 Really compelling use cases have been enabled through our silicon through the use of both vision and AI working hand in hand.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-4\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>[] Awesome, yeah. And just to kind of confirm that, I was actually at GopherCon last week and one of the keynotes - I think on the second day, or something - they used a drone with a Myriad chip in it to do some facial recognition. It was some cool stuff.<\/p><p>So let\u2019s dive into a little bit more about what you talked about - is there in these types of use cases where you\u2019re wanting to run your neural network in a drone, or in a camera, or whatever it is, explain a little bit the tension between portability and performance that we\u2019ve seen in the past, and the state of it now.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-5\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Sure. If you go the archive, or if you go to NIPS, or ICLR, or CVPR, leading academic vision conferences, we\u2019ll find there\u2019s a lot of work being done to optimize neural networks for things like ImageNet or MS COCO, or kind of academic data sets. And that\u2019s awesome in terms of pushing the envelope over the fields and advancing the science, and it\u2019s moving super-fast, right? So then typically when embedded engineers will start off on a problem, they have access to that sort of research and these sorts of models, and then they\u2019re gonna wanna do something that\u2019s gonna work for them and their device, right? And one of the things that they would find is a lot of the models that are available out there were tuned on ImageNet, which is great at recognizing 1,000 classes of images, and it can differentiate one sort of whale from a different type of porpoise, and this sort of stuff; very fine-grained classification on specific tasks\u2026<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-6\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Important problems, yeah.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-7\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah\u2026 Not so much in the real world, right? We have different problems to solve. So then in the real world we may care about \u201cHey, my robot wants to be able to recognize 100 common objects found in the home\u201d or \u201cIn this security camera we want to be able to recognize these different types of objects that are happening.\u201d So yeah, different problems\u2026 And often those problems are simpler than the thousand-class problem of an ImageNet. So one of the things we were talking about this morning is using techniques like model printing, and sparsification to \u2013 if you\u2019re doing what we called domain transfers, so you go from your thousand-class problem, say if you were taking ResNet-50, and you\u2019re then retraining that for your home robot which wants to recognize 100 images, you\u2019ll find that you can get away with a much simpler network, with a less representational capacity to solve that 100-image problem than the one you started off in the 1000-image problem.<\/p><p>So we were sharing some results and some techniques specifically around channel pruning, which is a very powerful technique when you are doing domain transfer to a simpler problem domain, and also looking at techniques like sparsification, which is introducing more zeroes into a neural network, because that\u2019s [unintelligible ] on platforms that support memory compression of neural network models. It\u2019ll enable those models to run much faster in bandwidth-limited devices such as those typically found on the edge.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-8\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Awesome. Let\u2019s say that I\u2019m working on one of these robotics problems, I\u2019m using a neural network and I want to pursue some of these methods to prune it down or optimize it for that setting or for that architecture\u2026 What\u2019s the process and the barriers that I would face as of now going into that and what\u2019s the state of the usability of these tools, and that sort of thing?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-9\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>That\u2019s a great question, because for sure, we were presenting a lot of works this morning saying \u201cHey, we were able to take a network and do pruning, and quantization, and sparsification, and go from 8-bit weights to 4-bit weights and this sort of stuff\u2026 But you know, straight up today, pretty non-trivial to repeat the results that we were showing this morning, right? To bridge that gap, we have a network in Intel that\u2019s part of the AI Products Group - as part of the Intel AI Products Group there is an open source project called Distiller. It\u2019s one of the resources listed in my slides, I think on the final slide, and I believe they\u2019ll get posted to O\u2019Reilly at some point\u2026<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-10\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>[] Yeah, we\u2019ll put them in the show links here, as well.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-11\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah. So there\u2019s a link to something on GitHub called Distiller, and there one of the things we\u2019re doing is if you went back maybe 12 months ago you would have found \u201cOh, this is an awesome quantization technique that somebody published, some grad student published a PyTorch fork or something with this\u2026 And then here\u2019s something else that was available in TensorFlow for quantization, and here\u2019s something else that was available in a different framework\u2026 What we were doing is really kind of taking all of those techniques that are available in a fairly fragmented way across the internet and trying to put them under one roof in a way that\u2019s a little bit easier to access. That was the goal of the Distiller project, to show that\u2026 And it\u2019s an ongoing project at Intel within AIPG to have this kind of set of tools. So they\u2019re available in PyTorch, and that\u2019s great, because PyTorch can export to ONNX, which is then widely available.<\/p><p>In addition to the work we\u2019re doing, it\u2019s entirely appropriate though to give a shout-out to the work the TensorFlow team are doing. Under TensorFlow Contrib there\u2019s a bunch of useful tools there on both quantization and on pruning as well\u2026 And there\u2019s a pretty strong ecosystem there, also showing a variety of techniques.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-12\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Okay, yeah. So it is at least to a point where I could get a model off of some repository - maybe in PyTorch, or whatever - and have some tooling that\u2019s publicly available to prune that down for certain architectures.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-13\">Cormac Brick<\/a><\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-14\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>What about prepping the model for maybe certain specialized hardware? You mentioned VPUs, and I know there\u2019s a lot of other people pursuing things around, of course, GPUs, but also FPGAs, and other things\u2026 What is that state of the art\u2026? Are these pruning methods and all of that tied into that world, or is that something totally separate?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-15\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>That\u2019s also a good question, and it was one of the goals of the talk today, to show that \u201cHey, here\u2019s four key techniques that you can use, that will work well on any hardware, and on some hardware it will work extra-well\u2026 But if you employ these techniques, you\u2019re not going to hurt your model\u2019s ability to run across a broad range of silicon\u201d And those techniques specifically are kind of model pruning, sparsification, quantizing a network to 8 bits, and then doing further quantization on weights to use kind of a lower bit depth.<\/p><p>So if you employ those four techniques, you will still have a model \u2013 you know, if you take a model and you represent it in ONNX, or in TensorFlow, you will still have a model that can work well on a wide variety of devices, but on some devices it\u2019s gonna work extra-well, right? \u2026because of different abilities to run quantized models at varying degrees of acceleration, and also different silicon will have varying degrees of (let\u2019s say) weight compression technology. Even in extreme cases, for sparsity, there\u2019s some silicon out there that can process sparse networks directly and in an accelerated fashion.<\/p><p>So against a variety of silicon you can employ these four techniques and get really good results across a range of silicon, and even better results in some other silicones. That was the core point.<\/p><p>To answer the second part of your question, in the final slide we were making the point as well that hey, if you set out with a single network and you know the piece of silicon you\u2019re running on, absolutely there\u2019s other techniques you can employ to really fit that piece of silicon as best as you can, to really make this one network shine on this combination of this network and this silicon\u2026 And there\u2019s been some very interesting work published on that in the last couple of months, and it\u2019s a pretty hot research topic now, showing how using \u2013 you may be familiar with [unintelligible ] so being able to use that type of techniques to refine a model or to learn a model that works really well on a particular version of silicon, with these types of performance claims and tradeoffs. That\u2019s a pretty active area of research, that\u2019s pretty interesting.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-16\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Awesome, awesome. I know that one of the things that I\u2019ve appreciated as I\u2019m hacking on things at home is that a lot of the stuff that you\u2019ve come out with through Movidius makes it really easy to experiment with neural networks on a lot of different types of devices, through the Neural Compute Stick, and other things\u2026 I was wondering if you had any interesting stories or customer experiences that you\u2019ve heard about of people enabling new sorts of things with these devices.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-17\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>[] Yeah, we\u2019ve really enjoyed the experience of launching the first version of the Neural Compute Stick based on Myriad 2, and it was great to get out there and meet lots of developers\u2026 And also, when we launched that - we announced it some time before, and we really launched it then at CVPR last year\u2026 Yeah, it was great to see what everybody was doing, but also to kind of show them, \u201cHey, AI at the edge is possible.\u201d If you go back 15 months or two years ago, people really associated AI with the cloud, right? So our first goal was to break down those perceived barriers, and for more people to be able to use AI and to see \u201cHey, AI at the edge is possible.\u201d That was our initial goal, and it was a great experience, very enjoyable talking to all the developers.<\/p><p>A couple of things we\u2019ve seen - we\u2019ve seen people use this, one of the software ambassadors for Intel used this to do a prototype water filter, so taking the guts of a microscope, putting that up to a camera, into a Raspberry Pi, with a Movidius Neural Compute Stick connected, and being able to show that you could actually use this to detect water impurities, so to have an entirely offline water impurity detection device that could be used effectively on premises, at the edge, with no cloud connection, or anything like this\u2026 Super-cool idea, and we were able to show that that\u2019s possible.<\/p><p>Equally, we\u2019ve had people putting them on a drone to detect sharks in the water, also doing prototype medical imagining to detect melanoma on skin, also kind of driven by image classification. Yeah, so those are just a few things, but there\u2019s been a lot of other fun projects posted on GitHub. I don\u2019t have a link to our models site and example site, but I can provide you with them for the blog page also.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-18\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Awesome. Yeah, we\u2019ll make sure that gets in our show notes, for sure. Well, I appreciate you taking time again. To wrap things up here, I was wondering - from your perspective, since you\u2019ve been working in this space for a while, what can we look forward to over the next couple of years with performing AI at the edge? What are you excited about and what do you think we\u2019ll see over the next couple years?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-19\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>I think we\u2019re definitely going to see a lot more silicon become available, both from Movidius Intel, and also from a bunch of competitors. I think that\u2019s gonna be really interesting, as inference silicon \u2013 those kind of metrics business people would track, like the number of ops per [unintelligible ] we can deliver, or the number of ops per dollar we can deliver\u2026 And we\u2019ll expect both of those metrics to progress at really fast paces over the next number of years. And if I look at what people are able to do with the first version of the Neural Compute Stick, with the capabilities that has, and \u2013 well, I can\u2019t disclose product roadmaps, but some visibility of the types of things we\u2019re gonna see, in terms of the volume of compute that various people can bring to market, at much lower price points and much lower power points\u2026 I\u2019m really excited to see how that\u2019s gonna play out, and the types of things people are gonna do with that. I think it\u2019s gonna be a very exciting space to watch in the next few years.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-20\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Awesome. Well, thank you again for taking time, and enjoy the rest of the conference.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/cbrick\/\" rel=\"external\" title=\"Cormac Brick\">                      <img alt=\"Cormac Brick Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/kwqJ\/avatar_small.png?v=63709812883\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-21\">Cormac Brick<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Thanks, Daniel.<\/p><\/div><\/div><div class=\"transcript-item transcript-item--break\"><div class=\"transcript-item-text\"><strong class=\"transcript-break\">Break<\/strong><p>[]<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-23\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Thanks for joining us, Mike. It\u2019s great to chat with you and meet you here at O\u2019Reilly AI. I\u2019ve heard about Michelangelo, this ML platform that you guys have developed at Uber, and I\u2019d love to hear a little bit more about it, but first, give us a little background of who you are and how you ended up where you are.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-24\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, thanks. Happy to be here. I currently am the product lead for ML Infrastructure at Uber. That encompasses a lot of things, most notably the Michelangelo platform. A little bit of background on me - I\u2019m an electrical engineer by training, and out of school I worked at Google; one of the places I got my ML chops, so to speak\u2026 Which is weird to say. I worked on the ads team at Google, specifically the ads auction group, and I was the product manager for all of the ML signals that go into the ads auction there\u2026 So these really real-time, high-scale, super-productionized ML systems that predict if you\u2019re gonna click an ad, and if this ad\u2019s gonna be relevant, and stuff like that. That\u2019s where I learned how to do ML right, and probably best in industry in terms of productionized machine learning.<\/p><p>Then about three years ago I joined Uber, where we started the Michelangelo \u2013 which is not named after me in any way\u2026<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-25\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>That\u2019s a shame.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-26\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, people ask me that question all the time\u2026 We started the Michelangelo platform, which helps data scientists and engineers across the company build ML systems, prototype, explore ML systems, build them, and then deploy them into production and serve predictions at scale.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-27\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>If you\u2019re in a company that\u2019s trying to build up their AI presence within the company, why would they need an ML platform? Why isn\u2019t Jupyter Notebooks everywhere just fine for people?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-28\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>That\u2019s a good question. The state of Uber\u2019s ML stuff about three years ago was that a lot of people were trying to do that. There were a lot of people \u2013 you know, grad students learned how to build their ML models in their grad school classes and whatever, and they have their own ways to do it, everybody has their own\u2026 I use R, I use Python\u2026 And we saw was that people were either trying to productionize an R model and run an R runtime in production at low latency, which is just very challenging - and people will cringe when they hear that today\u2026 Secondly, you will see data scientists that did have engineer support - they would build up these bespoke towers of infrastructure per use case basis, that would tend to be less well-built, just because they had lower resources, but duplicative of different pieces of infrastructure that people would build to serve these models in production across all the different ML use cases the company has\u2026 And then kind of the scariest is people just wouldn\u2019t get started at all, because some people wouldn\u2019t have a way to get their models into production.<\/p><p>So we saw the opportunity to build a common platform to help people have a unified way to build models, and to (this is the trickiest part) put those same models that they had prototyped on into production, to make those predictions\u2026 And along the way, bring a lot of data science best practices, build into the system reproducibility, common analyses, versioning and all that kind of good stuff, that is kind of like these data science best practices that aren\u2019t yet really well established. We have a lot of really well-established software engineering best practices that everybody knows - CI\/CD, version control and stuff like that\u2026 That stuff is not as well appreciated in the data science community, and it\u2019s just because a lot of this work is new. It\u2019s not like these guys don\u2019t understand the importance of it, but it\u2019s just like the best processes and the best patterns for building this stuff have not yet \u2013 we have not really converged on those yet.<\/p><p>So we\u2019ve spent a lot of effort to focus on where we think this stuff is going to go, and to help build the tools to empower data scientists to do the right thing from the beginning.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-29\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>[] Awesome. How many people are using Michelangelo at Uber these days?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-30\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>That\u2019s really hard to say. I would say we probably have more than \u2013 so this platform supports machine learning use cases across the company\u2026 Everything from fraud-related things, to predicting how long it\u2019s gonna take a car to get to you, to even ranking dishes in the Uber Eats app\u2026 All of the main ML stuff runs through this platform now. And this is just like an interesting kind of platform development challenge - we have a lot of people who kind of use it, and they\u2019re like \u201cHey, I kind of wanna build an ML thing\u201d, and they dabble and explore a couple little models they wanna make, but maybe they never end up fully deploying that model to production\u2026<\/p><p>So it\u2019s kind of tricky to say like \u201cHow many actual use cases do you have on this system?\u201d We know it\u2019s well over 100, but it\u2019s hard for us in the platform to say \u201cIs this something that this human is just using this as an experiment, or is it fully productionized and deployed across the whole company?\u201d That\u2019s an area that we\u2019ve just under-invested in a little bit, but we think there\u2019s a lot more to do there.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-31\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>As you\u2019ve seen people start to use the system, are there features of it that kind of surprised you in the sense of how people relied on them, or things that people needed that you didn\u2019t expect that they would need, or other things?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-32\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, that\u2019s a really good question, and I\u2019ve been reflecting on this a lot recently\u2026 You know, I\u2019m the product manager, so it\u2019s kind of my job, but\u2026 [laughs] The thing I would say that has gotten disproportionate adoption - given maybe even our under-investment into this, where you still could do a lot more in this space, but our users just adopted this overwhelmingly and they love it - is our feature store, which is part of the platform. Common problems for managing features for ML workflows are that you have to clean your data and transform your data, and combine it all, and also historically, into a training data set, so you can train your model. But then once your model is created, how do you do all those same transforms in the same way, the same pre-processing to that data in real-time when you deploy your model? So there\u2019s kind of this dual type of ETL that happens in different computing environments, that\u2019s really tricky, and\u2013<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-33\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, possibly on a variety of resources\u2026<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-34\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah. And we see a lot of vendor solutions here, but I feel like we don\u2019t see anybody really tackling that kind of stuff, and I think it\u2019s partially because it\u2019s not sexy at all to work on that stuff, and also just because it\u2019s super-hard to do it properly. We\u2019ve provided some nice ways for people to define their feature transforms to the platform, and then be confident that those transforms will happen consistently across both computer environments, real-time and offline\u2026<\/p><p>But I think the other interesting thing is we saw \u2013 let\u2019s take the Uber Eats world, for example. They probably have more than ten different models that they use to rank dishes and whatever they do. A lot of those models use the same kind of features, and before this feature store, data scientists didn\u2019t have any insight into \u201cHey, other people that were working on similar problems, what kind of feature pipelines have they built?\u201d And then when this feature store came along, now when a data scientist wants to start a new model, they can just look and see \u201cWhat features exist that are relevant for me? Let me just start my model exploration process with the X features that are most relevant to this problem from the beginning.\u201d So there\u2019s a whole new element of collaboration, visibility, feature sharing that was previously not there. I really don\u2019t see solutions in that space in the industry today either, so I think that\u2019s a really promising area.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-35\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>[] Cool. Yeah, I look forward to hearing more about that, and definitely if you publish anything about that, I\u2019ll be happy to post that on the show links here.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-36\">Mike Del Balso<\/a><\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-37\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>The other thing I was curious about, just from the fact that you mentioned before that the incentives for data scientists are kind of different and not always aligned with producing production-ready models and all of those things - how do you build up a team to build an ML platform where really you kind of need software engineering experience to be able to build something that\u2019s production-ready, but you need the knowledge and the expertise around machine learning to be able to understand what to build so it\u2019s gonna be relevant to the people you\u2019re building it for?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-38\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>I think one of the nice things is that we\u2019ve had a little bit of \u2013 the leadership in our organization has been relatively forward-thinking to be willing to fund the development of an ML platform much earlier than I think is common in the industry, and that\u2019s allowed us to get it wrong a couple times before we got it right\u2026 But we feel like we really got it really right now\u2026 And there\u2019s like a tension between data scientists wanting nimbleness and flexibility throughout their exploration and prototyping stages, and if you think of any productionized system, it\u2019s super-stable, so how do you accomplish both of those constraints? It\u2019s a challenge.<\/p><p>Some of the design philosophy that we\u2019re taking - and this is always developing - is we\u2019re trying to allow data scientists to work within our system using the tools that are most relevant for them. We\u2019d love for them to work in Jupyter Notebooks, and write all their models the way they normally would, where we can provide some helpful APIs for them - for example the feature store stuff to pull in their data, so they don\u2019t have to reimplement a whole bunch of work that already exists in terms of enterprise intelligence that\u2019s already been done. But after a certain point, when the prototyping stage is complete, if you think of this machine learning lifecycle, where it\u2019s like \u201cNow I wanna actually use this in production\u201d, and maybe it doesn\u2019t mean you\u2019re gonna launch it to the whole company and you\u2019re done with the project; it could just be like \u201cI wanna experiment with this on live traffic.\u201d<\/p><p>We focus on making it relatively low activation energy to take your prototype and transform it into something that can go into these productionized, well-engineered, hardened systems, that we can be confident will be stable from a systems perspective\u2026 And we still wanna give data scientists the ability to monitor these models that are in production for not just systems issues like whatever applies to typical microservices, but also the data science monitoring, how accurate this is modeled over time, are there any model drifts, stuff like that.<\/p><p>So there\u2019s a story for data scientists throughout the lifecycle, and a story for engineers throughout the lifecycle, and it balances\u2026 And the challenge is \u201cHow do you balance between those at the different stages, taking into account all of the priorities for both stakeholders throughout?\u201d<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-39\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Awesome. Yeah, that gives some great perspective. Well, to kind of end things out here, are there places online where people can find out more about what you guys have done, and maybe also some things that you\u2019ve put out there that you might wanna share?<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-40\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Yeah, that\u2019s a good question. We\u2019ve published a blog post about Michelangelo - I think October 2017 - and it\u2019s pretty easy if you just search \u201cmichelangelo ml platform\u201d on Google, you can find that. We\u2019ve published a lot of other pieces about related ML work we\u2019ve done, and I think we\u2019re likely to in the near future open up the [unintelligible ] a little bit more on Michelangelo, so stay tuned.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.datadan.io\/\" rel=\"external\" title=\"Daniel Whitenack\">                      <img alt=\"Daniel Whitenack Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/secure.gravatar.com\/avatar\/1ea0c31fbcab54853329d9f7bcdb6d6d.jpg?s=150&amp;d=mm\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-41\">Daniel Whitenack<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Cool. Awesome! We look forward to that. Thanks for joining, and enjoy the rest of the conference.<\/p><\/div><\/div><div class=\"transcript-item\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/www.linkedin.com\/in\/michaeldelbalso\/\" rel=\"external\" title=\"Mike Del Balso\">                      <img alt=\"Mike Del Balso Says\" class=\"transcript-item-speaker-image lazy\" height=\"36\" src=\"https:\/\/cdn.changelog.com\/uploads\/avatars\/people\/D1oj\/avatar_small.png?v=63709813157\" width=\"36\" \/><\/a><p class=\"transcript-item-speaker-name\"><a href=\"#transcript-42\">Mike Del Balso<\/a><\/p><\/div><div class=\"transcript-item-text\"><p>Thank you, I appreciate it.<\/p><\/div><\/div><div class=\"transcript-item transcript-item--break\"><div class=\"transcript-item-speaker\"><a href=\"https:\/\/changelog.com\/about\" title=\"Changelog Media\">                  \n<\/a><p class=\"transcript-item-speaker-name\">Changelog<\/p><\/div><div class=\"transcript-item-text\"><p>Our transcripts are <a href=\"https:\/\/github.com\/thechangelog\/transcripts\/blob\/master\/practicalai\/practical-ai-21.md\" rel=\"external\" title=\"View transcript source on GitHub\">open source on GitHub<\/a>. Improvements are welcome. \ud83d\udc9a<\/p><\/div><\/div><\/div><\/section><p><a class=\"button\" href=\"https:\/\/changelog.com\/practicalai\">View all episodes<\/a><\/p>","created_at":"2018-11-26T14:23:05+0000","updated_at":"2018-11-26T14:23:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":25,"domain_name":"changelog.com","preview_picture":"https:\/\/cdn.changelog.com\/uploads\/covers\/practical-ai-original.png?v=63688789033","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":718,"title":"JSON-LD Playground","url":"https:\/\/json-ld.org\/playground\/","content":"JSON-LD Playground\n    \n    <div class=\"container\"><p>Play around with JSON-LD markup by typing out\n        some JSON below and seeing what gets generated from it at the\n        bottom of the page. Pick any of the examples below to get started.\n        The playground uses the <a href=\"https:\/\/github.com\/digitalbazaar\/jsonld.js\">jsonld.js JSON-LD processor<\/a> which <a href=\"http:\/\/json-ld.org\/test-suite\/reports\/#subj_3\">fully conforms<\/a> to the JSON-LD <a href=\"http:\/\/json-ld.org\/spec\/latest\/json-ld\/\">Syntax<\/a> and <a href=\"http:\/\/json-ld.org\/spec\/latest\/json-ld-api\/\">API<\/a> specifications.\n      <\/p><p><strong>NOTE<\/strong>: This is the <strong>stable<\/strong> playground supporting <a href=\"https:\/\/www.w3.org\/TR\/2014\/REC-json-ld-20140116\/\">JSON-LD 1.0<\/a>. A <strong>development<\/strong> version that supports many recent <a href=\"https:\/\/json-ld.org\/spec\/latest\/json-ld\/\">W3C Community Group Draft JSON-LD 1.1<\/a> and <a href=\"https:\/\/github.com\/digitalbazaar\/jsonld.js\">jsonld.js<\/a> features is available on the <a href=\"https:\/\/json-ld.org\/playground-dev\/\">dev playground<\/a>.<\/p><\/div>","created_at":"2018-11-26T13:35:32+0000","updated_at":"2018-11-26T13:35:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"json-ld.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":717,"title":"How difficult is it to install stock Android on a device which has a modified version of the OS pre-installed?","url":"https:\/\/www.quora.com\/How-difficult-is-it-to-install-stock-Android-on-a-device-which-has-a-modified-version-of-the-OS-pre-installed","content":"<div class=\"header\"><div id=\"wFSTMRKB1\"><div class=\"QuestionArea\"><div id=\"wondWdHV1\"><div id=\"wondWdHV3\"><div class=\"TopicList QuestionTopicHorizontalList\" id=\"__w2_wondWdHV4_topics_list\"><div id=\"wondWdHV15\"><div class=\"QuestionTopicHorizontalListItems SimpleToggle TopicListItems2 Toggle QuestionTopicListItems\" id=\"__w2_wondWdHV16__truncated\"><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wondWdHV18_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wondWdHV23\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Android-Devices\" id=\"__w2_wondWdHV24_link\">Android Devices<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wondWdHV20_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wondWdHV25\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Android-operating-system\" id=\"__w2_wondWdHV26_link\">Android (operating system)<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wondWdHV22_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wondWdHV27\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Smartphones\" id=\"__w2_wondWdHV28_link\">Smartphones<\/a><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"wondWdHV5\"><div class=\"question_text_edit\"><h1>How difficult is it to install stock Android on a device which has a modified version of the OS pre-installed?<\/h1><\/div><\/div><div id=\"wondWdHV7\"><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wR141Ex69_answer\"><div><div id=\"wR141Ex610\"><div class=\"ContentHeader AnswerHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wR141Ex619\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Ian-C-Ward\" id=\"__w2_wR141Ex633_name_link\">Ian C. Ward<\/a>, Android user<\/div><div class=\"u-inline-block\" id=\"wR141Ex614\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/How-difficult-is-it-to-install-stock-Android-on-a-device-which-has-a-modified-version-of-the-OS-pre-installed\/answer\/Ian-C-Ward\" id=\"__w2_wR141Ex640_link\">Answered 252w ago<\/a> \u00b7 Author has 110 answers and 253.1k answer views<\/div><\/div><\/div><div id=\"__w2_wR141Ex612_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wR141Ex612_follow_button\"><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wR141Ex631_suggested_section\"><\/div><\/div><div id=\"__w2_wR141Ex69_answer_content\"><div class=\"ExpandedAnswer ExpandedContent\" id=\"__w2_wR141Ex613_expanded_content\"><div id=\"wR141Ex616\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">\"Stock\" generally means \"as provided by the vendor.\" In this sense, the modified version of Android that comes pre-installed on these devices is \"stock.\"<p>A better way to think of the Play Store editions of these phones is \"neat\" or un-modified Android. Only when Google sells the device is this un-modified version \"stock.\"<\/p><p>Both the modified and un-modified versions of these devices include drivers for the specific hardware that the device contains, which allow it to function.<br \/>Most drivers are proprietary\/closed source, so when Google makes a Play Edition device, they typically have to get these drivers from the hardware vendor.<\/p><p>If you want to install the \"neat\" version of Android on a device, you'll need to get the code from the Android Open Source Project (AOSP), compile it, and upload it to your device and hope that it boots. I would not recommend this to start out with.<\/p><p>For the beginner, the best bet is probably to find a custom ROM based on AOSP. There are a lot of these in forums at <a href=\"http:\/\/xda-developers.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"xda-developers.com\">xda-developers.com<\/a>.<\/p><p>Personally, I use CyanogenMod (<a href=\"http:\/\/cyanogenmod.org\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"cyanogenmod.org\">cyanogenmod.org<\/a>), since it attempts to remain true to AOSP, while improving it with crazy customizability. I had little issue installing it on my Samsung Galaxy S3, following the instructions on this page (but using the Odin method documented in the \"Helpful Tip\" section at the top):<br \/>http:\/\/wiki.cyanogenmod.org\/w\/In...<br \/>They even have an installer now, which makes the process pretty painless.<\/p><p>Then after you get a better feel for installing pre-compiled ROMs, you can pull the source and compile it yourself and upload to your phone. CyanogenMod also has instructions on how to do this:<br \/>http:\/\/wiki.cyanogenmod.org\/w\/De...<\/p><p>Once you've got a handle on building stable source for a phone, then you're probably ready to dig in and build your own ROM from scratch.<\/p><p>Granted, you can attempt to go straight to building your own ROM from scratch at the beginning; I just recommend this route as it will ease you into Android development and (hopefully) save you a bricked phone.<\/p><p>On a side note, I was initially a bit averse to switching to Android, since I was familiar with the iPhone and had a lot of apps and music, but now that I've switched, I'm glad I did and I don't expect to be going back anytime soon at all.<\/p><p>Happy hunting!<\/p><\/div><\/div><div id=\"wR141Ex624\"><div id=\"__w2_wR141Ex613_survey_wrapper\"><div id=\"wR141Ex634\"><div class=\"OutOfDateAnswerSurvey ComponentSurvey AnswerSurvey\" id=\"__w2_wR141Ex635_survey\"><div class=\"survey_wrapper\" id=\"__w2_wR141Ex635_survey_inner\"><div class=\"inner\"><div class=\"caption_wrapper\"><div class=\"locked\"><div class=\"caption\">Your feedback is private.<\/div><\/div><div class=\"question_wrapper\"><div class=\"flex_item\"><div class=\"icon_frame\"><div class=\"icon out_of_date_icon\"><\/div><div class=\"title_text\">Is this answer still relevant and up to date?<\/div><\/div><div class=\"flex_item button_group\" id=\"__w2_wR141Ex635_survey_buttons\"><input class=\"survey_button\" type=\"button\" value=\"Yes\" data-group=\"js-editable\" id=\"__w2_wR141Ex635_positive_action\" \/><input class=\"survey_button\" type=\"button\" value=\"No\" data-group=\"js-editable\" id=\"__w2_wR141Ex635_negative_action\" \/><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wekHQeA09_answer\"><div><div id=\"wekHQeA010\"><div class=\"ContentHeader AnswerHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wekHQeA019\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Justin-Sturgis\" id=\"__w2_wekHQeA033_name_link\">Justin Sturgis<\/a>, Founder Spaciety.com &amp; SimpleLotus, LLC<\/div><div class=\"u-inline-block\" id=\"wekHQeA014\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/How-difficult-is-it-to-install-stock-Android-on-a-device-which-has-a-modified-version-of-the-OS-pre-installed\/answer\/Justin-Sturgis\" id=\"__w2_wekHQeA040_link\">Answered 252w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wekHQeA012_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wekHQeA012_follow_button\"><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wekHQeA031_suggested_section\"><\/div><\/div><div id=\"__w2_wekHQeA09_answer_content\"><div class=\"ExpandedAnswer ExpandedContent\" id=\"__w2_wekHQeA013_expanded_content\"><div id=\"wekHQeA016\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">Not for everyone, but for most it is incredibly simple. The key is having an exploit to unlock the device, a stock ROM (software\/OS) that matches the device and a release of an appropriate stock kernel. Once you gain root access to the device and install a custom recovery, flashing vanilla Android or any other custom ROM is a piece of cake. I've run stock Android on my SGS4 for a bit.<p>Every device and every manufacturer is a little different so you will want to find the method for your specific device (carrier, nation, software, etc.)<\/p><p>The best place to start is <a href=\"http:\/\/xda-developers.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"xda-developers.com\">XDA-developers.com<\/a>. Even if you just want more specific details on the process with no intention of owning or rooting your own Android device, it's an interesting site and a great resource.<\/p><p>Obviously there is a lot of uncommon terminology above (root, recovery, ROM, kernel, etc.) which you can learn more about at XDA as well.<\/p><p>In case you are wondering, I've owned 12 Android devices including the G1\/Dream (the first Android device) and have rooted each of them. Picked up my 12th tonight, the Note 3 which I'll root tomorrow morning. I'm sure many other folks will come in here with more info and different perspectives, but feel free to ask if you have any questions or want more specific info.<\/p><\/div><\/div><div id=\"wekHQeA024\"><div id=\"__w2_wekHQeA013_survey_wrapper\"><div id=\"wekHQeA034\"><div class=\"OutOfDateAnswerSurvey ComponentSurvey AnswerSurvey\" id=\"__w2_wekHQeA035_survey\"><div class=\"survey_wrapper\" id=\"__w2_wekHQeA035_survey_inner\"><div class=\"inner\"><div class=\"caption_wrapper\"><div class=\"locked\"><div class=\"caption\">Your feedback is private.<\/div><\/div><div class=\"question_wrapper\"><div class=\"flex_item\"><div class=\"icon_frame\"><div class=\"icon out_of_date_icon\"><\/div><div class=\"title_text\">Is this answer still relevant and up to date?<\/div><\/div><div class=\"flex_item button_group\" id=\"__w2_wekHQeA035_survey_buttons\"><input class=\"survey_button\" type=\"button\" value=\"Yes\" data-group=\"js-editable\" id=\"__w2_wekHQeA035_positive_action\" \/><input class=\"survey_button\" type=\"button\" value=\"No\" data-group=\"js-editable\" id=\"__w2_wekHQeA035_negative_action\" \/><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wKufsDi29_answer\"><div><div id=\"wKufsDi210\"><div class=\"ContentHeader AnswerHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wKufsDi219\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Shane-Britton-3\" id=\"__w2_wKufsDi233_name_link\">Shane Britton<\/a>, Developer\/Recognized Contributor  at XDA Developers (2015-present)<\/div><div class=\"u-inline-block\" id=\"wKufsDi214\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/How-difficult-is-it-to-install-stock-Android-on-a-device-which-has-a-modified-version-of-the-OS-pre-installed\/answer\/Shane-Britton-3\" id=\"__w2_wKufsDi238_link\">Answered 40w ago<\/a> \u00b7 Author has 710 answers and 1m answer views<\/div><\/div><\/div><div id=\"__w2_wKufsDi212_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wKufsDi212_follow_button\"><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wKufsDi231_suggested_section\"><\/div><\/div><div id=\"__w2_wKufsDi29_answer_content\"><div class=\"ExpandedAnswer ExpandedContent\" id=\"__w2_wKufsDi213_expanded_content\"><div id=\"wKufsDi216\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">Not difficult at all. Moreover, it does not require root access or an in depth understanding of device flashing. Nearly all Android devices have stock firmware packages available for free online. Sometimes these packages are released by the manufacturers,sometimes they are leaked by the manufacturers, and other times Android developers and enthusiasts compile these packages for love of the Android community. The key component of a pure stock firmware package is that it is digitally signed by the manufacturer. This signature ensures that it can be flashed to an unrooted and locked device, as well as a device which is OEM unlocked and fully rooted. On Motorola device, for example, fastboot mode can be used to flash the device with a stock firmware package, irregardless of the modified state of the device. On Samsung devices this can all be done with a program known as Odin. On LG devices the LG Flash Tool does the job. Nearly all OEMs have a means with which to flash the device back to unmodified stock condition.<\/p><p class=\"ui_qtext_para\">For your particular device, to flash it back to stock condition, I recommend searching <a href=\"http:\/\/xda-developers.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"xda-developers.com\">XDA-Developers Android Forums<\/a> for your device model. There you will typically find a stock firmware package for your device as well as the resources and precise instructions for flashing your device.<\/p><\/div><\/div><div id=\"wKufsDi224\"><div id=\"__w2_wKufsDi213_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wIzdg06d9_answer\"><div><div id=\"wIzdg06d10\"><div class=\"ContentHeader AnswerHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wIzdg06d19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Josh-Hall-120\" id=\"__w2_wIzdg06d33_name_link\">Josh Hall<\/a>, MCSE A+ CCNA Engineering &amp; Information Technology (1999)<\/div><div class=\"u-inline-block\" id=\"wIzdg06d14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/How-difficult-is-it-to-install-stock-Android-on-a-device-which-has-a-modified-version-of-the-OS-pre-installed\/answer\/Josh-Hall-120\" id=\"__w2_wIzdg06d38_link\">Answered 55w ago<\/a> \u00b7 Author has 368 answers and 132.5k answer views<\/div><\/div><\/div><div id=\"__w2_wIzdg06d12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wIzdg06d12_follow_button\"><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wIzdg06d31_suggested_section\"><\/div><\/div><div id=\"__w2_wIzdg06d9_answer_content\"><div class=\"ExpandedAnswer ExpandedContent\" id=\"__w2_wIzdg06d13_expanded_content\"><div id=\"wIzdg06d16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">The answer 100% depends on your ability to find a raw firmware image for your device. If there isn't one availible then it will be beyond most people's ability to do. Also it will depend on if your bootloader is unlocked or not and if anyone has been able to obtain full root on that device. Without root permissions changing your firmware won't work as the phone will just reload it's backed up copy of the modified one it came with. If you mean modified like you bought a phone from someone already rooted and flashed then it will be fairly easy. Just go to XDA and search for your phone you will find everything u need<\/p><\/div><\/div><div id=\"wIzdg06d24\"><div id=\"__w2_wIzdg06d13_survey_wrapper\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-25T22:59:12+0000","updated_at":"2018-11-25T22:59:12+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.quora.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["css","typography"],"is_public":false,"id":716,"title":"How to use legible font sizes for all devices","url":"https:\/\/varvy.com\/mobile\/legible-font-size.html","content":"<h2>Text should be easy to read<\/h2><p>There are many screen sizes now so we must make the text of our pages legible (easily readable) on the smallest and biggest screens.<\/p><p>If you are getting the \"text too small to read\" or \"use legible font size\" errors, this article will help.<\/p><p>To check if your font size is too small try the <a href=\"https:\/\/varvy.com\/mobile\/\">Mobile SEO tool<\/a>.<\/p><h2>Make font size work for any size device<\/h2><p>There are a few ways to accomplish this. We will discuss the following options...\n<\/p><ol><li>Buy a mobile theme \/ template<\/li>\n<li>Use responsive design \/ media queries (css)<\/li>\n<li>Design different versions of your site<\/li>\n<\/ol><p>What option you choose for your webpage or blog will pretty much be determined by what time you are willing to put into it, and what experience you have (or are willing to gain).<\/p><h2>1. Buying a mobile theme or template<\/h2><p>Google recommends <sup><a href=\"https:\/\/www.google.com\/webmasters\/tools\/mobile-friendly\/\"><small>1<\/small><\/a><\/sup> responsive design. With a responsive design theme you will have the problem of font size already solved for you. <\/p><p>Good places to buy modern responsive themes and templates are..<\/p><ul><li><a href=\"http:\/\/www.shareasale.com\/r.cfm?b=346200&amp;u=869072&amp;m=28169&amp;urllink=&amp;afftrack=\">StudioPress<\/a>\n<\/li><li><a href=\"https:\/\/creativemarket.com\/themes?u=pinopat\" rel=\"nofollow\">Creative Market<\/a><\/li>\n<li><a href=\"https:\/\/themeforest.net\/page\/top_sellers?ref=pinopat\" rel=\"nofollow\">Theme Forest<\/a><\/li>\n<\/ul><h2>2. Use responsive design \/ media queries<\/h2><p>If you have experience with CSS and are not scared to learn a bit more of it, you can control in great detail how well readable your text is on any screen size.\n<\/p><h2>How font size works<\/h2><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-size-get-v.png\" alt=\"webpage getting font from css\" \/><p>The size of your font is declared in your CSS file. If no font size is declared the browser will use a default font size which is typically 16px.<\/p><p>The thing that controls your font size in CSS is called (not surprisingly) \"font-size\". It looks like this in your css...\n<\/p><div class=\"code\"><p>p{font-size:120%;}<\/p><\/div><p>Font size can use different unit of measures to declare how big your text should be - pixels (px), points (pt), EMs (em), and percent (%) in the example code above it uses percent.\n<\/p><p>Google recommends <sup><a href=\"https:\/\/developers.google.com\/speed\/docs\/insights\/UseLegibleFontSizes\"><small>2<\/small><\/a><\/sup> the following when it comes to font and responsive design:\n<\/p><ul><li>Use a base font size of 16 CSS pixels. Adjust the size as needed\nbased on properties of the font being used.<\/li>\n<li>Use sizes relative to the base size to define the typographic\nscale.<\/li>\n<li>Text needs vertical space between its characters and may need to\nbe adjusted for each font. The general recommendation is to use the\nbrowser default line-height of 1.2em.<\/li>\n<li>Restrict the number of fonts used and the typographic scale: too\nmany fonts and font sizes lead to messy and overly complex page\nlayouts.<\/li>\n<\/ul><p>Let's discuss each of the above recommendations...<\/p><h2>Use a base font size of 16 CSS pixels<\/h2><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-set-g.png\" alt=\"css file with 16px as base font size\" \/><p>This is saying to declare a font size for your page explicitly as 16px. To do this means we use CSS to state a page wide instruction that the font size is 16px. This is usually done using \"body\" or \"html\". This also tells us to \"adjust\" our fonts (which is explained in the next point). \n<\/p><h2>Use sizes relative to the base size to define the typographic\nscale.<\/h2><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-set-percent-g.png\" alt=\"css file using percents\" \/><p>Now that we have a font size for the whole page, we will need to adjust that font size so our H1s and h2s etc. are appropriately larger than the base font size of 16 pixels.<\/p><p>The way Google is recommending to do this is \"relative\" to the base font size. For our examples we will use percentages to accomplish this.<\/p><p>In the above image we see that our H1 font-size is declared as \"250%\". This basically just means that our H1 text will be displayed two and half times bigger than our base size of 16px. This is the \"relative\" part. It means that wherever we declare a font size, it is always declared as bigger, smaller, or the same size as our base font size of 16px.<\/p><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/mobile-responsive-v.png\" alt=\"font shown on mobile and desktop\" \/><p>By doing this we are ensuring that no matter what size screen our page is seen on, our font size relationships will always stay the same. Our H1 will always be two and a half the size of our base font whether shown on the tiniest phone or the most gigantic TV screen.<\/p><p>This will also play an important part of making sure that font is easily readable (legible) when a user is viewing the page on mobile or smaller screens.<\/p><h2>Text needs vertical space between its characters and may need to be adjusted for each font.<\/h2><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-lineheight-g.png\" alt=\"line heights displayed ideally and poorly\" \/><p>Another important aspect of readable font is amount of space between the lines of text. In CSS this is called \"line-height\".<\/p><p>Google recommends to use at least the browser default line-height of 1.2 which means that the space between the lines of text are at least 1.2 times the size of the text itself.<\/p><p>To make this happen we can declare the line-height in our CSS...<\/p><div class=\"code\"><p>p{font-size:120%;line-height:1.2;}<\/p><\/div><p>We often read our our own webpages on our desktop monitors and may have tight text and small line heights and it works fine. The problem occurs when you see that same text on a smaller screen and the text is nearly impossible to read without zooming.<\/p><p>By making sure our line-height is at least 1.2 we help ensure that our text is easier to read even on those small screens.<\/p><h2>List link issues<\/h2><p>Links need more room on mobile devices, and this can often become an issue when you have links in ordered or unordered lists.<\/p><p>To solve this issue try:<\/p><div class=\"code\"><p>ul{line-height:200%;}<\/p><\/div><p>This will make sure there is plenty room around the links in a list by increasing the line height of the list. You will have to tweak it for your own needs by changing \"200%\" to another value that fits your site design.<\/p><h2>Restrict the number of fonts used and the typographic scale<\/h2><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-size-few-g.png\" alt=\"font usage shown ideally and poorly\" \/><p>Too many fonts and font sizes lead to messy and overly complex page layouts. Overly complex page layouts may or not be readable on a desktop, but they sure won't be easily readable on a smaller screen.<\/p><p>Remember that your \"small\" text will be even smaller on a mobile device.<\/p><h2>How to make font readable on any device<\/h2><p>The size of our text is controlled by the <a href=\"https:\/\/varvy.com\/mobile\/media-queries.html\">media queries<\/a> we declare for different sized screens. We can say that for mobile users we want one size of text, and declare a different size of text for desktops.<\/p><p>Let's take a look at a simple example media query to control the size of our font.<\/p><div class=\"code\"><p>html{font-size:100%;} <br \/>@media(min-width:60em){html{font-size: 100%}}<\/p><\/div><p>In the above example we have the same font-size declared for large screens and small screens. The font size is \"100%\" and since we declared the base size of 16 pixels earlier, this means our font size will be 16 pixels on both desktops and smaller mobile devices. This size font may be shown differently on different devices, but it will always be nice and legible. <\/p><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-size-same-g.png\" alt=\"font size 16px on desktop and mobile devices\" \/><p>This is quite readable on both types of devices especially when we follow all the above listed recommendations. We will typically have legible text on both desktops and devices. In this case it works.<\/p><p>Where a problem may come along is when we style our desktop text to be rather small. In this situation the same text will be seen even smaller on mobile or smaller screens.<\/p><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-size-small-mobile-g.png\" alt=\"small font shown as unreadable on smaller screens\" \/><p>In this scenario we will need to change our media query. The first example code above had the same value (100%) for both desktops and mobile devices.<\/p><p>Let's look at a media query that would have smaller tight text on a desktop, but still have readable text on a smaller screen.\n<\/p><div class=\"code\"><p>html{font-size:110%;} <br \/>@media(min-width:60em){html{font-size: 90%}}<\/p><\/div><p>This media query is saying that text should be 90% (of our default 16px) on desktop, but on mobile it should be 110% (of our default 16px).<\/p><img src=\"denied:data:image\/png;base64,R0lGODlhAQABAAD\/ACwAAAAAAQABAAACADs=\" data-v=\"\/mobile\/img\/font-size-adjust-g.png\" alt=\"font use shown to work on both mobile and desktop\" \/><p>This means that the text is easily readable on different sized screens.<\/p><ul><li>To learn more about <a href=\"https:\/\/varvy.com\/mobile\/media-queries.html\">media queries<\/a>.\n<\/li><li>The above media query settings may or may not work for you. The values were only used to illustrate points.\n<\/li><\/ul><h2>3. Design different versions of your site<\/h2><p>If your website is not responsive, you likely are using a mobile url or dynamic serving for your mobile visitors.<\/p><p>In this scenario you are just creating different versions of your webpages, and can style the mobile page however you want using normal CSS.<\/p><p>It would be good practice to ensure your text is large enough (16px) and that your line heights are big enough (1.2) to allow text to be easily readable.<\/p><p><img src=\"https:\/\/varvy.com\/patrick-sexton.jpg\" width=\"50\" height=\"50\" alt=\"Patrick Sexton\" class=\"circle vert\" \/> by <\/p><br \/>","created_at":"2018-11-25T03:14:58+0000","updated_at":"2018-11-25T03:15:07+0000","published_by":["Patrick Sexton"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"varvy.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","autoload","composer"],"is_public":false,"id":715,"title":"Re-Introducing Composer &#8211; the Cornerstone of Modern PHP Apps","url":"https:\/\/www.sitepoint.com\/re-introducing-composer\/","content":"<p class=\"wp-special\"><em>This article was peer reviewed by <a href=\"https:\/\/www.sitepoint.com\/author\/wancheta\">Wern Ancheta<\/a>. Thanks to all of SitePoint\u2019s peer reviewers for making SitePoint content the best it can be!<\/em><\/p><p class=\"wp-special\"><strong>This popular article was updated on May 21st 2017 to provide a more comprehensive introduction into an essential, modern tool.<\/strong><\/p><hr \/><p>In this article, we will tackle the basics of Composer, and what makes it such a powerful and useful tool.<\/p><div class=\"widget maestro maestro-content-type-html hide-for-tablet-SP hide-for-mobile-SP\" id=\"maestro-636\"><p>--ADVERTISEMENT--<\/p><\/div><p>Before we go into detail, there are two things that we need to have in mind:<\/p><ul><li>What Composer is:\n<p>As we can see on their <a href=\"https:\/\/getcomposer.org\">website<\/a>: <em>\u201cComposer is a tool for dependency management in PHP. It allows you to declare the libraries your project depends on and it will manage (install\/update) them for you.\u201d<\/em><\/p><\/li>\n<li>What Composer is not:\n<p>Composer is not a package manager. It deals with packages but in a \u201cper project\u201d way. While it provides a global installation option, it does not install anything globally by default.<\/p><\/li>\n<\/ul><p>Essentially, Composer allows you to declare and manage every dependency of your PHP projects.<\/p><p>Let\u2019s now install Composer so we can see it in action.<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2016\/08\/1472279623logo-composer-transparent-badass.png\" alt=\"Composer Logo\" width=\"450\" \/><\/p><h2 id=\"installation\">Installation<\/h2><p>There are two ways of installing Composer: locally and globally. Because Composer is such a useful and widespread tool, I always recommend installing it globally, as you\u2019ll probably be working on more than one project at once.<\/p><p><em>Note that if you use an environment like <a href=\"http:\/\/www.sitepoint.com\/quick-tip-get-homestead-vagrant-vm-running\/\">Homestead Improved<\/a>, it\u2019s installed globally by default. If you\u2019re unfamiliar with Vagrant and Homestead, <a href=\"http:\/\/www.sitepoint.com\/re-introducing-vagrant-right-way-start-php\/\">this post<\/a> will clear things up, and if you\u2019d like to dive deeper, <a href=\"http:\/\/bit.ly\/phpenv-sp\">this book will reveal some wild mysteries to you<\/a>.<\/em><\/p><p>To install Composer, follow <a href=\"https:\/\/getcomposer.org\/download\/\">the instructions<\/a> for your operating system.<\/p><p>If you\u2019re on a Unix system, after installing it you can run the command:<\/p><pre class=\"bash language-bash\">mv composer.phar \/usr\/local\/bin\/composer&#13;\n<\/pre><p>This moves the <code>composer.phar<\/code> file to a directory that\u2019s on your path (the location of your filesystem where the OS looks for executable files by default). This way, you can use just the <code>composer<\/code> command instead of <code>composer.phar<\/code>.<\/p><p>Running the <code>composer<\/code> command shows the information page:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2017\/01\/1491890237Composer1.jpg\" alt=\"Composer main information\" \/><\/p><p>Some of Composer\u2019s more useful commands are:<\/p><ul><li><code>composer help &lt;command&gt;<\/code> \u2013 will show the help page for a given command.<\/li>\n<li><code>composer self update<\/code> \u2013 Updates composer to the latest available version.<\/li>\n<li><code>composer search<\/code> \u2013 Searches for packages.<\/li>\n<li><code>composer status<\/code> \u2013 Shows a list of dependencies that have been modified locally. If we install something from the source (using the <code>--prefer-source<\/code> option), we will end up with a clone of that package in the <code>\/vendor<\/code> folder. If we make some changes to that package, the <code>composer status<\/code> command will show us a git status for those changes.<\/li>\n<li><code>composer diagnose<\/code> \u2013 Diagnoses the system for common errors. This is especially useful for debugging, as it checks for things like connectivity to Packagist, free disk space, and git settings.<\/li>\n<\/ul><p>We will focus our attention on <code>composer init<\/code>, <code>composer create-project<\/code> and <code>composer require<\/code>.<\/p><h2 id=\"usingcomposer\">Using Composer<\/h2><p>To manage dependencies in a project, Composer uses a <code>json<\/code> file. This file describes all the dependencies, and holds some metadata as well.<\/p><p>In this example we will create a new PHP project in which we will need <a href=\"https:\/\/www.sitepoint.com\/tag\/phpunit\">PHPUnit<\/a> for our unit testing. There are several ways to use Composer to manage this dependency.<\/p><ul><li><code>composer init<\/code> will launch the wizard. The wizard will guide us through the generation of our <code>composer.json<\/code> file.<\/li>\n<li><code>composer search phpunit<\/code> will search for all packages available that have the keyword <code>phpunit<\/code>. After we choose the package we want, we can then use <code>composer require &lt;package&gt;<\/code>. This command will not only create the <code>composer.json<\/code> file, it will also download all the dependencies specified.<\/li>\n<li>Manually write the <code>composer.json<\/code> file (not recommended \u2013 easy to slip up with syntax and brings no apparent benefits). Here\u2019s an example:<\/li>\n<\/ul><pre class=\"javascript language-javascript\">{&#13;\n    \"require\": {&#13;\n        \"phpunit\/phpunit\": \"^4.8\"&#13;\n    }&#13;\n}&#13;\n<\/pre><ul><li>The <code>require<\/code> key takes objects that map package names (e.g. <code>phpunit\/phpunit<\/code>) to version constraints (e.g. <code>4.8<\/code>). Composer uses this information to search for the right set of files in package \u201crepositories\u201d that you register using the <code>repositories<\/code> key (more info <a href=\"https:\/\/getcomposer.org\/doc\/05-repositories.md\">here<\/a>) or on <a href=\"http:\/\/packagist.org\">Packagist<\/a>, the default package repository which we\u2019ll discuss a little later.<\/li>\n<\/ul><p>If you used <code>composer init<\/code> or the manual method, you need to run <code>composer install<\/code>. This will download all the dependencies needed for the packages described in the <code>composer.json<\/code> file.<\/p><p>Note that writing the file manually and then running <code>composer install<\/code> is the same as running <code>composer require &lt;package&gt;<\/code>.<\/p><blockquote><p>Using the <code>composer require &lt;package&gt;<\/code> approach is a recommended <a href=\"https:\/\/www.sitepoint.com\/mastering-composer-tips-tricks\/\">best practice<\/a>.<\/p><\/blockquote><p>The first time you run the <code>composer install<\/code> command, the dependencies are downloaded and their installed versions are registered in the <code>composer.lock<\/code> file. The next time someone runs <code>composer install<\/code> on the project, it will not get the latest available package versions, but the ones that are registered in the <code>composer.lock<\/code> file. This way, Composer ensures that your project doesn\u2019t break due to unexpected code changes in newer versions of the packages.<\/p><h3 id=\"updatingyourpackages\">Updating Your Packages<\/h3><p>The <code>composer.lock<\/code> file ensures that everyone in the project uses the same version of the packages by not allowing newer versions to be downloaded.<\/p><p>If for any reason we want to stick to a particular version, than we should mention the appropriate version in the <code>composer.json<\/code> file. For example:<\/p><pre class=\"javascript language-javascript\">  {&#13;\n      \"require\": {&#13;\n          \"phpunit\/phpunit\": \"4.8\"&#13;\n      }&#13;\n  }&#13;\n<\/pre><p>But what if we want to update our dependencies to the latest versions?<\/p><p>To update dependencies, Composer provides the <code>composer update<\/code> command. This command will not only download the latest versions of our dependencies, it will also update the <code>composer.lock<\/code> file to the new versions. Running <code>composer update<\/code> is actually the same as deleting <code>composer.lock<\/code> and then re-running <code>composer install<\/code>.<\/p><blockquote><p>It is generally recommended to commit the <code>composer.lock<\/code> file into the version control system when working on a project (app) and to ignore it when working on a package. A package should always strive to use the latest version of its dependencies, while an app is generally more fragile. More discussion about this and alternative approaches <a href=\"https:\/\/www.reddit.com\/r\/PHP\/comments\/2kt4cw\/composer_its_all_about_the_lock_file\/\">here<\/a>.<\/p><\/blockquote><h2 id=\"thevendorfolder\">The Vendor Folder<\/h2><p>Composer downloads all dependencies into the <code>\/vendor<\/code> folder, which it creates if it doesn\u2019t exist. It also creates a <code>vendor\/autoload.php<\/code> file which looks something like this:<\/p><pre class=\"php language-php\">&lt;?php&#13;\n&#13;\n\/\/ autoload.php @generated by Composer&#13;\n&#13;\nrequire_once __DIR__ . '\/composer\/autoload_real.php';&#13;\n&#13;\nreturn ComposerAutoloaderInit7a731d2067108a4739c3dd65be722f17::getLoader();&#13;\n<\/pre><p>Once you <code>include<\/code> or <code>require<\/code> this file in your code, you can use all the functionality provided by your dependencies. This is as simple as adding the following snippet to your code:<\/p><pre class=\"php language-php\">require __DIR__ . '\/vendor\/autoload.php';&#13;\n<\/pre><h2 id=\"autoloading\">Autoloading<\/h2><p>The autoloading provided by Composer to access the project\u2019s dependencies is very useful, and can also be used to access our own application\u2019s code. To do this, we need to register our code in the <code>composer.json<\/code> file using the <code>autoload<\/code> key.<\/p><pre class=\"javascript language-javascript\">{&#13;\n    \"autoload\": {&#13;\n        \"psr-4\": {\"Foo\\\\\": \"src\/\"}&#13;\n    }&#13;\n}&#13;\n<\/pre><p>First, Composer will register a <a href=\"https:\/\/www.sitepoint.com\/battle-autoloaders-psr-0-vs-psr-4\/\">PSR-4 autoloader<\/a> for the <code>Foo<\/code> namespace.<\/p><p>Then, we map the <code>Foo<\/code> namespace to the <code>\/src<\/code> folder. The <code>\/src<\/code> folder should be at the same level as the <code>\/vendor<\/code> folder (root of the project). After this, we need to re-generate the <code>vendor\/autoload.php<\/code> file. Composer provides a command for this: <code>composer dump-autoload<\/code>.<\/p><p><em>Pro-tip: <code>composer du<\/code> would also work, because composer runs any command which is uniquely identified by its first few letters. In <code>dump-autoload<\/code>\u2018s case, there\u2019s no other command in the Composer namespace that starts with <code>du<\/code>, so Composer knows that\u2019s the one we want. This works with all Composer commands.<\/em><\/p><p>This way we are adding our own code to the autoloader. In addition to PSR-4 autoloading, Composer also supports PSR-0, classmap, and files autoloading. Check out the <a href=\"https:\/\/getcomposer.org\/doc\/04-schema.md#autoload\">documentation<\/a> for more information on autoloading with Composer.<\/p><h2 id=\"configurationvaluesandminimumstability\">Configuration Values and Minimum Stability<\/h2><p>Sometimes, your project will depend on packages that don\u2019t have a stable version released. Composer defaults to preventing unstable packages from being installed.<\/p><p>In those cases, it\u2019s easy to add the <code>minimum-stability<\/code> flag into the <code>composer.json<\/code> file, and Composer will allow their installation.<\/p><pre class=\"bash language-bash\">composer config minimum-stability dev&#13;\n<\/pre><p>Almost all configuration values can be edited from the command line, as explained in <a href=\"https:\/\/getcomposer.org\/doc\/03-cli.md\">this section<\/a> of the docs.<\/p><h2 id=\"globalpackages\">Global Packages<\/h2><p>Some packages can come in handy if installed globally. Case in point, tools like <a href=\"https:\/\/www.sitepoint.com\/building-an-spress-svbtle-theme-responsive-static-blogs\/\">Spress<\/a> and <a href=\"https:\/\/www.sitepoint.com\/premium\/courses\/laravel-5-2930?aref=bskvorc\">Laravel<\/a> can have command line tools for generating new projects, re-running some common commands, etc. Likewise, a test suite like <a href=\"https:\/\/www.sitepoint.com\/tag\/phpunit\">PHPUnit<\/a> can benefit from being runnable from any part of the system, and makes the project it\u2019s testing much lighter.<\/p><p>In general, packages can be installed globally with the <code>global<\/code> option, like so:<\/p><pre class=\"bash language-bash\">composer global require phpunit\/phpunit&#13;\n<\/pre><p>However, <a href=\"https:\/\/www.sitepoint.com\/composer-global-require-considered-harmful\/\">according to some<\/a>, this is considered harmful. Your mileage may vary \u2013 we\u2019ve generally found that when a professional and modern approach of <a href=\"https:\/\/bit.ly\/phpenv-sp\">developing in an isolated environment<\/a> is followed, <code>global require<\/code> is 100% safe to use.<\/p><h2 id=\"packagist\">Packagist<\/h2><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2017\/01\/1491890263packagist.png\" alt=\"Packagist\" \/><\/p><p><a href=\"https:\/\/packagist.org\/\">Packagist<\/a> is Composer\u2019s main repository. This is where all the Composer packages are aggregated, which means that when we execute <code>composer search &lt;package&gt;<\/code> we are actually searching Packagist. The Packagist website also has information on all the packages and a useful search and sort engine. This way, PHP developers don\u2019t have to spend days searching for packages on Github.<\/p><p>Not only does Packagist aggregate PHP packages, it also encourages PHP developers to submit their own packages, strengthening the community.<\/p><h3 id=\"submittingapackagetopackagist\">Submitting a Package to Packagist<\/h3><p>Let\u2019s create a simple PHP package so we can see how easy it is to submit it to Packagist. We are going to create a simple class that converts inches to centimetres and centimetres to inches. Let\u2019s create a <code>\/src\/Converter\/Converter.php<\/code> file. Inside this file, we are going to have the following code:<\/p><pre class=\"php language-php\">&lt;?php &#13;\n&#13;\nnamespace Converter;&#13;\n&#13;\nclass Converter&#13;\n{&#13;\n    public function convertToCm($value)&#13;\n    {&#13;\n        return $value * 0.3937;&#13;\n    }&#13;\n&#13;\n    public function convertToInch($value)&#13;\n    {&#13;\n        return $value * 2.54;&#13;\n    }&#13;\n}&#13;\n<\/pre><p>As you can see, a very simple class. The second step is to navigate to the root of our project and run the <code>composer init<\/code> command to run the wizard and generate the <code>composer.json<\/code> file.<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2017\/01\/1491890259init.jpg\" alt=\"Run init on our package\" \/><\/p><p>The last thing we need to do is add the autoload key to our <code>composer.json<\/code> file. The final file should look like this:<\/p><pre class=\"javascript language-javascript\">{&#13;\n    \"name\": \"dailymatters\/inch-conv\",&#13;\n    \"description\": \"Inch To Cm converter\",&#13;\n    \"type\": \"library\",&#13;\n    \"license\": \"creative commons\",&#13;\n    \"authors\": [&#13;\n        {&#13;\n            \"name\": \"DailyMatters\",&#13;\n            \"email\": \"email@gmail.com\"&#13;\n        }&#13;\n    ],&#13;\n    \"minimum-stability\": \"dev\",&#13;\n    \"require\": {},&#13;\n    \"autoload\": {&#13;\n        \"psr-0\": {&#13;\n            \"Converter\": \"src\/\"&#13;\n        }&#13;\n    }&#13;\n}&#13;\n<\/pre><p>Let\u2019s now run <code>composer install<\/code> so that our autoloader can be created. Then, let\u2019s test our package. In the root of our project, let\u2019s create an <code>index.php<\/code> file.<\/p><pre class=\"php language-php\">&lt;?php&#13;\n&#13;\nrequire_once '\/vendor\/autoload.php';&#13;\n&#13;\nuse Converter\\Converter;&#13;\n&#13;\n$conv = new Converter;&#13;\n&#13;\n$result = $conv-&gt;convertToCm(1);&#13;\necho $result;&#13;\n&#13;\n$result = $conv-&gt;convertToInch(1);&#13;\necho $result;&#13;\n<\/pre><p>As you can see, we are using the <code>require_once '\/vendor\/autoload.php';<\/code> line to autoload our class and then using it to do some conversions.<\/p><p>Testing it on the command line with <code>php index.php<\/code> should produce the desired result.<\/p><p>To submit this package to Packagist, we need to push it to a Github repo. Then, we sign in to Packagist using our Github account, hit <code>Submit<\/code>, and point to our newly created Github repo.<\/p><blockquote><p>Remember to update Packagist whenever you update your repo in Github! You can either configure Packagist to do it automatically, or you can access your package on Packagist and click the <code>Update<\/code> button. This way, your package in Packagist will always mirror your package\u2019s Github state.<\/p><\/blockquote><h2 id=\"conclusion\">Conclusion<\/h2><p>Composer has revolutionized the way PHP developers work, and promises to continue doing so. It\u2019s a powerful and essential tool for any PHP developer\u2019s tool belt, and the cornerstone of many modern PHP applications.<\/p><p>Are you using Composer? Is anything about it unclear? Let us know in the comments below.<\/p>","created_at":"2018-11-24T02:16:11+0000","updated_at":"2018-11-24T02:16:15+0000","published_at":"2017-05-21T16:00:17+0000","annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"www.sitepoint.com","preview_picture":"https:\/\/www.sitepoint.com\/wp-content\/themes\/sitepoint\/assets\/images\/icon.php.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","php","psysh"],"is_public":false,"id":714,"title":"emacs-php\/psysh.el","url":"https:\/\/github.com\/emacs-php\/psysh.el","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<a href=\"https:\/\/melpa.org\/#\/psysh\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/b8ca6cf463e59ca7283e8b690e4c708313aa1e5c\/68747470733a2f2f6d656c70612e6f72672f7061636b616765732f70737973682d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"https:\/\/melpa.org\/packages\/psysh-badge.svg\" \/><\/a>\n<p><a href=\"http:\/\/psysh.org\/\" rel=\"nofollow\">PsySH<\/a>, PHP interactive shell (REPL)<\/p>\n<h2><a id=\"user-content-compatibility\" class=\"anchor\" aria-hidden=\"true\" href=\"#compatibility\"><\/a>Compatibility<\/h2>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<h3><a id=\"user-content-a-the-easy-way\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-the-easy-way\"><\/a>A: The easy way<\/h3>\n<pre>$ wget psysh.org\/psysh\n$ chmod +x psysh\n<\/pre>\n<p>And copy or make symlink to your <code>$PATH<\/code> dir.<\/p>\n<h3><a id=\"user-content-b-the-other-easy-way\" class=\"anchor\" aria-hidden=\"true\" href=\"#b-the-other-easy-way\"><\/a>B: The other easy way<\/h3>\n<p>Get Composer.  See <a href=\"https:\/\/getcomposer.org\/download\/\" rel=\"nofollow\">https:\/\/getcomposer.org\/download\/<\/a><\/p>\n<pre>$ composer g require psy\/psysh:@stable\n<\/pre>\n<h3><a id=\"user-content-c-project-local-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#c-project-local-repl\"><\/a>C: Project local REPL<\/h3>\n<p>Set <code>psysh-comint-buffer-process<\/code> (buffer local variable).<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq psysh-comint-buffer-process \"path\/to\/shell.php\")<\/pre><\/div>\n<p><code>shell.php<\/code> is for example:<\/p>\n<div class=\"highlight highlight-text-html-php\"><pre>#!\/usr\/bin\/env php\n&lt;?php\n\/\/ \u2193Namespace for your project\nnamespace Nyaan;\n\/\/ load Composer autoload file\nrequire_once __DIR__ . '\/vendor\/autoload.php';\n\/\/ load other initialize PHP files\n\/\/ require_once \u2026\necho __NAMESPACE__ . \" shell\\n\";\n$sh = new \\Psy\\Shell();\n\/\/ Set project namespace in REPL\nif (defined('__NAMESPACE__') &amp;&amp; __NAMESPACE__ !== '') {\n    $sh-&gt;addCode(sprintf(\"namespace %s;\", __NAMESPACE__));\n}\n$sh-&gt;run();\n\/\/ Termination message\necho \"Bye.\\n\";<\/pre><\/div>\n<p>See also <a href=\"http:\/\/cho-phper.hateblo.jp\/entry\/2015\/11\/10\/031000\" rel=\"nofollow\">http:\/\/cho-phper.hateblo.jp\/entry\/2015\/11\/10\/031000<\/a> <b>(Japanese)<\/b><\/p>\n<\/article>","created_at":"2018-11-24T02:11:37+0000","updated_at":"2018-11-24T02:11:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/31366612?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","psysh"],"is_public":false,"id":713,"title":"PsySH","url":"https:\/\/psysh.org\/","content":"<p>\nPsySH has tab completion for variable names, functions, classes, methods, properties, and even files!\nJust start typing, then hit <code>[tab]<\/code>. But you probably already guessed that.\n<\/p><p>\n<small>(Tab completion requires Readline- or Libedit-enabled PHP. But you really should have that anyway)<\/small>\n<\/p>","created_at":"2018-11-24T02:09:25+0000","updated_at":"2018-11-24T02:09:33+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"psysh.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","psysh"],"is_public":false,"id":712,"title":"Interactive PHP Debugging with PsySH","url":"https:\/\/www.sitepoint.com\/interactive-php-debugging-psysh\/","content":"<p>It\u2019s 1:00 a.m., the deadline for your web application\u2019s delivery is in 8 hours\u2026 and it\u2019s not working.  As you try to figure out what\u2019s going on, you fill your code with <code>var_dump()<\/code> and <code>die()<\/code> everywhere to see where the bug is\u2026<\/p><p>You are annoyed. Every time you want to try out a return value or variable assignment, you have to change your source code, execute your application, and see the results\u2026  In the end, you are unsure of whether or not you\u2019ve removed all those var_dumps from the code. Is this situation familiar to you?<\/p><h2 id=\"psysh-to-the-rescue\">PsySH to the rescue<\/h2><p><a href=\"http:\/\/psysh.org\/\">PsySH<\/a> is a Read-Eval-Print Loop (or <a href=\"http:\/\/en.wikipedia.org\/wiki\/Read%E2%80%93eval%E2%80%93print_loop\">REPL<\/a>).  You may have used a REPL before via your browser\u2019s javascript console.  If you have, you know that it possesses a lot of power and can be useful while debugging your JS code. <\/p><div class=\"widget maestro maestro-content-type-html hide-for-tablet-SP hide-for-mobile-SP\" id=\"maestro-636\"><p>--ADVERTISEMENT--<\/p><\/div><p>Talking about PHP, you might have used PHP\u2019s interactive console (<code>php -a<\/code>) before. There, you can write some code and the console will execute it as soon as you press enter:<\/p><pre class=\"language-\">php -a&#13;\nInteractive shell&#13;\n&#13;\nphp &gt; $a = 'Hello world!';&#13;\nphp &gt; echo $a;&#13;\nHello world!&#13;\nphp &gt;<\/pre><p>Unfortunately, the interactive shell is not a REPL since it lacks the \u201cP\u201d (print).  I had to execute an <code>echo<\/code> statement to see the contents of <code>$a<\/code>.  In a true REPL, we would have seen it immediately after assigning the value to it.<\/p><p>You can install PsySH globally either with a <code>composer g require<\/code>, or downloading the PsySH executable:<\/p><p><strong>Composer<\/strong><\/p><pre class=\"language-\">composer g require psy\/psysh:~0.1&#13;\npsysh<\/pre><p><strong>Direct download<\/strong> (Linux\/Mac)<\/p><pre class=\"language-\">wget psysh.org\/psysh&#13;\nchmod +x psysh&#13;\n.\/psysh<\/pre><p>Additionally, you can have it included per project with composer as you\u2019ll see later in this article.<\/p><p>Now let\u2019s play with PsySH a little.<\/p><pre class=\"language-\">.\/psysh                                                                                                                                             &#13;\n&#13;\nPsy Shell v0.1.11 (PHP 5.5.8 \u2014 cli) by Justin Hileman                                                                                                                                                                              &#13;\n&gt;&gt;&gt;<\/pre><p>The main help will be your best friend. It\u2019s what will give you all sorts of commands and their explanations:<\/p><pre class=\"language-\">&gt;&gt;&gt; help&#13;\n&#13;\n  help      Show a list of commands. Type `help [foo]` for information about [foo].      Aliases: ?&#13;\n  &#13;\n  ls        List local, instance or class variables, methods and constants.              Aliases: list, dir&#13;\n  &#13;\n  dump      Dump an object or primitive.&#13;\n  &#13;\n  doc       Read the documentation for an object, class, constant, method or property.   Aliases: rtfm, man &#13;\n  &#13;\n  show      Show the code for an object, class, constant, method or property.&#13;\n  &#13;\n  wtf       Show the backtrace of the most recent exception.                             Aliases: last-exception, wtf?&#13;\n  &#13;\n  trace     Show the current call stack.&#13;\n  &#13;\n  buffer    Show (or clear) the contents of the code input buffer.                       Aliases: buf&#13;\n  &#13;\n  clear     Clear the Psy Shell screen.&#13;\n  &#13;\n  history   Show the Psy Shell history.&#13;\n  &#13;\n  exit      End the current session and return to caller.                                Aliases: quit, q<\/pre><pre class=\"language-\">&gt;&gt;&gt; help ls&#13;\n&#13;\nUsage:&#13;\n&#13;\nls [--vars] [-c|--constants] [-f|--functions] [-k|--classes] [-I|--interfaces] [-t|--traits] [-p|--properties] [-m|--methods] [-G|--grep=\"...\"] [-i|--insensitive] [-v|--invert] [-g|--globals] [-n|--internal] [-u|--user] [-C|--&#13;\ncategory=\"...\"] [-a|--all] [-l|--long] [target]&#13;\n&#13;\nAliases: list, dir&#13;\n&#13;\nArguments:&#13;\n&#13;\n target             A target class or object to list.&#13;\n &#13;\n &#13;\nOptions:&#13;\n&#13;\n --vars             Display variables.&#13;\n &#13;\n --constants (-c)   Display defined constants.&#13;\n &#13;\n --functions (-f)   Display defined functions.&#13;\n &#13;\n --classes (-k)     Display declared classes.&#13;\n &#13;\n --interfaces (-I)  Display declared interfaces.&#13;\n &#13;\n --traits (-t)      Display declared traits.&#13;\n &#13;\n --properties (-p)  Display class or object properties (public properties by default).&#13;\n &#13;\n --methods (-m)     Display class or object methods (public methods by default).&#13;\n &#13;\n --grep (-G)        Limit to items matching the given pattern (string or regex).&#13;\n &#13;\n --insensitive (-i) Case-insensitive search (requires --grep).&#13;\n &#13;\n --invert (-v)      Inverted search (requires --grep).&#13;\n &#13;\n --globals (-g)     Include global variables.&#13;\n &#13;\n --internal (-n)    Limit to internal functions and classes.&#13;\n &#13;\n --user (-u)        Limit to user-defined constants, functions and classes.&#13;\n &#13;\n --category (-C)    Limit to constants in a specific category (e.g. \"date\").&#13;\n &#13;\n --all (-a)         Include private and protected methods and properties.&#13;\n &#13;\n --long (-l)        List in long format: includes class names and method signatures.&#13;\n &#13;\n &#13;\n Help:&#13;\n &#13;\n List variables, constants, classes, interfaces, traits, functions, methods, and properties.&#13;\n &#13;\n Called without options, this will return a list of variables currently in scope.&#13;\n &#13;\n If a target object is provided, list properties, constants and methods of that target. If a class, interface or trait name is passed instead, list constants and methods on that class.&#13;\n &#13;\n e.g. &#13;\n &#13;\n &gt;&gt;&gt; ls&#13;\n &gt;&gt;&gt; ls $foo&#13;\n &gt;&gt;&gt; ls -k --grep mongo -i&#13;\n &gt;&gt;&gt; ls -al ReflectionClass&#13;\n &gt;&gt;&gt; ls --constants --category date&#13;\n &gt;&gt;&gt; ls -l --functions --grep \/^array_.*\/&#13;\n &gt;&gt;&gt;<\/pre><p>Basically, what a REPL can do is:<\/p><pre class=\"language-php\">&gt;&gt;&gt; $a = 'hello';&#13;\n=&gt; \"hello\"&#13;\n&gt;&gt;&gt;<\/pre><p>Please note that if we compare PsySH against PHP\u2019s interactive console, PsySH  prints the <code>$a<\/code> value as soon as it is assigned.<\/p><p>A more complex example can be as follows:<\/p><pre class=\"language-\">&gt;&gt;&gt; function say($a) {&#13;\n...     echo $a;&#13;\n... }&#13;\n=&gt; null&#13;\n&gt;&gt;&gt; say('hello');&#13;\nhello&#13;\n=&gt; null&#13;\n&gt;&gt;&gt;<\/pre><p>I defined the  function <code>say()<\/code> and invoked it.  Those two <code>null<\/code> you see are because neither the function definition nor its execution returned a value (the function echoes the value).  Additionally, while defining the function, the prompt changed from <code>&gt;&gt;&gt;<\/code> to <code>...<\/code>.<\/p><p>Can we define a class and instantiate it?<\/p><pre class=\"language-\">&gt;&gt;&gt; class Foo&#13;\n... {&#13;\n...     protected $a;&#13;\n...&#13;\n...     public function setA($a) {&#13;\n...         $this-&gt;a = $a;&#13;\n...     }&#13;\n...&#13;\n...     public function getA() {&#13;\n...         return $this-&gt;a;&#13;\n...     }&#13;\n... }&#13;\n=&gt; null&#13;\n&gt;&gt;&gt; $foo = new Foo();&#13;\n=&gt; &lt;Foo #000000001dce50dd000000002dda326e&gt; {}&#13;\n&gt;&gt;&gt; $foo-&gt;setA('hello');&#13;\n=&gt; null&#13;\n&gt;&gt;&gt; $foo-&gt;getA();&#13;\n=&gt; \"hello\"&#13;\n&gt;&gt;&gt;<\/pre><p>When I instantiated Foo, the constructor returned a reference to the object.  This is why PsySh printed  <code>&lt;Foo #000000001dce50dd000000002dda326e&gt;<\/code>.  Now let\u2019s see what is interesting about PsySH and objects.<\/p><pre class=\"language-\">&gt;&gt;&gt; ls $foo&#13;\nClass Methods: getA, setA&#13;\n&gt;&gt;&gt;<\/pre><p>If by any chance you forgot which methods the class Foo has defined, you now have the answer.  Have you used a Linux OS or Mac command line interface? Then you might be familiar with the <code>ls<\/code> command.  Remember the <code>-la<\/code> options?<\/p><pre class=\"language-\">&gt;&gt;&gt; ls -la $foo&#13;\nClass Properties:&#13;\n&#13;\n  $a   \"hello\" &#13;\n  &#13;\n&#13;\nClass Methods:&#13;\n&#13;\n  getA   public function getA()&#13;\n  setA   public function setA($a)<\/pre><p>Sweet, isn\u2019t it?<\/p><p>PsySH\u2019s true power shines when integrated with a web application, so let\u2019s build one.<\/p><h2 id=\"demo-app\">Demo app<\/h2><p>I\u2019m going to implement a quick application to showcase the <a href=\"http:\/\/www.oodesign.com\/decorator-pattern.html\">decorator design pattern<\/a>. The UML class diagram of such a pattern is as follows: <br \/><img src=\"http:\/\/www.oodesign.com\/images\/design_patterns\/structural\/decorator-design-pattern-implementation-uml-class-diagram.png\" alt=\"Decorator Design Pattern\" title=\"\" \/><\/p><p>Do not worry if you do not know much about <a href=\"http:\/\/en.wikipedia.org\/wiki\/Unified_Modeling_Language\">UML<\/a> or <a href=\"http:\/\/en.wikipedia.org\/wiki\/Software_design_pattern\">design patterns<\/a>, understanding them is not required for this article. <\/p><p>Also for this project, I created a set of test cases.  Those test cases can be run by <a href=\"http:\/\/phpunit.de\/\">phpUnit<\/a>. Again, you do not have to be familiar with <a href=\"http:\/\/en.wikipedia.org\/wiki\/Unit_test\">Unit Testing<\/a> to understand this article.<\/p><p>The complete source code for this little application can be found at <a href=\"https:\/\/github.com\/sitepoint-examples\/PsySH\">https:\/\/github.com\/sitepoint-examples\/PsySH<\/a><\/p><p>First of all, let\u2019s define our <code>composer.json<\/code> file in order to declare a dependency on <strong>PsySH<\/strong>:<\/p><pre class=\"language-\">{&#13;\n    \"name\": \"example\/psysh\",&#13;\n    \"authors\": [&#13;\n        {&#13;\n            \"name\": \"John Doe\",&#13;\n            \"email\": \"john@doe.tst\"&#13;\n        }&#13;\n    ],&#13;\n    \"require\": {&#13;\n        \"psy\/psysh\": \"~0.1\"&#13;\n    },&#13;\n    \"autoload\": {&#13;\n        \"psr-4\": {\"Acme\\\\\": \"src\/\"}&#13;\n    }&#13;\n}<\/pre><p>After a <code>composer install<\/code> you should be good to go.<\/p><p>Please take a look at the following source code from the file <code>public\/decorator.php<\/code>. It will instantiate the SimpleWindow, DecoratedWindow, and TitledWindow objects to showcase the decorator pattern:<\/p><pre class=\"language-php\">&lt;?php&#13;\nchdir(dirname(__DIR__));&#13;\n&#13;\nrequire_once('vendor\/autoload.php');&#13;\n&#13;\nuse Acme\\Patterns\\Decorator\\SimpleWindow;&#13;\nuse Acme\\Patterns\\Decorator\\DecoratedWindow;&#13;\nuse Acme\\Patterns\\Decorator\\TitledWindow;&#13;\n&#13;\necho PHP_EOL . 'Simple Window' . PHP_EOL;&#13;\n&#13;\n$window = new SimpleWindow();&#13;\n&#13;\necho $window-&gt;render();&#13;\n&#13;\necho PHP_EOL . 'Decorated Simple Window' . PHP_EOL;&#13;\n&#13;\n$decoratedWindow = new DecoratedWindow($window);&#13;\n&#13;\necho $decoratedWindow-&gt;render();&#13;\n&#13;\necho PHP_EOL . 'Titled Simple Window' . PHP_EOL;&#13;\n&#13;\n$titledWindow = new TitledWindow($window);&#13;\n&#13;\necho $titledWindow-&gt;render();<\/pre><p>We can execute the code via PHP\u2019s CLI (command line interface) or through a webserver if one is configured.  We can use PHP\u2019s internal web server too.<\/p><h3 id=\"debugging-in-cli\">Debugging in cli<\/h3><p>The execution of the above code through the command line interface will look like this:<\/p><pre class=\"language-\">php public\/decorator.php &#13;\n&#13;\nSimple Window&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+&#13;\n&#13;\nDecorated Simple Window&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+&#13;\n&#13;\nTitled Simple Window&#13;\n+-------------+&#13;\n|Title        |&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+<\/pre><p>How can we interact with PsySH?  Just add <code>\\Psy\\Shell::debug(get_defined_vars());<\/code> anywhere on the code where you want to debug your application, typically where you would insert a <code>var_dump()<\/code> statement:<\/p><pre class=\"language-php\">&lt;?php&#13;\nchdir(dirname(__DIR__));&#13;\n&#13;\nrequire_once('vendor\/autoload.php');&#13;\n&#13;\n\/\/... a lot of code here&#13;\n&#13;\n$titledWindow = new TitledWindow($window);&#13;\n&#13;\necho $titledWindow-&gt;render();&#13;\n&#13;\n\\Psy\\Shell::debug(get_defined_vars()); \/\/we want to debug our application here!<\/pre><p>After saving the file, we will get the following output:<\/p><pre class=\"language-\">php public\/decorator.php &#13;\n&#13;\nSimple Window&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+&#13;\n&#13;\nDecorated Simple Window&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+&#13;\n&#13;\nTitled Simple Window&#13;\n+-------------+&#13;\n|Title        |&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+&#13;\n&#13;\nPsy Shell v0.1.11 (PHP 5.5.8 \u2014 cli) by Justin Hileman&#13;\n&gt;&gt;&gt;<\/pre><p>The script\u2019s execution will be suspended, and we now have PsySH\u2019s prompt to play with. I am passing <code>get_defined_vars()<\/code> as a parameter to <code>Psy\\Shell::debug()<\/code> so I have access to all defined variables inside the shell:<\/p><pre class=\"language-\">&gt;&gt;&gt; ls&#13;\nVariables: $_COOKIE, $_FILES, $_GET, $_POST, $_SERVER, $argc, $argv, $decoratedWindow, $titledWindow, $window&#13;\n&gt;&gt;&gt;<\/pre><p>Let\u2019s examine the <code>$window<\/code> variable:<\/p><pre class=\"language-\">&gt;&gt;&gt; ls -al $window&#13;\n&#13;\nClass Methods:&#13;\n  render   public function render()  &#13;\n&gt;&gt;&gt;<\/pre><p>Something nice about having PsySH inside an application is that we can examine the source code of an instantiated object.<\/p><pre class=\"language-\">&gt;&gt;&gt; show $window&#13;\nclass Acme\\Patterns\\Decorator\\SimpleWindow implements Acme\\Patterns\\Decorator\\Window&#13;\nclass SimpleWindow implements Window&#13;\n{&#13;\n    public function render()&#13;\n    {&#13;\n        $returnString = &lt;&lt;&lt;EOL&#13;\n+-------------+&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n|             |&#13;\n+-------------+&#13;\nEOL;&#13;\n        return $returnString . PHP_EOL;&#13;\n    }&#13;\n}&#13;\n&gt;&gt;&gt;<\/pre><p>So, <code>$window<\/code> is an instance of SimpleWindow, which implements the Window interface\u2026  I wonder what the source code for the Window interface looks like\u2026<\/p><pre class=\"language-\">&gt;&gt;&gt; show Acme\\Patterns\\Decorator\\Window&#13;\ninterface Acme\\Patterns\\Decorator\\Window&#13;\ninterface Window&#13;\n{&#13;\n    public function render();&#13;\n}&#13;\n&gt;&gt;&gt;<\/pre><p>Why do SimpleWindow and DecoratedWindow have the same output? Let\u2019s examine the <code>$decoratedWindow<\/code> object.<\/p><pre class=\"language-\">&gt;&gt;&gt; ls -al $decoratedWindow&#13;\n&#13;\nClass Properties:&#13;\n  $windowReference   &lt;Digitec\\Patterns\\Decorator\\SimpleWindow #000000003643d67700000000731101b7&gt;  &#13;\n&#13;\nClass Methods:&#13;\n  __construct          public function __construct(Digitec\\Patterns\\Decorator\\Window $windowReference)         &#13;\n  getWindowReference   public function getWindowReference()                                                    &#13;\n  render               public function render()                                                                &#13;\n  setWindowReference   public function setWindowReference(Digitec\\Patterns\\Decorator\\Window $windowReference)  &#13;\n&gt;&gt;&gt;<\/pre><p>This object is \u201cheavier\u201d than the SimpleWindow one, so the source code might be long\u2026 Let\u2019s see the source code for the <code>render()<\/code> method only:<\/p><pre class=\"language-\">&gt;&gt;&gt; show $decoratedWindow-&gt;render&#13;\npublic function render()&#13;\n    public function render()&#13;\n    {&#13;\n        return $this-&gt;getWindowReference()-&gt;render();&#13;\n    }<\/pre><p>The <code>getWindowReference()<\/code> method is invoked, and then it returns the result from the <code>render()<\/code> method.  Let\u2019s check the <code>getWindowReference()<\/code> source:<\/p><pre class=\"language-\">&gt;&gt;&gt; show $decoratedWindow-&gt;getWindowReference&#13;\npublic function getWindowReference()&#13;\n    public function getWindowReference()&#13;\n    {&#13;\n        return $this-&gt;windowReference;&#13;\n    }&#13;\n&gt;&gt;&gt;<\/pre><p>This method is returning the object\u2019s windowReference property, and as we saw from the <code>ls -al<\/code> command above, it is an instance of <code>Acme\\Patterns\\Decorator\\SimpleWindow<\/code>.  Of course, we could have just looked at how <code>DecoratedWindow::__construct()<\/code> works, but this is another way we can check.<\/p><h3 id=\"debugging-with-embedded-server\">Debugging with embedded server<\/h3><p>Unfortunately, debugging through a webserver like Apache is not supported. However, we can debug our application using PHP\u2019s embedded server:<\/p><pre class=\"language-\">$ cd public&#13;\n$ php -S localhost:8080&#13;\nPHP 5.5.8 Development Server started at Wed Jul 23 17:40:30 2014&#13;\nListening on http:\/\/localhost:8080&#13;\nDocument root is \/home\/action\/workspace\/lab\/PsySH\/public&#13;\nPress Ctrl-C to quit.<\/pre><p>The Development Server is now listening for connections on port 8080, so as soon as we request the <code>decorator.php<\/code> file through this web server (<a href=\"http:\/\/localhost:8080\/decorator.php\">http:\/\/localhost:8080\/decorator.php<\/a>), we should see the following:<\/p><pre class=\"language-\">Psy Shell v0.1.11 (PHP 5.5.8 \u2014 cli-server) by Justin Hileman&#13;\n&gt;&gt;&gt;<\/pre><p>We can begin playing with PsySH just as we did with the CLI<\/p><pre class=\"language-php\">&gt;&gt;&gt; ls -al&#13;\n&#13;\nVariables:&#13;\n  $_COOKIE           []                                                                              &#13;\n  $_FILES            []                                                                              &#13;\n  $_GET              []                                                                              &#13;\n  $_POST             []                                                                              &#13;\n  $_SERVER           Array(19)                                                                       &#13;\n  $decoratedWindow   &lt;Acme\\Patterns\\Decorator\\DecoratedWindow #0000000031ef2e3e000000003c2d3a90&gt;  &#13;\n  $titledWindow      &lt;Acme\\Patterns\\Decorator\\TitledWindow #0000000031ef2e39000000003c2d3a90&gt;     &#13;\n  $window            &lt;Acme\\Patterns\\Decorator\\SimpleWindow #0000000031ef2e3f000000003c2d3a90&gt;     &#13;\n  $_                 null                                                                            &#13;\n&gt;&gt;&gt; exit&#13;\nExit:  Goodbye.<\/pre><h3 id=\"debugging-with-unit-tests\">Debugging with unit tests<\/h3><p>As a good developer, you should write unit tests for your code as proof that it is working as expected. In the project\u2019s files you will find the <code>tests<\/code> folder, and if you have phpUnit installed, you can run the tests inside it.<\/p><pre class=\"language-\">cd tests\/&#13;\nphpunit&#13;\nPHPUnit 4.0.14 by Sebastian Bergmann.&#13;\n&#13;\nConfiguration read from \/home\/action\/workspace\/lab\/PsySH\/tests\/phpunit.xml&#13;\n&#13;\n...F+-------------+&#13;\n|Title        |&#13;\n&#13;\n&#13;\nTime: 66 ms, Memory: 4.50Mb&#13;\n&#13;\nThere was 1 failure:&#13;\n&#13;\n1) AcmeTest\\Patterns\\Decorator\\TitledWindowTest::testAddTitle&#13;\nFailed asserting that true is false.&#13;\n&#13;\n\/home\/action\/workspace\/lab\/PsySH\/tests\/Patterns\/Decorator\/TitledWindowTest.php:46&#13;\n                                     &#13;\nFAILURES!                            &#13;\nTests: 4, Assertions: 7, Failures: 1.<\/pre><p>Even though the code appears to run flawlessly, a test is failing. We can examine further by running just the failing test:<\/p><pre class=\"language-\">$ phpunit --verbose --debug --filter=TitledWindowTest::testAddTitle&#13;\nPHPUnit 4.0.14 by Sebastian Bergmann.&#13;\n&#13;\nConfiguration read from \/home\/action\/workspace\/lab\/PsySH\/tests\/phpunit.xml&#13;\n&#13;\n&#13;\nStarting test 'AcmeTest\\Patterns\\Decorator\\TitledWindowTest::testAddTitle'.&#13;\nF+-------------+&#13;\n|Title        |&#13;\n&#13;\n&#13;\nTime: 60 ms, Memory: 4.25Mb&#13;\n&#13;\nThere was 1 failure:&#13;\n&#13;\n1) AcmeTest\\Patterns\\Decorator\\TitledWindowTest::testAddTitle&#13;\nFailed asserting that true is false.&#13;\n&#13;\n\/home\/action\/workspace\/lab\/PsySH\/tests\/Patterns\/Decorator\/TitledWindowTest.php:46&#13;\n                                     &#13;\nFAILURES!                            &#13;\nTests: 1, Assertions: 1, Failures: 1.<\/pre><p>We have the test, file and line where the error is being generated.  Let\u2019s take a look at <code>TitledWindowTest.php<\/code><\/p><pre class=\"language-php\">&lt;?php&#13;\nnamespace AcmeTest\\Patterns\\Decorator;&#13;\n&#13;\nuse PHPUnit_Framework_TestCase;&#13;\nuse Acme\\Patterns\\Decorator\\TitledWindow;&#13;\nuse ReflectionMethod;&#13;\n&#13;\n&#13;\nclass TitledWindowTest extends PHPUnit_Framework_TestCase &#13;\n{&#13;\n    public function testRender()&#13;\n    {&#13;\n        \/* some test code here *\/&#13;\n        &#13;\n    }&#13;\n    &#13;\n    public function testAddTitle()&#13;\n    {&#13;\n        $renderString = 'bar';&#13;\n        &#13;\n        $window = $this-&gt;getMock('Acme\\Patterns\\Decorator\\SimpleWindow');&#13;\n        $window-&gt;expects($this-&gt;any())-&gt;method('render')-&gt;will($this-&gt;returnValue($renderString));&#13;\n        &#13;\n        $titledWindow = new TitledWindow($window);&#13;\n        &#13;\n        $reflectionMethod = new ReflectionMethod($titledWindow, 'addTitle');&#13;\n        $reflectionMethod-&gt;setAccessible(true);&#13;\n        &#13;\n        $rs = $reflectionMethod-&gt;invoke($titledWindow);&#13;\n        &#13;\n        $this-&gt;assertFalse(empty($rs));&#13;\n        &#13;\n    }&#13;\n}<\/pre><p>If you are unfamiliar with phpUnit, do not pay too much attention to the code.  In a nutshell, I\u2019m setting up everything to test the <code>TitledWindow::addTitle()<\/code> method, and expecting to receive a non empty value.<\/p><p>So, how do we use PsySh to check what is going on? Just add the <code>Shell::debug()<\/code> method as we did previously.<\/p><pre class=\"language-php\">&lt;?php&#13;\nnamespace DigitecTest\\Patterns\\Decorator;&#13;\n&#13;\nuse PHPUnit_Framework_TestCase;&#13;\nuse Digitec\\Patterns\\Decorator\\TitledWindow;&#13;\nuse ReflectionMethod;&#13;\n&#13;\nclass TitledWindowTest extends PHPUnit_Framework_TestCase &#13;\n{&#13;\n    public function testRender()&#13;\n    {&#13;\n        \/* Some test code here *\/&#13;\n    }&#13;\n    &#13;\n    public function testAddTitle()&#13;\n    {&#13;\n        $renderString = 'bar';&#13;\n        &#13;\n        $window = $this-&gt;getMock('Digitec\\Patterns\\Decorator\\SimpleWindow');&#13;\n        $window-&gt;expects($this-&gt;any())-&gt;method('render')-&gt;will($this-&gt;returnValue($renderString));&#13;\n        &#13;\n        $titledWindow = new TitledWindow($window);&#13;\n        &#13;\n        $reflectionMethod = new ReflectionMethod($titledWindow, 'addTitle');&#13;\n        $reflectionMethod-&gt;setAccessible(true);&#13;\n        &#13;\n        $rs = $reflectionMethod-&gt;invoke($titledWindow);&#13;\n        &#13;\n        \\Psy\\Shell::debug(get_defined_vars());&#13;\n        &#13;\n        $this-&gt;assertFalse(empty($rs));&#13;\n        &#13;\n    }&#13;\n}<\/pre><p>We are ready to rock! <\/p><pre class=\"language-\">$ phpunit --verbose --debug --filter=TitledWindowTest::testAddTitle&#13;\nPHPUnit 4.0.14 by Sebastian Bergmann.&#13;\n&#13;\nConfiguration read from \/home\/action\/workspace\/lab\/PsySH\/tests\/phpunit.xml&#13;\n&#13;\n&#13;\nStarting test 'AcmeTest\\Patterns\\Decorator\\TitledWindowTest::testAddTitle'.&#13;\nPsy Shell v0.1.11 (PHP 5.5.8 \u2014 cli) by Justin Hileman&#13;\n&gt;&gt;&gt;<\/pre><p>So in <code>$rs<\/code> we should have a string; let\u2019s see what we really have.<\/p><pre class=\"language-\">&gt;&gt;&gt; $rs&#13;\n=&gt; null<\/pre><p>Null value, no wonder the test failed\u2026 Let\u2019s check the source code of <code>TitledWindow::addTitle()<\/code>.  If we perform an <code>ls<\/code> command, we can see that we have that object\u2019s method available through the <code>$titledWindow<\/code> object.<\/p><pre class=\"language-\">&gt;&gt;&gt; show $titledWindow-&gt;addTitle&#13;\nprotected function addTitle()&#13;\n    protected function addTitle()&#13;\n    {&#13;\n        $returnString = &lt;&lt;&lt;EOL&#13;\n+-------------+&#13;\n|Title        |&#13;\nEOL;&#13;\n        echo $returnString . PHP_EOL;&#13;\n    }&#13;\n&gt;&gt;&gt;<\/pre><p>There is the bug. The method is echoing the value instead of returning it.  Even though the application seems to work right, through unit testing and PsySH we discovered a defect and we can now fix it.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>This article was not meant to be exhaustive in showcasing all the potential PsySH has.  There are other cool features (like \u2018doc\u2019) that you should try out.  PsySH alone may not be very useful, but if combined with other tools and your clever debugging powers, it can prove to be a valuable asset.<\/p>","created_at":"2018-11-24T02:08:23+0000","updated_at":"2018-11-24T02:08:33+0000","published_at":"2014-09-29T16:00:47+0000","annotations":[],"mimetype":"text\/html","reading_time":15,"domain_name":"www.sitepoint.com","preview_picture":"https:\/\/www.sitepoint.com\/wp-content\/themes\/sitepoint\/assets\/images\/icon.php.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php"],"is_public":false,"id":711,"title":"Inversion of Control &#8211; The Hollywood Principle","url":"https:\/\/www.sitepoint.com\/inversion-of-control-the-hollywood-principle\/","content":"<p>There\u2019s a consensus among programmers (myself included, so here\u2019s my own public <em>mea culpa<\/em>) that Inversion of Control (IoC) is nothing more than a synonym for plain old Dependency Injection (DI). There\u2019s a pretty intuitive reason that sustains this mindset: if the motivation behind DI is to promote the design of classes whose external collaborators are supplied by its surrounding context rather than inversely looking up them, the process can effectively been seen as a form of IoC.<\/p><p>But while the DI = IoC equation can be considered generally valid, the concept of inversion of control itself is actually much broader. In fact, it could be said that DI is a specific use case which exploits the benefits of IoC, but it\u2019s far from being the only one. This leads us back to the start; if DI is just a pattern that relies on the strengths of IoC, what\u2019s IoC really then? <\/p><p>Traditionally, application components have been designed to operate on and control the execution environment, an approach that delivers well to some extent. For instance, a logging module could be implemented to log data to a file, and how and when to log the data would be a entirely under control of the module. The log file (a part of the environment in this case) would be just an external, passive element with no influence on the way the module works. But let\u2019s say we need to extend the module\u2019s functionality and give it the ability for additionally logging data to a database, or eventually even via email. Upgrading the module to expose the extra functionality will make it grow in complexity, becoming more bloated as the logic required to attend to these additional duties is packaged behind the same API. The approach works, but won\u2019t scale at all.<\/p><div class=\"widget maestro maestro-content-type-html hide-for-tablet-SP hide-for-mobile-SP\" id=\"maestro-636\"><p>--ADVERTISEMENT--<\/p><\/div><p>This tangled situation can be sorted out in a fairly simple manner. Instead of making the module completely responsible for logging data to multiple endpoints, we can transfer the responsibility straight to the external environment. The module\u2019s implementation would remain ridiculously simple, limited to acting as a simple event dispatcher. On the flip side, the environment would be responsible for implementing all of the logic required to log data entirely independent from the module in question.<\/p><p>Not surprisingly, the process of inverting these responsibilities between components and the environment is formally known as Inversion of Control (or in a more relaxed jargon, <a href=\"http:\/\/en.wikipedia.org\/wiki\/Hollywood_principle\" title=\"Hollywood principle - Wikipedia, the free encyclopedia\" target=\"_blank\">The Hollywood Principle<\/a>), and its implementation can be a real boost when it comes to developing extensible, highly-decoupled program modules.<\/p><p>Of course, IoC is a language-agnostic paradigm, and as such it\u2019s possible to consume it in the PHP world without much fuss.         <\/p><h2>Achieving Inversion of Control \u2013 Observing Domain Objects<\/h2><p>IoC has indeed an ubiquitous presence, so it\u2019s pretty easy to find implementations of it production. The first use case that comes to mind is Dependency Injection, but there are many other cases equally demonstrative, especially when stepping on the terrain of Event-Driven Design. If you\u2019re wondering in what parallel universe IoC gets along with event handling mechanisms, consider a classic in the GoF repertoire: the Observer pattern. <\/p><p>Used nearly everywhere, even client-side via JavaScript, observers are a shining example of the IoC concept in action; there\u2019s a highly-decoupled subject focused on doing just a few narrow tasks without polluting the surrounding context while one or more external observers are responsible for implementing the logic required for handling the events triggered by the subject. How to handle the events, and even processing new ones, is entirely a responsibility of the observers rather than the subject.<\/p><p>An example might be a nice way to make my previous babbling a little bit clearer. So, let\u2019s say we\u2019ve implemented a primitive Domain Model which defines a one-to-many relationship between blog posts and comments. In this case we\u2019ll be deliberately ambitious and give the model the ability for firing off an email to notify the system administrator when a new comment is added to a post.<\/p><p>Honestly, implementing such a feature without appealing to IoC would literally be a tangled mess, as we\u2019d be asking the domain objects to doing something that\u2019s way outside their scope. Instead, we could take an IoC-based approach and define the domain classes as follows:<\/p><pre class=\"language-php\">&lt;?php&#13;\nnamespace Model;&#13;\n&#13;\ninterface PostInterface&#13;\n{&#13;\n    public function setTitle($title);&#13;\n    public function getTitle();&#13;\n    &#13;\n    public function setContent($content);&#13;\n    public function getContent();&#13;\n    &#13;\n    public function setComment(CommentInterface $comment);&#13;\n    public function getComments();&#13;\n}<\/pre><pre class=\"language-php\">&lt;?php&#13;\nnamespace Model;&#13;\n&#13;\nclass Post implements PostInterface, SplSubject&#13;\n{&#13;\n    private $title;&#13;\n    private $content;&#13;\n    private $comments  = [];&#13;\n    private $observers = [];&#13;\n&#13;\n    public function __construct($title, $content) {&#13;\n        $this-&gt;setTitle($title);&#13;\n        $this-&gt;setContent($content);&#13;\n    }&#13;\n    &#13;\n    public function setTitle($title) {&#13;\n        if (!is_string($title) &#13;\n            || strlen($title) &lt; 2&#13;\n            || strlen($title) &gt; 100) {&#13;\n            throw new InvalidArgumentException(&#13;\n                \"The post title is invalid.\");&#13;\n        }&#13;\n        $this-&gt;title = $title;&#13;\n        return $this;&#13;\n    }&#13;\n&#13;\n    public function getTitle() {&#13;\n        return $this-&gt;title;&#13;\n    }&#13;\n    &#13;\n    public function setContent($content) {&#13;\n        if (!is_string($content) || strlen($content) &lt; 10) {&#13;\n            throw new InvalidArgumentException(&#13;\n                \"The post content is invalid.\");&#13;\n        }&#13;\n        $this-&gt;content = $content;&#13;\n        return $this;&#13;\n    }&#13;\n    &#13;\n    public function getContent() {&#13;\n        return $this-&gt;content;&#13;\n    }&#13;\n&#13;\n    public function setComment(CommentInterface $comment) {&#13;\n        $this-&gt;comments[] = $comment;&#13;\n        $this-&gt;notify();&#13;\n    }&#13;\n    &#13;\n    public function getComments() {&#13;\n        return $this-&gt;comments;&#13;\n    }&#13;\n&#13;\n    public function attach(SplObserver $observer) { &#13;\n        $id = spl_object_hash($observer);&#13;\n        if (!isset($this-&gt;observers[$id])) {&#13;\n            $this-&gt;observers[$id] = $observer;&#13;\n        }&#13;\n        return $this;&#13;\n    }&#13;\n    &#13;\n    public function detach(SplObserver $observer) {&#13;\n        $id = spl_object_hash($observer);&#13;\n        if (!isset($this-&gt;observers[$id])) {&#13;\n            throw new RuntimeException(&#13;\n                \"Unable to detach the requested observer.\");&#13;\n        }&#13;\n        unset($this-&gt;observers[$id]); &#13;\n        return $this;&#13;\n    }&#13;\n    &#13;\n    public function notify() {&#13;\n        foreach ($this-&gt;observers as $observer) {    &#13;\n            $observer-&gt;update($this);&#13;\n        }&#13;\n    }&#13;\n}<\/pre><pre class=\"language-php\">&lt;?php&#13;\nnamespace Model;&#13;\n&#13;\ninterface CommentInterface&#13;\n{&#13;\n    public function setContent($content);&#13;\n    public function getContent();&#13;\n    &#13;\n    public function setAuthor($author);&#13;\n    public function getAuthor();&#13;\n}<\/pre><pre class=\"language-php\">&lt;?php&#13;\nnamespace Model;&#13;\n&#13;\nclass Comment implements CommentInterface&#13;\n{&#13;\n    private $content;&#13;\n    private $author;&#13;\n    &#13;\n    public function __construct($content, $author) {&#13;\n       $this-&gt;setContent($content);&#13;\n       $this-&gt;setAuthor($author);&#13;\n    }&#13;\n    &#13;\n    public function setContent($content) {&#13;\n        if (!is_string($content) || strlen($content) &lt; 10) {&#13;\n            throw new InvalidArgumentException(&#13;\n                \"The comment is invalid.\");&#13;\n        }&#13;\n        $this-&gt;content = $content;&#13;\n        return $this;&#13;\n    }&#13;\n    &#13;\n    public function getContent() {&#13;\n        return $this-&gt;content;&#13;\n    }&#13;\n    &#13;\n    public function setAuthor($author) {&#13;\n        if (!is_string($author) &#13;\n            || strlen($author) &lt; 2&#13;\n            || strlen($author) &gt; 50) {&#13;\n            throw new InvalidArgumentException(&#13;\n                \"The author is invalid.\");&#13;\n        }&#13;\n        $this-&gt;author = $author;&#13;\n        return $this;&#13;\n    }&#13;\n    &#13;\n    public function getAuthor() {&#13;\n        return $this-&gt;author;&#13;\n    }&#13;\n}<\/pre><p>The interaction between the <code>Post<\/code> and <code>Comment<\/code> classes is trivial, but the <code>Post<\/code> class deserves an in-depth look. Effectively, it has been designed as a \u201cclassic\u201d subject, hence providing the typical API which permits to attach\/detach and notify observers at will.<\/p><p>The most interesting facet of this process is the implementation of <code>setComment()<\/code> where the actual inversion of control takes place. The method just fires a \u201cpull\u201d update to all the registered observers whenever a comment is added. This means that all the logic required for sending out the email notification is delegated to one or more external observers, thus offloading the dirty work from <code>Post<\/code>, keeping it focused on just its own business logic.  <\/p><p>With this simple but effective schema of inversion of control in place, the only structure that needs to be added to the picture is at least one observer which should be responsible for dispatching the aforementioned email. To keep things easy to follow, I\u2019m going to implement the observer as a thin entity living and breathing in the service layer.<\/p><h2>Delegating Control to the External Environment \u2013 Implementing a Comment Notification Service<\/h2><p>Building an observer service capable of triggering an email notification when a new comment is added to a blog post is a simple process, reduced to defining a class that implements the pertaining <code>update()<\/code> method. If you\u2019re curious and want to see how the service in question looks, here it is:<\/p><pre class=\"language-php\">&lt;?php&#13;\nnamespace Service;&#13;\n&#13;\nclass CommentService implements SplObserver&#13;\n{&#13;\n    public function update(SplSubject $post) {&#13;\n        $subject = \"New comment posted!\";&#13;\n        $message = \"A comment has been made on a post entitled \" .&#13;\n            $post-&gt;getTitle();&#13;\n        $headers = \"From: \"Notification System\" &lt;notify@example.com&gt;rnMIME-Version: 1.0rn\";&#13;\n        if (!@mail(\"admin@example.com\", $subject, $message, $headers)) {&#13;\n            throw new RuntimeException(\"Unable to send the update.\");&#13;\n        }&#13;\n    }&#13;\n}<\/pre><p>The <code>CommentService<\/code> class does exactly what it\u2019s supposed to; it invokes its <code>update()<\/code> method to dispatch an email to the sysadmin each time a user drops a comment related to a given post.<\/p><p>It\u2019d be a lot easier to see the benefits brought by inversion of control in this situation if I showed you a script that puts all the sample classes to work, so below is some code:<\/p><pre class=\"language-php\">&lt;?php&#13;\nuse LibraryLoaderAutoloader,&#13;\n    ModelPost,&#13;\n    ModelComment,&#13;\n    ServiceCommentService;&#13;\n         &#13;\nrequire_once __DIR__ . \"\/Library\/Loader\/Autoloader.php\";&#13;\n$autoloader = new Autoloader();&#13;\n$autoloader-&gt;register();&#13;\n&#13;\n$post = new Post(&#13;\n    \"A sample post\",&#13;\n    \"This is the content of the sample post\"&#13;\n);&#13;\n&#13;\n$post-&gt;attach(new CommentService());&#13;\n&#13;\n$comment = new Comment(&#13;\n    \"A sample comment\",&#13;\n    \"Just commenting on the previous post\"&#13;\n);&#13;\n&#13;\n$post-&gt;setComment($comment);<\/pre><p>Quite possibly, and this is just my personal preference, I don\u2019t feel quite right injecting a service into the internals of a domain object. The powers that be have always proclaimed the domain layer must be agnostic about the service layer (unless you appeal to separated interfaces), and this one should lay down on an upper level interoperating with multiple clients. But in this case the approach isn\u2019t really that sinful considering the <code>Post<\/code> class is just a dummy container for the registered observers which are only consumed when triggering event updates. Moreover, taking into account how neatly the responsibilities between the service in question and the Post class have been inverted, my complaint should be considered pretty much a picky whim.<\/p><h2>Closing Thoughts<\/h2><p>Quite often considered an obscure, tangled concept, especially in PHP where many developers tend to intuitively associate the concept only with plain old Dependency Injection, Inversion of Control is a simple yet killer programming methodology which, when properly implemented, is a fantastic way for creating decoupled, orthogonal systems whose components can be easily tested in isolation.<\/p><p>If you\u2019re using Dependency Injection within your applications (you are, right?) then you should feel your coder\u2019s instincts pretty well satisfied as you\u2019re already exploiting the benefits that Inversion of Control provides. As I attempted to demonstrate before, however, there\u2019s a wide array of situations where the approach fits well other than just managing class dependencies the right way. Event-Driven Design is certainly a good example.<\/p><p><small>Image via <a title=\"Royalty Free Stock Photos at Fotolia.com\" href=\"http:\/\/us.fotolia.com\/?utm_source=sitepoint&amp;utm_medium=website_link&amp;utm=campaign=sitepoint\" target=\"_blank\">Fotolia<\/a><\/small><\/p><div class=\"Article_authorBio l-mv4 t-bg-white m-border l-pa3\"><div class=\"l-d-f l-pt3\"><a class=\"l-s5 l-pa0 l-mr3 l-round l-o-h\" href=\"https:\/\/www.sitepoint.com\/author\/agervasio\/\">\n            <img src=\"https:\/\/secure.gravatar.com\/avatar\/0140403b7f4a8b362e8d8d8ad6443960?s=96&amp;d=mm&amp;r=g\" width=\"96\" height=\"96\" alt=\"\" class=\"avatar avatar-96 wp-user-avatar wp-user-avatar-96 photo avatar-default\" \/><\/a><\/div><p>Alejandro Gervasio is a senior System Analyst from Argentina who has been involved in software development since the mid-80's. He has more than 12 years of experience in PHP development, 10 years in Java Programming, Object-Oriented Design, and most of the client-side technologies available out there.<\/p><\/div>","created_at":"2018-11-24T01:54:39+0000","updated_at":"2018-11-24T01:54:40+0000","published_at":"2012-12-07T14:30:01+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"www.sitepoint.com","preview_picture":"https:\/\/www.sitepoint.com\/wp-content\/themes\/sitepoint\/assets\/images\/icon.php.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":710,"title":"You should\u2019ve asked","url":"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/","content":"<p>Here is the english version of my now famous \u201cFallait demander\u201d ; soon to come as a book with other stories :<\/p>\n<p><img data-attachment-id=\"318\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/mentalload_cover\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg\" data-orig-size=\"1500,1756\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;1&quot;}\" data-image-title=\"MentalLoad_cover\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=256\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=656\" class=\"alignnone size-full wp-image-318\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=430\" alt=\"MentalLoad_cover.jpg\" width=\"430\" height=\"503\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=430 430w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=860 860w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=128 128w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=256 256w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/mentalload_cover.jpg?w=768 768w\" \/><\/p>\n<p>Preorders available <a href=\"https:\/\/www.barnesandnoble.com\/w\/the-mental-load-emma\/1128860695?ean=9781609809188\">here<\/a> or <a href=\"https:\/\/www.indiebound.org\/book\/9781609809188\">here <\/a>or <a href=\"https:\/\/www.amazon.co.uk\/Mental-Load-Feminist-Comic-ebook\/dp\/B07BD25KZD\">here <\/a>^_^<\/p>\n<div class=\"embed-amazon\"><iframe width=\"656\" height=\"550\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" src=\"https:\/\/read.amazon.co.uk\/kp\/card?preview=inline&amp;linkCode=kpd&amp;ref_=k4w_oembed_kW0aB6QB65fIW7&amp;asin=B07DWVYZXR&amp;tag=kpembed-20\">[embedded content]<\/iframe><\/div>\n<p>Thanks Una from\u00a0<a href=\"http:\/\/unadtranslation.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">unadtranslation.com<\/a>\u00a0for the translation \ud83d\ude42<\/p>\n<p><img data-attachment-id=\"42\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_001\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_001\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=656?w=656\" class=\"alignnone size-full wp-image-42\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=656\" alt=\"You should've asked_001\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_001.png?w=1024 1024w\" \/><img data-attachment-id=\"43\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_002\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_002\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=656?w=656\" class=\"alignnone size-full wp-image-43\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=656\" alt=\"You should've asked_002\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_002.png?w=1024 1024w\" \/><img data-attachment-id=\"44\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_003\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_003\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=656?w=656\" class=\"alignnone size-full wp-image-44\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=656\" alt=\"You should've asked_003\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_003.png?w=1024 1024w\" \/><img data-attachment-id=\"45\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_004\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_004\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=656?w=656\" class=\"alignnone size-full wp-image-45\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=656\" alt=\"You should've asked_004\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_004.png?w=1024 1024w\" \/><img data-attachment-id=\"46\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_005\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_005\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=656?w=656\" class=\"alignnone size-full wp-image-46\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=656\" alt=\"You should've asked_005\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_005.png?w=1024 1024w\" \/><img data-attachment-id=\"47\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_006\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_006\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=656?w=656\" class=\"alignnone size-full wp-image-47\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=656\" alt=\"You should've asked_006\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_006.png?w=1024 1024w\" \/><img data-attachment-id=\"48\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_007\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_007\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=656?w=656\" class=\"alignnone size-full wp-image-48\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=656\" alt=\"You should've asked_007\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_007.png?w=1024 1024w\" \/><img data-attachment-id=\"49\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_008\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_008\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=656?w=656\" class=\"alignnone size-full wp-image-49\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=656\" alt=\"You should've asked_008\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_008.png?w=1024 1024w\" \/><img data-attachment-id=\"50\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_009\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_009\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=656?w=656\" class=\"alignnone size-full wp-image-50\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=656\" alt=\"You should've asked_009\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_009.png?w=1024 1024w\" \/><img data-attachment-id=\"51\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_010\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_010\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=656?w=656\" class=\"alignnone size-full wp-image-51\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=656\" alt=\"You should've asked_010\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_010.png?w=1024 1024w\" \/><\/p>\n<p><img data-attachment-id=\"90\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_011-2\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_011\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=656?w=656\" class=\"alignnone size-full wp-image-90\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=656\" alt=\"You should've asked_011\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0111.png?w=1024 1024w\" \/><\/p>\n<p><img data-attachment-id=\"53\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_012\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_012\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=656?w=656\" class=\"alignnone size-full wp-image-53\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=656\" alt=\"You should've asked_012\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_012.png?w=1024 1024w\" \/><img data-attachment-id=\"54\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_013\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_013\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=656?w=656\" class=\"alignnone size-full wp-image-54\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=656\" alt=\"You should've asked_013\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_013.png?w=1024 1024w\" \/><img data-attachment-id=\"55\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_014\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_014\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=656?w=656\" class=\"alignnone size-full wp-image-55\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=656\" alt=\"You should've asked_014\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_014.png?w=1024 1024w\" \/><\/p>\n<p><img data-attachment-id=\"96\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_015-3\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=656&amp;h=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_015\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=656&amp;h=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=656&amp;h=656?w=656\" class=\"alignnone size-full wp-image-96\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=656&amp;h=656\" alt=\"\" width=\"656\" height=\"656\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=656&amp;h=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=1312&amp;h=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=150&amp;h=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=300&amp;h=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=768&amp;h=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_0152.png?w=1024&amp;h=1024 1024w\" \/><\/p>\n<p><img data-attachment-id=\"57\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_016\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_016\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=656?w=656\" class=\"alignnone size-full wp-image-57\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=656\" alt=\"You should've asked_016\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_016.png?w=1024 1024w\" \/><img data-attachment-id=\"58\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_017\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_017\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=656?w=656\" class=\"alignnone size-full wp-image-58\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=656\" alt=\"You should've asked_017\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_017.png?w=1024 1024w\" \/><img data-attachment-id=\"59\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_018\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_018\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=656?w=656\" class=\"alignnone size-full wp-image-59\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=656\" alt=\"You should've asked_018\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_018.png?w=1024 1024w\" \/><img data-attachment-id=\"60\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_019\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_019\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=656?w=656\" class=\"alignnone size-full wp-image-60\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=656\" alt=\"You should've asked_019\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_019.png?w=1024 1024w\" \/><img data-attachment-id=\"61\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_020\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_020\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=656?w=656\" class=\"alignnone size-full wp-image-61\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=656\" alt=\"You should've asked_020\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_020.png?w=1024 1024w\" \/><img data-attachment-id=\"62\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_021\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_021\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=656?w=656\" class=\"alignnone size-full wp-image-62\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=656\" alt=\"You should've asked_021\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_021.png?w=1024 1024w\" \/><img data-attachment-id=\"63\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_022\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_022\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=656?w=656\" class=\"alignnone size-full wp-image-63\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=656\" alt=\"You should've asked_022\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_022.png?w=1024 1024w\" \/><img data-attachment-id=\"64\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_023\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_023\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=656?w=656\" class=\"alignnone size-full wp-image-64\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=656\" alt=\"You should've asked_023\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_023.png?w=1024 1024w\" \/><img data-attachment-id=\"65\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_024\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_024\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=656?w=656\" class=\"alignnone size-full wp-image-65\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=656\" alt=\"You should've asked_024\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_024.png?w=1024 1024w\" \/><img data-attachment-id=\"66\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_025\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_025\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=656?w=656\" class=\"alignnone size-full wp-image-66\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=656\" alt=\"You should've asked_025\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_025.png?w=1024 1024w\" \/><img data-attachment-id=\"67\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_026\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_026\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=656?w=656\" class=\"alignnone size-full wp-image-67\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=656\" alt=\"You should've asked_026\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_026.png?w=1024 1024w\" \/><img data-attachment-id=\"68\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_027\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_027\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=656?w=656\" class=\"alignnone size-full wp-image-68\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=656\" alt=\"You should've asked_027\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_027.png?w=1024 1024w\" \/><img data-attachment-id=\"69\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_028\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_028\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=656?w=656\" class=\"alignnone size-full wp-image-69\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=656\" alt=\"You should've asked_028\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_028.png?w=1024 1024w\" \/><img data-attachment-id=\"70\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_029\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_029\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=656?w=656\" class=\"alignnone size-full wp-image-70\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=656\" alt=\"You should've asked_029\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_029.png?w=1024 1024w\" \/><img data-attachment-id=\"71\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_030\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_030\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=656?w=656\" class=\"alignnone size-full wp-image-71\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=656\" alt=\"You should've asked_030\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_030.png?w=1024 1024w\" \/><img data-attachment-id=\"72\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_031\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_031\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=656?w=656\" class=\"alignnone size-full wp-image-72\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=656\" alt=\"You should've asked_031\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_031.png?w=1024 1024w\" \/><img data-attachment-id=\"73\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_032\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_032\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=656?w=656\" class=\"alignnone size-full wp-image-73\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=656\" alt=\"You should've asked_032\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_032.png?w=1024 1024w\" \/><img data-attachment-id=\"74\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_033\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_033\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=656?w=656\" class=\"alignnone size-full wp-image-74\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=656\" alt=\"You should've asked_033\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_033.png?w=1024 1024w\" \/><img data-attachment-id=\"75\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_034\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_034\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=656?w=656\" class=\"alignnone size-full wp-image-75\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=656\" alt=\"You should've asked_034\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_034.png?w=1024 1024w\" \/><img data-attachment-id=\"76\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_035\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_035\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=656?w=656\" class=\"alignnone size-full wp-image-76\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=656\" alt=\"You should've asked_035\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_035.png?w=1024 1024w\" \/><img data-attachment-id=\"77\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_036\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_036\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=656?w=656\" class=\"alignnone size-full wp-image-77\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=656\" alt=\"You should've asked_036\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_036.png?w=1024 1024w\" \/><img data-attachment-id=\"78\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_037\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_037\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=656?w=656\" class=\"alignnone size-full wp-image-78\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=656\" alt=\"You should've asked_037\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_037.png?w=1024 1024w\" \/><img data-attachment-id=\"79\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_038\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_038\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=656?w=656\" class=\"alignnone size-full wp-image-79\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=656\" alt=\"You should've asked_038\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_038.png?w=1024 1024w\" \/><img data-attachment-id=\"80\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_039\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_039\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=656?w=656\" class=\"alignnone size-full wp-image-80\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=656\" alt=\"You should've asked_039\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_039.png?w=1024 1024w\" \/><img data-attachment-id=\"81\" data-permalink=\"https:\/\/english.emmaclit.com\/2017\/05\/20\/you-shouldve-asked\/you-shouldve-asked_040\/\" data-orig-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=656\" data-orig-size=\"2000,2000\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"You should\u2019ve asked_040\" data-image-description=\"\" data-medium-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=656?w=300\" data-large-file=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=656?w=656\" class=\"alignnone size-full wp-image-81\" src=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=656\" alt=\"You should've asked_040\" srcset=\"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=656 656w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=1312 1312w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=150 150w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=300 300w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=768 768w, https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png?w=1024 1024w\" \/><\/p>","created_at":"2018-11-23T17:28:37+0000","updated_at":"2018-11-23T17:28:37+0000","published_at":"2017-05-20T09:39:10+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"english.emmaclit.com","preview_picture":"https:\/\/emmacliten.files.wordpress.com\/2017\/05\/you-shouldve-asked_040.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":709,"title":"8BitDo","url":"http:\/\/www.8bitdo.com\/","content":"<a href=\"http:\/\/www.8bitdo.com\/\" class=\"bottomDiv\"><b><i class=\"fa fa-angle-down\"><\/i><\/b><\/a><a href=\"http:\/\/www.8bitdo.com\/\" class=\"topDiv\">\n<i class=\"fa fa-angle-up\"><\/i><\/a>\n<div id=\"bootstrap-touch-slider\" class=\"carousel bs-slider fade  control-round indicators-line marTop\" data-ride=\"carousel\" data-pause=\"hover\" data-interval=\"7000\"><div class=\"carousel-inner\" role=\"listbox\"><div class=\"item active\"><img srcset=\"images\/hot\/gbros-02-L.jpg\" alt=\"8BitDo\" src=\"src\" \/><a href=\"http:\/\/www.8bitdo.com\/gbros\">\n      <div class=\"bs-slider-overlay\"><div class=\"container mar_1\"><div class=\"row\"><div class=\"slide-text gbros mar slide1\"><p>Play your Switch with your original wired controllers like Gamecube, NES, SNES, SFC Classic Edition and Wii Classic. <\/p><\/div><\/div><\/div><\/div>\n      <\/a><\/div><div class=\"item\"><img srcset=\"images\/hot\/n30pro2-02-L.jpg\" alt=\"8BitDo\" src=\"src\" \/><a href=\"http:\/\/www.8bitdo.com\/n30pro-2\">\n      <div class=\"bs-slider-overlay\"><div class=\"container mar_1\"><div class=\"row\"><div class=\"slide-text n30pro2 mar slide1\"><p>Now meet N30 Pro 2.N30 Pro 2 refines every element of its predecessor with new features like motion controls, vibration, USB-C and turbo functionality. <\/p><\/div><\/div><\/div><\/div>\n      <\/a><\/div><div class=\"item\"><img srcset=\"images\/hot\/sn30pro-l-1.jpg\" alt=\"8BitDo\" src=\"src\" \/><a class=\"pcmag\" href=\"https:\/\/www.pcmag.com\/review\/364767\/8bitdo-sn30-pro\" target=\"_blank\"> <img src=\"http:\/\/www.8bitdo.com\/images\/NES30-Arcade-Stick\/EC_logo_lg.png\" alt=\"image\" \/><\/a><a href=\"http:\/\/www.8bitdo.com\/sn30-pro-g-classic-or-sn30-pro-sn\/\">\n      <div class=\"bs-slider-overlay\"><div class=\"container mar_1\"><div class=\"row\"><div class=\"slide-text hotbox-5 mar slide1\"><p>We set out to design a fully featured retro controller to play 30 years of video games. With inspiration from the classics. Designed with the utmost attention to detail. We paid extra attention to the most critical characteristics like the d-pad, to make sure it feels exactly like you remember it.<\/p><\/div><\/div><\/div><\/div>\n      <\/a><\/div><div class=\"item\"><img srcset=\"images\/hot\/sn30-l-1.jpg\" alt=\"8BitDo\" src=\"src\" \/><a href=\"http:\/\/www.8bitdo.com\/sn30-gp\/\">\n      \n      <\/a><\/div><div class=\"item\"><img srcset=\"images\/hot\/n30-2.4g-l.jpg\" alt=\"8BitDo\" src=\"src\" \/><a href=\"http:\/\/www.8bitdo.com\/n30-2-4g\/\">\n      <div class=\"bs-slider-overlay\"><div class=\"container mar_2\"><div class=\"row\"><div class=\"slide-text hotbox-1 mar\"><p>Play your NES Classic Wirelessly<\/p><\/div><\/div><\/div><\/div>\n      <\/a><\/div><div class=\"item\"><img srcset=\"images\/hot\/usb-rr-xbox-l.jpg\" alt=\"8BitDo\" src=\"src\" \/><a href=\"http:\/\/support.8bitdo.com\">\n      \n      <\/a><\/div><\/div><a class=\"left carousel-control\" href=\"#bootstrap-touch-slider\" role=\"button\" data-slide=\"prev\">  Previous <\/a><a class=\"right carousel-control\" href=\"#bootstrap-touch-slider\" role=\"button\" data-slide=\"next\">  Next <\/a><\/div>\n \n\n<section id=\"Products\" class=\"content-block team-2 team-2-2 product-list\"><\/section><section class=\"content-block-nopad footer-wrap-1-1 bg-deepocean\"><div class=\"container footer-1-1\"><div class=\"row\"><div class=\"col-sm-6 col-sm-offset-1\"><div class=\"row\"><div class=\"col-sm-6 col-md-6 col-lg-6\"><h4>Company<\/h4><ul><li> <a href=\"https:\/\/www.dropbox.com\/sh\/bt8yi5yz4gqoua1\/AABy11laLxX_PGEfto1xXf7da?dl=0\" target=\"_blank\">Press Kit <\/a> <\/li>\n              <li><\/li> \n            <\/ul><\/div><div class=\"col-sm-6 col-md-6 col-lg-6\"><h4>About<\/h4><ul><li>\n                <a href=\"http:\/\/www.8bitdo.com\/e3-2018\/\">E3 2018<\/a>\n              <\/li>\n              <li>\n                <\/li>\n              \n            <\/ul><\/div><\/div><\/div><\/div><\/div><\/section>","created_at":"2018-11-23T13:41:47+0000","updated_at":"2018-11-23T13:41:47+0000","annotations":[],"mimetype":"text\/html","language":"zh_CN","reading_time":0,"domain_name":"www.8bitdo.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":708,"title":"briarproject.org","url":"https:\/\/briarproject.org\/","created_at":"2018-11-23T12:30:19+0000","updated_at":"2018-11-23T12:30:19+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"briarproject.org"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":707,"title":"A fullscreen desktop application resembling a sci-fi computer interface","url":"https:\/\/news.ycombinator.com\/item?id=18509828","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D18509828\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"18509828\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/github.com\/GitSquared\/edex-ui\" class=\"storylink\">A fullscreen desktop application resembling a sci-fi computer interface<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=github.com\">github.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        255 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=nailer\" class=\"hnuser\">nailer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18509828\">4 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=18509828&amp;goto=item%3Fid%3D18509828\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=A%20fullscreen%20desktop%20application%20resembling%20a%20sci-fi%20computer%20interface&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=A%20fullscreen%20desktop%20application%20resembling%20a%20sci-fi%20computer%20interface\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=18509828&amp;auth=c89471699bc1801a51098dd5f6d26b9c5cd291b8\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=18509828\">60\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-22T17:57:11+0000","updated_at":"2018-11-22T17:57:11+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":706,"title":"browsh-org\/browsh","url":"https:\/\/github.com\/browsh-org\/browsh","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/browsh-org\/browsh\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/16c6da868e3f1f1a97987311b398f7f2aaf20ad8\/68747470733a2f2f7472617669732d63692e6f72672f62726f7773682d6f72672f62726f7773682e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/browsh-org\/browsh.svg?branch=master\" \/><\/a> <a href=\"https:\/\/twitter.com\/intent\/follow?screen_name=brow_sh\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/045ba45671a7043bda514a23c13146e443ebf4da\/68747470733a2f2f696d672e736869656c64732e696f2f747769747465722f666f6c6c6f772f62726f775f73682e7376673f7374796c653d736f6369616c266c6162656c3d466f6c6c6f77\" alt=\"Follow @brow_sh\" data-canonical-src=\"https:\/\/img.shields.io\/twitter\/follow\/brow_sh.svg?style=social&amp;label=Follow\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/49969623f58348592a23738746e6f34f7c63a412\/68747470733a2f2f7777772e62726f772e73682f6173736574732f696d616765732f62726f7773682d6865616465722e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/49969623f58348592a23738746e6f34f7c63a412\/68747470733a2f2f7777772e62726f772e73682f6173736574732f696d616765732f62726f7773682d6865616465722e6a7067\" alt=\"Browsh Logo\" data-canonical-src=\"https:\/\/www.brow.sh\/assets\/images\/browsh-header.jpg\" \/><\/a><\/p>\n<p><strong>A fully interactive, realtime, and modern text-based browser rendered to TTYs and browsers<\/strong><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/1091b6805623a9dbe3b2b06b60e81e876454feb3\/68747470733a2f2f6d656469612e67697068792e636f6d2f6d656469612f6262736d566b596a50644f4b48684d584f4f2f67697068792e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/1091b6805623a9dbe3b2b06b60e81e876454feb3\/68747470733a2f2f6d656469612e67697068792e636f6d2f6d656469612f6262736d566b596a50644f4b48684d584f4f2f67697068792e676966\" alt=\"Browsh Gif\" data-canonical-src=\"https:\/\/media.giphy.com\/media\/bbsmVkYjPdOKHhMXOO\/giphy.gif\" \/><\/a><\/p>\n<h2><a id=\"user-content-why-use-browsh\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-use-browsh\"><\/a>Why use Browsh?<\/h2>\n<p>Not all the world has good Internet.<\/p>\n<p>If you only have a 3kbps internet connection tethered from a phone,\nthen it's good to SSH into a server and browse the web through, say,\n<a href=\"https:\/\/github.com\/browsh-org\/browsh\/issues\/17\">elinks<\/a>. That way the\n<em>server<\/em> downloads the web pages and uses the limited bandwidth of an\nSSH connection to display the result. However, traditional text-based browsers\nlack JS and all other modern HTML5 support. Browsh is different\nin that it's backed by a real browser, namely headless Firefox,\nto create a purely text-based version of web pages and web apps. These can be easily\nrendered in a terminal or indeed, ironically, in another browser. Do note that currently the browser client doesn't have feature parity with the terminal client.<\/p>\n<p>Why not VNC? Well VNC is certainly one solution but it doesn't quite\nhave the same ability to deal with extremely bad Internet. Terminal\nBrowsh can also use MoSH to further reduce bandwidth and increase stability\nof the connection. Mosh offers features like automatic\nreconnection of dropped or roamed connections and diff-only screen updates.\nFurthermore, other than SSH or MoSH, terminal Browsh doesn't require a client\nlike VNC.<\/p>\n<p>One final reason to use terminal Browsh could be to offload the battery-drain of a modern\nbrowser from your laptop or low-powered device like a Raspberry Pi. If you're a CLI-native,\nthen you could potentially get a few more hours of life if your CPU-hungry browser\nis running somewhere else on mains electricity.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>Download a binary from the <a href=\"https:\/\/github.com\/browsh-org\/browsh\/releases\">releases<\/a> (~7MB).\nYou will need to have <a href=\"https:\/\/www.mozilla.org\/en-US\/firefox\/new\/\" rel=\"nofollow\">Firefox 63<\/a> (or higher) aleady installed.<\/p>\n<p>Or download and run the Docker image (~230MB) with:\n<code>docker run --rm -it browsh\/browsh<\/code><\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>Most keys and mouse gestures should work as you'd expect on a desktop\nbrowser.<\/p>\n<p>For full documentation click <a href=\"https:\/\/www.brow.sh\/docs\/introduction\/\" rel=\"nofollow\">here<\/a>.<\/p>\n<h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>To setup a dev env you will need NodeJS and Golang installed. If you get stuck\nsetting up your env, take a look in <code>.travis.yml<\/code>, it has to setup everything\nfrom scratch for every push to Github.<\/p>\n<p>I'd recommend <a href=\"https:\/\/github.com\/creationix\/nvm\">nvm<\/a> for NodeJS - note that\n<code>nvm install<\/code> will automatically parse the <code>.nvmrc<\/code> version in this repo to get\nthe correct NodeJS version. For Golang it's probably best to just use your OS's\npackage manager. The current Golang version being used is stored in <code>.travis.yml<\/code>.<\/p>\n<p>You'll then need to install the project dependencies. For the webextension, just\nrun: <code>npm install<\/code> inside the <code>webext\/<\/code> folder. For the CLI client you will first\nneed to install <code>dep<\/code>, there is a script for this in <code>interfacer\/contrib\/setup_dep.sh<\/code>.\nI don't fully understand Golang's best practices, but it seems you are forced to\nkeep your Go project's code under <code>$GOPATH\/src<\/code>, you might be able to get away\nwith symlinks. Anyway, to install the dependencies use: <code>dep ensure<\/code> inside the\n<code>interfacer\/<\/code> folder.<\/p>\n<p>Then the ideal setup for development is:<\/p>\n<ul><li>have Webpack watch the JS code so that it rebuilds automatically:\n<code>webpack --watch<\/code><\/li>\n<li>run the CLI client without giving it the responsibility to launch Firefox:\n<code>go run .\/interfacer\/src\/main.go --firefox.use-existing --debug<\/code><\/li>\n<li>have Mozilla's handy <code>web-ext<\/code> tool run Firefox and reinstall the\nwebextension everytime webpack rebuilds it: (in <code>webext\/dist<\/code>)\n<code>web-ext run --verbose<\/code><\/li>\n<\/ul><p>For generic Linux systems you can follow <a href=\"https:\/\/github.com\/browsh-org\/browsh\/blob\/master\/contrib\/setup_linux_build_environment.md\">this guide<\/a> on how to setup a build environment, that you may be able to adapt for other systems as well.<\/p>\n<p>Windows users can follow <a href=\"https:\/\/github.com\/browsh-org\/browsh\/blob\/master\/contrib\/setup_windows_build_environment.md\">this guide<\/a> in order to set up a build environment.<\/p>\n<p>Mac users may follow <a href=\"https:\/\/github.com\/browsh-org\/browsh\/blob\/master\/contrib\/setup_mac_build_environment.md\">this guide<\/a> that goes through the steps of setting up a build environment.<\/p>\n<h3><a id=\"user-content-communication\" class=\"anchor\" aria-hidden=\"true\" href=\"#communication\"><\/a>Communication<\/h3>\n<p>Questions about Brow.sh? Stuck trying to resolve a tricky issue? Connect with the Brow.sh community on <a href=\"https:\/\/gitter.im\/browsh\" rel=\"nofollow\">Gitter<\/a>!<\/p>\n<h2><a id=\"user-content-building-a-browsh-release\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-a-browsh-release\"><\/a>Building a Browsh release<\/h2>\n<p>If you'd like to build Browsh for a new package manager, or for any other reason,\nyou can use the script at <code>interfacer\/contrib\/build_browsh.go<\/code> as a guide. Note\nyou won't be able to build the web extension as Mozilla only allows one canonical\nversion of web extensions per version number. So the build script downloads the\nofficial web extension <code>.xpi<\/code> file from the Mozilla archives.<\/p>\n<h2><a id=\"user-content-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#tests\"><\/a>Tests<\/h2>\n<p>For the webextension: in <code>webext\/<\/code> folder, <code>npm test<\/code><br \/>For CLI unit tests: in <code>\/interfacer<\/code> run <code>go test src\/browsh\/*.go<\/code><br \/>For CLI E2E tests: in <code>\/interfacer<\/code> run <code>go test test\/tty\/*.go<\/code><br \/>For HTTP Service tests: in <code>\/interfacer<\/code> run <code>go test test\/http-server\/*.go<\/code><\/p>\n<h2><a id=\"user-content-special-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#special-thanks\"><\/a>Special Thanks<\/h2>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>GNU Lesser General Public License v2.1<\/p>\n<\/article>","created_at":"2018-11-22T17:56:17+0000","updated_at":"2018-11-22T17:56:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/35799511?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":705,"title":"GitSquared\/edex-ui","url":"https:\/\/github.com\/GitSquared\/edex-ui","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p style=\"text-align: center;\">\n  <br \/><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/logo.png\"><img alt=\"Logo\" src=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/logo.png\" \/><\/a>\n  <\/p>\n<table><thead><tr><th>Download<\/th>\n<th>Linux &amp; macOS<\/th>\n<th>Windows<\/th>\n<th>Dependencies<\/th>\n<th>Build system<\/th>\n<th>Source<\/th>\n<\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/github.com\/GitSquared\/edex-ui\/releases\"><img src=\"https:\/\/camo.githubusercontent.com\/0d613849411c69e31ec2b8379d5a9c7adcb52111\/68747470733a2f2f62616467656e2e6e65742f6769746875622f72656c656173652f476974537175617265642f656465782d7569\" alt=\"Release Badge\" data-canonical-src=\"https:\/\/badgen.net\/github\/release\/GitSquared\/edex-ui\" \/><\/a><\/td>\n<td><a href=\"https:\/\/travis-ci.org\/GitSquared\/edex-ui\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/54cfccc1665b27abf42f012fd30224c5858d1bf2\/68747470733a2f2f7472617669732d63692e6f72672f476974537175617265642f656465782d75692e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/GitSquared\/edex-ui.svg?branch=master\" \/><\/a><\/td>\n<td><a href=\"https:\/\/ci.appveyor.com\/project\/GitSquared\/edex-ui\/branch\/master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/de82c28b27e35d126324312ed6a4d68c64f65f31\/68747470733a2f2f63692e6170707665796f722e636f6d2f6170692f70726f6a656374732f7374617475732f6c656230363962726f3867776f6375372f6272616e63682f6d61737465723f7376673d74727565\" alt=\"Build status\" data-canonical-src=\"https:\/\/ci.appveyor.com\/api\/projects\/status\/leb069bro8gwocu7\/branch\/master?svg=true\" \/><\/a><\/td>\n<td><a href=\"https:\/\/dependabot.com\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/6d859eb98b7901e7f719addc5a472a90f2e1595d\/68747470733a2f2f6170692e646570656e6461626f742e636f6d2f6261646765732f7374617475733f686f73743d676974687562267265706f3d476974537175617265642f656465782d7569\" alt=\"Dependabot Status\" data-canonical-src=\"https:\/\/api.dependabot.com\/badges\/status?host=github&amp;repo=GitSquared\/edex-ui\" \/><\/a><\/td>\n<td><a href=\"https:\/\/snyk.io\/test\/github\/GitSquared\/edex-ui?targetFile=package.json\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/2f3407887d2dca5c12e6f2f4676d38089ed5741a\/68747470733a2f2f736e796b2e696f2f746573742f6769746875622f476974537175617265642f656465782d75692f62616467652e7376673f74617267657446696c653d7061636b6167652e6a736f6e\" alt=\"Known Vulnerabilities\" data-canonical-src=\"https:\/\/snyk.io\/test\/github\/GitSquared\/edex-ui\/badge.svg?targetFile=package.json\" \/><\/a><\/td>\n<td><a href=\"https:\/\/snyk.io\/test\/github\/GitSquared\/edex-ui?targetFile=src%2Fpackage.json\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/15bfa8b78236d7bef18666eabe36e5ba4b45541b\/68747470733a2f2f736e796b2e696f2f746573742f6769746875622f476974537175617265642f656465782d75692f62616467652e7376673f74617267657446696c653d7372632532467061636b6167652e6a736f6e\" alt=\"Known Vulnerabilities\" data-canonical-src=\"https:\/\/snyk.io\/test\/github\/GitSquared\/edex-ui\/badge.svg?targetFile=src%2Fpackage.json\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><p>eDEX-UI is a fullscreen desktop application resembling a sci-fi computer interface, heavily inspired from <a href=\"https:\/\/github.com\/seenaburns\/dex-ui\">DEX-UI<\/a> and the <a href=\"https:\/\/web.archive.org\/web\/20170511000410\/http:\/\/jtnimoy.com\/blogs\/projects\/14881671\" rel=\"nofollow\">TRON Legacy movie effects<\/a>. It runs the shell of your choice in a real terminal, and displays live information about your system. It was made to be used on large touchscreens but will work nicely on a regular desktop computer or perhaps a tablet PC or one of those funky 360\u00b0 laptops with touchscreens.<\/p>\n<p>I had no ideas for a name so i took DEX-UI and added a \"e\" for Electron. Deal with it.<\/p>\n<p>(<a href=\"https:\/\/aur.archlinux.org\/packages\/edex-ui\/\" rel=\"nofollow\">Official ArchLinux AUR package<\/a> maintained by <a href=\"https:\/\/github.com\/JesusCrie\">@JesusCrie<\/a>)<\/p>\n<h2><a id=\"user-content-screenshots\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshots\"><\/a>Screenshots<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/screenshot_default.png\"><img src=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/screenshot_default.png\" alt=\"Default screenshot\" \/><\/a><\/p>\n<p><em>(<a href=\"https:\/\/github.com\/dylanaraps\/neofetch\">neofetch<\/a> on eDEX-UI 1.0 with the default \"tron\" theme &amp; QWERTY keyboard)<\/em><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/screenshot_blade.png\"><img src=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/screenshot_blade.png\" alt=\"Blade screenshot\" \/><\/a><\/p>\n<p><em>(<a href=\"https:\/\/github.com\/browsh-org\/browsh\">browsh<\/a> showing the results of a DuckDuckGo image search on eDEX-UI 1.0 with the \"blade\" theme &amp; AZERTY keyboard)<\/em><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/screenshot_disrupted.png\"><img src=\"https:\/\/github.com\/GitSquared\/edex-ui\/raw\/master\/media\/screenshot_disrupted.png\" alt=\"Disrupted screenshot\" \/><\/a><\/p>\n<p><em>(<a href=\"https:\/\/github.com\/antonmedv\/fx\">fx<\/a> showing the results of a <a href=\"https:\/\/github.com\/GitSquared\/ipapi\">ipapi<\/a> query on eDEX-UI 1.0 with the experimental \"tron-disrupted\" theme, and the French AZERTY keyboard)<\/em><\/p>\n<h2><a id=\"user-content-qa\" class=\"anchor\" aria-hidden=\"true\" href=\"#qa\"><\/a>Q&amp;A<\/h2>\n<h4><a id=\"user-content-what-os-can-this-thing-run-on\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-os-can-this-thing-run-on\"><\/a>What OS can this thing run on?<\/h4>\n<p>Currently Windows, the latest macOS and any Linux distro that can run Chromium (AppImage package).<\/p>\n<h4><a id=\"user-content-is-this-a-real-terminal\" class=\"anchor\" aria-hidden=\"true\" href=\"#is-this-a-real-terminal\"><\/a>Is this a real terminal?<\/h4>\n<p>Yes. By default, eDEX runs bash on Linux and Powershell on Windows, but you can change that to any command in the settings.json file.<\/p>\n<h4><a id=\"user-content-i-dont-like-the-colorsthe-keyboard-layout-is-not-right-for-me\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-dont-like-the-colorsthe-keyboard-layout-is-not-right-for-me\"><\/a>I don't like the colors\/the keyboard layout is not right for me!<\/h4>\n<p>We got you covered! Check the <a href=\"https:\/\/github.com\/GitSquared\/edex-ui\/wiki\">repo's Wiki<\/a> for in-depth information about customizing eDEX.<\/p>\n<h4><a id=\"user-content-why-is-there-a-keyboard\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-is-there-a-keyboard\"><\/a>Why is there a keyboard?<\/h4>\n<p>eDEX-UI is meant to be used on touchscreens, even if it works well on regular displays! If you have a physical keyboard wired to your computer, pressing keys IRL will illuminate the virtual keyboard: <em>please remember to not type any passwords if you are recording your screen!<\/em><\/p>\n<h4><a id=\"user-content-whats-the-difference-between-this-and-the-original-dex-ui\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-the-difference-between-this-and-the-original-dex-ui\"><\/a>What's the difference between this and the original DEX-UI?<\/h4>\n<p>Seenaburns' DEX-UI was created <em>\"as an experiment or an art piece, not distributable software\"<\/em>. The goal of this project is to push Seena's vision forward by making such an interface usable in real-life scenarios.<\/p>\n<h4><a id=\"user-content-will-using-this-make-me-insanely-badass\" class=\"anchor\" aria-hidden=\"true\" href=\"#will-using-this-make-me-insanely-badass\"><\/a>Will using this make me insanely badass?<\/h4>\n<p><a href=\"https:\/\/78.media.tumblr.com\/35d4ef4447e0112f776b629bffd99188\/tumblr_mk4gf8zvyC1s567uwo1_500.gif\" rel=\"nofollow\">Yes.<\/a><\/p>\n<h2><a id=\"user-content-featured-in\" class=\"anchor\" aria-hidden=\"true\" href=\"#featured-in\"><\/a>Featured in...<\/h2>\n<ul><li><a href=\"https:\/\/www.linuxuprising.com\/2018\/11\/edex-ui-fully-functioning-sci-fi.html\" rel=\"nofollow\">Linux Uprising Blog<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/unixporn\/comments\/9ysbx7\/oc_a_little_project_that_ive_been_working_on\/\" rel=\"nofollow\">My post on r\/unixporn<\/a><\/li>\n<li><a href=\"https:\/\/korben.info\/une-interface-futuriste-pour-vos-ecrans-tactiles.html\" rel=\"nofollow\">Korben article (in french)<\/a><\/li>\n<li><a href=\"https:\/\/news.ycombinator.com\/item?id=18509828\" rel=\"nofollow\">Hacker News<\/a><\/li>\n<li><a href=\"https:\/\/twitter.com\/mikemaccana\/status\/1065615451940667396\" rel=\"nofollow\">This tweet that made me smile<\/a><\/li>\n<\/ul><h2><a id=\"user-content-useful-commands-for-the-nerds\" class=\"anchor\" aria-hidden=\"true\" href=\"#useful-commands-for-the-nerds\"><\/a>Useful commands for the nerds<\/h2>\n<h4><a id=\"user-content-starting-from-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#starting-from-source\"><\/a>Starting from source:<\/h4>\n<p>on *nix systems (You'll need the Xcode command line tools on macOS):<\/p>\n<ul><li>clone the repository<\/li>\n<li><code>npm run install-linux<\/code><\/li>\n<li><code>npm start<\/code><\/li>\n<\/ul><p>on Windows:<\/p>\n<ul><li>start cmd or powershell <strong>as administrator<\/strong><\/li>\n<li>clone the repository<\/li>\n<li><code>npm run install-windows<\/code><\/li>\n<li><code>npm start<\/code><\/li>\n<\/ul><h4><a id=\"user-content-building\" class=\"anchor\" aria-hidden=\"true\" href=\"#building\"><\/a>Building<\/h4>\n<p>Note: Due to native modules, you can only compile binaries for the OS you are compiling from.\nNote\u00b2: It is recommended to start building from a fresh clone to prevent the code minifier script from parsing an entire <code>node_modules<\/code> folder.<\/p>\n<ul><li>make a fresh clone (recommended)<\/li>\n<li><code>npm install<\/code> (NOT <code>install-linux<\/code> or <code>install-windows<\/code>)<\/li>\n<li><code>npm run build-linux<\/code> or <code>build-windows<\/code> or <code>build-darwin<\/code><\/li>\n<\/ul><p>Resulting binaries and assets will be in the <code>dist<\/code> folder.<\/p>\n<h2><a id=\"user-content-credits\" class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h2>\n<p>eDEX-UI's source code was primarily written by me, <a href=\"https:\/\/github.com\/GitSquared\">Squared<\/a>. If you want to get in touch with me or find other projects I'm involved in, check out <a href=\"https:\/\/squared.codebrew.fr\" rel=\"nofollow\">my website<\/a>.<\/p>\n<p><a href=\"https:\/\/github.com\/PixelyIon\">PixelyIon<\/a> helped me get started with Windows compatibility and offered some precious advice when I started to work on this project seriously.<\/p>\n<h2><a id=\"user-content-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#thanks\"><\/a>Thanks<\/h2>\n<p>Of course, eDEX would never have existed if I hadn't stumbled upon the amazing work of <a href=\"https:\/\/github.com\/seenaburns\">Seena<\/a> on <a href=\"https:\/\/reddit.com\/r\/unixporn\" rel=\"nofollow\">r\/unixporn<\/a>.<\/p>\n<p>This project uses a bunch of open-source libraries, frameworks and tools, see <a href=\"https:\/\/github.com\/GitSquared\/edex-ui\/network\/dependencies\">the full dependency graph<\/a>.<\/p>\n<p>I want to namely thank the developers behind <a href=\"https:\/\/github.com\/xtermjs\/xterm.js\">xterm.js<\/a>, <a href=\"https:\/\/github.com\/sebhildebrandt\/systeminformation\">systeminformation<\/a> and <a href=\"https:\/\/github.com\/joewalnes\/smoothie\">SmoothieCharts<\/a>.<\/p>\n<p>Huge thanks to <a href=\"https:\/\/github.com\/arscan\">Rob \"Arscan\" Scanlon<\/a> for making the fantastic <a href=\"https:\/\/github.com\/arscan\/encom-globe\">ENCOM Globe<\/a>, also inspired by the TRON: Legacy movie, and distributing it freely. His work really puts the icing on the cake.<\/p>\n<\/article>","created_at":"2018-11-22T17:55:44+0000","updated_at":"2018-11-22T17:55:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/24496417?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":704,"title":"jbmusso\/awesome-graph","url":"https:\/\/github.com\/jbmusso\/awesome-graph\/blob\/master\/README.md","content":"<p>A curated list of resources for graph databases and graph computing tools<\/p>\n<h2><a id=\"user-content-graph-databases\" class=\"anchor\" aria-hidden=\"true\" href=\"#graph-databases\"><\/a>Graph databases<\/h2>\n<ul><li><a href=\"https:\/\/bitnine.net\/agensgraph-2\/\" rel=\"nofollow\">AgensGraph<\/a> - multi-model graph database with SQL and Cypher support<\/li>\n<li><a href=\"https:\/\/www.arangodb.com\/\" rel=\"nofollow\">ArangoDB<\/a> - highly available Multi-Model NoSQL database<\/li>\n<li><a href=\"https:\/\/github.com\/blazegraph\/database\">Blazegraph<\/a> - GPU accelerated graph database<\/li>\n<li><a href=\"https:\/\/github.com\/cayleygraph\/cayley\">Cayley<\/a> - open source database written in Go<\/li>\n<li><a href=\"https:\/\/docs.microsoft.com\/en-us\/azure\/cosmos-db\/graph-introduction\" rel=\"nofollow\">CosmosDB<\/a> - cloud-based multi-model database with support for TinkerPop3<\/li>\n<li><a href=\"https:\/\/dgraph.io\" rel=\"nofollow\">Dgraph<\/a> - Fast, Transactional, Distributed Graph Database (open source, written in Go)<\/li>\n<li><a href=\"https:\/\/www.datastax.com\/products\/datastax-enterprise-graph\" rel=\"nofollow\">DSE Graph<\/a> - Graph layer on top of DataStax Enterprise (Cassandra, SolR, Spark)<\/li>\n<li><a href=\"https:\/\/grakn.ai\/\" rel=\"nofollow\">Grakn.AI<\/a> - a distributed hyper-relational database for knowledge-oriented systems, i.e. a distributed knowledge base<\/li>\n<li><a href=\"https:\/\/github.com\/google\/graphd\">Graphd<\/a> - the Metaweb\/Freebase Graph Repository<\/li>\n<li><a href=\"http:\/\/janusgraph.org\" rel=\"nofollow\">JanusGraph<\/a> - an open-source, distributed graph database with pluggable storage and indexing backends<\/li>\n<li><a href=\"https:\/\/memgraph.com\/\" rel=\"nofollow\">Memgraph<\/a> - High Performance, In-Memory, Transactional Graph Database<\/li>\n<li><a href=\"http:\/\/tinkerpop.apache.org\/docs\/currentg\/#neo4j-gremlin\" rel=\"nofollow\">Neo4j<\/a> - OLTP graph database<\/li>\n<li><a href=\"http:\/\/www.sparsity-technologies.com\/#sparksee\" rel=\"nofollow\">Sparksee<\/a> - makes space and performance compatible with a small footprint and a fast analysis of large networks<\/li>\n<li><a href=\"http:\/\/stardog.com\/\" rel=\"nofollow\">Stardog<\/a> - RDF graph database with OLTP and OLAP support<\/li>\n<li><a href=\"http:\/\/orientdb.com\/orientdb\/\" rel=\"nofollow\">OrientDB<\/a> - Distributed Multi-Model NoSQL Database with a Graph Database Engine<\/li>\n<\/ul><h3><a id=\"user-content-triple-stores\" class=\"anchor\" aria-hidden=\"true\" href=\"#triple-stores\"><\/a>Triple stores<\/h3>\n<ul><li><a href=\"https:\/\/franz.com\/agraph\/allegrograph\/\" rel=\"nofollow\">AllegroGraph<\/a> - high-performance, persistent graph database that scales to billions of quads<\/li>\n<li><a href=\"https:\/\/jena.apache.org\/\" rel=\"nofollow\">Apache Jena<\/a> - open source Java framework for building Semantic Web and Linked Data applications<\/li>\n<li><a href=\"http:\/\/rdf4j.org\/\" rel=\"nofollow\">Eclipse RDF4J<\/a> - (formerly known as Sesame) is an open source Java framework for processing RDF data. This includes parsing, storing, inferencing and querying of\/over such data. It offers an easy-to-use API that can be connected to all leading RDF storage solutions. It allows you to connect with SPARQL endpoints and create applications that leverage the power of linked data and Semantic Web.<\/li>\n<li><a href=\"http:\/\/graphdb.ontotext.com\/graphdb\/\" rel=\"nofollow\">GraphDB<\/a> - enterprise ready Semantic Graph Database, compliant with W3C Standards<\/li>\n<li><a href=\"https:\/\/virtuoso.openlinksw.com\/\" rel=\"nofollow\">Virtuoso<\/a> - a \"Data Junction Box\" that drives enterprise and individual agility by deriving a Semantic Web of Linked Data from existing data silos<\/li>\n<\/ul><h2><a id=\"user-content-graph-computing-frameworks\" class=\"anchor\" aria-hidden=\"true\" href=\"#graph-computing-frameworks\"><\/a>Graph computing frameworks<\/h2>\n<ul><li><a href=\"https:\/\/giraph.apache.org\/\" rel=\"nofollow\">Apache Giraph<\/a> - an iterative graph processing system built for high scalability<\/li>\n<li><a href=\"https:\/\/tinkerpop.apache.org\/\" rel=\"nofollow\">Apache TinkerPop<\/a> - a graph computing framework for both graph databases (OLTP) and graph analytic systems (OLAP)<\/li>\n<li><a href=\"https:\/\/spark.apache.org\/graphx\/\" rel=\"nofollow\">Apache Spark - GraphX<\/a> - Apache Spark's API for graphs and graph-parallel computation<\/li>\n<\/ul><h2><a id=\"user-content-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#languages\"><\/a>Languages<\/h2>\n<ul><li><a href=\"http:\/\/www.opencypher.org\/\" rel=\"nofollow\">Cypher<\/a><\/li>\n<li><a href=\"https:\/\/tinkerpop.apache.org\/gremlin.html\" rel=\"nofollow\">Gremlin<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/SPARQL\" rel=\"nofollow\">SPARQL<\/a><\/li>\n<\/ul><h2><a id=\"user-content-managed-hosting-services\" class=\"anchor\" aria-hidden=\"true\" href=\"#managed-hosting-services\"><\/a>Managed hosting services<\/h2>\n<ul><li><a href=\"https:\/\/docs.microsoft.com\/en-us\/azure\/cosmos-db\/graph-introduction\" rel=\"nofollow\">CosmosDB @ Microsoft<\/a><\/li>\n<li><a href=\"https:\/\/console.ng.bluemix.net\/catalog\/services\/ibm-graph\/\" rel=\"nofollow\">IBM Graph @ IBM Bluemix<\/a> - OLTP graph database as a service.<\/li>\n<li><a href=\"https:\/\/www.compose.com\/databases\/janusgraph\" rel=\"nofollow\">JanusGraph @ IBM Compose<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/awslabs\/dynamodb-janusgraph-storage-backend\">JanusGraph @ Amazon Web Services Labs<\/a><\/li>\n<li><a href=\"https:\/\/www.graphenedb.com\/\" rel=\"nofollow\">Neo4j @ Graphene<\/a><\/li>\n<li><a href=\"https:\/\/aws.amazon.com\/neptune\/\" rel=\"nofollow\">Neptune @ Amazon Web Services<\/a> - a fast, reliable, fully-managed graph database service that makes it easy to build and run applications that work with highly connected datasets<\/li>\n<\/ul><h2><a id=\"user-content-learning-materials\" class=\"anchor\" aria-hidden=\"true\" href=\"#learning-materials\"><\/a>Learning materials<\/h2>\n<h3><a id=\"user-content-official-documentations\" class=\"anchor\" aria-hidden=\"true\" href=\"#official-documentations\"><\/a>Official documentations<\/h3>\n<ul><li><a href=\"https:\/\/neo4j.com\/developer\/cypher-query-language\/\" rel=\"nofollow\">Cypher<\/a> - reference documentation<\/li>\n<li><a href=\"http:\/\/tinkerpop.apache.org\/docs\/current\/reference\/#traversal\" rel=\"nofollow\">Gremlin<\/a> - reference documentation<\/li>\n<\/ul><h3><a id=\"user-content-community-effort\" class=\"anchor\" aria-hidden=\"true\" href=\"#community-effort\"><\/a>Community effort<\/h3>\n<ul><li><a href=\"https:\/\/github.com\/krlawrence\/graph\">Graph Book<\/a> - TinkerPop3 centric book written by <a href=\"https:\/\/twitter.com\/gfxman\" rel=\"nofollow\">Kelvin R. Lawrence<\/a><\/li>\n<li><a href=\"http:\/\/sql2gremlin.com\/\" rel=\"nofollow\">SQL2Gremlin<\/a> - transition from SQL to Gremlin by <a href=\"https:\/\/twitter.com\/dkuppitz\" rel=\"nofollow\">Daniel Kuppitz<\/a><\/li>\n<li><a href=\"http:\/\/www.doanduyhai.com\/blog\/?p=13460\" rel=\"nofollow\">The Gremlin Compendium<\/a> - minimum survival kit for any Gremlin user, 10 blog post series by <a href=\"https:\/\/twitter.com\/doanduyhai\" rel=\"nofollow\">Doan DuyHai<\/a><\/li>\n<\/ul><h2><a id=\"user-content-conferences\" class=\"anchor\" aria-hidden=\"true\" href=\"#conferences\"><\/a>Conferences<\/h2>\n<ul><li><a href=\"http:\/\/graphconnect.com\/\" rel=\"nofollow\">Graph Connect<\/a> - powered by Neo4j<\/li>\n<li><a href=\"http:\/\/graphday.com\/\" rel=\"nofollow\">Graph Day<\/a> - an Independent Graph Conference from the Data Day folks<\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p><a href=\"https:\/\/creativecommons.org\/publicdomain\/zero\/1.0\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/60561947585c982aee67ed3e3b25388184cc0aa3\/687474703a2f2f6d6972726f72732e6372656174697665636f6d6d6f6e732e6f72672f70726573736b69742f627574746f6e732f38387833312f7376672f63632d7a65726f2e737667\" alt=\"CC0\" data-canonical-src=\"http:\/\/mirrors.creativecommons.org\/presskit\/buttons\/88x31\/svg\/cc-zero.svg\" \/><\/a><\/p>\n<p>To the extent possible under law, <a href=\"https:\/\/github.com\/jbmusso\">Jean-Baptiste Musso<\/a> has waived all copyright and related or neighboring rights to this work.<\/p>","created_at":"2018-11-22T03:07:27+0000","updated_at":"2018-11-22T03:07:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/203476?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":703,"title":"walmartlabs\/lacinia","url":"https:\/\/github.com\/walmartlabs\/lacinia\/blob\/master\/README.md","content":"<p><a href=\"https:\/\/clojars.org\/com.walmartlabs\/lacinia\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/97e1b55b88d20ede5a71d88f039fd634f814a186\/68747470733a2f2f696d672e736869656c64732e696f2f636c6f6a6172732f762f636f6d2e77616c6d6172746c6162732f6c6163696e69612e737667\" alt=\"Clojars Project\" data-canonical-src=\"https:\/\/img.shields.io\/clojars\/v\/com.walmartlabs\/lacinia.svg\" \/><\/a><\/p>\n<p><a href=\"https:\/\/circleci.com\/gh\/walmartlabs\/lacinia\/tree\/master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/c51057956def272c6834706f4efbae1b39aff73b\/68747470733a2f2f636972636c6563692e636f6d2f67682f77616c6d6172746c6162732f6c6163696e69612f747265652f6d61737465722e7376673f7374796c653d737667\" alt=\"CircleCI\" data-canonical-src=\"https:\/\/circleci.com\/gh\/walmartlabs\/lacinia\/tree\/master.svg?style=svg\" \/><\/a><\/p>\n<p><a href=\"http:\/\/lacinia.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">Lacinia Manual<\/a> |\n<a href=\"http:\/\/lacinia.readthedocs.io\/en\/latest\/tutorial\" rel=\"nofollow\">Lacinia Tutorial<\/a> |\n<a href=\"http:\/\/walmartlabs.github.io\/lacinia\/\" rel=\"nofollow\">API Documentation<\/a><\/p>\n<p>This library is a full implementation of\nFacebook's <a href=\"https:\/\/facebook.github.io\/graphql\" rel=\"nofollow\">GraphQL specification<\/a>.<\/p>\n<p>Lacinia should be viewed as roughly analogous to the\n<a href=\"https:\/\/github.com\/graphql\/graphql-js\/\">official reference JavaScript implementation<\/a>.\nIn other words, it is a backend-agnostic GraphQL query execution engine.\nLacinia is not an Object Relational Mapper ... it's simply the implementation of a contract\nsitting between the GraphQL client and your data.<\/p>\n<p>Lacinia features:<\/p>\n<ul><li>\n<p>An <a href=\"https:\/\/github.com\/edn-format\/edn\">EDN<\/a>-based schema language.<\/p>\n<\/li>\n<li>\n<p>High performance parser for GraphQL queries, built on <a href=\"http:\/\/www.antlr.org\/\" rel=\"nofollow\">Antlr4<\/a>.<\/p>\n<\/li>\n<li>\n<p>Efficient and asynchronous query execution.<\/p>\n<\/li>\n<li>\n<p>Full support for GraphQL types, interfaces, unions, enums, input objects, and custom scalars.<\/p>\n<\/li>\n<li>\n<p>Full support for GraphQL subscriptions.<\/p>\n<\/li>\n<li>\n<p>Full support of inline and named query fragments.<\/p>\n<\/li>\n<li>\n<p>Full support for GraphQL Schema Introspection.<\/p>\n<\/li>\n<\/ul><p>Lacinia has been developed with a set of core philosophies:<\/p>\n<ul><li>\n<p>Prefer data over macros and other tricks. Compose your schema in whatever mix of data and code works for you.<\/p>\n<\/li>\n<li>\n<p>Embrace Clojure: Use EDN data, keywords, functions, and persistent data structures.<\/p>\n<\/li>\n<li>\n<p>Keep it simple: You provide the schema and a handful of functions to resolve data, and Lacinia does the rest.<\/p>\n<\/li>\n<li>\n<p>Do the right thing: apply reasonable defaults without a lot of \"magic\".<\/p>\n<\/li>\n<\/ul><p>This library can be plugged into any Clojure HTTP pipeline.\nThe companion library <a href=\"https:\/\/github.com\/walmartlabs\/lacinia-pedestal\">lacinia-pedestal<\/a> provides\nfull HTTP support, including GraphQL subscriptions, for <a href=\"https:\/\/github.com\/pedestal\/pedestal\">Pedestal<\/a>.<\/p>\n<h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting Started<\/h2>\n<p>For more detailed documentation, <a href=\"http:\/\/lacinia.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">read the manual<\/a>.<\/p>\n<p>GraphQL starts with a schema definition of exposed types.<\/p>\n<p>A schema starts as an EDN file; the example below demonstrates a small subset\nof the available options:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:enums\n {:episode\n  {:description \"The episodes of the original Star Wars trilogy.\"\n   :values [:NEWHOPE :EMPIRE :JEDI]}}\n :objects\n {:droid\n  {:fields {:primary_functions {:type (list String)}\n            :id {:type Int}\n            :name {:type String}\n            :appears_in {:type (list :episode)}}}\n  :human\n  {:fields {:id {:type Int}\n            :name {:type String}\n            :home_planet {:type String}\n            :appears_in {:type (list :episode)}}}}\n :queries\n {:hero {:type (non-null :human)\n         :args {:episode {:type :episode}}\n         :resolve :get-hero}\n  :droid {:type :droid\n          :args {:id {:type String :default-value \"2001\"}}\n          :resolve :get-droid}}}<\/pre><\/div>\n<p>A schema alone describes what data is available to clients, but doesn't identify where\nthe data comes from; that's the job of a field resolver, provided by the\n:resolve key inside fields such as the :hero and :droid query.<\/p>\n<p>The values here, :get-hero and :get-droid, are placeholders; the startup code\nof the application will use\n<code>com.walmartlabs.lacinia.util\/attach-resolvers<\/code> to attach the actual\nfield resolver function.<\/p>\n<p>A field resolver is just a function which is passed the application context,\na map of arguments values, and a resolved value from a\nparent field.\nThe field resolver returns a value. If it's a scalar type, it should return a value\nthat conforms to the defined type in the schema.\nIf not, it's a type error.<\/p>\n<p>The field resolver is totally responsible for obtaining the data from whatever\nexternal store you use: whether it is a database, a web service, or something\nelse.<\/p>\n<p>It's important to understand that <em>every<\/em> field has a field resolver, even if\nyou don't define it explicitly.  If you don't supply a field resolver,\nLacinia provides a default field resolver, customized to the field.<\/p>\n<p>Here's what the <code>get-hero<\/code> field resolver might look like:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn get-hero [context arguments value]\n  (let [{:keys [episode]} arguments]\n    (if (= episode :NEWHOPE)\n      {:id 1000\n       :name \"Luke\"\n       :home_planet \"Tatooine\"\n       :appears_in [\"NEWHOPE\" \"EMPIRE\" \"JEDI\"]}\n      {:id 2000\n       :name \"Lando Calrissian\"\n       :home_planet \"Socorro\"\n       :appears_in [\"EMPIRE\" \"JEDI\"]})))<\/pre><\/div>\n<p>In this greatly simplified example, the field resolver can simply return the resolved value.\nField resolvers that return multiple values return a list, vector, or set of values.<\/p>\n<p>In real applications, a field resolver might execute a query against a database,\nor send a request to another web service.<\/p>\n<p>After attaching resolvers, it is necessary to compile the schema; this\nstep performs validations, provide defaults, and organizes the schema\nfor efficient execution of queries.<\/p>\n<p>This needs only be done once, in application startup code:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[clojure.edn :as edn]\n         '[com.walmartlabs.lacinia.util :refer [attach-resolvers]]\n         '[com.walmartlabs.lacinia.schema :as schema])\n(def star-wars-schema\n  (-&gt; \"schema.edn\"\n      slurp\n      edn\/read-string\n      (attach-resolvers {:get-hero get-hero\n                         :get-droid (constantly {})})\n      schema\/compile))<\/pre><\/div>\n<p>With the compiled application available, it can be used to execute\nrequests; this typically occurs inside a Ring handler function:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[com.walmartlabs.lacinia :refer [execute]]\n         '[clojure.data.json :as json])\n(defn handler [request]\n  {:status 200\n   :headers {\"Content-Type\" \"application\/json\"}\n   :body (let [query (get-in request [:query-params :query])\n               result (execute star-wars-schema query nil nil)]\n           (json\/write-str result)})<\/pre><\/div>\n<p>Lacinia doesn't know about the web tier at all, it just knows about\nparsing and executing queries against a compiled schema.\nA companion library, <a href=\"https:\/\/github.com\/walmartlabs\/lacinia-pedestal\">lacinia-pedestal<\/a>,\nis one way to expose your schema on the web.<\/p>\n<p>User queries are provided as the body of a request with the content type <code>application\/graphql<\/code>.\nThe GraphQL query language is designed to look familiar to someone who is versant in JSON.<\/p>\n<pre>{\n  hero {\n    id\n    name\n  }\n}\n<\/pre>\n<p>The <code>execute<\/code> function returns EDN data that can be easily converted to JSON.\nThe :data key contains the value requested for the <code>hero<\/code> query in the request.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:data\n  {:hero {:id 2000\n          :name \"Lando Calrissian\"}}}<\/pre><\/div>\n<p>This example request has no errors, and contained only a single query.\nGraphQL supports multiple queries in a single request.\nThere may be errors executing the query, Lacinia will process as much as\nit can, and will report errors in the :errors key.<\/p>\n<p>One of the benefits of GraphQL is that the client has the power to rename\nfields in the response:<\/p>\n<pre>{\n  hero(episode: NEWHOPE) {\n    movies: appears_in\n  }\n}\n<\/pre>\n<div class=\"highlight highlight-source-clojure\"><pre>{:data {:hero {:movies [:NEWHOPE :EMPIRE :JEDI]}}}<\/pre><\/div>\n<p>This is just an overview, far more detail is available\nin <a href=\"http:\/\/lacinia.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">the manual<\/a>.<\/p>\n<h2><a id=\"user-content-status\" class=\"anchor\" aria-hidden=\"true\" href=\"#status\"><\/a>Status<\/h2>\n<p>Although this library is used in production at Walmart, it is\nstill considered alpha software - subject to change.\nWe expect to stabilize it in the near future.<\/p>\n<p>To use this library with Clojure 1.8, you must include\na dependency on <a href=\"https:\/\/github.com\/tonsky\/clojure-future-spec\">clojure-future-spec<\/a>.<\/p>\n<p>More details are <a href=\"http:\/\/lacinia.readthedocs.io\/en\/latest\/clojure.html\" rel=\"nofollow\">in the manual<\/a>.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2017-2018 WalmartLabs<\/p>\n<p>Distributed under the Apache License, Version 2.0.<\/p>\n<p>Portions of the code are derived from\nthe <a href=\"https:\/\/github.com\/amalloy\/ordered\">ordered<\/a>\nand <a href=\"https:\/\/github.com\/amalloy\/useful\">useful<\/a> libraries, which are released under the terms\nof the <a href=\"https:\/\/github.com\/walmartlabs\/lacinia\/blob\/master\/LICENSE.ordered.txt\">Eclipse Public License - v 1.0<\/a>.<\/p>","created_at":"2018-11-22T02:49:34+0000","updated_at":"2018-11-22T02:49:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/768298?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":702,"title":"Slow Software","url":"https:\/\/www.inkandswitch.com\/slow-software.html","content":"<h2 id=\"what-feels-slow\">What feels slow<\/h2><p>What feels \"slow\" to users? We all have a sense for when software annoys us with delays. But to get a better handle on this problem, we'll complement these intuitions with academic research that answers the question rigorously.<\/p><p>Perceived speed is all about latency. Comparing the \"what feels slow\" findings from academia with measurements of real-world app latency tells us just how bad things really are.<\/p><h3 id=\"latency-not-throughput\">Latency not throughput<\/h3><p>When discussing software performance, we often hear about <em>throughput<\/em>. E.g. \"this web server can do 10,000 requests \/ second.\" But that's not how users perceive things. They care how long their particular web request takes, or how long a doc takes to open, or how responsive an app is to their clicks. These interactions are about <em>latency<\/em>.<small class=\"note\">\n        User perception\n        See <a id=\"ret-1\" href=\"#ref-1\">[1]<\/a> for more on the importance of latency in interactive systems.\n        Other factors influence how fast software feels. E.g. frame rates and feedback about long-running tasks. But we think latency is the most fundamental, and we're only going to have software that feels truly instantaneous if latency is very low.\n      <\/small> Latency is the critical metric we'll examine in this article.<\/p><h3 id=\"touch-interfaces\">Touch interfaces<\/h3><p>To start, let's look at user sensitivity to latency when using touch screens.<\/p><p>Researchers can test this with rigs that control exactly how much latency users see. They present a user with an interface having (say) 1ms latency and another with (say) 70ms and then ask them to perform operations like tapping a button. If the 70ms interface consistently feels slower than the 1ms interface, 70ms would be a \"noticeable difference.\"<small class=\"note\">\n        Just noticeable differences\n        For more on just noticeable latency differences and associated experimental rigs, see e.g. <a href=\"#ref-2\">[2]<\/a>.\n      <\/small><\/p><p>The smallest such noticeable difference is the budget one has before a given operation starts to feel slow to that user.<\/p><p>When dragging items on the screen, for example, users perceive latencies as low as ~2ms. The just noticeable latency varies by user and action being performed, but it's consistently very low.<small class=\"note\">\n        Drag latency perception\n        <a id=\"ret-2\" href=\"#ref-2\">[2]<\/a> reports 2\u201316ms for drags with a stylus.\n        <a id=\"ret-3\" href=\"#ref-3\">[3]<\/a>: 2\u201311ms for drags with finger.\n        <a id=\"ret-4\" href=\"https:\/\/www.inkandswitch.com\/ref-4\">[4]<\/a>: mean of 11ms for drags with finger.\n      <\/small><\/p><p>Inking on a tablet with a stylus gives similar results. Here the literature suggests users detect slowness between 20ms and 80ms of latency. In our own informal tests in the lab, latency towards 80ms feels very slow, and it takes something much closer to 20ms for a stylus to feel responsive while inking.<small class=\"note\">\n        Stylus latency perception\n        <a href=\"#ref-2\">[2]<\/a> reports 10\u201370ms for stylus scribbling.\n        <a id=\"ret-5\" href=\"#ref-5\">[5]<\/a>: 21\u201382ms for various stylus actions.\n      <\/small><\/p><p>The difference between low- and high-latency inking is clear when looking at contrasting examples side-by-side:<\/p><div class=\"caption\"><p class=\"caption-title\">Inking latency examples<\/p><p>Left: an iPad Pro and Notes app with ~15ms of end-to-end latency.<\/p><p>Right: a Samsung S3 and OneNote app with ~70ms of latency.<\/p><p>Videos slowed 16x.<\/p><\/div><div class=\"video\"><img class=\"play\" src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/play.svg\" alt=\"image\" \/><\/div><p>Another common operation on touch devices is tapping on buttons or links. Here tests suggests users on average notice latency as it goes beyond ~70ms (though it's likely lower for some individual users).<small class=\"note\">\n        Tapping latency perception\n        <a href=\"#ref-4\">[4]<\/a> reports 69ms mean just noticeable latency for finger tap across users.\n      <\/small><\/p><p>Here's an example showing two different latencies side-by-side:<\/p><div class=\"caption\"><p class=\"caption-title\">Tapping latency examples<\/p><p>Left: Opening a settings tab on an iPhone 6s with ~90ms of latency.<\/p><p>Right: Toggling a setting a Samsung S3 with ~330ms of latency.<\/p><p>Videos slowed 16x.<\/p><\/div><div class=\"video\"><img class=\"play\" src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/play.svg\" alt=\"image\" \/><\/div><p>How do modern apps fare compared to these latency thresholds for touch interactions? In terms of dragging with a finger, no current consumer system will consistently meet the low single digit millisecond level needed to satisfy all users.<small class=\"note\">\n        Drag performance\n        As we'll see below, &lt;10ms is not enough latency budget even for input hardware and displays, not to mention multiple layers of software.\n      <\/small> So all current touchscreen operating systems will leave at least some users feeling like the object they're dragging is lagging behind their finger.<\/p><p>For drawing with a stylus, a small number of systems get close to the latency levels needed to feel reasonably good. But most are well above these levels and\u2014as we'd expect\u2014feel very slow to users:\n      <\/p><div class=\"caption\"><p class=\"caption-title\">Inking latencies<\/p><p>Results from Ink &amp; Switch  tests on tablet inking latency.<\/p><p>Average latencies as measured from screen contact to start of corresponding pixel color change, rounded to nearest 5ms.<\/p><\/div><div class=\"table\"><table><tr><th class=\"text left\">Device<\/th>\n            <th class=\"text middle\">Program<\/th>\n            <th colspan=\"2\" class=\"mixed right\">Latency (ms)<\/th>\n          <\/tr><tr><td class=\"text left first\">iPad Pro<\/td>\n            <td class=\"text middle first\">Notes<\/td>\n            <td class=\"number middle first\">20<\/td>\n            <td class=\"bar right first\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">GoodNotes<\/td>\n            <td class=\"number middle\">30<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">Flutter<\/td>\n            <td class=\"number middle\">35<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left first\">Surface Pro<\/td>\n            <td class=\"text middle first\">OneNote<\/td>\n            <td class=\"number middle first\">25<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">SketchPad<\/td>\n            <td class=\"number middle\">30<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">Canvas<\/td>\n            <td class=\"number middle\">60<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left first\">Pixelbook<\/td>\n            <td class=\"text middle first\">Squid<\/td>\n            <td class=\"number middle first\">40<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">Canvas<\/td>\n            <td class=\"number middle\">60<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left first\">Samsung S3<\/td>\n            <td class=\"text middle first\">Squid<\/td>\n            <td class=\"number middle first\">60<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">Flutter<\/td>\n            <td class=\"number middle\">65<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">\n            <\/td><td class=\"text middle\">Canvas<\/td>\n            <td class=\"number middle\">75<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left last\">\n            <\/td><td class=\"text middle last\">LiveBoard<\/td>\n            <td class=\"number middle last\">80<\/td>\n            <td class=\"bar right last\">\n          <\/td><\/tr><\/table><\/div><p>While we don't have data on tapping latencies across devices, we'd expect them to be comparable to the inking latencies observed above.<small class=\"note\">\n        Inking vs tapping latencies\n        We expect these latencies to be comparable because they both test the touch input \u2192 screen update loop. That said, there are differences between the two paths so we wouldn't expect them to be exactly the same.\n      <\/small> Since the noticeable latency here is ~70ms, most systems should be able to feel responsive to taps. But it's also easy to find apps that perform much worse than the system's theoretical capability.<\/p><p>Overall, touch systems need to have very low latencies to feel responsive. Most devices and apps fail to perform at this level, and consequently feel varying degrees of slow to users.<\/p><h3 id=\"typing\">Typing<\/h3><p>There is some evidence that increased typing latency impairs users.<small class=\"note\">\n        Typing latency impact\n        In <a id=\"ret-6\" href=\"#ref-6\">[6]<\/a>, random amounts of latency are added to keying operations, which reduces typing performance. However it only assessed one distribution of latency. It also suggested skilled typists may acclimate to the increased latency.\n      <\/small> However, we're not aware of studies specifically measuring least noticeable end-to-end typing latencies. The tap latency numbers (noticeable at ~70ms) may be a useful benchmark because they also measure a discrete finger touch to visual update.<\/p><p>Here are some informal end-to-end keyboard latency measurements:<\/p><div class=\"caption\"><p class=\"caption-title\">Typing latencies<\/p><p>End-to-end latencies from start of keypress to a character appearing in app for a variety of machines.<\/p><p>Sources: <a href=\"https:\/\/danluu.com\/input-lag\/\">Computer latency: 1977-2017<\/a>, Ink &amp; Switch tests<\/p><\/div><div class=\"table\"><table><tr><th class=\"text left\">Computer<\/th>\n            <th colspan=\"2\" class=\"mixed right\">Latency (ms)<\/th>\n          <\/tr><tr><td class=\"text left first\">Apple IIe<\/td>\n            <td class=\"number middle first\">30<\/td>\n            <td class=\"bar right first\">\n          <\/td><\/tr><tr><td class=\"text left\">Commodore Pet 4016<\/td>\n            <td class=\"number middle\">60<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">iMac g4 OS 9<\/td>\n            <td class=\"number middle\">70<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Macbook Pro 2014<\/td>\n            <td class=\"number middle\">100<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Custom Haswell-e 24Hz<\/td>\n            <td class=\"number middle\">140<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Samsung S3<\/td>\n            <td class=\"number middle\">150<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Powerspec g405 Linux<\/td>\n            <td class=\"number middle\">170<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left last\">Symbolics 3620<\/td>\n            <td class=\"number middle last\">300<\/td>\n            <td class=\"bar right last\">\n          <\/td><\/tr><\/table><\/div><p>Rigorous assessments of the impact of keyboard latency would be great experiments for enterprising researchers. In any event, it seems likely the latency threshold for typing is below ~100ms for many users, and perhaps well below it.<\/p><h3 id=\"mousing\">Mousing<\/h3><p>The last input type we'll look at is mice. One experiment found user latency perception thresholds down to 34ms.<small class=\"note\">\n        Mouse latency impact\n        <a id=\"ret-7\" href=\"#ref-7\">[7]<\/a> measures latency sensitivity in a cursor dragging test and reports sensitivities in the range of 34\u2013137ms, with a mean of 65ms.\n      <\/small><\/p><p>Input latency of mice varies widely. Some setups achieve latencies in the single digit milliseconds range by combining high-performance hardware with careful, low-level programming. It's also possible to go beyond 100ms of end-to-latency with a combination of mediocre hardware and applications that introduce extra delays or buffers between input and display.<small class=\"note\">\n        Mouse latencies\n        <a href=\"#ref-7\">[7]<\/a> describes an optimized setup with end-to-end average mouse movement latencies of ~8ms.\n      <\/small><\/p><h3 id=\"applications\">Applications<\/h3><p>Application-level latencies measure how long it takes to complete app-specific actions like loading web pages:<\/p><div class=\"caption\"><p class=\"caption-title\">Application latency example<\/p><p>An example of application-level latency. It takes ~3000ms to load this NYTimes web page.<\/p><\/div><div class=\"video\"><img class=\"play\" src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/play.svg\" alt=\"image\" \/><\/div><p>What feels fast for application actions? It's hard to say exactly because actions are more complex and varied than simple inputs. The answer here probably also depends on what users are conditioned to expect (currently, it's usually slow software). But we can triangulate a rough number.\n      <small class=\"note\">\n        Latency literature\n        <a id=\"ret-8\" href=\"#ref-8\">[8]<\/a> surveys literature on the impact of latency to application users. It's a good starting point for a deeper dive on this topic.\n      <\/small><\/p><p>One reference is the typical ~70ms just noticeable difference mentioned above for tapping a touch screen. If you notice a delay between tapping a link and seeing a tap indicator, you can probably notice a similar delay between tapping a link and seeing a web page open.<\/p><p>Another data point is Google\u2019s RAIL model. This model claims that responses within 100ms \u201cfeel like the result is immediate\u201d and that higher latency \u201c[breaks] the connection between action and reaction\u201d.<small class=\"note\">\n        RAIL\n        <a href=\"https:\/\/developers.google.com\/web\/fundamentals\/performance\/rail\">Google's model<\/a>\n        Unfortunately they doesn't give a basis for their 100ms figure.\n      <\/small><\/p><p>You can informally test our own sensitivity in a terminal. Consider the command-line programs you use and how they feel, then try them with `time`. You'll notice the difference between e.g. ~15ms CLI responses (great!) and ~500ms (obviously slow).<\/p><div class=\"caption\"><p class=\"caption-title\">Terminal execution times<\/p><p>Examples of execution times for different terminal commands.<\/p><p>Even a command that takes half a second is distractingly slow.<\/p><\/div><img src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/terminal-times.png\" alt=\"image\" \/><p>As a final data point, consider that typical human reaction time from seeing a visual stimulus to taking a physical action is about 220ms.<small class=\"note\">\n        Reaction time\n        See <a href=\"https:\/\/web.archive.org\/web\/20100611222125\/http:\/\/biae.clemson.edu\/bpc\/bp\/Lab\/110\/reaction.htm\">Literature Review on Reaction Time<\/a>\n      <\/small> This value must be significantly more than noticeable latencies, because reactions involve observing something and then doing something.<\/p><p>Altogether we think this suggests action latencies should be ~100ms or less to avoid user perception of delay.<small class=\"note\">\n        Physiological effects\n        Software latency may be even more damaging than we consciously realize.\n        <a id=\"ret-9\" href=\"#ref-9\">[9]<\/a> documents unconscious, physiological effects in users exposed to increased latency.\n      <\/small><\/p><h3 id=\"real-world-apps\">Real-world apps<\/h3><p>How do current apps fair against this benchmark? Some do well. For example, many Unix command line programs run in under 100ms.<\/p><p>Most of the web does poorly. A ~1,000ms Google search result will feel faster than most of what you see online, though still noticeably delayed compared to a ~100ms interaction. And it's easy to find examples of pages that take ~5,000ms+ to load even on a good connection.<\/p><p>In the case of mobile and desktop, there are some apps that will consistently achieve &lt;100ms latency, such as the built-in calculator on iOS. But it's easy to find cases of productivity apps that significantly exceed this threshold even when they have (or should have) all data available locally. Consider the Slack example below:<\/p><div class=\"caption\"><p class=\"caption-title\">Slack latency example<\/p><p>Slack on an iPad Pro took ~220ms to change between two low-volume channels in the same workspace.<\/p><p>This is a long time considering that no network call is needed and the iPad Pro is perhaps the highest performance mobile device in the world.<\/p><p>Video slowed 8x.<\/p><\/div><div class=\"video\"><img class=\"play\" src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/play.svg\" alt=\"image\" \/><\/div><p>It's hard to draw a general conclusion about such a broad area as the action latency of all software. That said, it seems clear that while some apps do complete actions fast enough to feel instantaneous to users (less than ~100ms), many apps do not.<\/p><h2 id=\"where-slowness-comes-from\">Where slowness comes from<\/h2><p>So we've established that a lot of software is in fact slow. Where does all that time go (and what might we optimize)? We'll look at this next, starting with the first component in the chain: input devices.<\/p><h3 id=\"input-devices\">Input devices<\/h3><p>The first step in the pipeline that converts physical inputs to updates on screen is the input processing: converting contact with a touch screen, keyboard, or mouse to a digital signal for the operating system. Here we'll look at how long this step takes.<\/p><p>Let's start with keyboards:<\/p><div class=\"caption\"><p class=\"caption-title\">Keyboard latencies<\/p><p>Measured latencies from start of key press to signal reaching USB hub, rounded to nearest 5ms.<\/p><p>Source: <a href=\"https:\/\/danluu.com\/keyboard-latency\">Keyboard latency<\/a><\/p><\/div><div class=\"table\"><table><tr><th class=\"text left\">Keyboard<\/th>\n            <th colspan=\"2\" class=\"mixed right\">Latency (ms)<\/th>\n          <\/tr><tr><td class=\"text left first\">Apple Magic<\/td>\n            <td class=\"number middle first\">15<\/td>\n            <td class=\"bar right first\">\n          <\/td><\/tr><tr><td class=\"text left\">Das 3<\/td>\n            <td class=\"number middle\">25<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Kinesis Freestyle2<\/td>\n            <td class=\"number middle\">30<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Ergodox<\/td>\n            <td class=\"number middle\">40<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left\">Kinesis Advantage<\/td>\n            <td class=\"number middle\">50<\/td>\n            <td class=\"bar right\">\n          <\/td><\/tr><tr><td class=\"text left last\">Logitech MK360<\/td>\n            <td class=\"number middle last\">60<\/td>\n            <td class=\"bar right last\">\n          <\/td><\/tr><\/table><\/div><p>As you can see these keyboards easily take up 10s of milliseconds of latency budget on the very first step in the processing pipeline. That's out of a total budget of ~100ms or less!<small class=\"note\">\n        Typing and latency\n        See <a href=\"https:\/\/pavelfatin.com\/typing-with-pleasure\">Typing with Pleasure<\/a> for much greater detail on this topic.\n      <\/small><\/p><p>Mice can similarly introduce 10s of milliseconds of latency. Though the highest performance gaming mice will have latencies in the single digit millisecond range.<small class=\"note\">\n        Mouse click latency\n        Data for mouse click latency is a scattered.\n        <a href=\"https:\/\/www.gamersnexus.net\/guides\/2594-wireless-mouse-click-latency-analysis-vs-wired\">One example<\/a> with single digit millisecond latency.\n      <\/small><\/p><p>It's harder for us to measure the contribution of input devices specifically in mobile devices (because they're tightly integrated with other hardware components). That said, we can use a few of the common patterns in input device hardware to understand latencies in these as well as standalone devices.<\/p><h3 id=\"sample-rates\">Sample rates<\/h3><p>One common pattern is sample rates. In many input devices, the hardware \"scans\" or \"samples\" for new input on a periodic interval. For example, typical consumer touch screens sample for input at the rate of 60hz, or once every ~17ms. This means that in the worst cases input device latency will be at least ~17ms, and in the average case it can be no better than ~8ms.<\/p><p>All things being equal, higher scan rates can reduce input latency. High-end Apple mobile hardware samples touch and stylus input more frequently than 60hz, correspondingly reducing latency:<\/p><div class=\"caption\"><p class=\"caption-title\">Apple touch scan rates<\/p><p>Touch and stylus sample rates for modern Apple mobile hardware.<\/p><p>Source: <a href=\"https:\/\/developer.apple.com\/library\/archive\/documentation\/DeviceInformation\/Reference\/iOSDeviceCompatibility\/Displays\/Displays.html#\/\/apple_ref\/doc\/uid\/TP40013599-CH108-SW5\">Apple Documentation Archive<\/a><\/p><\/div><p><\/p><table><tr><th class=\"text left\">Device<\/th>\n            <th class=\"number middle\">Touch (hz)<\/th>\n            <th class=\"number right\">Stylus (hz)<\/th>\n          <\/tr><tr><td class=\"text left first\">iPhone 6\n            <\/td><td class=\"number middle first\">60\n            <\/td><td class=\"number right first\">\n          <\/td><\/tr><tr><td class=\"text left\">iPhone 7\n            <\/td><td class=\"number middle\">60\n            <\/td><td class=\"number right\">\n          <\/td><\/tr><tr><td class=\"text left\">iPhone 8\n            <\/td><td class=\"number middle\">60\n            <\/td><td class=\"number right\">\n          <\/td><\/tr><tr><td class=\"text left\">iPhone X\n            <\/td><td class=\"number middle\">120\n            <\/td><td class=\"number right\">\n          <\/td><\/tr><tr><td class=\"text left\">iPad Air 2\n            <\/td><td class=\"number middle\">60\n            <\/td><td class=\"number right\">\n          <\/td><\/tr><tr><td class=\"text left\">iPad Mini 4\n            <\/td><td class=\"number middle\">60\n            <\/td><td class=\"number right\">\n          <\/td><\/tr><tr><td class=\"text left last\">iPad Pro\n            <\/td><td class=\"number middle last\">120\n            <\/td><td class=\"number right last\">240\n          <\/td><\/tr><\/table><p>A similar source of latency is USB scanning. The USB protocol pulls input from the keyboard, so the keyboard needs to wait for the USB scan to send its key presses. Low speed USB scans at 125hz, introducing an unavoidable ~8ms max and ~4ms average delay. More recent USB versions scan at 1000hz or more, minimizing the latency impact.<\/p><p>There are many other potential sources of latency in input devices, for example debouncing in keyboards.<small class=\"note\">\n        Debouncing\n        See <a href=\"https:\/\/summivox.wordpress.com\/2016\/06\/03\/keyboard-matrix-scanning-and-debouncing\">Keyboard Matrix Scanning and Debouncing<\/a> for details on hardware and software implications of debouncing.\n      <\/small> We won't cover them all here, but emphasize the meta points that a) input devices themselves can introduce significant latency before any software processing happens and b) this may be due to multiple discrete causes that add up.<\/p><h3 id=\"displays-and-gpus\">Displays and GPUs<\/h3><p>The hardware at the other end of the pipeline are displays and graphics cards.<\/p><p>One source of latency here is the frame rate of the display. Since displays can't redraw constantly, this introduces unavoidable latency similar to the input scanning discussed above. If a screen updates (say) every 20ms, it adds 20ms of latency in the worst case and 10ms in the average case.<small class=\"note\">\n        Perceiving movement\n        Other factors influence how we perceive objects moving on screens. <a href=\"https:\/\/www.blurbusters.com\">Blur Busters<\/a> is a great resource. See e.g. <a href=\"https:\/\/www.blurbusters.com\/faq\/lcd-motion-artifacts\/\">LCD Motion Artifacts 101<\/a>.\n      <\/small><\/p><p>Most displays run at 60hz, though high end-devices and especially displays built for gaming run at 120hz, 144hz, and 240hz. Thus display frame rate alone usually contributes ~8ms average-case latency, though this can be reduced to a few milliseconds in the highest-frame-rate displays.<\/p><p>Another contribution to latency from displays is the time it takes them to physically change the color of pixels after they receive new pixel data. This time varies from low single digit milliseconds or less in high-end gaming displays to double digit milliseconds in less responsive LCDs.<small class=\"note\">\n        Display response times\n        These are hard to measure, but <a href=\"https:\/\/www.notebookcheck.net\">Notebook Check<\/a> has some illustrative data. See e.g. a <a href=\"https:\/\/www.notebookcheck.net\/Asus-ROG-Chimera-G703GI-i9-8950HK-GTX-1080-Full-HD-Laptop-Review.308366.0.html\">fast<\/a> and <a href=\"https:\/\/www.notebookcheck.net\/Google-Pixelbook-Chromebook-Review.262418.0.html\">slow<\/a> example.\n      <\/small><\/p><p>On modern high-end devices, a dedicated graphics unit (GPU) powers the display. GPUs produce the array of pixels for displays by e.g. compositing 2D windowing layers or rendering 3D virtual scenes. GPUs produce frames at a rate that depends on the GPU hardware, their interaction with application and framework code, and sometimes on synchronization logic with displays.<\/p><p>A related issue happens when application code is outright slow, and doesn't even send instructions to the GPU fast enough to take full advantage of it. This can lead to the GPU producing unique frames at a lower rate than it could if it did in fact have frequent instruction from the application. This is a common source of \"jank\" we see in 2D applications that render less than 60fps.<small class=\"note\">\n        Jank\n        \"Jank\" is hard to define, but you know it when you see it. Nathan Gitter defines it as \"visual glitches that are unexpected or distracting\" in <a href=\"https:\/\/medium.com\/@nathangitter\/designing-jank-free-apps-9f66d43b9c87\">Designing Jank-Free Apps<\/a>.\n      <\/small><\/p><h3 id=\"cycle-stacking\">Cycle stacking<\/h3><p>We've discussed at least three parts of the pipeline where latency accrues due to periodic activity: input scanning, GPU rendering loops, and display refresh cycles. It's important to note that these can stack in ways that essentially adds all of their latency together:<\/p><div class=\"caption\"><p class=\"caption-title\">Waiting for multiple cycles<\/p><p>A hypothetical latency cascade shows how waiting for successive hardware cycles can accumulate latency.<\/p><p>Dashed vertical lines indicate cycles the pipeline needs to wait for.<\/p><\/div><img src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/cycle-stacking.png\" alt=\"image\" \/><p>In order to move to the next step in a pipeline, we need to wait for the next cycle of that stage to come along. And the cycles may not be aligned. Misaligned cycles and an unfavorable initial input time can cause 10s of milliseconds of additional latency, a large amount relative to the latency budgets discussed above.<\/p><p>On the software side, runtime overhead is a catch-all for overhead from the operating system and other non-application code. We'll look at two important examples: garbage collection and scheduling.<\/p><p>First up is garbage collection (GC). GC is critical in the two most widely-used platforms in the world\u2014the web (JavaScript) and Android (Java).<\/p><p>GC can contribute a lot of latency at specific instances, especially relative to requirements for fast input latency. GCs on the order of 10ms wouldn't be surprising for JavaScript or Java runtimes. But that's the entire budget we have for dragging objects on a touch screen!<small class=\"note\">\n        GC latency impact\n        A GC may delay ~1 frame, not all frames. But like \"jank\" from missed frames, latency jitters are noticeable and annoying to users.\n      <\/small><\/p><p>There are ways to mitigate GC-induced latency. These include moving as much GC work as possible off of the main thread and optimizing the GC to require only small individual pauses. One can also use a language that trades off some of the convenience of GC for more predictable performance.<small class=\"note\">\n        GC latency mitigation\n        See for example the <a href=\"https:\/\/v8project.blogspot.com\/2016\/04\/jank-busters-part-two-orinoco.html\">V8 effort<\/a> to move more GC off the main thread and the <a href=\"https:\/\/blog.golang.org\/ismmkeynote\">Go work<\/a> to bring max GC pauses well below 1ms. Languages like Swift avoid arbitrary GC by using <a href=\"https:\/\/docs.swift.org\/swift-book\/LanguageGuide\/AutomaticReferenceCounting.html\">automatic reference counting<\/a>.\n      <\/small><\/p><p>Another potential source of overhead is operating system scheduling. Our application (and its dependencies in the OS) are not necessarily running all the time. Other programs may be scheduled in while ours is paused, even if for a very short time.<\/p><p>If we have a wall-clock budget of 10ms and our app takes 10ms \"end-to-end\" without accounting for any other programs running, we may well exceed our budget in wall clock time due to OS scheduling.<small class=\"note\">\n        CPU usage\n        Related to the scheduling issue is CPU usage. If your app meets its performance goals but requires ~100% of the CPU to do that, it may well annoy users. You'll drain batteries faster, make the device hotter, and perhaps trigger a noisy fan. In other words, even with user-facing performance held constant, lower CPU usage is a better user experience.\n      <\/small><\/p><p>Every program will ask for their time slices, and there are only so many CPU cores to go around.<\/p><h3 id=\"latency-by-design\">Latency by design<\/h3><p>A common source of latency on mobile interfaces is the design of the OS and apps themselves. There are some important interactions that can only be accomplished by literally waiting.<\/p><p>Android and iOS both make substantial use of \"long press\" to access context menus, which require that the user wait hundreds of milliseconds in the middle of their command gestures.<\/p><p>A related source is delays for disambiguation. For example, on mobile Safari there's a default 350ms delay between when the user taps a link and when the browser begins fetching the new page, in order to tell the difference between a link click and a double-tap zoom.<small class=\"note\">\n        iOS tap delays\n        See <a href=\"https:\/\/webkit.org\/blog\/5610\/more-responsive-tapping-on-ios\/\">More Response Tapping on iOS<\/a> for background on this delay as well as recent changes that allow application developers to work around the issue.\n      <\/small><\/p><h3 id=\"user-hostile-work\">User-hostile work<\/h3><p>A major source of latency for users on the web is user-hostile work, such as downloading trackers that surveil user activity and loading intrusive ads.<\/p><div class=\"caption\"><p class=\"caption-title\">Loading a news article<\/p><p>A ~500 word article on the Washington Post website needed 100s of HTTP requests and ~4400ms. Many of the requests are for surveillance and advertising. A small selection of the requests are shown.<\/p><\/div><img src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/washington-post-profile.png\" alt=\"image\" \/><p>There are many great articles on web bloat.<small class=\"note\">\n        Web bloat\n        See e.g. <a href=\"https:\/\/pxlnv.com\/blog\/bullshit-web\/\">The Bullshit Web<\/a>, <a href=\"http:\/\/idlewords.com\/talks\/website_obesity.htm\">The Website Obesity Crisis<\/a>, <a href=\"https:\/\/danluu.com\/web-bloat\/\">Web bloat<\/a>\n      <\/small> We'll just emphasize that the single biggest source of latency on many sites is downloading stuff that users don't want.<\/p><h3 id=\"application-code\">Application code<\/h3><p>The last source of latency we'll mention is perhaps the most obvious: the application. If the app spends a lot of CPU time processing a given input, or executing some action, it will be slow.<\/p><h3 id=\"putting-it-together\">Putting it together<\/h3><p>Let's look at an example of how latency can add up:<\/p><div class=\"caption\"><p class=\"caption-title\">Latency waterfall example<\/p><p>A hypothetical example of end-to-end latency from input to display.<\/p><p>Dashed vertical lines indicate cycles the pipeline needs to wait for.<\/p><\/div><img src=\"https:\/\/www.inkandswitch.com\/media\/slow-software\/input-latency-cascade.png\" alt=\"image\" \/><p>The example above is hypothetical but illustrative. It shows how there are a lot of layers that add latency and an application can exhibit high latency (much greater than frame length) even if it successfully runs at at the full frame rate.<\/p><h2 id=\"toward-fast-software\">Toward fast software<\/h2><p>There is a deep stack of technology that makes a modern computer interface respond to a user's requests. Even something as simple as pressing a key on a keyboard and having the corresponding character appear in a text input box traverses a lengthy, complex gauntlet of steps, from the scan rate of the keyboard, through the OS and framework processing layers, through the graphics card rendering and display refresh rate.<\/p><p>There is reason for this complexity, and yet we feel sad that computer users trying to be productive with these devices are so often left waiting, watching spinners, or even just with the slight but still perceptible sense that their devices simply can't keep up with them.<\/p><p>We believe fast software empowers users and makes them more productive. We know today's software often lets users down by being slow, and we want to do better. We hope this material is helpful for you as you work on your own software.<\/p><p>If you're interested in the topic of fast software, or have feedback on this article, we'd love to hear from you: <a href=\"https:\/\/twitter.com\/inkandswitch\">@inkandswitch<\/a> or <a href=\"mailto:hello@inkandswitch.com\">hello@inkandswitch.com<\/a><\/p><h2 id=\"references\">References<\/h2><p>\n        <a class=\"ref-number\" id=\"ref-1\" href=\"#ret-1\">[1]<\/a>\n        Y. Endo, Z. Wang, J. Chen, and M. Seltzer. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=10535754908454057825\">Using Latency to Evaluate Interactive System Performance<\/a>,\" In Proceedings of the USENIX 2nd Symposium on Operating Systems Design and Implementation, 1996.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-2\" href=\"#ret-2\">[2]<\/a>\n        A. Ng, M. Annett, P. Dietz, A. Gupta, and W. Bischof. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=8942755275099535739\">In the Blink of an Eye: Investigating Latency Perception During Stylus Interaction<\/a>,\" In Proceedings of the 32nd Annual ACM Conference on Human Factors in Computing Systems, 2014.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-3\" href=\"#ret-3\">[3]<\/a>\n        A. Ng, J. Lepinski, D. Wigdor, S. Sanders, and P. Dietz. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=17905801937419740410\">Designing for low-latency direct-touch input<\/a>,\" In Proceedings of the 25th Annual ACM Symposium on User Interface Software and Technology, 2012.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-4\" href=\"#ret-4\">[4]<\/a>\n        J. Deber, and R. Jota, C. Forlines, and D. Wigdor. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=16992791481343673285\">How Much Faster is Fast Enough?: User Perception of Latency &amp; Latency Improvements in Direct and Indirect Touch<\/a>,\" In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, 2015.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-5\" href=\"#ret-5\">[5]<\/a>\n        M. Annett, A. Ng, P. Dietz, W. Bischof, and A. Gupta. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=15159574178009198437\">How Low Should We Go?: Understanding the Perception of Latency While Inking<\/a>,\" In Proceedings of Graphics Interface 2014, 2014.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-6\" href=\"#ret-6\">[6]<\/a>\n        M. Annett, A. Ng, P. Dietz, W. Bischof, and A. Gupta. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=15159574178009198437\">How Low Should We Go?: Understanding the Perception of Latency While Inking<\/a>,\" In Proceedings of Graphics Interface 2014, 2014.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-7\" href=\"#ret-7\">[7]<\/a>\n        V. Forch, T. Franke, N. Rauh, and J. Krems. \"<a href=\"https:\/\/scholar.google.com\/scholar?q=Are+100+ms+Fast+Enough%3F+Characterizing+Latency+Perception+Thresholds+in+Mouse-Based+Interaction\">Are 100 ms Fast Enough? Characterizing Latency Perception Thresholds in Mouse-Based Interaction<\/a>,\" In Engineering Psychology and Cognitive Ergonomics: Cognition and Design, 2017.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-8\" href=\"#ret-8\">[8]<\/a>\n        J. Dabrowski and E. V. Munson. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=15646007677492381636\">40 years of searching for the best computer system response time<\/a>,\" Interacting with Computers 23.5, 2011.\n      <\/p><p>\n        <a class=\"ref-number\" id=\"ref-9\" href=\"#ret-9\">[9]<\/a>\n        M. Barreda-\u00c1ngeles, I. Arapakis, X. Bai, B. Cambazoglu, A. Pereda. \"<a href=\"https:\/\/scholar.google.com\/scholar?cluster=3612836859261052030\">Unconscious Physiological Effects of Search Latency on Users and Their Click Behaviour<\/a>,\" In 38th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2015.\n      <\/p>","created_at":"2018-11-22T02:45:16+0000","updated_at":"2018-11-22T02:45:16+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":20,"domain_name":"www.inkandswitch.com","preview_picture":"https:\/\/www.inkandswitch.com\/media\/slow-software\/input-latency-cascade.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["css","typography","fonts"],"is_public":false,"id":701,"title":"8 Definitive Web Font Stacks Article","url":"https:\/\/www.sitepoint.com\/eight-definitive-font-stacks\/","content":"<p><strong><em>\u201cWeb design is 95% typography.\u201d<\/em> \u2013 <a class=\"sublink\" href=\"http:\/\/informationarchitects.jp\/the-web-is-all-about-typography-period\/\">Oliver Reichenstein<\/a><\/strong><\/p><p><strong><em>\u201cTypography is not \u2018picking a cool font.'\u201d<\/em> \u2013 <a class=\"sublink\" href=\"http:\/\/files.jeffcroft.com\/presentations\/fowd_april_2007\/JeffCroft_FOWD_Workshop_Elegant_Web_Typography.pdf\">Jeff Croft<\/a><\/strong><\/p><p><strong><em>\u201cTypography is \u2026 about shaping text for optimal reading experience.\u201d<\/em> \u2013 <a class=\"sublink\" href=\"http:\/\/informationarchitects.jp\/webdesign-is-95-typography-partii\/\">Oliver Reichenstein<\/a><\/strong><\/p><div class=\"widget maestro maestro-content-type-html hide-for-tablet-SP hide-for-mobile-SP\" id=\"maestro-636\"><p>--ADVERTISEMENT--<\/p><\/div><p>Over the last few months, I\u2019ve spent more time than I intended on exploring the whole idea of fonts and typography for the Web. (My friend, typography expert <a class=\"sublink\" href=\"http:\/\/klepas.org\/\">Simon Pascal Klein<\/a>, writes, \u201cThe former is a stylized set of glyphs of characters,\u201d while \u201cthe other [is] the whole art of creating type and setting it into the written word.\u201d For more clarification and illumination, consult <a class=\"sublink\" href=\"http:\/\/jontangerine.com\/log\/2008\/08\/typeface--font\">Jon Tan<\/a> and <a class=\"sublink\" href=\"http:\/\/www.typophile.com\/node\/14701#comment-84393\">Mark Simonson<\/a>.)<\/p><p>In the process, I\u2019ve been considering the idea of <strong>font stacks<\/strong> \u2013 using the well-known <code>font-family<\/code> CSS property \u2013 to list as many different fonts as possible in order to optimize the web site experience for a maximum number of users.<\/p><h5>What About User Preferences?<\/h5><p>There\u2019s a strong argument for leaving well enough alone, and simply specifying serif or sans serif fonts, thus letting the user\u2019s own settings determine the font display. I can understand this philosophy, but personally I prefer exerting more control over the display of my sites.<\/p><p>Fonts matter more than we may realize; they\u2019re as important a choice in determining the visual impact and informational flow of a web page as the color scheme or the navigation layout. And as a designer (even an admitted amateur), I\u2019m unable to leave these crucial elements to whatever settings a user may have.<\/p><h5>Different Categories of Fonts<\/h5><p>Of course, there are more distinctions among typefaces than just the presence of a serif. The traditional categories, especially for web use, are: <strong>serif<\/strong>, <strong>sans serif<\/strong>, <strong>monospace<\/strong>, <strong>cursive\/script<\/strong>, and <strong>decorative<\/strong> or \u201cfantasy\u201d fonts, which are useful primarily in snazzy graphic creations.<\/p><p><a class=\"sublink\" href=\"http:\/\/www.danielmall.com\/archives\/2005\/09\/01\/a_man_of_many_faces.php\">Daniel Mall<\/a> has a useful list of categories for his typeface collection, including <em>pixel<\/em> (the really little guys), <em>symbol<\/em> (Wingdings), and <em>blackletter<\/em> (used in medieval manuscripts and heavy metal bands). I list these because a lot of very capable web designers are simply unaware of what exists beyond serif, sans serif, and monospace.<\/p><p>In fact, this is how I also viewed typography for years, until I came to understand that the more you learn about typography as a designer, the more informed choices you can make, and the better your sites will be. Rather than an afterthought, your typographical decisions should be at the forefront of your design, navigation, and structural choices in any site you design. The difference good typography makes in a site are often subtle, but equally profound.<\/p><h5>Expanding Your Typographic Options<\/h5><p>The idea behind creating recommended font stacks is simple: since most web designers don\u2019t know a great deal about font selection and typography for the Web, and have far too much on their plates to spend the time needed to learn, they need a one-stop shop of font stacks that provides a wide variety for all platforms \u2013 Windows, Mac and Linux.<\/p><p>Additionally, those designers who <em>do<\/em> understand enough about typography may feel like their creativity is limited by the restrictions of the standard \u201cweb safe\u201d set of fonts. Using font stacks is one possible way of increasing a designer\u2019s options.<\/p><p>The font stacks listed here are grouped together by the <em>universal<\/em> font that forms the base of that stack. A designer can therefore decide on a typographical look for their site, grab the appropriate font stack, and tweak it to suit their needs.<\/p><p>I hope that readers use this as a springboard for their own typographical research and experimentation, developing their own stacks to suit themselves and the needs of their site users.<\/p><h5>Introducing the Stacks<\/h5><p>Eight font stacks are listed, combining Windows, Mac, Linux, and Adobe fonts in groups that are <em>relatively<\/em> similar among themselves. Each stack takes the following format:<\/p><p><code>exact font, nearest alternative, platform-wide alternative(s), universal (cross-platform) choice(s), generic<\/code><\/p><p>A second structure is also listed \u2013 one that will often conflict with the first structure:<\/p><p><code>Windows fonts, Mac fonts, Linux fonts, universal, generic<\/code><\/p><p>In this article I\u2019ve listed three serif stacks, four sans serif stacks, and a monospace stack. Of course, the idea that these might be definitive is laughable, the title of this article notwithstanding. They are merely a starting point for experimentation. I\u2019m happy to update this article with changes if there\u2019s a strong enough argument to do so, and welcome your feedback in the comments.<\/p><p>Here are the font stacks as they currently stand. Copy and paste as you will.<\/p><p><strong>The Times New Roman-based serif stack:<\/strong><\/p><p><code>font-family: Cambria, \"Hoefler Text\", Utopia, \"Liberation Serif\", \"Nimbus Roman No9 L Regular\", Times, \"Times New Roman\", serif;<\/code><\/p><p><img src=\"http:\/\/i2.sitepoint.com\/graphics\/1697-tnrstack.thumb.jpg\" alt=\"\" width=\"400\" height=\"200\" \/><\/p><p><strong>A modern Georgia-based serif stack:<\/strong><\/p><p><code>font-family: Constantia, \"Lucida Bright\", Lucidabright, \"Lucida Serif\", Lucida, \"DejaVu Serif\", \"Bitstream Vera Serif\", \"Liberation Serif\", Georgia, serif;<\/code><\/p><p><img src=\"http:\/\/i2.sitepoint.com\/graphics\/1697-georgiastack.thumb.jpg\" alt=\"\" width=\"400\" height=\"215\" \/><\/p><p><strong>Go to page:<\/strong> <a href=\"https:\/\/www.sitepoint.com\/eight-definitive-font-stacks\">1<\/a> | <a href=\"https:\/\/www.sitepoint.com\/eight-definitive-font-stacks-2\/\">2<\/a> | <a href=\"https:\/\/www.sitepoint.com\/eight-definitive-font-stacks-3\/\">3<\/a><\/p>","created_at":"2018-11-20T21:40:54+0000","updated_at":"2018-11-20T21:41:01+0000","published_at":"2009-01-07T14:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"www.sitepoint.com","preview_picture":"https:\/\/www.sitepoint.com\/wp-content\/themes\/sitepoint\/assets\/images\/icon.html-css.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["business","marketing","startup"],"is_public":false,"id":700,"title":"Product Hunt 101: How To Launch Your Product From Early Idea To Revenue","url":"https:\/\/medium.com\/swlh\/product-hunt-101-how-to-launch-your-product-from-early-idea-to-revenue-c3f01864cdde","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"7b19\" class=\"graf graf--h3 graf--leading graf--title\">Product Hunt 101: How To Launch Your Product From Early Idea To\u00a0Revenue<\/h1><h2 id=\"e6cc\" class=\"graf graf--h4 graf-after--h3 graf--subtitle\">The Step by Step Process of How I Launched <a href=\"https:\/\/historysearch.com\" data-href=\"https:\/\/historysearch.com\" class=\"markup--anchor markup--h4-anchor\" rel=\"nofollow noopener\" target=\"_blank\">History Search<\/a> on Product Hunt and Got 1000 Sign Ups in 2\u00a0Days<\/h2><\/div><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"8a7d\" class=\"graf graf--figure graf--layoutFillWidth graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*m0UO4banJrJaLR5UciM1XQ.png\" data-width=\"2501\" data-height=\"1033\" data-is-featured=\"true\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*m0UO4banJrJaLR5UciM1XQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"17d3\" class=\"graf graf--p graf-after--figure\">3, 2, 1, launch\u2026 right? Not quite. When you\u2019re getting ready to launch on Product Hunt, you\u2019ve probably put months or even years of your life into this product. Failing the launch would be devastating, while doing it right can give you everything you need to turn your product into a business.<\/p><p id=\"bff0\" class=\"graf graf--p graf-after--p\">This article isn\u2019t focused on getting upvotes, but on maximising the impact of your launch in general. By focusing on impact you\u2019ll get the fundamentals that matter most for your long-term success.<\/p><p id=\"10ee\" class=\"graf graf--p graf-after--p\">Some background: I always felt annoyed with how hard it was to keep track of the thousands of webpages we come across every month. So around 4 years ago I got the idea to create a filesystem for the web. It\u2019s similar to the finder on a computer, but indexes text on webpages you visit. This way you can use any word on the page to retrieve it\u2014saving you a bunch of time every-time you need a page again.<\/p><figure id=\"c576\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*qQ2drr2_2wtD0mSTdz6AuQ.gif\" data-width=\"1274\" data-height=\"712\" data-action=\"zoom\" data-action-value=\"1*qQ2drr2_2wtD0mSTdz6AuQ.gif\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*qQ2drr2_2wtD0mSTdz6AuQ.gif\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"b23a\" class=\"graf graf--p graf-after--figure\">Prior to our launch I\u2019d spent around 1 year beta-testing. You don\u2019t need Product Hunt to get your first users. On the contrary, you actually need a base of beta users to launch successfully. We had around 3500 beta users, which I got through contacting people in Facebook groups about Product Hunt. I used a simple message:<\/p><\/div><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"f0c6\" class=\"graf graf--figure graf--layoutFillWidth graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*rxrLp5Hw_JMHLO3qoZ8how.png\" data-width=\"2233\" data-height=\"394\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*rxrLp5Hw_JMHLO3qoZ8how.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"e0de\" class=\"graf graf--p graf-after--figure\">The main benefit to this approach is that you can literally contact thousands of people who you know are interested in or are working on new products\u200a\u2014\u200asince they\u2019re in Product (Hunt) groups. You\u2019ll be surprised how many people are interested.<\/p><p id=\"f2ce\" class=\"graf graf--p graf-after--p\">When someone replied, they typically asked for more information, so I made sure to give them a very brief summary of what <a href=\"https:\/\/historysearch.com\" data-href=\"https:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> was.<\/p><\/div><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"472b\" class=\"graf graf--figure graf--layoutFillWidth graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*p7trWiSINpMF1JTD89XhbA.png\" data-width=\"2231\" data-height=\"240\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*p7trWiSINpMF1JTD89XhbA.png\" alt=\"image\" \/><\/div><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"d3c6\" class=\"graf graf--p graf-after--figure\">This was actually the very first version of our value prop and sales pitch. I then continued the conversation with those willing\/interested\u200a\u2014\u200aand used this information to improve the product and \u201cstory\u201d. Getting your story right is going to be critical during your launch.<\/p><p id=\"8eaf\" class=\"graf graf--p graf-after--p\">This concludes the section about building an early-beta user audience since this article is really aimed at the next stage: leveraging your Product Hunt launch to turn your work into a business (getting reoccurring users and revenue).<\/p><p id=\"c4cd\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Groups that worked for me<\/em><\/strong><\/p><blockquote id=\"68e2\" class=\"graf graf--blockquote graf-after--p\"><a href=\"https:\/\/www.facebook.com\/groups\/SaaSgrowthhacking\/\" data-href=\"https:\/\/www.facebook.com\/groups\/SaaSgrowthhacking\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">SaaS Growth Hacks<\/a><\/blockquote><blockquote id=\"142d\" class=\"graf graf--blockquote graf-after--blockquote\"><a href=\"https:\/\/www.facebook.com\/groups\/Pushstarter\/\" data-href=\"https:\/\/www.facebook.com\/groups\/Pushstarter\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Pushstart<\/a><\/blockquote><blockquote id=\"505f\" class=\"graf graf--blockquote graf-after--blockquote\"><a href=\"https:\/\/www.facebook.com\/groups\/saasmantra\/\" data-href=\"https:\/\/www.facebook.com\/groups\/saasmantra\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Saas Mantra<\/a><\/blockquote><blockquote id=\"c970\" class=\"graf graf--blockquote graf-after--blockquote\"><a href=\"https:\/\/www.facebook.com\/groups\/startupproductlaunches\/\" data-href=\"https:\/\/www.facebook.com\/groups\/startupproductlaunches\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Startup Product Launches<\/a><\/blockquote><blockquote id=\"26c6\" class=\"graf graf--blockquote graf-after--blockquote\"><a href=\"https:\/\/www.facebook.com\/groups\/producthunt\/\" data-href=\"https:\/\/www.facebook.com\/groups\/producthunt\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Product Hunt<\/a><\/blockquote><blockquote id=\"f230\" class=\"graf graf--blockquote graf-after--blockquote\"><a href=\"https:\/\/www.facebook.com\/groups\/saasproductsandmarketing\/\" data-href=\"https:\/\/www.facebook.com\/groups\/saasproductsandmarketing\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">SaaS Product &amp; Marketing<\/a><\/blockquote><h3 id=\"1bdd\" class=\"graf graf--h3 graf-after--blockquote\">What is Product Hunt About\u00a0Then?<\/h3><p id=\"6e02\" class=\"graf graf--p graf-after--h3\"><em class=\"markup--em markup--p-em\">Product Hunt is a community of tech enthusiasts and product makers. Every day new products are hunted and receive upvotes. The top 5 most upvoted products get blasted in a daily newsletter to +500k members the next day.<\/em><\/p><p id=\"2a23\" class=\"graf graf--p graf-after--p\">After spending countless days, nights and dollars on <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> the main thing I needed to know was \u201cdo people really need History Search?\u201d and were they willing to pay for it so I can continue working on it\u200a\u2014\u200aor will my hard work go in vain? Because that\u2019s what your first users are doing in reality: they\u2019re paying so that you can continue to work on your vision.<\/p><blockquote id=\"8e06\" class=\"graf graf--pullquote graf-after--p\"><div>Your first users aren\u2019t paying for your product\u200a\u2014\u200athey pay for you to continue working on your\u00a0vision<\/div><\/blockquote><p id=\"f0a5\" class=\"graf graf--p graf-after--pullquote\">Obviously it\u2019s better to finish as the top product of the day, since this will get your product in both the daily and weekly newsletter. However it\u2019s not all about coming in first during the initial launch of your product, it\u2019s about validation and building an engaged users base\u2014who are prepared to pay for your product. So as long as you get in the top 5 you\u2019ll be in the daily newsletter and you\u2019re fine in regards to reach.<\/p><h4 id=\"9b7a\" class=\"graf graf--h4 graf-after--p\">What we got from launching<\/h4><ul class=\"postList\"><li id=\"4502\" class=\"graf graf--li graf-after--h4\">Our average daily site traffic jumped from 156 users to around 2000 on launch day, and 2500 the day after, thanks to our<a href=\"https:\/\/www.cnet.com\/news\/history-search-browser-extension-keeps-archive-websites-you-visit\/\" data-href=\"https:\/\/www.cnet.com\/news\/history-search-browser-extension-keeps-archive-websites-you-visit\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\"> coverage in CNET<\/a> (more on getting news coverage later).<\/li><li id=\"3ddf\" class=\"graf graf--li graf-after--li\">We got 1000 new signups during launch day and the following day. We tripled our daily organic growth after the launch spike settled\u200a\u2014\u200afrom 10 before launch to 35 daily signups after the launch.<\/li><li id=\"d3ce\" class=\"graf graf--li graf-after--li\">On the launch day itself we had 30\u201340 paid conversions (we offered a 25% discount), with approximately 2 new paying customers daily since the launch.<\/li><li id=\"e162\" class=\"graf graf--li graf-after--li\">My<a href=\"https:\/\/www.producthunt.com\/@martijn_verbove\" data-href=\"https:\/\/www.producthunt.com\/@martijn_verbove\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\"> Product Hunt<\/a> followers increased significantly from less than 500 to almost 1000 (although I don\u2019t know if this will contribute to future launches in any way)<\/li><li id=\"72f2\" class=\"graf graf--li graf-after--li\">We currently have approximately 650 daily active users, which has been increasing about 8\u201315% per day since the launch.<\/li><li id=\"eb1d\" class=\"graf graf--li graf-after--li\">Our current conversion rate of paying users is around 5.8%, up from 3.5% on launch day.<\/li><\/ul><h3 id=\"d85e\" class=\"graf graf--h3 graf-after--li\">How Do You Know What\u2019s the Right Time to Launch on Product\u00a0Hunt?<\/h3><p id=\"60ca\" class=\"graf graf--p graf-after--h3\">My main piece of advice about deciding when to launch is that it\u2019s not about being done, but about being ready. The reason is simple: Your product is never going to be perfect when you launch it.<\/p><p id=\"59a5\" class=\"graf graf--p graf-after--p\">There are so many things in <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> that I want to improve. For example, I really dislike the way we onboard new users. I think it\u2019s really tricky and confusing. But instead of iterating on it for another 2 months we launched and got it done in one day.<\/p><p id=\"3c06\" class=\"graf graf--p graf-after--p\">When you launch the right way (without focusing on upvotes), you\u2019ll get more questions, suggestions and ideas than you can ever imagine. In our case this gave us the input we needed to create an onboarding experience that answered the most common questions.<\/p><p id=\"8fbe\" class=\"graf graf--p graf-after--p\">When you talk to people, you see it fall together. All of these things come up because you actually have <strong class=\"markup--strong markup--p-strong\">one day <\/strong>where you can talk to hundreds of people. In a way you have hundreds of minds working with you\u200a\u2014\u200athey tell you what\u2019s unclear about your product or message, what they love about it, and even ideas or use-cases you never thought of yourself. This doesn\u2019t mean you have to then develop everything people ask, more on that later.<\/p><h3 id=\"a0ac\" class=\"graf graf--h3 graf-after--p\">Get Your Story\u00a0Straight<\/h3><p id=\"c468\" class=\"graf graf--p graf-after--h3\">Before even thinking about launching, make sure you\u2019ve thoroughly thought about how to position your product. Why did we switch from \u201cYour Personal Search engine\u201d to \u201cYour Filesystem for the web\u201d? This was because \u201csearch engine\u201d leads many people to believe it was only public pages, and so made it seem like Google. While in reality <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> does technically do the same as Google, but it indexes the web pages you visit through an integration with your browser (via browser extensions)\u200a\u2014\u200aso this also makes it possible to find pages behind logins.<\/p><h4 id=\"6280\" class=\"graf graf--h4 graf-after--p\">What is it vs. What does it\u00a0do?<\/h4><p id=\"022b\" class=\"graf graf--p graf-after--h4\"><strong class=\"markup--strong markup--p-strong\">What it is <\/strong>and <strong class=\"markup--strong markup--p-strong\">What it does<\/strong> are two fundamentally different approaches to positioning, and you need both. In our case we have:<\/p><ul class=\"postList\"><li id=\"8044\" class=\"graf graf--li graf-after--p\">Your Filesystem for the web<\/li><li id=\"200f\" class=\"graf graf--li graf-after--li\">Your way back to a million websites<\/li><\/ul><p id=\"feb4\" class=\"graf graf--p graf-after--li\">It\u2019s important not to run this past too many people, since it will send you in a different direction every time. The main thing I learned is that your product is whatever you say it is\u200a\u2014\u200afor example. Is Uber a taxi company or is it the future of transportation?<\/p><h4 id=\"35a6\" class=\"graf graf--h4 graf-after--p\">Telling your\u00a0story<\/h4><p id=\"f21d\" class=\"graf graf--p graf-after--h4\">Your first launch isn\u2019t really about your product, it\u2019s about your story and vision. The story about how you got to where you are today is important to write down, not only for the outside world but also for yourself. You\u2019ll be able to better align your pitch if you know your story well. This is exactly what I did, and even though I spent about two weeks hesitating whether to publish it or not, I eventually did. Guess what?! It was one of the best decisions I could have made\u200a\u2014\u200apeople loved it and it even got published by The Startup. <a href=\"https:\/\/medium.com\/swlh\/moments-before-success-or-failure-d0f50ff148ea\" data-href=\"https:\/\/medium.com\/swlh\/moments-before-success-or-failure-d0f50ff148ea\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Here\u2019s a link to the article<\/a>.<\/p><blockquote id=\"7686\" class=\"graf graf--pullquote graf-after--p\"><div>Market Fit = Your vision + People who believe in\u00a0it<\/div><\/blockquote><p id=\"8cbc\" class=\"graf graf--p graf-after--pullquote graf--trailing\">Your product is never going to be perfect when you launch it. With <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a>, there are so many things that I want to improve. But at first it\u2019s all about finding the people who support you and believe in your vision.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"f83a\" class=\"graf graf--h3 graf--leading\">Getting Launch\u00a0Ready<\/h3><p id=\"5f8c\" class=\"graf graf--p graf-after--h3\"><strong class=\"markup--strong markup--p-strong\">You need a clear message. <\/strong>Our general tagline for <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> is \u201cyour way back to a million websites.\u201d However, we deliberately decided not to use this message for Product Hunt, because we knew that it\u2019s a more technical audience.<\/p><p id=\"6d2e\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Your tagline has to be just right<\/strong>. It\u2019s tempting to choose a funny tagline, and this might get you a lot of upvotes but:<\/p><ol class=\"postList\"><li id=\"bd08\" class=\"graf graf--li graf-after--p\">Upvotes don\u2019t give you anything unless they are from the people who are really interested in what you are offering.<\/li><li id=\"b356\" class=\"graf graf--li graf-after--li\">If you do get conversions based on your funny or attractive tagline, you\u2019ll have a very high churn\u200a\u2014\u200apeople will love it and think it\u2019s cool for a day, but think about it like an Instagram photo. You get a ton of likes but no sustained success. That\u2019s how it goes with funny products on Product Hunt.<\/li><\/ol><p id=\"3e51\" class=\"graf graf--p graf-after--li\"><strong class=\"markup--strong markup--p-strong\">You need an existing user base. <\/strong>It is very important that when you launch, you already have an existing user base. It doesn\u2019t have to be huge, but you have to have people who really know the product already. This is critical for two reasons. You need:<\/p><ol class=\"postList\"><li id=\"f2a7\" class=\"graf graf--li graf-after--p\">Their feedback.<\/li><li id=\"c992\" class=\"graf graf--li graf-after--li\">Their support.<\/li><\/ol><p id=\"3085\" class=\"graf graf--p graf-after--li\"><strong class=\"markup--strong markup--p-strong\">You need support. <\/strong>Before you launch, sit down and think: Who can I genuinely ask for support to get the word of my launch out on this day? If that list of people is very small, it\u2019s going to be difficult to launch. Because all of this is not just a way to get upvotes, it\u2019s to get reach. And to get reach, you need some support. So you need a group of people who really believe and support the thing you\u2019re doing.<\/p><p id=\"0237\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">You need a plan for getting press. <\/strong>This is extremely difficult for people who are launching for the first time. If you read an article about how to reach out to press, most of it is incorrect. Most of it is written by journalists, and they assume that the best way is by writing a custom email\u200a\u2014\u200awhich may work for larger brands, but not startups.<\/p><p id=\"cdbc\" class=\"graf graf--p graf-after--p\">When preparing for the <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> launch, I wrote custom emails to almost 20 people. I spent 2.5 weeks researching.<\/p><p id=\"432e\" class=\"graf graf--p graf-after--p\">How many journalists replied? One. And he just replied to say \u201cI don\u2019t write about these types of services anymore.\u201d<\/p><p id=\"90f0\" class=\"graf graf--p graf-after--p\">So you need to keep some things in mind when you\u2019re contacting journalists.<\/p><ol class=\"postList\"><li id=\"f6b6\" class=\"graf graf--li graf-after--p\">When you\u2019re writing to a journalist, <strong class=\"markup--strong markup--li-strong\">it\u2019s not about you, it\u2019s about the story.<\/strong> Why should their audience be interested in reading this? So you need to shift your perspective.<\/li><li id=\"86d5\" class=\"graf graf--li graf-after--li\">When you do reach out to the press, you DO need to mass email. But not in the blast way that most people do. Don\u2019t just buy a directory that you can buy on the web. You need a curated list. Analyze the specific people that you\u2019re sending your pitch\u200a\u2014\u200astory\u200a\u2014\u200ato.<\/li><li id=\"97b8\" class=\"graf graf--li graf-after--li\">Send them a link to your pressroom\u200a\u2014\u200adon\u2019t send them a press release. You can see ours here: <a href=\"https:\/\/historysearch.pr.co\" data-href=\"https:\/\/historysearch.pr.co\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">historysearch.pr.co<\/a>.<\/li><\/ol><p id=\"f000\" class=\"graf graf--p graf-after--li\">Send the story in the email with the press information 3\u20135 days before launching. This is the email I used:<\/p><figure id=\"0f1f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*xXE4XT2SDv2K5gnr\" data-width=\"1280\" data-height=\"934\" data-action=\"zoom\" data-action-value=\"0*xXE4XT2SDv2K5gnr\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*xXE4XT2SDv2K5gnr\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"f854\" class=\"graf graf--p graf-after--figure\">If journalists don\u2019t reply to your email 3\u20135 days after you pitch, it doesn\u2019t mean they\u2019re not going to cover you. I recommend sending a follow-up email in the evening of your launch\u200a\u2014\u200anot the same pitch\u200a\u2014\u200awith the result. You could even do it the next day so that you don\u2019t bombard the journalist with the same press release over and over.<\/p><p id=\"954d\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Don\u2019t be worried if you don\u2019t get any responses. <\/strong>Our story just appeared on CNET but they had never replied. I discovered the story because I couldn\u2019t figure out why our traffic was still increasing after the launch (usually it drops) and I used <a href=\"https:\/\/mention.com\" data-href=\"https:\/\/mention.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Mention<\/a> to find the article on <a href=\"https:\/\/www.cnet.com\/news\/history-search-browser-extension-keeps-archive-websites-you-visit\" data-href=\"https:\/\/www.cnet.com\/news\/history-search-browser-extension-keeps-archive-websites-you-visit\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">CNET<\/a>.<\/p><figure id=\"74d7\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*fa8vpobk4AQclv4C\" data-width=\"1266\" data-height=\"1110\" data-action=\"zoom\" data-action-value=\"0*fa8vpobk4AQclv4C\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*fa8vpobk4AQclv4C\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"d18e\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">Keep in mind: <\/strong>If journalists from publications like CNET, Mashable, etc, <em class=\"markup--em markup--p-em\">are <\/em>willing to write about you. Then they are helping you, you are <em class=\"markup--em markup--p-em\">not <\/em>helping them. I am extremely grateful that Stephen Shankland, the CNET journalist, decided, based on the story I wrote, to write about <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a>. He can write about a billion other things, and he chose us.<\/p><p id=\"50b2\" class=\"graf graf--p graf-after--p\">If you\u2019re lost about what to write in your emails to journalists, here\u2019s the basic formula for a pitch. I should actually say story and not pitch.<\/p><ol class=\"postList\"><li id=\"7f59\" class=\"graf graf--li graf-after--p\">Start with where you\u2019re coming from. How long have you been working on this? What is it?<\/li><li id=\"9be0\" class=\"graf graf--li graf-after--li\">The effort that you needed to invest to get where you are today.<\/li><li id=\"4c27\" class=\"graf graf--li graf-after--li\">Now that you\u2019ve invested this effort, <strong class=\"markup--strong markup--li-strong\">you have solved a problem.<\/strong><\/li><li id=\"e68f\" class=\"graf graf--li graf-after--li\">Don\u2019t forget the call to action (CTA). What do you want from the journalist? And why do you need his or her help?<\/li><\/ol><p id=\"79b9\" class=\"graf graf--p graf-after--li\">These are simple marketing rules, but I can guarantee that 99 out of 100 statements sent out to press do not have this last step (CTA).<\/p><p id=\"bb9f\" class=\"graf graf--p graf-after--p\">In my case, I was quite sincere, and I was willing to create a story that was interesting to their audience. I would like to get <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> into the world and to reach as many people as possible.<\/p><p id=\"3389\" class=\"graf graf--p graf-after--p graf--trailing\">So when it comes to press, be genuine. An event is as good as its best story, so tell yours in a way people want to read about it and you\u2019ve got a big chance getting covered.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"21db\" class=\"graf graf--h3 graf--leading\">Launch Controls<\/h3><h4 id=\"4f7e\" class=\"graf graf--h4 graf-after--h3\">Your listing: where the magic\u00a0happens<\/h4><p id=\"9121\" class=\"graf graf--p graf-after--h4\">It\u2019s super important to make sure the different components of your listing work together\u200a\u2014\u200asince space to tell your story is limited. Let\u2019s start with the first part of your message that appears in the Product Hunt feed.<\/p><figure id=\"9c95\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*6G6MrBjBMzdGSkdG\" data-width=\"1392\" data-height=\"221\" data-action=\"zoom\" data-action-value=\"0*6G6MrBjBMzdGSkdG\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*6G6MrBjBMzdGSkdG\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"f63c\" class=\"graf graf--p graf-after--figure\">What we did was use the thumbnail to show someone how it worked, and made sure the title, tagline (and don\u2019t forget the \u201ctags\u201d ex. Productivity) worked together. You see how we selected productivity to show up as first\u200a\u2014\u200athis is because it\u2019s a part of the story.<\/p><p id=\"1343\" class=\"graf graf--p graf-after--p\">Next are your slides. Remember people might be intrigued by your tagline but probably still don\u2019t understand how to use your product. That\u2019s why we mainly used this to show use cases.<\/p><figure id=\"f819\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*j_Kl6LWaZ0C-L414x_n6xg.gif\" data-width=\"1348\" data-height=\"758\" data-action=\"zoom\" data-action-value=\"1*j_Kl6LWaZ0C-L414x_n6xg.gif\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*j_Kl6LWaZ0C-L414x_n6xg.gif\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"70b7\" class=\"graf graf--p graf-after--figure\">You can then use your opening comment to introduce yourself and give more details, the image is too long to post here so I\u2019d recommend that you check out our listing here:<a href=\"https:\/\/www.producthunt.com\/posts\/history-search\" data-href=\"https:\/\/www.producthunt.com\/posts\/history-search\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\"> https:\/\/www.producthunt.com\/posts\/history-search<\/a> this will also give you an impression of what kind of responses and questions you can expect.<\/p><p id=\"55e0\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tips<\/strong><\/p><p id=\"dd1a\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">1 Size matters. Keep your slides under 2MB each or they\u2019ll have trouble loading for visitors on mobile or weak internet connections (<a href=\"https:\/\/imageoptim.com\/\" data-href=\"https:\/\/imageoptim.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">ImageOptim<\/a> worked great for me)<\/p><p id=\"f00e\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">2 YouTube videos in comments. Not many people know you can also post videos in the comment section. So I\u2019d keep the main slide to use-cases only and add your video in the comments.<\/p><p id=\"0b3f\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">3 OG Images. The first slide is automatically set as the image shown when people share your post to social media\u2014make it easy to load (less than 1 MB) and add slightly more context in the image.<\/p><figure id=\"9fb8\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*F1gjXIxvVfjYfa61b5eTOQ.png\" data-width=\"2498\" data-height=\"1160\" data-action=\"zoom\" data-action-value=\"1*F1gjXIxvVfjYfa61b5eTOQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*F1gjXIxvVfjYfa61b5eTOQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><h4 id=\"7830\" class=\"graf graf--h4 graf-after--figure\">Launch day: a basic structure to plan the\u00a0day<\/h4><p id=\"640f\" class=\"graf graf--p graf-after--h4\">Launch day itself is super hectic. You get distracted all the time. You need a plan\u200a\u2014\u200anot so that you stick to the plan perfectly, but so that you have something to reference in the midst of all the distraction. If you get a bug or an email or a news post that you have to answer on the fly, you know where to continue, because you\u2019ve created this plan.<\/p><p id=\"48b5\" class=\"graf graf--p graf-after--p\">I\u2019ve actually made the structure I used available in an Airtable (see below), it\u2019s nothing fancy but it\u2019s all you need to plan your launch. It\u2019s structured as followed:<\/p><ol class=\"postList\"><li id=\"d9dd\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Activities. <\/strong>These are the actions you take on your launch date.<\/li><li id=\"7339\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Content. <\/strong>This is at the core of your activities because if you\u2019re launching digitally, it\u2019s all about content.<\/li><li id=\"5b1a\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Messaging.<\/strong> Make sure the messages you are communicating in your content, social media, and elsewhere are consistent.<\/li><li id=\"3b8f\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Support. <\/strong>You need reach. And to get reach, you need support.<\/li><\/ol><figure id=\"feae\" class=\"graf graf--figure graf--iframe graf-after--li graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe data-width=\"800\" data-height=\"533\" width=\"700\" height=\"466\" src=\"https:\/\/medium.com\/media\/f232d9b224509a966753285e97729bad?postId=c3f01864cdde\" data-media-id=\"f232d9b224509a966753285e97729bad\" data-thumbnail=\"https:\/\/i.embed.ly\/1\/image?url=https%3A%2F%2Fstatic.airtable.com%2Fimages%2Foembed%2Fairtable.png&amp;key=a19fcc184b9711e1b4764040d3dc5c07\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><figcaption class=\"imageCaption\">Copy the base and use it to plan your own launch \ud83d\udc46<\/figcaption><\/div><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"b156\" class=\"graf graf--h3 graf--leading\">Launch day<\/h3><p id=\"ee9e\" class=\"graf graf--p graf-after--h3\">First, <strong class=\"markup--strong markup--p-strong\">don\u2019t look at your ranking. <\/strong>You worked really hard to make something, and you don\u2019t want to see that you are 3rd, 4th, 5th etc. on your launch date\u200a\u2014\u200ait will only distract you.<\/p><p id=\"8993\" class=\"graf graf--p graf-after--p\">With <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a>, people in the office looked at it, but I didn\u2019t. I made a plan for how this launch date was going to go.<\/p><p id=\"b61b\" class=\"graf graf--p graf-after--p\">Of all of these activities, engaging with your audience could be the most important. After years working on your startup, you might be the Albert Einstein of your product, but <strong class=\"markup--strong markup--p-strong\">you can\u2019t beat a million brains thinking together. <\/strong>Or even just a thousand brains. And the bottom line is that if you have a thousand people out there who are willing to talk about your product, talk with them.<\/p><p id=\"7595\" class=\"graf graf--p graf-after--p\">First of all, it\u2019s an amazing feeling to get support and hear from people who love your work. And second, you will learn things that you could never have imagined.<\/p><p id=\"b87c\" class=\"graf graf--p graf-after--p\">For example:<\/p><ul class=\"postList\"><li id=\"72a0\" class=\"graf graf--li graf-after--p\">Somebody asked \u201cI don\u2019t understand the difference between Google and <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a>,\u201d which gave me the idea for the very first slide on our improved onboarding page, which visualizes the way it works: it indexes the text on the website when you\u2019re there (even if it\u2019s behind a login) and Google or browsing history don\u2019t do this.<\/li><li id=\"17bd\" class=\"graf graf--li graf-after--li\">Somebody else had a very good point about our upgrade page.<\/li><li id=\"aea3\" class=\"graf graf--li graf-after--li\">And somebody was confused about importing history.<\/li><\/ul><p id=\"c722\" class=\"graf graf--p graf-after--li\">Your launch isn\u2019t actually about you giving something to the world. <strong class=\"markup--strong markup--p-strong\">It is actually you getting something: you get an opportunity to talk to people and get feedback<\/strong>. It\u2019s very important that you learn from the people who are interested and also from the people who are not interested.<\/p><p id=\"0489\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Especially the people who are not interested\u200a<\/strong>\u2014\u200athis is super important. Talk to those people, and don\u2019t try to convince them. You really want to get down to the core reason of why they like or don\u2019t like your product. And sometimes, someone will ask you about a use case you hadn\u2019t considered, like the guy who asked if we could use <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a> without indexing automatically. Because of his question, I realized that this is actually possible. You can pause the extension and then favorite a web page manually\u2014so there you go, another selling-point without developing anything new.<\/p><blockquote id=\"1cc2\" class=\"graf graf--pullquote graf-after--p\"><div>It takes time to understand what it is you really\u00a0built<\/div><\/blockquote><p id=\"2764\" class=\"graf graf--p graf-after--pullquote\"><strong class=\"markup--strong markup--p-strong\">Tips<\/strong><\/p><p id=\"8b28\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">1 Go Live. Go live in the morning. Social media notifies everyone of this and it will be boosted through your Facebook and Instagram followers.<\/p><p id=\"30c0\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">2 Wait to win. The daily Product Hunt race starts at 00:00 PDT, but I\u2019d recommend you wait an hour to see who launches on Product Hunt because if for ex. Google launches something new you\u2019ll have to compete with them.<\/p><p id=\"9c58\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">3 Don\u2019t expect anything of Reddit or Hacker news. Sure, there are ways to use these channels, but not unless you really build a strategy.<\/p><p id=\"3fa0\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">4 Engaging. When people engage with you, really talk to them, try to understand and learn because this is your chance.<\/p><p id=\"4007\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">5Traction on content matters most. Don\u2019t link dump, it\u2019s better to target and generate interaction in a few groups, because practically all social media only boosts content that get a lot of comments.<\/p><p id=\"c5b9\" class=\"graf graf--p graf-after--p\">A few groups that I\u2019ve been in and contain many supportive and entrepreneurial people:<\/p><ul class=\"postList\"><li id=\"a382\" class=\"graf graf--li graf-after--p\"><a href=\"https:\/\/www.facebook.com\/groups\/startupproductlaunches\/\" data-href=\"https:\/\/www.facebook.com\/groups\/startupproductlaunches\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Startup Product Launches<\/a><\/li><li id=\"4ead\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.facebook.com\/groups\/saasmantra\/\" data-href=\"https:\/\/www.facebook.com\/groups\/saasmantra\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Saas Mantra<\/a><\/li><li id=\"8ae4\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.facebook.com\/groups\/producthunt\/\" data-href=\"https:\/\/www.facebook.com\/groups\/producthunt\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Product Hunt<\/a><\/li><li id=\"5eec\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.facebook.com\/groups\/Pushstarter\/\" data-href=\"https:\/\/www.facebook.com\/groups\/Pushstarter\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Pushstart<\/a><\/li><li id=\"5f8b\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.facebook.com\/groups\/SaaSgrowthhacking\/\" data-href=\"https:\/\/www.facebook.com\/groups\/SaaSgrowthhacking\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">SaaS Growth Hacks<\/a><\/li><li id=\"0a3f\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.facebook.com\/groups\/saasproductsandmarketing\/\" data-href=\"https:\/\/www.facebook.com\/groups\/saasproductsandmarketing\/\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">SaaS Product &amp; Marketing<\/a><\/li><\/ul><h3 id=\"fea0\" class=\"graf graf--h3 graf-after--li\">You\u2019ve Launched, What\u2019s\u00a0Next?<\/h3><p id=\"2099\" class=\"graf graf--p graf-after--h3\">So did we get the validation we were looking for?<\/p><ul class=\"postList\"><li id=\"16dd\" class=\"graf graf--li graf-after--p\">Total traffic on launch day: ~2,000<\/li><li id=\"15a9\" class=\"graf graf--li graf-after--li\">Total traffic on CNET day (the next day): ~2,500<\/li><li id=\"a775\" class=\"graf graf--li graf-after--li\">We now have approximately 650 active daily users, which has been increasing around 15% per day!<\/li><li id=\"57d1\" class=\"graf graf--li graf-after--li\">Free to paying users conversion rate on launch day: 3.5%<\/li><li id=\"83e2\" class=\"graf graf--li graf-after--li\">Free to paying users conversion today: 5.8%<\/li><\/ul><p id=\"e48a\" class=\"graf graf--p graf-after--li\">Once you\u2019ve launched, the things you learn mean you can never go back to your pre-launch perspective. My most important takeaways?<\/p><ol class=\"postList\"><li id=\"095a\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">It\u2019s not about when you\u2019re done. It\u2019s about when you\u2019re ready. <\/strong>Launch when you\u2019re ready and take advantage of the incredible opportunity for feedback from the community to improve your product.<\/li><li id=\"5bd4\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Value the feedback<\/strong>. Beyond the upvotes or the press, the greatest value comes from the feedback we received from hundreds of people talking to us and helping shape our product.<\/li><li id=\"0629\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Your launch day isn\u2019t a single day<\/strong>. It\u2019s more like 5 days: 2 days before and 2 days after. Plan accordingly.<\/li><\/ol><p id=\"542c\" class=\"graf graf--p graf-after--li\">So what\u2019s next for <a href=\"http:\/\/historysearch.com\" data-href=\"http:\/\/historysearch.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">History Search<\/a>? As I said, feedback was the most valuable thing we received from our launch date. We\u2019re using the feedback around our onboarding process, to improve our onboarding. Stay tuned for a later article that goes deeper into that process.<\/p><h4 id=\"d75e\" class=\"graf graf--h4 graf-after--p\">Before you\u00a0go\u2026<\/h4><p id=\"bd52\" class=\"graf graf--p graf-after--h4\"><em class=\"markup--em markup--p-em\">If you found this article helpful, click the \u2764\ufe0f or \ud83d\udc4f button below or share the article on Facebook so your friends can benefit from it too.<\/em><\/p><figure id=\"750d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"https:\/\/medium.com\/swlh\" data-href=\"https:\/\/medium.com\/swlh\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*YqDjlKFwScoQYQ62DWEdig.png\" data-width=\"2000\" data-height=\"100\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*YqDjlKFwScoQYQ62DWEdig.png\" alt=\"image\" \/><\/a><\/div><\/div><\/figure><h4 id=\"c148\" class=\"graf graf--h4 graf-after--figure\">This story is published in <a href=\"https:\/\/medium.com\/swlh\" data-href=\"https:\/\/medium.com\/swlh\" class=\"markup--anchor markup--h4-anchor\" target=\"_blank\">The Startup<\/a>, Medium\u2019s largest entrepreneurship publication followed by +390,426\u00a0people.<\/h4><h4 id=\"869c\" class=\"graf graf--h4 graf-after--h4\">Subscribe to receive <a href=\"http:\/\/growthsupply.com\/the-startup-newsletter\/\" data-href=\"http:\/\/growthsupply.com\/the-startup-newsletter\/\" class=\"markup--anchor markup--h4-anchor\" rel=\"nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow nofollow nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow nofollow noopener nofollow noopener nofollow noopener nofollow nofollow noopener nofollow nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener noopener noopener nofollow noopener nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener noopener nofollow noopener noopener nofollow noopener nofollow noopener nofollow noopener noopener noopener noopener noopener noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener noopener nofollow noopener noopener nofollow noopener noopener nofollow noopener nofollow noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener noopener nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener nofollow noopener\" target=\"_blank\">our top stories\u00a0here<\/a>.<\/h4><figure id=\"2d04\" class=\"graf graf--figure graf-after--h4 graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"https:\/\/medium.com\/swlh\" data-href=\"https:\/\/medium.com\/swlh\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*ouK9XR4xuNWtCes-TIUNAw.png\" data-width=\"2000\" data-height=\"100\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ouK9XR4xuNWtCes-TIUNAw.png\" alt=\"image\" \/><\/a><\/div><\/div><\/figure><\/div><\/div><\/section>","created_at":"2018-11-20T14:57:13+0000","updated_at":"2018-11-20T14:57:29+0000","published_at":"2018-11-19T06:40:30+0000","published_by":["Martijn Verbove"],"annotations":[],"mimetype":"text\/html","reading_time":18,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*m0UO4banJrJaLR5UciM1XQ.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":699,"title":"How Federation Works","url":"https:\/\/blog.soykaf.com\/post\/how-federation-works\/","content":"<p>A lot of people join our support channel (#pleroma on freenode) with the same sentence: \u201cI just set up Pleroma but federation isn\u2019t working\u201d. Usually, this is not actually true, they just have an empty timeline. What they don\u2019t know is how posts actually get into your instance. This short post is about the mechanisms that can make a post use your instance.<\/p><p>This is Pleroma specific, other fediverse servers may use different systems, although they are all very similar.<\/p>\n<h2 id=\"you-follow-someone\">You follow someone<\/h2>\n<p>This is probably the easiest to understand: If you follow someone, their posts will reach your server, because their server will send the posts to your server.<\/p>\n<h2 id=\"someone-directly-mentions-you\">Someone directly mentions you<\/h2>\n<p>Someone directly mentioning you will send the post to your server, so this post will end up in the federated timeline.<\/p>\n<h2 id=\"a-post-is-the-op-of-a-reply-you-receive\">A post is the OP of a reply you receive<\/h2>\n<p>If a reply to any post that the server doesn\u2019t have arrives, Pleroma will try to fetch the original post. This is recursive, so if the last reply in a 200 post thread gets to you, your server will pull in the whole thread.<\/p>\n<h2 id=\"a-post-is-repeated-boosted\">A post is repeated \/ boosted.<\/h2>\n<p>This will pull in the repeated post.<\/p>\n<h2 id=\"a-post-is-favorited\">A post is favorited.<\/h2>\n<p>This is maybe less obvious, but Pleroma will pull in all the favorites of all the users that people are following on your instance as well. This is a mechanism that helps to make the federated timeline more likely to contain posts that you might like, but have not seen. Because of this, even small Pleroma instances often have a rather lively federated TL.<\/p>\n<h2 id=\"a-post-comes-in-through-the-relay-mechanism\">A post comes in through the relay mechanism<\/h2>\n<p>This is essentially the same as repeating.<\/p>\n<h2 id=\"a-post-is-explicitly-requested-through-the-search-interface\">A post is explicitly requested through the search interface<\/h2>\n<p>If you search for a post with the url of that post, Pleroma will try to pull it in.<\/p>\n<p>These are all the ways that a post can reach your instance. As you can see, you need to start interacting to make the federated TL actually fill up! I hope this will help a few people when setting up new instances.<\/p>","created_at":"2018-11-20T14:39:01+0000","updated_at":"2018-11-20T14:39:01+0000","published_at":"2018-05-04T10:46:00+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"blog.soykaf.com","preview_picture":"https:\/\/source.unsplash.com\/category\/technology\/2000x1322","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["graphviz"],"is_public":false,"id":698,"title":"GraphViz Examples and Tutorial","url":"https:\/\/graphs.grevian.org\/","content":"<h2>What is this?<\/h2><p>Some clever people recognized that CS Majors suck at drawing, but still often need to draw graphs. Those\n            noble souls\n            made a program to draw graphs for us called <a href=\"http:\/\/graphviz.org\/\">GraphViz<\/a>, it's free, open\n            source, and great,\n            but not incredibly easy to use, So I threw this web interface and tutorial on top of it to make it easy\n            for us to make graphs for our assignments.<\/p><h2>What?<\/h2><p>Convert a simple language that describes graphs, to pretty pictures of graphs.<\/p><div class=\"row\"><div class=\"col-xsm\"><pre>\ngraph {\n    a -- b;\n    b -- c;\n    a -- c;\n    d -- c;\n    e -- c;\n    e -- a;\n  }\n                \n            <\/pre><\/div><div class=\"col-sm\"><figure class=\"figure\"><img src=\"https:\/\/graphs.grevian.org\/resources\/static\/images\/example1.png\" class=\"dot-example figure-img img-fluid rounded\" alt=\"A generic square placeholder image with rounded corners in a figure.\" \/><\/figure><\/div><\/div><h2>Cool, Where do I start?<\/h2><p>Take a look at some <a href=\"https:\/\/graphs.grevian.org\/example\">Examples<\/a> to see how this works, Then open the\n            <a href=\"https:\/\/graphs.grevian.org\/reference\">reference<\/a> and the <a href=\"https:\/\/graphs.grevian.org\/graph\">Graph<\/a> page and go to work.\n        <\/p>","created_at":"2018-11-19T14:30:00+0000","updated_at":"2018-11-19T14:30:04+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"graphs.grevian.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","graphviz","org-mode","org-babel"],"is_public":false,"id":697,"title":"Inline GraphViz DOT evaluation for graphs using Emacs, org-mode and org-babel","url":"https:\/\/vxlabs.com\/2014\/12\/04\/inline-graphviz-dot-evaluation-for-graphs-using-emacs-org-mode-and-org-babel\/","content":"<p> With Emacs, <a href=\"http:\/\/orgmode.org\/\">org mode<\/a> and org-babel, it\u2019s possible to evaluate source code samples embedded in your org files and have the output of said evaluation appear inline. This makes for a beautiful literate programming environment. It also enables one to include graphs in one\u2019s documents (org mode, PDF, HTML presentations or blog posts) by using for example GraphViz.<\/p><p> This blog post (<a href=\"http:\/\/vxlabs.com\/2014\/05\/25\/emacs-24-with-prelude-org2blog-and-wordpress\/\">obviously authored using Emacs and Org mode<\/a>) contains short instructions for doing so.<\/p><p> First follow <a href=\"http:\/\/orgmode.org\/manual\/Languages.html\">the org-babel documentation<\/a> and enable source code evaluation for dot by adding the following to your Emacs init.el:<\/p><p>(org-babel-do-load-languages<br \/> \u2018org-babel-load-languages<br \/> \u2018((dot . t)))<\/p><p> Then create a new document, and add something like the following dot source code sample to it:<\/p><pre class=\"brush: plain; title: ; notranslate\" title=\"\">\n#+BEGIN_SRC dot :file dot_success.png :cmdline -Kdot -Tpng\n  digraph {\n  \/\/ graph from left to right\n  rankdir=LR;\n  splines=true;\n  node [shape=box];\n\n  id [label=\"Install Graphviz\"]\n  conf [label=\"Configure org-babel\"]\n  dot [label=\"DOT in org-mode\"]\n\n  id -&gt; conf\n  conf -&gt; dot\n  dot -&gt; \"Profit\"\n  dot -&gt; \"Success\" [style=dotted]\n  }\n#+END_SRC\n<\/pre><p> Now press <code>C-c C-c<\/code> to evaluate this code. Emacs will generate the configured output file <code>dot_success.png<\/code> and then link to it in an automatically created <code>#+RESULTS<\/code> section right below it.<\/p><p> You can press <code>C-c C-x C-v<\/code> to toggle display of inline images to see it directly in Emacs. Alternatively, use <code>M-x org-display-inline-images<\/code> to switch this on. Whenever you change the DOT source code (press <code>C-c<\/code> \u2018 to edit the dot source code in a separate buffer), just press <code>C-c C-c<\/code> to re-execute the updated source.<\/p><p> In any exported documents (for examlpe this blog post), only the output graph itself will appear, like this:<\/p><div class=\"figure\"><p><a href=\"https:\/\/i2.wp.com\/vxlabs.com\/wp-content\/uploads\/2014\/12\/wpid-dot_success.png\"><img src=\"https:\/\/i1.wp.com\/vxlabs.com\/wp-content\/uploads\/2014\/12\/wpid-dot_success-300x56.png?resize=300%2C56\" data-recalc-dims=\"1\" alt=\"image\" \/><\/a><\/p><\/div><p> Exported as a PDF with <code>C-e l o<\/code> it looks <a href=\"http:\/\/vxlabs.com\/wp-content\/uploads\/2014\/12\/wpid-vxlabs-orgmode-graphviz.pdf\">like this<\/a>.<\/p><p> In your Emacs, it should look like this:<\/p><div class=\"figure\"><p><a href=\"https:\/\/i2.wp.com\/vxlabs.com\/wp-content\/uploads\/2014\/12\/wpid-emacs-org-babel-dot.png\"><img src=\"https:\/\/i1.wp.com\/vxlabs.com\/wp-content\/uploads\/2014\/12\/wpid-emacs-org-babel-dot-300x245.png?resize=300%2C245\" data-recalc-dims=\"1\" alt=\"image\" \/><\/a><\/p><\/div><p> This approach should work for any of the many languages supported by org-babel. Let me know in the comments what you come up with!<\/p><div id=\"jp-relatedposts\" class=\"jp-relatedposts\"><h3 class=\"jp-relatedposts-headline\"><em>Related<\/em><\/h3><\/div>","created_at":"2018-11-19T13:40:44+0000","updated_at":"2018-11-19T14:30:29+0000","published_at":"2014-12-04T09:25:00+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"vxlabs.com","preview_picture":"https:\/\/s0.wp.com\/i\/blank.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["css","typography"],"is_public":false,"id":696,"title":"CSS System Font Stack Monospace v2","url":"https:\/\/www.client9.com\/css-system-font-stack-monospace-v2\/","content":"<p>Get a great monospace font without using web-fonts. See the latest monospace system font stack based on work by GitHub and Bootstrap.<\/p><p> \u00b7 <time itemprop=\"pubdate\" datetime=\"2018-02-24T21:50:56Z\">February 24, 2018<\/time><\/p><h2 id=\"h.mcm5t4tv7hb1\" class=\"h4 mt-4 mb-4\">What is this?<\/h2><p>For most systems the default monospace font is something horrible like Courier. This can easily improved by using an installed system font, that also avoids the overhead and hacks of using a web-font.<\/p><p>This an update to the <a href=\"https:\/\/www.client9.com\/css-system-font-stack-monospace-v1\/\">original Monospace V1 article<\/a>.<\/p><h2 id=\"h.1nhupvxpu0ds\" class=\"h4 mt-4 mb-4\">What\u2019s Changed?<\/h2><p>Since that first article, <a href=\"https:\/\/github.com\">GitHub.com<\/a> has been using a monospace system font stack for at least six months. Given the their focus on computer code and that <a href=\"https:\/\/www.alexa.com\/siteinfo\/github.com\">GitHub is the 64th largest site on the Internet<\/a> means they should know what a good monospace stack is. Likewise the UI framework <a href=\"https:\/\/getbootstrap.com\">Bootstrap<\/a> has switched to an all system font stack with the release of v4. It <a href=\"https:\/\/trends.builtwith.com\/docinfo\/Twitter-Bootstrap\">powers 3.6% of the Internet<\/a>.<\/p><p>But unlike the <a href=\"https:\/\/www.client9.com\/css-system-font-stack-sans-serif-v3\/\">san-serif font stacks<\/a>, there hasn\u2019t been as much discussion as to why or how these stacks were created and updated.<\/p><h2 id=\"h.ivugv8jbpbp3\" class=\"h4 mt-4 mb-4\">Github 2018<\/h2><p>As of 2018, GitHub.com uses the following for their monospace font:<\/p><pre>font-family:\n    \"SFMono-Regular\",\n    Consolas,\n    \"Liberation Mono\",\n    Menlo,\n    Courier,\n    monospace;<\/pre><p class=\"p-3 bg-light\">\nTHE QUICK BROWN FOX JUMPED OVER THE LAZY DOG.<br \/>The quick brown fox jumped over the lazy dog.<br \/>0123456789<br \/>11111.11<br \/>88888.88<br \/>0Oo il1I! Z2z 8$s5S\ud83d\ude3a\u203c\u270f\u26a0\ud83d\ude2d\u2934\u2935\n<\/p><p><a href=\"https:\/\/developer.apple.com\/fonts\/\">SFMono-Regular<\/a> isn\u2019t available on iOS Safari and isn\u2019t available on macOS Safari unless you do some <a href=\"https:\/\/simonfredsted.com\/1438\">hackery<\/a>. I figure 0.00001% of people actually do this, but those that do must really love this font. For most macOS and iOS users, this stack defaults to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Menlo_(typeface)\">Menlo<\/a>.<\/p><p>The reason to add <a href=\"https:\/\/en.wikipedia.org\/wiki\/Liberation_fonts\">Liberation Mono<\/a> is unclear. It must solve some problem for a Linux distribution but not clear how. Since Liberation Mono is similar to Courier, I would have assumed it\u2019s the default \u201cmonospace\u201d font for these platforms. But maybe not. Either way it\u2019s benign to put in the stack.<\/p><h2 id=\"h.7ejv3md8xsf\" class=\"h4 mt-4 mb-4\">Bootstrap v4<\/h2><p>With Bootstrap v4.0.0, the monospace stack is:<\/p><pre>font-family:\n    SFMono-Regular,\n    Menlo,\n    Monaco,\n    Consolas,\n    \"Liberation Mono\",\n    \"Courier New\",\n    monospace;<\/pre><p class=\"p-3 bg-light\">\nTHE QUICK BROWN FOX JUMPED OVER THE LAZY DOG.<br \/>The quick brown fox jumped over the lazy dog.<br \/>0123456789<br \/>11111.11<br \/>88888.88<br \/>0Oo il1I! Z2z 8$s5S\ud83d\ude3a\u203c\u270f\u26a0\ud83d\ude2d\u2934\u2935\n<\/p><p>More mysterious choices. Monaco has not been a primary font on macOS since 2009, and has never been part of iOS. Courier New is absolutely not as good as Courier for code or the screen. But with a semi-modern system, it\u2019s unlikely either Monaco or Courier New will ever get used.<\/p><h2 id=\"h.qz225hbft1n4\" class=\"h4 mt-4 mb-4\">Conclusion<\/h2><p>While I don\u2019t quite understand some of the choices in these font stack, both of these provide a great monospace font for anyone using macOS, iOS, Windows or Android from the last five years. I figure that\u2019s most of your audience.<\/p><p>Instead of tracking the latest browser\/linux font combinations and the SFMono soap opera, I\u2019m going to just use whatever Bootstrap uses and call it a day. It\u2019s going to delivery a great monospace font without incurring the wrath of a web-font.<\/p>","created_at":"2018-11-19T02:27:22+0000","updated_at":"2018-11-19T02:27:35+0000","published_at":"2018-02-24T21:50:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.client9.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":695,"title":"The New York Times > Page Not Found","url":"https:\/\/opinionator.blogs.nytimes.com\/2012\/08\/08\/hear-all-ye...","content":"<div class=\"columnGroup header insetV\"><p>We\u2019re sorry, we seem to have lost this page, but we don\u2019t want to lose you.<\/p><ul><li>\n<strong>Check the Archives.<\/strong> Most articles remain\nonline for seven days after publication. Articles back to 1851 are available\nthrough <a href=\"http:\/\/query.nytimes.com\/search\/query?srchst=nyt&amp;&amp;srcht=a&amp;srchr=n\">The New York Times Article Archive<\/a>. 1851 \u2013\npresent.\n<\/li>\n<li>\n<strong>Report the broken link.<\/strong> If you clicked on a\nheadline or other link on NYTimes.com, you can <a href=\"http:\/\/www.nytimes.com\/membercenter\/formblink.html\">report the missing page<\/a>.\n<\/li>\n<\/ul><\/div>","created_at":"2018-11-19T01:28:39+0000","updated_at":"2018-11-19T01:28:39+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"opinionator.blogs.nytimes.com","http_status":"404"},{"is_archived":0,"is_starred":0,"tags":["css","webdev","typography"],"is_public":false,"id":694,"title":"The New System Font Stack?","url":"https:\/\/bitsofco.de\/the-new-system-font-stack\/","content":"<p>A few months ago, I wrote about how you can use system fonts in the browser using the built-in keywords that work with the <code>font<\/code> shorthand property (see <a href=\"https:\/\/bitsofco.de\/using-system-fonts-in-the-browser\/\">Using System Fonts in the Browser<\/a>). These keywords can be useful, but they also have some drawbacks, the main one being that they only work with the shorthand <code>font<\/code> property. This means that we cannot use it in a list of font families with fallbacks or as a fallback itself.<\/p><p>Relatively recently, some websites and web applications have been adopting a new method for using system fonts in the browser. With this method, the fonts used by different systems are explicitly called themselves in the <code>font-family<\/code> property.<\/p><h2 id=\"thefontstack\">The Font Stack<\/h2><p>At the moment, there is not one universally-used sequence of fonts for this system font stack. However, the fonts used by different sites, although not exactly the same, overlap. These are the different fonts that have been used, and the devices (and browsers) they target -<\/p><table><thead><\/thead><tbody><tr class=\"odd\"><td><code>-apple-system<\/code> (San Francisco)<\/td>\n<td>iOS Safari, macOS Safari, macOS Firefox<\/td>\n<\/tr><tr class=\"even\"><td><code>BlinkMacSystemFont<\/code> (San Francisco)<\/td>\n<td>macOS Chrome<\/td>\n<\/tr><tr class=\"odd\"><td><code>Segoe UI<\/code><\/td>\n<td>Windows<\/td>\n<\/tr><tr class=\"even\"><td><code>Roboto<\/code><\/td>\n<td>Android, Chrome OS<\/td>\n<\/tr><tr class=\"odd\"><td><code>Oxygen<\/code> \/ <code>Oxygen-Sans<\/code><\/td>\n<td>KDE<\/td>\n<\/tr><tr class=\"even\"><td><code>Fira Sans<\/code><\/td>\n<td>Firefox OS<\/td>\n<\/tr><tr class=\"odd\"><td><code>Droid Sans<\/code><\/td>\n<td>Older versions of Android<\/td>\n<\/tr><tr class=\"even\"><td><code>Ubuntu<\/code><\/td>\n<td>Ubuntu<\/td>\n<\/tr><tr class=\"odd\"><td><code>Cantarell<\/code><\/td>\n<td>GNOME<\/td>\n<\/tr><tr class=\"even\"><td><code>Helvetica Neue<\/code><\/td>\n<td>macOS versions &lt; 10.11<\/td>\n<\/tr><tr class=\"odd\"><td><code>Arial<\/code><\/td>\n<td>Any<\/td>\n<\/tr><tr class=\"even\"><td><code>sans-serif<\/code><\/td>\n<td>Any<\/td>\n<\/tr><\/tbody><\/table><p>The order in which the fonts are defined in the <code>font-family<\/code> property is very important. Because we are just passing font names to the property, the font that will be used is the first font available on the user's machine that matches that name. This means that we have to craft the stack in such a way that the first font available to the user is their system font, and not some font they have installed on their machine.<\/p><p>We do this by starting with the most specific fonts possible, and moving on to fonts that are less guaranteed to be tied to a specific device. For example, the first font defined is always <code>-apple-system<\/code>. This will certainly only be applied to users on an iOS or macOS device. Further down the line, we have <code>Roboto<\/code>. This is a font that very well <em>could<\/em> be installed on a Mac user's device, but since the font stack started with the more specific <code>-apple-system<\/code>, we don't have to worry about this conflict.<\/p><h3 id=\"examples\">Examples<\/h3><p>Here is how this new system font stack is used across different websites. Starting with <strong>Wordpress 4.6<\/strong> -<\/p><pre class=\"language-language-css\">body {\n   font-family: -apple-system, \n                BlinkMacSystemFont, \n                \"Segoe UI\", \n                Roboto, \n                Oxygen-Sans, \n                Ubuntu, \n                Cantarell, \n                \"Helvetica Neue\", \n                sans-serif;\n}\n<\/pre><p><strong>Medium<\/strong> -<\/p><pre class=\"language-language-css\">body {\n  font-family: -apple-system, \n               BlinkMacSystemFont, \n               \"Segoe UI\", \n               Roboto, \n               Oxygen,\n               Ubuntu,\n               Cantarell,\n               \"Open Sans\",\n               \"Helvetica Neue\", \n               sans-serif;\n}\n<\/pre><p><strong>Ghost<\/strong> -<\/p><pre class=\"language-language-css\">body {\n  font-family: -apple-system, \n               BlinkMacSystemFont, \n               \"Segoe UI\", \n               Roboto, \n               Oxygen,\n               Ubuntu,\n               Cantarell,\n               \"Fira Sans\",\n               \"Droid Sans\",\n               \"Helvetica Neue\", \n               sans-serif;\n}\n<\/pre><p>Finally, <strong>GitHub<\/strong> -<\/p><pre class=\"language-language-css\">body {\n  font-family: -apple-system, \n               BlinkMacSystemFont, \n               \"Segoe UI\", \n               Roboto, \n               Helvetica,\n               Arial,\n               sans-serif,\n               \"Apple Color Emoji\",  \/* Emojis*\/\n               \"Segoe UI Emoji\", \/* Emojis*\/\n               \"Segoe UI Symbol\"; \/* Emojis*\/\n}\n<\/pre><p>Here is a comparison between the font stacks used by these sites -<\/p><table><thead><\/thead><tbody><tr class=\"odd\"><td><code>-apple-system<\/code><\/td>\n<td><code>-apple-system<\/code><\/td>\n<td><code>-apple-system<\/code><\/td>\n<td><code>-apple-system<\/code><\/td>\n<\/tr><tr class=\"even\"><td><code>BlinkMacSystemFont<\/code><\/td>\n<td><code>BlinkMacSystemFont<\/code><\/td>\n<td><code>BlinkMacSystemFont<\/code><\/td>\n<td><code>BlinkMacSystemFont<\/code><\/td>\n<\/tr><tr class=\"odd\"><td><code>\u201cSegoe UI\u201d<\/code><\/td>\n<td><code>\u201cSegoe UI\u201d<\/code><\/td>\n<td><code>\u201cSegoe UI\u201d<\/code><\/td>\n<td><code>\u201cSegoe UI\u201d<\/code><\/td>\n<\/tr><tr class=\"even\"><td><code>Roboto<\/code><\/td>\n<td><code>Roboto<\/code><\/td>\n<td><code>Roboto<\/code><\/td>\n<td><code>Roboto<\/code><\/td>\n<\/tr><tr class=\"odd\"><td><code>Oxygen-Sans<\/code><\/td>\n<td><code>Oxygen<\/code><\/td>\n<td><code>Oxygen<\/code><\/td>\n<td><code>Helvetica<\/code><\/td>\n<\/tr><tr class=\"even\"><td><code>Ubuntu<\/code><\/td>\n<td><code>Ubuntu<\/code><\/td>\n<td><code>Ubuntu<\/code><\/td>\n<td><code>Arial<\/code><\/td>\n<\/tr><tr class=\"odd\"><td><code>Cantarell<\/code><\/td>\n<td><code>Cantarell<\/code><\/td>\n<td><code>Cantarell<\/code><\/td>\n<td><code>sans-serif<\/code><\/td>\n<\/tr><tr class=\"even\"><td><code>\u201cHelvetica Neue\u201d<\/code><\/td>\n<td><code>\u201cOpen Sans\u201d<\/code><\/td>\n<td><code>\u201cFira Sans\u201d<\/code><\/td>\n<td> <\/td>\n<\/tr><tr class=\"odd\"><td><code>sans-serif<\/code><\/td>\n<td><code>\u201cHelvetica Neue\u201d<\/code><\/td>\n<td><code>\u201cDroid Sans\u201d<\/code><\/td>\n<td> <\/td>\n<\/tr><tr class=\"even\"><td> <\/td>\n<td><code>sans-serif<\/code><\/td>\n<td><code>\u201cHelvetica Neue\u201d<\/code><\/td>\n<td> <\/td>\n<\/tr><tr class=\"odd\"><td> <\/td>\n<td> <\/td>\n<td><code>sans-serif<\/code><\/td>\n<td> <\/td>\n<\/tr><\/tbody><\/table><h2 id=\"whyusethisfontstack\">Why use this Font Stack?<\/h2><p>There are a number of benefits to using system fonts in general, and to this particular method of using them.<\/p><h3 id=\"performance\">Performance<\/h3><p>The number one reason for using system fonts at all is performance. Fonts are typically one of the largest\/heaviest resources loaded on a website. If we can use a font already available on the user\u2019s machine, we can completely eliminate the need to fetch this resource, making load times noticeably faster.<\/p><h3 id=\"thenativefeel\">The Native Feel<\/h3><p>Another benefit of using system fonts is that it creates a feeling of the website being in a native application. This makes this method particularly useful for web applications, such as Wordpress or GitHub, rather than websites, as the former may be trying to emulate native applications.<\/p><h3 id=\"greatercontrol\">Greater Control<\/h3><p>In comparison to using the built-in keywords with the <code>font<\/code> shorthand property, this method offers us a lot more control when specifying our fonts. We can define which font families we want to favour. Additionally, the other typography-related properties are not affected since this method only applies to the <code>font-family<\/code> property.<\/p><h3 id=\"websitesdontneedtolookthesameeverywhere\">Websites Don't Need to Look the Same Everywhere<\/h3><p>Although the fact that different fonts will be shown for different users may seem like a drawback, in actual fact this doesn't matter. As <a href=\"https:\/\/twitter.com\/mrmrs_\/status\/775837789619417088\">mrmrs<\/a> aptly put it -<\/p><h2 id=\"whatarethedrawbacks\">What are the Drawbacks?<\/h2><p>Using this method also has some drawbacks to consider.<\/p><h3 id=\"namingconflicts\">Naming Conflicts<\/h3><p>Besides <code>-apple-system<\/code> and <code>BlinkMacSystemFont<\/code>, all we are doing is calling the actual name of fonts. This means that, if a user has a font by that same name installed on their machine that isn't their system font, that may be used instead. This is <a href=\"https:\/\/medium.design\/system-shock-6b1dc6d6596f\">what happened to Medium<\/a> when they first rolled out this method last year.<\/p><h3 id=\"maintenance\">Maintenance<\/h3><p>This list of fonts will have to be maintained and updated as the system fonts used by the various operating systems change. Hopefully, more operating systems will move towards Apple's method of having a generic name that will target the latest device font. But until then, a list will need to be periodically iterated on.<\/p><h2 id=\"usingthisfontstacktoday\">Using this Font Stack Today<\/h2><p>If you would like to use this font stack today, which fonts you use will be dependent on your users. Personally, I think the Wordpress method is the most suitable for the largest amount of users, and has been <a href=\"https:\/\/make.wordpress.org\/core\/2016\/07\/07\/native-fonts-in-4-6\/\">extensively tested by their team<\/a>.<\/p><pre class=\"language-language-css\">body {\n   font-family: -apple-system, \n                BlinkMacSystemFont, \n                \"Segoe UI\", \n                Roboto, \n                Oxygen-Sans, \n                Ubuntu, \n                Cantarell, \n                \"Helvetica Neue\", \n                sans-serif;\n}\n<\/pre><p>Whichever method you choose, make sure to test it across the different browsers (and versions) that are important to the particular site.<\/p>","created_at":"2018-11-18T21:05:02+0000","updated_at":"2018-11-18T21:05:32+0000","published_at":"2016-09-20T07:17:37+0000","published_by":["Ire Aderinokun"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"bitsofco.de","preview_picture":"https:\/\/bitsofco.de\/assets\/icons\/icon_128.png?v=db27bb16a3","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["typography"],"is_public":false,"id":693,"title":"Prac\u00adti\u00adcal Ty\u00adpog\u00adra\u00adphy | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=18477231","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D18477231\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"18477231\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/practicaltypography.com\/\" class=\"storylink\">Prac\u00adti\u00adcal Ty\u00adpog\u00adra\u00adphy<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=practicaltypography.com\">practicaltypography.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        147 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=beefhash\" class=\"hnuser\">beefhash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18477231\">9 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=18477231&amp;goto=item%3Fid%3D18477231\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Prac%C2%ADti%C2%ADcal%20Ty%C2%ADpog%C2%ADra%C2%ADphy&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Prac%C2%ADti%C2%ADcal%20Ty%C2%ADpog%C2%ADra%C2%ADphy\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=18477231&amp;auth=2fce2286060a24aae7e2ee7f4aa8d19da5ac1b59\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=18477231\">15\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-18T21:04:39+0000","updated_at":"2018-11-18T21:05:52+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["typography"],"is_public":false,"id":692,"title":"practicaltypography.com","url":"https:\/\/practicaltypography.com\/","created_at":"2018-11-18T21:04:06+0000","updated_at":"2018-11-18T21:06:07+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"practicaltypography.com"},{"is_archived":0,"is_starred":0,"tags":["databases","graphs","dgraph"],"is_public":false,"id":691,"title":"Graph databases 101","url":"https:\/\/changelog.com\/posts\/graph-databases-101","content":"<p><a href=\"https:\/\/dgraph.io\/\">Dgraph<\/a> founder <a href=\"https:\/\/twitter.com\/manishrjain\">Manish R Jain<\/a> gave us <mark>an excellent explanation of graph databases<\/mark> on <a href=\"https:\/\/changelog.com\/podcast\/322\">episode 322 of The Changelog<\/a>.<\/p><p><em>(Thanks to listener <a href=\"https:\/\/twitter.com\/MisterGenest\">Evan Genest<\/a> for highlighting this fact <a href=\"https:\/\/twitter.com\/MisterGenest\/status\/1063224954446069761\">on Twitter<\/a>.)<\/em><\/p><p>I extracted that section from our trusty <a href=\"https:\/\/changelog.com\/podcast\/322#transcript-34\">episode transcript<\/a> for your reading pleasure:<\/p><blockquote>\n<p>I think graphs are probably the simplest things to think about, really. People think about SQL tables - you have a row and you have some columns. Think of a graph as three columns where you have a subject, a predicate, and an object. If you put together a whole bunch of these things, you get a graph. A subject is essentially \u2013 think of it as an entity; a predicate is the relationship, and the object is either another entity, or a value.<\/p>\n<p>The subject could be, let\u2019s say, me. The relationship might be \u201clives in\u201d and object might be \u201cSan Francisco.\u201d Or it could be me - the name is Manish, and that\u2019s sort of like a property. So you just put together a whole bunch of these facts, or triples, and you get a graph.<\/p>\n<p>Then other people who live in San Francisco would have similar facts, and then you could run a graph query around \u201cHey, tell me all the people who live in San Francisco and who eat sushi.\u201d So you pick up all the people who live in San Francisco, you intersect with people in the world who eat sushi - which are completely different facts. You didn\u2019t create them as \u201cThis person lives in San Francisco AND eats sushi.\u201d This is something that we\u2019re doing on the fly. So you pick up all the people in San Francisco, you pick up all the people in the world who eat sushi, you intersect the two lists, and now you get people in San Francisco who eat sushi. Now you can take that result and say \u201cIntersect it with all the people who have been to Japan.\u201d You pick up another list of people who have been to Japan, intersect it with this, and now you get people who live in San Francisco, who eat sushi, and who have been to Japan.<\/p>\n<p>So the power of graphs is really in these joins that you can do, coming from just very simple facts.<\/p>\n<\/blockquote><p>Listen for yourself <a href=\"https:\/\/changelog.com\/podcast\/322#t=30:09\">starting at 30:09<\/a> or start <a href=\"https:\/\/changelog.com\/podcast\/322\">right from the top<\/a>.<\/p><p>Thanks, Manish! \ud83d\udc9a\ud83d\udc9a\ud83d\udc9a<\/p>","created_at":"2018-11-18T17:46:52+0000","updated_at":"2018-11-18T18:17:10+0000","published_at":"2018-11-17T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"changelog.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":690,"title":"A List of Reasons Why Our Brains Love Lists","url":"https:\/\/www.newyorker.com\/tech\/annals-of-technology\/a-list-of-reasons-why-our-brains-love-lists","content":"<p>\u201c<em class=\"small\">6 TITANIC SURVIVORS WHO SHOULD HAVE DIED<\/em>.\u201d \u201c<em class=\"small\">THESE 9 NAZI ATROCITIES WILL MAKE YOU LOSE FAITH IN HUMANITY<\/em>.\u201d \u201c<em class=\"small\">5 INSANE PLANS FOR FEEDING WEST BERLIN YOU WON\u2019T BELIEVE ARE REAL<\/em>.\u201d These are just some of the lists that the comic strip \u201c<a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/xkcd.com\/1283\/\" target=\"_blank\">XKCD<\/a>\u201d recently joked would result from retrofitting the twentieth century\u2019s most newsworthy events with modern, Internet-style headlines. Despite the growing derision of listicles exemplified by the comic, numbered lists\u2014a <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/en.wikipedia.org\/wiki\/The_Ninety-Five_Theses\" target=\"_blank\">venerable<\/a> media format\u2014have become one of the most ubiquitous ways to package content on the Web. Why do we find them so appealing?<\/p><p>The article-as-numbered-list has several features that make it inherently captivating: the headline catches our eye in a stream of content; it positions its subject within a pre\u00ebxisting category and classification system, like \u201ctalented animals\u201d; it spatially organizes the information; and it promises a story that\u2019s finite, whose length has been quantified upfront. Together, these create an easy reading experience, in which the mental heavy lifting of conceptualization, categorization, and analysis is completed well in advance of actual consumption\u2014a bit like sipping green juice instead of munching on a bundle of kale. And there\u2019s little that our brains crave more than effortlessly acquired data.<\/p><p>Whenever we encounter new information, our brains <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.sciencedirect.com\/science\/article\/pii\/S1364661306002117\" target=\"_blank\">immediately try<\/a> to make sense of it. Once <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.pnas.org\/content\/98\/2\/676.abstract\" target=\"_blank\">they figure out<\/a> what we\u2019re seeing in a physical sense, they work to provide personal context and decide if it\u2019s relevant enough to focus on further. The process is instantaneous: we don\u2019t even realize we\u2019ve made a choice in the time our minds have selected one path or another. Our gaze either stops, or we simply keep scanning. Recall a time when you were spacing out while skimming a stream of content and then, without quite knowing why, found yourself pausing to actually process the words. What made you stop and focus? On a physical level, the <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.sciencedirect.com\/science\/article\/pii\/S0042698901002504\" target=\"_blank\">answer is often simple<\/a>: difference. Whenever we\u2019re scanning the environment for nothing in particular, our visual system is arrested by the things that don\u2019t fit\u2014features that suddenly change or <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.cell.com\/trends\/cognitive-sciences\/\/retrieve\/pii\/S1364661306001616\" target=\"_blank\">somehow stand out<\/a> from the background. A headline that is graphically salient in some way has a greater chance of capturing our eye, and in an environment where dozens of headlines and stories vie for attention, numerals break up the visual field. Consider the contexts in which we\u2019re most likely to debate which article to read: a publication\u2019s home page, a Twitter feed, or a Facebook feed. Most of what we see is words and images (even though it often seems like Web pages or streams are composed of nothing but lists). In that context, numbers pop.<\/p><p>Once our attention has been ensnared, we still need to be sufficiently intrigued to read the story. In 2009, when <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.sciencedirect.com\/science\/article\/pii\/S0378216608002798\" target=\"_blank\">researchers at the University of Athens<\/a> examined actual readers\u2019 responses to headlines from English-language newspapers in the U.S. and U.K., ranging from hard news to tabloids, they found that people preferred headlines that were both creative and uninformative, like \u201c<em class=\"small\">THE SMELL OF CORRUPTION, THE SCENT OF TRUTH<\/em>\u201d or \u201c<em class=\"small\">FACE TO FAITH<\/em>.\u201d They not only rated them as more interesting over-all but also indicated that they would be more likely to read the corresponding stories. List-style headlines often provide that optimal balance of information and ambivalence, intriguing us just enough to click, on the chance that we\u2019ll come across something particularly relevant or exciting.<\/p><p>Once we click, lists tap into our preferred way of receiving and organizing information at a subconscious level; from an information-processing standpoint, they often hit our attentional sweet spot. When <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/dl.acm.org\/citation.cfm?id=165556\" target=\"_blank\">we process information<\/a>, we do so spatially. For instance, it\u2019s hard to memorize through brute force the groceries we need to buy. It\u2019s easier to remember everything if we write it down in bulleted, or numbered, points. Then, even if we forget the paper at home, it is easier for us to recall what was on it because we can think back to the location of the words themselves. Lists also appeal to our general tendency to categorize things\u2014in fact, it\u2019s hard for us <em class=\"\">not<\/em> to categorize something the moment we see it\u2014since they chunk information into short, distinct components. This type of organization facilitates both immediate understanding and later recall, as the neuroscientist Walter Kintsch <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/psycnet.apa.org\/index.cfm?fa=buy.optionToBuy&amp;id=1969-01986-001\" target=\"_blank\">pointed out<\/a> back in 1968. Because we can process information more easily when it\u2019s in a list than when it\u2019s clustered and undifferentiated, like in standard paragraphs, a list feels more intuitive. In other words, lists simply <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.ncbi.nlm.nih.gov\/pubmed\/18468944\" target=\"_blank\">feel better<\/a>.<\/p><p>But the list\u2019s deepest appeal, and the source of its staying power, goes beyond the fact that it feels good. In 2011, the psychologists Claude Messner and Michaela W\u00e4nke <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.sciencedirect.com\/science\/article\/pii\/S1057740810001312\" target=\"_blank\">investigated<\/a> what, if anything, could alleviate the so-called \u201cparadox of choice\u201d\u2014the phenomenon that the more information and options we have, the worse we feel. They concluded that we feel better when the amount of conscious work we have to do in order to process something is reduced; the faster we decide on something, whether it\u2019s what we\u2019re going to eat or what we\u2019re going to read, the happier we become. Within the context of a Web page or Facebook stream, with their many choices, a list is the easy pick, in part because it promises a definite ending: we think we know what we\u2019re in for, and the certainty is both <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/link.springer.com\/article\/10.1007%2FBF00122575\" target=\"_blank\">alluring and reassuring<\/a>. The more we know about something\u2014including precisely how much time it will consume\u2014the greater the chance we will commit to it. The process is self-reinforcing: we recall with pleasure that we were able to complete the task (of reading the article) instead of leaving it undone and that satisfaction, in turn, makes us more likely to click on lists again\u2014even ones <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.newyorker.com\/online\/blogs\/books\/2013\/01\/the-pleasure-of-reading-repulsive-writers.html\">we hate-read<\/a>. The social psychologist Robert Zajonc, who made his name studying the connection between emotion and cognition, <a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/psycnet.apa.org\/index.cfm?fa=search.displayRecord&amp;uid=1980-09733-001\" target=\"_blank\">argued<\/a> that the positive feeling of completion in and of itself is enough to inform future decisions. Preferences, goes his famous coinage, need no inferences.<\/p><p>In the current media environment, a list is perfectly designed for our brain. We are drawn to it intuitively, we process it more efficiently, and we retain it with little effort. Faced with a detailed discussion of policies toward China or five insane buildings under construction in Shanghai, we tend to choose the latter bite-sized option, even when we know we will not be entirely satisfied by it. And that\u2019s just fine, as long as we realize that our fast-food information diet is necessarily limited in content and nuance, and thus unlikely to contain the nutritional value of the more in-depth analysis of traditional articles that rely on paragraphs, not bullet points.<\/p><p><em class=\"\"><a class=\"ArticleBody__link___1FS03\" href=\"https:\/\/twitter.com\/mkonnikova\" target=\"_blank\">Maria Konnikova<\/a> is the author of \u201c<a class=\"ArticleBody__link___1FS03\" href=\"http:\/\/www.amazon.com\/dp\/0670026573\/?tag=thneyo0f-20\" target=\"_blank\" rel=\"nofollow\">Mastermind: How to Think Like Sherlock Holmes<\/a>.\u201d<\/em><\/p><p><em class=\"\">Photograph: Bert Hardy Advertising Archive\/Getty.<\/em><\/p>","created_at":"2018-11-16T20:54:51+0000","updated_at":"2018-11-16T20:54:51+0000","published_at":"2013-12-02T13:16:03+0000","published_by":["Maria Konnikova"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.newyorker.com","preview_picture":"https:\/\/media.newyorker.com\/photos\/590952641c7a8e33fb38a7f8\/16:9\/w_1200,h_630,c_limit\/woman-with-grocery-list-580.jpg","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","wordpress","phpunit"],"is_public":false,"id":689,"title":"Unit testing WordPress plugins with PHPUnit in PhpStorm \u2022 Yoast","url":"https:\/\/yoast.com\/dev-blog\/unit-testing-wordpress-plugins-with-phpunit-in-phpstorm\/","content":"<p>A nice way to optimize your test flow is by having PhpStorm run your unit tests for you. It can be quite some work to set this up for WordPress plugins, since they depend so heavily on WordPress. In this post, I\u2019ll go over all the steps you\u2019ll have to take to do this.<\/p><h3>1) Install PHPUnit<\/h3><p>To start unit testing your code you first have to install PHPUnit.<\/p><p>There are basically 3 ways to do this:<\/p><ul><li>Download the .phar file from the <a href=\"https:\/\/phar.phpunit.de\/\">phpunit.de website<\/a><\/li>\n<li>Install via composer in the composer.json file<\/li>\n<\/ul><pre class=\"brush: jscript; title: ; notranslate\" title=\"\">&#13;\n{ &#13;\n  \"require-dev\": {&#13;\n    \"phpunit\/phpunit\": \"5.0.*\"&#13;\n  }&#13;\n} &#13;\n<\/pre><ul><li>Install via homebrew, when you\u2019re using mac OSX: <code>brew install phpunit<\/code><\/li>\n<\/ul><p>More elaborate <a href=\"https:\/\/phpunit.de\/manual\/current\/en\/installation.html\">installation instructions<\/a> can be found on the website of PHPUnit.<\/p><p>In my situation and examples I\u2019ll use PHPUnit installed with homebrew.<\/p><h3>2) Set up tests for WordPress plugins<\/h3><p>There are a few things you need, to write Unit Tests for WordPress plugins.<br \/><\/p><h4>WP develop<\/h4><p>As we\u2019re developing a WordPress plugin, the code will be coupled to WordPress. This is not ideal for unit testing, but since WordPress has lots of global state and functions, it\u2019s virtually impossible (or at least quite impractical) to isolate the plugin code from it. This also means we want to run our unit tests on top of WP develop. A guideline on\u00a0how to setup WP develop unit tests can be found <a href=\"https:\/\/make.wordpress.org\/core\/handbook\/testing\/automated-testing\/phpunit\/\">on the WordPress website<\/a>.<\/p><h4>bootstrap.php<\/h4><p>To run PHPUnit you need to first bootstrap your environment. Since WordPress is part of our environment, we\u2019ll also require the WP develop <code>bootstrap.php<\/code> (you can find this in the <code>wp-develop<\/code> directory). We\u2019ll put the <code>bootstrap.php<\/code> file in a tests directory in the project root.<\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n\/\/ disable xdebug backtrace&#13;\nif ( function_exists( 'xdebug_disable' ) ) {&#13;\n  xdebug_disable();&#13;\n}&#13;\n&#13;\nif ( false !== getenv( 'WP_PLUGIN_DIR' ) ) {&#13;\n  define( 'WP_PLUGIN_DIR', getenv( 'WP_PLUGIN_DIR' ) );&#13;\n}&#13;\n&#13;\nif ( false !== getenv( 'WP_DEVELOP_DIR' ) ) {&#13;\n  require getenv( 'WP_DEVELOP_DIR' ) . 'tests\/phpunit\/includes\/bootstrap.php';&#13;\n}&#13;\n<\/pre><h4>phpunit.xml<\/h4><p>This is the configuration file where you tell PHPUnit in which directory the tests and the <code>bootstrap.php<\/code> are.<\/p><pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n&lt;phpunit bootstrap=\"tests\/bootstrap.php\"&gt;&#13;\n  &lt;testsuites&gt;&#13;\n    &lt;testsuite&gt;&#13;\n      &lt;directory prefix=\"test-\" suffix=\".php\"&gt;.\/tests\/&lt;\/directory&gt;&#13;\n    &lt;\/testsuite&gt;&#13;\n  &lt;\/testsuites&gt;&#13;\n&lt;\/phpunit&gt;&#13;\n<\/pre><p>In our situation the bootstrap file can be found under the <code>\/tests<\/code> directory. The target directory where PHPUnit will look for the file will be <code>\/tests<\/code> as well. According to convention, all the test files have to begin with <code>test-<\/code> and should end with <code>.php <\/code> (the extension) in order to be picked up by PHPUnit.<\/p><div id=\"attachment_617287\" class=\"wp-caption alignnone\"><a href=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-filestructure-1.png\" rel=\"attachment wp-att-617287\"><img class=\"wp-image-617287 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-filestructure-1-600x141.png\" alt=\"phpunit-filestructure\" width=\"600\" height=\"141\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-filestructure-1-600x141.png 600w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-filestructure-1-250x59.png 250w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-filestructure-1-768x181.png 768w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-filestructure-1.png 891w\" \/><\/a><p class=\"wp-caption-text\">We should end up in a file structure like this.<\/p><\/div><h3>3) Configure PhpStorm<\/h3><p>The first thing we need to do is make sure if we\u2019ve configured an interpreter under <i>Preferences<\/i> &gt; <i>Languages &amp; Frameworks<\/i> &gt; <i>PHP<\/i>.<\/p><div id=\"attachment_617236\" class=\"wp-caption alignnone\"><a href=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter.png\"><img class=\"wp-image-617236 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter-600x444.png\" alt=\"Don\u2019t forget to add &lt;code&gt;wp-develop&lt;\/code&gt; and phpunit to your include path for some autocompletion niceness.\" width=\"600\" height=\"444\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter-600x444.png 600w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter-250x185.png 250w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter-768x568.png 768w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter-203x150.png 203w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-interpreter.png 1576w\" \/><\/a><p class=\"wp-caption-text\">Choose a PHP interpreter (1) or select a PHP interpreter in your filesystem (2). Then add wp-develop and the PHPUnit framework (3) to have some autocompletion niceness.<\/p><\/div><p>Next we need to tell PHPStorm where it can find PHPUnit. You can do this under <i>Preferences<\/i> &gt; <i>Languages<\/i> &gt; <i>PHP<\/i> &gt; <i>PHPUnit<\/i>.<\/p><div id=\"attachment_617237\" class=\"wp-caption alignnone\"><a href=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar.png\"><img class=\"wp-image-617237 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar-600x453.png\" alt=\"Choose for the option \u2018Path to phpunit.phar\u2019 and make sure it points to the phar file you have downloaded before.\" width=\"600\" height=\"453\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar-600x453.png 600w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar-250x189.png 250w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar-768x580.png 768w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar-199x150.png 199w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-configure-phar.png 1532w\" \/><\/a><p class=\"wp-caption-text\">Fill in the path to <code>phpunit.phar<\/code><\/p><\/div><p>We can now start configuring a testrunner. You can do this under under <i>Run<\/i> &gt; <i>Edit Configurations\u2026<\/i> where you\u2019ll get a screen like this:<\/p><div id=\"attachment_617239\" class=\"wp-caption alignnone\"><a href=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration.png\"><img class=\"wp-image-617239 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration-600x379.png\" alt=\"phpunit-setup-debug-configuration\" width=\"600\" height=\"379\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration-600x379.png 600w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration-180x114.png 180w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration-250x158.png 250w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration-768x485.png 768w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-setup-debug-configuration-237x150.png 237w\" \/><\/a><p class=\"wp-caption-text\">You can configure PHPUnit in four steps.<\/p><\/div><p>There are four steps to take:<\/p><ol><li>Give the configuration a name, for example: PHP Unit.<\/li>\n<li>Check the option \u2018defined in the configuration file\u2019.<\/li>\n<li>Check \u2018use alternative configuration file\u2019 and set the path to the phpunit.xml file.<\/li>\n<li>Set environment variables, you can do this by clicking the \u2018\u2026\u2019 after the input field. You\u2019ll get the next screen:<\/li>\n<\/ol><div id=\"attachment_617240\" class=\"wp-caption alignnone\"><a href=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables.png\"><img class=\"wp-image-617240 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables-600x600.png\" alt=\"phpunit-environment-variables\" width=\"600\" height=\"600\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables-600x600.png 600w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables-250x250.png 250w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables-768x768.png 768w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables-150x150.png 150w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-environment-variables.png 1208w\" \/><\/a><p class=\"wp-caption-text\">Add <code>WP_DEVELOP_DIR<\/code> as an environment variable.<\/p><\/div><p>You have to enter \u2018WP_DEVELOP_DIR\u2019 as name and the path to the <code>wp-develop<\/code> directory we\u2019ve cloned before. We use the <code>WP_DEVELOP_DIR<\/code> environment variable in our <code>bootstrap.php<\/code> to require the WordPress unit test <code>bootstrap.php<\/code>.<\/p><h3>Run tests<\/h3><p>For this article we write an example test to show how PHPUnit works. The tests have to be put in the <code>\/tests<\/code> folder.<\/p><p>Our example class will look something like:<\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\nclass Foo_Test extends WP_UnitTestCase {&#13;\n&#13;\n  public function test_foo_is_foo() {&#13;\n    $this-&gt;assertTrue( 'foo' === 'foo' );&#13;\n  }&#13;\n}&#13;\n<\/pre><p>We extend the <code>WP_UnitTestCase<\/code> because we are in a WordPress context. If this is not the case you can use <code>PHPUnit_Framework_TestCase<\/code> instead.<\/p><p>You can run the test by pressing the play icon in the run toolbar or under &gt;<i>Run<\/i>&gt; <i>Run \u2018PHP Unit\u2019<\/i>. PhpStorm will now open a small window in which it will run your tests:<\/p><div id=\"attachment_617244\" class=\"wp-caption alignnone\"><a href=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-run-test.png\"><img class=\"wp-image-617244 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-run-test-600x141.png\" alt=\"phpunit-run-test\" width=\"600\" height=\"141\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-run-test-600x141.png 600w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-run-test-250x59.png 250w, https:\/\/yoast.com\/app\/uploads\/2016\/02\/phpunit-run-test-768x180.png 768w\" \/><\/a><p class=\"wp-caption-text\">PhpStorm will show you a window like this when it runs tests.<\/p><\/div><p>Congratulations. You\u2019ve setup PhpStorm to run your unit tests.<\/p>","created_at":"2018-11-16T04:13:27+0000","updated_at":"2018-11-16T04:13:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"yoast.com","preview_picture":"https:\/\/yoast.com\/app\/uploads\/2015\/11\/Template_Dev_blog_FI-1.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","wordpress","phpunit"],"is_public":false,"id":688,"title":"An Introduction To Automated Testing Of WordPress Plugins With PHPUnit","url":"https:\/\/www.smashingmagazine.com\/2017\/12\/automated-testing-wordpress-plugins-phpunit\/","content":"<p class=\"article__summary\">\n\t\t\t\t\t\t\n\t\t\t\t\t\t\tYou really don\u2019t want to spend hours manually testing every part of your WordPress plugin to ensure nothing is broken every time you deploy a new version \u2014 do you? In this tutorial, you will learn how to test efficiently with automated testing.\n\t\t\t\t\t\t\n\t        <\/p><div class=\"drop-caps\" aria-hidden=\"true\"><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/07bab6a876338626943c46d654f45aabe7e0e807\/47054\/images\/drop-caps\/w.svg\" alt=\"\" \/><img src=\"https:\/\/d33wubrfki0l68.cloudfront.net\/4b6dbee344f771b99984ae6663d6e49bdeddbdef\/348d0\/images\/drop-caps\/character-8.svg\" alt=\"\" \/><\/div><p>WordPress is a popular content management system for building websites because it is easy to get started with and a ton of themes and plugins are available to extend its feature set. The main reason WordPress has a lot of plugins and themes is because it's easy for developers of any level to start building one. Most of its developers are not experienced, and they do not write tests for their work, perhaps because of the following reasons:<\/p><ul><li>There aren't many discussions going on about unit testing, so they might not know that testing is even possible.<\/li>\n<li>They do not believe in the value of writing tests for their code, or they think it will slow them down.<\/li>\n<li>They believe that testing to see whether their plugin or theme works in the browser is enough.<\/li>\n<\/ul><p>In this tutorial, we will learn what automated testing is and its importance, get to know PHPUnit and WP-CLI, learn how to write a test and, finally, set up continuous automated testing with Travis CI.<\/p><p>We are opting to use Travis CI because it offers seamless integration with GitHub; you don't have to go to your repository and set up any connection between them. And it is free for public repositories. Unlike its competitors, such as Semaphore CI, GitLab CI and CircleCI, Travis CI does not offer a free private repository plan. However, none of its competitors offer seamless integration with GitHub as it does.<\/p><h3>What Is Automated Testing?<\/h3><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Software_testing\">According to Wikipedia<\/a>, automated testing, or test automation, is the use of special software (separate from the software being tested) to control the execution of tests and the comparison of actual outcomes with predicted outcomes. Test automation can automate some repetitive but necessary tasks in a formalized testing process already in place, or perform additional testing that would be difficult to do manually.<\/p><p>There are several types of testing. Of all, unit testing is the most popular. Unit tests verify that a block of code, function or class method does what it is intended to do. We'll be doing unit testing in this tutorial.<\/p><p>Automated testing helps to detect bugs so that they don't make their way to production. No doubt, a plugin coded and tested would take longer to complete than one that is not tested. However, the resulting plugin would contain fewer or no bugs.<\/p><p>Let\u2019s see a simple real-world example of how unit tests are invaluable and what we can use them for.<\/p><p>My WordPress <a href=\"https:\/\/mailoptin.io\/\">lead-generation plugin<\/a> has an <code>OptinThemesRepository<\/code> class, with an <code>add()<\/code> method for adding new opt-in form templates, and a <code>get()<\/code> method for the retrieval of an opt-in form template.<\/p><p>To ensure both <code>add()<\/code> and <code>get()<\/code> work as intended now and in the future, I wrote the test below.<\/p><pre class=\"language-php\">public function testAddGetMethods()\n{\n    $kick_optin_form = array(\n        'name' =&gt; 'Kick',\n        'optin_class' =&gt; 'kick',\n        'optin_type' =&gt; 'kick',\n        'screenshot' =&gt; MAILOPTIN_ASSETS_URL . 'img\/kick.png'\n    );\n    \/\/ add kick optin theme\n    OptinThemesRepository::add($kick_optin_form);\n    $result = OptinThemesRepository::get('kick');\n    $this-&gt;assertEquals($kick_optin_form, $result);\n}<\/pre><p>If, in future, this test started to fail, I would know there is a problem and would know the exact function, class method or spot in my plugin where it is occurring.<\/p><h3>Benefits Of Automated Testing<\/h3><p>Now that we know what automated testing is, let's see more benefits.<\/p><h4>Early Bug Detection<\/h4><p>While developing software, you can easily find bugs with automated testing tools. This can save a lot of time and effort in tracking down bugs.<\/p><h4>Higher Software Quality<\/h4><p>A tester with many years of experience can make mistakes when they have to prepare the same boring manual test scripts over and over again. Automated testing not only yields accurate results, but also saves time.<\/p><h4>Easy and Robust Reporting<\/h4><p>Automated testing tools can track each and every test script. The execution of each test script can be seen in visual logs. The visual log, or report, typically displays the number of test scripts executed and their status (for example, passed, failed or skipped), their reported bugs and hints on how to fix the bugs.<\/p><p>Before we go over how to set up and write tests, let's create a simple plugin to use as a case study.<\/p><h3>Building A WordPress Plugin<\/h3><p>We are going to build a simple plugin that displays Google and Bing webmaster verification meta tags in the header of WordPress\u2019 front end. The plugin is hosted <a href=\"https:\/\/github.com\/collizo4sky\/wp-meta-verify\">in my GitHub account<\/a>.<\/p><p>The code for this plugin below will go in the <code>wp-meta-verify.php<\/code> file.<\/p><pre class=\"language-php\">&lt;?php \nclass WP_Meta_Verify \n{\n    public function __construct()\n    {\n        add_action('wp_head', \\[$this, 'header_code']);\n    }\n    public function header_code()\n    {\n        $google_code = get_option('wpmv_google_code');\n        $bing_code = get_option('wpmv_google_code');\n        echo $this-&gt;google_site_verification($google_code);\n        echo $this-&gt;bing_site_verification($bing_code);\n    }\n    public function google_site_verification($code)\n    {\n        return \"&lt;meta name=\\\"google-site-verification\\\" content=\\\"$code\\\"&gt;\";\n    }\n    public function bing_site_verification($code)\n    {\n        return \"&lt;meta name=\\\"msvalidate.01\\\" content=\\\"$code\\\"&gt;\";\n    } \n} \nnew WP_Meta_Verify();<\/pre><p>You might notice that we didn't include a settings page in the plugin, where you would typically save the Google and Bing verification code to. I did this on purpose to keep this simple and to focus our attention on what matters most. However, <code>get_option('wpmv_google_code')<\/code> and <code>get_option('wpmv_bing_code')<\/code> assume that there is a settings page, and they retrieve the verification codes from there.<\/p><h3>Unit Testing A WordPress Plugin<\/h3><p><a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a> is the <em>de facto<\/em> testing tool for PHP, whereas <a href=\"http:\/\/wp-cli.org\/\">WP-CLI<\/a> is the official command line interface for WordPress.<\/p><p>Prior to WP-CLI, <a href=\"https:\/\/make.wordpress.org\/core\/handbook\/testing\/automated-testing\/phpunit\/\">setting up PHPUnit testing<\/a> for WordPress plugins was a pain. WP-CLI has a great <a href=\"https:\/\/make.wordpress.org\/cli\/handbook\/plugin-unit-tests\/\">guide on setting it up<\/a>; however, we will still go over the steps here.<\/p><h4>Install PHPUnit<\/h4><p>To <a href=\"https:\/\/phpunit.de\/manual\/current\/en\/installation.html\">install PHPUnit<\/a>, run the following commands.<\/p><pre class=\"language-php\">composer global require phpunit\/phpunit:5.*<\/pre><p><strong>Note:<\/strong> We are explicitly installing <code>5.x<\/code> because that's what WordPress supports when you\u2019re running PHP 7 or above, which I have on my machine. Install PHPUnit 4.8 if you are running PHP version 5.<\/p><p>Run <code>phpunit --version<\/code> to confirm it's been installed.<\/p><h4>Install WP-CLI<\/h4><p>To install WP-CLI, run the following commands.<\/p><pre class=\"language-bash\">curl -O https:\/\/raw.githubusercontent.com\/wp-cli\/builds\/gh-pages\/phar\/wp-cli.phar \nchmod +x wp-cli.phar \nsudo mv wp-cli.phar \/usr\/local\/bin\/wp<\/pre><p>Run <code>wp --info<\/code> to confirm its installation.<\/p><p>Having installed PHPUnit and WP-CLI, we will use the latter to set up the unit test for the plugin.<\/p><h4>Set Up Plugin Unit Test<\/h4><p>Change your terminal\u2019s directory to the root of your WordPress installation, and run the command below to generate the plugin test files.<\/p><pre class=\"language-bash\">wp scaffold plugin-tests wp-meta-verify<\/pre><p>Below is what the structure of the plugin will look like after the command above generates the test files.<\/p><pre>|-bin\/\n|----install-wp-tests.sh\n|-tests\/\n|----bootstrap.php\n|----test-sample.php\n|-.travis.yml\n|-phpcs.xml.dist\n|-phpunit.xml.dist\n|-wp-meta-verify.php<\/pre><p><strong>Note:<\/strong> By default, the <code>wp scaffold plugin-tests<\/code> command generates a Travis CI configuration file. You can specify a <code>--ci<\/code> flag to generate a configuration file for the CI service you use, like so: <code>wp scaffold plugin-tests --c gitlab<\/code>. As at the time of writing, only Travis CI, CircleCI and GitLab CI are supported.<\/p><p>Change your terminal\u2019s directory to your plugin\u2019s directory, and run the installation script:<\/p><pre class=\"language-bash\">cd path-to-wordpress-plugin<\/pre><pre class=\"sourceCode bash\">bin\/install-wp-tests.sh wordpress_test root '' localhost latest<\/pre><p>If you\u2019re like me, then your MySQL username is not <code>root<\/code>, and the password is not empty. For example, suppose the username is <code>homestead<\/code> and the password is <code>secret<\/code>. You would run the installation script like so:<\/p><pre class=\"language-bash\">bin\/install-wp-tests.sh wordpress_test homestead 'secret' localhost latest<\/pre><p>Run the <code>phpunit<\/code> command to run the default test in <code>tests\/test-sample.php<\/code>.<\/p><figure><a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/c92d3a38-73e8-41d9-95d8-756f7440e53e\/1-phpunit-command-large-opt.png\"><img src=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/ab04b6d1-8532-438b-b0a0-695b157ec43b\/1-phpunit-command-800w-opt.png\" width=\"800\" height=\"343\" alt=\"PHPUnit test result\" \/><\/a><figcaption>PHPUnit test result (<a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/c92d3a38-73e8-41d9-95d8-756f7440e53e\/1-phpunit-command-large-opt.png\">Large preview<\/a>)<\/figcaption><\/figure><h4>Write Our Plugin Tests<\/h4><p>Create a <code>test-wp-meta-verify.php<\/code> file in the <code>tests<\/code> folder. It will contain our plugin tests with the following <code>setUp<\/code> class.<\/p><pre class=\"language-php\">&lt;?php \nclass WP_Meta_VerifyTest extends WP_UnitTestCase\n{\n    public function setUp()\n    {\n        parent::setUp();\n        $this-&gt;class_instance = new WP_Meta_Verify();\n    }\n    public function test_google_site_verification()\n    {\n    }\n    public function test_bing_site_verification()\n    {\n    }\n}<\/pre><p>It is worth noting that in order for a method to be considered a unit test, it must be prefixed with <code>test<\/code>. A best practice is to add a <code>Test<\/code> suffix to every test class, although it is not required. See <code>WP_Meta_VerifyTest<\/code>.<\/p><p>Confused about what <code>setUp()<\/code> does? Just know that PHPUnit runs it once before each test method (and on fresh instances) of the test case class. There is also <code>tearDown()<\/code>, but it is run after each test method. There are also <code>setUpBeforeClass()<\/code> and <code>tearDownAfterClass()<\/code>, which run before and after each test case, respectively. A test case is basically a class that contains a number of test methods. See the <a href=\"https:\/\/make.wordpress.org\/core\/handbook\/testing\/automated-testing\/writing-phpunit-tests\/#advanced-topics\">WordPress Handbook<\/a> and the <a href=\"https:\/\/phpunit.de\/manual\/current\/en\/fixtures.html\">PHPUnit documentation<\/a> for more information.<\/p><p>From the class above, it is pretty obvious we are going to be writing tests for the <code>google_site_verification<\/code> and <code>bing_site_verification<\/code> methods of our plugin class.<\/p><pre class=\"language-php\">public function test_google_site_verification()\n{\n    $meta_tag = $this-&gt;class_instance-&gt;google_site_verification('B6wFaCRbzWE42SyxSvKUOyyPxZfJCb5g');\n    $expected = '&lt;meta name=\"google-site-verification\" content=\"B6wFaCRbzWE42SyxSvKUOyyPxZfJCb5g\"&gt;';\n    $this-&gt;assertEquals($expected, $meta_tag);\n}<\/pre><pre class=\"language-php\">public function test_bing_site_verification()\n{\n    $meta_tag = $this-&gt;class_instance-&gt;bing_site_verification('B6wFaCRbzWE42SyxSvKUOyyPxZfJCb5g');\n    $expected = '&lt;meta name=\"msvalidate.01\" content=\"B6wFaCRbzWE42SyxSvKUOyyPxZfJCb5g\"&gt;';\n    $this-&gt;assertEquals($expected, $meta_tag);\n}<\/pre><p>Basically, the tests will ensure that both methods return the correct meta tag when Google and Bing webmaster verification codes are passed to them as arguments.<\/p><p>Run <code>phpunit<\/code>, and you should see an output similar to the screenshot below.<\/p><figure><a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/c11f167f-3111-4452-b8ae-ca751b115d24\/2-phpunit-command-output-large-opt.png\"><img src=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/e49004d0-af58-428b-82be-ba9be27be047\/2-phpunit-command-output-800w-opt.png\" width=\"800\" height=\"347\" alt=\"PHPUnit output\" \/><\/a><figcaption>PHPUnit output (<a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/c11f167f-3111-4452-b8ae-ca751b115d24\/2-phpunit-command-output-large-opt.png\">Large preview<\/a>)<\/figcaption><\/figure><h3>Continuous Automated Testing With Travis CI<\/h3><p><a href=\"https:\/\/travis-ci.org\/\">Travis CI<\/a> is a hosted, distributed continuous integration service used to build and test software projects hosted on GitHub.<\/p><p>To use Travis CI, therefore, we have to publish our plugin on GitHub. Go ahead and <a href=\"https:\/\/help.github.com\/articles\/adding-an-existing-project-to-github-using-the-command-line\/\">do that now<\/a>. Feel free to <a href=\"https:\/\/github.com\/collizo4sky\/wp-meta-verify\">refer to mine<\/a>.<\/p><p>Thanks to WP-CLI, we already have it set up in our plugin, courtesy of the <code>.travis.yml<\/code> file.<\/p><p>I would like to mention that I adhere not to <a href=\"https:\/\/make.wordpress.org\/core\/handbook\/best-practices\/coding-standards\/php\/\">WordPress coding standards<\/a>, but rather to <a href=\"http:\/\/www.php-fig.org\/psr\/\">PHP Standards Recommendations<\/a>, and my plugins require at least PHP 5.4. In order for my <a href=\"https:\/\/travis-ci.org\/collizo4sky\/wp-meta-verify\/builds\/264324465\">builds not to fail<\/a>, I had to replace their matrix with the following in <code>.travis.yml<\/code> file.<\/p><pre class=\"language-markup\">matrix:\n    include:\n        - php: 7.1\n        env:\u00a0WP_VERSION=latest\n        - php: 7.0\n        env:\u00a0WP_VERSION=latest\n        - php: 5.6\n        env:\u00a0WP_VERSION=latest\n        - php: 5.6\n        env:\u00a0WP_VERSION=trunk\n        - php: 5.5\n        env:\u00a0WP_VERSION=latest\n        - php: 5.4\n        env:\u00a0WP_VERSION=latest<\/pre><p>Head over to <a href=\"https:\/\/travis-ci.org\/\">Travis CI<\/a> and sign in with your GitHub account. Follow the on-screen guide to add your GitHub repository.<\/p><p>After account synchronization with GitHub, scroll to your plugin\u2019s repository and activate.<\/p><figure><a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/9efa469e-8432-4e99-a234-522de79b6ca3\/3-phpunit-activate-plugin-repository-large-opt.png\"><img src=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/4e18ae68-527a-4409-83c4-20ce0714a114\/3-phpunit-activate-plugin-repository-800w-opt.png\" width=\"800\" height=\"242\" alt=\"Setting up Travis CI repository\" \/><\/a><figcaption>Setting up Travis CI repository (<a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/9efa469e-8432-4e99-a234-522de79b6ca3\/3-phpunit-activate-plugin-repository-large-opt.png\">Large preview<\/a>)<\/figcaption><\/figure><p>The next time you make a code change and push to GitHub, a build will be triggered on Travis CI.<\/p><figure><a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/5497aa2b-b4f4-47f1-9ee6-8add99c364d2\/4-phpunit-build-travisci-large-opt.png\"><img src=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/0e6aa73b-0627-49ed-873b-6d38bf54e5e6\/4-phpunit-build-travisci-800w-opt.png\" width=\"800\" height=\"501\" alt=\"\" \/><\/a><figcaption>Travis CI build result (<a href=\"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/5497aa2b-b4f4-47f1-9ee6-8add99c364d2\/4-phpunit-build-travisci-large-opt.png\">Large preview<\/a>)<\/figcaption><\/figure><p>I\u2019ve made available a <a href=\"https:\/\/travis-ci.org\/collizo4sky\/wp-meta-verify\/builds\/264329247\">successful build result<\/a> for your viewing pleasure.<\/p><h3>Wrapping Up<\/h3><p>It's no secret that a lot of developers, not just WordPress ones, do not write tests for their projects because they don't know about them. Even some of the experienced and advanced among us don't apparently because they consider it a waste of time.<\/p><p>Granted, setting up automated testing can be boring and time-consuming. Nevertheless, it's an investment that will ensure that few or no bugs creep into your software, thereby saving you the time and resources (including financial) that bugs in your software would have cost you.<\/p><p>Always write a test before implementing a feature, so that you don't forget or feel lazy to do it after the feature has been implemented.<\/p><p>I hope you now recognize the importance of writing tests and how to start writing one for your own WordPress plugin.<\/p><p>If you have any questions or comments, please let me know in the comments section.<\/p>","created_at":"2018-11-16T04:11:54+0000","updated_at":"2018-11-16T04:12:00+0000","published_at":"2017-12-14T14:38:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"www.smashingmagazine.com","preview_picture":"https:\/\/cloud.netlifyusercontent.com\/assets\/344dbf88-fdf9-42bb-adb4-46f01eedd629\/ab04b6d1-8532-438b-b0a0-695b157ec43b\/1-phpunit-command-800w-opt.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["php","doctrine","orm"],"is_public":false,"id":687,"title":"Getting Started with Doctrine - Object Relational Mapper (ORM) - Doctrine","url":"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/current\/tutorials\/getting-started.html","content":"<div class=\"dropdown show d-inline-block\"><p>\n                                                                                            \n                                                <a class=\"project-version-switcher dropdown-item\" href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/latest\/tutorials\/getting-started.html\">\n                                                    3.0\n                                                    \n                                                                                                            (upcoming)\n                                                                                                    <\/a>\n                                                                                            \n                                                <a class=\"project-version-switcher dropdown-item\" href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/2.7\/tutorials\/getting-started.html\">\n                                                    2.7\n                                                    \n                                                                                                            (upcoming)\n                                                                                                    <\/a>\n                                                                                            \n                                                <a class=\"project-version-switcher dropdown-item active\" href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/2.6\/tutorials\/getting-started.html\">\n                                                    2.6\n                                                                                                            (current)\n                                                    \n                                                                                                    <\/a>\n                                                                                    \n                                                                                    \n                                                                                            \n                                                <a class=\"dropdown-item\" href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/2.5\/tutorials\/getting-started.html\">2.5<\/a>\n                                                                                            \n                                                <a class=\"dropdown-item\" href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/2.4\/tutorials\/getting-started.html\">2.4<\/a><\/p><\/div><p>This guide covers getting started with the Doctrine ORM. After working\nthrough the guide you should know:<\/p><ul><li class=\"dash\">How to install and configure Doctrine by connecting it to a database<\/li>\n<li class=\"dash\">Mapping PHP objects to database tables<\/li>\n<li class=\"dash\">Generating a database schema from PHP objects<\/li>\n<li class=\"dash\">Using the <code>EntityManager<\/code> to insert, update, delete and find\n  objects in the database.<\/li>\n<\/ul><p>This guide is designed for beginners that haven't worked with Doctrine ORM\nbefore. There are some prerequisites for the tutorial that have to be\ninstalled:<\/p><ul><li class=\"dash\">PHP (latest stable version)<\/li>\n<li class=\"dash\">Composer Package Manager (<a href=\"http:\/\/getcomposer.org\/doc\/00-intro.md\">Install Composer\n <\/a>)<\/li>\n<\/ul><p>The code of this tutorial is <a href=\"https:\/\/github.com\/doctrine\/doctrine2-orm-tutorial\">available on Github<\/a>.<\/p><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>This tutorial assumes you work with <strong>Doctrine 2.6<\/strong> and above.\nSome of the code will not work with lower versions.<\/p>\n<\/td><\/tr><\/table><\/div><p>Doctrine 2 is an <a href=\"#object-relational-mapper-orm-http-en-wikipedia-org-wiki-object-relational-mapping\">object-relational mapper (ORM)\n&lt;http:\/\/en.wikipedia.org\/wiki\/Object-relational_mapping&gt;<\/a> for PHP 5.4+ that\nprovides transparent persistence for PHP objects. It uses the Data Mapper\npattern at the heart, aiming for a complete separation of your domain\/business\nlogic from the persistence in a relational database management system.<\/p><p>The benefit of Doctrine for the programmer is the ability to focus\non the object-oriented business logic and worry about persistence only\nas a secondary problem. This doesn't mean persistence is downplayed by Doctrine\n2, however it is our belief that there are considerable benefits for\nobject-oriented programming if persistence and entities are kept\nseparated.<\/p><p>Entities are PHP Objects that can be identified over many requests\nby a unique identifier or primary key. These classes don't need to extend any\nabstract base class or interface. An entity class must not be final\nor contain final methods. Additionally it must not implement\n<strong>clone<\/strong> nor <strong>wakeup<\/strong>, unless it <a href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/current\/cookbook\/implementing-wakeup-or-clone.html\">does so safely<\/a>.<\/p><p>An entity contains persistable properties. A persistable property\nis an instance variable of the entity that is saved into and retrieved from the database\nby Doctrine's data mapping capabilities.<\/p><p>For this Getting Started Guide for Doctrine we will implement the\nBug Tracker domain model from the\n<a href=\"http:\/\/framework.zend.com\/manual\/1.12\/en\/zend.db.adapter.html\">Zend\\_Db\\_Table<\/a>\ndocumentation. Reading their documentation we can extract the\nrequirements:<\/p><ul><li class=\"dash\"> A Bug has a description, creation date, status, reporter and\n   engineer<\/li>\n<li class=\"dash\"> A Bug can occur on different Products (platforms)<\/li>\n<li class=\"dash\"> A Product has a name.<\/li>\n<li class=\"dash\"> Bug reporters and engineers are both Users of the system.<\/li>\n<li class=\"dash\"> A User can create new Bugs.<\/li>\n<li class=\"dash\"> The assigned engineer can close a Bug.<\/li>\n<li class=\"dash\"> A User can see all his reported or assigned Bugs.<\/li>\n<li class=\"dash\"> Bugs can be paginated through a list-view.<\/li>\n<\/ul><p>Create a new empty folder for this tutorial project, for example\n<code>doctrine2-tutorial<\/code> and create a new file <code>composer.json<\/code> inside\nthat directory with the following contents:<\/p><div class=\"console\"><pre class=\"console\">{\n    \"require\": {\n        \"doctrine\/orm\": \"^2.6.2\",\n        \"symfony\/yaml\": \"2.*\"\n    },\n    \"autoload\": {\n        \"psr-0\": {\"\": \"src\/\"}\n    }\n}<\/pre><\/div><p>Install Doctrine using the Composer Dependency Management tool, by calling:<\/p><div class=\"console\"><pre class=\"console\">$ composer install<\/pre><\/div><p>This will install the packages Doctrine Common, Doctrine DBAL, Doctrine ORM,\ninto the <code>vendor<\/code> directory.<\/p><p>Add the following directories:<\/p><div class=\"console\"><pre class=\"console\">doctrine2-tutorial\n|-- config\n|   `-- xml\n|   `-- yaml\n`-- src<\/pre><\/div><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>The YAML driver is deprecated and will be removed in version 3.0.\nIt is strongly recommended to switch to one of the other mappings.<\/p>\n<\/td><\/tr><\/table><\/div><p>Doctrine's public interface is through the <code>EntityManager<\/code>. This class\nprovides access points to the complete lifecycle management for your entities,\nand transforms entities from and back to persistence. You have to\nconfigure and create it to use your entities with Doctrine 2. I\nwill show the configuration steps and then discuss them step by\nstep:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"22\">&lt;?php\n\/\/ bootstrap.php\nuse Doctrine\\ORM\\Tools\\Setup;\nuse Doctrine\\ORM\\EntityManager;\nrequire_once \"vendor\/autoload.php\";\n\/\/ Create a simple \"default\" Doctrine ORM configuration for Annotations\n$isDevMode = true;\n$config = Setup::createAnnotationMetadataConfiguration(array(__DIR__.\"\/src\"), $isDevMode);\n\/\/ or if you prefer yaml or XML\n\/\/$config = Setup::createXMLMetadataConfiguration(array(__DIR__.\"\/config\/xml\"), $isDevMode);\n\/\/$config = Setup::createYAMLMetadataConfiguration(array(__DIR__.\"\/config\/yaml\"), $isDevMode);\n\/\/ database configuration parameters\n$conn = array(\n    'driver' =&gt; 'pdo_sqlite',\n    'path' =&gt; __DIR__ . '\/db.sqlite',\n);\n\/\/ obtaining the entity manager\n$entityManager = EntityManager::create($conn, $config);<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-bafc1b9092783eb6d544808abc40a88fd3513e36-22\">22<\/a><\/td><\/tr><\/table><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>The YAML driver is deprecated and will be removed in version 3.0.\nIt is strongly recommended to switch to one of the other mappings.<\/p>\n<\/td><\/tr><\/table><\/div><p>The <code>require_once<\/code> statement sets up the class autoloading for Doctrine and\nits dependencies using Composer's autoloader.<\/p><p>The second block consists of the instantiation of the ORM\n<code>Configuration<\/code> object using the Setup helper. It assumes a bunch\nof defaults that you don't have to bother about for now. You can\nread up on the configuration details in the\n<a href=\"https:\/\/www.doctrine-project.org\/projects\/doctrine-orm\/en\/current\/reference\/configuration.html\">reference chapter on configuration<\/a>.<\/p><p>The third block shows the configuration options required to connect to\na database. In this case, we'll use a file-based SQLite database. All the\nconfiguration options for all the shipped drivers are given in the\n<a href=\"http:\/\/docs.doctrine-project.org\/projects\/doctrine-dbal\/en\/latest\/\">DBAL Configuration section of the manual<\/a>.<\/p><p>The last block shows how the <code>EntityManager<\/code> is obtained from a\nfactory method.<\/p><p>Doctrine has a command-line interface that allows you to access the SchemaTool,\na component that can generate a relational database schema based entirely on the\ndefined entity classes and their metadata. For this tool to work, a\n<code>cli-config.php<\/code> file must exist in the project root directory:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-da7cbfb681faad69925b8674469c9de87e64e002-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"5\">&lt;?php\n\/\/ cli-config.php\nrequire_once \"bootstrap.php\";\nreturn \\Doctrine\\ORM\\Tools\\Console\\ConsoleRunner::createHelperSet($entityManager);<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-da7cbfb681faad69925b8674469c9de87e64e002-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-da7cbfb681faad69925b8674469c9de87e64e002-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-da7cbfb681faad69925b8674469c9de87e64e002-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-da7cbfb681faad69925b8674469c9de87e64e002-5\">5<\/a><\/td><\/tr><\/table><p>Now call the Doctrine command-line tool:<\/p><div class=\"console\"><pre class=\"console\">$ vendor\/bin\/doctrine orm:schema-tool:create<\/pre><\/div><p>Since we haven't added any entity metadata in <code>src<\/code> yet, you'll see a message\nstating \"No Metadata Classes to process.\" In the next section, we'll create a\nProduct entity along with the corresponding metadata, and run this command again.<\/p><p>Note that as you modify your entities' metadata during the development process,\nyou'll need to update your database schema to stay in sync with the metadata.\nYou can easily recreate the database using the following commands:<\/p><div class=\"console\"><pre class=\"console\">$ vendor\/bin\/doctrine orm:schema-tool:drop --force\n$ vendor\/bin\/doctrine orm:schema-tool:create<\/pre><\/div><p>Or you can use the update functionality:<\/p><div class=\"console\"><pre class=\"console\">$ vendor\/bin\/doctrine orm:schema-tool:update --force<\/pre><\/div><p>The updating of databases uses a diff algorithm for a given\ndatabase schema. This is a cornerstone of the <code>Doctrine\\DBAL<\/code> package,\nwhich can even be used without the Doctrine ORM package.<\/p><p>We start with the simplest entity, the Product. Create a <code>src\/Product.php<\/code> file to contain the <code>Product<\/code>\nentity definition:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"28\">&lt;?php\n\/\/ src\/Product.php\nclass Product\n{\n    \/**\n     * @var int\n     *\/\n    protected $id;\n    \/**\n     * @var string\n     *\/\n    protected $name;\n    public function getId()\n    {\n        return $this-&gt;id;\n    }\n    public function getName()\n    {\n        return $this-&gt;name;\n    }\n    public function setName($name)\n    {\n        $this-&gt;name = $name;\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2a361833634b9b76a9b4155e66e54f0224b1fa2-28\">28<\/a><\/td><\/tr><\/table><p>When creating entity classes, all of the fields should be <code>protected<\/code> or <code>private<\/code>\n(not <code>public<\/code>), with getter and setter methods for each one (except <code>$id<\/code>).\nThe use of mutators allows Doctrine to hook into calls which\nmanipulate the entities in ways that it could not if you just\ndirectly set the values with <code>entity#field = foo;<\/code><\/p><p>The id field has no setter since, generally speaking, your code\nshould not set this value since it represents a database id value.\n(Note that Doctrine itself can still set the value using the\nReflection API instead of a defined setter function.)<\/p><p>The next step for persistence with Doctrine is to describe the\nstructure of the <code>Product<\/code> entity to Doctrine using a metadata\nlanguage. The metadata language describes how entities, their\nproperties and references should be persisted and what constraints\nshould be applied to them.<\/p><p>Metadata for an Entity can be configured using DocBlock annotations directly\nin the Entity class itself, or in an external XML or YAML file. This Getting\nStarted guide will demonstrate metadata mappings using all three methods,\nbut you only need to choose one.<\/p><div class=\"configuration-block jsactive clearfix\"><ul class=\"simple\"><li><em>PHP<\/em><div class=\"highlight-php\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"14\">&lt;?php\n\/\/ src\/Product.php\n\/**\n * @Entity @Table(name=\"products\")\n **\/\nclass Product\n{\n    \/** @Id @Column(type=\"integer\") @GeneratedValue **\/\n    protected $id;\n    \/** @Column(type=\"string\") **\/\n    protected $name;\n    \/\/ .. (other code)\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0c5c83ba85d89d86c897ba08a01b406077852d19-14\">14<\/a><\/td><\/tr><\/table><\/div><\/li><li><em>XML<\/em><div class=\"highlight-xml\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"14\">&lt;!-- config\/xml\/Product.dcm.xml --&gt;\n&lt;doctrine-mapping xmlns=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\"\n      xmlns:xsi=\"http:\/\/www.w3.org\/2001\/XMLSchema-instance\"\n      xsi:schemaLocation=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\n                          https:\/\/www.doctrine-project.org\/schemas\/orm\/doctrine-mapping.xsd\"&gt;\n      &lt;entity name=\"Product\" table=\"products\"&gt;\n          &lt;id name=\"id\" type=\"integer\"&gt;\n              &lt;generator strategy=\"AUTO\" \/&gt;\n          &lt;\/id&gt;\n          &lt;field name=\"name\" type=\"string\" \/&gt;\n      &lt;\/entity&gt;\n&lt;\/doctrine-mapping&gt;<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-89937f7650184e11e44d1b365a29aea536c19bb5-14\">14<\/a><\/td><\/tr><\/table><\/div><\/li><\/ul><\/div><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>The YAML driver is deprecated and will be removed in version 3.0.\nIt is strongly recommended to switch to one of the other mappings.<\/p>\n<table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"12\"># config\/yaml\/Product.dcm.yml\nProduct:\n  type: entity\n  table: products\n  id:\n    id:\n      type: integer\n      generator:\n        strategy: AUTO\n  fields:\n    name:\n      type: string<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-ee6aa2b406d9b4141d5a45300f9677bddd667c40-12\">12<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><\/div><p>The top-level <code>entity<\/code> definition specifies information about\nthe class and table name. The primitive type <code>Product#name<\/code> is\ndefined as a <code>field<\/code> attribute. The <code>id<\/code> property is defined with\nthe <code>id<\/code> tag.  It has a <code>generator<\/code> tag nested inside, which\nspecifies that the primary key generation mechanism should automatically\nuse the database platform's native id generation strategy (for\nexample, AUTO INCREMENT in the case of MySql, or Sequences in the\ncase of PostgreSql and Oracle).<\/p><p>Now that we have defined our first entity and its metadata,\nlet's update the database schema:<\/p><div class=\"console\"><pre class=\"console\">$ vendor\/bin\/doctrine orm:schema-tool:update --force --dump-sql<\/pre><\/div><p>Specifying both flags <code>--force<\/code> and <code>--dump-sql<\/code> will cause the DDL\nstatements to be executed and then printed to the screen.<\/p><p>Now, we'll create a new script to insert products into the database:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"13\">&lt;?php\n\/\/ create_product.php &lt;name&gt;\nrequire_once \"bootstrap.php\";\n$newProductName = $argv[1];\n$product = new Product();\n$product-&gt;setName($newProductName);\n$entityManager-&gt;persist($product);\n$entityManager-&gt;flush();\necho \"Created Product with ID \" . $product-&gt;getId() . \"\\n\";<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b5be97232c4ddb958a1d4e0e1952c4fb66af0296-13\">13<\/a><\/td><\/tr><\/table><p>Call this script from the command-line to see how new products are created:<\/p><div class=\"console\"><pre class=\"console\">$ php create_product.php ORM\n$ php create_product.php DBAL<\/pre><\/div><p>What is happening here? Using the <code>Product<\/code> class is pretty standard OOP.\nThe interesting bits are the use of the <code>EntityManager<\/code> service. To\nnotify the EntityManager that a new entity should be inserted into the database,\nyou have to call <code>persist()<\/code>. To initiate a transaction to actually <em>perform<\/em>\nthe insertion, you have to explicitly call <code>flush()<\/code> on the <code>EntityManager<\/code>.<\/p><p>This distinction between persist and flush is what allows the aggregation of\nall database writes (INSERT, UPDATE, DELETE) into one single transaction, which\nis executed when <code>flush()<\/code> is called. Using this approach, the write-performance\nis significantly better than in a scenario in which writes are performed on\neach entity in isolation.<\/p><p>Next, we'll fetch a list of all the Products in the database. Let's create a\nnew script for this:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"10\">&lt;?php\n\/\/ list_products.php\nrequire_once \"bootstrap.php\";\n$productRepository = $entityManager-&gt;getRepository('Product');\n$products = $productRepository-&gt;findAll();\nforeach ($products as $product) {\n    echo sprintf(\"-%s\\n\", $product-&gt;getName());\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-7350b7f5fc99224764363f04357f7143f3cd5a89-10\">10<\/a><\/td><\/tr><\/table><p>The <code>EntityManager#getRepository()<\/code> method can create a finder object (called\na repository) for every type of entity. It is provided by Doctrine and contains\nsome finder methods like <code>findAll()<\/code>.<\/p><p>Let's continue by creating a script to display the name of a product based on its ID:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"13\">&lt;?php\n\/\/ show_product.php &lt;id&gt;\nrequire_once \"bootstrap.php\";\n$id = $argv[1];\n$product = $entityManager-&gt;find('Product', $id);\nif ($product === null) {\n    echo \"No product found.\\n\";\n    exit(1);\n}\necho sprintf(\"-%s\\n\", $product-&gt;getName());<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-3e9150d7f488942a4a9ad2be3ab9b15338c405a1-13\">13<\/a><\/td><\/tr><\/table><p>Next we'll update a product's name, given its id. This simple example will\nhelp demonstrate Doctrine's implementation of the UnitOfWork pattern. Doctrine\nkeeps track of all the entities that were retrieved from the Entity Manager,\nand can detect when any of those entities' properties have been modified.\nAs a result, rather than needing to call <code>persist($entity)<\/code> for each individual\nentity whose properties were changed, a single call to <code>flush()<\/code> at the end of a\nrequest is sufficient to update the database for all of the modified entities.<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"17\">&lt;?php\n\/\/ update_product.php &lt;id&gt; &lt;new-name&gt;\nrequire_once \"bootstrap.php\";\n$id = $argv[1];\n$newName = $argv[2];\n$product = $entityManager-&gt;find('Product', $id);\nif ($product === null) {\n    echo \"Product $id does not exist.\\n\";\n    exit(1);\n}\n$product-&gt;setName($newName);\n$entityManager-&gt;flush();<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e3b85e42876f5b73e20012b3d2847b9b18d59623-17\">17<\/a><\/td><\/tr><\/table><p>After calling this script on one of the existing products, you can verify the\nproduct name changed by calling the <code>show_product.php<\/code> script.<\/p><p>We continue with the bug tracker example by creating the <code>Bug<\/code> and <code>User<\/code>\nclasses. We'll store them in <code>src\/Bug.php<\/code> and <code>src\/User.php<\/code>, respectively.<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"63\">&lt;?php\n\/\/ src\/Bug.php\n\/**\n * @Entity(repositoryClass=\"BugRepository\") @Table(name=\"bugs\")\n *\/\nclass Bug\n{\n    \/**\n     * @Id @Column(type=\"integer\") @GeneratedValue\n     * @var int\n     *\/\n    protected $id;\n    \/**\n     * @Column(type=\"string\")\n     * @var string\n     *\/\n    protected $description;\n    \/**\n     * @Column(type=\"datetime\")\n     * @var DateTime\n     *\/\n    protected $created;\n    \/**\n     * @Column(type=\"string\")\n     * @var string\n     *\/\n    protected $status;\n    public function getId()\n    {\n        return $this-&gt;id;\n    }\n    public function getDescription()\n    {\n        return $this-&gt;description;\n    }\n    public function setDescription($description)\n    {\n        $this-&gt;description = $description;\n    }\n    public function setCreated(DateTime $created)\n    {\n        $this-&gt;created = $created;\n    }\n    public function getCreated()\n    {\n        return $this-&gt;created;\n    }\n    public function setStatus($status)\n    {\n        $this-&gt;status = $status;\n    }\n    public function getStatus()\n    {\n        return $this-&gt;status;\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-31\">31<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-32\">32<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-33\">33<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-34\">34<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-35\">35<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-36\">36<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-37\">37<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-38\">38<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-39\">39<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-40\">40<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-41\">41<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-42\">42<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-43\">43<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-44\">44<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-45\">45<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-46\">46<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-47\">47<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-48\">48<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-49\">49<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-50\">50<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-51\">51<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-52\">52<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-53\">53<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-54\">54<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-55\">55<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-56\">56<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-57\">57<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-58\">58<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-59\">59<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-60\">60<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-61\">61<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-62\">62<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b2355de7261e145a6d719166ba4d3566d6eaa1a7-63\">63<\/a><\/td><\/tr><\/table><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"33\">&lt;?php\n\/\/ src\/User.php\n\/**\n * @Entity @Table(name=\"users\")\n *\/\nclass User\n{\n    \/**\n     * @Id @GeneratedValue @Column(type=\"integer\")\n     * @var int\n     *\/\n    protected $id;\n    \/**\n     * @Column(type=\"string\")\n     * @var string\n     *\/\n    protected $name;\n    public function getId()\n    {\n        return $this-&gt;id;\n    }\n    public function getName()\n    {\n        return $this-&gt;name;\n    }\n    public function setName($name)\n    {\n        $this-&gt;name = $name;\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-31\">31<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-32\">32<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4822e20dca360d8827fff869859a0602806273a4-33\">33<\/a><\/td><\/tr><\/table><p>All of the properties we've seen so far are of simple types (integer, string,\nand datetime). But now, we'll add properties that will store objects of\nspecific <em>entity types<\/em> in order to model the relationships between different\nentities.<\/p><p>At the database level, relationships between entities are represented by foreign\nkeys. But with Doctrine, you'll never have to (and never should) work with\nthe foreign keys directly. You should only work with objects that represent\nforeign keys through their own identities.<\/p><p>For every foreign key you either have a Doctrine ManyToOne or OneToOne\nassociation. On the inverse sides of these foreign keys you can have\nOneToMany associations. Obviously you can have ManyToMany associations\nthat connect two tables with each other through a join table with<br \/>two foreign keys.<\/p><p>Now that you know the basics about references in Doctrine, we can extend the\ndomain model to match the requirements:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"15\">&lt;?php\n\/\/ src\/Bug.php\nuse Doctrine\\Common\\Collections\\ArrayCollection;\nclass Bug\n{\n    \/\/ ... (previous code)\n    protected $products;\n    public function __construct()\n    {\n        $this-&gt;products = new ArrayCollection();\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d489172872c136eee932e0da955165b4ff889825-15\">15<\/a><\/td><\/tr><\/table><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"17\">&lt;?php\n\/\/ src\/User.php\nuse Doctrine\\Common\\Collections\\ArrayCollection;\nclass User\n{\n    \/\/ ... (previous code)\n    protected $reportedBugs;\n    protected $assignedBugs;\n    public function __construct()\n    {\n        $this-&gt;reportedBugs = new ArrayCollection();\n        $this-&gt;assignedBugs = new ArrayCollection();\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-014b94f1d825ad754121e906dc29e7525d4f54ce-17\">17<\/a><\/td><\/tr><\/table><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>Whenever an entity is created from the database, a <code>Collection<\/code>\nimplementation of the type <code>PersistentCollection<\/code> will be injected into\nyour entity instead of an <code>ArrayCollection<\/code>. This helps Doctrine ORM\nunderstand the changes that have happened to the collection that are\nnoteworthy for persistence.<\/p>\n<\/td><\/tr><\/table><\/div><div class=\"alert warning-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-exclamation-circle text-warning mr-2\"><\/i><\/td><td><p>Lazy load proxies always contain an instance of\nDoctrine's EntityManager and all its dependencies. Therefore a\nvar\\_dump() will possibly dump a very large recursive structure\nwhich is impossible to render and read. You have to use\n<code>Doctrine\\Common\\Util\\Debug::dump()<\/code> to restrict the dumping to a\nhuman readable level. Additionally you should be aware that dumping\nthe EntityManager to a Browser may take several minutes, and the\nDebug::dump() method just ignores any occurrences of it in Proxy\ninstances.<\/p>\n<\/td><\/tr><\/table><\/div><p>Because we only work with collections for the references we must be\ncareful to implement a bidirectional reference in the domain model.\nThe concept of owning or inverse side of a relation is central to\nthis notion and should always be kept in mind. The following\nassumptions are made about relations and have to be followed to be\nable to work with Doctrine 2. These assumptions are not unique to\nDoctrine 2 but are best practices in handling database relations\nand Object-Relational Mapping.<\/p><ul><li class=\"dash\"> In a one-to-one relation, the entity holding the foreign key of\n   the related entity on its own database table is <em>always<\/em> the owning\n   side of the relation.<\/li>\n<li class=\"dash\"> In a many-to-one relation, the Many-side is the owning side by\n   default because it holds the foreign key. Accordingly, the One-side\n   is the inverse side by default.<\/li>\n<li class=\"dash\"> In a many-to-one relation, the One-side can only be the owning side if\n   the relation is implemented as a ManyToMany with a join table, and the\n   One-side is restricted to allow only UNIQUE values per database constraint.<\/li>\n<li class=\"dash\"> In a many-to-many relation, both sides can be the owning side of\n   the relation. However, in a bi-directional many-to-many relation,\n   only one side is allowed to be the owning side.<\/li>\n<li class=\"dash\"> Changes to Collections are saved or updated, when the entity on\n   the <em>owning<\/em> side of the collection is saved or updated.<\/li>\n<li class=\"dash\"> Saving an Entity at the inverse side of a relation never\n   triggers a persist operation to changes to the collection.<\/li>\n<\/ul><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>Consistency of bi-directional references on the inverse side of a\nrelation have to be managed in userland application code. Doctrine\ncannot magically update your collections to be consistent.<\/p>\n<\/td><\/tr><\/table><\/div><p>In the case of Users and Bugs we have references back and forth to\nthe assigned and reported bugs from a user, making this relation\nbi-directional. We have to change the code to ensure consistency of\nthe bi-directional reference:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"31\">&lt;?php\n\/\/ src\/Bug.php\nclass Bug\n{\n    \/\/ ... (previous code)\n    protected $engineer;\n    protected $reporter;\n    public function setEngineer(User $engineer)\n    {\n        $engineer-&gt;assignedToBug($this);\n        $this-&gt;engineer = $engineer;\n    }\n    public function setReporter(User $reporter)\n    {\n        $reporter-&gt;addReportedBug($this);\n        $this-&gt;reporter = $reporter;\n    }\n    public function getEngineer()\n    {\n        return $this-&gt;engineer;\n    }\n    public function getReporter()\n    {\n        return $this-&gt;reporter;\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-40a5a79273fd3e3956abe51cde3498f6d81a5493-31\">31<\/a><\/td><\/tr><\/table><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"19\">&lt;?php\n\/\/ src\/User.php\nclass User\n{\n    \/\/ ... (previous code)\n    protected $reportedBugs;\n    protected $assignedBugs;\n    public function addReportedBug(Bug $bug)\n    {\n        $this-&gt;reportedBugs[] = $bug;\n    }\n    public function assignedToBug(Bug $bug)\n    {\n        $this-&gt;assignedBugs[] = $bug;\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-af14539eae207e7e951b7f66f813b9bc451ab597-19\">19<\/a><\/td><\/tr><\/table><p>I chose to name the inverse methods in past-tense, which should\nindicate that the actual assigning has already taken place and the\nmethods are only used for ensuring consistency of the references.\nThis approach is my personal preference, you can choose whatever\nmethod to make this work.<\/p><p>You can see from <code>User#addReportedBug()<\/code> and\n<code>User#assignedToBug()<\/code> that using this method in userland alone\nwould not add the Bug to the collection of the owning side in\n<code>Bug#reporter<\/code> or <code>Bug#engineer<\/code>. Using these methods and\ncalling Doctrine for persistence would not update the Collections'\nrepresentation in the database.<\/p><p>Only using <code>Bug#setEngineer()<\/code> or <code>Bug#setReporter()<\/code>\ncorrectly saves the relation information.<\/p><p>The <code>Bug#reporter<\/code> and <code>Bug#engineer<\/code> properties are\nMany-To-One relations, which point to a User. In a normalized\nrelational model, the foreign key is saved on the Bug's table, hence\nin our object-relation model the Bug is at the owning side of the\nrelation. You should always make sure that the use-cases of your\ndomain model should drive which side is an inverse or owning one in\nyour Doctrine mapping. In our example, whenever a new bug is saved\nor an engineer is assigned to the bug, we don't want to update the\nUser to persist the reference, but the Bug. This is the case with\nthe Bug being at the owning side of the relation.<\/p><p>Bugs reference Products by a uni-directional ManyToMany relation in\nthe database that points from Bugs to Products.<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"18\">&lt;?php\n\/\/ src\/Bug.php\nclass Bug\n{\n    \/\/ ... (previous code)\n    protected $products = null;\n    public function assignToProduct(Product $product)\n    {\n        $this-&gt;products[] = $product;\n    }\n    public function getProducts()\n    {\n        return $this-&gt;products;\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-2f7e9f23c9c7b6ba622825739f80a6f3e85e6370-18\">18<\/a><\/td><\/tr><\/table><p>We are now finished with the domain model given the requirements.\nLets add metadata mappings for the <code>Bug<\/code> entity, as we did for\nthe <code>Product<\/code> before:<\/p><div class=\"configuration-block jsactive clearfix\"><ul class=\"simple\"><li><em>PHP<\/em><div class=\"highlight-php\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"41\">&lt;?php\n\/\/ src\/Bug.php\n\/**\n * @Entity @Table(name=\"bugs\")\n **\/\nclass Bug\n{\n    \/**\n     * @Id @Column(type=\"integer\") @GeneratedValue\n     **\/\n    protected $id;\n    \/**\n     * @Column(type=\"string\")\n     **\/\n    protected $description;\n    \/**\n     * @Column(type=\"datetime\")\n     **\/\n    protected $created;\n    \/**\n     * @Column(type=\"string\")\n     **\/\n    protected $status;\n    \/**\n     * @ManyToOne(targetEntity=\"User\", inversedBy=\"assignedBugs\")\n     **\/\n    protected $engineer;\n    \/**\n     * @ManyToOne(targetEntity=\"User\", inversedBy=\"reportedBugs\")\n     **\/\n    protected $reporter;\n    \/**\n     * @ManyToMany(targetEntity=\"Product\")\n     **\/\n    protected $products;\n    \/\/ ... (other code)\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-31\">31<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-32\">32<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-33\">33<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-34\">34<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-35\">35<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-36\">36<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-37\">37<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-38\">38<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-39\">39<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-40\">40<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5cdd364ab67e444420c4f69c057d9f025e9f918b-41\">41<\/a><\/td><\/tr><\/table><\/div><\/li><li><em>XML<\/em><div class=\"highlight-xml\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"21\">&lt;!-- config\/xml\/Bug.dcm.xml --&gt;\n&lt;doctrine-mapping xmlns=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\"\n      xmlns:xsi=\"http:\/\/www.w3.org\/2001\/XMLSchema-instance\"\n      xsi:schemaLocation=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\n                          https:\/\/www.doctrine-project.org\/schemas\/orm\/doctrine-mapping.xsd\"&gt;\n    &lt;entity name=\"Bug\" table=\"bugs\"&gt;\n        &lt;id name=\"id\" type=\"integer\"&gt;\n            &lt;generator strategy=\"AUTO\" \/&gt;\n        &lt;\/id&gt;\n        &lt;field name=\"description\" type=\"text\" \/&gt;\n        &lt;field name=\"created\" type=\"datetime\" \/&gt;\n        &lt;field name=\"status\" type=\"string\" \/&gt;\n        &lt;many-to-one target-entity=\"User\" field=\"reporter\" inversed-by=\"reportedBugs\" \/&gt;\n        &lt;many-to-one target-entity=\"User\" field=\"engineer\" inversed-by=\"assignedBugs\" \/&gt;\n        &lt;many-to-many target-entity=\"Product\" field=\"products\" \/&gt;\n    &lt;\/entity&gt;\n&lt;\/doctrine-mapping&gt;<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-6e5d73e0484a4af5b7e8b1209a2be608956872a6-21\">21<\/a><\/td><\/tr><\/table><\/div><\/li><\/ul><\/div><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>The YAML driver is deprecated and will be removed in version 3.0.\nIt is strongly recommended to switch to one of the other mappings.<\/p>\n<table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"26\"># config\/yaml\/Bug.dcm.yml\nBug:\n  type: entity\n  table: bugs\n  id:\n    id:\n      type: integer\n      generator:\n        strategy: AUTO\n  fields:\n    description:\n      type: text\n    created:\n      type: datetime\n    status:\n      type: string\n  manyToOne:\n    reporter:\n      targetEntity: User\n      inversedBy: reportedBugs\n    engineer:\n      targetEntity: User\n      inversedBy: assignedBugs\n  manyToMany:\n    products:\n      targetEntity: Product<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d60746364acafea7eaed8d360dad46ff60333648-26\">26<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><\/div><p>Here we have the entity, id and primitive type definitions.\nFor the \"created\" field we have used the <code>datetime<\/code> type,<br \/>which translates the YYYY-mm-dd HH:mm:ss database format<br \/>into a PHP DateTime instance and back.<\/p><p>After the field definitions, the two qualified references to the\nuser entity are defined. They are created by the <code>many-to-one<\/code>\ntag. The class name of the related entity has to be specified with\nthe <code>target-entity<\/code> attribute, which is enough information for\nthe database mapper to access the foreign-table. Since\n<code>reporter<\/code> and <code>engineer<\/code> are on the owning side of a\nbi-directional relation, we also have to specify the <code>inversed-by<\/code>\nattribute. They have to point to the field names on the inverse\nside of the relationship. We will see in the next example that the <code>inversed-by<\/code>\nattribute has a counterpart <code>mapped-by<\/code> which makes that\nthe inverse side.<\/p><p>The last definition is for the <code>Bug#products<\/code> collection. It\nholds all products where the specific bug occurs. Again\nyou have to define the <code>target-entity<\/code> and <code>field<\/code> attributes\non the <code>many-to-many<\/code> tag.<\/p><p>Finally, we'll add metadata mappings for the <code>User<\/code> entity.<\/p><div class=\"configuration-block jsactive clearfix\"><ul class=\"simple\"><li><em>PHP<\/em><div class=\"highlight-php\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"33\">&lt;?php\n\/\/ src\/User.php\n\/**\n * @Entity @Table(name=\"users\")\n **\/\nclass User\n{\n    \/**\n     * @Id @GeneratedValue @Column(type=\"integer\")\n     * @var int\n     **\/\n    protected $id;\n    \/**\n     * @Column(type=\"string\")\n     * @var string\n     **\/\n    protected $name;\n    \/**\n     * @OneToMany(targetEntity=\"Bug\", mappedBy=\"reporter\")\n     * @var Bug[] An ArrayCollection of Bug objects.\n     **\/\n    protected $reportedBugs = null;\n    \/**\n     * @OneToMany(targetEntity=\"Bug\", mappedBy=\"engineer\")\n     * @var Bug[] An ArrayCollection of Bug objects.\n     **\/\n    protected $assignedBugs = null;\n    \/\/ .. (other code)\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-31\">31<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-32\">32<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-d746d1b9fd1b59dcdd0ebefa539b0d5f6c7fa826-33\">33<\/a><\/td><\/tr><\/table><\/div><\/li><li><em>XML<\/em><div class=\"highlight-xml\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"17\">&lt;!-- config\/xml\/User.dcm.xml --&gt;\n&lt;doctrine-mapping xmlns=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\"\n      xmlns:xsi=\"http:\/\/www.w3.org\/2001\/XMLSchema-instance\"\n      xsi:schemaLocation=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\n                          https:\/\/www.doctrine-project.org\/schemas\/orm\/doctrine-mapping.xsd\"&gt;\n     &lt;entity name=\"User\" table=\"users\"&gt;\n         &lt;id name=\"id\" type=\"integer\"&gt;\n             &lt;generator strategy=\"AUTO\" \/&gt;\n         &lt;\/id&gt;\n         &lt;field name=\"name\" type=\"string\" \/&gt;\n         &lt;one-to-many target-entity=\"Bug\" field=\"reportedBugs\" mapped-by=\"reporter\" \/&gt;\n         &lt;one-to-many target-entity=\"Bug\" field=\"assignedBugs\" mapped-by=\"engineer\" \/&gt;\n     &lt;\/entity&gt;\n&lt;\/doctrine-mapping&gt;<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b7408f05ac4512babea6f58f459fff38146aca3c-17\">17<\/a><\/td><\/tr><\/table><\/div><\/li><\/ul><\/div><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>The YAML driver is deprecated and will be removed in version 3.0.\nIt is strongly recommended to switch to one of the other mappings.<\/p>\n<table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"19\"># config\/yaml\/User.dcm.yml\nUser:\n  type: entity\n  table: users\n  id:\n    id:\n      type: integer\n      generator:\n        strategy: AUTO\n  fields:\n    name:\n      type: string\n  oneToMany:\n    reportedBugs:\n      targetEntity: Bug\n      mappedBy: reporter\n    assignedBugs:\n      targetEntity: Bug\n      mappedBy: engineer<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-e58b4f6f8d591444689aa6f25b092a740ab2e5b2-19\">19<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><\/div><p>Here are some new things to mention about the <code>one-to-many<\/code> tags.\nRemember that we discussed about the inverse and owning side. Now\nboth reportedBugs and assignedBugs are inverse relations, which\nmeans the join details have already been defined on the owning\nside. Therefore we only have to specify the property on the Bug\nclass that holds the owning sides.<\/p><p>Update your database schema by running:<\/p><div class=\"console\"><pre class=\"console\">$ vendor\/bin\/doctrine orm:schema-tool:update --force<\/pre><\/div><p>So far, we've seen the most basic features of the metadata definition language.\nTo explore additional functionality, let's first create new <code>User<\/code> entities:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"13\">&lt;?php\n\/\/ create_user.php\nrequire_once \"bootstrap.php\";\n$newUsername = $argv[1];\n$user = new User();\n$user-&gt;setName($newUsername);\n$entityManager-&gt;persist($user);\n$entityManager-&gt;flush();\necho \"Created User with ID \" . $user-&gt;getId() . \"\\n\";<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5e54bc3bd9fe68fdc8c85ac7039f2f2c115d605d-13\">13<\/a><\/td><\/tr><\/table><p>Now call:<\/p><div class=\"console\"><pre class=\"console\">$ php create_user.php beberlei<\/pre><\/div><p>We now have the necessary data to create a new Bug entity:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"32\">&lt;?php\n\/\/ create_bug.php &lt;reporter-id&gt; &lt;engineer-id&gt; &lt;product-ids&gt;\nrequire_once \"bootstrap.php\";\n$reporterId = $argv[1];\n$engineerId = $argv[2];\n$productIds = explode(\",\", $argv[3]);\n$reporter = $entityManager-&gt;find(\"User\", $reporterId);\n$engineer = $entityManager-&gt;find(\"User\", $engineerId);\nif (!$reporter || !$engineer) {\n    echo \"No reporter and\/or engineer found for the given id(s).\\n\";\n    exit(1);\n}\n$bug = new Bug();\n$bug-&gt;setDescription(\"Something does not work!\");\n$bug-&gt;setCreated(new DateTime(\"now\"));\n$bug-&gt;setStatus(\"OPEN\");\nforeach ($productIds as $productId) {\n    $product = $entityManager-&gt;find(\"Product\", $productId);\n    $bug-&gt;assignToProduct($product);\n}\n$bug-&gt;setReporter($reporter);\n$bug-&gt;setEngineer($engineer);\n$entityManager-&gt;persist($bug);\n$entityManager-&gt;flush();\necho \"Your new Bug Id: \".$bug-&gt;getId().\"\\n\";<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-31\">31<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-dd1ef5519515299a9a86d9d9aad723dcf20042da-32\">32<\/a><\/td><\/tr><\/table><p>Since we only have one user and product, probably with the ID of 1, we can\ncall this script as follows:<\/p><div class=\"console\"><pre class=\"console\">$ php create_bug.php 1 1 1\n<\/pre><\/div><p>See how simple it is to relate a Bug, Reporter, Engineer and Products?\nAlso recall that thanks to the UnitOfWork pattern, Doctrine will detect\nthese relations and update all of the modified entities in the database\nautomatically when <code>flush()<\/code> is called.<\/p><p>Using the previous examples we can fill up the database quite a\nbit. However, we now need to discuss how to query the underlying\nmapper for the required view representations. When opening the\napplication, bugs can be paginated through a list-view, which is\nthe first read-only use-case:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"19\">&lt;?php\n\/\/ list_bugs.php\nrequire_once \"bootstrap.php\";\n$dql = \"SELECT b, e, r FROM Bug b JOIN b.engineer e JOIN b.reporter r ORDER BY b.created DESC\";\n$query = $entityManager-&gt;createQuery($dql);\n$query-&gt;setMaxResults(30);\n$bugs = $query-&gt;getResult();\nforeach ($bugs as $bug) {\n    echo $bug-&gt;getDescription().\" - \".$bug-&gt;getCreated()-&gt;format('d.m.Y').\"\\n\";\n    echo \"    Reported by: \".$bug-&gt;getReporter()-&gt;getName().\"\\n\";\n    echo \"    Assigned to: \".$bug-&gt;getEngineer()-&gt;getName().\"\\n\";\n    foreach ($bug-&gt;getProducts() as $product) {\n        echo \"    Platform: \".$product-&gt;getName().\"\\n\";\n    }\n    echo \"\\n\";\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-02e6ba3eda645f6b5eef96fff3d7dffaccb6d58e-19\">19<\/a><\/td><\/tr><\/table><p>The DQL Query in this example fetches the 30 most recent bugs with\ntheir respective engineer and reporter in one single SQL statement.\nThe console output of this script is then:<\/p><div class=\"console\"><pre class=\"console\">Something does not work! - 02.04.2010\n    Reported by: beberlei\n    Assigned to: beberlei\n    Platform: My Product<\/pre><\/div><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p><strong>DQL is not SQL<\/strong><\/p>\n<p>You may wonder why we start writing SQL at the beginning of this\nuse-case. Don't we use an ORM to get rid of all the endless\nhand-writing of SQL? Doctrine introduces DQL which is best\ndescribed as <strong>object-query-language<\/strong> and is a dialect of\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Object_Query_Language\">OQL<\/a> and\nsimilar to <a href=\"http:\/\/www.hibernate.org\">HQL<\/a> or\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Java_Persistence_Query_Language\">JPQL<\/a>.\nIt does not know the concept of columns and tables, but only those\nof Entity-Class and property. Using the Metadata we defined before\nit allows for very short distinctive and powerful queries.<\/p>\n<p>An important reason why DQL is favourable to the Query API of most\nORMs is its similarity to SQL. The DQL language allows query\nconstructs that most ORMs don't: GROUP BY even with HAVING,\nSub-selects, Fetch-Joins of nested classes, mixed results with\nentities and scalar data such as COUNT() results and much more.\nUsing DQL you should seldom come to the point where you want to\nthrow your ORM into the dumpster, because it doesn't support some\nthe more powerful SQL concepts.<\/p>\n<p>Instead of handwriting DQL you can use the <code>QueryBuilder<\/code> retrieved\nby calling <code>$entityManager-&gt;createQueryBuilder()<\/code>. There are more\ndetails about this in the relevant part of the documentation.<\/p>\n<p>As a last resort you can still use Native SQL and a description of the\nresult set to retrieve entities from the database. DQL boils down to a<br \/>Native SQL statement and a <code>ResultSetMapping<\/code> instance itself. Using<br \/>Native SQL you could even use stored procedures for data retrieval, or<br \/>make use of advanced non-portable database queries like PostgreSql's<br \/>recursive queries.<\/p>\n<\/td><\/tr><\/table><\/div><p>In the previous use-case we retrieved the results as their\nrespective object instances. We are not limited to retrieving\nobjects only from Doctrine however. For a simple list view like the\nprevious one we only need read access to our entities and can\nswitch the hydration from objects to simple PHP arrays instead.<\/p><p>Hydration can be an expensive process so only retrieving what you need can<br \/>yield considerable performance benefits for read-only requests.<\/p><p>Implementing the same list view as before using array hydration we\ncan rewrite our code:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"18\">&lt;?php\n\/\/ list_bugs_array.php\nrequire_once \"bootstrap.php\";\n$dql = \"SELECT b, e, r, p FROM Bug b JOIN b.engineer e \".\n       \"JOIN b.reporter r JOIN b.products p ORDER BY b.created DESC\";\n$query = $entityManager-&gt;createQuery($dql);\n$bugs = $query-&gt;getArrayResult();\nforeach ($bugs as $bug) {\n    echo $bug['description'] . \" - \" . $bug['created']-&gt;format('d.m.Y').\"\\n\";\n    echo \"    Reported by: \".$bug['reporter']['name'].\"\\n\";\n    echo \"    Assigned to: \".$bug['engineer']['name'].\"\\n\";\n    foreach ($bug['products'] as $product) {\n        echo \"    Platform: \".$product['name'].\"\\n\";\n    }\n    echo \"\\n\";\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-f85425157a6879dd8e6c853198c4bae688679d36-18\">18<\/a><\/td><\/tr><\/table><p>There is one significant difference in the DQL query however, we\nhave to add an additional fetch-join for the products connected to\na bug. The resulting SQL query for this single select statement is\npretty large, however still more efficient to retrieve compared to\nhydrating objects.<\/p><p>The next Use-Case is displaying a Bug by primary key. This could be\ndone using DQL as in the previous example with a where clause,\nhowever there is a convenience method on the <code>EntityManager<\/code> that\nhandles loading by primary key, which we have already seen in the\nwrite scenarios:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"10\">&lt;?php\n\/\/ show_bug.php &lt;id&gt;\nrequire_once \"bootstrap.php\";\n$theBugId = $argv[1];\n$bug = $entityManager-&gt;find(\"Bug\", (int)$theBugId);\necho \"Bug: \".$bug-&gt;getDescription().\"\\n\";\necho \"Engineer: \".$bug-&gt;getEngineer()-&gt;getName().\"\\n\";<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-83ea7f659dcd7351105873491d448c95ea537deb-10\">10<\/a><\/td><\/tr><\/table><p>The output of the engineer\u2019s name is fetched from the database! What is happening?<\/p><p>Since we only retrieved the bug by primary key both the engineer and reporter\nare not immediately loaded from the database but are replaced by LazyLoading\nproxies. These proxies will load behind the scenes, when the first method\nis called on them.<\/p><p>Sample code of this proxy generated code can be found in the specified Proxy\nDirectory, it looks like:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"22\">&lt;?php\nnamespace MyProject\\Proxies;\n\/**\n * THIS CLASS WAS GENERATED BY THE DOCTRINE ORM. DO NOT EDIT THIS FILE.\n **\/\nclass UserProxy extends \\User implements \\Doctrine\\ORM\\Proxy\\Proxy\n{\n    \/\/ .. lazy load code here\n    public function addReportedBug($bug)\n    {\n        $this-&gt;_load();\n        return parent::addReportedBug($bug);\n    }\n    public function assignedToBug($bug)\n    {\n        $this-&gt;_load();\n        return parent::assignedToBug($bug);\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-0d0452f39e38ffe8cbc539cd967d95cf888c9f77-22\">22<\/a><\/td><\/tr><\/table><p>See how upon each method call the proxy is lazily loaded from the\ndatabase?<\/p><p>The call prints:<\/p><div class=\"console\"><pre class=\"console\">$ php show_bug.php 1\nBug: Something does not work!\nEngineer: beberlei<\/pre><\/div><div class=\"alert warning-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-exclamation-circle text-warning mr-2\"><\/i><\/td><td><p>Lazy loading additional data can be very convenient but the additional\nqueries create an overhead. If you know that certain fields will always\n(or usually) be required by the query then you will get better performance\nby explicitly retrieving them all in the first query.<\/p>\n<\/td><\/tr><\/table><\/div><p>For the next use-case we want to retrieve the dashboard view, a\nlist of all open bugs the user reported or was assigned to. This\nwill be achieved using DQL again, this time with some WHERE clauses\nand usage of bound parameters:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"19\">&lt;?php\n\/\/ dashboard.php &lt;user-id&gt;\nrequire_once \"bootstrap.php\";\n$theUserId = $argv[1];\n$dql = \"SELECT b, e, r FROM Bug b JOIN b.engineer e JOIN b.reporter r \".\n       \"WHERE b.status = 'OPEN' AND (e.id = ?1 OR r.id = ?1) ORDER BY b.created DESC\";\n$myBugs = $entityManager-&gt;createQuery($dql)\n                        -&gt;setParameter(1, $theUserId)\n                        -&gt;setMaxResults(15)\n                        -&gt;getResult();\necho \"You have created or assigned to \" . count($myBugs) . \" open bugs:\\n\\n\";\nforeach ($myBugs as $bug) {\n    echo $bug-&gt;getId() . \" - \" . $bug-&gt;getDescription().\"\\n\";\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-a51ddaa1379c4b2487fb7cbb47b0b187794bda33-19\">19<\/a><\/td><\/tr><\/table><p>Until now we only retrieved entities or their array representation.\nDoctrine also supports the retrieval of non-entities through DQL.\nThese values are called \"scalar result values\" and may even be\naggregate values using COUNT, SUM, MIN, MAX or AVG functions.<\/p><p>We will need this knowledge to retrieve the number of open bugs\ngrouped by product:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"11\">&lt;?php\n\/\/ products.php\nrequire_once \"bootstrap.php\";\n$dql = \"SELECT p.id, p.name, count(b.id) AS openBugs FROM Bug b \".\n       \"JOIN b.products p WHERE b.status = 'OPEN' GROUP BY p.id\";\n$productBugs = $entityManager-&gt;createQuery($dql)-&gt;getScalarResult();\nforeach ($productBugs as $productBug) {\n    echo $productBug['name'].\" has \" . $productBug['openBugs'] . \" open bugs!\\n\";\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-43e9d5730ad27defc43afed0262c14061691c5e8-11\">11<\/a><\/td><\/tr><\/table><p>There is a single use-case missing from the requirements, Engineers\nshould be able to close a bug. This looks like:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"10\">&lt;?php\n\/\/ src\/Bug.php\nclass Bug\n{\n    public function close()\n    {\n        $this-&gt;status = \"CLOSE\";\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-9811cc35891950a59934044714d8aeb2f358811a-10\">10<\/a><\/td><\/tr><\/table><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"10\">&lt;?php\n\/\/ close_bug.php &lt;bug-id&gt;\nrequire_once \"bootstrap.php\";\n$theBugId = $argv[1];\n$bug = $entityManager-&gt;find(\"Bug\", (int)$theBugId);\n$bug-&gt;close();\n$entityManager-&gt;flush();<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-adc4d9226edfbc1728da4e78af2271404c51a7bd-10\">10<\/a><\/td><\/tr><\/table><p>When retrieving the Bug from the database it is inserted into the\nIdentityMap inside the UnitOfWork of Doctrine. This means your Bug\nwith exactly this id can only exist once during the whole request\nno matter how often you call <code>EntityManager#find()<\/code>. It even\ndetects entities that are hydrated using DQL and are already\npresent in the Identity Map.<\/p><p>When flush is called the EntityManager loops over all the entities\nin the identity map and performs a comparison between the values\noriginally retrieved from the database and those values the entity\ncurrently has. If at least one of these properties is different the\nentity is scheduled for an UPDATE against the database. Only the\nchanged columns are updated, which offers a pretty good performance\nimprovement compared to updating all the properties.<\/p><p>For now we have not discussed how to separate the Doctrine query logic from your model.\nIn Doctrine 1 there was the concept of <code>Doctrine_Table<\/code> instances for this\nseparation. The similar concept in Doctrine2 is called Entity Repositories, integrating\nthe <a href=\"http:\/\/martinfowler.com\/eaaCatalog\/repository.html\">repository pattern<\/a> at the heart of Doctrine.<\/p><p>Every Entity uses a default repository by default and offers a bunch of convenience\nmethods that you can use to query for instances of that Entity. Take for example\nour Product entity. If we wanted to Query by name, we can use:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-5a6c5fa900939da9ebe82862490e1e06beec1b9b-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"3\">&lt;?php\n$product = $entityManager-&gt;getRepository('Product')\n                         -&gt;findOneBy(array('name' =&gt; $productName));<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5a6c5fa900939da9ebe82862490e1e06beec1b9b-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-5a6c5fa900939da9ebe82862490e1e06beec1b9b-3\">3<\/a><\/td><\/tr><\/table><p>The method <code>findOneBy()<\/code> takes an array of fields or association keys and the values to match against.<\/p><p>If you want to find all entities matching a condition you can use <code>findBy()<\/code>, for\nexample querying for all closed bugs:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"7\">&lt;?php\n$bugs = $entityManager-&gt;getRepository('Bug')\n                      -&gt;findBy(array('status' =&gt; 'CLOSED'));\nforeach ($bugs as $bug) {\n    \/\/ do stuff\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4ec3c33def76d76750c7f747e7867733459ac615-7\">7<\/a><\/td><\/tr><\/table><p>Compared to DQL these query methods are falling short of functionality very fast.\nDoctrine offers you a convenient way to extend the functionalities of the default <code>EntityRepository<\/code>\nand put all the specialized DQL query logic on it. For this you have to create a subclass\nof <code>Doctrine\\ORM\\EntityRepository<\/code>, in our case a <code>BugRepository<\/code> and group all\nthe previously discussed query functionality in it:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"43\">&lt;?php\n\/\/ src\/BugRepository.php\nuse Doctrine\\ORM\\EntityRepository;\nclass BugRepository extends EntityRepository\n{\n    public function getRecentBugs($number = 30)\n    {\n        $dql = \"SELECT b, e, r FROM Bug b JOIN b.engineer e JOIN b.reporter r ORDER BY b.created DESC\";\n        $query = $this-&gt;getEntityManager()-&gt;createQuery($dql);\n        $query-&gt;setMaxResults($number);\n        return $query-&gt;getResult();\n    }\n    public function getRecentBugsArray($number = 30)\n    {\n        $dql = \"SELECT b, e, r, p FROM Bug b JOIN b.engineer e \".\n               \"JOIN b.reporter r JOIN b.products p ORDER BY b.created DESC\";\n        $query = $this-&gt;getEntityManager()-&gt;createQuery($dql);\n        $query-&gt;setMaxResults($number);\n        return $query-&gt;getArrayResult();\n    }\n    public function getUsersBugs($userId, $number = 15)\n    {\n        $dql = \"SELECT b, e, r FROM Bug b JOIN b.engineer e JOIN b.reporter r \".\n               \"WHERE b.status = 'OPEN' AND e.id = ?1 OR r.id = ?1 ORDER BY b.created DESC\";\n        return $this-&gt;getEntityManager()-&gt;createQuery($dql)\n                             -&gt;setParameter(1, $userId)\n                             -&gt;setMaxResults($number)\n                             -&gt;getResult();\n    }\n    public function getOpenBugsByProduct()\n    {\n        $dql = \"SELECT p.id, p.name, count(b.id) AS openBugs FROM Bug b \".\n               \"JOIN b.products p WHERE b.status = 'OPEN' GROUP BY p.id\";\n        return $this-&gt;getEntityManager()-&gt;createQuery($dql)-&gt;getScalarResult();\n    }\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-15\">15<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-16\">16<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-17\">17<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-18\">18<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-19\">19<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-20\">20<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-21\">21<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-22\">22<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-23\">23<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-24\">24<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-25\">25<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-26\">26<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-27\">27<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-28\">28<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-29\">29<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-30\">30<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-31\">31<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-32\">32<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-33\">33<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-34\">34<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-35\">35<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-36\">36<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-37\">37<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-38\">38<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-39\">39<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-40\">40<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-41\">41<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-42\">42<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-b320b9eeb21c47d9842d00e0320b0735b5bd09fe-43\">43<\/a><\/td><\/tr><\/table><p>To be able to use this query logic through <code>$this-&gt;getEntityManager()-&gt;getRepository('Bug')<\/code>\nwe have to adjust the metadata slightly.<\/p><div class=\"configuration-block jsactive clearfix\"><ul class=\"simple\"><li><em>PHP<\/em><div class=\"highlight-php\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"9\">&lt;?php\n\/**\n * @Entity(repositoryClass=\"BugRepository\")\n * @Table(name=\"bugs\")\n **\/\nclass Bug\n{\n    \/\/...\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-87095733373b696a2e0c5b4ae79492a407086d6f-9\">9<\/a><\/td><\/tr><\/table><\/div><\/li><li><em>XML<\/em><div class=\"highlight-xml\"><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"9\">&lt;doctrine-mapping xmlns=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\"\n      xmlns:xsi=\"http:\/\/www.w3.org\/2001\/XMLSchema-instance\"\n      xsi:schemaLocation=\"http:\/\/doctrine-project.org\/schemas\/orm\/doctrine-mapping\n                          https:\/\/www.doctrine-project.org\/schemas\/orm\/doctrine-mapping.xsd\"&gt;\n      &lt;entity name=\"Bug\" table=\"bugs\" repository-class=\"BugRepository\"&gt;\n      &lt;\/entity&gt;\n&lt;\/doctrine-mapping&gt;<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-c50bf6d1d0082515a7a103306f4b3145a7d36e74-9\">9<\/a><\/td><\/tr><\/table><\/div><\/li><\/ul><\/div><div class=\"alert note-admonition bg-light-yellow text-dark border\"><table style=\"width: 100%;\"><tr><td class=\"align-top\" style=\"width: 10px;\"><i class=\"fas fa-sticky-note text-primary mr-2\"><\/i><\/td><td><p>The YAML driver is deprecated and will be removed in version 3.0.\nIt is strongly recommended to switch to one of the other mappings.<\/p>\n<table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-4eacca44ca4ee091e2d133f388bf22a3de7ebd06-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"3\">Bug:\n  type: entity\n  repositoryClass: BugRepository<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4eacca44ca4ee091e2d133f388bf22a3de7ebd06-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4eacca44ca4ee091e2d133f388bf22a3de7ebd06-3\">3<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><\/div><p>Now we can remove our query logic in all the places and instead use them through the EntityRepository.\nAs an example here is the code of the first use case \"List of Bugs\":<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"15\">&lt;?php\n\/\/ list_bugs_repository.php\nrequire_once \"bootstrap.php\";\n$bugs = $entityManager-&gt;getRepository('Bug')-&gt;getRecentBugs();\nforeach ($bugs as $bug) {\n    echo $bug-&gt;getDescription().\" - \".$bug-&gt;getCreated()-&gt;format('d.m.Y').\"\\n\";\n    echo \"    Reported by: \".$bug-&gt;getReporter()-&gt;getName().\"\\n\";\n    echo \"    Assigned to: \".$bug-&gt;getEngineer()-&gt;getName().\"\\n\";\n    foreach ($bug-&gt;getProducts() as $product) {\n        echo \"    Platform: \".$product-&gt;getName().\"\\n\";\n    }\n    echo \"\\n\";\n}<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-3\">3<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-4\">4<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-5\">5<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-6\">6<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-7\">7<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-8\">8<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-9\">9<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-10\">10<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-11\">11<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-12\">12<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-13\">13<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-14\">14<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-4280a6f673e0d5a77f877a9ba349252402a837bd-15\">15<\/a><\/td><\/tr><\/table><p>Using EntityRepositories you can avoid coupling your model with specific query logic.\nYou can also re-use query logic easily throughout your application.<\/p><p>The method <code>count()<\/code> takes an array of fields or association keys and the values to match against.\nThis provides you with a convenient and lightweight way to count a resultset when you don't need to\ndeal with it:<\/p><table class=\"code-block-table\"><tr><td class=\"line-number noselect\"><a href=\"#line-number-09e044fe4ecfe597e457db377a413d517d46f67e-1\">1<\/a><\/td><td class=\"code-line\" rowspan=\"3\">&lt;?php\n$productCount = $entityManager-&gt;getRepository(Product::class)\n                         -&gt;count(['name' =&gt; $productName]);<\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-09e044fe4ecfe597e457db377a413d517d46f67e-2\">2<\/a><\/td><\/tr><tr><td class=\"line-number noselect\"><a href=\"#line-number-09e044fe4ecfe597e457db377a413d517d46f67e-3\">3<\/a><\/td><\/tr><\/table><p>This tutorial is over here, I hope you had fun. Additional content\nwill be added to this tutorial incrementally, topics will include:<\/p><ul><li class=\"dash\">  More on Association Mappings<\/li>\n<li class=\"dash\">  Lifecycle Events triggered in the UnitOfWork<\/li>\n<li class=\"dash\">  Ordering of Collections<\/li>\n<\/ul><p>Additional details on all the topics discussed here can be found in\nthe respective manual chapters.<\/p>","created_at":"2018-11-14T13:13:59+0000","updated_at":"2018-11-14T13:14:05+0000","annotations":[],"mimetype":"text\/html","reading_time":32,"domain_name":"www.doctrine-project.org","preview_picture":"https:\/\/www.doctrine-project.org\/images\/og.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["procgen","gamedev"],"is_public":false,"id":686,"title":"procedural-generation","url":"https:\/\/www.grendel-games.com\/procedural-generation\/?lang=en","created_at":"2018-11-13T20:57:59+0000","updated_at":"2018-11-13T20:58:04+0000","published_by":[""],"annotations":[],"reading_time":0,"domain_name":"www.grendel-games.com"},{"is_archived":0,"is_starred":0,"tags":["procgen","unity","indiegames"],"is_public":false,"id":685,"title":"Infinite procedurally-generated city with the Wave Function Collapse algorithm","url":"https:\/\/news.ycombinator.com\/item?id=18443311","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D18443311\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"18443311\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/marian42.itch.io\/wfc\" class=\"storylink\">Infinite procedurally-generated city with the Wave Function Collapse algorithm<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=itch.io\">itch.io<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        249 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=guessmyname\" class=\"hnuser\">guessmyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18443311\">2 hours ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=18443311&amp;goto=item%3Fid%3D18443311\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Infinite%20procedurally-generated%20city%20with%20the%20Wave%20Function%20Collapse%20algorithm&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Infinite%20procedurally-generated%20city%20with%20the%20Wave%20Function%20Collapse%20algorithm\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=18443311&amp;auth=110293ded14cae3712214c3686b1fccf4365f260\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=18443311\">23\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>\n          \n      <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-13T20:51:56+0000","updated_at":"2018-11-13T20:52:10+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["emacs","spotify"],"is_public":false,"id":684,"title":"chiply\/spot4e","url":"https:\/\/github.com\/chiply\/spot4e","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>An emacs helm interface for searching Spotify and controlling\n  playback (on premium).<\/p>\n<p>Of note, spot4e controls Spotify via the Spotify API, not\n  dbus.  This means you can control Spotify playback on whichever\n  device Spotify is currently active.  For example, you can control\n  Spotify playback on your iPhone or Google Home if you wish.  You do\n  not have to have Spotify open on the computer running emacs.<\/p>\n<p>This feature also means that spot4e should work as-is on most\n  operating systems.<\/p>\n<h2><a id=\"user-content-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>Dependencies<\/h2>\n<p>spot4e depends upon lisp packages json, url, and helm to function properly.<\/p>\n<p>Also, this app is NOT a backdoor to Spotify.  You need to have a\n  premium account to use playback features!<\/p>\n<h2><a id=\"user-content-installation-instructions\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation-instructions\"><\/a>Installation instructions<\/h2>\n<p>This package is not yet on MELPA.<\/p>\n<p>To install, simply download the spot4e.el file and install however you\n  like.<\/p>\n<p>For example, clone the project to whichever location you prefer, and\n  add the directory to your load path.  I recommend placing the\n  following to your init file:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'load-path \"\/path\/to\/spot4e\")<\/pre><\/div>\n<h2><a id=\"user-content-initial-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#initial-setup\"><\/a>Initial setup<\/h2>\n<p>Because spot4e controls a given user\u2019s Spotify playback, it requires\n  each user allow spot4e to access user data.  Each user must authorize\n  spot4e through Spotify\u2019s <a href=\"https:\/\/developer.spotify.com\/web-api\/authorization-guide\/#authorization_code_flow\" rel=\"nofollow\">Authorization Code Flow<\/a>, so there is some\n  setup required:<\/p>\n<h3><a id=\"user-content-step-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-1\"><\/a>Step 1:<\/h3>\n<p>To credential the app, run the interactive function spot4e-authorize:<\/p>\n<p>M-x spot4e-authorize<\/p>\n<p>This will open a login page in your browser.<\/p>\n<h3><a id=\"user-content-step-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-2\"><\/a>Step 2<\/h3>\n<p>Log in (or click ok if you are already logged in).  You will be\n  redirected to Spotify\u2019s home page, where the URL contains your\n  authorization code:<\/p>\n<p><a href=\"https:\/\/www.spotify.com\/us\/?code=AUTH_CODE_HERE\" rel=\"nofollow\">https:\/\/www.spotify.com\/us\/?code=AUTH_CODE_HERE<\/a><\/p>\n<h3><a id=\"user-content-step-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-3\"><\/a>Step 3<\/h3>\n<p>Copy the this code and paste into the emacs minibuffer prompt \u201cEnter code\n  from URL\u201d and press enter<\/p>\n<h3><a id=\"user-content-step-4\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-4\"><\/a>Step 4<\/h3>\n<p>This authorization process has given you a refresh token, stored in\n  spot4e-refresh-token, which you will want to save for future sessions.<\/p>\n<p>I recommend storing this token in your emacs config file:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq spot4e-refresh-token \"REFRESH_TOKEN_HERE\")<\/pre><\/div>\n<p>Note: you can get your refresh token by simply typing\n  spot4e-refresh-token into any buffer, placing your cursor immediately\n  after the variable, and pressing C-u C-x C-e.  This evaluates the\n  variable and prints the output at point.<\/p>\n<h3><a id=\"user-content-step-5\" class=\"anchor\" aria-hidden=\"true\" href=\"#step-5\"><\/a>Step 5<\/h3>\n<p>This refresh token does not expire and will be good for future\n  sessions.  The access-token, however, does expire every hour, so I\n  recommend refreshing it periodically with the spot4e-refresh function.\n  I recommend adding the following to your config file as well:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(run-with-timer 0 (* 60 59) 'spot4e-refresh)<\/pre><\/div>\n<p>To reiterate from the summary, spot4e controls user\u2019s Spotify playback\n  through the Spotify API.  You must have an active Spotify session for\n  spot4e to trigger playback on the respective device.<\/p>\n<h2><a id=\"user-content-playback-control\" class=\"anchor\" aria-hidden=\"true\" href=\"#playback-control\"><\/a>Playback control<\/h2>\n<p>You can control Spotify playback with the following.<\/p>\n<table><tbody><tr><td>spot4e-message-currently-playing<\/td><td>message current track in minibuffer<\/td><\/tr><tr><td>spot4e-player-play<\/td><td>send play to Spotify playback<\/td><\/tr><tr><td>spot4e-player-pause<\/td><td>send pause to Spotify playback<\/td><\/tr><tr><td>spot4e-player-next<\/td><td>send next to Spotify playback<\/td><\/tr><tr><td>spot4e-player-previous<\/td><td>send previous to Spotify playback<\/td><\/tr><\/tbody><\/table><h2><a id=\"user-content-search\" class=\"anchor\" aria-hidden=\"true\" href=\"#search\"><\/a>Search<\/h2>\n<h3><a id=\"user-content-search-track-by-album--artist--or-track-name\" class=\"anchor\" aria-hidden=\"true\" href=\"#search-track-by-album--artist--or-track-name\"><\/a>Search track by album-, artist-, or track name<\/h3>\n<p>To start searching for tracks:<\/p>\n<p>M-x spot4e-search-tracks<\/p>\n<p>and begin searching for tracks.<\/p>\n<p>Of note, this interfaces with the tracks search endpoint which allows\n  you to search for tracks by artist, album, and track name.<\/p>\n<p>Once the items appear, you can go to next an previous with either the\n  C-n and C-p commands, respectively or the &lt;down&gt; and &lt;up&gt; keys,\n  respectively.<\/p>\n<p>To play a track, simply press Enter.  By default, the track plays in\n  the context of the album in which it appears, otherwise playback stops\n  after the song is finished.<\/p>\n<h3><a id=\"user-content-search-for-album-by-album--or-artist--name\" class=\"anchor\" aria-hidden=\"true\" href=\"#search-for-album-by-album--or-artist--name\"><\/a>Search for album by album- or artist- name<\/h3>\n<p>To start searching for albums:<\/p>\n<p>M-x spot4e-search-albums<\/p>\n<p>and begin searching for albums<\/p>\n<p>Navigate through tracks as usual and press enter to see album tracks.\n  Press enter on an album track to play the track.  You can navigate\n  back to search-albums with &lt;f2&gt;.<\/p>\n<h3><a id=\"user-content-search-for-artist-by-artist--name\" class=\"anchor\" aria-hidden=\"true\" href=\"#search-for-artist-by-artist--name\"><\/a>Search for artist by artist- name<\/h3>\n<p>To start searching for artists:<\/p>\n<p>M-x spot4e-search-artists<\/p>\n<p>and begin searching for artists<\/p>\n<p>Navigate through artists as usual and press enter to see artist albums.\n  Press enter on an artist album to see album tracks.  You can navigate\n  back to search-albums with &lt;f2&gt;.<\/p>\n<p>Press enter on album-track to play the track.<\/p>\n<h2><a id=\"user-content-browse\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse\"><\/a>Browse<\/h2>\n<h3><a id=\"user-content-browse-category--playlist\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-category--playlist\"><\/a>Browse category \u2013&gt; playlist<\/h3>\n<p>To start the category browser:<\/p>\n<p>M-x spot4e-helm-search-categories<\/p>\n<p>And browse categories.  Press enter on a category to see that\n  category\u2019s playlists.  Press enter on a playlist to see a category\u2019s\n  tracks.  Press enter on a track to play the track.<\/p>\n<p>A crude model for this browse interface is:<\/p>\n<p>category&lt;\u2013&gt;playlist&lt;\u2013&gt;track<\/p>\n<p>Note you can \u2018go back\u2019, let\u2019s say from playlists to category, by\n  pressing &lt;f2&gt;.  This makes browsing more natural - similar to what\n  you might experience in the Spotify app.<\/p>\n<h3><a id=\"user-content-browse-new-releases-by-album\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-new-releases-by-album\"><\/a>Browse new releases by album<\/h3>\n<p>To start the new releases browser:<\/p>\n<p>M-x spot4e-helm-search-new-releases<\/p>\n<p>And browse new releases.  Note these are albums.  Press enter on a new\n  release to see the tracks (note most of these new releases will be\n  singles, so there may only be one or two tracks).  Press enter on a\n  track to play the track.  Go back to new releases with &lt;f2&gt;.<\/p>\n<h3><a id=\"user-content-browse-featured-playlists\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-featured-playlists\"><\/a>Browse featured-playlists<\/h3>\n<p>To start browsing featured playlists:<\/p>\n<p>M-x spot4e-helm-search-featured-playlists<\/p>\n<p>And browse featured playlists.  Press enter on a playlist to view\n  it\u2019s tracks.  Press enter on a track to play the track or &lt;f2&gt; to go\n  back to playlists.<\/p>\n<h3><a id=\"user-content-browse-user-playlists\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-user-playlists\"><\/a>Browse user-playlists<\/h3>\n<p>To start browsing user owned and followed playlists:<\/p>\n<p>M-x spot4e-helm-search-user-playlists<\/p>\n<p>And browse user followed playlists.  Press enter on a playlist to view\n  it\u2019s tracks.  Press enter on a track to play the track or &lt;f2&gt; to go\n  back to playlists.<\/p>\n<h3><a id=\"user-content-browse-user-artists\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-user-artists\"><\/a>Browse user-artists<\/h3>\n<p>To start browsing user followed artists:<\/p>\n<p>M-x spot4e-helm-search-user-artists<\/p>\n<p>Navigate through artists as usual and press enter to see artist albums.\n  Press enter on an artist album to see album tracks.  You can navigate\n  back to search-albums with &lt;f2&gt;.<\/p>\n<p>Press enter on album-track to play the track.<\/p>\n<h3><a id=\"user-content-browse-user-tracks\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-user-tracks\"><\/a>Browse user-tracks<\/h3>\n<p>To start browsing user saved tracks:<\/p>\n<p>M-x spot4e-helm-search-user-tracks<\/p>\n<p>and browse user-saved tracks.<\/p>\n<p>Press enter on a track to play the track.<\/p>\n<h3><a id=\"user-content-browse-recommended-tracks-based-on-currently-playing-track\" class=\"anchor\" aria-hidden=\"true\" href=\"#browse-recommended-tracks-based-on-currently-playing-track\"><\/a>Browse recommended tracks based on currently playing track<\/h3>\n<p>Assuming a track is playing, start the recommended track browser with:<\/p>\n<p>M-x spot4e-helm-search-recommendations-track<\/p>\n<p>And browse recommendations.  Of note, if done repeatedly, the\n  recommended track list varies slightly, as I imagine there is some\n  randomization that occurs in Spotify\u2019s recommendation engine.<\/p>\n<p>Of note you can also get recommendations from the\n  spot4e-helm-search-tracks interface by pressing &lt;f2&gt;.  This way you\n  can get recommendations based upon specific songs you have in<\/p>\n<h2><a id=\"user-content-user-data\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-data\"><\/a>User data<\/h2>\n<h3><a id=\"user-content-save-tracks\" class=\"anchor\" aria-hidden=\"true\" href=\"#save-tracks\"><\/a>Save tracks<\/h3>\n<p>To save a track:<\/p>\n<p>spot4e-save.<\/p>\n<p>This will then appear in your saved tracks on Spotify (can access\n  these with spot4e-helm-search-user-tracks).  Note that you can also\n  save tracks from the spot4e-helm-search-tracks interface with\n  spot4e-save.<\/p>\n<p>This package was primarily inspired by Kris Jenkins\u2019s <a href=\"https:\/\/github.com\/krisajenkins\/helm-spotify\/blob\/master\/README.org\">helm-spotify<\/a>,\n  with the main differences being support for authentication and control\n  of playback on third-party devices.<\/p>\n<p>Credit is also due to the developers of the Spotify API, whose\n  excellent design made the development of this app straightforward.<\/p>\n<\/article>","created_at":"2018-11-13T18:38:47+0000","updated_at":"2018-11-13T18:38:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/36180664?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":683,"title":"Coder (PHP validation tool)","url":"https:\/\/www.drupal.org\/project\/coder","content":"<p> Please click \"I am not a robot\" to continue <\/p><p> Access to this page has been denied because we believe you are using automation tools to browse the website. <\/p><p> This may happen as a result of the following: <\/p><ul><li> Javascript is disabled or blocked by an extension (ad blockers for example) <\/li> <li> Your browser does not support cookies <\/li> <\/ul><p> Please make sure that Javascript and cookies are enabled on your browser and that you are not blocking them from loading. <\/p><p> Reference ID: #be074850-e748-11e8-ba0e-bf6d4d52b75b <\/p>","created_at":"2018-11-13T13:33:44+0000","updated_at":"2018-11-13T20:15:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.drupal.org","http_status":"403"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":682,"title":"(The only proper) PDO tutorial","url":"https:\/\/phpdelusions.net\/pdo","content":"<ol><li><a href=\"#why\">Why PDO?<\/a><\/li>\n<li><a href=\"#dsn\">Connecting. DSN<\/a><\/li>\n<li><a href=\"#query\">Running queries. PDO::query()<\/a><\/li>\n<li><a href=\"#prepared\">Prepared statements. Protection from SQL injections<\/a><\/li>\n<\/ol><ul><li><a href=\"#methods\">Binding methods<\/a><\/li>\n<li><a href=\"#bindable\">Query parts you can bind<\/a><\/li>\n<\/ul><a href=\"#multiexec\">Prepared statements. Multiple execution<\/a><a href=\"#dml\">Running SELECT INSERT, UPDATE, or DELETE statements<\/a><a href=\"#foreach\">Getting data out of statement. foreach()<\/a><a href=\"#fetch\">Getting data out of statement. fetch()<\/a><ul><li><a href=\"#returntypes\">Return types.<\/a><\/li>\n<\/ul><a href=\"#fetchcolumn\">Getting data out of statement. fetchColumn()<\/a><a href=\"#fetchall\">Getting data out of statement in dozens different formats. fetchAll()<\/a><ul><li><a href=\"#array\">Getting a plain array.<\/a><\/li>\n<li><a href=\"#column\">Getting a column.<\/a><\/li>\n<li><a href=\"#pairs\">Getting key-value pairs.<\/a><\/li>\n<li><a href=\"#unique\">Getting rows indexed by unique field<\/a><\/li>\n<li><a href=\"#group\">Getting rows grouped by some field<\/a><\/li>\n<\/ul><a href=\"#errors\">Error handling. Exceptions<\/a><ul><li><a href=\"#reporting_errors\">Reporting PDO errors<\/a><\/li>\n<li><a href=\"#catch\">Catching PDO exceptions<\/a><\/li>\n<\/ul><a href=\"#count\">Getting row count with PDO<\/a><a href=\"#affected\">Affected rows and insert id<\/a><a href=\"#like\">Prepared statements and LIKE clause<\/a><a href=\"#in\">Prepared statements and IN clause<\/a><a href=\"#identifiers\">Prepared statements and table names<\/a><a href=\"#limit\">A problem with LIMIT clause<\/a><a href=\"#transactions\">Transactions<\/a><a href=\"#call\">Calling stored procedures in PDO<\/a><a href=\"#multiquery\">Running multiple queries with PDO<\/a><a href=\"#emulation\">Emulation mode. PDO::ATTR_EMULATE_PREPARES<\/a><ul><li><a href=\"#emulation_on\">When emulation mode is turned ON<\/a><\/li>\n<li><a href=\"#emulation_off\">When emulation mode is turned OFF<\/a><\/li>\n<\/ul><a href=\"#mysqlnd\">Mysqlnd and buffered queries. Huge datasets. <\/a><a href=\"#comments\">Comments (194)<\/a><p>There are many tutorials on PDO already, but unfortunately, most of them fail to explain the real benefits of PDO, or even promote rather bad practices. The only two exceptions are <a href=\"http:\/\/www.phptherightway.com\/#pdo_extension\">phptherightway.com<\/a> and <a href=\"http:\/\/wiki.hashphp.org\/PDO_Tutorial_for_MySQL_Developers\">hashphp.org<\/a>, but they miss a lot of important information. As a result, half of PDO's features remain in obscurity and are almost never used by PHP developers, who, as a result, are constantly trying to reinvent the wheel which <em>already exists in PDO.<\/em><\/p><p>Unlike those, this tutorial is written by someone who has used PDO for many years, dug through it, and answered thousands questions on Stack Overflow (the <a href=\"http:\/\/stackoverflow.com\/help\/badges\/4220\/pdo\">sole gold PDO badge bearer<\/a>). Following <a href=\"https:\/\/phpdelusions.net\/\">the mission of this site<\/a>, this article will disprove various delusions and bad practices, while showing the right way instead.<\/p><p>Although this tutorial is based on <strong>mysql<\/strong> driver, the information, in general, is applicable for any driver supported.<\/p><h3 id=\"why\" class=\"title-h3\">Why PDO?<\/h3><p>First things first. Why <abbr title=\"PHP Data Objects\">PDO<\/abbr> at all?<\/p><p>PDO is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Database_abstraction_layer\">Database Access Abstraction Layer<\/a>. The abstraction, however, is two-fold: one is widely known but less significant, while another is obscure but of most importance.<\/p><p>Everyone knows that PDO offers unified interface to access <a href=\"http:\/\/php.net\/manual\/en\/pdo.drivers.php\">many different databases<\/a>. Although this feature is magnificent by itself, it doesn't make a big deal for the particular application, where only one database backend is used anyway. And, despite some rumors, it is impossible to switch database backends by changing a single line in PDO config - due to different SQL flavors (to do so, one needs to use an averaged query language like <a href=\"http:\/\/doctrine-orm.readthedocs.org\/projects\/doctrine-orm\/en\/latest\/reference\/dql-doctrine-query-language.html\">DQL<\/a>). Thus, for the average LAMP developer, this point is rather insignificant, and to him, PDO is just a more complicated version of familiar <code>mysql(i)_query()<\/code> function. However, it is not; it is much, much more.<\/p><p>PDO abstracts not only a database API, but also basic operations that otherwise have to be repeated hundreds of times in every application, making your code extremely <abbr title=\"Write Everything Twice\">WET<\/abbr>. <a href=\"https:\/\/phpdelusions.net\/pdo\/mysqli_comparison\">Unlike <em>mysql<\/em> and <em>mysqli<\/em><\/a>, both of which are low level bare APIs not intended to be used directly (but only as a building material for some higher level abstraction layer), <code>PDO<\/code> <em>is<\/em> such an abstraction already. Still incomplete though, but at least usable.<\/p><p>The real PDO benefits are:<\/p><ul><li><strong>security<\/strong> (<em>usable<\/em> prepared statements)<\/li>\n<li><strong>usability<\/strong> (many helper functions to automate routine operations)<\/li>\n<li><strong>reusability<\/strong> (unified API to access multitude of databases, from SQLite to Oracle)<\/li>\n<\/ul><p>Note that although PDO is the best out of native db drivers, for a modern web-application consider to use an ORM with a Query Builder, or any other higher level abstraction library, with only occasional fallback to vanilla PDO. Good ORMs are Doctrine, Eloquent, RedBean, and Yii::AR. Aura.SQL is a good example of a PDO wrapper with many additional features.<\/p><p>Either way, it's good to know the basic tools first. So, let's begin:<\/p><h3 id=\"dsn\" class=\"title-h3\">Connecting. DSN<\/h3><p>PDO has a fancy connection method called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Data_source_name\">DSN<\/a>. It's nothing complicated though - instead of one plain and simple list of options, PDO asks you to input different configuration directives in three different places:<\/p><ul><li><code>database driver<\/code>, <code>host<\/code>, <code>db (schema) name<\/code> and <code>charset<\/code>, as well as less frequently used <code>port<\/code> and <code>unix_socket<\/code> go into DSN;<\/li>\n<li><code>username<\/code> and <code>password<\/code> go to constructor;<\/li>\n<li>all other options go into options array.<\/li>\n<\/ul><p>where DSN is a semicolon-delimited string, consists of <code>param=value<\/code> pairs, that begins from the driver name and a colon:<\/p><pre>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0mysql:host=localhost;dbname=test;port=3306;charset=utf8mb4<br \/>driver^\u00a0\u00a0\u00a0\u00a0^\u00a0colon\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0^param=value\u00a0pair\u00a0\u00a0\u00a0\u00a0^semicolon\u00a0\u00a0<\/pre><p>Note that it's important to follow the proper format - <strong>no spaces or quotes or other decorations have to be used in DSN,<\/strong> but only parameters, values and delimiters, as shown in the <a href=\"http:\/\/php.net\/manual\/en\/pdo.construct.php\">manual<\/a>.<\/p><p>Here goes an example for mysql:<\/p><pre>$host\u00a0=\u00a0'127.0.0.1';<br \/>$db\u00a0\u00a0\u00a0=\u00a0'test';<br \/>$user\u00a0=\u00a0'root';<br \/>$pass\u00a0=\u00a0'';<br \/>$charset\u00a0=\u00a0'utf8mb4';$dsn\u00a0=\u00a0\"mysql:host=$host;dbname=$db;charset=$charset\";<br \/>$options\u00a0=\u00a0[<br \/>\u00a0\u00a0\u00a0\u00a0PDO::ATTR_ERRMODE\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0=&gt;\u00a0PDO::ERRMODE_EXCEPTION,<br \/>\u00a0\u00a0\u00a0\u00a0PDO::ATTR_DEFAULT_FETCH_MODE\u00a0=&gt;\u00a0PDO::FETCH_ASSOC,<br \/>\u00a0\u00a0\u00a0\u00a0PDO::ATTR_EMULATE_PREPARES\u00a0\u00a0\u00a0=&gt;\u00a0false,<br \/>];<br \/>try\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0$pdo\u00a0=\u00a0new\u00a0PDO($dsn,\u00a0$user,\u00a0$pass,\u00a0$options);<br \/>}\u00a0catch\u00a0(\\PDOException\u00a0$e)\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0throw\u00a0new\u00a0\\PDOException($e-&gt;getMessage(),\u00a0(int)$e-&gt;getCode());<br \/>}<\/pre><p>With all aforementioned variables properly set, we will have proper PDO instance in <code>$pdo<\/code> variable.<\/p><p>Important notes for the late mysql extension users:<\/p><ol><li>Unlike old <code>mysql_*<\/code> functions, which can be used anywhere in the code, <code>PDO<\/code> instance is stored in a regular variable, which means it can be inaccessible inside functions - so, one has to make it accessible, by means of passing it via function parameters or using more advanced techniques, such as IoC container.<\/li>\n<li>The connection has to be made only once! No connects in every function. No connects in every class constructor. Otherwise, multiple connections will be created, which will eventually kill your database server. Thus, a sole PDO instance has to be created and then used through whole script execution.<\/li>\n<li>It is very important to <strong>set charset through DSN<\/strong> - that's the only proper way because it tells PDO which charset is going to be used. Therefore forget about running <code>SET NAMES<\/code> query manually, either via <code>query()<\/code> or <code>PDO::MYSQL_ATTR_INIT_COMMAND<\/code>. Only if your PHP version is unacceptably outdated (namely below 5.3.6), you have to use <code>SET NAMES<\/code> query and always turn <a href=\"#emulation\">emulation mode<\/a> off. <\/li>\n<\/ol><p>More details regarding Mysql can be found in the corresponding chapter, <a href=\"https:\/\/phpdelusions.net\/pdo_examples\/connect_to_mysql\">Connecting to MySQL<\/a><\/p><h3 id=\"query\" class=\"title-h3\">Running queries. PDO::query()<\/h3><p>There are two ways to run a query in PDO.\nIf no variables are going to be used in the query, you can use the <a href=\"http:\/\/php.net\/manual\/en\/pdo.query.php\">PDO::query()<\/a> method. It will run your query and return special object of <a href=\"http:\/\/php.net\/manual\/en\/class.pdostatement.php\">PDOStatement class<\/a> which can be roughly compared to a resource, returned by <code>mysql_query()<\/code>, especially in the way you can get actual rows out of it:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0name\u00a0FROM\u00a0users');<br \/>while\u00a0($row\u00a0=\u00a0$stmt-&gt;fetch())<br \/>{<br \/>\u00a0\u00a0\u00a0\u00a0echo\u00a0$row['name']\u00a0.\u00a0\"\\n\";<br \/>}<\/pre><p>Also, the <code>query()<\/code> method allows us to use a neat method chaining for SELECT queries, which will be shown below.<\/p><h3 id=\"prepared\" class=\"title-h3\">Prepared statements. Protection from SQL injections<\/h3><p>This is the main and the only important reason why you were deprived from your beloved <code>mysql_query()<\/code> function and thrown into the harsh world of Data Objects: PDO has prepared statements support out of the box. Prepared statement is <strong>the only proper way to run a query<\/strong>, if any variable is going to be used in it. The reason why it is so important is explained in detail in <a href=\"https:\/\/phpdelusions.net\/sql_injection\">The Hitchhiker's Guide to SQL Injection prevention<\/a>.<\/p><p>So, for every query you run, if at least one variable is going to be used, you have to substitute it with a <strong>placeholder<\/strong>, then prepare your query, and then execute it, passing variables separately.<\/p><p>Long story short, it is not as hard as it seems. In most cases, you need only two functions - <a href=\"http:\/\/php.net\/manual\/en\/pdo.prepare.php\">prepare()<\/a> and <a href=\"http:\/\/php.net\/manual\/en\/pdostatement.execute.php\">execute()<\/a>.<\/p><p>First of all, you have to alter your query, adding placeholders in place of variables. Say, a code like this<\/p><pre>$sql\u00a0=\u00a0\"SELECT\u00a0*\u00a0FROM\u00a0users\u00a0WHERE\u00a0email\u00a0=\u00a0'$email'\u00a0AND\u00a0status='$status'\";<\/pre><p>will become<\/p><pre>$sql\u00a0=\u00a0'SELECT\u00a0*\u00a0FROM\u00a0users\u00a0WHERE\u00a0email\u00a0=\u00a0?\u00a0AND\u00a0status=?';<\/pre><p>or<\/p><pre>$sql\u00a0=\u00a0'SELECT\u00a0*\u00a0FROM\u00a0users\u00a0WHERE\u00a0email\u00a0=\u00a0:email\u00a0AND\u00a0status=:status';<\/pre><p>Note that PDO supports positional (<code>?<\/code>) and named (<code>:email<\/code>) placeholders, the latter always begins from a colon and can be written using letters, digits and underscores only.\nAlso note that <strong>no quotes<\/strong> have to be ever used around placeholders.<\/p><p>Having a query with placeholders, you have to prepare it, using the <code>PDO::prepare()<\/code> method. This function will return the same <code>PDOStatement<\/code> object we were talking about above, but <em>without any data attached to it<\/em>.<\/p><p>Finally, to get the query executed, you must run <code>execute()<\/code> method of this object, passing variables in it, in the form of array. And after that, you will be able to get the resulting data out of statement (if applicable):<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare('SELECT\u00a0*\u00a0FROM\u00a0users\u00a0WHERE\u00a0email\u00a0=\u00a0?\u00a0AND\u00a0status=?');<br \/>$stmt-&gt;execute([$email,\u00a0$status]);<br \/>$user\u00a0=\u00a0$stmt-&gt;fetch();<br \/>\/\/\u00a0or<br \/>$stmt\u00a0=\u00a0$pdo-&gt;prepare('SELECT\u00a0*\u00a0FROM\u00a0users\u00a0WHERE\u00a0email\u00a0=\u00a0:email\u00a0AND\u00a0status=:status');<br \/>$stmt-&gt;execute(['email'\u00a0=&gt;\u00a0$email,\u00a0'status'\u00a0=&gt;\u00a0$status]);<br \/>$user\u00a0=\u00a0$stmt-&gt;fetch();<\/pre><p>As you can see, for the positional placeholders, you have to supply a regular array with values, while for the named placeholders, it has to be an associative array, where keys have to match the placeholder names in the query. You cannot mix positional and named placeholders in the same query.<\/p><p>Please note that positional placeholders let you write shorter code, but are sensitive to the order of arguments (which have to be exactly the same as the order of the corresponding placeholders in the query). While named placeholders make your code more verbose, they allow random binding order.<\/p><p>Also note that despite a widespread delusion, no \"<code>:<\/code>\" in the keys is required.<\/p><p>After the execution you may start getting your data, using all supported methods, as described down in this article.<\/p><p>More examples can be found in the <a href=\"https:\/\/phpdelusions.net\/pdo_examples\">respective article<\/a>.<\/p><h4 id=\"methods\" class=\"title-h4\">Binding methods<\/h4><p>Passing data into <code>execute()<\/code> (like shown above) should be considered default and most convenient method.\nWhen this method is used, all values will be bound as <strong>strings<\/strong> (save for <code>NULL<\/code> values, that will be sent to the query as is, i.e. as SQL <code>NULL<\/code>), but most of time it's all right and won't cause any problem. <\/p><p>However, sometimes it's better to set the data type explicitly. Possible cases are:<\/p><ul><li><a href=\"#limit\">LIMIT<\/a> clause (or any other SQL clause that just cannot accept a string operand) if <a href=\"#emulation\">emulation mode<\/a> is turned ON. <\/li>\n<li>complex queries with non-trivial query plan that can be affected by a wrong operand type<\/li>\n<li>peculiar column types, like <code>BIGINT<\/code> or <code>BOOLEAN<\/code> that require an operand of exact type to be bound (note that in order to bind a BIGINT value with PDO::PARAM_INT you need a <a href=\"#mysqlnd\">mysqlnd<\/a>-based installation). <\/li>\n<\/ul><p>In such a case explicit binding have to be used, for which you have a choice of two functions, <a href=\"http:\/\/php.net\/manual\/en\/pdostatement.bindvalue.php\">bindValue()<\/a> and <a href=\"http:\/\/php.net\/manual\/en\/pdostatement.bindparam.php\">bindParam()<\/a>. The former one has to be preferred, because, unlike <code>bindParam()<\/code> it has no side effects to deal with.<\/p><h4 id=\"bindable\" class=\"title-h4\">Query parts you can bind<\/h4><p>It is very important to understand which query parts you can bind using prepared statements and which you cannot. In fact, the list is overwhelmingly short: only string and numeric literals can be bound. So you can tell that as long as your data can be represented in the query as a numeric or a quoted string literal - it can be bound. For all other cases you cannot use PDO prepared statements at all: neither an identifier, or a comma-separated list, or a part of a quoted string literal or whatever else arbitrary query part cannot be bound using a prepared statement.<\/p><p>Workarounds for the most frequent use cases can be found in the <a href=\"#like\">corresponding part of the article<\/a><\/p><h3 id=\"multiexec\" class=\"title-h3\">Prepared statements. Multiple execution<\/h3><p>Sometimes you can use prepared statements for the multiple execution of a prepared query. It is slightly faster than performing the same query again and again, as it does query parsing only once.\nThis feature would have been more useful if it was possible to execute a statement prepared in another PHP instance. But alas - it is not. So, you are limited to repeating the same query only within the same instance, which is seldom needed in regular PHP scripts and which is limiting the use of this feature to repeated inserts or updates:<\/p><pre>$data\u00a0=\u00a0[<br \/>\u00a0\u00a0\u00a0\u00a01\u00a0=&gt;\u00a01000,<br \/>\u00a0\u00a0\u00a0\u00a05\u00a0=&gt;\u00a0\u00a0300,<br \/>\u00a0\u00a0\u00a0\u00a09\u00a0=&gt;\u00a0\u00a0200,<br \/>];<br \/>$stmt\u00a0=\u00a0$pdo-&gt;prepare('UPDATE\u00a0users\u00a0SET\u00a0bonus\u00a0=\u00a0bonus\u00a0+\u00a0?\u00a0WHERE\u00a0id\u00a0=\u00a0?');<br \/>foreach\u00a0($data\u00a0as\u00a0$id\u00a0=&gt;\u00a0$bonus)<br \/>{<br \/>\u00a0\u00a0\u00a0\u00a0$stmt-&gt;execute([$bonus,\u00a0$id]);<br \/>}<\/pre><p>Note that this feature is a bit overrated. Not only it is needed too seldom to talk about, but the performance gain is not that big - query parsing is <i>real<\/i> fast these times.<\/p><p>Note that you can get this advantage only when <a href=\"#emulation\">emulation mode<\/a> is turned <strong>off<\/strong>.<\/p><h3 id=\"dml\" class=\"title-h3\">Running SELECT INSERT, UPDATE, or DELETE statements<\/h3><p>Come on folks.\nThere is absolutely nothing special in these queries. To PDO they all the same. It doesn't matter which query you are running.<\/p><p>Just like it was shown above, what you need is to prepare a query with placeholders, and then execute it, sending variables separately. Either for <code>DELETE<\/code> and <code>SELECT<\/code> query the process is essentially the same. The only difference is (as <abbr title=\"Data Manipulation Language, INSERT, UPDATE and DELETE queries\">DML<\/abbr> queries do not return any data), that you can use the method chaining and thus call <code>execute()<\/code> right along with <code>prepare()<\/code>:<\/p><pre>$sql\u00a0=\u00a0\"UPDATE\u00a0users\u00a0SET\u00a0name\u00a0=\u00a0?\u00a0WHERE\u00a0id\u00a0=\u00a0?\";<br \/>$pdo-&gt;prepare($sql)-&gt;execute([$name,\u00a0$id]);<\/pre><p>However, if you want to get the number of affected rows, the code will have to be the same boresome three lines:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"DELETE\u00a0FROM\u00a0goods\u00a0WHERE\u00a0category\u00a0=\u00a0?\");<br \/>$stmt-&gt;execute([$cat]);<br \/>$deleted\u00a0=\u00a0$stmt-&gt;rowCount();<\/pre><p>More examples can be found in the <a href=\"https:\/\/phpdelusions.net\/pdo_examples\">respective article<\/a>.<\/p><h3 id=\"foreach\" class=\"title-h3\">Getting data out of statement. foreach()<\/h3><p>The most basic and direct way to get multiple rows from a statement would be <code>foreach()<\/code> loop. Thanks to <a href=\"http:\/\/php.net\/manual\/en\/class.traversable.php\">Traversable<\/a> interface, <code>PDOStatement<\/code> can be iterated over by using <code>foreach()<\/code> operator:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0name\u00a0FROM\u00a0users');<br \/>foreach\u00a0($stmt\u00a0as\u00a0$row)<br \/>{<br \/>\u00a0\u00a0\u00a0\u00a0echo\u00a0$row['name']\u00a0.\u00a0\"\\n\";<br \/>}<\/pre><p>Note that this method is memory-friendly, as it doesn't load all the resulting rows in the memory but delivers them one by one (though keep in mind this <a href=\"#mysqlnd\">issue<\/a>).<\/p><h3 id=\"fetch\" class=\"title-h3\">Getting data out of statement. fetch()<\/h3><p>We have seen this function already, but let's take a closer look. It fetches a single row from database, and moves the internal pointer in the result set, so consequent calls to this function will return all the resulting rows one by one. Which makes this method a rough analogue to <code>mysql_fetch_array()<\/code> but it works in a slightly different way: instead of many separate functions (<code>mysql_fetch_assoc()<\/code>, <code>mysql_fetch_row()<\/code>, etc), there is only one, but its behavior can be changed by a parameter. There are many fetch modes in PDO, and we will discuss them later, but here are few for starter:<\/p><ul><li><code>PDO::FETCH_NUM<\/code> returns enumerated array<\/li>\n<li><code>PDO::FETCH_ASSOC<\/code> returns associative array<\/li>\n<li><code>PDO::FETCH_BOTH<\/code> - both of the above<\/li>\n<li><code>PDO::FETCH_OBJ<\/code> returns object<\/li>\n<li><code>PDO::FETCH_LAZY<\/code> allows all three (numeric associative and object) methods without memory overhead.<\/li>\n<\/ul><p>From the above you can tell that this function have to be used in two cases:<\/p><ol><li>\n<p>When only one row is expected - to get that only row. For example,<\/p>\n<pre>$row\u00a0=\u00a0$stmt-&gt;fetch(PDO::FETCH_ASSOC);<\/pre>\n<p>Will give you single row from the statement, in the form of associative array.<\/p>\n<\/li>\n<li>When we need to process the returned data somehow before use. In this case it have to be run through usual while loop, like one shown <a href=\"#query\">above<\/a>.<\/li>\n<\/ol><p>Another useful mode is <code>PDO::FETCH_CLASS<\/code>, which can create an object of particular class<\/p><pre>$news\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0*\u00a0FROM\u00a0news')-&gt;fetchAll(PDO::FETCH_CLASS,\u00a0'News');<\/pre><p>will produce an array filled with objects of News class, setting class properties from returned values. Note that in this mode<\/p><ul><li>properties are set <em>before<\/em> constructor call<\/li>\n<li>for all undefined properties <code>__set<\/code> magic method will be called<\/li>\n<li>if there is no <code>__set<\/code> method in the class, then new property will be created<\/li>\n<li>private properties will be filled as well, which is a bit unexpected but quite handy<\/li>\n<\/ul><p>Note that default mode is <code>PDO::FETCH_BOTH<\/code>, but you can change it using <code>PDO::ATTR_DEFAULT_FETCH_MODE<\/code> configuration option as shown in the connection example. Thus, once set, it can be omitted most of the time.<\/p><h4 id=\"returntypes\" class=\"title-h4\">Return types.<\/h4><p>Only when PDO is built upon <a href=\"#mysqlnd\">mysqlnd<\/a> <em>and<\/em> <a href=\"#emulation\">emulation mode<\/a> is <strong>off<\/strong>, then PDO will return <code>int<\/code> and <code>float<\/code> values with respective types. Say, if we create a table<\/p><pre>create\u00a0table\u00a0typetest\u00a0(string\u00a0varchar(255),\u00a0`int`\u00a0int,\u00a0`float`\u00a0float,\u00a0`null`\u00a0int);<br \/>insert\u00a0into\u00a0typetest\u00a0values('foo',1,1.1,NULL);<\/pre><p>And then query it from mysqlnd-based PDO with emulation turned off, the output will be<\/p><pre>array(4)\u00a0{<br \/>\u00a0\u00a0[\"string\"]\u00a0=&gt;\u00a0string(3)\u00a0\"foo\"<br \/>\u00a0\u00a0[\"int\"]\u00a0\u00a0\u00a0\u00a0=&gt;\u00a0int(1)<br \/>\u00a0\u00a0[\"float\"]\u00a0\u00a0=&gt;\u00a0float(1.1)<br \/>\u00a0\u00a0[\"null\"]\u00a0\u00a0\u00a0=&gt;\u00a0NULL<br \/>}<\/pre><p>Otherwise the familiar <code>mysql_fetch_array()<\/code> behavior will be followed - all values returned as strings with only <code>NULL<\/code> returned as <code>NULL<\/code>.<\/p><p>If for some reason you don't like this behavior and prefer the old style with strings and NULLs only, then you can use the following configuration option to override it:<\/p><pre>$pdo-&gt;setAttribute(PDO::ATTR_STRINGIFY_FETCHES,\u00a0true);\u00a0<\/pre><p>Note that for the <code>DECIMAL<\/code> type the string is always returned, due to nature of this type intended to retain the precise value, unlike deliberately non-precise FLOAT and DOUBLE types.<\/p><h3 id=\"fetchcolumn\" class=\"title-h3\">Getting data out of statement. fetchColumn()<\/h3><p>A neat helper function that returns value of the singe field of returned row. Very handy when we are selecting only one field:<\/p><pre>\/\/\u00a0Getting\u00a0the\u00a0name\u00a0based\u00a0on\u00a0id<br \/>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a0name\u00a0FROM\u00a0table\u00a0WHERE\u00a0id=?\");<br \/>$stmt-&gt;execute([$id]);<br \/>$name\u00a0=\u00a0$stmt-&gt;fetchColumn();\/\/\u00a0getting\u00a0number\u00a0of\u00a0rows\u00a0in\u00a0the\u00a0table\u00a0utilizing\u00a0method\u00a0chaining<br \/>$count\u00a0=\u00a0$pdo-&gt;query(\"SELECT\u00a0count(*)\u00a0FROM\u00a0table\")-&gt;fetchColumn();<\/pre><h3 id=\"fetchall\" class=\"title-h3\">Getting data out of statement in dozens different formats. fetchAll()<\/h3><p>That's most interesting function, with most astonishing features.\nMostly thanks to its existence one can call PDO a wrapper, as this function can automate many operations otherwise performed manually.<\/p><p><code>PDOStatement::fetchAll()<\/code> returns an array that consists of <b>all the rows<\/b> returned by the query.\nFrom this fact we can make two conclusions:<\/p><ol><li>This function should not be used, if many rows has been selected. In such a case conventional while loop ave to be used, fetching rows one by one instead of getting them all into array at once. \"Many\" means more than it is suitable to be shown on the average web page.<\/li>\n<li>This function is mostly useful in a modern web application that never outputs data right away during fetching, but rather passes it to template.<\/li>\n<\/ol><p>You'd be amazed, in how many different formats this function can return data in (and how little an average PHP user knows of them), all controlled by <code>PDO::FETCH_*<\/code> variables.\nSome of them are:<\/p><h4 id=\"array\" class=\"title-h4\">Getting a plain array.<\/h4><p>By default, this function will return just simple enumerated array consists of all the returned rows. Row formatting constants, such as <code>PDO::FETCH_NUM<\/code>, <code>PDO::FETCH_ASSOC<\/code>, <code>PDO::FETCH_OBJ<\/code> etc can change the row format. <\/p><pre>$data\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0name\u00a0FROM\u00a0users')-&gt;fetchAll(PDO::FETCH_ASSOC);<br \/>var_export($data);<br \/>\/*<br \/>array\u00a0(<br \/>\u00a0\u00a00\u00a0=&gt;\u00a0array('John'),<br \/>\u00a0\u00a01\u00a0=&gt;\u00a0array('Mike'),<br \/>\u00a0\u00a02\u00a0=&gt;\u00a0array('Mary'),<br \/>\u00a0\u00a03\u00a0=&gt;\u00a0array('Kathy'),<br \/>)*\/<\/pre><h4 id=\"column\" class=\"title-h4\">Getting a column.<\/h4><p>It is often very handy to get plain one-dimensional array right out of the query, if only one column out of many rows being fetched. Here you go:<\/p><pre>$data\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0name\u00a0FROM\u00a0users')-&gt;fetchAll(PDO::FETCH_COLUMN);<br \/>\/*\u00a0array\u00a0(<br \/>\u00a0\u00a00\u00a0=&gt;\u00a0'John',<br \/>\u00a0\u00a01\u00a0=&gt;\u00a0'Mike',<br \/>\u00a0\u00a02\u00a0=&gt;\u00a0'Mary',<br \/>\u00a0\u00a03\u00a0=&gt;\u00a0'Kathy',<br \/>)*\/<\/pre><h4 id=\"pairs\" class=\"title-h4\">Getting key-value pairs.<\/h4><p>Also extremely useful format, when we need to get the same column, but indexed not by numbers in order but by another field. Here goes <code>PDO::FETCH_KEY_PAIR<\/code> constant:<\/p><pre>$data\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0id,\u00a0name\u00a0FROM\u00a0users')-&gt;fetchAll(PDO::FETCH_KEY_PAIR);<br \/>\/*\u00a0array\u00a0(<br \/>\u00a0\u00a0104\u00a0=&gt;\u00a0'John',<br \/>\u00a0\u00a0110\u00a0=&gt;\u00a0'Mike',<br \/>\u00a0\u00a0120\u00a0=&gt;\u00a0'Mary',<br \/>\u00a0\u00a0121\u00a0=&gt;\u00a0'Kathy',<br \/>)*\/<\/pre><p>Note that you have to select only two columns for this mode, first of which have to be unique.<\/p><h4 id=\"unique\" class=\"title-h4\">Getting rows indexed by unique field<\/h4><p>Same as above, but getting not one column but full row, yet indexed by an unique field, thanks to\n<code>PDO::FETCH_UNIQUE<\/code> constant:<\/p><pre>$data\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0*\u00a0FROM\u00a0users')-&gt;fetchAll(PDO::FETCH_UNIQUE);<br \/>\/*\u00a0array\u00a0(<br \/>\u00a0\u00a0104\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'John',<br \/>\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Toyota',<br \/>\u00a0\u00a0),<br \/>\u00a0\u00a0110\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'Mike',<br \/>\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Ford',<br \/>\u00a0\u00a0),<br \/>\u00a0\u00a0120\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'Mary',<br \/>\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Mazda',<br \/>\u00a0\u00a0),<br \/>\u00a0\u00a0121\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'Kathy',<br \/>\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Mazda',<br \/>\u00a0\u00a0),<br \/>)*\/<\/pre><p>Note that first column selected have to be unique (in this query it is assumed that first column is id, but to be sure better list it explicitly).<\/p><h4 id=\"group\" class=\"title-h4\">Getting rows grouped by some field<\/h4><p><code>PDO::FETCH_GROUP<\/code> will group rows into a nested array, where indexes will be unique values from the first columns, and values will be arrays similar to ones returned by regular <code>fetchAll()<\/code>. The following code, for example, will separate boys from girls and put them into different arrays:<\/p><pre>$data\u00a0=\u00a0$pdo-&gt;query('SELECT\u00a0sex,\u00a0name,\u00a0car\u00a0FROM\u00a0users')-&gt;fetchAll(PDO::FETCH_GROUP);<br \/>array\u00a0(<br \/>\u00a0\u00a0'male'\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a00\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'John',<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Toyota',<br \/>\u00a0\u00a0\u00a0\u00a0),<br \/>\u00a0\u00a0\u00a0\u00a01\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'Mike',<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Ford',<br \/>\u00a0\u00a0\u00a0\u00a0),<br \/>\u00a0\u00a0),<br \/>\u00a0\u00a0'female'\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a00\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'Mary',<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Mazda',<br \/>\u00a0\u00a0\u00a0\u00a0),<br \/>\u00a0\u00a0\u00a0\u00a01\u00a0=&gt;\u00a0array\u00a0(<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'name'\u00a0=&gt;\u00a0'Kathy',<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0'car'\u00a0=&gt;\u00a0'Mazda',<br \/>\u00a0\u00a0\u00a0\u00a0),<br \/>\u00a0\u00a0),<br \/>)<\/pre><p>So, this is the ideal solution for such a popular demand like \"group events by date\" or \"group goods by category\". Some real life use cases:<\/p><ul><li><a href=\"http:\/\/stackoverflow.com\/q\/35317555\/285587\">How to multiple query results in order to reduce the query number?<\/a><\/li>\n<li><a href=\"http:\/\/stackoverflow.com\/a\/41263854\/285587\">List records grouped by category name<\/a> <\/li>\n<\/ul><p><strong>Other modes<\/strong><\/p><p>Of course, there is a <code>PDO::FETCH_FUNC<\/code> for the functional programming fans. <\/p><p>More modes are coming soon.<\/p><h3 id=\"errors\" class=\"title-h3\">Error handling. Exceptions<\/h3><p>Although there are several error handling modes in PDO, the only proper one is <code>PDO::ERRMODE_EXCEPTION<\/code>. So, one ought to always set it this way, either by adding this line after creation of PDO instance, <\/p><pre>$dbh-&gt;setAttribute(\u00a0PDO::ATTR_ERRMODE,\u00a0PDO::ERRMODE_EXCEPTION\u00a0);<\/pre><p>or as a connection option, as demonstrated in the example above. And this is <em>all<\/em> you need for the basic error reporting. <\/p><h4 id=\"reporting_errors\" class=\"title-h4\">Reporting PDO errors<\/h4><p><strong>TL;DR:<\/strong><br \/>Despite what all other tutorials say, <strong>you don't need a <code>try..catch<\/code> operator to report PDO errors<\/strong>. Catch an exception only if you have a handling scenario other than just reporting it. Otherwise just let it bubble up to a site-wide handler (note that you don't have to write one, there is a basic built-in handler in PHP, which is quite good).<\/p><p>The only exception (pun not intended) is the creation of the PDO instance, which in case of error might reveal the connection credentials (that would be the part of the stack trace). In order to hide them, we can wrap the connection code into a <code>try..catch<\/code> operator and then throw a new <code>ErrorException<\/code> that contains only the message but not the credentials.<\/p><p><strong>A long rant on the matter:<\/strong>   <\/p><p>Despite a widespread delusion, you should never catch errors to report them. A module (like a database layer) should not report its errors. This function has to be delegated to an application-wide handler. All we need is to raise an error (in the form of exception) - which we already did. That's all. Nor should you \"<em>always wrap your PDO operations in a <code>try\/catch<\/code><\/em>\" like the most popular tutorial from tutsplus recommends. Quite contrary, catching an exception should be rather an exceptional case (pun intended).<\/p><p>In fact, there is nothing special in PDO exceptions - they are errors all the same. Thus, you have to treat them exactly the same way as other errors. If you had an error handler before, you shouldn't create a dedicated one for PDO. If you didn't care - it's all right too, as PHP is good with basic error handling and will conduct PDO exceptions all right.<\/p><p>Exception handling is one of the problems with PDO tutorials. Being acquainted with exceptions for the first time when starting with PDO, authors consider exceptions dedicated to this library, and start diligently (but improperly) handling exceptions for PDO only. This is utter nonsense. If one paid no special attention to any exceptions before, they shouldn't have changed their habit for PDO. If one didn't use <code>try..catch<\/code> before, they should keep with that, eventually learning how to use exceptions and when it is suitable to catch them.<\/p><p>So now you can tell that the PHP manual is wrong, <a href=\"http:\/\/php.net\/manual\/en\/pdo.connections.php\">stating that<\/a><\/p><blockquote>\n<p>If your application does not catch the exception thrown from the PDO constructor, the default action taken by the zend engine is to terminate the script and display a back trace. This back trace will likely reveal the full database connection details, including the username and password.<\/p>\n<\/blockquote><p>However, <strong>there is no such thing as \"the displaying of a back trace\"!<\/strong> What zend engine <em>really<\/em> does is just convert an uncaught exception into a fatal error. And then this fatal error is treated <strong>like any other error<\/strong> - so it will be displayed only if appropriate <code>php.ini<\/code> directive is set.\nThus, although you may or you may not catch an exception, it has absolutely nothing to do with displaying sensitive information, because it's <strong>a totally different configuration setting<\/strong> in response to this. So, do not catch PDO exceptions to report them. Instead, configure your server properly:<\/p><p>On a development server just turn displaying errors on:<\/p><pre>ini_set('display_errors',\u00a01);<\/pre><p>While on a production server turn displaying errors off while logging errors on:<\/p><pre>ini_set('display_errors',\u00a00);<br \/>ini_set('log_errors',\u00a01);<\/pre><ul><li>keep in mind that <em>there are other errors that shouldn't be revealed to the user as well.<\/em>  <\/li>\n<\/ul><h4 id=\"catch\" class=\"title-h4\">Catching PDO exceptions<\/h4><p>You may want to catch PDO errors only in two cases:<\/p><ol><li>If you are writing a wrapper for PDO, and you want to augment the error info with some additional data, like query string. In this case, catch the exception, gather the required information, and <strong>re-throw another Exception<\/strong>.<\/li>\n<li>\n<p>If you have a <strong>certain scenario<\/strong> for handling errors in the particular part of code. Some examples are: <\/p>\n<ul><li>if the error can be bypassed, you can use try..catch for this. However, do not make it a habit. Empty catch in every aspect works as error suppression operator, <a href=\"http:\/\/programmers.stackexchange.com\/questions\/219788\/is-error-suppressing-bad-practice\">and so equally evil it is<\/a>.<\/li>\n<li>if there is an action that has to be taken in case of failure, i.e. <a href=\"#transactions\">transaction<\/a> rollback.<\/li>\n<li>if you are waiting for a particular error to handle. In this case, catch the exception, see if the error is one you're looking for, and then handle this one. Otherwise just throw it again - so it will bubble up to the handler in the usual way.<\/li>\n<\/ul><\/li>\n<\/ol><p>E.g.:<\/p><pre>try\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0$pdo-&gt;prepare(\"INSERT\u00a0INTO\u00a0users\u00a0VALUES\u00a0(NULL,?,?,?,?)\")-&gt;execute($data);<br \/>}\u00a0catch\u00a0(PDOException\u00a0$e)\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0$existingkey\u00a0=\u00a0\"Integrity\u00a0constraint\u00a0violation:\u00a01062\u00a0Duplicate\u00a0entry\";<br \/>\u00a0\u00a0\u00a0\u00a0if\u00a0(strpos($e-&gt;getMessage(),\u00a0$existingkey)\u00a0!==\u00a0FALSE)\u00a0{\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\/\/\u00a0Take\u00a0some\u00a0action\u00a0if\u00a0there\u00a0is\u00a0a\u00a0key\u00a0constraint\u00a0violation,\u00a0i.e.\u00a0duplicate\u00a0name<br \/>\u00a0\u00a0\u00a0\u00a0}\u00a0else\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0throw\u00a0$e;<br \/>\u00a0\u00a0\u00a0\u00a0}<br \/>}<\/pre><p>However, in general, no dedicated treatment for PDO exceptions is ever needed. In short, to have PDO errors properly reported:<\/p><ol><li>Set PDO in exception mode.<\/li>\n<li>Do not use <code>try..catch<\/code> to report errors. <\/li>\n<li>Configure PHP for proper error reporting\n<ul><li>on a <strong>live<\/strong> site set <code>display_errors=off<\/code> and <code>log_errors=on<\/code><\/li>\n<li>on a <strong>development<\/strong> site, you may want to set <code>display_errors=on<\/code><\/li>\n<li>of course, <code>error_reporting<\/code> has to be set to E_ALL in both cases<\/li>\n<\/ul><\/li>\n<\/ol><p>As a result, you will be always notified of all database errors without a single line of extra code! <a href=\"https:\/\/phpdelusions.net\/try-catch\">Further reading<\/a>.<\/p><h3 id=\"count\" class=\"title-h3\">Getting row count with PDO<\/h3><p>You don't needed it.<\/p><p>Although PDO offers a function for returning the number of rows found by the query, <code>PDOstatement::rowCount()<\/code>, you scarcely need it. Really.<\/p><p>If you think it over, you will see that this is a most misused function in the web. Most of time it is used not to <i>count<\/i> anything, but as a mere flag - just to see if there was any data returned. But for such a case you have the data itself! Just get your data, using either <code>fetch()<\/code> or <code>fetchAll()<\/code> - and it will serve as such a flag all right! Say, to see if there is any user with such a name, just select a row:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a01\u00a0FROM\u00a0users\u00a0WHERE\u00a0name=?\");<br \/>$stmt-&gt;execute([$name]);<br \/>$userExists\u00a0=\u00a0$stmt-&gt;fetchColumn();<\/pre><p>Exactly the same thing with getting either a single row or an array with rows:<\/p><pre>$data\u00a0=\u00a0$pdo-&gt;query(\"SELECT\u00a0*\u00a0FROM\u00a0table\")-&gt;fetchAll();<br \/>if\u00a0($data)\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0\/\/\u00a0You\u00a0have\u00a0the\u00a0data!\u00a0No\u00a0need\u00a0for\u00a0the\u00a0rowCount()\u00a0ever!<br \/>}<\/pre><p>Remember that here you don't need the <i>count<\/i>, the actual number of rows, but rather a boolean flag. So you got it.<\/p><p>Not to mention that the second most popular use case for this function should never be used at all. One should never use the <code>rowCount()<\/code> to count rows in database! Instead, one has to ask a database to count them, and return the result in a <b>single<\/b> row:<\/p><pre>$count\u00a0=\u00a0$pdo-&gt;query(\"SELECT\u00a0count(1)\u00a0FROM\u00a0t\")-&gt;fetchColumn();<\/pre><p>is the only proper way.<\/p><p>In essence:<\/p><ul><li>if you need to know how many rows in the table, use <code>SELECT COUNT(*)<\/code> query.<\/li>\n<li>if you need to know whether your query returned any data - check that data.<\/li>\n<li>if you still need to know how many rows has been returned by some query (though I hardly can imagine a case), then you can either use <code>rowCount()<\/code> or simply call <code>count()<\/code> on the array returned by <code>fetchAll()<\/code> (if applicable).<\/li>\n<\/ul><p>Thus you could tell that the <a href=\"http:\/\/stackoverflow.com\/a\/883382\/285587\/\">top answer for this question on Stack Overflow<\/a> is essentially pointless and harmful - a call to <code>rowCount()<\/code> could be never substituted with <code>SELECT count(*)<\/code> query - their purpose is essentially different, while running an extra query only to get the number of rows returned by other query makes absolutely no sense. <\/p><h3 id=\"affected\" class=\"title-h3\">Affected rows and insert id<\/h3><p>PDO is using the same function for returning both number of rows returned by SELECT statement and number of rows affected by <abbr title=\"Data Manipulation Language, INSERT, UPDATE and DELETE queries\">DML<\/abbr> queries - <code>PDOstatement::rowCount()<\/code>. Thus, to get the number of rows affected, just call this function after performing a query. <\/p><p>Another frequently asked question is caused by the fact that mysql won't update the row, if new value is the same as old one. Thus number of rows affected could differ from the number of rows matched by the WHERE clause. Sometimes it is required to know this latter number. <\/p><p>Although you can tell <code>rowCount()<\/code> to return the number of rows matched instead of rows affected by setting <code>PDO::MYSQL_ATTR_FOUND_ROWS<\/code> option to TRUE, but, as this is a connection-only option and thus you cannot change it's behavior during runtime, you will have to stick to only one mode for the application, which could be not very convenient. <\/p><p>Note that <code>PDO::MYSQL_ATTR_FOUND_ROWS<\/code> is not guaranteed to work, as it's described in the <a href=\"https:\/\/phpdelusions.net\/pdo#comment-276\">comment<\/a> below.<\/p><p>Unfortunately, there is no PDO counterpart for the <code>mysql(i)_info()<\/code> function which output can be easily parsed and desired number found. This is one of minor PDO drawbacks.<\/p><p>An auto-generated identifier from a sequence or auto_inclement field in mysql can be obtained from the <a href=\"http:\/\/php.net\/manual\/en\/pdo.lastinsertid.php\">PDO::lastInsertId<\/a> function. An answer to a frequently asked question, \"whether this function is safe to use in concurrent environment?\" is positive: yes, it is safe. Being just an interface to MySQL C API <a href=\"http:\/\/dev.mysql.com\/doc\/refman\/5.7\/en\/mysql-insert-id.html\">mysql_insert_id()<\/a> function it's perfectly safe. <\/p><h3 id=\"like\" class=\"title-h3\">Prepared statements and LIKE clause<\/h3><p>Despite PDO's overall ease of use, there are some gotchas anyway, and I am going to explain some.<\/p><p>One of them is using placeholders with <code>LIKE<\/code> SQL clause. At first one would think that such a query will do:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0name\u00a0LIKE\u00a0'%?%'\");<\/pre><p>but soon they will learn that it will produce an error. To understand its nature one has to understand that, <a href=\"#bindable\">like it was said above<\/a>, <i>a placeholder have to represent a complete data literal only<\/i> - a string or a number namely. And by no means can it represent either a part of a literal or some arbitrary SQL part. So, when working with LIKE, we have to prepare our <strong>complete literal<\/strong> first, and then send it to the query the usual way:<\/p><pre>$search\u00a0=\u00a0\"%$search%\";<br \/>$stmt\u00a0\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0name\u00a0LIKE\u00a0?\");<br \/>$stmt-&gt;execute([$search]);<br \/>$data\u00a0=\u00a0$stmt-&gt;fetchAll();<\/pre><h3 id=\"in\" class=\"title-h3\">Prepared statements and IN clause<\/h3><p>Just like it was said above, it is impossible to substitute an arbitrary query part with a placeholder. Any string you bind through a placeholder will be put into query as a single string literal. For example, a <em>string<\/em> <code>'1,2,3'<\/code> will be bound as a <em>string<\/em>, resulting in<\/p><pre>SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0column\u00a0IN\u00a0('1,2,3')<\/pre><p>making SQL to search for just <em>one<\/em> value.<\/p><p>To make it right, one needs separated values, to make a query look like<\/p><pre>SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0column\u00a0IN\u00a0('1','2','3')<\/pre><p>Thus, for the comma-separated values, like for <code>IN()<\/code> SQL operator, one must create a set of <code>?<\/code>s manually and put them into the query:<\/p><pre>$arr\u00a0=\u00a0[1,2,3];<br \/>$in\u00a0\u00a0=\u00a0str_repeat('?,',\u00a0count($arr)\u00a0-\u00a01)\u00a0.\u00a0'?';<br \/>$sql\u00a0=\u00a0\"SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0column\u00a0IN\u00a0($in)\";<br \/>$stm\u00a0=\u00a0$db-&gt;prepare($sql);<br \/>$stm-&gt;execute($arr);<br \/>$data\u00a0=\u00a0$stm-&gt;fetchAll();<\/pre><p>Not very convenient, but compared to mysqli it's <a href=\"https:\/\/phpdelusions.net\/pdo\/mysqli_comparison#in\"><em>amazingly<\/em> concise<\/a>.<\/p><p>In case there are other placeholders in the query, you could use <code>array_merge()<\/code> function to join all the variables into a single array, adding your other variables in the form of arrays, in the order they appear in your query:<\/p><pre>$arr\u00a0=\u00a0[1,2,3];<br \/>$in\u00a0\u00a0=\u00a0str_repeat('?,',\u00a0count($arr)\u00a0-\u00a01)\u00a0.\u00a0'?';<br \/>$sql\u00a0=\u00a0\"SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0foo=?\u00a0AND\u00a0column\u00a0IN\u00a0($in)\u00a0AND\u00a0bar=?\u00a0AND\u00a0baz=?\";<br \/>$stm\u00a0=\u00a0$db-&gt;prepare($sql);<br \/>$params\u00a0=\u00a0array_merge([$foo],\u00a0$arr,\u00a0[$bar,\u00a0$baz]);<br \/>$stm-&gt;execute($params);<br \/>$data\u00a0=\u00a0$stm-&gt;fetchAll();<\/pre><p>In case you are using named placeholders, the code would be a little more complex, as you have to create a sequence of the named placeholders, e.g. <code>:id0,:id1,:id2<\/code>. So the code would be:<\/p><pre>\/\/\u00a0other\u00a0parameters\u00a0that\u00a0are\u00a0going\u00a0into\u00a0query<br \/>$params\u00a0=\u00a0[\"foo\"\u00a0=&gt;\u00a0\"foo\",\u00a0\"bar\"\u00a0=&gt;\u00a0\"bar\"];$ids\u00a0=\u00a0[1,2,3];<br \/>$in\u00a0=\u00a0\"\";<br \/>foreach\u00a0($ids\u00a0as\u00a0$i\u00a0=&gt;\u00a0$item)<br \/>{<br \/>\u00a0\u00a0\u00a0\u00a0$key\u00a0=\u00a0\":id\".$i;<br \/>\u00a0\u00a0\u00a0\u00a0$in\u00a0.=\u00a0\"$key,\";<br \/>\u00a0\u00a0\u00a0\u00a0$in_params[$key]\u00a0=\u00a0$item;\u00a0\/\/\u00a0collecting\u00a0values\u00a0into\u00a0key-value\u00a0array<br \/>}<br \/>$in\u00a0=\u00a0rtrim($in,\",\");\u00a0\/\/\u00a0:id0,:id1,:id2$sql\u00a0=\u00a0\"SELECT\u00a0*\u00a0FROM\u00a0table\u00a0WHERE\u00a0foo=:foo\u00a0AND\u00a0id\u00a0IN\u00a0($in)\u00a0AND\u00a0bar=:bar\";<br \/>$stm\u00a0=\u00a0$db-&gt;prepare($sql);<br \/>$stm-&gt;execute(array_merge($params,$in_params));\u00a0\/\/\u00a0just\u00a0merge\u00a0two\u00a0arrays<br \/>$data\u00a0=\u00a0$stm-&gt;fetchAll();<\/pre><p>Luckily, for the named placeholders we don't have to follow the strict order, so we can merge our arrays in any order.<\/p><h3 id=\"identifiers\" class=\"title-h3\">Prepared statements and table names<\/h3><p>On Stack Overflow I've seen overwhelming number of PHP users implementing <a href=\"https:\/\/phpdelusions.net\/pdo\/lame_update\">the most fatal PDO code<\/a>, thinking that only data values have to be protected. But of course it is not.<\/p><p>Unfortunately, PDO has no placeholder for identifiers (table and field names), so a developer must manually format them. <\/p><p>For <strong>mysql<\/strong> to format an identifier, follow these two rules:<\/p><ul><li>Enclose identifier in backticks.<\/li>\n<li>Escape backticks inside by doubling them.<\/li>\n<\/ul><p>so the code would be:<\/p><pre>$table\u00a0=\u00a0\"`\".str_replace(\"`\",\"``\",$table).\"`\";<\/pre><p>After such formatting, it is safe to insert the <code>$table<\/code> variable into query.<\/p><p>For other databases rules will be different but it is essential to understand that using only delimiters is not enough - delimiters themselves should be escaped.<\/p><p>It is also important to always check dynamic identifiers against a list of allowed values. Here is a brief example:<\/p><pre>$orders\u00a0\u00a0=\u00a0[\"name\",\"price\",\"qty\"];\u00a0\/\/field\u00a0names<br \/>$key\u00a0\u00a0\u00a0\u00a0\u00a0=\u00a0array_search($_GET['sort'],$orders);\u00a0\/\/\u00a0see\u00a0if\u00a0we\u00a0have\u00a0such\u00a0a\u00a0name<br \/>$orderby\u00a0=\u00a0$orders[$key];\u00a0\/\/if\u00a0not,\u00a0first\u00a0one\u00a0will\u00a0be\u00a0set\u00a0automatically.\u00a0smart\u00a0enuf\u00a0:)<br \/>$query\u00a0\u00a0\u00a0=\u00a0\"SELECT\u00a0*\u00a0FROM\u00a0`table`\u00a0ORDER\u00a0BY\u00a0$orderby\";\u00a0\/\/value\u00a0is\u00a0safe<\/pre><p>Or, extending this approach for the INSERT\/UPDATE statements (as Mysql supports SET for both),<\/p><pre>$data\u00a0=\u00a0['name'\u00a0=&gt;\u00a0'foo',\u00a0'submit'\u00a0=&gt;\u00a0'submit'];\u00a0\/\/\u00a0data\u00a0for\u00a0insert<br \/>$allowed\u00a0=\u00a0[\"name\",\u00a0\"surname\",\u00a0\"email\"];\u00a0\/\/\u00a0allowed\u00a0fields<br \/>$values\u00a0=\u00a0[];<br \/>$set\u00a0=\u00a0\"\";<br \/>foreach\u00a0($allowed\u00a0as\u00a0$field)\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0if\u00a0(isset($data[$field]))\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$set.=\"`\".str_replace(\"`\",\u00a0\"``\",\u00a0$field).\"`\".\u00a0\"=:$field,\u00a0\";<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$values[$field]\u00a0=\u00a0$data[$field];<br \/>\u00a0\u00a0\u00a0\u00a0}<br \/>}<br \/>$set\u00a0=\u00a0substr($set,\u00a00,\u00a0-2);\u00a0<\/pre><p>This code will produce the correct sequence for SET operator that will contain only allowed fields and placeholders like this:<\/p><pre>`name`=:foo<\/pre><p>as well as <code>$values<\/code> array for <code>execute()<\/code>, which can be used like this<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"INSERT\u00a0INTO\u00a0users\u00a0SET\u00a0$set\");<br \/>$stmt-&gt;execute($values);<\/pre><p>Yes, it looks extremely ugly, but that is all PDO can offer. <\/p><h3 id=\"limit\" class=\"title-h3\">A problem with LIMIT clause<\/h3><p>Another problem is related to the SQL <code>LIMIT<\/code> clause. When in <a href=\"#emulation\">emulation mode<\/a> (which is on by default), PDO substitutes placeholders with actual data, instead of sending it separately. And with \"lazy\" binding (using array in <code>execute()<\/code>), PDO treats every parameter as a string. As a result, the prepared <code>LIMIT ?,?<\/code> query becomes <code>LIMIT '10', '10'<\/code> which is invalid syntax that causes query to fail.<\/p><p>There are two solutions:<\/p><p>One is <a href=\"#emulation\">turning emulation off<\/a> (as MySQL can sort all placeholders properly). To do so one can run this code:<\/p><pre>$conn-&gt;setAttribute(\u00a0PDO::ATTR_EMULATE_PREPARES,\u00a0false\u00a0);<\/pre><p>And parameters can be kept in <code>execute()<\/code>:<\/p><pre>$conn-&gt;setAttribute(\u00a0PDO::ATTR_EMULATE_PREPARES,\u00a0false\u00a0);<br \/>$stmt\u00a0=\u00a0$pdo-&gt;prepare('SELECT\u00a0*\u00a0FROM\u00a0table\u00a0LIMIT\u00a0?,\u00a0?');<br \/>$stmt-&gt;execute([$offset,\u00a0$limit]);<br \/>$data\u00a0=\u00a0$stmt-&gt;fetchAll();<\/pre><p>Another way would be to bind these variables explicitly while setting the proper param type:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare('SELECT\u00a0*\u00a0FROM\u00a0table\u00a0LIMIT\u00a0?,\u00a0?');<br \/>$stmt-&gt;bindParam(1,\u00a0$offset,PDO::PARAM_INT);<br \/>$stmt-&gt;bindParam(2,\u00a0$limit,PDO::PARAM_INT);<br \/>$stmt-&gt;execute();<br \/>$data\u00a0=\u00a0$stmt-&gt;fetchAll();<\/pre><p>One peculiar thing about <code>PDO::PARAM_INT<\/code>: for some reason it does not enforce the type casting. Thus, using it on a number that has a string type will cause the aforementioned error: <\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a01\u00a0LIMIT\u00a0?\");<br \/>$stmt-&gt;bindValue(1,\u00a0\"1\",\u00a0PDO::PARAM_INT);<br \/>$stmt-&gt;execute();<\/pre><p>But change <code>\"1\"<\/code> in the example to <code>1<\/code> - and everything will go smooth.<\/p><h3 id=\"transactions\" class=\"title-h3\">Transactions<\/h3><p>To successfully run a transaction, you have to make sure that error mode is set to exceptions, and learn three canonical methods:<\/p><ul><li><code>beginTransaction()<\/code> to start a transaction<\/li>\n<li><code>commit()<\/code> to commit one<\/li>\n<li><code>rollback()<\/code> to cancel all the changes you made since transaction start.<\/li>\n<\/ul><p>Exceptions are essential for transactions because they can be caught. So in case one of the queries failed, the execution will be stopped and moved straight to the catch block, where the whole transaction will be rolled back.<\/p><p>So a typical example would be like<\/p><pre>try\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0$pdo-&gt;beginTransaction();<br \/>\u00a0\u00a0\u00a0\u00a0$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"INSERT\u00a0INTO\u00a0users\u00a0(name)\u00a0VALUES\u00a0(?)\");<br \/>\u00a0\u00a0\u00a0\u00a0foreach\u00a0(['Joe','Ben']\u00a0as\u00a0$name)<br \/>\u00a0\u00a0\u00a0\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0$stmt-&gt;execute([$name]);<br \/>\u00a0\u00a0\u00a0\u00a0}<br \/>\u00a0\u00a0\u00a0\u00a0$pdo-&gt;commit();<br \/>}catch\u00a0(Exception\u00a0$e){<br \/>\u00a0\u00a0\u00a0\u00a0$pdo-&gt;rollback();<br \/>\u00a0\u00a0\u00a0\u00a0throw\u00a0$e;<br \/>}<\/pre><p>Please note the following important things:<\/p><ul><li>PDO error reporting mode should be set to <code>PDO::ERRMODE_EXCEPTION<\/code><\/li>\n<li>you have catch an <code>Exception<\/code>, not <code>PDOException<\/code>, as it doesn't matter what particular exception aborted the execution.<\/li>\n<li>you should re-throw an exception after rollback, to be notified of the problem the usual way.<\/li>\n<li>also make sure that a table engine supports transactions (i.e. for Mysql it should be InnoDB, not MyISAM)<\/li>\n<li>there are no Data definition language (DDL) statements that define or modify database schema among queries in your transaction, as such a query will cause an implicit commit<\/li>\n<\/ul><h3 id=\"call\" class=\"title-h3\">Calling stored procedures in PDO<\/h3><p>There is one thing about stored procedures any programmer stumbles upon at first: every stored procedure always returns <strong>one extra result set<\/strong>: one (or many) results with actual data and one just empty. Which means if you try to call a procedure and then proceed to another query, then <b>\"Cannot execute queries while other unbuffered queries are active\"<\/b> error will occur, because you have to clear that extra empty result first. Thus, after calling a stored procedure that is intended to return only one result set, just call <a href=\"http:\/\/php.net\/manual\/en\/pdostatement.nextrowset.php\"><code>PDOStatement::nextRowset()<\/code><\/a> once (of course after fetching all the returned data from statement, or it will be discarded):<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;query(\"CALL\u00a0bar()\");<br \/>$data\u00a0=\u00a0$stmt-&gt;fetchAll();<br \/>$stmt-&gt;nextRowset();<\/pre><p>While for the stored procedures returning many result sets the behavior will be the same as with <a href=\"#multiquery\">multiple queries execution<\/a>: <\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"CALL\u00a0foo()\");<br \/>$stmt-&gt;execute();<br \/>do\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0$data\u00a0=\u00a0$stmt-&gt;fetchAll();<br \/>\u00a0\u00a0\u00a0\u00a0var_dump($data);<br \/>}\u00a0while\u00a0($stmt-&gt;nextRowset()\u00a0&amp;&amp;\u00a0$stmt-&gt;columnCount());<\/pre><p>However, as you can see here is another trick have to be used: remember that extra result set? It  is so essentially <em>empty<\/em> that even an attempt to fetch from it will produce an error. So, we cannot use just <code>while ($stmt-&gt;nextRowset())<\/code>. Instead, we have to check also for empty result. For which purpose <code>PDOStatement::columnCount()<\/code> is just excellent.<\/p><p>This feature is one of essential differences between old mysql ext and modern libraries: after calling a stored procedure with <code>mysql_query()<\/code> there was no way to continue working with the same connection, because there is no <code>nextResult()<\/code> function for <code>mysql ext<\/code>. One had to close the connection and then open a new one again in order to run other queries after calling a stored procedure. <\/p><p>Calling a stored procedure is a rare case where <code>bindParam()<\/code> use is justified, as it's the only way to handle <code>OUT<\/code> and <code>INOUT<\/code> parameters. The example can be found in the <a href=\"http:\/\/php.net\/manual\/en\/pdo.prepared-statements.php#example-1009\">corresponding manual chapter<\/a>. However, <strong>for mysql it doesn't work.<\/strong> You have to resort to an <a href=\"http:\/\/stackoverflow.com\/a\/23749445\/285587\">SQL variable and an extra call<\/a>.<\/p><p>Note that for the different databases the syntax could be different as well. For example, to run a sored procedure against Microsoft SQL server, use the following format<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"EXEC\u00a0stored_procedure\u00a0?\u00a0?\");<\/pre><p>where ? marks are placeholders. Note that no braces should be used in the call.<\/p><h3 id=\"multiquery\" class=\"title-h3\">Running multiple queries with PDO<\/h3><p>When in <a href=\"#emulation\">emulation mode<\/a>, PDO can run mutiple queries in the same statement, either via query() or <code>prepare()\/execute()<\/code>. To access the result of consequent queries one has to use <a href=\"http:\/\/php.net\/manual\/en\/pdostatement.nextrowset.php\"><code>PDOStatement::nextRowset()<\/code><\/a>:  <\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a0?;SELECT\u00a0?\");<br \/>$stmt-&gt;execute([1,2]);<br \/>do\u00a0{<br \/>\u00a0\u00a0\u00a0\u00a0$data\u00a0=\u00a0$stmt-&gt;fetchAll();<br \/>\u00a0\u00a0\u00a0\u00a0var_dump($data);<br \/>}\u00a0while\u00a0($stmt-&gt;nextRowset());<\/pre><p>Within this loop you'll be able to gather all the related information from the every query, like affected rows, auto-generated id or errors occurred. <\/p><p>It is important to understand that at the point of <code>execute()<\/code> PDO will report the error <strong>for the first query only<\/strong>. But if error occurred at any of consequent queries, to get that error one has to iterate over results. Despite some <a href=\"https:\/\/bugs.php.net\/bug.php?id=61613\">ignorant opinions<\/a>, PDO can not and should not report all the errors at once. Some people just cannot grasp the problem at whole, and don't understand that error message is not the only outcome from the query. There could be a dataset returned, or some metadata like insert id. To get these, one has to iterate over resultsets, one by one. But to be able to throw an error immediately, PDO would have to iterate automatically, and thus <strong>discard some results<\/strong>. Which would be a clear nonsense.<\/p><p>Unlike <code>mysqli_multi_query()<\/code> PDO doesn't make an asynchronous call, so you can't  \"fire and forget\" - send bulk of queries to mysql and close connection, PHP will wait until last query gets executed.<\/p><h3 id=\"emulation\" class=\"title-h3\">Emulation mode. PDO::ATTR_EMULATE_PREPARES<\/h3><p>One of the most controversial PDO configuration options is <code>PDO::ATTR_EMULATE_PREPARES<\/code>. What does it do? PDO can run your queries in two ways: <\/p><ol><li>It can use a <b>real<\/b> or native prepared statement:<br \/>When prepare() is called, your query with placeholders gets sent to mysql as is, with all the question marks you put in (in case named placeholders are used, they are substituted with ?s as well), while actual data goes later, when execute() is called. <\/li>\n<li>It can use <b>emulated<\/b> prepared statement, when your query is sent to mysql as proper SQL, with all the data in place, <b>properly formatted<\/b>. In this case only one roundtrip to database happens, with <code>execute()<\/code> call. For some drivers (including mysql) emulation mode is turned <code>ON<\/code> by default.<\/li>\n<\/ol><p>Both methods has their drawbacks and advantages but, and - I have to stress on it - both being <b>equally secure<\/b>, if used properly. Despite rather appealing tone of the popular <a href=\"http:\/\/stackoverflow.com\/a\/12202218\/285587\">article on Stack Overflow<\/a>, in the end it says that <b>if you are using supported versions of PHP and MySQL properly, you are 100% safe<\/b>.\nAll you have to do is to set encoding in the DSN, as it shown in the <a href=\"#dsn\">example above<\/a>, and your emulated prepared statements will be as secure as real ones. <\/p><p>Note that when native mode is used, <strong>the data is never appears in the query<\/strong>, which is parsed by the engine as is, with all the placeholders in place. If you're looking into Mysql query log for your prepared query, you have to understand that it's just an artificial query that has been created solely for logging purpose, but not a real one that has been executed.<\/p><p>Other issues with emulation mode as follows:<\/p><h4 id=\"emulation_on\" class=\"title-h4\">When emulation mode is turned ON<\/h4><p>one can use a handy feature of named prepared statements - a placeholder with same name could be used any number of times in the same query, while corresponding variable have to be bound only once. For some obscure reason this functionality is disabled when emulation mode is off:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SELECT\u00a0*\u00a0FROM\u00a0t\u00a0WHERE\u00a0foo\u00a0LIKE\u00a0:search\u00a0OR\u00a0bar\u00a0LIKE\u00a0:search\");<br \/>$stmt-&gt;execute(['search']\u00a0=&gt;\u00a0\"%$search%\");`<\/pre><p>Also, when emulation is <code>ON<\/code>, PDO is able to run <a href=\"#multiquery\">multiple queries in one prepared statement<\/a>.<\/p><p>Also, as native prepared statements support only certain query types, you can run some queries with prepared statements only when emulation is <code>ON<\/code>. The following code will return table names in emulation mode and error otherwise:<\/p><pre>$stmt\u00a0=\u00a0$pdo-&gt;prepare(\"SHOW\u00a0TABLES\u00a0LIKE\u00a0?\");<br \/>$stmt-&gt;execute([\"%$name%\"]);<br \/>var_dump($stmt-&gt;fetchAll());<\/pre><h4 id=\"emulation_off\" class=\"title-h4\">When emulation mode is turned OFF<\/h4><p>One could bother not with parameter types, as mysql will sort all the types properly. Thus, even string can be bound to LIMIT parameters, as it was noted in the <a href=\"#limit\">corresponding chapter<\/a>. <\/p><p>Also, this mode will allow to use the advantage of <a href=\"#multiexec\">single prepare-multiple execute<\/a> feature.<\/p><p>It's hard to decide which mode have to be preferred, but for usability sake I would rather turn it <code>OFF<\/code>, to avoid a hassle with <code>LIMIT<\/code> clause. Other issues could be considered negligible in comparison. <\/p><h3 id=\"mysqlnd\" class=\"title-h3\">Mysqlnd and buffered queries. Huge datasets. <\/h3><p>Recently all PHP extensions that work with mysql database were updated based on a low-level library called <code>mysqlnd<\/code>, which replaced old <code>libmysql<\/code> client. Thus some changes in the PDO behavior, mostly described above and one that follows: <\/p><p>There is one thing called <a href=\"http:\/\/php.net\/manual\/en\/mysqlinfo.concepts.buffering.php\">buffered queries<\/a>. Although you probably didn't notice it, you were using them all the way. Unfortunately, here are bad news for you: unlike old PHP versions, where you were using buffered queries virtually for free, modern versions built upon <a href=\"http:\/\/php.net\/manual\/en\/book.mysqlnd.php\">mysqlnd driver<\/a> won't let you to do that anymore:<\/p><blockquote>\n<p>When using libmysqlclient as library PHP's memory limit won't count the memory used for result sets unless the data is fetched into PHP variables. <strong>With mysqlnd the memory accounted for will include the full result set.<\/strong><\/p>\n<\/blockquote><p>The whole thing is about a resultset, which stands for all the data found by the query. <\/p><p>When your SELECT query gets executed, there are two ways to deliver the results in your script: buffered and unbuffered one. When buffered method is used, all the data returned by the query <strong>gets copied in the script's memory<\/strong> at once. While in unbuffered mode a database server feeds the found rows one by one.<\/p><p>So you can tell that in buffered mode a resultset is always burdening up the memory on the server <em>even if fetching weren't started at all<\/em>. Which is why it is not advisable to select huge datasets if you don't need all the data from it. <\/p><p>Nonetheless, when old libmysql-based clients were used, this problem didn't bother PHP uers too much, because the memory consumed by the resultset didn't count in the the <code>memory_get_usage()<\/code> and <code>memory_limit<\/code>. <\/p><p>But with mysqlnd things got changed, and the resultset returned by the buffered query will be count towards both <code>memory_get_usage()<\/code> and <code>memory_limit<\/code>, no matter which way you choose to get the result:<\/p><pre>$pdo-&gt;setAttribute(PDO::MYSQL_ATTR_USE_BUFFERED_QUERY,\u00a0FALSE);<br \/>$stmt\u00a0=\u00a0$pdo-&gt;query(\"SELECT\u00a0*\u00a0FROM\u00a0Board\");<br \/>$mem\u00a0=\u00a0memory_get_usage();<br \/>while($row\u00a0=\u00a0$stmt-&gt;fetch());<br \/>echo\u00a0\"Memory\u00a0used:\u00a0\".round((memory_get_usage()\u00a0-\u00a0$mem)\u00a0\/\u00a01024\u00a0\/\u00a01024,\u00a02).\"M\\n\";$pdo-&gt;setAttribute(PDO::MYSQL_ATTR_USE_BUFFERED_QUERY,\u00a0TRUE);<br \/>$stmt\u00a0=\u00a0$pdo-&gt;query(\"SELECT\u00a0*\u00a0FROM\u00a0Board\");<br \/>$mem\u00a0=\u00a0memory_get_usage();<br \/>while($row\u00a0=\u00a0$stmt-&gt;fetch());<br \/>echo\u00a0\"Memory\u00a0used:\u00a0\".round((memory_get_usage()\u00a0-\u00a0$mem)\u00a0\/\u00a01024\u00a0\/\u00a01024,\u00a02).\"M\\n\";<\/pre><p>will give you (for my data)<\/p><pre>Memory\u00a0used:\u00a00.02M<br \/>Memory\u00a0used:\u00a02.39M<\/pre><p>which means that with buffered query the memory is consumed <strong>even if you're fetching rows one by one!<\/strong> <\/p><p>So, keep in mind that if you are selecting a really huge amount of data, always set <code>PDO::MYSQL_ATTR_USE_BUFFERED_QUERY<\/code> to <code>FALSE<\/code>. <\/p><p>Of course, there are some drawbacks, two minor ones: <\/p><ol><li>With unbuffered query you can't use <code>rowCount()<\/code> method (which is useless, as we learned <a href=\"#count\">above<\/a>) <\/li>\n<li>Moving (seeking) the current resultset internal pointer back and forth (which is useless as well).<\/li>\n<\/ol><p>And a rather important one:<\/p><ol><li>While an unbufferered query is active, you cannot execute any other query using the same connection, so use this mode wisely.<\/li>\n<\/ol><br \/><div><h3 class=\"uppercase mb-1\">Related articles:<\/h3><\/div>","created_at":"2018-11-12T23:43:23+0000","updated_at":"2018-11-12T23:43:23+0000","annotations":[],"mimetype":"text\/html","reading_time":41,"domain_name":"phpdelusions.net","preview_picture":"https:\/\/phpdelusions.net\/img\/phpdelusion-big.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":681,"title":"Micropub - IndieWeb","url":"https:\/\/indieweb.org\/Micropub","content":"<p><b><dfn>Micropub<\/dfn><\/b> is an open API standard (<a href=\"https:\/\/indieweb.org\/W3C\" title=\"W3C\">W3C<\/a> Recommendation) that is used to create, update, and delete posts on one's own domain using third-party clients, and supersedes both <a href=\"https:\/\/indieweb.org\/MetaWeblog\" title=\"MetaWeblog\">MetaWeblog<\/a> and <a href=\"https:\/\/indieweb.org\/AtomPub\" class=\"mw-redirect\" title=\"AtomPub\">AtomPub<\/a>. Web apps and native apps (e.g. iPhone, Android) can use Micropub to post and edit <a href=\"https:\/\/indieweb.org\/article\" title=\"article\">articles<\/a>, short <a href=\"https:\/\/indieweb.org\/note\" title=\"note\">notes<\/a>, <a href=\"https:\/\/indieweb.org\/reply\" title=\"reply\">comments<\/a>, <a href=\"https:\/\/indieweb.org\/like\" title=\"like\">likes<\/a>, <a href=\"https:\/\/indieweb.org\/photo\" title=\"photo\">photos<\/a>, <a href=\"https:\/\/indieweb.org\/event\" title=\"event\">events<\/a>, or other kinds of <a href=\"https:\/\/indieweb.org\/posts\" title=\"posts\">posts<\/a> to your own site.\n<\/p><p>(new!) <b>Submit an implementation report for Micropub to help the specification advance at W3C!<\/b>\n<\/p><ul><li> Start here: <a class=\"external free\" href=\"https:\/\/micropub.rocks\/\">https:\/\/micropub.rocks\/<\/a><\/li><\/ul><p><a href=\"https:\/\/indieweb.org\/File:w3c-spec-REC.svg\" class=\"image\"><img alt=\"w3c-spec-REC.svg\" src=\"https:\/\/indieweb.org\/images\/thumb\/1\/19\/w3c-spec-REC.svg\/696px-w3c-spec-REC.svg.png\" width=\"696\" height=\"117\" srcset=\"\/images\/thumb\/1\/19\/w3c-spec-REC.svg\/1044px-w3c-spec-REC.svg.png 1.5x, \/images\/thumb\/1\/19\/w3c-spec-REC.svg\/1392px-w3c-spec-REC.svg.png 2x\" \/><\/a>\n<\/p><div id=\"toc\" class=\"toc\"><p><\/p><h2>Contents<\/h2><ul><li class=\"toclevel-1 tocsection-1\"><a href=\"#Background\">1 Background<\/a><\/li>\n<li class=\"toclevel-1 tocsection-2\"><a href=\"#Syntax\">2 Syntax<\/a>\n<ul><li class=\"toclevel-2 tocsection-3\"><a href=\"#Overview\">2.1 Overview<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-4\"><a href=\"#How_to_implement\">3 How to implement<\/a>\n<ul><li class=\"toclevel-2 tocsection-5\"><a href=\"#Endpoint_Discovery\">3.1 Endpoint Discovery<\/a>\n<ul><li class=\"toclevel-3 tocsection-6\"><a href=\"#HTTP_Header\">3.1.1 HTTP Header<\/a><\/li>\n<li class=\"toclevel-3 tocsection-7\"><a href=\"#HTML_Head\">3.1.2 HTML Head<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-8\"><a href=\"#Authentication\">3.2 Authentication<\/a>\n<ul><li class=\"toclevel-3 tocsection-9\"><a href=\"#HTTP_Header_2\">3.2.1 HTTP Header<\/a><\/li>\n<li class=\"toclevel-3 tocsection-10\"><a href=\"#Form-Encoded_Body_Parameter\">3.2.2 Form-Encoded Body Parameter<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-11\"><a href=\"#Scope\">3.3 Scope<\/a><\/li>\n<li class=\"toclevel-2 tocsection-12\"><a href=\"#Verification\">3.4 Verification<\/a><\/li>\n<li class=\"toclevel-2 tocsection-13\"><a href=\"#Form-encoded_Microformats_Representation\">3.5 Form-encoded Microformats Representation<\/a><\/li>\n<li class=\"toclevel-2 tocsection-14\"><a href=\"#Examples_of_Creating_Objects\">3.6 Examples of Creating Objects<\/a>\n<ul><li class=\"toclevel-3 tocsection-15\"><a href=\"#Indicating_the_object_being_created\">3.6.1 Indicating the object being created<\/a><\/li>\n<li class=\"toclevel-3 tocsection-16\"><a href=\"#h-entry\">3.6.2 h-entry<\/a>\n<ul><li class=\"toclevel-4 tocsection-17\"><a href=\"#New_Note\">3.6.2.1 New Note<\/a>\n<ul><li class=\"toclevel-5 tocsection-18\"><a href=\"#Minimal_Example\">3.6.2.1.1 Minimal Example<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-4 tocsection-19\"><a href=\"#New_Reply\">3.6.2.2 New Reply<\/a><\/li>\n<li class=\"toclevel-4 tocsection-20\"><a href=\"#New_Repost\">3.6.2.3 New Repost<\/a><\/li>\n<li class=\"toclevel-4 tocsection-21\"><a href=\"#New_Article\">3.6.2.4 New Article<\/a><\/li>\n<li class=\"toclevel-4 tocsection-22\"><a href=\"#New_Bookmark\">3.6.2.5 New Bookmark<\/a><\/li>\n<li class=\"toclevel-4 tocsection-23\"><a href=\"#Adding_Files\">3.6.2.6 Adding Files<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-3 tocsection-24\"><a href=\"#h-event\">3.6.3 h-event<\/a><\/li>\n<li class=\"toclevel-3 tocsection-25\"><a href=\"#h-cite\">3.6.4 h-cite<\/a><\/li>\n<li class=\"toclevel-3 tocsection-26\"><a href=\"#Nested_Microformat_Objects\">3.6.5 Nested Microformat Objects<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-27\"><a href=\"#Servers\">4 Servers<\/a><\/li>\n<li class=\"toclevel-1 tocsection-28\"><a href=\"#Clients\">5 Clients<\/a><\/li>\n<li class=\"toclevel-1 tocsection-29\"><a href=\"#Libraries\">6 Libraries<\/a><\/li>\n<li class=\"toclevel-1 tocsection-30\"><a href=\"#Issues\">7 Issues<\/a><\/li>\n<li class=\"toclevel-1 tocsection-31\"><a href=\"#See_Also\">8 See Also<\/a><\/li>\n<\/ul><\/div><h2>Background<\/h2><p>The Micropub vocabulary is derived directly from the Microformats vocabulary. Micropub is meant to be a serialization of Microformats that can be submitted as an HTTP POST. The method for developing new Micropub vocabularies is to look at the Microformats representation and work backwards.\n<\/p><h2>Syntax<\/h2><p>Similar to how <a href=\"https:\/\/indieweb.org\/microformats\" title=\"microformats\">microformats<\/a> has a relatively small ruleset for parsing HTML documents into a data structure, Micropub defines a small set of rules to interpret HTTP POST and GET requests as Micropub commands. Where Microformats does not require changing the parsing rules to introduce new properties of an object such as an h-entry, Micropub similarly does not require changing parsing rules to interpret requests that may correspond to different post types, such as posting videos vs \"likes\".\n<\/p><p>The Micropub syntax describes how to interpret HTTP POST and GET requests into a useful action the server can take.\n<\/p><p>The full specification can be read at <a class=\"external free\" href=\"https:\/\/www.w3.org\/TR\/micropub\/\">https:\/\/www.w3.org\/TR\/micropub\/<\/a>\n<\/p><h3>Overview<\/h3><p>All Micropub requests to create posts are sent as form-encoded<a class=\"external autonumber\" href=\"http:\/\/www.w3.org\/TR\/html5\/forms.html#url-encoded-form-data\">[1]<\/a>, multipart form-data <a class=\"external autonumber\" href=\"http:\/\/www.w3.org\/TR\/html5\/forms.html#multipart-form-data\">[2]<\/a>, or JSON encoded HTTP requests. Responses typically do not include a response body, indicating the needed information (such as the URL of the created post) in HTTP headers. When a response body is required, it is returned as either form-encoded or JSON<a class=\"external autonumber\" href=\"https:\/\/tools.ietf.org\/html\/rfc7159\">[3]<\/a>, depending on the HTTP Accept<a class=\"external autonumber\" href=\"http:\/\/tools.ietf.org\/html\/rfc7231#section-5.3.2\">[4]<\/a> header.\n<\/p><h2>How to implement<\/h2><p>How to implement the Micropub API, both in a client that can discover an endpoint and publish to it, and on a server to support an endpoint to create\/update\/delete posts in response.\n<\/p><h3>Endpoint Discovery<\/h3><p>It should be possible to configure an API client by authenticating as your domain name using IndieAuth. After signing in, your domain needs a way to specify the API endpoint the client will use to create new posts.\n<\/p><p>Add a &lt;link&gt; tag in the HTML head of your home page, or send an HTTP Link header.\n<\/p><h4>HTTP Header<\/h4><pre>Link: &lt;https:\/\/example.com\/micropub&gt;; rel=\"micropub\"<\/pre><h4>HTML Head<\/h4><pre>&lt;link rel=\"micropub\" href=\"https:\/\/example.com\/micropub\"&gt;<\/pre><h3>Authentication<\/h3><p>Micropub requests are authenticated by including a Bearer Token in either the HTTP header or a form-encoded body parameter as described in the <a class=\"external text\" href=\"http:\/\/tools.ietf.org\/html\/rfc6750#section-2\">OAuth Bearer Token RFC<\/a>.\n<\/p><p>Authorization should be handled via the IndieAuth protocol, built on top of OAuth 2.0. See <a href=\"https:\/\/indieweb.org\/obtaining-an-access-token\" title=\"obtaining-an-access-token\">obtaining-an-access-token<\/a> for more details. An app that wants to post to a user's Micropub endpoint will need to obtain authorization from the user in order to get an access token.\n<\/p><h4>HTTP Header<\/h4><pre>\nAuthorization: Bearer XXXXXXXX\n<\/pre><h4>Form-Encoded Body Parameter<\/h4><pre>\naccess_token=XXXXXXXXX\n<\/pre><h3>Scope<\/h3><p>The client may request one or more scopes during the authorization request. It does this according to standard OAuth 2.0 techniques, by passing a space-separated list of scope names in the authorization request.\n<\/p><p>The authorization server must indicate to the user any scopes that are part of the request, whether or not the authorization server recognizes the scopes. The authorization server may also allow the user to add or remove scopes that the client requests.\n<\/p><p>For example, most Micropub servers require clients to obtain the \"create\" scope in order to create posts. (The legacy \"post\" scope previously encompassed both creating and updating posts.) However, some servers may require more granular scope requests, such as \"delete\" or \"create:video\". See <a href=\"https:\/\/indieweb.org\/scope\" title=\"scope\">scope<\/a> for more details and a list of all currently used values for scope.\n<\/p><h3>Verification<\/h3><p>A micropub client can verify that an access token is still valid (and other details) by querying the user's <a href=\"https:\/\/indieweb.org\/token_endpoint\" class=\"mw-redirect\" title=\"token endpoint\">token endpoint<\/a>. \n<\/p><p>Issues:\n<\/p><ul><li> this does not cover the case where the server has decided not to accept it anymore, e.g. because the user's access to the site has been removed\n<ul><li> if a micropub endpoint plans on rejecting a token in the future, it should ensure the token endpoint also rejects the token. For systems where the token endpoint and micropub endpoint are part of the same software, this is an implementation detail. If the token endpoint and micropub endpoint are on different systems, there needs to be a mechanism for the micropub endpoint to expire the token at the token endpoint. This is what's known as token revocation, and has been written up for OAuth 2: <a class=\"external free\" href=\"https:\/\/tools.ietf.org\/html\/rfc7009\">https:\/\/tools.ietf.org\/html\/rfc7009<\/a><\/li><\/ul><\/li><\/ul><h3>Form-encoded Microformats Representation<\/h3><p>For the simplicity of writing clients and servers, all Micropub endpoints must handle the standard form-encoded format. At a most basic level, you should be able to write an HTML form and set the form action to your own endpoint and use it to post to your site.\n<\/p><h3>Examples of Creating Objects<\/h3><p>How to create some of the common <a href=\"https:\/\/indieweb.org\/Category:PostType\" title=\"Category:PostType\">post types<\/a>\n<\/p><h4>Indicating the object being created<\/h4><p>To indicate the object being created, use a property called \"h\", (which would never be the name of a property of a microformats object), followed by the name of the microformats object. Examples:\n<\/p><ul><li> h=entry<\/li>\n<li> h=card<\/li>\n<li> h=event<\/li>\n<li> h=cite<\/li><\/ul><h4>h-entry<\/h4><p>The following properties may be included in a request to create a new <a href=\"https:\/\/indieweb.org\/h-entry\" title=\"h-entry\">h-entry<\/a>:\n<\/p><ul><li> name<\/li>\n<li> summary<\/li>\n<li> content<\/li>\n<li> published<\/li>\n<li> updated<\/li>\n<li> category = tag1, tag2, tag3 (sent as array syntax: <code>&amp;category[]=tag1&amp;category[]=tag2<\/code>)<\/li>\n<li> location\n<ul><li> as a <a class=\"external text\" href=\"http:\/\/tools.ietf.org\/html\/rfc5870\">Geo URI<\/a>, for example <code><a class=\"external free\" href=\"denied:geo:45.51533714,-122.646538633\">geo:45.51533714,-122.646538633<\/a><\/code>\n<ul><li> GeoURIs allow <a class=\"external text\" href=\"https:\/\/tools.ietf.org\/html\/rfc5870#section-6.4\">additional parameters<\/a> that may be used to include the title of a location: <code><a class=\"external free\" href=\"denied:geo:1,2;name=Somewhere\">geo:1,2;name=Somewhere<\/a><\/code><\/li><\/ul><\/li>\n<li> as a URL that contains an <a href=\"https:\/\/indieweb.org\/h-card\" title=\"h-card\">h-card<\/a><\/li>\n<li> My micropub client currently has a map to mark the specific location in LatLng values, and a text input for an actual place name. So my client is sending a request of the form \"&amp;location=1.23,%40-4.56:An%40Address\" <a href=\"https:\/\/indieweb.org\/User:Jonnybarnes.net\" class=\"mw-redirect\" title=\"User:Jonnybarnes.net\">Jonnybarnes.net<\/a> 07:57, 10 June 2014 (PDT)<\/li><\/ul><\/li>\n<li> in-reply-to<\/li>\n<li> repost-of<\/li>\n<li> syndication\n<ul><li> Pass one or more URLs pointing to places where this entry already exists. Can be used for <a href=\"https:\/\/indieweb.org\/PESOS\" title=\"PESOS\">PESOS<\/a> implementations.<\/li><\/ul><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/mp-syndicate-to\" title=\"mp-syndicate-to\">mp-syndicate-to<\/a> = https:\/\/myfavoritesocialnetwork.example\/aaronpk, https:\/\/archive.org\/, etc.\n<ul><li> This property is slightly different from the others since it is giving a command to the server rather than describing an object, which is why it is prefixed with \"mp-\".<\/li><\/ul><\/li><\/ul><h5>New Note<\/h5><p>Posting a new <a href=\"https:\/\/indieweb.org\/note\" title=\"note\">note<\/a> with tags, syndicating to myfavoritesocialnetwork:\n<\/p><ul><li> content<\/li>\n<li> category<\/li>\n<li> mp-syndicate-to<\/li>\n<li> published (optional, defaults to \"now\" if not present. Useful for writing offline and syncing later)<\/li><\/ul><pre>\nPOST \/micropub HTTP\/1.1\nHost: aaronparecki.com\nContent-type: application\/x-www-form-urlencoded\nh=entry\n&amp;content=The+%40Jawbone+UP%2C+my+favorite+of+the+%23quantifiedself+trackers%2C+finally+released+their+official+API%21+http%3A%2F%2Fjawbone.com%2Fup%2Fdeveloper%2F\n&amp;category[]=jawbone&amp;category[]=quantifiedself&amp;category[]=api\n&amp;mp-syndicate-to=https:\/\/myfavoritesocialnetwork.example\/aaronpk\n<\/pre><h6>Minimal Example<\/h6><pre>\nPOST \/micropub HTTP\/1.1\nHost: example.com\nContent-type: application\/x-www-form-urlencoded\nAuthorization: Bearer XXXXXXX\nh=entry\n&amp;content=Hello+World\n<\/pre><pre>\ncurl https:\/\/example.com\/micropub -d h=entry -d \"content=Hello World\" -H \"Authorization: Bearer XXXXXXX\"\n<\/pre><h5>New Reply<\/h5><p>Posting a new <a href=\"https:\/\/indieweb.org\/reply\" title=\"reply\">reply<\/a>, syndicating to myfavoritesocialnetwork\n<\/p><ul><li> content<\/li>\n<li> in-reply-to<\/li>\n<li> mp-syndicate-to<\/li>\n<li> published<\/li><\/ul><pre>\nPOST \/post\/new HTTP\/1.1\nHost: aaronparecki.com\nContent-type: application\/x-www-form-urlencoded\nh=entry\n&amp;content=%40BarnabyWalters+My+favorite+for+that+use+case+is+Redis.+It%27s+easy+to+set+up+and+use%2C+I+often+use+it+to+move+data+between+apps+written+in+different+languages+too.\n&amp;in-reply-to=http:\/\/waterpigs.co.uk\/notes\/4S0LMw\/\n&amp;mp-syndicate-to=https:\/\/myfavoritesocialnetwork.example\/aaronpk\n<\/pre><h5>New Repost<\/h5><p>Posting a new <a href=\"https:\/\/indieweb.org\/repost\" title=\"repost\">repost<\/a>, and adding additional tags.\n<\/p><ul><li> repost-of<\/li>\n<li> category<\/li><\/ul><pre>\nPOST \/micropub HTTP\/1.1\nHost: aaronparecki.com\nContent-type: application\/x-www-form-urlencoded\nh=entry\n&amp;repost-of=http:\/\/waterpigs.co.uk\/notes\/4S0LMw\/\n&amp;category=realtime\n<\/pre><h5>New Article<\/h5><p>Posting a new <a href=\"https:\/\/indieweb.org\/article\" title=\"article\">article<\/a>\n<\/p><ul><li> content<\/li>\n<li> name<\/li>\n<li> category<\/li>\n<li> published<\/li><\/ul><pre>\nPOST \/micropub HTTP\/1.1\nHost: aaronparecki.com\nContent-type: application\/x-www-form-urlencoded\nh=entry\n&amp;content=Now+that+I%27ve+been+%5Bhttp%3A%2F%2Faaronparecki.com%2Fevents+creating+a+list+of+events%5D+on+my+site+using+%5Bhttp%3A%2F%2Findiewebcamp.com%2Fp3k+p3k%5D%2C+it+would+be+great+if+I+could+get+a+more+calendar-like+view+of+that+list.+%0A%0ASince+I+live+in+Google+Calendar+every+day+anyway%2C+it+would+be+great+to+use+that+interface+to+browse+my+%23indieweb+events+as+well%21+Since+my+events+page+is+marked+up+with+%5Bhttp%3A%2F%2Fmicroformats.org%2Fwiki%2Fh-event+h-event+microformats%5D%2C+all+it+would+take+is+to+write+an+h-event+to+iCal+converter+script+to+generate+an+iCal+feed+from+my+list+of+events.+Then+I+could+just+subscribe+to+the+iCal+feed+from+within+Google+Calendar.%0A%0A%23%23%23+Bonus%3A+read%2Fwrite+access+to+indieweb+events+via+Google+Calendar%0A%0AEven+better+would+be+to+use+Google+Calendar+to+also+create+events+on+my+site.+Unfortunately+Google+Calendar+doesn%27t+support+CalDAV%2C+so+we+can%27t+do+it+that+way.+%28Of+course+I+could+use+Apple%27s+iCal+to+publish+directly%2C+but+that+also+means+I%27d+have+to+write+some+code+tot+speak+CalDAV%29.+%0A%0AInstead%2C+I+can+create+a+%22write-only%22+calendar+in+Google+Calendar%2C+and+have+p3k+subscribe+to+it.+Any+new+events+in+that+feed+would+be+moved+over+to+the+internal+events+page+and+deleted+from+the+Google+Calendar.\n&amp;name=Itching%3A+h-event+to+iCal+converter\n&amp;category[]=indieweb&amp;category[]=hevent&amp;category[]=events&amp;category[]=calendar&amp;category[]=p3k\n<\/pre><h5>New Bookmark<\/h5><p>Posting a new <a href=\"https:\/\/indieweb.org\/bookmark\" title=\"bookmark\">bookmark<\/a> with name, quote, and tags.\n<\/p><ul><li> bookmark-of<\/li>\n<li> name<\/li>\n<li> content<\/li>\n<li> category<\/li><\/ul><pre>\nPOST \/micropub HTTP\/1.1\nHost: aaronparecki.com\nContent-type: application\/x-www-form-urlencoded\nh=entry\n&amp;bookmark-of=https%3A%2F%2Fplus.google.com%2F%2BKartikPrabhu%2Fposts%2FUzKErSbfmHq\n&amp;name=To+everyone+who+is+complaining+about+Popular+Science+shutting+down+comments...\n&amp;content=%22Why+is+there+this+expectation+that+every+website+should+be+a+forum%3F+No+website+has+any+obligation+to+provide+a+space+for+your+rants.+Use+your+own+space+on+the+web+to+do+that.%22\n&amp;category[]=indieweb&amp;category[]=comments\n<\/pre><h5>Adding Files<\/h5><p>When a Micropub request includes a file, the entire request is sent in the <code>multipart\/form-data<\/code> encoding, and the file is named by content type, either \"audio\", \"video\" or \"photo\". A request may include one or more of these files.\n<\/p><p>When <a href=\"https:\/\/indieweb.org\/OwnYourGram\" title=\"OwnYourGram\">OwnYourGram<\/a> makes a Micropub request to post a video, it also sends a photo which is a thumbnail preview of the video.\n<\/p><p>In PHP, these files are accessible using the <code>$_FILES<\/code> array:\n<\/p><pre>$_FILES['video']\n$_FILES['photo']\n$_FILES['audio']\n<\/pre><p>see also: <a href=\"https:\/\/indieweb.org\/Micropub-brainstorming#Uploading_files_in_separate_requests\" title=\"Micropub-brainstorming\">Micropub-brainstorming#Uploading_files_in_separate_requests<\/a>\n<\/p><p>Note that there is no practical way to upload a file when the request body is JSON encoded. The <a href=\"https:\/\/indieweb.org\/media-endpoint\" class=\"mw-redirect\" title=\"media-endpoint\">media endpoint<\/a> needs to be used in that case.\n<\/p><p>curl example: \n<\/p><pre>$ curl -H 'Authorization: Bearer abc'\\\n       -F 'h=entry'\\\n       -F 'content=an image!'\\\n       -F 'photo=@\/home\/cweiske\/Bilder\/openid-id.bogo.png'\\\n       '<a class=\"external free\" href=\"http:\/\/known.bogo\/micropub\/endpoint'\">http:\/\/known.bogo\/micropub\/endpoint'<\/a>\n<\/pre><h4>h-event<\/h4><p>The following properties may be included in a request to create a new <a href=\"https:\/\/indieweb.org\/h-event\" title=\"h-event\">h-event<\/a>:\n<\/p><p>Posting a new event\n<\/p><ul><li> name<\/li>\n<li> summary<\/li>\n<li> description<\/li>\n<li> start<\/li>\n<li> end<\/li>\n<li> duration<\/li>\n<li> category<\/li>\n<li> location<\/li><\/ul><pre>\nPOST \/micropub HTTP\/1.1\nHost: aaronparecki.com\nContent-type: application\/x-www-form-urlencoded\nh=event\n&amp;name=IndieWeb Dinner at 21st Amendment\n&amp;description=In SF Monday evening? Join @caseorganic and I for an #indieweb dinner at 6pm! (Sorry for the short notice!)\n&amp;start=2013-09-30T18:00:00-07:00\n&amp;category=indieweb\n&amp;location=http:\/\/21st-amendment.com\/\n<\/pre><h4>h-cite<\/h4><p>The following properties may be included in a request to create a new <a href=\"https:\/\/indieweb.org\/h-cite\" title=\"h-cite\">h-cite<\/a>:\n<\/p><p>(The following list is from <a class=\"external text\" href=\"http:\/\/microformats.org\/wiki\/h-cite#Properties\">microformats.org\/wiki\/h-cite<\/a>)\n<\/p><ul><li> name<\/li>\n<li> published - date of publication of the work (not the date the h-cite was created)<\/li>\n<li> author - URL of an h-card<\/li>\n<li> url - a URL to access the cited work<\/li>\n<li> content - the content or partial content of the work itself, such as when including a blockquote snippet of a work<\/li><\/ul><h4>Nested Microformat Objects<\/h4><p>Whenever possible, nested Microformats objects should be avoided. A better alternative is to reference objects by their URLs. The most common example is including an h-card for a venue, such as checking in to a location or tagging a photo with a person or location. In these cases, it is better to reference the object by URL, creating it first if necessary.\n<\/p><p>For simple values such as an h-geo property with latitude and longitude, just use a <a class=\"external text\" href=\"http:\/\/en.wikipedia.org\/wiki\/Geo_URI\">Geo URI<\/a> such as <code><a class=\"external free\" href=\"denied:geo:37.786971,-122.399677\">geo:37.786971,-122.399677<\/a><\/code>\n<\/p><p>This technique has the advantage of ensuring that each object that is created has its own URL (each piece of data has its own link). This also gives the server an opportunity to handle each entity separately. E.g., rather than creating a duplicate of an existing venue, it may give back a link to one that was already created, possibly even merging in newly received data first.\n<\/p><p>For more complicated objects, it is better to first create an object on the target site so that it has its own URL, then reference that object's URL in the main request.\n<\/p><p>For example, creating a checkin post would involve two POST requests:\n<\/p><p>First create the venue by posting an h-card:\n<\/p><pre>\nPOST \/micropub\nh=card\n&amp;name=Ford+Food+and+Drink\n&amp;url=http:\/\/www.fordfoodanddrink.com\/\n&amp;street-address=2505 SE 11th Ave\n&amp;locality=Portland\n&amp;region=OR\n&amp;postal-code=97214\n&amp;geo=geo:45.5048473,-122.6549551\n&amp;tel=(503) 236-3023\n<\/pre><p>Response:\n<\/p><pre>\nHTTP\/1.1 201 Created\nLocation: http:\/\/example.com\/venue\/10\n<\/pre><p>Then create the checkin post:\n<\/p><pre>\nPOST \/micropub\nh=entry\n&amp;location=http:\/\/example.com\/venue\/10\n&amp;name=Working on Micropub\n&amp;category=indieweb\n<\/pre><p>Response:\n<\/p><pre>\nHTTP\/1.1 201 Created\nLocation: http:\/\/example.com\/entry\/1001\n<\/pre><ul><li> This technique has the advantage of ensuring that each object that is created has its own URL (each piece of data has its own link)<\/li>\n<li> Also gives the server an opportunity to handle each entity separately. E.g., rather than creating a duplicate of an existing venue, it may give back a link to one that was already created, possibly even merging in newly received data first.<\/li><\/ul><h2>Servers<\/h2><p>Main article: <a href=\"https:\/\/indieweb.org\/Micropub\/Servers\" title=\"Micropub\/Servers\">Micropub\/Servers<\/a><\/p><p>Examples of Micropub endpoints are listed at <a href=\"https:\/\/indieweb.org\/Micropub\/Servers\" title=\"Micropub\/Servers\">Micropub\/Servers<\/a>.\n<\/p><h2>Clients<\/h2><p>Main article: <a href=\"https:\/\/indieweb.org\/Micropub\/Clients\" title=\"Micropub\/Clients\">Micropub\/Clients<\/a><\/p><p>Click through to the main article for a list of sites and client applications that publish via Micropub as well as a table indicating implementation statuses. \n<\/p><h2>Libraries<\/h2><p>\nOpen source libraries &amp; implementations used to support micropub on the client app side and on the API endpoint side on the server:\n<\/p><ul><li> <a class=\"external text\" href=\"https:\/\/github.com\/aaronpk\/p3k-micropub\">p3k-micropub<\/a> <b>PHP<\/b> library used in <a href=\"https:\/\/indieweb.org\/p3k\" title=\"p3k\">p3k<\/a><\/li>\n<li> <a class=\"external text\" href=\"https:\/\/github.com\/taproot\/authentication\">taproot\/authentication<\/a> <b>PHP<\/b> library used in <a href=\"https:\/\/indieweb.org\/Taproot\" title=\"Taproot\">Taproot<\/a>, implements both client app and resource provider authentication<\/li>\n<li> <a class=\"external text\" href=\"https:\/\/github.com\/snarfed\/wordpress-micropub\">wordpress-micropub<\/a>: <a href=\"https:\/\/indieweb.org\/WordPress\" title=\"WordPress\">WordPress<\/a> plugin for server endpoint<\/li>\n<li> <a class=\"external text\" href=\"https:\/\/github.com\/kylewm\/flask-micropub\">flask-micropub<\/a>: <a href=\"https:\/\/indieweb.org\/Flask\" title=\"Flask\">Flask<\/a> (<a href=\"https:\/\/indieweb.org\/Python\" title=\"Python\">Python<\/a>) plugin that makes it easy to add IndieAuth authentication and Micropub authorization to a client application.<\/li>\n<li> <a class=\"external text\" href=\"https:\/\/github.com\/voxpelli\/node-micropub-express\">micropub-express<\/a> <b>Node.js<\/b> server library created and used by <img src=\"http:\/\/voxpelli.com\/avatar.jpg\" class=\"u-photo\" alt=\"\" \/><a href=\"https:\/\/indieweb.org\/User:Kodfabrik.se\" title=\"User:Kodfabrik.se\">Pelle Wessman<\/a> for voxpelli.com's micropub endpoint<\/li>\n<li> <a class=\"external text\" href=\"https:\/\/www.npmjs.com\/package\/micropub-helper\">micropub-helper<\/a> <b>Node.js<\/b> client library created and used by <a href=\"https:\/\/indieweb.org\/grant.codes\" class=\"mw-redirect\" title=\"grant.codes\">grant.codes<\/a><\/li>\n<li> <a class=\"external text\" href=\"https:\/\/www.drupal.org\/project\/indieweb\">IndieWeb for Drupal 8<\/a> A module which supports setting up a micropub server endpoint, written by <img src=\"https:\/\/realize.be\/vue\/images\/swentel-200.png\" class=\"u-photo\" alt=\"\" \/><a href=\"https:\/\/indieweb.org\/User:Realize.be\" title=\"User:Realize.be\">Kristof De Jaeger<\/a><\/li>\n<li> <a class=\"external text\" href=\"https:\/\/github.com\/am1t\/blotpub\">blotpub<\/a> <b>Node.js<\/b> An endpoint that accepts Micropub requests, creates a simple Blot posts and saves them to a configured Dropbox folder. This enables updating a <a class=\"external text\" href=\"https:\/\/blot.im\/\">Blot<\/a> blog through a Micropub client.<\/li><\/ul><h2>Issues<\/h2><p>Please use the <a class=\"external text\" href=\"https:\/\/github.com\/w3c\/Micropub\/issues\">Github repo<\/a> for issues and discussion related to the spec.\n<\/p><p>For issues that have been discussed and are ready to consider for inclusion in the spec, use the <a href=\"https:\/\/indieweb.org\/Micropub-brainstorming\" title=\"Micropub-brainstorming\">Micropub-brainstorming<\/a> page to collect examples and iterate on the ideas.\n<\/p><h2>See Also<\/h2><ul><li> <a class=\"external text\" href=\"http:\/\/micropub.net\/draft\/\">Latest Editor's Draft<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/Micropub-brainstorming\" title=\"Micropub-brainstorming\">Micropub-brainstorming<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/Micropub-extensions\" title=\"Micropub-extensions\">Micropub-extensions<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/post\" title=\"post\">post<\/a> - <a href=\"https:\/\/indieweb.org\/h-entry\" title=\"h-entry\">h-entry<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/event\" title=\"event\">event<\/a> - <a href=\"https:\/\/indieweb.org\/h-event\" title=\"h-event\">h-event<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/obtaining-an-access-token\" title=\"obtaining-an-access-token\">obtaining-an-access-token<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/token-endpoint\" title=\"token-endpoint\">token-endpoint<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/micropub-endpoint\" title=\"micropub-endpoint\">micropub-endpoint<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/access_token\" title=\"access token\">access_token<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/incremental_authorization\" title=\"incremental authorization\">incremental authorization<\/a><\/li>\n<li> <a href=\"https:\/\/indieweb.org\/micropub_media_endpoint\" title=\"micropub media endpoint\">micropub media endpoint<\/a><\/li><\/ul>","created_at":"2018-11-12T18:32:11+0000","updated_at":"2018-11-12T18:32:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"indieweb.org","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":[],"is_public":false,"id":680,"title":"PHP Warning: Missing argument 2 for wpdb::prepare()","url":"https:\/\/make.wordpress.org\/core\/2012\/12\/12\/php-warning-missing-argument-2-for-wpdb-prepare\/","content":"<p>Hello plugin or theme author! You possibly found this post after searching the Internet for the error above: \u201cPHP Warning: Missing argument 2 for wpdb::prepare().\u201d<\/p>\n<p>So, this is a new warning in 3.5. <strong>No sites are broken,<\/strong> everything is fine as before. But, this is indeed something you need to look at, because you may be exposing your users to a possible SQL injection vulnerability. Now that\u2019s no fun!<\/p>\n<p>First, if you\u2019re a user and you want to get rid of these errors, you should turn off the displaying of errors in PHP. There are many ways to do this, such as in php.ini, .htaccess, etc. For this, you can just put this in wp-config.php. (Note that hiding errors on production sites is good practice anyway.)<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n@ini_set('display_errors', 0);\n<\/pre>\n<p>If you\u2019re a user, you can stop here. <strong>(If you need more help, please don\u2019t comment here, try the helpful <a href=\"https:\/\/wordpress.org\/support\/\">Support Forums<\/a>.)<\/strong> Just be sure to send a link to this post to the developer of the theme or plugin referenced in the error.<\/p>\n<p>Now, developers: Here\u2019s how $wpdb-&gt;prepare() is supposed to work:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n$wpdb-&gt;prepare( \"SELECT * FROM table WHERE ID = %d AND name = %s\", $id, $name );\n<\/pre>\n<p>See how $id \u2014 an integer, presumably \u2014 was passed as the second argument? That corresponds to the first placeholder, %d. Then, $name (a string) was passed as the third argument, thus the second placeholder, %s. This makes sure your query is safe, and prevents something like <a href=\"http:\/\/xkcd.com\/327\/\">little bobby tables<\/a>. (Note: the comic is wrong, don\u2019t sanitize \u2014 always prepare your queries.)<\/p>\n<p>The problem is, a number of people were calling $wpdb-&gt;prepare() with only one argument, like so:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n$wpdb-&gt;prepare( \"SELECT COUNT(*) FROM table\" );\n<\/pre>\n<p>See, there\u2019s no parameter (%d, %s, or for floats, %f) in this query. This happens to work fine, but the prepare call isn\u2019t doing anything. You should instead the query directly, as there are no inputs.<\/p>\n<p>But here\u2019s where the problem lies:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n$wpdb-&gt;prepare( \"SELECT * FROM table WHERE id = $id\" );\n<\/pre>\n<p>See the problem? That query isn\u2019t secure! You may think you are \u201cpreparing\u201d this query, but you\u2019re not \u2014 you\u2019re passing $id directly into the query, unprepared. And <strong>this<\/strong>, right here, is why $wpdb-&gt;prepare() now issues a warning if it isn\u2019t called with more than one argument. Because you can\u2019t prepare a query without more than one argument. Here\u2019s a correct example:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n$wpdb-&gt;prepare( \"SELECT * FROM table WHERE id = %d\", $id );\n<\/pre>\n<p>This wasn\u2019t a decision done lightly. We don\u2019t like shoving PHP warnings into the faces of users and developers. But given the potential security risks, we wanted everyone to immediately look at how they are running queries. And, of course, always prepare them properly.<\/p>\n<p>For more: <a href=\"https:\/\/codex.wordpress.org\/Class_Reference\/wpdb\">wpdb Codex reference<\/a>, <a href=\"https:\/\/core.trac.wordpress.org\/ticket\/22262\">#22262<\/a>, and <a href=\"https:\/\/core.trac.wordpress.org\/changeset\/22429\">[22429]<\/a>.<\/p>\n<p class=\"o2-appended-tags\"><a href=\"https:\/\/make.wordpress.org\/core\/tag\/3-5\/\" class=\"tag\">#3-5<\/a>, <a href=\"https:\/\/make.wordpress.org\/core\/tag\/dev-notes\/\" class=\"tag\">#dev-notes<\/a>, <a href=\"https:\/\/make.wordpress.org\/core\/tag\/sql\/\" class=\"tag\">#sql<\/a>, <a href=\"https:\/\/make.wordpress.org\/core\/tag\/wpdb\/\" class=\"tag\">#wpdb<\/a><\/p>","created_at":"2018-11-12T03:53:03+0000","updated_at":"2018-11-12T03:53:03+0000","published_at":"2012-12-12T06:14:32+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"make.wordpress.org","preview_picture":"https:\/\/make.wordpress.org\/core\/files\/2015\/11\/wordpress-logo-notext-rgb.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["mastodon"],"is_public":false,"id":679,"title":"Node.js uWebSockets binding is not maintained and its source is not available \u00b7 Issue #6853 \u00b7 tootsuite\/mastodon","url":"https:\/\/github.com\/tootsuite\/mastodon\/issues\/6853","content":"Node.js uWebSockets binding is not maintained and its source is not available \u00b7 Issue #6853 \u00b7 tootsuite\/mastodon \u00b7 GitHub\n    \n  \n  \n    \n  \n        \n  <div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n    \n  \n  \n  <details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark\" open=\"open\"><summary aria-haspopup=\"dialog\" aria-label=\"Close dialog\">\n      \n    <\/summary><\/details>\n<p>Press h to open a hovercard with more details.<\/p>","created_at":"2018-11-11T15:48:56+0000","updated_at":"2018-11-13T20:14:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/17036990?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["json-ld","activitypub","java"],"is_public":false,"id":678,"title":"jsonld-java\/jsonld-java","url":"https:\/\/github.com\/jsonld-java\/jsonld-java","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><strong>JSONLD-Java is looking for a maintainer<\/strong><\/p>\n<p>This is a Java implementation of the <a href=\"https:\/\/www.w3.org\/TR\/2014\/REC-json-ld-20140116\/\" rel=\"nofollow\">JSON-LD 1.0 specification<\/a> and the <a href=\"https:\/\/www.w3.org\/TR\/2014\/REC-json-ld-api-20140116\/\" rel=\"nofollow\">JSON-LD-API 1.0 specification<\/a>.<\/p>\n<p><a href=\"https:\/\/travis-ci.org\/jsonld-java\/jsonld-java\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/8ce5048e1e4c2d0a62760445fe94233b62ab981d\/68747470733a2f2f7472617669732d63692e6f72672f6a736f6e6c642d6a6176612f6a736f6e6c642d6a6176612e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/jsonld-java\/jsonld-java.svg?branch=master\" \/><\/a> <a href=\"https:\/\/coveralls.io\/r\/jsonld-java\/jsonld-java?branch=master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/67eacc0ef483ccef6bd9efaa00dfc74ba3192714\/68747470733a2f2f636f766572616c6c732e696f2f7265706f732f6a736f6e6c642d6a6176612f6a736f6e6c642d6a6176612f62616467652e7376673f6272616e63683d6d6173746572\" alt=\"Coverage Status\" data-canonical-src=\"https:\/\/coveralls.io\/repos\/jsonld-java\/jsonld-java\/badge.svg?branch=master\" \/><\/a><\/p>\n<h2><a id=\"user-content-from-maven\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-maven\"><\/a>From Maven<\/h2>\n<pre>&lt;dependency&gt;\n    &lt;groupId&gt;com.github.jsonld-java&lt;\/groupId&gt;\n    &lt;artifactId&gt;jsonld-java&lt;\/artifactId&gt;\n    &lt;version&gt;0.12.2&lt;\/version&gt;\n&lt;\/dependency&gt;\n<\/pre>\n<h2><a id=\"user-content-code-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#code-example\"><\/a>Code example<\/h2>\n<div class=\"highlight highlight-source-java\"><pre>\/\/ Open a valid json(-ld) input file\nInputStream inputStream = new FileInputStream(\"input.json\");\n\/\/ Read the file into an Object (The type of this object will be a List, Map, String, Boolean,\n\/\/ Number or null depending on the root object in the file).\nObject jsonObject = JsonUtils.fromInputStream(inputStream);\n\/\/ Create a context JSON map containing prefixes and definitions\nMap context = new HashMap();\n\/\/ Customise context...\n\/\/ Create an instance of JsonLdOptions with the standard JSON-LD options\nJsonLdOptions options = new JsonLdOptions();\n\/\/ Customise options...\n\/\/ Call whichever JSONLD function you want! (e.g. compact)\nObject compact = JsonLdProcessor.compact(jsonObject, context, options);\n\/\/ Print out the result (or don't, it's your call!)\nSystem.out.println(JsonUtils.toPrettyString(compact));<\/pre><\/div>\n<h2><a id=\"user-content-processor-options\" class=\"anchor\" aria-hidden=\"true\" href=\"#processor-options\"><\/a>Processor options<\/h2>\n<p>The Options specified by the <a href=\"https:\/\/json-ld.org\/spec\/latest\/json-ld-api\/#the-jsonldoptions-type\" rel=\"nofollow\">JSON-LD API Specification<\/a> are accessible via the <code>com.github.jsonldjava.core.JsonLdOptions<\/code> class, and each <code>JsonLdProcessor.*<\/code> function has an optional input to take an instance of this class.<\/p>\n<h2><a id=\"user-content-controlling-network-traffic\" class=\"anchor\" aria-hidden=\"true\" href=\"#controlling-network-traffic\"><\/a>Controlling network traffic<\/h2>\n<p>Parsing JSON-LD will normally follow any external <code>@context<\/code> declarations.\nLoading these contexts from the network may in some cases not be desirable, or\nmight require additional proxy configuration or authentication.<\/p>\n<p>JSONLD-Java uses the <a href=\"https:\/\/hc.apache.org\/httpcomponents-client-ga\/index.html\" rel=\"nofollow\">Apache HTTPComponents Client<\/a> for these network connections,\nbased on the <a href=\"http:\/\/hc.apache.org\/httpcomponents-client-ga\/httpclient\/apidocs\/org\/apache\/http\/impl\/client\/SystemDefaultHttpClient.html\" rel=\"nofollow\">SystemDefaultHttpClient<\/a> which reads\nstandard Java properties like <code>http.proxyHost<\/code>.<\/p>\n<p>The default HTTP Client is wrapped with a\n<a href=\"https:\/\/hc.apache.org\/httpcomponents-client-ga\/httpclient-cache\/apidocs\/org\/apache\/http\/impl\/client\/cache\/CachingHttpClient.html\" rel=\"nofollow\">CachingHttpClient<\/a> to provide a\nsmall memory-based cache (1000 objects, max 128 kB each) of regularly accessed contexts.<\/p>\n<h3><a id=\"user-content-loading-contexts-from-classpath\" class=\"anchor\" aria-hidden=\"true\" href=\"#loading-contexts-from-classpath\"><\/a>Loading contexts from classpath<\/h3>\n<p>Your application might be parsing JSONLD documents which always use the same\nexternal <code>@context<\/code> IRIs. Although the default HTTP cache (see above) will\navoid repeated downloading of the same contexts, your application would still\ninitially be vulnerable to network connectivity.<\/p>\n<p>To bypass this issue, and even facilitate parsing of such documents in an\noffline state, it is possible to provide a 'warmed' cache populated\nfrom the classpath, e.g. loaded from a JAR.<\/p>\n<p>In your application, simply add a resource <code>jarcache.json<\/code> to the root of your\nclasspath together with the JSON-LD contexts to embed. (Note that you might\nhave to recursively embed any nested contexts).<\/p>\n<p>The syntax of <code>jarcache.json<\/code> is best explained by example:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>[\n  {\n    \"Content-Location\": \"http:\/\/www.example.com\/context\",\n    \"X-Classpath\": \"contexts\/example.jsonld\",\n    \"Content-Type\": \"application\/ld+json\"\n  },\n  {\n    \"Content-Location\": \"http:\/\/data.example.net\/other\",\n    \"X-Classpath\": \"contexts\/other.jsonld\",\n    \"Content-Type\": \"application\/ld+json\"\n  }\n]<\/pre><\/div>\n<p>(See also <a href=\"https:\/\/github.com\/jsonld-java\/jsonld-java\/blob\/master\/core\/src\/test\/resources\/jarcache.json\">core\/src\/test\/resources\/jarcache.json<\/a>).<\/p>\n<p>This will mean that any JSON-LD document trying to import the <code>@context<\/code>\n<code>http:\/\/www.example.com\/context<\/code> will instead be given\n<code>contexts\/example.jsonld<\/code> loaded as a classpath resource.<\/p>\n<p>The <code>X-Classpath<\/code> location is an IRI reference resolved relative to the\nlocation of the <code>jarcache.json<\/code> - so if you have multiple JARs with a\n<code>jarcache.json<\/code> each, then the <code>X-Classpath<\/code> will be resolved within the\ncorresponding JAR (minimizing any conflicts).<\/p>\n<p>Additional HTTP headers (such as <code>Content-Type<\/code> above) can be included,\nalthough these are generally ignored by JSONLD-Java.<\/p>\n<p>Unless overridden in <code>jarcache.json<\/code>, this <code>Cache-Control<\/code> header is\nautomatically injected together with the current <code>Date<\/code>, meaning that the\nresource loaded from the JAR will effectively never expire (the real HTTP\nserver will never be consulted by the Apache HTTP client):<\/p>\n<pre>Date: Wed, 19 Mar 2014 13:25:08 GMT\nCache-Control: max-age=2147483647\n<\/pre>\n<p>The mechanism for loading <code>jarcache.json<\/code> relies on\n<a href=\"http:\/\/docs.oracle.com\/javase\/7\/docs\/api\/java\/lang\/Thread.html#getContextClassLoader%28%29\" rel=\"nofollow\">Thread.currentThread().getContextClassLoader()<\/a>\nto locate resources from the classpath - if you are running on a command line,\nwithin a framework (e.g. OSGi) or Servlet container (e.g. Tomcat) this should\nnormally be set correctly. If not, try:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>ClassLoader oldContextCL = Thread.currentThread().getContextClassLoader();\ntry { \n    Thread.currentThread().setContextClassLoader(getClass().getClassLoader());\n    JsonLdProcessor.expand(input);   \/\/ or any other JsonLd operation\n} finally { \n    \/\/ Restore, in case the current thread was doing something else\n    \/\/ with the context classloader before calling our method\n    Thread.currentThread().setContextClassLoader(oldContextCL);\n}<\/pre><\/div>\n<p>To disable all remote document fetching, when using the default DocumentLoader, set the\nfollowing Java System Property to \"true\" using:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>System.setProperty(\"com.github.jsonldjava.disallowRemoteContextLoading\", \"true\");<\/pre><\/div>\n<p>You can also use the constant provided in DocumentLoader for the same purpose:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>System.setProperty(DocumentLoader.DISALLOW_REMOTE_CONTEXT_LOADING, \"true\");<\/pre><\/div>\n<p>Note that if you override DocumentLoader you should also support this setting for consistency and security.<\/p>\n<h3><a id=\"user-content-loading-contexts-from-a-string\" class=\"anchor\" aria-hidden=\"true\" href=\"#loading-contexts-from-a-string\"><\/a>Loading contexts from a string<\/h3>\n<p>Your application might be parsing JSONLD documents which reference external <code>@context<\/code> IRIs\nthat are not available as file URIs on the classpath. In this case, the <code>jarcache.json<\/code>\napproch will not work. Instead you can inject the literal context file strings through\nthe <code>JsonLdOptions<\/code> object, as follows:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>\/\/ Inject a context document into the options as a literal string\nDocumentLoader dl = new DocumentLoader();\nJsonLdOptions options = new JsonLdOptions();\n\/\/ ... the contents of \"contexts\/example.jsonld\"\nString jsonContext = \"{ \\\"@contxt\\\": { ... } }\";\ndl.addInjectedDoc(\"http:\/\/www.example.com\/context\",  jsonContext);\noptions.setDocumentLoader(dl);\nInputStream inputStream = new FileInputStream(\"input.json\");\nObject jsonObject = JsonUtils.fromInputStream(inputStream);\nMap context = new HashMap();\nObject compact = JsonLdProcessor.compact(jsonObject, context, options);\nSystem.out.println(JsonUtils.toPrettyString(compact));<\/pre><\/div>\n<h3><a id=\"user-content-customizing-the-apache-httpclient\" class=\"anchor\" aria-hidden=\"true\" href=\"#customizing-the-apache-httpclient\"><\/a>Customizing the Apache HttpClient<\/h3>\n<p>To customize the HTTP behaviour (e.g. to disable the cache or provide\n<a href=\"https:\/\/hc.apache.org\/httpcomponents-client-ga\/tutorial\/html\/authentication.html\" rel=\"nofollow\">authentication\ncredentials)<\/a>,\nyou may want to create and configure your own <code>CloseableHttpClient<\/code> instance, which can\nbe passed to a <code>DocumentLoader<\/code> instance using <code>setHttpClient()<\/code>. This document\nloader can then be inserted into <code>JsonLdOptions<\/code> using <code>setDocumentLoader()<\/code>\nand passed as an argument to <code>JsonLdProcessor<\/code> arguments.<\/p>\n<p>Example of inserting a credential provider (e.g. to load a <code>@context<\/code> protected\nby HTTP Basic Auth):<\/p>\n<div class=\"highlight highlight-source-java\"><pre>Object input = JsonUtils.fromInputStream(..);\nDocumentLoader documentLoader = new DocumentLoader();\n        \nCredentialsProvider credsProvider = new BasicCredentialsProvider();\ncredsProvider.setCredentials(\n        new AuthScope(\"localhost\", 443),\n        new UsernamePasswordCredentials(\"username\", \"password\"));\n       \nCacheConfig cacheConfig = CacheConfig.custom().setMaxCacheEntries(1000)\n        .setMaxObjectSize(1024 * 128).build();\nCloseableHttpClient httpClient = CachingHttpClientBuilder\n        .create()\n        \/\/ allow caching\n        .setCacheConfig(cacheConfig)\n        \/\/ Wrap the local JarCacheStorage around a BasicHttpCacheStorage\n        .setHttpCacheStorage(\n                new JarCacheStorage(null, cacheConfig, new BasicHttpCacheStorage(\n                        cacheConfig)))....\n\t\t\n        \/\/ Add in the credentials provider\n        .setDefaultCredentialsProvider(credsProvider);\n        \/\/ When you are finished setting the properties, call build\n        .build();\ndocumentLoader.setHttpClient(httpClient);\n        \nJsonLdOptions options = new JsonLdOptions();\noptions.setDocumentLoader(documentLoader);\n\/\/ .. and any other options        \nObject rdf = JsonLdProcessor.toRDF(input, options);<\/pre><\/div>\n<h2><a id=\"user-content-playground\" class=\"anchor\" aria-hidden=\"true\" href=\"#playground\"><\/a>PLAYGROUND<\/h2>\n<p>The <a href=\"https:\/\/github.com\/jsonld-java\/jsonld-java-tools\">jsonld-java-tools<\/a> repository contains a simple application which provides command line access to JSON-LD functions<\/p>\n<h3><a id=\"user-content-initial-clone-and-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#initial-clone-and-setup\"><\/a>Initial clone and setup<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>git clone git@github.com:jsonld-java\/jsonld-java-tools.git\nchmod +x .\/jsonldplayground<\/pre><\/div>\n<h3><a id=\"user-content-usage-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage-1\"><\/a>Usage<\/h3>\n<p>run the following to get usage details:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>.\/jsonldplayground --help<\/pre><\/div>\n<h2><a id=\"user-content-for-developers\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-developers\"><\/a>For Developers<\/h2>\n<h3><a id=\"user-content-compiling--packaging\" class=\"anchor\" aria-hidden=\"true\" href=\"#compiling--packaging\"><\/a>Compiling &amp; Packaging<\/h3>\n<p><code>jsonld-java<\/code> uses maven to compile. From the base <code>jsonld-java<\/code> module run <code>mvn clean install<\/code> to install the jar into your local maven repository.<\/p>\n<h3><a id=\"user-content-running-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-tests\"><\/a>Running tests<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>mvn test<\/pre><\/div>\n<p>or<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>mvn test -pl core<\/pre><\/div>\n<p>to run only core package tests<\/p>\n<h3><a id=\"user-content-code-style\" class=\"anchor\" aria-hidden=\"true\" href=\"#code-style\"><\/a>Code style<\/h3>\n<p>The JSONLD-Java project uses custom Eclipse formatting and cleanup style guides to ensure that Pull Requests are fairly simple to merge.<\/p>\n<p>These guides can be found in the \/conf directory and can be installed in Eclipse using \"Properties&gt;Java Code Style&gt;Formatter\", followed by \"Properties&gt;Java Code Style&gt;Clean Up\" for each of the modules making up the JSONLD-Java project.<\/p>\n<p>If you don't use Eclipse, then don't worry, your pull requests can be cleaned up by a repository maintainer prior to merging, but it makes the initial check easier if the modified code uses the conventions.<\/p>\n<h3><a id=\"user-content-submitting-pull-requests\" class=\"anchor\" aria-hidden=\"true\" href=\"#submitting-pull-requests\"><\/a>Submitting Pull Requests<\/h3>\n<p>Once you have made a change to fix a bug or add a new feature, you should commit and push the change to your fork.<\/p>\n<p>Then, you can open a pull request to merge your change into the master branch of the main repository.<\/p>\n<p>The Implementation Reports documenting the conformance of JSONLD-Java with JSONLD-1.0 are available at:<\/p>\n<p><a href=\"https:\/\/github.com\/jsonld-java\/jsonld-java\/tree\/master\/core\/reports\">https:\/\/github.com\/jsonld-java\/jsonld-java\/tree\/master\/core\/reports<\/a><\/p>\n<h3><a id=\"user-content-regenerating-implementation-report\" class=\"anchor\" aria-hidden=\"true\" href=\"#regenerating-implementation-report\"><\/a>Regenerating Implementation Report<\/h3>\n<p>Implementation Reports conforming to the <a href=\"http:\/\/json-ld.org\/test-suite\/reports\/#instructions-for-submitting-implementation-reports\" rel=\"nofollow\">JSON-LD Implementation Report<\/a> document can be regenerated using the following command:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>mvn test -pl core -Dtest=JsonLdProcessorTest -Dreport.format=&lt;format&gt;<\/pre><\/div>\n<p>Current possible values for <code>&lt;format&gt;<\/code> include JSON-LD (<code>application\/ld+json<\/code> or <code>jsonld<\/code>), NQuads (<code>text\/plain<\/code>, <code>nquads<\/code>, <code>ntriples<\/code>, <code>nq<\/code> or <code>nt<\/code>) and Turtle (<code>text\/turtle<\/code>, <code>turtle<\/code> or <code>ttl<\/code>). <code>*<\/code> can be used to generate reports in all available formats.<\/p>\n<p>This is the base package for JSONLD-Java. Integration with other Java packages are done in separate repositories.<\/p>\n<h2><a id=\"user-content-existing-integrations\" class=\"anchor\" aria-hidden=\"true\" href=\"#existing-integrations\"><\/a>Existing integrations<\/h2>\n<h2><a id=\"user-content-creating-an-integration-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-an-integration-module\"><\/a>Creating an integration module<\/h2>\n<h3><a id=\"user-content-create-a-repository-for-your-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-a-repository-for-your-module\"><\/a>Create a repository for your module<\/h3>\n<p>Create a GitHub repository for your module under your user account, or have a JSONLD-Java maintainer create one in the jsonld-java organisation.<\/p>\n<h2><a id=\"user-content-create-maven-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-maven-module\"><\/a>Create maven module<\/h2>\n<h3><a id=\"user-content-create-pomxml-for-your-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-pomxml-for-your-module\"><\/a>Create pom.xml for your module<\/h3>\n<p>Here is the basic outline for what your module's pom.xml should look like<\/p>\n<div class=\"highlight highlight-text-xml\"><pre>&lt;project xmlns=\"http:\/\/maven.apache.org\/POM\/4.0.0\" xmlns:xsi=\"http:\/\/www.w3.org\/2001\/XMLSchema-instance\"\n    xsi:schemaLocation=\"http:\/\/maven.apache.org\/POM\/4.0.0 http:\/\/maven.apache.org\/maven-v4_0_0.xsd\"&gt;\n  &lt;parent&gt;\n    &lt;groupId&gt;com.github.jsonld-java&lt;\/groupId&gt;\n    &lt;artifactId&gt;jsonld-java-parent&lt;\/artifactId&gt;\n    &lt;version&gt;0.12.2&lt;\/version&gt;\n  &lt;\/parent&gt;\n  &lt;modelVersion&gt;4.0.0&lt;\/modelVersion&gt;\n  &lt;artifactId&gt;jsonld-java-{your module}&lt;\/artifactId&gt;\n  &lt;version&gt;0.12.3-SNAPSHOT&lt;\/version&gt;\n  &lt;name&gt;JSONLD Java :: {your module name}&lt;\/name&gt;\n  &lt;description&gt;JSON-LD Java integration module for {RDF Library your module integrates}&lt;\/description&gt;\n  &lt;packaging&gt;jar&lt;\/packaging&gt;\n  &lt;developers&gt;\n    &lt;developer&gt;\n      &lt;name&gt;{YOU}&lt;\/name&gt;\n      &lt;email&gt;{YOUR EMAIL ADDRESS}&lt;\/email&gt;\n    &lt;\/developer&gt;\n  &lt;\/developers&gt;\n  &lt;dependencies&gt;\n    &lt;dependency&gt;\n      &lt;groupId&gt;${project.groupId}&lt;\/groupId&gt;\n      &lt;artifactId&gt;jsonld-java&lt;\/artifactId&gt;\n      &lt;version&gt;${project.version}&lt;\/version&gt;\n      &lt;type&gt;jar&lt;\/type&gt; \n      &lt;scope&gt;compile&lt;\/scope&gt; \n    &lt;\/dependency&gt;\n    &lt;dependency&gt;\n      &lt;groupId&gt;${project.groupId}&lt;\/groupId&gt;\n      &lt;artifactId&gt;jsonld-java&lt;\/artifactId&gt;\n      &lt;version&gt;${project.version}&lt;\/version&gt;\n      &lt;type&gt;test-jar&lt;\/type&gt;\n      &lt;scope&gt;test&lt;\/scope&gt;\n    &lt;\/dependency&gt;\n    &lt;dependency&gt;\n      &lt;groupId&gt;junit&lt;\/groupId&gt;\n      &lt;artifactId&gt;junit&lt;\/artifactId&gt;\n      &lt;scope&gt;test&lt;\/scope&gt;\n    &lt;\/dependency&gt;\n    &lt;dependency&gt;\n      &lt;groupId&gt;org.slf4j&lt;\/groupId&gt;\n      &lt;artifactId&gt;slf4j-jdk14&lt;\/artifactId&gt;\n      &lt;scope&gt;test&lt;\/scope&gt;\n    &lt;\/dependency&gt;\n  &lt;\/dependencies&gt;\n&lt;\/project&gt;<\/pre><\/div>\n<p>Make sure you edit the following:<\/p>\n<ul><li><code>project\/artifactId<\/code> : set this to <code>jsonld-java-{module id}<\/code>, where <code>{module id}<\/code> usually represents the RDF library you're integrating (e.g. <code>jsonld-java-jena<\/code>)<\/li>\n<li><code>project\/name<\/code> : set this to <code>JSONLD Java :: {Module Name}<\/code>, wher <code>{module name}<\/code> is usually the name of the RDF library you're integrating.<\/li>\n<li><code>project\/description<\/code><\/li>\n<li><code>project\/developers\/developer\/...<\/code> : Give youself credit by filling in the developer field. At least put your <code>&lt;name&gt;<\/code> in (<a href=\"http:\/\/maven.apache.org\/pom.html#Developers\" rel=\"nofollow\">see here for all available options<\/a>).<\/li>\n<li><code>project\/dependencies\/...<\/code> : remember to add any dependencies your project needs<\/li>\n<\/ul><h3><a id=\"user-content-import-into-your-favorite-editor\" class=\"anchor\" aria-hidden=\"true\" href=\"#import-into-your-favorite-editor\"><\/a>Import into your favorite editor<\/h3>\n<p>For Example: Follow the first few steps in the section above to import the whole <code>jsonld-java<\/code> project or only your new module into eclipse.<\/p>\n<h2><a id=\"user-content-create-rdfparser-implementation\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-rdfparser-implementation\"><\/a>Create RDFParser Implementation<\/h2>\n<p>The interface <code>com.github.jsonldjava.core.RDFParser<\/code> is used to parse RDF from the library into the JSONLD-Java internal RDF format. See the documentation in <a href=\"https:\/\/github.com\/jsonld-java\/jsonld-java\/blob\/core\/src\/main\/java\/com\/github\/jsonldjava\/core\/RDFParser.java\"><code>RDFParser.java<\/code><\/a> for details on how to implement this interface.<\/p>\n<h2><a id=\"user-content-create-triplecallback-implementation\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-triplecallback-implementation\"><\/a>Create TripleCallback Implementation<\/h2>\n<p>The interface <code>com.github.jsonldjava.core.JSONLDTripleCallback<\/code> is used to generate a representation of the JSON-LD input in the RDF library. See the documentation in <a href=\"https:\/\/github.com\/jsonld-java\/jsonld-java\/blob\/core\/src\/main\/java\/com\/github\/jsonldjava\/core\/JSONLDTripleCallback.java\"><code>JSONLDTripleCallback.java<\/code><\/a> for details on how to implement this interface.<\/p>\n<h2><a id=\"user-content-using-your-implementations\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-your-implementations\"><\/a>Using your Implementations<\/h2>\n<h3><a id=\"user-content-rdfparser\" class=\"anchor\" aria-hidden=\"true\" href=\"#rdfparser\"><\/a>RDFParser<\/h3>\n<p>A JSONLD RDF parser is a class that can parse your frameworks' RDF model\nand generate JSON-LD.<\/p>\n<p>There are two ways to use your <code>RDFParser<\/code> implementation.<\/p>\n<p>Register your parser with the <code>JSONLD<\/code> class and set <code>options.format<\/code> when you call <code>fromRDF<\/code><\/p>\n<div class=\"highlight highlight-source-java\"><pre>JSONLD.registerRDFParser(\"format\/identifier\", new YourRDFParser());\nObject jsonld = JSONLD.fromRDF(yourInput, new Options(\"\") {{ format = \"format\/identifier\" }});<\/pre><\/div>\n<p>or pass an instance of your <code>RDFParser<\/code> into the <code>fromRDF<\/code> function<\/p>\n<div class=\"highlight highlight-source-java\"><pre>Object jsonld = JSONLD.fromRDF(yourInput, new YourRDFParser());<\/pre><\/div>\n<h3><a id=\"user-content-jsonldtriplecallback\" class=\"anchor\" aria-hidden=\"true\" href=\"#jsonldtriplecallback\"><\/a>JSONLDTripleCallback<\/h3>\n<p>A JSONLD triple callback is a class that can populate your framework's\nRDF model from JSON-LD - being called for each triple (technically quad).<\/p>\n<p>Pass an instance of your <code>TripleCallback<\/code> to <code>JSONLD.toRDF<\/code><\/p>\n<div class=\"highlight highlight-source-java\"><pre>Object yourOutput = JSONLD.toRDF(jsonld, new YourTripleCallback());<\/pre><\/div>\n<h2><a id=\"user-content-integrate-with-your-framework\" class=\"anchor\" aria-hidden=\"true\" href=\"#integrate-with-your-framework\"><\/a>Integrate with your framework<\/h2>\n<p>Your framework might have its own system of readers and writers, where\nyou should register JSON-LD as a supported format. Remember that here\nthe \"parse\" direction is opposite of above, a 'reader' may be a class\nthat can parse JSON-LD and populate an RDF Graph.<\/p>\n<h2><a id=\"user-content-write-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#write-tests\"><\/a>Write Tests<\/h2>\n<p>It's helpful to have a test or two for your implementations to make sure they work and continue to work with future versions.<\/p>\n<h2><a id=\"user-content-write-readmemd\" class=\"anchor\" aria-hidden=\"true\" href=\"#write-readmemd\"><\/a>Write README.md<\/h2>\n<p>Write a <code>README.md<\/code> file with instrutions on how to use your module.<\/p>\n<h2><a id=\"user-content-submit-your-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#submit-your-module\"><\/a>Submit your module<\/h2>\n<p>Once you've <code>commit<\/code>ted your code, and <code>push<\/code>ed it into your github fork you can issue a <a href=\"https:\/\/help.github.com\/articles\/using-pull-requests\">Pull Request<\/a> so that we can add a reference to your module in this README file.<\/p>\n<p>Alternatively, we can also host your repository in the jsonld-java organisation to give it more visibility.<\/p>\n<h3><a id=\"user-content-2018-11-03\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-11-03\"><\/a>2018-11-03<\/h3>\n<h3><a id=\"user-content-2018-09-05\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-09-05\"><\/a>2018-09-05<\/h3>\n<ul><li>handle omit graph flag (Patch by @eroux)<\/li>\n<li>Release 0.12.1<\/li>\n<li>Make pruneBlankNodeIdentifiers false by default in 1.0 mode and always true in 1.1 mode (Patch by @eroux)<\/li>\n<li>Fix issue with blank node identifier pruning when @id is aliased (Patch by @eroux)<\/li>\n<li>Allow wildcard {} for @id in framing (Patch by @eroux)<\/li>\n<\/ul><h3><a id=\"user-content-2018-07-07\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-07-07\"><\/a>2018-07-07<\/h3>\n<ul><li>Fix tests setup for schema.org with HttpURLConnection that break because of the inability of HttpURLConnection to redirect from HTTP to HTTPS<\/li>\n<\/ul><h3><a id=\"user-content-2018-04-08\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-04-08\"><\/a>2018-04-08<\/h3>\n<ul><li>Release 0.12.0<\/li>\n<li>Encapsulate RemoteDocument and make it immutable<\/li>\n<\/ul><h3><a id=\"user-content-2018-04-03\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-04-03\"><\/a>2018-04-03<\/h3>\n<ul><li>Fix performance issue caused by not caching schema.org and others that use <code>Cache-Control: private<\/code> (Patch by @HansBrende)<\/li>\n<li>Cache classpath scans for jarcache.json to fix a similar performance issue<\/li>\n<li>Add internal shaded dependency on Google Guava to use maintained soft and weak reference maps rather than adhoc versions<\/li>\n<li>Make JsonLdError a RuntimeException to improve its use in closures<\/li>\n<li>Bump minor version to 0.12 to reflect the API incompatibility caused by JsonLdError and protected field change and hiding in JarCacheStorage<\/li>\n<\/ul><h3><a id=\"user-content-2018-01-25\" class=\"anchor\" aria-hidden=\"true\" href=\"#2018-01-25\"><\/a>2018-01-25<\/h3>\n<ul><li>Fix resource leak in JsonUtils.fromURL on unsuccessful requests (Patch by @plaplaige)<\/li>\n<\/ul><h3><a id=\"user-content-2017-11-15\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-11-15\"><\/a>2017-11-15<\/h3>\n<ul><li>Ignore UTF BOM (Patch by @christopher-johnson)<\/li>\n<\/ul><h3><a id=\"user-content-2017-08-26\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-08-26\"><\/a>2017-08-26<\/h3>\n<ul><li>Release 0.11.1<\/li>\n<li>Fix @embed:@always support (Patch by @dr0i)<\/li>\n<\/ul><h3><a id=\"user-content-2017-08-24\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-08-24\"><\/a>2017-08-24<\/h3>\n<ul><li>Release 0.11.0<\/li>\n<\/ul><h3><a id=\"user-content-2017-08-22\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-08-22\"><\/a>2017-08-22<\/h3>\n<ul><li>Add implicit \"flag only\" subframe to fix incomplete list recursion (Patch by @christopher-johnson)<\/li>\n<li>Support pruneBlankNodeIdentifiers framing option in 1.1 mode (Patch by @fsteeg and @eroux)<\/li>\n<li>Support new @embed values (Patch by @eroux)<\/li>\n<\/ul><h3><a id=\"user-content-2017-07-11\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-07-11\"><\/a>2017-07-11<\/h3>\n<ul><li>Add injection of contexts directly into DocumentLoader (Patch by @ryankenney)<\/li>\n<li>Fix N-Quads content type (Patch by @NicolasRouquette)<\/li>\n<li>Add JsonUtils.fromJsonParser (Patch by @dschulten)<\/li>\n<\/ul><h3><a id=\"user-content-2017-02-16\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-02-16\"><\/a>2017-02-16<\/h3>\n<ul><li>Make literals compare consistently (Patch by @stain)<\/li>\n<li>Release 0.10.0<\/li>\n<\/ul><h3><a id=\"user-content-2017-01-09\" class=\"anchor\" aria-hidden=\"true\" href=\"#2017-01-09\"><\/a>2017-01-09<\/h3>\n<ul><li>Propagate causes for JsonLdError instances where they were caused by other Exceptions<\/li>\n<li>Remove schema.org hack as it appears to work again now...<\/li>\n<li>Remove deprecated and unused APIs<\/li>\n<li>Bump version to 0.10.0-SNAPSHOT per the removed\/changed APIs<\/li>\n<\/ul><h3><a id=\"user-content-2016-12-23\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-12-23\"><\/a>2016-12-23<\/h3>\n<ul><li>Release 0.9.0<\/li>\n<li>Fixes schema.org support that is broken with Apache HTTP Client but works with java.net.URL<\/li>\n<\/ul><h3><a id=\"user-content-2016-05-20\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-05-20\"><\/a>2016-05-20<\/h3>\n<ul><li>Fix reported NPE in JsonLdApi.removeDependents<\/li>\n<\/ul><h3><a id=\"user-content-2016-05-18\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-05-18\"><\/a>2016-05-18<\/h3>\n<ul><li>Release 0.8.3<\/li>\n<li>Fix @base in remote contexts corrupting the local context<\/li>\n<\/ul><h3><a id=\"user-content-2016-04-23\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-04-23\"><\/a>2016-04-23<\/h3>\n<ul><li>Support @default inside of sets for framing<\/li>\n<\/ul><h3><a id=\"user-content-2016-02-29\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-02-29\"><\/a>2016-02-29<\/h3>\n<ul><li>Fix ConcurrentModificationException in the implementation of the Framing API<\/li>\n<\/ul><h3><a id=\"user-content-2016-02-17\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-02-17\"><\/a>2016-02-17<\/h3>\n<ul><li>Re-release version 0.8.2 with the refactoring work actually in it. 0.8.1 is identical in functionality to 0.8.0<\/li>\n<li>Release version 0.8.1<\/li>\n<li>Refactor JSONUtils and DocumentLoader to move most of the static logic into JSONUtils, and deprecate the DocumentLoader versions<\/li>\n<\/ul><h3><a id=\"user-content-2016-02-10\" class=\"anchor\" aria-hidden=\"true\" href=\"#2016-02-10\"><\/a>2016-02-10<\/h3>\n<ul><li>Release version 0.8.0<\/li>\n<\/ul><h3><a id=\"user-content-2015-11-19\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-11-19\"><\/a>2015-11-19<\/h3>\n<ul><li>Replace deprecated HTTPClient code with the new builder pattern<\/li>\n<li>Chain JarCacheStorage to any other HttpCacheStorage to simplify the way local caching is performed<\/li>\n<li>Bump version to 0.8.0-SNAPSHOT as some interface method parameters changed, particularly, DocumentLoader.setHttpClient changed to require CloseableHttpClient that was introduced in HttpClient-4.3<\/li>\n<\/ul><h3><a id=\"user-content-2015-11-16\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-11-16\"><\/a>2015-11-16<\/h3>\n<ul><li>Bump dependencies to latest versions, particularly HTTPClient that is seeing more use on 4.5\/4.4 than the 4.2 series that we have used so far<\/li>\n<li>Performance improvements for serialisation to N-Quads by replacing string append and replace with StringBuilder<\/li>\n<li>Support setting a system property, com.github.jsonldjava.disallowRemoteContextLoading, to \"true\" to disable remote context loading.<\/li>\n<\/ul><h3><a id=\"user-content-2015-09-30\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-09-30\"><\/a>2015-09-30<\/h3>\n<ul><li>Release 0.7.0<\/li>\n<\/ul><h3><a id=\"user-content-2015-09-27\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-09-27\"><\/a>2015-09-27<\/h3>\n<ul><li>Move Tools, Clerezza and RDF2GO modules out to separate repositories. The Tools repository had a circular build dependency with Sesame, while the other modules are best located and managed in separate repositories<\/li>\n<\/ul><h3><a id=\"user-content-2015-08-25\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-08-25\"><\/a>2015-08-25<\/h3>\n<ul><li>Remove Sesame-2.7 module in favour of sesame-rio-jsonld for Sesame-2.8 and 4.0<\/li>\n<li>Fix bug where parsing did not fail if content was present after the end of a full JSON top level element<\/li>\n<\/ul><h3><a id=\"user-content-2015-03-12\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-03-12\"><\/a>2015-03-12<\/h3>\n<ul><li>Compact context arrays if they contain a single element during compaction<\/li>\n<li>Bump to Sesame-2.7.15<\/li>\n<\/ul><h3><a id=\"user-content-2015-03-01\" class=\"anchor\" aria-hidden=\"true\" href=\"#2015-03-01\"><\/a>2015-03-01<\/h3>\n<ul><li>Use jopt-simple for the playground cli to simplify the coding and improve error messages<\/li>\n<li>Allow RDF parsing and writing using all of the available Sesame Rio parsers through the playground cli<\/li>\n<li>Make the httpclient dependency OSGi compliant<\/li>\n<\/ul><h3><a id=\"user-content-2014-12-31\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-12-31\"><\/a>2014-12-31<\/h3>\n<ul><li>Fix locale sensitive serialisation of XSD double\/decimal typed literals to always be Locale.US<\/li>\n<li>Bump to Sesame-2.7.14<\/li>\n<li>Bump to Clerezza-0.14<\/li>\n<\/ul><h3><a id=\"user-content-2014-11-14\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-11-14\"><\/a>2014-11-14<\/h3>\n<ul><li>Fix identification of integer, boolean, and decimal in RDF-JSONLD with useNativeTypes<\/li>\n<li>Release 0.5.1<\/li>\n<\/ul><h3><a id=\"user-content-2014-10-29\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-10-29\"><\/a>2014-10-29<\/h3>\n<ul><li>Add OSGi metadata to Jar files<\/li>\n<li>Bump to Sesame-2.7.13<\/li>\n<\/ul><h3><a id=\"user-content-2014-07-14\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-07-14\"><\/a>2014-07-14<\/h3>\n<ul><li>Release version 0.5.0<\/li>\n<li>Fix Jackson parse exceptions being propagated through Sesame without wrapping as RDFParseExceptions<\/li>\n<\/ul><h3><a id=\"user-content-2014-07-02\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-07-02\"><\/a>2014-07-02<\/h3>\n<ul><li>Fix use of Java-7 API so we are still Java-6 compatible<\/li>\n<li>Ensure that Sesame RDFHandler endRDF and startRDF are called in SesameTripleCallback<\/li>\n<\/ul><h3><a id=\"user-content-2014-06-30\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-06-30\"><\/a>2014-06-30<\/h3>\n<ul><li>Release version 0.4.2<\/li>\n<li>Bump to Sesame-2.7.12<\/li>\n<li>Remove Jena integration module, as it is now maintained by Jena team in their repository<\/li>\n<\/ul><h3><a id=\"user-content-2014-04-22\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-04-22\"><\/a>2014-04-22<\/h3>\n<ul><li>Release version 0.4<\/li>\n<li>Bump to Sesame-2.7.11<\/li>\n<li>Bump to Jackson-2.3.3<\/li>\n<li>Bump to Jena-2.11.1<\/li>\n<\/ul><h3><a id=\"user-content-2014-03-26\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-03-26\"><\/a>2014-03-26<\/h3>\n<ul><li>Bump RDF2GO to version 5.0.0<\/li>\n<\/ul><h3><a id=\"user-content-2014-03-24\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-03-24\"><\/a>2014-03-24<\/h3>\n<ul><li>Allow loading remote @context from bundled JAR cache<\/li>\n<li>Support JSON array in @context with toRDF<\/li>\n<li>Avoid exception on @context with default @language and unmapped key<\/li>\n<\/ul><h3><a id=\"user-content-2014-02-24\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-02-24\"><\/a>2014-02-24<\/h3>\n<ul><li>Javadoc some core classes, JsonLdProcessor, JsonLdApi, and JsonUtils<\/li>\n<li>Rename some core classes for consistency, particularly JSONUtils to JsonUtils and JsonLdTripleCallback<\/li>\n<li>Fix for a Context constructor that wasn't taking base into account<\/li>\n<\/ul><h3><a id=\"user-content-2014-02-20\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-02-20\"><\/a>2014-02-20<\/h3>\n<ul><li>Fix JsonLdApi mapping options in framing algorithm (Thanks Scott Blomquist @sblom)<\/li>\n<\/ul><h3><a id=\"user-content-2014-02-06\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-02-06\"><\/a>2014-02-06<\/h3>\n<ul><li>Release version 0.3<\/li>\n<li>Bump to Sesame-2.7.10<\/li>\n<li>Fix Jena module to use new API<\/li>\n<\/ul><h3><a id=\"user-content-2014-01-29\" class=\"anchor\" aria-hidden=\"true\" href=\"#2014-01-29\"><\/a>2014-01-29<\/h3>\n<ul><li>Updated to final Recommendation<\/li>\n<li>Namespaces supported by Sesame integration module<\/li>\n<li>Initial implementation of remote document loading<\/li>\n<li>Bump to Jackson-2.3.1<\/li>\n<\/ul><h3><a id=\"user-content-2013-11-22\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-11-22\"><\/a>2013-11-22<\/h3>\n<ul><li>updated jena writer<\/li>\n<\/ul><h3><a id=\"user-content-2013-11-07\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-11-07\"><\/a>2013-11-07<\/h3>\n<ul><li>Integration packages renamed com.github.jsonldjava.sesame,\ncom.github.jsonldjava.jena etc. (Issue #76)<\/li>\n<\/ul><h3><a id=\"user-content-2013-10-07\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-10-07\"><\/a>2013-10-07<\/h3>\n<ul><li>Matched class names to Spec<\/li>\n<\/ul><ul><li>Renamed <code>JSONLDException<\/code> to <code>JsonLdError<\/code><\/li>\n<li>Renamed <code>JSONLDProcessor<\/code> to <code>JsonLdApi<\/code><\/li>\n<li>Renamed <code>JSONLD<\/code> to <code>JsonLdProcessor<\/code><\/li>\n<li>Renamed <code>ActiveContext<\/code> to <code>Context<\/code><\/li>\n<li>Renamed <code>Options<\/code> to <code>JsonLdOptions<\/code><\/li>\n<\/ul><ul><li>All context related utility functions moved to be members of the <code>Context<\/code> class<\/li>\n<\/ul><h3><a id=\"user-content-2013-09-30\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-09-30\"><\/a>2013-09-30<\/h3>\n<ul><li>Fixed JSON-LD to Jena to handle of BNodes<\/li>\n<\/ul><h3><a id=\"user-content-2013-09-02\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-09-02\"><\/a>2013-09-02<\/h3>\n<ul><li>Add RDF2Go integration<\/li>\n<li>Bump Sesame and Clerezza dependency versions<\/li>\n<\/ul><h3><a id=\"user-content-2013-06-18\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-06-18\"><\/a>2013-06-18<\/h3>\n<ul><li>Bump to version 0.2<\/li>\n<li>Updated Turtle integration<\/li>\n<li>Added Caching of contexts loaded from URI<\/li>\n<li>Added source formatting eclipse config<\/li>\n<li>Fixed up seasame integration package names<\/li>\n<li>Replaced depreciated Jackson code<\/li>\n<\/ul><h3><a id=\"user-content-2013-05-19\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-05-19\"><\/a>2013-05-19<\/h3>\n<ul><li>Added Turtle RDFParser and TripleCallback<\/li>\n<li>Changed Maven groupIds to <code>com.github.jsonld-java<\/code> to match github domain.<\/li>\n<li>Released version 0.1<\/li>\n<\/ul><h3><a id=\"user-content-2013-05-16\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-05-16\"><\/a>2013-05-16<\/h3>\n<ul><li>Updated core code to match <a href=\"http:\/\/json-ld.org\/spec\/latest\/json-ld-api\/\" rel=\"nofollow\">JSON-LD 1.0 Processing Algorithms and API \/ W3C Editor's Draft 14 May 2013<\/a><\/li>\n<li>Deprecated JSONLDSerializer in favor of the RDFParser interface to better represent the purpose of the interface and better fit in with the updated core code.<\/li>\n<li>Updated the JSONLDTripleCallback to better fit with the updated code.<\/li>\n<li>Updated the Playground tool to support updated core code.<\/li>\n<\/ul><h3><a id=\"user-content-2013-05-07\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-05-07\"><\/a>2013-05-07<\/h3>\n<ul><li>Changed base package names to com.github.jsonldjava<\/li>\n<li>Reverted version to 0.1-SNAPSHOT to allow version incrementing pre 1.0 while allowing a 1.0 release when the json-ld spec is finalised.<\/li>\n<li>Turned JSONLDTripleCallback into an interface.<\/li>\n<\/ul><h3><a id=\"user-content-2013-04-18\" class=\"anchor\" aria-hidden=\"true\" href=\"#2013-04-18\"><\/a>2013-04-18<\/h3>\n<ul><li>Updated to Sesame 2.7.0, Jena 2.10.0, Jackson 2.1.4<\/li>\n<li>Fixing a character encoding issue in the JSONLDProcessorTests<\/li>\n<li>Bumping to 1.0.1 to reflect dependency changes<\/li>\n<\/ul><h3><a id=\"user-content-2012-10-30\" class=\"anchor\" aria-hidden=\"true\" href=\"#2012-10-30\"><\/a>2012-10-30<\/h3>\n<ul><li>Brought the implementation up to date with the reference implementation (minus the normalization stuff)<\/li>\n<li>Changed entry point for the functions to the static functions in the JSONLD class<\/li>\n<li>Changed the JSONLDSerializer to an abstract class, requiring the implementation of a \"parse\" function. The JSONLDSerializer is now passed to the JSONLD.fromRDF function.<\/li>\n<li>Added JSONLDProcessingError class to handle errors more efficiently<\/li>\n<\/ul><ul><li>The <code>Context<\/code> class is a <code>Map<\/code> and many of the options are stored as values of the map. These could be made into variables, whice should speed things up a bit (the same with the termDefinitions variable inside the Context).<\/li>\n<li>some sort of document loader interface (with a mockup for testing) is required<\/li>\n<\/ul><\/article>","created_at":"2018-11-10T17:41:50+0000","updated_at":"2018-11-10T17:41:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/4359856?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["json-ld","activitypub","clojure"],"is_public":false,"id":677,"title":"SnootyMonkey\/clj-json-ld","url":"https:\/\/github.com\/SnootyMonkey\/clj-json-ld","content":"<p><a href=\"https:\/\/www.mozilla.org\/MPL\/2.0\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/c214fab1f19e907ac5a8f50a3e8dbdc0c26c9f16\/687474703a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d504c2d626c75652e7376673f7374796c653d666c6174\" alt=\"MPL License\" data-canonical-src=\"http:\/\/img.shields.io\/badge\/license-MPL-blue.svg?style=flat\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/SnootyMonkey\/clj-json-ld\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7ea6b3dfcdf4c83057b3a41a6b8bd3c562ff0f75\/687474703a2f2f696d672e736869656c64732e696f2f7472617669732f536e6f6f74794d6f6e6b65792f636c6a2d6a736f6e2d6c642e7376673f7374796c653d666c6174\" alt=\"Build Status\" data-canonical-src=\"http:\/\/img.shields.io\/travis\/SnootyMonkey\/clj-json-ld.svg?style=flat\" \/><\/a><\/p><blockquote>\n<p>\"Data is messy and disconnected. JSON-LD organizes and connects it, creating a better Web.\"<\/p>\n<\/blockquote><p>clj-json-ld is the Clojure library for <a href=\"http:\/\/json-ld.org\/\" rel=\"nofollow\">JSON-LD<\/a> (JavaScript Object Notation for Linking Data).<\/p><ul><li><a href=\"#introduction\">Introduction<\/a>\n<ul><li><a href=\"#benefits-of-linked-data\">Benefits of Linked Data<\/a><\/li>\n<li><a href=\"#benefits-of-json-ld\">Benefits of JSON-LD<\/a><\/li>\n<li><a href=\"#json-ld-101\">JSON-LD 101<\/a><\/li>\n<li><a href=\"#capabilities-of-clj-json-ld\">Capabilities of clj-json-ld<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#installation\">Installation<\/a><\/li>\n<li><a href=\"#usage\">Usage<\/a><\/li>\n<li><a href=\"#testing\">Testing<\/a><\/li>\n<li><a href=\"#development-and-contributing\">Development and Contributing<\/a>\n<ul><li><a href=\"#branches\">Branches<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#acknowledgements\">Acknowledgements<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2>Introduction<\/h2><p>clj-json-ld implements the <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/\" rel=\"nofollow\">JSON LD 1.0 Processing Algorithms and API<\/a>\nas a <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/#conformance\" rel=\"nofollow\">conforming JSON-LD Processor<\/a>.<\/p><h3>Benefits of Linked Data<\/h3><p><a href=\"http:\/\/www.w3.org\/DesignIssues\/LinkedData.html\" rel=\"nofollow\">Linked Data<\/a> is a way to create a network of standards-based machine interpretable data across different documents, APIs and Web sites. It allows an application to start at one piece of Linked Data, and follow embedded links to other pieces of Linked Data that come from different sites and APIs across the Web.<\/p><h3>Benefits of JSON-LD<\/h3><p>JSON-LD is a lightweight syntax to serialize Linked Data in <a href=\"http:\/\/www.ietf.org\/rfc\/rfc4627.txt\" rel=\"nofollow\">JSON<\/a>. Its design allows existing JSON to be interpreted as Linked Data with minimal changes.<\/p><p>JSON-LD is a way to use Linked Data in API programming environments, to build interoperable APIs, and to store Linked Data in JSON-based database engines (most NoSQL databases). Since JSON-LD is 100% compatible with JSON, the large number of JSON parsers, libraries and databases available today can be reused.<\/p><p>JSON-LD provides:<\/p><ul><li>a universal identifier mechanism for JSON objects,<\/li>\n<li>a way to disambiguate keys shared among different JSON documents,<\/li>\n<li>a mechanism in which a value in a JSON object may refer to a JSON object on a different site on the Web,<\/li>\n<li>the ability to annotate strings in JSON objects with their language,<\/li>\n<li>a way to associate data types with string values such as dates and times,<\/li>\n<li>and a facility to express one or more directed graphs, such as a social network, or a taxonomy, in a single document.<\/li>\n<\/ul><p>The JSON-LD syntax is designed to not disturb already deployed systems running on JSON, and provides a smooth upgrade path from JSON to JSON-LD. Since the shape of JSON data varies wildly, JSON-LD features mechanisms to reshape documents into a deterministic structure which simplifies their processing.<\/p><p>JSON-LD is designed to be usable directly as JSON, with no knowledge of <a href=\"http:\/\/www.w3.org\/TR\/2014\/PR-rdf11-concepts-20140109\/\" rel=\"nofollow\">RDF<\/a>. It is an alternative syntax for the same underlying data model as RDF, so it is also designed to be transformable to\/from RDF, if desired, for use with other Linked Data technologies like <a href=\"http:\/\/www.w3.org\/TR\/rdf-sparql-query\/\" rel=\"nofollow\">SPARQL<\/a>.<\/p><p>Developers who require any of the facilities listed above or need to serialize\/deserialize an RDF Graph or RDF Dataset in a JSON-based syntax will find JSON-LD of interest.<\/p><h3>JSON-LD 101<\/h3><p>Let's take a look at some very simple JSON about a book that you might get back from an API:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"name\": \"Myth of Sisyphus\",\n  \"author\": \"Albert Camus\",\n  \"location\": \"http:\/\/amazon.com\/Myth-Sisyphus-Albert-Camus\/dp\/7500133340\/\",\n  \"image\": \"http:\/\/ecx.images-amazon.com\/images\/I\/61hJVrZgBBL.jpg\"\n}<\/pre><\/div><p>A different API might provide this JSON about the very same book:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"author\": \"Albert Camus\",\n  \"title\": \"Myth of Sisyphus\",\n  \"image\": \"myth.png\",\n  \"location\": \"3rd Floor, Manning Bldg.\",\n  \"lang\": \"en-US\"\n}<\/pre><\/div><p>As a human, it's easy to deduce what this is all about. We have the title of the book, the author, and different locations and images of the book.<\/p><p>As a computer algorithm however, this is all quite vague. Is <code>name<\/code> the same as <code>title<\/code>? Is <code>location<\/code> a place or a URL? Is <code>image<\/code> the name of a file? A full URL? A relative URL? A base-64 encoded image? What's <code>lang<\/code> mean?<\/p><p>The same JSON documents converted to JSON-LD documents remove much of this ambiguity:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"http:\/\/www.schema.org\/name\": \"Myth of Sisyphus\",\n  \"http:\/\/www.schema.org\/author\": \"Albert Camus\",\n  \"http:\/\/www.schema.org\/url\": {\"@id\": \"http:\/\/amazon.com\/Myth-Sisyphus-Albert-Camus\/dp\/7500133340\/\"}, \u2190 The '@id' keyword means 'This value is an identifier that is an IRI (URL)'\n  \"http:\/\/www.schema.org\/image\": {\"@id\": \"http:\/\/ecx.images-amazon.com\/images\/I\/61hJVrZgBBL.jpg\"} \u2190 The '@id' keyword means 'This value is an identifier that is an IRI (URL)'\n}<\/pre><\/div><p>and:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"http:\/\/www.schema.org\/author\": \"Albert Camus\",\n  \"http:\/\/www.schema.org\/name\": \"Myth of Sisyphus\",\n  \"http:\/\/www.schema.org\/image\": \"myth.png\",\n  \"http:\/\/www.schema.org\/contentLocation\": \"3rd Floor, Manning Library\",\n  \"http:\/\/www.schema.org\/inLanguage\": \"en-US\"\n}<\/pre><\/div><p>Our ambiguity is cleared up nicely by using JSON-LD documents. The <code>@id<\/code> term tells us that <code>image<\/code> is a URL in one case, but not in the other. Using a common schema, <a href=\"http:\/\/www.schema.org\/Book\" rel=\"nofollow\">schema.org's book schema<\/a> in this case, has brought together <code>name<\/code> and <code>title<\/code> as referring to the same thing, but the <code>location<\/code> values as referring to different things, a virtual location in one case and a physical location in the other.<\/p><p>But who wants their JSON documents to look like this? A computer algorithm that's dealing with JSON documents from different sources does, but you don't want to read and write your JSON documents like this all the time. With JSON-LD we can move this metadata out of the JSON document and into a context.<\/p><p>The contexts for the above JSON-LD documents look like this:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"@context\": {\n    \"name\": \"http:\/\/schema.org\/name\",  \u2190 This means that 'name' is shorthand for 'http:\/\/schema.org\/name' \n    \"author\": \"http:\/\/www.schema.org\/author\",  \u2190 This means that 'author' is shorthand for 'http:\/\/schema.org\/author' \n    \"location\": {\n      \"@id\": \"http:\/\/schema.org\/url\",  \u2190 This means that 'location' is shorthand for 'http:\/\/schema.org\/url' \n      \"@type\": \"@id\"  \u2190 This means that a string value associated with 'location' should be interpreted as an identifier that is an IRI (a URL)\n    },\n    \"image\": {\n      \"@id\": \"http:\/\/schema.org\/image\",  \u2190 This means that 'image' is shorthand for 'http:\/\/schema.org\/image' \n      \"@type\": \"@id\"  \u2190 This means that a string value associated with 'image' should be interpreted as an identifier that is an IRI (a URL)\n    }\n  }\n}<\/pre><\/div><p>and:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"@context\": {\n    \"author\": \"http:\/\/www.schema.org\/author\",  \u2190 This means that 'author' is shorthand for 'http:\/\/schema.org\/author' \n    \"title\": \"http:\/\/www.schema.org\/name\",  \u2190 This means that 'title' is shorthand for 'http:\/\/schema.org\/name' \n    \"image\": \"http:\/\/schema.org\/image\", \u2190 This means that 'image' is shorthand for 'http:\/\/schema.org\/image', and is not an IRI (a URL)\n    \"location\": \"http:\/\/www.schema.org\/contentLocation\",  \u2190 This means that 'location' is shorthand for 'http:\/\/schema.org\/contentLocation', and is not an IRI (a URL)\n    \"lang\": \"http:\/\/www.schema.org\/inLanguage\"  \u2190 This means that 'lang' is shorthand for 'http:\/\/schema.org\/inLanguage' \n  }\n}<\/pre><\/div><p>If that context is then embedded in its own section of the JSON-LD document, or placed in an accessible location online, say at <code>http:\/\/the-site.org\/contexts\/book.jsonld<\/code>, then the original simpler JSON can be used, but with the semantic ambiguity removed. This is achieved with the addition of a <code>@context<\/code> property to make it a JSON-LD document.<\/p><p>The context can be included by reference:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"@context\": \"http:\/\/the-site.org\/contexts\/book.jsonld\",\n  \"name\": \"Myth of Sisyphus\",\n  \"author\": \"Albert Camus\",\n  \"location\": \"http:\/\/amazon.com\/Myth-Sisyphus-Albert-Camus\/dp\/7500133340\/\",\n  \"image\": \"http:\/\/ecx.images-amazon.com\/images\/I\/61hJVrZgBBL.jpg\"\n}<\/pre><\/div><p>Or the context can be directly embedded:<\/p><div class=\"highlight highlight-source-json\"><pre>{\n  \"@context\": {\n    \"author\": \"http:\/\/www.schema.org\/author\", \n    \"title\": \"http:\/\/www.schema.org\/name\", \n    \"image\": \"http:\/\/schema.org\/image\",\n    \"location\": \"http:\/\/www.schema.org\/contentLocation\",\n    \"lang\": \"http:\/\/www.schema.org\/inLanguage\" \n  },\n  \"author\": \"Albert Camus\",\n  \"title\": \"Myth of Sisyphus\",\n  \"image\": \"myth.png\",\n  \"location\": \"3rd Floor, Manning Bldg.\",\n  \"lang\": \"en-US\"\n}<\/pre><\/div><p>Or the context can be left out completely and provided by an HTTP Link Header:<\/p><pre>HTTP\/1.1 200 OK\nContent-Type: application\/json\nLink: &lt;http:\/\/the-site.org\/contexts\/book.jsonld&gt;; rel=\"http:\/\/www.w3.org\/ns\/json-ld#context\"; type=\"application\/ld+json\"\n{\n  \"name\": \"Myth of Sisyphus\",\n  \"author\": \"Albert Camus\",\n  \"location\": \"http:\/\/amazon.com\/Myth-Sisyphus-Albert-Camus\/dp\/7500133340\/\",\n  \"image\": \"http:\/\/ecx.images-amazon.com\/images\/I\/61hJVrZgBBL.jpg\"\n}\n<\/pre><p>Or maybe the JSON provider is not interested in providing JSON-LD in any form. In that case, we can create the context for the JSON data ourselves and provide it directly to the JSON-LD processor.<\/p><h3>Capabilities of clj-json-ld<\/h3><p>A JSON-LD processor, like clj-json-ld, helps you transform your JSON-LD documents in between different valid JSON-LD formats, some more explicit for semantic reasoning by algorithms, and others more readable and compact for humans and for transmission.<\/p><p>clj-json-ld can perform the <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/#expanded-document-form\" rel=\"nofollow\">expansion<\/a>, <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/#compacted-document-form\" rel=\"nofollow\">compaction<\/a>, and <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/#flattened-document-form\" rel=\"nofollow\">flattening<\/a> operations defined in the <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/#expansion-algorithm\" rel=\"nofollow\">expansion<\/a>, <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/#compaction-algorithm\" rel=\"nofollow\">compaction<\/a> and <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/#flattening-algorithm\" rel=\"nofollow\">flattening<\/a> sections of the <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/\" rel=\"nofollow\">processing specification<\/a>.<\/p><p>\ud83d\udcdd Show the document expanded, compacted and flattened.<\/p><h2>Installation<\/h2><p>\ud83d\udca3 This library is currently a WIP stub and is not yet usable. Stay tuned!<\/p><h2>Usage<\/h2><p>At the Clojure REPL:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(require '[clj-json-ld.core :as json-ld])<\/pre><\/div><p>In your Clojure namespace:<\/p><div class=\"highlight highlight-source-clojure\"><pre>(ns your-app.core\n  (:require [clj-json-ld.core :as :json-ld]))  <\/pre><\/div><p>\u26a0\ufe0f The <code>flatten<\/code> function specified in <a href=\"http:\/\/www.w3.org\/TR\/json-ld-api\/\" rel=\"nofollow\">JSON-LD 1.0 Processing Algorithms and API<\/a> conflicts with Clojure core's <a href=\"https:\/\/clojuredocs.org\/clojure.core\/flatten\" rel=\"nofollow\">flatten<\/a> function, so unless you use <code>:as<\/code> in your require, or exclude Clojure's <code>flatten<\/code> from your namespace with <code>(:refer-clojure :exclude [flatten])<\/code>, you'll see a warning about the replacement:<\/p><pre>WARNING: flatten already refers to: #'clojure.core\/flatten in namespace: user, being replaced by: #'clj-json-ld.core\/flatten\n<\/pre><p>More detailed usage instructions are in the API documentation.<\/p><h2>Testing<\/h2><p>The <a href=\"http:\/\/json-ld.org\/test-suite\/\" rel=\"nofollow\">JSON-LD Test Suite<\/a> is used to verify <a href=\"http:\/\/json-ld.org\/test-suite\/reports\/\" rel=\"nofollow\">JSON-LD Processor conformance<\/a>. clj-json-ld uses the JSON-LD Test Suite, as well as some additional tests.<\/p><p>Tests are run in continuous integration of the <code>master<\/code> and <code>dev<\/code> branches on <a href=\"https:\/\/travis-ci.org\/path\/posthere.io\" rel=\"nofollow\">Travis CI<\/a>:<\/p><p><a href=\"https:\/\/travis-ci.org\/SnootyMonkey\/clj-json-ld\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7ea6b3dfcdf4c83057b3a41a6b8bd3c562ff0f75\/687474703a2f2f696d672e736869656c64732e696f2f7472617669732f536e6f6f74794d6f6e6b65792f636c6a2d6a736f6e2d6c642e7376673f7374796c653d666c6174\" alt=\"Build Status\" data-canonical-src=\"http:\/\/img.shields.io\/travis\/SnootyMonkey\/clj-json-ld.svg?style=flat\" \/><\/a><\/p><p>To run the tests locally, you need to clone the <a href=\"https:\/\/github.com\/json-ld\/tests\">JSON-LD tests GitHub repository<\/a> as a peer directory to this repository:<\/p><div class=\"highlight highlight-text-shell-session\"><pre>git clone https:\/\/github.com\/json-ld\/tests json-ld-tests<\/pre><\/div><p>Then run the tests from this repository with:<\/p><div class=\"highlight highlight-text-shell-session\"><pre>lein midje<\/pre><\/div><h2>Development and Contributing<\/h2><p>If you'd like to enhance clj-json-ld, please fork it <a href=\"https:\/\/github.com\/SnootyMonkey\/clj-json-ld\">on GitHub<\/a>. If you'd like to contribute back your enhancements (awesome!), please submit your pull requests to the <code>dev<\/code> branch. I promise to look at every pull request and incorporate it, or at least provide feedback on why if I won't.<\/p><ul><li>Do your best to conform to the coding style that's here... I like it.<\/li>\n<li>Use 2 soft spaces for indentation.<\/li>\n<li>Don't leave trailing spaces after lines.<\/li>\n<li>Don't leave trailing new lines at the end of files.<\/li>\n<li>Write comments.<\/li>\n<li>Write tests.<\/li>\n<li>Don't submit \u00fcber pull requests, keep your changes atomic.<\/li>\n<li>Have fun!<\/li>\n<\/ul><h3>Branches<\/h3><p>There are 2 long lived branches in the repository:<\/p><p><a href=\"https:\/\/github.com\/SnootyMonkey\/clj-json-ld\/tree\/master\">master<\/a> - mainline, picked up by <a href=\"https:\/\/travis-ci.org\/SnootyMonkey\/clj-json-ld\" rel=\"nofollow\">continual integration on Travis-CI<\/a>, named releases are tagged with: vX.X.X<\/p><p><a href=\"https:\/\/github.com\/SnootyMonkey\/clj-json-ld\/tree\/dev\">dev<\/a> - development mainline, picked up by <a href=\"https:\/\/travis-ci.org\/SnootyMonkey\/clj-json-ld\" rel=\"nofollow\">continual integration on Travis-CI<\/a><\/p><p>Additional short lived feature branches will come and go.<\/p><h2>Acknowledgements<\/h2><p>Thank you to the creators of the <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/\" rel=\"nofollow\">JSON-LD 1.0 W3C Recommendation<\/a> for their hard work in creating the specification and the comprehensive test suite. Thank you to the authors and editors for the very clear spec writing.<\/p><p>Portions of the <a href=\"#benefits-of-linked-data\">Benefits of Linked Data<\/a> and <a href=\"#benefits-of-json-ld\">Benefits of JSON-LD<\/a> sections of this README document are lifted with slight modifications from the <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/\" rel=\"nofollow\">JSON-LD 1.0 W3C Recommendation<\/a>.<\/p><p>Thank you to <a href=\"https:\/\/github.com\/gkellogg\">Gregg Kellog<\/a>, author of the <a href=\"https:\/\/github.com\/ruby-rdf\/json-ld\/\">json-ld Ruby processor<\/a>, and <a href=\"https:\/\/github.com\/dlongley\">Dave Longley<\/a>, author of the <a href=\"https:\/\/github.com\/digitalbazaar\/pyld\">pyld Python processor<\/a>, for providing prior implementations that were useful to study, particularly in the area of testing.<\/p><h2>License<\/h2><p>clj-json-ld is distributed under the <a href=\"http:\/\/www.mozilla.org\/MPL\/2.0\/\" rel=\"nofollow\">Mozilla Public License v2.0<\/a>.<\/p><p>Copyright \u00a9 2014-2015 <a href=\"http:\/\/snootymonkey.com\/\" rel=\"nofollow\">Snooty Monkey, LLC<\/a><\/p>","created_at":"2018-11-10T17:40:25+0000","updated_at":"2018-11-10T17:40:43+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/3689846?s=400&v=4","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["css","html"],"is_public":false,"id":676,"title":"Advanced CSS Theming with Custom Properties and JavaScript","url":"https:\/\/www.sitepoint.com\/css-theming-custom-properties-javascript\/","content":"<p><strong>Throughout this tutorial on CSS theming, we\u2019ll be using CSS custom properties (also known as CSS variables) to implement dynamic themes for a simple HTML page. We\u2019ll create dark and light example themes, then write JavaScript to switch between the two when the user clicks a button.<\/strong><\/p><p>Just like in typical programming languages, variables are used to hold or store values. In CSS, they\u2019re typically used to store colors, font names, font sizes, length units, etc. They can then be referenced and reused in multiple places in the stylesheet. Most developers refer to \u201cCSS variables\u201d, but the official name is <strong>custom properties<\/strong>.<\/p><p>CSS custom properties make it possible to modify variables that can be referenced throughout the stylesheet. Previously, this was only possible with CSS preprocessors such as Sass.<\/p><h2 id=\"understandingrootandvar\">Understanding :root and <code>var()<\/code><\/h2><div class=\"widget maestro maestro-content-type-html hide-for-tablet-SP hide-for-mobile-SP\" id=\"maestro-636\"><p>--ADVERTISEMENT--<\/p><\/div><p>Before creating our dynamic theming example, let\u2019s understand the essential basics of custom properties.<\/p><p>A <a href=\"https:\/\/www.w3.org\/TR\/css-variables\/#custom-property\">custom property<\/a> is a property whose name starts with two hyphens (<strong>\u2014<\/strong>) like <code>--foo<\/code>. They define variables that can be referenced using <a href=\"https:\/\/www.w3.org\/TR\/css-variables\/#funcdef-var\"><code>var()<\/code><\/a>. Let\u2019s consider this example:<\/p><pre class=\"css language-css\">:root {&#13;\n  --bg-color: #000;&#13;\n  --text-color: #fff;&#13;\n}&#13;\n<\/pre><p>Defining custom properties within the <code>:root<\/code> selector means they are available in the global document space to all elements. <strong><code>:root<\/code><\/strong> is a CSS pseudo class which matches the root element of the document \u2014 the <code>&lt;html&gt;<\/code> element. It\u2019s similar to the <code>html<\/code> selector, but with higher <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/Specificity\">specificity<\/a>.<\/p><p>You can access the value of a <code>:root<\/code> custom property anywhere in the document:<\/p><pre class=\"css language-css\">div {&#13;\n  color: var(--text-color);&#13;\n  background-color: var(--bg-color);&#13;\n}&#13;\n<\/pre><p>You can also include a fallback value with your CSS variable. For example:<\/p><pre class=\"css language-css\">div {&#13;\n  color: var(--text-color, #000);&#13;\n  background-color: var(--bg-color, #fff);&#13;\n}&#13;\n<\/pre><p>If a custom property isn\u2019t defined, their fallback value is used instead.<\/p><p>Defining custom properties inside a CSS selector other than the <code>:root<\/code> or <code>html<\/code> selector makes the variable available to matching elements and their children.<\/p><h2 id=\"csscustompropertiesvspreprocessorvariables\">CSS Custom Properties vs Preprocessor Variables<\/h2><p>CSS pre-processors such as Sass are often used to aid front-end web development. Among the other useful features of preprocessors are variables. But what\u2019s the difference between Sass variables and CSS custom properties?<\/p><ul><li>CSS custom properties are natively parsed in modern browsers. Preprocessor variables require compilation into a standard CSS file and all variables are converted to values.<\/li>\n<li>Custom properties can be accessed and modified by JavaScript. Preprocessor variables are compiled once and only their final value is available on the client.<\/li>\n<\/ul><h2 id=\"writingasimplehtmlpage\">Writing a Simple HTML Page<\/h2><p>Let\u2019s start by creating a folder for our project:<\/p><pre class=\"bash language-bash\">$ mkdir css-variables-theming&#13;\n<\/pre><p>Next, add an <code>index.html<\/code> inside the project\u2019s folder:<\/p><pre class=\"bash language-bash\">$ cd css-variables-theming&#13;\n$ touch index.html&#13;\n<\/pre><p>And add the following content:<\/p><pre class=\"markup language-markup\">&lt;nav class=\"navbar\"&gt;Title&lt;\/nav&gt;&#13;\n&lt;div class=\"container\"&gt;&#13;\n  &lt;div&gt;&#13;\n    &lt;input type=\"button\" value=\"Light\/Dark\" id=\"toggle-theme\" \/&gt;&#13;\n  &lt;\/div&gt;&#13;\n    &lt;h2 class=\"title\"&gt;What is Lorem Ipsum?&lt;\/h2&gt;&#13;\n    &lt;p class=\"content\"&gt;Lorem Ipsum is simply dummy text of the printing and typesetting industry...&lt;\/p&gt;&#13;\n&lt;\/div&gt;&#13;\n&lt;footer&gt;&#13;\n  Copyright 2018&#13;\n&lt;\/footer&gt;&#13;\n<\/pre><p>We are adding a navigation bar using a <code>&lt;nav&gt;<\/code> tag, a footer, and a container <code>&lt;div&gt;<\/code> that contains a button (that will be used to switch between light and dark themes) and some dummy <em>Lorem Ipsum<\/em> text.<\/p><h2 id=\"writingbasiccssforourhtmlpage\">Writing Basic CSS for Our HTML Page<\/h2><p>Now let\u2019s style our page. In the same file using an inline <code>&lt;style&gt;<\/code> tag in the <code>&lt;head&gt;<\/code> add the following CSS styles:<\/p><pre class=\"markup language-markup\">&lt;style&gt;&#13;\n* {&#13;\n  margin: 0;&#13;\n}&#13;\nhtml{&#13;\n  height: 100%;&#13;\n}&#13;\nbody{&#13;\n  height: 100%;&#13;\n  font-family: -apple-system, BlinkMacSystemFont\u201cSegoe UI\u201d, \u201cRoboto\u201d, \u201cOxygen\u201d, \u201cUbuntu\u201d, \u201cCantarell\u201d,&#13;\n \u201cFira Sans\u201d, \u201cDroid Sans\u201d, \u201cHelvetica Neue\u201d,sans-serif;  &#13;\n  display: flex;&#13;\n  flex-direction: column;&#13;\n}&#13;\nnav{&#13;\n  background: hsl(350, 50%, 50%);&#13;\n  padding: 1.3rem;&#13;\n  color: hsl(350, 50%, 10%);&#13;\n}&#13;\n.container{&#13;\n  flex: 1;&#13;\n  background:hsl(350, 50%, 95%);&#13;\n  padding: 1rem;&#13;\n}&#13;\np.content{&#13;\n  padding: 0.7rem;&#13;\n  font-size: 0.7rem;&#13;\n  color: hsl(350, 50%, 50%);&#13;\n}&#13;\n.container h2.title{&#13;\n  padding: 1rem;&#13;\n  color: hsl(350, 50%, 20%);&#13;\n}&#13;\nfooter{&#13;\n  background: hsl(350, 93%, 88%);&#13;\n  padding: 1rem;&#13;\n}&#13;\ninput[type=button] {&#13;\n  color:hsl(350, 50%, 20%);&#13;\n  padding: 0.3rem;&#13;\n  font-size: 1rem;&#13;\n}&#13;\n&lt;\/style&gt;&#13;\n<\/pre><p>CSS3 HSL (Hue, Saturation, Lightness) notation is used to define colors. The hue is the angle on a color circle and the example uses 350 for red. All page colors use differing variations by changing the saturation (percentage of color) and lightness (percentage).<\/p><p>Using HSL allows us to easily try different main colors for the theme by only changing the hue value. We could also use a CSS variable for the hue value and switch the color theme by changing a single value in the stylesheet or dynamically altering it with JavaScript.<\/p><p>This is a screen shot of the page:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2018\/08\/1535598717example-page.png\" alt=\"Example HTML page\" width=\"294\" height=\"510\" class=\"aligncenter size-full wp-image-168070\" \/><\/p><p>Check out the related pen:<\/p><p data-height=\"300\" data-theme-id=\"6441\" data-slug-hash=\"yxVJod\" data-default-tab=\"html,result\" data-user=\"SitePoint\" data-pen-title=\"CSS Theming 1\" class=\"codepen\">See the Pen <a href=\"https:\/\/codepen.io\/SitePoint\/pen\/yxVJod\/\">CSS Theming 1<\/a> by SitePoint (<a href=\"https:\/\/codepen.io\/SitePoint\">@SitePoint<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p><p>Let\u2019s use a CSS variable for holding the value of the hue of all colors in the page. Add a global CSS variable in the <code>:root<\/code> selector at the top of the <code>&lt;style&gt;<\/code> tag:<\/p><pre class=\"css language-css\">:root{&#13;\n  --main-hue : 350;&#13;\n}&#13;\n<\/pre><p>Next, we replace all hard-coded <em>350<\/em> values in <em>hsl()<\/em> colors with the <code>--main-hue<\/code> variable. For example, this is the <em>nav<\/em> selector:<\/p><pre class=\"css language-css\">nav{&#13;\n  background: hsl(var(--main-hue) , 50%, 50%);&#13;\n  padding: 1.3rem;&#13;\n  color: hsl(var(--main-hue), 50%, 10%);&#13;\n}&#13;\n<\/pre><p>Now if you want to specify any color other than red, you can just assign the corresponding value to <code>--main-hue<\/code>. These are some examples:<\/p><pre class=\"css language-css\">:root{&#13;\n  --red-hue: 360;&#13;\n  --blue-hue: 240;&#13;\n  --green-hue: 120;&#13;\n  --main-hue : var(--red-hue);&#13;\n}&#13;\n<\/pre><p>We are defining three custom properties for red, blue and green, then assigning the <code>--red-hue<\/code> variable to <code>--main-hue<\/code>.<\/p><p>This a screen shot of pages with different values for <code>--main-hue<\/code>:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2018\/08\/1535599018different-themes.png\" alt=\"Pages with different themes\" width=\"882\" height=\"510\" class=\"aligncenter size-full wp-image-168072\" \/><\/p><p>CSS custom properties offer a couple of benefits:<\/p><ul><li>A value can be defined in a single place.<\/li>\n<li>That value can be named appropriately to aid maintenance.<\/li>\n<li>The value can be dynamically altered using JavaScript. For example, the <code>--main-hue<\/code> can be set to any value between 0 and 360.<\/li>\n<\/ul><p>Using JavaScript to dynamically set the value of <code>--main-hue<\/code> from a set of predefined values or user submitted value for hue (it should be between <em>0<\/em> and <em>360<\/em>) we can provide the user with many possibilities for colored themes.<\/p><p>The following line of code will set the value of <code>--main-hue<\/code> to 240 (blue):<\/p><pre class=\"javascript language-javascript\">document.documentElement.style.setProperty('--main-hue', 240);&#13;\n<\/pre><p>Check out the following pen, which shows a full example that allows you to dynamically switch between red, blue and green colored themes:<\/p><p data-height=\"300\" data-theme-id=\"6441\" data-slug-hash=\"yxVJEB\" data-default-tab=\"css,result\" data-user=\"SitePoint\" data-pen-title=\"CSS Theming 2\" class=\"codepen\">See the Pen <a href=\"https:\/\/codepen.io\/SitePoint\/pen\/yxVJEB\/\">CSS Theming 2<\/a> by SitePoint (<a href=\"https:\/\/codepen.io\/SitePoint\">@SitePoint<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p><p>This is a screen shot of the page from the pen:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2018\/08\/1535599604theme-variations.png\" alt=\"A page with different themes\" width=\"933\" height=\"516\" class=\"aligncenter size-full wp-image-168073\" \/><\/p><h2 id=\"addingacssdarktheme\">Adding a CSS Dark Theme<\/h2><p>Now let\u2019s provide a dark theme for this page. For more control over the colors of different entities, we need to add more variables.<\/p><p>Going through the page\u2019s styles, we can replace all HSL colors in different selectors with variables after defining custom properties for the corresponding colors in <code>:root<\/code>:<\/p><pre class=\"css language-css\">:root{&#13;\n  \/*...*\/&#13;\n  --nav-bg-color: hsl(var(--main-hue) , 50%, 50%);&#13;\n  --nav-text-color: hsl(var(--main-hue), 50%, 10%);&#13;\n  --container-bg-color: hsl(var(--main-hue) , 50%, 95%);&#13;\n  --content-text-color: hsl(var(--main-hue) , 50%, 50%);&#13;\n  --title-color: hsl(var(--main-hue) , 50%, 20%);&#13;\n  --footer-bg-color: hsl(var(--main-hue) , 93%, 88%);&#13;\n  --button-text-color: hsl(var(--main-hue), 50%, 20%);&#13;\n}&#13;\n<\/pre><p>Appropriate names for the custom properties have been used. For example, <code>--nav-bg-color<\/code> refers to <em>the color of the nav background<\/em>, while <code>--nav-text-color<\/code> refers to <em>the color of nav foreground\/text<\/em>.<\/p><p>Now duplicate the <code>:root<\/code> selector with its content, but add a theme attribute with a <em>dark<\/em> value:<\/p><pre class=\"css language-css\">:root[theme='dark']{&#13;\n  \/*...*\/&#13;\n}&#13;\n<\/pre><p>This theme will be activated if a <em>theme<\/em> attribute with a <em>dark<\/em> value is added to the <code>&lt;html&gt;<\/code> element.<\/p><p>We can now play with the values of these variables manually, by reducing the lightness value of the HSL colors to provide a dark theme, or we can use other techniques such as CSS filters like <code>invert()<\/code> and <code>brightness()<\/code>, which are commonly used to adjust the rendering of images but can also be used with any other element.<\/p><p>Add the following code to <code>:root[theme='dark']<\/code>:<\/p><pre class=\"css language-css\">:root[theme='dark'] {&#13;\n  --red-hue: 360;&#13;\n  --blue-hue: 240;&#13;\n  --green-hue: 120;&#13;\n  --main-hue: var(--blue-hue);&#13;\n  --nav-bg-color: hsl(var(--main-hue), 50%, 90%);&#13;\n  --nav-text-color: hsl(var(--main-hue), 50%, 10%);&#13;\n  --container-bg-color: hsl(var(--main-hue), 50%, 95%);&#13;\n  --content-text-color: hsl(var(--main-hue), 50%, 50%);&#13;\n  --title-color: hsl(--main-hue, 50%, 20%);&#13;\n  --footer-bg-color: hsl(var(--main-hue), 93%, 88%);&#13;\n  --button-text-color: hsl(var(--main-hue), 50%, 20%);&#13;\n  filter: invert(1) brightness(0.6);&#13;\n}&#13;\n<\/pre><p>The <code>invert()<\/code> filter inverts all the colors in the selected elements (every element in this case). The value of inversion can be specified in percentage or number. A value of <code>100%<\/code> or <code>1<\/code> will completely invert the hue, saturation, and lightness values of the element.<\/p><p>The <code>brightness()<\/code> filter makes an element brighter or darker. A value of <code>0<\/code> results in a completely dark element.<\/p><p>The <code>invert()<\/code> filter makes some elements very bright. These are toned down by setting <code>brightness(0.6)<\/code>.<\/p><p>A dark theme with different degrees of darkness:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2018\/08\/1535599870dark-theme.png\" alt=\"A dark theme\" width=\"933\" height=\"516\" class=\"aligncenter size-full wp-image-168075\" \/><\/p><h2 id=\"switchingthemeswithjavascript\">Switching Themes with JavaScript<\/h2><p>Let\u2019s now use JavaScript to switch between the dark and light themes when a user clicks the <em>Dark\/Light<\/em> button. In <code>index.html<\/code> add an inline <code>&lt;script&gt;<\/code> before the closing <code>&lt;\/body&gt;<\/code> with the following code:<\/p><pre class=\"javascript language-javascript\">const toggleBtn = document.querySelector(\"#toggle-theme\");&#13;\ntoggleBtn.addEventListener('click', e =&gt; {&#13;\n  console.log(\"Switching theme\");&#13;\n  if(document.documentElement.hasAttribute('theme')){&#13;\n    document.documentElement.removeAttribute('theme');&#13;\n  }&#13;\n  else{&#13;\n    document.documentElement.setAttribute('theme', 'dark');&#13;\n  }&#13;\n});&#13;\n<\/pre><p><strong>Document.documentElement<\/strong> refers to the the root DOM Element of the document \u2014 that is, <code>&lt;html&gt;<\/code>. This code checks for the existence of a <em>theme<\/em> attribute using the <code>.hasAttribute()<\/code> method and adds the attribute with a <em>dark<\/em> value if it doesn\u2019t exist, causing the switch to the dark theme. Otherwise, it removes the attribute, which results in switching to the light theme.<\/p><h2 id=\"changingcsscustompropertieswithjavascript\">Changing CSS Custom Properties with JavaScript<\/h2><p>Using JavaScript, we can access custom properties and change their values dynamically. In our example, we hard-coded the brightness value, but it could be dynamically changed. First, add a slider input in the HTML page next to the <em>dark\/light<\/em> button:<\/p><pre class=\"markup language-markup\"> &lt;input type=\"range\" id=\"darknessSlider\" name=\"darkness\" value=\"1\" min=\"0.3\" max=\"1\" step=\"0.1\" \/&gt;&#13;\n<\/pre><p>The slider starts at <code>1<\/code> and allows the user to reduce it to <code>0.3<\/code> in steps of <code>0.1<\/code>.<\/p><p>Next, add a custom property for the darkness amount with an initial value of <code>1<\/code> in <code>:root[theme='dark']<\/code>:<\/p><pre class=\"css language-css\">:root[theme='dark']{&#13;\n  \/*...*\/&#13;\n  --theme-darkness: 1;&#13;\n}&#13;\n<\/pre><p>Change the <code>brightness<\/code> filter to this custom property instead of the hard-coded value:<\/p><pre class=\"css language-css\">filter: invert(1) brightness(var(--theme-darkness));&#13;\n<\/pre><p>Finally, add the following code to synchronize the value of <code>--theme-darkness<\/code> with the slider value:<\/p><pre class=\"javascript language-javascript\">const darknessSlider = document.querySelector(\"#darknessSlider\");&#13;\ndarknessSlider.addEventListener('change', (e)=&gt;{&#13;\n  const val = darknessSlider.value&#13;\n  document.documentElement.style.setProperty('--theme-darkness', val);&#13;\n});&#13;\n<\/pre><p>We\u2019re listening for the change event of the slider and setting the value of <code>--theme-darkness<\/code> accordingly using the <code>setProperty()<\/code> method.<\/p><p>We can also apply the <code>brightness<\/code> filter to the light theme. Add the <code>--theme-darkness<\/code> custom property in the top of the <code>:root<\/code> selector:<\/p><pre class=\"css language-css\">:root{&#13;\n  \/*...*\/&#13;\n  --theme-darkness: 1;&#13;\n}&#13;\n<\/pre><p>Then add a <code>brightness<\/code> filter in the bottom of the same selector:<\/p><pre class=\"css language-css\">:root{&#13;\n  \/*...*\/&#13;\n  filter: brightness(var(--theme-darkness));&#13;\n}&#13;\n<\/pre><p>You can find the code of this example in the following pen:<\/p><p data-height=\"300\" data-theme-id=\"6441\" data-slug-hash=\"eLBzxJ\" data-default-tab=\"css,result\" data-user=\"SitePoint\" data-pen-title=\"CSS Theming\" class=\"codepen\">See the Pen <a href=\"https:\/\/codepen.io\/SitePoint\/pen\/eLBzxJ\/\">CSS Theming<\/a> by SitePoint (<a href=\"https:\/\/codepen.io\/SitePoint\">@SitePoint<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p><p>Here\u2019s a screenshot of the dark theme of the final example:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2018\/08\/1535600202dark-theme-screenshot.png\" alt=\"A screen shot of the dark theme\" width=\"933\" height=\"516\" class=\"aligncenter size-full wp-image-168076\" \/><\/p><p>Here\u2019s a screenshot of the light theme:<\/p><p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2018\/08\/1535600271light-theme-screenshot.png\" alt=\"A screenshot of the light theme\" width=\"933\" height=\"516\" class=\"aligncenter size-full wp-image-168078\" \/><\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>In this tutorial, we\u2019ve seen how to use CSS custom properties to create themes and switch dynamically between them. We\u2019ve used the HSL color scheme, which allows us to specify colors with hue, saturation and lightness values and CSS filters (<code>invert<\/code> and <code>brightness<\/code>) to create a dark version of a light theme.<\/p><p>Here are some links for further reading if you want to learn more about CSS theming:<\/p><ul><li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/Using_CSS_variables#Basic_usage\">Using CSS custom properties (variables)<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/HSL_and_HSV\">HSL and HSV<\/a><\/li>\n<li><a href=\"https:\/\/css-tricks.com\/css-custom-properties-theming\/\">CSS Custom Properties and Theming<\/a><\/li>\n<li><a href=\"https:\/\/medium.com\/@mwichary\/dark-theme-in-a-day-3518dde2955a\">Dark theme in a day<\/a><\/li>\n<\/ul>","created_at":"2018-11-09T18:24:53+0000","updated_at":"2018-11-09T18:28:34+0000","published_at":"2018-08-30T16:00:00+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"www.sitepoint.com","preview_picture":"https:\/\/www.sitepoint.com\/wp-content\/themes\/sitepoint\/assets\/images\/icon.html-css.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["javascript","procgen"],"is_public":false,"id":675,"title":"The HTML Canvas Guide","url":"https:\/\/flaviocopes.com\/canvas\/","content":"<p><img src=\"https:\/\/flaviocopes.com\/canvas\/banner.png\" alt=\"Banner\" \/><\/p><p>The HTML canvas is an HTML tag, <code>&lt;canvas&gt;<\/code>, which is an element where we can draw to using the Canvas API.<\/p>\n<h2 id=\"create-a-canvas\">Create a canvas<\/h2>\n<p>Creating a canvas is as simple as dropping a <code>&lt;canvas&gt;&lt;\/canvas&gt;<\/code> in a blank HTML file:<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2009.23.22.png\" alt=\"\" \/><\/p>\n<p>You don\u2019t see anything in the page because the canvas is an invisible element. Let\u2019s add some border:<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2009.23.43.png\" alt=\"\" \/><\/p>\n<p>Chrome automatically adds an 8px margin to the <code>body<\/code> element. This is why our border looks like a frame, and you can remove that margin by setting<\/p>\n<pre class=\"language-css\">body {\n  margin: 0;\n}\n<\/pre>\n<p>We\u2019ll leave the default for now.<\/p>\n<p>Our canvas is now reachable from JavaScript using the <a href=\"https:\/\/flaviocopes.com\/selectors-api\/\">DOM Selectors API<\/a>, so we can use <code>document.querySelector()<\/code>:<\/p>\n<pre class=\"language-js\">const canvas = document.querySelector('canvas')\n<\/pre>\n<h2 id=\"change-the-background-color-of-a-canvas\">Change the background color of a canvas<\/h2>\n<p>You do that in CSS:<\/p>\n<pre class=\"language-css\">canvas {\n  background-color: lightblue;\n}\n<\/pre>\n<h2 id=\"resize-a-canvas\">Resize a canvas<\/h2>\n<p>You can set the width and height in CSS:<\/p>\n<pre class=\"language-css\">canvas {\n  border: 1px solid black;\n  width: 100%;\n  height: 100%;\n}\n<\/pre>\n<p>and in this way the canvas will expand to fill all the outer element size.<\/p>\n<p>If you put the canvas as a first level element in the HTML, the above code will expand the canvas to fit the entire body.<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2009.27.05.png\" alt=\"\" \/><\/p>\n<p>The body is not filling the entire window size. To fill the entire page instead we need to use JavaScript:<\/p>\n<pre class=\"language-js\">canvas.width = window.innerWidth\ncanvas.height = window.innerHeight\n<\/pre>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2009.51.51.png\" alt=\"\" \/><\/p>\n<p>If you now remove the body margin and set the background of the canvas using CSS, we can fill our entire page with the canvas and we can start drawing on it:<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2009.54.20.png\" alt=\"\" \/><\/p>\n<p>If the window resizes we need to recalculate the canvas width as well, using a <strong>debounce<\/strong> to avoid calling too many times our canvas resizing (the <code>resize<\/code> event can be called hundreds of times as you move the window with the mouse, for example):<\/p>\n<pre class=\"language-js\">const debounce = (func) =&gt; {\n  let timer\n  return (event) =&gt; {\n    if (timer) { clearTimeout(timer) }\n    timer = setTimeout(func, 100, event)\n  }\n}\nwindow.addEventListener('resize', debounce(() =&gt; {\n  canvas.width = window.innerWidth\n  canvas.height = window.innerHeight\n}))\n<\/pre>\n<h2 id=\"get-a-context-from-the-canvas\">Get a context from the canvas<\/h2>\n<p>We want to draw to the canvas.<\/p>\n<p>To do this, we need to get a context:<\/p>\n<pre class=\"language-js\">const c = canvas.getContext('2d')\n<\/pre>\n<blockquote>\n<p>Some assign the context to a variable named <code>c<\/code>, some <code>ctx<\/code> - it\u2019s a common way to shortcut \u201ccontext\u201d<\/p>\n<\/blockquote>\n<p>The <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/HTMLCanvasElement\/getContext\"><code>getContext()<\/code><\/a> method returns a drawing context on the canvas, according to the type that you pass as parameter.<\/p>\n<p>Valid values are<\/p>\n<ul><li><code>2d<\/code>, the one we\u2019ll use<\/li>\n<li><code>webgl<\/code> to use WebGL version 1<\/li>\n<li><code>webgl2<\/code> to use WebGL version 2<\/li>\n<li><code>bitmaprenderer<\/code> to use with <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/ImageBitmap\">ImageBitmap<\/a><\/li>\n<\/ul><p>Based on the context type, you can pass a second parameter to <code>getContext()<\/code> to specify additional options.<\/p>\n<p>In the case of the <code>2d<\/code> context, we basically have one parameter we can use in all browsers, and it\u2019s <code>alpha<\/code>, a boolean that defaults to true. If set to false, the browser knows the canvas does not have a transparent background and can speed up rendering.<\/p>\n<h2 id=\"draw-elements-to-a-canvas\">Draw elements to a canvas<\/h2>\n<p>With the context we can now draw elements.<\/p>\n<p>We have several methods to do so. We can draw:<\/p>\n<ul><li>text<\/li>\n<li>lines<\/li>\n<li>rectangles<\/li>\n<li>paths<\/li>\n<li>images<\/li>\n<\/ul><p>and for each of those elements we can alter the fill, the stroke, the gradient, the pattern, the shadow, rotate them, scale and perform a lot of operations.<\/p>\n<p>Let\u2019s start with the simplest thing: a rectangle.<\/p>\n<p>The <code>fillRect(x, y, width, height)<\/code> method serves this purpose:<\/p>\n<pre class=\"language-js\">c.fillRect(100, 100, 100, 100)\n<\/pre>\n<p>This is going to draw a black rectangle of 100 x 100 pixels, starting from position x 100 and y 100:<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2010.39.01.png\" alt=\"\" \/><\/p>\n<p>You can color the rectangle by using the <code>fillStyle()<\/code> method, passing any valid CSS color string:<\/p>\n<pre class=\"language-js\">c.fillStyle = 'white'\nc.fillRect(100, 100, 100, 100)\n<\/pre>\n<p>You can get creative now and draw many things in this way:<\/p>\n<pre class=\"language-js\">for (let i = 0; i &lt; 60; i++) {\n  for (let j = 0; j &lt; 60; j++) {\n    c.fillStyle = `rgb(${i * 5}, ${j * 5}, ${(i+j) * 50})`\n    c.fillRect(j * 20, i * 20, 10, 10)\n  }\n}\n<\/pre>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2010.47.55.png\" alt=\"\" \/><\/p>\n<p>or<\/p>\n<pre class=\"language-js\">for (let i = 0; i &lt; 60; i++) {\n  for (let j = 0; j &lt; 60; j++) {\n    c.fillStyle = `rgb(${i * 5}, ${j * 5}, ${(i+j) * 50})`\n    c.fillRect(j * 20, i * 20, 20, 20)\n  }\n}\n<\/pre>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2010.48.21.png\" alt=\"\" \/><\/p>\n<h2 id=\"drawing-elements\">Drawing elements<\/h2>\n<p>As mentioned you can draw many things:<\/p>\n<ul><li>text<\/li>\n<li>lines<\/li>\n<li>rectangles<\/li>\n<li>paths<\/li>\n<li>images<\/li>\n<\/ul><p>Let\u2019s just see a few of them, rectangles and text, to get the gist of how things work. You can find the API for all the rest that you need <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/CanvasRenderingContext2D\">here<\/a>.<\/p>\n<h3 id=\"changing-the-colors\">Changing the colors<\/h3>\n<p>Use the <code>fillStyle<\/code> and <code>strokeStyle<\/code> properties to change the fill and stroke colors of any figure. They accept any valid CSS color, including strings and RGB calculations:<\/p>\n<pre class=\"language-js\">c.strokeStyle = `rgb(255, 255, 255)`\nc.fillStyle = `white`\n<\/pre>\n<h3 id=\"rectangles\">Rectangles<\/h3>\n<ul><li>clearRect(x, y, width, height)<\/li>\n<li>fillRect(x, y, width, height)<\/li>\n<li>strokeRect(x, y, width, height)<\/li>\n<\/ul><p>We saw <code>fillRect()<\/code> in the previous section. <code>strokeRect()<\/code> is similar in how it\u2019s called, but instead of filling a rect, it just draws the stroke using the current stroke style (which can be changed using the <code>strokeStyle<\/code> context property):<\/p>\n<pre class=\"language-js\">const c = canvas.getContext('2d')\nfor (let i = 0; i &lt; 61; i++) {\n  for (let j = 0; j &lt; 61; j++) {\n    c.strokeStyle = `rgb(${i * 5}, ${j * 5}, ${(i+j) * 50})`\n    c.strokeRect(j * 20, i * 20, 20, 20)\n  }\n}\n<\/pre>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2011.00.10.png\" alt=\"\" \/><\/p>\n<p><code>clearRect()<\/code> sets an area as transparent:<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2011.02.47.png\" alt=\"\" \/><\/p>\n<h3 id=\"text\">Text<\/h3>\n<p>Drawing text is similar to rectangles. You have 2 methods<\/p>\n<ul><li>fillText(text, x, y)<\/li>\n<li>strokeText(text, x, y)<\/li>\n<\/ul><p>which let you write text on the canvas.<\/p>\n<p><code>x<\/code> and <code>y<\/code> refer to the bottom-left corner.<\/p>\n<p>You change the font family and size using the <code>font<\/code> property of the canvas:<\/p>\n<pre class=\"language-js\">c.font = '148px Courier New'\n<\/pre>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2011.15.51.png\" alt=\"\" \/><\/p>\n<p>There are other properties you can change, related to text (* = default):<\/p>\n<ul><li><code>textAlign<\/code> (start*, end, left, right, center)<\/li>\n<li><code>textBaseline<\/code> (top, hanging, middle, alphabetic*, ideographic, bottom)<\/li>\n<li><code>direction<\/code> (ltr, rtl, inherit*)<\/li>\n<\/ul><h3 id=\"lines\">Lines<\/h3>\n<p>To draw a line you first call the <code>beginPath()<\/code> method, then you provide a starting point with <code>moveTo(x, y)<\/code>, and then you call <code>lineTo(x, y)<\/code> to make the line to that new coordinates set. You finally call <code>stroke()<\/code>:<\/p>\n<pre class=\"language-js\">c.beginPath()\nc.moveTo(10, 10)\nc.lineTo(300, 300)\nc.stroke()\n<\/pre>\n<p>The line is going to be colored according to the <code>c.strokeStyle<\/code> property value.<\/p>\n<h2 id=\"a-more-complex-example\">A more complex example<\/h2>\n<p>This code creates a canvas that generates 800 circles:<\/p>\n<p><img src=\"https:\/\/flaviocopes.com\/canvas\/Screen%20Shot%202018-10-28%20at%2014.38.54.png\" alt=\"\" \/><\/p>\n<p>Every circle is perfectly contained in the canvas, and its radius is randomized.<\/p>\n<p>Any time you resize the window, the elements are regenerated.<\/p>\n<p>You can play around with on <a href=\"https:\/\/codepen.io\/flaviocopes\/pen\/qJvyVp\">Codepen<\/a>.<\/p>\n<pre class=\"language-js\">const canvas = document.querySelector('canvas')\ncanvas.width = window.innerWidth\ncanvas.height = window.innerHeight\nconst c = canvas.getContext('2d')\nconst circlesCount = 800\nconst colorArray = [\n  '#046975',\n  '#2EA1D4',\n  '#3BCC2A',\n  '#FFDF59',\n  '#FF1D47'\n]\nconst debounce = (func) =&gt; {\n  let timer\n  return (event) =&gt; {\n    if (timer) { clearTimeout(timer) }\n    timer = setTimeout(func, 100, event)\n  }\n}\nwindow.addEventListener('resize', debounce(() =&gt; {\n  canvas.width = window.innerWidth\n  canvas.height = window.innerHeight\n  init()\n}))\nconst init = () =&gt; {\n  for (let i = 0; i &lt; circlesCount; i++) {\n    const radius = Math.random() * 20 + 1\n    const x = Math.random() * (innerWidth - radius  * 2) + radius\n    const y = Math.random() * (innerHeight - radius  * 2) + radius\n    const dx = (Math.random() - 0.5) * 2\n    const dy = (Math.random() - 0.5) * 2\n    const circle = new Circle(x, y, dx, dy, radius)\n    circle.draw()\n  }\n}\nconst Circle = function(x, y, dx, dy, radius) {\n  this.x = x\n  this.y = y\n  this.dx = dx\n  this.dy = dy\n  this.radius = radius\n  this.minRadius = radius\n  this.color = colorArray[Math.floor(Math.random() * colorArray.length)]\n  this.draw = function() {\n    c.beginPath()\n    c.arc(this.x, this.y, this.radius, 0, Math.PI * 2, false)\n    c.strokeStyle = 'black'\n    c.stroke()\n    c.fillStyle = this.color\n    c.fill()\n  }\n}\ninit()\n<\/pre>\n<h2 id=\"another-example-animating-elements-on-the-canvas\">Another example: animating elements on the canvas<\/h2>\n<p>Based on the example above, we animate the elements using a loop. Every circle has its own \u201clife\u201d and moves within the borders of the canvas. When the border is reached it bounces back:<\/p>\n<p data-height=\"265\" data-theme-id=\"0\" data-slug-hash=\"MPxBBb\" data-default-tab=\"js,result\" data-user=\"flaviocopes\" data-pen-title=\"HTML Canvas fun with circles, not interactive\" data-preview=\"true\" class=\"codepen\">See the Pen <a href=\"https:\/\/codepen.io\/flaviocopes\/pen\/MPxBBb\/\">HTML Canvas fun with circles, not interactive<\/a> by Flavio Copes (<a href=\"https:\/\/codepen.io\/flaviocopes\">@flaviocopes<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p>\n<p>We achieve this by using <a href=\"https:\/\/flaviocopes.com\/requestanimationframe\/\"><code>requestAnimationFrame()<\/code><\/a> and slightly moving the image at every frame rendering iteration.<\/p>\n<h2 id=\"interact-with-the-elements-on-the-canvas\">Interact with the elements on the canvas<\/h2>\n<p>Here is the above example expanded to let you interact with the circles using the mouse.<\/p>\n<p>When you hover the canvas, the items near your mouse will increase in size, and they will return back to normal when you move somewhere else:<\/p>\n<p data-height=\"265\" data-theme-id=\"0\" data-slug-hash=\"YJgarj\" data-default-tab=\"js,result\" data-user=\"flaviocopes\" data-pen-title=\"HTML Canvas fun with circles\" data-preview=\"true\" class=\"codepen\">See the Pen <a href=\"https:\/\/codepen.io\/flaviocopes\/pen\/YJgarj\/\">HTML Canvas fun with circles<\/a> by Flavio Copes (<a href=\"https:\/\/codepen.io\/flaviocopes\">@flaviocopes<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p>\n<p>How does it work? Well, first I track the mouse position using 2 variables:<\/p>\n<pre class=\"language-js\">let mousex = undefined\nlet mousey = undefined\nwindow.addEventListener('mousemove', (e) =&gt; {\n  mousex = e.x\n  mousey = e.y\n})\n<\/pre>\n<p>Then we use those variables inside the update() method of Circle, to determine if the radius should increase (or decrease):<\/p>\n<pre class=\"language-js\">if (mousex - this.x &lt; distanceFromMouse &amp;&amp; mousex - this.x &gt; -distanceFromMouse &amp;&amp; mousey - this.y &lt; distanceFromMouse &amp;&amp; mousey - this.y &gt; -distanceFromMouse) {\n  if (this.radius &lt; maxRadius) this.radius += 1\n} else {\n  if (this.radius &gt; this.minRadius) this.radius -= 1\n}\n<\/pre>\n<p><code>distanceFromMouse<\/code> is a value expressed in pixels (set to 200) that defines how far we want the circles to react to the mouse.<\/p>\n<h2 id=\"performance\">Performance<\/h2>\n<p>If you try to edit those projects above and add a bunch more circles and moving parts, you\u2019ll probably notice performance issues. Browsers consume a lot of energy to render the canvas with animations and interactions, so pay attention so that the experience is not ruined on less performant machines than yours.<\/p>\n<p>In particular I had issues when trying to create a similar experience with emojis rather than circles, and I found that text takes a lot more power to render, and so it was sluggish pretty quickly.<\/p>\n<p data-height=\"265\" data-theme-id=\"0\" data-slug-hash=\"vVPagr\" data-default-tab=\"js,result\" data-user=\"flaviocopes\" data-pen-title=\"HTML Canvas fun with Emojis\" data-preview=\"true\" class=\"codepen\">See the Pen <a href=\"https:\/\/codepen.io\/flaviocopes\/pen\/vVPagr\/\">HTML Canvas fun with Emojis<\/a> by Flavio Copes (<a href=\"https:\/\/codepen.io\/flaviocopes\">@flaviocopes<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p>\n<p><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Canvas_API\/Tutorial\/Optimizing_canvas\">This page on MDN<\/a> lists many performance tips.<\/p>\n<h2 id=\"closing-words\">Closing words<\/h2>\n<p>This was just an introduction to the possibilities of Canvas, an amazing tool that you can use to create incredible experiences on your web pages.<\/p>","created_at":"2018-11-09T18:17:32+0000","updated_at":"2018-11-09T18:17:47+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"flaviocopes.com","preview_picture":"https:\/\/flaviocopes.com\/canvas\/banner.png","http_status":"200"},{"is_archived":0,"is_starred":0,"tags":["json-ld","activitypub","activitystreams"],"is_public":false,"id":674,"title":"JSON-LD: Building Meaningful Data APIs - via @codeship","url":"https:\/\/blog.codeship.com\/json-ld-building-meaningful-data-apis\/","content":"<p> Reading Time: 7 minutes<\/p><p>Everybody loves <a href=\"http:\/\/json.org\/\">JSON<\/a>!<\/p><p>However, JSON by itself is pretty meaningless. Well. It has meaning, but only to the original creator of that format. They can share that meaning via documentation, conversation, or usage, but those often exist in places far removed from the JSON document you\u2019re looking at.<\/p><p>What if the meaning itself was encoded directly into the document? What if every key in a JSON document had its own unique identity which you could look up (because web!) and read about its meaning? Wouldn\u2019t that be slick?<\/p><p>Well. It\u2019s quite possible, actually.<\/p><p>Here\u2019s what I mean. Below is a pretty typical bit of \u201cperson\u201d information written in JSON with English key names (to match this article\u2019s content):<\/p><pre class=\"language-bash\">{\n  \"first_name\": \"Benjamin\",\n  \"last_name\": \"Young\",\n  \"alias\": \"BigBlueHat\",\n  \"email\": \"byoung@bigbluehat.com\"\n}<\/pre><p>You likely know what all that means (more or less). However, your software doesn\u2019t, unless you coded it exactly to those string names. While there\u2019s some possibility we picked the same keys (yay us!), there\u2019s a higher likelihood we didn\u2019t. We probably fix our <code>:coffee:<\/code> differently too. Who knows.<\/p><p>Currently, we\u2019d need to write some code that took my JSON and turned the <code>first_name<\/code> key into your <code>firstName<\/code> key (or into <code>first<\/code> which is inside your <code>name<\/code> object or into <code>given<\/code> or\u2026 whatever). We\u2019d then need to do that again the next time we found someone else with information about a person that used different key names. Rinse. Repeat.<\/p><p>What if we could match <code>first_name<\/code>, <code>name.first<\/code>, <code>given<\/code>, and even an obfuscated <code>sadfeqpoif<\/code> key into a meaningful <em>Thing<\/em> which made each of these \u201cmean\u201d the same thing? It can be done!<\/p><p><img src=\"https:\/\/cloudant.com\/wp-content\/uploads\/json-ld-button.png\" alt=\"JSON-LD logo\" \/><\/p><p>Meet <a href=\"http:\/\/json-ld.org\/\">JSON-LD<\/a>! JSON-LD stands for \u201cJSON for Linked Data.\u201d It\u2019s a specification for encoding contextualized meaning into otherwise meaningless JSON documents.<\/p><p>Here\u2019s an example:<\/p><pre class=\"language-bash\">{\n  \"@context\": {\n    \"@vocab\": \"http:\/\/schema.org\/\",\n    \"first_name\": \"givenName\",\n    \"last_name\": \"familyName\",\n    \"alias\": \"alternateName\",\n    \"email\": \"email\"\n  },\n  \"first_name\": \"Benjamin\",\n  \"last_name\": \"Young\",\n  \"alias\": \"BigBlueHat\",\n  \"email\": \"byoung@bigbluehat.com\"\n}<\/pre><p>In this <em>upgraded<\/em> version of the earlier JSON document, you\u2019ll see a new <code>@context<\/code> object. This object handles the mapping from my ad hoc terminology to the <a href=\"http:\/\/schema.org\/\">Schema.org<\/a> vocabulary which is referenced via value of <code>@vocab<\/code>.<\/p><p>The key\/value pairs below the <code>@vocab<\/code> key map directly to meaningful URLs at <code>http:\/\/schema.org\/<\/code>. Now, my keys have meaning. Each key now has:<\/p><ul><li>an <strong>identity<\/strong> in the form of a URL: <code>last_name<\/code> now maps to <a href=\"http:\/\/schema.org\/familyName\">http:\/\/schema.org\/familyName<\/a> for instance.<\/li><li>a <strong>definition<\/strong> found by visiting the URL for <code>familyName<\/code>: <em>\u201cFamily name. In the U.S., the last name of a Person. This can be used along with givenName instead of the name property.\u201d<\/em><\/li><\/ul><p>Pretty handy, no? Now. If you found that JSON \u2014 or got it back from my API \u2014 you could know what the keys meant by looking up their URLs.<\/p><h2>Mapping to Meaning<\/h2><p>Knowing what keys mean within a JSON document is already an upgrade over the status quo, but let\u2019s not stop there!<\/p><p>Here\u2019s an example of the same JSON using the Schema.org terms (key names) directly:<\/p><pre class=\"language-bash\">{\n  \"@context\": \"http:\/\/schema.org\/\",\n  \"givenName\": \"Benjamin\",\n  \"familyName\": \"Young\",\n  \"alternateName\": \"BigBlueHat\",\n  \"email\": \"byoung@bigbluehat.com\"\n}<\/pre><p>This Schema.org JSON example has the same meaning as the earlier version. \u201cBigBlueHat\u201d in both cases means an <a href=\"http:\/\/schema.org\/alternateName\">alternateName<\/a> for Benjamin Young.<\/p><p>In the world of JSON-LD, my JSON with <code>first_name<\/code> and this JSON with <code>givenName<\/code> are equivalent, identical, the exact same, they match, and son. Crazy, right?<\/p><h2>Coding for the @context<\/h2><p>So wait. We\u2019ve got two different JSON documents which mean the same thing. Fabulous. However, most current JSON code for processing these likely looks a bit like this:<\/p><pre class=\"language-javascript\">var doc = {...}; \/\/ one of the documents above\nvar first = doc.first_name || doc.givenName;\nconsole.log('First Name', first);<\/pre><p>How does the <code>@context<\/code> object help get around that? We\u2019re still just looking for the strings not the Things.<\/p><p>We have two options:<\/p><ul><li>expansion<\/li><li>compaction<\/li><\/ul><h3>Expansion<\/h3><p>Expansion in JSON-LD parlance means taking those meaningful, URL-based names and making them (at least within the processing code) actually <em>be<\/em> the URLs. Here\u2019s the output from the expansion process done on both the above contextualized examples (remember! They\u2019re identicial now!):<\/p><pre class=\"language-ruby\">[\n  {\n    \"http:\/\/schema.org\/alternateName\": [\n      {\n        \"@value\": \"BigBlueHat\"\n      }\n    ],\n    \"http:\/\/schema.org\/email\": [\n      {\n        \"@value\": \"byoung@bigbluehat.com\"\n      }\n    ],\n    \"http:\/\/schema.org\/givenName\": [\n      {\n        \"@value\": \"Benjamin\"\n      }\n    ],\n    \"http:\/\/schema.org\/familyName\": [\n      {\n        \"@value\": \"Young\"\n      }\n    ]\n  }\n]<\/pre><p>Yeah. Sorry. This is the sausage making bit.<\/p><p>The above is the internal, identical representation in a JSON-LD processing system of both the above documents. It\u2019s structured that way because there\u2019s much more data that can (and likely will, in a real-life example) be added to those arrays, objects, etc.<\/p><p>However, you\u2019re likely not currently inspired to switch all your JSON processing code to use this format. No worries! That\u2019s where <em>compaction<\/em> comes in!<\/p><h3>Compaction<\/h3><p>JSON-LD processors will take the above two example docs, turn them into that <em>expanded<\/em> format, and then (if asked) turn them into a much more human-friendly <em>compacted<\/em> variation that looks like this:<\/p><pre class=\"language-bash\">{\n  \"@context\": {\n    \"@vocab\": \"http:\/\/schema.org\/\",\n    \"first_name\": \"givenName\",\n    \"last_name\": \"familyName\",\n    \"alias\": \"alternateName\",\n    \"email\": \"email\"\n  },\n  \"alias\": \"BigBlueHat\",\n  \"email\": \"byoung@bigbluehat.com\",\n  \"last_name\": \"Young\",\n  \"first_name\": \"Benjamin\"\n}<\/pre><p>\u201cWait, what?\u201d I hear you say. \u201cYou just copy\/pasted the first example again! You, just now.\u201d<\/p><p>Actually, I didn\u2019t. Don\u2019t trust me? Check it out in the <a href=\"http:\/\/tinyurl.com\/z77crzk\">JSON-LD Playground<\/a>. Here\u2019s a screenshot for the doubters who don\u2019t want to click:<\/p><p><img src=\"http:\/\/1npo9l3lml0zvr6w62acc3t1-wpengine.netdna-ssl.com\/wp-content\/uploads\/2016\/03\/Recontextualized-JSON-LD.png\" alt=\"Recontextualized JSON-LD\" width=\"1192\" height=\"731\" class=\"aligncenter size-full wp-image-3142\" srcset=\"https:\/\/1npo9l3lml0zvr6w62acc3t1-wpengine.netdna-ssl.com\/wp-content\/uploads\/2016\/03\/Recontextualized-JSON-LD.png 1192w, https:\/\/1npo9l3lml0zvr6w62acc3t1-wpengine.netdna-ssl.com\/wp-content\/uploads\/2016\/03\/Recontextualized-JSON-LD-300x184.png 300w, https:\/\/1npo9l3lml0zvr6w62acc3t1-wpengine.netdna-ssl.com\/wp-content\/uploads\/2016\/03\/Recontextualized-JSON-LD-768x471.png 768w, https:\/\/1npo9l3lml0zvr6w62acc3t1-wpengine.netdna-ssl.com\/wp-content\/uploads\/2016\/03\/Recontextualized-JSON-LD-1024x628.png 1024w\" \/><\/p><p>Yep. That\u2019s the \u201cstraight\u201d Schema.org example (on the left), re-contextualized via my idosyncratic <code>@context<\/code> (seen on the right) with the fully realized and re-contextualized output document at the bottom.<\/p><p>So, yeah. Round tripped someone else\u2019s JSON into my own key names. Nice, yeah? Let\u2019s code for that.<\/p><p><a href=\"http:\/\/cta-redirect.hubspot.com\/cta\/redirect\/1169977\/964db6a6-69da-4366-afea-b129019aff07\"><img src=\"https:\/\/no-cache.hubspot.com\/cta\/default\/1169977\/964db6a6-69da-4366-afea-b129019aff07.png\" alt=\"Sign up for a free Codeship Account\" \/><\/a> <\/p><h3>The codes<\/h3><p>Here\u2019s the magic using <a href=\"https:\/\/github.com\/digitalbazaar\/jsonld.js\">json-ld.js<\/a>:<\/p><pre class=\"language-javascript\">var recontextualized = {};\njsonld.compact(schema_org_doc, my_context,\n  function(err, compacted) {\n    recontextualized = compacted;\n  });<\/pre><p>Running that code, I now have this output you just saw above. Not much code for so much magic, eh?<\/p><p>Here\u2019s a live environment with the content from this example, so you can test it out: <a class=\"jsbin-embed\" href=\"http:\/\/jsbin.com\/finore\/embed?js,output\">JS Bin on jsbin.com<\/a><\/p><h3>Meaningful People APIs<\/h3><p>Now that we can transform simple JSON documents into magically meaningful JSON-LD documents, let\u2019s look at what this gets us in practice.<\/p><p>Let\u2019s look at three JSON API endpoints from three different people-information-providers (commonly called Social Networks):<\/p><ul><li>Meetup<\/li><li>Twitter<\/li><li>LinkedIn<\/li><\/ul><p>First, here\u2019s the full idiosyncratic <code>@context<\/code> to cover the content I want \u201cnormalized\u201d from these three social networks:<\/p><pre class=\"language-bash\">{\n  \"@vocab\": \"http:\/\/schema.org\/\",\n  \"first_name\": \"givenName\",\n  \"last_name\": \"familyName\",\n  \"alias\": \"alternateName\",\n  \"job_title\": \"jobTitle\",\n  \"city\": \"addressLocality\",\n  \"country\": \"addressCountry\"\n}<\/pre><p>Below are the URLs, plain-old-JSON, and the site-specific custom <code>@context<\/code> needed to map their key names to the Schema.org standard URLs. There\u2019s also a JSBIN workspace you can fork for each of them!<\/p><h3>Meetup<\/h3><p>URL: <code>https:\/\/api.meetup.com\/2\/member\/19524571?&amp;sign=true&amp;photo-host=public&amp;page=20&amp;only=country,city,link,bio,name<\/code><\/p><p>JSON:<\/p><pre class=\"language-bash\">{\n  \"link\": \"http:\/\/www.meetup.com\/members\/19524571\",\n  \"name\": \"Benjamin Young\",\n  \"country\": \"us\",\n  \"bio\": \"aka BigBlueHat -=- Developer, Web, &amp; Open Source Advocate, Invited Expert in the Annotation and Digital Publishing Working Groups at the W3C. Previously an inventor and evangelist for IBM's Cloudant, Couchbase, and also CTO at InnoVenture.\",\n  \"city\": \"Greenville\"\n}<\/pre><p>Meetup Mapping <code>@context<\/code>:<\/p><pre class=\"language-bash\">{\n  \"city\": \"http:\/\/schema.org\/addressLocality\",\n  \"country\": \"http:\/\/schema.org\/addressCountry\",\n  \"bio\": \"http:\/\/schema.org\/description\",\n  \"name\": \"http:\/\/schema.org\/name\"\n}<\/pre><p><a class=\"jsbin-embed\" href=\"http:\/\/jsbin.com\/wenanu\/embed?js,output\">JS Bin on jsbin.com<\/a><\/p><h4>Twitter<\/h4><p>URL: <code>https:\/\/api.twitter.com\/1.1\/users\/show.json?screen_name=bigbluehat&amp;user_id=15841047<\/code><\/p><p>JSON <em>(only partial results below \u2019cause it\u2019s huge!)<\/em><\/p><pre class=\"language-bash\">{\n  \"id\": 15841047,\n  \"id_str\": \"15841047\",\n  \"name\": \"bigbluehat\",\n  \"screen_name\": \"bigbluehat\",\n  \"location\": \"Greenville, SC\",\n  \"profile_location\": null,\n  \"description\": \"inventor &amp; evangelist - I :heart: hypothes.is, @couchdb, open source, open communities. I organize @RESTFest &amp; @OpenUpstate.\"\n  ...\n}<\/pre><p>Twitter Mapping <code>@context<\/code>:<\/p><pre class=\"language-bash\">{\n  \"name\": \"http:\/\/schema.org\/alternateName\",\n  \"screen_name\": \"http:\/\/schema.org\/alternateName\",\n  \"description\": \"http:\/\/schema.org\/description\",\n  \"location\": \"http:\/\/schema.org\/location\"\n}<\/pre><p><a class=\"jsbin-embed\" href=\"http:\/\/jsbin.com\/gejewo\/embed?js,output\">JS Bin on jsbin.com<\/a><\/p><h4>LinkedIn<\/h4><p>URL: <code>https:\/\/api.linkedin.com\/v1\/people\/~?format=json<\/code><\/p><p>JSON:<\/p><pre class=\"language-bash\">{\n  \"firstName\": \"Benjamin\",\n  \"headline\": \"Invited Expert, W3C\",\n  \"id\": \"Ol-pwbI97V\",\n  \"lastName\": \"Young\",\n  \"siteStandardProfileRequest\": {\n    \"url\": \"https:\/\/www.linkedin.com\/profile\/view?id=AAoAAAAAx4oBxNlPOjsrspSms5FMi7Tx0c-EBSk&amp;authType=name&amp;authToken=XFwT&amp;trk=api*a3227641*s3301901*\"\n  }\n}<\/pre><p>LinkedIn Mapping <code>@context<\/code>:<\/p><pre class=\"language-bash\">{\n  \"firstName\": \"http:\/\/schema.org\/givenName\",\n  \"lastName\": \"http:\/\/schema.org\/familyName\",\n  \"headline\": \"http:\/\/schema.org\/jobTitle\"\n}<\/pre><p><a class=\"jsbin-embed\" href=\"http:\/\/jsbin.com\/godaro\/embed?js,output\">JS Bin on jsbin.com<\/a><\/p><h2>Conclusion<\/h2><p>I now have three different but meaningfully normalized variations of\u2026 myself. I can merge these in a number of ways depending on which data store I\u2019m using. The keys all match now. They all have meaningful identifiers that, when followed, result in documentation for the key names. That\u2019s far more than I got from the raw <code>GET<\/code> requests to these various APIs.<\/p><p>Next time you sit down to code up some JSON document \u2014 even one you think only you will use \u2014 consider mapping your key names to meaningful vocabulary terms via JSON-LD. Your future self will thank you.<\/p>","created_at":"2018-11-09T15:48:33+0000","updated_at":"2018-11-09T15:48:51+0000","published_at":"2016-03-17T11:15:22+0000","published_by":["Benjamin Young"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"blog.codeship.com","preview_picture":"http:\/\/blog.codeship.com\/wp-content\/uploads\/2016\/03\/Codeship_JSON-LD-Building-Meaningful-Data-APIs.jpg","http_status":"200"},{"is_archived":1,"is_starred":1,"tags":[],"is_public":false,"id":673,"title":"Start Your Own ISP","url":"https:\/\/startyourownisp.com","content":"<p>This site is dedicated to helping you start your own Internet Service Provider. Specifically this guide is about building a Wireless ISP (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Wireless_Internet_service_provider\">WISP<\/a>).<\/p><p>This guide is focused on the very earliest stages of starting a WISP - determining feasibility up through connecting the first few customers. There are many challenges that will come up at 100, 1,000 or 10,000 customers that are not (yet) covered in this guide.<\/p><p><a href=\"https:\/\/riot.im\/app\/#\/room\/#startyourownisp:matrix.org\"><strong>Join the discussion!<\/strong><\/a> Chat with me (the author) and others interested in this kind of thing here: <a href=\"https:\/\/riot.im\/app\/#\/room\/#startyourownisp:matrix.org\">#startyourownisp:matrix.org<\/a>.<\/p><div class=\"admonition note\"><p class=\"admonition-title\">This site is a work in progress!<\/p><p>Only some of the content is up so far and there\u2019s still some bugs in the interface. Check out the <a href=\"https:\/\/startyourownisp.com\/posts\/changelog\/\">changelog<\/a> to see what\u2019s new.<\/p><\/div><h2 id=\"getting-started\">Getting Started<\/h2><p><a href=\"https:\/\/startyourownisp.com\/posts\/introduction\/\"><strong>What is a WISP?<\/strong><\/a>\nAnd why might you want to build one? Also defines some terminology.<\/p><p><a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1jjUYOQMuZ4cRyTv1M5X8HGtrKFiW96_gNlfG8BBzRN0\/edit?usp=sharing\"><strong>Costs<\/strong><\/a>\nWhat does it cost to build a wireless Internet Service Provider? (Link to a Google Sheet that you can copy and customize.)<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/about-me\/\"><strong>About Me<\/strong><\/a>\nWho am I? Why am I doing this?<\/p><h2 id=\"step-by-step-guide\">Step by Step Guide<\/h2><p><a href=\"https:\/\/startyourownisp.com\/posts\/location-location-location\/\"><strong>Step 1: Evaluate an Area<\/strong><\/a>: Make sure your area is a good candidate for a Wireless Internet network.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/fiber-provider\/\"><strong>Step 2: Find a Fiber Provider<\/strong><\/a>: Find a building where you can purchase a fiber connection and use the rooftop to start your wireless network.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/relay-sites\/\"><strong>Step 3: Find Relay Sites<\/strong><\/a>:\nExtend your network wirelessly toward your customers.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/hardware-platform\/\"><strong>Step 4: Pick a Hardware Platform<\/strong><\/a>:\nEvaluate available options for wireless hardware.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/billing-customer-management\/\"><strong>Step 5: Billing and Customer Management<\/strong><\/a>:\nMake sure you\u2019re able to get paid and support your customers.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/network-topology\/\"><strong>Step 6: Network Topology<\/strong><\/a>:\nDesign your network topology to make your network reliable and scalable. Routers, switches, IP addresses, VLANs, etc.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/build-your-infrastructure\/\"><strong>Step 7: Build your Infrastructure<\/strong><\/a>:\nInstall hardware for your fiber connection and your relay sites.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/install-a-customer\/\"><strong>Step 8: Install a Customer<\/strong><\/a>:\nGet your first customer online!<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/marketing\/\"><strong>Step 9: Marketing<\/strong><\/a>:\nLet people know about your service so they can experience a better Internet connection!<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/maintenance\/\"><strong>Step 10: Maintenance<\/strong><\/a>:\nKeep your network running smoothly, even in bad weather.<\/p><h2 id=\"miscellaneous\">Miscellaneous<\/h2><p><a href=\"https:\/\/startyourownisp.com\/posts\/form477\/\"><strong>Form 477: How to prepare and file with the FCC<\/strong><\/a>\nForm 477 is used by the FCC to determine which providers are servicing which areas. ISPs must file this form twice a year.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/tool-list\/\"><strong>Tools you\u2019ll want to have<\/strong><\/a>\nA list of the tools you\u2019ll need to install relays sites and customers.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/aim-backhaul\/\"><strong>Aim a Backhaul<\/strong><\/a>\nA guide describing the proper techniques for aiming backhauls. Designed to be printed out and taken to the site for reference.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/backhaul-picker\/\"><strong>Backhaul List<\/strong><\/a>\nIf you just need to get a solid wireless connection from Point A to Point B then use this list to pick the right equipment and get it set up.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/channel-planning\/\"><strong>RF Basics and Channel Planning<\/strong><\/a>\nAvoid self interference by carefully choosing channels for your access points and backhauls.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/mdus\/\"><strong>MDUs (Multiple Dwelling Units)<\/strong><\/a>\nBest practices for providing service to apartment buildings, condos, attached townhomes, etc.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/guide-to-google-earth\/\"><strong>Guide to Google Earth<\/strong><\/a>\nSome tips and tricks for using Google Earth to plan and build your network.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/roof-and-ladder-safety\/\"><strong>Roof and Ladder Safety<\/strong><\/a>\nStay safe out there!<\/p><aside class=\"copyright\" role=\"note\"><div style=\"text-align: center;\">\n\t\t\t\t\tSponsored by <a href=\"https:\/\/broadbandnow.com\/\">BroadbandNow<\/a>.<br \/>Looking to donate? <a href=\"https:\/\/www.patreon.com\/startyourownisp\">Patreon<\/a> or BTC 1JKa1Kdrp3r4xPSXBRJ6nPC6YYdLcqQ4Bp. Many thanks!\n\t\t\t\t\t<br \/>\u00a9 2018 startyourownisp.com \u2013\n\t\t\t\t\t\n\t\t\t\t\tDocumentation built with\n\t\t\t\t\t<a href=\"https:\/\/www.gohugo.io\" target=\"_blank\">Hugo<\/a>\n\t\t\t\t\tusing the\n\t\t\t\t\t<a href=\"http:\/\/github.com\/digitalcraftsman\/hugo-material-docs\" target=\"_blank\">Material<\/a> theme.\n\t\t\t\t<\/div>\n\t\t\t<\/aside>","created_at":"2018-11-08T02:21:55+0000","updated_at":"2018-11-08T02:22:31+0000","starred_at":"2018-11-08T02:22:31+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"startyourownisp.com","preview_picture":"https:\/\/startyourownisp.com\/images\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":672,"title":"Gabriel Weinberg's answer to What is the revenue generation model for DuckDuckGo?","url":"https:\/\/www.quora.com\/What-is-the-revenue-generation-model-for-DuckDuckGo\/answer\/Gabriel-Weinberg","content":"<div id=\"wt6gwkNk43\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">DuckDuckGo has been a profitable company since 2014 without storing or sharing any personal information on people using our search engine. As we like to say, what you search on DuckDuckGo is private, even from us! We\u2019re proud to have a business model for a web-based business that\u2019s profitable without making your personal information the product. I\u2019m happy to tell you all about how we make it work (and how other companies can, too).<\/p><p class=\"ui_qtext_para\">Though first, if you\u2019re not familiar with DuckDuckGo, we are an Internet privacy company that empowers you to seamlessly take control of your personal information online, without any tradeoffs. We operate a search engine alternative to Google at <a href=\"http:\/\/duckduckgo.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\">http:\/\/duckduckgo.com<\/a>, and offer additional <a href=\"https:\/\/duckduckgo.com\/app\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"duckduckgo.com\">apps and extensions<\/a> to protect you from Google, Facebook and other trackers, no matter where you go on the Internet.<\/p><p class=\"ui_qtext_para\"><b>The Big Myth<\/b><\/p><p class=\"ui_qtext_para\">It\u2019s actually a big myth that search engines need to track your personal search history to make money or deliver quality search results. Almost all of the money search engines make (including Google) is based on the keywords you type in, without knowing anything about you, including your search history or the seemingly endless amounts of additional data points they have collected about registered and non-registered users alike.<\/p><p class=\"ui_qtext_para\">In fact, search advertisers buy search ads by bidding on keywords, not people. It makes intuitive sense, too. If you search for \u2018car\u2019, you are more likely to respond to a car ad than something you searched for last week.<\/p><p class=\"ui_qtext_para\">This keyword-based advertising is our primary business model. When you search on DuckDuckGo, we can show you an ad based on the keywords you type in. That\u2019s it. And it works. Our <a href=\"https:\/\/duckduckgo.com\/privacy\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"duckduckgo.com\">privacy policy<\/a>, in a nutshell, is to not collect or share any personal information at all. Every time you search on DuckDuckGo it is as if you were there for the first time \u2013 <a href=\"https:\/\/www.quora.com\/How-anonymous-is-DuckDuckGo\/answer\/Gabriel-Weinberg\">anonymous<\/a>.<\/p><p class=\"ui_qtext_para\">For example, if you type in \u2018dishwasher\u2019 you will get a dishwasher ad.<\/p><div class=\"ui_qtext_image_outer\"><div class=\"ui_qtext_image_wrapper\"><img class=\"landscape ui_qtext_image zoomable_in zoomable_in_feed lazy_loaded lazy_loading\" src=\"https:\/\/qph.fs.quoracdn.net\/main-qimg-9da3e4647664ed2031d44124ea7ecae0\" alt=\"image\" \/><\/div><\/div><p class=\"ui_qtext_para\">Google also makes most of their money via this same type of keyword-based advertising that doesn\u2019t require any search-history tracking.<\/p><p class=\"ui_qtext_para\">So why do they track it all then? Because Google is not really a search company; they are an advertising company. On Google, your searches are tracked, mined, and packaged up into a data profile for advertisers to follow you around the Internet through intrusive and annoying ever-present banner ads, using Google\u2019s <a href=\"https:\/\/www.doubleclickbygoogle.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"doubleclickbygoogle.com\">massive<\/a> <a href=\"https:\/\/www.google.com\/adsense\/start\/#\/?modal_active=none\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">ad<\/a> <a href=\"https:\/\/www.google.com\/admob\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">networks<\/a>, embedded across millions of sites and apps.<\/p><p class=\"ui_qtext_para\">This is why if you search for something on Google, you may start seeing ads for it everywhere. Using the internet doesn\u2019t have to feel like you\u2019re being watched, listened to, and monitored.<\/p><p class=\"ui_qtext_para\"><b>Google, Facebook, and The Creepy Line<\/b><\/p><p class=\"ui_qtext_para\">Eric Schmidt, former Google CEO and Chairman, <a href=\"https:\/\/www.youtube.com\/watch?v=o-rvER6YTss\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"youtube.com\">famously said<\/a> \u201c<i>Google\u2019s policy on a lot of these things is to get right up to the creepy line, but not cross it.<\/i>\u201d But for most people, that line was crossed by Google, Facebook, and others long ago.<\/p><p class=\"ui_qtext_para\">Alarmingly, Google now deploys hidden trackers on 76% of websites across the web to monitor your behavior and Facebook has hidden trackers on about 25% of websites, according to the <a href=\"https:\/\/webtap.princeton.edu\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"princeton.edu\">Princeton Web Transparency &amp; Accountability Project<\/a>. It is likely that Google and\/or Facebook are watching you on most sites you visit, in addition to tracking you when using their products.<\/p><div class=\"ui_qtext_image_outer\"><div class=\"ui_qtext_image_wrapper\"><img class=\"landscape ui_qtext_image zoomable_in zoomable_in_feed lazy_loaded lazy_loading\" src=\"https:\/\/qph.fs.quoracdn.net\/main-qimg-109807effc93555406ede32b7b7d235b\" alt=\"image\" \/><\/div><\/div><p class=\"ui_qtext_para\">As a result, these two companies have amassed huge data profiles on individuals, which can include interests, past purchases, search, browsing and location history, and much more. This personal data is stored indefinitely and used for invasive targeted advertising that can follow you around the Internet.<\/p><p class=\"ui_qtext_para\">This advertising system is designed to enable hyper-targeting, which has many unintended consequences that have dominated the headlines in recent years, such as the ability for bad actors to use the system to influence elections, to exclude groups in a way that facilitates discrimination, and to expose your personal data to companies you\u2019ve never even heard of.<\/p><p class=\"ui_qtext_para\">The operative question is, though, is all of this tracking necessary to make substantial profits? Is this the only way to run a profitable digital consumer focused service company? Not in my opinion. The fact is, these companies would still be wildly profitable if, for example, they dropped all of these hidden trackers across the web and limited the amount of data they keep to only what is most necessary.<\/p><p class=\"ui_qtext_para\">Yes, this additional tracking probably helps them compete with each other and adds some incremental revenue, but I believe the vast majority of their revenue would still exist if the tracking dial was turned way down, and they backed far away from the creepy line.<\/p><p class=\"ui_qtext_para\">The reason is simple: Google and Facebook are the undisputed champions of audience and reach across the internet, something advertisers will always pay for. Their business models don\u2019t need to be this invasive.<\/p><p class=\"ui_qtext_para\">It is a choice to squeeze every last ounce of profit at the expense of privacy, democracy and society. A choice they don\u2019t have to make. Without all this tracking, I\u2019m confident they would still be among the most profitable companies in the world, and we\u2019d all be better off.<\/p><p class=\"ui_qtext_para\"><b>Anonymous Affiliates<\/b><\/p><p class=\"ui_qtext_para\">As mentioned, DuckDuckGo is profitable based mostly on keyword-based search ads, though we have always been on the search for other ways to anonymously make money so that we can reduce the dependence on advertising. The only other way we\u2019ve found so far, which currently accounts for a much smaller portion of our revenue, is non-tracking affiliate partnerships with Amazon and eBay.<\/p><p class=\"ui_qtext_para\">When you visit those sites through DuckDuckGo, including when using <a href=\"https:\/\/duckduckgo.com\/bang\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"duckduckgo.com\">!bangs<\/a>, and subsequently make a purchase, we receive a small commission. This mechanism operates anonymously and there is no personally identifiable information exchanged between us and Amazon or eBay. These partnerships also don't affect the ranking of search results. The reason we can do this in an anonymous way with Amazon and eBay, though not with other retailers, is because Amazon and eBay run their own affiliate networks.<\/p><p class=\"ui_qtext_para\"><b>What Other Companies Can Do<\/b><\/p><p class=\"ui_qtext_para\">At the beginning of this answer, I noted that other companies using an advertising business model could follow a similar path to DuckDuckGo. Here are a few actionable things companies can do to remain profitable without tracking the maximum amount of information possible on consumers:<\/p><ul><li>Favor interest-based advertising instead of hyper-targeted advertising. For us, that is basing ads just on the keywords people type in. For others, that could mean basing ads on the content on the page and not on the individual viewing the page.<\/li><li>Sell advertising directly based on such interests, avoiding going through the hyper-targeted advertising systems of Google and Facebook.<\/li><li>Consider using an anonymous affiliate system like DuckDuckGo does as described above. This can help you get away from as much advertising on your pages.<\/li><\/ul><p class=\"ui_qtext_para\">Our vision is to set a new standard of trust online. The Internet shouldn't feel so creepy and getting the privacy you deserve online should be as simple as closing the blinds.<\/p><\/div><\/div><div id=\"wt6gwkNk48\"><div id=\"__w2_wt6gwkNk42_survey_wrapper\"><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T17:08:05+0000","updated_at":"2018-11-08T02:45:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.quora.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["emacs","mu4e"],"is_public":false,"id":671,"title":"Email templates in mu4e with yasnippet","url":"http:\/\/pragmaticemacs.com\/emacs\/email-templates-in-mu4e-with-yasnippet\/","content":"<p> This is the second in a <a href=\"http:\/\/pragmaticemacs.com\/mu4e-tutorials\/\">series of posts<\/a> on using <code>mu4e<\/code> for email in emacs. When I <a href=\"http:\/\/pragmaticemacs.com\/emacs\/master-your-inbox-with-mu4e-and-org-mode\/\">moved from thunderbird to mu4e<\/a>, there were some thunderbird extensions that I missed. One was quicktext which I used to create simple email templates for common sorts of emails. This is easily replicated in emacs using <a href=\"http:\/\/pragmaticemacs.com\/emacs\/smart-text-templates-with-yasnippet\/\">yasnippet<\/a>. <\/p>\n<p> As a simple example, I made a snippet to expand the key <code>all<\/code> into <\/p>\n<blockquote>\n<p> Hi all, <\/p>\n<p> \u2026 <\/p>\n<p> Cheers, <\/p>\n<p> Ben <\/p>\n<\/blockquote>\n<p> The snippet is <\/p>\n<div class=\"org-src-container\"><pre class=\"src src-snippet\"># -*- mode: snippet -*-\n# name: hi all\n# key: all\n# --\nHi all,\n$0\nCheers,\n Ben\n<\/pre><\/div>\n<p> This snippet is saved in the <code>message-mode<\/code> subdirectory of my <code>snippets<\/code> directory, since <code>message-mode<\/code> is the major mode for email composition in <code>mu4e<\/code>. <\/p>\n<p> We can be a bit smarter than this with a snippet that takes the name of the email recipient and adds that to the template. The following is close to <a href=\"http:\/\/blog.binchen.org\/posts\/how-to-use-yasnippets-to-produce-email-templates-in-emacs.html\">binchen\u2019s instructions<\/a> with a few small modifications. There are two parts to this: a snippet that expands out as normal, and a lisp function that it calls when it expands to extract the recipient\u2019s name. <\/p>\n<p> First, the lisp function <\/p>\n<div class=\"org-src-container\"><pre class=\"src src-elisp\">;; function to return first name of email recipients\n;; used by yasnippet\n;; inspired by\n;;http:\/\/blog.binchen.org\/posts\/how-to-use-yasnippets-to-produce-email-templates-in-emacs.html\n(defun bjm\/mu4e-get-names-for-yasnippet ()\n  \"Return comma separated string of names for an email\"\n  (interactive)\n  (let ((email-name \"\") str email-string email-list email-name2 tmpname)\n    (save-excursion\n      (goto-char (point-min))\n      ;; first line in email could be some hidden line containing NO to field\n      (setq str (buffer-substring-no-properties (point-min) (point-max))))\n    ;; take name from TO field - match series of names\n    (when (string-match \"^To: \\\"?\\\\(.+\\\\)\" str)\n      (setq email-string (match-string 1 str)))\n    ;;split to list by comma\n    (setq email-list (split-string email-string \" *, *\"))\n    ;;loop over emails\n    (dolist (tmpstr email-list)\n      ;;get first word of email string\n      (setq tmpname (car (split-string tmpstr \" \")))\n      ;;remove whitespace or \"\"\n      (setq tmpname (replace-regexp-in-string \"[ \\\"]\" \"\" tmpname))\n      ;;join to string\n      (setq email-name\n            (concat email-name \", \" tmpname)))\n    ;;remove initial comma\n    (setq email-name (replace-regexp-in-string \"^, \" \"\" email-name))\n    ;;see if we want to use the name in the FROM field\n    ;;get name in FROM field if available, but only if there is only\n    ;;one name in TO field\n    (if (&lt; (length email-list) 2)\n        (when (string-match \"^\\\\([^ ,\\n]+\\\\).+writes:$\" str)\n          (progn (setq email-name2 (match-string 1 str))\n                 ;;prefer name in FROM field if TO field has \"@\"\n                 (when (string-match \"@\" email-name)\n                   (setq email-name email-name2))\n                 )))\n    email-name))\n<\/pre><\/div>\n<p> This function takes the first name of the email recipient from the \u201cTo:\u201d field of the message. It also looks for a name in the \u201cJoe Bloggs writes:\u201d text that mu4e generates when you reply to an email. This is populated from the \u201cFrom:\u201d field of the email being replied to and sometimes gives a better match for the name. The function compares the two name strings and prefers the \u201cTo:\u201d name unless it contains an \u201c@\u201d in which case it chooses the \u201cFrom:\u201d name (this is the addition I made to <i>binchen<\/i>\u2018s version). You could make this more sophisticated but it works pretty well for me as is. <\/p>\n<p> Now we need a snippet to expand: <\/p>\n<div class=\"org-src-container\"><pre class=\"src src-snippet\"># -*- mode: snippet -*-\n# name: dear name\n# key: dear\n# --\nDear ${1:`(bjm\/mu4e-get-names-for-yasnippet)`},\n$0\nBest wishes,\n Ben\n<\/pre><\/div>\n<p> Note how our function is straightforwardly called by the snippet to give \u201cDear NAME\u201d. <\/p>\n<p> Hopefully this gives you some ideas of how to make useful email templates with yasnippet. In future posts I\u2019ll talk about how I have added attachment reminders and delayed sending features to mu4e. <\/p>","created_at":"2018-11-07T17:01:43+0000","updated_at":"2018-11-13T20:13:10+0000","published_at":"2016-01-11T20:54:00+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"pragmaticemacs.com","preview_picture":"https:\/\/i1.wp.com\/pragmaticemacs.com\/wp-content\/uploads\/2015\/05\/blog-logo-55532ec6v1_site_icon.png?fit=512%2C512","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":670,"title":"Welcome to Instapaper!","url":"http:\/\/blog.instapaper.com\/private\/85060389946\/tumblr_n585uton3C1qz8fxi","content":"<p>Instapaper is the best way to save and read web content. There are a number of convenient ways to save articles to Instapaper. Let\u2019s walkthrough your options.<\/p><p><b>Saving from your desktop<\/b><\/p><p><b>Using your Browser<\/b><\/p><p>\nThe most convenient way to save articles to Instapaper is with our\u00a0<a href=\"https:\/\/t.umblr.com\/redirect?z=https%3A%2F%2Fchrome.google.com%2Fwebstore%2Fdetail%2Finstapaper%2Fldjkgaaoikpmhmkelcgkgacicjfbofhh%3Fhl%3Den-GB&amp;t=MGVmZGNjZGU0ODcwOWI1YzMzMDA4MTBiOGMzZTRhM2Y2MjU0MDcxYywwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Chrome extension<\/a>,\u00a0<a href=\"https:\/\/t.umblr.com\/redirect?z=https%3A%2F%2Fsafari-extensions.apple.com%2Fdetails%2F%3Fid%3Dcom.instapaper.extension-CAM49M58WK&amp;t=MDhhMGEyODEwMzRmNzBjMTZkZjgxYzdjNWE2Nzg0YzY0Njk4NTNjOSwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Safari extension<\/a>,\u00a0<a href=\"https:\/\/t.umblr.com\/redirect?z=https%3A%2F%2Faddons.mozilla.org%2Fen-US%2Ffirefox%2Faddon%2Finstapaper-official%2F&amp;t=ZDdjY2FmNDZmMDAxN2RhZGU1NzJlYmEyNDhjNmNkYjE0OWFiYjY1ZCwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Firefox extension<\/a>, or\u00a0<a href=\"https:\/\/t.umblr.com\/redirect?z=http%3A%2F%2Fwww.instapaper.com%2Fsave&amp;t=ZGEzNzEwODg3Mzk1MDhmMDQ3N2UxZWEzMjU2MTY3MzlhYTg3MjI5ZSwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">bookmarklet<\/a>. When you\u2019re visiting a webpage, you can save it with one click.<\/p><figure data-orig-width=\"500\" data-orig-height=\"280\" class=\"tmblr-full\"><img src=\"https:\/\/66.media.tumblr.com\/fa716b2e7895973cc1b8de59f84179ff\/tumblr_inline_ojqp3l2Jog1tjlbd3_500.gif\" alt=\"image\" data-orig-width=\"500\" data-orig-height=\"280\" width=\"500\" height=\"280\" \/><\/figure><p><b>Email Articles<\/b><\/p><p>You can send links to articles, videos, and even long emails directly to Instapaper via email. To get started, add <a href=\"https:\/\/t.umblr.com\/redirect?z=http%3A%2F%2Fwww.instapaper.com%2Fsave%2Femail&amp;t=OWZlMmNjMWZmYmZmYjlhNzdhNzU1MjkzNDAwZjRiYjRkNGE4MzJlOCwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">your unique\u00a0Instapaper\u00a0email\u00a0address<\/a>\u00a0to your contacts.<\/p><p><b>Browse<\/b><\/p><p>Head over to Instapaper's\u00a0<a href=\"https:\/\/t.umblr.com\/redirect?z=http%3A%2F%2Fwww.instapaper.com%2Fbrowse&amp;t=YjJlZjNlNGEwYmQwZjAwZjExYTQzMTc5ZWFhNGFiOTdiZDlhYjhlNCwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Browse<\/a> section to find articles chosen by our editors &amp; algorithms, and shared by your friends.<\/p><p><b>Saving from iOS<\/b><\/p><p>If you download the <a href=\"https:\/\/t.umblr.com\/redirect?z=https%3A%2F%2Fitunes.apple.com%2Fus%2Fapp%2Finstapaper%2Fid288545208%3Fmt%3D8&amp;t=MTY2NGZmODZkZTA1OWI5ZTI0MDkwOGU5MjhmYTU4NTc1MDRjMDY2ZiwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Instapaper iPhone or iPad app<\/a>, you can save directly from your mobile device via the system share sheet!<\/p><p><b>Connect Third Party Apps<\/b><\/p><p>Save directly to Instapaper from <a href=\"https:\/\/t.umblr.com\/redirect?z=http%3A%2F%2Fpystaging03.ec2.instapaper.com%2Fsave%2Fapps&amp;t=MTBjMmU3MTQzODQ3NjhmYWU5YzJkMDBiNmE3YjEyNTYwZGQ4M2JjOSwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">hundreds of iPhone and iPad apps<\/a>.<\/p><p><b>Safari Bookmarklet<\/b><\/p><p>Install the Instapaper bookmarklet in Safari\u00a0to save while you browse. After\u00a0installing the bookmarklet, you can\u00a0save to Instapaper by selecting the bookmarklet from\u00a0your bookmarks menu.<\/p><p><b>Saving from Android<\/b><\/p><p>Once you install the <a href=\"https:\/\/t.umblr.com\/redirect?z=https%3A%2F%2Fplay.google.com%2Fstore%2Fapps%2Fdetails%3Fid%3Dcom.instapaper.android&amp;t=YzRhNWM0MTA2ZTE1ZjM2MWUyYWExY2ZjODQ5MDdhZjYzYTVlYWVlMywwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Instapaper Android<\/a> app, saving is as easy as pressing the \u201cShare\u201d option from your browser or favorite apps and choosing the Instapaper option.<\/p><p><b>Some great Instapaper features<\/b><\/p><p><b>Highlights<\/b><\/p><p>With Instapaper you can highlight and save text from any article. To create a highlight just select some text and press the \u201cHighlight\u201d option.\u00a0Your highlights are immediately synced to all your mobile devices and Instapaper.com.<\/p><p><b>Text control<\/b><\/p><p>Instapaper can be tailored to your reading preferences. When viewing an article, tap the \u201cAa\"\u00a0icon for controls to\u00a0adjust font, text size, color, margins, and line spacing.<\/p><p><b>Sharing<\/b><\/p><p>There are a lot of options for sharing from Instapaper. Share links directly to Twitter, Facebook, Tumblr, Evernote, Pinboard, and many other app. Email a link, a selection, or full text of an article.<\/p><p><b>Archive<\/b><\/p><p>Finished reading this article? Move it to the Archive.\u00a0<a href=\"https:\/\/t.umblr.com\/redirect?z=https%3A%2F%2Fwww.instapaper.com%2Fsubscription&amp;t=OTQ2MmY3ZDE2ZmIxMjVkMzE0OTY5ZTM3MDkwOTkzYTUzOGE4YTI1ZiwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">Instapaper subscribers<\/a>\u00a0can\u00a0search the full article text of all articles in Instapaper.<\/p><p><b>Questions? Comments?<\/b><\/p><p>Please feel free to shoot us an email at <a href=\"https:\/\/t.umblr.com\/redirect?z=mailto%3Asupport%40help.instapaper.com&amp;t=YTg5MTNkNWUyNzUzNDNiODJiOWQ4YTMwYmM0ZWY1YzY1NmUwMDhjZSwwaXBUaDU1bA%3D%3D&amp;b=t%3AOZjQZ-Pa6uMEdOyfetS5_g&amp;p=http%3A%2F%2Fblog.instapaper.com%2Fpost%2F85060389946&amp;m=1\">support@help.instapaper.com<\/a>, or tweet to us\u00a0<a href=\"http:\/\/twitter.com\/InstapaperHelp\">@InstapaperHelp<\/a>.\u00a0Thank you for using Instapaper!<\/p><p>- Instapaper Team<\/p>","created_at":"2018-11-07T16:54:48+0000","updated_at":"2018-11-07T16:54:48+0000","published_at":"2014-05-07T22:16:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"blog.instapaper.com","preview_picture":"https:\/\/staticinstapaper.s3.amazonaws.com\/images\/insta_logo_med.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":668,"title":"learnbyexample\/Command-line-text-processing","url":"https:\/\/github.com\/learnbyexample\/Command-line-text-processing\/blob\/master\/gnu_awk.md","content":"<h2><a id=\"user-content-gnu-awk\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnu-awk\"><\/a>GNU awk<\/h2>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#field-processing\">Field processing<\/a>\n<ul><li><a href=\"#default-field-separation\">Default field separation<\/a><\/li>\n<li><a href=\"#specifying-different-input-field-separator\">Specifying different input field separator<\/a><\/li>\n<li><a href=\"#specifying-different-output-field-separator\">Specifying different output field separator<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#filtering\">Filtering<\/a>\n<ul><li><a href=\"#idiomatic-print-usage\">Idiomatic print usage<\/a><\/li>\n<li><a href=\"#field-comparison\">Field comparison<\/a><\/li>\n<li><a href=\"#regular-expressions-based-filtering\">Regular expressions based filtering<\/a><\/li>\n<li><a href=\"#fixed-string-matching\">Fixed string matching<\/a><\/li>\n<li><a href=\"#line-number-based-filtering\">Line number based filtering<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#case-insensitive-filtering\">Case Insensitive filtering<\/a><\/li>\n<li><a href=\"#changing-record-separators\">Changing record separators<\/a>\n<ul><li><a href=\"#paragraph-mode\">Paragraph mode<\/a><\/li>\n<li><a href=\"#multicharacter-rs\">Multicharacter RS<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#substitute-functions\">Substitute functions<\/a><\/li>\n<li><a href=\"#inplace-file-editing\">Inplace file editing<\/a><\/li>\n<li><a href=\"#using-shell-variables\">Using shell variables<\/a><\/li>\n<li><a href=\"#multiple-file-input\">Multiple file input<\/a><\/li>\n<li><a href=\"#control-structures\">Control Structures<\/a>\n<ul><li><a href=\"#if-else-and-loops\">if-else and loops<\/a><\/li>\n<li><a href=\"#next-and-nextfile\">next and nextfile<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#multiline-processing\">Multiline processing<\/a><\/li>\n<li><a href=\"#two-file-processing\">Two file processing<\/a>\n<ul><li><a href=\"#comparing-whole-lines\">Comparing whole lines<\/a><\/li>\n<li><a href=\"#comparing-specific-fields\">Comparing specific fields<\/a><\/li>\n<li><a href=\"#getline\">getline<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#creating-new-fields\">Creating new fields<\/a><\/li>\n<li><a href=\"#dealing-with-duplicates\">Dealing with duplicates<\/a><\/li>\n<li><a href=\"#lines-between-two-regexps\">Lines between two REGEXPs<\/a>\n<ul><li><a href=\"#all-unbroken-blocks\">All unbroken blocks<\/a><\/li>\n<li><a href=\"#specific-blocks\">Specific blocks<\/a><\/li>\n<li><a href=\"#broken-blocks\">Broken blocks<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#arrays\">Arrays<\/a><\/li>\n<li><a href=\"#awk-scripts\">awk scripts<\/a><\/li>\n<li><a href=\"#miscellaneous\">Miscellaneous<\/a>\n<ul><li><a href=\"#fpat-and-fieldwidths\">FPAT and FIELDWIDTHS<\/a><\/li>\n<li><a href=\"#string-functions\">String functions<\/a><\/li>\n<li><a href=\"#executing-external-commands\">Executing external commands<\/a><\/li>\n<li><a href=\"#printf-formatting\">printf formatting<\/a><\/li>\n<li><a href=\"#redirecting-print-output\">Redirecting print output<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#gotchas-and-tips\">Gotchas and Tips<\/a><\/li>\n<li><a href=\"#further-reading\">Further Reading<\/a><\/li>\n<\/ul><br \/><div class=\"highlight highlight-source-shell\"><pre>$ awk --version | head -n1\nGNU Awk 4.1.3, API: 1.1 (GNU MPFR 3.1.4, GNU MP 6.1.0)\n$ man awk\nGAWK(1)                        Utility Commands                        GAWK(1)\nNAME\n       gawk - pattern scanning and processing language\nSYNOPSIS\n       gawk [ POSIX or GNU style options ] -f program-file [ -- ] file ...\n       gawk [ POSIX or GNU style options ] [ -- ] program-text file ...\nDESCRIPTION\n       Gawk  is  the  GNU Project's implementation of the AWK programming lan\u2010\n       guage.  It conforms to the definition of  the  language  in  the  POSIX\n       1003.1  Standard.   This version in turn is based on the description in\n       The AWK Programming Language, by Aho, Kernighan, and Weinberger.   Gawk\n       provides  the additional features found in the current version of Brian\n       Kernighan's awk and a number of GNU-specific extensions.\n...<\/pre><\/div>\n<p><strong>Prerequisites and notes<\/strong><\/p>\n<ul><li>familiarity with programming concepts like variables, printing, control structures, arrays, etc<\/li>\n<li>familiarity with regular expressions\n<ul><li>if not, check out <strong>ERE<\/strong> portion of <a href=\"https:\/\/github.com\/learnbyexample\/Command-line-text-processing\/blob\/master\/gnu_sed.md#regular-expressions\">GNU sed regular expressions<\/a> which is close enough to features available in <code>gawk<\/code><\/li>\n<\/ul><\/li>\n<li>this tutorial is primarily focussed on short programs that are easily usable from command line, similar to using <code>grep<\/code>, <code>sed<\/code>, etc<\/li>\n<li>see <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/\" rel=\"nofollow\">Gawk: Effective AWK Programming<\/a> manual for complete reference, has information on other <code>awk<\/code> versions as well as notes on POSIX standard<\/li>\n<\/ul><br \/><h2><a id=\"user-content-field-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#field-processing\"><\/a>Field processing<\/h2>\n<br \/><h4><a id=\"user-content-default-field-separation\" class=\"anchor\" aria-hidden=\"true\" href=\"#default-field-separation\"><\/a>Default field separation<\/h4>\n<ul><li><code>$0<\/code> contains the entire input record\n<ul><li>default input record separator is newline character<\/li>\n<\/ul><\/li>\n<li><code>$1<\/code> contains the first field text\n<ul><li>default input field separator is one or more of continuous space, tab or newline characters<\/li>\n<\/ul><\/li>\n<li><code>$2<\/code> contains the second field text and so on<\/li>\n<li><code>$(2+3)<\/code> result of expressions can be used, this one evaluates to <code>$5<\/code> and hence gives fifth field\n<ul><li>similarly if variable <code>i<\/code> has value <code>2<\/code>, then <code>$(i+3)<\/code> will give fifth field<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Expressions.html\" rel=\"nofollow\">gawk manual - Expressions<\/a><\/li>\n<\/ul><\/li>\n<li><code>NF<\/code> is a built-in variable which contains number of fields in the current record\n<ul><li>so, <code>$NF<\/code> will give last field<\/li>\n<li><code>$(NF-1)<\/code> will give second last field and so on<\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat fruits.txt\nfruit   qty\napple   42\nbanana  31\nfig     90\nguava   6\n$ # print only first field\n$ awk '{print $1}' fruits.txt\nfruit\napple\nbanana\nfig\nguava\n$ # print only second field\n$ awk '{print $2}' fruits.txt\nqty\n42\n31\n90\n6<\/pre><\/div>\n<br \/><h4><a id=\"user-content-specifying-different-input-field-separator\" class=\"anchor\" aria-hidden=\"true\" href=\"#specifying-different-input-field-separator\"><\/a>Specifying different input field separator<\/h4>\n<ul><li>by using <code>-F<\/code> command line option<\/li>\n<li>by setting <code>FS<\/code> variable<\/li>\n<li>See <a href=\"#fpat-and-fieldwidths\">FPAT and FIELDWIDTHS<\/a> section for other ways of defining input fields<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # second field where input field separator is :\n$ echo 'foo:123:bar:789' | awk -F: '{print $2}'\n123\n$ # last field\n$ echo 'foo:123:bar:789' | awk -F: '{print $NF}'\n789\n$ # first and last field\n$ # note the use of , and space between output fields\n$ echo 'foo:123:bar:789' | awk -F: '{print $1, $NF}'\nfoo 789\n$ # second last field\n$ echo 'foo:123:bar:789' | awk -F: '{print $(NF-1)}'\nbar\n$ # use quotes to avoid clashes with shell special characters\n$ echo 'one;two;three;four' | awk -F';' '{print $3}'\nthree<\/pre><\/div>\n<ul><li>Regular expressions based input field separator<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ echo 'Sample123string54with908numbers' | awk -F'[0-9]+' '{print $2}'\nstring\n$ # first field will be empty as there is nothing before '{'\n$ echo '{foo}   bar=baz' | awk -F'[{}= ]+' '{print $1}'\n$ echo '{foo}   bar=baz' | awk -F'[{}= ]+' '{print $2}'\nfoo\n$ echo '{foo}   bar=baz' | awk -F'[{}= ]+' '{print $3}'\nbar<\/pre><\/div>\n<ul><li>default input field separator is one or more of continuous space, tab or newline characters (will be termed as whitespace here on)\n<ul><li>exact same behavior if <code>FS<\/code> is assigned single space character<\/li>\n<\/ul><\/li>\n<li>in addition, leading and trailing whitespaces won't be considered when splitting the input record<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ printf ' a    ate b\\tc   \\n'\n a    ate b     c\n$ printf ' a    ate b\\tc   \\n' | awk '{print $1}'\na\n$ printf ' a    ate b\\tc   \\n' | awk '{print NF}'\n4\n$ # same behavior if FS is assigned to single space character\n$ printf ' a    ate b\\tc   \\n' | awk -F' ' '{print $1}'\na\n$ printf ' a    ate b\\tc   \\n' | awk -F' ' '{print NF}'\n4\n$ # for anything else, leading\/trailing whitespaces will be considered\n$ printf ' a    ate b\\tc   \\n' | awk -F'[ \\t]+' '{print $2}'\na\n$ printf ' a    ate b\\tc   \\n' | awk -F'[ \\t]+' '{print NF}'\n6<\/pre><\/div>\n<ul><li>assigning empty string to FS will split the input record character wise<\/li>\n<li>note the use of command line option <code>-v<\/code> to set FS<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ echo 'apple' | awk -v FS= '{print $1}'\na\n$ echo 'apple' | awk -v FS= '{print $2}'\np\n$ echo 'apple' | awk -v FS= '{print $NF}'\ne\n$ # detecting multibyte characters depends on locale\n$ printf 'hi\ud83d\udc4d how are you?' | awk -v FS= '{print $3}'\n\ud83d\udc4d<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Field-Splitting-Summary.html#Field-Splitting-Summary\" rel=\"nofollow\">gawk manual - Field Splitting Summary<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/30405694\/default-field-separator-for-awk\" rel=\"nofollow\">stackoverflow - explanation on default FS<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/362550\/how-to-remove-line-if-it-contains-a-character-exactly-once\" rel=\"nofollow\">unix.stackexchange - filter lines if it contains a particular character only once<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/24516141\/awk-processing-2-files-with-different-field-separators\" rel=\"nofollow\">stackoverflow - Processing 2 files with different field separators<\/a><\/li>\n<\/ul><br \/><h4><a id=\"user-content-specifying-different-output-field-separator\" class=\"anchor\" aria-hidden=\"true\" href=\"#specifying-different-output-field-separator\"><\/a>Specifying different output field separator<\/h4>\n<ul><li>by setting <code>OFS<\/code> variable<\/li>\n<li>also gets added between every argument to <code>print<\/code> statement\n<ul><li>use <a href=\"#printf-formatting\">printf<\/a> to avoid this<\/li>\n<\/ul><\/li>\n<li>default is single space<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # statements inside BEGIN are executed before processing any input text\n$ echo 'foo:123:bar:789' | awk 'BEGIN{FS=OFS=\":\"} {print $1, $NF}'\nfoo:789\n$ # can also be set using command line option -v\n$ echo 'foo:123:bar:789' | awk -F: -v OFS=':' '{print $1, $NF}'\nfoo:789\n$ # changing a field will re-build contents of $0\n$ echo ' a      ate b   ' | awk '{$2 = \"foo\"; print $0}' | cat -A\na foo b$\n$ # $1=$1 is an idiomatic way to re-build when there is nothing else to change\n$ echo 'foo:123:bar:789' | awk -F: -v OFS='-' '{print $0}'\nfoo:123:bar:789\n$ echo 'foo:123:bar:789' | awk -F: -v OFS='-' '{$1=$1; print $0}'\nfoo-123-bar-789\n$ # OFS is used to separate different arguments given to print\n$ echo 'foo:123:bar:789' | awk -F: -v OFS='\\t' '{print $1, $3}'\nfoo     bar\n$ echo 'Sample123string54with908numbers' | awk -F'[0-9]+' '{$1=$1; print $0}'\nSample string with numbers<\/pre><\/div>\n<br \/><h2><a id=\"user-content-filtering\" class=\"anchor\" aria-hidden=\"true\" href=\"#filtering\"><\/a>Filtering<\/h2>\n<br \/><h4><a id=\"user-content-idiomatic-print-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#idiomatic-print-usage\"><\/a>Idiomatic print usage<\/h4>\n<ul><li><code>print<\/code> statement with no arguments will print contents of <code>$0<\/code><\/li>\n<li>if condition is specified without corresponding statements, contents of <code>$0<\/code> is printed if condition evaluates to true<\/li>\n<li><code>1<\/code> is typically used to represent always true condition and thus print contents of <code>$0<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat poem.txt\nRoses are red,\nViolets are blue,\nSugar is sweet,\nAnd so are you.\n$ # displaying contents of input file(s) similar to 'cat' command\n$ # equivalent to using awk '{print $0}' and awk '1'\n$ awk '{print}' poem.txt\nRoses are red,\nViolets are blue,\nSugar is sweet,\nAnd so are you.<\/pre><\/div>\n<br \/><h4><a id=\"user-content-field-comparison\" class=\"anchor\" aria-hidden=\"true\" href=\"#field-comparison\"><\/a>Field comparison<\/h4>\n<ul><li>Each block of statements within <code>{}<\/code> can be prefixed by an optional condition so that those statements will execute only if condition evaluates to true<\/li>\n<li>Condition specified without corresponding statements will lead to printing contents of <code>$0<\/code> if condition evaluates to true<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # if first field exactly matches the string 'apple'\n$ awk '$1==\"apple\"{print $2}' fruits.txt\n42\n$ # print first field if second field &gt; 35\n$ # NR&gt;1 to avoid the header line\n$ # NR built-in variable contains record number\n$ awk 'NR&gt;1 &amp;&amp; $2&gt;35{print $1}' fruits.txt\napple\nfig\n$ # print header and lines with qty &lt; 35\n$ awk 'NR==1 || $2&lt;35' fruits.txt\nfruit   qty\nbanana  31\nguava   6<\/pre><\/div>\n<ul><li>If the above examples are too confusing, think of it as syntactical sugar<\/li>\n<li>Statements are grouped within <code>{}<\/code>\n<ul><li>inside <code>{}<\/code>, we have a <code>if<\/code> control structure<\/li>\n<li>Like <code>C<\/code> language, braces not needed for single statements within <code>if<\/code>, but consider that <code>{}<\/code> is used for clarity<\/li>\n<li>From this explicit syntax, remove the outer <code>{}<\/code>, <code>if<\/code> and <code>()<\/code> used for <code>if<\/code><\/li>\n<\/ul><\/li>\n<li>As we'll see later, this allows to mash up few lines of program compactly on command line itself\n<ul><li>Of course, for medium to large programs, it is better to put the code in separate file. See <a href=\"#awk-scripts\">awk scripts<\/a> section<\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # awk '$1==\"apple\"{print $2}' fruits.txt\n$ awk '{\n         if($1 == \"apple\"){\n            print $2\n         }\n       }' fruits.txt\n42\n$ # awk 'NR==1 || $2&lt;35' fruits.txt\n$ awk '{\n         if(NR==1 || $2&lt;35){\n            print $0\n         }\n       }' fruits.txt\nfruit   qty\nbanana  31\nguava   6<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Truth-Values-and-Conditions.html\" rel=\"nofollow\">gawk manual - Truth Values and Conditions<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Precedence.html\" rel=\"nofollow\">gawk manual - Operator Precedence<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/359697\/print-columns-in-awk-by-header-name\" rel=\"nofollow\">unix.stackexchange - filtering columns by header name<\/a><\/li>\n<\/ul><br \/><h4><a id=\"user-content-regular-expressions-based-filtering\" class=\"anchor\" aria-hidden=\"true\" href=\"#regular-expressions-based-filtering\"><\/a>Regular expressions based filtering<\/h4>\n<ul><li>the <em>REGEXP<\/em> is specified within <code>\/\/<\/code> and by default acts upon <code>$0<\/code><\/li>\n<li>See also <a href=\"https:\/\/stackoverflow.com\/questions\/17908555\/printing-with-sed-or-awk-a-line-following-a-matching-pattern\" rel=\"nofollow\">stackoverflow - lines around matching regexp<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # all lines containing the string 'are'\n$ # same as: grep 'are' poem.txt\n$ awk '\/are\/' poem.txt\nRoses are red,\nViolets are blue,\nAnd so are you.\n$ # negating REGEXP, same as: grep -v 'are' poem.txt\n$ awk '!\/are\/' poem.txt\nSugar is sweet,\n$ # same as: grep 'are' poem.txt | grep -v 'so'\n$ awk '\/are\/ &amp;&amp; !\/so\/' poem.txt\nRoses are red,\nViolets are blue,\n$ # lines starting with 'a' or 'b'\n$ awk '\/^[ab]\/' fruits.txt\napple   42\nbanana  31\n$ # print last field of all lines containing 'are'\n$ awk '\/are\/{print $NF}' poem.txt\nred,\nblue,\nyou.<\/pre><\/div>\n<ul><li>strings can be used as well, which will be interpreted as <em>REGEXP<\/em> if necessary<\/li>\n<li>Allows <a href=\"#using-shell-variables\">using shell variables<\/a> instead of hardcoded <em>REGEXP<\/em>\n<ul><li>that section also notes difference between using <code>\/\/<\/code> and string<\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk '$0 !~ \"are\"' poem.txt\nSugar is sweet,\n$ awk '$0 ~ \"^[ab]\"' fruits.txt\napple   42\nbanana  31\n$ # also helpful if search strings have the \/ delimiter character\n$ cat paths.txt\n\/foo\/a\/report.log\n\/foo\/y\/power.log\n$ awk '\/\\\/foo\\\/a\\\/\/' paths.txt\n\/foo\/a\/report.log\n$ awk '$0 ~ \"\/foo\/a\/\"' paths.txt\n\/foo\/a\/report.log<\/pre><\/div>\n<ul><li><em>REGEXP<\/em> matching against specific field<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # if first field contains 'a'\n$ awk '$1 ~ \/a\/' fruits.txt\napple   42\nbanana  31\nguava   6\n$ # if first field contains 'a' and qty &gt; 20\n$ awk '$1 ~ \/a\/ &amp;&amp; $2 &gt; 20' fruits.txt\napple   42\nbanana  31\n$ # if first field does NOT contain 'a'\n$ awk '$1 !~ \/a\/' fruits.txt\nfruit   qty\nfig     90<\/pre><\/div>\n<br \/><h4><a id=\"user-content-fixed-string-matching\" class=\"anchor\" aria-hidden=\"true\" href=\"#fixed-string-matching\"><\/a>Fixed string matching<\/h4>\n<ul><li>to search a string literally, <code>index<\/code> function can be used instead of <em>REGEXP<\/em>\n<ul><li>similar to <code>grep -F<\/code><\/li>\n<\/ul><\/li>\n<li>the function returns the starting position and <code>0<\/code> if no match found<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat eqns.txt\na=b,a-b=c,c*d\na+b,pi=3.14,5e12\ni*(t+9-g)\/8,4-a+b\n$ # no output since '+' is meta character, would need '\/a\\+b\/'\n$ awk '\/a+b\/' eqns.txt\n$ # same as: grep -F 'a+b' eqns.txt\n$ awk 'index($0,\"a+b\")' eqns.txt\na+b,pi=3.14,5e12\ni*(t+9-g)\/8,4-a+b\n$ # much easier than '\/i\\*\\(t\\+9-g\\)\/'\n$ awk 'index($0,\"i*(t+9-g)\")' eqns.txt\ni*(t+9-g)\/8,4-a+b\n$ # check only last field\n$ awk -F, 'index($NF,\"a+b\")' eqns.txt\ni*(t+9-g)\/8,4-a+b\n$ # index not needed if entire field\/line is being compared\n$ awk -F, '$1==\"a+b\"' eqns.txt\na+b,pi=3.14,5e12<\/pre><\/div>\n<ul><li>return value is useful to match at specific position<\/li>\n<li>for ex: at start\/end of line<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # start of line\n$ awk 'index($0,\"a+b\")==1' eqns.txt\na+b,pi=3.14,5e12\n$ # end of line\n$ # length function returns number of characters, by default acts on $0\n$ awk 'index($0,\"a+b\")==length()-length(\"a+b\")+1' eqns.txt\ni*(t+9-g)\/8,4-a+b\n$ # to avoid repetitions, save the search string in variable\n$ awk -v s=\"a+b\" 'index($0,s)==length()-length(s)+1' eqns.txt\ni*(t+9-g)\/8,4-a+b<\/pre><\/div>\n<br \/><h4><a id=\"user-content-line-number-based-filtering\" class=\"anchor\" aria-hidden=\"true\" href=\"#line-number-based-filtering\"><\/a>Line number based filtering<\/h4>\n<ul><li>Built-in variable <code>NR<\/code> contains total records read so far<\/li>\n<li>Use <code>FNR<\/code> if you need line numbers separately for <a href=\"#multiple-file-processing\">multiple file processing<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # same as: head -n2 poem.txt | tail -n1\n$ awk 'NR==2' poem.txt\nViolets are blue,\n$ # print 2nd and 4th line\n$ awk 'NR==2 || NR==4' poem.txt\nViolets are blue,\nAnd so are you.\n$ # same as: tail -n1 poem.txt\n$ # statements inside END are executed after processing all input text\n$ awk 'END{print}' poem.txt\nAnd so are you.\n$ awk 'NR==4{print $2}' fruits.txt\n90<\/pre><\/div>\n<ul><li>for large input, use <code>exit<\/code> to avoid unnecessary record processing<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ seq 14323 14563435 | awk 'NR==234{print; exit}'\n14556\n$ # sample time comparison\n$ time seq 14323 14563435 | awk 'NR==234{print; exit}'\n14556\nreal    0m0.004s\nuser    0m0.004s\nsys     0m0.000s\n$ time seq 14323 14563435 | awk 'NR==234{print}'\n14556\nreal    0m2.167s\nuser    0m2.280s\nsys     0m0.092s<\/pre><\/div>\n<ul><li>See also <a href=\"https:\/\/unix.stackexchange.com\/questions\/325985\/how-to-print-lines-number-15-and-25-out-of-each-50-lines\" rel=\"nofollow\">unix.stackexchange - filtering list of lines from every X number of lines<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-case-insensitive-filtering\" class=\"anchor\" aria-hidden=\"true\" href=\"#case-insensitive-filtering\"><\/a>Case Insensitive filtering<\/h2>\n<div class=\"highlight highlight-source-shell\"><pre>$ # same as: grep -i 'rose' poem.txt\n$ awk -v IGNORECASE=1 '\/rose\/' poem.txt\nRoses are red,\n$ # for small enough set, can also use REGEXP character class\n$ awk '\/[rR]ose\/' poem.txt\nRoses are red,\n$ # another way is to use built-in string function 'tolower'\n$ awk 'tolower($0) ~ \/rose\/' poem.txt\nRoses are red,<\/pre><\/div>\n<br \/><h2><a id=\"user-content-changing-record-separators\" class=\"anchor\" aria-hidden=\"true\" href=\"#changing-record-separators\"><\/a>Changing record separators<\/h2>\n<ul><li><code>RS<\/code> to change input record separator<\/li>\n<li>default is newline character<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ s='this is a sample string'\n$ # space as input record separator, printing all records\n$ printf \"$s\" | awk -v RS=' ' '{print NR, $0}'\n1 this\n2 is\n3 a\n4 sample\n5 string\n$ # print all records containing 'a'\n$ printf \"$s\" | awk -v RS=' ' '\/a\/'\na\nsample<\/pre><\/div>\n<ul><li><code>ORS<\/code> to change output record separator<\/li>\n<li>gets added to every <code>print<\/code> statement\n<ul><li>use <a href=\"#printf-formatting\">printf<\/a> to avoid this<\/li>\n<\/ul><\/li>\n<li>default is newline character<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ seq 3 | awk '{print $0}'\n1\n2\n3\n$ # note that there is empty line after last record\n$ seq 3 | awk -v ORS='\\n\\n' '{print $0}'\n1\n2\n3\n$ # dynamically changing ORS\n$ # ?: ternary operator to select between two expressions based on a condition\n$ # can also use: seq 6 | awk '{ORS = NR%2 ? \" \" : RS} 1'\n$ seq 6 | awk '{ORS = NR%2 ? \" \" : \"\\n\"} 1'\n1 2\n3 4\n5 6\n$ seq 6 | awk '{ORS = NR%3 ? \"-\" : \"\\n\"} 1'\n1-2-3\n4-5-6<\/pre><\/div>\n<br \/><h4><a id=\"user-content-paragraph-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#paragraph-mode\"><\/a>Paragraph mode<\/h4>\n<ul><li>When <code>RS<\/code> is set to empty string, one or more consecutive empty lines is used as input record separator<\/li>\n<li>Can also use regular expression <code>RS=\\n\\n+<\/code> but there are subtle differences, see <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Multiple-Line.html\" rel=\"nofollow\">gawk manual - multiline records<\/a>. Important points from that link quoted below<\/li>\n<\/ul><blockquote>\n<p>However, there is an important difference between \u2018RS = \"\"\u2019 and \u2018RS = \"\\n\\n+\"\u2019. In the first case, leading newlines in the input data file are ignored, and if a file ends without extra blank lines after the last record, the final newline is removed from the record. In the second case, this special processing is not done<\/p>\n<\/blockquote>\n<blockquote>\n<p>Now that the input is separated into records, the second step is to separate the fields in the records. One way to do this is to divide each of the lines into fields in the normal manner. This happens by default as the result of a special feature. When RS is set to the empty string and FS is set to a single character, the newline character always acts as a field separator. This is in addition to whatever field separations result from FS<\/p>\n<\/blockquote>\n<blockquote>\n<p>When FS is the null string (\"\") or a regexp, this special feature of RS does not apply. It does apply to the default field separator of a single space: \u2018FS = \" \"\u2019<\/p>\n<\/blockquote>\n<p>Consider the below sample file<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat sample.txt\nHello World\nGood day\nHow are you\nJust do-it\nBelieve it\nToday is sunny\nNot a bit funny\nNo doubt you like it too\nMuch ado about nothing\nHe he he<\/pre><\/div>\n<ul><li>Filtering paragraphs<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # print all paragraphs containing 'it'\n$ # if extra newline at end is undesirable, can use\n$ # awk -v RS= '\/it\/{print c++ ? \"\\n\" $0 : $0}' sample.txt\n$ awk -v RS= -v ORS='\\n\\n' '\/it\/' sample.txt\nJust do-it\nBelieve it\nToday is sunny\nNot a bit funny\nNo doubt you like it too\n$ # based on number of lines in each paragraph\n$ awk -F'\\n' -v RS= -v ORS='\\n\\n' 'NF==1' sample.txt\nHello World\n$ awk -F'\\n' -v RS= -v ORS='\\n\\n' 'NF==2 &amp;&amp; \/do\/' sample.txt\nJust do-it\nBelieve it\nMuch ado about nothing\nHe he he\n<\/pre><\/div>\n<ul><li>Re-structuring paragraphs<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # default FS is one or more of continuous space, tab or newline characters\n$ # default OFS is single space\n$ # so, $1=$1 will change it uniformly to single space between fields\n$ awk -v RS= '{$1=$1} 1' sample.txt\nHello World\nGood day How are you\nJust do-it Believe it\nToday is sunny Not a bit funny No doubt you like it too\nMuch ado about nothing He he he\n$ # a better usecase\n$ awk 'BEGIN{FS=\"\\n\"; OFS=\". \"; RS=\"\"; ORS=\"\\n\\n\"} {$1=$1} 1' sample.txt\nHello World\nGood day. How are you\nJust do-it. Believe it\nToday is sunny. Not a bit funny. No doubt you like it too\nMuch ado about nothing. He he he\n<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/unix.stackexchange.com\/questions\/359717\/select-line-with-empty-line-above-and-under\" rel=\"nofollow\">unix.stackexchange - filtering line surrounded by empty lines<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/46142118\/converting-regex-to-sed-or-grep-regex\" rel=\"nofollow\">stackoverflow - excellent example and explanation of RS and FS<\/a><\/li>\n<\/ul><br \/><h4><a id=\"user-content-multicharacter-rs\" class=\"anchor\" aria-hidden=\"true\" href=\"#multicharacter-rs\"><\/a>Multicharacter RS<\/h4>\n<ul><li>Some marker like <code>Error<\/code> or <code>Warning<\/code> etc<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat report.log\nblah blah\nError: something went wrong\nmore blah\nwhatever\nError: something surely went wrong\nsome text\nsome more text\nblah blah blah\n$ awk -v RS='Error:' 'END{print NR-1}' report.log\n2\n$ awk -v RS='Error:' 'NR==1' report.log\nblah blah\n$ # filter 'Error:' block matching particular string\n$ # to preserve formatting, use: '\/whatever\/{print RS $0}'\n$ awk -v RS='Error:' '\/whatever\/' report.log\n something went wrong\nmore blah\nwhatever\n$ # blocks with more than 3 lines\n$ # splitting string with 3 newlines will yield 4 fields\n$ awk -F'\\n' -v RS='Error:' 'NF&gt;4{print RS $0}' report.log\nError: something surely went wrong\nsome text\nsome more text\nblah blah blah\n<\/pre><\/div>\n<ul><li>Regular expression based <code>RS<\/code>\n<ul><li>the <code>RT<\/code> variable will contain string matched by <code>RS<\/code><\/li>\n<\/ul><\/li>\n<li>Note that entire input is treated as single string, so <code>^<\/code> and <code>$<\/code> anchors will apply only once - not every line<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ s='Sample123string54with908numbers'\n$ printf \"$s\" | awk -v RS='[0-9]+' 'NR==1'\nSample\n$ # note the relationship between record and separators\n$ printf \"$s\" | awk -v RS='[0-9]+' '{print NR \" : \" $0 \" - \" RT}'\n1 : Sample - 123\n2 : string - 54\n3 : with - 908\n4 : numbers - \n$ # need to be careful of empty records\n$ printf '123string54with908' | awk -v RS='[0-9]+' '{print NR \" : \" $0}'\n1 : \n2 : string\n3 : with\n$ # and newline at end of input\n$ printf '123string54with908\\n' | awk -v RS='[0-9]+' '{print NR \" : \" $0}'\n1 : \n2 : string\n3 : with\n4 : \n<\/pre><\/div>\n<ul><li>Joining lines based on specific end of line condition<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat msg.txt\nHello there.\nIt will rain to-\nday. Have a safe\nand pleasant jou-\nrney.\n$ # join lines ending with - to next line\n$ # by manipulating RS and ORS\n$ awk -v RS='-\\n' -v ORS= '1' msg.txt\nHello there.\nIt will rain today. Have a safe\nand pleasant journey.\n$ # by manipulating ORS alone, sub function covered in later sections\n$ awk '{ORS = sub(\/-$\/,\"\") ? \"\" : \"\\n\"} 1' msg.txt\nHello there.\nIt will rain today. Have a safe\nand pleasant journey.\n$ # easier: perl -pe 's\/-\\n\/\/' msg.txt as newline is still part of input line<\/pre><\/div>\n<ul><li>processing null terminated input<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ printf 'foo\\0bar\\0' | cat -A\nfoo^@bar^@$\n$ printf 'foo\\0bar\\0' | awk -v RS='\\0' '{print}'\nfoo\nbar<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Records.html#Records\" rel=\"nofollow\">gawk manual - Records<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/304457\/slurp-mode-in-awk\" rel=\"nofollow\">unix.stackexchange - Slurp-mode in awk<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/45102651\/how-to-grep-double-quote-followed-by-a-string-at-same-time\/45102962#45102962\" rel=\"nofollow\">stackoverflow - using RS to count number of occurrences of a given string<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-substitute-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#substitute-functions\"><\/a>Substitute functions<\/h2>\n<ul><li>Use <code>sub<\/code> string function for replacing first occurrence<\/li>\n<li>Use <code>gsub<\/code> for replacing all occurrences<\/li>\n<li>By default, <code>$0<\/code> which contains input record is modified, can specify any other field or variable as needed<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # replacing first occurrence\n$ echo '1-2-3-4-5' | awk '{sub(\"-\", \":\")} 1'\n1:2-3-4-5\n$ # replacing all occurrences\n$ echo '1-2-3-4-5' | awk '{gsub(\"-\", \":\")} 1'\n1:2:3:4:5\n$ # return value for sub\/gsub is number of replacements made\n$ echo '1-2-3-4-5' | awk '{n=gsub(\"-\", \":\"); print n} 1'\n4\n1:2:3:4:5\n$ # \/\/ format is better suited to specify search REGEXP\n$ echo '1-2-3-4-5' | awk '{gsub(\/[^-]+\/, \"abc\")} 1'\nabc-abc-abc-abc-abc\n$ # replacing all occurrences only for third field\n$ echo 'one;two;three;four' | awk -F';' '{gsub(\"e\", \"E\", $3)} 1'\none two thrEE four<\/pre><\/div>\n<ul><li>Use <code>gensub<\/code> to return the modified string unlike <code>sub<\/code> or <code>gsub<\/code> which modifies inplace<\/li>\n<li>it also supports back-references and ability to modify specific match<\/li>\n<li>acts upon <code>$0<\/code> if target is not specified<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # replace second occurrence\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\":\", \"-\", 2)} 1'\nfoo:123-bar:baz\n$ # use REGEXP as needed\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"XYZ\", 2)} 1'\nfoo:XYZ:bar:baz\n$ # or print the returned string directly\n$ echo 'foo:123:bar:baz' | awk '{print gensub(\":\", \"-\", 2)}'\nfoo:123-bar:baz\n$ # replace third occurrence\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"XYZ\", 3)} 1'\nfoo:123:XYZ:baz\n$ # replace all occurrences, similar to gsub\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"XYZ\", \"g\")} 1'\nXYZ:XYZ:XYZ:XYZ\n$ # target other than $0\n$ echo 'foo:123:bar:baz' | awk -F: -v OFS=: '{$1=gensub(\/o\/, \"b\", 2, $1)} 1'\nfob:123:bar:baz<\/pre><\/div>\n<ul><li>back-reference examples<\/li>\n<li>use <code>\\\"<\/code> within double-quotes to represent <code>\"<\/code> character in replacement string<\/li>\n<li>use <code>\\\\1<\/code> to represent <code>\\1<\/code> - the first captured group and so on<\/li>\n<li><code>&amp;<\/code> or <code>\\0<\/code> will back-reference entire matched string<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # replacing last occurrence without knowing how many occurrences are there\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/(.*):\/, \"\\\\1-\", 1)} 1'\nfoo:123:bar-baz\n$ echo 'foo and bar and baz land good' | awk '{$0=gensub(\/(.*)and\/, \"\\\\1XYZ\", 1)} 1'\nfoo and bar and baz lXYZ good\n$ # use word boundaries as necessary\n$ echo 'foo and bar and baz land good' | awk '{$0=gensub(\/(.*)\\&lt;and\\&gt;\/, \"\\\\1XYZ\", 1)} 1'\nfoo and bar XYZ baz land good\n$ # replacing last but one\n$ echo '456:foo:123:bar:789:baz' | awk '{$0=gensub(\/(.*):(.*:)\/, \"\\\\1-\\\\2\", 1)} 1'\n456:foo:123:bar-789:baz\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"\\\"&amp;\\\"\", \"g\")} 1'\n\"foo\":\"123\":\"bar\":\"baz\"<\/pre><\/div>\n<ul><li>saving quotes in variables - to avoid escaping double quotes or having to use octal code for single quotes<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"\\047&amp;\\047\", \"g\")} 1'\n'foo':'123':'bar':'baz'\n$ echo 'foo:123:bar:baz' | awk -v sq=\"'\" '{$0=gensub(\/[^:]+\/, sq\"&amp;\"sq, \"g\")} 1'\n'foo':'123':'bar':'baz'\n$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"\\\"&amp;\\\"\", \"g\")} 1'\n\"foo\":\"123\":\"bar\":\"baz\"\n$ echo 'foo:123:bar:baz' | awk -v dq='\"' '{$0=gensub(\/[^:]+\/, dq\"&amp;\"dq, \"g\")} 1'\n\"foo\":\"123\":\"bar\":\"baz\"<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/String-Functions.html\" rel=\"nofollow\">gawk manual - String-Manipulation Functions<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Gory-Details.html\" rel=\"nofollow\">gawk manual - escape processing<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-inplace-file-editing\" class=\"anchor\" aria-hidden=\"true\" href=\"#inplace-file-editing\"><\/a>Inplace file editing<\/h2>\n<ul><li>Use this option with caution, preferably after testing that the <code>awk<\/code> code is working as intended<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat greeting.txt\nHi there\nHave a nice day\n$ awk -i inplace '{gsub(\"e\", \"E\")} 1' greeting.txt\n$ cat greeting.txt\nHi thErE\nHavE a nicE day<\/pre><\/div>\n<ul><li>Multiple input files are treated individually and changes are written back to respective files<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat f1\nI ate 3 apples\n$ cat f2\nI bought two bananas and 3 mangoes\n$ awk -i inplace '{gsub(\"3\", \"three\")} 1' f1 f2\n$ cat f1\nI ate three apples\n$ cat f2\nI bought two bananas and three mangoes<\/pre><\/div>\n<ul><li>to create backups of original file, set <code>INPLACE_SUFFIX<\/code> variable<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk -i inplace -v INPLACE_SUFFIX='.bkp' '{gsub(\"three\", \"3\")} 1' f1\n$ cat f1\nI ate 3 apples\n$ cat f1.bkp\nI ate three apples<\/pre><\/div>\n<ul><li>See <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Extension-Sample-Inplace.html\" rel=\"nofollow\">gawk manual - Enabling In-Place File Editing<\/a> for implementation details<\/li>\n<\/ul><br \/><h2><a id=\"user-content-using-shell-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-shell-variables\"><\/a>Using shell variables<\/h2>\n<ul><li>when <code>awk<\/code> code is part of shell program and shell variable needs to be passed as input to <code>awk<\/code> code<\/li>\n<li>for example:\n<ul><li>command line argument passed to shell script, which is in turn passed on to <code>awk<\/code><\/li>\n<li>control structures in shell script calling <code>awk<\/code> with different search strings<\/li>\n<\/ul><\/li>\n<li>See also <a href=\"https:\/\/stackoverflow.com\/questions\/19075671\/how-do-i-use-shell-variables-in-an-awk-script\" rel=\"nofollow\">stackoverflow - How do I use shell variables in an awk script?<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # examples tested with bash shell\n$ f='apple'\n$ awk -v word=\"$f\" '$1==word' fruits.txt\napple   42\n$ f='fig'\n$ awk -v word=\"$f\" '$1==word' fruits.txt\nfig     90\n$ q='20'\n$ awk -v threshold=\"$q\" 'NR==1 || $2&gt;threshold' fruits.txt\nfruit   qty\napple   42\nbanana  31\nfig     90<\/pre><\/div>\n<ul><li>accessing shell environment variables<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # existing environment variable\n$ awk 'BEGIN{print ENVIRON[\"PWD\"]}'\n\/home\/learnbyexample\n$ awk 'BEGIN{print ENVIRON[\"SHELL\"]}'\n\/bin\/bash\n$ # defined along with awk code\n$ word='hello world' awk 'BEGIN{print ENVIRON[\"word\"]}'\nhello world\n$ # using ENVIRON also prevents awk's interpretation of escape sequences\n$ s='a\\n=c'\n$ foo=\"$s\" awk 'BEGIN{print ENVIRON[\"foo\"]}'\na\\n=c\n$ awk -v foo=\"$s\" 'BEGIN{print foo}'\na\n=c<\/pre><\/div>\n<ul><li>passing <em>REGEXP<\/em><\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Computed-Regexps.html\" rel=\"nofollow\">gawk manual - Using Dynamic Regexps<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ s='are'\n$ # for: awk '!\/are\/' poem.txt\n$ awk -v s=\"$s\" '$0 !~ s' poem.txt\nSugar is sweet,\n$ # for: awk '\/are\/ &amp;&amp; !\/so\/' poem.txt\n$ awk -v s=\"$s\" '$0 ~ s &amp;&amp; !\/so\/' poem.txt\nRoses are red,\nViolets are blue,\n$ r='[^-]+'\n$ echo '1-2-3-4-5' | awk -v r=\"$r\" '{gsub(r, \"abc\")} 1'\nabc-abc-abc-abc-abc\n$ # escape sequence has to be doubled when string is interpreted as REGEXP\n$ s='foo and bar and baz land good'\n$ echo \"$s\" | awk '{$0=gensub(\"(.*)\\\\&lt;and\\\\&gt;\", \"\\\\1XYZ\", 1)} 1'\nfoo and bar XYZ baz land good\n$ # hence passing as variable should be\n$ r='(.*)\\\\&lt;and\\\\&gt;'\n$ echo \"$s\" | awk -v r=\"$r\" '{$0=gensub(r, \"\\\\1XYZ\", 1)} 1'\nfoo and bar XYZ baz land good\n$ # or use ENVIRON\n$ r='(.*)\\&lt;and\\&gt;'\n$ echo \"$s\" | r=\"$r\" awk '{$0=gensub(ENVIRON[\"r\"], \"\\\\1XYZ\", 1)} 1'\nfoo and bar XYZ baz land good<\/pre><\/div>\n<br \/><h2><a id=\"user-content-multiple-file-input\" class=\"anchor\" aria-hidden=\"true\" href=\"#multiple-file-input\"><\/a>Multiple file input<\/h2>\n<ul><li>Example to show difference between <code>NR<\/code> and <code>FNR<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # NR for overall record number\n$ awk 'NR==1' poem.txt greeting.txt\nRoses are red,\n$ # FNR for individual file's record number\n$ # same as: head -q -n1 poem.txt greeting.txt\n$ awk 'FNR==1' poem.txt greeting.txt\nRoses are red,\nHi thErE<\/pre><\/div>\n<ul><li>Constructs to do some processing before starting each file as well as at the end<\/li>\n<li><code>BEGINFILE<\/code> - to add code to be executed before start of each input file<\/li>\n<li><code>ENDFILE<\/code> - to add code to be executed after processing each input file<\/li>\n<li><code>FILENAME<\/code> - file name of current input file being processed<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # similar to: tail -n1 poem.txt greeting.txt\n$ awk 'BEGINFILE{print \"file: \"FILENAME}\n       ENDFILE{print $0\"\\n------\"}' poem.txt greeting.txt\nfile: poem.txt\nAnd so are you.\n------\nfile: greeting.txt\nHavE a nicE day\n------<\/pre><\/div>\n<ul><li>And of course, there can be usual <code>awk<\/code> code<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'BEGINFILE{print \"file: \"FILENAME}\n       FNR==1;\n       ENDFILE{print \"------\"}' poem.txt greeting.txt\nfile: poem.txt\nRoses are red,\n------\nfile: greeting.txt\nHi thErE\n------\n$ awk 'BEGINFILE{c++; print \"file: \"FILENAME}\n       FNR==2;\n       END{print \"\\nTotal input files: \"c}' poem.txt greeting.txt\nfile: poem.txt\nViolets are blue,\nfile: greeting.txt\nHavE a nicE day\nTotal input files: 2<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/ARGC-and-ARGV.html\" rel=\"nofollow\">gawk manual - Using ARGC and ARGV<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Auto_002dset.html#index-ARGIND-variable\" rel=\"nofollow\">gawk manual - ARGIND<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Auto_002dset.html#index-ERRNO-variable\" rel=\"nofollow\">gawk manual - ERRNO<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/a\/43473385\/4082052\" rel=\"nofollow\">stackoverflow - Finding common value across multiple files<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-control-structures\" class=\"anchor\" aria-hidden=\"true\" href=\"#control-structures\"><\/a>Control Structures<\/h2>\n<ul><li>Syntax is similar to <code>C<\/code> language and single statements inside control structures don't require to be grouped within <code>{}<\/code><\/li>\n<li>See <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Statements.html\" rel=\"nofollow\">gawk manual - Control Statements<\/a> for details<\/li>\n<\/ul><p>Remember that by default there is a loop that goes over all input records and constructs like <code>BEGIN<\/code> and <code>END<\/code> fall outside that loop<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat nums.txt\n42\n-2\n10101\n-3.14\n-75\n$ awk '{sum += $1} END{print sum}' nums.txt\n10062.9\n$ # uninitialized variables will have empty string\n$ printf '' | awk '{sum += $1} END{print sum}'\n$ # so either add '0' or use unary '+' operator to convert to number\n$ printf '' | awk '{sum += $1} END{print +sum}'\n0\n$ awk '{sum += $1} END{print sum+0}' \/dev\/null\n0<\/pre><\/div>\n<ul><li>See also <a href=\"https:\/\/unix.stackexchange.com\/questions\/421904\/regression-with-unary-plus\" rel=\"nofollow\">unix.stackexchange - change in behavior of unary + with gawk version 4.2.0<\/a><\/li>\n<\/ul><br \/><h4><a id=\"user-content-if-else-and-loops\" class=\"anchor\" aria-hidden=\"true\" href=\"#if-else-and-loops\"><\/a>if-else and loops<\/h4>\n<ul><li>We have already seen simple <code>if<\/code> examples in <a href=\"#filtering\">Filtering<\/a> section<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Switch-Statement.html\" rel=\"nofollow\">gawk manual - Switch<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # same as: sed -n '\/are\/ s\/so\/SO\/p' poem.txt\n$ # remember that sub\/gsub returns number of substitutions made\n$ awk '\/are\/{if(sub(\"so\", \"SO\")) print}' poem.txt\nAnd SO are you.\n$ # of course, can also use\n$ awk '\/are\/ &amp;&amp; sub(\"so\", \"SO\")' poem.txt\nAnd SO are you.\n$ # if-else example\n$ awk 'NR&gt;1{if($2&gt;40) $0=\"+\"$0; else $0=\"-\"$0} 1' fruits.txt\nfruit   qty\n+apple   42\n-banana  31\n+fig     90\n-guava   6<\/pre><\/div>\n<ul><li>ternary operator<\/li>\n<li>See also <a href=\"https:\/\/stackoverflow.com\/a\/29784278\/4082052\" rel=\"nofollow\">stackoverflow - finding min and max value of a column<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat nums.txt\n42\n-2\n10101\n-3.14\n-75\n$ # changing -ve to +ve and vice versa\n$ # same as: awk '{if($0 ~ \/^-\/) sub(\/^-\/,\"\"); else sub(\/^\/,\"-\")} 1' nums.txt\n$ awk '{$0 ~ \/^-\/ ? sub(\/^-\/,\"\") : sub(\/^\/,\"-\")} 1' nums.txt\n-42\n2\n-10101\n3.14\n75\n$ # can also use: awk '!sub(\/^-\/,\"\"){sub(\/^\/,\"-\")} 1' nums.txt<\/pre><\/div>\n<ul><li>for loop<\/li>\n<li>similar to <code>C<\/code> language, <code>break<\/code> and <code>continue<\/code> statements are also available<\/li>\n<li>See also <a href=\"https:\/\/stackoverflow.com\/questions\/38491676\/how-can-i-find-the-missing-integers-in-a-unique-and-sequential-list-one-per-lin\" rel=\"nofollow\">stackoverflow - find missing numbers from sequential list<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'BEGIN{for(i=2; i&lt;11; i+=2) print i}'\n2\n4\n6\n8\n10\n$ # looping each field\n$ s='scat:cat:no cat:abdicate:cater'\n$ echo \"$s\" | awk -F: -v OFS=: '{for(i=1;i&lt;=NF;i++) if($i==\"cat\") $i=\"CAT\"} 1'\nscat:CAT:no cat:abdicate:cater\n$ # can also use sub function\n$ echo \"$s\" | awk -F: -v OFS=: '{for(i=1;i&lt;=NF;i++) sub(\/^cat$\/,\"CAT\",$i)} 1'\nscat:CAT:no cat:abdicate:cater<\/pre><\/div>\n<ul><li>while loop<\/li>\n<li>do-while is also available<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'BEGIN{i=2; while(i&lt;11){print i; i+=2}}'\n2\n4\n6\n8\n10\n$ # recursive substitution\n$ # here again return value of sub\/gsub is useful\n$ echo 'titillate' | awk '{while( gsub(\/til\/, \"\") ) print}'\ntilate\nate<\/pre><\/div>\n<br \/><h4><a id=\"user-content-next-and-nextfile\" class=\"anchor\" aria-hidden=\"true\" href=\"#next-and-nextfile\"><\/a>next and nextfile<\/h4>\n<ul><li><code>next<\/code> will skip rest of statements and start processing next line of current file being processed\n<ul><li>there is a loop by default which goes over all input records, <code>next<\/code> is applicable for that<\/li>\n<li>it is similar to <code>continue<\/code> statement within loops<\/li>\n<\/ul><\/li>\n<li>it is often used in <a href=\"#two-file-processing\">Two file processing<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # here 'next' is used to skip processing header line\n$ awk 'NR==1{print; next} \/a.*a\/{$0=\"*\"$0} \/[eiou]\/{$0=\"-\"$0} 1' fruits.txt\nfruit   qty\n-apple   42\n*banana  31\n-fig     90\n-*guava   6<\/pre><\/div>\n<ul><li><code>nextfile<\/code> is useful to skip remaining lines from current file being processed and move on to next file<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # same as: head -q -n1 poem.txt greeting.txt fruits.txt\n$ awk 'FNR&gt;1{nextfile} 1' poem.txt greeting.txt fruits.txt\nRoses are red,\nHi thErE\nfruit   qty\n$ # specific field\n$ awk 'FNR&gt;2{nextfile} {print $1}' poem.txt greeting.txt fruits.txt\nRoses\nViolets\nHi\nHavE\nfruit\napple\n$ # similar to 'grep -il'\n$ awk -v IGNORECASE=1 '\/red\/{print FILENAME; nextfile}' *\ncolors_1.txt\ncolors_2.txt\npoem.txt\n$ awk -v IGNORECASE=1 '$1 ~ \/red\/{print FILENAME; nextfile}' *\ncolors_1.txt\ncolors_2.txt<\/pre><\/div>\n<br \/><h2><a id=\"user-content-multiline-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#multiline-processing\"><\/a>Multiline processing<\/h2>\n<ul><li>Processing consecutive lines<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat poem.txt\nRoses are red,\nViolets are blue,\nSugar is sweet,\nAnd so are you.\n$ # match two consecutive lines\n$ awk 'p~\/are\/ &amp;&amp; \/is\/{print p ORS $0} {p=$0}' poem.txt\nViolets are blue,\nSugar is sweet,\n$ # if only the second line is needed\n$ awk 'p~\/are\/ &amp;&amp; \/is\/; {p=$0}' poem.txt\nSugar is sweet,\n$ # match three consecutive lines\n$ awk 'p2~\/red\/ &amp;&amp; p1~\/blue\/ &amp;&amp; \/is\/{print p2} {p2=p1; p1=$0}' poem.txt\nRoses are red,\n$ # common mistake\n$ sed -n '\/are\/{N;\/is\/p}' poem.txt\n$ # would need something like this and not practical to extend for other cases\n$ sed '$!N; \/are.*\\n.*is\/p; D' poem.txt\nViolets are blue,\nSugar is sweet,<\/pre><\/div>\n<p>Consider this sample input file<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat range.txt\nfoo\nBEGIN\n1234\n6789\nEND\nbar\nBEGIN\na\nb\nc\nEND\nbaz<\/pre><\/div>\n<ul><li>extracting lines around matching line<\/li>\n<li>See also <a href=\"https:\/\/stackoverflow.com\/questions\/17908555\/printing-with-sed-or-awk-a-line-following-a-matching-pattern\" rel=\"nofollow\">stackoverflow - lines around matching regexp<\/a><\/li>\n<li>how <code>n &amp;&amp; n--<\/code> works:\n<ul><li>need to note that right hand side of <code>&amp;&amp;<\/code> is processed only if left hand side is <code>true<\/code><\/li>\n<li>so for example, if initially <code>n=2<\/code>, then we get\n<ul><li><code>2 &amp;&amp; 2; n=1<\/code> - evaluates to <code>true<\/code><\/li>\n<li><code>1 &amp;&amp; 1; n=0<\/code> - evaluates to <code>true<\/code><\/li>\n<li><code>0 &amp;&amp; <\/code> - evaluates to <code>false<\/code> ... no decrementing <code>n<\/code> and hence will be <code>false<\/code> until <code>n<\/code> is re-assigned non-zero value<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # similar to: grep --no-group-separator -A1 'BEGIN' range.txt\n$ awk '\/BEGIN\/{n=2} n &amp;&amp; n--' range.txt\nBEGIN\n1234\nBEGIN\na\n$ # only print the line after matching line\n$ # can also use: awk '\/BEGIN\/{n=1; next} n &amp;&amp; n--' range.txt\n$ awk 'n &amp;&amp; n--; \/BEGIN\/{n=1}' range.txt\n1234\na\n$ # generic case: print nth line after match\n$ awk 'n &amp;&amp; !--n; \/BEGIN\/{n=3}' range.txt\nEND\nc\n$ # print second line prior to matched line\n$ awk '\/END\/{print p2} {p2=p1; p1=$0}' range.txt\n1234\nb\n$ # save all lines in an array for generic case\n$ awk '\/END\/{print a[NR-3]} {a[NR]=$0}' range.txt\nBEGIN\na\n$ # or use the reversing trick\n$ tac range.txt | awk 'n &amp;&amp; !--n; \/END\/{n=3}' | tac\nBEGIN\na<\/pre><\/div>\n<ul><li>Checking if multiple strings are present at least once in entire input file<\/li>\n<li>If there are lots of strings to check, use arrays<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # can also use BEGINFILE instead of FNR==1\n$ awk 'FNR==1{s1=s2=0} \/is\/{s1=1} \/are\/{s2=1} s1&amp;&amp;s2{print FILENAME; nextfile}' *\npoem.txt\nsample.txt\n$ awk 'FNR==1{s1=s2=0} \/foo\/{s1=1} \/report\/{s2=1} s1&amp;&amp;s2{print FILENAME; nextfile}' *\npaths.txt<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/stackoverflow.com\/questions\/49112877\/delete-line-if-line-matches-foo-line-above-matches-bar-and-line-below-match\" rel=\"nofollow\">stackoverflow - delete line based on content of previous\/next lines<\/a><\/li>\n<li><a href=\"https:\/\/softwareengineering.stackexchange.com\/questions\/47806\/examples-of-finite-state-machines\" rel=\"nofollow\">softwareengineering - FSM examples<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Finite-state_machine\" rel=\"nofollow\">wikipedia - FSM<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-two-file-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#two-file-processing\"><\/a>Two file processing<\/h2>\n<ul><li>We'll use awk's associative arrays (key-value pairs) here\n<ul><li>key can be number or string<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Arrays.html\" rel=\"nofollow\">gawk manual - Arrays<\/a><\/li>\n<\/ul><\/li>\n<li>Unlike <a href=\"https:\/\/github.com\/learnbyexample\/Command-line-text-processing\/blob\/master\/sorting_stuff.md#comm\">comm<\/a> the input files need not be sorted and comparison can be done based on certain field(s) as well<\/li>\n<\/ul><br \/><h4><a id=\"user-content-comparing-whole-lines\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparing-whole-lines\"><\/a>Comparing whole lines<\/h4>\n<p>Consider the following test files<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat colors_1.txt\nBlue\nBrown\nPurple\nRed\nTeal\nYellow\n$ cat colors_2.txt\nBlack\nBlue\nGreen\nRed\nWhite<\/pre><\/div>\n<ul><li>common lines and lines unique to one of the files<\/li>\n<li>For two files as input, <code>NR==FNR<\/code> will be true only when first file is being processed<\/li>\n<li>Using <code>next<\/code> will skip rest of code when first file is processed<\/li>\n<li><code>a[$0]<\/code> will create unique keys (here entire line content is used as key) in array <code>a<\/code>\n<ul><li>just referencing a key will create it if it doesn't already exist, with value as empty string (will also act as zero in numeric context)<\/li>\n<\/ul><\/li>\n<li><code>$0 in a<\/code> will be true if key already exists in array <code>a<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # common lines\n$ # same as: grep -Fxf colors_1.txt colors_2.txt\n$ awk 'NR==FNR{a[$0]; next} $0 in a' colors_1.txt colors_2.txt\nBlue\nRed\n$ # lines from colors_2.txt not present in colors_1.txt\n$ # same as: grep -vFxf colors_1.txt colors_2.txt\n$ awk 'NR==FNR{a[$0]; next} !($0 in a)' colors_1.txt colors_2.txt\nBlack\nGreen\nWhite\n$ # reversing the order of input files gives\n$ # lines from colors_1.txt not present in colors_2.txt\n$ awk 'NR==FNR{a[$0]; next} !($0 in a)' colors_2.txt colors_1.txt\nBrown\nPurple\nTeal\nYellow<\/pre><\/div>\n<br \/><h4><a id=\"user-content-comparing-specific-fields\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparing-specific-fields\"><\/a>Comparing specific fields<\/h4>\n<p>Consider the sample input file<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat marks.txt\nDept    Name    Marks\nECE     Raj     53\nECE     Joel    72\nEEE     Moi     68\nCSE     Surya   81\nEEE     Tia     59\nECE     Om      92\nCSE     Amy     67<\/pre><\/div>\n<ul><li>single field<\/li>\n<li>For ex: only first field comparison by using <code>$1<\/code> instead of <code>$0<\/code> as key<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat list1\nECE\nCSE\n$ # extract only lines matching first field specified in list1\n$ awk 'NR==FNR{a[$1]; next} $1 in a' list1 marks.txt\nECE     Raj     53\nECE     Joel    72\nCSE     Surya   81\nECE     Om      92\nCSE     Amy     67\n$ # if header is needed as well\n$ awk 'NR==FNR{a[$1]; next} FNR==1 || $1 in a' list1 marks.txt\nDept    Name    Marks\nECE     Raj     53\nECE     Joel    72\nCSE     Surya   81\nECE     Om      92\nCSE     Amy     67<\/pre><\/div>\n<ul><li>multiple fields<\/li>\n<li>create a string by adding some character between the fields to act as key\n<ul><li>for ex: to avoid matching two field values <code>abc<\/code> and <code>123<\/code> to match with two other field values <code>ab<\/code> and <code>c123<\/code><\/li>\n<li>by adding character, say <code>_<\/code>, the key would be <code>abc_123<\/code> for first case and <code>ab_c123<\/code> for second case<\/li>\n<li>this can still lead to false match if input data has <code>_<\/code><\/li>\n<li>there is also a built-in way to do this using <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Multidimensional.html#Multidimensional\" rel=\"nofollow\">gawk manual - Multidimensional Arrays<\/a><\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat list2\nEEE Moi\nCSE Amy\nECE Raj\n$ # extract only lines matching both fields specified in list2\n$ awk 'NR==FNR{a[$1\"_\"$2]; next} $1\"_\"$2 in a' list2 marks.txt\nECE     Raj     53\nEEE     Moi     68\nCSE     Amy     67\n$ # uses SUBSEP as separator, whose default value is non-printing character \\034\n$ awk 'NR==FNR{a[$1,$2]; next} ($1,$2) in a' list2 marks.txt\nECE     Raj     53\nEEE     Moi     68\nCSE     Amy     67<\/pre><\/div>\n<ul><li>field and value comparison<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat list3\nECE 70\nEEE 65\nCSE 80\n$ # extract line matching Dept and minimum marks specified in list3\n$ awk 'NR==FNR{d[$1]; m[$1]=$2; next} $1 in d &amp;&amp; $3 &gt;= m[$1]' list3 marks.txt\nECE     Joel    72\nEEE     Moi     68\nCSE     Surya   81\nECE     Om      92<\/pre><\/div>\n<br \/><h4><a id=\"user-content-getline\" class=\"anchor\" aria-hidden=\"true\" href=\"#getline\"><\/a>getline<\/h4>\n<ul><li><code>getline<\/code> is an alternative way to read from a file and could be faster than <code>NR==FNR<\/code> method for some cases<\/li>\n<li>But use it with caution\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Getline.html\" rel=\"nofollow\">gawk manual - getline<\/a> for details, especially about corner cases, errors, etc<\/li>\n<li><a href=\"https:\/\/web.archive.org\/web\/20170524214527\/http:\/\/awk.freeshell.org\/AllAboutGetline\" rel=\"nofollow\">getline caveats<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Close-Files-And-Pipes.html\" rel=\"nofollow\">gawk manual - Closing Input and Output Redirections<\/a> if you have to start from beginning of file again<\/li>\n<\/ul><\/li>\n<li><code>getline<\/code> return value: <code>1<\/code> if record is found, <code>0<\/code> if end of file, <code>-1<\/code> for errors such as file not found (use <code>ERRNO<\/code> variable to get details)<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # replace mth line in poem.txt with nth line from nums.txt\n$ # return value handling is not shown here, but should be done ideally\n$ awk -v m=3 -v n=2 'BEGIN{while(n-- &gt; 0) getline s &lt; \"nums.txt\"}\n                     FNR==m{$0=s} 1' poem.txt\nRoses are red,\nViolets are blue,\n-2\nAnd so are you.\n$ # without getline, but slower due to NR==FNR check for every line processed\n$ awk -v m=3 -v n=2 'NR==FNR{if(FNR==n){s=$0; nextfile} next}\n                     FNR==m{$0=s} 1' nums.txt poem.txt\nRoses are red,\nViolets are blue,\n-2\nAnd so are you.\n$ # Note that if nums.txt has less than n lines:\n$ # getline version will use last line of nums.txt if any\n$ # NR==FNR version will give empty string as 's' would be uninitialized<\/pre><\/div>\n<ul><li>Another use case is if two files are to be processed simultaneously<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # print line from fruits.txt if corresponding line from nums.txt is +ve number\n$ # the return value check ensures corresponding line number comparison\n$ awk -v file='nums.txt' '(getline num &lt; file)==1 &amp;&amp; num&gt;0' fruits.txt\nfruit   qty\nbanana  31\n$ # without getline, but has to save entire file in array\n$ awk 'NR==FNR{n[FNR]=$0; next} n[FNR]&gt;0' nums.txt fruits.txt\nfruit   qty\nbanana  31<\/pre><\/div>\n<ul><li>error handling<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'NR==FNR{n[FNR]=$0; next} n[FNR]&gt;0' xyz.txt fruits.txt\nawk: fatal: cannot open file 'xyz.txt' for reading (No such file or directory)\n$ awk -v file='xyz.txt' '{ e=(getline num &lt; file);\n                           if(e&lt;0){print file \": \" ERRNO; exit} }\n                         e==1 &amp;&amp; num&gt;0' fruits.txt\nxyz.txt: No such file or directory<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/stackoverflow.com\/questions\/42239179\/fastest-way-to-find-lines-of-a-text-file-from-another-larger-text-file-in-bash\" rel=\"nofollow\">stackoverflow - Fastest way to find lines of a text file from another larger text file<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/320651\/read-numbers-from-control-file-and-extract-matching-line-numbers-from-the-data-f\" rel=\"nofollow\">unix.stackexchange - filter lines based on line numbers specified in another file<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/45036019\/how-to-filter-the-values-from-selected-columns-and-rows\" rel=\"nofollow\">stackoverflow - three file processing to extract a matrix subset<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/294145\/merging-two-files-one-column-at-a-time\" rel=\"nofollow\">unix.stackexchange - column wise merging<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/40595990\/print-many-specific-rows-from-a-text-file-using-an-index-file\" rel=\"nofollow\">stackoverflow - extract specific rows from a text file using an index file<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-creating-new-fields\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-new-fields\"><\/a>Creating new fields<\/h2>\n<ul><li>Number of fields in input record can be changed by simply manipulating <code>NF<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # reducing fields\n$ echo 'foo,bar,123,baz' | awk -F, -v OFS=, '{NF=2} 1'\nfoo,bar\n$ # creating new empty field(s)\n$ echo 'foo,bar,123,baz' | awk -F, -v OFS=, '{NF=5} 1'\nfoo,bar,123,baz,\n$ # assigning to field greater than NF will create empty fields as needed\n$ echo 'foo,bar,123,baz' | awk -F, -v OFS=, '{$7=42} 1'\nfoo,bar,123,baz,,,42<\/pre><\/div>\n<ul><li>adding a field based on existing fields<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # adding a new 'Grade' field\n$ awk 'BEGIN{OFS=\"\\t\"; g[9]=\"S\"; g[8]=\"A\"; g[7]=\"B\"; g[6]=\"C\"; g[5]=\"D\"}\n      {NF++; $NF = NR==1 ? \"Grade\" : g[int($(NF-1)\/10)]} 1' marks.txt\nDept    Name    Marks   Grade\nECE     Raj     53      D\nECE     Joel    72      B\nEEE     Moi     68      C\nCSE     Surya   81      A\nEEE     Tia     59      D\nECE     Om      92      S\nCSE     Amy     67      C\n$ # can also use split (covered in a later section)\n$ # array assignment: split(\"DCBAS\",g,\/\/)\n$ # index adjustment: g[int($(NF-1)\/10)-4]<\/pre><\/div>\n<ul><li>two file example<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat list4\nRaj class_rep\nAmy sports_rep\nTia placement_rep\n$ awk -v OFS='\\t' 'NR==FNR{r[$1]=$2; next}\n         {NF++; $NF = NR==1 ? \"Role\" : $NF=r[$2]} 1' list4 marks.txt\nDept    Name    Marks   Role\nECE     Raj     53      class_rep\nECE     Joel    72\nEEE     Moi     68\nCSE     Surya   81\nEEE     Tia     59      placement_rep\nECE     Om      92\nCSE     Amy     67      sports_rep<\/pre><\/div>\n<br \/><h2><a id=\"user-content-dealing-with-duplicates\" class=\"anchor\" aria-hidden=\"true\" href=\"#dealing-with-duplicates\"><\/a>Dealing with duplicates<\/h2>\n<ul><li>default value of uninitialized variable is <code>0<\/code> in numeric context and empty string in text context\n<ul><li>and evaluates to <code>false<\/code> when used conditionally<\/li>\n<\/ul><\/li>\n<\/ul><p><em>Illustration to show default numeric value and array in action<\/em><\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ printf 'mad\\n42\\n42\\ndam\\n42\\n'\nmad\n42\n42\ndam\n42\n$ printf 'mad\\n42\\n42\\ndam\\n42\\n' | awk '{print $0 \"\\t\" int(a[$0]); a[$0]++}'\nmad     0\n42      0\n42      1\ndam     0\n42      2\n$ # only those entries with second column value zero will be retained\n$ printf 'mad\\n42\\n42\\ndam\\n42\\n' | awk '!a[$0]++'\nmad\n42\ndam<\/pre><\/div>\n<ul><li>first, examples that retain only first copy of duplicates<\/li>\n<li>See also <a href=\"https:\/\/stackoverflow.com\/questions\/47774779\/add-letter-to-second-third-fourth-occurrence-of-a-string\" rel=\"nofollow\">stackoverflow - add a letter to duplicate entries<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat duplicates.txt\nabc  7   4\nfood toy ****\nabc  7   4\ntest toy 123\ngood toy ****\n$ # whole line\n$ awk '!seen[$0]++' duplicates.txt\nabc  7   4\nfood toy ****\ntest toy 123\ngood toy ****\n$ # particular column\n$ awk '!seen[$2]++' duplicates.txt\nabc  7   4\nfood toy ****\n$ # total count\n$ awk '!seen[$2]++{c++} END{print +c}' duplicates.txt\n2<\/pre><\/div>\n<ul><li>if input is so large that integer numbers can overflow<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Arbitrary-Precision-Integers.html\" rel=\"nofollow\">gawk manual - Arbitrary-Precision Integer Arithmetic<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # avoid unnecessary counting altogether\n$ awk '!($2 in seen); {seen[$2]}' duplicates.txt\nabc  7   4\nfood toy ****\n$ # use arbitrary-precision integers, limited only by available memory\n$ awk -M '!($2 in seen){c++} {seen[$2]} END{print +c}' duplicates.txt\n2<\/pre><\/div>\n<ul><li>For multiple fields, separate them using <code>,<\/code> or form a string with some character in between\n<ul><li>choose a character unlikely to appear in input data, else there can be false matches<\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk '!seen[$2\"_\"$3]++' duplicates.txt\nabc  7   4\nfood toy ****\ntest toy 123\n$ # can also use simulated multidimensional array\n$ # SUBSEP, whose default is \\034 non-printing character, is used as separator\n$ awk '!seen[$2,$3]++' duplicates.txt\nabc  7   4\nfood toy ****\ntest toy 123<\/pre><\/div>\n<ul><li>retaining specific numbered copy<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # second occurrence of duplicate\n$ awk '++seen[$2]==2' duplicates.txt\nabc  7   4\ntest toy 123\n$ # third occurrence of duplicate\n$ awk '++seen[$2]==3' duplicates.txt\ngood toy ****<\/pre><\/div>\n<ul><li>retaining only last copy of duplicate<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # reverse the input line-wise, retain first copy and then reverse again\n$ tac duplicates.txt | awk '!seen[$2]++' | tac\nabc  7   4\ngood toy ****<\/pre><\/div>\n<ul><li>filtering based on duplicate count<\/li>\n<li>allows to emulate <a href=\"https:\/\/github.com\/learnbyexample\/Command-line-text-processing\/blob\/master\/sorting_stuff.md#uniq\">uniq<\/a> command for specific fields<\/li>\n<li>See also <a href=\"https:\/\/unix.stackexchange.com\/questions\/362571\/filter-out-paths-from-a-text-file-that-are-deeper-than-their-immediate-predecces\" rel=\"nofollow\">unix.stackexchange - retain only parent directory paths<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # all duplicates based on 1st column\n$ awk 'NR==FNR{a[$1]++; next} a[$1]&gt;1' duplicates.txt duplicates.txt\nabc  7   4\nabc  7   4\n$ # all duplicates based on 3rd column\n$ awk 'NR==FNR{a[$3]++; next} a[$3]&gt;1' duplicates.txt duplicates.txt\nabc  7   4\nfood toy ****\nabc  7   4\ngood toy ****\n$ # more than 2 duplicates based on 2nd column\n$ awk 'NR==FNR{a[$2]++; next} a[$2]&gt;2' duplicates.txt duplicates.txt\nfood toy ****\ntest toy 123\ngood toy ****\n$ # only unique lines based on 3rd column\n$ awk 'NR==FNR{a[$3]++; next} a[$3]==1' duplicates.txt duplicates.txt\ntest toy 123<\/pre><\/div>\n<br \/><h2><a id=\"user-content-lines-between-two-regexps\" class=\"anchor\" aria-hidden=\"true\" href=\"#lines-between-two-regexps\"><\/a>Lines between two REGEXPs<\/h2>\n<ul><li>This section deals with filtering lines bound by two <em>REGEXP<\/em>s (referred to as blocks)<\/li>\n<li>For simplicity the two <em>REGEXP<\/em>s usually used in below examples are the strings <strong>BEGIN<\/strong> and <strong>END<\/strong><\/li>\n<\/ul><br \/><h4><a id=\"user-content-all-unbroken-blocks\" class=\"anchor\" aria-hidden=\"true\" href=\"#all-unbroken-blocks\"><\/a>All unbroken blocks<\/h4>\n<p>Consider the below sample input file, which doesn't have any unbroken blocks (i.e <strong>BEGIN<\/strong> and <strong>END<\/strong> are always present in pairs)<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat range.txt\nfoo\nBEGIN\n1234\n6789\nEND\nbar\nBEGIN\na\nb\nc\nEND\nbaz<\/pre><\/div>\n<ul><li>Extracting lines between starting and ending <em>REGEXP<\/em><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # include both starting\/ending REGEXP\n$ # can also use: awk '\/BEGIN\/,\/END\/' range.txt\n$ # which is similar to sed -n '\/BEGIN\/,\/END\/p'\n$ # but not suitable to extend for other cases\n$ awk '\/BEGIN\/{f=1} f; \/END\/{f=0}' range.txt\nBEGIN\n1234\n6789\nEND\nBEGIN\na\nb\nc\nEND\n$ # exclude both starting\/ending REGEXP\n$ # can also use: awk '\/BEGIN\/{f=1; next} \/END\/{f=0} f' range.txt\n$ awk '\/END\/{f=0} f; \/BEGIN\/{f=1}' range.txt\n1234\n6789\na\nb\nc<\/pre><\/div>\n<ul><li>Include only start or end <em>REGEXP<\/em><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # include only starting REGEXP\n$ awk '\/BEGIN\/{f=1} \/END\/{f=0} f' range.txt\nBEGIN\n1234\n6789\nBEGIN\na\nb\nc\n$ # include only ending REGEXP\n$ awk 'f; \/END\/{f=0} \/BEGIN\/{f=1}' range.txt\n1234\n6789\nEND\na\nb\nc\nEND<\/pre><\/div>\n<ul><li>Extracting lines other than lines between the two <em>REGEXP<\/em>s<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk '\/BEGIN\/{f=1} !f; \/END\/{f=0}' range.txt\nfoo\nbar\nbaz\n$ # the other three cases would be\n$ awk '\/END\/{f=0} !f; \/BEGIN\/{f=1}' range.txt\n$ awk '!f; \/BEGIN\/{f=1} \/END\/{f=0}' range.txt\n$ awk '\/BEGIN\/{f=1} \/END\/{f=0} !f' range.txt<\/pre><\/div>\n<br \/><h4><a id=\"user-content-specific-blocks\" class=\"anchor\" aria-hidden=\"true\" href=\"#specific-blocks\"><\/a>Specific blocks<\/h4>\n<ul><li>Getting first block<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk '\/BEGIN\/{f=1} f; \/END\/{exit}' range.txt\nBEGIN\n1234\n6789\nEND\n$ # use other tricks discussed in previous section as needed\n$ awk '\/END\/{exit} f; \/BEGIN\/{f=1}' range.txt\n1234\n6789<\/pre><\/div>\n<ul><li>Getting last block<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # reverse input linewise, change the order of REGEXPs, finally reverse again\n$ tac range.txt | awk '\/END\/{f=1} f; \/BEGIN\/{exit}' | tac\nBEGIN\na\nb\nc\nEND\n$ # or, save the blocks in a buffer and print the last one alone\n$ # ORS contains output record separator, which is newline by default\n$ seq 30 | awk '\/4\/{f=1; b=$0; next} f{b=b ORS $0} \/6\/{f=0} END{print b}'\n24\n25\n26<\/pre><\/div>\n<ul><li>Getting blocks based on a counter<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # all blocks\n$ seq 30 | sed -n '\/4\/,\/6\/p'\n4\n5\n6\n14\n15\n16\n24\n25\n26\n$ # get only 2nd block\n$ # can also use: seq 30 | awk -v b=2 '\/4\/{c++} c==b{print; if(\/6\/) exit}'\n$ seq 30 | awk -v b=2 '\/4\/{c++} c==b; \/6\/ &amp;&amp; c==b{exit}'\n14\n15\n16\n$ # to get all blocks greater than 'b' blocks\n$ seq 30 | awk -v b=1 '\/4\/{f=1; c++} f &amp;&amp; c&gt;b; \/6\/{f=0}'\n14\n15\n16\n24\n25\n26<\/pre><\/div>\n<ul><li>excluding a particular block<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # excludes 2nd block\n$ seq 30 | awk -v b=2 '\/4\/{f=1; c++} f &amp;&amp; c!=b; \/6\/{f=0}'\n4\n5\n6\n24\n25\n26<\/pre><\/div>\n<br \/><h4><a id=\"user-content-broken-blocks\" class=\"anchor\" aria-hidden=\"true\" href=\"#broken-blocks\"><\/a>Broken blocks<\/h4>\n<ul><li>If there are blocks with ending <em>REGEXP<\/em> but without corresponding start, <code>awk '\/BEGIN\/{f=1} f; \/END\/{f=0}'<\/code> will suffice<\/li>\n<li>Consider the modified input file where starting <em>REGEXP<\/em> doesn't have corresponding ending<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat broken_range.txt\nfoo\nBEGIN\n1234\n6789\nEND\nbar\nBEGIN\na\nb\nc\nbaz\n$ # the file reversing trick comes in handy here as well\n$ tac broken_range.txt | awk '\/END\/{f=1} f; \/BEGIN\/{f=0}' | tac\nBEGIN\n1234\n6789\nEND<\/pre><\/div>\n<ul><li>But if both kinds of broken blocks are present, accumulate the records and print accordingly<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat multiple_broken.txt\nqqqqqqq\nBEGIN\nfoo\nBEGIN\n1234\n6789\nEND\nbar\nEND\n0-42-1\nBEGIN\na\nBEGIN\nb\nEND\nxyzabc\n$ awk '\/BEGIN\/{f=1; buf=$0; next}\n       f{buf=buf ORS $0}\n       \/END\/{f=0; if(buf) print buf; buf=\"\"}' multiple_broken.txt\nBEGIN\n1234\n6789\nEND\nBEGIN\nb\nEND<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/stackoverflow.com\/questions\/38972736\/how-to-select-lines-between-two-patterns\" rel=\"nofollow\">stackoverflow - select lines between two regexps<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/295600\/deleting-lines-between-rows-in-a-text-file-using-awk-or-sed\" rel=\"nofollow\">unix.stackexchange - print only blocks with lines &gt; n<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/a\/335523\/109046\" rel=\"nofollow\">unix.stackexchange - print a block only if it contains matching string<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/347368\/grep-with-range-and-pass-three-filters\" rel=\"nofollow\">unix.stackexchange - print a block matching two different strings<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/404175\/using-awk-to-print-lines-from-one-match-through-a-second-instance-of-a-separate\" rel=\"nofollow\">unix.stackexchange - extract block up to 2nd occurrence of ending REGEXP<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-arrays\" class=\"anchor\" aria-hidden=\"true\" href=\"#arrays\"><\/a>Arrays<\/h2>\n<p>We've already seen examples using arrays, some more examples discussed in this section<\/p>\n<ul><li>array looping<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # average marks for each department\n$ awk 'NR&gt;1{d[$1]+=$3; c[$1]++} END{for(i in d)print i, d[i]\/c[i]}' marks.txt\nECE 72.3333\nEEE 63.5\nCSE 74<\/pre><\/div>\n<ul><li>Sorting<\/li>\n<li>See <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Controlling-Scanning.html#Controlling-Scanning\" rel=\"nofollow\">gawk manual - Predefined Array Scanning Orders<\/a> for more details<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # by default, keys are traversed in random order\n$ awk 'BEGIN{a[\"z\"]=1; a[\"x\"]=12; a[\"b\"]=42; for(i in a)print i, a[i]}'\nx 12\nz 1\nb 42\n$ # index sorted ascending order as strings\n$ awk 'BEGIN{PROCINFO[\"sorted_in\"] = \"@ind_str_asc\";\n       a[\"z\"]=1; a[\"x\"]=12; a[\"b\"]=42; for(i in a)print i, a[i]}'\nb 42\nx 12\nz 1\n$ # value sorted ascending order as numbers\n$ awk 'BEGIN{PROCINFO[\"sorted_in\"] = \"@val_num_asc\";\n       a[\"z\"]=1; a[\"x\"]=12; a[\"b\"]=42; for(i in a)print i, a[i]}'\nz 1\nx 12\nb 42<\/pre><\/div>\n<ul><li>deleting array elements<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat list5\nCSE     Surya   75\nEEE     Jai     69\nECE     Kal     83\n$ # update entry if a match is found\n$ # else append the new entries\n$ awk '{ky=$1\"_\"$2} NR==FNR{upd[ky]=$0; next}\n        ky in upd{$0=upd[ky]; delete upd[ky]} 1;\n        END{for(i in upd)print upd[i]}' list5 marks.txt\nDept    Name    Marks\nECE     Raj     53\nECE     Joel    72\nEEE     Moi     68\nCSE     Surya   75\nEEE     Tia     59\nECE     Om      92\nCSE     Amy     67\nECE     Kal     83\nEEE     Jai     69<\/pre><\/div>\n<ul><li>true multidimensional arrays<\/li>\n<li>length of sub-arrays need not be same. See <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Arrays-of-Arrays.html#Arrays-of-Arrays\" rel=\"nofollow\">gawk manual - Arrays of Arrays<\/a> for details<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'NR&gt;1{d[$1][$2]=$3} END{for(i in d[\"ECE\"])print i}' marks.txt\nJoel\nRaj\nOm\n$ awk -v f='CSE' 'NR&gt;1{d[$1][$2]=$3} END{for(i in d[f])print i, d[f][i]}' marks.txt\nSurya 81\nAmy 67<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Arrays.html\" rel=\"nofollow\">gawk manual - all array topics<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/396855\/is-there-an-easy-way-to-count-characters-in-words-in-file-from-terminal\" rel=\"nofollow\">unix.stackexchange - count words based on length<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/a\/326215\/109046\" rel=\"nofollow\">unix.stackexchange - filtering specific lines<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-awk-scripts\" class=\"anchor\" aria-hidden=\"true\" href=\"#awk-scripts\"><\/a>awk scripts<\/h2>\n<ul><li>For larger programs, save the code in a file and use <code>-f<\/code> command line option<\/li>\n<li><code>;<\/code> is not needed to terminate a statement<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Options.html#Options\" rel=\"nofollow\">gawk manual - Command-Line Options<\/a> for other related options<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat buf.awk\n\/BEGIN\/{\n    f=1\n    buf=$0\n    next\n}\nf{\n    buf=buf ORS $0\n}\n\/END\/{\n    f=0\n    if(buf)\n        print buf\n    buf=\"\"\n}\n$ awk -f buf.awk multiple_broken.txt\nBEGIN\n1234\n6789\nEND\nBEGIN\nb\nEND<\/pre><\/div>\n<ul><li>Another advantage is that single quotes can be freely used<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ echo 'foo:123:bar:baz' | awk '{$0=gensub(\/[^:]+\/, \"\\047&amp;\\047\", \"g\")} 1'\n'foo':'123':'bar':'baz'\n$ cat quotes.awk\n{\n    $0 = gensub(\/[^:]+\/, \"'&amp;'\", \"g\")\n}\n1\n$ echo 'foo:123:bar:baz' | awk -f quotes.awk\n'foo':'123':'bar':'baz'<\/pre><\/div>\n<ul><li>If the code has been first tried out on command line, add <code>-o<\/code> option to get a pretty printed version<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk -o -v OFS='\\t' 'NR==FNR{r[$1]=$2; next}\n         {NF++; if(FNR==1)$NF=\"Role\"; else $NF=r[$2]} 1' list4 marks.txt\nDept    Name    Marks   Role\nECE     Raj     53      class_rep\nECE     Joel    72\nEEE     Moi     68\nCSE     Surya   81\nEEE     Tia     59      placement_rep\nECE     Om      92\nCSE     Amy     67      sports_rep<\/pre><\/div>\n<p>File name can be passed along <code>-o<\/code> option, otherwise by default <code>awkprof.out<\/code> will be used<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cat awkprof.out\n        # gawk profile, created Tue Oct 24 15:10:02 2017\n        # Rule(s)\n        NR == FNR {\n                r[$1] = $2\n                next\n        }\n        {\n                NF++\n                if (FNR == 1) {\n                        $NF = \"Role\"\n                } else {\n                        $NF = r[$2]\n                }\n        }\n        1 {\n                print $0\n        }\n$ # note that other command line options have to be provided as usual\n$ # for ex: awk -v OFS='\\t' -f awkprof.out list4 marks.txt<\/pre><\/div>\n<br \/><h2><a id=\"user-content-miscellaneous\" class=\"anchor\" aria-hidden=\"true\" href=\"#miscellaneous\"><\/a>Miscellaneous<\/h2>\n<br \/><h4><a id=\"user-content-fpat-and-fieldwidths\" class=\"anchor\" aria-hidden=\"true\" href=\"#fpat-and-fieldwidths\"><\/a>FPAT and FIELDWIDTHS<\/h4>\n<ul><li><code>FS<\/code> allows to define field separator<\/li>\n<li>In contrast, <code>FPAT<\/code> allows to define what should the fields be made up of<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Splitting-By-Content.html\" rel=\"nofollow\">gawk manual - Defining Fields by Content<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ s='Sample123string54with908numbers'\n$ # define fields to be one or more consecutive digits\n$ echo \"$s\" | awk -v FPAT='[0-9]+' '{print $1, $2, $3}'\n123 54 908\n$ # define fields to be one or more consecutive alphabets\n$ echo \"$s\" | awk -v FPAT='[a-zA-Z]+' '{print $1, $2, $3, $4}'\nSample string with numbers<\/pre><\/div>\n<ul><li>For simpler <strong>csv<\/strong> input having quoted strings if fields themselves have <code>,<\/code> in them, using <code>FPAT<\/code> is reasonable approach<\/li>\n<li>Use a proper parser if input can have other cases like newlines in fields\n<ul><li>See <a href=\"https:\/\/unix.stackexchange.com\/a\/238192\" rel=\"nofollow\">unix.stackexchange - using csv parser<\/a> for a sample program in <code>perl<\/code><\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ s='foo,\"bar,123\",baz,abc'\n$ echo \"$s\" | awk -F, '{print $2}'\n\"bar\n$ echo \"$s\" | awk -v FPAT='\"[^\"]*\"|[^,]*' '{print $2}'\n\"bar,123\"<\/pre><\/div>\n<ul><li>if input has well defined fields based on number of characters, <code>FIELDWIDTHS<\/code> can be used to specify width of each field<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk -v FIELDWIDTHS='8 3' -v OFS= '\/fig\/{$2=35} 1' fruits.txt\nfruit   qty\napple   42\nbanana  31\nfig     35\nguava   6\n$ # without FIELDWIDTHS\n$ awk '\/fig\/{$2=35} 1' fruits.txt\nfruit   qty\napple   42\nbanana  31\nfig 35\nguava   6<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Fixed-width-data.html\" rel=\"nofollow\">gawk manual - Processing Fixed-Width Data<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/368574\/modify-records-in-fixed-width-files\" rel=\"nofollow\">unix.stackexchange - Modify records in fixed-width files<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/321559\/extracting-data-with-awk-when-some-lines-have-empty-missing-values\" rel=\"nofollow\">unix.stackexchange - detecting empty fields in fixed width files<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/37450880\/how-do-i-filter-tab-separated-input-by-the-count-of-fields-with-a-given-value\" rel=\"nofollow\">stackoverflow - count number of times value is repeated each line<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/46932189\/how-do-you-skip-characters-with-fieldwidths-in-gnu-awk-4-2\" rel=\"nofollow\">stackoverflow - skip characters with FIELDWIDTHS in GNU Awk 4.2<\/a><\/li>\n<\/ul><br \/><h4><a id=\"user-content-string-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#string-functions\"><\/a>String functions<\/h4>\n<ul><li><code>length<\/code> function - returns length of string, by default acts on <code>$0<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ seq 8 13 | awk 'length()==1'\n8\n9\n$ awk 'NR==1 || length($1)&gt;4' fruits.txt\nfruit   qty\napple   42\nbanana  31\nguava   6\n$ # character count and not byte count is calculated, similar to 'wc -m'\n$ printf 'hi\ud83d\udc4d' | awk '{print length()}'\n3\n$ # use -b option if number of bytes are needed\n$ printf 'hi\ud83d\udc4d' | awk -b '{print length()}'\n6<\/pre><\/div>\n<ul><li><code>split<\/code> function - similar to <code>FS<\/code> splitting input record into fields<\/li>\n<li>use <code>patsplit<\/code> function to get results similar to <code>FPAT<\/code><\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/gawk.html#index-split_0028_0029-function\" rel=\"nofollow\">gawk manual - Split function<\/a><\/li>\n<li>See also <a href=\"https:\/\/unix.stackexchange.com\/questions\/372253\/awk-command-to-delimit-the-second-column\" rel=\"nofollow\">unix.stackexchange - delimit second column<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # 1st argument is string to be split\n$ # 2nd argument is array to save results, indexed from 1\n$ # 3rd argument is separator, default is FS\n$ s='foo,1996-10-25,hello,good'\n$ echo \"$s\" | awk -F, '{split($2,d,\"-\"); print \"Month is: \" d[2]}'\nMonth is: 10\n$ # using regular expression to define separator\n$ # return value is number of fields after splitting\n$ s='Sample123string54with908numbers'\n$ echo \"$s\" | awk '{n=split($0,s,\/[0-9]+\/); for(i=1;i&lt;=n;i++)print s[i]}'\nSample\nstring\nwith\nnumbers\n$ # use 4th argument if separators are needed as well\n$ echo \"$s\" | awk '{n=split($0,s,\/[0-9]+\/,seps); for(i=1;i&lt;n;i++)print seps[i]}'\n123\n54\n908\n$ # single row to multiple rows based on splitting last field\n$ s='foo,baz,12:42:3'\n$ echo \"$s\" | awk -F, '{n=split($NF,a,\":\"); NF--; for(i=1;i&lt;=n;i++) print $0,a[i]}'\nfoo baz 12\nfoo baz 42\nfoo baz 3<\/pre><\/div>\n<ul><li><code>substr<\/code> function allows to extract specified number of characters from given string\n<ul><li>indexing starts with <code>1<\/code><\/li>\n<\/ul><\/li>\n<li>See <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/gawk.html#index-substr_0028_0029-function\" rel=\"nofollow\">gawk manual - substr function<\/a> for corner cases and details<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # 1st argument is string to be worked on\n$ # 2nd argument is starting position\n$ # 3rd argument is number of characters to be extracted\n$ echo 'abcdefghij' | awk '{print substr($0,1,5)}'\nabcde\n$ echo 'abcdefghij' | awk '{print substr($0,4,3)}'\ndef\n$ # if 3rd argument is not given, string is extracted until end\n$ echo 'abcdefghij' | awk '{print substr($0,6)}'\nfghij\n$ echo 'abcdefghij' | awk -v OFS=':' '{print substr($0,2,3), substr($0,6,3)}'\nbcd:fgh\n$ # if only few characters are needed from input line, can use empty FS\n$ echo 'abcdefghij' | awk -v FS= '{print $3}'\nc\n$ echo 'abcdefghij' | awk -v FS= '{print $3, $5}'\nc e<\/pre><\/div>\n<br \/><h4><a id=\"user-content-executing-external-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#executing-external-commands\"><\/a>Executing external commands<\/h4>\n<ul><li>External commands can be issued using <code>system<\/code> function<\/li>\n<li>Output would be as usual on <code>stdout<\/code> unless redirected while calling the command<\/li>\n<li>Return value of <code>system<\/code> depends on <code>exit<\/code> status of executed command, see <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/I_002fO-Functions.html\" rel=\"nofollow\">gawk manual - Input\/Output Functions<\/a> for details<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'BEGIN{system(\"echo Hello World\")}'\nHello World\n$ wc poem.txt\n 4 13 65 poem.txt\n$ awk 'BEGIN{system(\"wc poem.txt\")}'\n 4 13 65 poem.txt\n$ awk 'BEGIN{system(\"seq 10 | paste -sd, &gt; out.txt\")}'\n$ cat out.txt\n1,2,3,4,5,6,7,8,9,10\n$ ls xyz.txt\nls: cannot access 'xyz.txt': No such file or directory\n$ echo $?\n2\n$ awk 'BEGIN{s=system(\"ls xyz.txt\"); print \"Status: \" s}'\nls: cannot access 'xyz.txt': No such file or directory\nStatus: 2\n$ cat f2\nI bought two bananas and three mangoes\n$ echo 'f1,f2,odd.txt' | awk -F, '{system(\"cat \" $2)}'\nI bought two bananas and three mangoes<\/pre><\/div>\n<br \/><h4><a id=\"user-content-printf-formatting\" class=\"anchor\" aria-hidden=\"true\" href=\"#printf-formatting\"><\/a>printf formatting<\/h4>\n<ul><li>Similar to <code>printf<\/code> function in <code>C<\/code> and shell built-in command<\/li>\n<li>use <code>sprintf<\/code> function to save result in variable instead of printing<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Printf.html\" rel=\"nofollow\">gawk manual - printf<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk '{sum += $1} END{print sum}' nums.txt\n10062.9\n$ # note that ORS is not appended and has to be added manually\n$ awk '{sum += $1} END{printf \"%.2f\\n\", sum}' nums.txt\n10062.86\n$ awk '{sum += $1} END{printf \"%10.2f\\n\", sum}' nums.txt\n  10062.86\n$ awk '{sum += $1} END{printf \"%010.2f\\n\", sum}' nums.txt\n0010062.86\n$ awk '{sum += $1} END{printf \"%d\\n\", sum}' nums.txt\n10062\n$ awk '{sum += $1} END{printf \"%+d\\n\", sum}' nums.txt\n+10062\n$ awk '{sum += $1} END{printf \"%e\\n\", sum}' nums.txt\n1.006286e+04<\/pre><\/div>\n<ul><li>to refer argument by positional number (starts with 1), use <code>&lt;num&gt;$<\/code><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # can also use: awk 'BEGIN{printf \"hex=%x\\noct=%o\\ndec=%d\\n\", 15, 15, 15}'\n$ awk 'BEGIN{printf \"hex=%1$x\\noct=%1$o\\ndec=%1$d\\n\", 15}'\nhex=f\noct=17\ndec=15\n$ # adding prefix to hex\/oct numbers\n$ awk 'BEGIN{printf \"hex=%1$#x\\noct=%1$#o\\ndec=%1$d\\n\", 15}'\nhex=0xf\noct=017\ndec=15<\/pre><\/div>\n<ul><li>strings<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # prefix remaining width with spaces\n$ awk 'BEGIN{printf \"%6s:%5s\\n\", \"foo\", \"bar\"}'\n   foo:  bar\n$ # suffix remaining width with spaces\n$ awk 'BEGIN{printf \"%-6s:%-5s\\n\", \"foo\", \"bar\"}'\nfoo   :bar  \n$ # truncate\n$ awk 'BEGIN{printf \"%.2s\\n\", \"foobar\"}'\nfo<\/pre><\/div>\n<ul><li>avoid using <code>printf<\/code> without format specifier<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk 'BEGIN{s=\"solve: 5 % x = 1\"; printf s}'\nawk: cmd. line:1: fatal: not enough arguments to satisfy format string\n    `solve: 5 % x = 1'\n               ^ ran out for this one\n$ awk 'BEGIN{s=\"solve: 5 % x = 1\"; printf \"%s\\n\", s}'\nsolve: 5 % x = 1<\/pre><\/div>\n<ul><li>See also <a href=\"https:\/\/stackoverflow.com\/questions\/49135518\/linux-csv-file-concatenate-columns-into-one-column\" rel=\"nofollow\">stackoverflow - concatenating columns in middle<\/a><\/li>\n<\/ul><br \/><h4><a id=\"user-content-redirecting-print-output\" class=\"anchor\" aria-hidden=\"true\" href=\"#redirecting-print-output\"><\/a>Redirecting print output<\/h4>\n<ul><li>redirecting to file instead of stdout using <code>&gt;<\/code><\/li>\n<li>similar to behavior in shell, if file already exists it is overwritten\n<ul><li>use <code>&gt;&gt;<\/code> to append to an existing file without deleting content<\/li>\n<\/ul><\/li>\n<li>however, unlike shell, subsequent redirections to same file will append to it<\/li>\n<li>See also <a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Close-Files-And-Pipes.html\" rel=\"nofollow\">gawk manual - Closing Input and Output Redirections<\/a> if you have too many redirections<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ seq 6 | awk 'NR%2{print &gt; \"odd.txt\"; next} {print &gt; \"even.txt\"}'\n$ cat odd.txt\n1\n3\n5\n$ cat even.txt\n2\n4\n6\n$ awk 'NR==1{col1=$1\".txt\"; col2=$2\".txt\"; next}\n       {print $1 &gt; col1; print $2 &gt; col2}' fruits.txt\n$ cat fruit.txt\napple\nbanana\nfig\nguava\n$ cat qty.txt\n42\n31\n90\n6<\/pre><\/div>\n<ul><li>redirecting to shell command<\/li>\n<li>this is useful if you have different things to redirect to different commands, otherwise it can be done as usual in shell acting on <code>awk<\/code>'s output<\/li>\n<li>all redirections to same command gets combined as single input to that command<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # same as: echo 'foo good 123' | awk '{print $2}' | wc -c\n$ echo 'foo good 123' | awk '{print $2 | \"wc -c\"}'\n5\n$ # to avoid newline character being added to print\n$ echo 'foo good 123' | awk -v ORS= '{print $2 | \"wc -c\"}'\n4\n$ # assuming no format specifiers in input\n$ echo 'foo good 123' | awk '{printf $2 | \"wc -c\"}'\n4\n$ # same as: echo 'foo good 123' | awk '{printf $2 $3 | \"wc -c\"}'\n$ echo 'foo good 123' | awk '{printf $2 | \"wc -c\"; printf $3 | \"wc -c\"}'\n7<\/pre><\/div>\n<p><strong>Further Reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/I_002fO-Functions.html\" rel=\"nofollow\">gawk manual - Input\/Output Functions<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Redirection.html\" rel=\"nofollow\">gawk manual - Redirecting Output of print and printf<\/a><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/html_node\/Two_002dway-I_002fO.html\" rel=\"nofollow\">gawk manual - Two-Way Communications with Another Process<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/321679\/gawk-inplace-and-stdout\" rel=\"nofollow\">unix.stackexchange - inplace editing as well as stdout<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/45098279\/write-blocks-in-a-text-file-to-multiple-new-files\" rel=\"nofollow\">stackoverflow - redirect blocks to separate files<\/a><\/li>\n<\/ul><br \/><h2><a id=\"user-content-gotchas-and-tips\" class=\"anchor\" aria-hidden=\"true\" href=\"#gotchas-and-tips\"><\/a>Gotchas and Tips<\/h2>\n<ul><li>using <code>$<\/code> for variables<\/li>\n<li>only input record <code>$0<\/code> and field contents <code>$1<\/code>, <code>$2<\/code> etc need <code>$<\/code><\/li>\n<li>See also <a href=\"https:\/\/unix.stackexchange.com\/questions\/291126\/why-does-awk-print-the-whole-line-when-i-want-it-to-print-a-variable\" rel=\"nofollow\">unix.stackexchange - Why does awk print the whole line when I want it to print a variable?<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # wrong\n$ awk -v word=\"apple\" '$1==$word' fruits.txt\n$ # right\n$ awk -v word=\"apple\" '$1==word' fruits.txt\napple   42<\/pre><\/div>\n<ul><li>dos style line endings<\/li>\n<li>See also <a href=\"https:\/\/unix.stackexchange.com\/questions\/399560\/using-awk-to-select-rows-with-specific-value-in-specific-column\" rel=\"nofollow\">unix.stackexchange - filtering when last column has \\r<\/a><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # no issue with unix style line ending\n$ printf 'foo bar\\n123 789\\n' | awk '{print $2, $1}'\nbar foo\n789 123\n$ # dos style line ending causes trouble\n$ printf 'foo bar\\r\\n123 789\\r\\n' | awk '{print $2, $1}'\n foo\n 123\n$ # easy to deal by simply setting appropriate RS\n$ # note that ORS would still be newline character only\n$ printf 'foo bar\\r\\n123 789\\r\\n' | awk -v RS='\\r\\n' '{print $2, $1}'\nbar foo\n789 123<\/pre><\/div>\n<ul><li>relying on default initial value<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # step 1 - works for single file\n$ awk '{sum += $1} END{print sum}' nums.txt\n10062.9\n$ # step 2 - change to work for multiple file\n$ awk '{sum += $1} ENDFILE{print FILENAME, sum}' nums.txt\nnums.txt 10062.9\n$ # step 3 - check with multiple file input\n$ # oops, default numerical value '0' for sum works only once\n$ awk '{sum += $1} ENDFILE{print FILENAME, sum}' nums.txt &lt;(seq 3)\nnums.txt 10062.9\n\/dev\/fd\/63 10068.9\n$ # step 4 - correctly initialize variables\n$ awk 'BEGINFILE{sum=0} {sum += $1} ENDFILE{print FILENAME, sum}' nums.txt &lt;(seq 3)\nnums.txt 10062.9\n\/dev\/fd\/63 6<\/pre><\/div>\n<ul><li>use unary operator <code>+<\/code> to force numeric conversion<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ awk '{sum += $1} END{print FILENAME, sum}' nums.txt\nnums.txt 10062.9\n$ awk '{sum += $1} END{print FILENAME, sum}' \/dev\/null\n\/dev\/null \n$ awk '{sum += $1} END{print FILENAME, +sum}' \/dev\/null\n\/dev\/null 0<\/pre><\/div>\n<ul><li>concatenate empty string to force string comparison<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ echo '5 5.0' | awk '{print $1==$2 ? \"same\" : \"different\", \"string\"}'\nsame string\n$ echo '5 5.0' | awk '{print $1\"\"==$2 ? \"same\" : \"different\", \"string\"}'\ndifferent string<\/pre><\/div>\n<ul><li>beware of expressions going -ve for field calculations<\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ cat misc.txt\nfoo\ngood bad ugly\n123 xyz\na b c d\n$ # trying to delete last two fields\n$ awk '{NF -= 2} 1' misc.txt\nawk: cmd. line:1: (FILENAME=misc.txt FNR=1) fatal: NF set to negative value\n$ # dynamically change it depending on number of fields\n$ awk '{NF = (NF&lt;=2) ? 0 : NF-2} 1' misc.txt\ngood\na b\n$ # similarly, trying to access 3rd field from end\n$ awk '{print $(NF-2)}' misc.txt\nawk: cmd. line:1: (FILENAME=misc.txt FNR=1) fatal: attempt to access field -1\n$ awk 'NF&gt;2{print $(NF-2)}' misc.txt\ngood\nb<\/pre><\/div>\n<ul><li>If input is ASCII alone, simple trick to improve speed<\/li>\n<li>For simple non-regex based column filtering, using <a href=\"https:\/\/github.com\/learnbyexample\/Command-line-text-processing\/blob\/master\/miscellaneous.md#cut\">cut<\/a> command might give faster results\n<ul><li>See <a href=\"https:\/\/stackoverflow.com\/questions\/46882557\/how-to-split-columns-faster-in-python\/46883120#46883120\" rel=\"nofollow\">stackoverflow - how to split columns faster<\/a> for example<\/li>\n<\/ul><\/li>\n<\/ul><div class=\"highlight highlight-source-shell\"><pre>$ # all words containing exactly 3 lowercase a\n$ time awk -F'a' 'NF==4{cnt++} END{print +cnt}' \/usr\/share\/dict\/words\n1019\nreal    0m0.075s\n$ time LC_ALL=C awk -F'a' 'NF==4{cnt++} END{print +cnt}' \/usr\/share\/dict\/words\n1019\nreal    0m0.045s<\/pre><\/div>\n<br \/><h2><a id=\"user-content-further-reading\" class=\"anchor\" aria-hidden=\"true\" href=\"#further-reading\"><\/a>Further Reading<\/h2>\n<ul><li>Manual and related\n<ul><li><code>man awk<\/code> and <code>info awk<\/code> for quick reference from command line<\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/gawk\/manual\/gawk.html#SEC_Contents\" rel=\"nofollow\">gawk manual<\/a> for complete reference, extensions and more<\/li>\n<li><a href=\"http:\/\/www.faqs.org\/faqs\/computer-lang\/awk\/faq\/\" rel=\"nofollow\">awk FAQ<\/a> - from 2002, but plenty of information, especially about all the various <code>awk<\/code> implementations<\/li>\n<\/ul><\/li>\n<li>What's up with different <code>awk<\/code> versions?\n<ul><li><a href=\"https:\/\/unix.stackexchange.com\/questions\/29576\/difference-between-gawk-vs-awk\" rel=\"nofollow\">unix.stackexchange - brief explanation<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/awk\/comments\/4omosp\/differences_between_gawk_nawk_mawk_and_posix_awk\/\" rel=\"nofollow\">Differences between gawk, nawk, mawk, and POSIX awk<\/a><\/li>\n<li><a href=\"http:\/\/www.catonmat.net\/download\/awk.cheat.sheet.txt\" rel=\"nofollow\">cheat sheet for awk\/nawk\/gawk<\/a><\/li>\n<\/ul><\/li>\n<li>Tutorials and Q&amp;A\n<ul><li><a href=\"https:\/\/code.snipcademy.com\/tutorials\/shell-scripting\/awk\/introduction\" rel=\"nofollow\">code.snipcademy - gentle intro<\/a><\/li>\n<li><a href=\"https:\/\/www.funtoo.org\/Awk_by_Example,_Part_1\" rel=\"nofollow\">funtoo - using examples<\/a><\/li>\n<li><a href=\"http:\/\/www.grymoire.com\/Unix\/Awk.html\" rel=\"nofollow\">grymoire - detailed tutorial<\/a> - covers information about different <code>awk<\/code> versions as well<\/li>\n<li><a href=\"http:\/\/www.catonmat.net\/series\/awk-one-liners-explained\" rel=\"nofollow\">catonmat - one liners explained<\/a><\/li>\n<li><a href=\"https:\/\/blog.jpalardy.com\/posts\/why-learn-awk\/\" rel=\"nofollow\">Why Learn AWK?<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/tagged\/awk?sort=votes&amp;pageSize=15\" rel=\"nofollow\">awk Q&amp;A on stackoverflow<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/tagged\/awk?sort=votes&amp;pageSize=15\" rel=\"nofollow\">awk Q&amp;A on unix.stackexchange<\/a><\/li>\n<\/ul><\/li>\n<li>Alternatives\n<ul><li><a href=\"https:\/\/www.gnu.org\/software\/datamash\/alternatives\/\" rel=\"nofollow\">GNU datamash<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/lh3\/bioawk\">bioawk<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/gelisam\/hawk\/blob\/master\/doc\/README.md\">hawk<\/a> - based on Haskell<\/li>\n<li><a href=\"https:\/\/github.com\/johnkerl\/miller\">miller<\/a> - similar to awk\/sed\/cut\/join\/sort for name-indexed data such as CSV, TSV, and tabular JSON\n<ul><li>See this <a href=\"https:\/\/news.ycombinator.com\/item?id=10066742\" rel=\"nofollow\">ycombinator news<\/a> for other tools like this<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li>miscellaneous\n<ul><li><a href=\"https:\/\/unix.stackexchange.com\/questions\/303044\/when-to-use-grep-less-awk-sed\" rel=\"nofollow\">unix.stackexchange - When to use grep, sed, awk, perl, etc<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/e36freak\/awk-libs\">awk-libs<\/a> - lots of useful functions<\/li>\n<li><a href=\"https:\/\/github.com\/TheMozg\/awk-raycaster\">awkaster<\/a> - Pseudo-3D shooter written completely in awk using raycasting technique<\/li>\n<li><a href=\"https:\/\/awk.js.org\/\" rel=\"nofollow\">awk REPL<\/a> - live editor on browser<\/li>\n<\/ul><\/li>\n<li>examples for some of the stuff not covered in this tutorial\n<ul><li><a href=\"https:\/\/unix.stackexchange.com\/questions\/372816\/awk-get-random-lines-of-file-satisfying-a-condition\" rel=\"nofollow\">unix.stackexchange - rand\/srand<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/224969\/current-date-in-awk\" rel=\"nofollow\">unix.stackexchange - strftime<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/222146\/awk-does-not-end\/222150#222150\" rel=\"nofollow\">unix.stackexchange - ARGC and ARGV<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/46904447\/strange-output-while-comparing-engineering-numbers-in-awk\" rel=\"nofollow\">stackoverflow - arbitrary precision integer extension<\/a><\/li>\n<li><a href=\"https:\/\/stackoverflow.com\/questions\/3683110\/how-to-make-calculations-on-hexadecimal-numbers-with-awk\" rel=\"nofollow\">stackoverflow - recognizing hexadecimal numbers<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/223727\/splitting-file-for-every-10000-numbers-not-lines\/223739#223739\" rel=\"nofollow\">unix.stackexchange - sprintf and close<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/72469\/gawk-passing-arrays-to-functions\" rel=\"nofollow\">unix.stackexchange - user defined functions and array passing<\/a><\/li>\n<li><a href=\"https:\/\/unix.stackexchange.com\/questions\/408742\/count-number-of-columns-in-csv-files-and-rename-if-less-than-11-columns\" rel=\"nofollow\">unix.stackexchange - rename csv files based on number of fields in header row<\/a><\/li>\n<\/ul><\/li>\n<\/ul>","created_at":"2018-11-07T16:54:47+0000","updated_at":"2018-11-07T16:54:47+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":56,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/17766317?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":669,"title":"Global warming","url":"https:\/\/en.wikipedia.org\/wiki\/Global_warming","content":"<a class=\"mw-jump-link\" href=\"#mw-head\">Jump to navigation<\/a>\n\t\t\t\t<a class=\"mw-jump-link\" href=\"#p-search\">Jump to search<\/a>\n\t\t\t\t<div id=\"mw-content-text\" lang=\"en\" dir=\"ltr\" class=\"mw-content-ltr\"><div class=\"mw-parser-output\">\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">This page is about the current warming of the Earth's climate system. \"<a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change\" title=\"Climate change\">Climate change<\/a>\" can also refer to climate trends at any point in Earth's history. For other uses see <a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_(disambiguation)\" class=\"mw-disambig\" title=\"Global warming (disambiguation)\">Global warming (disambiguation)<\/a>.<\/div>\n<p class=\"mw-empty-elt\">\n\n<\/p><p class=\"mw-empty-elt\">\n<\/p>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Global_Temperature_Anomaly.svg\" class=\"image\"><img alt=\"refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f8\/Global_Temperature_Anomaly.svg\/330px-Global_Temperature_Anomaly.svg.png\" width=\"330\" height=\"290\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f8\/Global_Temperature_Anomaly.svg\/495px-Global_Temperature_Anomaly.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f8\/Global_Temperature_Anomaly.svg\/660px-Global_Temperature_Anomaly.svg.png 2x\" data-file-width=\"951\" data-file-height=\"835\" \/><\/a>  <div class=\"thumbcaption\">Global mean surface-temperature change from 1880 to 2017, relative to the 1951\u20131980 mean. The 1951\u20131980 mean is 14.19 \u00b0C (57.54 \u00b0F).<sup id=\"cite_ref-landocean_1-0\" class=\"reference\"><a href=\"#cite_note-landocean-1\">[1]<\/a><\/sup> The black line is the global annual mean, and the red line is the five-year <a href=\"https:\/\/en.wikipedia.org\/wiki\/Local_regression\" title=\"Local regression\">local regression<\/a> line. The blue <a href=\"https:\/\/en.wikipedia.org\/wiki\/Error_bar\" title=\"Error bar\">uncertainty bars<\/a> show a 95% confidence interval.<\/div><\/div><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:All_forcing_agents_CO2_equivalent_concentration.png\" class=\"image\"><img alt=\"Refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/3\/3e\/All_forcing_agents_CO2_equivalent_concentration.png\/330px-All_forcing_agents_CO2_equivalent_concentration.png\" width=\"330\" height=\"239\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/3\/3e\/All_forcing_agents_CO2_equivalent_concentration.png\/495px-All_forcing_agents_CO2_equivalent_concentration.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/3\/3e\/All_forcing_agents_CO2_equivalent_concentration.png\/660px-All_forcing_agents_CO2_equivalent_concentration.png 2x\" data-file-width=\"1423\" data-file-height=\"1032\" \/><\/a>  <div class=\"thumbcaption\">Future CO<sub>2<\/sub> projections, including all forcing agents' atmospheric CO<sub>2<\/sub>-equivalent concentrations (in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Parts_per_million\" class=\"mw-redirect\" title=\"Parts per million\">parts-per-million<\/a>-by-volume (ppmv)) according to four RCPs (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Representative_Concentration_Pathways\" class=\"mw-redirect\" title=\"Representative Concentration Pathways\">Representative Concentration Pathways<\/a>).<\/div><\/div><\/div>\n<p><b>Global warming<\/b> is a long-term rise in the average temperature of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Earth\" title=\"Earth\">Earth<\/a>'s <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate\" title=\"Climate\">climate system<\/a>, an aspect of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change\" title=\"Climate change\">climate change<\/a> shown by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Temperature_measurement\" title=\"Temperature measurement\">temperature measurements<\/a> and by multiple effects of the warming.<sup id=\"cite_ref-AR5_WG1_SPM_p4_2-0\" class=\"reference\"><a href=\"#cite_note-AR5_WG1_SPM_p4-2\">[2]<\/a><\/sup><sup id=\"cite_ref-EPA_myths_3-0\" class=\"reference\"><a href=\"#cite_note-EPA_myths-3\">[3]<\/a><\/sup> The term commonly refers to the mainly human-caused observed warming since pre-industrial times and its projected continuation,<sup id=\"cite_ref-AR5_Glossary_GW_4-0\" class=\"reference\"><a href=\"#cite_note-AR5_Glossary_GW-4\">[4]<\/a><\/sup> though there were also much earlier periods of global warming.<sup id=\"cite_ref-AR5_WG1_Ch_5_5-0\" class=\"reference\"><a href=\"#cite_note-AR5_WG1_Ch_5-5\">[5]<\/a><\/sup> In the modern context the terms are commonly used interchangeably,<sup id=\"cite_ref-Shaftel_2016_6-0\" class=\"reference\"><a href=\"#cite_note-Shaftel_2016-6\">[6]<\/a><\/sup> but <i>global warming<\/i> more specifically relates to worldwide surface temperature increases; while <i>climate change<\/i> is any regional or global statistically identifiable persistent change in the state of climate which lasts for decades or longer, including warming or cooling.<sup id=\"cite_ref-NOAA_diff_GW_&amp;_CC_7-0\" class=\"reference\"><a href=\"#cite_note-NOAA_diff_GW_&amp;_CC-7\">[7]<\/a><\/sup><sup id=\"cite_ref-AR5_Glossary_CC_8-0\" class=\"reference\"><a href=\"#cite_note-AR5_Glossary_CC-8\">[8]<\/a><\/sup> Many of the observed warming changes since the 1950s are unprecedented in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Instrumental_temperature_record\" title=\"Instrumental temperature record\">instrumental temperature record<\/a>, and in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Historical_climatology\" title=\"Historical climatology\">historical<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Paleoclimatology\" title=\"Paleoclimatology\">paleoclimate<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Proxy_(climate)\" title=\"Proxy (climate)\">proxy records<\/a> of climate change over thousands to millions of years.<sup id=\"cite_ref-AR5_WG1_SPM_p4_2-1\" class=\"reference\"><a href=\"#cite_note-AR5_WG1_SPM_p4-2\">[2]<\/a><\/sup><\/p><p>In 2013, the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">Intergovernmental Panel on Climate Change<\/a> (IPCC) <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Fifth_Assessment_Report\" title=\"IPCC Fifth Assessment Report\">Fifth Assessment Report<\/a> concluded, \"It is <em>extremely likely<\/em> that human influence has been the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Attribution_of_recent_climate_change\" title=\"Attribution of recent climate change\">dominant cause<\/a> of the observed warming since the mid-20th century.\"<sup id=\"cite_ref-9\" class=\"reference\"><a href=\"#cite_note-9\">[9]<\/a><\/sup> The largest human influence has been the emission of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas\" title=\"Greenhouse gas\">greenhouse gases<\/a> such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide\" title=\"Carbon dioxide\">carbon dioxide<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Methane\" title=\"Methane\">methane<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Nitrous_oxide\" title=\"Nitrous oxide\">nitrous oxide<\/a>. In view of the dominant role of human activity in causing it, the phenomenon is sometimes called \"<em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Human_impact_on_the_environment\" title=\"Human impact on the environment\">anthropogenic<\/a><\/em> global warming\" or \"<em>anthropogenic<\/em> climate change\". <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_model\" title=\"Climate model\">Climate model<\/a> projections summarized in the report indicated that during the 21st century, the global surface temperature is likely to rise a further 0.3 to 1.7\u00a0\u00b0C (0.5 to 3.1\u00a0\u00b0F) to  2.6 to 4.8\u00a0\u00b0C (4.7 to 8.6\u00a0\u00b0F) depending on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Representative_Concentration_Pathways\" class=\"mw-redirect\" title=\"Representative Concentration Pathways\">rate of greenhouse gas emissions<\/a>.<sup id=\"cite_ref-10\" class=\"reference\"><a href=\"#cite_note-10\">[10]<\/a><\/sup> These findings have been recognized by the national science academies of the major industrialized nations<sup id=\"cite_ref-11\" class=\"reference\"><a href=\"#cite_note-11\">[11]<\/a><\/sup><sup id=\"cite_ref-13\" class=\"reference\"><a href=\"#cite_note-13\">[a]<\/a><\/sup> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_opinion_on_climate_change\" title=\"Scientific opinion on climate change\">are not disputed by any scientific body of national or international standing<\/a>.<sup id=\"cite_ref-14\" class=\"reference\"><a href=\"#cite_note-14\">[13]<\/a><\/sup><sup id=\"cite_ref-15\" class=\"reference\"><a href=\"#cite_note-15\">[14]<\/a><\/sup><\/p><p>Future climate change and associated impacts will differ from <a href=\"https:\/\/en.wikipedia.org\/wiki\/Regional_effects_of_global_warming\" title=\"Regional effects of global warming\">region to region<\/a>.<sup id=\"cite_ref-16\" class=\"reference\"><a href=\"#cite_note-16\">[15]<\/a><\/sup><sup id=\"cite_ref-17\" class=\"reference\"><a href=\"#cite_note-17\">[16]<\/a><\/sup> Anticipated <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming\" title=\"Effects of global warming\">effects<\/a> include <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sea_level_rise\" title=\"Sea level rise\">rising sea levels<\/a>, changing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Precipitation\" title=\"Precipitation\">precipitation<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Desertification\" title=\"Desertification\">expansion of deserts<\/a> in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Subtropics\" title=\"Subtropics\">subtropics<\/a>.<sup id=\"cite_ref-Zeng_L17401_18-0\" class=\"reference\"><a href=\"#cite_note-Zeng_L17401-18\">[17]<\/a><\/sup> Warming is expected to be greater over land than over the oceans and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_sea_ice_decline\" title=\"Arctic sea ice decline\">greatest in the Arctic<\/a>, with the continuing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Retreat_of_glaciers_since_1850\" title=\"Retreat of glaciers since 1850\">retreat of glaciers<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Permafrost\" title=\"Permafrost\">permafrost<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sea_ice\" title=\"Sea ice\">sea ice<\/a>. Other likely changes include more frequent <a href=\"https:\/\/en.wikipedia.org\/wiki\/Extreme_weather\" title=\"Extreme weather\">extreme weather<\/a> events such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Heat_wave\" title=\"Heat wave\">heat waves<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Drought\" title=\"Drought\">droughts<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wildfires\" class=\"mw-redirect\" title=\"Wildfires\">wildfires<\/a>, heavy rainfall with floods, and heavy snowfall;<sup id=\"cite_ref-19\" class=\"reference\"><a href=\"#cite_note-19\">[18]<\/a><\/sup><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ocean_acidification\" title=\"Ocean acidification\">ocean acidification<\/a>; and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Extinction_risk_from_global_warming\" title=\"Extinction risk from global warming\">species extinctions<\/a> due to shifting temperature regimes. Effects significant to humans include the threat to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_agriculture\" title=\"Climate change and agriculture\">food security<\/a> from decreasing crop yields and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_migrant\" title=\"Environmental migrant\">abandonment of populated areas<\/a> due to rising sea levels.<sup id=\"cite_ref-20\" class=\"reference\"><a href=\"#cite_note-20\">[19]<\/a><\/sup><sup id=\"cite_ref-21\" class=\"reference\"><a href=\"#cite_note-21\">[20]<\/a><\/sup> Because the climate system has a large \"<a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_inertia\" title=\"Climate inertia\">inertia<\/a>\" and greenhouse gases will remain in the atmosphere for a long time, many of these effects will persist for not only decades or centuries, but tens of thousands of years.<sup id=\"cite_ref-22\" class=\"reference\"><a href=\"#cite_note-22\">[21]<\/a><\/sup><\/p><p>Possible societal responses to global warming include <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_mitigation\" title=\"Climate change mitigation\">mitigation<\/a> by emissions reduction, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_adaptation\" title=\"Climate change adaptation\">adaptation<\/a> to its effects, building systems <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_resilience\" title=\"Climate resilience\">resilient<\/a> to its effects, and possible future <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_engineering\" title=\"Climate engineering\">climate engineering<\/a>. Most countries are parties to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/United_Nations_Framework_Convention_on_Climate_Change\" title=\"United Nations Framework Convention on Climate Change\">United Nations Framework Convention on Climate Change<\/a> (UNFCCC),<sup id=\"cite_ref-23\" class=\"reference\"><a href=\"#cite_note-23\">[22]<\/a><\/sup> whose ultimate objective is to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Avoiding_Dangerous_Climate_Change\" class=\"mw-redirect\" title=\"Avoiding Dangerous Climate Change\">prevent dangerous anthropogenic climate change<\/a>.<sup id=\"cite_ref-24\" class=\"reference\"><a href=\"#cite_note-24\">[23]<\/a><\/sup> Parties to the UNFCCC have agreed that deep cuts in emissions are required<sup id=\"cite_ref-unfccc_2_degrees_celsius_target_25-0\" class=\"reference\"><a href=\"#cite_note-unfccc_2_degrees_celsius_target-25\">[24]<\/a><\/sup> and that global warming should be limited to well below 2.0\u00a0\u00b0C (3.6\u00a0\u00b0F) compared to pre-industrial levels,<sup id=\"cite_ref-27\" class=\"reference\"><a href=\"#cite_note-27\">[b]<\/a><\/sup> with efforts made to limit warming to 1.5\u00a0\u00b0C (2.7\u00a0\u00b0F).<sup id=\"cite_ref-28\" class=\"reference\"><a href=\"#cite_note-28\">[26]<\/a><\/sup> Some scientists call into question climate adaptation feasibility, with higher emissions scenarios,<sup id=\"cite_ref-Hansen_2013_29-0\" class=\"reference\"><a href=\"#cite_note-Hansen_2013-29\">[27]<\/a><\/sup> or the two degree temperature target.<sup id=\"cite_ref-Steffen2018_30-0\" class=\"reference\"><a href=\"#cite_note-Steffen2018-30\">[28]<\/a><\/sup><\/p><p>Public reactions to global warming and concern about its effects are also increasing. A global 2015 <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pew_Research_Center\" title=\"Pew Research Center\">Pew Research Center<\/a> report showed that a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Median\" title=\"Median\">median<\/a> of 54% of all respondents asked consider it \"a very serious problem\". Significant regional differences exist, with Americans and Chinese (whose economies are responsible for <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_countries_by_carbon_dioxide_emissions\" title=\"List of countries by carbon dioxide emissions\">the greatest annual CO<sub>2<\/sub> emissions<\/a>) among the least concerned.<sup id=\"cite_ref-31\" class=\"reference\"><a href=\"#cite_note-31\">[29]<\/a><\/sup><\/p>\n\n\n<h2>Observed temperature changes<\/h2>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Instrumental_temperature_record\" title=\"Instrumental temperature record\">Instrumental temperature record<\/a><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/6\/69\/Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png\/320px-Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png\" width=\"320\" height=\"198\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/6\/69\/Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png\/480px-Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/6\/69\/Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png\/640px-Annual_Mean_Temperature_Change_for_Land_and_for_Ocean_NASA_GISTEMP_2017_October.png 2x\" data-file-width=\"1130\" data-file-height=\"700\" \/><\/a>  <div class=\"thumbcaption\">Annual (thin lines) and five-year <a href=\"https:\/\/en.wikipedia.org\/wiki\/Local_regression\" title=\"Local regression\">lowess smooth<\/a> (thick lines) for the temperature anomalies averaged over the Earth\u2019s land area (red line) and sea surface temperature anomalies (blue line) averaged over the part of the ocean that is free of ice at all times (open ocean).<\/div><\/div><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:2000_Year_Temperature_Comparison.png\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c1\/2000_Year_Temperature_Comparison.png\/320px-2000_Year_Temperature_Comparison.png\" width=\"320\" height=\"240\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c1\/2000_Year_Temperature_Comparison.png\/480px-2000_Year_Temperature_Comparison.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c1\/2000_Year_Temperature_Comparison.png\/640px-2000_Year_Temperature_Comparison.png 2x\" data-file-width=\"1449\" data-file-height=\"1088\" \/><\/a>  <div class=\"thumbcaption\">Two millennia of mean surface temperatures according to different reconstructions from <a href=\"https:\/\/en.wikipedia.org\/wiki\/Proxy_(climate)\" title=\"Proxy (climate)\">climate proxies<\/a>, each smoothed on a decadal scale, with the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Instrumental_temperature_record\" title=\"Instrumental temperature record\">instrumental temperature record<\/a> overlaid in black.<\/div><\/div><\/div>\n<p>Multiple independently produced datasets confirm that from 1880 to 2012 the global average (land and ocean) surface temperature increased by 0.85 [0.65 to 1.06]\u00a0\u00b0C.<sup id=\"cite_ref-32\" class=\"reference\"><a href=\"#cite_note-32\">[30]<\/a><\/sup> From 1906 to 2005, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_average_surface_temperature\" class=\"mw-redirect\" title=\"Global average surface temperature\">Earth's average surface temperature<\/a> rose by 7002273890000000000\u26600.74\u00b10.18\u00a0\u00b0C. The rate of warming almost doubled in the last half of that period (7002273279999999999\u26600.13\u00b10.03\u00a0\u00b0C\u00a0per decade, against 7002273219999999999\u26600.07\u00b10.02\u00a0\u00b0C\u00a0per decade).<sup id=\"cite_ref-33\" class=\"reference\"><a href=\"#cite_note-33\">[31]<\/a><\/sup> Although the popular press often reports the increase of the average near-surface atmospheric temperature as <i>the<\/i> measure of global warming, most of the additional energy stored in the climate system since 1970 has accumulated in the oceans. The rest has melted ice and warmed the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Continent\" title=\"Continent\">continents<\/a> and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atmosphere\" title=\"Atmosphere\">atmosphere<\/a>.<sup id=\"cite_ref-34\" class=\"reference\"><a href=\"#cite_note-34\">[32]<\/a><\/sup><sup id=\"cite_ref-35\" class=\"reference\"><a href=\"#cite_note-35\">[c]<\/a><\/sup><\/p><p>Since 1979, the average temperature of the lower <a href=\"https:\/\/en.wikipedia.org\/wiki\/Troposphere\" title=\"Troposphere\">troposphere<\/a> has increased between 0.12 and 0.135\u00a0\u00b0C (0.216 and 0.243\u00a0\u00b0F)\u00a0per\u00a0decade, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Satellite_temperature_measurements\" title=\"Satellite temperature measurements\">satellite temperature measurements<\/a> confirm.<sup id=\"cite_ref-:1_36-0\" class=\"reference\"><a href=\"#cite_note-:1-36\">[33]<\/a><\/sup><sup id=\"cite_ref-:0_37-0\" class=\"reference\"><a href=\"#cite_note-:0-37\">[34]<\/a><\/sup><a href=\"https:\/\/en.wikipedia.org\/wiki\/Proxy_(climate)\" title=\"Proxy (climate)\">Climate proxies<\/a> show the temperature to have been relatively stable over the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Temperature_record_of_the_past_1000_years\" title=\"Temperature record of the past 1000 years\">one or two thousand years<\/a> before 1850, with regionally varying fluctuations such as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Medieval_Warm_Period\" title=\"Medieval Warm Period\">Medieval Warm Period<\/a> and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Little_Ice_Age\" title=\"Little Ice Age\">Little Ice Age<\/a>.<sup id=\"cite_ref-38\" class=\"reference\"><a href=\"#cite_note-38\">[35]<\/a><\/sup><\/p><p>The warming evident in the instrumental temperature record is consistent with a wide range of observations, as documented by many independent scientific groups.<sup id=\"cite_ref-kennedy_evidence_for_global_warming_39-0\" class=\"reference\"><a href=\"#cite_note-kennedy_evidence_for_global_warming-39\">[36]<\/a><\/sup> Examples include <a href=\"https:\/\/en.wikipedia.org\/wiki\/Current_sea_level_rise\" class=\"mw-redirect\" title=\"Current sea level rise\">sea level rise<\/a>,<sup id=\"cite_ref-40\" class=\"reference\"><a href=\"#cite_note-40\">[37]<\/a><\/sup> widespread melting of snow and land ice,<sup id=\"cite_ref-41\" class=\"reference\"><a href=\"#cite_note-41\">[38]<\/a><\/sup> increased <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ocean_heat_content\" title=\"Ocean heat content\">heat content of the oceans<\/a>,<sup id=\"cite_ref-kennedy_evidence_for_global_warming_39-1\" class=\"reference\"><a href=\"#cite_note-kennedy_evidence_for_global_warming-39\">[36]<\/a><\/sup> increased <a href=\"https:\/\/en.wikipedia.org\/wiki\/Specific_humidity\" class=\"mw-redirect\" title=\"Specific humidity\">humidity<\/a>,<sup id=\"cite_ref-kennedy_evidence_for_global_warming_39-2\" class=\"reference\"><a href=\"#cite_note-kennedy_evidence_for_global_warming-39\">[36]<\/a><\/sup> and the earlier <a href=\"https:\/\/en.wikipedia.org\/wiki\/Phenology\" title=\"Phenology\">timing<\/a> of spring events,<sup id=\"cite_ref-42\" class=\"reference\"><a href=\"#cite_note-42\">[39]<\/a><\/sup> e.g., the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Flowering\" class=\"mw-redirect\" title=\"Flowering\">flowering<\/a> of plants.<sup id=\"cite_ref-43\" class=\"reference\"><a href=\"#cite_note-43\">[40]<\/a><\/sup> The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Probability\" title=\"Probability\">probability<\/a> that these changes could have occurred by chance is virtually zero.<sup id=\"cite_ref-kennedy_evidence_for_global_warming_39-3\" class=\"reference\"><a href=\"#cite_note-kennedy_evidence_for_global_warming-39\">[36]<\/a><\/sup><\/p>\n<h3>Regional trends and short-term fluctuations<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">See also: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Regional_effects_of_global_warming\" title=\"Regional effects of global warming\">Regional effects of global warming<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cold_blob_(North_Atlantic)\" title=\"Cold blob (North Atlantic)\">Cold blob (North Atlantic)<\/a><\/div>\n<div class=\"thumb tleft\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:GISS_temperature_2000-09.png\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f3\/GISS_temperature_2000-09.png\/320px-GISS_temperature_2000-09.png\" width=\"320\" height=\"186\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f3\/GISS_temperature_2000-09.png\/480px-GISS_temperature_2000-09.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f3\/GISS_temperature_2000-09.png\/640px-GISS_temperature_2000-09.png 2x\" data-file-width=\"1078\" data-file-height=\"626\" \/><\/a>  <div class=\"thumbcaption\">Difference between average temperature in 2000\u20132009 compared to the 1951-1980 period, showing strong arctic amplification.<\/div><\/div><\/div>\n<p>Global warming refers to global averages. It is not uniform around the world: effects can vary by region.<sup id=\"cite_ref-WarmingHole-SE-USA_44-0\" class=\"reference\"><a href=\"#cite_note-WarmingHole-SE-USA-44\">[41]<\/a><\/sup> Since 1979, global average land temperatures have increased about twice as fast as global average ocean temperatures (7002273399999999999\u26600.25\u00a0\u00b0C\u00a0per decade against 7002273279999999999\u26600.13\u00a0\u00b0C\u00a0per decade).<sup id=\"cite_ref-45\" class=\"reference\"><a href=\"#cite_note-45\">[42]<\/a><\/sup> Ocean temperatures increase more slowly than land temperatures because of the larger <a href=\"https:\/\/en.wikipedia.org\/wiki\/Heat_capacity\" title=\"Heat capacity\">heat capacity<\/a> of the oceans and because oceans lose more heat by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Evaporation\" title=\"Evaporation\">evaporation<\/a>.<sup id=\"cite_ref-46\" class=\"reference\"><a href=\"#cite_note-46\">[43]<\/a><\/sup> Since the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Industrial_Revolution\" title=\"Industrial Revolution\">beginning of industrialisation<\/a> in the 18th century, the temperature difference between the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hemispheres_of_the_Earth\" class=\"mw-redirect\" title=\"Hemispheres of the Earth\">hemispheres<\/a> has increased due to melting of sea ice and snow in the North, and because there is more land in the Northern Hemisphere.<sup id=\"cite_ref-47\" class=\"reference\"><a href=\"#cite_note-47\">[44]<\/a><\/sup>\nIn the past 100 years, average Arctic temperatures have increased at almost twice the rate of the rest of the world.<sup id=\"cite_ref-48\" class=\"reference\"><a href=\"#cite_note-48\">[45]<\/a><\/sup> This has been referred to as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_amplification\" class=\"mw-redirect\" title=\"Arctic amplification\">Arctic amplification<\/a>.\n<\/p><p>Although more greenhouse gases are emitted in the Northern than in the Southern Hemisphere, this does not contribute to the difference in warming because the major greenhouse gases persist long enough to diffuse within and between the two hemispheres.<sup id=\"cite_ref-49\" class=\"reference\"><a href=\"#cite_note-49\">[46]<\/a><\/sup><\/p><p>There are different ways in which a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_forcing\" class=\"mw-redirect\" title=\"Climate forcing\">climate can be forced<\/a> to change, but because the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate\" title=\"Climate\">climate system<\/a> has large <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_inertia#Thermal_inertia\" title=\"Climate inertia\">thermal inertia<\/a>, it can take centuries\u00a0\u2013 or even longer\u00a0\u2013 for the climate to fully adjust. One <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_commitment\" title=\"Climate commitment\">climate commitment<\/a> study concluded that if greenhouse gases were stabilized at year 2000 levels, surface temperatures would still increase by about 0.5\u00a0\u00b0C,<sup id=\"cite_ref-50\" class=\"reference\"><a href=\"#cite_note-50\">[47]<\/a><\/sup> and another found that if they were stabilized at 2005 levels, surface warming could exceed a whole degree Celsius. Some of this surface warming would be driven by past natural forcings which have not yet reached <a href=\"https:\/\/en.wikipedia.org\/wiki\/Earth%27s_energy_budget#Climate_forcings_and_global_warming\" title=\"Earth's energy budget\">equilibrium in the climate system<\/a>. One study using a highly simplified climate model indicates these past natural forcings may account for as much as 64% of the committed 2050 surface warming, and their influence will fade with time compared to the human contribution.<sup id=\"cite_ref-51\" class=\"reference\"><a href=\"#cite_note-51\">[48]<\/a><\/sup><\/p><p>Global temperature is subject to short-term fluctuations that overlay long-term trends, and can temporarily mask or magnify them.<sup id=\"cite_ref-52\" class=\"reference\"><a href=\"#cite_note-52\">[49]<\/a><\/sup><sup id=\"cite_ref-53\" class=\"reference\"><a href=\"#cite_note-53\">[50]<\/a><\/sup> The relative stability in surface temperature from 2002 to 2009, which has since been dubbed the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_hiatus\" title=\"Global warming hiatus\">global warming hiatus<\/a> by the media and some scientists,<sup id=\"cite_ref-54\" class=\"reference\"><a href=\"#cite_note-54\">[51]<\/a><\/sup> may be an example of such an episode.<sup id=\"cite_ref-55\" class=\"reference\"><a href=\"#cite_note-55\">[52]<\/a><\/sup><sup id=\"cite_ref-56\" class=\"reference\"><a href=\"#cite_note-56\">[53]<\/a><\/sup> 2015 updates to account for differing methods of ocean surface temperature measurements show a positive trend over the recent decade.<sup id=\"cite_ref-57\" class=\"reference\"><a href=\"#cite_note-57\">[54]<\/a><\/sup><sup id=\"cite_ref-58\" class=\"reference\"><a href=\"#cite_note-58\">[55]<\/a><\/sup><\/p>\n<h3>Warmest years vs. overall trend<\/h3>\n<p>Sixteen of the seventeen warmest years on record have occurred since 2000.<sup id=\"cite_ref-59\" class=\"reference\"><a href=\"#cite_note-59\">[56]<\/a><\/sup> While record-breaking years attract considerable public interest, individual years are less significant than the overall trend. Some climatologists have criticized the attention that the popular press gives to \"warmest year\" statistics. In particular, ocean oscillations such as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/El_Ni\u00f1o\u2013Southern_Oscillation\" title=\"El Ni\u00f1o\u2013Southern Oscillation\">El Ni\u00f1o\u2013Southern Oscillation<\/a> (ENSO) can cause temperatures of a given year to be abnormally warm or cold for reasons unrelated to the overall trend of climate change. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gavin_Schmidt\" title=\"Gavin Schmidt\">Gavin Schmidt<\/a> stated: \"the long-term trends or the expected sequence of records are far more important than whether any single year is a record or not.\"<sup id=\"cite_ref-Gavin_Schmidt_201501_60-0\" class=\"reference\"><a href=\"#cite_note-Gavin_Schmidt_201501-60\">[57]<\/a><\/sup><\/p>\n<h2>Initial causes of temperature changes (external forcings)<\/h2>\n<div class=\"thumb tmulti tright\"><div class=\"thumbinner\"><div class=\"tsingle\"><div class=\"thumbimage\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Carbon_Dioxide_800kyr.svg\" class=\"image\"><img alt=\"refer to caption and adjacent text\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/11\/Carbon_Dioxide_800kyr.svg\/300px-Carbon_Dioxide_800kyr.svg.png\" width=\"300\" height=\"203\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/11\/Carbon_Dioxide_800kyr.svg\/450px-Carbon_Dioxide_800kyr.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/11\/Carbon_Dioxide_800kyr.svg\/600px-Carbon_Dioxide_800kyr.svg.png 2x\" data-file-width=\"509\" data-file-height=\"344\" \/><\/a><\/div><div class=\"thumbcaption\">CO<sub>2<\/sub> concentrations over the last 800,000 years<\/div><\/div><div class=\"tsingle\"><div class=\"thumbimage\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Greenhouse_Effect.svg\" class=\"image\"><img alt=\"refer to caption and adjacent text\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/58\/Greenhouse_Effect.svg\/300px-Greenhouse_Effect.svg.png\" width=\"300\" height=\"231\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/58\/Greenhouse_Effect.svg\/450px-Greenhouse_Effect.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/58\/Greenhouse_Effect.svg\/600px-Greenhouse_Effect.svg.png 2x\" data-file-width=\"750\" data-file-height=\"577\" \/><\/a><\/div><div class=\"thumbcaption\">Greenhouse effect schematic showing energy flows between space, the atmosphere, and Earth's surface. Energy exchanges are expressed in watts per square metre (W\/m<sup>2<\/sup>).<\/div><\/div><\/div><\/div>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Attribution_of_recent_climate_change\" title=\"Attribution of recent climate change\">Attribution of recent climate change<\/a><\/div>\n<p>By itself, the climate system may generate random changes in global temperatures for years to decades at a time, but long-term changes emanate only from so-called <i>external forcings.<\/i><sup id=\"cite_ref-61\" class=\"reference\"><a href=\"#cite_note-61\">[58]<\/a><\/sup><sup id=\"cite_ref-62\" class=\"reference\"><a href=\"#cite_note-62\">[59]<\/a><\/sup><sup id=\"cite_ref-63\" class=\"reference\"><a href=\"#cite_note-63\">[60]<\/a><\/sup> These forcings are \"external\" to the climate system, but not necessarily external to Earth.<sup id=\"cite_ref-64\" class=\"reference\"><a href=\"#cite_note-64\">[61]<\/a><\/sup> Examples of external forcings include changes in the composition of the atmosphere (e.g., increased concentrations of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas\" title=\"Greenhouse gas\">greenhouse gases<\/a>), <a href=\"https:\/\/en.wikipedia.org\/wiki\/Solar_luminosity\" title=\"Solar luminosity\">solar luminosity<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Volcano\" title=\"Volcano\">volcanic<\/a> eruptions, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Orbital_forcing\" title=\"Orbital forcing\">variations in Earth's orbit<\/a> around the Sun.<sup id=\"cite_ref-Hegerl-2001_65-0\" class=\"reference\"><a href=\"#cite_note-Hegerl-2001-65\">[62]<\/a><\/sup><\/p>\n<h3>Greenhouse gases<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main articles: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas\" title=\"Greenhouse gas\">Greenhouse gas<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_effect\" title=\"Greenhouse effect\">Greenhouse effect<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Radiative_forcing\" title=\"Radiative forcing\">Radiative forcing<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide_in_Earth%27s_atmosphere\" title=\"Carbon dioxide in Earth's atmosphere\">Carbon dioxide in Earth's atmosphere<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Earth%27s_energy_budget\" title=\"Earth's energy budget\">Earth's energy budget<\/a><\/div>\n<p>The greenhouse effect is the process by which <a href=\"https:\/\/en.wikipedia.org\/wiki\/Absorption_(electromagnetic_radiation)\" title=\"Absorption (electromagnetic radiation)\">absorption<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Emission_spectrum\" title=\"Emission spectrum\">emission<\/a> of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Infrared\" title=\"Infrared\">infrared<\/a> radiation by gases in a planet's <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atmosphere\" title=\"Atmosphere\">atmosphere<\/a> warm its lower atmosphere and surface. It was proposed by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Joseph_Fourier\" title=\"Joseph Fourier\">Joseph Fourier<\/a> in 1824, discovered in 1860 by <a href=\"https:\/\/en.wikipedia.org\/wiki\/John_Tyndall\" title=\"John Tyndall\">John Tyndall<\/a>,<sup id=\"cite_ref-66\" class=\"reference\"><a href=\"#cite_note-66\">[63]<\/a><\/sup> was first investigated quantitatively by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Svante_Arrhenius\" title=\"Svante Arrhenius\">Svante Arrhenius<\/a> in 1896,<sup id=\"cite_ref-:4_67-0\" class=\"reference\"><a href=\"#cite_note-:4-67\">[64]<\/a><\/sup> and the hypothesis was reported in the popular press as early as 1912.<sup id=\"cite_ref-68\" class=\"reference\"><a href=\"#cite_note-68\">[65]<\/a><\/sup><sup id=\"cite_ref-69\" class=\"reference\"><a href=\"#cite_note-69\">[66]<\/a><\/sup> The scientific description of global warming was further developed in the 1930s through 1960s by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Guy_Stewart_Callendar\" title=\"Guy Stewart Callendar\">Guy Stewart Callendar<\/a>.<sup id=\"cite_ref-70\" class=\"reference\"><a href=\"#cite_note-70\">[67]<\/a><\/sup><sup id=\"cite_ref-71\" class=\"reference\"><a href=\"#cite_note-71\">[68]<\/a><\/sup><\/p>\n<link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r856303569\" \/><div class=\"thumb tmulti tleft\"><div class=\"thumbinner\"><div class=\"tsingle\"><div class=\"thumbimage\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Annual_greenhouse_gas_emissions_by_sector,_in_2010_(color)_png.png\" class=\"image\"><img alt=\"refer to caption and image description\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/5c\/Annual_greenhouse_gas_emissions_by_sector%2C_in_2010_%28color%29_png.png\/220px-Annual_greenhouse_gas_emissions_by_sector%2C_in_2010_%28color%29_png.png\" width=\"220\" height=\"152\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/5c\/Annual_greenhouse_gas_emissions_by_sector%2C_in_2010_%28color%29_png.png\/330px-Annual_greenhouse_gas_emissions_by_sector%2C_in_2010_%28color%29_png.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/5c\/Annual_greenhouse_gas_emissions_by_sector%2C_in_2010_%28color%29_png.png\/440px-Annual_greenhouse_gas_emissions_by_sector%2C_in_2010_%28color%29_png.png 2x\" data-file-width=\"1200\" data-file-height=\"831\" \/><\/a><\/div><div class=\"thumbcaption\">Annual world greenhouse gas emissions, in 2010, by sector<\/div><\/div><div class=\"tsingle\"><div class=\"thumbimage\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg\" class=\"image\"><img alt=\"refer to caption and image description\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/0a\/Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg\/220px-Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg.png\" width=\"220\" height=\"143\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/0a\/Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg\/330px-Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/0a\/Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg\/440px-Percentage_share_of_global_cumulative_energy-related_carbon_dioxide_emissions_between_1751_and_2012_across_different_regions.svg.png 2x\" data-file-width=\"738\" data-file-height=\"481\" \/><\/a><\/div><div class=\"thumbcaption\">Percentage share of global cumulative energy-related CO<sub>2<\/sub> emissions between 1751 and 2012 across different regions<\/div><\/div><\/div><\/div>\n<p>On Earth, an atmosphere containing naturally occurring amounts of greenhouse gases causes air temperature near the surface to be warmer by about 33\u00a0\u00b0C (59\u00a0\u00b0F) than it would be in their absence.<sup id=\"cite_ref-72\" class=\"reference\"><a href=\"#cite_note-72\">[69]<\/a><\/sup><sup id=\"cite_ref-73\" class=\"reference\"><a href=\"#cite_note-73\">[d]<\/a><\/sup> Without the Earth's atmosphere, the Earth's average temperature would be well below the freezing temperature of water.<sup id=\"cite_ref-Blue_74-0\" class=\"reference\"><a href=\"#cite_note-Blue-74\">[70]<\/a><\/sup> The major greenhouse gases are <a href=\"https:\/\/en.wikipedia.org\/wiki\/Water_vapor\" title=\"Water vapor\">water vapour<\/a>, which causes about 36\u201370% of the greenhouse effect; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide\" title=\"Carbon dioxide\">carbon dioxide<\/a> (CO<sub>2<\/sub>), which causes 9\u201326%; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Methane\" title=\"Methane\">methane<\/a> (CH<sub>4<\/sub>), which causes 4\u20139%; and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ozone\" title=\"Ozone\">ozone<\/a> (O<sub>3<\/sub>), which causes 3\u20137%.<sup id=\"cite_ref-75\" class=\"reference\"><a href=\"#cite_note-75\">[71]<\/a><\/sup><sup id=\"cite_ref-76\" class=\"reference\"><a href=\"#cite_note-76\">[72]<\/a><\/sup><sup id=\"cite_ref-77\" class=\"reference\"><a href=\"#cite_note-77\">[73]<\/a><\/sup><\/p><p>Human activity since the Industrial Revolution has increased the amount of greenhouse gases in the atmosphere, leading to increased <a href=\"https:\/\/en.wikipedia.org\/wiki\/Radiative_forcing\" title=\"Radiative forcing\">radiative forcing<\/a> from CO<sub>2<\/sub>, methane, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Tropospheric_ozone\" title=\"Tropospheric ozone\">tropospheric ozone<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Chlorofluorocarbon\" title=\"Chlorofluorocarbon\">CFCs<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Nitrous_oxide\" title=\"Nitrous oxide\">nitrous oxide<\/a>. According to work published in 2007, the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas#Anthropogenic_greenhouse_gases\" title=\"Greenhouse gas\">concentrations<\/a> of CO<sub>2<\/sub> and methane had increased by 36% and 148% respectively since 1750.<sup id=\"cite_ref-78\" class=\"reference\"><a href=\"#cite_note-78\">[74]<\/a><\/sup> These levels are much higher than at any time during the last 800,000 years, the period for which reliable data has been extracted from <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ice_core\" title=\"Ice core\">ice cores<\/a>.<sup id=\"cite_ref-79\" class=\"reference\"><a href=\"#cite_note-79\">[75]<\/a><\/sup><sup id=\"cite_ref-80\" class=\"reference\"><a href=\"#cite_note-80\">[76]<\/a><\/sup><sup id=\"cite_ref-81\" class=\"reference\"><a href=\"#cite_note-81\">[77]<\/a><\/sup><sup id=\"cite_ref-82\" class=\"reference\"><a href=\"#cite_note-82\">[78]<\/a><\/sup> Less direct geological evidence indicates that CO<sub>2<\/sub> values higher than this were last seen about 20 million years ago.<sup id=\"cite_ref-83\" class=\"reference\"><a href=\"#cite_note-83\">[79]<\/a><\/sup><\/p><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fossil_fuel\" title=\"Fossil fuel\">Fossil fuel<\/a> burning has produced about three-quarters of the increase in CO<sub>2<\/sub> from human activity over the past 20 years. The rest of this increase is caused mostly by changes in land-use, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Deforestation_and_climate_change\" title=\"Deforestation and climate change\">particularly deforestation<\/a>.<sup id=\"cite_ref-84\" class=\"reference\"><a href=\"#cite_note-84\">[80]<\/a><\/sup> Another significant non-fuel source of anthropogenic CO<sub>2<\/sub> emissions is the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Calcination\" title=\"Calcination\">calcination<\/a> of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Limestone\" title=\"Limestone\">limestone<\/a> for <a href=\"https:\/\/en.wikipedia.org\/wiki\/Clinker_(cement)\" title=\"Clinker (cement)\">clinker<\/a> production, a chemical process which releases CO<sub>2<\/sub>.<sup id=\"cite_ref-85\" class=\"reference\"><a href=\"#cite_note-85\">[81]<\/a><\/sup> Estimates of global CO<sub>2<\/sub> emissions in 2011 from fossil fuel combustion, including cement production and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gas_flaring\" class=\"mw-redirect\" title=\"Gas flaring\">gas flaring<\/a>, was 34.8 billion tonnes (9.5 \u00b1 0.5 PgC), an increase of 54% above emissions in 1990. Coal burning was responsible for 43% of the total emissions, oil 34%, gas 18%, cement 4.9% and gas flaring 0.7%.<sup id=\"cite_ref-86\" class=\"reference\"><a href=\"#cite_note-86\">[82]<\/a><\/sup><\/p><p>In May 2013, it was reported that readings for CO<sub>2<\/sub> taken at the world's primary benchmark site in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mauna_Loa_Observatory\" title=\"Mauna Loa Observatory\">Mauna Loa<\/a> surpassed 400 <a href=\"https:\/\/en.wikipedia.org\/wiki\/Parts_per_million\" class=\"mw-redirect\" title=\"Parts per million\">ppm<\/a>. According to professor <a href=\"https:\/\/en.wikipedia.org\/wiki\/Brian_Hoskins\" title=\"Brian Hoskins\">Brian Hoskins<\/a>, this is likely the first time CO<sub>2<\/sub> levels have been this high for about 4.5 million years.<sup id=\"cite_ref-87\" class=\"reference\"><a href=\"#cite_note-87\">[83]<\/a><\/sup><sup id=\"cite_ref-88\" class=\"reference\"><a href=\"#cite_note-88\">[84]<\/a><\/sup> Monthly global CO<sub>2<\/sub> concentrations exceeded 400 ppm in March 2015, probably for the first time in several million years.<sup id=\"cite_ref-89\" class=\"reference\"><a href=\"#cite_note-89\">[85]<\/a><\/sup> On 12 November 2015, <a href=\"https:\/\/en.wikipedia.org\/wiki\/NASA\" title=\"NASA\">NASA scientists<\/a> reported that human-made carbon dioxide <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide_in_Earth%27s_atmosphere\" title=\"Carbon dioxide in Earth's atmosphere\">continues to increase<\/a> above levels not seen in hundreds of thousands of years; currently, about half of the carbon dioxide released from the burning of fossil fuels is not absorbed by vegetation and the oceans and remains in the atmosphere.<sup id=\"cite_ref-NASA-20151112-ab_90-0\" class=\"reference\"><a href=\"#cite_note-NASA-20151112-ab-90\">[86]<\/a><\/sup><\/p>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:CO2_emission_pie_chart.svg\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/ca\/CO2_emission_pie_chart.svg\/220px-CO2_emission_pie_chart.svg.png\" width=\"220\" height=\"220\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/ca\/CO2_emission_pie_chart.svg\/330px-CO2_emission_pie_chart.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/ca\/CO2_emission_pie_chart.svg\/440px-CO2_emission_pie_chart.svg.png 2x\" data-file-width=\"540\" data-file-height=\"540\" \/><\/a>  <div class=\"thumbcaption\">Global <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_countries_by_carbon_dioxide_emissions\" title=\"List of countries by carbon dioxide emissions\">carbon dioxide emissions<\/a> by country<\/div><\/div><\/div>\n<p>Over the last three decades of the twentieth century, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gross_domestic_product\" title=\"Gross domestic product\">gross domestic product<\/a> per capita and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Population_growth\" title=\"Population growth\">population growth<\/a> were the main drivers of increases in greenhouse gas emissions.<sup id=\"cite_ref-91\" class=\"reference\"><a href=\"#cite_note-91\">[87]<\/a><\/sup> CO<sub>2<\/sub> emissions are continuing to rise due to the burning of fossil fuels and land-use change.<sup id=\"cite_ref-nrc2008_92-0\" class=\"reference\"><a href=\"#cite_note-nrc2008-92\">[88]<\/a><\/sup><sup id=\"cite_ref-worldbank2010_93-0\" class=\"reference\"><a href=\"#cite_note-worldbank2010-93\">[89]<\/a><\/sup><sup class=\"reference\">:71<\/sup> Emissions can be <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas#Regional_and_national_attribution_of_emissions\" title=\"Greenhouse gas\">attributed to different regions<\/a>. Attributions of emissions due to land-use change are subject to considerable uncertainty.<sup id=\"cite_ref-94\" class=\"reference\"><a href=\"#cite_note-94\">[90]<\/a><\/sup><sup id=\"cite_ref-liverman_95-0\" class=\"reference\"><a href=\"#cite_note-liverman-95\">[91]<\/a><\/sup><sup class=\"reference\">:289<\/sup><\/p><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_scenario#Emissions_scenarios\" title=\"Climate change scenario\">Emissions scenarios<\/a>, estimates of changes in future emission levels of greenhouse gases, have been projected that depend upon uncertain economic, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sociology\" title=\"Sociology\">sociological<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Technology\" title=\"Technology\">technological<\/a>, and natural developments.<sup id=\"cite_ref-96\" class=\"reference\"><a href=\"#cite_note-96\">[92]<\/a><\/sup> In most scenarios, emissions continue to rise over the century, while in a few, emissions are reduced.<sup id=\"cite_ref-97\" class=\"reference\"><a href=\"#cite_note-97\">[93]<\/a><\/sup><sup id=\"cite_ref-98\" class=\"reference\"><a href=\"#cite_note-98\">[94]<\/a><\/sup> Fossil fuel reserves are abundant, and will not limit carbon emissions in the 21st century.<sup id=\"cite_ref-99\" class=\"reference\"><a href=\"#cite_note-99\">[95]<\/a><\/sup>\nEmission scenarios, combined with modelling of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_cycle\" title=\"Carbon cycle\">carbon cycle<\/a>, have been used to produce estimates of how atmospheric concentrations of greenhouse gases might change in the future. Using the six IPCC <a href=\"https:\/\/en.wikipedia.org\/wiki\/Special_Report_on_Emissions_Scenarios\" title=\"Special Report on Emissions Scenarios\">SRES<\/a> \"marker\" scenarios, models suggest that by the year 2100, the atmospheric concentration of CO<sub>2<\/sub> could range between 541 and 970 ppm.<sup id=\"cite_ref-100\" class=\"reference\"><a href=\"#cite_note-100\">[96]<\/a><\/sup><\/p>\n<h3>Aerosols and soot<\/h3>\n<div class=\"thumb tleft\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:ShipTracks_MODIS_2005may11.jpg\" class=\"image\"><img alt=\"Refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/ee\/ShipTracks_MODIS_2005may11.jpg\/220px-ShipTracks_MODIS_2005may11.jpg\" width=\"220\" height=\"192\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/ee\/ShipTracks_MODIS_2005may11.jpg\/330px-ShipTracks_MODIS_2005may11.jpg 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/ee\/ShipTracks_MODIS_2005may11.jpg\/440px-ShipTracks_MODIS_2005may11.jpg 2x\" data-file-width=\"4502\" data-file-height=\"3937\" \/><\/a>  <div class=\"thumbcaption\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ship_tracks\" title=\"Ship tracks\">Ship tracks<\/a> can be seen as lines in these clouds over the Atlantic Ocean on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/East_Coast_of_the_United_States\" title=\"East Coast of the United States\">East Coast of the United States<\/a>. Atmospheric particles from these and other sources could have a large effect on climate through the aerosol indirect effect.<\/div><\/div><\/div>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_dimming\" title=\"Global dimming\">Global dimming<\/a>, a gradual reduction in the amount of global direct <a href=\"https:\/\/en.wikipedia.org\/wiki\/Irradiance\" title=\"Irradiance\">irradiance<\/a> at the Earth's surface, was observed from 1961 until at least 1990.<sup id=\"cite_ref-101\" class=\"reference\"><a href=\"#cite_note-101\">[97]<\/a><\/sup><a href=\"https:\/\/en.wikipedia.org\/wiki\/Particulates#Climate_effects\" title=\"Particulates\">Solid and liquid particles<\/a> known as <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Aerosol\" title=\"Aerosol\">aerosols<\/a><\/i>, produced by volcanoes and human-made <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pollutant\" title=\"Pollutant\">pollutants<\/a>, are thought to be the main cause of this dimming. They exert a cooling effect by increasing the reflection of incoming sunlight. The effects of the products of fossil fuel combustion\u00a0\u2013 CO<sub>2<\/sub> and aerosols\u00a0\u2013 have partially offset one another in recent decades, so that net warming has been due to the increase in non-CO<sub>2<\/sub> greenhouse gases such as methane.<sup id=\"cite_ref-pmid10944197_102-0\" class=\"reference\"><a href=\"#cite_note-pmid10944197-102\">[98]<\/a><\/sup> Radiative forcing due to aerosols is temporally limited due to the processes that remove aerosols from the atmosphere. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wet_deposition\" class=\"mw-redirect\" title=\"Wet deposition\">Removal by clouds and precipitation<\/a> gives tropospheric aerosols an <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atmospheric_lifetime\" class=\"mw-redirect\" title=\"Atmospheric lifetime\">atmospheric lifetime<\/a> of only about a week, while <a href=\"https:\/\/en.wikipedia.org\/wiki\/Stratosphere\" title=\"Stratosphere\">stratospheric<\/a> aerosols can remain for a few years. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide\" title=\"Carbon dioxide\">Carbon dioxide<\/a> has a lifetime of a century or more, and as such, changes in aerosols will only delay climate changes due to carbon dioxide.<sup id=\"cite_ref-103\" class=\"reference\"><a href=\"#cite_note-103\">[99]<\/a><\/sup><a href=\"https:\/\/en.wikipedia.org\/wiki\/Black_carbon\" title=\"Black carbon\">Black carbon<\/a> is second only to carbon dioxide for its contribution to global warming (contribution being estimated at 17 to 20%, whereas carbon dioxide contributes 40 to 45% to global warming<sup id=\"cite_ref-104\" class=\"reference\"><a href=\"#cite_note-104\">[100]<\/a><\/sup><sup id=\"cite_ref-105\" class=\"reference\"><a href=\"#cite_note-105\">[101]<\/a><\/sup>).<sup id=\"cite_ref-106\" class=\"reference\"><a href=\"#cite_note-106\">[102]<\/a><\/sup><\/p><p>In addition to their direct effect by scattering and absorbing solar radiation, aerosols have indirect effects on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Earth%27s_energy_budget\" title=\"Earth's energy budget\">Earth's radiation budget<\/a>. Sulfate aerosols act as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cloud_condensation_nuclei\" title=\"Cloud condensation nuclei\">cloud condensation nuclei<\/a> and thus lead to clouds that have more and smaller cloud droplets. These clouds reflect solar radiation more efficiently than clouds with fewer and larger droplets, a phenomenon known as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Twomey_effect\" title=\"Twomey effect\">Twomey effect<\/a>.<sup id=\"cite_ref-107\" class=\"reference\"><a href=\"#cite_note-107\">[103]<\/a><\/sup> This effect also causes droplets to be of more uniform size, which reduces <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cloud_physics#Collision-coalescence\" title=\"Cloud physics\">growth of raindrops<\/a> and makes the cloud more reflective to incoming sunlight, known as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Albrecht_effect\" title=\"Albrecht effect\">Albrecht effect<\/a>.<sup id=\"cite_ref-108\" class=\"reference\"><a href=\"#cite_note-108\">[104]<\/a><\/sup> Indirect effects are most noticeable in marine stratiform clouds, and have very little radiative effect on convective clouds. Indirect effects of aerosols represent the largest uncertainty in radiative forcing.<sup id=\"cite_ref-109\" class=\"reference\"><a href=\"#cite_note-109\">[105]<\/a><\/sup><\/p><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Soot\" title=\"Soot\">Soot<\/a> may either cool or warm Earth's <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate\" title=\"Climate\">climate system<\/a>, depending on whether it is airborne or deposited. Atmospheric soot directly absorbs solar radiation, which heats the atmosphere and cools the surface. In isolated areas with high soot production, such as rural India, as much as 50% of surface warming due to greenhouse gases may be masked by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atmospheric_brown_cloud\" class=\"mw-redirect\" title=\"Atmospheric brown cloud\">atmospheric brown clouds<\/a>.<sup id=\"cite_ref-110\" class=\"reference\"><a href=\"#cite_note-110\">[106]<\/a><\/sup> When deposited, especially on glaciers or on ice in arctic regions, the lower surface <a href=\"https:\/\/en.wikipedia.org\/wiki\/Albedo\" title=\"Albedo\">albedo<\/a> can also directly heat the surface.<sup id=\"cite_ref-111\" class=\"reference\"><a href=\"#cite_note-111\">[107]<\/a><\/sup> The influences of atmospheric particles, including black carbon, are most pronounced in the tropics and sub-tropics, particularly in Asia, while the effects of greenhouse gases are dominant in the extratropics and southern hemisphere.<sup id=\"cite_ref-112\" class=\"reference\"><a href=\"#cite_note-112\">[108]<\/a><\/sup><\/p>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers,_1975-2013.png\" class=\"image\"><img alt=\"Refer to caption and adjacent text\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/b\/b4\/Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers%2C_1975-2013.png\/220px-Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers%2C_1975-2013.png\" width=\"220\" height=\"166\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/b\/b4\/Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers%2C_1975-2013.png\/330px-Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers%2C_1975-2013.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/b\/b4\/Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers%2C_1975-2013.png\/440px-Changes_in_total_solar_irradiance_and_monthly_sunspot_numbers%2C_1975-2013.png 2x\" data-file-width=\"2318\" data-file-height=\"1746\" \/><\/a>  <div class=\"thumbcaption\">Changes in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Total_solar_irradiance\" class=\"mw-redirect\" title=\"Total solar irradiance\">total solar irradiance<\/a> (TSI) and monthly <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sunspot\" title=\"Sunspot\">sunspot<\/a> numbers since the mid-1970s.<\/div><\/div><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Radiative-forcings.svg\" class=\"image\"><img alt=\"Refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/b\/bb\/Radiative-forcings.svg\/220px-Radiative-forcings.svg.png\" width=\"220\" height=\"176\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/b\/bb\/Radiative-forcings.svg\/330px-Radiative-forcings.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/b\/bb\/Radiative-forcings.svg\/440px-Radiative-forcings.svg.png 2x\" data-file-width=\"600\" data-file-height=\"480\" \/><\/a>  <div class=\"thumbcaption\">Contribution of natural factors and human activities to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Radiative_forcing\" title=\"Radiative forcing\">radiative forcing<\/a> of climate change.<sup id=\"cite_ref-ipcc_radiative_forcing_113-0\" class=\"reference\"><a href=\"#cite_note-ipcc_radiative_forcing-113\">[109]<\/a><\/sup> Radiative forcing values are for the year 2005, relative to the pre-industrial era (1750).<sup id=\"cite_ref-ipcc_radiative_forcing_113-1\" class=\"reference\"><a href=\"#cite_note-ipcc_radiative_forcing-113\">[109]<\/a><\/sup> The contribution of solar irradiance to radiative forcing is 5% of the value of the combined radiative forcing due to increases in the atmospheric concentrations of carbon dioxide, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atmospheric_methane\" title=\"Atmospheric methane\">methane<\/a> and nitrous oxide.<sup id=\"cite_ref-epa_solar_activity_114-0\" class=\"reference\"><a href=\"#cite_note-epa_solar_activity-114\">[110]<\/a><\/sup><\/div><\/div><\/div>\n<h3>Solar activity<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Solar_activity_and_climate\" title=\"Solar activity and climate\">Solar activity and climate<\/a><\/div>\n<p>Since 1978, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Solar_variation\" class=\"mw-redirect\" title=\"Solar variation\">solar irradiance<\/a> has been measured by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Satellite\" title=\"Satellite\">satellites<\/a>.<sup id=\"cite_ref-115\" class=\"reference\"><a href=\"#cite_note-115\">[111]<\/a><\/sup><\/p><p>Climate models have been used to examine the role of the Sun in recent climate change.<sup id=\"cite_ref-116\" class=\"reference\"><a href=\"#cite_note-116\">[112]<\/a><\/sup> Models are unable to reproduce the rapid warming observed in recent decades when only taking into account variations in solar output and volcanic activity.<sup id=\"cite_ref-Schmidt_et_al_2014_117-0\" class=\"reference\"><a href=\"#cite_note-Schmidt_et_al_2014-117\">[113]<\/a><\/sup><sup id=\"cite_ref-Fyfe_et_al_2016_118-0\" class=\"reference\"><a href=\"#cite_note-Fyfe_et_al_2016-118\">[114]<\/a><\/sup><\/p><p>Another line of evidence for the Sun's non-attributability is the differing temperature changes at different levels in the Earth's atmosphere.<sup id=\"cite_ref-119\" class=\"reference\"><a href=\"#cite_note-119\">[115]<\/a><\/sup> According to basic physical principles, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_effect\" title=\"Greenhouse effect\">the greenhouse effect<\/a> produces warming of the lower atmosphere (the troposphere), but cooling of the upper atmosphere (the stratosphere).<sup id=\"cite_ref-120\" class=\"reference\"><a href=\"#cite_note-120\">[116]<\/a><\/sup><sup id=\"cite_ref-Randel2009_121-0\" class=\"reference\"><a href=\"#cite_note-Randel2009-121\">[117]<\/a><\/sup> If solar variations were responsible for the observed warming, warming of both the troposphere and the stratosphere would be expected.<sup id=\"cite_ref-122\" class=\"reference\"><a href=\"#cite_note-122\">[118]<\/a><\/sup><\/p>\n<h3>Variations in Earth's orbit<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Milankovitch_cycles\" title=\"Milankovitch cycles\">Milankovitch cycles<\/a><\/div>\n<p>The tilt of the Earth\u2019s axis and the shape of its orbit around the Sun vary slowly over tens of thousands of years. This changes climate by changing the seasonal and latitudinal distribution of incoming <a href=\"https:\/\/en.wikipedia.org\/wiki\/Insolation\" class=\"mw-redirect\" title=\"Insolation\">solar energy<\/a> at Earth's surface.<sup id=\"cite_ref-123\" class=\"reference\"><a href=\"#cite_note-123\">[119]<\/a><\/sup>\nDuring the last few thousand years, this phenomenon contributed to a slow cooling trend at high latitudes of the Northern Hemisphere during summer, a trend that was reversed by greenhouse-gas-induced warming during the 20th century.<sup id=\"cite_ref-124\" class=\"reference\"><a href=\"#cite_note-124\">[120]<\/a><\/sup><sup id=\"cite_ref-125\" class=\"reference\"><a href=\"#cite_note-125\">[121]<\/a><\/sup><sup id=\"cite_ref-126\" class=\"reference\"><a href=\"#cite_note-126\">[122]<\/a><\/sup><sup id=\"cite_ref-127\" class=\"reference\"><a href=\"#cite_note-127\">[123]<\/a><\/sup> Orbital cycles favorable for glaciation are not expected within the next 50,000 years.<sup id=\"cite_ref-128\" class=\"reference\"><a href=\"#cite_note-128\">[124]<\/a><\/sup><sup id=\"cite_ref-129\" class=\"reference\"><a href=\"#cite_note-129\">[125]<\/a><\/sup><\/p>\n<h2>Climate change feedback<\/h2>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main articles: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_feedback\" title=\"Climate change feedback\">Climate change feedback<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_sensitivity\" title=\"Climate sensitivity\">Climate sensitivity<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_amplification\" class=\"mw-redirect\" title=\"Arctic amplification\">Arctic amplification<\/a><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:NORTH_POLE_Ice_(19626661335).jpg\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/6\/64\/NORTH_POLE_Ice_%2819626661335%29.jpg\/220px-NORTH_POLE_Ice_%2819626661335%29.jpg\" width=\"220\" height=\"147\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/6\/64\/NORTH_POLE_Ice_%2819626661335%29.jpg\/330px-NORTH_POLE_Ice_%2819626661335%29.jpg 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/6\/64\/NORTH_POLE_Ice_%2819626661335%29.jpg\/440px-NORTH_POLE_Ice_%2819626661335%29.jpg 2x\" data-file-width=\"7360\" data-file-height=\"4912\" \/><\/a>  <div class=\"thumbcaption\">The dark ocean surface reflects only 6 percent of incoming solar radiation, whereas sea ice reflects 50 to 70 percent.<sup id=\"cite_ref-130\" class=\"reference\"><a href=\"#cite_note-130\">[126]<\/a><\/sup><\/div><\/div><\/div>\n<p>A range of <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Feedback\" title=\"Feedback\">feedbacks<\/a><\/i> govern the response of the system to changes in forcings. Positive feedbacks increase the response of the climate system to an initial forcing, while negative feedbacks reduce it.<sup id=\"cite_ref-jackson_feedbacks_131-0\" class=\"reference\"><a href=\"#cite_note-jackson_feedbacks-131\">[127]<\/a><\/sup><\/p><p>There are a range of feedbacks in the climate system, including <a href=\"https:\/\/en.wikipedia.org\/wiki\/Water_vapor_feedback\" class=\"mw-redirect\" title=\"Water vapor feedback\">water vapour<\/a>, changes in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ice\u2013albedo_feedback\" title=\"Ice\u2013albedo feedback\">ice\u2013albedo<\/a> (snow and ice cover affect how much the Earth's surface absorbs or reflects incoming sunlight), clouds, and changes in the Earth's carbon cycle (e.g., the release of carbon from soil).<sup id=\"cite_ref-132\" class=\"reference\"><a href=\"#cite_note-132\">[128]<\/a><\/sup> The main negative feedback is via <a href=\"https:\/\/en.wikipedia.org\/wiki\/Radiative_cooling\" title=\"Radiative cooling\">radiative cooling<\/a> to space as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Infrared_radiation\" class=\"mw-redirect\" title=\"Infrared radiation\">infrared radiation<\/a>,<sup id=\"cite_ref-133\" class=\"reference\"><a href=\"#cite_note-133\">[129]<\/a><\/sup> which increases strongly with increasing temperature<sup id=\"cite_ref-134\" class=\"reference\"><a href=\"#cite_note-134\">[130]<\/a><\/sup><\/p><p>Feedbacks are an important factor in determining the sensitivity of the climate system to increased atmospheric greenhouse gas concentrations. Other factors being equal, a higher <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_sensitivity\" title=\"Climate sensitivity\">climate sensitivity<\/a><\/i> means that more warming will occur for a given increase in greenhouse gas forcing.<sup id=\"cite_ref-135\" class=\"reference\"><a href=\"#cite_note-135\">[131]<\/a><\/sup> Uncertainty over the effect of feedbacks is a major reason why different climate models project different magnitudes of warming for a given forcing scenario. More research is needed to understand the role of clouds<sup id=\"cite_ref-jackson_feedbacks_131-1\" class=\"reference\"><a href=\"#cite_note-jackson_feedbacks-131\">[127]<\/a><\/sup> and carbon cycle feedbacks in climate projections.<sup id=\"cite_ref-136\" class=\"reference\"><a href=\"#cite_note-136\">[132]<\/a><\/sup><\/p><p>An observation based study on future climate change, on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Soil_carbon_feedback\" title=\"Soil carbon feedback\">soil carbon feedback<\/a>, conducted since 1991 in Harvard, suggests the pending release of about 190 petagrams of soil carbon, the equivalent of the past two decades of greenhouse gas emissions from fossil fuel burning, up through the year 2100 from the top 1-meter of Earth's soils, due to changes in microbial communities under elevated temperatures.<sup id=\"cite_ref-137\" class=\"reference\"><a href=\"#cite_note-137\">[133]<\/a><\/sup><sup id=\"cite_ref-138\" class=\"reference\"><a href=\"#cite_note-138\">[134]<\/a><\/sup> Climate models do not account for this possible feedback mechanism. Another study conducted by Harvard researchers suggests that increased water vapor injected into the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Stratosphere\" title=\"Stratosphere\">stratosphere<\/a>, due to rising temperatures, increases <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ozone_depletion\" title=\"Ozone depletion\">ozone depletion<\/a>, subsequently raising the odds of skin cancer and damaging crops.<sup id=\"cite_ref-139\" class=\"reference\"><a href=\"#cite_note-139\">[135]<\/a><\/sup><\/p><p>Arctic temperatures have increased at almost twice the rate of the rest of the world.<sup id=\"cite_ref-140\" class=\"reference\"><a href=\"#cite_note-140\">[136]<\/a><\/sup> This appears to arise both from an intensification of poleward heat transport and more directly from changes in the local net radiation balance.<sup id=\"cite_ref-Alexeev_141-0\" class=\"reference\"><a href=\"#cite_note-Alexeev-141\">[137]<\/a><\/sup> Some examples of feedbacks thought to contribute include the reduction of snow cover and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_sea_ice_decline\" title=\"Arctic sea ice decline\">sea ice<\/a>, changes in atmospheric and ocean circulation, the presence of anthropogenic soot in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic\" title=\"Arctic\">Arctic<\/a> environment, increases in cloud cover, and water vapor. According to the 2013 IPCC report, models often tend to underestimate Arctic amplification.<sup id=\"cite_ref-IPCC_2013_142-0\" class=\"reference\"><a href=\"#cite_note-IPCC_2013-142\">[138]<\/a><\/sup><sup id=\"cite_ref-143\" class=\"reference\"><a href=\"#cite_note-143\">[139]<\/a><\/sup> Studies have linked the rapidly warming Arctic to a vanishing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cryosphere\" title=\"Cryosphere\">cryosphere<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Extreme_weather\" title=\"Extreme weather\">extreme weather<\/a> in mid-latitudes.<sup id=\"cite_ref-144\" class=\"reference\"><a href=\"#cite_note-144\">[140]<\/a><\/sup><sup id=\"cite_ref-145\" class=\"reference\"><a href=\"#cite_note-145\">[141]<\/a><\/sup> and changes in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Jet_stream\" title=\"Jet stream\">jet stream<\/a>.<sup id=\"cite_ref-146\" class=\"reference\"><a href=\"#cite_note-146\">[142]<\/a><\/sup><\/p>\n<h2>Climate models<\/h2>\n<link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r856303569\" \/><div class=\"thumb tmulti tright\"><div class=\"thumbinner\"><div class=\"tsingle\"><div class=\"thumbimage\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Global_Warming_Predictions.png\" class=\"image\"><img alt=\"refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/a\/aa\/Global_Warming_Predictions.png\/300px-Global_Warming_Predictions.png\" width=\"300\" height=\"214\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/a\/aa\/Global_Warming_Predictions.png\/450px-Global_Warming_Predictions.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/a\/aa\/Global_Warming_Predictions.png 2x\" data-file-width=\"528\" data-file-height=\"377\" \/><\/a><\/div><div class=\"thumbcaption\">Calculations of global warming prepared in or before 2001 from a range of climate models under the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Special_Report_on_Emissions_Scenarios\" title=\"Special Report on Emissions Scenarios\">SRES<\/a> A2 emissions scenario, which assumes no action is taken to reduce emissions and regionally divided economic development.<\/div><\/div><div class=\"tsingle\"><div class=\"thumbimage\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century,_based_on_SRES_emissions_scenario_A1B.png\" class=\"image\"><img alt=\"refer to caption and image description\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e6\/Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century%2C_based_on_SRES_emissions_scenario_A1B.png\/300px-Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century%2C_based_on_SRES_emissions_scenario_A1B.png\" width=\"300\" height=\"161\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e6\/Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century%2C_based_on_SRES_emissions_scenario_A1B.png\/450px-Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century%2C_based_on_SRES_emissions_scenario_A1B.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e6\/Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century%2C_based_on_SRES_emissions_scenario_A1B.png\/600px-Projected_change_in_annual_mean_surface_air_temperature_from_the_late_20th_century_to_the_middle_21st_century%2C_based_on_SRES_emissions_scenario_A1B.png 2x\" data-file-width=\"2811\" data-file-height=\"1507\" \/><\/a><\/div><div class=\"thumbcaption\">Projected change in annual mean surface air temperature from the late 20th century to the middle 21st century, based on a medium <a href=\"https:\/\/en.wikipedia.org\/wiki\/Economics_of_global_warming#Emissions_scenarios\" title=\"Economics of global warming\">emissions scenario<\/a> (SRES A1B).<sup id=\"cite_ref-147\" class=\"reference\"><a href=\"#cite_note-147\">[143]<\/a><\/sup> This scenario assumes that no future policies are adopted to limit greenhouse gas emissions. Image credit: <a href=\"https:\/\/en.wikipedia.org\/wiki\/NOAA\" class=\"mw-redirect\" title=\"NOAA\">NOAA<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Geophysical_Fluid_Dynamics_Laboratory\" title=\"Geophysical Fluid Dynamics Laboratory\">GFDL<\/a>.<sup id=\"cite_ref-148\" class=\"reference\"><a href=\"#cite_note-148\">[144]<\/a><\/sup><\/div><\/div><\/div><\/div>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_climate_model\" class=\"mw-redirect\" title=\"Global climate model\">Global climate model<\/a><\/div>\n<p>A <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_model\" title=\"Climate model\">climate model<\/a> is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mathematical_models\" class=\"mw-redirect\" title=\"Mathematical models\">representation<\/a> of the physical, chemical and biological processes that affect the climate system.<sup id=\"cite_ref-149\" class=\"reference\"><a href=\"#cite_note-149\">[145]<\/a><\/sup><\/p><p>Although researchers attempt to include as many processes as possible, simplifications of the actual climate system are inevitable because of the constraints of available computer power and limitations in knowledge of the climate system. Results from models can also vary due to different greenhouse gas inputs and the model's climate sensitivity. For example, the uncertainty in IPCC's 2007 projections is caused by (1) the use of multiple models<sup id=\"cite_ref-meehl_2007_global_warming_projections_150-0\" class=\"reference\"><a href=\"#cite_note-meehl_2007_global_warming_projections-150\">[146]<\/a><\/sup> with differing sensitivity to greenhouse gas concentrations,<sup id=\"cite_ref-151\" class=\"reference\"><a href=\"#cite_note-151\">[147]<\/a><\/sup> (2) the use of differing estimates of humanity's future greenhouse gas emissions,<sup id=\"cite_ref-meehl_2007_global_warming_projections_150-1\" class=\"reference\"><a href=\"#cite_note-meehl_2007_global_warming_projections-150\">[146]<\/a><\/sup> (3) any additional emissions from climate feedbacks that were not included in the models IPCC used to prepare its report, i.e., greenhouse gas releases from permafrost.<sup id=\"cite_ref-NSDIC_permafrost_study_152-0\" class=\"reference\"><a href=\"#cite_note-NSDIC_permafrost_study-152\">[148]<\/a><\/sup><\/p><p>The models do not assume the climate will warm due to increasing levels of greenhouse gases. Instead the models predict how greenhouse gases will interact with radiative transfer and other physical processes. Warming or cooling is thus a result, not an assumption, of the models.<sup id=\"cite_ref-153\" class=\"reference\"><a href=\"#cite_note-153\">[149]<\/a><\/sup><\/p><p>Clouds and their effects are especially difficult to predict. Improving the models' representation of clouds is therefore an important topic in current research.<sup id=\"cite_ref-154\" class=\"reference\"><a href=\"#cite_note-154\">[150]<\/a><\/sup> Another prominent research topic is expanding and improving representations of the carbon cycle.<sup id=\"cite_ref-155\" class=\"reference\"><a href=\"#cite_note-155\">[151]<\/a><\/sup><sup id=\"cite_ref-156\" class=\"reference\"><a href=\"#cite_note-156\">[152]<\/a><\/sup><sup id=\"cite_ref-157\" class=\"reference\"><a href=\"#cite_note-157\">[153]<\/a><\/sup><\/p><p>Models are also used to help investigate the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Attribution_of_recent_climate_change\" title=\"Attribution of recent climate change\">causes of recent climate change<\/a> by comparing the observed changes to those that the models project from various natural and human causes. Although these models do not unambiguously attribute the warming that occurred from approximately 1910 to 1945 to either natural variation or human effects, they do indicate that the warming since 1970 is dominated by anthropogenic greenhouse gas emissions.<sup id=\"cite_ref-Hegerl-2001_65-1\" class=\"reference\"><a href=\"#cite_note-Hegerl-2001-65\">[62]<\/a><\/sup><\/p><p>The physical realism of models is tested by examining their ability to simulate contemporary or past climates.<sup id=\"cite_ref-158\" class=\"reference\"><a href=\"#cite_note-158\">[154]<\/a><\/sup> Climate models produce a good match to observations of global temperature changes over the last century, but do not simulate all aspects of climate.<sup id=\"cite_ref-159\" class=\"reference\"><a href=\"#cite_note-159\">[155]<\/a><\/sup> Not all <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming\" title=\"Effects of global warming\">effects of global warming<\/a> are accurately predicted by the climate models used by the IPCC. Observed <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_shrinkage\" class=\"mw-redirect\" title=\"Arctic shrinkage\">Arctic shrinkage<\/a> has been faster than that predicted.<sup id=\"cite_ref-160\" class=\"reference\"><a href=\"#cite_note-160\">[156]<\/a><\/sup> Precipitation increased proportionally to atmospheric humidity, and hence significantly faster than global climate models predict.<sup id=\"cite_ref-161\" class=\"reference\"><a href=\"#cite_note-161\">[157]<\/a><\/sup><sup id=\"cite_ref-162\" class=\"reference\"><a href=\"#cite_note-162\">[158]<\/a><\/sup> Since 1990, sea level has also risen considerably faster than models predicted it would.<sup id=\"cite_ref-163\" class=\"reference\"><a href=\"#cite_note-163\">[159]<\/a><\/sup> The 2017 United States published <a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Climate_Assessment\" title=\"National Climate Assessment\">National Climate Assessment<\/a>, notes that \"climate models may still be underestimating or missing relevant feedback processes\".<sup id=\"cite_ref-164\" class=\"reference\"><a href=\"#cite_note-164\">[160]<\/a><\/sup><\/p>\n<h2>Effects<\/h2>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming\" title=\"Effects of global warming\">Effects of global warming<\/a><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Projections_of_global_mean_sea_level_rise_by_Parris_et_al._(2012).png\" class=\"image\"><img alt=\"Refer to caption and adjacent text\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/54\/Projections_of_global_mean_sea_level_rise_by_Parris_et_al._%282012%29.png\/350px-Projections_of_global_mean_sea_level_rise_by_Parris_et_al._%282012%29.png\" width=\"350\" height=\"197\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/54\/Projections_of_global_mean_sea_level_rise_by_Parris_et_al._%282012%29.png\/525px-Projections_of_global_mean_sea_level_rise_by_Parris_et_al._%282012%29.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/54\/Projections_of_global_mean_sea_level_rise_by_Parris_et_al._%282012%29.png\/700px-Projections_of_global_mean_sea_level_rise_by_Parris_et_al._%282012%29.png 2x\" data-file-width=\"1355\" data-file-height=\"761\" \/><\/a>  <div class=\"thumbcaption\">Projections of global mean sea level rise by Parris and others.<sup id=\"cite_ref-parris_sea_level_rise_165-0\" class=\"reference\"><a href=\"#cite_note-parris_sea_level_rise-165\">[161]<\/a><\/sup> Probabilities have not been assigned to these projections.<sup id=\"cite_ref-166\" class=\"reference\"><a href=\"#cite_note-166\">[162]<\/a><\/sup> Therefore, none of these projections should be interpreted as a \"best estimate\" of future sea level rise. Image credit: NOAA.<\/div><\/div><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:6m_Sea_Level_Rise.jpg\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/5c\/6m_Sea_Level_Rise.jpg\/250px-6m_Sea_Level_Rise.jpg\" width=\"250\" height=\"125\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/5c\/6m_Sea_Level_Rise.jpg\/375px-6m_Sea_Level_Rise.jpg 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/5\/5c\/6m_Sea_Level_Rise.jpg\/500px-6m_Sea_Level_Rise.jpg 2x\" data-file-width=\"2048\" data-file-height=\"1024\" \/><\/a>  <div class=\"thumbcaption\">Map of the Earth with a six-meter sea level rise represented in red.<\/div><\/div><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Glacier_Mass_Balance.png\" class=\"image\"><img alt=\"refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e5\/Glacier_Mass_Balance.png\/250px-Glacier_Mass_Balance.png\" width=\"250\" height=\"190\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e5\/Glacier_Mass_Balance.png\/375px-Glacier_Mass_Balance.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e5\/Glacier_Mass_Balance.png\/500px-Glacier_Mass_Balance.png 2x\" data-file-width=\"658\" data-file-height=\"500\" \/><\/a>  <div class=\"thumbcaption\">Sparse records indicate that glaciers have been retreating since the early 1800s. In the 1950s measurements began that allow the monitoring of glacial mass balance, reported to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/World_Glacier_Monitoring_Service\" title=\"World Glacier Monitoring Service\">World Glacier Monitoring Service<\/a> (WGMS) and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Snow_and_Ice_Data_Center\" title=\"National Snow and Ice Data Center\">National Snow and Ice Data Center<\/a> (NSIDC).<\/div><\/div><\/div>\n<h3>Biosphere<\/h3>\n<p>Overall, it is expected that climate change will result in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Extinction\" title=\"Extinction\">extinction<\/a> of many species and reduced diversity of ecosystems.<sup id=\"cite_ref-167\" class=\"reference\"><a href=\"#cite_note-167\">[163]<\/a><\/sup> Rising temperatures have been found to push bees to their physiological limits, and could cause the extinction of bee populations.<sup id=\"cite_ref-168\" class=\"reference\"><a href=\"#cite_note-168\">[164]<\/a><\/sup> A 2012 study concluded that continued ocean uptake of CO<sub>2<\/sub> affects the brains and central nervous system of certain fish species and this impacts their ability to hear, smell, and evade predators. The study authors note, \"We've now established it isn't simply the acidification of the oceans that is causing disruption\u00a0\u2013 as is the case with shellfish and plankton with chalky skeletons\u00a0\u2013 but the actual dissolved CO<sub>2<\/sub> itself is damaging the fishes' nervous systems.\"<sup id=\"cite_ref-169\" class=\"reference\"><a href=\"#cite_note-169\">[165]<\/a><\/sup><\/p>\n<h3>Environmental<\/h3>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Endangered_arctic_-_starving_polar_bear.jpg\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1a\/Endangered_arctic_-_starving_polar_bear.jpg\/220px-Endangered_arctic_-_starving_polar_bear.jpg\" width=\"220\" height=\"124\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1a\/Endangered_arctic_-_starving_polar_bear.jpg\/330px-Endangered_arctic_-_starving_polar_bear.jpg 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1a\/Endangered_arctic_-_starving_polar_bear.jpg\/440px-Endangered_arctic_-_starving_polar_bear.jpg 2x\" data-file-width=\"3009\" data-file-height=\"1693\" \/><\/a>  <div class=\"thumbcaption\">As the climate change melts sea ice, the U.S. Geological Survey projects that two-thirds of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Polar_bear\" title=\"Polar bear\">polar bears<\/a> will disappear by 2050.<sup id=\"cite_ref-170\" class=\"reference\"><a href=\"#cite_note-170\">[166]<\/a><\/sup><sup id=\"cite_ref-171\" class=\"reference\"><a href=\"#cite_note-171\">[167]<\/a><\/sup><\/div><\/div><\/div>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main articles: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Physical_impacts_of_climate_change\" title=\"Physical impacts of climate change\">Physical impacts of climate change<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_ecosystems\" title=\"Climate change and ecosystems\">Climate change and ecosystems<\/a><\/div>\n<p>The environmental effects of global warming are broad and far reaching. They include the following diverse effects:\n<\/p>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_sea_ice_decline\" title=\"Arctic sea ice decline\">Arctic sea ice decline<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sea_level_rise\" title=\"Sea level rise\">sea level rise<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Retreat_of_glaciers_since_1850\" title=\"Retreat of glaciers since 1850\">retreat of glaciers<\/a>: Global warming has led to decades of shrinking and thinning in a warm climate that has put the Arctic sea ice in a precarious position, it is now vulnerable to atmospheric anomalies.<sup id=\"cite_ref-172\" class=\"reference\"><a href=\"#cite_note-172\">[168]<\/a><\/sup> Projections of declines in Arctic sea ice vary.<sup id=\"cite_ref-173\" class=\"reference\"><a href=\"#cite_note-173\">[169]<\/a><\/sup><sup id=\"cite_ref-174\" class=\"reference\"><a href=\"#cite_note-174\">[170]<\/a><\/sup> Recent projections suggest that Arctic summers could be ice-free (defined as ice extent less than 1 million <a href=\"https:\/\/en.wikipedia.org\/wiki\/Square_km\" class=\"mw-redirect\" title=\"Square km\">square km<\/a>) as early as 2025\u20132030.<sup id=\"cite_ref-175\" class=\"reference\"><a href=\"#cite_note-175\">[171]<\/a><\/sup> The sea level rise since 1993 has been estimated to have been on average 2.6\u00a0mm and 2.9\u00a0mm per year \u00b1 0.4\u00a0mm. Additionally, sea level rise has accelerated from 1995 to 2015.<sup id=\"cite_ref-176\" class=\"reference\"><a href=\"#cite_note-176\">[172]<\/a><\/sup> Over the 21st century, the IPCC projects for a high emissions scenario, that global mean <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sea_level_rise\" title=\"Sea level rise\">sea level<\/a> could rise by 52\u201398\u00a0cm.<sup id=\"cite_ref-IPCC_SLR2013_177-0\" class=\"reference\"><a href=\"#cite_note-IPCC_SLR2013-177\">[173]<\/a><\/sup><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming#Extreme_weather\" title=\"Effects of global warming\">Extreme weather<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Physical_impacts_of_climate_change#Extreme_events\" title=\"Physical impacts of climate change\">extreme events<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Tropical_cyclones_and_climate_change\" title=\"Tropical cyclones and climate change\">tropical cyclones<\/a>: Data analysis of extreme events from 1960 until 2010 suggests that <a href=\"https:\/\/en.wikipedia.org\/wiki\/Drought\" title=\"Drought\">droughts<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Heat_wave\" title=\"Heat wave\">heat waves<\/a> appear simultaneously with increased frequency.<sup id=\"cite_ref-178\" class=\"reference\"><a href=\"#cite_note-178\">[174]<\/a><\/sup> Extremely wet or dry events within the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Monsoon\" title=\"Monsoon\">monsoon<\/a> period have increased since 1980.<sup id=\"cite_ref-179\" class=\"reference\"><a href=\"#cite_note-179\">[175]<\/a><\/sup> Projections suggest a probable increase in the frequency and severity of some extreme weather events, such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Heat_wave\" title=\"Heat wave\">heat waves<\/a>.<sup id=\"cite_ref-180\" class=\"reference\"><a href=\"#cite_note-180\">[176]<\/a><\/sup><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_ecosystems\" title=\"Climate change and ecosystems\">Ecosystem changes<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_oceans\" title=\"Effects of global warming on oceans\">changes in ocean properties<\/a>: In terrestrial <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ecosystem\" title=\"Ecosystem\">ecosystems<\/a>, the earlier timing of spring events, as well as poleward and upward shifts in plant and animal ranges, have been linked with high confidence to recent warming.<sup id=\"cite_ref-ipcc.ch_181-0\" class=\"reference\"><a href=\"#cite_note-ipcc.ch-181\">[177]<\/a><\/sup> It is expected that most ecosystems will be affected by higher atmospheric CO<sub>2<\/sub> levels, combined with higher global temperatures.<sup id=\"cite_ref-182\" class=\"reference\"><a href=\"#cite_note-182\">[178]<\/a><\/sup><a href=\"https:\/\/en.wikipedia.org\/wiki\/Desertification\" title=\"Desertification\">Expansion of deserts<\/a> in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Subtropics\" title=\"Subtropics\">subtropics<\/a> is probably linked to global warming.<sup id=\"cite_ref-Zeng_L17401_18-1\" class=\"reference\"><a href=\"#cite_note-Zeng_L17401-18\">[17]<\/a><\/sup> The physical effect of global warming on <a href=\"https:\/\/en.wikipedia.org\/wiki\/Oceans\" class=\"mw-redirect\" title=\"Oceans\">oceans<\/a> include an increase in acidity, and a reduction of oxygen levels (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Ocean_deoxygenation\" title=\"Ocean deoxygenation\">ocean deoxygenation<\/a>).<sup id=\"cite_ref-ncadac_2013_climate_183-0\" class=\"reference\"><a href=\"#cite_note-ncadac_2013_climate-183\">[179]<\/a><\/sup><sup id=\"cite_ref-:2_184-0\" class=\"reference\"><a href=\"#cite_note-:2-184\">[180]<\/a><\/sup> Increases in atmospheric CO<sub>2<\/sub> concentrations have led to an increase in dissolved CO<sub>2<\/sub> and thus <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ocean_acidification\" title=\"Ocean acidification\">ocean acidity<\/a>, measured by lower <a href=\"https:\/\/en.wikipedia.org\/wiki\/PH\" title=\"PH\">pH<\/a> values.<sup id=\"cite_ref-ncadac_2013_climate_183-1\" class=\"reference\"><a href=\"#cite_note-ncadac_2013_climate-183\">[179]<\/a><\/sup> Ocean acidification threatens damage to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_issues_with_coral_reefs\" title=\"Environmental issues with coral reefs\">coral reefs<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Fishery\" title=\"Fishery\">fisheries<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Protected_species\" class=\"mw-redirect\" title=\"Protected species\">protected species<\/a>, and other <a href=\"https:\/\/en.wikipedia.org\/wiki\/Natural_resource\" title=\"Natural resource\">natural resources<\/a> of value to society.<sup id=\"cite_ref-ncadac_2013_climate_183-2\" class=\"reference\"><a href=\"#cite_note-ncadac_2013_climate-183\">[179]<\/a><\/sup><sup id=\"cite_ref-185\" class=\"reference\"><a href=\"#cite_note-185\">[181]<\/a><\/sup> Without substantial actions to reduce the rate of global warming, land based ecosystems are at risk of major ecological shifts, transforming composition and structure.<sup id=\"cite_ref-186\" class=\"reference\"><a href=\"#cite_note-186\">[182]<\/a><\/sup><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Long-term_effects_of_global_warming\" title=\"Long-term effects of global warming\">Long-term effects of global warming<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Runaway_climate_change\" title=\"Runaway climate change\">runaway climate change<\/a>: On the timescale of centuries to millennia, the magnitude of global warming will be determined primarily by anthropogenic CO<sub>2<\/sub> emissions.<sup id=\"cite_ref-long-term_effects_of_global_warming_187-0\" class=\"reference\"><a href=\"#cite_note-long-term_effects_of_global_warming-187\">[183]<\/a><\/sup> This is due to carbon dioxide's very long lifetime in the atmosphere.<sup id=\"cite_ref-long-term_effects_of_global_warming_187-1\" class=\"reference\"><a href=\"#cite_note-long-term_effects_of_global_warming-187\">[183]<\/a><\/sup> Long-term effects also include a response from the Earth's crust, due to ice melting and deglaciation, in a process called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Post-glacial_rebound\" title=\"Post-glacial rebound\">post-glacial rebound<\/a>, when land masses are no longer depressed by the weight of ice. This could lead to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Landslide\" title=\"Landslide\">landslides<\/a> and increased seismic and volcanic activities. Tsunamis could be generated by submarine landslides caused by warmer ocean water thawing ocean-floor permafrost or releasing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Clathrate_hydrate\" title=\"Clathrate hydrate\">gas hydrates<\/a>.<sup id=\"cite_ref-188\" class=\"reference\"><a href=\"#cite_note-188\">[184]<\/a><\/sup><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Abrupt_climate_change\" title=\"Abrupt climate change\">Abrupt climate change<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cold_blob_(North_Atlantic)\" title=\"Cold blob (North Atlantic)\">cold blob (North Atlantic)<\/a>: Climate change could result in global, large-scale changes in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Natural_environment\" title=\"Natural environment\">natural<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Society\" title=\"Society\">social systems<\/a>.<sup id=\"cite_ref-189\" class=\"reference\"><a href=\"#cite_note-189\">[185]<\/a><\/sup> Examples include ocean acidification caused by increased atmospheric concentrations of carbon dioxide, and the long-term melting of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ice_sheet\" title=\"Ice sheet\">ice sheets<\/a>, which contributes to sea level rise.<sup id=\"cite_ref-190\" class=\"reference\"><a href=\"#cite_note-190\">[186]<\/a><\/sup> Some large-scale changes could occur <a href=\"https:\/\/en.wikipedia.org\/wiki\/Abrupt_climate_change\" title=\"Abrupt climate change\">abruptly<\/a>, i.e., over a short time period, and might also be <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming#Irreversibilities\" title=\"Effects of global warming\">irreversible<\/a>. Examples of abrupt climate change are the rapid release of methane and carbon dioxide from <a href=\"https:\/\/en.wikipedia.org\/wiki\/Permafrost\" title=\"Permafrost\">permafrost<\/a>, which would lead to amplified global warming. Another example is the possibility for the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atlantic_Meridional_Overturning_Circulation\" class=\"mw-redirect\" title=\"Atlantic Meridional Overturning Circulation\">Atlantic Meridional Overturning Circulation<\/a> to slow- or shutdown (see also <a href=\"https:\/\/en.wikipedia.org\/wiki\/Shutdown_of_thermohaline_circulation\" title=\"Shutdown of thermohaline circulation\">shutdown of thermohaline circulation<\/a>).<sup id=\"cite_ref-ccsp_abrupt_climate_change_191-0\" class=\"reference\"><a href=\"#cite_note-ccsp_abrupt_climate_change-191\">[187]<\/a><\/sup><sup id=\"cite_ref-192\" class=\"reference\"><a href=\"#cite_note-192\">[188]<\/a><\/sup> This could trigger cooling in the North <a href=\"https:\/\/en.wikipedia.org\/wiki\/Atlantic\" class=\"mw-redirect\" title=\"Atlantic\">Atlantic<\/a>, Europe, and North America.<sup id=\"cite_ref-193\" class=\"reference\"><a href=\"#cite_note-193\">[189]<\/a><\/sup><sup id=\"cite_ref-194\" class=\"reference\"><a href=\"#cite_note-194\">[190]<\/a><\/sup> It would particularly affect areas such as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/British_Isles\" title=\"British Isles\">British Isles<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/France\" title=\"France\">France<\/a> and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Nordic_countries\" title=\"Nordic countries\">Nordic countries<\/a>, which are warmed by the <a href=\"https:\/\/en.wikipedia.org\/wiki\/North_Atlantic_drift\" class=\"mw-redirect\" title=\"North Atlantic drift\">North Atlantic drift<\/a>.<sup id=\"cite_ref-195\" class=\"reference\"><a href=\"#cite_note-195\">[191]<\/a><\/sup><sup id=\"cite_ref-196\" class=\"reference\"><a href=\"#cite_note-196\">[192]<\/a><\/sup><\/li><\/ul><h3>Social systems<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Further information: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_humans\" title=\"Effects of global warming on humans\">Effects of global warming on humans<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_human_health\" title=\"Effects of global warming on human health\">Effects of global warming on human health<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_national_security\" class=\"mw-redirect\" title=\"Climate change and national security\">Climate change and national security<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_refugee\" class=\"mw-redirect\" title=\"Climate refugee\">Climate refugee<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_adaptation\" title=\"Climate change adaptation\">Climate change adaptation<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Economics_of_global_warming\" title=\"Economics of global warming\">Economics of global warming<\/a><\/div>\n<p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_climate_change_on_humans\" class=\"mw-redirect\" title=\"Effects of climate change on humans\">effects of climate change on human systems<\/a>, mostly due to warming or shifts in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Precipitation\" title=\"Precipitation\">precipitation<\/a> patterns, or both, have been detected worldwide. The future social impacts of climate change will be uneven across the world.<sup id=\"cite_ref-197\" class=\"reference\"><a href=\"#cite_note-197\">[193]<\/a><\/sup> Many risks are expected to increase with higher magnitudes of global warming.<sup id=\"cite_ref-198\" class=\"reference\"><a href=\"#cite_note-198\">[194]<\/a><\/sup> All regions are at risk of experiencing negative impacts.<sup id=\"cite_ref-199\" class=\"reference\"><a href=\"#cite_note-199\">[195]<\/a><\/sup> Low-latitude, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Developing_countries\" class=\"mw-redirect\" title=\"Developing countries\">less developed areas<\/a> face the greatest risk.<sup id=\"cite_ref-200\" class=\"reference\"><a href=\"#cite_note-200\">[196]<\/a><\/sup> A study from 2015 concluded that economic growth (gross domestic product) of poorer countries is much more impaired with projected future climate warming, than previously thought.<sup id=\"cite_ref-201\" class=\"reference\"><a href=\"#cite_note-201\">[197]<\/a><\/sup> In small islands and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mega_delta\" class=\"mw-redirect\" title=\"Mega delta\">mega deltas<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Inundation\" title=\"Inundation\">inundation<\/a> as a result of sea level rise is expected to threaten vital infrastructure and human settlements.<sup id=\"cite_ref-202\" class=\"reference\"><a href=\"#cite_note-202\">[198]<\/a><\/sup><sup id=\"cite_ref-203\" class=\"reference\"><a href=\"#cite_note-203\">[199]<\/a><\/sup> This could lead to issues of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_refugee\" class=\"mw-redirect\" title=\"Climate refugee\">homelessness<\/a> in countries with low-lying areas such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bangladesh\" title=\"Bangladesh\">Bangladesh<\/a>, as well as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Statelessness\" title=\"Statelessness\">statelessness<\/a> for populations in countries such as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Maldives\" title=\"Maldives\">Maldives<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Tuvalu\" title=\"Tuvalu\">Tuvalu<\/a>.<sup id=\"cite_ref-204\" class=\"reference\"><a href=\"#cite_note-204\">[200]<\/a><\/sup><\/p><p>Examples of impacts of global warming on humans include:\n<\/p>\n<ul><li>A <a href=\"https:\/\/en.wikipedia.org\/wiki\/Meta-analysis\" title=\"Meta-analysis\">meta-analysis<\/a> concluded in 2014 that each degree of temperature rise will increase <a href=\"https:\/\/en.wikipedia.org\/wiki\/Violence\" title=\"Violence\">violence<\/a> by up to 20%, which includes fist fights, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Violent_crime\" title=\"Violent crime\">violent crimes<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Civil_disorder\" title=\"Civil disorder\">civil unrest<\/a>, or <a href=\"https:\/\/en.wikipedia.org\/wiki\/War\" title=\"War\">wars<\/a>.<sup id=\"cite_ref-205\" class=\"reference\"><a href=\"#cite_note-205\">[201]<\/a><\/sup><\/li>\n<li>Estimates in 2015 based on the IPCC A1B emission scenario from additional greenhouse gases released from permafrost, found associated impact damages to the economy to be US$43 trillion.<sup id=\"cite_ref-206\" class=\"reference\"><a href=\"#cite_note-206\">[202]<\/a><\/sup><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Crop_productivity\" class=\"mw-redirect\" title=\"Crop productivity\">Crop production<\/a> will probably be negatively affected in low latitude countries, while effects at northern latitudes may be positive or negative.<sup id=\"cite_ref-207\" class=\"reference\"><a href=\"#cite_note-207\">[203]<\/a><\/sup> Global warming of around 4.6\u00a0\u00b0C relative to pre-industrial levels could pose a large risk to global and regional food security.<sup id=\"cite_ref-208\" class=\"reference\"><a href=\"#cite_note-208\">[204]<\/a><\/sup> The impact of climate change on <a href=\"https:\/\/en.wikipedia.org\/wiki\/Crop_productivity\" class=\"mw-redirect\" title=\"Crop productivity\">crop productivity<\/a> for the four major crops was negative for wheat and maize and neutral for soy and rice in the years 1960-2013.<sup id=\"cite_ref-Fifth_Assessment_Report_(AR5),_WGII,_Climate_Change_2014:_Impacts,_Adaptation,_and_Vulnerability,_Chapter_7:_Food_Security_and_Food_Production_Systems_209-0\" class=\"reference\"><a href=\"#cite_note-Fifth_Assessment_Report_(AR5),_WGII,_Climate_Change_2014:_Impacts,_Adaptation,_and_Vulnerability,_Chapter_7:_Food_Security_and_Food_Production_Systems-209\">[205]<\/a><\/sup> While crop production has increased in some mid-latitude regions such as the UK and Northeast China, economic losses due to extreme weather events have increased globally.<sup id=\"cite_ref-:3_210-0\" class=\"reference\"><a href=\"#cite_note-:3-210\">[206]<\/a><\/sup> See also <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_agriculture\" title=\"Climate change and agriculture\">Climate change and agriculture<\/a>.<\/li>\n<li>Generally <a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_human_health\" title=\"Effects of global warming on human health\">impacts on public health<\/a> will be more negative than positive.<sup id=\"cite_ref-211\" class=\"reference\"><a href=\"#cite_note-211\">[207]<\/a><\/sup><sup id=\"cite_ref-lancet2009_212-0\" class=\"reference\"><a href=\"#cite_note-lancet2009-212\">[208]<\/a><\/sup><sup id=\"cite_ref-lancet2015_213-0\" class=\"reference\"><a href=\"#cite_note-lancet2015-213\">[209]<\/a><\/sup> Impacts include: the effects of extreme weather, leading to injury and loss of life;<sup id=\"cite_ref-214\" class=\"reference\"><a href=\"#cite_note-214\">[210]<\/a><\/sup> and indirect effects, such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Undernutrition\" class=\"mw-redirect\" title=\"Undernutrition\">undernutrition<\/a> brought on by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Crop_failure\" class=\"mw-redirect\" title=\"Crop failure\">crop failures<\/a>.<sup id=\"cite_ref-lancet2009_212-1\" class=\"reference\"><a href=\"#cite_note-lancet2009-212\">[208]<\/a><\/sup><sup id=\"cite_ref-lancet2015_213-1\" class=\"reference\"><a href=\"#cite_note-lancet2015-213\">[209]<\/a><\/sup><sup id=\"cite_ref-215\" class=\"reference\"><a href=\"#cite_note-215\">[211]<\/a><\/sup> There has been a shift from cold- to heat-related mortality in some regions as a result of warming.<sup id=\"cite_ref-:3_210-1\" class=\"reference\"><a href=\"#cite_note-:3-210\">[206]<\/a><\/sup> A 2018 study of data from the U.S. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Centers_for_Disease_Control_and_Prevention\" title=\"Centers for Disease Control and Prevention\">Centers for Disease Control and Prevention<\/a> connected temperature rise to increased numbers of suicides.<sup id=\"cite_ref-216\" class=\"reference\"><a href=\"#cite_note-216\">[212]<\/a><\/sup> The study revealed that hotter days could increase suicide rates and could cause approximately 26,000 more suicides in the U.S. by 2050.<sup id=\"cite_ref-217\" class=\"reference\"><a href=\"#cite_note-217\">[213]<\/a><\/sup><\/li>\n<li>Livelihoods of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Circumpolar_peoples\" title=\"Circumpolar peoples\">indigenous peoples of the Arctic<\/a> have been altered by climate change, and there is emerging evidence of climate change impacts on livelihoods of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Indigenous_peoples\" title=\"Indigenous peoples\">indigenous peoples<\/a> in other regions. Regional impacts of climate change are now observable at more locations than before, on all continents and across ocean regions.<sup id=\"cite_ref-:3_210-2\" class=\"reference\"><a href=\"#cite_note-:3-210\">[206]<\/a><\/sup><\/li><\/ul><h3>Regional<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Regional_effects_of_global_warming\" title=\"Regional effects of global warming\">Regional effects of global warming<\/a><\/div>\n<p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic\" title=\"Arctic\">Arctic<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Africa\" title=\"Africa\">Africa<\/a>, small islands and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Asia\" title=\"Asia\">Asian<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/River_delta\" title=\"River delta\">megadeltas<\/a> are regions that are likely to be especially affected by future climate change.<sup id=\"cite_ref-218\" class=\"reference\"><a href=\"#cite_note-218\">[214]<\/a><\/sup><a href=\"https:\/\/en.wikipedia.org\/wiki\/Africa\" title=\"Africa\">Africa<\/a> is one of the most vulnerable continents to climate variability and change because of multiple existing stresses and low <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_adaptation\" class=\"mw-redirect\" title=\"Climate adaptation\">adaptive capacity<\/a>.<sup id=\"cite_ref-us_epa_regional_impacts_219-0\" class=\"reference\"><a href=\"#cite_note-us_epa_regional_impacts-219\">[215]<\/a><\/sup> Existing stresses include poverty, political conflicts, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ecosystem\" title=\"Ecosystem\">ecosystem<\/a> degradation. By 2050, between 350 million and 600 million people are projected to experience increased <a href=\"https:\/\/en.wikipedia.org\/wiki\/Water_stress\" class=\"mw-redirect\" title=\"Water stress\">water stress<\/a> due to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change\" title=\"Climate change\">climate change<\/a> (see <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Africa\" title=\"Climate change in Africa\">Climate change in Africa<\/a>).<sup id=\"cite_ref-us_epa_regional_impacts_219-1\" class=\"reference\"><a href=\"#cite_note-us_epa_regional_impacts-219\">[215]<\/a><\/sup> Climate variability and change is projected to severely compromise <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_agriculture\" title=\"Climate change and agriculture\">agricultural production<\/a>, including access to food, across Africa.<sup id=\"cite_ref-us_epa_regional_impacts_219-2\" class=\"reference\"><a href=\"#cite_note-us_epa_regional_impacts-219\">[215]<\/a><\/sup> Research projects that regions even may become uninhabitable, due to a high <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wet-bulb_temperature\" title=\"Wet-bulb temperature\">wet-bulb temperature<\/a>.<sup id=\"cite_ref-220\" class=\"reference\"><a href=\"#cite_note-220\">[216]<\/a><\/sup><\/p>\n<h2>Responses<\/h2>\n<h3>Mitigation<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_mitigation\" title=\"Climate change mitigation\">Climate change mitigation<\/a><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_(PBL).png\" class=\"image\"><img alt=\"Refer to caption and image description\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/d\/d8\/Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_%28PBL%29.png\/300px-Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_%28PBL%29.png\" width=\"300\" height=\"327\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/d\/d8\/Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_%28PBL%29.png\/450px-Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_%28PBL%29.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/d\/d8\/Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_%28PBL%29.png\/600px-Limiting_global_warming_to_2_degrees_Celsius_-_options_to_reduce_greenhouse_gas_emissions_%28PBL%29.png 2x\" data-file-width=\"1471\" data-file-height=\"1604\" \/><\/a>  <div class=\"thumbcaption\">The graph on the right shows three \"pathways\" to meet the UNFCCC's 2\u00a0\u00b0C target, labelled \"global technology\", \"decentralized solutions\", and \"consumption change\". Each pathway shows how various measures (e.g., improved energy efficiency, increased use of renewable energy) could contribute to emissions reductions. Image credit: PBL <a href=\"https:\/\/en.wikipedia.org\/wiki\/Netherlands_Environmental_Assessment_Agency\" title=\"Netherlands Environmental Assessment Agency\">Netherlands Environmental Assessment Agency<\/a>.<sup id=\"cite_ref-221\" class=\"reference\"><a href=\"#cite_note-221\">[217]<\/a><\/sup><\/div><\/div><\/div>\n<p>Mitigation of climate change are actions to reduce greenhouse gas emissions, or enhance the capacity of carbon sinks to absorb greenhouse gases from the atmosphere.<sup id=\"cite_ref-222\" class=\"reference\"><a href=\"#cite_note-222\">[218]<\/a><\/sup> There is a large potential for future reductions in emissions by a combination of activities, including <a href=\"https:\/\/en.wikipedia.org\/wiki\/Energy_conservation\" title=\"Energy conservation\">energy conservation<\/a> and increased <a href=\"https:\/\/en.wikipedia.org\/wiki\/Efficient_energy_use\" title=\"Efficient energy use\">energy efficiency<\/a>; the use of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Low-carbon_energy\" class=\"mw-redirect\" title=\"Low-carbon energy\">low-carbon energy<\/a> technologies, such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Renewable_energy\" title=\"Renewable energy\">renewable energy<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Nuclear_power\" title=\"Nuclear power\">nuclear energy<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_capture_and_storage\" title=\"Carbon capture and storage\">carbon capture and storage<\/a>;<sup id=\"cite_ref-autogenerated2007_223-0\" class=\"reference\"><a href=\"#cite_note-autogenerated2007-223\">[219]<\/a><\/sup><sup id=\"cite_ref-2014_mitigation_options_224-0\" class=\"reference\"><a href=\"#cite_note-2014_mitigation_options-224\">[220]<\/a><\/sup> and enhancing carbon sinks through, for example, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Reforestation\" title=\"Reforestation\">reforestation<\/a> and preventing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Deforestation\" title=\"Deforestation\">deforestation<\/a>.<sup id=\"cite_ref-autogenerated2007_223-1\" class=\"reference\"><a href=\"#cite_note-autogenerated2007-223\">[219]<\/a><\/sup><sup id=\"cite_ref-2014_mitigation_options_224-1\" class=\"reference\"><a href=\"#cite_note-2014_mitigation_options-224\">[220]<\/a><\/sup>\nA 2015 report by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Citibank\" title=\"Citibank\">Citibank<\/a> concluded that transitioning to a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Low_carbon_economy\" class=\"mw-redirect\" title=\"Low carbon economy\">low carbon economy<\/a> would yield positive return on investments.<sup id=\"cite_ref-225\" class=\"reference\"><a href=\"#cite_note-225\">[221]<\/a><\/sup><\/p><p>Near- and long-term trends in the global energy system are inconsistent with limiting global warming at below 1.5 or 2\u00a0\u00b0C, relative to pre-industrial levels.<sup id=\"cite_ref-226\" class=\"reference\"><a href=\"#cite_note-226\">[222]<\/a><\/sup><sup id=\"cite_ref-progress_to_2_degrees_C_227-0\" class=\"reference\"><a href=\"#cite_note-progress_to_2_degrees_C-227\">[223]<\/a><\/sup> Pledges made as part of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/2010_United_Nations_Climate_Change_Conference\" title=\"2010 United Nations Climate Change Conference\">Canc\u00fan agreements<\/a> are broadly consistent with having a likely chance (66 to 100% probability) of limiting global warming (in the 21st century) at below 3\u00a0\u00b0C, relative to pre-industrial levels.<sup id=\"cite_ref-progress_to_2_degrees_C_227-1\" class=\"reference\"><a href=\"#cite_note-progress_to_2_degrees_C-227\">[223]<\/a><\/sup><\/p><p>In limiting warming at below 2\u00a0\u00b0C, more stringent emission reductions in the near-term would allow for less rapid reductions after 2030.<sup id=\"cite_ref-228\" class=\"reference\"><a href=\"#cite_note-228\">[224]<\/a><\/sup> Many integrated models are unable to meet the 2\u00a0\u00b0C target if pessimistic assumptions are made about the availability of mitigation technologies.<sup id=\"cite_ref-229\" class=\"reference\"><a href=\"#cite_note-229\">[225]<\/a><\/sup><\/p>\n<h3>Adaptation<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_adaptation\" title=\"Climate change adaptation\">Climate change adaptation<\/a><\/div>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_adaptation\" title=\"Climate change adaptation\">Climate change adaptation<\/a> is another policy response. The adaptation may be planned, either in reaction to or anticipation of global warming, or spontaneous, i.e., without government intervention.<sup id=\"cite_ref-230\" class=\"reference\"><a href=\"#cite_note-230\">[226]<\/a><\/sup> Planned adaptation is already occurring on a limited basis.<sup id=\"cite_ref-autogenerated2007_223-2\" class=\"reference\"><a href=\"#cite_note-autogenerated2007-223\">[219]<\/a><\/sup> The barriers, limits, and costs of future adaptation are not fully understood.<sup id=\"cite_ref-autogenerated2007_223-3\" class=\"reference\"><a href=\"#cite_note-autogenerated2007-223\">[219]<\/a><\/sup> Environmental organizations and public figures have emphasized changes in the climate and the risks they entail, while promoting adaptation to changes in infrastructural needs and emissions reductions.<sup id=\"cite_ref-231\" class=\"reference\"><a href=\"#cite_note-231\">[227]<\/a><\/sup><\/p><p>Adaptation is especially important in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Developing_countries\" class=\"mw-redirect\" title=\"Developing countries\">developing countries<\/a> since those countries are predicted to bear the brunt of the effects of global warming.<sup id=\"cite_ref-232\" class=\"reference\"><a href=\"#cite_note-232\">[228]<\/a><\/sup> That is, the capacity and potential for humans to adapt (called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Adaptive_capacity\" title=\"Adaptive capacity\">adaptive capacity<\/a>) is unevenly distributed across different regions and populations, and developing countries generally have less capacity to adapt.<sup id=\"cite_ref-233\" class=\"reference\"><a href=\"#cite_note-233\">[229]<\/a><\/sup><\/p>\n<h3>Climate engineering<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_engineering\" title=\"Climate engineering\">Climate engineering<\/a><\/div>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_engineering\" title=\"Climate engineering\">Climate engineering<\/a> (sometimes called <i>geoengineering<\/i> or <i>climate intervention<\/i>) is the deliberate modification of the climate. It has been investigated as a possible response to global warming, e.g. by NASA<sup id=\"cite_ref-234\" class=\"reference\"><a href=\"#cite_note-234\">[230]<\/a><\/sup> and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Royal_Society\" title=\"Royal Society\">Royal Society<\/a>.<sup id=\"cite_ref-235\" class=\"reference\"><a href=\"#cite_note-235\">[231]<\/a><\/sup> Techniques under research fall generally into the categories <a href=\"https:\/\/en.wikipedia.org\/wiki\/Solar_radiation_management\" title=\"Solar radiation management\">solar radiation management<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide_removal\" title=\"Carbon dioxide removal\">carbon dioxide removal<\/a>, although various other schemes have been suggested. A study from 2014 investigated the most common climate engineering methods and concluded they are either ineffective or have potentially severe side effects and cannot be stopped without causing rapid climate change.<sup id=\"cite_ref-236\" class=\"reference\"><a href=\"#cite_note-236\">[232]<\/a><\/sup><\/p>\n<h2>Society and culture<\/h2>\n<h3>Political discussion<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Politics_of_global_warming\" title=\"Politics of global warming\">Politics of global warming<\/a><\/div>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Further information: <a href=\"https:\/\/en.wikipedia.org\/wiki\/2011_United_Nations_Climate_Change_Conference\" title=\"2011 United Nations Climate Change Conference\">2011<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/2012_United_Nations_Climate_Change_Conference\" title=\"2012 United Nations Climate Change Conference\">2012<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/2013_United_Nations_Climate_Change_Conference\" title=\"2013 United Nations Climate Change Conference\">2013<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/2015_United_Nations_Climate_Change_Conference\" title=\"2015 United Nations Climate Change Conference\">2015<\/a> sessions of <a href=\"https:\/\/en.wikipedia.org\/wiki\/United_Nations_Climate_Change_Conference\" class=\"mw-redirect\" title=\"United Nations Climate Change Conference\">United Nations Climate Change Conference<\/a><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_(vertical)_1.svg\" class=\"image\"><img alt=\"refer to caption\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/d\/de\/Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_%28vertical%29_1.svg\/220px-Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_%28vertical%29_1.svg.png\" width=\"220\" height=\"286\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/d\/de\/Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_%28vertical%29_1.svg\/330px-Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_%28vertical%29_1.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/d\/de\/Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_%28vertical%29_1.svg\/440px-Stabilizing_the_atmospheric_concentration_of_carbon_dioxide_at_a_constant_level_would_require_emissions_to_be_effectively_eliminated_%28vertical%29_1.svg.png 2x\" data-file-width=\"457\" data-file-height=\"594\" \/><\/a>  <div class=\"thumbcaption\">Article 2 of the UN Framework Convention refers explicitly to \"stabilization of greenhouse gas concentrations\".<sup id=\"cite_ref-237\" class=\"reference\"><a href=\"#cite_note-237\">[233]<\/a><\/sup> To stabilize the atmospheric concentration of CO<sup><br \/><sub>2<\/sub>, emissions worldwide would need to be dramatically reduced from their present level.<sup id=\"cite_ref-238\" class=\"reference\"><a href=\"#cite_note-238\">[234]<\/a><\/sup><\/sup><\/div><\/div><\/div>\n<p>Most countries in the world are parties to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/United_Nations_Framework_Convention_on_Climate_Change\" title=\"United Nations Framework Convention on Climate Change\">United Nations Framework Convention on Climate Change<\/a> (UNFCCC).<sup id=\"cite_ref-239\" class=\"reference\"><a href=\"#cite_note-239\">[235]<\/a><\/sup>\nThe ultimate objective of the Convention is to prevent dangerous human interference of the climate system.<sup id=\"cite_ref-240\" class=\"reference\"><a href=\"#cite_note-240\">[236]<\/a><\/sup>\nAs stated in the Convention, this requires that greenhouse gas concentrations are stabilized in the atmosphere at a level where <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_ecosystems\" title=\"Climate change and ecosystems\">ecosystems<\/a> can adapt naturally to climate change, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_agriculture\" title=\"Climate change and agriculture\">food production<\/a> is not threatened, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Economics_of_global_warming\" title=\"Economics of global warming\">economic development<\/a> can proceed in a sustainable fashion.<sup id=\"cite_ref-241\" class=\"reference\"><a href=\"#cite_note-241\">[237]<\/a><\/sup> The Framework Convention was agreed on in 1992, but global emissions have risen since then.<sup id=\"cite_ref-242\" class=\"reference\"><a href=\"#cite_note-242\">[238]<\/a><\/sup><\/p><p>During negotiations, the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Group_of_77\" title=\"Group of 77\">G77<\/a> (a lobbying group in the United Nations representing 133 <a href=\"https:\/\/en.wikipedia.org\/wiki\/Developing_country\" title=\"Developing country\">developing countries<\/a>)<sup id=\"cite_ref-dessai_243-0\" class=\"reference\"><a href=\"#cite_note-dessai-243\">[239]<\/a><\/sup><sup class=\"reference\">:4<\/sup> pushed for a mandate requiring developed countries to \"[take] the lead\" in reducing their emissions.<sup id=\"cite_ref-244\" class=\"reference\"><a href=\"#cite_note-244\">[240]<\/a><\/sup> This was justified on the basis that the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Developed_country\" title=\"Developed country\">developed countries<\/a>' emissions had contributed most to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas#Cumulative_and_historical_emissions\" title=\"Greenhouse gas\">cumulation<\/a> of greenhouse gases in the atmosphere, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas#Annual_and_per_capita_emissions\" title=\"Greenhouse gas\">per-capita emissions<\/a> (i.e., emissions per head of population) were still relatively low in developing countries, and the emissions of developing countries would grow to meet their development needs.<sup id=\"cite_ref-liverman_95-1\" class=\"reference\"><a href=\"#cite_note-liverman-95\">[91]<\/a><\/sup><sup class=\"reference\">:290<\/sup><\/p><p>This mandate was sustained in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Kyoto_Protocol\" title=\"Kyoto Protocol\">Kyoto Protocol<\/a> to the Framework Convention,<sup id=\"cite_ref-liverman_95-2\" class=\"reference\"><a href=\"#cite_note-liverman-95\">[91]<\/a><\/sup><sup class=\"reference\">:290<\/sup> which entered into legal effect in 2005.<sup id=\"cite_ref-kyoto_245-0\" class=\"reference\"><a href=\"#cite_note-kyoto-245\">[241]<\/a><\/sup> In ratifying the Kyoto Protocol, most developed countries accepted legally binding commitments to limit their emissions. These first-round commitments expired in 2012.<sup id=\"cite_ref-kyoto_245-1\" class=\"reference\"><a href=\"#cite_note-kyoto-245\">[241]<\/a><\/sup>\nUnited States President <a href=\"https:\/\/en.wikipedia.org\/wiki\/George_W._Bush\" title=\"George W. Bush\">George W. Bush<\/a> rejected the treaty on the basis that \"it exempts 80% of the world, including major population centres such as China and India, from compliance, and would cause serious harm to the US economy\".<sup id=\"cite_ref-dessai_243-1\" class=\"reference\"><a href=\"#cite_note-dessai-243\">[239]<\/a><\/sup><sup class=\"reference\">:5<\/sup><\/p><p>At the <a href=\"https:\/\/en.wikipedia.org\/wiki\/2009_United_Nations_Climate_Change_Conference\" title=\"2009 United Nations Climate Change Conference\">15th UNFCCC Conference of the Parties<\/a>, held in 2009 at <a href=\"https:\/\/en.wikipedia.org\/wiki\/Copenhagen\" title=\"Copenhagen\">Copenhagen<\/a>, several UNFCCC Parties produced the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Copenhagen_Accord\" title=\"Copenhagen Accord\">Copenhagen Accord<\/a>.<sup id=\"cite_ref-246\" class=\"reference\"><a href=\"#cite_note-246\">[242]<\/a><\/sup><sup id=\"cite_ref-NYT-20150525_247-0\" class=\"reference\"><a href=\"#cite_note-NYT-20150525-247\">[243]<\/a><\/sup> Parties associated with the Accord (140 countries, as of November 2010)<sup id=\"cite_ref-unep_248-0\" class=\"reference\"><a href=\"#cite_note-unep-248\">[244]<\/a><\/sup><sup class=\"reference\">:9<\/sup> aim to limit the future increase in global mean temperature to below 7002275149999999999\u26602\u00a0\u00b0C.<sup id=\"cite_ref-249\" class=\"reference\"><a href=\"#cite_note-249\">[245]<\/a><\/sup> The 16th Conference of the Parties (COP16) was held at <a href=\"https:\/\/en.wikipedia.org\/wiki\/Canc\u00fan\" title=\"Canc\u00fan\">Canc\u00fan<\/a> in 2010. It produced an agreement, not a binding treaty, that the Parties should take urgent action to reduce greenhouse gas emissions to meet a goal of limiting global warming to 7002275149999999999\u26602\u00a0\u00b0C above pre-industrial temperatures. It also recognized the need to consider strengthening the goal to a global average rise of 7002274649999999999\u26601.5\u00a0\u00b0C.<sup id=\"cite_ref-250\" class=\"reference\"><a href=\"#cite_note-250\">[246]<\/a><\/sup><\/p>\n<h3>Scientific discussion<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_opinion_on_climate_change\" title=\"Scientific opinion on climate change\">Scientific opinion on climate change<\/a><\/div>\n<p>The discussion continues in scientific articles that are peer-reviewed and assessed by scientists who\nwork in the relevant fields and participate in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">Intergovernmental Panel on Climate Change<\/a>. The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_consensus\" title=\"Scientific consensus\">scientific consensus<\/a> as of 2013 stated in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Fifth_Assessment_Report\" title=\"IPCC Fifth Assessment Report\">IPCC Fifth Assessment Report<\/a> is that it \"is extremely likely that human influence has been the dominant cause of the observed warming since the mid-20th century\".<sup id=\"cite_ref-ipcc_251-0\" class=\"reference\"><a href=\"#cite_note-ipcc-251\">[247]<\/a><\/sup>\nA 2008 report by the U.S. <a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Academy_of_Sciences\" title=\"National Academy of Sciences\">National Academy of Sciences<\/a> stated that most scientists by then agreed that observed warming in recent decades was primarily caused by human activities increasing the amount of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas\" title=\"Greenhouse gas\">greenhouse gases<\/a> in the atmosphere.<sup id=\"cite_ref-nrc2008_92-1\" class=\"reference\"><a href=\"#cite_note-nrc2008-92\">[88]<\/a><\/sup> In 2005 the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Royal_Society\" title=\"Royal Society\">Royal Society<\/a> stated that while the overwhelming majority of scientists were in agreement on the main points, some individuals and organizations opposed to the consensus on urgent action needed to reduce greenhouse gas emissions had tried to undermine the science and work of the IPCC.<sup id=\"cite_ref-royal_society_misleading_arguments_252-0\" class=\"reference\"><a href=\"#cite_note-royal_society_misleading_arguments-252\">[248]<\/a><\/sup> National science academies have called on world leaders for policies to cut global emissions.<sup id=\"cite_ref-253\" class=\"reference\"><a href=\"#cite_note-253\">[249]<\/a><\/sup><\/p><p>In 2018, the IPCC published <a href=\"https:\/\/en.wikipedia.org\/wiki\/Special_Report_on_Global_Warming_of_1.5\u00baC\" class=\"mw-redirect\" title=\"Special Report on Global Warming of 1.5\u00baC\">SR15<\/a>, which warned that if current rate of greenhouse gas emissions are not mitigated, major crises could occur by 2040 as the planet warms by 2.7 degrees Fahrenheit (1.5 degrees Celsius). The report said that preventing such crises will require a swift transformation of the global economy that has \"no documented historic precedent.\"<sup id=\"cite_ref-254\" class=\"reference\"><a href=\"#cite_note-254\">[250]<\/a><\/sup><\/p><p>In the scientific literature, there is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_opinion_on_climate_change\" title=\"Scientific opinion on climate change\">strong consensus that global surface temperatures have increased<\/a> in recent decades and that the trend is caused mainly by human-induced emissions of greenhouse gases.<sup id=\"cite_ref-255\" class=\"reference\"><a href=\"#cite_note-255\">[251]<\/a><\/sup> No scientific body of national or international standing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_opinion_on_climate_change#Opposing\" title=\"Scientific opinion on climate change\">disagrees with this view<\/a>.<sup id=\"cite_ref-The_MIT_Press_256-0\" class=\"reference\"><a href=\"#cite_note-The_MIT_Press-256\">[252]<\/a><\/sup><sup id=\"cite_ref-AQAonAAPG_257-0\" class=\"reference\"><a href=\"#cite_note-AQAonAAPG-257\">[253]<\/a><\/sup> In November 2017, a second <a href=\"https:\/\/en.wikipedia.org\/wiki\/World_Scientists%27_Warning_to_Humanity\" title=\"World Scientists' Warning to Humanity\">warning to humanity<\/a> signed by 15,364 scientists from 184 countries stated that \"the current trajectory of potentially catastrophic climate change due to rising greenhouse gases from burning fossil fuels, deforestation, and agricultural production\u00a0\u2013 particularly from farming ruminants for meat consumption\" is \"especially troubling\".<sup id=\"cite_ref-258\" class=\"reference\"><a href=\"#cite_note-258\">[254]<\/a><\/sup> A July 2017 study published in <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_Research_Letters\" title=\"Environmental Research Letters\">Environmental Research Letters<\/a><\/i> asserts that the most significant action individuals could make to mitigate their own carbon footprint is to have fewer children, followed by living vehicle free, forgoing air travel and adopting a plant-based diet.<sup id=\"cite_ref-259\" class=\"reference\"><a href=\"#cite_note-259\">[255]<\/a><\/sup><\/p>\n<h3>Public opinion and disputes<\/h3>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main articles: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_denial\" title=\"Climate change denial\">Climate change denial<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_controversy\" title=\"Global warming controversy\">Global warming controversy<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Media_coverage_of_climate_change\" class=\"mw-redirect\" title=\"Media coverage of climate change\">Media coverage of climate change<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Public_opinion_on_climate_change\" class=\"mw-redirect\" title=\"Public opinion on climate change\">Public opinion on climate change<\/a><\/div>\n<div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Ms._magazine_Cover_-_Spring_2007.jpg\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c9\/Ms._magazine_Cover_-_Spring_2007.jpg\/170px-Ms._magazine_Cover_-_Spring_2007.jpg\" width=\"170\" height=\"228\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c9\/Ms._magazine_Cover_-_Spring_2007.jpg\/255px-Ms._magazine_Cover_-_Spring_2007.jpg 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c9\/Ms._magazine_Cover_-_Spring_2007.jpg\/340px-Ms._magazine_Cover_-_Spring_2007.jpg 2x\" data-file-width=\"1662\" data-file-height=\"2233\" \/><\/a>  <div class=\"thumbcaption\">Global warming was the cover story in this 2007 issue of <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ms._(magazine)\" title=\"Ms. (magazine)\">Ms.<\/a><\/i> magazine<\/div><\/div><\/div>\n<p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_controversy\" title=\"Global warming controversy\">global warming controversy<\/a> refers to a variety of disputes, substantially more pronounced in the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Media_coverage_of_climate_change\" class=\"mw-redirect\" title=\"Media coverage of climate change\">popular media<\/a> than in the scientific literature,<sup id=\"cite_ref-260\" class=\"reference\"><a href=\"#cite_note-260\">[256]<\/a><\/sup><sup id=\"cite_ref-merchants_doubt_261-0\" class=\"reference\"><a href=\"#cite_note-merchants_doubt-261\">[257]<\/a><\/sup> regarding the nature, causes, and consequences of global warming. The disputed issues include the causes of increased <a href=\"https:\/\/en.wikipedia.org\/wiki\/Instrumental_temperature_record\" title=\"Instrumental temperature record\">global average air temperature<\/a>, especially since the mid-20th century, whether this warming trend is unprecedented or within normal climatic variations, whether humankind has contributed significantly to it, and whether the increase is completely or partially an artefact of poor measurements. Additional disputes concern estimates of climate sensitivity, predictions of additional warming, and what the consequences of global warming will be.\n<\/p><p>In the United States from about 1990 onwards, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Conservatism_in_the_United_States\" title=\"Conservatism in the United States\">American conservative<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Think_tank\" title=\"Think tank\">think tanks<\/a> had begun challenging the legitimacy of global warming as a social problem. They <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_denial\" title=\"Climate change denial\">challenged the scientific evidence<\/a>, argued that <a href=\"https:\/\/en.wikipedia.org\/wiki\/Benefits_of_global_warming\" class=\"mw-redirect\" title=\"Benefits of global warming\">global warming would have benefits<\/a>, and asserted that proposed solutions would do more harm than good.<sup id=\"cite_ref-262\" class=\"reference\"><a href=\"#cite_note-262\">[258]<\/a><\/sup> Some people dispute aspects of climate change science.<sup id=\"cite_ref-royal_society_misleading_arguments_252-1\" class=\"reference\"><a href=\"#cite_note-royal_society_misleading_arguments-252\">[248]<\/a><\/sup><sup id=\"cite_ref-263\" class=\"reference\"><a href=\"#cite_note-263\">[259]<\/a><\/sup> Organizations such as the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Libertarianism\" title=\"Libertarianism\">libertarian<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Competitive_Enterprise_Institute\" title=\"Competitive Enterprise Institute\">Competitive Enterprise Institute<\/a>, conservative commentators, and some companies such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/ExxonMobil\" title=\"ExxonMobil\">ExxonMobil<\/a> have challenged IPCC climate change scenarios, funded scientists who disagree with the scientific consensus, and provided their own projections of the economic cost of stricter controls.<sup id=\"cite_ref-264\" class=\"reference\"><a href=\"#cite_note-264\">[260]<\/a><\/sup><sup id=\"cite_ref-265\" class=\"reference\"><a href=\"#cite_note-265\">[261]<\/a><\/sup><sup id=\"cite_ref-MSNBC_01-12_266-0\" class=\"reference\"><a href=\"#cite_note-MSNBC_01-12-266\">[262]<\/a><\/sup><sup id=\"cite_ref-267\" class=\"reference\"><a href=\"#cite_note-267\">[263]<\/a><\/sup> On the other hand, some fossil fuel companies have scaled back their efforts in recent years,<sup id=\"cite_ref-268\" class=\"reference\"><a href=\"#cite_note-268\">[264]<\/a><\/sup> or even called for policies to reduce global warming.<sup id=\"cite_ref-269\" class=\"reference\"><a href=\"#cite_note-269\">[265]<\/a><\/sup> Global oil companies have begun to acknowledge climate change exists and is caused by human activities and the burning of fossil fuels.<sup id=\"cite_ref-270\" class=\"reference\"><a href=\"#cite_note-270\">[266]<\/a><\/sup><\/p><p>The global warming problem came to international public attention in the late 1980s. Polling groups began to track opinions on the subject, at first mainly in the United States.<sup id=\"cite_ref-271\" class=\"reference\"><a href=\"#cite_note-271\">[267]<\/a><\/sup> The longest consistent polling, by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gallup_(company)\" title=\"Gallup (company)\">Gallup<\/a> in the US, found relatively small deviations of 10% or so from 1998 to 2015 in opinion on the seriousness of global warming, but with increasing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Polarization_(politics)\" title=\"Polarization (politics)\">polarization<\/a> between those concerned and those unconcerned.<sup id=\"cite_ref-272\" class=\"reference\"><a href=\"#cite_note-272\">[268]<\/a><\/sup><\/p><p>Due to confusing media coverage in the early 1990s, issues such as ozone depletion and climate change were often mixed up, affecting public understanding of these issues.<sup id=\"cite_ref-Newell2006_273-0\" class=\"reference\"><a href=\"#cite_note-Newell2006-273\">[269]<\/a><\/sup> According to a 2010 survey of Americans, a majority thought that the ozone layer and spray cans contribute to global warming.<sup id=\"cite_ref-274\" class=\"reference\"><a href=\"#cite_note-274\">[270]<\/a><\/sup> Although there are a few <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ozone_depletion#Ozone_depletion_and_global_warming\" title=\"Ozone depletion\">areas of linkage<\/a>, the relationship between the two is not strong. Reduced stratospheric ozone has had a slight cooling influence on surface temperatures, while increased tropospheric ozone has had a somewhat larger warming effect.<sup id=\"cite_ref-275\" class=\"reference\"><a href=\"#cite_note-275\">[271]<\/a><\/sup> However, the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Chlorofluorocarbon\" title=\"Chlorofluorocarbon\">CFC<\/a>'s used in spray cans are powerful greenhouse gases, with some estimates attributing CFC emissions during the 70s to have caused almost half of the global warming for that decade.<sup id=\"cite_ref-276\" class=\"reference\"><a href=\"#cite_note-276\">[272]<\/a><\/sup><\/p><p>By 2010, with 111 countries surveyed, Gallup determined that there had been a substantial decrease since 2007\u20132008 in the number of Americans and Europeans who viewed global warming as a serious threat. In the US, just a little over half the population (53%) viewed it as a serious concern for either themselves or their families; this was 10 points below the 2008 poll (63%). Latin America had the biggest rise in concern: 73% said global warming was a serious threat to their families.<sup id=\"cite_ref-277\" class=\"reference\"><a href=\"#cite_note-277\">[273]<\/a><\/sup> This global poll also found that people were more likely to attribute global warming to human activities than to natural causes, except in the US where nearly half (47%) of the population attributed global warming to natural causes.<sup id=\"cite_ref-278\" class=\"reference\"><a href=\"#cite_note-278\">[274]<\/a><\/sup><\/p><p>A March\u2013May 2013 survey by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pew_Research_Center_for_the_People_%26_the_Press\" class=\"mw-redirect\" title=\"Pew Research Center for the People &amp; the Press\">Pew Research Center for the People &amp; the Press<\/a> polled 39 countries about global threats. According to 54% of those questioned, global warming featured top of the perceived global threats.<sup id=\"cite_ref-279\" class=\"reference\"><a href=\"#cite_note-279\">[275]<\/a><\/sup><\/p>\n<h2>History<\/h2>\n<div role=\"note\" class=\"hatnote navigation-not-searchable\">Main article: <a href=\"https:\/\/en.wikipedia.org\/wiki\/History_of_climate_change_science\" title=\"History of climate change science\">History of climate change science<\/a><\/div>\n<p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/History_of_climate_change_science\" title=\"History of climate change science\">history of climate change science<\/a> began in the early 19th century when <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ice_ages\" class=\"mw-redirect\" title=\"Ice ages\">ice ages<\/a> and other natural changes in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Paleoclimate\" class=\"mw-redirect\" title=\"Paleoclimate\">paleoclimate<\/a> were first suspected and the natural <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_effect\" title=\"Greenhouse effect\">greenhouse effect<\/a> first identified.<sup id=\"cite_ref-:4_67-1\" class=\"reference\"><a href=\"#cite_note-:4-67\">[64]<\/a><\/sup> In the late 19th century, scientists first argued that human emissions of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gases\" class=\"mw-redirect\" title=\"Greenhouse gases\">greenhouse gases<\/a> could change the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate\" title=\"Climate\">climate<\/a>. In the 1960s, the warming effect of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide\" title=\"Carbon dioxide\">carbon dioxide<\/a> gas became increasingly convincing.<sup id=\"cite_ref-Weart_27_280-0\" class=\"reference\"><a href=\"#cite_note-Weart_27-280\">[276]<\/a><\/sup> By the 1990s, as a result of improving fidelity of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Computer_models\" class=\"mw-redirect\" title=\"Computer models\">computer models<\/a> and observational work confirming the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Milankovitch_theory\" class=\"mw-redirect\" title=\"Milankovitch theory\">Milankovitch theory<\/a> of the ice ages, a consensus position formed: greenhouse gases were deeply involved in most climate changes and human caused emissions were bringing discernible global warming. Since the 1990s, scientific research on climate change has included multiple disciplines and has expanded.<sup id=\"cite_ref-Weart_Hansen_281-0\" class=\"reference\"><a href=\"#cite_note-Weart_Hansen-281\">[277]<\/a><\/sup> Research during this period has been summarized in the Assessment Reports by the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">Intergovernmental Panel on Climate Change<\/a>.\n<\/p>\n<h3>Terminology<\/h3>\n<p>In the 1950s, research suggested increasing temperatures, and a 1952 newspaper reported \"climate change\". This phrase next appeared in a November 1957 report in <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Times_of_Northwest_Indiana\" title=\"The Times of Northwest Indiana\">The Hammond Times<\/a><\/i> which described <a href=\"https:\/\/en.wikipedia.org\/wiki\/Roger_Revelle\" title=\"Roger Revelle\">Roger Revelle<\/a>'s research into the effects of increasing human-caused CO<sub>2<\/sub> emissions on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_effect\" title=\"Greenhouse effect\">greenhouse effect<\/a>, \"a large scale global warming, with radical climate changes may result\". Both phrases were used only occasionally until 1975, when <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wallace_Smith_Broecker\" title=\"Wallace Smith Broecker\">Wallace Smith Broecker<\/a> published a scientific paper on the topic, \"Climatic Change: Are We on the Brink of a Pronounced Global Warming?\" The phrase began to come into common use, and in 1976 <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mikhail_Budyko\" title=\"Mikhail Budyko\">Mikhail Budyko<\/a>'s statement that \"a global warming up has started\" was widely reported.<sup id=\"cite_ref-Weart_27_280-1\" class=\"reference\"><a href=\"#cite_note-Weart_27-280\">[276]<\/a><\/sup> Other studies, such as a 1971 <a href=\"https:\/\/en.wikipedia.org\/wiki\/Massachusetts_Institute_of_Technology\" title=\"Massachusetts Institute of Technology\">MIT<\/a> report, referred to the human impact as \"inadvertent climate modification\", but an influential 1979 <a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Academy_of_Sciences\" title=\"National Academy of Sciences\">National Academy of Sciences<\/a> study headed by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Jule_Gregory_Charney\" title=\"Jule Gregory Charney\">Jule Charney<\/a> followed Broecker in using <i>global warming<\/i> for rising surface temperatures, while describing the wider effects of increased CO<sub>2<\/sub> as <i>climate change<\/i>.<sup id=\"cite_ref-nasacbaon_282-0\" class=\"reference\"><a href=\"#cite_note-nasacbaon-282\">[278]<\/a><\/sup><\/p><p>In 1986 and November 1987, NASA climate scientist <a href=\"https:\/\/en.wikipedia.org\/wiki\/James_Hansen\" title=\"James Hansen\">James Hansen<\/a> gave testimony to Congress on global warming. There were increasing heatwaves and drought problems in the summer of 1988, and when Hansen testified in the Senate on 23 June he sparked worldwide interest.<sup id=\"cite_ref-Weart_Hansen_281-1\" class=\"reference\"><a href=\"#cite_note-Weart_Hansen-281\">[277]<\/a><\/sup> He said, \"global warming has reached a level such that we can ascribe with a high degree of confidence a cause and effect relationship between the greenhouse effect and the observed warming.\"<sup id=\"cite_ref-283\" class=\"reference\"><a href=\"#cite_note-283\">[279]<\/a><\/sup> Public attention increased over the summer, and <i>global warming<\/i> became the dominant popular term, commonly used both by the press and in public discourse.<sup id=\"cite_ref-nasacbaon_282-1\" class=\"reference\"><a href=\"#cite_note-nasacbaon-282\">[278]<\/a><\/sup><\/p><p>In a 2008 NASA article on usage, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Erik_M._Conway\" title=\"Erik M. Conway\">Erik M. Conway<\/a> defined <i>global warming<\/i> as \"the increase in Earth's average surface temperature due to rising levels of greenhouse gases\", while <i>climate change<\/i> was \"a long-term change in the Earth's climate, or of a region on Earth\". Because effects such as changing patterns of rainfall and rising sea levels would probably have more impact than temperatures alone, he considered <i>global climate change<\/i> a more scientifically accurate term, and like the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">Intergovernmental Panel on Climate Change<\/a>, the NASA website emphasized this wider context.<sup id=\"cite_ref-nasacbaon_282-2\" class=\"reference\"><a href=\"#cite_note-nasacbaon-282\">[278]<\/a><\/sup><\/p>\n<h2>See also<\/h2>\n\n<div class=\"div-col columns column-width\">\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Anthropocene\" title=\"Anthropocene\">Anthropocene<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_agriculture\" title=\"Climate change and agriculture\">Climate change and agriculture<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_the_Arctic\" title=\"Climate change in the Arctic\">Climate change in the Arctic<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_the_coal_industry\" title=\"Environmental impact of the coal industry\">Environmental impact of the coal industry<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Geologic_temperature_record\" title=\"Geologic temperature record\">Geologic temperature record<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_cooling\" title=\"Global cooling\">Global cooling<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Glossary_of_climate_change\" title=\"Glossary of climate change\">Glossary of climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas_emissions_accounting\" title=\"Greenhouse gas emissions accounting\">Greenhouse gas emissions accounting<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_countries_by_carbon_dioxide_emissions\" title=\"List of countries by carbon dioxide emissions\">List of countries by carbon dioxide emissions<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Holocene_extinction\" title=\"Holocene extinction\">Holocene extinction<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Index_of_climate_change_articles\" title=\"Index of climate change articles\">Index of climate change articles<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Planetary_boundaries\" title=\"Planetary boundaries\">Planetary boundaries<\/a>, climate change\/global warming is one of them<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_opinion_on_climate_change\" title=\"Scientific opinion on climate change\">Scientific opinion on climate change<\/a><\/li><\/ul><\/div>\n<h2>Notes<\/h2>\n<div class=\"reflist\">\n<div class=\"mw-references-wrap\"><ol class=\"references\"><li id=\"cite_note-13\"><b><a href=\"#cite_ref-13\">^<\/a><\/b> The 2001 joint statement was signed by the national academies of science of Australia, Belgium, Brazil, Canada, the Caribbean, the People's Republic of China, France, Germany, India, Indonesia, Ireland, Italy, Malaysia, New Zealand, Sweden, and the UK.<sup id=\"cite_ref-12\" class=\"reference\"><a href=\"#cite_note-12\">[12]<\/a><\/sup> The 2005 statement added Japan, Russia, and the US. The 2007 statement added Mexico and South Africa. The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Network_of_African_Science_Academies\" title=\"Network of African Science Academies\">Network of African Science Academies<\/a> and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Polish_Academy_of_Sciences\" title=\"Polish Academy of Sciences\">Polish Academy of Sciences<\/a> have issued separate statements. Professional scientific societies include <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Astronomical_Society\" title=\"American Astronomical Society\">American Astronomical Society<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Chemical_Society\" title=\"American Chemical Society\">American Chemical Society<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Geophysical_Union\" title=\"American Geophysical Union\">American Geophysical Union<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Institute_of_Physics\" title=\"American Institute of Physics\">American Institute of Physics<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Meteorological_Society\" title=\"American Meteorological Society\">American Meteorological Society<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Physical_Society\" title=\"American Physical Society\">American Physical Society<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/American_Quaternary_Association\" title=\"American Quaternary Association\">American Quaternary Association<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Australian_Meteorological_and_Oceanographic_Society\" title=\"Australian Meteorological and Oceanographic Society\">Australian Meteorological and Oceanographic Society<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Canadian_Foundation_for_Climate_and_Atmospheric_Sciences\" title=\"Canadian Foundation for Climate and Atmospheric Sciences\">Canadian Foundation for Climate and Atmospheric Sciences<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Canadian_Meteorological_and_Oceanographic_Society\" title=\"Canadian Meteorological and Oceanographic Society\">Canadian Meteorological and Oceanographic Society<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/European_Academy_of_Sciences_and_Arts\" title=\"European Academy of Sciences and Arts\">European Academy of Sciences and Arts<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/European_Geosciences_Union\" title=\"European Geosciences Union\">European Geosciences Union<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/European_Science_Foundation\" title=\"European Science Foundation\">European Science Foundation<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Geological_Society_of_America\" title=\"Geological Society of America\">Geological Society of America<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Geological_Society_of_Australia\" title=\"Geological Society of Australia\">Geological Society of Australia<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Geological_Society_of_London\" title=\"Geological Society of London\">Geological Society of London<\/a>-Stratigraphy Commission, <a href=\"https:\/\/en.wikipedia.org\/wiki\/InterAcademy_Council\" class=\"mw-redirect\" title=\"InterAcademy Council\">InterAcademy Council<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Union_of_Geodesy_and_Geophysics\" title=\"International Union of Geodesy and Geophysics\">International Union of Geodesy and Geophysics<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Union_for_Quaternary_Research\" title=\"International Union for Quaternary Research\">International Union for Quaternary Research<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nagt.org\/index.html\">National Association of Geoscience Teachers<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/United_States_National_Research_Council\" class=\"mw-redirect\" title=\"United States National Research Council\">National Research Council (US)<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Royal_Meteorological_Society\" title=\"Royal Meteorological Society\">Royal Meteorological Society<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/World_Meteorological_Organization\" title=\"World Meteorological Organization\">World Meteorological Organization<\/a>.\n<\/li>\n<li id=\"cite_note-27\"><b><a href=\"#cite_ref-27\">^<\/a><\/b> Earth has already experienced almost 1\/2 of the 2.0\u00a0\u00b0C (3.6\u00a0\u00b0F) described in the Canc\u00fan Agreement. In the last 100 years, Earth's average surface temperature increased by about 0.8\u00a0\u00b0C (1.4\u00a0\u00b0F) with about two thirds of the increase occurring over just the last three decades.<sup id=\"cite_ref-AmericasClimateChoices-2011-FullReport_26-0\" class=\"reference\"><a href=\"#cite_note-AmericasClimateChoices-2011-FullReport-26\">[25]<\/a><\/sup><\/li>\n<li id=\"cite_note-35\"><b><a href=\"#cite_ref-35\">^<\/a><\/b> Scientific journals use \"global warming\" to describe an increasing global average temperature just at earth's surface, and most of these authorities further limit \"global warming\" to such increases caused by human activities or increasing greenhouse gases.\n<\/li>\n<li id=\"cite_note-73\"><b><a href=\"#cite_ref-73\">^<\/a><\/b> The greenhouse effect produces an average worldwide temperature <i>increase<\/i> of about 33\u00a0\u00b0C (59\u00a0\u00b0F) compared to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Black_body\" title=\"Black body\">black body<\/a> predictions without the greenhouse effect, not an average <i>surface temperature<\/i> of 33\u00a0\u00b0C (91\u00a0\u00b0F). The average worldwide surface temperature is about 14\u00a0\u00b0C (57\u00a0\u00b0F).\n<\/li>\n<\/ol><\/div><\/div>\n<h2>Citations<\/h2>\n<div class=\"reflist columns references-column-width\">\n<ol class=\"references\"><li id=\"cite_note-landocean-1\"><b><a href=\"#cite_ref-landocean_1-0\">^<\/a><\/b> <a rel=\"nofollow\" class=\"external free\" href=\"http:\/\/berkeleyearth.lbl.gov\/auto\/Global\/Land_and_Ocean_summary.txt\">http:\/\/berkeleyearth.lbl.gov\/auto\/Global\/Land_and_Ocean_summary.txt<\/a> Berkeley Earth \u2013 Land and Ocean Summary\n<\/li>\n<li id=\"cite_note-AR5_WG1_SPM_p4-2\">^ <a href=\"#cite_ref-AR5_WG1_SPM_p4_2-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-AR5_WG1_SPM_p4_2-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation report\">Stocker, T.F.; Qin, D.; Plattner, G.-K.; Tignor, M.; Allen, S.K.; Boschung, J.; Nauels, A.; Xia, Y.; Bex, V.; Midgley, P.M. (2013). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/wg1\/WG1AR5_SPM_FINAL.pdf\">\"The Physical Science Basis \u2013 Summary for Policymakers\"<\/a> (PDF). IPCC WGI AR5 (Report). p.\u00a04. <q>Warming of the climate system is unequivocal, and since the 1950s, many of the observed changes are unprecedented over decades to millennia. The atmosphere and ocean have warmed, the amounts of snow and ice have diminished, sea level has risen, and the concentrations of greenhouse gases have increased.<\/q><\/cite>\n<\/li>\n<li id=\"cite_note-EPA_myths-3\"><b><a href=\"#cite_ref-EPA_myths_3-0\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.epa.gov\/ghgemissions\/myths-vs-facts-denial-petitions-reconsideration-endangerment-and-cause-or-contribute\">\"Myths vs. Facts: Denial of Petitions for Reconsideration of the Endangerment and Cause or Contribute Findings for Greenhouse Gases under Section 202(a) of the Clean Air Act\"<\/a>. U.S. Environmental Protection Agency. Retrieved 7 August 2017. <q>The U.S. Global Change Research Program, the National Academy of Sciences, and the Intergovernmental Panel on Climate Change (IPCC) have each independently concluded that warming of the climate system in recent decades is \"unequivocal\". This conclusion is not drawn from any one source of data but is based on multiple lines of evidence, including three worldwide temperature datasets showing nearly identical warming trends as well as numerous other independent indicators of global warming (e.g., rising sea levels, shrinking Arctic sea ice).<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-AR5_Glossary_GW-4\"><b><a href=\"#cite_ref-AR5_Glossary_GW_4-0\">^<\/a><\/b> <cite class=\"citation report\">Mach, Katharine J.; Serge, Planton; von Stechow, Christoph (2014). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/syr\/AR5_SYR_FINAL_Annexes.pdf\">\"Annex II Glossary\"<\/a> (PDF). IPCC SYR AR5 (Report). p.\u00a0124. <q>Global warming refers to the gradual increase, observed or projected, in global surface temperature, as one of the consequences of radiative forcing caused by anthropogenic emissions. {WGIII}<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-AR5_WG1_Ch_5-5\"><b><a href=\"#cite_ref-AR5_WG1_Ch_5_5-0\">^<\/a><\/b> <cite class=\"citation report\">Masson-Delmotte, Val\u00e9rie; Schulz, Michael (2013). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/wg1\/WG1AR5_Chapter05_FINAL.pdf\">\"5: Information from Paleoclimate Archives\"<\/a> (PDF). IPCC WGI AR5 (Report). pp.\u00a0389, 399\u2013400. <q>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Paleocene\u2013Eocene_Thermal_Maximum\" title=\"Paleocene\u2013Eocene Thermal Maximum\">PETM<\/a> [around 55.5\u201355.3 million years ago] was marked by ... global warming of 4\u00b0C to 7\u00b0C .....  <a href=\"https:\/\/en.wikipedia.org\/wiki\/Deglaciation\" title=\"Deglaciation\">Deglacial<\/a> global warming occurred in two main steps from 17.5 to 14.5 ka [thousand years ago] and 13.0 to 10.0 ka.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-Shaftel_2016-6\"><b><a href=\"#cite_ref-Shaftel_2016_6-0\">^<\/a><\/b> <cite class=\"citation web\">Shaftel, Holly (January 2016). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/climate.nasa.gov\/resources\/global-warming\">\"What's in a name? Weather, global warming and climate change\"<\/a>. <i>NASA Climate Change: Vital Signs of the Planet<\/i>. Retrieved 12 October 2018. <q>'Climate change' and 'global warming' are often used interchangeably but have distinct meanings. .... Global warming refers to the upward temperature trend across the entire Earth since the early 20th century .... Climate change refers to a broad range of global phenomena ...[which] include the increased temperature trends described by global warming<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-NOAA_diff_GW_&amp;_CC-7\"><b><a href=\"#cite_ref-NOAA_diff_GW_&amp;_CC_7-0\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.climate.gov\/news-features\/climate-qa\/whats-difference-between-global-warming-and-climate-change\">\"What's the difference between global warming and climate change?\"<\/a>. <i>NOAA Climate.gov<\/i>. 17 June 2015. Retrieved 15 October 2018. <q><i>Global warming<\/i> refers only to the Earth\u2019s rising surface temperature, while <i>climate change<\/i> includes warming and the 'side effects' of warming\u2014like melting glaciers, heavier rainstorms, or more frequent drought. Said another way, global warming is one symptom of the much larger problem of human-caused climate change.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-AR5_Glossary_CC-8\"><b><a href=\"#cite_ref-AR5_Glossary_CC_8-0\">^<\/a><\/b> <cite class=\"citation report\">Mach, Katharine J.; Serge, Planton; von Stechow, Christoph (2014). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/syr\/AR5_SYR_FINAL_Annexes.pdf\">\"Annex II Glossary\"<\/a> (PDF). IPCC SYR AR5 (Report). p.\u00a0120. <q>Climate change refers to a change in the state of the climate that can be identified (e.g., by using statistical tests) by changes in the mean and\/or the variability of its properties and that persists for an extended period, typically decades or longer. Climate change may be due to natural internal processes or external forcings such as modulations of the solar cycles, volcanic eruptions and persistent anthropogenic changes in the composition of the atmosphere or in land use. .... {WGI, II, III}<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-9\"><b><a href=\"#cite_ref-9\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/wg1\/WG1AR5_SPM_FINAL.pdf\">\"IPCC, Climate Change 2013: The Physical Science Basis \u2013 Summary for Policymakers (AR5 WG1)\"<\/a> (PDF). Intergovernmental Panel on Climate Change. p.\u00a017. <q>It is extremely likely that human influence has been the dominant cause of the observed warming since the mid-20th century.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-10\"><b><a href=\"#cite_ref-10\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/wg1\/WG1AR5_TS_FINAL.pdf\">\"IPCC, Climate Change 2013: The Physical Science Basis -Technical Summary\"<\/a> (PDF). Intergovernmental Panel on Climate Change. pp.\u00a089\u201390.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-11\"><b><a href=\"#cite_ref-11\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/nationalacademies.org\/onpi\/06072005.pdf\">\"Joint Science Academies' Statement\"<\/a> (PDF). Retrieved 6 January 2014.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-12\"><b><a href=\"#cite_ref-12\">^<\/a><\/b> <cite class=\"citation news\">Kirby, Alex (17 May 2001). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/news.bbc.co.uk\/1\/hi\/sci\/tech\/1335872.stm\">\"Science academies back Kyoto\"<\/a>. BBC News. Retrieved 27 July 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-14\"><b><a href=\"#cite_ref-14\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/climate.nasa.gov\/scientific-consensus\/\">\"Scientific consensus: Earth's climate is warming\"<\/a>. <i>Climate Change: Vital Signs of the Planet<\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/NASA\" title=\"NASA\">NASA<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20180628042920\/https:\/\/climate.nasa.gov\/scientific-consensus\/\">Archived<\/a> from the original on 28 June 2018. Retrieved 7 August 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-15\"><b><a href=\"#cite_ref-15\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20170807111205\/https:\/\/www.opr.ca.gov\/s_listoforganizations.php\">\"List of Organizations\"<\/a>. The Governor's Office of Planning &amp; Research, State of California. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.opr.ca.gov\/s_listoforganizations.php\">the original<\/a> on 7 August 2017. Retrieved 7 August 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-16\"><b><a href=\"#cite_ref-16\">^<\/a><\/b> <cite class=\"citation web\">Field, Christopher B.; Barros, Vicente R.; Mach, Katharine J.; Mastrandrea, Michael D.;  et al. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/wg2\/WGIIAR5-TS_FINAL.pdf\">\"IPCC, Climate Change 2014: Impacts, Adaptation, and Vulnerability \u2013 Technical Summary\"<\/a> (PDF). Intergovernmental Panel on Climate Change. pp.\u00a044\u201346.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-17\"><b><a href=\"#cite_ref-17\">^<\/a><\/b> Solomon <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ts.html\">Technical Summary<\/a>,\n<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/tssts-5-3.html\">Section TS.5.3: Regional-Scale Projections<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>.\n<\/li>\n<li id=\"cite_note-Zeng_L17401-18\">^ <a href=\"#cite_ref-Zeng_L17401_18-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-Zeng_L17401_18-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation journal\">Zeng, Ning; Yoon, Jinho (1 September 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/onlinelibrary.wiley.com\/doi\/10.1029\/2009GL039699\/abstract\">\"Expansion of the world's deserts due to vegetation-albedo feedback under global warming\"<\/a>. <i>Geophysical Research Letters<\/i>. <b>36<\/b> (17): L17401. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2009GeoRL..3617401Z\">2009GeoRL..3617401Z<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2009GL039699\">10.1029\/2009GL039699<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Serial_Number\" title=\"International Standard Serial Number\">ISSN<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.worldcat.org\/issn\/1944-8007\">1944-8007<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-19\"><b><a href=\"#cite_ref-19\">^<\/a><\/b> On snowfall:\n<ul><li><cite class=\"citation news\">Christopher Joyce (15 February 2010). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.npr.org\/templates\/story\/story.php?storyId=123671588\">\"Get This: Warming Planet Can Mean More Snow\"<\/a>. NPR.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/phys.org\/news\/2011-03-global-snowstorms-scientists.html\">\"Global warming means more snowstorms: scientists\"<\/a>. 1 March 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.skepticalscience.com\/Record-snowfall-disproves-global-warming.htm\">\"Does record snowfall disprove global warming?\"<\/a>. 9 July 2010. Retrieved 14 December 2014.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li><\/ul><\/li>\n<li id=\"cite_note-20\"><b><a href=\"#cite_ref-20\">^<\/a><\/b> <cite class=\"citation journal\">Battisti, David S.; Naylor, Rosamond L. (9 January 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/science.sciencemag.org\/content\/323\/5911\/240\">\"Historical Warnings of Future Food Insecurity with Unprecedented Seasonal Heat\"<\/a>. <i>Science<\/i>. <b>323<\/b> (5911): 240\u201344. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1164363\">10.1126\/science.1164363<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Serial_Number\" title=\"International Standard Serial Number\">ISSN<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.worldcat.org\/issn\/0036-8075\">0036-8075<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/19131626\">19131626<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-21\"><b><a href=\"#cite_ref-21\">^<\/a><\/b> <a href=\"#CITEREFUS_NRC2012\">US NRC 2012<\/a>, p.\u00a026\n<\/li>\n<li id=\"cite_note-22\"><b><a href=\"#cite_ref-22\">^<\/a><\/b> <cite class=\"citation journal\">Clark, Peter U. (8 February 2016). \"Consequences of twenty-first-century policy for multi-millennial climate and sea-level change\". <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Nature_Climate_Change\" title=\"Nature Climate Change\">Nature Climate Change<\/a><\/i>. <b>6<\/b> (4): 360\u201369. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2016NatCC...6..360C\">2016NatCC...6..360C<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2FNCLIMATE2923\">10.1038\/NCLIMATE2923<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-23\"><b><a href=\"#cite_ref-23\">^<\/a><\/b> <cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/essential_background\/convention\/status_of_ratification\/items\/2631.php\">\"Status of Ratification of the Convention\"<\/a>. UNFCCC Secretariat: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bonn\" title=\"Bonn\">Bonn<\/a>, Germany: United Nations Framework Convention on Climate Change. 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Most countries in the world are Parties to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/United_Nations_Framework_Convention_on_Climate_Change\" title=\"United Nations Framework Convention on Climate Change\">United Nations Framework Convention on Climate Change<\/a> (UNFCCC), which has adopted the 7002275149999999999\u26602\u00a0\u00b0C limit. As of 25 November 2011, there are 195 parties (194 states and 1 regional economic integration organization (the <a href=\"https:\/\/en.wikipedia.org\/wiki\/European_Union\" title=\"European Union\">European Union<\/a>)) to the UNFCCC.\n<\/li>\n<li id=\"cite_note-24\"><b><a href=\"#cite_ref-24\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140108192827\/http:\/\/unfccc.int\/essential_background\/convention\/items\/6036.php\">\"First steps to a safer future: Introducing The United Nations Framework Convention on Climate Change\"<\/a>. United Nations Framework Convention on Climate Change. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/essential_background\/convention\/items\/6036.php\">the original<\/a> on 8 January 2014. Retrieved 7 August 2017. <q>Preventing \"dangerous\" human interference with the climate system is the ultimate aim of the UNFCCC.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-unfccc_2_degrees_celsius_target-25\"><b><a href=\"#cite_ref-unfccc_2_degrees_celsius_target_25-0\">^<\/a><\/b> \n<cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/resource\/docs\/2010\/cop16\/eng\/07a01.pdf#page=2\">\"Conference of the Parties\u00a0\u2013 Sixteenth Session: Decision 1\/CP.16: The Cancun Agreements: Outcome of the work of the Ad Hoc Working Group on Long-term Cooperative Action under the Convention (English): Paragraph 4\"<\/a> (PDF). UNFCCC Secretariat: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bonn\" title=\"Bonn\">Bonn<\/a>, Germany: United Nations Framework Convention on Climate Change. 2011: 3.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> \"(...) deep cuts in global greenhouse gas emissions are required according to science, and as documented in the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, with a view to reducing global greenhouse gas emissions so as to hold the increase in global average temperature below 7002275149999999999\u26602\u00a0\u00b0C above preindustrial levels\"\n<\/li>\n<li id=\"cite_note-AmericasClimateChoices-2011-FullReport-26\"><b><a href=\"#cite_ref-AmericasClimateChoices-2011-FullReport_26-0\">^<\/a><\/b> \n<cite class=\"citation book\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nap.edu\/openbook.php?record_id=12781&amp;page=1\"><i>America's Climate Choices<\/i><\/a>. Washington, DC: The National Academies Press. 2011. p.\u00a015. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-309-14585-5\" title=\"Special:BookSources\/978-0-309-14585-5\">978-0-309-14585-5<\/a>. <q>The average temperature of the Earth's surface increased by about 1.4\u00a0\u00b0F (0.8\u00a0\u00b0C) over the past 100 years, with about 1.0\u00a0\u00b0F (0.6\u00a0\u00b0C) of this warming occurring over just the past three decades.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-28\"><b><a href=\"#cite_ref-28\">^<\/a><\/b> \n<ul><li><cite class=\"citation news\">Sutter, John D.; Berlinger, Joshua (12 December 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.cnn.com\/2015\/12\/12\/world\/global-climate-change-conference-vote\/\">\"Final draft of climate deal formally accepted in Paris\"<\/a>. CNN. Retrieved 12 December 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite class=\"citation news\">Vaughan, A. (12 December 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.theguardian.com\/environment\/2015\/dec\/12\/paris-climate-deal-key-points\">\"Paris climate deal: key points at a glance\"<\/a>. <i>The Guardian<\/i>. London and Manchester. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20151213005658\/https:\/\/www.theguardian.com\/environment\/2015\/dec\/12\/paris-climate-deal-key-points\">Archived<\/a> from the original on 13 December 2015. Retrieved 12 December 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Archived .<\/li><\/ul><\/li>\n<li id=\"cite_note-Hansen_2013-29\"><b><a href=\"#cite_ref-Hansen_2013_29-0\">^<\/a><\/b> <cite class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/James_Hansen\" title=\"James Hansen\">James Hansen<\/a>; Makiko Sato; Gary Russell; Pushker Kharecha (September 2013). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/rsta.royalsocietypublishing.org\/content\/371\/2001\/20120294\">\"Climate sensitivity, sea level and atmospheric carbon dioxide\"<\/a>. <i>Royal Society Publishing<\/i>. <b>371<\/b> (2001): 20120294. <a href=\"https:\/\/en.wikipedia.org\/wiki\/ArXiv\" title=\"ArXiv\">arXiv<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/arxiv.org\/abs\/1211.4846\">1211.4846<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2013RSPTA.37120294H\">2013RSPTA.37120294H<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1098%2Frsta.2012.0294\">10.1098\/rsta.2012.0294<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC3785813\">3785813<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-Steffen2018-30\"><b><a href=\"#cite_ref-Steffen2018_30-0\">^<\/a><\/b> <cite class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Will_Steffen\" title=\"Will Steffen\">Steffen<\/a> et al. (2018). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pnas.org\/content\/early\/2018\/07\/31\/1810141115\">\"Trajectories of the Earth System in the Anthropocene\"<\/a>. <i>PNAS<\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.1810141115\">10.1073\/pnas.1810141115<\/a>.<\/cite>CS1 maint: Uses authors parameter (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:CS1_maint:_Uses_authors_parameter\" title=\"Category:CS1 maint: Uses authors parameter\">link<\/a>) CS1 maint: Explicit use of et al. (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:CS1_maint:_Explicit_use_of_et_al.\" title=\"Category:CS1 maint: Explicit use of et al.\">link<\/a>) <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-31\"><b><a href=\"#cite_ref-31\">^<\/a><\/b> <cite class=\"citation web\">Stokes, Bruce; Wike, Richard; Carle, Jill (5 November 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pewglobal.org\/2015\/11\/05\/global-concern-about-climate-change-broad-support-for-limiting-emissions\/\">\"Global Concern about Climate Change, Broad Support for Limiting Emissions\"<\/a>. Pew Research Center's Global Attitudes Project. Retrieved 7 August 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-32\"><b><a href=\"#cite_ref-32\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatechange2013.org\/images\/report\/WG1AR5_ALL_FINAL.pdf\">\"Climate Change 2013: The Physical Science Basis, IPCC Fifth Assessment Report (WGI AR5)\"<\/a> (PDF). <i>WGI AR5<\/i>. IPCC AR5. 2013. p.\u00a05.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-33\"><b><a href=\"#cite_ref-33\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch3s3-es.html\">\"Climate Change 2007: Working Group I: The Physical Science Basis\"<\/a>. IPCC AR4. 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-34\"><b><a href=\"#cite_ref-34\">^<\/a><\/b> <cite class=\"citation report\">Rhein, M.; Rintoul, S.R. (2013). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatechange2013.org\/images\/report\/WG1AR5_Chapter03_FINAL.pdf\">\"3: Observations: Ocean\"<\/a> (PDF). IPCC WGI AR5 (Report). p.\u00a0257. <q>Ocean warming dominates the global energy change inventory. Warming of the ocean accounts for about 93% of the increase in the Earth's energy inventory between 1971 and 2010 (high confidence), with warming of the upper (0 to 700 m) ocean accounting for about 64% of the total. Melting ice (including Arctic sea ice, ice sheets and glaciers) and warming of the continents and atmosphere account for the remainder of the change in energy.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-:1-36\"><b><a href=\"#cite_ref-:1_36-0\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/vortex.nsstc.uah.edu\/data\/msu\/v6.0\/tlt\/uahncdc_lt_6.0.txt\">\"UAH v6.0 TLT data\"<\/a> (trend data at bottom of file). National Space Science and Technology Center. Retrieved 3 February 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-:0-37\"><b><a href=\"#cite_ref-:0_37-0\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.remss.com\/measurements\/upper-air-temperature\">\"Upper Air Temperature: Decadal Trends\"<\/a>. Remote Sensing Systems. Retrieved 3 February 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-38\"><b><a href=\"#cite_ref-38\">^<\/a><\/b> Jansen <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch6.html\">Ch. 6, Palaeoclimate<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch6s6-6.html#6-6-1\">Section 6.6.1.1: What Do Reconstructions Based on Palaeoclimatic Proxies Show?<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc-wg1.unibe.ch\/publications\/wg1-ar4\/ar4-wg1-chapter6.pdf\">pp. 466\u201378<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20100524082417\/http:\/\/www.ipcc-wg1.unibe.ch\/publications\/wg1-ar4\/ar4-wg1-chapter6.pdf\">Archived<\/a> 24 May 2010 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>., in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>.\n<\/li>\n<li id=\"cite_note-kennedy_evidence_for_global_warming-39\">^ <a href=\"#cite_ref-kennedy_evidence_for_global_warming_39-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-kennedy_evidence_for_global_warming_39-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-kennedy_evidence_for_global_warming_39-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-kennedy_evidence_for_global_warming_39-3\"><sup><i><b>d<\/b><\/i><\/sup><\/a> <cite id=\"CITEREFKennedy2010\" class=\"citation journal\">Kennedy, J.J.;  et al. (2010). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ncdc.noaa.gov\/bams-state-of-the-climate\/2009.php\">\"How do we know the world has warmed? in: 2. Global Climate, in: State of the Climate in 2009\"<\/a>. <i>Bull. Amer. Meteor. Soc<\/i>. <b>91<\/b> (7): 26.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-40\"><b><a href=\"#cite_ref-40\">^<\/a><\/b> \n<cite id=\"CITEREFKennedy,_C.2012\" class=\"citation journal\">Kennedy, C. (10 July 2012). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatewatch.noaa.gov\/article\/2012\/2011-state-of-the-climate-sea-level\">\"ClimateWatch Magazine &gt;&gt; State of the Climate: 2011 Global Sea Level\"<\/a>. NOAA Climate Services Portal.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-41\"><b><a href=\"#cite_ref-41\">^<\/a><\/b> \n<cite class=\"citation book\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/spm.html\">\"Summary for Policymakers\"<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/spmsspm-direct-observations.html\"><i>Direct Observations of Recent Climate Change<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>\n<\/li>\n<li id=\"cite_note-42\"><b><a href=\"#cite_ref-42\">^<\/a><\/b> \n<cite class=\"citation book\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/spm.html\">\"Summary for Policymakers\"<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/spmsspm-b.html\"><i>B. Current knowledge about observed impacts of climate change on the natural and human environment<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_AR4_WG22007\">IPCC AR4 WG2 2007<\/a>\n<\/li>\n<li id=\"cite_note-43\"><b><a href=\"#cite_ref-43\">^<\/a><\/b> \n<cite id=\"CITEREFRosenzweig,_C.\" class=\"citation book\">Rosenzweig, C.;  et al. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch1.html\">\"Ch 1: Assessment of Observed Changes and Responses in Natural and Managed Systems\"<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch1s1-3-5-1.html\"><i>Sec 1.3.5.1 Changes in phenology<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_AR4_WG22007\">IPCC AR4 WG2 2007<\/a>, p.\u00a099\n<\/li>\n<li id=\"cite_note-WarmingHole-SE-USA-44\"><b><a href=\"#cite_ref-WarmingHole-SE-USA_44-0\">^<\/a><\/b> <cite class=\"citation news\">Leavenworth, Stuart (15 February 2018). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.mcclatchydc.com\/news\/nation-world\/national\/article200169249.html\">\"Snow-covered beaches? Chilly iguanas? They are part of a mysterious 'hole' in global warming\"<\/a>. <i>McClatchy Washington Bureau<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-45\"><b><a href=\"#cite_ref-45\">^<\/a><\/b> Trenberth <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch3.html\">Chap 3, Observations: Atmospheric Surface and Climate Change<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch3s3-es.html\">Executive Summary<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar4\/wg1\/ar4-wg1-chapter3.pdf\">p. 237<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>.\n<\/li>\n<li id=\"cite_note-46\"><b><a href=\"#cite_ref-46\">^<\/a><\/b> \n<cite id=\"CITEREFSuttonDongGregory2007\" class=\"citation journal\">Sutton, Rowan T.; Dong, Buwen; Gregory, Jonathan M. (16 January 2007). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.agu.org\/pubs\/crossref\/2007\/2006GL028164.shtml\">\"Land\/sea warming ratio in response to climate change: IPCC AR4 model results and comparison with observations\"<\/a>. <i>Geophysical Research Letters<\/i>. <b>34<\/b> (2): L02701. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2007GeoRL..3402701S\">2007GeoRL..3402701S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2006GL028164\">10.1029\/2006GL028164<\/a>. Retrieved 19 September 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-47\"><b><a href=\"#cite_ref-47\">^<\/a><\/b> <cite id=\"CITEREFFeulnerRahmstorf,_StefanLevermann,_AndersVolkwardt,_Silvia2013\" class=\"citation journal\">Feulner, Georg; Rahmstorf, Stefan; Levermann, Anders; Volkwardt, Silvia (March 2013). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/journals.ametsoc.org\/doi\/abs\/10.1175\/JCLI-D-12-00636.1\">\"On the Origin of the Surface Air Temperature Difference Between the Hemispheres in Earth's Present-Day Climate\"<\/a>. <i>Journal of Climate<\/i>. <b>26<\/b>: 130325101629005. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2013JCli...26.7136F\">2013JCli...26.7136F<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1175%2FJCLI-D-12-00636.1\">10.1175\/JCLI-D-12-00636.1<\/a>. Retrieved 25 April 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-48\"><b><a href=\"#cite_ref-48\">^<\/a><\/b> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/tssts-3-1-2.html\">TS.3.1.2 Spatial Distribution of Changes in Temperature, Circulation and Related Variables \u2013 AR4 WGI Technical Summary<\/a>\n<\/li>\n<li id=\"cite_note-49\"><b><a href=\"#cite_ref-49\">^<\/a><\/b> Ehhalt <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/127.htm\">Chapter 4: Atmospheric Chemistry and Greenhouse Gases<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20120123063256\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/127.htm\">Archived<\/a> 23 January 2012 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/139.htm#4231\">Section 4.2.3.1: Carbon monoxide (CO) and hydrogen (H<sub>2<\/sub>)<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20120409122116\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/139.htm\">Archived<\/a> 9 April 2012 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/pdf\/TAR-04.pdf\">p. 256<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20120117054647\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/pdf\/TAR-04.pdf\">Archived<\/a> 17 January 2012 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>. in <a href=\"#CITEREFIPCC_TAR_WG12001\">IPCC TAR WG1 2001<\/a>.\n<\/li>\n<li id=\"cite_note-50\"><b><a href=\"#cite_ref-50\">^<\/a><\/b> <cite id=\"CITEREFMeehlWashington,_Warren_M.Collins,_William_D.Arblaster,_Julie_M.2005\" class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gerald_Meehl\" title=\"Gerald Meehl\">Meehl, Gerald A.<\/a>; Washington, Warren M.; Collins, William D.; Arblaster, Julie M.; Hu, Aixue; Buja, Lawrence E.; Strand, Warren G.; Teng, Haiyan (18 March 2005). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/science.sciencemag.org\/content\/307\/5716\/1769\">\"How Much More Global Warming and Sea Level Rise\"<\/a> (PDF). <i>Science<\/i>. <b>307<\/b> (5716): 1769\u201372. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2005Sci...307.1769M\">2005Sci...307.1769M<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1106663\">10.1126\/science.1106663<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/15774757\">15774757<\/a>. Retrieved 11 February 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-51\"><b><a href=\"#cite_ref-51\">^<\/a><\/b> <cite class=\"citation web\">Wigley, T. M. L.;  et al. (2005). <a rel=\"nofollow\" class=\"external text\" href=\"ftp:\/\/ftp.soest.hawaii.edu\/coastal\/Climate%20Articles\/Wigley_2005%20Sea%20level%20commitment.pdf\">\"The Climate Change Commitment\"<\/a> (PDF). School of Ocean and Earth Science and Technology, University of Hawaii at Manoa. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2005Sci...307.1766W\">2005Sci...307.1766W<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1103934\">10.1126\/science.1103934<\/a>. <q>Even if atmospheric composition were fixed today, global mean temperature and sea level rise would continue due to oceanic thermal inertia. These constant-composition (CC) commitments and their uncertainties are quantified. Constant-emissions (CE) commitments are also considered. The CC warming commitment could exceed 1C. The CE warming commitment is 2 to 6C by the year 2400.\" (...) \"A breakdown of the natural and anthropogenic components of the CC commitment, together with uncertainties arising from ocean mixing (Kz) uncertainties, is given in table S1. Past natural forcing (inclusion of which is the default case here) has a marked effect. The natural forcing component is surprisingly large, 64% of the total commitment in 2050, reducing to 52% by 2400.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-52\"><b><a href=\"#cite_ref-52\">^<\/a><\/b> <cite class=\"citation journal\">Florian S\u00e9vellec &amp; Sybren S. Drijfhout (2018). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nature.com\/articles\/s41467-018-05442-8\">\"A novel probabilistic forecast system predicting anomalously warm 2018-2022 reinforcing the long-term global warming trend\"<\/a>. <i>Nature Communications<\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fs41467-018-05442-8\">10.1038\/s41467-018-05442-8<\/a>.<\/cite>CS1 maint: Uses authors parameter (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:CS1_maint:_Uses_authors_parameter\" title=\"Category:CS1 maint: Uses authors parameter\">link<\/a>) <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-53\"><b><a href=\"#cite_ref-53\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.washingtonpost.com\/energy-environment\/2018\/08\/14\/next-five-years-will-be-anomalously-warm-scientists-predict\">\"The next five years will be 'anomalously warm,' scientists predict\"<\/a>. <i>The Washington Post<\/i>. 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-54\"><b><a href=\"#cite_ref-54\">^<\/a><\/b> <cite class=\"citation journal\">England, Matthew H.; McGregor, Shayne; Spence, Paul; Meehl, Gerald A.; Timmermann, Axel; Cai, Wenju; Sen Gupta, Alex; McPhaden, Michael J.; Purich, Ariaan; Santoso, Agus (9 February 2014). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nature.com\/articles\/nclimate2106\">\"Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Nature_Climate_Change\" title=\"Nature Climate Change\">Nature Climate Change<\/a><\/i>. <b>4<\/b>: 222\u201327. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2014NatCC...4..222E\">2014NatCC...4..222E<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fnclimate2106\">10.1038\/nclimate2106<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-55\"><b><a href=\"#cite_ref-55\">^<\/a><\/b> <cite id=\"CITEREFKnightKenneyFollandHarris2009\" class=\"citation journal\">Knight, J.; Kenney, J.J.; Folland, C.; Harris, G.; Jones, G.S.; Palmer, M.; Parker, D.; Scaife, A.; Stott, P. (August 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.metoffice.gov.uk\/media\/pdf\/j\/j\/global_temperatures_09.pdf\">\"Do Global Temperature Trends Over the Last Decade Falsify Climate Predictions? [in \"State of the Climate in 2008\"]\"<\/a> (PDF). <i>Bull. Amer. Meteor. Soc<\/i>. <b>90<\/b> (8): S75\u201379. Retrieved 13 August 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-56\"><b><a href=\"#cite_ref-56\">^<\/a><\/b> <cite class=\"citation book\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/webarchive.nationalarchives.gov.uk\/20101209213753\/http:\/www.metoffice.gov.uk\/climatechange\/policymakers\/policy\/slowdown.html\"><i>Global temperature slowdown\u00a0\u2013 not an end to climate change<\/i><\/a>. UK Met Office. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.metoffice.gov.uk\/climatechange\/policymakers\/policy\/slowdown.html\">the original<\/a> on 9 December 2010. Retrieved 20 March 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-57\"><b><a href=\"#cite_ref-57\">^<\/a><\/b> <cite class=\"citation web\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gavin_Schmidt\" title=\"Gavin Schmidt\">Schmidt, Gavin<\/a> (4 June 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.realclimate.org\/index.php\/archives\/2015\/06\/noaa-temperature-record-updates-and-the-hiatus\/\">\"NOAA temperature record updates and the 'hiatus'\"<\/a>. RealClimate.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-58\"><b><a href=\"#cite_ref-58\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20180327091221\/http:\/\/www.noaanews.noaa.gov\/stories2015\/noaa-analysis-journal-science-no-slowdown-in-global-warming-in-recent-years.html\">\"<i>Science<\/i> publishes new NOAA analysis: Data show no recent slowdown in global warming\"<\/a>. NOAA. 4 June 2015. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.noaanews.noaa.gov\/stories2015\/noaa-analysis-journal-science-no-slowdown-in-global-warming-in-recent-years.html\">the original<\/a> on 27 March 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-59\"><b><a href=\"#cite_ref-59\">^<\/a><\/b> <cite class=\"citation news\">Mooney, Chris (18 January 2017). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.washingtonpost.com\/news\/energy-environment\/wp\/2017\/01\/18\/u-s-scientists-officially-declare-2016-the-hottest-year-on-record-that-makes-three-in-a-row\/\">\"U.S. scientists officially declare 2016 the hottest year on record. That makes three in a row\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Washington_Post\" title=\"The Washington Post\">The Washington Post<\/a><\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-Gavin_Schmidt_201501-60\"><b><a href=\"#cite_ref-Gavin_Schmidt_201501_60-0\">^<\/a><\/b> <cite class=\"citation web\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gavin_Schmidt\" title=\"Gavin Schmidt\">Schmidt, Gavin<\/a> (22 January 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.realclimate.org\/index.php\/archives\/2015\/01\/thoughts-on-2014-and-ongoing-temperature-trends\/\">\"Thoughts on 2014 and ongoing temperature trends\"<\/a>. RealClimate. Retrieved 4 September 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-61\"><b><a href=\"#cite_ref-61\">^<\/a><\/b> \n<cite class=\"citation web\">Group (28 November 2004). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.realclimate.org\/index.php\/archives\/2004\/11\/forcings\/\">\"Forcings (filed under: Glossary)\"<\/a>. RealClimate.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-62\"><b><a href=\"#cite_ref-62\">^<\/a><\/b> \n<cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20121025225858\/http:\/\/www.c2es.org\/docUploads\/PewSB1-Attribution-SMALL_102606.pdf\">\"Science Brief 1: The Causes of Global Climate Change\"<\/a> (PDF). Arlington, Virginia: Pew Center on Global Climate Change \/ Center for Climate and Energy Solutions. September 2006: 2. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.c2es.org\/docUploads\/PewSB1-Attribution-SMALL_102606.pdf\">the original<\/a> (PDF) on 25 October 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-63\"><b><a href=\"#cite_ref-63\">^<\/a><\/b> <cite class=\"citation journal\">Brown, Patrick T.; Li, Wenhong; Jiang, Jonathan H.; Su, Hui (7 December 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/journals.ametsoc.org\/doi\/abs\/10.1175\/JCLI-D-15-0384.1\">\"Unforced Surface Air Temperature Variability and Its Contrasting Relationship with the Anomalous TOA Energy Flux at Local and Global Spatial Scales\"<\/a>. <i>Journal of Climate<\/i>. <b>29<\/b> (3): 925\u201340. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2016JCli...29..925B\">2016JCli...29..925B<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1175%2FJCLI-D-15-0384.1\">10.1175\/JCLI-D-15-0384.1<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Serial_Number\" title=\"International Standard Serial Number\">ISSN<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.worldcat.org\/issn\/0894-8755\">0894-8755<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-64\"><b><a href=\"#cite_ref-64\">^<\/a><\/b> \n<a href=\"#CITEREFUS_NRC2012\">US NRC 2012<\/a>, p.\u00a09\n<\/li>\n<li id=\"cite_note-Hegerl-2001-65\">^ <a href=\"#cite_ref-Hegerl-2001_65-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-Hegerl-2001_65-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> Hegerl <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch9.html\">Chapter 9: Understanding and Attributing Climate Change<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch9s9-4-1-5.html\">Section 9.4.1.5: The Influence of Other Anthropogenic and Natural Forcings<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>, pp.\u00a0690\u2013691. \"Recent estimates indicate a relatively small combined effect of natural forcings on the global mean temperature evolution of the second half of the 20th century, with a small net cooling from the combined effects of solar and volcanic forcings.\" <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar4\/wg1\/ar4-wg1-chapter9.pdf\">p. 690<\/a>\n<\/li>\n<li id=\"cite_note-66\"><b><a href=\"#cite_ref-66\">^<\/a><\/b> <cite id=\"CITEREFTyndall1861\" class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/John_Tyndall\" title=\"John Tyndall\">Tyndall, John<\/a> (1 January 1861). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/rstl.royalsocietypublishing.org\/content\/151\/1.full.pdf+html\">\"On the Absorption and Radiation of Heat by Gases and Vapours, and on the Physical Connection of Radiation, Absorption, and Conduction\"<\/a>. <i>Philosophical Magazine<\/i>. 4. <b>22<\/b>: 169\u201394, 273\u201385. Retrieved 8 May 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-:4-67\">^ <a href=\"#cite_ref-:4_67-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-:4_67-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation web\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Spencer_R._Weart\" title=\"Spencer R. Weart\">Weart, Spencer<\/a> (2008). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/history.aip.org\/climate\/co2.htm\">\"The Carbon Dioxide Greenhouse Effect\"<\/a>. <i>The Discovery of Global Warming<\/i>. American Institute of Physics. Retrieved 21 April 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-68\"><b><a href=\"#cite_ref-68\">^<\/a><\/b> <cite class=\"citation web\">Molina, Francis (March 1912). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/books.google.com\/books?id=Tt4DAAAAMBAJ&amp;pg=PA339#v=onepage&amp;q&amp;f=false\">\"Remarkable Weather of 1911:The Effect of Combustion of Coal on Climate -- What Scientists Predict for the Future\"<\/a>. Popular Mechanics. pp.\u00a0339\u2013342. Retrieved 13 Oct 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-69\"><b><a href=\"#cite_ref-69\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/trove.nla.gov.au\/newspaper\/article\/100645214\">\"Coal Consumption Affecting Climate\"<\/a>. The Braidwood Dispatch and Mining Journal (New South Wales). 17 July 1912. p.\u00a04. Retrieved 13 Oct 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-70\"><b><a href=\"#cite_ref-70\">^<\/a><\/b> <cite class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Guy_Stewart_Callendar\" title=\"Guy Stewart Callendar\">Callendar, G. S.<\/a> (April 1938). \"The artificial production of carbon dioxide and its influence on temperature\". <i>Quarterly Journal of the Royal Meteorological Society<\/i>. <b>64<\/b> (275): 223\u201340. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/1938QJRMS..64..223C\">1938QJRMS..64..223C<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1002%2Fqj.49706427503\">10.1002\/qj.49706427503<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-71\"><b><a href=\"#cite_ref-71\">^<\/a><\/b> The Callendar Effect: the life and work of Guy Stewart Callendar (1898\u20131964) Amer Meteor Soc., Boston. <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-1-878220-76-9\" title=\"Special:BookSources\/978-1-878220-76-9\">978-1-878220-76-9<\/a>\n<\/li>\n<li id=\"cite_note-72\"><b><a href=\"#cite_ref-72\">^<\/a><\/b> <cite id=\"CITEREFLe_Treut\" class=\"citation book\">Le Treut;  et al. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch1.html\">\"Chapter 1: Historical Overview of Climate Change Science\"<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/faq-1-2.html\"><i>FAQ 1.1<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar4\/wg1\/ar4-wg1-chapter1.pdf\">p. 97<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>: \"To emit 240 W m\u20132, a surface would have to have a temperature of around \u221219\u00a0\u00b0C. This is much colder than the conditions that actually exist at the Earth's surface (the global mean surface temperature is about 14\u00a0\u00b0C). Instead, the necessary \u221219\u00a0\u00b0C is found at an altitude about 5 km above the surface.\"\n<\/li>\n<li id=\"cite_note-Blue-74\"><b><a href=\"#cite_ref-Blue_74-0\">^<\/a><\/b> <cite class=\"citation news\">Blue, Jessica. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20160415195934\/http:\/\/education.nationalgeographic.org\/encyclopedia\/greenhouse-effect\/\">\"What is the Natural Greenhouse Effect?\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Geographic_(magazine)\" class=\"mw-redirect\" title=\"National Geographic (magazine)\">National Geographic<\/a><\/i>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/education.nationalgeographic.org\/encyclopedia\/greenhouse-effect\/\">the original<\/a> on 15 April 2016. Retrieved 1 January 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-75\"><b><a href=\"#cite_ref-75\">^<\/a><\/b> <cite id=\"CITEREFKiehlTrenberth1997\" class=\"citation journal\">Kiehl, J.T.; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Kevin_Trenberth\" class=\"mw-redirect\" title=\"Kevin Trenberth\">Trenberth, Kevin E.<\/a> (1997). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20080624223905\/http:\/\/www.atmo.arizona.edu\/students\/courselinks\/spring04\/atmo451b\/pdf\/RadiationBudget.pdf\">\"Earth's Annual Global Mean Energy Budget\"<\/a> (PDF). <i>Bulletin of the American Meteorological Society<\/i>. <b>78<\/b> (2): 197\u2013208. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/1997BAMS...78..197K\">1997BAMS...78..197K<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1175%2F1520-0477%281997%29078%3C0197%3AEAGMEB%3E2.0.CO%3B2\">10.1175\/1520-0477(1997)078&lt;0197:EAGMEB&gt;2.0.CO;2<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Serial_Number\" title=\"International Standard Serial Number\">ISSN<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.worldcat.org\/issn\/1520-0477\">1520-0477<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.atmo.arizona.edu\/students\/courselinks\/spring04\/atmo451b\/pdf\/RadiationBudget.pdf\">the original<\/a> (PDF) on 24 June 2008. Retrieved 21 April 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-76\"><b><a href=\"#cite_ref-76\">^<\/a><\/b> <cite class=\"citation web\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gavin_Schmidt\" title=\"Gavin Schmidt\">Schmidt, Gavin<\/a> (6 April 2005). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.realclimate.org\/index.php?p=142\">\"Water vapour: feedback or forcing?\"<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/RealClimate\" title=\"RealClimate\">RealClimate<\/a>. Retrieved 21 April 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-77\"><b><a href=\"#cite_ref-77\">^<\/a><\/b> \n<cite class=\"citation web\">Russell, Randy (16 May 2007). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.windows.ucar.edu\/tour\/link=\/earth\/climate\/greenhouse_effect_gases.html&amp;edu=high\">\"The Greenhouse Effect &amp; Greenhouse Gases\"<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/University_Corporation_for_Atmospheric_Research\" title=\"University Corporation for Atmospheric Research\">University Corporation for Atmospheric Research<\/a> Windows to the Universe. Retrieved 27 December 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-78\"><b><a href=\"#cite_ref-78\">^<\/a><\/b> <cite class=\"citation web\">EPA (2007). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20090510053004\/http:\/\/www.epa.gov\/climatechange\/science\/recentac.html\">\"Recent Climate Change: Atmosphere Changes\"<\/a>. United States Environmental Protection Agency Climate Change Science Program. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.epa.gov\/climatechange\/science\/recentac.html\">the original<\/a> on 10 May 2009. Retrieved 21 April 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-79\"><b><a href=\"#cite_ref-79\">^<\/a><\/b> <cite id=\"CITEREFSpahniJ\u00e9r\u00f4me_ChappellazThomas_F._StockerLaetitia_Loulergue2005\" class=\"citation journal\">Spahni, Renato; J\u00e9r\u00f4me Chappellaz; Thomas F. Stocker; Laetitia Loulergue; Gregor Hausammann; Kenji Kawamura; Jacqueline Fl\u00fcckiger; Jakob Schwander; Dominique Raynaud; Val\u00e9rie Masson-Delmotte; Jean Jouzel (November 2005). \"Atmospheric Methane and Nitrous Oxide of the Late Pleistocene from Antarctic Ice Cores\". <i>Science<\/i>. <b>310<\/b> (5752): 1317\u201321. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2005Sci...310.1317S\">2005Sci...310.1317S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1120132\">10.1126\/science.1120132<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/16311333\">16311333<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-80\"><b><a href=\"#cite_ref-80\">^<\/a><\/b> <cite id=\"CITEREFSiegenthaler2005\" class=\"citation journal\">Siegenthaler, Urs;  et al. (November 2005). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.sciencemag.org\/cgi\/content\/abstract\/310\/5752\/1313\">\"Stable Carbon Cycle\u2013Climate Relationship During the Late Pleistocene\"<\/a> (PDF). <i>Science<\/i>. <b>310<\/b> (5752): 1313\u201317. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2005Sci...310.1313S\">2005Sci...310.1313S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1120130\">10.1126\/science.1120130<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/16311332\">16311332<\/a>. Retrieved 25 August 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-81\"><b><a href=\"#cite_ref-81\">^<\/a><\/b> <cite id=\"CITEREFPetit1999\" class=\"citation journal\">Petit, J. R.;  et al. (3 June 1999). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.daycreek.com\/dc\/images\/1999.pdf\">\"Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica\"<\/a> (PDF). <i>Nature<\/i>. <b>399<\/b> (6735): 429\u201336. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/1999Natur.399..429P\">1999Natur.399..429P<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2F20859\">10.1038\/20859<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20171117025013\/http:\/\/www.daycreek.com\/dc\/images\/1999.pdf\">Archived<\/a> (PDF) from the original on 17 November 2017. Retrieved 27 December 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-82\"><b><a href=\"#cite_ref-82\">^<\/a><\/b> <cite class=\"citation journal\">L\u00fcthi, Dieter; Le Floch, Martine; Bereiter, Bernhard; Blunier, Thomas; Barnola, Jean-Marc; Siegenthaler, Urs; Raynaud, Dominique; Jouzel, Jean; Fischer, Hubertus; Kawamura, Kenji; Stocker, Thomas F. (15 May 2008). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nature.com\/articles\/nature06949\">\"High-resolution carbon dioxide concentration record 650,000\u2013800,000\u2009years before present\"<\/a>. <i>Nature<\/i>. <b>453<\/b> (7193): 379\u2013382. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2008Natur.453..379L\">2008Natur.453..379L<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fnature06949\">10.1038\/nature06949<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/18480821\">18480821<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-83\"><b><a href=\"#cite_ref-83\">^<\/a><\/b> <cite id=\"CITEREFPearsonPalmer2000\" class=\"citation journal\">Pearson, Paul Nicholas; Palmer, Martin R. (17 August 2000). \"Atmospheric carbon dioxide concentrations over the past 60 million years\". <i>Nature<\/i>. <b>406<\/b> (6797): 695\u201399. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2F35021000\">10.1038\/35021000<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/10963587\">10963587<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-84\"><b><a href=\"#cite_ref-84\">^<\/a><\/b> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/005.htm\">Summary for Policymakers<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20160307220102\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/005.htm\">Archived<\/a> 7 March 2016 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/006.htm\">Concentrations of atmospheric greenhouse gases ...<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20040118133758\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/006.htm\">Archived<\/a> 18 January 2004 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/pdf\/WG1_TAR-FRONT.pdf\">p. 7<\/a>, in <a href=\"#CITEREFIPCC_TAR_WG12001\">IPCC TAR WG1 2001<\/a>.\n<\/li>\n<li id=\"cite_note-85\"><b><a href=\"#cite_ref-85\">^<\/a><\/b> IPCC (2007) AR4. Climate Change 2007: Working Group III: Mitigation of Climate Change, section 7.4.5.1. <a rel=\"nofollow\" class=\"external free\" href=\"https:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch7s7-4-5.html\">https:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch7s7-4-5.html<\/a>\n<\/li>\n<li id=\"cite_note-86\"><b><a href=\"#cite_ref-86\">^<\/a><\/b> <cite id=\"CITEREFLe_Qu\u00e9r\u00e9AndresBodenConway2012\" class=\"citation journal\">Le Qu\u00e9r\u00e9, C.; Andres, R.J.; Boden, T.; Conway, T.; Houghton, R. A.; House, J. I.; Marland, G.; Peters, G. P.; van der Werf, G.; Ahlstr\u00f6m, A.; Andrew, R. M.; Bopp, L.; Canadell, J. G.; Ciais, P.; Doney, S. C.; Enright, C.; Friedlingstein, P.; Huntingford, C.; Jain, A. K.; Jourdain, C.; Kato, E.; Keeling, R.F.; Klein Goldewijk, K.; Levis, S.; Levy, P.; Lomas, M.; Poulter, B.; Raupach, M.R.; Schwinger, J.; Sitch, S.; Stocker, B. D.; Viovy, N.; Zaehle, S.; Zeng, N. (2 December 2012). \"The global carbon budget 1959\u20132011\". <i>Earth System Science Data Discussions<\/i>. <b>5<\/b> (2): 1107\u20131157. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2012ESSDD...5.1107L\">2012ESSDD...5.1107L<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.5194%2Fessdd-5-1107-2012\">10.5194\/essdd-5-1107-2012<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-87\"><b><a href=\"#cite_ref-87\">^<\/a><\/b> <cite class=\"citation news\">Amos, Jonathan (10 May 2013). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.bbc.co.uk\/news\/science-environment-22486153\">\"Carbon dioxide passes symbolic mark\"<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/BBC\" title=\"BBC\">BBC<\/a>. Retrieved 27 May 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-88\"><b><a href=\"#cite_ref-88\">^<\/a><\/b> <cite class=\"citation news\">Clark, Pilita (10 May 2013). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ft.com\/cms\/s\/0\/e00ba374-b9a4-11e2-bc57-00144feabdc0.html\">\"CO<sub>2<\/sub> at highest level for millions of years\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Financial_Times\" title=\"Financial Times\">Financial Times<\/a><\/i>. Retrieved 27 May 2013. (Subscription required (help)).<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-89\"><b><a href=\"#cite_ref-89\">^<\/a><\/b> <cite class=\"citation web\">Schiermeier, Quirin (7 July 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nature.com\/news\/climate-scientists-discuss-future-of-their-field-1.17917\">\"Climate scientists discuss future of their field\"<\/a>. Nature.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-NASA-20151112-ab-90\"><b><a href=\"#cite_ref-NASA-20151112-ab_90-0\">^<\/a><\/b> <cite class=\"citation web\">Buis, Alan; Ramsayer, Kate; Rasmussen, Carol (12 November 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.jpl.nasa.gov\/news\/news.php?feature=4769\">\"A Breathing Planet, Off Balance\"<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/NASA\" title=\"NASA\">NASA<\/a>. Retrieved 13 November 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-91\"><b><a href=\"#cite_ref-91\">^<\/a><\/b> Rogner, H.-H., <i>et al<\/i>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch1.html\">Chap. 1, Introduction<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch1s1-3-1-2.html\">Section 1.3.1.2: Intensities<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG32007\">IPCC AR4 WG3 2007<\/a>.\n<\/li>\n<li id=\"cite_note-nrc2008-92\">^ <a href=\"#cite_ref-nrc2008_92-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-nrc2008_92-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation web\">NRC (2008). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/dels.nas.edu\/resources\/static-assets\/materials-based-on-reports\/booklets\/climate_change_2008_final.pdf\">\"Understanding and Responding to Climate Change\"<\/a> (PDF). Board on Atmospheric Sciences and Climate, US National Academy of Sciences. p.\u00a02. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20171011182257\/http:\/\/dels.nas.edu\/resources\/static-assets\/materials-based-on-reports\/booklets\/climate_change_2008_final.pdf\">Archived<\/a> (PDF) from the original on 11 October 2017. Retrieved 9 November 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-worldbank2010-93\"><b><a href=\"#cite_ref-worldbank2010_93-0\">^<\/a><\/b> <cite class=\"citation book\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20100305081236\/http:\/\/econ.worldbank.org\/WBSITE\/EXTERNAL\/EXTDEC\/EXTRESEARCH\/EXTWDRS\/EXTWDR2010\/0%2C%2CcontentMDK%3A21969137~menuPK%3A5287816~pagePK%3A64167689~piPK%3A64167673~theSitePK%3A5287741%2C00.html\"><i>World Development Report 2010: Development and Climate Change<\/i><\/a>. The International Bank for Reconstruction and Development \/ The World Bank, 1818 H Street NW, Washington, DC. 2010. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1596%2F978-0-8213-7987-5\">10.1596\/978-0-8213-7987-5<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-8213-7987-5\" title=\"Special:BookSources\/978-0-8213-7987-5\">978-0-8213-7987-5<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/econ.worldbank.org\/WBSITE\/EXTERNAL\/EXTDEC\/EXTRESEARCH\/EXTWDRS\/EXTWDR2010\/0,,contentMDK:21969137~menuPK:5287816~pagePK:64167689~piPK:64167673~theSitePK:5287741,00.html\">the original<\/a> on 5 March 2010. Retrieved 6 April 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-94\"><b><a href=\"#cite_ref-94\">^<\/a><\/b> Banuri <i>et al.<\/i>, Chapter 3: Equity and Social Considerations, Section 3.3.3: Patterns of greenhouse gas emissions, and Box 3.1, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/ipccreports\/sar\/wg_III\/ipcc_sar_wg_III_full_report.pdf\">pp. 92\u201393<\/a> in <a href=\"#CITEREFIPCC_SAR_WG31996\">IPCC SAR WG3 1996<\/a>.\n<\/li>\n<li id=\"cite_note-liverman-95\">^ <a href=\"#cite_ref-liverman_95-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-liverman_95-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-liverman_95-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> <cite id=\"CITEREFLiverman2009\" class=\"citation journal\">Liverman, Diana M. (April 2009). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.sciencedirect.com\/science\/article\/pii\/S0305748808001023\">\"Conventions of climate change: constructions of danger and the dispossession of the atmosphere\"<\/a> (PDF). <i>Journal of Historical Geography<\/i>. <b>35<\/b> (2): 279\u201396. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1016%2Fj.jhg.2008.08.008\">10.1016\/j.jhg.2008.08.008<\/a>. Retrieved 10 May 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-96\"><b><a href=\"#cite_ref-96\">^<\/a><\/b> Fisher <i>et al<\/i>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch3.html\">Chapter 3: Issues related to mitigation in the long-term context<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch3s3-1.html\">Section 3.1: Emissions scenarios: Issues related to mitigation in the long term context<\/a> in <a href=\"#CITEREFIPCC_AR4_WG32007\">IPCC AR4 WG3 2007<\/a>.\n<\/li>\n<li id=\"cite_note-97\"><b><a href=\"#cite_ref-97\">^<\/a><\/b> Morita, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg3\/068.htm\">Chapter 2: Greenhouse Gas Emission Mitigation Scenarios and Implications<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg3\/081.htm#2514\">Section 2.5.1.4: Emissions and Other Results of the SRES Scenarios<\/a>, in <a href=\"#CITEREFIPCC_TAR_WG32001\">IPCC TAR WG3 2001<\/a>.\n<\/li>\n<li id=\"cite_note-98\"><b><a href=\"#cite_ref-98\">^<\/a><\/b> Rogner <i>et al<\/i>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch1.html\">Ch. 1: Introduction<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch1s1-3-2-2.html\">Figure 1.7<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG32007\">IPCC AR4 WG3 2007<\/a>.\n<\/li>\n<li id=\"cite_note-99\"><b><a href=\"#cite_ref-99\">^<\/a><\/b> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg3\/003.htm\">Summary for Policymakers<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg3\/004.htm\">Introduction, paragraph 6<\/a>, in <a href=\"#CITEREFIPCC_TAR_WG32001\">IPCC TAR WG3 2001<\/a>.\n<\/li>\n<li id=\"cite_note-100\"><b><a href=\"#cite_ref-100\">^<\/a><\/b> Prentence <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/095.htm\">Chapter 3: The Carbon Cycle and Atmospheric Carbon Dioxide<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/096.htm\">Executive Summary<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20091207172617\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/096.htm\">Archived<\/a> 7 December 2009 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>., in <a href=\"#CITEREFIPCC_TAR_WG12001\">IPCC TAR WG1 2001<\/a>.\n<\/li>\n<li id=\"cite_note-101\"><b><a href=\"#cite_ref-101\">^<\/a><\/b> <cite class=\"citation book\">Solomon, S.; D. Qin; M. Manning; Z. Chen; M. Marquis; K.B. Averyt; M. Tignor; H.L. Miller, eds. (2007). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch3s3-4-4-2.html\">\"3.4.4.2 Surface Radiation\"<\/a>. <i>Climate Change 2007: Working Group I: The Physical Science Basis<\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-88009-1\" title=\"Special:BookSources\/978-0-521-88009-1\">978-0-521-88009-1<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-pmid10944197-102\"><b><a href=\"#cite_ref-pmid10944197_102-0\">^<\/a><\/b> <cite id=\"CITEREFHansenSatoRuedyLacis2000\" class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/James_Hansen\" title=\"James Hansen\">Hansen, James<\/a>; Sato, Makiko; Ruedy, Reto; Lacis, Andrew; Oinas, Valdar (29 August 2000). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC27611\">\"Global warming in the twenty-first century: An alternative scenario\"<\/a>. <i>Proc. Natl. Acad. Sci. USA<\/i>. <b>97<\/b> (18): 9875\u201380. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2000PNAS...97.9875H\">2000PNAS...97.9875H<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.170278997\">10.1073\/pnas.170278997<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC27611\">27611<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/10944197\">10944197<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-103\"><b><a href=\"#cite_ref-103\">^<\/a><\/b> <cite id=\"CITEREFRamanathanCarmichael2008\" class=\"citation journal\">Ramanathan, V.; Carmichael, G. (2008). \"Global and Regional Climate Changes due to Black Carbon\". <i>Nature Geoscience<\/i>. <b>1<\/b> (4): 221\u201327. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2008NatGe...1..221R\">2008NatGe...1..221R<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fngeo156\">10.1038\/ngeo156<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-104\"><b><a href=\"#cite_ref-104\">^<\/a><\/b> Statement made by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mark_Z._Jacobson\" title=\"Mark Z. Jacobson\">Mark Jacobson<\/a> of the Atmosphere \/ Energy Program at <a href=\"https:\/\/en.wikipedia.org\/wiki\/Stanford_University\" title=\"Stanford University\">Stanford University<\/a> in the documentary \"Sea Blind\"\n<\/li>\n<li id=\"cite_note-105\"><b><a href=\"#cite_ref-105\">^<\/a><\/b> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.rivm.nl\/en\/Documents_and_publications\/Common_and_Present\/Newsmessages\/2016\/Documentary_Sea_Blind_on_Dutch_Television\">Sea Blind<\/a>\n<\/li>\n<li id=\"cite_note-106\"><b><a href=\"#cite_ref-106\">^<\/a><\/b> V. Ramanathan and G. Carmichael, <i>supra<\/i> note 1, at 221 (\". . . emissions of black carbon are the second strongest contribution to current global warming, after carbon dioxide emissions.\") Numerous scientists also calculate that black carbon may be second only to CO<sub>2<\/sub> in its contribution to climate change, including Tami C. Bond &amp; Haolin Sun, <i>Can Reducing Black Carbon Emissions Counteract Global Warming<\/i>, <i>Environmental Science &amp; Technology<\/i> (2005), at 5921 (\"BC is the second or third largest individual warming agent, following carbon dioxide and methane.\"); <i>and<\/i> J. Hansen, <i>A Brighter Future<\/i>, 53 <i>Climate Change<\/i> 435 (2002), <i>available at<\/i> <a rel=\"nofollow\" class=\"external free\" href=\"http:\/\/pubs.giss.nasa.gov\/docs\/2002\/2002_Hansen_1.pdf\">http:\/\/pubs.giss.nasa.gov\/docs\/2002\/2002_Hansen_1.pdf<\/a> (calculating the climate forcing of BC at 1.0\u00b10.5 W\/m<sup>2<\/sup>).\n<\/li>\n<li id=\"cite_note-107\"><b><a href=\"#cite_ref-107\">^<\/a><\/b> <cite id=\"CITEREFTwomey1977\" class=\"citation journal\">Twomey, S. (1 July 1977). \"The Influence of Pollution on the Shortwave Albedo of Clouds\". <i>J. Atmos. Sci<\/i>. <b>34<\/b> (7): 1149\u201352. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/1977JAtS...34.1149T\">1977JAtS...34.1149T<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1175%2F1520-0469%281977%29034%3C1149%3ATIOPOT%3E2.0.CO%3B2\">10.1175\/1520-0469(1977)034&lt;1149:TIOPOT&gt;2.0.CO;2<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Serial_Number\" title=\"International Standard Serial Number\">ISSN<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.worldcat.org\/issn\/1520-0469\">1520-0469<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-108\"><b><a href=\"#cite_ref-108\">^<\/a><\/b> <cite id=\"CITEREFAlbrecht1989\" class=\"citation journal\">Albrecht, Bruce A. (15 September 1989). \"Aerosols, Cloud Microphysics, and Fractional Cloudiness\". <i>Science<\/i>. <b>245<\/b> (4923): 1227\u201339. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/1989Sci...245.1227A\">1989Sci...245.1227A<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.245.4923.1227\">10.1126\/science.245.4923.1227<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/17747885\">17747885<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-109\"><b><a href=\"#cite_ref-109\">^<\/a><\/b> IPCC, \"<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ipcc.ch\/ipccreports\/tar\/wg1\/pdf\/TAR-05.PDF\">Aerosols, their Direct and Indirect Effects<\/a>\", pp. 291\u201392 in <a href=\"#CITEREFIPCC_TAR_WG12001\">IPCC TAR WG1 2001<\/a>.\n<\/li>\n<li id=\"cite_note-110\"><b><a href=\"#cite_ref-110\">^<\/a><\/b> <cite class=\"citation journal\">Ramanathan, V.; Chung, C.; Kim, D.; Bettge, T.; Buja, L.; Kiehl, J. T.; Washington, W. M.; Fu, Q.; Sikka, D. R.; Wild, M. (2005). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pnas.org\/content\/102\/15\/5326.full\">\"Atmospheric brown clouds: Impacts on South Asian climate and hydrological cycle\"<\/a> (Full free text). <i>Proceedings of the National Academy of Sciences<\/i>. <b>102<\/b> (15): 5326\u201333. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2005PNAS..102.5326R\">2005PNAS..102.5326R<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.0500656102\">10.1073\/pnas.0500656102<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC552786\">552786<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/15749818\">15749818<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-111\"><b><a href=\"#cite_ref-111\">^<\/a><\/b> \n<cite class=\"citation web\">Ramanathan, V.;  et al. (2008). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20110718201256\/http:\/\/www.rrcap.unep.org\/abc\/impact\/files\/ABC_Report_Summary_Final.pdf\">\"Report Summary\"<\/a> (PDF). <i>Atmospheric Brown Clouds: Regional Assessment Report with Focus on Asia<\/i>. United Nations Environment Programme. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.rrcap.unep.org\/abc\/impact\/files\/ABC_Report_Summary_Final.pdf\">the original<\/a> (PDF) on 18 July 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-112\"><b><a href=\"#cite_ref-112\">^<\/a><\/b> \n<cite class=\"citation web\">Ramanathan, V.;  et al. (2008). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20110718201515\/http:\/\/www.rrcap.unep.org\/abc\/publication\/Part%20III.pdf\">\"Part III: Global and Future Implications\"<\/a> (PDF). <i>Atmospheric Brown Clouds: Regional Assessment Report with Focus on Asia<\/i>. United Nations Environment Programme. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.rrcap.unep.org\/abc\/publication\/Part%20III.pdf\">the original<\/a> (PDF) on 18 July 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-ipcc_radiative_forcing-113\">^ <a href=\"#cite_ref-ipcc_radiative_forcing_113-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-ipcc_radiative_forcing_113-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/spm.html\">Summary for Policymakers<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/spmsspm-human-and.html\">Human and Natural Drivers of Climate Change<\/a>, Figure SPM.2, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>.\n<\/li>\n<li id=\"cite_note-epa_solar_activity-114\"><b><a href=\"#cite_ref-epa_solar_activity_114-0\">^<\/a><\/b> <cite class=\"citation book\">\"3.2.2 Solar Irradiance\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.epa.gov\/climatechange\/endangerment\/comments\/volume3.html#2-2\"><i>Volume 3: Attribution of Observed Climate Change<\/i><\/a>. Endangerment and Cause or Contribute Findings for Greenhouse Gases under Section 202(a) of the Clean Air Act. EPA's Response to Public Comments. US Environmental Protection Agency. 2009. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20110616210216\/http:\/\/www.epa.gov\/climatechange\/endangerment\/comments\/volume3.html\">Archived<\/a> from the original on 16 June 2011. Retrieved 23 June 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-115\"><b><a href=\"#cite_ref-115\">^<\/a><\/b> \n<a href=\"#CITEREFUS_NRC2008\">US NRC 2008<\/a>, p.\u00a06\n<\/li>\n<li id=\"cite_note-116\"><b><a href=\"#cite_ref-116\">^<\/a><\/b> Hegerl, <i>et al<\/i>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch9.html\">Chapter 9: Understanding and Attributing Climate Change<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/faq-9-2.html\">Frequently Asked Question 9.2: Can the Warming of the 20th century be Explained by Natural Variability?<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>.\n<\/li>\n<li id=\"cite_note-Schmidt_et_al_2014-117\"><b><a href=\"#cite_ref-Schmidt_et_al_2014_117-0\">^<\/a><\/b> <cite class=\"citation journal\">Schmidt, Gavin A.;  et al. (March 2014). \"Reconciling warming trends\". <i>Nature Geoscience<\/i>. <b>7<\/b> (3): 158\u2013160. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2014NatGe...7..158S\">2014NatGe...7..158S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fngeo2105\">10.1038\/ngeo2105<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-Fyfe_et_al_2016-118\"><b><a href=\"#cite_ref-Fyfe_et_al_2016_118-0\">^<\/a><\/b> <cite class=\"citation journal\">Fyfe, John C.;  et al. (March 2016). \"Making sense of the early-2000s warming slowdown\". <i>Nature Climate Change<\/i>. <b>6<\/b> (3): 224\u2013228. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2016NatCC...6..224F\">2016NatCC...6..224F<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fnclimate2938\">10.1038\/nclimate2938<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-119\"><b><a href=\"#cite_ref-119\">^<\/a><\/b> <cite class=\"citation web\">Simmon, R.; D. Herring (November 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climate.gov\/#understandingClimate\/presentationLibrary\">\"Notes for slide number 7, titled \"Satellite evidence also suggests greenhouse gas warming\", in presentation, \"Human contributions to global climate change\"\"<\/a>. U.S. National Oceanic and Atmospheric Administration's Climate Services. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20110703210729\/http:\/\/www.climate.gov\/\">Archived<\/a> from the original on 3 July 2011. Retrieved 23 June 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-120\"><b><a href=\"#cite_ref-120\">^<\/a><\/b> Hegerl <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch9.html\">Chapter 9: Understanding and Attributing Climate Change<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/faq-9-2.html\">Frequently Asked Question 9.2: Can the Warming of the 20th century be Explained by Natural Variability?<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>. \n<\/li>\n<li id=\"cite_note-Randel2009-121\"><b><a href=\"#cite_ref-Randel2009_121-0\">^<\/a><\/b> <cite id=\"CITEREFRandel,_William_J.ShineAustinBarnett2009\" class=\"citation journal\">Randel, William J.; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Keith_Shine\" title=\"Keith Shine\">Shine, Keith P.<\/a>; Austin, John;  et al. (2009). \"An update of observed stratospheric temperature trends\". <i>Journal of Geophysical Research<\/i>. <b>114<\/b> (D2): D02107. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2009JGRD..11402107R\">2009JGRD..11402107R<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2008JD010421\">10.1029\/2008JD010421<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-122\"><b><a href=\"#cite_ref-122\">^<\/a><\/b> <a href=\"#CITEREFUSGCRP2009\">USGCRP 2009<\/a>, p.\u00a020\n<\/li>\n<li id=\"cite_note-123\"><b><a href=\"#cite_ref-123\">^<\/a><\/b> <cite class=\"citation book\">R.S. Bradley; K.R. Briffa; J. Cole; M.K. Hughes; T.J. Osborn (2003). \"The climate of the last millennium\".  In K.D. Alverson; R.S. Bradley; T.F. Pederson. <i>Paleoclimate, global change and the future<\/i>. Springer. pp.\u00a0105\u201341. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/3-540-42402-4\" title=\"Special:BookSources\/3-540-42402-4\">3-540-42402-4<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-124\"><b><a href=\"#cite_ref-124\">^<\/a><\/b> <cite class=\"citation journal\">Kaufman, D. S.; Schneider, D. P.; McKay, N. P.; Ammann, C. M.; Bradley, R. S.; Briffa, K. R.; Miller, G. H.; Otto-Bliesner, B. L.; Overpeck, J. T.; Vinther, B. M.; Abbott, M.; Axford, M.; Bird, Y.; Birks, B.; Bjune, H. J. B.; Briner, A. E.; Cook, J.; Chipman, T.; Francus, M.; Gajewski, P.; Geirsdottir, K.; Hu, A.; Kutchko, F. S.; Lamoureux, B.; Loso, S.; MacDonald, M.; Peros, G.; Porinchu, M.; Schiff, D.; Seppa, C.; Seppa, H.; Arctic Lakes 2k Project Members (2009). \"Recent Warming Reverses Long-Term Arctic Cooling\". <i>Science<\/i>. <b>325<\/b> (5945): 1236\u201339. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2009Sci...325.1236K\">2009Sci...325.1236K<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1173983\">10.1126\/science.1173983<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/19729653\">19729653<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-125\"><b><a href=\"#cite_ref-125\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20110427235538\/http:\/\/www2.ucar.edu\/news\/846\/arctic-warming-overtakes-2000-years-natural-cooling\">\"Arctic Warming Overtakes 2,000 Years of Natural Cooling\"<\/a>. UCAR. 3 September 2009. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www2.ucar.edu\/news\/846\/arctic-warming-overtakes-2000-years-natural-cooling\">the original<\/a> on 27 April 2011. Retrieved 8 June 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-126\"><b><a href=\"#cite_ref-126\">^<\/a><\/b> <cite class=\"citation news\">Bello, David (4 September 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.scientificamerican.com\/article.cfm?id=global-warming-reverses-arctic-cooling\">\"Global Warming Reverses Long-Term Arctic Cooling\"<\/a>. <i>Scientific American<\/i>. Retrieved 8 June 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-127\"><b><a href=\"#cite_ref-127\">^<\/a><\/b> <cite class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Michael_E._Mann\" title=\"Michael E. Mann\">Mann, Michael E.<\/a>; Zhang, Zhihua; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Malcolm_K._Hughes\" title=\"Malcolm K. Hughes\">Hughes, Malcolm K.<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Raymond_S._Bradley\" title=\"Raymond S. Bradley\">Bradley, Raymond S.<\/a>; Miller, Sonya K.; Rutherford, Scott; Ni, Fenbiao (9 September 2008). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC2527990\">\"Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia\"<\/a>. <i>Proceedings of the National Academy of Sciences<\/i>. <b>105<\/b> (36): 13252\u201357. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2008PNAS..10513252M\">2008PNAS..10513252M<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.0805721105\">10.1073\/pnas.0805721105<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC2527990\">2527990<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/18765811\">18765811<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-128\"><b><a href=\"#cite_ref-128\">^<\/a><\/b> <cite class=\"citation journal\">Berger, Andr\u00e9; Loutre, Marie-France (23 August 2002). \"CLIMATE: An Exceptionally Long Interglacial Ahead?\". <i>Science<\/i>. <b>297<\/b> (5585): 1287\u201388. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1076120\">10.1126\/science.1076120<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/12193773\">12193773<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-129\"><b><a href=\"#cite_ref-129\">^<\/a><\/b> <cite class=\"citation book\">Masson-Delmotte V. M.;  et al. (2013). \"Information from paleoclimate archives\".  In Stocker, T.F.;  et al. <i>Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change<\/i>. Cambridge University Press. pp.\u00a0383\u2013464. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-1-107-66182-0\" title=\"Special:BookSources\/978-1-107-66182-0\">978-1-107-66182-0<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-130\"><b><a href=\"#cite_ref-130\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/nsidc.org\/cryosphere\/seaice\/processes\/albedo.html\">\"Thermodynamics: Albedo\"<\/a>. <i>NSIDC<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-jackson_feedbacks-131\">^ <a href=\"#cite_ref-jackson_feedbacks_131-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-jackson_feedbacks_131-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> \n<cite id=\"CITEREFJackson,_R.A._Jenkins2012\" class=\"citation journal\">Jackson, R.; A. Jenkins (17 November 2012). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/climate.nasa.gov\/uncertainties\">\"Vital signs of the planet: global climate change and global warming: uncertainties\"<\/a>. Earth Science Communications Team at NASA's Jet Propulsion Laboratory \/ California Institute of Technology.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-132\"><b><a href=\"#cite_ref-132\">^<\/a><\/b> \n<cite id=\"CITEREFRiebeek,_H.2011\" class=\"citation journal\">Riebeek, H. (16 June 2011). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.earthobservatory.nasa.gov\/Features\/CarbonCycle\/page5.php\">\"The Carbon Cycle: Feature Articles: Effects of Changing the Carbon Cycle\"<\/a>. Earth Observatory, part of the EOS Project Science Office located at NASA Goddard Space Flight Center.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-133\"><b><a href=\"#cite_ref-133\">^<\/a><\/b> \n<cite id=\"CITEREFUS_National_Research_Council2003\" class=\"citation book\">US National Research Council (2003). \"Ch. 1 Introduction\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nap.edu\/openbook.php?record_id=10850&amp;page=19\"><i>Understanding Climate Change Feedbacks<\/i><\/a>. Washington, DC: National Academies Press. p.\u00a019.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-134\"><b><a href=\"#cite_ref-134\">^<\/a><\/b> <cite id=\"CITEREFLindsey,_R.2009\" class=\"citation journal\">Lindsey, R. (14 January 2009). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/earthobservatory.nasa.gov\/Features\/EnergyBalance\/page4.php\">\"Earth's Energy Budget, in: Climate and Earth's Energy Budget: Feature Articles\"<\/a>. Earth Observatory, part of the EOS Project Science Office, located at NASA Goddard Space Flight Center: 4.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-135\"><b><a href=\"#cite_ref-135\">^<\/a><\/b> \n<cite id=\"CITEREFUS_National_Research_Council2006\" class=\"citation book\">US National Research Council (2006). \"Ch. 1 Introduction to Technical Chapters\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/books.nap.edu\/openbook.php?record_id=11676&amp;page=26\"><i>Surface Temperature Reconstructions for the Last 2,000 Years<\/i><\/a>. Washington, DC: National Academies Press. pp.\u00a026\u201327.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-136\"><b><a href=\"#cite_ref-136\">^<\/a><\/b> <cite id=\"CITEREFAMS_Council2012\" class=\"citation journal\">AMS Council (20 August 2012). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ametsoc.org\/policy\/2012climatechange.html\">\"2012 American Meteorological Society (AMS) Information Statement on Climate Change\"<\/a>. Boston, Massachusetts.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-137\"><b><a href=\"#cite_ref-137\">^<\/a><\/b> <cite class=\"citation news\">Mooney, Chris (5 October 2017). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.washingtonpost.com\/news\/energy-environment\/wp\/2017\/10\/05\/one-of-the-oldest-climate-change-experiments-has-led-to-a-troubling-conclusion\">\"One of the oldest climate change experiments has led to a troubling conclusion\"<\/a>. <i>The Washington Post<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-138\"><b><a href=\"#cite_ref-138\">^<\/a><\/b> <cite class=\"citation journal\">Melillo, J.M.; Frey, S.D.; DeAngelis, K.M.; Werner, W.J.; Bernard, M.J.; Bowles, F.P.; Pold, G.; Knorr, M.A.; Grandy, A.S. (6 October 2017). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/science.sciencemag.org\/content\/358\/6359\/101\">\"Long-term pattern and magnitude of soil carbon feedback to the climate system in a warming world\"<\/a>. <i>Science<\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2017Sci...358..101M\">2017Sci...358..101M<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.aan2874\">10.1126\/science.aan2874<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-139\"><b><a href=\"#cite_ref-139\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/news.harvard.edu\/gazette\/story\/2012\/07\/concerns-about-climate-change-health\/\">\"Concerns about climate change, health\"<\/a>. Harvard University. 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-140\"><b><a href=\"#cite_ref-140\">^<\/a><\/b> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/tssts-3-1-2.html\">TS.3.1.2 Spatial Distribution of Changes in Temperature, Circulation and Related Variables \u2013 AR4 WGI Technical Summary<\/a>\n<\/li>\n<li id=\"cite_note-Alexeev-141\"><b><a href=\"#cite_ref-Alexeev_141-0\">^<\/a><\/b> <cite class=\"citation journal\">Alexeev V. A., Langen P. L., Bates J. R. (2005). \"Polar amplification of surface warming on an aquaplanet in \"ghost forcing\" experiments without sea ice feedbacks\". <i>Climate Dynamics<\/i>. <b>24<\/b> (7\u20138): 655\u2013666. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2005ClDy...24..655A\">2005ClDy...24..655A<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1007%2Fs00382-005-0018-3\">10.1007\/s00382-005-0018-3<\/a>.<\/cite>CS1 maint: Multiple names: authors list (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:CS1_maint:_Multiple_names:_authors_list\" title=\"Category:CS1 maint: Multiple names: authors list\">link<\/a>) <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-IPCC_2013-142\"><b><a href=\"#cite_ref-IPCC_2013_142-0\">^<\/a><\/b> <cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatechange2013.org\/images\/report\/WG1AR5_Chapter11_FINAL.pdf\">\"IPCC AR5 \u2013 Near-term Climate Change: Projections and Predictability (Chapter 11 \/ page 983 )\"<\/a> (PDF). 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-143\"><b><a href=\"#cite_ref-143\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/climate.nasa.gov\/news\/927\/arctic-amplification\">\"Arctic amplification\"<\/a>. <i>NASA<\/i>. 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-144\"><b><a href=\"#cite_ref-144\">^<\/a><\/b> <cite class=\"citation journal\">Francis Jennifer A (2012). \"Evidence linking Arctic amplification to extreme weather in mid-latitudes\". <i>Geophysical Research Letters<\/i>. <b>39<\/b>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2012GeoRL..39.6801F\">2012GeoRL..39.6801F<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2012GL051000\">10.1029\/2012GL051000<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-145\"><b><a href=\"#cite_ref-145\">^<\/a><\/b> <cite class=\"citation journal\">Vladimir Petoukhov &amp; Vladimir A. Semenov (November 2010). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/onlinelibrary.wiley.com\/doi\/10.1029\/2009JD013568\/abstract\">\"A link between reduced Barents-Kara sea ice and cold winter extremes over northern continents\"<\/a>. <i>Journal of Geophysical Research: Atmospheres<\/i>. <b>115<\/b> (21). <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2010JGRD..11521111P\">2010JGRD..11521111P<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2009JD013568\">10.1029\/2009JD013568<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-146\"><b><a href=\"#cite_ref-146\">^<\/a><\/b> <cite class=\"citation journal\">Francis, J. A.; Vavrus, S. J. (2012). \"Evidence linking Arctic amplification to extreme weather in mid-latitudes\". <i>Geophysical Research Letters<\/i>. <b>39<\/b> (6). <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2012GeoRL..39.6801F\">2012GeoRL..39.6801F<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2012GL051000\">10.1029\/2012GL051000<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-147\"><b><a href=\"#cite_ref-147\">^<\/a><\/b> \n<cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.gfdl.noaa.gov\/cms-filesystem-action\/user_files\/kd\/pdf\/gfdlhighlight_vol1n6.pdf\">\"Patterns of greenhouse warming\"<\/a> (PDF). <i>GFDL Climate Modeling Research Highlights<\/i>. Princeton, New Jersey: The National Oceanic and Atmospheric Administration (NOAA) Geophysical Fluid Dynamics Laboratory (GFDL). <b>1<\/b> (6). January 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, revision 2 February 2007, 8:50.08 AM.\n<\/li>\n<li id=\"cite_note-148\"><b><a href=\"#cite_ref-148\">^<\/a><\/b> \n<cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.gfdl.noaa.gov\/patterns-of-greenhouse-warming-ar4\">\"NOAA GFDL Climate Research Highlights Image Gallery: Patterns of Greenhouse Warming\"<\/a>. NOAA Geophysical Fluid Dynamics Laboratory (GFDL). 9 October 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-149\"><b><a href=\"#cite_ref-149\">^<\/a><\/b> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/annexessglossary-a-d.html\">Glossary A\u2013D<\/a>: \"Climate Model\", in <a href=\"#CITEREFIPCC_AR4_SYR2007\">IPCC AR4 SYR 2007<\/a>.\n<\/li>\n<li id=\"cite_note-meehl_2007_global_warming_projections-150\">^ <a href=\"#cite_ref-meehl_2007_global_warming_projections_150-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-meehl_2007_global_warming_projections_150-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> \n<cite id=\"CITEREFMeehl,_G._A.\" class=\"citation book\">Meehl, G. A.;  et al. \"Ch 10: Global Climate Projections\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch10s10-5-4-6.html\"><i>Sec 10.5.4.6 Synthesis of Projected Global Temperature at Year 2100]<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>\n<\/li>\n<li id=\"cite_note-151\"><b><a href=\"#cite_ref-151\">^<\/a><\/b> \n<cite id=\"CITEREFKarl,_TR2009\" class=\"citation book\">Karl, TR;  et al., eds. (2009). \"Global Climate Change\". <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20120915115254\/http:\/\/www.globalchange.gov\/publications\/reports\/scientific-assessments\/us-impacts\/full-report\/global-climate-change\"><i>Global Climate Change Impacts in the United States<\/i><\/a>. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-14407-0\" title=\"Special:BookSources\/978-0-521-14407-0\">978-0-521-14407-0<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.globalchange.gov\/publications\/reports\/scientific-assessments\/us-impacts\/full-report\/global-climate-change\">the original<\/a> on 15 September 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-NSDIC_permafrost_study-152\"><b><a href=\"#cite_ref-NSDIC_permafrost_study_152-0\">^<\/a><\/b> <cite id=\"CITEREFSchaeferZhangBruhwilerBarrett2011\" class=\"citation journal\">Schaefer, Kevin; Zhang, Tingjun; Bruhwiler, Lori; Barrett, Andrew P. (24 January 2011). \"Amount and timing of permafrost carbon release in response to climate warming\". <i>Tellus Series B<\/i>. <b>63<\/b> (2): 165\u201380. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2011TellB..63..165S\">2011TellB..63..165S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1111%2Fj.1600-0889.2011.00527.x\">10.1111\/j.1600-0889.2011.00527.x<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-153\"><b><a href=\"#cite_ref-153\">^<\/a><\/b> <cite class=\"citation book\">Hansen, James (2000). \"Climatic Change: Understanding Global Warming\".  In Lanza, Robert. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/books.google.com\/?id=sx6DFr8rbpIC&amp;dq=robert+lanza&amp;printsec=frontcover\"><i>One World: The Health &amp; Survival of the Human Species in the 21st century<\/i><\/a>. Health Press (New Mexico). pp.\u00a0173\u201390. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-929173-33-3\" title=\"Special:BookSources\/0-929173-33-3\">0-929173-33-3<\/a>. Retrieved 18 August 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-154\"><b><a href=\"#cite_ref-154\">^<\/a><\/b> Stocker <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/260.htm\">Chapter 7: Physical Climate Processes and Feedbacks<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/271.htm\">Section 7.2.2: Cloud Processes and Feedbacks<\/a>, in <a href=\"#CITEREFIPCC_TAR_WG12001\">IPCC TAR WG1 2001<\/a>.\n<\/li>\n<li id=\"cite_note-155\"><b><a href=\"#cite_ref-155\">^<\/a><\/b> <cite id=\"CITEREFTornHarte,_John2006\" class=\"citation journal\">Torn, Margaret; Harte, John (26 May 2006). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.webpages.uidaho.edu\/envs501\/downloads\/Torn%20%26%20Harte%202006.pdf\">\"Missing feedbacks, asymmetric uncertainties, and the underestimation of future warming\"<\/a> (PDF). <i>Geophysical Research Letters<\/i>. <b>33<\/b> (10): L10703. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2006GeoRL..3310703T\">2006GeoRL..3310703T<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2005GL025540\">10.1029\/2005GL025540<\/a>. Retrieved 4 March 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-156\"><b><a href=\"#cite_ref-156\">^<\/a><\/b> <cite id=\"CITEREFHarteSaleskaShih2006\" class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/John_Harte_(scientist)\" title=\"John Harte (scientist)\">Harte, John<\/a>; Saleska, Scott; Shih, Tiffany (30 October 2006). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/iopscience.iop.org\/article\/10.1088\/1748-9326\/1\/1\/014001\/meta\">\"Shifts in plant dominance control carbon-cycle responses to experimental warming and widespread drought\"<\/a>. <i>Environmental Research Letters<\/i>. <b>1<\/b> (1): 014001. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2006ERL.....1a4001H\">2006ERL.....1a4001H<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1088%2F1748-9326%2F1%2F1%2F014001\">10.1088\/1748-9326\/1\/1\/014001<\/a>. Retrieved 2 May 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-157\"><b><a href=\"#cite_ref-157\">^<\/a><\/b> <cite id=\"CITEREFSchefferBrovkin,_VictorCox,_Peter2006\" class=\"citation journal\">Scheffer, Marten; Brovkin, Victor; Cox, Peter (26 May 2006). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pik-potsdam.de\/~victor\/recent\/scheffer_etal_T_CO2_GRL_in_press.pdf\">\"Positive feedback between global warming and atmospheric CO<sub>2<\/sub> concentration inferred from past climate change\"<\/a> (PDF). <i>Geophysical Research Letters<\/i>. <b>33<\/b> (10): L10702. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2006GeoRL..3310702S\">2006GeoRL..3310702S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2005gl025044\">10.1029\/2005gl025044<\/a>. Retrieved 4 May 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-158\"><b><a href=\"#cite_ref-158\">^<\/a><\/b> Randall <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch8.html\">Chapter 8, Climate Models and Their Evaluation<\/a>, Sec. FAQ 8.1 in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>.\n<\/li>\n<li id=\"cite_note-159\"><b><a href=\"#cite_ref-159\">^<\/a><\/b> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20040604005623\/http:\/\/www.grida.no\/climate\/ipcc_tar\/wg1\/pdf\/WG1_TAR-FRONT.pdf\">Technical Summary<\/a>, p. 54, in <a href=\"#CITEREFIPCC_TAR_WG12001\">IPCC TAR WG1 2001<\/a>.\n<\/li>\n<li id=\"cite_note-160\"><b><a href=\"#cite_ref-160\">^<\/a><\/b> <cite id=\"CITEREFStroeve,_J.2007\" class=\"citation journal\">Stroeve, J.;  et al. (2007). \"Arctic sea ice decline: Faster than forecast\". <i>Geophysical Research Letters<\/i>. <b>34<\/b> (9): L09501. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2007GeoRL..3409501S\">2007GeoRL..3409501S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2007GL029703\">10.1029\/2007GL029703<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-161\"><b><a href=\"#cite_ref-161\">^<\/a><\/b> <cite id=\"CITEREFWentz,_F._J.2007\" class=\"citation journal\">Wentz, F. J.;  et al. (2007). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.sciencemag.org\/cgi\/content\/abstract\/317\/5835\/233\">\"How Much More Rain Will Global Warming Bring?\"<\/a>. <i>Science<\/i>. <b>317<\/b> (5835): 233\u201335. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2007Sci...317..233W\">2007Sci...317..233W<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1140746\">10.1126\/science.1140746<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/17540863\">17540863<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-162\"><b><a href=\"#cite_ref-162\">^<\/a><\/b> <cite id=\"CITEREFLiepertPrevidi,_Michael2009\" class=\"citation journal\">Liepert, Beate G.; Previdi, Michael (2009). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/journals.ametsoc.org\/doi\/abs\/10.1175\/2008JCLI2472.1\">\"Do Models and Observations Disagree on the Rainfall Response to Global Warming?\"<\/a>. <i>Journal of Climate<\/i>. <b>22<\/b> (11): 3156\u201366. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2009JCli...22.3156L\">2009JCli...22.3156L<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1175%2F2008JCLI2472.1\">10.1175\/2008JCLI2472.1<\/a>. <q>Recently analyzed satellite-derived global precipitation datasets from 1987 to 2006 indicate an increase in global-mean precipitation of 1.1%\u20131.4% decade\u22121. This trend corresponds to a hydrological sensitivity (HS) of 7% K\u22121 of global warming, which is close to the Clausius\u2013Clapeyron (CC) rate expected from the increase in saturation water vapor pressure with temperature. Analysis of two available global ocean evaporation datasets confirms this observed intensification of the atmospheric <a href=\"https:\/\/en.wikipedia.org\/wiki\/Water_cycle\" title=\"Water cycle\">water cycle<\/a>. The observed hydrological sensitivity over the past 20-yr period is higher by a factor of 5 than the average HS of 1.4% K\u22121 simulated in state-of-the-art coupled atmosphere\u2013ocean climate models for the twentieth and twenty-first centuries.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-163\"><b><a href=\"#cite_ref-163\">^<\/a><\/b> <cite class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Stefan_Rahmstorf\" title=\"Stefan Rahmstorf\">Rahmstorf, Stefan<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Anny_Cazenave\" title=\"Anny Cazenave\">Cazenave, Anny<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/John_A._Church\" title=\"John A. Church\">Church, John A.<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/James_Hansen\" title=\"James Hansen\">Hansen, James E.<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ralph_Keeling\" title=\"Ralph Keeling\">Keeling, Ralph F.<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/David_Parker_(climatologist)\" title=\"David Parker (climatologist)\">Parker, David E.<\/a>; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Richard_Somerville\" title=\"Richard Somerville\">Somerville, Richard C. J.<\/a> (4 May 2007). \"Recent Climate Observations Compared to Projections\". <i>Science<\/i>. <b>316<\/b> (5825): 709. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2007Sci...316..709R\">2007Sci...316..709R<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1136843\">10.1126\/science.1136843<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/17272686\">17272686<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-164\"><b><a href=\"#cite_ref-164\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/science2017.globalchange.gov\/chapter\/15\/\">\"Chapter 15: Potential Surprises: Compound Extremes and Tipping Elements\"<\/a>. <i>US National Climate Assessment<\/i>. 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-parris_sea_level_rise-165\"><b><a href=\"#cite_ref-parris_sea_level_rise_165-0\">^<\/a><\/b> 4. Global Mean Sea Level Rise Scenarios, in: Main Report, in <a href=\"#CITEREFParrisothers2012\">Parris &amp; others 2012<\/a>, p.\u00a012\n<\/li>\n<li id=\"cite_note-166\"><b><a href=\"#cite_ref-166\">^<\/a><\/b> \nExecutive Summary, in <a href=\"#CITEREFParrisothers2012\">Parris &amp; others 2012<\/a>, p.\u00a01\n<\/li>\n<li id=\"cite_note-167\"><b><a href=\"#cite_ref-167\">^<\/a><\/b> Schneider <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch19.html\">Chapter 19: Assessing Key Vulnerabilities and the Risk from Climate Change<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch19s19-3-4.html\">Section 19.3.4: Ecosystems and biodiversity<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG22007\">IPCC AR4 WG2 2007<\/a>.\n<\/li>\n<li id=\"cite_note-168\"><b><a href=\"#cite_ref-168\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.sciencedaily.com\/releases\/2018\/06\/180628105009.htm\">\"Climate change linked to potential population decline in bees\"<\/a>. ScienceDaily. 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-169\"><b><a href=\"#cite_ref-169\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.sciencedaily.com\/releases\/2012\/01\/120120184233.htm\">\"Carbon dioxide is 'driving fish crazy'\"<\/a>. ScienceDaily. 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-170\"><b><a href=\"#cite_ref-170\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nwf.org\/en\/Educational-Resources\/Wildlife-Guide\/Mammals\/Polar-Bear\">\"Global Warming and Polar Bears - National Wildlife Federation\"<\/a>. Retrieved 16 October 2017. <q>As climate change melts sea ice, the U.S. Geological Survey projects that two thirds of polar bears will disappear by 2050.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-171\"><b><a href=\"#cite_ref-171\">^<\/a><\/b> <cite class=\"citation web\">Amstrup, S. C.; Marcot, B. G.; Douglas, D. C. (2013). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/alaska.usgs.gov\/science\/biology\/polar_bears\/pdfs\/Amstrupetal_2008_AGU_Ch14_WithCitation.pdf\">\"A Bayesian Network Modeling Approach to Forecasting the 21st Century Worldwide Status of Polar Bears\"<\/a> (PDF). <i>Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications<\/i>. Geophysical Monograph Series 180. pp.\u00a0213\u2013268). <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F180GM14\">10.1029\/180GM14<\/a>.<\/cite>CS1 maint: Uses authors parameter (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:CS1_maint:_Uses_authors_parameter\" title=\"Category:CS1 maint: Uses authors parameter\">link<\/a>) <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-172\"><b><a href=\"#cite_ref-172\">^<\/a><\/b> <cite class=\"citation journal\">Zhang, Jinlun (11 June 2008). \"What drove the dramatic arctic sea ice retreat during summer 2007?\". <i>Geophysical Research Letters<\/i>. <b>35<\/b>: 1\u20135. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2008GeoRL..3511505Z\">2008GeoRL..3511505Z<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2008gl034005\">10.1029\/2008gl034005<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-173\"><b><a href=\"#cite_ref-173\">^<\/a><\/b> \n<cite id=\"CITEREFMeehl,_G.A.\" class=\"citation book\">Meehl, G.A.;  et al. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch10.html\">\"Ch 10: Global Climate Projections\"<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/ch10s10-3-2-4.html#10-3-3\"><i>Sec 10.3.3.1 Changes in Sea Ice Cover<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_AR4_WG12007\">IPCC AR4 WG1 2007<\/a>, p.\u00a0770\n<\/li>\n<li id=\"cite_note-174\"><b><a href=\"#cite_ref-174\">^<\/a><\/b> <cite id=\"CITEREFWangOverland,_J._E.2009\" class=\"citation journal\">Wang, M.; Overland, J. E. (2009). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.pmel.noaa.gov\/publications\/search_abstract.php?fmContributionNum=3261\">\"A sea ice free summer Arctic within 30 years?\"<\/a>. <i>Geophys. Res. Lett<\/i>. <b>36<\/b> (7). <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2009GeoRL..36.7502W\">2009GeoRL..36.7502W<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2009GL037820\">10.1029\/2009GL037820<\/a>. Retrieved 2 May 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-175\"><b><a href=\"#cite_ref-175\">^<\/a><\/b> \n<cite class=\"citation journal\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.metoffice.gov.uk\/research\/news\/sea-ice-2012\">\"Arctic sea ice 2012\"<\/a>. Exeter, UK: Met Office.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-176\"><b><a href=\"#cite_ref-176\">^<\/a><\/b> <cite class=\"citation journal\">Watson, Christopher S.; White, Neil J.; Church, John A.; King, Matt A.; Burgette, Reed J.; Legresy, Benoit (11 May 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nature.com\/nclimate\/journal\/vaop\/ncurrent\/full\/nclimate2635.html\">\"Unabated global mean sea-level rise over the satellite altimeter era\"<\/a>. <i>Nature Climate Change<\/i>. <b>5<\/b>: 565\u201368. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2015NatCC...5..565W\">2015NatCC...5..565W<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fnclimate2635\">10.1038\/nclimate2635<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-IPCC_SLR2013-177\"><b><a href=\"#cite_ref-IPCC_SLR2013_177-0\">^<\/a><\/b> <cite class=\"citation web\">Churchs, John; Clark, Peter. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatechange2013.org\/images\/uploads\/WGIAR5_WGI-12Doc2b_FinalDraft_Chapter13.pdf\">\"Chapter 13: Sea Level Change \u2013 Final Draft Underlying Scientific-Technical Assessment\"<\/a> (PDF). IPCC Working Group I. Retrieved 21 January 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-178\"><b><a href=\"#cite_ref-178\">^<\/a><\/b> <cite class=\"citation web\">Bell, Brian (31 August 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/news.uci.edu\/2015\/08\/31\/uci-study-finds-dramatic-increase-in-concurrent-droughts-heat-waves\/\">\"UCI study finds dramatic increase in concurrent droughts, heat waves\"<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/University_of_California,_Irvine\" title=\"University of California, Irvine\">University of California, Irvine<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-179\"><b><a href=\"#cite_ref-179\">^<\/a><\/b> <cite class=\"citation journal\">Ogburn, Stephanie Paige (29 April 2014). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.scientificamerican.com\/article\/indian-monsoons-are-becoming-more-extreme\/\">\"Indian Monsoons Are Becoming More Extreme\"<\/a>. Scientific American.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-180\"><b><a href=\"#cite_ref-180\">^<\/a><\/b> <cite class=\"citation book\">\"D. Future Climate Extremes, Impacts, and Disaster Losses, in: Summary for policymakers\". <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/docs.google.com\/file\/d\/0B1gFp6Ioo3akYklZcWkwWHJud00\/edit?pli=1\"><i>Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_SREX2012\">IPCC SREX 2012<\/a>, pp.\u00a09\u201313\n<\/li>\n<li id=\"cite_note-ipcc.ch-181\"><b><a href=\"#cite_ref-ipcc.ch_181-0\">^<\/a><\/b> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/spm.html\">Synthesis Report Summary for Policymakers<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/spms1.html\">Section 1: Observed changes in climate and their effects<\/a>, in <a href=\"#CITEREFIPCC_AR4_SYR2007\">IPCC AR4 SYR 2007<\/a>.\n<\/li>\n<li id=\"cite_note-182\"><b><a href=\"#cite_ref-182\">^<\/a><\/b> Fischlin, <i>et al<\/i>., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/contents.html\">Chapter 4: Ecosystems, their Properties, Goods and Services<\/a>,\n<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar4\/wg2\/ar4-wg2-chapter4.pdf\">Executive Summary, p. 213<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG22007\">IPCC AR4 WG2 2007<\/a>. Executive summary not present in on-line text; see pdf.\n<\/li>\n<li id=\"cite_note-ncadac_2013_climate-183\">^ <a href=\"#cite_ref-ncadac_2013_climate_183-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-ncadac_2013_climate_183-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-ncadac_2013_climate_183-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> Ocean Acidification, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ncadac.globalchange.gov\/download\/NCAJan11-2013-publicreviewdraft-chap2-climate.pdf\">Ch. 2. Our Changing Climate<\/a>, in <a href=\"#CITEREFNCADAC2013\">NCADAC 2013<\/a>, pp.\u00a069\u201370\n<\/li>\n<li id=\"cite_note-:2-184\"><b><a href=\"#cite_ref-:2_184-0\">^<\/a><\/b> <cite class=\"citation journal\">Deutsch;  et al. (2011). \"Climate-Forced Variability of Ocean Hypoxia\". <i>Science<\/i>. <b>333<\/b>: 336\u201339. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2011Sci...333..336D\">2011Sci...333..336D<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1126%2Fscience.1202422\">10.1126\/science.1202422<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/21659566\">21659566<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-185\"><b><a href=\"#cite_ref-185\">^<\/a><\/b> * <a href=\"#CITEREFUNEP2010\">UNEP 2010<\/a>\n<ul><li>5. Ocean acidification, in <a href=\"#CITEREFGoodothers2010\">Good &amp; others 2010<\/a>, pp.\u00a073\u201381<\/li>\n<li><a href=\"#CITEREFIAP2009\">IAP 2009<\/a><\/li><\/ul><\/li>\n<li id=\"cite_note-186\"><b><a href=\"#cite_ref-186\">^<\/a><\/b> <cite class=\"citation news\">Sarah Kaplan (2018). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.washingtonpost.com\/science\/2018\/08\/30\/climate-change-could-render-many-earths-ecosystems-unrecognizable\">\"Climate change could render many of Earth's ecosystems unrecognizable\"<\/a>. <i>The Washington Post<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-long-term_effects_of_global_warming-187\">^ <a href=\"#cite_ref-long-term_effects_of_global_warming_187-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-long-term_effects_of_global_warming_187-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> \n<ul><li>Summary, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nap.edu\/openbook.php?record_id=12877&amp;page=14\">pp. 14\u201319<\/a>, in <a href=\"#CITEREFNational_Research_Council2011\">National Research Council 2011<\/a><\/li>\n<li>FAQ 12.3, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatechange2013.org\/images\/uploads\/WGIAR5_WGI-12Doc2b_FinalDraft_Chapter12.pdf\">Chapter 12: Long-term Climate Change: Projections, Commitments and Irreversibility<\/a>, in <a href=\"#CITEREFIPCC_AR5_WG12013\">IPCC AR5 WG1 2013<\/a>, pp. 88\u201389 (pp. 90\u201391 of PDF chapter)<\/li><\/ul><\/li>\n<li id=\"cite_note-188\"><b><a href=\"#cite_ref-188\">^<\/a><\/b> <cite class=\"citation journal\">Bill McGuire (2010). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/rsta.royalsocietypublishing.org\/content\/368\/1919\/2311\">\"Climate forcing of geological and geomorphological hazards\"<\/a>. <i>Philosophical Transactions of the Royal Society A<\/i>. Royal Society. <b>368<\/b>: 2311\u201315. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2010RSPTA.368.2311M\">2010RSPTA.368.2311M<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1098%2Frsta.2010.0077\">10.1098\/rsta.2010.0077<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-189\"><b><a href=\"#cite_ref-189\">^<\/a><\/b> \n<cite id=\"CITEREFSmith,_J._B.\" class=\"citation book\">Smith, J. B.;  et al. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg2\/657.htm\">\"Ch. 19. Vulnerability to Climate Change and Reasons for Concern: A Synthesis\"<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg2\/679.htm\"><i>Sec 19.6. Extreme and Irreversible Effects<\/i><\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_TAR_WG22001\">IPCC TAR WG2 2001<\/a>\n<\/li>\n<li id=\"cite_note-190\"><b><a href=\"#cite_ref-190\">^<\/a><\/b> <cite class=\"citation journal\">Smith, Joel B.; Schneider, Stephen H.; Oppenheimer, Michael; Yohe, Gary W.; Hare, William; Mastrandrea, Michael D.; Patwardhan, Anand; Burton, Ian; Corfee-Morlot, Jan; Magadza, Chris H. D.; F\u00fcssel, Hans-Martin; Pittock, A. Barrie; Rahman, Atiq; Suarez, Avelino; van Ypersele, Jean-Pascal (17 March 2009). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC2648893\">\"Assessing dangerous climate change through an update of the Intergovernmental Panel on Climate Change (IPCC) 'reasons for concern'\"<\/a>. <i>Proceedings of the National Academy of Sciences<\/i>. <b>106<\/b> (11): 4133\u201337. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2009PNAS..106.4133S\">2009PNAS..106.4133S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.0812355106\">10.1073\/pnas.0812355106<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC2648893\">2648893<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/19251662\">19251662<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-ccsp_abrupt_climate_change-191\"><b><a href=\"#cite_ref-ccsp_abrupt_climate_change_191-0\">^<\/a><\/b> \n<cite id=\"CITEREFClark,_P._U.2008\" class=\"citation book\">Clark, P. U.;  et al. (December 2008). \"Executive Summary\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.globalchange.gov\/browse\/reports\/sap-34-abrupt-climate-change\"><i>Abrupt Climate Change. A Report by the U.S. Climate Change Science Program and the Subcommittee on Global Change Research<\/i><\/a>. Reston, Virginia: U.S. Geological Survey.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, pp. 1\u20137. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatescience.gov\/Library\/sap\/sap3-4\/final-report\/default.htm\">Report website<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20130504113820\/http:\/\/www.climatescience.gov\/Library\/sap\/sap3-4\/final-report\/default.htm\">Archived<\/a> 4 May 2013 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wayback_Machine\" title=\"Wayback Machine\">Wayback Machine<\/a>.\n<\/li>\n<li id=\"cite_note-192\"><b><a href=\"#cite_ref-192\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.bbc.co.uk\/news\/science-environment-21549643\">\"Siberian permafrost thaw warning sparked by cave data\"<\/a>. BBC. 22 February 2013. Retrieved 24 February 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-193\"><b><a href=\"#cite_ref-193\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20050113152254\/http:\/\/www.sciencedaily.com\/releases\/2004\/12\/041219153611.htm\">\"Shutdown Of Circulation Pattern Could Be Disastrous, Researchers Say\"<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/ScienceDaily\" class=\"mw-redirect\" title=\"ScienceDaily\">ScienceDaily<\/a>. 20 December 2004. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.sciencedaily.com\/releases\/2004\/12\/041219153611.htm\">the original<\/a> on 2005-01-13.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-194\"><b><a href=\"#cite_ref-194\">^<\/a><\/b> <cite class=\"citation journal\">Link, Peter Michael; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Richard_Tol\" title=\"Richard Tol\">Tol, Richard S.J.<\/a> (September 2004). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/core.ac.uk\/download\/pdf\/7079902.pdf\">\"Possible Economic Impacts of a Shutdown of the Thermohaline Circulation: an Application of <i>FUND<\/i>\"<\/a> (PDF). <i>Portuguese Economic Journal<\/i> (PDF). <b>3<\/b> (2): 99\u2013114. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1007%2Fs10258-004-0033-z\">10.1007\/s10258-004-0033-z<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-195\"><b><a href=\"#cite_ref-195\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.weatheronline.co.uk\/reports\/wxfacts\/North-Atlantic-Drift-Gulf-Stream.htm\">\"Weather Facts: North Atlantic Drift (Gulf Stream)\"<\/a>. Weather Online UK.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-196\"><b><a href=\"#cite_ref-196\">^<\/a><\/b> <cite class=\"citation journal\">Bischof, Barbie; Mariano, Arthur J.; Ryan, Edward H. (2003). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/oceancurrents.rsmas.miami.edu\/atlantic\/north-atlantic-drift.html\">\"Ocean Surface Currents: The North Atlantic Drift Current\"<\/a>. Rosenstiel School of Marine and Atmospheric Science, University of Miami.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-197\"><b><a href=\"#cite_ref-197\">^<\/a><\/b> FAQ 7 and 8, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Volume-FAQs_FGD.pdf\">Volume-wide Frequently Asked Questions (FAQs)<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Volume-FAQs_FGD.pdf\">8 July 2014<\/a>), pp. 2\u20133, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-198\"><b><a href=\"#cite_ref-198\">^<\/a><\/b> Oppenheimer, M., <i>et al<\/i>., Section 19.6.3: Updating Reasons for Concern, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap19_FGDall.pdf\">Chapter 19: Emergent risks and key vulnerabilities<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap19_FGDall.pdf\">8 July 2014<\/a>), pp. 39\u201346, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-199\"><b><a href=\"#cite_ref-199\">^<\/a><\/b> Field, C., <i>et al<\/i>., B-3: Regional Risks and Potential for Adaptation, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-TS_FGDall.pdf\">Technical Summary<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-TS_FGDall.pdf\">8 July 2014<\/a>), pp. 27\u201330, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-200\"><b><a href=\"#cite_ref-200\">^<\/a><\/b> Oppenheimer, M., <i>et al<\/i>., Section 19.6.3: Updating Reasons for Concern, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap19_FGDall.pdf\">Chapter 19: Emergent risks and key vulnerabilities<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap19_FGDall.pdf\">8 July 2014<\/a>), pp. 42\u201343, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-201\"><b><a href=\"#cite_ref-201\">^<\/a><\/b> <cite class=\"citation news\">Nuccitelli, Dana (26 January 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.theguardian.com\/environment\/climate-consensus-97-per-cent\/2015\/jan\/26\/climate-change-could-impact-poor-much-more-than-previously-thought\">\"Climate change could impact the poor much more than previously thought\"<\/a>. <i>The Guardian<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-202\"><b><a href=\"#cite_ref-202\">^<\/a><\/b> <cite class=\"citation book\"><a href=\"#CITEREFIPCC_AR4_SYR2007\">IPCC AR4 SYR 2007<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/mains3-3-3.html\"><i>3.3.3 Especially affected systems, sectors and regions<\/i><\/a>. Synthesis report.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-203\"><b><a href=\"#cite_ref-203\">^<\/a><\/b> <cite class=\"citation book\">Mimura, N.;  et al. (2007). \"Executive summary\".  In Parry, M.L.;  et al. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch16s16-es.html\"><i>Chapter 16: Small Islands<\/i><\/a>. Climate change 2007: impacts, adaptation and vulnerability: contribution of Working Group II to the fourth assessment report of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press (CUP): Cambridge: Print version: CUP. This version: IPCC website. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0521880106\" title=\"Special:BookSources\/0521880106\">0521880106<\/a>. Retrieved 15 September 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-204\"><b><a href=\"#cite_ref-204\">^<\/a><\/b> <cite class=\"citation web\">Park, Susin (May 2011). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20130502223251\/http:\/\/www.unhcr.org\/4df9cb0c9.pdf\">\"Climate Change and the Risk of Statelessness: The Situation of Low-lying Island States\"<\/a> (PDF). United Nations High Commissioner for Refugees. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.unhcr.org\/4df9cb0c9.pdf\">the original<\/a> (PDF) on 2013-05-02. Retrieved 13 April 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-205\"><b><a href=\"#cite_ref-205\">^<\/a><\/b> <cite class=\"citation web\">Mooney, Chris (22 October 2014). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.washingtonpost.com\/blogs\/wonkblog\/wp\/2014\/10\/22\/the-surprisingly-strong-link-between-climate-change-and-violence\/\">\"There's a surprisingly strong link between climate change and violence\"<\/a>. <i>The Washington Post<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-206\"><b><a href=\"#cite_ref-206\">^<\/a><\/b> <cite class=\"citation journal\">Hope, Chris; Schaefer, Kevin (21 September 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nature.com\/nclimate\/journal\/vaop\/ncurrent\/full\/nclimate2807.html\">\"Economic impacts of carbon dioxide and methane released from thawing permafrost\"<\/a>. <i>Nature<\/i>. <b>6<\/b>: 56\u201359. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2016NatCC...6...56H\">2016NatCC...6...56H<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fnclimate2807\">10.1038\/nclimate2807<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-207\"><b><a href=\"#cite_ref-207\">^<\/a><\/b> Porter, J.R., <i>et al<\/i>., Executive summary, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap7_FGDall.pdf\">Chapter 7: Food security and food production systems<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140709050337\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap7_FGDall.pdf\">8 July 2014<\/a>), p. 3, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-208\"><b><a href=\"#cite_ref-208\">^<\/a><\/b> Reference temperature period converted from late-20th century to pre-industrial times (approximated in the source as 1850\u20131900).* Assessment Box SPM-1 (p. 14) and B-2. Sectoral Risks and Potential for Adaptation: Food security and food production systems (p. 18), in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WG2AR5_SPM_FINAL.pdf\">Summary for Policymakers<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WG2AR5_SPM_FINAL.pdf\">8 July 2014<\/a>), in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-Fifth_Assessment_Report_(AR5),_WGII,_Climate_Change_2014:_Impacts,_Adaptation,_and_Vulnerability,_Chapter_7:_Food_Security_and_Food_Production_Systems-209\"><b><a href=\"#cite_ref-Fifth_Assessment_Report_(AR5),_WGII,_Climate_Change_2014:_Impacts,_Adaptation,_and_Vulnerability,_Chapter_7:_Food_Security_and_Food_Production_Systems_209-0\">^<\/a><\/b> <cite class=\"citation book\">R. Porter, John; Xie, Liyong (2014). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/assessment-report\/ar5\/wg2\/WGIIAR5-Chap7_FINAL.pdf\"><i>Fifth Assessment Report (AR5), WGII, Climate Change 2014: Impacts, Adaptation, and Vulnerability, Chapter 7: Food Security and Food Production Systems<\/i><\/a> (PDF). Intergovermental Pannel on Climate Change. pp.\u00a0491\u2013492. Retrieved 29 July 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-:3-210\">^ <a href=\"#cite_ref-:3_210-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-:3_210-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-:3_210-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> Cramer, Wolfgang, <i>et al<\/i>., Executive summary, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap18_FGDall.pdf\">Chapter 18: Detection and attribution of observed impacts<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap18_FGDall.pdf\">8 July 2014<\/a>), pp. 3\u20134, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-211\"><b><a href=\"#cite_ref-211\">^<\/a><\/b> Smith, K.R., <i>et al<\/i>., FAQ 11.2, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap11_FGDall.pdf\">Chapter 11: Human health: impacts, adaptation, and co-benefits<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap11_FGDall.pdf\">8 July 2014<\/a>), p. 37, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-lancet2009-212\">^ <a href=\"#cite_ref-lancet2009_212-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-lancet2009_212-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation journal\">Costello, Anthony; Abbas, Mustafa; Allen, Adriana; Ball, Sarah; Bell, Sarah; Bellamy, Richard; Friel, Sharon; Groce, Nora; Johnson, Anne; Kett, Maria; Lee, Maria; Levy, Caren; Maslin, Mark; McCoy, David; McGuire, Bill; Montgomery, Hugh; Napier, David; Pagel, Christina; Patel, Jinesh; de Oliveira, Jose Antonio Puppim; Redclift, Nanneke; Rees, Hannah; Rogger, Daniel; Scott, Joanne; Stephenson, Judith; Twigg, John; Wolff, Jonathan; Patterson, Craig (May 2009). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.thelancet.com\/journals\/lancet\/article\/PIIS0140-6736%2809%2960935-1\/fulltext\">\"Managing the health effects of climate change\"<\/a>. <i>The Lancet<\/i>. <b>373<\/b> (9676): 1693\u2013733. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1016%2FS0140-6736%2809%2960935-1\">10.1016\/S0140-6736(09)60935-1<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-lancet2015-213\">^ <a href=\"#cite_ref-lancet2015_213-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-lancet2015_213-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation journal\">Watts, Nick; Adger, W Neil; Agnolucci, Paolo; Blackstock, Jason; Byass, Peter; Cai, Wenjia; Chaytor, Sarah; Colbourn, Tim; Collins, Mat; Cooper, Adam; Cox, Peter M; Depledge, Joanna; Drummond, Paul; Ekins, Paul; Galaz, Victor; Grace, Delia; Graham, Hilary; Grubb, Michael; Haines, Andy; Hamilton, Ian; Hunter, Alasdair; Jiang, Xujia; Li, Moxuan; Kelman, Ilan; Liang, Lu; Lott, Melissa; Lowe, Robert; Luo, Yong; Mace, Georgina; Maslin, Mark; Nilsson, Maria; Oreszczyn, Tadj; Pye, Steve; Quinn, Tara; Svensdotter, My; Venevsky, Sergey; Warner, Koko; Xu, Bing; Yang, Jun; Yin, Yongyuan; Yu, Chaoqing; Zhang, Qiang; Gong, Peng; Montgomery, Hugh; Costello, Anthony (November 2015). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.thelancet.com\/journals\/lancet\/article\/PIIS0140-6736%2815%2960854-6\/fulltext\">\"Health and climate change: policy responses to protect public health\"<\/a>. <i>The Lancet<\/i>. <b>386<\/b> (10006): 1861\u2013914. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1016%2FS0140-6736%2815%2960854-6\">10.1016\/S0140-6736(15)60854-6<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/26111439\">26111439<\/a>. Retrieved 4 January 2016.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-214\"><b><a href=\"#cite_ref-214\">^<\/a><\/b> Smith, K.R., <i>et al<\/i>., Section 11.4: Direct Impacts of Climate and Weather on Health, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap11_FGDall.pdf\">Chapter 11: Human health: impacts, adaptation, and co-benefits<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap11_FGDall.pdf\">8 July 2014<\/a>), pp. 10\u201313, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-215\"><b><a href=\"#cite_ref-215\">^<\/a><\/b> Smith, K.R., <i>et al<\/i>., Section 11.6.1. Nutrition, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap11_FGDall.pdf\">Chapter 11: Human health: impacts, adaptation, and co-benefits<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140708062559\/http:\/\/ipcc-wg2.gov\/AR5\/images\/uploads\/WGIIAR5-Chap11_FGDall.pdf\">8 July 2014<\/a>), pp. 10\u201313, in <a href=\"#CITEREFIPCC_AR5_WG2_A2014\">IPCC AR5 WG2 A 2014<\/a>\n<\/li>\n<li id=\"cite_note-216\"><b><a href=\"#cite_ref-216\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/amp.usatoday.com\/amp\/817731002\">\"Global warming risk: Rising temperatures from climate change linked to rise in suicides\"<\/a>. <i>USA Today<\/i>. 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-217\"><b><a href=\"#cite_ref-217\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.theatlantic.com\/science\/archive\/2018\/07\/high-temperatures-cause-suicide-rates-to-increase\/565826\/\">\"Climate Change May Cause 26,000 More U.S. Suicides by 2050\"<\/a>. <i>The Atlantic<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-218\"><b><a href=\"#cite_ref-218\">^<\/a><\/b> \n<cite class=\"citation book\">Core Writing Team;  et al., eds. (2007d). \"3.3.3 Especially affected systems, sectors and regions\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/mains3-3-3.html\"><i>Synthesis report<\/i><\/a>. Climate Change 2007: Synthesis Report. A Contribution of Working Groups I, II, and III to the Fourth Assessment Report of the Integovernmental Panel on Climate Change (IPCC). Geneva, Switzerland: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">Intergovernmental Panel on Climate Change<\/a>. Retrieved 15 September 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-us_epa_regional_impacts-219\">^ <a href=\"#cite_ref-us_epa_regional_impacts_219-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-us_epa_regional_impacts_219-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-us_epa_regional_impacts_219-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> \n<img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/12px-PD-icon.svg.png\" width=\"12\" height=\"12\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/18px-PD-icon.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/24px-PD-icon.svg.png 2x\" data-file-width=\"196\" data-file-height=\"196\" \/>\u00a0This article incorporates <a href=\"https:\/\/en.wikipedia.org\/wiki\/Copyright_status_of_work_by_the_U.S._government\" title=\"Copyright status of work by the U.S. government\">public domain material<\/a>\u00a0from the\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/US_Environmental_Protection_Agency\" class=\"mw-redirect\" title=\"US Environmental Protection Agency\">US Environmental Protection Agency<\/a> document:\u00a0<cite class=\"citation\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.epa.gov\/climatechange\/impacts-adaptation\/international.html\"><i>International Impacts &amp; Adaptation: Climate Change: US EPA<\/i><\/a>, US Environmental Protection Agency (US EPA), 14 June 2012<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-220\"><b><a href=\"#cite_ref-220\">^<\/a><\/b> <cite class=\"citation\">Steven C. Sherwood and Matthew Huber (2010), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pnas.org\/content\/107\/21\/9552\">\"An adaptability limit to climate change due to heat stress\"<\/a>, <i>PNAS<\/i>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.0913352107\">10.1073\/pnas.0913352107<\/a><\/cite>CS1 maint: Uses authors parameter (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:CS1_maint:_Uses_authors_parameter\" title=\"Category:CS1 maint: Uses authors parameter\">link<\/a>) <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-221\"><b><a href=\"#cite_ref-221\">^<\/a><\/b> \n<cite id=\"CITEREFPBL_Netherlands_Environment_Agency2012\" class=\"citation book\">PBL Netherlands Environment Agency (15 June 2012). \"Figure 6.14, in: Chapter 6: The energy and climate challenge\".  In van Vuuren, D.; M. Kok. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pbl.nl\/sites\/default\/files\/cms\/publicaties\/pbl-2012-roads-from-rio-pathways-to-achieve-global-sustainability-goals-by-2050_0.pdf\"><i>Roads from Rio+20<\/i><\/a> (PDF). <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-90-78645-98-6\" title=\"Special:BookSources\/978-90-78645-98-6\">978-90-78645-98-6<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, p. 177, Report no: 500062001. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pbl.nl\/en\/publications\/2012\/roads-from-rio20\">Report website.<\/a>\n<\/li>\n<li id=\"cite_note-222\"><b><a href=\"#cite_ref-222\">^<\/a><\/b> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.globalchange.gov\/climate-change\/glossary#letter_m\">Mitigation<\/a>, in <a href=\"#CITEREFUSGCRP2015\">USGCRP 2015<\/a>\n<\/li>\n<li id=\"cite_note-autogenerated2007-223\">^ <a href=\"#cite_ref-autogenerated2007_223-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-autogenerated2007_223-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-autogenerated2007_223-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-autogenerated2007_223-3\"><sup><i><b>d<\/b><\/i><\/sup><\/a> IPCC, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/spm.html\">Synthesis Report Summary for Policymakers<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/spms4.html\">Section 4: Adaptation and mitigation options<\/a>, in <a href=\"#CITEREFIPCC_AR4_SYR2007\">IPCC AR4 SYR 2007<\/a>.\n<\/li>\n<li id=\"cite_note-2014_mitigation_options-224\">^ <a href=\"#cite_ref-2014_mitigation_options_224-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-2014_mitigation_options_224-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> Edenhofer, O., <i>et al<\/i>., Table TS.3, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_technical-summary.pdf\">Technical summary<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141230092610\/http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_technical-summary.pdf\">30 December 2014)<\/a>, in: <a href=\"#CITEREFIPCC_AR5_WG32014\">IPCC AR5 WG3 2014<\/a>, p.\u00a068\n<\/li>\n<li id=\"cite_note-225\"><b><a href=\"#cite_ref-225\">^<\/a><\/b> <cite class=\"citation news\">Nuccitelli, Dana (31 August 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.theguardian.com\/environment\/climate-consensus-97-per-cent\/2015\/aug\/31\/citi-report-slowing-global-warming-would-save-tens-of-trillions-of-dollars\">\"Citi report: slowing global warming would save tens of trillions of dollars\"<\/a>. <i>The Guardian<\/i>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-226\"><b><a href=\"#cite_ref-226\">^<\/a><\/b> Clarke, Leon, <i>et al<\/i>., Executive summary, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_chapter6.pdf\">Chapter 6: Assessing Transformation Pathways<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141230092610\/http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_chapter6.pdf\">30 December 2014)<\/a>, in: <a href=\"#CITEREFIPCC_AR5_WG32014\">IPCC AR5 WG3 2014<\/a>, p.\u00a0418\n<\/li>\n<li id=\"cite_note-progress_to_2_degrees_C-227\">^ <a href=\"#cite_ref-progress_to_2_degrees_C_227-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-progress_to_2_degrees_C_227-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> SPM4.1: Long-term mitigation pathways, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/report.mitigation2014.org\/spm\/ipcc_wg3_ar5_summary-for-policymakers.pdf\">Summary for Policymakers<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141227021430\/http:\/\/report.mitigation2014.org\/spm\/ipcc_wg3_ar5_summary-for-policymakers.pdf\">27 December 2014)<\/a>, in: <a href=\"#CITEREFIPCC_AR5_WG32014\">IPCC AR5 WG3 2014<\/a>, pp.\u00a010\u201313\n<\/li>\n<li id=\"cite_note-228\"><b><a href=\"#cite_ref-228\">^<\/a><\/b> Edenhofer, O., <i>et al<\/i>., TS.3.1.2: Short- and long-term requirements of mitigation pathways, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_technical-summary.pdf\">Technical summary<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141230092610\/http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_technical-summary.pdf\">30 December 2014)<\/a>, in: <a href=\"#CITEREFIPCC_AR5_WG32014\">IPCC AR5 WG3 2014<\/a>, pp.\u00a055\u201356\n<\/li>\n<li id=\"cite_note-229\"><b><a href=\"#cite_ref-229\">^<\/a><\/b> Edenhofer, O., <i>et al<\/i>., TS.3.1.3: Costs, investments and burden sharing, in: <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_technical-summary.pdf\">Technical summary<\/a> (archived <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141230092610\/http:\/\/report.mitigation2014.org\/report\/ipcc_wg3_ar5_technical-summary.pdf\">30 December 2014)<\/a>, in: <a href=\"#CITEREFIPCC_AR5_WG32014\">IPCC AR5 WG3 2014<\/a>, p.\u00a058\n<\/li>\n<li id=\"cite_note-230\"><b><a href=\"#cite_ref-230\">^<\/a><\/b> Smit <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg2\/641.htm\">Chapter 18: Adaptation to Climate Change in the Context of Sustainable Development and Equity<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/climate\/ipcc_tar\/wg2\/644.htm#1823\">Section 18.2.3: Adaptation Types and Forms<\/a>, in <a href=\"#CITEREFIPCC_TAR_WG22001\">IPCC TAR WG2 2001<\/a>.\n<\/li>\n<li id=\"cite_note-231\"><b><a href=\"#cite_ref-231\">^<\/a><\/b> <cite class=\"citation pressrelease\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.globalchange.gov\/news\/new-assessment-national-regional-impacts-climate-change\">\"New Report Provides Authoritative Assessment of National, Regional Impacts of Global Climate Change\"<\/a> (Press release). <a href=\"https:\/\/en.wikipedia.org\/wiki\/U.S._Global_Change_Research_Program\" title=\"U.S. Global Change Research Program\">U.S. Global Change Research Program<\/a>. 16 June 2009. Retrieved 14 January 2016.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-232\"><b><a href=\"#cite_ref-232\">^<\/a><\/b> Cole, Daniel A. \"Climate Change, Adaptation, and Development\", 26 UCLA J. ENVTL. L. &amp; POL'Y 1, 3 (2008)\n<\/li>\n<li id=\"cite_note-233\"><b><a href=\"#cite_ref-233\">^<\/a><\/b> <cite class=\"citation book\">Schneider, S.H.; Semenov, S.; Patwardhan, A.; Burton, I.; Magadza, C.H.D.; Oppenheimer, M.; Pittock, A.B.; Rahman, A.; Smith, J.B.; Suarez, A.; Yamin, F. (2007).  Parry, M.L.; Canziani, O.F.; Palutikof, J.P.; van der Linden, P.J.; Hanson, C.E., eds. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch19s19-es.html\"><i>Executive summary. In (book chapter): Chapter 19: Assessing Key Vulnerabilities and the Risk from Climate Change. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change<\/i><\/a>. Print version: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cambridge_University_Press\" title=\"Cambridge University Press\">Cambridge University Press<\/a>, Cambridge, UK. This version: IPCC website. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-88010-7\" title=\"Special:BookSources\/978-0-521-88010-7\">978-0-521-88010-7<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20100502171249\/http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/ch19s19-es.html\">Archived<\/a> from the original on 2 May 2010. Retrieved 2010-04-06.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-234\"><b><a href=\"#cite_ref-234\">^<\/a><\/b> <cite class=\"citation web\">Lane, Lee; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ken_Caldeira\" title=\"Ken Caldeira\">Caldeira, Ken<\/a>; Chatfield, Robert; Langhoff, Stephanie (April 2007). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20090531173021\/http:\/\/event.arc.nasa.gov\/main\/home\/reports\/SolarRadiationCP.pdf\">\"Workshop on managing solar radiation\"<\/a> (PDF). <a href=\"https:\/\/en.wikipedia.org\/wiki\/NASA\" title=\"NASA\">NASA<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/event.arc.nasa.gov\/main\/home\/reports\/SolarRadiationCP.pdf\">the original<\/a> (PDF) on 31 May 2009. Retrieved 23 May 2009.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-235\"><b><a href=\"#cite_ref-235\">^<\/a><\/b> <cite class=\"citation pressrelease\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20110624054716\/http:\/\/royalsociety.org\/Stop-emitting-CO2-or-geoengineering-could-be-our-only-hope\/\">\"Stop emitting CO2 or geoengineering could be our only hope\"<\/a> (Press release). The Royal Society. 28 August 2009. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/royalsociety.org\/Stop-emitting-CO2-or-geoengineering-could-be-our-only-hope\/\">the original<\/a> on 24 June 2011. Retrieved 14 June 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-236\"><b><a href=\"#cite_ref-236\">^<\/a><\/b> <cite class=\"citation journal\">Keller, David P.; Feng, Ellias Y.; Oschlies, Andreas (January 2014). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nature.com\/ncomms\/2014\/140225\/ncomms4304\/full\/ncomms4304.html\">\"Potential climate engineering effectiveness and side effects during a high carbon dioxide-emission scenario\"<\/a>. <i>Nature Communications<\/i>. <b>5<\/b>: 3304. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2014NatCo...5E3304K\">2014NatCo...5E3304K<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1038%2Fncomms4304\">10.1038\/ncomms4304<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC3948393\">3948393<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/24569320\">24569320<\/a>. Retrieved 31 March 2014. <q>We find that even when applied continuously and at scales as large as currently deemed possible, all methods are, individually, either relatively ineffective with limited (&lt;8%) warming reductions, or they have potentially severe side effects and cannot be stopped without causing rapid climate change.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-237\"><b><a href=\"#cite_ref-237\">^<\/a><\/b> Quoted in <a href=\"#CITEREFIPCC_SAR_SYR1996\">IPCC SAR SYR 1996<\/a>, \"Synthesis of Scientific-Technical Information Relevant to Interpreting Article 2 of the UN Framework Convention on Climate Change\", paragraph 4.1, p. 8 (<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/climate-changes-1995\/ipcc-2nd-assessment\/2nd-assessment-en.pdf\">pdf p. 18<\/a>.) \n<\/li>\n<li id=\"cite_note-238\"><b><a href=\"#cite_ref-238\">^<\/a><\/b> <cite class=\"citation book\">Morgan, M. Granger; Dowlatabadi, Hadi; Henrion, Max; Keith, David; Lempert, Robert; McBride, Sandra; Small, Mitchell; Wilbanks, Thomas (2009). \"Non-Technical Summary: BOX NT.1 Summary of Climate Change Basics\". <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/downloads.globalchange.gov\/sap\/sap5-2\/sap5-2-final-report-all.pdf\"><i>Synthesis and Assessment Product 5.2: Best practice approaches for characterizing, communicating, and incorporating scientific uncertainty in decisionmaking. A Report by the U.S. Climate Change Science Program and the Subcommittee on Global Change Research<\/i><\/a> (PDF). Washington, D.C.: National Oceanic and Atmospheric Administration. p.\u00a011. Retrieved 1 June 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-239\"><b><a href=\"#cite_ref-239\">^<\/a><\/b> \n<cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/essential_background\/items\/2877.php\">\"Essential Background\"<\/a>. United Nations Framework Convention on Climate Change. n.d. Retrieved 18 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-240\"><b><a href=\"#cite_ref-240\">^<\/a><\/b> \n<cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20051028023600\/http:\/\/unfccc.int\/essential_background\/convention\/background\/items\/1353.php\">\"Full text of the Convention, Article 2\"<\/a>. United Nations Framework Convention on Climate Change. n.d. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/essential_background\/convention\/background\/items\/1353.php\">the original<\/a> on 28 October 2005. Retrieved 18 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-241\"><b><a href=\"#cite_ref-241\">^<\/a><\/b> Rogner <i>et al.<\/i>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch1.html\">Chapter 1: Introduction<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/ch1s1-es.html\">Executive summary<\/a>, in <a href=\"#CITEREFIPCC_AR4_WG32007\">IPCC AR4 WG3 2007<\/a>.\n<\/li>\n<li id=\"cite_note-242\"><b><a href=\"#cite_ref-242\">^<\/a><\/b> <cite class=\"citation journal\">Raupach, Michael R.; Marland, Gregg; Ciais, Philippe; Le Qu\u00e9r\u00e9, Corinne; Canadell, Josep G.; Klepper, Gernot; Field, Christopher B. (12 June 2007). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pnas.org\/content\/104\/24\/10288.long\">\"Global and regional drivers of accelerating CO<sub>2<\/sub> emissions\"<\/a> (Free full text). <i>Proceedings of the National Academy of Sciences<\/i>. <b>104<\/b> (24): 10288\u201393. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2007PNAS..10410288R\">2007PNAS..10410288R<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1073%2Fpnas.0700609104\">10.1073\/pnas.0700609104<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Serial_Number\" title=\"International Standard Serial Number\">ISSN<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.worldcat.org\/issn\/0027-8424\">0027-8424<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Central\" title=\"PubMed Central\">PMC<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC1876160\">1876160<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/PubMed_Identifier\" class=\"mw-redirect\" title=\"PubMed Identifier\">PMID<\/a>\u00a0<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.ncbi.nlm.nih.gov\/pubmed\/17519334\">17519334<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-dessai-243\">^ <a href=\"#cite_ref-dessai_243-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-dessai_243-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation web\">Dessai, Suraje (December 2001). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20120610013556\/http:\/\/www.tyndall.ac.uk\/sites\/default\/files\/wp12.pdf\">\"The climate regime from The Hague to Marrakech: Saving or sinking the Kyoto Protocol?\"<\/a> (PDF). <i>Tyndall Centre Working Paper 12<\/i>. Tyndall Centre. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.tyndall.ac.uk\/sites\/default\/files\/wp12.pdf\">the original<\/a> (PDF) on 10 June 2012. Retrieved 5 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-244\"><b><a href=\"#cite_ref-244\">^<\/a><\/b> <cite id=\"CITEREFGrubb,_M.2003\" class=\"citation journal\">Grubb, M. (July\u2013September 2003). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ynccf.net\/pdf\/CDM\/The_economic_of_Kyoto_protocol.pdf\">\"The Economics of the Kyoto Protocol\"<\/a> (PDF). <i>World Economics<\/i>. <b>4<\/b> (3): 144\u201345. Retrieved 25 March 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-kyoto-245\">^ <a href=\"#cite_ref-kyoto_245-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-kyoto_245-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> \n<cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/kyoto_protocol\/items\/2830.php\">\"Kyoto Protocol\"<\/a>. United Nations Framework Convention on Climate Change. n.d. Retrieved 21 May 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-246\"><b><a href=\"#cite_ref-246\">^<\/a><\/b> \n<cite class=\"citation book\">M\u00fcller, Benito (February 2010). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.oxfordenergy.org\/wpcms\/wp-content\/uploads\/2011\/03\/EV49-Copenhagen2009Failureorfinalwake-upcallforourleaders-BenitoMuller-2010.pdf\"><i>Copenhagen 2009: Failure or final wake-up call for our leaders? EV 49<\/i><\/a> (PDF). <a href=\"https:\/\/en.wikipedia.org\/wiki\/Oxford_Institute_for_Energy_Studies\" title=\"Oxford Institute for Energy Studies\">Oxford Institute for Energy Studies<\/a>. p.\u00a0i. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-1-907555-04-6\" title=\"Special:BookSources\/978-1-907555-04-6\">978-1-907555-04-6<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20170710081944\/https:\/\/www.oxfordenergy.org\/wpcms\/wp-content\/uploads\/2011\/03\/EV49-Copenhagen2009Failureorfinalwake-upcallforourleaders-BenitoMuller-2010.pdf\">Archived<\/a> (PDF) from the original on 10 July 2017. Retrieved 18 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-NYT-20150525-247\"><b><a href=\"#cite_ref-NYT-20150525_247-0\">^<\/a><\/b> \n<cite class=\"citation news\">Rudd, Kevin (25 May 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nytimes.com\/2015\/05\/26\/opinion\/kevin-rudd-paris-cant-be-another-copenhagen.html\">\"Paris Can't Be Another Copenhagen\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_New_York_Times\" title=\"The New York Times\">The New York Times<\/a><\/i>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20180203110636\/https:\/\/www.nytimes.com\/2015\/05\/26\/opinion\/kevin-rudd-paris-cant-be-another-copenhagen.html\">Archived<\/a> from the original on 3 February 2018. Retrieved 26 May 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-unep-248\"><b><a href=\"#cite_ref-unep_248-0\">^<\/a><\/b> \n<cite class=\"citation book\">\"Technical summary\". <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20170227095017\/http:\/\/www.unep.org\/publications\/ebooks\/emissionsgapreport\/pdfs\/The_EMISSIONS_GAP_REPORT.pdf\"><i>The Emissions Gap Report: Are the Copenhagen Accord pledges sufficient to limit global warming to 7002275149999999999\u26602\u00a0\u00b0C or 7002274649999999999\u26601.5\u00a0\u00b0C? A preliminary assessment (advance copy)<\/i><\/a> (PDF). United Nations Environment Programme. November 2010. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.unep.org\/publications\/ebooks\/emissionsgapreport\/pdfs\/The_EMISSIONS_GAP_REPORT.pdf\">the original<\/a> (PDF) on 27 February 2017. Retrieved 11 May 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> This publication is also available in <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.unep.org\/publications\/ebooks\/emissionsgapreport\/\">e-book format<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/webarchive.loc.gov\/all\/20101125051326\/http:\/\/www.unep.org\/publications\/ebooks\/emissionsgapreport\/\">Archived<\/a> 25 November 2010 at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Library_of_Congress\" title=\"Library of Congress\">Library of Congress<\/a> Web Archives\n<\/li>\n<li id=\"cite_note-249\"><b><a href=\"#cite_ref-249\">^<\/a><\/b> \n<cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/documentation\/documents\/advanced_search\/items\/3594.php?rec=j&amp;priref=600005735#beg\">\"Decision 2\/CP. 15 Copenhagen Accord. In: Report of the Conference of the Parties on its fifteenth session, held in Copenhagen from 7 to 19\u00a0December\u00a02009. Addendum. Part Two: Action taken by the Conference of the Parties at its fifteenth session\"<\/a> (PDF). United Nations Framework Convention on Climate Change. 30 March 2010. p.\u00a05. Retrieved 17 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-250\"><b><a href=\"#cite_ref-250\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/unfccc.int\/files\/meetings\/cop_16\/application\/pdf\/cop16_lca.pdf\">\"Outcome of the work of the Ad Hoc Working Group on long-term Cooperative Action under the Convention\"<\/a> (PDF). Presidencia De La Rep\u00fablica, M\u00e9xico. 11 December 2010. p.\u00a02. Retrieved 12 January 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-ipcc-251\"><b><a href=\"#cite_ref-ipcc_251-0\">^<\/a><\/b> <cite class=\"citation book\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/report\/ar5\/wg1\/\">\"Summary for Policymakers\"<\/a>. <i>Detection and Attribution of Climate Change<\/i>. IPCC. <q>\u00abIt is extremely likely that human influence has been the dominant cause of the observed warming since the mid-20th century\u00bb (page 15) and \u00abIn this Summary for Policymakers, the following terms have been used to indicate the assessed likelihood of an outcome or a result: (...) extremely likely: 95\u2013100%\u00bb (page 2).<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, in <a href=\"#CITEREFIPCC_AR5_WG12013\">IPCC AR5 WG1 2013<\/a>.\n<\/li>\n<li id=\"cite_note-royal_society_misleading_arguments-252\">^ <a href=\"#cite_ref-royal_society_misleading_arguments_252-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-royal_society_misleading_arguments_252-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> \n<cite class=\"citation book\">Royal Society (13 April 2005). \"Letter from The Royal Society: A Guide to Facts and Fictions About Climate Change: Misleading arguments: Many scientists do not think that climate change is a problem. Some scientists have signed petitions stating that climate change is not a problem. ... There are some individuals and organisations, some of which are funded by the US oil industry, that seek to undermine the science of climate change and the work of the IPCC. They appear motivated in their arguments by opposition to the United Nations Framework Convention on Climate Change and the Kyoto Protocol, which seek urgent action to tackle climate change through a reduction in greenhouse gas emissions.\". <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/publications.parliament.uk\/pa\/ld200506\/ldselect\/ldeconaf\/12\/12we24.htm\"><i>Economic Affairs \u2013 Written Evidence<\/i><\/a>. The Economics of Climate Change, the Second Report of the 2005\u20132006 session, produced by the UK Parliament House of Lords Economics Affairs Select Committee. UK Parliament. Retrieved 9 July 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> This document is also available in <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/royalsociety.org\/Facts-and-fictions-about-climate-change\/\">PDF format<\/a>\n<\/li>\n<li id=\"cite_note-253\"><b><a href=\"#cite_ref-253\">^<\/a><\/b> <cite class=\"citation web\">Academia Brasileira de Ci\u00e9ncias (Brazil); Royal Society of Canada; Chinese Academy of Sciences; Acad\u00e9mie des Sciences (France); Deutsche Akademie der Naturforscher Leopoldina (Germany); Indian National Science Academy; Accademia Nazionale dei Lincei (Italy); Science Council of Japan, Academia Mexicana de Ciencias; Russian Academy of Sciences; Academy of Science of South Africa; Royal Society (United Kingdom); National Academy of Sciences (United States of America) (May 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nationalacademies.org\/includes\/G8+5energy-climate09.pdf\">\"G8+5 Academies' joint statement: Climate change and the transformation of energy technologies for a low carbon future\"<\/a> (PDF). The National Academies of Sciences, Engineering, and Medicine. Retrieved 5 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-254\"><b><a href=\"#cite_ref-254\">^<\/a><\/b>  <cite class=\"citation news\">Davenport, Carol (October 7, 2018). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nytimes.com\/2018\/10\/07\/climate\/ipcc-climate-report-2040.html\">\"Major Climate Report Describes a Strong Risk of Crisis as Early as 2040\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_New_York_Times\" title=\"The New York Times\">The New York Times<\/a><\/i>. Retrieved October 10, 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-255\"><b><a href=\"#cite_ref-255\">^<\/a><\/b> <cite class=\"citation journal\">Cook, John;  et al. (13 April 2016). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/iopscience.iop.org\/article\/10.1088\/1748-9326\/11\/4\/048002\/meta\">\"Consensus on consensus: a synthesis of consensus estimates on human-caused global warming\"<\/a>. <i>Environmental Research Letters<\/i>. <b>11<\/b> (4): 048002. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2016ERL....11d8002C\">2016ERL....11d8002C<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1088%2F1748-9326%2F11%2F4%2F048002\">10.1088\/1748-9326\/11\/4\/048002<\/a>. Retrieved 21 July 2016.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-The_MIT_Press-256\"><b><a href=\"#cite_ref-The_MIT_Press_256-0\">^<\/a><\/b> <cite class=\"citation book\">DiMento, Joseph F. C.; Doughman, Pamela M. (2007). <i>Climate Change: What It Means for Us, Our Children, and Our Grandchildren<\/i>. The MIT Press. p.\u00a068. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-262-54193-0\" title=\"Special:BookSources\/978-0-262-54193-0\">978-0-262-54193-0<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-AQAonAAPG-257\"><b><a href=\"#cite_ref-AQAonAAPG_257-0\">^<\/a><\/b> <cite id=\"CITEREFBrigham-Grette2006\" class=\"citation journal\">Brigham-Grette, Julie;  et al. (September 2006). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/onlinelibrary.wiley.com\/doi\/10.1029\/2006EO360008\/abstract\">\"Petroleum Geologists' Award to Novelist Crichton Is Inappropriate\"<\/a> (PDF). <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Eos_(journal)\" class=\"mw-redirect\" title=\"Eos (journal)\">Eos<\/a><\/i>. <b>87<\/b> (36): 364. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2006EOSTr..87..364B\">2006EOSTr..87..364B<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2006EO360008\">10.1029\/2006EO360008<\/a>. Retrieved 23 January 2007. <q>The AAPG stands alone among scientific societies in its denial of human-induced effects on global warming.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-258\"><b><a href=\"#cite_ref-258\">^<\/a><\/b> <cite class=\"citation journal\">Ripple, William J.; Wolf, Christopher; Newsome, Thomas M.; Galetti, Mauro; Alamgir, Mohammed; Crist, Eileen; Mahmoud, Mahmoud I.; Laurance, William F. (13 November 2017). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/academic.oup.com\/bioscience\/article\/doi\/10.1093\/biosci\/bix125\/4605229\">\"World Scientists' Warning to Humanity: A Second Notice\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/BioScience\" title=\"BioScience\">BioScience<\/a><\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1093%2Fbiosci%2Fbix125\">10.1093\/biosci\/bix125<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-259\"><b><a href=\"#cite_ref-259\">^<\/a><\/b> <cite class=\"citation news\">Perkins, Sid (11 July 2017). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.sciencemag.org\/news\/2017\/07\/best-way-reduce-your-carbon-footprint-one-government-isn-t-telling-you-about\">\"The best way to reduce your carbon footprint is one the government isn't telling you about\"<\/a>. <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Science_(journal)\" title=\"Science (journal)\">Science<\/a><\/i>. Retrieved 29 November 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-260\"><b><a href=\"#cite_ref-260\">^<\/a><\/b> <cite class=\"citation journal\">Boykoff, Maxwell T.; <a href=\"https:\/\/en.wikipedia.org\/wiki\/Jules_Boykoff\" title=\"Jules Boykoff\">Boykoff, Jules M.<\/a> (July 2004). \"Balance as bias: global warming and the US prestige press\". <i>Global Environmental Change Part A<\/i>. <b>14<\/b> (2): 125\u201336. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1016%2Fj.gloenvcha.2003.10.001\">10.1016\/j.gloenvcha.2003.10.001<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-merchants_doubt-261\"><b><a href=\"#cite_ref-merchants_doubt_261-0\">^<\/a><\/b> <cite class=\"citation book\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Naomi_Oreskes\" title=\"Naomi Oreskes\">Oreskes, Naomi<\/a>; Conway, Erik. <i>Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming<\/i> (first ed.). Bloomsbury Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-1-59691-610-4\" title=\"Special:BookSources\/978-1-59691-610-4\">978-1-59691-610-4<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-262\"><b><a href=\"#cite_ref-262\">^<\/a><\/b> McCright, Aaron M.; Dunlap, Riley E., \"Challenging Global Warming as a Social Problem: An Analysis of the Conservative Movement's Counter-Claims\", <i>Social Problems<\/i>, November 2000, Vol. 47 Issue 4, pp 499\u2013522 <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.jstor.org\/pss\/3097132\">in JSTOR<\/a>\n<\/li>\n<li id=\"cite_note-263\"><b><a href=\"#cite_ref-263\">^<\/a><\/b> <cite class=\"citation web\">Weart, S. (July 2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.aip.org\/history\/climate\/public2.htm\">\"The Public and Climate Change (cont.\u00a0\u2013 since 1980). Section: After 1988\"<\/a>. American Institute of Physics. Retrieved 5 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><ul><li><cite class=\"citation web\">SEPP (n.d.). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20080511011611\/http:\/\/www.sepp.org\/FAQ\/faq.html\">\"Frequently Asked Questions About Climate Change\"<\/a>. Science &amp; Environmental Policy Project (SEPP). Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.sepp.org\/FAQ\/faq.html\">the original<\/a> on 11 May 2008. Retrieved 5 May 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li><\/ul><\/li>\n<li id=\"cite_note-264\"><b><a href=\"#cite_ref-264\">^<\/a><\/b> <cite class=\"citation news\">Begley, Sharon (13 August 2007). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.newsweek.com\/id\/32482\">\"The Truth About Denial\"<\/a>. <i>Newsweek<\/i>. Retrieved 13 August 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-265\"><b><a href=\"#cite_ref-265\">^<\/a><\/b> <cite class=\"citation news\">Adams, David (20 September 2006). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.theguardian.com\/environment\/2006\/sep\/20\/oilandpetrol.business\">\"Royal Society tells Exxon: stop funding climate change denial\"<\/a>. <i>The Guardian<\/i>. London. Retrieved 9 August 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-MSNBC_01-12-266\"><b><a href=\"#cite_ref-MSNBC_01-12_266-0\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20070618152340\/http:\/\/www.msnbc.msn.com\/id\/16593606\">\"Exxon cuts ties to global warming skeptics\"<\/a>. MSNBC. 12 January 2007. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.msnbc.msn.com\/id\/16593606\">the original<\/a> on 18 June 2007. Retrieved 2 May 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-267\"><b><a href=\"#cite_ref-267\">^<\/a><\/b> <cite class=\"citation news\">Sandell, Clayton (3 January 2007). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/abcnews.go.com\/Technology\/Business\/story?id=2767979&amp;page=1\">\"Report: Big Money Confusing Public on Global Warming\"<\/a>. ABC. Retrieved 27 April 2007.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-268\"><b><a href=\"#cite_ref-268\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.usatoday.com\/weather\/climate\/globalwarming\/2007-05-18-greenpeace-exxon_N.htm\">\"Greenpeace: Exxon still funding climate skeptics\"<\/a>. <i>USA Today<\/i>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Reuters\" title=\"Reuters\">Reuters<\/a>. 18 May 2007. Retrieved 21 January 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-269\"><b><a href=\"#cite_ref-269\">^<\/a><\/b> <cite class=\"citation pressrelease\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.csrwire.com\/press\/press_release\/23395-Global-Warming-Resolutions-at-U-S-Oil-Companies-Bring-Policy-Commitments-from-Leaders-and-Record-High-Votes-at-Laggards\">\"Global Warming Resolutions at U.S. Oil Companies Bring Policy Commitments from Leaders, and Record High Votes at Laggards\"<\/a> (Press release). Ceres. 13 May 2004. Retrieved 4 March 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-270\"><b><a href=\"#cite_ref-270\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.uwosh.edu\/es\/climate-change\/oil-company-positions-on-the-reality-and-risk-of-climate-change\">\"Oil Company Positions on the Reality and Risk of Climate Change\"<\/a>. Environmental Studies, University of Oshkosh \u2013 Wisconsin. Retrieved 27 March 2016.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-271\"><b><a href=\"#cite_ref-271\">^<\/a><\/b> <cite class=\"citation web\">Weart, S. (February 2015). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/history.aip.org\/climate\/public2.htm#L_0446\">\"The Public and Climate Change (cont.\u00a0\u2013 since 1980). Section: after 1988\"<\/a>. American Institute of Physics. Retrieved 18 August 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-272\"><b><a href=\"#cite_ref-272\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.gallup.com\/poll\/1615\/environment.aspx\">\"Environment\"<\/a>. Gallup. 2015. Retrieved 18 August 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-Newell2006-273\"><b><a href=\"#cite_ref-Newell2006_273-0\">^<\/a><\/b> <cite class=\"citation book\">Peter Newell (December 14, 2006). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/books.google.com\/books?id=ing21MGmh5UC\"><i>Climate for Change: Non-State Actors and the Global Politics of the Greenhouse<\/i><\/a>. Cambridge University Press. p.\u00a080. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-02123-4\" title=\"Special:BookSources\/978-0-521-02123-4\">978-0-521-02123-4<\/a>. Retrieved July 30, 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-274\"><b><a href=\"#cite_ref-274\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.yaleclimateconnections.org\/2010\/11\/communicating-with-american-public\">\"Yale Researcher Anthony Leiserowitz On Studying, Communicating with American Public\"<\/a>. Yale Climate Connections. 2010.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-275\"><b><a href=\"#cite_ref-275\">^<\/a><\/b> <cite id=\"CITEREFShindellFaluvegiLacisHansen2006\" class=\"citation journal\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Drew_Shindell\" title=\"Drew Shindell\">Shindell, Drew<\/a>; Faluvegi, Greg; Lacis, Andrew; <a href=\"https:\/\/en.wikipedia.org\/wiki\/James_Hansen\" title=\"James Hansen\">Hansen, James<\/a>; Ruedy, Reto; Aguilar, Elliot (28 April 2006). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.image.ucar.edu\/idag\/Papers\/Shindell_tropospheric_ozone.pdf\">\"Role of tropospheric ozone increases in 20th-century climate change\"<\/a> (PDF). <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Journal_of_Geophysical_Research\" title=\"Journal of Geophysical Research\">Journal of Geophysical Research<\/a><\/i>. <b>111<\/b> (D8): D08302. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2006JGRD..11108302S\">2006JGRD..11108302S<\/a>. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1029%2F2005JD006348\">10.1029\/2005JD006348<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-276\"><b><a href=\"#cite_ref-276\">^<\/a><\/b> <cite class=\"citation news\"><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.nytimes.com\/interactive\/2018\/08\/01\/magazine\/climate-change-losing-earth.html\">\"Losing Earth: The Decade We Almost Stopped Climate Change, chapter 2 You Scientists Win 1985\"<\/a>. <i>The New York Times<\/i>. 1 August 2018.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-277\"><b><a href=\"#cite_ref-277\">^<\/a><\/b> <cite class=\"citation web\">Pugliese, Anita (20 April 2011). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.gallup.com\/poll\/147203\/Fewer-Americans-Europeans-View-Global-Warming-Threat.aspx\">\"Fewer Americans, Europeans View Global Warming as a Threat\"<\/a>. Gallup. Retrieved 22 April 2011.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-278\"><b><a href=\"#cite_ref-278\">^<\/a><\/b> \n<cite class=\"citation web\">Ray, Julie; Anita Pugliese (22 April 2011). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.gallup.com\/poll\/147242\/Worldwide-Blame-Climate-Change-Falls-Humans.aspx\">\"Worldwide, Blame for Climate Change Falls on Humans\"<\/a>. <i>Gallup.Com<\/i>. Retrieved 3 May 2011. <q>People nearly everywhere, including majorities in developed Asia and Latin America, are more likely to attribute global warming to human activities rather than natural causes. The U.S. is the exception, with nearly half (47%) \u2013 and the largest percentage in the world \u2013 attributing global warming to natural causes.<\/q><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-279\"><b><a href=\"#cite_ref-279\">^<\/a><\/b> <cite class=\"citation web\"><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.pewglobal.org\/2013\/06\/24\/climate-change-and-financial-instability-seen-as-top-global-threats\/\">\"Climate Change and Financial Instability Seen as Top Global Threats\"<\/a>. Pew Research Center for the People &amp; the Press. 24 June 2013.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-Weart_27-280\">^ <a href=\"#cite_ref-Weart_27_280-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-Weart_27_280-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation web\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Spencer_R._Weart\" title=\"Spencer R. Weart\">Weart, Spencer R.<\/a> (February 2014). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/history.aip.org\/climate\/public.htm#S2\">\"The Discovery of Global Warming; The Public and Climate Change: Suspicions of a Human-Caused Greenhouse (1956\u20131969)\"<\/a>. American Institute of Physics. Retrieved 12 May 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, and <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.aip.org\/history\/climate\/public.htm#M_27_\">footnote 27<\/a>\n<\/li>\n<li id=\"cite_note-Weart_Hansen-281\">^ <a href=\"#cite_ref-Weart_Hansen_281-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-Weart_Hansen_281-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <cite class=\"citation web\">Weart, Spencer R. (February 2014). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/history.aip.org\/climate\/public2.htm#L_0575\">\"The Discovery of Global Warming; The Public and Climate Change: The Summer of 1988\"<\/a>. American Institute of Physics. Retrieved 12 May 2015.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li id=\"cite_note-nasacbaon-282\">^ <a href=\"#cite_ref-nasacbaon_282-0\"><sup><i><b>a<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-nasacbaon_282-1\"><sup><i><b>b<\/b><\/i><\/sup><\/a> <a href=\"#cite_ref-nasacbaon_282-2\"><sup><i><b>c<\/b><\/i><\/sup><\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Erik_M._Conway\" title=\"Erik M. Conway\">Erik Conway<\/a>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nasa.gov\/topics\/earth\/features\/climate_by_any_other_name.html\">\"What's in a Name? Global Warming vs. Climate Change\"<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/NASA\" title=\"NASA\">NASA<\/a>, 5 December 2008\n<\/li>\n<li id=\"cite_note-283\"><b><a href=\"#cite_ref-283\">^<\/a><\/b> U.S. Senate, Committee on Energy and Natural Resources, \"Greenhouse Effect and Global Climate Change, part 2\" 100th Cong., 1st sess., 23 June 1988, p. 44.\n<\/li>\n<\/ol><\/div>\n<h2>References<\/h2>\n<div class=\"refbegin columns references-column-width\">\n<ul><li><cite id=\"CITEREFGoodothers2010\" class=\"citation\">Good, P.;  et al. (2010), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/archive.theccc.org.uk\/aws2\/4th%20Budget\/fourthbudget_supporting_research_reviewofclimatescience_developements_since_IPPC_AR4.pdf\"><i>An updated review of developments in climate science research since IPCC AR4. A report by the AVOID consortium<\/i><\/a> (PDF), London, UK: Committee on Climate Change<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, p.\u00a014. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.theccc.org.uk\/publication\/the-fourth-carbon-budget-reducing-emissions-through-the-2020s-2\/\">Report website.<\/a><\/li>\n<li><cite id=\"CITEREFIAP2009\" class=\"citation\">IAP (June 2009), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.interacademies.net\/10878\/13951.aspx\"><i>Interacademy Panel (IAP) Member Academies Statement on Ocean Acidification<\/i><\/a><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>, Secretariat: TWAS (the Academy of Sciences for the Developing World), Trieste, Italy.<\/li>\n<li><cite id=\"CITEREFIEA2009\" class=\"citation book\">IEA (2009). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.worldenergyoutlook.org\/media\/weowebsite\/2009\/WEO2009.pdf\"><i>World Energy Outlook 2009<\/i><\/a> (PDF). Paris, France: International Energy Agency (IEA). <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-92-64-06130-9\" title=\"Special:BookSources\/978-92-64-06130-9\">978-92-64-06130-9<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite id=\"CITEREFIPCC_AR4_SYR2007\" class=\"citation book\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">IPCC AR4 SYR<\/a> (2007).  Core Writing Team; Pachauri, R.K; Reisinger, A., eds. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/syr\/en\/contents.html\"><i>Climate Change 2007: Synthesis Report<\/i><\/a>. Contribution of Working Groups I, II and III to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Fourth_Assessment_Report\" title=\"IPCC Fourth Assessment Report\">Fourth Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. IPCC. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/92-9169-122-4\" title=\"Special:BookSources\/92-9169-122-4\">92-9169-122-4<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite id=\"CITEREFIPCC_AR4_WG12007\" class=\"citation book\">IPCC AR4 WG1 (2007).  Solomon, S.; Qin, D.; Manning, M.; Chen, Z.; Marquis, M.; Averyt, K.B.; Tignor, M.; Miller, H.L., eds. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg1\/en\/contents.html\"><i>Climate Change 2007: The Physical Science Basis<\/i><\/a>. Contribution of Working Group I to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Fourth_Assessment_Report\" title=\"IPCC Fourth Assessment Report\">Fourth Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-88009-1\" title=\"Special:BookSources\/978-0-521-88009-1\">978-0-521-88009-1<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-70596-7\" title=\"Special:BookSources\/978-0-521-70596-7\">978-0-521-70596-7<\/a>)<\/li>\n<li><cite id=\"CITEREFIPCC_AR4_WG22007\" class=\"citation book\">IPCC AR4 WG2 (2007).  Parry, M.L.; Canziani, O.F.; Palutikof, J.P.; van der Linden, P.J.; Hanson, C.E., eds. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg2\/en\/contents.html\"><i>Climate Change 2007: Impacts, Adaptation and Vulnerability<\/i><\/a>. Contribution of Working Group II to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Fourth_Assessment_Report\" title=\"IPCC Fourth Assessment Report\">Fourth Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-88010-7\" title=\"Special:BookSources\/978-0-521-88010-7\">978-0-521-88010-7<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-70597-4\" title=\"Special:BookSources\/978-0-521-70597-4\">978-0-521-70597-4<\/a>)<\/li>\n<li><cite id=\"CITEREFIPCC_AR4_WG32007\" class=\"citation book\">IPCC AR4 WG3 (2007).  Metz, B.; Davidson, O.R.; Bosch, P.R.; Dave, R.; Meyer, L.A., eds. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/ar4\/wg3\/en\/contents.html\"><i>Climate Change 2007: Mitigation of Climate Change<\/i><\/a>. Contribution of Working Group III to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Fourth_Assessment_Report\" title=\"IPCC Fourth Assessment Report\">Fourth Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-88011-4\" title=\"Special:BookSources\/978-0-521-88011-4\">978-0-521-88011-4<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-70598-1\" title=\"Special:BookSources\/978-0-521-70598-1\">978-0-521-70598-1<\/a>)<\/li>\n<li><cite id=\"CITEREFIPCC_AR5_WG12013\" class=\"citation\">IPCC AR5 WG1 (2013),  Stocker, T.F.;  et al., eds., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/report\/ar5\/wg1\/\"><i>Climate Change 2013: The Physical Science Basis. Working Group 1 (WG1) Contribution to the Intergovernmental Panel on Climate Change (IPCC) 5th Assessment Report (AR5)<\/i><\/a>, Cambridge University Press<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climatechange2013.org\/\">Climate Change 2013 Working Group 1 website.<\/a><\/li>\n<li><cite id=\"CITEREFIPCC_AR5_WG2_A2014\" class=\"citation\">IPCC AR5 WG2 A (2014),  Field, C.B.;  et al., eds., <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/report\/ar5\/wg2\/\"><i>Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects (GSA). Contribution of Working Group II (WG2) to the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC)<\/i><\/a>, Cambridge University Press, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140416051047\/http:\/\/www.ipcc.ch\/report\/ar5\/wg2\/\">archived<\/a> from the original on 16 April 2014<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Archived<\/li>\n<li><cite id=\"CITEREFIPCC_AR5_WG32014\" class=\"citation\">IPCC AR5 WG3 (2014),  Edenhofer, O.;  et al., eds., <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141127222605\/http:\/\/www.ipcc.ch\/report\/ar5\/wg3\/\"><i>Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III (WG3) to the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC)<\/i><\/a>, Cambridge University Press, archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/report\/ar5\/wg3\/\">the original<\/a> on 27 November 2014<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Also available at <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20141230092610\/http:\/\/mitigation2014.org\/report\/publication\/\">mitigation2014.org<\/a>.<\/li>\n<li><cite id=\"CITEREFIPCC_SAR_SYR1996\" class=\"citation journal\">IPCC SAR SYR (1996). \"Climate Change 1995: A report of the Intergovernmental Panel on Climate Change\". <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Second_Assessment_Report\" title=\"IPCC Second Assessment Report\">Second Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. IPCC.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/pdf\/climate-changes-1995\/ipcc-2nd-assessment\/2nd-assessment-en.pdf\">pdf<\/a>. <small>The \"Full Report\", consisting of \"The IPCC Second Assessment Synthesis of Scientific-Technical Information Relevant to Interpreting Article 2 of the UN Framework Convention on Climate Change\" and the Summaries for Policymakers of the three Working Groups.<\/small><\/li>\n<li><cite id=\"CITEREFIPCC_SAR_WG31996\" class=\"citation book\">IPCC SAR WG3 (1996).  Bruce, J.P.; Lee, H.; Haites, E.F., eds. <i>Climate Change 1995: Economic and Social Dimensions of Climate Change<\/i>. Contribution of Working Group III to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Second_Assessment_Report\" title=\"IPCC Second Assessment Report\">Second Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-56051-9\" title=\"Special:BookSources\/0-521-56051-9\">0-521-56051-9<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-56854-4\" title=\"Special:BookSources\/0-521-56854-4\">0-521-56854-4<\/a>) <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/ipccreports\/sar\/wg_III\/ipcc_sar_wg_III_full_report.pdf\">pdf<\/a>.<\/li>\n<li><cite id=\"CITEREFIPCC_SREX2012\" class=\"citation journal\">IPCC SREX (2012).  Field, C.B.;  et al., eds. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20121219055946\/http:\/\/ipcc-wg2.gov\/SREX\/report\/full-report\/\">\"Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX)\"<\/a>. Cambridge University Press. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ipcc-wg2.gov\/SREX\/report\/full-report\/\">the original<\/a> on 19 December 2012.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Summary for Policymakers <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ipcc.ch\/publications_and_data\/publications_and_data_reports.shtml\">Summary for Policymakers<\/a>.<\/li>\n<li><cite id=\"CITEREFIPCC_TAR_WG12001\" class=\"citation book\">IPCC TAR WG1 (2001).  Houghton, J.T.; Ding, Y.; Griggs, D.J.; Noguer, M.; van der Linden, P.J.; Dai, X.; Maskell, K.; Johnson, C.A., eds. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20160330214626\/http:\/\/grida.no\/publications\/other\/ipcc_tar\/?src=%2Fclimate%2Fipcc_tar%2Fwg1%2Findex.htm\"><i>Climate Change 2001: The Scientific Basis<\/i><\/a>. Contribution of Working Group I to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Third_Assessment_Report\" title=\"IPCC Third Assessment Report\">Third Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-80767-0\" title=\"Special:BookSources\/0-521-80767-0\">0-521-80767-0<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/publications\/other\/ipcc%5Ftar\/?src=\/climate\/ipcc_tar\/wg1\/index.htm\">the original<\/a> on 30 March 2016.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-01495-6\" title=\"Special:BookSources\/0-521-01495-6\">0-521-01495-6<\/a>)<\/li>\n<li><cite id=\"CITEREFIPCC_TAR_WG22001\" class=\"citation book\">IPCC TAR WG2 (2001).  McCarthy, J. J.; Canziani, O. F.; Leary, N. A.; Dokken, D. J.; White, K. S., eds. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20160514070847\/http:\/\/www.grida.no\/publications\/other\/ipcc_tar\/?src=%2Fclimate%2Fipcc_tar%2Fwg2%2Findex.htm\"><i>Climate Change 2001: Impacts, Adaptation and Vulnerability<\/i><\/a>. Contribution of Working Group II to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Third_Assessment_Report\" title=\"IPCC Third Assessment Report\">Third Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-80768-9\" title=\"Special:BookSources\/0-521-80768-9\">0-521-80768-9<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/publications\/other\/ipcc%5Ftar\/?src=\/climate\/ipcc_tar\/wg2\/index.htm\">the original<\/a> on 14 May 2016.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-01500-6\" title=\"Special:BookSources\/0-521-01500-6\">0-521-01500-6<\/a>)<\/li>\n<li><cite id=\"CITEREFIPCC_TAR_WG32001\" class=\"citation book\">IPCC TAR WG3 (2001).  Metz, B.; Davidson, O.; Swart, R.; Pan, J., eds. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20170227094845\/http:\/\/www.grida.no\/publications\/other\/ipcc_tar\/?src=%2Fclimate%2Fipcc_tar%2Fwg3%2Findex.htm\"><i>Climate Change 2001: Mitigation<\/i><\/a>. Contribution of Working Group III to the <a href=\"https:\/\/en.wikipedia.org\/wiki\/IPCC_Third_Assessment_Report\" title=\"IPCC Third Assessment Report\">Third Assessment Report<\/a> of the Intergovernmental Panel on Climate Change. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-80769-7\" title=\"Special:BookSources\/0-521-80769-7\">0-521-80769-7<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.grida.no\/publications\/other\/ipcc%5Ftar\/?src=\/climate\/ipcc_tar\/wg3\/index.htm\">the original<\/a> on 27 February 2017.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> (pb: <link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-521-01502-2\" title=\"Special:BookSources\/0-521-01502-2\">0-521-01502-2<\/a>)<\/li>\n<li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/12px-PD-icon.svg.png\" width=\"12\" height=\"12\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/18px-PD-icon.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/24px-PD-icon.svg.png 2x\" data-file-width=\"196\" data-file-height=\"196\" \/>\u00a0This article incorporates <a href=\"https:\/\/en.wikipedia.org\/wiki\/Copyright_status_of_work_by_the_U.S._government\" title=\"Copyright status of work by the U.S. government\">public domain material<\/a>\u00a0from the\u00a0US Global Change Research Program (<a href=\"https:\/\/en.wikipedia.org\/wiki\/USGCRP\" class=\"mw-redirect\" title=\"USGCRP\">USGCRP<\/a>) document:\u00a0<cite id=\"CITEREFNCADAC2013\" class=\"citation\">NCADAC (11 January 2013), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/ncadac.globalchange.gov\/\"><i>Federal Advisory Committee Draft Climate Assessment. A report by the National Climate Assessment Development Advisory Committee (NCADAC)<\/i><\/a>, Washington, DC<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite id=\"CITEREFUSGCRP2015\" class=\"citation\">USGCRP (2015), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.globalchange.gov\/climate-change\/glossary\"><i>Glossary<\/i><\/a>, Washington, DC: U.S. Global Change Research Program (USGCRP), retrieved 20 January 2014<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/*\/http:\/\/www.globalchange.gov\/climate-change\/glossary\">Archived url<\/a>.<\/li>\n<li><cite id=\"CITEREFNational_Research_Council2011\" class=\"citation\">National Research Council (2011), <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140327000317\/http:\/\/www.nap.edu\/catalog.php?record_id=12877\"><i>Climate Stabilization Targets: Emissions, Concentrations, and Impacts over Decades to Millennia<\/i><\/a>, Washington, DC: National Academies Press, archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nap.edu\/catalog.php?record_id=12877\">the original<\/a> on 27 March 2014<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite id=\"CITEREFNational_Research_Council2010\" class=\"citation book\">National Research Council (2010). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20140529161102\/http:\/\/www.nap.edu\/catalog.php?record_id=12782\"><i>America's Climate Choices: Panel on Advancing the Science of Climate Change;<\/i><\/a>. Washington, DC: The National Academies Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/0-309-14588-0\" title=\"Special:BookSources\/0-309-14588-0\">0-309-14588-0<\/a>. Archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nap.edu\/catalog.php?record_id=12782\">the original<\/a> on 29 May 2014.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite id=\"CITEREFParrisothers2012\" class=\"citation\">Parris, A.;  et al. (6 December 2012), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/cpo.noaa.gov\/sites\/cpo\/Reports\/2012\/NOAA_SLR_r3.pdf\"><i>Global Sea Level Rise Scenarios for the US National Climate Assessment. NOAA Tech Memo OAR CPO-1<\/i><\/a> (PDF), NOAA Climate Program Office<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Report <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/cpo.noaa.gov\/AllNews\/TabId\/315\/ArtMID\/668\/ArticleID\/80\/Global-Sea-Level-Rise-Scenarios-for-the-United-States-National-Climate-Assessment.aspx\">website.<\/a><\/li>\n<li><cite id=\"CITEREFUNEP2010\" class=\"citation\">UNEP (2010), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/webarchive.loc.gov\/all\/20150407111527\/http:\/\/unep.org\/dewa\/Portals\/67\/pdf\/Ocean_Acidification.pdf\"><i>UNEP Emerging Issues: Environmental Consequences of Ocean Acidification: A Threat to Food Security<\/i><\/a> (PDF), Nairobi, Kenya: United Nations Environment Programme (UNEP), archived from <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.unep.org\/dewa\/Portals\/67\/pdf\/Ocean_Acidification.pdf\">the original<\/a> (PDF) on 7 April 2015<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20131205011808\/http:\/\/www.unepmap.org\/index.php?module=news\">Report summary.<\/a><\/li>\n<li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/12px-PD-icon.svg.png\" width=\"12\" height=\"12\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/18px-PD-icon.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/6\/62\/PD-icon.svg\/24px-PD-icon.svg.png 2x\" data-file-width=\"196\" data-file-height=\"196\" \/>\u00a0This article incorporates <a href=\"https:\/\/en.wikipedia.org\/wiki\/Copyright_status_of_work_by_the_U.S._government\" title=\"Copyright status of work by the U.S. government\">public domain material<\/a>\u00a0from the\u00a0US Global Change Research Program (<a href=\"https:\/\/en.wikipedia.org\/wiki\/USGCRP\" class=\"mw-redirect\" title=\"USGCRP\">USGCRP<\/a>) document:\u00a0<cite id=\"CITEREFUSGCRP2009\" class=\"citation book\">USGCRP (2009).  Karl, T.R.; Melillo. J.; Peterson, T.; Hassol, S.J., eds. <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.globalchange.gov\/publications\/reports\/scientific-assessments\/us-impacts\"><i>Global Climate Change Impacts in the United States<\/i><\/a>. Cambridge University Press. <a href=\"https:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\" title=\"International Standard Book Number\">ISBN<\/a>\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Special:BookSources\/978-0-521-14407-0\" title=\"Special:BookSources\/978-0-521-14407-0\">978-0-521-14407-0<\/a>.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>. Public-domain status of this report can be found on p. 4 of <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/library.globalchange.gov\/downloads\/download.php?id=102\">PDF<\/a><\/li>\n<li><cite id=\"CITEREFUS_NRC2008\" class=\"citation journal\">US NRC (2008). <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/dels.nas.edu\/Materials\/Booklets\/Lines-of-Evidence\">\"Understanding and responding to climate change: Highlights of National Academies Reports, 2008 edition, produced by the US National Research Council (US NRC)\"<\/a>. Washington, DC: National Academy of Sciences.<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/><\/li>\n<li><cite id=\"CITEREFUS_NRC2012\" class=\"citation journal\">US NRC (2012). <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.scribd.com\/doc\/98458016\/Climate-Change-Lines-of-Evidence\">\"Climate Change: Evidence, Impacts, and Choices\"<\/a>. US National Research Council (US NRC).<\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/> Also available as <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/nas-sites.org\/americasclimatechoices\/files\/2012\/06\/19014_cvtx_R1.pdf\">PDF<\/a><\/li>\n<li><cite id=\"CITEREFZeebe2012\" class=\"citation\">Zeebe, R.E. (May 2012), <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.soest.hawaii.edu\/oceanography\/faculty\/zeebe_files\/Publications\/ZeebeAR12.pdf\">\"History of Seawater Carbonate Chemistry, Atmospheric CO<sub>2<\/sub>, and Ocean Acidification\"<\/a> (PDF), <i>Annual Review of Earth and Planetary Sciences<\/i>, <b>40<\/b>, pp.\u00a0141\u201365, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bibcode\" title=\"Bibcode\">Bibcode<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/adsabs.harvard.edu\/abs\/2012AREPS..40..141Z\">2012AREPS..40..141Z<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1146%2Fannurev-earth-042711-105521\">10.1146\/annurev-earth-042711-105521<\/a><\/cite><link rel=\"mw-deduplicated-inline-style\" href=\"denied:mw-data:TemplateStyles:r861714446\" \/>.<\/li><\/ul><\/div>\n<h2>External links<\/h2>\n<div role=\"navigation\" aria-labelledby=\"sister-projects\" class=\"metadata plainlinks plainlist mbox-small\">\n    <div>Find more about<b>Global warming<\/b>at Wikipedia's <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wikipedia:Wikimedia_sister_projects\" title=\"Wikipedia:Wikimedia sister projects\">sister projects<\/a>\n    <\/div>\n    <ul><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/0\/06\/Wiktionary-logo-v2.svg\/27px-Wiktionary-logo-v2.svg.png\" width=\"27\" height=\"27\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/0\/06\/Wiktionary-logo-v2.svg\/41px-Wiktionary-logo-v2.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/0\/06\/Wiktionary-logo-v2.svg\/54px-Wiktionary-logo-v2.svg.png 2x\" data-file-width=\"391\" data-file-height=\"391\" \/><a href=\"https:\/\/en.wiktionary.org\/wiki\/Global_warming\" class=\"extiw\" title=\"wikt:Global warming\">Definitions<\/a> from Wiktionary\n            <\/li><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/4a\/Commons-logo.svg\/20px-Commons-logo.svg.png\" width=\"20\" height=\"27\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/4a\/Commons-logo.svg\/30px-Commons-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/4a\/Commons-logo.svg\/40px-Commons-logo.svg.png 2x\" data-file-width=\"1024\" data-file-height=\"1376\" \/><a href=\"https:\/\/commons.wikimedia.org\/wiki\/Category:Global_warming\" class=\"extiw\" title=\"c:Category:Global warming\">Media<\/a> from Wikimedia Commons\n            <\/li><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/2\/24\/Wikinews-logo.svg\/27px-Wikinews-logo.svg.png\" width=\"27\" height=\"15\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/2\/24\/Wikinews-logo.svg\/41px-Wikinews-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/2\/24\/Wikinews-logo.svg\/54px-Wikinews-logo.svg.png 2x\" data-file-width=\"759\" data-file-height=\"415\" \/><a href=\"https:\/\/en.wikinews.org\/wiki\/Special:Search\/Global_warming\" class=\"extiw\" title=\"n:Special:Search\/Global warming\">News<\/a> from Wikinews\n            <\/li><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/fa\/Wikiquote-logo.svg\/23px-Wikiquote-logo.svg.png\" width=\"23\" height=\"27\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/fa\/Wikiquote-logo.svg\/35px-Wikiquote-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/fa\/Wikiquote-logo.svg\/46px-Wikiquote-logo.svg.png 2x\" data-file-width=\"300\" data-file-height=\"355\" \/><a href=\"https:\/\/en.wikiquote.org\/wiki\/Global_warming\" class=\"extiw\" title=\"q:Global warming\">Quotations<\/a> from Wikiquote\n            <\/li><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/4\/4c\/Wikisource-logo.svg\/26px-Wikisource-logo.svg.png\" width=\"26\" height=\"27\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/4\/4c\/Wikisource-logo.svg\/39px-Wikisource-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/4\/4c\/Wikisource-logo.svg\/51px-Wikisource-logo.svg.png 2x\" data-file-width=\"410\" data-file-height=\"430\" \/><a href=\"https:\/\/en.wikisource.org\/wiki\/Global_warming\" class=\"extiw\" title=\"s:Global warming\">Texts<\/a> from Wikisource\n            <\/li><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/fa\/Wikibooks-logo.svg\/27px-Wikibooks-logo.svg.png\" width=\"27\" height=\"27\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/fa\/Wikibooks-logo.svg\/41px-Wikibooks-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/fa\/Wikibooks-logo.svg\/54px-Wikibooks-logo.svg.png 2x\" data-file-width=\"300\" data-file-height=\"300\" \/><a href=\"https:\/\/en.wikibooks.org\/wiki\/Climate_Change\" class=\"extiw\" title=\"b:Climate Change\">Textbooks<\/a> from Wikibooks\n            <\/li><li><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1b\/Wikiversity-logo-en.svg\/27px-Wikiversity-logo-en.svg.png\" width=\"27\" height=\"24\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1b\/Wikiversity-logo-en.svg\/41px-Wikiversity-logo-en.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1b\/Wikiversity-logo-en.svg\/54px-Wikiversity-logo-en.svg.png 2x\" data-file-width=\"1000\" data-file-height=\"900\" \/><a href=\"https:\/\/en.wikiversity.org\/wiki\/Global_warming\" class=\"extiw\" title=\"v:Global warming\">Learning resources<\/a> from Wikiversity\n            <\/li>\n    <\/ul><\/div>\n<table role=\"presentation\" class=\"mbox-small plainlinks sistersitebox\"><tbody><tr><td class=\"mbox-image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/9\/91\/Wikiversity-logo.svg\/40px-Wikiversity-logo.svg.png\" width=\"40\" height=\"32\" class=\"noviewer\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/9\/91\/Wikiversity-logo.svg\/60px-Wikiversity-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/9\/91\/Wikiversity-logo.svg\/80px-Wikiversity-logo.svg.png 2x\" data-file-width=\"1000\" data-file-height=\"800\" \/><\/td>\n<td class=\"mbox-text plainlist\"><a href=\"https:\/\/en.wikiversity.org\/wiki\/How_things_work_college_course\/Global_warming_quizzes\" class=\"extiw\" title=\"v:How things work college course\/Global warming quizzes\">Wikiversity hosts a testbank of quizzes to assess students' understanding of this article<\/a><\/td><\/tr><\/tbody><\/table><h3>Research<\/h3>\n<ul><li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.giss.nasa.gov\/\">NASA Goddard Institute for Space Studies<\/a> \u2013 Global change research<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ncdc.noaa.gov\/sotc\">NOAA State of the Climate Report<\/a> \u2013 U.S. and global monthly state of the climate reports<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/dels.nas.edu\/Climate\/Climate-Change\/Reports-Academies-Findings\">Climate Change at the National Academies<\/a>\u00a0\u2013 repository for reports<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.nature.com\/nclimate\/index.html\">Nature Reports Climate Change<\/a> \u2013 free-access web resource<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.metoffice.gov.uk\/climate-guide\">Met Office: Climate Guide<\/a> \u2013 UK National Weather Service<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/edgcm.columbia.edu\/\">Educational Global Climate Modelling<\/a> (EdGCM) \u2013 research-quality climate change simulator<\/li><\/ul><h3>Educational<\/h3>\n<ul><li><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/science2017.globalchange.gov\/\">Climate Science Special Report<\/a> United States 2017<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/climate.nasa.gov\/evidence\/\">NASA: Climate change: How do we know?<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/climate.jpl.nasa.gov\/\">Global Climate Change: NASA's Eyes on the Earth<\/a> \u2013 NASA, JPL, Caltech<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ncdc.noaa.gov\/indicators\/\">Global Climate Change Indicators<\/a> \u2013 NOAA<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.climate.gov\/#understandingClimate\">NOAA Climate Services<\/a> \u2013 NOAA<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.skepticalscience.com\/\">Skeptical Science: Getting skeptical about global warming skepticism<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20170517034730\/https:\/\/www2.ucar.edu\/global-warming-climate-change\/faq\">Understanding Climate Change \u2013 Frequently Asked Questions<\/a> \u2013 UCAR<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.jpl.nasa.gov\/news\/news.php?feature=6701\">Global Carbon Dioxide Circulation<\/a> (<a href=\"https:\/\/en.wikipedia.org\/wiki\/NASA\" title=\"NASA\">NASA<\/a>; 13 December 2016)<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.worldbank.org\/en\/news\/feature\/2012\/11\/18\/Climate-change-report-warns-dramatically-warmer-world-this-century\">The World Bank \u2013 Climate Change \u2013 A 4 Degree Warmer World \u2013 We must and can avoid it<\/a><\/li>\n<li>Climate change tutorial by Prof. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Myles_Allen\" title=\"Myles Allen\">Myles Allen<\/a> (Oxford), March 2018: Parts <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.youtube.com\/watch?time_continue=8&amp;v=CqjbTMz5Hro\">1<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.youtube.com\/watch?time_continue=104&amp;v=Tqx99ea-DrM\">2<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.youtube.com\/watch?time_continue=13&amp;v=2Hh8iurCvcA\">3<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.youtube.com\/watch?time_continue=3&amp;v=QFlcmsFOv6Y\">4<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.youtube.com\/watch?time_continue=2&amp;v=aMjzpaNLQ0Y\">5<\/a> (45 min. total); <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.eci.ox.ac.uk\/news\/2018\/0410.html\">background &amp; slide deck<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/climatestate.com\/2018\/07\/29\/experts-explain-recent-heat-waves-and-atmospheric-changes\/\">Experts Discuss Recent Heat Waves and Atmospheric Changes<\/a> (July 2018)<\/li><\/ul><div role=\"navigation\" class=\"navbox\" aria-labelledby=\"Authority_control_frameless_&amp;#124;text-top_&amp;#124;10px_&amp;#124;alt=Edit_this_at_Wikidata_&amp;#124;link=https&amp;#58;\/\/www.wikidata.org\/wiki\/Q7942&amp;#124;Edit_this_at_Wikidata\"><table class=\"nowraplinks hlist navbox-inner\"><tbody><tr><th id=\"Authority_control_frameless_&amp;#124;text-top_&amp;#124;10px_&amp;#124;alt=Edit_this_at_Wikidata_&amp;#124;link=https&amp;#58;\/\/www.wikidata.org\/wiki\/Q7942&amp;#124;Edit_this_at_Wikidata\" scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Help:Authority_control\" title=\"Help:Authority control\">Authority control<\/a> <a href=\"https:\/\/www.wikidata.org\/wiki\/Q7942\" title=\"Edit this at Wikidata\"><img alt=\"Edit this at Wikidata\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/73\/Blue_pencil.svg\/10px-Blue_pencil.svg.png\" width=\"10\" height=\"10\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/73\/Blue_pencil.svg\/15px-Blue_pencil.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/73\/Blue_pencil.svg\/20px-Blue_pencil.svg.png 2x\" data-file-width=\"600\" data-file-height=\"600\" \/><\/a><\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Biblioth\u00e8que_nationale_de_France\" title=\"Biblioth\u00e8que nationale de France\">BNF<\/a>: <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/catalogue.bnf.fr\/ark:\/12148\/cb12156561p\">cb12156561p<\/a> <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/data.bnf.fr\/ark:\/12148\/cb12156561p\">(data)<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Integrated_Authority_File\" title=\"Integrated Authority File\">GND<\/a>: <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/d-nb.info\/gnd\/4344515-9\">4344515-9<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Diet_Library\" title=\"National Diet Library\">NDL<\/a>: <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/id.ndl.go.jp\/auth\/ndlna\/00923848\">00923848<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n<div role=\"navigation\" class=\"navbox\" aria-labelledby=\"Global_warming_and_climate_change\"><table class=\"nowraplinks hlist collapsible expanded navbox-inner\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div class=\"plainlinks hlist navbar mini\"><ul><li class=\"nv-view\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template:Global_warming\" title=\"Template:Global warming\"><abbr title=\"View this template\">v<\/abbr><\/a><\/li><li class=\"nv-talk\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template_talk:Global_warming\" title=\"Template talk:Global warming\"><abbr title=\"Discuss this template\">t<\/abbr><\/a><\/li><li class=\"nv-edit\"><a class=\"external text\" href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Template:Global_warming&amp;action=edit\"><abbr title=\"Edit this template\">e<\/abbr><\/a><\/li><\/ul><\/div><div id=\"Global_warming_and_climate_change\"><a class=\"mw-selflink selflink\">Global warming<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change\" title=\"Climate change\">climate change<\/a><\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Temperatures\">Temperatures<\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Brightness_temperature\" title=\"Brightness temperature\">Brightness temperature<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effective_temperature\" title=\"Effective temperature\">Effective temperature<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Geologic_temperature_record\" title=\"Geologic temperature record\">Geologic record<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_hiatus\" title=\"Global warming hiatus\">Hiatus<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Historical_climatology\" title=\"Historical climatology\">Historical climatology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Instrumental_temperature_record\" title=\"Instrumental temperature record\">Instrumental record<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Paleoclimatology\" title=\"Paleoclimatology\">Paleoclimatology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Paleotempestology\" title=\"Paleotempestology\">Paleotempestology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Proxy_(climate)\" title=\"Proxy (climate)\">Proxy data<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Temperature_record_of_the_past_1000_years\" title=\"Temperature record of the past 1000 years\">Record of the past 1,000 years<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Satellite_temperature_measurements\" title=\"Satellite temperature measurements\">Satellite measurements<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Causes\">Causes<\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks navbox-subgroup\"><tbody><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Anthropogenic_climate_change\" class=\"mw-redirect\" title=\"Anthropogenic climate change\">Anthropogenic<\/a><div>(caused by human activity)<\/div><\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Attribution_of_recent_climate_change\" title=\"Attribution of recent climate change\">Attribution of recent climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_aviation\" title=\"Environmental impact of aviation\">Aviation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Biofuel\" title=\"Biofuel\">Biofuel<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Black_carbon\" title=\"Black carbon\">Black carbon<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide\" title=\"Carbon dioxide\">Carbon dioxide<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Deforestation_and_climate_change\" title=\"Deforestation and climate change\">Deforestation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Earth%27s_energy_budget\" title=\"Earth's energy budget\">Earth's energy budget<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Earth%27s_energy_budget\" title=\"Earth's energy budget\">Earth's radiation balance<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ecocide\" title=\"Ecocide\">Ecocide<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fossil_fuel\" title=\"Fossil fuel\">Fossil fuel<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_dimming\" title=\"Global dimming\">Global dimming<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_potential\" title=\"Global warming potential\">Global warming potential<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_effect\" title=\"Greenhouse effect\">Greenhouse effect<\/a><\/li>\n<li>(<a href=\"https:\/\/en.wikipedia.org\/wiki\/Infrared_window\" title=\"Infrared window\">Infrared window<\/a>)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Greenhouse_gas\" title=\"Greenhouse gas\">Greenhouse gases<\/a><\/li>\n<li>(<a href=\"https:\/\/en.wikipedia.org\/wiki\/Halocarbon\" title=\"Halocarbon\">Halocarbons<\/a>)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Land_use,_land-use_change,_and_forestry\" title=\"Land use, land-use change, and forestry\">Land use, land-use change, and forestry<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Radiative_forcing\" title=\"Radiative forcing\">Radiative forcing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Tropospheric_ozone\" title=\"Tropospheric ozone\">Tropospheric ozone<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Urban_heat_island\" title=\"Urban heat island\">Urban heat island<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Natural_environment\" title=\"Natural environment\">Natural<\/a><\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Albedo\" title=\"Albedo\">Albedo<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Bond_event\" title=\"Bond event\">Bond events<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_oscillation\" title=\"Climate oscillation\">Climate oscillations<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_sensitivity\" title=\"Climate sensitivity\">Climate sensitivity<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Cloud_forcing\" title=\"Cloud forcing\">Cloud forcing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Cosmic_ray\" title=\"Cosmic ray\">Cosmic rays<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_feedback\" title=\"Climate change feedback\">Feedbacks<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Glacial_period\" title=\"Glacial period\">Glaciation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_cooling\" title=\"Global cooling\">Global cooling<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Milankovitch_cycles\" title=\"Milankovitch cycles\">Milankovitch cycles<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Thermohaline_circulation\" title=\"Thermohaline circulation\">Ocean variability<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Atlantic_multidecadal_oscillation\" title=\"Atlantic multidecadal oscillation\">AMO<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/El_Ni\u00f1o\u2013Southern_Oscillation\" title=\"El Ni\u00f1o\u2013Southern Oscillation\">ENSO<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Indian_Ocean_Dipole\" title=\"Indian Ocean Dipole\">IOD<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Pacific_decadal_oscillation\" title=\"Pacific decadal oscillation\">PDO<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Orbital_forcing\" title=\"Orbital forcing\">Orbital forcing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Solar_cycle\" title=\"Solar cycle\">Solar variation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Volcanism\" title=\"Volcanism\">Volcanism<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_model\" title=\"Climate model\">Models<\/a><\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/General_circulation_model\" title=\"General circulation model\">Global climate model<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"History\">History<\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/History_of_climate_change_science\" title=\"History of climate change science\">History of climate change science<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Atmospheric_thermodynamics\" title=\"Atmospheric thermodynamics\">Atmospheric thermodynamics<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Svante_Arrhenius\" title=\"Svante Arrhenius\">Svante Arrhenius<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/James_Hansen\" title=\"James Hansen\">James Hansen<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Charles_David_Keeling\" title=\"Charles David Keeling\">Charles David Keeling<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Opinion_and_climate_change\">Opinion and climate change<\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks navbox-subgroup\"><tbody><tr><th scope=\"row\" class=\"navbox-group\">General<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_ethics\" title=\"Environmental ethics\">Environmental ethics<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Media_coverage_of_global_warming\" title=\"Media coverage of global warming\">Media coverage of climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Public_opinion_on_global_warming\" title=\"Public opinion on global warming\">Public opinion on climate change<\/a><\/li>\n<li>(<a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_in_popular_culture\" title=\"Global warming in popular culture\">Popular culture<\/a>)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_opinion_on_climate_change\" title=\"Scientific opinion on climate change\">Scientific opinion on climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_scientists_who_disagree_with_the_scientific_consensus_on_global_warming\" title=\"List of scientists who disagree with the scientific consensus on global warming\">Scientists who disagree with the mainstream assessment<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_denial\" title=\"Climate change denial\">Climate change denial<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_conspiracy_theory\" title=\"Global warming conspiracy theory\">Global warming conspiracy theory<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_opinion_by_country\" title=\"Climate change opinion by country\">By country &amp; region<\/a><\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Africa\" title=\"Climate change in Africa\">Africa<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_the_Arctic\" title=\"Climate change in the Arctic\">Arctic<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Argentina\" title=\"Climate change in Argentina\">Argentina<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Australia\" title=\"Climate change in Australia\">Australia<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Bangladesh\" title=\"Climate change in Bangladesh\">Bangladesh<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_in_Belgium\" title=\"Global warming in Belgium\">Belgium<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Canada\" title=\"Climate change in Canada\">Canada<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_China\" title=\"Climate change in China\">China<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Europe\" title=\"Climate change in Europe\">Europe<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_the_European_Union\" title=\"Climate change in the European Union\">European Union<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Finland\" title=\"Climate change in Finland\">Finland<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Grenada\" title=\"Climate change in Grenada\">Grenada<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Japan\" title=\"Climate change in Japan\">Japan<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_in_Luxembourg\" title=\"Global warming in Luxembourg\">Luxembourg<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_New_Zealand\" title=\"Climate change in New Zealand\">New Zealand<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_warming_in_Norway\" title=\"Global warming in Norway\">Norway<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Russia\" title=\"Climate change in Russia\">Russia<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Scotland\" title=\"Climate change in Scotland\">Scotland<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_South_Korea\" title=\"Climate change in South Korea\">South Korea<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Sweden\" title=\"Climate change in Sweden\">Sweden<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_Tuvalu\" title=\"Climate change in Tuvalu\">Tuvalu<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_the_United_Kingdom\" title=\"Climate change in the United Kingdom\">United Kingdom<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_in_the_United_States\" title=\"Climate change in the United States\">United States<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Politics\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Politics_of_global_warming\" title=\"Politics of global warming\">Politics<\/a><\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks navbox-subgroup\"><tbody><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Clean_Power_Plan\" title=\"Clean Power Plan\">Clean Power Plan<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_denial\" title=\"Climate change denial\">Climate change denial<\/a><\/li>\n<li>(<a href=\"https:\/\/en.wikipedia.org\/wiki\/Manufactured_controversy\" title=\"Manufactured controversy\">Manufactured controversy<\/a>)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Intergovernmental_Panel_on_Climate_Change\" title=\"Intergovernmental Panel on Climate Change\">Intergovernmental Panel on Climate Change (IPCC)<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/March_for_Science\" title=\"March for Science\">March for Science<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/People%27s_Climate_March\" title=\"People's Climate March\">People's Climate March<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/United_Nations_Framework_Convention_on_Climate_Change\" title=\"United Nations Framework Convention on Climate Change\">United Nations Framework Convention on Climate Change (UNFCCC \/ FCCC)<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Global_climate_regime\" title=\"Global climate regime\">Global climate regime<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Potential_effects_and_issues\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming\" title=\"Effects of global warming\">Potential effects and issues<\/a><\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks navbox-subgroup\"><tbody><tr><th scope=\"row\" class=\"navbox-group\">General<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Abrupt_climate_change\" title=\"Abrupt climate change\">Abrupt climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Anoxic_event\" title=\"Anoxic event\">Anoxic event<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_dipole_anomaly\" title=\"Arctic dipole anomaly\">Arctic dipole anomaly<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_haze\" title=\"Arctic haze\">Arctic haze<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Arctic_methane_emissions\" title=\"Arctic methane emissions\">Arctic methane emissions<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_agriculture\" title=\"Climate change and agriculture\">Climate change and agriculture<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_ecosystems\" title=\"Climate change and ecosystems\">Climate change and ecosystems<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_gender\" title=\"Climate change and gender\">Climate change and gender<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_and_poverty\" title=\"Climate change and poverty\">Climate change and poverty<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Drought\" title=\"Drought\">Drought<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Economics_of_global_warming\" title=\"Economics of global warming\">Economics of global warming<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_climate_change_on_plant_biodiversity\" title=\"Effects of climate change on plant biodiversity\">Effects on plant biodiversity<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_human_health\" title=\"Effects of global warming on human health\">Effects on health<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_humans\" title=\"Effects of global warming on humans\">Effects on humans<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_marine_mammals\" title=\"Effects of global warming on marine mammals\">Effects on marine mammals<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_migrant\" title=\"Environmental migrant\">Environmental migrant<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Extinction_risk_from_global_warming\" title=\"Extinction risk from global warming\">Extinction risk from global warming<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fisheries_and_climate_change\" title=\"Fisheries and climate change\">Fisheries and climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Forest_dieback\" title=\"Forest dieback\">Forest dieback<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change,_industry_and_society\" title=\"Climate change, industry and society\">Industry and society<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Iris_hypothesis\" title=\"Iris hypothesis\">Iris hypothesis<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Megadrought\" title=\"Megadrought\">Megadrought<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ocean_acidification\" title=\"Ocean acidification\">Ocean acidification<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ozone_depletion\" title=\"Ozone depletion\">Ozone depletion<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Physical_impacts_of_climate_change\" title=\"Physical impacts of climate change\">Physical impacts<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Polar_stratospheric_cloud\" title=\"Polar stratospheric cloud\">Polar stratospheric cloud<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Regime_shift\" title=\"Regime shift\">Regime shift<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Retreat_of_glaciers_since_1850\" title=\"Retreat of glaciers since 1850\">Retreat of glaciers since 1850<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Runaway_climate_change\" title=\"Runaway climate change\">Runaway climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sea_level_rise\" title=\"Sea level rise\">Sea level rise<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Season_creep\" title=\"Season creep\">Season creep<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Shutdown_of_thermohaline_circulation\" title=\"Shutdown of thermohaline circulation\">Shutdown of thermohaline circulation<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">By country<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_Australia\" title=\"Effects of global warming on Australia\">Australia<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_South_Asia\" title=\"Effects of global warming on South Asia\">South Asia<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Effects_of_global_warming_on_India\" class=\"mw-redirect\" title=\"Effects of global warming on India\">India<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_adaptation_in_Nepal\" title=\"Climate change adaptation in Nepal\">Nepal<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Climate_Assessment\" title=\"National Climate Assessment\">United States<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Mitigation\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_mitigation\" title=\"Climate change mitigation\">Mitigation<\/a><\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks navbox-subgroup\"><tbody><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Kyoto_Protocol\" title=\"Kyoto Protocol\">Kyoto Protocol<\/a><\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Clean_Development_Mechanism\" title=\"Clean Development Mechanism\">Clean Development Mechanism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Joint_Implementation\" title=\"Joint Implementation\">Joint Implementation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Bali_Road_Map\" title=\"Bali Road Map\">Bali Road Map<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/2009_United_Nations_Climate_Change_Conference\" title=\"2009 United Nations Climate Change Conference\">2009 United Nations Climate Change Conference<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Governmental<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/European_Climate_Change_Programme\" title=\"European Climate Change Programme\">European Climate Change Programme<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/G8_Climate_Change_Roundtable\" title=\"G8 Climate Change Roundtable\">G8 Climate Change Roundtable<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/United_Kingdom_Climate_Change_Programme\" title=\"United Kingdom Climate Change Programme\">United Kingdom Climate Change Programme<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Paris_Agreement\" title=\"Paris Agreement\">Paris Agreement<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/United_States_withdrawal_from_the_Paris_Agreement\" title=\"United States withdrawal from the Paris Agreement\">United States withdrawal<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Regional_climate_change_initiatives_in_the_United_States\" title=\"Regional climate change initiatives in the United States\">Regional climate change initiatives in the United States<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_climate_change_initiatives\" title=\"List of climate change initiatives\">List of climate change initiatives<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Emissions reduction<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_credit\" title=\"Carbon credit\">Carbon credit<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon-neutral_fuel\" title=\"Carbon-neutral fuel\">Carbon-neutral fuel<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_offset\" title=\"Carbon offset\">Carbon offset<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_tax\" title=\"Carbon tax\">Carbon tax<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Emissions_trading\" title=\"Emissions trading\">Emissions trading<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fossil_fuel_phase-out\" title=\"Fossil fuel phase-out\">Fossil fuel phase-out<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_neutrality\" title=\"Carbon neutrality\">Carbon-free<\/a> energy<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_capture_and_storage\" title=\"Carbon capture and storage\">Carbon capture and storage<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Efficient_energy_use\" title=\"Efficient energy use\">Efficient energy use<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Low-carbon_economy\" title=\"Low-carbon economy\">Low-carbon economy<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Nuclear_power\" title=\"Nuclear power\">Nuclear power<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Renewable_energy\" title=\"Renewable energy\">Renewable energy<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Personal<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Individual_action_on_climate_change\" title=\"Individual action on climate change\">Individual action on climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Simple_living\" title=\"Simple living\">Simple living<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Other<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_dioxide_removal\" title=\"Carbon dioxide removal\">Carbon dioxide removal<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Carbon_sink\" title=\"Carbon sink\">Carbon sink<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_mitigation\" title=\"Climate change mitigation\">Climate action<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_climate_change_initiatives\" title=\"List of climate change initiatives\">Climate Action Plan<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_mitigation_scenarios\" title=\"Climate change mitigation scenarios\">Climate change mitigation scenarios<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_engineering\" title=\"Climate engineering\">Climate engineering<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Individual_and_political_action_on_climate_change\" title=\"Individual and political action on climate change\">Individual and political action on climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Reducing_emissions_from_deforestation_and_forest_degradation\" title=\"Reducing emissions from deforestation and forest degradation\">Reducing emissions from deforestation and forest degradation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Reforestation\" title=\"Reforestation\">Reforestation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Urban_reforestation\" title=\"Urban reforestation\">Urban reforestation<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks collapsible collapsed navbox-subgroup\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div id=\"Proposed_adaptations\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_adaptation\" title=\"Climate change adaptation\">Proposed adaptations<\/a><\/div><\/th><\/tr><tr><td colspan=\"2\" class=\"navbox-list navbox-odd\"><div><table class=\"nowraplinks navbox-subgroup\"><tbody><tr><th scope=\"row\" class=\"navbox-group\">Strategies<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Dam\" title=\"Dam\">Damming<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Glacial_lake\" title=\"Glacial lake\">glacial lakes<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Desalination\" title=\"Desalination\">Desalination<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Drought_tolerance\" title=\"Drought tolerance\">Drought tolerance<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Irrigation\" title=\"Irrigation\">Irrigation investment<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Rainwater_tank\" title=\"Rainwater tank\">Rainwater storage<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sustainable_development\" title=\"Sustainable development\">Sustainable development<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Weather_modification\" title=\"Weather modification\">Weather modification<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Programmes<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Avoiding_Dangerous_Climate_Change_(2005_conference)\" title=\"Avoiding Dangerous Climate Change (2005 conference)\">Avoiding dangerous climate change<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Land_allocation_decision_support_system\" title=\"Land allocation decision support system\">Land allocation decision support system<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><\/tbody><\/table><\/div><\/td><\/tr><tr><td class=\"navbox-abovebelow\" colspan=\"2\"><div>\n<ul><li><img alt=\"List-Class article\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/d\/db\/Symbol_list_class.svg\/16px-Symbol_list_class.svg.png\" title=\"List-Class article\" width=\"16\" height=\"16\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/d\/db\/Symbol_list_class.svg\/23px-Symbol_list_class.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/d\/db\/Symbol_list_class.svg\/31px-Symbol_list_class.svg.png 2x\" data-file-width=\"180\" data-file-height=\"185\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Glossary_of_climate_change\" title=\"Glossary of climate change\">Glossary<\/a><\/li>\n<li><img alt=\"List-Class article\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/d\/db\/Symbol_list_class.svg\/16px-Symbol_list_class.svg.png\" title=\"List-Class article\" width=\"16\" height=\"16\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/d\/db\/Symbol_list_class.svg\/23px-Symbol_list_class.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/d\/db\/Symbol_list_class.svg\/31px-Symbol_list_class.svg.png 2x\" data-file-width=\"180\" data-file-height=\"185\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Index_of_climate_change_articles\" title=\"Index of climate change articles\">Index<\/a><\/li>\n<li><img alt=\"Category\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/16px-Folder_Hexagonal_Icon.svg.png\" title=\"Category\" width=\"16\" height=\"14\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/24px-Folder_Hexagonal_Icon.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/32px-Folder_Hexagonal_Icon.svg.png 2x\" data-file-width=\"36\" data-file-height=\"31\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:Climate_change\" title=\"Category:Climate change\">Climate change<\/a><\/li>\n<li><img alt=\"Category\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/16px-Folder_Hexagonal_Icon.svg.png\" title=\"Category\" width=\"16\" height=\"14\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/24px-Folder_Hexagonal_Icon.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/32px-Folder_Hexagonal_Icon.svg.png 2x\" data-file-width=\"36\" data-file-height=\"31\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:Global_warming\" title=\"Category:Global warming\">Global warming<\/a><\/li>\n<li><img alt=\"Portal\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/f\/fd\/Portal-puzzle.svg\/16px-Portal-puzzle.svg.png\" title=\"Portal\" width=\"16\" height=\"14\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/f\/fd\/Portal-puzzle.svg\/24px-Portal-puzzle.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/f\/fd\/Portal-puzzle.svg\/32px-Portal-puzzle.svg.png 2x\" data-file-width=\"32\" data-file-height=\"28\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Portal:Global_warming\" title=\"Portal:Global warming\">Portal<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n<div role=\"navigation\" class=\"navbox\" aria-labelledby=\"Simple_living\"><table class=\"nowraplinks collapsible autocollapse navbox-inner\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div class=\"plainlinks hlist navbar mini\"><ul><li class=\"nv-view\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template:Simple_living\" title=\"Template:Simple living\"><abbr title=\"View this template\">v<\/abbr><\/a><\/li><li class=\"nv-talk\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template_talk:Simple_living\" title=\"Template talk:Simple living\"><abbr title=\"Discuss this template\">t<\/abbr><\/a><\/li><li class=\"nv-edit\"><a class=\"external text\" href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Template:Simple_living&amp;action=edit\"><abbr title=\"Edit this template\">e<\/abbr><\/a><\/li><\/ul><\/div><div id=\"Simple_living\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Simple_living\" title=\"Simple living\">Simple living<\/a><\/div><\/th><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Practices<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Barter\" title=\"Barter\">Barter<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Cord-cutting\" title=\"Cord-cutting\">Cord-cutting<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/DIY_ethic\" title=\"DIY ethic\">DIY ethic<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Downshifting_(lifestyle)\" title=\"Downshifting (lifestyle)\">Downshifting<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Dry_toilet\" title=\"Dry toilet\">Dry toilet<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Forest_gardening\" title=\"Forest gardening\">Forest gardening<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Freeganism\" title=\"Freeganism\">Freeganism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Frugality\" title=\"Frugality\">Frugality<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gift_economy\" title=\"Gift economy\">Gift economy<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Intentional_community\" title=\"Intentional community\">Intentional community<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Local_currency\" title=\"Local currency\">Local currency<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Low-impact_development_(UK)\" title=\"Low-impact development (UK)\">Low-impact development<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Minimalism\" title=\"Minimalism\">Minimalism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/No_frills\" title=\"No frills\">No frills<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Off-the-grid\" title=\"Off-the-grid\">Off-the-grid<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Permaculture\" title=\"Permaculture\">Permaculture<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Self-sufficiency\" class=\"mw-redirect\" title=\"Self-sufficiency\">Self-sufficiency<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Subsistence_agriculture\" title=\"Subsistence agriculture\">Subsistence agriculture<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sustainable_living\" title=\"Sustainable living\">Sustainable living<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sustainable_sanitation\" title=\"Sustainable sanitation\">Sustainable sanitation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Veganism\" title=\"Veganism\">Veganism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Vegetarianism\" title=\"Vegetarianism\">Vegetarianism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Tax_resistance\" title=\"Tax resistance\">War tax resistance<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/WWOOF\" title=\"WWOOF\">WWOOF<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Religious and spiritual<\/th><td class=\"navbox-list navbox-even hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Asceticism\" title=\"Asceticism\">Asceticism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Aparigraha\" title=\"Aparigraha\">Aparigraha<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Cynicism_(philosophy)\" title=\"Cynicism (philosophy)\">Cynicism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Detachment_(philosophy)\" title=\"Detachment (philosophy)\">Detachment<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Distributism\" title=\"Distributism\">Distributism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Jesus_movement\" title=\"Jesus movement\">Jesus movement<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mendicant\" title=\"Mendicant\">Mendicant<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sati_(Buddhism)\" title=\"Sati (Buddhism)\">Mindfulness<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Monasticism\" title=\"Monasticism\">Monasticism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/New_Monasticism\" title=\"New Monasticism\">New Monasticism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Plain_dress\" title=\"Plain dress\">Plain dress<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Plain_people\" title=\"Plain people\">Plain people<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Quakers\" title=\"Quakers\">Quakers<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Rastafari\" title=\"Rastafari\">Rastafari<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Temperance_(virtue)\" title=\"Temperance (virtue)\">Temperance<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Testimony_of_Simplicity\" class=\"mw-redirect\" title=\"Testimony of Simplicity\">Testimony of simplicity<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Tolstoyan_movement\" title=\"Tolstoyan movement\">Tolstoyan movement<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Secular movements<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Back-to-the-land_movement\" title=\"Back-to-the-land movement\">Back-to-the-land<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Car-free_movement\" title=\"Car-free movement\">Car-free<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Movement_for_Compassionate_Living\" title=\"Movement for Compassionate Living\">Compassionate living<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_movement\" title=\"Environmental movement\">Environmental<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Hippie\" title=\"Hippie\">Hippie<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Slow_movement_(culture)\" title=\"Slow movement (culture)\">Slow<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Small_house_movement\" class=\"mw-redirect\" title=\"Small house movement\">Small house<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Transition_town\" title=\"Transition town\">Transition town<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Open_Source_Ecology\" title=\"Open Source Ecology\">Open Source Ecology<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Notable writers<\/th><td class=\"navbox-list navbox-even hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Wendell_Berry\" title=\"Wendell Berry\">Wendell Berry<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ernest_Callenbach\" title=\"Ernest Callenbach\">Ernest Callenbach<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/G._K._Chesterton\" title=\"G. K. Chesterton\">G. K. Chesterton<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Duane_Elgin\" title=\"Duane Elgin\">Duane Elgin<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mahatma_Gandhi\" title=\"Mahatma Gandhi\">Mahatma Gandhi<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Richard_Gregg_(social_philosopher)\" title=\"Richard Gregg (social philosopher)\">Richard Gregg<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Tom_Hodgkinson\" title=\"Tom Hodgkinson\">Tom Hodgkinson<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Harlan_Hubbard\" title=\"Harlan Hubbard\">Harlan Hubbard<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Satish_Kumar\" title=\"Satish Kumar\">Satish Kumar<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Helen_Nearing\" title=\"Helen Nearing\">Helen Nearing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Scott_Nearing\" title=\"Scott Nearing\">Scott Nearing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Peace_Pilgrim\" title=\"Peace Pilgrim\">Peace Pilgrim<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Nick_Rosen\" title=\"Nick Rosen\">Nick Rosen<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Dugald_Semple\" title=\"Dugald Semple\">Dugald Semple<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/E._F._Schumacher\" title=\"E. F. Schumacher\">E. F. Schumacher<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Henry_David_Thoreau\" title=\"Henry David Thoreau\">Henry David Thoreau<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Leo_Tolstoy\" title=\"Leo Tolstoy\">Leo Tolstoy<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Modern-day adherents<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mark_Boyle_(Moneyless_Man)\" title=\"Mark Boyle (Moneyless Man)\">Mark Boyle<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Jim_Merkel\" title=\"Jim Merkel\">Jim Merkel<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Peace_Pilgrim\" title=\"Peace Pilgrim\">Peace Pilgrim<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Suelo\" title=\"Suelo\">Suelo<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Thomas_(activist)\" title=\"Thomas (activist)\">Thomas<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Media<\/th><td class=\"navbox-list navbox-even hlist\"><div>\n<ul><li>\"<a href=\"https:\/\/en.wikipedia.org\/wiki\/Anekdote_zur_Senkung_der_Arbeitsmoral\" title=\"Anekdote zur Senkung der Arbeitsmoral\">Anekdote zur Senkung der Arbeitsmoral<\/a>\"<\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Escape_from_Affluenza\" title=\"Escape from Affluenza\">Escape from Affluenza<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Good_Life_(1975_TV_series)\" title=\"The Good Life (1975 TV series)\">The Good Life<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Moon_and_the_Sledgehammer\" title=\"The Moon and the Sledgehammer\">The Moon and the Sledgehammer<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mother_Earth_News\" title=\"Mother Earth News\">Mother Earth News<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Power_of_Half\" title=\"The Power of Half\">The Power of Half<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Small_Is_Beautiful\" title=\"Small Is Beautiful\">Small Is Beautiful<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Walden\" title=\"Walden\">Walden<\/a><\/i><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Related topics<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Affluenza\" title=\"Affluenza\">Affluenza<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Agrarianism\" title=\"Agrarianism\">Agrarianism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Anarcho-primitivism\" title=\"Anarcho-primitivism\">Anarcho-primitivism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Anti-consumerism\" title=\"Anti-consumerism\">Anti-consumerism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Appropriate_technology\" title=\"Appropriate technology\">Appropriate technology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Bohemianism\" title=\"Bohemianism\">Bohemianism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Consumerism\" title=\"Consumerism\">Consumerism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Deep_ecology\" title=\"Deep ecology\">Deep ecology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Degrowth\" title=\"Degrowth\">Degrowth<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ecological_footprint\" title=\"Ecological footprint\">Ecological footprint<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Food_miles\" title=\"Food miles\">Food miles<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Front_Porch_Republic\" title=\"Front Porch Republic\">Front Porch Republic<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Green_anarchism\" title=\"Green anarchism\">Green anarchism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Eudaimonia\" title=\"Eudaimonia\">The good life<\/a><\/li>\n<li><a class=\"mw-selflink selflink\">Global warming<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Hedonophobia\" title=\"Hedonophobia\">Hedonophobia<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Intentional_living\" title=\"Intentional living\">Intentional living<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Itinerant_groups_in_Europe\" title=\"Itinerant groups in Europe\">Itinerant<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Low-technology\" class=\"mw-redirect\" title=\"Low-technology\">Low-technology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Nonviolence\" title=\"Nonviolence\">Nonviolence<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Peak_oil\" title=\"Peak oil\">Peak oil<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sustainability\" title=\"Sustainability\">Sustainability<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Work\u2013life_balance\" title=\"Work\u2013life balance\">Work\u2013life balance<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n<div role=\"navigation\" class=\"navbox\" aria-labelledby=\"Human_impact_on_the_environment\"><table class=\"nowraplinks hlist collapsible autocollapse navbox-inner\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"3\"><div class=\"plainlinks hlist navbar mini\"><ul><li class=\"nv-view\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template:Human_impact_on_the_environment\" title=\"Template:Human impact on the environment\"><abbr title=\"View this template\">v<\/abbr><\/a><\/li><li class=\"nv-talk\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template_talk:Human_impact_on_the_environment\" title=\"Template talk:Human impact on the environment\"><abbr title=\"Discuss this template\">t<\/abbr><\/a><\/li><li class=\"nv-edit\"><a class=\"external text\" href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Template:Human_impact_on_the_environment&amp;action=edit\"><abbr title=\"Edit this template\">e<\/abbr><\/a><\/li><\/ul><\/div><div id=\"Human_impact_on_the_environment\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Human_impact_on_the_environment\" title=\"Human impact on the environment\">Human impact on the environment<\/a><\/div><\/th><\/tr><tr><th scope=\"row\" class=\"navbox-group\">General<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_issue\" title=\"Environmental issue\">Environmental issues<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_environmental_issues\" title=\"List of environmental issues\">List of issues<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Human_impact_on_the_environment\" title=\"Human impact on the environment\">Human impact<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_assessment\" title=\"Environmental impact assessment\">Impact assessment<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Planetary_boundaries\" title=\"Planetary boundaries\">Planetary boundaries<\/a><\/li><\/ul><\/div><\/td><td class=\"navbox-image\" rowspan=\"4\"><div><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Planetary_boundaries.svg\" class=\"image\" title=\"Planetary boundaries\"><img alt=\"Planetary boundaries\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/70\/Planetary_boundaries.svg\/140px-Planetary_boundaries.svg.png\" width=\"140\" height=\"115\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/70\/Planetary_boundaries.svg\/210px-Planetary_boundaries.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/70\/Planetary_boundaries.svg\/280px-Planetary_boundaries.svg.png 2x\" data-file-width=\"1417\" data-file-height=\"1163\" \/><\/a><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Causes<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_agriculture\" title=\"Environmental impact of agriculture\">Agriculture<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_fishing\" title=\"Environmental impact of fishing\">fishing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_irrigation\" title=\"Environmental impact of irrigation\">irrigation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_meat_production\" title=\"Environmental impact of meat production\">meat production<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_cocoa_production\" class=\"mw-redirect\" title=\"Environmental impact of cocoa production\">cocoa production<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Social_and_environmental_impact_of_palm_oil\" title=\"Social and environmental impact of palm oil\">palm oil<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_the_energy_industry\" title=\"Environmental impact of the energy industry\">Energy industry<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Indirect_land_use_change_impacts_of_biofuels\" title=\"Indirect land use change impacts of biofuels\">biofuels<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_biodiesel\" title=\"Environmental impact of biodiesel\">biodiesel<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_the_coal_industry\" title=\"Environmental impact of the coal industry\">coal<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_electricity_generation\" title=\"Environmental impact of electricity generation\">electricity generation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_nuclear_power\" title=\"Environmental impact of nuclear power\">nuclear power<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_the_oil_shale_industry\" title=\"Environmental impact of the oil shale industry\">oil shale<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_the_petroleum_industry\" title=\"Environmental impact of the petroleum industry\">petroleum<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_reservoirs\" title=\"Environmental impact of reservoirs\">reservoirs<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_wind_power\" title=\"Environmental impact of wind power\">wind power<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Genetic_pollution\" title=\"Genetic pollution\">Genetic pollution<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Industrialization\" class=\"mw-redirect\" title=\"Industrialization\">Industrialization<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Land_use#Environment\" title=\"Land use\">Land use<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Manufacturing\" title=\"Manufacturing\">Manufacturing<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_cleaning_agents\" title=\"Environmental impact of cleaning agents\">cleaning agents<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_concrete\" title=\"Environmental impact of concrete\">concrete<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_plastics\" class=\"mw-redirect\" title=\"Environmental impact of plastics\">plastics<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_nanotechnology\" class=\"mw-redirect\" title=\"Environmental impact of nanotechnology\">nanotechnology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_paint\" title=\"Environmental impact of paint\">paint<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_paper\" title=\"Environmental impact of paper\">paper<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_pesticides\" title=\"Environmental impact of pesticides\">pesticides<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_pharmaceuticals_and_personal_care_products\" title=\"Environmental impact of pharmaceuticals and personal care products\">pharmaceuticals and personal care<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_mining\" title=\"Environmental impact of mining\">Mining<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Overdrafting\" title=\"Overdrafting\">Overdrafting<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Overfishing\" title=\"Overfishing\">Overfishing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Overgrazing\" title=\"Overgrazing\">Overgrazing<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Overexploitation\" title=\"Overexploitation\">Overexploitation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Human_overpopulation\" title=\"Human overpopulation\">Overpopulation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Particulates\" title=\"Particulates\">Particulates<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Pollution\" title=\"Pollution\">Pollution<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Quarry\" title=\"Quarry\">Quarrying<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_reservoirs\" title=\"Environmental impact of reservoirs\">Reservoirs<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_tourism\" class=\"mw-redirect\" title=\"Environmental impact of tourism\">Tourism<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_transport\" title=\"Environmental impact of transport\">Transport<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_aviation\" title=\"Environmental impact of aviation\">aviation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_roads\" class=\"mw-redirect\" title=\"Environmental impact of roads\">roads<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_shipping\" title=\"Environmental impact of shipping\">shipping<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Urbanization\" title=\"Urbanization\">Urbanization<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Urban_sprawl\" title=\"Urban sprawl\">urban sprawl<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_impact_of_war\" title=\"Environmental impact of war\">war<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Effects<\/th><td class=\"navbox-list navbox-odd\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Biodiversity_threats\" class=\"mw-redirect\" title=\"Biodiversity threats\">Biodiversity threats<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Biodiversity_loss\" title=\"Biodiversity loss\">Biodiversity loss<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change\" title=\"Climate change\">Climate change<\/a>\n<ul><li><a class=\"mw-selflink selflink\">Global warming<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Runaway_climate_change\" title=\"Runaway climate change\">Runaway climate change<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_issues_with_coral_reefs\" title=\"Environmental issues with coral reefs\">Coral reefs<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Deforestation\" title=\"Deforestation\">Deforestation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Defaunation\" title=\"Defaunation\">Defaunation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Desertification\" title=\"Desertification\">Desertification<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ecocide\" title=\"Ecocide\">Ecocide<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Erosion\" title=\"Erosion\">Erosion<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_degradation\" title=\"Environmental degradation\">Environmental degradation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Freshwater_cycle\" class=\"mw-redirect\" title=\"Freshwater cycle\">Freshwater cycle<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Habitat_destruction\" title=\"Habitat destruction\">Habitat destruction<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Holocene_extinction\" title=\"Holocene extinction\">Holocene extinction<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Human_impact_on_the_nitrogen_cycle\" title=\"Human impact on the nitrogen cycle\">Nitrogen cycle<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Land_degradation\" title=\"Land degradation\">Land degradation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Land_surface_effects_on_climate\" title=\"Land surface effects on climate\">Land surface effects on climate<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Green_belt\" title=\"Green belt\">Loss of green belts<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Phosphorus_cycle#Human_influences\" title=\"Phosphorus cycle\">Phosphorus cycle<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ocean_acidification#Possible_impacts\" title=\"Ocean acidification\">Ocean acidification<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ozone_depletion\" title=\"Ozone depletion\">Ozone depletion<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Resource_depletion\" title=\"Resource depletion\">Resource depletion<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Water_degradation\" class=\"mw-redirect\" title=\"Water degradation\">Water degradation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Water_scarcity\" title=\"Water scarcity\">Water scarcity<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Mitigation<\/th><td class=\"navbox-list navbox-even\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Birth_control\" title=\"Birth control\">Birth control<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Cleaner_production\" title=\"Cleaner production\">Cleaner production<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_change_mitigation\" title=\"Climate change mitigation\">Climate change mitigation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Climate_engineering\" title=\"Climate engineering\">Climate engineering<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Ecological_engineering\" title=\"Ecological engineering\">Ecological engineering<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_engineering\" title=\"Environmental engineering\">Environmental engineering<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Environmental_mitigation\" title=\"Environmental mitigation\">Environmental mitigation<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Industrial_ecology\" title=\"Industrial ecology\">Industrial ecology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mitigation_banking\" title=\"Mitigation banking\">Mitigation banking<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Organic_farming\" title=\"Organic farming\">Organic farming<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Reforestation\" title=\"Reforestation\">Reforestation<\/a>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Urban_reforestation\" title=\"Urban reforestation\">urban<\/a><\/li><\/ul><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Restoration_ecology\" title=\"Restoration ecology\">Restoration ecology<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sustainable_consumption\" title=\"Sustainable consumption\">Sustainable consumption<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Waste_minimisation\" title=\"Waste minimisation\">Waste minimization<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><td class=\"navbox-abovebelow\" colspan=\"3\"><div>\n<ul><li>\u00a0<img alt=\"Commons page\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/4a\/Commons-logo.svg\/12px-Commons-logo.svg.png\" title=\"Commons page\" width=\"12\" height=\"16\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/4a\/Commons-logo.svg\/18px-Commons-logo.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/4a\/Commons-logo.svg\/24px-Commons-logo.svg.png 2x\" data-file-width=\"1024\" data-file-height=\"1376\" \/><b><a href=\"https:\/\/commons.wikimedia.org\/wiki\/Category:Environmental_impact\" class=\"extiw\" title=\"commons:Category:Environmental impact\">Commons<\/a><\/b><\/li>\n<li>\u00a0<img alt=\"Category\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/16px-Folder_Hexagonal_Icon.svg.png\" title=\"Category\" width=\"16\" height=\"14\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/24px-Folder_Hexagonal_Icon.svg.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/en\/thumb\/4\/48\/Folder_Hexagonal_Icon.svg\/32px-Folder_Hexagonal_Icon.svg.png 2x\" data-file-width=\"36\" data-file-height=\"31\" \/><a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:Environmental_impact_by_source\" title=\"Category:Environmental impact by source\">Category<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:Environmental_impact_by_country\" title=\"Category:Environmental impact by country\">by country<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:Environmental_impact_assessment\" title=\"Category:Environmental impact assessment\">assessment<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Category:Environmental_mitigation\" title=\"Category:Environmental mitigation\">mitigation<\/a><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n<p class=\"mw-empty-elt\">\n<\/p>\n\n\n\n\n<\/div><noscript><img src=\"https:\/\/en.wikipedia.org\/wiki\/Special:CentralAutoLogin\/start?type=1x1\" alt=\"\" title=\"\" width=\"1\" height=\"1\" \/><\/noscript><\/div>\t\t\t\t\t<div class=\"printfooter\">\n\t\t\t\t\t\tRetrieved from \"<a dir=\"ltr\" href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Global_warming&amp;oldid=867575737\">https:\/\/en.wikipedia.org\/w\/index.php?title=Global_warming&amp;oldid=867575737<\/a>\"\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<div class=\"visualClear\">\n\t\t\t\t\t\t\t<\/div>","created_at":"2018-11-07T16:54:47+0000","updated_at":"2018-11-07T16:54:47+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":104,"domain_name":"en.wikipedia.org","preview_picture":"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/f\/f8\/Global_Temperature_Anomaly.svg\/1200px-Global_Temperature_Anomaly.svg.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":664,"title":"FZF &amp; RipGrep - Navigate with bash faster than ever before","url":"http:\/\/owen.cymru\/fzf-ripgrep-navigate-with-bash-faster-than-ever-before\/","content":"<p>I've always had <a href=\"https:\/\/github.com\/junegunn\/fzf\">fzf<\/a> and <a href=\"https:\/\/github.com\/BurntSushi\/ripgrep\">ripgrep<\/a> on my radar, and I've finally gotten around to using them together. Good lord it makes a world of difference, especially when added to Vim as well as Bash. <\/p><p>Add the following snippet to your ~\/.bashrc, this add's fzf keybindings to bash and gets fzf to use ripgrep by default for faster searching.<\/p><pre>[ -f ~\/.fzf.bash ] &amp;&amp; source ~\/.fzf.bash\nexport FZF_DEFAULT_COMMAND='rg --files --no-ignore --hidden --follow -g \"!{.git,node_modules}\/*\" 2&gt; \/dev\/null'\nexport FZF_CTRL_T_COMMAND=\"$FZF_DEFAULT_COMMAND\"\nbind -x '\"\\C-p\": vim $(fzf);'\n<\/pre><p>Okay now what can you do? <\/p><ul><li><strong>Ctrl + r<\/strong> - search through bash history with fzf<\/li>\n<li><strong>Ctrl + p<\/strong> - edit a file in vim from fzf<\/li>\n<li><strong>mv dir\/**<\/strong> - expand a directory with (**) and select from fzf<\/li>\n<li><strong>Alt + c<\/strong> - change directory from fzf - see the update at the bottom for faster search with <code>bfs<\/code>.<\/li>\n<li><strong>Ctrl + t<\/strong> - insert file from fzf into command<\/li>\n<\/ul><p>Neat right! Now if you are a vim user there is more, add the fzf plugin to your ~\/.vimrc, along with this snippet. Obviously  customise the bindings, and excludes \/ includes to your workflows!<\/p><pre>let g:rg_command = '\n  \\ rg --column --line-number --no-heading --fixed-strings --ignore-case --no-ignore --hidden --follow --color \"always\"\n  \\ -g \"*.{js,json,php,md,styl,jade,html,config,py,cpp,c,go,hs,rb,conf}\"\n  \\ -g \"!{.git,node_modules,vendor}\/*\" '\ncommand! -bang -nargs=* F call fzf#vim#grep(g:rg_command .shellescape(&lt;q-args&gt;), 1, &lt;bang&gt;0)\n<\/pre><p>You now have a killer free text search with <code>:F<\/code> that uses ripgrep and is faster than any I've seen before.<\/p><p>I've done more with this, but want to leave it there for now. There are even more goodies in my <a href=\"https:\/\/github.com\/bag-man\/dotfiles\/blob\/master\/vimrc\">.vimrc<\/a>, and in my <a href=\"https:\/\/github.com\/bag-man\/dotfiles\/blob\/master\/bashrc\">.bashrc<\/a>, including auto installing fzf and ripgrep (admittedly hackily) from the .vimrc and a nice snippet that uses fzf for git logs. <\/p><p>Just made this nice snippet for tmux:<\/p><pre>tm() {\n  [[ -n \"$TMUX\" ]] &amp;&amp; change=\"switch-client\" || change=\"attach-session\"\n  if [ $1 ]; then \n     tmux $change -t \"$1\" 2&gt;\/dev\/null || (tmux new-session -d -s $1 &amp;&amp; tmux $change -t \"$1\"); return\n  fi\n  session=$(tmux list-sessions -F \"#{session_name}\" 2&gt;\/dev\/null | fzf --exit-0) &amp;&amp;  tmux $change -t \"$session\" || echo \"No sessions found.\"\n}\n<\/pre><ul><li><code>tm irc<\/code> it will attach to the irc session (if it exists), else it will create it.<\/li>\n<li><code>tm<\/code> will open fzf and let you select the session<\/li>\n<\/ul><p>What is awesome is this lets you switch and create sessions from inside of other sessions.<\/p><p>And here is a nice snippet for looking through git logs:<\/p><pre>fzf_log() {  \n  hash=$(git log --color=always --format=\"%C(auto)%h%d %s %C(black)%C(bold)%cr\" \"$@\" |  fzf | awk '{print $1}')\n  echo $hash | xclip\n  git showtool $hash\n}\n<\/pre><p>It will let you select a commit, and display the diff off it, and put the commit hash on your clipboard using xclip. I also have it set to use icdiff via <code>git showtool<\/code>.<\/p><p>A few people have pointed out that doing this doesn't work for fzf's Ctrl+T completion or the Alt+C completion. Thanks to <a href=\"https:\/\/twitter.com\/mikepqr\">@mikepqr<\/a>, for letting me know the work around for Ctrl+T. I have updated the configurations at the top of the post with what he sent me. <\/p><p>Alt+C has proved a bit tricky to get working with ripgrep, as rg doesn't natively support searching just for directories. Instead I have been using <a href=\"https:\/\/github.com\/tavianator\/bfs\">bfs<\/a> by adding the following to me config. <\/p><pre>export FZF_ALT_C_COMMAND=\"cd ~\/; bfs -type d -nohidden | sed s\/^\\.\/~\/\"  \n<\/pre><p>After a bit of quick testing it appears to run faster and give slightly nicer results that the <a href=\"https:\/\/github.com\/BurntSushi\/ripgrep\/issues\/388\">ripgrep solutions<\/a> I've tried, although it does mean installing another tool!<\/p><p>:wq<\/p><p>BTC | 15FYeMgQsETViUcc8JtSmoHmCAmADy2Ymn <br \/>LTC | LVdVyd3R1S719ZRApaAY1RgLJ5vYgEGGxW <br \/>ARK | ANKmYTHV7QdkcLz2R65uMzEvcRRdz2iRQm <\/p>","created_at":"2018-11-07T16:54:46+0000","updated_at":"2018-11-07T16:54:46+0000","published_at":"2017-02-12T15:23:54+0000","published_by":["Owen Garland"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"owen.cymru","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":665,"title":"Getting started with emacs for clojure development","url":"http:\/\/www.aidanf.net\/posts\/getting-started-with-emacs-for-clojure-development","content":"<p>Emacs is a great tool for clojure development. In fact you often hear people raving about the marriage of emacs and clojure without any clear explanation of if or why it's better than any other editor that you could use for clojure development.<\/p><p>I'm not going to get into why emacs is good for clojure development here. The aim of this post is to help someone who is looking to try it get up and running quickly, so they can try it for themselves.<\/p><p>I'm assuming you already have emacs installed and <a href=\"http:\/\/www.aidanf.net\/posts\/getting-setup-with-clojure\">clojure setup<\/a>. I won't go into the basics of using emacs here, but I'll outline a few of the features that you may find handy for working with clojure code.<\/p><h2>Emacs Live<\/h2><p>The quickest way to get your emacs up and running is to install <a href=\"https:\/\/github.com\/overtone\/emacs-live\">emacs-live<\/a>. Emacs live is<\/p><blockquote><div>\"An opinionated set of defaults for getting started with a specific focus on live coding with\u00a0<a href=\"http:\/\/overtone.github.io\/\">Overtone<\/a>and\u00a0<a href=\"http:\/\/github.com\/quil\/quil\">Quil<\/a>.\" <\/div><\/blockquote><p>It provides an excellent default setup for working with clojure.<\/p><h2>Emacs Modes for clojure development<\/h2><p>Among the most important modes are <a href=\"https:\/\/github.com\/clojure-emacs\/clojure-mode\">Clojure-mode<\/a> and <a href=\"http:\/\/www.emacswiki.org\/ParEdit\">Paredit<\/a> along with repl integration via <a href=\"https:\/\/github.com\/clojure-emacs\/cider\">Cider<\/a>.<\/p><h3>Paredit<\/h3><p>One of the most common complaints I hear from people who haven't tried clojure is about the abundant parentheses and how difficult they are to manage. Managing and matching parentheses is really a problem for the text editor, not the programmer. If you have trouble with unmatched parentheses, then either your text editor is not up to the task or you're not using it correctly.<\/p><p>Paredit mode is an emacs mode that takes care of the minutiae of working with s-expressions. Some of the things it does for you include:<\/p><ul><li>Ensures your parentheses are always balanced. E.g. It won't allow you to create an opening bracket without the corresponding bracket, or it won't let you delete a single bracket - you have to delete the entire expression.\u00a0<\/li><li>Lets you cut, paste, move etc s-expressions in their entirety. You can move the entire sub-tree of an expression somewhere else or wrap an entire expression with another one.<\/li><\/ul><p>If you've never used Paredit before, it can be a little off-putting. You need to move from thinking in terms of text-tokens to thinking in terms of expressions. Once you get used to it though, the parentheses that so many people are wary of fade into the background. They become easy to manage.<\/p><p>To get a feel for Paredit, lets edit a small clojure file and play around with the brackets.<\/p><p>Lets create a new clojure project...<\/p><p>...and then open the generated core.clj file.<\/p><p id=\"yui_3_17_2_1_1411661917196_164775\">If you've setup emacs-live you should see the following list of active modes when you open this file:<\/p><p>\n<img src=\"http:\/\/www.aidanf.net\/images\/pareditmode.png\" alt=\"Active modes for a .clj file: Clojure, Undo-Tree, yas, VHl, AC, Paredit\" \/><br \/><b>Active modes for a .clj file: Clojure, Undo-Tree, yas, VHl, AC, Paredit<\/b>\n<\/p>\n<p>Now, try to delete a closing \")\". You can't. Navigate to the opening \"(\" of the println form. Press C-k to delete it. Notice that it deletes the entire form, but not the entire line - the last \")\" remains. Keep, pressing C-k - you can't delete that closing \")\"!<\/p><p>Press C-y to paste the println form back to where it was. Now highlight the println form (Press C-space on the opening \"(\" and then move the cursor to the closing \")\" ). Now if you enter an opening bracket \"(\" it will wrap that entire expression with brackets. There's a lot more stuff you can do with Paredit mode that facilitates working with lisp-forms - take some time to get comfortable with it.<\/p><h3>Clojure-mode<\/h3><p>This provides indentation, highlighting and navigation for your clojure code. You don't really need to learn to use it - just know that you need what it provides.<\/p><h3>Cider<\/h3><p>Cider provides a set of commands for starting and running a repl from emacs, and then sending code for evaluation to that repl. This is core to emacs clojure development. You get into a workflow of constantly evaluating and re-evaluating functions as you iterively develop them and incrementally build up a working program in the repl.\u00a0<\/p><p>Let's continue with our example project from above and launch a repl from emacs.<\/p><p><code>M-x nrepl-jack-in<\/code> <\/p>\n<p>will start a repl (This command may be cider-jack-in instead of nrepl-jack-in, depending on which version of cider you have installed). Once it's started you should be able to enter clojure code in the repl and see the results immediately.<\/p>\n<p>Let's evaluate some of our clojure code in the repl. The handiest way to do this is <code>C-c C-e<\/code>. This will evaluate the previous form in the text window and print the results.<\/p>\n<p>Lets navigate to the space just after the last closing \")\". Now if you run <code>C-c C-e<\/code> you should see the results of evaluating that function in your status window i.e. (<code>#hello.core\/foo<\/code>).<\/p>\n<p>Now switch you the repl window and change to the hello.core namespace.<\/p>\n<p><code>(ns hello.core)<\/code><\/p>\n<p>You can now run the foo function from the repl.<\/p>\n<p><code>(foo \"Goodbye\")<\/code><\/p>\n<p>Let's say our foo function is broken and we want to debug it. We'll set x in the repl and then we can step into our function and evaluate the forms in the function individually.<\/p>\n<p>In the repl enter <\/p>\n<p><code>(def x \"Broken\")<\/code><\/p>\n<p>Now go back to the core.clj file and navigate the cursor to the space after the closing bracket of the angle println form. If we press <code>C-c C-e<\/code> we should see \"Broken Hello, World!\"in the repl<\/p>\n<p>This is a very simple example, but in this way we can step through functions evaluating forms, seeing what every form in a function evaluates to for a given input. We can re-evaluate a function at any time to update the code that the repl is running. This gives us a lot of power to inspect how our code is behaving and to interively build up a working program in the repl.<\/p>\n<p>Here are a few more commands that are useful for interacting with the repl and evaluating code. You can find the full list of available commands in the <a href=\"https:\/\/github.com\/clojure-emacs\/cider#keyboard-shortcuts\">cider documentation<\/a>.<\/p><table><tr><td colspan=\"2\"><i>Buffer Shortcuts<\/i><\/td>\n    <\/tr><tr><td><b>C-c C-e<\/b><\/td>\n        <td>Evaluate the form preceding the cursor and print the result to the echo area<\/td>\n    <\/tr><tr><td><b>C-c C-r<\/b><\/td>\n        <td>Evaluate the highlighted region<\/td>\n    <\/tr><tr><td><b>C-c M-n<\/b><\/td>\n        <td>Switch the namespace of the repl buffer to the namespace of the current buffer<\/td>\n    <\/tr><tr><td><b>C-c C-z<\/b><\/td>\n        <td>Switch to the repl buffer (and back to previous buffer)<\/td>\n    <\/tr><tr><td><b>C-c C-k<\/b><\/td>\n        <td>Load (all) the current buffer<\/td>\n    <\/tr><tr><td><b>C-c C-d<\/b><\/td>\n        <td>Display the doc string for the symbol under the cursor<\/td>\n    <\/tr><tr><td colspan=\"2\"><i>Repl Shortcuts<\/i><\/td>\n    <\/tr><tr><td><b>M-p<\/b><\/td>\n        <td>previous repl command<\/td>\n    <\/tr><tr><td><b>M-n<\/b><\/td>\n        <td>next repl command<\/td>\n    <\/tr><tr><td><b>Tab<\/b><\/td>\n        <td>Autocomplete current symbol<\/td>\n    <\/tr><tr><td><b>C-c C-z<\/b><\/td>\n        <td>Back to previous buffer<\/td>\n    <\/tr><\/table><p>So that's a very quick and basic tour of how to get started with emacs for clojure development. The best thing to do now is to jump in and start using emacs for editing clojure code, while at the same time reading the paredit and cider documentation and incorporating them into your workflow as you go along.<\/p>","created_at":"2018-11-07T16:54:46+0000","updated_at":"2018-11-07T16:54:46+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"www.aidanf.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":666,"title":"overtone\/emacs-live","url":"https:\/\/github.com\/overtone\/emacs-live","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"http:\/\/github.com\/downloads\/overtone\/emacs-live\/emacs-live.png\"><img src=\"http:\/\/github.com\/downloads\/overtone\/emacs-live\/emacs-live.png\" alt=\"image\" \/><\/a><\/p>\n<pre>Energy starts surging through your fingertips. Sparks fly, making\nthe shadows grow and flicker ominously around you. You arch your\nback and open your eyes as if for the first time. The text appears\nmore vibrant and colourful than you remember. Your fingers dance on\nthe keyboard commanding the cursor with a joyful precision and\ncontrol.\nYou lean back and marvel as a shimmering swirl of syntactic and\nsemantic structures project out of the screen and intertwine with a\nfractal beauty. You watch as meaning recursively unfolds into deeper\nmeaning live in front of your eyes. You feel a deep and lasting\nsynchronicity form as the boundaries between you and your Emacs\nwash away. You and your Emacs Live.\nM-x start-hacking.\n<\/pre>\n<p>An opinionated set of defaults for getting started with a specific focus\non live coding with <a href=\"http:\/\/overtone.github.io\" rel=\"nofollow\">Overtone<\/a> and\n<a href=\"http:\/\/github.com\/quil\/quil\">Quil<\/a>. However, it's not just a one trick\npony. It also happens to be:<\/p>\n<ul><li>a jolly good generic Clojure hacking config<\/li>\n<li>a nice structured approach to organising your Emacs config<\/li>\n<li>modular in that functionality is organised by discrete <em>packs<\/em><\/li>\n<li>a decent starting point for live coding in general<\/li>\n<li>a goldmine of config snippets to plunder and add to your own config<\/li>\n<\/ul><p>So, wherever you are in the multiverse, Emacs Live is ready to join you\nin battle against the evil friction of poor text editor workflows.<\/p>\n<pre>\"Power of the horse, full force!\"\n             The Space Stallions.\n<\/pre>\n<h3><a id=\"user-content-requires-emacs-244\" class=\"anchor\" aria-hidden=\"true\" href=\"#requires-emacs-244\"><\/a>Requires Emacs 24.4<\/h3>\n<p>Emacs Live is only compatible with Emacs 24.4 and above.<\/p>\n<h3><a id=\"user-content-easy-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#easy-install\"><\/a>Easy Install<\/h3>\n<p>If you're the kind of shoot-from-the-hip brogrammer that doesn't mind\nexecuting random scripts from the interwebs, then the easiest way to\ninstall Emacs Live is to run the following which will safely preserve\nany Emacs configs you already have:<\/p>\n<pre>bash &lt;(curl -fksSL https:\/\/raw.github.com\/overtone\/emacs-live\/master\/installer\/install-emacs-live.sh)\n<\/pre>\n<p>Note: you should always read through random scripts before executing\nthem!<\/p>\n<h3><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting Started<\/h3>\n<p>The (only ever-so-slightly) more involved way to install is to follow\nthese steps:<\/p>\n<ol><li>Move aside <code>~\/.emacs<\/code>, <code>~\/.emacs.el<\/code> or <code>~\/.emacs.d<\/code> if they\ncurrently exist.<\/li>\n<li>Download the zip bundle (or clone the repository with git) and move\nand rename to <code>~\/.emacs.d<\/code><\/li>\n<li>Launch Emacs version 24+<\/li>\n<li>Live code your hat off!<\/li>\n<\/ol><h3><a id=\"user-content-clojure-hacking\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure-hacking\"><\/a>Clojure Hacking<\/h3>\n<p>If you wish to hack with Clojure projects such as\n<a href=\"http:\/\/github.com\/overtone\/overtone\">Overtone<\/a> and\n<a href=\"http:\/\/github.com\/quil\/quil\">Quil<\/a> you'll need to install\n<a href=\"https:\/\/github.com\/technomancy\/leiningen\/wiki\/Upgrading\">Leiningen 2<\/a>\nand you're ready to roll.<\/p>\n<p>Make sure you create a profile in <code>~\/.lein\/profiles.clj<\/code>. It should\ncontain the following:<\/p>\n<pre>{:user {:plugins [[cider\/cider-nrepl \"0.12.0\"]\n                  [refactor-nrepl    \"2.2.0\"]]}}\n<\/pre>\n<p>Now you're ready to start a repl in a Clojure project with <code>lein repl<\/code>\nand connect to it from Emacs with <code>M-x cider-connect<\/code> (supplying the\nhost and the correct port) for full Emacs REPL\/autocompletion joy.<\/p>\n<h3><a id=\"user-content-screenshots\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshots\"><\/a>Screenshots<\/h3>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/overtone\/emacs-live\/blob\/master\/screenshots\/live-coding-config-in-use.png\"><img src=\"https:\/\/github.com\/overtone\/emacs-live\/raw\/master\/screenshots\/live-coding-config-in-use.png\" alt=\"Screenshot 1\" title=\"Live Coding Config Screenshot 1\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/overtone\/emacs-live\/blob\/master\/screenshots\/live-coding-config-in-use-2.png\"><img src=\"https:\/\/github.com\/overtone\/emacs-live\/raw\/master\/screenshots\/live-coding-config-in-use-2.png\" alt=\"Screenshot 2\" title=\"Live Coding Config Screenshot 2\" \/><\/a><\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#video\"><\/a>Video<\/h3>\n<p>Here's a video showing the config in use:\n<a href=\"http:\/\/vimeo.com\/22798433\" rel=\"nofollow\">Quick Intro to Live Programming with Overtone<\/a><\/p>\n<h2><a id=\"user-content-live-packs\" class=\"anchor\" aria-hidden=\"true\" href=\"#live-packs\"><\/a>Live Packs<\/h2>\n<p>Emacs Live is powered by a number of special packs. Packs are\ndirectories which are used to store isolated sets of functionality and\nconfiguration. These may be symlinks or git submodules depending on how\nyou choose to store and manage your dot emacs.<\/p>\n<h3><a id=\"user-content-pack-structure\" class=\"anchor\" aria-hidden=\"true\" href=\"#pack-structure\"><\/a>Pack Structure<\/h3>\n<p>Each pack consists of three components: the <code>init.el<\/code>, <code>config<\/code> dir and\n<code>lib<\/code> dir. The <code>init.el<\/code> file is loaded first and it is here that you\ncan run arbitrary elisp. However, it is recommended that you organise\nyour pack by placing library code in the lib dir and individual config\nfiles in the config dir. Emacs Live provides helper fns to make it easy\nfor you to load config files and for you to add lib dirs to your load\npath. See the section on helper fns below.<\/p>\n<h3><a id=\"user-content-shipped-packs\" class=\"anchor\" aria-hidden=\"true\" href=\"#shipped-packs\"><\/a>Shipped Packs<\/h3>\n<p>Emacs live ships with a few packs:<\/p>\n<h4><a id=\"user-content-foundation\" class=\"anchor\" aria-hidden=\"true\" href=\"#foundation\"><\/a>Foundation<\/h4>\n<p>A set of defaults to create a clutter free, friendly and more dynamic\nEmacs foundation. Also adds fuzzy matching autocomplete functionality\nfor most of the Emacs minibuffer actions - such as finding files, calling\nfunctions and switching buffers.<\/p>\n<h4><a id=\"user-content-colour\" class=\"anchor\" aria-hidden=\"true\" href=\"#colour\"><\/a>Colour<\/h4>\n<p>Colour highlighting in two flavours - cyberpunk and gandalf. Use\n<code>color-theme-cyberpunk<\/code> and <code>color-theme-gandalf<\/code> to switch between the\ntwo themes. Currently cyberpunk has seen more love - patch requests\naccepted for appropriate improvements to Gandalf.<\/p>\n<h4><a id=\"user-content-clojure\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure\"><\/a>Clojure<\/h4>\n<p>A set of goodies to get you hacking Clojure like a pro.<\/p>\n<ul><li>Clojure Mode (with fancy <code>(\u03bb [a] (+ a 5))<\/code> and <code>\u0192(+ % 5)<\/code> prettifications)<\/li>\n<li>cider (for communicating with nREPL servers)<\/li>\n<li>Auto completion (configured to work with nREPL for inline auto\ncompletion of documentation)<\/li>\n<li>Tailor-made colour theme<\/li>\n<li>Fancy highlighting of sexps on eval<\/li>\n<li>Rainbow parens and delimiters (to allow you to easily navigate a sea\nof parens)<\/li>\n<\/ul><h4><a id=\"user-content-lang\" class=\"anchor\" aria-hidden=\"true\" href=\"#lang\"><\/a>Lang<\/h4>\n<p>A number of extra language modes for your joy. Languages include:<\/p>\n<ul><li>Markdown<\/li>\n<li>Yaml<\/li>\n<li>Ruby<\/li>\n<li>SuperCollider<\/li>\n<li>JavaScript<\/li>\n<li>Extempore<\/li>\n<li>GLSL Shaders<\/li>\n<\/ul><h4><a id=\"user-content-power\" class=\"anchor\" aria-hidden=\"true\" href=\"#power\"><\/a>Power<\/h4>\n<p>A boost of fantastic functionality for your live-coding\nfingertips. Highlights include:<\/p>\n<ul><li>The amazing undo-tree (live-code with confidence!)<\/li>\n<li>Textmate-like snippets<\/li>\n<li>Refheap mode for pasting snippets to refheap.com<\/li>\n<li>Github Gist support<\/li>\n<li>Quick jump mode for accessing previous locations<\/li>\n<li>Ace jump mode for jumping directly to any symbol on the screen with 2\nor three keystrokes.<\/li>\n<\/ul><h3><a id=\"user-content-loading-packs\" class=\"anchor\" aria-hidden=\"true\" href=\"#loading-packs\"><\/a>Loading Packs<\/h3>\n<p>By default, Emacs live will load the packs in the following order:<\/p>\n<ol><li><code>foundation-pack<\/code><\/li>\n<li><code>colour-pack<\/code><\/li>\n<li><code>lang-pack<\/code><\/li>\n<li><code>power-pack<\/code><\/li>\n<li><code>git-pack<\/code><\/li>\n<li><code>org-pack<\/code><\/li>\n<li><code>clojure-pack<\/code><\/li>\n<li><code>bindings-pack<\/code><\/li>\n<\/ol><p>However, you may create a <code>~\/.emacs-live.el<\/code> file to override this\nbehaviour. Simply set the var live-packs to a list of symbols\nrepresenting the packs you'd like to load up (the order will be\nhonoured). For example to only load the foundation and colour packs:<\/p>\n<pre>(live-use-packs '(stable\/foundation-pack stable\/colour-pack))\n<\/pre>\n<p>If just you wish to load your own packs after the default packs then\nsimply use <code>live-add-packs<\/code>:<\/p>\n<pre>(live-add-packs '(~\/.live-packs\/yourname-pack))\n<\/pre>\n<p>Packs are expected to reside in <code>~\/.emacs.d\/packs\/<\/code> unless you specify\nthem with absolute paths in which case the absolute path with be\nhonoured.<\/p>\n<h3><a id=\"user-content-creating-your-own-packs\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-your-own-packs\"><\/a>Creating your own Packs<\/h3>\n<p>Emacs Live provides a couple of useful helper fns which you can use\nwithin your own live packs:<\/p>\n<ul><li><code>live-pack-lib-dir<\/code> this returns the path of the lib dir for the\ncurrent pack<\/li>\n<li><code>live-load-config-file<\/code> loads a config file located in the config dir\nof the current pack<\/li>\n<\/ul><p>It is recommended that you place your own personal packs in an external\ndirectory. See the <code>user\/template-pack<\/code>'s README for more information.<\/p>\n<h2><a id=\"user-content-mailing-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#mailing-list\"><\/a>Mailing List<\/h2>\n<p>I'm very happy to hear any feedback regarding this config. The idea is\nfor you to use it to get started and give you a platform to start\nediting it and turning it into something personal.<\/p>\n<p>Please make any comments on the Emacs Live mailing list:<\/p>\n<p><a href=\"http:\/\/groups.google.com\/group\/emacs-live\" rel=\"nofollow\">http:\/\/groups.google.com\/group\/emacs-live<\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:54:46+0000","updated_at":"2018-11-07T16:54:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/373359?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":667,"title":"hemanth\/functional-programming-jargon","url":"https:\/\/github.com\/hemanth\/functional-programming-jargon\/blob\/master\/readme.md","content":"<p>Functional programming (FP) provides many advantages, and its popularity has been increasing as a result. However, each programming paradigm comes with its own unique jargon and FP is no exception. By providing a glossary, we hope to make learning FP easier.<\/p>\n<p>Examples are presented in JavaScript (ES2015). <a href=\"https:\/\/github.com\/hemanth\/functional-programming-jargon\/wiki\/Why-JavaScript%3F\">Why JavaScript?<\/a><\/p>\n<p><em>This is a <a href=\"https:\/\/github.com\/hemanth\/functional-programming-jargon\/issues\/20\">WIP<\/a>; please feel free to send a PR ;)<\/em><\/p>\n<p>Where applicable, this document uses terms defined in the <a href=\"https:\/\/github.com\/fantasyland\/fantasy-land\">Fantasy Land spec<\/a><\/p>\n<p><strong>Translations<\/strong><\/p>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#arity\">Arity<\/a><\/li>\n<li><a href=\"#higher-order-functions-hof\">Higher-Order Functions (HOF)<\/a><\/li>\n<li><a href=\"#closure\">Closure<\/a><\/li>\n<li><a href=\"#partial-application\">Partial Application<\/a><\/li>\n<li><a href=\"#currying\">Currying<\/a><\/li>\n<li><a href=\"#auto-currying\">Auto Currying<\/a><\/li>\n<li><a href=\"#function-composition\">Function Composition<\/a><\/li>\n<li><a href=\"#continuation\">Continuation<\/a><\/li>\n<li><a href=\"#purity\">Purity<\/a><\/li>\n<li><a href=\"#side-effects\">Side effects<\/a><\/li>\n<li><a href=\"#idempotent\">Idempotent<\/a><\/li>\n<li><a href=\"#point-free-style\">Point-Free Style<\/a><\/li>\n<li><a href=\"#predicate\">Predicate<\/a><\/li>\n<li><a href=\"#contracts\">Contracts<\/a><\/li>\n<li><a href=\"#category\">Category<\/a><\/li>\n<li><a href=\"#value\">Value<\/a><\/li>\n<li><a href=\"#constant\">Constant<\/a><\/li>\n<li><a href=\"#functor\">Functor<\/a><\/li>\n<li><a href=\"#pointed-functor\">Pointed Functor<\/a><\/li>\n<li><a href=\"#lift\">Lift<\/a><\/li>\n<li><a href=\"#referential-transparency\">Referential Transparency<\/a><\/li>\n<li><a href=\"#equational-reasoning\">Equational Reasoning<\/a><\/li>\n<li><a href=\"#lambda\">Lambda<\/a><\/li>\n<li><a href=\"#lambda-calculus\">Lambda Calculus<\/a><\/li>\n<li><a href=\"#lazy-evaluation\">Lazy evaluation<\/a><\/li>\n<li><a href=\"#monoid\">Monoid<\/a><\/li>\n<li><a href=\"#monad\">Monad<\/a><\/li>\n<li><a href=\"#comonad\">Comonad<\/a><\/li>\n<li><a href=\"#applicative-functor\">Applicative Functor<\/a><\/li>\n<li><a href=\"#morphism\">Morphism<\/a>\n<ul><li><a href=\"#endomorphism\">Endomorphism<\/a><\/li>\n<li><a href=\"#isomorphism\">Isomorphism<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#setoid\">Setoid<\/a><\/li>\n<li><a href=\"#semigroup\">Semigroup<\/a><\/li>\n<li><a href=\"#foldable\">Foldable<\/a><\/li>\n<li><a href=\"#lens\">Lens<\/a><\/li>\n<li><a href=\"#type-signatures\">Type Signatures<\/a><\/li>\n<li><a href=\"#algebraic-data-type\">Algebraic data type<\/a>\n<ul><li><a href=\"#sum-type\">Sum type<\/a><\/li>\n<li><a href=\"#product-type\">Product type<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#option\">Option<\/a><\/li>\n<li><a href=\"#totality\">Totality<\/a><\/li>\n<li><a href=\"#functional-programming-libraries-in-javascript\">Functional Programming Libraries in JavaScript<\/a><\/li>\n<\/ul><h2><a id=\"user-content-arity\" class=\"anchor\" aria-hidden=\"true\" href=\"#arity\"><\/a>Arity<\/h2>\n<p>The number of arguments a function takes. From words like unary, binary, ternary, etc. This word has the distinction of being composed of two suffixes, \"-ary\" and \"-ity.\" Addition, for example, takes two arguments, and so it is defined as a binary function or a function with an arity of two. Such a function may sometimes be called \"dyadic\" by people who prefer Greek roots to Latin. Likewise, a function that takes a variable number of arguments is called \"variadic,\" whereas a binary function must be given two and only two arguments, currying and partial application notwithstanding (see below).<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const sum = (a, b) =&gt; a + b\nconst arity = sum.length\nconsole.log(arity) \/\/ 2\n\/\/ The arity of sum is 2<\/pre><\/div>\n<h2><a id=\"user-content-higher-order-functions-hof\" class=\"anchor\" aria-hidden=\"true\" href=\"#higher-order-functions-hof\"><\/a>Higher-Order Functions (HOF)<\/h2>\n<p>A function which takes a function as an argument and\/or returns a function.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const filter = (predicate, xs) =&gt; xs.filter(predicate)<\/pre><\/div>\n<div class=\"highlight highlight-source-js\"><pre>const is = (type) =&gt; (x) =&gt; Object(x) instanceof type<\/pre><\/div>\n<div class=\"highlight highlight-source-js\"><pre>filter(is(Number), [0, '1', 2, null]) \/\/ [0, 2]<\/pre><\/div>\n<h2><a id=\"user-content-closure\" class=\"anchor\" aria-hidden=\"true\" href=\"#closure\"><\/a>Closure<\/h2>\n<p>A closure is a scope which retains variables available to a function when it's created. This is important for\n<a href=\"#partial-application\">partial application<\/a> to work.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const addTo = (x) =&gt; {\n    return (y) =&gt; {\n        return x + y\n    }\n}<\/pre><\/div>\n<p>We can call <code>addTo<\/code> with a number and get back a function with a baked-in <code>x<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var addToFive = addTo(5)<\/pre><\/div>\n<p>In this case the <code>x<\/code> is retained in <code>addToFive<\/code>'s closure with the value <code>5<\/code>. We can then call <code>addToFive<\/code> with the <code>y<\/code>\nand get back the desired number.<\/p>\n<pre>addToFive(3) \/\/ =&gt; 8\n<\/pre>\n<p>This works because variables that are in parent scopes are not garbage-collected as long as the function itself is retained.<\/p>\n<p>Closures are commonly used in event handlers so that they still have access to variables defined in their parents when they\nare eventually called.<\/p>\n<p><strong>Further reading<\/strong><\/p>\n<ul><li><a href=\"http:\/\/stackoverflow.com\/questions\/220658\/what-is-the-difference-between-a-closure-and-a-lambda\" rel=\"nofollow\">Lambda Vs Closure<\/a><\/li>\n<li><a href=\"http:\/\/stackoverflow.com\/questions\/111102\/how-do-javascript-closures-work\" rel=\"nofollow\">How do JavaScript Closures Work?<\/a><\/li>\n<\/ul><h2><a id=\"user-content-partial-application\" class=\"anchor\" aria-hidden=\"true\" href=\"#partial-application\"><\/a>Partial Application<\/h2>\n<p>Partially applying a function means creating a new function by pre-filling some of the arguments to the original function.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Helper to create partially applied functions\n\/\/ Takes a function and some arguments\nconst partial = (f, ...args) =&gt;\n  \/\/ returns a function that takes the rest of the arguments\n  (...moreArgs) =&gt;\n    \/\/ and calls the original function with all of them\n    f(...args, ...moreArgs)\n\/\/ Something to apply\nconst add3 = (a, b, c) =&gt; a + b + c\n\/\/ Partially applying `2` and `3` to `add3` gives you a one-argument function\nconst fivePlus = partial(add3, 2, 3) \/\/ (c) =&gt; 2 + 3 + c\nfivePlus(4) \/\/ 9<\/pre><\/div>\n<p>You can also use <code>Function.prototype.bind<\/code> to partially apply a function in JS:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const add1More = add3.bind(null, 2, 3) \/\/ (c) =&gt; 2 + 3 + c<\/pre><\/div>\n<p>Partial application helps create simpler functions from more complex ones by baking in data when you have it. <a href=\"#currying\">Curried<\/a> functions are automatically partially applied.<\/p>\n<h2><a id=\"user-content-currying\" class=\"anchor\" aria-hidden=\"true\" href=\"#currying\"><\/a>Currying<\/h2>\n<p>The process of converting a function that takes multiple arguments into a function that takes them one at a time.<\/p>\n<p>Each time the function is called it only accepts one argument and returns a function that takes one argument until all arguments are passed.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const sum = (a, b) =&gt; a + b\nconst curriedSum = (a) =&gt; (b) =&gt; a + b\ncurriedSum(40)(2) \/\/ 42.\nconst add2 = curriedSum(2) \/\/ (b) =&gt; 2 + b\nadd2(10) \/\/ 12\n<\/pre><\/div>\n<h2><a id=\"user-content-auto-currying\" class=\"anchor\" aria-hidden=\"true\" href=\"#auto-currying\"><\/a>Auto Currying<\/h2>\n<p>Transforming a function that takes multiple arguments into one that if given less than its correct number of arguments returns a function that takes the rest. When the function gets the correct number of arguments it is then evaluated.<\/p>\n<p>lodash &amp; Ramda have a <code>curry<\/code> function that works this way.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const add = (x, y) =&gt; x + y\nconst curriedAdd = _.curry(add)\ncurriedAdd(1, 2) \/\/ 3\ncurriedAdd(1) \/\/ (y) =&gt; 1 + y\ncurriedAdd(1)(2) \/\/ 3<\/pre><\/div>\n<p><strong>Further reading<\/strong><\/p>\n<ul><li><a href=\"http:\/\/fr.umio.us\/favoring-curry\/\" rel=\"nofollow\">Favoring Curry<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=m3svKOdZijA\" rel=\"nofollow\">Hey Underscore, You're Doing It Wrong!<\/a><\/li>\n<\/ul><h2><a id=\"user-content-function-composition\" class=\"anchor\" aria-hidden=\"true\" href=\"#function-composition\"><\/a>Function Composition<\/h2>\n<p>The act of putting two functions together to form a third function where the output of one function is the input of the other.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const compose = (f, g) =&gt; (a) =&gt; f(g(a)) \/\/ Definition\nconst floorAndToString = compose((val) =&gt; val.toString(), Math.floor) \/\/ Usage\nfloorAndToString(121.212121) \/\/ '121'<\/pre><\/div>\n<h2><a id=\"user-content-continuation\" class=\"anchor\" aria-hidden=\"true\" href=\"#continuation\"><\/a>Continuation<\/h2>\n<p>At any given point in a program, the part of the code that's yet to be executed is known as a continuation.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const printAsString = (num) =&gt; console.log(`Given ${num}`)\nconst addOneAndContinue = (num, cc) =&gt; {\n  const result = num + 1\n  cc(result)\n}\naddOneAndContinue(2, printAsString) \/\/ 'Given 3'<\/pre><\/div>\n<p>Continuations are often seen in asynchronous programming when the program needs to wait to receive data before it can continue. The response is often passed off to the rest of the program, which is the continuation, once it's been received.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const continueProgramWith = (data) =&gt; {\n  \/\/ Continues program with data\n}\nreadFileAsync('path\/to\/file', (err, response) =&gt; {\n  if (err) {\n    \/\/ handle error\n    return\n  }\n  continueProgramWith(response)\n})<\/pre><\/div>\n<h2><a id=\"user-content-purity\" class=\"anchor\" aria-hidden=\"true\" href=\"#purity\"><\/a>Purity<\/h2>\n<p>A function is pure if the return value is only determined by its\ninput values, and does not produce side effects.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const greet = (name) =&gt; `Hi, ${name}`\ngreet('Brianne') \/\/ 'Hi, Brianne'<\/pre><\/div>\n<p>As opposed to each of the following:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>window.name = 'Brianne'\nconst greet = () =&gt; `Hi, ${window.name}`\ngreet() \/\/ \"Hi, Brianne\"<\/pre><\/div>\n<p>The above example's output is based on data stored outside of the function...<\/p>\n<div class=\"highlight highlight-source-js\"><pre>let greeting\nconst greet = (name) =&gt; {\n  greeting = `Hi, ${name}`\n}\ngreet('Brianne')\ngreeting \/\/ \"Hi, Brianne\"<\/pre><\/div>\n<p>... and this one modifies state outside of the function.<\/p>\n<h2><a id=\"user-content-side-effects\" class=\"anchor\" aria-hidden=\"true\" href=\"#side-effects\"><\/a>Side effects<\/h2>\n<p>A function or expression is said to have a side effect if apart from returning a value, it interacts with (reads from or writes to) external mutable state.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const differentEveryTime = new Date()<\/pre><\/div>\n<div class=\"highlight highlight-source-js\"><pre>console.log('IO is a side effect!')<\/pre><\/div>\n<h2><a id=\"user-content-idempotent\" class=\"anchor\" aria-hidden=\"true\" href=\"#idempotent\"><\/a>Idempotent<\/h2>\n<p>A function is idempotent if reapplying it to its result does not produce a different result.<\/p>\n<pre>f(f(x)) \u224d f(x)\n<\/pre>\n<div class=\"highlight highlight-source-js\"><pre>Math.abs(Math.abs(10))<\/pre><\/div>\n<div class=\"highlight highlight-source-js\"><pre>sort(sort(sort([2, 1])))<\/pre><\/div>\n<h2><a id=\"user-content-point-free-style\" class=\"anchor\" aria-hidden=\"true\" href=\"#point-free-style\"><\/a>Point-Free Style<\/h2>\n<p>Writing functions where the definition does not explicitly identify the arguments used. This style usually requires <a href=\"#currying\">currying<\/a> or other <a href=\"#higher-order-functions-hof\">Higher-Order functions<\/a>. A.K.A Tacit programming.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Given\nconst map = (fn) =&gt; (list) =&gt; list.map(fn)\nconst add = (a) =&gt; (b) =&gt; a + b\n\/\/ Then\n\/\/ Not points-free - `numbers` is an explicit argument\nconst incrementAll = (numbers) =&gt; map(add(1))(numbers)\n\/\/ Points-free - The list is an implicit argument\nconst incrementAll2 = map(add(1))<\/pre><\/div>\n<p><code>incrementAll<\/code> identifies and uses the parameter <code>numbers<\/code>, so it is not points-free.  <code>incrementAll2<\/code> is written just by combining functions and values, making no mention of its arguments.  It <strong>is<\/strong> points-free.<\/p>\n<p>Points-free function definitions look just like normal assignments without <code>function<\/code> or <code>=&gt;<\/code>.<\/p>\n<h2><a id=\"user-content-predicate\" class=\"anchor\" aria-hidden=\"true\" href=\"#predicate\"><\/a>Predicate<\/h2>\n<p>A predicate is a function that returns true or false for a given value. A common use of a predicate is as the callback for array filter.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const predicate = (a) =&gt; a &gt; 2\n;[1, 2, 3, 4].filter(predicate) \/\/ [3, 4]<\/pre><\/div>\n<h2><a id=\"user-content-contracts\" class=\"anchor\" aria-hidden=\"true\" href=\"#contracts\"><\/a>Contracts<\/h2>\n<p>A contract specifies the obligations and guarantees of the behavior from a function or expression at runtime. This acts as a set of rules that are expected from the input and output of a function or expression, and errors are generally reported whenever a contract is violated.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Define our contract : int -&gt; int\nconst contract = (input) =&gt; {\n  if (typeof input === 'number') return true\n  throw new Error('Contract violated: expected int -&gt; int')\n}\nconst addOne = (num) =&gt; contract(num) &amp;&amp; num + 1\naddOne(2) \/\/ 3\naddOne('some string') \/\/ Contract violated: expected int -&gt; int<\/pre><\/div>\n<h2><a id=\"user-content-category\" class=\"anchor\" aria-hidden=\"true\" href=\"#category\"><\/a>Category<\/h2>\n<p>A category in category theory is a collection of objects and morphisms between them. In programming, typically types\nact as the objects and functions as morphisms.<\/p>\n<p>To be a valid category 3 rules must be met:<\/p>\n<ol><li>There must be an identity morphism that maps an object to itself.\nWhere <code>a<\/code> is an object in some category,\nthere must be a function from <code>a -&gt; a<\/code>.<\/li>\n<li>Morphisms must compose.\nWhere <code>a<\/code>, <code>b<\/code>, and <code>c<\/code> are objects in some category,\nand <code>f<\/code> is a morphism from <code>a -&gt; b<\/code>, and <code>g<\/code> is a morphism from <code>b -&gt; c<\/code>;\n<code>g(f(x))<\/code> must be equivalent to <code>(g \u2022 f)(x)<\/code>.<\/li>\n<li>Composition must be associative\n<code>f \u2022 (g \u2022 h)<\/code> is the same as <code>(f \u2022 g) \u2022 h<\/code><\/li>\n<\/ol><p>Since these rules govern composition at very abstract level, category theory is great at uncovering new ways of composing things.<\/p>\n<p><strong>Further reading<\/strong><\/p>\n<ul><li><a href=\"https:\/\/bartoszmilewski.com\/2014\/10\/28\/category-theory-for-programmers-the-preface\/\" rel=\"nofollow\">Category Theory for Programmers<\/a><\/li>\n<\/ul><h2><a id=\"user-content-value\" class=\"anchor\" aria-hidden=\"true\" href=\"#value\"><\/a>Value<\/h2>\n<p>Anything that can be assigned to a variable.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>5\nObject.freeze({name: 'John', age: 30}) \/\/ The `freeze` function enforces immutability.\n;(a) =&gt; a\n;[1]\nundefined<\/pre><\/div>\n<h2><a id=\"user-content-constant\" class=\"anchor\" aria-hidden=\"true\" href=\"#constant\"><\/a>Constant<\/h2>\n<p>A variable that cannot be reassigned once defined.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const five = 5\nconst john = Object.freeze({name: 'John', age: 30})<\/pre><\/div>\n<p>Constants are <a href=\"#referential-transparency\">referentially transparent<\/a>. That is, they can be replaced with the values that they represent without affecting the result.<\/p>\n<p>With the above two constants the following expression will always return <code>true<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>john.age + five === ({name: 'John', age: 30}).age + (5)<\/pre><\/div>\n<h2><a id=\"user-content-functor\" class=\"anchor\" aria-hidden=\"true\" href=\"#functor\"><\/a>Functor<\/h2>\n<p>An object that implements a <code>map<\/code> function which, while running over each value in the object to produce a new object, adheres to two rules:<\/p>\n<p><strong>Preserves identity<\/strong><\/p>\n<pre>object.map(x =&gt; x) \u224d object\n<\/pre>\n<p><strong>Composable<\/strong><\/p>\n<pre>object.map(compose(f, g)) \u224d object.map(g).map(f)\n<\/pre>\n<p>(<code>f<\/code>, <code>g<\/code> are arbitrary functions)<\/p>\n<p>A common functor in JavaScript is <code>Array<\/code> since it abides to the two functor rules:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>;[1, 2, 3].map(x =&gt; x) \/\/ = [1, 2, 3]<\/pre><\/div>\n<p>and<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const f = x =&gt; x + 1\nconst g = x =&gt; x * 2\n;[1, 2, 3].map(x =&gt; f(g(x))) \/\/ = [3, 5, 7]\n;[1, 2, 3].map(g).map(f)     \/\/ = [3, 5, 7]<\/pre><\/div>\n<h2><a id=\"user-content-pointed-functor\" class=\"anchor\" aria-hidden=\"true\" href=\"#pointed-functor\"><\/a>Pointed Functor<\/h2>\n<p>An object with an <code>of<\/code> function that puts <em>any<\/em> single value into it.<\/p>\n<p>ES2015 adds <code>Array.of<\/code> making arrays a pointed functor.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Array.of(1) \/\/ [1]<\/pre><\/div>\n<h2><a id=\"user-content-lift\" class=\"anchor\" aria-hidden=\"true\" href=\"#lift\"><\/a>Lift<\/h2>\n<p>Lifting is when you take a value and put it into an object like a <a href=\"#pointed-functor\">functor<\/a>. If you lift a function into an <a href=\"#applicative-functor\">Applicative Functor<\/a> then you can make it work on values that are also in that functor.<\/p>\n<p>Some implementations have a function called <code>lift<\/code>, or <code>liftA2<\/code> to make it easier to run functions on functors.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const liftA2 = (f) =&gt; (a, b) =&gt; a.map(f).ap(b) \/\/ note it's `ap` and not `map`.\nconst mult = a =&gt; b =&gt; a * b\nconst liftedMult = liftA2(mult) \/\/ this function now works on functors like array\nliftedMult([1, 2], [3]) \/\/ [3, 6]\nliftA2(a =&gt; b =&gt; a + b)([1, 2], [3, 4]) \/\/ [4, 5, 5, 6]<\/pre><\/div>\n<p>Lifting a one-argument function and applying it does the same thing as <code>map<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const increment = (x) =&gt; x + 1\nlift(increment)([2]) \/\/ [3]\n;[2].map(increment) \/\/ [3]<\/pre><\/div>\n<h2><a id=\"user-content-referential-transparency\" class=\"anchor\" aria-hidden=\"true\" href=\"#referential-transparency\"><\/a>Referential Transparency<\/h2>\n<p>An expression that can be replaced with its value without changing the\nbehavior of the program is said to be referentially transparent.<\/p>\n<p>Say we have function greet:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const greet = () =&gt; 'Hello World!'<\/pre><\/div>\n<p>Any invocation of <code>greet()<\/code> can be replaced with <code>Hello World!<\/code> hence greet is\nreferentially transparent.<\/p>\n<h2><a id=\"user-content-equational-reasoning\" class=\"anchor\" aria-hidden=\"true\" href=\"#equational-reasoning\"><\/a>Equational Reasoning<\/h2>\n<p>When an application is composed of expressions and devoid of side effects, truths about the system can be derived from the parts.<\/p>\n<h2><a id=\"user-content-lambda\" class=\"anchor\" aria-hidden=\"true\" href=\"#lambda\"><\/a>Lambda<\/h2>\n<p>An anonymous function that can be treated like a value.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>;(function (a) {\n  return a + 1\n})\n;(a) =&gt; a + 1<\/pre><\/div>\n<p>Lambdas are often passed as arguments to Higher-Order functions.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>;[1, 2].map((a) =&gt; a + 1) \/\/ [2, 3]<\/pre><\/div>\n<p>You can assign a lambda to a variable.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const add1 = (a) =&gt; a + 1<\/pre><\/div>\n<h2><a id=\"user-content-lambda-calculus\" class=\"anchor\" aria-hidden=\"true\" href=\"#lambda-calculus\"><\/a>Lambda Calculus<\/h2>\n<p>A branch of mathematics that uses functions to create a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lambda_calculus\" rel=\"nofollow\">universal model of computation<\/a>.<\/p>\n<h2><a id=\"user-content-lazy-evaluation\" class=\"anchor\" aria-hidden=\"true\" href=\"#lazy-evaluation\"><\/a>Lazy evaluation<\/h2>\n<p>Lazy evaluation is a call-by-need evaluation mechanism that delays the evaluation of an expression until its value is needed. In functional languages, this allows for structures like infinite lists, which would not normally be available in an imperative language where the sequencing of commands is significant.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const rand = function*() {\n  while (1 &lt; 2) {\n    yield Math.random()\n  }\n}<\/pre><\/div>\n<div class=\"highlight highlight-source-js\"><pre>const randIter = rand()\nrandIter.next() \/\/ Each execution gives a random value, expression is evaluated on need.<\/pre><\/div>\n<h2><a id=\"user-content-monoid\" class=\"anchor\" aria-hidden=\"true\" href=\"#monoid\"><\/a>Monoid<\/h2>\n<p>An object with a function that \"combines\" that object with another of the same type.<\/p>\n<p>One simple monoid is the addition of numbers:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>1 + 1 \/\/ 2<\/pre><\/div>\n<p>In this case number is the object and <code>+<\/code> is the function.<\/p>\n<p>An \"identity\" value must also exist that when combined with a value doesn't change it.<\/p>\n<p>The identity value for addition is <code>0<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>1 + 0 \/\/ 1<\/pre><\/div>\n<p>It's also required that the grouping of operations will not affect the result (associativity):<\/p>\n<div class=\"highlight highlight-source-js\"><pre>1 + (2 + 3) === (1 + 2) + 3 \/\/ true<\/pre><\/div>\n<p>Array concatenation also forms a monoid:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>;[1, 2].concat([3, 4]) \/\/ [1, 2, 3, 4]<\/pre><\/div>\n<p>The identity value is empty array <code>[]<\/code><\/p>\n<div class=\"highlight highlight-source-js\"><pre>;[1, 2].concat([]) \/\/ [1, 2]<\/pre><\/div>\n<p>If identity and compose functions are provided, functions themselves form a monoid:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const identity = (a) =&gt; a\nconst compose = (f, g) =&gt; (x) =&gt; f(g(x))<\/pre><\/div>\n<p><code>foo<\/code> is any function that takes one argument.<\/p>\n<pre>compose(foo, identity) \u224d compose(identity, foo) \u224d foo\n<\/pre>\n<h2><a id=\"user-content-monad\" class=\"anchor\" aria-hidden=\"true\" href=\"#monad\"><\/a>Monad<\/h2>\n<p>A monad is an object with <a href=\"#pointed-functor\"><code>of<\/code><\/a> and <code>chain<\/code> functions. <code>chain<\/code> is like <a href=\"#functor\"><code>map<\/code><\/a> except it un-nests the resulting nested object.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Implementation\nArray.prototype.chain = function (f) {\n  return this.reduce((acc, it) =&gt; acc.concat(f(it)), [])\n}\n\/\/ Usage\nArray.of('cat,dog', 'fish,bird').chain((a) =&gt; a.split(',')) \/\/ ['cat', 'dog', 'fish', 'bird']\n\/\/ Contrast to map\nArray.of('cat,dog', 'fish,bird').map((a) =&gt; a.split(',')) \/\/ [['cat', 'dog'], ['fish', 'bird']]<\/pre><\/div>\n<p><code>of<\/code> is also known as <code>return<\/code> in other functional languages.\n<code>chain<\/code> is also known as <code>flatmap<\/code> and <code>bind<\/code> in other languages.<\/p>\n<h2><a id=\"user-content-comonad\" class=\"anchor\" aria-hidden=\"true\" href=\"#comonad\"><\/a>Comonad<\/h2>\n<p>An object that has <code>extract<\/code> and <code>extend<\/code> functions.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const CoIdentity = (v) =&gt; ({\n  val: v,\n  extract () {\n    return this.val\n  },\n  extend (f) {\n    return CoIdentity(f(this))\n  }\n})<\/pre><\/div>\n<p>Extract takes a value out of a functor.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>CoIdentity(1).extract() \/\/ 1<\/pre><\/div>\n<p>Extend runs a function on the comonad. The function should return the same type as the comonad.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>CoIdentity(1).extend((co) =&gt; co.extract() + 1) \/\/ CoIdentity(2)<\/pre><\/div>\n<h2><a id=\"user-content-applicative-functor\" class=\"anchor\" aria-hidden=\"true\" href=\"#applicative-functor\"><\/a>Applicative Functor<\/h2>\n<p>An applicative functor is an object with an <code>ap<\/code> function. <code>ap<\/code> applies a function in the object to a value in another object of the same type.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Implementation\nArray.prototype.ap = function (xs) {\n  return this.reduce((acc, f) =&gt; acc.concat(xs.map(f)), [])\n}\n\/\/ Example usage\n;[(a) =&gt; a + 1].ap([1]) \/\/ [2]<\/pre><\/div>\n<p>This is useful if you have two objects and you want to apply a binary function to their contents.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Arrays that you want to combine\nconst arg1 = [1, 3]\nconst arg2 = [4, 5]\n\/\/ combining function - must be curried for this to work\nconst add = (x) =&gt; (y) =&gt; x + y\nconst partiallyAppliedAdds = [add].ap(arg1) \/\/ [(y) =&gt; 1 + y, (y) =&gt; 3 + y]<\/pre><\/div>\n<p>This gives you an array of functions that you can call <code>ap<\/code> on to get the result:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>partiallyAppliedAdds.ap(arg2) \/\/ [5, 6, 7, 8]<\/pre><\/div>\n<h2><a id=\"user-content-morphism\" class=\"anchor\" aria-hidden=\"true\" href=\"#morphism\"><\/a>Morphism<\/h2>\n<p>A transformation function.<\/p>\n<h3><a id=\"user-content-endomorphism\" class=\"anchor\" aria-hidden=\"true\" href=\"#endomorphism\"><\/a>Endomorphism<\/h3>\n<p>A function where the input type is the same as the output.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ uppercase :: String -&gt; String\nconst uppercase = (str) =&gt; str.toUpperCase()\n\/\/ decrement :: Number -&gt; Number\nconst decrement = (x) =&gt; x - 1<\/pre><\/div>\n<h3><a id=\"user-content-isomorphism\" class=\"anchor\" aria-hidden=\"true\" href=\"#isomorphism\"><\/a>Isomorphism<\/h3>\n<p>A pair of transformations between 2 types of objects that is structural in nature and no data is lost.<\/p>\n<p>For example, 2D coordinates could be stored as an array <code>[2,3]<\/code> or object <code>{x: 2, y: 3}<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Providing functions to convert in both directions makes them isomorphic.\nconst pairToCoords = (pair) =&gt; ({x: pair[0], y: pair[1]})\nconst coordsToPair = (coords) =&gt; [coords.x, coords.y]\ncoordsToPair(pairToCoords([1, 2])) \/\/ [1, 2]\npairToCoords(coordsToPair({x: 1, y: 2})) \/\/ {x: 1, y: 2}<\/pre><\/div>\n<h2><a id=\"user-content-setoid\" class=\"anchor\" aria-hidden=\"true\" href=\"#setoid\"><\/a>Setoid<\/h2>\n<p>An object that has an <code>equals<\/code> function which can be used to compare other objects of the same type.<\/p>\n<p>Make array a setoid:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Array.prototype.equals = function (arr) {\n  const len = this.length\n  if (len !== arr.length) {\n    return false\n  }\n  for (let i = 0; i &lt; len; i++) {\n    if (this[i] !== arr[i]) {\n      return false\n    }\n  }\n  return true\n}\n;[1, 2].equals([1, 2]) \/\/ true\n;[1, 2].equals([0]) \/\/ false<\/pre><\/div>\n<h2><a id=\"user-content-semigroup\" class=\"anchor\" aria-hidden=\"true\" href=\"#semigroup\"><\/a>Semigroup<\/h2>\n<p>An object that has a <code>concat<\/code> function that combines it with another object of the same type.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>;[1].concat([2]) \/\/ [1, 2]<\/pre><\/div>\n<h2><a id=\"user-content-foldable\" class=\"anchor\" aria-hidden=\"true\" href=\"#foldable\"><\/a>Foldable<\/h2>\n<p>An object that has a <code>reduce<\/code> function that can transform that object into some other type.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const sum = (list) =&gt; list.reduce((acc, val) =&gt; acc + val, 0)\nsum([1, 2, 3]) \/\/ 6<\/pre><\/div>\n<h2><a id=\"user-content-lens\" class=\"anchor\" aria-hidden=\"true\" href=\"#lens\"><\/a>Lens<\/h2>\n<p>A lens is a structure (often an object or function) that pairs a getter and a non-mutating setter for some other data\nstructure.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Using [Ramda's lens](http:\/\/ramdajs.com\/docs\/#lens)\nconst nameLens = R.lens(\n  \/\/ getter for name property on an object\n  (obj) =&gt; obj.name,\n  \/\/ setter for name property\n  (val, obj) =&gt; Object.assign({}, obj, {name: val})\n)<\/pre><\/div>\n<p>Having the pair of get and set for a given data structure enables a few key features.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>const person = {name: 'Gertrude Blanch'}\n\/\/ invoke the getter\nR.view(nameLens, person) \/\/ 'Gertrude Blanch'\n\/\/ invoke the setter\nR.set(nameLens, 'Shafi Goldwasser', person) \/\/ {name: 'Shafi Goldwasser'}\n\/\/ run a function on the value in the structure\nR.over(nameLens, uppercase, person) \/\/ {name: 'GERTRUDE BLANCH'}<\/pre><\/div>\n<p>Lenses are also composable. This allows easy immutable updates to deeply nested data.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ This lens focuses on the first item in a non-empty array\nconst firstLens = R.lens(\n  \/\/ get first item in array\n  xs =&gt; xs[0],\n  \/\/ non-mutating setter for first item in array\n  (val, [__, ...xs]) =&gt; [val, ...xs]\n)\nconst people = [{name: 'Gertrude Blanch'}, {name: 'Shafi Goldwasser'}]\n\/\/ Despite what you may assume, lenses compose left-to-right.\nR.over(compose(firstLens, nameLens), uppercase, people) \/\/ [{'name': 'GERTRUDE BLANCH'}, {'name': 'Shafi Goldwasser'}]<\/pre><\/div>\n<p>Other implementations:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/calmm-js\/partial.lenses\">partial.lenses<\/a> - Tasty syntax sugar and a lot of powerful features<\/li>\n<li><a href=\"http:\/\/www.kovach.me\/nanoscope\/\" rel=\"nofollow\">nanoscope<\/a> - Fluent-interface<\/li>\n<\/ul><h2><a id=\"user-content-type-signatures\" class=\"anchor\" aria-hidden=\"true\" href=\"#type-signatures\"><\/a>Type Signatures<\/h2>\n<p>Often functions in JavaScript will include comments that indicate the types of their arguments and return values.<\/p>\n<p>There's quite a bit of variance across the community but they often follow the following patterns:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ functionName :: firstArgType -&gt; secondArgType -&gt; returnType\n\/\/ add :: Number -&gt; Number -&gt; Number\nconst add = (x) =&gt; (y) =&gt; x + y\n\/\/ increment :: Number -&gt; Number\nconst increment = (x) =&gt; x + 1<\/pre><\/div>\n<p>If a function accepts another function as an argument it is wrapped in parentheses.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ call :: (a -&gt; b) -&gt; a -&gt; b\nconst call = (f) =&gt; (x) =&gt; f(x)<\/pre><\/div>\n<p>The letters <code>a<\/code>, <code>b<\/code>, <code>c<\/code>, <code>d<\/code> are used to signify that the argument can be of any type. The following version of <code>map<\/code> takes a function that transforms a value of some type <code>a<\/code> into another type <code>b<\/code>, an array of values of type <code>a<\/code>, and returns an array of values of type <code>b<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ map :: (a -&gt; b) -&gt; [a] -&gt; [b]\nconst map = (f) =&gt; (list) =&gt; list.map(f)<\/pre><\/div>\n<p><strong>Further reading<\/strong><\/p>\n<h2><a id=\"user-content-algebraic-data-type\" class=\"anchor\" aria-hidden=\"true\" href=\"#algebraic-data-type\"><\/a>Algebraic data type<\/h2>\n<p>A composite type made from putting other types together. Two common classes of algebraic types are <a href=\"#sum-type\">sum<\/a> and <a href=\"#product-type\">product<\/a>.<\/p>\n<h3><a id=\"user-content-sum-type\" class=\"anchor\" aria-hidden=\"true\" href=\"#sum-type\"><\/a>Sum type<\/h3>\n<p>A Sum type is the combination of two types together into another one. It is called sum because the number of possible values in the result type is the sum of the input types.<\/p>\n<p>JavaScript doesn't have types like this but we can use <code>Set<\/code>s to pretend:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ imagine that rather than sets here we have types that can only have these values\nconst bools = new Set([true, false])\nconst halfTrue = new Set(['half-true'])\n\/\/ The weakLogic type contains the sum of the values from bools and halfTrue\nconst weakLogicValues = new Set([...bools, ...halfTrue])<\/pre><\/div>\n<p>Sum types are sometimes called union types, discriminated unions, or tagged unions.<\/p>\n<p>There's a <a href=\"https:\/\/github.com\/paldepind\/union-type\">couple<\/a> <a href=\"https:\/\/github.com\/puffnfresh\/daggy\">libraries<\/a> in JS which help with defining and using union types.<\/p>\n<p>Flow includes <a href=\"https:\/\/flow.org\/en\/docs\/types\/unions\/\" rel=\"nofollow\">union types<\/a> and TypeScript has <a href=\"https:\/\/www.typescriptlang.org\/docs\/handbook\/enums.html\" rel=\"nofollow\">Enums<\/a> to serve the same role.<\/p>\n<h3><a id=\"user-content-product-type\" class=\"anchor\" aria-hidden=\"true\" href=\"#product-type\"><\/a>Product type<\/h3>\n<p>A <strong>product<\/strong> type combines types together in a way you're probably more familiar with:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ point :: (Number, Number) -&gt; {x: Number, y: Number}\nconst point = (x, y) =&gt; ({ x, y })<\/pre><\/div>\n<p>It's called a product because the total possible values of the data structure is the product of the different values. Many languages have a tuple type which is the simplest formulation of a product type.<\/p>\n<p>See also <a href=\"https:\/\/en.wikipedia.org\/wiki\/Set_theory\" rel=\"nofollow\">Set theory<\/a>.<\/p>\n<h2><a id=\"user-content-option\" class=\"anchor\" aria-hidden=\"true\" href=\"#option\"><\/a>Option<\/h2>\n<p>Option is a <a href=\"#sum-type\">sum type<\/a> with two cases often called <code>Some<\/code> and <code>None<\/code>.<\/p>\n<p>Option is useful for composing functions that might not return a value.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Naive definition\nconst Some = (v) =&gt; ({\n  val: v,\n  map (f) {\n    return Some(f(this.val))\n  },\n  chain (f) {\n    return f(this.val)\n  }\n})\nconst None = () =&gt; ({\n  map (f) {\n    return this\n  },\n  chain (f) {\n    return this\n  }\n})\n\/\/ maybeProp :: (String, {a}) -&gt; Option a\nconst maybeProp = (key, obj) =&gt; typeof obj[key] === 'undefined' ? None() : Some(obj[key])<\/pre><\/div>\n<p>Use <code>chain<\/code> to sequence functions that return <code>Option<\/code>s<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ getItem :: Cart -&gt; Option CartItem\nconst getItem = (cart) =&gt; maybeProp('item', cart)\n\/\/ getPrice :: Item -&gt; Option Number\nconst getPrice = (item) =&gt; maybeProp('price', item)\n\/\/ getNestedPrice :: cart -&gt; Option a\nconst getNestedPrice = (cart) =&gt; getItem(cart).chain(getPrice)\ngetNestedPrice({}) \/\/ None()\ngetNestedPrice({item: {foo: 1}}) \/\/ None()\ngetNestedPrice({item: {price: 9.99}}) \/\/ Some(9.99)<\/pre><\/div>\n<p><code>Option<\/code> is also known as <code>Maybe<\/code>. <code>Some<\/code> is sometimes called <code>Just<\/code>. <code>None<\/code> is sometimes called <code>Nothing<\/code>.<\/p>\n<h2><a id=\"user-content-totality\" class=\"anchor\" aria-hidden=\"true\" href=\"#totality\"><\/a>Totality<\/h2>\n<h3><a id=\"user-content-total-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#total-functions\"><\/a>Total functions<\/h3>\n<p>The total function is just like a function in math - it will always return a result of the expected type for expected inputs and will always terminate. The easiest example is <code>identity<\/code> function:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ identity :: a -&gt; a\nconst identity = a =&gt; a<\/pre><\/div>\n<p>or <code>negate<\/code>:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ negate :: Boolean -&gt; Boolean\nconst negate = value =&gt; !value<\/pre><\/div>\n<p>Such functions always meet the requirements, you just can't provide such arguments, which satisfy inputs but break outputs.<\/p>\n<h3><a id=\"user-content-partial-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#partial-functions\"><\/a>Partial functions<\/h3>\n<p>It's a function which violates the requirements to be total - it may return an unexpected result with some inputs or it may never terminate. Partial functions add cognitive overhead, they are harder to reason about and they can lead to runtime errors. Some examples:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ example 1: sum of the list\n\/\/ sum :: [Number] -&gt; Number\nconst sum = arr =&gt; arr.reduce((a, b) =&gt; a + b)\nsum([1, 2, 3]) \/\/ 6\nsqrt([]) \/\/ TypeError: Reduce of empty array with no initial value\n\/\/ example 2: get the first item in list\n\/\/ first :: [A] -&gt; A\nconst first = a =&gt; a[0]\nfirst([42]) \/\/ 42\nfirst([]) \/\/ undefined\n\/\/or even worse:\nfirst([[42]])[0] \/\/ 42\nfirst([])[0] \/\/ Uncaught TypeError: Cannot read property '0' of undefined\n\/\/ example 3: repeat function N times\n\/\/ times :: Number -&gt; (Number -&gt; Number) -&gt; Number\nconst times = n =&gt; fn =&gt; n &amp;&amp; (fn(n), times(n - 1)(fn))\ntimes(3)(console.log)\n\/\/ 3\n\/\/ 2\n\/\/ 1\ntimes(-1)(console.log)\n\/\/ RangeError: Maximum call stack size exceeded<\/pre><\/div>\n<h3><a id=\"user-content-avoiding-partial-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#avoiding-partial-functions\"><\/a>Avoiding partial functions<\/h3>\n<p>Partial functions are dangerous, you can sometimes get the expected result, sometimes the wrong result, and sometimes your function can't stop the calculations at all. The input of partial functions should be always checked, and it can be hard to track all edge cases through entire applications, the easiest way to deal with it it's just to convert all partial functions to the total. General advice can be the usage of <code>Option<\/code> type, providing default values for edge cases and checking function conditions to make them always terminate:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ example 1: sum of the list\n\/\/ we can provide default value so it will always return result\n\/\/ sum :: [Number] -&gt; Number\nconst sum = arr =&gt; arr.reduce((a, b) =&gt; a + b, 0)\nsum([1, 2, 3]) \/\/ 6\nsqrt([]) \/\/ 0\n\/\/ example 2: get the first item in list\n\/\/ change result to Option\n\/\/ first :: [A] -&gt; Option A\nconst first = a =&gt; a.length ? Some(a[0]) : None()\nfirst([[42]]).map(a =&gt; console.log(a)) \/\/ 42\nfirst([]).map(a =&gt; console.log(a)) \/\/ console.log won't execute at all\n\/\/our previous worst case\nfirst([[42]]).map(a =&gt; console.log(a[0])) \/\/ 42\nfirst([]).map(a =&gt; console.log(a[0])) \/\/ won't execte, so we won't have error here\n\/\/ more of that, you will know by function return type (Option)\n\/\/ that you should use `.map` method to access the data and you will never forget\n\/\/ to check your input because such check become built-in into the function\n\/\/ example 3: repeat function N times\n\/\/ we should make function always terminate by changing conditions:\n\/\/ times :: Number -&gt; (Number -&gt; Number) -&gt; Number\nconst times = n =&gt; fn =&gt; n &gt; 0 &amp;&amp; (fn(n), times(n - 1)(fn))\ntimes(3)(console.log)\n\/\/ 3\n\/\/ 2\n\/\/ 1\ntimes(-1)(console.log)\n\/\/ won't execute anything<\/pre><\/div>\n<p>If you will change all your functions from partial to total, it can prevent you from having runtime exceptions, will make code easier to reason about and easier to maintain.<\/p>\n<h2><a id=\"user-content-functional-programming-libraries-in-javascript\" class=\"anchor\" aria-hidden=\"true\" href=\"#functional-programming-libraries-in-javascript\"><\/a>Functional Programming Libraries in JavaScript<\/h2>\n<hr \/><p><strong>P.S:<\/strong> This repo is successful due to the wonderful <a href=\"https:\/\/github.com\/hemanth\/functional-programming-jargon\/graphs\/contributors\">contributions<\/a>!<\/p>","created_at":"2018-11-07T16:54:46+0000","updated_at":"2018-11-07T16:54:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":20,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/18315?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":662,"title":"Five minute guide to better typography","url":"http:\/\/pierrickcalvez.com\/journal\/a-five-minutes-guide-to-better-typography","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:54:45+0000","updated_at":"2018-11-07T16:54:45+0000","annotations":[],"reading_time":0,"domain_name":"pierrickcalvez.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":663,"title":"Ask HN: Your experience with realty investment startups like RealtyShares?","url":"https:\/\/news.ycombinator.com\/item?id=15652430","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15652430\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15652430\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=15652430\" class=\"storylink\">Ask HN: Your experience with realty investment startups like RealtyShares?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        62 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=gk1\" class=\"hnuser\">gk1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15652430\">12 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15652430&amp;goto=item%3Fid%3D15652430\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Your%20experience%20with%20realty%20investment%20startups%20like%20RealtyShares%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Your%20experience%20with%20realty%20investment%20startups%20like%20RealtyShares%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15652430&amp;auth=837e5c85e07b97baf2b1584ae41b14e005c4bd14\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15652430\">60\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Startups like RealtyShares and Fundrise have been popping up, all promoting great returns and diversification... If you considered or invested with these platforms, how was your experience?<\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:54:45+0000","updated_at":"2018-11-07T16:54:45+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":659,"title":"Six Steps to Put Christensen's Jobs-to-be-Done Theory into Practice","url":"https:\/\/www.forbes.com\/sites\/stephenwunker\/2012\/02\/07\/six-steps-to-put-christensens-jobs-to-be-done-theory-into-practice\/","content":"<p class=\"speakable-paragraph\">Nine of the ten most valuable companies in America can trace their greatness to reconceiving a market\u2019s boundaries.\u00a0 From \n <a href=\"https:\/\/www.forbes.com\/companies\/exxonmobil\/\" target=\"_self\" data-ga-track=\"InternalLink:\/companies\/exxonmobil\/\" rel=\"nofollow noopener noreferrer\">ExxonMobil<\/a> \n  to \n <a href=\"https:\/\/www.forbes.com\/companies\/apple\/\" target=\"_self\" data-ga-track=\"InternalLink:\/companies\/apple\/\" rel=\"nofollow noopener noreferrer\">Apple<\/a> \n  to Wal-Mart, these firms expanded markets that others saw as static.\u00a0 How can companies in today\u2019s economy \u2013 seemingly hemmed in on all sides by hyper-competitive markets \u2013 break into new ground?<\/p><p>The wrong place to start is by asking customers what they want.\u00a0 Overwhelmingly, customers will answer based on how a market exists today.\u00a0 As Henry Ford reputedly said of his industry, \u201cIf I asked customers what they wanted, they would have said a faster horse.\u201d\u00a0 Instead, you need to look deeper and examine underlying needs.<\/p><p>Clayton Christensen, the famed Harvard <a href=\"http:\/\/www.forbes.com\/business\/\" target=\"_self\" data-ga-track=\"InternalLink:http:\/\/www.forbes.com\/business\/\">Business<\/a> School Professor known for coining the term \u201cdisruptive innovation,\u201d believes that one of his most enduring legacies will be an idea he first put forward in his 2003 book <em><a title=\"The Innovator's Solution\" href=\"http:\/\/www.amazon.com\/Innovators-Solution-Creating-Sustaining-Successful\/dp\/1578518520\" target=\"_blank\" rel=\"nofollow noopener noreferrer\" data-ga-track=\"ExternalLink:http:\/\/www.amazon.com\/Innovators-Solution-Creating-Sustaining-Successful\/dp\/1578518520\">The Innovator\u2019s Solution<\/a>: <\/em>don\u2019t sell products and services to customers, but rather try to help people <a title=\"HBR Jobs-to-be-Done\" href=\"http:\/\/hbswk.hbs.edu\/item\/5170.html\" target=\"_blank\" rel=\"nofollow noopener noreferrer\" data-ga-track=\"ExternalLink:http:\/\/hbswk.hbs.edu\/item\/5170.html\">address their jobs-to-be-done<\/a>.\u00a0 This seemingly simple idea has profound implications for re-framing industries.\u00a0 As I saw in years of consulting with Christensen to companies giant and small, it can revolutionize how firms compete.\u00a0 But the concept can also be tough to put into practice.\u00a0 A six-step process provides a rigorous way of defining the jobs you can address.\u00a0 Once those challenges are tightly defined, it is much easier to generate bold ideas for new solutions.<\/p><p><strong>1. \u00a0What are the high-level jobs-to-be-done?<\/strong><\/p><p>Rather than looking just at what people buy, examine the needs that arise during their lives. \u00a0Sometimes the job is much broader than the product or service that is bought.\u00a0 For instance, why did I take five small children to a movie on Sunday afternoon?\u00a0 Because on a rainy day I needed to get them out of house for a few hours.\u00a0 Could movie theaters expand their addressable market by emphasizing how they can occupy kids?\u00a0 What if the room used for the 20<sup>th<\/sup> screen was adapted instead for inexpensive play like a children's gym?<\/p><p><strong>2. \u00a0What are the current approaches and what pain points result?<\/strong><\/p><p>Jobs-to-be-done can sprawl across dozens of industry categories.\u00a0 Clearly a company can\u2019t address each job, but by looking broadly it can re-define its true \u201ccompetition.\u201d\u00a0 After it understands the full landscape, it can focus narrowly.\u00a0 Theaters may not want to invest in indoor playgrounds, but they need to see playgrounds as a rival every bit as real as a multiplex a few miles away.\u00a0 By understanding the pain points associated with competitive offerings, a business can better invest in emphasizing its distinctive strengths.<\/p><p><strong>3. \u00a0What benchmarks exist in the full range of competing offerings and analogies?<\/strong><\/p><p>Companies should always compare themselves to directly comparable firms, but they should not be seduced by the simplicity of that exercise.\u00a0 Through examining all that the full set of rivals and analogous offerings can do, they can get excellent ideas for their own business.\u00a0 For instance, a movie theater could learn from Disney World about how to market merchandise to children and how to entertain people in lines.<\/p><p><strong>4. \u00a0What performance criteria do customers use?<\/strong><\/p><p>Much psychological research has shown that even horribly complicated decisions are often reduced to a small handful of criteria that people can keep in mind at any one time.\u00a0 What are they for your industry?\u00a0 What adjectives describe a good solution?\u00a0 Asking customers these questions can open up surprising routes for improving current solutions or marketing existing offerings more effectively.<\/p><p><strong>5. \u00a0What prevents new solutions from being adopted?<\/strong><\/p><p>Managers are often too enamored of their own ideas.\u00a0 Unfortunately, even compelling ideas can take a long time to catch on.\u00a0 Indoor plumbing took 4,500 years from its invention to become widely adopted.\u00a0 Really, is your idea <em>better than indoor plumbing? \u00a0<\/em>Think in a disciplined fashion about all the obstacles hindering adoption of new solutions in your industry.\u00a0 Talk to customers about how they made a decision to adopt a recent innovation \u2013 not innovations in general, as that can average out important details, but rather a specific case study.<\/p><p><strong>6. \u00a0What value would success create for customers?<\/strong><\/p><p>By understanding the value that lies in resolving a pain point, you can see how many degrees of freedom you have to engineer a new solution.\u00a0 For instance, if resolving an issue on construction sites could avoid 30 minutes of downtime twice a week, and that time is valued at $600 \/ hour for the crew, then you get a sense of potential price and cost of a new solution. \u00a0Remember that value can be defined by money, time, convenience, peace of mind, and other metrics.<\/p><p>Re-framing a market can be a potent engine of business growth.\u00a0 These six steps help to translate Christensen\u2019s theory of jobs-to-be-done \u2013 an immensely powerful concept \u2013 into specific ideas for action.<\/p><p><em><a title=\"Jobs to be Done book chapter\" href=\"http:\/\/www.jobsroadmap.com\" target=\"_blank\" rel=\"nofollow noopener noreferrer\" data-ga-track=\"ExternalLink:http:\/\/www.jobsroadmap.com\">Click here to request a book chapter describing in further detail how to deploy the Jobs to be Done approach.<\/a><\/em><\/p>","created_at":"2018-11-07T16:54:44+0000","updated_at":"2018-11-07T16:54:44+0000","published_by":["Stephen Wunker"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.forbes.com","preview_picture":"https:\/\/thumbor.forbes.com\/thumbor\/600x315\/https%3A%2F%2Fi.forbesimg.com%2Fmedia%2Fassets%2Fforbes_1200x1200.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":660,"title":"Clojure Guides | Getting Started With Clojure | Clojure Tutorials | Clojure Documentation","url":"http:\/\/clojure-doc.org\/","content":"<div class=\"row\"><div class=\"col-md-4\"><p>Tutorials aimed at new users.<\/p><\/div><div class=\"col-md-4\"><p>Comprehensive guides on every aspect of the core language.<\/p><\/div><div class=\"col-md-4\"><p>\n      This material is not covered by the Clojure Contributor Agreement and is developed using <a href=\"https:\/\/github.com\/clojuredocs\/cds#how-to-contribute\">pull-requests on GitHub<\/a>.\n    <\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-4\"><p>Guides covering areas outside of the core language.<\/p><\/div><div class=\"col-md-4\"><p>Subject-specific tutorials and guides.<\/p><\/div><\/div>","created_at":"2018-11-07T16:54:44+0000","updated_at":"2018-11-07T16:54:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"clojure-doc.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":661,"title":"Disable anti-forgery check \u00b7 Clojure - building a simple web app","url":"https:\/\/practicalli.github.io\/clojure-webapps\/project-url-shortner\/disable-anti-forgery-check.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:54:44+0000","updated_at":"2018-11-07T16:54:44+0000","annotations":[],"reading_time":0,"domain_name":"practicalli.github.io","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":658,"title":"SVG Animations","url":"http:\/\/shop.oreilly.com\/product\/0636920045335.do","content":"<section id=\"product-image-detail-links\"><a href=\"https:\/\/covers.oreillystatic.com\/images\/0636920045335\/lrg.jpg\" target=\"_blank\"><img id=\"product-image\" src=\"https:\/\/covers.oreillystatic.com\/images\/0636920045335\/cat.gif\" alt=\"image\" \/><\/a><section id=\"product-detail\"><h3>From Common UX Implementations to Complex Responsive Animation<\/h3><section id=\"publisher-release-length\"><p><strong>Publisher:<\/strong> <a href=\"https:\/\/www.safaribooksonline.com\/library\/publisher\/oreilly-media-inc\/?utm_medium=referral&amp;utm_campaign=publisher&amp;utm_source=oreilly&amp;utm_content=catalog&amp;utm_content=catalog\">O'Reilly Media<\/a><\/p><p><strong>Release Date:<\/strong> March 2017<\/p><p><strong>Pages:<\/strong> 246<\/p><\/section><section id=\"product-links\"><a id=\"safari-link\" href=\"https:\/\/www.safaribooksonline.com\/library\/view\/svg-animations\/9781491939697\/?utm_medium=referral&amp;utm_campaign=publisher&amp;utm_source=oreilly&amp;utm_content=buybox\">\n\t\t\t\t\t<h4>\n\t\t\t\t\t\t\n\t\t\t\t\t\t\tRead\n\t\t\t\t\t\t\n\t\t\t\t\t\ton Safari with\u00a0a\u00a010-day\u00a0trial\n\t\t\t\t\t<\/h4>\n\t\t\t\t\tStart your free trial now\n\t\t\t\t<\/a><a id=\"buy-on-amazon\" class=\"btn\" href=\"https:\/\/www.amazon.com\/_\/dp\/1491939702?tag=oreilly20-20\">Buy on Amazon<\/a><\/section><section id=\"product-description\"><p>SVG is extremely powerful, with its reduced HTTP requests and crispness on any display. It becomes increasingly more interesting as you explore its capabilities for responsive animation and performance boons. When you animate SVG, you must be aware of normal image traits like composition, color, implementation, and optimization. But when you animate, it increases the complexity of each of these factors exponentially.\n<\/p><p>\nThis practical book takes a deep dive into how you can to solve these problems with stability, performance, and creativity in mind.\n<\/p><ul><li>Learn how to make SVG cross-browser compatible, backwards compatible, optimized, and responsive\n<\/li><li>Plan and debug animation\n<\/li><li>Make a complex animation responsive, as many sites are responsive\n<\/li><li>Profile each animation technique in terms of performance so that you know what you're getting in to with each library or native technology\n<\/li><\/ul><\/section><\/section><\/section><section id=\"toc\"><h3>Table of Contents<\/h3><div id=\"toc-content\"><ol id=\"chapters\"><li class=\"chapter\" title=\"1\">\n   <h3>\n    \n     Chapter 1\n    \n    The Anatomy of an SVG\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"1.1\">\n     <h4>\n      SVG DOM Syntax\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"1.2\">\n     <h4>\n      viewBox and preserveAspectRatio\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"1.3\">\n     <h4>\n      Drawing Shapes\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"1.4\">\n     <h4>\n      Responsive SVG, Grouping, and Drawing Paths\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"1.5\">\n     <h4>\n      SVG on Export, Recommendations, and Optimization\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"1.6\">\n     <h4>\n      Reduce Path Points\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"1.7\">\n     <h4>\n      Optimization Tools\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"2\">\n   <h3>\n    \n     Chapter 2\n    \n    Animating with CSS\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"2.1\">\n     <h4>\n      Animating with SVG\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"2.2\">\n     <h4>\n      Benefits of Drawing with SVG\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"2.3\">\n     <h4>\n      Silky-Smooth Animation\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"3\">\n   <h3>\n    \n     Chapter 3\n    \n    CSS Animation and Hand-Drawn SVG Sprites\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"3.1\">\n     <h4>\n      Keyframe Animation with steps() and SVG Sprites, Two Ways\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"4\">\n   <h3>\n    \n     Chapter 4\n    \n    Creating a Responsive SVG Sprite\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"4.1\">\n     <h4>\n      SVG Sprites and CSS for Responsive Development\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"4.2\">\n     <h4>\n      Grouping and DRYing It Out\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"4.3\">\n     <h4>\n      The viewBox Trick\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"4.4\">\n     <h4>\n      Responsive Animation\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"5\">\n   <h3>\n    \n     Chapter 5\n    \n    UI\/UX Animations with No External Libraries\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"5.1\">\n     <h4>\n      Context-Shifting in UX Patterns\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"5.2\">\n     <h4>\n      Putting It All Together\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"6\">\n   <h3>\n    \n     Chapter 6\n    \n    Animating Data Visualizations\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"6.1\">\n     <h4>\n      Why Use Animation in Data Visualization?\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"6.2\">\n     <h4>\n      D3 with CSS Animation Example\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"6.3\">\n     <h4>\n      Chartist with CSS Animation Example\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"6.4\">\n     <h4>\n      Animating with D3\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"7\">\n   <h3>\n    \n     Chapter 7\n    \n    A Comparison of Web Animation Technologies\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"7.1\">\n     <h4>\n      Native Animation\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"7.2\">\n     <h4>\n      External Libraries\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"7.3\">\n     <h4>\n      Not Suggested\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"7.4\">\n     <h4>\n      React-Specific Workflows\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"7.5\">\n     <h4>\n      Covering Ground\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"8\">\n   <h3>\n    \n     Chapter 8\n    \n    Animating with GreenSock\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"8.1\">\n     <h4>\n      Basic GreenSock Syntax\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"9\">\n   <h3>\n    \n     Chapter 9\n    \n    GreenSock\u2019s Timeline\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"9.1\">\n     <h4>\n      A Simple Timeline\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"9.2\">\n     <h4>\n      Relative Labels\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"9.3\">\n     <h4>\n      Nested and Master Timelines\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"10\">\n   <h3>\n    \n     Chapter 10\n    \n    MorphSVG and Motion Along a Path\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"10.1\">\n     <h4>\n      MorphSVG\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"10.2\">\n     <h4>\n      Motion Along a Path\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"11\">\n   <h3>\n    \n     Chapter 11\n    \n    Stagger Effects, Tweening HSL, and SplitText for Text Animation\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"11.1\">\n     <h4>\n      Staggered Animations\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"11.2\">\n     <h4>\n      Relative HSL Color Animation\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"12\">\n   <h3>\n    \n     Chapter 12\n    \n    DrawSVG and Draggable\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"12.1\">\n     <h4>\n      Draggable\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"12.2\">\n     <h4>\n      DrawSVG\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"13\">\n   <h3>\n    \n     Chapter 13\n    \n    Mo.js\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"13.1\">\n     <h4>\n      Base Premises\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"13.2\">\n     <h4>\n      Shapes\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"13.3\">\n     <h4>\n      Shape Motion\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"14\">\n   <h3>\n    \n     Chapter 14\n    \n    React-Motion\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"14.1\">\n     <h4>\n      &lt;Motion \/&gt;\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"14.2\">\n     <h4>\n      &lt;StaggeredMotion \/&gt;\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"15\">\n   <h3>\n    \n     Chapter 15\n    \n    Animating the Unanimatable: Motion with Attributes and Bare-Metal Implementations\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"15.1\">\n     <h4>\n      requestAnimationFrame()\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"15.2\">\n     <h4>\n      GreenSock\u2019s AttrPlugin\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"15.3\">\n     <h4>\n      Practical Application: Animating the viewBox\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"16\">\n   <h3>\n    \n     Chapter 16\n    \n    Responsive Animation\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"16.1\">\n     <h4>\n      Some Quick Responsive Tips\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"16.2\">\n     <h4>\n      GreenSock and Responsive SVG\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"16.3\">\n     <h4>\n      Responsive SVG, with or without GreenSock\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"16.4\">\n     <h4>\n      Responsive Reorganization by Updating the ViewBox\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"16.5\">\n     <h4>\n      Responsive Reorganization with Multiple SVGs and Media Queries\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n  <li class=\"chapter\" title=\"17\">\n   <h3>\n    \n     Chapter 17\n    \n    Designing, Prototyping, and Animation in Component Libraries\n   <\/h3>\n   <ol><li class=\"sect1\" title=\"17.1\">\n     <h4>\n      Designing an Animation\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"17.2\">\n     <h4>\n      Prototyping\n     <\/h4>\n    <\/li>\n    <li class=\"sect1\" title=\"17.3\">\n     <h4>\n      Animations in Component Libraries\n     <\/h4>\n    <\/li>\n   <\/ol><\/li>\n <\/ol><\/div><\/section>","created_at":"2018-11-07T16:54:43+0000","updated_at":"2018-11-07T16:54:43+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"shop.oreilly.com","preview_picture":"http:\/\/covers.oreilly.com\/images\/0636920045335\/bkt.gif","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":657,"title":"Skyrim rendered in text | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=15661490","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15661490\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15661490\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/medium.com\/@filiph\/skyrim-rendered-in-text-1899548ab2c4\" class=\"storylink\">Skyrim rendered in text<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=medium.com\">medium.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        459 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=ellinokon\" class=\"hnuser\">ellinokon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15661490\">12 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15661490&amp;goto=item%3Fid%3D15661490\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Skyrim%20rendered%20in%20text&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Skyrim%20rendered%20in%20text\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15661490&amp;auth=10cb629cf1f4da5952fc60f865986cf9224cc38c\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15661490\">106\u00a0comments<\/a>              <\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"15665214\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=superflyguy\" class=\"hnuser\">superflyguy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665214\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15669276\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=milankovic\" class=\"hnuser\">milankovic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15669276\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15669656\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=O4epegb\" class=\"hnuser\">O4epegb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15669656\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15680494\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=milankovic\" class=\"hnuser\">milankovic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15680494\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666022\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mizza\" class=\"hnuser\">Mizza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666022\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667908\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kbenson\" class=\"hnuser\">kbenson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667908\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668488\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hdhzy\" class=\"hnuser\">hdhzy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668488\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667174\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smnscu\" class=\"hnuser\">smnscu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667174\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668241\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668241\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15669811\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kakarot\" class=\"hnuser\">kakarot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15669811\">12 months ago<\/a><\/p><br \/><div class=\"comment\">I just want to say that I was thoroughly impressed by the vision you sought to achieve, and equally impressed with the presentation. Your 5 years of research were not wasted!<p>The way the game flowed downwards complements the mechanics quite well, allowing you to backtrack if you forgot an important detail, and providing a moment of suspense as you wait to see what is materializing.<\/p><p>This begs for the addition of more components than just text, choices, and the dice. \"Minigames\", as I believe you phrased it. Think something somewhere between Undertale and what you have now.<\/p><p>The writing was strong, varied, fun and descriptive without being too prosy. My only complaint in that respect was that the early dialogue for your character is a bit terse and constrained, but it seemed to get better late game.<\/p><p>Combat was good, with interesting choices. There was a moment of intense satisfaction from the realization that I indeed felt that sense of agency you were aiming for.<\/p><p>The UX itself was inspired. Very clean look. Good bit of polishing in the cracks and crevices, and it shows. From the help icons next to the choices, to the detailed and generous colored artwork, to the health display, to the way you slowly introduce features in a way that seems natural to the story and not forced.<\/p><p>And the dice. Oh, the dice. Probably the coolest thing I've seen in a while. You actually found a way to make a rudimentary random number generator look visually appealing, and created a sense of tension that ebbs and flows. This is extremely hard to pull off in the digital format, because with physical dice you have this connection to the universe.<\/p><p>Whether you roll high or low, you feel a sense of fairness as the universe does not play favorites. You can see and feel all of the factors that determine the outcome, like wind, air pressure, the texture and weight of the die and surface. You also feel like you can influence the outcome to a degree by adjusting things like how you roll, how high you drop, blowing on \/ kissing your hand, etc. With digital dice, it usually feels hollow and fake, as you know the only thing determining whether you win is a single line of code. You don't feel any agency.<\/p><p>The idea of having more or less hearts depending on the likelihood of success sets the initial level of tension. This wanes to make room for the excitement of chance, of expectation of reward.<\/p><p>As the hearts begin to slow and lock into place one by one, and never the same boring way, the tension rapidly increases to its peak before even more rapidly falling flat and dissolving into either excitement or disappointment. This entire experience is modulated by the lightly colored text below the prompt that hints at how likely you are to succeed. <i>Highly probable<\/i>? The disappointment stings that much worse. <i>Not likely<\/i>? The satisfaction from nailing that strike is peerless.<\/p><p>This is probably my favorite mechanic of the entire game as it provides the backbone for an experience designed around manipulating your thoughts and emotions. Like slot machines at a casino, the moment of anticipation is greater than the reward, and affects your behavior and aversion to risk, leading you to sometimes feel that it is best to stick to the shadows and stay quiet, and other times playing you for the fool by causing you to get too bold in your actions before putting you back in line with a terrible roll. It would serve the experience well to focus on maximizing the dynamics it introduces to the world.<\/p><p>I could keep going, but I'll stop. Great game, I subscribed to the mailing list and eagerly await the sequel. If the scope of the sequel is large enough, it will set the bar quite high and pave the way for others to innovate.<\/p><p>If you could use a few hours of help a week with programming, design, or development of new mechanics, I would love to help you pioneer this interesting new twist on an old genre.\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666399\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666399\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663870\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Aardwolf\" class=\"hnuser\">Aardwolf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663870\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666000\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osaariki\" class=\"hnuser\">osaariki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666000\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667005\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=katastic\" class=\"hnuser\">katastic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667005\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664309\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stickydink\" class=\"hnuser\">stickydink<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664309\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664722\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Something1234\" class=\"hnuser\">Something1234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664722\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667833\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=khedoros1\" class=\"hnuser\">khedoros1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667833\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666568\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666568\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666936\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shakna\" class=\"hnuser\">shakna<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666936\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666992\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666992\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667043\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shakna\" class=\"hnuser\">shakna<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667043\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667081\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667081\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666954\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Something1234\" class=\"hnuser\">Something1234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666954\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667865\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=khedoros1\" class=\"hnuser\">khedoros1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667865\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667014\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667014\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662380\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=le-mark\" class=\"hnuser\">le-mark<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662380\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664938\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664938\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668554\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=subwayclub\" class=\"hnuser\">subwayclub<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668554\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667611\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=phlakaton\" class=\"hnuser\">phlakaton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667611\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666103\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kevinmchugh\" class=\"hnuser\">kevinmchugh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666103\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663536\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pwaivers\" class=\"hnuser\">pwaivers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663536\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664903\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tarboreus\" class=\"hnuser\">tarboreus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664903\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664979\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664979\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665948\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=davidkuhta\" class=\"hnuser\">davidkuhta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665948\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666767\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=euyyn\" class=\"hnuser\">euyyn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666767\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15731116\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tarboreus\" class=\"hnuser\">tarboreus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15731116\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665690\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brlewis\" class=\"hnuser\">brlewis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665690\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15731910\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tarboreus\" class=\"hnuser\">tarboreus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15731910\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664236\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=c22\" class=\"hnuser\">c22<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664236\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665043\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665043\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664218\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DigitalJack\" class=\"hnuser\">DigitalJack<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664218\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664545\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway7645\" class=\"hnuser\">throwaway7645<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664545\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668903\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdrothrock\" class=\"hnuser\">sdrothrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668903\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664779\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=8_hours_ago\" class=\"hnuser\">8_hours_ago<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664779\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665988\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=optimuspaul\" class=\"hnuser\">optimuspaul<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665988\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667064\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gnulinux\" class=\"hnuser\">gnulinux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667064\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662216\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bitL\" class=\"hnuser\">bitL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662216\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662696\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=duckarmada\" class=\"hnuser\">duckarmada<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662696\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663291\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelbuckbee\" class=\"hnuser\">michaelbuckbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663291\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664462\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ratacat\" class=\"hnuser\">ratacat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664462\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664597\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=astrobe_\" class=\"hnuser\">astrobe_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664597\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664870\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664870\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666855\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ioddly\" class=\"hnuser\">ioddly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666855\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665792\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=qplex\" class=\"hnuser\">qplex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665792\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664019\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e12e\" class=\"hnuser\">e12e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664019\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668136\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andreareina\" class=\"hnuser\">andreareina<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668136\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15669404\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=indigochill\" class=\"hnuser\">indigochill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15669404\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664830\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CountHackulus\" class=\"hnuser\">CountHackulus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664830\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665008\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=oxide\" class=\"hnuser\">oxide<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665008\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663175\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=guard0g\" class=\"hnuser\">guard0g<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663175\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664001\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wavefunction\" class=\"hnuser\">wavefunction<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664001\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663229\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=52-6F-62\" class=\"hnuser\">52-6F-62<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663229\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668772\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Jaruzel\" class=\"hnuser\">Jaruzel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668772\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664221\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bduerst\" class=\"hnuser\">bduerst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664221\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662326\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Rooster61\" class=\"hnuser\">Rooster61<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662326\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664096\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=denisw\" class=\"hnuser\">denisw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664096\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664591\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AdmiralAsshat\" class=\"hnuser\">AdmiralAsshat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664591\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664865\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=komali2\" class=\"hnuser\">komali2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664865\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668276\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kilburn\" class=\"hnuser\">kilburn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668276\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15669550\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=indigochill\" class=\"hnuser\">indigochill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15669550\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662392\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawlessone\" class=\"hnuser\">lawlessone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662392\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664784\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LoSboccacc\" class=\"hnuser\">LoSboccacc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664784\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666489\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amaranth\" class=\"hnuser\">amaranth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666489\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664353\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=PaulHoule\" class=\"hnuser\">PaulHoule<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664353\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664168\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=NetStrikeForce\" class=\"hnuser\">NetStrikeForce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664168\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664194\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=VectorLock\" class=\"hnuser\">VectorLock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664194\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662229\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hmhrex\" class=\"hnuser\">hmhrex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662229\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665434\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brlewis\" class=\"hnuser\">brlewis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665434\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15675551\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=scott_karana\" class=\"hnuser\">scott_karana<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15675551\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668224\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=grrowl\" class=\"hnuser\">grrowl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668224\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664708\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nategri\" class=\"hnuser\">nategri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664708\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666110\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=btym\" class=\"hnuser\">btym<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666110\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666417\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kermittd\" class=\"hnuser\">kermittd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666417\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664567\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Romanulus\" class=\"hnuser\">Romanulus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664567\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662104\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=atriches\" class=\"hnuser\">atriches<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662104\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662388\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=haspok\" class=\"hnuser\">haspok<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662388\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662444\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pjc50\" class=\"hnuser\">pjc50<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662444\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664073\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dkersten\" class=\"hnuser\">dkersten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664073\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664207\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wlesieutre\" class=\"hnuser\">wlesieutre<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664207\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665239\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wlesieutre\" class=\"hnuser\">wlesieutre<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665239\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665249\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665249\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665617\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrec\" class=\"hnuser\">mrec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665617\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15666182\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15666182\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667271\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrec\" class=\"hnuser\">mrec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667271\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668500\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hdhzy\" class=\"hnuser\">hdhzy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668500\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668861\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrec\" class=\"hnuser\">mrec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668861\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15669034\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hdhzy\" class=\"hnuser\">hdhzy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15669034\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667390\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Zarathruster\" class=\"hnuser\">Zarathruster<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667390\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15662873\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jamesrcole\" class=\"hnuser\">jamesrcole<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15662873\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663218\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=PostOnce\" class=\"hnuser\">PostOnce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663218\">12 months ago<\/a><\/p><br \/><div class=\"comment\">I still play all the time and there's some truth to them all being the same. They get more and more polished on a UI level, and more realistic looking, but remain static mechanically.<p>For example, Skyrim doesn't do much mechanically that Ultima Underworld didn't. Ultima proper did all of that from a different perspective for a long while before that, too. The Witcher 3 is my absolute favorite RPG at least since Skyrim and probably supercedes it, but again, it's just super polished, it doesn't do much that's really new, that I haven't seen before in an RPG or other game. You pick speech options from dialogue trees, loot dead things, and fetch quest items, but you're in a cardboard world. The monsters just wait for you to come slaughter them, they're set-pieces whose purpose is to sit there until you arrive, then bite you. The world reacts largely in a single specific way: you win.<\/p><p>Actually, I thing modern RPGs have regressed, Ultima VI's keyword-based dialogue was better than our modern dialogue trees. You could speak in natural language about a variety of topics, some of which were hinted at by the NPC with highlighted words, and others you had to figure out for yourself. (you could just type \"troll\" or you could type \"What can you tell me about the troll?\" or \"I have seen a troll, and need advice, can you help?\", it was more fun and interactive)<\/p><p>I'm hyped for Underworld Ascendant, because just maybe they'll break some new ground for a change. \"A living ecosystem\", but with people who can (and have) backed up that claim. I don't see much else on the horizon -- the indies are our only hopes! AAA can't afford to lose now, so it's all cookie cutter sequels there, only the indies can afford the risk.<\/p><p>I think a huge, huge part of the problem is that games cost so much and carry so much risk now. Ultima Underworld got a full sequel _ELEVEN MONTHS LATER_. You could move faster and try more then, because art didn't cost so much, there was no motion capture, no voice acting, it was cheap to make a game. If a sequel flops now you can't just try again, you lost $75,000,000 dollars and are now bankrupt.<\/p><p>Anyway, here's a link to a <i>free<\/i>, breathtakingly good book on the history of CRPGs from the 70s through today: <a href=\"https:\/\/crpgbook.wordpress.com\/\" rel=\"nofollow\">https:\/\/crpgbook.wordpress.com\/<\/a>\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664897\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=freeflight\" class=\"hnuser\">freeflight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664897\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15667259\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdfin\" class=\"hnuser\">sdfin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15667259\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15668834\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=PostOnce\" class=\"hnuser\">PostOnce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15668834\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15663751\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lph\" class=\"hnuser\">lph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15663751\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15664293\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=megaman22\" class=\"hnuser\">megaman22<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15664293\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15665171\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=filiph\" class=\"hnuser\">filiph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15665171\">12 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:54:42+0000","updated_at":"2018-11-07T16:54:42+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":656,"title":"nikitavoloboev\/my-mac-os","url":"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<blockquote>\n<p>List of applications and tools that make my macOS experience even more amazing<\/p>\n<\/blockquote>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/nikitavoloboev\/my-mac-os\/master\/mac.png\"><img src=\"https:\/\/raw.githubusercontent.com\/nikitavoloboev\/my-mac-os\/master\/mac.png\" alt=\"\" \/><\/a><\/p>\n<blockquote>\n<p>All the apps are in one desktop because there is delay in switching between macOS multiple desktops<\/p>\n<\/blockquote>\n<h5><a id=\"user-content-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#contents\"><\/a>Contents<\/h5>\n<ul><li><a href=\"#applications\">Applications<\/a>\n<ul><li><a href=\"#productivity\">Productivity<\/a><\/li>\n<li><a href=\"#code\">Code<\/a><\/li>\n<li><a href=\"#social\">Social<\/a><\/li>\n<li><a href=\"#writing\">Writing<\/a><\/li>\n<li><a href=\"#design\">Design<\/a><\/li>\n<li><a href=\"#music\">Music<\/a><\/li>\n<li><a href=\"#images\">Images<\/a><\/li>\n<li><a href=\"#video\">Video<\/a><\/li>\n<li><a href=\"#utilities\">Utilities<\/a><\/li>\n<li><a href=\"#browsers\">Browsers<\/a>\n<ul><li><a href=\"#safari-extensions\">Safari Extensions<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#command-line-apps\">Command Line Apps<\/a><\/li>\n<li><a href=\"#preference-panes\">Preference Panes<\/a><\/li>\n<li><a href=\"#my-wonderful-world-of-ios-\">My wonderful world of iOS<\/a><\/li>\n<li><a href=\"#similar-setups\">Similar Setups<\/a><\/li>\n<li><a href=\"#related\">Related<\/a><\/li>\n<li><a href=\"#contributing\">Contributing<\/a><\/li>\n<\/ul><h2><a id=\"user-content-applications\" class=\"anchor\" aria-hidden=\"true\" href=\"#applications\"><\/a>Applications<\/h2>\n<p>I use a lot of applications on my mac. Below is a list of the ones that I love and use the most in my day to day life.<\/p>\n<p>I also share <a href=\"https:\/\/github.com\/nikitavoloboev\/dotfiles#readme\">my dotfiles<\/a> together with my <a href=\"https:\/\/github.com\/nikitavoloboev\/my-ios#readme\">iOS setup<\/a>. And I made a <a href=\"https:\/\/t.me\/macOSautomation\" rel=\"nofollow\">Telegram group<\/a> to discuss all things macOS\/iOS.<\/p>\n<h3><a id=\"user-content-productivity\" class=\"anchor\" aria-hidden=\"true\" href=\"#productivity\"><\/a>Productivity<\/h3>\n<h4><a id=\"user-content-alfred---launcher\" class=\"anchor\" aria-hidden=\"true\" href=\"#alfred---launcher\"><\/a><a href=\"https:\/\/www.alfredapp.com\" rel=\"nofollow\">Alfred<\/a> - Launcher<\/h4>\n<ul><li>Alfred is a powerful launcher that you can program to show anything you want. It saved me a lot of time in my life.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/fe64b2d40c502b9e397831e86da4b5ee14ef0eaa\/68747470733a2f2f692e696d6775722e636f6d2f757a71317735672e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/fe64b2d40c502b9e397831e86da4b5ee14ef0eaa\/68747470733a2f2f692e696d6775722e636f6d2f757a71317735672e706e67\" width=\"700\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/uzq1w5g.png\" \/><\/a><\/p>\n<ul><li>\n<p>It has a great <a href=\"http:\/\/www.alfredforum.com\/\" rel=\"nofollow\">community<\/a> and <a href=\"https:\/\/github.com\/learn-anything\/alfred-workflows#readme\">amazing workflows<\/a> that you can use.<\/p>\n<\/li>\n<li>\n<p>I wrote <a href=\"https:\/\/medium.com\/@nikitavoloboev\/writing-alfred-workflows-in-go-2a44f62dc432\" rel=\"nofollow\">an article<\/a> on how anyone can start developing workflows of their own using Go language and <a href=\"https:\/\/github.com\/deanishe\/awgo\">AwGo<\/a> library.<\/p>\n<\/li>\n<\/ul><h4><a id=\"user-content-karabiner---keyboard-remapping\" class=\"anchor\" aria-hidden=\"true\" href=\"#karabiner---keyboard-remapping\"><\/a><a href=\"https:\/\/pqrs.org\/osx\/karabiner\/\" rel=\"nofollow\">Karabiner<\/a> - Keyboard remapping<\/h4>\n<ul><li>Karabiner is an absolutely amazing app that lets you remap keys at a very low level on macOS.<\/li>\n<li>I have completely remapped my keyboard with it and every key on my keyboard is a custom modifier key that I can program to do what I want.<\/li>\n<li>For example you can make caps lock into an escape key when pressed once but if you hold it, it becomes a <a href=\"http:\/\/brettterpstra.com\/2017\/06\/15\/a-hyper-key-with-karabiner-elements-full-instructions\/\" rel=\"nofollow\">hyper key<\/a>. Hyper key means that a key now serves two purposes, once when pressed alone and once when held down. So for example, for remapping caps lock, we can remap it to act as escape when pressed alone once but if we hold down on it, it becomes <code>\u2318 + \u2303<\/code> modifier key. So <code>Caps Lock + F<\/code> becomes <code>\u2318 + \u2303 + F<\/code>. And so on.<\/li>\n<li>I take this idea further and define these kind of hyper keys for <strong>every single key on my keyboard<\/strong>.<\/li>\n<li>I describe how I use Karabiner in detail <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/macOS\/apps\/karabiner\/karabiner\" rel=\"nofollow\">here<\/a> and I generate <a href=\"https:\/\/github.com\/nikitavoloboev\/dotfiles\/blob\/master\/karabiner\/karabiner.edn\">my configuration<\/a> with a <a href=\"https:\/\/github.com\/yqrashawn\/GokuRakuJoudo#readme\">Karabiner DSL<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-keyboard-maestro---automation-tool\" class=\"anchor\" aria-hidden=\"true\" href=\"#keyboard-maestro---automation-tool\"><\/a><a href=\"https:\/\/www.keyboardmaestro.com\/main\/\" rel=\"nofollow\">Keyboard Maestro<\/a> - Automation tool<\/h4>\n<ul><li>Keyboard Maestro is essentially an IDE for automation. You create macros of actions that you can then easily call from Karabiner.<\/li>\n<li>It has a <a href=\"https:\/\/forum.keyboardmaestro.com\/\" rel=\"nofollow\">wonderful community<\/a> that is happy to help with whatever you are trying to achieve.<\/li>\n<li>I share <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/macOS\/apps\/keyboard-maestro\/km-macros\" rel=\"nofollow\">all the macros I use<\/a> with the app.<\/li>\n<\/ul><h4><a id=\"user-content-2do---flexible-task-manager\" class=\"anchor\" aria-hidden=\"true\" href=\"#2do---flexible-task-manager\"><\/a><a href=\"http:\/\/www.2doapp.com\/mac\" rel=\"nofollow\">2Do<\/a> - Flexible task manager<\/h4>\n<ul><li>I love GTD methodology. This app is phenomenal with helping me <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/macOS\/apps\/2do\" rel=\"nofollow\">organize my tasks<\/a>. Here is how my sidebar looks:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/9af81f62e463c6cd40e67ea7fa6b0e7e5fd54d4b\/68747470733a2f2f692e696d6775722e636f6d2f445559344f794f2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/9af81f62e463c6cd40e67ea7fa6b0e7e5fd54d4b\/68747470733a2f2f692e696d6775722e636f6d2f445559344f794f2e706e67\" width=\"150\" data-canonical-src=\"https:\/\/i.imgur.com\/DUY4OyO.png\" alt=\"image\" \/><\/a><\/p>\n<ul><li>It also has global quick add with a hotkey. Together with lists, priorities, powerful search and a lot more.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/bcc0e9b606259d778d634009134ebf405342b9e1\/68747470733a2f2f692e696d6775722e636f6d2f62396c474673302e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/bcc0e9b606259d778d634009134ebf405342b9e1\/68747470733a2f2f692e696d6775722e636f6d2f62396c474673302e706e67\" width=\"400\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/b9lGFs0.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-trello---project-management-tool\" class=\"anchor\" aria-hidden=\"true\" href=\"#trello---project-management-tool\"><\/a><a href=\"https:\/\/trello.com\" rel=\"nofollow\">Trello<\/a> - Project management tool<\/h4>\n<ul><li>\n<p>I use the app a lot to track various projects I have in my life.<\/p>\n<\/li>\n<li>\n<p>I share many boards I made <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/sharing\/my-trello\" rel=\"nofollow\">publicly<\/a>. Here is an example of a public board for tracking various <a href=\"https:\/\/trello.com\/b\/cu32qF3q\" rel=\"nofollow\">things I want to learn<\/a>.\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/d6e94a02e35bbe9b6f993b1427492b8cb0d704b6\/68747470733a2f2f692e696d6775722e636f6d2f377341565839452e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/d6e94a02e35bbe9b6f993b1427492b8cb0d704b6\/68747470733a2f2f692e696d6775722e636f6d2f377341565839452e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/7sAVX9E.png\" \/><\/a><\/p>\n<\/li>\n<li>\n<p>I assign myself to the cards I am working on now and filter to see only them by <a href=\"https:\/\/trello.com\/shortcuts\" rel=\"nofollow\">pressing Q key<\/a>. I mark cards I want to work on next with <code>Next<\/code> tag.<\/p>\n<\/li>\n<\/ul><h4><a id=\"user-content-1password---password-manager\" class=\"anchor\" aria-hidden=\"true\" href=\"#1password---password-manager\"><\/a><a href=\"https:\/\/1password.com\" rel=\"nofollow\">1Password<\/a> - Password manager<\/h4>\n<ul><li>Generate all of my passwords with it and keep everything in a secured and encrypted vault kept secure by my one master password.<\/li>\n<li>No longer need to remember passwords and I now have a unique password for every website that I am signed up on whilst <a href=\"https:\/\/support.1password.com\/one-time-passwords\/\" rel=\"nofollow\">activating two factor authentication<\/a> wherever possible.<\/li>\n<\/ul><h4><a id=\"user-content-timing---automatic-time-tracker\" class=\"anchor\" aria-hidden=\"true\" href=\"#timing---automatic-time-tracker\"><\/a><a href=\"https:\/\/timingapp.com\/?\" rel=\"nofollow\">Timing<\/a> - Automatic time tracker<\/h4>\n<ul><li>I use Timing app to fully automate tracking my time on my computer.<\/li>\n<li>The fact that I can combine active and passive tracking together is very powerful and allows me to be more aware of where I spend my time.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/95ebdeba507ed9f53d83e3c1874ea47e87b1c9a8\/68747470733a2f2f692e696d6775722e636f6d2f435335626f6a612e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/95ebdeba507ed9f53d83e3c1874ea47e87b1c9a8\/68747470733a2f2f692e696d6775722e636f6d2f435335626f6a612e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/CS5boja.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-mindnode---interactive-mind-mapping\" class=\"anchor\" aria-hidden=\"true\" href=\"#mindnode---interactive-mind-mapping\"><\/a><a href=\"https:\/\/mindnode.com\" rel=\"nofollow\">MindNode<\/a> - Interactive Mind Mapping<\/h4>\n<ul><li>An application which allowed me to originally create all of the maps for <a href=\"https:\/\/learn-anything.xyz\" rel=\"nofollow\">Learn Anything<\/a>.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/learn-anything\/learn-anything\/master\/media\/header.png\"><img src=\"https:\/\/raw.githubusercontent.com\/learn-anything\/learn-anything\/master\/media\/header.png\" alt=\"\" \/><\/a><\/p>\n<ul><li>I also use MindNode to visualize topics and their relations as an <a href=\"https:\/\/my.mindnode.com\/dPmKpWZgTpZsaG4Rf1yaz7rXbYhnuqPDDpdss4w2#1594.8,-1084.6,1\" rel=\"nofollow\">interactive mind map<\/a>.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/950e1403e84f851cd53bb63659d6adca84b14db4\/68747470733a2f2f692e696d6775722e636f6d2f757a4e4e59697a2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/950e1403e84f851cd53bb63659d6adca84b14db4\/68747470733a2f2f692e696d6775722e636f6d2f757a4e4e59697a2e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/uzNNYiz.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-typinator---text-expansions\" class=\"anchor\" aria-hidden=\"true\" href=\"#typinator---text-expansions\"><\/a><a href=\"http:\/\/www.ergonis.com\/products\/typinator\/\" rel=\"nofollow\">Typinator<\/a> - Text expansions<\/h4>\n<ul><li>I use the app to <a href=\"https:\/\/medium.com\/@nikitavoloboev\/write-once-never-write-again-c2fa1f6c4e8\" rel=\"nofollow\">fully automate writing repetitive text<\/a>.<\/li>\n<li>I share all the Typinator sets I made with the app <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/typinator\">here<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-bettertouchtool---mac-input-customizer\" class=\"anchor\" aria-hidden=\"true\" href=\"#bettertouchtool---mac-input-customizer\"><\/a><a href=\"https:\/\/www.boastr.net\/\" rel=\"nofollow\">BetterTouchTool<\/a> - Mac input customizer<\/h4>\n<ul><li>I use this app a lot for mapping <a href=\"https:\/\/medium.com\/@nikitavoloboev\/take-control-of-your-touchpad-on-macos-45c581f542e0#.7n1ye6vze\" rel=\"nofollow\">various trackpad gestures<\/a> to hotkeys and actions both globally and per specific apps.<\/li>\n<li>I love scrolling through my tabs in Safari with three finger swipes left and right as well as opening and closing tabs with swiping up and down respectively.<\/li>\n<li>I share <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/btt#readme\">all the different gestures I have setup to use with the app<\/a> that you can view and download.<\/li>\n<li>I also use the app to modify how my TouchBar looks.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/a2a785176bec761631366af85bc046691ee6a3c6\/68747470733a2f2f692e696d6775722e636f6d2f427142477558712e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/a2a785176bec761631366af85bc046691ee6a3c6\/68747470733a2f2f692e696d6775722e636f6d2f427142477558712e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/BqBGuXq.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-pdf-expert---pdf-readereditor\" class=\"anchor\" aria-hidden=\"true\" href=\"#pdf-expert---pdf-readereditor\"><\/a><a href=\"https:\/\/pdfexpert.com\/\" rel=\"nofollow\">PDF Expert<\/a> - PDF reader\/editor<\/h4>\n<ul><li>I read a lot of PDFs like books, research papers and the like. This app is a huge upgrade over Preview app that I used before.<\/li>\n<li>Multiple tabs, sepia mode, very nice annotation tools, great search, performance.<\/li>\n<\/ul><h4><a id=\"user-content-fantastical---calendar\" class=\"anchor\" aria-hidden=\"true\" href=\"#fantastical---calendar\"><\/a><a href=\"https:\/\/flexibits.com\/fantastical\" rel=\"nofollow\">Fantastical<\/a> - Calendar<\/h4>\n<ul><li>I use the app to manage events in my life.<\/li>\n<li>I take great use of Fantastical's natural language input and I use <a href=\"https:\/\/medium.com\/@nikitavoloboev\/fantastical-natural-input-text-expansions-3ea8cf7ccac3#.pv5937ncr\" rel=\"nofollow\">many Typinator expansions<\/a> to ease this process.<\/li>\n<li>I always view my events from <code>Week<\/code> view. And show 5 days only with all 24h shown for all days. This lets me have a perspective over what I have to do now. What deadlines I have to complete soon. And gives me the freedom to adjust my schedule in light of upcoming deadlines and events.<\/li>\n<\/ul><h4><a id=\"user-content-contexts---window-switcher\" class=\"anchor\" aria-hidden=\"true\" href=\"#contexts---window-switcher\"><\/a><a href=\"https:\/\/contexts.co\" rel=\"nofollow\">Contexts<\/a> - Window switcher<\/h4>\n<ul><li>Allows me to fuzzy search through all the currently active windows that I have.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/91d9cf29b1d9f256e82fb9931223df2597c360a7\/68747470733a2f2f692e696d6775722e636f6d2f68776c56684d4b2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/91d9cf29b1d9f256e82fb9931223df2597c360a7\/68747470733a2f2f692e696d6775722e636f6d2f68776c56684d4b2e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/hwlVhMK.png\" \/><\/a><\/p>\n<ul><li>Makes jumping to the right window I need effortless. I often may have many VS Code instances with different projects running and this lets me switch to the project I need in seconds.<\/li>\n<\/ul><h4><a id=\"user-content-taskpaper---plain-text-todos\" class=\"anchor\" aria-hidden=\"true\" href=\"#taskpaper---plain-text-todos\"><\/a><a href=\"https:\/\/www.taskpaper.com\" rel=\"nofollow\">TaskPaper<\/a> - Plain text TODOs<\/h4>\n<ul><li>Use the app to <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/research\/solving-problems\" rel=\"nofollow\">dissect problems I am trying to solve into smaller and smaller parts I can work on<\/a>.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/145309ade0ea1c993172f7f738872f2ba933d8fb\/68747470733a2f2f692e696d6775722e636f6d2f4d7946496630392e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/145309ade0ea1c993172f7f738872f2ba933d8fb\/68747470733a2f2f692e696d6775722e636f6d2f4d7946496630392e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/MyFIf09.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-dictionary\" class=\"anchor\" aria-hidden=\"true\" href=\"#dictionary\"><\/a><a href=\"http:\/\/www.wikiwand.com\/en\/Dictionary_(software)\" rel=\"nofollow\">Dictionary<\/a><\/h4>\n<ul><li>Comes natively with macOS and I started to love using it for exploring and searching through Wikipedia.<\/li>\n<li>It is incredibly fast to make the searches and it gives quick autosuggestions for any query I type that I can then select with up and down arrows.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/16a8fb9cd4833cbc785408420878a2af5c37c019\/68747470733a2f2f692e696d6775722e636f6d2f4e6a34657276352e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/16a8fb9cd4833cbc785408420878a2af5c37c019\/68747470733a2f2f692e696d6775722e636f6d2f4e6a34657276352e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/Nj4erv5.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-bartender---menu-bar-organizer\" class=\"anchor\" aria-hidden=\"true\" href=\"#bartender---menu-bar-organizer\"><\/a><a href=\"https:\/\/www.macbartender.com\/\" rel=\"nofollow\">Bartender<\/a> - Menu bar organizer<\/h4>\n<ul><li>Allows you to customize and hide the contents of your menu bar and improve the aesthetics of your OS. Here is how mine looks:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/8108aff63714146ed16742dabcbe03fed4c9f834\/68747470733a2f2f692e696d6775722e636f6d2f684f7a4a69444b2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/8108aff63714146ed16742dabcbe03fed4c9f834\/68747470733a2f2f692e696d6775722e636f6d2f684f7a4a69444b2e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/hOzJiDK.png\" \/><\/a><\/p>\n<ul><li>My menu bar is always hidden so I rarely ever use it and I show most needed info on the <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/btt#readme\">TouchBar<\/a> instead.<\/li>\n<\/ul><h4><a id=\"user-content-focus---block-distracting-websites\" class=\"anchor\" aria-hidden=\"true\" href=\"#focus---block-distracting-websites\"><\/a><a href=\"https:\/\/heyfocus.com\" rel=\"nofollow\">Focus<\/a> - Block distracting websites<\/h4>\n<ul><li>I use <a href=\"https:\/\/heyfocus.com\" rel=\"nofollow\">Focus app<\/a> as a recurring <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pomodoro_Technique\" rel=\"nofollow\">pomodoro timer<\/a> which lets me work in sessions of 80 minutes with 10 min breaks by blocking <a href=\"https:\/\/gist.github.com\/nikitavoloboev\/57340528a6dfb3eb498dce647ed6841f\">distracting websites<\/a> and apps.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/561e4fc54c2693bf46e408a00c5556855b89047c\/68747470733a2f2f692e696d6775722e636f6d2f423954736178662e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/561e4fc54c2693bf46e408a00c5556855b89047c\/68747470733a2f2f692e696d6775722e636f6d2f423954736178662e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/B9Tsaxf.png\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/38e8a056a0a764be799bf39991361db277d9ba36\/68747470733a2f2f692e696d6775722e636f6d2f723666663643722e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/38e8a056a0a764be799bf39991361db277d9ba36\/68747470733a2f2f692e696d6775722e636f6d2f723666663643722e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/r6ff6Cr.png\" \/><\/a><\/p>\n<ul><li>If I go to any of my blocked websites during my <code>Focus<\/code> time, I am greeted with this message.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/cf95b5698b89c24115b93afada0eff9c15830592\/68747470733a2f2f692e696d6775722e636f6d2f4a4d4c347977682e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/cf95b5698b89c24115b93afada0eff9c15830592\/68747470733a2f2f692e696d6775722e636f6d2f4a4d4c347977682e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/JML4ywh.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-annotate---create-and-annotate-screenshots\" class=\"anchor\" aria-hidden=\"true\" href=\"#annotate---create-and-annotate-screenshots\"><\/a><a href=\"https:\/\/itunes.apple.com\/us\/app\/annotate-capture-screenshot\/id918207447?mt=12\" rel=\"nofollow\">Annotate<\/a> - Create and annotate screenshots<\/h4>\n<ul><li>Best annotation tool I found to exist. After you make a screenshot, it allows for quick edits. Adding arrows, some text, blurring parts of the image as well as ability to quickly save the image or drag it to Dropzone to upload to Imgur to share the link quickly with anyone I want.<\/li>\n<\/ul><h4><a id=\"user-content-popclip---ios-like-mouse-text-selection-popover\" class=\"anchor\" aria-hidden=\"true\" href=\"#popclip---ios-like-mouse-text-selection-popover\"><\/a><a href=\"https:\/\/pilotmoon.com\/popclip\/\" rel=\"nofollow\">Popclip<\/a> - iOS like mouse text selection popover<\/h4>\n<ul><li>Another great utility I cannot live without, brings up a quick menu whenever some text is selected on which I can do a number of quick actions, like searching on Google, DuckDuckGo, YouTube, Dictionary, Reddit, Google Images or it can translate text selected, copy it or even make a SnippetsLab snippet from it. Here is how it looks for me:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/2db3c4deec5eb8170155e6eeac849b38d395279b\/68747470733a2f2f692e696d6775722e636f6d2f52366b734a69332e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/2db3c4deec5eb8170155e6eeac849b38d395279b\/68747470733a2f2f692e696d6775722e636f6d2f52366b734a69332e706e67\" width=\"400\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/R6ksJi3.png\" \/><\/a><\/p>\n<ul><li>I share all the extensions I use with it <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/popclip#readme.md\">here<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-noizio---ambient-sounds\" class=\"anchor\" aria-hidden=\"true\" href=\"#noizio---ambient-sounds\"><\/a><a href=\"http:\/\/noiz.io\/\" rel=\"nofollow\">Noizio<\/a> - Ambient sounds<\/h4>\n<ul><li>I never listen to music when I want to focus on something as it is distracting. Instead I listen to sounds of rain, train or fire.<\/li>\n<\/ul><h4><a id=\"user-content-hammerspoon---automation-tool\" class=\"anchor\" aria-hidden=\"true\" href=\"#hammerspoon---automation-tool\"><\/a><a href=\"http:\/\/www.hammerspoon.org\" rel=\"nofollow\">Hammerspoon<\/a> - Automation tool<\/h4>\n<ul><li>Powerful automation engine I am still exploring. My config for it can be seen <a href=\"https:\/\/github.com\/nikitavoloboev\/dotfiles\/blob\/master\/hammerspoon\/\">here<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-reeder---rss-reeder\" class=\"anchor\" aria-hidden=\"true\" href=\"#reeder---rss-reeder\"><\/a><a href=\"http:\/\/reederapp.com\/mac\/\" rel=\"nofollow\">Reeder<\/a> - RSS Reeder<\/h4>\n<ul><li>The app I use to keep up with my RSS feeds. I use RSS to follow my favorite blogs, stay up to date on new podcast episodes and even track some software releases.<\/li>\n<li>I use <a href=\"https:\/\/www.inoreader.com\" rel=\"nofollow\">Inoreader<\/a> to sync <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/research\/blogs\" rel=\"nofollow\">blogs I follow<\/a> between phone and mac.<\/li>\n<li>Here is how Reeder looks like for me:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b6c3fd5d198e8c79c482d8bb12106e3d42b40a24\/68747470733a2f2f692e696d6775722e636f6d2f554256754149432e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/b6c3fd5d198e8c79c482d8bb12106e3d42b40a24\/68747470733a2f2f692e696d6775722e636f6d2f554256754149432e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/UBVuAIC.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-dropzone---drag-and-drop-actions\" class=\"anchor\" aria-hidden=\"true\" href=\"#dropzone---drag-and-drop-actions\"><\/a><a href=\"https:\/\/aptonic.com\" rel=\"nofollow\">Dropzone<\/a> - Drag and drop actions<\/h4>\n<ul><li>A quick utility that pops up when I drag some file into it or open it with a hotkey, use it to upload images to <a href=\"http:\/\/imgur.com\" rel=\"nofollow\">Imgur<\/a>, dropping files into a preset number of folders that I made quickly as well as going to them in an instant.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/a1be1a41a52dcd59d86e4f036a42fba229375a8e\/68747470733a2f2f692e696d6775722e636f6d2f35496e784d72452e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/a1be1a41a52dcd59d86e4f036a42fba229375a8e\/68747470733a2f2f692e696d6775722e636f6d2f35496e784d72452e706e67\" width=\"300\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/5InxMrE.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-little-snitch---control-incomingoutgoing-network-traffic\" class=\"anchor\" aria-hidden=\"true\" href=\"#little-snitch---control-incomingoutgoing-network-traffic\"><\/a><a href=\"https:\/\/www.obdev.at\/products\/littlesnitch\/index.html\" rel=\"nofollow\">Little Snitch<\/a> - Control incoming\/outgoing network traffic<\/h4>\n<ul><li>Amazing networking tool that gives you a clear picture of what connections are incoming to your computer and what are outgoing.<\/li>\n<li>Takes a bit of time to set it up correctly and is quite an insightful experience first turning it on and having it notify every couple of seconds that some app is trying to send data to some server and whether you want to allow that.<\/li>\n<li>This is essential if you want to take control of what information gets sent out from your computer and what connections have the right to connect to your data.<\/li>\n<\/ul><h4><a id=\"user-content-beardedspice---extend-mac-media-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#beardedspice---extend-mac-media-keys\"><\/a><a href=\"https:\/\/github.com\/beardedspice\/beardedspice\">BeardedSpice<\/a> - Extend Mac Media Keys<\/h4>\n<ul><li>Enhance play\/pause as well as previous\/next playback keys to not only work in iTunes and Spotify but work for SoundCloud, YouTube and many other services.<\/li>\n<\/ul><h4><a id=\"user-content-transmission---bittorrent-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#transmission---bittorrent-client\"><\/a><a href=\"https:\/\/www.transmissionbt.com\/\" rel=\"nofollow\">Transmission<\/a> - BitTorrent client<\/h4>\n<ul><li>A torrent client that I use. Very minimal in its UI but very powerful and has all the features that I need.<\/li>\n<\/ul><h4><a id=\"user-content-transmit---transfer-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#transmit---transfer-files\"><\/a><a href=\"https:\/\/www.panic.com\/transmit\/\" rel=\"nofollow\">Transmit<\/a> - Transfer files<\/h4>\n<ul><li>I sometimes use this app to quickly send files from my local file system to the cloud and get a shareable link I can send to people.<\/li>\n<li>I use <a href=\"https:\/\/github.com\/bigluck\/alfred2-transmit\">this Alfred workflow<\/a> to quickly open the cloud storage I need.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/880dd9a154874c14453a6c5a29e278c95f8bd257\/68747470733a2f2f692e696d6775722e636f6d2f546643736f6b742e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/880dd9a154874c14453a6c5a29e278c95f8bd257\/68747470733a2f2f692e696d6775722e636f6d2f546643736f6b742e706e67\" width=\"400\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/TfCsokt.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-paprika-recipe-manager\" class=\"anchor\" aria-hidden=\"true\" href=\"#paprika-recipe-manager\"><\/a><a href=\"https:\/\/itunes.apple.com\/us\/app\/paprika-recipe-manager-3\/id1303222628?ls=1&amp;mt=12\" rel=\"nofollow\">Paprika Recipe Manager<\/a><\/h4>\n<ul><li>Use the app to store the many <a href=\"https:\/\/github.com\/nikitavoloboev\/vegan-recipes#readme\">vegan recipes<\/a> I like to make.<\/li>\n<\/ul><h3><a id=\"user-content-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#code\"><\/a>Code<\/h3>\n<h4><a id=\"user-content-vs-code---code-editor\" class=\"anchor\" aria-hidden=\"true\" href=\"#vs-code---code-editor\"><\/a><a href=\"https:\/\/github.com\/Microsoft\/vscode\">VS Code<\/a> - Code editor<\/h4>\n<ul><li>My favorite editor that I use to write code in. I use <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/text-editors\/vs-code\/vs-code-extensions\" rel=\"nofollow\">many extensions<\/a> for it.<\/li>\n<li>My config for it can be found <a href=\"https:\/\/github.com\/nikitavoloboev\/dotfiles\/blob\/master\/vscode\/settings.json\">here<\/a>.<\/li>\n<li>I use <a href=\"https:\/\/github.com\/fabiospampinato\/vscode-monokai-night#readme\">VS Code Monokai Night<\/a> theme with <a href=\"https:\/\/github.com\/tonsky\/FiraCode\">Fira Code<\/a> font. Here is how it looks:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/56cad12c3b255993242a3b3a20833858cd519387\/68747470733a2f2f692e696d6775722e636f6d2f684f67535137462e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/56cad12c3b255993242a3b3a20833858cd519387\/68747470733a2f2f692e696d6775722e636f6d2f684f67535137462e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/hOgSQ7F.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-iterm---terminal-emulator\" class=\"anchor\" aria-hidden=\"true\" href=\"#iterm---terminal-emulator\"><\/a><a href=\"https:\/\/www.iterm2.com\/\" rel=\"nofollow\">iTerm<\/a> - Terminal Emulator<\/h4>\n<ul><li>Use Zsh as my shell together with <a href=\"https:\/\/github.com\/getantibody\/antibody\">Antibody<\/a> to install <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/unix\/shell\/zsh\/zsh-plugins\" rel=\"nofollow\">Zsh plugins I use<\/a>.<\/li>\n<li>Assigned w + j with <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/macOS\/apps\/karabiner\/karabiner\" rel=\"nofollow\">Karabiner<\/a> to open the app from Keyboard Maestro in seconds.<\/li>\n<li>I made my own <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/iterm#readme\">Monokai Night theme<\/a> for it.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/86c0b418cbaad4888efca773d7fe00ab20ca4880\/68747470733a2f2f692e696d6775722e636f6d2f53516271504c4e2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/86c0b418cbaad4888efca773d7fe00ab20ca4880\/68747470733a2f2f692e696d6775722e636f6d2f53516271504c4e2e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/SQbqPLN.png\" \/><\/a><\/p>\n<ul><li>I also love using <a href=\"https:\/\/github.com\/isometry\/alfred-tty\">this workflow<\/a> to quickly switch between active tty sessions.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/5d621e18bff2152a473e594b246ecf20cc3c47b5\/68747470733a2f2f692e696d6775722e636f6d2f6b646f307838592e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/5d621e18bff2152a473e594b246ecf20cc3c47b5\/68747470733a2f2f692e696d6775722e636f6d2f6b646f307838592e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/kdo0x8Y.png\" \/><\/a><\/p>\n<ul><li>I go over how I use the app <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/macOS\/apps\/iterm\" rel=\"nofollow\">here<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-sublime-text---text-editor\" class=\"anchor\" aria-hidden=\"true\" href=\"#sublime-text---text-editor\"><\/a><a href=\"https:\/\/www.sublimetext.com\" rel=\"nofollow\">Sublime Text<\/a> - Text Editor<\/h4>\n<ul><li>Use this editor in addition to VS Code and Neovim for its blazing fast speed of opening files.<\/li>\n<li>I use it primarily to edit markdown files like <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/other\/wiki-workflow\" rel=\"nofollow\">my wiki<\/a>. I also edit config files and open large and small files for quick edits.<\/li>\n<li>I use <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/text-editors\/sublime-text\/sublime-text-plugins\" rel=\"nofollow\">many plugins<\/a> together with <a href=\"https:\/\/github.com\/dempfi\/ayu\">Ayu theme<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-dash---api-documentation-browser\" class=\"anchor\" aria-hidden=\"true\" href=\"#dash---api-documentation-browser\"><\/a><a href=\"https:\/\/kapeli.com\/dash\" rel=\"nofollow\">Dash<\/a> - API Documentation Browser<\/h4>\n<ul><li>Allows you to download any docset that you might want to use, search for any method, class or anything that you need very quickly, comes with the amazing <a href=\"https:\/\/www.alfredapp.com\/blog\/productivity\/dash-quicker-api-documentation-search\/\" rel=\"nofollow\">Alfred workflow<\/a> to simplify the process of searching for the right things.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/d09106a5a1c71c8b924898f507af2a6840ddbd68\/68747470733a2f2f692e696d6775722e636f6d2f6449746e4235322e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/d09106a5a1c71c8b924898f507af2a6840ddbd68\/68747470733a2f2f692e696d6775722e636f6d2f6449746e4235322e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/dItnB52.png\" \/><\/a><\/p>\n<ul><li>I also use <a href=\"https:\/\/github.com\/nikitavoloboev\/small-workflows\/tree\/master\/dash-profile-switch#readme\">this workflow<\/a> I made to quickly switch between Dash profiles.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/f2fed603bea7fc0e0705d715ee0cb36dce14efff\/68747470733a2f2f692e696d6775722e636f6d2f67437a3473397a2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/f2fed603bea7fc0e0705d715ee0cb36dce14efff\/68747470733a2f2f692e696d6775722e636f6d2f67437a3473397a2e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/gCz4s9z.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-snippetslab---snippet-manager\" class=\"anchor\" aria-hidden=\"true\" href=\"#snippetslab---snippet-manager\"><\/a><a href=\"https:\/\/www.renfei.org\/snippets-lab\/\" rel=\"nofollow\">SnippetsLab<\/a> - Snippet manager<\/h4>\n<ul><li>I use the app to manage my own personal library of snippets. I prefix all snippets I make. For example vim snippets are prefixed with <code>vim:<\/code>. Git related snippets with <code>git:<\/code> and so on.<\/li>\n<li>This lets me then use the <a href=\"https:\/\/www.renfei.org\/snippets-lab\/press-release\/whats-new\/osx-1.6\" rel=\"nofollow\">Alfred workflow<\/a> that the app's author provides to search for these snippets insanely fast.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b32a79736c93506b1e847466c6f88381c5431837\/68747470733a2f2f692e696d6775722e636f6d2f514372626674542e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/b32a79736c93506b1e847466c6f88381c5431837\/68747470733a2f2f692e696d6775722e636f6d2f514372626674542e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/QCrbftT.png\" \/><\/a><\/p>\n<p>Here is how my library looks:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/fcba70a0a2f5954a6a40e4dc7b0e88811bbcc628\/68747470733a2f2f692e696d6775722e636f6d2f46376a314d536b2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/fcba70a0a2f5954a6a40e4dc7b0e88811bbcc628\/68747470733a2f2f692e696d6775722e636f6d2f46376a314d536b2e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/F7j1MSk.png\" \/><\/a><\/p>\n<ul><li>I share my entire library of snippets as <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/sharing\/my-gists\" rel=\"nofollow\">GitHub Gists<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-tower---git-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#tower---git-client\"><\/a><a href=\"https:\/\/www.git-tower.com\" rel=\"nofollow\">Tower<\/a> - Git client<\/h4>\n<ul><li>Use the app to help me version control any project I am working on. Love the ability to search through commits, check out branches with ease and searching through all the Git repositories I have on my mac with quick search.\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/19911444dc1c8065e8ac94f224a29de679681341\/68747470733a2f2f692e696d6775722e636f6d2f457a394d6d4c652e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/19911444dc1c8065e8ac94f224a29de679681341\/68747470733a2f2f692e696d6775722e636f6d2f457a394d6d4c652e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/Ez9MmLe.png\" \/><\/a><\/li>\n<\/ul><h4><a id=\"user-content-paw---http-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#paw---http-client\"><\/a><a href=\"https:\/\/paw.cloud\" rel=\"nofollow\">Paw<\/a> - HTTP client<\/h4>\n<ul><li>Use the tool to quickly make HTTP requests and test out API endpoints.<\/li>\n<\/ul><h4><a id=\"user-content-neovim---text-editor\" class=\"anchor\" aria-hidden=\"true\" href=\"#neovim---text-editor\"><\/a><a href=\"https:\/\/neovim.io\" rel=\"nofollow\">Neovim<\/a> - Text Editor<\/h4>\n<ul><li>The best text editing experience you will ever get. It is worth learning it as you will start thinking about text editing differently.<\/li>\n<li>I use some kind of vim bindings in any app I use and if there is a vim plugin for the app, I often use it. I use <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/text-editors\/vim\/vim\" rel=\"nofollow\">vim editor<\/a> often when connecting to remote sessions and when I need a console editor.<\/li>\n<li>My <a href=\"https:\/\/github.com\/neovim\/neovim\">Neovim<\/a> config can be seen <a href=\"https:\/\/github.com\/nikitavoloboev\/dotfiles\/blob\/master\/nvim\/init.vim\">here<\/a>. I use <a href=\"https:\/\/github.com\/nikitavoloboev\/vim-monokai-night#readme\">Monokai Night theme<\/a> with <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/text-editors\/vim\/vim-plugins\" rel=\"nofollow\">many plugins<\/a>. Here is how it looks:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/46b4c468567032c2b9130758f7d147089bf84a30\/68747470733a2f2f692e696d6775722e636f6d2f734c58427676372e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/46b4c468567032c2b9130758f7d147089bf84a30\/68747470733a2f2f692e696d6775722e636f6d2f734c58427676372e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/sLXBvv7.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-bee---github-issue-tracker\" class=\"anchor\" aria-hidden=\"true\" href=\"#bee---github-issue-tracker\"><\/a><a href=\"https:\/\/www.neat.io\/bee\/\" rel=\"nofollow\">Bee<\/a> - GitHub Issue Tracker<\/h4>\n<ul><li>Use the app to stay on top of all issues across my GitHub repos. All issues I want to work on soon are put into <strong>Next<\/strong> queue and I work through each one, one by one.<\/li>\n<\/ul><h3><a id=\"user-content-social\" class=\"anchor\" aria-hidden=\"true\" href=\"#social\"><\/a>Social<\/h3>\n<h4><a id=\"user-content-textual---irc-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#textual---irc-client\"><\/a><a href=\"https:\/\/www.codeux.com\/textual\/\" rel=\"nofollow\">Textual<\/a> - IRC Client<\/h4>\n<ul><li>I love IRC and this is the best macOS client for it.<\/li>\n<li>I created my own <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/textual#readme\">Monokai Night theme<\/a> for it that I love.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/76988867b76167a941e7b958ef870a650ea737db\/68747470733a2f2f692e696d6775722e636f6d2f4c706f7a6b38582e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/76988867b76167a941e7b958ef870a650ea737db\/68747470733a2f2f692e696d6775722e636f6d2f4c706f7a6b38582e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/Lpozk8X.png\" \/><\/a><\/p>\n<ul><li>The app also has an awesome channel search feature that I use a lot.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/ab637751ab2365a69482eea4164bdb84de816237\/68747470733a2f2f692e696d6775722e636f6d2f434d42573851662e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/ab637751ab2365a69482eea4164bdb84de816237\/68747470733a2f2f692e696d6775722e636f6d2f434d42573851662e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/CMBW8Qf.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-telegram---messenger\" class=\"anchor\" aria-hidden=\"true\" href=\"#telegram---messenger\"><\/a><a href=\"https:\/\/desktop.telegram.org\/\" rel=\"nofollow\">Telegram<\/a> - Messenger<\/h4>\n<ul><li>The app has a very clean interface, a native client for macOS and stickers. Chats are not encrypted by default and don't use <a href=\"https:\/\/security.stackexchange.com\/questions\/49782\/is-telegram-secure\" rel=\"nofollow\">approved cryptography<\/a> but the client is amazing and fast.<\/li>\n<\/ul><h4><a id=\"user-content-spark---email-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#spark---email-client\"><\/a><a href=\"https:\/\/sparkmailapp.com\" rel=\"nofollow\">Spark<\/a> - Email client<\/h4>\n<ul><li>Love how it smartly categorizes emails by categories.<\/li>\n<li>I approach all of my email tasks in GTD style. Keeping my email Inbox close to 0 at all times.<\/li>\n<\/ul><h4><a id=\"user-content-tweetbot---twitter-client\" class=\"anchor\" aria-hidden=\"true\" href=\"#tweetbot---twitter-client\"><\/a><a href=\"http:\/\/tapbots.com\/tweetbot\/mac\/\" rel=\"nofollow\">Tweetbot<\/a> - Twitter client<\/h4>\n<ul><li>I use the app to stay up to date and communicate on Twitter.<\/li>\n<li>Tweetbot is also blocked for me during my <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/focusing\/focusing\" rel=\"nofollow\">Focus time<\/a>. I can only post tweets but I can't visit the app to see my feeds.<\/li>\n<li>Twitter is my primary way to stay up to date on things and is my favorite social network after GitHub. I curate and publicly share all <a href=\"https:\/\/twitter.com\/nikitavoloboev\/lists\" rel=\"nofollow\">Twitter lists<\/a> I personally use.<\/li>\n<li>I have Tweetbot open in full screen with four columns (Mentions\/<a href=\"https:\/\/twitter.com\/nikitavoloboev\/lists\/top\" rel=\"nofollow\">Top<\/a>\/<a href=\"https:\/\/twitter.com\/nikitavoloboev\/lists\/ml\" rel=\"nofollow\">ML<\/a>\/<a href=\"https:\/\/twitter.com\/nikitavoloboev\/lists\/other1\" rel=\"nofollow\">Other<\/a>). Here is how that looks for me:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/5544eb1df53c14afb9ea5fbbcdc85c6050d154c2\/68747470733a2f2f692e696d6775722e636f6d2f6e3630626c38772e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/5544eb1df53c14afb9ea5fbbcdc85c6050d154c2\/68747470733a2f2f692e696d6775722e636f6d2f6e3630626c38772e6a7067\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/n60bl8w.jpg\" \/><\/a><\/p>\n<h3><a id=\"user-content-writing\" class=\"anchor\" aria-hidden=\"true\" href=\"#writing\"><\/a>Writing<\/h3>\n<h4><a id=\"user-content-day-one---digital-journal\" class=\"anchor\" aria-hidden=\"true\" href=\"#day-one---digital-journal\"><\/a><a href=\"http:\/\/dayoneapp.com\/\" rel=\"nofollow\">Day One<\/a> - Digital journal<\/h4>\n<ul><li>Day One is my digital life journal.<\/li>\n<li>My entire life's experiences lives encrypted in the journal. Adding photographs I made, the thoughts I had and events that happened to me.<\/li>\n<li>I document and review <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/looking-back\/looking-back\" rel=\"nofollow\">my life<\/a> publicly too.<\/li>\n<li>The app has integration with <a href=\"https:\/\/ifttt.com\/day_one\" rel=\"nofollow\">IFTTT<\/a> which I use to automatically log all <a href=\"https:\/\/twitter.com\/nikitavoloboev\" rel=\"nofollow\">my tweets<\/a> and <a href=\"https:\/\/instagram.com\/nikitavoloboev\" rel=\"nofollow\">Instagram posts<\/a>.<\/li>\n<li>If you don't journal, I suggest you to start, it is a very powerful mind cleanser and acts as a wonderful history record of your life.<\/li>\n<\/ul><h4><a id=\"user-content-ulysses---writing-app\" class=\"anchor\" aria-hidden=\"true\" href=\"#ulysses---writing-app\"><\/a><a href=\"http:\/\/www.ulyssesapp.com\/\" rel=\"nofollow\">Ulysses<\/a> - Writing app<\/h4>\n<ul><li>The app I use to write all my <a href=\"https:\/\/medium.com\/@nikitavoloboev\" rel=\"nofollow\">Medium<\/a> articles in.<\/li>\n<li>All your writing in one place is the motto and the underlying design of the app. I hook up my <a href=\"https:\/\/github.com\/nikitavoloboev\/knowledge\">wiki<\/a> as an external folder and I love how it creates an index of everything that I can search over:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/f32d349d4bf010a99c6312dbc559966170ca4afc\/687474703a2f2f692e696d6775722e636f6d2f416131375243512e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/f32d349d4bf010a99c6312dbc559966170ca4afc\/687474703a2f2f692e696d6775722e636f6d2f416131375243512e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"http:\/\/i.imgur.com\/Aa17RCQ.png\" \/><\/a><\/p>\n<ul><li>It also has pretty awesome export functions that let you export the text you write to PDF, ePub and HTML. I mostly use it's ability to export text to Medium.<\/li>\n<li>I use <a href=\"https:\/\/styles.ulyssesapp.com\/bundle\/Monokai+Night\/5bbcb619b3dafa6b4e6d6bd4\" rel=\"nofollow\">Monokai Night<\/a> theme when I write. Here is how it looks:<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/531b8be7a1fd50d824f5b078ec752a255586f834\/68747470733a2f2f692e696d6775722e636f6d2f357966786c75502e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/531b8be7a1fd50d824f5b078ec752a255586f834\/68747470733a2f2f692e696d6775722e636f6d2f357966786c75502e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/5yfxluP.png\" \/><\/a><\/p>\n<ul><li>I describe my thoughts and approach to writing <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/writing\/writing\" rel=\"nofollow\">here<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-marked---preview-rendered-markdown-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#marked---preview-rendered-markdown-files\"><\/a><a href=\"http:\/\/marked2app.com\" rel=\"nofollow\">Marked<\/a> - Preview rendered markdown files<\/h4>\n<ul><li>If I ever need to preview a readme or any other markdown file that I wrote or just want to read, this app is the best application for that I found. It also features live updating and quite a lot of customizable features.<\/li>\n<\/ul><h3><a id=\"user-content-design\" class=\"anchor\" aria-hidden=\"true\" href=\"#design\"><\/a>Design<\/h3>\n<h4><a id=\"user-content-sketch---design-tool\" class=\"anchor\" aria-hidden=\"true\" href=\"#sketch---design-tool\"><\/a><a href=\"https:\/\/www.sketchapp.com\/\" rel=\"nofollow\">Sketch<\/a> - Design tool<\/h4>\n<ul><li>Use the app to quickly prototype new designs. I also enjoy using <a href=\"https:\/\/www.figma.com\" rel=\"nofollow\">Figma<\/a> a lot.<\/li>\n<\/ul><h4><a id=\"user-content-sip---collect-organize--share-colors\" class=\"anchor\" aria-hidden=\"true\" href=\"#sip---collect-organize--share-colors\"><\/a><a href=\"https:\/\/sipapp.io\/\" rel=\"nofollow\">Sip<\/a> - Collect, organize &amp; share colors<\/h4>\n<ul><li>A great color picker I use to collect my favorite colors and color schemes.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/bea9dcb24b37bad974f30da4fea23885de1c0711\/68747470733a2f2f692e696d6775722e636f6d2f5a53695a3551412e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/bea9dcb24b37bad974f30da4fea23885de1c0711\/68747470733a2f2f692e696d6775722e636f6d2f5a53695a3551412e706e67\" width=\"400\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/ZSiZ5QA.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-pixelsnap---measure-everything-on-screen\" class=\"anchor\" aria-hidden=\"true\" href=\"#pixelsnap---measure-everything-on-screen\"><\/a><a href=\"https:\/\/getpixelsnap.com\/\" rel=\"nofollow\">PixelSnap<\/a> - Measure everything on screen<\/h4>\n<ul><li>Use it to quickly get measurements of objects and distances between elements.<\/li>\n<\/ul><h3><a id=\"user-content-music\" class=\"anchor\" aria-hidden=\"true\" href=\"#music\"><\/a>Music<\/h3>\n<h4><a id=\"user-content-spotify---music-streaming\" class=\"anchor\" aria-hidden=\"true\" href=\"#spotify---music-streaming\"><\/a><a href=\"https:\/\/www.spotify.com\/us\/\" rel=\"nofollow\">Spotify<\/a> - Music streaming<\/h4>\n<ul><li>Found a <a href=\"https:\/\/open.spotify.com\/user\/nikitavoloboev\" rel=\"nofollow\">lot of great music<\/a> with this application and the phenomenal <a href=\"http:\/\/alfred-spotify-mini-player.com\/\" rel=\"nofollow\">Alfred Workflow<\/a> makes using the application an absolute joy.<\/li>\n<li>Quickly finding artists, songs I want to listen, instantly adding the song playing to my <a href=\"https:\/\/open.spotify.com\/user\/nikitavoloboev\/playlist\/0ERn0U4qZIKC8Dy7RrMMsn?\" rel=\"nofollow\">Likes<\/a> playlist or any other playlist I want, seeing what other songs the artist has and more.<\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/01b9a6b785a948f323fb760496f763b55448680a\/68747470733a2f2f692e696d6775722e636f6d2f744e54685647302e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/01b9a6b785a948f323fb760496f763b55448680a\/68747470733a2f2f692e696d6775722e636f6d2f744e54685647302e706e67\" width=\"500\" alt=\"img\" data-canonical-src=\"https:\/\/i.imgur.com\/tNThVG0.png\" \/><\/a><\/p>\n<h3><a id=\"user-content-images\" class=\"anchor\" aria-hidden=\"true\" href=\"#images\"><\/a>Images<\/h3>\n<h4><a id=\"user-content-pixave---imagegifvideo-organizer\" class=\"anchor\" aria-hidden=\"true\" href=\"#pixave---imagegifvideo-organizer\"><\/a><a href=\"http:\/\/www.littlehj.com\/mac\" rel=\"nofollow\">Pixave<\/a> - Image\/GIF\/Video organizer<\/h4>\n<ul><li>\n<p>I use Pixave as my own personal image library. I sort images in groups and I have powerful search I can use to find any image I need.<\/p>\n<\/li>\n<li>\n<p>Here is how my library looks like:\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/5f2d714e8b179508f5fd489ce62854af19e58960\/68747470733a2f2f692e696d6775722e636f6d2f4d6a47467653622e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/5f2d714e8b179508f5fd489ce62854af19e58960\/68747470733a2f2f692e696d6775722e636f6d2f4d6a47467653622e6a7067\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/MjGFvSb.jpg\" \/><\/a><\/p>\n<\/li>\n<li>\n<p>The red tag means that the image was uploaded to my <a href=\"https:\/\/instagram.com\/prettiways\" rel=\"nofollow\">Instagram page<\/a>.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-video\" class=\"anchor\" aria-hidden=\"true\" href=\"#video\"><\/a>Video<\/h3>\n<h4><a id=\"user-content-iina---video-player\" class=\"anchor\" aria-hidden=\"true\" href=\"#iina---video-player\"><\/a><a href=\"https:\/\/github.com\/lhc70000\/iina\">IINA<\/a> - Video player<\/h4>\n<ul><li>Open source alternative to VLC built specifically for macOS.<\/li>\n<li>It is based on <a href=\"https:\/\/github.com\/mpv-player\/mpv\">mpv<\/a> and has a more modern and native look than VLC.<\/li>\n<\/ul><h4><a id=\"user-content-kap---screen-recorder\" class=\"anchor\" aria-hidden=\"true\" href=\"#kap---screen-recorder\"><\/a><a href=\"https:\/\/github.com\/wulkano\/kap\">Kap<\/a> - Screen recorder<\/h4>\n<ul><li>An open source screen recorder I use to record GIFs.<\/li>\n<li>Has keyboard support so I can quickly start and end recording of the GIF in one hotkey.<\/li>\n<\/ul><h3><a id=\"user-content-utilities\" class=\"anchor\" aria-hidden=\"true\" href=\"#utilities\"><\/a>Utilities<\/h3>\n<h4><a id=\"user-content-hazel---automated-file-organization\" class=\"anchor\" aria-hidden=\"true\" href=\"#hazel---automated-file-organization\"><\/a><a href=\"https:\/\/www.noodlesoft.com\" rel=\"nofollow\">Hazel<\/a> - Automated File Organization<\/h4>\n<ul><li>Use it to fully automate my filing process and automatically commit changes in <a href=\"https:\/\/github.com\/learn-anything\/curated-lists#readme\">many curated lists<\/a>.<\/li>\n<li>I share most Hazel rules and how I use them <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/hazel#readme\">here<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-daisydisk---visualize-taken-disk-space\" class=\"anchor\" aria-hidden=\"true\" href=\"#daisydisk---visualize-taken-disk-space\"><\/a><a href=\"https:\/\/daisydiskapp.com\" rel=\"nofollow\">DaisyDisk<\/a> - Visualize taken disk space<\/h4>\n<ul><li>A great tool to quickly get a visual glance over what is taking up your disk space and where.<\/li>\n<\/ul><h4><a id=\"user-content-istat-menus---mac-system-monitoring-from-menu-bar\" class=\"anchor\" aria-hidden=\"true\" href=\"#istat-menus---mac-system-monitoring-from-menu-bar\"><\/a><a href=\"https:\/\/bjango.com\/mac\/istatmenus\/\" rel=\"nofollow\">iStat Menus<\/a> - Mac system monitoring from menu bar<\/h4>\n<ul><li>Great system monitoring tools which I use to quickly see my CPU\/network\/memory usage. I find it very valuable to know if my CPU is being abused by some application and if so by which. I then use <a href=\"https:\/\/github.com\/vitorgalvao\/alfred-workflows\/tree\/master\/ProcessControl#readme\">ProcessControl<\/a> workflow to quickly kill off misbehaving processes. Often times iStat Menus is deactivated for me to preserve battery.<\/li>\n<\/ul><h4><a id=\"user-content-podcastmenu---puts-overcast-in-menu-bar\" class=\"anchor\" aria-hidden=\"true\" href=\"#podcastmenu---puts-overcast-in-menu-bar\"><\/a><a href=\"https:\/\/github.com\/insidegui\/PodcastMenu\">PodcastMenu<\/a> - Puts Overcast in menu bar<\/h4>\n<ul><li>Allows you to listen and control playback of podcasts from <a href=\"https:\/\/overcast.fm\" rel=\"nofollow\">Overcast<\/a> right from your menu bar.<\/li>\n<\/ul><h3><a id=\"user-content-browsers\" class=\"anchor\" aria-hidden=\"true\" href=\"#browsers\"><\/a>Browsers<\/h3>\n<h4><a id=\"user-content-safari\" class=\"anchor\" aria-hidden=\"true\" href=\"#safari\"><\/a><a href=\"https:\/\/www.apple.com\/lae\/safari\/\" rel=\"nofollow\">Safari<\/a><\/h4>\n<ul><li>My favorite browser. It is incredibly fast, doesn't have the baggage of all the Google Chrome tracking and is native to macOS so it is extremely optimized and doesn't kill your battery.<\/li>\n<li>One powerful plus that Chrome does have over safari is the amount of Chrome extensions that you can get. Fortunately the ones most needed for me do <a href=\"https:\/\/github.com\/learn-anything\/safari-extensions#readme\">exist on Safari<\/a>.<\/li>\n<\/ul><h5><a id=\"user-content-safari-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#safari-extensions\"><\/a>Safari Extensions<\/h5>\n<ul><li><a href=\"https:\/\/github.com\/flipxfx\/sVim\">sVim<\/a> - Adds a custom layer of keybinds you can customize to personalize your browsing experience. <a href=\"https:\/\/gist.github.com\/c26e6a05e4e426e0542e55b7513b581c\">My config<\/a>.<\/li>\n<li><a href=\"https:\/\/itunes.apple.com\/nl\/app\/wipr\/id1320666476?l=en&amp;mt=12\" rel=\"nofollow\">Wipr<\/a> - Ad content blocker that uses <a href=\"https:\/\/developer.apple.com\/library\/content\/documentation\/Extensions\/Conceptual\/ContentBlockingRules\/Introduction\/Introduction.html\" rel=\"nofollow\">Safari native content blocking API's<\/a>.<\/li>\n<li><a href=\"https:\/\/agilebits.com\/onepassword\/extensions\" rel=\"nofollow\">1Password<\/a> - <a href=\"https:\/\/1password.com\" rel=\"nofollow\">1Password<\/a> is a phenomenal password manager, this extension just gives a seamless interaction of it with the browser.<\/li>\n<li><a href=\"https:\/\/cascadea.app\" rel=\"nofollow\">Cascadea<\/a> - Custom styles extension. Use it to create dark themes for apps that don't have a dark mode option. Also allows importing of <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/web\/browsers\/stylish-themes\" rel=\"nofollow\">Stylish themes<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-google-chrome\" class=\"anchor\" aria-hidden=\"true\" href=\"#google-chrome\"><\/a><a href=\"https:\/\/www.google.com\/chrome\/\" rel=\"nofollow\">Google Chrome<\/a><\/h4>\n<ul><li>I only use Chrome for web development due to its superior Chrome Dev tools.<\/li>\n<li>I also love the variety of <a href=\"https:\/\/github.com\/learn-anything\/chrome-extensions#readme\">Chrome extensions<\/a> that people have built and shared.<\/li>\n<\/ul><h4><a id=\"user-content-beaker-browser\" class=\"anchor\" aria-hidden=\"true\" href=\"#beaker-browser\"><\/a><a href=\"https:\/\/beakerbrowser.com\" rel=\"nofollow\">Beaker Browser<\/a><\/h4>\n<ul><li>Experimental Peer to Peer browser. Experimenting with using it more and more as I find the <a href=\"https:\/\/www.youtube.com\/watch?v=-ep0ZIe6i10\" rel=\"nofollow\">technology and the implications of it fascinating<\/a>.<\/li>\n<\/ul><h4><a id=\"user-content-firefox-developer-edition\" class=\"anchor\" aria-hidden=\"true\" href=\"#firefox-developer-edition\"><\/a><a href=\"https:\/\/www.mozilla.org\/en-US\/firefox\/developer\/\" rel=\"nofollow\">Firefox Developer Edition<\/a><\/h4>\n<ul><li>Use it for web development for its superior Dev Tools in debugging CSS Grid. Prefer Chrome Dev tools still.<\/li>\n<\/ul><h2><a id=\"user-content-command-line-apps\" class=\"anchor\" aria-hidden=\"true\" href=\"#command-line-apps\"><\/a>Command Line Apps<\/h2>\n<p>I use <a href=\"https:\/\/nixos.org\/nix\/\" rel=\"nofollow\">Nix<\/a> package manager to install all the <a href=\"https:\/\/gist.github.com\/nikitavoloboev\/3fbe13ce427132d0297f411b62f49034\">packages I use on my system<\/a>.<\/p>\n<p>I curate a list of <a href=\"https:\/\/github.com\/learn-anything\/command-line-tools#readme\">interesting CLI tools<\/a> you can use. Below are some command line tools I personally use and love.<\/p>\n<ul><li><a href=\"https:\/\/github.com\/junegunn\/fzf\">fzf<\/a> - Command-line fuzzy finder.<\/li>\n<li><a href=\"https:\/\/github.com\/BurntSushi\/ripgrep\">ripgrep<\/a> - Search text for patterns fast.<\/li>\n<li><a href=\"https:\/\/github.com\/sharkdp\/fd\">fd<\/a> - Simple, fast and user-friendly alternative to 'find'.<\/li>\n<li><a href=\"https:\/\/github.com\/ogham\/exa\">exa<\/a> - Replacement for ls written in rust.<\/li>\n<li><a href=\"https:\/\/github.com\/sharkdp\/bat\">bat<\/a> - Cat clone with wings.<\/li>\n<li><a href=\"https:\/\/github.com\/apex\/up\">up<\/a> - Deploy infinitely scalable serverless apps, APIs, and sites in seconds to AWS.<\/li>\n<li><a href=\"https:\/\/github.com\/stedolan\/jq\">jq<\/a> - JSON processor.<\/li>\n<li><a href=\"https:\/\/github.com\/git\/git\">git<\/a> - Version control.<\/li>\n<li><a href=\"https:\/\/curl.haxx.se\/docs\/manpage.html\" rel=\"nofollow\">curl<\/a> - Transfer data from or to a server.<\/li>\n<li><a href=\"https:\/\/github.com\/rg3\/youtube-dl\">youtube-dl<\/a> - Download videos from YouTube and other video sites.<\/li>\n<li><a href=\"https:\/\/github.com\/tmux\/tmux\">tmux<\/a> - Terminal multiplexer.<\/li>\n<li><a href=\"https:\/\/direnv.net\/\" rel=\"nofollow\">direnv<\/a> - Environment switcher for the shell.<\/li>\n<li><a href=\"https:\/\/github.com\/hishamhm\/htop\">htop<\/a> - Interactive text-mode process viewer for Unix systems.<\/li>\n<li><a href=\"https:\/\/github.com\/jakubroztocil\/httpie\">httpie<\/a> - HTTP client.<\/li>\n<li><a href=\"https:\/\/github.com\/dflemstr\/rq\">rq<\/a> - Tool for doing record analysis and transformation.<\/li>\n<li><a href=\"https:\/\/github.com\/jgm\/pandoc\">pandoc<\/a> - Universal markup converter.<\/li>\n<li><a href=\"https:\/\/github.com\/sindresorhus\/trash\">trash<\/a> - Move files and folders to the trash.<\/li>\n<li><a href=\"https:\/\/github.com\/MrRio\/vtop\">vtop<\/a> - Graphical activity monitor.<\/li>\n<li><a href=\"https:\/\/github.com\/cjbassi\/gotop\">gotop<\/a> - Terminal based graphical activity monitor inspired by gtop and vtop.<\/li>\n<li><a href=\"https:\/\/github.com\/curl\/curl\">curl<\/a> - Transfer data, supports various protocols.<\/li>\n<li><a href=\"https:\/\/github.com\/gleitz\/howdoi\">howdoi<\/a> - Instant coding answers.<\/li>\n<li><a href=\"https:\/\/github.com\/asciinema\/asciinema\">asciinema<\/a> - Terminal session recorder.<\/li>\n<li><a href=\"https:\/\/github.com\/tldr-pages\/tldr\">tldr<\/a> - Simplified and community-driven man pages.<\/li>\n<li><a href=\"https:\/\/github.com\/eddieantonio\/imgcat\">imgcat<\/a> - Like <a href=\"http:\/\/www.linfo.org\/cat.html\" rel=\"nofollow\">cat<\/a> but for images.<\/li>\n<li><a href=\"https:\/\/github.com\/KittyKatt\/screenFetch\">screenfetch<\/a> - Fetches system\/theme information in terminal.<\/li>\n<li><a href=\"https:\/\/github.com\/gohugoio\/hugo\">hugo<\/a> - Fast and flexible static site generator.<\/li>\n<li><a href=\"https:\/\/github.com\/cespare\/reflex\">reflex<\/a> - Run a command when files change.<\/li>\n<li><a href=\"https:\/\/github.com\/cortesi\/modd\">modd<\/a> - Flexible tool for responding to file system changes.<\/li>\n<li><a href=\"https:\/\/github.com\/facebookincubator\/create-react-app\">create-react-app<\/a> - Create React apps with no build configuration.<\/li>\n<li><a href=\"https:\/\/github.com\/zeit\/now-cli\">now<\/a> - Real time global deployments served over HTTP\/2.<\/li>\n<li><a href=\"https:\/\/github.com\/yarnpkg\/yarn\">yarn<\/a> - Fast, reliable, and secure dependency management.<\/li>\n<li><a href=\"https:\/\/github.com\/github\/hub\">hub<\/a> - GitHub wrapper.<\/li>\n<li><a href=\"http:\/\/www.entrproject.org\/\" rel=\"nofollow\">entr<\/a> - Run arbitrary commands when files change.<\/li>\n<li><a href=\"https:\/\/github.com\/BurntSushi\/xsv\">xsv<\/a> - Fast CSV command line toolkit written in Rust.<\/li>\n<li><a href=\"https:\/\/ivarch.com\/programs\/pv.shtml\" rel=\"nofollow\">pv<\/a> - Pipe Viewer.<\/li>\n<li><a href=\"https:\/\/github.com\/rgcr\/m-cli\">m-cli<\/a> - Useful utils for macOS.<\/li>\n<li><a href=\"https:\/\/github.com\/dbcli\/pgcli\">pgcli<\/a> - Postgres CLI with autocompletion and syntax highlighting.<\/li>\n<li><a href=\"https:\/\/github.com\/mas-cli\/mas\">mas<\/a> - CLI for mac app store.<\/li>\n<li><a href=\"https:\/\/github.com\/cgag\/loc\">loc<\/a> - Count lines of code quickly.<\/li>\n<li><a href=\"https:\/\/godoc.org\/github.com\/jason0x43\/go-alfred\/alfred\" rel=\"nofollow\">alfred<\/a> - Manage Go-based Alfred workflows.<\/li>\n<li><a href=\"https:\/\/github.com\/dylanaraps\/neofetch\">neofetch<\/a> - System information tool.<\/li>\n<li><a href=\"https:\/\/github.com\/nikitavoloboev\/license-up\">license-up<\/a> - Create a license quickly for your project.<\/li>\n<li><a href=\"https:\/\/github.com\/jedisct1\/piknik\">piknik<\/a> - Copy\/paste anything over the network.<\/li>\n<li><a href=\"https:\/\/github.com\/Gabriel439\/bench\">bench<\/a> - Command-line benchmark tool.<\/li>\n<li><a href=\"https:\/\/github.com\/motemen\/ghq\">ghq<\/a> - Manage remote repository clones.<\/li>\n<li><a href=\"https:\/\/github.com\/zkat\/npx\">npx<\/a> - Execute npm package binaries.<\/li>\n<li><a href=\"https:\/\/github.com\/cortesi\/devd\">devd<\/a> - Local webserver for developers.<\/li>\n<li><a href=\"https:\/\/github.com\/rauchg\/wifi-password\">wifi-password<\/a> - Get the password of the WiFi you're on.<\/li>\n<li><a href=\"https:\/\/github.com\/sindresorhus\/fkill-cli\">fkill<\/a> - Fabulously kill processes.<\/li>\n<\/ul><h2><a id=\"user-content-preference-panes\" class=\"anchor\" aria-hidden=\"true\" href=\"#preference-panes\"><\/a>Preference Panes<\/h2>\n<ul><li><a href=\"https:\/\/gpgtools.org\/\" rel=\"nofollow\">GPG Suite<\/a> - Encrypt, decrypt, sign and verify files or messages.<\/li>\n<li><a href=\"https:\/\/github.com\/Lord-Kamina\/SwiftDefaultApps\/releases\/latest\">SwiftDefaultApps<\/a> - Set the default application used for various URL schemes, file extensions, file types, MIME types, and Uniform Type Identifiers.<\/li>\n<li><a href=\"https:\/\/www.bresink.com\/osx\/TinkerTool.html\" rel=\"nofollow\">TinkerTool<\/a> - Gives you access to additional preference settings Apple has built into macOS.<\/li>\n<\/ul><p>I also <a href=\"https:\/\/imgur.com\/a\/rlkOyUg\" rel=\"nofollow\">share screenshots of all the changes to System Preferences I have made<\/a>.<\/p>\n<h2><a id=\"user-content-desktop-screenshot\" class=\"anchor\" aria-hidden=\"true\" href=\"#desktop-screenshot\"><\/a>Desktop Screenshot<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/37a5eae47e0cfca0eafc049382486c0bc1306d2e\/68747470733a2f2f692e696d6775722e636f6d2f754b737539346f2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/37a5eae47e0cfca0eafc049382486c0bc1306d2e\/68747470733a2f2f692e696d6775722e636f6d2f754b737539346f2e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/uKsu94o.png\" \/><\/a><\/p>\n<blockquote>\n<p>Using <a href=\"https:\/\/github.com\/KittyKatt\/screenFetch\">screenfetch<\/a><\/p>\n<\/blockquote>\n<h2><a id=\"user-content-alfred-launcher\" class=\"anchor\" aria-hidden=\"true\" href=\"#alfred-launcher\"><\/a>Alfred launcher<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/e0f00acdaf2595fc2fa04e246fe63b1aa7188987\/68747470733a2f2f692e696d6775722e636f6d2f6b3268727572582e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/e0f00acdaf2595fc2fa04e246fe63b1aa7188987\/68747470733a2f2f692e696d6775722e636f6d2f6b3268727572582e706e67\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/k2hrurX.png\" \/><\/a><\/p>\n<blockquote>\n<p>Searching <a href=\"https:\/\/github.com\/nikitavoloboev\/alfred-learn-anything#readme\">Alfred Learn Anything<\/a>. Using <a href=\"https:\/\/www.alfredapp.com\/extras\/theme\/PQVZpeg4Zi\/\" rel=\"nofollow\">Monokai Night theme<\/a>.<\/p>\n<\/blockquote>\n<h2><a id=\"user-content-monokai-night-themes\" class=\"anchor\" aria-hidden=\"true\" href=\"#monokai-night-themes\"><\/a>Monokai Night themes<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/0fa0153550c4b55fa752f02d939f7a3a897ad512\/68747470733a2f2f692e696d6775722e636f6d2f50735a4d7572352e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/0fa0153550c4b55fa752f02d939f7a3a897ad512\/68747470733a2f2f692e696d6775722e636f6d2f50735a4d7572352e6a7067\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/PsZMur5.jpg\" \/><\/a><\/p>\n<blockquote>\n<p>Having a <a href=\"https:\/\/wiki.nikitavoloboev.xyz\/design\/monokai-night-themes\" rel=\"nofollow\">unified theme<\/a> across apps is amazing<\/p>\n<\/blockquote>\n<h2><a id=\"user-content-launchpad\" class=\"anchor\" aria-hidden=\"true\" href=\"#launchpad\"><\/a>Launchpad<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/7dc0258d3a7fc296112c32e2cbe3aac990b7c6a7\/68747470733a2f2f692e696d6775722e636f6d2f394b4d445966422e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/7dc0258d3a7fc296112c32e2cbe3aac990b7c6a7\/68747470733a2f2f692e696d6775722e636f6d2f394b4d445966422e6a7067\" alt=\"\" data-canonical-src=\"https:\/\/i.imgur.com\/9KMDYfB.jpg\" \/><\/a><\/p>\n<blockquote>\n<p>I add a <a href=\"https:\/\/www.imore.com\/add-space-your-mac-dock\" rel=\"nofollow\">small space in my Dock<\/a> to separate apps I always have open with apps I seldom open<\/p>\n<\/blockquote>\n<p>If you found this interesting, I also have <a href=\"https:\/\/github.com\/nikitavoloboev\/my-ios#readme\">similar repository<\/a> going over what applications I use on iOS\/WatchOS as well as how and why I use them.<\/p>\n<a href=\"https:\/\/github.com\/nikitavoloboev\/my-ios#readme\">\n    <img width=\"250\" src=\"https:\/\/camo.githubusercontent.com\/91f79d05972c5d9cb451d51db0c1754b9d8be446\/68747470733a2f2f692e696d6775722e636f6d2f547636397233332e6a7067\" data-canonical-src=\"https:\/\/i.imgur.com\/Tv69r33.jpg\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/github.com\/nikitavoloboev\/my-ios#readme\">\n    <img width=\"250\" src=\"https:\/\/camo.githubusercontent.com\/b6daabc281eca96d25680df0d494abc820162ddd\/68747470733a2f2f692e696d6775722e636f6d2f774e7a37505a522e6a7067\" data-canonical-src=\"https:\/\/i.imgur.com\/wNz7PZR.jpg\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/github.com\/nikitavoloboev\/my-ios#readme\">\n    <img width=\"250\" src=\"https:\/\/camo.githubusercontent.com\/5d2a61521486a800e3371501ebe192b32d664bf3\/68747470733a2f2f692e696d6775722e636f6d2f6a6e724b4564532e6a7067\" data-canonical-src=\"https:\/\/i.imgur.com\/jnrKEdS.jpg\" alt=\"image\" \/><\/a>\n<h2><a id=\"user-content-similar-setups\" class=\"anchor\" aria-hidden=\"true\" href=\"#similar-setups\"><\/a>Similar Setups<\/h2>\n<p>Here you can find more setups by other people that you can take ideas and inspiration from.<\/p>\n<ul><li><a href=\"https:\/\/works-for-me.github.io\/\" rel=\"nofollow\">Works for me<\/a> - Collection of developer toolkits.<\/li>\n<li><a href=\"https:\/\/usesthis.com\" rel=\"nofollow\">Use This Interviews<\/a> - What do people use to get stuff done?<\/li>\n<li><a href=\"https:\/\/github.com\/obahareth\/my-mac-os\">Omar Bahareth's my-mac-os<\/a> - Another my-mac-os.<\/li>\n<\/ul><h2><a id=\"user-content-related\" class=\"anchor\" aria-hidden=\"true\" href=\"#related\"><\/a>Related<\/h2>\n<h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>If you shared a similar personal setup to this, be it for Windows, Linux or anything else, you can add it in <a href=\"#similar-setups\">Similar Setups<\/a> section.<\/p>\n<p>I love finding new awesome tools and apps. If you have a favorite tool or app that you think I missed, please <a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/issues\/new\">say it<\/a>.<\/p>\n<h2><a id=\"user-content-thank-you\" class=\"anchor\" aria-hidden=\"true\" href=\"#thank-you\"><\/a>Thank you<\/h2>\n<p>You can support me on <a href=\"https:\/\/www.patreon.com\/nikitavoloboev\" rel=\"nofollow\">Patreon<\/a> or look into <a href=\"https:\/\/nikitavoloboev.xyz\/projects\" rel=\"nofollow\">other projects<\/a> I shared.<\/p>\n<p><a href=\"https:\/\/github.com\/nikitavoloboev\/my-mac-os\/blob\/master\/LICENSE\"><img src=\"https:\/\/camo.githubusercontent.com\/b6422cd160bd28e40c183e62a512f91fbf3d0d7f\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d3061306130612e7376673f7374796c653d666c617426636f6c6f72413d306130613061\" alt=\"MIT\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-MIT-0a0a0a.svg?style=flat&amp;colorA=0a0a0a\" \/><\/a> <a href=\"https:\/\/twitter.com\/nikitavoloboev\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/eda7d53f5d690b75efca61bb693916b7b9e8c0cb\/687474703a2f2f6269742e6c792f6e696b697477656574\" alt=\"Twitter\" data-canonical-src=\"http:\/\/bit.ly\/nikitweet\" \/><\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:54:41+0000","updated_at":"2018-11-07T16:54:41+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/6391776?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":655,"title":"The Side Project Marketing Checklist","url":"https:\/\/www.sideprojectchecklist.com\/marketing-checklist\/","content":"<p>The <a href=\"https:\/\/www.sideprojectchecklist.com\/marketing-checklist\">Side Project Marketing Checklist<\/a> is a comprehensive, chronologically ordered list of marketing tactics and ideas that you can try with your next side project. The list is free and open source so feel free to <a href=\"https:\/\/github.com\/portable-cto\/side-project-marketing\/issues\">suggest your own additions anytime<\/a>.<\/p><h2 id=\"pre-launch\">Pre-Launch<\/h2><p>This section pairs nicely with the <a href=\"https:\/\/www.sideprojectchecklist.com\/sales-checklist\">Pre-Launch section of the Sales Checklist<\/a>, so be sure to check that out too.<\/p><blockquote>\n  <p><em>I have been up against tough competition all my life. I wouldn\u2019t know how to get along without it.<\/em> - Walt Disney<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">\n    <p>Make a list of competitors (see <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/competition-tools\/\">this list of tools to research competitors<\/a>).<\/p>\n    <h5 id=\"competitor-data-to-collect\">Competitor data to collect<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Link to their website.<\/li>\n      <li class=\"task-list-item\">\u201cOne-liners\u201d and taglines competitors use.<\/li>\n      <li class=\"task-list-item\">Pricing and business model.<\/li>\n      <li class=\"task-list-item\">Blog\/RSS link.<\/li>\n      <li class=\"task-list-item\">Social media links.<\/li>\n      <li class=\"task-list-item\">Key employees on social media.<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Subscribe to RSS feeds of your competitors\u2019 blogs.<\/li>\n  <li class=\"task-list-item\">Follow your competitors and their key employees on social media.<\/li>\n  <li class=\"task-list-item\">\n    <p>Monitor competitors\u2019 websites for changes.<\/p>\n    <h5 id=\"site-monitoring-tools\">Site monitoring tools<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"http:\/\/sitealerts.com\/\">Site Alerts<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/versionista.com\/\">Versionista<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/pagedelta.com\/\">Page Delta<\/a><\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Subscribe to industry newsletters, magazines, trade journals etc.<\/li>\n  <li class=\"task-list-item\">Use <a href=\"https:\/\/www.sideprojectchecklist.com\/2018\/google-trends-research\/\">Google Trends<\/a> to do initial demand and search volume research.<\/li>\n<\/ul><blockquote>\n  <p><em>If we knew what we were doing it wouldn\u2019t be called research.<\/em> - Albert Einstein<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Find people in your contact list who might be future customers. Get their feedback on your idea.<\/li>\n  <li class=\"task-list-item\">Attend meetups or conferences for your target market.<\/li>\n  <li class=\"task-list-item\">Meet another entrepreneur who has a similar or competitive product.<\/li>\n  <li class=\"task-list-item\">\n    <p>Make a list of your competitors\u2019 customers:<\/p>\n    <h5 id=\"finding-competitors-customers\">Finding competitors\u2019 customers<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Check their Twitter followers, mentions.<\/li>\n      <li class=\"task-list-item\">Look through their Facebook fans.<\/li>\n      <li class=\"task-list-item\">Find people who follow them on Linkedin.<\/li>\n      <li class=\"task-list-item\">Check their website. Some companies list customer stories on their websites.<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Reach out to your competitors\u2019 customers, find out what they like\/don\u2019t like.<\/li>\n  <li class=\"task-list-item\">Create an \u201cearly access\u201d list for potential future customers.<\/li>\n<\/ul><h3 id=\"public-relations-preparations\">Public Relations Preparations<\/h3><ul class=\"task-list\"><li class=\"task-list-item\">Create list of tech, startup, and industry blogs.<\/li>\n  <li class=\"task-list-item\">Create list of local small business journals (eg: <a href=\"http:\/\/www.chicagobusiness.com\/section\/Contact-Us&amp;template=mobile\">Crain\u2019s Chicago<\/a>).<\/li>\n  <li class=\"task-list-item\">Create list of local bloggers and journalists in your industry.<\/li>\n  <li class=\"task-list-item\">Create a \u201cMedia Kit\u201d page (<a href=\"https:\/\/www.kickresume.com\/mediakit\/\">check out this example<\/a>).<\/li>\n<\/ul><h3 id=\"website-setup\">Website Setup<\/h3><ul class=\"task-list\"><li class=\"task-list-item\">Come up with a name and domain name.<\/li>\n  <li class=\"task-list-item\">Write a site tagline and elevator pitch.<\/li>\n  <li class=\"task-list-item\">Create a logo (see <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/logos\/\">this list of logo creation tools<\/a>).<\/li>\n  <li class=\"task-list-item\">Set up a landing page (see <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/landing-page-tools\/\">this list of landing page tools<\/a>).<\/li>\n  <li class=\"task-list-item\">Create \u201cAbout\u201d and \u201cContact\u201d pages.<\/li>\n  <li class=\"task-list-item\">\n    <p>Create Pricing page (see <a href=\"https:\/\/taprun.com\/methods\/\">this list of pricing models<\/a>):<\/p>\n    <h5 id=\"pricing-ideas\">Pricing ideas<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Create a free or trial tier for your paid product.<\/li>\n      <li class=\"task-list-item\">Offer a 100% satisfaction\/money-back guarantee.<\/li>\n      <li class=\"task-list-item\">Make product invite-only to start.<\/li>\n      <li class=\"task-list-item\">Offer free\/discounted access for early adopters\/beta testers.<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Add social media follow links to landing page.<\/li>\n  <li class=\"task-list-item\">\n    <p>Set up analytics to learn about who signs up, bounces, etc.<\/p>\n    <h5 id=\"analytics-platforms\">Analytics platforms<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"https:\/\/analytics.google.com\">Google Analytics<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/mixpanel.com\/\">Mixpanel<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.kissmetrics.com\/\">Kiss Metrics<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.google.com\/webmasters\/tools\">Webmaster Tools<\/a><\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">\n    <p>Set up error logging, monitoring, capturing, etc.<\/p>\n    <h5 id=\"error-monitoring-platforms\">Error monitoring platforms<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"https:\/\/www.bugsnag.com\/\">Bugsnag<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/rollbar.com\/\">Rollbar<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/sentry.io\/welcome\/\">Sentry<\/a><\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">\n    <p>Set up on-site messaging to capture leads and engage customers as they land on your site (see our <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/customer-chat-tools\">list of live customer chat tools<\/a>).<\/p>\n  <\/li>\n  <li class=\"task-list-item\">Test your website for common SEO, speed and security problems like broken links, missing page titles and insecure pages (see <a href=\"https:\/\/www.checkbot.io\">Checkbot<\/a>).<\/li>\n<\/ul><h3 id=\"email-setup\">Email Setup<\/h3><p><em><a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/email-marketing-tools\/\">See this list of email marketing tools<\/a> for services that make managing email easier.<\/em><\/p><blockquote>\n  <p><em>Email is the Jason Bourne of online: somebody\u2019s always trying to kill it. It can\u2019t be done.<\/em> - Unknown<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Set up email address to send and receive emails.<\/li>\n  <li class=\"task-list-item\">Set up email list and signup form.<\/li>\n  <li class=\"task-list-item\">Create a standard email template for your brand.<\/li>\n  <li class=\"task-list-item\">Create transactional emails for when users sign up\/purchase.<\/li>\n<\/ul><h3 id=\"blog-setup\">Blog Setup<\/h3><blockquote>\n  <p><em>Content Marketing is all the marketing that\u2019s left.<\/em> - Seth Godin<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Choose a blogging platform (see this list of <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/blogging-platforms\">blogging platforms for your side project<\/a>).<\/li>\n  <li class=\"task-list-item\">\n    <p>Research keywords that you\u2019d like your site\/blog to rank for.<\/p>\n    \n    <ul><li><a href=\"https:\/\/ahrefs.com\/\">Ahrefs<\/a><\/li>\n      <li><a href=\"https:\/\/www.spyfu.com\/\">SpyFu<\/a><\/li>\n      <li><a href=\"https:\/\/www.semrush.com\/\">SEMrush<\/a><\/li>\n      <li><a href=\"http:\/\/keywordtool.io\/\">KeywordTool.io<\/a><\/li>\n      <li><a href=\"https:\/\/moz.com\/explorer\">Moz Keyword Explorer<\/a><\/li>\n      <li><a href=\"https:\/\/kwfinder.com\/\">KWFinder<\/a><\/li>\n      <li><a href=\"https:\/\/adwords.google.com\/home\/tools\/keyword-planner\/\">Google Keyword Research<\/a><\/li>\n      <li><a href=\"https:\/\/neilpatel.com\/ubersuggest\/\">Ubersuggest<\/a><\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Create anchor posts or pages for keywords you\u2019d like to rank for.<\/li>\n  <li class=\"task-list-item\">Have a blog post brainstorming session (see this <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/blog-post-ideas\">big list of blog post ideas<\/a>).<\/li>\n  <li class=\"task-list-item\">Add email signup form or link to all blog posts.<\/li>\n  <li class=\"task-list-item\">Add social media follow links to all blog posts.<\/li>\n<\/ul><blockquote>\n  <p><em>Social media is the ultimate equalizer. It gives a voice and a platform to anyone willing to engage.<\/em> - Amy Jo Martin, CEO of Digital Royalty<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Search for availability of names on social networks using <a href=\"http:\/\/shareasale.com\/r.cfm?b=170332&amp;u=1653894&amp;m=21886&amp;urllink=&amp;afftrack=\">KnowEm<\/a>.<\/li>\n  <li class=\"task-list-item\">\n    <p>Choose the social media accounts you\u2019ll use.<\/p>\n    \n    <ul class=\"task-list\"><li class=\"task-list-item\">Facebook<\/li>\n      <li class=\"task-list-item\">Twitter<\/li>\n      <li class=\"task-list-item\">Linkedin<\/li>\n      <li class=\"task-list-item\">Pinterest<\/li>\n      <li class=\"task-list-item\">Instagram<\/li>\n      <li class=\"task-list-item\">Snapchat<\/li>\n      <li class=\"task-list-item\">YouTube<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Standardize profile image, background photo, links and call to action across social channels.<\/li>\n  <li class=\"task-list-item\">Put premium content behind \u201c<a href=\"https:\/\/www.paywithatweet.com\/\">pay with a tweet<\/a>\u201d.<\/li>\n<\/ul><h2 id=\"post-launch\">Post-Launch<\/h2><p>This section pairs nicely with the <a href=\"https:\/\/www.sideprojectchecklist.com\/sales-checklist\">Post-Launch section of the Sales Checklist<\/a>. Sales and marketing are really two sides of the same coin, so you\u2019ll probably need a little of both.<\/p><h3 id=\"free-promotional-channels\">Free Promotional Channels<\/h3><blockquote>\n  <p><em>I don\u2019t care how much money you have, free stuff is always a good thing.<\/em> - Queen Latifah<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Post your product on directories and review sites (Matt McCaffrey has compiled <a href=\"https:\/\/github.com\/mmccaff\/PlacesToPostYourStartup\">a great list on Github<\/a> or you can use a service like <a href=\"https:\/\/www.instaaa.com\/\">Instaaa<\/a>).<\/li>\n  <li class=\"task-list-item\">Write and distribute a Press Release (here\u2019s a <a href=\"http:\/\/www.wajeez.com\/writing-quick-press-releases\/\">comprehensive guide from Wajeez<\/a>, or you can <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/press-release-tools\/\">use one of these distribution tools<\/a>).<\/li>\n  <li class=\"task-list-item\">Write and distribute an eBook, exchange it for email signup.<\/li>\n  <li class=\"task-list-item\">Write and distribute a white paper, exchange it for email signup.<\/li>\n  <li class=\"task-list-item\">Give free access to influential bloggers in the industry.<\/li>\n  <li class=\"task-list-item\">Build a \u201cbest of\u201d page with your best blog posts that you wrote or contributed to other sites (ProBlogger calls this a \u201c<a href=\"https:\/\/www.sideprojectchecklist.com\/2018\/sneeze-page\">Sneeze Page<\/a>\u201d).<\/li>\n  <li class=\"task-list-item\">Make sure all blog posts have high quality images (see <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/free-stock-images\/\">this list of free stock image sites<\/a>).<\/li>\n  <li class=\"task-list-item\">Create an online course or guide around your product\/industry (be sure to <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/course-creation-tools\">check out this list of course management tools<\/a>).<\/li>\n  <li class=\"task-list-item\">Run an <a href=\"https:\/\/www.reddit.com\/r\/startups\/comments\/7aap5u\/influencer_marketing_summary_of_12_articles\/\">influencer marketing campaign<\/a>.<\/li>\n  <li class=\"task-list-item\">\n    <p>Run a link-building campaign.<\/p>\n    <h5 id=\"link-building-strategies\">Link building strategies<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"http:\/\/backlinko.com\/skyscraper-technique\">Skyscraper campaign<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"http:\/\/pointblankseo.com\/egobait-guide\">Ego baiting<\/a><\/li>\n      <li class=\"task-list-item\">Interview\/expert Q&amp;A post<\/li>\n      <li class=\"task-list-item\">Guest posting<\/li>\n      <li class=\"task-list-item\">Broken link building<\/li>\n      <li class=\"task-list-item\">Get interviewed on podcasts<\/li>\n      <li class=\"task-list-item\">Become a source on <a href=\"https:\/\/www.helpareporter.com\/\">Help a Reporter Out<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"http:\/\/backlinko.com\/link-building\">More strategies on Backlinko.com<\/a><\/li>\n    <\/ul><h5 id=\"link-building-tools\">Link building tools<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"https:\/\/ahrefs.com\/\">Ahrefs<\/a> - Two-week free trial<\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/majestic.com\/\">Majestic<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.ranksignals.com\/\">Rank Signals<\/a> - Free<\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.semrush.com\/\">SEMrush<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/serpstat.com\/\">Serpstat<\/a> - Free limited account<\/li>\n    <\/ul><\/li>\n<\/ul><h3 id=\"paid-promotional-channels\">Paid Promotional Channels<\/h3><blockquote>\n  <p><em>You must spend money to make money.<\/em> - Plautus, Ancient Roman playwright<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Try a few social, search, and commission-based advertising platforms (see <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/online-advertising\/\">this big list of advertising platforms if you aren\u2019t sure where to start<\/a>).<\/li>\n  <li class=\"task-list-item\">Sponsor a local meetup or conference for your target customers.<\/li>\n  <li class=\"task-list-item\">Sponsor podcasts your customers might be listening to.<\/li>\n  <li class=\"task-list-item\">Sponsor\/advertise an industry newsletter (check out <a href=\"https:\/\/newsletter.city\/\">Newsletter.city<\/a>).<\/li>\n  <li class=\"task-list-item\">\n    <p>Set up a user referral marketing system.<\/p>\n    <h5 id=\"referral-marketing-platforms\">Referral marketing platforms<\/h5>\n    <ul><li><a href=\"https:\/\/github.com\/harrystech\/prelaunchr\">Harry\u2019s Prelaunchr (free, open source)<\/a><\/li>\n      <li><a href=\"https:\/\/clevertap.com\/\">CleverTap (offers a free tier)<\/a><\/li>\n      <li><a href=\"http:\/\/expectreferrals.com\/pricing\/\">Expect Referrals<\/a><\/li>\n      <li><a href=\"https:\/\/www.hirefrederick.com\/\">Frederick<\/a><\/li>\n      <li><a href=\"http:\/\/www.geniusreferrals.com\/\">Genius Referrals<\/a><\/li>\n      <li><a href=\"https:\/\/invitebox.com\/\">InviteBox<\/a><\/li>\n      <li><a href=\"http:\/\/www.invitereferrals.com\/\">InviteReferrals<\/a><\/li>\n      <li><a href=\"https:\/\/www.referralcandy.com\/\">Referral Candy<\/a><\/li>\n      <li><a href=\"https:\/\/viral-loops.com\/\">Viral Loops<\/a><\/li>\n      <li><a href=\"https:\/\/girafi.co\/\">Girafi<\/a><\/li>\n      <li><a href=\"https:\/\/campaignedapp.com\/\">Campaigned<\/a><\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Run an engagement contest with prizes or free products for winners.<\/li>\n  <li class=\"task-list-item\">Run <a href=\"https:\/\/retargeter.com\/what-is-retargeting-and-how-does-it-work\">retargeting ads<\/a> to target users who have been to your site before.<\/li>\n<\/ul><h2 id=\"recurring\">Recurring<\/h2><h3 id=\"blogging\">Blogging<\/h3><blockquote>\n  <p><em>Blogging is like work, but without coworkers thwarting you at every turn.<\/em> - Scott Adams, Creator of Dilbert<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Build\/update publishing calendar for your blog.<\/li>\n  <li class=\"task-list-item\">Regularly post blog posts on your blog(s).<\/li>\n  <li class=\"task-list-item\">Solicit guest posts from early customers and fans of your product.<\/li>\n  <li class=\"task-list-item\">\n    <p>Repurpose existing blog posts:<\/p>\n    <h5 id=\"repurposing-blog-posts\">Repurposing blog posts<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Record\/post video of you reading the post on YouTube.<\/li>\n      <li class=\"task-list-item\">Turn posts into a podcast.<\/li>\n      <li class=\"task-list-item\">Create an infographic based on the post.<\/li>\n      <li class=\"task-list-item\">Create a presentation of your post (see the <a href=\"#presentation-software\">list of presentation software below<\/a>).<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">\n    <p>Promote your blog content:<\/p>\n    <h5 id=\"blog-promotion-techniques\">Blog promotion techniques<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Send post to your email list.<\/li>\n      <li class=\"task-list-item\">Promote on your social media.<\/li>\n      <li class=\"task-list-item\">Email friends and relatives, ask them to share if relevant.<\/li>\n      <li class=\"task-list-item\">Send to other bloggers for feedback, ask to share if they like it.<\/li>\n      <li class=\"task-list-item\">Add your latest blog post or landing page to your email signature.<\/li>\n    <\/ul><\/li>\n<\/ul><h3 id=\"email\">Email<\/h3><blockquote>\n  <p><em>Any email that contains the words \u2018important\u2019 or \u2018urgent\u2019 never are, and annoy me to the point of not replying out of principle.<\/em> - Markus Persson, aka \u201cNotch\u201d, creator of Minecraft<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Send a regular email newsletter with blog posts, use cases, customer stories, etc.<\/li>\n  <li class=\"task-list-item\">Promote email list on social media.<\/li>\n  <li class=\"task-list-item\">Send 20 cold emails per week to connect with early customers and get direct feedback.<\/li>\n  <li class=\"task-list-item\">Send new users a personal email introducing yourself.<\/li>\n<\/ul><blockquote>\n  <p><em>We have technology, finally, that for the first time in human history allows people to really maintain rich connections with much larger numbers of people.<\/em> \u2013 Pierre Omidyar, Founder of eBay<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">\n    <p>Schedule regular social media content (try these <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/social-media-tools\">social media tools to help automate the social media process<\/a>).<\/p>\n    <h5 id=\"content-ideas\">Content ideas<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Show what happening \u201cbehind the scenes\u201d (eg: pictures of your workspace, in-development features, etc.).<\/li>\n      <li class=\"task-list-item\">Blog posts from your blog (or partners\u2019 blogs).<\/li>\n      <li class=\"task-list-item\">Special offers or discounts.<\/li>\n      <li class=\"task-list-item\">Relevant news, advice, or blogs.<\/li>\n      <li class=\"task-list-item\">Funny or inspiring quotes.<\/li>\n      <li class=\"task-list-item\">Polls or questions for your followers.<\/li>\n      <li class=\"task-list-item\">Suggested books.<\/li>\n      <li class=\"task-list-item\">Useful tools or websites.<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Join Facebook and Linkedin groups where your product might be beneficial.<\/li>\n  <li class=\"task-list-item\">Send exclusive offers to LinkedIn\/Facebook Group owners.<\/li>\n  <li class=\"task-list-item\">Join Twitter chats related to your industry\/product.<\/li>\n  <li class=\"task-list-item\">Give early users discounts for taking pics with your product.<\/li>\n<\/ul><h3 id=\"public-relations\">Public Relations<\/h3><blockquote>\n  <p><em>The art of publicity is a black art; but it has come to stay, and every year adds to its potency.<\/em> \u2013 Thomas Paine<\/p>\n<\/blockquote><ul class=\"task-list\"><li class=\"task-list-item\">Ask bloggers with list articles to add your site to their content.<\/li>\n  <li class=\"task-list-item\">Reach out to small business journals, reporters, bloggers. Inform and ask for coverage.<\/li>\n  <li class=\"task-list-item\">Reach out to podcasters, try to get featured on their show.<\/li>\n  <li class=\"task-list-item\">Find professors and students with <code class=\"highlighter-rouge\">.edu<\/code> sites to link to your content (good for SEO).<\/li>\n  <li class=\"task-list-item\">Find reviews or lists of similar products. Ask to be added or leave a comment about your product.<\/li>\n  <li class=\"task-list-item\">\n    <p>Guest post on other blogs.<\/p>\n    <h5 id=\"guest-posting\">Guest posting<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\">Build a list of blogs that are a good fit for your product and accept guest posts. Save their contact form\/information.<\/li>\n      <li class=\"task-list-item\">Write a few posts on your own blog first (to use as a demonstration).<\/li>\n      <li class=\"task-list-item\">Create a list of \u201cpitches,\u201d blog post ideas with a title and one paragraph summary that might grab bloggers.<\/li>\n      <li class=\"task-list-item\">Pitch the blogs one idea each. See what they respond to.<\/li>\n      <li class=\"task-list-item\">Help promote your posts via social media, email lists, etc.<\/li>\n      <li class=\"task-list-item\">Reach back out in a month and try another pitch. Try to become a \u201cregular\u201d<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Leave non-spammy comments on blog posts related to your industry or product.<\/li>\n  <li class=\"task-list-item\">Join and participate in forums related to your product or industry.<\/li>\n<\/ul><h3 id=\"external-sites\">External Sites<\/h3><ul class=\"task-list\"><li class=\"task-list-item\">\n    <p>Answer relevant questions on Q&amp;A community sites (check back regularly)<\/p>\n    <h5 id=\"qa-communities\">Q&amp;A Communities<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"https:\/\/stackoverflow.com\/\">Stack Overflow<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.quora.com\/\">Quora<\/a><\/li>\n      <li class=\"task-list-item\">Industry or geographically specific forums<\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.producthunt.com\/\">Product Hunt<\/a> For more information on how to launch here, <a href=\"https:\/\/blog.producthunt.com\/how-to-launch-on-product-hunt-7c1843e06399\">read this<\/a>.<\/li>\n    <\/ul><\/li>\n  <li class=\"task-list-item\">Leave comments on Slideshare presentations.<\/li>\n  <li class=\"task-list-item\">Review industry-related books on Amazon, Goodreads.<\/li>\n  <li class=\"task-list-item\">Register to be <a href=\"https:\/\/www.sideprojectchecklist.com\/2018\/f5bot-tracking\">notified of topics on Reddit and Hacker News with F5Bot<\/a>.<\/li>\n  <li class=\"task-list-item\">\n    <p>Create\/publish presentations relevant to your industry.<\/p>\n    <h5 id=\"presentation-software\">Presentation Software<\/h5>\n    <ul class=\"task-list\"><li class=\"task-list-item\"><a href=\"https:\/\/docs.google.com\/presentation\/\">Google Slides<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/prezi.com\/\">Prezi<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/slides.com\/\">Slides.com<\/a><\/li>\n      <li class=\"task-list-item\"><a href=\"https:\/\/www.slideshare.net\/\">Slideshare<\/a><\/li>\n    <\/ul><\/li>\n<\/ul><h2 id=\"optimizations\">Optimizations<\/h2><ul class=\"task-list\"><li class=\"task-list-item\">Run a customer poll (can also generate content for your blog or social media channels).<\/li>\n  <li class=\"task-list-item\">Create another side project to promote your product (<a href=\"https:\/\/medium.com\/swlh\/side-product-marketing-is-the-new-king-a75c4ed0c0c5\">read more<\/a>).<\/li>\n  <li class=\"task-list-item\">A\/B test your landing\/payment pages (check out this <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/what-is-ab-testing\/\">introduction to A\/B testing<\/a>).<\/li>\n  <li class=\"task-list-item\">A\/B test email newsletters and promotions.<\/li>\n  <li class=\"task-list-item\">Implement <a href=\"https:\/\/dev.twitter.com\/cards\/overview\">Twitter cards<\/a> on your blog posts.<\/li>\n  <li class=\"task-list-item\">Implement <a href=\"https:\/\/developers.google.com\/search\/docs\/guides\/mark-up-content\">rich snippets in Google search results<\/a>.<\/li>\n  <li class=\"task-list-item\">Analyze user signup flow (<a href=\"http:\/\/www.useronboard.com\/\">check out the teardowns here<\/a>).<\/li>\n  <li class=\"task-list-item\">Test your website on multiple platforms, make sure speed is good.<\/li>\n  <li class=\"task-list-item\">Use <a href=\"https:\/\/website.grader.com\/\">Website Grader<\/a> to pinpoint website improvements.<\/li>\n  <li class=\"task-list-item\">Create and track weekly traffic and growth goals.<\/li>\n  <li class=\"task-list-item\">Time social media posts and email newsletters to when your audience is most likely to respond.<\/li>\n  <li class=\"task-list-item\">Make sure each page on your site has a clear call-to-action.<\/li>\n  <li class=\"task-list-item\">Audit and improve your conversion rate (see <a href=\"http:\/\/www.conversionchecklist.org\/\">this checklist for detailed steps you can take<\/a>)<\/li>\n  <li class=\"task-list-item\">Set up automatic analytics reports to be emailed to you each week.<\/li>\n  <li class=\"task-list-item\">Experiment with various signup form locations, colors, and sizes.<\/li>\n  <li class=\"task-list-item\">Add \u201cExit Intent\u201d popup to your blog\/site.<\/li>\n  <li class=\"task-list-item\">Create an FAQs page.<\/li>\n  <li class=\"task-list-item\">Verify your website\u2019s tags are correct with <a href=\"https:\/\/get.google.com\/tagassistant\/\">Google Tag Assistant<\/a>.<\/li>\n  <li class=\"task-list-item\">Watch a first time user use your product.<\/li>\n  <li class=\"task-list-item\">Collect and display testimonials from happy users.<\/li>\n  <li class=\"task-list-item\">Survey or poll your users for feedback (see our <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/survey-tools\">list of survey tools for lots of great options at any budget<\/a>).<\/li>\n  <li class=\"task-list-item\">Use <a href=\"https:\/\/monkeytest.it\/\">Monkey Test It<\/a> to find bugs and broken links of your site. Set up an alert to notify you on breakage.<\/li>\n  <li class=\"task-list-item\">Use <a href=\"https:\/\/www.checkbot.io\">Checkbot<\/a> on your website to test for and fix SEO, speed and security problems.<\/li>\n<\/ul>","created_at":"2018-11-07T16:54:40+0000","updated_at":"2018-11-07T16:54:40+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"www.sideprojectchecklist.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":654,"title":"How Do I Make Money Using the Product Advertising API?","url":"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/HowDoIMakeMoneyUsingA2S.html","content":"<div id=\"main-col-body\"><table summary=\"Breadcrumbs\"><tr><td>\n                              <p><a href=\"https:\/\/docs.aws.amazon.com\/index.html\">AWS Documentation<\/a> \u00bb <a href=\"http:\/\/aws.amazon.com\/archives\/Product%20Advertising%20API\">Amazon Product Advertising API Docs<\/a> \u00bb <a href=\"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/index.html\">Developer Guide<\/a> \u00bb <a href=\"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/ProgrammingGuide.html\">Programming Guide<\/a> \u00bb <a href=\"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/CHAP_A2STerminologyandBasicConcepts.html\">Product Advertising API Terminology and Basic Concepts<\/a> \u00bb How Do I Make Money Using the Product Advertising API?<\/p>\n                           <\/td>\n                        <\/tr><\/table><p>Use the Product Advertising API to advertise Amazon products with Amazon Associates\n                        and earn\n                        referral fees when customers buy qualifying products.\n                     <\/p><p>You need to register for an Amazon Associates account and ensure that you include\n                        your Associate Tag in all requests. Product Advertising API returns URLs that have\n                        your Associate Tag. When a customer visits an \n                        Amazon marketplace through a tagged link and buys a product, you earn referral fees.\n                        For more information, see <a href=\"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/becomingAssociate.html\">Sign up as an Amazon Associate<\/a>.\n                     <\/p><\/div><div id=\"main-col-footer\"><p><a target=\"_top\" href=\"https:\/\/docs.aws.amazon.com\/general\/latest\/gr\/docconventions.html\">Document Conventions<\/a><\/p><p><a class=\"awstoc\" accesskey=\"p\" href=\"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/WhatIsA2S.html\">\u00ab Previous <\/a><a class=\"awstoc\" accesskey=\"n\" href=\"https:\/\/docs.aws.amazon.com\/AWSECommerceService\/latest\/DG\/ItemsforSale.html\">Next \u00bb<\/a><\/p><p>\u00a9 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.<\/p><\/div>","created_at":"2018-11-07T16:54:25+0000","updated_at":"2018-11-07T16:54:25+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"docs.aws.amazon.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":651,"title":"I Used to Get Distracted by Email. So I Built a Tool to Stay Focused. - Indie Hackers","url":"https:\/\/www.indiehackers.com\/interview\/i-used-to-get-distracted-by-email-so-i-built-a-tool-to-stay-focused-c0a3d64f27","content":"<h2>Hello! What's your background, and what are you working on?<\/h2><p>Hi! I'm Peter Hartree, and I'm the developer of <a href=\"https:\/\/inboxwhenready.org\">Inbox When Ready<\/a>, a Chrome extension for Gmail which helps people cultivate a better email workflow.<\/p><p>I made my first website when I was 12 and worked as a freelance web developer during my teens. I gained a bachelor's degree in philosophy, then worked as a freelance contractor for a few years. I now work (nearly) full time in a remote role at <a href=\"https:\/\/80000hours.org\">80,000 Hours<\/a>, a nonprofit which helps people have a big positive social impact with their careers.<\/p><p>Inbox When Ready started out as a \"scratch my own itch\" weekend project. Two years on, the extension has 2,000 weekly active users and revenue of $1.2K\/month. People who find the extension a good fit reclaim 30-60 minutes of focused work each week they use it.<\/p><h2>What motivated you to get started with Inbox When Ready?<\/h2><p>During my undergrad I read about the <a href=\"http:\/\/consc.net\/papers\/extended.html\">extended mind thesis<\/a>, which encourages us to think of computers as extensions of our biological minds. Since then, I've thought quite a bit about \"<a href=\"https:\/\/medium.com\/@peterhartree\/extended-mind-design-f8388056f77d\">extended mind design<\/a>\" \u2014 I think this is a useful frame for thinking about many techniques and tools commonly tagged as \"productivity advice\", \"lifehacking\", or \"metaskills\".<\/p><p>A key function of your extended mind is to help you allocate your attention effectively. Our daily attention environment is surprisingly hostile by default, because a lot of actors have <a href=\"https:\/\/www.1843magazine.com\/features\/the-scientists-who-make-apps-addictive\">strong incentives<\/a> to <a href=\"https:\/\/soundcloud.com\/the_rsa\/are-digital-technologies-making-politics-impossible\">misdirect our attention<\/a>. Recognizing this, I began observing how my attention shifts during the day and taking measures to make my attention environment more amenable to what Cal Newport calls \"<a href=\"http:\/\/calnewport.com\/books\/deep-work\/\">Deep Work<\/a>\".<\/p><p>One morning I was doing some programming and I needed to ask my client a question. I opened Gmail with the intention of emailing the client, but my attention was immediately derailed by some new messages in my inbox. When I remembered my original intention (perhaps 30 minutes later), I realized that Gmail would be much less of an attentional liability for me <em>if only my inbox were hidden by default<\/em>.<\/p><p>I hacked a Chrome extension to implement this functionality the following day. I used it for a few weeks and found it was a big win for my focus. I went from seeing my inbox 10-20 times a day unintentionally to just once or twice intentionally. I published the extension on the <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/inbox-when-ready-for-gmai\/cdedhgmbfjhobfnphaoihdfmnjidcpim?hl=en\">Chrome Web Store<\/a>, shared it with a couple of friends, then went back to my day job. That was October 2015.<\/p><h2>What went into building the initial product?<\/h2><p>In the months that followed I began receiving a trickle of thank you's, feature requests, and bug reports. In spring 2016 I realised that the extension had accumulated ~25 weekly active users and a handful of positive reviews from complete strangers.<\/p><p>So I decided to block out 10 days in the summer to work on the extension some more. My plan was to fix some bugs, setup basic analytics, make a marketing page, and do some user research. Fundamentally, I wanted to get a better sense of whether this could be something that a lot of people would want.<\/p><p>Over the summer I emailed and Skyped with several dozen people who said they loved the extension. The early analytics data looked promising \u2014 about 30% of people who tried the extension were still using it after two weeks, and a lot of users were keeping their inboxes hidden for several hours each day. This made me think I should keep working on the project, so I blocked out another 30 days during the autumn to focus on product development (75%) and user acquisition (25%).<\/p><a class=\"screenshot-wrapper\" href=\"https:\/\/inboxwhenready.org\">\n  <img alt=\"Inbox When Ready Screencast\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/content\/inbox-when-ready__screencast.gif\" \/><\/a><p>Up to this point, the extension simply made your Gmail inbox hidden by default and added a show\/hide button to the Gmail UI.<\/p><p>After talking to a bunch of users, I came to realise that my initial vision for the extension \u2014 protect your focus, hide your inbox by default \u2014 was too narrow. I realised that a lot of people have horrible inbox workflows, and could benefit from an interface that actively tried to help them cultivate something better. So my vision for the extension became \"protect your focus, cultivate a better email workflow\", and I started experimenting with a handful of features to address this user need.<\/p><p>Feature development at this stage was guided by a mix of qualitative feedback and my general reading and intuition. I realised, though, that it was going to be hard to decide which new features were \"keepers\" without detailed quantitative usage data.<\/p><p>So I invested several day's work in setting up usage analytics and feature flag infrastructure that would let me A\/B test new features. I was hoping to conduct epistemically robust experiments to gauge the effect of new features on key behaviors like the number of times people clicked \"Show Inbox\" each day, and the amount of time they left their inboxes visible.<\/p><p>This investment turned out to be somewhat premature \u2014 my userbase was too small to power many of the tests I wanted to run. So most of my decisions during this period were guided by qualitative feedback and less robust quantitative measures such as \"what fraction of users enable and keep using this feature\".<\/p><h2>How have you attracted users and grown Inbox When Ready?<\/h2><p>I got most of my first 100 users via manual outreach. I made a list of several hundred people who I thought would be interested in trying <a href=\"https:\/\/inboxwhenready.org\">Inbox When Ready<\/a> and contacted them individually, mostly by email. To make the list I:<\/p><ul><li>thought about all my friends, colleagues and acquaintances who might be interested<\/li>\n<li>searched the Chrome Web Store, Twitter, and Google for people who had said lukewarm or negative things about an extension which addresses a similar user problem in a different way<\/li>\n<li>searched the same places for people who were clearly Gmail users interested in email productivity.<\/li>\n<\/ul><p>I wrote a short email with a clear subject line. I sent the emails via <a href=\"https:\/\/mailgun.com\">Mailgun's free SMTP service<\/a>, which makes it easy to track open rates and spam complaints (I received zero spam complaints, which I took to mean I was emailing the right kind of people in a reasonable way). About 20% of recipients gave the extension a try. Crucially, a decent fraction of these people kept using the extension and began endorsing it on social media and the Chrome Web Store.<\/p><p>After passing the 100 weekly active users mark I kept going with manual outreach but shifted my focus to \"supernodes\". This led to endorsements from people like Tristan Harris, Ezra Klein, and Eric Barker, a trickle of press coverage, and a feature on <a href=\"https:\/\/www.producthunt.com\/posts\/inbox-when-ready\">Product Hunt<\/a>.<\/p><p>I also added a few features to the extension to encourage people to tell their friends: a dialog that shows after you hide your inbox 20 times, and an automatic email that comes after you load the extension 50 times.<\/p><p>Here's a look at Inbox When Ready's usage data over time:<\/p><p><\/p><table><thead><tr><td>Month<\/td>\n        <td>Active Users<\/td>\n      <\/tr><\/thead><tbody><tr><td data-full-value=\"June 1, 2016\">Jun<\/td>\n        <td>0<\/td>\n      <\/tr><tr><td data-full-value=\"July 1, 2016\">Jul<\/td>\n        <td>26<\/td>\n      <\/tr><tr><td data-full-value=\"August 1, 2016\">Aug<\/td>\n        <td>22<\/td>\n      <\/tr><tr><td data-full-value=\"September 1, 2016\">Sep<\/td>\n        <td>225<\/td>\n      <\/tr><tr><td data-full-value=\"October 1, 2016\">Oct<\/td>\n        <td>404<\/td>\n      <\/tr><tr><td data-full-value=\"November 1, 2016\">Nov<\/td>\n        <td>810<\/td>\n      <\/tr><tr><td data-full-value=\"December 1, 2016\">Dec<\/td>\n        <td>844<\/td>\n      <\/tr><tr><td data-full-value=\"January 1, 2017\">Jan<\/td>\n        <td>788<\/td>\n      <\/tr><tr><td data-full-value=\"February 1, 2017\">Feb<\/td>\n        <td>1206<\/td>\n      <\/tr><tr><td data-full-value=\"March 1, 2017\">Mar<\/td>\n        <td>1402<\/td>\n      <\/tr><tr><td data-full-value=\"April 1, 2017\">Apr<\/td>\n        <td>1519<\/td>\n      <\/tr><tr><td data-full-value=\"May 1, 2017\">May<\/td>\n        <td>1644<\/td>\n      <\/tr><tr><td data-full-value=\"June 1, 2017\">Jun<\/td>\n        <td>1760<\/td>\n      <\/tr><tr><td data-full-value=\"July 1, 2017\">Jul<\/td>\n        <td>1744<\/td>\n      <\/tr><tr><td data-full-value=\"August 1, 2017\">Aug<\/td>\n        <td>1850<\/td>\n      <\/tr><tr><td data-full-value=\"September 1, 2017\">Sep<\/td>\n        <td>1959<\/td>\n      <\/tr><tr><td data-full-value=\"October 1, 2017\">Oct<\/td>\n        <td>2078<\/td>\n      <\/tr><\/tbody><\/table><h2>What's your business model, and how have you grown your revenue?<\/h2><p><a href=\"https:\/\/inboxwhenready.org\">Inbox When Ready<\/a> hit 1,000 weekly active users in winter 2016. By this point, I was confident that a decent fraction of these users were getting some real value from the extension \u2014 on the order of 30-60 minutes of focused work reclaimed every week. I also felt like the product could become 2-5x better with another 6 months of full time work, and potentially help tens of thousands of users realise wins of this magnitude. So it was time to think about my long-term commitment to the project, and with that, monetisation.<\/p><p>As a first step, I made a Patreon page and emailed it to a few hundred of my most engaged users. A handful of users kindly pledged their support, but it was clear within a fortnight that a donation-based model was not going to be sustainable.<\/p><p>I looked at how other popular Gmail extensions raise revenue, and decided to go for a simple freemium model. The free edition of the extension adds an \"I'm using an Inbox When Ready\" signature to outgoing email \u2014 if you don't like that, you can pay $36\/year for the PRO edition. Everyone gets a 14-day free trial of the PRO edition when they sign up.<\/p><p>I don't love this model \u2014 the signature, as a growth hack, doesn't sit comfortably with my views on the ethics of attention (it strikes me as a moderately objectionable form of attention pollution, since it relies on interrupting the focus of email recipients).<\/p><p>On the plus side, the signature lets me provide all features to all users, rather than making a subset of features \"paid only\". It also serves as a fairly compelling \"please upgrade\" stick for professional users, which is the subset of users I feel okay about asking to pay. Overall the signature has been working quite well, but I hope to figure out something better soon. In the meantime, I invite people who can't pay to claim a <a href=\"https:\/\/inboxwhenready.org\/upgrade\/#concessions\">free PRO license<\/a>.<\/p><p>Monthly revenue has been fairly stable since launch. The monthly variation is mostly explained by variations in the number of people who visit the Inbox When Ready website or the Chrome Web Store listing, which itself spikes from time to time due to press coverage or influencer mentions.<\/p><div class=\"interactive-chart\" data-description=\"March through September 2017\" data-hover-prefix=\"$\" data-hover-suffix=\"MR\" data-title=\"Monthly Revenue\" data-y-axis-prefix=\"$\"><table><thead><tr><td>Month<\/td>\n        <td>Revenue<\/td>\n      <\/tr><\/thead><tbody><tr><td data-full-value=\"March 1, 2017\">March<\/td>\n        <td>108<\/td>\n      <\/tr><tr><td data-full-value=\"April 1, 2017\">April<\/td>\n        <td>918<\/td>\n      <\/tr><tr><td data-full-value=\"May 1, 2017\">May<\/td>\n        <td>1514<\/td>\n      <\/tr><tr><td data-full-value=\"June 1, 2017\">June<\/td>\n        <td>1152<\/td>\n      <\/tr><tr><td data-full-value=\"July 1, 2017\">July<\/td>\n        <td>972<\/td>\n      <\/tr><tr><td data-full-value=\"August 1, 2017\">August<\/td>\n        <td>1908<\/td>\n      <\/tr><tr><td data-full-value=\"September 1, 2017\">September<\/td>\n        <td>1008<\/td>\n      <\/tr><\/tbody><\/table><p>Note: March was a partial month.<\/p><\/div><p>I track all the time I spend working on <a href=\"https:\/\/inboxwhenready.org\">Inbox When Ready<\/a>, so I can tell you that as of October 1, 2017, I've spent 568.5 hours working on the project, and my all-time effective wage has been about $9\/hour.<\/p><p>For my personal finances, the trend in my marginal hourly wage is what matters. It's heading in a good direction; in the past 3 months it's grown from ~$50 to ~$60\/hour.<\/p><p>From a narrowly financial perspective, it'll take a while to recoup the initial time investment, such that my all time effective wage looks similar to what I could have earned by just doing freelance work. But this project was never mostly about the money, and overall I am already delighted by and grateful for the returns I've received on my time, in the form of learning, friends, and fun.<\/p><h2>What are your goals for the future?<\/h2><p>This time next year, <a href=\"https:\/\/inboxwhenready.org\">Inbox When Ready<\/a> should be at least 2x better and have at least 10,000 weekly active users. My dream outcome for this project would be to usefully influence the designers of Gmail and other major email clients (which between them have over a billion monthly active users!).<\/p><p>In comparison to that possibility, my personal goals seem fairly trivial. But they include: keep learning (about product development and user acquisition in particular) and have fun helping strangers (I get thank you emails every week, which is really satisfying). The extra income is nice too, and could be fairly life changing if things go well.<\/p><p>The biggest challenge is that I'm severely time constrained. I can work on this project about 3 days\/month, and this won't change anytime soon, since I'm strongly committed to my day job at 80,000 Hours.<\/p><p>My default plan is just to push forward on product development and user growth in the time I have available. My progress will be slower than I'd like, but that's okay.<\/p><p>I'm actively looking for potential collaborators \u2014 I'd love to hire or partner with someone who can work on product development and\/or marketing and growth. I've also had a few acquisition enquiries, and I would consider selling if I found someone that I trusted to take the project in a good direction.<\/p><h2>What are the biggest challenges you've faced and obstacles you've overcome?<\/h2><p>Time constraints are my biggest persistent challenge. I try to stay focused on the highest value activities and say \"no\" to everything else. In general I prefer automation to outsourcing.<\/p><p>I've put quite a lot of work into keeping the customer support burden manageable, mainly by addressing the root causes of common queries and optimising the handling of enquiries with templates and scripts. I usually spend fewer than 30 minutes a week on customer support, and I'm dragging my feet on outsourcing because it turns out I love doing it, and people seem to appreciate the fact that they're getting support directly from the maker.<\/p><p>In the last six months, user growth has been a bit slower than I would like. There's a lot I can do on the product development side to address this \u2014 onboarding could be 2x better, and I've a long list of ideas that could make the extension more useful for more people and thereby improve long term retention.<\/p><p>At the top of the funnel, I've yet to find a repeatable activity of the form \"do X and Y new people will try the extension\". If I could figure this out, I'd be able to invest more heavily in product development (e.g. hire someone) with greater confidence.<\/p><h2>If you had to start over, what would you do differently?<\/h2><p>If I were doing this again, I'd start charging soon after passing the 100 weekly active users milestone.<\/p><p>It's just such useful feedback that people are willing to pay, and paid users, once invested, often take the time to provide especially valuable written feedback. I'm pretty sure the product would be better, and have many more users, if I'd started charging sooner.<\/p><h2>Have you found anything particularly helpful or advantageous?<\/h2><p>I went through Y Combinator with 80,000 Hours in 2015, and this experience helped a lot. A lot of YC's <a href=\"https:\/\/blog.ycombinator.com\/ycs-essential-startup-advice\/\">basic advice<\/a> is just extremely valuable, and the feedback you get during the program is a kind of training which helps internalize it. For people just starting out, \"launch before you're ready\", \"find people that love your product\" and \"you can always do less\" strike me as especially valuable YC maxims.<\/p><p>In general, staying focused on the highest value activities is very hard and very important. My own process for this is to make daily, weekly, monthly and quarterly plans, and at each level of planning to search the possibility space and then do <a href=\"http:\/\/lesswrong.com\/lw\/h5e\/fermi_estimates\/\">Fermi estimates<\/a> on the costs and benefits of all the things I might work on that seem promising. These estimates often reveal big differences in expected value that are hard to see intuitively.<\/p><p>These books helped me a lot, and I would recommend them to all Indie Hackers:<\/p><ul><li><a href=\"https:\/\/leanstack.com\/running-lean-book\/\">Running Lean<\/a> by Ash Maurya \u2014 for his emphasis on speed, learning and focus.<\/li>\n<li><a href=\"https:\/\/strategyzer.com\/books\/value-proposition-design\">Value Proposition Design<\/a> by Alex Osterwalder \u2014 for the advice on user research and ideation.<\/li>\n<li><a href=\"https:\/\/www.amazon.com\/Traction-Startup-Achieve-Explosive-Customer\/dp\/1591848369\">Traction<\/a> by Gabriel Weinberg \u2014 if you want to start a business, it's worth thinking about distribution before you start making.<\/li>\n<li><a href=\"https:\/\/www.amazon.com\/Getting-Results-Agile-Way-Personal\/dp\/0984548203\">Getting Results the Agile Way<\/a> by J.D. Meier \u2014 I've read embarrassingly many books on personal productivity. This is my favourite.<\/li>\n<\/ul><h2>What's your advice for indie hackers who are just starting out?<\/h2><p>If you're still searching for a problem to work on, Spencer Greenberg has a <a href=\"https:\/\/www.youtube.com\/watch?v=LiVyRtS_d9o\">great talk on deliberate innovation<\/a>. Among other things he points out that you'll be better placed to generate new and valuable ideas if you cultivate a rare conjunction of skills and domain expertise.<\/p><p>In general I'd love to see more indie hackers think about how their work could help address some of the world's <a href=\"https:\/\/80000hours.org\/articles\/cause-selection\/\">most pressing problems<\/a>. Even if you're not well placed to work on these problems directly, you may be able to use your skills to make people who are working in these areas much more effective.<\/p><h2>Where can we go to learn more?<\/h2><p>You can find out more and try the extension at the <a href=\"https:\/\/inboxwhenready.org\">Inbox When Ready website<\/a>.<\/p><p>I'm on <a href=\"https:\/\/twitter.com\/peterhartree\">Twitter<\/a>, <a href=\"https:\/\/medium.com\/@peterhartree\">Medium<\/a>, and my personal homepage is <a href=\"https:\/\/www.peterhartree.co.uk\">here<\/a>.<\/p><p>I'd be glad to answer questions in the comments. :)<\/p><p>\n  \u2014\n  <a id=\"ember32257757\" href=\"https:\/\/www.indiehackers.com\/peterhartree?id=H2opxMMmDyU9pUNz8c8S18cTQiT2\" class=\"user-link__link ember-view\">    <img class=\"user-avatar__img\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/avatars\/72x72_H2opxMMmDyU9pUNz8c8S18cTQiT2\" alt=\"image\" \/>\n      \n          Peter Hartree\n      \n\n    \n<\/a>\n, Creator of Inbox When Ready\n<\/p><h2>Want to build your own business like Inbox When Ready?<\/h2><p>\n  You should join the <a id=\"ember32257759\" href=\"https:\/\/www.indiehackers.com\/\" class=\"ember-view\">Indie Hackers community<\/a>! \ud83e\udd17\n<\/p><p>\n  We're a few thousand founders helping each other build profitable businesses\n  and side projects. Come share what you're working on and get feedback from\n  your peers.\n<\/p><p>\n  Not ready to get started on your product yet? No problem. The community is a\n  great place to meet people, learn, and get your feet wet. Feel free to\n  <a id=\"ember32257760\" href=\"https:\/\/www.indiehackers.com\/\" class=\"ember-view\">just browse<\/a>!\n<\/p><p>\n  \u2014\n  <a id=\"ember32257762\" href=\"https:\/\/www.indiehackers.com\/csallen?id=ibTLPyjwVebnZjMGKvz6ztarnuV2\" class=\"user-link__link ember-view\">    <img class=\"user-avatar__img\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/avatars\/72x72_ibTLPyjwVebnZjMGKvz6ztarnuV2\" alt=\"image\" \/>\n      \n          Courtland Allen\n      \n\n    \n<\/a>\n, Indie Hackers founder\n<\/p>","created_at":"2018-11-07T16:54:24+0000","updated_at":"2018-11-07T16:54:24+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":13,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/580157013239-us-east-1-chromeless.s3.amazonaws.com\/interview-shareables\/c0a3d64f27.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":652,"title":"Hack Your Own Band With Clojure and Overtone","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=qUU7WeUiR5o","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/qUU7WeUiR5o?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:54:24+0000","updated_at":"2018-11-07T16:54:24+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/qUU7WeUiR5o\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":653,"title":"Calling clojure from java","url":"https:\/\/stackoverflow.com\/questions\/2181774\/calling-clojure-from-java","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        155&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>Most of the top google hits for \"calling clojure from java\" are outdated and recommend using <code>clojure.lang.RT<\/code> to compile the source code. Could you help with a clear explanation of how to call Clojure from Java assuming you have already built a jar from the Clojure project and included it in the classpath?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2181774\/revisions\" title=\"show all edits to this post\">edited May 31 '15 at 5:11<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/843660\/dskrvk\">dskrvk<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            771718&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Feb 2 '10 at 3:51&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/90801\/arthur-ulfeldt\">Arthur Ulfeldt<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            72.5k19163249&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"9\">&#13;\n                                9 Answers&#13;\n                                9&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        149&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>Update<\/strong>: Since this answer was posted, some of the tools available have changed. After the original answer, there is an update including information on how to build the example with current tools.<\/p>\n<p>It isn't quite as simple as compiling to a jar and calling the internal methods. There do seem to be a few tricks to make it all work though. Here's an example of a simple Clojure file that can be compiled to a jar:<\/p>\n<pre>(ns com.domain.tiny\n  (:gen-class\n    :name com.domain.tiny\n    :methods [#^{:static true} [binomial [int int] double]]))\n(defn binomial\n  \"Calculate the binomial coefficient.\"\n  [n k]\n  (let [a (inc n)]\n    (loop [b 1\n           c 1]\n      (if (&gt; b k)\n        c\n        (recur (inc b) (* (\/ (- a b) b) c))))))\n(defn -binomial\n  \"A Java-callable wrapper around the 'binomial' function.\"\n  [n k]\n  (binomial n k))\n(defn -main []\n  (println (str \"(binomial 5 3): \" (binomial 5 3)))\n  (println (str \"(binomial 10042 111): \" (binomial 10042 111)))\n)\n<\/pre>\n<p>If you run it, you should see something like:<\/p>\n<pre>(binomial 5 3): 10\n(binomial 10042 111): 49068389575068144946633777...\n<\/pre>\n<p>And here's a Java program that calls the <code>-binomial<\/code> function in the <code>tiny.jar<\/code>.<\/p>\n<pre>import com.domain.tiny;\npublic class Main {\n    public static void main(String[] args) {\n        System.out.println(\"(binomial 5 3): \" + tiny.binomial(5, 3));\n        System.out.println(\"(binomial 10042, 111): \" + tiny.binomial(10042, 111));\n    }\n}\n<\/pre>\n<p>It's output is:<\/p>\n<pre>(binomial 5 3): 10.0\n(binomial 10042, 111): 4.9068389575068143E263\n<\/pre>\n<p>The first piece of magic is using the <code>:methods<\/code> keyword in the <code>gen-class<\/code> statement. That seems to be required to let you access the Clojure function something like static methods in Java.<\/p>\n<p>The second thing is to create a wrapper function that can be called by Java. Notice that the second version of <code>-binomial<\/code> has a dash in front of it.<\/p>\n<p>And of course the Clojure jar itself must be on the class path. This example used the Clojure-1.1.0 jar.<\/p>\n<p><strong>Update<\/strong>: This answer has been re-tested using the following tools:<\/p>\n<ul><li>Clojure 1.5.1<\/li>\n<li>Leiningen 2.1.3<\/li>\n<li>JDK 1.7.0 Update 25<\/li>\n<\/ul><p><strong>The Clojure Part<\/strong><\/p>\n<p>First create a project and associated directory structure using Leiningen:<\/p>\n<pre>C:\\projects&gt;lein new com.domain.tiny\n<\/pre>\n<p>Now, change to the project directory.<\/p>\n<pre>C:\\projects&gt;cd com.domain.tiny\n<\/pre>\n<p>In the project directory, open the <code>project.clj<\/code> file and edit it such that the contents are as shown below.<\/p>\n<pre>(defproject com.domain.tiny \"0.1.0-SNAPSHOT\"\n  :description \"An example of stand alone Clojure-Java interop\"\n  :url \"http:\/\/clarkonium.net\/2013\/06\/java-clojure-interop-an-update\/\"\n  :license {:name \"Eclipse Public License\"\n  :url \"http:\/\/www.eclipse.org\/legal\/epl-v10.html\"}\n  :dependencies [[org.clojure\/clojure \"1.5.1\"]]\n  :aot :all\n  :main com.domain.tiny)\n<\/pre>\n<p>Now, make sure all of the dependencies (Clojure) are available.<\/p>\n<pre>C:\\projects\\com.domain.tiny&gt;lein deps\n<\/pre>\n<p>You may see a message about downloading the Clojure jar at this point.<\/p>\n<p>Now edit the Clojure file <code>C:\\projects\\com.domain.tiny\\src\\com\\domain\\tiny.clj<\/code> such that it contains the Clojure program shown in the original answer. (This file was created when Leiningen created the project.)<\/p>\n<p>Much of the magic here is in the namespace declaration. The <code>:gen-class<\/code> tells the system to create a class named <code>com.domain.tiny<\/code> with a single static method called <code>binomial<\/code>, a function taking two integer arguments and returning a double. There are two similarly named functions <code>binomial<\/code>, a traditional Clojure function, and <code>-binomial<\/code> and wrapper accessible from Java. Note the hyphen in the function name <code>-binomial<\/code>. The default prefix is a hyphen, but it can be changed to something else if desired. The <code>-main<\/code> function just makes a couple of calls to the binomial function to assure that we are getting the correct results. To do that, compile the class and run the program.<\/p>\n<pre>C:\\projects\\com.domain.tiny&gt;lein run\n<\/pre>\n<p>You should see output shown in the original answer.<\/p>\n<p>Now package it up in a jar and put it someplace convenient. Copy the Clojure jar there too.<\/p>\n<pre>C:\\projects\\com.domain.tiny&gt;lein jar\nCreated C:\\projects\\com.domain.tiny\\target\\com.domain.tiny-0.1.0-SNAPSHOT.jar\nC:\\projects\\com.domain.tiny&gt;mkdir \\target\\lib\nC:\\projects\\com.domain.tiny&gt;copy target\\com.domain.tiny-0.1.0-SNAPSHOT.jar target\\lib\\\n        1 file(s) copied.\nC:\\projects\\com.domain.tiny&gt;copy \"C:&lt;path to clojure jar&gt;\\clojure-1.5.1.jar\" target\\lib\\\n        1 file(s) copied.\n<\/pre>\n<p><strong>The Java Part<\/strong><\/p>\n<p>Leiningen has a built-in task, <code>lein-javac<\/code>, that should be able to help with the Java compilation. Unfortunately, it seems to be broken in version 2.1.3. It can't find the installed JDK and it can't find the Maven repository. The paths to both have embedded spaces on my system. I assume that is the problem. Any Java IDE could handle the compilation and packaging too. But for this post, we're going old school and doing it at the command line.<\/p>\n<p>First create the file <code>Main.java<\/code> with the contents shown in the original answer.<\/p>\n<p>To compile java part<\/p>\n<pre>javac -g -cp target\\com.domain.tiny-0.1.0-SNAPSHOT.jar -d target\\src\\com\\domain\\Main.java\n<\/pre>\n<p>Now create a file with some meta-information to add to the jar we want to build. In <code>Manifest.txt<\/code>, add the following text<\/p>\n<pre>Class-Path: lib\\com.domain.tiny-0.1.0-SNAPSHOT.jar lib\\clojure-1.5.1.jar\nMain-Class: Main\n<\/pre>\n<p>Now package it all up into one big jar file, including our Clojure program and the Clojure jar.<\/p>\n<pre>C:\\projects\\com.domain.tiny\\target&gt;jar cfm Interop.jar Manifest.txt Main.class lib\\com.domain.tiny-0.1.0-SNAPSHOT.jar lib\\clojure-1.5.1.jar\n<\/pre>\n<p>To run the program:<\/p>\n<pre>C:\\projects\\com.domain.tiny\\target&gt;java -jar Interop.jar\n(binomial 5 3): 10.0\n(binomial 10042, 111): 4.9068389575068143E263\n<\/pre>\n<p>The output is essentially identical to that produced by Clojure alone, but the result has been converted to a Java double.<\/p>\n<p>As mentioned, a Java IDE will probably take care of the messy compilation arguments and the packaging.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2187427\/revisions\" title=\"show all edits to this post\">edited Jun 25 '13 at 18:41<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 2 '10 at 20:27&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/193509\/clartaq\">clartaq<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            3,66122838&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        110&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>As of Clojure 1.6.0, there is a new preferred way to load and invoke Clojure functions. This method is now preferred to calling RT directly (and supersedes many of the other answers here). The javadoc is <a href=\"http:\/\/clojure.github.io\/clojure\/javadoc\/clojure\/java\/api\/package-summary.html\">here<\/a> - the main entry point is <code>clojure.java.api.Clojure<\/code>.<\/p>\n<p>To lookup and call a Clojure function:<\/p>\n<pre>IFn plus = Clojure.var(\"clojure.core\", \"+\");\nplus.invoke(1, 2);\n<\/pre>\n<p>Functions in <code>clojure.core<\/code> are automatically loaded. Other namespaces can be loaded via require:<\/p>\n<pre>IFn require = Clojure.var(\"clojure.core\", \"require\");\nrequire.invoke(Clojure.read(\"clojure.set\"));\n<\/pre>\n<p><code>IFn<\/code>s can be passed to higher order functions, e.g. the example below passes <code>plus<\/code> to <code>read<\/code>:<\/p>\n<pre>IFn map = Clojure.var(\"clojure.core\", \"map\");\nIFn inc = Clojure.var(\"clojure.core\", \"inc\");\nmap.invoke(inc, Clojure.read(\"[1 2 3]\"));\n<\/pre>\n<p>Most <code>IFn<\/code>s in Clojure refer to functions. A few, however, refer to non-function data values. To access these, use <code>deref<\/code> instead of <code>fn<\/code>:<\/p>\n<pre>IFn printLength = Clojure.var(\"clojure.core\", \"*print-length*\");\nIFn deref = Clojure.var(\"clojure.core\", \"deref\");\nderef.invoke(printLength);\n<\/pre>\n<p>Sometimes (if using some other part of the Clojure runtime), you may need to ensure that the Clojure runtime is properly initialized - calling a method on the Clojure class is sufficient for this purpose. If you do not need to call a method on Clojure, then simply causing the class to load is sufficient (in the past there has been a similar recommendation to load the RT class; this is now preferred):<\/p>\n<pre>Class.forName(\"clojure.java.api.Clojure\") \n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/23555959\/revisions\" title=\"show all edits to this post\">edited May 12 '14 at 18:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered May 9 '14 at 3:03&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/7671\/alex-miller\">Alex Miller<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            52.7k2497147&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        33&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>EDIT<\/strong> This answer was written in 2010, and worked at that time. See Alex Miller's answer for more modern solution.<\/p>\n<p>What kind of code are calling from Java? If you have class generated with gen-class, then simply call it. If you want to call function from script, then look to <a href=\"http:\/\/en.wikibooks.org\/wiki\/Clojure_Programming\/Tutorials_and_Tips#Invoking_Clojure_from_Java\" rel=\"nofollow noreferrer\">following example<\/a>.<\/p>\n<p>If you want to evaluate code from string, inside Java, then you can use following code:<\/p>\n<pre>import clojure.lang.RT;\nimport clojure.lang.Var;\nimport clojure.lang.Compiler;\nimport java.io.StringReader;\npublic class Foo {\n  public static void main(String[] args) throws Exception {\n    \/\/ Load the Clojure script -- as a side effect this initializes the runtime.\n    String str = \"(ns user) (defn foo [a b]   (str a \\\" \\\" b))\";\n    \/\/RT.loadResourceScript(\"foo.clj\");\n    Compiler.load(new StringReader(str));\n    \/\/ Get a reference to the foo function.\n    Var foo = RT.var(\"user\", \"foo\");\n    \/\/ Call it!\n    Object result = foo.invoke(\"Hi\", \"there\");\n    System.out.println(result);\n  }\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2182608\/revisions\" title=\"show all edits to this post\">edited Jun 19 '17 at 16:48<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 2 '10 at 8:09&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/18627\/alex-ott\">Alex Ott<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            26.1k34871&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        12&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>EDIT:<\/strong> I wrote this answer almost three years ago. In Clojure 1.6 there is a proper API exactly for the purpose of calling Clojure from Java. Please Alex Miller's answer for up to date information: <a href=\"https:\/\/stackoverflow.com\/a\/23555959\/202121\">https:\/\/stackoverflow.com\/a\/23555959\/202121<\/a><\/p>\n<p><strong>Original answer from 2011:<\/strong><\/p>\n<p>As I see it, the simplest way (if you don't generate a class with AOT compilation) is to use clojure.lang.RT to access functions in clojure. With it you can mimic what you would have done in Clojure (no need to compile things in special ways):<\/p>\n<pre>;; Example usage of the \"bar-fn\" function from the \"foo.ns\" namespace from Clojure\n(require 'foo.ns)\n(foo.ns\/bar-fn 1 2 3)\n<\/pre>\n<p>And in Java:<\/p>\n<pre>\/\/ Example usage of the \"bar-fn\" function from the \"foo.ns\" namespace from Java\nimport clojure.lang.RT;\nimport clojure.lang.Symbol;\n...\nRT.var(\"clojure.core\", \"require\").invoke(Symbol.intern(\"foo.ns\"));\nRT.var(\"foo.ns\", \"bar-fn\").invoke(1, 2, 3);\n<\/pre>\n<p>It is a bit more verbose in Java, but I hope it's clear that the pieces of code are equivalent.<\/p>\n<p>This should work as long as Clojure and the source files (or compiled files) of your Clojure code is on the classpath.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/6410926\/revisions\" title=\"show all edits to this post\">edited May 23 '17 at 12:34<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jun 20 '11 at 12:13&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/202121\/raek\">raek<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,7081315&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        10&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I agree with clartaq's answer, but I felt that beginners could also use:<\/p>\n<ul><li>step-by-step information on how to actually get this running<\/li>\n<li>information that's current for Clojure 1.3 and recent versions of leiningen.  <\/li>\n<li>a Clojure jar that also includes a main function, so it can be run standalone <em>or<\/em> linked as a library.<\/li>\n<\/ul><p>So I covered all that in <a href=\"http:\/\/walkwithoutrhythm.net\/blog\/2012\/03\/26\/how-to-call-clojure-1-dot-3-functions-from-java\/\" rel=\"noreferrer\">this blog post<\/a>.  <\/p>\n<p>The Clojure code looks like this:<\/p>\n<pre>(ns ThingOne.core\n (:gen-class\n    :methods [#^{:static true} [foo [int] void]]))\n(defn -foo [i] (println \"Hello from Clojure. My input was \" i))\n(defn -main [] (println \"Hello from Clojure -main.\" ))\n<\/pre>\n<p>The leiningen 1.7.1 project setup looks like this:<\/p>\n<pre>(defproject ThingOne \"1.0.0-SNAPSHOT\"\n  :description \"Hello, Clojure\"\n  :dependencies [[org.clojure\/clojure \"1.3.0\"]]\n  :aot [ThingOne.core]\n  :main ThingOne.core)\n<\/pre>\n<p>The Java code looks like this:<\/p>\n<pre>import ThingOne.*;\nclass HelloJava {\n    public static void main(String[] args) {\n        System.out.println(\"Hello from Java!\");\n        core.foo (12345);\n    }\n}\n<\/pre>\n<p>Or you can also get all the code from <a href=\"https:\/\/github.com\/sandover\/HelloJavaClojure\" rel=\"noreferrer\">this project on github<\/a>. <\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Mar 29 '12 at 17:58&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/45850\/sandover\">sandover<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            5,00011311&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        3&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>This works with Clojure 1.5.0:<\/p>\n<pre>public class CljTest {\n    public static Object evalClj(String a) {\n        return clojure.lang.Compiler.load(new java.io.StringReader(a));\n    }\n    public static void main(String[] args) {\n        new clojure.lang.RT(); \/\/ needed since 1.5.0        \n        System.out.println(evalClj(\"(+ 1 2)\"));\n    }\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Nov 14 '13 at 15:55&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1941928\/kim-taegyoon\">KIM Taegyoon<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,1531512&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>If the use case is to include a JAR built with Clojure in a Java application, I have found having a separate namespace for the interface between the two worlds to be beneficial: <\/p>\n<pre>(ns example-app.interop\n  (:require [example-app.core :as core])\n;; This example covers two-way communication: the Clojure library \n;; relies on the wrapping Java app for some functionality (through\n;; an interface that the Clojure library provides and the Java app\n;; implements) and the Java app calls the Clojure library to perform \n;; work. The latter case is covered by a class provided by the Clojure lib.\n;; \n;; This namespace should be AOT compiled.\n;; The interface that the java app can implement\n(gen-interface\n  :name com.example.WeatherForecast\n  :methods [[getTemperature [] Double]])\n;; The class that the java app instantiates\n(gen-class\n  :name com.example.HighTemperatureMailer\n  :state state\n  :init init\n  ;; Dependency injection - take an instance of the previously defined\n  ;; interface as a constructor argument\n  :constructors {[com.example.WeatherForecast] []}\n  :methods [[sendMails [] void]])\n(defn -init [weather-forecast]\n  [[] {:weather-forecast weather-forecast}])\n;; The actual work is done in the core namespace\n(defn -sendMails\n  [this]\n  (core\/send-mails (.state this)))\n<\/pre>\n<p>The core namespace can use the injected instance to accomplish its tasks:<\/p>\n<pre>(ns example-app.core)\n(defn send-mails \n  [{:keys [weather-forecast]}]\n  (let [temp (.getTemperature weather-forecast)] ...)) \n<\/pre>\n<p>For testing purposes, the interface can be stubbed:<\/p>\n<pre>(example-app.core\/send-mails \n  (reify com.example.WeatherForecast (getTemperature [this] ...)))\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 8 '16 at 15:35&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1479173\/jstaffans\">jstaffans<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Other technique that works also with other languages on top of JVM is to declare an interface for functions you want to call and then use 'proxy' function to create instance that implemennts them.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Feb 17 '10 at 8:58&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/117220\/petr-gladkikh\">Petr Gladkikh<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,32821223&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        -1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>You can also use AOT compilation to create class files representing your clojure code.  Read the documentation about compilation, gen-class and friends in the Clojure API docs for the details about how to do this, but in essence you will create a class that calls clojure functions for each method invocation.<\/p>\n<p>Another alternative is to use the new defprotocol and deftype functionality, which will also require AOT compilation but provide better performance.  I don't know the details of how to do this yet, but a question on the mailing list would probably do the trick.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Feb 2 '10 at 17:07&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/131935\/rosejn\">rosejn<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            616168&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-11-07T16:54:24+0000","updated_at":"2018-11-07T16:54:24+0000","annotations":[],"mimetype":"text\/html","reading_time":13,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":649,"title":"Home \u00b7 ENSIME","url":"http:\/\/ensime.github.io\/","content":"<p>ENSIME brings Scala and Java IDE-like features to your favourite text editor, such as:<\/p><ul><li>Show inferred types.<\/li>\n  <li>Contextual completion.<\/li>\n  <li>Semantic highlighting with <code class=\"highlighter-rouge\">implicit<\/code> expansions.<\/li>\n  <li>Jump to source code or documentation.<\/li>\n  <li>Refactorings (rename, organise imports).<\/li>\n  <li>Errors and warnings in your code (i.e. \u201cred squigglies\u201d).<\/li>\n<\/ul><p>ENSIME is <a href=\"https:\/\/www.gnu.org\/philosophy\/free-sw.en.html\">Libre Software<\/a> and appeals to hackers, minimalists and connoisseurs - engineers and artists who craft their own exquisite tools and help their neighbour. If you would rather use a mainstream IDE, you will prefer <a href=\"https:\/\/www.jetbrains.com\/help\/idea\/2016.1\/creating-and-running-your-scala-application.html\">IntelliJ IDEA<\/a> or <a href=\"http:\/\/scala-ide.org\/\">Scala IDE<\/a>. We don\u2019t claim to be better, just different.<\/p><p>If you would like to learn more, please read our <a href=\"http:\/\/ensime.github.io\/getting_started\">Getting Started<\/a> section.<\/p><p>Please consider <a href=\"http:\/\/ensime.github.io\/sponsor\">sponsoring a developer<\/a> to ensure that ENSIME remains sustainable.<\/p><p><a href=\"https:\/\/github.com\/ensime\/ensime.github.io\/edit\/master\/index.md\">[edit]<\/a> \u2190\n    <small><i>Every page has an edit button, click it to improve the documentation.<\/i><\/small>\n  <\/p>","created_at":"2018-11-07T16:54:22+0000","updated_at":"2018-11-07T16:54:22+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"ensime.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":650,"title":"Bootstrapping In a Nutshell \u2013 Ep #2: How to Validate Your Idea By Talking to Customers \u2013 Indie Hackers","url":"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-part-2-how-to-validate-ideas-by-talking-to-customers-47cab94448","content":"<div class=\"medium-insert-images medium-insert-images-full\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C47cab94448\/validation.jpg\" alt=\"\" class=\"\" \/><\/figure><\/div><p class=\"\"><i>Bootstrapping In a Nutshell is a series of articles where you'll learn how to bootstrap a new online business from scratch through a real-world project.<\/i><br \/><\/p><p class=\"\">Now that the subconscious part of your brain will wake you up at 1:42 in the morning with an <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-1-how-to-select-your-domain-name-for-seo-a5eb5879a7\" target=\"_blank\" rel=\"noopener noreferrer\">awesome domain name\u00a0made for SEO<\/a>, you can move on the most important part of your job.<\/p><p class=\"\">Which is also, unfortunately, the most difficult one: talking to customers to validate your idea.<br \/><\/p><h2>Why founders never get this right<\/h2><p class=\"\">Yes. You read well. Validating your idea by talking to customers before doing any building is the most difficult part of bootstrapping a business.<\/p><p class=\"\">For several reasons:<\/p><ul><li>You may have no connections with your audience yet, no network, and no friends interested in your idea. So you'll have to work your ass off to build these.<br \/><\/li><li>You don't want to get out of your comfort zone by talking to real people. It's scary. They may\u00a0burn your beloved idea into ashes. They can even make fun of you publicly on a forum (yes, that did happen to me \u2014 I'm glad you asked).<br \/><\/li><li>You want to avoid a feeling of rejection if nobody responds to your post.<\/li><li>You absolutely want to work on your idea because you <i>know<\/i>\u00a0it will work! So you avoid any confrontation with the real world as long as possible to avoid this harsh moment when the market says, \"No, I decided otherwise. Sorry bro.\"<\/li><\/ul><p class=\"\">The thing is... If you're reading this article right now, it means that you want to make money out of your project so you must make a product that's useful.<\/p><blockquote><p>Make something people want \u2014 Paul Graham<\/p><\/blockquote><p class=\"\">To be sure it's something people want, you must involve potential customers from day one, even before building the product. This way they can guide you along the way to create a product that will be customer-centric: a key to success.<\/p><p class=\"\">That part of the process is the most difficult one to get right for all the reasons I mentioned above. But you can handle it easily with this little secret that will change your whole perspective about this.<\/p><h2>The secret to becoming rejection-proof<\/h2><p class=\"\">The most important thing you must do to become good at this is to separate yourself from your idea. When people say, \"That's the shittiest idea I've ever heard,\" they aren't talking about you. They don't know you. So don't take it personally. They're talking about the idea you suggested, that's it.<\/p><p class=\"\">When you can separate your idea from yourself, you become unstoppable and also very flexible. Indeed, you must learn not to like an idea too much, because it will change and evolve over time based on customers feedback.<\/p><p class=\"\">Keep in mind that this part is the most difficult for every founder, because it can instantly kill all your fantasies about your idea. You'll have to base your decisions not on what you think the world wants but on what your customers want.\u00a0The difference between what you think and what they actually want is huge.<\/p><p class=\"\">This difference is what makes an entrepreneur succeed or miserably fail.<\/p><p class=\"\">If talking to people makes you feel like you're falling out of the sky and crashing to the ground, you're on the right track.<\/p><h2>How to find your customers online<\/h2><p class=\"\">First of all, if your idea has a market, you should be able to easily find some website, blog, or forums that gather the type of people that may be interested in your idea.<\/p><p class=\"\">If not, you should create this community first, before anything else. Because building a product starts by building a relationship with your customers.<\/p><p class=\"\">Here is a quick list of some online points of contact you can use to talk to potential customers:<\/p><ul><li>Social platforms: Facebook, Twitter, Instagram, LinkedIn, Pinterest, Snapchat, Youtube, etc.<\/li><li>Forums: use the advanced search prefix \"forum: &lt;subject&gt;\" in Google to filter the results to forums only.<\/li><li>Blogs: use\u00a0the advanced search prefix \"blog: &lt;subject&gt;\" in Google to filter the results to blogs only.<\/li><li>Quora: use the advanced search prefix \"site:quora.com &lt;subject&gt;\" in Google to filter the results to Quora only. Ask questions there, answer questions, contact directly people that asked some questions.<\/li><li>Competitors websites &amp; forums.<\/li><\/ul><p class=\"\">This is not an exhaustive list. I don't want to create one because it's impossible. You must put yourself in the shoes of your users and ask yourself: where do they hang out online (or even offline)? You can also find some inspiration from the excellent <a href=\"https:\/\/www.sideprojectchecklist.com\/marketing-checklist\/\" target=\"_blank\" rel=\"noopener noreferrer\">Side Project Checklist<\/a>.<\/p><p class=\"\">No matter which channel you pick, remember to stay polite, follow the rules (most forums state their rules explicitly, for example), and build a relationship first with people here. It takes time. A lot. Be friendly, makes friends along the way. Then, tell them about your idea and ask for feedback. Do not post any link, simply talk about the idea, make it grow, shape it based on the feedback and on what you learn.<\/p><h2>A real-world example<\/h2><p class=\"\">I asked Google where I should go for my <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-0-how-to-select-your-next-business-idea-bd6f7c7a9c\" target=\"_blank\" rel=\"noopener noreferrer\">blockbuster movies newsletter idea<\/a>:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C47cab94448\/24ab7b8859.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">So last week, I registered on the movieforums website (@jemo) and started to post respecting the rules and presenting myself. On this forum, there is an elegant and elaborate system of reputation, people can upvote your posts. Also, you need to have 25 posts before you can post any link to avoid bots and aggressive spam. That's great!<\/p><p class=\"\">I'm not here just to get feedback on my idea, I'm here to stay, because I love movies. Period. So all these restrictions are perfect, indeed they ensure a good quality of the community.<\/p><p class=\"\">I learned a lot from this forum in just a couple of weeks by talking to people. Here are some things I learned:<\/p><ul><li>People are passionate about movies. There are thousands of people talking about and rating movies they just watched. People write huge blog posts about them, it's crazy! That's not a surprise, but I didn't think it was so deep. People love movies.<\/li><li>People purchase DVD and Blu-rays several times a week. I asked myself why? With Netflix and the streaming technology, why do they still use disks? Well because they just love putting a DVD into a player. It's like a ritual. It's an emotional connection to a physical object (like people loving books).<\/li><li>Netflix still runs <a href=\"https:\/\/dvd.netflix.com\" target=\"_blank\" rel=\"noopener noreferrer\">its DVD service<\/a>! Their <a href=\"https:\/\/dvd.netflix.com\/HowItWorks\" target=\"_blank\" rel=\"noopener noreferrer\">landing page<\/a> is really well crafted. \"Happiness delivered to you\". Emotions. Benefits for the user, no features talk, no contract talk, they bring you happiness through the mail. Awesome.<br \/><\/li><li>I learned that Netflix is profitable with their DVD service even at $4.99 \/ mo because of some \"tricks\". Mailing in volume is cheap. Plus people don't see the movie immediately when they receive it. And there is the mailing time too: even if the mailing is free for the user, it still takes 2 days on average to go back to Netflix and 2 more days so they can send you the next movie on your list. Therefore, in total, you can see maybe 10 movies per month if you see it right away and send it back immediately (which is never the case \u2014 we're busy people). Those are interesting insights.<\/li><li>I learned that you can pre-order DVD and Blu-rays at a discounted price on Amazon before they're released. That is a very good value proposition for our business like \"Never miss a blockbuster. Save money by pre-ordering the movie you didn't see at the theater\". It makes my brain think. Love it. Also, it's easy to use affiliate marketing in the newsletter to make money out of it.<\/li><li>I learned that \"blockbuster\" was <a href=\"https:\/\/en.wikipedia.org\/wiki\/Blockbusters_(U.S._game_show)\" target=\"_blank\" rel=\"noopener noreferrer\">a famous TV show<\/a>\u00a0which means that this keyword may have several meanings for people searching for it. So I should use it carefully.<\/li><\/ul><p class=\"\">I started a <a href=\"https:\/\/www.movieforums.com\/community\/showthread.php?t=51569\" target=\"_blank\" rel=\"noopener noreferrer\">new thread about blockbusters<\/a> to see if I can have discussions with blockbusters fans like me. I will send them some private message about my idea so I respect the forum rules (no marketing, no spam). I will probably start other threads and talk about forgetting to go to see a movie when it's at the theater for example, I may get some insights from that too.<\/p><p class=\"\"><i>Update:<\/i>\u00a0Well, in fact, while writing this article a new very interesting answer came today on this thread:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C47cab94448\/0118d72613.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">Missing a movie that was playing at the theater may be a pain because you'll have to wait several months before it appears on the Netflix streaming service or for the DVD\/Blu-ray to be released. My newsletter could fix that problem by sending to your inbox the new movies each week so you never miss one. That can be part of my value proposition, why not. We can also imagine some kind of reminders by email \u2014 when you want to see a movie you just tap on a \"send me a reminder\" button... Interesting!<\/p><p class=\"\">Let's get back to the forum usage. I want you to keep in mind one thing here. I hate marketing. I hate advertisement. I hate when people come to a forum to drop a link to their website to get traffic. So you should never do that. NEVER.<\/p><p class=\"\">I add new posts on this forum that I enjoy writing and talk about with the people there. Because I genuinely love movies. I always open up a new thread with something interesting and new for the forum that brings value to the community.<\/p><p class=\"\">For example, even if I'm a big fan of blockbusters I'm wondering about the fact that indie movies have absolutely no chance to get to the theater and be watched by people when you compare the marketing budget they have against big productions. Maybe a startup (or a newsletter!) that promote only \"indie\" movies could be a good thing to do?<\/p><p class=\"\">Did you see what just happened? By using my initial idea and some interaction with my potential customers, idea generation kicked in! So really, I've confirmed that <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-0-how-to-select-your-next-business-idea-bd6f7c7a9c\" target=\"_blank\" rel=\"noopener noreferrer\">your first idea doesn't really matter<\/a>, like I told you!<\/p><h2>How founders can always get this right<\/h2><p class=\"\">At this point, you should talk to your customers. Over and over before purchasing your domain name, before creating a landing page, before anything else. Just talk about your idea, try to shape it, make it grow, make it mature. Try to manually contact at least 100 people to talk about it. That's a lot, but that's what will make your results more trustworthy.<\/p><p class=\"\">The time you invest here is not lost \u2014 it will make you save time in the long run by not working on a product that no one will want at the end. This is a huge advantage, really. So if you think \"I'm losing my time talking to guys I don't even know instead of building something and show it to them\", please, do yourself a favor:<\/p><p class=\"\"><b>Rewire your brain.<\/b><\/p><p class=\"\">Think about it this way: \"I'm saving time and pain by carefully involving users in my product from the start. This way I ensure that I spend my precious time and talent on things that matter to launch a successful product.\"<\/p><h2>How to know when your idea is validated<\/h2><p class=\"\">Now in term of numbers. It's very difficult to put a threshold on how to know if your idea is validated. You can ask 100 people and 15 can be interested, is that enough? I don't know. This part is extremely difficult because you can ask the wrong people so your conversion rate will be very low, even if your idea aligns with the right people.<\/p><p class=\"\">For example, you can talk to 100 random people in the street and have 5 people interested in your idea. Should you pursue? Should you try to hit 10, 15, or 30 interested people? Moreover, you can ask the right people but fall into an extremist subset of your market that doesn't represent the real market. Like you talked to 10 Linux users, and 9 of them are interested in your awesome new Linux-based idea! But it turns out that these users are Linux extremists so you fool yourself.<\/p><p class=\"\">As you should put a threshold on this because you go nowhere otherwise, I suggest you put it on 10%. So you'll have to contact 100 people and have 10 of them that are interested in your idea. But what does interested mean?<\/p><p class=\"\">Again, that's something quite difficult to address. If you can pre-sell a product, that should mean 10 pre-orders. So people should actually pay for your product before it's even built. For a newsletter, it should be 10 subscribers that actually put their emails in the form of a landing page.<\/p><p class=\"\">Now let's be clear here: having a 10% conversion rate for a newsletter is far easier than for pre-ordering a product. Indeed people put their email easily on a newsletter landing page, but getting them to pull out their wallets is another story.<\/p><p class=\"\">Based on all this knowledge it's quite difficult to determine if your idea is validated or not upfront, only experience will tell you. I take 10% for my newsletter idea, it seems reasonable, but you may adjust and take 5% for pre-selling your ebook.<\/p><p class=\"\">Another way to think about it is by doing some CAC (Customer Acquisition Cost) and LTV (Life-Time Value of a customer) calculations. For example, let's say you want to sell an ebook. You put a landing page with a pre-sell order at $9. If you drive traffic to your website with some ads, and each click costs you $0.15, you can compute a minimum conversion rate so you're profitable with your ads (or at least don't lose money). If each visitor costs you $0.15, this means that at least 1 out of $9 \/ $0.15 = 60 visitors should pre-order your book, so you should have a conversion rate of at least 1.67%.<\/p><p class=\"\">Anyway, judging if an idea is validated or not is really difficult. Determine a threshold and try to stick with it. Test multiple ideas, listen to your potential customers, let them talk, be friendly, make friends along the way and be flexible!<\/p><p class=\"\">That's what I'm gonna do. The next step after talking to your customers is to make them actually take action with your product, engage with it. For this purpose you'll need to create a landing page.<\/p><p class=\"\"><i>In the next article, I'll show you how to create a landing page with a custom domain name through <a href=\"http:\/\/carrd.co\" target=\"_blank\" rel=\"noopener noreferrer\">carrd.co<\/a>. Write your thoughts about my articles in the comments: what can I do to improve them? What do you love the most about them? Thanks!<\/i><\/p><blockquote><p><b>ABOUT ME<\/b>: I'm a 32 yo software engineer and entrepreneur. I use my free time to start new online businesses. Actually I'm learning front-end development while building an audience on twitter, <a href=\"https:\/\/twitter.com\/jeremymouzin\" target=\"_blank\" rel=\"noopener noreferrer\">follow me<\/a> ;-).<\/p><p><b>NEXT ARTICLE<\/b>: <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-3-how-to-create-your-landing-page-25-off-carrd-co-3651e8c05f\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #3: How to Create Your Landing Page<\/a><\/p><p><b>FULL ARTICLES LIST<\/b><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-0-how-to-select-your-next-business-idea-bd6f7c7a9c\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #0: How to Select Your Next Business Idea<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-1-how-to-select-your-domain-name-for-seo-a5eb5879a7\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #1: How to Select Your Domain Name for SEO<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-2-how-to-validate-your-idea-by-talking-to-customers-47cab94448\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #2: How to Validate Your Idea by Talking to Customers<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-3-how-to-create-your-landing-page-25-off-carrd-co-3651e8c05f\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #3: How to Create Your Landing Page<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-4-how-to-set-up-google-analytics-for-your-landing-page-191a88ea92\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #4: How to Set Up Google Analytics For Your Landing Page<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-5-how-to-drive-quality-traffic-to-your-landing-page-44f14d0c66\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #5: How to Drive Quality Traffic to Your Landing Page<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/edit-article\/@Hercule\/bootstrapping-in-a-nutshell-ep-6-advice-for-beginners-and-conclusion-b604c96ab0\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #6: Advice for beginners and conclusion<\/a><br \/><\/p><\/blockquote>","created_at":"2018-11-07T16:54:22+0000","updated_at":"2018-11-07T16:54:22+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":12,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C47cab94448\/validation.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":646,"title":"Startup Stash - Curated resources and tools for startups","url":"http:\/\/startupstash.com\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:54:21+0000","updated_at":"2018-11-07T16:54:21+0000","annotations":[],"reading_time":0,"domain_name":"startupstash.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":647,"title":"An Old Article I Wrote","url":"https:\/\/cdsmith.wordpress.com\/2011\/01\/09\/an-old-article-i-wrote\/","content":"<div class=\"post-box\">\n\t\t\t<div id=\"post-224\" class=\"post-224 post type-post status-publish format-standard hentry category-computer-science category-haskell\">\n\t\t\t\t<div class=\"meta clear\">\n\t\t\t\t\t<div class=\"author\">January 9, 2011 \/ cdsmith<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"post-header\">\n\t\t\t\t <h1>An Old Article I\u00a0Wrote<\/h1>\n\t\t\t\t \t\t\t\t<\/div>\n\t\t\t\t<div class=\"entry clear\">\n\t\t\t\t\t\t\t\t\t\t<p>I was asked by someone to put this old piece of writing of mine somewhere more permanent, so here it is. \u00a0A few comments off the top:<\/p>\n<ul><li>This isn\u2019t new. \u00a0I wrote it years ago.<\/li>\n<li>Yes, I\u2019m a bit embarrassed at some of the wording, as most people are when they read content they wrote a long time ago. \u00a0There are things that I\u2019d say differently if I wrote something like this today. \u00a0There is clumsy wording, and there are poor choices of examples. \u00a0I\u2019m not taking this opportunity to revise the writing; this is exactly as I wrote it all those years ago.<\/li>\n<li>If pressed for an answer, though, yes I still do believe in the central conclusions of the article. \u00a0Namely:\n<ul><li>That \u201cstatic typing\u201d and \u201cdynamic typing\u201d are two concepts that are fundamentally unrelated to each other, and just happen to share a word.<\/li>\n<li>That \u201cstatic types\u201d are, at their core a tool for writing and maintaining computer-checked proofs about code<\/li>\n<li>That \u201cdynamic types\u201d are, at their core, there to make unit testing less tedious, and are a tool for finding bugs.<\/li>\n<li>That the two are related in the way I outline: that one establishes lower bounds on correctness of code, while the other establishes upper bounds, and that questions about their real world use should come down to the possibility and effectiveness of addressing certain kinds of bugs by either computer-checked proof, or testing.<\/li>\n<\/ul><\/li>\n<\/ul><p>With that said, here is the original article:<\/p>\n<hr \/><blockquote>\n<h3>What To Know Before Debating Type Systems<\/h3>\n<p>I would be willing to place a bet that most computer programmers have, on multiple occasions, expressed an opinion about the desirability of certain kinds of type systems in programming languages. Contrary to popular conception, that\u2019s a great thing! Programmers who care about their tools are the same programmers who care about their work, so I hope the debate rages on.<\/p>\n<p>There are a few common misconceptions, though, that confuse these discussions. This article runs through those I\u2019ve encountered that obscure the most important parts of the debate. My goal is to build on a shared understanding of some of the basic issues, and help people get to the interesting parts more quickly.<\/p>\n<h3>Classifying Type Systems<\/h3>\n<p>Type systems are commonly classified by several words, of which the most common are \u201cstatic,\u201d \u201cdynamic,\u201d \u201cstrong,\u201d and \u201cweak.\u201d In this section, I address the more common kinds of classification. Some are useful, and some are not.<\/p>\n<h4>Strong and Weak Typing<\/h4>\n<p>Probably the most common way type systems are classified is \u201cstrong\u201d or \u201cweak.\u201d This is unfortunate, since these words have nearly no meaning at all. It is, to a limited extent, possible to compare two languages with very similar type systems, and designate one as having the strong<strong>er<\/strong> of those two systems. Beyond that, the words mean nothing at all.<\/p>\n<p>Therefore: I give the following general definitions for strong and weak typing, at least when used as absolutes:<\/p>\n<ul><li><strong>Strong typing:<\/strong> A type system that I like and feel comfortable with<\/li>\n<li><strong>Weak typing:<\/strong> A type system that worries me, or makes me feel uncomfortable<\/li>\n<\/ul><p>What about when the phrase is used in a more limited sense? Then strong typing, depending on the speaker or author, may mean anything on the spectrum from \u201cstatic\u201d to \u201csound,\u201d both of which are defined below.<\/p>\n<h4>Static and Dynamic Types<\/h4>\n<p>This is very nearly the\u00a0<em>only<\/em> common classification of type systems that has real meaning. As a matter of fact, it\u2019s significance is frequently under-estimated. I realize that may sound ridiculous; but this theme will recur throughout this article. Dynamic and static type systems are two\u00a0<em>completely<\/em> different things, whose goals happen to partially overlap.<\/p>\n<p>A static type system is a mechanism by which a compiler examines source code and assigns labels (called \u201ctypes\u201d) to pieces of the syntax, and then uses them to infer something about the program\u2019s behavior. A dynamic type system is a mechanism by which a compiler generates code to keep track of the sort of data (coincidentally, also called its \u201ctype\u201d) used by the program. The use of the same word \u201ctype\u201d in each of these two systems is, of course, not really entirely coincidental; yet it is best understood as having a sort of weak historical significance. Great confusion results from trying to find a world view in which \u201ctype\u201d really means the same thing in both systems. It doesn\u2019t. The better way to approach the issue is to recognize that:<\/p>\n<ul><li>Much of the time, programmers are trying to solve the same problem with static and dynamic types.<\/li>\n<li>Nevertheless, static types are not limited to problems solved by dynamic types.<\/li>\n<li>Nor are dynamic types limited to problems that can be solved with static types.<\/li>\n<li>At their core, these two techniques are not the same thing at all.<\/li>\n<\/ul><p>Observing the second of these four simple facts is a popular pass-time in some circles. Consider\u00a0<a href=\"http:\/\/web.archive.org\/web\/20080822101209\/http:\/\/perl.plover.com\/yak\/typing\/notes.html\">this set of presentation notes<\/a>, with a rather complicated \u201cthe type system found my infinite loop\u201d comment. From a theoretical perspective, preventing infinite loops is in a very deep sense the most basic possible thing you can do with static types! The simply-typed lambda calculus, on which all other type systems are based, proves that programs terminate in a finite amount of time. Indeed, the more interesting question is how to usefully extend the type system to be able to describe programs that don\u2019t terminate! Finding infinite loops, though, is not in the class of things most people associate with \u201ctypes,\u201d so it\u2019s surprising. It is, indeed, provably impossible with dynamic types (that\u2019s called the halting problem; you\u2019ve probably heard of it!). But it\u2019s nothing special for static types. Why? Because they are an entirely different thing from dynamic types.<\/p>\n<p>The dichotomy between static and dynamic types is somewhat misleading. Most languages, even when they claim to be dynamically typed, have some static typing features. As far as I\u2019m aware,\u00a0<em>all<\/em> languages have some dynamic typing features. However, most languages can be characterized as choosing one or the other. Why? Because of the first of the four facts listed above: many of the problems solved by these features overlap, so building in strong versions of both provides little benefit, and significant cost.<\/p>\n<h4>Other Distinctions<\/h4>\n<p>There are many other ways to classify type systems. These are less common, but here are some of the more interesting ones:<\/p>\n<ul><li><strong>Sound types.<\/strong> A sound type system is one that provides some kind of guarantee. It is a well-defined concept relating to static type systems, and has proof techniques and all those bells and whistles. Many modern type systems are sound; but older languages like C often do not have sound type systems by design; their type systems are just designed to give warnings for common errors. The concept of a sound type system can be imperfectly generalized to dynamic type systems as well, but the exact definition there may vary with usage.<\/li>\n<li><strong>Explicit\/Implicit Types.<\/strong> When these terms are used, they refer to the extent to which a compiler will reason about the static types of parts of a program. All programming languages have some form of reasoning about types. Some have more than others. ML and Haskell have implicit types, in that no (or very few, depending on the language and extensions in use) type declarations are needed. Java and Ada have very explicit types, and one is constantly declaring the types of things. All of the above have (relatively, compared to C and C++, for example) strong static type systems.<\/li>\n<li><strong>The Lambda Cube.<\/strong> Various distinctions between static type systems are summarized with an abstraction called the \u201clambda cube.\u201d Its definition is beyond the scope of this article, but it basically looks at whether the system provides certain features: parametric types, dependent types, or type operators. Look\u00a0<a href=\"http:\/\/web.archive.org\/web\/20080822101209\/http:\/\/citeseer.ist.psu.edu\/barendregt92lambda.html\">here<\/a> for more information.<\/li>\n<li><strong>Structural\/Nominal Types.<\/strong> This distinction is generally applied to static types with subtyping. Structural typing means a type is assumed whenever it is possible to validly assume it. For example, a record with fields called x, y, and z might be automatically considered a subtype of one with fields x and y. With nominal typing, there would be no such assumed relationship unless it were declared somewhere.<\/li>\n<li><strong>Duck Typing.<\/strong> This is a word that\u2019s become popular recently. It refers to the dynamic type analogue of structural typing. It means that rather than checking a tag to see whether a value has the correct general type to be used in some way, the runtime system merely checks that it supports all of the operations performed on it. Those operations may be implemented differently by different types.<\/li>\n<\/ul><p>This is but a small sample, but this section is too long already.<\/p>\n<h3>Fallacies About Static and Dynamic Types<\/h3>\n<p>Many programmers approach the question of whether they prefer static or dynamic types by comparing some languages they know that use both techniques. This is a reasonable approach to most questions of preference. The problem, in this case, is that most programmers have limited experience, and haven\u2019t tried a lot of languages. For context, here, six or seven doesn\u2019t count as \u201ca lot.\u201d On top of that, it requires more than a cursory glance to really see the benefit of these two very different styles of programming. Two interesting consequences of this are:<\/p>\n<ul><li>Many programmers have used very poor statically typed languages.<\/li>\n<li>Many programmers have used dynamically typed languages very poorly.<\/li>\n<\/ul><p>This section, then, brings up some of the consequences of this limited experience: things many people assume about static or dynamic typing that just ain\u2019t so.<\/p>\n<h4>Fallacy: Static types imply type declarations<\/h4>\n<p>The thing most obvious about the type systems of Java, C, C++, Pascal, and many other widely-used \u201cindustry\u201d languages is not that they are statically typed, but that they are explicitly typed. In other words, they require lots of type declarations. (In the world of less explicitly typed languages, where these declarations are optional, they are often called \u201ctype annotations\u201d instead. You may find me using that word.) This gets on a lot of people\u2019s nerves, and programmers often turn away from statically typed languages for this reason.<\/p>\n<p>This has nothing to do with static types. The first statically typed languages were explicitly typed by necessity. However, type inference algorithms \u2013 techniques for looking at source code with no type declarations at all, and deciding what the types of its variables are \u2013 have existed for many years now. The ML language, which uses it, is among the older languages around today. Haskell, which improves on it, is now about 15 years old. Even C# is now adopting the idea, which will raise a lot of eyebrows (and undoubtedly give rise to claims of its being \u201cweakly typed\u201d \u2014 see definition above). If one does not like type declarations, one is better off describing that accurately as not liking explicit types, rather than static types.<\/p>\n<p>(This is not to say that type declarations are always bad; but in my experience, there are few situations in which I\u2019ve wished to see them required. Type inference is generally a big win.)<\/p>\n<h4>Fallacy: Dynamically typed languages are weakly typed<\/h4>\n<p>The statement made at the beginning of this thread was that many programmers have used dynamically typed languages poorly. In particular, a lot of programmers coming from C often treat dynamically typed languages in a manner similar to what made sense for C prior to ANSI function prototypes. Specifically, this means adding lots of comments, long variable names, and so forth to obssessively track the \u201ctype\u201d information of variables and functions.<\/p>\n<p>Doing this prevents a programmer from realizing the benefits of dynamic typing. It\u2019s like buying a new car, but refusing to drive any faster than a bicycle. The car is horrible; you can\u2019t get up the mountain trails, and it requires gasoline on top of everything else. Indeed, a car is a pretty lousy excuse for a bicycle! Similarly, dynamically typed languages are pretty lousy excuses for statically typed languages.<\/p>\n<p>The trick is to compare dynamically typed languages when used in ways that fit in with their design and goals. Dynamically typed languages have all sorts of mechanisms to fail immediately and clearly if there is a runtime error, with diagnostics that show you exactly how it happened. If you program with the same level of paranoia appropriate to C \u2013 where a simple bug may cause a day of debugging \u2013 you will find that it\u2019s tough, and you won\u2019t be actually using your tools.<\/p>\n<p>(As a side comment, and certainly a more controversial one, the converse is equally true; it doesn\u2019t make sense to do the same kinds of exhaustive unit testing in Haskell as you\u2019d do in Ruby or Smalltalk. It\u2019s a waste of time. It\u2019s interesting to note that the whole TDD movement comes from people who are working in dynamically typed languages\u2026 I\u2019m not saying that unit testing is a bad idea with static types; only that it\u2019s entirely appropriate to scale it back a little.)<\/p>\n<h4>Fallacy: Static types imply upfront design or waterfall methods<\/h4>\n<p>Some statically typed languages are also designed to enforce someone\u2019s idea of a good development process. Specifically, they often require or encourage that you specify the whole interface to something in one place, and then go write the code. This can be annoying if one is writing code that evolves over time or trying out ideas. It sometimes means changing things in several different places in order to make one tweak. The worst form of this I\u2019m aware of (though done mainly for pragmatic reasons rather than ideological ones) is C and C++ header files. Pascal has similar aims, and requires that all variables for a procedure or function be declared in one section at the top. Though few other languages enforce this separation in quite the same way or make it so hard to avoid, many do encourage it.<\/p>\n<p>It is absolutely true that these language restrictions can get in the way of software development practices that are rapidly gaining acceptance, including agile methodologies. It\u2019s also true that they have nothing to do with static typing. There is nothing in the core ideas of static type systems that has anything to do with separating interface from implementation, declaring all variables in advance, or any of these other organizational restrictions. They are sometimes carry-overs from times when it was considered normal for programmers to cater to the needs of their compilers. They are sometimes ideologically based decisions. They are not static types.<\/p>\n<p>If one doesn\u2019t want a language deciding how they should go about designing their code, it would be clearer to say so. Expressing this as a dislike for static typing confuses the issue.<\/p>\n<p>This fallacy is often stated in different terms: \u201cI like to do exploratory programming\u201d is the popular phrase. The idea is that since everyone knows statically typed languages make you do everything up front, they aren\u2019t as good for trying out some code and seeing what it\u2019s like. Common tools for exploratory programming include the REPL (read-eval-print loop), which is basically an interpreter that accepts statements in the language a line at a time, evaluates them, and tells you the result. These tools are quite useful, and they exist for many languages, both statically and dynamically typed. They don\u2019t exist (or at least are not widely used) for Java, C, or C++, which perpetuates the unfortunate myth that they only work in dynamically typed languages. There may be advantages for dynamic typing in exploratory programming (in fact, there certainly are\u00a0<em>some<\/em> advantages, anyway), but it\u2019s up to someone to explain what they are, rather than just to imply the lack of appropriate tools or language organization.<\/p>\n<h4>Fallacy: Dynamically typed languages provide no way to find bugs<\/h4>\n<p>A common argument leveled at dynamically typed languages is that failures will occur for the customer, rather than the developer. The problem with this argument is that it very rarely occurs in reality, so it\u2019s not very convincing. Programs written in dynamically typed languages don\u2019t have far higher defect rates than programs written in languages like C++ and Java.<\/p>\n<p>One can debate the reasons for this, and there are good arguments to be had there. One reason is that the average skill level of programmers who know Ruby is higher than those who know Java, for example. One reason is that C++ and Java have relatively poor static type systems. Another reason, though, is testing. As mentioned in the aside above, the whole unit testing movement basically came out of dynamically typed languages. It has some definite disadvantages over the guarantees provided by static types, but it also has some advantages; static type systems can\u2019t check nearly as many properties of code as testing can. Ignoring this fact when talking to someone who really knows Ruby will basically get you ignored in turn.<\/p>\n<h4>Fallacy: Static types imply longer code<\/h4>\n<p>This fallacy is closely associated with the one above about type declarations. Type declarations are the reason many people associated static types with a lot of code. However, there\u2019s another side to this. Static types often allow one to write much more concise code!<\/p>\n<p>This may seem like a surprising claim, but there\u2019s a good reason. Types carry information, and that information can be used to resolve things later on and prevent programmers from needing to write duplicate code. This doesn\u2019t show up often in simple examples, but a really excellent case is found in the Haskell standard library\u2019s\u00a0<code>Data.Map<\/code> module. This module implements a balanced binary search tree, and it contains a function whose type signature looks like this:<\/p>\n<pre>lookup :: (Monad m, Ord k) =&gt; k -&gt; Map k a -&gt; m a<\/pre>\n<p>This is a magical function. It says that I can look something up in a\u00a0<code>Map<\/code> and get back the result. Simple enough, but here\u2019s the trick: what do I do if the result isn\u2019t there? Common answers might include returning a special \u201cnothing\u201d value, or aborting the current computation and going to an error handler, or even terminating the whole program. The function above does any of the above! Here\u2019s how I compare the result against a special nothing value:<\/p>\n<pre>case (lookup bobBarker employees) of&#13;\n    Nothing     -&gt; hire bobBarker&#13;\n    Just salary -&gt; pay bobBarker salary<\/pre>\n<p>How does Haskell know that I want to choose the option of getting back\u00a0<code>Nothing<\/code> when the value doesn\u2019t exist, rather than raising some other kind of error? It\u2019s because I wrote code afterward to compare the result against<code>Nothing<\/code>! If I had written code that didn\u2019t immediately handle the problem but was called from somewhere that handled errors three levels up the stack, then\u00a0<code>lookup<\/code> would have failed that way instead, and I\u2019d be able to write seven or eight consecutive lookup statements and compute something with the results without having to check for\u00a0<code>Nothing<\/code> all the time. This completely dodges the very serious \u201cexception versus return value\u201d debate in handling failures in many other languages. This debate has no answer. Return values are great if you want to check them now; exceptions are great if you want to handle them several levels up. This code simply goes along with whatever you write the code to do.<\/p>\n<p>The details of this example are specific to Haskell, but similar examples can be constructed in many statically typed languages. There is no evidence that code in ML or Haskell is any longer than equivalent code in Python or Ruby. This is a good thing to remember before stating, as if it were obviously true, that statically typed languages require more code. It\u2019s not obvious, and I doubt if it\u2019s true.<\/p>\n<h3>Benefits of Static Types<\/h3>\n<p>My experience is that the biggest problems in the static\/dynamic typing debate occur in failing to understand the issues and potential of static types. The next two sections, then, are devoted to explaining this position in detail. This section works upward from the pragmatic perspective, while the next develops it into its full form.<\/p>\n<p>There are a number of commonly cited advantages for static typing. I am going to list them in order from\u00a0<em>least<\/em> to\u00a0<em>most<\/em> significant. (This helps the general structure of working up to the important stuff.)<\/p>\n<h4>Performance<\/h4>\n<p>Performance is the gigantic red herring of all type system debates. The knowledge of the compiler in a statically typed language can be used in a number of ways, and improving performance is one of them. It\u2019s one of the least important, though, and one of the least interesting.<\/p>\n<p>For most computing environments, performance is the problem of two decades ago. Last decade\u2019s problem was already different, and this decades problems are at least 20 years advanced beyond performance being the main driver of technology decisions. We have new problems, and performance is not the place to waste time.<\/p>\n<p>(On the other hand, there are a few environments where performance still matters. Languages in use there are rarely dynamically typed, but I\u2019m not interested enough in them to care much. If you do, maybe this is your corner of the type system debate.)<\/p>\n<h4>Documentation<\/h4>\n<p>If, indeed, performance is irrelevant, what does one look to next? One answer is documentation. Documentation is an important aspect of software, and static typing can help.<\/p>\n<p>Why? Because documentation isn\u2019t just about comments. It\u2019s about everything that helps people understand software. Static type systems build ideas that help explain a system and what it does. The capture information about the inputs and outputs of various functions and modules. This is exactly the set of information needed in documentation. Clearly, if all of this information is written in comments, there is a pretty good chance it will eventually become out of date. If this information is written in identifier names, it will be nearly impossible to fit it all in. It turns out that type information is a very nice place to keep this information.<\/p>\n<p>That\u2019s the boring view. As everyone knows, though, it\u2019s better to have self- documenting code than code that needs a lot of comments (even if it has them!). Conveniently enough, most languages with interesting static type systems have type inference, which is directly analogous to self-documenting code. Information about the correct way to use a piece of code is extracted from the code itself (i.e., it\u2019s self-documenting), but then verified and presented in a convenient format. It\u2019s documentation that doesn\u2019t need to be maintained or even written, but is available on demand even without reading the source code.<\/p>\n<h4>Tools and Analysis<\/h4>\n<p>Things get way more interesting than documentation, though. Documentation is writing for human beings, who are actually pretty good at understanding code anyway. It\u2019s great that the static type system can help, but it doesn\u2019t do anything fundamentally new.<\/p>\n<p>Fundamentally new things happen when type systems help computer programs to understand code. Perhaps I need to explain myself here. After all, a wise man (Martin Fowler, IIRC) one said:<\/p>\n<p>\u201cAny fool can write code that a computer can understand. Good programmers write code that humans can understand.\u201dI don\u2019t disagree with Martin Fowler, but we have different definitions of\u00a0<em>understand<\/em> in mind. Getting a computer to follow code step by step is easy. Getting a computer to analyze it and answer more complex questions about it is a different thing entirely, and it is very hard.<\/p>\n<p>We often want our development tools to understand code. This is a big deal. I\u2019ll turn back to Martin Fowler, who\u00a0<a href=\"http:\/\/web.archive.org\/web\/20080822101209\/http:\/\/www.martinfowler.com\/bliki\/DynamicTyping.html\">points this out as well<\/a>.<\/p>\n<h4>Correctness<\/h4>\n<p>Ultimately, though, the justification for static typing has to come back to writing correct code. Correctness, of course, is just the program doing \u201cwhat you want.\u201d<\/p>\n<p>This is a really tough problem; perhaps the toughest of all. The theory of computation has a result called Rice\u2019s Theorem, which essentially says this: Given an arbitrary program written in a general purpose programming language, it is impossible to write a computer program that determines anything about the program\u2019s output. If I\u2019m teaching an intro to programming class and assign my students to write \u201chello world\u201d, I can\u2019t program a grader to determine if they did so or not. There will be some programs for which the answer is easy; if the program never makes any I\/O calls, then the answer is no. If the program consists of a single print statement, it\u2019s easy to check if the answer is yes. However, there will be some complicated programs for which my grader can never figure out the answer. (A minor but important technical detail: one can\u2019t run the program and wait for it to finish, because the program might never finish!) This is true of any statement about programs, including some more interesting ones like \u201cdoes this program ever finish?\u201d or \u201cdoes this program violate my security rules?\u201d<\/p>\n<p>Given that we can\u2019t actually check the correctness of a program, there are two approaches that help us make approximations:<\/p>\n<ul><li><strong>Testing:<\/strong> establishes upper bounds on correctness<\/li>\n<li><strong>Proof:<\/strong> establishes lower bounds on correctness<\/li>\n<\/ul><p>Of course, we care far more about lower bounds than upper bounds. The problem with proofs, though, is the same as the problem with documentation. Proving correctness is\u00a0easy only somewhat insanely difficult when you have a static body of code to prove things about. When the code is being maintained by three programmers and changing seven times per day, maintaining the correctness proofs falls behind. Static typing here plays exactly the same role as it does with documentation. If (and this is a big if) you can get your proofs of correctness to follow a certain form that can be reproduced by machine, the computer itself can be the prover, and let you know if the change you just made breaks the proof of correctness. The \u201ccertain form\u201d is called structural induction (over the syntax of the code), and the prover is called a type checker.<\/p>\n<p>An important point here is that static typing does not preclude proving correctness in the traditional way, nor testing the program. It is a technique to handle those cases in which testing might be guaranteed to succeed so they don\u2019t need testing; and similarly, to provide a basis from which the effort of manual proof can be saved for those truly challenging areas in which it is necessary.<\/p>\n<h4>Dynamic Typing Returns<\/h4>\n<p>Certainly dynamic typing has answers to this. Dynamically typed languages can sometimes perform rather well (see Dylan), sometimes have great tools (see Smalltalk), and I\u2019m sure they occasionally have good documentation as well, though the hunt for an example is too much for me right now. These are not knock-down arguments for static typing, but they are worth being aware of.<\/p>\n<p>The correctness case is particularly enlightening. Just as static types strengthened our proofs of correctness by making them easier and automatic, dynamic typing improves testing by making it easier and more effective. It simply makes the code fail more spectacularly. I find it amusing when novice programmers believe their main job is preventing programs from crashing. I imagine this spectacular failure argument wouldn\u2019t be so appealing to such a programmer. More experienced programmers realize that correct code is great, code that crashes could use improvement, but incorrect code that doesn\u2019t crash is a horrible nightmare.<\/p>\n<p>It is through testing, then, that dynamically typed languages establish correctness. Recall that testing establishes only upper bounds on correctness. (Dijkstra said it best: \u201cProgram testing can be used to show the presence of bugs, but never to show their absence.\u201d) The hope is that if one tries hard enough and still fails to show the presence of bugs, then their absence becomes more likely. If one can\u2019t seem to prove any better upper bound, then perhaps the correctness really is 100%. Indeed, there is probably at some correlation in that direction.<\/p>\n<h3>What is a Type?<\/h3>\n<p>This is as good a point as any to step back and ask the fundamental question: what is a type? I\u2019ve already mentioned that I think there are two answers. One answer is for static types, and the other is for dynamic types. I am considering the question for static types.<\/p>\n<p>It is dangerous to answer this question too quickly. It is dangerous because we risk excluding some things as types, and missing their \u201ctype\u201d nature because we never look for it. Indeed, the definition of a type that I will eventually give is extremely broad.<\/p>\n<h4>Problems with Common Definitions<\/h4>\n<p>One common saying, quoted often in an attempt to reconcile static and dynamic typing, goes something like this: Statically typed languages assign types to variables, while dynamically typed languages assign types to values. Of course, this doesn\u2019t actually define types, but it is already clearly and obviously wrong. One could fix it, to some extent, by saying \u201cstatically typed languages assign types to expressions, \u2026\u201d Even so, the implication that these types are fundamentally the same thing as the dynamic version is quite misleading.<\/p>\n<p>What is a type, then? When a typical programmer is asked that question, they may have several answers. Perhaps a type is just a set of possible values. Perhaps it is a set of operations (a very structural-type-ish view, to be sure). There could be arguments in favor of each of these. One might make a list: integers, real numbers, dates, times, and strings, and so on. Ultimately, though, the problem is that these are all symptoms rather than definitions. Why is a type a set of values? It\u2019s because one of the things we want to prove about our program is that it doesn\u2019t calculate the square roots of a string. Why is a type a set of operations? It\u2019s because one of the things we want to know is whether our program attempts to perform an impossible operation.<\/p>\n<p>Let\u2019s take a look at another thing we often want to know: does our web application stick data from the client into SQL queries without escaping special characters first? If this is what we want to know, then these becomes types.\u00a0<a href=\"http:\/\/web.archive.org\/web\/20080822101209\/http:\/\/blog.moertel.com\/articles\/2006\/10\/18\/a-type-based-solution-to-the-strings-problem\">This article by Tom Moertel<\/a> builds this on top of Haskell\u2019s type system. So far, it looks like a valid definition of \u201ctype\u201d is as follows: something we want to know.<\/p>\n<h4>A Type System<\/h4>\n<p>Clearly that\u2019s not a satisfactory definition of a type. There are plenty of things we want to know that types can\u2019t tell us. We want to know whether our program is correct, but I already said that types provide conservative lower bounds on correctness, and don\u2019t eliminate the need for testing or manual proof. What makes a type a type, then? The other missing component is that a type is part of a type system.<\/p>\n<p>Benjamin Pierce\u2019s book\u00a0<a href=\"http:\/\/web.archive.org\/web\/20080822101209\/http:\/\/www.amazon.com\/dp\/0262162091\">Types and Programming Languages<\/a> is far ans away the best place to read up on the nitty gritty details of static type systems, at least if you are academically inclined. I\u2019ll quote his definition.<\/p>\n<p><em>A type system is a tractable syntactic method for proving the absence of certain program behaviors by classifying phrases according to the kinds of values they compute.<\/em>This is a complex definition, but the key ideas are as follows:<\/p>\n<ul><li><em>syntactic method .. by classifying phrases<\/em>: A type system is necessarily tied to the syntax of the language. It is a set of rules for working bottom up from small to large phrases of the language until you reach the result.<\/li>\n<li><em>proving the absence of certain program behaviors<\/em>: This is the goal. There is no list of \u201ccertain\u201d behaviors, though. The word just means that for any specific type system, there will be a list of things that it proves. What it proves is left wide open. (Later on in the text: \u201c\u2026 each type system comes with a definition of the behaviors it aims to prevent.\u201d)<\/li>\n<li><em>tractable<\/em>: This just means that the type system finishes in a reasonable period of time. Without wanting to put words in anyone\u2019s mouth, I think it\u2019s safe to say most people would agree that it\u2019s a mistake to include this in the definition of a type system. Some languages even have undecidable type systems. Nevertheless, it is certainly a common goal; one doesn\u2019t expect the compiler to take two years to type-check a program, even if the program will run for two years.<\/li>\n<\/ul><p>The remainder of the definition is mainly unimportant. The \u201ckinds of values they compute\u201d is basically meaningless unless we know what kinds we might choose from, and the answer is any kind at all.<\/p>\n<p>An example looks something like that. Given the expression\u00a0<code>5 + 3<\/code>, a type checker may look at 5 and infer that it\u2019s an integer. It may look at 3 and infer it\u2019s an integer. It may then look at the + operator, and know that when + is applied to two integers, the result is an integer. Thus it\u2019s proven the absence of program behaviors (such as adding an integer to a string) by working up from the basic elements of program syntax.<\/p>\n<h4>Examples of Unusual Type Systems<\/h4>\n<p>That was a pretty boring example, and one that plays right into a trap: thinking of \u201ctype\u201d as meaning the same thing it does in a dynamic type system. Here are some more interesting problems being solved with static types.<\/p>\n<ul><li><a href=\"https:\/\/docs.google.com\/viewer?url=http%3A%2F%2Fweb.archive.org%2Fweb%2F20080822101209%2Fhttp%3A%2F%2Fwiki.di.uminho.pt%2Ftwiki%2Fpub%2FPersonal%2FXana%2FWebHome%2Freport.pdf\">http:\/\/wiki.di.uminho.pt\/twiki\/pub\/Personal\/Xana\/WebHome\/report.pdf<\/a>. Uses types to ensure that the correct kinds of data are gotten out of a relational database. Via the type system, the compiler ends up understanding how to work with concepts like functional dependencies and normal forms, and can statically prove levels of normalization.<\/li>\n<li><a href=\"https:\/\/docs.google.com\/viewer?url=http%3A%2F%2Fweb.archive.org%2Fweb%2F20080822101209%2Fhttp%3A%2F%2Fwww.cs.bu.edu%2F~hwxi%2Facademic%2Fpapers%2Fpldi98.pdf\">http:\/\/www.cs.bu.edu\/~hwxi\/academic\/papers\/pldi98.pdf<\/a>. Uses an extension to ML\u2019s type system to prove that arrays are never accessed out of bounds. This is an unusually hard problem to solve without making the languages that solve it unusable, but it\u2019s a popular one to work on.<\/li>\n<li><a href=\"https:\/\/docs.google.com\/viewer?url=http%3A%2F%2Fweb.archive.org%2Fweb%2F20080822101209%2Fhttp%3A%2F%2Fwww.cis.upenn.edu%2F~stevez%2Fpapers%2FLZ06a.pdf\">http:\/\/www.cis.upenn.edu\/~stevez\/papers\/LZ06a.pdf<\/a>. This is great. This example uses Haskell\u2019s type system to let someone define a security policy for a Haskell program, in Haskell, and then proves that the program properly implements that security. If a programmer gets security wrong, the compiler will complain rather than opening up a potential security bug in the system.<\/li>\n<li><a href=\"https:\/\/docs.google.com\/viewer?url=http%3A%2F%2Fweb.archive.org%2Fweb%2F20080822101209%2Fhttp%3A%2F%2Fwww.brics.dk%2FRS%2F01%2F16%2FBRICS-RS-01-16.pdf\">http:\/\/www.brics.dk\/RS\/01\/16\/BRICS-RS-01-16.pdf<\/a>. Just in case you thought type systems only solved easy problems, this bit of Haskell gets the type system to prove two central theorems about the simply typed lambda calculus, a branch of computation theory!<\/li>\n<\/ul><p>The point of these examples is to point out that type systems can solve all sorts of programming problems. For each of these type systems, concepts of types are created that represent the ideas needed to accomplish this particular task with the type system. Some problems solved by static type systems look nothing like the intuitive idea of a type. A buggy security check isn\u2019t normally considered a type error, but only because not many people use languages with type systems that solve that problem.<\/p>\n<p>To reiterate the point above, it\u2019s important to understand how limiting it is to insist, as many people do, that the dynamic typing definition of a \u201ctype\u201d is applied to static typing as well. One would miss the chance to solve several real-world problems mentioned above.<\/p>\n<h4>The True Meaning of Type<\/h4>\n<p>So what is a type? The only true definition is this: a type is a label used by a type system to prove some property of the program\u2019s behavior. If the type checker can assign types to the whole program, then it succeeds in its proof; otherwise it fails and points out why it failed. This is a definition, then, but it doesn\u2019t tell us anything of fundamental importance. Some further exploration leads us to insight about the fundamental trade-offs involved in using a static type checker.<\/p>\n<p>If you were looking at things the right way, your ears may have perked up a few sections back, when I said that Rice\u2019s Theorem says we can\u2019t determine anything about the output of a program. Static type systems prove properties of code, but it almost appears that Rice\u2019s Theorem means we can\u2019t prove anything of interest with a computer. If true, that would be an ironclad argument against static type systems. Of course, it\u2019s not true. However, it is very nearly true. What Rice\u2019s Theorem says is that we can\u2019t determine anything. (Often the word \u201cdecide\u201d is used; they mean the same thing here.) It didn\u2019t say we can\u2019t prove anything. It\u2019s an important distinction!<\/p>\n<p>What this distinction means is that a static type system is a conservative estimate. If it accepts a program, then we know the program has the properties proven by that type checker. If it fails\u2026 then we don\u2019t know anything. Possibly the program doesn\u2019t have that property, or possibly the type checker just doesn\u2019t know how to prove it. Furthermore, there is an ironclad mathematical proof that a type checker of any interest at all is\u00a0<em>always<\/em>conservative. Building a type checker that doesn\u2019t reject any correct programs isn\u2019t just difficult; it\u2019s impossible.<\/p>\n<p>That, then, is the trade-off. We get assurance that the program is correct (in the properties checked by this type checker), but in turn we must reject some interesting programs. To continue the pattern, this is the diametric opposite of testing. With testing, we are assured that we\u2019ll never fail a correct program. The trade-off is that for any program with an infinite number of possible inputs (in other words, any interesting program), a test suite may still accept programs that are not correct \u2013 even in just those properties that are tested.<\/p>\n<h3>Framing the Interesting Debate<\/h3>\n<p>That last paragraph summarizes the interesting part of the debate between static and dynamic typing. The battleground on which this is fought out is framed by eight questions, four for each side:<\/p>\n<ol><li>For what interesting properties of programs can we build static type systems?<\/li>\n<li>How close can we bring those type systems to the unattainable ideal of never rejecting a correct program?<\/li>\n<li>How easy can it be made to program in a language with such a static type system?<\/li>\n<li>What is the cost associated with accidentally rejecting a correct computer program?<\/li>\n<li>For what interesting properties of programs can we build test suites via dynamic typing?<\/li>\n<li>How close can we bring those test suites to the unattainable ideal of never accepting a broken program?<\/li>\n<li>How easy can it be made to write and execute test suites for programs?<\/li>\n<li>What is the cost associated with accidentally accepting an incorrect computer program?<\/li>\n<\/ol><p>If you knew the answer to those eight questions, you could tell us all, once and for all, where and how we ought to use static and dynamic typing for our programming tasks.<\/p><\/blockquote>\n<div id=\"atatags-370373-5be318bdbebb0\">\n        \n    <\/div>\t\t<div class=\"wpcnt\">\n\t\t\t<div class=\"wpa wpmrec\">\n\t\t\t\tAdvertisements\n\t\t\t\t<div class=\"u\">\t\t<div>\n\t\t<div id=\"atatags-26942-5be318bdbebe1\">\n\t\t\t\n\t\t<\/div><\/div>\t\t<div>\n\t\t<div id=\"atatags-114160-5be318bdbebe3\">\n\t\t\t\n\t\t<\/div><\/div><\/div>\n\t\t\t\t\n\t\t\t<\/div>\n\t\t<\/div>\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<div class=\"post-footer clear\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"category\">Filed under <a href=\"https:\/\/cdsmith.wordpress.com\/category\/computer-science\/\" rel=\"category tag\">Computer Science<\/a>, <a href=\"https:\/\/cdsmith.wordpress.com\/category\/haskell\/\" rel=\"category tag\">Haskell<\/a><\/div>\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n\t\t<div class=\"pagination post single clear\">\n\t\t\t<div class=\"alignleft\"><a href=\"https:\/\/cdsmith.wordpress.com\/2011\/01\/09\/request-uri-handling-in-http-a-primer\/\" rel=\"prev\">\u2190 Request URI Handling in HTTP: A\u00a0Primer<\/a><\/div>\n\t\t\t<div class=\"alignright\"><a href=\"https:\/\/cdsmith.wordpress.com\/2011\/01\/16\/haskells-own-dll-hell\/\" rel=\"next\">Haskell\u2019s Own DLL\u00a0Hell \u2192<\/a><\/div>\n\t\t<\/div>\n\t\t\t\n\n\n\n\t\n\t<div>\n\t<\/div>","created_at":"2018-11-07T16:54:21+0000","updated_at":"2018-11-07T16:54:21+0000","published_at":"2011-01-10T04:20:19+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":33,"domain_name":"cdsmith.wordpress.com","preview_picture":"https:\/\/s0.wp.com\/i\/blank.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":648,"title":"How to load and export variables from an .env file in Makefile?","url":"https:\/\/stackoverflow.com\/questions\/44628206\/how-to-load-and-export-variables-from-an-env-file-in-makefile","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        6&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>What is the best way to use a <code>.env<\/code> in a Makefile, i.e. loading that file and exporting all variables for subshells in <code>make<\/code>?<\/p>\n<p>It would be great if the proposed solution would work with <code>make<\/code> only, e.g. not using any third party tools. Also <code>.env<\/code> files support multiline variables like:<\/p>\n<p><code>\nFOO=\"this\\nis\\na\\nmultiline\\nvar\"\n<\/code><\/p>\n<p>this is why <a href=\"https:\/\/stackoverflow.com\/questions\/30300830\/load-env-file-in-makefile\">this solution<\/a> is probably not adequate.<\/p><\/div><div class=\"user-action-time\">&#13;\n        asked Jun 19 '17 at 10:32&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2881414\/bastian-venthur\">Bastian Venthur<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            798826&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"1\">&#13;\n                                1 Answer&#13;\n                                1&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        3&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Make does not offer any way to read a content of the file to some variable. So, I consider it impossible to achieve the result without using external tools. However, if I am wrong, I'd be glad to learn some new trick.<\/p>\n<p>So, let's assume there are two files, <code>.env<\/code>, being a technically correct shell file:<\/p>\n<pre>FOO=bar\nBAR=\"notfoo\" # comment\n   #comment\nMULTILINE=\"This\\nis\\nSparta!\"\n# comment\n<\/pre>\n<p>and <code>script.sh<\/code>:<\/p>\n<pre>#!\/bin\/bash\necho FOO=${FOO}\necho BAR=${BAR}\necho -e ${MULTILINE}\n<\/pre>\n<p>One solution is to include the <code>.env<\/code> file, then make sure variables are exported:<\/p>\n<pre>include .env\n$(eval export $(shell sed -ne 's\/ *#.*$\/\/; \/.\/ s\/=.*$$\/\/ p' .env))\nall:\n    .\/script.sh\n<\/pre>\n<p>Because of different treatment of quotes by shell and make, you will see the quotes in output.<\/p>\n<p>You can avoid that by reprocessing the variables by make:<\/p>\n<pre>include .env\nVARS:=$(shell sed -ne 's\/ *\\#.*$$\/\/; \/.\/ s\/=.*$$\/\/ p' .env )\n$(foreach v,$(VARS),$(eval $(shell echo export $(v)=\"$($(v))\")))\nall:\n    .\/script.sh\n<\/pre>\n<p>but then the multiline variable become a one-liner.<\/p>\n<p>Finally, you can generate a temporary file to be processed by bash and source it before any command is run:<\/p>\n<pre>SHELL=bash\nall: .env-export\n    . .env-export &amp;&amp; .\/script.sh\n.env-export: .env\n    sed -ne '\/^export \/ {p;d}; \/.*=\/ s\/^\/export \/ p' .env &gt; .env-export\n<\/pre>\n<p>Oh, new lines got messed in this case in multiline variable. You need to additionally quote them.<\/p>\n<p>Finally, you can add <code>export<\/code> to .env using above <code>sed<\/code> command, and do:<\/p>\n<pre>SHELL=bash\n%: .env-export\n    . .env-export &amp;&amp; make -f secondary \"$@\"\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jun 19 '17 at 18:07&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2989411\/arturfh\">ArturFH<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,326824&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-11-07T16:54:21+0000","updated_at":"2018-11-07T16:54:21+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":641,"title":"How I Went from Indie Lurker to Indie Hacker \u2013 Indie Hackers","url":"https:\/\/www.indiehackers.com\/@lynnetye\/how-i-went-from-indie-lurker-to-indie-hacker-d1042ffa5f","content":"<div class=\"medium-insert-images medium-insert-images-full\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/IH-article--sprinklers.jpg\" alt=\"\" class=\"\" \/><\/figure><\/div><p class=\"\">Many of you reading this are indie lurkers.<br \/><\/p><p class=\"\">You subscribe to the Indie Hackers newsletter, share inspiring interviews with friends, and visit the forum to see what people are talking about. I bet a lot of you upvote forum posts and start typing out comments, too... only to change your mind and delete, delete, delete.\u00a0<\/p><p class=\"\">I know this because that's exactly what I did for months as an indie lurker.<\/p><p class=\"\">But a couple of weeks ago that changed. I launched my product\u00a0<a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_intro\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a>, something I built all on my own, and got it to the top\u00a0of Hacker News and Product Hunt, netting almost 30,000 pageviews in a few days.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/Screen%20Shot%202017-09-21%20at%204.33.23%20PM.png\" alt=\"\" class=\"\" \/><figcaption class=\"\">Google Analytics behavior for the 14 days following my Hacker News post.<\/figcaption><\/figure><\/div><p class=\"\">There is <i>a lot<\/i> to say about what comes after that initial traffic spike (the post-launch trough of sorrow as they call it), but today I'll talk about what came before it.<\/p><p class=\"\">Below, I discuss the six major barriers that I had to break through in order to go from indie lurker to indie hacker.\u00a0<\/p><h2>Trust that slow progress is better than no progress.<\/h2><p class=\"\">I moved to San Francisco to go to grad school at UCSF and I didn't even know what a startup was until after I dropped out two years later. I was <i>amazed<\/i> by how people casually started companies, raised money, pivoted, folded, and then started new companies. It seemed like entrepreneurship was in everyone's DNA but mine.<\/p><p class=\"\">Even after working at my first tech startup, I still didn't see founder potential in myself. I wondered what I could have been had I been born into a family of business owners instead of a family of academics. I was 26 years old and had worked hard my whole life pursuing a career I no longer had interest in. It was a pity party mixed with mania, and I felt so behind compared to everyone else. I felt like I'd never catch up.<\/p><p class=\"\">Then I decided to embrace that.<\/p><p class=\"\">I would trust that slow progress is better than no progress.<\/p><p class=\"\">I decided I should learn to code. I mentally prepared for months of confusion and Lean Cuisines, signed up for a coding bootcamp (which I didn't actually finish), and started learning how to make something out of nothing. It was hard and I might have cried several times, but I knew this is what it'd take to catch up.<br \/><\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/IH-article--lean-cuisines.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">I don't even want to know how many Lean Cuisines I ate in 2015...\u00a0\ud83d\ude33<\/figcaption><\/figure><\/div><p>Even after I was able to get contract work, I still didn't have the confidence to start a business of my own. I didn't know when I'd feel ready or what kind of business I'd start, but I did know that whatever it was, I would need a marketing website.<\/p><p>I set out to find clients who needed a web developer to build a new website from scratch, but already had designers and a marketing team for me to learn from. I was looking for opportunities to get more exposure and breadth, and essentially bootstrap my education on how to become a founder.<\/p><p>Ironically, I needed\u00a0<a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_progress\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a>\u00a0to help me find jobs that aligned with my personal goals and values, but it didn't exist yet.<\/p><p class=\"\">It took longer using the old-school method of waiting and hoping for the right opportunity to come my way, but I did eventually find work that paid me to round out my skill set and gain more confidence.<br \/><\/p><p class=\"\">We all want to start at the finish line, but we can't. So start at the starting line and just focus on moving in the right direction. It doesn't matter how long it takes you.<\/p><h2>Make sacrifices now to be a founder in the future.<\/h2><p class=\"\">I needed more than just confidence to try and start a business. I needed money. I needed time.\u00a0There are tons of people who can work on passion projects after work or on the weekends, but I am not one of them. I've never been able to split my attention and focus on many things at once.<\/p><p class=\"\">As a freelancer, I knew that I could start and stop at any time, but I needed enough of a cushion so that I could work for a long stretch of uninterrupted time. People suggested I find an angel investor or raise capital, but... I didn't even have an idea yet. I needed a financial situation that would give me time to make mistakes and learn things slowly.<\/p><p class=\"\">So I decided to hustle hard, double up on client work, and save money myself.\u00a0<\/p><p class=\"\">I worked over the holidays and on weekends, and went back to\u00a0poor grad student living (I was living off that $32k\/year stipend in San Francisco not too long ago!). I stopped eating out, stopped drinking, and stopped buying things I didn't absolutely need. I even stopped taking Ubers and Lyfts and started biking everywhere.<\/p><p class=\"\">I made sacrifices early so that I could later give starting a business my best shot. It took me more than 18 months, but eventually I saved up enough money to give me about 12 months of runway.<\/p><p class=\"\">Everyone has a number of tolls to pay in order to get to where they're trying to go. It helped me to just focus on paying each of them, one at a time.<\/p><h2>Stop trying to come up with a genius idea.<\/h2><p class=\"\">I was pumped to start my journey as a first-time founder. I knew how to code, I had money saved up, and felt confident that I'd be able to persevere through tough times. However,\u00a0I was missing something pretty important.\u00a0<\/p><p class=\"\">I didn't have an idea!<\/p><p class=\"\">Even as a web developer, I didn't see the world as a sea of problems that could be solved with technology. I'd often complain about things (why is it so hard to share bookmarks, I wish Spotify let me cut songs together in a playlist, it's so annoying when I forget my passwords!), but never thought whether I could engineer solutions myself. Idea generation was a muscle I'd never worked out before.<\/p><p>So I gave myself homework to write down ideas every day, no matter how bad they were. In addition to writing down ideas, I also got in the habit of validating ideas to see if they were any good. Unfortunately, I ruled out each of them because I'd find out that someone else had already thought of it before.<br \/><\/p><p class=\"\">Then I had an incredibly important eureka moment.<\/p><p class=\"\"><b>Do not rule out ideas that there are already solutions for.\u00a0<\/b>I was listening to <a href=\"https:\/\/www.indiehackers.com\/podcast\/010-laura-roeder-of-meetedgar\" target=\"_blank\" rel=\"noopener noreferrer\">Laura Roeder<\/a>\u00a0talk about this while driving and nearly pulled over, it was such an a-ha moment for me.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/Screen%20Shot%202017-09-24%20at%205.09.49%20PM.png\" alt=\"\" \/><\/figure><\/div><p class=\"\">It is one of the most recurring lessons on Indie Hackers, but it still took me a long time to understand it. If you look through each <a href=\"https:\/\/www.indiehackers.com\/podcast\" target=\"_blank\" rel=\"noopener noreferrer\">podcast<\/a> episode and search for \"competit\" (for competitor, competition, competitive), you'll hear <a href=\"https:\/\/www.indiehackers.com\/podcast\/028-wes-bos\" target=\"_blank\" rel=\"noopener noreferrer\">Wes Bos<\/a>, <a href=\"https:\/\/www.indiehackers.com\/podcast\/008-nathan-barry-of-convertkit\" target=\"_blank\" rel=\"noopener noreferrer\">Nathan Barry<\/a>, <a href=\"https:\/\/www.indiehackers.com\/podcast\/019-todd-garland-of-buysellads\" target=\"_blank\" rel=\"noopener noreferrer\">Todd Garland<\/a>, and others talk about the <i>advantages<\/i> of tackling a problem that already has solutions.<\/p><p class=\"\">I didn't need a genius, never-before-done idea. In fact, I didn't even want one.<br \/><\/p><p class=\"\">When I told people about\u00a0<a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_idea\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a>, a product that helps engineers find jobs, everyone would name competitors. There are behemoths like Monster.com, LinkedIn, and Indeed, and at least a hundred more in the space. Dozens focus specifically on engineering recruiting and hiring. Previously, I would've given up on the idea immediately, but this time I didn't rule it out. Even in this crowded market, I still couldn't find an engineering team that shared my values.<\/p><p class=\"\">I went further in validating my idea, making sure it was something I really wanted to work on.<\/p><p class=\"\"><b>Am I personally familiar with the problem I'm solving? <\/b>Very.\u00a0I literally spent the last couple of years working as an engineer, experiencing the pains of finding jobs myself.<\/p><p class=\"\"><b>Do I know if customers will pay? <\/b>Yes! And this is part of the brilliance of not ruling out ideas that there are already solutions for. There are many competitors already making money from\u00a0companies wanting to recruit, hire, and retain engineers.<\/p><p><b>Do I know of any effective channels to reach my target audience?<\/b>\u00a0Tons. As an engineer myself, I know that engineers hang out on Twitter, bootcamp alum mailing lists, and Hacker News.<\/p><p><b>Will I enjoy working on this every day for the next 2 years?<\/b>\u00a0I <i>love<\/i> meeting new people, talking about company culture, helping people find happiness and fulfillment, and learning about organizational psychology. I get to call this work?!<\/p><p><b>Is this a winner-take-all market? <\/b>No way, Jose. It's a big pie and I only need a little slice. Nobody else has to lose in order for me to win.<\/p><p><b>If things go well, can I easily scale? <\/b>Yes, I can automate the process of onboarding teams to my website by having them write their own profiles. It shouldn't require much, if any, extra money or manpower on my part.<\/p><p><b>Can I still walk away with a win even if my idea fails?\u00a0<\/b>Absolutely. I will learn so much talking to different engineering teams and building a product from scratch. More importantly, I can <a href=\"https:\/\/medium.com\/@rrhoover\/building-a-startup-build-an-audience-first-9fbba4f1fa15\" target=\"_blank\" rel=\"noopener noreferrer\">build an audience<\/a> through <a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_idea\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a> and can take that with me even if the idea itself flops. Worst case scenario, I'm already researching which teams I'll apply to if this doesn't work out.<\/p><p class=\"\">For anyone that isn't an idea person (\ud83d\ude4b) coming up with an idea is a massive barrier. I admittedly still suck at it. But you'll immediately be better at it once you stop ruling out your non-genius ideas.<\/p><h2>Take pride in showing people the ugly behind-the-scenes.<\/h2><p>Before I started building anything for <a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_pride\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a>, I spent weeks interviewing engineers, meeting with engineering managers, getting coffees with technical recruiters, and researching personality tests, dating sites, and match-making algorithms. Meanwhile, I was keeping up my indie lurker status.\u00a0<\/p><p>I'd see other people ask for feedback about their products, and I'd think, \"Wow. I'd be so embarrassed to show anyone that!\" I was certain that my early versions would be prettier, better put together, and more thought out.<\/p><p class=\"\">I promised myself to ask for feedback at the end of each week, and every week, I'd make an excuse and push it back. I kept delaying.<\/p><p class=\"\">I spent almost <i>3 months<\/i> working in a vacuum, obsessing over things like cool hover effects without having shown my website to anyone. I didn't even know if anyone wanted what I was making! I worked alone for so long that I went a little crazy. Don't do what I did.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/IH-article--all-alone.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Don't isolate yourself like this guy either.<\/figcaption><\/figure><\/div><p class=\"\">After months of lurking, I finally got the courage to make <a href=\"https:\/\/www.indiehackers.com\/forum\/post\/-KoiK8pNNyrTxjWc3lgY\" target=\"_blank\" rel=\"noopener noreferrer\">my first post on the Indie Hackers forum<\/a>. I had 15 replies (okay, 6 were mine), but the feedback was so helpful and it sucked me out of the dark and lonely hole I had been working in.<br \/><\/p><p class=\"\">I wish I had asked for feedback sooner. I wasted time and energy building out features that no one even wanted and spent many weeks feeling lost, spinning my wheels.<\/p><p class=\"\">So what if your design sucks? Ask for design help! Not sure what your business model will be? Then ask for feedback about different pricing models! Don't be afraid to put yourself out there.<\/p><p class=\"\"><b>If you're doing it alone, you're doing it wrong.<\/b><\/p><p class=\"\">If you're stuck, show us the ugly behind-the-scenes and tell us what specifically you need help with. Fellow makers and founders can help you get unstuck.<\/p><p class=\"\">The Indie Hackers community is your tribe. It's our tribe. Everyone is working through some set of challenges and remembers exactly what it feels like to get feedback on an early idea. You might not feel ready to put yourself out there, but I promise you that your product will never be good without your doing it.<\/p><h2>Don't rely on one big launch.<\/h2><p class=\"\">I had never launched a product before and was incredibly worried that I'd mess it up. It was stressful thinking about when I'd launch, what I needed to get done before I launched, and how I'd recover if everyone who saw my product ended up hating it.<\/p><p>To get over my own anxiety, I decided to do a series of mini-launches first.\u00a0<\/p><p>My first post in the IH forum? I absolutely count that as a mini-launch. I incorporated everyone's feedback, graduated my key-values.herokuapp.com domain to keyvalues.io, and a month later,\u00a0I mini-launched again with\u00a0<a href=\"https:\/\/www.indiehackers.com\/forum\/post\/-KqZ35js1O4C41FwE_jY\" target=\"_blank\" rel=\"noopener noreferrer\">my second IH forum post<\/a>.<\/p><p>The next week, I posted in a Facebook group for Dev Bootcamp alums and then the week after that, in a stealth-mode all-women's forum.<br \/><\/p><p>Launch, get feedback, improve your product, and repeat.<\/p><p>Truth is, you don't want to go from zero traffic to 10k pageviews overnight. Thinking this way will give you unnecessary stress wrinkles, and it also makes it hard to prepare for.<\/p><p class=\"\">By removing the intimidating, anxiogenic, self-imposed deadline to launch <a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_mini\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a>, I let myself reap the benefits of the Always Be Launching lifestyle. Each mini-launch let me practice answering tough questions, fix bugs that people reported, and make improvements to the design, UX, and marketing of my website.<\/p><p>You don't have to start with one big launch, so free yourself from thinking this way.<\/p><h2>Prepare for your first big launch.<\/h2><p>I heard tons of horrors stories about failed launches: misconfigured websites, servers crashing, only getting negative feedback, and of course the worst one, having a product launch be a total dud. As a first-timer, I was sure there were other things I didn't even know to worry about!<\/p><p class=\"\">It's hard to defend against the unknown, so I decided to focus on what I did know.<\/p><p class=\"\">I set up Google Analytics and Amplitude to make sure I was tracking basic user behavior. I didn't want to drive a lot of traffic to my site and miss the opportunity to understand how people were interacting with it.<\/p><p class=\"\">I also read\u00a0<a href=\"https:\/\/levels.io\/hoodmaps\/#launch\" target=\"_blank\" rel=\"noopener noreferrer\">Pieter Levels' launch advice<\/a>\u00a0and took it to heart. <a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_launch\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a> is a static website hosted on Heroku, and I set up server-side caching using\u00a0Amazon CloudFront to make sure my server wouldn't crash.\u00a0I had identified Hacker News as being a good channel months earlier (during my validation phase), and I decided to post in \"Show HN\" where I thought I'd have a better chance of getting traction.<\/p><p class=\"\">I jumped right into the comments. I introduced myself and provided some context based on the questions I had been getting from my previous mini-launches.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/Screen%20Shot%202017-09-24%20at%2011.24.08%20AM.png\" alt=\"\" \/><\/figure><\/div><p class=\"\">I ended up getting\u00a0<a href=\"https:\/\/news.ycombinator.com\/item?id=15178283\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values on the front page of HN<\/a>\u00a0where it reached #3 and stayed on the <a href=\"http:\/\/hnrankings.info\/15178283\/\" target=\"_blank\" rel=\"noopener noreferrer\">front page for about 14 hours<\/a>!<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/hacker-news-top-3_09-05-2017.png\" alt=\"\" class=\"\" \/><\/figure><\/div><p class=\"\">Another thing I did to prepare for my HN launch was rewrite\u00a0<a href=\"https:\/\/www.keyvalues.com\/about\/?ref=indielurker\" target=\"_blank\" rel=\"noopener noreferrer\">my About page<\/a>. I knew this would be a good idea but vastly underestimated just how important it'd turn out to be.\u00a0<br \/><\/p><p class=\"\">I did some research on what makes a good About page and made sure to hit the major marks.\u00a0Open with the single sentence you want everyone to read. Include a photo of yourself. Make it personal. Provide a call to action to get in contact, subscribe, or both.<\/p><p><a href=\"https:\/\/www.keyvalues.com\/about\/?ref=indielurker\" target=\"_blank\" rel=\"noopener noreferrer\">My About page<\/a> ended up being the 2nd most visited page on Key Values.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/IH-article--about-page.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Site content behavior on Google Analytics between September 5th and 19th, 2017.<\/figcaption><\/figure><\/div><p class=\"\">There are plenty of things to worry about when doing your first big launch. My advice? Identify, analyze, and prepare as best you can. You won't be able to prepare for everything and that's okay.<\/p><p class=\"\">Two things happened that I didn't prepare for at all.<\/p><p class=\"\">First, someone posted <a href=\"https:\/\/www.producthunt.com\/posts\/key-values\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values to Product Hunt<\/a> on my behalf and I wasn't sure what to do. While it ended up on the front page all day, I wished I could have coordinated my own Product Hunt launch. (You'll see that I didn't reply to anyone because I didn't even have access to comment on my own product. \ud83d\ude02) It still feels like a missed opportunity, but hey, lesson learned!<\/p><p class=\"\">Second,\u00a0I hadn't prepared how I'd present Key Values to the press because it hadn't even crossed my mind. Matthew Hughes, a tech journalist, reached out to me over Twitter and since I didn't have a plan, I ended up forgetting some things I wanted to communicate. It ended up alright though and he published an\u00a0<a href=\"https:\/\/thenextweb.com\/dd\/2017\/09\/06\/job-site-wants-match-developers-employers-based-culture-values\/#.tnw_ezyPB6c5\" target=\"_blank\" rel=\"noopener noreferrer\">article about Key Values in The Next Web<\/a>!<\/p><p class=\"\">I prepared for my launch to the best of my ability, but looking back, can think of a few things I'd do differently if I was granted a redo. The takeaway though is that launches are a combination of preparation and luck. Focus on the things you <i>can<\/i> control and cross your fingers for everything you can't.<\/p><h2>Final thoughts about launching my first product.<\/h2><p class=\"\">I did it! I launched <a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_final\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a>. I built something I love and put it out there for other people to see and use.\u00a0It feels like a rite of passage that I wasn't sure I'd ever go through, and I couldn't be more proud or happy.<\/p><p class=\"\">When people ask me how I started Key Values, I tell them it was a side project that quickly turned into my full-time passion some time back in May. While this isn't untrue, it also makes it sound like something that happened overnight. Trust me, it didn't.<\/p><p class=\"\">In some ways, Key Values started in 2012 when I left academic research to find a more fast-paced and risky profession.<\/p><p class=\"\">It started at Homejoy (my first job at a startup) where I realized that a single line of code can be as impactful as having hundreds of 1-on-1 meetings.<\/p><p class=\"\">It started when I made yet another career change to become a web developer, because I wanted to be able to start and finish projects without relying on others.<\/p><p class=\"\">It started when I got inspired to create and innovate something of my own after months of lurking on Indie Hackers.<\/p><p class=\"\">It started when I learned how to adapt to the different working styles, goals, and values of my various clients.<\/p><p class=\"\">It started when I became comfortable asking for feedback and engaging with my new community of indie hackers, founders, and entrepreneurs.<\/p><p class=\"\"><b>No one starts right at the finish line.<\/b><\/p><p class=\"\">A lot of people are now asking me about what I'm going to do next. I've entered the post-launch trough of sorrow and the truth is that I'm still figuring out my next steps. I'm experimenting with different marketing strategies and exploring ways to turn <a href=\"https:\/\/www.keyvalues.com\/?ref=indielurker_final\" target=\"_blank\" rel=\"noopener noreferrer\">Key Values<\/a> into a revenue-generating business (I haven't started charging for my service). While I don't have all the answers right now, I'm optimistic and trust that slow progress is better than no progress.<\/p><p class=\"\">I'll let you know how it goes. \ud83d\ude09<\/p><p>If you want to get in touch, you can email me (<a href=\"mailto:lynne@keyvalues.com\" target=\"_blank\" rel=\"noopener noreferrer\">lynne@keyvalues.com<\/a>), find me on Twitter (<a href=\"https:\/\/twitter.com\/lynnetye?lang=en\" target=\"_blank\" rel=\"noopener noreferrer\">@lynnetye<\/a>,\u00a0<a href=\"https:\/\/twitter.com\/keyvaluesio?lang=en\" target=\"_blank\" rel=\"noopener noreferrer\">@keyvaluesio<\/a>), or better yet, stop lurking and leave a comment in the Indie Hackers forum!<br \/><\/p>","created_at":"2018-11-07T16:54:19+0000","updated_at":"2018-11-07T16:54:19+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":16,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/t1EK5yv5EbOdEkzkYn77pgB39a62\/@lynnetye%7Cd1042ffa5f\/IH-article--sprinklers.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":642,"title":"Chapter 3: Finding Micro-SaaS Business Ideas - Tyler Tringas","url":"https:\/\/tylertringas.com\/finding-micro-saas-business-ideas\/","content":"<p><a href=\"http:\/\/tylertringas.com\/good-micro-saas-ideas\/\">&lt;&lt; Back to\u00a0Chapter 2: What Makes a Good Micro-SaaS Idea?<\/a><\/p>\n<p>As I mentioned before, the most important part of ending up with a good Micro-SaaS idea is throwing out a lot of them. But there are a few good ways to prime the pump and look for new ideas. Here are a few methodologies that I\u2019ve developed or generalized from other Micro-SaaS entrepreneurs.<\/p>\n   <div class=\"ck_form ck_minimal\">   <div class=\"ck_form_fields\">     <h3 class=\"ck_form_title\">Micro-SaaS Idea Generation Mini-Course<\/h3>     <div class=\"ck_description\">       <p>Get 5 days of exercises in your inbox to help you come up with your own profitable Micro-SaaS business idea. Even if you've already subscribed to the newsletter drop your email in here to start the course.<\/p>     <\/div>                 <form class=\"ck_subscribe_form\" action=\"https:\/\/api.convertkit.com\/landing_pages\/118253\/subscribe\" data-remote=\"true\">       <div><input type=\"hidden\" value=\"{&quot;form_style&quot;:&quot;minimal&quot;,&quot;embed_style&quot;:&quot;inline&quot;,&quot;embed_trigger&quot;:&quot;scroll_percentage&quot;,&quot;scroll_percentage&quot;:&quot;70&quot;,&quot;delay_seconds&quot;:&quot;10&quot;,&quot;display_position&quot;:&quot;br&quot;,&quot;display_devices&quot;:&quot;all&quot;,&quot;days_no_show&quot;:&quot;15&quot;,&quot;converted_behavior&quot;:&quot;show&quot;}\" \/><input type=\"hidden\" name=\"id\" value=\"118253\" \/><input type=\"hidden\" name=\"ck_form_recaptcha\" value=\"\" \/><div class=\"ck_errorArea\">                <\/div>       <div class=\"ck_control_group ck_email_field_group\">         <label class=\"ck_label\" for=\"ck_emailField\">Email Address<\/label>         <input type=\"email\" name=\"email\" class=\"ck_email_address\" placeholder=\"Email Address\" required=\"required\" \/><\/div>       <div class=\"ck_control_group ck_captcha2_h_field_group ck-captcha2-h\">         <input type=\"text\" name=\"captcha2_h\" class=\"ck-captcha2-h\" placeholder=\"We use this field to detect spam bots. If you fill this in, you will be marked as a spammer.\" \/><\/div>                  <button class=\"subscribe_button ck_subscribe_button btn fields\">         Start Course       <\/button>                           <a class=\"ck_powered_by\" href=\"https:\/\/convertkit.com\/?utm_source=dynamic&amp;utm_medium=referral&amp;utm_campaign=poweredby&amp;utm_content=form\">Powered by ConvertKit<\/a>            <\/div><\/form>   <\/div>  <\/div>       \n<h3>Scratch your own itch<\/h3>\n<p>Eat your own dog food as they say. This is the classic example of how to build a product company and it does work. The most commonly cited example is of course 37signals, a design firm that built a project management tool, Basecamp, for their own internal needs. They realized other firms might need a similar tool and started selling it as SaaS. The software business quickly eclipsed the design business and now they can hardly be called \u201cMicro\u201d with millions of customers. Here is the <a href=\"https:\/\/medium.com\/@jasonfried\/basecamp-the-origin-story-f509fdd725f8#.rhxmmqqba\">origin story<\/a>.<\/p>\n<p>Another great example is Josh Pigford, the founder of <a href=\"https:\/\/baremetrics.com\/\">Baremetrics<\/a>. Josh was running another Micro-SaaS business and was struggling to get some basic financial metrics like Monthly Recurring Revenue out of his payment processor.he hacked together a minimal product that anybody, but first and foremost his own business, could use to instantly get SaaS metrics from their payment processor. Realizing lots of other business owners would have a similar problem, he opened it up as a product and\u00a0shot up to $8,000\/month in six months.<\/p>\n<p>This is a well tested strategy but it has two big problems at this point in time. First the idea of scratching your own itch has been around and popular for a while and second the type of people who build software businesses tend to be somewhat similar in their needs. There are only so many itches to scratch for a software entrepreneur and at this point a lot of them have been sufficiently scratched. Slightly paradoxically, this is in fact part of the reason why now is such a good time to build a Micro-SaaS business. Scratch your own itch has been around so long, that for almost everything someone trying to build a software business needs, there are 15 other software businesses selling pretty great solutions for it. Once you hit on an original idea, you can often use those other itch-scratchers\u2019 off the shelf solutions rather than having to build your own internal tools.<\/p>\n<p>Pieter Levels, the creator of <a href=\"https:\/\/nomadlist.com\/\">Nomad List<\/a>, has a good solution to the over-saturation of scratch your own itch. In his book <a href=\"https:\/\/makebook.io\">MAKE<\/a>\u00a0he suggests\u00a0that if you live an interesting or unique life, you will generate new and unique itches to scratch. In his case he was living a digital nomad lifestyle, traveling around the world in a backpack while building a new startup every month. He needed to figure out the next best place to live each month comparing data like cost of living and internet speed for cities all over the world. So he scratched his own itch, crowd-sourcing a basic dataset and built a site to help other digital nomads connect and solve the problem.<\/p>\n<p>If something about your life is really unique, then you can use <em>scratch your own itch<\/em> in that area. If you are a freelance graphic designer in San Francisco who likes yoga and green\u00a0juice, you\u2019re going to need a different strategy.<\/p>\n<h3>Niche versions of existing SaaS<\/h3>\n<p>A lot of big funded SaaS startups have built generalized software that many different types of businesses are using\u00a0for their own specific needs. These SaaS businesses have to target a wide audience because of the shear volume of sales their company size and valuations require. Looking at generalized products like Salesforce, Quickbooks, Basecamp, Freshbooks, Shopify and WordPress can be a good source of ideas for niche SaaS products.<\/p>\n<p>A great example is <a href=\"https:\/\/www.mindbodyonline.com\/yoga\">Mindbody<\/a>, which is custom SaaS for yoga studios. Everything about the app is customized specifically for this niche: online class schedules, tracking teachers, accepting payments and subscriptions. A\u00a0yoga studio could probably\u00a0cobble together all of those features from a few existing big SaaS products but when there is a single product that works for them out of the box it easily clears the 5x better hurdle. And I think you would be surprised at how low the bar is here. Even something as simple a building a simple CRM with data fields and workflows pre-built for a particular industry could be enough of a UX improvement to sell well.<\/p>\n<p>Haven\u2019t seen\u00a0any great examples yet for this, but I think there is a big opportunity in just building a simple CRM for specific industries. Even if it\u2019s less powerful that Salesforce, a simple contact and pipeline tool with workflows and a data model (even just having the right words and terminology built in) could sell well enough to be a healthy Micro-SaaS. Salesforce for wedding photographers. Freshbooks for dentists. Play around with combinations until something passes the Meat Grinder.<\/p>\n<h3>Repeatable freelance and consulting\u00a0work<\/h3>\n<p>SaaS, and software more generally, has been stealing work from freelancers since the beginning. Nobody wants to be in the business of reinventing wheels,\u00a0so if you see a specific type of job that businesses are paying contractors to do over and over again from scratch, there may be an opportunity to build a SaaS solution. This was the method that generated the idea for Storemapper. There was no good plug-and-play solution for a store locator app on the market and several of my own clients asked me to build them one. I could have made something from scratch and sent them a bill for $1,000 \u2013 $2,000 each but it was a broad enough need to warrant investigation as a Micro-SaaS. Another thing to look for is a big change in complexity. I did some research into other store locators and found that companies with a store locator would have a relatively simple site, something easily built without a developer on Squarespace, WordPress or Shopify, but the store locator would be a big step change in complexity. So I reasoned that you might have a lot of customers who could launch their entire site without a developer, except for the store locator. Great opportunity.<\/p>\n<p>The best way to sniff out these kinds of opportunities is to actually become a consultant in a particular field and really learn what customers need.<\/p>\n<h3>Automate manual tasks<\/h3>\n<p>Sometimes you can use the sections from Chapter 2 on what makes a great Micro-SaaS idea, to generate new ideas. One of the best attributes of a Micro-SaaS business is automation. So you can start from there and look around for things that could be automated, and then test to see if they would be good overall businesses.<\/p>\n<p>A great example is <a href=\"https:\/\/stunning.co\/\">Stunning <\/a>\u00a0which\u00a0is an app that handles <em>dunning<\/em>, an odd term for the painful process of following up with customers whose credit card on file has expired and getting them to update it. Without Stunning you are forced manually check this each customer, setup an alert if their regular monthly payment fails, and hound them until they update their payment info. With Storemapper I did this by hand for the first year. It was awful and I was not very good at keeping up with it until one day I realized fully 20% of my \u201cpaying\u201d customers were not actually paying me due to some kind of billing issue. Stunning plugs directly into Stripe and automatically starts sending follow-up emails and in-app messages to customers until they update their account. It\u2019s a great business model.<\/p>\n<h3>Build upon growing platforms<\/h3>\n<p>Like automating manual tasks, this source of ideas works hand in hand with what makes a good business model. A majority of the successful Micro-SaaS businesses I have encountered are plugged in to or enhance a rapidly growing platform. Baremetrics and Stunning built on Stripe a massively popular new payment processor. Mailparser, a Micro-SaaS for parsing data out of emails, got many of its early clients through Zapier a growing business automation tool.<\/p>\n<p>We\u2019ll discuss this platform strategy in more detail in a following chapter on acquiring customers but again you can work backwards for a source of ideas. Keep a look out for rapidly growing platforms, check their forums and Twitter mentions to see if there is one big feature or use case that many users are clamoring for but the company doesn\u2019t seem likely to build themselves.<\/p>\n<p>The best version of this is a platform bolt-on that can use that integration as a beachhead, but isn\u2019t too tightly interwoven with that one platform. If you can also add value as a standalone product <em>and<\/em> integrate with other competing platforms, you have an even better idea.<\/p>\n<p><a href=\"http:\/\/tylertringas.com\/shipping-a-saas-minimum-viable-product\/\">Next\u00a0Chapter 4: From Idea to Minimum Viable Product &gt;&gt;<\/a><\/p>\n <div class=\"ck_form_container ck_inline\" data-ck-version=\"6\">    <div class=\"ck_form ck_vertical_subscription_form\">   <div class=\"ck_form_content\">     <h3 class=\"ck_form_title\">Building Micro-SaaS Businesses<\/h3>     <div class=\"ck_description\">                <img src=\"https:\/\/convertkit.s3.amazonaws.com\/subscription_forms\/images\/004\/813\/077\/standard\/book.png?1423213667\" alt=\"Book\" \/><p>All the stuff I did wrong building a six figure Micro-SaaS business and how to do it better yourself. Anecdotes, pseudo-philosophy and self-deprecation.<\/p>     <\/div>   <\/div>    <div class=\"ck_form_fields\">                <form class=\"ck_subscribe_form\" action=\"https:\/\/forms.convertkit.com\/landing_pages\/1427\/subscribe\" data-remote=\"true\">       <div><input type=\"hidden\" value=\"{&quot;embed_style&quot;:&quot;inline&quot;,&quot;embed_trigger&quot;:&quot;scroll_percentage&quot;,&quot;scroll_percentage&quot;:&quot;70&quot;,&quot;delay_seconds&quot;:&quot;10&quot;,&quot;display_position&quot;:&quot;br&quot;,&quot;display_devices&quot;:&quot;all&quot;,&quot;days_no_show&quot;:&quot;15&quot;,&quot;form_style&quot;:&quot;full&quot;,&quot;converted_behavior&quot;:&quot;show&quot;}\" \/><input type=\"hidden\" name=\"id\" value=\"1427\" \/><input type=\"hidden\" name=\"ck_form_recaptcha\" value=\"\" \/><div class=\"ck_errorArea\">                <\/div>       <div class=\"ck_control_group ck_first_name_field_group\">         <label class=\"ck_label\" for=\"ck_firstNameField\">First Name<\/label>         <input type=\"text\" name=\"first_name\" class=\"ck_first_name\" id=\"ck_firstNameField\" \/><\/div>       <div class=\"ck_control_group ck_email_field_group\">         <label class=\"ck_label\" for=\"ck_emailField\">Email Address<\/label>         <input type=\"email\" name=\"email\" class=\"ck_email_address\" required=\"required\" \/><\/div>       <div class=\"ck_control_group ck_captcha2_h_field_group ck-captcha2-h\">         <label class=\"ck_label\" for=\"ck_captcha2_h\">We use this field to detect spam bots. If you fill this in, you will be marked as a spammer.<\/label>         <input type=\"text\" name=\"captcha2_h\" class=\"ck-captcha2-h\" \/><\/div>                  <button class=\"subscribe_button ck_subscribe_button btn fields\">         Subscribe For New Chapters       <\/button>                           <a class=\"ck_powered_by\" href=\"https:\/\/convertkit.com\/?utm_source=dynamic&amp;utm_medium=referral&amp;utm_campaign=poweredby&amp;utm_content=form\">Powered by ConvertKit<\/a>            <\/div><\/form>   <\/div>  <\/div>  <\/div>","created_at":"2018-11-07T16:54:19+0000","updated_at":"2018-11-07T16:54:19+0000","published_at":"2016-10-28T21:27:34+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"tylertringas.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":643,"title":"Clojure for the Brave and True","url":"https:\/\/www.braveclojure.com\/foreword\/","content":"<p class=\"BodyFirst\">As you read this hilarious book, you will at some point experience a very serious moment. It is the moment you admit to yourself that programming is more enjoyable after knowing some Clojure. It is also the moment that your investment in this book, in money and time, comes back to you\u2014with interest.<\/p><p class=\"Body\">Humor has a certain relationship to seriousness. It is appropriate to\u00a0joke about serious things, but only after the right amount of time has passed. For example, it took years for me to be able to crack a smile when I remember my favorite uncle\u2019s last words: \u201cHold my beer.\u201d<\/p><p class=\"Body\">This book works in the opposite way. It points out really funny things for the right amount of time <em>before<\/em>, and perhaps even <em>during<\/em>, the serious event\u2014that moment you realize you enjoy programming more because of Clojure. It does this without obscuring the deep, technical aspects of Clojure programming that you will learn.<\/p><p class=\"Body\">This approach is refreshing because most of the programming books I\u2019ve read are drier than a camel\u2019s fart. We are fortunate that Daniel is a brilliant programmer and writer and that his wife Jess is an equally brilliant illustrator. We are especially fortunate that they both went insane and decided to write a book at exactly the same time.<\/p><p class=\"Body\">Clojure is the topic of this book, but in a way it\u2014or perhaps its creator, Rich Hickey\u2014is also one of the authors, since Clojure is the most elegant programming language ever designed. Like the concept of brunch, Clojure is so elegant that it\u2019s difficult to tell anyone anything about it without somehow improving them.<\/p><p class=\"Body\">Elegance is a quality regularly ascribed to many dialects in the family of programming languages known collectively as Lisp, of which Clojure is one. All Lisps descend from a set of simple and beautiful discoveries made by the mathematician John McCarthy in 1958.<\/p><p class=\"Body\">Since 1958, there have been many Lisps and Lisp books. There are many more Lisps and books to come. As artifacts of the past and future, each are right for the unique combination of constraints and desires faced and fancied by their authors, in their respective times.<\/p><p class=\"Body\">I find Clojure, and this particular book about it, especially right for the present. I hope you will too.<\/p><p class=\"BodyFirst para-style-override-1\">Alan Dipert<\/p>","created_at":"2018-11-07T16:54:19+0000","updated_at":"2018-11-07T16:54:19+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.braveclojure.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":644,"title":"All the Linear Algebra You Need for AI","url":"https:\/\/github.com\/fastai\/fastai\/blob\/master\/tutorials\/linalg_pytorch.ipynb","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:54:19+0000","updated_at":"2018-11-07T16:54:19+0000","annotations":[],"reading_time":0,"domain_name":"github.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":645,"title":"The Weird and Wonderful Characters of Clojure","url":"https:\/\/yobriefca.se\/blog\/2014\/05\/19\/the-weird-and-wonderful-characters-of-clojure\/","content":"<blockquote class=\"warning\"><div>This ancient article has been superseded by a <a href=\"https:\/\/clojure.org\/guides\/weird_characters\">much better and up to date article on the Clojure website<\/a><\/div><\/blockquote><blockquote><p>A reference collection of characters used in Clojure that are difficult to \"google\". Descriptions sourced from various blogs, <a href=\"http:\/\/stackoverflow.com\/questions\/tagged\/clojure\">StackOverflow<\/a>, <a href=\"http:\/\/en.wikibooks.org\/wiki\/Learning_Clojure\">Learning Clojure<\/a> and the <a href=\"http:\/\/clojure.org\/documentation\">official Clojure docs<\/a> - sources attributed where necessary. Type the symbols into the box below to search (or use <code>CTRL-F<\/code>). Sections not in any particular order but related items are grouped for ease. If I'm wrong or missing anything worthy of inclusion tweet me <a href=\"http:\/\/twitter.com\/kouphax\">@kouphax<\/a> or mail me at <a href=\"https:\/\/yobriefca.se\/cdn-cgi\/l\/email-protection#e389828e8690a39a8c81918a86858082cd9086\">[email\u00a0protected]<\/a>.<\/p>\n<\/blockquote><h2><code>#<\/code> - Dispatch macro<\/h2><p>You'll see this macro character beside another e.g. <code>#(<\/code> or <code>#\"<\/code>. This topic will act as a bit preamble before looking at your specific case.<\/p><p><code>#<\/code> is the dispatch macro, a reader macro that tells the Clojure reader (the thing that takes a file of Clojure text and parses it for consumption in the compiler) to go and look at another <strong>read table<\/strong> for the definition of the next character - in essence this allows extending default reader behaviour.<\/p><p>Clojure doesn't provide support for creating reader macros but it is possible through <a href=\"http:\/\/briancarper.net\/blog\/449\/\">a bit of hackery<\/a>.<\/p><p>If you see <code>#<\/code> <strong>at the end<\/strong> of a symbol then this is used to automatically generate a new symbol. This is useful inside macros to keep macro specifics from leaking into the userspace. A regular <code>let<\/code> will fail in a macro definition<\/p>\n<pre class=\"clojure\">user=&gt; (defmacro m [] `(let [x 1] x))\n#'user\/m\nuser=&gt; (m)\nCompilerException java.lang.RuntimeException: Can't let qualified name: user\/x, compiling:(NO_SOURCE_PATH:1)\n<\/pre><p>Instead you need to append <code>#<\/code> to the end of the variable name and let Clojure generate a unique symbol for it.<\/p>\n<pre class=\"clojure\">user=&gt; (defmacro m [] `(let [x# 1] x#))\n#'user\/m\nuser=&gt; (m)\n1\nuser=&gt;\n<\/pre><p>If we expand this macro we can see the <code>gensym<\/code>'d name<\/p>\n<pre class=\"clojure\">user=&gt; (macroexpand '(m))\n(let* [x__681__auto__ 1] x__681__auto__)\n<\/pre><p>Another place you'll see the <code>#<\/code> is in <a href=\"http:\/\/clojure.org\/reader#The%20Reader--Tagged%20Literals\">tagged literals<\/a>. Most commonly you'll see this use in <a href=\"https:\/\/github.com\/edn-format\/edn\">EDN<\/a> (extensible data notation - a rich data format that can be used in Clojure) and in ClojureScript (<code>#js<\/code>). Search for <code>#inst<\/code>, <code>#uuid<\/code> or <code>#js<\/code> for some more info.<\/p>\n<ul><li><a href=\"http:\/\/clojure.org\/reader\">Clojure Documentation - Reader<\/a><\/li>\n  <li><a href=\"http:\/\/briancarper.net\/blog\/449\/\">Clojure Reader Macros<\/a><\/li>\n  <li><a href=\"http:\/\/clojuredocs.org\/clojure_core\/clojure.core\/gensym\">ClojureDocs - gensym<\/a><\/li>\n<\/ul><hr \/><h2><code>#{<\/code> - Set macro<\/h2><p>See the dispatch (<code>#<\/code>) macro for additional details.<\/p><p><code>#{<\/code> defines a set (a collection of unique values) specifically a <code>hash-set<\/code>. The following are equivalent,<\/p>\n<pre class=\"clojure\">user=&gt; #{1 2 3 4}\n#{1 2 3 4}\nuser=&gt; (hash-set 1 2 3 4)\n#{1 2 3 4}\n<\/pre><p>Attempting to create a <code>set<\/code> using this literal form will throw if there are duplicates. Instead the <code>hash-set<\/code> function should be used on a vector<\/p>\n<pre class=\"clojure\">user=&gt; #{1 2 3 4 1}\nIllegalArgumentException Duplicate key: 1  clojure.lang.PersistentHashSet.createWithCheck (PersistentHashSet.java:68)\nuser=&gt; (set [1 2 3 4 1]) ; convert vector to set, removing duplicates\n#{1 2 3 4}\n<\/pre>\n<ul><li><a href=\"http:\/\/clojure.org\/data_structures#Data%20Structures-Sets\">Clojure Documentation: Sets<\/a><\/li>\n<\/ul><hr \/><h2><code>#_<\/code> - Discard macro<\/h2><p>See the dispatch (<code>#<\/code>) macro for additional details.<\/p><p><code>#_<\/code> tells the reader to ignore the next form completely.<\/p>\n<pre class=\"clojure\">user=&gt; [1 2 3 #_ 4 5]\n[1 2 3 5]\n<\/pre><p>The docs suggest that \"The form following <code>#_<\/code> is completely skipped by the reader. (This is a more complete removal than the comment macro which yields nil).\". This can prove useful for debugging situations or for multline comments. I've never used it.<\/p>\n<ul><li><a href=\"http:\/\/clojure.org\/reader\">Clojure Documentation - Reader<\/a><\/li>\n<\/ul><hr \/><h2><code>#\"<\/code> - Regular Expression macro<\/h2><p>See the dispatch (<code>#<\/code>) macro for additional details.<\/p><p><code>#\"<\/code> indicates the start of a regular expression pattern.<\/p>\n<pre class=\"clojure\">user=&gt; (re-matches #\"^test$\" \"test\")\n\"test\"\n<\/pre><p>This form is compiled at <em>read time<\/em> into a <code>java.util.regex.Pattern<\/code>.<\/p>\n<ul><li><a href=\"http:\/\/clojure.org\/other_functions#Other%20Useful%20Functions%20and%20Macros-Regex%20Support\">Clojure Documentation: Regex Support<\/a><\/li>\n<\/ul><hr \/><h2><code>#(<\/code> - Function macro<\/h2><p>See the dispatch (<code>#<\/code>) macro for additional details.<\/p><p><code>#(<\/code> begins the short hand syntax for an inline function definition. The following 2 bits of code are similar,<\/p>\n<pre class=\"clojure\">; anonymous function takin a single argument and printing it\n(fn [line] (println line)) ;\n; anonymous function takin a single argument and printing it - shorthand\n#(println %)\n<\/pre><p>The macro expands the shorthand syntax into a function definition whose arity (the number of arguments it takes) is defined by how the <code>%<\/code> placeholders are declared. See the <code>%<\/code> character for discussion around arity.<\/p>\n<pre class=\"clojure\">user=&gt; (macroexpand `#(println %))\n(fn* [arg] (clojure.core\/println arg)) ; argument names shortened for clarity\n<\/pre>\n<hr \/><h2><code>#'<\/code> - Var macro<\/h2><p><code>#'<\/code> is the var quote. It is the same a the <code>var<\/code> method,<\/p>\n<pre class=\"clojure\">user=&gt; (def nine 9)\n#'user\/nine\nuser=&gt; nine\n9\nuser=&gt; (var nine)\n#'user\/nine\nuser=&gt; #'nine\n#'user\/nine\n<\/pre><p>When used it will attempt to return the referenced var. This is useful when you want to talk about the reference\/declaration instead of the value it represents. See the use of <code>meta<\/code> in the metadata (<code>^<\/code>) discussion.<\/p>\n<ul><li><a href=\"http:\/\/clojure.org\/special_forms#var\">Clojure Official Documentation: Special Forms<\/a><\/li>\n<\/ul><hr \/><h2><code>#inst<\/code>, <code>#uuid<\/code> &amp; <code>#js<\/code> etc. - tagged literals<\/h2><p>Commonly found in EDN and ClojureScript this use of <code>#<\/code> is called the tagged literal. Look at this example,<\/p>\n<pre class=\"clojure\">user=&gt; (java.util.Date.)\n#inst \"2014-05-19T19:12:37.925-00:00\"\n<\/pre><p>When we create a new date it is represented as a tagged literal, or in this case a tagged string. We can use Clojures <code>read-string<\/code> to read this back (or use it directly)<\/p>\n<pre class=\"clojure\">user=&gt; (type #inst \"2014-05-19T19:12:37.925-00:00\")\njava.util.Date\n(read-string \"#inst \\\"2014-05-19T19:12:37.925-00:00\\\"\")\n#inst \"2014-05-19T19:12:37.925-00:00\"\nuser=&gt; (type (read-string \"#inst \\\"2014-05-19T19:12:37.925-00:00\\\"\"))\njava.util.Date\n<\/pre><p>A tagged literal tells the reader how to parse the literal value. Other common uses include <code>#uuid<\/code> for generating UUIDs and in the ClojureScript world an extremely common use of tagged literals is <code>#js<\/code> which can be used to convert ClojureScript data structures into JavaScript structures directly.<\/p>\n<ul><li><a href=\"https:\/\/github.com\/edn-format\/edn#tagged-elements\">EDN Tagged Elements<\/a><\/li>\n<\/ul><hr \/><h2><code>%<\/code> - Argument placeholder<\/h2><p><code>%<\/code> is not a macro but a placeholder for use in the <code>#(<\/code> macro. It represents an argument that will be passed into the function when it is expanded.<\/p>\n<pre class=\"clojure\">user=&gt; (macroexpand `#(println %))\n(fn* [arg] (clojure.core\/println arg)) ; takes a single arg, uses it once\nuser=&gt; (macroexpand `#(println % %))\n(fn* [arg] (clojure.core\/println arg arg)) ; takes a single arg, uses it twice\n<\/pre><p>Numbers can be placed directly after the <code>%<\/code> to indicate the arguments position. Numbers are also used by the <code>#(<\/code> macro to determine the number of arguments to pass in.<\/p>\n<pre class=\"clojure\">user=&gt; (macroexpand `#(println %1 %2))\n(fn* [arg1 arg2] (clojure.core\/println arg1 arg2)) ; takes 2 args\nuser=&gt; (macroexpand `#(println %4))\n(fn* [arg1 arg2 arg3 arg4] (clojure.core\/println arg4)) ; takes 4 args doesn't use 3\n<\/pre><p>So you don't have to use the arguments but you do need to declare them in the order you'd expect an external caller to pass them in.<\/p><p><code>%<\/code> and <code>%1<\/code> can be used interchangably,<\/p>\n<pre class=\"clojure\">user=&gt; (macroexpand `#(println % %1)) ; use both % and %1\n(fn* [arg1] (clojure.core\/println arg1 arg1)) ; still only takes 1 argument\n<\/pre>\n<hr \/><h2><code>@<\/code> - Deref macro<\/h2><p><code>@<\/code> is the deref macro, it is the shorthand equivalent of the <code>deref<\/code> function so these 2 forms are the same,<\/p>\n<pre class=\"clojure\">user=&gt; (def x (atom 1))\n#'user\/x\nuser=&gt; @x\n1\nuser=&gt; (deref x)\n1\nuser=&gt;\n<\/pre><p><code>@<\/code> is used to get the current value of a reference. The above example uses <code>@<\/code> to get the current value of an <a href=\"http:\/\/clojure.org\/atoms\">atom<\/a> but <code>@<\/code> can be applied to other things such as <code>future<\/code>s, <code>delay<\/code>s, <code>promise<\/code>s etc. to force computation and potentially block.<\/p>\n<hr \/><h2><code>^<\/code> - Metadata<\/h2><p><code>^<\/code> is the metadata marker. Metadata is a map of values (with shorthand option) that can be attached to various forms in Clojure. This provides extra information for these forms and can be used for documentation, compilation warnings, typehints and other features.<\/p>\n<pre class=\"clojure\">user=&gt; (def ^{ :debug true } five 5) ; meta map with single boolean value\n#'user\/five\n<\/pre><p>We can access the metadata by the <code>meta<\/code> method which should be executed against the declaration itself (rather than the returned value).<\/p>\n<pre class=\"clojure\">user=&gt; (def ^{ :debug true } five 5)\n#'user\/five\nuser=&gt; (meta #'five)\n{:ns #&lt;Namespace user&gt;, :name five, :column 1, :debug true, :line 1, :file \"NO_SOURCE_PATH\"}\n<\/pre><p>As we have a single value here we can use a shorthand notation for declaring the metadata <code>^:name<\/code> which is useful for flags as the value will be set to true.<\/p>\n<pre class=\"clojure\">user=&gt; (def ^:debug five 5)\n#'user\/five\nuser=&gt; (meta #'five)\n{:ns #&lt;Namespace user&gt;, :name five, :column 1, :debug true, :line 1, :file \"NO_SOURCE_PATH\"}\n<\/pre><p>Another use of <code>^<\/code> is for type hints. These are used to tell the compiler what type the value will be and allow it to perform type specific optimisations thus potentially making resultant code a bit faster.<\/p>\n<pre class=\"clojure\">user=&gt; (def ^Integer five 5)\n#'user\/five\nuser=&gt; (meta #'five)\n{:ns #&lt;Namespace user&gt;, :name five, :column 1, :line 1, :file \"NO_SOURCE_PATH\", :tag java.lang.Integer}\n<\/pre><p>We can see in that example the <code>:tag<\/code> property is set.<\/p><p>You can also stack the shorthand notations,<\/p>\n<pre class=\"clojure\">user=&gt; (def ^Integer ^:debug ^:private five 5)\n#'user\/five\nuser=&gt; (meta #'five)\n{:ns #&lt;Namespace user&gt;, :name five, :column 1, :private true, :debug true, :line 1, :file \"NO_SOURCE_PATH\", :tag java.lang.Integer}\n<\/pre>\n<ul><li><a href=\"http:\/\/clojure.org\/metadata\">Clojure Official Documentation: Metadata<\/a><\/li>\n  <li><a href=\"http:\/\/en.wikibooks.org\/wiki\/Learning_Clojure\/Meta_Data\">Learning Clojure: Meta Data<\/a><\/li>\n<\/ul><hr \/><h2><code>'<\/code> - Quote macro<\/h2><p>Can be used against symbols as part of a dispatch macro (see <code>#'<\/code>). Also used to quote forms and prevent their evaluation as with the <code>quote<\/code> function.<\/p>\n<pre class=\"clojure\">user=&gt; (1 3 4) ; fails as it tries to evaluate 1 as a function\nClassCastException java.lang.Long cannot be cast to clojure.lang.IFn  user\/eval925 (NO_SOURCE_FILE:1)\nuser=&gt; '(1 3 4) ; quote\n(1 3 4)\nuser=&gt; (quote (1 2 3)) ; using the longer quote method\n(1 2 3)\nuser=&gt;\n<\/pre>\n<ul><li><a href=\"http:\/\/clojure.org\/special_forms#quote\">Clojure Official Documentation<\/a><\/li>\n<\/ul><hr \/><h2><code>;<\/code> - Comment<\/h2><p><code>;<\/code> is a comment. In fact its a comment <strong>macro<\/strong> that takes all input from its starting point to the end of the line and ensures the reader ignore it.<\/p>\n<pre class=\"clojure\">user=&gt; (def x \"x\") ; this is a comment\n#'user\/x\nuser=&gt; ; this is a comment too\n&lt;returns nothing&gt;\n<\/pre>\n<hr \/><h2><code>:<\/code> - Keyword<\/h2><p><code>:<\/code> is the indicator for a Keyword which is an interned string that provides fast comparison and lower memory overhead.<\/p>\n<pre class=\"clojure\">user=&gt; (type :test)\nclojure.lang.Keyword\n<\/pre><p>Alternativley you can use <code>keyword<\/code> to create a keyword from a string<\/p>\n<pre class=\"clojure\">user=&gt; (keyword \"test\")\n:test\n<\/pre><p>A neat thing about keywords is they also implement <code>IFn<\/code> and can act as functions for extracting values from maps which is very nice.<\/p>\n<pre class=\"clojure\">user=&gt; (def my-map { :one 1 :two 2 })\n#'user\/my-map\nuser=&gt; (:one my-map) ; get the value for :one by invoking it as function\n1\nuser=&gt; (:three my-map) ; it can safely access non-keys\nnil\nuser=&gt; (:three my-map 3) ; it can return a default if specified\n3\n<\/pre>\n<ul><li><a href=\"http:\/\/clojure.org\/data_structures#Data%20Structures-Keywords\">Clojure Official Documentation<\/a><\/li>\n<\/ul><hr \/><h2><code>::<\/code> - Qualified keyword<\/h2><p><code>::<\/code> is used to fully qualify a keyword with the current namespace.<\/p>\n<pre class=\"clojure\">user=&gt; :my-keyword\n:my-keyword\nuser=&gt; ::my-keyword\n:user\/my-keyword\nuser=&gt; (= ::my-keyword :my-keyword)\nfalse\n<\/pre><p>I have found this useful when creating macros. If I want to ensure a macro, that calls another method in the macro namespace, correctly expands to call the method I have used ::my-method to refer to the fully qualified name.<\/p>\n<ul><li><a href=\"http:\/\/stackoverflow.com\/questions\/5771168\/what-is-the-used-for-in-clojure\">What is the :: used for in clojure?<\/a><\/li>\n<\/ul><hr \/><h2><code>\/<\/code> - Namespace separator<\/h2><p>Can be the division function <code>\/<\/code> but can also act as a separator in a symbol name to break apart the symbol name and the namespace it resides in <code>my-namepace\/utils<\/code>. This allows symbols to be fully qualified to prevent collisions or spread.<\/p>\n<ul><li><a href=\"http:\/\/clojure.org\/reader\">Cloure Official Documentation<\/a><\/li>\n<\/ul><hr \/><h2><code>$<\/code> - Inner class reference<\/h2><p>Used to reference inner classes and interfaces in Java. Seperates the container class name and the inner class name,<\/p>\n<pre class=\"clojure\">(:import (basex.core BaseXClient$EventNotifier)\n(defn- build-notifier [notifier-action]\n  (reify BaseXClient$EventNotifier\n    (notify [this value]\n      (notifier-action value))))\n<\/pre><p><code>EventNotifier<\/code> is an inner interface of the <code>BaseXClient<\/code> class which is an imported Java class.<\/p>\n<ul><li><a href=\"http:\/\/blog.jayfields.com\/2011\/01\/clojure-using-java-inner-classes.html\">Clojure: Using Java Inner Classes<\/a><\/li>\n<\/ul><hr \/><h2><code>-&gt; -&gt;&gt; some-&gt; cond-&gt; as-&gt;<\/code> etc. - Threading macros<\/h2><p>These are threading macros. Almost all of them take an initial value and <strong>thread<\/strong> this value through a number of forms. Lets imagine (for reasons unknown) we wanted to take a number, find the square root, cast it to an int, then a string then back to an integer again. We could write it like this,<\/p>\n<pre class=\"clojure\">user=&gt; (Integer. (str (int (Math\/sqrt 25))))\n5\n<\/pre><p>The threading macro allows us to unravel this deep nesting,<\/p>\n<pre class=\"clojure\">user=&gt; (-&gt; 25 (Math\/sqrt) int str Integer.)\n5\n<\/pre><p>Or if you prefer multiline and consistent brackettering<\/p>\n<pre class=\"clojure\">(-&gt; 25\n    (Math\/sqrt)\n    (int)\n    (str)\n    (Integer.))\n<\/pre><p>What the macro does is take the value returned from each expression and push it in as the first argument to the next one.<\/p><p><code>-&gt;&gt;<\/code> is the same but different. Rather than push the last value in as the <strong>first<\/strong> argument it passes it in as the <strong>last<\/strong> argument.<\/p><p>The \"etc.\" in the title refers to the fact there are a whole host of threading macros that perform variations on the same theme (<code>cond-&gt;<\/code>, <code>some-&gt;<\/code>, <code>as-&gt;<\/code> and their <code>-&gt;&gt;<\/code> equivalents). There is also an entire library, <a href=\"https:\/\/github.com\/rplevy\/swiss-arrows\">swiss-arrows<\/a>, dedicated to the threading macros.<\/p>\n<ul><li><a href=\"http:\/\/blog.fogus.me\/2009\/09\/04\/understanding-the-clojure-macro\/\">Understanding the Clojure -&gt; macro<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/67571d1844e7b9a0cab6089245d7e5cde208c67e\/changes.md\">Clojure Changelog<\/a><\/li>\n<\/ul><hr \/><h2><code>~<\/code> - Unquote macro<\/h2><p>See <code>`<\/code> (syntax quote) for additional information.<\/p><p><code>~<\/code> is unquote. That is within as syntax quoted (<code>`<\/code>) block <code>~<\/code> will <strong>unquote<\/strong> the associated symbol i.e. resolve it in the current context.<\/p>\n<pre class=\"clojure\">user=&gt; (def five 5) ; create a named ref representing the number 5\n#'user\/five\nuser=&gt; five ; five will yeild its internal value\n5\nuser=&gt; `five ; syntax quoting five will fully resolve the SYMBOL\nuser\/five\nuser=&gt; `~five ; within a syntax quoted block ~ wil resolve the value in the current context\n5\n<\/pre><p>This forms the meat and potatoes of creating macros which are, to be highly reductionist, functions that return blocks of syntax with parts evaluated in varying contexts.<\/p>\n<ul><li><a href=\"http:\/\/www.braveclojure.com\/writing-macros\/\">Clojure for the Brave and True - Writing Macros<\/a><\/li>\n  <li><a href=\"http:\/\/aphyr.com\/posts\/305-clojure-from-the-ground-up-macros\">Clojure from the ground up: macros<\/a><\/li>\n  <li><a href=\"http:\/\/clojure.org\/macros\">Clojure Official Documentation<\/a><\/li>\n<\/ul><hr \/><p>See <code>`<\/code> (syntax quote) and <code>~<\/code> (unquote) for additional information.<\/p><p><code><a href=\"https:\/\/yobriefca.se\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"730d33\">[email\u00a0protected]<\/a><\/code> is unquote-splicing. Where unquote (<code>~<\/code>) deals with single values (or treats its attached item as a single item) <code><a href=\"https:\/\/yobriefca.se\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"126c52\">[email\u00a0protected]<\/a><\/code> works on lists and expands them out into multiple statements. Think JavaScripts <code>.apply<\/code> method that takes an array and expands it out as arguments to the applied function.<\/p>\n<pre class=\"clojure\">user=&gt; (def three-and-four (list 3 4))\n#'user\/three-and-four\nuser=&gt; `(1 ~three-and-four) ; treates as a single statement produces a nested list\n(1 (3 4))\nuser=&gt; `(1 <a href=\"https:\/\/yobriefca.se\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"4d330d\">[email\u00a0protected]<\/a>three-and-four) ; expand out as seperate statements\n(1 3 4)\n<\/pre><p>Again this gives us a lot of power in macros.<\/p>\n<ul><li><a href=\"http:\/\/www.braveclojure.com\/writing-macros\/\">Clojure for the Brave and True - Writing Macros<\/a><\/li>\n  <li><a href=\"http:\/\/aphyr.com\/posts\/305-clojure-from-the-ground-up-macros\">Clojure from the ground up: macros<\/a><\/li>\n  <li><a href=\"http:\/\/clojure.org\/macros\">Clojure Official Documentation<\/a><\/li>\n<\/ul><hr \/><h2><code>`<\/code> - Syntax quote<\/h2><p>See <code><a href=\"https:\/\/yobriefca.se\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"106e50\">[email\u00a0protected]<\/a><\/code> (unquote splicing) and <code>~<\/code> (unquote) for additional information.<\/p><p><code>`<\/code> is the syntax quote. When used on a symbol it resolves the symbol in the current context,<\/p>\n<pre class=\"clojure\">user=&gt; (def five 5)\n#'user\/five\nuser=&gt; `five\nuser\/five\n<\/pre><p>When used with lists (remember every thing in Clojure is data) it forms a <strong>template<\/strong> for the data structure and won't immediately resolve it.<\/p>\n<pre class=\"clojure\">user=&gt; (1 2 3)\nClassCastException java.lang.Long cannot be cast to clojure.lang.IFn  user\/eval832 (NO_SOURCE_FILE:1)\nuser=&gt; `(1 2 3)\n(1 2 3)\n<\/pre><p>You'll see this most often in the context of macros. We can write one now,<\/p>\n<pre class=\"clojure\">user=&gt; (defmacro debug [body]\n  #_=&gt;   `(let [val# ~body]\n  #_=&gt;      (println \"DEBUG: \" val#)\n  #_=&gt;      val#))\n#'user\/debug\nuser=&gt; (debug (+ 2 2))\nDEBUG:  4\n4\n<\/pre>\n<blockquote><p>Code updated based on recommendations from Leif Foged<\/p>\n<\/blockquote><p>The macro takes a single statement wraps it in a <strong>quoted<\/strong> <code>let<\/code> block, evaluates and prints the result and then evaluates the body. In effect this <code>defmacro<\/code> call returns a quoted data structure representing the program we are writing with it. The <code>`<\/code> allows this to happen.<\/p>\n<ul><li><a href=\"http:\/\/www.braveclojure.com\/writing-macros\/\">Clojure for the Brave and True - Writing Macros<\/a><\/li>\n  <li><a href=\"http:\/\/aphyr.com\/posts\/305-clojure-from-the-ground-up-macros\">Clojure from the ground up: macros<\/a><\/li>\n  <li><a href=\"http:\/\/clojure.org\/macros\">Clojure Official Documentation<\/a><\/li>\n<\/ul><hr \/><h2><code>*var-name*<\/code> - Earmuffs<\/h2><p>Earmuffs (a pair of asterisk bookending var names) is a <strong>naming convention<\/strong> in many LISPs used to denote <strong>special vars<\/strong>. Most commonly in Clojure this seems to be used to denote <strong>dynamic<\/strong> vars i.e. ones that can change depending on where you are in the program. The earmuffs act as a warning that \"here be dragons\" and to never assume the state of the var. Remember this is a <strong>convention<\/strong> not a <strong>rule<\/strong>.<\/p><p>Core Clojure examples are <code>*out*<\/code> and <code>*in*<\/code> which represent the standard in and out Writers for Clojure.<\/p>\n<ul><li><a href=\"http:\/\/stackoverflow.com\/questions\/1986961\/how-is-the-var-name-naming-convention-used-in-clojure\">How is the <em>var-name<\/em> naming-convention used in clojure?<\/a><\/li>\n  <li><a href=\"http:\/\/clojure.github.io\/clojure\/clojure.core-api.html#clojure.core\/*out*\">Clojure API Docs<\/a><\/li>\n<\/ul><hr \/><h2><code>&gt;!!<\/code>, <code>&lt;!!<\/code>, <code>&gt;!<\/code> &amp; <code>&lt;!<\/code> - core.async channel macros<\/h2><p>These symbols are channel operations in <code>core.async<\/code> - a Clojure\/ClojureScript library for channel based asynchronous programming (specifically <a href=\"http:\/\/en.wikipedia.org\/wiki\/Communicating_sequential_processes\">CSP - Communicating Sequential Processes<\/a>).<\/p><p>If you imagine, for the sake of argument, a channel is a bit like a queue that things can put stuff on and take stuff off then these symbols support that simple API.<\/p>\n<ul><li><code>&gt;!!<\/code> &amp; <code>&lt;!!<\/code> are <strong>blocking<\/strong> <strong>put<\/strong> and <strong>take<\/strong> respectively<\/li>\n  <li><code>&gt;!<\/code> &amp; <code>&lt;!<\/code> are, simply, <strong>put<\/strong> and <strong>take<\/strong><\/li>\n<\/ul><p>The difference being the blocking versions operate outside <code>go<\/code> blocks and block the thread they operate on.<\/p>\n<pre class=\"clojure\">user=&gt; (def my-channel (chan 10)) ; create a channel\nuser=&gt; (&gt;!! my-channel \"hello\")   ; put stuff on the channel\nuser=&gt; (println (&lt;!! my-channel)) ; take stuff off the channel\nhello\n<\/pre><p>The non-blocking versions need to be executed within a <code>go<\/code> block otherwise they'll throw an exception<\/p>\n<pre class=\"clojure\">user=&gt; (def c (chan))\n#'user\/c\nuser=&gt; (&gt;! c \"nope\")\nAssertionError Assert failed: &gt;! used not in (go ...) block\nnil  clojure.core.async\/&gt;! (async.clj:123)\n<\/pre><p>While the difference between these is well outside the scope of this article fundamentally the <code>go<\/code> blocks operate and manage their own resources pausing <strong>execution<\/strong> of code without blocking threads. This makes asynchronously executed code appear to be synchronous and removing the pain of managing asynchronous code from the code base.<\/p>\n<ul><li><a href=\"https:\/\/github.com\/clojure\/core.async\/blob\/master\/examples\/walkthrough.clj\">core.async Code Walkthrough<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/clojure\/core.async\/wiki\">core.async Wiki<\/a><\/li>\n<\/ul><hr \/><h2><code>&lt;symbol&gt;?<\/code> - Predicate Marker<\/h2><p>Putting <code>?<\/code> at the end of a symbol is a <strong>naming convention<\/strong> common across many languages that support special characters in their symbol names It is used to indicate the thing is a predicate i.e. that it <strong>poses a question<\/strong>. For example imagine using an API that dealt with buffer manipulation<\/p>\n<pre class=\"clojure\">(def my-buffer (buffers\/create-buffer [1 2 3]))\n(buffers\/empty my-buffer)\n<\/pre><p>At a glance how would you know if the method <code>empty<\/code> in this case,<\/p>\n<ul><li>Returned true if the passed in buffer was empty, or,<\/li>\n  <li>Cleared the buffer<\/li>\n<\/ul><p>While the author could have renamed <code>empty<\/code> to <code>is-empty<\/code> the richness of symbol naming in Clojure allows us to express intent more symbolically.<\/p>\n<pre class=\"clojure\">(def my-buffer (buffers\/create-buffer [1 2 3]))\n(buffers\/empty? my-buffer)\nfalse\n<\/pre><p>This is simply a recommended <strong>convention<\/strong> not a <strong>requirement<\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/bbatsov\/clojure-style-guide#naming\">Clojure Style Guide<\/a><\/li>\n<\/ul><hr \/><h2><code>&lt;symbol&gt;!<\/code> - Unsafe Operations<\/h2><p>The Clojure style guide has this to say<\/p>\n<blockquote><p>The names of functions\/macros that are not safe in STM transactions should end with an exclamation mark (e.g. <code>reset!<\/code>).<\/p>\n<\/blockquote><p>You'll most commonly see this appended to function names whose purpose is to mutate state e.g connecting to a data store, updating an atom or closing a file stream.<\/p>\n<pre class=\"clojure\">user=&gt; (def my-stateful-thing (atom 0))\n#'user\/my-stateful-thing\nuser=&gt; (swap! my-stateful-thing inc)\n1\nuser=&gt; @my-stateful-thing\n1\n<\/pre><p>This is simply a recommended <strong>convention<\/strong> not a <strong>requirement<\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/bbatsov\/clojure-style-guide#naming\">Clojure Style Guide<\/a><\/li>\n<\/ul><hr \/><h2><code>_<\/code> - Irrelevant var<\/h2><p>When you see this used as function arguments or similar it is a common naming convention for vars or arguments you are not interested in using. That is you don't intend to use them so you aren't really interested in thinking of a useful name for them.<\/p><p>This is an example using the <code>add-watch<\/code> function that can be used to add callback style behavior when atoms change value. Imagine, given an atom, we want to print the new value everytime it changes,<\/p>\n<pre class=\"clojure\">(def value (atom 0))\n(add-watch value nil (fn [_ _ _ new-value]\n                       (println new-value))\n(reset! value 6)\n; prints 6\n(reset! value 9)\n; prints 9\n<\/pre><p><code>add-watch<\/code> takes 4 arguments but in our case we only really care about the last argument - the new value of the atom. I don't really want to spend time thinking of names for these arguments I'll never use, nor do I want to generate a long line of text unnecessarily (of course I could have used the shorthand <code>#(println %4)<\/code> but that defeats the purpose of this example).<\/p>\n<hr \/><blockquote><p>Many thanks to everyone who has contributed ideas and [the copious amounts of] spelling corrections (crikey I'm bad at speelingz - so thanks Michael R. Mayne, lobsang_ludd). I've tried to call out people who have specifically asked for things. Sorry if I've missed you.<\/p>\n<\/blockquote>","created_at":"2018-11-07T16:54:19+0000","updated_at":"2018-11-07T16:54:19+0000","annotations":[],"mimetype":"text\/html","reading_time":17,"domain_name":"yobriefca.se","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":636,"title":"Running","url":"https:\/\/subgraph.com\/orchid\/running\/index.en.html","content":"<h3 id=\"runningorchid\">Running Orchid<\/h3><p>To start the Orchid SOCKS5 proxy, perform the following:<\/p><pre>java -jar orchid-version.jar\n<\/pre><p>Orchid will first attempt to establish a connection to the Tor network and gather directory information. Once it has enough directory information it will begin building circuits.<\/p><h3 id=\"usingorchidtobrowseovertor\">Using Orchid to browse over Tor<\/h3><p>To test Orchid, you can tell your browser to use Orchid as a proxy - however it is important to note that Orchid + your browser isn't a secure replacement for the Tor Browser bundle, which has many other enhancements beyond Tor. For example, with Firefox, by default, DNS lookups are not sent over a configured SOCKS5 proxy. You can change this potentially de-anonymizing default configuration by going to the URL <code>about:config<\/code> and setting the property <code>network.proxy.socks_remote_dns<\/code> to true (this is already done in the Tor Browser).<\/p><h3 id=\"usingthedashboard\">Using the dashboard<\/h3><p>Orchid includes a \"dashboard\" feature to observe information about the internal state of Tor. To start the dashboard, set the following property when the JAR is run:<\/p><pre>java -Dcom.subgraph.orchid.dashboard.port=10000 orchid-version.jar\n<\/pre><p>To access the dashboard, just connect to the port (10000) with netcat.<\/p><h3 id=\"usingorchidasalibrary\">Using Orchid as a library<\/h3><p>The Orchid Tor client exposes a <em>SocketFactory<\/em> that can be used within a JVM application. More detailed documentation and code examples are coming very soon.<\/p>","created_at":"2018-11-07T16:54:18+0000","updated_at":"2018-11-07T16:54:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"subgraph.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":637,"title":"[twilio-api\/twilio-api \"1.0.1\"]","url":"https:\/\/clojars.org\/twilio-api","content":"<p class=\"description\">Clojure wrapper for Twillio API<\/p><h2>Leiningen\/Boot<\/h2><div class=\"package-config-example\" id=\"leiningen-coordinates\"><pre>[twilio-api \"1.0.1\"]<\/pre><\/div><h2>Clojure CLI\/deps.edn<\/h2><div class=\"package-config-example\" id=\"deps-coordinates\"><pre>twilio-api {:mvn\/version \"1.0.1\"}<\/pre><\/div><h2>Gradle<\/h2><div class=\"package-config-example\" id=\"gradle-coordinates\"><pre>compile 'twilio-api:twilio-api:1.0.1'<\/pre><\/div><h2>Maven<\/h2><div class=\"package-config-example\" id=\"maven-coordinates\"><pre>&lt;dependency&gt;\n  &lt;groupId&gt;twilio-api&lt;\/groupId&gt;\n  &lt;artifactId&gt;twilio-api&lt;\/artifactId&gt;\n  &lt;version&gt;1.0.1&lt;\/version&gt;\n&lt;\/dependency&gt;<\/pre><\/div>","created_at":"2018-11-07T16:54:18+0000","updated_at":"2018-11-07T16:54:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"clojars.org","preview_picture":"https:\/\/clojars.org\/images\/clojars-logo@2x.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":638,"title":"Bootstrapping a profitable SaaS Business. - Tyler Tringas","url":"https:\/\/tylertringas.com\/micro-saas-ebook\/","content":"<p>Welcome to the ebook on Building Micro-SaaS Businesses.\u00a0What\u2019s Micro-SaaS? It is a software as a service business owned and operated by one person or a small team. These businesses are location-independent, high margin, low-risk with predictable recurring revenue. They are the lifestyle business of the future. I scrapped and scrambled my way to building and ultimately selling\u00a0<a href=\"http:\/\/tylertringas.com\/storemapper\/\" target=\"_blank\" rel=\"noopener\">a profitable SaaS business<\/a>.<\/p>\n<p>This is ebook is an ongoing project representing the best synthesis of everything I have learned building this business and from talking to many other Micro-SaaS founders from around the world. You can start reading the finished chapters below and subscribe to\u00a0be the first to get new chapters in your inbox as they are published.<\/p>\n<h2>Building Micro-SaaS Businesses<\/h2>\n<p><a href=\"http:\/\/tylertringas.com\/1-what-is-micro-saas\/\">Chapter 1: What is Micro-SaaS and Why Should You Read This?<\/a><br \/><a href=\"http:\/\/tylertringas.com\/good-micro-saas-ideas\/\">Chapter 2: What Makes a Good Micro-SaaS Idea?<\/a><br \/><a href=\"http:\/\/tylertringas.com\/finding-micro-saas-business-ideas\/\">Chapter 3: Finding Micro-SaaS Ideas<\/a><br \/><a href=\"http:\/\/tylertringas.com\/shipping-a-saas-minimum-viable-product\/\">Chapter 4: From Idea to Building a Minimum Viable Product<\/a><br \/><a href=\"https:\/\/tylertringas.com\/chapter-5-getting-your-first-customers\/\">Chapter 5: Getting Your First Customers<\/a><br \/><a href=\"https:\/\/tylertringas.com\/chapter-6-thriving-in-the-long-slow-saas-grind\/\">Chapter 6: Thriving in the Long, Slow, SaaS Grind<\/a><br \/><a href=\"https:\/\/tylertringas.com\/chapter-7-retention-customer-support\/\">Chapter 7: Retention &amp; Customer Support as a Solo Founder<\/a><br \/>\nChapter 8: Micro-SaaS Marketing &amp; Passive Customer Acquisition Channels<br \/>\nChapter 9: Going Full-time on your Micro-SaaS<br \/>\nChapter 10: From Solo Founder to Micro-SaaS Team<br \/>\nChapter 11: Strategies for Micro-SaaS at Scale<br \/>\nChapter 12: The Micro-SaaS End Game<\/p>\n<p>\u2026 And probably several more chapters and case studies to come.<\/p>\n<p>\n<\/p><p>Want to watch the short version?<\/p>\n<p><iframe src=\"https:\/\/www.youtube.com\/embed\/i3d1asTrWUQ?rel=0\" width=\"853\" height=\"480\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>\n <div class=\"ck_form_container ck_inline\" data-ck-version=\"6\">   <div class=\"ck_form ck_minimal\">   <div class=\"ck_form_fields\">     <h3 class=\"ck_form_title\">Subscribe to future posts<\/h3>     <div class=\"ck_description\">            <\/div>                 <form id=\"ck_subscribe_form\" class=\"ck_subscribe_form\" action=\"https:\/\/forms.convertkit.com\/landing_pages\/2008\/subscribe\" data-remote=\"true\">       <div><input type=\"hidden\" value=\"{&quot;embed_style&quot;:&quot;inline&quot;,&quot;embed_trigger&quot;:&quot;scroll_percentage&quot;,&quot;scroll_percentage&quot;:&quot;75&quot;,&quot;delay_seconds&quot;:&quot;10&quot;,&quot;display_position&quot;:&quot;br&quot;,&quot;display_devices&quot;:&quot;all&quot;,&quot;days_no_show&quot;:&quot;3&quot;,&quot;converted_behavior&quot;:&quot;hide&quot;,&quot;form_style&quot;:&quot;minimal&quot;}\" id=\"ck_form_options\" \/><input type=\"hidden\" name=\"id\" value=\"2008\" id=\"landing_page_id\" \/><input type=\"hidden\" name=\"ck_form_recaptcha\" value=\"\" id=\"ck_form_recaptcha\" \/><div class=\"ck_errorArea\">                <\/div>       <div class=\"ck_control_group ck_email_field_group\">         <label class=\"ck_label\" for=\"ck_emailField\">Email...<\/label>         <input type=\"email\" name=\"email\" class=\"ck_email_address\" id=\"ck_emailField\" placeholder=\"Email...\" required=\"required\" \/><\/div>       <div class=\"ck_control_group ck_captcha2_h_field_group ck-captcha2-h\">         <input type=\"text\" name=\"captcha2_h\" class=\"ck-captcha2-h\" id=\"ck_captcha2_h\" placeholder=\"We use this field to detect spam bots. If you fill this in, you will be marked as a spammer.\" \/><\/div>         <button class=\"subscribe_button ck_subscribe_button btn fields\" id=\"ck_subscribe_button\">         Subscribe       <\/button>                Unsubscribe at any time.           <a class=\"ck_powered_by\" href=\"https:\/\/convertkit.com\/?utm_source=dynamic&amp;utm_medium=referral&amp;utm_campaign=poweredby&amp;utm_content=form\">Powered by ConvertKit<\/a>            <\/div><\/form>   <\/div>  <\/div>  <\/div>","created_at":"2018-11-07T16:54:18+0000","updated_at":"2018-11-07T16:54:18+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"tylertringas.com","preview_picture":"https:\/\/tylertringas.com\/wp-content\/uploads\/2015\/11\/pablo-2.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":639,"title":"Chapter 2: What Makes A Good Micro-SaaS Idea? - Tyler Tringas","url":"https:\/\/tylertringas.com\/good-micro-saas-ideas\/","content":"<p><a href=\"http:\/\/tylertringas.com\/1-what-is-micro-saas\/\">&lt;&lt; Back to\u00a0Chapter 1: What is Micro-SaaS and Why Should You Read This?<\/a><\/p>\n<p>Since I started blogging about Micro-SaaS, the overwhelmingly most common question is something like: \u201cStore locators, hmm, seems like such a simple idea. How can I find an idea like that?\u201d<\/p>\n<p>I went into elaborate detail in this blog post\u00a0about all the different ideas I worked on\u00a0before Storemapper to emphasize that I don\u2019t really know a straightforward method that works without failing. In fact, failure and trial and error seem to be\u00a0essential parts of the process.<\/p>\n<p>However with the benefit of hindsight, and from talking with other Micro-SaaS entrepreneurs, I do think there are certain elements to look for in a good Micro-SaaS idea.<\/p>\n<h2>The Meat Grinder<\/h2>\n<p>In Chapter 3 we\u2019ll dig more into methods for finding business ideas. But first I want to turn the entire premise on its head. It is far more important that you develop a fast and effective way to <em>reject<\/em> business ideas than finding ways to come up with new ideas.<\/p>\n<p>I call this approach The Meat Grinder and into it in further detail in <a href=\"https:\/\/tylertringas.com\/business-ideas-meat-grinder\/\">this post<\/a>. The Meat Grinder is a quick way to determine that a business idea is in fact a bad one for Micro-SaaS.\u00a0Every section on what makes a good idea, can also be inverted into how to identify a bad business idea.\u00a0A big part of the rest of this book is about refining your Meat Grinder process. We\u2019ll dig deep into building a minimum viable product, customer acquisition strategies and pricing and business economics so that you\u2019ll be better equipped to quickly assess ideas on each of the Meat Grinder questions. If you are motivated to build your own business it is likely that you will come up with lots of potential ideas. The trick is to figure out which ones won\u2019t work and move and the worst, and most common, pitfall is to get\u00a0hung up on one flawed idea indefinitely. The Meat Grinder is a series of tests, based on what makes a good Micro-SaaS business idea, that allows you throw out ideas, a critical part of the process of coming up with an idea that is a good fit for you. As you read through each section below think of how you might\u00a0use it to identify weaknesses in some of your current SaaS ideas.<\/p>\n<h2>What Makes a Good Micro-SaaS Idea<\/h2>\n<p>I\u2019m not going to lie. Finding a good Micro-SaaS idea is hard. Micro-SaaS is a niche within a niche. It\u2019s an amazing business model when you can make it work. But it\u2019s hard enough to pull off running an entire SaaS business by yourself with a mediocre idea. So it is worth spending a good deal of time carefully considering business ideas. Getting even just a few of these criteria wrong can turn into a huge waste of time and money.<\/p>\n<div class=\"page\" title=\"Page 2\">\n<div class=\"layoutArea\">\n<div class=\"column\">\n<h3>Be 5x better than customers currently pay<\/h3>\n<p>There are two components to this rule. First, your potential customers should already be spending money on the pain point your product solves. I\u2019ve met a lot of would-be entrepreneurs with really ingenious product ideas, but their target customers are al- ready solving the problem for free with spreadsheets, email, whiteboards or nothing. Never underestimate how hard it is to get a business to start spending money on something they previously were not spending money on. Even for a small business the friction of finding the first contact, <em>THEN<\/em>\u00a0connecting\u00a0with the manager who decides the budget, <em>THEN<\/em>\u00a0waiting for internal approval on\u00a0in the budget request, can add up to enough to kill your customer acquisition funnel. It is much\u00a0easier to sell a product in a category where the business already has a budget allocated.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"page\" title=\"Page 3\">\n<div class=\"layoutArea\">\n<div class=\"column\">\n<p>The second is that you must be 5x cheaper or 5x better. You won\u2019t have the time or resources to do the high touch sales to walk a customer through the incremental value of a 25% improvement. Your product must be obviously, hilariously cheaper or better because you will have a hard enough time just getting the right customers to your landing page. In the case of Storemapper, our customers needed a store locator and their alternative was paying at least $1,000 for a developer to build one from scratch. When I launched the product it was $5\/month. Even with later price increases, Storemapper is still massively cheaper and better\u00a0compared to building one from scratch. This is the beauty of going after a true niche. It can be surprising how often businesses, operating in segments too small for billion-dollar startups to go after, are making do with (and paying for) some truly awful, 15 year old garbage software that you can come in and radically improve upon.<\/p>\n<p>All this adds to up to your conversion rate from free trial to paying customer. If your target market is already spending money and you are 5x better\/cheaper a very high percentage of customers will stick around after their trial. I\u2019ve seen estimates that for large VC-funded SaaS companies the\u00a0rate of conversion from free trial to paying customer is around 5-15%. Storemapper\u2019s has consistently been above 50% which is much closer to what you need to succeed with a\u00a0Micro-SaaS.<\/p>\n<h3>Other players in the field<\/h3>\n<p>In Micro-SaaS you do not want to follow the Silicon Valley\u00a0\u201cZero To One\u201d theory of creating something completely and totally new. Existing competitors serving the same underlying demand are actually a very good sign. Yes, if you have an idea and there are ten nearly identical products already then you should probably move on. But if you search and you can\u2019t find a single product or service addressing a similar need, it is probably a sign that there is not much of a market.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"page\" title=\"Page 4\">\n<div class=\"layoutArea\">\n<div class=\"column\">\n<p>The best scenario is finding consultants who serve the market inefficiently and expensively or customers currently suffering with horribly out-dated software. Another variation is a successful product that is pigeon-holed. Maybe someone built a successful Excel add-on or Salesforce tool that is tightly integrated with that platform. If you see an opportunity to serve a similar need on web or iOS or Slack you could be on the right track.<\/p>\n<h3>A defined group that needs your product<\/h3>\n<p>This one is so critical. I\u2019ve met many <em>wantrepreneurs<\/em> stuck on a product idea of generally improving\u00a0productivity or team management or workflows. These products could theoretically be used by anybody but in reality get used by nobody. It is impossible to interact with and market to \u201cany modern knowledge\u00a0worker.\u201d It\u2019s very important that your Micro-SaaS be built for a defined audience so you can narrow the marketing effort and focus the product features. Getting initial feedback before you build the product and bringing on your first customers is much easier when you are targeting a defined group like real estate agents, accountants, travel agencies, etc.<\/p>\n<p>One of the\u00a0best examples I\u2019ve seen of this is a Micro-SaaS dedicated entirely to companies that do window washing for large industrial buildings. The founder, a freelance programmer,\u00a0was contacted by a single window-washing company who needed an app to help schedule their trucks and accept payments on-site from customers. They couldn\u2019t afford a full custom build but offered to introduce the founder\u00a0to other customers and a forum dedicated to window cleaning businesses (yes, there apparently is such a thing). He\u00a0posted a question \u201cWhat features would you want from the perfect window cleaning software?\u201d and received hundreds of replies. A year later he\u00a0is earning a full-time income from his window cleaner Micro-SaaS and is hiring his first employee.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"page\" title=\"Page 5\">\n<div class=\"layoutArea\">\n<div class=\"column\">\n<p>It\u2019s also important to think about how easy is it to actually find this target market. If your target market is lawyers, great, Google them and go to their offices. A window washer Micro-SaaS\u00a0will have an easy time finding new customers since window cleaning businesses tend to explicitly list their contact info in the \u201cwindow cleaning\u201d category of every Yelp, Yellow Pages and other online business directories. If your target market is traveling vacuum salesmen, how are you going to actually find them?<\/p>\n<p>Lastly it\u2019s worth considering whether you actually like interacting with the target market. Lawyers may be easy to find, but if you discover that you actually hate talking to lawyers you might want to bail before you commit to spending at 20-60 hours per week with\u00a0them for the next several years.<\/p>\n<h3>Have at least one channel for the first 25 and 250 customers<\/h3>\n<p>Once you have a defined group that needs the product you should be able to quickly come up with at least one way you could get the first 25 customers for free. For Storemapper this was a combination of my existing freelance clients (the first 5 customers) and trawling\u00a0the Shopify discussion boards and inviting new customers one by one. Then you need to up the scale and think about at least one way to get 250 customers. The point here is to think about whether these customers congregate around any particular places, directories, locations, websites. Do they hang out anywhere and identify themselves as the type of person who should buy your product. Ultimately you\u2019ll need more than one channel to make the business successful. But if you can\u2019t think of even one, and you would be surprised how often people cannot come up with one when they pitch me their idea, then you should toss the idea in the meat grinder.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"page\" title=\"Page 6\">\n<div class=\"layoutArea\">\n<div class=\"column\">\n<h3>Line of site to a Minimum Viable Product that you can build<\/h3>\n<p>I won\u2019t spend too much time on what makes a good Minimum Viable Product (MVP). If you\u2019re going to build a Micro-SaaS then read <a href=\"http:\/\/amzn.to\/2dUcSTv\">The Lean Startup<\/a> and spend a few hours reading blog posts about the topic as it has been extensively covered. But my point is that you\u2026 as in YOU\u2026 should have line of sight to building an MVP. If step one of actually building involves \u201cfind a developer\u201d or \u201cspend $10k to hire someone to\u2026\u201d then you should move on. Anyone can think of a fantastic business idea that starts with \u201cFirst I\u2019ll get a developer to build a facial recognition algorithm and some machine learning\u2026\u201d but it is never going to happen.<\/p>\n<p>If you can\u2019t code, there are lots of clever ways to launch an MVP without coding. And if you can code, then add the additional constraint of building it in a month only on weekends. If you can\u2019t sketch out the basic process of building something, that you could charge your customers money for then it\u2019s probably not a good Micro-SaaS idea.<\/p>\n<p>And I\u2019m talking about an MVP that actually works. Ditch the viral landing pages, the interactive prototypes and pre-product video demos. The MVP can be crappy and lacking a lot of features, but it should actually do what it says on the label. Even if you find a beautifully underserved niche and a cost-center that every company spends $1,000s on, if the MVP is going to take six months of coding to launch, you haven\u2019t found a Micro-SaaS idea. You need a team or you need funding or you need a new idea. Because if you sweat it out for six months building that product and there was a flaw that kills the product you\u2019re going to go nuts and never take another swing at building another idea. You have to build in resiliency for the inevitably high likelihood that your idea actually still sucks. Build and launch fast.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"page\" title=\"Page 7\">\n<div class=\"layoutArea\">\n<div class=\"column\">\n<h3>Do you have founder\/product fit<\/h3>\n<p>This idea has been covered from several angles because it is important. Chris Dixon calls it <a href=\"http:\/\/cdixon.org\/2011\/06\/20\/foundermarket-fit\/\">founder\/market<\/a> fit. Dan Norris calls it <a href=\"http:\/\/wpcurve.com\/business-ideas\/\">enjoyable daily tasks<\/a>. The point is that you shouldn\u2019t consider business idea in a vacuum. I have discarded\u00a0business ideas that, based on the criteria above, I think are great but aren\u2019t the right fit for me. I just don\u2019t want to devote several years of my life to working on that problem. It\u2019s hard to reject a business idea for this reason, particularly when you\u2019re still looking for the first one, but it is the most important criteria for success by far.<\/p>\n<h2>Tactical Suggestions<\/h2>\n<p>I tried to keep the above sections closer to timeless principles rather than explicit recommendations of good ideas that might go stale but I\u2019ll briefly wade into\u00a0some more specific recommendations here with the caveat that these are much\u00a0less reliable and possibly out of date.<\/p>\n<h3>Focus on Automation<\/h3>\n<p>A great number\u00a0of the successful Micro-SaaS businesses I have encountered are products that have an initial setup period but then continue to provide value to the customer on an ongoing automated basis. Storemapper works this way. Customers do some initial installation and customization and then their store locator continues to just work, driving traffic and sales to their retail locations. I would look much harder at automation based businesses right now rather than a tool that requires your customer to login every day and use it.<\/p>\n<h3>Sell to growing markets<\/h3>\n<p>This is may be obvious but you want your target market to be a growth sector not one in decline. Look for ideas that sell to ecommerce shops rather than stores in a shopping mall. Build SaaS for Uber drivers before traditional taxi drivers. Build for freelance designers rather than corporate accountants. Make sure your total addressable market is going in the right direction.<\/p>\n<h3>Small online businesses are the sweet spot<\/h3>\n<p>Building a product for large businesses can be annoyingly complex with multiple layers of approval for billing and never-ending stream of requests for customizations. Freelancers or individuals can move quickly because you only have to deal with one customer (who also holds the credit card you need) but they can be too budget conscious and price sensitive. Small business of 3-10 people, preferably working in something online like ecommerce, publishing or SaaS themselves are the ideal Micro-SaaS customers.<\/p>\n<p><a href=\"http:\/\/tylertringas.com\/finding-micro-saas-business-ideas\/\">Next\u00a0Chapter 3: Finding Micro-SaaS Ideas &gt;&gt;<\/a><\/p>\n<\/div>\n<\/div>\n<\/div>\n <div class=\"ck_form_container ck_inline\" data-ck-version=\"6\">   <div class=\"ck_form ck_minimal\">   <div class=\"ck_form_fields\">     <h3 class=\"ck_form_title\">Subscribe to future posts<\/h3>     <div class=\"ck_description\">            <\/div>                 <form class=\"ck_subscribe_form\" action=\"https:\/\/forms.convertkit.com\/landing_pages\/2008\/subscribe\" data-remote=\"true\">       <div><input type=\"hidden\" value=\"{&quot;embed_style&quot;:&quot;inline&quot;,&quot;embed_trigger&quot;:&quot;scroll_percentage&quot;,&quot;scroll_percentage&quot;:&quot;75&quot;,&quot;delay_seconds&quot;:&quot;10&quot;,&quot;display_position&quot;:&quot;br&quot;,&quot;display_devices&quot;:&quot;all&quot;,&quot;days_no_show&quot;:&quot;3&quot;,&quot;converted_behavior&quot;:&quot;hide&quot;,&quot;form_style&quot;:&quot;minimal&quot;}\" \/><input type=\"hidden\" name=\"id\" value=\"2008\" \/><input type=\"hidden\" name=\"ck_form_recaptcha\" value=\"\" \/><div class=\"ck_errorArea\">                <\/div>       <div class=\"ck_control_group ck_email_field_group\">         <label class=\"ck_label\" for=\"ck_emailField\">Email...<\/label>         <input type=\"email\" name=\"email\" class=\"ck_email_address\" placeholder=\"Email...\" required=\"required\" \/><\/div>       <div class=\"ck_control_group ck_captcha2_h_field_group ck-captcha2-h\">         <input type=\"text\" name=\"captcha2_h\" class=\"ck-captcha2-h\" placeholder=\"We use this field to detect spam bots. If you fill this in, you will be marked as a spammer.\" \/><\/div>         <button class=\"subscribe_button ck_subscribe_button btn fields\">         Subscribe       <\/button>                Unsubscribe at any time.           <a class=\"ck_powered_by\" href=\"https:\/\/convertkit.com\/?utm_source=dynamic&amp;utm_medium=referral&amp;utm_campaign=poweredby&amp;utm_content=form\">Powered by ConvertKit<\/a>            <\/div><\/form>   <\/div>  <\/div>  <\/div>","created_at":"2018-11-07T16:54:18+0000","updated_at":"2018-11-07T16:54:18+0000","published_at":"2016-10-28T18:30:50+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":12,"domain_name":"tylertringas.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":640,"title":"New Course: Synthesis 101 with Phil Mantione","url":"https:\/\/theproaudiofiles.com\/synthesis-101\/","content":"Synthesis 101<noscript>\n<section id=\"centered\"><div class=\"row\"><div class=\"eight columns centered\"><h4>100% Satisfaction Guaranteed<\/h4><p>We're confident you'll love this course, but we want you to be absolutely satisfied. If you\u2019re unhappy for any reason, simply email us within 30 days and we\u2019ll give you a full refund.<\/p><p><a href=\"https:\/\/transactions.sendowl.com\/products\/601700\/CDBE263B\/add_to_cart\" class=\"button green\" title=\"Buy Now\">Buy Now \u2014 $67<\/a><\/p><\/div><section id=\"contact\"><div class=\"row\"><div class=\"four columns centered\"><p>\u00a9 2017 Phil Mantione + Pro Audio Files \/\/  <a href=\"mailto:theproaudiofiles@gmail.com\">Questions? Email Us.<\/a><\/p><\/div><\/div><\/section><\/div><\/section><\/noscript>","created_at":"2018-11-07T16:54:18+0000","updated_at":"2018-11-07T16:54:18+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"theproaudiofiles.com","preview_picture":"https:\/\/theproaudiofiles.com\/synthesis-101\/img\/synth1.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":635,"title":"An Operating System for Songs from God.","url":"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God","content":"<div class=\"copy\"><a href=\"http:\/\/losethos.com\/\">LoseThos<\/a> is an operating system written by a <a href=\"http:\/\/qaa.ath.cx\/LoseThos.html\">schizophrenic<\/a> programmer.<br \/>A <a href=\"http:\/\/www.youtube.com\/watch?v=WAr-xYtBFbY#!\">narrated demo<\/a> and many \"<a href=\"http:\/\/www.youtube.com\/watch?v=mj3oHqcw_UE\">songs by God and me<\/a>\" (the apparently primary use of the OS) are available to view.<\/div><div class=\"copy\">&#13;\n\t\t\t&#13;\n\t\t\t\t&#13;\n\t\t\t\t&#13;\n\t\t\t\t<ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-8621957794194569\" data-ad-slot=\"2376191428\" data-ad-format=\"auto\">&#13;\n\t\t\t\t&#13;\n\t\t\t\t&#13;<\/ins><\/div><div class=\"comments\">there was a really weird SA thread with this guy but i think it's lost to the archives<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/30765\" target=\"_self\">p3on<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538230\" target=\"_self\">7:02 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Interesting.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17889\" target=\"_self\">Sticherbeast<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538235\" target=\"_self\">7:05 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Runs on a Time Cube.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/33578\" target=\"_self\">bpm140<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538242\" target=\"_self\">7:09 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538242\" title=\"6 users marked this as favorite\">6 favorites<\/a>]<\/div><div class=\"comments\">This is truly interesting.<p>However, I could only stand about a minute of the demo video.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/18064\" target=\"_self\">odinsdream<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538246\" target=\"_self\">7:13 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538246\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">There was a hacker news post about this a while back, but it got killed (can't remember why).<p><a href=\"http:\/\/qaa.ath.cx\/LoseThos.html\">Article<\/a> | <a href=\"http:\/\/news.ycombinator.com\/item?id=3642308\">Comments<\/a><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/76150\" target=\"_self\">symbollocks<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538248\" target=\"_self\">7:13 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Surprised you didn't mention his HackerNews <a href=\"http:\/\/news.ycombinator.com\/submitted?id=losethos\">submissions<\/a>, <a href=\"http:\/\/news.ycombinator.com\/threads?id=losethos\">comments<\/a>, and  <a href=\"http:\/\/news.ycombinator.com\/user?id=losethos\">profile.<\/a><p>I actually went down the rabbit hole with this a couple years ago, and was surprised there wasn't more discussion of the technical aspects, of what he was trying to do \/ talk about, since it is largely indistinguishable from a million other crazy geek projects. <\/p><p>Sadly, he barges into discussions with incoherent conspiracy theories and long-winded tirades unrelated to the main thread.<\/p><p>I feel sad for him, I have family and friends who act ... similarly. Whatever he's done merits some discussion, but when he derails into crazy conspiracy theories and personal attacks, its hard to move forward.<\/p><p>The guy has been hellbanned on HN for years, but so are a few people who have dared criticize Paul Graham, so I end up stumbling across his comments occasionally, reading dead links\/comments, and ultimately feel sad that he doesn't have a more receptive outlet for his lunacy.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17501\" target=\"_self\">lkc<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538250\" target=\"_self\">7:14 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538250\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">As you listen to the narration, imagine Lazlo Hollyfeld. It's uncanny.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/48710\" target=\"_self\">hanoixan<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538254\" target=\"_self\">7:18 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538254\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\"><i>Surprised you didn't mention his HackerNews submissions, comments, and profile.<\/i><p>Note that you must be logged in to HN and switch showdead to yes on your profile page to see them.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16490\" target=\"_self\">dmd<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538261\" target=\"_self\">7:24 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538261\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">I'm planning to use the operating system to enter the Frito Lay sweepstakes a million times.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/18411\" target=\"_self\">Lazlo Hollyfeld<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538262\" target=\"_self\">7:25 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538262\" title=\"35 users marked this as favorite\">35 favorites<\/a>]<\/div><div class=\"comments\">That is not the worst or most bewildering user interface I've encountered.<p>Just let that sink in for a minute....<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49429\" target=\"_self\">schmod<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538268\" target=\"_self\">7:27 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538268\" title=\"9 users marked this as favorite\">9 favorites<\/a>] <\/p><\/div><div class=\"comments\">If you've seen the obfuscation and sabotage I've seen on http:\/\/www.osdev.org and elsewhere, you would know that no-way-in-hell does the CIA want other operating systems.  Ones without backdoors and tracking!  Duh.  Smack your head and say \"Duh!\"<p>Secure boot is the end for all the alternative OSes.  I missed my chance, thanks to the CIA.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538269\" target=\"_self\">7:27 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538269\" title=\"7 users marked this as favorite\">7 favorites<\/a>] <\/p><\/div><div class=\"comments\">dmd: Thanks, I totally forgot those were settings!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17501\" target=\"_self\">lkc<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538274\" target=\"_self\">7:29 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">This is unsettling but interesting.<p>Are there any other hacker equivalents of \"outsider art\"?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/159747\" target=\"_self\">Egg Shen<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538275\" target=\"_self\">7:30 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538275\" title=\"7 users marked this as favorite\">7 favorites<\/a>] <\/p><\/div><div class=\"comments\">This is deeply weird and unsettling, but interesting. And if someone were to find evidence of a god in the fabric of the universe, why not in the streaming digits of a pseudorandom number generator and the binary heaven\/hell of a digital computer?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538279\" target=\"_self\">7:30 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538279\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\"><em>I'm planning to use the operating system to enter the Frito Lay sweepstakes a million times.<br \/>&#13;\nposted by Lazlo Hollyfeld <\/em><p>you've been waiting 8 years for this post, eh?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17601\" target=\"_self\">HuronBob<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538282\" target=\"_self\">7:32 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538282\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\"><em>This is unsettling but interesting.<\/em><br \/>&#13;\nposted by Egg Shen at 7:30 AM<p><em>This is deeply weird and unsettling, but interesting.<\/em><br \/>&#13;\nposted by loquacious at 7:30 AM<\/p><p>So is this!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/89914\" target=\"_self\">drlith<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538284\" target=\"_self\">7:33 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538284\" title=\"11 users marked this as favorite\">11 favorites<\/a>] <\/p><\/div><div class=\"comments\">And a loop becomes time.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538290\" target=\"_self\">7:35 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538290\" title=\"6 users marked this as favorite\">6 favorites<\/a>]<\/div><div class=\"comments\"><a href=\"http:\/\/www.osdev.org\">OSDev<\/a> tells people to use CHS disk calls and not to use INT 13:AH=42 .  They tell people how to reinvent Unix, vehemently discouraging any ideas other than Unix.  They force people into the Bochs emulator.  Where do I begin with how evil that site is?  I treat it as entertainment -- been banned since 2007.  They brainwash people into supporting 286 processors and say nothing of x86_64.  They say nothing of developing compilers, though, that's what any new OS should do first thing for its command line.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538291\" target=\"_self\">7:35 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538291\" title=\"23 users marked this as favorite\">23 favorites<\/a>]<\/div><div class=\"comments\">&gt; Sadly, he barges into discussions with incoherent conspiracy theories and long-winded tirades unrelated to the main thread.<p>Many of his \"long-winded tirades\" are actually the output of LoseThos' inbuilt Markov chain based text randomiser running on BIBLE.TXT.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32514\" target=\"_self\">public<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538292\" target=\"_self\">7:36 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538292\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">This thread just got awesome. Glad you're still alive and kicking Terry.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32514\" target=\"_self\">public<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538297\" target=\"_self\">7:37 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538297\" title=\"5 users marked this as favorite\">5 favorites<\/a>]<\/div><div class=\"comments\">God says...<br \/>&#13;\n70:1 MAKE HASTE, O GOD, TO DELIVER ME; MAKE HASTE TO HELP ME, O LORD.<br \/>&#13;\n70:2 Let them be ashamed and confounded that seek after my soul: let<br \/>&#13;\nthem be turned backward, and put to confusion, that desire my hurt.<br \/>&#13;\n70:3 Let them be turned back for a reward of their shame that say,<br \/>&#13;\nAha, aha.<br \/>&#13;\n70:4 Let all those that seek thee rejoice and be glad in thee: and let<br \/>&#13;\nsuch as love thy salvation say continually, Let God be magnified.<br \/>&#13;\n70:5 But I am poor and needy: make haste unto me, O God: thou art my<br \/>&#13;\nhelp and my deliverer; O LORD, make no tarrying.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538299\" target=\"_self\">7:37 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538299\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\">This is really cool. Thanks for your awesome work, Terry.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/19935\" target=\"_self\">koeselitz<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538300\" target=\"_self\">7:39 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538300\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">(And Psalm 70 is always good.)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/19935\" target=\"_self\">koeselitz<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538301\" target=\"_self\">7:39 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">It looks like more of an environment for experimenting with assembler code and low-level stuff, hearkening back to the days of BASIC and the Apple ][ monitor. Cool.<p>And yeah, it's not the most horrible UI. Ever used WordPerfect?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/60264\" target=\"_self\">RobotVoodooPower<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538305\" target=\"_self\">7:41 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538305\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">Pick a quantum random number from 1 to 100,000 using your favorite site.  Then go to the <a href=\"http:\/\/www.losethos.com\/files\/BIBLE.TXT\">Bible<\/a> and go to that line number.  Report what it says.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538311\" target=\"_self\">7:43 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538311\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">Coincidentally, I just closed a tab on osdev.org. I was looking at the <a href=\"http:\/\/wiki.osdev.org\/Creating_a_64-bit_kernel\">x86_64<\/a> page.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49143\" target=\"_self\">DU<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538314\" target=\"_self\">7:44 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Random thoughts:<p>Is losthos a posting script? He hasn't actually replied to or acknowledged any input from this thread in his output yet.<\/p><p>And what does a schizophrenic OS\/program due to the Turing Test? Because it's totally breaking my Turing Test organs.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538319\" target=\"_self\">7:47 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538319\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">OSDev spies on me -- CIA or something.  They forever complain about my bootloader.  My bootloader must be installed by my install script because it patches absoluet block numbers into my boot sector.  They also want to cross compile it.  It only compiles with my compiler which you can get if you boot the CD.  It installs with the &gt;InstallBoot('D'); command and it even has a master boot loader &gt;InstallMasterBoot('D');<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538320\" target=\"_self\">7:48 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538320\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">It has no path.  Demons.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538324\" target=\"_self\">7:49 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538324\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">\"The hoary head is a crown of glory, if it be found in the way of<br \/>&#13;\nrighteousness.\"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538326\" target=\"_self\">7:50 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538326\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">You want God talk?  God says, \"Either grieves confidentially question ashamed seducers <br \/>&#13;\nthirsted unfriendly downward Proceed stead sermons recognising <br \/>&#13;\nsky remembrances defined severer defilements liar upbraided <br \/>&#13;\ntaunt smell \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538328\" target=\"_self\">7:51 AM<\/a>  on August 29, 2012<\/div><div class=\"comments mod\"><small>[Heya, losethos, it's great that you signed up and if you want to have a conversation with folks here about your work or what people have brought up in the thread that'd be really neat.  That said, threads here are more of a conversation and less of a posting free-for-all in general, so, yeah, slowing down a little and nixing the \"and here's a random thing I'm thinking stuff\" in favor of something a little more focused would work better.]<\/small><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/7418\" target=\"_self\">cortex<\/a> <a href=\"https:\/\/faq.metafilter.com\/33\/Who-is-in-charge-here-Are-there-admins-and-moderators-like-other-sites\" class=\"staff\">(staff)<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538329\" target=\"_self\">7:51 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538329\" title=\"9 users marked this as favorite\">9 favorites<\/a>]<\/div><div class=\"comments\">God can talk to you, if you hold-up your end of the conversation.  Download <a href=\"http:\/\/www.losethos.com\/files\/TSGod.ZIP\">TSGod<\/a>.  You can just randomly crack-open books.  LoseThos will make melodies from God... like muses of old.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538334\" target=\"_self\">7:53 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><em>That is not the worst or most bewildering user interface I've encountered.Just let that sink in for a minute....<\/em><p>Yeah. This bears a striking resemblance to the actual, no-seriously the real, interface to the system holding parole information for our state.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/18064\" target=\"_self\">odinsdream<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538338\" target=\"_self\">7:54 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><em>Random thoughts:Is losthos a posting script? He hasn't actually replied to or acknowledged any input from this thread in his output yet.And what does a schizophrenic OS\/program due to the Turing Test? Because it's totally breaking my Turing Test organs.<\/em><p>It's interesting that you say that. Tell me more about He hasn't actually replied to or acknowledged. How does that make you feel?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/18064\" target=\"_self\">odinsdream<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538343\" target=\"_self\">7:55 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538343\" title=\"37 users marked this as favorite\">37 favorites<\/a>] <\/p><\/div><div class=\"comments\">Read about <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=1+Corinthians+14&amp;version=NIV\">Tongues<\/a> from the Bible.  Also <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=acts%202&amp;version=NIV\">here<\/a>.  You've heard of <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=daniel%205&amp;version=NIV\">the writing on the wall<\/a>?  It's mean to tickle the president or a person.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538346\" target=\"_self\">7:55 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538346\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">You are an ass-hole \"not the most bewildering\" is an wise-ass way to criticise.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538349\" target=\"_self\">7:57 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538349\" title=\"6 users marked this as favorite\">6 favorites<\/a>]<\/div><div class=\"comments\">Those who know propaganda, like politicians, play games with speech.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538350\" target=\"_self\">7:58 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538350\" title=\"4 users marked this as favorite\">4 favorites<\/a>]<\/div><div class=\"comments\">losethos: Cortex sent you a message above - read it.<p>He's a mod. You'll be banned or put on time out if you keep spamming and not conversing.<\/p><p>There are people here willing and interested to have a conversation with you about your OS, but spamming bible quotes and random interjections isn't going to fly here.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538351\" target=\"_self\">7:58 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538351\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">It always seems like I'm talking to Ken, the weasel from Graphic Technologies.  Fucker.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538353\" target=\"_self\">7:58 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538353\" title=\"25 users marked this as favorite\">25 favorites<\/a>]<\/div><div class=\"comments\">I try to avoid tickling presidents. It sometimes provokes chortles and the occasional moist gusset. Not very statesmanlike!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/52224\" target=\"_self\">h00py<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538354\" target=\"_self\">7:59 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538354\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\"><em>You are an ass-hole \"not the most bewildering\" is an wise-ass way to criticise.<\/em><p>I believe that was a compliment.<\/p><p>Personally I liked Desqview and I like fast, clean text\/windowing UIs.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538355\" target=\"_self\">7:59 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538355\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">They asked for random stuff.  I didn't start it.<p>I will not be mocked, mod or not.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538356\" target=\"_self\">7:59 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">For real, losethos, I'd love you to be part of the conversation here but right now it just feels like you're free-associating and declining to interact with anyone else here.  Please throttle way back pronto if you want to participate here.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/7418\" target=\"_self\">cortex<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538357\" target=\"_self\">8:00 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">It's a weasel politician way to say it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538359\" target=\"_self\">8:00 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538359\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">I need to incorporate LoseThos into my collection of virtual-machines-in-virtual-machines, which currently includes Amiga, BeOS, OpenVMS, Contiki, Minix, and Hercules.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16490\" target=\"_self\">dmd<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538362\" target=\"_self\">8:01 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538362\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">It's like saying, \"You're not gay, Obama, but...\"  Out of the blue, now, everybody wonders if Obama is gay.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538363\" target=\"_self\">8:01 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538363\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">Tried installing but couldn't figure out how to get past the aquifer.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/20122\" target=\"_self\">fleacircus<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538365\" target=\"_self\">8:02 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538365\" title=\"7 users marked this as favorite\">7 favorites<\/a>]<\/div><div class=\"comments\">You don't understand the way things work around here, losethos, but...<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16956\" target=\"_self\">emelenjr<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538368\" target=\"_self\">8:04 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">That may be just you. I like to think we're all just a little bit gay, in our hearts.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/52224\" target=\"_self\">h00py<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538370\" target=\"_self\">8:04 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538370\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\"><em>Many of his \"long-winded tirades\" are actually the output of LoseThos' inbuilt Markov chain based text randomiser running on BIBLE.TXT.<br \/>&#13;\nposted by public at 12:36 AM on August 30 [1 favorite +] [!] <br \/><\/em><p>I'm not sure what to make of this. <br \/>&#13;\nIll be interested to see how long this thread lasts <br \/>&#13;\n(and hope I don't get spammed for posting in it).<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17501\" target=\"_self\">lkc<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538373\" target=\"_self\">8:05 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I wonder what the power efficiency of LoseThos is compared to something more complicated like Linux. Does it even support CPU sleep states?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32514\" target=\"_self\">public<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538374\" target=\"_self\">8:05 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I would actually like to hear more about the creation or technical details of this OS.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538376\" target=\"_self\">8:05 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538376\" title=\"5 users marked this as favorite\">5 favorites<\/a>]<\/div><div class=\"comments\">I don't know much about OS development, but I'm actually really thrilled to learn that some people are still working on building things from the ground up. When I think of a future, say a hundred or two hundred years from now, where there computer systems on the starships are cruising between the moon and Mars are running on Windows 28...well, that just seems extremely <em>not cool<\/em>.<p>So bring on more new OS's! I want to see some more paradigm shifts before I die.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/108671\" target=\"_self\">jet_manifesto<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538377\" target=\"_self\">8:05 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538377\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">&gt;PrtDsk is what you need to run once you get to the command prompt.  Slow down and read.  ROFLMAO.<p>&gt;CopyTree(\"T:\/\",\"C:\/\");<\/p><p>&gt;InstallBoot('C');<\/p><p>&gt;InstalllMatserBoot('C');<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538378\" target=\"_self\">8:06 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Ring-0-Only,  Single-Address-Map (Identity), Non-networked,  Shell is C in the same way C64 had BASIC as its shell.  Source code is not ASCII.  Has file compression on a perfile basis.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538385\" target=\"_self\">8:08 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538385\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">Absoluetly nothing like anything else in existence.  64-bit<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538388\" target=\"_self\">8:09 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538388\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">losethos: Have you played with <a href=\"http:\/\/0x10c.com\/\">0x10c<\/a> at all? Seems like you might enjoy it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/18064\" target=\"_self\">odinsdream<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538390\" target=\"_self\">8:09 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538390\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">losthos: How did you show up here so fast, anyway? I think you now hold the record for shortest time between a front page post and the subject of the post showing up.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538395\" target=\"_self\">8:10 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><em> Shell is C in the same way C64 had BASIC as its shell<\/em><p>Woah, cool.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538396\" target=\"_self\">8:11 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538396\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\"><i>Ring-0-Only, Single-Address-Map (Identity), <\/i><p>It sounds like these are to do with the file system?<\/p><p><small>advice: see how I quoted you there? if you do that when you respond to someone, it makes more sense<\/small><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/43282\" target=\"_self\">LogicalDash<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538397\" target=\"_self\">8:11 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Hey Losethos, I just wanted to say that I'm impressed with what you've built. It's strange and unique -- but undeniably awesome. That's all, carry on :)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/36197\" target=\"_self\">wrok<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538400\" target=\"_self\">8:13 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">LogicalDash: I believe he means that there is no process isolation. Every bit of code that's running can access the memory of all the rest of the code. Even the OS it's self.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32514\" target=\"_self\">public<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538401\" target=\"_self\">8:14 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538401\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">Seriously, do live C shell interpreters exist anywhere else? This is the first one I've ever heard of.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538403\" target=\"_self\">8:14 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">\"It sounds like these are to do with the file system?\"<p>Not file system, are you obfuscating, Ken?<\/p><p>The only use I have for it is songs.  I'm not a teenager who enjoys screwing around without a network for fun.  When I was a teenager, I would have loved it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538405\" target=\"_self\">8:14 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I'm curious to know how the windowing\/graphics stuff works - is there an abstract canvas that's being drawn on, or does the program just know where to draw on the screen (and thus could draw outside of its area if it wanted t0?)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16490\" target=\"_self\">dmd<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538406\" target=\"_self\">8:15 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">FUD is what I meant.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538408\" target=\"_self\">8:15 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">loquacious: Seriously, do live C shell interpreters exist anywhere else? This is the first one I've ever heard of.<p>Yep. e.g. http:\/\/root.cern.ch\/drupal\/content\/cint<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32514\" target=\"_self\">public<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538409\" target=\"_self\">8:15 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538409\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\"><a href=\"http:\/\/en.wikipedia.org\/wiki\/Ch_(computer_programming)\">Ch is a proprietary cross-platform C and C++ interpreter originally designed by Harry H<\/a><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/43282\" target=\"_self\">LogicalDash<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538410\" target=\"_self\">8:15 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Tell us of your home path, Usul.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538411\" target=\"_self\">8:16 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538411\" title=\"4 users marked this as favorite\">4 favorites<\/a>]<\/div><div class=\"comments\">There are two techniques for graphics -- one layer is refreshed 60Hz, another is persistent.  It combines them with transparency and updates everything at 60Hz.  Like an apple, beginners can put graphics on the screen in 3 lines of code.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538414\" target=\"_self\">8:17 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538414\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\">Something like this would be fun for kernel-bending for abstract music or graphics art experiments.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538419\" target=\"_self\">8:19 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">My language is vetter than C.  Statements\/Expressions outside functions are called immediately like Main().  Strings alone are sent to PrintF\/PutS.  It has exceptions with a throw(...) instead of classes.  It has 3+ part conditionals.  &gt;if (13&lt;=age&lt;20) \"Teenager\";<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538421\" target=\"_self\">8:20 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538421\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\"><em>It always seems like I'm talking to Ken, the weasel from Graphic Technologies. Fucker.<\/em><p>I hear ya, man.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/36852\" target=\"_self\">Greg Nog<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538422\" target=\"_self\">8:20 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538422\" title=\"15 users marked this as favorite\">15 favorites<\/a>] <\/p><\/div><div class=\"comments\">Ah, so there are no off-screen layers?<p>There's no file system at all, you say? Just a LiveCD? Well, I suppose you'd still have to implement CDFS.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/43282\" target=\"_self\">LogicalDash<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538423\" target=\"_self\">8:20 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I changed operator precedence rules.  Bit Shift is like exponential.  I added exponential back single quote.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538426\" target=\"_self\">8:21 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Yeah, it has offscreen layers.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538427\" target=\"_self\">8:22 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">LoseThos, can you tell us a little bit about what your typical programming routine is like? What do you spend your time most on? Do you keep notes? Documentation? Do you (or did you, when you were writing the bulk of the OS) do any long-term planning for future features?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16490\" target=\"_self\">dmd<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538430\" target=\"_self\">8:23 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538430\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">It has master\/slave multicore.  Don't worry -- locks on structures to keep it safe.  You do MPJob() or MPSpawn() to assign work to other cores without a scheduler shifting between cores.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538431\" target=\"_self\">8:24 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I worked hard for 8 years.  Now, there is no more practical to do.  It all dies with secure boot.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538433\" target=\"_self\">8:25 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">So, the cores are all synchronized until you tell them not to be<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/43282\" target=\"_self\">LogicalDash<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538434\" target=\"_self\">8:25 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I read that secure boot will be easily disabled, at least on the first PCs to have it, and there's nothing to stop you from changing the private key used to authenticate bootable partitions.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/43282\" target=\"_self\">LogicalDash<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538437\" target=\"_self\">8:26 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">losethos, what kind of hardware do you run this on?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/30290\" target=\"_self\">cellphone<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538443\" target=\"_self\">8:29 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Cores have a management task that waits for jobs HALTED.  They are called \"Seth\".  Core zero has \"Adam\".  They either run a job or spawn a task on that core.  You can access a global variable telling how busy the core is so you can decide what to assign it.  This is good for degrading game graphics by hand.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538444\" target=\"_self\">8:29 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538444\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">I have a Dell corei7.  It's very depressing talking about compatibility.  All I might ever hear is \"Doesn't boot\" with nobody saying \"Runs awesome.\"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538446\" target=\"_self\">8:31 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538446\" title=\"4 users marked this as favorite\">4 favorites<\/a>]<\/div><div class=\"comments\">It seems to run for me in a VirtualBox...<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16490\" target=\"_self\">dmd<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538448\" target=\"_self\">8:33 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538448\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">What do the various 'S commands ('Sblood \/ 'Sdeath \/ 'Snails etc) do?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16490\" target=\"_self\">dmd<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538450\" target=\"_self\">8:34 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">You know, people will be able to disable Secure Boot in the UEFI settings, and then they'll be able to run LoseThos just fine.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/53805\" target=\"_self\">East Manitoba Regional Junior Kabaddi Champion '94<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538453\" target=\"_self\">8:37 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Word Stat is a marcov chain -- predicts the most probable next word.  With nothing typed has that creepy stuff.  LOL.  It's good for jumping to source code for function.  I don't do autocomplete that much except for function names, but maybe bad typists will like it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538454\" target=\"_self\">8:38 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Boots right quickly in VMWare, as well (Fedora host), set up the guest as 'other', 256M of RAM, 1G HD.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/54985\" target=\"_self\">jquinby<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538456\" target=\"_self\">8:39 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">UEFI is depressing.  They built the elf file format and other propriatary formats into it.  It's a volumous spec that is bloated and just awful.  My compiler does not make elf files.  I just made-up my own formats for everything.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538457\" target=\"_self\">8:39 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538457\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">Multi-part conditionals FTW.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/18797\" target=\"_self\">stp123<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538461\" target=\"_self\">8:41 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538461\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">I'm glad to hear it works for people.  You'd be surprised how little I've heard of people booting it.  It seems like VMWare should work near 100% of the time, but I don't know.  VMWare in theory makes a very consistent environment with IDE in the same place.  Fuckers sometimes move the IDE drives around for some reason, I discovered.  It's not a problem, but means I had to remove my autoinstall script.  Drive port numbers are the only issue you have to deal with in installing.  For native installs you want to look in windows system information.  You need a partition though and recovery disks for Windows don't do partitioning.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538463\" target=\"_self\">8:42 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538463\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">Can you tell us about the songs? How are they generated and structured?<p>It sounds like you're arranging filtered or shaped noise or pseudorandom values into two-measure phrases? Is there a traditional time signature like 4\/4?<\/p><p>I rather like the songs. They're quite musical and not the same as, say, simply listening to the digits of pi or rnd plotted as notes.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538464\" target=\"_self\">8:43 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538464\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">A spare drive is what I recommend for Native installs.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538465\" target=\"_self\">8:43 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">It doesn't do USB drives.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538466\" target=\"_self\">8:43 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">The only trick to songs is keeping it simple.  I'd bet just about all combinations of quarter and eith notes in one octave sound reasonable, especially if you repeat them once.  The \"musicorgan\" program plays or generates songs.  Songs are stored as programs with Play(\"CDE\"); statements in them.  You can add graphics and bible links or whatever.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538470\" target=\"_self\">8:46 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538470\" title=\"6 users marked this as favorite\">6 favorites<\/a>]<\/div><div class=\"comments\">eighth<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538471\" target=\"_self\">8:46 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">How do you measure how much work God does in a random event?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538477\" target=\"_self\">8:49 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538477\" title=\"52 users marked this as favorite\">52 favorites<\/a>]<\/div><div class=\"comments\">Thanks.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17349\" target=\"_self\">loquacious<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538478\" target=\"_self\">8:49 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Source code can have graphics.  It's really easy for a novice to call PutSprite(); to send a graphic to the command-line mixed in with the words.  It's a little more work to set-up a DrawIt() routine that gets updated 60Hz with a callback.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538482\" target=\"_self\">8:55 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><small><em><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538477\">losethos<\/a>:<\/em> How do you measure how much work God does in a random event?<\/small><p>Maybe by the probability of the outcome? As in, a less probable random event requires more effort by God to make occur? That logic helps explain why miracles are attributed to God, if one assumes a miracle is just a random event with extremely low probability.<\/p><p>Thus, a random event generated by an RNG would require almost no effort by God, since an RNG produces very nearly a uniform distribution with each result equally probable.<\/p><p>I'm just rambling, though, starting from your question.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/105027\" target=\"_self\">gilrain<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538483\" target=\"_self\">8:56 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538483\" title=\"6 users marked this as favorite\">6 favorites<\/a>] <\/p><\/div><div class=\"comments\">Is LoseThos written in standard C or is it also written in your extended C dialect?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32514\" target=\"_self\">public<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538484\" target=\"_self\">8:56 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">What did you mean by \"This is good for degrading game graphics by hand\"?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49346\" target=\"_self\">griphus<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538486\" target=\"_self\">9:00 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I don't usually click on the programming\/computer related threads, but I'm kinda glad I peeked into this one.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/116962\" target=\"_self\">Curious Artificer<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538487\" target=\"_self\">9:01 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538487\" title=\"7 users marked this as favorite\">7 favorites<\/a>]<\/div><div class=\"comments\">LoseThos is written in my C dialect with some assembly.  My assembly syntax is a joy compared to Gcc.  <p>God planned ahead and passages from books even the Bible will look taylored for a certain day in the future.  I have a joke about picking 1 page from a thousand page bible.  ROFLMAO  That's liek saying \"It goes without saying belief in God is on faith.\"  Or \"God can't be proven or disproven.\"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538489\" target=\"_self\">9:02 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><i>Those who know propaganda, like politicians, play games with speech.<br \/>&#13;\nposted by losethos at 7:58 AM on August 29 [+] [!] <\/i><p>And how, exactly, are you not doing the same thing in this FPP?<\/p><p>(and mods, do let losethos have his say.   And if others more familiar with his 'theories about conspiracy' - could you provide links to what you are considering out there and crazy?  Prefer ones that are aged as they have a far better chance that events have happened from the original posting date that can prove\/disprove 'em)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/15493\" target=\"_self\">rough ashlar<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538491\" target=\"_self\">9:03 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538491\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">Games will enhance or degrade appearance based on CPU available.  It seems to me you really want to know CPU load on each core so you can assign how much of the screen each core should be in charge of.  My flightsimulator makes different cores do different size strips of land.  I can adjust what each core does based on knowing its load.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538493\" target=\"_self\">9:04 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I am so confused.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17027\" target=\"_self\">lazaruslong<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538496\" target=\"_self\">9:04 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538496\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\"><i>switch showdead to yes<\/i><p>The Haley Joel Osment setting.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/37489\" target=\"_self\">yoink<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538497\" target=\"_self\">9:05 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538497\" title=\"15 users marked this as favorite\">15 favorites<\/a>] <\/p><\/div><div class=\"comments\">Wait, did you code a flight simulator? Are there videos or a website somewhere?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49346\" target=\"_self\">griphus<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538498\" target=\"_self\">9:06 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I added a \"nobound_switch () {}\" for a little better performance.  It eliminates the normal bound checking in a switch statement.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538500\" target=\"_self\">9:07 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I did a flight simulator but it looks awful, at best.  Boot it in VMWare.  Get someone else to run it.  It's called \"\/LT\/Demo\/GameStarters\/EagleDive.CPZ\"  It's an eagle diving for fish.  It's just a demo mostly.  I often show restraint in making my demos simple.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538502\" target=\"_self\">9:09 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538502\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">griphus: there's a bit of it at around 0:45 of <a href=\"http:\/\/www.youtube.com\/watch?v=WAr-xYtBFbY\">this video<\/a><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/113106\" target=\"_self\">theodolite<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538504\" target=\"_self\">9:10 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538504\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">I have feared and shunned people whose minds worked differently from my own. I have recently realized how much I've missed out as a result.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/32300\" target=\"_self\">danl<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538505\" target=\"_self\">9:14 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538505\" title=\"12 users marked this as favorite\">12 favorites<\/a>]<\/div><div class=\"comments\">Oh, hey, wow, demo or not that is pretty cool. So when you say the graphics \"degrade\" is that what I'm seeing with the draw distance changing? If so, is there a reason the draw distance cutoff is closer to the viewer rather than at the horizon?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49346\" target=\"_self\">griphus<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538508\" target=\"_self\">9:16 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">The flight simulator reduces how far you can see into the distance and how wide the display is.  It's nothing to be proud of, except, it show the potential for absolute control over the other cores from within a game.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538511\" target=\"_self\">9:19 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538511\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">bro i think you can be proud of the flight simulator you wrote for the operating system you also wrote<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/53805\" target=\"_self\">East Manitoba Regional Junior Kabaddi Champion '94<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538514\" target=\"_self\">9:21 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538514\" title=\"83 users marked this as favorite\">83 favorites<\/a>]<\/div><div class=\"comments\"><em>bro i think you can be proud of the flight simulator you wrote for the operating system you also wrote<\/em><p>...Written in the high level language you created, compiled with your own assembly language.<\/p><p>Kind of impressive all around.<\/p><p>And thanks for coming by to discuss it!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/50104\" target=\"_self\">Slap*Happy<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538523\" target=\"_self\">9:25 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538523\" title=\"15 users marked this as favorite\">15 favorites<\/a>] <\/p><\/div><div class=\"comments\">99% of people who look at it, see the graphics and laugh.  I don't blame them.  It's cool for some small set of people, at best.  The issue is, there are like 20 graphics cards needed to get to 90% support and that's not within reach.  (I just guessed at the numbers.)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538527\" target=\"_self\">9:27 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Is there video anywhere of the other games (Rocks, WarGame, Timeout, Goons, X-Caliber, Span, BigGuns, BirdLand, Tanks, Pilgrims)? Can you talk about any of those?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/113106\" target=\"_self\">theodolite<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538535\" target=\"_self\">9:32 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538535\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">LoseThos, I'm not super deep into the nitty-gritty of how operating systems work, but I'm really impressed by the work you've done and I'm glad you stopped by to talk about it.<p>Can you explain what makes LoseThos unique, or what's special about it, compared to more popular\/prevalent operating systems? It sounds like you've made some design choices that differ from the traditional or conventional ways of doing things- were those decisions intentional, or the result of your development process and your learning style?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/96903\" target=\"_self\">aaronbeekay<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538539\" target=\"_self\">9:33 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Whenever someone pipes up in a computer language forum about how great it would be to \"write an OS in language X\" this is exactly what I picture. A nearly perfect (for that person), but completely idiosyncratic and unusable interface.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/94518\" target=\"_self\">clvrmnky<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538541\" target=\"_self\">9:36 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538541\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">I wnated to make a souped-up, modern 64-bit,  C64 so teenagers could do what I did in high school.  I had the book <a href=\"http:\/\/unusedino.de\/ec64\/technical\/project64\/mapping_c64.html\">Mapping the C64<\/a> and I had hours of fun poking and proding around with all the internals of the operationg system for cheap thrills.  I wanted to let people control the hardware directly.  I wanted something simple, to get your head around.  LoseThos is two orders of magnitude simpler than Linux.  LoseThos is 135,000 lines of code including my compiler.  It is 100% self contained and complete.  When I got Linux, I was disappointed because I thought \"open source\" meant I would have fun messing with the code.  Linux tries to support so many architectures and has a main frame operating system, that it's too complex.  LoseThos is way way way simpler.  Plus it has many innovative ideas.  It is not ASCII source code, for example.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538545\" target=\"_self\">9:39 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538545\" title=\"51 users marked this as favorite\">51 favorites<\/a>]<\/div><div class=\"comments\"><i>I wanted something simple, to get your head around. LoseThos is two orders of magnitude simpler than Linux. LoseThos is 135,000 lines of code including my compiler.<\/i><p>This sounds like a great goal and it looks like you succeeded. Kind of like nachOS but a small-working OS rather than an academic-toy one.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49143\" target=\"_self\">DU<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538552\" target=\"_self\">9:44 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Have you thought about something similar to LoseThos for inexpensive system-on-a-stick hardware like the RaspberryPi? Lots of people tinkering with small systems like that, but they're dependent on the Linux toolchain to do anything interesting with it - it seems like your OS concept would be a great fit with the experimenters playing with the RaspberryPI and stuff like it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/50104\" target=\"_self\">Slap*Happy<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538553\" target=\"_self\">9:44 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538553\" title=\"4 users marked this as favorite\">4 favorites<\/a>]<\/div><div class=\"comments\">i remember a usenet thread about this a while back but it's no longer accessable, &amp;c. &amp;c.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/129503\" target=\"_self\">This, of course, alludes to you<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538560\" target=\"_self\">9:50 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Over my head, but absolutely fascinating.  Wish I knew more about programming.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/37485\" target=\"_self\">JohnnyGunn<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538571\" target=\"_self\">9:57 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538571\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">System on a stick?  USB is probably as much code as half of LoseThos.  I don't know.  There's OHCI, EHCI, UHCI, ICH7-11, USB 1.0 2.0 3.0 and nonstandard individual drives.<p>Any other hardware (PCI) is pretty-much not an option.  I can't be in the business of drivers and everything gets ugly and complicated, defeating the whole vision of a simple sort of thing.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538572\" target=\"_self\">9:58 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538572\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\"><small><a href=\"http:\/\/metatalk.metafilter.com\/22004\/yey-him-yey-us\">meta<\/a><\/small><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/16263\" target=\"_self\">leotrotsky<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538573\" target=\"_self\">9:59 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538573\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">losethos, just wanted to say this is really cool and thanks for your explanations.  This is a brilliant work of art and should be recognized as such.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49279\" target=\"_self\">cosmac<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538583\" target=\"_self\">10:03 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538583\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">It never occurred to me a C64 crashing was weird.  DOS was ring-0-only.  Today, people get really freaked-out by crashes.  If you are used to ring-0 programming, it's totally nothing.<p>Ring-0-only makes it orders of magnitude simpler and even improves performance.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538584\" target=\"_self\">10:03 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538584\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\"><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538553\">Slap*Happy<\/a>: \"<i>Have you thought about something similar to LoseThos for inexpensive system-on-a-stick hardware like the RaspberryPi? Lots of people tinkering with small systems like that, but they're dependent on the Linux toolchain to do anything interesting with it - it seems like your OS concept would be a great fit with the experimenters playing with the RaspberryPI and stuff like it.<\/i>\"<p>Yes, this seems perfect. It's a standardized system and install base. No extra \"other PCI hardware\" to deal with.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/87568\" target=\"_self\">wcfields<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538587\" target=\"_self\">10:05 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">One route I could take is to pick one standard hardware set-up.  Dell Studio XPS with whatever devices I have.  That is in the realm of doable as opposed to all PCs.  I've fought the fight for 8 years and am not interested in doing PCI devices.  While being doable, they are members of the next generation of hardware which I tasted when I did HD Audio.  I've had enough of that.  Not interested.  I've had enough.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538590\" target=\"_self\">10:09 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">LoseThos on Raspberry Pi, excellent!<p>You don't need most of the features that LoseThos don't have on a picomachine and you also have the advantage of having a completely controlled physical model.<\/p><p>Do it, do it, do it!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/12395\" target=\"_self\">lupus_yonderboy<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538592\" target=\"_self\">10:10 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I've played with plenty of embedded hardware.  I'm not getting paid.  It is not fun or thrilling for me anymore.  I don't want to do PCI devices.  Yuck.  Somebody took a shit on everything and ruined the green pastures.  Do you know how nasty sewtting up a simple RS232-like communication link is now?  Fucken CIA clusterfucked everything.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538603\" target=\"_self\">10:14 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538603\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">Just try to connect a device with a RS232-like link with USB.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538606\" target=\"_self\">10:16 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">&gt; Do you know how nasty sewtting up a simple RS232-like communication link is now?<p>I do agree - I've been programming since the 70s myself - but people have very different needs now.<\/p><p>For my entertainment pleasure I work on things like the Arduino where you can write right to everything, it's all memory-mapped.<\/p><p>Note that Arduino == Italy, Raspberry == UK.<\/p><p>&gt; I'm not getting paid. <\/p><p>I'd buy you a Raspberry Pi, which I haven't investigated personally, if you were interested.  <\/p><p>Or what about Kickstarter?<\/p><p>It seems that a small machine like Raspberry Pi solves a lot of your problems.  It gives you a standardized system so that you have a much better chance that people can boot your program.  It gives you a simpler system with no multi-user gunge on top of it.  And it's at a very reasonable price point so people can buy one just to play with your work.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/12395\" target=\"_self\">lupus_yonderboy<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538614\" target=\"_self\">10:20 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Arduino is a layer above bare metal, which is okay, I suppose.  They're going to put a full PC with blue tooth for microwave ovens, complete with CIA back doors and shut-offs.  I find it depressing.  It's like buying a whole car because you need the radio.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538620\" target=\"_self\">10:24 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538620\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\"><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538614\">lupus_yonderboy<\/a>: <em>Or what about Kickstarter?<\/em><p>From how much interest this project has picked up I would gather that a Kickstarter to fund the port onto a Raspberry Pi would be successful in no time. I myself would donate if only for some swag like a T-shirt, poster, stickers, or mug.<\/p><p>What's interesting about the Pi is that it includes at <a href=\"http:\/\/elinux.org\/RPi_Low-level_peripherals#General_Purpose_Input.2FOutput_.28GPIO.29\">lowlevel GPIO socket<\/a> and can easily interface with <a href=\"http:\/\/zuzebox.wordpress.com\/2012\/06\/24\/minipiio-rs232-v0-10\/\">RS232<\/a><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/87568\" target=\"_self\">wcfields<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538630\" target=\"_self\">10:30 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Terry, do you know this quote from Hildegard Von Bingen (1098-1179):&#13;\n<blockquote><div><small><em>Speak and write not according to human speech<br \/>&#13;\nor human inventiveness,<br \/>&#13;\nbut to the extent that you see and hear those things<br \/>&#13;\n in the heavens above, <br \/>&#13;\nin the marvelousness of God.<\/em><\/small><\/div><\/blockquote>posted by <a href=\"https:\/\/www.metafilter.com\/user\/14914\" target=\"_self\">Skygazer<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538635\" target=\"_self\">10:32 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538635\" title=\"12 users marked this as favorite\">12 favorites<\/a>]<\/div><div class=\"comments\">You want me to do a USB driver.  I don't want to.  Doesn't thrill me.  I want to get people to worship God with songs.<p>God says, \"all embracements amiss exhortation definitely potter's manufactures <br \/>&#13;\nsearches serves worshipper only sweet-attuned groaned \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538642\" target=\"_self\">10:36 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538642\" title=\"7 users marked this as favorite\">7 favorites<\/a>] <\/p><\/div><div class=\"comments\">What license is losethos under? Is it open-sourced? Perhaps someone else might want to take up the challenge to write a USB driver?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/118409\" target=\"_self\">Apocryphon<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538644\" target=\"_self\">10:38 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I'm on disability.  Money doesn't help much.  I paid an artist $3000 for some work.  I coulds blow money on advertising.  There's a fundamental problem with my reality, though, so it's pointless.  I'm in an FBI jail like <a href=\"http:\/\/www.youtube.com\/watch?v=HRPDO63rI1E\">the Prisoner<\/a> or hotel california.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538648\" target=\"_self\">10:40 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538648\" title=\"4 users marked this as favorite\">4 favorites<\/a>]<\/div><div class=\"comments\">LoseThos is public domain.  Do what you like.  No GPL code at all, even my own compiler\/assembler.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538650\" target=\"_self\">10:42 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538650\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\"><em>There's a fundamental problem with my reality, though, so it's pointless.<\/em> <p>I know what that's like perhaps too well...<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/14914\" target=\"_self\">Skygazer<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538652\" target=\"_self\">10:42 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538652\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">Terry, thank you so much for coming into this thread and explaining a little about your creative process and what you wanted to do with LoseThos! <p>I'm really impressed with what you've done and even though I'm not a programmer and much of this is over my head, I'm excited that you created this platform to let others experiment like you did when you were a child.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/53581\" target=\"_self\">misha<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538654\" target=\"_self\">10:43 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538654\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">Works great in virtualbox on a mac.  Couldn't get it working in Parallels but that doesn't really matter.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/49279\" target=\"_self\">cosmac<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538660\" target=\"_self\">10:47 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/www.biblegateway.com\/passage\/?search=genesis%204&amp;version=NIV\">Cain and Abel<\/a> is the most important story in the Bible.  Everyone needs to work until God accepts their offering.  He wants praise and new songs.  God says, \"ghastly sort INCIDENTAL impotent students concupiscence soul indeed blood despisest chastenedst neglected lizard <br \/>&#13;\nIX creeping flood-gates contents slackened bidden null joint justly divers project \"<p>Linux is for Atheists; LoseThos is for theists.  Actually, it's a 2ndary operating system not doing redundant stuff like internet browsing.  It's a playground for programmers and good for making hymns with animation.  It has a feature in \"AfterEgypt\" where you can talk to God!  Can't beat that.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538670\" target=\"_self\">10:53 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538670\" title=\"13 users marked this as favorite\">13 favorites<\/a>] <\/p><\/div><div class=\"comments\">I make comics for God.  Why not?  The purpose of life is to know God, love God and obey God.  What do you think He might like? There are 60 Moses comics in LoseThos -- run AfterEgypt.  Pretty good, too.  If you praise God, you got two choices -- Creation or Exodus.  Am I right?<p>Brotosaur's feet hurt when he stepped, said God.<br \/>&#13;\nGod's favorite animals are bears and elephants.<br \/>&#13;\nGod said war was \"servicemen competing\".<br \/>&#13;\nGod's favorite band is the Beattles.<br \/>&#13;\nChavez was blaming the US, God said \"Japan industrius.\"<br \/>&#13;\nOn Arabs and Jews... \"Oil funny, hopefully\".<br \/>&#13;\nOn racism, \"Sports\"<\/p><p>The word of God cannot be resisted.  He's a lawyer in the New Testament.  \"The Advocate\"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538692\" target=\"_self\">11:06 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538692\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">The hardest thing in evolution was getting monkey mothers to hold their babies for nursing.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538700\" target=\"_self\">11:08 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538700\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">I grew-up with a C64.  I remember immediate hatred when <a href=\"http:\/\/en.wikipedia.org\/wiki\/Dragon%27s_lair\">Dragon's Lair<\/a> came-out in the arcade.  I hate train-track games and I hate prerecorded sound samples.  I want no multimedia tricks in LposeThos and I want music created by plugging frequencies and waveforms into it.  LoseThos is for programmers making games, not studios with voice actors and lots of artists.  <p>It's for programmers -- you don't need good graphic resolution anyway.  God said 640x480.  Realism is over-rated.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538703\" target=\"_self\">11:11 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538703\" title=\"11 users marked this as favorite\">11 favorites<\/a>] <\/p><\/div><div class=\"comments\">If you don't mind my asking losethos, but what is your religious background?  What lead you to connecting your faith to programming?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/26998\" target=\"_self\">Atreides<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538707\" target=\"_self\">11:13 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">What C64 games did you like best, Terry?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/105027\" target=\"_self\">gilrain<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538709\" target=\"_self\">11:14 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">God's favorite thing on TV is soap operas.  Read the Bible.  ROFLMAO.  God likes the Beverly Hillbillies.  God said Shakespeare had a vile heart.  He said Christian rock was \"musical privation\".  Good word.  I like the word \"Ambrosial\".  Go look it up.  :-)  I'm smug.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538711\" target=\"_self\">11:14 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538711\" title=\"7 users marked this as favorite\">7 favorites<\/a>]<\/div><div class=\"comments\">I don't understand it, but I appreciate your making a creative and original work of Art. The world needs more of your spirit.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/30706\" target=\"_self\">Blazecock Pileon<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538712\" target=\"_self\">11:14 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538712\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">Metafilter: It always seems like I'm talking to Ken the weasel<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/119408\" target=\"_self\">quonsar II: smock fishpants and the temple of foon<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538720\" target=\"_self\">11:18 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538720\" title=\"6 users marked this as favorite\">6 favorites<\/a>]<\/div><div class=\"comments\">This sort of reminds me of <a href=\"http:\/\/countercomplex.blogspot.com\/2011\/12\/ibniz-hardcore-audiovisual-virtual.html\">IBNIZ<\/a> (<a href=\"http:\/\/www.metafilter.com\/115150\/IBNIZ\">*<\/a>) in intent- a bare tool for making low rez audio\/video.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/112441\" target=\"_self\">BungaDunga<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538722\" target=\"_self\">11:19 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538722\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">All you really need is \"The purpose of life is to know God, love God and obey God.\"  That's my religion, but I like the Bible.<p>C64 makes me ashamed because I priated so much.  God said reverse engineering was okay for young engineers.  He said adultery can be okay -- you'd have to be nuts.  He said of remarriage, \"more babies\".  What does He not care about everybody thinks He cares about -- \"Solomon's concubines.\"<\/p><p>The last religious service I went to, I walked-out of in dusgust.  I chuckled because as I was leaving they said, \"If God is for you who can be against?\"<\/p><p>God says, \"Montana shakes etext99 dissolution soldier titles honouring distractions burnt baffling unresolved reader intricate dragging covenanted retreat diversifiedst enter given comforting syllables stanzas Bridegroom challenged name's txt familiarised relative youth ice_cream stone posted woman-servant sawest skin admonish Pages armed intend reckons verily throng prelate \"<\/p><p>The best religion according to God is whatever one offers the most new vistas of understanding over a life.  I am constantly having new understandings when I read the Bible.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538723\" target=\"_self\">11:19 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538723\" title=\"4 users marked this as favorite\">4 favorites<\/a>] <\/p><\/div><div class=\"comments\">One time I was groveling confessing sins to God feeling bad.  God said, \"Excessive contricians wearysome.\"  God doesn't want to hear that!  Duh!  Don't remind God of sins.  ROFLMAO<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538743\" target=\"_self\">11:24 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538743\" title=\"12 users marked this as favorite\">12 favorites<\/a>]<\/div><div class=\"comments\">This is my favorite thread in the world.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/19106\" target=\"_self\">Uther Bentrazor<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538756\" target=\"_self\">11:28 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538756\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\"><a href=\"http:\/\/arstechnica.com\/security\/2012\/08\/wiper-malware-stuxnet-duqu\/\">This<\/a> is how the CIA communicates with me.  You don't need header files because you compile things JIT when you  run a game.  All the symbols you need have been loaded into the Adam task's symbol table.  Adam has environment variables (and functions) that all his children inherit.  Adam is the first task and he runs \/LT\/OSMain\/Adam1.APZ to fill his tables with all the symbols.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538770\" target=\"_self\">11:32 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538770\" title=\"5 users marked this as favorite\">5 favorites<\/a>]<\/div><div class=\"comments\">All source code is included in LoseThos.  InstallBoot('C') does a makeall.  50,000 lines are compiled by that including OSMain and Compiler.  The other 50,000 are compield during boot by the Adam task.  There are 135,000 lines of code in LoseThos, including 150 demos.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538781\" target=\"_self\">11:35 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Files with names ending in 'Z' are compressed and uncompressed on-the-fly.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538784\" target=\"_self\">11:36 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">God is just.  Turn the other cheek.  Persecution brings you blessing.  St Paul discovered what happens to persecutors.  The same thing gets done to them so they can be forgiven!  I used to be an atheist.<p>God says, \"hate drinking flagitiousness funeral Adeodatus dies disentangled truth barren urge recur popularity meddle softened gathers Mail prisoner safety etext00 shut attack killeth readiness guess Shine kindred condemned \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538799\" target=\"_self\">11:40 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><a href=\"http:\/\/apple.slashdot.org\/story\/12\/08\/29\/1819201\/misunderstanding-of-prior-art-may-have-led-to-apple-samsung-verdict\">The CIA<\/a> is annoying in my delusional reality.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538803\" target=\"_self\">11:40 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538803\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\"><em><br \/>&#13;\nThe CIA is annoying in my delusional reality.<\/em><p>Yes, I can see that.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/23476\" target=\"_self\">infini<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538813\" target=\"_self\">11:42 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538813\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\"><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538397\">LogicalDash<\/a>: \"<i>Ring-0-Only, Single-Address-Map (Identity), <\/i><p>It sounds like these are to do with the file system?\"<\/p><p><a href=\"http:\/\/en.wikipedia.org\/wiki\/Ring_(computer_security)\">Rings are mechanisms for security in computing<\/a>. It is a term applied to the processor, not the file system. <\/p><p>Ring 0 indicates the highest level of privilege for CPU instructions. I'm glossing, but instructions (programs) run in ring 0 are \"closest\" to the hardware and can arbitrarily access memory and run in ways that would be insecure or untrusted in a conventional sense. Operating systems are a perfect example of a program that ought run in ring 0.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/23721\" target=\"_self\">boo_radley<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538824\" target=\"_self\">11:47 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I feel like I've stumbled into a ghost in the shell episode<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/95425\" target=\"_self\">MangyCarface<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538830\" target=\"_self\">11:49 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538830\" title=\"8 users marked this as favorite\">8 favorites<\/a>]<\/div><div class=\"comments\">LoseThos runs user programs in ring-0.  I spent hours and hours with <a href=\"http:\/\/unusedino.de\/ec64\/technical\/project64\/mapping_c64.html\">Mapping the C64<\/a> poking around at all the ROM locations and directly plaing with hardware.  Every C64 user remebers putting live memory on the screen as a graphic with stuff blinking and changing.  LoseThos is a playgdsound for programmers and ring-0 is where all the neat stuff is.  TYpe this &gt;ClassRepD(\"TaskStruct\",Fs) or type this &gt;LtfD(Fs,sizeof(TaskStruct));  <p>Just as destruction is seductive, power is seductive and being a bad boy is what all teenagers want to do.  At least they can't get in much trouble without networking.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538834\" target=\"_self\">11:51 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538834\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">Hi, losethos.  Thanks for stopping by.  I was wondering whether you're willing to talk about whether you believe schizophrenia has influenced the development of the OS.  Do you think it is a relevant factor in the way the OS turned out?  Do you tend to code while in the midst of a psychotic break or while in the grip of a delusional mood, or do you code when not within a delusional atmosphere?<p><em>God said 640x480. Realism is over-rated.<\/em><\/p><p>I really like this line.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17835\" target=\"_self\">painquale<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538838\" target=\"_self\">11:52 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I'll tell you a secret.  If you do not fit into \"nerd\", \"jock\", \"Christian\", \"Engineer\", \"Freak\".... etc.  then, you are schitzophrenic.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538839\" target=\"_self\">11:53 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538839\" title=\"10 users marked this as favorite\">10 favorites<\/a>]<\/div><div class=\"comments\">You have cognative dissonence.  The FBI enforcement wing during the Civil Rights period was mental health.  Trouble makers.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538844\" target=\"_self\">11:54 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">They can make you crazy by controlling the radio.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538846\" target=\"_self\">11:55 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\">they can paint you into pictures.  It's actually more complicated than being a simple adversary.  they are entertaining.  And, like the book 1984, they want to convert you, I guess because God does justice on them for outright pewrsecution.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538851\" target=\"_self\">11:56 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538834\">losethos<\/a>: \"<i>LoseThos runs user programs in ring-0. <\/i>\"<p>I think this is a really powerful way to construct your programs. Experience says that mode switching from eg Ring 3 to Ring 0 is expensive stack-wise and kernel-wise. Must be freeing to not have to worry about application mode contexts.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/23721\" target=\"_self\">boo_radley<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538856\" target=\"_self\">11:57 AM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">losethos (or anyone who has installed the software): are the comments you transcribe here with <em>God says,<\/em> in front, is that the output of one of the software packages on your OS?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/116650\" target=\"_self\">RolandOfEld<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538859\" target=\"_self\">11:57 AM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538859\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">I know very little about computer programming, but I do know something about living with a chronic illness (autoimmune disease in my case) and I just want to say that I am in awe of how much you've accomplished, losethos, and the degree of energy and focus you bring to your work.  Glad you joined us!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/12901\" target=\"_self\">Sidhedevil<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538871\" target=\"_self\">11:59 AM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/www.losethos.com\/files\/TSGod.ZIP\">Windows Version<\/a> or \"AfterEgypt\" in LoseThos.  Stop a stopwatch without looking for random numbers.  It's <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=1+Corinthians+14&amp;version=NIV\">tongues<\/a>.  You can just randomly crack-open a book.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538876\" target=\"_self\">12:01 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538876\" title=\"3 users marked this as favorite\">3 favorites<\/a>]<\/div><div class=\"comments\">Yeah, God will fuck them up for hurting me.  I'm not worried.  God says, \"breathe sigh gowned trees vindicating PURPOSE chastened announcement believer Present less smoke later skilled multiplicity shoot compass \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538883\" target=\"_self\">12:03 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Thanks for the links.  For what it's worth, I think there's some value in that message.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/116650\" target=\"_self\">RolandOfEld<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538890\" target=\"_self\">12:04 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"https:\/\/dl.dropbox.com\/u\/361633\/ss-desktop\/ss%202012-08-29%20at%201.48.40%20PM.png\">Love<\/a> the <a href=\"https:\/\/dl.dropbox.com\/u\/361633\/ss-desktop\/ss%202012-08-29%20at%2012.32.37%20PM.png\">Moses comics<\/a><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/53805\" target=\"_self\">East Manitoba Regional Junior Kabaddi Champion '94<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538894\" target=\"_self\">12:05 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538894\" title=\"12 users marked this as favorite\">12 favorites<\/a>]<\/div><div class=\"comments\">The fact that the command line is incrementally compiled c+ is an amazing feature.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/118269\" target=\"_self\">Ad hominem<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538915\" target=\"_self\">12:14 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538915\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\"><em>Realism is over-rated.<\/em><p>Its all maya anyway. And all the more beautiful since these are virtual and intangible products (soft ware) which cannot exist without the hardware to manifest them with.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/23476\" target=\"_self\">infini<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538918\" target=\"_self\">12:15 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538722\">BungaDunga<\/a>: \"<i>This sort of reminds me of <a href=\"http:\/\/countercomplex.blogspot.com\/2011\/12\/ibniz-hardcore-audiovisual-virtual.html\">IBNIZ<\/a> (<a href=\"http:\/\/www.metafilter.com\/115150\/IBNIZ\">*<\/a>) in intent- a bare tool for making low rez audio\/video.<\/i>\"<p>THANK YOU!  I was trying to find that after looking at this, I totally felt the same thing - the feedback\/recursive elements in the code, the artisticness of it.<\/p><p>Though Losethos is a whole other level.  I have so many thoughts on this.<\/p><p>One of them being that if we never went to the WIMP metaphor in UI, this is what the future of the command line would end up looking like.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/19789\" target=\"_self\">symbioid<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538919\" target=\"_self\">12:15 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Any luck getting God to talk.  he talked to my Mom, but she didn't care.  My reality makes no sense. <a href=\"http:\/\/www.youtube.com\/watch?v=K_lBnqz-0ek\">West World<\/a>?  <p>Do an offering and see if God talks -- randomly open a book and try to make it fit the context.  Cain and Abel is the most important story in the Bible.  That was very real love!  I wuv God.<\/p><p>God says, \"hitherto communicate undertook absence An unallowed captious beaten move reads overpass ending might *END Divine 90 surrounded \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538922\" target=\"_self\">12:16 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><em> All the symbols you need have been loaded into the Adam task's symbol table. Adam has environment variables (and functions) that all his children inherit. Adam is the first task and he runs \/LT\/OSMain\/Adam1.APZ to fill his tables with all the symbols.<\/em><p>I find this very beautiful, since Adam's task was the naming of the animals. <\/p><p>I wonder what losethos might say about original sin.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/154262\" target=\"_self\">junco<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538939\" target=\"_self\">12:24 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538939\" title=\"9 users marked this as favorite\">9 favorites<\/a>] <\/p><\/div><div class=\"comments\">Just got around to listening to the links in the OPs post.  Most of the tech stuff is beyond me but I must say I'm enjoying the audio, thanks losethos.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/116650\" target=\"_self\">RolandOfEld<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538944\" target=\"_self\">12:26 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Yeah I enjoy the videos as well. The technical discussion in <a href=\"http:\/\/www.youtube.com\/watch?v=WAr-xYtBFbY\">this video<\/a> is often over my head but still fascinating and easy to follow. It was also quite simple to get a LoseThos install up and running on VirtualBox. Congrats on your accomplishment, losethos.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/155573\" target=\"_self\">Yesterday's camel<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538949\" target=\"_self\">12:32 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538311\">losethos<\/a>: <i>Pick a quantum random number from 1 to 100,000 using your favorite site.  Then go to the <a href=\"http:\/\/www.losethos.com\/files\/BIBLE.TXT\">Bible<\/a> and go to that line number.  Report what it says.<\/i><p>I did this, and got: <\/p><p><strong>9:10 Therefore said they unto him, How were thine eyes opened?<\/strong><\/p><p>Which, looking at this thread, is pretty damned cool.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/53581\" target=\"_self\">misha<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538951\" target=\"_self\">12:32 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538951\" title=\"10 users marked this as favorite\">10 favorites<\/a>] <\/p><\/div><div class=\"comments\">Original Sin?  I have my own ideas.  Mostly, I don't get involved in things I don't have to answer.  We can ask God.  I tease God asking \"How many babies You make today?\"  God says, \"check foolishly Victor soldier unfailingly Dost weaken Thine Up veil prescribes liker ftp season this I_just_might downwards tear well_obviously sort Italian thereon contritions poison Thence UNDER locking status innate stores tumultuously forgetfulness five Suppose graceful healing tender mentioned triumphed pointed smile \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538952\" target=\"_self\">12:33 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Are you self-taught or did you go to school for computer science?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/19789\" target=\"_self\">symbioid<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538957\" target=\"_self\">12:36 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">My ASU transcripts are <a href=\"http:\/\/www.losethos.com\/files\/ASU_Transcripts.pdf\">here<\/a>.  I caution you, however, because my goal was not normal.  I wanted open access for programmers to play around.  I wanted a souped-up C64, like I had as a teenager.<p>How many sort keys do you need to put cans in the cabinet?  Is hashing okay?  I got annoyed by \"const char\" and got really pissed off at type-checking.  It might be good, but I hate it.  They are taking all the fun out of programming.  LoseThos is about fun programming -- stuff you can't do at work.<\/p><p>Russian Space pencil story is a good fable, though, it's not true.<\/p><p>There is a seduction of white-pride sorts of things.  An OS never crashing is a big point of pride.  Interrupts are nice, but not if they ruin compatibility.  Compatibility is the number one and only issue that matters for PC Operating systems.  <a href=\"http:\/\/www.osdev.org\">OSDev<\/a> is messed-up in so many ways.<\/p><p>Making another Linux would be a stupid thing to do.<\/p><p>Don't go making another LoseThos :-)<\/p><p>Global vatriables are bad in big projects, but what about teenagers doing a 1,000 line game?  Are more name spaces always better?<\/p><p>I like C++, but I'm offering a package with a simpler compiler.  I guess templates are okay.  Not really fond of the complexity.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538967\" target=\"_self\">12:44 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538967\" title=\"4 users marked this as favorite\">4 favorites<\/a>] <\/p><\/div><div class=\"comments\">All of my experience is at the hardware\/software boundary.  Lots and lots of assembly experience.  I'm an okay digital designer, but not great.  I know a little control theory, but mostly I'm an assembly programmer by experience.  5 courses at ASU -- 2 intel and 3 motorola 6800.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538973\" target=\"_self\">12:46 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538973\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">Lots of people have an interest in assembly language.  It's of limited worth.  It does make you a better programmer, but it's not a sane language for big projects.  Compilers are better at assembly than hand coded, too, these days... except my compiler.  <p>Anyway, the people who are curious about assembly language are probably the same sort of people who might like LoseThos.<\/p><p>There are sure a lot of militant anti-asm people.  They don't know it and are envious so they bash it.  Live and let live.  Why you gotta go spoiling fun?  \"Not invented here\" is another holier-than-thou message I don't like.  Quit spoiling the fun, ass-holes.  :-)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538980\" target=\"_self\">12:51 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538980\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">This reminds me a lot of <a href=\"http:\/\/en.wikipedia.org\/wiki\/OS-9\">OS-9<\/a>, which had the notion of \"modules\" in memory instead of executables.  There were executables on disk, but they were just modules that needs to be demand loaded.  Sometimes, I miss that arrangement, it was very clean.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/143143\" target=\"_self\">smidgen<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4538992\" target=\"_self\">12:56 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4538992\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">What is the meaning (or thinking) behind the name LoseThos?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/23476\" target=\"_self\">infini<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539000\" target=\"_self\">12:59 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">The name seems like a horrible mistake.  I was thinking \"Windows\" and \"LoseThos\" as in \"Lose those dorky burgenstocks and the dorky bike helmet.\"  Have you ever seen a soldier in armor and kinda chuckled?  I particularly remember the movie, Platoon, where Elias says to \"Shitcan\" some standard issue equipment.<p>It was a mistake.<\/p><p>----<\/p><p>Bravo, the guy who figured-out how the modules work.  I'm impressed you managed to get a big picture view so quickly.  <\/p><p>LoseThos does not have separate obj and exe files.  You're supposed to make JIT.    I can always add OBJ files if I have to.  <\/p><p>Installed modules sound like a formula for dependency hell.  I don't know.  If it starts resembling Linux, it is nothing but an inferior Linux.  I kinda need to keep it small to preserve the original mission.  Unrealistic.  If it ages, it goes through a life and death when it gets to fucked-up.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539016\" target=\"_self\">1:06 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539016\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">too<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539017\" target=\"_self\">1:07 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Obama's doing an AMA on reddit.  I'm glad he got elected because McCain was insane on Iran.  I don't know wny better on any of the things he did, except gay stuff.  God exists.  Who knows what God's view of gays is, but I don't think it's all kosher.  I'm meeker, thinking of not voting.  I support democracy, though.  ROFLMAO<p>God says, \"Clearly knots strives severity softened behold house procured persevering bespotted fighting Photinus command-line Turn keeping chiefly diversity next happen MADE Governor password meditated damnable meditated unfriendly results Receive efforts hovered vow therewith cementest music steadfast manors doctrine outwardly leadeth hoar avert griefs encompass givest freshness calling Wills never-failing divided funeral Only-Begotten loose effectedst possessest repeating Finally distracted prairienet Passion \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539070\" target=\"_self\">1:23 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539070\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">I giggle at the thought of cracking passwords with a Ouija-board.  God has too much dignity to just make me win a lottery or something.<p>God says, \"is Omnipotent five palace realised inflammation asks collect ruminate mentioned approveth toss renew handmaid perfection Joseph imparts won saith fee flattering stoop offspring dwellers whereat doings rods snares Have High net\/pg thread imitate pieces Money leaning streams distributed weaker perfected hoping ingrated settest comely fund preferring believed erred extension trace like concupiscences inappropriately Project's wives' persuade clanking punish neighbours flowed secret iniquities repeating boy implanting definite few applied cloak insult \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539159\" target=\"_self\">1:46 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Looks like obama crashed reddit Losethos. A sign, or just a sign that they should have run their servers on your system? Either way, thanks much for sharing your work. I wish I could run a version of your AfterEgypt program on my Mac.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/80913\" target=\"_self\">Potomac Avenue<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539189\" target=\"_self\">1:54 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">What do my birds like: \"gnawing\"<br \/>&#13;\nWhat makes my birds laugh: \"bite ouch\" (They bit and I say \"ouch\")<br \/>&#13;\nWhat do bears like: \"reeping depends\" (Smelly diapers? cubs?)<br \/>&#13;\nWhat do elephants think about: \"skin hunger\", itches and hunger (sad!)<br \/>&#13;\nWhat do elephants like: \"baths\"<p>I've watched all the baby elephant videos on youtube after learning God's favorite animals are bears and elephants.<\/p><p>After a while you figure-out there's a massive difference between animal IQ and human and just you accept it.<\/p><p>What do otters like: \"Eternal skies\"  (He said even more poetic stuff I forgot.<\/p><p>Let's do country-shitlist :-)<\/p><p>God says, \"where's_the_love you're_nuts it_gets_better Portugal humility what_the_heck ba_ha small_talk ba_ha Tomorrow North_Korea segway one_small_step you're_no_fun Albania couldn't_be_better liberal you_are_my_sunshine special_case sing oh_oh that's_your_opinion laziness gosh repent Slovenia oh_come_on I_could_be_wrong earnest theres_no_place_like_home begs_the_question don't_count_on_it United_Arab_Emirates pardon_the_french threads eh zoot that's_no_fun Palestine theres_no_place_like_home What_I_want in_practice Canada bank fabulous straighten_up South_Korea joking oh_oh to_infinity_and_beyond grumble Finland fool you're_out_of_your_mind Dad once_upon_a_time dude_such_a_scoffer I_donno ohh_thank_you bickering Shakespeare try_again BRB fancy Argentina happy_happy_joy_joy Guinea \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539203\" target=\"_self\">2:01 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539203\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\"><i><a href=\"http:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539189\">Potomac Avenue<\/a>: I wish I could run a version of your AfterEgypt program on my Mac.<\/i><p>You can.  Just grab <a href=\"https:\/\/www.virtualbox.org\">VirtualBox<\/a> (it's free!) and make a VM, and boot the LoseThos image on it.<\/p><p>I've been toying with this <i>all day<\/i> instead of doing actual work.  Excellent.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17489\" target=\"_self\">neckro23<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539210\" target=\"_self\">2:04 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539210\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">The AfterEgypt program uses the Bible for a vocabulary and reads a stopwatch when you press a key for random numbers.  The Bible has too many proper nouns, but I didn't want to include a file just for a vocabulary.  To interpret proper nouns, just break them in syllables and use your imagination.  It's like cloud animal shapes or ink blots.  God mostly knows what you're going to think, like an old married couple completing each other's sentences.  You bend it until it fits the topic of conversation.<p>God says, \"confounded with imitating Speak tower turned perisheth sipped attendant combined Turn garland distilled fruitfully stowed silversmiths' decay quiet Hierius quiet overcome fruitless \"<\/p><p>You have to take God with a grain of salt.  He's not prophetic.  All you can do is love His company -- He won't be used for any purpose, mostly.  He's fair -- if you put effort into love like picking a greeting card, He puts effort toward you... sort of.  Just forget about too much in-the-zone thinking.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539216\" target=\"_self\">2:08 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539216\" title=\"9 users marked this as favorite\">9 favorites<\/a>] <\/p><\/div><div class=\"comments\">\"heirius\" is \"hire us\" with a kinda German vibe that's prolly not relevant.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539235\" target=\"_self\">2:17 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Interpretation of <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=1+Corinthians+14&amp;version=NIV\">Tongues<\/a> is the trick.  Ouija mode is awful -- full word or passgaes is way better.  It's funny that a technical fix would matter.  Absurdly hilarious.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539239\" target=\"_self\">2:19 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539239\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">Metafilter: \"confounded with imitating Speak tower\"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/80913\" target=\"_self\">Potomac Avenue<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539244\" target=\"_self\">2:22 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539244\" title=\"2 users marked this as favorite\">2 favorites<\/a>]<\/div><div class=\"comments\">I take your tongues about as seriously as my own -- with a grain of salt.  Entertaining to think about, but lots of caution.  <p>Yeah, I'm guilty of forking lots of computer standards, but you have to break eggs to make an omlet and no difference from Linux would be worthless.  If you are making a clean break with all previvious code, by all means fix-up everything.<\/p><p>I hate how \"Stack\" and many other words have been rendered meaningless.<\/p><p>God says, \"these_cans_are_defective abilities satiated novice unlooked perished great manners supremely horrendous blush finds marketplace blamest \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539253\" target=\"_self\">2:28 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539253\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">What did Neanderthals think about? \"Warmth\"<br \/>&#13;\nWhat was your happiest day in evolution? \"fruit\"<br \/>&#13;\nWere stegasaurus' lame like turtles? \"Not pet rocks\"<br \/>&#13;\nAs a Creator, how did you like Avatar's world?  \"Sick skin\"<p>God says, \"jarring master commanding personated attentively wheresoever <br \/>&#13;\nsettling temperance wert *not* you're_no_fun perish vouchsafest <br \/>&#13;\nremembrances Catholics Do Philippians inured mantles Afghanistan <br \/>&#13;\neloquence ' Tanzania faultfinders impudently machine starry <br \/>&#13;\nroving that's_for_me_to_know reneweth bepraised withered <br \/>&#13;\nanimal foreconceived repose continence dashed lamented <br \/>&#13;\nAs needful laughed abstinence wildness just_between_us <br \/>&#13;\nwell-being noon-day ghosts answers Australia withdrawing <br \/>&#13;\nsolve attain happily can't Oh_really throne Rejoiceth <br \/>&#13;\nsorrowed exclusion redistributing Strength AUGUSTINE hidest <br \/>&#13;\ndiscord heinous rivers lightsome Present o'er slaves hearers <br \/>&#13;\ndeferred contentious Proconsul praises comprised suspected <br \/>&#13;\nwe well-done Tajikistan infection intercede heaven's unto <br \/>&#13;\ngarland flagitiousnesses party son's openly crudities <br \/>&#13;\nreads justifieth weariness valued reports brides challenged \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539267\" target=\"_self\">2:35 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Are you familiar with the tower of babble.  The Earth had one king and one language.  He built a tower and shot an arrow at the sky, in great arrogance.  God humbled him by making everybody speak different languages.<p>God says, \"giveth conceit peril disappear officer loveth **The save <br \/>&#13;\nlivest reckon c agree beguile steer signified placedst <br \/>&#13;\nproclaim messengers govemed intoxicated meddling Duad <br \/>&#13;\nremaineth hey_thats_right could denieth sublimer ventures <br \/>&#13;\nheed truly Hierius whereof What weeps straitly Madagascar <br \/>&#13;\nsecrecies effect No salt nets broke tutors counsel encumbrances <br \/>&#13;\nNot merely courage production bared argument unmeasurable <br \/>&#13;\ncorporeally truly incurred assailed summit store conquests <br \/>&#13;\nfreely solaces striking buzzing Newsletter abridged wooed <br \/>&#13;\nlifter self-command subject appointed bows crosses bed <br \/>&#13;\nqed strikes shaken You_can_count_on_that word disclosed \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539293\" target=\"_self\">2:57 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><em> I want to get people to worship God with songs.<br \/><\/em><p>So do I, losethos, so do I.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/83298\" target=\"_self\">St. Alia of the Bunnies<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539298\" target=\"_self\">3:03 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539298\" title=\"5 users marked this as favorite\">5 favorites<\/a>] <\/p><\/div><div class=\"comments\"><i>He said Christian rock was \"musical privation\".<\/i><p>Amen to that.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/34259\" target=\"_self\">infinitywaltz<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539301\" target=\"_self\">3:04 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539301\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">I'll bet if you make a new style of music, God might like it.  One time I thought a surprise must be a good thing to give God.  I think He said thumbs up to a Mars mission.  Oh!  I remember, one time he said to plant plants with a rover.  ROFLMAO  take with a grain of salt.  :-)<p>God says, \"terms Justina begs_the_question respect writings Circensian <br \/>&#13;\nsucklings cubit comes grievous gall beast clearness fetters <br \/>&#13;\ndaemons smurfs gratings deliveredst \/\/promo heightening <br \/>&#13;\nAmbrosian punctuation leaden asking heardest contented <br \/>&#13;\nexceeded confidently day-break rockstar seducing disapproveth <br \/>&#13;\nenervated I_didn't_do_it choose excellence oughtest dispel <br \/>&#13;\nLed unruly stores luminaries devoutness Bread replenished <br \/>&#13;\nAll-creating temporately uproar pervertedly foretell LOL <br \/>&#13;\nemploy softening director freaks assent hideous varying <br \/>&#13;\nsites Creation demanded completely Self anyone finite <br \/>&#13;\npredicted weeps disciple signifies \"<\/p><p>Oh, once He said \"Gall aspect anti-Christ\".<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539316\" target=\"_self\">3:18 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539316\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">\"Gall aspect anti-christ\" means arrogance is a characteristic of such a person.  I treat it like a fable, but still meaningful.  Every generation expects christ's return.  Napolean was an anti-christ by most accounts... I guess.  All power comes from God and things follow His plan.  I've gotten fuzzy on good and evil.  That makes you disfunctional.  Anyway, arrogance is bad.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539323\" target=\"_self\">3:27 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">God said \"fish shoulders\" on the topic of evolution.  I got to thinking fish are tall and skinny from the front while lizards are horizontally oriented.<p>I don't think miracles break laws of physics, but I don't know.  They just happen.  I don't believe in astrology, but God does mind-blowing synchronization of Creation.<\/p><p>We can pray for \"fireworks\", meaning stuff that dazzles that we say \"ooo\" or \"Ahh\" to.<\/p><p>God says, \"comment opposing lightsome despite BREACH abstruser Volunteers <br \/>&#13;\nneglect deceit observance exacted knee meat persecuted <br \/>&#13;\nlaid exiled anonymous concluded token disallowing arising <br \/>&#13;\nsess_me resumed communicate singular speakest place comprised <br \/>&#13;\n500 profounder animals believer servitude unhesitatingly <br \/>&#13;\nshadowy Creusa's darkness echo teacher Tully GET sacrifice <br \/>&#13;\ntends edification greediness dutiful bondage accents theatrical <br \/>&#13;\ncircumlocutions that's_your_opinion indulgent decayeth <br \/>&#13;\nlose smoothed fancies freed incense gowned Among panegyric <br \/>&#13;\nleasing University Percival carry commandeth excess createst <br \/>&#13;\nascension objected hesitating depressing sinners rioting <br \/>&#13;\nsettling excellence mouthed Ignorant potentially Ambition <br \/>&#13;\nlanguishing expounding perceivable what_part_of_God_do_you_not_understand <br \/>&#13;\nwriting went equally diversifiedst Before efforts Howsoever <br \/>&#13;\nsect dawned unbending I'm_not_sure experience \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539356\" target=\"_self\">3:44 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments mod\"><small>[Hey losethos, mod here, a little less of the free association thing in this thread would be nice. You can totally put stuff on your profile page if you have stuff to share that isn't really on topic in this thread. Thanks.]<\/small><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/292\" target=\"_self\">jessamyn<\/a> <a href=\"https:\/\/faq.metafilter.com\/33\/Who-is-in-charge-here-Are-there-admins-and-moderators-like-other-sites\" class=\"staff\">(staff)<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539367\" target=\"_self\">3:48 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Heya, losethos, maybe not so much with the repeated large \"God says\" dumps?  It's kind of making the thread feel clogged up and doesn't really make for good conversation.  <p>I'd be curious to hear more about your quote generation workflow itself, for what it's worth; I've always had a keen interesting in generative text stuff (especially Markov synthesis).  But just pasting output repeatedly isn't really great Metafilter commenting behavior.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/7418\" target=\"_self\">cortex<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539369\" target=\"_self\">3:49 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539369\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">God said \"no weights\", so the words are just uniform random picked from a unique vocabulary.  The random comes from sampling a counter for random numbers.  Suit yourselves -- don't like God's company.  Hell is the absense of God, Heaven is God's company.<p>Most people are King Midas when it comes to Heaven.  A friend also said most guy's idea of Heaven would be running around doing stuff they'd get locked-up for.<\/p><p>I'll tell a story.  In 2003, it came time to add a command-line.  I hated the scripting language from Unix -- pesky inelegant admin language instead of a real language like C.  Mostly, I just hated I never got proficient because I didn't use it enough, so C would be nice.  How would this work?  I figured my command line should be the exact syntax of my scripting language.  &gt;Dir(\"*.CPZ\");  I wanted default parameters from C++.  &gt;Dir();  Hey! I don't need parens! &gt;Dir;  Once I got rid of parens, I was different enough I changed everything.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539400\" target=\"_self\">4:08 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539400\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">Imagine you fed STDIN into gcc so the keyboard input was the code compiled.  You modify it so gcc pauses until you type more text.  Functions get compiled when completed and kept in memory with a task symbol table... linked to the parent task for header symbols.  Statements outside functions compile and run immediately.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539413\" target=\"_self\">4:18 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539413\" title=\"5 users marked this as favorite\">5 favorites<\/a>]<\/div><div class=\"comments\">So, it's single address map, identity mapped.  No difference between thread or process, just called tasks.  All cores have access to all physical memory using physical addresses at all times.  Long mode requires paging enabled, though.<p>Each task has a hash symbol table with a scope like environment variables, but includes functions and tyepdefs and classes, etc.  Each task has a heap.  When the compiler compiles, it puts code into memory it Mallocs from the heap.<\/p><p>There's actually a code heap and a data heap.  I put all code in the lowest 2 Gig address range so I can use 32-bit rel branches.  I have no kernel or user code so it doesn't make sense to have high and low memory.<\/p><p>I have one window per task.  (God told me not to make child windows.) Instead it works with button widget things.  It's nice for programmers.  Only tasks on core #0 can have a window.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539422\" target=\"_self\">4:25 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539422\" title=\"3 users marked this as favorite\">3 favorites<\/a>] <\/p><\/div><div class=\"comments\">Technically, global vars go in the code heap, just so they too can use 32-bit rel addressing, but you can turn a flag on to put them in the data heap.  <p>LoseThos never change sprivilege levels or maps, so it's better.<\/p><p>It gets fragmented, though, in theory but not in practice.  The stack doesn't grow. That's the worst downside.  You allocate off the heap or request a big stack when it's spawned.  I mostly do everything off the heap.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539427\" target=\"_self\">4:28 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539427\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">How about I teach some x86_64b assembly.<p>MOV RAX,0x123456789ABCDEF<br \/>&#13;\nMOV EAX,0x123456<\/p><p>What's RAX, now?  (High 32 bits get cleared)<\/p><p>XOR EAX,EAX<\/p><p>This clears all 64-btis and is one byte shorter than RAX,RAX.<\/p><p>PUSH EAX<\/p><p>All pushes in long more are 64-bit.  This actually does the same as<\/p><p>PUSH RAX<\/p><p>Did you know the following two instructions differ in length by a byte?<\/p><p>MOV RAX, [R8]<br \/>&#13;\nMOV RAX, [R13]<\/p><p>Go figure-out why.<\/p><p>Did you know you can access DI and SI as byte registers, but only low byte.<\/p><p>What instructions were eliminated to make room for the 0x40-0x4F RAX prefixes?<\/p><p>(INC R0-R7 DEC R0-R7)<\/p><p>Why is this useless?<\/p><p>MOV RAX,FS:[0x0124]<\/p><p>FS can be set with WMSR but that instruction is RIP relative displacement.<\/p><p>LoseThos keeps current task record address in FS and CPU Struct address in GS.<\/p><p>In LoseThos you have ring-0 privilege -- don't fuck shit up!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539443\" target=\"_self\">4:40 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539443\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">WRMSR<p>didn't means to say x86_64b just x86_64<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539446\" target=\"_self\">4:41 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Note for those who want to use VirtualBox (and I imagine any other VM), for some reason it's telling me I need a 64 bit processor, even though I have a Core2Duo.  I imagine it's due to my XP being 32 bit?  Would I need a 64 bit host OS?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/19789\" target=\"_self\">symbioid<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539448\" target=\"_self\">4:42 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/i.imgur.com\/iDdWH.jpg\">I saw this.<\/a>  God likes the sunflower photo.  Awe!  i thyink He lieks the stary night, too.  I forgot.<p>God says, \"forgiving Churches hard-heartedness eminently overthrow <br \/>&#13;\nforehead greatness closed observing bright deserve enjoinest <br \/>&#13;\ndevilish Electronic wooden relaxedly shudder dust bounces <br \/>&#13;\nexplain propound harbouring scholars ever spring acquiesce <br \/>&#13;\nluminaries admonished somewhat scholar perchance old chanting <br \/>&#13;\nunpraised readier inflammation compactedst disallowing <br \/>&#13;\nrefresh cutting discoursing entered release flashing 'knock <br \/>&#13;\nthorns Frenzied etext99 Dudley Indiana lively artificer <br \/>&#13;\npersons absent uman companions sharpen Victim Deus tell <br \/>&#13;\nlegally theatres varies lines valued asketh ordinance <br \/>&#13;\nhealth-giving brittle manifest designs example manifoldly <br \/>&#13;\nreckoned Well toward hope curiosities laying flights tolerating <br \/>&#13;\ncelestial rovings derides vengeance conversing unabiding <br \/>&#13;\nnotices idleness arms fifty Look corrupting echoes moulded <br \/>&#13;\nsuspect manifested proper neglecting testified succession <br \/>&#13;\nXI sons alone assent whithersoever mercies parted foreshower <br \/>&#13;\nmanufactures churches table Since drunken weak fowls providedst <br \/>&#13;\nlawfully Onesiphorus sixth philosophy magnitudes whatsoever <br \/>&#13;\nsoundness disposed steal unjust lilies vanquished furnace <br \/>&#13;\ntaughtest mget walking effort transferred A skill prescribes <br \/>&#13;\ndeclared two-edged grammar-learning carelessly baseness <br \/>&#13;\nFaithful Comforter intercourse interposing fulfilled transformed <br \/>&#13;\ntoiling weakened mysteriously works patiently watchings <br \/>&#13;\nconferring cast MS around prepares assistance fables vowing \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539453\" target=\"_self\">4:46 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I did 800x600 4bit.  I got a new machine didn't have that mode.  640x480 was similar.  God said to just do 640x480.  It turns-out that kept me from getting side tracked all these years.  Also, you need a GPU for higher modes or you will have a VW Bug engine in a semi.  (Too little horsepower.)  I cannot do all the different drivers.<p>God's fabvorite car is a Beemer.<\/p><p>26 <br \/>&#13;\nToward the faithful you are faithful; to the honest you are honest; <br \/>&#13;\n27 <br \/>&#13;\nToward the sincere, sincere; but to the perverse you are devious. <br \/>&#13;\n28 <br \/>&#13;\nHumble people you save; haughty eyes you bring low. <br \/>&#13;\n29 <\/p><p>Read <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=1%20kings%2022&amp;version=NIV\">1 kings 22<\/a>.<br \/>&#13;\nAnd read <a href=\"http:\/\/www.biblegateway.com\/passage\/?search=numbers%2011&amp;version=NIV\">Numbers 11<\/a>.  They're hilarious.  Of course God can lie :-)<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539461\" target=\"_self\">4:51 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I saw <a href=\"http:\/\/www.techfounder.net\/2012\/08\/30\/developers-vs-bigcorp\/\">this<\/a>.  Got a Martin Luther.  ROFLMAO.  <p>Why not meditate on the wisdom of God's 640x480 rule... that might not be law.<\/p><p>LoseThos is for recreational programming -- programmers making games by themselves without artists.  It is oddly satisfying with big pixels because it looks like shit for everybody and you can come closer to ideal for that environment.  For programmers only is a strange niche -- tiny.  I did tons of pointless programs in high school, for my own thrills.  <\/p><p>Photorealism is graphic and panders to base nature of humans.  640x480 is innocent.  How many of you are horrified by modern games, longing for a more innocent time?<\/p><p>I don't know if this happens but many crappy books have really nice covers.  Game play is cool.<\/p><p>Big games are too risky for companies to fail, so little risk is done... not that a sane 3rd party would release a game with crappy graphics.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539474\" target=\"_self\">5:01 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539474\" title=\"7 users marked this as favorite\">7 favorites<\/a>] <\/p><\/div><div class=\"comments\">God quoted this one time on sects.<p>have been very great. The interested and active zeal of religious<br \/>&#13;\nteachers can be dangerous and troublesome only where there is either but<br \/>&#13;\none sect tolerated in the society, or where the whole of a large society<br \/>&#13;\nis divided into two or three great sects; the teachers of each acting by<br \/>&#13;\nconcert, and under a regular discipline and subordination. But that zeal<br \/>&#13;\nmust be altogether innocent, where the society is divided into two or<br \/>&#13;\nthree hundred, or, perhaps, into as many thousand small sects, of which<br \/>&#13;\nno one could be considerable enough to disturb the public tranquillity.<br \/>&#13;\nThe teachers of each sect, seeing themselves surrounded on all sides<br \/>&#13;\nwith more adversaries than friends, would be obliged to learn that<br \/>&#13;\ncandour and moderation which are so seldom to be found among the<br \/>&#13;\nteachers of those great sects, whose tenets, being supported by the<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539477\" target=\"_self\">5:05 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I saw <a href=\"http:\/\/runtime-era.blogspot.com\/2012\/08\/simple-password-encrypted-files-w-gnupg.html\">this.<\/a>  LoseThos has postfix typecasting because it was way cleaner to implement in my compiler.  I have no idea what all the C++ casts do, but I know what mine does.<p>If you're used to the freedom of assembly language you might find yourself doing typecasts.  I don't respect any clown would would be so bold as to pass judgement on me.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539483\" target=\"_self\">5:09 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><a href=\"http:\/\/www.youtube.com\/watch?v=g8LBnZOjHqA&amp;feature=related\">These birds are awesome!<\/a>  I said God must be proud of birds.  Most colorful things, can talk, can sing, can fly, penguins, ostriches...<p>God said formations.  Awe yeah, huh!<\/p><p>I did a nice poem in LoseThos.  <a href=\"http:\/\/en.wikisource.org\/wiki\/To_A_Waterfowl\">To a waterfowl.<\/a><br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539488\" target=\"_self\">5:16 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539488\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">God says, \"upright tell slaves necessaries tolerated unveiled gladly Lord's rioting contented murdering trustees existence mistresses Thou crudities nowhere province consigned writings <br \/>&#13;\nforgetteth ye dull derision fixed \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539504\" target=\"_self\">5:26 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">\"To the perverse He is devious\"  ROFLMAO<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539505\" target=\"_self\">5:27 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/forum.osdev.org\/viewtopic.php?f=1&amp;t=25729\">This<\/a> makes me laugh.  everybody starts-out full of pride in PC OSDeving.  Soon, you'll be humbly making your best guesses on things.  Just don't be a fool and mock wiser people.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539512\" target=\"_self\">5:33 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">Another luddite chiming in to say that this has been a most fascinating thread. <p>Thank you losethos for your OS and also your participation here. Much of what you say is beyond my ken, but I still like it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/7721\" target=\"_self\">carsonb<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539515\" target=\"_self\">5:34 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">You'll be asking about PCI interrupts, next, I'll bet.  ROFLMAO  Nobody figures that out, properly.  Maybe on their own machine.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539517\" target=\"_self\">5:34 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">It's easy to do PCI interrupts impropperly, but good enough so you fool yourself, ken.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539536\" target=\"_self\">5:45 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539536\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\">'kay so I have HD Audio.  On my machine it's either A,B,C,or D.  Fuck it, huh?  ROFLMAO  That's about your speed.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539539\" target=\"_self\">5:48 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">yeah, so you gotta figure-out what bus function and device it's located at.  PCIBIOS reports device classes but they're not very standardize, excpet there aren't that many possible ICH chips.  VMWare had to be a prick and not be like an ICH chip.  So, suppose you figure-out bus-dev-fun.  How do you know what pin and what's sharing with that IRQ pin?  I give-up.  I poll HD AUdio.  It doesn't figure-out bus-dev-fun either.  It doesn't work in VMWare when you tell it bus-dev0fun.  You can do PCIRep; in losethos and get a report that is nice.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539560\" target=\"_self\">6:05 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\">I can get God to talk with the MHz stop-watch-random-number technique, or I can go to a quantum random number site that would be proven if you watched my port.<p>God says, \"holy burnt throughout expends purposed reside drive drive \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539570\" target=\"_self\">6:13 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">God says line: 79019<p>said, 22:2 The kingdom of heaven is like unto a certain king, which<br \/>&#13;\nmade a marriage for his son, 22:3 And sent forth his servants to call<br \/>&#13;\nthem that were bidden to the wedding: and they would not come.<\/p><p>22:4 Again, he sent forth other servants, saying, Tell them which are<br \/>&#13;\nbidden, Behold, I have prepared my dinner: my oxen and my fatlings are<br \/>&#13;\nkilled, and all things are ready: come unto the marriage.<\/p><p>22:5 But they made light of it, and went their ways, one to his farm,<br \/>&#13;\nanother to his merchandise: 22:6 And the remnant took his servants,<br \/>&#13;\nand entreated them spitefully, and slew them.<\/p><p>22:7 But when the king heard thereof, he was wroth: and he sent forth<br \/>&#13;\nhis armies, and destroyed those murderers, and burned up their city.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539574\" target=\"_self\">6:14 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Hooya!<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539575\" target=\"_self\">6:15 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><a href=\"http:\/\/ca.news.yahoo.com\/rabbi-warns-dutch-populist-wilders-over-ritual-slaughter-205244682.html\">This<\/a> is about Jewish animal sacrifices.  God said it was \"primitive\".  That's a slippery-slope.  It's kinda strange logic, but I, too, can be easily confused.  Catholicism is strange.  I used to be a contemptuous atheist, but that was wrong.<p>This doesn't give confidence in leaders -- hides stuff?  And Jesus approves?<\/p><p>25 At that time Jesus said, \u201cI praise you, Father, Lord of heaven and earth, because you have hidden these things from the wise and learned, and revealed them to little children. 26 Yes, Father, for this is what you were pleased to do.<\/p><p>27 \u201cAll things have been committed to me by my Father. No one knows the Son except the Father, and no one knows the Father except the Son and those to whom the Son chooses to reveal him.<\/p><p>28 \u201cCome to me, all you who are weary and burdened, and I will give you rest. 29 Take my yoke upon you and learn from me, for I am gentle and humble in heart, and you will find rest for your souls. 30 For my yoke is easy and my burden is light.\u201d<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539588\" target=\"_self\">6:27 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">Watching the video, the ability to put in non-text elements into the code (graphics), and have the compiler recognize them and incorporate them into the program, is remarkable. It really upends a lot of assumptions on how higher level languages and development environments should behave. <p>I can see immediate applications in getting rid of interface builders as a separate system from the code editor (not really a text editor anymore, is it?), but I wonder if it could be taken further,  if you can toss in data sets entire or network maps, and manipulate their contents on the fly programmatically... truly modeless computing.<\/p><p>Did you come up with that on your own as part of LoseThos, or did you re-implement it from another project\/language\/OS somewhere?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/50104\" target=\"_self\">Slap*Happy<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539603\" target=\"_self\">6:38 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539603\" title=\"2 users marked this as favorite\">2 favorites<\/a>] <\/p><\/div><div class=\"comments\">So God said His favorite car was a beemer.  What other honest answer could He give?<p>What would Jesus drive?  Jesus might have had expensive cloths, but his ride was a borrowed donkey.<\/p><p>God says, \"surmount Suffer weariness shower instillest 'it recite createth \"<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539606\" target=\"_self\">6:41 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539606\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\">DoItOnce eh? I get it. I like it.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/107644\" target=\"_self\">anarch<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539639\" target=\"_self\">7:11 PM<\/a>  on August 29, 2012<\/div><div class=\"comments\"><em>Photorealism is graphic and panders to base nature of humans. 640x480 is innocent. How many of you are horrified by modern games, longing for a more innocent time?<\/em><p>I agree! However, I do not want a return to the past, which had its own problems. I think that the idea of using graphics to communicate meaning rather than as their own is coming back, though. Are there any games you like, <strong>losethos<\/strong>, new or old?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/44769\" target=\"_self\">ignignokt<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539678\" target=\"_self\">7:38 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\"><i>Note for those who want to use VirtualBox (and I imagine any other VM), for some reason it's telling me I need a 64 bit processor, even though I have a Core2Duo. I imagine it's due to my XP being 32 bit? Would I need a 64 bit host OS?<\/i><p>losethos is telling me my core2duo sl9400 isn't 64-bit... in virtualbox... in linux (agnostic). oh well.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/102698\" target=\"_self\">ennui.bz<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539738\" target=\"_self\">8:25 PM<\/a>  on August 29, 2012 <\/p><\/div><div class=\"comments\">I'd imagine to run a 64-bit virtualbox instance, not only does your processor have to be 64bit, but it would have to support hardware virtualization (VT-x or AMD-V).<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/156346\" target=\"_self\">dereferenced<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539818\" target=\"_self\">9:15 PM<\/a>  on August 29, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539818\" title=\"1 user marked this as favorite\">1 favorite<\/a>]<\/div><div class=\"comments\"><em>The name seems like a horrible mistake. I was thinking \"Windows\" and \"LoseThos\" as in \"Lose those dorky burgenstocks and the dorky bike helmet.\" Have you ever seen a soldier in armor and kinda chuckled? I particularly remember the movie, Platoon, where Elias says to \"Shitcan\" some standard issue equipment.It was a mistake.<\/em><p>Given I feel this might've slipped by some people (though of course I cannot be sure):<\/p><p>Windows = Winthos<\/p><p>and thus, the opposite, LoseThos, which can also work as a statement for losing all the extraneous material.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/125761\" target=\"_self\">solarion<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4539963\" target=\"_self\">3:18 AM<\/a>  on August 30, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4539963\" title=\"1 user marked this as favorite\">1 favorite<\/a>] <\/p><\/div><div class=\"comments\"><i>I'd imagine to run a 64-bit virtualbox instance, not only does your processor have to be 64bit, but it would have to support hardware virtualization (VT-x or AMD-V).<\/i><p>the sl9400 supports vt-x, and i'm running a x86_64 kernel so... something with virtualbox?  ohwell x 2.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/102698\" target=\"_self\">ennui.bz<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4540117\" target=\"_self\">7:02 AM<\/a>  on August 30, 2012 <\/p><\/div><div class=\"comments\">Terry, I'm pretty clueless when it comes to programming, so I can't comment much on LoseThos.  But some things you've said here...<p><em>God can talk to you, if you hold-up your end of the conversation. Download TSGod. You can just randomly crack-open books. LoseThos will make melodies from God... like muses of old.<\/em><\/p><p>This is just...beautiful.  Holding up one's own end of a conversation can be difficult no matter who one is talking to.  I love that you've created something to help you do that.<\/p><p><em>You have to take God with a grain of salt.<\/em><\/p><p>We all have friends like that, don't we?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/128720\" target=\"_self\">MissySedai<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4540126\" target=\"_self\">7:04 AM<\/a>  on August 30, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4540126\" title=\"4 users marked this as favorite\">4 favorites<\/a>] <\/p><\/div><div class=\"comments\"><i>the sl9400 supports vt-x, and i'm running a x86_64 kernel so... something with virtualbox? ohwell x 2.<\/i><p>I had this same problem trying to run it under VirtualBox on my Windows machine last night.  Mac with Intel i5 was no problem, but on Windows 7 (x64) with Athlon 64 X2 it complained about not being 64-bit.  Hardware virtualization support is probably the culprit.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/17489\" target=\"_self\">neckro23<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4540474\" target=\"_self\">9:46 AM<\/a>  on August 30, 2012 <\/p><\/div><div class=\"comments\">Regarding VT-x: you might want to check your BIOS settings. I've dealt with a few computers that had it as an option in there, and for some reason it tends to default to off.<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/154692\" target=\"_self\">reprise the theme song and roll the credits<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4541128\" target=\"_self\">2:24 PM<\/a>  on August 30, 2012<\/div><div class=\"comments\">I wish I could delete the \"Martin Luther\" comment, earlier.  I don't know what possessed me, maybe God.  My thinking was along the lines of \"Eh? A critic?\", kinda in a funny way.  Do I sound Catholic?  The best I can explain is that I graduated from Catholicism and from atheism.  Now, I talk with God!  Put it this way -- In Heaven, how long before you never talk about Jesus or the Bible?  Think about it.  Fanticise about being in Heaven.  Before long, Jesus and the Bible are ancient history.  Right?<br \/>posted by <a href=\"https:\/\/www.metafilter.com\/user\/162429\" target=\"_self\">losethos<\/a> at <a href=\"https:\/\/www.metafilter.com\/119424\/An-Operating-System-for-Songs-from-God#4541843\" target=\"_self\">8:36 PM<\/a>  on August 30, 2012 [<a href=\"https:\/\/www.metafilter.com\/favorited\/2\/4541843\" title=\"5 users marked this as favorite\">5 favorites<\/a>]<\/div><div class=\"comments\">&#13;\n\t\t\t\t&#13;\n\t\t\t\t\t&#13;\n\t\t\t\t\t&#13;\n\t\t\t\t\t<ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-8621957794194569\" data-ad-slot=\"2376191428\" data-ad-format=\"auto\">&#13;\n\t\t\t\t\t&#13;\n\t\t\t\t\t&#13;\n&#13;<\/ins><\/div>","created_at":"2018-11-07T16:54:17+0000","updated_at":"2018-11-07T16:54:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":77,"domain_name":"www.metafilter.com","preview_picture":"http:\/\/cdn.mefi.us\/images\/mefi\/apple-touch-icon.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":634,"title":"Persisting the changes of range objects after selection in HTML","url":"https:\/\/stackoverflow.com\/questions\/13949059\/persisting-the-changes-of-range-objects-after-selection-in-html","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        15&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>Is there a way to save the changes like changing the background of HTML text that span over multiple tags so that when it is loaded again the changes made should be reflected in the HTML page.<\/p>\n<p>EDIT: Detailed explanation.<\/p>\n<p>When the HTML page is loaded, the text is selected and highlighted using the range object and the executeCommand:<\/p>\n<pre>             document.execCommand(\"BackColor\", false, 'yellow');\n<\/pre>\n<p>The changes (highlighting the text as yellow) remain until the page is reloaded. But when the page is reloaded these changes are not there. What i want is to save somehow these changes like in local DB sqlite so that when page is reloaded\/refreshed the changes in HTML page should appear.<\/p>\n<p>Any idea how to do it. Do i need to save its range start offset and end offset which can be used to create range next time the page is loaded. Please give your insights.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/13949059\/revisions\" title=\"show all edits to this post\">edited Dec 19 '12 at 10:27<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Dec 19 '12 at 9:21&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/252929\/satish\">Satish<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            6303921&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"3\">&#13;\n                                3 Answers&#13;\n                                3&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        55&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>For each selection, you could serialize the selected range to character offsets and deserialize it again on reload using something like this:<\/p>\n<p>Demo: <a href=\"http:\/\/jsfiddle.net\/WeWy7\/3\/\">http:\/\/jsfiddle.net\/WeWy7\/3\/<\/a><\/p>\n<p>Code:<\/p>\n<pre>var saveSelection, restoreSelection;\nif (window.getSelection &amp;&amp; document.createRange) {\n    saveSelection = function(containerEl) {\n        var range = window.getSelection().getRangeAt(0);\n        var preSelectionRange = range.cloneRange();\n        preSelectionRange.selectNodeContents(containerEl);\n        preSelectionRange.setEnd(range.startContainer, range.startOffset);\n        var start = preSelectionRange.toString().length;\n        return {\n            start: start,\n            end: start + range.toString().length\n        };\n    };\n    restoreSelection = function(containerEl, savedSel) {\n        var charIndex = 0, range = document.createRange();\n        range.setStart(containerEl, 0);\n        range.collapse(true);\n        var nodeStack = [containerEl], node, foundStart = false, stop = false;\n        while (!stop &amp;&amp; (node = nodeStack.pop())) {\n            if (node.nodeType == 3) {\n                var nextCharIndex = charIndex + node.length;\n                if (!foundStart &amp;&amp; savedSel.start &gt;= charIndex &amp;&amp; savedSel.start &lt;= nextCharIndex) {\n                    range.setStart(node, savedSel.start - charIndex);\n                    foundStart = true;\n                }\n                if (foundStart &amp;&amp; savedSel.end &gt;= charIndex &amp;&amp; savedSel.end &lt;= nextCharIndex) {\n                    range.setEnd(node, savedSel.end - charIndex);\n                    stop = true;\n                }\n                charIndex = nextCharIndex;\n            } else {\n                var i = node.childNodes.length;\n                while (i--) {\n                    nodeStack.push(node.childNodes[i]);\n                }\n            }\n        }\n        var sel = window.getSelection();\n        sel.removeAllRanges();\n        sel.addRange(range);\n    }\n} else if (document.selection) {\n    saveSelection = function(containerEl) {\n        var selectedTextRange = document.selection.createRange();\n        var preSelectionTextRange = document.body.createTextRange();\n        preSelectionTextRange.moveToElementText(containerEl);\n        preSelectionTextRange.setEndPoint(\"EndToStart\", selectedTextRange);\n        var start = preSelectionTextRange.text.length;\n        return {\n            start: start,\n            end: start + selectedTextRange.text.length\n        }\n    };\n    restoreSelection = function(containerEl, savedSel) {\n        var textRange = document.body.createTextRange();\n        textRange.moveToElementText(containerEl);\n        textRange.collapse(true);\n        textRange.moveEnd(\"character\", savedSel.end);\n        textRange.moveStart(\"character\", savedSel.start);\n        textRange.select();\n    };\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/13950376\/revisions\" title=\"show all edits to this post\">edited Jan 9 '14 at 23:44<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Dec 19 '12 at 10:33&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/96100\/tim-down\">Tim Down<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            245k56371458&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Without knowing more about the context, it is hard to give an exact answer, but yes it would be possible, but it will be quite complex for most cases. Depending on the usecase, there are a few ways to go.<\/p>\n<p><strong>Cookies or Local storage<\/strong><\/p>\n<p>You could use some sort of client-side storage (cookies, local storage or similar) and save information about what elements were modified and how. Whenever the page is reloaded you read that storage and apply the changes. How to implement it will depend on how those changes are made, and will be to0 extensive to cover in a single SO-answer I'm afraid.<\/p>\n<p><strong>Server-side storage<\/strong><\/p>\n<p>If you know who each user is (you have some form of authentication), whenever they change the appearance of something (however that is made), you make an ajax-request to the server and save those changes to a database. On every subsequent page load, you would then have to check what use is making the request, do a lookup in your database to see if they've made any changes, and in that case apply them accordingly.<\/p>\n<p>Common for both the client- and server-side storage solutions is that they will be quite extensive to implement I believe.<\/p>\n<p><strong>Browser plugin<\/strong><\/p>\n<p>Another way to go would be to make use of plugins like <a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/greasemonkey\/\" rel=\"nofollow\">Greasemonkey for Firefox<\/a> that allow the user to customize the way a webpage is rendered. Those customizations will be persistent across page loads.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Dec 19 '12 at 9:39&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/678801\/christofer-eliasson\">Christofer Eliasson<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            23.8k55890&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Using character offsets doesn't work if the cursor is at the beginning of a new paragraph. The approach below walks the DOM node and counts all nodes towards the offset. It also handles start and end individually to make sure that the selection remembers its exact position. Here is an updated version that I use in a major project (see functions at end):<\/p>\n<pre>\/*\n Gets the offset of a node within another node. Text nodes are\n counted a n where n is the length. Entering (or passing) an\n element is one offset. Exiting is 0.\n *\/\nvar getNodeOffset = function(start, dest) {\n  var offset = 0;\n  var node = start;\n  var stack = [];\n  while (true) {\n    if (node === dest) {\n      return offset;\n    }\n    \/\/ Go into children\n    if (node.firstChild) {\n      \/\/ Going into first one doesn't count\n      if (node !== start)\n        offset += 1;\n      stack.push(node);\n      node = node.firstChild;\n    }\n    \/\/ If can go to next sibling\n    else if (stack.length &gt; 0 &amp;&amp; node.nextSibling) {\n      \/\/ If text, count length (plus 1)\n      if (node.nodeType === 3)\n        offset += node.nodeValue.length + 1;\n      else\n        offset += 1;\n      node = node.nextSibling;\n    }\n    else {\n      \/\/ If text, count length\n      if (node.nodeType === 3)\n        offset += node.nodeValue.length + 1;\n      else\n        offset += 1;\n      \/\/ No children or siblings, move up stack\n      while (true) {\n        if (stack.length &lt;= 1)\n          return offset;\n        var next = stack.pop();\n        \/\/ Go to sibling\n        if (next.nextSibling) {\n          node = next.nextSibling;\n          break;\n        }\n      }\n    }\n  }\n};\n\/\/ Calculate the total offsets of a node\nvar calculateNodeOffset = function(node) {\n  var offset = 0;\n  \/\/ If text, count length\n  if (node.nodeType === 3)\n    offset += node.nodeValue.length + 1;\n  else\n    offset += 1;\n  if (node.childNodes) {\n    for (var i=0;i&lt;node.childNodes.length;i++) {\n      offset += calculateNodeOffset(node.childNodes[i]);\n    }\n  }\n  return offset;\n};\n\/\/ Determine total offset length from returned offset from ranges\nvar totalOffsets = function(parentNode, offset) {\n  if (parentNode.nodeType == 3)\n    return offset;\n  if (parentNode.nodeType == 1) {\n    var total = 0;\n    \/\/ Get child nodes\n    for (var i=0;i&lt;offset;i++) {\n      total += calculateNodeOffset(parentNode.childNodes[i]);\n    }\n    return total;\n  }\n  return 0;\n};\nvar getNodeAndOffsetAt = function(start, offset) {\n  var node = start;\n  var stack = [];\n  while (true) {\n    \/\/ If arrived\n    if (offset &lt;= 0)\n      return { node: node, offset: 0 };\n    \/\/ If will be within current text node\n    if (node.nodeType == 3 &amp;&amp; (offset &lt;= node.nodeValue.length))\n      return { node: node, offset: Math.min(offset, node.nodeValue.length) };\n    \/\/ Go into children (first one doesn't count)\n    if (node.firstChild) {\n      if (node !== start)\n        offset -= 1;\n      stack.push(node);\n      node = node.firstChild;\n    }\n    \/\/ If can go to next sibling\n    else if (stack.length &gt; 0 &amp;&amp; node.nextSibling) {\n      \/\/ If text, count length\n      if (node.nodeType === 3)\n        offset -= node.nodeValue.length + 1;\n      else\n        offset -= 1;\n      node = node.nextSibling;\n    }\n    else {\n      \/\/ No children or siblings, move up stack\n      while (true) {\n        if (stack.length &lt;= 1) {\n          \/\/ No more options, use current node\n          if (node.nodeType == 3)\n            return { node: node, offset: Math.min(offset, node.nodeValue.length) };\n          else\n            return { node: node, offset: 0 };\n        }\n        var next = stack.pop();\n        \/\/ Go to sibling\n        if (next.nextSibling) {\n          \/\/ If text, count length\n          if (node.nodeType === 3)\n            offset -= node.nodeValue.length + 1;\n          else\n            offset -= 1;\n          node = next.nextSibling;\n          break;\n        }\n      }\n    }\n  }\n};\nexports.save = function(containerEl) {\n  \/\/ Get range\n  var selection = window.getSelection();\n  if (selection.rangeCount &gt; 0) {\n    var range = selection.getRangeAt(0);\n    return {\n      start: getNodeOffset(containerEl, range.startContainer) + totalOffsets(range.startContainer, range.startOffset),\n      end: getNodeOffset(containerEl, range.endContainer) + totalOffsets(range.endContainer, range.endOffset)\n    };\n  }\n  else\n    return null;\n};\nexports.restore = function(containerEl, savedSel) {\n  if (!savedSel)\n    return;\n  var range = document.createRange();\n  var startNodeOffset, endNodeOffset;\n  startNodeOffset = getNodeAndOffsetAt(containerEl, savedSel.start);\n  endNodeOffset = getNodeAndOffsetAt(containerEl, savedSel.end);\n  range.setStart(startNodeOffset.node, startNodeOffset.offset);\n  range.setEnd(endNodeOffset.node, endNodeOffset.offset);\n  var sel = window.getSelection();\n  sel.removeAllRanges();\n  sel.addRange(range);\n};\n<\/pre>\n<p>This only works on modern browsers (IE 9+ at least).<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/38752483\/revisions\" title=\"show all edits to this post\">edited Jan 3 '17 at 11:04<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/96100\/tim-down\">Tim Down<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            245k56371458&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 3 '16 at 19:53&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2162273\/clayton-grassick\">Clayton Grassick<\/a>&#13;\n        &#13;<\/div>","created_at":"2018-11-07T16:54:15+0000","updated_at":"2018-11-07T16:54:15+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":632,"title":"monome","url":"https:\/\/monome.org\/","content":"<iframe src=\"https:\/\/player.vimeo.com\/video\/266741634?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>changes. travels. is open to possibilities.<\/p><p>norns is many sound instruments. it connects to grids, MIDI and other objects. norns lets you define its behavior with scripts and DSP.<\/p><p>past. present. future. norns is many tomorrows.<\/p><ul class=\"detail-list\"><li>linux-powered open-source sound computer<\/li>\n<li>2 audio in, 2 audio out, headphone<\/li>\n<li>4 USB device ports: grids, MIDI, WIFI, etc<\/li>\n<li>bright white pixels, endless rotary encoders, key switches<\/li>\n<li>milled aluminum assembly, hard anodized<\/li>\n<li>internal battery, charge by USB<\/li>\n<li>6.2\" x 4.0\" x 0.8\"<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/128746859?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>an open, interactive instrument that allows the user to imagine and define its function.<\/p><p>tactile and visual feedback. decoupled by design. clear response and intuitive play. when you push a key the input data is sent to the computer (be it norns, ansible, or normal laptop). the computer considers this, then perhaps sends back some light data, or changes a sound. use is determined by the chosen application running on the computer: sequencer, sample cutter, tone map, polyrhythm machine are just a few of the many possibilities, and we\u2019ve put together studies to help you create your own.<\/p><ul class=\"detail-list\"><li>16 x 8 backlit keys<\/li>\n<li>bright warm white light<\/li>\n<li>two-tone aluminum assembly<\/li>\n<li>usb powered<\/li>\n<li>10.2\" x 5.4\" x 0.6\"<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/182119406?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>rotational controls encircled by dense segmented lighting, completely open for definition.<\/p><p>like the grid, tactile and visual feedback. but instead of a multitude of binary points, the arc has four incredibly precise continuous controls with visual indication to match. connects to computers (norns, ansible, laptops) to become a sequencer, physics simulation, granular instrument, or anything you can create yourself.<\/p><ul class=\"detail-list\"><li>four high resolution optical encoders<\/li>\n<li>bright warm white light<\/li>\n<li>two-tone aluminum assembly<\/li>\n<li>usb powered<\/li>\n<li>10.2\" x 2.7\" x 0.6\"<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/129271731?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>algorithmic ecosystem. teletype is a dynamic, musical event triggering platform. edit scripts with a USB keyboard which plugs into the front panel. syntax is simple and easy to learn with the provided tutorials, video, and reference pages. teletype runs without the keyboard attached: you might treat editing as precomposition, or leave the keyboard plugged in for live coding.<\/p><ul class=\"detail-list\"><li>format: eurorack<\/li>\n<li>width: 18hp<\/li>\n<li>depth: 26mm (\u201cskiff friendly\u201d)<\/li>\n<li>power: 12v: 72ma, -12v: 12ma, 5v: none<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/214221486?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>far communicator, speaks openly.<\/p><p>ansible connect various USB devices (grid, arc, MIDI) to the modular environment. several modes provide different sequencers and control approaches.<\/p><ul class=\"detail-list\"><li>format: eurorack<\/li>\n<li>width: 6hp<\/li>\n<li>depth: 39mm (\u201cskiff friendly\u201d)<\/li>\n<li>power: 12v: 200ma, -12v: 19ma, 5v: none<\/li>\n<\/ul><\/div>","created_at":"2018-11-07T16:53:56+0000","updated_at":"2018-11-07T16:53:56+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"monome.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":633,"title":"How to Conduct a SWOT Analysis for Your Small Business","url":"https:\/\/www.thebalancesmb.com\/swot-analysis-for-small-business-2951706","content":"<div id=\"article-meta_1-0\" class=\"comp article-meta mntl-block\" data-tracking-container=\"true\"><p>By\n<a href=\"https:\/\/www.thebalancesmb.com\/alyssa-gregory-2951140\" class=\"byline__link\">Alyssa Gregory<\/a><\/p><p>Updated August 13, 2018<\/p><\/div><div id=\"mntl-sc-page_1-0\" class=\"comp mntl-sc-page mntl-block structured-content expert-content\" data-sc-sticky-offset=\"80\" data-sc-ad-label-height=\"20\" data-sc-offset=\"200\" data-sc-ad-track-spacing=\"100\" data-sc-min-track-height=\"250\" data-sc-max-track-height=\"600\" data-sc-breakpoint=\"54em\" data-sc-load-immediate=\"4\" data-sc-content-positions=\"[300, 1, 1, 1250, 1, 1, 1, 1]\" data-bind-scroll-on-start=\"true\"><div id=\"mntl-sc-block_1-0\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>A SWOT analysis is a strategic planning tool that helps a business owner identify his or her own strengths and weaknesses, as well as any opportunities and threats that may exist in a specific business situation. A SWOT analysis is most commonly used as part of a <a href=\"https:\/\/www.thebalancesmb.com\/how-to-write-a-small-business-marketing-plan-2951749\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">marketing plan<\/a>, but it is also a good tool for general business strategizing, and to use as a starting point for team discussions.<\/p><\/div><div id=\"mntl-sc-block_1-0-2\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>When conducted thoroughly, a SWOT analysis can uncover a wealth of information and can be useful in a number of situations. This article will walk you through how to conduct a SWOT analysis and provide some tips that will help you use the tool effectively.\u200b<\/p><\/div><figure id=\"mntl-sc-block_1-0-4\" class=\"comp mntl-sc-block mntl-sc-block-image figure-landscape figure-high-res\"><div class=\"figure-media\"><div class=\"img-placeholder\"><img data-srcset=\"https:\/\/www.thebalancesmb.com\/thmb\/wcD-gb5aTDdavO7-GDQNxHxmIM8=\/300x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 300w, https:\/\/www.thebalancesmb.com\/thmb\/AQ7xLxpV_jEIVMCpxtJXt4mWH0Y=\/600x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 600w, https:\/\/www.thebalancesmb.com\/thmb\/WkYFlVQHF9TWe-HamGIqDzJE52g=\/900x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 900w, https:\/\/www.thebalancesmb.com\/thmb\/qy3XebJYepEM1t_rYP_WNrbW9nQ=\/1500x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 1500w\" src=\"https:\/\/www.thebalancesmb.com\/thmb\/9CAr1q093sjzzJZ3PtbG82fekiE=\/1500x1000\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png\" class=\"lazyload\" data-click-tracked=\"true\" data-img-lightbox=\"true\" data-expand=\"300\" id=\"mntl-sc-block-image_1-0-4\" data-tracking-container=\"true\" alt=\"image\" \/><\/div><\/div>\n<figcaption id=\"mntl-figure-caption_1-0\" class=\"comp mntl-figure-caption figure-article-caption\">\n\u00a9\u00a0The Balance\n<\/figcaption><\/figure><div id=\"mntl-sc-block_1-0-6\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>A SWOT matrix is usually depicted as a square divided into four quadrants. Each quadrant represents one element of the SWOT analysis -- Strengths, Weaknesses, Opportunities, Threats.<\/p><\/div><div id=\"mntl-sc-block_1-0-9\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>The easiest way to start filling in each quadrant in the matrix above is by answering a series of questions. Use the list below to get started, focusing on the questions that are most relevant to your business and current situation.<\/p><\/div><div id=\"mntl-sc-block_1-0-11\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Strengths:<\/strong> For this quadrant, think about the attributes of yourself and your business that will help you achieve your objective. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-13\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>What do you do well?<\/li><li>What are your unique skills?<\/li><li>What expert or specialized knowledge do you have?<\/li><li>What experience do you have?<\/li><li>What do you do better than your competitors?<\/li><li>Where are you most profitable in your business?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-15\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Weaknesses:<\/strong> For this quadrant, think about the attributes of yourself and your business that could hurt your progress in achieving your objective. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-17\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>In what areas do you need to improve?<\/li><li>What resources do you lack?<\/li><li>What parts of your business are not very profitable?<\/li><li>Where do you need further education and\/or experience?<\/li><li>What costs you time and\/or money?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-19\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Opportunities:<\/strong> For this quadrant, think about the external conditions that will help you achieve your objective. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-21\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>What are the <a href=\"https:\/\/www.thebalancesmb.com\/four-common-small-business-goals-to-inspire-you-2951810\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">business goals<\/a> you are currently working towards?<\/li><li>How can you do more with your existing customers or clients?<\/li><li>How can you <a href=\"https:\/\/www.thebalancesmb.com\/low-cost-technology-solutions-2951652\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"2\">use technology<\/a> to enhance your business?<\/li><li>Are there new target audiences you have the potential to reach?<\/li><li>Are there related products and services that provide an opportunity for your business?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-23\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Threats:<\/strong> For this quadrant, think about the external conditions that could damage your business's performance. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-25\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>What obstacles do you face?<\/li><li>What are the strengths of your biggest competitors?<\/li><li>What are your competitors doing that you're not?<\/li><li>What's going on in the economy?<\/li><li>What's going on in the industry?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-28\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>One of the most important parts of your SWOT analysis is using the data you compiled to identify <a href=\"https:\/\/www.thebalancesmb.com\/swot-example-strengths-weaknesses-opportunities-threats-2947985\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">new strategies<\/a> and goals for your business. For example, you can:<\/p><\/div><div id=\"mntl-sc-block_1-0-30\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>Create a plan to build up your strengths even more<\/li><li>List ways you can work on building up your weaknesses<\/li><li>Set <a href=\"https:\/\/www.thebalancesmb.com\/smart-goal-setting-101-2951829\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">SMART goals<\/a> for each of the opportunities you identified<\/li><li>Devise a plan to use your strengths to decrease the threats you identified<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-32\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>Then, look for ways to combine data from different quadrants in even more ways:<\/p><\/div><div id=\"mntl-sc-block_1-0-34\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>Explore how you can combine your strengths and opportunities to develop new strategies<\/li><li>Try combining strengths and threats to identify threats you can eliminate<\/li><li>Look at your weaknesses and opportunities to create a list of areas ready for improvement<\/li><li>Make a list of areas to avoid that fall under weaknesses and threats<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-36\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>Once you understand how to compile your <a href=\"https:\/\/www.thebalancesmb.com\/every-business-needs-a-good-swot-analysis-1794147\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">SWOT data<\/a> and find ways to use it strategically, the SWOT analysis will be a tool that you can use over and over in your business to explore new opportunities and improve your decision-making process.<\/p><\/div><\/div>","created_at":"2018-11-07T16:53:56+0000","updated_at":"2018-11-07T16:53:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"www.thebalancesmb.com","preview_picture":"https:\/\/www.thebalancesmb.com\/thmb\/wKFp9D0ZGgVDUJYEo-2UF9wrS6g=\/1080x1080\/filters:fill(auto,1)\/money-business-social-share-default-5ac795acfa6bcc00376a1bdb.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":630,"title":"Understanding HBase and BigTable - DZone","url":"https:\/\/dzone.com\/articles\/understanding-hbase-and-bigtab","content":"<div class=\"content-html\" itemprop=\"text\"><p>The hardest part about learning Hbase (the open source implementation of Google's BigTable), is just wrapping your mind around the concept of what it actually is. <\/p>\n<p>I find it rather unfortunate that these two great systems contain the words <i>table<\/i> and <i>base<\/i> in their names, which tend to cause confusion among RDBMS indoctrinated individuals (like myself). <\/p>\n<p>This article aims to describe these distributed data storage systems from a conceptual standpoint. After reading it, you should be better able to make an educated decision regarding when you might want to use Hbase vs when you'd be better off with a \"traditional\" database. <\/p>\n\n<h2>It's all in the terminology <\/h2>\n<p>Fortunately, <a href=\"http:\/\/labs.google.com\/papers\/bigtable.html\" rel=\"nofollow\" title=\"http:\/\/labs.google.com\/papers\/bigtable.html\" class=\"external text\">Google's BigTable Paper<\/a> clearly explains what BigTable actually is. Here is the first sentence of the \"Data Model\" section: <\/p>\n<blockquote><div>\n A Bigtable is a sparse, distributed, persistent multidimensional sorted map. \n<\/div><\/blockquote>\n<p><i>Note: At this juncture I like to give readers the opportunity to collect any brain matter which may have left their skulls upon reading that last line.<\/i> <\/p>\n<p>The BigTable paper continues, explaining that: <\/p>\n<blockquote><div>\n The map is indexed by a row key, column key, and a timestamp; each value in the map is an uninterpreted array of bytes. \n<\/div><\/blockquote>\n<p>Along those lines, the <a href=\"http:\/\/wiki.apache.org\/hadoop\/Hbase\/HbaseArchitecture\" rel=\"nofollow\" title=\"http:\/\/wiki.apache.org\/hadoop\/Hbase\/HbaseArchitecture\" class=\"external text\">HbaseArchitecture<\/a> page of the Hadoop wiki posits that: <\/p>\n<blockquote><div>\n HBase uses a data model very similar to that of Bigtable. Users store data rows in labelled tables. A data row has a sortable key and an arbitrary number of columns. The table is stored sparsely, so that rows in the same table can have crazily-varying columns, if the user likes. \n<\/div><\/blockquote>\n<p>Although all of that may seem rather cryptic, it makes sense once you break it down a word at a time. I like to discuss them in this sequence: map, persistent, distributed, sorted, multidimensional, and sparse. <\/p>\n<p>Rather than trying to picture a complete system all at once, I find it easier to build up a mental framework piecemeal, to ease into it... <\/p>\n<h2>map <\/h2> \n<p>At its core, Hbase\/BigTable is a <a href=\"http:\/\/en.wikipedia.org\/wiki\/Associative_array\" rel=\"nofollow\" title=\"http:\/\/en.wikipedia.org\/wiki\/Associative_array\" class=\"external text\">map<\/a>. Depending on your programming language background, you may be more familiar with the terms associative array (PHP), dictionary (Python), Hash (Ruby), or Object (JavaScript). <\/p>\n<p>From the wikipedia article, a map is \"an abstract data type composed of a collection of keys and a collection of values, where each key is associated with one value.\" <\/p>\n<p>Using <a href=\"http:\/\/en.wikipedia.org\/wiki\/JSON\" rel=\"nofollow\" title=\"http:\/\/en.wikipedia.org\/wiki\/JSON\" class=\"external text\">JavaScript Object Notation<\/a>, here's an example of a simple map where all the values are just strings: <\/p>\n<pre class=\"js\">{<br \/>\"zzzzz\" : \"woot\",<br \/>\"xyz\" : \"hello\",<br \/>\"aaaab\" : \"world\",<br \/>\"1\" : \"x\",<br \/>\"aaaaa\" : \"y\"<br \/>}<br \/><\/pre>\n<h2>persistent <\/h2>\n<p>Persistence merely means that the data you put in this special map \"persists\" after the program that created or accessed it is finished. This is no different in concept than any other kind of persistent storage such as a file on a filesystem. Moving along... <\/p>\n<h2>distributed <\/h2>\n<p>Hbase and BigTable are built upon distributed filesystems so that the underlying file storage can be spread out among an array of independent machines. <\/p>\n<p>Hbase sits atop either <a href=\"http:\/\/hadoop.apache.org\/core\/docs\/r0.16.4\/hdfs_design.html\" rel=\"nofollow\" title=\"http:\/\/hadoop.apache.org\/core\/docs\/r0.16.4\/hdfs_design.html\" class=\"external text\">Hadoop's Distributed File System<\/a> (HDFS) or Amazon's <a href=\"http:\/\/en.wikipedia.org\/wiki\/Amazon_S3\" rel=\"nofollow\" title=\"http:\/\/en.wikipedia.org\/wiki\/Amazon_S3\" class=\"external text\">Simple Storage Service<\/a> (S3), while a BigTable makes use of the <a href=\"http:\/\/labs.google.com\/papers\/gfs.html\" rel=\"nofollow\" title=\"http:\/\/labs.google.com\/papers\/gfs.html\" class=\"external text\">Google File System<\/a> (GFS). <\/p>\n<p>Data is <a href=\"http:\/\/en.wikipedia.org\/wiki\/Replication_%28computer_science%29\" rel=\"nofollow\" title=\"http:\/\/en.wikipedia.org\/wiki\/Replication_%28computer_science%29\" class=\"external text\">replicated<\/a> across a number of participating nodes in an analogous manner to how data is striped across discs in a <a href=\"http:\/\/en.wikipedia.org\/wiki\/RAID\" rel=\"nofollow\" title=\"http:\/\/en.wikipedia.org\/wiki\/RAID\" class=\"external text\">RAID<\/a> system. <\/p>\n<p>For the purpose of this article, we don't really care which distributed filesystem implementation is being used. The important thing to understand is that it <i>is<\/i> distributed, which provides a layer of protection against, say, a node within the cluster failing. <\/p>\n<h2>sorted <\/h2>\n<p>Unlike most map implementations, in Hbase\/BigTable the key\/value pairs are kept in strict alphabetical order. That is to say that the row for the key \"aaaaa\" should be right next to the row with key \"aaaab\" and very far from the row with key \"zzzzz\". <\/p>\n<p>Continuing our JSON example, the sorted version looks like this: <\/p>\n<pre class=\"javascript\">{<br \/>\"1\" : \"x\",<br \/>\"aaaaa\" : \"y\",<br \/>\"aaaab\" : \"world\",<br \/>\"xyz\" : \"hello\",<br \/>\"zzzzz\" : \"woot\"<br \/>}<br \/><\/pre>\n<p>Because these systems tend to be so huge and distributed, this sorting feature is actually very important. The spacial propinquity of rows with like keys ensures that when you must scan the table, the items of greatest interest to you are near each other. <\/p>\n<p>This is important when choosing a row key convention. For example, consider a table whose keys are domain names. It makes the most sense to list them in reverse notation (so \"com.jimbojw.www\" rather than \"www.jimbojw.com\") so that rows about a subdomain will be near the parent domain row. <\/p>\n<p>Continuing the domain example, the row for the domain \"mail.jimbojw.com\" would be right next to the row for \"www.jimbojw.com\" rather than say \"mail.xyz.com\" which would happen if the keys were regular domain notation. <\/p>\n<p>It's important to note that the term \"sorted\" when applied to Hbase\/BigTable does not mean that \"values\" are sorted. There is no automatic indexing of anything other than the keys, just as it would be in a plain-old map implementation. <\/p>\n<h2>multidimensional <\/h2>\n<p>Up to this point, we haven't mentioned any concept of \"columns\", treating the \"table\" instead as a regular-old hash\/map in concept. This is entirely intentional. The word \"column\" is another loaded word like \"table\" and \"base\" which carries the emotional baggage of years of RDBMS experience. <\/p>\n<p>Instead, I find it easier to think about this like a multidimensional map - a map of maps if you will. Adding one dimension to our running JSON example gives us this: <\/p>\n<pre class=\"javascript\">{<br \/>\"1\" : {<br \/>\"A\" : \"x\",<br \/>\"B\" : \"z\"<br \/>},<br \/>\"aaaaa\" : {<br \/>\"A\" : \"y\",<br \/>\"B\" : \"w\"<br \/>},<br \/>\"aaaab\" : {<br \/>\"A\" : \"world\",<br \/>\"B\" : \"ocean\"<br \/>},<br \/>\"xyz\" : {<br \/>\"A\" : \"hello\",<br \/>\"B\" : \"there\"<br \/>},<br \/>\"zzzzz\" : {<br \/>\"A\" : \"woot\",<br \/>\"B\" : \"1337\"<br \/>}<br \/>}<br \/><\/pre>\n<p>In the above example, you'll notice now that each key points to a map with exactly two keys: \"A\" and \"B\". From here forward, we'll refer to the top-level key\/map pair as a \"row\". Also, in BigTable\/Hbase nomenclature, the \"A\" and \"B\" mappings would be called \"Column Families\". <\/p>\n<p>A table's column families are specified when the table is created, and are difficult or impossible to modify later. It can also be expensive to add new column families, so it's a good idea to specify all the ones you'll need up front. <\/p>\n<p>Fortunately, a column family may have any number of columns, denoted by a column \"qualifier\" or \"label\". Here's a subset of our JSON example again, this time with the column qualifier dimension built in: <\/p>\n<pre class=\"javascript\">{<br \/>\/\/ ...<br \/>\"aaaaa\" : {<br \/>\"A\" : {<br \/>\"foo\" : \"y\",<br \/>\"bar\" : \"d\"<br \/>},<br \/>\"B\" : {<br \/>\"\" : \"w\"<br \/>}<br \/>},<br \/>\"aaaab\" : {<br \/>\"A\" : {<br \/>\"foo\" : \"world\",<br \/>\"bar\" : \"domination\"<br \/>},<br \/>\"B\" : {<br \/>\"\" : \"ocean\"<br \/>}<br \/>},<br \/>\/\/ ...<br \/>}<br \/><\/pre>\n<p>Notice that in the two rows shown, the \"A\" column family has two columns: \"foo\" and \"bar\", and the \"B\" column family has just one column whose qualifier is the empty string (\"\"). <\/p>\n<p>When asking Hbase\/BigTable for data, you must provide the full column name in the form \"<code>&lt;family&gt;:&lt;qualifier&gt;<\/code>\". So for example, both rows in the above example have three columns: \"A:foo\", \"A:bar\" and \"B:\". <\/p>\n<p>Note that although the column families are static, the columns themselves are not. Consider this expanded row: <\/p>\n<pre class=\"js\">{<br \/>\/\/ ...<br \/>\"zzzzz\" : {<br \/>\"A\" : {<br \/>\"catch_phrase\" : \"woot\",<br \/>}<br \/>}<br \/>}<br \/><\/pre>\n<p>In this case, the \"zzzzz\" row has exactly one column, \"A:catch_phrase\". Because each row may have any number of different columns, there's no built-in way to query for a list of all columns in all rows. To get that information, you'd have to do a full table scan. You <i>can<\/i> however query for a list of all column families since these are immutable (more-or-less). <\/p>\n<p>The final dimension represented in Hbase\/BigTable is time. All data is versioned either using an integer timestamp (seconds since the epoch), or another integer of your choice. The client may specify the timestamp when inserting data. <\/p>\n<p>Consider this updated example utilizing arbitrary integral timestamps: <\/p>\n<pre class=\"javascript\">{<br \/>\/\/ ...<br \/>\"aaaaa\" : {<br \/>\"A\" : {<br \/>\"foo\" : {<br \/>15 : \"y\",<br \/>4 : \"m\"<br \/>},<br \/>\"bar\" : {<br \/>15 : \"d\",<br \/>}<br \/>},<br \/>\"B\" : {<br \/>\"\" : {<br \/>6 : \"w\"<br \/>3 : \"o\"<br \/>1 : \"w\"<br \/>}<br \/>}<br \/>},<br \/>\/\/ ...<br \/>}<br \/><\/pre>\n<p>Each column family may have its own rules regarding how many versions of a given cell to keep (a cell is identified by its rowkey\/column pair) In most cases, applications will simply ask for a given cell's data, without specifying a timestamp. In that common case, Hbase\/BigTable will return the most recent version (the one with the highest timestamp) since it stores these in reverse chronological order. <\/p>\n<p>If an application asks for a given row at a given timestamp, Hbase will return cell data where the timestamp is less than or equal to the one provided. <\/p>\n<p>Using our imaginary Hbase table, querying for the row\/column of \"aaaaa\"\/\"A:foo\" will return \"y\" while querying for the row\/column\/timestamp of \"aaaaa\"\/\"A:foo\"\/10 will return \"m\". Querying for a row\/column\/timestamp of \"aaaaa\"\/\"A:foo\"\/2 will return a null result. <\/p>\n<h2>sparse <\/h2>\n<p>The last keyword is sparse. As already mentioned, a given row can have any number of columns in each column family, or none at all. The other type of sparseness is row-based gaps, which merely means that there may be gaps between keys. <\/p>\n<p>This, of course, makes perfect sense if you've been thinking about Hbase\/BigTable in the map-based terms of this article rather than perceived similar concepts in RDBMS's. <\/p>\n<h2>And that's about it <\/h2>\n<p>Well, I hope that helps you understand conceptually what the Hbase data model feels like. <\/p>\n<p>As always, I look forward to your thoughts, comments and suggestions. <\/p><\/div><div class=\"content-html\" itemprop=\"text\"><a>\n                        <img class=\"pub-image\" width=\"420\" itemprop=\"image\" src=\"src\" alt=\"image\" \/><\/a><\/div>","created_at":"2018-11-07T16:53:55+0000","updated_at":"2018-11-07T16:53:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"dzone.com","preview_picture":"https:\/\/dzone.com\/themes\/dz20\/images\/ArticleImg_8.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":631,"title":"Insomnia","url":"https:\/\/insomnia.rest\/","content":"<section class=\"no-margin padding-top-lg padding-bottom-lg\"><div class=\"container center\"><div class=\"row\"><div class=\"col-12\"><h2 class=\"text-xxl\">Powerful HTTP and GraphQL tool belt<\/h2><p class=\"text-lg\"><strong>Free<\/strong> and <strong>open source<\/strong> on Mac, Windows, and Linux<\/p><\/div><\/div><div class=\"row\"><div class=\"col-4 padding-top\"><h3>Create HTTP requests<\/h3><p>Specify URL, payload, headers, and authorization all in one place. Then just hit send.<\/p><\/div><div class=\"col-4 padding-top\"><h3>View response details<\/h3><p>Get all the details on every response. View status code, body, headers, cookies, and more!<\/p><\/div><div class=\"col-4 padding-top\"><h3>Organize everything<\/h3><p>Create workspaces or folders, drag-and-drop requests, and easily import and export your data.<\/p><\/div><\/div><\/div><\/section><section class=\"dark no-margin padding-bottom-lg padding-top-lg\"><div class=\"container center\"><div class=\"row\"><div class=\"col-12\"><h2 class=\"text-xxl\">Much more than an HTTP client<\/h2><p class=\"text-lg\">Advanced authentication helpers, templating, and request chaining help get things done faster.<\/p><p><a class=\"button button--big\" href=\"https:\/\/insomnia.rest\/download\">Download for Desktop<\/a><\/p><\/div><\/div><\/div><\/section><section class=\"no-margin padding-top-lg padding-bottom-lg\"><div class=\"container\"><div class=\"row row-center-y\"><div class=\"col-5\"><h3 class=\"text-xl\">Reusable Values<\/h3><p>Reuse API keys or session IDs. <strong>Define environment variables<\/strong> globally or switch between sub-environments for a seamless development\/production workflow.<\/p><\/div><\/div><div class=\"row row-center-y padding-top\"><div class=\"col-5\"><h3 class=\"text-xl\">Code Snippet Generation<\/h3><p>Generate http code for <strong>over thirty language libraries<\/strong>, including Curl, NodeJS, Go, Swift, Python, Java, C, and others.<\/p><\/div><\/div><div class=\"row row-center-y padding-top\"><div class=\"col-5\"><h3 class=\"text-xl\">Beautiful Interface<\/h3><p>Get started quickly with Insomnia's intuitive interface, and choose from <strong>nine unique themes<\/strong> to tailor the experience to <i>you<\/i>.<\/p><\/div><\/div><\/div><\/section><section class=\"light no-margin padding-bottom-lg padding-top-lg\"><div class=\"container\"><div class=\"row\"><div class=\"col-12 center\"><h2>More than 400,000 developers trust Insomnia<\/h2><div class=\"padding-top\"><p><a class=\"button button--big\" href=\"https:\/\/insomnia.rest\/download\">Download App<\/a>\u00a0\u00a0<a class=\"button button--big button--no-outline\" href=\"https:\/\/insomnia.rest\/teams\">Team Edition<\/a><\/p><\/div><\/div><\/div><\/div><\/section><section class=\"dark no-margin padding-bottom-lg padding-top-lg\"><div class=\"container\"><div class=\"row\"><div class=\"col-12 center\"><h2 class=\"text-xl\">Still not convinced?<\/h2><p>Maybe this big ol' list of features will help<\/p><\/div><\/div><div class=\"row feature-list\"><div class=\"col-6\"><ul><li><a href=\"https:\/\/insomnia.rest\/graphql\/\">GraphQL<\/a> support<\/li><li>OAuth 1.0 and 2.0 auth<\/li><li>Multipart form builder<\/li><li>Query parameter builder<\/li><li>Plugin System<\/li><li>SSL client certificates<\/li><li>JSONPath and XPath<\/li><li>Response history<\/li><li>Data import\/export<\/li><li>Rendered HTML preview<\/li><li>Image and SVG preview<\/li><li>AWS authentication<\/li><li>Configurable proxy<\/li><li>Color themes<\/li><li>Cloud sync and sharing<\/li><\/ul><\/div><div class=\"col-6\"><ul><li>Import from <code>curl<\/code><\/li><li>Digest, Basic, NTLM Auth<\/li><li>Nunjucks templating<\/li><li>Configurable timeout<\/li><li>HAR import<\/li><li>Swagger import<\/li><li>Request filtering<\/li><li>Toggle SSL validation<\/li><li>Keyboard shortcuts<\/li><li>Usable at almost all sizes<\/li><li>NTLM authentication<\/li><li>Responsive interface<\/li><li>Autocomplete Hints<\/li><li>Redirect chain visualization<\/li><li>Mac, Windows and Linux<\/li><\/ul><\/div><\/div><br \/><div class=\"center\"><p>Go on, give it a try. You won't regret it.<\/p><br \/><a class=\"button button--big\" href=\"https:\/\/insomnia.rest\/download\">Download for Desktop<\/a><\/div><br \/><\/div><\/section>","created_at":"2018-11-07T16:53:55+0000","updated_at":"2018-11-07T16:53:55+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"insomnia.rest","preview_picture":"https:\/\/insomnia.rest\/images\/twitter-promo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":627,"title":"PythonInMusic - Python Wiki","url":"https:\/\/wiki.python.org\/moin\/PythonInMusic","content":"<p class=\"line874\">This page is divided in four sections: Music software written in Python, Music software supporting Python, Music programming in Python, and a category of unsorted (may still fit in the above) <\/p><div><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/bluemindo.codingteam.net\/\">Bluemindo<\/a><\/strong> - Bluemindo is a really simple but powerful audio player in Python\/PyGTK, using Gstreamer. Bluemindo is a free (as in freedom) software, released under GPLv3, only. <\/p><p class=\"line891\"><strong><a class=\"https\" href=\"https:\/\/github.com\/holizz\/cplay\">cplay<\/a><\/strong> - a curses front-end for various audio players <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/edna.sourceforge.net\/\">edna<\/a><\/strong> - an MP3 server, edna allows you to access your MP3 collection from any networked computer. The web pages are dynamically constructed, adjusting to directory structure and the files in those directories. This is much nicer than using simple directory indexing. Rather than directly serving up an MP3, the software serves up a playlist. This gets passed to your player (e.g. <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/WinAmp\">WinAmp<\/a>) which turns around with an HTTP request to stream the MP3. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/listengnome.free.fr\/\">Listen<\/a><\/strong> - Music management and playback for GNOME <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/getmediacore.com\/\">MediaCore Audio\/Podcast Player and CMS<\/a><\/strong> - Web based CMS for music management in video, audio and podcast form. All audio, video, and podcasts added to the system are playable from any browser. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.mellowood.ca\/mma\/\">MMA<\/a><\/strong> - Musical Midi Accompaniment. If you follow the above link you will find that Pymprovisator is no longer developed due to the fact that there is this similar, but more powerful GPL Python software. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/peyote.sourceforge.net\/\">Peyote<\/a><\/strong> - Peyote is an audio player with friendly MC-like interface. Peyote is designed specifically for work easy with cue sheets. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/pymprovisator.sourceforge.net\/\">Pymprovisator<\/a><\/strong> - Pymprovisator is a program that emulates the program Band in a Box from PG Music. You can think in it like the electronic version of the books+CD from Jamey Aebersold. You set the basic parameters in a song: title, style, key, chords sequence,... and the program will generate a Midi file with the correct accompaniment. (dev suspended) <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.anti-particle.com\/pymps.shtml\">Pymps<\/a><\/strong> - Pymps is the PYthon Music Playing System - a web based mp3\/ogg jukebox. It's written in Python and utilises the PostgreSQL database. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/albertz.github.com\/music-player\/\">MusicPlayer<\/a><\/strong> - <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/MusicPlayer\">MusicPlayer<\/a> is a high-quality music player implemented in Python, using FFmpeg and <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PortAudio\">PortAudio<\/a>. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/home.gna.org\/pymserv\/\">Pymserv<\/a><\/strong> - PyMServ is a graphical client for mserv, a music server. It is written in Python using pygtk and gconf to store prefs. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.luga.de\/pytone\/\">Pytone<\/a><\/strong> - Pytone is a music jukebox written in Python with a curses based GUI. While providing advanced features like crossfading and multiple players, special emphasis is put on ease of use, turning <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PyTone\">PyTone<\/a> into an ideal jukebox system for use at parties. <\/p><p class=\"line891\"><strong><a class=\"https\" href=\"https:\/\/github.com\/quodlibet\/quodlibet\/\">Quod Libet<\/a><\/strong> - Quod Libet is a GTK+-based audio player written in Python. It lets you make playlists based on regular expressions. It lets you display and edit any tags you want in the file. And it lets you do this for all the file formats it supports -- Ogg Vorbis, FLAC, MP3, Musepack, and MOD. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/turcanator.tangentcode.com\/\">TheTurcanator<\/a><\/strong> - a small midi piano tutor for windows and mac. Includes CoreMIDI wrapper written in pyrex. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/linuxband.org\/\">LinuxBand<\/a><\/strong> - <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/LinuxBand\">LinuxBand<\/a> is a GUI front-end for MMA (Musical MIDI Accompaniment). Type in the chords, choose the groove and <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/LinuxBand\">LinuxBand<\/a> will play a musical accompaniment for you. <\/p><\/div><div><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/github.com\/jiaaro\/pydub\">pydub<\/a><\/strong> - Pydub is a simple and easy high level interface based on ffmpeg and influenced by jquery. It manipulates audio, adding effects, id3 tags, slicing, concatenating audio tracks. Supports python 2.6, 2.7, 3.2, 3.3 <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/audiere.sourceforge.net\/\">audiere<\/a><\/strong> - Audiere is a high-level audio API. It can play Ogg VorbisAU, MP3, FLACAS, uncompressed WAV, AIFF, MOD, S3M, XM, and ITAN files. For audio output, Audiere supports <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/DirectSound\">DirectSound<\/a> or WinMM in Windows, OSS on Linux and Cygwin, and SGI AL on IRIX. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.ar.media.kyoto-u.ac.jp\/members\/david\/softwares\/audiolab\/\">audiolab<\/a><\/strong> - audiolab is a small Python package (now part of <a class=\"http\" href=\"http:\/\/scikits.appspot.com\/audiolab\">scikits<\/a>) to import data from audio files to numpy arrays and export data from numpy arrays to audio files. It uses libsndfile from Erik Castro de Lopo for the underlying IO, which supports many different audio formats: <a class=\"http\" href=\"http:\/\/www.mega-nerd.com\/libsndfile\/\">http:\/\/www.mega-nerd.com\/libsndfile\/<\/a> <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/gstreamer.freedesktop.org\/modules\/gst-python.html\">GStreamer Python Bindings<\/a><\/strong> - GStreamer is a big multimedia library, it is very simple to use it with these python bindings. Many applications rely on it (Exaile, Pitivi, Jokosher, Listen usw.). Online documentation can be found on <a class=\"http\" href=\"http:\/\/pygstdocs.berlios.de\/\">http:\/\/pygstdocs.berlios.de\/<\/a> <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/improviser.onderstekop.nl\/\">improviser<\/a><\/strong> - Automatic music generation software. Experiments in musical content generation. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/python-musical\/\">python-musical<\/a><\/strong> - Python library for music theory, synthesis, and playback. Contains a collection of audio wave generators and filters powered by numpy. Also contains a pythonic music theory library for handling notes, chords, scales. Can load, save, and playback audio. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/loopjam.wordpress.com\/\">LoopJam<\/a><\/strong> - Instant 1 click remixing of sample loops, able to boost your creativity and multiply your sample loop library. Remix audio loops on a slice level, apply up to 9 FX to individual slices or create countless versions using LJ's auto-remix feature (jam) which re-arranges the audio loop forming musical patterns. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.cerlsoundgroup.org\/Loris\/\">Loris<\/a><\/strong> - Loris is an Open Source C++ class library implementing analysis, manipulation, and synthesis of digitized sounds using the Reassigned Bandwidth-Enhanced Additive Sound Model. Loris supports modified resynthesis and manipulations of the model data, such as time- and frequency-scale modification and sound morphing. Loris includes support and wrapper code for building extension modules for various scripting languages (Python, Tcl, Perl). <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/musickit.sourceforge.net\/\">MusicKit<\/a><\/strong> - The <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/MusicKit\">MusicKit<\/a> is an object-oriented software system for building music, sound, signal processing, and MIDI applications. It has been used in such diverse commercial applications as music sequencers, computer games, and document processors. Professors and students in academia have used the <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/MusicKit\">MusicKit<\/a> in a host of areas, including music performance, scientific experiments, computer-aided instruction, and physical modeling. PyObjC is required to use this library in Python. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/ekyo.nerim.net\/software\/pyogg\/index.html\">pyao<\/a><\/strong> - pyao provides Python bindings for <a class=\"http\" href=\"http:\/\/xiph.org\/ao\/\">libao<\/a>, a cross-platform audio output library. It supports audio output on Linux (OSS, ALSA, <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PulseAudio\">PulseAudio<\/a>, esd), MacOS X, Windows, *BSD and some more. There are ready-to-use packages in <a class=\"http\" href=\"http:\/\/packages.debian.org\/search?keywords=python-pyao\">Debian<\/a>\/<a class=\"http\" href=\"http:\/\/ns2.canonical.com\/search?searchon=names&amp;keywords=python-pyao\">Ubuntu<\/a>, and Audio output is as easy as: <code class=\"backtick\">import\u00a0ao;\u00a0pcm\u00a0=\u00a0ao.Audio<\/code><code class=\"backtick\">Device(\"pulse\");\u00a0pcm.play(data)<\/code> <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/people.csail.mit.edu\/hubert\/pyaudio\/\">pyAudio<\/a><\/strong> - <a href=\"https:\/\/wiki.python.org\/moin\/PyAudio\">PyAudio<\/a> provides Python bindings for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PortAudio\">PortAudio<\/a>, the cross-platform audio I\/O library. Using <a href=\"https:\/\/wiki.python.org\/moin\/PyAudio\">PyAudio<\/a>, you can easily use Python to play and record audio on a variety of platforms. Seems to be a successor of fastaudio, a once popular binding for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PortAudio\">PortAudio<\/a> <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/pyfluidsynth\/\">pyFluidSynth<\/a><\/strong> - Python bindings for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/FluidSynth\">FluidSynth<\/a>, a MIDI synthesizer that uses <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/SoundFont\">SoundFont<\/a> instruments. This module contains Python bindings for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/FluidSynth\">FluidSynth<\/a>. <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/FluidSynth\">FluidSynth<\/a> is a software synthesizer for generating music. It works like a MIDI synthesizer. You load patches, set parameters, then send NOTEON and NOTEOFF events to play notes. Instruments are defined in <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/SoundFonts\">SoundFonts<\/a>, generally files with the extension SF2. <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/FluidSynth\">FluidSynth<\/a> can either be used to play audio itself, or you can call a function that returns chunks of audio data and output the data to the soundcard yourself. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/pygame.org\/\">Pygame<\/a><\/strong> - Pygame is a set of Python modules designed for writing games. It is written on top of the excellent SDL library. This allows you to create fully featured games and multimedia programs in the Python language. Pygame is highly portable and runs on nearly every platform and operating system. .ogg .wav .midi .mod .xm .mp3. Sound output. midi input and output. Load sounds into numeric and numpy arrays. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/pymedia.org\">PyMedia<\/a><\/strong> - (Not updated since 2006) <a href=\"https:\/\/wiki.python.org\/moin\/PyMedia\">PyMedia<\/a> is a Python module for the multimedia purposes. It provides rich and simple interface for the digital media manipulation( wav, mp3, ogg, avi, divx, dvd, cdda etc ). It includes parsing, demutiplexing, multiplexing, coding and decoding. It can be compiled for Windows, Linux and cygwin. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/pyo\/\">pyo<\/a><\/strong> - pyo is a Python module containing classes for a wide variety of audio signal processing types. With pyo, user will be able to include signal processing chains directly in Python scripts or projects, and to manipulate them in real time through the interpreter. Tools in pyo module offer primitives, like mathematical operations on audio signal, basic signal processing (filters, delays, synthesis generators, etc.), but also complex algorithms to create sound granulation and others creative audio manipulations. pyo supports OSC protocol (Open Sound Control), to ease communications between softwares, and MIDI protocol, for generating sound events and controlling process parameters. pyo allows creation of sophisticated signal processing chains with all the benefits of a mature, and wildly used, general programming language. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/zyne\/\">Zyne<\/a><\/strong> - Zyne is a Python modular synthesizer using pyo as its audio engine. Zyne comes with more than 10 builtin modules implementing different kind of synthesis engines and provides a simple API to create your own custom modules. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/soundgrain\/\">Soundgrain<\/a><\/strong> - Soundgrain is a graphical interface where users can draw and edit trajectories to control granular sound synthesis modules. Soundgrain is written with Python and <a href=\"https:\/\/wiki.python.org\/moin\/WxPython\">WxPython<\/a> and uses pyo as its audio engine. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/andy.bigwhitebox.org\/pyper\/\">Pyper<\/a><\/strong> - (Not updated since early 2005) Pyper is a musical development environment. It allows you to write Python scripts that generates music in real-time. Pyper uses <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/QuickTime\">QuickTime<\/a> Musical Instruments for synthesis. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/pysonic.sourceforge.net\/\">pySonic<\/a><\/strong> - (Not updated since 2005) pySonic is a Python wrapper around the high performance, cross platform, but closed source, <a class=\"http\" href=\"http:\/\/www.fmod.org\/\">FMOD sound library<\/a>. You get all the benefits of the FMOD library, but in a Pythonic, object oriented package. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/sndobj.sourceforge.net\/\">PySndObj<\/a><\/strong> - The Sound Object Library is an object-oriented audio processing library. It provides objects for synthesis and processing of sound that can be used to build applications for computer-generated music. The core code, including soundfile and text input\/output, is fully portable across several platforms. Platform-specific code includes realtime audio IO and MIDI input support for Linux (OSS,ALSA and Jack), Windows (MME and ASIO), MacOS X (<a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/CoreAudio\">CoreAudio<\/a>, but no MIDI at moment), Silicon Graphics (Irix) machines and any Open Sound System-supported UNIX. The <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/SndObj\">SndObj<\/a> library also exists as Python module, aka <a href=\"https:\/\/wiki.python.org\/moin\/PySndObj\">PySndObj<\/a>. The programming principles for Python <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/SndObj\">SndObj<\/a> programming are similar to the ones used in C++. It is also possible to use the Python interpreter for on-the-fly synthesis programming. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/home.arcor.de\/mdoege\/pysynth\/\">PySynth<\/a><\/strong> - A simple music synthesizer. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.speech.kth.se\/snack\/\">Snack<\/a><\/strong> - (last update: December 2005) The Snack Sound Toolkit is designed to be used with a scripting language such as Tcl\/Tk or Python. Using Snack you can create powerful multi-platform audio applications with just a few lines of code. Snack has commands for basic sound handling, such as playback, recording, file and socket I\/O. Snack also provides primitives for sound visualization, e.g. waveforms and spectrograms. It was developed mainly to handle digital recordings of speech (being developped at the KTH music&amp;speech department), but is just as useful for general audio. Snack has also successfully been applied to other one-dimensional signals. The combination of Snack and a scripting language makes it possible to create sound tools and applications with a minimum of effort. This is due to the rapid development nature of scripting languages. As a bonus you get an application that is cross-platform from start. It is also easy to integrate Snack based applications with existing sound analysis software. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/github.com\/danilobellini\/audiolazy\/\">AudioLazy<\/a><\/strong> - Real-Time Expressive Digital Signal Processing (DSP) Package for Python, using any Python iterable as a [-1;1] range audio source. Has time-variant linear filters as well as LTI filters using Z-Transform equations like <code>1\u00a0-\u00a0z\u00a0**\u00a0-1<\/code>, as well as analysis (ZCR \/ zero crossing rate, LPC \/ Linear Predictive Coding, AMDF, etc.), synthesis (table lookup, ADSR, etc.), ear modeling (Patterson-Holdsworth with gammatone filters and ERB models), and multiple implementation of common filters (lowpass, highpass, comb, resonator), among several other resources (e.g. Lagrange polynomial interpolation, simple converters among MIDI pitch \/ frequency \/ string). Works mainly with <code>Stream<\/code> instances for its signal outputs, a generator-like (lazy) iterable with elementwise\/broadcast-style operators similar to the Numpy array operators. Integrated with Matplotlib for LTI filter plotting, although it doesn't require Matplotlib nor Numpy for computation, DSP or I\/O. Emphasizes sample-based processing while keeping block-based processing easy to be done, this package can also be seen as a highly enhanced itertools. Pure Python, multiplatform, compatible with Python 2.7 and 3.2+, uses <a href=\"https:\/\/wiki.python.org\/moin\/PyAudio\">PyAudio<\/a> for audio I\/O (if needed). Can be used together with Scipy, Sympy, music21 and several other packages, none required for DSP computation based on Python iterables. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/python-sounddevice.rtfd.org\/\">sounddevice<\/a><\/strong> - This module provides bindings for the <a class=\"http\" href=\"http:\/\/www.portaudio.com\/\">PortAudio<\/a> library (using <a class=\"https\" href=\"https:\/\/cffi.rtfd.org\/\">CFFI<\/a>) and a few convenience functions to play and record <a class=\"http\" href=\"http:\/\/www.numpy.org\/\">NumPy<\/a> arrays containing audio signals. <\/p><\/div><div><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/eyed3.nicfit.net\/\">eyed3<\/a><\/strong> - eyeD3 is a Python module and program for processing ID3 tags. Information about mp3 files (i.e bit rate, sample frequency, play time, etc.) is also provided. The formats supported are ID3 v1.0\/v1.1 and v2.3\/v2.4. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/quodlibet\/wiki\/Development\/Mutagen\">mutagen<\/a><\/strong> - Mutagen is a Python module to handle audio metadata. It supports ASF, FLAC, M4A, Monkey's Audio, MP3, Musepack, Ogg FLAC, Ogg Speex, Ogg Theora, Ogg Vorbis, True Audio, <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/WavPack\">WavPack<\/a> and OptimFROG audio files. All versions of ID3v2 are supported, and all standard ID3v2.4 frames are parsed. It can read Xing headers to accurately calculate the bitrate and length of MP3s. ID3 and APEv2 tags can be edited regardless of audio format. It can also manipulate Ogg streams on an individual packet\/page level. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/id3-py.sourceforge.net\/\">ID3.py<\/a><\/strong> - This module allows one to read and manipulate so-called ID3 informational tags on MP3 files through an object-oriented Python interface. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.nedbatchelder.com\/code\/modules\/id3reader.html\">id3reader.py<\/a><\/strong> - Id3reader.py is a Python module that reads ID3 metadata tags in MP3 files. It can read ID3v1, ID3v2.2, ID3v2.3, or ID3v2.4 tags. It does not write tags at all. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.mpgedit.org\/mpgedit\/\">mpgedit<\/a><\/strong> - mpgedit is an MPEG 1 layer 1\/2\/3 (mp3), MPEG 2, and MPEG 2.5 audio file editor that is capable of processing both Constant Bit Rate (CBR) and Variable Bit Rate (VBR) encoded files. mpgedit can cut an input MPEG file into one or more output files, as well as join one or more input MPEG files into a single output file. Since no file decoding \/ encoding occurs during editing, there is no audio quality loss when editing with mpgedit. A Python development toolkit enables Python developers to utilize the core mpgedit API, providing access to mp3 file playback, editing and indexing functionality. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/sourceforge.net\/projects\/m3ute2\/\">m3ute2<\/a><\/strong> - m3ute2 is program for copying, moving, and otherwise organizing M3U playlists and directories. m3ute2 can also generate detailed reports about lists of files. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/sourceforge.net\/projects\/mmpython\/\">mmpython<\/a><\/strong> - MMPython is a Media Meta Data retrieval framework. It retrieves metadata from mp3, ogg, avi, jpg, tiff and other file formats. Among others it thereby parses ID3v2, ID3v1, EXIF, IPTC and Vorbis data into an object oriented struture. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/doc.freevo.org\/2.0\/SourceDoc\/KaaMetadata\">KaaMetadata<\/a><\/strong> Sucessor of MMPython. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/github.com\/myers\/pyid3\">PyID3<\/a><\/strong> - pyid3 is a pure Python library for reading and writing id3 tags (version 1.0, 1.1, 2.3, 2.4, readonly support for 2.2). What makes this better than all the others? Testing! This library has been tested against some 200+ MB of just tags. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/beets.radbox.org\/\">beets<\/a><\/strong> - music tag correction and cataloging tool. Consists of both a command-line interface for music manipulation and a library for building related tools. Can automatically correct tags using the <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/MusicBrainz\">MusicBrainz<\/a> database. <\/p><p class=\"line862\">see also: <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PySonic\">PySonic<\/a> for programmable MP3 playback <\/p><\/div><div><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.pygame.org\/docs\/ref\/midi.html\">pygame.midi<\/a><\/strong> - is a portmidi wrapper orginally based on the pyportmidi wrapper. Also pygame.music can play midi files. Can get input from midi devices and can output to midi devices. For osx, linux and windows. New with pygame 1.9.0. <em>python -m pygame.examples.midi --output<\/em> <\/p><p class=\"line891\"><strong><a class=\"https\" href=\"https:\/\/sourceforge.net\/projects\/pymidi\/\">pyMIDI<\/a><\/strong> - Provides object oriented programmatic manipulation of MIDI streams. Using this framework, you can read MIDI files from disk, build new MIDI streams, process, or filter preexisting streams, and write your changes back to disk. If you install this package on a Linux platform with alsalib, you can take advantage of the ALSA kernel sequencer, which provides low latency scheduling and receiving of MIDI events. SWIG is required to compile the ALSA extension sequencer extension. Although OS-X and Windows provide similar sequencer facilities, the current version of the API does not yet support them. Some buggs are remaining in this package (for example when trying to delete a track), it has not been updated since 2006. This package is by Giles Hall. A sourceforge download. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/groups.google.com\/groups?hl=en&amp;selm=3C0F67FE.96E7CAE8@alum.mit.edu\">midi.py<\/a><\/strong> - (DEAD LINK) - Python MIDI classes: meaningful data structures that represent MIDI events and other objects. You can read MIDI files to create such objects, or generate a collection of objects and use them to write a MIDI file. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.pjb.com.au\/midi\/MIDI.html\">MIDI.py<\/a><\/strong> - This module offers functions: concatenate_scores(), grep(), merge_scores(), mix_scores(), midi2opus(), midi2score(), opus2midi(), opus2score(), play_score(), score2midi(), score2opus(), score2stats(), score_type(), segment(), timeshift() and to_millisecs(). Uses Python3. There is a call-compatible Lua module. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/sourceforge.net\/projects\/uncassist\">PMIDI<\/a><\/strong> - The PMIDI library allows the generation of short MIDI sequences in Python code.The interface allows a programmer to specify songs, instruments, measures, and notes. Playback is handled by the Windows MIDI stream API so proper playback timing is handled by the OS rather than by client code. The library is especially useful for generating earcons. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/gitorious.org\/portmidizero\/mainline\/trees\/master\">portmidizero<\/a><\/strong> - portmidizero is a simple ctypes wrapper for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PortMidi\">PortMidi<\/a> in pure Python. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/chordrecognizer.sourceforge.net\/\">PyChoReLib<\/a><\/strong> - Python Chord Recognition Library. This is a library that implements the transformation from a list of notenames to a chord name. The system can be taught new chords by example: tell it that ['c', 'e', 'g'] is called a 'C' chord, and using its built-in music knowledge it immediately recognizes all major triads in all keys and all inversions\/permutations. Comes with a real-time midi-input demo program (needs <a href=\"https:\/\/wiki.python.org\/moin\/PyPortMidi\">PyPortMidi<\/a>). <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.hyperreal.org\/~est\/python\/MIDI\/\">PyMIDI<\/a><\/strong> - The MIDI module provides MIDI input parsers for Python. Package not updated since 2000. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/web.media.mit.edu\/~harrison\/pyportmidi.html\">PyPortMidi<\/a><\/strong> - <a href=\"https:\/\/wiki.python.org\/moin\/PyPortMidi\">PyPortMidi<\/a> is a Python wrapper for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PortMidi\">PortMidi<\/a>. <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PortMidi\">PortMidi<\/a> is a cross-platform C library for realtime MIDI control. Using <a href=\"https:\/\/wiki.python.org\/moin\/PyPortMidi\">PyPortMidi<\/a>, you can send and receive MIDI data in realtime from Python. Besides using <a href=\"https:\/\/wiki.python.org\/moin\/PyPortMidi\">PyPortMidi<\/a> to communicate to synthesizers and the like, it is possible to use <a href=\"https:\/\/wiki.python.org\/moin\/PyPortMidi\">PyPortMidi<\/a> as a way to send MIDI messages between software packages on the same computer. <a href=\"https:\/\/wiki.python.org\/moin\/PyPortMidi\">PyPortMidi<\/a> is now maintained at <a class=\"http\" href=\"http:\/\/bitbucket.org\/aalex\/pyportmidi\/\">http:\/\/bitbucket.org\/aalex\/pyportmidi\/<\/a> <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.mxm.dk\/products\/public\/pythonmidi\">PythonMIDI<\/a><\/strong> - The Python Midi package is a collection of classes handling Midi in and output in the Python programming language. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.sci.ccny.cuny.edu\/~brinkman\/software\/pyseq\/\">PySeq<\/a><\/strong> - Python bindings for ALSA using ctypes <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.quitte.de\/milk.html\">milk<\/a><\/strong> - Superceding the older <a class=\"http\" href=\"http:\/\/www.quitte.de\/nam.html\">Nam<\/a>, milk provides Python with classes representing key MIDI sequencer components: MIDI I\/O, <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/EventLists\">EventLists<\/a>, Plugins and a realtime Flow class. The components can be freely interconnected in a fashion very similar to physical MIDI cabling, however the milk event system is not limited to MIDI events alone; you can define your own extensions should the need arise. Website says it is unpolished and unfinished. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/trac2.assembla.com\/pkaudio\/wiki\/pyrtmidi\">pyrtmidi<\/a><\/strong> - rtmidi provides realtime MIDI input\/output across Linux (ALSA), Macintosh OS X, SGI, and Windows (Multimedia Library) operating systems. It is very fast, has a clean and pythonic interface, and supports virtual ports, according to author Patrick Kidd. In fact it is a wrapper for Gary Scavone's rtmidi from <a class=\"http\" href=\"http:\/\/www.music.mcgill.ca\/~gary\/rtmidi\/\">here<\/a>, rather than the address on this website: <\/p><p class=\"line891\"><strong><a class=\"https\" href=\"https:\/\/github.com\/superquadratic\/rtmidi-python\/\">rtmidi-python<\/a><\/strong> - Another <a class=\"http\" href=\"http:\/\/www.music.mcgill.ca\/~gary\/rtmidi\/\">RtMidi<\/a> wrapper. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.sabren.net\/code\/python\/midi\/\">winmidi.pyd<\/a><\/strong> - A demo? of a Python extension interfacing to the native windows midi libs that developed from <a class=\"http\" href=\"http:\/\/www.sabren.net\/rants\/2000\/01\/20000129a.php3\">earlier<\/a> attempts. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/groups.google.com\/group\/pymidiwin\/files\">win32midi<\/a><\/strong> - Some Python samples to demonstrate how to output MIDI stream on MS windows platform. Unlike previous links, these samples playback MIDI by directly calling the Win32 MIDI APIs without an intermediate portable library. It provides a simple player class for playing with MIDI sound using the synthesizer on the soundcard\/onboard soundchip. A sample script is provided for testing it out. As it is still a work in progress, bugs are expected. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/midiutil\/\">midiutil<\/a><\/strong> - A pure Python library for generating Midi files <\/p><p class=\"line891\"><strong><a class=\"https\" href=\"https:\/\/github.com\/kroger\/pyknon\">Pyknon<\/a><\/strong> - Pyknon is a simple music library for Python hackers. With Pyknon you can generate Midi files quickly and reason about musical proprieties. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.desfonema.com.ar\/software.html\">Desfonema Sequencer<\/a><\/strong> - A tracker minded MIDI sequencer for Linux (ALSA) written in Python\/PyGTK <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/code.google.com\/p\/python-music-gen\/\">python-music-gen<\/a><\/strong> - Simple library to generate midi patterns from numbers. Useful for building generative music tools. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/sevkeifert.blogspot.com\/2015\/11\/simple-graphical-interface-for.html\">fluidsynth-gui<\/a><\/strong> - Graphical User Interface for <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/FluidSynth\">FluidSynth<\/a>, and an alternative to Qsynth. <\/p><\/div><div><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.reaper.fm\">REAPER<\/a><\/strong> - \"Audio Production Without Limits\": REAPER is a professional digital audio workstation (DAW) for Windows, OS X and WINE. It comes with an uncrippled evaluation licence and supports advanced audio and MIDI recording, arranging and mixing. The support of several plugin formats (like VST, DX and AU) as well as the extremely flexible routing capabilities make it a powerful production suite. Since version 3.12 REAPER is scriptable with Python, allowing access to internal actions and parts of the API. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.ableton.com\">Ableton Live<\/a><\/strong> - Award-winning commercial music creation, production and performance platform for Mac OS and Windows. Live is far and away one of the most interesting and groundbreaking audio recording and sequencing tools to come along in the past five years. Live uses Python internally and an experimental API has been exposed at <a class=\"http\" href=\"http:\/\/code.google.com\/p\/liveapi\/\">this site<\/a>, and there is a discussion group <a class=\"http\" href=\"http:\/\/groups.google.com\/group\/liveapi\">here<\/a>. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/csounds.com\/stevenyi\/blue\/index.html\">blue<\/a><\/strong> - blue is a Java program for use with Csound. It's interface is much like a digital multitrack, but differs in that there timelines within timelines (polyObjects). This allows for a compositional organization in time that seems to me to be very intuitive, informative, and flexible. soundObjects are the building blocks within blue's score timeline. soundObjects can be lists of notes, algorithmic generators, Python script code, csound instrument definitions, and whatever plugins that are developed for blue. these soundObjects may be text based, but they can be completely GUI based as well. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.jokosher.org\/\">Jokosher<\/a><\/strong> - Jokosher is a simple yet powerful multi-track studio. With it you can create and record music, podcasts and more, all from an integrated simple environment. Jokosher is written in Python and uses the GNOME platform and the GTK widget set. The audio engine is powered by GStreamer, and we use Cairo for some of the graphics. <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/www.jeskola.net\/buzz\">Jeskola Buzz Modular<\/a><\/strong> - Buzz is a modular audio host that saw the beginning of it's development in 1997 leading the way in open plugin-format hosting (pre-VST) and a unique spin on modular routing using modules called Machines in the form of Generators, Effects, Control machines. Buzz's implementation of Python comes through the use of the Control plugin called <a class=\"nonexistent\" href=\"https:\/\/wiki.python.org\/moin\/PyBuzz\">PyBuzz<\/a>, a fully customizable and assignable meta-editor created by Leonard Ritter (creator of the Lunar audio library and the Linux modular host, Aldrin). A discussion group can be found <a class=\"http\" href=\"http:\/\/www.buzzchurch.com\">here<\/a> <\/p><p class=\"line891\"><strong><a class=\"http\" href=\"http:\/\/sourceforge.net\/projects\/libmodsynth\/\">PyDAW<\/a><\/strong> - PyDAW is a powerful pattern-based DAW and plugin suite for producing electronic music.  The UI is written entirely in Python\/PyQt, and the audio engine in C. <\/p><\/div>","created_at":"2018-11-07T16:53:54+0000","updated_at":"2018-11-07T16:53:54+0000","annotations":[],"mimetype":"text\/html","reading_time":18,"domain_name":"wiki.python.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":628,"title":"How I do my Computing","url":"https:\/\/stallman.org\/stallman-computing.html","content":"<div class=\"comp\"><h2>My computer<\/h2><p>\n        I use a Thinkpad T400s computer, which has a free\n        initialization program (libreboot) and a free operating system\n        (Trisquel GNU\/Linux). It was not sold that way by Lenovo,\n        however; small businesses buy them used, recondition them, and\n        install the free software.  This is one of the\n        computers <a href=\"https:\/\/www.fsf.org\/resources\/hw\/systems\">\n        endorsed by the FSF<\/a>.\n    <\/p><p>\n        Before using such Thinkpads, I used the Lemote Yeeloong for\n        several years.  At the time, it was the only laptop one could\n        buy that could run a free initialization program and a free\n        operating system.  But it was never sold with a free operating\n        system.\n    <\/p><p>\n        Before that, I used an OLPC for some weeks.  The OLPC uses\n        a nonfree firmware blob for the WiFi, so I could not use the internal\n        WiFi device.  No big problem, I used an external WiFi adaptor.<br \/>I stopped using it because the OLPC project decided to make their\n        machine support Windows, so I did not want to appear to endorse it\n        by visibly carrying it around.  I could have continued using it\n        privately with its free software installation, but I had no need\n        for another computer to use only privately.\n    <\/p><p>\n        The results that seemed likely, millions of children running Windows\n        on the OLPC, have not occurred.  Instead we see millions of children\n        running Windows on the Intel Classmate, or nowadays a Chromebook\n\tthat sends the child's personal data to Google.\n    <\/p><p>\n        Before that I used machines that ran completely free GNU\/Linux\n        systems but had nonfree BIOSes.  I tried for about 8 years to find a\n        way to avoid the nonfree BIOS in some commercial machine.\n    <\/p><h2>GNU\/Linux distro<\/h2><p>\n        I do not have a preferred GNU\/Linux distro.  I recommend all the\n        <a href=\"http:\/\/www.gnu.org\/distros\">ethical distros<\/a> \u2014\n        namely, those that are 100% free software.\n    <\/p><p>\n        I've chosen not to have any preferences among those ethical distros.\n        In fact, I am not in a position to judge them on other criteria:\n        even to try them all would be a lot work of that I have no need to do.\n    <\/p><h2>What I do on my computer<\/h2><p>\n        Mostly I use a text console, for convenience's sake.  Most of\n        my work is editing text and that is more efficient on a text\n        console. On the text console, the touchpad can't cause me any\n        trouble if I touch it by accident.\n    <\/p><p>\n        I do use X11 for tasks that need a graphical interface.\n        I have no preferred graphical environment or window manager.\n        Since my interest in using graphical environments is small,\n        I don't want to spend time comparing them.\n    <\/p><p>\n        This is not an ethical issue, just my own personal preference.  On\n        the ethical level, I think it is important for free software to\n        provide convenient free graphical user interface software, which is\n        why the GNU Project launched three projects to develop that.  The\n        third, <a href=\"http:\/\/gnome.org\/\">GNOME<\/a>, was successful, so we\n        never needed a fourth one.\n    <\/p><p>\n        I spend most of my time editing in Emacs.  I read and send mail with\n        Emacs using M-x rmail and C-x m.  I have no experience with any\n        other email client programs.  In principle I would be glad to know\n        about other free email clients, but learning about them is not a\n        priority for me and I don't have time.\n    <\/p><p>\n        I edit the pages on this site with Emacs also, although volunteer\n        helpers install the political notes and urgent notes.\n        I have no experience with other ways of maintaining web sites.\n        In principle I would be glad to know about other ways, but learning\n        about them is low priority for me and I have other things to do.\n    <\/p><h2>How this site is maintained<\/h2><p>\n        This site is maintained in a very simple way.  I edit the pages such\n        as this one manually as HTML.  I only know simple HTML; others who\n        know more wrote the parts at the top and bottom of pages, and the\n        more complex formatting on the home page.  Volunteer helpers install\n        the political notes every day after receiving the text from me by\n        email.  A cron job \"rolls over\" the political notes page every two\n        months.  The photo galleries are generated with <a href=\"http:\/\/www.netpurgatory.com\/downloads.html#thumbnail\">this\n        perl script<\/a>.  The search feature on the site is done with\n        <a href=\"https:\/\/stallman.org\/site-search\/StallmanOrgSearch.tar.gz\">this code<\/a>.\n    <\/p><p>\n        An <a href=\"http:\/\/contemporary-home-computing.org\/RUE\/\">\n        explanation of the concept of designing<\/a> a \"user experience\"\n        which also shows why I find it loathsome.\n        This is why I want stallman.org to remain simple: not a \"user\n        experience\" but rather a place where I present certain information,\n        views and action opportunities to you.\n    <\/p><p>\n        Would you like to help do this?  Write to rms at the site gnu.org.\n    <\/p><h2>How I use the Internet<\/h2><ul><li>\n            <p>\n                I have used the Internet since it first existed.\n                I never used UUCP, though occasionally I sent emails to\n                addresses that involved transmission via UUCP.\n        <\/p><\/li>\n        <li>\n            <p>\n                I am careful in how I connect to the internet.\n                <\/p><p>Specifically, I refuse to connect through portals that would\n                require me to identify myself, or to run any nontrivial\n                <a href=\"http:\/\/gnu.org\/philosophy\/javascript-trap.html\">nonfree\n                Javascript code<\/a>. I don't mind giving an identity that isn't\n                really me, if that works.\n            <\/p><p>\n                I often connect in a person's home.  The <em>person<\/em>\n                of course knows who I am, but that does not bother me.  What I\n                would object to is putting my identity in a database that can be\n                searched.  I prevent that by changing my mac address at each\n                location.\n        <\/p><\/li>\n        <li>\n            <p>\n                <a name=\"internetuse\" id=\"internetuse\">I am careful in how I use the Internet<\/a>.\n                <\/p><p>I generally do not connect to web sites from my own machine,\n                aside from a few sites I have some special relationship with.\n                I usually fetch web pages from other sites by sending mail to a\n                program (see https:\/\/git.savannah.gnu.org\/git\/womb\/hacks.git)\n                that fetches them, much like wget, and then mails them back to\n                me.  Then I look at them using a web browser, unless it is easy\n                to see the text in the HTML page directly.  I usually try lynx\n                first, then a graphical browser if the page needs it\n                (using konqueror, which won't fetch from other sites in such a\n                situation).\n            <\/p><p>\n                I occasionally also browse unrelated sites using IceCat via Tor.\n                Except for rare cases, I do not identify myself to them. I think\n                that is enough to prevent my browsing from being connected with\n                me. IceCat blocks tracking tags and most fingerprinting methods.\n            <\/p><p>\n                I never pay for anything on the Web.  Anything on the net that\n                requires payment, I don't do.  (I made an exception for the fees\n                for the stallman.org domain, since that is connected with me\n                anyway.)  I also avoid paying with <a href=\"https:\/\/stallman.org\/rms-lifestyle.html#cards\">\n                credit cards<\/a>.  For freedom's sake, insist on paying cash.\n                When a business pressures you to pay in an identified way, that\n                means your help as a citizen is needed: say, \"If you won't take\n                my cash, no sale!\"\n        <\/p><\/li>\n        <li>\n            <p>\n                I would not mind paying for a copy of <a href=\"https:\/\/stallman.org\/ebooks.pdf\">an\n                e-book<\/a> or <a href=\"https:\/\/stallman.org\/internet-music.html\">music recording<\/a>\n                on the Internet if I could do so anonymously, and it were\n                ethical in other ways (no DRM or EULA).  But that option almost\n                never exists. I keep looking for ways to make it exist.\n        <\/p><\/li>\n        <li>\n            <p>\n                For searching, I have mostly used DuckDuckGo for the past\n                few years.  It does work with JS disabled, but you have\n                to follow a link before you search.\n            <\/p><p>\n                I also sometimes use ixquick.com.  My usual\n                precautions should stop them from knowing it is me.\n            <\/p><p>\n                I no longer user Google search, not even occasionally, because\n                it sends me a broken CAPTCHA.  I suspect the reason it tries to\n                send me a CAPTCHA is that I am coming through Tor.\n                I would answer the CAPTCHA if that worked.\n            <\/p><p>\n                I suspect that the reason the CAPTCHA is broken is that it\n                depends on nonfree Javascript.  I won't run that in my computer.\n                I am not willing to let Google see where I am, so I won't bypass\n                Tor. Therefore, I can't use Google search any more.\n        <\/p><\/li>\n    <\/ul><h2>Social media<\/h2><ul><li>\n            <p>\n                <a name=\"social-networking\">I do not use any social networking\n                sites<\/a> because that way of working is inconvenient for me.\n                That doesn't mean I think they are all unethical.  Some are,\n                some are not. Social networking sites raise their own set of\n                ethical issues, completely different from the ethical issues of\n                distributing software (<a href=\"http:\/\/gnu.org\/philosophy\/free-software-even-more-important.html\">\n                free vs proprietary<\/a>), and there are big differences between\n                them.\n            <\/p><p>\n                <a name=\"twitter\" id=\"twitter\">I have a Twitter account<\/a> called\n                rmspostcomments, which I use to log in on other sites to post\n                comments on articles. I never post on Twitter.  Someone made an\n                account stallman_feed which I'm told posts something about my\n                political notes. Any other Twitter account that claims to be\n                mine is an impostor.\n            <\/p><p>\n                The rms account on gnusocial.no repeats the political notes from\n                this site, but I do not post on it directly.  That site runs\n                <a href=\"http:\/\/gnu.org\/software\/social\">GNU Social<\/a>.\n            <\/p><p>\n                Aside from those two, any account on a social networking site\n                that says it is mine is an impostor.\n            <\/p><p>\n                I do not post on 4chan.  I have nothing against it in principle,\n                but I am told a lot of the posts nowadays are right-wing bigotry\n                which I condemn totally.\n                I have occasionally answered questions for interviews for 4chan,\n                but I have never posted anything there. Any posting there that\n                says it is by me is by an impostor.\n        <\/p><\/li>\n        <li>\n            <p>\n                I have <a name=\"facebook\" id=\"facebook\">never had a Facebook account<\/a>, or a\n                Google+ account.  Some impostor created a Facebook account using\n                my name.  The page is not mine.  The Google+ account using my\n                name is also not mine.\n            <\/p><p id=\"realname\">\n                I reject Facebook because it requires each used (i.e., person\n                used by Facebook) to have just one account, which means that all\n                the person's activities are grouped together.  It also insists\n                on knowing the person's usual name, and it is starting to demand\n                a series of different photos.\n            <\/p><p>\n                I am proud to identify myself when stating my views; I can\n                afford to do that because I am in a fairly safe position.  There\n                are people who rationally fear reprisals (from employers,\n                gangsters, right-wing extremists, or the state) if they sign\n                their name to their views. For their sake, let's reject any\n                social networking site which insists on connecting an account to\n                a person's real identity.\n            <\/p><p>\n                Of course, <a href=\"https:\/\/stallman.org\/facebook.html\">Facebook is bad for\n                many other reasons as well.<\/a>\n            <\/p><p>\n                Google+ formerly required knowing the user's real name, but\n                no longer.  However, it does require identification in the form\n                of a phone number. Meanwhile, Google+ has another unacceptable\n                injustice: it requires\n                <a href=\"https:\/\/www.gnu.org\/philosophy\/javascript-trap.html\">\n                running nonfree Javascript code<\/a> to post a message.\n        <\/p><\/li>\n    <\/ul><h2>E-mail service<\/h2><p id=\"email\">\n        People sometimes ask me to recommend an email service.\n        The two ethical issues for an email service are (1) whether you can use\n        it without running any nonfree software (including nonfree Javascript\n        code from the site), and (2) whether it respects your privacy.\n        <\/p><p>For issue 1, see <a href=\"https:\/\/www.fsf.org\/resources\/webmail-systems\">\n        the FSF's page<\/a>.\n        On issue 2, I have no way to verify that any email\n        service is satisfactory.  Therefore, I have no recommendation to\n        offer.\n        <\/p><p>However, I can suggest that it may be wise to use an email service\n        that is not connected with your search engine.  That way you can be\n        almost sure that your email contents don't influence your search\n        results.  You shouldn't identify yourself to your search engine in\n        any case.\n    <\/p><h2>Programming languages<\/h2><ul><li>\n            <p>\n                <a name=\"lisp\" id=\"lisp\">The most powerful programming language is Lisp.<\/a>\n                If you don't know Lisp (or its variant, Scheme), you don't know\n                what it means for a programming language to be powerful and\n                elegant. Once you learn Lisp, you will see what is lacking in\n                most other languages.\n            <\/p><p>\n                Unlike most languages today, which are focused on defining\n                specialized data types, Lisp provides a few data types which are\n                general.  Instead of defining specific types, you build\n                structures from these types.  Thus, rather than offering a way\n                to define a list-of-this type and a list-of-that type, Lisp has\n                one type of lists which can hold any sort of data.\n            <\/p><p>\n                Where other languages allow you to define a function to search a\n                list-of-this, and sometimes a way to define a generic\n                list-search function that you can instantiate for list-of-this,\n                Lisp makes it easy to write a function that will search any list\n                \u2014 and provides a range of such functions.\n            <\/p><p>\n                In addition, functions and expressions in Lisp are represented\n                as data in a way that makes it easy to operate on them.\n            <\/p><p>\n                When you start a Lisp system, it enters a read-eval-print loop.\n                Most other languages have nothing comparable to `read',\n                nothing comparable to `eval', and nothing comparable to `print'.\n                What gaping deficiencies!\n            <\/p><p>\n                While I love the power of Lisp, I am not a devotee of functional\n                programming.  I see nothing bad about side effects and I do not\n                make efforts to avoid them unless there is a practical reason.\n                There is code that is natural to write in a functional way, and\n                code that is more natural with side effects, and I do not\n                campaign about the question. I limit my campaigning to issues of\n                freedom and justice, such as to eliminate nonfree software from\n                the world.\n            <\/p><p>\n                Lisp is no harder to understand than other languages.  So if you\n                have never learned to program, and you want to start, start with\n                Lisp.  If you learn to edit with Emacs, you can learn Lisp by\n                writing editing commands for Emacs.  You can use the\n                Introduction to Programming in Emacs Lisp to learn with: it is <a href=\"http:\/\/gnu.org\/philosophy\/free-doc.html\">free as in\n                freedom<\/a>, and you can <a href=\"http:\/\/shop.fsf.org\/category\/books\">order printed copies<\/a>\n                from the FSF.\n            <\/p><p>\n                You can learn Scheme (and a lot of deep ideas about programming)\n                from Structure and Interpretation of Computer Programs by\n                Abelson and Sussman.  That book is now free\/libre although the\n                printed copies do not say so.\n                <\/p><p>Please <a href=\"https:\/\/stallman.org\/amazon.html\">don't buy books (or anything)\n                from Amazon!<\/a>\n        <\/p><\/li>\n        <li>\n            <p>\n                My favorite programming languages are Lisp and C.  However,\n                since around 1992 I have worked mainly on free software activism,\n                which means I am too busy to do much programming.  Around 2008 I\n                stopped doing programming projects.  As a result, I have not\n                had time or occasion to learn newer languages such as Perl,\n                Python, PHP or Ruby.\n            <\/p><p>\n                I read a book about Java, and found it an elegant further\n                development from C.  But I have never used it.  I did write some\n                code in Java once, but the code was in C and Lisp (I simply\n                happened to be in Java at the time ;-).\n        <\/p><\/li>\n        <li>\n            <p>\n                By contrast, I find C++ quite ugly.<\/p><p>The flaws of C++, as I recall from when I studied the matter\n                around 1990, include syntax and semantics.  As for syntax, its\n                grammar is ambiguous, and it is gratuitously incompatible with C,\n                which blocks the smooth upgrade path from C to C++.\n            <\/p><p>\n                As for semantics, the abstract object facility of C++ is\n                designed around the case where the real type of an object is\n                known at compile time.  However, in that case, abstract objects\n                are equivalent to a naming convention for functions to call.\n                The case where abstract objects add real power to a language is\n                when the type is not known until run time.  C++ does handle that,\n                but it seems to be an afterthought, a poor relation.\n            <\/p><p>\n                I suspect that I would find plenty of ugliness in the template\n                library, but I don't know.  That was added to C++ after I\n                studied it.\n        <\/p><\/li>\n        <li>\n            <p>\n                I skimmed documentation of Python after people told me it was\n                fundamentally similar to Lisp.  My conclusion is that that is\n                not so. `read', `eval', and `print' are all missing in Python.\n        <\/p><\/li>\n    <\/ul><h2>How to learn programming<\/h2><p>\n        First, read a textbook about programming in some language, then\n        manuals for several programming languages <a href=\"#lisp\">including\n        Lisp<\/a>. If this makes natural intuitive sense to you, that\n        indicates your mind is well-adapted towards programming.\n    <\/p><p>\n        If they don't make intuitive sense to you, I suggest you do something\n        other than programming. You might be able to do programming to some\n        degree with a struggle, but if you find it a struggle you won't be\n        very good at it.  What's the point of programming if it is a struggle\n        instead of a fascination?\n    <\/p><p>\n        After that, you need to read the source code of real programs (or\n        parts of them) and figure out what they do.  Then start writing\n        changes in them, to add features, or fix bugs if you can find out\n        about specific bugs to fix.  Ask some good programmers who are\n        familiar with the code of those programs to read and critique your\n        changes.\n    <\/p><p>\n        If you fix a bug in a free program that people are developing, the\n        developers are likely to be glad to get fixes from you and will tell\n        you the way to write them to make them good to install.  Look at their\n        TODO list for features you would like to implement.  You will find it\n        is a great satisfaction when the developers incorporate your changes.\n        <\/p><p>Do this over and over and you will become good at developing software.\n    <\/p><p>\n        Please use your programming capability only for good, not for evil.\n        Don't develop nonfree software, or <a href=\"https:\/\/gnu.org\/philosophy\/who-does-that-server-really-serve.html\">\n        service as a software substitute<\/a>.  Design systems not to collect\n        personal information, and to allow anonymous use.\n    <\/p><h2>Non-free software issues<\/h2><p>\n        I firmly refuse to install non-free software or tolerate its\n        installed presence on my computer or on computers set up for me.\n    <\/p><p>\n        However, if I am visiting somewhere and the machines available\n        nearby happen to contain non-free software, through no doing of\n        mine, I don't refuse to touch them.  I will use them briefly for\n        tasks such as browsing.  This limited usage doesn't give my assent\n        to the software's license, or make me responsible its being present\n        in the computer, or make me the possessor of a copy of it, so I\n        don't see an ethical obligation to refrain from this.  Of course, I\n        explain to the local people why they should migrate the machines to\n        free software, but I don't push them hard, because annoying them\n        is not the way to convince them.\n    <\/p><p>\n        Likewise, I don't need to worry about what software is in a kiosk,\n        pay phone, or ATM that I am using.  I hope their owners migrate them\n        to free software, for their sake, but there's no need for me to\n        refuse to touch them until then.  (I do consider what those machines\n        and their owners might do with my personal data, but that's a\n        different issue, which would arise just the same even if they did\n        use free software.  My response to that issue is to minimize those\n        activities which give them any data about me.)\n    <\/p><p>\n        That's my policy about using a machine once in a while.\n        If I were to use it for an hour every day, that would no longer\n        be \"once in a while\" \u2014 it would be regular use.\n        At that point, I would start to feel the heavy hand of\n        any nonfree software in that computer, and feel the duty\n        to arrange to use a liberated computer instead.\n    <\/p><p>\n        Likewise, if I were to ask or lead someone to set up a computer for\n        me to use, that would make me ethically responsible for its software\n        load.  In such a case I insist on free software, just as if the\n        machine were my own property.\n    <\/p><p>\n        As for microwave ovens and other appliances, if updating software is\n        not a normal part of use of the device, then it is not a computer.\n        In that case, I think the user need not take cognizance of whether\n        the device contains a processor and software, or is built some other\n        way.  However, if it has an \"update firmware\" button, that means\n        installing different software is a normal part of use, so it is a\n        computer.\n    <\/p><p>\n        Skype (or any nonfree noninteroperable communication program) is a\n        special case because of its network effect.  Using Skype to talk\n        with someone else who is using Skype is encouraging the other to use\n        nonfree software.  Doing so regularly is pressuring the other to use\n        nonfree software.  So I refuse to use Skype under any circumstances\n        (See <a href=\"https:\/\/stallman.org\/skype.html\">more information<\/a>.)\n    <\/p><h2>Streaming medias and DRM issues<\/h2><ul><li>\n            <p>\n                <a name=\"streaming\" id=\"streaming\">Streaming<\/a> media dis-services such as\n                Netflix and Spotify require nonfree client programs that impose\n                digital restrictions mechanisms (<a href=\"http:\/\/DefectiveByDesign.org\/\">DRM<\/a>) intended to stop\n                the user from saving a copy of the data being streamed through\n                her own computer.  You should never use DRM that you can't break,\n                so you should not use these dis-services unless you can break\n                their DRM.\n            <\/p><p>\n                An additional injustice of these and other streaming client\n                programs is that they impose unjust contracts (EULAs) which\n                restrict the user more strictly than copyright law itself.  I do\n                not agree to EULAs, period, and I urge you to join me in\n                rejecting them.\n            <\/p><p>\n                These streaming dis-services are malicious technology designed\n                to make people antisocial.  (If you don't have a copy, you can't\n                share copies.)  Rejecting them is of the highest ethical\n                priority.\n            <\/p><p>\n                A friend once asked me to watch a video with her that she was\n                going to display on her computer using Netflix.  I declined,\n                saying that Netflix was such a threat to freedom that I could\n                not treat it as anything but an enemy.\n            <\/p><p>\n                Out, out, damned Spotify!  Flick off, Netflix!\n        <\/p><\/li>\n        <li>\n            <p>\n                Every product with <a href=\"http:\/\/DefectiveByDesign.org\">Digital\n                Restrictions Management<\/a> (DRM) is an attack on your freedom.\n            <\/p><p>\n                Therefore, one should not buy or tolerate any product with DRM\n                handcuffs unless one personally possesses the means to break the\n                handcuffs.  For instance, don't use encrypted DVDs unless you\n                have DeCSS or another comparable free program.  And never use\n                a Bluray disk unless you find a way to break its handcuffs.\n                Don't use the <a href=\"https:\/\/stallman.org\/ebooks.pdf\"> Amazon Swindle<\/a>\n                or other e-book readers that trample readers' freedoms.\n                Don't use music or video streaming \"services\" that impose DRM.\n                (If they require a nonfree client program, it is probably for\n                DRM or some sort of surveillance of users.)\n        <\/p><\/li>\n    <\/ul><h2>Miscellaneous<\/h2><ul><li>\n            <p>\n                I never used Unix (not even for a minute) until after I decided\n                to develop a free replacement for it (<a href=\"http:\/\/www.gnu.org\/gnu\/the-gnu-project.html\">the GNU\n                system<\/a>).  I chose that design to follow because it was\n                portable and seemed fairly clean.  I was never a fan of Unix;\n                I had some criticisms of it too.  But it was ok overall as a\n                model.\n        <\/p><\/li>\n        <li>\n            <p>\n                In the mid 90s I had bad hand pain, so bad that most of the day\n                I could only type with one finger.  The FSF hired typists for me\n                part of the day, and part of the day I tolerated the pain.\n                After a few years I found out that this was due to the hard keys\n                of my keyboard.  I switched to a keyboard with lighter key\n                pressure and the problem mostly went away.\n            <\/p><p>\n                My problem was not carpal tunnel syndrome: I avoid that by keeping\n                my wrists pretty straight as I type.  There are several kinds of\n                hand injuries that can be caused by repetitive stress; don't\n                assume you have the one you heard of.\n        <\/p><\/li>\n        <li>\n            <p>\n                I find it bizarre that people use the term \"coding\" to mean\n                programming.  For decades, we used the word \"coding\" for the\n                work of low-level staff in a business programming team.\n                The designer would write a detailed flow chart, then the\n                \"coders\" would write code to implement the flow chart.  This is\n                quite different from what we did and do in the hacker\n                community -- with us, one person designs the program and writes\n                its code as a single activity.  When I developed GNU programs,\n                that was programming, but it was definitely not coding.\n            <\/p><p>\n                Since I don't think the recent fad for \"coding\" is an\n                improvement, I have decided not to adopt it.  I don't use the\n                term \"coding\", except if I were talking about a business\n                programming team which has coders.\n        <\/p><\/li>\n        <li>\n            Why I coined the name <a href=\"https:\/\/stallman.org\/articles\/posix.html\">POSIX<\/a>.\n        <\/li>\n    <\/ul><\/div><p>\n        Verbatim copying and distribution of this entire article is\n        permitted in any medium, provided this notice is preserved.\n    <\/p>","created_at":"2018-11-07T16:53:54+0000","updated_at":"2018-11-07T16:53:54+0000","annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"stallman.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":629,"title":"Noun Project - Icons for Everything","url":"https:\/\/thenounproject.com","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:53:54+0000","updated_at":"2018-11-07T16:53:54+0000","annotations":[],"reading_time":0,"domain_name":"thenounproject.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":626,"title":"What is the best software to take notes with while reading online?","url":"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online","content":"<div class=\"header\"><div id=\"wuKKl5oc1\"><div class=\"QuestionArea\"><div id=\"wiNyfgOo1\"><div id=\"wiNyfgOo3\"><div class=\"TopicList QuestionTopicHorizontalList\" id=\"__w2_wiNyfgOo4_topics_list\"><div id=\"wiNyfgOo15\"><div class=\"QuestionTopicHorizontalListItems QuestionTopicListItems TopicListItems2 SimpleToggle Toggle\" id=\"__w2_wiNyfgOo16__truncated\"><div class=\"QuestionTopicListItem TopicListItem topic_pill\" id=\"__w2_wiNyfgOo18_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wiNyfgOo25\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Bookmarks-15\" id=\"__w2_wiNyfgOo26_link\">Bookmarks<\/a><\/div><\/div><\/div><div class=\"QuestionTopicListItem TopicListItem topic_pill\" id=\"__w2_wiNyfgOo20_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wiNyfgOo27\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Notes\" id=\"__w2_wiNyfgOo28_link\">Notes<\/a><\/div><\/div><\/div><div class=\"QuestionTopicListItem TopicListItem topic_pill\" id=\"__w2_wiNyfgOo22_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wiNyfgOo29\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Startups\" id=\"__w2_wiNyfgOo30_link\">Startups<\/a><\/div><\/div><\/div><div class=\"QuestionTopicListItem TopicListItem topic_pill\" id=\"__w2_wiNyfgOo24_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wiNyfgOo31\"><a class=\"HoverMenu TopicNameLink topic_name\" href=\"https:\/\/www.quora.com\/topic\/Reading-1\" id=\"__w2_wiNyfgOo32_link\">Reading<\/a><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"wiNyfgOo5\"><div class=\"question_text_edit\"><h1>What is the best software to take notes with while reading online?<\/h1><\/div><\/div><div id=\"wiNyfgOo7\"><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wshqO0zc9_answer\"><div><div id=\"wshqO0zc10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wshqO0zc19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Dane-Rossenrode\" id=\"__w2_wshqO0zc32_name_link\">Dane Rossenrode<\/a>, works at Touchdreams<\/div><div class=\"u-inline-block\" id=\"wshqO0zc14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answer\/Dane-Rossenrode\" id=\"__w2_wshqO0zc40_link\">Answered 74w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wshqO0zc12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wshqO0zc12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wshqO0zc9_answer_content\"><div id=\"__w2_wshqO0zc13_expanded_content\"><div id=\"wshqO0zc16\"><div class=\"OriginallyAnsweredBanner\">Originally Answered: <a class=\"question_link\" href=\"https:\/\/www.quora.com\/What-is-the-best-website-for-taking-notes-online?no_redirect=1\" target=\"_top\" id=\"__w2_wshqO0zc35_link\">What is the best website for taking notes online?<\/a><\/div><\/div><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">Obviously it depends on what type of notes\u2026 for small, quick lists, <a href=\"http:\/\/keep.google.com\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">Google keep<\/a> works exceptionally well.<\/p><p class=\"ui_qtext_para\">For detailed notes with all sorts of insertions (drawings, numbered lists, pictures, etc.) <a href=\"https:\/\/evernote.com\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"evernote.com\">Evernote<\/a> or MS <a href=\"https:\/\/www.onenote.com\/hrd\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"onenote.com\">OneNote<\/a> are great.<\/p><p class=\"ui_qtext_para\">But for longer, detailed notes which only need to be written (no pictures or diagrams), one really stands out:<\/p><ul><li><a href=\"http:\/\/workflowy.com\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"workflowy.com\">WorkFlowy<\/a><\/li><\/ul><p class=\"ui_qtext_para\">It's incredibly fast and intuitive, and is basically an infinite list of lists. Go try it out!<\/p><\/div><\/div><div id=\"wshqO0zc24\"><div id=\"__w2_wshqO0zc13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wc4y4LoE9_answer\"><div><div id=\"wc4y4LoE10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wc4y4LoE19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Anshul-Jain-13\" id=\"__w2_wc4y4LoE32_name_link\">Anshul Jain<\/a>, MBA General Management, Indian Institute of Management, Ahmedabad (2012)<\/div><div class=\"u-inline-block\" id=\"wc4y4LoE14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answer\/Anshul-Jain-13\" id=\"__w2_wc4y4LoE37_link\">Updated 62w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wc4y4LoE12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wc4y4LoE12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wc4y4LoE9_answer_content\"><div id=\"__w2_wc4y4LoE13_expanded_content\"><div id=\"wc4y4LoE16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">I have used Evernote, Diigo, Pocket, OneNote. But increasingly, I have started \u201cwatching\u201d online rather than \u201creading\u201d online and I find that these tools fail to deliver on the new-age media.<\/p><p class=\"ui_qtext_para\">If you are also in the same boat, I recommend using ReClipped extension (<a href=\"https:\/\/chrome.google.com\/webstore\/detail\/reclipped-create-video-cl\/gbnebpdekafhpcipejfhabfghccgfnbh\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">ReClipped - create video clips instantly<\/a>) while studying from videos.<\/p><ol><li>While you are watching the video, ReClipped lets you create highlight reel on the sides, in no time.<\/li><li>You can write simple notes and tag them for easy cataloging of ideas<\/li><li>You can even share your notes with friends on any social media platform<\/li><li>Search your notes with keywords and tags<\/li><\/ol><p class=\"ui_qtext_para\">Revisiting most important parts in a video lecture is the key to greater retention and that is what ReClipped aspires to achieve.<\/p><p class=\"ui_qtext_para\">You would have noticed that creating highlight reel is not useful for just note taking, it can be a boon for cataloging inspirations and ideas from any video you watch.<\/p><p class=\"ui_qtext_para\">Here is a short video on how to create notes on videos with ReClipped extension.<\/p><div class=\"ui_qtext_embed thumbnail\" data-video-provider=\"youtube\" data-embed=\"&lt;iframe width=&quot;100%&quot; height=&quot;100%&quot; src=&quot;https:\/\/www.youtube.com\/embed\/EiAOfRqwXCY?wmode=opaque&amp;amp;autoplay=1&amp;amp;autohide=1&amp;amp;iv_load_policy=3&amp;amp;enablejsapi=1&quot; frameborder=&quot;0&quot; allow=&quot;autoplay; encrypted-media&quot; allowfullscreen&gt;&lt;\/iframe&gt;\" data-yt-id=\"EiAOfRqwXCY\"><\/div><\/div><div id=\"wc4y4LoE24\"><div id=\"__w2_wc4y4LoE13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wdjSGF5y9_answer\"><div><div id=\"wdjSGF5y10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wdjSGF5y19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Seb-Paquet\" id=\"__w2_wdjSGF5y33_name_link\">Seb Paquet<\/a>, Culture Hacker &amp; Research Developer at Element AI (2016-present)<\/div><div class=\"u-inline-block\" id=\"wdjSGF5y14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answer\/Seb-Paquet\" id=\"__w2_wdjSGF5y40_link\">Answered 428w ago<\/a> \u00b7 Author has 1.8k answers and 2.4m answer views<\/div><\/div><\/div><div id=\"__w2_wdjSGF5y12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wdjSGF5y12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wdjSGF5y9_answer_content\"><div id=\"__w2_wdjSGF5y13_expanded_content\"><div id=\"wdjSGF5y16\"><div class=\"OriginallyAnsweredBanner\">Originally Answered: <a class=\"question_link\" href=\"https:\/\/www.quora.com\/Whats-the-best-way-to-take-notes-on-a-webpage?no_redirect=1\" target=\"_top\" id=\"__w2_wdjSGF5y35_link\">What's the best way to take notes on a webpage?<\/a><\/div><\/div><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"> <a href=\"https:\/\/www.quora.com\/topic\/Diigo\">Diigo<\/a>.<\/div><\/div><div id=\"wdjSGF5y24\"><div id=\"__w2_wdjSGF5y13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wXG65I339_answer\"><div><div id=\"wXG65I3310\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wXG65I3319\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Ravi-Shankar-7\" id=\"__w2_wXG65I3332_name_link\">Ravi Shankar<\/a>, works at Infosys<\/div><div class=\"u-inline-block\" id=\"wXG65I3314\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answer\/Ravi-Shankar-7\" id=\"__w2_wXG65I3340_link\">Answered 226w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wXG65I3312_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wXG65I3312_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wXG65I339_answer_content\"><div id=\"__w2_wXG65I3313_expanded_content\"><div id=\"wXG65I3316\"><div class=\"OriginallyAnsweredBanner\">Originally Answered: <a class=\"question_link\" href=\"https:\/\/www.quora.com\/Whats-the-best-way-to-take-notes-on-a-webpage?no_redirect=1\" target=\"_top\" id=\"__w2_wXG65I3337_link\">What's the best way to take notes on a webpage?<\/a><\/div><\/div><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\">There are many ways to take notes on a webpage. I usually use, for bookmarks type <a href=\"https:\/\/www.google.com\/bookmarks\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">Google Bookmarks<\/a>. For short Note taking <a href=\"https:\/\/keep.google.com\/keep\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">Google Keep<\/a> and for a paragraph <a href=\"https:\/\/evernote.com\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"evernote.com\">Remember everything with Evernote, Skitch and our other great apps.<\/a> and Microsoft Onenote in my office. But office fails in sync with its own android mobile app. Evernote is better in sync but its used for only long annotations. I also use delicious for bookmarking but not very useful as its a old bookmarking tool. Right now there is no collaborative right tool in market with all the combined features. Also for note taking there is no proper improvements stand for so long period. Always we need to depend on google products due to its long standing in market.<\/div><\/div><div id=\"wXG65I3324\"><div id=\"__w2_wXG65I3313_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wQVppiHs9_answer\"><div><div id=\"wQVppiHs10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wQVppiHs19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\">Anonymous<\/div><div class=\"u-inline-block\" id=\"wQVppiHs14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answers\/94846770\" id=\"__w2_wQVppiHs33_link\">Answered 12w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wQVppiHs12_follow_button_section\"><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wQVppiHs9_answer_content\"><div id=\"__w2_wQVppiHs13_expanded_content\"><div id=\"wQVppiHs16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">Being fresh out of college, I know how important it is to use laptops and other technology in school. No one can deny the speed at which you can take notes, compared to writing by hand.<\/p><p class=\"ui_qtext_para\">You control what you want<a href=\"https:\/\/go-rewind.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"go-rewind.com\"> GO Rewind<\/a> to remember and what it should ignore. Your Data is stored on our dedicated ISO 27001 and HIPAA certified private cloud servers.<\/p><\/div><\/div><div id=\"wQVppiHs24\"><div id=\"__w2_wQVppiHs13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wxa4Qc8i54_answer\"><div><div id=\"wxa4Qc8i55\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wxa4Qc8i64\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Ivana-Stefanoska-1\" id=\"__w2_wxa4Qc8i78_name_link\">Ivana Stefanoska<\/a>, Owner (2015-present)<\/div><div class=\"u-inline-block\" id=\"wxa4Qc8i59\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answer\/Ivana-Stefanoska-1\" id=\"__w2_wxa4Qc8i82_link\">Answered 12w ago<\/a> \u00b7 Author has 746 answers and 258.5k answer views<\/div><\/div><\/div><div id=\"__w2_wxa4Qc8i57_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wxa4Qc8i57_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wxa4Qc8i54_answer_content\"><div id=\"__w2_wxa4Qc8i58_expanded_content\"><div id=\"wxa4Qc8i61\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">Work, even life, is all about productivity, and those who work long hours each day are bound to run into situations where things simply fall out of your head for no apparent reason at all. Whether we get stressed, or there is too much information to go around, forgetting things can add more to the stressful experience, especially if it is something important, or an idea that we would have loved to explore more. The only solution towards keeping your mind somewhat free and organized is to start practicing taking notes.<\/p><p class=\"ui_qtext_para\">Whatever you do\/see, <b>GO Rewind <\/b>saves it and allows you to rewind to every moment in time on your screen. You\u2019ll never have to fear losing anything again.<\/p><\/div><\/div><div id=\"wxa4Qc8i69\"><div id=\"__w2_wxa4Qc8i58_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wJJpxfcD9_answer\"><div><div id=\"wJJpxfcD10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wJJpxfcD19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\">Anonymous<\/div><div class=\"u-inline-block\" id=\"wJJpxfcD14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answers\/94841572\" id=\"__w2_wJJpxfcD33_link\">Answered 12w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wJJpxfcD12_follow_button_section\"><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wJJpxfcD9_answer_content\"><div id=\"__w2_wJJpxfcD13_expanded_content\"><div id=\"wJJpxfcD16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">Since most campuses allow laptops in classrooms these days, you may find that typing notes during lecture is a faster way to capture notes. And in a way that you can read them later! I am using <a href=\"https:\/\/go-rewind.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"go-rewind.com\">Go Rewind<\/a> tool is great data collector, so you don\u2019t have to worry if you miss something.<\/p><\/div><\/div><div id=\"wJJpxfcD24\"><div id=\"__w2_wJJpxfcD13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_w7bkRXGI9_answer\"><div><div id=\"w7bkRXGI10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"w7bkRXGI19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Suresh-Maidaragi-1\" id=\"__w2_w7bkRXGI33_name_link\">Suresh Maidaragi<\/a>, Senior Software Engineer (2014-present)<\/div><div class=\"u-inline-block\" id=\"w7bkRXGI14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/What-is-the-best-software-to-take-notes-with-while-reading-online\/answer\/Suresh-Maidaragi-1\" id=\"__w2_w7bkRXGI37_link\">Answered 28w ago<\/a> \u00b7 Author has 90 answers and 28.8k answer views<\/div><\/div><\/div><div id=\"__w2_w7bkRXGI12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_w7bkRXGI12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_w7bkRXGI9_answer_content\"><div id=\"__w2_w7bkRXGI13_expanded_content\"><div id=\"w7bkRXGI16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--regular\"><div class=\"ui_qtext_expanded\"><div class=\"ui_qtext_image_outer\"><div class=\"ui_qtext_image_wrapper\"><img class=\"landscape ui_qtext_image zoomable_in zoomable_in_feed\" src=\"https:\/\/qph.fs.quoracdn.net\/main-qimg-58edfaa140144e2378d4f16bd194f4e3\" alt=\"image\" \/><\/div><\/div><p class=\"ui_qtext_para\"><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=in.notepad\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"google.com\">Notes offline (Voice note, OCR, Bar code scanner) - Apps on Google Play<\/a><\/p><p class=\"ui_qtext_para\">One of the best app with all latest features like<\/p><ol><li>Voice notes, voice to text\/text to voice<\/li><li>Speak out<\/li><li>Drawing\/Handwriting<\/li><li>Change the note style as like microsoft document<\/li><li>Share via other apps<\/li><li>Scan the text from piece of paper (OCR scanner)<\/li><li>Scan product barcode\/QR code<\/li><\/ol><p class=\"ui_qtext_para\">All this features offline with\/without internet<\/p><\/div><\/div><div id=\"w7bkRXGI24\"><div id=\"__w2_w7bkRXGI13_survey_wrapper\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:53:53+0000","updated_at":"2018-11-07T16:53:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.quora.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":625,"title":"Selling My Bootstrapped SaaS Business \u2013 Indie Hackers","url":"https:\/\/www.indiehackers.com\/@tylertringas\/selling-my-bootstrapped-saas-business-26c8406e1d","content":"<div class=\"medium-insert-images medium-insert-images-full\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/1N9ZPzh8xNWZ3qGHXaPBWv2kDIJ3\/@tylertringas%7C26c8406e1d\/map-desk-2.jpg\" alt=\"\" class=\"\" \/><\/figure><\/div><p class=\"\">Five years ago I built and launched the first version of a SaaS app on a single flight from San Francisco to Buenos Aires. Slowly and steadily, <a href=\"https:\/\/www.storemapper.co\/\" target=\"_blank\" rel=\"noopener noreferrer\">Storemapper<\/a>\u00a0grew into a healthy location-independent business for one person and then later a small dedicated remote team. At parties, I would describe it as, \u201cnot a startup; a healthy growing internet small business.\u201d This year, almost exactly five years after launching, I sold the business for what, to someone growing up middle class in Florida, is a life-changing amount of money that will enable all kinds of exciting new projects and adventures. From start to finish, it has been an exciting ride. With the sale concluded, I wanted to share as much as I could about the process of building a business that can be sold and how I sold it.<\/p><p>There\u2019s always a risk that these posts turn into a 5,000-word humblebrag. But I really do think it\u2019s worth a read because, unlike most business acquisition stories, which often feel like an out of the blue stroke of good luck, the way that I sold Storemapper feels very replicable for other entrepreneurs. When I spoke to someone two years ago about what it would look like if I ever sold the business I would say, \u201cI\u2019m not trying to sell it now, but if I ever did it would probably look this\u2026\u201d And six months ago I would tell a few folks privately, \u201cI think that one of the people I met recently might be the one to buy Storemapper and if they do it will probably go like this\u2026\u201d And, then basically when it all went down it looked more or less like\u2026 that. There wasn\u2019t some single huge stroke of good luck, though of course, I got lucky in the little ways that every successful business has to. An excellent outcome, but also a perfectly reasonable and achievable one that I think can serve as something of a template for other bootstrapped entrepreneurs.<\/p><p>This is a long and detailed post. If you find it useful or interesting, you will probably like my other posts on <a href=\"http:\/\/microsaas.co\" target=\"_blank\" rel=\"noopener noreferrer\">building Micro-SaaS businesses<\/a>.<\/p><h2>First the obvious: why sell your software business?<\/h2><p>The past two years of my life have been pretty amazing due in no small part to owning my own SaaS business. Two years ago, my girlfriend and I set off on a year backpacking around the world: driving the coasts of Croatia, South Africa, and New Zealand; going on safari in the Serengeti and practicing yoga in Bali; diving in Fiji and soaking up the Mediterranean lifestyle in an apartment in Barcelona. Throughout the process, thanks primarily to an awesome remote team, Storemapper was able to keep running and growing even as I took long stretches of time off the grid (hiking Mt. Kilimanjaro or rock climbing in El Potrero Chico). Last year we moved back to Washington, DC and I felt like I was living the dream: being able to afford to live in a world-class city while still maintaining complete control over my time.<\/p><p>I did warn you this might come out humble-braggy. But it\u2019s important to highlight that none of the normal pressures to sell a business were present. I wasn\u2019t burned out, on the contrary life was pretty good and running Storemapper only took about 10 hours per week of my time and attention. I wasn\u2019t crunched for cash; Storemapper was profitable and I was saving away the bulk of the profits each month. So in light of that, why sell the business? Here was my thought process.<\/p><h3>Running the numbers<\/h3><p>From a purely financial perspective, owning your own stable, profitable SaaS business is incredible. Unlike the wild seasonal swings of running an e-commerce store or the feast-or-famine revenues of selling books or info-products, once you have a reasonably large SaaS customer base you tend to have predictable revenues, stable costs and a low risk of any of that changing dramatically month to month. Say the business is already profitable and in one month you add $1,000 in new monthly subscribers. You keep that $1,000. You can have a reasonable expectation that next month you will also get most of that $1,000 and so on for the foreseeable future. On top of that, the value of your SaaS business is a multiple of that number so that extra $1,000 in subscriptions may have added $36,000 or more to the total value of your business. If your business is solid and predictable and you \u201crun the numbers\u201d on selling <em>this<\/em> month versus six months from now you will typically arrive at the conclusion that you could keep the business, pocket the six months of extra profit <em>and then<\/em> sell the business for <em>more<\/em> than you could this month. Essentially the very positive math of operating SaaS will never lead you to conclude that you should sell your business just yet. A spreadsheet will almost always tell you to hold the SaaS and reconsider selling it in six months and it will generally be correct.<\/p><p>Unless of course, it all blows up somewhere in those six months.<\/p><p>And that\u2019s the key consideration. When you estimate \u201chow will my SaaS business be doing in 6-12 months\u201d the answer is very bi-modal. There\u2019s a very very large chance that it\u2019s doing a bit better and a very small but non-zero chance the whole thing blows up in some unforeseen way. At least that was my estimate at the time and that\u2019s why a key financial decision for me to sell the business was diversification.<\/p><h3>Diversification and peace of mind<\/h3><p>As an entrepreneur, the risks and rewards of your financial future are extremely concentrated. In my case, as Storemapper succeeded I had slowly wound down freelancing and other sources of income so that I was 100% reliant on Storemapper revenues for income. There wasn\u2019t a lot to complain about there. Business was good and I was still saving much of what I made from Storemapper, which is more than I can say for the last paycheck I ever had. But I sat down and did a detailed visualization of what it would be like for some <em>black swan<\/em> event to force me to completely shut down Storemapper. Even though it seemed extremely unlikely I found that my loss aversion was pretty high. I\u2019d been working on Storemapper for five years and I really wanted to lock in a solid win. One of the most important skills for a poker player is knowing when to end a big run and take chips off the table.<\/p><p>Leaving aside a \u201cstrategic acquisition offer\u201d for ludicrous amounts of money (more on that later), most of the time the financial spreadsheet of selling your business doesn\u2019t add up to a convincing case to sell. Almost everyone will end up selling for non-financial reasons like getting burned out or bored, badly needing the money or a million other reasons other than \u201cI did the math.\u201d<\/p><p>Diversification felt like one of the better reasons to sell a business.<\/p><p>I felt at the time that the benefits of diversification, reducing my overall risk, outweighed the opportunity cost of the potential money I could earn by continuing to run the business.<\/p><p>But okay, this is getting a bit boring\u2026<\/p><h3>Momentum for other projects<\/h3><p>Diversification was honestly just the line of thinking that allowed me to rationalize selling Storemapper. The real driver was largely non-financial. I have often pointed out that Storemapper was a means to an end, not necessarily a passion project. I have always said that I did not want my legacy to be the King of Store Locators. Building Storemapper was a way for me to learn a ton about building and running a business and a way to achieve financial freedom. That\u2019s not meant as an insult by the way. Crafting a great product that works well and makes customers happy can be extremely fulfilling. But I felt it was important, to be honest with myself about what my motivations were and over time, once Storemapper became financially successful, those motivations changed.<\/p><p>As soon as Storemapper reached a point of providing me a full-time income I began optimizing to reduce my time requirements, rather than trying to maximize revenue and growth. I started looking for another project a bit further up the hierarchy of needs. A little over a year ago I found it and started working part-time with the team at <a href=\"https:\/\/www.maptia.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Maptia<\/a>. After working with the team for nearly a year, while still running Storemapper, <a href=\"https:\/\/tylertringas.com\/joining-maptia\/\" target=\"_blank\" rel=\"noopener noreferrer\">I knew<\/a> Maptia was where I wanted to really put my focus and energy. Even though Storemapper was as close to a Four Hour Work Week business as I think is really feasible, it still takes up quite a bit of your background thought processes and mental bandwidth. I never seriously considered selling Storemapper until I had some next thing that I knew I wanted to throw myself into. I think Derek Sivers calls this the <a href=\"https:\/\/sivers.org\/tarzan\" target=\"_blank\" rel=\"noopener noreferrer\">Tarzan approach<\/a> of waiting until you grab hold of the next \u201cvine\u201d before you let go of the last one. I\u2019m not sure I would have gone through with the deal if I didn\u2019t have a clue what the next thing would be, but the prospect of really building momentum for projects that I wanted to invest more time and energy into was the main reason for selling the business.<\/p><p>That\u2019s why earlier this year, after five years of telling people flatly \u201cNo\u201d when they asked if I would sell Storemapper, I started to become open to the idea. Ultimately the prospect of shifting my whole focus and energy into new projects was what motivated me to close the deal.<\/p><p>I\u2019m also finding that the increased bandwidth, and of course the cash, from selling the business has built momentum for other projects like <a href=\"https:\/\/tylertringas.com\" target=\"_blank\" rel=\"noopener noreferrer\">my blog<\/a>, angel investing, which I will probably dabble in once I formulate a real strategy for it, and a few other ideas kicking around. Having a finite chunk of money and no income, rather than an open-ended income stream, has been a surprisingly good kick in the butt to start some more ambitious projects.<\/p><h2>So, let\u2019s talk about the actual process<\/h2><p>A lot of people wanted to know about the actual nuts and bolts of selling a business so here are a few FAQs. The answers are, um, a little on the dry side so if you\u2019re not fascinated by the nitty-gritty of how to execute a sale, skip to the next section.<\/p><p><strong>How did you find the buyer, or did they find you?<\/strong><\/p><p>Being <a href=\"https:\/\/tylertringas.com\/storemapper\/\" target=\"_blank\" rel=\"noopener noreferrer\">transparent<\/a> and making the business visible by blogging, tweeting and podcasting, brought a number of potential buyers to my inbox. I would estimate at least 50-75 people have emailed me over five years. A good chunk emailed in response to a single post on the NYC Ruby Meetup group, entitled \u201c<a href=\"https:\/\/www.meetup.com\/NYC-rb\/messages\/57549452\/\" target=\"_blank\" rel=\"noopener noreferrer\">UnSexy, Profitable SaaS business for sale or partner<\/a>,\u201d where I briefly considered selling the business about a year after it launched. I\u2019ve been amazed at how years later that single post continued to generate random inbound interest. Initially, I took quite a number of calls, primarily out of curiosity and a huge number of those inquiries were complete wastes of time. There is a large pool of buyers out there scanning the web looking for distressed or under-valued properties trying to pick up things for a steal. If I wanted to sell Storemapper for 6x <em>monthly<\/em> profits I could have had 20 offers in a day. But most inquiries were not serious buyers willing to pay a reasonable multiple for a profitable SaaS business. A few offers during that stretch were legitimate and I almost considered selling, but then I ran the numbers and decided to keep the business. Good thing too since the business grew by more than 20x in the next four years.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/tylertringas.com\/wp-content\/uploads\/2017\/09\/Screenshot-2017-09-29-11.02.02-768x733.png\" alt=\"\" srcset=\"https:\/\/tylertringas.com\/wp-content\/uploads\/2017\/09\/Screenshot-2017-09-29-11.02.02-768x733.png 768w, https:\/\/tylertringas.com\/wp-content\/uploads\/2017\/09\/Screenshot-2017-09-29-11.02.02-300x286.png 300w, https:\/\/tylertringas.com\/wp-content\/uploads\/2017\/09\/Screenshot-2017-09-29-11.02.02-1024x977.png 1024w, https:\/\/tylertringas.com\/wp-content\/uploads\/2017\/09\/Screenshot-2017-09-29-11.02.02.png 1582w\" \/><\/figure><\/div><p>I kid you not this post, by uh \u201cuser 1.\u201d, brought in at least 30 pretty serious inquiries.<\/p><p>In March of this year, I went to my first <a href=\"http:\/\/www.microconf.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">MicroConf<\/a>, a conference dedicated to bootstrapped software businesses, which is where I met Kevin from <a href=\"https:\/\/www.sureswiftcapital.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">SureSwift Capital<\/a>. In the follow-up to MicroConf I had discussions with three serious potential buyers and, somewhat by coincidence, I was connected with two more potential buyers around the same time. These were either individuals that I knew I had sold businesses before or who had raised a fund dedicated to buying businesses like mine. Serious buyers, not just someone poking around looking for something on the cheap. Over the subsequent months, I made a concerted effort to schedule calls and keep each one of the discussions moving forward at roughly the same pace. Two of the discussions pretty quickly concluded as not viable but the other three progressed to either a verbal offer or a letter of intent (more on that below) and one became the deal with SureSwift that concluded about five months after initially meeting Kevin at MicroConf.<\/p><p><strong>How did the initial discussion go? What did you disclose?<\/strong><\/p><p>While none of them lead to a deal, taking those calls from early inquiries in 2013 was helpful to me in understanding the kinds of questions a potential buyer might ask and help me to think through what I would and wouldn\u2019t like to disclose on an initial call. Some people might play this process very close their chest, asking a potential buyer to sign an NDA and only disclosing the bare minimum information. I had already taken the path of disclosing a huge amount of Storemapper\u2019s financial info, even going so far as to post a live public dashboard with figures like revenue growth, churn and customer lifetime value. So I didn\u2019t really have much to hide. This simplified the process substantially. In an initial call, I was able to summarize the publicly available info and give an overview of our costs and profitability (which were not public) which is more than enough for a buyer to make an initial assessment of whether they are seriously interested and even to discuss preliminary pricing for a deal.<\/p><p><strong>How did you discuss pricing?<\/strong><\/p><p>If you ask around at conferences or find the right blog posts you\u2019ll learn about a metric called Seller Discretionary Earnings (SDE) or Seller Discretionary Income (SDI). I first heard these terms from Thomas Smale, who runs a brokerage focusing on bootstrapped businesses like mine. Essentially it is the profitability of the business, but adding back in things like the owner\u2019s salary and various other expenses that make sense (from a tax perspective) to charge to the business but that wouldn\u2019t be a cost the new owner would be obliged to pay. Most discussions of pricing are framed as a combination of two components: the SDE and the multiple.<\/p><p>Simple example: the business has a top line revenue of $200k. After operating costs like employees, your email marketing software and other SaaS you use for the business and servers, you\u2019re left with $90k. You add another $20k of expenses to your business like your cell phone bill, your annual Macbook Pro upgrade and so on, then you pay yourself the remaining $70k as a salary. Your SDE would be the $90k and a reasonable valuation of the business would be some multiple of that $90k.<\/p><p>So the default approach for discussing pricing is to agree on the business\u2019s SDE and then haggle over the multiple. This approach seems sensible, and it probably works well if you are working with a broker with a very rigorous and well-tested methodology for what does and doesn\u2019t count for SDE, but I really did not like it as a communication tool. The problem is that is can still be very difficult to pin down all the edges cases that affect SDE. For example, an SDE is an annual number, spanning 12 months\u2026 but which 12 months? Is it the previous (or \u201ctrailing\u201d) 12 months? But a fast-growing business that added 50% of its customers in the last quarter and a totally flat-lined business could have the same trailing twelve months SDE, while the former would clearly be much more valuable. What if a deal takes 6 months to close, which is not uncommon, does the SDE update every month automatically or is it frozen at some point in negotiations? Why not use a forward-looking 12 months? When fellow founders ask \u201cwhat\u2019s the MRR of your business?\u201d we are usually referring to the current level of subscriptions, but current MRR isn\u2019t something that shows up in the historical bookkeeping; you have to calculate it across all your current subscribers. If you\u2019re looking forward do you factor in churn? What about expected growth? Is it a 12-month forecast (what you really expect to happen) or a 12-month pro forma (current customers minus churn plus expansion MRR)? What expenses really are excluded from SDE? Does that $6,000 marketing experiment you ran 11 months ago and then abandoned count against you? Should you wait 2 months until it gets pushed out of the trailing 12 months metric? And so on, and so on\u2026<\/p><p>Much like when you are pitching a client on how many \u201chours\u201d a project will take, the multiple is just a polite proxy for what both parties are really thinking about: the actual amount of money being exchanged. Once I started taking the idea of selling the business very seriously I explained why I felt the \u201cmultiple\u201d discussion was a problematic proxy and just moved to speaking in whole dollar amounts. If we close in the next X months, my price is Y. I recommend this approach.<\/p><p>At the end of the day, the valuation of a business is where the buyer\u2019s willingness to pay and the seller\u2019s willingness to sell overlap. The SDE + Multiple framing is just a way to attempt to simplify communication, but I think just speaking in whole dollar amounts is more direct and sensible.<\/p><p><strong>So how does the process actually kick off?<\/strong><\/p><p>In my case, I had a few preliminary calls with buyers. I was very transparent about the business since I already had nothing to hide. In the initial calls, we didn\u2019t explicitly discuss a sale price, although we did allude to a range of multiples just to make sure we were on the same general page about deal size.<\/p><p>For my part, I was also interviewing each potential buyer about the non-monetary factors. Most importantly, what did they plan to do with the business and would there be opportunities for my team? But also making sure they seemed technically competent so that the handover wouldn\u2019t become a giant headache. I asked a few of them for references that I could speak to from people they had worked with or other companies they had acquired. It\u2019s a perfectly reasonable request and is immensely helpful.<\/p><p>Once you generally feel that this could be a serious buyer, the next step is for the buyer to prepare a <em>letter of intent<\/em> (LOI). An LOI is a legal document with no legal weight whatsoever\u2026 or at least that is my perspective, I am not a lawyer. The goal of it is for the buyer to start to put some details in writing about what they are willing to offer for the business and to lay out the terms of how you will approach <em>due diligence<\/em>. Due diligence is the process where the buyer really starts to dig into the details of the company and it will vary enormously from buyer to buyer. In the LOI you will typically describe the timeline for due diligence and the buyer may ask for a period of exclusivity, say 30-60 days, in which they will conduct due diligence and you will agree not to talk to other buyers.<\/p><p><em>You do not have to agree to exclusivity.<\/em> As soon as you get an LOI from a serious buyer you can and should immediately begin looking for another buyer. Having a strong backup buyer is the absolute best way to ensure you get a good deal and have a strong negotiating position all the way to the end.<\/p><p>After the LOI and through due diligence you are still mostly operating on good faith. Nobody has any legal obligation to buy or sell and both parties can still walk away or radically change the deal terms. But once due diligence is over, the next step is the important one. After due diligence is when you will start preparing the final documents that are actually legally binding. In Storemapper\u2019s case, it was an <em>asset purchase agreement<\/em> where I agreed to sell all of the operating assets to SureSwift Capital. Tax and legal considerations will determine whether the buyer wants to acquire the entire corporate entity (LLC, corporation, etc) or just the assets, but an asset sale seems to be by far the most common. So in my case after the sale, I was left still owning the empty shell of an LLC and bank accounts associated with Storemapper. Which is totally fine.<\/p><p>This document is the legally binding one where you want to make sure you get all the tax and legal advice you need before signing. Then you sign it.<\/p><p><strong>How do you actually conclude the sale?<\/strong><\/p><p>This is by far the most common question I\u2019ve gotten but it has a pretty simple answer. The asset purchase agreement will more or less lay out the terms. You will use an <em>escrow agent<\/em>, a law firm or dedicated escrow firm. Both the buyer and seller sign a simple document with the escrow agent who then accepts a wire transfer from the buyer. The escrow agent then confirms to you that they have the money in their account and you start a pre-defined process of handing over control of the assets of the business. Once things like the domain, the servers, and various passwords are handed over, the buyer has a short period of time to review them and make sure you haven\u2019t been completely misleading them. Then they give the escrow agent the thumbs up, and the escrow agent releases the money to you.<\/p><p>Then you have sold your business.<\/p><p>In our case, this process was incredibly smooth for two reasons. First, SureSwift has acquired a number of businesses already and provided a very comprehensive list of things to prepare for the day of, much of which I would not have thought of on my own. Second, at Storemapper we often use platforms rather than hand-crafting our own solutions. For example, handing over control of the app was a matter of two clicks on Heroku, but it would be a different story if Storemapper was running on some artisanal private server co-located with every other software project I\u2019d ever started.<\/p><p><strong>Did you use a broker?<\/strong><\/p><p>In the end, I did not. I know several entrepreneurs who used brokers and recommend it highly. I have met a few brokers in this space who seem very smart and who I really considered working with. Here\u2019s the main reason why you might not want to use a broker and why I ultimately didn\u2019t. I was an opportunistic seller, not a motivated one.<\/p><p>Brokers will argue that they are totally aligned with your interest because their fee is a percentage of the deal size, so they are incentivized to get you the best deal possible. On a single deal basis that is true, but on an organizational level, a brokerage is not incentivized to get each client the absolute best possible deal they can. They are incentivized to create maximum liquidity in the marketplace. As an organization, they will earn far more if every deal takes half as long, so they do twice as many deals than if they take an unlimited amount of time and get every client a (let\u2019s say hypothetically) 30% better offer. As a seller that 30% could be a massive difference but for the broker it\u2019s better to just get more deals done. I\u2019m not saying any of them would be deceptive or deny this, just that brokers are much better suited for motivated sellers who definitely want to get the business sold in a reasonable timeframe. I\u2019ll go into this later, but my whole strategy for selling the business was to put myself in a position where I <em>didn\u2019t have<\/em> to sell; where I would be perfectly happy to run the business myself for another year or five years if the deal wasn\u2019t exactly what I wanted. This is not a good client for a broker and a good broker will tell you as much.<\/p><p>If you want a reasonable deal on a reasonable timeline, get a broker. If you want a phenomenal deal, you may <em>still<\/em> want a broker if you feel like having someone walk you through the process is worth the fee. But I did not use one.<\/p><p><strong>So do you just walk away as soon as the deal closes?<\/strong><\/p><p>No, almost all deals like this have a transition period where you are still involved with the business and slowly handover all the responsibilities to the buyer\u2019s team. 3-6 months seems to be pretty normal. In my case, the first 30 days or so after the sale were exactly the same as the previous 30 days, where I was still managing the day to day as I had before.<\/p><h2>How I chose SureSwift<\/h2><p>I had never heard of SureSwift Capital until I met Kevin at MicroConf in Las Vegas, but they ended up being the ideal acquirer for Storemapper. I\u2019ll walk through the reasons I chose them, not because I\u2019m contractually obligated to compliment them now, but because I think the reasoning, which goes beyond just the price of the deal, is pretty broadly applicable.<\/p><ul><li>A very important feature, in some regards, simply because it\u2019s somewhat rare in the bootstrapped or Micro-SaaS market, is that SureSwift is a long-term player. They have raised a fund to go out and acquire lots of bootstrapped profitable businesses like mine. They had closed on quite a few deals before buying Storemapper and I knew they intended to close on many more after. This means they have reputational skin in the game. Throughout the process I could talk to founders they have acquired businesses from before and I knew that if they tried to completely screw me over, then they wouldn\u2019t be able to acquire more businesses in the future. This is a huge point of differentiation from a solo buyer, using their own money, who might only acquire one business per decade and not be that concerned about their reputation. SureSwift needs to maintain an impeccable reputation in the market or else they would not be able to deploy the capital they raised. As a first time seller there is a constant feeling of being at an informational disadvantage and having someone with long-term skin in the game really helps balance that fear.<\/li><li>It was a great opportunity for the Storemapper team. A big part of Storemapper\u2019s enterprise value was that the small and awesome team mostly ran it entirely on their own. It was important to me that after an acquisition they would have lots of opportunities and also not be pressured into making drastic changes. SureSwift offered exactly that. They manage a portfolio of software businesses and were equally happy to have the employees stay exactly in their current roles at Storemapper or to expand their roles across the portfolio. It\u2019s hard to overstate how important those assurances have been for a smooth closing and transition.<\/li><li>They had a competent team in place to accept the business. Throughout the acquisition process I also made sure to counter-interview the buyers to try to find out how well they would likely manage the handover. Were they used to working with highly remote\/distributed teams? Did they have people in place who understood the tech stack we used? Even if they made a higher offer I really was not interested in selling Storemapper to an organization with a CTO who only codes in .NET and who wants to get everybody on the phone every other day for a two hour check-in.<\/li><li>Lastly, they had similar priorities on what was and wasn\u2019t important in the transaction. A classic move from a potential buyer, that I\u2019ve experienced in other businesses, is to make all kinds of assurances about how interested they are very early on in the discussion and then at the very end of the deal process the lawyers and accountants dig in with a million nit-picky irrelevant quibbles; then they come back with a ridiculous low ball offer. This was <em>not<\/em> how SureSwift operated. I more or less asked this point blank and then verified by speaking to other founders that had been through the process with them, that they really did focus on the key important metrics and didn\u2019t sweat the small stuff.<\/li><\/ul><h2>What made Storemapper sell-able<\/h2><p>Kevin from SureSwift is probably better equipped to answer precisely why it made sense for them to buy Storemapper, but I think it\u2019s worth my taking a stab at why I think it was a sell-able business.<\/p><p>There are many things that are required, being necessary but not sufficient, of a sell-able business but only one thing makes a buyer want to buy your business: they need to see an opportunity. Entrepreneurs sometimes adopt a strange language when discussing acquisitions that seems to assume that acquirers are just out there as a natural part of the world, and if they create a \u201cgood business\u201d someone <em>should<\/em> want to buy it. But acquirers of all kinds are not simply walking around looking to pay full market price to take over good businesses. They need to see an opportunity. Either they need to feel they are getting a good deal on the business by paying below what it\u2019s worth, or they need to see a pathway to the business performing much better under their ownership. The first one, deal hunters or vultures, are by far the dominant mentality of potential acquirers because it\u2019s the easiest and simplest way to make sure an acquisition is a good investment. As a potential seller you want to sniff out quickly if someone is just poking around looking for distressed assets and move. The best reason for a business to be sell-able is for the buyer to see an opportunity for the business to flourish under their ownership and be a worthwhile investment even when paying full market value for the business; a true win-win. Here are a few examples of good reasons for a buyer to want to buy your business:<\/p><ul><li><strong>Budget dynamics<\/strong>: there is a well known budget dynamic in large SaaS businesses that they often lose money when they grow quickly. If you spend $50 to get a customer for your $40 ebook, you\u2019re doing it wrong. But it\u2019s perfectly rational to spend $50 to acquire a customer who subscribes to your $10\/month SaaS plan, who then stays paying for 24 months. But in this case you actually lose money in the short-term, the faster you grow the faster you rack up losses even when your business is doing great. A small bootstrapped startup might not have the budget to sustain this but a well-funded acquirer might be perfectly happy to plow money into customer acquisition, growing the enterprise value of the business quickly, even if it means short-term losses.<\/li><li><strong>Customer synergies:<\/strong> an acquirer may have a product that is complementary to an overlapping customer base. They may be able to cross-sell their product to your customers and vice versa, reducing their customer acquisition costs and creating up-sell opportunities. Cross-selling and up-selling are immensely more efficient than finding completely new customers so an acquirer may see buying your business as a far more effective use of funds than burning through the same money on marketing.<\/li><li><strong>Cost synergies:<\/strong> less fun than the others but still common. An acquirer may have a team or overhead that allows them to run your product for much less than you currently do. This doesn\u2019t always have to be the \u201cchop shop\u201d strategy where they would fire some of your employees to cut costs. For example Storemapper never got very involved in paid Facebook or Adwords ads primarily because we just weren\u2019t big enough to justify having a full time person dedicated to managing it. A larger acquirer might be able to amortize the cost of a smart marketer across a larger business.<\/li><li><strong>Core competencies:<\/strong> businesses as they grow often require different core competencies. A very early stage business needs to be able to connect deeply with customers\u2019 needs and iterate quickly to achieve product-market fit. A more mature business may no longer need to iterate quickly but rather scale up its existing activities. These can be completely different skillsets that different owners excel at, which can make a business sell-able under the right circumstances.<\/li><\/ul><p>Any one of these, or hopefully several plus others I haven\u2019t mentioned, creates the seed of an opportunity to buy\/sell a business. However there are a many other factors that, in combination with an opportunity, make a business sell-able. Ultimately it is the buyers\u2019 opinions on these that matter, but here is my take on the key factors to making your business sell-able.<\/p><p><strong>An awesome team:<\/strong> first and foremost is building a team that can keep the business growing without you. Too often founders want to insinuate themselves into every part of the business, making the entire structure dependent on their engaged focus. Instead you want to build a team and a product that are self-sufficient without you. A buyer can\u2019t possibly learn exactly how active you are day-to-day before a sale, so a great team gives them confidence that the whole thing won\u2019t collapse after you walk away.<\/p><p><strong>Predictable, diversified revenue base:<\/strong> the reason for being of a business is to generate revenue but not all revenue is created equal when it comes to buying and selling a business. A buyer is buying your revenue streams and different revenue streams have different degrees of risk to a buyer. Obviously customers with recurring subscriptions are more valuable than one-off purchasers. But also customers that have been paying for 30 months continuously are more valuable than a customer who subscribed last month. A customer base made up of a diverse group of \u201cboring\u201d businesses is better than a customer base composed entirely of fidget spinner manufacturers. $10,000 per month composed of 1,000 companies paying $10 is more robust than two companies paying $5,000. The more predictable and diversified the revenue base, the more sell-able the business.<\/p><p><strong>Documentation and processes:<\/strong> basically a business needs to have its shit together to be sell-able. Buyers can to some extent be choosy and they can and should pass on buying a great business that will be a nightmare to run. Having thorough documentation and good internal processes may not really increase the value of your business per se, but not having them can be a deal-breaker to an acquirer. You want them to be impressed by how painless it will be to take over operating the business.<\/p><p>This is an incomplete list but I believe some combination of all of these factors lead to Storemapper being a very sell-able business. It had an awesome team who had mostly been with the company for more than two years and who managed the day to day business without me having to be involved. The examples of me going completely offline for 10 days in Tanzania and Fiji really drove this home. It had great documentation and processes and a solid predictable, customer base with recurring low-churn revenue. It was also very easy to see how a buyer with different skillsets could invest time, money or even just more focus to really increase the already growing revenue.<\/p><h2>Lessons on how to sell a bootstrapped SaaS business<\/h2><p>The book <a href=\"https:\/\/www.amazon.com\/Built-Sell-Creating-Business-Without\/dp\/1591845823\/ref=sr_1_1?ie=UTF8&amp;qid=1506625929&amp;sr=8-1&amp;keywords=built+to+sell\" target=\"_blank\" rel=\"noopener noreferrer\">Built to Sell<\/a> is an essential read for every entrepreneur. It lays out clearly and simply some of the key counter-intuitive features of selling a business versus running a business. Here are some of the most salient lessons from that book and elsewhere.<\/p><p><strong>Run your business to sell it, before you want to sell it<\/strong><\/p><p>I can\u2019t emphasize this enough. If you think you will ever want to sell your business, the time to start preparing is now. Don\u2019t wait until you are actually ready to sell it. Making a business sell-able requires systematically making small or big decisions differently and optimizing for certain qualities in your business that can\u2019t be bolted on last minute or even over the course of a few months. Your business will be sold on a track record, every selling point about your business will be greatly enhanced by being able to point to 48 consecutive months of that same thing. Through some combination of books and blog posts I was fortunate enough to internalize this lesson early on and acted accordingly. It turns out that creating a sell-able business is also very similar to running a lifestyle business, focusing on taking yourself out of the day to day and optimizing to recapture your time and freedom. The process of removing yourself from a business is also a big part of making that business sell-able.<\/p><p><strong>Focus on financial, not strategic acquirers<\/strong><\/p><p>First let\u2019s differentiate between the two broad categories of potential buyers of a business: strategic and financial acquirers. A strategic acquisition is the kind you hear most about in the news: Facebook buying Instagram, Google buying Nest. Strategic acquisitions can be driven by a variety of motivations such as buying a competitor before they grow too large or a proprietary piece of intellectual property. The reasons can be varied but the common thread is that the actual finances of the business are a secondary concern. In a strategic acquisition, businesses can be purchased<\/p><p>That\u2019s the definitional difference but here is the key difference between the two kinds of acquisitions: a strategic acquisition is mostly about getting lucky whereas a financial acquisition can be engineered by a founder. In both instances you have to run a great company, but beyond that there isn\u2019t much you can do to increase your odds of a bolt from the blue strategic acquisition offer for 25x times your revenue. But you can structure your business in such a way that it is highly attractive to financial acquirers and greatly increase your odds of a successful transaction. Learn what makes a successful financial acquisition from this post and elsewhere. Get a sense of what the very high end of a financial acquisition might look like and shoot for that.<\/p><p><strong>Acquisition offers are a hidden benefit of radical transparency<\/strong><\/p><p>I joined the radical transparency train after being inspired by and learning a ton from the founders of Buffer, Baremetrics, Keen, Patrick McKenzie and many more. I felt at the time that being transparent even with the financial metrics of the business added an appropriate level of credibility and was helpful to readers. Over time, as copycat apps started popping up, I began to question whether being so transparent was really a good idea, although they never made a material dent in the business. However being public about the numbers behind Storemapper certainly increased the number of acquisition inquiries I received over the years. People knew the revenue and growth and could immediately tell whether Storemapper was roughly in the range of businesses they were looking for. Between that and reading my transparent blogging about the strategy and challenges, we only needed to briefly discuss the costs and profitability before they would have enough information to credibly make an initial offer. This hugely accelerated the speed at which I could have acquisition discussions. Having multiple buyers involved at the same time really helped me feel confident to press forward with the deal. In the end transparency definitely was a big net benefit.<\/p><p><strong>Go to conferences<\/strong><\/p><p>This one is pretty self-explanatory but despite all the blogging and sporadic inbound interest in Storemapper, I only reached a critical mass of serious acquisition offers after attending my first Micro-Conf in Las Vegas.<\/p><h2>How to get the best deal possible<\/h2><p>Obviously I have only sold one business, so this isn\u2019t some proven methodology for selling businesses that I\u2019ve honed over years. But, I\u2019m synthesizing advise I got from other founders and that I felt generally worked well for me.<\/p><p><strong>Be an opportunistic seller<\/strong><\/p><p>The absolute most important factor in getting a great deal in a sale is not <em>having<\/em> to sell. All of the work here happens well before the sale. Craft your business in a way that you would be perfectly happy to run it indefinitely. Get your work\/life balance in order and your stress level under control. Go into a sale process with the idea that if you don\u2019t get exactly the offer you want, you are 100% willing to just wait it out and run your business for another year. Running yourself into the ground, getting burned out or bored of your business, or falling into a cash crunch, or any other situation where you really <em>need<\/em> to sell the business is an excellent way to get lured into accepting a low-ball offer.<\/p><p><strong>Be aggressive and ready to walk<\/strong><\/p><p>Not having to sell allows you adopt an aggressive \u2018take it or leave it\u2019 attitude to a sale negotiation. You want to be extremely polite, nobody wants to buy a business from a jerk, but continually remind the buyer, \u201clook if we\u2019re not on the same page here, no big deal, let\u2019s just save each other some time and move on.\u201d Smart buyers can sense if you are a really motivated seller and can use all kinds of tactics, waiting until the last minute of negotiations when you have already invested a lot of time and energy into the deal, to try to gain last minute leverage. The sale process can be long and grueling. With every hour you spend working on a deal, your brain starts to become more and more committed to getting the deal done. You start envisioning your life post-sale before it closes. But, that is a sunk cost fallacy. Throughout the process I repeatedly made myself sit down and visualize the process of walking away from the whole deal if necessary and that gave me the confidence to stand my ground, from actually turning down a pretty big offer to holding firm on several small points, to get the best deal I think was reasonably possible.<\/p><p><strong>Have multiple offers in parallel<\/strong><\/p><p>Being aggressive and ready to walk away from one particular negotiation is made immensely easier if you have multiple offers on the table. It\u2019s important to make an effort to develop those offers at roughly the same time. From the time you meet a buyer to the time you have a firm offer may be many months. When you\u2019re in late-stage negotiations, having another firm vaguely interested in talking about an acquisition is not nearly as useful as having three LOIs in the same week. You can view each thread with a buyer as a conversion funnel and try to keep as many conversations moving forward in parallel.<\/p><p><strong>Purposefully under-optimize<\/strong><\/p><p>The first few recommendation are all about the seller\u2019s mindset, but it\u2019s important not to neglect the buyer\u2019s psychology. One thing I would recommend is to purposefully under-optimize certain aspects of your business. A business that is fully optimized in every aspect, with the juice wrung out of every possible avenue of improvement, is not actually attractive to a buyer. They want to see areas where they can come in and bring their own skills and resources to quickly accelerate the business. For example, I really do not understand paid search advertising (Google Adwords, Facebook, etc) so I purposefully put almost no effort into it as a customer acquisition channel for Storemapper. I would highlight that fact to buyers saying, \u201cDespite the fact that revenue is growing well, I have put almost no effort in paid ads, if you have an Adwords guru on staff, I\u2019m sure they can fire up a profitable paid advertising funnel in no time to further increase growth.\u201d<\/p><p><strong>Build a narrative for the buyer<\/strong><\/p><p>At the end of the day buyers need to tell themselves a story about acquiring your business. If they have raised outside funding and have partners or a board, they need to be able to tell that story to defend their investment to their partners. If they are a solo buyer they still need to tell that story to their spouse or to themselves. Almost everything in this post was part of an explicit narrative that I built and gave to potential buyers. For me it looked something like this:<\/p><p>\u201cI\u2019ve been running this business for five years. It\u2019s profitable, growing and has product-market fit. I have a great team who handle most of the day to day operations so I\u2019m perfectly willing to run it indefinitely. However, because I\u2019ve run this with a lifestyle business approach, certain aspects are very under-optimized and its still performing well below its full potential. The right buyer, who I, and more importantly my team, would like working with, could pay a comparatively high price for the business and still generate a very strong return on investment.\u201d<\/p><p>Build your own narrative and sell that story to the buyer.<\/p><h2>Conclusion<\/h2><p>Wow, thanks for hanging in there all the way to the end. If you have any follow-up questions, fire them at me <a href=\"https:\/\/twitter.com\/tylertringas\" target=\"_blank\" rel=\"noopener noreferrer\">on Twitter<\/a>. If you are little earlier on in the Micro-SaaS process, feel free to check out my ebook <a href=\"https:\/\/tylertringas.com\/micro-saas-ebook\/\" target=\"_blank\" rel=\"noopener noreferrer\">Building Micro-SaaS Businesses<\/a>.<\/p>","created_at":"2018-11-07T16:53:52+0000","updated_at":"2018-11-07T16:53:52+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":38,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/1N9ZPzh8xNWZ3qGHXaPBWv2kDIJ3\/@tylertringas%7C26c8406e1d\/map-desk-2.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":620,"title":"How do Third-Party \"tracking cookies\" work?","url":"https:\/\/stackoverflow.com\/questions\/13897472\/how-do-third-party-tracking-cookies-work","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        86&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I have read this question here: \n<a href=\"https:\/\/stackoverflow.com\/questions\/8794981\/how-do-internet-advertisers-use-third-party-cookies\">How Do Internet Advertisers Use Third-Party Cookies?<\/a>\non how third-party tracking cookies work, but am still very confused.\nI don't understand how if I visit Website A (a normal website with ads) how Website B (an advertising website) can assign my computer an ID, and then figure out that I was on website A, and other websites after it that have its ads.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/13897472\/revisions\" title=\"show all edits to this post\">edited May 23 '17 at 11:55<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Dec 16 '12 at 0:27&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1395137\/user093203920\">User093203920<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,09252048&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"1\">&#13;\n                                1 Answer&#13;\n                                1&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        192&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>First, cookies are set and retrieved through HTTP headers. If your browser sends a request to <a href=\"http:\/\/example.com\">http:\/\/example.com<\/a>, then the response might come back with a header that says <code>Set-Cookie: foo=bar<\/code>. Your browser stores this cookie, and on any subsequent requests to <a href=\"http:\/\/example.com\">http:\/\/example.com<\/a>, your browser will send <code>foo=bar<\/code> in the <code>Cookie<\/code> header. (Or at least until the cookie expires or is deleted.) The browser sends the <code>foo=bar<\/code> cookie with <em>any<\/em> request to <a href=\"http:\/\/example.com\">http:\/\/example.com<\/a>, regardless of who initiated the request or what the context is. If <a href=\"http:\/\/example2.com\">http:\/\/example2.com<\/a> contains the tag <code>&lt;img src=\"http:\/\/example.com\/img.jpg\"&gt;<\/code>, then  the browser will send the cookie <code>foo=bar<\/code> when it fetches <a href=\"http:\/\/example.com\/img.jpg\">http:\/\/example.com\/img.jpg<\/a>, even though <a href=\"http:\/\/example2.com\">http:\/\/example2.com<\/a> is responsible for the request being sent.<\/p>\n<p>So, if website A contains an ad that is served by website B, then website B can set a cookie in your browser. For example, maybe website A uses <code>&lt;iframe src=\"http:\/\/websiteB.com\/ad.html&gt;&lt;\/iframe&gt;<\/code> to serve the ad from website B. Then when your browser goes to fetch <code>http:\/\/websiteB.com\/ad.html<\/code>, the response will come back with a <code>Set-Cookie<\/code> header that sets a cookie with some unique random string. If website C also includes an ad from website B, then that unique cookie will be sent when the ad on website C is fetched from website B.<\/p>\n<p>As far as how website B knows which actual website you're visiting, there are a variety of ways. In some cases, when the browser sends a request to one website, it tells the website which website you're coming from. So when the browser goes to fetch <code>http:\/\/websiteB.com\/ad.html<\/code>, it might include the HTTP header <code>Referer: http:\/\/websiteA.com<\/code> that tells website B that the request was initiated by website A. Every time website B sees the unique random string that it assigned to you, it can check the Referer header to add to its log of where you've been. If website A is cooperating with website B, A can just directly tell B that you're coming from website A. For example, website A could include the ad from website B by using <code>&lt;iframe src=\"http:\/\/websiteB.com\/ad.html?referer=websiteA.com\"&gt;<\/code>, and then website B will see the referer in the query string.<\/p>\n<p>Does that help? Are there particular parts of the answer you linked that don't make sense to you?<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Dec 16 '12 at 1:55&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1867669\/emily\">Emily<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            4,45011412&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-11-07T16:53:49+0000","updated_at":"2018-11-07T16:53:49+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":621,"title":"Keep your Slack distractions under control with Emacs","url":"http:\/\/endlessparentheses.com\/keep-your-slack-distractions-under-control-with-emacs.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:53:49+0000","updated_at":"2018-11-07T16:53:49+0000","annotations":[],"reading_time":0,"domain_name":"endlessparentheses.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":622,"title":"monome","url":"https:\/\/monome.org\/","content":"<iframe src=\"https:\/\/player.vimeo.com\/video\/266741634?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>changes. travels. is open to possibilities.<\/p><p>norns is many sound instruments. it connects to grids, MIDI and other objects. norns lets you define its behavior with scripts and DSP.<\/p><p>past. present. future. norns is many tomorrows.<\/p><ul class=\"detail-list\"><li>linux-powered open-source sound computer<\/li>\n<li>2 audio in, 2 audio out, headphone<\/li>\n<li>4 USB device ports: grids, MIDI, WIFI, etc<\/li>\n<li>bright white pixels, endless rotary encoders, key switches<\/li>\n<li>milled aluminum assembly, hard anodized<\/li>\n<li>internal battery, charge by USB<\/li>\n<li>6.2\" x 4.0\" x 0.8\"<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/128746859?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>an open, interactive instrument that allows the user to imagine and define its function.<\/p><p>tactile and visual feedback. decoupled by design. clear response and intuitive play. when you push a key the input data is sent to the computer (be it norns, ansible, or normal laptop). the computer considers this, then perhaps sends back some light data, or changes a sound. use is determined by the chosen application running on the computer: sequencer, sample cutter, tone map, polyrhythm machine are just a few of the many possibilities, and we\u2019ve put together studies to help you create your own.<\/p><ul class=\"detail-list\"><li>16 x 8 backlit keys<\/li>\n<li>bright warm white light<\/li>\n<li>two-tone aluminum assembly<\/li>\n<li>usb powered<\/li>\n<li>10.2\" x 5.4\" x 0.6\"<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/182119406?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>rotational controls encircled by dense segmented lighting, completely open for definition.<\/p><p>like the grid, tactile and visual feedback. but instead of a multitude of binary points, the arc has four incredibly precise continuous controls with visual indication to match. connects to computers (norns, ansible, laptops) to become a sequencer, physics simulation, granular instrument, or anything you can create yourself.<\/p><ul class=\"detail-list\"><li>four high resolution optical encoders<\/li>\n<li>bright warm white light<\/li>\n<li>two-tone aluminum assembly<\/li>\n<li>usb powered<\/li>\n<li>10.2\" x 2.7\" x 0.6\"<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/129271731?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>algorithmic ecosystem. teletype is a dynamic, musical event triggering platform. edit scripts with a USB keyboard which plugs into the front panel. syntax is simple and easy to learn with the provided tutorials, video, and reference pages. teletype runs without the keyboard attached: you might treat editing as precomposition, or leave the keyboard plugged in for live coding.<\/p><ul class=\"detail-list\"><li>format: eurorack<\/li>\n<li>width: 18hp<\/li>\n<li>depth: 26mm (\u201cskiff friendly\u201d)<\/li>\n<li>power: 12v: 72ma, -12v: 12ma, 5v: none<\/li>\n<\/ul><\/div><iframe src=\"https:\/\/player.vimeo.com\/video\/214221486?color=ffffff&amp;title=0&amp;byline=0&amp;portrait=0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><div class=\"col-container\"><p>far communicator, speaks openly.<\/p><p>ansible connect various USB devices (grid, arc, MIDI) to the modular environment. several modes provide different sequencers and control approaches.<\/p><ul class=\"detail-list\"><li>format: eurorack<\/li>\n<li>width: 6hp<\/li>\n<li>depth: 39mm (\u201cskiff friendly\u201d)<\/li>\n<li>power: 12v: 200ma, -12v: 19ma, 5v: none<\/li>\n<\/ul><\/div>","created_at":"2018-11-07T16:53:49+0000","updated_at":"2018-11-07T16:53:49+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"monome.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":623,"title":"GPG (Pete's notes)","url":"https:\/\/nets.ucar.edu\/nets\/intro\/staff\/siemsen\/tools\/gpg.html","content":"GPG (Pete's notes)\n<ul><li><a href=\"#gen\">General<\/a><\/li>\n  <li><a href=\"#gmail\">GMail<\/a><\/li>\n  <li><a href=\"#resources\">Resources<\/a><\/li>\n  <li><a href=\"#installGPGTools\">Install GPGTools<\/a><\/li>\n  <li><a href=\"#mailing\">Mailing lists for GPGTools problems (Lighthouse)<\/a><\/li>\n  <li><a href=\"#config\">Configure GPG<\/a><\/li>\n  <li><a href=\"#generate\">Generate PGP keys<\/a><\/li>\n  <li><a href=\"#keyservers\">Register public key with key servers<\/a><\/li>\n  <li><a href=\"#ucar\">Register public key with UCAR Security<\/a><\/li>\n  <li><a href=\"#keys\">Managing keys (keyrings and the keychain)<\/a>\n  <ul><li><a href=\"#mancl\">Managing keys with command lines<\/a><\/li>\n    <li><a href=\"#mangka\">Managing keys with GPG_Keychain_Access<\/a><\/li>\n    <li><a href=\"#mans\">Managing keys with Seahorse<\/a><\/li>\n    <li><a href=\"#pubring\">Updating UCAR keys in pubring<\/a><\/li>\n  <\/ul><\/li>\n<li><a href=\"#store\">Store PGP keys on a flashdrive<\/a>\n<ul><li><a href=\"#partition\">Erase and partition the flash drive<\/a><\/li>\n  <li><a href=\"#diskimage\">Copy the gnupg directory to the flash drive<\/a><\/li>\n  <li><a href=\"#ssh\">Copy the SSH id_rsa file to the flash drive<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#gpg-agent\">gpg-agent<\/a><\/li>\n<li><a href=\"#abkey\">ABKey: MacCPC integration with Address Book<\/a><\/li>\n<li><a href=\"#trouble\">Troubleshooting<\/a><\/li>\n<\/ul><h3><a name=\"gen\" id=\"gen\">General<\/a><\/h3>\n<p> UCAR security policy requires that all sysadmins be\ncapable of reading and writing email messages that are\nencrypted with PGP.  GPG is the Gnu (free) version of\nPGP.  On the Mac, you want to install GPGTools in\norder to get the command-line \"gpg\" command.<\/p>\n<p>Note: GPGMail is an add-on for Apple Mail. It is a component of\nGPGTools. I use Gmail to read mail, so I don't use GPGMail.  <\/p>\n<h3><a name=\"gmail\" id=\"gmail\">GMail<\/a><\/h3>\nWith GMail, there seem to be two ways to deal with PGP\nmessages:\n<ol><li>Mailvelope<\/li>\n  <li>Use the Unix command-line to run gpg<\/li>\n<\/ol><p>I had some success with Mailvelope, but it only handles\nmail messages, not files outside of a mail reader, so I\nsettled on gpg.<\/p>\n<h4><a name=\"decrypt\" id=\"decrypt\">decrypt with gpg<\/a><\/h4>\n<p>With GMail, to decrypt a message, do this:\n<\/p><ol><li>Make sure you have the keyring mounted, otherwise insert\n  the USB drive and mount the gnupg virtual drive<\/li>\n  <li>In GMail, do \"show\u00a0original\"<\/li>\n  <li>Highlight the PGP part (including the \"BEGIN\u00a0PGP\u00a0MESSAGE\" and \"END\u00a0PGP\u00a0MESSAGE\"\n  lines) and copy it with Command-C<\/li>\n  <li>Go to a Terminal window<\/li>\n  <li>\u00a0<p>cat &gt;ttt.txt<\/p><\/li>\n  <li>Paste the PGP part into the file<\/li>\n  <li>Control-D<\/li>\n  <li>\u00a0<p>gpg2 --decrypt ttt.txt<\/p><\/li>\n<\/ol><h4><a name=\"encrypt\" id=\"encrypt\">encrypt with gpg<\/a><\/h4>\n<p>With GMail, to decrypt a message, do this:\n<\/p><ol><li>Go to a Terminal window<\/li>\n  <li>\u00a0<p>gpg -ae<\/p><\/li>\n  <li>When it prompts, enter the recipient's email address<\/li>\n  <li>When it prompts, enter \"y\" to confirm that you want to use the key<\/li>\n  <li>When it prompts, enter the text of the message you want\n  to send<\/li>\n  <li>It'll output a PGP part. Cut\/paste it, (including the\n  \"BEGIN\u00a0PGP\u00a0MESSAGE\" and\n  \"END\u00a0PGP\u00a0MESSAGE\" lines) into the GMail \"compose\" window<\/li>\n<\/ol><h3><a name=\"resources\" id=\"resources\">Resources<\/a><\/h3>\n<p>\nThere are several resources for information about PGP and GPG.  Here\nare the ones I liked best:\n<\/p>\n<ul><li><a href=\"https:\/\/wiki.ucar.edu\/display\/sysadmin\/PGP+for+UCAR+Sysadmins\">PGP for UCAR Sysadmins<\/a><\/li>\n  <li><a href=\"https:\/\/wiki.ucar.edu\/display\/sysadmin\/Keyserver+Use\">Keyserver Use<\/a><\/li>\n  <li><a href=\"https:\/\/wiki.ucar.edu\/display\/sysadmin\/Computer+Security+Master+Key\">Computer Security Master Key<\/a><\/li>\n<\/ul><h3><a name=\"installGPGTools\" id=\"installGPGTools\">Install GPGTools<\/a><\/h3>\n<p>\nAs of 2013-01, install GPGTools by downloading the installer from\n<a href=\"http:\/\/www.gpgtools.org\/\">http:\/\/www.gpgtools.org\/<\/a>.\n<\/p>\nClicking on GPGTools.mpkg will take you through the\nsteps.  It installs\n<ul><li><code><strong>\/usr\/local\/bin\/gpg2<\/strong>, and links\n  to it so \"gpg\" is the same as \"gpg2\".<\/code><\/li>\n  <li><code><strong>\/Applications\/GPG Keychain Access.app<\/strong><\/code><\/li>\n  <li><code><strong>~\/Library\/Mail\/Bundles\/GPGMail.mailbundle<\/strong><\/code><\/li>\n  <li>several services (See\n  System Preferences-&gt;Keyboard-&gt;Keyboard Shortcuts-&gt;Services-&gt;Text\n  and look for ones named \"OpenPGP\")<\/li>\n  <li><code><strong>\/Library\/LaunchAgents\/org.gpgtools.macgpg2.gpg-agent.plist<\/strong><\/code>,\n  which causes launchd to start gpg-agent as a daemon at\n  login time.  See <a href=\"#gpg-agent\">Managing my secret key with gpg-agent<\/a>\n  <\/li>\n  <li>\/usr\/local\/MacGPG2\/share\/man<\/li>\n<\/ul>\nWhen it's done installing, you should be able to do\n<p>gpg --version<\/p>\n...and get 2.0.27 or later.\n<h3><a name=\"mailing\" id=\"mailing\">Mailing lists for GPGTools problems\n(Lighthouse)<\/a><\/h3>\n<p>I have an account at Lighthouse that lets me access the mailing\nlists for the GPGTools.  See 1Password.<\/p>\n<h3><a name=\"config\" id=\"config\">Configure GPG<\/a><\/h3>\n<p>\nEdit the <strong>~\/.gnupg\/gpg.conf<\/strong> file.  Set:\n<\/p><p>default-key E4BA9BEC\nkeyserver hkp:\/\/wwwkeys.pgp.net\nkeyserver-options auto-key-retrieve no-include-revoked<\/p>\n<h3><a name=\"generate\" id=\"generate\">Generate PGP keys<\/a><\/h3>\nCreate your key files in ~\/gnupg.\n<p>cisl-irving$ <strong>gpg2 --gen-key<\/strong>\n  gpg (GnuPG) 1.4.3; Copyright (C) 2006 Free Software Foundation, Inc.\n  This program comes with ABSOLUTELY NO WARRANTY.\n  This is free software, and you are welcome to redistribute it\n  under certain conditions. See the file COPYING for details.\n  Please select what kind of key you want:\n  (1) DSA and Elgamal (default)\n  (2) DSA (sign only)\n  (5) RSA (sign only)\n  Your selection? <strong>1<\/strong>\n  DSA keypair will have 1024 bits.\n  ELG-E keys may be between 1024 and 4096 bits long.\n  What keysize do you want? (2048)\n  Requested keysize is 2048 bits\n  Please specify how long the key should be valid.\n  0 = key does not expire\n  &lt;n&gt;  = key expires in n days\n  &lt;n&gt;w = key expires in n weeks\n  &lt;n&gt;m = key expires in n months\n  &lt;n&gt;y = key expires in n years\n  Key is valid for? (0) <strong>0<\/strong>\n  Key does not expire at all\n  Is this correct? (y\/N) <strong>y<\/strong>\n  You need a user ID to identify your key; the software constructs the user ID\n  from the Real Name, Comment and Email Address in this form:\n  \"Heinrich Heine (Der Dichter) &lt;heinrichh@duesseldorf.de&gt;\"\n  Real name: <strong>Pete Siemsen<\/strong>\n  Email address: <strong>siemsen@ucar.edu<\/strong>\n  Comment:\n  You selected this USER-ID:\n  \"Pete Siemsen &lt;siemsen@ucar.edu&gt;\"\n  Change (N)ame, (C)omment, (E)mail or (O)kay\/(Q)uit? <strong>o<\/strong>\n  You need a Passphrase to protect your secret key.\n  We need to generate a lot of random bytes. It is a good idea to perform\n  some other action (type on the keyboard, move the mouse, utilize the\n  disks) during the prime generation; this gives the random number\n  generator a better chance to gain enough entropy.\n  +++++++++++++++++++++++++.+++++++++++++++++++++++++++++++++++\n  We need to generate a lot of random bytes. It is a good idea to perform\n  some other action (type on the keyboard, move the mouse, utilize the\n  disks) during the prime generation; this gives the random number\ngenerator a better chance to gain enough entropy.\n+++++++++++++++.+++++.+++++..++++b+.+++++++++++++++b+++++k..x\ngpg: \/Users\/siemsen\/.gnupg\/trustdb.gpg: trustdb created\ngpg: key E4BA9BEC marked as ultimately trusted\npublic and secret key created and signed.\ngpg: checking the trustdb\ngpg: 3 marginal(s) needed, 1 complete(s) needed, PGP trust model\ngpg: depth: 0  valid:   1  signed:   0  trust: 0-, 0q, 0n, 0m, 0f, 1u\npub   1024D\/E4BA9BEC 2006-05-23\nKey fingerprint = D311 1402 4026 21AF 2F81  D861 4172 CF6A E4BA 9BEC\nuid                  Pete Siemsen &lt;siemsen@ucar.edu&gt;\nsub   2048g\/331EA98A 2006-05-23\ncisl-irving$<\/p>\n<p>\nVerify your keys with\n<\/p><p>gpg --fingerprint<\/p>\n<p>\nGenerate a public key with\n<\/p><p>gpg --export --armor siemsen@ucar.edu &gt; ~\/.gnupg\/my-key.asc<\/p>\n<h3><a name=\"keyservers\" id=\"keyservers\">Register public key with key servers<\/a><\/h3>\n<p>\nTo allow other users to get your public key, you want to register it\nwith keyservers.\n<\/p>\n<p>\nIf\/when the UCAR hkp server is up, register my public key with the UCAR keyserver:\n<\/p><p>gpg --send-keys --keyserver hkp:\/\/keyserver.ucar.edu E4BA9BEC<\/p>\n<p>\nRegister my public key with the public keyservers:\n<\/p><p>gpg --send-keys --keyserver pgp.mit.edu E4BA9BEC<\/p>\n<h3><a name=\"ucar\" id=\"ucar\">Register public key with UCAR Security<\/a><\/h3>\n<p>\nFirst, register your public key with the UCAR keyserver as described\nabove.  Then get a paper copy of the form and fill in the fields with\nthe key ID and fingerprint displayed by this command:\n<\/p><p>gpg --fingerprint<\/p>\n<p>\nHint: the fingerprint is 10 blocks of 4 hexadecimal characters, and\nthe key ID is the last 8 characters of the fingerprint.\n<\/p>\n<p> Once accepted, the UCAR security guys will sign the key\nthat you stored in the UCAR keyserver, which verifies\nthat it is valid.  Then they'll send you a test email\nmessage that is encrypted.  You'll have to read the\nmessage and respond to it to demonstrate that you can\ndo PGP email.  <\/p>\n<h3><a name=\"keys\" id=\"keys\">Managing keys (keyrings and the keychain)<\/a><\/h3>\n<p>Each person has a public key.  To make your public key available so\nothers can send you encrypted mail, you should publicize your public\nkey on a keyserver, and\/or put it on a webpage somewhere.  To read\nsigned or encrypted mail sent to you, you need to have the sender's\npublic key.  To ease the problem of accessing a lot of other people's\npublic keys, gpg maintains a \"public\u00a0keyring\" in\n~\/.gnupg\/pubring.gpg.  You store people's public keys in your public\nkeyring as you learn them, and the mail reader uses the keys to\ndecrypt incoming mail and\/or to verify signatures in incoming mail\nmessages.<\/p>\n<p>The GPG keyring is different than the Mac \"keychain\", even though\nthe Mac keychain can store PGP keys.  It seems that people don't use\nthe Mac keychain to store public keys - they use the GPG keyring\ninstead.  As I understand it, there's no benefit to storing public\nkeys in the keychain.<\/p>\n<p>The GPG keyring is stored in ~\/.gnupg\/pubring.gpg.  I need a copy\nof it on each machine that I readmail on.<\/p>\n<p>There is also a secret GPG keyring, named secring.gpg, in which you\nstore your secret key.  It is needed when you want to sign or encrypt\nan outgoing mail message.  UCAR policy says to store your secret key\non an encrypted disk image on a flash drive, so I do that.  To\nminimize the mounting\/unmounting of that drive, I copy my secret to\ngpg-agent when I log in, using a bash function I wrote named \"ssho\".\nThe details of this are described in the <a href=\"#gpg-agent\">gpg-agent section<\/a>.<\/p>\n<p> To facilitate sharing keys easily, people can register their keys\nin a public keyserver.  I use one of two keyservers:\n<strong>keyserver.ucar.edu<\/strong> or <strong>pgp.mit.edu<\/strong>.\nThe first is the keyserver maintained by the UCAR security guys, and\nholds the keys of all the UCAR sysadmins.  The second is the MIT one\nthat is public.  <\/p>\n<p> GPG can be configured to look up keys on one keyserver\nautomatically.  GPG won't forward on key lookups - it only talks to\none keyserver.  I configured GPG to look them up from the UCAR\nkeyserver.  <\/p>\n<p> Keys that you retrieve from a keyserver might be bullshit, so\nafter you retrieve a key you have to assign it a \"validity\" or\n\"trustworthiness\" level.  You only have to import a person's public\nkey and then trust it once - the trust level you assign is stored with\nthe key in your public keyring.  <\/p>\n<p> Every once in a while, you'll want to update the GPG keyring, aka\n~\/.gnupg\/pubring.gpg so that you can verify signed messages.  As of\n2016-08-02, I receive signed messages from these people: <\/p>\n<ul><li>Colin Randall<\/li>\n  <li>Alex Hsia<\/li>\n  <li>Kaustubh Gadkari<\/li>\n  <li>Rich Johnson<\/li>\n<\/ul><p>As of 2014-04-07, I have their keys in my GPG keyring, so I can verify\ntheir signed messages.  If I receive a signed message for someone that\nI don't have in my keyring, here's how to add them:<\/p>\n<p>first use the person's email address to find the their key ID:\n<\/p><p>cisl-irving$ <strong>gpg2 --search-keys --keyserver hkp:\/\/wwwkeys.pgp.net Alex_Hsia@noaa.gov<\/strong>\n  gpg: searching for \"Alex?Hsia@noaa.gov\" from hkp server wwwkeys.pgp.net\n  (1)     Alex Hsia &lt;Alex.Hsia@noaa.gov&gt;\n  1024 bit DSA key 80C14108, created: 2003-03-27\n  (2)     Alex Hsia &lt;alex@boulder.noaa.gov&gt;\n  1024 bit DSA key 1D663A0A, created: 2003-03-27\n  Keys 1-2 of 2 for \"Alex?Hsia@noaa.gov\".  Enter number(s), N)ext, or Q)uit &gt; q\n  cisl-irving$<\/p>\n  <p> Then, use the their key ID to load their public key into the\n  keyring:\n  <\/p><p>cisl-irving$ <strong>gpg2 --recv-keys --keyserver hkp:\/\/wwwkeys.pgp.net 0x80C14108<\/strong>\n    gpg: requesting key 80C14108 from hkp server wwwkeys.pgp.net\n    gpg: key 80C14108: public key \"Alex Hsia &lt;Alex.Hsia@noaa.gov&gt;\" imported\n    gpg: 3 marginal(s) needed, 1 complete(s) needed, PGP trust model\n    gpg: depth: 0  valid:   1  signed:   1  trust: 0-, 0q, 0n, 0m, 0f, 1u\n    gpg: depth: 1  valid:   1  signed:   0  trust: 0-, 0q, 0n, 0m, 1f, 0u\n    gpg: next trustdb check due at 2011-01-22\n    gpg: Total number processed: 1\n    gpg:               imported: 1\n    cisl-irving$<\/p>\n    <p> After you do that, if you click on another email message in Apple\n    Mail and click back on the original message, Mail should show\n    \"<strong>The\u00a0message\u00a0has\u00a0been\u00a0signed\u00a0with\u00a0PGP<\/strong>\"\n    instead \"<strong>Missing\u00a0Key\u00a00xaaaaaaaa<\/strong>\".  <\/p>\n    <h4><a name=\"mancl\" id=\"mancl\">Managing keys with command lines<\/a><\/h4>\n    <p>\n    To see the keys in your local keyring:\n    <\/p><p>gpg2 --list-keys<\/p>\n    <p>\n    To load a key into the local keyring, cut the key id from wherever and\n    paste it into one of these:\n    <\/p><p>gpg2 --recv-keys --keyserver hkp:\/\/wwwkeys.pgp.net 0xaaaaaaaa<br \/>gpg2 --recv-keys --keyserver hkp:\/\/keyserver.ucar.edu 0xaaaaaaaa<\/p>\n    <p>\n    To sign a key that's in your keyring:\n    <\/p><p>gpg2 --sign-key AAAAAAAA<\/p>\n    <p>\n    To assign a trust level to a key, it's easiest to use\n    GPG Keychain Access.\n    <\/p>\n    <h4><a name=\"mangka\" id=\"mangka\">Managing keys with GPG_Keychain_Access<\/a><\/h4>\n    <p>\n    Use the <code><b>GPG Keychain Access<\/b><\/code> app to\n    display\/import\/export\/delete them.  The alternative is the\n    \"gpg\" command.\n    <\/p>\n    <p>\n    If you want to assign a trust level (a.k.a.) validity, highlight an\n    entry and do <strong>GPG<\/strong>-&gt;<strong>Update\u00a0Trust\u00a0Database...<\/strong>.\n    It'll open a Terminal window to ask for the\n    trust level - kludgy.  Then, it doesn't let you assign trust levels\n    for a specific key, but only for whatever keys don't have trust\n    levels defined - so how do you change the trust level of a key?\n    <\/p>\n    <h4><a name=\"mans\" id=\"mans\">Managing keys with Seahorse<\/a><\/h4>\n    <p>\n    Because GPG_Keychain_Access seems so kludgy, I downloaded Seahorse\n    from\n    <a href=\"http:\/\/seahorse.sourceforge.net\/\">http:\/\/seahorse.sourceforge.net\/<\/a>.\n    <\/p>\n    <p>\n    I got\n    http:\/\/ftp.gnome.org\/pub\/GNOME\/sources\/seahorse\/0.8\/seahorse-0.9.1.tar.gz\n    and Stuffit\u00a0Expander created a seahorse-0.9.1 directory under\n    Desktop.  I copied it to \/usr\/src.  When I tried to build it, it\n    complained because XML::Parser wasn't in the Perl libraries.  Seahorse\n    seems to have lots of other dependencies, so I gave up.  Sigh.\n    <\/p>\n    <h4><a name=\"pubring\" id=\"pubring\">Updating UCAR keys in pubring<\/a><\/h4>\n<p>On 2017-07-19, Rich Johnson emaied out an encrypted message\ncontaining a file named \"msg.asc\". It contained an updated list of\nkeys for UCAR personnel. I decrypted it with <\/p>\n    \n    <p>\"gpg2 --decrypt smp.asc &gt; msggg.asc\"<\/p>\n<p>Then I started the GPG Keychain app, and dragged msggg.asc into the\napp window. This updated my ~\/.gnupg\/pubring.gpg file.  Yet the app\nrequires that the gnupg drive be mounted. I don't understand yexactly\nwhat it's doing.<\/p>\n          \n    <h3><a name=\"store\" id=\"store\">Initialize a \"PETES KEYS\" flash drive<\/a><\/h3>\n    <p> I keep my SSH keys and my GPG secret keyring on a flash\n    drive.  This section describes how to write my secrets\n    to an empty flash drive.  <\/p>\n    <h4><a name=\"partition\" id=\"partition\">Erase and partition the flash drive<\/a><\/h4>\n    <ol><li>Insert the flash drive into the Mac<\/li>\n      <li>Start Disk\u00a0Utility<\/li>\n      <li>Select the flash drive<\/li>\n      <li>select the <strong>Partition<\/strong> tab<\/li>\n      <li>For <strong>Volume\u00a0Scheme<\/strong>, select <strong>1\u00a0Partition<\/strong><\/li>\n      <li>Set the <strong>Name<\/strong> to <strong>PETES\u00a0KEYS<\/strong><\/li>\n      <li>Set the <strong>Format<\/strong> to <strong>MS-DOS\u00a0File\u00a0System\u00a0(FAT16)<\/strong><\/li>\n      <li>Click <strong>Partition<\/strong><\/li>\n      <li>Click <strong>Partition<\/strong> in the pull-down verification box<\/li>\n    <\/ol><h4><a name=\"diskimage\" id=\"diskimage\">Copy the gnupg directory to the flash drive<\/a><\/h4>\n    <p>\n    CSAC requires that sysadmins keep their GPG keys on a flashdrive.\n    Greg\u00a0Woods described how to do it on Macs in\n    <a href=\"https:\/\/secureweb.scd.ucar.edu\/twiki\/bin\/view\/CSAC\/PGPUSBFlashFilesystems\">PGPUSBFlashFilesystems<\/a>.\n    Greg's description mentions FAT32,\n    but I use FAT16 because it's just as good for drives smaller\n    than 128GB, it's what the Mac Disk Utility can create, and\n    according to some documentation it avoids weirdness with Finder.\n    <\/p>\n    <p>\n    NCAR best practice is to store my GPG secret key on a flashdrive\n    in an encrypted form, so that if I lose the flashdrive, others\n    won't be able to get the key.  They recommend creating an\n    AES-encrypted disk image on the flashdrive.\n    <\/p>\n    <ol><li>Insert the flash drive into the Mac<\/li>\n      <li>Start Disk Utility<\/li>\n      <li><strong>File<\/strong>-&gt;<strong>New<\/strong>-&gt;<strong>Blank Disk Image...<\/strong><\/li>\n      <li>Set <strong>Save As<\/strong> to <strong>gnupg<\/strong><\/li>\n      <li>Set <strong>Where<\/strong> to <strong>PETES KEYS<\/strong><\/li>\n      <li>Set <strong>Size<\/strong> to <strong>2.5 MB<\/strong><\/li>\n      <li>Set <strong>Encryption<\/strong> to <strong>AES-128<\/strong><\/li>\n      <li>Leave <strong>Format<\/strong> as <strong>read\/write disk image<\/strong><\/li>\n      <li>Click <strong>Create<\/strong><\/li>\n      <li>set the password to one you'll remember<\/li>\n    <\/ol><p>\n    At this point, you're either creating your primary flashdrive, or\n    making a backup flashdrive.  Either way, you need to have a secret\n    keyring file (secring.gpg) to put on the new flashdrive.  This\n    file is usually stored only on flashdrives - it is only on the\n    Mac's hard disk while you are in the process of creating flashdrives.  When you\n    first create your primary flashdrive, your secring.gpg file is\n    sitting in ~\/.gnupg because you created it as described in the\n    <a href=\"#gen\">Generate\u00a0PGP\u00a0keys<\/a> section of this\n    document.  When you are making a backup flashdrive, you need to\n    put a temporary copy of the file onto the Mac's hard disk like so:\n    <\/p><ol><li>insert your primary flashdrive<\/li>\n      <li>mount the <strong>gnupg<\/strong> disk image<\/li>\n      <li>cp \/Volumes\/gnupg\/secring.gpg ~\/.gnupg<\/li>\n      <li>eject the <strong>gnupg<\/strong> disk image<\/li>\n      <li>eject the primary flashdrive<\/li>\n    <\/ol><p>\n    Now comes the one-time step: copying the secret key to the flash\n    drive and deleting it from the Mac.  From this point forward,\n    you'll need to have your flashdrive inserted in order to sign or\n    decrypt email.  Also, to initialize new flashdrives, you'll have\n    to copy the gnupg disk image from one flashdrive to another - like\n    my ssh key, it's not anywhere on the Mac's hard disk.\n    <\/p>\n    <ol><li>mount the new flashdrive<\/li>\n      <li>mount the <strong>gnupg<\/strong> disk image<\/li>\n      <li>cp ~\/.gnupg\/secring.gpg \/Volumes\/gnupg<\/li>\n      <li>rm ~\/.gnupg\/secring.gpg<\/li>\n    <\/ol><p>\n    Then configure GPG to get the secret key from the flashdrive.\n    Edit <strong>~\/.gnupg\/gpg.conf<\/strong> and set\n  <\/p><p>#<br \/># I added these 2006-06-02 to implement getting my secret key from my<br \/># flashdrive.  See<br \/>#<br \/>#  http:\/\/netserver.ucar.edu\/intro\/staff\/siemsen\/tools\/gpg.html#diskimage<br \/>#<br \/>no-default-keyring<br \/>keyring ~\/.gnupg\/pubring.gpg<br \/>secret-keyring \/Volumes\/gnupg\/secring.gpg<br \/><\/p>\n  <h4><a name=\"ssh\" id=\"ssh\">Copy the SSH id_rsa file to the flash drive<\/a><\/h4>\n  <p>\n  See my\n  <a href=\"https:\/\/nets.ucar.edu\/nets\/intro\/staff\/siemsen\/linux\/usb\/flashdrive.html\">USB flash drive notes<\/a> details.\n  <\/p>\n  <h3><a name=\"trouble\" id=\"trouble\">Troubleshooting<\/a><\/h3>\n  <dl><dt><b>Mail doesn't decrypt encrypted messages<\/b><\/dt>\n    <dd><p>Under Mail -&gt; Preferences, GPGMail, under Reading, make\n      sure \"Enable OpenPGP\/MIME\" is checked.  <\/p><\/dd>\n      <dt><b>GPG_CONFIG_ERROR_TITLE<\/b> and\/or <b>GPG_CONFIG_ERROR_MESSAGE<\/b><\/dt>\n      <dd><p>This happened to me when I started Mail and I had a symbolic link named\n        \"~\/.gnupg\/secring.gpg\" that pointed to \"\/Volumes\/gnupg\/secring.gpg\"\n        (on a USB drive)\n        and I didn't have the USB drive mounted and the \"gnupg\" dmg mounted.\n        Mounting them and restarting Mail made the error go away.\n        <\/p><\/dd>\n        <dt><b>GPG Config Error - Ups, something went wrong.  GPG was found\n          on your system but there seems to be a problem with the config\n          file.  Please contact the GPGTools team http:\/\/www.gpgtools.org\/about\/html.<\/b><\/dt>\n          <dd><p>This happened to me when installing GPGTools and when trying to\n            start <code><b>\/Applications\/GPG Keychain Access.app<\/b><\/code> when I had a symbolic link named\n            \"~\/.gnupg\/secring.gpg\" that pointed to \"\/Volumes\/gnupg\/secring.gpg\"\n            (on a USB drive)\n            and I didn't have the USB drive mounted and the \"gnupg\" dmg mounted.\n            Mounting them and restarting Mail made the error go away.\n            <\/p><\/dd>\n            <dt><b>Missing Key<\/b><\/dt>\n            <dd>\n              <p>If, when you click \"Verify\", Mail displays\n              \"<em><strong>Missing\u00a0Key\u00a00xaaaaaaaa<\/strong><\/em>\", then you\n              need to load the person's public key into your local public keyring.  I use the command\n              line.  You can get the person's key from a keyserver, or as a\n              block of text that you get from the person's webpage or something.<\/p>\n            <\/dd>\n            <dt><b>No Secret Key<\/b><\/dt>\n            <dd>\n              <p>If, when you click \"Decrypt\", Mail displays\n              \"<em><b>&lt;username&gt; no secret key<\/b><\/em>\", then you probably don't have\n              the flashdrive mounted.  Put the flash drive into the slot and do\n            <\/p><p>open \/Volumes\/PETES\\ KEYS\\gnupg.dmg<\/p>\n            <p>Enter the password.  Then try again.\n            <\/p><\/dd>\n            <dt><b>PGP Encryption Failed<\/b><\/dt>\n            <dd>\n              <p>If, when signing or encrypting an an outbound message, you get a dialog\n              box that says\n              <\/p><blockquote><div><em>\n              <b>PGP Encryption Failed<\/b>No valid personal key has been found.  Either none<br \/>was selected, or selected key is revoked, disabled or<br \/>has expired.\n              <\/em><\/div><\/blockquote>\n              <p>\n              Mail couldn't find your personal secret key, probably because you\n              don't have the flashdrive mounted.  Put the flash drive into the slot and do\n            <\/p><p>open \/Volumes\/PETES\\ KEYS\\gnupg.dmg<\/p>\n            <p>\n            Enter the password.  Then try again.\n            <\/p><\/dd>\n          <\/dl><h3><a name=\"gpg-agent\">gpg-agent<\/a><\/h3>\n          <p>Briefly, gpg-agent is a superset of ssh-agent - it stores\n          secret (private) keys in a process so you don't have to keep\n          a sensitive file mounted all the time.<\/p>\n          <p>As described above, UCAR policy says my secret key has to\n          be stored in an inconvenient place: in a secret keyring on an\n          encrypted drive on a USB flash drive.  This means that to use\n          PGP, I have to go through this Royal Pain:<\/p><ol><li>physically insert the USB drive<\/li>\n            <li>open it and double-click on the gnupg drive<\/li>\n            <li>type a password to decrypt the subdrive<\/li>\n            <li>start Mail<\/li>\n            <li>don't pull out the USB drive<\/li>\n          <\/ol><p>To make it worse, I have to tell Mail\/GPGMail where the\n          secret keyring is by configuring my gpg.conf file with\n          \"secret-keyring \/Volumes\/gnupg\/secring.gpg\".  If\n          that drive isn't mounted when I start Mail, it will generate error\n          messages and Mail becomes unreliable (crashes every so often).\n          <\/p>\n          <p>This is so onerous that people wouldn't use PGP at all, so\n          someone came up with a solution: gpg-agent.  Gpg-agent is a\n          daemon that stores your secret key so you have to go through\n          the Royal Pain much less often.  You endure the Royal Pain\n          once at login time to save the key in gpg-agent.  Then\n          Mail\/GPGMail and the rest of the gpg system gets the key from\n          gpg-agent.  The GPGTools folks anticipated that you'll do\n          this, so they set up gpg-agent for you - it's started as a\n          launchd daemon at boot time.  The trick is to get your secret\n          key into gpg-agent at login time.  <\/p>\n          <p>So we assume gpg-agent is already running, and we just\n          have to load our secret key into it.  The new Less Royal\n          Pain procedure is<\/p>\n          <ol><li>physically insert the USB drive<\/li>\n            <li>open it and double-click on the gnupg drive<\/li>\n            <li>type a password to decrypt the subdrive<\/li>\n            <li>load the secret key into gpg-agent<\/li>\n            <li>unmount the subdrive<\/li>\n            <li>unmount the USB drive<\/li>\n            <li>pull out the USB drive<\/li>\n            <li>start and stop Mail as often as you like<\/li>\n          <\/ol><h3><a name=\"abeyk\" id=\"abeyk\">ABKey: MacGPG integration with Address Book<\/a><\/h3>\n          <p>\n          ABKey is a plug-in for Address Book that shows GPG information in\n          Address\u00a0Book entries.  See\n          <a href=\"http:\/\/www.far-blue.co.uk\/projects\/keymanager.html\">\n          http:\/\/www.far-blue.co.uk\/projects\/keymanager.html<\/a>.\n          It works, but as the author warns, you have to move the mouse over an\n          email address in an Address\u00a0Book entry to activate it.\n          <\/p>\n          <hr \/><span>\n          Address comments or questions about this Web page to the\n          <a href=\"mailto:siemsen@ucar.edu\">siemsen@ucar.edu<\/a>.<br \/>Last modified: Wed Jul 19 15:46:56 MDT 2017 \n          <\/span>","created_at":"2018-11-07T16:53:49+0000","updated_at":"2018-11-07T16:53:49+0000","annotations":[],"mimetype":"text\/html","reading_time":18,"domain_name":"nets.ucar.edu","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":624,"title":"How to Conduct a SWOT Analysis for Your Small Business","url":"https:\/\/www.thebalancesmb.com\/swot-analysis-for-small-business-2951706","content":"<div id=\"article-meta_1-0\" class=\"comp article-meta mntl-block\" data-tracking-container=\"true\"><p>By\n<a href=\"https:\/\/www.thebalancesmb.com\/alyssa-gregory-2951140\" class=\"byline__link\">Alyssa Gregory<\/a><\/p><p>Updated August 13, 2018<\/p><\/div><div id=\"mntl-sc-page_1-0\" class=\"comp mntl-sc-page mntl-block structured-content expert-content\" data-sc-sticky-offset=\"80\" data-sc-ad-label-height=\"20\" data-sc-offset=\"200\" data-sc-ad-track-spacing=\"100\" data-sc-min-track-height=\"250\" data-sc-max-track-height=\"600\" data-sc-breakpoint=\"54em\" data-sc-load-immediate=\"4\" data-sc-content-positions=\"[300, 1, 1, 1250, 1, 1, 1, 1]\" data-bind-scroll-on-start=\"true\"><div id=\"mntl-sc-block_1-0\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>A SWOT analysis is a strategic planning tool that helps a business owner identify his or her own strengths and weaknesses, as well as any opportunities and threats that may exist in a specific business situation. A SWOT analysis is most commonly used as part of a <a href=\"https:\/\/www.thebalancesmb.com\/how-to-write-a-small-business-marketing-plan-2951749\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">marketing plan<\/a>, but it is also a good tool for general business strategizing, and to use as a starting point for team discussions.<\/p><\/div><div id=\"mntl-sc-block_1-0-2\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>When conducted thoroughly, a SWOT analysis can uncover a wealth of information and can be useful in a number of situations. This article will walk you through how to conduct a SWOT analysis and provide some tips that will help you use the tool effectively.\u200b<\/p><\/div><figure id=\"mntl-sc-block_1-0-4\" class=\"comp mntl-sc-block mntl-sc-block-image figure-landscape figure-high-res\"><div class=\"figure-media\"><div class=\"img-placeholder\"><img data-srcset=\"https:\/\/www.thebalancesmb.com\/thmb\/wcD-gb5aTDdavO7-GDQNxHxmIM8=\/300x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 300w, https:\/\/www.thebalancesmb.com\/thmb\/AQ7xLxpV_jEIVMCpxtJXt4mWH0Y=\/600x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 600w, https:\/\/www.thebalancesmb.com\/thmb\/WkYFlVQHF9TWe-HamGIqDzJE52g=\/900x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 900w, https:\/\/www.thebalancesmb.com\/thmb\/qy3XebJYepEM1t_rYP_WNrbW9nQ=\/1500x0\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png 1500w\" src=\"https:\/\/www.thebalancesmb.com\/thmb\/9CAr1q093sjzzJZ3PtbG82fekiE=\/1500x1000\/filters:no_upscale():max_bytes(150000):strip_icc()\/swot-analysis-for-small-business-2951706-v02-5b3b809dc9e77c001afe8f7e.png\" class=\"lazyload\" data-click-tracked=\"true\" data-img-lightbox=\"true\" data-expand=\"300\" id=\"mntl-sc-block-image_1-0-4\" data-tracking-container=\"true\" alt=\"image\" \/><\/div><\/div>\n<figcaption id=\"mntl-figure-caption_1-0\" class=\"comp mntl-figure-caption figure-article-caption\">\n\u00a9\u00a0The Balance\n<\/figcaption><\/figure><div id=\"mntl-sc-block_1-0-6\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>A SWOT matrix is usually depicted as a square divided into four quadrants. Each quadrant represents one element of the SWOT analysis -- Strengths, Weaknesses, Opportunities, Threats.<\/p><\/div><div id=\"mntl-sc-block_1-0-9\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>The easiest way to start filling in each quadrant in the matrix above is by answering a series of questions. Use the list below to get started, focusing on the questions that are most relevant to your business and current situation.<\/p><\/div><div id=\"mntl-sc-block_1-0-11\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Strengths:<\/strong> For this quadrant, think about the attributes of yourself and your business that will help you achieve your objective. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-13\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>What do you do well?<\/li><li>What are your unique skills?<\/li><li>What expert or specialized knowledge do you have?<\/li><li>What experience do you have?<\/li><li>What do you do better than your competitors?<\/li><li>Where are you most profitable in your business?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-15\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Weaknesses:<\/strong> For this quadrant, think about the attributes of yourself and your business that could hurt your progress in achieving your objective. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-17\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>In what areas do you need to improve?<\/li><li>What resources do you lack?<\/li><li>What parts of your business are not very profitable?<\/li><li>Where do you need further education and\/or experience?<\/li><li>What costs you time and\/or money?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-19\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Opportunities:<\/strong> For this quadrant, think about the external conditions that will help you achieve your objective. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-21\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>What are the <a href=\"https:\/\/www.thebalancesmb.com\/four-common-small-business-goals-to-inspire-you-2951810\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">business goals<\/a> you are currently working towards?<\/li><li>How can you do more with your existing customers or clients?<\/li><li>How can you <a href=\"https:\/\/www.thebalancesmb.com\/low-cost-technology-solutions-2951652\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"2\">use technology<\/a> to enhance your business?<\/li><li>Are there new target audiences you have the potential to reach?<\/li><li>Are there related products and services that provide an opportunity for your business?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-23\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p><strong>Threats:<\/strong> For this quadrant, think about the external conditions that could damage your business's performance. Questions to consider:<\/p><\/div><div id=\"mntl-sc-block_1-0-25\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>What obstacles do you face?<\/li><li>What are the strengths of your biggest competitors?<\/li><li>What are your competitors doing that you're not?<\/li><li>What's going on in the economy?<\/li><li>What's going on in the industry?<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-28\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>One of the most important parts of your SWOT analysis is using the data you compiled to identify <a href=\"https:\/\/www.thebalancesmb.com\/swot-example-strengths-weaknesses-opportunities-threats-2947985\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">new strategies<\/a> and goals for your business. For example, you can:<\/p><\/div><div id=\"mntl-sc-block_1-0-30\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>Create a plan to build up your strengths even more<\/li><li>List ways you can work on building up your weaknesses<\/li><li>Set <a href=\"https:\/\/www.thebalancesmb.com\/smart-goal-setting-101-2951829\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">SMART goals<\/a> for each of the opportunities you identified<\/li><li>Devise a plan to use your strengths to decrease the threats you identified<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-32\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>Then, look for ways to combine data from different quadrants in even more ways:<\/p><\/div><div id=\"mntl-sc-block_1-0-34\" class=\"comp mntl-sc-block mntl-sc-block-html\"><ul><li>Explore how you can combine your strengths and opportunities to develop new strategies<\/li><li>Try combining strengths and threats to identify threats you can eliminate<\/li><li>Look at your weaknesses and opportunities to create a list of areas ready for improvement<\/li><li>Make a list of areas to avoid that fall under weaknesses and threats<\/li><\/ul><\/div><div id=\"mntl-sc-block_1-0-36\" class=\"comp mntl-sc-block mntl-sc-block-html\"><p>Once you understand how to compile your <a href=\"https:\/\/www.thebalancesmb.com\/every-business-needs-a-good-swot-analysis-1794147\" data-component=\"link\" data-source=\"inlineLink\" data-type=\"internalLink\" data-ordinal=\"1\">SWOT data<\/a> and find ways to use it strategically, the SWOT analysis will be a tool that you can use over and over in your business to explore new opportunities and improve your decision-making process.<\/p><\/div><\/div>","created_at":"2018-11-07T16:53:49+0000","updated_at":"2018-11-07T16:53:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"www.thebalancesmb.com","preview_picture":"https:\/\/www.thebalancesmb.com\/thmb\/wKFp9D0ZGgVDUJYEo-2UF9wrS6g=\/1080x1080\/filters:fill(auto,1)\/money-business-social-share-default-5ac795acfa6bcc00376a1bdb.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":618,"title":"Word vectors are awesome but you don\u2019t need a neural network to find them","url":"https:\/\/news.ycombinator.com\/item?id=15502859","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15502859\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15502859\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/multithreaded.stitchfix.com\/blog\/2017\/10\/18\/stop-using-word2vec\/\" class=\"storylink\">Word vectors are awesome but you don\u2019t need a neural network to find them<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=stitchfix.com\">stitchfix.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        276 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=blopeur\" class=\"hnuser\">blopeur<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15502859\">on Oct 18, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15502859&amp;goto=item%3Fid%3D15502859\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Word%20vectors%20are%20awesome%20but%20you%20don%E2%80%99t%20need%20a%20neural%20network%20to%20find%20them&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Word%20vectors%20are%20awesome%20but%20you%20don%E2%80%99t%20need%20a%20neural%20network%20to%20find%20them\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15502859&amp;auth=49b9a62899d89f23281cc586d3a36343b2c7b496\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15502859\">66\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:53:48+0000","updated_at":"2018-11-07T16:53:48+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":619,"title":"git git git git git","url":"http:\/\/caiustheory.com\/git-git-git-git-git\/","content":"<header><time datetime=\"2017-09-26T11:00:23Z\">2017-09-26 11:00:23<\/time><\/header><p>Ever found you've accidentally entered too many <code>git<\/code>s in your terminal and wondered if there's a solution to it? I quite often type <code>git<\/code> then go away and come back, then type a full <code>git status<\/code> after it. This leads to a lovely (annoying) error out the box:<\/p><div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">$ git git status\ngit: 'git' is not a git command. See 'git --help'.<\/pre><\/div><p>What a git.<\/p><p>My initial thought was overriding the <code>git<\/code> binary in my <code>$PATH<\/code> and having it strip any leading arguments that match <code>git<\/code>, so we end up running just the <code>git status<\/code> at the end of the arguments. An easier way is to just use <a href=\"https:\/\/git-scm.com\/docs\/git-config\"><code>git-config<\/code><\/a>'s <code>alias.*<\/code> functionality to expand the first argument being <code>git<\/code> to a shell command.<\/p><div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">git config --global alias.git '!exec git'<\/pre><\/div><p>Which adds the following git config to your <code>.gitconfig<\/code> file<\/p><div class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">[alias]\n  git = !exec git<\/pre><\/div><p>And then you'll find you can <code>git git<\/code> to your heart's content<\/p><div class=\"highlight\"><pre class=\"language-shell\" data-lang=\"shell\">$ git sha\ncc9c642663c0b63fba3964297c13ce9b61209313\n$ git git sha\ncc9c642663c0b63fba3964297c13ce9b61209313\n$ git git git git git git git git git git git git git git git git git git git git git git git git git git sha\ncc9c642663c0b63fba3964297c13ce9b61209313<\/pre><\/div><p>(<code>git sha<\/code> is an alias for <code>git rev-parse HEAD<\/code>.)<\/p><p>See what other git alias' I have in my <a href=\"https:\/\/github.com\/caius\/zshrc\/blob\/master\/dotfiles\/gitconfig\"><code>~\/.gitconfig<\/code><\/a>, and laugh at all the typo corrections I have in there. (Yes, git provides autocorrection if you enable it, but I'm used to these typos working!)<\/p><p>Now <code>git<\/code> back to doing useful things!<\/p><footer><p class=\"license\">\n        This work is licensed under a <a rel=\"license\" href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/4.0\/\">Creative Commons Attribution-ShareAlike 4.0 International License<\/a>.\n      <\/p>\n  <\/footer>","created_at":"2018-11-07T16:53:48+0000","updated_at":"2018-11-07T16:53:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"caiustheory.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":617,"title":"How We Use Quora For Lead Generation and Marketing Software Services \u2013 Romexsoft","url":"https:\/\/www.romexsoft.com\/blog\/quora-marketing\/","content":"<h2>In this guide, I\u2019ll share the exact tips for using Quora as content marketing and lead generation tool we\u2019ve used at Romexsoft, plus the results we\u2019ve gotten. &#13;\n<\/h2><p>Quora is a big deal these days.<\/p><p>You\u2019ve heard the buzz \u2013 it can bring massive traffic to your corporate blog, leverage your authority in the niche and drive qualified leads.<\/p><p>But does marketing on Quora work for software companies?<\/p><p>Indeed, it does.<\/p><p>In this guide, I\u2019ll share the exact tips we\u2019ve used at Romexsoft, plus the results we\u2019ve gotten. And, boy, those were surprising!<\/p><p>On to the practicalities!<\/p><p><strong>What You Will Learn Today:<\/strong><\/p><ol><li>What exactly is Quora and why should you care?<\/li>\n<li>Why Quora is hot right now?<\/li>\n<li>Why Quora is great for B2B Marketing.<\/li>\n<li>How we started using Quora<\/li>\n<li>How to create an attractive Quora profile.<\/li>\n<li>How to navigate the platform and use it to your advantage<\/li>\n<li>How to use Quora for lead generation.<\/li>\n<li>How to use Quora for content marketing.<\/li>\n<li>Useful add-ons.<\/li>\n<li>Things to watch out for.<\/li>\n<li>Recap.<\/li>\n<\/ol><h3>What exactly is Quora and why should you care?<\/h3><p>Quora is a popular questions &amp; answers platform introduced by two ex-facebook employees \u2013 D\u2019Angelo and Charlie Cheever in 2009. From that time Quora had its ups and downs from being extremely niche and popular in 2011 and then a slight downfall in 2013 with a steady growth during 2015 up till now. In December 2016, Quora\u2019s co-founder and CEO Adam D\u2019Angelo revealed that Quora hit <a href=\"http:\/\/techcrunch.com\/2016\/01\/27\/quora-80-million-users\/\">80 million unique monthly visitors<\/a> by simply <a href=\"https:\/\/www.quora.com\/How-many-people-use-Quora-7\">answering the question on Quora<\/a> in his monthly session with more than <a href=\"http:\/\/blogs.wsj.com\/digits\/2014\/04\/09\/with-80-million-in-new-capital-quora-still-has-no-business-model\/\">500,000 topics on the platform<\/a>.<\/p><p>It\u2019s a cool platform apart from Ask.com and Yahoo Answers, where anyone can ask and respond to questions about basically anything in the world, even things like \u201c<a href=\"https:\/\/www.quora.com\/Whats-it-like-to-play-basketball-with-President-Obama\">What\u2019s it like to play basketball with President Obama?<\/a>\u201d and, amazingly enough, Obama will hop into answering.<\/p><p>Moreover, Quora is quickly indexed by Google which gives the platform an additional value and increases the chance for searches to stumble on your writings.<\/p><p>There are many discussion and articles on the web how Quora does wonders for thought leadership, helps to raise brand awareness and generates leads. We all already know that using Quora is beneficial in many ways. And it is unquestionable whether to use it or not.<\/p><p>Yet, I\u2019d like to tackle a more specific angle: how we can use Quora for B2B marketing in outsourcing software development niche to attract new leads and enhance our content efforts based on my experience.<\/p><h3>Why Quora is Hot Right Now?<\/h3><ol><li>You can build authority in your niche or industry.<\/li>\n<li>Drive long-term traffic to your website.<\/li>\n<li>Demonstrate your expertise and reputation on almost any topic.<\/li>\n<li>Promote your content by sharing topic-focused boards on your profile.<\/li>\n<li>Get insights from experts in any industry.<\/li>\n<li>Do market research of your industry.<\/li>\n<li>Perform competitor analysis.<\/li>\n<li>Target your questions to specific Quora users.<\/li>\n<li>Start your Quora blog for content syndication.<\/li>\n<li>Search for specific questions or keywords related to your domain to stay on top of the industry buzz.<\/li>\n<\/ol><h3>Why Quora is Great for B2B Marketing<\/h3><p>For B2B marketing Quora can be the perfect channel to leverage your expertise and offer your knowledge to solve your client concerns and problems.<\/p><p>All questions are answered by real people as Quora requires real name registration and the community is here to vet and react over your answers.<\/p><p>So you have a team. All of them are experts in different fields e.g. design, marketing, testing, etc. with tons of professional tips and secrets to spill. Ask them to create profiles and spend a few minutes a day browsing and answering questions.<\/p><p>This way you can build a trustworthy personal brand, but your name is still tied to your company name, and it gets a certain amount of exposure too.<\/p><h3>How We Started Using Quora<\/h3><p>We began using Quora extensively for marketing in December 2015 with two employees accounts. Our main goal was to generate leads and promote existing content. As for today, my personal account gained over 6,185 views!<\/p><p>Quora became one of the most quality traffic sources for our website.<\/p><p>Ready to get started with Quora too? Read on!<\/p><h3>How to Create an Attractive Quora Profile (In Case You Don\u2019t Have One Yet)<\/h3><p>To create your account, you can either login with Facebook or Google+ or use your email to sign up. By setting up your profile, think about interests and topics you would like to contribute to, a snappy tagline and profile content. Quora shows only the first 50 characters of your profile as a tagline above your answer.<\/p><p>Your short bio will appear along with the answers you write. Mention your brand or website name in the beginning, as most of the users will only see your tagline.<\/p><p>Don\u2019t forget to include links to your website\/blog and primary social media accounts. In my case, I linked my personal Facebook &amp; LinkedIn, though Twitter is linked to my company account.<\/p><p>Apart from your main bio, you can set topic-related bios. Click <strong>Profile page<\/strong> and choose <strong>Knows about<\/strong> section, where you can add and describe all the experiences you have.<\/p><p>You can go ahead and complete your Quora profile with additional information by adding your areas of expertises, interests, location, employment, education and so on.<\/p><h3>How to navigate the platform and use it to your advantage<\/h3><p><strong>Search.<\/strong> Search is the one of the most powerful tools at Quora, and it is the perfect tool for\u2026market research. You can learn the exact things people in your industry look for and struggle with. On top, you can keep track on all the talks relevant to your industry via email by clicking the Settings button and configure your Email Notifications. You\u2019ll get a quick summary newsletter for review daily.<\/p><p><strong>Following Topics.<\/strong> Find and start the following topic feeds relevant to your industry. For example, subscribe to <a href=\"https:\/\/www.quora.com\/topic\/Outsourcing\">Outsourcing<\/a> if you are a software development company like us. Start typing in your interest and Quora will auto-provide you with suggested topic feeds and questions to check out. Click the button \u201csearch\u201d to see all the existing content for your keyword.<\/p><p>As Romexsoft is a software development biz, mainly in the outsourcing niche, the topics we are mostly interested in are Outsourcing, Inbound Marketing, Startups, IT Outsourcing, Healthcare IT, Healthcare Technology, FinTech Startup, Edutech, Automated Testing, etc.<\/p><p>User search and follow\/subscribe. There is an option to look for users, mostly experts in your niche. You can follow them or subscribe to receive notifications about their newest answers. In software development sphere I recommend following other opinion leaders, trends setters and bloggers with useful and interesting content to bring more value to your account and network coverage.<\/p><p>Here\u2019re some great guys from my feed:<\/p><ul><li><a href=\"https:\/\/www.quora.com\/profile\/Quincy-Larson\">Quincy Larson<\/a><\/li>\n<li><a href=\"https:\/\/www.quora.com\/profile\/John-L-Miller\">John L. Miller<\/a><\/li>\n<li><a href=\"https:\/\/www.quora.com\/profile\/Jason-M.-Lemkin\">Jason M. Lemkin<\/a><\/li>\n<li><a href=\"https:\/\/www.quora.com\/profile\/J%C3%A9r%C3%B4me-Cukier\">J\u00e9r\u00f4me Cukier<\/a><\/li>\n<li><a href=\"https:\/\/www.quora.com\/profile\/Auren-Hoffman\">Auren Hoffman<\/a><\/li>\n<li>And yours truly \u2013 <a href=\"https:\/\/www.quora.com\/profile\/Valentyna-Zalevska\">Valentyna Zalevska<\/a><\/li>\n<\/ul><p>Blog posts search and follow. For specific blog posts on Quora, search for blog name or author to follow and receive updates for new posts. All your follow actions are tracked by Quora, by logging in \u2013 you will receive notifications about the latest activities around the topics, users and blogs you\u2019ve followed\/subscribed to.<\/p><h3>How To Use Quora for Lead Generation<\/h3><p>Answer questions. Quora is not the place to shamelessly self-promote or bring zero value to your answer. Poor answers can get collapsed or downvoted by the community. Always aim to write longer and more detailed answers. Formatting and structure matters as well, especially for easier \u201cscanning\u201d: use paragraph breaks, bullet points, numbered lists, bold or italics highlights, adding relevant images or a video is a plus too.<br \/><a href=\"https:\/\/www.romexsoft.com\/subscribe-to-romexsoft-blog\/\" target=\"_blank\"> <img src=\"https:\/\/www.romexsoft.com\/wp-content\/uploads\/2018\/01\/Sign-up-for-Industry-Updates_Romexsoft.png\" width=\"350\" height=\"350\" alt=\"image\" \/><\/a><br \/>In search Quora answers will display as a short description with just the first few lines visible and Read more tag. Just like in a blog post, the first lines should be catchy enough for the reader to click through and view the complete answer.<\/p><p>Try to provide a detailed answer as in-depth; educational answers receive more community love and upvotes. Yet do not aim at 1000+ words. Leave those texts for a blog post. Include reference links and data when applicable as in the example above. I linked to the blog article after the key points and conclusion for those, who\u2019d love to know more.<\/p><p><strong>Link-building<\/strong>. You shouldn\u2019t always link your company website or blog. Shameless self-promotion and spamming will not win you any friends. Bring value first. Answer the questions just for the sake of sharing your expertise. Credit and cite others as well as long as the material backs up your tip.<\/p><p><strong>Mentions<\/strong>. Mentioning other Quora writers and users is another good thing to do. You can tag those, who may have more experience on the subject or previously covered the same topics. They will receive a mention notification and check out your answer and profile. This way you can build your trustful network of influencers and connect with amazing peeps in your niche.<\/p><p>Quality answers will have more views and get more chances to be upvoted and shared across the web. Also people, who loved your answer can follow or subscribe to your profile or head straight to your blog for more content. All these actions lead to bigger conversion rate on your website.<\/p><p><strong>Share your answers on other social networks<\/strong>. Most of your prospects look for a solution on your website or elsewhere regarding your processes, services and approach to work. Try sharing your answers on social media. Google will index your answers even better as you accumulate social proof. Quora allows \u201cAuto-sharing\u201d once you\u2019ve connected other social networks on your profile.<\/p><p><strong>Asking questions<\/strong>. Quora is the perfect source to gather all sort of data. You can ask the question you are currently struggling with or researching. As you can see in the example below, I recently asked about automation tools people use in the daily work, because we are preparing a new blog post about automation testing tools and I\u2019d love to hear the community feedback.<\/p><p>Ask to answer feature allows you to request an answer from 15 popular writers on your topic giving you a chance to get the most qualified answer.<\/p><p>Quora encourages to ask questions you want to answer \u2013 a \u201cwin-win\u201d case to show your competence in certain topics and hear other opinions as well.<\/p><p><strong>Analytics &amp; ROI<\/strong>. On Quora you can see how many people have viewed your answers and profiles of people, who followed your question. You can easily connect with your followers or message them directly with related content offers.<\/p><p>On top, there is a separate analytics section called \u201cStats.\u201d This free analytics tool provides you with the views, upvotes and shares for all the categories \u2013 questions, answers, posts and all the content you\u2019ve contributed.<\/p><p>Views and upvotes are quite vanity metrics for me. The only thing I find useful is to identify the key topics which may be the content candidates by the number of views.<\/p><p>The thing people forget about when starting to use Quora is to track ROI coming from the answers on Quora in Google Analytics. I use GA, but if you are using other tracking tools, please, let me know in the comments below.<\/p><p>It\u2019s quite and easy step a tracking campaign in GA: Go to \u201cReporting tab\u201d, then choose \u201cSocial\u201d under \u201cAcquisition.\u201d Then go to \u201cNetwork Referrals\u201d. Quora will appear as one of your sources if you already have started your activity there. Furthermore, by clicking on Quora source, you can see which pages are getting the most traffic from Quora and what answers include those referrals. I can quickly see which blog posts turned out to be the most popular and which answers on Quora I gave mentioning this link.<\/p><p>If you are setting your conversion tracking in GA, you can set the funnels and events to track if your customers came from Quora or other sources. For more info on that, you can go to a <a href=\"http:\/\/blog.wishpond.com\/post\/115675435946\/how-we-answer-quora-questions-to-drive-traffic-to-our-blogwebsite\">brilliant piece by Cara Tarbaj<\/a> at <a href=\"http:\/\/blog.wishpond.com\/\">Wishpond blog<\/a>.<\/p><p><strong>Message prospects and followers<\/strong>. Private messaging is a powerful feature for your lead generation strategy. You can send a message to your followers and prospects if the privacy settings allow doing so. Make sure you choose to receive messages from anyone in your <a href=\"https:\/\/www.quora.com\/settings\/privacy\">privacy settings on Quora<\/a> if you\u2019d like to connect with more prospects.<\/p><p>There are two ways to reach your marketing pre-qualified leads as they are already interested in your answer\/question\/topic.<\/p><ul><li>send a thank you note to your followers;<\/li>\n<li>send a direct follow-up after your answer to the question with targeted offers and link to your landing page;<\/li>\n<li>connect with them on LinkedIn with a Quora reference.<\/li>\n<\/ul><h3>How To Use Quora for Content Marketing<\/h3><p><strong>Discover content ideas<\/strong>. Type your targeted keywords in search and Quora will help you discover the top questions asked by the community. You can easily find trending topics for your blog and other content marketing materials. It takes your time to dive in and research what are the most viewed and followed questions on Quora in your niche and how you can plan your content strategy around them. The most upvoted and viewed questions can be a good signal of valuable content to produce.<\/p><p>You can filter your search results by type of questions, answers or posts and also by the timeline.<\/p><p><strong>Quora blog<\/strong>. One of the smart ways to get even more exposure from Quora is to start your blog apart from doing the questions-and-answers thing. Go to your profile and navigate to Blogs, create your unique name, URL and a catchy description.<\/p><p>Add your content, just like on Medium, Facebook Notes or LinkedIn Pulse. In general, Quora posts get higher rankings in search results and more exposure from around the network.<\/p><p><em><strong>Quick Tip:<\/strong> Repurpose your blog content, either from your personal or your company blog that you contributed to. Review your old content on Quora and give it a second traffic boost.<\/em><\/p><p><strong>Create your company page<\/strong>. Use one of Quora features \u2013 building a company page, to leverage your brand\u2019s visibility. This feature is particularly useful for product software development companies as your company can track the questions about the product and answer them directly under the company page on Quora. Software outsourcing companies providing services can benefit from this page by creating an additional visibility and also answering basic questions potential prospects may have. Offer customer support to folks who are not your customers\u2026.and they are likely to become them!<\/p><p>I advise creating a simple FAQ about your company with the most popular inquiries and ask your current customers to post references. This way, you establish social proof and authority, plus generate a long-term traffic source showing high in search results for branded keywords.<\/p><p>See a <a href=\"https:\/\/www.quora.com\/topic\/Quora\">Quora topic example<\/a> below with FAQ about the company itself and customers feedback.<\/p><h3>Interesting add-ons<\/h3><p><strong>\u201cTop writers.\u201d<\/strong> Quora actively encourages users to answer more questions and receive \u201cMost Viewed Writer\u201d accolade in return. To get this privileged badge, you need to accumulate the highest number of answers views over the past 30 days. The rankings consist of top 10 writers for each topic. The good news is \u2013 you can become a top writer for numerous topics.<\/p><p>By reaching more than 5000 views on my answers, I became the most viewed writer in Code Outsourcing. This is a great way for you to get noticed by other users, who can request you to answer their questions and connect with you. The badge indicates the level of your authority on Quora.<\/p><p>Online sessions. The great Quora feature for promotion and knowledge sharing are online sessions. Tech stars, presidents and more, can go live for an hour or two to answer questions on Quora, either personal or professional ones.<\/p><p><a href=\"https:\/\/www.quora.com\/profile\/Sheryl-Sandberg\/session\/18\/\">The last session<\/a> I attended was with <a href=\"https:\/\/www.quora.com\/profile\/Sheryl-Sandberg\">Sheryl Sandberg<\/a>, COO at Facebook, where she intended to pick the best ten questions to answer. This is a great way to talk and reach out to opinion leaders and influencers in your industry. By the way, in small font below the answer, you can identify the person who asked the question, which stands for a perfect opportunity to think of smarter questions to get noticed by others and even potentially get reached out to by the session speaker.<\/p><p><strong>Quora app<\/strong>. Good news for you \u2013 you can keep track of your views and answers by using a Quora app. It is available for Windows, Android and iOS and has pretty neat design and navigation.<\/p><p><strong>Knowledge badge<\/strong>. An experimental feature introduced by Quora is Knowledge Prizes (currently on beta), which aims to inspire experts to answer important questions online and receive a financial contribution for that. Technology topics only are in bet now, which is an obvious sign of their extreme popularity. Some questions are sponsored by companies, who want to get expert opinions and rewards the best ones with a prize. It\u2019s up to a winner on how to spend the money, though. These days Quora recommends transferring the prize to partner charities. You can read more about Knowledge Prizes on <a href=\"https:\/\/blog.quora.com\/Knowledge-Prizes-Unlock-answers-to-important-questions\">Quora blog<\/a>.<\/p><h3>Things Watch Out For<\/h3><p>Quora is not the place for self-promotion or heavy promotion of your website, products or services. It has no tolerance towards spam and aggressive link-building to leverage your SEO rankings. Quora was <a href=\"https:\/\/www.quora.com\/about\">created to help people<\/a> find the information they need. Up till now it remains free source and doesn\u2019t have a business model to earn profits.<\/p><p>Mention your website in your bio description is more than enough. When people like your posts and content, they can see where are you working and what services you provide. If you blogged about something relevant to the answer on Quora, please, still give the fairly comprehensive answer with a link to Read more in other sources. Quora attractively links external sources by showing up the entire titles.<\/p><p>Sometimes your answer can literally \u201ccollapse\u201d and it will not be shown in the main answer thread. It means that either your answer was downvoted, too short, requires improvements or you haven\u2019t passed the real name verification process.<\/p><h3>Recap<\/h3><p>One more tool for your content marketing efforts and lead generation won\u2019t help your business; there is always need to set goals and objectives why you are using Quora. You can limit yourself on how much time you spend on Quora weekly and create a checklist of what are the steps to take.<\/p><p>My strategy is to use Quora three times per week no more than 20 minutes for the following activities:<\/p><ul><li>search for questions with the most followers or views and answer them;<\/li>\n<li>follow and answer the new questions in my expert topics by checking notifications, when possible, include the links back to the website;<\/li>\n<li>upvote and comment the answers I like;<\/li>\n<li>follow new active people and influencers by searching for the top answers in chosen topics;<\/li>\n<li>send and respond to messages;<\/li>\n<li>content repurposing from the company blog;<\/li>\n<li>constantly improving the profile by adding new info on expert topics and include new targeted links.<\/li>\n<\/ul><p>I hope that this post will boost your future marketing on Quora. If you already use the platform, let me know the results in the comments below.<\/p><p>P.S. Are you interested in other aspects of digital marketing and how we do it at Romexsoft? Then you definitely should read <a href=\"https:\/\/blog.digitalinbound.com\/internet-marketing-blog\/expert-interview-series-nataliia-bubniuk-of-romexsoft.com-on-creating-a-digital-marketing-strategist\">an interview with our\u00a0Digital Marketing Strategist \u2013 Nataliia Bubniuk.<\/a><\/p>","created_at":"2018-11-07T16:53:47+0000","updated_at":"2018-11-07T16:53:47+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":16,"domain_name":"www.romexsoft.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":616,"title":"Real-World Startup Sales and Marketing for the Lost and the Clueless (Part 5 of The Epic Guide) \u2013 Indie Hackers","url":"https:\/\/www.indiehackers.com\/@cliffordoravec\/real-world-startup-sales-and-marketing-for-the-lost-and-the-clueless-8f2df4804c","content":"<div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/zoolander-files-in-computer.jpeg\" alt=\"\" class=\"\" \/><figcaption class=\"\">The customers are *on* the internet? It's so simple!<\/figcaption><\/figure><\/div><p class=\"\">Or\u200a\u2014\u200a\u201cHow to Get Real Traffic and Real Customers in Just 1 Hour a Day!\u201d<\/p><p><br \/>Wow, it\u2019s been a while!<\/p><p>I know, I know!<\/p><p>In fact, it\u2019s been months since the last installation of The Epic Guide.<\/p><p>All I can say is that I certainly hope that you\u2019re all happy with yourselves.<\/p><p>It\u2019s because of all of you super fans of The Epic Guide that I\u2019ve had to basically go into hiding all summer long to focus on <a href=\"https:\/\/cliffordoravec.com\/the-epic-guide-to-bootstrapping-a-startup\" target=\"_blank\" rel=\"noopener noreferrer\"><b>writing this damned book<\/b><\/a>\u200a\u2014\u200awhile still <a href=\"https:\/\/gettamboo.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><b>running and growing Tamboo<\/b><\/a> on top of it.<\/p><p>In the process of being around-the-clock heads down, I\u2019ve grown out a mean Thor-inspired hair style and aesthetic (not to mention a growing case of agoraphobia) that I think would make even Howard Hughes himself envious and proud.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/thor.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Write a book they said! It'll be \"good for you\" and \"fun\" they said!<\/figcaption><\/figure><\/div><p id=\"39fa\">One thing of note that was actually pretty super fun is that <a href=\"https:\/\/www.indiehackers.com\/podcast\/009-clifford-oravec-of-tamboo\" data-href=\"https:\/\/www.indiehackers.com\/podcast\/009-clifford-oravec-of-tamboo\" rel=\"noopener noreferrer\" target=\"_blank\">I had a chat with the super awesome Courtland Allen<strong> on the kick-ass IndieHackers podcast.<\/strong><\/a>\u00a0If you\u2019re not familiar with <a href=\"https:\/\/www.indiehackers.com\/\" data-href=\"https:\/\/www.indiehackers.com\" rel=\"noopener noreferrer\" target=\"_blank\">Indie Hackers<\/a> or <a href=\"https:\/\/medium.com\/@csallen\" data-href=\"https:\/\/medium.com\/@csallen\" data-anchor-type=\"2\" data-user-id=\"35d2978e0b2f\" data-action-value=\"35d2978e0b2f\" data-action=\"show-user-card\" data-action-type=\"hover\" target=\"_blank\" rel=\"noopener noreferrer\">Courtland Allen<\/a>, you <strong>need<\/strong> to check them both out (and follow them!).<\/p><p id=\"b900\">But enough about me!<\/p><p id=\"eb31\">Let\u2019s have some <em>more<\/em> \u201cfun\u201d and pick up where we left off in <a href=\"https:\/\/medium.com\/@cliffordoravec\/ewww-your-startup-funnel-is-leaking-all-over-me-part-4-of-the-epic-guide-to-bootstrapping-d08e0db6b167\" data-href=\"https:\/\/medium.com\/@cliffordoravec\/ewww-your-startup-funnel-is-leaking-all-over-me-part-4-of-the-epic-guide-to-bootstrapping-d08e0db6b167\" target=\"_blank\" rel=\"noopener noreferrer\">Part 4 of The Epic Guide<\/a>.\u00a0(If this is your first foray with <em>The Epic Guide<\/em>, you may want to start with <a href=\"https:\/\/medium.com\/@cliffordoravec\/the-epic-guide-to-bootstrapping-a-saas-startup-from-scratch-by-yourself-part-1-4d834e1df8c1#.uut3kmbyv\" data-href=\"https:\/\/medium.com\/@cliffordoravec\/the-epic-guide-to-bootstrapping-a-saas-startup-from-scratch-by-yourself-part-1-4d834e1df8c1#.uut3kmbyv\" target=\"_blank\" rel=\"noopener noreferrer\">Part 1<\/a> and work your way back here\u200a\u2014\u200aafter all, it\u2019s an ongoing <em>epic<\/em>\u2026)<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/dumb-and-dumber-iou.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Look, see this? That's quit your job money. $10k MRR. Might wanna hang onto that one.<\/figcaption><\/figure><\/div><p id=\"33c7\"><strong>Stop pretending that whatever you\u2019re doing is somehow going to \u201cmake you money\u201d.<\/strong><\/p><p id=\"4ea0\">If you haven\u2019t been making mini \"fuck you money\" off of your MVP yet, it might be time to step back and <em>think<\/em> about what it is that you <em>have<\/em> been doing. Because one thing\u2019s for certain\u200a\u2014\u200aif you keep doing whatever it is you <em>are\u00a0<\/em>doing\u200a\u2014\u200ayou\u2019re just going to get more of the same \u201cresults\u201d.<\/p><p id=\"8d78\">Let me see if I can guess what\u2019s been going on here.<\/p><p id=\"2cd2\">You\u2019re probably in one of two camps at this point:<\/p><ol><li id=\"b44f\"><strong>You\u2019ve done absolutely nothing of any real or substantial value when it comes to marketing or selling your startup.<\/strong> Maybe you tried \u201cmarketing\u201d your startup on Reddit once or twice, sure. But since then? Not so much. And cold emails? Maybe you <em>thought<\/em> of sending that one that one time, but you kinda just never hit send and fucked around reading up on blockchain and cryptocurrencies instead (which gave you this <strong>great idea<\/strong> for a brand new startup, by the way!)<\/li><li id=\"b861\"><strong>You\u2019ve been going \u201ctransparent\u201d with your startup and writing blog posts about all of the geeky things you\u2019ve been wasting your time on.<\/strong>People seem to <em>love<\/em> your posts about how you\u2019re still at $0 MRR and how \u201chard it is\u201d to actually build a startup. You\u2019ve been getting some pretty good \u201ctraction\u201d on Twitter and Hacker Noon with your content, but you\u2019re not really reaching anyone in your target market (or on the off chance that you are, they\u2019re not forking over their credit card, so it\u2019s more or less the same). But hey! You\u2019re \u201cbuilding an audience\u201d, right?!<\/li><\/ol><p id=\"294a\">What\u2019s <strong>downright dangerous<\/strong> about finding yourself bunked up in either of these two camps is that\u200a\u2014\u200aeven though it <em>feels<\/em> like you\u2019re doing something\u200a\u2014\u200a<strong>you\u2019re just busying yourself with bullshit.<\/strong><\/p><p id=\"069b\">If there\u2019s one thing you take away from this, make sure it\u2019s this:<\/p><p><b>If you aren\u2019t reaching and resonating with actual people in your actual target market, you\u2019re just \u201cplaying startup\u201d.<\/b><\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/our-pets-heads.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">What the hell are we doing here? We gotta get out of this town!<\/figcaption><\/figure><\/div><p id=\"7a91\"><strong>To a place where the beer flows like wine.<\/strong><\/p><p id=\"a9e8\">I\u2019m going to let you in on a little secret here.<\/p><p id=\"7dce\">You want to know the <em>real<\/em> reason you\u2019re not getting anywhere?<\/p><p id=\"c510\">It\u2019s because your <em>goals<\/em> are all wrong.<\/p><p id=\"2b6d\">No, that John Denver was not full of shit. You\u2019ve just been looking in the wrong places.<\/p><p id=\"f8c4\">You see, if you\u2019re like most first-time founders, you\u2019ve probably set some (\u201creasonably\u201d sounding) goals for yourself like this:<\/p><ol><li id=\"b42d\">Get 100 visitors a day to my website.<\/li><li id=\"c20f\">Add 10 new people to my newsletter weekly.<\/li><li id=\"bfcc\">Get 1 new trial signup a day.<\/li><\/ol><p id=\"576f\">(and so on)<\/p><p id=\"207e\">The problem with these \u201cgoals\u201d\u200a\u2014\u200aand the reason that you\u2019re getting your ass kicked\u200a\u2014\u200ais that they are not, in fact, <em>goals<\/em>.<\/p><p id=\"87c0\"><strong>You<\/strong> <strong>cannot<\/strong> get 100 visitors a day to your website.<\/p><p id=\"cf70\"><strong>You cannot<\/strong> get 10 new people to sign up to your newsletter a day.<\/p><p id=\"4702\"><strong>You cannot <\/strong>get 1 new trial signup a day.<\/p><p id=\"4dee\">Because <strong>those are not goals.<\/strong><\/p><p id=\"f4b8\">Those are <strong>events<\/strong> that are based on <strong>what other people <em>choose<\/em> to do<\/strong>.<\/p><p id=\"3dc9\">And you have <strong>zero<\/strong> control over them.<\/p><p id=\"32c2\">So for you \u201cset a goal\u201d of getting 100 visitors a day\u200a\u2014\u200ayou might as well \u201cset a goal\u201d of winning the lottery (or finding a briefcase full of money).<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/crazy-pills.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">But I have \"traffic goals\" and \"MRR targets\" to hit!!!<\/figcaption><\/figure><\/div><p id=\"e8c8\">The first step on your way out of this is to realize that <strong>you can only set goals for things you have direct control over.<\/strong><\/p><p id=\"8269\">So let\u2019s fix this up a bit. Let\u2019s say that we have goals like the following:<\/p><ol><li id=\"eb34\">Find 10 potential customers a day and send cold emails to each of them.<\/li><li id=\"33d6\">Write at least 1 piece of content a day and promote it in at least 3 places.<\/li><li id=\"8a15\">Promote my startup in at least 5 different places a day.<\/li><li id=\"18ed\">Create and run a new ad for 1 week every week.<\/li><li id=\"664d\">Find at least 4 people talking about or asking questions about topics related to my startup a day and write them a detailed, personal answer.<\/li><\/ol><p id=\"061f\">Well, now.<\/p><p id=\"0c9b\">These <strong>we can control<\/strong>, can\u2019t we?<\/p><p id=\"3546\"><strong>You can control<\/strong> whether or not you choose to screw around watching the price of Ethereum bounce around or instead find yourself 10 people who might be interested in your product that you could then send 10 emails to.<\/p><p id=\"61e5\"><strong>You can control <\/strong>whether or not you decide to slack off on Slack all day or instead find yourself 4 people asking questions about something related to your startup and write them some useful and helpful responses.<\/p><p id=\"1ebf\">Now sure, you still have <strong>no control<\/strong> over the <em>outcomes<\/em> of those activities. Just because you send 10 cold emails does not mean you\u2019ll get 10 happy customers\u200a\u2014\u200aor even a single response. Just because you wrote an article does not mean you\u2019ll get 100 visitors\u200a\u2014\u200aor even one.<\/p><p id=\"42bb\"><strong>But that\u2019s not the point.<\/strong><\/p><p id=\"d292\">The point is that rather than get hung up on things you have no control over and to expect any kind of \u201cgoals\u201d you\u2019ve set around such things to \u201cmagically happen\u201d\u200a\u2014\u200athat you\u2019re instead focusing your energy and your efforts on things you can shape and influence\u200a\u2014\u200aand you\u2019re <em>holding yourself accountable for doing actual, real work.<\/em><\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/40-hour-work-week.png\" alt=\"\" class=\"\" \/><figcaption class=\"\">The good news? You can work on whatever you want during the second 40.<\/figcaption><\/figure><\/div><p id=\"57e1\"><strong>It\u2019s time to do work, son.<\/strong><\/p><p id=\"7747\">Okay, so let\u2019s say you\u2019re bought into this whole \u201cmeasure activities instead of goals\u201d business.<\/p><p id=\"afe8\">How exactly do we make money with that approach?<\/p><p id=\"5816\">First thing\u2019s first, you\u2019re going to have to do some <strong>work <\/strong>(I hear you Millennials hate that word, so sorry if I just shocked some of you off of your parents\u2019 couch while you were reading this).<\/p><p id=\"d5c3\">When you start out doing this, it\u2019s going to take some time to get the hang and feel of it all. Expect to put in some hours until things start to \u201cclick\u201d.<\/p><p id=\"f9d4\">But once they click?<\/p><p id=\"fbf6\">You could probably do most of this in an hour or less a day.<\/p><p id=\"e415\">For reals.<\/p><p id=\"2663\">Okay, so let\u2019s get to it.<\/p><p id=\"6399\">For this to work for you, you have to know what you want to get out of it.<\/p><p id=\"1fbc\">You have to know if you want to build traffic to your website to drive trials or opt-ins or if you want to close some cash-in-hand sales.<\/p><p id=\"04ab\">These things are your <strong>expected outcomes.<\/strong><\/p><p id=\"e456\">You <em>expect<\/em> that if you do these activities <em>right<\/em>, that you will have (at least some of) these <em>outcomes<\/em> as a result.<\/p><p id=\"7d40\">Now, you don\u2019t set goals on your outcomes at this stage (that comes later).<\/p><p id=\"4301\">At this stage, you just need to identify the thing you want at the end of the day.<\/p><p id=\"dd8c\">From there, you work your way <strong>backwards<\/strong> to determine <strong>what kind<\/strong> of activities you should be doing to <em>realize<\/em> those outcomes.<\/p><p id=\"63fe\">If you want to <strong>close some sales<\/strong>, you\u2019re going to be focused on hand-to-hand sales combat activities, typically starting from scratch in the \u201clead generation\u201d and \u201cprospecting\u201d stages, doing things such as:<\/p><ul><li id=\"d177\">Using search engines and social networks to find businesses\u200a\u2014\u200aand the individuals in them\u200a\u2014\u200awho seem to be \u201ca good fit\u201d for what you offer and adding them to your CRM<\/li><li id=\"e92f\">Building opt-in offers and lead magnets to capture leads<\/li><li id=\"34e0\">Doing cold outreach (aka cold emailing or cold calling)<\/li><li id=\"bc5c\">Performing educational outreach (webinars, etc.)<\/li><li id=\"0ef9\">Generating referrals through partnerships and networking<\/li><\/ul><p id=\"bfaf\">Your expected outcomes at this stage with these activities would be:<\/p><ul><li id=\"9dff\">Get responses from people you\u2019ve contacted<\/li><li id=\"a54a\">Set appointments\/meetings to talk with the people who responded<\/li><\/ul><p id=\"d31d\">If you want <strong>to build traffic traffic<\/strong> or <strong>get trial signups<\/strong>, you\u2019re going to be focused on marketing activities, such as:<\/p><ul><li id=\"2521\">Using search engines and social networks to find websites, forums, Slack groups, and even other social networks where people who would use your product go for information or entertainment<\/li><li id=\"6d73\">Promoting your startup on <em>relevant <\/em>websites your target market visits through educational (and entertaining!) content marketing, guest posts, recommendations, participating in forums, and even taking out ads<\/li><li id=\"f3c9\">Finding \u201crelated\u201d products or services that your target market uses that you can \u201ctie in\u201d your offering to through techniques such as integration marketing<\/li><li id=\"1fed\">Getting your product listed on <em>relevant<\/em> \u201clists\u201d\u200a\u2014\u200asuch as \u201cThe top 10 best tools for X\u201d as well as on <em>relevant<\/em> directories\u200a\u2014\u200asuch as Capterra and Product Hunt.<\/li><li id=\"e970\">Creating (and promoting!) <em>relevant <\/em>content that you can use to get backlinks and list for specific keywords to improve your site\u2019s SEO <em>for those keywords<\/em><\/li><li id=\"c9e9\">Taking out <em>relevant<\/em> ads on Google or anywhere else your target market is likely to search for what you have to offer<\/li><li id=\"54db\">Taking out <em>relevant <\/em>ads on Facebook, Twitter, or anywhere else your target market is likely to <em>be<\/em> and <em>be interested in<\/em> what you have to offer<\/li><li id=\"77be\">Answering questions on Q&amp;A sites such as Quora, Product Hunt Ask, and (especially!) on \u201cniche\u201d forums specific to your target market.<\/li><li id=\"cec2\">Finding <em>relevant<\/em> publications (old-school industry-specific ones or even just simply Medium ones) and submitting <em>relevant<\/em> content to them that educates and informs their readers about a topic they give a shit about (and which you\u2019re <em>perfectly<\/em> situated to help them with!)<\/li><li id=\"cda8\">Sponsoring <em>relevant<\/em> newsletters that your target market subscribes to<\/li><li id=\"e31d\">Doing interviews on <em>relevant<\/em> podcasts that your target market listens to about topics that your target market would find useful, interesting, and helpful<\/li><li id=\"1bed\">Engaging in <em>relevant<\/em> social media channels where your target market participates and consumes information and where you can <em>naturally<\/em>promote your articles, content, interviews, posts, answers, etc. to an audience that would actually <em>want to see and read those things<\/em><\/li><\/ul><p id=\"0994\">Your expected outcome at this stage of the game with these activities would simply be <strong>traffic\u200a\u2014\u200abut not just <em>any<\/em> traffic as you\u2019ll soon see.<\/strong> (You can worry about optimizing for conversions after you actually <em>have<\/em> traffic you can optimize.)<\/p><p id=\"8ab2\">And these are just ideas to get you started.<\/p><p id=\"9928\">There is <em>so much more<\/em> that you can do.<\/p><p id=\"68d9\">But start somewhere you must!<\/p><p id=\"bf8a\">So now you need to <strong>pick 2\u20133 things that you will do daily for a month <\/strong>(to start; once you get better at this, add more to your plate as you wish \ud83d\ude1c).<\/p><p id=\"8e9c\">[Ideally you\u2019ll want to pick activities that you <em>think<\/em> (<em>read: <\/em><strong><em>wild ass guess<\/em><\/strong>) will have the biggest potential payoff. It shouldn\u2019t be completely random.]<\/p><p id=\"fca8\">Let\u2019s say that your list looks like this:<\/p><ol><li id=\"16f0\">Find 5 businesses a day that are a fit for my offering and send each of them an email to qualify their interest in getting\/solving (this specific benefit\/pain point).<\/li><li id=\"35a1\">Find 3 people asking questions on a forum or Q&amp;A site specifically related to my product offering and give them the <em>best<\/em> answer they have ever seen to their question.<\/li><li id=\"fe7e\">Write 1 piece of \u201cquick\u201d content that talks about one <em>very <\/em>specific topic my target market is interested in in under 30 minutes and promote it in at least 3 different places (relevant Facebook Groups, Reddit, Medium, \u201clink\u201d sites, forums, Slack groups, Twitter, etc.)<\/li><\/ol><p id=\"b163\">Do you think you could find the time to do <em>that<\/em> in 1 to 2 hours a day, every day?<\/p><p id=\"9341\">No?<\/p><p id=\"d6f8\">Well, if you never <em>try<\/em>, then you\u2019re right\u200a\u2014\u200ayou\u2019ll never do it.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/2014-12-derek-zoolander12.0.jpeg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Better learn how to turn left, bruh.<\/figcaption><\/figure><\/div><p id=\"c4c9\"><strong>Up and to the right (very slowly)<\/strong><\/p><p id=\"8597\">With <em>everything you do<\/em>\u200a\u2014\u200aask yourself\u200a\u2014\u200a\u201cHow many outcomes can I reasonably expect this to get me?\u201d<\/p><p id=\"24bb\">(A little secret? No one knows.)<\/p><p id=\"9a6f\">So you just guess.<\/p><p id=\"e995\">And you write it down.<\/p><p id=\"9b71\">For example, if you write a piece of content, you might think you can get 10\u201350 people to read it if you post it to a specific link sharing site your target market frequents.<\/p><p id=\"b4b6\">Great. Let\u2019s write that down and build a spreadsheet (we\u2019re all nerds here, right?).<\/p><p id=\"356c\">On the far left, you write down the \u201ccategory\u201d of the activity\u200a\u2014\u200asuch as \u201cContent Marketing\u201d.<\/p><p id=\"1308\">Then, a cell over, you write down the <em>specific <\/em>activity you\u2019re planning on doing\u200a\u2014\u200a\u201cWrite a blog post about how to simply purchase ordering (plugging Startup Awesome).\u201d<\/p><p id=\"7bb5\">Then, a cell over from that, write down the <em>specific place <\/em>you\u2019re going to promote that thing (if applicable). (When dealing with promoting the same piece of content in different places, create rows for each place and separate estimates for each place.)<\/p><p id=\"bae2\">One more cell over, write down what type of outcome you\u2019re expecting\u200a\u2014\u200ain this case \u201cTraffic\u201d.<\/p><p id=\"b80f\">Then, one cell from that, write down how many expected outcomes you might get (your wild ass guess).<\/p><p id=\"5581\">Use one \u201csheet\u201d per month and record your daily activities for each month <em>religiously<\/em>.<\/p><p id=\"9313\">Then, as you actually <strong>execute <\/strong>on those activities (aka, that nasty word you hate\u200a\u2014\u200a\u201cwork\u201d), write down how many outcomes <em>you actually got <\/em>in a cell at the end of each row.<\/p><p id=\"edc0\">So something like this:<\/p><p id=\"cc4c\"><strong>Category | Activity | Place Promoted | Type of Outcome | Expected Result | Actual Result<\/strong><\/p><p id=\"2691\">Using something like Google Sheets (free!), it\u2019s pretty simple to get sums and totals from this information based on \u201cCategory\u201d, \u201cPlace Promoted\u201d, or \u201cType of Outcome\u201d so you can see what\u2019s working (and what\u2019s not).<\/p><p id=\"138e\">[Pro tip: Look for patterns.]<\/p><p id=\"815a\">You\u2019ll start to \u201cget a feel\u201d for how much you should <em>actually<\/em> expect from different kinds of activities and promotional placements. This is critical, as it helps you to dial in your (sometimes ridiculous) expectations. Once you\u2019re better \u201ccalibrated\u201d, you\u2019ll be able to make better judgment calls about where you should put your efforts and how much you should expect from them.<\/p><p id=\"545c\">Do this for a month.<\/p><p id=\"320a\">Don\u2019t expect everything you do to work well\u200a\u2014\u200aor even work at all.<\/p><p id=\"e3f6\">But keep at it.<\/p><p id=\"e4c2\">And at the end of the month, pick the winners.<\/p><p id=\"9585\">If you have certain Categories that are really paying off, <strong>double down on them.<\/strong><\/p><p id=\"2044\">If you have certain Categories that aren\u2019t moving the needle, <strong>dump them and replace them with something else.<\/strong><\/p><p id=\"946e\">The same thing with Placements\u200a\u2014\u200ause what works and ditch the rest.<\/p><p id=\"fd8e\">The next month, carry forward the winners and mix in the replacements.<\/p><p id=\"9531\">Lather, rinse, and repeat until finished.<\/p><p id=\"a8d9\">It\u2019s slog, but the little numbers add up (especially when you start \u201cdialing in\u201d your \u201cwhat works compass\u201d).<\/p><p id=\"6497\">To get the sales and the marketing impressions you need to build your startup, you need to either knock it out of the park in one fell swoop (you probably have a better chance of being struck by lightning) or gradually building up hundreds and thousands of cold outreaches and marketing impressions over time.<\/p><p id=\"a453\">In our example above, if you did that <em>exact<\/em> routine <em>every day<\/em>, after just one month, you will have sent 150 cold emails, have written 90 targeted answers (and plugs for your startup!), and have written 30 content pieces that you promoted in 90 different places.<\/p><p id=\"9741\">And that\u2019s just in one month.<\/p><p id=\"5879\">Imagine if you do this <em>every month.<\/em><\/p><p id=\"0437\">It\u2019s cause and effect.<\/p><p id=\"99d3\">If you don\u2019t <strong>do<\/strong>, you don\u2019t <strong>get.<\/strong><\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/zoolander-center-for-kids.png\" alt=\"\" class=\"\" \/><figcaption class=\"\">How can we be expected to teach founders to learn how to sell or market if they can't even fit their product inside their market?!<\/figcaption><\/figure><\/div><p id=\"9c2e\"><strong>But what if I don\u2019t know how to sell good or market good (or do other stuff good too)?!<\/strong><\/p><p id=\"a4e5\">Ahhh!<\/p><p id=\"bef1\">So here we are.<\/p><p id=\"d150\">You don\u2019t know how to sell or market good?<\/p><p id=\"7916\">Guess what?<\/p><p id=\"c808\">I don\u2019t want to hear your excuses! Your sales and marketing must be at least\u2026 three times bigger than this!<\/p><p id=\"4007\">(Okay, okay, I\u2019ll back up.)<\/p><p id=\"b767\">Here\u2019s the thing that I\u2019ll tell you that the startup gurus won\u2019t: There is no \u201cright\u201d or \u201cwrong\u201d answer when it comes to sales and marketing\u200a\u2014\u200athe only thing that matters is <em>what works.<\/em><\/p><p id=\"fcc4\">And one more thing I\u2019ll admit that they won\u2019t: I can\u2019t help you sell or market to your customer base\u200a\u2014\u200abecause I don\u2019t know who they are.<\/p><p id=\"e3f9\">There is no \u201cstandard\u201d way to do this (even if there are \u201cwell understood\u201d patterns and practices).<\/p><p id=\"8eb6\">Because the truth is that all the \u201ctemplates\u201d and \u201ccase studies\u201d in the world aren\u2019t going to help you if you don\u2019t understand your customer.<\/p><p id=\"6a8b\">That\u2019s because <em>everything you do<\/em>\u200a\u2014\u200ain sales or in marketing\u200a\u2014\u200a<em>must be based on, about, and specifically tailored for your buyer.<\/em><\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/sell-me-this-pen.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Go for it, Slick. I double dog dare you.<\/figcaption><\/figure><\/div><p id=\"651b\"><strong>It\u2019s about them\u200a\u2014\u200anot about you, stupid.<\/strong><\/p><p id=\"f3f1\">In order to craft <em>truly effective<\/em> sales or marketing strategies or tactics, <em>you must first know your buyer.<\/em><\/p><p id=\"2860\">You must know where they \u201ccongregate\u201d.<\/p><p id=\"8d73\">You must know <em>how they talk.<\/em><\/p><p id=\"042b\">You must know what they\u2019re <em>interested in.<\/em><\/p><p id=\"ccf0\">You must know what they <em>give a shit about.<\/em><\/p><p id=\"e548\">You must know <em>what kind of content they like to consume\u200a\u2014\u200aand share with their peers.<\/em><\/p><p id=\"f14f\">You must know who they <em>look up to, respect, and trust.<\/em><\/p><p id=\"c866\">You must know how they <em>like<\/em> to be \u201cengaged\u201d.<\/p><p id=\"4e09\">You must know <em>what they\u2019re trying to accomplish.<\/em><\/p><p id=\"5a6a\">And you must know <em>how they like to buy.<\/em><\/p><p id=\"99a7\">Then\u200a\u2014\u200aand only then\u200a\u2014\u200acan you <strong>hope<\/strong> to be able to \u201ccraft a message\u201d that <strong>resonates<\/strong> with them.<\/p><p id=\"f1af\">Your goal\u200a\u2014\u200ain any copy that you use\u200a\u2014\u200amust be to establish rapport and to (subliminally) communicate with them that \u201cI get you and I get what you\u2019re trying to do and all the <em>shit<\/em> that you\u2019ve been going through trying to get that done.\u201d<\/p><p id=\"d95c\">You do that by <em>talking like they talk.<\/em><\/p><p id=\"2cd7\">You do that by <em>referencing things that only an \u201cinsider\u201d would \u201cget\u201d.<\/em><\/p><p id=\"8735\">You do that by <em>reaching them where they\u2019re open to suggestion and influence\u2014 the places where they\u2019re asking questions or trying to learn.<\/em><\/p><p id=\"c001\">Then\u200a\u2014\u200aand only then\u200a\u2014\u200acan you <strong>offer a suggestion<\/strong> that they will register and <em>consider<\/em>.<\/p><p id=\"7e42\">How do you \u201clearn\u201d how to do this effectively?<\/p><p id=\"d84d\"><strong>You practice empathy.<\/strong><\/p><p id=\"2d83\">You practice \u201cputting yourself in their shoes.\u201d<\/p><p id=\"964f\">You practice \u201cfeeling\u201d how they feel about what they\u2019re trying to accomplish.<\/p><p id=\"29ec\">You practice \u201cfeeling\u201d how <em>they<\/em> \u201cfeel\u201d when they see and read your cold email\/content marketing\/social media posts, etc.<\/p><p id=\"5c9b\">Basically, you project your mind into theirs until you \u201cget them\u201d.<\/p><p id=\"6fbe\">You do this by <em>immersing yourself<\/em> in their world until you \u201cget a feel\u201d for \u201chow they think\u201d by reading their comments, posts, frustrations, and insights.<\/p><p id=\"8fd7\">Until you somehow vicariously \u201cbecome one of them\u201d.<\/p><p id=\"d9ba\">And then you just <strong>give them what they want.<\/strong><\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/dumb-and-dumber-airport-limo.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">Yes, it's stupidly obvious. But sometimes you need to be.<\/figcaption><\/figure><\/div><p id=\"283a\"><strong>Be an idiot (when it\u2019s smart to be one)<\/strong><\/p><p id=\"5db3\">Chances are, if you\u2019re like most developers, you\u2019re going to get <strong>bored<\/strong> with this.<\/p><p id=\"8e61\">After all, it\u2019s basically saying <em>the same thing <\/em><strong>over and over again.<\/strong><\/p><p id=\"bdff\">In a hundred different places.<\/p><p id=\"e83e\">But that\u2019s what you have to do.<\/p><p id=\"bb77\">I do it, and so should you.<\/p><p id=\"5d54\">Never assume people \u201cget\u201d or \u201cunderstand\u201d what you can do for them.<\/p><p id=\"dba6\">Spell it the fuck out.<\/p><p id=\"6494\">And yes. You\u2019re going to be repeating yourself. A lot.<\/p><p id=\"7202\">Get used to it.<\/p><p id=\"3dcd\">Every article, every post, every answer\u200a\u2014\u200athey\u2019re all meant to reach <strong>new people who never heard of you before.<\/strong><\/p><p id=\"a6b1\">So you have to <strong>tell them who you are and what you do.<\/strong><\/p><p id=\"7a70\"><strong>Every time.<\/strong><\/p><p id=\"3e9b\">Here\u2019s an example of <a href=\"https:\/\/www.quora.com\/What-should-I-measure-to-get-user-behaviour-results-on-a-one-page-web-design\" data-href=\"https:\/\/www.quora.com\/What-should-I-measure-to-get-user-behaviour-results-on-a-one-page-web-design\" rel=\"noopener noreferrer\" target=\"_blank\">an answer I posted on Quora<\/a> for a question that I felt <a href=\"https:\/\/gettamboo.com\/\" data-href=\"https:\/\/gettamboo.com\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>Tamboo<\/strong><\/a> would be able to help with:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/tamboo-quora-answer.png\" alt=\"\" \/><\/figure><\/div><p id=\"9e27\">Now, there\u2019s a few things going on here that I think I should point out, as I think they\u2019ll help you \u201cgrok\u201d how to sell and market \u201cbetter\u201d:<\/p><ol><li id=\"b5f6\">Use some form or varient of an <a href=\"https:\/\/en.wikipedia.org\/wiki\/AIDA_%28marketing%29\" data-href=\"https:\/\/en.wikipedia.org\/wiki\/AIDA_(marketing)\" rel=\"noopener noreferrer\" target=\"_blank\">AIDA<\/a> construct with whatever you write.<\/li><li id=\"ff4e\">Use visuals where you can and where it makes sense. Humans are inherently visual and their attention will go where they pretty pictures are. Make sure you have pretty pictures to show.<\/li><li id=\"fd2b\">Show proof. Don\u2019t just tell\u200a\u2014\u200a<strong>show.<\/strong><\/li><li id=\"56d2\">Use a conversational tone as if you were talking to a friend\u200a\u2014\u200adon\u2019t be overly \u201cprofessional\u201d. <strong>Professionalism is for amateurs.<\/strong><\/li><li id=\"684f\">Be yourself, but be unique\u200a\u2014\u200adon\u2019t \u201cblend in\u201d. You <strong>must<\/strong> have a different \u201cvoice\u201d than everyone else. You must \u201cstand out\u201d in some way. <strong>Don\u2019t be a robot.<\/strong><\/li><li id=\"ef28\">Be clear about what you\u2019re saying\u200a\u2014\u200aexplain it like they\u2019re 5 (without being patronizing). Don\u2019t assume they know as much as you do, and don\u2019t make it intimidating to follow along. Also, don\u2019t assume they\u2019re total morons at the same time.<\/li><li id=\"f518\">Don\u2019t be pushy\u200a\u2014\u200ajust be suggestive. Craft your CTAs using \u201crecommendations\u201d and \u201csuggestions\u201d. Don\u2019t tell them what to do\u200a\u2014\u200asimply suggest how they can take what you\u2019ve said to the next logical step.<\/li><li id=\"3fac\">Don\u2019t be skimpy \u2014whatever you say must be something they care about and must be better than everything else they\u2019ve been reading. You must <strong>one-up everyone else in some way\u200a\u2014\u200aalways.<\/strong><\/li><li id=\"5b1d\">Don\u2019t ever forget\u200a\u2014\u200a<strong>IT\u2019S ABOUT THEM. <\/strong>Answer their question. Give them the information they came for. And only after you\u2019ve done that, go in for the soft \u201cask\u201d.<\/li><\/ol><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/seinfeld-who-are-these-people.jpg\" alt=\"\" class=\"\" \/><figcaption class=\"\">And what's the deal with lamp shades anyways?<\/figcaption><\/figure><\/div><p id=\"907a\"><strong>One more secret: You don\u2019t want \u201ctraffic\u201d.<\/strong><\/p><p id=\"3365\"><em>Are you crazy?!<\/em><\/p><p id=\"add9\">Don\u2019t get me wrong\u200a\u2014\u200atraffic is important.<\/p><p id=\"db31\">Without it, you\u2019re not going to go very far.<\/p><p id=\"a74e\">But you can have <strong>a lot of bad traffic<\/strong> and it\u2019s not going to get you anywhere.<\/p><p id=\"a282\">Pop quiz: Would you rather have 1,000 visitors at a 1% conversion rate or 100 visitors at a 10% conversion rate?<\/p><p id=\"589d\">On the surface, they\u2019re same thing: You get 10 conversions at the end of both.<\/p><p id=\"eaa6\">But the second one (100 visitors at 10%) is <strong>easier to grow<\/strong>.<\/p><p id=\"e804\">To get to <em>20 customers<\/em>, it\u2019s the difference between having to get 2,000 visitors as opposed to just 200.<\/p><p id=\"2d43\">Ahhh!<\/p><p id=\"239f\">And that\u2019s the dirty little secret\u200a\u2014\u200a<strong>you need targeted traffic.<\/strong><\/p><p id=\"0b13\">You don\u2019t want just <em>any<\/em> traffic \u2014after all, visitors != conversions.<\/p><p id=\"08a7\">You want <em>traffic from people who give a shit about what you\u2019re offering.<\/em><\/p><p id=\"ad15\">Not people who are just \u201ccurious\u201d because they saw your latest startup blog post.<\/p><p id=\"d608\">You want people who you <em>resonated<\/em> with\u200a\u2014\u200awho think \u201cyou get me!\u201d and are <em>genuinely interested in your offering.<\/em><\/p><p id=\"eb00\">So how do you do that?<\/p><p id=\"09fb\">You find ways to send <strong>warm traffic<\/strong> to your homepage (or if you\u2019re super slick, to a custom built landing page specific for the type of traffic you\u2019re sending it).<\/p><p id=\"5387\">Basically, you don\u2019t try to sell on your marketing site. <strong>You try to close.<\/strong><\/p><p id=\"64bc\">And you can only do that if you\u2019ve <strong>already had a conversation with the people going there.<\/strong><\/p><p id=\"bc19\">Whatever article, blog post, forum discussion, podcast interview, or Q&amp;A answer you put out there needs to <strong>warm up your prospect for your offer.<\/strong><\/p><p id=\"40db\">You need to \u201chook them\u201d with that \u201cexternal\u201d content\u200a\u2014\u200aand then\u200a\u2014\u200aand only then\u200a\u2014\u200ado you encourage them to go to your marketing site to take the next logical step.<\/p><p id=\"e117\">This is <strong>warm traffic.<\/strong><\/p><p id=\"b5e0\">They are primed. They are ready. Because you\u2019ve done the work to get them to that state\u200a\u2014\u200aand that\u2019s why they\u2019re <strong>more likely<\/strong> to take whatever action it is you want them to take.<\/p><p id=\"3b61\">A lot better than just \u201cspraying and praying\u201d with whatever traffic you can get your hands on, right?<\/p><p id=\"597a\">And with that final thought for you to ponder on, I\u2019m going bring this installment of <em>The Epic Guide <\/em>to a close.<\/p><p id=\"18df\">In the next Part, we\u2019re going to go in depth about why your website sucks and why it\u2019s costing you customers (enjoy!).<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/Do-You-Realize-What-You-Have-Done-Funny-Jim-Carrey-Meme-Image.jpeg\" alt=\"\" class=\"\" \/><figcaption class=\"\">My friend here is a little slow!<\/figcaption><\/figure><\/div><p id=\"1418\"><strong>I almost forgot!<\/strong><\/p><p id=\"e346\">If you love <em>The Epic Guide<\/em> and want even more down-and-dirty, no-frills, battle-tested startup warfare strategy and tactics, <a href=\"http:\/\/cliffordoravec.com\/the-epic-guide-to-bootstrapping-a-startup\" data-href=\"http:\/\/cliffordoravec.com\/the-epic-guide-to-bootstrapping-a-startup\" rel=\"noopener noreferrer\" target=\"_blank\">be sure to check out <em>The Epic Guide: The Book!<\/em><\/a><\/p><p id=\"3197\">Or, if you want to hear my voice inside of your head (I\u2019m not judging!) in what could only be comporable to Samuel L. Jackson\u2019s character in <em>Pulp Fiction<\/em>reading you startup bedtime stories, <a href=\"https:\/\/saasingear.com\/\" data-href=\"https:\/\/saasingear.com\" rel=\"noopener noreferrer\" target=\"_blank\">sign up for my upcoming podcast (aptly titled) SaaS In Gear!<\/a><\/p><p id=\"0fda\">And of course, if you\u2019ve been working hard on building your startup or online business but need a little \u201chelp\u201d to understand what your website visitors are <em>really <\/em>doing, <a href=\"https:\/\/gettamboo.com\/\" data-href=\"https:\/\/gettamboo.com\" rel=\"noopener noreferrer\" target=\"_blank\">don\u2019t hesitate to give Tamboo a try on your website<\/a>. You\u2019ll get all the secret tools marketers like me use to increase conversions and create even more potent sales pages\u200a\u2014\u200aincluding funnels, visitor recordings, and heatmaps.<\/p><p id=\"90a7\">And last but not least!<\/p><p id=\"58d0\">Be sure to follow me on <a href=\"https:\/\/medium.com\/@cliffordoravec\" data-href=\"https:\/\/medium.com\/@cliffordoravec\" target=\"_blank\" rel=\"noopener noreferrer\">Medium<\/a> and on <a href=\"https:\/\/twitter.com\/cliffordoravec\" data-href=\"https:\/\/twitter.com\/cliffordoravec\" rel=\"noopener noreferrer\" target=\"_blank\">Twitter<\/a> at @cliffordoravec and give me a shout out if you like what I\u2019m doing here!<\/p>","created_at":"2018-11-07T16:53:46+0000","updated_at":"2018-11-07T16:53:46+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":22,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/ucpi8JlaUVUCbGPzfikMLvYa7Z23\/@cliffordoravec%7C8f2df4804c\/zoolander-files-in-computer.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":614,"title":"Is it Windy in Here?","url":"https:\/\/heredragonsabound.blogspot.com\/search?updated-max=2016-10-24T10:11:00-07:00&max-results=7&start=44&by-date=false","content":"<div class=\"post-body entry-content\" id=\"post-body-2114242896415638596\">\nOne of the things I'd like to add to the map generation is realistic biomes -- where grass, forest, desert, etc. naturally occurs on the map.\u00a0 The biome in a location is largely a function of the amount of water available and the temperature.\u00a0 This is often illustrated with a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Biome#Whittaker_.281962.2C_1970.2C_1975.29_biome-types\">Whittaker diagram<\/a>.<p><\/p><div class=\"separator\">\n<a href=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/2\/29\/PrecipitationTempBiomes.jpg\"><img border=\"0\" height=\"388\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/2\/29\/PrecipitationTempBiomes.jpg\" width=\"400\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nTemperature is pretty straightforward (warmer near the equator and colder near the poles), but the amount of precipitation is more complex.\u00a0 Martin O'Leary's map generator just has a fixed amount of precipitation at every location.\u00a0 (It's only used to generate rivers.)\u00a0 My notion of an interesting map has a variety of biomes, so I wanted something more complex.\u00a0 I could use <a href=\"https:\/\/heredragonsabound.blogspot.com\/2016\/10\/is-it-noisy-in-here.html\">noise<\/a> to create the precipitation pattern, but I also want the map to be plausible.\u00a0 Using noise could lead to some odd maps: jungle next to deserts, an isolated pocket of heavy rainfall deep in a mountain range, etc.\u00a0 The human mind is pretty good at rationalizing patterns, but I wanted something a bit more realistic.\u00a0 I decide to simulate a very simplified rainfall model:\u00a0 moisture enters the atmosphere by evaporating from the oceans, is blown onto land by the trade winds, and precipitates out where conditions are favorable.<p>The first step in creating this rainfall model is wind.\u00a0 Real wind models are of course very complex, but my concern is primarily with how the surface winds are affected by the land geography.\u00a0 At a very simple level, this is just how the rise and fall of the land changes the direction and speed of the wind.\u00a0 I experimented with a variety of different approaches, but in the end settled on two simple rules:\n<br \/><\/p><ol><li>\u00a0Wind slows down when going uphill and speeds up when going downhill.<\/li>\n<li>\u00a0Wind turns away from obstructions.<\/li>\n<\/ol>\nThe simulation of the wind is a flow model.\u00a0 The wind in a location is represented a vector of the wind force and direction.\u00a0 At every time step of the model, the wind in every location is propagated forward based upon its direction.\u00a0 If the new location for the wind is higher, then some of the wind is turned aside while some of the wind continues upward but at less force.\u00a0 How the wind splits depends upon the steepness and direction of the slope.\u00a0 Conversely, wind going downward increases in force and turns towards the downward slope.\u00a0 All the wind entering a location is then summed to get the new wind force and direction for that location.\u00a0 This repeats until all the wind has propagated across the map and disappears out the edges.<p>Here is an example of wind interacting with the surface geography:<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/1.bp.blogspot.com\/-_GogKs2QbK4\/WADuWZSYhwI\/AAAAAAAALEM\/1wZisiVBoMwBPTuZqywSx4xFGvLZk6IOQCLcB\/s1600\/ddtABgd.jpg\"><img border=\"0\" height=\"382\" src=\"https:\/\/1.bp.blogspot.com\/-_GogKs2QbK4\/WADuWZSYhwI\/AAAAAAAALEM\/1wZisiVBoMwBPTuZqywSx4xFGvLZk6IOQCLcB\/s400\/ddtABgd.jpg\" width=\"400\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThe red arrows represent the wind vectors at each location.\u00a0 In this example, the island rises up out of the sea and presents an obstacle to the incoming wind.\u00a0 The wind splits around the island and the wind that goes over the island is slowed.\u00a0 The result is a \"wind shadow\" on the leeward side of the island.\u00a0 The peninsula has a similar effect and the two features funnel the wind between them.\u00a0 (Note that there's actually wind everywhere; for display purposes I'm not showing arrows where the wind is very weak.)<p><\/p><div class=\"separator\">\n<a href=\"https:\/\/2.bp.blogspot.com\/-j80HVW8erXM\/WADwHEiobEI\/AAAAAAAALEY\/O9cKwhoUvqwKQXYNL0oLnc4DiWy6bJh-wCLcB\/s1600\/reRp7lY.jpg\"><img border=\"0\" height=\"640\" src=\"https:\/\/2.bp.blogspot.com\/-j80HVW8erXM\/WADwHEiobEI\/AAAAAAAALEY\/O9cKwhoUvqwKQXYNL0oLnc4DiWy6bJh-wCLcB\/s640\/reRp7lY.jpg\" width=\"590\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nHere's a more zoomed out view, showing the wind patterns up a mountainous coast.\u00a0  \nThe high hills along the sea make it difficult for the wind to reach the interior.\u00a0 Along the southern part of the coast the fairly steep shore turns the \nwind northward.  It flows up the coast until it enters an inlet, where \nthe land to the north turns it inward.  It flows up the inlet until it \nis forced up the slopes and onto the land.  In the upper left you can \nsee how the wind tends to flow up the rills and valleys where the path \nis easiest.  Once upon onto the generally flat land above, the wind \nstarts to spread out.<p>Tweaking the rules on how slopes affect the wind can lead to stronger and\/or more chaotic patterns.<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/4.bp.blogspot.com\/-nT_nRASQXuc\/WADwtCPJj1I\/AAAAAAAALEc\/lUVxOp1r2HEa9_sjSdDhkgPyL9wMt1UsgCLcB\/s1600\/3wpKZEp.jpg\"><img border=\"0\" height=\"584\" src=\"https:\/\/4.bp.blogspot.com\/-nT_nRASQXuc\/WADwtCPJj1I\/AAAAAAAALEc\/lUVxOp1r2HEa9_sjSdDhkgPyL9wMt1UsgCLcB\/s640\/3wpKZEp.jpg\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThis picture illustrates how the wind can form very distinct and complex patterns.\u00a0 This map has gentle slopes and a number of large rounded hills.  Wind \nhits the land almost directly, where it gets filtered by the coast into a\n number of strong streams.  These flow inland, generally following the \ntopography, turning, splitting, and rejoining.  In places the prevailing\n wind blows at 90 degrees or more to the original trade wind.  Weaker \nwinds (not depicted here) can be quite chaotic, flowing every which way.<p>I'm generally pleased with the wind model -- it generates complex and plausible results.\u00a0 The real drawback is that it is fairly slow to run.\u00a0 Because the wind vectors can split into neighboring locations and because a map based on Voronoi partitions is irregular, it can take thousands of time steps for wind to blow completely across a map with rough terrain.\u00a0 At the moment, the wind model is the slowest step in the map generation.\n<\/p><div>\n<\/div><div class=\"post-body entry-content\" id=\"post-body-5466769363956167420\">\nMany of the procedural map generators you'll find on the Internet generate maps of islands, like this example from <a href=\"http:\/\/www-cs-students.stanford.edu\/~amitp\/game-programming\/polygon-map-generation\/\">Red Blob Games<\/a>:<p><\/p><div class=\"separator\">\n<a href=\"http:\/\/www-cs-students.stanford.edu\/~amitp\/game-programming\/polygon-map-generation\/voronoi-land-water.png\"><img border=\"0\" height=\"320\" src=\"https:\/\/lh3.googleusercontent.com\/proxy\/cXkhNOKoP8BK_gKidtlSr_8LUa1ZXQo9B59f48h7oAww52OdHkjNl0cLO8GfXTVcYm54Mxx8HJdiqj3Sqe09TYxXc_gtQ-Jmkc9YZwkNY2bpAHaVf1Q1bt5V8Pd6qF6ztVTrhWpmONU7geHNbmxTJDo63pOODdYq9k7t3rM=s0-d\" width=\"320\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nIslands are nice because they're self-contained, and having all the edges of the map as water avoids problems with land running off the map.\u00a0 Island maps are also handy for something like a role-playing campaign because they set a natural boundary to the story.\u00a0 But Martin O'Leary's <a href=\"http:\/\/mewo2.com\/notes\/terrain\/\">fantasy map generator<\/a> creates maps that look like a portion of the coastline of a bigger country, like this early example from my map generator:<br \/><div class=\"separator\">\n<a href=\"https:\/\/3.bp.blogspot.com\/--210VvlZTiY\/V_lafwCYhRI\/AAAAAAAALA8\/7KyzHK3MphQTZe5Kn1sUCfVgFCRKhMSRwCLcB\/s1600\/Image8.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/3.bp.blogspot.com\/--210VvlZTiY\/V_lafwCYhRI\/AAAAAAAALA8\/7KyzHK3MphQTZe5Kn1sUCfVgFCRKhMSRwCLcB\/s640\/Image8.png\" width=\"620\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThese maps look pretty authentic.\u00a0 But although I like the look of these maps, I also wanted to be able to generate island maps.\u00a0 But Martin's generator doesn't do islands. So how do you generate an island, anyway?<p>To start with, you can simply plop a blob down in the middle of the map.<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/3.bp.blogspot.com\/-VSQ1peA7wLY\/V_mLCBD3_lI\/AAAAAAAALBM\/6ZcTsAXbRtQQmdx8_TdKpfYymu5EYA23wCLcB\/s1600\/Image2.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/3.bp.blogspot.com\/-VSQ1peA7wLY\/V_mLCBD3_lI\/AAAAAAAALBM\/6ZcTsAXbRtQQmdx8_TdKpfYymu5EYA23wCLcB\/s640\/Image2.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThat's certainly an island, but it's a pretty unsatisfying sort of island.\u00a0 A more interesting island should have a more irregular (but natural-looking) shape, possibly with some outlying pieces.\u00a0 I don't really care about perfectly simulating the natural processes that lead to islands.\u00a0 I just want an island shape that is interesting and plausible.\u00a0 So how can we generate an <i>interesting<\/i> island?<p>Let's back up a step.\u00a0 Before we worry about generating an interesting island, let's consider how we can generate an interesting landscape -- without worrying about whether it is an island.\u00a0 One common approach is to use <a href=\"https:\/\/flafla2.github.io\/2014\/08\/09\/perlinnoise.html\">noise<\/a>.\u00a0 A <a href=\"http:\/\/libnoise.sourceforge.net\/tutorials\/tutorial5.html\">common technique<\/a> in <a href=\"http:\/\/www.redblobgames.com\/articles\/noise\/introduction.html\">procedural generation<\/a> is to create a heightmap based on a noise function, usually overlayed at several frequencies.\u00a0 The result is not always entirely plausible, but our brains interpret it as \"natural looking\":<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/2.bp.blogspot.com\/-giX5OAT--Ew\/V_mWDE-j1nI\/AAAAAAAALBg\/9PZSy4mNz2UzvxywLQACbNhIv3wM8V23gCLcB\/s1600\/Image3.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/2.bp.blogspot.com\/-giX5OAT--Ew\/V_mWDE-j1nI\/AAAAAAAALBg\/9PZSy4mNz2UzvxywLQACbNhIv3wM8V23gCLcB\/s640\/Image3.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\n(That view, by the way, uses the 3-D viewer from <a href=\"https:\/\/callumprentice.github.io\/apps\/webgl_terrain\/index.html\">this page<\/a>. \u00a0 You can play around on that page generating noise-based landscapes.)<p>Unfortunately, the landscapes generated that way aren't often islands:<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/4.bp.blogspot.com\/-uHO2PYw9zjc\/V_mWDCcTmxI\/AAAAAAAALBc\/VOErDmEf82IjxN_80na8QyHGtX1r1bJQACLcB\/s1600\/Image4.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/4.bp.blogspot.com\/-uHO2PYw9zjc\/V_mWDCcTmxI\/AAAAAAAALBc\/VOErDmEf82IjxN_80na8QyHGtX1r1bJQACLcB\/s640\/Image4.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nA simple way to turn these landscapes into islands is to apply a circular mask to the landscape.\u00a0 Anything outside the mask is pushed down into the sea. Voila!\u00a0 An island:<p><\/p><div class=\"separator\">\n<a href=\"https:\/\/4.bp.blogspot.com\/-A8V2aN6sPQo\/V_mZEUn34UI\/AAAAAAAALBw\/yC0uzQbVpOg2sZAcd83MTvbBxm319gOKACLcB\/s1600\/Image6.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/4.bp.blogspot.com\/-A8V2aN6sPQo\/V_mZEUn34UI\/AAAAAAAALBw\/yC0uzQbVpOg2sZAcd83MTvbBxm319gOKACLcB\/s640\/Image6.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/><div class=\"separator\">\n<a href=\"https:\/\/2.bp.blogspot.com\/-6fW-XEShz3E\/V_mZETe-jVI\/AAAAAAAALBs\/hMMg7Ee3MtMC0StWAddEGRW2bAWBCYYawCEw\/s1600\/Image5.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/2.bp.blogspot.com\/-6fW-XEShz3E\/V_mZETe-jVI\/AAAAAAAALBs\/hMMg7Ee3MtMC0StWAddEGRW2bAWBCYYawCEw\/s640\/Image5.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThat's a little better -- it's definitely an island, and has some interest within the boundaries of the island -- but the circular outline is too apparent.\u00a0 A circle just isn't a very interesting island mask, largely because it's so obviously unnatural.\u00a0 We know how to create an interesting landscape using noise, so our problem is really how to create an interesting island mask.<p>An initial idea is to make the mask more interesting by pushing the circle inward in some places, and outward in other places, in some sort of natural-looking way:<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/4.bp.blogspot.com\/-sQyi6p7CP2U\/V_mbkkapoaI\/AAAAAAAALB8\/qeE0XgzcPwcsCbMAm6foiAAm1EYt5vIIgCLcB\/s1600\/Image7.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/4.bp.blogspot.com\/-sQyi6p7CP2U\/V_mbkkapoaI\/AAAAAAAALB8\/qeE0XgzcPwcsCbMAm6foiAAm1EYt5vIIgCLcB\/s640\/Image7.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nThat's an improvement -- still a blob, but at least more interesting than a perfect circle.\u00a0 So how can we do this pushing in and out effect?<p>The approach I took was to create a circular island and then <i>perturb<\/i> it.\u00a0 The idea is to randomly move all the pieces of the island around -- to perturb their locations.\u00a0 In pseudo-code it might look something like this:<br \/><\/p><pre class=\"brush: js\">for (x = 0; x &lt; mapWidth; x++) {\n       for (y = 0; y &lt; mapHeight; y++) {\n           if (circularMask(x,y,radius) == true) {\n                 \/\/ Circular mask says to put some land at\n                 \/\/ (x,y).  Instead, let's perturb that\n                 \/\/ location and put land in the perturbed\n                 \/\/ location instead.\n                 x = x + randomOffset();\n                 y = y + randomOffset();\n                 world[x, y] = land;\n           }\n       }\n    }<\/pre>\n<br \/>\nLet's try it:<p><\/p><div class=\"separator\">\n<a href=\"https:\/\/1.bp.blogspot.com\/-U3kTEHDzk84\/V_riYGJoIpI\/AAAAAAAALCU\/DjP2u2qsgHgFFJot8lnziJVb8-IZtNXxQCLcB\/s1600\/Image1.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/1.bp.blogspot.com\/-U3kTEHDzk84\/V_riYGJoIpI\/AAAAAAAALCU\/DjP2u2qsgHgFFJot8lnziJVb8-IZtNXxQCLcB\/s640\/Image1.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nWell, that doesn't look <i>quite <\/i>right.\u00a0 The problem is that we're just randomly throwing every bit of the island around the map.\u00a0 Some of it ends up in the center of the map, but the rest goes every which way.\u00a0 We want neighboring parts of the island to move together -- so that if one piece of the shore gets pushed out, so do its neighbors.\u00a0 Instead of moving each piece of the island independently, we want to move pieces that are near each other approximately the same way.<p>Luckily for us, that's exactly what <a href=\"https:\/\/heredragonsabound.blogspot.com\/2016\/10\/is-it-noisy-in-here.html\">noise<\/a> does.\u00a0 The value of the noise function depends upon the parameters you give it and the value varies smoothly as those parameters change.\u00a0 If we use our x,y coordinates as the parameters, then nearby points will be similarly perturbed.\u00a0 So let's revise our code to use noise to perturb the island locations:<br \/><\/p><pre class=\"brush: js\">    for (x = 0; x &lt; mapWidth; x++) {\n       for (y = 0; y &lt; mapHeight; y++) {\n           if (circularMask(x,y,radius) == true) {\n                 x = x + noise(x,y);\n                 y = y + noise(x,y);\n                 world[x, y] = land;\n           }\n       }\n    }<\/pre>\n<br \/>\nLet's try it:<br \/><div class=\"separator\">\n<a href=\"https:\/\/4.bp.blogspot.com\/-9F4Eucy60XM\/V_rkOY5LbVI\/AAAAAAAALCg\/XOnX510Gr00ayEKy_2q1T8SwQkzaXzqWgCLcB\/s1600\/Image8.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/4.bp.blogspot.com\/-9F4Eucy60XM\/V_rkOY5LbVI\/AAAAAAAALCg\/XOnX510Gr00ayEKy_2q1T8SwQkzaXzqWgCLcB\/s640\/Image8.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nAlthough this looks much better, there's a (possibly non-obvious) problem.\u00a0 In the algorithm above, I'm adding the same noise to both the x and the y coordinates.\u00a0 So points will only be perturbed along a diagonal -- the same distance and direction in both x and y.\u00a0 To address this, we need to modify the algorithm so that we have separate noise sources for the two dimensions:<p><\/p><pre class=\"brush: js\">    noiseX = new NoiseGenerator();\n    noiseY = new NoiseGenerator();\n    for (x = 0; x &lt; mapWidth; x++) {\n       for (y = 0; y &lt; mapHeight; y++) {\n           if (circularMask(x,y,radius) == true) {\n                 x = x + noiseX(x,y);\n                 y = y + noiseY(x,y);\n                 world[x, y] = land;\n           }\n       }\n    }<\/pre>\n<br \/><div class=\"separator\">\n<a href=\"https:\/\/2.bp.blogspot.com\/-_uSi0PdkUJk\/V_wONswt0PI\/AAAAAAAALCw\/-YBAzVxrwgEvd7A6U2i8KpbWG4CcmTGRACLcB\/s1600\/Image9.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/2.bp.blogspot.com\/-_uSi0PdkUJk\/V_wONswt0PI\/AAAAAAAALCw\/-YBAzVxrwgEvd7A6U2i8KpbWG4CcmTGRACLcB\/s640\/Image9.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThe cool thing about generating islands this way is that we can create quite a bit of interesting diversity just by tweaking the noise we use to perturb the island.\u00a0 For example, if we use a low-frequency noise, the perturbations are smooth:<br \/><div class=\"separator\">\n<a href=\"https:\/\/1.bp.blogspot.com\/-dC66mCO4JRs\/V_wPnyiYPkI\/AAAAAAAALDA\/_79EzVJWtdYapFmtbHusVPBmGeJJGaUZACLcB\/s1600\/Image12.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/1.bp.blogspot.com\/-dC66mCO4JRs\/V_wPnyiYPkI\/AAAAAAAALDA\/_79EzVJWtdYapFmtbHusVPBmGeJJGaUZACLcB\/s640\/Image12.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n\u00a0If we use a high-frequency noise, the perturbations are jagged:<br \/><div class=\"separator\">\n<a href=\"https:\/\/2.bp.blogspot.com\/-f6PkZYvtW1M\/V_wPatjQZaI\/AAAAAAAALC8\/7J2fS0ofKfk16rOLysb9ER_zds50BIRzwCLcB\/s1600\/Image11.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/2.bp.blogspot.com\/-f6PkZYvtW1M\/V_wPatjQZaI\/AAAAAAAALC8\/7J2fS0ofKfk16rOLysb9ER_zds50BIRzwCLcB\/s640\/Image11.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nWe can also vary the magnitude of the perturbations.\u00a0 A low frequency perturbation with a stronger magnitude gives us a stretched out blob:<br \/><div class=\"separator\">\n<a href=\"https:\/\/3.bp.blogspot.com\/-apo1biKfzt4\/V_wgrhXRDSI\/AAAAAAAALDQ\/pFcfWv6WuyEHZcQiYvPe_9flJTcfi6WoQCLcB\/s1600\/Image13.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/3.bp.blogspot.com\/-apo1biKfzt4\/V_wgrhXRDSI\/AAAAAAAALDQ\/pFcfWv6WuyEHZcQiYvPe_9flJTcfi6WoQCLcB\/s640\/Image13.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nA high frequency perturbation with a stronger magnitude gives us a wildly chaotic coastline:<br \/><div class=\"separator\">\n<a href=\"https:\/\/1.bp.blogspot.com\/-ehL5NVuGMvs\/V_wg_0sFCBI\/AAAAAAAALDU\/DTBuI4Wc2g4bqUY3BP8cDJo5D3l-6z_DACLcB\/s1600\/Image14.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/1.bp.blogspot.com\/-ehL5NVuGMvs\/V_wg_0sFCBI\/AAAAAAAALDU\/DTBuI4Wc2g4bqUY3BP8cDJo5D3l-6z_DACLcB\/s640\/Image14.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nThese two adjustments alone can produce a nice variety of plausible island shapes, but you can imagine other variants.\u00a0 Adding octaves of noise can create some interesting shapes: <br \/><div class=\"separator\">\n<a href=\"https:\/\/1.bp.blogspot.com\/-8AMcy2GjQcY\/V_wkVURHL5I\/AAAAAAAALDg\/WMpa5b2zTCgQyCdrmS_Oka3qhMmi3TZwQCLcB\/s1600\/Image16.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/1.bp.blogspot.com\/-8AMcy2GjQcY\/V_wkVURHL5I\/AAAAAAAALDg\/WMpa5b2zTCgQyCdrmS_Oka3qhMmi3TZwQCLcB\/s640\/Image16.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nOr you could try adding more noise in one direction than the other:<br \/><div class=\"separator\">\n<a href=\"https:\/\/3.bp.blogspot.com\/-BkQoDLDZJsE\/V_wkVRX8nrI\/AAAAAAAALDk\/H8BKYCjrJxA7B60-00kc1UAhUCcvWoAvgCEw\/s1600\/Image15.png\"><img border=\"0\" height=\"640\" src=\"https:\/\/3.bp.blogspot.com\/-BkQoDLDZJsE\/V_wkVRX8nrI\/AAAAAAAALDk\/H8BKYCjrJxA7B60-00kc1UAhUCcvWoAvgCEw\/s640\/Image15.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\nIn these examples so far, the land I've been perturbing has been a flat circle.\u00a0 More complexity and interest is added when the land being perturbed is more interesting -- when it has been filled in with noise, mountains, etc.<p><\/p><div>\n<\/div><div class=\"post-body entry-content\" id=\"post-body-5205705313157228012\">\nI'm not going to write a detailed explanation of noise and its use in procedural generation.\u00a0 There's <a href=\"http:\/\/catlikecoding.com\/unity\/tutorials\/simplex-noise\/\">plenty<\/a> of <a href=\"http:\/\/asserttrue.blogspot.com\/2011\/12\/perlin-noise-in-javascript_31.html\">material<\/a> on <a href=\"https:\/\/flafla2.github.io\/2014\/08\/09\/perlinnoise.html\">that topic<\/a> for the interested.\u00a0 I'll summarize by saying that for procedural generation, truly random numbers don't always work well.\u00a0 Often it works better to use a random number generator that varies smoothly according to some input parameters, aka \"noise\".\u00a0 This idea was famously popularized in 1985 by Ken Perlin when he invented \"Perlin noise\".\u00a0 (He won an Academy Award for this work, so don't tell me that computer science isn't glamorous.)\u00a0 Noise is often illustrated with this sort of \"cloud\" picture:<p><\/p><div class=\"separator\">\n<a href=\"https:\/\/4.bp.blogspot.com\/-aVC4csLlxr8\/V_bufTYX8YI\/AAAAAAAALAY\/DHOIK3j1yDkJ1m2OtEd-X5QLCvT2rXWdgCLcB\/s1600\/perlin-noise.png\"><img border=\"0\" src=\"https:\/\/4.bp.blogspot.com\/-aVC4csLlxr8\/V_bufTYX8YI\/AAAAAAAALAY\/DHOIK3j1yDkJ1m2OtEd-X5QLCvT2rXWdgCLcB\/s1600\/perlin-noise.png\" alt=\"image\" \/><\/a><\/div>\nHere the dimensions (x, y) are the input parameters, and the noise output is a number from 0 to 1 shown as a shade from black to white.\u00a0 You can see that the noise (shade) varies smoothly in interesting ways as you move across the photo.<p>What I will talk about here are a couple of pitfalls, issues and technical details about noise generation (specifically Perlin and Simplex noise) that I have not seen much mention of elsewhere.<\/p><p>The first pitfall is that some implementations of Perlin noise or Simplex noise on the Web are not correct.\u00a0 For example, the first result in a Google search for \"perlin noise javascript\" is <a href=\"http:\/\/asserttrue.blogspot.com\/2011\/12\/perlin-noise-in-javascript_31.html\">this page<\/a>.\u00a0 If you read through the code on that page you'll notice a curious thing:\u00a0 there are no initialization parameters and no use of random numbers.\u00a0 If you use that code, you'll generate exactly the same Perlin noise every time you use it!\u00a0 That's not exactly an error, but it's probably not how you expect it to work.<\/p><p>The <a href=\"https:\/\/gist.github.com\/banksean\/304522\">second result<\/a> in the Google search has a more serious problem.\u00a0 Take a look at the first twenty lines of the code and see if you see spot the problem:<\/p><p><\/p><blockquote class=\"tr_bq\">\n<pre>\/\/ Ported from Stefan Gustavson's java implementation\n\/\/ http:\/\/staffwww.itn.liu.se\/~stegu\/simplexnoise\/simplexnoise.pdf\n\/\/ Read Stefan's excellent paper for details on how this code works.\n\/\/\n\/\/ Sean McCullough banksean@gmail.com\n\n\/**\n * You can pass in a random number generator object if you like.\n * It is assumed to have a random() method.\n *\/\nvar ClassicalNoise = function(r) { \/\/ Classic Perlin noise in 3D, for comparison \n   if (r == undefined) r = Math;\n   this.grad3 = [[1,1,0],[-1,1,0],[1,-1,0],[-1,-1,0], \n                                 [1,0,1],[-1,0,1],[1,0,-1],[-1,0,-1], \n                                 [0,1,1],[0,-1,1],[0,1,-1],[0,-1,-1]]; \n   this.p = [];\n   for (var i=0; i&lt;256; i++) {\n   this.p[i] = Math.floor(r.random()*256);\n   }\n   \/\/ To remove the need for index wrapping, double the permutation table length \n   this.perm = []; \n   for(var i=0; i&lt;512; i++) {\n  this.perm[i]=this.p[i &amp; 255];\n  }\n};<\/pre>\n<\/blockquote>\nThe good news is that this code is using a random number generator to initialize the noise generator, so you won't get the same results every time you run it.\u00a0 The bad news is that this code is filling a permutation table with random numbers.\u00a0 This table is supposed to be shuffled numbers from 0 to 255.\u00a0 Instead it's some random selection of numbers from 0 to 255.\u00a0 Most of the time that will work pretty well.\u00a0 Just hope you aren't the unlucky user who ends up with the table of 255 ones! <p>Assuming you find a correct implementation, there are also issues with the return value of Perlin and Simplex noise functions.\u00a0 To start with, some implementations -- like the first one above -- return values from 0 to 1, while others -- like the second one above -- return values from -1 to 1.\u00a0 So you need to be check that before using.\u00a0 But a more serious issue with the return value is that it doesn't <i>really <\/i>range from -1 to 1.\u00a0 Ken Perlin stated that his original implementation returned values from -1 to 1, but it actually returned values (for the two-dimensional case) from about -.70 to +.70.\u00a0 This was (mostly) fixed in the design of Simplex noise, and some implementations of Perlin noise correct the range as well.\u00a0 But if you're not certain, you should check the range of your noise function.<\/p><p>For advanced students, a further complication with return values arises when we consider using <a href=\"https:\/\/eev.ee\/blog\/2016\/05\/29\/perlin-noise\/#octaves\">octaves<\/a>.\u00a0 The basic idea of octaves is to layer noise at different frequencies and different scales.\u00a0 This can create a more natural result, and is how noise is most often used in procedural generation.\u00a0 However, you need to be aware that this changes the range of the noise output.\u00a0 The random noise tends to cancel out, so the range of the output narrows.\u00a0 For example, if we have a noise source that has a range of -1 to 1 and we combine 6 layers with a persistence of 0.5, the resulting noise has a range of about -0.82 to 0.82.<\/p><p>(Even more advanced students will realize that I actually lied in that previous paragraph.\u00a0 I'll get back to that in a moment.)<\/p><p>Unfortunately, range isn't the only issue with the return values of the noise functions.\u00a0 It turns out that the output of these noise functions is not uniformly distributed.\u00a0 If you make a histogram of the output of a Perlin noise function, it looks something like this (copied from <a href=\"https:\/\/eev.ee\/blog\/2016\/05\/29\/perlin-noise\/#some-properties\">here<\/a>):<\/p><p><\/p><div class=\"separator\">\n<a href=\"https:\/\/3.bp.blogspot.com\/-lfp7Mu0YFuE\/V_b5b_uaBZI\/AAAAAAAALAo\/BhWfnxAKf00yM-stlcnE7KZMoSSz1oSFwCLcB\/s1600\/perlin-histogram.png\"><img border=\"0\" height=\"171\" src=\"https:\/\/3.bp.blogspot.com\/-lfp7Mu0YFuE\/V_b5b_uaBZI\/AAAAAAAALAo\/BhWfnxAKf00yM-stlcnE7KZMoSSz1oSFwCLcB\/s640\/perlin-histogram.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nThis shows that numbers near zero are much more likely than numbers near the edges of the range.\u00a0 If you use a noise function to (say) create mountains, and your rule is that you have a mountain wherever the noise is greater than 0.5, you may be surprised to see that your mountains cover only about 5% of the map instead of the expected 25%!<p>(Now that we've seen the distribution of noise, let me return now to my earlier lie.\u00a0 It's not exactly true that when you layer noise in octaves the output range is reduced.\u00a0 Summing noise functions makes the output distribution taller and steeper.\u00a0 On a practical level, this means that the range of the output numbers will narrow.\u00a0 But it's still possible -- just very unlikely -- to get larger numbers.\u00a0 Layering 6 octaves with a persistence of 0.5, the expected range is about -0.82 to 0.82, but theoretically you could get a number from approximately -2 to 2.\u00a0 If a number outside of the expected range will break your code, you should check for that and clip anything outside that range.)<\/p><p>I've recommended testing your noise function to understand it's range, so let's talk now about how to do that.\u00a0 Here's some Javascript code to generate a hundred million values of a one-dimensional noise function and check the output range by finding the highest and lowest returned values.\u00a0 Can you see the problem with this code?<\/p><p>\u00a0\u00a0\u00a0 var pn = new Noise();<br \/>\u00a0\u00a0\u00a0 var minVal = 9999;<br \/>\u00a0\u00a0\u00a0 var maxVal = -9999;<br \/>\u00a0\u00a0\u00a0 for(var i=0; i &lt; 100000000;i++) {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var val = pn.noise(i);<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 minVal = Math.min(minVal, val);<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 maxVal = Math.max(maxVal, val);<br \/>\u00a0\u00a0\u00a0 };<br \/>\u00a0\u00a0\u00a0 console.log('Noise range was '+minVal+' to '+maxVal);<\/p><p>Don't feel bad if you don't.\u00a0 There's no real error in this code -- the problem is in the noise generation code.\u00a0 That starts off like this:<\/p><p>\u00a0\u00a0\u00a0 Noise.prototype.noise = function(x) {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var bx0 = Math.floor(x) % 255;<br \/>\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 [...]<\/p><p>As you can see, the whole number portion of the input \"x\" is taken modulo 255.\u00a0 As a result, this noise function has a periodicity of 256 -- it starts repeating after 256 units in the \"x\" direction. \u00a0 We think we're testing 100 million values, but we're actually only testing 256!\u00a0 This periodicity is related to the 256 value permutation table we saw earlier.\u00a0 In theory, a noise function can use a longer permutation table for a longer periodicity, but in practice you'll rarely see that implemented.<\/p><p>The consequence of this periodicity is that you need to scale your input parameters to the range [0, 255].\u00a0 If you want to use noise to generate a landscape that is 10,000 units on a side, you need to divide your x and y coordinates by 40 before feeding them into the noise function.\u00a0 (There's no periodicity with the fractional portion of the input parameters, so you can simply divide by the appropriate number.)<\/p><p>A less well-understood consequence concerns the octave function.\u00a0 The typical implementation of octaves multiplies the input parameters by 2 every octave, so the noise being layered starts to repeat after the 8th octave (2^8 == 256).\u00a0 In practice this is inconsequential because that many octaves of noise are rarely used.\u00a0 But if for some reason you really need that many octaves, consider extending your noise generator to a greater periodicity.<\/p><p>In a future post I'll talk about ridged noise and at some point I'll get around to cleaning up the Javascript implementation I'm using and post it somewhere.\n<\/p><div>\n<\/div><div class=\"post-body entry-content\" id=\"post-body-5106231915928272086\">\nIn the late summer of 2016, Martin O'Leary made a posting on his web blog of a <a href=\"http:\/\/mewo2.com\/notes\/terrain\/\">fantasy map generator<\/a> written in Javascript.<p><\/p><table cellpadding=\"0\" class=\"tr-caption-container\" style=\"margin: auto; border-spacing: 0px;\"><tbody><tr><td><a href=\"http:\/\/mewo2.com\/deserts\/map.png\"><img height=\"640\" src=\"https:\/\/mewo2.com\/deserts\/map.png\" width=\"640\" alt=\"image\" \/><\/a><\/td><\/tr><tr><td class=\"tr-caption\">An example map from Martin O'Leary's fantasy map generator<\/td><\/tr><\/tbody><\/table><br \/>\nI was quite intrigued by Martin's map generator.\u00a0 (So was a lot of the Internet.\u00a0 And even <a href=\"http:\/\/news.nationalgeographic.com\/2016\/08\/bot-fantasy-map-generator\/\">National Geographic<\/a>.)\u00a0\u00a0 But unlike most folks, I accepted Martin's invitation to grab the code and start playing with it.<p><\/p><table cellpadding=\"0\" class=\"tr-caption-container\" style=\"margin: auto; border-spacing: 0px;\"><tbody><tr><td><a href=\"https:\/\/4.bp.blogspot.com\/-LKG2FIMCiqs\/V_bYUpzFNdI\/AAAAAAAALAI\/sO76Hd7dHzE_3pF2sqhOkKlF9prgf4dLgCLcB\/s1600\/Image1.png\"><img height=\"640\" src=\"https:\/\/4.bp.blogspot.com\/-LKG2FIMCiqs\/V_bYUpzFNdI\/AAAAAAAALAI\/sO76Hd7dHzE_3pF2sqhOkKlF9prgf4dLgCLcB\/s640\/Image1.png\" width=\"622\" alt=\"image\" \/><\/a><\/td><\/tr><tr><td class=\"tr-caption\">An example map from my modified map generator<\/td><\/tr><\/tbody><\/table><br \/>\nFor the past month or so, I've continued to modify and extend Martin's map generator.\u00a0 His maps reminded me of something you might see inside the front cover of an old fantasy paperback, so I made them look more like an old book page.\u00a0 I added color, and different ways to display the maps.\u00a0<p>Along the way I've run into some interesting problems and insights.\u00a0 I've posted some of these to the <a href=\"https:\/\/www.reddit.com\/r\/proceduralgeneration\/\">procedural generation Reddit<\/a>, but I don't want to overstay my welcome there with constant posting, so I decided to start this blog up to capture some of my work in creating this map generator.\u00a0 If that's the sort of thing you find interesting, I welcome your attention!\u00a0 My posting here will no doubt be erratic and meandering, so I encourage you to subscribe through RSS, email or some other method to be notified of new material.\n<\/p><div>\n<\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:53:39+0000","updated_at":"2018-11-07T16:53:39+0000","published_at":"2016-10-16T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":17,"domain_name":"heredragonsabound.blogspot.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":615,"title":"How to Get Startup Ideas: The Meat Grinder Approach - Tyler Tringas","url":"https:\/\/tylertringas.com\/business-ideas-meat-grinder\/","content":"<p>There are a million books and blog posts on how to get startup and business ideas. Many people think the biggest road block to successful entrepreneurial life is having that one great idea. Once you get even a modicum of success, lots of people will start asking you about how you got an idea like that, and for any tips and tricks on how to \u201ccome up\u201d with a similar idea.<\/p>\n<p>The concept is extremely tantalizing and sounds so imminently teachable that it\u2019s a favorite tool of lifestyle business spammers everywhere. 7 Step Guide to Profitable Business Ideas. Join My Webinar on Finding Your Dream Business Plan.<\/p>\n<p>But this is entirely the wrong way to think about it.<\/p>\n<p>First of all you should be coming up with at least five possible business ideas every day. This part should be basically effortless. People trying to sell this part are scamming you.<\/p>\n<p>If you are going to be a successful entrepreneur at all you should innately be looking around you at your life and the lives of others, thinking what are their problems. What are their desires. What do they spend money on. Which of those things are broken or could be done massively better or cheaper or faster. You should be <strong>constantly<\/strong> thinking this way. It should annoy people who spend a lot of time around you.<\/p>\n<p>If you\u2019re not doing that, you\u2019re probably not going to be an entrepreneur\u2026 sorry. It\u2019s okay. There are lots of other great life paths but this one isn\u2019t for you.<\/p>\n<p>There is one common exception to the rule. You might be hung up on one idea, and that stops the process of thinking of new ideas. It\u2019s cool, you just need to build yourself a better meat grinder.<\/p>\n<h3>The secret to coming up with a successful business idea is putting hundreds of ideas through the meat grinder.<\/h3>\n<p>\n<\/p><p>The process goes like this. You see a problem. You quickly sketch the outline of what a business that solves that problem might look like. Then you put that business through a brutal meat grinder of questions. You\u2019re actively trying to poke holes in the idea and stress test it. Beat it up. Be actively trying to discard it and move onto the next one. You need to refine the meat grinder, get really good and fast at obliterating ideas.<\/p>\n<p>And then move on to the next one. If it fails even one part of the meat grinder, ditch it. The worst thing you can do is hold on to that one idea, which is actually fatally flawed, and talk about it with your friends for decades never actually getting to the one that you could have actually built and launched.<\/p>\n<p>A note: the meat grinder is not useful if you are looking for genuine moonshot ideas. If you want to build the next Facebook or Tesla by all means ignore this list and chase your dreams. But just know that unless you are currently well-networked in Silicon Valley, the time of getting VC funding for \u201cjust and idea\u201d is over.<\/p>\n<p>If however you want to build a profitable, successful business that could replace your job, build a lifestyle, help a lot of people and potentially grow into a big sustainable and impactful business, use the meat grinder.<\/p>\n<p>Here are some of the basic questions in my meat grinder for business ideas.<\/p>\n <div class=\"ck_form_container ck_inline\" data-ck-version=\"6\">   <div class=\"ck_form ck_minimal\">   <div class=\"ck_form_fields\">     <h3 class=\"ck_form_title\">Subscribe to future posts<\/h3>     <div class=\"ck_description\">            <\/div>                 <form id=\"ck_subscribe_form\" class=\"ck_subscribe_form\" action=\"https:\/\/forms.convertkit.com\/landing_pages\/2008\/subscribe\" data-remote=\"true\">       <div><input type=\"hidden\" value=\"{&quot;embed_style&quot;:&quot;inline&quot;,&quot;embed_trigger&quot;:&quot;scroll_percentage&quot;,&quot;scroll_percentage&quot;:&quot;75&quot;,&quot;delay_seconds&quot;:&quot;10&quot;,&quot;display_position&quot;:&quot;br&quot;,&quot;display_devices&quot;:&quot;all&quot;,&quot;days_no_show&quot;:&quot;3&quot;,&quot;converted_behavior&quot;:&quot;hide&quot;,&quot;form_style&quot;:&quot;minimal&quot;}\" id=\"ck_form_options\" \/><input type=\"hidden\" name=\"id\" value=\"2008\" id=\"landing_page_id\" \/><input type=\"hidden\" name=\"ck_form_recaptcha\" value=\"\" id=\"ck_form_recaptcha\" \/><div class=\"ck_errorArea\">                <\/div>       <div class=\"ck_control_group ck_email_field_group\">         <label class=\"ck_label\" for=\"ck_emailField\">Email...<\/label>         <input type=\"email\" name=\"email\" class=\"ck_email_address\" id=\"ck_emailField\" placeholder=\"Email...\" required=\"required\" \/><\/div>       <div class=\"ck_control_group ck_captcha2_h_field_group ck-captcha2-h\">         <input type=\"text\" name=\"captcha2_h\" class=\"ck-captcha2-h\" id=\"ck_captcha2_h\" placeholder=\"We use this field to detect spam bots. If you fill this in, you will be marked as a spammer.\" \/><\/div>         <button class=\"subscribe_button ck_subscribe_button btn fields\" id=\"ck_subscribe_button\">         Subscribe       <\/button>                Unsubscribe at any time.           <a class=\"ck_powered_by\" href=\"https:\/\/convertkit.com\/?utm_source=dynamic&amp;utm_medium=referral&amp;utm_campaign=poweredby&amp;utm_content=form\">Powered by ConvertKit<\/a>            <\/div><\/form>   <\/div>  <\/div>  <\/div>     \n<h2>Can I make this?<\/h2>\n<p>Like literally can <strong>I<\/strong> make this. Me, just me. Or failing that, me and a friend with some skills that I know I can conscript for a weekend with guilt and promises of free beer. If the idea is a great one but requires building 3D-printed prototypes. And you don\u2019t have the first clue how to do that. And you don\u2019t know anyone who does. Move on.<\/p>\n<p>If you start getting dozens of ideas that require building 3D-printed prototypes. Then do a deep dive and learn how. Or go to 3D printing Meetups about it and meet someone who knows how and wants to help. Then start over on those ideas.<\/p>\n<p>This is one reason why I\u2019m a huge fan of building a first version software-ish business without writing any code. Everybody has one great idea for an app but because most of them can\u2019t code, they get stuck waiting around for a software developer to fall into their lap. If you can figure out a way to build a crappy version of your \u201capp\u201d without actually coding, you can get past this first hurdle. If it involves facial image recognition algorithms and you can\u2019t even write HTML\u2026 Next!<\/p>\n<h2>Are people currently spending money on it?<\/h2>\n<p>If people are not currently spending money on the problem that your product solves. Ditch it. Otherwise it better be just unbelievably better than what they currently use. Companies currently using spreadsheets or free things? You better be <strong>at minimum 10x<\/strong> better to convince them to start spending money on a totally new thing they already solve for free. If you\u2019re selling to businesses of more than one person, don\u2019t underestimate the shear enormous friction of someone having to go to the boss and convince them to start spending money on something they currently don\u2019t.<\/p>\n<p>What\u2019s that? Your customers are currently not even trying to solve this problem? It\u2019s a problem they don\u2019t even know they have? OMFG, Next!<\/p>\n<h2>How will I get the first 25 customers? Then the next 250?<\/h2>\n<p>Viral word of mouth or put it up in the Apple App Store are not valid answers here. If I can\u2019t think of an easy way to find and get in front of the first 25 customers, the idea is toast. Good answer: I\u2019m building an app for construction companies so I will go on Yelp, find 50 of them, call them and go to their office to pitch them the business. Bad Answer: I\u2019ll launch my music app for tween girls in the App Store and tell my friends to tell their younger sisters. Then it will go viral!<\/p>\n<p>Then I start to think about the next 250 customers. Is there at least one scalable way to reach large groups of potential customers. Do they hang out in places online or in physical places. If your product is for traveling vacuum salesmen\u2026 ditch it! How are you ever going to find them and literally get your idea in front of their eye balls?<\/p>\n<p>By the way, these numbers are geared towards B2B businesses. If you want to sell direct to consumers, multiply them both by 10x.<\/p>\n<h2>If it works, will it be sustainable?<\/h2>\n<p>This is where I start thinking a bit about the end game. Let\u2019s assume I can build the product and can get the first 25 and next 250 customers. At that point, would I have some kind of sustainable competitive advantage? Would I be profitable? What are the switching costs and\/or can my business lock in customers for the long term?<\/p>\n<p>Too often people have great business ideas that are a small variation on another company\u2019s ideas. After all the sweat of building a product from scratch. And after all the hustle of building a customer base from nothing. You\u2019ve just proven to their much larger competitor that there is a market for this feature. They launch the new feature in a week and squash you. Make sure that you won\u2019t just be doing a larger company\u2019s R&amp;D work for them.<\/p>\n<h2>Am I the person to build this business?<\/h2>\n<p>Sometimes I think of great business ideas that pass all these basic tests, but I realize I\u2019m just not the right person to build it. Maybe I think that I just don\u2019t actually like the day to day process of what that business would entail. If you\u2019re going to build an app for lawyers, make sure you don\u2019t absolutely hate talking to lawyers, because you know you\u2019re going to be doing a lot of that.<\/p>\n<p>For example, I travel constantly. I\u2019m always thinking of new travel gear ideas. Really substantial improvements on the things I carry in my backpack. But actually building an ecommerce business requires doing lots of \u201cfulfillment\u201d yourself in the beginning. That means sitting somewhere with your inventory piled up all around you, packing it up and shipping it out every day. Not exactly compatible with a minimalist traveling lifestyle. Ditched.<\/p>\n<h2>Just the beginning<\/h2>\n<p>This is not even close to comprehensive. Building on and refining your meat grinder is the path to finding a profitable business idea. Make it your mission to <strong>discard<\/strong> five business ideas per day for a week. If something comes out the other side, and you really just can\u2019t get rid of it. Odds are you\u2019ve got something worth your time. You already know that <strong>you<\/strong> can go out and build it, so build a first version and launch it. If it flops, figure out what you can learn from that to improve your meat grinder and get back to grinding.<\/p>\n<h2>Got an idea and want me to put it through the meat grinder?<\/h2>\n<p>Tweet it at me.<\/p>\n<blockquote class=\"twitter-tweet\" data-lang=\"en\">\n<p dir=\"ltr\" lang=\"en\">I have discovered the secret to profitable business ideas and it is a kitchen appliance <a href=\"https:\/\/t.co\/BYpRLG1eLF\">https:\/\/t.co\/BYpRLG1eLF<\/a><\/p>\n<p>\u2014 Tyler Tringas (@tylertringas) <a href=\"https:\/\/twitter.com\/tylertringas\/status\/723854809338617856\">April 23, 2016<\/a><\/p><\/blockquote>\n<p>\n <\/p><div class=\"ck_form_container ck_inline\" data-ck-version=\"6\">   <div class=\"ck_form ck_minimal\">   <div class=\"ck_form_fields\">     <h3 class=\"ck_form_title\">Subscribe to future posts<\/h3>     <div class=\"ck_description\">            <\/div>                 <form class=\"ck_subscribe_form\" action=\"https:\/\/forms.convertkit.com\/landing_pages\/2008\/subscribe\" data-remote=\"true\">       <div><input type=\"hidden\" value=\"{&quot;embed_style&quot;:&quot;inline&quot;,&quot;embed_trigger&quot;:&quot;scroll_percentage&quot;,&quot;scroll_percentage&quot;:&quot;75&quot;,&quot;delay_seconds&quot;:&quot;10&quot;,&quot;display_position&quot;:&quot;br&quot;,&quot;display_devices&quot;:&quot;all&quot;,&quot;days_no_show&quot;:&quot;3&quot;,&quot;converted_behavior&quot;:&quot;hide&quot;,&quot;form_style&quot;:&quot;minimal&quot;}\" \/><input type=\"hidden\" name=\"id\" value=\"2008\" \/><input type=\"hidden\" name=\"ck_form_recaptcha\" value=\"\" \/><div class=\"ck_errorArea\">                <\/div>       <div class=\"ck_control_group ck_email_field_group\">         <label class=\"ck_label\" for=\"ck_emailField\">Email...<\/label>         <input type=\"email\" name=\"email\" class=\"ck_email_address\" placeholder=\"Email...\" required=\"required\" \/><\/div>       <div class=\"ck_control_group ck_captcha2_h_field_group ck-captcha2-h\">         <input type=\"text\" name=\"captcha2_h\" class=\"ck-captcha2-h\" placeholder=\"We use this field to detect spam bots. If you fill this in, you will be marked as a spammer.\" \/><\/div>         <button class=\"subscribe_button ck_subscribe_button btn fields\">         Subscribe       <\/button>                Unsubscribe at any time.           <a class=\"ck_powered_by\" href=\"https:\/\/convertkit.com\/?utm_source=dynamic&amp;utm_medium=referral&amp;utm_campaign=poweredby&amp;utm_content=form\">Powered by ConvertKit<\/a>            <\/div><\/form>   <\/div>  <\/div>  <\/div>","created_at":"2018-11-07T16:53:39+0000","updated_at":"2018-11-07T16:53:39+0000","published_at":"2016-04-23T12:23:40+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"tylertringas.com","preview_picture":"https:\/\/tylertringas.com\/wp-content\/uploads\/2016\/04\/pablo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":612,"title":"My Transformation: How I lost 66 pounds and gained a 6 pack in 8 months.","url":"http:\/\/debarghyadas.com\/writes\/transformation\/","content":"<p>On June 2016, I was 23 years old, 5' 9\" and weighed 95kg. I knew I'd let go of myself.  I was a year and a half out of college, and I had no idea how to adult. Putting my professional life first was the only way I knew how to function, but when the fixed duration problem sets of college times became an infinite stream of work, I kept pushing myself further and further. The longer I worked, the more I needed to relax in my downtime. As someone who's always found comfort in friends and food, it was no surprise that it had come to a point where midnight department store runs to pick up chocolates and junk food had become a daily fixture. As the weeks ticked on, so did the pounds. When the snow started melting, the thick peacoats of a New York City winter couldn't disguise my ignorance any more. I had to do something about it.<\/p><p>In the next 8 months, I lost 30kg (66lbs) and became the fittest I've ever been. I measured in at 8% body fat and had a 6 pack for the first time in my life. Using data and YouTube videos to guide me, I changed the way I ate and started training while trying to balance a fair share of fun and a full-time job.<\/p>\n<div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/beforeafter.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    Originally posted on Reddit's <a href=\"https:\/\/www.reddit.com\/r\/progresspics\/comments\/5h9uop\/m2459_209lbs_152lbs_57lbs_6_months_first_time_in\/\">\/r\/progresspics<\/a> where it received over 360 upvotes. <br \/>All body fat measurements are taken on a bioelectrical impedance scale and are likely inaccurate.\n  <\/div>\n<\/div>\n<div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/finalshirtless.jpg\" alt=\"image\" \/><img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/mar_2017.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    <b>Left:<\/b> In my final form. Picture credit: <a href=\"https:\/\/www.instagram.com\/nyphotony\/\">Richard Storm<\/a>. <br \/><b>Right:<\/b> At maintenance calories for about a month. A newly acquired <a href=\"https:\/\/en.wikipedia.org\/wiki\/Apollo%27s_belt\">Adonis belt<\/a>.<br \/>All body fat measurements are taken on a bioelectrical impedance scale and are likely inaccurate.\n  <\/div>\n<\/div>\n<div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/facegains.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    Face differences at various times between early 2016 and late 2016.\n  <\/div>\n<\/div>\n\n<p>In the beginning, my goal was to religiously do a 12 week training program. I had a fixed end date in mind, and I'd hit the gym and eat correctly until that date. By the time I'd reached that date, I began enjoying the change, and I wanted to see how far I could go. It took about 16 more weeks to get to a six pack, and it's been 8 months since I started as of the time of writing this. There were five big changes to be made:<\/p>\n<ul><li><strong>Data and Measurement<\/strong>: Track progress to keep motivated and to be able to adjust and plan better.<\/li>\n<li><strong>Nutrition<\/strong>: I used an intermittent fasting and ketogenic diet approach, and several supplements.<\/li>\n<li><strong>Training<\/strong>: A thought out 4-days-a-week training plan.<\/li>\n<li><strong>Grooming<\/strong>: Better, well-fitting clothes, lenses over glasses, and a clean shave.<\/li>\n<li><strong>Motivation<\/strong>: Learn and stay motivated from people who've done this before. Fitness Youtubers really helped.<\/li>\n<\/ul><div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/four.jpg\" alt=\"image\" \/><img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/one.jpg\" alt=\"image\" \/><img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/three.jpg\" alt=\"image\" \/><img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/two.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    More after pictures. Also viewable <a href=\"http:\/\/imgur.com\/a\/e1fMn\">here<\/a>.\n  <\/div>\n<\/div>\n<p>Looking at the mirror every day trying to spot a visible change is not a very scientific way to go about measuring change. I realized a few weeks in that I needed to approach this problem more methodically. One of my Computer Science professors in college once said something that really stuck with me.<\/p>\n<blockquote>\n<p>How can you improve if you cannot measure?<\/p>\n<\/blockquote>\n<p><a href=\"http:\/\/www.johnstonefitness.com\/\">JohnStoneFitness<\/a>, one of the people I'd started following online for fitness advice and motivation, preaches data-driven training. He had religiously tracked his food intake, calories, macronutrient breakdown, weight, body fat, and front and side profiles for almost 6 years in his <a href=\"http:\/\/www.johnstonefitness.com\/my-transformation\/food-logs\/\">food logs<\/a>. Every few weeks, he'd take body measurements too. While his methods may have been extreme, I found tracking your progress broadly (beyond just weight) and often to be extremely helpful, if not necessary, in gauging your progress and estimating how long it'll take you to achieve your goal. Seeing the scale tick down even 0.1kg every day can be immensely motivating. Here's my transformation by the numbers:<\/p>\n<table><thead><tr><th>Before<\/th>\n<th>After<\/th>\n<th>Difference<\/th>\n<\/tr><\/thead><tbody><tr><td><strong>Vitals<\/strong><\/td>\n<\/tr><tr><td>Time<\/td>\n<td>June 1, 2016<\/td>\n<td>Jan 28, 2017<\/td>\n<td>+241days<\/td>\n<\/tr><tr><td>Height<\/td>\n<td>5' 9\"<\/td>\n<td>5' 9\"<\/td>\n<td>-<\/td>\n<\/tr><tr><td>Weight<\/td>\n<td>95kg (209lbs)<\/td>\n<td>64.9kg (143lbs)<\/td>\n<td>-30.1kg (66lbs)<\/td>\n<\/tr><tr><td>Body Fat Percentage<\/td>\n<td>25%<\/td>\n<td>8.1%<\/td>\n<td>-16.9%<\/td>\n<\/tr><tr><td>Body Mass Index (BMI)<\/td>\n<td>31.3<\/td>\n<td>21.4<\/td>\n<td>-9.9<\/td>\n<\/tr><tr><td><strong>Measurements<\/strong><\/td>\n<\/tr><tr><td>Waist<\/td>\n<td>37.5\"<\/td>\n<td>27\"<\/td>\n<td>-10.5\"<\/td>\n<\/tr><tr><td>Hip<\/td>\n<td>42\"<\/td>\n<td>34.5\"<\/td>\n<td>-7.5\"<\/td>\n<\/tr><tr><td>Chest<\/td>\n<td>42\"<\/td>\n<td>37\"<\/td>\n<td>-5\"<\/td>\n<\/tr><tr><td>Bicep<\/td>\n<td>15.5\"<\/td>\n<td>13.5\"<\/td>\n<td>-2\"<\/td>\n<\/tr><tr><td>Thigh<\/td>\n<td>27\"<\/td>\n<td>22\"<\/td>\n<td>-5\"<\/td>\n<\/tr><tr><td>Calves<\/td>\n<td>14.5\"<\/td>\n<td>13.5\"<\/td>\n<td>-1\"<\/td>\n<\/tr><tr><td><strong>Primary Lifts<\/strong><\/td>\n<\/tr><tr><td>Bench Press Max<\/td>\n<td>~135lbs<\/td>\n<td>202lbs<\/td>\n<td>+67lbs<\/td>\n<\/tr><tr><td>Deadlift Max<\/td>\n<td>N\/A<\/td>\n<td>260lbs<\/td>\n<td>-<\/td>\n<\/tr><tr><td>Squat<\/td>\n<td>~95lbs<\/td>\n<td>190lbs<\/td>\n<td>+95lbs<\/td>\n<\/tr><tr><td><strong>Cardio<\/strong><\/td>\n<\/tr><tr><td>Running (5k)<\/td>\n<td>~35:00+<\/td>\n<td>~22:10<\/td>\n<td>-12:50<\/td>\n<\/tr><tr><td>Running (10k)<\/td>\n<td>N\/A<\/td>\n<td>46:40<\/td>\n<td>-<\/td>\n<\/tr><\/tbody><\/table><div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/progresschart.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    Weight progress from July 19th 2016 to Feb 1, 2017. I use a 10-day moving average, and I interpolate the weights on days I was on extended vacation. There were several weight loss plateaus, but the average rate was between 0.1 and 0.15 kg a day, or between 1.5 to 2.3lbs a week.\n  <\/div>\n<\/div>\n<p>In order to record and track your progress, I used, and would recommend:<\/p>\n<ul><li><a href=\"https:\/\/www.amazon.com\/AccuFitness-MT05-MyoTape-Body-Measure\/dp\/B000G7YW7Y\/ref=sr_1_4_a_it?ie=UTF8&amp;qid=1487755922&amp;sr=8-4&amp;keywords=body+tape+measure\">Body Tape Measure<\/a>: For taking measurements in your key body areas.<\/li>\n<li><a href=\"https:\/\/www.amazon.com\/Balance-Accuracy-Digital-Accurate-Measurements\/dp\/B019928F74\/ref=sr_1_3_a_it?ie=UTF8&amp;qid=1487755957&amp;sr=8-3&amp;keywords=body+fat+weighing+scale\">Body Fat Weighing Scale<\/a>: These scales estimate your body fat with bioelectrical impedance, by measuring how small amounts of electricity flows through your body. They are not accurate in measuring your absolute fat, but can track relative body fat change well. A typical one that costs $40 measures primarily in the legs, whereas the <a href=\"https:\/\/www.amazon.com\/HBF-514C-Full-Body-Sensor-Composition-Monitor\/dp\/B00DL5WTZK\/ref=sr_1_2_a_it?ie=UTF8&amp;qid=1487756114&amp;sr=8-2&amp;keywords=fat+scale+hand\">more accurate one that measures full body<\/a> will put you back $90.<\/li>\n<li><a href=\"http:\/\/www.exrx.net\/Calculators\/OneRepMax.html\">One Rep Max Calculator<\/a>: If you can lift X lbs of weight Y times, how much can you lift 1 time?<\/li>\n<\/ul><p>Tracking metrics daily, besides being motivating, can help you very accurately predict how long you have left until you reach your weight loss goals. Almost universally, the normal rate of weight loss is 0.1 to 0.2kg a day, or between <strong>1.5 - 3lbs a week<\/strong>. Any quicker than this is considered unhealthy weight loss unless you're morbidly obese. Anything lower than 1lb should be a cause for concern, and lead you to tweak your plan. The important caveats are:<\/p>\n<ul><li>You almost always will lose weight extra quickly in the first week, which is water weight loss, and is not to be counted in any prediction or average calculation.<\/li>\n<li>I've personally noticed daily fluctuations go from +-2lbs. In order to get as consistent a reading as possible, weigh yourself in the morning post-dump.<\/li>\n<\/ul>\n<p>It quickly became apparent to me that losing weight would be 80% about eating correctly. A misconception I'd held is that it involved a lot of cardiovascular activity, like running. This isn't true. Twenty minutes, almost 5kms of running, burns about 200 calories, which is about as much as a <a href=\"https:\/\/www.fatsecret.com\/calories-nutrition\/subway\/chocolate-chip-cookie\">Subway cookie<\/a>. All that effort and sweat from running a 5k is equivalent to a mere cookie. <em>That's<\/em> why nutrition is critical to fat loss.<\/p>\n<h3 id=\"old-nutrition-habits\">Old Nutrition Habits<\/h3>\n<p>I had an Indian upbringing, and relationship between an Indian and their food is divine. Growing up, I was a fat kid. In Bengali culture, the opposite of thin is \"healthy\". Food is a central part of celebration, and being fat is regarded as the norm. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bengali_cuisine#Mish.C5.A3i_.28sweets.29\">Sweets<\/a> make up a critical part of the Bengali diet. In fact, in the city of Kolkata, where I'm from, <a href=\"http:\/\/rchiips.org\/nfhs\/FCTS\/WB\/Kolkata.pdf\">40% of adults are overweight and over 20% of men have high blood sugar<\/a>. I was rarely ever denied food growing up. I had about 3 main meals and 3 snack meals every day. The philosophy with food has always been<\/p>\n<blockquote>\n<p>Eat till you can't eat any more... and then have dessert.<\/p>\n<\/blockquote>\n<h3 id=\"nutrition-theories-and-tactics-101\">Nutrition Theories and Tactics 101<\/h3>\n<p>When I started looking for a better way to eat, I realized most of my preconceptions of what healthy food was were wrong. The <a href=\"http:\/\/blogs-images.forbes.com\/michaelpellmanrowland\/files\/2016\/11\/USDA_Food_Pyramid.gif?\">food pyramid<\/a> we'd learnt growing up was obsolete and replaced by a very different distribution of nutrients, according to this <a href=\"https:\/\/www.hsph.harvard.edu\/nutritionsource\/healthy-eating-plate\/\">Harvard Health study<\/a>.<\/p>\n<p>The nutrition principle most fitness folks follow is <strong>Calories In Calories Out (CICO)<\/strong>. According to this theory, everyone has a <strong>Total Daily Energy Expenditure (TDEE)<\/strong>, which depends on your gender, age, height, weight, muscle mass and physical activity. There are plenty of <a href=\"https:\/\/mytdee.com\">calculators<\/a> online to help you calculate what your TDEE is. Mine was initially ~2400 calories. The theory postulates that if you consume more calories than your TDEE, you gain weight, and if you consume less, you lose weight. One pound of weight is 3500 calories, so if you incur a 500 calorie deficit on your daily TDEE, you'll lose a consistent pound a week. Of course, this has its critiques and may be an oversimplification, but seems to work in practice.<\/p>\n<p>Another popular theory is that of the <strong>Ketogenic diet<\/strong>. What this diet states, contrary to CICO, is that all calories aren't made equal, and some are better for you than others. On a ketogenic diet, you almost completely cut out carbohydrates and sugar (a form of carbs) from your diet, consuming less than 50g of it daily. When carbohydrates are consumed, it breaks into sugar which provides the body with energy. If you don't use this energy, your body secretes insulin to control your sugar, depositing it into fat. The theory claims that if you cut out carbs, your body will break down fat for energy and your insulin will never spike, thereby preventing you from putting on weight. Does this mean you can consume as much of the other macronutrients (proteins and fats) as you want? Probably not. That being said, this diet has also seen its fair share of success.<\/p>\n<p><strong>Intermittent fasting (IF)<\/strong> is not so much of a theory as it is a principle to conveniently eat less without being too hungry. Most fitness professionals have the time and patience to diligently weigh, track and record the details of their daily diet, usually using tools like <a href=\"https:\/\/www.myfitnesspal.com\/\">MyFitnessPal<\/a>. While this works for them, it can be overbearing and time-consuming for a normal person. Often, all the work involved in just tracking calories is not sustainable and people struggle to do it for a sustained period of time. IF is a simple heuristic that lets you control your calories by simply minimizing the time you allow yourself to eat. Typically, in IF, you give yourself an 8 hour eating window, usually from around noon - 8pm. Because we tend to not be too hungry in the morning, and can delay our first meal with some black coffee, following IF is generally not very hard and doesn't feel too restrictive.<\/p>\n<h3 id=\"new-nutrition-habits\">New Nutrition Habits<\/h3>\n<p>Because I had excess body fat, my diet was centered around what bodybuilders called <strong>cutting<\/strong> - when you try to lose weight and maintain muscle mass. This is in contrast to a diet a lanky person would take, called <strong>bulking<\/strong>, when you try to gain weight and muscle mass.\nMy general approach to nutrition involved incorporating ideas from all of these theories and tactics:<\/p>\n<ul><li><strong>CICO<\/strong> I began to incorporate CICO principles a few months into my progress with Keto and IF. Initially, I tried to use MyFitnessPal to get an idea of the calorie intake of my most common meals in a week, and their macronutrient breakdown (fat-carb-protein). I left myself some leeway and aimed to hit 1600-1800 calories a day. After some time, I felt like I could accurately eyeball calories and breakdowns for new meals as well.<\/li>\n<li><strong>Keto<\/strong> I cut out carbohydrates completely, including all rices, breads, noodles, pastas, etc. Sugars, desserts, and all processed foods were a complete no-no. Later on, I re-incorporated healthier <a href=\"http:\/\/www.diabetes.org\/food-and-fitness\/food\/what-can-i-eat\/understanding-carbohydrates\/glycemic-index-and-diabetes.html\">low-glycemic index carbs<\/a> like brown rice, farro and quinoa. My diet consisted primarily of vegetables and meats and healthy fats (nuts, avocado, and the occasional oil, butter).<\/li>\n<li><strong>IF<\/strong> I would start my morning with a black coffee, and eat my first meal at lunch, maybe a snack after, and dinner at 630-7pm usually.<\/li>\n<li><strong>Protein Intake<\/strong> Conventional bodybuilding wisdom dictates that you consume 1g+ of protein per pound of body weight daily, but I usually could not eat more than 150g.<\/li>\n<\/ul><p>Notably, I did not stop drinking. I usually consumed 4-5 drinks a week. I made sure to usually stick to neat hard liquors and wine. When I used a mixer, I'd make sure it was either diet coke or just soda. A drink averages out to about 100 calories, which, to me, was a small price to pay in a week for the lifestyle benefit.<\/p>\n<p>I often have cravings. My best solution for savory cravings were usually some nuts, cheese, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pork_rind\">pork rinds<\/a>. For sugar cravings, the ultimate no-carb solution for me was the <a href=\"http:\/\/www.questnutrition.com\/?gclid=Cj0KEQiA25rFBRC8rfyX1vjeg7YBEiQAFIb3b0qhTXZvcWbJdXEBjDtA8_2io9cjLGAmB6ZlEgeZ-C0aAst08P8HAQ\">Quest Bar<\/a>. I may have eaten over 100 of these during these 8 months and they're delicious.<\/p>\n<h3 id=\"supplements\">Supplements<\/h3>\n<p>Supplements, in retrospect, were by far the most overrated aspect of a transformation. At worst, they are an expensive placebo that keeps you motivated because you spent money on things. At best, they accelerate your progress by maybe 10%. I've used 4 supplements:<\/p>\n<ul><li><strong>Grenade Thermo Detonator<\/strong>: A scary name, but essentially a dosage of caffeine from concentrated green tea extract that's meant to be a fat-burner\/pre-workout pill. It gives you some energy to focus, and suppresses your appetite. It is not very different from simply taking a packet of coffee. I used these for the first 4 months. You can find them <a href=\"https:\/\/www.amazon.com\/Grenade-Detonator-Powerful-Thermogenic-Capsules\/dp\/B0055G4PO6\">here<\/a><\/li>\n<li><strong>Multivitamin<\/strong>: Just to ensure you get your vitamin requirements for the day, but there is <a href=\"http:\/\/www.theatlantic.com\/health\/archive\/2013\/07\/the-vitamin-myth-why-we-think-we-need-supplements\/277947\/?1\">evidence to show<\/a> this might not really be necessary. You can find the ones I used <a href=\"https:\/\/www.amazon.com\/Optimum-Nutrition-Opti-Men-Multivitamin-Supplement\/dp\/B00K2RJAR0\/ref=sr_1_1_s_it?s=hpc&amp;ie=UTF8&amp;qid=1487362220&amp;sr=1-1&amp;keywords=on+vitamins\">here<\/a>.<\/li>\n<li><strong>Fish Oil<\/strong>: I heard from many sources that omega-3s are good for you so I thought I'd get some <a href=\"https:\/\/www.amazon.com\/Optimum-Nutrition-Fish-Oil-Softgels\/dp\/B003RDGNY4\/ref=sr_1_1_s_it?s=hpc&amp;ie=UTF8&amp;qid=1487362270&amp;sr=1-1&amp;keywords=on+fish+oil\">here<\/a>.<\/li>\n<li><strong>Whey Protein<\/strong>: If you don't consume enough protein a day and are on a weight loss plan, you will lose muscle with your weight. 1 - 1.5g\/lb of body weight is the general recommendation. Because it's not always easy to make sure you get that much pure protein, a protein shake is an easy way to quickly gulp down 50g of protein regardless of how hungry\/full you are. I initially started with the most popular <a href=\"https:\/\/www.amazon.com\/Optimum-Nutrition-Standard-Protein-Chocolate\/dp\/B000QSNYGI\/ref=sr_1_1_s_it?s=hpc&amp;ie=UTF8&amp;qid=1487362313&amp;sr=1-1&amp;keywords=on+whey+gold+standard\">brand<\/a> but realized that there were just as good and <a href=\"https:\/\/us.myprotein.com\/sports-nutrition\/impact-whey-protein\/10852500.html\">far cheaper options<\/a> out there (use discount codes for the latter). I did an  un-comprehensive <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1CvBHxfDkwYOXwHFXCFS9RNflsXF-0R1ayY4oZQS0Mc4\/edit#gid=0\">comparison of protein prices<\/a> too.<\/li>\n<\/ul><p>I do not represent any of these brands.<\/p>\n<div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/supplements.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    These were the four supplements I took. From top right clockwise: Grenade, Fish oil, Whey Protein, Multivitamin.\n  <\/div>\n<\/div>\n<p>One thing I learnt was that many people who lose weight gain it back because their <strong>Total Daily Energy Expenditure (TDEE)<\/strong> goes down after they lose weight, so when they resume their original nutrition habits, they put it all on again. One way to prevent this is to increase your TDEE by building muscle while you lose weight. Having more muscle, for both males and females, means your body expends more energy, allowing you to maintain your weight whilst eating much more.<\/p>\n<p>There are an infinite number of variations and ways to train, some negligibly better than others. My philosophy was to choose something I enjoyed and could stick with. This is what I did:<\/p>\n<ul><li>Hit the gym 3-4 days a week for about an hour.<\/li>\n<li>Start with 10 minutes of cardio, with 50 minutes for lifting.<\/li>\n<li>Cycle the following lift routine, commonly called the <strong>bro split<\/strong>: Chest+Triceps, Back+Biceps, Legs, and Shoulders+Abs.<\/li>\n<li><strong>Compound exercises<\/strong> are far harder and far more effective than isolation movements because they hit many body parts simultaneously. Each day has a corresponding compound exercise. Chest: <strong>Bench Press<\/strong>. Back: <strong>Deadlift<\/strong>. Legs: <strong>Squat<\/strong>. Shoulders: <strong>Overhead Press<\/strong>.<\/li>\n<li>For your compound movement, work first on perfecting form. Do 4-5 sets of 5-6 repetitions.<\/li>\n<li>After the compound exercise, hit 4 exercises in the training session for the muscle groups allotted to the day. Do 3 sets of 8-12 repetitions. You can look up an exhaustive list on <a href=\"https:\/\/www.bodybuilding.com\/exercises\/\">bodybuilding.com<\/a>.<\/li>\n<li>Each week, you should be focusing on either improving form of your exercise, or using more weight, or more sets or more repetitions than you were previously. This concept is called <strong>progressive overload<\/strong>.<\/li>\n<\/ul><p>Many more experienced lifters say that their mistake in their early lifting career was <a href=\"http:\/\/fitness.stackexchange.com\/questions\/4585\/why-do-compound-exercises-build-more-mass-than-isolation-exercises\">not doing enough compound exercises<\/a>.<\/p>\n<div class=\"imgcap\">\n  <img src=\"http:\/\/debarghyadas.com\/writes\/assets\/transform\/compounds.jpg\" alt=\"image\" \/><div class=\"thecap\">\n    The four primary compound movements.\n  <\/div>\n<\/div>\n<p>Besides the daily tracking, the source of motivation that really kept me going were fitness YouTubers. Here are my favorite fitness channels:<\/p>\n<ul><li><a href=\"https:\/\/www.youtube.com\/channel\/UCe0TLA0EsQbE-MjuHXevj2A\">AthleanX<\/a>: One of the most popular fitness channels on YouTube that keeps producing great content.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCPxMZIFE856tbTfdkdjzTSQ\">BeerBiceps<\/a>: The best Indian-specific fitness channel.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCRFUw9QQy7ypUkIe-HOVdZw\">StudentAesthetics<\/a>: Dutch guy. Focuses less on the science and advice, and more on balancing lifestyle with fitness. Amazing production quality.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCAcQW_4qZ12wNt3T0M9b-Vw\">VitruvianPhysique<\/a>: Canadian guy living in Australia. Very informative, scientific and funny.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCU7J3bxynfGRz--EFZ9-Dvw\">RobLipsett<\/a>: Irish guy. Just a generally chill lad.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCaHx0T1LWrVKWF1XfWWLSuw\">OmarIsuf<\/a>: A popular, old channel. Collaborates often, and focuses on powerlifting as well as bodybuilding.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCduKuJToxWPizJ7I2E6n1kA\">BroScienceLife<\/a>: The most popular satire fitness channel.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UCKf0UqBiCQI4Ol0To9V0pKQ\">BuffDudes<\/a>: Twins.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UC68TLK0mAEzUyHx5x5k-S1Q\">JeffNippard<\/a>: Very scientific.<\/li>\n<li><a href=\"https:\/\/www.youtube.com\/channel\/UC260Bk8ms5CvOdh-zZkmrbw\">TerronBeckham<\/a>: NFL player Odell Beckham's cousin. Extremely strong.<\/li>\n<\/ul><p>Many of these channels are mid-way between small channels and super-popular channels. I've found the 25-150k subscriber range to be ideal because the Youtubers have proven themselves but are still eager to get better.\nAll of these channels feature people who I believe to be <em>completely natural<\/em>, and do not take anabolic steroids. There are plenty of channels featuring people that are on steroids, but they sets unrealistic expectations and give incorrect guidance.\nAlthough some of these channels are sponsored, I believe that their advice is authentic and not sponsor-influenced. They do not try to sell you stuff.<\/p>\n<p>Other sources of inspiration include:<\/p>\n<ul><li><a href=\"https:\/\/www.reddit.com\/r\/progresspics\">r\/progresspics<\/a>: Transformations of real people.<\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/Fitness\">r\/fitness<\/a>: Great advice from real people not trying to sell you things.<\/li>\n<li><a href=\"http:\/\/www.johnstonefitness.com\/my-transformation\/food-logs\/\">JohnStoneFitness<\/a>: Very rich source of data.<\/li>\n<\/ul>\n<p>In the beginning, everything about this personal goal seemed difficult. The more I looked online, the harder and more confusing it seemed. The fitness industry makes a living on convincing you that it is hard. There may be a 1000 diets to choose from and a 1000 micro-optimizations you could make in the gym, but in the end it all boils down to science and consistency. If you know the scientific truths, and can consistently maintain better nutritional and training habits, you're guaranteed to succeed.<\/p>\n<p>By no means do I embrace fitness as the best lifestyle to adopt. I would never consciously judge other people by their appearance. However, I've personally noticed a significant effect on how people interact with me and a huge confidence boost too. Please feel free to reach out to me with any questions I might not have answered in this post.<\/p>\n<ul><li><a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1zq5EnUBuZHHb7btgL_9YWxmCYqy2JYEQe7xo_5sT1iU\/edit?usp=sharing\">My Google Doc Weight Loss and Measurement Log<\/a><\/li>\n<li><a href=\"https:\/\/tdeecalculator.net\/\">TDEE Calculator<\/a>: How many calories can you eat daily and maintain weight?<\/li>\n<li><a href=\"http:\/\/www.strengthstandards.co\/#\/\">Strength Standards<\/a>: How do your lifts compare to others with your weight and gender?<\/li>\n<li><a href=\"http:\/\/www.heartbreakhill.org\/age_graded.htm\">Running Standards<\/a>: How do your running times compare with others of your height and gender?<\/li>\n<li><a href=\"http:\/\/www.myfitnesspal.com\/food\/calorie-chart-nutrition-facts\">MyFitnessPal<\/a> and <a href=\"https:\/\/www.google.com\/search?q=apple+calories&amp;oq=apple+calories&amp;aqs=chrome.0.0l6.2630j0j4&amp;sourceid=chrome&amp;ie=UTF-8\">Google for calories<\/a>: Instant calorie facts and nutrition data<\/li>\n<li><a href=\"http:\/\/www.exrx.net\/Calculators\/OneRepMax.html\">One Rep Max Calculator<\/a>: If you can lift X lbs of weight Y times, how much can you lift 1 time?<\/li>\n<\/ul>\n  I love hearing feedback! If you didn't like something, let me know in the comments and feel free reach out to me. If you did, you can <a target=\"_blank\" href=\"https:\/\/twitter.com\/intent\/tweet?url=http:\/\/debarghyadas.com\/writes\/transformation\/&amp;text=My Transformation: How I lost 66 pounds and gained a 6 pack in 8 months.&amp;via=debarghya_das\">share it with your followers in one click<\/a> or <a href=\"https:\/\/twitter.com\/debarghya_das\">follow me on Twitter<\/a>!","created_at":"2018-11-07T16:53:38+0000","updated_at":"2018-11-07T16:53:38+0000","annotations":[],"mimetype":"text\/html","reading_time":16,"domain_name":"debarghyadas.com","preview_picture":"http:\/\/debarghyadas.com\/writes\/assets\/transform\/finalshirtless.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":613,"title":"Not everything is an expression","url":"http:\/\/www.rntz.net\/post\/2016-06-06-not-everything-is-an-expression.html","content":"<header><p>\"<em>Everything is an expression<\/em>\" is a slogan of several programming languages,\nparticularly those descended from the Lisp family. It is meant to contrast with\nlanguages that distinguish <em>statements<\/em> from <em>expressions<\/em>, such as Python,\nJavascript, and C.<\/p><p>Favoring expressions over statements can simplify code. Instead of the bulky:<\/p><pre>if increment:\n    do_something_with(x + 1)\nelse:\n    do_something_with(x)\n<\/pre><p>you can write:<\/p><pre>do_something_with(x + if increment then 1 else 0)\n<\/pre><p>However, this is not an essay about the benefits of replacing statements with\nexpressions.<\/p><p>Rather, I claim that <strong>not<\/strong> everything is an expression! The idea that there\nare only expressions ignores the importance of other <em>syntax classes<\/em>. I believe\nLisp macros are underpowered because of this ignorance.<\/p><h2>1 What are syntax classes?<\/h2><p>I will explain what I mean by <em>syntax class<\/em> in two ways, first with an\nabstract definition, then with an example. Skip the first if it's confusing;\nskip the second if you don't need it.<\/p><h3>Syntax classes, abstractly<\/h3><p>By <em>syntax class<\/em> I mean <strong>a distinct type in our language's\nabstract syntax tree<\/strong>, or <strong>a semantically interesting nonterminal symbol in\nour language's grammar<\/strong>. In most grammars, many nonterminals are <em>semantically\nboring<\/em>, and exist to specify banal syntactic details \u2014 operator precedence,\nwhere commas go in a list, the format of numeric literals. But some nonterminals\nrepresent things like <em>expressions<\/em>, <em>statements<\/em>, <em>declarations<\/em>, <em>modules<\/em>,\n<em>patterns<\/em> or <em>variables<\/em>; the \"top-level\" categories in a broad overview of the\nlanguage's syntax. These are syntax classes.<sup class=\"fnlink\"><a href=\"#fn-1\">1<\/a><\/sup><\/p><h3>Syntax classes, by example<\/h3><pre>def twice_last(seq):\n    last, *rest = reversed(seq)\n    return last * 2<\/pre><p>This (rather unidiomatic) Python code is colored by syntax class: expressions in red, statements in\nblue, declarations in purple, and patterns in green.<\/p><p>Different syntax classes represent distinct concepts:<\/p><ul><li><p>Expressions are <em>evaluated<\/em> to compute their\nvalue. Literal values, variables, binary operations, and function calls are\nexamples of expressions.<\/p><\/li>\n<li><p>Statements are <em>executed<\/em> to take some action.\nVariable assignments, loops, conditionals, and raising exceptions are\nexamples of statements.<\/p><\/li>\n<li><p>Declarations <em>alter our environment<\/em>, often by\ndefining something. Function and type definitions, imports, and variable\ndeclarations (in a language that distinguishes them from assignments, such\nas C or Javascript) are declarations.<\/p><\/li>\n<li><p>Patterns tell us how to bind variables. They get\n<em>matched<\/em> against values. For example, in Python:<\/p>\n<pre>x,y = (1,2)\n<\/pre>\n<p>This matches the pattern <code class=\"sc-pat\">x,y<\/code> against the value\n<code class=\"sc-expr\">(1,2)<\/code>. The match succeeds, binding <code>x<\/code> to <code>1<\/code>\nand <code>y<\/code> to <code>2<\/code>. If instead we wrote:<\/p>\n<pre>x,y = (1,2,3)\n<\/pre>\n<p>Then the match would fail, since <code class=\"sc-pat\">x,y<\/code> expects a\n2-element sequence, and <code class=\"sc-expr\">(1,2,3)<\/code> has 3 elements.\nThis raises a <code>ValueError<\/code>.<\/p>\n<p>Even an ordinary variable assignment like <code class=\"sc-stmt\">x = 3<\/code> involves a\npattern: the varia\u00adble <code class=\"sc-pat\">x<\/code>. <code class=\"sc-pat\">x<\/code> is simply a pattern that always matches, binding <code>x<\/code>\nto whatever it is matched against.<\/p><\/li>\n<\/ul><p>This is not an exhaustive list. These classes are fuzzy. Different languages\ntreat them differently. Many languages collapse two or more into one category.\nLisps don't distinguish between statements and expressions; Python combines\nstatements and declarations; Haskell lacks statements entirely. Some languages\n(Prolog, Forth, sed, SQL, BNF) aren't easy to describe in these terms at all!<\/p><h2>2 The power of patterns<\/h2><p>Most mainstream languages only have very simple patterns. But in some languages,\nparticularly those descended from the ML family, patterns are quite powerful and\npattern-matching is a core construct. Some examples, in\n:<\/p><pre>fibonacci 0 = 1\nfibonacci 1 = 1\nfibonacci n = fibonacci (n-1) + fibonacci (n-2)\nsum [] = 0               -- list is empty\nsum (x:xs) = x + sum xs  -- list is `x' followed by `xs'\n-- Finds all adjacent pairs:\n-- adjacent [1,2,3] == [(1,2), (2,3)]\nadjacent (x:y:rest) = (x,y) : adjacent (y:rest)\nadjacent _ = []\n-- Searching a binary search tree\ndata Tree = Empty | Node Int Tree Tree\ninTree x Empty = False\ninTree x (Node v l r) =\n    case compare x v of     -- matches on (compare x v)\n      EQ -&gt; True            -- x = v, EQ for \"equals\"\n      LT -&gt; inTree x l      -- x &lt; v, LT for \"less than\"\n      GT -&gt; inTree x r      -- x &gt; v, GT for \"greater than\"\n<\/pre><pre>fun fibonacci 0 = 1\n  | fibonacci 1 = 1\n  | fibonacci n = fibonacci (n-1) + fibonacci (n-2)\nfun sum [] = 0\n  | sum (x::xs) = x + sum xs\n(* Finds all adjacent pairs:\n * adjacent [1,2,3] == [(1,2), (2,3)] *)\nfun adjacent (x::y::rest) = (x,y) :: adjacent (y::rest)\n  | adjacent _ = []\n(* Searching a binary search tree *)\ndatatype tree = Empty | Node of int * tree * tree\nfun inTree x Empty = false\n  | inTree x (Node(v,l,r)) =\n    case Int.compare (x,v)\n     of EQUAL =&gt; true\n      | LESS =&gt; inTree x l\n      | GREATER =&gt; inTree x r\n<\/pre><pre>let rec fibonacci = function\n  | 0 -&gt; 1\n  | 1 -&gt; 1\n  | n -&gt; fibonacci (n-1) + fibonacci (n-2)\nlet rec sum = function\n  | [] -&gt; 0\n  | (x::xs) -&gt; x + sum xs\n(* Finds all adjacent pairs:\n * adjacent [1;2;3] == [(1,2); (2,3)] *)\nlet rec adjacent = function\n  | (x::y::rest) -&gt; (x,y) :: adjacent (y::rest)\n  | _ -&gt; []\n(* Searching a binary search tree *)\ntype tree = Empty | Node of int * tree * tree\nlet rec inTree x t = match t with\n  | Empty -&gt; false\n  | Node(v,l,r) -&gt; x == v\n                || inTree x (if x <\/pre><pre>fn fib(n: int) -&gt; int {\n    match n {\n    0 =&gt; 1,\n    1 =&gt; 1,\n    n =&gt; fib(n-1) + fib(n-2),\n}\n}\nfn sum(l: &amp;[int]) -&gt; int {\n    match l {\n    [] =&gt; 0,\n    [x, ..xs] =&gt; x + sum(xs),\n}\n}\n\/\/ Finds all adjacent pairs:\n\/\/ adjacent(&amp;[1,2,3]) == vec![(2,3), (1,2)]\nfn adjacent(l: &amp;[int]) -&gt; Vec&lt;(int,int)&gt; {\n    match l {\n    [x, .. xs @ [y, ..]] =&gt; adjacent(xs) + &amp;[(x, y)],\n    _ =&gt; vec![]\n}\n}\n\/\/ Searching a binary search tree\nenum Tree { Empty, Node(int, Box&lt;Tree&gt;, Box&lt;Tree&gt;) }\nfn inTree(x: int, t: &amp;Tree) -&gt; bool {\n    match t {\n        &amp;Empty =&gt; false,\n        &amp;Node(v, ref l, ref r) =&gt; match x.cmp(&amp;v) {\n            Equal =&gt; true,\n            Less =&gt; inTree(x, *l),\n            Greater =&gt; inTree(x, *r),\n        }\n    }\n}\n<\/pre><pre>(define\/match (fibonacci n)\n  [((or 0 1)) 1]\n  [(n) (+ (fibonacci (- n 1)) (fibonacci (- n 2)))])\n(define\/match (sum l)\n  [('()) 0]\n  [((cons x xs)) (+ 1 (sum xs))])\n;; Finds adjacent pairs:\n;; (adjacent '(a b c)) == '((a b) (b c))\n(define\/match (adjacent l)\n  [((cons x (and xs (cons y _))))\n   (cons (list x y) (adjacent xs))]\n  [(_) '()])\n;; Searching a binary search tree\n(struct node (value left right) #:transparent)\n(define\/match (in-tree x tree)\n  [(x (node v l r))\n   (or (= x v) (in-tree x (if (<\/pre><p>Many language features are simply special-cases of pattern-matching. We've\nalready seen that Python's tuple unpacking is a form of pattern-matching.\nOptional arguments and rest arguments (<code>*args<\/code> in Python &amp; Ruby) are\npatterns matched against a list of arguments. Keyword arguments are patterns\nmatched against a <em>dictionary<\/em> of arguments. <code>if<\/code> is just pattern-matching on\nbooleans:<\/p><pre>if CONDITION { THEN-CASE ... }\nelse { ELSE-CASE ... }\n<\/pre><p>Is the same as:<\/p><pre>case CONDITION of\n  True -&gt; THEN-CASE ...\n  False -&gt; ELSE-CASE ...\n<\/pre><p>Many useful domain-specific languages are based on pattern-matching. Regular\nexpressions express patterns of strings; <code>sed<\/code> and <code>awk<\/code> employ this to great\neffect. <code>yacc<\/code> uses a more powerful cousin of this idea, extending patterns to\ngrammars. The <a href=\"http:\/\/tinlizzie.org\/ometa\/\">OMeta<\/a> language attempts to push the\npower of pattern-matching as far as possible in a general-purpose setting.<\/p><h2>3 Pattern-matching in Lisps<\/h2><p>Pattern-matching is not native to Lisp as it is to the ML\nfamily. Neither Common Lisp nor Scheme nor Clojure have a general-purpose\npattern-matching form built-in, much less integrated into function\ndefinition.<sup class=\"fnlink\"><a href=\"#fn-2\">2<\/a><\/sup><\/p><p>But adding language features is Lisp's strong point: we can\nimplement pattern-matching as a macro!<sup class=\"fnlink\"><a href=\"#fn-3\">3<\/a><\/sup>\nAs a toy example, here's an <code>if-match<\/code> macro, where <code>(if-match pat expr then\nelse)<\/code> runs <code>then<\/code> if <code>pat<\/code> matches <code>expr<\/code> and otherwise runs <code>else<\/code>. In\n:<\/p><pre>(defmacro if-match [pat expr then else]\n  (cond\n    ;; `pat' is a variable\n    (symbol? pat) `(let [~pat ~expr] ~then)\n    ;; `pat' is a literal atom\n    (not (seq? pat)) `(if (= '~pat ~expr) ~then ~else)\n    ;; `pat' is (cons A B) with subpatterns A, B\n    (= 'cons (first pat))\n      (let [[_ p-first p-rest] pat\n            tmp (gensym)]\n        `(let [~tmp ~expr]\n           (if (not (seq? ~tmp)) ~else\n             (if-match ~p-first (first ~tmp)\n               (if-match ~p-rest (rest ~tmp) ~then ~else)\n               ~else))))\n    :else (throw (Exception. \"Unrecognized pattern\"))))\n<\/pre><pre>(defmacro if-match (pat expr then else)\n  (cond\n    ;; `pat' is a variable\n    ((and pat (symbolp pat)) `(let ((,pat ,expr)) ,then))\n    ;; `pat' is a literal atom\n    ((atom pat) `(if (equalp ',pat ,expr) ,then ,else))\n    ;; `pat' is (cons A B) with subpatterns A, B\n    ((eq 'cons (car pat))\n      (destructuring-bind (_ p-car p-cdr) pat\n        (let ((tmp (gensym)))\n          `(let ((,tmp ,expr))\n             (if (not (consp ,tmp)) ,else\n               (if-match ,p-car (car ,tmp)\n                 (if-match ,p-cdr (cdr ,tmp) ,then ,else)\n                 ,else))))))\n    (t (error \"Unrecognized pattern\"))))\n<\/pre><pre>(define-syntax if-match\n  (syntax-rules (cons)\n    ;; pattern is (cons A B) with subpatterns A, B\n    [(_ (cons p-car p-cdr) expr then else)\n      (let ((x expr))\n        (if (not (pair? x)) else\n          (if-match p-car (car x)\n            (if-match p-cdr (cdr x) then else) else)))]\n    [(_ pat expr then else)\n      (<a href=\"http:\/\/okmij.org\/ftp\/Scheme\/macros.html#macro-symbol-p\">symbol??<\/a> pat    ;; deep magic\n        ;; pattern is a variable\n        (let ((pat expr)) then)\n        ;; otherwise, pattern is a literal\n        (if (equal? 'pat expr) then else))]))\n<\/pre><pre>(require syntax\/parse\/define)\n(define-syntax-parser if-match\n  #:literals (cons)\n  ;; pattern is a variable\n  [(_ x:id expr then else) #'(let ((x expr)) then)]\n  ;; pattern is (cons A B) with subpatterns A, B\n  [(_ (cons p-car p-cdr) expr then else)\n    #'(let ((x expr))\n        (if (not (pair? x)) else\n          (if-match p-car (car x)\n            (if-match p-cdr (cdr x) then else) else)))]\n  ;; pattern is an atomic literal\n  [(_ (~and lit (~not (_ ...))) expr then else)\n    #'(if (equal? 'lit expr) then else)])\n<\/pre><p class=\"ex-macro racket inactive\">(Of course, in Racket you could just use <a href=\"http:\/\/docs.racket-lang.org\/reference\/match.html\">match<\/a>, but now we're\ngetting ahead of ourselves.)<\/p><p>We can use it like so:<\/p><pre>(if-match 0 0 true false)                 ; --&gt; true\n(if-match 0 1 true false)                 ; --&gt; false\n(if-match (cons 0 x) '(0 hi there) x nil) ; --&gt; (hi there)\n(if-match (cons 0 x) '(1 hi there) x nil) ; --&gt; nil\n<\/pre><pre>(if-match 0 0 t nil)                      ; --&gt; T\n(if-match 0 1 t nil)                      ; --&gt; NIL\n(if-match (cons 0 x) '(0 hi there) x nil) ; --&gt; (HI THERE)\n(if-match (cons 0 x) '(1 hi there) x nil) ; --&gt; NIL\n<\/pre><pre>(if-match 0 0 #t #f)                     ; --&gt; #t\n(if-match 0 1 #t #f)                     ; --&gt; #t\n(if-match (cons 0 x) '(0 hi there) x #f) ; --&gt; (hi there)\n(if-match (cons 0 x) '(1 hi there) x #f) ; --&gt; #f\n<\/pre><p>So far, so good, right? <em>But<\/em>...<\/p><h3>The Library Problem<\/h3><p>Now that I've defined my handy-dandy pattern-matching macro <code>if-match<\/code>, I'm\n<em>stuck<\/em> with whatever patterns I thought up when I wrote it. If I come up with a\nnew pattern, I have to rewrite <code>if-match<\/code> to support it.<\/p><p>This sucks! Consider: Hilary the Hacker writes a spiffy lazy stream library,\nwhile Charlie the Coder writes a fast B-tree library. Naturally, both want to\nsupport pattern-matching. They can either:<\/p><ol><li><p>Each implement their own custom pattern-matching macros. This duplicates\nwork, defeats standardization, and makes it impossible to pattern-match on a\nstream and a tree at the same time.<\/p><\/li>\n<li><p>Bug me to include support for Hilary's streams and Charlie's trees in my\n<code>if-match<\/code> library. This introduces irrelevant dependencies, and over time, a\nbig ugly monolithic ball of code will accrete.<sup id=\"fnref-4\" class=\"fnlink\"><a href=\"#fn-4\">4<\/a><\/sup><\/p><\/li>\n<\/ol><p>The whole meaning of <em>modularity<\/em> is avoiding this rock-and-a-hard-place choice\nbetween duplicating work and centralizing it.<\/p><h3>Generalized macros<\/h3><p>The problem with <code>if-match<\/code> is that it requires you to think of everything in\nadvance \u2014 exactly the difficulty <em>macros<\/em> are supposed to solve! Macros\nare Lisp's standard lan\u00adguage ext\u00adension mechanism, and patterns are\njust a small, domain-specific lan\u00adguage.<\/p><p>So why can't we just define new patterns <em>with macros<\/em>?<\/p><p>Here's how this might work. Suppose we already know how to match the following\npatterns:<\/p><table id=\"pattern-table\"><tbody><tr><td> <code class=\"sc-pat\">nil<\/code> <\/td> <td> Matches an empty list. <\/td> <\/tr><tr><td> <code class=\"sc-pat\">(cons <i>A<\/i> <i>B<\/i>)<\/code> <\/td>\n<td> Matches a list whose first element matches <code class=\"sc-pat\"><i>A<\/i><\/code>\n<br \/>and whose remainder matches <code class=\"sc-pat\"><i>B<\/i><\/code>. <\/td>\n<\/tr><\/tbody><\/table><p>It would be handy to have a pattern <code class=\"sc-pat\">(list\n<i>A<\/i><sub>1<\/sub> ... <i>A<sub>n<\/sub><\/i>)<\/code>, which matches an\n<i>n<\/i>-element list whose 1<sup>st<\/sup> element matches <code class=\"sc-pat\"><i>A<\/i><sub>1<\/sub><\/code>, whose 2<sup>nd<\/sup> element matches\n<code class=\"sc-pat\"><i>A<\/i><sub>2<\/sub><\/code>, ... and so on up to <code class=\"sc-pat\"><i>A<sub>n<\/sub><\/i><\/code>. For example, <code class=\"sc-pat\">(list x 2 y)<\/code> would match <code>(1 2 3)<\/code>, binding <code>x<\/code> to <code>1<\/code> and\n<code>y<\/code> to <code>3<\/code>.<\/p><p>I'm proposing that this be done using a <em>macro for patterns<\/em>, with something\nlike the following mock- syntax:<\/p><pre>(defpattern list [&amp; element-pats]\n  (if (empty? element-pats) 'nil\n    (let [[p &amp; ps] element-pats]\n      `(cons ~p (list ~@ps)))))\n<\/pre><pre>(defpattern list (&amp;rest element-pats)\n  (if (not element-pats) 'nil\n    (destructuring-bind (p &amp;rest ps) element-pats\n      `(cons ,p (list ,@ps)))))\n<\/pre><pre>(define-pattern-syntax list\n  (syntax-rules ()\n    [(_) nil]\n    [(_ p ps ...) (cons p (list ps ...))]))\n<\/pre><p>This solves Hilary and Charlie's library problem: instead of rolling their own\nincompatible pattern-matchers, or asking me to rewrite <code>if-match<\/code>, they\nseparately extend <code>if-match<\/code> using <code class=\"ex-defpattern clojure&#10;lisp\">defpattern<\/code><code class=\"ex-defpattern scheme&#10;inactive\">define-pattern-syntax<\/code>. These extensions are pulled into scope\nwhen I import their libraries, just as a macro would be.<sup id=\"fnref-5\" class=\"fnlink\"><a href=\"#fn-5\">5<\/a><\/sup><\/p><p>This idea is not original to me: The <a href=\"http:\/\/racket-lang.org\/\">Racket<\/a>\nlanguage's <a href=\"http:\/\/docs.racket-lang.org\/guide\/match.html\"><code>match<\/code><\/a> macro is\nextensible via\n<a href=\"http:\/\/docs.racket-lang.org\/reference\/match.html#%28form._%28%28lib._racket%2Fmatch..rkt%29._define-match-expander%29%29\"><code>define-match-expander<\/code><\/a>.\nThe <a href=\"https:\/\/github.com\/m2ym\/optima\">optima<\/a> library for Common Lisp exposes a\n<code>defpattern<\/code> macro as well. I don't know of a Clojure library that uses this\napproach <em>per se<\/em>, but <a href=\"https:\/\/github.com\/clojure\/core.match\"><code>core.match<\/code><\/a> is\nextensible in <a href=\"https:\/\/github.com\/clojure\/core.match\/wiki\/Advanced-usage\">other<\/a>\n<a href=\"https:\/\/github.com\/clojure\/core.match\/wiki\/Extending-match-for-new-Patterns\">ways<\/a>.<\/p><p>Ultimately, this is a partial success story for Lisp. ML-style pattern-matching\ncan be assimilated and macro-ified. With some work, it can even be made\nmacro-extensible. This design is not yet idiomatic among Lisps, but hopefully in\ntime it will be. And in any language <em>but<\/em> a Lisp, would this kind of extension\neven be thinkable?<sup id=\"fnref-6\" class=\"fnlink\"><a href=\"#fn-6\">6<\/a><\/sup><\/p><h2>4 DSLs as syntax classes<\/h2><p>Hold on \u2014 isn't this an essay about syntax classes, and how not everything\nis an expression? What does the design of pattern-matching macros have to do\nwith <em>that<\/em>? Let's rephrase our progress so far in terms of syntax classes:<\/p><p>Pattern-matching doesn't just add a new kind of expression, it adds a new\n<em>syntax class<\/em>: patterns. The problem with <code>if-match<\/code> is that it doesn't allow\nmodularly extending the syntax of patterns. We already know how to modularly\nextend <em>expression<\/em> syntax: macros! By extending macros to cover <em>patterns<\/em> as\nwell, we solve the library problem.<\/p><p>Now that we're thinking in terms of syntax classes, the problem and solution\nhave become much clearer!<\/p><p>But there's nothing special about patterns. Many domain-specific languages are\nsimply new syntax classes on top of an existing language. Microsoft's\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Language_Integrated_Query\">LINQ<\/a> adds a syntax\nclass for <em>queries<\/em>, for example. Expressions embed queries to be performed, and\nqueries embed expressions representing (e.g.) predicates to be filtered on or\nkeys to be looked up. Parser-generators are another example: typically, we write\nparsing rules in a special syntax, like BNF-inspired <code>yacc<\/code> rules or PEG parsing\nexpressions. Into these rules we embed <em>semantic actions<\/em> \u2014 expressions\nthat run when a parse rule matches.<\/p><p>The <a href=\"http:\/\/people.csail.mit.edu\/gregs\/ll1-discuss-archive-html\/msg01539.html\">canonical\nuses<\/a>\nof Lisp macros are: 1. changing order of evaluation, 2. creating new binding or\ndefinition forms, and 3. making domain-specific languages. LINQ and parsing are\nclear-cut examples of case 3. But the obvious way to implement a DSL as a macro,\nas we saw with <code>if-match<\/code>, hard-codes the form of the new syntax class, giving\nup on extensibility. Because of this, I think Lisp macros aren't good enough at\nmaking domain-specific languages.<\/p><p>Consider: Each of the pattern-matching libraries I listed above had to design\nand implement a extension mechanism from scratch \u2014 even though Lisp\n<em>already has<\/em> an extension mechanism: macros! And patterns are just one of\n<em>many<\/em> useful domain-specific languages we might desire. Introducing a new\nextensible syntax class shouldn't require us to reinvent the wheel! Lisp doesn't\nhave a standard way to define new syntax classes, one which makes them\nmacro-extensible <em>by default<\/em>. It should.<sup id=\"fnref-7\" class=\"fnlink\"><a href=\"#fn-7\">7<\/a><\/sup><\/p><p><\/p><aside><small>This essay first appeared in\n<a href=\"https:\/\/codewords.recurse.com\/issues\/two\/not-everything-is-an-expression\">Code Words<\/a>.\nThanks to Jari Takkala and Aki Yamada for editing this article.<\/small><\/aside><ol id=\"footnotes\"><li id=\"fn-1\">\n  <p>In <cite>Practical Foundations for Programming Languages<\/cite>, these are\n    called \"sorts\". I'd love to hear whether other textbooks have a name for this\n    concept; perhaps I'm missing some piece of standard terminology.\n    <a class=\"fnback\" href=\"#fnref-1\">\u21a9<\/a><\/p>\n<\/li>\n<li id=\"fn-2\">\n  <p>Common Lisp has\n    <a href=\"http:\/\/www.lispworks.com\/documentation\/HyperSpec\/Body\/m_destru.htm#destructuring-bind\"><code>destructuring-bind<\/code><\/a>\n    and Clojure\n    has <a href=\"http:\/\/clojure.org\/special_forms#Special%20Forms--Binding%20Forms%20%28Destructuring%29\">structural\n    binding<\/a>, but those don't backtrack and can't match literals\n    (like <code>0<\/code> or <code>'t<\/code>), so you can't use them for\n    case-analysis. Scheme has <code>syntax-rules<\/code>\n    and <code>syntax-case<\/code> (in R6RS), but they're only useful for defining\n    macros.\n    <a class=\"fnback\" href=\"#fnref-2\">\u21a9<\/a><\/p>\n<\/li>\n<li id=\"fn-3\">\n  <p> Indeed, there are many libraries implementing pattern-matching macros.\n    Here's a <a href=\"http:\/\/www.cliki.net\/pattern%20matching\">pageful<\/a> for\n    Common Lisp. Clojure has\n    <a href=\"https:\/\/github.com\/clojure\/core.match\">core.match<\/a>. Scheme\n    doesn't really have a library ecosystem,\n    but <a href=\"http:\/\/okmij.org\/ftp\/Scheme\/macros.html#match-case-simple\">here's\n    a file you can use<\/a>. Many Schemes also provide their own libraries; for\n    example, Racket has a powerful\n    <a href=\"http:\/\/docs.racket-lang.org\/guide\/match.html\"><code>match<\/code><\/a>\n    macro in the standard library.\n    <a href=\"#fnref-3\" class=\"fnback\">\u21a9<\/a>\n  <\/p>\n<\/li>\n<li id=\"fn-4\">\n  <p>Moreover, suppose Hilary's stream library itself uses pattern-matching.\n    Then we have a circular dependency: <code>if-match<\/code> depends on\n    Hilary's streams, and Hilary's streams depend on <code>if-match<\/code>! Few\n    languages support this kind of library-level mutual recursion.\n    <a href=\"#fnref-4\" class=\"fnback\">\u21a9<\/a>\n  <\/p>\n<\/li>\n<li id=\"fn-5\">\n  <p>Of course, the devil is in the details: how is <code>defpattern<\/code> \/\n    <code>define-pattern-syntax<\/code> implemented? How does\n    <code>if-match<\/code> ask what extensions have been defined via\n    <code>defpattern<\/code>? How do we avoid namespace collisions? <a href=\"http:\/\/www.rntz.net\/post\/2014-06-27-monoids-scope-extensibility.html\">This\n    article<\/a> lays out some of the theory behind my approach to answering\n    these questions, but I don't have all the answers yet. See also footnotes <a href=\"#fn-6\">6<\/a> and <a href=\"#fn-7\">7<\/a>. <a href=\"#fnref-5\" class=\"fnback\">\u21a9<\/a>\n  <\/p>\n<\/li>\n<li id=\"fn-6\">\n  <p>Actually, yes. Quite a bit of work has been done on non-Lisp languages with\n    extensible syntax, most notably the\n    <a href=\"http:\/\/www.student.informatik.tu-darmstadt.de\/~xx00seba\/projects\/sugarj\/\">SugarJ<\/a>\n    project. <a href=\"https:\/\/github.com\/rntz\/moxy\">Moxy<\/a>, the language I built at\n    <a href=\"http:\/\/www.hackerschool.com\">Hacker School<\/a>, is another (as yet\n    undocumented) example.\n    <a href=\"#fnref-6\" class=\"fnback\">\u21a9<\/a>\n  <\/p>\n<\/li>\n<li id=\"fn-7\">\n  <p>In fact, once upon a time, PLT Scheme (the language which became Racket)\n    did! It was called\n    <a href=\"http:\/\/cs.brown.edu\/~sk\/Publications\/Papers\/Published\/kfd-macro-to-gen-prog\/\">McMicMac<\/a>,\n    and I only discovered it after I was nearly done writing this essay. Its\n    notion of extensible syntax class was called a <em>vocabulary<\/em>, and\n    extensions were <em>micros<\/em> (by analogy with \"macros\").\n    It <a href=\"http:\/\/lists.racket-lang.org\/users\/archive\/2003-June\/002978.html\">apparently\n    had problems<\/a>, however, and was replaced\n    by <a href=\"https:\/\/www.cs.utah.edu\/plt\/publications\/macromod.pdf\">an\n    ancestor of Racket's current macro system<\/a>, which sadly lacks the ability\n    to define vocabularies.\n    <a class=\"fnback\" href=\"#fnref-7\">\u21a9<\/a>\n  <\/p>\n<\/li>\n<\/ol><\/header>","created_at":"2018-11-07T16:53:38+0000","updated_at":"2018-11-07T16:53:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":15,"domain_name":"www.rntz.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":611,"title":"Doomsday prepping for less crazy folk","url":"http:\/\/lcamtuf.coredump.cx\/prep\/","content":"Doomsday prepping for less crazy folk\n<p>Written by \n<a href=\"mailto:lcamtuf@coredump.cx\">lcamtuf@coredump.cx<\/a>, last update Jan 2018. \nTwitter: <a href=\"https:\/\/twitter.com\/lcamtuf\">@lcamtuf<\/a>. More about the author\n<a href=\"https:\/\/lcamtuf.blogspot.com\/2015\/03\/on-journeys.html\">here<\/a>.<\/p> \n<h2>1. Introduction [<a href=\"#1\">link<\/a>]<\/h2>\n<p>\nThe prepper culture begs to be taken with a grain of salt. In a sense, it has\nall the makings of a doomsday cult: a tribe of unkempt misfits who hoard gold\nbullion, study herbalism, and preach about the imminent collapse of our society.\n<\/p>\n<p>\nToday, we see such worries as absurd. It's not that life-altering disasters are\nrare: every year, we hear about millions of people displaced by wildfires, earthquakes,\nhurricanes, or floods. Heck, not a decade goes by without at least one first-class\ndemocracy lapsing into armed conflict or fiscal disarray. But having grown up in a period\nof unprecedented prosperity and calm, we take our way of life for granted - and find\nit difficult to believe that an episode of bad weather or a currency crisis could\ndestroy almost everything we worked for to date.\n<\/p>\n<p>\nI suspect that we dismiss such hazards not only because they seem surreal, but also because\nworrying about them makes us feel helpless and lost. What's more, we follow the same instincts\nto tune out far more pedestrian and avoidable risks; for example, \nmost of us don't plan ahead for losing a job, for dealing with a week-long water outage, or\nfor surviving the night if our home goes up in smoke.\n<\/p>\n<p>\nFor many, the singular strategy for dealing with such dangers is to pray for the\ngovernment to bail us out. But no matter if our elected officials prefer to school us with\npassages from\n<a href=\"http:\/\/smile.amazon.com\/dp\/0156334607\">Milton\nFriedman<\/a> or from\n<a href=\"http:\/\/smile.amazon.com\/dp\/0486477487\" title=\"I'm sorry... I'm really sorry!\">Thomas\nPiketty<\/a>, the hard truth is that no state can provide a robust safety net for all\nof life's likely contingencies; in most places, government-run social programs are severely deficient in funding, in\nefficiency, and in scope. Large-scale disasters pit us against even worse odds; from New Orleans in 2005 to\nFukushima in 2011, there are countless stories of people left behind due to political dysfunction, poorly\nallocated resources, or lost paperwork.\n<\/p>\n<p>\nAnd so, the purpose of this guide is to combat the mindset of learned helplessness by\npromoting simple, level-headed, personal preparedness techniques that are easy to\nimplement, don't cost much, and will probably help you cope with whatever life throws your way.\n<\/p>\n<p>\nOh, one thing: in contrast to most other docs of its kind, this page is pure, unadulterated labor of love; there\nare no affiliate links, paid product placements, or ads anywhere in the guide.\n<\/p>\n<h2>2. Mapping out the unknown [<a href=\"#2\">link<\/a>]<\/h2>\n<p>\nEffective preparedness can be simple, but it has to be rooted in an honest and\nsystematic review of the risks you are likely to face. Plenty of excited newcomers begin\nby shopping for ballistic vests and night vision goggles; they would be better\nserved by grabbing a fire extinguisher, some bottled water, and then putting the rest of\ntheir money in a rainy-day fund.\n<\/p>\n<p>\nTo maintain sanity while trying to enumerate risks, I found that it's best to focus on\nbroad outcomes instead of trying to track down every single way for things to go south.\nSay, it should not matter if you are laid off because of a downsizing, because\nyour new boss hates you, or because they finally catch you stealing paperclips. The\noutcome is the same: you are out of a job and urgently need a way to pay your bills.\n<\/p>\n<p>\nAnother insidious distraction is the desire to immediately figure out how to respond to all the scenarios\nwe end up dreaming of. Let's save that for later; by prematurely focusing on the second half of the\nproblem, we may end up glossing over some of the less tractable scenarios - or make\nhaphazard assumptions that will cloud our judgment in other ways.\n<\/p>\n<p>\nI also found that to come up with a rational threat model, we need to think of \"risk\" as a product of\nboth the probability and the consequences of a given event. By that metric, stubbed toes\nand zombie outbreaks are equally uninteresting; one of them has nearly zero significance,\nthe other, nearly zero odds.\n<\/p>\n<p>\nWhat else? Ah, right: the final piece of advice I have is to keep things uncomplicated. There are\npopular doomsday predictions that deal with cutting-edge particle physics, god-like computer\nhackers, vast government conspiracies, or extraterrestrial messages hidden in pop songs. I suppose\nwe can't <i>really<\/i> rule that stuff out, but historical data suggests that there's a lot more\nmerit in worrying about falling off a ladder or getting hit by a car.\n<\/p>\n<p>\nAll right! With these caveats in mind, let's go over some canonical scenarios that are worth thinking about.\n<\/p>\n<h3>2.1. Problem space #1: Small-scale events [<a href=\"#2.1\">link<\/a>]<\/h3>\n<p>\nIt's always fun to speculate about solar flares and supervolcanoes; it's far more mind-numbing to\nseriously evaluate the consequences of backed up sewage or burst water mains. But in reality,\nsuch unglamorous, small-scale incidents are far more likely to disrupt and reshape our\nlives.\n<\/p>\n<p>\nBroadly speaking, disastrous outcomes of such humdrum contingencies can be divided into\nseveral groups:\n<\/p>\n<ul><li>\n<p>\n<b>Insolvency.<\/b>\n  If a person over the age of 40 tells you that they have never lost a job, they are\n  pretty lucky (or lying). Yet, the risk is seldom taken seriously; many middle-class,\n  single-income families would be in deep trouble if it ever took them more than 2-3 months\n  to find a new, equally well-paying gig.\n<\/p>\n<\/li><li>\n<p>\n<b>Disrupted access to water, food, energy, or transportation.<\/b>\n  Substantial and prolonged outages happen everywhere; many of us will experience\n  at least one at some point in our lives. A week without electricity may be just\n  inconvenient and scary, especially in a high-rise or\n  in a seedy neighborhood; but even a single hot day without potable water is universally\n  bad news.\n<\/p>\n<\/li><li>\n<p>\n<b>Loss of shelter.<\/b>\n  Every year, there are over 350,000 house fires in the United States. Such accidents\n  usually aren't deadly - but if you are unlucky, they can leave you stranded in the middle\n  of the night in your PJs, with no documents or credit cards at hand.\n<\/p>\n<\/li><li>\n<p>\n<b>Unintentional injury.<\/b>\n  Largely preventable and predictable incidents - such as falls, vehicle collisions, and poisonings -\n  account for some 40 million ER visits annually. And lest you say people are simply too\n  quick to rush to the hospital, said incidents also result in about 100,000 US\n  deaths every year.\n<\/p>\n<\/li><li>\n<p>\n<b>Intentionally inflicted harm.<\/b>\n  Violent crime is essentially <i>normal<\/i> almost everywhere in the world.\n  In the US in the 90s, your lifetime likelihood of victimization was\n  estimated to be around 80%; the odds of suffering criminal injury hovered at 40%. Today, the numbers\n  are likely much better - but life-threatening encounters remain a very real risk.\n<\/p>\n<\/li><li>\n<p>\n<b>Debilitating illness or death.<\/b>\n  It's going to get you; maybe next week, maybe in 50 years. We can't really predict the day,\n  but we can understand and meaningfully manage the impact it will have on those who depend on us -\n  say, our stay-at-home partners or young kids.\n<\/p>\n<\/li><\/ul><p>\nAll in all, the risks discussed in this section have three defining characteristics: they are relatively\nlikely to happen; are strikingly easy to mitigate (we'll get into that soon); and tend to be so\nunglamorous that they seldom make the cut in any \"serious\" guide to emergency preparedness.\n<\/p>\n<h3>2.2. Problem space #2: Mass calamities [<a href=\"#2.2\">link<\/a>]<\/h3>\n<p>\nIf an errant backhoe took out the utilities for your block, you would probably head to the\ngrocery store to pick up bottled water (and use their restrooms, too). But if a\nonce-in-a-century storm damaged major roads and left half the city without running water, your\noptions wouldn't be as clear-cut.\n<\/p>\n<p>\nThat's why we have to look at larger-scale emergencies through somewhat different lens, taking into\naccount their likely magnitude, duration, and the nature of the forces at play. Some of the\nplausible scenarios to think about include:\n<\/p>\n<ul><li>\n<p>\n<b>Natural disasters.<\/b>\n  Common examples include floods, hurricanes, earthquakes, wildfires, and heatwaves. In some\n  regions, such events are very rare; in others, they are almost guaranteed every decade or two.\n<\/p>\n<\/li><li>\n<p>\n<b>Industrial accidents.<\/b>\n  Many people live in the proximity of heavy industries - say, refineries, freight railroads, or power plants.\n  Depending on the type of industrial facilities nearby, you may want to evaluate the potential\n  consequences of upwind and upstream explosions or chemical spills.\n<\/p>\n<\/li><li>\n<p>\n<b>Social unrest.<\/b>\n  Riots are a distinct risk in many urban and suburban areas around the world. When angry mobs\n  take it to the streets, widespread arson and violent crime are not unheard of, sometimes going\n  on for days or weeks.\n<\/p>\n<\/li><li>\n<p>\n<b>Economic crises.<\/b>\n  All highly developed countries go through cyclic recessions and periods of high unemployment;\n  the US had about ten big ones in the past 100 years alone. Sometimes, such events\n  are accompanied by bank runs and collapses of financial institutions; other times,\n  they involve hyperinflation, product rationing, and currency controls.\n<\/p>\n<\/li><li>\n<p>\n<b>Pandemic.<\/b>\n  It's been a while since the highly developed world experienced a devastating outbreak, but it\n  may be premature to flat out dismiss the risk. In 1918, an unusual strain of flu managed to kill 75\n  million people. We aren't necessarily better prepared for a similar event today.\n<\/p>\n<\/li><li>\n<p>\n<b>Terrorism or conventional war.<\/b>\n  We think we would see it coming - but history shows that such events tend to catch nations\n  off guard. These phenomena are noteworthy not only because of their immediate death toll,\n  which can be relatively low - but because of the far-reaching and long-term socioeconomic\n  disruption they can cause.\n<\/p>\n<\/li><\/ul><p>\nMost of us will probably not get tangled up in a large-scale disaster of any sort, but it\nis only wise to hedge your bets. There are countless examples to demonstrate that such events\nhappen often and can strike close to home - say:\n<\/p>\n<ul><li>\n<p>\nThe EU debt crisis, from 2009 onward. A series of events that led to staggering unemployment rates\nin Greece, deposit confiscations in Cyprus, and uncertain prospects for the entire\neurozone.\n<\/p>\n<\/li><li> \n<p>\nHurricane Sandy in 2012. Damaged more than 300,000 homes in the US, left millions \nwithout power - some of them for up to two weeks.\n<\/p>\n<\/li><li> \n<p>\nThe California wildfires of 2007. The blaze destroyed thousands of homes and forced\n1 million SoCal residents to seek improvised shelter at stadiums and schools.\n<\/p>\n<\/li><li>\n<p>\nHurricanes Katrina and Rita in 2005. Put 80% of New Orleans under water, stranded\ncountless motorists when 3 million people attempted to evacuate.\n<\/p>\n<\/li><li>\n<p>\nA decade of social unrest in France, from 2005 on. A series of widespread\nurban riots led to more than 13,000 cars being torched and hundreds of structures\ndamaged or destroyed.\n<\/p>\n<\/li><li>\n<p>\nThe European heatwave of 2003. A seemingly benign weather anomaly killed 70,000 -\nmostly in highly developed and wealthy countries, such as France and Spain.\n<\/p>\n<\/li><li>\n<p>\nThe Los Angeles riots of 1992. An epic rampage that resulted in about 2,000 injuries,\n4,500 structures being looted, and some 1,000 buildings set ablaze.\n<\/p>\n<\/li><\/ul><p>\nPreparing for such emergencies may seem exceedingly difficult, but much can be\naccomplished with very simple tools. For example, if you own a car, always keeping\nthe tank at least half full, and having a small box of rudimentary supplies in your\ntrunk, can go a long way. Again, we'll go over many other low-cost preparedness\nstrategies soon.\n<\/p>\n<h3>2.3. Problem space #3: The zombie apocalypse [<a href=\"#2.3\">link<\/a>]<\/h3>\n<p>\nMany of the best-known works of science fiction explore the possibility that one day,\nperhaps soon, the very fabric of our society may simply unravel - and that those who\nsurvive The Event would be forced to fend for themselves in a ravaged and hostile world.\n<\/p>\n<p>\nTo connect with the reader, the writers of such post-apocalyptic fiction skillfully play up\ncontemporary fears, often implying that we may be on the verge of an extinction event or a\nfundamental societal shift. The usual literary themes include:\n<\/p>\n<ul class=\"compact\"><li> <p>Uniquely virulent and deadly diseases (<i>28 Days Later<\/i>, <i>The Stand<\/i>),<\/p>\n<\/li><li> <p>Preposterously violent weather anomalies (<i>The Day After Tomorrow<\/i>, <i>Snowpiercer<\/i>),<\/p>\n<\/li><li> <p>Unimpeded emergence of totalitarian techno-dystopias (<i>Blade Runner<\/i>, <i>Gatacca<\/i>, <i>The Hunger Games<\/i>),<\/p>\n<\/li><li> <p>Global thermonuclear war (<i>Alas Babylon<\/i>, <i>The Day After<\/i>, <i>Fallout<\/i>),<\/p>\n<\/li><li> <p>Robots deciding to kill all humans (<i>Battlestar Galactica<\/i>, <i>Terminator<\/i>),<\/p>\n<\/li><li> <p>Impertinent extraterrestrials (<i>Edge of Tomorrow<\/i>, <i>Independence Day<\/i>),<\/p>\n<\/li><li> <p>Hand-wavy ecosystem collapses (<i>Children of Men<\/i>, <i>The Road<\/i>, <i>Interstellar<\/i>),<\/p>\n<\/li><li> <p>Cosmic collisions (<i>Lucifer's Hammer<\/i>),<\/p>\n<\/li><li> <p>Something with magnets (<i>The Core<\/i>, <i>One Second After<\/i>).<\/p>\n<\/li><\/ul><p>\nIt's no wonder that all this vivid imagery keeps many preppers preoccupied with civilization-ending events.\nSome of their worries are based on patently absurd or exaggerated science; some are valid, but rather\nunlikely to materialize within the span of our lives; and many others boil down to interesting but \nsomewhat idle speculation, devoid of quantifiable risk or historical precedent.\n<\/p>\n<p>\nAt least in theory, the recipe for surviving civilizational collapse is simple: you need to get away\nfrom other people and become self-sufficient. It's fairly clear that deprived of their industrial\nbackbone, most of our cities and suburbs wouldn't be able to support even a tiny fraction of their\ncurrent population densities - and would become horrid death traps. Living off the grid shields\nyou from all but the worst doomsday events.\n<\/p>\n<p>\nMy advice? Heck yeah, if farming is your cup of tea, buy a plot of land in the countryside.\nBut if you'd rather live the rest of your days without having to skin deer or plow a field,\nfocus on better quantified and substantiated risks - and don't let asteroids or space zombies\nkeep you up at night.\n<\/p>\n<h2>3. The prepared lifestyle [<a href=\"#3\">link<\/a>]<\/h2>\n<p>\nOf course, there is no use in worrying about the hypotheticals if we're not going to actually tackle the\nrisks. But it pays to be smart about it, too: when it comes to emergency preparedness, the\nright mindset can matter a lot more than a small fortune spent on ninja gear and canned cheese.\nSo, before going on a shopping spree, let's talk about several simple, everyday principles\nthat can shield you from harm.\n<\/p>\n<p>\nAs you will undoubtedly notice,\nmuch of the content in this section has little to do with hardcore prepping; some of the chapters\ntouch on seemingly banal topics, such as financial planning, community building, or the prevention\nof burglaries and car wrecks. You have heard most of this advice before - but if you are serious\nabout dealing with adversities and shielding yourself and your family from harm, you need to\ninternalize these rules, understand where they are coming from, and actually live by them every\nday.\n<\/p>\n<p>\n(Of course, before we dive in,\n let's make one thing clear: this guide is an opinion piece, not revealed truth. While \nI did my best to make it thoughtful and accurate, it comes with no guarantees - and you should do\nyour own homework before taking life advice from people on the Internet.)\n<\/p>\n<h3>3.1. Prepper commandment #1: Save some money [<a href=\"#3.1\">link<\/a>]<\/h3>\n<p>\nThere are those who are born into familial wealth or who display supernatural business acumen.\nThen there is the rest of us, perhaps having robust and satisfying careers, but ultimately\ntiptoeing the line between middle-class prosperity and crushing poverty.\nIt may be a matter of our employer going out of business, it may be a shift in\nthe job market, an illness, or a legal dispute - but in all likelihood, it would not take much to\nsend us over the edge. I have friends who lived paycheck-to-paycheck on cozy Silicon Valley\nsalaries of $100k+ a year, only to lose their cars and homes in the midst of the 2007\nfinancial crisis - having found out the hard way that unemployment benefits in the San Francisco\nBay Area max out at $450 a week.\n<\/p>\n<p>\nI can't claim to have good advice for people who are already in a tough spot: if your\nhousehold earnings are well below median, you may simply have no disposable income to build\na personal safety net. But for most other folks, the ability to prepare for the zero-income\ncontingency is well within reach - and it would be unwise not to give it a go. Sure, even a lifetime\nof belt-tightening won't make the average middle-class family fabulously rich. But rainy-day funds\nwork in a different way: their purpose is to get you through a rough spell, not to pay for a\nmansion or a fancy car. Since the amount needed is directly proportional to how much you currently\nmake, it makes relatively little difference if your household brings in $70k or $140k a year.\nEither way, if you set aside 10% of every post-tax paycheck, you should have a 6-month financial\nsafety net established within 3 years and a change.\n<\/p>\n<p>\nThis, of course, is easier said than done. We tend to scale up our living expenses in proportion to\nearned income, so even in the $100k+ bracket, people living paycheck-to-paycheck are not a rare sight.\nAnd it's usually not the big-ticket stuff that gets them: we're far more likely to overspend on all\nthe smaller, habitual purchases, because their cumulative cost is less apparent - and potential\nsavings are much easier to miss. The patterns to look for will depend on your lifestyle and on how\nmuch you make, but here are several suggestions for where to search for that 10%:\n<\/p>\n<ul><li>\n<p>\n<b>Groceries.<\/b>\nTry to shop at less expensive grocery stores and try out lower-shelf brands - especially when\nit comes to commodities such as cooking oil, paper towels, milk, seltzer water, flour, sugar,\nor salt. Table salt tastes and works the same, whether you paid $1 at Walmart or $15 for a\nSherpa-approved Himalayan variety at Whole Foods. Groceries eat up a good chunk of our\nmonthly budgets, so even seemingly inconsequential savings tend to add up very fast.\n<\/p>\n<\/li><li>\n<p>\n<b>Restaurants &amp; taxis.<\/b>\nMany young, urban-dwelling folks frequently dine at restaurants, go to clubs, or take cabs\naround town. If that sounds like you, don't give it up, but scale back slightly by\ndesignating several days a week strictly for public transport and home food. Contemporary\nfrozen dinners can be surprisingly tasty, and Netflix has some good shows.\n<\/p>\n<\/li><li>\n<p>\n<b>Gadget upgrades.<\/b>\nIf your older phone, laptop, or a TV set are still working fine, keep them for another year or two.\nSure, it's fun to play with a new toy, but the excitement wears off quickly, and being a bit\nbehind probably won't make your life feel hollow and pointless. All things considered, you are not impressing\nyour friends that much by showing them that you had $500 or $1,000 to throw away.\n<\/p>\n<\/li><li>\n<p>\n<b>Subscription services.<\/b>\nSmall monthly fees add up to gargantuan sums over the years. Do you really need cable\nTV, or can you watch most of the same shows online for less? Are you still paying\nfor a landline or for that AOL account? How often are you using that gym membership? Can you\ntry out lower speed for your Internet service? Or slightly increase the deductible on your\ncar?\n<\/p>\n<\/li><li>\n<p>\n<b>Home improvement.<\/b>\nEven if you are renting, you can save big bucks by learning how to reupholster or refinish\nfurniture, patch drywall, paint walls and trim, install tile and cabinets, fix or replace faucets,\nunclog drains, and take care of other simple DIY jobs. Home improvement can be incredibly\nrewarding, too.\n<\/p>\n<\/li><li>\n<p>\n<b>Vacation.<\/b>\nIf you're itching to go to a popular destination like Vegas, Hawaii, or Disneyland - how about\na camping trip? For kids and adults alike, sleeping in a tent and playing around a campfire\ncan be more enjoyable and less stressful than waiting in line to meet a middle-aged dude\ndressed as Donald Duck.\n<\/p>\n<\/li><li>\n<p>\n<b>Extra cash.<\/b>\nIf you ever get a bonus or a raise at work, don't squander it by reflexively spending\nmore on mundane, everyday stuff. Put back that fancy bottle of barrel-aged olive oil,\nthat pair of ergonomic toe socks, or that wi-fi enabled toaster. Smile, set the proceeds away,\nand go about your life as if nothing ever changed.\n<\/p>\n<\/li><\/ul><p>\nSomewhat counterintuitively, saving money is not just about cutting down expenses; seeing\na higher balance on a checking account tends to instinctively make us less frugal, too. To\ncounter this trend, I found it helpful to set up a small, daily transfer to\na savings account, in an amount that blends in with daily purchases - say, $10 or so. This\nmethod takes much less planning and mental discipline than trying to make one big deposit every\nmonth. And hey - when you are comfortable with $10, you can painlessly test the limits by\ngently ramping the amount up.\n<\/p>\n<p>\nNow, a word of caution: beware of debt. Many of us are taught that owing money is normal, even\ndesirable; indeed, for middle-class folks, some forms of indebtedness may be difficult\nto avoid. But unnecessarily accrued debt cuts into your bottom line in two insidious ways.\nFirst of all, monthly installment payments limit your flexibility in an emergency - so if your\nincome shrinks, your savings will be depleted at a merciless and non-negotiable rate. Secondly,\nhigh-interest loans, such as credit cards, amount to giving out a good chunk of your income\nwithout earning anything useful in return. They are akin to voluntarily accepting a pay cut.\n<\/p>\n<p>\nAs a matter of practicality, don't worry too much about your existing mortgages or student\nloans: they are difficult to repay early, tend to have very low interest, and confer special\ntax benefits. But use your initial savings to pay off credit card balances, and do it quick.\nBe careful with new obligations, too. Unless you already have a very generous safety net, a\nhome loan that eats up more than 15% of your paycheck over the course of 30 years is a\nvery risky deal; and going over 30% is almost certainly dumb, at least as far as financial\ncontinuity planning goes.\n<\/p>\n<p>\nAs for the optimal size of your emergency fund, there are no hard rules. Around six months'\nworth of post-tax earnings should provide a very comfortable cushion for short- and medium-term\ndisasters of all sorts. Even in a deeper crisis, six months is plenty of time to scale down\nexpenses, find a new line of work, or regroup in some other way.\n<\/p>\n<p>\nNow, chances are, if you follow the advice contained in this section, you will not feel an urge to ramp\nup spending the moment you hit the 6-month mark. So, keep going; with the initial rainy-day fund\nestablished and good fiscal habits in place, you can start treating the extra funds with more\nflexibility - for example, \"borrowing\" against them to self-finance larger purchases, or aiming\nto retire a bit earlier and a bit more comfortably than the government expects you to. And by the\ntime your savings are sufficient to get you through a full year of unemployment, I bet that your\noutlook on life, work, and personal finance will change in an\n<a href=\"https:\/\/thebillfold.com\/a-story-of-a-fuck-off-fund-648401263659\">interesting way<\/a>.\n<\/p>\n<p>\nOh, one more thing: when trying to reform your fiscal habits, it can be useful to think of money as\na unit of debt owed to you by the society, awarded in exchange for your hard work (or as a consequence\nof some other fortuitous event). Despite the\npopular saying, when allocated wisely, that money can buy you safety, comfort, influence, or true\nfriendship and happiness. It would be foolish to squander it on trinkets - just as it would be\nfoolish to take it with you to the grave. My advice is simple: make every transaction count.\n<\/p>\n<h3>3.2. Prepper commandment #2: Don't lose what you saved [<a href=\"#3.2\">link<\/a>]<\/h3>\n<p>\nIf all goes well, your rainy-day fund will eventually grow big enough for you to face a wonderful\nand important question: how do I keep all that capital safe? Although it may seem like a remote\nconcern, events such as bank collapses, market crashes, and currency devaluations happen all over the\nworld with near-clockwork regularity - and there are few things more infuriating and disenfranchising\nthan finding out that the fruits of many years of your labor have been wiped out by a market panic or\nan administrative decree.\n<\/p>\n<p>\nThe answer to the question of safeguarding your wealth lies in the solution\nto another riddle: the mechanism by which the society determines the worth of a piece of money to\nbegin with. It's a puzzle central not only to everyday financial planning, but also to any attempts\nto decipher and meaningfully evaluate countless mainstream conspiracy theories and doomsday\npredictions related to the financial world.\n<\/p>\n<p>\nSo, let's start from the beginning! Throughout much of the recorded history, the monetary systems of the\nwestern world employed so-called <i>commodity money<\/i>, generally settling on coins minted out of\nsilver or gold. The two metals were favored because of their nearly universal appeal, and because of their\ninherently constrained, labor-intensive supply. In this system, early prices likely reflected\nthe worth of a particular good compared to the valuation of the coin as a non-monetary commodity. Over\ntime, the exact \"melt value\" of the coins started to matter less, and the currency functioned as a more\nabstract medium of exchange - but its precious metal content stabilized the economy by ensuring that\nthe coinage had an inherent and lasting value, even if the issuing state simply vanished from the map.\n<\/p>\n<p>\nBy the 19th century, many European countries moved on to a more flexible model where coins were made out\nof cheaper metals, and banknotes were printed on paper or cloth. To encourage the use of these new instruments\nand to establish their value, the governments promised to freely exchange such intrinsically worthless tokens\nfor a predefined amount of gold. In other words, as long as people had faith in their rulers, the fundamental\nmechanics of this new <i>representative currency<\/i> remained roughly the same as before.\n<\/p>\n<p>\nIn theory, both of these systems looked simple and robust. But there was another, somethwat subversive\nforce at play:\nin the 17th century, many European states have witnessed the emergence of fractional-reserve banks.\nThese private ventures operated according to a simple scheme: they accepted people's money for\nsafekeeping, promising to pay a premium on every deposit made. To meet these obligations and to make\na profit, the banks then used the pooled deposits to make high-interest loans to other\nfolks. The financiers figured out that under normal circumstances and when operating at a sufficient\nscale, they needed only a very modest reserve - well under 10% of all deposited money - to be able to\nservice the usual volume and size of withdrawals requested by their customers. The rest could\nbe loaned out.\n<\/p>\n<p>\nThe very curious consequence of fractional-reserve banking was that it pulled new money out of thin air. The funds were\nsimultaneously accounted for in the statements shown to the depositor, evidently available for withdrawal or\ntransfer at any\ntime; and given to third-party borrowers, who could spend them on just about anything. Heck, the borrowers\ncould deposit the proceeds in another bank, creating even more money along the way! Whatever they did,\nthe sum of all funds in the monetary system now appeared much higher than the value of all coins and\nbanknotes issued by the government.\n<\/p>\n<p>\nOf course, no new money was being created in any physical sense: all that banks were doing was\nengaging in a bit of creative accounting - the sort of which would probably land you in jail if you \nattempted it in any other comparably vital field of enterprise. If too many depositors were to\nask for their money back, or if too many loans were to go bad, the banking system would fold.\nFortunes would evaporate in a puff of accounting smoke, and with the disappearance of vast\nquantities of quasi-fictitious (\"broad\") money, the wealth of the entire nation would shrink.\n<\/p>\n<p>\nIn the early 20th century, the world kept witnessing just that; a series of bank runs and economic contractions\nforced the governments around the globe to act. At that stage, outlawing fractional-reserve\nbanking was no longer politically or economically tenable; a simpler alternative was to let\ngo of gold and move to <i>fiat money<\/i> - a currency implemented as an abstract social construct representing\nindebtedness, with no predefined connection to the physical realm. A new breed of economists saw the role of the government\nnot in trying to peg the value of money to an inflexible commodity, but in manipulating its supply\nto smooth out economic hiccups or to stimulate growth. Depending on who you ask today,\ncontemporary monetary policies - especially in the era of bank bailouts and debt-fueled GDP boosting -\nare either a brilliant way to stabilize free markets and promote wealth, or a reckless charade\nthat papers over systemic problems and sets us up for serious trouble in the coming years.\n<\/p>\n<p>\nThat question aside, the obvious peril of fiat money is that in the long haul, its value is determined\nstrictly by people's willingness to accept a piece of paper in exchange for their trouble; that\nwillingness, in turn, is conditioned solely on their belief that the same piece of paper would buy them\nsomething nice a week, a month, or a year from now. It follows that a simple crisis of confidence could make a\ncurrency nearly worthless overnight. A prolonged period of hyperinflation and subsequent austerity in Germany and Austria\nwas one of the precipitating factors that led to World War II. In\nmore recent times, dramatic episodes of hyperinflation plagued the fiat currencies of Israel (1984),\nMexico (1988), Poland (1990), Yugoslavia (1994), Bulgaria (1996), Turkey (2002), Zimbabwe (2009),\nVenezuela (2016), and several other nations around the globe.\n<\/p>\n<p>\nFor the United States, the switch to fiat money came relatively late, in 1971. To stop the\ndollar from plunging like a rock, the Nixon administration employed a clever trick:\nthey ordered the freeze of wages and prices for the 90 days that immediately followed the move.\nPeople went on about their lives and paid the usual for eggs or milk - and by the time\nthe freeze ended, they were accustomed to the idea that the \"new\", free-floating dollar is worth about\nthe same as the old, gold-backed one. A robust economy and favorable geopolitics did the rest, and so far,\nthe American adventure with fiat currency has been rather uneventful - perhaps except for the fact\nthat the price of gold itself skyrocketed from $35 per troy ounce in 1971 to $850 in 1980 (or, from\n$210 to $2,500 in today's dollars).\n<\/p>\n<p>\nWell, one thing did change: now better positioned to freely tamper with the supply of money, the regulators\nin accord with the bankers adopted a policy of creating it at a rate that slightly outstripped the organic\ngrowth in economic activity. They did this to induce a small, steady degree of inflation, believing that doing\nso would discourage people from hoarding cash and force them to reinvest it for the betterment of the society.\nSome critics point out that such a policy functions as a \"backdoor\" tax on savings that happens to align with\nthe regulators' less noble interests; still, either way: in the US and most other developed nations, the\npurchasing power of any money kept under a mattress will drop at a rate of somewhere between 2 to 10% a year.\n<\/p>\n<p>\nAll right. Financial systems are messy. Fiat money and fractional-reserve banking, although\nwildly successful, can fail in interesting and horrific ways. Foreign trade, neglected here but absolutely\nvital for most European countries, adds even more variables to the mix. So, let's talk about what can be done\nto protect your rainy-day funds against some of the most likely or most talked-about risks.\n<\/p>\n<h4>3.2.1. Fiscal challenge #1: Dealing with \"normal\" inflation [<a href=\"#3.2.1\">link<\/a>]<\/h4>\n<p>\nAt a rate of 4%, inflation will halve the purchasing power of your savings in about 17 years; at 6%,\nthe process will take just 11 years and a change. Worse yet, depending on your location and lifestyle choices,\nthe inflation rate you experience can be much higher than the nation-wide government numbers imply. For example,\nthe skyrocketing housing prices in the SF Bay Area have halved the purchasing power of some renters and would-be\nhomeowners in a matter of five years or so.\n<\/p>\n<p>\nIn any case, while it's seldom a prepper-grade emergency, you need to tackle the problem of inflation\nsooner than later - or face huge and unnecessary losses down the road.\n<\/p>\n<p>\nFor our parents, the solution was simple: they had to take their money to a bank. The returns were usually sufficient\nto offset the loss, and since the value of their money already depended on the health of the financial system, they\nweren't facing that much added risk. But today, the trick no longer works: people are skittish about the state of\nthe economy and are trying to play it safe, so banks already have more deposits than they can use - and offer\nnear-zero interest rates across much of the developed world.\n<\/p>\n<p>\nThere are many other ways to get returns on your capital, but most are associated with limited liquidity or\nsignificant outlay costs. One well-known exception are publicly traded companies. Businesses usually go public because\nthey want to expand their operations - say, build a new factory or hire more workers. Instead of getting an\nexpensive loan, they put themselves up for sale, allowing people to purchase and trade fractional ownership in the company.\nThe investors' willingness to pay for this privilege depends on two factors: the intrinsic value of the enterprise\n(its assets, debts, operating profits) and the \"hype premium\" - the faith in the company's long-term prospects and \nthe health of the entire industry. For some companies, the intrinsic value is modest, and the premium is huge;\ntheir shares are usually subject to violent price swings on even seemingly minor macroeconomic news. For other,\nless exciting businesses, the situation may be the opposite.\n<\/p>\n<p>\nThe conundrum of owning stock is that it serves as a hedge against inflation only in an otherwise viable economy.\nAt the first sight of serious economic trouble, the premiums paid on corporate stocks take a nosedive and not recover for\nmonths or years; in a genuine downturn, the intrinsic value of many companies will also shrink. Since a downturn\nis probably the time when you will need your rainy-day money the most, it's important to play it safe. Putting somewhere\naround 30-40% of your emergency stash into the stock market may be a good call. Going all in is a very risky bet,\nsince in an economic crisis, it's not rare to see stock indices plunge 50%.\n<\/p>\n<p>\nThe fundamental rule is to not be greedy: within the scope of this guide, your goal should\nbe to preserve capital, not to take crazy risks. If you are tempted to put your money into Tesla, Twitter, or\nsome penny stock mentioned by your third cousin, you are not thinking straight. Pick about 10-20 boring\ncompanies that seem to be <a href=\"http:\/\/smile.amazon.com\/dp\/0060555661\">valued fairly<\/a>,\nthat are free of crippling debt, and that have robust prospects for the coming years. Stay clear of financial enterprises,\nof highly speculative sectors such as biotech or solar power, and of heavily regulated industries that lack the flexibility to\ndeal with sudden economic shifts (say, airlines). Relatively safe picks can be found in no-frills\ndomains: basic chemicals, staple electronic components, profitable freight railways, mechanical assembly\nmanufacturing, home and office supplies, and so on.\n<\/p>\n<p>\nIt is worth noting that many personal finance experts advise against hand-picking your investments. Instead,\nthey advocate a process known as \"indexing\": buying into an investment vehicle comprising hundreds of stocks, \nstructured to represent the stock market as a whole. The proponents of indexing have a point: most people who try to pick\nindividual winners in the stock market usually fare no better than an index fund. But in the context of prepping, I think\nthis is advice is flawed. To remain calm in tumultuous times, it is important to maintain a firm grasp of the merits of your\ninvestments. One can convincingly reason about the financial condition, the valuation, or the long-term prospects of a paper mill;\nthe same can't be said of an S&amp;P 500 index fund - which, among other things, contains the shares of\nabout a hundred global financial conglomerates.\n<\/p>\n<h4>3.2.2. Fiscal challenge #2: Having no access to your bank account [<a href=\"#3.2.2\">link<\/a>]<\/h4>\n<p>\nSooner or later, you may find yourself unable to access your bank deposits for a couple of days or\nweeks. It could be a matter of IT trouble at your bank, of a lost wallet, or of being a victim of identity\ntheft. Heck, take Greece or Cyprus: when the confidence in the nation's financial institutions is shattered, it's\neasy to get caught up in government-imposed bank closures and withdrawal controls. (Folks in the United\nStates may also recall the forced closure of Washington Mutual in 2008, or several state-level \"bank holidays\"\nimposed to combat bank runs during the savings &amp; loan crisis back in the 80s.)\n<\/p>\n<p>\nFor short-term survival, simple solutions work best: just keep about 2-4 weeks' worth of cash somewhere at hand;\nhave enough money on you to get you back home when traveling, too. Of course, be mindful of the risk of burglary, so if\nyou're keeping the funds at home, pick an unobvious location for the stash; more about that soon.\n<\/p>\n<p>\nAs for the remainder of your money, I suggest splitting it across two largely unrelated financial institutions with\ndifferent risk profiles - say,\na big national bank and a local credit union. As long as the deposits are insured by the government (as they\nnormally are in the US and in Europe, up to a per-account limit), this\napproach greatly increases the availability of your money, and probably doesn't expose you to any substantially new dangers.\nKeeping all your savings outside the banking system is an option, too, but it's not necessarily a smart choice.\nWith fiat currencies, this move does not truly insulate you from that many longer-term risks, but adds the very real\npossibility of losing all your funds to fire or theft.\n<\/p>\n<p>\nPerhaps of note: over the past 5 years, many European governments have moved to\nseverely restrict the use of cash by imposing per-transaction or daily limits as low as $1,000.\nThe stated reasons have to do with money laundering and tax evasion, but such measures also\nmake it harder to store or use physical currency. Be mindful of that caveat.\n<\/p>\n<h4>3.2.3. Fiscal challenge #3: Dodging hyperinflation and \"bail-ins\" [<a href=\"#3.2.3\">link<\/a>]<\/h4>\n<p>\nAsk a financial advisor about the possibility, and they will probably recommend keeping some of your funds\noverseas. But the odds aren't great of correctly picking a currency with more staying power\nthan the one in which you get paid. Historically, the Swiss franc had a reputation for being an exceptionally\nsafe choice, in part because of being the last major currency still quasi-pegged to gold; but Switzerland\nabolished this requirement in a referendum in 2000.\n<\/p>\n<p>\nAnother complication is that even if you make the right call, many governments impose onerous reporting\nrequirements on foreign assets - and especially in times of economic hardship, they treat them with\nsuspicion and contempt. Host countries are also more cavalier about confiscating foreign deposits,\nas evidenced by the Cypriot \"bail-ins\" in 2013. Lastly, the public associates overseas accounts with\ntax evasion and money laundering, so it may be difficult to garner any sympathy for your case when\nthings go wrong.\n<\/p>\n<p>\nOne school of thought popular in the prepper community is to convert some of your savings into commodity\nmetals: copper, tin, silver, platinum, palladium, and the likes. All of them are easy to store, last\nindefinitely, and will certainly hold value far better than a fiat currency in free fall. On the flip side,\nyou may still need to accept substantial loss: an economic collapse will disrupt industrial demand,\ncausing the prices of many such commodities to slump.\n<\/p>\n<p>\nThis brings us to gold: this metal occupies an interesting niche, because its value is driven chiefly\nnot by industrial applications, but by direct consumer demand and by its status as a mainstream financial\ninstrument. In fact, investors and governments alike frequently flock to it in times of economic uncertainty and\nstagnation,\nas they did in the wake of the financial crisis of 2007. Of course, this goes both ways: should the economy\npick up steam, the demand may decrease and the currently elevated prices of gold may fall closer to their\nhistorical, inflation-adjusted average of $800 per troy oz. Still, the metal is an interesting and reliable\nhedge against economic disasters, especially given that it is very easily bought and sold. If you are worried\nabout hyperinflation, you may want to convert some of your savings into this shiny commodity, although I\nwouldn't go over 20-30% or so.\n<\/p>\n<p>\nBecause of its very high value-to-volume ratio, physical gold is stored and moved around very easily, but keeping substantial\namounts at home can be ill-advised; theft is a very real risk, and most insurance policies will not adequately\ncover the loss. Safe deposit boxes at a local bank, available for around $20 a year, are usually a better\nalternative - although they come with some trade-offs; for example, the access to deposit boxes was\nrestricted by the government during the Greek debt crisis in 2015. Non-bank storage services do not have\nthat problem, but cost quite a bit more.\n<\/p>\n<h4>3.2.4. Fiscal challenge #4: Oh no! Zombie apocalypse! [<a href=\"#3.2.4\">link<\/a>]<\/h4>\n<p>\nThe global financial system simply disappears. Mutant hordes roam the earth!\n<\/p>\n<p>\nThe preservation of wealth after a civilization-ending event is a popular topic of idle banter in the\nprepper community. Some folks believe that commodities such as silver and gold would return as the basis of\na primitive post-apocalyptic economy - and if we're wildly speculating, in the longer haul, that seems like\na fair guess. But in the immediate aftermath, it seems more likely that economic activity would be minimal and\nlimited to barter or communal ledgers. Nobody would want to exchange a candy bar for a gold coin if they\ncan't be sure about being able to find any other food.\n<\/p>\n<p>\nThis may sound like a good argument for putting all your money into freeze-dried meals, medicine, shovels,\nand other survival supplies. But of course, that decision would become a huge liability should the\napocalypse not come, or simply not come soon enough: you probably can't pay a roofer or a dentist with a\npallet loaded with ammo, cigarettes, and canned ham.\n<\/p>\n<p>\nHere's my advice: keep the bulk of your savings in cash, stocks, and other assets you can easily liquidate\nor put to use today; even if you genuinely worry about the apocalypse, plan to spend no more than 2-4% of\nyour money on\nessential prepper supplies. Sure, when the zombies come, your financial instruments will almost certainly\nbecome worthless; but you better believe that the value of your survival gear will increase 100-fold.\nZombies or not, your net worth will be safe. Your delicious, tasty brains - well, that's something to\nworry about!\n<\/p>\n<h3>3.3. Prepper commandment #3: Learn new skills [<a href=\"#3.3\">link<\/a>]<\/h3>\n<p>\nIn the 90s, it seemed that you couldn't go wrong by getting into professional journalism,\nopening a video rental store or an arcade, or selling calculators, encyclopedias, disposable\ncameras, answering machines, and audio CDs. We would be very naive to hope that the next twenty\nyears will not bring similarly dramatic disruption to many of the seemingly cozy professions of today.\n<\/p>\n<p>\nSo, here's another unorthodox prescription for building a comprehensive preparedness plan: develop\nuseful and marketable secondary skills. A simple and enjoyable way of doing so is to pick a\nhobby you can get passionate about - and then work hard, be very honest about your own mistakes and\nshortcomings, and try to get better at it every week. You shouldn't\nseek immediate profits, since progressing from a hobby to a paid occupation inevitably takes away some\nof the fun; but try to gravitate toward pursuits that could conceivably morph into viable career\nchoices within a decade or so. If you have a family, help your spouse and children pursue thoughtful\nhobbies of their own, too.\n<\/p>\n<p>\nThe value of such a step extends beyond the mere task of shielding you from glacial shifts in the job market: if\na major disaster suddenly cripples the local economy, there may be no more jobs for insurance claims\nadjusters or account executives, but carpentry or metalworking skills could be in high demand for the coming\nyear or two. You\ncan never predict it exactly, but the more you can do, the better you can cope with\nwhatever adventures come your way.\n<\/p>\n<p>\nWhen it comes to recommendations,\nthere is no short list of hobbies that are objectively better than the rest; the selection is vast,\nand the right choice will inevitably depend on your own interests, natural talents, the space you have\navailable, and on countless other constraints. That said, here are some fairly popular options that may\nbe worth thinking about:\n<\/p>\n<ul class=\"compact\"><li> <p>Woodworking and carpentry,<\/p>\n<\/li><li> <p>Metalworking, knife making, gunsmithing,<\/p>\n<\/li><li> <p>Glassblowing, pottery,<\/p>\n<\/li><li> <p>Sewing, embroidery, leather crafting, toy making,<\/p>\n<\/li><li> <p>Amateur electronics or robotics,<\/p>\n<\/li><li> <p>Computer programming,<\/p>\n<\/li><li> <p>Graphical design,<\/p>\n<\/li><li> <p>Creative writing or online journalism,<\/p>\n<\/li><li> <p>Farming or hunting (in rural regions),<\/p>\n<\/li><li> <p>Community education and local charity work.<\/p>\n<\/li><\/ul><p>\nNot all of these hobbies can be turned into well-paying gigs unless you truly excel at them - but they\nare guaranteed to be challenging, meaningful, and fun. The Internet gives you ample opportunities to\nlearn from others, compare notes, and get feedback on your work - all without prematurely subjecting\nyourself to the pressures of the commercial marketplace.\n<\/p>\n<p>\nOf course, marketable hobbies aside, some of the more determined preppers pursue interests such as\nmartial arts, pro marksmanship, bushcraft, paramilitary combat tactics, and other extreme survival strategies.\nSuch talents can be quite useful in a couple of plausible if unlikely scenarios - but their major\ndisadvantage is that during a simple economic downturn, they won't put bread on the table or pay your bills.\nIt's not unwise to give them some consideration, but be sure to balance it with more pragmatic skills.\n<\/p>\n<h3>3.4. Prepper commandment #4: Don't hurt yourself [<a href=\"#3.4\">link<\/a>]<\/h3>\n<p>\nYup. Unintentional injury may seem like a topic unbecoming a true prepper, but it will be hard to live out a\npost-apocalyptic Mad Max fantasy with a bum leg or a broken neck; and more prosaically, serious prior injury\nmay limit your ability to provide for yourself and your family, confront a robber, or get out of a burning\nhome. It may seem like a far-fetched worry, but the lifetime probability of suffering serious harm is much\ngreater than we intuitively suspect.\n<\/p>\n<p>\nNow, there are some dangers to life and limb that we simply can't predict or prevent: the occasional falling piano,\nthe murderous roommate, the untimely stroke. Then there are the risks we take willingly, accepting\nthe inherent and unavoidable trade-offs of our hobbies or jobs: the possibility of being snatched by a giant squid\nwhile snorkeling off the coast of California, or the near-certainty of lung fibrosis from toiling in a \nsugar mine. These are the things we can't or don't want to give up - and that's perfectly fine.\n<\/p>\n<p>\nBut there are also the \"hold my beer\" moments: the unnecessary displays of bravado, overconfidence, or\nthoughtlessness. We recognize them from cringeworthy yet hilarious Youtube clips of people getting hurt -\nbut we are guilty of the same: we get honked at for carelessly changing lanes, we end up climbing more than\na fair share of rickety ladders and office chairs, and every now and then, we all get a bit cavalier with\nlawnmowers, escalators, ATVs, or other power tools.\n<\/p>\n<p>\nWe keep getting away with all that goofiness, and that only serves to make us more certain that our own\ntransgressions carry no serious risks. But government statistics tell a strikingly different tale: in the US\nalone, unintentional injuries result in 40 million ER visits and 100,000 deaths every year. Heck, accidental\ninjury is the leading cause of death for people between the ages of 1 and 45 - far ahead of cancer, heart disease,\ngun violence, and other pop culture bogeymen. And the injuries themselves are very prosaic, too: all you see\nare falls, cuts, burns, vehicular collisions, poisonings, and so forth. The \"idiots\" we sometimes watch on\nYoutube are us. They just happen to have been caught on camera on the day their luck ran out, perhaps aided by\ndowning a couple of beers.\n<\/p>\n<p>\nIn the end, ladders, cars, and space heaters are a much greater threat to your well-being than a gun-totting\nrobber or an army of zombie marauders could ever be. So, gleaned from accident statistics, here are some of\nthe familiar-sounding but crucial survival tips. It may sound unlikely, but if something appears on this\nlist, it's responsible for quite a few gruesome deaths or injuries every year; take it to heart.\n<\/p>\n<h4>3.4.1. Safety tip #1: Don't be a moron when working at heights [<a href=\"#3.4.1\">link<\/a>]<\/h4>\n<p>\nSure, the lightbulb needs changing and that office chair is really close, but it would take you only 15 seconds more\nto bring a more sturdy stool from another room, so don't take chances if you don't have to. Similarly,\nhaving someone hold a wobbly ladder for you or securing it with some rope can be a minor hurdle - but it's gonna be\nmuch less of a hurdle than dealing with a compound fracture or a dent in your skull.\n<\/p>\n<p>\nWhatever you do, keep in mind that you\nprobably experienced most of your falls as a kid, and that can profoundly skew your perception of danger: falling five feet\nand landing on your side is different when you weigh 50 lbs, and different when you are 180 lbs.\n<\/p>\n<h4>3.4.2. Safety tip #2: Drive defensively and stay calm on the road [<a href=\"#3.4.2\">link<\/a>]<\/h4>\n<p>\nGot a car? Don't go too fast, keep a three-second distance to the vehicle in front of you, and always scan\nfor cross traffic when approaching intersections or making turns: other drivers may be less attentive than\nyou. Be very careful when changing lanes, do it slowly, and be sure to adjust your mirrors to eliminate blind spots\n(you don't really need to see the sides of your car). Slow down for cars stopped in other lanes -\nthey may be letting a pedestrian through. Wear seat belts, keep children in fitting car seats,\nget some rest on longer trips, and don't talk on the phone - it doesn't matter that it's hands-free.\nAvoid frequent rides with people who drive badly, too.\n<\/p>\n<p>\nInterestingly, driving and safeguarding your finances have something in common: when you end up hurting\nanother person in an at-fault accident, they may go after your savings or real estate to recoup medical\nexpenses, lost wages, and other costs. So, in addition to driving defensively, have a look at your insurance\npolicy. The minimum liability coverage mandated by the state can be as low as $15,000; relatively few victims\nwill settle with the insurer for that amount if they think that taking you to court could net them ten times as much.\nBumping your limit to $250,000 is usually pretty cheap. In fact, if you have collision coverage,\nyou can more than make up for it by increasing your deductible to $1,000.\n<\/p>\n<p>\nBiking on public roads? Wear a helmet and bright-colored clothing, stay well clear of the doors of parked\nvehicles, move in a straight line instead of weaning in and out of the traffic,\nand watch for cars trying to make right turns. Scan for cross traffic at intersections \nand signal all turns. Don't run red lights - it's not worth the seconds shaved off your commute. Don't ride\nwithout holding the handlebars; it makes you look very hip, but when you do that, a single rock or an\nunnoticed pothole can throw you right under the wheels of a passing bus.\n<\/p>\n<h4>3.4.3. Safety tip #3: Show respect to dangerous machines and chemicals [<a href=\"#3.4.3\">link<\/a>]<\/h4>\n<p>\nRead up on the safety rules pertaining to your weekend hobbies and to your daytime job; ask others about their\nhorror stories, too. Be very careful around select power tools, such as chainsaws, table saws,\nangle grinders, lathes, and nail guns - they have quite a penchant for taking fingers off and eyes out. (Table\nsaws in particular are responsible for about 30,000 serious injuries annually, or about 40% of all workshop\naccidents.)\n<\/p><p>\nIn general, don't horse around any heavy machinery; if it's weighs more than you, it can kill you \nin the blink of an eye. This is also important for children: teach them\nnot to play behind or under parked cars. And if you have a toddler, use straps to secure rickety dressers and\nother tall, heavy furniture.\n<\/p>\n<p>\nAt home, be very alert around deep fryers and pots of boiling water. Wear eye protection when working with\ndrain cleaners, bleach, and other caustic substances. Learn about the overdose risks of paracetamol\n(ibuprofen is a much safer pick) and take a critical look at your prescription drugs. If you're ever doing DIY\nelectrical work, learn how to do it properly, and get a non-contact voltage probe to double-check\nfor live wires before you touch anything.\n<\/p>\n<h4>3.4.4. Safety tip #4: Don't die in a fire [<a href=\"#3.4.4\">link<\/a>]<\/h4>\n<p>\nInstall smoke detectors and keep them operational; if they are going off too often, try to move them\nfarther away from the kitchen, or switch to photoelectric sensors, which are much less sensitive to minor\ncooking mishaps. There are many wired and battery-operated units to choose from.\n<\/p>\n<p>\nGet at least two good-sized ABC fire extinguishers (5-10 lbs or so) and keep one in\nyour bedroom. Learn how to \n<a href=\"https:\/\/www.youtube.com\/watch?v=kz2SKZgK9Kk\">deal with oil fires<\/a>, don't stockpile excess\nflammable materials, and be very careful when\n<a href=\"https:\/\/www.youtube.com\/watch?v=Joj7pzmz7yk\">pouring flammable liquids<\/a> near open flames.\nBe smart about where you put space heaters and candles, and don't smoke in bed. \n<a href=\"https:\/\/www.youtube.com\/watch?v=xr6b9b8FYKk\">Water your Christmas tree<\/a> and use LED tree lights.\nUnplug devices with <a href=\"https:\/\/www.youtube.com\/watch?v=ixIOEPnsgbI\">lithium-polymer batteries<\/a>\nwhen leaving home. Don't put grills next to siding-clad walls and don't overload extension cords.\nWhen cooking, stay in the kitchen or set a timer to remind yourself that the oven is on.\n<\/p>\n<p>\nFolks in rural areas who live far away from any fire stations, but who have a pond, a creek, or a pool\non their property, may want to look at a\n<a href=\"https:\/\/www.amazon.com\/Pressure-HPFC-2065HR-Fire-Fighting-Cart\/dp\/B00K2T9RAS\/\">fire pump kit<\/a>.\nIt's a fairly expensive tool ($1,000 or so), but it can rapidly put out even a large blaze - provided that\nyou have a pond, a pool, or a cistern nearby.\n<\/p>\n<h4>3.4.5. Safety tip #5: Just in case, keep your senses razor-sharp [<a href=\"#3.4.5\">link<\/a>]<\/h4>\n<p>\nIf you're regularly drinking or doing psychoactive drugs, try to kick the habit - or at least scale it back.\nI'm not going to give you a talk about the evils of cannabis, but by the end of the day,\nimpaired judgment is impaired judgment. It inherently increases the odds of waltzing into trouble or\ngetting hurt.\n<\/p>\n<h3>3.5. Prepper commandment #5: Don't become a victim [<a href=\"#3.5\">link<\/a>]<\/h3>\n<p>\nIn addition to the dangers of poor financial planning and the ever-present specter of\nunintentional injury, another threat we should reckon with is becoming a victim of a crime.\nAlthough the risk is not as pervasive as the challenges discussed earlier in this chapter,\nit still earns a distinction as one of the things that many readers will almost certainly have to face at\nsome point in their lives.\n<\/p>\n<p>\nAsk a hardcore prepper for advice on this matter, and they will probably tell you to start practicing hand-to-hand\ncombat, get a knife, or carry a gun. But in reality, we need a more nuanced and proportional approach to threats,\nand one that emphasizes avoidance and de-escalation, rather than the ability to resolve each and every\nconflict with a single well-placed shot. Sure, a self-defense weapon can save your butt in some life-or-death\nsituations, but these are comparatively rare; such a tool won't deter a pickpocketer, won't stop a\nburglar from ransacking your place while you are at work, and won't prevent a hacker from emptying your bank\naccount while you're busy watching the reruns of <i>Gilmore Girls<\/i>.\n<\/p>\n<p>\nIn other words, while this guide certainly doesn't have an anti-self-defense or an anti-firearm\nslant, we'll try to take a broader view before delving into the comparative merits of\nkarate chops, blade weapons, firearms, stun guns, and pepper spray.\n<\/p>\n<h4>3.5.1. Defense tip #1: Practice situational awareness when on foot [<a href=\"#3.5.1\">link<\/a>]<\/h4>\n<p>\nWhen walking around town, keep scanning your surroundings and be mindful of people around you. If your spidey\nsenses are tingling, just bail: make a sharp turn and sprint away. Don't worry that some random dude\nclosing up on you at night might be perplexed or offended by your move.\nDon't freeze up if somebody hollers or authoritatively barks an order at you - complying is a very\npowerful instinct, and some assailants know to exploit it.\n<\/p>\n<p>\nOf course, some street-savvy readers may consider it to be in poor form to sprint away from a threat, and may be\ninclined to confront the danger and see where that takes them. It's an OK choice if you are well-armed or\nphysically fit, but is certainly not a way to minimize the probability of harm.\n<\/p>\n<h4>3.5.2. Defense tip #2: Make yourself pickpocket-proof [<a href=\"#3.5.2\">link<\/a>]<\/h4>\n<p>\nIn shopping malls, on mass transit, and in other crowded settings, don't carry your most precious\nvaluables in front or back pockets; a purse is also a clear no-no. Inner pockets of jackets, and breast- or\nknee-level pockets of pants and shirts, are much harder to muck with. Discreet, slim waist\npacks or under-the-garments neck wallets work even better. Emphasize to your friends that you are\nwearing such fashion accessories only ironically; who knows, maybe you will start a trend.\n<\/p>\n<h4>3.5.3. Defense tip #3: Protect your valuables at home [<a href=\"#3.5.3\">link<\/a>]<\/h4>\n<p>\nMost break-ins are purely opportunistic: thieves are in and out within five minutes, quickly rummaging\nthrough all the places where people usually keep valuable stuff. You can bank on them going through\nevery nook and cranny of your bedroom, looking under the mattress, peeking into every drawer -\nand grabbing everything that looks shiny and is easy to lift. Their usual targets include phones,\ncameras, tablets, laptops, jewelry, firearms, loose cash, checkbooks, credit cards, and prescription\nmeds. Vital documents that may be useful for identity theft or benefits fraud, such as drivers\nlicenses, passports, and SSN cards, are also a fair game.\n<\/p>\n<p>\nBreak-ins are difficult to prevent, especially in suburban single-family homes with secluded backyards and\nstreet-level windows and doors; tall fences and window bars can work, but they are expensive and\ntend to draw the ire of your neighbors. The most cost-effective solution may be to keep your windows\nand doors closed when away, but beyond that, just optimize for hassle-free\noutcomes. You can leave some less important goodies in plain sight - say, some cheap jewelry, a modest\namount of cash, and a beat-up phone - and put all the real valuables in a much less obvious or less\naccessible spot.\nA heavy safe will usually do;\n<a href=\"http:\/\/smile.amazon.com\/s\/ref=nb_sb_noss?url=search-alias%3Daps&amp;field-keywords=diversion+safe\">diversion\nsafes<\/a> are pretty cool, too, if you trust yourself not to accidentally throw them away.\n<\/p>\n<p>\nWe'll talk about home security equipment later on - but in general, devices such as alarm systems, cameras,\nor sophisticated locks play a lesser role in keeping your belongings safe; a well-trained dog can work wonders,\nbut especially in an urban or a suburban setting, such a pet is a substantial commitment (and can turn out to be\na lazy bum).\n<\/p>\n<p>\nIf there's another powerful and low-cost\nburglary prevention tool at your disposal, it's being careful not to attract targeted theft. Be mindful\nof who you invite into your home, who handles your keys, and how much you signal about your financial status to\nyour more distant family, random acquaintances, or strangers parked across the street. If you have children, teach\nthem about the virtues of modesty, too. If you tell them that the thieves will probably take their game console,\nthey may even listen to what you have to say.\n<\/p>\n<p>\nAdvertising your wealth aside,\nanother sure way to invite burglars is to make it seem that your house is unoccupied: packages piling up in front,\nan overflowing mailbox, an empty driveway, all lights turned off at night. Asking a neighbor to park a car \nin your driveway, putting some lights on a \n<a href=\"http:\/\/smile.amazon.com\/dp\/B005WQIDHY\">multi-cycle timer<\/a> ($8), and having someone pick up your mail,\nare just several examples of low-cost solutions that are worth trying out whenever going on a longer trip.\n<\/p>\n<h4>3.5.4. Defense tip #4: Plan for dangerous encounters [<a href=\"#3.5.4\">link<\/a>]<\/h4>\n<p>\nSometimes, you can't avoid a confrontation. Figure out how to react if somebody asks you to get into a car,\ndemands cash, or barges into your occupied home. It would be a complete shock for you, but they\nhave probably done it before - so you gotta rehearse if you want to have the upper hand. Even if they have a \nweapon trained on you, it's a game of confidence and wits, not just physical force.\n<\/p>\n<p>\nFor muggings, keeping several $10 or $20 bills in your front pocket (and having real valuables somewhere else)\ncan be enough to send them on their way; in busy locations, you should also be able to just\nignore the mugger and briskly walk away. For more serious incidents, it may be useful to respond with something\nthat is non-threatening but catches the assailant off guard. Simply feigning a panic attack or initiating\na startling conversation (<i>\"hey, are you a friend of CJ? Pretty sure we've met last year!\"<\/i>)\ncan throw them off balance - allowing you to fight back or get away. Of course, you also need a plan for\nthat next step; that's where your running skills, your bare-hands self-defense talents, or \nyour weapons proficiency can come into play. But again, you need to actively practice and develop approaches\nthat have a chance of working in real life; there's no verbal diversion strategy in the world that would\ngive you enough time to fumble through your purse to find an old, gummed up can of pepper spray.\n<\/p>\n<p>\nFor home intrusions, you should try to act out various scenarios: say, confronting the bad guy or trying\nto escape. Take into account that break-ins can happen at different times of day or night. Remember that\nintruders may have varied intents; some folks may be homicidal or high, but most will just\nwant your laptop and will be very worried about getting hurt. Many robbers work in pairs, too.\n<\/p>\n<p>\nWhether you like it or not, you may eventually have to defend yourself, so be\nsure to understand the law. You shouldn't take such advice from random people on the Internet, but as far\nas I can tell, in much of the US and in many other western countries, you have no duty to run away from an attacker\nand can use deadly force if you have a reasonable and immediate reason to fear for your life or the\nlives of others. But there are exceptions; for example, despite recent reforms, a duty to retreat\nexists in some form in several northeastern states and in some corners of the Midwest. There are also\ndifferences in how seemingly similar self-defense statutes get interpreted by the police, by prosecutors,\nand by courts in different parts of the world.\n<\/p>\n<h4>3.5.5. Defense tip #5: Don't be an easy target online [<a href=\"#3.5.5\">link<\/a>]<\/h4>\n<p>\nYour cyber-life matters - or at the very least, your bank account password and your credit\ncard numbers do. To avoid falling prey to hackers, \nkeep your software up-to-date, choose decent and unique passwords for all\nimportant websites, don't install sketchy freebies, and don't fall for legit-looking but\nunexpected messages and prompts. When in doubt, just leave the site you were on, do some web searches\nto understand what's going on, and maybe try again some time later.\n<\/p>\n<p>\nBeware of financial scams. If you get an IM from a friend on an\nunannounced overseas trip urgently asking for a loan, call them or check in with their family\nfirst. And no, there's no inheritance waiting for you in Nigeria, your chatroom bride-to-be in\nUkraine is not short on cash, and you are not the 1,000,000th visitor to <i>timecube.com<\/i>. \n<\/p>\n<p>\nOh, one more thing: don't log into your bank, e-mail, or any other sensitive services from\nother people's computers; if you absolutely have to, change your password as soon as you get\nhome. It's not that your acquitances are evil, but they may be more sloppy than you when it\ncomes to keeping their devices safe.\n<\/p>\n<h4>3.5.6. Defense tip #6: Don't make enemies [<a href=\"#3.5.6\">link<\/a>]<\/h4>\n<p>\nIf you're doing something that's morally reprehensible or socially unwelcome, you\nare greatly increasing the odds of getting hurt. It doesn't matter if you think it's perfectly legal:\nif you are a monumental and malicious jerk, a bored prosecutor will probably dream up a \n<a href=\"https:\/\/smile.amazon.com\/dp\/1594035229\">felony\ncharge<\/a> to hit you with. Or perhaps they won't, but one of the people you wronged will lose it and\ntake justice into their own hands. In other words, if you want to escape harm, don't mess with\nothers out of malice, jealousy, boredom, or for petty personal gain.\n<\/p>\n<h3>3.6. Prepper commandment #6: Get in shape [<a href=\"#3.6\">link<\/a>]<\/h3>\n<p>\nIn the United States, about one in three adults is obese - that is, are overweight to the\npoint where the condition likely interferes with their health or their daily lives. And while many\nfolks in the prepper community tend to grossly overstate the importance of tip-top physical fitness,\nthere is no denying that obesity is a very real foe. For example, among low-BMI individuals, the\nincidence of diabetes hovers around 1-4%, but the same number skyrockets to 50-80% for obese folks.\nMany other, serious metabolic and cardiovascular diseases follow the same curve - and can make it\nvery difficult for the affected families to cope even with fairly prosaic and short-lived emergencies.\n<\/p>\n<p>\nIn all likelihood, if you are obese or slowly getting there, you know quite well that losing some weight is not really the hard part:\nif you were to stop eating for a week, you would likely shed 5-10 pounds or more. But it would\nbe a miserable experience, and one almost guaranteed to be followed by an even faster rebound. So, the real\nchallenge of weight management is coming up with a long-term strategy that does not amount to torture -\nand does not leave you constantly craving for familiar foods.\n<\/p>\n<p>\nAlas, most of the popular diets make this task awfully hard: they force their followers to abandon a\nlifetime of dietary habits, taste preferences, and eating schedules - and stuff themselves full of kale, turnips,\nquinoa, acai berry, or whatever else happens to be this week's \"fat-fighting superfood\". To add insult to injury,\nmost of the nutrition fads are not actually backed by real, reproducible science; suffice to say that in the 70s, \n<a href=\"https:\/\/cdn.theatlantic.com\/assets\/media\/img\/posts\/2015\/06\/sugar_221\/6a2d1191c.jpg\">table sugar<\/a>\nwas widely touted as a dieting aid. Even today, weight loss advice tends to revolve around robustly debunked\nconcepts - say, the existence negative calorie foods, the alleged superiority of low-carb but high-fat diets, the evils of\nHFCS and aspartame, or the significance of eating meals on a particular schedule thoroughout the day.\n<\/p>\n<p>\nJust as importantly, our innate nutritional instincts can be badly misguided, too: for example,\ncontrary to common wisdom, bananas are not really healthier than potatoes, and the bulk nutritional qualities\nof a glass of apple juice are pretty close to those of a can of Sprite. Heck, good ol' butter has fewer calories than\nolive or coconut oil, so a \"healthy\" bruschetta is not far off from a less-reputable southern delicacy:\ndeep-fried butter on a stick. It gets better: a supposedly nutritious\nburrito from Chipotle easily packs four times as many calories as a greasy burger from McDonald's, while a\nloaded coffee at Starbucks is about the same as downing two hot dogs with a heaping\nside of mashed potatoes to boot. The end result is a\ntruly abysmal track record for most weight loss regimes; the long-term success rate for people who try to slim down\nis estimated to be somewhere between 5 and 20%.\n<\/p>\n<p>\nMy advice is that if you want to lose weight, it's best to stay away from celebrity-endorsed diets,\ndubious nutrition claims, and rigid, unworkable plans that seek to control your every urge. While every\nsituation is different, here are several ways to eat less while still staying happy:\n<\/p>\n<ul><li> \n<p>\nStart your breakfast with\n<a href=\"http:\/\/smile.amazon.com\/dp\/B008OLCOF6\">high-quality protein powder<\/a> and \n<a href=\"http:\/\/smile.amazon.com\/dp\/B0040IWMTA\">insoluble dietary fibers<\/a> (about\n20-30 grams each; you can mix them together in a cup of cold water). There is fairly clear evidence\nthat fiber and protein can increase satiety and reduce cravings thoroughout the day. Sure, it's not \"natural\",\nbut it beats making implausible resolutions to organize your life around low-calorie, fiber-rich meals - especially\nif you don't like veggies or don't usually cook your food.\n<\/p>\n<\/li><li> \n<p>\nBeyond this, stick to your favorite foods and don't feel pressured to skip regular meals - but cut <b>all<\/b>\nportions in half, even if it means throwing a half-eaten burger out. Don't go back for seconds, too.\nIt will feel wrong the first couple of times, but it's surprisingly easy to do. That's because portion control\nis almost completely psychological; your blood hormone and nutrient levels go up only some time after you cleaned your\nplate. Eating more slowly can make this step a lot easier, too.\n<\/p>\n<\/li><li> \n<p>\nFor habitual snacking in front of a computer or a TV, see if you can substantially reduce calories while still\nsticking to satiating and tasty treats. This can be easier than it sounds: say, helping yourself to a nice\nserving of <a href=\"http:\/\/smile.amazon.com\/dp\/B01HZ7L4EG\">salted popcorn<\/a> (110 kcal), preparing a cup of\n<a href=\"http:\/\/smile.amazon.com\/dp\/B000RPYXZ6\/\">buttery mashed potatoes<\/a> (110 kcal), grabbing some\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00MFCEB2E\/\">quick oatmeal<\/a> (130 kcal), \nor sipping some hot <a href=\"http:\/\/smile.amazon.com\/dp\/B000EM8UJC\">instant\nchicken soup<\/a> (50-80 kcal), is an excellent alternative to Cheetos, M&amp;Ms, or even supposedly healthy peanuts\n(easily 600-800 kcal). If you enjoy pickles or raw sauerkraut, they are extremely low-calorie, so have as\nmuch as you want; in the same vein, carrots are a pretty guilt-free choice. Chewing gum can keep you occupied\nbetween meals, and if you are downing multiple cans of sugary drinks a day, artificially-sweetened sodas offer a\n<a href=\"http:\/\/examine.com\/faq\/is-diet-soda-bad-for-you\/\">good alternative<\/a>.\n<\/p>\n<\/li><li>\n<p>\nWhen buying food, don't fall for \"diet\", \"reduced fat\", \"low sugar\", or \"low carbs\" ice cream, yogurts, cakes, pizza,\npasta, and so on - the differences are so minor that you might as well have the real thing and stop\nfooling yourself. Watch out for deceptive portion sizes, too. For example, Cheetos are labeled as 150 kcal\nper \"serving\", but there are almost 10 servings in a regular bag! Frozen fries are another great example:\nthey look pretty low-cal until you realize that a serving is just 10-15 pieces or so - certainly not enough\nto make you feel full.\n<\/p>\n<\/li><li>\n<p>\nWhile dieting, try to drink a bit more and start taking\n<a href=\"http:\/\/smile.amazon.com\/dp\/B003G4BP5G\">OTC multivitamin supplements<\/a>; they don't offer\nclear-cut benefits under normal circumstances, but they can compensate for mineral and vitamin\ndeficiencies if you never had a particularly thoughtful diet, and are about to start eating less than you\nwere accustomed to.\n<\/p>\n<\/li><li>\n<p>\nDon't hit the gym. Hold off with intense workouts at least until you are close to your target weight.\nDaily exercise schedules are hard to keep for more than a couple weeks, especially if you lead a busy\nlife; on top of that, a drastic increase in physical activity can trigger cravings or upend your\nnutritional needs. If you are itching to burn some extra fat, incorporate less punishing activities\ninto your daily routine - say, walking or leisurely biking to work.\n<\/p>\n<\/li><li> \n<p>Be in this for the long haul. Effortlessly losing 1-2 lbs per week while slowly developing better\nhabits is far more meaningful than starving yourself for a month to get immediate but short-lived\nresults. Get an accurate bathroom scale, take daily measurements first thing in the morning,\ncalculate key milestones, and put it all in a spreadsheet to keep yourself honest and motivated.\nIt will probably take 6-9 months to get the outcome you want; daily or weekly weight\nfluctuations are almost completely meaningless, but you should be seeing\na consistent and predictable biweekly drop.\n<\/p>\n<\/li><\/ul><p>\nI can't promise that this approach will work for everyone, but after trying countless other methods,\nI found that regime quite enjoyable, effective, and worry-free; at the ripe old age of 35 and with three kids in tow,\nI ended up losing over 65 pounds - going down from a BMI of 29 to 20 or so.\n<\/p>\n<p>\nOf course, if you are very obese or have any serious health conditions, such as diabetes or CVD, talk\nto a doctor first. In such situations, aggressive dieting can carry additional risks and calls for\nsome monitoring along the way; a routine blood test or an ECG shouldn't cost much. And it goes without\nsaying that if your diet makes you feel listless or sick, it's definitely time to stop right away!\n<\/p>\n<p>\nNow, when it comes to fitness per se,\nI firmly believe that there is no need to go overboard; good health is far more important\nthan Rambo skills. While getting buff may be a fun pastime for some young folks, there are\nvery few emergencies that would force you to run 30 miles or climb a 20 foot wall. Being able to\nwalk or bike for several hours is likely good enough to deal with all practical scenarios we talked\nabout thus far.\n<\/p>\n<h3>3.7. Prepper commandment #7: Make friends with neighbors [<a href=\"#3.7\">link<\/a>]<\/h3>\n<p>\nThe funny thing about disasters is that they seldom unfold precisely as planned. Perhaps you will lose a job\nand get robbed the same week. Perhaps in the middle of a prolonged outage, you will find out that some of\nyour emergency supplies have been misplaced, damaged, or spoiled. Maybe your plan to walk a mile to get\ndrinking water from a river will get foiled by a broken limb. And maybe a brilliant home defense\nstrategy will prove worthless when standing eye to eye with an angry mob of rioters armed with rocks.\n<\/p>\n<p>\nIn trying times, people always come together and find strength in local communities. Even if you\ndon't expect it, you will almost certainly be able to count on the kindness of strangers. But your\nodds can be greatly improved by getting to know your neighbors ahead of the time, by cultivating\ntrust and mutual respect, and by getting a sense of each others' toolkits and skills. In a grim\nsituation, being on good terms with a doctor or a veterinarian can quite literally save your life.\nAnd heck, some rural communities in the US even maintain communal stashes of emergency supplies!\n<\/p>\n<p>\nOf course, it goes both ways: you will almost certainly find it harder to get help if your neighbors still resent\nyou for puking on their doormat and constantly partying at night; so once again, not being a\njerk to other people is not just good manners, but a very legitimate survival skill. Even if it's not\nreally your nature, say \"hi\", engage in small talk, and offer to help with minor hurdles every now and\nthen. Bring your neighbors a pie or some donuts, add them on Facebook, and try to find common interests.\nSocialize with coworkers who live nearby, too. Even if the zombie apocalypse never comes, it\nstill won't be a waste of your time.\n<\/p>\n<p>\nNow, many \"true\" preppers would tell you to keep mum about your plans, so that in an emergency, you don't have\nto fend off armies of freeloaders begging for a slice of your meager supplies - or worse yet, trying to take them by\nforce. I think that this attitude is short-sighted; sure, it makes sense not to broadcast your plans\nto the entire world, and there is no conceivable benefit to posting Facebook selfies with your stash\nof freeze-dried food or with a pile of cash. But the clear value of convincing some of your friends to\nstart prepping greatly outweighs the distant possibility\nthat one of them will attempt to raid your home the moment the power goes out.\n<\/p>\n<h3>3.8. Prepper commandment #8: Write down a response plan [<a href=\"#3.8\">link<\/a>]<\/h3>\n<p>\nTo cope with a true emergency, it's not enough to know the risks and sit on a pile of overpriced\nsurvival gear: you need to plan ahead. If your house is on fire, there may be no time to rifle through\nfolders to gather all your vital documents; if the floodwaters are rising or a chemical tanker overturns\non a nearby highway, it may a bit late to start thinking about refueling your car. And if you're\nstranded on a rural road in a broken-down vehicle, you may sorely regret not putting any drinking\nwater in the trunk.\n<\/p>\n<p>\nIn fact, even in situations that don't unfold in such a dramatic way, sketching out a plan can help you\noptimize spending and spot potential problems early on. For example, there may be little merit to stockpiling\n50 lbs of rice if you can't possibly store enough water to cook it all. But then, a quick look at the map\nmay reveal that there is a freshwater reservoir within a\nbiking distance of your home. Great - maybe all you need is a bike basket and \na pair of spare inner tubes.\n<\/p>\n<p>\nSo, here's your homework: make a list of all the major threats discussed in this guide, alongside with\nany other contingencies you worry about. For each and every one of them, draft a detailed, step-by-step\npreparedness and response plan that sounds right to <b>you<\/b>, and that you can see yourself\nactually putting into action. Don't obsess over getting all the details right;\nwe'll try to refine your choices in section 4, and even then, it may take several iterations to\nsettle on an approach you are really happy with. For now, simply list all the noteworthy dangers,\njot down some initial answers to the following questions, and see where that takes you.\n<\/p>\n<h4>3.8.1. Threat assessment topic #1: Is it a priority? [<a href=\"#3.8.1\">link<\/a>]<\/h4>\n<p>\nDo some research to make sure you are not wasting your time on implausible risks. How likely is it that you\nwould have to face this particular danger, and how much damage can it conceivably cause? For example, do\nyou live in a 100-year flood zone? In the path of tropical storms? In a high-crime neighborhood?\nBe sure to search around and study publicly available resources; reaching out to local emergency response\norganizations can be a good plan, too. Try to focus on reputable sources; the science in doomsday movies and\non conspiracy websites seldom checks out.\n<\/p>\n<p>\n(If you decide that you don't care about a particular risk, skip the remaining questions and just move on to\nthe next problem on your list.)\n<\/p>\n<h4>3.8.2. Threat assessment topic #2: What do you need to be ready? [<a href=\"#3.8.2\">link<\/a>]<\/h4>\n<p>\nDo you need to secure any supplies or make other arrangements to prepare for this scenario? If you\nneed to stockpile items, do you have enough room?\nHow long would these supplies last in an emergency, and how often would you need to replace them in\nstorage? Are there any additional steps that you want to take to be in a better place a month,\na year, or five years from now? Make a detailed list and tally the costs.\n<\/p>\n<p>\nBe critical of any assumptions you are making in your plan: sure, you have seen it all in the movies,\nbut search for statistics or historical accounts to better understand how such events typically unfold\nin real life. Don't assume that all contingencies are covered by a generic home insurance policy, too: for\nexample, earthquake and flood coverage is often sold separately (and costs a lot); and in any case, even\nif the insurance will eventually pay, you still need a short-term survival strategy to deal with the loss.\n<\/p>\n<p>\nWhen it comes to dangers such as break-ins, fires, or earthquakes, be sure to walk around the house and\ntake note of anything that unnecessarily exacerbates the risk. Perhaps throwing out old junk,\nreorganizing the contents of kitchen cabinets, adding \n<a href=\"http:\/\/smile.amazon.com\/dp\/B0142G91U2\">earthquake latches<\/a>,\nor fixing a broken lock would be a better use of your time than ordering space-age\nprepper gadgets from Amazon.\n<\/p>\n<p>\nFor natural disasters common in your region, you can find a lot of relevant (if US-specific) advice on the\nwebpages of <a href=\"http:\/\/emergency.cdc.gov\/disasters\/\">CDC<\/a>,\n<a href=\"http:\/\/www.fema.gov\/plan-prepare\">FEMA<\/a>, and the\n<a href=\"http:\/\/www.redcross.org\/prepare\">American Red Cross<\/a>. You can also have a look at the excellent,\ndetailed hazard maps available from FEMA, USGS, NOAA, and several other agencies.\n<\/p>\n<h4>3.8.3. Threat assessment topic #3: How, exactly, will you respond? [<a href=\"#3.8.3\">link<\/a>]<\/h4>\n<p>\nIf this particular calamity strikes, what's the step-by-step playbook you would follow, and how\nmuch time would you have to carry out the plan? Will you be sheltering at home, or is this a situation\nwhere you need to evacuate? If the answer is \"it depends\", what would be the key factors,\nand how quickly would you need to make the final call? If you pick the wrong option, how bad would\nit be?\n<\/p>\n<p>\nBugging out puts you at a tremendous disadvantage, so it's almost always better to dig in; but if\nyou have to leave, also ask yourself this:\n<\/p>\n<ul><li><p><b>...where would I go?<\/b> Is the location sufficiently far away to be unaffected by the\n  event you're running away from? If you are driving, will you have enough gas to get there?\n  (Again, would it help to have a habit of keeping the tank at least half full?) Are the roads\n  likely to be congested or blocked? How many people will think of exactly the same location?\n  What if the destination proves to be inaccessible, overcrowded, or unsafe?\n<\/p>\n<\/li><li><p><b>...what do I need to bring?<\/b> How much can you realistically take with you\n  when leaving by car, by bike, or on foot? What are the most important items, and will\n  you be able to grab them quickly enough? To simplify things, would it make sense to maintain a\n  small cache of supplies in the trunk of your car or at a friend's place - and if yes, what\n  should be in that box?\n<\/p>\n<\/li><\/ul>\n<h4>3.8.4. Threat assessment topic #4: What if you are somewhere unexpected? [<a href=\"#3.8.4\">link<\/a>]<\/h4>\n<p>\nIf you have a daytime job or go to school, there is a good chance that the event you are\npreparing for could unfold while you are at that location - or somewhere on the way. Do you need a\nspecial plan to handle this possibility? Can you get back home easily? Do you need water,\nfood, money, or clothes to make that trip? How likely is it that you would be hurt\nor stranded somewhere?\n<\/p>\n<p>\nSeparately from this, consider the risks of less frequent but more dangerous trips.\nIf you're sightseeing a particularly inhospitable or remote destination, what do you\nneed to survive if, say, you get lost or your car breaks down?\n<\/p>\n<h4>3.8.5. Threat assessment topic #5: Do you need to talk to your family? [<a href=\"#3.8.5\">link<\/a>]<\/h4>\n<p>\nIf you have a spouse, walk them through your plans and make sure they can access the\nessential supplies and know how to use them in your absence. If you have children, give them the\nvery basics as well. For example, in case of a fire, they should know the safest way out\nwithout having to wait for you; tell them how to react to home intrusions and\nmedical emergencies, too.\n<\/p>\n<p>\nIf you are separated from others and need to meet someplace other than home,\nmake sure that everybody will remember where the meeting point is - and will know\nwhat to do if you can't make it on time. Keeping some instructions in a wallet can go\na long way.\n<\/p>\n<h4>3.8.6. Threat assessment topic #6: How can your plans backfire or fail? [<a href=\"#3.8.6\">link<\/a>]<\/h4>\n<p>\nTry to identify the aspects of your plan that are most likely to go wrong, and come up with\nviable alternatives. Take a hard look at any new problems you are creating, too:\nfor example, if you want to store gas in your garage, it will probably help in an evacuation,\nbut will also increase the odds of accidentally starting a serious fire. Or, consider a \nfar more prosaic case: if\nyou are planning to stockpile batteries or bottles of insect repellent, you should make sure\nthey can't leak and spoil other, more vital supplies nearby.\n<\/p>\n<h3>3.9. Prepper commandment #9: Plans have flaws, so also write a will [<a href=\"#3.9\">link<\/a>]<\/h3>\n<p>\nPeople die, sometimes unexpectedly and in silly, undignified ways. Dying is seldom a\npleasant affair and you generally don't get to choose the \"how\" or \"when\", but there is no\nbetter way to spoil your final moments than the sudden realization that you <i>really<\/i>\nscrewed over the people you loved the most - and that it's too late to fix that now.\n<\/p>\n<p>\nSo, if you have children, a stay-at-home spouse, or any other people who may be dependent\non you, it makes sense to write a will. Even if you don't have much of an estate to\ndispose of, your will can provide instructions for the custody of minor kids,\npotentially shielding them from abusive relatives or from foster care. This can be\nparticularly important for expats, whose closest surviving family members\nmay be in another country, difficult for the court to pinpoint or communicate with.\n<\/p>\n<p>\nIn most jurisdictions, to draft a will, you don't need a lawyer; the only skill that\ncomes handy is the ability to express yourself clearly and unambiguously. There are\ncountless state- and country-specific templates available online; in many cases, to\ncarry legal weight, the will just needs to be co-signed by disinterested parties acting\nas witnesses - or cheaply notarized.\n<\/p>\n<p>\nWhere possible, you should name an executor in the will - a trusted person who will follow\nyour instructions and take care of the formalities. If you don't, one will be\nprobably appointed by the court. This will be costly and may lead to disappointing outcomes.\n<\/p>\n<p>\nWhen writing this document, describe your \"ideal\" scenario, but also think about all the\ncomplications that may crop up and derail your plan. For example, what if both you and your\nspouse die, but your children survive? Or, who should get what share of the money if\nyour spouse is badly hurt and can't resume caring for the kids, so they end up in the\ncustody of a relative? What if your designated executor or custodian is unable or unwilling\nto perform the duties? And in an extreme case, if there are no surviving relatives,\ndo you have any favorite charity?\n<\/p>\n<p>\nOf course, for the will to be executed, it needs to be found. It makes sense to keep one\ncopy in an intuitive location in your home, because that's where people will be looking for it\nfirst; but if there's a fire or a flood, that copy may be lost. So, make another witnessed or\nnotarized copy and give it to the executor or to a close family member who doesn't live with you.\nSome folks don't recommend creating multiple legally binding copies of the same will, since it\nmay cause some confusion, but from a disaster preparedness perspective, it's a smart call.\n<\/p>\n<p>\nIn many states, even with an uncontested will, it may take many months for the probate process\nto be wrapped up. If you are the sole provider for your family, make sure that they\nwill have the means to survive in the meantime. The right kind of a shared bank account (\"joint\ntenancy with right of survivorship\") can do the trick. There are several other approaches\nto this problem, too, but they tend to be more dicey from the legal perspective - or more\ncostly and more time-consuming to set up.\n<\/p>\n<h2>4. Prepper gear &amp; supplies [<a href=\"#4\">link<\/a>]<\/h2>\n<p>\nYes! That's the part where he stops blabbing about personal finance and wills, and we all get to\nshop for nunchucks and throwing knives!\n<\/p>\n<h3>4.1. Water [<a href=\"#4.1\">link<\/a>]<\/h3>\n<p>\n...Gotcha. But yup, let's talk about the supplies and tools that may enable you to prepare for\ncommon emergencies in a cost-effective way.\n<\/p>\n<p>\nEven with adequate shelter and with limited physical activity, losing access to water means certain\ndeath after a couple of truly rotten days. Thankfully, in the developed\nworld, this is a very uncommon fate: emergencies that leave communities without potable water are not rare,\nbut when they happen, the government practically trips over itself to immediately restore service\nor to get water trucks on the road. \n<\/p>\n<p>\nAt the same time, it's not entirely crazy to worry that in some circumstances, the response may\nnot come quickly enough; heck, the Department of Homeland Security says that for the\nfirst 72 hours after a disaster,\n<a href=\"http:\/\/www.dhs.gov\/how-do-i\/prepare-my-family-disaster\">you may be on your own<\/a>\nand should have enough supplies to survive. The odds of ending up in a real pickle may be modest,\nbut the stakes are extremely high - and compared to the complexity of preparing to some\nother contingencies, the cost of stockpiling some drinking water is practically nil. If nothing\nelse, when a calamity strikes, you would have one less thing to worry about.\n<\/p>\n<p>\nAn argument can be made that even in an emergency, potable water is never too far away; after all,\nmost human settlements have been erected near natural reservoirs: rivers, lakes, or easily\nreached underground aquifers. But this is an oversimplification. In\nrural areas, water supply can be fairly meager and vulnerable to weather fluctuations and\nother cyclic phenomena. For cities, it is true that many of them are seated on the banks of major\nlakes or rivers, but suburban sprawl can easily put some residents 10-20 miles away from the\nnearest reservoir; on top of that, some of the 20th century settlements in semi-arid and\ndesert climates rely on water hauled from tens or hundreds of miles away.\n<\/p>\n<p>\nHeck, even if you do have a nearby water source, it may take surprisingly little to spoil it: for example,\nafter an unusually powerful storm, floodwaters can carry toxic sewage from treatment plants and into\nrivers and lakes. All in all, stockpiling some amount of drinking water is just a smart,\nlow-effort prepper strategy, especially in areas with an elevated likelihood of large-scale\nnatural disasters or industrial accidents.\n<\/p>\n<p>\nThis brings us to the \"how\". In most cases, the absolute minimum water intake is somewhere around\none quart per person per day; but note that this assumes no weather extremes, no substantial\nexertion, and no immediate hygiene needs. When these assumptions hold true, storing about 1.5\nto 2 gallons per household member - enough for perhaps up to a week - should provide a viable\nif modest buffer for short-term emergencies. Store-bought gallon jugs are pretty cheap,\nhassle-free, and easy to squeeze in just about anywhere; if you keep them away from sunlight\nand heat, they should last 5+ years before needing to be rotated or thrown out. Don't try to\nsave a buck by reusing milk or juice jugs, though: they are almost impossible to clean properly\nand may end up supporting bacterial growth.\n<\/p>\n<p>\nWell, that's the bare minimum - but if you have a garage, a basement, or other unused storage\nspace, I would actually recommend going a bit further and grabbing one or two\n<a href=\"http:\/\/smile.amazon.com\/Scepter-04933-Water-Can-5-Gallon\/dp\/B000MTI0GA\/\">5 gallon\ncans<\/a> ($18) per every household member. Although multi-week water outages are very unlikely,\nthis simply gives you a more comfortable safety margin: if something goes awfully wrong and\nit becomes clear that help is not coming any time soon, you will still have time to look for\nalternatives or evacuate. A reserve also puts you in a better situation if it's unusually\nhot or if you have any urgent hygiene needs. The cans are very easy to use: wash them with a \nsmall amount of regular, non-scented laundry bleach, rinse, and fill up with tap water.\nRotate the contents every 2-4 years or so.\n<\/p>\n<p>\nIf you want to prepare for more outlandish, long-term disasters, or if you have a large family,\nyou can make realistic plans only if you live in a single-family home. This\ndecision alone may give you access to 50-100 gallons of water sitting in the water heater\n(unless it's a tankless design). Next, if you have some backyard space, you may opt for\nrelatively inexpensive\n<a href=\"https:\/\/www.zoro.com\/eagle-transport-drum-open-head-575-gal-yellow-1656\/i\/G7611895\/\">55 gallon\nplastic barrels<\/a> ($60), taking up about as much space a small curbside trash can. Another \ncommon option, costing about the same per gallon stored, is a \n<a href=\"http:\/\/smile.amazon.com\/275-Gallon-Tote-Metal-Cage\/dp\/B003SW6WHM\">275 gallon tote<\/a> ($325).\nSuch solutions can easily provide water for an entire family for up to several months at\n2 quarts per person per day, with some\nallowance for laundry and hygiene needs. Some preppers stockpile even more - but really, if you\nwaited this long and the conditions are still dire, it may be high time to hit the road and find\nsome other place to live.\n<\/p>\n<p>\n(Of course, there are rural places that get water only seasonally, and where huge\nstorage cisterns or rainwater collection systems are already a necessity; but within the scope\nof this guide, we're looking beyond what's normally needed to subsist.)\n<\/p>\n<p>\nWhile home storage of water is not hugely complicated, things get a bit dicey when you have\nto evacuate - or if you end up being stranded away from home. If you have a car, your best\nbet is to put together a small emergency supplies box that, among other essentials, houses\none or two 1-gallon jugs of water - and keep it in your trunk at all times. But without a\ncar, your prospects are less cheerful: in case of a widespread disaster, your range will be\nseverely limited, and even if you take some modest amount water with you, you will need to\nreach a more hospitable location within 1-2 days. A bicycle, a plan, and a good map\nwill help. A\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00X5UDZ2M\">folding cart<\/a> or a\nan inconspicuous box of supplies kept at work may be viable choices, too.\n<\/p>\n<p>\nAs for drinking untreated water: contrary to popular beliefs, in temperate climates, you are\ngenerally not taking huge risks by drinking from a backcountry lake or a creek; if it looks\nand smells all right, it's quite likely fine. On the flip side, a bout of diarrhea is probably\nthe last thing you want to experience in such a situation, so it's good to take precautions\nif you can. Boiling your drinking water is a very robust method of eradicating microscopic\nwildlife (more about that soon). When boiling is not an option, adding several drops of regular,\nold-fashioned \nlaundry bleach per gallon of water, then letting it sit for 30-60 minutes, will have a roughly\ncomparable effect. Note that bleach has a limited shelf life; you will need to rotate it every\n5 years or so. When on the\ngo, <a href=\"http:\/\/www.amazon.com\/Aquatabs-Water-Purification-Tablets-pack\/dp\/B00GG19KWQ\">sodium\ndichloroisocyanurate<\/a> pills can be more convenient than liquid bleach and work just as well.\n<\/p>\n<p>\nThere are several other water purification techniques, including iodine, various types\nof filters, or even ridiculously expensive desalination systems.\nThe recent <a href=\"https:\/\/en.wikipedia.org\/wiki\/Flint_water_crisis\">Flint water crisis<\/a>\nhighlights the value of permanently installed and well-maintained RO units. But for \nshort-term survival, my take is that most of the filtration systems targeted at preppers\nincrease costs without offering clear-cut benefits. And really, don't overthink it: if the choice is between dehydration\nand drinking straight from a scummy pond, drink from the pond.\n<\/p>\n<h3>4.2. Food [<a href=\"#4.2\">link<\/a>]<\/h3>\n<p>\nYou can survive several weeks without food, but you won't be having a very good time. Food is\ncostly, its supply is fairly easily disrupted, and it's a resource that the government may be\nmuch less inclined to deliver to your doorstep when things go wrong. So, with a variety of\nreasonable scenarios to worry about - anything from natural disasters to economic downturns - it\njust makes sense to be able to feed yourself even if you can't buy groceries for a while.\n<\/p>\n<p>\nOf course, everybody has some non-perishable food around the house, but it's much better to create a\ndedicated stash: this way, you can count on the supplies always being at hand, and you can stockpile\nsomething more nutritious than stale crackers, a suspect bottle of olive oil, and a rusty can of tomato\nsauce. With a well thought-out stockpile of ready-made food, it's also a lot easier to hit the road.\n<\/p>\n<p>\nFor short-term survival, the highest priority is just energy - preferably in a form that doesn't\ncost much, lasts forever, requires no preparation, and takes up little space. Generally,\nyou're looking at high-sugar or high-fat foods. Some of the common zero-preparation options include:\n<\/p>\n<ul><li>\n<p>\n<b>High-sugar energy bars.<\/b> \nGrocery store brands have limited shelf life, but several prepper-targeted,\n<a href=\"http:\/\/smile.amazon.com\/gp\/product\/B00SVX00NU\">Mylar-packed varieties<\/a> can last\n5-10 years. Such products are inexpensive (~300 kcal per dollar), convenient, and energy-dense\n(~2,000 kcal per pound). On the flip side, they are probably pretty nauseating as your primary\nfood. Imagine living solely off Jelly Bellies for a week.\n<\/p>\n<\/li><li>\n<p>\n<b>Dry survival rations.<\/b>\nSold under several brands, including\n<a href=\"http:\/\/smile.amazon.com\/gp\/product\/B00WT6M4BG\">Datrex<\/a>,\n<a href=\"http:\/\/smile.amazon.com\/gp\/product\/B0089EOPDM\">ER Bar<\/a>,\n<a href=\"http:\/\/smile.amazon.com\/gp\/product\/B00MDY91HY\">S.O.S.<\/a>,\n<a href=\"http:\/\/smile.amazon.com\/gp\/product\/B00NGYGCH2\">Grizzly<\/a>,\nand more. Biscuit-like, less sugary and with a more agreeable taste than energy bars -\nsomewhat reminiscent of shortbread. In my book, S.O.S. and Datrex taste best.\nVery inexpensive (~550 kcal per dollar) and should last 5-10 years when stored in a\ncool place. A tolerable choice for short- \nto medium-term nutrition in an emergency. Easy to pack, giving you ~2,200 kcal\nper pound.\n<\/p>\n<\/li><li>\n<p>\n<b>Canned meat, veggies, or fruit.<\/b>\nStorage life in excess of 20 years (regardless of \"best by\" dates).\nTasty, relatively cheap (~200-300 kcal per dollar), and the choice is pretty broad.\nFruits, veggies, and soups are not very energy-dense (~200 kcal per pound), making them\nimpractical for hiking or bugging out; on the flip side, the syrup may provide some additional\nhydration. Meats fare much better, tipping the scales at around 1,500 kcal per pound.\nCanned foods are a good option for longer-term planning, provided that you have enough\nshelf space.\n<\/p>\n<\/li><li>\n<p>\n<b>Freeze-dried dinners.<\/b>\nPopular with hikers. Extremely lightweight (up to 2,500 kcal per pound) and surprisingly\ntasty. The most reputable brand is\n<a href=\"http:\/\/smile.amazon.com\/s\/?&amp;rh=i%3Aaps%2Ck%3A%22mountain+house%22+%22pouch%22%2Cp_85%3A2470955011\">Mountain House<\/a>.\nFairly expensive on a calorie basis (100-150 kcal per $1), but you get a choice of\nraspberry crumble, chicken with dumplings, bacon and eggs, and everything in between.\nStorage life in excess of 15 years. The drawback is that they need some boiling water to\nreconstitute (cold water will also work, but not make a tasty meal).\n<\/p>\n<\/li><li>\n<p>\n<b>Military MREs.<\/b>\nI don't find them particularly tasty, but they are popular among preppers. \nA bit on the heavy side (usually around 1,100 kcal per pound).\nPortable warm food with a ton of different menus available - although for the best price,\nyou usually need to buy a variety box and can't cherry-pick. Moderately expensive \n(~150 kcal per dollar). Shelf life around 5-7 years, depending on the manufacturer and on storage\nconditions. A good source is <a href=\"http:\/\/theepicenter.com\/\">TheEpicenter.com<\/a>.\n<\/p>\n<\/li><li>\n<p>\n<b>Nuts.<\/b>\nAt ~3,000 kcal per pound, nuts are surprisingly energy-dense for something that can be eaten as-is.\nRelatively cheap (~400 kcal per dollar). The major drawback is a relatively short shelf life,\nprobably not exceeding a year or two. Nuts also do not provide complete nutrition, but that's only\na concern in the long haul.\n<\/p>\n<\/li><li>\n<p>\n<b>Vegetable shortening.<\/b>\nExtremely cheap and energy-dense (2,500 kcal per dollar, 4,100 per pound), making it a unique choice\nwhen space or money is in critically short supply. In contrast to other common fats,\nstore-brought <a href=\"https:\/\/smile.amazon.com\/dp\/B00I8G7QYQ\/\">cans of vegetable shortening<\/a> should stay\nfine for 4 years or more. The product is very bland, but it's perfectly palatable when spread on crackers, mixed\nwith bacon bits, and so forth. It does not provide complete nutrition - but again, that won't harm you in the\nshort haul.\n<\/p>\n<\/li><\/ul><p>\nFor folks who want to focus on the most plausible risks, I suggest stockpiling dry survival rations\nto last around 2-3 weeks; seven 3,600 kcal packets per person should do the trick.\nThat may sound like a lot, but keep in mind that it may take a while for stores to reopen after an earthquake or a\nflood - so it's not a bad plan to play it safe. And while eating ration bars for a week sounds bland, you may\nbe in no mood for home cooking when you have to fix a collapsed roof and the utilities are cut off.\n<\/p>\n<p>\nFor those who are worried about less likely, longer-term contingencies - or who want to limit their future\ngrocery expenses in case of a financial shortfall - a more varied stockpile to cover 3-4 months is a reasonable\nchoice. At that timescale, it's still smart to begin with some number of hassle-free survival rations, but it is\nimportant complement them with a more palatable menu: freeze-dried or canned meals, MREs, or cheap home-made food.\nCrisco aside, some of the nutritious and easily stored staples include <a href=\"http:\/\/smile.amazon.com\/dp\/B00967K6CA\">Mylar-bagged,\noxygen-scavenged<\/a> white rice, white flour, dried beans and grains, instant mashed potatoes and oatmeal, pasta, sugar, honey,\npowdered milk, salt, spices, and so forth; when stored properly, all of them can last 5 years or more. \n<a href=\"http:\/\/www.augasonfarms.com\/Fruit-Vegetables\">Freeze-dried<\/a>\nor garden-grown fruit and veggies can add some flavor to your post-disaster cooking, too. Heck, you\ncan even buy supposedly tasty\n<a href=\"http:\/\/www.augasonfarms.com\/Butter-Powder-10-Can-upc-000946903632\">butter<\/a>,\n<a href=\"http:\/\/www.augasonfarms.com\/Freeze-Dried-Chicken-Breast-Chunks-Real-Meat--10-Can-upc-078716111709\">chicken breast<\/a>, \n<a href=\"http:\/\/smile.amazon.com\/dp\/B003RC5FQ2\">canned bacon<\/a>, and\n<a href=\"http:\/\/smile.amazon.com\/dp\/B000M7V1F4\">bread-like crackers<\/a>\nwith 10+ year shelf life (although they are not cheap).\n<\/p>\n<p>\nAs for the appropriate calorie intake: almost all adults can function normally on 1,500 calories a\nday for extended periods of time, although they will slowly lose weight (probably not more than\nhalf a pound a week). If you are skinny or if you're aiming for surviving many months with no access to other\nfood, budgeting 2,000-2,200 kcal per day is a safer bet. Since you would have to exercise portion control,\nit pays to focus on satiating foods: high-fiber, high-protein, or high-fat.\n<\/p>\n<p>\nFor hardcore preppers convinced that they may be left with no access to food for a very long time,\nit would be also important to maintain a robust intake of protein; somewhere around 40-50 g per\nday is believed to be optimal, although you certainly don't need to observe it religiously.\nReady-made protein-rich foods include some energy bars, most dry ration packets, some\nfreeze-dried dinners, canned meat, and military MREs; smaller amounts\ncan be found in some veggies, too. You can also stockpile\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0015AQL1Q\">protein powder<\/a> - it's bland but relatively\ncheap ($1 per day). \n<a href=\"http:\/\/smile.amazon.com\/dp\/B0002YSQ5U\">Freeze-dried scrambled eggs<\/a>,\n<a href=\"http:\/\/smile.amazon.com\/dp\/B004VUJDYE\">powdered milk<\/a>,\nand related products, including long shelf life\n<a href=\"http:\/\/smile.amazon.com\/\/dp\/B00TEMS4U8\/\">canned<\/a> or\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00FTETVO8\">powdered cheeses<\/a> and \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00FTETGAW\">pancake mixes<\/a>, work\nwell, too. As mentioned earlier, protein and dietary fiber also have a well-established satiating\neffect, helping you maintain a balanced diet - which can otherwise be tricky when snacking on\nhigh-calorie foods. Oh: having some OTC multivitamins may be a good plan, especially\nto supplement vitamin C.\n<\/p>\n<p>\nOf course, some \"doomsday\" preppers worry about even more exotic, post-apocalyptic scenarios\nmentioned in section 2.3, basically aiming for indefinite self-sufficiency. I don't think it's\na particularly sound concern, but if the prospect of\na civilizational collapse keeps you up at night, my best advice is to move to a rural community\nwhere you could farm, fish, or hunt. Some urban survivalists fantasize about trapping local\nsquirrels, pigeons, or raccoons - but they would run out of food very fast. Small urban and\nsuburban gardens are usually difficult to maintain and don't produce enough to feed a family,\ntoo.\n<\/p>\n<h3>4.3. Fuel and electricity [<a href=\"#4.3\">link<\/a>]<\/h3>\n<p>\nIt takes just a single downed power line to knock out your furnace, AC unit, cooktop, refrigerator,\nand to make the lights go out; and when such an outage happens due to a larger-scale natural\ndisaster, repairs can easily take days or even weeks. We think of fuel as a more dependable resource, but\nif 1979 is any guide, you only need  <a href=\"https:\/\/en.wikipedia.org\/wiki\/1979_energy_crisis\">one\nwell-timed revolution<\/a> in the Middle East to make it nearly impossible to fill up a car in some\nparts of the United States. Of course, such events are usually inconvenient, not disastrous, so it's\nperfectly fine not to dwell on them in your plans. At the same time, it doesn't hurt to take a closer\nlook at what's at stake - and what the potential solutions may be.\n<\/p>\n<p>\nOf all the plausible scenarios, another major oil crisis would probably hit most car-owning families\nthe hardest, limiting their ability to get food or to take care of other, everyday needs. Generally\nspeaking, there is no simple fix: keeping a gallon or two in your garage won't make much of a\ndifference, while maintaining significant reserves of gas for personal use can be done safely (and\nlegally) only if you own a large, rural plot of land. Electric vehicles,\nespecially if charged from rooftop solar panels, can offer a wonderful backup in some parts of\nthe world, but they carry a very hefty price tag. The best workaround may be the least inspired\none: if you own a car, you can always keep your tank at least half full (a familiar mantra by now),\nand have enough food and other essentials to be able to wait out the worst.\n<\/p>\n<p>\nA service interruption or a fuel crisis that takes your cooktop out of service for a week or two\nis the other hardship perhaps worth worrying about. It's not just about eating well: in an\nemergency, the ability to boil water is one of the best methods of making it safe to drink. While\nthe owners of rural homes with 500 gallon propane tanks may have little to worry about, the rest of\nus would not be having fun. For those who cook using municipal natural gas, a simple backup is a\n<a href=\"http:\/\/smile.amazon.com\/dp\/B005T0SN0K\">small, countertop electric burner<\/a>, costing\nabout $15. Conversely, for people with electric ranges, a portable \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00B4FY8YO\">camping stove<\/a> ($13) and a handful of dirt-cheap\n1 lb propane tanks ($4 at any hardware store) can be a safe, no-hassle choice. A\npound of propane can boil around 12 gallons of water; the entire setup is also very easy to put in a\nbackpack if you ever need to leave - so it's basically worth getting either way.\n<\/p>\n<p>\nNo heating in the middle of a particularly nasty winter can be problematic, too - although it's mostly a\nmatter of comfort, not survival. In most places, with robust shelter and adequate clothing, bedding, and food, it's\nfairly hard to freeze to death at home (but note that the cold may make some infections or medical\nconditions worse; you may have to worry about frozen water pipes, too). The situation can become a lot more dire if you are on\nfoot in the middle of nowhere, so truly hardcore, wilderness-minded preppers may have something to ponder about; but\nhauling a sufficient amount of fuel is typically impractical to begin with, so their best bet would be warm clothing,\nimprovised shelter, and the ability to build a fire. We'll talk about that in the section that\ndeals with camping supplies.\n<\/p>\n<p>\nIn some parts of the world, extreme heat can be far more dangerous than cold. When AC is not an option,\nit's usually possible to avoid trouble by staying in the shade, drinking a lot, and limiting\nphysical activity. If it gets really nasty, the best way to cool yourself is to wet your clothing\nand hair, then stand in front of a running fan. You have a bigger problem if you happen to be\nstranded in a broken down car somewhere in the middle of a desert - but carrying some water\nand several other supplies in your trunk should help a great deal. More about that soon.\n<\/p>\n<p>\nDuring a prolonged blackout, keeping flashlights, radios, and cell phones running can become a \nchallenge of its own. The most cost-effective approach is to stick to devices that can take\nregular AA, AAA, or PP3 (9V) batteries; modern alkaline cells have very long shelf lives (10 years),\ncan be  <a href=\"http:\/\/smile.amazon.com\/dp\/B00466TDS4\">bought cheaply<\/a> in bulk, and will do the\ntrick for almost all portable electronics you can think of. Of course,\nmodern power-hungry smartphones are a notable exception to this rule. For that, you can always try \n<a href=\"http:\/\/smile.amazon.com\/dp\/B017GQ7OEA\">solar chargers<\/a> ($50) - they work\nwell, but are a bit fragile and may not perform too amazingly in wooded areas or on overcast\ndays. You can also go for\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00H4JDHRS\">hand-crank<\/a> generators ($30, labor-intensive)\nor <a href=\"http:\/\/smile.amazon.com\/dp\/B00D2ZQ6ZI\">AA power banks<\/a> ($15, mixed reviews). But\nultimately, also allow for the possibility of not being able to call others for a while.\n<\/p>\n<p>\nFor powering more serious electrical equipment, a generator is a popular choice for people living \nin the backwoods. That said, this option comes with an interesting trade-off: if you were ever\nto face a contingency that may last for a longer while, it may be more important to conserve\nfuel for driving, cooking, or heating, than to use it for keeping the lights on. A fully-fledged\nsolar installation helps you\navoid such dilemmas, but costs an arm and a leg. A possible compromise is a jury-rigged\nsolar setup done at a smaller scale:\nif you hook up <a href=\"http:\/\/smile.amazon.com\/dp\/B00B8L6EFA\">100W panel<\/a> ($160) to a\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00K8E0WAG\">deep-cycle lead-acid battery<\/a> ($60) and a \n<a href=\"http:\/\/smile.amazon.com\/dp\/B004NZI4K6\">low-cost inverter<\/a> ($30), you gain the ability to\nrecharge laptops and phones, or even power several desk lamps, a decent-size fan,\nor a small refrigerator. The whole contraption costs around $250 and is easy to stow away if\nyou're not very short on space.\n<\/p>\n<p>\n(If you take this route, charge the lead-acid battery fully, bag it, and then put it in the fridge\nor in another cool place; this will extend its shelf life by many years. Li-poly and regular \nnon-rechargeable batteries benefit from being stored at low temperatures, too. But whatever you do, don't\nfreeze any of them!)\n<\/p>\n<p>\nThat's probably about it... well, all right: this section focused chiefly\non the immediate consequences of an outage, but a severe fuel crisis or a long-lasting power grid\nfailure would have profound, cascading effects on the entire economy - probably including out-of-control\nunemployment, high inflation, product shortages, and more. That said, these are the outcomes we can already\nprepare for by other means. As for extreme\npreppers who aspire to long-term energy self-sufficiency, I think it's going to be a\ndifficult feat: even with a solar installation, under constant cycling, the batteries may not last\nmuch longer than 5 years. Short of finding a cheap Soviet \n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Radioisotope_thermoelectric_generator\">RTG<\/a>\non eBay, they may simply have to adapt to living without electricity or gas.\n<\/p>\n<h3>4.4. Electronics [<a href=\"#4.4\">link<\/a>]<\/h3>\n<p>\nAlthough this may sound very disappointing to a typical geek,\nthe list of genuinely useful prepper electronics is pretty short. Here's what I can honestly recommend:\n<\/p>\n<ul><li>\n<p><b>A large pen drive.<\/b>\nComputer hardware failures are far more common than space zombies or mutant superbugs. \nBecause of this, one of your best investments can be a \n<a href=\"http:\/\/smile.amazon.com\/dp\/B007YX9OGW\">decent 128 GB pen drive<\/a> ($30) with\na copy of all your important files; in case of bank mix-ups, throw in copies of\nrecent account statements, too. And hey, if want to feel like\na cyber-ninja - you can always grab a\n<a href=\"https:\/\/dumps.wikimedia.org\/enwiki\/latest\/enwiki-latest-pages-articles-multistream.xml.bz2\">copy of\nWikipedia<\/a>. It will undoubtedly come handy for rebuilding the civilization, and it's just 12 GB.\n<\/p>\n<\/li><li> \n<p><b>Flashlights.<\/b>\nUnless you are living in a rural area, you don't need an eye-searing torch that \nchews through ten boxes of batteries in a day. Get two small, high-quality AA flashlights that\ngive you at least 20 hours on low power; keep one near your bed, and another\nin your car or in an emergency stash. For a low-cost option,\ntry <a href=\"http:\/\/smile.amazon.com\/dp\/B00I14HLLS\">Fenix E12<\/a> ($25). If you want\n100+ hours of battery life and don't mind the price tag, check out\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00KA50G68\">Fenix LD22<\/a> ($55).\n<\/p>\n<\/li><li>\n<p><b>A lantern.<\/b>\nNot essential, but \nuseful for preparing food, dining, reading, and other fun blackout activities where a narrow beam\nwould be less comfortable than omnidirectional light. I like\n<a href=\"http:\/\/smile.amazon.com\/dp\/B004AJLUGK\">this device<\/a> ($35) - it's small,\nfairly inexpensive, and very robust.\n<\/p>\n<\/li><li>\n<p><b>An old-fashioned radio receiver.<\/b>\nA battery-operated AM\/FM radio will be a good way to stay in the loop if cell\nnetworks and the Internet are down, and the civilization is temporarily banished\nback to the dark ages (aka the 90s). A cheap, brand-name model, such as\n<a href=\"http:\/\/smile.amazon.com\/dp\/B012F0DGYE\">Sony ICFP26<\/a> ($18), will do just fine.\n<\/p>\n<\/li><li>\n<p><b>Handheld FRS\/GMRS radios.<\/b>\nMany preppers obsess about long-distance communications, but in a typical emergency, chatting\nwith people 100 miles away is not a priority. In contrast, a hand-held two-way radio\ncan be very useful for keeping in touch with your friends and family during any prolonged\noutage. Again, pick\na device that accepts the kind of batteries you can stockpile cheaply. Expect a range\nof 2-3 miles in rural regions, and less than a mile in highly urbanized areas -\nno matter what the manufacturer claims. With all that in mind, \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00M1BP3T4\">Olympia R500<\/a> ($55) is a good choice.\n<\/p>\n<p>\nHam radio is a longer-distance option, but generally requires a license; if you are interested\nin amateur radio, getting a Technician license in the United States is \n<a href=\"http:\/\/www.arrl.org\/getting-licensed\">fairly straightforward<\/a>, and can help\nyou coordinate local response to more substantial emergencies.\n<\/p>\n<\/li><li>\n<p><b>A thermometer that won't run out of juice.<\/b>\nResponding to serious emergencies can be stressful and physically taxing, making it easy to catch\nnasty infections along the way. To know how bad things have gotten, it's good to have a reliable way to\ntake body temperature; keep in mind that many low-cost axillary thermometers use LR41\nbatteries, and that you probably don't have any spares lying around. One good choice is\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00MUK6M82\">this<\/a> ($35). A traditional glass\nthermometer will also work, but is more fragile.\n<\/p>\n<\/li><li>\n<p><b>A small space heater and a fan.<\/b>\nSometimes useful for coping with temperature extremes. Decide for yourself if you\nneed it. If yes, pretty much any make and model will do.\n<\/p>\n<\/li><li>\n<p><b>A dashcam or a security camera for your home.<\/b> We'll talk about these later on.\n<\/p><\/li><\/ul><p>\nExcept for flashlights, don't keep any batteries in your emergency electronics, as to minimize the danger of an electrolyte leak messing up the device.\n<\/p>\n<p>\nOf course, there are many other high-tech gadgets popular among some of the more affluent \nand paranoid preppers\n- anything from satellite phones, to night vision goggles, to heated insoles. In all\nlikelihood, none of that is worth the cost. If I had to pick two extravagant\n\"doomsday\" accessories that could conceivably be useful to some people if something\ntruly awful happens to the world, I'd go with\na <a href=\"http:\/\/smile.amazon.com\/dp\/B002EOSQII\">waterproof hiking GPS unit<\/a> and\na portable Geiger counter. Both\nare powered by AAA batteries and cost around $140. (We'll talk about Geiger counters\nand their relative merits a bit later on.)<\/p>\n<p>\nInspired by science fiction books and a handful of real-world incidents, some\nhardcore survivalists worry that their portable electronics or vehicles could be\ndisabled by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Electromagnetic_pulse\">EMP weapons<\/a>\nor solar flares. The concern over solar flares is misplaced; the threat of EMP is a tad\nmore valid, but even if EMP warfare came to pass, small electronics and quasi-shielded automotive\ncircuitry would probably not be permanently affected by anything other than a\nclose blast. Power plants and transmission lines are a different story. Heck, in 1989, solar\nflares knocked out  <a href=\"https:\/\/en.wikipedia.org\/wiki\/March_1989_geomagnetic_storm\">a good\nchunk of the Canadian power grid<\/a>. Still, for that, a surge protector works better than tinfoil.\n<\/p>\n<h3>4.5. Essential tools [<a href=\"#4.5\">link<\/a>]<\/h3>\n<p>\nWhile not all urban-dwelling readers may be particularly interested in DIY work, there are times when\nit's hard to call a handyman. So, almost every home will benefit from having a well-maintained \"emergency\"\ntoolbox containing several items useful for performing basic repairs and dealing with other minor incidents:\n<\/p>\n<ul class=\"compact\"><li> <p>A cheap pocket knife (<a href=\"http:\/\/smile.amazon.com\/dp\/B0074FI28Q\">link<\/a>, $20),<\/p>\n<\/li><li> <p>A basic, no-frills multitool (<a href=\"http:\/\/smile.amazon.com\/dp\/B005DI0XM4\">link<\/a>, $30),<\/p>\n<\/li><li> <p>A small adjustable wrench (<a href=\"http:\/\/smile.amazon.com\/dp\/B000NIK9S2\">link<\/a>, $8),<\/p>\n<\/li><li> <p>At least 100 ft of paracord or UHMWPE rope (<a href=\"http:\/\/smile.amazon.com\/dp\/B00ACL4KJO\">link<\/a>, $10),<\/p>\n<\/li><li> <p>A big roll of duct tape (<a href=\"http:\/\/smile.amazon.com\/dp\/B00DVBPPGY\">link<\/a>, $10),<\/p>\n<\/li><li> <p>A rudimentary sewing kit (<a href=\"http:\/\/smile.amazon.com\/dp\/B001682BQQ\">link<\/a>, $5),<\/p>\n<\/li><li> <p>A decent pen or a mechanical pencil (<a href=\"http:\/\/smile.amazon.com\/dp\/B000WGD13U\">link<\/a>, $12),<\/p>\n<\/li><li> <p>A pack of writing pads (<a href=\"http:\/\/smile.amazon.com\/dp\/B00QSR9BT0\">link<\/a>, $6),<\/p>\n<\/li><li> <p>Sharp, large scissors (<a href=\"http:\/\/smile.amazon.com\/dp\/B001BKHHGS\">link<\/a>, $5),<\/p>\n<\/li><li> <p>Needle-point tweezers (<a href=\"http:\/\/smile.amazon.com\/dp\/B000FFYMYU\">link<\/a>, $10),<\/p>\n<\/li><li> <p>A compact but sturdy hammer (<a href=\"http:\/\/smile.amazon.com\/dp\/B0042D6HLW\">link<\/a>, $15),<\/p>\n<\/li><li> <p>A box of 1\" nails (<a href=\"http:\/\/smile.amazon.com\/dp\/B00LP3Z0NS\">link<\/a>, $2),<\/p>\n<\/li><li> <p>A usable can opener (<a href=\"http:\/\/smile.amazon.com\/dp\/B000AX13US\">link<\/a>, $7),<\/p>\n<\/li><li> <p>A sealed bag with several boxes of matches ($5).<\/p>\n<\/li><\/ul><p>\nHaving a <a href=\"http:\/\/smile.amazon.com\/dp\/B01EBX4WOG\">medium-size bucket<\/a> ($8) at home is a\nmust, too. If you own a bicycle and are expecting to use it in emergencies, it would be wise to\nthrow in a <a href=\"http:\/\/smile.amazon.com\/dp\/B0012Q41HY\">bike tool<\/a> ($20),\nseveral <a href=\"http:\/\/smile.amazon.com\/dp\/B000IZGFCE\">tire levers<\/a> ($5),\na <a href=\"http:\/\/smile.amazon.com\/dp\/B0016HSA4O\">patch kit<\/a> ($5), \none or two spare tubes ($10), and a \n<a href=\"http:\/\/smile.amazon.com\/dp\/B007G8EXH0\">portable pump<\/a> ($10).\nFinally, for those who are worried about the decidedly unlikely prospect of having to escape home\nand fight off radscorpions in the wilderness, a\na <a href=\"http:\/\/smile.amazon.com\/dp\/B0002YTO7E\">lightweight hatchet<\/a> ($25),\na <a href=\"http:\/\/smile.amazon.com\/dp\/B0001IX7OW\">folding saw<\/a> ($20),\na <a href=\"http:\/\/smile.amazon.com\/dp\/B00XJE9MJE\">larger fixed-blade knife<\/a> ($24),\na <a href=\"http:\/\/smile.amazon.com\/dp\/B014IEHPYC\">folding shovel<\/a> ($25),\na <a href=\"http:\/\/smile.amazon.com\/dp\/B001ID4ZY0\">compass<\/a> ($9),\nand some matches or a lighter in a waterproof container\ncan come handy in several ways.\n<\/p>\n<p>\nA subset of this is also worth keeping in a car. It's not just about zombies or life-and-death situations:\nif you hit something in a parking lot and your bumper cover comes off or your liftgate won't stay\nshut, it's nice to be able to tie it down and get back on the road. Similarly, a shovel can help you\nget back on the road after getting stuck in snow or mud. But speaking of survival: a pocket\nknife, kept within reach (e.g., in the center console), can be used to cut\nseat belts if you get into a wreck; and in a pinch, it will double as a self-defense tool.\n<a href=\"http:\/\/smile.amazon.com\/dp\/B001PTGOKA\">Belt cutters<\/a> can also fulfill that first task,\nand may be easier to operate if you are hurt - although they are less useful for other purposes.\nI'd also\nrecommend getting an <a href=\"http:\/\/smile.amazon.com\/dp\/B0046HJ950\">automatic center punch<\/a> -\nit's a neat $7 tool that effortlessly shatters tempered glass (i.e., side and rear windows) when\nthe doors won't budge. It <a href=\"https:\/\/www.youtube.com\/watch?v=l6C5Ytc5Soc\">works way better<\/a>\nthan many of the specialized car escape devices sold on the Internet.\n<\/p>\n<p>\nIf you own a house, especially in a region prone to earthquakes or tropical storms, you should\nprobably have a sledgehammer, a chainsaw (with a charged battery or some fuel at hand), bolt cutters, and\na pry bar. These heavy tools are essential for clearing debris and getting to whatever's underneath. Keep\nthem far from your other supplies: if your primary stash gets pinned under other junk, you can use the tools\nto get it out. Don't store pry bars and similar equipment in plain sight; robbers often use found\ntools to force patio doors, to pop safes, or\n<a href=\"https:\/\/www.google.com\/search?q=cut+wall+burglary\">worse<\/a>.\n<\/p>\n<p>\nNo matter where you live, it's also nice to have some materials at hand to patch up broken windows on a stormy\nnight. <a href=\"http:\/\/smile.amazon.com\/dp\/B00CX5WVH0\">Window security film<\/a> can be used to keep broken\nglass in place, while a roll of thick plastic sheeting or tarp from a hardware store can come handy for temporary\nrepairs; space permitting, you may also want to keep several wooden planks. For suburban and rural homes in regions\nprone to extreme weather, pressboard and sandbags may be worthwhile, too.\nTraditional sandbags tend to be extremely labor-intensive to fill, so plan accordingly; \n<a href=\"http:\/\/hydrabarrier.com\/collections\/hydrabarrier\/products\/hydrabarrier-ultra-6\">water-filled\nbarriers<\/a> are much easier to deploy, but cost more.\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0085S2DEU\">Absorbent sock-style barriers<\/a> can be used to deal with\nminor flooding, but only up to an inch of water or so; they may be less useful for inclement weather,\nbut may prove indispensable for dealing with backed-up sewage or similar ills.\n<\/p>\n<p>\nAnd now, for something completely different: during a longer water outage, you won't be able to flush your\ntoilet - a little-appreciated but grave hygiene risk. When living in a single-family home, you should probably\nget a shovel and a pickax: they are useful in the backyard either way, but if push comes to shove, such tools\nallow you to dig out a latrine and address the sanitation problem in a fairly sustainable way. Of course, dumping\nbagged human waste into trash will work for a while, too, but it quickly becomes a liability.\n<\/p>\n<h3>4.6. Camping equipment [<a href=\"#4.6\">link<\/a>]<\/h3>\n<p>\nMany hardcore preppers spend their time fantasizing about heroic survival in the endless, pristine wilderness,\nequipped with nothing more than a bug out bag, a trusty rifle, and their own iron will.\nBut even in far more realistic situations, being able to set up a camp can be a valuable skill. During mass evacuations,\nthere is always a good chance of being stuck on a congested highway for a day or two, or reaching your destination only\nto find out that all the hotels and motels are full. The benefits are clear for some\nsmall-scale emergencies, too: if backed-up sewage makes your home uninhabitable for a while, setting up a tent\nin a friend's backyard can be much cheaper than staying in a hotel for several weeks.\n<\/p>\n<p>\nAbove all, the nice thing about it is that camping gear doesn't need to just sit in your closet, collecting dust\non the off chance that something bad may happen a decade from now. You can simply grab it\nand head out for the weekend every now and then; camping is fun, doubly so for kids. It's also a great opportunity\nto test some of your other equipment, and spot potential flaws in your preparedness plans.\n<\/p>\n<p>\nEspecially if you're living in the suburbs or in the countryside, I recommend having the following items\nat hand:\n<\/p>\n<ul><li><p><b>Maps.<\/b>\nEssential for finding hospitable destinations, getting there, and finding alternative routes if something goes\nwrong. I suggest getting both a country road atlas and a more detailed map of your county or state. If you have\na car, just keep them there - they may also come handy if you get lost and your cell phone dies.\n<\/p>\n<\/li><li><p><b>Weather-appropriate clothing.<\/b>\nA well-maintained stash of warm clothes, including \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00B0I7T00\">waterproof ponchos<\/a> and rain boots. In a pinch, you can also use\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00IDBOZ5Y\">metallized Mylar blankets<\/a> ($0.80 a piece): tie them with some\ntape to make improvised rainproof, windproof, or thermally insulating clothing and hats, shoe liners,\nand more.\nThe blankets cost very little and take up virtually no room, so I strongly suggest keeping some in your\ncar. If you're stranded in an inhospitable place, they could save your life.\n<\/p>\n<\/li><li><p><b>Shelter.<\/b>\nThe usual combo is an appropriately-sized, waterproofed tent, a set of sleeping bags &amp; pads, and possibly\nsome <a href=\"http:\/\/smile.amazon.com\/dp\/B002PW8SFK\">compressible or inflatable pillows<\/a> for comfort.\nIn cold weather or during heavy rainfall, it's also possible to shelter in your vehicle, with Mylar blankets\nserving as a substitute for sleeping bags; and on a hot day, the same material and some rope can be used\nto improvise a shade. Again, the blankets are worth having in your car at all times.\n<\/p>\n<\/li><li><p><b>Food prep gear.<\/b>\nAs discussed in section 4.3, one of the best ways to cook food \nor to sterilize drinking water on the go is to have several portable propane tanks,\na <a href=\"http:\/\/smile.amazon.com\/dp\/B00B4FY8YO\">miniature stove<\/a> ($13),\nand a lightweight <a href=\"http:\/\/smile.amazon.com\/dp\/B007OJKI2U\">covered pot<\/a> ($28). \nSome <a href=\"http:\/\/smile.amazon.com\/dp\/B001LF3IC0\">plastic dinnerware<\/a> ($13) can be\na nice touch.\nMake sure that the stove fits your propane tanks; you may need a \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00CFPISZW\">Lindal valve adapter<\/a> ($20) if not.\n<\/p>\n<\/li><li><p><b>Mosquito repellent.<\/b>\nYep. End up in the wrong part of the woods and these bastards can suck you dry. Go for\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0054NFYDG\">40% DEET<\/a> ($5) or \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00HJTH06M\">20% picaridin<\/a> ($6). Actually, get both.\n<\/p>\n<\/li><\/ul><p>\nOf course, there is also the luxury option: buying a camper or an RV. Such vehicles are frowned upon\nin many cities, but if you are in a rural area, if it's within your budget, and if you can see yourself\nusing it regularly, go for it - and don't look back.\n<\/p>\n<h3>4.7. Hygiene, health, environmental protection [<a href=\"#4.7\">link<\/a>]<\/h3>\n<p>\nHumans are pretty fragile. In many of the more serious scenarios discussed in this guide, your\nsurvival or well-being may be critically dependent on the ability to deal with incidental medical\nemergencies, to curb the spread of communicable diseases, or to maintain sanitary conditions at home.\n<\/p>\n<p>\nOf course, there are situations where prompt medical attention is simply a necessity; for\nexample, although it may be theoretically possible for an untrained enthusiast equipped with\nan anatomy handbook to perform appendectomy, the odds of the patient surviving are pretty damn\nlow. That said, outside the domain of major surgery, the outlook is not necessarily as grim -\nso even when professional help is not available right away, not all hope may be lost.\n<\/p>\n<h4>4.7.1. Health &amp; hygiene topic #1: Basic cleaning supplies [<a href=\"#4.7.1\">link<\/a>]<\/h4>\n<p>\nBefore we dive into heady stuff, let's talk good housekeeping!\n<\/p>\n<p>\nIt pays to be prepared for nasty spills or sanitation emergencies. In addition to some of the tools \ndiscussed in earlier chapters, a good starting point is a large box of 13 gallon \n<a href=\"http:\/\/smile.amazon.com\/dp\/B001UB44SM\/\">trash bags<\/a> ($10-$30),\nanother box of thick, 42 gallon <a href=\"http:\/\/smile.amazon.com\/dp\/B001BBHILG\">contractor bags<\/a> ($30),\na set of <a href=\"http:\/\/smile.amazon.com\/dp\/B004Q6XN6C\">disposable vinyl gloves<\/a> ($4) plus\na pair of tough <a href=\"http:\/\/smile.amazon.com\/dp\/B004A9KI7W\">PVC or nitrile ones<\/a> ($11),\nsome rubbing alcohol ($6), and a bottle of regular laundry bleach ($3). Rubbing alcohol is an excellent\nsolvent and a rapid-evaporating disinfectant; bleach is a potent, long-lasting biocide and\na great odor neutralizer. Oh - as noted earlier, an\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0085S2DEU\">absorbent sock-style barrier<\/a> may\nbe useful for containing particularly nasty spills, too.\n<\/p>\n<p>\nWhen it comes to comfort and personal hygiene, I would also suggest stockpiling a carton of toilet paper, a\npack of <a href=\"http:\/\/smile.amazon.com\/dp\/B007N4D4ZO\">old-school soap bars<\/a>, a bottle of\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00GI0O06G\">no-rinse body wash<\/a> and\n<a href=\"http:\/\/smile.amazon.com\/dp\/B000EHNE0M\">shampoo<\/a> (to conserve water), some antiperspirant,\ntoothpaste and a toothbrush, nail clippers or scissors, and several disposable razors. At home, such items\ncan come handy only during prolonged emergencies; but as a part of your in-car kit, they are invaluable if you\nare ever stuck somewhere or forget to take your toiletries with you on a camping trip.\n<\/p>\n<p>\nAs mentioned earlier, water outages can create an unpleasant problem with the disposal of human\nwaste. While a bucket with a trash bag can be used as an impromptu toilet, to keep the conditions\nsanitary, some form of waste treatment is a must. A cheap option is pouring a layer of\nclumping (bentonite-based) kitty litter over it after every use; other budget solutions include\ncement and lime. More expensive choices are\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00H9R0JLM\">gelling agents<\/a> and\n<a href=\"http:\/\/smile.amazon.com\/dp\/B009KP9CY6\">RV waste digestants<\/a>.\n<\/p>\n<p>\nLaundry is another (if slightly less pressing) problem that many preppers may have to reckon with.\nWell-chosen antiperspirants and <a href=\"http:\/\/smile.amazon.com\/dp\/B000UDDT9G\/\">BZK-based\nantimicrobal sprays<\/a> do wonders to control bodily odors and extend the life of undergarments.\nBeyond that, careful hand-washing and rinsing techniques help minimize waste - but when there is\nno running water, doing laundry is still going to be a rare luxury for most.\n<\/p>\n<h4>4.7.2. Health &amp; hygiene topic #2: Visibility on the road [<a href=\"#4.7.2\">link<\/a>]<\/h4>\n<p>\nIn addition to assorted tools and hygiene supplies, your car kit should probably include\na <a href=\"http:\/\/smile.amazon.com\/dp\/B0030A85LS\">high-visibility vest<\/a> ($7); if your vehicle gives up\nlife, it's better to be visible when trying to revive it or walking to get help. I'm less convinced about\nthe benefits of carrying flares or flare guns; while they can be useful in serious emergencies, they\nalso pose some fire risk.\n<\/p>\n<h4>4.7.3. Health &amp; hygiene topic #3: Medical education [<a href=\"#4.7.3\">link<\/a>]<\/h4>\n<p>\nBefore you even think about self-medicating or treating wounds, you should get a reasonably systematic\nunderstanding of emergency medicine. I recommend getting\n<a href=\"http:\/\/smile.amazon.com\/dp\/0762780703\">\"Wilderness Medicine: Beyond First Aid\"<\/a> ($10): it\nis accessible, focuses on situations where diagnostic and treatment facilities are limited, and\ngoes well beyond the basics. Just as importantly, it avoids weird spiritual, homeopathic, or\nnaturopathic claims that often creep up in prepper books.\n<\/p>\n<p>\nTwo other publications I can recommend are\n<a href=\"http:\/\/smile.amazon.com\/dp\/0942364155\">\"Where There is No Doctor\"<\/a> and\n<a href=\"http:\/\/smile.amazon.com\/dp\/0942364058\">\"Where There is No Dentist\"<\/a> by the \nHesperian Foundation ($17 a pop for a paperback, but also available online for free).\nThey pay less attention to contemporary meds or nuanced emergency procedures,\nand spend more time on holistic, community-oriented care for practitioners in some of the most\nimpoverished regions of the world. This probably makes the publications worthwhile for\nhardcore preppers who worry about widespread, long-term cataclysms.\n<\/p>\n<p>\nNow, even after reading all these, don't get too cocky: there are good reasons why it takes about\n10 years to become a doctor, and why it involves not just reading a book or two, but also dissecting\ncadavers and watching other medics at work. When trying to render medical aid, not understanding the limits\nof your knowledge can literally kill.\n<\/p>\n<h4>4.7.4. Health &amp; hygiene topic #4: Common meds [<a href=\"#4.7.4\">link<\/a>]<\/h4>\n<p>\nWhen living in squalor conditions and running short on supplies, even seemingly prosaic medical conditions can\nbecome life-threatening. For example, in less developed countries, otherwise non-lethal diarrheal\ndiseases cause almost 2.5 million deaths every year. The reason is simple: without proper care,\nthe disease makes it easy for the victims to \nget terminally dehydrated or succumb to severe electrolyte imbalance.\n<\/p>\n<p>\nAnd it's not all about dying, too: a nasty toothache or a debilitating allergy can make it very\ndifficult to stay productive and alert. With all that in mind, my list of essential and easily available\nprepper medicines includes\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00JF9G1N2\">ibuprofen<\/a> ($10, pain relief),\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00G9E1GYA\">cetirizine<\/a> ($15, allergy management), \n<a href=\"https:\/\/www.payless-petproducts.com\/amox500.html\">amoxicillin<\/a> ($15, broad-spectrum antibiotic),\n<a href=\"http:\/\/smile.amazon.com\/dp\/B002RL8FJS\">loperamide<\/a> ($10, anti-diarrheal), \n<a href=\"http:\/\/smile.amazon.com\/dp\/B004FJFF2A\">meclizine<\/a> ($7, prevents vomiting), \n<a href=\"http:\/\/smile.amazon.com\/dp\/B006HSREVC\">miconazole nitrate<\/a> ($10, treats fungal skin infections), \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00D2PAXAM\">bacitracin ointment<\/a> ($5, bacterial skin infections),\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00O133T00\">topical lidocaine<\/a> ($20, anesthetic),\nand <a href=\"http:\/\/smile.amazon.com\/dp\/B001KYA8KW\">hydrocortisone cream<\/a> ($6, anti-itch). \nFor disinfecting your hands and cleaning wounds,\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0057Z24WK\">benzalkonium chloride wipes<\/a> ($4) can work pretty well;\nfor burns, many people swear by\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00HZTLZDU\">hydrogel dressings<\/a> ($5) or\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00G9XR186\">hydrogel creams<\/a> ($13), too.\nFinally, for treating severe dehydration, try\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00FOTDVS0\">oral rehydration packs<\/a> ($30).\n<\/p>\n<p>\nManufacturers' expiration dates on all meds are very conservative, but even the US\ngovernment\n<a href=\"http:\/\/www.fda.gov\/EmergencyPreparedness\/Counterterrorism\/MedicalCountermeasures\/MCMLegalRegulatoryandPolicyFramework\/ucm411446.htm\">intentionally\nignores them<\/a> for its stockpiles of shelf-stable drugs. As far as I can tell, when stored in a\nfridge, all of these products should be good for 5+ years.\n<\/p>\n<p>\nOf course, if you need any prescription meds to survive, try to get a reserve. Except for\nthe case of narcotic painkillers, most doctors should be quite willing to help. In contrast to\nthe shelf-stable substances discussed earlier on, be very careful with expiration dates if your\nstockpile includes any easily perishable drugs (say, insulin).\n<\/p>\n<p>\nThe aforementioned short list aside, there are other substances that may treat conditions such as\nanaphylaxis (epinephrine), or diseases such as tuberculosis and leprosy (isoniazid),\nuncomplicated appendicitis (levofloxacin), cholera and anthrax (doxycycline),\nmalaria (artemisinin), typhoid fever, skin infections, and MRSA (TMP\/SMX), systemic\nfungal infections (fluconazole), intestinal parasites (albendazole), and so on.\nIn many places, you could legitimately stockpile them without prescription, and some\nprepper guides will encourage you to do so; but if you are truly concerned about such\nemergencies, be very aware of the difficulty of properly diagnosing unfamiliar diseases,\nand the serious health problems that many of these products may cause.\n<\/p>\n<p>\nA handful of OTC dietary supplements may be useful for treating some chronic conditions\nin situations where prescription drugs are not available - but if you want to\nlearn more, be prepared to wade through a sea of low-quality research and outright\nquackery. Preliminary but somewhat plausible\nfindings include the apparent antidepressant properties of saffron and fish oil,\nthe beneficial effects of curcumin on some types of chronic pain, or the\nutility of berberine and salacia reticulata in managing type 2 diabetes. Again, tread carefully;\n<i><a href=\"http:\/\/examine.com\/supplements\/\">examine.com<\/a><\/i> is a good starting point for getting the\ndata behind some of the claims.\n<\/p>\n<p>\nAh, one more thing: for you car kit, I recommend getting some ibuprofen and\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0011865IQ\">caffeine pills<\/a> ($8). Sure,\nif you are getting sleepy, you should pull over and get some rest - but if you really can't, \ncaffeine can help you stay alert. Note that it's possible to overdose; the symptoms are\ntypically just very unpleasant, but in rare cases, can be life-threatening.\n<\/p>\n<p>\n(For folks who consume coffee or caffeinated beverages daily, and who have\nnear-complete tolerance to caffeine, \n<a href=\"https:\/\/nootropicscity.com\/product\/adrafinil-capsules\/\">adrafinil pills<\/a>\ncan be a viable emergency choice. Adrafinil is a powerful wakefulness-promoting agent\nwith no stimulant properties, and is not a regulated substance in the US and many other\nparts of the world. Note that daily long-term use carries some risk of liver toxicity\nand other side effects. Do your homework.)\n<\/p>\n<h4>4.7.5. Health &amp; hygiene topic #5: Dental emergencies [<a href=\"#4.7.5\">link<\/a>]<\/h4>\n<p>\nBlame modern diets, blame our longevity, or blame the mistakes of mother nature - but the bottom line is that for most humans,\ndental problems are a question of \"when\", not \"if\". And when excruciating pain strikes at an inopportune time, it's really no\nlaughing matter: in absence of adequate medical care, tooth problems have been known to push some people to the verge of suicide.\n<\/p>\n<p>\nUnfortunately for preppers, the management of serious pain in an emergency situation is tricky: virtually all the\npotent painkillers have narcotic properties and are very illegal to buy or possess without a prescription. Fixing the\nunderlying tooth problem can be similarly elusive: you need expert knowledge and a collection of expensive and bulky\ntools: high-speed drills, suction units, and so forth. In the end, the best preparedness strategy is just prevention:\ntake good care of your mouth and stop by for a routine checkup at least once a year.\n<\/p>\n<p>\nOf course, while this approach reduces the odds of being blindsided by a painful problem, it does not make the risk go away:\na chipped tooth or a painful abscess can strike with virtually no warning. If you can't see a dentist right away, OTC painkillers\ncan offer partial relief, but no amount of ibuprofen will let you forget about an exposed nerve. Topical\n<a href=\"https:\/\/smile.amazon.com\/dp\/B0042U5VUS\/\">benzocaine ointments<\/a> may work better in some situations, but they don't\nlast very long. Sometimes, swishing some cold water in your mouth, or sucking in air through a carefully positioned straw, can offer\ndecent relief. In the longer haul, amoxicillin can clear up many dental infections, while\n<a href=\"http:\/\/www.ebay.com\/itm\/Dharma-Research-Dental-Zinc-Oxide-Eugenol-Cement-42-5-Gram-Powder-30-ML-Liquid-\/111661278732\">zinc\noxide \/ eugenol cement<\/a> can be used for emergency repair of damaged teeth. There are some reports that repeated treatments of\ncavities with\n<a href=\"http:\/\/www.ebay.com\/itm\/100-Silver-Nitrate-Double-Dipped-6-Applicator-Sticks-100-pack-NEW-\/272316796901\">silver nitrate<\/a>\ncan be beneficial, too - but be aware that the substance is caustic and tends to semi-permanently stain skin (and anything else it\ncomes into contact with).\n<\/p>\n<h4>4.7.6. Health &amp; hygiene topic #6: Hemorrhage and wound management [<a href=\"#4.7.6\">link<\/a>]<\/h4>\n<p>\nSevere bleeding is one of the major causes of death following an injury. It's an emergency\non its own, and knowing how to manage it until help arrives can make all the difference between\nlife and death.\n<\/p>\n<p>\nA decent hemorrhage kit should probably include a generous amount of bandages,\na <a href=\"http:\/\/smile.amazon.com\/dp\/B004U4EB3E\">tourniquet<\/a> ($5-$15), \n<a href=\"https:\/\/www.rescue-essentials.com\/quikclot-combat-gauze-z-folded-military\/\">clotting gauze<\/a> ($40), and some duct tape. \nYou should read the manuals and consult an up-to-date first-aid guide, but the basic\nidea is to apply lots of pressure to any profusely bleeding wounds. This can be done\nwith bandages, clothing, duct tape, or even your elbow, knee, or hips. Clotting gauze or sponges,\nwhen pushed into the wound cavity, can help stop bleeding more quickly and stabilize the\nvictim. Tourniquets used to be frowned upon in the past, but when dealing with major\ntrauma to a limb, they sure beat bleeding out to death; it's just that they cause some\ntissue injury, and if kept on for too long, necrosis may set in and the limb may have to go.\n<\/p>\n<p>\nDon't just keep that stuff in your home; it's actually more important to have such a kit in your car (ideally in\nthe center console) and carry something equivalent when biking, hiking, climbing, hunting,\nor engaging into other injury-prone sports.\nThere are some <a href=\"http:\/\/smile.amazon.com\/dp\/B003BS2PW4\">lightweight ready-made kits<\/a>\n($19) that may do the trick, although making your own is always a better choice.\n<\/p>\n<p>\nIn addition to such immediately necessary supplies,\nsome prepper guides recommend purchasing sutures, along with tissue forceps and hemostatic clamps.\nSuch equipment may be useful for neatly closing major wounds in situations where bandages won't do -\nbut suturing correctly requires a fair amount of practice and know-how.\nFor gnarly cuts, \n<a href=\"http:\/\/www.ebay.com\/itm\/Skin-Stapler-Preloaded-Sterile-Surgical-35-Wide-Staples-Each-35W-\/381692826001?epid=1279818616&amp;hash=item58deaa5191:g:JvQAAOSwZ1BXfABd\">skin staplers<\/a> or\n<a href=\"http:\/\/smile.amazon.com\/dp\/B004WFXCSQ\">skin closure strips<\/a> +\n<a href=\"http:\/\/smile.amazon.com\/dp\/B002GXAZO0\">benzoin swabs<\/a>\ntend to involve less hassle - and are harder to mess up.\n<\/p>\n<p>\nFor longer-term wound management, in addition to hydrogels and benzalkonium chloride disinfectants\ndiscussed earlier on, there is some utility to\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00N54ZEHM\">non-adherent dressings<\/a> ($30) and\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0070P2Q64\">abdominal pads<\/a> ($7). In a pinch,\nwhen dealing with high-drainage wounds, sanitary pads should also do the trick.\n<\/p>\n<h4>4.7.7. Health &amp; hygiene topic #7: Broken or crushed limbs [<a href=\"#4.7.7\">link<\/a>]<\/h4>\n<p>\nYour emergency medicine book will go into more details about setting bones, applying splints,\nor even doing field amputations with a knife and a saw. But even just to deal with a sprained ankle, a\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0050R7ZJC\">folding cane<\/a> ($14) may be good to have somewhere in your\nstash. Beyond that, bandages are useful for improvising splints; in areas where improvisation may be\ndifficult - say, in the desert or up in the mountains - \n<a href=\"http:\/\/smile.amazon.com\/dp\/B001J5H92C\">portable folding splints<\/a> ($10)\ncan be handy, too.\n<\/p>\n<h4>4.7.8. Health &amp; hygiene topic #8: Respiratory and environmental protection [<a href=\"#4.7.8\">link<\/a>]<\/h4>\n<p>\nWhen it comes to transmittable diseases, your best bet is avoiding exposure: if there's something really\nnasty making rounds in your community, stay home - or at the very least, avoid public transport and other\ncrowds. With that in mind, \n<a href=\"http:\/\/smile.amazon.com\/dp\/B0002YKBV2\">N95 masks<\/a> ($2 a piece) probably provide adequate protection against most\nairborne diseases. The other major transmission vector is hand contact, so don't touch other people,\navoid public-use surfaces, wear gloves, and resist the instinct to touch your face without first washing\nor disinfecting your hands. We subconsciously touch our faces a lot more frequently than we suspect.\n<\/p>\n<p>\nIf you worry about releases from chemical plants or overturned ammonia tankers,\n<a href=\"http:\/\/smile.amazon.com\/dp\/B009POHLRC\">3M multi-gas cartridges<\/a> ($17) and\n<a href=\"http:\/\/smile.amazon.com\/dp\/B007JZ1MK6\">half facepieces<\/a> ($12) may offer robust protection when sized\nand fitted properly. That said, in most cases, it's more important to develop a plan for sealing your home;\nwalk around and take note of any crawl space inlets, bathroom and kitchen exhausts, chimneys,\nfireplaces, and any other gaps. In an emergency, you can cover them with trash bags and duct tape.\n<\/p>\n<p>\nAlthough we are entering the realm of extremely unlikely events, if you genuinely worry about\nencountering an overturned chemical tanker while driving down the highway,\n<a href=\"http:\/\/www.ebay.com\/itm\/281211595242\">3M 5512 escape respirators<\/a> ($16) offer decent short-term\nprotection against many threats. In addition to low price, their major advantage is their small size; you\ncould conceivably have one for every occupant, and just store that gear in the vehicle.\n<\/p>\n<p>\nOf course, some extreme preppers will settle for nothing less than a military-grade gas mask rated for\nchemical, biological, and nuclear warfare. While it is true that such masks offer much better protection\nagainst nerve agents and similar extremely harmful substances, it's not a likely concern in most parts\nof the world - and either way, it's doubtful that you would have enough time to suit up once the symptoms\nkick in.\n<\/p>\n<p>\nOn a perhaps more realistic note: since many of the common and harmful industrial gases are highly\nwater-soluble, breathing through a wet towel or other moist piece of cloth, or draping it to form a\ntent over an infant's car seat or a stroller, might offer some very short-term protection when\nall other options fail.\n<\/p>\n<h4>4.7.9. Health &amp; hygiene topic #9: Vector control [<a href=\"#4.7.9\">link<\/a>]<\/h4>\n<p>\nIn some parts of the world, mosquitoes, flies, and other biting insects are major vectors for extremely\nserious diseases, such as malaria, African sleeping sickness, dengue fever, or lymphatic filariasis. The same is not\na grave concern in more temperate climates; although malaria used to be a problem in some of the southern states,\nit has been eradicated by 1951. Today, some of the most serious incidents in the US and in Europe are the occassional\ncases of Lyme borreliosis or the West Nile virus.\n<\/p>\n<p>\nStill, if you are worried about the situation changing for the worse, repellents such as DEET and picaridin can\nprovide the first line of defense. Beyond that, more radical solutions may include\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00RB1YM5U\">electric bug zappers<\/a>\n(especially when coupled with mosquito attractants, such as octenol or lactic acid),\n<a href=\"http:\/\/smile.amazon.com\/dp\/B001ANQVZE\">permethrin<\/a> or\n<a href=\"http:\/\/smile.amazon.com\/dp\/B0035HA0AK\">pyrethrin<\/a> insecticide sprays\n(applied to clothes or to the perimeter), <a href=\"http:\/\/www.amazon.com\/dp\/B000J0MP18\/\">mesh jackets<\/a>, \nwindow screens,\nand <a href=\"http:\/\/smile.amazon.com\/dp\/B014NT0QUW\">bed nets<\/a>. For crawling insects, borax and\ndiatomaceous earth can act as a deadly barrier, too.\n<\/p>\n<p>\nWhile serious zoonotic diseases can be also spread by birds, rodents, or even cats and dogs, such vectors\nare more easily controlled or avoided by maintaining sanitary conditions at home; in extreme cases,\ntraps or poisons such as bromadiolone may do the trick, but unless you are living in a place prone\nto infestations, I wouldn't worry too much.\n<\/p>\n<h4>4.7.10. Health &amp; hygiene topic #10: Dealing with nuclear fallout [<a href=\"#4.7.10\">link<\/a>]<\/h4>\n<p>\nThere are quite a few pop culture myths surrounding the dangers of nuclear incidents, contributing to a defeatist attitude\namong even some of the most hardened preppers. But in reality, such events are a lot more survivable than portrayed in\nfiction - and perhaps more importantly, the world that awaits the survivors would not necessarily be all that bleak. A good way to explore this topic is a book titled\n<a href=\"http:\/\/oism.org\/nwss\/nwss.pdf\"><i>\"Nuclear War Survival Skills\"<\/i><\/a>. It sounds goofy, but it's been written\nby the folks who worked on the Manhattan Project, and is as close to scientific truth as you can get; plus, it is not copyrighted\nand can be downloaded for free.\n<\/p>\n<p>\nBut to summarize, let's start with the eponymous threat of nuclear war. A typical ICBM strike is likely to kill most people within a 1 to 10 mile\nradius of the explosion, with most perishing due to the blast wave and intense heat, not gamma rays. In fact, as witnessed in\nHiroshima and Nagasaki, when one sees a flash of light, ducking behind cover is quite likely to save their life. But more\nimportantly, even an \"all out\" nuclear exchange with another superpower would leave most of the United States unscathed. It would also\nnot turn the planet into a post-apocalyptic wasteland - at least not any more than the\nhundreds of nuclear tests already conducted in the twentieth century.\n<\/p>\n<p>\nThe fallout threat tends to be overblown, too. For one, air bursts, which are preferred because of their improved blast radius,\ndo not produce that much of it; far more tends to be released during nuclear power plant meltdowns or ground bursts.\nIn any case, whatever gets kicked up in the air can travel hundreds of miles before settling down. So, in the aftermath of\nan incident, getting indoors and sealing your home should be your first instinct. Basic respiratory protection can help, too.\n<\/p>\n<p>\nAt first, the falling dust will be extremely radioactive and even short-term exposure might be lethal - so your best bet\nwould be to seek shelter in the basement, or near the center of mass of any other building. This maximizes the mass between\nyou and the outside world, shielding you from the gamma rays produced by whatever happens to settle on the roof and the outside\nwalls. The density and volume of the shielding material matters more than anything else. It does not have to be lead; mattresses\nand bulky furniture should do.\n<\/p>\n<p>\nLuckily for the survivors, the highly radioactive isotopes present in the fallout are also very-short lived; the intensity of radiation\nwill likely drop ten-fold within 6-8 hours, and will decrease a hundred-fold within two days. If you wait a week or two, it\nshould be quite safe to venture out. Of course, it's still best to stay indoors for as long as possible, and when heading out,\nit's good to keep the trips short, to wear <a href=\"http:\/\/smile.amazon.com\/dp\/B00VKX6HA8\">disposable coveralls<\/a> ($11), and\nto take care not to track any residues into your home. But you don't need to lock yourself in an underground vault.\n<\/p>\n<p>\nThis brings us to an interesting question: if the radioactivity decays so quickly, why aren't people moving back to\nFukushima or Chernobyl? Well... spare for several small hotspots, the exclusion zones are safe to walk around,\nbut it can be more dangerous to drink contaminated water or eat local wildlife or crops.\nOur body may end up using some of the longer-lived radioisotopes as biological building blocks - which would expose us\nto low-grade radiation, at an <i>extremely<\/i> close range, for the rest of our lives. While it is usually not a death\nsentence, this phenomena is bound to produce a measurable spike in mortality across any sufficiently \nlarge population. It is more humane and more socially acceptable to keep people out.\n<\/p>\n<p>\nStill, this does not mean that the survivors of a nuclear war would have to choose between starvation and death. Water and food stored in closed\ncontainers will not become radioactive - people would just have to be mindful of the dust on top. Crops can be grown after removing \nseveral inches of topsoil, and most rivers, streams, and creeks become safe quite rapidly (shallow bodies of standing water are a\ndifferent animal). In other words, with basic precautions, it's quite possible to thrive in the aftermath of even the\nworst nuclear war. All it takes is some luck and a bit of knowledge; blast-proof bunkers are not a must.\n<\/p>\n<p>\nIn fact, one of the most significant health consequences of nuclear accidents is also very easy to manage: it's the release of\ncopious amounts of radioactive iodine, a short-lived substance that gets absorbed by the thyroid gland. To deal with this\nissue, people in the affected areas are typically offered potassium iodide pills; this temporarily saturates\nthyroid and prevents any further uptake of iodine for a couple of days or weeks. Such tablets are \n<a href=\"http:\/\/smile.amazon.com\/dp\/B00006NT3A\">available over-the-counter<\/a> and dirt cheap ($7), so\nit's not a bad idea to have some at hand.\n<\/p>\n<p>\nOther than that, there isn't much that can be done to limit the damage caused by serious exposure to radiation. Some animal\nstudies suggest that pyrroloquinoline quinone (PQQ), an OTC dietary supplement, can have fairly pronounced\nradioprotective benefits. While the evidence is very preliminary, the substance is believed to be pretty safe,\nso you can certainly \n<a href=\"http:\/\/smile.amazon.com\/dp\/B010MOGNUE\">grab some<\/a> just in case. Another potentially beneficial\nOTC products along the same lines\nare <a href=\"http:\/\/smile.amazon.com\/dp\/B000UZXYZS\">diindolylmethane<\/a> (DIM) and\n<a href=\"https:\/\/smile.amazon.com\/dp\/B00KT3H13C\">n-acetylcysteine<\/a> (NAC).\n<\/p>\n<p>\nWith all this out of the way, let's get back to a gadget mentioned a bit earlier in this guide: Geiger counters.\nUnless you're an emergency responder, you may not really need one. That said, such a device could conceivably\nhelp you stay informed and keep your family and friends at ease - and I don't mean just the remote possibility of\na nuclear war. Consider all the bogus rumors of contaminated\nwater and food in the aftermath of Fukushima; a radiation meter could have put any such speculation to rest.\n<\/p>\n<p>\nOf course, to use a Geiger counter effectively, you need to know how to interpret the results. Although some\ncontroversy exists, the prevailing view is that the effects of radiation are cumulative in a fairly linear fashion;\nhere's what you can expect once you hit a particular dose:\n<\/p>\n<ul><li><p><b>100 mSv (millisieverts):<\/b> no immediate effects. A marginal but measurable increase in\nthe likelihood of developing cancer later in life (+0.5%). This dose was \nreceived by a handful of residents during the Fukushima disaster in 2011.<\/p>\n<\/li><li><p><b>1 Sv:<\/b> mild radiation sickness and a substantial increase in lifetime\ncancer risk (around +6%). A dose that might have been\nreceived by some subset of residents near Chernobyl in 1986.<\/p>\n<\/li><li><p><b>2 Sv:<\/b> severe radiation sickness with vomiting and hair loss. Dangerous\nbut with good chances of survival. A pretty substantial risk of cancer down the line (+12%).<\/p>\n<\/li><li><p><b>5 Sv:<\/b> typically lethal, although there is a chance of recovery with\nproper medical treatment. Survivors are highly likely to develop cancer later on.<\/p>\n<\/li><li><p><b>10+ Sv:<\/b> nearly always fatal. A dose received by some of the first responders\nat Chernobyl.<\/p>\n<\/li><\/ul><p>\nSome Geiger counters can keep track of the cumulative dose for you, but most will\nsimply display the rate at which you are getting exposed - typically in microsieverts\nper hour (\u00b5Sv\/h) or microroentgens per hour (\u00b5R\/h; for gamma radiation,\n1 R\/h \u2248 1\/100 Sv\/h).\nTo figure out what the reading means, you need to do the math: for example, at 600 \u00b5Sv\/h,\nyou will hit 100 mSv within a week, and 1 Sv within about two months. (The usual background\nrate from natural sources hovers around 0.1-0.2 \u00b5Sv\/h.)\n<\/p>\n<p>\nFor folks interested in getting a nice, compact Geiger counter,\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00YQARZ5S\">Radex One<\/a> ($120) is pretty hard to beat;\nit is tiny, inexpensive, and can be hooked up to a PC to continuously monitor the environment (and send e-mail or SMS alerts). \nThe one caveat is that similarly to many other low-cost units,\nthis device maxes out at 1 mSv\/h - enough to know that something is very wrong, but not enough to tell if\nyou're going to receive a life-threatening dose in an hour or somewhere within the next six months. In other words, some of the more hardcore\npreppers may want to invest in a more capable unit, such as ADM-300 (which goes all the way to 100 Sv\/h) or RAD-60R (3 Sv\/h).\nDecommissioned military and civilian devices in excellent working condition can be found on eBay for around $200.\n<\/p>\n<h3>4.8. Self-defense and personal security [<a href=\"#4.8\">link<\/a>]<\/h3>\n<p>\nAs discussed earlier in this guide, we face surprisingly high lifetime odds of becoming victims of\nburglary, assault, or other major crime. To deal with this danger, the guide promotes a handful of passive\nrisk avoidance and loss minimization strategies, with the bulk of this advice found in section 3.5.\nBut it would be dishonest to claim that such methods will always shield us from harm - so in this\nchapter, let's have a look at some of the tools that serve as the last line of defense when \nall other approaches fail.\n<\/p>\n<h4>4.8.1. Defense topic #1: Protecting your personal property [<a href=\"#4.8.1\">link<\/a>]<\/h4>\n<p>\nIt is fiendishly difficult to safeguard your belongings when you're not home. When dealing\nwith opportunistic burglaries, a heavy, bolted-down safe, ideally ordered directly from a reputable\nmanufacturer, is probably your best defensive tool. Against sophisticated adversaries on a\ntargeted job, almost all bets are off; in such cases, \noperational security (section 3.5.3) is more important than any amount of high-tech gear.\n<\/p>\n<p>\nNow, when asked about the best way to make a residence burglar-proof, most people would\nprobably mention getting an alarm system. But alarm systems are fairly weak deterrents\nagainst theft; most statistics suggest that they reduce the likelihood of a break-in \nby around 50%. So, do the math: take the costs of installing an alarm system (probably around\n$2,000 for a comprehensive solution), plus the ongoing monitoring fees (easily $200-$500\/year),\nand then contrast these numbers with the likely loss in case of one or two break-ins over the\nnext several decades. Keep in mind that even if the numbers are favorable, a high-quality safe\n($500+) may still be a more cost-effective approach.\n<\/p>\n<p>\nAlarm systems aside, cameras are another popular security tool. They do relatively little to\ndeter theft, but can document all sorts of problematic encounters - and in the event of a burglary,\nperhaps improve the odds of recovering stolen goods.\nDecent wi-fi cameras start at $100 a piece; many models can record to a local SD card, although\nhaving a centralized DVR unit ($200+), ideally stowed away in an inconspicuous place, will\nmake the system more robust.\n<\/p>\n<p>\nSome preppers advocate \"hardening\" the perimeter of your home. The returns on this investment will\nvary; for example, high-security locks and reinforced doors may be worthwhile in high-rise\napartment buildings, where the front door may be the only way in. For single-family homes, the\nburglars will probably not bother with the locks at all; forcing open a bathroom window takes\nmuch less work. Tall fences and locked backyard gates can help, although their benefits are\nlimited in rural areas or in shady neighborhoods.\n<\/p>\n<p>\nThe other tools worth mentioning here may be remote motion detectors, such as\n<a href=\"http:\/\/smile.amazon.com\/dp\/B002ISZD2C\">passive IR monitors<\/a> or \n<a href=\"http:\/\/smile.amazon.com\/dp\/B002WV19DM\">beam sensors<\/a>. They won't do squat\nwhen you're not on premises, but when you're home, they can give you an advance warning\nabout unwanted visitors. It's a remote concern, but the sensors are relatively cheap, so\njust do what feels right.\n<\/p>\n<p>\nAs for cars: there is no hope. Don't leave anything of substantial value in the vehicle, and\nif the car itself is expensive, have it insured against theft (setting your deductible to $1,000\nor more  keep the premiums low). Avoid tempting the thieves\nin any way: countless car windows have been smashed over a $5 bill and some coins left\nin the cup holder. Put spare change somewhere else.\n<\/p>\n<h4>4.8.2. Defense topic #2: Limiting liability for car accidents [<a href=\"#4.8.2\">link<\/a>]<\/h4>\n<p>\nVehicular accidents are depressingly common; while defensive driving can limit your\nrisk, the possibility of injuring another person or causing property damage never really\ngoes away. When you are involved in a car wreck in unclear circumstances, or when your\nstatements do not match the words of another driver, video evidence may be the best way\nto escape criminal charges or to resolve civil claims.\n<\/p>\n<p>\nIt used to be that dashcams were prohibitively expensive; but today, the prices\nstart at $50, so it makes sense to give the devices a try. I can recommend\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00X528FNE\">Rexing V1<\/a> ($100), but there are\ncountless other options to choose from. The bottom line is, if you own a car, it's\nprobably the most affordable and meaningful liability insurance policy you can get.\n<\/p>\n<h4>4.8.3. Defense topic #3: Fighting for your life [<a href=\"#4.8.3\">link<\/a>]<\/h4>\n<p>\nIn theory, a person's natural right to self-defense is broadly recognized in much of the\nwestern world - but in practice, different societies look at it in very different ways. Today,\nin much of Europe, the very notion that one private citizen could lawfully harm another human\nbeing is met with suspicion and distaste; the control over life and death is more willingly\ndelegated to the agents of the state.\n<\/p>\n<p>\nThese differences manifest in how the regulators around the globe approach anything from\nknives to pepper spray - but of course, no topic is more contentious than firearms. The debate about\nthe social benefits and costs of gun ownership is \n<a href=\"https:\/\/lcamtuf.blogspot.com\/2015\/10\/subjective-explainer-gun-debate-in-us.html\">hopelessly\npolarizing and clouded by emotion<\/a>; I'm certain that roughly half of the folks reading this document\nhave a very visceral, negative reaction to the very idea that a private person should be allowed\nto carry a gun - and I do not honestly expect to change their minds.\n<\/p>\n<p>\nThat said, when we look beyond the dogma, the underlying facts paint an\nan <a href=\"https:\/\/lcamtuf.blogspot.com\/2015\/06\/a-bit-more-on-firearms-in-us.html\">incredibly\nnuanced picture<\/a> of the right to bear arms - putting into question many of the deeply-held and seemingly\ncommon-sense beliefs. For example, despite the striking ubiquity of legally owned\nfirearms in the United States - about one per every resident - the country actually enjoys\nlower per-capita rates of suicide, robbery, assault, or rape, compared to some of its esteemed European\npeers. This simple fact deals a serious blow to the creed that more guns translate to more crime.\n<\/p>\n<p>\nOf course, this observation ignores one important fact: compared to EU countries, the United States\nsuffers from a markedly elevated (but rapidly falling) rate of homicide. About two thirds of them are committed with\nguns - but lest we jump to conclusions, the non-firearm-related murder rate alone puts America well\nahead of most of Europe, suggesting that the cause may have more to do with societal differences than with\nthe availability of a particular tool. A finding that supports this theory is the fact that upward of 80%\nof US gun homicides \n<a href=\"http:\/\/www.cdc.gov\/nchs\/data\/nvsr\/nvsr61\/nvsr61_06.pdf\">trace back to gang activity and drug trade<\/a>,\noften within the disadvantaged or impoverished strata of the society that are far less prominent on the old continent.\nAnother telling observation is that comparisons of overall murder rates across US states or across EU countries\nwith vastly different firearms ownership profiles \n<a href=\"http:\/\/www.objectobot.com\/wp-content\/uploads\/2012\/12\/us_rate.jpg\">don't reveal a convincing\ncorrelation<\/a> between the two variables - something you would expect to see if legally owned\nguns had a causative relationship with violent crime.\n<\/p>\n<p>\nThere is also a more utilitarian way of looking at it. From an individualistic, survival-focused\npoint of view, the social costs (real or imagined) are basically moot: if you live in a place where guns\nare readily available to criminals, it's hard to think of a violent confrontation where not owning a\nfirearm would put you at an advantage. There is a body of fairly compelling research showing that\ndefensive gun uses happen in the US at a rate of somewhere \n<a href=\"http:\/\/www.nap.edu\/read\/18319\/chapter\/1\">between 500,000 and 1,500,000<\/a> times a year,\nwith almost all confrontations resolved without firing a single shot. The inherent dangers of owning\na firearm are often overstated by gun control advocates, too: unintentional injury or death due to\nhaving a gun at home is\n<a href=\"http:\/\/www.nssf.org\/pdf\/research\/iir_injurystatistics2013.pdf\">surprisingly rare<\/a>.\n<\/p>\n<p>\nAll in all, it's OK to reject armed self-defense (or shun guns in particular) on religious or moral\ngrounds - but doing so is probably not a particularly rational decision within the scope of this\nguide. From a rational standpoint, you should always pick the tools that are best suited for the scenarios\nyou anticipate (provided that the state allows you to). Of course, a firearm is not always the\nanswer, so let's take a broader look at some of the most popular options for shooing away looters or\ndefending yourself:\n<\/p>\n<ul><li>\n<p><b>Bare hands.<\/b>\nNo deterrent effect, but surprisingly effective when a confrontation can't be avoided -\n especially when facing a single assailant.\nReserved for people who are physically fit and willing to invest a fair amount of time into training. One of the most pragmatic\nand widely-taught schools is Krav Maga, and there's certainly no harm in checking it out.\n<\/p>\n<\/li><li>\n<p><b>Pepper spray.<\/b>\nAn excellent, temporarily incapacitating weapon - very difficult to resist and capable of buying you just\nenough time to escape. Works quickly and reliably at distances up to perhaps 10 feet; can also stop some animal\nattacks. Usually not heavily regulated, making it easy to obtain and carry even in places\nthat frown upon other forms of armed self-defense (but check the laws). Pepper spray becomes less effective in\nstrong wind; there is also some risk of blowback, but this is mitigated in narrow-stream products,\nsuch as <a href=\"http:\/\/smile.amazon.com\/dp\/B004NKSPR8\">Sabre Pepper Gel<\/a> ($18).\n<\/p>\n<\/li><li>\n<p><b>Knives.<\/b>\nVery lethal and dangerous at close quarters, but only provided that you have the element of\nsurprise on your side. They require some degree of physical fitness and training to use well.\nTheir value is diminished when facing multiple assailants or dealing with a gun-wielding\nindividual: even if you stab them, you are probably still gonna get shot. In a handful of places,\ncarrying a knife may be illegal or subject to somewhat confusing\nrestrictions, so perform due diligence if you want to take this route.\n<\/p>\n<\/li><li>\n<p><b>Stun guns.<\/b>\nA nice idea, in theory. Unfortunately, most products require direct body contact with the attacker,\nare easily foiled by clothing, and may be less effective against people who are well-built,\nunusually agitated, or just high on drugs. While it's a popular self-defense choice, I can't honestly\nrecommend it over pepper spray.\n<\/p>\n<p>\n(Projectile-type stun guns, such as Taser C2, are far better and work at a distance - but\nyou need to have perfect aim on first try. They also cost quite a bit.)\n<\/p>\n<\/li><li>\n<p><b>Firearms.<\/b>\nA very effective and supremely intimidating weapon, with lethality ranging from\n20% for handguns to 80% for shotguns; for a novice user, the effective range is somewhere\nbetween 10 and 100 yards. Guns are heavily regulated in much of the world, but widely\navailable in the US - although there are several states or municipalities that make it\nvery difficult to get a permit unless you are\n<a href=\"http:\/\/www.laweekly.com\/news\/sheriff-lee-baca-and-the-gun-gift-connection-2612907\">a celebrity\nor a prominent donor<\/a>.\n<\/p>\n<p>\nEven if your local government does not issue carry permits or erects other bureaucratic roadblocks, \nhaving a firearm in your home can provide\na very effective deterrent in case of home intrusions. That said, a gun is also a responsibility:\nsimilarly to a chainsaw, you are either serious about following the safety rules, or somebody may get\nhurt.\n<\/p>\n<\/li><\/ul><p>\nInterestingly, the legal bar for claiming self-defense is typically no different whether you are\nusing a less-lethal weapon or lethal force. But of course, the legal and psychological consequences of\nbeing wrong can be far more severe if you kill a person, versus just making their eyes itch. There are\nno easy answers, so do some soul-searching first. If you can't imagine killing another person to protect\nyour family - and living with the consequences - don't get a knife or a gun.\n<\/p>\n<h4>4.8.4. Defense topic #4: Understanding firearms [<a href=\"#4.8.4\">link<\/a>]<\/h4>\n<p>\nIf you are contemplating getting a gun for home defense or for more outlandish survival scenarios,\nthe first choice you will face is between a handgun, a shotgun, and a rifle. Here's what you need to know:\n<\/p>\n<ul><li> <p><b>Handguns.<\/b>\nThis category encompasses a wide selection of small, lower-powered firearms that can be easily carried\nwithout attracting attention. Most have a fairly modest stopping power, so-so ballistics,\nand require quite a bit of\npractice to accurately hit anything more than 10 yards away. A telling statistic is that in shootouts,\nthe police have a hit rate somewhere between 10% and 30%; contrary to what some gun control supporters\nclaim, an  average policeman does not get that much practice, and probably trains less than your typical\ngun enthusiast - but these numbers are still something to keep in mind.\n<\/p>\n<p>\nHandguns are typically regulated more heavily than other firearms, in part because they account for\nthe overwhelming majority of all gun crime. If you can get one, your basic choice is between:\n<\/p>\n<ul><li><p><b>Semi-automatic pistols.<\/b> Typically capable of firing\nsomewhere around 6-17 rounds from a removable magazine; reloading is very fast, provided that you carry\nanother magazine with you. Their user interface is relatively complex, and some knowledge is needed\nto deal with potential misfires, jams, or to avoid negligent discharges - although the probability of\nany such issues is generally very low. (It's wise to avoid discount ammo and magazines.)\n<\/p>\n<\/li><li><p><b>Revolvers.<\/b> A more ancient type of a repeating firearm, usually holding 5-8 rounds in\na non-removable cylinder. They are functionally simple, very dependable, and pretty accurate - but\ntake more time to reload if it ever comes to that.\n<\/p>\n<\/li><\/ul><p>\nYou may be tempted to go for the most lightweight and highest-powered handgun you can find, but you\nwould have to cope with punishing recoil and potentially blinding muzzle flash, so it's not always a\ngood call. In home defense situations, 9mm pistols and .38 Special+P revolvers are probably\nthe sweet spot. There are countless models to choose from, but the bottom line is that you can't go particularly\nwrong with Glock, Beretta, SIG Sauer, Ruger, Smith &amp; Wesson, CZ, Heckler &amp; Koch, or Springfield Armory.\n<\/p>\n<p>\nIf you need some inspiration, perhaps the most iconic and time-tested semi-auto design is the 1911; it was\noriginally a pretty bulky and heavy all-steel firearm, but many manufacturers now make\n<a href=\"http:\/\/www.springfield-armory.com\/products\/1911-emp-9-mm\/\">lighter, scaled-down variants<\/a> that are OK for\ncarrying around. Alternatives include a growing number of more lightweight, polymer-frame guns, such\nas <a href=\"https:\/\/us.glock.com\/products\/model\/g43\">Glock 43<\/a>,\n<a href=\"https:\/\/www.smith-wesson.com\/firearms\/mp-9-shield-0\">M&amp;P Shield<\/a>, or\n<a href=\"https:\/\/www.sigsauer.com\/store\/p365-nitron-micro-compact.html\">SIG P365<\/a>.\nFor revolvers, \n<a href=\"http:\/\/www.ruger.com\/products\/lcr\/overview.html\">Ruger LCR<\/a> and\n<a href=\"https:\/\/www.smith-wesson.com\/firearms\/model-642\">Smith &amp; Wesson 642<\/a> are probably two very reasonable\npicks to haul around. Some \n<a href=\"https:\/\/www.chiappafirearms.com\/p\/id\/302\/product\/Rhino-Revolver-40SAR.php\">very futuristic<\/a> if a tad more expensive\nrevolvers are made by Chiappa, too. Whatever you do, avoid Taurus.\n<\/p>\n<\/li><li> <p><b>Rifles.<\/b>\nLong, heavy guns, often with detachable magazines housing anywhere from 4 to 30 rounds. Rifles\nfire high-velocity projectiles capable of accurately striking distant targets - and even a complete novice\nshould be able to hit targets 25-50 yards away. With plenty of practice (and expensive optics), some\nrifles allow reliable hits at 1,000 yards or so.\n<\/p>\n<p>\nThe most popular category are semi-automatic rifles, including the scary-looking \n<a href=\"http:\/\/ruger.com\/products\/ar556\/models.html\">AR-15 clones<\/a>,\nthe much less villainous <a href=\"http:\/\/ruger.com\/products\/mini14RanchRifle\/models.html\">Ruger Mini-14<\/a>, and\na bunch of in-between choices, such as \n<a href=\"https:\/\/www.americanrifleman.org\/articles\/2015\/12\/12\/nra-gun-of-the-week-ares-defense-scr-rifle\/\">Ares SCR<\/a>.\nAll three are known for reliability and good accuracy. Semi-automatics typically\nfire relatively small but high-energy projectiles (e.g., .223 Remington \/ 5.56x45mm NATO); such projectiles\nare still suitable for home defense and hunting small to medium game.\n<\/p>\n<p>\nAnother popular pick are bolt-action rifles, including \n<a href=\"http:\/\/www.remington.com\/rifles\/bolt-action\/model-700\/model-700-cdl\">Remington 700<\/a>,\n<a href=\"http:\/\/www.winchesterguns.com\/products\/rifles\/model-70.html\">Winchester 70<\/a>,\n<a href=\"http:\/\/www.ruger.com\/products\/HawkeyeStandard\/models.html\">Ruger Hawkeye<\/a>,\nand \n<a href=\"http:\/\/www.ruger.com\/products\/precisionRifle\/models.html\">Ruger Precision Rifle<\/a>.\nAlthough there is a lot of variety, many are chambered for larger cartridges ideal for hunting big game \n(from .243 Winchester to .50 BMG) and are more suited for long-range shooting. For home defense,\noverpenetration becomes a significant concern.\n<\/p>\n<p>\nIn the US, long guns are subject to fewer restrictions than handguns, chiefly because of their negligible\nrole in street crime; that said, \"assault weapons\" - i.e., semi-automatic rifles with scary-looking cosmetic\nfeatures, such as barrel shrouds or forward grips - have been a subject of recurring moral panics\nand various state- or municipality-level restrictions and bans.\n<\/p>\n<p>\nPerhaps interestingly, there is a handful of rifles chambered for handgun ammunition. Canonical examples include\n<a href=\"http:\/\/www.ruger.com\/products\/77Series77357\/models.html\">Ruger 77 series<\/a>,\n<a href=\"https:\/\/www.henryrifles.com\/rifles\/big-boy\/\">Henry Big Boy<\/a>, and some of the\nmodern-day clones of Winchester Models 1873 and 1892\n(e.g., <a href=\"http:\/\/www.personaldefenseworld.com\/2014\/10\/gun-review-taylors-1892-alaskan\/#taylor-1892-alaskan-takedown-1\">Chiappa 1892 Alaskan<\/a>).\nIn the prepper context, their appeal is that you only need to\nkeep one kind of ammo for two types of firearms. Putting a handgun caliber in a rifle gives you greatly\nimproved accuracy, virtually no recoil, comparatively quiet operation, and somewhat improved range - but going past\n100 yards is still going to be a stretch.\n<\/p>\n<\/li><li> <p><b>Shotguns.<\/b>\nLong, heavy, large-bore weapons with tremendous stopping power, variously firing one large metal\nslug or a swarm of high-energy pellets; less-lethal rubber batons are also sometimes used for\ncrowd or animal control. Aimed as easily as rifles, but because of the less ergonomic shapes of\nprojectiles and their lower velocities, the effective range of a shotgun doesn't extend far\nbeyond 100-200 yards. In places such as Australia or the UK, buying a shotgun is subject to\nfewer restrictions than other types of firearms.\n<\/p>\n<p>\nExcellent for home defense, where they combine high stopping power and accuracy with reduced\nrisk of overpenetration. Also popular for waterfowl hunting (where multiple projectiles reduce\nthe need for very precise aim), deer hunting, and for encounters with\npredators such as mountain lions, bears, and the like.\n<\/p>\n<p>\nMost shotguns pack a substantial amount of recoil and can easily bruise your shoulder -\nalthough lower-powered loads or smaller-gauge\nvariants (e.g., 20 ga) can be operated by small-framed or younger shooters.\nCapacity typically ranges from 2 to 8 shells. A very affordable, popular, and reliable\ntype of a shotgun is pump-action (reloaded by racking the slide); two of the best-known\nexamples are \n<a href=\"http:\/\/www.remington.com\/shotguns\/pump-action\/model-870\/model-870-express-synthetic-tactical-7-round\">Remington 870<\/a>\nand\n<a href=\"http:\/\/www.mossberg.com\/product\/590-7-shot-50778\/\">Mossberg 590<\/a>.\n<\/p>\n<\/li><\/ul><p>\nWhen it comes to self-defense, hollow point ammunition (HP \/ JHP) is preferred for rifles and handguns;\ncompared to solid metal bullets, it deforms and falls apart more readily on impact - therefore\nreducing the risk of overpenetration and collateral damage.\n<\/p>\n<p>\nIf you decide to get a gun, you must learn how to operate it safely: sign up for a basic course\nor have someone truly competent take you to the range. The basic safety rules can be summed up as:\n<\/p>\n<ul><li><p><b>Rule #1:<\/b> When picking up a gun, always assume that it's loaded. Don't trust others and \ndefinitely don't trust your own memory. Always double-check. If you don't know how to operate\na particular gun, ask before touching it. In firearms with removable magazines, always check the\nchamber <u>after<\/u> removing the mag.<\/p>\n<\/li><li><p><b>Rule #2:<\/b> Assume that you will eventually mess up rule #1, so when handling a gun, always keep it\npointed in a safe direction. \"Sweeping\" other people is a no-no, so is looking down the barrel.\nAt the shooting range, the safe direction is down range, toward the backstop.\nAt home, just be mindful of neighbors and people in other rooms.<\/p>\n<\/li><li><p><b>Rule #3:<\/b> Assume that you will eventually break rules #1 and #2, so keep your finger off the\ntrigger unless the gun is aimed and you're ready to fire.<\/p>\n<\/li><li><p><b>Rule #4:<\/b> When firing, know your target and what's beyond it. You don't want to shoot\na drunk family member in the middle of the night, or have a bullet go through your target and strike\na bystander.<\/p>\n<\/li><\/ul><p>\nWith the rules internalized, you are extremely unlikely to cause unintentional harm. \nKeep practicing at least twice a month until you get good, and then go to the range at\nleast several times a year. Try to use\npractice ammo with lead-free primers and clean bullets (e.g., RUAG Copper Matrix,\nMagtech Clean Range, Winchester Super Clean, Federal Ballisticlean,\nRemington Disintegrator, Federal American Eagle TMJ, Federal Power-Shok Copper) and avoid\ntracking lead residues from indoor ranges back home -\nespecially if you have small kids. Always wear hearing and eye protection, too.\n<\/p>\n<p>\nKeep your firearms in a quick-access safe if you have young children or expect people with kids\nto visit you every now and then. A safe is also a good way to deter opportunistic theft; a clever\nhiding place will also do, but criminals often have the same ideas as you and know where to look.\nBe smart about picking the right safe and placing it sensibly: you don't want to have to walk to the\nother end of the house, or to fiddle with keys or rotary dials, when every second counts.\nElectronic combination or biometric safes are usually pretty good, unless you go for the bargain bin\n(don't).\n<\/p>\n<h2>5. Organizing and tracking your supplies [<a href=\"#5\">link<\/a>]<\/h2>\n<p>\nRobust bookkeeping is an essential component of any preparedness plan. Without a neat\nspreadsheet to go back to, you will eventually lose track of the stuff you have, won't be able to locate\ncritical supplies in a pinch, and won't know if and when your batteries, food, or meds have\ngone bad.\n<\/p>\n<p>\nTo get it right, make a list of all your emergency gear, along with expiration dates where applicable.\nNext, go through the list marking all the \"stays home\" stuff - the supplies that are impractical\nto haul around or not particularly essential when evacuating. Make sure that all the tactical gear\n- such as flashlights, fire extinguishers, first-aid kits, and self-defense weapons - are in a\nlogical and easily reached place. For the remaining \"stay home\" items,\njust find an unobtrusive location, stow them away, and write the spot down in your spreadsheet.\n<\/p>\n<p>\nThe task of organizing the essentials you want to travel with is a bit\nmore involved. For car owners, I suggest preparing two separate kits:\n<\/p>\n<ul><li><p><b>Boxed evacuation essentials.<\/b>\nCamping and survival supplies to get you through at least one week, in case you need to leave home\nand can't be sure about finding a hospitable location right away. The gear should be boxed\nor bagged to make it easy to load into your vehicle. Include some amount of water and food,\nand make sure that the entire kit actually fits into the car. For many US cars,\n<a href=\"http:\/\/smile.amazon.com\/dp\/B007TOJ9DY\">56 quart storage totes<\/a> ($19) work very well.\n<\/p>\n<\/li><li> <p><b>A get-me-home box.<\/b>\nA container always kept in the trunk, small enough so that it doesn't hinder your normal\nuse of the car, but substantial enough to help you survive several days (or cope with other,\nmore prosaic roadside emergencies). The kit should include 1-2 gallons of water, Mylar\nblankets, rope, and other car supplies discussed earlier in the guide. A\n<a href=\"http:\/\/smile.amazon.com\/dp\/B00BI9AM3C\">collapsible water bottle<\/a> and a\n<a href=\"http:\/\/smile.amazon.com\/dp\/B01DWPOPB4\">folding daypack<\/a> can be useful if you have\nto walk on foot from a broken car to the nearest town. Throwing in\nsome cash - just enough to pay for gas, a meal, a motel room, or a ride home - is also a\ngood plan.\n<\/p>\n<\/li><\/ul><p>\nWhile you're at it, also check that you have a spare tire, a jack, a lug wrench, and a box of extra\nfuses. They come optional on some very cheap cars, and may be simply missing when you buy a used one.\nSome people find out too late.\n<\/p>\n<p>\nFolks without a car are at a marked disadvantage, but should still try to put together\na 72-hour \"bug out bag\" - and ideally, keep it somewhere within a walking or biking\ndistance of their home (say, at work or at a friend's place). It's best to keep it light;\nsome cash, 2-4 quarts of water, a 3,600 kcal emergency ration, and a raincoat will be almost\ncertainly more useful than a gun and a collection of throwing knives.\n<\/p>\n<p>\nEither way, when done with the list, be sure to re-read the response plans you drafted earlier on\nand cross-reference them with this spreadsheet. Iterate until you're happy with both, then\nprint out the docs and place them somewhere intuitive. In a stressful situation, you will\nbe able to quickly review the printouts to confirm that you are not missing anything.\n<\/p>\n<h2>6. Final words [<a href=\"#6\">link<\/a>]<\/h2>\n<p>\nHey, nice! You made it to the very end! So, here is a closing thought...\n<\/p>\n<p>\nRational prepping is meant to give you confidence to go about your business, knowing that you are\nwell-equipped to weather out adversities. But it should not be about convincing yourself that the\ncollapse is just around the bend, and letting that thought consume and disrupt your life.\n<\/p>\n<p>\nStay positive: the world is <a href=\"http:\/\/smile.amazon.com\/dp\/1250057671\">probably not ending<\/a>,\nand there is a good chance that it will be\nan even better place for our children than it is for us. But the universe is a harsh mistress, and\nthere is only so much faith we should be putting in good fortune, in benevolent governments, or in\nthe wonders of modern technology. So, always have a backup plan.\n<\/p>\n<p>\n<i>\n<b>Follow the author on Twitter:<\/b> <a href=\"https:\/\/twitter.com\/lcamtuf\">@lcamtuf<\/a>.<br \/><b>Send in your feedback, questions, or kind words:<\/b>\n<a href=\"mailto:lcamtuf@coredump.cx\">lcamtuf@coredump.cx<\/a>.\n<\/i>\n<\/p>\n<p><i>Your lucky number is: 18908986<\/i><\/p>","created_at":"2018-11-07T16:53:37+0000","updated_at":"2018-11-07T16:53:37+0000","annotations":[],"mimetype":"text\/html","reading_time":146,"domain_name":"lcamtuf.coredump.cx","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":610,"title":"Working on an Android tablet, 2017 edition","url":"https:\/\/bergie.iki.fi\/blog\/working-on-android-2017\/","content":"<p>Back in 2013 I was <a href=\"https:\/\/bergie.iki.fi\/blog\/working-on-android\/\">working exclusively on an Android tablet<\/a>. Then with the <a href=\"https:\/\/bergie.iki.fi\/blog\/noflo-kickstarter-launch\/\">NoFlo Kickstarter<\/a> I needed a device with a desktop browser. What followed were brief periods working on a Chromebook, on a 12\u201d MacBook, and even an iPad Pro.<\/p><p>But from April 2016 onwards I\u2019ve been again working with an Android device. Some people have asked me about my setup, and so here is an update.<\/p><p><a href=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_travelers_notebook.jpg\"><img src=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_travelers_notebook_small.jpg\" alt=\"Information technology\" \/><\/a><\/p><h2 id=\"why-work-on-a-tablet\">Why work on a tablet?<\/h2><p>When I started on this path in 2013, using a tablet for \u201creal work\u201d was considered crazy. While every story on tablet productivity still brings out the people claiming <em>it is not a real computer for real work<\/em>, using <em>tablets for real work<\/em> is becoming more and more common.<\/p><p>A big contributor to this has been the plethora of work-oriented tablets and convertibles released since then. Microsoft\u2019s popular <a href=\"https:\/\/en.m.wikipedia.org\/wiki\/Microsoft_Surface\">Surface Pro line<\/a> brought the PC to tablet form factor, and Apple\u2019s <a href=\"https:\/\/en.m.wikipedia.org\/wiki\/IPad_Pro\">iPad Pro devices<\/a> gave the iPad a keyboard.<\/p><p>Here are couple of great posts talking about how it feels to work on an iPad:<\/p><ul><li><a href=\"http:\/\/mattgemmell.com\/rediscovering-the-ipad\/\">Rediscovering the iPad<\/a><\/li>\n  <li><a href=\"http:\/\/www.macworld.com\/article\/3130710\/ios\/when-traveling-my-ipad-is-essential-and-my-mac-is-the-add-on.html\">When traveling, my iPad is essential and my Mac is the add-on<\/a><\/li>\n  <li><a href=\"https:\/\/www.macstories.net\/stories\/one-year-of-ipad-pro\/\">A Computer for Everything: One Year of iPad Pro<\/a><\/li>\n  <li><a href=\"https:\/\/brooksreview.net\/2016\/12\/evovling-ipad-desktop-usage\/\">Evolving iPad Desktop Usage<\/a><\/li>\n  <li><a href=\"https:\/\/medium.learningbyshipping.com\/my-tablet-has-stickers-8f7ab9022ebd#.vqpn9n2fi\">My Tablet Has Stickers<\/a><\/li>\n<\/ul><p>With all the activity going on, one could claim using a tablet for work has been normalized. But why work on a tablet instead of a \u201creal computer\u201d? Here are some reasons, at least for me:<\/p><h3 id=\"free-of-legacy-cruft\">Free of legacy cruft<\/h3><p>Desktop operating systems have become clunky. Window management. File management. Multiple ways to discover, install, and uninstall applications. Broken notification mechanisms.<\/p><p>With a tablet you can bypass pretty much all of that, and jump into a simpler, cleaner interface designed for the modern connected world.<\/p><p>I think this is also the reason driving some developers back to Linux and <a href=\"http:\/\/swaywm.org\">tiling window managers<\/a> \u2014 cutting manual tweaking and staying focused.<\/p><h3 id=\"amazing-endurance\">Amazing endurance<\/h3><p>Admittedly, laptop battery life has increased a lot since 2013. But with some manufacturers using this an excuse to ship thinner devices, tablets still win the endurance game.<\/p><p>With my current work tablet, I\u2019m customarily getting 12 or more hours of usage. This means I can power through the typical long days of a startup founder without having to plug in. And when traveling, I really don\u2019t have to care where power sockets are located on trains, airplanes, and conference centers.<\/p><p>Low power usage also means that I can really get a lot of more runtime by utilizing the <a href=\"http:\/\/www.macworld.com\/article\/3034575\/hardware\/anker-powercore-20100-review-a-top-performing-usb-c-battery-pack.html\">mobile battery pack<\/a> I originally bought to use with my phone. While I\u2019ve never actually had to try this, back-of-the-envelope math claims I should be able to get a full workweek from the combo without plugging in.<\/p><h3 id=\"work-and-play\">Work and play<\/h3><p>The other aspect of using a tablet is that it becomes a very nice content consumption device after I\u2019m done working. Simply disconnect the keyboard and lean back, and the same device you used for writing software becomes a great e-reader, video player, or a gaming machine.<\/p><p><a href=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_spacex.jpg\"><img src=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_spacex_small.jpg\" alt=\"Livestreaming a SpaceX launch\" \/><\/a><\/p><p>This combined with the battery life has meant that I\u2019ve actually stopped carrying a Kindle with me. While an e-ink screen is still nicer to read, not needing an extra device has its benefits, especially for a frequent one-bag traveller.<\/p><h2 id=\"the-setup\">The setup<\/h2><p>I\u2019m writing this on a <a href=\"https:\/\/en.m.wikipedia.org\/wiki\/Pixel_C\">Pixel C<\/a>, a 10.2\u201d Android tablet made by Google. I got the device last spring when there were developer discounts available at ramp-up to the Android 7 release, and have been using it full-time since.<\/p><h3 id=\"software\">Software<\/h3><p><a href=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/android_homescreen_2017.png\"><img src=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/android_homescreen_2017_small.png\" alt=\"My Android homescreen\" \/><\/a><\/p><p>Surprisingly little has changed in my software use <a href=\"https:\/\/bergie.iki.fi\/blog\/working-on-android\/\">since 2013<\/a> \u2014 I still spend the most of the time writing software in either <a href=\"https:\/\/flowhub.io\">Flowhub<\/a> or terminal. Here are the apps I use on daily basis:<\/p><ul><li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.sonelli.juicessh\">JuiceSSH<\/a> for mosh access to my remote development servers<\/li>\n  <li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.termux\">Termux<\/a> for local and offline development<\/li>\n  <li><a href=\"https:\/\/flowhub.io\">Flowhub<\/a> for visual programming<\/li>\n  <li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=io.thegrid.app\">The Grid<\/a> for updating my various websites<\/li>\n  <li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.Slack\">Slack<\/a>, <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.google.android.talk\">Hangouts<\/a>, and <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.google.android.apps.inbox\">Inbox by Gmail<\/a> for communications<\/li>\n  <li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.google.android.apps.docs\">Google Drive<\/a> and the associated applications for budgeting and planning<\/li>\n  <li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.android.chrome\">Chrome<\/a> for web<\/li>\n<\/ul><p>Looking back to the situation in early 2013, the biggest change is that <strong>Slack<\/strong> has pretty much killed work email.<\/p><p><strong>Termux<\/strong> is a new app that has done a lot to improve the local development situation. By starting the app you get a very nice Linux chroot environment where a lot of software is only a quick <code class=\"highlighter-rouge\">apt install<\/code> away.<\/p><p>Since much of my non-Flowhub work is done in <em>tmux<\/em> and <em>vim<\/em>, I get the exactly same working environment on both local chroot and cloud machines by simply installing <a href=\"https:\/\/github.com\/bergie\/dotfiles\">my dotfiles<\/a> on each of them.<\/p><h3 id=\"keyboard\">Keyboard<\/h3><p><a href=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_laptop.jpg\"><img src=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_laptop_small.jpg\" alt=\"Laptop tablet\" \/><\/a><\/p><p>When I\u2019m on the road I\u2019m using the <a href=\"http:\/\/www.anandtech.com\/show\/9972\/the-google-pixel-c-review\/7\">Pixel C keyboard<\/a>. This doubles as a screen protector, and provides a reasonable laptop-like typing environment. It attaches to the tablet with very strong magnets and allows a good amount of flexibility on the screen angles.<\/p><p>However, when stationary, no laptop keyboard compares to a real mechanical keyboard. When I\u2019m in the office I use a <a href=\"http:\/\/www.cultofmac.com\/290750\/filco-minila-air-bluetooth-keyboard-review\/\">Filco MiniLa Air<\/a>, a bluetooth keyboard with quiet-ish Cherry MX brown switches.<\/p><p><a href=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_desktop.jpg\"><img src=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_desktop_small.jpg\" alt=\"Desktop tablet\" \/><\/a><\/p><p>This tenkeyless (60%) keyboard is extremely comfortable to type on. However, the sturdy metal case means that it is a little too big and heavy to carry on a daily basis.<\/p><p>In practice I\u2019ve only taken the mechanical keyboard with me when there has been a longer trip where I know that I\u2019ll be doing a lot of typing. To solve this, I\u2019m actually <a href=\"https:\/\/www.instagram.com\/p\/BP0lNxJDng_\/?taken-by=henribergius\">looking to build<\/a> a more compact custom mechanical keyboard so I could always have it with me. (<strong>Update<\/strong>: <a href=\"http:\/\/bergie.iki.fi\/blog\/atreus-build-log\/\">here is the keyboard I built<\/a>).<\/p><h2 id=\"comparison-with-ios\">Comparison with iOS<\/h2><p>So, why work on Android instead of getting an iPad Pro? I\u2019ve actually worked on both, and here are my reasons:<\/p><ul><li><strong>Communications between apps<\/strong>: while iOS has extensions now, the ability to send data from an app to another is still a hit-or-miss. Android had intents from day one, meaning pretty much any app can talk to any other app<\/li>\n  <li><strong>Standard charging<\/strong>: all of my other devices charge with the same USB-C chargers and cables. iPads still use the proprietary Lightnight plug, requiring custom dongles for everything<\/li>\n  <li><strong>Standard accessories<\/strong>: this boils down to USB-C just like charging. With Android I can plug in a network adapter or even a mouse, and it\u2019ll just work<\/li>\n  <li><strong>Ecosystem lock-in<\/strong>: we\u2019re moving to a world where everything \u2014 from household electronics to cars \u2014 is either locked to the Apple ecosystem or following standards. I don\u2019t want to be locked to a single vendor for everything digital<\/li>\n  <li><strong>Browser choice<\/strong>: with iOS you only get one web renderer, the rather dated Safari. On Android I can choose between Chrome, Firefox, or any other browser that has been ported to the platform<\/li>\n<\/ul><p>Of course, iOS has its own benefits. Apple has a stronger stance on privacy than Google. And there is more well-made tablet software available for iPads than Android. But when almost everything I use is available on the web, this doesn\u2019t matter that much.<\/p><h2 id=\"the-future\">The future<\/h2><p><a href=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_cbase.jpg\"><img src=\"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_cbase_small.jpg\" alt=\"Hacking on the c-base patio\" \/><\/a><\/p><p>As a software developer working on Android tablets, the weakest point of the platform is still that there are <em>no browser developer tools<\/em> available. This was a problem in 2013, and it is still a problem now.<\/p><p>From my conversations with some Chrome developers, it seems Google has very little interest in addressing this. However, there is a bright spot: the new breed of convertible Chromebooks being released now. And they run Android apps:<\/p><ul><li><a href=\"https:\/\/www.wired.com\/2016\/09\/chromebooks-totally-transform-laptop-design\/\">How Chromebooks Are About to Totally Transform Laptop Design<\/a><\/li>\n  <li><a href=\"https:\/\/chromeunboxed.com\/detachable-chromebooks-pixel-c-and-the-future-of-chrome-os\/\">Detachable Chromebooks, Pixel C And The Future Of Chrome OS<\/a><\/li>\n  <li><a href=\"https:\/\/arstechnica.com\/gadgets\/2017\/02\/samsungs-chromebook-pro-a-thoughtful-marriage-of-android-and-chrome-os\/\">Samsung\u2019s Chromebook Pro gives me hope in Chrome OS\u2014thanks to Android\u2019s help<\/a><\/li>\n<\/ul><p>Chrome OS is another clean, legacy free, modern computing interface. With these new devices you get the combination of a full desktop browser and the ability to run all Android tablet software.<\/p><p>The Samsung Chromebook Pro\/Plus mentioned above is definitely interesting. A high-res 12\u201d screen and a digital pen which I see as something very promising for <a href=\"https:\/\/flowhub.io\">visual programming<\/a> purposes.<\/p><p>However, given that I already have a great mechanical keyboard, I\u2019d love a device that shipped without an attached keyboard. We\u2019ll see what kind of devices get out later this year.<\/p>","created_at":"2018-11-07T16:53:13+0000","updated_at":"2018-11-07T16:53:13+0000","published_at":"2017-02-10T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"bergie.iki.fi","preview_picture":"https:\/\/d2vqpl3tx84ay5.cloudfront.net\/android-tablet-2017\/pixel_c_desktop.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":609,"title":"exa \u00b7 a modern replacement for ls","url":"https:\/\/the.exa.website\/","content":"<div class=\"row main-introduction sep\"><section class=\"main-lede left\"><p>You list files hundreds of times a day.\nWhy spend your time squinting at black and white text?<\/p><p><strong>exa<\/strong> is an improved file lister with more features and better defaults.\nIt uses colours to distinguish file types and metadata.\nIt knows about symlinks, extended attributes, and Git.\nAnd it\u2019s small, fast, and just one single binary.<\/p><p><b>Most recent version:<\/b><br \/><a class=\"internal\" href=\"https:\/\/the.exa.website\/releases\/v0.8.0\">v0.8.0, released on September 30<sup>th<\/sup><\/a><\/p><\/section><section id=\"js-terminal\" class=\"terminal-monitor right\"><pre class=\"terminal\">exa --long --header --git\nPermissions Size User Date Modified Git Name\n.rw-r--r--  1.9k ben   1 Feb  3:14   -- build.rs\n.rw-r--r--   19k ben  16 Mar 17:36   -M Cargo.lock\n.rw-r--r--  1.1k ben  15 Mar 23:25   -- Cargo.toml\ndrwxr-xr-x     - ben  15 Mar 23:25   -- contrib\ndrwxr-xr-x     - ben   1 Feb  3:14   -- devtools\n.rwxr-xr-x  1.4M ben   1 Feb  3:14   -- exa-linux-x86_64\n.rwxr-xr-x  1.2M ben   1 Feb  3:15   -- exa-macos-x86_64\n.rw-r--r--  1.1k ben   1 Feb  3:15   -- LICENCE\n.rw-r--r--  3.1k ben   1 Feb  3:15   -- Makefile\n.rw-r--r--   102 ben   1 Feb  3:15   -- MD5SUMS\n.rw-r--r--  6.0k ben   1 Feb  3:15   -- README.md\n.rw-r--r--  454k ben   1 Feb  3:15   -- screenshots.png\n.rw-r--r--   118 ben   1 Feb  3:15   -- SHA1SUMS\ndrwxr-xr-x     - ben  30 Apr 17:39   -N src\ndrwxr-xr-x     - ben  16 Mar 17:32   -- target\n.rw-r--r--   24k ben  16 Mar 17:58   -M Vagrantfile\ndrwxr-xr-x     - ben   1 Feb  3:16   -- xtests<\/pre><\/section><\/div><section class=\"main-features sep\"><div class=\"feature-row row -fill\"><div class=\"feature-box feature -colours\"><p class=\"content\"><b class=\"title\">All the colours:<\/b>\nDifferent types of file and data will be <a class=\"internal\" href=\"https:\/\/the.exa.website\/features\/colours\">coloured differently<\/a>,\nand the user and group columns will be <a class=\"internal\" href=\"https:\/\/the.exa.website\/features\/long-view#owners-and-groups\">highlighted for the current user<\/a>.<\/p><\/div><div class=\"feature-box feature -fast\"><p class=\"content\"><b class=\"title\">It\u2019s fast:<\/b>\nexa queries files in parallel, giving you performance on par with <code>ls<\/code>.<\/p><\/div><\/div><div class=\"feature-row row -fill\"><div class=\"feature-box feature -tree\"><p class=\"content\"><b class=\"title\">Tree view:<\/b>\nNot only is the standard <strong>tree<\/strong> tool <a class=\"internal\" href=\"https:\/\/the.exa.website\/features\/tree-view\">built-in<\/a>, but it\u2019ll show you your files\u2019 information alongside the hierarchy.<\/p><\/div><div class=\"feature-box feature -git\"><p class=\"content\"><b class=\"title\">Git support:<\/b>\nView the <a class=\"internal\" href=\"https:\/\/the.exa.website\/features\/git\">staged and unstaged status<\/a> of every file, right there in the standard view. Also works in tree view for a high-level overview of your repository.<\/p><\/div><div class=\"feature-box feature -wide\"><p class=\"content\"><b class=\"title\">Wide view:<\/b>\nHow many columns can you display in your terminal at once? <a class=\"internal\" href=\"https:\/\/the.exa.website\/features\/long-view#long-grid\">How many do you need?<\/a><\/p><\/div><\/div><\/section><section class=\"main-installation sep\"><div class=\"row -fill radio\"><div class=\"radio-inner\"><div class=\"os-picker\"><p>All downloads\nCompile from source<\/p><p>Arch\nFedora\nGentoo\nmacOS\nNixOS\nopenSUSE<\/p><\/div><div id=\"js-in-arch\" class=\"split -middle\"><div><p>sudo pacman -S exa<\/p><\/div><p>On Arch, install the <a href=\"https:\/\/www.archlinux.org\/packages\/community\/x86_64\/exa\/\" title=\"exa in the Arch package repository\">exa<\/a> package.<\/p><\/div><div id=\"js-in-fedora\" class=\"split -middle\"><div><p>sudo dnf install exa<\/p><\/div><p>On Fedora, install the <a href=\"https:\/\/apps.fedoraproject.org\/packages\/rust-exa\" title=\"exa in the Fedora package repository\">rust-exa<\/a> package.<\/p><\/div><div id=\"js-in-gentoo\" class=\"split -middle\"><div><p>emerge sys-apps\/exa<\/p><\/div><p>On Gentoo, install the <a href=\"https:\/\/packages.gentoo.org\/packages\/sys-apps\/exa\" title=\"exa in the Gentoo package repository\">sys-apps\/exa<\/a> package.<\/p><\/div><div id=\"js-in-macos\" class=\"split -middle\"><div><p>brew install exa<\/p><\/div><p>If you have <a href=\"https:\/\/brew.sh\" title=\"Homebrew, a package manager for macOS\">Homebrew<\/a> installed, you can download and install the <a href=\"http:\/\/formulae.brew.sh\/formula\/exa\" title=\"exa in the Homebrew package repository\">exa<\/a> formula.<\/p><\/div><div id=\"js-in-nixos\" class=\"split -middle\"><div><p>nix-env -i exa<\/p><\/div><p>On NixOS, install the <a href=\"https:\/\/github.com\/NixOS\/nixpkgs\/blob\/master\/pkgs\/tools\/misc\/exa\/default.nix\" title=\"exa in the NixOS package repository\">exa<\/a> package.<\/p><\/div><div id=\"js-in-opensuse\" class=\"split -middle\"><div><p>sudo zypper install exa<\/p><\/div><p>On openSUSE, install the <a href=\"https:\/\/software.opensuse.org\/package\/exa\" title=\"exa in the openSUSE package repository\">exa<\/a> package.<\/p><\/div><section id=\"js-in-downloads\" class=\"split\"><p class=\"downloads\">Select your OS for<br \/>more specific instructions<\/p><ul><li><code><a href=\"https:\/\/github.com\/ogham\/exa\/releases\/download\/v0.8.0\/exa-linux-x86_64-0.8.0.zip\" title=\"The latest Linux exa binary\">exa-linux-x86_64-0.8.0.zip<\/a><\/code><\/li><li><code><a href=\"https:\/\/github.com\/ogham\/exa\/releases\/download\/v0.8.0\/exa-macos-x86_64-0.8.0.zip\" title=\"The latest macOS exa binary\">exa-macos-x86_64-0.8.0.zip<\/a><\/code><\/li><\/ul><\/section><section id=\"js-in-compile\" class=\"compile\"><div class=\"split -middle\"><p>If you have the right programs installed, you can build the latest development version from source.<\/p><\/div><ol><li>Download and install <a href=\"https:\/\/www.rust-lang.org\/\" title=\"The Rust programming language\">Rust<\/a> for your platform.<\/li><li>Install <code>libgit2<\/code> and <code>cmake<\/code>.<\/li><li>To download the latest version, run:<code> git clone https:\/\/github.com\/\u00adogham\/exa.git<\/code><\/li><li>Run <code>make install<\/code> in the new directory to compile and install exa into <code>\/usr\/local\/bin<\/code>.<\/li><\/ol><\/section><\/div><\/div><\/section><div class=\"main-questions sep\"><section class=\"faq-section\"><div class=\"row -cols -flip\"><div class=\"text\"><h3 class=\"question\">Is this a drop-in replacement for <code>ls<\/code>?<\/h3><p class=\"answer\">No \u2014 exa has, in my opinion, much saner defaults than <code>ls<\/code>, so while the available command-line options are similar, they are not <strong>exactly<\/strong> the same. Most of the common options will work consistently, though. For example, exa prints <a class=\"internal\" href=\"https:\/\/the.exa.website\/features\/long-view#sizes\">human-readable file sizes<\/a> by default, so the <code>-h<\/code> option no longer applies.<\/p><h3 class=\"question\">Why would I want Git in my file listing anyway?<\/h3><p class=\"answer\">Because you get to see the Git information alongside everything else. Yes, it\u2019s already trivial to just run <code>git status<\/code> when you\u2019re in your shell. But when you need to double-check exactly which files have been staged, you <strong>really<\/strong> need to see it clearly!<\/p><h3 class=\"question\">Windows support?<\/h3><p class=\"answer\">Soon\u2122.<\/p><h3 class=\"question\">What licence does exa use?<\/h3><p class=\"answer\">It uses the <a href=\"https:\/\/choosealicense.com\/licenses\/mit\/\" title=\"The permissions of the MIT licence\">MIT licence<\/a>. The code is open-source \u2014 development happens <a href=\"https:\/\/github.com\/ogham\/exa\" title=\"exa on GitHub\">over on GitHub<\/a>, and <a class=\"internal\" href=\"https:\/\/the.exa.website\/contributing\">contributions are gratefully accepted<\/a>!<\/p><\/div><\/div><\/section><\/div><footer class=\"main-footer\"><img class=\"icon\" src=\"https:\/\/the.exa.website\/static\/img\/twitter.svg\" alt=\"\" width=\"16\" height=\"16\" \/>\nwith\n<a href=\"https:\/\/twitter.com\/dot_slash_exa\" title=\"exa on Twitter\">@dot_slash_exa<\/a>\non Twitter\n\u00b7\n<img class=\"icon\" src=\"https:\/\/the.exa.website\/static\/img\/octocat.svg\" alt=\"\" width=\"16\" height=\"16\" \/><a href=\"https:\/\/github.com\/ogham\/exa\" title=\"exa on GitHub\">ogham\/exa<\/a>\non GitHub<\/footer>","created_at":"2018-11-07T16:53:12+0000","updated_at":"2018-11-07T16:53:12+0000","published_by":["@cairnrefinery"],"annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"the.exa.website","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":607,"title":"Crafting Plausible Maps","url":"https:\/\/mythcreants.com\/blog\/crafting-plausible-maps\/","content":"<input class=\"jpibfi\" type=\"hidden\" \/><div class=\"content-wrap\">Building a map for a fantasy setting involves a lot of details \u2013 most of them fun! Art styles, fonts, and icons need to be chosen. But some mapping concerns go beyond mere aesthetics. If you\u2019re building a sizable chunk of continent on an <a title=\"Should Your Fantasy World Resemble Earth?\" href=\"https:\/\/mythcreants.com\/blog\/should-your-fantasy-world-resemble-earth\/\">Earth-like world<\/a>,<a href=\"#\" title=\"Which is generally a good idea, unless exploring a bizarre world is the point of the story.\">*<\/a> you\u2019re going to need to keep <a title=\"Tweaking Your World\u2019s Geology\" href=\"https:\/\/mythcreants.com\/blog\/tweaking-your-worlds-geology\/\">geology<\/a> in mind. \n<h2><strong>Mountains<\/strong><\/h2>\n<p> Mountains are the skeletal system of your world. Unless you already know the exact shape of your continent or island, sketch your mountains before your coastlines.<a href=\"#\" title=\"Or be ready to revise the coasts when mountains start bumping into them.\">*<\/a> Mountains back each other up and form chains, ranges, and ridges. They take their shape from their creation: <\/p>\n<ul><li>Most mountains are the children of lusty continental plates, birthed from scandalous collisions between landmasses too attracted to each other for their own good. The surface crumples upward where the plates collide. Even when one plate slides beneath the other, the friction and pressure send mountains shooting upward.<\/li>\n<li>Volcanic mountains can form near the edge of the top plate during this long, slow collision. Volcanoes are formed from the recycled remnants of the bottom plate.<a href=\"#\" title=\"Technically speaking, the plate that is being subducted.\">*<\/a><\/li>\n<li>Plates burning themselves as they migrate over <a title=\"Wikipedia: Mantle Plume\" href=\"https:\/\/en.wikipedia.org\/wiki\/Mantle_plume\">mantle plumes<\/a><a href=\"#\" title=\"Abnormally hot regions within the Earth\u2019s mantle. The mantle has a complex structure and is not just a featureless hot ball of molten rock.\">*<\/a> deep in the planet can also give rise to volcano-blisters. These eventually form chains of mountains as the plate slowly moves over the hotspot.<\/li>\n<li>Impacts between planets and asteroids form circular rings of mountains. Sometimes an impact crater boasts a second ring or lone peak at the center of the crater. If the crater is big enough and the world is Earth-like, it will likely be a lake or sea.<\/li>\n<\/ul><p> All four of these methods form chains, ridges, long plateaus, or rings. Lone mountains are rare and almost always volcanic in origin.<a href=\"#\" title=\"So Mount Doom from Lord of the Rings is not geologically implausible, being a volcano and all. Its too-obvious name and its physics-defying lava leave something to be desired, however.\">*<\/a> Hill country, worn plateaus, badlands, and other rugged-but-not-mountainous terrain is typically the eroded remains of ancient, \u201cdead\u201d mountains. <\/p>\n<h3>On Your Map:<\/h3>\n<p> Place mountains in rows or long blobs. If you\u2019ve already drawn your coastlines, extend the mountain chains past them, forming peninsulas and island chains. On land, their foothills should be visible: scatter a few hills here and there at the margins of the mountains. A few ranges of lone hills (old, worn out mountain ranges) are a good idea. The occasional lone mountain is okay; it\u2019s probably a volcano or the central peak inside a mountain-ringed impact crater. <\/p>\n<p> If your goal is realistic topography, remember that mountain chains look like long fused ridges at continental scales. Most individual mountains are only discernible when the map is zoomed in to the scale of a small European country or U.S. state.<a href=\"#\" title=\"But this is primarily a choice of artistic style. Portraying a continent\u2019s mountain ranges as a series of, say, twenty individual peaks may not be technically accurate, but it gets your point across. You only need to worry about the size of individual mountains when you\u2019re interested in high levels of topographical accuracy.\">*<\/a> <\/p>\n<h2><strong>Coastlines<\/strong><\/h2>\n<p> Coastlines appear to be the singular, end-all defining feature of any landmass \u2013 and to human experience, they are. But a planet\u2019s rock is stalwart and little concerned with its surroundings, whether <a title=\"Tour of Our Terraformed Solar System: Jupiter and Saturn\" href=\"https:\/\/mythcreants.com\/blog\/tour-of-the-terraformed-solar-system-jupiter-and-zohal\/\">gaseous atmosphere<\/a> or <a title=\"Spicing Up Your World\u2019s Chemical Makeup\" href=\"https:\/\/mythcreants.com\/blog\/spicing-up-your-worlds-chemical-makeup\/\">liquid ocean<\/a>. On human timescales (thousands of years or less) water is largely irrelevant to solid land. An island is just a mountain surrounded by ocean, and a mountain is just an island with the water drained down. Any <a title=\"Wikipedia: Contour Line\" href=\"https:\/\/en.wikipedia.org\/wiki\/Contour_line\">contour line<\/a> on a <a title=\"Wikipedia: Topographic Map\" href=\"https:\/\/en.wikipedia.org\/wiki\/Topographic_map\">topographic map<\/a> or any depth line on a nautical chart could theoretically be a coastline. So outline your mountains first. Then for your coastlines, just add water. <\/p>\n<p> When water is poured over a landscape formed of continents and mountains, what does it do? Land isn\u2019t arbitrarily arranged on a planet. Plate tectonics sort rock into lighter and heavier, and clumps the lighter rock together as continents. And at any level of magnification closer than a full global picture, coastlines will have a similar, fractal look. <\/p>\n<p> This fractal nature means you can focus on the coast of a small territory or examine the margins of a continent, and the contours<a href=\"#\" title=\"The general \u201cjaggedness\u201d of the coastline\">*<\/a> should be about the same. In fact, if you aren\u2019t feeling very inspired, you can take a zoomed-in island from Earth<a href=\"#\" title=\"Using, for example, a random Google Maps search.\">*<\/a> and make it your fantasy continent, or vice versa.<a href=\"#\" title=\"Though in the latter case, you may end up with a small island that looks exactly like, say, North America \u2013 and people are going to recognize North America.\">*<\/a> <\/p>\n<p> But there are some quirky exceptions to the fractal behavior of coastlines, and some places where the shore may look distinctly different: <\/p>\n<ul><li>Land sorts into large continents from a distant enough perspective, and tidal flats and beaches complicate matters when zoomed far in.<\/li>\n<li>Mountains, glaciers, and (most importantly) a combination of the two affect coastlines. The glacier-carved fjords, inlets, and island ranges of Chile, British Columbia, and Norway are complex, convoluted affairs. And these regions don\u2019t look the same at all levels of magnification.<\/li>\n<li>Coastlines also get a bit weird around river deltas. The land bulges outward, but this bulge is carved up by multiple river channels.<\/li>\n<li>Certain flat regions, like North America\u2019s eastern seaboard, can have complex chains of ever-shifting barrier islands.<\/li>\n<li>Continents that haven\u2019t had sea-level ice for hundreds of millions of years might have smooth-looking coastlines without a lot of inlets or islands; Africa and Australia are perfect examples.<\/li>\n<\/ul><h3>On Your Map:<\/h3>\n<ul><li>Respect the fact that land likes to clump together into continents; don\u2019t arbitrarily blend land and sea at large scales.<a href=\"#\" title=\"The Earthsea setting is surprisingly non-Earth-like in this regard.\">*<\/a><\/li>\n<li>Make mountainous coastlines in cold areas \u2013 or areas that were once glaciated \u2013 especially \u201c<a title=\"Wikipedia: Slartibartfast (designer of Norway's coastline)\" href=\"https:\/\/en.wikipedia.org\/wiki\/Slartibartfast\">fiddly<\/a>.\u201d Add fjords, inlets, outlying islands, that sort of thing.<\/li>\n<li>Some occasional long, skinny barrier islands running parallel to the coast but smoother than it, can add a realistic feel.<\/li>\n<\/ul><p> Otherwise just make sure that most of your coastlines are rough or jagged. <\/p>\n<h2><strong>Rivers<\/strong><\/h2>\n<p> Mountains and coastlines follow general patterns but few hard-and-fast rules. Not so with rivers; the courses of rivers are dictated by simple logic and never deviate without a lot of magic. Even the most vigorous hand-waving won\u2019t save you from ridicule if one of your rivers flows uphill or follows a circular loop. <\/p>\n<p> Rivers are like coastlines in that their shape is (arguably) fractal, and their placement is a matter of contour lines. Except that rivers are perpendicular to the contours rather than parallel. Rivers always move along the easiest available path from high elevation to low. And the fractal pattern of rivers is directional: rivers will always merge as they flow toward the coast, never split.<a href=\"#\" title=\"Exception: deltas that form in flat coastlines. These are small compared to the course of the river on a whole.\">*<\/a> <\/p>\n<p> Rivers often find themselves trapped at the bottoms of massive canyons eroded over millions of years, prisons of their own making. However, rivers are not permanently fixed in place like mountains; they can migrate. Over flat country, <a title=\"Patternbank: The Alluvial Valley of the Lower Mississippi River \u2013 Harold Fisk\" href=\"http:\/\/blog.patternbank.com\/the-alluvial-valley-of-the-lower-mississippi-river-harold-fisk\/\">meandering rivers shift their course<\/a> by growing continually wider loops.<a href=\"#\" title=\"The river arc\u2019s outer bank is continually eroded and the inner bank grown with sediment deposits.\">*<\/a> Eventually these loops come full circle and \u201cpinch off,\u201d straightening the river once more. But in these scenarios the river stays within its basin. In a mountainous region, a river may change course if rockslides or glaciers block it, forcing the waters to find an entirely new route to the sea. And rivers can be tamed by humans. Even ancient humans frequently diverted rivers along different courses, typically for purposes of irrigation or flood regulation. <\/p>\n<p> Many rivers will be fed by regions with seasonally variant rainfall<a href=\"#\" title=\"Monsoon-watered regions of Africa, for example.\">*<\/a> or spring melts, and have periods of very low or high (flood-level) flow. As an example, the <a title=\"Wikipedia: Fraser River\" href=\"https:\/\/en.wikipedia.org\/wiki\/Fraser_River\">Fraser<\/a> is a fairly short river in Western Canada, but during the wet spring, snowmelt combined with massive rainfall gives this river higher flow volumes than the Mississippi, North America\u2019s longest river. Some rivers \u2013 especially those winding through hot deserts \u2013 may be completely dry and \u201cdead\u201d for part of the year. Or, alternatively, completely frozen. <\/p>\n<h3>On Your Map:<\/h3>\n<p> Here are some river placement rules, in order of importance from \u201cavoid deviation at all costs\u201d to \u201cexplain yourself if you get creative\u201d: <\/p>\n<ul><li>Rivers <em>never<\/em> completely cross a continent or other landmass; they never start in the same body of water they empty into!<a href=\"#\" title=\"Connected oceans count as the same body of water for this purpose. Rivers never flow directly from the Atlantic to the Pacific, for example. And no, the Panama Canal doesn\u2019t count as a river.\">*<\/a><\/li>\n<li>Rivers never form loops. If they appear to, these are actually ring-shaped lakes.<\/li>\n<li>Major rivers will always end at an ocean or, occasionally, a lake.<a href=\"#\" title=\"Or another river, but this doesn\u2019t count as an \u201cend\u201d per se; the river just merges its identity, Borg-like, with the river collective.\">*<\/a> Smaller rivers are like people: they can die long, slow, painful deaths in attempted crossings of a vast desert.<\/li>\n<li>Rivers always join together as they flow toward the sea; they never branch out. Exceptions are made in flat deltas, where the river deposits islands of silt as it empties into an ocean or lake. These are typically small (no bigger than a large city) but can fan out over large distances if the terrain is flat enough. For example, <a title=\"Wikipedia: Bangladesh\" href=\"https:\/\/en.wikipedia.org\/wiki\/Bangladesh\">Bangladesh<\/a>, the world\u2019s eighth-most populous country, consists almost entirely of a single conjoined delta of three large rivers.<\/li>\n<li>Rivers tend to start at mountains \u2013 or occasionally lakes, wetlands, or hills. Mountains catch more water, and they hold significant water as ice that can melt and feed rivers during the spring. If there aren\u2019t any mountains handy, use a forest; these imply sufficient water to found a river.<\/li>\n<\/ul><h2><strong>Lakes<\/strong><\/h2>\n<p> Lakes are generally just a consequence of rivers getting delayed in depressed portions of a continent on their way to the sea. Any time a bowl-shaped, walled-off region exists inside a landmass, its fate depends on how much rain it gets and whether it has a river pouring into it. <\/p>\n<p> If the depressed region is a desert, it will be a dry <a title=\"Wikipedia: Endorheic Basin\" href=\"https:\/\/en.wikipedia.org\/wiki\/Endorheic_basin\">endorheic basin<\/a>.<a href=\"#\" title=\"A drainage area completely internal to a continent; no outlet to an ocean\">*<\/a> It may have a small salt lake in its bottom, because salts will accumulate over millions of years with no way to escape to an ocean. Or it could have a more recently-filled fresh or brackish lake. If it is rainy or fed by a river, it will fill up until the water can spill out at its lowest point, making a freshwater lake that is part of a larger drainage area. <\/p>\n<p> There\u2019s no real limit on lake size, though continents don\u2019t usually have giant holes in them by default. Earth\u2019s largest lakes have one of two backstories: they were carved recently by glaciers,<a href=\"#\" title=\"Example: the Great Lakes of North America\">*<\/a> or they were formed when pieces of continents began to rift apart.<a href=\"#\" title=\"The Great Lakes of Africa\">*<\/a> Rarely, they are <a title=\"Wikipedia: Annular Lake\" href=\"https:\/\/en.wikipedia.org\/wiki\/Annular_lake\">water-filled impact craters<\/a>. <\/p>\n<p> Cold temperate and subarctic regions tend to be peppered with little lakes, connected by a tangle of streams and rivers. This is especially true for cold regions that are rocky over short scales but flat over larger scales. Rockiness means the rivers can\u2019t just carve a straight line through soft ground to their goal. They\u2019re stuck filtering through a series of pools on their way to the ocean. The <a title=\"Wikipedia: Canadian Shield\" href=\"https:\/\/en.wikipedia.org\/wiki\/Canadian_Shield\">Canadian Shield<\/a><a href=\"#\" title=\"Comprises most of east-central Canada, and extends south into the northern two-thirds of Minnesota, \u201cland of 10,000 lakes.\u201d\">*<\/a> is so dense with lakes and connecting rivers that its \u201cland\u201d is nearly half water. <\/p>\n<p> Like rivers, lakes can shift their boundaries. But when lakes drain or fill, they tend to impact a much larger area. These regional apocalypses can be caused by earthquakes, ice dams, or human intervention.<a href=\"#\" title=\"Lake Mead wasn\u2019t a lake until the Hoover Dam was built. And China\u2019s more recent Three Gorges Dam created a lake that displaced millions of people.\">*<\/a> Consider these as great sources of historical trauma for your world. Any fertile river valley that becomes a lakebed has a good chance of drowning an entire civilization. And any inland sea of freshwater that dries up is also going to bring down a civilization or two, if for different reasons. Both of these things have happened throughout humanity\u2019s early history \u2013 if not always on fantastically grand scales. <\/p>\n<p> Less apocalyptic but still interesting are gradual or seasonal changes to a lake\u2019s shoreline. Many lakes, especially in Africa and Australia, fill during monsoons but shrink or <a title=\"Wikipedia: Lake Eyre\" href=\"https:\/\/en.wikipedia.org\/wiki\/Lake_Eyre\">empty out<\/a> during the dry season. This is common in flat regions with shallow, gradual coastlines. Lakes in the mountains might gain or lose similar amounts of water, but their shores are typically so steep they wouldn\u2019t look any different on a map. <\/p>\n<h3>On Your Map:<\/h3>\n<ul><li>Treat the coasts of large lakes like a continent\u2019s outer coastline. Any mountain range that ends at a large lake should poke into it a bit, perhaps ending in a few islands. Small lakes are more like little inlets or channels, filling the spaces between mountains and often mimicking rivers in their long, narrow shape.<a href=\"#\" title=\"Washington State\u2019s Lake Chelan\">*<\/a><\/li>\n<li>Consider dotted or blurred lines for some lakes, since the shores are not always stable. In regions with glacial ice dams or seasonal monsoon rains, lakes could empty and fill over the course of a year.<\/li>\n<li>You can pepper cold, wet regions with lakes, especially if the land is flat enough.<a href=\"#\" title=\"Most of central Canada.\">*<\/a> But don\u2019t put as many lakes in hot, dry regions unless they\u2019re fed by large rivers. Consider placing salt flats here instead.<\/li>\n<li>In flat country, consider decorating meandering rivers with elbow lakes. These are remnants of the river\u2019s previous courses; they look like curved pieces of pinched-off river.<\/li>\n<li>Feel free to string multiple lakes along the course of a river, like beads on a string. This is especially common in cold regions or on rivers held back by dams (natural or human-made).<\/li>\n<\/ul><h2>Vegetation<\/h2>\n<p> Not all maps bother to show vegetation cover, but very large plants warrant enough importance to human travel and habitation that fantasy maps will almost always mark forests. More detailed maps may also differentiate farmers\u2019 fields, barren tundra, deserts, prairie, ice sheets, and marshes. I have a <a title=\"How to Color Your Map Using SCIENCE!\" href=\"https:\/\/mythcreants.com\/blog\/how-to-color-your-map-using-science\/\">full post on placing vegetation<\/a>, but here I\u2019ll give a quick summary: <\/p>\n<p> <a title=\"How to Color Your Map Using SCIENCE!\" href=\"https:\/\/mythcreants.com\/blog\/how-to-color-your-map-using-science\/\"><img class=\"alignleft size-medium wp-image-4572\" src=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181-273x236.png\" alt=\"Interplay of Climate and Terrain\" width=\"273\" height=\"236\" srcset=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181-273x236.png 273w, https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181-300x260.png 300w, https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181-350x303.png 350w, https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181-548x474.png 548w, https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181.png 600w\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/mythcreants.com\/blog\/crafting-plausible-maps\/\" data-jpibfi-post-title=\"Crafting Plausible Maps\" data-jpibfi-src=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2014\/04\/Interplay-of-Climate-and-Terrain-FEATURE_IMAGE_MAYBE-e1397534778181-273x236.png\" \/><\/a> <\/p>\n<p> Climate zones are critical to a region\u2019s vegetation, and the biggest predictor for climate is latitude.<a href=\"#\" title=\"How far north or south are you, and which hemisphere are you in? Just because Earth neglects its southern hemisphere doesn\u2019t mean you have to.\">*<\/a> Any terrestrial planet will have bands of wet and dry. The climate tends to be wet near the equator and again in temperate regions.<a href=\"#\" title=\"Centered at roughly 60% of the distance to the poles, though feel free to tweak this for different Earth-like planets.\">*<\/a> Conversely, most land is dry about a third of the way to the poles<a href=\"#\" title=\"Centered around latitude 30 on Earth, but again, feel free to vary this.\">*<\/a> and dry<a href=\"#\" title=\"Dry in the sense that precipitation is low. But it\u2019s too cold for anything to get the chance to truly dry out.\">*<\/a> again at the poles themselves. <\/p>\n<p> Mountains tend to be colder and wetter than their surroundings, and frequently one side of a mountain range will be wetter or milder than the other. If very different eco-regions don\u2019t have mountains separating them, their transition will be gradual. For example, there\u2019s rarely a cut-and-dry boundary between temperate forest and desert on flat ground. Instead there\u2019s a gradual transition first to savannah, then grassland, then arid scrubland, and finally high desert. <\/p>\n<h2><strong>Settlements<\/strong><\/h2>\n<p> Humans (especially pre-industrial, agricultural humans) like to settle near rivers. All the earliest agricultural civilizations grew up along rivers.<a href=\"#\" title=\"Specifically: the Nile, Tigris\/Euphrates, Ganges, Indus, Yangtze and Yellow.\">*<\/a> Fresh, moving water provided something to drink, irrigation for crops, power for watermills<a href=\"#\" title=\"Though fantasy authors misattribute many Renaissance technologies and fashions as medieval, waterwheels were common power-generating structures (mechanical, not electrical, of course) throughout Europe in early Medieval times.\">*<\/a> and highways for high-volume trade. In the ancient world, it was <em>much<\/em> easier to trade large quantities of goods via water than it was overland. This remained true until \u2013 well, until forever. It\u2019s still true. Ocean coasts and rivers have tradeoffs: both offer good shipping routes, but rivers are sources of fresh water and fertile sediment while oceans are superior for fishing and whaling. Cities that become large and prosperous require not only water and resources but also trade. <\/p>\n<p> Humans also construct some fortifications or settlements in defensible positions, such as atop hills and cliffs or on islands. But these structures will still be close to their centers of civilization (river valleys and coastlines), especially at first. Some fortresses may be found guarding remote mountain passes as well, once agricultural city-states grow into large nations. <\/p>\n<p> <a title=\"Should You Use Non-Humans in Your Setting?\" href=\"https:\/\/mythcreants.com\/blog\/should-you-use-non-humans-in-your-setting\/\">Different humanoids<\/a> \u2013 and different human cultures \u2013 could find different regions attractive for settlement. In high-stereotype fantasy, dwarves prefer underground or mountainous areas and elves prefer great forests.<a href=\"#\" title=\"Alternatively, elves simply don\u2019t cut down the forests they settle in, whereas humans presumably do.\">*<\/a> Drow and sun-sensitive orcs, goblins, or trolls may require underground dwellings. But access to the lands above may be essential for resources or ventilation, necessitating <a title=\"Wikipedia: Karst\" href=\"https:\/\/en.wikipedia.org\/wiki\/Karst\">karst landscapes<\/a> rich with caves. Even garden-variety humans can grow large cities in nontraditional areas. This is especially possible once modern technology, magic, or domesticating a new <a title=\"Using Animals for Transport and Labor\" href=\"https:\/\/mythcreants.com\/blog\/using-animals-for-transport-and-labor\/\">herd animal<\/a> makes it viable. For example, in North America we see cities in barren inland regions<a href=\"#\" title=\"like Phoenix or Denver\">*<\/a> due to modern technologies such as air conditioning; plus, highways and railways have become more important trade routes than rivers. <\/p>\n<h3>On Your Map:<\/h3>\n<p> Put most of your cities on rivers and\/or coastlines unless you have a story-relevant reason not to. Don\u2019t completely disregard this in science fiction, either. We humans might not be so reliant on rivers as we once were, but we still <em>like<\/em> them. The biggest cities tend to show up where trade routes intersect \u2013 at branches in rivers or on the coastline near an important mountain pass. <\/p>\n<h2><strong>Roads<\/strong><\/h2>\n<p> Roads must connect important centers of habitation and trade, and they also follow some of the same rules as rivers \u2013 if not so strictly. Roads will often track a river or coastline, because these are easy places to build roads<a href=\"#\" title=\"compared to, say, directly over some mountains\">*<\/a> and most people will be living there anyway. Unlike rivers and coastlines, roads do occasionally need to pass over mountains. Humans will find the lowest and easiest possible routes through mountain ranges, and these routes will be important choke points for defense. <\/p>\n<h3>On Your Map:<\/h3>\n<p> If you choose to include roads, make sure you keep their style visually distinct from rivers, since they\u2019ll be drawn alongside each other quite a lot. Otherwise, just play connect-the-dots with cities, and make roads sparser and more squiggly over hills or mountains. <\/p>\n<h2><strong>Boundaries<\/strong><\/h2>\n<p> Political boundaries are a special case. For starters, they\u2019re drawn only onto maps and rarely marked on the face of the world itself. Boundaries have more in common with abstract human creations like \u201cjustice\u201d and \u201csin\u201d than with concrete<a href=\"#\" title=\"Sometimes literally.\">*<\/a> creations such as roads and towers. Yet they do follow certain rules, just like natural features and the less abstract trappings of civilization. <\/p>\n<p> Boundaries often follow obvious natural barriers like mountains and rivers. But river boundaries have a different flavor than boundaries marked by a high mountain range, empty desert, or large lake. Though a river marks an arbitrary line that a boundary can follow, by default an agricultural civilization will fill up as much of a river valley as it can. Thus a boundary along a river is an \u201cartificial\u201d divide, likely separating two halves of the same culture. Rivers are common places for imperial conquerors to divide land, but they are almost never natural cultural boundaries. Not to mention, rivers on flat ground shift over time \u2013 so they can be unreliable for demarcation. <\/p>\n<p> Boundaries that evolve naturally between two nations are more commonly mountains<a href=\"#\" title=\"often high mountain ridges dividing drainage basins; \u201canti-rivers\u201d\">*<\/a>, arbitrary lines in uninhabited deserts, choke-points on peninsulas, or wide bodies of water such as lakes or sea channels. Nations don\u2019t always align perfectly to regions divided by these features \u2013 but there\u2019s a good correlation. This correlation is strongest where the nations evolved naturally and weren\u2019t given arbitrarily-drawn boundaries by imperial powers carving up the world. <\/p>\n<h3>On Your Map:<\/h3>\n<p> Before placing boundaries, ask yourself whether the boundaries were placed by locals or by external empires. <\/p>\n<p> If these boundaries arose naturally between the local inhabitants, force your boundaries into uninhabited or difficult-to-traverse regions. If boundaries have been imposed by outsiders<a href=\"#\" title=\"Not just traditional imperialist powers, but also, for example, kingdoms that fracture as children of the dead queen inherit separate pieces. Anyone placing a boundary while looking at a map in a throne room has the temptation to follow natural demarcations like rivers. Anyone placing a boundary while living in a fertile river valley knows better.\">*<\/a> at some point in history,<a href=\"#\" title=\"As history progresses this becomes increasingly the case\">*<\/a> rivers and arbitrary straight lines should be more common. <\/p>\n<h2>Behold: The Realm of Badmap!<\/h2>\n<p> Getting a feel for how maps should look can take time, especially since there are so few hard-and-fast rules.<a href=\"#\" title=\"Except with rivers! Follow those damn river rules.\">*<\/a> That\u2019s where analyzing other maps \u2013 both real and fictional \u2013 can help. Most maps are pretty reasonable affairs; it\u2019s rare to find one so ridiculously ill-conceived that it can illustrate all the \u201cdon\u2019ts\u201d of mapmaking in one place. <\/p>\n<p> And so I\u2019ve made that exact thing for your viewing displeasure! Can you point out all the very bad, not good features on the following map of the Realm of Badmap?<a href=\"#\" title=\"Most of the flaws are geological in nature, but I apologize to the designers among us \u2013 I couldn\u2019t resist the choice of font. And the names!\">*<\/a> <\/p>\n<p> <img class=\"aligncenter wp-image-8749 size-full\" src=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1.png\" alt=\"Badmap\" width=\"660\" height=\"660\" srcset=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1.png 660w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-300x300.png 300w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-100x100.png 100w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-350x350.png 350w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-273x273.png 273w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-500x500.png 500w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-32x32.png 32w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-64x64.png 64w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-96x96.png 96w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1-128x128.png 128w\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/mythcreants.com\/blog\/crafting-plausible-maps\/\" data-jpibfi-post-title=\"Crafting Plausible Maps\" data-jpibfi-src=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/Badmap1.png\" \/><\/p>\n<p> I haven\u2019t included boundaries or roads (they\u2019re not always needed), but everything else is there. I\u2019ve used the amazing <a title=\"StarRaven's Sketchy Cartography Brushes\" href=\"https:\/\/starraven.deviantart.com\/art\/Sketchy-Cartography-Brushes-198264358\">Sketchy Cartography Brushes<\/a> by <a title=\"Deviant Art: StarRaven (User)\" href=\"https:\/\/starraven.deviantart.com\/\">Deviant Art\u2019s StarRaven<\/a>, which I highly recommend. I apologize for exploiting the brushes for this work of cartography evil. It had to be done. For, uh\u2026 for Science. <\/p>\n<h2>A Reasonable Map<\/h2>\n<p> Here\u2019s a variant on the same general landmass but with reasonable features. I\u2019ve tried to be as consistent as possible with the art style so that the only major differences are geological and climatological.<a href=\"#\" title=\"Though I couldn\u2019t live with myself if I used Papyrus \u2013 the evil, evil font \u2013 for the supposedly reasonable version.\">*<\/a> <\/p>\n<p> <img class=\"aligncenter wp-image-8750 size-full\" src=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands.png\" alt=\"ReasonableLands\" width=\"660\" height=\"660\" srcset=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands.png 660w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-300x300.png 300w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-100x100.png 100w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-350x350.png 350w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-273x273.png 273w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-500x500.png 500w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-32x32.png 32w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-64x64.png 64w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-96x96.png 96w, https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands-128x128.png 128w\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/mythcreants.com\/blog\/crafting-plausible-maps\/\" data-jpibfi-post-title=\"Crafting Plausible Maps\" data-jpibfi-src=\"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/01\/ReasonableLands.png\" \/><\/p>\n<p> This map still isn\u2019t \u201cperfect\u201d because there is no singular perfect; the permutations of acceptable are vast. In fact, it\u2019s statistically likely for a map of any given region to be <em>not<\/em> perfect. Even on Earth, curious \u201c<a title=\"Five Crazy Things That Look Like Fantasy \u2013 but Aren\u2019t\" href=\"https:\/\/mythcreants.com\/blog\/five-crazy-things-that-look-like-fantasy-but-arent\/\">un-Earth-like<\/a>\u201d features do occur here and there. So for detailed continent-sized maps, I advise consciously adding an odd feature or two. The map I drew has some borderline-believable features. The big southern river exits to the coast through a mountain range.<a href=\"#\" title=\"This happens on Earth \u2013 for example the Columbia and Fraser rivers \u2013 but it\u2019s uncommon.\">*<\/a> The barrier islands are a bit far from the coastline in the northwest. And the vegetation patterns could suggest a reversed (compared to Earth) rotation of the planet.<a href=\"#\" title=\"This last feature depends on the map\u2019s scale, which I have purposefully left ambiguous. It can be inferred from the northern tundra and southern desert that it\u2019s probably the size of a small continent.\">*<\/a> <\/p>\n<p> Be prepared to explain the existence of any anomaly on your map. Semi-plausible oddities can add interest to your world\u2019s geological history or aesthetic, or cause some confusion among your world\u2019s scholars. Why do those rivers seem to cross each other? Because of elaborate irrigation canals built by the Ancients making artificial under\/over paths for the sacred streams. Why are there only foothills on one side of the mountain range? Because of the way the continent buckled to form them, or possibly the Earth God was just lazy that day. <\/p>\n<p> But hopefully your maps will keep mostly to the plausible and believable, thanks in part to the guidelines and examples of this post. It would be a shame to craft a perfect story or game only to be mocked for your world\u2019s impossible river or misplaced mountains. <\/p> <\/div>\n<div class=\"action\">\n<p> P.S. Our bills are paid by our wonderful patrons. <a href=\"https:\/\/www.patreon.com\/mythcreants\">Could you chip in?<\/a> <\/p>\n<\/div>\n<p class=\"tags\">Read more about <a href=\"https:\/\/mythcreants.com\/blog\/tag\/maps\/\" rel=\"tag\">Maps<\/a> <\/p>\n<div class=\"ctx-subscribe-container ctx-personalization-container ctx_default_placement ctx-clearfix\"><div class=\"ctx-social-container ctx_default_placement ctx-clearfix\"><div class=\"ctx-module-container ctx_default_placement ctx-clearfix\">  \u00a0 <p class=\"clear\">\u00a0<\/p>\n<\/div><\/div><\/div>","created_at":"2018-11-07T16:53:11+0000","updated_at":"2018-11-07T16:53:11+0000","published_at":"2015-02-08T00:00:00+0000","published_by":["Brandon Kier"],"annotations":[],"mimetype":"text\/html","reading_time":17,"domain_name":"mythcreants.com","preview_picture":"https:\/\/mythcreants.com\/wp-content\/uploads\/2015\/02\/Map-Post-Title-Image-Resized.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":608,"title":"How to Use Quora to Increase Your Business Exposure","url":"https:\/\/www.socialmediaexaminer.com\/how-to-use-quora-to-increase-your-business-exposure\/","content":"<p><a href=\"https:\/\/www.socialmediaexaminer.com\/category\/how-to\/\" target=\"_blank\"><img class=\"alignright\" title=\"social media how to\" src=\"https:\/\/www.socialmediaexaminer.com\/images\/how-to-pose.png\" alt=\"social media how to\" width=\"190\" height=\"166\" \/><\/a>Want to build visibility for your business?<\/p><p>Have you considered Quora?<\/p><p>Quora is a useful place to share knowledge about your company and industry.<\/p><p>In this article I\u2019ll share how to successfully <strong>use Quora to gain exposure and manage your brand<\/strong>.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-exposure-480.png\" alt=\"use quora to increase business exposure\" width=\"480\" height=\"250\" \/><p class=\"wp-caption-text\">Discover how to use Quora to increase business exposure.<\/p><\/div><h3>Listen to this article:<\/h3><p><\/p><a href=\"http:\/\/traffic.libsyn.com\/smeshow\/SocialMediaExaminer-081.mp3\" class=\"external\" target=\"_blank\">http:\/\/traffic.libsyn.com\/smeshow\/SocialMediaExaminer-081.mp3<\/a><p class=\"powerpress_links powerpress_links_mp3\">Listen now: <a href=\"http:\/\/traffic.libsyn.com\/smeshow\/SocialMediaExaminer-081.mp3\" class=\"powerpress_link_pinw external\" target=\"_blank\" title=\"Play in new window\" rel=\"nofollow\">Play in new window<\/a> | <a href=\"http:\/\/traffic.libsyn.com\/smeshow\/SocialMediaExaminer-081.mp3\" class=\"powerpress_link_d external\" title=\"Download\" rel=\"nofollow\" download=\"SocialMediaExaminer-081.mp3\" target=\"_blank\">Download<\/a><\/p><h3>#1: Complete Your Profile to Increase Visibility<\/h3><p>Because\u00a0the Quora community has zero tolerance for self-promotion and spam, there are many factors to consider when setting up your presence.<\/p><p><a href=\"https:\/\/www.socialmediaexaminer.com\/is-quora-the-next-big-social-media-site\/\" target=\"_blank\">Quora<\/a> gives you tremendous exposure by <a href=\"http:\/\/www.quora.com\/What-is-Quoras-policy-on-displaying-profile-bios-and-topic-bios-next-to-answers\" target=\"_blank\" class=\"external\">showing your bio next to each answer<\/a> you post on the forum. Additionally, you can <strong>create multiple bios so each one appropriately fits the different topics<\/strong>. <strong>Add clickable links to the bio<\/strong>, as well.<\/p><p><a href=\"http:\/\/www.seosmarty.com\/crafting-an-about-page-that-proves-the-authority-in-you\/\" target=\"_blank\" class=\"external\">Your bio<\/a> is an excellent means of branding, and should be both professional and catchy. It should also be really short, since Quora will show only the first 50 characters of your whole tagline (your name and bio) above the answer you post. <strong>Mention your brand name as close to the beginning as you can and make the most of your 50 characters<\/strong>.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-bio-01.png\" alt=\"multiple bios on quora\" width=\"480\" height=\"470\" \/><p class=\"wp-caption-text\">The first 50 characters of your name and bio will be displayed above your answer on Quora. Users can click More to see the rest.<\/p><\/div><p>Quora users also see your professional tagline when you follow them, so the language you use affects how you build contacts on the platform.<\/p><p><strong>Complete your Quora account by adding a well-written detailed About Me section, areas of expertise, interests, cities you know about, schools and colleges you attended, previous companies and your other social media accounts<\/strong>.<\/p><p>To <strong>link your profile to your Facebook, LinkedIn and Twitter accounts<\/strong>, visit Settings on Quora. Just click on your name in the upper right.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand.png\" alt=\"connect additional social profiles on quora\" width=\"480\" height=\"268\" \/><p class=\"wp-caption-text\">Go to Settings to connect your social media accounts to Quora, so they are displayed on your profile page.<\/p><\/div><p>One more thing: Quora has an account <a href=\"http:\/\/www.quora.com\/What-does-it-mean-to-be-verified-on-Quora-What-are-verified-profiles-How-does-Quora-verify-accounts\" target=\"_blank\" class=\"external\">verification option<\/a>, but similar to Twitter, the verification is passive. You cannot request it. You must wait for Quora to approach you and make the offer to get a \u201cVerified\u201d label.<\/p><h3>#2: Understand Your Company Topic<\/h3><p>Quora is organized by topics. Whenever you search, you\u2019ll be invited to check out topics, which are discussions organized by a major keyword. In many cases that keyword is a brand name.<\/p><p>Quora is structured like Wikipedia, as pages can be created and edited by any user. Therefore, if you don\u2019t already have a company page on Quora, create one. Just <strong>start typing your company name on your About page<\/strong>. If it doesn\u2019t come up, Quora will suggest creating it as a topic.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-05.png\" alt=\"create a topic on quora\" width=\"480\" height=\"300\" \/><p class=\"wp-caption-text\">If your business isn\u2019t already a topic on Quora, create it by adding it to your profile.<\/p><\/div><p>After you <strong>add your company to your account, click the link to customize it<\/strong>.<strong> Add the logo and description<\/strong>.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-08.jpg\" alt=\"customize your profile on quora\" width=\"480\" height=\"189\" \/><p class=\"wp-caption-text\">After you add your brand topic, customize it with your logo and a description.<\/p><\/div><p><strong>Remember, other Quora users are able to edit your company page, so keep an eye on it regularly<\/strong>.<\/p><p>To further build up your company presence, <strong>request other people review the topic<\/strong>. It\u2019s another channel for your brand ambassadors to share their positive experience and spread the word.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-07.png\" alt=\"request topic reviews on quora\" width=\"480\" height=\"446\" \/><p class=\"wp-caption-text\">If applicable, see if others want to review your business by clicking Request.<\/p><\/div><p>To invite people you know to review your topic on Quora, just type in their names.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-09.png\" alt=\"invite topic reviews on quora\" width=\"480\" height=\"224\" \/><p class=\"wp-caption-text\">Invite your customers and ambassadors to review your topic on Quora.<\/p><\/div><p>If your business is already being discussed on Quora, you can still <strong>customize the topic page with an image and description and direct the conversation toward a positive sentiment<\/strong>.<\/p><p>Also to monitor your business and manage its online reputation, <strong>subscribe to the topic to see questions in your feed and pin the topic to always see it on your dashboard<\/strong>.<\/p><h3>#3: Be Active to Build Your Following<\/h3><p>Once you set up your account, go on Quora and examine the content. <strong>Read through various questions and answers, follow people who participate in discussions and upvote their updates<\/strong>.<\/p><p>Your upvotes appear on your public Quora profile, so this is a good way to show there\u2019s some relevant activity going on, even if your account is relatively new. Your follows and upvotes appear in users\u2019 Quora notifications, so you have a much higher chance of getting followed back.<\/p><div><a href=\"https:\/\/www.socialmediaexaminer.com\/smmworld\/?smesite=world19-site-middlearticlead\" target=\"_blank\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2018\/11\/SME-Large-Surfer-MiddleOfPageAdNov7.png\" alt=\"image\" \/><\/a><\/div><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-03-b.jpg\" alt=\"upvoted answers on quora\" width=\"480\" height=\"312\" \/><p class=\"wp-caption-text\">Follow and upvote other people\u2019s answers, so they see your activity and follow you back.<\/p><\/div><h3>#4: Use Quora as a Research and Brainstorming Tool<\/h3><p>When exploring Quora, <strong>take lots of notes to better understand the way your topic is covered on the platform<\/strong>. <strong>Seek out in-depth case studies and solid content ideas<\/strong>.<\/p><p>Quora search can be filtered to only display the questions, which is a way to get inspiration for <a href=\"https:\/\/www.socialmediaexaminer.com\/how-to-craft-headlines-that-draw-people-to-your-content\/\" target=\"_blank\">article headlines<\/a>.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-04.png\" alt=\"questions on quora\" width=\"480\" height=\"296\" \/><p class=\"wp-caption-text\">Scroll through popular questions in your industry to come up with article titles and content ideas.<\/p><\/div><p>Once you have ideas from your Quora research, <strong><a href=\"http:\/\/www.startablog123.com\/how-to-write-a-blog-post\/\" target=\"_blank\" class=\"external\">write articles for your website<\/a> that answer popular industry questions<\/strong>.<\/p><h3>#5: Explore Quora for Competitor Research<\/h3><p>Since the Quora community is continuously discussing different products and challenges, it\u2019s an excellent tool for performing <a href=\"https:\/\/www.socialmediaexaminer.com\/research-competition-with-social-media\/\" target=\"_blank\">competitor research<\/a>.<\/p><p><a href=\"http:\/\/www.quora.com\/On-Quora-how-do-I-follow-a-topic-and-how-does-topic-following-work\" target=\"_blank\" class=\"external\">Follow any topic<\/a> on Quora, so questions and answers posted in that topic will appear in your feed and you\u2019re able to stay on top of discussions around your competitors, brands in neighboring niches, partner companies and more.<\/p><p><strong>Search Quora for your competitors\u2019 brand names and browse the discussions<\/strong>. Then <strong>follow the relevant threads<\/strong>.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-10.png\" alt=\"search on quora\" width=\"480\" height=\"361\" \/><p class=\"wp-caption-text\">Do a navigational search on Quora to see what your competitors are talking about.<\/p><\/div><p><strong>Be careful when you chime in on threads that discuss your competitors<\/strong>. You don\u2019t want to post that people should check out your site instead of the competition. However, participating in a generic discussion with your own take never hurts.<\/p><p><strong>Pro tip<\/strong>: I am using <a href=\"http:\/\/www.cyfe.com\/\" class=\"external\" target=\"_blank\">Cyfe<\/a> to spy on my competitors on Quora as well as all over the world. I have a separate dashboard that monitors my competitors\u2019 recent mentions and stats.<\/p><h3>#6: Talk to Your Customers<\/h3><p><a href=\"https:\/\/www.socialmediaexaminer.com\/improve-customer-relationships-on-social-media\/\" target=\"_blank\">Communicating with your customers<\/a> is one of the greatest benefits of being active on Quora. Since people are already discussing you and your competitors, personally responding to topics may give you a competitive advantage. It also may win over new customers.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-11.png\" alt=\"comment on relevant questions on quora\" width=\"480\" height=\"437\" \/><p class=\"wp-caption-text\">If you\u2019re active on Quora, you can comment, giving you an advantage. Here, the founder of BuzzSumo chimes in on a discussion of BuzzSumo alternatives.<\/p><\/div><p>Treat your company image on Quora very carefully. Since your company name is in your bio, you\u2019re answering questions as the official representative. It\u2019s a huge responsibility that should not be taken lightly.<\/p><h3>#7: Share Your Quora Expertise on Other Social Media Networks<\/h3><p>Quora allows you to <strong>auto-share your answers on <a href=\"https:\/\/www.socialmediaexaminer.com\/twitter-marketing-resources\/\" target=\"_blank\">Twitter<\/a>, <a href=\"https:\/\/www.socialmediaexaminer.com\/essential-facebook-marketing-resources\/\" target=\"_blank\">Facebook<\/a>, <a href=\"https:\/\/www.socialmediaexaminer.com\/linkedin-marketing-resources\/\" target=\"_blank\">LinkedIn<\/a>, Tumblr and WordPress.com<\/strong>. To do this, just <strong>connect any or all of the above to your Quora account using <a href=\"http:\/\/www.quora.com\/settings\" target=\"_blank\" class=\"external\">Settings<\/a><\/strong>.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-02.png\" alt=\"cross promote answers on quora\" width=\"480\" height=\"241\" \/><p class=\"wp-caption-text\">Connect your social accounts to Quora to share answers on other social media platforms.<\/p><\/div><p>Sharing answers on other sites brings your expertise outside of Quora, while helping you build up your other accounts.<\/p><h3>#8: Analyze Stats to Determine What Works Best<\/h3><p>Quora\u2019s free analytics allow you to analyze the content you put on the platform.<\/p><p>This feature will help you find your most popular content in terms of clicks and see what works best with the Quora community.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-brand-12.png\" alt=\"analytics on quora\" width=\"480\" height=\"280\" \/><p class=\"wp-caption-text\">Use Quora\u2019s free analytics to determine popular content.<\/p><\/div><p><strong>Use Quora analytics to review your most clicked, shared and upvoted content within different periods of time<\/strong>: last 7 days, last 30 days, last 3 months and all-time.<\/p><p><strong>Conclusion<\/strong><\/p><p>According to the <a href=\"http:\/\/www.huffingtonpost.com\/jim-yu\/pinterest-and-quora-for-b_b_3046967.html\" target=\"_blank\" class=\"external\">Huffington Post<\/a>, most of the top brands are being discussed on Quora. In 2013, 84% of the top 100 brands had at least one discussion thread dedicated to them. Nowadays, it\u2019s hard to find a brand with no Quora mentions.<\/p><p>To get the most out of Quora, <strong>make sure your account is complete and professional, read through the discussions to get a feel for popular topics in your industry, communicate with others and share your answers on other social media sites<\/strong>.<\/p><p><strong>What do you think? Do you use Quora to build your brand and exposure? How often do you post? What other ways do you use Quora?<\/strong> Please share your tips in the comments.<\/p><div class=\"wp-caption alignnone\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2015\/04\/as-quora-exposure-pinterestimage.png\" alt=\"how to use quora to build business exposure\" width=\"210\" height=\"315\" \/><p class=\"wp-caption-text\">Tips for using Quora to increase business exposure.<\/p><\/div><div><a href=\"https:\/\/www.socialmediaexaminer.com\/smmworld\/?smesite=world19-site-endarticlead\" target=\"_blank\"><img src=\"https:\/\/www.socialmediaexaminer.com\/wp-content\/uploads\/2018\/11\/SME-Green-End-of-PageNov7.png\" alt=\"image\" \/><\/a><\/div>","created_at":"2018-11-07T16:53:11+0000","updated_at":"2018-11-07T16:53:11+0000","published_at":"2015-04-14T02:00:13+0000","published_by":["Ann Smarty"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"www.socialmediaexaminer.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":606,"title":"Getting Started with Clojure","url":"https:\/\/www.twilio.com\/blog\/2016\/02\/getting-started-with-clojure.html","content":"<div class=\"rich-text\"><p>JavaScript has been my go-to language for quite some time now and before that it was Ruby. Both of these languages have a lovely dash of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Functional_programming\">functional programming<\/a>\u00a0that I\u2019ve toyed with here and there. It seems functional concepts are becoming more and more prominent amongst the JS community with projects like <a href=\"https:\/\/github.com\/rackt\/redux\">Redux <\/a>and <a href=\"https:\/\/facebook.github.io\/immutable-js\/\">Immutable.js<\/a>\u00a0gaining popularity. I recently decided to take the dive into functional programming. A friend recommended the <a href=\"http:\/\/clojure.org\/\">Clojure<\/a>\u00a0programming language and learning it has been a real hoot.<\/p><p><img alt=\"owl.gif\" class=\"&quot; aligncenter\" data-alt=\"owl.gif\" data-embedtype=\"image\" data-format=\"left\" data-id=\"29281\" data-mce-contenteditable=\"false\" src=\"https:\/\/s3.amazonaws.com\/com.twilio.prod.twilio-docs\/original_images\/WjGl1KFBqTGi_OmkZLnRbCEN1ZTFta3F_uSzn4DOrogRFS6kzvu1Aj01pcP3Cw_LKAvAQK0_2-.png\" title=\"\" \/><\/p><p>Clojure was created by Rich Hickey who had a very strong <a href=\"http:\/\/clojure.org\/rationale\">rationale<\/a>\u00a0for its birth. The language is designed to compile down to bytecode and run on a virtual machine so it\u2019s not OS dependent. The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Java_virtual_machine\">JVM<\/a>\u00a0is the the most popular target but it can also compile down to <a href=\"https:\/\/github.com\/clojure\/clojurescript\">JavaScript<\/a>\u00a0or the <a href=\"https:\/\/github.com\/clojure\/clojure-clr\">Microsoft CLR<\/a>. Thanks to strong <a href=\"https:\/\/en.wikipedia.org\/wiki\/Interoperation\">interoperation<\/a>\u00a0support you can easily take advantage of your target platform\u2019s host language and features. We\u2019ll see an example of this later on.<\/p><p>In this post we\u2019re going to start with the basics of Clojure and its de facto build tool <a href=\"http:\/\/leiningen.org\/\">Leiningen<\/a>. We\u2019ll wrap up by sending our first SMS with Clojure.<\/p><p><em>I\u2019m absolutely still an apprentice so feel free to help out below in the comments if something can be improved. I\u2019ve found that the Clojure community is super friendly and willing to help. Heck the language is <a href=\"https:\/\/github.com\/clojure\/clojure\">community developed<\/a>.<\/em><\/p><h2 class=\"c1 c5\">Time To Get Technical<\/h2><p>Let\u2019s start by installing Leiningen. Not only does it have the best mascot logo but might just be the best build tool I\u2019ve ever used.<br \/><img alt=\"\" class=\"&quot; aligncenter\" data-alt=\"\" data-embedtype=\"image\" data-format=\"left\" data-id=\"29282\" data-mce-contenteditable=\"false\" src=\"https:\/\/s3.amazonaws.com\/com.twilio.prod.twilio-docs\/original_images\/F4WbU_2edhWB-cD5zcxZEazHBtGWJYI8NVguDj8IOT0bfENCxoTlYJpg1Da7nuOwzn_CYEXjRQ.png\" title=\"\" \/><\/p><p>The Clojure compiler is actually just a JAR file and Leiningen will take care of downloading it and doing all the heavy <code>javac<\/code> flag and classpath lifting.<\/p><p>Follow the <a href=\"http:\/\/leiningen.org\/#install\">instructions<\/a>\u00a0based on your platform. There is also an installer for Windows and a Homebrew formula if you\u2019re on a Mac.<\/p><p>We can confirm it\u2019s installed by running the following.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">lein repl\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>This will start up a Clojure Read Eval Print Loop (REPL) and give you a nice little prompt.<\/p><p>Clojure is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lisp_(programming_language)\">lisp<\/a>\u00a0which means it uses tons of parenthesis and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Polish_notation\">prefix notation<\/a>. Calls to functions start with the name of the function and then the arguments. So to use the <code><a href=\"https:\/\/clojuredocs.org\/clojure.core\/+\">+<\/a><\/code> function to add <code>2 + 2<\/code> we can type the following into our REPL.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(+ 2 2)\n; 4\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>This construct is called a <code>form<\/code> and\u00a0the \u201cwords\u201d inside of it are known as <a href=\"http:\/\/stackoverflow.com\/questions\/2320348\/symbols-in-clojure\">symbols<\/a>.\u00a0We can easily chain forms by adding more parenthesis.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(+ 2 (+ 2 2))\n; 6\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Notice we don\u2019t need to use commas to separate symbols \u2013 they are optional and treated as whitespace. The syntax will take some getting used to but I\u2019ve found it to be very expressive.<\/p><p>The <a href=\"http:\/\/clojure.org\/data_structures\">native data structures<\/a>\u00a0are powerful\u00a0and are used everywhere. The most common we\u2019ll be using are the vector and the map.<\/p><p>The vector is what we might think of as an array. It is a list of elements enclosed in square brackets. <code>[1 2 3 \"a\" \"b\" \"c\"]<\/code>. We can generally ask for the <code><a href=\"https:\/\/clojuredocs.org\/clojure.core\/first\">first<\/a><\/code> item out or the <code><a href=\"https:\/\/clojuredocs.org\/clojure.core\/rest\">rest<\/a><\/code> to perform a recursive\u00a0operation on the vector.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(first [1 2 3])\n; 1\n(rest [1 2 3])\n; (2 3)\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>We can also access an element at an index using the <code>get<\/code> function.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(get [\"Cats\" \"Dogs\" \"Fish\"] 1)\n; Dogs\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Let\u2019s iterate over a vector and call a function on each element using <code><a href=\"https:\/\/clojuredocs.org\/clojure.core\/doseq\">doseq<\/a><\/code>. This function\u00a0expects a symbol and a collection that it will iterate over and bind to the symbol. It then calls your provided form for each item.\u00a0We\u2019re also using the <code>println<\/code> function which, as you guessed, prints a line to the console along with the <code>str<\/code> function which concatenates strings.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(doseq [name [\"Betty\" \"Sue\" \"Amy\"]]\n  (println (str \"Hello, \" name)))\n; Hello, Betty\n; Hello, Sue\n; Hello, Amy\n; nil\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>The map is our dictionary structure in Clojure. It\u2019s created with curly braces and a set of keys and values separated by whitespace. This one\u2019s a little tricky to get used to since we\u2019re not visually associating the key and the value.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(get {\"name\" \"Bob Barker\" \"game\" \"The Price is Right\"} \"name\")\n; \"Bob Barker\"\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>It\u2019s common to use Clojure keywords as the key for a map. A keyword is a symbolic identifier that evaluates to itself. We denote a keyword with a colon.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(get {:name \"Bob Barker\" :game \"The Price is Right\"} :game)\n; The Price is Right\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Since this is so common we can actually use the keyword as a function and pass it a map to pull out the value.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(:name {:name \"Bob Barker\" :game \"The Price is Right\"})\n; Bob Barker\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>The last building block we\u2019ll learn is how to define our own function. To do this we use the <code><a href=\"https:\/\/clojuredocs.org\/clojure.core\/defn\">defn<\/a><\/code> macro. The form of this is <code>(defn name-of-function [args] (body))<\/code>. The following function <code>say-hello<\/code> will take a name and print out a personalized greeting.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(defn say-hello\n  [name]\n  (println (str \"Hello, \" name)))\n(say-hello \"Mary Poppins\")\n; Hello, Mary Poppins\n; nil\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Our function can take an undefined number of arguments using the <code>&amp;<\/code> operator. This causes all the arguments passed in to be squished into one variable as a collection.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(defn say-hello\n  [&amp; people]\n  (doseq [person people] (println (str \"Hello, \" person))))\n(say-hello \"Albert Einstein\" \"Grace Hopper\" \"Ada Lovelace\")\n; Hello, Albert Einstein\n; Hello, Grace Hopper\n; Hello, Ada Lovelace\n; nil\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Finally let\u2019s have a go at the Java interoperation that\u2019s available. We\u2019ll use <a href=\"https:\/\/docs.oracle.com\/javase\/8\/docs\/api\/java\/util\/Date.html\">Java\u2019s Date class<\/a>\u00a0to get the current time. To do this in Java you\u2019d run the following.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">System.out.println(new java.util.Date());\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>In Clojure we pass the full path and package name to the <code>new<\/code> function.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(new java.util.Date)\n; #inst \"2015-11-09T22:29:37.530-00:00\"\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>This is done so much that there is a nice shortcut around it. Note the trailing <code>.<\/code>.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(java.util.Date.)\n; #inst \"2015-11-09T22:29:37.530-00:00\"\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>We can call methods on the returned object with the .methodToCall\u00a0function.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(.toString (java.util.Date.))\n; Mon Nov 09 18:18:24 EST 2015\n; This syntax comes from the base `.` function\n(. (java.util.Date.) toString)\n; Mon Nov 09 18:18:55 EST 2015\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Even better, if we want to chain a bunch of Java method calls the <code>..<\/code> function makes this easy.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(.. (java.util.Date.) toString toUpperCase)\n; MON NOV 09 18:19:06 EST 2015\n<\/pre><\/div><\/div><div class=\"rich-text\"><h2 class=\"c1 c5\">Hello, World!<\/h2><p>We now have enough of a basis to create our first Clojure app. We\u2019ll leave the REPL and use Leiningen to generate our skeleton. The <code>quit<\/code> command should take you out of the REPL.<\/p><p>Create a new directory where you want to work and <code>cd<\/code> to it in your shell. Now let\u2019s unleash Leiningen on our new directory and have it create a blank app project for us.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">lein new app hello-clojure\ncd hello-clojure\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>There are two files in here that we\u2019ll be working with. The first is <code>project.clj<\/code> which is our build file for Leiningen and the second is our source file at <code>src\/hello_clojure\/core.clj<\/code>. We\u2019ll come back to the build file in a bit but for now let\u2019s open up the source file in your favorite editor.<\/p><p>The top two lines set up our namespace and tell the compiler to generate a class for our app since Java uses classes and we\u2019re running on the JVM. Following that is our main function definition which is the entry point to our app. The string after the declaration is used to document what this function does. The rest should look familiar from above.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(ns hello-clojure.core\n  (:gen-class))\n(defn -main\n  \"I don't do a whole lot ... yet.\"\n  [&amp; args]\n  (println \"Hello, World!\"))\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>We can tell Leiningen to compile and run our application by running<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">lein run\n# Hello, World\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Even better we can painlessly build an <a href=\"http:\/\/stackoverflow.com\/a\/11947093\/959953\">uberjar<\/a>\u00a0that we can distribute and run on a standard JVM.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">lein uberjar\n# Compiling hello-clojure.core\n# Created \/Users\/ezaneski\/Codez\/hello-clojure\/target\/uberjar uberjar\/hello-clojure-0.1.0-SNAPSHOT.jar\n# Created \/Users\/ezaneski\/Codez\/hello-clojure\/target\/uberjar\/hello-clojure-0.1.0-SNAPSHOT-standalone.jar\njava -jar target\/uberjar\/hello-clojure-0.1.0-SNAPSHOT-standalone.jar\n# Hello, World\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Wow that was absurdly easy.<\/p><h2 class=\"c1 c5\">Sending an SMS<\/h2><p>Our program is rad but doesn\u2019t do much. Let\u2019s fix that by making it send a text message using the <a href=\"https:\/\/www.twilio.com\/docs\/api\/rest\/sending-messages\">Twilio API<\/a>. You\u2019ll need your Twilio account credentials and a Twilio phone number. If you don\u2019t have an account yet you can sign up for a free one <a href=\"https:\/\/www.twilio.com\/try-twilio\">here<\/a>.<\/p><p>An awesome <a href=\"https:\/\/twitter.com\/owainlewis\">community member<\/a> has created a <a href=\"https:\/\/github.com\/owainlewis\/twilio\">Twilio Clojure library<\/a> that makes sending an SMS a breeze. To add the dependency to our app open up <code>project.clj<\/code> And add <code>[twilio-api \"1.0.1\"]<\/code> to the <code>:dependencies<\/code> vector.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(defproject hello-clojure \"0.1.0-SNAPSHOT\"\n  :description \"FIXME: write description\"\n  :url \"http:\/\/example.com\/FIXME\"\n  :license {:name \"Eclipse Public License\"\n            :url \"http:\/\/www.eclipse.org\/legal\/epl-v10.html\"}\n  :dependencies [[org.clojure\/clojure \"1.7.0\"]\n                 [twilio-api \"1.0.1\"]]\n  :main ^:skip-aot hello-clojure.core\n  :target-path \"target\/%s\"\n  :profiles {:uberjar {:aot :all}})\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>We can have Leiningen download the dependency by running the following.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">lein deps\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Open up <code>core.clj<\/code> again and replace it with the following.\u00a0You\u2019ll notice a that we\u2019re using an <code>@<\/code> in front of the call to <code>twilio\/send-sms<\/code>. It\u2019s a bit out of the scope of this post but that\u2019s actually a shortcut for the <code><a href=\"https:\/\/clojuredocs.org\/clojure.core\/deref\">deref<\/a><\/code> function which causes our program to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Blocking_(computing)\">block<\/a> and wait for an asynchronous operation to complete \u2013 such as the HTTP request to Twilio\u2019s API. If we don\u2019t block on the HTTP request, the program will stop running before it\u2019s made due to it being asynchronous.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">(ns hello-clojure.core\n  (:require [twilio.core :as twilio])\n  (:gen-class))\n(defn -main\n  [&amp; args]\n  (twilio\/with-auth \"YOUR_TWILIO_ACCOUNT_SID\" \"YOUR_TWILIO_AUTH_TOKEN\"\n    @(twilio\/send-sms\n       {:From \"YOUR_TWILIO_PHONE_NUMBER\"\n        :To \"YOUR_PHONE_NUMBER\"\n        :Body \"Hello from Clojure!\"})))\n<\/pre><\/div><\/div><div class=\"rich-text\"><p>Run the code one last time and take out your phone.<\/p><\/div><div class=\"markdown\"><div class=\"codehilite\"><pre class=\"twlo-code\">lein run\n<\/pre><\/div><\/div><div class=\"rich-text\"><p><a data-rel=\"lightbox-0\" href=\"https:\/\/s3.amazonaws.com\/com.twilio.prod.twilio-docs\/original_images\/7e0EvlBD7nxZu.gif\" title=\"\"><img alt=\"7e0EvlBD7nxZu\" class=\"alignnone size-full wp-image-16802\" data-alt=\"7e0EvlBD7nxZu\" data-embedtype=\"image\" data-format=\"left\" data-id=\"29283\" data-mce-contenteditable=\"false\" height=\"285\" src=\"https:\/\/s3.amazonaws.com\/com.twilio.prod.twilio-docs\/original_images\/7e0EvlBD7nxZu.gif\" width=\"500\" \/><\/a><\/p><h2 class=\"c1 c5\">Next Steps<\/h2><p>We\u2019ve explored some of the basics of Clojure and seen how we can take advantage of the Java interop. We also wrote a function that sends an SMS using an stellar community created library.\u00a0I think we can all agree that Leiningen lives up to its tagline of \u201c\u2026 automating Clojure projects without setting your hair on fire\u201d.<\/p><p>Hopefully this post has inspired you to give Clojure a try. It is a much different paradigm but pretty fun to work with.<\/p><p>For more reading and learning check out the book <a href=\"http:\/\/www.braveclojure.com\/\">Clojure for the Brave and True<\/a>\u00a0by Daniel Higginbotham. And the <a href=\"https:\/\/aphyr.com\/tags\/Clojure-from-the-ground-up\">Clojure from the ground up<\/a>\u00a0blog post series by Kyle Kingsbury.<\/p><p>Be sure to let me know what your thoughts on the language are!<\/p><p>Twitter: <a href=\"https:\/\/twitter.com\/eddiezane\">@eddiezane<\/a><\/p><\/div><div class=\"article-coauthors\"><p>Authors<\/p><ul class=\"article-coauthors-list\"><li>\n          <img src=\"https:\/\/www.gravatar.com\/avatar\/a804cc0d352d0c61f2ebf320429e8c97?d=https%3A%2F%2Fwww.twilio.com%2Fmarketing%2Fbundles%2Fcompany-brand%2Fimg%2Flogos%2Fred%2Ftwilio-mark-red.png&amp;s=40\" alt=\"image\" \/><a href=\"https:\/\/www.twilio.com\/blog\/author\/ezaneski\">Eddie Zaneski<\/a>\n        <\/li>\n        \n      <\/ul><\/div><div class=\"build_more\"><ul class=\"build_more__list build_more__list\"><li class=\"build_more__list-item\">\n        <a href=\"https:\/\/www.twilio.com\/docs\/video\/javascript\" class=\"build_more__link\">JavaScript Platform Overview<\/a>\n      <\/li>\n      \n      <li class=\"build_more__list-item\">\n        <a href=\"https:\/\/www.twilio.com\/docs\/voice\/client\/javascript\/sdk-11-12-migration-guide\" class=\"build_more__link\">Twilio Client JavaScript SDK 1.1 &amp; 1.2 Migration Guide<\/a>\n      <\/li>\n      \n      <li class=\"build_more__list-item\">\n        <a href=\"https:\/\/www.twilio.com\/blog\/2018\/04\/hosting-live-webinars-javascript-nodejs-adonisjs-twilio-video.html\" class=\"build_more__link\">Hosting Live Webinars with JavaScript, Node.js, AdonisJs and Twilio Programmable Video<\/a>\n      <\/li>\n      \n      <li class=\"build_more__list-item\">\n        <a href=\"https:\/\/www.twilio.com\/blog\/2014\/07\/extending-the-reach-of-browser-based-communications-introducing-twilio-client-javascript-sdk-v1-2-nt.html\" class=\"build_more__link\">Extending The Reach of Browser-Based Communications: Introducing Twilio Client Javascript SDK v1.2<\/a>\n      <\/li>\n      \n      <li class=\"build_more__list-item\">\n        <a href=\"https:\/\/www.twilio.com\/blog\/2012\/12\/building-a-real-time-sms-voting-app-part-2-socket-io-and-highcharts-js.html\" class=\"build_more__link\">Building a Real-time SMS Voting App Part 2: Socket.io and Highcharts.js<\/a>\n      <\/li>\n      \n    <\/ul><\/div><noscript><div>Please enable JavaScript to view the <a href=\"https:\/\/disqus.com\/?ref_noscript\" rel=\"nofollow\">comments powered by Disqus.<\/a><\/div><\/noscript>","created_at":"2018-11-07T16:53:10+0000","updated_at":"2018-11-07T16:53:10+0000","published_at":"2016-02-01T00:00:00+0000","published_by":["Eddie Zaneski"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"www.twilio.com","preview_picture":"https:\/\/s3.amazonaws.com\/com.twilio.prod.twilio-docs\/original_images\/GettingStartedWithClojureFeatured.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":605,"title":"overtone\/shadertone","url":"https:\/\/github.com\/overtone\/shadertone","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><pre>                     __              __          __\n               _____\/ \/_  ____ _____\/ \/__  _____\/ \/_____  ____  ___\n              \/ ___\/ __ \\\/ __ `\/ __  \/ _ \\\/ ___\/ __\/ __ \\\/ __ \\\/ _ \\\n             (__  ) \/ \/ \/ \/_\/ \/ \/_\/ \/  __\/ \/  \/ \/_\/ \/_\/ \/ \/ \/ \/  __\/\n            \/____\/_\/ \/_\/\\__,_\/\\__,_\/\\___\/_\/   \\__\/\\____\/_\/ \/_\/\\___\/\n                                                                                 .\n<\/pre>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/overtone\/shadertone\/raw\/master\/readme_header.jpg\"><img src=\"https:\/\/github.com\/overtone\/shadertone\/raw\/master\/readme_header.jpg\" alt=\"HeaderImage\" \/><\/a><\/p>\n<p>A Clojure library designed to mix\nMusical Synthesis via <a href=\"https:\/\/github.com\/overtone\/overtone\">Overtone<\/a> and\nOpenGL shaders a la <a href=\"http:\/\/www.shadertoy.com\" rel=\"nofollow\">Shadertoy<\/a>.<\/p>\n<p><a href=\"https:\/\/github.com\/overtone\/overtone\">Overtone<\/a> is a toolkit for\ndesigning sound synthesizers and coding musical performances.  Based\non <a href=\"http:\/\/clojure.org\/\" rel=\"nofollow\">Clojure<\/a> and\n<a href=\"http:\/\/supercollider.sourceforge.net\/\" rel=\"nofollow\">Supercollider<\/a>, it is a\npowerful combination of a general-purpose, interactive, multi-threaded\nlanguage built on the JVM with a state of the art, realtime sound\nsynthesis server.<\/p>\n<p><a href=\"http:\/\/www.shadertoy.com\" rel=\"nofollow\">Shadertoy<\/a> is a website that exhibits some\namazing dynamic imagery--all procedurally generated via WebGL.  It is\na fantastic example of how limitations and constraints allow you to\nfocus and create great artwork.  All the imagery is created via a\nfunctional per-pixel program called a Fragment Shader that is provided\nonly a few inputs like the pixel position and current time.  The\nimagery is generated inside your Graphics Processing Unit (GPU) and\noffloads your CPU.  Visuals are more powerful than words, here.  Take\na look at some fun examples:\n<a href=\"https:\/\/www.shadertoy.com\/view\/MsXGz8\" rel=\"nofollow\">isolines<\/a>,\n<a href=\"https:\/\/www.shadertoy.com\/view\/4dX3Rn\" rel=\"nofollow\">flower<\/a>, and\n<a href=\"https:\/\/www.shadertoy.com\/view\/MslGRn\" rel=\"nofollow\">electron<\/a>.<\/p>\n<p>This library provides a way for you to create music and visuals that\nwork together.  The example inspiration for Shadertone was <a href=\"http:\/\/vimeo.com\/51993089\" rel=\"nofollow\">this\nlive-coding video by h3xl3r<\/a>.  H3xl3r edits\nthe fragment shader visual code live while the music plays.  This\nlibrary gives you the ability to hack on <em>both<\/em> the visual and audio\ncode at the same time.<\/p>\n<p>I've created a few screencast demos that you can watch:<\/p>\n<ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=_8T15N3ZvYc\" rel=\"nofollow\">Basic features overview<\/a> from the <a href=\"https:\/\/github.com\/overtone\/shadertone\/blob\/master\/examples\/00demo_intro_tour.clj\">intro tour<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=X7EMPYlGgFc\" rel=\"nofollow\">Livecoding demo<\/a> which you can try via <code>lein run<\/code><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=kyL3xc7MzR0\" rel=\"nofollow\">A quick teaser<\/a> from the <a href=\"https:\/\/github.com\/overtone\/shadertone\/blob\/master\/examples\/02demo_disco.clj\">disco example<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=UMg8Td5Gqhk\" rel=\"nofollow\">Video preview<\/a> showing the very early stages of shadertone.<\/li>\n<\/ul><p><a href=\"http:\/\/meta-ex.com\/\" rel=\"nofollow\">Meta-eX<\/a> has used Shadertone in live\nperformance.  Here are a few videos from their May, 2013 Kiev #hotcode\nconference performance that show how Shadertone can be used: <a href=\"http:\/\/vimeo.com\/67487486\" rel=\"nofollow\">after\nparty 1<\/a>, <a href=\"http:\/\/vimeo.com\/67487485\" rel=\"nofollow\">main\nconference<\/a>, <a href=\"http:\/\/www.youtube.com\/watch?v=CjFCSTQbJx0\" rel=\"nofollow\">after party\n1<\/a>, <a href=\"http:\/\/www.youtube.com\/watch?v=CjFCSTQbJx0\" rel=\"nofollow\">after party\n2<\/a>.  Latest news &amp; videos\ncan always be found at <a href=\"http:\/\/meta-ex.com\" rel=\"nofollow\">their website<\/a><\/p>\n<p><a href=\"http:\/\/www.repl-electric.com\/\" rel=\"nofollow\">Repl Electric<\/a> has also used\nShadertone in performance.  Here is a wonderful recreation of <a href=\"http:\/\/vimeo.com\/95988263\" rel=\"nofollow\">\"The\nStars\" livecoded on Vimeo<\/a>.<\/p>\n<p>If you're interested, here's <a href=\"https:\/\/groups.google.com\/forum\/?fromgroups=#!topic\/overtone\/7bQSJUUviBw\" rel=\"nofollow\">the original\nannouncement<\/a>\non the Overtone Google Group.<\/p>\n<p>I hope this library allows you to create some happiness.  Enjoy!<\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>The library is just coming together, so expect change.  There are two\nmain ways to use the code.  If you want the latest code or are just\nexploring what shadertone can do, you should clone the repo from\ngithub.  But, if you have a project idea and want to use shadertone as\npart of that project, you can specify the current version in your\nLeiningen project.clj to download the library from clojars.<\/p>\n<h3><a id=\"user-content-option-1-clone-this-repository\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-1-clone-this-repository\"><\/a>Option 1: clone this repository<\/h3>\n<ol><li>See top of <a href=\"https:\/\/github.com\/overtone\/shadertone\">https:\/\/github.com\/overtone\/shadertone<\/a> for details on how to clone the repo.<\/li>\n<li>Bring up your favorite REPL and step through the demo files in the examples directory.  There are examples of all of the various shadertone features in the demos.<\/li>\n<li>When you want to create your own code, make some files like examples\/yourcode.clj and examples\/yourshader.glsl.  See below.<\/li>\n<li>Start with something basic in both files to be certain it works.<\/li>\n<li>You can edit both files \"live\".  Shadertone will watch for edits to your active shader and load them when you save the file.<\/li>\n<\/ol><h3><a id=\"user-content-option-2-download-from-clojars\" class=\"anchor\" aria-hidden=\"true\" href=\"#option-2-download-from-clojars\"><\/a>Option 2: download from clojars<\/h3>\n<p>In your project.clj, add <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/ecd01422fb9653eb57a6a27674578b05de1c139b\/687474703a2f2f636c6f6a6172732e6f72672f736861646572746f6e652f6c61746573742d76657273696f6e2e737667\"><img src=\"https:\/\/camo.githubusercontent.com\/ecd01422fb9653eb57a6a27674578b05de1c139b\/687474703a2f2f636c6f6a6172732e6f72672f736861646572746f6e652f6c61746573742d76657273696f6e2e737667\" alt=\"Clojars Project\" data-canonical-src=\"http:\/\/clojars.org\/shadertone\/latest-version.svg\" \/><\/a> to your <code>:dependencies<\/code>.<\/p>\n<p>See <a href=\"https:\/\/github.com\/rogerallen\/sot\">https:\/\/github.com\/rogerallen\/sot<\/a> for a simple example of using\nshadertone via Leiningen.  Use the example directory in this repository for ideas.<\/p>\n<h3><a id=\"user-content-writing-fragment-shaders\" class=\"anchor\" aria-hidden=\"true\" href=\"#writing-fragment-shaders\"><\/a>Writing Fragment Shaders<\/h3>\n<p><a href=\"http:\/\/www.opengl.org\/wiki\/Fragment_Shader\" rel=\"nofollow\">Fragment Shaders<\/a> are\ncode that is executed for every pixel displayed in the window.  The\noutput is the color of that pixel, which you store in\n<code>gl_FragColor<\/code>. To create different imagery, you take the current\npixel's position (from <code>gl_FragCoord<\/code>) and use that as input.  To have\nOvertone affect the color, you'll want to use inputs like\n<code>iOvertoneVolume<\/code>.  The syntax is C-like and given the math-processing\npower of modern GPUs, it is somewhat astonishing what you can do in\n1\/60th of second.<\/p>\n<p>The code needs to be in a GLSL format with a <code>main()<\/code> routine that\nreturns the <code>gl_FragColor<\/code>.  Shadertoy.com has changed their\nsyntax to make translation just a bit more work.  You will need to\ntranslate Image shaders formatted like this:<\/p>\n<div class=\"highlight highlight-source-c\"><pre>void mainImage( out vec4 fragColor, in vec2 fragCoord ) {\n  ...\n  fragColor = vec4(...);\n}<\/pre><\/div>\n<p>To GLSL fragment shaders formatted like this:<\/p>\n<div class=\"highlight highlight-source-c\"><pre>void main(void) {\n  ...\n  gl_FragColor = vec4(...);\n}<\/pre><\/div>\n<p>One of the creators of <a href=\"http:\/\/www.shadertoy.com\" rel=\"nofollow\">www.shadertoy.com<\/a>, I\u00f1igo Qu\u00edlez, has a website\nthat explains a variety of techniques to create interesting Fragment\nShaders at <a href=\"http:\/\/www.iquilezles.org\/www\/index.htm\" rel=\"nofollow\">http:\/\/www.iquilezles.org\/www\/index.htm<\/a>.  He also has some\ninteresting live coding demos here\n<a href=\"http:\/\/www.iquilezles.org\/live\/index.htm\" rel=\"nofollow\">http:\/\/www.iquilezles.org\/live\/index.htm<\/a>.<\/p>\n<h3><a id=\"user-content-shader-inputs\" class=\"anchor\" aria-hidden=\"true\" href=\"#shader-inputs\"><\/a>Shader Inputs<\/h3>\n<p>Here are the inputs you have available to your fragment shader.<\/p>\n<h4><a id=\"user-content-from-wwwshadertoycom\" class=\"anchor\" aria-hidden=\"true\" href=\"#from-wwwshadertoycom\"><\/a>From <a href=\"http:\/\/www.shadertoy.com\" rel=\"nofollow\">www.shadertoy.com<\/a>:<\/h4>\n<p>There are several implicit values that Shadertone provides, just like\nthe website.  You will have to define other inputs at the top of your\nshader.<\/p>\n<pre>uniform vec3      iResolution;     \/\/ viewport resolution (in pixels)\nuniform float     iGlobalTime;     \/\/ shader playback time (in seconds)\nuniform float     iChannelTime[4]; \/\/ channel playback time (in seconds)\nuniform vec3      iChannelResolution[4]; \/\/ channel width, height, 1\nuniform vec4      iMouse;          \/\/ mouse pixel coords. xy: current (if MLB down), zw: click\nuniform samplerXX iChannel0..3;    \/\/ input channel. XX = 2D\/Cube\nuniform vec4      iDate;           \/\/ (year, month, day, time in seconds)\n<\/pre>\n<h4><a id=\"user-content-overtone-inputs\" class=\"anchor\" aria-hidden=\"true\" href=\"#overtone-inputs\"><\/a>Overtone inputs:<\/h4>\n<p>There are three inputs from Overtone that you can use.<\/p>\n<p>First is a tap on the output volume.  It is sent to the variable\n<code>iOvertoneVolume<\/code> which you should define at the top of your shader if\nyou start your window with <code>shadertone.tone\/start<\/code>.<\/p>\n<pre>uniform float iOvertoneVolume; \/\/ tap of system volume\n<\/pre>\n<p>Next, are two special textures.  First, if you use the keyword\n:overtone-audio as a texture name, Overtone will send the output\nsound's frequency spectrum (FFT) and audio waveform data to the\niChannel texture you specify.  This is\nsimilar to <a href=\"http:\/\/www.shadertoy.com's\" rel=\"nofollow\">www.shadertoy.com's<\/a> audio textures.  The first row of the\ntexture contains the FFT data.  The second row of the texture contains\nthe current waveform data.<\/p>\n<p>Second, if you use the :previous-frame keyword as a texture name,\nShadertone will capture the framebuffers you render and allow you to\nuse this as input for the next frame.<\/p>\n<p>As an example, if you want these textures data to go to the first two\ntextures, add the argument\n<code>:textures [:overtone-audio :previous-frame]<\/code>\nto your <code>shadertone.tone\/start<\/code> call.  Then, in your\nglsl code use <code>texture2D(iChannel0,uv).r<\/code> to access the audio and\n<code>texture2D(iChannel1,uv).rgb<\/code> to access the previous frame.<\/p>\n<p>But that isn't all.  If you want to create your own inputs from\nOvertone to Shadertone, you can create atoms that hold either a\nfloating-point number or a vector of 1-4 values.<\/p>\n<p>For example, this clojure code<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def my-rgb (atom [0.3 0.1 0.5]))\n(t\/start \"shaders\/rgb.glsl\" :user-data { \"iRGB\" my-rgb})<\/pre><\/div>\n<p>Will adjust this shader code live<\/p>\n<div class=\"highlight highlight-source-c\"><pre>uniform vec3 iRGB;\nvoid main(void)\n{\n    gl_FragColor = vec4(iRGB, 1.0);\n}<\/pre><\/div>\n<p>Via a call like this:<\/p>\n<pre>(swap! my-rgb (fn [x] [0.55 0.95 0.75]))\n<\/pre>\n<h5><a id=\"user-content-synth-inputs\" class=\"anchor\" aria-hidden=\"true\" href=\"#synth-inputs\"><\/a>Synth Inputs<\/h5>\n<p>You can <code>tap<\/code> a synth value and easily communicate that\nvalue to your GLSL fragment shader.  See the \"vvv\" synth example in\nthe 00_demo_intro_tour.clj demo and also the core.clj usage.<\/p>\n<p>In a nutshell, you define your synth and add a call to <code>(tap \"a\" 60 a)<\/code> where <code>\"a\"<\/code> is the name of your tap, it is updated 60 times per\nsecond and the <code>a<\/code> synth variable is some interesting value you want\nto communicate.<\/p>\n<p>Then in your <code>start<\/code> call, the <code>:user-data<\/code> map should include a\nreference to that <code>tap<\/code>.  Like <code>:user-data { \"iA\" (atom {:synth v :tap \"a\"}) }<\/code> to call the value <code>iA<\/code> as input to your GLSL shader, if your\nsynth was instantiated as <code>v<\/code>.<\/p>\n<p>Finally, in your GLSL fragment shader, define your new input value as\n<code>uniform float iA;<\/code> and use <code>iA<\/code> to control your shader in interesting\nways.<\/p>\n<h3><a id=\"user-content-clojure-api\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure-api\"><\/a>Clojure API<\/h3>\n<p>To use Shadertone, the main routines you will use are in the\n<code>shadertoy.tone<\/code> namespace.  Use <code>start<\/code> or <code>start-fullscreen<\/code> to bring\nup a window.  Use <code>stop<\/code> to bring it down.  Calling <code>start<\/code> while a\nwindow is active will stop it gracefully before starting up the new\nwindow.<\/p>\n<p><code>start<\/code> takes a few parameters<\/p>\n<ul><li>the first parameter is either the filename of a glsl shader or an atom containing a string with the glsl code.<\/li>\n<li><code>:width<\/code> &amp; <code>:height<\/code> are the window size in pixels<\/li>\n<li><code>:title<\/code> is the window title<\/li>\n<li><code>:display-sync-hz<\/code> is for setting the refresh rate of the window.  By default it is 60 Hz.<\/li>\n<li><code>:textures<\/code> are a vector of texture filenames or the\nspecial texture keywords.  Up to 4 textures are allowed.<\/li>\n<li><code>:user-data<\/code> is a map of strings to atoms.  The strings must match a\nuniform variable name in your shader.  The atoms are used to\ncommunicate values to those uniform variable names.<\/li>\n<li><code>:user-fn<\/code> is for complete custom control of the application.\nNormally, you should not need to override this, but you can if you\nneed to.<\/li>\n<\/ul><p><code>start-fullscreen<\/code> takes just a subset of these, but the meanings are\nthe same: <code>:display-sync-hz<\/code>, <code>:textures<\/code>, <code>:user-data<\/code>, and <code>:user-fn<\/code><\/p>\n<p>The base api is in <code>shadertone.shader<\/code> and does not\ndepend on Overtone.  This is useful when you don't need sound accompaniment and\ncould be useful for other interactive ideas.<\/p>\n<h3><a id=\"user-content-lisp-like-glsl\" class=\"anchor\" aria-hidden=\"true\" href=\"#lisp-like-glsl\"><\/a>Lisp-like GLSL<\/h3>\n<p>You can program your GLSL in a lisp-like language.  It is a simple, straight translation\nlayer from lisp to a GLSL string.  For those hoping for Clojure instead of lisp,\nI welcome your suggestions and help, but a Clojure-to-C compiler was way too much work for me at this time.<\/p>\n<p>See <a href=\"https:\/\/github.com\/overtone\/shadertone\/blob\/master\/examples\/03demo_translate.clj\">https:\/\/github.com\/overtone\/shadertone\/blob\/master\/examples\/03demo_translate.clj<\/a> for an example of how you can use this.\nIn addition, there are tests in <a href=\"https:\/\/github.com\/overtone\/shadertone\/blob\/master\/test\/shadertone\/translate_test.clj\">https:\/\/github.com\/overtone\/shadertone\/blob\/master\/test\/shadertone\/translate_test.clj<\/a> that could be informative.<\/p>\n<p>To access this functionality, import the <code>shadertone.translate<\/code> namespace and use <code>defshader<\/code> to compile your\nlisp into a string for passing to the <code>start<\/code> function.<\/p>\n<h4><a id=\"user-content-special-forms-and-how-they-relate-to-glsl-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#special-forms-and-how-they-relate-to-glsl-code\"><\/a>Special forms and how they relate to GLSL code<\/h4>\n<ul><li>define functions\n<code>(defn &lt;return-type&gt; &lt;function-name&gt; &lt;function-args-vector&gt; &lt;body-stmt1&gt; ... )<\/code><\/li>\n<li>function calls\n<code>(&lt;name&gt; &lt;arg1&gt; &lt;arg2&gt; ... )<\/code><\/li>\n<li>variable creation\/assignment\n<code>(uniform &lt;type&gt; &lt;name&gt;)<\/code>\n<code>(setq &lt;type&gt; &lt;name&gt; &lt;statement&gt;)<\/code>\n<code>(setq &lt;name&gt; &lt;statement&gt;)<\/code><\/li>\n<li>for(;;) {}\n<code>(forloop [ &lt;init-stmt&gt; &lt;test-stmt&gt; &lt;step-stmt&gt; ] &lt;body-stmt1&gt; ... )<\/code><\/li>\n<li>while() {}\n<code>(while &lt;test-stmt&gt; &lt;body-stmt1&gt; ... )<\/code><\/li>\n<li>if() {}\n<code>(if &lt;test&gt; &lt;stmt&gt;)<\/code>\n<code>(if &lt;test&gt; (do &lt;body-stmt1&gt; ...))<\/code><\/li>\n<li>if() {} else {}\n<code>(if &lt;test&gt; &lt;stmt&gt; &lt;else-stmt&gt;)<\/code>\n<code>(if &lt;test&gt; (do &lt;body-stmt1&gt; ...) (do &lt;else-stmt1&gt; ...))<\/code><\/li>\n<li>switch () { case integer: ... break; ... default: ... }\n<code>(switch &lt;test&gt; &lt;case-int-1&gt; &lt;case-stmt-1&gt; ...)<\/code>\ncases can only be integer or :default keyword<\/li>\n<li>break;\n<code>(break)<\/code><\/li>\n<li>continue;\n<code>(continue)<\/code><\/li>\n<li>return value;\n<code>(return &lt;statement&gt;)<\/code><\/li>\n<\/ul><h3><a id=\"user-content-reading-pixels-from-the-framebuffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#reading-pixels-from-the-framebuffer\"><\/a>Reading Pixels from the Framebuffer<\/h3>\n<p>You can read back a pixel value from your frame either via\n<code>(shadertone.shader\/pixel)<\/code> function or by accessing the\n<code>@shadertone.shader\/pixel-value<\/code> atom directly.  With this feature,\nyour fragment-shader could control your synths!  Thanks to Circu Virtu\non the Google Group for this idea.<\/p>\n<p>Enable this feature via <code>(pixel-read-enable! x y)<\/code> where x and y are\nvalid locations within your window.  Be careful as there is no error\nchecking here.<\/p>\n<p>Disable this access via <code>(pixel-read-disable!)<\/code><\/p>\n<h2><a id=\"user-content-changes\" class=\"anchor\" aria-hidden=\"true\" href=\"#changes\"><\/a>Changes<\/h2>\n<ul><li>\n<p><strong>0.2.6 - Released Sept ?, 2015<\/strong><\/p>\n<ul><li>Bugfix: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/31\">Issue #31<\/a> - Fix unary math ops and note the shadertoy.com syntax change.  Thanks to <a href=\"https:\/\/github.com\/hlolli\">hlolli<\/a><\/li>\n<\/ul><\/li>\n<li>\n<p><strong>0.2.5 - Released May 26, 2015<\/strong><\/p>\n<\/li>\n<li>\n<p><strong>0.2.4 - Released Dec 11, 2014<\/strong><\/p>\n<ul><li>Enhancement: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/16\">Issue #16<\/a> - Add iChannelResolution.<\/li>\n<li>Enhancement: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/18\">Issue #18<\/a> - Add pixel read feature.  Thanks to Circu Virtu.<\/li>\n<li>Enhancement: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/14\">Issue #14<\/a> - No external change since LWJGL is limited to one window, but code has been improved by reducing the use of global state a bit.<\/li>\n<li>Enhancement: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/23\">Issue #23<\/a> - Generally improve error handling &amp; robustness.  Thanks to <a href=\"https:\/\/github.com\/josephwilk\">josephwilk<\/a><\/li>\n<\/ul><\/li>\n<li>\n<p><strong>0.2.3 - Released Mar 8, 2014<\/strong><\/p>\n<\/li>\n<li>\n<p>Enhancement: Update to LWJGL 2.9.1<\/p>\n<\/li>\n<li>\n<p>Bugfix: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/22\">Issue #22<\/a> - Fix support for Intel Graphics chips.  Thanks to <a href=\"https:\/\/github.com\/josephwilk\">josephwilk<\/a><\/p>\n<\/li>\n<li>\n<p>Bugfix: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/20\">Issue #20<\/a> - Fix fullscreen support on Retina Macs.  Thanks to <a href=\"https:\/\/github.com\/johnjelinek\">johnjelinek<\/a><\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/overtone\/shadertone\/issues?milestone=4\">0.2.2<\/a> - Released Nov 25, 2013<\/strong><\/p>\n<\/li>\n<li>\n<p>Update: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/19\">Issue #19<\/a> - Overtone 0.9.1<\/p>\n<\/li>\n<li>\n<p>Bugfix: uninitialized <code>rotx<\/code> variable in simplecube.glsl example.<\/p>\n<\/li>\n<li>\n<p><strong>0.2.1 - Released Sep 1, 2013<\/strong><\/p>\n<\/li>\n<li>\n<p>Bugfix: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/15\">Issue #15<\/a> - Automatic file reloading fail on cygwin.<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/overtone\/shadertone\/issues?milestone=2&amp;state=closed\">0.2.0<\/a> - Released Aug 1, 2013<\/strong><\/p>\n<\/li>\n<li>\n<p>Enhancement: Add ability to 'tap' a synth and communicate that to a shader via :user-data<\/p>\n<\/li>\n<li>\n<p>Enhancement: create shaders in a lisp-like language (Issue #1)<\/p>\n<\/li>\n<li>\n<p>Enhancement: \"lein run\" does something now! Added redFrik tweet-inspired demo.<\/p>\n<\/li>\n<li>\n<p>Enhancement: Increased resolution of waveform and FFT texture to 4096 from 512.<\/p>\n<\/li>\n<li>\n<p>Enhancement: Add Spectrogram example using FFT texture and :previous-frame texture.<\/p>\n<\/li>\n<li>\n<p>Enhancement: Update to LWJGL 2.9.0<\/p>\n<\/li>\n<li>\n<p>Bugfix: fixup native library handling (Issues <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/8\">#8<\/a>, <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/13\">#13<\/a>)<\/p>\n<\/li>\n<li>\n<p>Bugfix: <a href=\"https:\/\/github.com\/overtone\/shadertone\/issues\/12\">Issue #12<\/a> - Mac window positioning broken on LWJGL 2.9.0<\/p>\n<\/li>\n<li>\n<p><strong><a href=\"https:\/\/github.com\/overtone\/shadertone\/issues?milestone=1&amp;state=closed\">0.1.0<\/a> - Released May 5, 2013<\/strong><\/p>\n<\/li>\n<li>\n<p>Initial Release<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-acknowledgements\" class=\"anchor\" aria-hidden=\"true\" href=\"#acknowledgements\"><\/a>Acknowledgements<\/h2>\n<p>Many thanks to those who have helped directly and indirectly to make\nShadertone.  Specifically, I'd like to thank:<\/p>\n<ul><li>Sam Aaron for helping organize the code and developing the public interface.<\/li>\n<li>I\u00f1igo Qu\u00edlez for creating Shadertoy.<\/li>\n<li>Fredrik Olofsson (@redFrik) for his fantastic sctweets.<\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2013-2014 Roger Allen and <a href=\"https:\/\/github.com\/overtone\/shadertone\/contributors\">other\ncontributors<\/a>.<\/p>\n<p>Permission is hereby granted, free of charge, to any person obtaining\na copy of this software and associated documentation files (the\n\"Software\"), to deal in the Software without restriction, including\nwithout limitation the rights to use, copy, modify, merge, publish,\ndistribute, sublicense, and\/or sell copies of the Software, and to\npermit persons to whom the Software is furnished to do so, subject to\nthe following conditions:<\/p>\n<p>The above copyright notice and this permission notice shall be\nincluded in all copies or substantial portions of the Software.<\/p>\n<p>THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND,\nEXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF\nMERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND\nNONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE\nLIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION\nOF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION\nWITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.<\/p>\n<p>For full license information, see the LICENSE file.<\/p>\n<\/article>","created_at":"2018-11-07T16:53:08+0000","updated_at":"2018-11-07T16:53:08+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/373359?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":602,"title":"Beautiful Racket by Matthew Butterick","url":"https:\/\/beautifulracket.com\/","content":"<p>Thank you for your comment<\/p><div id=\"doc\" class=\"home\"><div id=\"a_7yQ3A2\"><div id=\"a_Vldpk\"><img id=\"polboxes\" src=\"https:\/\/beautifulracket.com\/img\/polboxes-placeholder.png\" alt=\"image\" \/><\/div><\/div><h3 class=\"dept\" id=\"start\"><a href=\"#start\">Start<\/a><\/h3><ul class=\"toc start\"><li><a href=\"https:\/\/beautifulracket.com\/foreword.html\">Fore\u00adword by Matthew Flatt<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/introduction.html\">Intro\u00adduc\u00adtion<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/setup.html\">Setup<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/acknowledgments.html\">Acknowl\u00adedg\u00adments<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/about-the-author.html\">About the author<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/legal.html\">Legal<\/a><\/li><\/ul><h3 class=\"dept\" id=\"this-book-is-not-free\"><a href=\"#this-book-is-not-free\">This book is not free<\/a><\/h3><ul class=\"toc\"><li><a href=\"https:\/\/beautifulracket.com\/how-to-pay.html\">How to pay for this book<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/why-you-should-pay.html\">Why you should pay<\/a><\/li><\/ul><h3 class=\"dept\" id=\"tutorials\"><a href=\"#tutorials\">Tutorials<\/a><\/h3><ul class=\"toc tutorials\"><li><a href=\"https:\/\/beautifulracket.com\/stacker\/intro.html\">Make a lan\u00adguage in one hour: stacker<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/funstacker\/intro.html\">Learn some func\u00adtional pro\u00adgram\u00adming: funstacker<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/stackerizer\/intro.html\">Dive deeper into macros: stackerizer<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/bf\/intro.html\">Fol\u00adlow the gram\u00admar: bf<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/jsonic\/intro.html\">Extend a data for\u00admat: jsonic<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/jsonic-2\/intro.html\">Level up: jsonic revis\u00adited<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/jsonic-3\/intro.html\">Fin\u00adish\u00ading moves: jsonic<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/wires\/intro.html\">Imag\u00adine a lan\u00adguage: wires<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/basic\/intro.html\">Go with the flow: basic<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/basic-2\/intro.html\">Into the rapids: more basic<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/basic-3\/intro.html\">Clos\u00ading the loop: basic<\/a><\/li><\/ul><h3 class=\"dept\" id=\"explainers\"><a href=\"#explainers\">Explainers<\/a><\/h3><ul class=\"toc explainers\"><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/booleans-and-conditionals.html\">Booleans &amp; con\u00addi\u00adtion\u00adals<\/a><\/li><li><a class=\"difficulty-3\" href=\"https:\/\/beautifulracket.com\/explainer\/continuations.html\">Con\u00adtin\u00adu\u00ada\u00adtions<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/contracts.html\">Con\u00adtracts<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/data-structures.html\">Data struc\u00adtures<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/equality.html\">Equal\u00adity<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/errors-and-exceptions.html\">Errors &amp; excep\u00adtions<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/evaluation.html\">Eval\u00adu\u00ada\u00adtion<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/functions.html\">Func\u00adtions<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/hygiene.html\">Hygiene<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/identifiers.html\">Iden\u00adti\u00adfiers<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/importing-and-exporting.html\">Import\u00ading &amp; export\u00ading<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/interposition-points.html\">Inter\u00adpo\u00adsi\u00adtion points<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/lang-line.html\">The #lang line<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/lists.html\">Lists<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/loops.html\">Loops<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/macros.html\">Macros<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/modules.html\">Mod\u00adules<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/numbers.html\">Num\u00adbers<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/pairs.html\">Pairs<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/parameters.html\">Para\u00adme\u00adters<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/recursion.html\">Recur\u00adsion<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/repl.html\">The REPL<\/a><\/li><li><a class=\"difficulty-1\" href=\"https:\/\/beautifulracket.com\/explainer\/stringlike-types.html\">String\u00adlike types<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/syntax-objects.html\">Syn\u00adtax objects<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/syntax-patterns.html\">Syn\u00adtax pat\u00adterns<\/a><\/li><li><a class=\"difficulty-2\" href=\"https:\/\/beautifulracket.com\/explainer\/unit-testing.html\">Unit test\u00ading<\/a><\/li><\/ul><h3 class=\"dept\" id=\"appendix\"><a href=\"#appendix\">Appendix<\/a><\/h3><ul class=\"toc appendix\"><li><a href=\"https:\/\/beautifulracket.com\/appendix\/domain-specific-languages.html\">Domain-spe\u00adcific lan\u00adguages<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/appendix\/from-br-to-racket-base.html\">From br to racket\/base<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/appendix\/glossary.html\">Glos\u00adsary<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/appendix\/master-recipe.html\">Mas\u00adter recipe<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/appendix\/racketcon.html\">The mak\u00ading of Beau\u00adti\u00adful Racket<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/appendix\/resources.html\">Resources<\/a><\/li><li><a href=\"https:\/\/beautifulracket.com\/appendix\/why-racket-why-lisp.html\">Why Racket? Why Lisp?<\/a><\/li><\/ul><\/div>","created_at":"2018-11-07T16:53:07+0000","updated_at":"2018-11-07T16:53:07+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"beautifulracket.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":603,"title":"How can I recall the argument of the previous bash command?","url":"https:\/\/stackoverflow.com\/questions\/3371294\/how-can-i-recall-the-argument-of-the-previous-bash-command","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        228&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>Is there a way in Bash to recall the argument of the previous command?<\/p>\n<p>I usually do <code>vi file.c<\/code> followed by <code>gcc file.c<\/code>. <\/p>\n<p>Is there a way in Bash to recall the argument of the previous command?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/3371294\/revisions\" title=\"show all edits to this post\">edited Mar 12 '16 at 11:59<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2314737\/user2314737\">user2314737<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            13.9k105067&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Jul 30 '10 at 12:16&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/406689\/the-coder\">The Coder<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,143384&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"7\">&#13;\n                                7 Answers&#13;\n                                7&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        412&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>You can use <code>$_<\/code> or <code>!$<\/code> to recall the last argument of the previous command.<\/p>\n<p>Also <code>Alt + .<\/code> can be used to recall the last argument of any of the previous commands.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jul 30 '10 at 12:17&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/227665\/codaddict\">codaddict<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            341k63429489&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        127&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>If the previous command had two arguments, like this<\/p>\n<pre>ls a.txt b.txt\n<\/pre>\n<p>and you wanted the first one, you could type<\/p>\n<pre>!:1\n<\/pre>\n<p>giving<\/p>\n<pre>a.txt\n<\/pre>\n<p>Or if you wanted both, you could type<\/p>\n<pre>!:1-2\n<\/pre>\n<p>giving<\/p>\n<pre>a.txt b.txt\n<\/pre>\n<p>You can extend this to any number of arguments, eg:<\/p>\n<pre>!:10-12\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jul 30 '10 at 13:10&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/20570\/robert-gowland\">Robert Gowland<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            4,39153452&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        55&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>In the command line you can press <kbd>esc<\/kbd>-<kbd>.<\/kbd> or <kbd>alt<\/kbd>+<kbd>.<\/kbd><\/p>\n<p>It cycles through the previous arguments you used.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/3374007\/revisions\" title=\"show all edits to this post\">edited May 19 '15 at 17:31<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1136008\/a-d\">A.D.<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            2,94113046&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jul 30 '10 at 17:35&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/450991\/antonio-mano\">Antonio Mano<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        48&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><code>!!:n<\/code> where <code>n<\/code> is the 0-based position of the argument you want.<\/p>\n<p>For example:<\/p>\n<pre>echo 'one' 'two'\n# \"one two\"\necho !!:2\n# \"two\"\n<\/pre>\n<p>The <code>!<\/code> prefix is used to access previous commands.<\/p>\n<p>Other useful commands:<\/p>\n<ul><li><code>!$<\/code> - last argument from previous command<\/li>\n<li><code>!^<\/code> - first argument (after the program\/built-in\/script) from previous command<\/li>\n<li><code>!!<\/code> - previous command (often pronounced \"bang bang\")<\/li>\n<li><code>!n<\/code> - command number <code>n<\/code> from <code>history<\/code><\/li>\n<li><code>!pattern<\/code> - most recent command matching <code>pattern<\/code><\/li>\n<li><code>!!:s\/find\/replace<\/code> - last command, substitute <code>find<\/code> with <code>replace<\/code><\/li>\n<\/ul><p><a href=\"http:\/\/www.washington.edu\/computing\/unix\/history.html\">More info on command history<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Apr 15 '16 at 18:59&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/400550\/johntron\">Johntron<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,35311721&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        22&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>If you know the number given in the history for a particular command, you can pretty much take any argument in that command using following terms.<\/p>\n<p>Use following to take the second argument from the third command in the history,<\/p>\n<p><code>!3:2<\/code><\/p>\n<p>Use following to take the third argument from the fifth last command in the history,<\/p>\n<p><code>!-5:3<\/code><\/p>\n<p>Using a minus sign, you ask it to traverse from the last command of the history.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Apr 23 '14 at 17:51&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3440170\/madisz\">Madisz<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        14&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Yes, you can use <code>!$<\/code> to recall the last argument of the preceding command.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jul 30 '10 at 12:18&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/101258\/justin-ethier\">Justin Ethier<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            91.9k44198259&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        8&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>!* runs a new command with all previous arguments.  <\/p>\n<pre>ls \/tmp\ncd !*\n#you are now in \/tmp\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/42281955\/revisions\" title=\"show all edits to this post\">edited Feb 16 '17 at 18:40<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 16 '17 at 18:30&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/6768257\/noah\">Noah<\/a>&#13;\n        &#13;<\/div>","created_at":"2018-11-07T16:53:07+0000","updated_at":"2018-11-07T16:53:07+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":604,"title":"Bootstrapping In a Nutshell \u2013 Ep #3: How to Create Your Landing Page (\ud83c\udf81 25% OFF carrd.co \ud83c\udf81) \u2013 Indie Hackers","url":"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-part-3-how-to-create-your-landing-page-3651e8c05f","content":"<div class=\"medium-insert-images medium-insert-images-full\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/carrd_hero.jpg\" alt=\"\" class=\"\" \/><\/figure><\/div><p class=\"\"><i>Bootstrapping In a Nutshell is a series of articles where you'll learn how to bootstrap a new online business from scratch through a real-world project.<\/i><br \/><\/p><blockquote><p class=\"\"><b>Get 25% OFF on <a href=\"http:\/\/carrd.co\" target=\"_blank\" rel=\"noopener noreferrer\">carrd.co<\/a> with the discount code: NUTSHELL25! (expires end of November).\u00a0<\/b>Redeem it by entering it in the \"Coupon code\" box at the bottom of the upgrade page.<br \/><\/p><\/blockquote><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/9b4fcd4b3d.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">To really <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-2-how-to-validate-your-idea-by-talking-to-customers-47cab94448\" target=\"_blank\" rel=\"noopener noreferrer\">validate your idea<\/a> you need people to take action. People might tell you they'll purchase or subscribe to your product, but there's a big difference between what they say and what they do \u2014 there's nothing forcing them to stick to their commitment when your product is actually released. So to make them take action, you need a landing page.<\/p><h2>The purpose of a landing page<\/h2><p class=\"\">A landing page is useful in several ways:<\/p><ul><li>It propels your idea into the real world. This forces you to work on your copywriting to be sure you do a good job explaining the purpose of your product.<\/li><li>It lets you gather immediate feedback about your idea and connect to people that may be interested in it. Collecting emails to have one-on-one conversations is imperative for getting insights for success.<\/li><li>It lets you really validate your idea in a real-world environment before actually coding your product. By putting a call-to-action on the page you force people to make a decision: is this product important enough for me to enter my email address or not?<\/li><\/ul><p class=\"\">Using a landing page is only the first step to validate an idea but it's a critical one. If you don't collect any email or pre-order, well you know that your idea sucks. The good thing is that you saved time by not building this product. You can move on.<\/p><p class=\"\">If you collect some emails but not that much, you can still connect to these people by starting a conversation about their needs and pains. These conversations will lead to creating a product that's useful for them and probably other people. It will give you insights to create a product people want.<\/p><p class=\"\">And if you collect lots of emails or pre-orders and people are crazy about your product, well you know that you have a market and can start building your product with more confidence.<\/p><h2>How to create a landing page<\/h2><p class=\"\">There are tons of services out there to create a landing page. But a very well-designed one is <a href=\"http:\/\/carrd.co\" target=\"_blank\" rel=\"noopener noreferrer\">carrd.co<\/a>. The pro plan price is $19 \/ year ($14 \/ year with coupon code NUTSHELL25), which is extremely cheap. Don't hesitate to purchase it.<\/p><p class=\"\">A landing page must be simple. There is a famous designer quote that says \"Less is More\" and you should definitely live by it. Simplicity sells. People are busy and don't have time to read tons of text to understand your product.<\/p><p class=\"\">So you should stick with the bare minimum: a good picture, a title, a tagline, and a call-to-action button for the email form. It's not a coincidence that the template designs available on Carrd are so simple:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/ecc79e31b6.png\" class=\"\" alt=\"image\" \/><figcaption class=\"\">Templates available on carrd.co to start your landing page<\/figcaption><\/figure><\/div><p class=\"\">Before rushing to select your template, you must first carefully craft the title and tagline you'll use.<\/p><p class=\"\">As I'm not a copywriting superhero (yet) I did my research. I read <a href=\"https:\/\/kopywritingkourse.com\/how-to-create-a-great-tagline\/\" target=\"_blank\" rel=\"noopener noreferrer\">this article<\/a>, which says in a nutshell:<\/p><ul><li>Step 1: dump out your entire business in a few sentences<\/li><li>Step 2: trim it down<\/li><li>Step 3: trim it down some more<\/li><\/ul><p class=\"\">As I said: less is more. But be careful here. Successfully finding a good title and tagline is not easy. That's why there are entire courses about copywriting \u2014 it's a full-time job. So try to do your best, iterate, and test your copywriting by talking to customers.<\/p><h2>Your value proposition<\/h2><p class=\"\">The value proposition of your business is the reason why your customers should use your product. Most of the time, it's because using your product will make them save time or money, or both. Ask yourself this question: why would someone want to use my product?<\/p><p class=\"\">For my blockbuster movies newsletter idea, it's to save time. It's more convenient to receive a list of all the movies I'm interested in in my inbox. First, the service filters out all the non-blockbuster movies, and second, it comes to me: I don't have to manually check the theater website or some cinema news website.<\/p><p class=\"\">This won't fix any major pain points for my customers so I'm not expecting any kind of huge success with this idea. Nevertheless, if I can collect the email addresses of a few movie fans I'll be able to open up some conversations and get great insights to move on the next idea or to reshape this one.<\/p><p class=\"\">So to know what your value proposition is, just answer this question: why would someone use your product?<\/p><h2>Exercise: write a title and a tagline<\/h2><p class=\"\">I did my homework and started at step 1 for my idea. I wrote a paragraph that explains what my product is:<\/p><p class=\"\"><i>My product will let you receive an email for free once a week with all the newly released blockbusters available in theaters so you never miss one. You can watch trailers, pre-order DVD and Blu-rays to save money right from your inbox.<\/i><\/p><p class=\"\">Let's continue this process with step 2: trim it down.<\/p><p class=\"\"><i>Receive an email every week for free with the latest blockbusters so you never miss them.<\/i><\/p><p class=\"\">I totally removed the watch trailer and pre-order part because my value proposition for now (I can change it later) is to never miss a blockbuster. Let's jump to step 3: trim it down some more.<\/p><p><i>Never miss a blockbuster. Get the latest blockbusters releases in your inbox. Every week. For Free.<\/i><\/p><p class=\"\">Put your value proposition \u2014 the benefit to the user \u2014 before everything else in your title. And then use your tagline to explain what your product is all about. Use short sentences. You should be able to tweet it!<\/p><p class=\"\">Now that we have the content for our website, let's build it!<\/p><h2>10 steps process to create a landing page with carrd.co<\/h2><p class=\"\">Let's build a landing page with a custom domain name, because, you know, we are serious!<\/p><h3>Step 1 \u2014 Domain name<\/h3><p class=\"\">Purchase your domain name. I asked for some help on this part in my <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-part-1-how-to-select-your-domain-name-for-seo-a5eb5879a7\" target=\"_blank\" rel=\"noopener noreferrer\">second article<\/a> and @shanecleveland suggested \"upcominghits\". I like it a lot (thanks Shane!), but the term hits can be used for songs too, and I do movies. So I thought upcomingmovies would be good. But a little bit too long.<\/p><p class=\"\">Maybe bigmovies? I searched for available domains, and of course the .com was already taken. So I took a look at it out of curiosity. Damn. A porn website! They're everywhere, dammit! I don't want to be associated with that in any way.<\/p><p class=\"\">So I switched back to upcomingmovies. The .com redirects to a facebook page (wtf?) but the .io is available. I take a look at Google search suggestions:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/915f5ebfe3.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">Jesus! I just typed 3 letters and boom! Incoming movies 1st. Awesome. I checked the Google Keyword Planner tool to look at the traffic:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/b93fb48789.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">Holy crap! Good numbers too here \u2014 let's go for upcomingmovies.io!<\/p><p class=\"\">A general note here: don't overthink your name too much. The numbers above <i>should not<\/i> have any impact on my rank in Google or the traffic on my website because of the <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-part-1-how-to-select-your-domain-name-for-seo-a5eb5879a7\" target=\"_blank\" rel=\"noopener noreferrer\">EMD update<\/a>. But my future content marketing strategy should definitely take that into account.<\/p><p class=\"\">By the way, Google <a href=\"https:\/\/en.wikipedia.org\/wiki\/History_of_Google#Name\" target=\"_blank\" rel=\"noopener noreferrer\">should have been named Googol<\/a>, did you know that? That didn't stop them to succeed. Your product is more important than your name or website, you can still change them later.<\/p><h3>Step 2 \u2014 Email address<\/h3><p class=\"\">Create an email address with your new domain. Usually you can use <b>hello@yourdomain.tld<\/b>, for me it's hello@upcomingmovies.io: shot me an email now to tell me you read this article \ud83d\udc4d.<\/p><h3>Step 3 \u2014 DNS configuration<\/h3><p class=\"\">Follow the instructions in <a href=\"https:\/\/carrd.co\/guides\/using-a-custom-domain#add-host-records\" target=\"_blank\" rel=\"noopener noreferrer\">this guide<\/a>, basically:<\/p><ul><li>Make your A record points to 108.168.244.214<\/li><li>Add a CNAME www record points to \"yourdomain.tld.\" (don't forget to add the trailing '.' at the end, some registrars adds it by default, others don't).<\/li><\/ul><p class=\"\">After 24 hours you should be able to connect to your website through SSL. Using SSL is better for SEO. When you type yourdomain.tld in the address bar it should use SSL by default automatically and redirect you to https:\/\/yourdomain.tld.<\/p><h3 class=\"\">Step 4 \u2014 Collect email addresses<\/h3><p class=\"\"><b>Create a MailChimp account<\/b> with your brand new email address. MailChimp is a service that will allow you to collect email addresses and send your newsletter easily. There are tons of other services, but I know this one and it is compatible natively with carrd.co so I pick it up. Plus it's free until 2000 subscribers!<\/p><p class=\"\"><a href=\"https:\/\/mailchimp.com\/signup\/\" target=\"_blank\" rel=\"noopener noreferrer\">Sign up for MailChimp<\/a>. Use your new email address, for me hello@upcomingmovies.io. Follow the instructions. Confirm your account etc. When everything is set up. Log in and <b>create a new list<\/b>. Then click on <b>Settings &gt; List names and defaults<\/b>:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/d082d8ae80.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">Fill the fields at the bottom and write somewhere your List ID, it's here:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/264fa2a3ec.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">MailChimp is packed with a ton of options for lists. For example you should select <b>Send a final welcome email<\/b>\u00a0and design this email carefully to welcome your new subscribers. You should also select <b>Send unsubscribe confirmations to subscribers<\/b>.<br \/><\/p><p class=\"\">In general, browse the knowledge database of MailChimp, there are tons of options you should learn about to make your newsletter awesome and stay in touch with your clients.<\/p><h3 class=\"\">Step 5 \u2014 Get your MailChimp API key to collect emails<\/h3><p class=\"\">Now, <a href=\"https:\/\/kb.mailchimp.com\/integrations\/api-integrations\/about-api-keys?_ga=2.256495053.792576225.1507818780-1322800729.1476084317\" target=\"_blank\" rel=\"noopener noreferrer\" class=\"\">create an API key<\/a>, label it carrd.co for example. You'll need it later to configure your website email form.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/49f3421536.png\" class=\"\" alt=\"image\" \/><\/figure><\/div><h3 class=\"\">Step 6 \u2014 Get an awesome template<\/h3><p class=\"\">Go to <a href=\"http:\/\/carrd.co\" target=\"_blank\" rel=\"noopener noreferrer\">carrd.co<\/a> and select a template with the \"pro\" label at the top right corner. Yes we take the pro label because we're going to purchase this service for $14 \/ year thanks to the <b>NUTSHELL25<\/b> coupon code, it's a steal. Trust me. Select the template you want, hover over it and click on <b>Select<\/b>.<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/15605dbe78.png\" alt=\"image\" \/><\/figure><\/div><h3>Step 7 \u2014 Design your landing page<\/h3><p class=\"\">Accept the trial period, click on <b>Select + Start the trial<\/b>. Read the instructions about the interface.<\/p><p class=\"\">Set a great picture for your background, change the title to your value proposition. Set the tagline. Customize everything to your needs but respect the KISS principle: Keep It Stupid and Simple. Your website should be able to load fast and be easy to understand. Put only one call-to-action button.<\/p><p class=\"\">Here is my project:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/6b16e6b645.jpg\" alt=\"image\" \/><\/figure><\/div><p class=\"\"><i>Optional<\/i>: I created a twitter account <a href=\"https:\/\/twitter.com\/@_moviesupcoming\" target=\"_blank\" rel=\"noopener noreferrer\">@_moviesupcoming<\/a> for this landing page so people can reach me out this way. You can create accounts for other social media platforms.<br \/><\/p><p class=\"\">Last but not least: go to your carrd dashboard &gt; Account &gt; click on <b>Select a pro plan<\/b>\u00a0and at the bottom enter your coupon code <b>NUTSHELL25<\/b>, hit apply and purchase the standard or pro plan with 25% OFF! You're welcome!<\/p><h3>Step 8 \u2014 Connect your email form to MailChimp<\/h3><p class=\"\">Let's do the technical part now. Edit your website, select your email form, then in the left panel, <b>select MailChimp Signup<\/b> and then <b>enter your API keys<\/b> and <b>List Id<\/b>. Plus check the <b>double\u00a0opt-in <\/b>checkbox like on this screenshot:<\/p><div class=\"medium-insert-images medium-insert-images-auto\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/384c15d293.png\" class=\"\" alt=\"image\" \/><\/figure><\/div><p class=\"\">Then <b>click Done<\/b>! Your email form is now hooked with your MailChimp account! Well done!<\/p><h3>Step 9 \u2014 Publish your landing page<\/h3><p class=\"\">Now configure the custom domain, <b>click on the publish button<\/b> at the top right corner:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/d1572d7708.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\">And <b>add your title and description<\/b>, select <b>Use my own domain<\/b> and <b>put your naked domain name<\/b> below like this:<\/p><div class=\"medium-insert-images medium-insert-images-inline\"><figure><img src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/df96ffea45.png\" alt=\"image\" \/><\/figure><\/div><p class=\"\"><b>Click on Publish Changes<\/b>\u00a0and then on <b>View Site<\/b> to see your site up and running!<\/p><h3>Step 10 \u2014 Test your landing page<\/h3><p class=\"\">Test the email form:<\/p><ol><li>On your website, enter your personal email and tap on your call-to-action button (Subscribe).<\/li><li>Go to your inbox and confirm your subscription via the link sent by MailChimp.<\/li><li>Go to your MailChimp account and check if your email has been added to your list.<\/li><\/ol><p class=\"\"><b>Important notes (or how I lost several hours trying to figure this out)<\/b><\/p><ul><li>If you don't confirm your account via the email MailChimp sends you when you subscribe, you won't see your email in your MailChimp list AND you won't be able to subscribe again on your website (error: Sorry, that email has already been subscribed).<\/li><li>To make your tests easier, never unsubscribe yourself from the list, but use MailChimp to delete your email from the list when your status is subscribed.<\/li><li>If you accidentally unsubscribe from your list, you won't be able to delete your email from the MailChimp List interface. You must resubscribe first from the email that confirmed you were correctly unsubscribed! Then you can delete your email through the MailChimp List interface because your status will be back to subscribed.<\/li><\/ul><p class=\"\">After that you can adjust the design of your website whenever your want and improve it. Mess around with the carrd interface to learn it. You can do lots of things but remember to keep everything simple at the end. Go straight to the point!<\/p><p class=\"\">Also I strongly encourage you to design your signup emails and double opt-in emails carefully to maximize the conversion rates. Make them personal and warm. Care about your new subscribers, provide them values immediately.<\/p><p class=\"\"><i>In the next article you will learn how to set up Google Analytics so you can record your conversion rate between new visitors to new subscribers over time.<\/i><\/p><blockquote><p><b>ABOUT ME<\/b>: I'm a 32 yo software engineer and entrepreneur. I use my free time to start new online businesses.\u00a0Actually I'm learning front-end development while building an audience on twitter, <a href=\"https:\/\/twitter.com\/jeremymouzin\" target=\"_blank\" rel=\"noopener noreferrer\">follow me<\/a> ;-).<\/p><p><b>NEXT ARTICLE<\/b>: <a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-4-how-to-set-up-google-analytics-for-your-landing-page-191a88ea92\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #4:\u00a0How to Set Up Google Analytics For Your Landing Page<\/a><\/p><p><b>FULL ARTICLES LIST<\/b><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-0-how-to-select-your-next-business-idea-bd6f7c7a9c\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #0: How to Select Your Next Business Idea<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-1-how-to-select-your-domain-name-for-seo-a5eb5879a7\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #1: How to Select Your Domain Name for SEO<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-2-how-to-validate-your-idea-by-talking-to-customers-47cab94448\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #2: How to Validate Your Idea by Talking to Customers<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-3-how-to-create-your-landing-page-25-off-carrd-co-3651e8c05f\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #3: How to Create Your Landing Page<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-4-how-to-set-up-google-analytics-for-your-landing-page-191a88ea92\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #4: How to Set Up Google Analytics For Your Landing Page<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/@Hercule\/bootstrapping-in-a-nutshell-ep-5-how-to-drive-quality-traffic-to-your-landing-page-44f14d0c66\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #5: How to Drive Quality Traffic to Your Landing Page<\/a><\/p><p><a href=\"https:\/\/www.indiehackers.com\/edit-article\/@Hercule\/bootstrapping-in-a-nutshell-ep-6-advice-for-beginners-and-conclusion-b604c96ab0\" target=\"_blank\" rel=\"noopener noreferrer\">Ep #6: Advice for beginners and conclusion<\/a><br \/><\/p><\/blockquote>","created_at":"2018-11-07T16:53:07+0000","updated_at":"2018-11-07T16:53:07+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":11,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/articles\/wMSY0xCPqCRqfmk5pe508YY2yTp2\/@Hercule%7C3651e8c05f\/ecc79e31b6.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":600,"title":"Mullvad VPN","url":"https:\/\/www.mullvad.net\/en\/","content":"<h3>Team<\/h3><p>Amagicom is 100% owned by founders Fredrik Str\u00f6mberg and Daniel Berntsson who are actively involved in the company. The rest of the team includes Robin L\u00f6vgren, Simon Andersson, Linus, Richard Mitra, Sanny Mitra, David Marby, Odd, Andrej Mihajlov, Janito Ferreira Filho, Elad Yarom and Jan Jonsson.<\/p>","created_at":"2018-11-07T16:53:06+0000","updated_at":"2018-11-07T16:53:06+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.mullvad.net","preview_picture":"http:\/\/mullvad.net\/static\/press\/mullvad-logo-lamp.384386f4d7a2.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":601,"title":"Why I\u2019m Learning Perl 6 \u2013 Evan Miller","url":"http:\/\/www.evanmiller.org\/why-im-learning-perl-6.html","content":"<p>By <a href=\"http:\/\/www.evanmiller.org\/\">Evan Miller<\/a><\/p><p><em>July 25, 2017<\/em><\/p><p>Want to hear a programming joke?<\/p><p>Perl\u00a06.<\/p><p>Want to know my new favorite programming language?<\/p><p>Perl\u00a06.<\/p><p>Shhhh just clone master and run this:<\/p><pre>use v6.d.PREVIEW;\nmy $channel = Channel.new;\nmy @ten_tasks = (^10).map: {\n    start {\n        my $thread = $*THREAD.id;\n        await $channel;\n        say \"HOLY MOLEY SOMEONE STOLE MY THREAD\" if $thread != $*THREAD.id;\n    }\n}\n$channel.send(\"Ring ring\") for ^10;\n$channel.close;\nawait @ten_tasks;\n<\/pre><p>Guess what it outputs?<\/p><p>\nNothing! Just kidding, you\u2019ll see a lot of distressed messages from tasks that\npassed out wait for the phone to ring, and woke up wearing someone else\u2019s OS\nthread.<sup><a href=\"#notes\">1<\/a><\/sup><\/p><p>\nWhy is this important? Concurrency is hard and if you want M:N thread\nmultiplexing (i.e. WEB SCALE CODE, where application threads aren\u2019t pinned to\npthreads) your options today are precisely Erlang, Go, .NET, and Perl\u00a06.\n<\/p><p>\nIf you\u2019re feeling <a href=\"http:\/\/www.evanmiller.org\/why-i-program-in-erlang.html\">confused<\/a> by Erlang, <a href=\"http:\/\/www.evanmiller.org\/four-days-of-go.html\">put off<\/a> by Go, and indifferent to .NET,\ntake a look at Perl\u00a06. Seriously. Sure, it was in development longer than\n<em>Boyhood<\/em> \u2014 begun in 2000, two years before the film, and released in 2015,\ntwo years after it \u2014 but maybe if somebody works on something for 15\nyears, they might actually be making something good. I\u2019m not saying it\u2019s true in\nthis case (and probably not in the case of <em>Boyhood<\/em>) but at least it\u2019s\npossible, right?\n<\/p><p>\nIt\u2019s funny, fifteen years ago everyone was saying how Perl was shooting itself\nin the foot with a massive language redesign, because the world was clearly\ngoing to switch to Python and Ruby. Well everyone did switch to Python and\nRuby, but now everyone\u2019s got the itch to switch again because it turns out\nPython and Ruby weren\u2019t designed for concurrent code execution. Maybe shooting\nyourself in the foot isn\u2019t such a bad strategy, if what you need are bionic\nrobo-feet.\n<\/p><p>\nI\u2019m not going to review Perl\u00a06 as a language here because it\u2019s a massive spec\nwith a lot of dark corners and weird edge cases, but I do want to share some things I\u2019ve\ndiscovered about\nthe language run-time. <a href=\"https:\/\/www.moarvm.org\/\">MoarVM<\/a>, the Perl\u00a06 virtual machine, is a fantastic piece\nof technology.\n<\/p><p>\nEvent loop-aware scheduler? Check. Continuations? You got \u2019em. What happens\nwhen you mix the two? Well, you can spin up hundreds of worker routines and\nhave them wait for messages (from, say, the network) without tying up any OS\nthreads. When the routine gets a message, it continues executing on whatever OS\nthread happens to be available. Hence all the <code>\"HOLY MOLEY\"<\/code> in the opening\nexample. It\u2019s like Go or Erlang with more dollar signs, or Nginx or Node.js\nwithout the callback cacciatore.<sup><a href=\"#notes\">2<\/a><\/sup><\/p><p>\nMoarVM is written in nice clean C, and full of pleasant surprises if you care to\ndig. There\u2019s no GIL, so unlike Those Other Languages, Perl\u00a06 bytecode can\nexecute concurrently. The Unicode is great, getting <code>\"Stra\u00dfe\".uc<\/code> right and\nusing a clever constant-time algorithm for addressing codepoint-combined\ngraphemes. The object model and garbage collector play nicely with the FFI, so\ninterfacing with C is a cinch. The code base is well-written, well-organized,\nwell-documented, and well-tested; little things, like functions having the\nright size and variables having good names, make it a pleasure to read.\n<\/p><p>\nA quick way to judge a language implementation is by inspecting its string\nconcatenation function. If concat is implemented as a <code>realloc<\/code> and <code>memcpy<\/code>,\nwell, the upstairs lights probably aren\u2019t set to full brightness. Behold, in\nMoarVM strings are broken into strands, and strands can repeat without taking\nup more memory. So this expression:\n<\/p><pre>\"All work and no play makes Jack a dull boy\" x 1000\n<\/pre><p>\nDoesn\u2019t make a thousand copies of the string, or even make a thousand pointers\nto the same string. It\u2019s stored as a short string and a repetition count, and\nthe VM string functions know how to operate efficiently on strings with repeated\nsubstrings. Pretty clever, right? Make a hundred million copies; RAM usage\nbarely budges.<sup><a href=\"#notes\">3<\/a><\/sup><\/p><p>\nI won\u2019t try to sell you on Perl\u00a06 here \u2014 I still don\u2019t entirely\nunderstand the language, to be honest \u2014 but having spent a few days poking\naround its run-time, I will tell you that the virtual machine is a promising\npiece of systems work. Incidentally MoarVM is a relatively new part of the\nPerl\u00a06 stack, having supplanted ParrotVM, the thing that was supposed to run\nAll The Scripting Languages; MoarVM has a tighter focus on running just Perl\u00a06,\nand looks like a much stronger foundation for the future.\n<\/p><p>\nIf you work on implementing programming languages, take a Sunday afternoon and\ncheck out the <a href=\"https:\/\/github.com\/MoarVM\/MoarVM\">MoarVM code<\/a>; I suspect even hardened VM hackers can learn at\nleast something from its scheduler and string handling. If you\u2019re drawing up a\nlist of programming languages to try out, do your future self a favor and put\nPerl\u00a06 on the menu. I can\u2019t tell you if the language will be to your taste, but\nI can tell you that the chef was famous once, and the kitchen has all the\ningredients it needs to make something really special.\n<\/p><hr \/><h3>Notes<\/h3><ol><li><p>Actually, the program will output nothing if you\u2019re using a release\n        of Perl\u00a06 prior to Rakudo Star 2017.07; a caching bug previously caused <code>$*THREAD.id<\/code>\n        to report an out-of-date value.<\/p><\/li>\n    <li><p>For the curious: Go switches contexts on\n        I\/O, message passes, and function calls. The Erlang VM switches contexts on\n        I\/O, message passes, and every 4,000 VM instructions. MoarVM switches\n        contexts on calls to <code>await<\/code>, which can include message\n        passes and asynchronous I\/O promises. Nginx and Node.js never\n        switch context, which is why all the callbacks are necessary.<\/p><\/li>\n    <li><p>There\u2019s a hard string-length limit of 2<sup>32<\/sup>, so don't go\ntoo crazy with this feature. Also, Perl\u00a06\u2019s regex \/ grammar engine seems to\nflatten everything before matching, so you don't yet get the full benefit of\nthe compressed representation. If you're looking for a project idea\u2026<\/p><\/li>\n<\/ol><hr \/><p><em>You\u2019re reading <a href=\"http:\/\/www.evanmiller.org\/\">evanmiller.org<\/a>, a random collection of math, tech, and musings. If you liked this you might also enjoy:\n    <a href=\"http:\/\/www.evanmiller.org\/a-review-of-perl-6.html\">A Review of Perl 6<\/a>\n        <a href=\"http:\/\/www.evanmiller.org\/elixir-ram-and-the-template-of-doom.html\">Elixir RAM and the Template of Doom<\/a>\n        <a href=\"http:\/\/www.evanmiller.org\/you-cant-dig-upwards.html\">You Can\u2019t Dig Upwards<\/a>\n        <a href=\"http:\/\/www.evanmiller.org\/why-im-betting-on-julia.html\">Why I\u2019m Betting on Julia<\/a>\n    <\/em>\n<\/p><hr \/><p><em>Get new articles as they\u2019re published, via <a href=\"https:\/\/twitter.com\/EvMill\">Twitter<\/a> or <a href=\"http:\/\/www.evanmiller.org\/news.xml\">RSS<\/a>.<\/em><\/p><hr \/><p><em>Want to look for statistical patterns in your MySQL, PostgreSQL, or SQLite database? My desktop statistics software <strong><a href=\"http:\/\/www.wizardmac.com\/\">Wizard<\/a><\/strong> can help you analyze <strong>more data in less time<\/strong> and <strong>communicate discoveries visually<\/strong> without spending days struggling with pointless command syntax. Check it out!<\/em><\/p><div><a href=\"http:\/\/www.wizardmac.com\/\"><img height=\"128\" width=\"128\" src=\"http:\/\/www.evanmiller.org\/images\/index\/wizard-nonpro.png\" alt=\"image\" \/><\/a><br \/><strong><a href=\"http:\/\/www.wizardmac.com\/\">Wizard<\/a><\/strong><br \/>Statistics the Mac way<\/div><hr \/><p><a href=\"http:\/\/www.evanmiller.org\/\">Back to Evan Miller\u2019s home page<\/a> \n\u2013 <a href=\"http:\/\/www.evanmiller.org\/news.xml\">Subscribe to RSS<\/a>\n\u2013 <a href=\"http:\/\/twitter.com\/EvMill\">Twitter<\/a> \n\u2013 <a href=\"https:\/\/www.youtube.com\/channel\/UC6PJEAiKFzEfq4stBoZz07Q\">YouTube<\/a> \n<\/p><hr \/>","created_at":"2018-11-07T16:53:06+0000","updated_at":"2018-11-07T16:53:06+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.evanmiller.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":598,"title":"Facebook Messenger Lite: Free Calls & Messages APKs - APKMirror","url":"https:\/\/www.apkmirror.com\/apk\/facebook-2\/messenger-lite\/","content":"<p>Have the APK file for an alpha, beta, or staged rollout update? Just drop it below, fill in any details you know, and we'll do the rest!<\/p><p class=\"modal-spacing\">NOTE: Every APK file is manually reviewed by the <a href=\"http:\/\/www.androidpolice.com\" target=\"_blank\">AndroidPolice<\/a> team before being posted to the site.<\/p><p>ERROR:<\/p><p>SUCCESS: Your files have been uploaded, please check if there were any errors.<\/p>","created_at":"2018-11-07T16:53:05+0000","updated_at":"2018-11-07T16:53:05+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.apkmirror.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":599,"title":"Ask HN: Where can I host a 1TB DB relatively cheaply?","url":"https:\/\/news.ycombinator.com\/item?id=15361724","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15361724\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15361724\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=15361724\" class=\"storylink\">Ask HN: Where can I host a 1TB DB relatively cheaply?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        58 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=anothernameiluv\" class=\"hnuser\">anothernameiluv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15361724\">on Sept 28, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15361724&amp;goto=item%3Fid%3D15361724\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Where%20can%20I%20host%20a%201TB%20DB%20relatively%20cheaply%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Where%20can%20I%20host%20a%201TB%20DB%20relatively%20cheaply%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15361724&amp;auth=e437efa7f3f7913278203e5aca0570d4f78ed1b2\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15361724\">34\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>I have a 1TB DB that will grow over time and have been looking for a host that is relatively cheap. Everything I see is hosting with SSD so getting 1TB and larger is somewhat expensive. Are there any (cheap) HDD hosts out there?<\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:53:05+0000","updated_at":"2018-11-07T16:53:05+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":597,"title":"The Nikon D3400 is a bit better cheap dSLR than before","url":"https:\/\/www.cnet.com\/reviews\/nikon-d3400-review\/","content":"<p> \tThe <a href=\"https:\/\/www.cnet.com\/reviews\/nikon-d3300-review\/\">Nikon D3300<\/a> has long been my go-to recommendation for a cheap dSLR, but after two years it's usually time to slap a new coat of paint on consumer products. Nikon's 2016 update to that camera, the D3400, has some small enhancements to bring it up to date, but nothing vital. It remains a great value -- fast with very good photo and video quality and a new kit lens optimized for shooting via the LCD -- but its weak wireless file transfer barely feels like a step up from nothing.<\/p><p> \tAlong with the camera, Nikon has kits with two versions of two new kit lenses. The  AF-P lenses  incorporate stepper motors like Canon's STM lenses for smoother and quieter focus Live View, the LCD-based view on a dSLR. There are two versions of each, 18-55mm and 70-300mm, one with optical image stabilization (Nikon's Vibration Reduction, or VR) and one without; the names differ solely by the \"VR\" designation.<\/p><p> \tNikon's  offering two kits of the D3400, available in different regions. The primary option is bundled with the 18-55mm VR lens for $650 and \u00a3490 (not sold in Australia); \u00a3470 and AU$700 with the non-VR version of the lens (which isn't sold in the US). The other option is a dual-lens kit with the VR version of the 18-55mm lens but the non-VR version of the 70-300mm. While that's a silly configuration -- stabilization on the lens where you don't need it and no stabilization on the one where you do -- it allows Nikon to hit its just-under-$1,000 price of $999.95. That dual-lens kit doesn't seem to be available in the UK and Australia, but you'll be able to get a body-only version in those regions for \u00a3400 and about AU$575.<\/p><h3>Sharper cookie<\/h3><p> \t Overall, the camera delivers very good photo and video quality for its price. Thanks to its new sensor -- the same resolution as the D3300, but without an anti-aliasing filter to slightly unsharpen the image -- the D3400 delivers slightly sharper photos in low light than its predecessor. They don't seem to have better tonal range in shadows and highlights than before, but the extra sharpness makes a difference.<\/p><div class=\"shortcode gallery\"><div data-item=\"image\"><a href=\"https:\/\/www.cnet.com\/pictures\/nikon-d3400-full-resolution-photo-samples\/\" data-component=\"imageGalleryModal\" data-image-gallery-modal-options=\"{&quot;url&quot;:&quot;\\\/pictures\\\/nikon-d3400-full-resolution-photo-samples\\\/xhr\\\/&quot;,&quot;trackUrl&quot;:&quot;\\\/pictures\\\/nikon-d3400-full-resolution-photo-samples\\\/&quot;,&quot;index&quot;:1,&quot;cssPath&quot;:&quot;https:\\\/\\\/cnet2.cbsistatic.com\\\/fly\\\/css\\\/image_gallery\\\/single.desktop-3944ac2815-rev.css&quot;}\">\n                <figure class=\"img\"><img src=\"https:\/\/cnet4.cbsistatic.com\/img\/Y0rqOrg6m5Dj869QS7jERW-w0a8=\/724x407\/2016\/10\/13\/48a40635-2a59-47e0-b11a-61109cdc8ead\/d3400-dsc-1213-3200.jpg\" class=\"\" alt=\"d3400-dsc-0288-100.jpg\" height=\"407\" width=\"724\" \/><\/figure><div class=\"title\">Nikon D3400 full-resolution photo samples<p>11 Photos<\/p><\/div>\n            <\/a><\/div><\/div><p> \t The automatic white balance isn't great: It displayed a purple cast under our  LED test lights and a moderate cyan cast in real daylight, so photos look a bit too cool. But under most circumstances, you won't really notice it if you're not hypersensitive to color accuracy or comparing  to a more accurate camera. This type of issue isn't unusual, though.<\/p><p> \tYou also can't get significantly better noise-reduction results by processing the raw files; the JPEG processing is pretty good. However, you can recover a bit of highlight detail in higher ISO-sensitivity shots, boost underexposures  for low ISO-sensitivity photos and fix the aforementioned white balance issues in raw.<\/p><p> \tAnd thanks to the AF-P lenses, Nikon now has improved -- faster and smoother -- autofocus during movies, which makes them look a lot better for  fans of automatic.<\/p><h4>Analysis samples<\/h4><figure class=\"image image-large pull-none hasCaption shortcode\"><div class=\"originalImage\" data-open=\"modal\" data-target=\"#4771545d-ef69-4703-8919-523e52f7a116\" data-component=\"modalEnlargeImage\"><img src=\"https:\/\/cnet1.cbsistatic.com\/img\/ht8ZGpYaXw1QjEL_xIFccoFNMBo=\/970x0\/2016\/10\/11\/4771545d-ef69-4703-8919-523e52f7a116\/d3400-low-iso-1333.jpg\" class=\"\" alt=\"d3400-low-iso-1333.jpg\" height=\"0\" width=\"970\" \/><a class=\"clickToEnlarge\">Enlarge Image<\/a><\/div><figcaption>JPEGs look clean through ISO 1600. And with nothing to compare them to, the white balance looks OK. \n                                                    Lori Grunin\/CNET\n                                                <\/figcaption><\/figure><figure class=\"image image-large pull-none hasCaption shortcode\"><div class=\"originalImage\" data-open=\"modal\" data-target=\"#601a948c-ad77-42f3-b4c7-5c619bc7aa0f\" data-component=\"modalEnlargeImage\"><img src=\"https:\/\/cnet2.cbsistatic.com\/img\/_pKdoym0BAzc1Eiy1fYYlqWRwAc=\/970x0\/2016\/10\/11\/601a948c-ad77-42f3-b4c7-5c619bc7aa0f\/d3400-high-iso-1333.jpg\" class=\"\" alt=\"d3400-high-iso-1333.jpg\" height=\"0\" width=\"970\" \/><a class=\"clickToEnlarge\">Enlarge Image<\/a><\/div><figcaption>JPEGs still look  good at ISO 3200, which gives you some latitude if you need to shoot action in daylight with the narrow-aperture kit lens. By ISO 6400 you start to see some detail smearing. However, the balance between detail and noise reduction in JPEGs (it's a trade-off) remains decent all the way through the camera's sensitivity range. \n                                                    Lori Grunin\/CNET\n                                                <\/figcaption><\/figure><figure class=\"image image-large pull-none hasCaption shortcode\"><div class=\"originalImage\" data-open=\"modal\" data-target=\"#313155d6-f062-492b-8dbd-eb52451a7464\" data-component=\"modalEnlargeImage\"><img src=\"https:\/\/cnet1.cbsistatic.com\/img\/TZipguVpIffHjteJuh6JCJ6ScDs=\/970x0\/2016\/10\/11\/313155d6-f062-492b-8dbd-eb52451a7464\/d3400-mwbvawb-1333.jpg\" class=\"\" alt=\"d3400-mwbvawb-1333.jpg\" height=\"0\" width=\"970\" \/><a class=\"clickToEnlarge\">Enlarge Image<\/a><\/div><figcaption>When you compare to a neutral, manually white-balanced shot, you can see the difference. \n                                                    Lori Grunin\/CNET\n                                                <\/figcaption><\/figure><figure class=\"image image-large pull-none hasCaption shortcode\"><div class=\"originalImage\" data-open=\"modal\" data-target=\"#2c3ebc76-a5b7-4cac-b836-5e3d32801c66\" data-component=\"modalEnlargeImage\"><img src=\"https:\/\/cnet1.cbsistatic.com\/img\/DN3wYO5sjCo7LMqUgzslCMYaLsU=\/970x0\/2016\/10\/11\/2c3ebc76-a5b7-4cac-b836-5e3d32801c66\/d3400-awb-1333.jpg\" class=\"\" alt=\"d3400-awb-1333.jpg\" height=\"0\" width=\"970\" \/><a class=\"clickToEnlarge\">Enlarge Image<\/a><\/div><figcaption>The camera's cool automatic white balance for daylight shots actually makes the photos look flatter and lower contrast than the raw versions; it's usually the opposite. \n                                                    Lori Grunin\/CNET\n                                                <\/figcaption><\/figure><figure class=\"image image-large pull-none hasCaption shortcode\"><div class=\"originalImage\" data-open=\"modal\" data-target=\"#d4260898-6a6d-4f51-8560-a20fb805eb4e\" data-component=\"modalEnlargeImage\"><img src=\"https:\/\/cnet3.cbsistatic.com\/img\/XgdyJoiABCsCadGJ6II5VU1aG70=\/970x0\/2016\/10\/11\/d4260898-6a6d-4f51-8560-a20fb805eb4e\/d3400-color-1333.jpg\" class=\"\" alt=\"d3400-color-1333.jpg\" height=\"0\" width=\"970\" \/><a class=\"clickToEnlarge\">Enlarge Image<\/a><\/div><figcaption>The colors generally look bright and saturated. Thankfully, manually white balancing with the camera is very easy if you feel the need. \n                                                    Lori Grunin\/CNET\n                                                <\/figcaption><\/figure><p> I don't have comparable performance results for the D3300, but overall the D3400 is sufficiently fast at focusing and shooting -- including continuous shooting -- for typical kids, pets and travel <a href=\"https:\/\/www.cnet.com\/topics\/photography\/\" data-annotation=\"true\" data-component=\"linkTracker\" data-link-tracker-options=\"{&quot;action&quot;:&quot;inline-annotation|Photography|CNET_CAT_TOPIC|225&quot;}\">photography<\/a>, as well as faster than many similarly priced <a href=\"https:\/\/www.cnet.com\/topics\/cameras\/\" data-annotation=\"true\" data-component=\"linkTracker\" data-link-tracker-options=\"{&quot;action&quot;:&quot;inline-annotation|Cameras|CNET_CAT_TOPIC|726&quot;}\">cameras<\/a>. It can autofocus during continuous shooting, though speed and accuracy depends upon which continuous autofocus mode you choose and how fast the subject is moving. It can't autoexpose at the same time, though, which means that it fixes the exposure on the first shot in the burst. So for subjects moving through daylight and shadows you'll probably get some that are under- or overexposed, depending upon initial conditions.<\/p>","created_at":"2018-11-07T16:53:04+0000","updated_at":"2018-11-07T16:53:04+0000","published_at":"2016-10-16T05:00:00+0000","published_by":["Lori Grunin"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"www.cnet.com","preview_picture":"https:\/\/cnet3.cbsistatic.com\/img\/Khap0K1N45RNIYrUo8NVt_06g4c=\/830x467\/2016\/09\/28\/12852994-eb5d-4288-9527-b1be806189ab\/12-nikon-d3400.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":593,"title":"Why isn't there an iOS email client that supports PGP or OpenGPG encryption?","url":"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption","content":"<div class=\"header\"><div id=\"wFqQIHXf1\"><div class=\"QuestionArea\"><div id=\"wSeq3YqN1\"><div id=\"wSeq3YqN3\"><div class=\"QuestionTopicHorizontalList TopicList\" id=\"__w2_wSeq3YqN4_topics_list\"><div id=\"wSeq3YqN14\"><div class=\"TopicListItems2 Toggle SimpleToggle QuestionTopicHorizontalListItems QuestionTopicListItems\" id=\"__w2_wSeq3YqN15__truncated\"><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wSeq3YqN17_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wSeq3YqN24\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/Pretty-Good-Privacy-PGP\" id=\"__w2_wSeq3YqN25_link\">Pretty Good Privacy (PGP)<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wSeq3YqN19_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wSeq3YqN26\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/Encryption\" id=\"__w2_wSeq3YqN27_link\">Encryption<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wSeq3YqN21_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wSeq3YqN28\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/iOS\" id=\"__w2_wSeq3YqN29_link\">iOS<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_wSeq3YqN23_topic_list_item\"><div class=\"u-inline-block\"><div id=\"wSeq3YqN30\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/Email\" id=\"__w2_wSeq3YqN31_link\">Email<\/a><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"wSeq3YqN5\"><div class=\"question_text_edit\"><h1>Why isn't there an iOS email client that supports PGP or OpenGPG encryption?<\/h1><\/div><\/div><div id=\"wSeq3YqN7\"><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wOsbrLGL9_answer\"><div><div id=\"wOsbrLGL10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wOsbrLGL19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Will-Price-1\" id=\"__w2_wOsbrLGL32_name_link\">Will Price<\/a>, co-founder of PGP Corporation, VP Engineering<\/div><div class=\"u-inline-block\" id=\"wOsbrLGL14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption\/answer\/Will-Price-1\" id=\"__w2_wOsbrLGL46_link\">Answered 265w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wOsbrLGL12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wOsbrLGL12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wOsbrLGL9_answer_content\"><div id=\"__w2_wOsbrLGL13_expanded_content\"><div id=\"wOsbrLGL16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">Gmail doesn't support PGP encryption because it is contrary to the goal of scanning your email content for advertisement targeting. If your email was encrypted client to client, that would be very undesirable from Google's perspective. At one point many years ago they briefly turned on PGP Signature support as a labs project, but that was quickly buried and the patient has not been seen since. Encryption I suspect has never been seriously considered.<p>Apple Mail doesn't doesn't support PGP encryption because as Bertrand Serlet, who was in charge at Apple then, said to me in 2002, everyone he knows who knows how to use PGP email has a Masters degree in Computer Science. In other words, it's just too hard to use. Of course, that is exactly the kind of problem Apple could have helped to solve in principle.<\/p><p>PGP Corporation as it were never produced an iOS product at all and very high-level internal forces (quite possibly funded by the NSA given the recent revelation that the NSA was funding subversive elements in companies that created encryption products) were causing the company to stop producing products for the consumer market. PGP Corporation no longer exists and it is safe to say that its acquirer Symantec wont be producing an email client for iOS.<\/p><p>GPG is really not even relevant to the discussion as it's never been a usable code base for anything other than its original command line roots and thus both the legality of using any of it on iOS combined with the technical challenge of making it work means one might as well start from scratch.<\/p><p>Creating a PGP email client today without the resources of Apple, Google, or the former PGP Corporation would be cost prohibitive at best. Either this needs to be an intrinsic part of a popular email client of which there are only two on iOS -- Gmail and Mail, or it needs to be built into the OS itself below the email client which is clearly never going to happen. Other than that, a third party would need to write an entire email client to do this right and that is just not realistic anymore.<\/p><p>Email as a technology is simply too complex these days for a startup to build from scratch, and iOS does not allow extending core technologies like email \u2013 there is no such thing as a driver or a proxy or a plugin anymore \u2013 which is for the best from a security perspective but completely eliminates classes of solutions that enabled PGP in the past.<\/p><p>So here we are in 2013 with no way for normal people (especially those without a degree in Computer Science) to use PGP email encryption. I would say the NSA was therefore uniquely successful in their goals over the last decade. One can only hope that this can be turned around over the next decade.<\/p><\/div><\/div><div id=\"wOsbrLGL24\"><div id=\"__w2_wOsbrLGL13_survey_wrapper\"><div id=\"wOsbrLGL34\"><div class=\"AnswerSurvey ComponentSurvey OutOfDateAnswerSurvey\" id=\"__w2_wOsbrLGL35_survey\"><div class=\"survey_wrapper\" id=\"__w2_wOsbrLGL35_survey_inner\"><div class=\"inner\"><div class=\"caption_wrapper\"><div class=\"locked\"><div class=\"caption\">Your feedback is private.<\/div><\/div><div class=\"question_wrapper\"><div class=\"flex_item\"><div class=\"icon_frame\"><div class=\"icon out_of_date_icon\"><\/div><div class=\"title_text\">Is this answer still relevant and up to date?<\/div><\/div><div class=\"flex_item button_group\" id=\"__w2_wOsbrLGL35_survey_buttons\"><input class=\"survey_button\" type=\"button\" value=\"Yes\" data-group=\"js-editable\" id=\"__w2_wOsbrLGL35_positive_action\" \/><input class=\"survey_button\" type=\"button\" value=\"No\" data-group=\"js-editable\" id=\"__w2_wOsbrLGL35_negative_action\" \/><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_w4UnBamM9_answer\"><div><div id=\"w4UnBamM10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"w4UnBamM19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/William-Emmanuel-Yu\" id=\"__w2_w4UnBamM32_name_link\">William Emmanuel Yu<\/a>, Now that everything is on public networks ... Cryptography is a must!<\/div><div class=\"u-inline-block\" id=\"w4UnBamM14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption\/answer\/William-Emmanuel-Yu\" id=\"__w2_w4UnBamM37_link\">Updated 208w ago<\/a> \u00b7 Author has 7.2k answers and 11.3m answer views<\/div><\/div><\/div><div id=\"__w2_w4UnBamM12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_w4UnBamM12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_w4UnBamM9_answer_content\"><div id=\"__w2_w4UnBamM13_expanded_content\"><div id=\"w4UnBamM16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">As mentioned, pleasure check out <a href=\"http:\/\/ipgmail.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"ipgmail.com\">iPGMail<\/a> and <a href=\"https:\/\/itunes.apple.com\/us\/app\/opengp\/id414003727?mt=8\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"apple.com\">oPenGP<\/a> if you want to use PGP or OpenPGP.<\/div><\/div><div id=\"w4UnBamM24\"><div id=\"__w2_w4UnBamM13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_w1wleFR39_answer\"><div><div id=\"w1wleFR310\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"w1wleFR319\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Wyllys-Ingersoll\" id=\"__w2_w1wleFR332_name_link\">Wyllys Ingersoll<\/a><\/div><div class=\"u-inline-block\" id=\"w1wleFR314\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption\/answer\/Wyllys-Ingersoll\" id=\"__w2_w1wleFR337_link\">Answered 207w ago<\/a><\/div><\/div><\/div><div id=\"__w2_w1wleFR312_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_w1wleFR312_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_w1wleFR39_answer_content\"><div id=\"__w2_w1wleFR313_expanded_content\"><div id=\"w1wleFR316\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">There is a PGP app for iOS:\u00a0 <a href=\"https:\/\/ipgmail.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"ipgmail.com\">iPGMail<\/a><p>It is a clean implementation of RFC 4880 in native Objective-C and its been in the iTunes app store for over 3 years.\u00a0 It is continually updated, improved and bugs are actively addressed. <\/p><p>The crypto library used is OpenSSL libcrypto (no copyleft issues), it did not require writing crypto ciphers from scratch.\u00a0 It works well on iPhone and iPads running iOS 6 or later.<\/p><\/div><\/div><div id=\"w1wleFR324\"><div id=\"__w2_w1wleFR313_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_w5BksXS19_answer\"><div><div id=\"w5BksXS110\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"w5BksXS119\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Fabian-Suchanek\" id=\"__w2_w5BksXS132_name_link\">Fabian Suchanek<\/a><\/div><div class=\"u-inline-block\" id=\"w5BksXS114\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption\/answer\/Fabian-Suchanek\" id=\"__w2_w5BksXS144_link\">Answered 25w ago<\/a><\/div><\/div><\/div><div id=\"__w2_w5BksXS112_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_w5BksXS112_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_w5BksXS19_answer_content\"><div id=\"__w2_w5BksXS113_expanded_content\"><div id=\"w5BksXS116\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">In the meantime, there are new apps available that do what you are asking for: Canary Mail and Kryptomail. Both are fully fledged email clients that have PGP encryption built in. As soon as the public key of a recipient is known to the app, outgoing messages are automatically encrypted. Incoming messages are automatically decrypted. Both happens transparently to the user; the only difference is that a small lock is shown.<\/p><p class=\"ui_qtext_para\">I have not tried out Kryptomail (because it is too expensive for my purposes), but I am trying out Canary mail. In my opinion, it is not as sleek as the default Mail app, but it does what it should.<\/p><\/div><\/div><div id=\"w5BksXS124\"><div id=\"__w2_w5BksXS113_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wNwBc9CU9_answer\"><div><div id=\"wNwBc9CU10\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wNwBc9CU19\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Lukas-Neumann\" id=\"__w2_wNwBc9CU33_name_link\">Lukas Neumann<\/a>, co-founder of mynigma.org<\/div><div class=\"u-inline-block\" id=\"wNwBc9CU14\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption\/answer\/Lukas-Neumann\" id=\"__w2_wNwBc9CU44_link\">Updated 208w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wNwBc9CU12_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wNwBc9CU12_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wNwBc9CU9_answer_content\"><div id=\"__w2_wNwBc9CU13_expanded_content\"><div id=\"wNwBc9CU16\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">[Disclaimer: I built an email client for iOS which implements its own encryption protocol]<p><a href=\"https:\/\/www.quora.com\/profile\/Will-Price-1\">Will Price<\/a> answer is great and hits the nail on the head.<\/p><p>Additionally, most PGP libraries have copyleft licenses which are incompatible with Apple's current licensing policy.<\/p><p>In order to add PGP to an email client, the developer would need to implement protocols, ciphers, and key management from scratch. No body wants to do this and it makes little sense to duplicate work which is already done.<\/p><p>At Mynigma we implemented our own encryption format and automated the whole key management, which makes the usage of secure encryption a lot easier.<\/p><\/div><\/div><div id=\"wNwBc9CU24\"><div id=\"__w2_wNwBc9CU13_survey_wrapper\"><\/div><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wSeq3YqN48_answer\"><div><div id=\"wSeq3YqN49\"><div class=\"AnswerHeader ContentHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wSeq3YqN58\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Nic-Whittle\" id=\"__w2_wSeq3YqN71_name_link\">Nic Whittle<\/a><\/div><div class=\"u-inline-block\" id=\"wSeq3YqN53\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/Why-isnt-there-an-iOS-email-client-that-supports-PGP-or-OpenGPG-encryption\/answer\/Nic-Whittle\" id=\"__w2_wSeq3YqN76_link\">Answered 26w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wSeq3YqN51_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wSeq3YqN51_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wSeq3YqN48_answer_content\"><div id=\"__w2_wSeq3YqN52_expanded_content\"><div id=\"wSeq3YqN55\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\"><p class=\"ui_qtext_para\">You can use EasyPGP which integrates pgp encryption into the iOS keyboard, making it simple and easy to encrypt\/decrypt within any app.<\/p><p class=\"ui_qtext_para\"><a href=\"http:\/\/easypgpapp.com\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"easypgpapp.com\">EasyPGP - PGP for iOS<\/a><\/p><\/div><\/div><div id=\"wSeq3YqN63\"><div id=\"__w2_wSeq3YqN52_survey_wrapper\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:53:02+0000","updated_at":"2018-11-07T16:53:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.quora.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":594,"title":"How to verify an imported GPG key","url":"https:\/\/serverfault.com\/questions\/569911\/how-to-verify-an-imported-gpg-key","content":"<blockquote>\n  <p>When I do this, I'm given the message \"This key is not certified with a trusted signature\". Is there anyway to make it trusted and better yet what's the proper way for doing so?<\/p>\n<\/blockquote><p>A \"trusted signature\" is a signature from a key that you trust, either because (a) you have personally verified that it belongs to the person to whom it claims to belong, or (b) because it has been signed by a key that you trust, possibly through a series of intermediate keys.<\/p><p>You can edit the trust level of keys by running \"gpg --edit-key \", and then using the <code>trust<\/code> command.  <a href=\"http:\/\/www.gnupg.org\/gph\/en\/manual\/x334.html\" rel=\"noreferrer\">This section<\/a> of the GPG manual discusses key trust, and it's worth a read: good security is hard.<\/p><p>Note that the warning \"This key is not certified with a trusted signature\" basically means, \"this thing could have been signed by anybody\".  I can create a key that claims to be for \"Internet Systems Consortium, Inc. (Signing key, 2013) \", and sign things with it, and GPG will happily confirm that yes, the things I signed were signed with my key.  To avoid this problem, you would presumably download the ISC GPG key from <a href=\"https:\/\/www.isc.org\/downloads\/software-support-policy\/openpgp-key\/\" rel=\"noreferrer\">the website<\/a> and either trust it ultimately (\"I believe this entity can certify itself\") or sign it with your ultimately-trusted private key.  Without proper management of key trust, signature verification is mostly theater.<\/p><blockquote>\n  <p>Find out when it expires? <\/p>\n<\/blockquote><p>Running <code>gpg -k &lt;keyid&gt;<\/code> will show you when a given key expires.  For example, I created a key that expires tomorrow, and <code>gpg -k &lt;keyid&gt;<\/code> gives me:<\/p><pre>$ gpg -k 0xD4C2B757C3FAE256\npub   2048R\/0xD4C2B757C3FAE256 2014-01-26 [expires: 2014-01-27]\nuid                 [ultimate] Test User &lt;testuser@example.com&gt;\nsub   2048R\/0xE87A56CDCC670D7A 2014-01-26 [expires: 2014-01-27]\n<\/pre><p>You can see that the expiration dates on subkeys are clearly marked.  Note that subkeys used for signing and encryption may have different expiration dates from the primary key.  You can read more about subkeys <a href=\"https:\/\/wiki.debian.org\/subkeys\" rel=\"noreferrer\">here<\/a>.<\/p><blockquote>\n  <p>In fact does GPG tells me when the key I've imported has already expired when I do a \"gpg --verify\"?<\/p>\n<\/blockquote><p>Yes, GPG will notify you about an expired key.  Note that this does not necessarily represent a problem: the signature was valid when the document was signed.<\/p><blockquote>\n  <p>Update the key. Do I have to delete the key and re-import when this happens?<\/p>\n<\/blockquote><p>You should have you GPG environment configured to use a keyserver, and periodically run <code>gpg --refresh-keys<\/code>.  This will update any keys in your keyring with new information from the keyserver, which may include:<\/p><ul><li>new expiration dates<\/li>\n<li>additional signatures on the key<\/li>\n<\/ul><p>If a person or organization begins using a new key, you would just add it to your keychain -- you would not need to delete the existing key.<\/p>","created_at":"2018-11-07T16:53:02+0000","updated_at":"2018-11-07T16:53:02+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"serverfault.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/serverfault\/img\/apple-touch-icon@2.png?v=9b1f48ae296b","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":595,"title":"A Taste of Rust \u2013 Evan Miller","url":"http:\/\/www.evanmiller.org\/a-taste-of-rust.html","content":"<p>By <a href=\"http:\/\/www.evanmiller.org\/\">Evan Miller<\/a><\/p><p><em>May 14, 2015<\/em><\/p><p><em><strong>Update<\/strong>: Steve Klabnik has <a href=\"https:\/\/news.ycombinator.com\/item?id=9546757\">responded<\/a> to some of the issues I brought up here. It looks like I did not fully appreciate Rust\u2019s optimization levels, among other things. When you\u2019re done with this, go read his comments!<\/em><\/p><p>The word <em>rust<\/em> is the last thing I want to think about when I am hungry and\/or fantasizing about building something that will endure through the ages. As a name for a programming language, \u201cRust\u201d doesn\u2019t exactly inspire confidence, but to take your mind off of ancient AMC Pacers and eating a bucket full of rusty nails, I\u2019ll let you in on a couple of folk etymologies for the language, only one of which I just made up.<\/p><p>The first etymology is an <a href=\"https:\/\/www.reddit.com\/r\/rust\/comments\/27jvdt\/internet_archaeology_the_definitive_endall_source\/\">apocryphal IRC chat log<\/a> from the language\u2019s creator. It is not made-up, at least not by me.<\/p><code>&lt;graydon&gt; I think I named it after fungi. rusts are amazing creatures.<br \/>&lt;graydon&gt; Five-lifecycle-phase heteroecious parasites. I mean, that\u2019s just _crazy_.<br \/>&lt;graydon&gt; talk about over-engineered for survival&lt;jonanin&gt; what does that mean? :]<br \/><\/code><p>Isn\u2019t a death-resistant parasitic fungus so much better to contemplate than tiny bits of corroded metal lacerating the insides of your mouth?<\/p><p>The second etymology is my own interpretation, but I think it works well \u2014 so well, in fact, that after reading the next three paragraphs you\u2019ll probably forget all about ineradicable fungal diseases, as well as the tetanus-infused blood dribbling down your chin.<\/p><p>If you\u2019ve ever seen a dark-brown skyscraper, it was probably made from a special steel called Corrosive Tensile Steel. Corten steel (sometimes spelled, with all the guile of an early programming language, COR-TEN) is deliberately designed to oxidize; the resulting rust forms a protective, weather-proof coating on the steel, so from an engineering perspective, the building never needs to be painted (or eaten).<\/p><p>Corten is used in a very famous <a href=\"https:\/\/en.wikipedia.org\/wiki\/Chicago_Picasso\">Picasso sculpture<\/a>, as well as in the Barclays Center in Brooklyn, the U.S. Steel building in Pittsburgh, and the Daley Center Building in Chicago. Buildings clad in Corten start out a neutral gray color, then turn bright orange after a couple of rains (usually discoloring nearby gutters and sidewalks in the process), and after a year or so, settle into a dark brown.<\/p><div><img src=\"http:\/\/www.evanmiller.org\/images\/a-taste-of-rust\/picasso1.jpg\" height=\"400\" alt=\"image\" \/>\u00a0<img src=\"http:\/\/www.evanmiller.org\/images\/a-taste-of-rust\/picasso2.jpg\" height=\"400\" alt=\"image\" \/><p><em>A Picasso sculpture made of Corten, in two stages of oxidation.<br \/>The building behind the sculpture is clad in the same material.<\/em><\/p><\/div><p>I like to think of Rust, the programming language, as a similarly earth-toned protective layer, but over a piece of software\u00a0instead of a piece of steel. The Rust language won\u2019t win any beauty contests, but it is impervious to the various elements that may ravage a multi-threaded software program. Initially you might complain about Rust\u2019s requirements and appearance, but in time it will feel normal, and eventually you\u2019ll start having smug thoughts about the reduced costs of long-term maintenance. So there you have it \u2014 Rust, the Corten steel of modern programming languages.<\/p><p>I admit, I only started learning Rust a few weeks ago. But I had so much fun <a href=\"http:\/\/www.evanmiller.org\/four-days-of-go.html\">learning Go last month<\/a> that I decided to write another programming language review. Rust 1.0 (final) is scheduled to be released tomorrow; what follows are my impressions of the Rust 1.0 nightly builds from reading docs and blog posts and writing a small library in Rust. YMMV, but hopefully you\u2019ll enjoy the ride\u2026 in my AMC Pacer made out of Corten.<\/p><h2>Iron and oxygen<\/h2><p>Rust is a systems language. I\u2019m not sure what that term means, but it seems to imply some combination of native code compilation, not being Fortran, and making no mention of category theory in the documentation. For what it\u2019s worth, there\u2019s a graduate student writing an operating system kernel in Rust, so whatever your definition of a systems language, Rust probably meets it.<\/p><p>The Wikipedia page on Rust lists sixteen \u201cinfluences,\u201d or sources of linguistic inspiration, half of which I had heard of before. Rust\u2019s influences, in alphabetical order, are: Alef, C#, C++, Cyclone, Erlang, Haskell, Hermes, Limbo, Newsqueak, NIL, OCaml, Python, Ruby, Scheme, Standard ML, and Swift.<\/p><p>I like the idea of a polyglot programming language that takes ideas from many places, the more obscure the better. (I especially appreciate the nihilistic audacity of a language named NIL.) I\u2019m not sure if the Wikipedia list is by any means complete or accurate, but I could detect features lifted from at least a few on the list. In practice, I found that while Rust\u2019s functional influences are an ergonomic improvement over traditional systems languages, the gluttony of influences has created some tension in the language design between competing paradigms.<\/p><p>Here\u2019s a quick example that may leave you scratching your head. Rust\u2019s <code>if<\/code> statement is also an expression, so you can do something like this:<\/p><pre>let x = if something &gt; 0 { 2 } else { 4 };\n<\/pre><p>Pretty great, right? But then, this won\u2019t compile:<\/p><pre>let x = if something &gt; 0 { 2 };\n<\/pre><p>The reason is that the phantom <code>else<\/code> clause evaluates to an empty tuple <code>()<\/code>, which is type-incompatible with 2. (Rust is strongly, statically typed with type inference.) Fair enough. Rather counterintuitively, the following line <em>will<\/em> compile:<\/p><pre>let x = if something &gt; 0 { () };\n<\/pre><p>The <code>if<\/code> clause and the phantom <code>else<\/code> clause have the same type <code>()<\/code>, so the above line generates no qualms from the compiler. While I understand the appeal of an <code>if<\/code> expression, juggling functional and imperative programming styles in a strongly typed universe creates some oddities such as the above.<\/p><p>As another example of the clash of paradigms, although Rust supports systems-and-arrays programming, it has renounced C-style <code>for<\/code> loops (initialize, test, increment) in favor of The Iterator; the documentation goes even further, renouncing indexed accesses altogether in favor of index-less Iteration. Iterators are a great and reusable way to encapsulate your blah-blah-blah-blah-blah, but of course they\u2019re impossible to optimize. The value of <code>.next()<\/code> remains a kind of holy mystery shrouded behind a non-pure function call, so anytime you have an iterator-style loop, the compiler is going to be left twiddling its thumbs.<\/p><p>As a simple demonstration, here\u2019s an idiomatic Rust program that uses an iterator:<\/p><pre>fn main() {\n    let mut sum = 0.0;\n    for i in 0..10000000 {\n        sum += i as f64;\n    }\n    println!(\u201cSum: {}\u201d, sum);\n}\n<\/pre><p>It runs about five times slower than the equivalent program that uses a <code>while<\/code> loop:<\/p><pre>fn main() {\n    let mut sum = 0.0;\n    let mut i = 0;\n    while i &lt; 10000000 {\n        sum += i as f64;\n        i += 1;\n    }\n    println!(\u201cSum: {}\u201d, sum);\n}\n<\/pre><p>The emphasis on iterators causes me extra sadness as Rust has contiguous-memory, natively-typed arrays, which would otherwise be perfect candidates for optimization, and Rust uses LLVM on the backend, which can optimize loops eight ways to next Sunday; but all that is thrown out the window in favor of The Noble Iterator. On the other hand, there\u2019s a SIMD module in the standard library, so maybe they care about CPU design after all? I\u2019m not sure. Rust feels like it\u2019s being pulled between the demands of efficiency (exemplified by its native-everything and choice of LLVM) and the modern definition of beautiful code (which I take to be, \u201clots of chained function calls\u201d).<\/p><p>One final example of Rust\u2019s unresolved identity before I move on to more interesting things. In addition to iterators, Rust has lazy evaluation of lists, so you can do things like create an infinite list, as in the following:<\/p><pre>let y = (1..)\n<\/pre><p>\u2026then chain together \u201citerator adaptors,\u201d like this example from the documentation:<\/p><pre>for i in (1..).step_by(5).take(5) {\n    println!(\u201c{}\u201d, i);\n}\n<\/pre><p>I\u2019ve personally never needed a list larger than the size of the known universe, but to each his own. Sipping from the chalice of infinity may feel like you\u2019re dining beneath the golden shields of Valhalla, but if you\u2019re coming from not-a-Haskell, you might be surprised at what code actually gets executed, and when. But then Haskell programmers won\u2019t necessarily feel at home, either; Rust does not have tail-call optimization, or any facilities for marking functions as pure, so the compiler can\u2019t do the sort of functional optimization that Haskell programmers have come to expect out of Scotland.<\/p><p>The Rust language creators apparently felt the multi-paradigm tension, too, and have ripped out a number of previously-announced features during the design and development of Rust; I\u2019ll discuss some ex-features (including green threads and garbage collection) later on. But for now I want to go over the fun stuff.<\/p><h2>Rust: The fun stuff<\/h2><p>Most tutorials on Rust start out by discussing Rust\u2019s concepts of memory safety, pointer ownership, borrowed references, and the like. They\u2019re important concepts in the language, but their immediate introduction is a bit like showing up at a party and being told in haste which bathrooms are off-limits. (Spoiler: most of them.) I generally like to take my coat off first.<\/p><p>So for the moment I\u2019m going to skip all the stuff about memory safety and references and threads and tell you some of my favorite (other) parts about the language, in no particular order.<\/p><p>Native types in Rust are intelligently named: <code>i32<\/code>, <code>u32<\/code>, <code>f32<\/code>, <code>f64<\/code>, and so on. These are, indisputably, the correct names for native types.<\/p><p>Destructuring assignment: awesome. All languages which don\u2019t have it should adopt it. Simple example:<\/p><pre>let ((a, b), c) = ((1, 2), 3);\n\/\/ a, b, and c are now bound\nlet (d, (e, f)) = ((4, 5), 6)\n\/\/ compile-time error: mismatched types\n<\/pre><p>More complex example:<\/p><pre>\/\/ What is a rectangle?\nstruct Rectangle { origin: (u32, u32), size: (u32, u32) }\n\/\/ Create an instance\nlet r1 = Rect { origin: (1, 2), size: (6, 5) };\n\/\/ Now use destructuring assignment\nlet Rect { origin: (x, y), size: (width, height) } = r1;\n\/\/ x, y, width, and height are now bound\n<\/pre><p>The <code>match<\/code> keyword: also awesome, basically a switch with destructuring assignment.<\/p><p>Like <code>if<\/code>, <code>match<\/code> is an expression, not a statement, so it can be used as an rvalue. But unlike <code>if<\/code>, <code>match<\/code> doesn\u2019t suffer from the phantom-<code>else<\/code> problem. The compiler uses type-checking to guarantee that the match will match <em>something<\/em> \u2014 or speaking more precisely, the compiler will complain and error out if a no-match situation is possible. (You can always use a bare underscore <code>_<\/code> as a catchall expression.)<\/p><p>The <code>match<\/code> keyword is an excellent language feature, but there are a couple of shortcomings that prevent my full enjoyment of it. The first is that the only way to express equality in the destructured assignments is to use guards. That is, you can\u2019t do this:<\/p><pre>match (1, 2, 3) {\n    (a, a, a) =&gt; \u201cequal!\u201d,\n  _ =&gt; \u201cnot equal!\u201d,\n}\n<\/pre><p>Instead, you have to do this:<\/p><pre>match (1, 2, 3) {\n    (a, b, c) if a == b &amp;&amp; b == c =&gt; \u201cequal!\u201d,\n  _ =&gt; \u201cnot equal!\u201d,\n}\n<\/pre><p>Erlang allows the former pattern, which makes for much more succinct code than requiring separate assignments for things that end up being the same anyway. It would be handy if Rust offered a similar facility.<\/p><p>My second qualm with Mr. Match Keyword is the way the compiler determines completeness. I said before \u201cthe compiler will complain and error out if a no-match situation is possible,\u201d but it would be better if that statement read <em>if and only if<\/em>. Rust uses something called Algebraic Data Types to analyze the match patterns, which sounds fancy and I only sort-of understand it. But in its analysis, the compiler only looks at types and discrete enumerations; it cannot, for example, tell whether every possible integer value has been considered. This construction, for instance, results in a compiler error:<\/p><pre>match 100 {\n    y if y &gt; 0 =&gt; \u201cpositive\u201d,\n    y if y == 0 =&gt; \u201czero\u201d,\n    y if y &lt; 0 =&gt; \u201cnegative\u201d,\n};\n<\/pre><p>The pattern list looks pretty exhaustive to me, but Rust wouldn\u2019t know it. I\u2019m that sure someone who is versed in type theory will send me an email explain how what I want is impossible unless P=NP, or something like that, but all I\u2019m saying is, it\u2019d be a nice feature to have. Are \u201cAlgebraic Data Values\u201d a thing? They should be.<\/p><p>It\u2019s a small touch, but Rust lets you nest function definitions, like so:<\/p><pre>fn a_function() {\n    fn b_function() {\n        fn c_function() {\n        }\n        c_function(); \/\/ works\n    }\n    b_function(); \/\/ works\n    c_function(); \/\/ error: unresolved name `c_function`\n}\n<\/pre><p>Neat, huh? With other languages, I\u2019m never quite sure where to put helper functions. I\u2019m usually wary of factoring code into small, \u201cbeautiful\u201d functions because I\u2019m afraid they\u2019ll end up under the couch cushions, or behind the radiator next to my car keys. With Rust, I can build up a kind of organic tree of function definitions, each scoped to the place where they\u2019re actually going to be used, and promote them up the tree as they take on the Platonic form of Reusable Code.<\/p><p>Mutability rules: also great. Variables are immutable by default, and mutable with the <code>mut<\/code> keyword. It took me a little while to come to grips with the mutable reference operator <code>&amp;mut<\/code>, but <code>&amp;mut<\/code> and I now have a kind of respectful understanding, I think. Data, by the way, inherits the mutability of its enclosing structure. This is in contrast to C, where I feel like I have to write <code>const<\/code> in about 8 different places just to be double-extra sure, only to have someone else\u2019s cast operator make a mockery of all my precautions.<\/p><p>Functions in Rust are dispatched statically, if the actual data type is known as compile-time, or dynamically, if only the interface is known. (Interfaces in Rust are called \u201ctraits\u201d.) As an added bonus, there\u2019s a feature called \u201ctype erasure\u201d so you can force a dynamic dispatch to prevent compiling lots of pointlessly specialized functions. This is a good compromise between flexibility and performance, while remaining more or less transparent to the typical user.<\/p><p>Is resource acquisition the same thing as initialization? I\u2019m not sure, but C++ programmers will appreciate Rust\u2019s capacity for RAII-style programming, the happy place where all your memory is freed and all your file descriptors are closed in the course of normal object deallocation. You don\u2019t need to explicitly close or free most things in Rust, even in a deferred manner as in Go, because the Rust compiler figures it out in advance. The RAII pattern works well here because (like C++) Rust doesn\u2019t have a garbage collector, so you won\u2019t have open file descriptors floating around all week waiting for Garbage Pickup Day.<\/p><p>On the whole, even leaving aside the borrowed references \/ bathroom situation, which I promise I\u2019ll talk about in a minute, there\u2019s enough fun stuff going on in Rust to keep me interested as a performant, systems-y language. Many of its features and design decisions should be emulated in other languages. But there are a few areas where I wish Rust had taken cues from other languages, that is, languages in addition to the Official Influences mentioned on the Wikipedia page.<\/p><h2>The not-so-fun stuff<\/h2><p>When I\u2019m trying to get a grip on a new language, I can usually count on assignment as a rudimentary building block from which greater epistemological edifices may be built.<\/p><p>But assignment in Rust is not a totally trivial topic. Thanks to Rust\u2019s memory checking, this program produces an error:<\/p><pre>enum MyValue { Digit(i32) }\nfn main() {\n    let x = MyValue::Digit(10);\n    let y = x;\n    let z = x;\n}\n<\/pre><p>The reason is that <code>z<\/code> might (later) mess with <code>x<\/code>, leaving <code>y<\/code> in an invalid state (a consequence of Rust\u2019s strict memory checking \u2014 more on that later). Fair enough. But then changing the top of the file to:<\/p><pre>#[derive(Copy, Clone)]\nenum MyValue { Digit(i32) }\n<\/pre><p>makes the program compile. Now, instead of binding <code>y<\/code> to the value represented by <code>x<\/code>, the assignment operator copies the value of <code>x<\/code> to <code>y<\/code>. So <code>z<\/code> can have no possible effect on <code>y<\/code>, and the memory-safety gods are happy.<\/p><p>To me it seems a little strange that a pragma on the <code>enum<\/code> affects its assignment semantics. It makes it difficult to reason about a program without first reading all the definitions. The binding-copy duality, by the way, is another artifact of Rust\u2019s mixed-paradigm heritage. Whereas functional languages tend to use bindings to map variable names to their values in the invisible spirit world of no-copy immutable data, imperative languages take names at face value, and assignment is always a copy (copying a pointer, perhaps, but still a copy). By attempting to straddle both worlds, Rust rather inelegantly overloads the assignment operator to mean either binding or copying. The meaning of = depends, rather Clintonesquely, on the context.<\/p><p>A related oddity is that assignment, evaluated as an expression, returns <code>()<\/code>. I have no idea why that is the case, or how that could ever be more useful than returning the newly-assigned rvalue.<\/p><p>Another feature where a pragma seems out of place is inlining, although in this case I can\u2019t really blame Rust for it. You can tell Rust you think a function should be inlined with a <code>#inline<\/code> pragma above the function (similar to C\u2019s <code>inline<\/code> keyword). I am not an expert in compilation and linking, so maybe this is a dumb idea, but I\u2019ve always thought it should be up to the caller to say which functions they\u2019d like inlined, and the function should just mind its own business. (Did someone in the crowd just say \u201cAmen\u201d?)<\/p><p>Rust is being developed side-by-side with a browser rendering engine called Servo, written in you-know-what. Since the purpose of a browser rendering engine is to convert text files (i.e. HTML, CSS, and JavaScript) into something more attractive, it seemed like a safe bet that Rust would have excellent support for parsing things. Sadly, Rust is not a target for my <a href=\"http:\/\/www.colm.net\/open-source\/ragel\/\">favorite parser generator<\/a>, and the lexers in Servo don\u2019t look much better than C-style state machines, with lots of matching (switching) on character literals. There are newish libraries on GitHub implementing LALR and PEG parser generators for Rust, but on the whole, I am not left itching to write my next parser in Rust.<\/p><p>Concurrency in Rust is straightforward: you get a choice between OS threads, OS threads, and OS threads. When Rust was originally announced, the team had ambitions to pursue the multi-core MacGuffin with a green-threaded actor model, but they found out that it\u2019s very hard to do green threads with native code compilation. (Go, for example, does native code compilation, but switches contexts only on I\/O; Erlang can context-switch during a long-running computation, as well as on I\/O, but only because it\u2019s running in a virtual machine that can count the number of byte-code ops.)<\/p><p>Nonetheless, OS threads are very well understood (by\u2026 people), and so Rust has taken up a different challenge: eliminating crashes from otherwise traditional, \u201cboring\u201d multi-threaded programs. In the sense that Rust has abandoned its original vision in favor of pursuing more modest and achievable goals, Rust does feel rather grown-up. The team, I think, made the right decision focusing their efforts on memory safety; it lowers the learning curve for the target audience, and gives Rust a corner booth in the increasingly crowded marketplace for programming languages.<\/p><p>Of course, lest the Real Programmers in the audience begin to worry, it\u2019s still possible to crash a Rust program. The <code>unsafe<\/code> keyword is a kind of fireman\u2019s axe that you\u2019re allowed to use if you promise you know what you\u2019re doing; it lets you blit, bswap, and bit-twiddle away. <code>unsafe<\/code> is a good approach, balancing safety with flexibility, and in general limits application crashes to areas of code explicitly marked unsafe.<\/p><p>But as with the <code>match<\/code> keyword, which as in the example above can possess impossible beliefs about integers, the Rust compiler is at times more cautious than intelligent. A common pattern in parallel computing, for example, is to concurrently perform an operation over non-overlapping parts of a shared data array. As a simple (single-threaded) example, the following code will not compile in Rust:<\/p><pre>fn main() {\n    let mut v = vec![1, 2]; \/\/ a mutable vector\n    let a = &amp;mut v[0]; \/\/ mutable reference to 1st element\n    let b = &amp;mut v[1]; \/\/ mutable reference to 2nd element\n}\n<\/pre><p>It\u2019s a perfectly valid code, and accessing non-overlapping slices of an array is obviously safe. But the above program will cause the Rust compiler to hem and haw about competing reference to <code>v<\/code>.<\/p><p>However, Rust\u2019s parallel-computation cabal has a kind of secret handshake called <code>split_at_mut<\/code>. Although the Rust compiler is unable to reason about whether specific index values overlap, it is willing to believe that <code>split_at_mut<\/code> produces non-overlapping slices, so this equivalent program compiles just fine:<\/p><pre>fn main() {\n    let mut v = vec![1, 2];\n    let (a_slice, b_slice) = v.split_at_mut(1);\n    let (a, b) = (&amp;mut a_slice[0], &amp;mut b_slice[0]);\n}\n<\/pre><p>Most parallel computations end up using <code>split_at_mut<\/code>. In the interest of code clarity, it would be nice if the Rust compiler got rid of the <code>split_at_mut<\/code> secret password and could reason sanely about slice literals and array indexes. For the record, the Nim language manages to get this right, and the Rust folks might look to it for inspiration.<\/p><p>(It is also worth noting that processing non-overlapping slices in parallel is destined to come into mortal conflict with The Iterator, which is by its nature sequential. In that sense parallel processing is another victim of Rust\u2019s multi-paradigm nature; the language would be easier to parallelize if it were more aggressively imperative, or more stringently functional, but by occupying the middle ground between The Array and The Function, it loses out on well-known techniques from both camps.)<\/p><p>Because of Rust\u2019s strict memory-checking requirements, error messages play a rather crucial role in learning to write Rust code that is safe and error-free (the same thing, outside an <code>unsafe<\/code> block). Rust\u2019s error messages are quite specific, perhaps to a fault; they tend to have a lot of jargon, which I suppose is necessary because Rust has its own private universe of memory-related concepts. But in learning Rust, I often felt lost, alone, and confused amid all the reprinted code and messages about block suffixes.<\/p><p>The error messages can also be, shall we say, overly effusive. For example, the official way to print a line is the <code>println!<\/code> macro; if you have some kind of error in the arguments, Rust treats you to a full exposition of how and where the <code>println!<\/code> macro was expanded, which is knowledge that I never particularly aspired to possess. Although the programming profession has, over the years, cultivated a reputation for providing little to no information about where things went wrong, it is in this case possible to have too much of a good thing. Perhaps when they\u2019re done writing a browser engine, the next Big Rust Project should be an IDE that reads Rust error messages just draws a giant crimson arrow through my code that says \u201cMove this line here, you dummy.\u201d<\/p><h2>C and back again<\/h2><p>As my first Rust project, I thought it would be very funny to fork the Servo web rendering engine, and try to implement a terminal-based web browser \u00e0 la Lynx or w3m. I figured I could get a rough proof-of-concept in a weekend.<\/p><p>Several naps later, I realized that the idea of working on a web browser led inevitably to thoughts about the meaning of life, so I decided instead to take on, well, a more modest and achievable project. Thus was born, in its place, <em>Excalibur: Destroyer of Spreadsheets<\/em>.<\/p><p>(Excalibur is the second installment in Terminally Insane Programming, my neo-Gothic terminal revival movement. The first project was <em>Hecate: The Hex Editor From Hell<\/em>, <a href=\"https:\/\/github.com\/evanmiller\/hecate\">now available in source code form<\/a>. One day I hope these projects will form either a trilogy, a trilogy of trilogies, or ideally, a trilogy of trilogies of trilogies, at which point the earth will be engulfed in calamities and the true nature of the universe will be revealed to mankind as a one-line Mathematica program.)<\/p><p>Excalibur\u2019s <em>raison d\u2019etre<\/em> is this: I do a lot of work with spreadsheets, and often want to see if all the formulas in a spreadsheet look right. To see a formula in any spreadsheet program \u2014 and they all use the VisiCalc interface from 1975 \u2014 you have to put your cursor over the cell and look at the top of the screen, which means you can only inspect one formula at a time. I\u2019d like a program where the formulas appear below each value, something like this:<\/p><div><img src=\"http:\/\/www.evanmiller.org\/images\/a-taste-of-rust\/excalibur.png\" width=\"512\" alt=\"image\" \/><\/div><p>(ASCII art done in Vim, my new favorite mockup tool.)<\/p><p>I figured this would be a good test of interfacing with C libraries \u2014 <a href=\"https:\/\/sourceforge.net\/projects\/libxls\/\">libxls<\/a> for reading spreadsheets, <a href=\"https:\/\/github.com\/nsf\/termbox\">termbox<\/a> for making a terminal program \u2014 so I was off to the races. And I don\u2019t mean data races.<\/p><p>The first task was to wrap up libxls, that is, expose the library functions for reading Excel spreadsheets to Rust somehow. Rust has a decent FFI mechanism, so I didn\u2019t have to actually write any glue code in C. However, at times I felt like I was jumping through an unnecessary number of hoops just to call a C function and convert its return value to something Rusty.<\/p><p>Definitions for standard C type (<code>int<\/code>, <code>long<\/code>, <code>void *<\/code>) are nested deep into packages, and can be hard to find. For example, the C <code>int<\/code> and <code>char<\/code> types are located, rather improbably, at <code>libc::types::os::arch::c95::c_int<\/code> and <code>libc::types::os::arch::c95::c_char<\/code>. The symbols are also accessible from the top-level <code>libc<\/code> namespace, so you don\u2019t actually have to type out a complicated module hierarchy every time, but the online documentation only lists the specific types at their five-layers-deep locations. If you want a full list, bring a hard hat and coveralls.<\/p><p>Rust and C have somewhat different concepts of strings, and so converting them is another minor chore. Here\u2019s a function that calls <code>xls_getVersion()<\/code>, which returns a C string, and converts it to a Rust String:<\/p><pre>pub fn version() -&gt; String {\n    let c_string = unsafe { CStr::from_ptr(xls_getVersion()) };\n    str::from_utf8(c_string.to_bytes()).unwrap().to_string()\n}\n<\/pre><p>I found a variant of that code with a big green check next to it on StackOverflow. I understand what it\u2019s doing, but five function calls (<code>from_ptr<\/code>, <code>from_utf8<\/code>, <code>to_bytes<\/code>, <code>unwrap<\/code>, <code>to_string<\/code>) just to convert a C string to a Rust string seems like an excessive amount of ceremony, at least for a language that keeps using the word \u201csystems\u201d on its website.<\/p><p>Once you get the types squared away, actually calling C functions is straightforward, albeit somewhat tedious. Rust won\u2019t read C header files, so you have to manually declare each function you want to call by wrapping it in an <code>extern<\/code> block, like this:<\/p><pre>extern {\n    fn xls_getVersion() -&gt; *const libc::c_char;\n}\n<\/pre><p>That part is easy, but things get a little trickier if a library\u2019s C API expects you to access struct fields directly rather than use accessor functions. In that case you have to mirror all of the structs manually, as with this C struct in libxls:<\/p><pre>struct st_sheet_data {\n    DWORD filepos;\n    BYTE visibility;\n    BYTE type;\n    BYTE* name;\n}\n<\/pre><p>And its doppelg\u00e4nger in Rust:<\/p><pre>#[repr(C)]\nstruct NativeWorksheetData {\n    filepos : u32,\n    visibility : u8, \n    ws_type : u8,\n    name : *const i8,\n}\n<\/pre><p>The <code>#[repr(C)]<\/code> bit at the top tells Rust that you want a Rust struct that is bit-compatible with the same struct declared in C. It\u2019s a great language feature, and there\u2019s a variant that lets you get a \u201cpacked\u201d structure (a common pattern in blit-heavy C libraries). In practice, though, the feature has its limits. For example, the libxls API has an array-of-structs in a few places (array of worksheets, array of rows, array of cells, etc.); because Rust doesn\u2019t believe in pointer arithmetic, I found myself manually writing the pointer arithmetic logic, which was totally unsafe, and non-trivial on account of how C aligns structures and fields in memory. The resulting code feels like a crash waiting to happen.<\/p><p>With all the pointer-hopping I was doing, I was surprised to find out how the <code>unsafe<\/code> keyword actually works. My initial belief was that a function that does something <code>unsafe<\/code> must, itself, be <code>unsafe<\/code>, but in reality it\u2019s more of an honor code system. A function that does something <code>unsafe<\/code> (e.g. pointer arithmetic or calling C libraries) can be exported as safe. This may come as a surprise if you are using someone else\u2019s Rust libraries, as it\u2019s possible they\u2019re doing something <code>unsafe<\/code> under the hood and just not telling you about it.<\/p><p>Of course, the lax rules about un-safety are rather convenient if you want to do something <code>unsafe<\/code> in a trait (interface) implementation. If the <code>unsafe<\/code> keyword propagated up the call chain, then you\u2019d be unable to implement RAII or iterators with any unsafe logic, because in Rust, safe and unsafe functions have incompatible method signatures, and RAII and iterator methods are marked safe. So I\u2019m grateful that I can construct a Sheet Iterator, a Row Iterator, and a Cell Iterator in my spreadsheet library; they purport to be safe, and are eligible for all the abstract wonders of Iteration, but in reality they call out to C code that could incinerate the user\u2019s hard drive. In that sense, the <code>unsafe<\/code> keyword is merely advisory signage, like those warnings near prisons about picking up hitchhikers.<\/p><p>Actually building a Rust program that links to C involved one final, minor annoyance. The Rust documentation advises you to build things with Rust\u2019s do-it-all package manager, Cargo, but until a few days ago, Cargo didn\u2019t understand linker flags, so I couldn\u2019t tell it where to find libxls on my hard drive. The issue has been sort-of addressed with the new <code>cargo rustc<\/code> command. My library is building, for now, and I\u2019ll refrain from any invective about it.<\/p><p>Once my libxls wrapper was finished, built, and running, I was pleased with the result. Rust lets me call C functions and access C structs, but then expose the results at a higher level of abstraction with iterators and automatic handle-closing via RAII. I particularly liked using Rust\u2019s type system to express the kinds of values represented by a spreadsheet cell (StringValue, DoubleValue, FormulaValue, and so on); as a consequence, Rust client code for libxls is quite a bit cleaner than the equivalent C code, which requires a lot of branching to extract different kinds of data. For the benefit of posterity, I\u2019ve put the <a href=\"https:\/\/github.com\/evanmiller\/rust-xls\">rust-xls library code on GitHub<\/a>, although it probably won\u2019t work for you until I get a tiny libxls patch accepted upstream.<\/p><p>By way of confession, despite my best intentions, I did not actually get around to implementing <em>Excalibur: Destroyer of Spreadsheets<\/em>. An exposition of that experience will have to wait for another installment. Part of the reason is that I don\u2019t actually need <em>Excalibur<\/em> all that badly, but the other part is that it took me longer than expected to feel comfortable writing Rust code. Although Rust looks like just-another-curly-brace-language, it can be surprisingly difficult to write correct (compilable) code. At every turn, you must satisfy, and make periodic sacrifices to, the gods of memory ownership.<\/p><h2>The bathroom at the end of the universe<\/h2><p>Thus far I have deliberately avoided talking too much about Rust\u2019s memory-ownership model. The reason is that learning it is a pain, and it can feel like a set of mysterious rules as to when variables can be assigned or used as arguments to function calls. I wondered for a while whether Rust\u2019s hype and insurgent popularity was purely due to how difficult it was to write \u2014 a sort of fixed-gear bicycle for the computing world \u2014 and whether Rust programs achieved memory safety simply by being bathed in the tears of programmers.<\/p><p>But after working for a while within The Memory Ownership Rules, I\u2019ve developed a kind of grudging respect for them. Rather inundate you with code samples here, I\u2019ll try to explain how memory ownership works at a high level, so when you sit down with a proper tutorial, you (hopefully) won\u2019t think that Rust\u2019s memory requirements are totally arbitrary and dumb, like I did at first.<\/p><p>To understand where The Rules come from, it helps to understand how other languages manage memory.<\/p><p>In modern languages, memory management usually happens with garbage collection or reference counting. Garbage collection is convenient because the programmer can usually bang out code without thinking too hard about memory allocation. Reference-counting requires a little more work on the part of the programmer, but has more predictable performance without those nasty CPU spikes while the garbage collector cleans house.<\/p><p>Rust initially had a special kind of garbage-collected variable, denoted with <code>@<\/code>, but the designers got rid of it in the interest of simplifying the language. Rust does have reference-counted data types in the form of <code>Rc<\/code> and <code>Arc<\/code>, but most memory is actually not refcounted \u2014 at least not at run-time.<\/p><p>Rust\u2019s memory is essentially <em>reference-counted at compile-time, rather than run-time<\/em>, with a constraint that the refcount cannot exceed 1. Blocks of code \u201cown\u201d the references (Rust calls them bindings), and in the normal case, objects are deallocated as they go out of scope. To change ownership of the reference (binding), you have to either \u201cmove\u201d the reference to another variable or block of code, or make a copy of the underlying data. Attempting to access a variable that has been moved results in a compiler error.<\/p><p>In normal (run-time) reference-counted languages, the language needs some kind of support for weak references to avoid reference cycles and therefore memory leaks. Weak references do not contribute to the reference count, so they must be carefully managed in order to prevent accessing a deallocated object. Typically, an \u201cowning\u201d object will have a strong reference to an \u201cowned\u201d object, and the owned object will have a weak back-reference to the owner. (It\u2019s up to the programmer to ensure that the owning object scrubs its name off of all its stuff before it gets deallocated.)<\/p><p>Weak references are analogous to Rust\u2019s concept of \u201cborrowing\u201d. The compiler ensures that a weak reference (borrower) never outlives a strong reference (owner), so your code will never be able to access an deallocated object. Rust goes a bit further with \u201clifetimes\u201d, which lets you express the longevity of a reference in terms of other references \u2014 usually in the context of a function call or data structure. It sounds rather abstract, but in practice it just means something like \u201cthis function returns a value that lives as long as the thing you passed it\u201d, which ought to be a cause for joy for anyone who has worked on a large, pointer-happy C++ or Objective-C program. If you\u2019re feeling lazy, there\u2019s a special lifetime called <code>static<\/code>, which means the thing lives as long as the program is running.<\/p><p>By doing all the reference accounting at compile-time, Rust has \u201czero-cost abstractions\u201d, including structs that are bit-compatible with C (with <code>#[repr(C)]<\/code>, as above), and extremely predictable performance. By having a maximum refcount of 1 \u2014 which embodies the Rust\u2019s idea of ownership \u2014 Rust heads off all kinds of problems in concurrency, where (otherwise) multiple threads with references to the same object might cause data races, and produce the kinds of subtle bugs that plague multi-threaded C++ programs.<\/p><p>I\u2019m not sure if the ownership rule is actually helpful in single-threaded contexts, but it at least makes sense in light of Rust\u2019s green-threaded heritage. Even though the green-threaded architecture has been abandoned, the Rust designers have been thinking carefully about memory safety in concurrent environments since the language\u2019s inception, and this thinking is apparent in the memory model.<\/p><p>So there you have it, at least as I understand it \u2014 the Rust memory model is compile-time reference-counting, with extra rules in place to prevent data races, and added facilities (<code>Rc<\/code>, <code>Arc<\/code>) for doing old-fashioned reference-counting at run-time. The memory model imposes constraints on the code you write, and it takes time to come to terms with all the rules, but the resulting code has a kind of structural clarity that is lost in the equivalent C++ programs, where invalid pointers could lurk anywhere. I haven\u2019t worked with the language enough to give a personal bug-count testimonial, but I am inclined to believe Mozilla\u2019s claim that 50% of the security vulnerabilities found in Firefox would have been mitigated by Rust. The language feels like a metal that is harder to work with, but once it is beaten into shape, lasts that much longer.<\/p><h2>More is less<\/h2><p>And yet I keep thinking back to the Daley Center Building, that brown Corten structure sitting behind the Picasso sculpture in the pictures that I showed you at the beginning of the article.<\/p><div><img src=\"http:\/\/www.evanmiller.org\/images\/a-taste-of-rust\/daley.jpg\" alt=\"image\" \/><p><em>The Richard J. Daley Center Building, formerly the Chicago Civic Center.<br \/>The Picasso sculpture is visible at the bottom of the photograph.<\/em><\/p><\/div><p>Unveiled in 1965, the building was designed by a student of Mies van der Rohe, the famous \u201cless is more\u201d messiah of architectural modernism. Like the best works of Mies, the Daley Center has a kind of spare, Spartan, almost severe beauty to it. Every line serves a purpose, and nothing feels confused or out of place. The building adheres to the \u201cform follows function\u201d mantra of modernism: the underlying structure is evident from the outward appearance, and nothing is shrouded or hidden away. You know it is Corten because it is brown; it was once a shameless shade of bright orange, as oxidation is simply a reality of the material.<\/p><p>I have a nagging sadness about the Rust language because in some ways, it comes very close to expressing the formal purity of that building. It makes the reality of the computer evident in the structure of the language (native types, structs, pointers, arrays), and tames that reality so that it does not present the traditional dangers of systems programming (via Rust\u2019s compile-time checks). The small amount of Rust code that I wrote was remarkably clear in its purpose and intent, and was satisfying in its safety guarantees. (That is, apart from my Evel Knievel-style pointer jumping discussed earlier.)<\/p><p>But then reading more example code, I can\u2019t help but think back to Rust\u2019s fundamental identity problem. It\u2019s really not sure if Rust wants to be a functional language or not, and I have to wonder if the emphasis on iterators, builder patterns, and (semi-)functional programming is a kind of pragmatic response to the constraints imposed by Rust\u2019s memory-ownership rules. My fear is that newcomers to the language will try to get around the demands of the borrow-checker simply by writing things in a pseudo-functional, lots-of-chained-method-calls style, for which Rust is not all that well-designed, and which spurns, rather than celebrates, the realities of The Machine.<\/p><p>And that, I think, is the heart of the problem with Rust, in spite of its many virtues. If a large piece of software is the computing equivalent of a skyscraper, then you might think of a systems language as a style (school) of architecture. The major breakthrough of Mies\u2019s International style was to show what heights could be achieved simply by respecting, and celebrating, the physicality of glass and steel. Some people loved the aesthetic, others despised all the black and brown boxes jutting into their skylines; but regardless of opinion, the style had a kind of sure-footedness and sense of identity that is hard not to admire. That sure-footedness is lacking in Rust. It feels like it has two systems languages inside it \u2014 one functional, one imperative \u2014 and they would achieve fuller expression if only they didn\u2019t have to bend to each other. As long as Rust is trying to accommodate both The Array and The Function, it will never fully exploit the capabilities of The Machine, and its potential heights will remain undiscovered. It won\u2019t have the Sartrean self-confidence of a more narrowly defined language, and, in practical terms, I\u2019ll never quite be sure what the = operator is doing.<\/p><p>In that sense, perhaps we should try to forget my faux etymology about Corrosive Tensile steel and minimalist skyscrapers, and look back to that IRC chat log for the more accurate origin of Rust\u2019s name. A <em>rust<\/em> is a fungus with up to five distinct morphologies, or forms. (Humans have two: a sperm or egg, then that angsty ball of flesh.) I\u2019ll let Rust\u2019s creator explain the rest:<\/p><code>&lt;graydon&gt; fungi are amazingly robust<br \/>&lt;graydon&gt; to start, they are distributed organisms. not single cellular, but also no single point of failure.<br \/>&lt;graydon&gt; then depending on the fungi, they have more than just the usual 2 lifecycle phases of critters like us (somatic and gamete)&lt;jonanin&gt; ohhh<br \/>&lt;jonanin&gt; those kind of phases&lt;graydon&gt; they might have 3, 4, or 5 lifecycle stages. several of which might cross back on one another (meet and reproduce, restart the lineage) and\/or self-reproduce or reinfect&lt;jonanin&gt; but i mean<br \/>&lt;jonanin&gt; you have haploid gametes and diploid somatic cells right? what else could there be?&lt;graydon&gt; and in rusts, some of them actually alternate between multiple different hosts. so a crop failure or host death of one sort doesn\u2019t kill off the line.<br \/>&lt;graydon&gt; they can double up!<br \/>&lt;graydon&gt; http:\/\/en.wikipedia.org\/wiki\/Dikaryon<br \/>&lt;graydon&gt; it\u2019s madness. basically like someone was looking at sexual reproduction and said \u201cnah, way too failure-prone, let\u2019s see how many other variations we can do in parallel\u201d&lt;jonanin&gt; I can\u2019t really understand that lol.<br \/><\/code><br \/><hr \/><p><em>You\u2019re reading <a href=\"http:\/\/www.evanmiller.org\/\">evanmiller.org<\/a>, a random collection of math, tech, and musings. If you liked this you might also enjoy:\n    <a href=\"http:\/\/www.evanmiller.org\/a-review-of-perl-6.html\">A Review of Perl 6<\/a>\n        <a href=\"http:\/\/www.evanmiller.org\/four-days-of-go.html\">Four Days of Go<\/a>\n        <a href=\"http:\/\/www.evanmiller.org\/why-i-program-in-erlang.html\">Why I Program in Erlang<\/a>\n    <\/em>\n<\/p><hr \/><p><em>Get new articles as they\u2019re published, via <a href=\"https:\/\/twitter.com\/EvMill\">Twitter<\/a> or <a href=\"http:\/\/www.evanmiller.org\/news.xml\">RSS<\/a>.<\/em><\/p><hr \/><p><em>Want to look for statistical patterns in your MySQL, PostgreSQL, or SQLite database? My desktop statistics software <strong><a href=\"http:\/\/www.wizardmac.com\/\">Wizard<\/a><\/strong> can help you analyze <strong>more data in less time<\/strong> and <strong>communicate discoveries visually<\/strong> without spending days struggling with pointless command syntax. Check it out!<\/em><\/p><div><a href=\"http:\/\/www.wizardmac.com\/\"><img height=\"128\" width=\"128\" src=\"http:\/\/www.evanmiller.org\/images\/index\/wizard-nonpro.png\" alt=\"image\" \/><\/a><br \/><strong><a href=\"http:\/\/www.wizardmac.com\/\">Wizard<\/a><\/strong><br \/>Statistics the Mac way<\/div><hr \/><p><a href=\"http:\/\/www.evanmiller.org\/\">Back to Evan Miller\u2019s home page<\/a> \n\u2013 <a href=\"http:\/\/www.evanmiller.org\/news.xml\">Subscribe to RSS<\/a>\n\u2013 <a href=\"http:\/\/twitter.com\/EvMill\">Twitter<\/a> \n\u2013 <a href=\"https:\/\/www.youtube.com\/channel\/UC6PJEAiKFzEfq4stBoZz07Q\">YouTube<\/a> \n<\/p><hr \/>","created_at":"2018-11-07T16:53:02+0000","updated_at":"2018-11-07T16:53:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":36,"domain_name":"www.evanmiller.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":596,"title":"anonymity - What are the advantages of paying for a VPN service as opposed to hosting your own? - Information Security Stack Exchange","url":"https:\/\/security.stackexchange.com\/questions\/148750\/what-are-the-advantages-of-paying-for-a-vpn-service-as-opposed-to-hosting-your-o","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>For a person looking to use a VPN solution for personal use, what are the pros and cons of paying for a VPN service as opposed to just hosting a VPN on your own rented, in the cloud server, except for the obvious managed vs unmanaged argument?<\/p><\/div><h2 data-answercount=\"3\">&#13;\n                                3 Answers&#13;\n                                3&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>VPNs are designed around the concept of trust between 2 or more parties, and were intended for corporate\/enterprise use.<\/p>\n<p>The popularity of offering \"Free VPN\" or \"Hosted VPN\" solutions to the consumer market has dramatically increased. A lot of people seem to forget there has to be a \"trust\" involved. You wouldn't want to have some random person on your home network, so why would you tolerate that on a virtual up scale of that idea?<\/p>\n<p>Hosting your own private VPN solution has a key distinct advantage, <strong>you trust yourself<\/strong>. You know how data is handled, you know who can view VPN data as it's relayed, you can ensure it's quality, reliability, and anonymity. Most people will use a VPS (Virtual Private Server) for hosting a VPN.<\/p>\n<p>Using a third-party hosted solution does come at a cost of some trust, corporations are legally bound entities which <em>may<\/em> be required to hand over subscriber information. Some self-hosted solutions over come this with payments by bitcoins (for the virtual private server or physical hardware). In some countries, a corporation <em>may also<\/em> be required to log usage by their clients. Beyond this, you will have to address each company's privacy policy and terms of use.<\/p>\n<p>Tor was the answer for the consumer's idea of anonymity on the internet. This again has some issues with trust, mainly that entry\/exit nodes have full implicit trust. It is not made apparent if a \"rogue\" node was to log your connections, or pass your information in the clear.<\/p>\n<p>If you are not overly concerned about \"trust\", here are some bullet points to think about:<\/p>\n<p><strong>Self-Hosted<\/strong><\/p>\n<ul><li>Security and anonymity to your standards<\/li>\n<li>Can be faster, as you are the only one using the service<\/li>\n<li>Must keep up to date with patches\/security of software<\/li>\n<li>Your information is known to the hoster, which could be a third-party<\/li>\n<li>Only your hosting platform is aware of your IP address<\/li>\n<\/ul><p><strong>Out Sourced Hosting<\/strong><\/p>\n<ul><li>Cheaper, and generally more reliable\/constant service<\/li>\n<li>Don't have to worry about patches\/security of software<\/li>\n<li>Simple payment plans, and generally cheaper<\/li>\n<li>What happens to your data in between is anyone's guess<\/li>\n<li>Your data\/connections might be logged for legal reasons<\/li>\n<li>Your data might be altered on the fly for advertisement revenues<\/li>\n<\/ul><p><strong>Tor<\/strong><\/p>\n<ul><li>Easy, usable by the general population<\/li>\n<li>Free, no real costs<\/li>\n<li>What entry\/exit points do with your data, is anyone's guess<\/li>\n<li>Slow, due to the number of relays, and non-profit architecture<\/li>\n<\/ul><\/div>","created_at":"2018-11-07T16:53:02+0000","updated_at":"2018-11-07T16:53:02+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"security.stackexchange.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/security\/img\/apple-touch-icon@2.png?v=497726d850f9","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":591,"title":"Laptop Battery Driver\/Firmware Problems \/ Newbie Corner \/ Arch Linux Forums","url":"https:\/\/bbs.archlinux.org\/viewtopic.php?id=170055","content":"<p>The chargers both indicate that they provide the correct voltage\/current, but I do not have a way to verify if they are not outputting what they should be.\u00a0 I'm going to look into updating the BOIS and see if that has any effect.<\/p><p>Also: Requested output from the newer computer (an ASUS Q400A).<\/p><p>uname:<\/p><div class=\"codebox\"><pre>Linux parzyck_asus_arch64 3.11.1-1-ARCH #1 SMP PREEMPT Sat Sep 14 19:30:21 CEST 2013 x86_64 GNU\/Linux<\/pre><\/div><p>lspci:<\/p><div class=\"codebox\"><pre>00:00.0 Host bridge: Intel Corporation 3rd Gen Core processor DRAM Controller (rev 09)\n00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)\n00:14.0 USB controller: Intel Corporation 7 Series\/C210 Series Chipset Family USB xHCI Host Controller (rev 04)\n00:16.0 Communication controller: Intel Corporation 7 Series\/C210 Series Chipset Family MEI Controller #1 (rev 04)\n00:1a.0 USB controller: Intel Corporation 7 Series\/C210 Series Chipset Family USB Enhanced Host Controller #2 (rev 04)\n00:1b.0 Audio device: Intel Corporation 7 Series\/C210 Series Chipset Family High Definition Audio Controller (rev 04)\n00:1c.0 PCI bridge: Intel Corporation 7 Series\/C210 Series Chipset Family PCI Express Root Port 1 (rev c4)\n00:1c.1 PCI bridge: Intel Corporation 7 Series\/C210 Series Chipset Family PCI Express Root Port 2 (rev c4)\n00:1c.3 PCI bridge: Intel Corporation 7 Series\/C210 Series Chipset Family PCI Express Root Port 4 (rev c4)\n00:1d.0 USB controller: Intel Corporation 7 Series\/C210 Series Chipset Family USB Enhanced Host Controller #1 (rev 04)\n00:1f.0 ISA bridge: Intel Corporation HM76 Express Chipset LPC Controller (rev 04)\n00:1f.2 SATA controller: Intel Corporation 7 Series Chipset Family 6-port SATA Controller [AHCI mode] (rev 04)\n00:1f.3 SMBus: Intel Corporation 7 Series\/C210 Series Chipset Family SMBus Controller (rev 04)\n02:00.0 Network controller: Intel Corporation Centrino Wireless-N 2230 (rev c4)\n03:00.0 Ethernet controller: Qualcomm Atheros AR8151 v2.0 Gigabit Ethernet (rev c0)<\/pre><\/div><p>lscpu:<\/p><div class=\"codebox\"><pre>Architecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                8\nOn-line CPU(s) list:   0-7\nThread(s) per core:    2\nCore(s) per socket:    4\nSocket(s):             1\nNUMA node(s):          1\nVendor ID:             GenuineIntel\nCPU family:            6\nModel:                 58\nModel name:            Intel(R) Core(TM) i7-3632QM CPU @ 2.20GHz\nStepping:              9\nCPU MHz:               1958.000\nBogoMIPS:              4391.66\nVirtualization:        VT-x\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              256K\nL3 cache:              6144K\nNUMA node0 CPU(s):     0-7<\/pre><\/div>","created_at":"2018-11-07T16:53:01+0000","updated_at":"2018-11-07T16:53:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"bbs.archlinux.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":592,"title":"Little Planet Procedural","url":"http:\/\/alanluo.com\/procgen\/midterm.html","content":"<h3 id=\"a-handbook-to-little-planet-procedural\">A handbook to \"Little Planet Procedural\"<\/h3><ol><li><a href=\"#introduction\">Introduction<\/a><\/li>\n<li><a href=\"#concepts\">Concepts<\/a>\n<ol><li><a href=\"#noise\">Noise<\/a><\/li>\n<li><a href=\"#more-dimensions\">More Dimensions<\/a><\/li>\n<\/ol><\/li>\n<li><a href=\"#mountains-and-hills\">Mountains and Hills<\/a>\n<ol><li><a href=\"#octaves\">Octaves<\/a><\/li>\n<li><a href=\"#an-observation\">An Observation<\/a><\/li>\n<\/ol><\/li>\n<li><a href=\"#rivers\">Rivers<\/a><\/li>\n<li><a href=\"#sky\">Sky<\/a>\n<ol><li><a href=\"#color-theory\">Color Theory<\/a><\/li>\n<li><a href=\"#color-in-detail\">Color in Detail<\/a><\/li>\n<li><a href=\"#triangulation\">Triangulation<\/a><\/li>\n<\/ol><\/li>\n<li><a href=\"#clouds\">Clouds<\/a><\/li>\n<li><a href=\"#celestial-objects\">Celestial Objects<\/a>\n<ol><li><a href=\"#stars\">Stars<\/a><\/li>\n<li><a href=\"#planet\">Planet<\/a><\/li>\n<li><a href=\"#sun\">Sun<\/a><\/li>\n<\/ol><\/li>\n<li><a href=\"#trees\">Trees<\/a>\n<ol><li><a href=\"#little-trees\">Little Trees<\/a><\/li>\n<li><a href=\"#big-trees\">Big Trees<\/a><\/li>\n<\/ol><\/li>\n<li><a href=\"#conclusion-and-future-work\">Conclusion and Future Work<\/a><\/li>\n<\/ol><p>This is a handbook meant to explain the concepts and execution behind the \"Little Planet Procedural\" project, my directed study midterm project.<\/p><p>As some background, the project is meant as a case study of how code can create diverse and infinite art through basic techniques of procedural generation. It is inspired by the likes of No Man's Sky.<\/p><p>Each time the page is refreshed, a new planet is generated. There are night scenes and day scenes.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image1.png\" alt=\"sample image\" \/><\/div><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image16.png\" alt=\"sample image2\" \/><\/div><p>The project is written entirely with raw HTML5 Canvas.<\/p><h2 id=\"concepts\">Concepts<\/h2><h3 id=\"noise\">Noise<\/h3><p>There is, for the most part, only one concept used throughout the entire project. To see this, let's first consider the following problem. Say we want to generate some hills. How should we go about doing so? We want something that looks kind of like this:<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image2.jpg\" alt=\"some hills\" \/><\/div><p>We want there to be some randomness, but we also want an organic nature to the terrain. In other words, we want local similarity, but global randomness. Generating random points is no good, since it gives us too much local variation.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image3.jpg\" alt=\"random points\" \/><\/div><p>We could mediate this by stretching it all the way out. We call this interpolation.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image4.jpg\" alt=\"lerping\" \/><\/div><p>Technically, with a more complex interpolation that smoothes out the ridges, this method could achieve feasible terrain. In fact, there are an infinite number of ways to generate realistic terrain. However, we will only look at one, called Perlin Noise. The details of Perlin Noise are complex and will not be explored. However, the result is seen in the first image, which is generated using Perlin Noise.<\/p><h3 id=\"more-dimensions\">More Dimensions<\/h3><p>It turns out we can extend this reasoning to higher dimensions. What if we want to generate three-dimensional terrain? We'd want to achieve the same local similarity and global variance on a two-dimensional input. Rather than representing this as a three-dimensional terrain, we can treat map the third dimension onto the continuum from black to white. This gives us cloud-like patterns.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image5.jpg\" alt=\"2D perlin noise\" \/><\/div><p>This might look familiar. It's used in a lot of movies and games to create special effects. In fact, Ken Perlin, the creator of Perlin Noise, won an academy award for his work!<\/p><p>We refer to the \"dimension\" of Perlin noise as the dimension of the input. So, 2D terrain is 1D noise, generated by a dimension of input and a dimension of output. 3D terrain is 2D noise.<\/p><p>The cool thing is that the third dimension does not have to be physical. We could lift a plane of 2D noise through 3D space to give the impression that the terrain is changing, but we are really just using 3D noise with time as a variable.<\/p><p>With this intuition, we can formally define noise.<\/p><ul><li>N-dimensional noise is a function from <strong>R<\/strong><sup>n<\/sup> to the continuous interval [0, 1]<\/li>\n<\/ul><p>To manipulate noise, we treat it like we treat sine funcitons. A noise function <em>f<\/em> of a vector <strong>x<\/strong> can take the form a<em>f<\/em>(b<strong>x<\/strong>)+<strong>c<\/strong>. This will make the amplitude vary from 0 to a, stretch by a factor of b, and offset by c.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image6.jpg\" alt=\"modified noise\" \/><\/div><p>In this case, the image on the right is the same noise on the left, but stretched out.<\/p><p>Now that these basic concepts are down, let's get rolling!<\/p><h2 id=\"mountains-and-hills\">Mountains and Hills<\/h2><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image7.jpg\" alt=\"hills and mountains\" \/><\/div><p>The hills and mountains are both generated by one-dimensional Perlin Noise. There are two things to notice. Firstly, the mountains have much more chaotic noise than the hills. Secondly, both have \"layers\" of color.<\/p><h3 id=\"octaves\">Octaves<\/h3><p>The variation is generated through what we call octaves. Like any function, Perlin Noise follows the concept of constructive interference. Suppose we have a sine wave. See how smooth it is?<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image8.png\" alt=\"a sine wave\" \/><\/div><p>Now suppose we want to introduce some variation to it. One thing we could do is superimpose a smaller sine wave on top - one that has both a smaller wavelength and amplitude.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image9.png\" alt=\"a messy sine wave\" \/><\/div><p>We can do this again and again to get more and more \"noisy\" (haha) functions. (If you add them infinitely you get something somewhat bizarre - but that's a different story.)<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image10.png\" alt=\"a messier sine wave\" \/><\/div><p>We call these \"octaves\" because the first harmonic of a signal, or twice the fundamental frequency, or the wave <code>sin(2x)<\/code> corresponding to the wave <code>sin(x)<\/code>, corresponds to the musical sound one octave up. For example, the note A4 on a piano is 440Hz, while A5 is 880Hz, A3 is 220Hz, and so on.<\/p><p>We can similarly create octaves for Perlin Noise. Implementation-wise, I write one function for drawing and filling in noise:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function make1DNoise(axis, amplitude, scale, params) {\n    var newNoise = [];\n    for(var i=0; i&lt;CANVAS_WIDTH; i++) {\n        newNoise.push({x: i, y:axis+amplitude*\n        params.noiseFunction(scale*i, params.zaxis)});  \n    }\n    newNoise.push(LOWER_LEFT); newNoise.push(LOWER_RIGHT);\n    ctx.fillStyle = params.fillColor;\n    fillPath(newNoise);\n}<\/pre><\/div><p>(<code>fillPath<\/code> is not a default canvas function - it's a function I wrote to make and fill a path from an array of points)<\/p><p>There are a few things to notice here. First of all, I pass in axis, amplitude, and scale as arguments, corresponding very similarly to the constants that manipulate a sine function. Secondly, I pass in a params variable that holds a number of useful things. For instance, I can step through the other dimension of the noise even if I am only using 1D noise, a feature that will become useful soon. More importantly, I reference this thing called <code>noiseFunction<\/code>. What's that about?<\/p><p>It turns out that in scripting language, Javascript included, everything is pretty much the same. Objects are functions and vice versa. So, I can pass in a function as an argument of a function, and save myself some time. I essentially wrote two noise functions: one for mountains, and one for hills.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function mountainNoise(x, z) {\n    return simplex.noise2D(z, x)+0.5*simplex.noise2D(0, 2*x\n    +0.25*simplex.noise2D(0, 4*x)+0.125*simplex.noise2D(0, 8*x);\n}\nfunction hillNoise(x, z) {\n    return simplex.noise2D(z, x);\n}<\/pre><\/div><p>(Simplex Noise is just a variant of Perlin Noise)<\/p><p>I pass these functions in as parameters of my previous function, and thus can use the same function for drawing mountains and hills. Notice that the mountains have four octaves and the hills one, corresponding to their very different noise shapes.<\/p><h3 id=\"an-observation\">An Observation<\/h3><p>You might also notice that there are bands of color on the hills and on the mountains.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image11.jpg\" alt=\"color bands\" \/><\/div><p>To see how these work, let's first make the following observation. Let's say we have a plane intersecting a field of 2D noise. In otherworse, let's take a slice out of the 2D noise.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image12.jpg\" alt=\"color bands\" \/><\/div><p>If we do this, we're essentially fixing one variable of the input. In other words, we only have one remaining variable. So we've essentially reduced the dimension to 1. If we then project this sliced noise, we will see that it is 1D.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image13.jpg\" alt=\"color bands\" \/><\/div><p>Since 2D noise follows our rule of local similarity, by taking a series of adjacent slices out of 2D noise, we can thus construct a series of similar but not identical noise functions. So, the bands of color on the mountains and hills are generated by setting slightly different fixed y-values as inputs for a 2D noise function.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">for(var i=0; i&lt;3; i++) {\n    make1DNoise(430, 200, 0.005, {noiseFunction:mountainNoise,\n    fillColor:toHslString(mountainshade), zaxis:0.07*i}); \n    mountainshade.s-=0.05; mountainshade.l-=0.04;\n}<\/pre><\/div><p>The z-axis parameter of the draw function changes how much the 1D noise is stepped through, or the position of the slicing plane. In this case, we move the plane by 0.07 each step. The last line just slighly varies the color of each new layer drawn - more on this later.<\/p><h2 id=\"rivers\">Rivers<\/h2><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image14.jpg\" alt=\"rivers\" \/><\/div><p>The rivers are pretty much done the same way as the mountains. They are generated starting from the minimum point of the noise. Two sloped lines are drawn from this point. One-octave Perlin Noise extrudes on both sides. This is repeated four times to create four bands of color.<\/p><h2 id=\"sky\">Sky<\/h2><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image15.jpg\" alt=\"rivers\" \/><\/div><p>With the sky, we need to consider color schemes. Recall that we have two types of scenes: night and day.<\/p><h3 id=\"color-theory\">Color Theory<\/h3><p>First, we need to introduce some basic ideas of color theory. We can arrange the primary colors around a circle equally spaced from each other, then add all the in between colors. We call this the color wheel.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image17.jpg\" alt=\"color wheel\" \/><\/div><p>One paradigm for choosing aesthetically pleasing color schemes is to choose colors that are as evenly spaced around the wheel as possible in order to maintain a constant relative coloring. For instnace, green, orange, and blue would be a bad choice because blue is farther from orange than it is from green.<\/p><p>We can obtain a number of color schemes by choosing colors this way. <strong>Analagous<\/strong> schemes choose one color and other similar colors.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image18.jpg\" alt=\"analagous\" \/><\/div><p><strong>Complementary<\/strong> schemes choose one color and other similar colors.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image19.jpg\" alt=\"complementary\" \/><\/div><p><strong>Triad<\/strong> schemes choose three evenly spaced colors.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image20.jpg\" alt=\"triad\" \/><\/div><p>There are many other schemes, but we won't delve into them here. In our case, we use analagous schemes for night scenes, and complementary schemes in day scenes.<\/p><h3 id=\"color-in-detail\">Color in Detail<\/h3><p>Most colors in the program are processed as hsb values rather than as rgb. Most computer colors are in the form of three bytes: one value each for red, green, and blue, coorresponding to the range from 0 to 255.<\/p><p>However, I process colors as a hue, a saturation, and a brightness. This is a much easier way to describe what a color \"looks like\" instead of tinkering with rgb values. Hue describes the angle position on the color wheel, with 0 = 360 degrees referring to red. Saturation describes how much color there is, with 0 being grayscale, and 1 being entirely colored. Brightness describes how much white there is, with 0 being totally black, and 1 being totally white.<\/p><p>I wrote a function that converts an hsb (sometimes known as hsl) color input to a usable string for canvas.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function toHslString(color) {\n    return \"hsl(\"+(color.h%1.0)*360+\", \n    \"+(color.s%1.0)*100+\"%, \"+(color.l%1.0)*100+\"%)\";\n}<\/pre><\/div><p>Notice that taking the hue mod 1.0 makes the wheel into a circle.<\/p><p>The details of the colors are mostly tinkered with through experimentation. There are two different schemes - one for day, one for night. I've left the full code here.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">if(data.time==\"night\") {\n    baseColor = randomColor({brightness:\"dark\"});\n    mountainColor = {h:baseColor.h+Math.random()*0.2-0.1, \n                         s:baseColor.s+Math.random()*0.2-0.1,\n                         l:baseColor.l+Math.random()*0.1+0.15};\n    hillColor = {h:mountainColor.h+Math.random()*0.2-0.1, \n                     s:mountainColor.s+Math.random()*0.2-0.1,\n                     l:mountainColor.l+Math.random()*0.1+0.15}; \n    riverColor = {h:baseColor.h,\n                    s:baseColor.s-Math.random()*0.1-0.05,\n                    l:baseColor.l+Math.random()*0.1-0.05};\n    skyColor2 = {\n        h:baseColor.h,\n        s:baseColor.s,\n        l:baseColor.l-0.2\n    };\n    cloudColor={h:baseColor.h, s:0.4, l:0.2};\n    treeColor={h:baseColor.h+0.5, s:0.4, l:0.2};\n    leafColor={h:treeColor.h+0.5, s:0.8, l:0.6};\n    planetColor={h:hillColor.h, s:0.4, l:0.4};\n} else if(data.time==\"day\") {\n    baseColor = randomColor({brightness:\"medium\"});\n    mountainColor = {h:baseColor.h+0.4+Math.random()*0.1, \n                         s:0.2+Math.random()*0.2,\n                         l:baseColor.l+Math.random()*0.1-0.05};\n    hillColor = {h:mountainColor.h+Math.random()*0.2-0.1, \n                     s:0.4+Math.random()*0.2,\n                     l:baseColor.l+Math.random()*0.1};\n    riverColor = {h:baseColor.h,\n                    s:baseColor.s-Math.random()*0.1-0.05,\n                    l:baseColor.l+Math.random()*0.1+0.2\n                    };\n    skyColor2 = {\n        h:baseColor.h,\n        s:baseColor.s,\n        l:baseColor.l-0.2\n    };\n    cloudColor={h:baseColor.h, s:0.3, l:0.9};\n    treeColor={h:baseColor.h-0.25, s:0.6, l:0.4};\n    leafColor={h:treeColor.h+0.5, s:0.8, l:0.6};\n    planetColor={h:treeColor.h+0.5, s:0.8, l:0.6};\n}<\/pre><\/div><h3 id=\"triangulation\">Triangulation<\/h3><p>You'll notice the sky is partitioned into triangular cells. This is done with an algorithm called Delaunay Triangulation. A Delaunay triangulation, informally, is a triangulation of a set of points that makes every triangle as close to equilateral as possible. It's a way of triangulating points to make them look \"nice and even.\"<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image21.png\" alt=\"triangulation\" \/><\/div><p>(I took this image from Wikipedia.)<\/p><p>For the sky, I generate a gradient. Then, I generate a bunch of random points, and also add some points on the edges. I throw all of this in to a Delaunay triangulation algorithm to get a set of triangles. For each triangle, I fill the entire triangle with the color at the pixel at the center of the triangle.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">var grd=ctx.createLinearGradient(0,CANVAS_HEIGHT,0,0);\ngrd.addColorStop(0,colors.skyColor);\ngrd.addColorStop(1,colors.skyColor2);\nctx.fillStyle=grd;\nctx.fill();\nvar trianglePoints = [];\ntrianglePoints.push(\n    [0, CANVAS_HEIGHT], [0, 0], \n    [CANVAS_WIDTH, CANVAS_HEIGHT], [CANVAS_WIDTH, 0]);\n    \nfor(var i=0; i&lt;15; i++) { \/\/add some stuff on top and bototm\n    trianglePoints.push([Math.random()*CANVAS_WIDTH, 0]);\n    trianglePoints.push([Math.random()*CANVAS_WIDTH, CANVAS_HEIGHT]);\n}\nfor(var i=0; i&lt;10; i++) { \/\/add some stuff on the sides\n    trianglePoints.push([0, Math.random()*CANVAS_HEIGHT]);\n    trianglePoints.push([CANVAS_WIDTH, Math.random()*CANVAS_HEIGHT]);\n}\nfor(var i=0; i&lt;50; i++) { \/\/add some stuff in the middle\n    trianglePoints.push([Math.random()*CANVAS_WIDTH,\n     Math.random()*CANVAS_HEIGHT]);\n}\nfor( \/* do this for each triangle *\/) {\n    var center;\n    center.x = (newtriangle[0].x+\n                newtriangle[1].x+\n                newtriangle[2].x)\/3;\n    center.y = (newtriangle[0].y+\n                newtriangle[1].y+\n                newtriangle[2].y)\/3;\n    var centercolor = ctx.getImageData(center.x, center.y, 1, 1).data;\n    \n    var fillcolor = \"rgb(\"+centercolor[0]+\", \n                    \"+centercolor[1]+\", \"+centercolor[2]+\")\";\n    ctx.fillStyle = fillcolor;\n    \n    fillPath(newtriangle);\n}<\/pre><\/div><h2 id=\"clouds\">Clouds<\/h2><p>The clouds are also generated using Perlin Noise. We essentially use a threshold function to chop off the noise past a certain point. To see this, imagine we intersect the field of noise with a plane, like before, but this time on the outputs.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image22.jpg\" alt=\"clouds1\" \/><\/div><p>If we take this and view it from the top, we see our cloud patterns.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image23.jpg\" alt=\"clouds2\" \/><\/div><p>You'll also notice that the noise in this image seems to be much cleaner than the real clouds. That's because we also introduce an element of variation on the threshold. Also, in order to simulate actual clouds, we stretch out the noise horizontally by 10x. Also, we draw the clouds at 40% opacity.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function makeClouds(threshold, offset, variance) {\n    ctx.globalAlpha = 0.4;\n    ctx.beginPath();\n    for(var i=0; i&lt;CANVAS_WIDTH; i++) {\n        for(var j=0; j&lt;CANVAS_HEIGHT; j++) {\n            var noiseValue = simplex.noise2D(i*0.001+offset, \n                                             j*0.01+offset);\n            if(noiseValue&gt;params.threshold\n                         +Math.random()*params.variance) {\n                drawPixel({x:i, y:j}, colors.cloudColor);\n            }\n        }\n    }\n    ctx.globalAlpha = 1.0;\n}<\/pre><\/div><h2 id=\"celestial-objects\">Celestial Objects<\/h2><p>There are a number of celestial objects in the scenes. At day, a sun is generated. At night, a planet and stars are generated.<\/p><h3 id=\"stars\">Stars<\/h3><p>We generate a bunch of stars at random points in the sky. Most stars are just little cirlces, but each star has a random chance of being a \"big star.\" The big stars are meant to simulate twinkling, and are modelled as an intersection of two thin rectangles, with a gradient fill.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">\/\/5% chance to make a big star\nif(Math.random()&lt;0.05){ \n    ctx.beginPath();\n    \n    var starwidth = Math.random()*7+3;\n    ctx.rect(starx-1, stary-starwidth, 2, 2*starwidth);\n    ctx.rect(starx-starwidth, stary-1, 2*starwidth, 2);\n    var grd=ctx.createRadialGradient(starx, stary, 3, \n                                     starx, stary, starwidth+5);\n    grd.addColorStop(0,\"white\");\n    grd.addColorStop(1,\"rgba(1, 1, 1, 0.0)\");\n    ctx.fillStyle=grd;\n    ctx.fill();\n}<\/pre><\/div><h3 id=\"planet\">Planet<\/h3><p>The planet is pretty much hard-coded in. In future work I may add some variance. For now, I just generate a circle and a bunch of ellipses at an angle to simulate light. I use a method very similar to that I used for the clouds to introduce some texture, as well.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image24.png\" alt=\"planet\" \/><\/div><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function makePlanet(position, radius, params) {\n    ctx.beginPath();\n    ctx.arc(position.x, position.y, radius, 0, 2*Math.PI);\n    ctx.fillStyle = colors.planetColor;\n    ctx.fill();\n    ctx.save();\n    ctx.clip();\n    \/\/in a square around the planet\n    var xposmax = position.x+radius;\n    var yposmax = position.y+radius;\n    ctx.globalCompositeOperation = 'overlay';\n    for(var xpos=position.x-radius; xpos&lt;xposmax; xpos++) {\n        for(var ypos=position.y-radius; ypos&lt;yposmax; ypos++) {\n            if(simplex.noise2D(xpos, ypos)&gt;0.1+Math.random()*0.2) {\n                drawPixel({x:xpos, y:ypos}, 'rgba(0, 0, 0, 0.05)');\n            }\n            if(simplex.noise2D(xpos*0.03, ypos*0.03)&gt;\n                                    0.1+Math.random()*0.2) {\n                drawPixel({x:xpos, y:ypos}, 'rgba(0, 0, 0, 0.05)');\n            }\n        }\n    }\n    ctx.fillStyle=\"rgba(255, 255, 255, 0.1)\";\n    ctx.beginPath();\n    ctx.ellipse(position.x+60, position.y-60, 60, \n            40, 45 * Math.PI\/180, 0, 2 * Math.PI);\n    ctx.fill(); ctx.beginPath();\n    ctx.ellipse(position.x+40, position.y-40, 80, \n            60, 45 * Math.PI\/180, 0, 2 * Math.PI);\n    ctx.fill(); ctx.beginPath();\n    ctx.ellipse(position.x+20, position.y-20, 100, \n            80, 45 * Math.PI\/180, 0, 2 * Math.PI);\n    ctx.fill();\n    ctx.globalCompositeOperation = 'source-over';\n    ctx.restore();\n}\n<\/pre><\/div><h3 id=\"sun\">Sun<\/h3><p>The sun is just a circle with another circle around it.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function makeSun(position, params) {\n    ctx.beginPath();\n    ctx.arc(position.x, position.y, params.innerradius, 0, 2*Math.PI);\n    ctx.globalAlpha = 0.95;\n    ctx.fillStyle=toHslString({h:baseColor.h, s:0.3, l:0.8});\n    ctx.fill();\n    ctx.arc(position.x, position.y, params.outerradius, 0, 2*Math.PI);\n    ctx.globalAlpha = 0.5;\n    ctx.fillStyle=toHslString({h:baseColor.h, s:0.3, l:0.9});\n    ctx.fill();\n    ctx.globalAlpha = 1.0;\n}<\/pre><\/div><h2 id=\"trees\">Trees<\/h2><p>The trees have two components. The program has two canvas components. The first shows the scene. Whenever a tree is clicked on, the second shows a fractal tree.<\/p><h3 id=\"little-trees\">Little Trees<\/h3><p>We should first consider the goal in making trees. We model the trunk as a rectangle, and the leaves as a blob. So, what exactly is a blob? How can we create one? Essentially, we'd like the blob to look random if you see it from a distance, but still maintain its local circular structure. Sound familiar?<\/p><p>Turns out, all we do is take a loop of Perlin noise. In the past, we've been inputting lines to the noise function. However, nothing stops us from putting in a circle.<\/p><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image26.png\" alt=\"fractal tree\" \/><\/div><p>Essentially, we make a circle out of a bunch of vertices. Then, for every vertex on the circle, we input its coordinate into a noise function and extrude it by that much.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">var pointcount = 30, radius = 10, blobpoints;\nfor(var j=0; j&lt;pointcount; j++) {\/\/map points onto a circle\n    var prepos = {\n        x:leafcenter.x+radius*Math.cos(j*(2*Math.PI)\/pointcount),\n        y:leafcenter.y+radius*Math.sin(j*(2*Math.PI)\/pointcount)};\n    newradius = radius + 5*simplex.noise2D(prepos.x, prepos.y);\n    \n    var newpos = {\n        x:leafcenter.x+newradius*Math.cos(j*(2*Math.PI)\/pointcount),\n        y:leafcenter.y+newradius*Math.sin(j*(2*Math.PI)\/pointcount) };\n    blobpoints.push(newpos);\n}\nfillPath(blobpoints);<\/pre><\/div><h3 id=\"big-trees\">Big Trees<\/h3><div class=\"figure\"><img src=\"http:\/\/alanluo.com\/procgen\/img\/image25.png\" alt=\"fractal tree\" \/><\/div><p>Clicking on the trees is supposed to simulate \"zooming in\" on the tree. So, clicking on the same tree twice gives the same \"zoomed\" tree. This is done through an L-system. We're not going to go into L-systems in depth. However, we can describe one as a type of recursive drawing function. In this case, we write a function where a line segment is drawn (the \"trunk\" of the tree), followed by two additional segments that sprout off towards either side. However, each call of this function sets off two new calls of the function, resulting in a constant branching.<\/p><p>In order to achieve variation while making the same tree translate to the same L-system, we introduce an angle variation that is seeded by the position of the tree.<\/p><div class=\"sourceCode\"><pre class=\"sourceCode javascript\">function makeBigTree(newseed) {\n    resetCanvas();\n    setRandomSeet(newseed);\n    branch(50);\n}\nfunction branch(len){\n  var theta = random()*(Math.PI\/3);\n  drawLine({x:0, y:0}, {x:0, y:len}, ctx2);\n  ctx2.translate(0, len);\n  len *= 0.66;\n  if (len &gt; 2) {\n    ctx2.save();\n    ctx2.rotate(theta);\n    branch(len);\n    ctx2.restore();\n    ctx2.save();\n    ctx2.rotate(-theta);\n    branch(len);\n    ctx2.restore();\n  }\n}\ncanvas.addEventListener('mousedown', doMouseDown, false);\nfunction doMouseDown(evt) {\n    var mousePos = getMousePos(canvas, evt);\n    for(var i=0; i&lt;clickboxes.length; i++) {\n        if(mousePos.x&gt;clickboxes[i].left \n        &amp;&amp; mousePos.x&lt;clickboxes[i].right \n       &amp;&amp; mousePos.y&gt;clickboxes[i].top  \n       &amp;&amp; mousePos.y&lt;clickboxes[i].bottom) {\n            clickboxes[i].action(i);\n        }\n    }\n}<\/pre><\/div><p>We use the same property of Javascript functions as objects as before here. <code>clickboxes<\/code> is an array that stores rectangular bounding boxes of each tree. Each bounding box has a corresponding action, which is a function property. The action is created by the L-system drawing function seeded by the clickbox index.<\/p><h2 id=\"conclusion-and-future-work\">Conclusion and Future Work<\/h2><p>Overall, this has been a fun project. It's showcased the ability for a relatively small set of techniques to generate some really cool looking stuff. There are still many, many techniques that I learned about that did not make it into the final project, including fractals and cellular automata.<\/p><p>If I do additional work on this, I'd like to:<\/p><ul><li>Implement scene types: e.g. presets for deserts, grasslands...<\/li>\n<li>Use a wider range of techniques<\/li>\n<li>Make the world bigger, adding star systems and such<\/li>\n<\/ul>","created_at":"2018-11-07T16:53:01+0000","updated_at":"2018-11-07T16:53:01+0000","annotations":[],"mimetype":"text\/html","reading_time":17,"domain_name":"alanluo.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":590,"title":"The Homely Mutt \/ Steve Losh","url":"http:\/\/stevelosh.com\/blog\/2012\/10\/the-homely-mutt\/","content":"<p>Now that <a href=\"http:\/\/sparrowmailapp.com\/\">Sparrow<\/a> is <a href=\"http:\/\/www.theverge.com\/2012\/7\/20\/3172222\/google-buys-sparrow-mail\">effectively dead<\/a> many of its users will\nbe looking for a new email client.  If you\u2019re not afraid of the terminal you may\nwant to give <a href=\"http:\/\/www.mutt.org\/\">Mutt<\/a> a try.<\/p><p>Mutt certainly isn\u2019t the prettiest email client around, and its\nsetup\/configuration process is one of the ugliest out there.  But once you get\nit set up it\u2019s got a lot of advantages over many other email clients.<\/p><p>In this post I\u2019ll show you how to set up Mutt on OS X like I do.<\/p><p>This setup is going to be specific to the way I work with email.  Notably:<\/p><ul><li>I have a Google Apps account that provides my steve@stevelosh.com email\naddress.<\/li>\n<li>I have many other email addresses, but they all simply forward to my main one.<\/li>\n<li>All mail I send comes from steve@stevelosh.com.<\/li>\n<li>I store my contacts in the OS X address book.<\/li>\n<li>All email comes into my inbox (or to a folder for a specific mailing list).<\/li>\n<li>Once I\u2019m done with an email, I remove it from my inbox and it lives in the\n\u201cAll Mail\u201d archive.  I don\u2019t sort email into folders after it arrives.<\/li>\n<li>I sometimes read email offline and mark it for deletion, then sync that\ndeletion back to the server once I get online again.<\/li>\n<li>Sometimes I write email without an internet connection and send it once I get\nconnected again.<\/li>\n<\/ul><p>My email setup is tailored around those requirements, so that\u2019s what it does\nbest.  Mutt is very configurable though, so if you work differently you can\nprobably bend it to make it do what you want.<\/p><p>In particular, extending this setup to work with multiple email accounts\nwouldn\u2019t be too much trouble.  I used to work with two separate accounts until\nI said \u201cscrew it, I\u2019ll just use the one\u201d.<\/p><h2 id=\"other-guides-and-resources\">Other Guides and Resources<\/h2><p>I\u2019ve used a lot of other guides to figure out how to get this giant Rube\nGoldberg machine of an email client working.  Here are a few of them:<\/p><ul><li><a href=\"http:\/\/thomas.pelletier.im\/2010\/10\/low-memory-mail-client\/\">http:\/\/thomas.pelletier.im\/2010\/10\/low-memory-mail-client\/<\/a><\/li>\n<li><a href=\"http:\/\/www.andrews-corner.org\/mutt.html\">http:\/\/www.andrews-corner.org\/mutt.html<\/a><\/li>\n<li><a href=\"http:\/\/jstorimer.com\/shells\/2010\/01\/19\/using-mutt-with-gmail-on-osx.html\">http:\/\/jstorimer.com\/shells\/2010\/01\/19\/using-mutt-with-gmail-on-osx.html<\/a><\/li>\n<li><a href=\"http:\/\/www.vijaykiran.com\/2010\/01\/27\/mutt-for-gmail-imap-on-mac-os-x\/\">http:\/\/www.vijaykiran.com\/2010\/01\/27\/mutt-for-gmail-imap-on-mac-os-x\/<\/a><\/li>\n<li><a href=\"http:\/\/hynek.me\/articles\/my-mutt-gmail-setup\/\">http:\/\/hynek.me\/articles\/my-mutt-gmail-setup\/<\/a><\/li>\n<li><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Mutt\">https:\/\/wiki.archlinux.org\/index.php\/Mutt<\/a><\/li>\n<li><a href=\"http:\/\/linsec.ca\/Using_mutt_on_OS_X\">http:\/\/linsec.ca\/Using_mutt_on_OS_X<\/a><\/li>\n<li><a href=\"http:\/\/www.mutt.org\/doc\/manual\/manual.html\">http:\/\/www.mutt.org\/doc\/manual\/manual.html<\/a><\/li>\n<li><a href=\"http:\/\/pbrisbin.com\/posts\/two_accounts_in_mutt\">http:\/\/pbrisbin.com\/posts\/two_accounts_in_mutt<\/a><\/li>\n<li><a href=\"http:\/\/upsilon.cc\/~zack\/blog\/posts\/2011\/01\/how_to_use_Notmuch_with_Mutt\/\">http:\/\/upsilon.cc\/~zack\/blog\/posts\/2011\/01\/how_to_use_Notmuch_with_Mutt\/<\/a><\/li>\n<\/ul><h2 id=\"overview\">Overview<\/h2><p>I\u2019m going to give it to you straight: getting this whole contraption set up is\ngoing to take at least an hour from start to finish, not counting the time it\u2019ll\ntake to download all of your email and install stuff.  Set aside an evening if\nyou\u2019re serious about this.<\/p><p>It\u2019s an investment, and you might not want to make it.  If not, go use\nThunderbird, er, Sparrow, er, I don\u2019t know, the Gmail web interface or\nsomething.<\/p><p>Mutt on its own doesn\u2019t do very much, so we\u2019re going to combine it with a few\nother things to get the job done.  Here\u2019s a bird\u2019s eye view of what it\u2019ll look\nlike when we\u2019re done:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/what-the-mutt.png\" alt=\"Diagram\" \/><\/p><p>If this diagram doesn\u2019t make you run screaming you might just be masochistic\nenough to make it through the initial setup of Mutt.  If you do, you\u2019ll be\nrewarded with email bliss that won\u2019t go away when Google or Facebook decide to\ntoss some money around.<\/p><p>First thing\u2019s first: we\u2019re going to pull down our email from Gmail to our local\nmachine.  All of it.  It\u2019ll take a while the first time you sync, but has a few\nbenefits.<\/p><h3 id=\"why-local-email\">Why Local Email?<\/h3><p>Having a local copy of all of your email means you\u2019ve always got access to it,\nno matter where you are.  Looking for that one person\u2019s address they emailed you\nsix years ago when you\u2019re trying to find their house and you don\u2019t have an\ninternet connection?  No problem, it\u2019s on your hard drive.<\/p><p>This also acts as a backup in case Google ever decides to kill your Gmail\naccount.  It\u2019ll be stored in a common format that a lot of programs can read, so\nyou\u2019ve got a safety net.  And the email is stored as normal files, so if you use\nsomething like Time Machine or <a href=\"http:\/\/www.backblaze.com\/partner\/af3574\">Backblaze<\/a> that\u2019s yet another free backup.<\/p><p>In this setup <strong>all of your email is stored as plain text<\/strong>.  If you want it\nencrypted just use OS X\u2019s full-disk encryption and you\u2019re set.<\/p><p>I use <a href=\"http:\/\/offlineimap.org\/\">offlineimap<\/a> to pull email down from Gmail and get it on my hard drive.\nOfflineimap will also sync any changes you make to this local copy back up to\nGmail.<\/p><h3 id=\"the-alternative\">The Alternative<\/h3><p>You may not care as much about reading your email offline as I do.  If you can\ntolerate always needing an internet connection to read your mail, you can skip\nthis painful section and follow <a href=\"http:\/\/empt1e.blogspot.com\/2009\/10\/using-mutt-with-gmail-imap-complete.html\">this guide<\/a> instead.<\/p><p>You\u2019ll probably still find the other sections of this post interesting though.<\/p><h3 id=\"installing-offlineimap\">Installing offlineimap<\/h3><p>I\u2019ve gone through a number of laptops in the past few years, and each time\nI spend a painful half hour or so screwing around with the latest version of\nofflineimap\u2019s backwards-incompatible changes.<\/p><p>If you\u2019re determined to run the latest version of offlineimap, you can install\nit with pip or something.  If you just want to download your fucking email and\nget on with your life, you can follow the instructions I\u2019ve laid out for you\nhere:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">git clone git:\/\/github.com\/spaetz\/offlineimap.git\ncd offlineimap\ngit checkout 679c491c56c981961e18aa43b31955900491d7a3\npython setup.py install<\/pre><\/div><p>That\u2019s the version I\u2019m using.  It works.  You can use a newer one if you want,\nbut expect to spend some time figuring out how to fix the configuration in this\npost to work with whatever breaking changes have been made since then.  The last\ntime I tried this I got to rewrite all my nametrans stuff.  That was fun.<\/p><h3 id=\"configuring-offlineimap\">Configuring offlineimap<\/h3><p>Once you\u2019ve got offlineimap installed, you\u2019ll need to create\na <code>~\/.offlineimaprc<\/code> file.  You can keep it in your dotfiles repo and symlink it\ninto place if you want.  Here\u2019s a sample to get you started:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">[general]\nui = TTY.TTYUI\naccounts = SteveLosh\npythonfile=~\/.mutt\/offlineimap.py\nfsync = False\n[Account SteveLosh]\nlocalrepository = SteveLosh-Local\nremoterepository = SteveLosh-Remote\nstatus_backend = sqlite\npostsynchook = notmuch new\n[Repository SteveLosh-Local]\ntype = Maildir\nlocalfolders = ~\/.mail\/steve-stevelosh.com\nnametrans = lambda folder: {'drafts':  '[Gmail]\/Drafts',\n                            'sent':    '[Gmail]\/Sent Mail',\n                            'flagged': '[Gmail]\/Starred',\n                            'trash':   '[Gmail]\/Trash',\n                            'archive': '[Gmail]\/All Mail',\n                            }.get(folder, folder)\n[Repository SteveLosh-Remote]\nmaxconnections = 1\ntype = Gmail\nremoteuser = steve@stevelosh.com\nremotepasseval = get_keychain_pass(account=\"steve@stevelosh.com\", server=\"imap.gmail.com\")\nrealdelete = no\nnametrans = lambda folder: {'[Gmail]\/Drafts':    'drafts',\n                            '[Gmail]\/Sent Mail': 'sent',\n                            '[Gmail]\/Starred':   'flagged',\n                            '[Gmail]\/Trash':     'trash',\n                            '[Gmail]\/All Mail':  'archive',\n                            }.get(folder, folder)\nfolderfilter = lambda folder: folder not in ['[Gmail]\/Trash',\n                                             'Nagios',\n                                             'Django',\n                                             'Flask',\n                                             '[Gmail]\/Important',\n                                             '[Gmail]\/Spam',\n                                             ]<\/pre><\/div><p>It\u2019s kind of long, so let\u2019s go through it line by line and see what\u2019s going on.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">[general]\nui = TTY.TTYUI\naccounts = SteveLosh\npythonfile=~\/.mutt\/offlineimap.py\nfsync = False<\/pre><\/div><p>First we tell offlineimap to use the <code>TTY.TTYUI<\/code> ui.  Yes, this program that\nsyncs your email has multiple user interfaces.  I guess if you can\u2019t decide what\ncolor the bikeshed should be you can just build a whole bunch of bikesheds\ninstead.<\/p><p>Then we specify the accounts.  There\u2019s only one because as I said before:\nI only use a single email account that all my addresses forward to.  If you\nwanted to have many, you\u2019d change this line.<\/p><p>The <code>pythonfile<\/code> is just a file that offlineimap will parse (as Python) before\nloading the rest of the config, so you can define custom helper functions more\neasily.  We\u2019ll see more of this later.<\/p><p>We\u2019re also telling offlineimap that it doesn\u2019t need to fsync after every single\noperation.  This will speed things up, and since it\u2019s just a local copy it\u2019s\ntypically not a big deal if we lose an email here and there from a crash (it\u2019ll\njust be synced the next time anyway).<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">[Account SteveLosh]\nlocalrepository = SteveLosh-Local\nremoterepository = SteveLosh-Remote\nstatus_backend = sqlite<\/pre><\/div><p>This next section hooks up a few things.  First, it tells offlineimap which\nlocal and remote repositories to use for the account.  Manual configuration\ninstead of sane defaults is a recurring theme we\u2019ll see throughout this process.<\/p><p>Hey, I titled the entry \u201cThe <em>Homely<\/em> Mutt\u201d for a reason.<\/p><p>We\u2019re also going to use a SQLite-based cache for this account.  If you don\u2019t\nalready have SQLite you\u2019ll want to get it with <code>brew install sqlite<\/code>.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">[Repository SteveLosh-Local]\ntype = Maildir\nlocalfolders = ~\/.mail\/steve-stevelosh.com\nnametrans = lambda folder: {'drafts':  '[Gmail]\/Drafts',\n                            'sent':    '[Gmail]\/Sent Mail',\n                            'flagged': '[Gmail]\/Starred',\n                            'trash':   '[Gmail]\/Trash',\n                            'archive': '[Gmail]\/All Mail',\n                            }.get(folder, folder)<\/pre><\/div><p>Now we\u2019re getting to the meat of the configuration.  This \u201clocal repository\u201d is\ngoing to be the mail as it sits on our hard drive.  We\u2019re going to use the\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Maildir\">Maildir format<\/a> because it plays nicely with Mutt (and tons of other\nstuff).<\/p><p>Then we specify the path where we\u2019re going to keep the mail.  This is going to\ntake a lot of space if you\u2019ve got a lot of mail.  Attachments are downloaded\ntoo.  When I said you\u2019re getting an offline copy of all your email I meant <em>all<\/em>\nof it.<\/p><p>I think offlineimap needs the <code>~\/.mail<\/code> directory created for it.  It\u2019s been\na while since I did this, so I might be wrong, but if it complains about not\nbeing able to access the mail folders just go ahead and <code>mkdir ~\/.mail<\/code>.<\/p><p>Next we have the craziest part of the configuration: name translation.<\/p><p>Here\u2019s the issue: offlineimap needs to know how to translate the names of\nfolders on the IMAP server to folder names on your hard drive.<\/p><p>Also, Gmail doesn\u2019t actually use <em>folders<\/em> but its own concept called \u201clabels\u201d.\nBut since the IMAP protocol doesn\u2019t know about labels, it fakes them by making\nthem appear to be folders.<\/p><p>User-created labels in Gmail (like \u201cMercurial\u201d or \u201cClients\u201d) will appear as\nfolders with those names through IMAP.<\/p><p>Built-in, special Gmail folders have names that start with <code>[Gmail]\/<\/code>.  We need\nto turn those into something sane for our hard drive, so that\u2019s what this\nnametrans setting is for.  It\u2019s a Python function that takes the remote folder\nname and returns the name that should be used on your local hard drive.<\/p><p>Yes, you read that right.  This is Python code embedded in the right hand side\nof an INI file\u2019s setting assignment.  I am not fucking with you, this is\nseriously how you do it.  Go ahead and crack open that beer now.<\/p><p>So the \u201cSent Mail\u201d folder in your Gmail account will be synced to\n<code>~\/.mail\/steve-stevelosh.com\/sent<\/code>.  Cool.<\/p><p>(No, I don\u2019t know what would happen if you created a label called <code>[Gmail]\/All\nMail<\/code> in Gmail.  If you try please let me know, but I take no responsibility if\nit ends with all your email being deleted.)<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">[Repository SteveLosh-Remote]\nmaxconnections = 1\ntype = Gmail\nremoteuser = steve@stevelosh.com\nremotepasseval = get_keychain_pass(account=\"steve@stevelosh.com\", server=\"imap.gmail.com\")\nrealdelete = no\nnametrans = lambda folder: {'[Gmail]\/Drafts':    'drafts',\n                            '[Gmail]\/Sent Mail': 'sent',\n                            '[Gmail]\/Starred':   'flagged',\n                            '[Gmail]\/Trash':     'trash',\n                            '[Gmail]\/All Mail':  'archive',\n                            }.get(folder, folder)\nfolderfilter = lambda folder: folder not in ['[Gmail]\/Trash',\n                                             'Nagios',\n                                             'Django',\n                                             'Flask',\n                                             '[Gmail]\/Important',\n                                             '[Gmail]\/Spam',\n                                             ]<\/pre><\/div><p>Finally, the home stretch.  The last section described the folder on our local\nhard drive, and this one describes our Gmail account.<\/p><p>First, we tell offlineimap to only ever use a single connection at a time.  You\ncan try increasing this number for better performance, but in my experience\nGoogle is not afraid to enforce its rate limits and would cut me off fairly\noften when I tried that.  Just leave it at one if you want to be safe.<\/p><p>Next is the type.  Luckily offlineimap provides a <code>Gmail<\/code> type that handles\na lot of the craziness that is Gmail\u2019s IMAP setup.  Nice.<\/p><p>Then we have the username.  Nothing special here, except that if you have\na non-apps account (i.e.: an actual vanilla Gmail account) you may or may not\nneed to include the <code>@gmail.com<\/code> in the username.  I don\u2019t know.  If one doesn\u2019t\nwork, just try the other.<\/p><p>Next we have <code>remotepasseval<\/code>.  This is a bit of Python code (drink!) that\nshould return the password for the account.<\/p><p>What is this <code>get_keychain_pass<\/code> function?  Well, remember when we saw the\n<code>pythonfile<\/code> setting back in the general section?  It\u2019s a function defined in\nthere.  I\u2019ll talk about that in the next section, for now just accept that it\nworks.<\/p><p>Next we set <code>realdelete<\/code> to no.  If this is set to yes, then deleting an email\nin your inbox would actually delete it entirely.  When you set it to no, then\ndeleting an email from your inbox (or any label\u2019s folder) will leave it in\nGmail\u2019s All Mail.<\/p><p>If you want to really delete an email, you\u2019ll need to delete it from All Mail\n(which is named archive on our local filesystem, remember?).  I feel like this\nis a good compromise.  I rarely care about actually deleting mail, given that\nI have many unused gigabytes available on Gmail.<\/p><p>Next we have another nametrans setting.  This is a Python function (drink!) just\nlike the one for the local repository, except it goes in the other direction.\nIt takes the name of a local folder and returns the name of the folder on the\nIMAP server.  Knowing this, it should be easy to understand this setting.<\/p><p>Finally, we have <code>folderfilter<\/code>.  This is a Python function (drink!) that takes\na <strong>remote<\/strong> folder name and returns <code>True<\/code> if that folder should be synced, or\n<code>False<\/code> if it should <strong>not<\/strong> be synced.  I\u2019ve chosen to skip syncing my Spam and\nTrash folders, as well as a few mailing list labels I don\u2019t check all that\noften.  Customize this to your own taste.<\/p><h3 id=\"retrieving-passwords\">Retrieving Passwords<\/h3><p>We\u2019re almost ready, but there\u2019s one more thing we need to do, and that\u2019s\nimplement a secure way for offlineimap to get access to our Gmail password.<\/p><p>If you don\u2019t care too much about security, you <em>can<\/em> configure offlineimap with\na plaintext password right in the config file.  But don\u2019t do that.  It\u2019ll only\ntake a minute to do this securely.<\/p><p>First, you need to add your Gmail password into your OS X keychain.  Open the\nKeychain Access app and press the <code>+<\/code> button:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/keychain-1.png\" alt=\"Keychain 1\" \/><\/p><p>Then fill out the form.  The \u201cKeychain Item Name\u201d should be\n<code>http:\/\/imap.gmail.com<\/code>.  The \u201cAccount Name\u201d should be your email address.  The\npassword should be your password:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/keychain-2.png\" alt=\"Keychain 2\" \/><\/p><p>Press \u201cAdd\u201d.  Now repeat the process for the SMTP server.  The \u201cKeychain Item\nName\u201d should be <code>smtp:\/\/smtp.gmail.com<\/code>.  The \u201cAccount Name\u201d should be your\nemail address.  The password should be your password:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/keychain-3.png\" alt=\"Keychain 3\" \/><\/p><p>Now we need to create the <code>offlineimap.py<\/code> file we pointed offlineimap to\nearlier.  It needs to contain the <code>get_keychain_pass<\/code> function, which takes an\n<code>account<\/code> and <code>server<\/code> and return the password.  Here\u2019s the file I\u2019m using:<\/p><div class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">#!\/usr\/bin\/python\nimport re, subprocess\ndef get_keychain_pass(account=None, server=None):\n    params = {\n        'security': '\/usr\/bin\/security',\n        'command': 'find-internet-password',\n        'account': account,\n        'server': server,\n        'keychain': '\/Users\/sjl\/Library\/Keychains\/login.keychain',\n    }\n    command = \"sudo -u sjl %(security)s -v %(command)s -g -a %(account)s -s %(server)s %(keychain)s\" % params\n    output = subprocess.check_output(command, shell=True, stderr=subprocess.STDOUT)\n    outtext = [l for l in output.splitlines()\n               if l.startswith('password: ')][0]\n    return re.match(r'password: \"(.*)\"', outtext).group(1)<\/pre><\/div><p>In a nutshell, it uses <code>\/usr\/bin\/security<\/code> to retrieve the password.  Read\nthrough the code if you\u2019re curious.<\/p><p>This is not <em>completely<\/em> secure, but it\u2019s better than having your password in\na plaintext file in your home directory.<\/p><p>Whew!  Time to actually run this thing and pull down our email!<\/p><h3 id=\"running-offlineimap\">Running offlineimap<\/h3><p>Assuming everything is in place, open a terminal and run offlineimap:<\/p><pre>offlineimap\n<\/pre><p>Go read a book, because this is going to pull down all the email (with\nattachments) in any folders you didn\u2019t exclude in the config file.<\/p><p><strong>If there\u2019s an error, stop and figure out what went wrong<\/strong>.  Remember,\nofflineimap is a <em>two-way<\/em> sync, so there\u2019s always the possibility it\u2019ll eat\nyour email if you seriously mess something up!  I wish it had a\n<code>--dont-touch-remote<\/code> option you could use as a safety net for the original\nsync, but it doesn\u2019t, so be careful!<\/p><p>In the future you can use <code>offlineimap -q<\/code> to run it in \u201cquick mode\u201d.  It\u2019ll\nperform fewer checks but will generally be much faster.<\/p><p>If you want to set up offlineimap to run every 5 minutes or so, you can use\n<code>launchd<\/code>.  <code>cron<\/code> does not work for some reason.  I\u2019m not entirely sure why.<\/p><p>Personally I actually <em>like<\/em> having to press a key to fetch new mail.  It\u2019s less\nof a distraction than having new mail rolling in all the time.  I can get new\nemail when I\u2019m ready to actually look at it, rather than having it nagging me\nall the time.<\/p><p>The great part about offlineimap is that once you\u2019ve got it configured and it\u2019s\nsuccessfully run once, it\u2019s pretty much rock solid from then on.  You can run it\noften, <code>Ctrl-c<\/code> it, put the laptop to sleep in the middle of a run, or <code>kill -9<\/code>\nit, and it still won\u2019t lose emails.  On the next sync it\u2019ll fix anything that\u2019s\nmissing.<\/p><h2 id=\"mutt\">Mutt!<\/h2><p>Now that you\u2019ve got your email on your computer, it\u2019s finally time to start\nusing Mutt itself!<\/p><h3 id=\"installing\">Installing<\/h3><p>Mutt can be installed in a bunch of different ways, but the easiest is through\nHomebrew:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">brew install mutt --sidebar-patch<\/pre><\/div><p>The sidebar patch is a third-party patch that adds a sidebar to Mutt.  I don\u2019t\nknow why it\u2019s not in core Mutt because it\u2019s insanely useful.  Oh well, at least\nHomebrew makes it simple to get.<\/p><p>That\u2019s pretty much it for installation, but don\u2019t get too relaxed because you\u2019re\nfar from done.<\/p><h3 id=\"configuring\">Configuring<\/h3><p>Mutt is <em>very<\/em> configurable.  This is great once you\u2019ve become a power user and\nwant to mold it to your will, but terrible when you\u2019re just getting started.<\/p><p>Mutt settings are kept in a <code>~\/.muttrc<\/code> file.  If this file doesn\u2019t exist Mutt\nwill look for <code>~\/.mutt\/muttrc<\/code> (note the lack of a dot in the filename), so you\ncan put it there if you prefer.<\/p><p>Let\u2019s start by creating a basic <code>~\/.muttrc<\/code> piece by piece (a lot of this was\ntaken from <a href=\"http:\/\/pbrisbin.com\/posts\/two_accounts_in_mutt\">this article<\/a>).<\/p><p>Once you\u2019ve got a bit of Mutt under your belt you\u2019ll want to read <a href=\"http:\/\/www.mutt.org\/doc\/manual\/manual-6.html\">the\ndocumentation<\/a> for these settings, but for now just use them to keep\nthings sane.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Paths ----------------------------------------------\nset folder           = ~\/.mail               # mailbox location\nset alias_file       = ~\/.mutt\/alias         # where to store aliases\nset header_cache     = ~\/.mutt\/cache\/headers # where to store headers\nset message_cachedir = ~\/.mutt\/cache\/bodies  # where to store bodies\nset certificate_file = ~\/.mutt\/certificates  # where to store certs\nset mailcap_path     = ~\/.mutt\/mailcap       # entries for filetypes\nset tmpdir           = ~\/.mutt\/temp          # where to keep temp files\nset signature        = ~\/.mutt\/sig           # my signature file<\/pre><\/div><p>Here we tell Mutt where to find the various folders it needs.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Basic Options --------------------------------------\nset wait_key = no        # shut up, mutt\nset mbox_type = Maildir  # mailbox type\nset timeout = 3          # idle time before scanning\nset mail_check = 0       # minimum time between scans\nunset move               # gmail does that\nset delete               # don't ask, just do\nunset confirmappend      # don't ask, just do!\nset quit                 # don't ask, just do!!\nunset mark_old           # read\/new is good enough for me\nset beep_new             # bell on new mails\nset pipe_decode          # strip headers and eval mimes when piping\nset thorough_search      # strip headers and eval mimes before searching<\/pre><\/div><p>These are some basic options to make Mutt behave a bit more sanely.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Sidebar Patch --------------------------------------\nset sidebar_delim   = '  \u2502'\nset sidebar_visible = yes\nset sidebar_width   = 24\ncolor sidebar_new color221 color233<\/pre><\/div><p>These options are specific to the sidebar patch.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Status Bar -----------------------------------------\nset status_chars  = \" *%A\"\nset status_format = \"\u2500\u2500\u2500[ Folder: %f ]\u2500\u2500\u2500[%r%m messages%?n? (%n new)?%?d? (%d to delete)?%?t? (%t tagged)? ]\u2500\u2500\u2500%&gt;\u2500%?p?( %p postponed )?\u2500\u2500\u2500\"<\/pre><\/div><p>This gives us a pretty status bar with the information we care about (and none\nof the stuff we don\u2019t).<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Header Options -------------------------------------\nignore *                                # ignore all headers\nunignore from: to: cc: date: subject:   # show only these\nunhdr_order *                           # some distros order things by default\nhdr_order from: to: cc: date: subject:  # and in this order<\/pre><\/div><p>These options hide some of the extra email headers we don\u2019t care about when\nviewing and composing email.<\/p><p>Now it\u2019s time to fill on our account details:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Account Settings -----------------------------------\n# Default inbox.\nset spoolfile = \"+steve-stevelosh.com\/INBOX\"\n# Alternate email addresses.\nalternates sjl@pculture.org still\\.?life@gmail.com steve@ladyluckblues.com steve@pculture.org\n# Mailboxes to show in the sidebar.\nmailboxes +steve-stevelosh.com\/INBOX \\\n          +steve-stevelosh.com\/vim \\\n          +steve-stevelosh.com\/clojure \\\n          +steve-stevelosh.com\/python \\\n          +steve-stevelosh.com\/mercurial \\\n          +steve-stevelosh.com\/archive \\\n          +steve-stevelosh.com\/sent \\\n          +steve-stevelosh.com\/drafts \\\n# Other special folders.\nset mbox      = \"+steve-stevelosh.com\/archive\"\nset postponed = \"+steve-stevelosh.com\/drafts\"<\/pre><\/div><p>Most of those should be self-explanatory.  Fill in the appropriate values for\nyour mail folder(s).<\/p><p>We\u2019ll add more as we go through the next few sections, but that\u2019s enough to get\nus started.<\/p><h3 id=\"running\">Running<\/h3><p>Now that you\u2019ve got Mutt configured you can run it:<\/p><p>I like to always be in my <code>~\/Desktop<\/code> folder when in Mutt, so that when I save\nemails or attachments they go there by default.  I have a little shell function\nset up that <code>cd<\/code>s there for me before running Mutt:<\/p><div class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">alias mutt 'cd ~\/Desktop &amp;&amp; mutt'<\/pre><\/div><p>If you run the <a href=\"http:\/\/ridiculousfish.com\/shell\/\">new fish shell<\/a>, this is going to cause problems later\n(long story, but it\u2019s related to the <code>read<\/code> builtin).  Do yourself a favor and\nhead those confusing issues off at the pass with a fish function:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">function mutt\n    bash --login -c 'cd ~\/Desktop; \/usr\/local\/bin\/mutt' $argv;\nend<\/pre><\/div><p>Remember that if you use another shell like this you\u2019ll want to set up any\naliases and your <code>PATH<\/code> for that shell properly (probably identically to your\nmain shell).<\/p><h2 id=\"reading-email\">Reading Email<\/h2><p>Once you start Mutt you should be looking at a list of the email in your inbox.\nIf so: congratulations!  If not: stop and figure out what went wrong.<\/p><h3 id=\"the-index\">The Index<\/h3><p>When viewing a folder, Mutt presents you with a list of your email.  This view\nis called the \u201cindex\u201d:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-index.png\" alt=\"Mutt's Index\" \/><\/p><p>This entry isn\u2019t meant be a guide to setting up Mutt on OS X.  For a full guide\non how to <em>use<\/em> Mutt, you can Google around for some tutorials, or just learn as\nyou go with <code>?<\/code>.  The <code>?<\/code> key will show you a list of all the keys you can use\nwherever you currently are, and what they do.<\/p><p>Let\u2019s add a few lines to our <code>~\/.muttrc<\/code> to make the index view behave a bit\nmore nicely:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Index View Options ---------------------------------\nset date_format = \"%m\/%d\"\nset index_format = \"[%Z]  %D  %-20.20F  %s\"\nset sort = threads                         # like gmail\nset sort_aux = reverse-last-date-received  # like gmail\nset uncollapse_jump                        # don't collapse on an unread message\nset sort_re                                # thread based on regex\nset reply_regexp = \"^(([Rr][Ee]?(\\[[0-9]+\\])?: *)?(\\[[^]]+\\] *)?)*\"<\/pre><\/div><p>I won\u2019t go into what those do here.  You can read the documentation if you\u2019re\ncurious.<\/p><p>Quit and rerun Mutt to see your changes.  Mutt is a very lightweight program so\nthis should be fast.<\/p><p>Let\u2019s also add a few key bindings in the index to make it easier to use:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Index Key Bindings ---------------------------------\nbind index gg       first-entry\nbind index G        last-entry\nbind index R        group-reply\nbind index &lt;tab&gt;    sync-mailbox\nbind index &lt;space&gt;  collapse-thread\n# Ctrl-R to mark all as read\nmacro index \\Cr \"T~U&lt;enter&gt;&lt;tag-prefix&gt;&lt;clear-flag&gt;N&lt;untag-pattern&gt;.&lt;enter&gt;\" \"mark all messages as read\"\n# Sync email\nmacro index O \"&lt;shell-escape&gt;offlineimap&lt;enter&gt;\"           \"run offlineimap to sync all mail\"\nmacro index o \"&lt;shell-escape&gt;offlineimap -qf INBOX&lt;enter&gt;\" \"run offlineimap to sync inbox\"\n# Saner copy\/move dialogs\nmacro index C \"&lt;copy-message&gt;?&lt;toggle-mailboxes&gt;\" \"copy a message to a mailbox\"\nmacro index M \"&lt;save-message&gt;?&lt;toggle-mailboxes&gt;\" \"move a message to a mailbox\"<\/pre><\/div><p>Remember to quit and rerun Mutt for them to take effect.<\/p><p>We\u2019re going to use <code>j<\/code> and <code>k<\/code> to move around, so we may as well support Vim\nkeys like <code>gg<\/code> and <code>G<\/code> too.  We\u2019ll use <code>R<\/code> for reply all, since that comes in\nhandy fairly often.  <code>Ctrl-R<\/code> will mark all messages in the current folder as\nread.<\/p><p>Don\u2019t worry if you don\u2019t understand how all these bindings and macros work right\nnow.  You can read the documentation later.<\/p><p>The <code>tab<\/code> key is going to \u201ccommit\u201d changes we\u2019ve made in Mutt (like deleting an\nemail) to our local Maildir folder.  Once those changes are in the Maildir\nfolder offlineimap will sync them to the server the next time it runs.  This is\nnice because it lets us recover if we accidentally do something stupid like\ndeleting the wrong email.<\/p><p><strong>Note:<\/strong> Mutt will also sync changes for a folder when you switch to\na different folder, and when you quit Mutt, so be aware of those.<\/p><p>The <code>space<\/code> key will toggle collapsing of threads, which can be convenient when\nviewing mailing lists (or any conversations with many messages).<\/p><p>The <code>o<\/code> and <code>O<\/code> keys will run offlineimap to sync mail.  Like I said before,\nI prefer having to press a button to grab mail instead of it constantly grabbing\nand nagging me.  <code>o<\/code> will sync only the inbox (fast), and <code>O<\/code> will sync\neverything (much slower).<\/p><p>Finally we rebind <code>C<\/code> and <code>M<\/code> to perform the same operations they usually do,\nbut in a more user-friendly manner.<\/p><p>While we\u2019re at it, let\u2019s add a way to navigate around the sidebar so we can\nswitch folders:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Sidebar Navigation ---------------------------------\nbind index,pager &lt;down&gt;   sidebar-next\nbind index,pager &lt;up&gt;     sidebar-prev\nbind index,pager &lt;right&gt;  sidebar-open<\/pre><\/div><p>We\u2019re binding the <code>up<\/code> and <code>down<\/code> arrow keys to switch between folders, and\n<code>right<\/code> to \u201center\u201d a folder.  Give it a try.<\/p><p>We don\u2019t need the arrows because we can navigate with <code>j<\/code> and <code>k<\/code>, but if you\nprefer to rebind them to something else feel free.<\/p><p>Practice moving between folders and around in the list, then we\u2019ll move on to\nactually reading emails.<\/p><h3 id=\"the-pager\">The Pager<\/h3><p>Press <code>return<\/code> in the index to open the selected email.  This view is called the\npager:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-pager.png\" alt=\"Mutt's Pager\" \/><\/p><p>Like before, let\u2019s add a few settings:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Pager View Options ---------------------------------\nset pager_index_lines = 10 # number of index lines to show\nset pager_context = 3      # number of context lines to show\nset pager_stop             # don't go to next message automatically\nset menu_scroll            # scroll in menus\nset tilde                  # show tildes like in vim\nunset markers              # no ugly plus signs\nset quote_regexp = \"^( {0,4}[&gt;|:#%]| {0,4}[a-z0-9]+[&gt;|]+)+\"\nalternative_order text\/plain text\/enriched text\/html<\/pre><\/div><p>This is a good, sane starting point.  And now for a few extra key bindings:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Pager Key Bindings ---------------------------------\nbind pager k  previous-line\nbind pager j  next-line\nbind pager gg top\nbind pager G  bottom\nbind pager R  group-reply\n# View attachments properly.\nbind attach &lt;return&gt; view-mailcap<\/pre><\/div><p>The first few make scrolling behave like it does in the index.  We\u2019re also going\nto use the same key for reply all here.  Consistency will make it easier to get\nMutt into your fingers.<\/p><p>Don\u2019t worry about the last one \u2014 that\u2019s to make sure Mutt treats attachments\nproperly.<\/p><p>Go ahead and try reading some emails.  Remember that <code>?<\/code> will always give you\na list of keys and their functions.<\/p><h3 id=\"attachments\">Attachments<\/h3><p>Now that we\u2019re all set for reading plain text email, it\u2019s time to deal with\nattachments.<\/p><p>When you\u2019re in the pager view reading an email with attachments, you can press\n<code>v<\/code> to view a list of them:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-attachments.png\" alt=\"Attachment List\" \/><\/p><p>Scroll through the list with <code>j<\/code> and <code>k<\/code> and press <code>return<\/code> to view one.  But\nfirst we need to tell Mutt how to view things that aren\u2019t text!<\/p><p>For that we need to create a <code>~\/.mutt\/mailcap<\/code> file.  Here\u2019s a sample to get you\nstarted:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># MS Word documents\napplication\/msword; ~\/.mutt\/view_attachment.sh %s \"-\" '\/Applications\/TextEdit.app'\n# Images\nimage\/jpg; ~\/.mutt\/view_attachment.sh %s jpg\nimage\/jpeg; ~\/.mutt\/view_attachment.sh %s jpg\nimage\/pjpeg; ~\/.mutt\/view_attachment.sh %s jpg\nimage\/png; ~\/.mutt\/view_attachment.sh %s png\nimage\/gif; ~\/.mutt\/view_attachment.sh %s gif\n# PDFs\napplication\/pdf; ~\/.mutt\/view_attachment.sh %s pdf\n# HTML\ntext\/html; ~\/.mutt\/view_attachment.sh %s html\n# Unidentified files\napplication\/octet-stream; ~\/.mutt\/view_attachment.sh %s \"-\"<\/pre><\/div><p>The <code>view_attachment.sh<\/code> script is from <a href=\"http:\/\/linsec.ca\/Using_mutt_on_OS_X#mailcap\">here<\/a>.  Here\u2019s a link to\n<a href=\"https:\/\/bitbucket.org\/sjl\/dotfiles\/src\/tip\/mutt\/view_attachment.sh\">my copy<\/a> in case that site ever goes down.  Grab the script,\nchmod it to executable, and stick it in <code>~\/.mutt<\/code>.<\/p><p>You can poke around and figure out how it works, or you can just not worry about\nit and get on with life.  I recommend the latter (at least for now).<\/p><p>Now you can press <code>return<\/code> to open an attachment in the proper program.<\/p><h3 id=\"urls\">URLs<\/h3><p>One thing you\u2019ll probably want to do while reading email is open links.  Many\nterminal programs like iTerm2 let you command-click on a link to open it, but\nthis is Mutt!  We shouldn\u2019t have to use the mouse!<\/p><p>We\u2019re going to use a small helper program called urlview to make it easy to open\nURLs in email.  First, install it with <code>brew install urlview<\/code>.  Then make\na <code>~\/.urlview<\/code> file with the following contents:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">COMMAND open %s<\/pre><\/div><p>This tells urlview what command to use to open a URL.  We\u2019re just going to use\nthe OS X <code>open<\/code> command to do the right thing.<\/p><p>Next, add the following line to your <code>~\/.muttrc<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">macro pager \\Cu \"|urlview&lt;enter&gt;\" \"call urlview to open links\"<\/pre><\/div><p>Now when you\u2019re reading an email with links in it you can press <code>Ctrl-u<\/code> to open\nurlview.  You\u2019ll see a screen like this:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-urls.png\" alt=\"urlview screen\" \/><\/p><p>Navigate with <code>j<\/code>, <code>k<\/code>, <code>gg<\/code>, <code>G<\/code>, or <code>\/<\/code> and press <code>return<\/code> when the desired\nlink is selected.  That link will be filled in at the bottom of the screen in\ncase you want to edit it, and you can press <code>return<\/code> one more time to actually\nopen it in your default browser.<\/p><p>That about wraps it up for reading email.  Now it\u2019s time to write some!<\/p><p>Writing email is one of the best parts of Mutt.  First let\u2019s add a few settings\nto get things nice and sane:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\"># Compose View Options -------------------------------\nset realname = \"Steve Losh\"          # who am i?\nset envelope_from                    # which from?\nset sig_dashes                       # dashes before sig\nset edit_headers                     # show headers when composing\nset fast_reply                       # skip to compose when replying\nset askcc                            # ask for CC:\nset fcc_attach                       # save attachments with the body\nunset mime_forward                   # forward attachments as part of body\nset forward_format = \"Fwd: %s\"       # format of subject when forwarding\nset forward_decode                   # decode when forwarding\nset attribution = \"On %d, %n wrote:\" # format of quoting header\nset reply_to                         # reply to Reply to: field\nset reverse_name                     # reply as whomever it was to\nset include                          # include message in replies\nset forward_quote                    # include message in forwards<\/pre><\/div><p>You can reply to an email with <code>r<\/code> in the index or pager, or start a fresh one\nwith <code>m<\/code>.<\/p><p>There\u2019s actually not a lot to say about writing mail, because Mutt itself\ndoesn\u2019t handle it!  Mutt passes control off to the text editor of your choice.\nJust specify your editor in your <code>~\/.muttrc<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">set editor = \"vim\"        # Use terminal Vim to compose email.\nset editor = \"mvim -f\"    # Use MacVim to compose email.\nset editor = \"subl -w\"    # Use Sublime Text 2 to compose email.<\/pre><\/div><p>Any command that takes a filename and doesn\u2019t return until you\u2019re done can be\nused here.<\/p><p>This is fantastic because it means you can use an editor you\u2019re already\ncomfortable and fast in to write email instead of learning yet another set of\nshortcuts.<\/p><p>Once you save the email in your editor and close it, Mutt will present you with\na menu that looks like this:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-send-1.png\" alt=\"Sending Screen\" \/><\/p><p>You can press <code>e<\/code> to go back and edit the mail, <code>a<\/code> to add attachments, and so\non (the options are listed at the top of the screen).<\/p><p>Before we can continue we need to tell Mutt how to send email.  Press <code>q<\/code> to\ndiscard the email for now.<\/p><p>Mutt does have (some) built-in SMTP support, but we\u2019re going to use a separate\nprogram to do our sending for a few reasons.<\/p><p>First, Mutt\u2019s SMTP support was considered \u201cexperimental\u201d the last time\nI checked.  Sending email is kind of important, so we\u2019ll stick with something\ntried and true.<\/p><p>Second, we want a method that won\u2019t require our password in a plaintext config\nfile.<\/p><p>Go ahead and install the <code>msmtp<\/code> program through Homebrew with <code>brew install\nmsmtp<\/code>.<\/p><p>Next we\u2019re going to need to create a <code>~\/.msmtprc<\/code> file with the following\ncontents:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">account stevelosh\nhost smtp.gmail.com\nport 587\nprotocol smtp\nauth on\nfrom steve@stevelosh.com\nuser steve@stevelosh.com\ntls on\ntls_trust_file ~\/.mutt\/Equifax_Secure_CA.cert\naccount default : stevelosh<\/pre><\/div><p><code>msmtp<\/code> will look in your keychain for your SMTP password, which we added\nearlier.  No plaintext passwords!<\/p><p>The other \u201cinteresting\u201d bit here is the <code>tls_trust_file<\/code>.  We\u2019re going to be\nconnecting to Gmail\u2019s SMTP server over SSL, and <code>msmtp<\/code> needs to know if it can\ntrust the certificate that the server on the other end is sending back.<\/p><p>Copy the following and paste it into the path <code>tls_trust_file<\/code> is set to:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">-----BEGIN CERTIFICATE-----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\n70+sB3c4\n-----END CERTIFICATE-----<\/pre><\/div><p>If you\u2019re paranoid and don\u2019t trust that I\u2019m giving you the right cert (or that\nsomeone has hacked my site and changed it), you can generate it yourself.  I\u2019ll\nleave that to you \u2014 if you care enough about it you\u2019ll figure it out.<\/p><p>Now we need to tell Mutt to use msmtp.  Add the following to your <code>~\/.muttrc<\/code>\nfile:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">set from     = \"steve@stevelosh.com\"\nset sendmail = \"\/usr\/local\/bin\/msmtp -a stevelosh\"\nset sendmail_wait = 0\nunset record<\/pre><\/div><p>The <code>-a stevelosh<\/code> will need to change to whatever you named your account in the\nmsmtp config.<\/p><p>The <code>unset record<\/code> line tells Mutt to not append a copy of every email you send\nto a file on your hard drive.  Gmail will save the emails you send in the sent\nfolder, so you\u2019ll get the the next time you sync with offlineimap anyway.<\/p><p>The <code>sendmail_wait<\/code> line tells Mutt to wait for the msmtp program to finish\nsending the mail before returning control, instead of running it in the\nbackground.  This makes it obvious if there\u2019s a problem sending a message, which\nI prefer to silent, backgrounded failures.<\/p><p>Now you can send email!  Awesome!<\/p><p>Once you\u2019ve composed a test email and saved it you\u2019ll be presented with a screen\nlike this (which we saw in the previous section):<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-send-1.png\" alt=\"Sending Screen\" \/><\/p><p>The keys you need are listed along the top.  Pressing <code>y<\/code> now will invoke msmtp\nand send your email!<\/p><p>You\u2019ll see \u201cSending message\u2026\u201d at the bottom of the screen while msmtp is\nworking.  If there\u2019s a problem, Mutt will tell you the error.  Figure it out\nbefore moving on.<\/p><h2 id=\"postponing-drafts\">Postponing Drafts<\/h2><p>Sometimes I like to read and respond to email without an internet connection,\nthen actually send the replies when I get back to civilization.<\/p><p>Since all of my email is stored locally, reading it offline is trivial.<\/p><p>To reply or compose offline, I use Mutt\u2019s \u201cpostpone\u201d feature.  First I\u2019ve added\nthe following line to my <code>~\/.muttrc<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">bind compose p postpone-message<\/pre><\/div><p>Now when I\u2019m at the sending screen instead of pressing <code>y<\/code> to send I can press\n<code>p<\/code> to postpone the message.  This places the message in the drafts folder.<\/p><p>The drafts folder in Mutt is just a normal folder of emails like any other.\nWhen you sync with offlineimap your postponed email will get pushed up to Gmail\nas a draft.  You can edit it in the Gmail web interface if you like, and those\nedits will sync back down too.<\/p><p>If you want to edit your drafts (postponed messages) locally, you need to\n\u201crecall\u201d them.  I have the following line in my <code>~\/.muttrc<\/code> to set the key to\nthe same one as I used to postpone it in the first place:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">bind index p recall-message<\/pre><\/div><p>You can also use <code>m<\/code> to start writing a new email, and mutt will prompt you if\nthere are existing postponed messages.<\/p><p>Once you hit <code>p<\/code> (or <code>m<\/code> and then select \u201cyes\u201d) Mutt will show you a list of the\npostponed messages.  Select one and press <code>return<\/code> to start editing it again.\nIf there\u2019s only one it will skip this step and simply open it for you\nimmediately.<\/p><p>Once you\u2019re done editing you can either postpone it again or send it as usual.<\/p><p>Unfortunately you have to go through the \u201crecall \u2192 edit \u2192 send\u201d process each\ntime.  As far as I know there\u2019s no way to simply run down a list of postponed\nemails sending each one with a single keystroke.  But that\u2019s not <em>too<\/em> bad, and\nit\u2019s great to be able to work offline like this.<\/p><p>Once you send the postponed email (through Mutt or the Gmail web interface) it\ndisappears from the drafts folder and the postponed list as you would expect.<\/p><p>Next we\u2019ll want to get Mutt to autocomplete our contacts from the OS X address\nbook.  I like using the OS X address book because it automatically syncs between\nmy laptops and phone, so I only need to maintain one address list.<\/p><h3 id=\"autocompleting\">Autocompleting<\/h3><p>Unfortunately I\u2019ve got some bad news for you: you\u2019re going to need to install\nXCode.<\/p><p>No, not the command-line developer tools.  The full XCode.  I\u2019m sorry, but trust\nme when I say it\u2019s going to save you a lot of pain, so just grumble to yourself\na bit and do it.<\/p><p>Okay, now that you\u2019ve got XCode you can install the <code>contacts<\/code> program through\nHomebrew with <code>brew install contacts<\/code>.<\/p><p><code>contacts<\/code> is a command-line program that you can use to query your address\nbook.  To tell Mutt how to use it add the following lines to your\n<code>~\/.mutt\/muttrc<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">set query_command = \"contacts -Sf '%eTOKEN%n' '%s' | sed -e 's\/TOKEN\/\\t\/g'\"\nbind editor &lt;Tab&gt; complete-query\nbind editor ^T    complete<\/pre><\/div><p>Now when you\u2019re filling out an email address field you can type a few characters\nand hit Tab to get a screen like this:<\/p><p><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2012\/10\/mutt-contacts-1.png\" alt=\"Contacts\" \/><\/p><p>You can use <code>j<\/code> and <code>k<\/code> to select an item, press return to complete it.  Press\n<code>q<\/code> if you\u2019ve changed your mind and want to cancel the completion.  Look at the\ntop of the screen for more handy little keys you can use here.<\/p><p>If there\u2019s only one item in the list Mutt won\u2019t bother showing you this screen\nand will just complete it right away.<\/p><p>This completion searches more than just the email address.  It\u2019ll also search\nthe names and possibly other fields from the address book entries as well.<\/p><h3 id=\"adding-contacts\">Adding Contacts<\/h3><p>What about adding contacts to your address book?  Any contacts you add on your\nphone will automatically be synced, but what if you\u2019re reading your mail in Mutt\nand just want to add the sender as a contact without leaving your command line?<\/p><p>For this I use a little script that <a href=\"http:\/\/simx.me\/\">Simone Manganelli<\/a> wrote for me\ncalled <code>addcontact<\/code>.  You can <a href=\"https:\/\/bitbucket.org\/sjl\/dotfiles\/src\/tip\/bin\/addcontact\">get it here<\/a> and stick it in your\n<code>$PATH<\/code> somewhere.  It\u2019s just a command-line utility that you can use like this:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">$ addcontact Steve Losh steve@stevelosh.com\n$ addcontact \"Steve Losh\" steve@stevelosh.com\n$ addcontact Steve Losh work steve@pculture.org<\/pre><\/div><p>As you can see, it\u2019s pretty flexible.<\/p><p><strong>Note:<\/strong> This utility always adds a new contact record, so if you add someone\nthat\u2019s already in there you\u2019re going to get a duplicate entry.  If that happens\nyou can search for the entries in Contacts.app, select the duplicates, and use\n\u201cCard \u2192 Merge Selected Cards\u201d to combine them.<\/p><p>Okay, so we can now add contacts from the command line.  You could set up a Mutt\nmacro to automatically add senders without too much trouble, but I don\u2019t do\nthat.  If I want to add a contact I just hit <code>!<\/code> and type out the shell command.\nIt\u2019s not that much work, and sometimes the name in the \u201cFrom:\u201d field is in\na weird format like \u201cLast, First Middle\u201d instead of \u201cFirst Last\u201d, so this gives\nme a chance to clean it up before I add it.<\/p><p>If you have more than a screen full of email in any given folder, you\u2019re going\nto want a way to search through it.  Mutt\u2019s built-in searching is a good start,\nbut I also use another program to get a bit more power.<\/p><p>There are two main ways to search your email in Mutt: plain searching and\n\u201climiting\u201d.<\/p><p>Plain searching is done with the <code>\/<\/code> key.  It works similarly to Vim or less:\nyou press <code>\/<\/code>, type your query, and press enter to perform the search.  You can\nuse <code>n<\/code> to move to the next match, and the search will loop around to the top if\nit hits the bottom.<\/p><p>I have the following lines in my <code>~\/.muttrc<\/code> to bind <code>N<\/code> to go to the <em>previous<\/em>\nmatch:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">bind index N search-opposite\nbind pager N search-opposite<\/pre><\/div><p>Normally the <code>N<\/code> key marks a message as unread (or \u201cnew\u201d).  I personally never\nwant to do that.  Unread mail should be \u201cmail that has not been read\u201d.  If you\nuse that feature you\u2019ll want to rebind it to something else.<\/p><p>Two things about queries:<\/p><ol><li>They are regular expressions (actually it\u2019s more powerful than than, see <a href=\"http:\/\/www.mutt.org\/doc\/manual\/manual-4.html#ss4.2\">the\ndocumentation<\/a> for more information).<\/li>\n<li>They only search the To and Subject fields (<em>not<\/em> the message bodies!).<\/li>\n<\/ol><p>I generally use this kind of searching when I see the email I want to open on\nthe screen but don\u2019t feel like pressing <code>j<\/code> or <code>k<\/code> forty times to move through\nthe list.  When I\u2019m actually trying to find a message I can\u2019t already see on the\nscreen I use limiting.<\/p><h3 id=\"vanilla-limiting\">Vanilla Limiting<\/h3><p>Limiting is the other way Mutt provides for searching mail.  It\u2019s done with the\n<code>l<\/code> key by default.<\/p><p>Like <code>\/<\/code>, <code>l<\/code> will ask you for a pattern.  But instead of simply moving you to\nthe next message that matches the pattern, Mutt will <em>hide<\/em> messages that don\u2019t\nmatch.<\/p><p>This lets you see all the ones that match in a single list.  This list works\njust like a normal list.  You can search it with <code>\/<\/code>, tag things, and so on as\nyou normally would.  It\u2019s really quite nice once you get used to it.<\/p><p>To remove the limiting once you\u2019re done, you can limit to the special value\n<code>all<\/code>.  I\u2019ve added a line to my <code>~\/.muttrc<\/code> so I can do that with a single key:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">macro index a \"&lt;limit&gt;all\\n\" \"show all messages (undo limit)\"<\/pre><\/div><p><strong>Note:<\/strong> This shadows the <code>create-alias<\/code> function which I never use.<\/p><p>Limit queries work exactly like search queries, so you can do powerful stuff\nlike <code>~f arthur ~C honza ~s api<\/code> (\u201climit to messages from \u2018arthur\u2019, to or cced\nto \u2018honza\u2019, containing \u2018api\u2019 in the subject\u201d).<\/p><p>By now you\u2019re probably wondering how to search the full text of messages.  There\nare two ways: one simple and slow, the other complicated and fast.<\/p><p>First is the simple, slow way.  You can use <code>~B<\/code> in your searches and limits to\nsearch inside the entire message.  If your folder only has a hundred messages\nthis works great.  But once you have a few more (my archive has about 30,000 at\nthe moment, and I prune it fairly often) it quickly becomes too slow to be\nusable.<\/p><p>I use a program called <a href=\"http:\/\/notmuchmail.org\/\">notmuch<\/a> to index and search my email.  It\u2019s blazingly\nfast and works pretty well.<\/p><p>First, install it with <code>brew install notmuch<\/code>.  Now you need\na <code>~\/.notmuch-config<\/code> file.  Run <code>notmuch setup<\/code> to generate one.  It\u2019s pretty\nstraightforward.  When it asks you for the path to your archive, that\u2019s the path\nto the folder containing all your individual IMAP folders\n(<code>~\/.mail\/steve-stevelosh.com<\/code> in my setup).<\/p><p>Now run <code>notmuch new<\/code> to perform the initial index.  It might take a while if\nyou have a lot of folders.  Subsequent indexes will be much faster since they\u2019ll\nonly reindex whatever changed since the last time.<\/p><p>You\u2019ll want to reindex when your email changes, so add a hook in your\n<code>~\/.offlineimaprc<\/code> to run <code>notmuch new<\/code> whenever offlineimap syncs your email:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">[Account SteveLosh]\nlocalrepository = SteveLosh-Local\nremoterepository = SteveLosh-Remote\nstatus_backend = sqlite\npostsynchook = notmuch new<\/pre><\/div><p>Now you can use <code>notmuch search foo<\/code> to search your mail for \u201cfoo\u201d from the\ncommand line.  Try a couple of queries and make sure it works.<\/p><p>Now we need to hook notmuch into Mutt.  There are a number of different ways to\ndo this, all of them hacky and ugly.  I\u2019ll describe how I do it.<\/p><p>A quick overview of how this is going to work:<\/p><ol><li>You\u2019ll press a key in Mutt to activate searching.<\/li>\n<li>You\u2019ll type your query and press return.<\/li>\n<li>Your mail will be searched with <code>notmuch<\/code>.<\/li>\n<li>\u03a4he resulting messages will be symlinked into a temporary maildir folder.<\/li>\n<li>That temporary folder will be opened in Mutt.<\/li>\n<\/ol><p>First get <a href=\"https:\/\/github.com\/honza\/mutt-notmuch-py\">mutt-notmuch-py<\/a>.  The original <code>mutt-notmuch<\/code> is a Perl script\nwith many external requirements that are a pain to install, and it doesn\u2019t work\non OS X.  <code>mutt-notmuch-py<\/code> is a Python script with zero external requirements.<\/p><p><code>mutt-notmuch-py<\/code> is going to handle steps 2 through 4 in the list.  Get the\nscript into your <code>$PATH<\/code> somehow and then run it:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">$ mutt-notmuch-py -G ~\/.mail\/temporary\/search\nQuery: foo\n$ ls ~\/.mail\/temporary\/search\ncur new<\/pre><\/div><p>The <code>-G<\/code> tells it to not perform certain Gmail-specific stuff that we don\u2019t\nneed.  The path is where it will create the temporary maildir folder with the\nresults.  Each time you run it this folder will be wiped clean before the new\nresults are linked into it.<\/p><p>Now we need to handle the first and last points in the list.  I have the\nfollowing mapping in my <code>~\/.muttrc<\/code>:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">macro index S \"&lt;enter-command&gt;unset wait_key&lt;enter&gt;&lt;shell-escape&gt;mutt-notmuch-py ~\/.mail\/temporary\/search&lt;enter&gt;&lt;change-folder-readonly&gt;+temporary\/search&lt;enter&gt;\" \"search mail (using notmuch)\"<\/pre><\/div><p>That\u2019s a lot to take in, so let\u2019s see how it works piece by piece:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">macro index S<\/pre><\/div><p>We\u2019re going to use the <code>S<\/code> key to perform a full search of all of our mail.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">&lt;enter-command&gt;unset wait_key&lt;enter&gt;<\/pre><\/div><p>Unset the <code>wait_key<\/code> Mutt option to prevent Mutt from asking us to press a key\nafter the search has finished.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">&lt;shell-escape&gt;mutt-notmuch-py -G ~\/.mail\/temporary\/search&lt;enter&gt;<\/pre><\/div><p>Run <code>mutt-notmuch-py<\/code>.  Control will pass over to that script and it will ask\nyou for your query, run the search, set up the results folder, and then hand\ncontrol back to Mutt.<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">&lt;change-folder-readonly&gt;+temporary\/search&lt;enter&gt;<\/pre><\/div><p>Now we change over to the temporary folder in Mutt, and we\u2019re now looking at\na list of search results!  Since this is a real Maildir folder like any other\none we can use the full range of tools to interact with it (searching, limiting,\netc).<\/p><p>Finally, let\u2019s get this search results folder in our sidebar so we can see where\nwe are visually at all times:<\/p><div class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">mailboxes +steve-stevelosh.com\/INBOX \\\n          +steve-stevelosh.com\/vim \\\n          +steve-stevelosh.com\/clojure \\\n          ...\n          +temporary\/search \\<\/pre><\/div><p>Now the search results folder can be navigated like any other one.  That\u2019s it\nfor email searching!  Now you should have a setup that you can use in real life\nto manage your email.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>Mutt is definitely a beast.  It\u2019s old, crufty, and ugly, but if you spend the\ntime to set it up and learn to use it you\u2019ll be rewarded with a fast, powerful,\ncustomizable environment for working with your email.<\/p><p>A wonderful trend these days is that more and more sites are including the\nability to respond to comments and such by simply replying to their notification\nemails.  This means that often you can reply to Facebook emails, comment on\nGitHub pull requests, and respond to Bitbucket issues all without leaving the\ncomfort of your finely-tuned email client.<\/p><p>Mutt\u2019s not for everyone, but if you deal with a lot of email and have some time\nto spend you should give it a try.  You just might learn to love the old dog.<\/p>","created_at":"2018-11-07T16:53:00+0000","updated_at":"2018-11-07T16:53:00+0000","annotations":[],"mimetype":"text\/html","reading_time":42,"domain_name":"stevelosh.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":589,"title":"Vim as E-Mail Editor","url":"http:\/\/www.256bit.org\/~chrisbra\/cms\/vim_as_e-mail_editor.html","content":"Vim is a great editor and can also used for writing Emails. I am using mutt as my mail user agent and have configured several scripts for using Vim with mutt.<br \/>You need to configure Vim to support filetype plugins. Vim by default can detect mails when used from vim and will set the buffers filetype to mail and load mail specific scripts. For this make sure, your .vimrc contains at least a statement such as<p>filetype plugin on<\/p><p>That's already it. All the coming scripts, if not explained otherwise can now be placed into a directory ~\/.vim\/ftplugin\/mail and Vim will automatically load them, when using Vim for writing Emails.<br \/><\/p><a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/format.vim\">format.vim<\/a>(<a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/format.vim.html\">html<\/a>)<p>This file let's Vim guess what kind of text you are writing and sets some formatoptions. So it sets different formatoptions, depending on whether you are editing the header of the Email, are fixing some quotations or edit a patch. \u00a0Credits go to  <a class=\"\" href=\"http:\/\/groups.google.com\/group\/vim_use\/msg\/f59e5c1adc6be2b3\">Teemu Likonen<\/a>.<\/p><a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/posting.vim\">posting.vim<\/a>(<a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/posting.vim.html\">html<\/a>)<p>Allows to automatically format your mail (e.g. strip quoted signatures, add Greeting, change and clean subject). <br \/>Note: For using the Greeting feature, make sure you define the variables b:Signoff and b:Name.<br \/>I explicitly set those variables in my .muttrc like this:<\/p><p>set editor='vim --cmd \"let b:Signoff=\\\"Gr\u00fc\u00dfe\\\"|let b:Name=\\\"Christian\\\"\"'<\/p><p>Using folder-hooks, you can configure those variables to your liking, e.g. have english greeting when opening a mailinglist folder.<\/p><p>Check out the mappings, that are defined in this script:<br \/>\\mdq \u00a0 Deletes quoted signatures<br \/>\\mds \u00a0 \u00a0Deletes old subject<br \/>\\mct \u00a0 \u00a0Change the subject (adds the (was...) around it)<br \/>\\mcs \u00a0 \u00a0Clean subject (removes the (was....) part)<\/p><a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/lbdbq.vim\">lbdbq.vim<\/a>(<a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/lbdbq.vim.html\">html<\/a>)<p>This is Vim-Script <a class=\"\" href=\"http:\/\/www.vim.org\/scripts\/script.php?script_id=1757\">Plugin 1757<\/a> \u00a0and lets you use <a class=\"\" href=\"http:\/\/www.spinnaker.de\/lbdb\/lbdbq.html\">lbdbq<\/a> for querying various adressbooks. I mainly use it for querying <a class=\"\" href=\"http:\/\/abook.sourceforge.net\/\">abook<\/a> and mutt's aliases file. Simply enter the search term and press \\lb. Vim will query lbdbq for your search-term and present a dialog of adresses to use. <\/p><a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/lbdbq_complete.vim\">lbdbq_complete.vim<\/a>(<a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/lbdbq_complete.vim.html\">html<\/a>)<p>This is an addon to the lbdq.vim script from above. It adds a omni-completion, so you can enter search-terms and press &lt;Ctrl-X&gt;&lt;Ctrl-U&gt; in insert mode to have Vim display a small completion menu for the adresses available.<\/p><a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/stl.vim\">stl.vim<\/a>(<a class=\"\" href=\"http:\/\/www.256bit.org\/~chrisbra\/cms\/Z-files\/stl.vim.html\">html<\/a>)<p>Displays the Author of the Email to whom you are responding in the statusline. Nothing more to say.<\/p><br \/>Use the Vim Plugin<a class=\"\" href=\"http:\/\/www.vim.org\/scripts\/script.php?script_id=2796\">CheckAttach<\/a>(<a class=\"\" href=\"https:\/\/github.com\/chrisbra\/CheckAttach\">github<\/a>)<br \/>Reminds you, if you want to attach files to your mails and you can select the files using filename completion. You can also attach files manually using the :AttachFiles command. Supports<a class=\"\" href=\"http:\/\/ranger.nongnu.org\/\">ranger<\/a>as file browser.","created_at":"2018-11-07T16:52:37+0000","updated_at":"2018-11-07T16:52:37+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"www.256bit.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":587,"title":"How to present a GitHub project for your resume","url":"https:\/\/thehftguy.com\/2016\/10\/24\/heres-how-to-make-a-good-github-project-for-your-resume\/","content":"<div class=\"pd-rating\" id=\"pd_rating_holder_8362781_post_3796\"><br \/><h1>Introduction<\/h1>\n<p>Companies ask for a GitHub profile. Recruiters ask for a GitHub profile. The question \u201c<em>Do you contribute to open-source?<\/em>\u201d is now one of the most common questions asked in phone screens.<\/p>\n<p>If people want a GitHub, we shall give them a GitHub. This article will explain how to present a GitHub project for use in a resume.<\/p>\n<p>The given advice can be read from two point of views. As a candidate, it is what to write to introduce and present a software (not necessary on GitHub). As an interviewer (or a fellow developer), it is what to look for to judge the experience of the developer(s) and the quality of a software.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_3948\" class=\"wp-caption aligncenter\"><img data-attachment-id=\"3948\" data-permalink=\"https:\/\/thehftguy.com\/2016\/10\/24\/heres-how-to-make-a-good-github-project-for-your-resume\/github-submit-application\/\" data-orig-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png\" data-orig-size=\"545,435\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"github-submit-application\" data-image-description=\"\" data-medium-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png?w=300&amp;h=239\" data-large-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png?w=545\" class=\"aligncenter size-medium wp-image-3948\" src=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png?w=300&amp;h=239\" alt=\"submit application form with a github link\" width=\"300\" height=\"239\" srcset=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png?w=300&amp;h=239 300w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png?w=150&amp;h=120 150w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-submit-application1.png 545w\" \/><figcaption class=\"wp-caption-text\">When having a GitHub is mandatory, just like having a name.<\/figcaption><\/figure><h1>Link to a specific project<\/h1>\n<p>Put a link to your GitHub in your resume and every application forms you have to fill.<\/p>\n<p>That link must send directly to a project. Never link to the root of your GitHub profile, it doesn\u2019t show anything useful and it\u2019s hard to navigate from there.<\/p>\n<p>It means that you must have ONE project to show. A single demonstration project is enough, don\u2019t need more.<\/p>\n<p>This project will be the \u201c<em>landing page<\/em>\u201d in web buzzword. This is the first page the employer will see. They will rarely go past it (and they shouldn\u2019t have to) so the page should be a good enough by itself. If they go past it, it\u2019s only because the page grabbed their interests and they wanted to see more.<\/p>\n<p>We\u2019ll write the project page to give a good first impression and show off skills as a software engineer.<\/p>\n<h1>Project Structure<\/h1>\n<p>A software project can be judged in 5 seconds by looking at the directory structure.<\/p>\n<p>An inexperienced developer is easy to spot. His project doesn\u2019t have any structure. Files are either in unpredictable places or all in the top directory.<\/p>\n<p><strong>There is one project structure to rule them all. There MUST be separate directories for source, test, libraries, compiled binaries, etc\u2026<\/strong><\/p>\n<p>Whether the naming convention will be \u201c<em>doc<\/em>\u201d or \u201c<em>docs<\/em>\u201d is an unimportant detail. For example, here are <a href=\"https:\/\/github.com\/kriasoft\/Folder-Structure-Conventions\" target=\"_blank\" rel=\"noopener\">Simple Folder Structure Conventions<\/a> for GitHub projects:<\/p>\n<pre><code>.\n\u251c\u2500\u2500 build                   # Compiled files (alternatively `dist`)\n\u251c\u2500\u2500 docs                    # Documentation files (alternatively `doc`)\n\u251c\u2500\u2500 src                     # Source files (alternatively `lib` or `app`)\n\u251c\u2500\u2500 test                    # Automated tests (alternatively `spec` or `tests`)\n\u251c\u2500\u2500 tools                   # Tools and utilities\n\u251c\u2500\u2500 LICENSE\n\u2514\u2500\u2500 README.md\n<\/code><\/pre>\n<figure data-shortcode=\"caption\" id=\"attachment_4111\" class=\"wp-caption alignnone\"><img data-attachment-id=\"4111\" data-permalink=\"https:\/\/thehftguy.com\/2016\/10\/24\/heres-how-to-make-a-good-github-project-for-your-resume\/github-project-structure\/\" data-orig-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=809\" data-orig-size=\"1002,500\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"github-project-structure\" data-image-description=\"\" data-medium-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=809?w=300\" data-large-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=809?w=809\" class=\"alignnone size-full wp-image-4111\" src=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=809\" alt=\"software project structure\" srcset=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=809 809w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=150 150w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=300 300w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png?w=768 768w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-project-structure.png 1002w\" \/><figcaption class=\"wp-caption-text\">A well-organized project<\/figcaption><\/figure><h1>Have a README<\/h1>\n<p>Have a README to:<\/p>\n<ul><li>Describe the purpose of the project<\/li>\n<li>Screenshots\/videos<\/li>\n<li>Usage<\/li>\n<li>Link to the installer\/webpage<\/li>\n<li>\u2026<\/li>\n<\/ul><h1>Have screenshots in the readme<\/h1>\n<p><strong>A picture is worth a thousand words.<\/strong><\/p>\n<p>People are not going to install the application just to see it. Give them screenshots.<\/p>\n<h1>Have videos in the readme<\/h1>\n<p><strong>A picture is worth a thousand words. A video is worth a thousand pictures.<\/strong><\/p>\n<p>There is nothing better than a video when it comes to giving a demonstration or showing off an application.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_3846\" class=\"wp-caption alignnone\"><img data-attachment-id=\"3846\" data-permalink=\"https:\/\/thehftguy.com\/2016\/10\/24\/heres-how-to-make-a-good-github-project-for-your-resume\/github-snake-gif\/\" data-orig-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-snake-gif.gif?w=809\" data-orig-size=\"600,370\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"github-snake-gif\" data-image-description=\"\" data-medium-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-snake-gif.gif?w=809?w=300\" data-large-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-snake-gif.gif?w=809?w=600\" class=\"alignnone size-full wp-image-3846\" src=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-snake-gif.gif?w=809\" alt=\"snake game preview animated gif\" \/><figcaption class=\"wp-caption-text\">Great demo from <a href=\"https:\/\/github.com\/DyegoCosta\/snake-game\" target=\"_blank\" rel=\"noopener\">a random snake project on GitHub<\/a><\/figcaption><\/figure><p><strong>Note<\/strong>: GitHub does not allow to embed video files in the readme, use animated gif instead.<\/p>\n<h1>Link to a website or an installer<\/h1>\n<p>Link to the site if it\u2019s a web application project. Of course, <strong>a web application should be running somewhere and publicly accessible, that\u2019s the point of web applications<\/strong>.<\/p>\n<p>Link to the installer if it\u2019s a desktop application project. It\u2019s unlikely that the user will install it but that looks professional, that\u2019s how desktop applications are distributed after all.<\/p>\n<h1>Integrate GitHub tools<\/h1>\n<p>GitHub has a rich ecosystem of free tools for building, packaging, testing and much more. All these tools are mandatory for professional software development.<\/p>\n<p>It used to be hell to setup the tooling but now everything is readily available for free through GitHub and the setup is dead simple. There is no excuse to not use the tooling.<\/p>\n<p><img data-attachment-id=\"3871\" data-permalink=\"https:\/\/thehftguy.com\/2016\/10\/24\/heres-how-to-make-a-good-github-project-for-your-resume\/github-integration-icons\/\" data-orig-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png?w=809\" data-orig-size=\"508,94\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"github-integration-icons\" data-image-description=\"&lt;p&gt;github integration icons screenshot&lt;\/p&gt;&#10;\" data-medium-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png?w=809?w=300\" data-large-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png?w=809?w=508\" class=\"size-full wp-image-3871 aligncenter\" src=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png?w=809\" alt=\"github-integration-icons\" srcset=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png 508w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png?w=150 150w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-integration-icons.png?w=300 300w\" \/><\/p>\n<p>This one is a sample C++ project for a Connect Four. From left to right:<\/p>\n<ol><li>Build on Linux (<a href=\"https:\/\/travis-ci.org\/\" target=\"_blank\" rel=\"noopener\">Travis CI<\/a>)<\/li>\n<li>Build on Windows (<a href=\"https:\/\/www.appveyor.com\/\" target=\"_blank\" rel=\"noopener\">AppVeyor<\/a>)<\/li>\n<li>Unit tests and coverage analysis (<a href=\"https:\/\/coveralls.io\/\" target=\"_blank\" rel=\"noopener\">Coveralls<\/a>)<\/li>\n<\/ol><h1>What about the source code?<\/h1>\n<p>\u201c<em>Nobody cares about your code. It was quite a shocking moment when I learned this in my programming career. I would take great care in polishing my code only to find out nobody actually cares. It\u2019s not the code that counts, it\u2019s the product. <\/em>\u201d \u2014 <a href=\"https:\/\/mortoray.com\/2015\/04\/20\/nobody-cares-about-your-code\/\">Source<\/a><\/p>\n<p>A paragraph to explain the purpose of the application is 10 times faster than guessing it. A quick start video of an [non-trivial] application is 100 times faster than figuring it out. A design diagram is 1000 times faster than reverse engineering the application. All of these could be achieved by reading the source code, at the cost of orders of magnitude more time and headache. It\u2019s extremely slow and difficult to read code (or should we say to <strong><em>decode code<\/em><\/strong>). It should only ever be a last resort.<\/p>\n<p>Lesson #1: Noone cares about your source code. Noone is gonna read it.<\/p>\n<p>Lesson #2: Don\u2019t expect people to read it. <strong>Don\u2019t force them to<\/strong>.<\/p>\n<h1>What if I don\u2019t have big projects to show?<\/h1>\n<p>Good. Smaller projects are easier to show, easier to explain and easier to understand for the interviewer. For instance, everyone can grasp a good old Connect Four.<\/p>\n<p>It is not a trivial project despite what it looks like at first. Write a decent UI, put some colors, allow a two players option, add a \u201c<em>hint<\/em>\u201d to show the best next move, add an AI to play against.<\/p>\n<p>While the game is conceptually simple there is a lot of work to turn it into a good and polished software. That leaves plenty of depth to talk about in a face-to-face interview.<\/p>\n<p>Did you know that the first player in a connect four game always wins? [if playing perfectly] Did you know that the second player can always draw the game, if the first player doesn\u2019t take the middle position as his first move?<\/p>\n<p>Source: <a href=\"http:\/\/www.informatik.uni-trier.de\/~fernau\/DSL0607\/Masterthesis-Viergewinnt.pdf\" target=\"_blank\" rel=\"noopener\">A Knowledge-based Approach of Connect-Four<\/a>, <em>The Game is Solved: White Wins<\/em>, Victor Allis<\/p>\n<h1>Do interviewers really look at GitHub?<\/h1>\n<p>As a matter of fact: <strong>No, they don\u2019t.<\/strong><\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_4010\" class=\"wp-caption aligncenter\"><img data-attachment-id=\"4010\" data-permalink=\"https:\/\/thehftguy.com\/2016\/10\/24\/heres-how-to-make-a-good-github-project-for-your-resume\/github-traffic\/\" data-orig-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=809\" data-orig-size=\"980,279\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"github-traffic\" data-image-description=\"\" data-medium-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=809?w=300\" data-large-file=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=809?w=809\" class=\"size-full wp-image-4010 aligncenter\" src=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=809\" alt=\"github traffic statistics\" srcset=\"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=809 809w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=150 150w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=300 300w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png?w=768 768w, https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-traffic.png 980w\" \/><figcaption class=\"wp-caption-text\">GitHub Traffic Statistics<\/figcaption><\/figure><p>We\u2019ve done the tests. Here are the statistics after sending a bunch of resumes. The 3 views are from myself, I accessed the project while writing this article, without being authenticated to GitHub. Oops, my bad.<\/p>\n<p>From personal experience from the last time I looked for a job. After a dozen of phone interviews (1 dev per call) and a couple of on-sites (4 to 7 devs per on-site), there was only 1 visit to my profile.<\/p>\n<p><strong>Conclusion: Noone cares about GitHub. Noone is gonna read it. <\/strong>Everyone is gonna ask for it nonetheless, cause it\u2019s hype.<\/p>\n<p><strong>Bonus<\/strong>: Since noone will check the link they\u2019re given, you too can refuse to participate in the GitHub masquerade by answer all GitHub request with <a href=\"https:\/\/github.com\/leachim6\/hello-world\">the ultimate hello world repository<\/a>. This is a repository of hello world in hundreds of languages.<\/p>\n<h1>Cheat Sheet<\/h1>\n<ol><li>Structure the project<\/li>\n<li>Have a README<\/li>\n<li>Write a paragraph to explain the purpose of the project<\/li>\n<li>Put screenshots and videos<\/li>\n<li>Distribute an installer (desktop app) or give the website (web app)<\/li>\n<li>Integrate development tools (CI, unit test, packager, etc\u2026)<\/li>\n<\/ol><p>That\u2019s good practices for software projects. It\u2019s not limited to GitHub.<\/p>\n<div id=\"s3gt_translate_tooltip\" class=\"s3gt_translate_tooltip\">\n<div class=\"s3gt_translate_tooltip\">\n\t\t<div class=\"wpcnt\">\n\t\t\t<div class=\"wpa wpmrec\">\n\t\t\t\tAdvertisements\n\t\t\t\t<div class=\"u\">\t\t<div>\n\t\t<div id=\"atatags-1080840231-5be31853babc8\">\n\t\t\t\n\t\t<\/div><\/div>\t\t<div>\n\t\t<div id=\"atatags-1080840234-5be31853babcb\">\n\t\t\t\n\t\t<\/div><\/div><\/div>\n\t\t\t\t\n\t\t\t<\/div>\n\t\t<\/div><div id=\"atatags-1080840235-5be31853bac04\">\n        \n    <\/div>\t\t\t\t\t<\/div><\/div><\/div>","created_at":"2018-11-07T16:52:36+0000","updated_at":"2018-11-07T16:52:36+0000","published_at":"2016-10-24T13:06:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"thehftguy.com","preview_picture":"https:\/\/thehftguy.files.wordpress.com\/2016\/10\/github-snake-gif.gif","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":588,"title":"Passwords Evolved: Authentication Guidance for the Modern Era","url":"https:\/\/www.troyhunt.com\/passwords-evolved-authentication-guidance-for-the-modern-era\/","content":"<p>In the beginning, things were simple: you had two strings (a username and a password) and if someone knew both of them, they could log in. Easy.<\/p><p>But the ecosystem in which they were used was simple too, for example in <a href=\"https:\/\/www.wired.com\/2012\/01\/computer-password\/\">MIT's Time-Sharing Computer<\/a>, considered to be the first computer system to use passwords:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Massachusetts-Institute-of-Technology.jpg\" alt=\"Massachusetts Institute of Technology\" \/><\/p><p>We're talking back in the 60's here so a fair bit has happened since then. Up until the last couple of decades, we had a small number of accounts and very limited connectivity which made for a pretty simple threat landscape. Your \"adversaries\" were those in the immediate vicinity, that is people who could gain direct physical access to the system. Over time that extended to remote users who could dial in - I mean <em>literally<\/em> dial in via phone - and that threat landscape grew. You pretty much know the story from here: more connectivity, more accounts, more threat actors and particularly in recent years, more data breaches. Suddenly, the simple premise of matching strings no longer seems like such a good idea.<\/p><p>A couple of months ago I wrote about <a href=\"https:\/\/www.troyhunt.com\/password-reuse-credential-stuffing-and-another-1-billion-records-in-have-i-been-pwned\/\">Password reuse, credential stuffing and another billion records in Have I been pwned<\/a> (HIBP). Here we have a situation where there's a 10-figure number of credentials sitting there waiting for evildoers to start testing them against any site of their choosing and that presents a very interesting challenge: how do we defend against this? I mean you're trying to run your online system and someone has <em>valid credentials for some of your users<\/em>, how are you going to stop them from getting in? The simple string matching of the 60's just isn't going to cut it.<\/p><p>There's a lot more to how authentication has evolved than just the rise and rise of credential stuffing though, many other aspects of how we logon to systems has also changed. In some cases, this has led to once-held \"truths\" about how we create and manage accounts to be totally flipped on their head, yet we still see modern organisations applying the patterns of yesterday to the threats of today. This post sets out to address this gap and talk about how we <em>should<\/em> be designing this critical part of our systems today. My hope is that in times where a company says \"we're doing this screwy thing because security\", this post becomes the resource that well-wishers direct them to.<\/p><h2 id=\"listentoyourgovernmentsandsmarttechcompanies\">Listen to Your Governments (and Smart Tech Companies)<\/h2><p>Let me start by referencing others because there's a lot of great material out there from recent times which I'm going to draw on. I want to lay these out early to be clear that a lot of the guidance I'll outline below is not the personal views of one individual, rather it's direct from the likes of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Institute_of_Standards_and_Technology\">National Institute of Standards and Technologies<\/a> (NIST). In fact, their recently released <a href=\"https:\/\/www.nist.gov\/itl\/tig\/special-publication-800-63-3\">Digital Identity Guidelines<\/a> from only last month (this is SP 800-63 for those interested in the details) were arguably the catalyst for me finally putting this together because there's so much good stuff in there.<\/p><p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/National_Cyber_Security_Centre_(United_Kingdom)\">National Cyber Security Centre<\/a> (NCSC) in the UK government is another excellent resource I'll be drawing on. They've consistently put out really thoughtful pieces on the topic at hand and refreshingly, it's an example of a government department really \"getting it\" when it comes to modern day tech.<\/p><p>I'll also be referring to <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2016\/06\/Microsoft_Password_Guidance-1.pdf\">Microsoft's Password Guidance<\/a> paper from the Identity Protection Team. The first page of the document talks about Microsoft having a \"unique vantage point to understand the role of passwords in account takeover\" due to them seeing 10 <em>million<\/em> attacks every day so yeah, they have some experience in this area! This is a very practical document that's been put together by people who put a huge amount of thought into keeping online accounts safe.<\/p><p>I'm sure there are other such examples too and I welcome people to add these to the comments section below. That's just a few sources and you'll see many others referenced throughout the remainder of this post. Let's get into it!<\/p><p>I want to start with a more philosophical look at how authentication usually works: you're not logged on so you have no access to anything then you logon and you have full access to everything that your account should have rights too. There are no grey areas, it's either one or the other.<\/p><p>One of the really neat things we're seeing in many aspects of infosec these days is the recognition of threat or confidence levels, that is that sometimes we're more confident in a particular risk scenario (i.e. logging on) than at other times. For example, rather than just allowing multiple logon attempts to an account and then locking it out completely after say, 5 failed attempts (notice how this is always an odd number?), perhaps after 3 attempts a CAPTCHA should be presented. This recognises that confidence in the user has dropped and steps should be taken to ensure an automated attack isn't being mounted. It's tough on bots, but a minor inconvenience for users and you can still lock out after X more failed attempts if necessary.<\/p><p>Likewise, once someone is successfully authenticated, should they have full access to all features? For example, if it took a few goes to get the password right and they've come in from a previously unseen browser in a different country to usual, should they have unbridled access to everything? Or should they be limited to basic features and must verify they still control the registered email address before doing anything of significance?<\/p><p>These are simple examples but the thought process I'm trying to get going is that we can be a lot smarter than the traditional binary authentication state that still prevails in the vast majority of systems today. You'll see this theme pop up a few times during the remainder of this blog post.<\/p><h2 id=\"longerisusuallystronger\">Longer is (Usually) Stronger<\/h2><p>Let's get into the nuts and bolts of things and we'll start with something easy: this is not a good password policy:<\/p><p>That first sentence in that pop-up is probably one of the most common poor password anti-patterns going, that is a short arbitrary length. It kills password managers (more on them soon), it kills pass phrases and subsequently, it kills usability. On that last point, the tweet above is from a 2016 blog post of mine on <a href=\"https:\/\/www.troyhunt.com\/security-insanity-how-we-keep-failing-at-the-basics\/\">how we keep failing at the basics<\/a> and along with Etihad's bad policy (which incidentally, <a href=\"https:\/\/twitter.com\/EtihadAirways\/status\/748626413306150912\">they allegedly do \"because security\"<\/a>), I show how PayPal effectively locked me out of my account due to a similar policy.<\/p><p>In addition to the problems mentioned above, short arbitrary limits like this regularly cause people to speculate that password storage is insufficient. When cryptographically hashed, all passwords are stored with the same fixed length so an arbitrary limit such as the one above may indicate the password is stored in a plain text and the column only allows 10 characters. Now that's not necessarily the case (<del>Microsoft limits their accounts to 16 characters and I'm confident they have very robust password storage<\/del>), but you can see how it causes the aforementioned speculation.<\/p><p>Edit: See <a href=\"https:\/\/www.troyhunt.com\/passwords-evolved-authentication-guidance-for-the-modern-era\/#comment-3437375825\">Ariel Gordon's comment below<\/a> regarding Microsoft no longer imposing this limit.<\/p><p>So how long <em>should<\/em> you allow a password to be? No, not \"as long as you want\" because there is a size at which you have other problems. For example, at over 4MB you'd exceed <a href=\"https:\/\/msdn.microsoft.com\/en-us\/library\/system.web.configuration.httpruntimesection.maxrequestlength(v=vs.110).aspx\">the default ASP.NET max request size<\/a>. Here's NIST's view:<\/p><blockquote>\n<p>Verifiers SHOULD permit subscriber-chosen memorized secrets at least 64 characters in length<\/p>\n<\/blockquote><p>No reasonable person is going to use a website with a 64-character password limit then turn around and say \"this site's security is crap because they didn't let me use more than 64 characters in my password\". But just to be sure, make it 100. Or 200. Or stick with NIST's thinking and make it 256, it doesn't matter because it's going to hash down to the same number of characters anyway.<\/p><p>NIST also makes another important if not obvious point when it comes to password length:<\/p><blockquote>\n<p>Truncation of the secret SHALL NOT be performed<\/p>\n<\/blockquote><p>This is really the simplest of concepts: don't have a short arbitrary password length and don't chop characters off the end of a password provided by a user. At the very least, an organisation defending this position should say \"we know it's bad, there's legacy reasons, we'll put it on the road map to be rectified\".<\/p><h2 id=\"allcharactersarespecialbutyoudonthavetohavespecialcharacters\">All Characters Are Special (But You Don't Have to Have Special Characters)<\/h2><p>I want to look at two different aspects of special characters and I'll start with this one:<\/p><p>Typically, certain characters are disallowed as a means of defending against potential attacks. For example, angle brackets may be used in XSS attacks and an apostrophe may be used in SQL injection attacks. However, both of these arguments show serious shortcomings in the security profile of the site in question because firstly, passwords should never be re-displayed in the UI where an XSS risk could be exploited and secondly, because they should never be sent to the database without being hashed which would mean only alphanumeric characters prevail. Plus of course you have output encoding and parameterisation even if you <em>were<\/em> inappropriately handling passwords by re-displaying them in the UI and saving them in plain text to the database.<\/p><p>NIST is pretty clear on this - don't do it:<\/p><blockquote>\n<p>All printing ASCII [RFC 20] characters as well as the space character SHOULD be acceptable in memorized secrets. Unicode [ISO\/ISC 10646] characters SHOULD be accepted as well.<\/p>\n<\/blockquote><p>It's also worth noting their point on Unicode characters; there are many precedents of sites restricting perfectly valid language characters simply because they don't fit into what a site's developers consider \"normal\". Then, of course, there are passwords like this:<\/p><pre>The \u2744\ufe0f \ud83c\udf1f \ud83d\udd26 \u26aa on the mountain \ud83c\udf19 \ud83c\udf20. \ud83d\ude45\ud83c\udffb a\ud83d\udc63 to \ud83d\udc1d \ud83d\udc40. A \ud83c\udff0 of \ud83d\ude22, and it \ud83d\udc40 like\u261d\ufe0f\ufe0f the \ud83d\udc51. The \ud83d\udca8 is \ud83d\udc3a like this \ud83c\udf00 \u2744\ufe0f \u2614\ufe0f \ud83c\udfe0. \ud83d\ude45\ud83c\udffb keep it in, \u2601\ufe0f \ud83d\udca1 \u261d\ufe0f\ufe0f tried.<\/pre><p>If someone <em>really<\/em> wants to have a password that's an emoji representation of the first verse of \"Let It Go\" from Frozen, good on 'em!<\/p><p>The other aspect of special characters is this from NIST:<\/p><blockquote>\n<p>Verifiers SHOULD NOT impose other composition rules (e.g., requiring mixtures of different character types or prohibiting consecutively repeated characters) for memorized secrets<\/p>\n<\/blockquote><p>Wait - what?! You mean people aren't required to have lowercase, uppercase, numbers and symbols?! This goes against so much of conventional password wisdom and it's not until you step back and think about it logically that it really makes sense. To demonstrate this, I recently wrote about <a href=\"https:\/\/www.troyhunt.com\/password-strength-indicators-help-people-make-dumb-choices\/\">how password strength indicators help people make ill-informed choices<\/a>, the premise of which was that mathematics alone (which is what most password strength meters use), does not help us make strong passwords. Requirements such as these leave scope to use passwords such as \"Passw0rd!\" and rule out passwords such as lowercase passphrases.<\/p><p>Microsoft has precisely the same guidance as NIST in their aforementioned documentation:<\/p><blockquote>\n<p>Eliminate character-composition requirements<\/p>\n<\/blockquote><p>What tends to happen when there are requirements around password complexity is that people first try something basic then they tweak characters until it comes up to the minimum requirement of the site. Microsoft explains the problem as follows:<\/p><blockquote>\n<p>Most people use similar patterns (i.e. capital letter in the first position, a symbol in the last, and a number in the last 2). Cyber criminals know this, so they run their dictionary attacks using the common substitutions, such as \"$\" for \"s\", \"@\" for \"a,\" \"1\" for \"l\" and so on.<\/p>\n<\/blockquote><p>This will almost certainly still be a bad password and almost certainly one they've previously used in other places too so all that's been achieved here is that the user has been put through a level of frustration in order to still arrive at a bad password!<\/p><h2 id=\"passwordhintsaredefinitelyout\">Password Hints Are <em>Definitely<\/em> Out<\/h2><p>Anecdotally, password hints are far less frequent today than what they used to be. This was the premise of \"hey, people forget passwords, let's make it easier for them to remember\" and it meant that at signup, as well as providing a password you could provide a <em>hint<\/em> as to what the password actually is. The problem is that this hint is shown to unauthenticated users because that's the precisely the point at which it's needed. The other problem is that because this is usually a user-provided piece of data, it's probably going to be terrible.<\/p><p>Adobe stored password hints in their database that was disclosed back in 2013. Just to illustrate the terribleness of these hints, I went back to take a look at the data and thought I'd highlight a few of them here:<\/p><ul><li>my name<\/li>\n<li>adobe<\/li>\n<li>usual<\/li>\n<li>password<\/li>\n<li>email<\/li>\n<\/ul><p>These were all stored in plain text too so think of what that meant once the system was compromised. For obvious reasons, NIST thinks these are a bad idea:<\/p><blockquote>\n<p>Memorized secret verifiers SHALL NOT permit the subscriber to store a \u201chint\u201d that is accessible to an unauthenticated claimant.<\/p>\n<\/blockquote><p>Like I said, they're rare to see in an online system these days anyway but just in case you were thinking about doing this, don't!<\/p><h2 id=\"embracepasswordmanagers\">Embrace Password Managers<\/h2><p>I've been going on about the value of password managers for a long time now, since early 2011 actually when I wrote about how <a href=\"https:\/\/www.troyhunt.com\/only-secure-password-is-one-you-cant\/\">the only secure password is the one you can\u2019t remember<\/a>. The premise is simple and I'll boil down to a few bullet points:<\/p><ol><li>We know that passwords must be \"strong\", that is that they shouldn't be predictable or readily brute forced so in other words, the longer and more random, the better.<\/li>\n<\/ol><ul><li>We know that passwords shouldn't be reused because disclosure by one service puts the user's other services at risk. This is the whole credential stuffing problem I referred to earlier.<\/li>\n<li>People cannot create strong, unique passwords across all their services using only their brain to remember which one they used where.<\/li>\n<\/ul><p>Now some people will argue that a password manager means putting all your eggs in one basket and they're right; if that basket gets compromised it's going to be bad news. But this is an exceptionally rare event compared to the compromise of an individual service which consequently exposes credentials. Further to that and as I wrote more recently, <a href=\"https:\/\/www.troyhunt.com\/password-managers-dont-have-to-be-perfect-they-just-have-to-be-better-than-not-having-one\/\">password managers don't have to be perfect, they just have to be better than not having one<\/a>.<\/p><p>But this post isn't intended to provide guidance to individuals about how to secure their accounts, rather it's aimed at people building services. This means that regardless of your personal views on password managers, you shouldn't be doing this:<\/p><p>This is typical short-sightedness and I could easily point to dozens of other tweets of a similar nature. It entirely neglects the 3 bullet points I outlined earlier and what it's essentially doing is saying \"hey, create a password you can easily remember and yeah, it'll be weak and probably the same as your other ones, but do it because security\".<\/p><p>The NCSC is quite explicit about this and has the following in the infographic accompanying their <a href=\"https:\/\/www.ncsc.gov.uk\/guidance\/password-guidance-simplifying-your-approach\">Password Guidance: Simplifying Your Approach<\/a> resource:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Allow-users-to-securely-record-and-store-their-passwords.png\" alt=\"Allow users to securely record and store their passwords\" \/><\/p><p>In other words, don't break password managers and don't trot out lines like in the tweet above. But the NCSC goes even further, providing the following recommendation targeted clearly at how organisations enable their staff to create and manage passwords in a secure fashion:<\/p><blockquote>\n<p>You should also provide appropriate facilities to store recorded passwords, with protection appropriate to the sensitivity of the information being secured. Storage could be physical (for example secure cabinets) or technical (such as password management software), or a combination of both. The important thing is that your organisation provides a sanctioned mechanism to help users manage passwords, as this will deter users from adopting insecure \u2018hidden\u2019 methods to manage password overload.<\/p>\n<\/blockquote><p>Have a think about the environment you work in at present - do they have password strength criteria? Do they possibly have annual training or posters on the wall, each encouraging unique and complex passwords? Inevitably there's at least <em>some<\/em> control and education around this, but do they actually provide you with a \"sanctioned mechanism\" to help you achieve those objectives? Almost certainly not and I know this because it's one of the questions I often ask when I run training in organisations. It's amazing that this gap prevails.<\/p><h2 id=\"letthempastepasswords\">Let Them Paste Passwords<\/h2><p>Closely related to the use of password managers is the ability to paste passwords into the login screen. There are plenty of password managers that can auto-fill credentials, but there are occasions where either pasting is still necessary or where a service blocks a password that hasn't been typed in character by character (easily identified with a bit of JavaScript). It means we have situations like this:<\/p><p>Back in 2014, I wrote about <a href=\"https:\/\/www.troyhunt.com\/the-cobra-effect-that-is-disabling\/\">the \u201cCobra Effect\u201d that is disabling paste on password fields<\/a>. I explain the meaning of this term in the blog post but in short, a cobra effect occurs when an attempted solution to a problem makes the problem worse. When a website blocks the pasting of passwords in an attempt to improve security, they force some users to weaken their passwords to the point where they're dumbed down to easily typed versions.<\/p><p>The NCSC <a href=\"https:\/\/www.ncsc.gov.uk\/blog-post\/let-them-paste-passwords\">is behind me on this one too<\/a>:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/copy-paste-wall-code.jpg\" alt=\"Anti-password-pasting code on a wall\" \/><\/p><p>They even coined a term for the anti-pattern - SPP or \"Stop Pasting Passwords\" - and they go on to debunk common myths around the \"risks\" of pasting passwords. They also reference my cobra effect article mentioned earlier which is nice endorsement from the British Government!<\/p><p>NIST echoes the NCSC's position with this statement:<\/p><blockquote>\n<p>Verifiers SHOULD permit claimants to use \u201cpaste\u201d functionality when entering a memorized secret. This facilitates the use of password managers, which are widely used and in many cases increase the likelihood that users will choose stronger memorized secrets.<\/p>\n<\/blockquote><p>Let there be no doubt about it: there is unanimous support from both sides of the Atlantic for encouraging the use of password managers as well as not actively blocking them.<\/p><p>I had an easy way of remembering how long I'd been using a password for in my previous corporate job: I'd take the number in it that I incremented every time the company forced a quarterly password rotation and divide it by 4. I got about 6 years out of that particular password, only ever changing 1 or 2 characters at a time. If you're working in an environment that mandates regular password changes, you're very likely doing the same thing because it's an easy human control to deal with a technology requirement that's seen as an impediment.<\/p><p>Let's think through the rationale of this approach for a moment: the premise of a regular password change is that should that password be compromised, forcing a change means it is no longer valid, ergo it cannot be used by malicious parties. The problem is, attackers have got up to 3 months in the example I gave earlier or in some cases, even longer:<\/p><p>Think about it: attackers don't generally just sit around waiting, they get on with the business of exploiting accounts. This is the same flawed logic which led to <a href=\"https:\/\/www.troyhunt.com\/breach-concealment-is-not-a-security-strategy\/\">Lifeboat covering up their breach last year<\/a> based on the following screwy rationale:<\/p><blockquote>\n<p>When this happened [in] early January we figured the best thing for our players was to quietly force a password reset without letting the hackers know they had limited time to act<\/p>\n<\/blockquote><p>Per the tweet above, the NCSC agrees that forcibly rotating passwords is a modern-day security anti-pattern saying this about the practice in <a href=\"https:\/\/www.ncsc.gov.uk\/guidance\/password-guidance-simplifying-your-approach\">their password guidance documentation<\/a>:<\/p><blockquote>\n<p>carries no real benefits as stolen passwords are generally exploited immediately<\/p>\n<\/blockquote><p>And it makes its way into the associated infographic too:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Only-ask-users-to-change-their-passwords-on-indication.png\" alt=\"Only ask users to change their passwords on indication of suspicion of compromise\" \/><\/p><p>Microsoft is also on the same page here:<\/p><blockquote>\n<p>Password expiration policies do more harm than good, because these policies drive users to very predictable passwords composed of sequential words and numbers which are closely related to each other (that is, the next password can be predicted based on the previous password). Password change offers no containment benefits cyber criminals almost always use credentials as soon as they compromise them.<\/p>\n<\/blockquote><p>Now this is not to say \"don't force password cycling and do nothing else\", rather it's a recognition that there should be a broader, more <em>evolved<\/em> approach to password management. For example, the NCSC also recommends the following:<\/p><ol><li>Monitoring logins to detect unusual use<\/li>\n<\/ol><ul><li>Notifying users with details of attempted logins, successful or unsuccessful; they should report any for which they were not responsible<\/li>\n<\/ul><p>Which brings us neatly to the next section:<\/p><h2 id=\"notifyusersofabnormalbehaviour\">Notify Users of Abnormal Behaviour<\/h2><p>This is an important part of the \"evolved\" component of authentication and you may well have seen this in action before. I recently logged in to Yammer from my new Lenovo Yoga 910, a machine I'd not previously used with the service. Shortly afterwards, I received this notification:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Yammer-message.png\" alt=\"It looks like you recently logged in to Yammer from a new application or browser\" \/><\/p><p>I later logged into Dropbox via the browser with the same machine and immediately received this:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Dropbox-message.png\" alt=\"We noticed you logged into Dropbox using Chrome on Windows 10 at 03:07 PM GMT-08:00 from Queensland, Australia\" \/><\/p><p>Now neither of these <em>stopped<\/em> me from logging in and indeed I could have been a bad guy holding the victim's legitimate credentials. But they let me know as soon as it happened and that's enormously valuable because per the messages above, I could now go and boot an intruder out of my account if necessary. Dropbox actually showed me devices I'd auth'd as far back as 5 years ago and gave me the ability to de-auth them:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Unauth-Dropbox-device.png\" alt=\"Unauth Dropbox device\" \/><\/p><p>Being able to identify account behaviour in this fashion is enormously useful, for example seeing which devices are presently logged in to my Facebook account:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Current-devices-logged-in-to-Facebook.png\" alt=\"Current devices logged in to Facebook\" \/><\/p><p>As with Dropbox and Yammer, there's also the ability to now log any of these out which means it gives the rightful account owner back some control in the event that their account has been inappropriately accessed. Many modern services do this; GitHub is a great example and they go into a lot more detail including the IP address and the specific security event, for example a 2FA request or the creation of a public key.<\/p><p>Here's another scenario:<\/p><p>This too, makes a lot of sense and when you start considering the various scenarios you could proactively notify users about, you start to realise how much you can do with really very little effort.<\/p><h2 id=\"blockpreviouslybreachedpasswords\">Block Previously Breached Passwords<\/h2><p>Getting back to the whole credential stuffing thing for a moment, once passwords are disclosed they must be considered \"burned\", that is they should never be used again. Ever. Once they're out there in the wild, an untold number of other parties now have those credentials which therefore <em>significantly<\/em> heightens the risk anyone uses them now faces. Imagine having access to a <em>billion<\/em> email address and password pairs taken from actual data breaches <a href=\"https:\/\/www.troyhunt.com\/password-reuse-credential-stuffing-and-another-1-billion-records-in-have-i-been-pwned\/\">as I highlighted in the credential stuffing post<\/a>:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/05\/Combos-in-list-1.png\" alt=\"Combo List\" \/><\/p><p>NIST talks about the problem as follows:<\/p><blockquote>\n<p>When processing requests to establish and change memorized secrets, verifiers SHALL compare the prospective secrets against a list that contains values known to be commonly-used, expected, or compromised. For example, the list MAY include, but is not limited to: Passwords obtained from previous breach corpuses.<\/p>\n<\/blockquote><p>In layman's terms, this means that when someone registers or changes their password, you should be checking to ensure it's not a password that's previously appeared in a data breach. It doesn't matter that it may not have been the user who is presently registering that used the password in the breach, the mere fact that it has now been leaked publicly increases the chances of it being used in an attack. They also mention that the password shouldn't be a dictionary word or a \"context-specific word\"; when I wrote about <a href=\"https:\/\/www.troyhunt.com\/data-from-connected-cloudpets-teddy-bears-leaked-and-ransomed-exposing-kids-voice-messages\/\">CloudPets leaving their database publicly exposed<\/a>, I pointed out how even bcrypt hashes were easily crackable by using a small password dictionary including words such as \"cloudpets\". Don't let people use a password which is the name of the service they're signing up to <em>because they will!<\/em><\/p><h2 id=\"beyondthebasics\">Beyond the Basics<\/h2><p>There's a lot more I consciously decided not to delve into here. For example, the various multi-step verification mechanisms which are available and indeed which NIST speaks to in their documentation. I wanted to focus on the absolute fundamentals in this post but certainly layering additional defences such as an authenticator app on top of everything above greatly improves the security position, it's just a shame barely anyone uses them:<\/p><p>Another consideration is the ability to actually identify and defend against a highly automated attack in the first place. Think about it - how well-equipped are you to identify a sudden influx of login attempts such as via a credential stuffing attack? Would it be a matter of waiting until a whole bunch of people reported their accounts pwned or would you know as soon as it started happening? And if you <em>did<\/em> identify an attack, could you rapidly deploy defences to mitigate that risk? For example, by toggling the security level on Cloudflare:<\/p><p><img src=\"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Cloudflare-under-attack.png\" alt=\"Cloudflare under attack\" \/><\/p><p>I also didn't delve into password <em>storage<\/em>, instead deciding to focus on the more immediately visible components of how websites deal with credentials. So I don't leave that totally untouched, check out <a href=\"https:\/\/www.owasp.org\/index.php\/Password_Storage_Cheat_Sheet\">OWASP's Password Storage Cheat Sheet<\/a> for guidance there. Also have a good read of <a href=\"https:\/\/blogs.dropbox.com\/tech\/2016\/09\/how-dropbox-securely-stores-your-passwords\/\">how Dropbox securely stores your passwords<\/a> which is a very interesting piece combining both a modern-day approach to hashing and the use of encryption.<\/p><p>It goes much deeper than just what I've covered here, the point is that what I've outlined above should be the modern \"normal\", but it's not necessarily where you stop.<\/p><h2 id=\"summary\">Summary<\/h2><p>Here's the bigger picture of what all this guidance from governments and tech companies alike is recognising: security is increasingly about a composition of controls which when combined, improve the overall security posture of a service. What you'll see across this post is a collection of recommendations which all help contribute to a more robust solution by virtue of complementing one another. That may mean that individual recommendations such as dropping complexity requirements look odd, but when you consider the way humans tended to deal with that (they'd just choose bad passwords with a combination of character types) alongside guidance such as blocking previously breached passwords, things start to make a lot more sense.<\/p><p>Now there's just one more thing: as good as all this guidance is, practically implementing it can be somewhat trickier. I'm going to follow this post up with another one next week that will introduce something all new; something that will make it easier for site owners to protect their subscribers. It's something I've invested quite a bit of effort in over the last few weeks and I'm going to give it away for free so check back soon and I'll also update this post with a link to the resource as soon as it goes public. Stay tuned!<\/p><p><strong>Edit:<\/strong> As promised, here's that new resource: <a href=\"https:\/\/www.troyhunt.com\/introducing-306-million-freely-downloadable-pwned-passwords\/\">Introducing 306 Million Freely Downloadable Pwned Passwords<\/a><\/p><p><a class=\"tag\" href=\"https:\/\/www.troyhunt.com\/tag\/security\/\">Security<\/a>&#13;\n                <a class=\"tag\" href=\"https:\/\/www.troyhunt.com\/tag\/passwords\/\">Passwords<\/a>&#13;<\/p>","created_at":"2018-11-07T16:52:36+0000","updated_at":"2018-11-07T16:52:36+0000","published_at":"2017-07-26T07:40:03+0000","published_by":["Troy Hunt"],"annotations":[],"mimetype":"text\/html","reading_time":22,"domain_name":"www.troyhunt.com","preview_picture":"https:\/\/www.troyhunt.com\/content\/images\/2017\/07\/Massachusetts-Institute-of-Technology.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":586,"title":"Functors & Categories","url":"https:\/\/medium.com\/javascript-scene\/functors-categories-61e031bac53f","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"f2b0\" class=\"graf graf--h3 graf--leading graf--title\">Functors &amp; Categories<\/h1><h2 id=\"17ac\" class=\"graf graf--h4 graf-after--h3 graf--subtitle\">Composing Software<\/h2><figure id=\"70f5\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*uVpU7iruzXafhU2VLeH4lw.jpeg\" data-width=\"5151\" data-height=\"2346\" data-action=\"zoom\" data-action-value=\"1*uVpU7iruzXafhU2VLeH4lw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*uVpU7iruzXafhU2VLeH4lw.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Smoke Art Cubes to Smoke\u200a\u2014\u200aMattysFlicks\u200a\u2014\u200a(CC BY\u00a02.0)<\/figcaption><\/div><\/figure><blockquote id=\"dd42\" class=\"graf graf--blockquote graf-after--figure\"><div>Note: This is part of the \u201cComposing Software\u201d series on learning functional programming and compositional software techniques in JavaScript ES6+ from the ground up. Stay tuned. There\u2019s a lot more of this to come!<br \/><a href=\"https:\/\/medium.com\/javascript-scene\/reduce-composing-software-fe22f0c39a1d#.w4y0mlpcs\" data-href=\"https:\/\/medium.com\/javascript-scene\/reduce-composing-software-fe22f0c39a1d#.w4y0mlpcs\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">&lt; Previous<\/a> | <a href=\"https:\/\/medium.com\/javascript-scene\/the-rise-and-fall-and-rise-of-functional-programming-composable-software-c2d91b424c8c#.2dfd6n6qe\" data-href=\"https:\/\/medium.com\/javascript-scene\/the-rise-and-fall-and-rise-of-functional-programming-composable-software-c2d91b424c8c#.2dfd6n6qe\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">&lt;&lt; Start over at Part 1<\/a> | <a href=\"https:\/\/medium.com\/javascript-scene\/functional-mixins-composing-software-ffb66d5e731c\" data-href=\"https:\/\/medium.com\/javascript-scene\/functional-mixins-composing-software-ffb66d5e731c\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">Next &gt;&gt;<\/a><\/div><\/blockquote><p id=\"e41d\" class=\"graf graf--p graf-after--blockquote\">A <strong class=\"markup--strong markup--p-strong\">functor data type<\/strong> is something you can map over. It\u2019s a container which has an interface which can be used to apply a function to the values inside it. When you see a functor, you should think <em class=\"markup--em markup--p-em\">\u201cmappable\u201d.<\/em> Functor types are typically represented as an object with a\u00a0<code class=\"markup--code markup--p-code\">.map()<\/code> method that maps from inputs to outputs while preserving structure. In practice, \u201cpreserving structure\u201d means that the return value is the same type of functor (though values inside the container may be a different type).<\/p><p id=\"e048\" class=\"graf graf--p graf-after--p\">A functor supplies a box with zero or more things inside, and a mapping interface. An array is a good example of a functor, but many other kinds of objects can be mapped over as well, including promises, streams, trees, objects, etc\u2026 JavaScript\u2019s built in array and promise objects act like functors. For collections (arrays, streams, etc\u2026),\u00a0<code class=\"markup--code markup--p-code\">.map()<\/code> typically iterates over the collection and applies the given function to each value in the collection, but not all functors iterate. Functors are really about applying a function in a specific context.<\/p><p id=\"b67c\" class=\"graf graf--p graf-after--p\">Promises use the name\u00a0<code class=\"markup--code markup--p-code\">.then()<\/code> instead of\u00a0<code class=\"markup--code markup--p-code\">.map()<\/code>. You can usually think of\u00a0<code class=\"markup--code markup--p-code\">.then()<\/code> as an asynchronous\u00a0<code class=\"markup--code markup--p-code\">.map()<\/code>method, except when you have a nested promise, in which case it automatically unwraps the outer promise. Again, for values which are not promises,\u00a0<code class=\"markup--code markup--p-code\">.then()<\/code> acts like an asynchronous\u00a0<code class=\"markup--code markup--p-code\">.map()<\/code>. For values which are promises themselves,\u00a0<code class=\"markup--code markup--p-code\">.then()<\/code>acts like the\u00a0<code class=\"markup--code markup--p-code\">.chain()<\/code> method from monads (sometimes also called\u00a0<code class=\"markup--code markup--p-code\">.bind()<\/code> or\u00a0<code class=\"markup--code markup--p-code\">.flatMap()<\/code>). So, promises are not quite functors, and not quite monads, but in practice, you can usually treat them as either. Don't worry about what monads are, yet. Monads are a kind of functor, so you need to learn functors first.<\/p><p id=\"4175\" class=\"graf graf--p graf-after--p\">Lots of libraries exist that will turn a variety of other things into functors, too.<\/p><p id=\"8cb5\" class=\"graf graf--p graf-after--p\">In Haskell, the functor type is defined as:<\/p><pre id=\"3588\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">fmap :: (a -&gt; b) -&gt; f a -&gt; f b<\/code><\/pre><p id=\"4acf\" class=\"graf graf--p graf-after--pre\">Given a function that takes an <code class=\"markup--code markup--p-code\">a<\/code> and returns a <code class=\"markup--code markup--p-code\">b<\/code> and a functor with zero or more <code class=\"markup--code markup--p-code\">a<\/code>s inside it: <code class=\"markup--code markup--p-code\">fmap<\/code> returns a box with zero or more <code class=\"markup--code markup--p-code\">b<\/code>s inside it. The <code class=\"markup--code markup--p-code\">f a<\/code> and <code class=\"markup--code markup--p-code\">f b<\/code> bits can be read as \u201ca functor of <code class=\"markup--code markup--p-code\">a<\/code>\u201d and \u201ca functor of <code class=\"markup--code markup--p-code\">b<\/code>\u201d, meaning <code class=\"markup--code markup--p-code\">f a<\/code> has <code class=\"markup--code markup--p-code\">a<\/code>s inside the box, and <code class=\"markup--code markup--p-code\">f b<\/code> has <code class=\"markup--code markup--p-code\">b<\/code>s inside the box.<\/p><p id=\"abf1\" class=\"graf graf--p graf-after--p\">Using a functor is easy\u200a\u2014\u200ajust call <code class=\"markup--code markup--p-code\">map()<\/code>:<\/p><pre id=\"69d0\" class=\"graf graf--pre graf-after--p\">const f = [1, 2, 3];<br \/>f.map(double); \/\/ [2, 4, 6]<\/pre><h3 id=\"8b38\" class=\"graf graf--h3 graf-after--pre\">Functor Laws<\/h3><p id=\"c712\" class=\"graf graf--p graf-after--h3\">Categories have two important properties:<\/p><ol class=\"postList\"><li id=\"0caa\" class=\"graf graf--li graf-after--p\">Identity<\/li><li id=\"41a0\" class=\"graf graf--li graf-after--li\">Composition<\/li><\/ol><p id=\"2b3a\" class=\"graf graf--p graf-after--li\">Since a functor is a mapping between categories, functors must respect identity and composition. Together, they\u2019re known as the functor laws.<\/p><h3 id=\"faa7\" class=\"graf graf--h3 graf-after--p\">Identity<\/h3><p id=\"6d64\" class=\"graf graf--p graf-after--h3\">If you pass the identity function (<code class=\"markup--code markup--p-code\">x =&gt; x<\/code>) into <code class=\"markup--code markup--p-code\">f.map()<\/code>, where <code class=\"markup--code markup--p-code\">f<\/code> is any functor, the result should be equivalent to (have the same meaning as) <code class=\"markup--code markup--p-code\">f<\/code>:<\/p><pre id=\"0ab2\" class=\"graf graf--pre graf-after--p\">const f = [1, 2, 3];<br \/>f.map(x =&gt; x); \/\/ [1, 2, 3]<\/pre><h3 id=\"e36d\" class=\"graf graf--h3 graf-after--pre\">Composition<\/h3><p id=\"8b7e\" class=\"graf graf--p graf-after--h3\">Functors must obey the composition law: <code class=\"markup--code markup--p-code\">F.map(x =&gt; f(g(x)))<\/code> is equivalent to <code class=\"markup--code markup--p-code\">F.map(g).map(f)<\/code>.<\/p><p id=\"a707\" class=\"graf graf--p graf-after--p\">Function Composition is the application of one function to the result of another, e.g., given an <code class=\"markup--code markup--p-code\">x<\/code> and the functions, <code class=\"markup--code markup--p-code\">f<\/code> and <code class=\"markup--code markup--p-code\">g<\/code>, the composition <code class=\"markup--code markup--p-code\">(f \u2218 g)(x)<\/code> (usually shortened to <code class=\"markup--code markup--p-code\">f \u2218 g<\/code> - the <code class=\"markup--code markup--p-code\">(x)<\/code> is implied) means <code class=\"markup--code markup--p-code\">f(g(x))<\/code>.<\/p><p id=\"d19d\" class=\"graf graf--p graf-after--p\">A lot of functional programming terms come from category theory, and the essence of category theory is composition. Category theory is scary at first, but easy. Like jumping off a diving board or riding a roller coaster. Here\u2019s the foundation of category theory in a few bullet points:<\/p><ul class=\"postList\"><li id=\"17c3\" class=\"graf graf--li graf-after--p\">A category is a collection of objects and arrows between objects (where \u201cobject\u201d can mean literally anything).<\/li><li id=\"a60c\" class=\"graf graf--li graf-after--li\">Arrows are known as morphisms. Morphisms can be thought of and represented in code as functions.<\/li><li id=\"64b8\" class=\"graf graf--li graf-after--li\">For any group of connected objects, <code class=\"markup--code markup--li-code\">a -&gt; b -&gt; c<\/code>, there must be a composition which goes directly from <code class=\"markup--code markup--li-code\">a -&gt; c<\/code>.<\/li><li id=\"3fd2\" class=\"graf graf--li graf-after--li\">All arrows can be represented as compositions (even if it\u2019s just a composition with the object\u2019s identity arrow). All objects in a category have identity arrows.<\/li><\/ul><p id=\"978b\" class=\"graf graf--p graf-after--li\">Say you have a function <code class=\"markup--code markup--p-code\">g<\/code> that takes an <code class=\"markup--code markup--p-code\">a<\/code> and returns a <code class=\"markup--code markup--p-code\">b<\/code>, and another function <code class=\"markup--code markup--p-code\">f<\/code> that takes a <code class=\"markup--code markup--p-code\">b<\/code> and returns a <code class=\"markup--code markup--p-code\">c<\/code>; there must also be a function <code class=\"markup--code markup--p-code\">h<\/code> that represents the composition of <code class=\"markup--code markup--p-code\">f<\/code> and <code class=\"markup--code markup--p-code\">g<\/code>. So, the composition from <code class=\"markup--code markup--p-code\">a -&gt; c<\/code>, is the composition <code class=\"markup--code markup--p-code\">f \u2218 g<\/code> (<code class=\"markup--code markup--p-code\">f<\/code> <em class=\"markup--em markup--p-em\">after<\/em> <code class=\"markup--code markup--p-code\">g<\/code>). So, <code class=\"markup--code markup--p-code\">h(x) = f(g(x))<\/code>. Function composition works right to left, not left to right, which is why <code class=\"markup--code markup--p-code\">f \u2218 g<\/code> is frequently called <code class=\"markup--code markup--p-code\">f<\/code> <em class=\"markup--em markup--p-em\">after<\/em> <code class=\"markup--code markup--p-code\">g<\/code>.<\/p><p id=\"344e\" class=\"graf graf--p graf-after--p\">Composition is associative. Basically that means that when you\u2019re composing multiple functions (morphisms if you\u2019re feeling fancy), you don\u2019t need parenthesis:<\/p><pre id=\"2cd3\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">h\u2218(g\u2218f) = (h\u2218g)\u2218f = h\u2218g\u2218f<\/code><\/pre><p id=\"5c46\" class=\"graf graf--p graf-after--pre\">Let\u2019s take another look at the composition law in JavaScript:<\/p><p id=\"56f9\" class=\"graf graf--p graf-after--p\">Given a functor, <code class=\"markup--code markup--p-code\">F<\/code>:<\/p><pre id=\"a8bd\" class=\"graf graf--pre graf-after--p\">const F = [1, 2, 3];<\/pre><p id=\"d42b\" class=\"graf graf--p graf-after--pre\">The following are equivalent:<\/p><pre id=\"1f86\" class=\"graf graf--pre graf-after--p\">F.map(x =&gt; f(g(x)));<\/pre><pre id=\"d2be\" class=\"graf graf--pre graf-after--pre\">\/\/ is equivalent to...<\/pre><pre id=\"746a\" class=\"graf graf--pre graf-after--pre\">F.map(g).map(f);<\/pre><h3 id=\"f642\" class=\"graf graf--h3 graf-after--pre\">Endofunctors<\/h3><p id=\"c775\" class=\"graf graf--p graf-after--h3\">An endofunctor is a functor that maps from a category back to the same category.<\/p><p id=\"a1a5\" class=\"graf graf--p graf-after--p\">A functor can map from category to category: <code class=\"markup--code markup--p-code\">X -&gt; Y<\/code><\/p><p id=\"f6e7\" class=\"graf graf--p graf-after--p\">An endofunctor maps from a category to the same category: <code class=\"markup--code markup--p-code\">X -&gt; X<\/code><\/p><p id=\"e73f\" class=\"graf graf--p graf-after--p\">A monad is an endofunctor. Remember:<\/p><blockquote id=\"e418\" class=\"graf graf--blockquote graf--startsWithDoubleQuote graf-after--p\"><div><em class=\"markup--em markup--blockquote-em\">\u201cA monad is just a monoid in the category of endofunctors. What\u2019s the problem?\u201d<\/em><\/div><\/blockquote><p id=\"ac4b\" class=\"graf graf--p graf-after--blockquote\">Hopefully that quote is starting to make a little more sense. We\u2019ll get to monoids and monads later.<\/p><h3 id=\"9ef0\" class=\"graf graf--h3 graf-after--p\">Build Your Own\u00a0Functor<\/h3><p id=\"5ce5\" class=\"graf graf--p graf-after--h3\">Here\u2019s a simple example of a functor:<\/p><pre id=\"81d6\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">const Identity = value =&gt; ({<br \/>  map: fn =&gt; Identity(fn(value))<br \/>});<\/code><\/pre><p id=\"5426\" class=\"graf graf--p graf-after--pre\">As you can see, it satisfies the functor laws:<\/p><pre id=\"289f\" class=\"graf graf--pre graf-after--p\">\/\/ trace() is a utility to let you easily inspect<br \/>\/\/ the contents.<br \/>const trace = x =&gt; {<br \/>  console.log(x);<br \/>  return x;<br \/>};<\/pre><pre id=\"8190\" class=\"graf graf--pre graf-after--pre\">const u = Identity(2);<\/pre><pre id=\"9c29\" class=\"graf graf--pre graf-after--pre\">\/\/ Identity law<br \/>u.map(trace);             \/\/ 2<br \/>u.map(x =&gt; x).map(trace); \/\/ 2<\/pre><pre id=\"d9a9\" class=\"graf graf--pre graf-after--pre\">const f = n =&gt; n + 1;<br \/>const g = n =&gt; n * 2;<\/pre><pre id=\"b7e1\" class=\"graf graf--pre graf-after--pre\">\/\/ Composition law<br \/>const r1 = u.map(x =&gt; f(g(x)));<br \/>const r2 = u.map(g).map(f);<\/pre><pre id=\"7d11\" class=\"graf graf--pre graf-after--pre\">r1.map(trace); \/\/ 5<br \/>r2.map(trace); \/\/ 5<\/pre><p id=\"3f1c\" class=\"graf graf--p graf-after--pre\">Now you can map over any data type, just like you can map over an array. Nice!<\/p><p id=\"01f0\" class=\"graf graf--p graf-after--p\">That\u2019s about as simple as a functor can get in JavaScript, but it\u2019s missing some features we expect from data types in JavaScript. Let\u2019s add them. Wouldn\u2019t it be cool if the <code class=\"markup--code markup--p-code\">+<\/code> operator could work for number and string values?<\/p><p id=\"6da0\" class=\"graf graf--p graf-after--p\">To make that work, all we need to do is implement\u00a0<code class=\"markup--code markup--p-code\">.valueOf()<\/code> -- which also seems like a convenient way to unwrap the value from the functor:<\/p><pre id=\"2f47\" class=\"graf graf--pre graf-after--p\">const Identity = value =&gt; ({<br \/>  map: fn =&gt; Identity(fn(value)),<\/pre><pre id=\"b71e\" class=\"graf graf--pre graf-after--pre\">  valueOf: () =&gt; value,<br \/>});<\/pre><pre id=\"7580\" class=\"graf graf--pre graf-after--pre\">const ints = (Identity(2) + Identity(4));<br \/>trace(ints); \/\/ 6<\/pre><pre id=\"a97d\" class=\"graf graf--pre graf-after--pre\">const hi = (Identity('h') + Identity('i'));<br \/>trace(hi); \/\/ \"hi\"<\/pre><p id=\"147e\" class=\"graf graf--p graf-after--pre\">Nice. But what if we want to inspect an <code class=\"markup--code markup--p-code\">Identity<\/code> instance in the console? It would be cool if it would say <code class=\"markup--code markup--p-code\">\"Identity(value)\"<\/code>, right. Let's add a\u00a0<code class=\"markup--code markup--p-code\">.toString()<\/code> method:<\/p><pre id=\"4012\" class=\"graf graf--pre graf-after--p\">toString: () =&gt; `Identity(${value})`,<\/pre><p id=\"8dd4\" class=\"graf graf--p graf-after--pre\">Cool. We should probably also enable the standard JS iteration protocol. We can do that by adding a custom iterator:<\/p><pre id=\"27b7\" class=\"graf graf--pre graf-after--p\">[Symbol.iterator]: function* () {<br \/>  yield value;<br \/>}<\/pre><p id=\"d79a\" class=\"graf graf--p graf-after--pre\">Now this will work:<\/p><pre id=\"b194\" class=\"graf graf--pre graf-after--p\">\/\/ [Symbol.iterator] enables standard JS iterations:<br \/>const arr = [6, 7, ...Identity(8)];<br \/>trace(arr); \/\/ [6, 7, 8]<\/pre><p id=\"83b2\" class=\"graf graf--p graf-after--pre\">What if you want to take an <code class=\"markup--code markup--p-code\">Identity(n)<\/code> and return an array of Identities containing <code class=\"markup--code markup--p-code\">n + 1<\/code>, <code class=\"markup--code markup--p-code\">n + 2<\/code>, and so on? Easy, right?<\/p><pre id=\"6623\" class=\"graf graf--pre graf-after--p\">const fRange = (<br \/>  start,<br \/>  end<br \/>) =&gt; Array.from(<br \/>  { length: end - start + 1 },<br \/>  (x, i) =&gt; Identity(i + start)<br \/>);<\/pre><p id=\"1324\" class=\"graf graf--p graf-after--pre\">Ah, but what if you want this to work with any functor? What if we had a spec that said that each instance of a data type must have a reference to its constructor? Then you could do this:<\/p><pre id=\"2e02\" class=\"graf graf--pre graf-after--p\">const fRange = (<br \/>  start,<br \/>  end<br \/>) =&gt; Array.from(<br \/>  { length: end - start + 1 },\/\/ change `Identity` to `start.constructor`<br \/>  (x, i) =&gt; start.constructor(i + start)<br \/>);<\/pre><pre id=\"9985\" class=\"graf graf--pre graf-after--pre\">const range = fRange(Identity(2), 4);<br \/>range.map(x =&gt; x.map(trace)); \/\/ 2, 3, 4<\/pre><p id=\"1e5f\" class=\"graf graf--p graf-after--pre\">What if you want to test to see if a value is a functor? We could add a static method on <code class=\"markup--code markup--p-code\">Identity<\/code> to check. We should throw in a static\u00a0<code class=\"markup--code markup--p-code\">.toString()<\/code> while we're at it:<\/p><pre id=\"882b\" class=\"graf graf--pre graf-after--p\">Object.assign(Identity, {<br \/>  toString: () =&gt; 'Identity',<br \/>  is: x =&gt; typeof x.map === 'function'<br \/>});<\/pre><p id=\"d92d\" class=\"graf graf--p graf-after--pre\">Let\u2019s put all this together:<\/p><pre id=\"b14d\" class=\"graf graf--pre graf-after--p\">const Identity = value =&gt; ({<br \/>  map: fn =&gt; Identity(fn(value)),<br \/>  valueOf: () =&gt; value,<br \/>  toString: () =&gt; `Identity(${value})`,<br \/>  [Symbol.iterator]: function* () {<br \/>    yield value;<br \/>  },<br \/>  constructor: Identity<br \/>});<\/pre><pre id=\"15be\" class=\"graf graf--pre graf-after--pre\">Object.assign(Identity, {<br \/>  toString: () =&gt; 'Identity',<br \/>  is: x =&gt; typeof x.map === 'function'<br \/>});<\/pre><p id=\"d9db\" class=\"graf graf--p graf-after--pre\">Note you don\u2019t need all this extra stuff for something to qualify as a functor or an endofunctor. It\u2019s strictly for convenience. All you <em class=\"markup--em markup--p-em\">need<\/em> for a functor is a\u00a0<code class=\"markup--code markup--p-code\">.map()<\/code> interface that satisfies the functor laws.<\/p><h3 id=\"1dac\" class=\"graf graf--h3 graf-after--p\">Why Functors?<\/h3><p id=\"075c\" class=\"graf graf--p graf-after--h3\">Functors are great for lots of reasons. Most importantly, they\u2019re an abstraction that you can use to implement lots of useful things in a way that works with any data type. For instance, what if you want to kick off a chain of operations, but only if the value inside the functor is not <code class=\"markup--code markup--p-code\">undefined<\/code> or <code class=\"markup--code markup--p-code\">null<\/code>?<\/p><pre id=\"d103\" class=\"graf graf--pre graf-after--p\">\/\/ Create the predicate<br \/>const exists = x =&gt; (x.valueOf() !== undefined &amp;&amp; x.valueOf() !== null);<\/pre><pre id=\"8e78\" class=\"graf graf--pre graf-after--pre\">const ifExists = x =&gt; ({<br \/>  map: fn =&gt; exists(x) ? x.map(fn) : x<br \/>});<\/pre><pre id=\"a82a\" class=\"graf graf--pre graf-after--pre\">const add1 = n =&gt; n + 1;<br \/>const double = n =&gt; n * 2;<\/pre><pre id=\"d51e\" class=\"graf graf--pre graf-after--pre\">\/\/ Nothing happens...<br \/>ifExists(Identity(undefined)).map(trace);<br \/>\/\/ Still nothing...<br \/>ifExists(Identity(null)).map(trace);<\/pre><pre id=\"71d4\" class=\"graf graf--pre graf-after--pre\">\/\/ 42<br \/>ifExists(Identity(20))<br \/>  .map(add1)<br \/>  .map(double)<br \/>  .map(trace)<br \/>;<\/pre><p id=\"2711\" class=\"graf graf--p graf-after--pre\">Of course, functional programming is all about composing tiny functions to create higher level abstractions. What if you want a generic map that works with any functor? That way you can partially apply arguments to create new functions.<\/p><p id=\"f53d\" class=\"graf graf--p graf-after--p\">Easy. Pick your favorite auto-curry, or use this magic spell from before:<\/p><pre id=\"cfdc\" class=\"graf graf--pre graf-after--p\">const curry = (<br \/>  f, arr = []<br \/>) =&gt; (...args) =&gt; (<br \/>  a =&gt; a.length === f.length ?<br \/>    f(...a) :<br \/>    curry(f, a)<br \/>)([...arr, ...args]);<\/pre><p id=\"12f0\" class=\"graf graf--p graf-after--pre\">Now we can customize map:<\/p><pre id=\"a2ac\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">const map = curry((fn, F) =&gt; F.map(fn));<\/code><\/pre><pre id=\"94aa\" class=\"graf graf--pre graf-after--pre\"><code class=\"markup--code markup--pre-code\">const double = n =&gt; n * 2;<\/code><\/pre><pre id=\"66eb\" class=\"graf graf--pre graf-after--pre\"><code class=\"markup--code markup--pre-code\">const mdouble = map(double);<br \/>mdouble(Identity(4)).map(trace); \/\/ 8<\/code><\/pre><h3 id=\"58f8\" class=\"graf graf--h3 graf-after--pre\">Conclusion<\/h3><p id=\"2ac7\" class=\"graf graf--p graf-after--h3\">Functors are things we can map over. More specifically, a functor is a mapping from category to category. A functor can even map from a category back to the same category (i.e., an <em class=\"markup--em markup--p-em\">endofunctor<\/em>).<\/p><p id=\"2727\" class=\"graf graf--p graf-after--p\">A category is a collection of objects, with arrows between objects. Arrows represent morphisms (aka functions, aka compositions). Each object in a category has an identity morphism (<code class=\"markup--code markup--p-code\">x =&gt; x<\/code>). For any chain of objects <code class=\"markup--code markup--p-code\">A -&gt; B -&gt; C<\/code> there must exist a composition <code class=\"markup--code markup--p-code\">A -&gt; C<\/code>.<\/p><p id=\"dc41\" class=\"graf graf--p graf-after--p\">Functors are great higher-order abstractions that allow you to create a variety of generic functions that will work for any data type.<\/p><p id=\"06d3\" class=\"graf graf--p graf-after--p\"><a href=\"https:\/\/medium.com\/javascript-scene\/functional-mixins-composing-software-ffb66d5e731c\" data-href=\"https:\/\/medium.com\/javascript-scene\/functional-mixins-composing-software-ffb66d5e731c\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Next: Functional Mixins &gt;<\/strong><\/a><\/p><h3 id=\"fd11\" class=\"graf graf--h3 graf-after--p\">Level Up Your Skills with Live 1:1 Mentorship<\/h3><p id=\"2db8\" class=\"graf graf--p graf-after--h3\">DevAnywhere is the fastest way to level up to advanced JavaScript skills:<\/p><ul class=\"postList\"><li id=\"5aa2\" class=\"graf graf--li graf-after--p\">Live lessons<\/li><li id=\"58c2\" class=\"graf graf--li graf-after--li\">Flexible hours<\/li><li id=\"b984\" class=\"graf graf--li graf-after--li\">1:1 mentorship<\/li><li id=\"cdc0\" class=\"graf graf--li graf-after--li\">Build real production apps<\/li><\/ul><figure id=\"154a\" class=\"graf graf--figure graf-after--li graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"https:\/\/devanywhere.io\/\" data-href=\"https:\/\/devanywhere.io\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"nofollow\"><img class=\"graf-image\" data-image-id=\"1*pskrI-ZjRX_Y0I0zZqVTcQ.png\" data-width=\"2470\" data-height=\"1390\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*pskrI-ZjRX_Y0I0zZqVTcQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/devanywhere.io\/\" data-href=\"https:\/\/devanywhere.io\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener noopener nofollow noopener nofollow noopener noopener noopener nofollow noopener nofollow noopener\" target=\"_blank\">https:\/\/devanywhere.io\/<\/a><\/figcaption><\/div><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"b215\" class=\"graf graf--p graf--leading\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Eric Elliott<\/em><\/strong><em class=\"markup--em markup--p-em\"> is the author of <\/em><a href=\"http:\/\/pjabook.com\" data-href=\"http:\/\/pjabook.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener noopener noopener nofollow noopener nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">\u201cProgramming JavaScript Applications\u201d<\/em><\/a><em class=\"markup--em markup--p-em\"> (O\u2019Reilly), and cofounder of <\/em><a href=\"https:\/\/devanywhere.io\/\" data-href=\"https:\/\/devanywhere.io\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">DevAnywhere.io<\/em><\/a><em class=\"markup--em markup--p-em\">. He has contributed to software experiences for <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Adobe Systems<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Zumba Fitness<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">The Wall Street Journal<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">ESPN<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">BBC<\/em><\/strong><em class=\"markup--em markup--p-em\">, and top recording artists including <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Usher<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Frank Ocean<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Metallica<\/em><\/strong><em class=\"markup--em markup--p-em\">, and many more.<\/em><\/p><p class=\"graf graf--p graf-after--p graf--trailing\"><em class=\"markup--em markup--p-em\">He works anywhere he wants with the most beautiful woman in the world.<\/em><\/p><\/div><\/div><\/section>","created_at":"2018-11-07T16:52:35+0000","updated_at":"2018-11-07T16:52:35+0000","published_at":"2017-03-17T06:10:29+0000","published_by":["Eric Elliott"],"annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*uVpU7iruzXafhU2VLeH4lw.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":585,"title":"Growth Doesn\u2019t Start With Experiments.","url":"https:\/\/thinkgrowth.org\/growth-doesnt-start-with-experiments-63bb4b117bd2?gi=12ec2965bee1","content":"<figure id=\"e362\" class=\"graf graf--figure graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*3BivqdaopgSoguk8KiH8PA.jpeg\" data-width=\"1400\" data-height=\"800\" data-action=\"zoom\" data-action-value=\"1*3BivqdaopgSoguk8KiH8PA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*3BivqdaopgSoguk8KiH8PA.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"d8f4\" class=\"graf graf--p graf-after--h4\">Spend an hour on <a href=\"https:\/\/growthhackers.com\/growth-studies\" data-href=\"https:\/\/growthhackers.com\/growth-studies\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Growth Hackers<\/a> and you\u2019ll find lots and lots of comments, tactics, and studies about companies that grow. Take HubSpot, for example:<\/p><blockquote id=\"8d19\" class=\"graf graf--blockquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cHubSpot uses webinars to educate users, generate buzz, and attract thousands of new website visitors\u200a\u2014\u200aa number of whom will eventually become customers.\u201d [<a href=\"https:\/\/growthhackers.com\/growth-studies\/hubspot\" data-href=\"https:\/\/growthhackers.com\/growth-studies\/hubspot\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener\" target=\"_blank\">Source<\/a>]<\/div><\/blockquote><p id=\"9867\" class=\"graf graf--p graf-after--blockquote\">Webinars worked for HubSpot, so they should work for you too, right? Webinars <em class=\"markup--em markup--p-em\">work.<\/em><\/p><p id=\"020c\" class=\"graf graf--p graf-after--p\">It\u2019s easy to draw ideas from a bucket of things that other companies did. Unfortunately for you, other companies are not your company. Their teams made decisions in the context of <em class=\"markup--em markup--p-em\">their<\/em> company and not yours. You\u2019ll quickly understand this as you watch your exact reproductions return mediocre results.<\/p><blockquote id=\"5f59\" class=\"graf graf--pullquote graf-after--p\"><div>Growth doesn\u2019t come from reproducing the experiments of successful companies, it comes from understanding relationships.<\/div><\/blockquote><p id=\"3156\" class=\"graf graf--p graf-after--pullquote\">In my experience, my most impactful learnings emerged from testing the relationships between a <strong class=\"markup--strong markup--p-strong\">segment, <\/strong>an<strong class=\"markup--strong markup--p-strong\"> action<\/strong>, and a<strong class=\"markup--strong markup--p-strong\"> theme. <\/strong>Those learnings turned into bets the larger business could run with. Once you begin to frame questions in this context, every experiment you run and all research you conduct will have targeted business impact.<\/p><figure id=\"9a87\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*GBBc9irIH6-87bmkD2hO0g.jpeg\" data-width=\"1400\" data-height=\"800\" data-action=\"zoom\" data-action-value=\"1*GBBc9irIH6-87bmkD2hO0g.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*GBBc9irIH6-87bmkD2hO0g.jpeg\" alt=\"image\" \/><\/div><\/figure><h4 id=\"13ad\" class=\"graf graf--h4 graf-after--figure\">Segment<\/h4><p id=\"58b2\" class=\"graf graf--p graf-after--h4\">A group of customers or users who share a similar set of properties. One segment could be <em class=\"markup--em markup--p-em\">customers who work in construction<\/em>. Another segment could be <em class=\"markup--em markup--p-em\">customers who have paid you more than $60.<\/em> If you desire, you can add more traits to the list to create an even more focused segment (like <em class=\"markup--em markup--p-em\">customers who have been paying for more than 6 months but not more than 12 months and who show lower-than-average usage and haven\u2019t talked to the support team<\/em>).<\/p><h4 id=\"3217\" class=\"graf graf--h4 graf-after--p\">Action<\/h4><p id=\"6ff0\" class=\"graf graf--p graf-after--h4\">A simple, measurable behavior. This can be something like <em class=\"markup--em markup--p-em\">requesting a demo from the sales team <\/em>or perhaps <em class=\"markup--em markup--p-em\">signing up with a discount code. <\/em>Actions are often easier to track if you build a software product because you can get very granular with things like <em class=\"markup--em markup--p-em\">exploring additional pricing details <\/em>or <em class=\"markup--em markup--p-em\">creating their first marketing campaign.<\/em><\/p><h4 id=\"409d\" class=\"graf graf--h4 graf-after--p\">Theme<\/h4><p id=\"1dc6\" class=\"graf graf--p graf-after--h4\">This is the one that people often struggle with. A theme is a thing that you think is true about your customers in a general sense. Usually, people forget to generalize and they skip right to a specific example of the theme. Below are some examples of the difference.<\/p><figure id=\"38b2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*UHF4ZmzMpYC6ia6k1jTy9Q.jpeg\" data-width=\"1400\" data-height=\"800\" data-action=\"zoom\" data-action-value=\"1*UHF4ZmzMpYC6ia6k1jTy9Q.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*UHF4ZmzMpYC6ia6k1jTy9Q.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"2021\" class=\"graf graf--p graf-after--figure\">The power of these three things comes when they are combined into a single relationship statement:<\/p><blockquote id=\"42cf\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cHow can we test the influence of [THEME] on how often people [ACTION] for those people who [SEGMENT]?\u201d<\/div><\/blockquote><p id=\"a3c2\" class=\"graf graf--p graf-after--pullquote\">Starting to make sense? How about if we fill in the blanks:<\/p><blockquote id=\"73d9\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cHow can we test the influence of social pressure on how often people request a demo for those people who are first-time visitors?\u201d<\/div><\/blockquote><p id=\"4715\" class=\"graf graf--p graf-after--pullquote\">At this point, magic happens. There is no limit on what research you do or what testing tactics you use. You can now excuse experiment failures without making assumptions about why they failed. <strong class=\"markup--strong markup--p-strong\">You are beginning to draw a map of relationships that can be used now and in future projects.<\/strong><\/p><h3 id=\"00c6\" class=\"graf graf--h3 graf-after--p\">Themes\u200a\u2014\u200awhere do they come\u00a0from?<\/h3><p id=\"297e\" class=\"graf graf--p graf-after--h3\">As I mentioned above, themes are the most challenging part of this framework. They are a bit abstract and force you to think in terms of commonalities instead of specific details. Even I face challenges wrapping my head around themes from time to time.<\/p><p id=\"07f9\" class=\"graf graf--p graf-after--p\">In most businesses, there is a need to increase the frequency of certain things that customers do, whether it\u2019s paying money, leaving reviews, asking for help, creating content, or simply returning to your store to look around. Since you already have an <strong class=\"markup--strong markup--p-strong\">action <\/strong>and some form of a <strong class=\"markup--strong markup--p-strong\">segment<\/strong>, all that\u2019s missing is a <strong class=\"markup--strong markup--p-strong\">theme<\/strong>. So, where do themes live?<\/p><p id=\"749a\" class=\"graf graf--p graf-after--p\">A very common mistake is to ideate the themes from nowhere. Usually, a team will brainstorm for hours and generate a list of clever ideas based on bias or past experience. You\u2019ve probably heard this a thousand times, but there is a better way, and that\u2019s to <a href=\"https:\/\/www.inc.com\/steve-blank\/key-to-success-getting-out-of-building.html\" data-href=\"https:\/\/www.inc.com\/steve-blank\/key-to-success-getting-out-of-building.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">get out of the building<\/a> and talk to customers.<\/p><p id=\"9cc6\" class=\"graf graf--p graf-after--p\">Start talking to people from each segment who <strong class=\"markup--strong markup--p-strong\">didn\u2019t<\/strong> do something.<\/p><p id=\"891b\" class=\"graf graf--p graf-after--p\">Ask them why they didn\u2019t do that action. Ask them about their experience up to that point. Ask them a whole bunch of other stuff if you\u2019re curious (but stay organized and specific).<\/p><p id=\"0073\" class=\"graf graf--p graf-after--p\">Once you have all of your answers, group them by generalized stories:<\/p><figure id=\"b5c4\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*08-ZKiqtjhnG9Qbw.jpg\" data-width=\"800\" data-height=\"655\" data-action=\"zoom\" data-action-value=\"0*08-ZKiqtjhnG9Qbw.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*08-ZKiqtjhnG9Qbw.jpg\" alt=\"image\" \/><\/div><\/figure><p id=\"fcc4\" class=\"graf graf--p graf-after--figure\">On the data analytics side (if you have that stuff), you can compile a list of things that users <strong class=\"markup--strong markup--p-strong\">did<\/strong> do instead. This might give the user statements some weight or perhaps reveal strange patterns of use.<\/p><p id=\"2b15\" class=\"graf graf--p graf-after--p\">Now you have a list of things people said validated by a list of things people did. Together, these make up <strong class=\"markup--strong markup--p-strong\">themes<\/strong>.<\/p><h3 id=\"39b6\" class=\"graf graf--h3 graf-after--p\">Once you have all three, start brainstorming<\/h3><p id=\"c9a0\" class=\"graf graf--p graf-after--h3\">Pull together a group of coworkers who build for or talk with customers (including but not limited to engineering, marketing, sales, services, or support) and present them with a prompt:<\/p><figure id=\"bc54\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*KdGzhyu1JWg-N7zt.jpg\" data-width=\"800\" data-height=\"623\" data-action=\"zoom\" data-action-value=\"0*KdGzhyu1JWg-N7zt.jpg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*KdGzhyu1JWg-N7zt.jpg\" alt=\"image\" \/><\/div><\/figure><p id=\"1f2b\" class=\"graf graf--p graf-after--figure\">Sometimes there is only one way to test this. Sometimes there are hundreds of ways. You may find yourself getting through 5 or 10 different prompts during the brainstorm, but you\u2019ll end up with a hyper-focused list of experiment and research ideas.<\/p><p id=\"4224\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">(Want a more granular example of how this works from beginning to end? <\/em><a href=\"https:\/\/medium.com\/@geoffdaigle\/for-the-growth-experts-a-very-granular-example-ba8b8cfec143\" data-href=\"https:\/\/medium.com\/@geoffdaigle\/for-the-growth-experts-a-very-granular-example-ba8b8cfec143\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Check out the expert\u2019s document<\/em><\/a><em class=\"markup--em markup--p-em\">.)<\/em><\/p><h3 id=\"d534\" class=\"graf graf--h3 graf-after--p\">Now you\u2019re asking the right questions<\/h3><p id=\"35f3\" class=\"graf graf--p graf-after--h3\">One more time so you don\u2019t forget this part:<\/p><blockquote id=\"8884\" class=\"graf graf--pullquote graf-after--p\"><div>Your goal is to find ways to test the relationship between a segment<strong class=\"markup--strong markup--pullquote-strong\">, <\/strong>an<strong class=\"markup--strong markup--pullquote-strong\"> action<\/strong>, and a<strong class=\"markup--strong markup--pullquote-strong\">\u00a0theme.<\/strong><\/div><\/blockquote><p id=\"525b\" class=\"graf graf--p graf-after--pullquote\">Never again will you copy and paste the work of another successful company without the consideration of your funnels. All of your ideas are now coming from the reality of what your customers do and feel, backed up by the truths that you pulled from past research and experimentation.<\/p><p id=\"2cad\" class=\"graf graf--p graf-after--p\">Sometime in the future, when you have spent hours doing your best work, you\u2019ll find an email in your inbox from the CEO with a chart attached.<\/p><p id=\"1268\" class=\"graf graf--p graf-after--p graf--trailing\">That chart goes up and to the right.<\/p>","created_at":"2018-11-07T16:52:34+0000","updated_at":"2018-11-07T16:52:34+0000","published_at":"2017-08-14T14:31:00+0000","published_by":["Geoff Daigle"],"annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"thinkgrowth.org","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*3BivqdaopgSoguk8KiH8PA.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":584,"title":"That One Privacy Site | Detailed VPN Comparison Chart","url":"https:\/\/thatoneprivacysite.net\/vpn-comparison-chart\/","content":"<div class=\"wpdt-c\">\n    \n<input type=\"hidden\" id=\"wdtNonceFrontendEdit\" name=\"wdtNonceFrontendEdit\" value=\"d7c07984cf\" \/><input type=\"hidden\" name=\"_wp_http_referer\" value=\"\/vpn-comparison-chart\/\" \/><input type=\"hidden\" id=\"table_1_desc\" value=\"{&quot;tableId&quot;:&quot;table_1&quot;,&quot;tableType&quot;:&quot;manual&quot;,&quot;selector&quot;:&quot;#table_1&quot;,&quot;responsive&quot;:true,&quot;editable&quot;:false,&quot;inlineEditing&quot;:false,&quot;infoBlock&quot;:true,&quot;globalSearch&quot;:true,&quot;showRowsPerPage&quot;:true,&quot;popoverTools&quot;:false,&quot;hideBeforeLoad&quot;:true,&quot;number_format&quot;:2,&quot;decimalPlaces&quot;:2,&quot;spinnerSrc&quot;:&quot;https:\\\/\\\/thatoneprivacysite.net\\\/wp-content\\\/plugins\\\/wpdatatables\\\/assets\\\/\\\/img\\\/spinner.gif&quot;,&quot;groupingEnabled&quot;:false,&quot;tableWpId&quot;:&quot;17&quot;,&quot;dataTableParams&quot;:{&quot;sDom&quot;:&quot;BT\\u003C\\u0027clear\\u0027\\u003Elf\\u003C\\u0027wdtscroll\\u0027t\\u003Eip&quot;,&quot;bSortCellsTop&quot;:false,&quot;bFilter&quot;:true,&quot;bPaginate&quot;:true,&quot;aLengthMenu&quot;:[[1,5,10,25,50,100,-1],[1,5,10,25,50,100,&quot;All&quot;]],&quot;iDisplayLength&quot;:10,&quot;columnDefs&quot;:[{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:false,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;wdt_ID&quot;,&quot;origHeader&quot;:&quot;wdt_ID&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[],&quot;className&quot;:&quot;numdata integer  column-wdt_id&quot;,&quot;aTargets&quot;:[0]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;vpnservice&quot;,&quot;origHeader&quot;:&quot;vpnservice&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[],&quot;className&quot;:&quot; column-vpnservice&quot;,&quot;aTargets&quot;:[1]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;jurisdictionbasedincountry&quot;,&quot;origHeader&quot;:&quot;jurisdictionbasedincountry&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[],&quot;className&quot;:&quot; column-jurisdictionbasedincountry&quot;,&quot;aTargets&quot;:[2]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;jurisdictionfourteeneyes&quot;,&quot;origHeader&quot;:&quot;jurisdictionfourteeneyes&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Not Disclosed&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Fourteen&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Nine&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Five&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Owned&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Cooperative&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;}],&quot;className&quot;:&quot; column-jurisdictionfourteeneyes&quot;,&quot;aTargets&quot;:[3]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;jurisdictionfreedomstatus&quot;,&quot;origHeader&quot;:&quot;jurisdictionfreedomstatus&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Not Disclosed&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Owned&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;}],&quot;className&quot;:&quot; column-jurisdictionfreedomstatus&quot;,&quot;aTargets&quot;:[4]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;logginglogstraffic&quot;,&quot;origHeader&quot;:&quot;logginglogstraffic&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-logginglogstraffic&quot;,&quot;aTargets&quot;:[5]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;logginglogsdnsrequests&quot;,&quot;origHeader&quot;:&quot;logginglogsdnsrequests&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-logginglogsdnsrequests&quot;,&quot;aTargets&quot;:[6]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;logginglogstimestamps&quot;,&quot;origHeader&quot;:&quot;logginglogstimestamps&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-logginglogstimestamps&quot;,&quot;aTargets&quot;:[7]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;logginglogsbandwidth&quot;,&quot;origHeader&quot;:&quot;logginglogsbandwidth&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-logginglogsbandwidth&quot;,&quot;aTargets&quot;:[8]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;logginglogsipaddress&quot;,&quot;origHeader&quot;:&quot;logginglogsipaddress&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-logginglogsipaddress&quot;,&quot;aTargets&quot;:[9]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;activismanonymouspaymentmethod&quot;,&quot;origHeader&quot;:&quot;activismanonymouspaymentmethod&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Email&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;}],&quot;className&quot;:&quot; column-activismanonymouspaymentmethod&quot;,&quot;aTargets&quot;:[10]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;activismacceptscash&quot;,&quot;origHeader&quot;:&quot;activismacceptscash&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-activismacceptscash&quot;,&quot;aTargets&quot;:[11]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;activismacceptsgiftcards&quot;,&quot;origHeader&quot;:&quot;activismacceptsgiftcards&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-activismacceptsgiftcards&quot;,&quot;aTargets&quot;:[12]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;activismacceptsbitcoin&quot;,&quot;origHeader&quot;:&quot;activismacceptsbitcoin&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-activismacceptsbitcoin&quot;,&quot;aTargets&quot;:[13]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;activismpgpkeyavailable&quot;,&quot;origHeader&quot;:&quot;activismpgpkeyavailable&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-activismpgpkeyavailable&quot;,&quot;aTargets&quot;:[14]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;givesbacktoprivacycauses&quot;,&quot;origHeader&quot;:&quot;givesbacktoprivacycauses&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-givesbacktoprivacycauses&quot;,&quot;aTargets&quot;:[15]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;activismmeetsprivacytoolsiocriteria&quot;,&quot;origHeader&quot;:&quot;activismmeetsprivacytoolsiocriteria&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-activismmeetsprivacytoolsiocriteria&quot;,&quot;aTargets&quot;:[16]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;leakprotection1stpartydnsservers&quot;,&quot;origHeader&quot;:&quot;leakprotection1stpartydnsservers&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-leakprotection1stpartydnsservers&quot;,&quot;aTargets&quot;:[17]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;leakprotectionipv6supportedblocked&quot;,&quot;origHeader&quot;:&quot;leakprotectionipv6supportedblocked&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-leakprotectionipv6supportedblocked&quot;,&quot;aTargets&quot;:[18]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;protocolsoffersopenvpn&quot;,&quot;origHeader&quot;:&quot;protocolsoffersopenvpn&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-protocolsoffersopenvpn&quot;,&quot;aTargets&quot;:[19]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationsupportsmultihop&quot;,&quot;origHeader&quot;:&quot;obfuscationsupportsmultihop&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-obfuscationsupportsmultihop&quot;,&quot;aTargets&quot;:[20]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationsupportstcpport443&quot;,&quot;origHeader&quot;:&quot;obfuscationsupportstcpport443&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-obfuscationsupportstcpport443&quot;,&quot;aTargets&quot;:[21]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationsupportsobfsproxy&quot;,&quot;origHeader&quot;:&quot;obfuscationsupportsobfsproxy&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-obfuscationsupportsobfsproxy&quot;,&quot;aTargets&quot;:[22]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationsupportssocks&quot;,&quot;origHeader&quot;:&quot;obfuscationsupportssocks&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-obfuscationsupportssocks&quot;,&quot;aTargets&quot;:[23]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationsupportsssltunnel&quot;,&quot;origHeader&quot;:&quot;obfuscationsupportsssltunnel&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-obfuscationsupportsssltunnel&quot;,&quot;aTargets&quot;:[24]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationsupportssshtunnel&quot;,&quot;origHeader&quot;:&quot;obfuscationsupportssshtunnel&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-obfuscationsupportssshtunnel&quot;,&quot;aTargets&quot;:[25]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;obfuscationotherproprietaryprotocols&quot;,&quot;origHeader&quot;:&quot;obfuscationotherproprietaryprotocols&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;}],&quot;className&quot;:&quot; column-obfuscationotherproprietaryprotocols&quot;,&quot;aTargets&quot;:[26]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;portblockingauthsmtp&quot;,&quot;origHeader&quot;:&quot;portblockingauthsmtp&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Some&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-portblockingauthsmtp&quot;,&quot;aTargets&quot;:[27]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;portblockingp2p&quot;,&quot;origHeader&quot;:&quot;portblockingp2p&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Some&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-portblockingp2p&quot;,&quot;aTargets&quot;:[28]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;float&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;speedsusserveraveragembps&quot;,&quot;origHeader&quot;:&quot;speedsusserveraveragembps&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:80,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:80,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:50,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:0,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;&quot;}],&quot;className&quot;:&quot;numdata float  column-speedsusserveraveragembps&quot;,&quot;aTargets&quot;:[29]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;float&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;speedsinternationalserveraveragembps&quot;,&quot;origHeader&quot;:&quot;speedsinternationalserveraveragembps&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:20,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:20,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:0,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;&quot;}],&quot;className&quot;:&quot;numdata float  column-speedsinternationalserveraveragembps&quot;,&quot;aTargets&quot;:[30]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;serversdedicatedorvirtual&quot;,&quot;origHeader&quot;:&quot;serversdedicatedorvirtual&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Dedicated&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Mix&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Virtual&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-serversdedicatedorvirtual&quot;,&quot;aTargets&quot;:[31]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;securityweakestdataencryption&quot;,&quot;origHeader&quot;:&quot;securityweakestdataencryption&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;192&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;160&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;128&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;256&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;Blowfish&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;MPPE&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;64&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;196&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-securityweakestdataencryption&quot;,&quot;aTargets&quot;:[32]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;securitystrongestdataencryption&quot;,&quot;origHeader&quot;:&quot;securitystrongestdataencryption&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;192&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;160&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;128&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;256&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;Blowfish&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;MPPE&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;64&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;196&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-securitystrongestdataencryption&quot;,&quot;aTargets&quot;:[33]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;securityweakesthandshakeencryption&quot;,&quot;origHeader&quot;:&quot;securityweakesthandshakeencryption&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;4096&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;2048&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;RC4&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;v2&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;1024&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-securityweakesthandshakeencryption&quot;,&quot;aTargets&quot;:[34]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;securitystrongesthandshakeencryption&quot;,&quot;origHeader&quot;:&quot;securitystrongesthandshakeencryption&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;4096&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;2048&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;RC4&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;v2&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;1024&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-securitystrongesthandshakeencryption&quot;,&quot;aTargets&quot;:[35]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;availabilityofconnections&quot;,&quot;origHeader&quot;:&quot;availabilityofconnections&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:5,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:5,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot;numdata integer  column-availabilityofconnections&quot;,&quot;aTargets&quot;:[36]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;availabilityofcountries&quot;,&quot;origHeader&quot;:&quot;availabilityofcountries&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot;numdata integer  column-availabilityofcountries&quot;,&quot;aTargets&quot;:[37]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;availabilityofservers&quot;,&quot;origHeader&quot;:&quot;availabilityofservers&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot;numdata integer  column-availabilityofservers&quot;,&quot;aTargets&quot;:[38]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;supportlinuxmanualconfig&quot;,&quot;origHeader&quot;:&quot;supportlinuxmanualconfig&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;Partial&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;}],&quot;className&quot;:&quot; column-supportlinuxmanualconfig&quot;,&quot;aTargets&quot;:[39]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;websiteofpersistentcookies&quot;,&quot;origHeader&quot;:&quot;websiteofpersistentcookies&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot;numdata integer  column-websiteofpersistentcookies&quot;,&quot;aTargets&quot;:[40]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;websiteofexternaltrackers&quot;,&quot;origHeader&quot;:&quot;websiteofexternaltrackers&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot;numdata integer  column-websiteofexternaltrackers&quot;,&quot;aTargets&quot;:[41]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;websiteofproprietaryapis&quot;,&quot;origHeader&quot;:&quot;websiteofproprietaryapis&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:10,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot;numdata integer  column-websiteofproprietaryapis&quot;,&quot;aTargets&quot;:[42]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;websiteserversslrating&quot;,&quot;origHeader&quot;:&quot;websiteserversslrating&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;T&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;F&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;D&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#f9cb9c&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;C&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;B&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#b6d7a8&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;A&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-websiteserversslrating&quot;,&quot;aTargets&quot;:[43]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;websitesslcertissuedto&quot;,&quot;origHeader&quot;:&quot;websitesslcertissuedto&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;contains_not&quot;,&quot;cellVal&quot;:&quot;Self&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;contains&quot;,&quot;cellVal&quot;:&quot;Self&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-websitesslcertissuedto&quot;,&quot;aTargets&quot;:[44]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;float&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;pricingmonthannualpricing&quot;,&quot;origHeader&quot;:&quot;pricingmonthannualpricing&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:6.67,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:6.67,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:4.17,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot;numdata float  column-pricingmonthannualpricing&quot;,&quot;aTargets&quot;:[45]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;float&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;pricingconnectionmonth&quot;,&quot;origHeader&quot;:&quot;pricingconnectionmonth&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:4,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:4,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:1.12,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot;numdata float  column-pricingconnectionmonth&quot;,&quot;aTargets&quot;:[46]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;pricingfreetrial&quot;,&quot;origHeader&quot;:&quot;pricingfreetrial&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Free&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-pricingfreetrial&quot;,&quot;aTargets&quot;:[47]},{&quot;sType&quot;:&quot;formatted-num&quot;,&quot;wdtType&quot;:&quot;int&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;pricingrefundperioddays&quot;,&quot;origHeader&quot;:&quot;pricingrefundperioddays&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;gteq&quot;,&quot;cellVal&quot;:7,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:7,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;lt&quot;,&quot;cellVal&quot;:3,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:-1,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot;numdata integer  column-pricingrefundperioddays&quot;,&quot;aTargets&quot;:[48]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;ethicscontradictoryloggingpolicies&quot;,&quot;origHeader&quot;:&quot;ethicscontradictoryloggingpolicies&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-ethicscontradictoryloggingpolicies&quot;,&quot;aTargets&quot;:[49]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;ethicsfalselyclaims100effective&quot;,&quot;origHeader&quot;:&quot;ethicsfalselyclaims100effective&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-ethicsfalselyclaims100effective&quot;,&quot;aTargets&quot;:[50]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;ethicsincentivizessocialmediaspam&quot;,&quot;origHeader&quot;:&quot;ethicsincentivizessocialmediaspam&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-ethicsincentivizessocialmediaspam&quot;,&quot;aTargets&quot;:[51]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;policiesforbidsspam&quot;,&quot;origHeader&quot;:&quot;policiesforbidsspam&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Some&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-policiesforbidsspam&quot;,&quot;aTargets&quot;:[52]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;policiesrequiresethicalcopy&quot;,&quot;origHeader&quot;:&quot;policiesrequiresethicalcopy&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Some&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-policiesrequiresethicalcopy&quot;,&quot;aTargets&quot;:[53]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;policiesrequiresfulldisclosure&quot;,&quot;origHeader&quot;:&quot;policiesrequiresfulldisclosure&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Some&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#ffd666&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;},{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;Yes&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#57bb8a&quot;}],&quot;className&quot;:&quot; column-policiesrequiresfulldisclosure&quot;,&quot;aTargets&quot;:[54]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;affiliatespracticeethicalcopy&quot;,&quot;origHeader&quot;:&quot;affiliatespracticeethicalcopy&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-affiliatespracticeethicalcopy&quot;,&quot;aTargets&quot;:[55]},{&quot;sType&quot;:&quot;string&quot;,&quot;wdtType&quot;:&quot;string&quot;,&quot;bVisible&quot;:true,&quot;orderable&quot;:true,&quot;searchable&quot;:true,&quot;InputType&quot;:&quot;text&quot;,&quot;name&quot;:&quot;affiliatesgivefulldisclosure&quot;,&quot;origHeader&quot;:&quot;affiliatesgivefulldisclosure&quot;,&quot;notNull&quot;:false,&quot;conditionalFormattingRules&quot;:[{&quot;ifClause&quot;:&quot;eq&quot;,&quot;cellVal&quot;:&quot;No&quot;,&quot;action&quot;:&quot;setCellColor&quot;,&quot;setVal&quot;:&quot;#e67c73&quot;}],&quot;className&quot;:&quot; column-affiliatesgivefulldisclosure&quot;,&quot;aTargets&quot;:[56]}],&quot;bAutoWidth&quot;:false,&quot;order&quot;:[[1,&quot;asc&quot;]],&quot;ordering&quot;:true,&quot;buttons&quot;:[],&quot;bProcessing&quot;:false,&quot;serverSide&quot;:true,&quot;ajax&quot;:{&quot;url&quot;:&quot;https:\\\/\\\/thatoneprivacysite.net\\\/wp-admin\\\/admin-ajax.php?action=get_wdtable&amp;table_id=17&quot;,&quot;type&quot;:&quot;POST&quot;},&quot;sPaginationType&quot;:&quot;full_numbers&quot;,&quot;oSearch&quot;:{&quot;bSmart&quot;:false,&quot;bRegex&quot;:false,&quot;sSearch&quot;:&quot;&quot;}},&quot;tabletWidth&quot;:&quot;1024&quot;,&quot;mobileWidth&quot;:&quot;480&quot;,&quot;advancedFilterEnabled&quot;:true,&quot;advancedFilterOptions&quot;:{&quot;aoColumns&quot;:[{&quot;type&quot;:&quot;null&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;wdt_ID&quot;,&quot;displayHeader&quot;:&quot;wdt_ID&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;AceVPN&quot;,&quot;label&quot;:&quot;AceVPN&quot;},{&quot;value&quot;:&quot;ActiVPN&quot;,&quot;label&quot;:&quot;ActiVPN&quot;},{&quot;value&quot;:&quot;AirVPN&quot;,&quot;label&quot;:&quot;AirVPN&quot;},{&quot;value&quot;:&quot;Anonine&quot;,&quot;label&quot;:&quot;Anonine&quot;},{&quot;value&quot;:&quot;AnonVPN&quot;,&quot;label&quot;:&quot;AnonVPN&quot;},{&quot;value&quot;:&quot;Anonymizer&quot;,&quot;label&quot;:&quot;Anonymizer&quot;},{&quot;value&quot;:&quot;AnonymousVPN&quot;,&quot;label&quot;:&quot;AnonymousVPN&quot;},{&quot;value&quot;:&quot;Astrill&quot;,&quot;label&quot;:&quot;Astrill&quot;},{&quot;value&quot;:&quot;Avast Secureline&quot;,&quot;label&quot;:&quot;Avast Secureline&quot;},{&quot;value&quot;:&quot;Avira Phantom VPN&quot;,&quot;label&quot;:&quot;Avira Phantom VPN&quot;},{&quot;value&quot;:&quot;AzireVPN&quot;,&quot;label&quot;:&quot;AzireVPN&quot;},{&quot;value&quot;:&quot;BeeVPN&quot;,&quot;label&quot;:&quot;BeeVPN&quot;},{&quot;value&quot;:&quot;Betternet&quot;,&quot;label&quot;:&quot;Betternet&quot;},{&quot;value&quot;:&quot;BlackVPN&quot;,&quot;label&quot;:&quot;BlackVPN&quot;},{&quot;value&quot;:&quot;Blockless&quot;,&quot;label&quot;:&quot;Blockless&quot;},{&quot;value&quot;:&quot;BolehVPN&quot;,&quot;label&quot;:&quot;BolehVPN&quot;},{&quot;value&quot;:&quot;Boxpn&quot;,&quot;label&quot;:&quot;Boxpn&quot;},{&quot;value&quot;:&quot;BTGuard&quot;,&quot;label&quot;:&quot;BTGuard&quot;},{&quot;value&quot;:&quot;Buffered&quot;,&quot;label&quot;:&quot;Buffered&quot;},{&quot;value&quot;:&quot;CactusVPN&quot;,&quot;label&quot;:&quot;CactusVPN&quot;},{&quot;value&quot;:&quot;Celo&quot;,&quot;label&quot;:&quot;Celo&quot;},{&quot;value&quot;:&quot;ChillGlobal&quot;,&quot;label&quot;:&quot;ChillGlobal&quot;},{&quot;value&quot;:&quot;CitizenVPN&quot;,&quot;label&quot;:&quot;CitizenVPN&quot;},{&quot;value&quot;:&quot;CloakVPN&quot;,&quot;label&quot;:&quot;CloakVPN&quot;},{&quot;value&quot;:&quot;CrypticVPN&quot;,&quot;label&quot;:&quot;CrypticVPN&quot;},{&quot;value&quot;:&quot;CryptoHippie&quot;,&quot;label&quot;:&quot;CryptoHippie&quot;},{&quot;value&quot;:&quot;CryptoStorm&quot;,&quot;label&quot;:&quot;CryptoStorm&quot;},{&quot;value&quot;:&quot;CyberGhost&quot;,&quot;label&quot;:&quot;CyberGhost&quot;},{&quot;value&quot;:&quot;DathoVPN&quot;,&quot;label&quot;:&quot;DathoVPN&quot;},{&quot;value&quot;:&quot;DefenceVPN&quot;,&quot;label&quot;:&quot;DefenceVPN&quot;},{&quot;value&quot;:&quot;Disconnect.me&quot;,&quot;label&quot;:&quot;Disconnect.me&quot;},{&quot;value&quot;:&quot;DooVPN&quot;,&quot;label&quot;:&quot;DooVPN&quot;},{&quot;value&quot;:&quot;DotVPN&quot;,&quot;label&quot;:&quot;DotVPN&quot;},{&quot;value&quot;:&quot;Doublehop&quot;,&quot;label&quot;:&quot;Doublehop&quot;},{&quot;value&quot;:&quot;EarthVPN&quot;,&quot;label&quot;:&quot;EarthVPN&quot;},{&quot;value&quot;:&quot;Encrypt.me&quot;,&quot;label&quot;:&quot;Encrypt.me&quot;},{&quot;value&quot;:&quot;EtherealVPN&quot;,&quot;label&quot;:&quot;EtherealVPN&quot;},{&quot;value&quot;:&quot;ExpressVPN&quot;,&quot;label&quot;:&quot;ExpressVPN&quot;},{&quot;value&quot;:&quot;Faceless.ME&quot;,&quot;label&quot;:&quot;Faceless.ME&quot;},{&quot;value&quot;:&quot;FinchVPN&quot;,&quot;label&quot;:&quot;FinchVPN&quot;},{&quot;value&quot;:&quot;FlowVPN&quot;,&quot;label&quot;:&quot;FlowVPN&quot;},{&quot;value&quot;:&quot;FlyVPN&quot;,&quot;label&quot;:&quot;FlyVPN&quot;},{&quot;value&quot;:&quot;FoxyProxy&quot;,&quot;label&quot;:&quot;FoxyProxy&quot;},{&quot;value&quot;:&quot;Freedom-IP&quot;,&quot;label&quot;:&quot;Freedom-IP&quot;},{&quot;value&quot;:&quot;Freedome&quot;,&quot;label&quot;:&quot;Freedome&quot;},{&quot;value&quot;:&quot;FrootVPN&quot;,&quot;label&quot;:&quot;FrootVPN&quot;},{&quot;value&quot;:&quot;FrostVPN&quot;,&quot;label&quot;:&quot;FrostVPN&quot;},{&quot;value&quot;:&quot;GetFlix&quot;,&quot;label&quot;:&quot;GetFlix&quot;},{&quot;value&quot;:&quot;GhostPath&quot;,&quot;label&quot;:&quot;GhostPath&quot;},{&quot;value&quot;:&quot;GooseVPN&quot;,&quot;label&quot;:&quot;GooseVPN&quot;},{&quot;value&quot;:&quot;GoTrusted&quot;,&quot;label&quot;:&quot;GoTrusted&quot;},{&quot;value&quot;:&quot;GoVPN&quot;,&quot;label&quot;:&quot;GoVPN&quot;},{&quot;value&quot;:&quot;Hide My IP&quot;,&quot;label&quot;:&quot;Hide My IP&quot;},{&quot;value&quot;:&quot;Hide.me&quot;,&quot;label&quot;:&quot;Hide.me&quot;},{&quot;value&quot;:&quot;HideALLIP&quot;,&quot;label&quot;:&quot;HideALLIP&quot;},{&quot;value&quot;:&quot;HideIPVPN&quot;,&quot;label&quot;:&quot;HideIPVPN&quot;},{&quot;value&quot;:&quot;HideMyAss&quot;,&quot;label&quot;:&quot;HideMyAss&quot;},{&quot;value&quot;:&quot;Hola!VPN&quot;,&quot;label&quot;:&quot;Hola!VPN&quot;},{&quot;value&quot;:&quot;Hotspot Shield&quot;,&quot;label&quot;:&quot;Hotspot Shield&quot;},{&quot;value&quot;:&quot;HotVPN&quot;,&quot;label&quot;:&quot;HotVPN&quot;},{&quot;value&quot;:&quot;IBVPN&quot;,&quot;label&quot;:&quot;IBVPN&quot;},{&quot;value&quot;:&quot;Identity Cloaker&quot;,&quot;label&quot;:&quot;Identity Cloaker&quot;},{&quot;value&quot;:&quot;In-Disguise&quot;,&quot;label&quot;:&quot;In-Disguise&quot;},{&quot;value&quot;:&quot;IncognitoVPN&quot;,&quot;label&quot;:&quot;IncognitoVPN&quot;},{&quot;value&quot;:&quot;Insorg&quot;,&quot;label&quot;:&quot;Insorg&quot;},{&quot;value&quot;:&quot;Integrity.st&quot;,&quot;label&quot;:&quot;Integrity.st&quot;},{&quot;value&quot;:&quot;IntelliVPN&quot;,&quot;label&quot;:&quot;IntelliVPN&quot;},{&quot;value&quot;:&quot;Internetz.me&quot;,&quot;label&quot;:&quot;Internetz.me&quot;},{&quot;value&quot;:&quot;IntroVPN&quot;,&quot;label&quot;:&quot;IntroVPN&quot;},{&quot;value&quot;:&quot;IPinator&quot;,&quot;label&quot;:&quot;IPinator&quot;},{&quot;value&quot;:&quot;IPredator&quot;,&quot;label&quot;:&quot;IPredator&quot;},{&quot;value&quot;:&quot;IPVanish&quot;,&quot;label&quot;:&quot;IPVanish&quot;},{&quot;value&quot;:&quot;Ironsocket&quot;,&quot;label&quot;:&quot;Ironsocket&quot;},{&quot;value&quot;:&quot;Ivacy&quot;,&quot;label&quot;:&quot;Ivacy&quot;},{&quot;value&quot;:&quot;IVPN&quot;,&quot;label&quot;:&quot;IVPN&quot;},{&quot;value&quot;:&quot;KeepSolid VPN Unlimited&quot;,&quot;label&quot;:&quot;KeepSolid VPN Unlimited&quot;},{&quot;value&quot;:&quot;Kepard&quot;,&quot;label&quot;:&quot;Kepard&quot;},{&quot;value&quot;:&quot;LeVPN&quot;,&quot;label&quot;:&quot;LeVPN&quot;},{&quot;value&quot;:&quot;LibertyShield&quot;,&quot;label&quot;:&quot;LibertyShield&quot;},{&quot;value&quot;:&quot;LibertyVPN&quot;,&quot;label&quot;:&quot;LibertyVPN&quot;},{&quot;value&quot;:&quot;LimeVPN&quot;,&quot;label&quot;:&quot;LimeVPN&quot;},{&quot;value&quot;:&quot;LiquidVPN&quot;,&quot;label&quot;:&quot;LiquidVPN&quot;},{&quot;value&quot;:&quot;Mullvad&quot;,&quot;label&quot;:&quot;Mullvad&quot;},{&quot;value&quot;:&quot;My Expat Network&quot;,&quot;label&quot;:&quot;My Expat Network&quot;},{&quot;value&quot;:&quot;My Private Network&quot;,&quot;label&quot;:&quot;My Private Network&quot;},{&quot;value&quot;:&quot;MyIP.io&quot;,&quot;label&quot;:&quot;MyIP.io&quot;},{&quot;value&quot;:&quot;MyVPN.Pro&quot;,&quot;label&quot;:&quot;MyVPN.Pro&quot;},{&quot;value&quot;:&quot;NetShade&quot;,&quot;label&quot;:&quot;NetShade&quot;},{&quot;value&quot;:&quot;Newshosting&quot;,&quot;label&quot;:&quot;Newshosting&quot;},{&quot;value&quot;:&quot;NEXTGenVPN&quot;,&quot;label&quot;:&quot;NEXTGenVPN&quot;},{&quot;value&quot;:&quot;NolimitVPN&quot;,&quot;label&quot;:&quot;NolimitVPN&quot;},{&quot;value&quot;:&quot;NordVPN&quot;,&quot;label&quot;:&quot;NordVPN&quot;},{&quot;value&quot;:&quot;Norton WiFi Privacy&quot;,&quot;label&quot;:&quot;Norton WiFi Privacy&quot;},{&quot;value&quot;:&quot;NVPN&quot;,&quot;label&quot;:&quot;NVPN&quot;},{&quot;value&quot;:&quot;OctaneVPN&quot;,&quot;label&quot;:&quot;OctaneVPN&quot;},{&quot;value&quot;:&quot;OneVPN&quot;,&quot;label&quot;:&quot;OneVPN&quot;},{&quot;value&quot;:&quot;Opera (Browser) VPN&quot;,&quot;label&quot;:&quot;Opera (Browser) VPN&quot;},{&quot;value&quot;:&quot;OverPlay&quot;,&quot;label&quot;:&quot;OverPlay&quot;},{&quot;value&quot;:&quot;OVPN.com&quot;,&quot;label&quot;:&quot;OVPN.com&quot;},{&quot;value&quot;:&quot;oVPN.to&quot;,&quot;label&quot;:&quot;oVPN.to&quot;},{&quot;value&quot;:&quot;Perfect Privacy&quot;,&quot;label&quot;:&quot;Perfect Privacy&quot;},{&quot;value&quot;:&quot;Private Internet Access&quot;,&quot;label&quot;:&quot;Private Internet Access&quot;},{&quot;value&quot;:&quot;PrivatePackets.io&quot;,&quot;label&quot;:&quot;PrivatePackets.io&quot;},{&quot;value&quot;:&quot;PrivateTunnel&quot;,&quot;label&quot;:&quot;PrivateTunnel&quot;},{&quot;value&quot;:&quot;PrivateVPN&quot;,&quot;label&quot;:&quot;PrivateVPN&quot;},{&quot;value&quot;:&quot;ProtonVPN&quot;,&quot;label&quot;:&quot;ProtonVPN&quot;},{&quot;value&quot;:&quot;ProXPN&quot;,&quot;label&quot;:&quot;ProXPN&quot;},{&quot;value&quot;:&quot;Proxy.sh&quot;,&quot;label&quot;:&quot;Proxy.sh&quot;},{&quot;value&quot;:&quot;PRQ&quot;,&quot;label&quot;:&quot;PRQ&quot;},{&quot;value&quot;:&quot;PureVPN&quot;,&quot;label&quot;:&quot;PureVPN&quot;},{&quot;value&quot;:&quot;RA4W VPN&quot;,&quot;label&quot;:&quot;RA4W VPN&quot;},{&quot;value&quot;:&quot;RogueVPN&quot;,&quot;label&quot;:&quot;RogueVPN&quot;},{&quot;value&quot;:&quot;RootVPN&quot;,&quot;label&quot;:&quot;RootVPN&quot;},{&quot;value&quot;:&quot;SaferVPN&quot;,&quot;label&quot;:&quot;SaferVPN&quot;},{&quot;value&quot;:&quot;SecureVPN.com&quot;,&quot;label&quot;:&quot;SecureVPN.com&quot;},{&quot;value&quot;:&quot;SecureVPN.to&quot;,&quot;label&quot;:&quot;SecureVPN.to&quot;},{&quot;value&quot;:&quot;SecurityKISS&quot;,&quot;label&quot;:&quot;SecurityKISS&quot;},{&quot;value&quot;:&quot;Seed4.Me&quot;,&quot;label&quot;:&quot;Seed4.Me&quot;},{&quot;value&quot;:&quot;Seedboxes.cc&quot;,&quot;label&quot;:&quot;Seedboxes.cc&quot;},{&quot;value&quot;:&quot;ShadeYou&quot;,&quot;label&quot;:&quot;ShadeYou&quot;},{&quot;value&quot;:&quot;SigaVPN&quot;,&quot;label&quot;:&quot;SigaVPN&quot;},{&quot;value&quot;:&quot;SlickVPN&quot;,&quot;label&quot;:&quot;SlickVPN&quot;},{&quot;value&quot;:&quot;Smart DNS Proxy&quot;,&quot;label&quot;:&quot;Smart DNS Proxy&quot;},{&quot;value&quot;:&quot;SmartVPN&quot;,&quot;label&quot;:&quot;SmartVPN&quot;},{&quot;value&quot;:&quot;Speedify&quot;,&quot;label&quot;:&quot;Speedify&quot;},{&quot;value&quot;:&quot;Spotflux&quot;,&quot;label&quot;:&quot;Spotflux&quot;},{&quot;value&quot;:&quot;Steganos&quot;,&quot;label&quot;:&quot;Steganos&quot;},{&quot;value&quot;:&quot;StrongVPN&quot;,&quot;label&quot;:&quot;StrongVPN&quot;},{&quot;value&quot;:&quot;SumRando&quot;,&quot;label&quot;:&quot;SumRando&quot;},{&quot;value&quot;:&quot;SunVPN&quot;,&quot;label&quot;:&quot;SunVPN&quot;},{&quot;value&quot;:&quot;SuperVPN&quot;,&quot;label&quot;:&quot;SuperVPN&quot;},{&quot;value&quot;:&quot;SurfEasy&quot;,&quot;label&quot;:&quot;SurfEasy&quot;},{&quot;value&quot;:&quot;SwitchVPN&quot;,&quot;label&quot;:&quot;SwitchVPN&quot;},{&quot;value&quot;:&quot;TGVPN&quot;,&quot;label&quot;:&quot;TGVPN&quot;},{&quot;value&quot;:&quot;TigerVPN&quot;,&quot;label&quot;:&quot;TigerVPN&quot;},{&quot;value&quot;:&quot;Torguard&quot;,&quot;label&quot;:&quot;Torguard&quot;},{&quot;value&quot;:&quot;TorrentPrivacy&quot;,&quot;label&quot;:&quot;TorrentPrivacy&quot;},{&quot;value&quot;:&quot;TorVPN&quot;,&quot;label&quot;:&quot;TorVPN&quot;},{&quot;value&quot;:&quot;TotalVPN&quot;,&quot;label&quot;:&quot;TotalVPN&quot;},{&quot;value&quot;:&quot;Traceless.me&quot;,&quot;label&quot;:&quot;Traceless.me&quot;},{&quot;value&quot;:&quot;Trust.Zone&quot;,&quot;label&quot;:&quot;Trust.Zone&quot;},{&quot;value&quot;:&quot;TunnelBear&quot;,&quot;label&quot;:&quot;TunnelBear&quot;},{&quot;value&quot;:&quot;Tunnelr&quot;,&quot;label&quot;:&quot;Tunnelr&quot;},{&quot;value&quot;:&quot;TVWhenAway&quot;,&quot;label&quot;:&quot;TVWhenAway&quot;},{&quot;value&quot;:&quot;Unblock VPN&quot;,&quot;label&quot;:&quot;Unblock VPN&quot;},{&quot;value&quot;:&quot;Unblock-Us&quot;,&quot;label&quot;:&quot;Unblock-Us&quot;},{&quot;value&quot;:&quot;Unlocator&quot;,&quot;label&quot;:&quot;Unlocator&quot;},{&quot;value&quot;:&quot;UnoTelly&quot;,&quot;label&quot;:&quot;UnoTelly&quot;},{&quot;value&quot;:&quot;Unseen Online&quot;,&quot;label&quot;:&quot;Unseen Online&quot;},{&quot;value&quot;:&quot;Unspyable&quot;,&quot;label&quot;:&quot;Unspyable&quot;},{&quot;value&quot;:&quot;VanishedVPN&quot;,&quot;label&quot;:&quot;VanishedVPN&quot;},{&quot;value&quot;:&quot;VikingVPN&quot;,&quot;label&quot;:&quot;VikingVPN&quot;},{&quot;value&quot;:&quot;VIP72&quot;,&quot;label&quot;:&quot;VIP72&quot;},{&quot;value&quot;:&quot;Virtual Shield&quot;,&quot;label&quot;:&quot;Virtual Shield&quot;},{&quot;value&quot;:&quot;VPN Gate&quot;,&quot;label&quot;:&quot;VPN Gate&quot;},{&quot;value&quot;:&quot;VPN Land&quot;,&quot;label&quot;:&quot;VPN Land&quot;},{&quot;value&quot;:&quot;VPN Master&quot;,&quot;label&quot;:&quot;VPN Master&quot;},{&quot;value&quot;:&quot;VPN Shield&quot;,&quot;label&quot;:&quot;VPN Shield&quot;},{&quot;value&quot;:&quot;VPN.ac&quot;,&quot;label&quot;:&quot;VPN.ac&quot;},{&quot;value&quot;:&quot;VPN.Asia&quot;,&quot;label&quot;:&quot;VPN.Asia&quot;},{&quot;value&quot;:&quot;VPN.cc&quot;,&quot;label&quot;:&quot;VPN.cc&quot;},{&quot;value&quot;:&quot;VPN.ht&quot;,&quot;label&quot;:&quot;VPN.ht&quot;},{&quot;value&quot;:&quot;VPN.sh&quot;,&quot;label&quot;:&quot;VPN.sh&quot;},{&quot;value&quot;:&quot;VPN4All&quot;,&quot;label&quot;:&quot;VPN4All&quot;},{&quot;value&quot;:&quot;VPNArea&quot;,&quot;label&quot;:&quot;VPNArea&quot;},{&quot;value&quot;:&quot;VPNAUS&quot;,&quot;label&quot;:&quot;VPNAUS&quot;},{&quot;value&quot;:&quot;VPNBaron&quot;,&quot;label&quot;:&quot;VPNBaron&quot;},{&quot;value&quot;:&quot;VPNJack&quot;,&quot;label&quot;:&quot;VPNJack&quot;},{&quot;value&quot;:&quot;VPNMe&quot;,&quot;label&quot;:&quot;VPNMe&quot;},{&quot;value&quot;:&quot;VPNSecure&quot;,&quot;label&quot;:&quot;VPNSecure&quot;},{&quot;value&quot;:&quot;VPNShazam&quot;,&quot;label&quot;:&quot;VPNShazam&quot;},{&quot;value&quot;:&quot;VPNTraffic&quot;,&quot;label&quot;:&quot;VPNTraffic&quot;},{&quot;value&quot;:&quot;VPNTunnel&quot;,&quot;label&quot;:&quot;VPNTunnel&quot;},{&quot;value&quot;:&quot;VPNUK&quot;,&quot;label&quot;:&quot;VPNUK&quot;},{&quot;value&quot;:&quot;VyprVPN&quot;,&quot;label&quot;:&quot;VyprVPN&quot;},{&quot;value&quot;:&quot;WASEL Pro&quot;,&quot;label&quot;:&quot;WASEL Pro&quot;},{&quot;value&quot;:&quot;Whoer&quot;,&quot;label&quot;:&quot;Whoer&quot;},{&quot;value&quot;:&quot;WifiMask&quot;,&quot;label&quot;:&quot;WifiMask&quot;},{&quot;value&quot;:&quot;WindScribe&quot;,&quot;label&quot;:&quot;WindScribe&quot;},{&quot;value&quot;:&quot;WiTopia&quot;,&quot;label&quot;:&quot;WiTopia&quot;},{&quot;value&quot;:&quot;WorldVPN&quot;,&quot;label&quot;:&quot;WorldVPN&quot;},{&quot;value&quot;:&quot;Zenmate&quot;,&quot;label&quot;:&quot;Zenmate&quot;},{&quot;value&quot;:&quot;ZoogVPN&quot;,&quot;label&quot;:&quot;ZoogVPN&quot;},{&quot;value&quot;:&quot;ZorroVPN&quot;,&quot;label&quot;:&quot;ZorroVPN&quot;},{&quot;value&quot;:&quot;ZPN&quot;,&quot;label&quot;:&quot;ZPN&quot;}],&quot;origHeader&quot;:&quot;vpnservice&quot;,&quot;displayHeader&quot;:&quot;VPN SERVICE&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:1,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;Australia&quot;,&quot;label&quot;:&quot;Australia&quot;},{&quot;value&quot;:&quot;Barbados&quot;,&quot;label&quot;:&quot;Barbados&quot;},{&quot;value&quot;:&quot;Belize&quot;,&quot;label&quot;:&quot;Belize&quot;},{&quot;value&quot;:&quot;Bosnia&quot;,&quot;label&quot;:&quot;Bosnia&quot;},{&quot;value&quot;:&quot;British Indian Ocean&quot;,&quot;label&quot;:&quot;British Indian Ocean&quot;},{&quot;value&quot;:&quot;British Virgin Islands&quot;,&quot;label&quot;:&quot;British Virgin Islands&quot;},{&quot;value&quot;:&quot;Bulgaria&quot;,&quot;label&quot;:&quot;Bulgaria&quot;},{&quot;value&quot;:&quot;Canada&quot;,&quot;label&quot;:&quot;Canada&quot;},{&quot;value&quot;:&quot;Cyprus&quot;,&quot;label&quot;:&quot;Cyprus&quot;},{&quot;value&quot;:&quot;Czech Republic&quot;,&quot;label&quot;:&quot;Czech Republic&quot;},{&quot;value&quot;:&quot;Denmark&quot;,&quot;label&quot;:&quot;Denmark&quot;},{&quot;value&quot;:&quot;Finland&quot;,&quot;label&quot;:&quot;Finland&quot;},{&quot;value&quot;:&quot;France&quot;,&quot;label&quot;:&quot;France&quot;},{&quot;value&quot;:&quot;Germany&quot;,&quot;label&quot;:&quot;Germany&quot;},{&quot;value&quot;:&quot;Gibraltar&quot;,&quot;label&quot;:&quot;Gibraltar&quot;},{&quot;value&quot;:&quot;Greece&quot;,&quot;label&quot;:&quot;Greece&quot;},{&quot;value&quot;:&quot;Hong Kong&quot;,&quot;label&quot;:&quot;Hong Kong&quot;},{&quot;value&quot;:&quot;Hungary&quot;,&quot;label&quot;:&quot;Hungary&quot;},{&quot;value&quot;:&quot;Iceland&quot;,&quot;label&quot;:&quot;Iceland&quot;},{&quot;value&quot;:&quot;India&quot;,&quot;label&quot;:&quot;India&quot;},{&quot;value&quot;:&quot;Ireland&quot;,&quot;label&quot;:&quot;Ireland&quot;},{&quot;value&quot;:&quot;Israel&quot;,&quot;label&quot;:&quot;Israel&quot;},{&quot;value&quot;:&quot;Italy&quot;,&quot;label&quot;:&quot;Italy&quot;},{&quot;value&quot;:&quot;Japan&quot;,&quot;label&quot;:&quot;Japan&quot;},{&quot;value&quot;:&quot;Malaysia&quot;,&quot;label&quot;:&quot;Malaysia&quot;},{&quot;value&quot;:&quot;Malta&quot;,&quot;label&quot;:&quot;Malta&quot;},{&quot;value&quot;:&quot;Mauritius&quot;,&quot;label&quot;:&quot;Mauritius&quot;},{&quot;value&quot;:&quot;Moldova&quot;,&quot;label&quot;:&quot;Moldova&quot;},{&quot;value&quot;:&quot;Morocco&quot;,&quot;label&quot;:&quot;Morocco&quot;},{&quot;value&quot;:&quot;Netherlands&quot;,&quot;label&quot;:&quot;Netherlands&quot;},{&quot;value&quot;:&quot;Northern Cyprus&quot;,&quot;label&quot;:&quot;Northern Cyprus&quot;},{&quot;value&quot;:&quot;Norway&quot;,&quot;label&quot;:&quot;Norway&quot;},{&quot;value&quot;:&quot;Panama&quot;,&quot;label&quot;:&quot;Panama&quot;},{&quot;value&quot;:&quot;Poland&quot;,&quot;label&quot;:&quot;Poland&quot;},{&quot;value&quot;:&quot;Romania&quot;,&quot;label&quot;:&quot;Romania&quot;},{&quot;value&quot;:&quot;Seychelles&quot;,&quot;label&quot;:&quot;Seychelles&quot;},{&quot;value&quot;:&quot;Singapore&quot;,&quot;label&quot;:&quot;Singapore&quot;},{&quot;value&quot;:&quot;Slovakia&quot;,&quot;label&quot;:&quot;Slovakia&quot;},{&quot;value&quot;:&quot;Sweden&quot;,&quot;label&quot;:&quot;Sweden&quot;},{&quot;value&quot;:&quot;Switzerland&quot;,&quot;label&quot;:&quot;Switzerland&quot;},{&quot;value&quot;:&quot;Taiwan&quot;,&quot;label&quot;:&quot;Taiwan&quot;},{&quot;value&quot;:&quot;UAE&quot;,&quot;label&quot;:&quot;UAE&quot;},{&quot;value&quot;:&quot;UK&quot;,&quot;label&quot;:&quot;UK&quot;},{&quot;value&quot;:&quot;USA&quot;,&quot;label&quot;:&quot;USA&quot;}],&quot;origHeader&quot;:&quot;jurisdictionbasedincountry&quot;,&quot;displayHeader&quot;:&quot;JURISDICTION Based In (Country)&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Country&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;Cooperative&quot;,&quot;label&quot;:&quot;Cooperative&quot;},{&quot;value&quot;:&quot;Five&quot;,&quot;label&quot;:&quot;Five&quot;},{&quot;value&quot;:&quot;Fourteen&quot;,&quot;label&quot;:&quot;Fourteen&quot;},{&quot;value&quot;:&quot;Nine&quot;,&quot;label&quot;:&quot;Nine&quot;},{&quot;value&quot;:&quot;No&quot;,&quot;label&quot;:&quot;No&quot;},{&quot;value&quot;:&quot;Not Disclosed&quot;,&quot;label&quot;:&quot;Not Disclosed&quot;},{&quot;value&quot;:&quot;Owned&quot;,&quot;label&quot;:&quot;Owned&quot;}],&quot;origHeader&quot;:&quot;jurisdictionfourteeneyes&quot;,&quot;displayHeader&quot;:&quot;JURISDICTION Fourteen Eyes?&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Eyes&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;No&quot;,&quot;label&quot;:&quot;No&quot;},{&quot;value&quot;:&quot;Not Disclosed&quot;,&quot;label&quot;:&quot;Not Disclosed&quot;},{&quot;value&quot;:&quot;Owned&quot;,&quot;label&quot;:&quot;Owned&quot;},{&quot;value&quot;:&quot;Yes&quot;,&quot;label&quot;:&quot;Yes&quot;}],&quot;origHeader&quot;:&quot;jurisdictionfreedomstatus&quot;,&quot;displayHeader&quot;:&quot;JURISDICTION Enemy of the Internet&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Enemy&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;logginglogstraffic&quot;,&quot;displayHeader&quot;:&quot;LOGGING Logs Traffic&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;logginglogsdnsrequests&quot;,&quot;displayHeader&quot;:&quot;LOGGING Logs DNS Requests&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;logginglogstimestamps&quot;,&quot;displayHeader&quot;:&quot;LOGGING Logs Timestamps&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;logginglogsbandwidth&quot;,&quot;displayHeader&quot;:&quot;LOGGING Logs Bandwidth&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;logginglogsipaddress&quot;,&quot;displayHeader&quot;:&quot;LOGGING Logs IP Address&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;activismanonymouspaymentmethod&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM Anonymous Payment Method&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;activismacceptscash&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM Accepts Cash&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;activismacceptsgiftcards&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM Accepts Gift Cards&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;activismacceptsbitcoin&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM Accepts Crypto Currency&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;activismpgpkeyavailable&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM PGP Key Available&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;givesbacktoprivacycauses&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM Gives back to Privacy Causes&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;activismmeetsprivacytoolsiocriteria&quot;,&quot;displayHeader&quot;:&quot;ACTIVISM Meets PrivacyTools IO Criteria&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;leakprotection1stpartydnsservers&quot;,&quot;displayHeader&quot;:&quot;LEAK PROTECTION 1st Party DNS Servers&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;leakprotectionipv6supportedblocked&quot;,&quot;displayHeader&quot;:&quot;LEAK PROTECTION IPv6 Supported \\\/ Blocked&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;protocolsoffersopenvpn&quot;,&quot;displayHeader&quot;:&quot;PROTOCOLS Offers OpenVPN&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationsupportsmultihop&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Supports Multihop&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationsupportstcpport443&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Supports TCP Port 443&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationsupportsobfsproxy&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Supports Obfsproxy&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationsupportssocks&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Supports SOCKS&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationsupportsssltunnel&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Supports SSL Tunnel&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationsupportssshtunnel&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Supports SSH Tunnel&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;obfuscationotherproprietaryprotocols&quot;,&quot;displayHeader&quot;:&quot;OBFUSCATION Other Proprietary Protocols&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;No&quot;,&quot;label&quot;:&quot;No&quot;},{&quot;value&quot;:&quot;Some&quot;,&quot;label&quot;:&quot;Some&quot;},{&quot;value&quot;:&quot;Yes&quot;,&quot;label&quot;:&quot;Yes&quot;}],&quot;origHeader&quot;:&quot;portblockingauthsmtp&quot;,&quot;displayHeader&quot;:&quot;PORT BLOCKING Auth SMTP&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;SMTP&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;No&quot;,&quot;label&quot;:&quot;No&quot;},{&quot;value&quot;:&quot;Some&quot;,&quot;label&quot;:&quot;Some&quot;},{&quot;value&quot;:&quot;Yes&quot;,&quot;label&quot;:&quot;Yes&quot;}],&quot;origHeader&quot;:&quot;portblockingp2p&quot;,&quot;displayHeader&quot;:&quot;PORT BLOCKING P2P&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;P2P&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;speedsusserveraveragembps&quot;,&quot;displayHeader&quot;:&quot;SPEEDS US Server Average (%)&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;speedsinternationalserveraveragembps&quot;,&quot;displayHeader&quot;:&quot;SPEEDS Int\\u0027l Server Average (%)&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;serversdedicatedorvirtual&quot;,&quot;displayHeader&quot;:&quot;SERVERS Dedicated or Virtual&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;AES-128&quot;,&quot;label&quot;:&quot;AES-128&quot;},{&quot;value&quot;:&quot;AES-196&quot;,&quot;label&quot;:&quot;AES-196&quot;},{&quot;value&quot;:&quot;AES-256&quot;,&quot;label&quot;:&quot;AES-256&quot;},{&quot;value&quot;:&quot;Blowfish-128&quot;,&quot;label&quot;:&quot;Blowfish-128&quot;},{&quot;value&quot;:&quot;Blowfish-256&quot;,&quot;label&quot;:&quot;Blowfish-256&quot;},{&quot;value&quot;:&quot;Blowfish-512&quot;,&quot;label&quot;:&quot;Blowfish-512&quot;},{&quot;value&quot;:&quot;Blowfish-64&quot;,&quot;label&quot;:&quot;Blowfish-64&quot;}],&quot;origHeader&quot;:&quot;securityweakestdataencryption&quot;,&quot;displayHeader&quot;:&quot;SECURITY Default Data Encryption&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Def Enc&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;AES-128&quot;,&quot;label&quot;:&quot;AES-128&quot;},{&quot;value&quot;:&quot;AES-196&quot;,&quot;label&quot;:&quot;AES-196&quot;},{&quot;value&quot;:&quot;AES-256&quot;,&quot;label&quot;:&quot;AES-256&quot;},{&quot;value&quot;:&quot;Blowfish-128&quot;,&quot;label&quot;:&quot;Blowfish-128&quot;},{&quot;value&quot;:&quot;Blowfish-256&quot;,&quot;label&quot;:&quot;Blowfish-256&quot;},{&quot;value&quot;:&quot;Blowfish-512&quot;,&quot;label&quot;:&quot;Blowfish-512&quot;},{&quot;value&quot;:&quot;Blowfish-64&quot;,&quot;label&quot;:&quot;Blowfish-64&quot;},{&quot;value&quot;:&quot;MPPE-128&quot;,&quot;label&quot;:&quot;MPPE-128&quot;}],&quot;origHeader&quot;:&quot;securitystrongestdataencryption&quot;,&quot;displayHeader&quot;:&quot;SECURITY Strongest Data Encryption&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Str Enc&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;DH-2048&quot;,&quot;label&quot;:&quot;DH-2048&quot;},{&quot;value&quot;:&quot;MS-CHAPv2&quot;,&quot;label&quot;:&quot;MS-CHAPv2&quot;},{&quot;value&quot;:&quot;MSCHAP-v2&quot;,&quot;label&quot;:&quot;MSCHAP-v2&quot;},{&quot;value&quot;:&quot;RSA-1024&quot;,&quot;label&quot;:&quot;RSA-1024&quot;},{&quot;value&quot;:&quot;RSA-2048&quot;,&quot;label&quot;:&quot;RSA-2048&quot;},{&quot;value&quot;:&quot;RSA-4096&quot;,&quot;label&quot;:&quot;RSA-4096&quot;},{&quot;value&quot;:&quot;RSA-RC4&quot;,&quot;label&quot;:&quot;RSA-RC4&quot;}],&quot;origHeader&quot;:&quot;securityweakesthandshakeencryption&quot;,&quot;displayHeader&quot;:&quot;SECURITY Weakest Handshake Encryption&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Weak HS&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;CA-4096&quot;,&quot;label&quot;:&quot;CA-4096&quot;},{&quot;value&quot;:&quot;DH-2048&quot;,&quot;label&quot;:&quot;DH-2048&quot;},{&quot;value&quot;:&quot;MSCHAP-v2&quot;,&quot;label&quot;:&quot;MSCHAP-v2&quot;},{&quot;value&quot;:&quot;RSA-1024&quot;,&quot;label&quot;:&quot;RSA-1024&quot;},{&quot;value&quot;:&quot;RSA-2048&quot;,&quot;label&quot;:&quot;RSA-2048&quot;},{&quot;value&quot;:&quot;RSA-4096&quot;,&quot;label&quot;:&quot;RSA-4096&quot;}],&quot;origHeader&quot;:&quot;securitystrongesthandshakeencryption&quot;,&quot;displayHeader&quot;:&quot;SECURITY Strongest Handshake Encryption&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Str HS&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;availabilityofconnections&quot;,&quot;displayHeader&quot;:&quot;AVAILABILITY # of Connections&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;availabilityofcountries&quot;,&quot;displayHeader&quot;:&quot;AVAILABILITY # of Countries&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;availabilityofservers&quot;,&quot;displayHeader&quot;:&quot;AVAILABILITY # of Servers&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;No&quot;,&quot;label&quot;:&quot;No&quot;},{&quot;value&quot;:&quot;Partial&quot;,&quot;label&quot;:&quot;Partial&quot;},{&quot;value&quot;:&quot;Yes&quot;,&quot;label&quot;:&quot;Yes&quot;}],&quot;origHeader&quot;:&quot;supportlinuxmanualconfig&quot;,&quot;displayHeader&quot;:&quot;SUPPORT Linux (Manual Config)&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Linux&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;websiteofpersistentcookies&quot;,&quot;displayHeader&quot;:&quot;WEBSITE # of Persistent Cookies&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;websiteofexternaltrackers&quot;,&quot;displayHeader&quot;:&quot;WEBSITE # of External Trackers&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;websiteofproprietaryapis&quot;,&quot;displayHeader&quot;:&quot;WEBSITE # of Proprietary APIs&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;A&quot;,&quot;label&quot;:&quot;A&quot;},{&quot;value&quot;:&quot;A+&quot;,&quot;label&quot;:&quot;A+&quot;},{&quot;value&quot;:&quot;A-&quot;,&quot;label&quot;:&quot;A-&quot;},{&quot;value&quot;:&quot;B&quot;,&quot;label&quot;:&quot;B&quot;},{&quot;value&quot;:&quot;C&quot;,&quot;label&quot;:&quot;C&quot;},{&quot;value&quot;:&quot;F&quot;,&quot;label&quot;:&quot;F&quot;},{&quot;value&quot;:&quot;T&quot;,&quot;label&quot;:&quot;T&quot;}],&quot;origHeader&quot;:&quot;websiteserversslrating&quot;,&quot;displayHeader&quot;:&quot;WEBSITE Server SSL Rating&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;SSL&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;checkbox&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:[{&quot;value&quot;:&quot;CloudFlare&quot;,&quot;label&quot;:&quot;CloudFlare&quot;},{&quot;value&quot;:&quot;Incapsula&quot;,&quot;label&quot;:&quot;Incapsula&quot;},{&quot;value&quot;:&quot;No SSL Cert&quot;,&quot;label&quot;:&quot;No SSL Cert&quot;},{&quot;value&quot;:&quot;Self&quot;,&quot;label&quot;:&quot;Self&quot;}],&quot;origHeader&quot;:&quot;websitesslcertissuedto&quot;,&quot;displayHeader&quot;:&quot;WEBSITE SSL Cert issued to&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;Cert&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;pricingmonthannualpricing&quot;,&quot;displayHeader&quot;:&quot;PRICING $ \\\/ Month (Annual Pricing)&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;pricingconnectionmonth&quot;,&quot;displayHeader&quot;:&quot;PRICING $ \\\/ Connection \\\/ Month&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;pricingfreetrial&quot;,&quot;displayHeader&quot;:&quot;PRICING Free Trial&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;number-range&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;pricingrefundperioddays&quot;,&quot;displayHeader&quot;:&quot;PRICING Refund Period (Days)&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:-1,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;ethicscontradictoryloggingpolicies&quot;,&quot;displayHeader&quot;:&quot;ETHICS Contradictory Logging Policies&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;ethicsfalselyclaims100effective&quot;,&quot;displayHeader&quot;:&quot;ETHICS Falsely Claims 100% Effective&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;ethicsincentivizessocialmediaspam&quot;,&quot;displayHeader&quot;:&quot;ETHICS Incentivizes Social Media Spam&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;policiesforbidsspam&quot;,&quot;displayHeader&quot;:&quot;POLICIES Forbids Spam&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;policiesrequiresethicalcopy&quot;,&quot;displayHeader&quot;:&quot;POLICIES Requires Ethical Copy&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;policiesrequiresfulldisclosure&quot;,&quot;displayHeader&quot;:&quot;POLICIES Requires Full Disclosure&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;affiliatespracticeethicalcopy&quot;,&quot;displayHeader&quot;:&quot;AFFILIATES Practice Ethical Copy&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;},{&quot;type&quot;:&quot;text&quot;,&quot;possibleValuesType&quot;:&quot;read&quot;,&quot;values&quot;:null,&quot;origHeader&quot;:&quot;affiliatesgivefulldisclosure&quot;,&quot;displayHeader&quot;:&quot;AFFILIATES Give Full Disclosure&quot;,&quot;possibleValuesAddEmpty&quot;:false,&quot;possibleValuesAjax&quot;:10,&quot;defaultValue&quot;:&quot;&quot;,&quot;exactFiltering&quot;:false,&quot;filterLabel&quot;:&quot;&quot;,&quot;checkboxesInModal&quot;:0,&quot;linkButtonLabel&quot;:&quot;&quot;}],&quot;bUseColVis&quot;:true},&quot;filterInForm&quot;:false,&quot;serverSide&quot;:true,&quot;autoRefreshInterval&quot;:0,&quot;processing&quot;:true,&quot;fnServerData&quot;:true,&quot;columnsFixed&quot;:0,&quot;sumFunctionsLabel&quot;:&quot;&quot;,&quot;avgFunctionsLabel&quot;:&quot;&quot;,&quot;minFunctionsLabel&quot;:&quot;&quot;,&quot;maxFunctionsLabel&quot;:&quot;&quot;,&quot;columnsDecimalPlaces&quot;:{&quot;wdt_ID&quot;:-1,&quot;vpnservice&quot;:-1,&quot;jurisdictionbasedincountry&quot;:-1,&quot;jurisdictionfourteeneyes&quot;:-1,&quot;jurisdictionfreedomstatus&quot;:-1,&quot;logginglogstraffic&quot;:-1,&quot;logginglogsdnsrequests&quot;:-1,&quot;logginglogstimestamps&quot;:-1,&quot;logginglogsbandwidth&quot;:-1,&quot;logginglogsipaddress&quot;:-1,&quot;activismanonymouspaymentmethod&quot;:-1,&quot;activismacceptscash&quot;:-1,&quot;activismacceptsgiftcards&quot;:-1,&quot;activismacceptsbitcoin&quot;:-1,&quot;activismpgpkeyavailable&quot;:-1,&quot;givesbacktoprivacycauses&quot;:-1,&quot;activismmeetsprivacytoolsiocriteria&quot;:-1,&quot;leakprotection1stpartydnsservers&quot;:-1,&quot;leakprotectionipv6supportedblocked&quot;:-1,&quot;protocolsoffersopenvpn&quot;:-1,&quot;obfuscationsupportsmultihop&quot;:-1,&quot;obfuscationsupportstcpport443&quot;:-1,&quot;obfuscationsupportsobfsproxy&quot;:-1,&quot;obfuscationsupportssocks&quot;:-1,&quot;obfuscationsupportsssltunnel&quot;:-1,&quot;obfuscationsupportssshtunnel&quot;:-1,&quot;obfuscationotherproprietaryprotocols&quot;:-1,&quot;portblockingauthsmtp&quot;:-1,&quot;portblockingp2p&quot;:-1,&quot;speedsusserveraveragembps&quot;:-1,&quot;speedsinternationalserveraveragembps&quot;:-1,&quot;serversdedicatedorvirtual&quot;:-1,&quot;securityweakestdataencryption&quot;:-1,&quot;securitystrongestdataencryption&quot;:-1,&quot;securityweakesthandshakeencryption&quot;:-1,&quot;securitystrongesthandshakeencryption&quot;:-1,&quot;availabilityofconnections&quot;:-1,&quot;availabilityofcountries&quot;:-1,&quot;availabilityofservers&quot;:-1,&quot;supportlinuxmanualconfig&quot;:-1,&quot;websiteofpersistentcookies&quot;:-1,&quot;websiteofexternaltrackers&quot;:-1,&quot;websiteofproprietaryapis&quot;:-1,&quot;websiteserversslrating&quot;:-1,&quot;websitesslcertissuedto&quot;:-1,&quot;pricingmonthannualpricing&quot;:-1,&quot;pricingconnectionmonth&quot;:-1,&quot;pricingfreetrial&quot;:-1,&quot;pricingrefundperioddays&quot;:-1,&quot;ethicscontradictoryloggingpolicies&quot;:-1,&quot;ethicsfalselyclaims100effective&quot;:-1,&quot;ethicsincentivizessocialmediaspam&quot;:-1,&quot;policiesforbidsspam&quot;:-1,&quot;policiesrequiresethicalcopy&quot;:-1,&quot;policiesrequiresfulldisclosure&quot;:-1,&quot;affiliatespracticeethicalcopy&quot;:-1,&quot;affiliatesgivefulldisclosure&quot;:-1},&quot;columnsThousandsSeparator&quot;:{&quot;wdt_ID&quot;:0,&quot;availabilityofconnections&quot;:1,&quot;availabilityofcountries&quot;:1,&quot;availabilityofservers&quot;:1,&quot;websiteofpersistentcookies&quot;:1,&quot;websiteofexternaltrackers&quot;:1,&quot;websiteofproprietaryapis&quot;:1,&quot;pricingrefundperioddays&quot;:1},&quot;sumColumns&quot;:[],&quot;avgColumns&quot;:[],&quot;sumAvgColumns&quot;:[],&quot;conditional_formatting_columns&quot;:[&quot;jurisdictionfourteeneyes&quot;,&quot;jurisdictionfreedomstatus&quot;,&quot;logginglogstraffic&quot;,&quot;logginglogsdnsrequests&quot;,&quot;logginglogstimestamps&quot;,&quot;logginglogsbandwidth&quot;,&quot;logginglogsipaddress&quot;,&quot;activismanonymouspaymentmethod&quot;,&quot;activismacceptscash&quot;,&quot;activismacceptsgiftcards&quot;,&quot;activismacceptsbitcoin&quot;,&quot;activismpgpkeyavailable&quot;,&quot;givesbacktoprivacycauses&quot;,&quot;activismmeetsprivacytoolsiocriteria&quot;,&quot;leakprotection1stpartydnsservers&quot;,&quot;leakprotectionipv6supportedblocked&quot;,&quot;protocolsoffersopenvpn&quot;,&quot;obfuscationsupportsmultihop&quot;,&quot;obfuscationsupportstcpport443&quot;,&quot;obfuscationsupportsobfsproxy&quot;,&quot;obfuscationsupportssocks&quot;,&quot;obfuscationsupportsssltunnel&quot;,&quot;obfuscationsupportssshtunnel&quot;,&quot;obfuscationotherproprietaryprotocols&quot;,&quot;portblockingauthsmtp&quot;,&quot;portblockingp2p&quot;,&quot;speedsusserveraveragembps&quot;,&quot;speedsinternationalserveraveragembps&quot;,&quot;serversdedicatedorvirtual&quot;,&quot;securityweakestdataencryption&quot;,&quot;securitystrongestdataencryption&quot;,&quot;securityweakesthandshakeencryption&quot;,&quot;securitystrongesthandshakeencryption&quot;,&quot;availabilityofconnections&quot;,&quot;availabilityofcountries&quot;,&quot;availabilityofservers&quot;,&quot;supportlinuxmanualconfig&quot;,&quot;websiteofpersistentcookies&quot;,&quot;websiteofexternaltrackers&quot;,&quot;websiteofproprietaryapis&quot;,&quot;websiteserversslrating&quot;,&quot;websitesslcertissuedto&quot;,&quot;pricingmonthannualpricing&quot;,&quot;pricingconnectionmonth&quot;,&quot;pricingfreetrial&quot;,&quot;pricingrefundperioddays&quot;,&quot;ethicscontradictoryloggingpolicies&quot;,&quot;ethicsfalselyclaims100effective&quot;,&quot;ethicsincentivizessocialmediaspam&quot;,&quot;policiesforbidsspam&quot;,&quot;policiesrequiresethicalcopy&quot;,&quot;policiesrequiresfulldisclosure&quot;,&quot;affiliatespracticeethicalcopy&quot;,&quot;affiliatesgivefulldisclosure&quot;],&quot;timeFormat&quot;:&quot;h:i A&quot;,&quot;datepickFormat&quot;:&quot;dd\\\/mm\\\/yy&quot;}\" \/><table id=\"table_1\" class=\"scroll responsive display nowrap data-t data-t wpDataTable\" data-described-by=\"table_1_desc\" data-wpdatatable_id=\"17\"><thead><tr><th class=\"wdtheader sort numdata integer\">wdt_ID<\/th>        <th data-class=\"expand\" class=\"wdtheader sort\">VPN SERVICE<\/th>        <th class=\"wdtheader sort\">JURISDICTION Based In (Country)<\/th>        <th class=\"wdtheader sort\">JURISDICTION Fourteen Eyes?<\/th>        <th class=\"wdtheader sort\">JURISDICTION Enemy of the Internet<\/th>        <th class=\"wdtheader sort\">LOGGING Logs Traffic<\/th>        <th class=\"wdtheader sort\">LOGGING Logs DNS Requests<\/th>        <th class=\"wdtheader sort\">LOGGING Logs Timestamps<\/th>        <th class=\"wdtheader sort\">LOGGING Logs Bandwidth<\/th>        <th class=\"wdtheader sort\">LOGGING Logs IP Address<\/th>        <th class=\"wdtheader sort\">ACTIVISM Anonymous Payment Method<\/th>        <th class=\"wdtheader sort\">ACTIVISM Accepts Cash<\/th>        <th class=\"wdtheader sort\">ACTIVISM Accepts Gift Cards<\/th>        <th class=\"wdtheader sort\">ACTIVISM Accepts Crypto Currency<\/th>        <th class=\"wdtheader sort\">ACTIVISM PGP Key Available<\/th>        <th class=\"wdtheader sort\">ACTIVISM Gives back to Privacy Causes<\/th>        <th class=\"wdtheader sort\">ACTIVISM Meets PrivacyTools IO Criteria<\/th>        <th class=\"wdtheader sort\">LEAK PROTECTION 1st Party DNS Servers<\/th>        <th class=\"wdtheader sort\">LEAK PROTECTION IPv6 Supported \/ Blocked<\/th>        <th class=\"wdtheader sort\">PROTOCOLS Offers OpenVPN<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Supports Multihop<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Supports TCP Port 443<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Supports Obfsproxy<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Supports SOCKS<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Supports SSL Tunnel<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Supports SSH Tunnel<\/th>        <th class=\"wdtheader sort\">OBFUSCATION Other Proprietary Protocols<\/th>        <th class=\"wdtheader sort\">PORT BLOCKING Auth SMTP<\/th>        <th class=\"wdtheader sort\">PORT BLOCKING P2P<\/th>        <th class=\"wdtheader sort numdata float\">SPEEDS US Server Average (%)<\/th>        <th class=\"wdtheader sort numdata float\">SPEEDS Int'l Server Average (%)<\/th>        <th class=\"wdtheader sort\">SERVERS Dedicated or Virtual<\/th>        <th class=\"wdtheader sort\">SECURITY Default Data Encryption<\/th>        <th class=\"wdtheader sort\">SECURITY Strongest Data Encryption<\/th>        <th class=\"wdtheader sort\">SECURITY Weakest Handshake Encryption<\/th>        <th class=\"wdtheader sort\">SECURITY Strongest Handshake Encryption<\/th>        <th class=\"wdtheader sort numdata integer\">AVAILABILITY # of Connections<\/th>        <th class=\"wdtheader sort numdata integer\">AVAILABILITY # of Countries<\/th>        <th class=\"wdtheader sort numdata integer\">AVAILABILITY # of Servers<\/th>        <th class=\"wdtheader sort\">SUPPORT Linux (Manual Config)<\/th>        <th class=\"wdtheader sort numdata integer\">WEBSITE # of Persistent Cookies<\/th>        <th class=\"wdtheader sort numdata integer\">WEBSITE # of External Trackers<\/th>        <th class=\"wdtheader sort numdata integer\">WEBSITE # of Proprietary APIs<\/th>        <th class=\"wdtheader sort\">WEBSITE Server SSL Rating<\/th>        <th class=\"wdtheader sort\">WEBSITE SSL Cert issued to<\/th>        <th class=\"wdtheader sort numdata float\">PRICING $ \/ Month (Annual Pricing)<\/th>        <th class=\"wdtheader sort numdata float\">PRICING $ \/ Connection \/ Month<\/th>        <th class=\"wdtheader sort\">PRICING Free Trial<\/th>        <th class=\"wdtheader sort numdata integer\">PRICING Refund Period (Days)<\/th>        <th class=\"wdtheader sort\">ETHICS Contradictory Logging Policies<\/th>        <th class=\"wdtheader sort\">ETHICS Falsely Claims 100% Effective<\/th>        <th class=\"wdtheader sort\">ETHICS Incentivizes Social Media Spam<\/th>        <th class=\"wdtheader sort\">POLICIES Forbids Spam<\/th>        <th class=\"wdtheader sort\">POLICIES Requires Ethical Copy<\/th>        <th class=\"wdtheader sort\">POLICIES Requires Full Disclosure<\/th>        <th class=\"wdtheader sort\">AFFILIATES Practice Ethical Copy<\/th>        <th class=\"wdtheader sort\">AFFILIATES Give Full Disclosure<\/th>    <\/tr><\/thead><tbody><tr id=\"table_17_row_0\"><td>194<\/td>\n                    <td>Doublehop<\/td>\n                    <td>Seychelles<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Email<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>Dedicated<\/td>\n                    <td>AES-256<\/td>\n                    <td>AES-256<\/td>\n                    <td>RSA-4096<\/td>\n                    <td>RSA-4096<\/td>\n                    <td>6<\/td>\n                    <td>5<\/td>\n                    <td>0<\/td>\n                    <td>Yes<\/td>\n                    <td>1<\/td>\n                    <td>0<\/td>\n                    <td>0<\/td>\n                    <td>A+<\/td>\n                    <td>CloudFlare<\/td>\n                    <td>2.75<\/td>\n                    <td>0.45<\/td>\n                    <td>No<\/td>\n                    <td>365<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n            <\/td><\/tr><tr id=\"table_17_row_1\"><td>2<\/td>\n                    <td>AceVPN<\/td>\n                    <td>USA<\/td>\n                    <td>Five<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>\n                    <\/td><td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>Some<\/td>\n                    <td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>2<\/td>\n                    <td>22<\/td>\n                    <td>0<\/td>\n                    <td>\n                    <\/td><td>4<\/td>\n                    <td>1<\/td>\n                    <td>10<\/td>\n                    <td>A<\/td>\n                    <td>Self<\/td>\n                    <td>4.59<\/td>\n                    <td>2.30<\/td>\n                    <td>No<\/td>\n                    <td>7<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n            <\/td><\/tr><tr id=\"table_17_row_2\"><td>3<\/td>\n                    <td>ActiVPN<\/td>\n                    <td>France<\/td>\n                    <td>Nine<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Email<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>5<\/td>\n                    <td>13<\/td>\n                    <td>20<\/td>\n                    <td>\n                    <\/td><td>0<\/td>\n                    <td>0<\/td>\n                    <td>0<\/td>\n                    <td>A+<\/td>\n                    <td>Self<\/td>\n                    <td>3.34<\/td>\n                    <td>0.67<\/td>\n                    <td>No<\/td>\n                    <td>0<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n            <\/td><\/tr><tr id=\"table_17_row_3\"><td>4<\/td>\n                    <td>AirVPN<\/td>\n                    <td>Italy<\/td>\n                    <td>Fourteen<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Email<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>90.56<\/td>\n                    <td>5.99<\/td>\n                    <td>\n                    <\/td><td>AES-256<\/td>\n                    <td>AES-256<\/td>\n                    <td>RSA-4096<\/td>\n                    <td>RSA-4096<\/td>\n                    <td>5<\/td>\n                    <td>19<\/td>\n                    <td>215<\/td>\n                    <td>Yes<\/td>\n                    <td>3<\/td>\n                    <td>0<\/td>\n                    <td>0<\/td>\n                    <td>A+<\/td>\n                    <td>Self<\/td>\n                    <td>4.93<\/td>\n                    <td>0.99<\/td>\n                    <td>Yes<\/td>\n                    <td>3<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n            <\/tr><tr id=\"table_17_row_4\"><td>5<\/td>\n                    <td>Anonine<\/td>\n                    <td>Seychelles<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Email<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>5<\/td>\n                    <td>22<\/td>\n                    <td>80<\/td>\n                    <td>\n                    <\/td><td>6<\/td>\n                    <td>2<\/td>\n                    <td>29<\/td>\n                    <td>C<\/td>\n                    <td>Self<\/td>\n                    <td>3.99<\/td>\n                    <td>0.80<\/td>\n                    <td>No<\/td>\n                    <td>0<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n            <\/td><\/tr><tr id=\"table_17_row_5\"><td>6<\/td>\n                    <td>Anonymizer<\/td>\n                    <td>USA<\/td>\n                    <td>Five<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>Email<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>6<\/td>\n                    <td>0<\/td>\n                    <td>0<\/td>\n                    <td>\n                    <\/td><td>14<\/td>\n                    <td>4<\/td>\n                    <td>19<\/td>\n                    <td>A-<\/td>\n                    <td>Self<\/td>\n                    <td>6.67<\/td>\n                    <td>1.11<\/td>\n                    <td>Yes<\/td>\n                    <td>45<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n            <\/td><\/tr><tr id=\"table_17_row_6\"><td>7<\/td>\n                    <td>Astrill<\/td>\n                    <td>Seychelles<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>\n                    <\/td><td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Some<\/td>\n                    <td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>AES-128<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>2<\/td>\n                    <td>54<\/td>\n                    <td>254<\/td>\n                    <td>\n                    <\/td><td>4<\/td>\n                    <td>1<\/td>\n                    <td>2<\/td>\n                    <td>A-<\/td>\n                    <td>Self<\/td>\n                    <td>5.83<\/td>\n                    <td>2.92<\/td>\n                    <td>Yes<\/td>\n                    <td>7<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n            <\/tr><tr id=\"table_17_row_7\"><td>9<\/td>\n                    <td>Avast Secureline<\/td>\n                    <td>Czech Republic<\/td>\n                    <td>Cooperative<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>AES-256<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>1<\/td>\n                    <td>19<\/td>\n                    <td>27<\/td>\n                    <td>\n                    <\/td><td>0<\/td>\n                    <td>0<\/td>\n                    <td>16<\/td>\n                    <td>A<\/td>\n                    <td>Self<\/td>\n                    <td>6.67<\/td>\n                    <td>6.67<\/td>\n                    <td>Yes<\/td>\n                    <td>30<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Yes<\/td>\n                    <td>Some<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>No<\/td>\n            <\/tr><tr id=\"table_17_row_8\"><td>10<\/td>\n                    <td>AzireVPN<\/td>\n                    <td>Sweden<\/td>\n                    <td>Fourteen<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>Yes<\/td>\n                    <td>\n                    <\/td><td>Yes<\/td>\n                    <td>\n                    <\/td><td>Yes<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>Dedicated<\/td>\n                    <td>AES-256<\/td>\n                    <td>AES-256<\/td>\n                    <td>RSA-4096<\/td>\n                    <td>RSA-4096<\/td>\n                    <td>5<\/td>\n                    <td>4<\/td>\n                    <td>26<\/td>\n                    <td>Yes<\/td>\n                    <td>1<\/td>\n                    <td>0<\/td>\n                    <td>0<\/td>\n                    <td>A+<\/td>\n                    <td>Self<\/td>\n                    <td>4.11<\/td>\n                    <td>0.82<\/td>\n                    <td>Yes<\/td>\n                    <td>7<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>No<\/td>\n            <\/tr><tr id=\"table_17_row_9\"><td>11<\/td>\n                    <td>BeeVPN<\/td>\n                    <td>Denmark<\/td>\n                    <td>Nine<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>Email<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>Yes<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>No<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>No<\/td>\n                    <td>0.00<\/td>\n                    <td>0.00<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>AES-256<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>1<\/td>\n                    <td>4<\/td>\n                    <td>0<\/td>\n                    <td>\n                    <\/td><td>4<\/td>\n                    <td>1<\/td>\n                    <td>12<\/td>\n                    <td>C<\/td>\n                    <td>Self<\/td>\n                    <td>7.17<\/td>\n                    <td>7.17<\/td>\n                    <td>No<\/td>\n                    <td>0<\/td>\n                    <td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n                    <\/td><td>\n            <\/td><\/tr><\/tbody><tfoot><tr><td class=\"wdtheader sort numdata integer\">                <\/td><td class=\"wdtheader sort\">VPN SERVICE<\/td>                <td class=\"wdtheader sort\">JURISDICTION Based In (Country)<\/td>                <td class=\"wdtheader sort\">JURISDICTION Fourteen Eyes?<\/td>                <td class=\"wdtheader sort\">JURISDICTION Enemy of the Internet<\/td>                <td class=\"wdtheader sort\">LOGGING Logs Traffic<\/td>                <td class=\"wdtheader sort\">LOGGING Logs DNS Requests<\/td>                <td class=\"wdtheader sort\">LOGGING Logs Timestamps<\/td>                <td class=\"wdtheader sort\">LOGGING Logs Bandwidth<\/td>                <td class=\"wdtheader sort\">LOGGING Logs IP Address<\/td>                <td class=\"wdtheader sort\">ACTIVISM Anonymous Payment Method<\/td>                <td class=\"wdtheader sort\">ACTIVISM Accepts Cash<\/td>                <td class=\"wdtheader sort\">ACTIVISM Accepts Gift Cards<\/td>                <td class=\"wdtheader sort\">ACTIVISM Accepts Crypto Currency<\/td>                <td class=\"wdtheader sort\">ACTIVISM PGP Key Available<\/td>                <td class=\"wdtheader sort\">ACTIVISM Gives back to Privacy Causes<\/td>                <td class=\"wdtheader sort\">ACTIVISM Meets PrivacyTools IO Criteria<\/td>                <td class=\"wdtheader sort\">LEAK PROTECTION 1st Party DNS Servers<\/td>                <td class=\"wdtheader sort\">LEAK PROTECTION IPv6 Supported \/ Blocked<\/td>                <td class=\"wdtheader sort\">PROTOCOLS Offers OpenVPN<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Supports Multihop<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Supports TCP Port 443<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Supports Obfsproxy<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Supports SOCKS<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Supports SSL Tunnel<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Supports SSH Tunnel<\/td>                <td class=\"wdtheader sort\">OBFUSCATION Other Proprietary Protocols<\/td>                <td class=\"wdtheader sort\">PORT BLOCKING Auth SMTP<\/td>                <td class=\"wdtheader sort\">PORT BLOCKING P2P<\/td>                <td class=\"wdtheader sort numdata float\">SPEEDS US Server Average (%)<\/td>                <td class=\"wdtheader sort numdata float\">SPEEDS Int'l Server Average (%)<\/td>                <td class=\"wdtheader sort\">SERVERS Dedicated or Virtual<\/td>                <td class=\"wdtheader sort\">SECURITY Default Data Encryption<\/td>                <td class=\"wdtheader sort\">SECURITY Strongest Data Encryption<\/td>                <td class=\"wdtheader sort\">SECURITY Weakest Handshake Encryption<\/td>                <td class=\"wdtheader sort\">SECURITY Strongest Handshake Encryption<\/td>                <td class=\"wdtheader sort numdata integer\">AVAILABILITY # of Connections<\/td>                <td class=\"wdtheader sort numdata integer\">AVAILABILITY # of Countries<\/td>                <td class=\"wdtheader sort numdata integer\">AVAILABILITY # of Servers<\/td>                <td class=\"wdtheader sort\">SUPPORT Linux (Manual Config)<\/td>                <td class=\"wdtheader sort numdata integer\">WEBSITE # of Persistent Cookies<\/td>                <td class=\"wdtheader sort numdata integer\">WEBSITE # of External Trackers<\/td>                <td class=\"wdtheader sort numdata integer\">WEBSITE # of Proprietary APIs<\/td>                <td class=\"wdtheader sort\">WEBSITE Server SSL Rating<\/td>                <td class=\"wdtheader sort\">WEBSITE SSL Cert issued to<\/td>                <td class=\"wdtheader sort numdata float\">PRICING $ \/ Month (Annual Pricing)<\/td>                <td class=\"wdtheader sort numdata float\">PRICING $ \/ Connection \/ Month<\/td>                <td class=\"wdtheader sort\">PRICING Free Trial<\/td>                <td class=\"wdtheader sort numdata integer\">PRICING Refund Period (Days)<\/td>                <td class=\"wdtheader sort\">ETHICS Contradictory Logging Policies<\/td>                <td class=\"wdtheader sort\">ETHICS Falsely Claims 100% Effective<\/td>                <td class=\"wdtheader sort\">ETHICS Incentivizes Social Media Spam<\/td>                <td class=\"wdtheader sort\">POLICIES Forbids Spam<\/td>                <td class=\"wdtheader sort\">POLICIES Requires Ethical Copy<\/td>                <td class=\"wdtheader sort\">POLICIES Requires Full Disclosure<\/td>                <td class=\"wdtheader sort\">AFFILIATES Practice Ethical Copy<\/td>                <td class=\"wdtheader sort\">AFFILIATES Give Full Disclosure<\/td>        <\/tr><\/tfoot><\/table><\/div>\n\n\n\n<p>PLEASE BE PATIENT AS IT MAY TAKE SOME TIME FOR THE SHEET TO LOAD<\/p>\n<p>Green = \u201cGenerally good\u201d, Yellow = \u201cSomething of concern\u201d, Red = \u201cSomething major of concern\u201d, Blank = Undefined or for reader\u2019s knowledge only.<\/p>\n\n<p>See what these fields mean <a href=\"https:\/\/thatoneprivacysite.net\/vpn-glossary\/\">here<\/a>.<\/p>\n\n<p>You can download the VPN Comparison Chart in other formats below:<\/p>\n<p>Excel \u2013 <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1L72gHJ5bTq0Djljz0P-NCAaURrXwsR1MsLpVmAt3bwg\/export?format=xlsx\">xlsx<\/a> | LibreOffice \u2013 <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1L72gHJ5bTq0Djljz0P-NCAaURrXwsR1MsLpVmAt3bwg\/export?format=ods\">ods<\/a> | CSV \u2013 <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1L72gHJ5bTq0Djljz0P-NCAaURrXwsR1MsLpVmAt3bwg\/export?format=csv\">csv<\/a><\/p>\n\n<p>Color blind?\u00a0 There is a version of the table for you <a href=\"https:\/\/thatoneprivacysite.net\/detailed-vpn-comparison-chart-for-the-color-blind\">here<\/a>.<\/p>\n\n<p>I make every effort to keep the data on the VPN Comparison Chart up to date.\u00a0 However, parts of it could potentially be incorrect for various reasons \u2013 including if a given VPN service is not transparent and does not make the data available on their official site.\u00a0 Please do your own research if this is of concern.\u00a0 Note also that if a company does not make available certain data on their website, I may assume the worst and a default field will be used.\u00a0 This in no way means I am making a claim about a service, but rather noting the data they lack.<\/p>\n<p>If you are with a VPN company and you\u2019d like to request an update to your information, feel free to contact me citing a proper source on the official website and I will be happy to update your data!<\/p>\n\n<p><em>(Data last updated on 8\/13\/18)<\/em><\/p>\n\n<p><em>If you like the project and find my work useful, please consider <a href=\"https:\/\/thatoneprivacysite.net\/donate\/\">donating<\/a> \u2013 your generous contributions help pay for the hosting, tools, and time I need to do my research and keep the data fresh.<br \/><\/em><\/p>","created_at":"2018-11-07T16:52:33+0000","updated_at":"2018-11-07T16:52:33+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":5,"domain_name":"thatoneprivacysite.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":583,"title":"Privacy Tools | Encryption against global mass surveillance","url":"https:\/\/www.privacytools.io\/","content":"<div class=\"jumbotron p-5\"><p class=\"lead\">You are being watched. Private and state-sponsored organizations are monitoring and recording your online activities. privacytools.io provides knowledge and tools to protect your privacy against global mass surveillance.<\/p><\/div><p>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"Translations of privacytools.io were done by volunteers. They might be not up to date.\">Language:<\/a>\n <a href=\"https:\/\/privacytools.twngo.xyz\/\">\u7e41\u9ad4\u4e2d\u6587<\/a>\n <a href=\"https:\/\/victorhck.gitlab.io\/privacytools-es\/\">Espa\u00f1ol<\/a>\n <a href=\"https:\/\/privacytools.it-sec.rocks\/\">Deutsch<\/a>\n <a href=\"https:\/\/privacytools-it.github.io\/\">Italiano<\/a>\n <a href=\"https:\/\/privacytools.ru\">\u0420\u0443\u0441\u0441\u043a\u0438\u0439<\/a>\n<\/p><blockquote class=\"blockquote\">\n<p>\n<a href=\"http:\/\/www.ted.com\/talks\/glenn_greenwald_why_privacy_matters\" title=\"Glenn Greenwald - Why privacy matters - TED Talk\"><img src=\"https:\/\/www.privacytools.io\/img\/layout\/Glenn-Greenwald-Why-privacy-matters.jpg\" class=\"img-fluid float-right ml-2\" alt=\"Glenn Greenwald: Why privacy matters\" \/><\/a>\nOver the last 16 months, as I've debated this issue around the world, every single time somebody has said to me, \"I don't really worry about invasions of privacy because I don't have anything to hide.\" I always say the same thing to them. I get out a\npen, I write down my email address. I say, \"Here's my email address. What I want you to do when you get home is email me the passwords to all of your email accounts, not just the nice, respectable work one in your name, but all of them, because I\nwant to be able to just troll through what it is you're doing online, read what I want to read and publish whatever I find interesting. After all, if you're not a bad person, if you're doing nothing wrong, you should have nothing to hide.\" <strong>Not a single person has taken me up on that offer.<\/strong><\/p>\n<\/blockquote><h4>Read also:<\/h4><ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Nothing_to_hide_argument\">Nothing to hide argument (Wikipedia)<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/privacy\/comments\/3hynvp\/how_do_you_counter_the_i_have_nothing_to_hide\/\">How do you counter the \"I have nothing to hide?\" argument? (reddit.com)<\/a><\/li>\n<li><a href=\"https:\/\/papers.ssrn.com\/sol3\/papers.cfm?abstract_id=998565\">'I've Got Nothing to Hide' and Other Misunderstandings of Privacy (Daniel J. Solove - San Diego Law Review)<\/a><\/li>\n<\/ul><img src=\"https:\/\/www.privacytools.io\/img\/layout\/UKUSA.png\" class=\"img-fluid float-right\" alt=\"UKUSA Agreement\" \/><p>The UKUSA Agreement is an agreement between the United Kingdom, United States, Australia, Canada, and New Zealand to cooperatively collect, analyze, and share intelligence. Members of this group, known as the <a href=\"http:\/\/www.giswatch.org\/en\/communications-surveillance\/unmasking-five-eyes-global-surveillance-practices\">Five Eyes<\/a>,\nfocus on gathering and analyzing intelligence from different parts of the world. While Five Eyes countries have agreed to <a href=\"http:\/\/www.pbs.org\/newshour\/rundown\/an-exclusive-club-the-five-countries-that-dont-spy-on-each-other\/\">not spy on each other<\/a> as adversaries, leaks by Snowden have revealed that some Five Eyes members monitor each other\u2019s citizens and <a href=\"http:\/\/www.theguardian.com\/uk\/2013\/jun\/21\/gchq-cables-secret-world-communications-nsa\">share intelligence<\/a> to <a href=\"http:\/\/www.theguardian.com\/politics\/2013\/jun\/10\/nsa-offers-intelligence-british-counterparts-blunkett\">avoid breaking domestic laws<\/a> that prohibit them from spying on their own citizens. The Five Eyes alliance also cooperates with groups of third-party countries to share intelligence (forming the Nine Eyes and Fourteen Eyes), however Five Eyes and third-party countries can and do\nspy on each other.<\/p><div class=\"row mb-3\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-danger\"><p><\/p><h3 class=\"h5\">Five Eyes<\/h3><div class=\"card-body\">1. Australia<br \/>2. Canada<br \/>3. New Zealand<br \/>4. United Kingdom<br \/>5. United States of America<\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Nine Eyes<\/h3><div class=\"card-body\">6. Denmark<br \/>7. France<br \/>8. Netherlands<br \/>9. Norway<\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-secondary\"><p><\/p><h3 class=\"h5\">Fourteen Eyes<\/h3><div class=\"card-body\">10. Belgium<br \/>11. Germany<br \/>12. Italy<br \/>13. Spain<br \/>14. Sweden<\/div><\/div><\/div><\/div><h3 id=\"usa\" class=\"anchor\">Why is it not recommended to choose a US-based service?<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/layout\/USA.png\" class=\"img-fluid float-right\" alt=\"USA\" \/><p>Services based in the United States are not recommended because of the country\u2019s surveillance programs, use of <a href=\"https:\/\/www.eff.org\/issues\/national-security-letters\/faq\">National Security Letters<\/a> (NSLs) and accompanying gag orders, which\nforbid the recipient from talking about the request. This combination allows the government to <a href=\"https:\/\/www.schneier.com\/blog\/archives\/2013\/08\/more_on_the_nsa.html\">secretly force<\/a> companies to grant complete access to customer data and\ntransform the service into a tool of mass surveillance.<\/p><p>An example of this is <a href=\"http:\/\/en.wikipedia.org\/wiki\/Lavabit#Suspension_and_gag_order\">Lavabit<\/a> \u2013 a discontinued secure email service created by Ladar Levison. The FBI <a href=\"https:\/\/motherboard.vice.com\/en_us\/article\/nzz888\/lavabit-founder-ladar-levison-discusses-his-federal-battle-for-privacy\">requested<\/a> Snowden\u2019s records after finding out that he used the service. Since Lavabit did not keep logs and email content was stored encrypted, the FBI served a subpoena (with a gag order) for the service\u2019s SSL keys. Having the SSL keys would allow them to access\ncommunications (both metadata and unencrypted content) in real time for all of Lavabit\u2019s customers, not just Snowden's.<\/p><p>Ultimately, Levison turned over the SSL keys and <a href=\"http:\/\/www.theguardian.com\/commentisfree\/2014\/may\/20\/why-did-lavabit-shut-down-snowden-email\">shut down<\/a> the service at the same time. The US government then <a href=\"http:\/\/www.cnbc.com\/id\/100962389\">threatened Levison with arrest<\/a>,\nsaying that shutting down the service was a violation of the court order.<\/p><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/www.bestvpn.com\/the-ultimate-privacy-guide\/#avoidus\">Avoid all US and UK based services<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Surespot#History\">Proof that warrant canaries work based on the surespot example.<\/a><\/li>\n<li><a href=\"http:\/\/en.wikipedia.org\/wiki\/UKUSA_Agreement\">http:\/\/en.wikipedia.org\/wiki\/UKUSA_Agreement<\/a><\/li>\n<li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Lavabit#Suspension_and_gag_order\">http:\/\/en.wikipedia.org\/wiki\/Lavabit#Suspension_and_gag_order<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law\">https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law<\/a><\/li>\n<li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Portal:Mass_surveillance\">http:\/\/en.wikipedia.org\/wiki\/Portal:Mass_surveillance<\/a><\/li>\n<\/ul><p><strong>All providers listed here are outside the US, use encryption, accept Bitcoin, support OpenVPN and have a no logging policy.<\/strong><\/p><br \/><div class=\"table-responsive\"><table class=\"table sortable-theme-bootstrap\" data-sortable=\"\"><thead><tr><th data-sortable=\"true\">Sortable VPN Providers Table<\/th>\n<th data-sortable=\"true\">Yearly Price<\/th>\n<th data-sortable=\"true\">Free Trial<\/th>\n<th data-sortable=\"true\" title=\"Number of Servers\"># Servers<\/th>\n<th data-sortable=\"true\">Jurisdiction<\/th>\n<th data-sortable=\"false\">Website<\/th>\n<\/tr><\/thead><tbody><tr><td data-value=\"AirVPN\">\n<a href=\"https:\/\/airvpn.org\/\"><img alt=\"AirVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/AirVPN.gif\" width=\"200\" height=\"70\" \/><\/a><\/td>\n<td data-value=\"70\">54 \u20ac<\/td>\n<td>Yes<\/td>\n<td>162<\/td>\n<td> Italy<\/td>\n<td><a href=\"https:\/\/airvpn.org\/\">AirVPN.org<\/a><\/td>\n<\/tr><tr><td data-value=\"AzireVPN\">\n<a href=\"https:\/\/www.azirevpn.com\/\"><img alt=\"AzireVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/AzireVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"58\">45 \u20ac<\/td>\n<td>Yes<\/td>\n<td>5<\/td>\n<td> Sweden<\/td>\n<td><a href=\"https:\/\/www.azirevpn.com\/\">AzireVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"blackVPN\">\n<a href=\"https:\/\/www.blackvpn.com\/\"><img alt=\"blackVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/blackVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"128\">99 \u20ac<\/td>\n<td>Yes<\/td>\n<td>27<\/td>\n<td> Hong Kong<\/td>\n<td><a href=\"https:\/\/www.blackvpn.com\/\">blackVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"Cryptostorm\">\n<a href=\"https:\/\/cryptostorm.is\/\"><img alt=\"Cryptostorm\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Cryptostorm.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"52\">$ 52<\/td>\n<td>Yes<\/td>\n<td>18<\/td>\n<td> Iceland<\/td>\n<td><a href=\"https:\/\/cryptostorm.is\/\">Cryptostorm.is<\/a><\/td>\n<\/tr><tr><td data-value=\"EarthVPN\">\n<a href=\"http:\/\/www.earthvpn.com\/\"><img alt=\"EarthVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/EarthVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"52\">39,99 \u20ac<\/td>\n<td>No<\/td>\n<td>432<\/td>\n<td> Northern Cyprus<\/td>\n<td><a href=\"http:\/\/www.earthvpn.com\/\">EarthVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"ExpressVPN\">\n<a href=\"https:\/\/www.expressvpn.com\/\"><img alt=\"ExpressVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/ExpressVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"100\">$ 99.95<\/td> \n<td>Yes<\/td>\n<td>145<\/td>\n<td> British Virgin Islands<\/td>\n<td><a href=\"https:\/\/www.expressvpn.com\/\">ExpressVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"FrootVPN\">\n<a href=\"https:\/\/www.frootvpn.com\/\"><img alt=\"FrootVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/FrootVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"36\">$ 35.88<\/td>\n<td>No<\/td>\n<td>27<\/td>\n<td> Sweden<\/td>\n<td><a href=\"https:\/\/www.frootvpn.com\/\">FrootVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"hide.me\">\n<a href=\"https:\/\/hide.me\/\"><img alt=\"hide.me\" src=\"https:\/\/www.privacytools.io\/img\/provider\/hide.me.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"65\">$ 65.04<\/td>\n<td>Yes<\/td>\n<td>88<\/td>\n<td> Malaysia<\/td>\n<td><a href=\"https:\/\/hide.me\/\">hide.me<\/a><\/td>\n<\/tr><tr><td data-value=\"IVPN\">\n<a href=\"https:\/\/www.ivpn.net\/\"><img alt=\"IVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/IVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"100\">$ 100<\/td>\n<td>Yes<\/td>\n<td>21<\/td>\n<td> Gibraltar<\/td>\n<td><a href=\"https:\/\/www.ivpn.net\/\">IVPN.net<\/a><\/td>\n<\/tr><tr><td data-value=\"Mullvad\">\n<a href=\"https:\/\/mullvad.net\/\"><img alt=\"Mullvad\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Mullvad.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"78\">60 \u20ac<\/td>\n<td>Yes<\/td>\n<td>168<\/td>\n<td> Sweden<\/td>\n<td><a href=\"https:\/\/mullvad.net\/\">Mullvad.net<\/a><\/td>\n<\/tr><tr><td data-value=\"NordVPN\">\n<a href=\"https:\/\/nordvpn.com\/\"><img alt=\"NordVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/NordVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"69\">$ 69<\/td>\n<td>Yes<\/td>\n<td>475<\/td>\n<td> Panama<\/td>\n<td><a href=\"https:\/\/nordvpn.com\/\">NordVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"OVPN.com\">\n<a href=\"https:\/\/www.ovpn.com\/\"><img alt=\"OVPN.com\" src=\"https:\/\/www.privacytools.io\/img\/provider\/OVPN.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"84\">$ 84<\/td>\n<td>Yes<\/td>\n<td>39<\/td>\n<td> Sweden<\/td>\n<td><a href=\"https:\/\/www.ovpn.com\/\">OVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"Perfect Privacy\">\n<a href=\"https:\/\/www.perfect-privacy.com\/\"><img alt=\"Perfect Privacy\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Perfect-Privacy.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"162\">124.95 \u20ac<\/td>\n<td>No<\/td>\n<td>41<\/td>\n<td> Panama<\/td>\n<td><a href=\"https:\/\/www.perfect-privacy.com\/\">Perfect-Privacy.com<\/a><\/td>\n<\/tr><tr><td data-value=\"ProtonVPN\">\n<a href=\"https:\/\/protonvpn.com\/\"><img alt=\"ProtonVPN\" src=\"https:\/\/www.privacytools.io\/img\/provider\/ProtonVPN.png\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"0\">Free<\/td>\n<td>Yes<\/td>\n<td>249<\/td>\n<td> Switzerland<\/td>\n<td><a href=\"https:\/\/protonvpn.com\/\">ProtonVPN.com<\/a><\/td>\n<\/tr><tr><td data-value=\"Proxy.sh\">\n<a href=\"https:\/\/proxy.sh\/\"><img alt=\"Proxy.sh\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Proxy.sh.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"90\">$ 90<\/td>\n<td>No<\/td>\n<td>300<\/td>\n<td> Seychelles<\/td>\n<td><a href=\"https:\/\/proxy.sh\/\">Proxy.sh<\/a><\/td>\n<\/tr><tr><td data-value=\"Trust.Zone\">\n<a href=\"https:\/\/trust.zone\/\"><img alt=\"Trust.Zone\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Trust.Zone.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"40\">$ 39.95<\/td>\n<td>Yes<\/td>\n<td>48<\/td>\n<td> Seychelles<\/td>\n<td><a href=\"https:\/\/trust.zone\/\">Trust.Zone<\/a><\/td>\n<\/tr><tr><td data-value=\"VPN.ht\">\n<a href=\"https:\/\/vpn.ht\/\"><img alt=\"VPN.ht\" src=\"https:\/\/www.privacytools.io\/img\/provider\/VPN.ht.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"40\">$ 39.99<\/td>\n<td>No<\/td>\n<td>122<\/td>\n<td> Hong Kong<\/td>\n<td><a href=\"https:\/\/vpn.ht\/\">VPN.ht<\/a><\/td>\n<\/tr><tr><td data-value=\"VPNArea\">\n<a href=\"https:\/\/vpnarea.com\/\"><img alt=\"VPNArea\" src=\"https:\/\/www.privacytools.io\/img\/provider\/vpnarea.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"59\">$ 59.00<\/td>\n<td>Yes<\/td>\n<td>204<\/td>\n<td> Bulgaria<\/td>\n<td><a href=\"https:\/\/vpnarea.com\/\">VPNArea.com<\/a><\/td>\n<\/tr><tr><td data-value=\"VPNTunnel\">\n<a href=\"https:\/\/vpntunnel.com\/\"><img alt=\"VPNTunnel\" src=\"https:\/\/www.privacytools.io\/img\/provider\/VPNTunnel.gif\" width=\"200\" height=\"70\" \/><\/a>\n<\/td>\n<td data-value=\"36\">$ 35.88<\/td>\n<td>No<\/td>\n<td>80<\/td>\n<td> Seychelles<\/td>\n<td><a href=\"https:\/\/vpntunnel.com\/\">VPNTunnel.com<\/a><\/td>\n<\/tr><\/tbody><\/table><\/div><p><strong>Note: Using a VPN provider will not make you anonymous. But it will give you a better privacy. A VPN is not a tool for illegal activities. Don't rely on a \"no log\" policy.<\/strong><\/p><div class=\"container\"><div class=\"row\"><div class=\"col-md-6\"><h3>Our VPN Provider Criteria<\/h3><ul><li>Operating outside the USA or other Five Eyes countries.\n<br \/>More: <a href=\"https:\/\/www.bestvpn.com\/the-ultimate-privacy-guide\/#avoidus\">Avoid all US and UK based services.<\/a><\/li>\n<li>OpenVPN software support.<\/li>\n<li>Accepts Bitcoin, cash, debit cards or cash cards as a payment method.<\/li>\n<li>No personal information is required to create an account. Only username, password and <a href=\"#email\"><i class=\"fas fa-link\"> Email.<\/i><\/a><\/li>\n<\/ul><p>We're not affiliated with any of the above-listed VPN providers. This way can give you honest recommendations.<\/p><h3>More VPN Provider<\/h3><p>\n<a href=\"https:\/\/thatoneprivacysite.net\/vpn-comparison-chart\/\">\nSpreadsheet with unbiased, independently verifiable data on over 100 VPN services.\n<\/a>\n<a href=\"https:\/\/www.reddit.com\/r\/privacytoolsIO\/comments\/43d4zs\/you_guys_might_be_interested_in_this_a\/\">\n(Join the discussion on Reddit)\n<\/a>\n<\/p><\/div><div class=\"col-md-6\"><h3>Related VPN information<\/h3><ul><li><a href=\"https:\/\/torrentfreak.com\/vpn-services-keep-anonymous-2018\/\">Which VPN Services Keep You Anonymous in 2018?<\/a><br \/>(<strong>Note:<\/strong> The providers listed first in the TorrentFreaks article are sponsored)<\/li>\n<li><a href=\"http:\/\/torrentfreak.com\/how-to-make-vpns-even-more-secure-120419\/\">How To Make VPNs Even More Secure<\/a><\/li>\n<li><a href=\"https:\/\/vikingvpn.com\/blogs\/off-topic\/beware-of-vpn-marketing-and-affiliate-programs\">Beware of False Reviews - VPN Marketing and Affiliate Programs<\/a><\/li>\n<li><a href=\"https:\/\/www.goldenfrog.com\/take-back-your-internet\/articles\/7-myths-about-vpn-logging-and-anonymity\">I am Anonymous When I Use a VPN - 7 Myths Debunked<\/a><br \/>(<strong>Note:<\/strong> While this is a good read, they also use the article for self-promotion)<\/li>\n<li><a href=\"http:\/\/torrentfreak.com\/proxy-sh-vpn-provider-monitored-traffic-to-catch-hacker-130930\/\">Proxy.sh VPN Provider Sniffed Server Traffic to Catch Hacker<\/a><\/li>\n<li><a href=\"https:\/\/proxy.sh\/panel\/knowledgebase.php?action=displayarticle&amp;id=5\">Ethical policy - All of the reasons why Proxy.sh might enable logging<\/a><\/li>\n<li><a href=\"https:\/\/www.ivpn.net\/privacy\">IVPN.net will collect your email and IP address after sign up<\/a><br \/>Read the <a data-toggle=\"tooltip\" data-placement=\"top\" data-original-title=\"The IP collected at signup is only used for a few seconds by our fraud module and then discarded, it is not stored. Storing them would significantly increase our own liability and certainly would not be in our interest. You're absolutely welcome to signup using Tor or a VPN.\">Email statement<\/a> from IVPN.<\/li>\n<li><a href=\"https:\/\/medium.com\/@blackVPN\/no-logs-6d65d95a3016\">blackVPN announced to delete connection logs after disconnection<\/a><\/li>\n<li><a href=\"https:\/\/gist.github.com\/kennwhite\/1f3bc4d889b02b35d8aa\">Don't use LT2P IPSec, use other protocols.<\/a><\/li>\n<\/ul><\/div><\/div><\/div><img src=\"https:\/\/www.privacytools.io\/img\/layout\/warrant-canary.jpg\" class=\"img-fluid float-right\" alt=\"Warrant Canary Example\" \/><p>A warrant canary is a posted document stating that an organization has not received any secret subpoenas during a specific period of time. If this document fails to be updated during the specified time then the user is to assume that the service has\nreceived such a subpoena and should stop using the service.<\/p><h4>Warrant Canary Examples:<\/h4><ol><li><a href=\"https:\/\/proxy.sh\/canary\">https:\/\/proxy.sh\/canary<\/a><\/li>\n<li><a href=\"https:\/\/www.ivpn.net\/resources\/canary.txt\">https:\/\/www.ivpn.net\/resources\/canary.txt<\/a><\/li>\n<li><a href=\"https:\/\/www.vpnsecure.me\/files\/canary.txt\">https:\/\/www.vpnsecure.me\/files\/canary.txt<\/a><\/li>\n<li><a href=\"https:\/\/www.bolehvpn.net\/canary.txt\">https:\/\/www.bolehvpn.net\/canary.txt<\/a><\/li>\n<li><a href=\"https:\/\/lokun.is\/canary.txt\">https:\/\/lokun.is\/canary.txt<\/a><\/li>\n<li><a href=\"https:\/\/www.ipredator.se\/static\/downloads\/canary.txt\">https:\/\/www.ipredator.se\/static\/downloads\/canary.txt<\/a><\/li>\n<\/ol><h4>Related Warrant Canary Information<\/h4><ul><li><a href=\"https:\/\/www.eff.org\/de\/deeplinks\/2014\/04\/warrant-canary-faq\">Warrant Canary Frequently Asked Questions<\/a><\/li>\n<li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Warrant_canary#Companies_and_organizations_with_warrant_canaries\">Companies and organizations with warrant canaries<\/a><\/li>\n<li><a href=\"https:\/\/www.schneier.com\/blog\/archives\/2015\/03\/australia_outla.html\">Warrant canary criticism by Bruce Schneier and an example of a law against warrant canaries.<\/a><\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Tor Browser<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Tor-Project.png\" alt=\"Tor Browser\" class=\"panel-item\" \/>Tor Browser is your choice if you need an extra layer of anonymity. It's a modified version of Firefox, it comes with pre-installed privacy add-ons, encryption and an advanced proxy.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/www.torproject.org\/\">Website:\ntorproject.org\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/expyuzz4wqqyqhjn.onion Requires specific software to access: torproject.org\" href=\"http:\/\/expyuzz4wqqyqhjn.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><p>OS: Windows, macOS, Linux, <a href=\"https:\/\/mike.tig.as\/onionbrowser\/\">iOS<\/a>, <a href=\"https:\/\/www.torproject.org\/docs\/android.html.en\">Android<\/a>, <a href=\"https:\/\/github.com\/torbsd\/openbsd-ports\">OpenBSD.<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Mozilla Firefox<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Firefox.png\" alt=\"Mozilla Firefox\" class=\"panel-item\" \/>Firefox is fast, reliable, open source and respects your privacy. Don't forget to adjust the settings according to our\nrecommendations: <a href=\"#webrtc\"><i class=\"fas fa-link\"> WebRTC<\/i><\/a> and <a href=\"#about_config\"><i class=\"fas fa-link\"> about:config<\/i><\/a> and get the <a href=\"#addons\"><i class=\"fas fa-link\"> privacy add-ons<\/i><\/a>.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/www.firefox.com\/\">Website:\nfirefox.com\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, iOS, BSD.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Brave<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Brave.png\" alt=\"Brave\" class=\"panel-item\" \/>The new open source browser \"Brave\" automatically blocks ads and trackers, making it faster and safer than your current browser. Brave is based on Chromium.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/www.brave.com\/\">Website:\nbrave.com\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, iOS.<\/p><\/div><\/div><\/div><\/div><\/div><p><strong>Your Browser sends information that makes you unique amongst millions of users and therefore easy to identify.<\/strong><\/p><img src=\"https:\/\/www.privacytools.io\/img\/layout\/browser-fingerprint.jpg\" class=\"img-fluid float-right\" alt=\"Fingerprint image\" \/><p>When you visit a web page, your browser voluntarily sends information about its configuration, such as available fonts, browser type, and add-ons. If this combination of information is unique, it may be possible to identify and track you without using\ncookies. EFF created a Tool called <a href=\"https:\/\/panopticlick.eff.org\/\">Panopticlick<\/a> to test your browser to see how unique it is.<\/p><p>\n<a class=\"btn btn-warning\" href=\"https:\/\/panopticlick.eff.org\/\">\nTest your Browser now\n<\/a>\n<\/p><p>You need to find what <strong>most browsers<\/strong> are reporting, and then use those variables to bring your browser in the same population. This means having the same fonts, plugins, and extensions installed as the large installed base. You should\nhave a <a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/uaswitcher\/\">spoofed user agent string<\/a> to match what the large userbase has. You need to have the same settings enabled and disabled, such as DNT and WebGL. You need your browser to\nlook as common as everyone else. Disabling JavaScript, using Linux, or even the TBB, will make your browser stick out from the masses.<\/p><p>Modern web browsers have not been architected to assure personal web privacy. Rather than worrying about being fingerprinted, it seems more practical to use <a href=\"#addons\"><i class=\"fas fa-link\"> free software plugins<\/i><\/a> like Privacy Badger, uBlock Origin and Disconnect.\nThey not only respect your freedom, but your privacy also. You can get much further with these than trying to manipulate your browser's fingerprint.<\/p><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/panopticlick.eff.org\/static\/browser-uniqueness.pdf\">How Unique Is Your Web Browser? Peter Eckersley, EFF.<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/privacytoolsIO\/comments\/35pqyl\/new_section_browser_fingerprint_is_your_browser\/\">Join our discussion on reddit.com about browser fingerprinting.<\/a><\/li>\n<li><a href=\"#addons\"><i class=\"fas fa-link\"> Our Firefox privacy add-ons section.<\/i><\/a><\/li>\n<li><a href=\"https:\/\/www.browserleaks.com\/\">BrowserLeaks.com<\/a> - Web browser security testing tools, that tell you what exactly personal identity data may be leaked without any permissions when you surf the Internet.<\/li>\n<\/ul><p><strong>WebRTC is a new communication protocol that relies on JavaScript that can leak your actual IP address from behind your VPN.<\/strong><\/p><p>While software like NoScript prevents this, it's probably a good idea to block this protocol directly as well, just to be safe.<\/p><p>\n<a class=\"btn btn-warning\" href=\"https:\/\/www.privacytools.io\/webrtc.html\">\nTest your Browser now\n<\/a>\n<\/p><h3>How to disable WebRTC in Firefox?<\/h3><p>In short: Set \"media.peerconnection.enabled\" to \"false\" in \"about:config\".<\/p><p><strong>Explained:<\/strong><\/p><ol class=\"long-string-list\"><li>Enter \"about:config\" in the firefox address bar and press enter.<\/li>\n<li>Press the button \"I'll be careful, I promise!\"<\/li>\n<li>Search for \"media.peerconnection.enabled\"<\/li>\n<li>Double click the entry, the column \"Value\" should now be \"false\"<\/li>\n<li>Done. Do the WebRTC leak test again.<\/li>\n<\/ol><p>If you want to make sure every single WebRTC related setting is really disabled change these settings:<\/p><ol class=\"long-string-list\"><li>media.peerconnection.turn.disable = true<\/li>\n<li>media.peerconnection.use_document_iceservers = false<\/li>\n<li>media.peerconnection.video.enabled = false<\/li>\n<li>media.peerconnection.identity.timeout = 1<\/li>\n<\/ol><p>Now you can be 100% sure WebRTC is disabled.<\/p><p>\n<a class=\"btn btn-primary\" href=\"https:\/\/www.privacytools.io\/webrtc.html\">\nTest your Browser again\n<\/a>\n<\/p><h3>How to fix the WebRTC Leak in Google Chrome?<\/h3><p>WebRTC cannot be fully disabled in Chrome, however it is possible to change its routing settings (and prevent leaks) using an extension. Two open source solutions include <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/webrtc-leak-prevent\/eiadekoaikejlgdbkbdfeijglgfdalml\">WebRTC Leak Prevent<\/a> (options may need to be changed depending on the scenario), and <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/ublock-origin\/cjpalhdlnbpafiamejdnhcphjbkeiagm\">uBlock Origin<\/a> (select \"Prevent WebRTC from leaking local IP addresses\" in Settings).<\/p><h3>What about other browsers?<\/h3><p>Chrome on iOS, Internet Explorer and Safari does not implement WebRTC yet. <a href=\"#browser\"><i class=\"fas fa-link\"> But we recommend using Firefox on all devices.<\/i><\/a><\/p><p><strong>Improve your privacy with these excellent Firefox add-ons.<\/strong><\/p><h3>Stop Tracking with \"Privacy Badger\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/Privacy-Badger.gif\" class=\"img-fluid float-left\" alt=\"Privacy Badger\" \/><p><strong>Privacy Badger<\/strong> is a browser add-on that stops advertisers and other third-party trackers from secretly tracking where you go and what pages you look at on the web. Privacy Badger learns about trackers as you browse by inspecting external resources websites request. Please be aware that Google Analytics is considered first-party by Privacy Badger, which means that Google Analytics will not be blocked if you do not use another blocker, such as uBlock Origin.\n<a href=\"https:\/\/www.eff.org\/privacybadger\">https:\/\/www.eff.org\/privacybadger\/<\/a><\/p><h3>Block Ads and Trackers with \"uBlock Origin\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/uBlock.gif\" class=\"img-fluid float-left\" alt=\"uBlock\" \/><p>An efficient <a href=\"https:\/\/github.com\/gorhill\/uBlock\/wiki\/Blocking-mode\">wide-spectrum-blocker<\/a> that's easy on memory, and yet can load and enforce thousands more filters than other popular blockers out there. It has no monetization strategy and\nis completely <a href=\"https:\/\/github.com\/gorhill\/uBlock\/\">open source<\/a>. We recommend Firefox but uBlock Origin also works in other browsers such as Safari, Opera, and Chromium. Unlike AdBlock Plus, uBlock does not allow so-called <a href=\"https:\/\/adblockplus.org\/acceptable-ads\">\"acceptable ads\"<\/a>.\n<br \/><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/ublock-origin\/\">https:\/\/addons.mozilla.org\/firefox\/addon\/ublock-origin\/<\/a>\n<\/p><h3>Automatically Delete Cookies with \"Cookie AutoDelete\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/Cookie-AutoDelete.gif\" class=\"img-fluid float-left\" alt=\"Cookie AutoDelete\" \/><p>Automatically removes cookies when they are no longer used by open browser tabs. With the cookies, lingering sessions, as well as information used to spy on you, will be expunged.\n<br \/><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/cookie-autodelete\">https:\/\/addons.mozilla.org\/firefox\/addon\/cookie-autodelete\/<\/a><\/p><h3>Encryption with \"HTTPS Everywhere\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/HTTPS-Everywhere.gif\" class=\"img-fluid float-left\" alt=\"HTTPS Everywhere\" \/><p>A Firefox, Chrome, and Opera extension that encrypts your communications with many major websites, making your browsing more secure. A collaboration between The Tor Project and the Electronic Frontier Foundation.\n<br \/><a href=\"https:\/\/www.eff.org\/https-everywhere\">https:\/\/www.eff.org\/https-everywhere<\/a><\/p><h3>Block Content Delivery Networks with \"Decentraleyes\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/Decentraleyes.gif\" class=\"img-fluid float-left\" alt=\"Decentraleyes\" \/><p>Emulates Content Delivery Networks locally by intercepting requests, finding the required resource and injecting it into the environment. This all happens instantaneously, automatically, and no prior configuration is required. <i>Source code: <a href=\"https:\/\/git.synz.io\/Synzvato\/decentraleyes\">GitLab (self-hosted)<\/a>.<\/i>\n<br \/><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/decentraleyes\/\">https:\/\/addons.mozilla.org\/firefox\/addon\/decentraleyes\/<\/a><\/p><p><strong>The following add-ons require quite a lot of interaction from the user to get things working. Some sites will not work properly until you have configured the add-ons.<\/strong><\/p><h3>Stop cross-site requests with \"uMatrix\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/uMatrix.gif\" class=\"img-fluid float-left\" alt=\"uMatrix\" \/><p>Many websites integrate features which let other websites track you, such as Facebook Like Buttons or Google Analytics. uMatrix gives you control over the requests that websites make to other websites. This gives you greater and more fine-grained control\nover the information that you leak online.\n<br \/><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/umatrix\/\">https:\/\/addons.mozilla.org\/firefox\/addon\/umatrix\/<\/a>\n<\/p><h3>Be in total control with \"NoScript Security Suite\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/NoScript.gif\" class=\"img-fluid float-left\" alt=\"NoScript\" \/><p>\nHighly customizable plugin to selectively allow Javascript, Java, and Flash to run only on websites you trust. Not for casual users, it requires technical knowledge to configure.\n<br \/><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/noscript\/\">https:\/\/addons.mozilla.org\/firefox\/addon\/noscript\/<\/a>\n<\/p><p>This is a collection of privacy-related <strong>about:config<\/strong> tweaks. We'll show you how to enhance the privacy of your Firefox browser.<\/p><h3>Preparation:<\/h3><ol><li>Enter \"about:config\" in the firefox address bar and press enter.<\/li>\n<li>Press the button \"I'll be careful, I promise!\"<\/li>\n<li>Follow the instructions below...<\/li>\n<\/ol><h3>Getting started:<\/h3><ol class=\"long-string-list\"><li>privacy.firstparty.isolate = true\n<ul><li>A result of the <a href=\"https:\/\/wiki.mozilla.org\/Security\/Tor_Uplift\">Tor Uplift<\/a> effort, this preference isolates all browser identifier sources (e.g. cookies) to the first party domain, with the goal of preventing tracking across different domains. (Don't do this if you are using the Firefox Addon \"Cookie AutoDelete\" with Firefox v58 or below.)<\/li>\n<\/ul><\/li>\n<li>privacy.resistFingerprinting = true\n<ul><li>A result of the <a href=\"https:\/\/wiki.mozilla.org\/Security\/Tor_Uplift\">Tor Uplift<\/a> effort, this preference makes Firefox more resistant to browser fingerprinting.<\/li>\n<\/ul><\/li>\n<li>privacy.trackingprotection.enabled = true\n<ul><li>This is Mozilla\u2019s new built-in tracking protection. It uses Disconnect.me filter list, which is redundant if you are already using uBlock Origin 3rd party filters, therefore you should set it to false if you are using the add-on functionalities.<\/li>\n<\/ul><\/li>\n<li>browser.cache.offline.enable = false\n<ul><li>Disables offline cache.<\/li>\n<\/ul><\/li>\n<li>browser.safebrowsing.malware.enabled = false\n<ul><li>Disable Google Safe Browsing malware checks. Security risk, but privacy improvement.<\/li>\n<\/ul><\/li>\n<li>browser.safebrowsing.phishing.enabled = false\n<ul><li>Disable Google Safe Browsing and phishing protection. Security risk, but privacy improvement.<\/li>\n<\/ul><\/li>\n<li>browser.send_pings = false\n<ul><li>The attribute would be useful for letting websites track visitors\u2019 clicks. <\/li>\n<\/ul><\/li>\n<li>browser.sessionstore.max_tabs_undo = 0\n<ul><li>Even with Firefox set to not remember history, your closed tabs are stored temporarily at Menu -&gt; History -&gt; Recently Closed Tabs.<\/li>\n<\/ul><\/li>\n<li>browser.urlbar.speculativeConnect.enabled = false\n<ul><li>Disable preloading of autocomplete URLs. Firefox preloads URLs that autocomplete when a user types into the address bar, which is a concern if URLs are suggested that the user does not want to connect to. <a href=\"https:\/\/www.ghacks.net\/2017\/07\/24\/disable-preloading-firefox-autocomplete-urls\/\">Source<\/a><\/li>\n<\/ul><\/li>\n <li>dom.battery.enabled = false\n<ul><li>Website owners can track the battery status of your device. <a href=\"https:\/\/www.reddit.com\/r\/privacytoolsIO\/comments\/3fzbgy\/you_may_be_tracked_by_your_battery_status_of_your\/\">Source<\/a><\/li>\n<\/ul><\/li>\n<li>dom.event.clipboardevents.enabled = false\n<ul><li>Disable that websites can get notifications if you copy, paste, or cut something from a web page, and it lets them know which part of the page had been selected.<\/li>\n<\/ul><\/li>\n<li>geo.enabled = false\n<ul><li>Disables geolocation.<\/li>\n<\/ul><\/li>\n<li>media.eme.enabled = false\n<ul><li>Disables playback of DRM-controlled HTML5 content, which, if enabled, automatically downloads the Widevine Content Decryption Module provided by Google Inc. <a href=\"https:\/\/support.mozilla.org\/en-US\/kb\/enable-drm#w_opt-out-of-cdm-playback-uninstall-cdms-and-stop-all-cdm-downloads\">Details<\/a><\/li>\n<li>DRM-controlled content that requires the Adobe Flash or Microsoft Silverlight NPAPI plugins will still play, if installed and enabled in Firefox.<\/li>\n<\/ul><\/li>\n<li>media.gmp-widevinecdm.enabled = false\n<ul><li>Disables the Widevine Content Decryption Module provided by Google Inc., used for the playback of DRM-controlled HTML5 content. <a href=\"https:\/\/support.mozilla.org\/en-US\/kb\/enable-drm#w_disable-the-google-widevine-cdm-without-uninstalling\">Details<\/a><\/li>\n<\/ul><\/li>\n<li>media.navigator.enabled = false\n<ul><li>Websites can track the microphone and camera status of your device.<\/li>\n<\/ul><\/li>\n<li>network.cookie.cookieBehavior = 1\n<ul><li>Disable cookies<\/li>\n<li>0 = Accept all cookies by default<\/li>\n<li>1 = Only accept from the originating site (block third-party cookies)<\/li>\n<li>2 = Block all cookies by default<\/li>\n<\/ul><\/li>\n<li>network.cookie.lifetimePolicy = 2\n<ul><li>cookies are deleted at the end of the session<\/li>\n<li>0 = Accept cookies normally<\/li>\n<li>1 = Prompt for each cookie<\/li>\n<li>2 = Accept for current session only<\/li>\n<li>3 = Accept for N days<\/li>\n<\/ul><\/li>\n<li>network.http.referer.trimmingPolicy = 2\n<ul><li>Send only the scheme, host, and port in the <code>Referer<\/code> header<\/li>\n<li>0 = Send the full URL in the <code>Referer<\/code> header<\/li>\n<li>1 = Send the URL without its query string in the <code>Referer<\/code> header<\/li>\n<li>2 = Send only the scheme, host, and port in the <code>Referer<\/code> header<\/li>\n<\/ul><\/li>\n<li>network.http.referer.XOriginPolicy = 2\n<ul><li>Only send <code>Referer<\/code> header when the full hostnames match. (Note: if you notice significant breakage, you might try <code>1<\/code> combined with an <code>XOriginTrimmingPolicy<\/code> tweak below.) <a href=\"https:\/\/feeding.cloud.geek.nz\/posts\/tweaking-referrer-for-privacy-in-firefox\/\">Source<\/a><\/li>\n<li>0 = Send <code>Referer<\/code> in all cases<\/li>\n<li>1 = Send <code>Referer<\/code> to same eTLD sites<\/li>\n<li>2 = Send <code>Referer<\/code> only when the full hostnames match<\/li>\n<\/ul><\/li>\n<li>network.http.referer.XOriginTrimmingPolicy = 2\n<ul><li>When sending <code>Referer<\/code> across origins, only send scheme, host, and port in the <code>Referer<\/code> header of cross-origin requests. <a href=\"https:\/\/feeding.cloud.geek.nz\/posts\/tweaking-referrer-for-privacy-in-firefox\/\">Source<\/a><\/li>\n<li>0 = Send full url in <code>Referer<\/code><\/li>\n<li>1 = Send url without query string in <code>Referer<\/code><\/li>\n<li>2 = Only send scheme, host, and port in <code>Referer<\/code><\/li>\n<\/ul><\/li>\n<li>webgl.disabled = true\n<ul><li>WebGL is a potential security risk. <a href=\"http:\/\/security.stackexchange.com\/questions\/13799\/is-webgl-a-security-concern\">Source<\/a><\/li>\n<\/ul><\/li>\n<li>browser.sessionstore.privacy_level = 2\n<ul><li>This preference controls when to store extra information about a session: contents of forms, scrollbar positions, cookies, and POST data. <a href=\"http:\/\/kb.mozillazine.org\/Browser.sessionstore.privacy_level\">Details<\/a><\/li>\n<li>0 = Store extra session data for any site. (Default starting with Firefox 4.)<\/li>\n<li>1 = Store extra session data for unencrypted (non-HTTPS) sites only. (Default before Firefox 4.)<\/li>\n<li>2 = Never store extra session data.<\/li>\n<\/ul><\/li>\n<li>network.IDN_show_punycode = true\n<ul><li>Not rendering IDNs as their Punycode equivalent leaves you open to phishing attacks that can be very difficult to notice. <a href=\"https:\/\/krebsonsecurity.com\/2018\/03\/look-alike-domains-and-visual-confusion\/#more-42636\">Source<\/a><\/li>\n<\/ul><\/li>\n<li>extensions.blocklist.url = https:\/\/blocklists.settings.services.mozilla.com\/v1\/blocklist\/3\/%20\/%20\/\n<ul><li>Limit the amount of identifiable information sent when requesting the Mozilla harmful extension blocklist.<\/li>\n<li>Optionally, the blocklist can be disabled entirely by setting <code>extensions.blocklist.enabled<\/code> to false for increased privacy, but decreased security. <a href=\"https:\/\/old.reddit.com\/r\/privacytoolsIO\/comments\/9uqeew\/firefox_tip_sanitize_firefox_blocklist_url_so_it\/\">Source<\/a><\/li>\n<\/ul><\/li>\n<\/ol><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/ffprofile.com\/\">ffprofile.com<\/a> - Helps you to create a Firefox profile with the defaults you like.<\/li>\n<li><a href=\"http:\/\/kb.mozillazine.org\/Category:Security_and_privacy-related_preferences\">mozillazine.org<\/a> - Security and privacy-related preferences. <\/li>\n<li><a href=\"https:\/\/github.com\/pyllyukko\/user.js\">user.js Firefox hardening stuff<\/a> - This is a user.js configuration file for Mozilla Firefox that's supposed to harden Firefox's settings and make it more secure.<\/li>\n<li><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/privacy-settings\/\">Privacy Settings<\/a> - A Firefox addon to alter built-in privacy settings easily with a toolbar panel.<\/li>\n<\/ul><p><strong>All providers listed here are operating outside the US and support <a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"When sending or receiving emails, if both the sending and receiving servers support TLS encryption, the email is sent between servers using an encrypted connection.\">SMTP TLS.<\/a> The table is sortable. Never trust any company with your privacy, always encrypt.<\/strong><\/p><br \/><div class=\"table-responsive\"><table class=\"table sortable-theme-bootstrap\" data-sortable=\"\"><thead><tr><th data-sortable=\"true\">Email Service<\/th>\n<th data-sortable=\"true\">URL<\/th>\n<th data-sortable=\"true\">Since<\/th>\n<th data-sortable=\"true\">Server<\/th>\n<th data-sortable=\"true\">Storage<\/th>\n<th data-sortable=\"true\">Price \/ Year<\/th>\n<th data-sortable=\"true\">Bitcoin<\/th>\n<th data-sortable=\"true\">Encryption<\/th>\n<th data-sortable=\"true\">Own Domain<\/th>\n<\/tr><\/thead><tbody><tr><td data-value=\"ProtonMail\">\n<img alt=\"Protonmail\" src=\"https:\/\/www.privacytools.io\/img\/provider\/ProtonMail.ch.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.protonmail.ch\" href=\"https:\/\/www.protonmail.ch\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/protonirockerxow.onion. Requires specific software to access: torproject.org\" href=\"https:\/\/protonirockerxow.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2013\">2013<\/td>\n<td> Switzerland<\/td>\n<td data-value=\"500\">500 MB<\/td>\n<td data-value=\"0\">Free<\/td>\n<td data-value=\"1\">Accepted<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"disroot.org\">\n<img alt=\"Disroot\" src=\"https:\/\/www.privacytools.io\/img\/provider\/disroot.org.png\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/disroot.org\" href=\"https:\/\/disroot.org\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2015\">2015<\/td>\n<td> Netherlands<\/td>\n<td data-value=\"2000\">2 GB<\/td>\n<td data-value=\"1\">Free<\/td>\n<td data-value=\"1\">Accepted<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"Tutanota\">\n<img alt=\"Tutanota\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Tutanota.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.tutanota.com\" href=\"https:\/\/www.tutanota.com\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2011\">2011<\/td>\n<td> Germany<\/td>\n<td data-value=\"1000\">1 GB<\/td>\n<td data-value=\"0\">Free<\/td>\n<td data-value=\"0\">No<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"Mailfence\">\n<img alt=\"Mailfence\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Mailfence.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.mailfence.com\" href=\"https:\/\/www.mailfence.com\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2013\">2013<\/td>\n<td> Belgium<\/td>\n<td data-value=\"500\">500 MB<\/td>\n<td data-value=\"0\">Free<\/td>\n<td data-value=\"1\">Accepted<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"mailbox.org\">\n<img alt=\"mailbox.org\" src=\"https:\/\/www.privacytools.io\/img\/provider\/mailbox.org.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.mailbox.org\" href=\"https:\/\/www.mailbox.org\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2014\">2014<\/td>\n<td> Germany<\/td>\n<td data-value=\"2000\">2 GB<\/td>\n<td data-value=\"13\">12 \u20ac<\/td>\n<td data-value=\"1\">Accepted<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"Posteo\">\n<img alt=\"Posteo\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Posteo.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.posteo.de\" href=\"https:\/\/www.posteo.de\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2009\">2009<\/td>\n<td> Germany<\/td>\n<td data-value=\"2000\">2 GB<\/td>\n<td data-value=\"13\">12 \u20ac<\/td>\n<td data-value=\"0\">No<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"0\">No<\/td>\n<\/tr><tr><td data-value=\"Runbox\">\n<img alt=\"Runbox\" src=\"https:\/\/www.privacytools.io\/img\/provider\/runbox.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/runbox.com\" href=\"https:\/\/runbox.com\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"1999\">1999<\/td>\n<td> Norway<\/td>\n<td data-value=\"1000\">1 GB<\/td>\n<td data-value=\"20\">$ 19.95<\/td>\n<td data-value=\"0\">No<\/td>\n<td data-value=\"0\">No<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"Neomailbox\">\n<img alt=\"Neomailbox\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Neomailbox.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.neomailbox.com\" href=\"https:\/\/www.neomailbox.com\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2003\">2003<\/td>\n<td> Switzerland<\/td>\n<td data-value=\"1000\">1 GB<\/td>\n<td data-value=\"50\">$ 49.95<\/td>\n<td data-value=\"1\">Accepted<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"StartMail\">\n<img alt=\"StartMail\" src=\"https:\/\/www.privacytools.io\/img\/provider\/StartMail.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.startmail.com\" href=\"https:\/\/www.startmail.com\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2014\">2014<\/td>\n<td> Netherlands<\/td>\n<td data-value=\"10000\">10 GB<\/td>\n<td data-value=\"60\">$ 59.95<\/td>\n<td data-value=\"0\">Accepted<\/td>\n<td data-value=\"1\">Built-in<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><tr><td data-value=\"Kolab Now\">\n<img alt=\"Kolab Now\" src=\"https:\/\/www.privacytools.io\/img\/provider\/Kolab-Now.gif\" width=\"200\" height=\"70\" \/><\/td>\n<td>\n<a data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"https:\/\/www.kolabnow.com\" href=\"https:\/\/www.kolabnow.com\"><img alt=\"WWW\" src=\"https:\/\/www.privacytools.io\/img\/layout\/www.png\" width=\"35\" \/><\/a>\n<\/td>\n<td data-value=\"2010\">2010<\/td>\n<td> Switzerland<\/td>\n<td data-value=\"2048\">2 GB<\/td>\n<td data-value=\"60\">$ 60<\/td>\n<td data-value=\"1\">Accepted<\/td>\n<td data-value=\"0\">No<\/td>\n<td data-value=\"1\">Yes<\/td>\n<\/tr><\/tbody><\/table><\/div><br \/><h3>Interesting Email Providers Under Development<\/h3><ul><li><a href=\"https:\/\/www.confidantmail.org\/\">Confidant Mail<\/a> - An open-source non-SMTP cryptographic email system optimized for large file attachments. It is a secure and spam-resistant alternative to regular email and online file drop services. It\nuses <a href=\"https:\/\/theprivacyguide.org\/tutorials\/gpg.html\">GNU Privacy Guard (GPG)<\/a> for content encryption and authentication, and TLS 1.2 with ephemeral keys for transport encryption.<\/li>\n<\/ul><h3>Become Your Own Email Provider with Mail-in-a-Box<\/h3><a href=\"https:\/\/mailinabox.email\/\"><img src=\"https:\/\/www.privacytools.io\/img\/provider\/Mail-in-a-Box.gif\" class=\"img-fluid float-left\" alt=\"Mail-in-a-Box\" \/><\/a><p>Take it a step further and get control of your email with this easy-to-deploy mail server in a box. Mail-in-a-Box lets you become your own mail service provider in a few easy steps. It\u2019s sort of like making your own Gmail, but one you control from top\nto bottom. Technically, Mail-in-a-Box turns a fresh cloud computer into a working mail server. But you don\u2019t need to be a technology expert to set it up. <strong>More: <a href=\"https:\/\/mailinabox.email\/\">https:\/\/mailinabox.email\/<\/a><\/strong><\/p><h3>Privacy Email Tools<\/h3><ul><li><a href=\"http:\/\/www.gpg4usb.org\/\">gpg4usb<\/a> - A very easy to use and small portable editor to encrypt and decrypt any text-message or -file. For Windows and Linux. <a href=\"https:\/\/theprivacyguide.org\/tutorials\/gpg.html\">GPG tutorial<\/a>.<\/li>\n<li><a href=\"https:\/\/www.mailvelope.com\/\">Mailvelope<\/a> - A browser extension that enables the exchange of encrypted emails following the <a href=\"https:\/\/theprivacyguide.org\/tutorials\/pgp.html\">OpenPGP encryption standard<\/a>.<\/li>\n<li><a href=\"https:\/\/www.enigmail.net\/\">Enigmail<\/a> - A security extension to Thunderbird and Seamonkey. It enables you to write and receive email messages signed and\/or encrypted with the <a href=\"https:\/\/theprivacyguide.org\/tutorials\/pgp.html\">OpenPGP standard<\/a>.<\/li>\n<li><a href=\"https:\/\/addons.mozilla.org\/thunderbird\/addon\/torbirdy\/\">TorBirdy<\/a> - This extension configures Thunderbird to make connections over the Tor anonymity network.<\/li>\n<li><a href=\"https:\/\/emailprivacytester.com\/\">Email Privacy Tester<\/a> - This tool will send an Email to your address and perform privacy-related tests.<\/li>\n<\/ul><h3>Related Information<\/h3><ul><li><a href=\"http:\/\/www.wired.com\/2011\/10\/ecpa-turns-twenty-five\/\">Aging \u2018Privacy\u2019 Law Leaves Cloud E-Mail Open to Cops<\/a> - Data stored in the cloud for longer than 6 months is considered abandoned and may be accessed by intelligence agencies without\na warrant. Learning: Use an external email client like Thunderbird or Enigmail, download your emails and store them locally. Never leave them on the server.<\/li>\n<li><a href=\"https:\/\/www.eff.org\/deeplinks\/2012\/04\/may-firstriseup-server-seizure-fbi-overreaches-yet-again\">With May First\/Riseup Server Seizure, FBI Overreaches Yet Again<\/a><\/li>\n<li><a href=\"http:\/\/www.autistici.org\/ai\/crackdown\/\">Autistici\/Inventati server compromised<\/a> - The cryptographic services offered by the Autistici\/Inventati server have been compromised on 15th June 2004. It was discovered on 21st June 2005. One year\nlater. During an enquiry on a single mailbox, the Postal Police may have tapped for a whole year every user's private communication going through the server autistici.org\/inventati.org.<\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Thunderbird<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Thunderbird.png\" alt=\"Thunderbird\" class=\"panel-item\" \/>Mozilla Thunderbird is a free, open source, cross-platform email, news, and chat client developed by the Mozilla Foundation. Thunderbird is an email, newsgroup, news feed, and chat (XMPP, IRC, Twitter) client.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/mozilla.org\/thunderbird\">Website:\nmozilla.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, BSD.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Claws Mail<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Claws-Mail.png\" alt=\"Claws Mail\" class=\"panel-item\" \/>Claws Mail is a free and open source, GTK+-based email and news client. It offers easy configuration and an abundance of features. It is included with Gpg4win, an encryption suite for Windows.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"http:\/\/www.claws-mail.org\/\">Website:\nclaws-mail.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, BSD, Solaris, Unix.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/github.com\/k9mail\/k-9\/releases\">K-9 Mail<\/a> - An independent mail application for Android. It supports both POP3 and IMAP mailboxes, but only supports push mail for IMAP.<\/li>\n<li><a href=\"https:\/\/www.gnupg.org\/\">GNU Privacy Guard<\/a> - Email Encryption. GnuPG is a GPL Licensed alternative to the PGP suite of cryptographic software. <a href=\"https:\/\/theprivacyguide.org\/tutorials\/gpg.html\">Tutorial.<\/a> Use <a href=\"https:\/\/gpgtools.org\/\">GPGTools for macOS.<\/a><\/li>\n<li><a href=\"https:\/\/www.mailpile.is\/\">Mailpile (Beta)<\/a> - A modern, fast web-mail client with user-friendly encryption and privacy features.<\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">I2P-Bote<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/I2P.png\" alt=\"I2P-Bote\" class=\"panel-item\" \/>I2P-Bote is a fully decentralized and distributed email system. It supports different identities and does not expose email headers. Currently (2015), it is still\nin beta version and can be accessed via its web application interface or IMAP and SMTP. All bote-mails are transparently end-to-end encrypted and, optionally, signed by the sender's private key.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/i2pbote.xyz\/\">Website:\ni2pbote.xyz\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, BSD \/ Solaris.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Bitmessage<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Bitmessage.png\" alt=\"Bitmessage\" class=\"panel-item\" \/>Bitmessage is a P2P communications protocol used to send encrypted messages to another person or to many subscribers. It is decentralized and trustless,\nmeaning that you need-not inherently trust any entities like root certificate authorities. It uses strong authentication which means that the sender of a message cannot be spoofed, and it aims to hide \"non-content\" data.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/bitmessage.org\/\">Website:\nbitmessage.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">RetroShare<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/RetroShare.png\" alt=\"RetroShare\" class=\"panel-item\" \/>Retroshare creates encrypted connections to your friends. Nobody can spy on you. Retroshare is completely decentralized. This means there are no central servers. It is entirely Open-Source and free. There are no costs, no ads and no Terms of Service.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"http:\/\/retroshare.net\">Website:\nretroshare.net\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><\/div><p><strong> If you are currently using search engines like Google, Bing or Yahoo you should pick an alternative here. <\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">searx<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/searx.jpg\" alt=\"searx\" class=\"panel-item\" \/>An <a href=\"https:\/\/github.com\/asciimoo\/searx\">open source<\/a> metasearch engine, aggregating the results of other search engines while not storing information about its users. No logs, no ads and no tracking.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/searx.me\/\">Website:\nsearx.me\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/ulrn6sryqaifefld.onion Requires specific software to access: torproject.org\" href=\"http:\/\/ulrn6sryqaifefld.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/github.com\/asciimoo\/searx\/wiki\/Searx-instances\">List of Instances<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">StartPage<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/StartPage.png\" alt=\"StartPage\" class=\"panel-item\" \/>Google search results, with complete privacy protection. Behind StartPage is a european company that has been obsessive about privacy since 2006.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/www.startpage.com\/\">Website:\nstartpage.com\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">DuckDuckGo<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/DuckDuckGo.jpg\" alt=\"DuckDuckGo\" class=\"panel-item\" \/>The search engine that doesn't track you. Some of DuckDuckGo's code is free software hosted at <a href=\"https:\/\/github.com\/duckduckgo\">GitHub<\/a>, but the core is proprietary.  <a href=\"#ukusa\">The company is based in the USA.<\/a><\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/duckduckgo.com\/\">Website:\nduckduckgo.com\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/3g2upl4pq6kufc4m.onion Requires specific software to access: torproject.org\" href=\"http:\/\/3g2upl4pq6kufc4m.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Firefox Addon<\/h3><ul><li>\n<a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/google-search-link-fix\/\">Google search link fix<\/a> - Firefox extension that prevents Google and Yandex search pages from modifying search result links when you click them. This is useful when\ncopying links but it also helps privacy by preventing the search engines from recording your clicks. (<a href=\"https:\/\/github.com\/palant\/searchlinkfix\">Open Source<\/a>)\n<\/li>\n<\/ul><h3>Worth Mentioning<\/h3><ul><li>\n<a href=\"https:\/\/www.findx.com\/\">findx<\/a> - Open Source. No logging. No tracking. Transparent algorithms. Hosted in Europe.\n<\/li>\n<li>\n<a href=\"https:\/\/www.qwant.com\/\">Qwant<\/a> - Qwant's philosophy is based on two principles: no user tracking and no filter bubble. Qwant was launched in France in February 2013. <a href=\"https:\/\/www.qwant.com\/privacy\">Privacy Policy.<\/a>\n<\/li>\n<li>\n<a href=\"https:\/\/metager.de\/en\/\">MetaGer<\/a> - An open source metasearch engine, which is based in Germany. It focuses on protecting the user's privacy.\n<\/li>\n<li>\n<a href=\"https:\/\/www.ixquick.eu\/\">ixquick.eu<\/a> - Returns the top results from multiple search engines. Based in the Netherlands.\n<\/li>\n<\/ul><p><strong>If you are currently using an Instant Messenger like WhatsApp, Viber, LINE, Telegram or Threema you should pick an alternative here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Mobile: Signal<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Signal.png\" alt=\"Mobile: Signal\" class=\"panel-item\" \/>Signal is a mobile app developed by Open Whisper Systems. The app provides instant messaging, as well as voice and video calling.\nAll communications are end-to-end encrypted. Signal is free and open source, enabling anyone to verify its security by auditing the code. The development team is supported by community donations and grants. There are no advertisements,\nand it doesn't cost anything to use.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/signal.org\">Website:\nsignal.org\n<\/a>\n<\/p><p>OS: Android, iOS, macOS, Windows, Linux, Web<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Wire<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/wire.png\" alt=\"Wire\" class=\"panel-item\" \/>Wire is an app developed by Wire Swiss GmbH.\nThe Wire app allows users to exchange end-to-end encrypted instant messages, as well as make voice and video calls. Wire is free and open source, enabling anyone to verify its security by auditing the code.\nThe development team is backed by Iconical and they will monetize in the future with premium features\/services.<br \/><b>Caution:<\/b> The company keeps a list of all the users you contact until you delete your account.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/get.wire.com\">Website:\nget.wire.com\n<\/a>\n<\/p><p>OS: Android, iOS, macOS, Windows, Linux, Web<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Desktop: Ricochet<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Ricochet.png\" alt=\"Desktop: Ricochet\" class=\"panel-item\" \/>Ricochet uses the <a href=\"#browser\"><i class=\"fas fa-link\"> Tor network<\/i><\/a> to reach your contacts without relying on messaging servers. It creates a hidden service, which is used to rendezvous with\nyour contacts without revealing your location or IP address. Instead of a username, you get a unique address that looks like <em>ricochet:rs7ce36jsj24ogfw<\/em>. Other Ricochet users can use this address to send a contact request - asking to be\nadded to your contacts list.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/ricochet.im\/\">Website:\nricochet.im\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/www.chatsecure.org\">ChatSecure<\/a> - ChatSecure is a free and open source messaging app that features OTR encryption over XMPP. <\/li>\n<li><a href=\"https:\/\/crypto.cat\/\">Cryptocat<\/a> - Encrypted open source messenger. Supports file sharing and multiple devices. Available for Windows, Linux and macOS.<\/li>\n<li><a href=\"http:\/\/kontalk.org\/\">Kontalk<\/a> - A community-driven instant messaging network. Supports end-to-end encryption. Both client-to-server and server-to-server channels are fully encrypted.<\/li>\n<li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=eu.siacs.conversations\">Conversations<\/a> - An open source Jabber\/XMPP client for Android 4.4+ smartphones. Supports end-to-end encryption with either OMEMO or openPGP. There is also <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=eu.siacs.conversations.legacy\">Conversations Legacy<\/a> which still supports OTR.<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Off-the-Record_Messaging#Client_support\">List of OTR Clients - Wikipedia<\/a><\/li>\n<li><a href=\"https:\/\/matrix.org\/\">The Matrix protocol<\/a> and it's various clients such as <a href=\"https:\/\/about.riot.im\/\">Riot.im<\/a>. A successor of XMPP.<\/li>\n<\/ul><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/motherboard.vice.com\/read\/ricochet-encrypted-messenger-tackles-metadata-problem-head-on\">Ricochet, the Messenger That Beats Metadata, Passes Security Audit | Motherboard<\/a><\/li>\n<li><a href=\"https:\/\/firstlook.org\/theintercept\/2015\/07\/14\/communicating-secret-watched\/\">Chatting in Secret While We're All Being Watched - firstlook.org<\/a><\/li>\n<li><a href=\"https:\/\/signal.org\/android\/apk\/\">Advanced users with special needs can download the Signal APK directly. Most users should not do this under normal circumstances.<\/a><\/li>\n<li><a href=\"https:\/\/motherboard.vice.com\/en_us\/article\/secure-messaging-app-wire-stores-everyone-youve-ever-contacted-in-plain-text\">Secure Messaging App Wire Stores Everyone You've Ever Contacted in Plain Text | Motherboard<\/a><\/li>\n<li><a href=\"https:\/\/security.stackexchange.com\/a\/49802\">Telegram encryption isn't secure.<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/privacytoolsIO\/comments\/6r655i\/telegram_isnt_safe\/\">Telegram is marketed as a private messenger, but it's not private by default.<\/a><\/li>\n<\/ul><p><strong>If you are currently using a Video &amp; Voice Messenger like Skype, Viber or Google Hangouts you should pick an alternative here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Mobile: Signal<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Signal.png\" alt=\"Mobile: Signal\" class=\"panel-item\" \/>Signal is a mobile app developed by Open Whisper Systems. The app provides instant messaging, as well as voice and video calling.\nAll communications are end-to-end encrypted. Signal is free and open source, enabling anyone to verify its security by auditing the code. The development team is supported by community donations and grants. There are no advertisements,\nand it doesn't cost anything to use.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/signal.org\">Website:\nsignal.org\n<\/a>\n<\/p><p>OS: Android, iOS, macOS, Windows, Linux, Web<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Wire<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/wire.png\" alt=\"Wire\" class=\"panel-item\" \/>Wire is an app developed by Wire Swiss GmbH.\nThe Wire app allows users to exchange end-to-end encrypted instant messages, as well as make voice and video calls. Wire is free and open source, enabling anyone to verify its security by auditing the code.\nThe development team is backed by Iconical and they will monetize in the future with premium features\/services.<br \/><b>Caution:<\/b> The company keeps a list of all the users you contact until you delete your account.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/get.wire.com\">Website:\nget.wire.com\n<\/a>\n<\/p><p>OS: Android, iOS, macOS, Windows, Linux, Web<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Linphone<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Linphone.png\" alt=\"Linphone\" class=\"panel-item\" \/>Linphone is an open source SIP Phone and a free voice over IP service, available on mobile and desktop environments and on web browsers. It supports ZRTP for end-to-end encrypted voice and video communication.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"http:\/\/www.linphone.org\/\">Website:\nlinphone.org\n<\/a>\n<\/p><p>OS: iOS, Android, Windows Phone, Linux, Windows, macOS, Browser (Web)<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/jitsi.org\/\">Jitsi<\/a> - Jitsi is a free and open source multiplatform voice (VoIP), videoconferencing and instant messaging application.<\/li>\n<li><a href=\"https:\/\/tox.chat\/\">Tox<\/a> - A free and open-source, peer-to-peer, encrypted instant messaging and video calling software.<\/li>\n<li><a href=\"http:\/\/www.ring.cx\/\">Ring (formerly SFLphone)<\/a> - Gives you a full control over your communications and an unmatched level of privacy.<\/li>\n<\/ul><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/signal.org\/android\/apk\/\">Advanced users with special needs can download the Signal APK directly. Most users should not do this under normal circumstances.<\/a><\/li>\n<li><a href=\"https:\/\/motherboard.vice.com\/en_us\/article\/secure-messaging-app-wire-stores-everyone-youve-ever-contacted-in-plain-text\">Secure Messaging App Wire Stores Everyone You've Ever Contacted in Plain Text | Motherboard<\/a><\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">OnionShare<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/OnionShare.png\" alt=\"OnionShare\" class=\"panel-item\" \/>OnionShare is an open source tool that lets you securely and anonymously share a file of any size. It works by starting a web server, making it accessible as a Tor onion service, and generating an unguessable URL to access and download the files. It doesn't require setting up a server on the internet somewhere or using a third-party filesharing service. You host the file on your own computer and use a Tor onion service to make it temporarily accessible over the internet.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/onionshare.org\/\">Website:\nonionshare.org\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/elx57ue5uyfplgva.onion\/ Requires specific software to access: torproject.org\" href=\"http:\/\/elx57ue5uyfplgva.onion\/\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Magic Wormhole<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Magic-Wormhole.png\" alt=\"Magic Wormhole\" class=\"panel-item\" \/>Get things from one computer to another, safely.This package provides a library and a command-line tool named wormhole, which makes it possible to get arbitrary-sized files and directories (or short pieces of text) from one computer to another. The two endpoints are identified by using identical \"wormhole codes\": in general, the sending machine generates and displays the code, which must then be typed into the receiving machine. The codes are short and human-pronounceable, using a phonetically-distinct wordlist. The receiving side offers tab-completion on the codewords, so usually only a few characters must be typed. Wormhole codes are single-use and do not need to be memorized.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"http:\/\/magic-wormhole.io\/\">Website:\nmagic-wormhole.io\n<\/a>\n<\/p><p>OS: cross-platform <a href=\"https:\/\/github.com\/warner\/magic-wormhole#license-compatibility\">(python)<\/a><\/p><\/div><\/div><\/div><\/div><\/div><p><strong>If you are currently using Cloud Storage Services like Dropbox, Google Drive, Microsoft OneDrive or Apple iCloud you should pick an alternative here.<\/strong><\/p><div class=\"row mb-4\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Nextcloud - Choose your hoster<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/Nextcloud.png\" alt=\"Nextcloud - Choose your hoster\" class=\"panel-item\" \/>Similar functionally to the widely used Dropbox, with the difference being that Nextcloud is free and open-source, and thereby allowing anyone to install and operate it without charge on a private server, with no limits on storage space or the number of connected clients.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/nextcloud.com\/\">Website:\nnextcloud.com\n<\/a>\n<\/p><p>Client OS: Windows, macOS, Linux, BSD, Unix, iOS, Android, Fire OS. Server: Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Least Authority S4 - For Experts<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/S4.jpg\" alt=\"Least Authority S4 - For Experts\" class=\"panel-item\" \/>S4 (Simple Secure Storage Service) is Least Authority's verifiably secure off-site backup system for individuals and businesses. 100% client-side encryption and open source transparency. 250GB for $9.95\/month or 5TB for $25.95\/month. Servers are hosted with Amazon S3 in the US.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/leastauthority.com\/\">Website:\nleastauthority.com\n<\/a>\n<\/p><p>OS: <a href=\"https:\/\/tahoe-lafs.org\/trac\/tahoe-lafs\/wiki\/Installation\">Linux<\/a>, <a href=\"https:\/\/tahoe-lafs.org\/trac\/tahoe-lafs\/browser\/trunk\/docs\/quickstart.rst\"> Windows, macOS, OpenSolaris, BSD.<\/a> (Installation for advanced users)<\/p><\/div><\/div><\/div><\/div><\/div><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/cryptomator.org\/\">Cryptomator<\/a> - Free client-side AES encryption for your cloud files. Open source software: No backdoors, no registration.<\/li>\n<\/ul><p><strong>If you are currently using a Cloud Storage Services like Dropbox, Google Drive, Microsoft OneDrive or Apple iCloud you should think about hosting it on your own.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Pydio<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Pydio.png\" alt=\"Pydio\" class=\"panel-item\" \/>Pydio is open source software that turns instantly any server (on-premise, NAS, cloud IaaS or PaaS) into a file sharing platform for your company. It is an alternative\nto SaaS Boxes and Drives, with more control, safety and privacy, and favorable TCOs.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/pydio.com\/\">Website:\npydio.com\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, iOS, Android.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Tahoe-LAFS<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Tahoe-LAFS.png\" alt=\"Tahoe-LAFS\" class=\"panel-item\" \/>Tahoe-LAFS is a Free and Open decentralized cloud storage system. It distributes your data across multiple servers. Even if some of the servers fail or are\ntaken over by an attacker, the entire file store continues to function correctly, preserving your privacy and security.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/www.tahoe-lafs.org\/\">Website:\ntahoe-lafs.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Nextcloud<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/Nextcloud.png\" alt=\"Nextcloud\" class=\"panel-item\" \/>Similar functionally to the widely used Dropbox, with the difference being that Nextcloud is free and open-source, and thereby allowing anyone to install and operate it without charge on a private server, with no limits on storage space or the number of connected clients.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/nextcloud.com\/\">Website:\nnextcloud.com\n<\/a>\n<\/p><p>Client OS: Windows, macOS, Linux, BSD, Unix, iOS, Android, Fire OS. Server: Linux.<\/p><\/div><\/div><\/div><\/div><\/div><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">SparkleShare<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/SparkleShare.png\" alt=\"SparkleShare\" class=\"panel-item\" \/>SparkleShare creates a special folder on your computer. You can add remotely hosted folders (or \"projects\") to this folder. These projects will be automatically\nkept in sync with both the host and all of your peers when someone adds, removes or edits a file.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"http:\/\/sparkleshare.org\/\">Website:\nsparkleshare.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Syncthing<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Syncthing.png\" alt=\"Syncthing\" class=\"panel-item\" \/>Syncthing replaces proprietary sync and cloud services with something open, trustworthy and decentralized. Your data is your data alone and you deserve to\nchoose where it is stored, if it is shared with some third-party and how it's transmitted over the Internet.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/syncthing.net\/\">Website:\nsyncthing.net\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, BSD, Solaris.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li>\n<a href=\"https:\/\/git-annex.branchable.com\/\">git-annex<\/a> - Allows managing files with git, without checking the file contents into git. While that may seem paradoxical, it is useful when dealing with files larger than git can currently easily handle,\nwhether due to limitations in memory, time, or disk space.\n<\/li>\n<\/ul><p><strong>If you are currently using a password manager software like 1Password, LastPass, Roboform or iCloud Keychain you should pick an alternative here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Master Password - Cross-platform<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Master-Password.jpg\" alt=\"Master Password - Cross-platform\" class=\"panel-item\" \/>Master Password is based on an ingenious password generation algorithm that guarantees your passwords can never be lost. Its passwords aren't stored:\nthey are generated on-demand from your name, the site and your master password. No syncing, backups or internet access needed.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/masterpassword.app\/\">Website:\nmasterpassword.app\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, iOS, Android, Web.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">KeePass \/ KeePassX - Local<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/KeePass.png\" alt=\"KeePass \/ KeePassX - Local\" class=\"panel-item\" \/>KeePass is a free open source password manager, which helps you to manage your passwords in a secure way. All passwords in one database, which is locked with one\nmaster key or a key file. The databases are encrypted using the best and most secure encryption algorithms currently known: AES and Twofish. See also: <a href=\"http:\/\/www.keepassx.org\/\">KeePassX (not regularly updated)<\/a> and <a href=\"https:\/\/keepassxc.org\/\">KeePassXC<\/a>.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"http:\/\/keepass.info\/download.html\">Website:\nkeepass.info\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, iOS, Android, BSD.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">LessPass - Browser<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/LessPass.png\" alt=\"LessPass - Browser\" class=\"panel-item\" \/>LessPass is a free and open source password manager that generates unique passwords for websites, email accounts, or anything else based on a master password and information you know. No sync needed. Uses PBKDF2 and SHA-256. It's advised to use the browser addons for more security.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/lesspass.com\/\">Website:\nlesspass.com\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li>\n<a href=\"https:\/\/www.privacytools.io\/password.html\">Secure Password Generator<\/a> - generates a unique set of custom, high quality, cryptographic-strength password strings which are safe for you to use.\n<\/li>\n<li>\n<a href=\"http:\/\/www.supergenpass.com\/\">SuperGenPass<\/a> - A master password and the domain name of the Web site you are visiting is used as the \u201cseed\u201d for a one-way hash algorithm (base-64 MD5). The output of this algorithm is your generated password.\nYou remember one password (your \u201cmaster password\u201d), and SGP uses it to generate unique, complex passwords for the Web sites you visit. Your generated passwords are never stored or transmitted, so you can use SGP on as many computers as you like without\nhaving to \u201csync\u201d anything.\n<\/li>\n<li>\n<a href=\"http:\/\/pwsafe.org\/\">Password Safe<\/a> - Whether the answer is one or hundreds, Password Safe allows you to safely and easily create a secured and encrypted username\/password list. With Password Safe all you have to do is create and remember\na single \"Master Password\" of your choice in order to unlock and access your entire username\/password list.\n<\/li>\n<\/ul><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=yzGzB-yYKcc\">Edward Snowden on Passwords - YouTube<\/a><\/li>\n<\/ul><p><strong>If you are currently using a calendar and or contacts synchronization service like Google Sync or iCloud you should pick an alternative here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Nextcloud<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/Nextcloud.png\" alt=\"Nextcloud\" class=\"panel-item\" \/>NextCloud is a suite of client-server software for creating and using file hosting services. This includes calendar sync via CalDAV and contacts sync via CardDAV. Nextcloud is free and open-source, thereby allowing anyone to install and operate it without charge on a private server.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/nextcloud.com\/\">Website:\nnextcloud.com\n<\/a>\n<\/p><p>Client OS: Windows, macOS, Linux, BSD, Unix, iOS, Android, Fire OS. Server: Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Email Providers<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/misc\/email.png\" alt=\"Email Providers\" class=\"panel-item\" \/>Many email providers also offer calendar and or contacts sync services. Refer to the <a href=\"#email\">Email Providers section<\/a> to choose an email provider and check if they also offer calendar and or contacts sync.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"#email\">Website:\nEmail Providers section\n<\/a>\n<\/p><p>OS: depends on email provider.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">EteSync<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/etesync.png\" alt=\"EteSync\" class=\"panel-item\" \/>EteSync is a secure, end-to-end encrypted and journaled personal information (e.g. contacts and calendar) cloud synchronization and backup for Android and any OS that supports CalDAV\/CardDAV. It costs $14 per year to use, or you can host the server yourself for free.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/www.etesync.com\/\">Website:\netesync.com\n<\/a>\n<\/p><p>Client OS: Android, Web. Server OS: Linux.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li>\n<a href=\"https:\/\/fruux.com\/\">fruux<\/a> - a unified contacts\/calendaring system that works across platforms and devices.\n<\/li>\n<li>\n<a href=\"https:\/\/signal.org\/blog\/flock\/\">Flock<\/a> - a discontinued calendar and contacts sync service by Open Whisper Systems. Despite being shut down the <a href=\"https:\/\/github.com\/signalapp\/Flock\">source code<\/a> is still available on GitHub.\n<\/li>\n<li>\n<a href=\"#cloud\">cloud backups<\/a> - consider regularly exporting your calendar and or contacts and backing them up on a separate storage drive or uploading them to cloud storage (ideally after <a href=\"#encrypt\">encrypting<\/a> them).\n<\/li>\n<\/ul><h3>Who is required to hand over the encryption keys to authorities?<\/h3><p>Mandatory <a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law\">key disclosure laws<\/a> require individuals to turn over encryption keys to law enforcement conducting a criminal investigation. How these laws are implemented (who may be legally compelled to assist) vary from nation to nation, but a warrant\nis generally required. Defenses against key disclosure laws include steganography and encrypting data in a way that provides plausible deniability.<\/p><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Steganography\">Steganography<\/a> involves hiding sensitive information (which may be encrypted) inside of ordinary data (for example, encrypting an image file and then hiding it in an audio file). With plausible deniability, data is encrypted in a way that prevents an\nadversary from being able to prove that the information they are after exists (for example, one password may decrypt benign data and another password, used on the same file, could decrypt sensitive data).<\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-danger\"><p><\/p><h3 class=\"h5\">Key disclosure laws apply<\/h3><div class=\"card-body\">1.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Antigua_and_Barbuda\">Antigua and Barbuda<\/a><br \/>2.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Australia\">Australia<\/a><br \/>3.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Canada\">Canada<\/a><br \/>4.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#France\">France<\/a><br \/>5.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#India\">India<\/a><br \/>6.<a href=\"https:\/\/edri.org\/norway-introduces-forced-biometric-authentication\/\">Norway<\/a><br \/>7.<a href=\"https:\/\/www.bloomberg.com\/news\/articles\/2018-03-20\/telegram-loses-bid-to-stop-russia-from-getting-encryption-keys\">Russia<\/a><br \/>8.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#South_Africa\">South Africa<\/a><br \/>9.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#United_Kingdom\">United Kingdom<\/a><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Key disclosure laws may apply<\/h3><div class=\"card-body\">1.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Belgium\">Belgium *<\/a><br \/>2.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Finland\">Finland *<\/a><br \/>3.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#New_Zealand\">New Zealand<\/a>(unclear)<br \/>4.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Sweden\">Sweden<\/a>(proposed)<br \/>5.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#The_Netherlands\">The Netherlands *<\/a><br \/>6.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#United_States\">United States<\/a>(see related information)<\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Key disclosure laws don't apply<\/h3><div class=\"card-body\">1.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Czech_Republic\">Czech Republic<\/a><br \/>2.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Poland\">Poland<\/a><br \/>3.<a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law#Germany\">Germany<\/a><\/div><\/div><\/div><\/div><p> * (people who know how to access a system may be ordered to share their knowledge, <strong>however, this doesn't apply to the suspect itself or family members.<\/strong>)<\/p><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Key_disclosure_law\">Wikipedia page on key disclosure law<\/a><\/li>\n<li><a href=\"https:\/\/law.stackexchange.com\/questions\/1523\/can-a-us-citizen-be-required-to-provide-the-authentication-key-for-encrypted-dat\">law.stackexchange.com question about key disclosure law in US<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=Jt7D4AIfqlQ\">DEFCON 20: Crypto and the Cops: the Law of Key Disclosure and Forced Decryption<\/a><\/li>\n<\/ul><p><strong>If you are currently not using encryption software for your hard disk, emails or file archives you should pick an encryption software here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">VeraCrypt - Disk Encryption<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/VeraCrypt.png\" alt=\"VeraCrypt - Disk Encryption\" class=\"panel-item\" \/>VeraCrypt is a source-available freeware utility used for on-the-fly encryption. It can create a virtual encrypted disk within a file or encrypt a partition\nor the entire storage device with pre-boot authentication. VeraCrypt is a fork of the discontinued TrueCrypt project. It was initially released on June 22, 2013. According to its developers, security improvements have been implemented and issues\nraised by the initial TrueCrypt code audit have been addressed.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/veracrypt.fr\/\">Website:\nveracrypt.fr\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">GNU Privacy Guard - Email Encryption<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/GnuPG.png\" alt=\"GNU Privacy Guard - Email Encryption\" class=\"panel-item\" \/>GnuPG is a GPL Licensed alternative to the PGP suite of cryptographic software. GnuPG is compliant with RFC 4880, which is the current IETF standards track specification\nof OpenPGP. Current versions of PGP (and Veridis' Filecrypt) are interoperable with GnuPG and other OpenPGP-compliant systems. GnuPG is a part of the Free Software Foundation's GNU software project, and has received major funding from the German\ngovernment. <a href=\"https:\/\/gpgtools.org\/\">GPGTools for macOS<\/a>.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/gpgtools.org\/\">Website:\ngpgtools.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, BSD.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">PeaZip - File Archive Encryption<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/PeaZip.png\" alt=\"PeaZip - File Archive Encryption\" class=\"panel-item\" \/>PeaZip is a free and open-source file manager and file archiver made by Giorgio Tani. It supports its native PEA archive format (featuring compression, multi volume\nsplit and flexible authenticated encryption and integrity check schemes) and other mainstream formats, with special focus on handling open formats. It supports 181 file extensions (as of version 5.5.1).<\/p><p><strong>macOS alternative:<\/strong> <a href=\"http:\/\/www.kekaosx.com\/\">Keka<\/a> is a free file archiver.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"http:\/\/www.peazip.org\/\">Website:\npeazip.org\n<\/a>\n<\/p><p>OS: Windows, Linux, BSD.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/cryptomator.org\/\">Cryptomator<\/a> - Free client-side AES encryption for your cloud files. Open source software: No backdoors, no registration.<\/li>\n<li><a href=\"https:\/\/minilock.io\/\">miniLock<\/a> - Browser plugin for Google Chrome \/ Chromium to encrypt files using a secret passphrase. Easy to use. From the developer of Cryptocat.<\/li>\n<li><a href=\"https:\/\/www.aescrypt.com\/\">AES Crypt<\/a> - Using a powerful 256-bit encryption algorithm, AES Crypt can safely secure your most sensitive files. For Windows, macOS, Linux and Android.<\/li>\n<li><a href=\"https:\/\/diskcryptor.net\/\">DiskCryptor<\/a> - A full disk and partition encryption system for Windows including the ability to encrypt the partition and disk on which the OS is installed.<\/li>\n<li><a href=\"https:\/\/gitlab.com\/cryptsetup\/cryptsetup\/\">Linux Unified Key Setup (LUKS)<\/a> - A full disk encryption system for Linux using dm-crypt as the disk encryption backend. Included by default in Ubuntu. Available for Windows and Linux.\n<\/li><\/ul><p><strong>If you are currently browsing <a href=\"https:\/\/en.wikipedia.org\/wiki\/Surface_Web\">clearnet<\/a> and want to access the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Dark_web\">dark web<\/a>, this section is for you.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Tor Browser<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Tor-Project.png\" alt=\"Tor Browser\" class=\"panel-item\" \/>The Tor network is a group of volunteer-operated servers that allows people to improve their privacy and security on the Internet. Tor's users employ this network by connecting through a series of virtual tunnels rather than making a direct connection, thus allowing both organizations and individuals to share information over public networks without compromising their privacy. Tor is an effective censorship circumvention tool.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/www.torproject.org\/\">Website:\ntorproject.org\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/expyuzz4wqqyqhjn.onion Requires specific software to access: torproject.org\" href=\"http:\/\/expyuzz4wqqyqhjn.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><p>OS: Windows, macOS, Linux, <a href=\"https:\/\/mike.tig.as\/onionbrowser\/\">iOS<\/a>, <a href=\"https:\/\/www.torproject.org\/docs\/android.html.en\">Android<\/a>, <a href=\"https:\/\/github.com\/torbsd\/openbsd-ports\">OpenBSD.<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">I2P Anonymous Network<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/I2P.png\" alt=\"I2P Anonymous Network\" class=\"panel-item\" \/>The Invisible Internet Project (I2P) is a computer network layer that allows applications to send messages to each other pseudonymously and securely. Uses include anonymous\nWeb surfing, chatting, blogging and file transfers. The software that implements this layer is called an I2P router and a computer running I2P is called an I2P node. The software is free and open source and is published under multiple licenses.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/geti2p.net\/\">Website:\ngeti2p.net\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, BSD \/ Solaris.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">The Freenet Project<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Freenet.png\" alt=\"The Freenet Project\" class=\"panel-item\" \/>Freenet is a peer-to-peer platform for censorship-resistant communication. It uses a decentralized distributed data store to keep and deliver information, and\nhas a suite of free software for publishing and communicating on the Web without fear of censorship. Both Freenet and some of its associated tools were originally designed by Ian Clarke, who defined Freenet's goal as providing freedom of speech\non the Internet with strong anonymity protection.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/freenetproject.org\/\">Website:\nfreenetproject.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/zeronet.io\/\">ZeroNet<\/a> - Open, free and uncensorable websites, using Bitcoin cryptography and BitTorrent network.<\/li>\n<li><a href=\"http:\/\/retroshare.net\/\">RetroShare<\/a> - Open Source cross-platform, Friend-2-Friend and secure decentralised communication platform.<\/li>\n<li><a href=\"https:\/\/gnunet.org\/\">GNUnet<\/a> - GNUnet provides a strong foundation of free software for a global, distributed network that provides security and privacy.<\/li>\n<li><a href=\"https:\/\/ipfs.io\/\">IPFS<\/a> - A peer-to-peer hypermedia protocol to make the web faster, safer, and more open. (<a href=\"https:\/\/github.com\/privacytoolsIO\/privacytools.io\/pull\/361#issuecomment-344414022\">important privacy warning<\/a>)<\/li>\n<\/ul><p><strong> If you are currently using Social Networks like Facebook, Twitter or Google+, you should pick an alternative here. <\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">diaspora*<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/diaspora.png\" alt=\"diaspora*\" class=\"panel-item\" \/>diaspora* is based on three key philosophies: Decentralization, freedom and privacy. It is intended to address privacy concerns related to centralized social\n networks by allowing users set up their own server (or \"pod\") to host content; pods can then interact to share status updates, photographs, and other social data.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/diasporafoundation.org\/\">Website:\ndiasporafoundation.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Friendica<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Friendica.png\" alt=\"Friendica\" class=\"panel-item\" \/>Friendica has an emphasis on extensive privacy settings and easy server installation. It aims to federate with as many other social networks as possible. Currently,\nFriendica users can integrate contacts from Facebook, Twitter, Diaspora, GNU social, App.net, Pump.io and other services in their social streams.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"http:\/\/friendi.ca\/\">Website:\nfriendi.ca\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">GNU social<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/GNU-social.png\" alt=\"GNU social\" class=\"panel-item\" \/>While offering functionality similar to Twitter, GNU social seeks to provide the potential for open, inter-service and distributed communications between\nmicroblogging communities. Enterprises and individuals can install and control their own services and data. Notable public deployments are <a href=\"https:\/\/quitter.se\/\">quitter.se<\/a> and <a href=\"https:\/\/gnusocial.no\">gnusocial.no<\/a>.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/gnu.io\/social\/try\/\">Website:\ngnu.io\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/mastodon.social\/about\">Mastodon<\/a> - An open sourced, free and decentralized social network built on top of GNU Social.<\/li>\n<\/ul><h3>Related Information<\/h3><ul><li><a href=\"https:\/\/addons.mozilla.org\/firefox\/addon\/facebook-container\/\">Facebook Container by Mozilla<\/a> - Prevent Facebook from tracking you around the web.<\/li>\n<li><a href=\"https:\/\/www.facebook.com\/help\/delete_account\">Delete your Facebook account<\/a> - Direct link to delete your Facebook account without being able to reactivate it again.<\/li>\n<li><a href=\"http:\/\/deletefacebook.com\/\">How To Permanently Delete A Facebook Account<\/a> - This guide will take you through a smooth and successful Facebook account deletion.<\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Njalla - Domain Registration<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/provider\/Njalla.png\" alt=\"Njalla - Domain Registration\" class=\"panel-item\" \/>Njalla only needs your email or jabber address in order to register a domain name for you. Created by people from The Pirate Bay and IPredator VPN. Accepted Payments: Bitcoin, Litecoin, Monero, DASH, Bitcoin Cash and PayPal. A privacy-aware domain registration service.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/njal.la\/\">Website:\nnjal.la\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/njalladnspotetti.onion Requires specific software to access: torproject.org\" href=\"http:\/\/njalladnspotetti.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">DNSCrypt - Tool<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/DNSCrypt.png\" alt=\"DNSCrypt - Tool\" class=\"panel-item\" \/>A protocol for securing communications between a client and a DNS resolver. The DNSCrypt protocol uses high-speed high-security elliptic-curve cryptography and\nis very similar to DNSCurve, but focuses on securing communications between a client and its first-level resolver.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/dnscrypt.info\/\">Website:\ndnscrypt.info\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">OpenNIC - Service<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/OpenNIC.png\" alt=\"OpenNIC - Service\" class=\"panel-item\" \/>OpenNIC is an alternate network information center\/alternative DNS root which lists itself as an alternative to ICANN and its registries. Like all alternative\nroot DNS systems, OpenNIC-hosted domains are unreachable to the vast majority of the Internet. Only specific configuration in one's DNS resolver makes these reachable, and very few Internet service providers have this configuration.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"http:\/\/www.opennicproject.org\/\">Website:\nopennicproject.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/github.com\/quidsup\/notrack\">NoTrack<\/a> - A network-wide DNS server which blocks Tracking sites. Currently works in Debian and Ubuntu.<\/li>\n<li><a href=\"https:\/\/namecoin.info\/\">Namecoin<\/a> - A decentralized DNS open source information registration and transfer system based on the Bitcoin cryptocurrency.<\/li>\n<li><a href=\"https:\/\/pi-hole.net\/\">Pi-hole<\/a> - A network-wide DNS server for the Raspberry Pi. Blocks advertising and tracking domains for all devices on your network.<\/li>\n<\/ul><p><strong>If you are currently using an application like Evernote, Google Keep or Microsoft OneNote you should pick an alternative here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Laverna<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Laverna.png\" alt=\"Laverna\" class=\"panel-item\" \/>Laverna is a JavaScript note-taking web application with a Markdown editor and encryption support. The application stores all your notes in your browser databases, which is good for security and privacy reasons, because only you have access.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/laverna.cc\/\">Website:\nlaverna.cc\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Web.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Standard Notes<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/StandardNotes.png\" alt=\"Standard Notes\" class=\"panel-item\" \/>Standard Notes is a simple and private notes app that makes your notes easy and available everywhere you are. Features end-to-end encryption on every platform, and a powerful desktop experience with themes and custom editors.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/standardnotes.org\/\">Website:\nstandardnotes.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, iOS, Android, Web.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Turtl<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Turtl.png\" alt=\"Turtl\" class=\"panel-item\" \/>Turtl lets you take notes, bookmark websites, and store documents for sensitive projects. From sharing passwords with your coworkers to tracking research on an article you're writing, Turtl keeps it all safe from everyone but you and those you share with.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/turtlapp.com\/\">Website:\nturtlapp.com\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"http:\/\/paperwork.rocks\/\">Paperwork<\/a> - Open source and self-hosted solution. For PHP \/ MySQL servers.<\/li>\n<li><a href=\"https:\/\/orgmode.org\">Org-mode<\/a> - A major mode for GNU Emacs. Org-mode is for keeping notes, maintaining TODO lists, planning projects, and authoring documents with a fast and effective plain-text system. <\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">PrivateBin<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/PrivateBin.png\" alt=\"PrivateBin\" class=\"panel-item\" \/>PrivateBin is a minimalist, open source online pastebin where the server has zero knowledge of pasted data. Data is encrypted\/decrypted in the browser using 256bit AES. It is the improved version of ZeroBin.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/privatebin.net\/\">Website:\nprivatebin.net\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Zerobin.net<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/ZeroBin.png\" alt=\"Zerobin.net\" class=\"panel-item\" \/>Zerobin is a minimalist, open source online pastebin where the server has zero knowledge of pasted data. Data is encrypted\/decrypted in the browser using 256bit AES.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/zerobin.net\/\">Website:\nzerobin.net\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/zerobinqmdqd236y.onion Requires specific software to access: torproject.org\" href=\"http:\/\/zerobinqmdqd236y.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Ghostbin<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Ghostbin.png\" alt=\"Ghostbin\" class=\"panel-item\" \/>Ghostbin supports encryption, expiration, sessions, grant users to edit your notes and pastes up to one megabyte. You can also create your own account to keep track of your pastes.<br \/>The software is good but ghostbin.com's security is <a href=\"https:\/\/github.com\/privacytoolsIO\/privacytools.io\/issues\/454\">worrisome<\/a>.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/ghostbin.com\/\">Website:\nghostbin.com\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/disroot.org\/\">disroot.org<\/a> - Free privacy-friendly service that offers PrivateBin and other applications.<\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Etherpad<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Etherpad.png\" alt=\"Etherpad\" class=\"panel-item\" \/>Etherpad is a highly customizable Open Source online editor providing collaborative editing in really real-time. Etherpad allows you to edit documents collaboratively in real-time, much like a live multi-player editor that runs in your browser. Write articles, press releases, to-do lists, etc.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"http:\/\/etherpad.org\/\">Website:\netherpad.org\n<\/a>\n<\/p><p>OS: Windows, macOS, Linux.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Write.as<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/WriteAs.png\" alt=\"Write.as\" class=\"panel-item\" \/>Write.as is a cross-platform, privacy-oriented blogging platform. It's anonymous by default, letting you publish without signing up. If you create an account, it doesn't require any personal information. No ads, distraction-free, and built on a sustainable business model.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/write.as\/\">Website:\nwrite.as\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/writeas7pm7rcdqg.onion Requires specific software to access: torproject.org\" href=\"http:\/\/writeas7pm7rcdqg.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><p>OS: Windows, macOS, Linux, Android, iOS, Chrome OS, Web.<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">ProtectedText<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/ProtectedText.png\" alt=\"ProtectedText\" class=\"panel-item\" \/>ProtectedText is an open source web application. It encrypts and decrypts text in the browser, and password (or its hash) is never sent to the server - so that text can't be decrypted even if requested by authorities. No cookies, no sessions, no registration, no users tracking.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/www.protectedtext.com\/\">Website:\nprotectedtext.com\n<\/a>\n<\/p><p>OS: All Browsers.<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"https:\/\/ethercalc.net\/\">EtherCalc<\/a> - EtherCalc is a web spreadsheet. Data is saved on the web, and people can edit the same document at the same time. Changes are instantly reflected on all screens. Work together on inventories, survey forms, list management, brainstorming sessions.<\/li>\n<li><a href=\"https:\/\/disroot.org\/\">disroot.org<\/a> - Free privacy-friendly service that offers Etherpad, EtherCalc and PrivateBin.<\/li>\n<li><a href=\"https:\/\/dudle.inf.tu-dresden.de\/privacy\/\">dudle<\/a> - An online scheduling application, which is free and OpenSource. Schedule meetings or make small online polls. No email collection or the need of registration.<\/li>\n<\/ul><p><strong>If you are currently using a operating system like Microsoft Windows or Apple macOS you should pick an alternative here.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Qubes OS<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Qubes-OS.png\" alt=\"Qubes OS\" class=\"panel-item\" \/>Qubes is an open-source operating system designed to provide strong security for desktop computing. Qubes is based on Xen, the X Window System, and Linux, and can run most Linux applications and utilize most of the Linux drivers.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/www.qubes-os.org\/\">Website:\nqubes-os.org\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/qubesos4rrrrz6n4.onion Requires specific software to access: torproject.org\" href=\"http:\/\/qubesos4rrrrz6n4.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Debian<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Debian.png\" alt=\"Debian\" class=\"panel-item\" \/>Debian is a Unix-like computer operating system and a Linux distribution that is composed entirely of free and open-source software, most of which is under the GNU General Public License, and packaged by a group of individuals known as the Debian project.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/www.debian.org\/\">Website:\ndebian.org\n<\/a>\n<a class=\"mb-1\" data-toggle=\"tooltip\" data-placement=\"bottom\" data-original-title=\"http:\/\/sejnfjrq6szgca7v.onion Requires specific software to access: torproject.org\" href=\"http:\/\/sejnfjrq6szgca7v.onion\"><img alt=\"Tor\" src=\"https:\/\/www.privacytools.io\/img\/layout\/tor.png\" width=\"35\" \/><\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Trisquel<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Trisquel.png\" alt=\"Trisquel\" class=\"panel-item\" \/>Trisquel is a Linux-based operating system derived from Ubuntu. The project aims for a fully free software system without proprietary software or firmware and uses Linux-libre, a version of the Linux kernel with the non-free code (binary blobs) removed.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"http:\/\/trisquel.info\/\">Website:\ntrisquel.info\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Warning<\/h3><ul><li><a href=\"#win10\"><i class=\"fas fa-link\"> Don't use Windows 10 - It's a privacy nightmare<\/i><\/a><\/li>\n<\/ul><h3>Worth Mentioning<\/h3><ul><li><a href=\"http:\/\/www.openbsd.org\/\">OpenBSD<\/a> - A project that produces a free, multi-platform 4.4BSD-based UNIX-like operating system. Emphasizes portability, standardization, correctness, proactive security and integrated cryptography.<\/li>\n<li><a href=\"https:\/\/www.archlinux.org\/\">Arch Linux<\/a> - A simple, lightweight Linux distribution. It is composed predominantly of free and open-source software, and supports community involvement. <a href=\"https:\/\/www.parabola.nu\/\">Parabola<\/a> is a\ncompletely open source version of Arch Linux.<\/li>\n<li><a href=\"https:\/\/www.whonix.org\/\">Whonix<\/a> - A Debian GNU\/Linux based security-focused Linux distribution. It aims to provide privacy, security and anonymity on the internet. The operating system consists of two virtual machines, a \"Workstation\"\nand a Tor \"Gateway\". All communication are forced through the Tor network to accomplish this.<\/li>\n<li><a href=\"https:\/\/subgraph.com\/\">Subgraph OS<\/a> - Another Debian based Linux distribution, it features security hardening which makes it more resistant to security vulnerabilities. Subgraph runs many desktop applications in a security sandbox to limit their risk in case of compromise.\nBy default, it anonymizes Internet traffic by sending it through the Tor network. Note: It is still in alpha, and much testing and bug fixing still has to be done.<\/li>\n<\/ul><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Tails<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Tails.gif\" alt=\"Tails\" class=\"panel-item\" \/>Tails is a live operating system, that starts on almost any computer from a DVD, USB stick, or SD card. It aims at preserving privacy and anonymity, and helps to: Use the Internet anonymously and circumvent censorship; Internet connections go through the Tor network; leave no trace on the computer; use state-of-the-art cryptographic tools to encrypt files, emails and instant messaging.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/tails.boum.org\/\">Website:\ntails.boum.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">KNOPPIX<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/KNOPPIX.png\" alt=\"KNOPPIX\" class=\"panel-item\" \/>Knoppix is an operating system based on Debian designed to be run directly from a CD \/ DVD (Live CD) or a USB flash drive (Live USB), one of the first of its kind for any operating system. When starting a program, it is loaded from the removable medium and decompressed into a RAM drive. The decompression is transparent and on-the-fly.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"http:\/\/www.knopper.net\/knoppix\/\">Website:\nknopper.net\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Puppy Linux<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Puppy-Linux.png\" alt=\"Puppy Linux\" class=\"panel-item\" \/>Puppy Linux operating system is a lightweight Linux distribution that focuses on ease of use and minimal memory footprint. The entire system can be run from RAM with current versions generally taking up about 210 MB, allowing the boot medium to be removed after the operating system has started.<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"http:\/\/puppylinux.org\/\">Website:\npuppylinux.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"http:\/\/distro.ibiblio.org\/tinycorelinux\/\">Tiny Core Linux<\/a> - A minimal Linux operating system focusing on providing a base system using BusyBox and FLTK. The distribution is notable for its size (15 MB) and minimalism, with additional functionality\nprovided by extensions.<\/li>\n<\/ul><p><strong>Even though the source code of the following OS is provided, installing Google Apps may compromise your setup. The MicroG project can serve as a FLOSS replacement, depending on your threat model.<\/strong><\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">LineageOS<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/LineageOS.png\" alt=\"LineageOS\" class=\"panel-item\" \/>LineageOS is a free and open-source operating system for smartphones and tablets, based on the official releases of Android by Google. It is the continuation of the CyanogenMod project.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/www.lineageos.org\/\">Website:\nlineageos.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Sailfish OS<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/Sailfish-OS.png\" alt=\"Sailfish OS\" class=\"panel-item\" \/>Sailfish OS is a mobile operating system combining the Linux kernel for a particular hardware platform use, the open-source Mer core middleware, a proprietary UI contributed by Jolla, and other third-party components.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/sailfishos.org\/\">Website:\nsailfishos.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"http:\/\/www.replicant.us\/\">Replicant<\/a> - An open-source operating system based on Android, aiming to replace all proprietary components with free software.<\/li>\n<li><a href=\"http:\/\/omnirom.org\/\">OmniROM<\/a> - A free software operating system for smartphones and tablet computers, based on the Android mobile platform.<\/li>\n<li><a href=\"https:\/\/microg.org\/\">MicroG<\/a> - A project that aims to reimplement the proprietary Google Play Services in the Android operating system with a FLOSS replacement.\n<\/li><\/ul><p><strong>Improve your privacy with these add-ons for Android.<\/strong><\/p><h3>Control your traffic with \"NetGuard\"<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/netguard.gif\" class=\"img-fluid float-left\" alt=\"NetGuard\" \/><p><strong>NetGuard<\/strong> provides simple and advanced ways to block certain apps' access to the internet without the help of root privileges. Applications and addresses can individually be allowed or denied access to your Wi-Fi and\/or mobile connection, allowing you to control which apps are able to call home or not.\n<a href=\"https:\/\/www.netguard.me\/\">https:\/\/www.netguard.me\/<\/a><\/p><p><strong>The following add-ons require root access, which makes your device more vulnerable. Proceed with caution.<\/strong><\/p><h3>Manage your apps' permissions with XPrivacyLua<\/h3><img src=\"https:\/\/www.privacytools.io\/img\/addons\/xprivacylua.gif\" class=\"img-fluid float-left\" alt=\"XPrivacyLua\" \/><p>Revoking Android permissions from apps often let apps crash or malfunction. <strong>XPrivacyLua<\/strong> solves this by feeding apps fake data instead of real data.\n<a href=\"https:\/\/lua.xprivacy.eu\/\">https:\/\/lua.xprivacy.eu\/<\/a><\/p><p> Requirements: <a href=\"https:\/\/forum.xda-developers.com\/apps\/magisk\/official-magisk-v7-universal-systemless-t3473445\">Magisk (xda-developers)<\/a>\n<a href=\"https:\/\/forum.xda-developers.com\/xposed\">Xposed Framework (xda-developers)<\/a>\n<\/p><div class=\"row mb-2\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">OpenWrt<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/OpenWrt.png\" alt=\"OpenWrt\" class=\"panel-item\" \/>OpenWrt is an operating system (in particular, an embedded operating system) based on the Linux kernel, primarily used on embedded devices to route network traffic. The main components are the Linux kernel, util-linux, uClibc and BusyBox. All components have been optimized for size, to be small enough for fitting into the limited storage and memory available in home routers.<\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/openwrt.org\/\">Website:\nopenwrt.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">pfSense<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/pfSense.png\" alt=\"pfSense\" class=\"panel-item\" \/>pfSense is an open source firewall\/router computer software distribution based on FreeBSD. It is installed on a computer to make a dedicated firewall\/router for a network and is noted for its reliability and offering features often only found in expensive commercial firewalls. pfSense is commonly deployed as a perimeter firewall, router, wireless access point, DHCP server, DNS server, and as a VPN endpoint.<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/pfsense.org\/\">Website:\npfsense.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">LibreCMC<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/tools\/LibreCMC.png\" alt=\"LibreCMC\" class=\"panel-item\" \/>LibreCMC is a GNU\/Linux-libre distribution for computers with minimal resources, such as the Ben Nanonote, ath9k-based Wi-Fi routers, and other hardware with emphasis on free software. The project's current goal is to aim for compliance with the GNU Free System Distribution Guidelines (GNU FSDG) and ensure that the project continues to meet these requirements set forth by the Free Software Foundation (FSF)..<\/p><div><p>\n<a class=\"btn btn-warning mb-1\" href=\"https:\/\/librecmc.org\">Website:\nlibrecmc.org\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><\/div><h3>Worth Mentioning<\/h3><ul><li><a href=\"http:\/\/www.openbsd.org\/\">OpenBSD<\/a> - A project that produces a free, multi-platform 4.4BSD-based UNIX-like operating system. Emphasizes portability, standardization, correctness, proactive security and integrated cryptography.<\/li>\n<li><a href=\"http:\/\/www.dd-wrt.com\/\">DD-WRT<\/a> - A is Linux-based firmware for wireless routers and wireless access points. It is compatible with several models of routers and access points.<\/li>\n<\/ul><p><strong>Microsoft introduced a lot of new features in Windows 10 such as Cortana. However, most of them are violating your privacy.<\/strong><\/p><img src=\"https:\/\/www.privacytools.io\/img\/layout\/Windows-10-Privacy.jpg\" class=\"img-fluid float-right\" alt=\"Windows 10 Privacy\" \/><ol><li><strong>Data syncing is by default enabled.<\/strong>\n<ul><li>Browsing history and open websites.<\/li>\n<li>Apps settings.<\/li>\n<li>WiFi hotspot names and passwords.<\/li>\n<\/ul><\/li>\n<li><strong>Your device is by default tagged with a unique advertising ID.<\/strong>\n<ul><li>Used to serve you with personalized advertisements by third-party advertisers and ad networks.<\/li>\n<\/ul><\/li>\n<li><strong>Cortana can collect any of your data.<\/strong>\n<ul><li>Your keystrokes, searches and mic input.<\/li>\n<\/ul><ul><li>Calendar data.<\/li>\n<\/ul><ul><li>Music you listen to.<\/li>\n<\/ul><ul><li>Credit Card information.<\/li>\n<\/ul><ul><li>Purchases.<\/li>\n<\/ul><\/li>\n<li><strong>Microsoft can collect any personal data.<\/strong>\n<ul><li>Your identity.<\/li>\n<\/ul><ul><li>Passwords.<\/li>\n<\/ul><ul><li>Demographics.<\/li>\n<\/ul><ul><li>Interests and habits.<\/li>\n<\/ul><ul><li>Usage data.<\/li>\n<\/ul><ul><li>Contacts and relationships.<\/li>\n<\/ul><ul><li>Location data.<\/li>\n<\/ul><ul><li>Content like emails, instant messages, caller list, audio and video recordings.<\/li>\n<\/ul><\/li>\n<li><strong>Your data can be shared.<\/strong>\n<ul><li>When downloading Windows 10, you are authorizing Microsoft to share any of above-mentioned data with any third-party, with or without your consent.<\/li>\n<\/ul><\/li>\n<\/ol><p>\n<a class=\"btn btn-warning\" href=\"http:\/\/www.winprivacy.de\/english-home\/\">\nDownload: W10Privacy\n<\/a>\n<\/p>This tool uses some known methods that attempt to disable major tracking features in Windows 10.<br \/><div class=\"container\"><div class=\"row\"><div class=\"col-md-6\"><h3>Some good news<\/h3><ul><li><a href=\"https:\/\/github.com\/crazy-max\/WindowsSpyBlocker\/releases\">WindowsSpyBlocker<\/a> - Open source tool that blocks data collection.<\/li>\n<li><a href=\"http:\/\/www.ghacks.net\/2015\/08\/14\/comparison-of-windows-10-privacy-tools\/\">Comparison of Windows 10 Privacy tools<\/a> - ghacks.net<\/li>\n<li><a href=\"https:\/\/fix10.isleaked.com\/\">Fix Windows 10 privacy.<\/a> - fix10.isleaked.com<\/li>\n<li><a href=\"http:\/\/arstechnica.com\/information-technology\/2015\/08\/windows-10-doesnt-offer-much-privacy-by-default-heres-how-to-fix-it\/\">Windows 10 doesn\u2019t offer much privacy by default: Here\u2019s how to fix it. <\/a> - Ars Technica.<\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/Windows10\/comments\/3f38ed\/guide_how_to_disable_data_logging_in_w10\">Guide: How to disable data logging in W10.<\/a><\/li>\n<\/ul><\/div><div class=\"col-md-6\"><h3>More bad news<\/h3><ul><li><a href=\"http:\/\/thehackernews.com\/2016\/02\/microsoft-windows10-privacy.html\">\nWindows 10 Sends Your Data 5500 Times Every Day Even After Tweaking Privacy Settings<\/a> - The Hacker News.<\/li>\n<li><a href=\"http:\/\/arstechnica.com\/information-technology\/2015\/08\/even-when-told-not-to-windows-10-just-cant-stop-talking-to-microsoft\/\">Even when told not to, Windows 10 just can\u2019t stop talking to Microsoft. It's no wonder that privacy activists are up in arms.<\/a> - Ars Technica.<\/li>\n<li><a href=\"https:\/\/www.techdirt.com\/articles\/20150820\/06171332012\/windows-10-reserves-right-to-block-pirated-games-unauthorized-hardware.shtml\">Windows 10 Reserves The Right To Block Pirated Games And 'Unauthorized' Hardware.<\/a> - Techdirt.<\/li>\n<\/ul><\/div><\/div><\/div><blockquote>\n<p>Arguing that you don't care about the right to privacy because you have nothing to hide is no different than saying you don't care about free speech because you have nothing to say.<\/p>\n<footer>Edward Snowden on <cite title=\"Just days left to kill mass surveillance under Section 215 of the Patriot Act. We are Edward Snowden and the ACLU\u2019s Jameel Jaffer. AUA.\"><a href=\"https:\/\/www.reddit.com\/r\/IAmA\/comments\/36ru89\/just_days_left_to_kill_mass_surveillance_under\/crglgh2\">Reddit<\/a><\/cite><\/footer><\/blockquote><blockquote>\n<p>The NSA has built an infrastructure that allows it to intercept almost everything. With this capability, the vast majority of human communications are automatically ingested without targeting. If I wanted to see your emails or your wife's phone, all\nI have to do is use intercepts. I can get your emails, passwords, phone records, credit cards. I don't want to live in a society that does these sort of things... I do not want to live in a world where everything I do and say is recorded. That is\nnot something I am willing to support or live under. <\/p>\n<footer>Edward Snowden in <cite title=\"Edward Snowden, NSA files source: 'If they want to get you, in time they will\"><a href=\"http:\/\/www.theguardian.com\/world\/2013\/jun\/09\/nsa-whistleblower-edward-snowden-why\">The Guardian<\/a><\/cite><\/footer><\/blockquote><blockquote>\n<p>We all need places where we can go to explore without the judgmental eyes of other people being cast upon us, only in a realm where we\u2019re not being watched can we really test the limits of who we want to be. It\u2019s really in the private realm where\ndissent, creativity and personal exploration lie.<\/p>\n<footer>Glenn Greenwald in <cite title=\"Glenn Greenwald On Why Privacy Is Vital, Even If You 'Have Nothing To Hide\"><a href=\"http:\/\/www.huffingtonpost.com\/2014\/06\/20\/glenn-greenwald-privacy_n_5509704.html\">Huffington Post<\/a><\/cite><\/footer><\/blockquote><h3>Guides<\/h3><ul><li><a href=\"https:\/\/ssd.eff.org\/\"><strong>Surveillance Self-Defense by EFF<\/strong><\/a> - Guide to defending yourself from surveillance by using secure technology and developing careful practices.<\/li>\n<li><a href=\"https:\/\/cryptoseb.pw\/paper\"><strong>The Crypto Paper<\/strong><\/a> - Privacy, Security and Anonymity for Every Internet User.<\/li>\n<li><a href=\"https:\/\/emailselfdefense.fsf.org\/en\/\"><strong>Email Self-Defense by FSF<\/strong><\/a> - A guide to fighting surveillance with GnuPG encryption.<\/li>\n<li><a href=\"https:\/\/www.bestvpn.com\/the-ultimate-privacy-guide\/\"><strong>The Ultimate Privacy Guide<\/strong><\/a> - Excellent privacy guide written by the creators of the bestVPN.com website.<\/li>\n<li><a href=\"https:\/\/www.ivpn.net\/privacy-guides\"><strong>IVPN Privacy Guides<\/strong><\/a> - These privacy guides explain how to obtain vastly greater freedom, privacy and anonymity through compartmentalization and isolation.<\/li>\n<li><a href=\"http:\/\/fried.com\/privacy\"><strong>The Ultimate Guide to Online Privacy<\/strong><\/a> - Comprehensive \"Ninja Privacy Tips\" and 150+ tools.<\/li>\n<\/ul><h3>Information<\/h3><ul><li><a href=\"https:\/\/www.reddit.com\/r\/privacytoolsIO\/wiki\/index\"><strong>r\/privacytoolsIO Wiki<\/strong><\/a> - Our Wiki on reddit.com.<\/li>\n<li><a href=\"https:\/\/www.grc.com\/securitynow.htm\"><strong>Security Now!<\/strong><\/a> - Weekly Internet Security Podcast by Steve Gibson and Leo Laporte.<\/li>\n<li><a href=\"http:\/\/www.jupiterbroadcasting.com\/show\/techsnap\/\"><strong>TechSNAP<\/strong><\/a> - Weekly Systems, Network, and Administration Podcast. Every week TechSNAP covers the stories that impact those of us in the tech industry.<\/li>\n<li><a href=\"https:\/\/keybase.io\/\"><strong>Keybase.io<\/strong><\/a> - Get a public key, safely, starting just with someone's social media username.<\/li>\n<li><a href=\"https:\/\/freedom.press\/\"><strong>Freedom of the Press Foundation<\/strong><\/a> - Supporting and defending journalism dedicated to transparency and accountability since 2012.<\/li>\n<li><a href=\"http:\/\/www.erfahrungen.com\/mit\/anonymisierung\/t\/\"><strong>Erfahrungen.com<\/strong><\/a> - German review aggregator website of privacy-related services.<\/li>\n<li><a href=\"https:\/\/tosdr.org\/\"><strong>Terms of Service; Didn't Read<\/strong><\/a> - \u201cI have read and agree to the Terms\u201d is the biggest lie on the web. We aim to fix that.<\/li>\n<\/ul><h3>Tools<\/h3><ul><li><a href=\"http:\/\/ipleak.net\/\"><strong>ipleak.net<\/strong><\/a> - IP\/DNS Detect - What is your IP, what is your DNS, what informations you send to websites.<\/li>\n<li><a href=\"http:\/\/www.ghacks.net\/2015\/12\/28\/the-ultimate-online-privacy-test-resource-list\/\"><strong>The ultimate Online Privacy Test Resource List<\/strong><\/a> - A collection of Internet sites that check whether your web browser leaks information.<\/li>\n<li><a href=\"https:\/\/prism-break.org\/\"><strong>PRISM Break<\/strong><\/a> - We all have a right to privacy, which you can exercise today by encrypting your communications and ending your reliance on proprietary services.<\/li>\n<li><a href=\"https:\/\/securityinabox.org\/\"><strong>Security in-a-Box<\/strong><\/a> - A guide to digital security for activists and human rights defenders throughout the world.<\/li>\n<li><a href=\"http:\/\/alternativeto.net\/?license=opensource&amp;platform=self-hosted&amp;sort=likes\"><strong>AlternativeTo.net<\/strong><\/a> - Great collection of open source online and self-hosted software sorted by likes.<\/li>\n<li><a href=\"https:\/\/securedrop.org\/\"><strong>SecureDrop<\/strong><\/a> - An open-source whistleblower submission system that media organizations can use to securely accept documents from and communicate with anonymous sources. It was originally created\nby the late Aaron Swartz and is currently managed by Freedom of the Press Foundation.<\/li>\n<li><a href=\"https:\/\/pack.resetthenet.org\/\"><strong>Reset The Net - Privacy Pack<\/strong><\/a> - Help fight to end mass surveillance. Get these tools to protect yourself and your friends.<\/li>\n<li><a href=\"http:\/\/www.secfirst.org\/\"><strong>Security First<\/strong><\/a> - Umbrella is an Android app that provides all the advice needed to operate safely in a hostile environment.<\/li>\n<li><a href=\"https:\/\/addons.mozilla.org\/en-US\/firefox\/addon\/block-cloudflare-mitm-attack\/\"><strong>Block Cloudflare MiTM Attack<\/strong><\/a> - Firefox add-on to detect and block corporate MITM attack.\n<\/li><\/ul><h3>Copy URL and Description<\/h3><p>\nFor easy copy and paste. Share this text snippet.\n<\/p><p>\nIt's important for a website like privacytools.io to be up-to-date. Keep an eye on software updates of the applications listed here. Follow recent news about providers that are recommended. We try our best to keep up but we're not perfect and the internet\nis changing fast. If you find an error, or you think a provider should not be listed here, or a qualified service provider is missing or a browser plugin is not the best choice anymore and anything else... <strong>Talk to us please.<\/strong> <\/p><br \/><div class=\"row\"><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-success\"><p><\/p><h3 class=\"h5\">Make suggestions on Reddit<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/icons\/reddit.png\" alt=\"Make suggestions on Reddit\" class=\"panel-item\" \/>Our active community subreddit to stay up to date or to make suggestions. Join now! <a href=\"#\" data-toggle=\"tooltip\" data-placement=\"top\" data-original-title=\"Warrant Canary has been removed by reddit\">(Info)<\/a><\/p><div><p>\n<a class=\"btn btn-success mb-1\" href=\"https:\/\/www.reddit.com\/r\/privacytoolsIO\/\">Website:\nreddit.com\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-primary\"><p><\/p><h3 class=\"h5\">Follow on Twitter<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/icons\/twitter.png\" alt=\"Follow on Twitter\" class=\"panel-item\" \/>Get the latest privacy-related updates from our Twitter Feed. Follow now!<\/p><div><p>\n<a class=\"btn btn-primary mb-1\" href=\"https:\/\/twitter.com\/privacytoolsIO\">Website:\ntwitter.com\n<\/a>\n<\/p><\/div><\/div><\/div><\/div><div class=\"col-xl-4 col-lg-6 col-md-12 mb-2\"><div class=\"card card-warning\"><p><\/p><h3 class=\"h5\">Develop on GitHub<\/h3><div class=\"card-body\"><p><img src=\"https:\/\/www.privacytools.io\/img\/icons\/github.png\" alt=\"Develop on GitHub\" class=\"panel-item\" \/>The complete website source code is available on GitHub. Join our developer team!<\/p><\/div><\/div><\/div><\/div><p>This is a community project and we're aiming to deliver the best information available for a better privacy. Thank you for participating. This project needs you.<\/p><p><\/p><h4>No Ads, No Google Analytics, No Affiliates, No Cross-Site Requests<\/h4><p>\n<a href=\"http:\/\/www.wtfpl.net\/\" data-toggle=\"tooltip\" data-placement=\"top\" data-original-title=\"This work is free. You can redistribute it and\/or modify it under the terms of the ''Do What The Fuck You Want To Public License'', Version 2,&#10;as published by Sam Hocevar.\"><img alt=\"WTFPL\" src=\"https:\/\/www.privacytools.io\/img\/layout\/wtfpl.png\" \/>WTFPL<\/a>\n<a href=\"http:\/\/www.kopimi.com\/\" data-toggle=\"tooltip\" data-placement=\"top\" data-original-title=\"kopimi (copyme), symbol showing that you want to be copied. use kopimi in your own fancy. kopimi may be put on homepages or blogs, in books, in software, as sound logos in music or whatever.\"><img alt=\"kopimi\" src=\"https:\/\/www.privacytools.io\/img\/layout\/kopimi.gif\" \/>kopimi (copyme)<\/a>\n<a href=\"https:\/\/www.privacytools.io\/donate.html\" data-toggle=\"tooltip\" data-placement=\"top\" data-original-title=\"Please support this project by donating Bitcoins. We are ad free and not affiliated with any providers. Your donation will cover our cost for server and domain.\"><img alt=\"Bitcoin\" src=\"https:\/\/www.privacytools.io\/img\/layout\/Bitcoin.png\" \/>Donate:<\/a> 1N4bKCx3kpvUTBhsvbmthPDssN8ATPEhgt or use <a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=5G9CBU8P6F4K2\">PayPal.<\/a>\n<\/p><p><em>privacytools.io is a socially motivated website that provides information for protecting your data security and privacy. never trust any company with your privacy, always encrypt.<\/em><\/p>","created_at":"2018-11-07T16:52:32+0000","updated_at":"2018-11-07T16:52:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":51,"domain_name":"www.privacytools.io","preview_picture":"https:\/\/www.privacytools.io\/img\/layout\/ogimage.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":581,"title":"The Woodnotes Guide to the Mutt Email Client","url":"http:\/\/www.therandymon.com\/woodnotes\/mutt\/using-mutt.html","content":"<a name=\"tex2html43\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node1.html\" id=\"tex2html43\">\n<img width=\"37\" height=\"24\" border=\"0\" alt=\"next\" src=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/next.png\" style=\"text-align: bottom;\" \/><\/a> \n<img width=\"26\" height=\"24\" border=\"0\" alt=\"up\" src=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/up_g.png\" style=\"text-align: bottom;\" \/><img width=\"63\" height=\"24\" border=\"0\" alt=\"previous\" src=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/prev_g.png\" style=\"text-align: bottom;\" \/><a name=\"tex2html41\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node1.html\" id=\"tex2html41\">\n<img width=\"65\" height=\"24\" border=\"0\" alt=\"contents\" src=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/contents.png\" style=\"text-align: bottom;\" \/><\/a>  \n<br \/><b> Next:<\/b> <a name=\"tex2html44\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node1.html\" id=\"tex2html44\">Contents<\/a>\n \u00a0 <b>  <a name=\"tex2html42\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node1.html\" id=\"tex2html42\">Contents<\/a><\/b> \n<hr \/><ul><li><a name=\"tex2html45\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node1.html\" id=\"tex2html45\">Contents<\/a>\n<\/li><li><a name=\"tex2html46\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node2.html\" id=\"tex2html46\">Introduction - Why Mutt?<\/a>\n<\/li><li><a name=\"tex2html47\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node3.html\" id=\"tex2html47\">Usage: The Basics<\/a>\n<ul><li><a name=\"tex2html48\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node4.html\" id=\"tex2html48\">Getting Help<\/a>\n<\/li><li><a name=\"tex2html49\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node5.html\" id=\"tex2html49\">Reading Messages<\/a>\n<ul><li><a name=\"tex2html50\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node6.html\" id=\"tex2html50\">Your Inbox (Index View)<\/a>\n<\/li><li><a name=\"tex2html51\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node7.html\" id=\"tex2html51\">Sorting and Threading Messages in the Index View<\/a>\n<\/li><li><a name=\"tex2html52\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node8.html\" id=\"tex2html52\">Reading Messages (Pager View)<\/a>\n<\/li><li><a name=\"tex2html53\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node9.html\" id=\"tex2html53\">Deleting Messages<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html54\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node10.html\" id=\"tex2html54\">Composing Messages and Saving Drafts (Postponing Messages)<\/a>\n<\/li><li><a name=\"tex2html55\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node11.html\" id=\"tex2html55\">Attachments<\/a>\n<ul><li><a name=\"tex2html56\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node12.html\" id=\"tex2html56\">Forwarding Attachments<\/a>\n<\/li><li><a name=\"tex2html57\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node13.html\" id=\"tex2html57\">Deleting and Undeleting Attachments<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html58\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node14.html\" id=\"tex2html58\">Printing Messages<\/a>\n<\/li><li><a name=\"tex2html59\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node15.html\" id=\"tex2html59\">Reading Mail in Other Folders<\/a>\n<ul><li><a name=\"tex2html60\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node16.html\" id=\"tex2html60\">Special case: IMAP Shared Folders: <\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html61\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node17.html\" id=\"tex2html61\">Saving Messages to Files<\/a>\n<\/li><li><a name=\"tex2html62\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node18.html\" id=\"tex2html62\">Operating on Multiple Messages<\/a>\n<ul><li><a name=\"tex2html63\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node19.html\" id=\"tex2html63\">Selection Patterns<\/a>\n<\/li><li><a name=\"tex2html64\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node20.html\" id=\"tex2html64\">Tagging and Flags<\/a>\n<\/li><li><a name=\"tex2html65\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node21.html\" id=\"tex2html65\">Limiting Messages Shown<\/a>\n<\/li><li><a name=\"tex2html66\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node22.html\" id=\"tex2html66\">Using Patterns to Archive Groups of Messages<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html67\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node23.html\" id=\"tex2html67\">Using the Address Book<\/a>\n<\/li><\/ul><br \/><\/li><li><a name=\"tex2html68\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node24.html\" id=\"tex2html68\">Configuring Mutt<\/a>\n<ul><li><a name=\"tex2html69\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node25.html\" id=\"tex2html69\">Basic Configuration<\/a>\n<\/li><li><a name=\"tex2html70\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node26.html\" id=\"tex2html70\">Setting Up Your Mail Accounts<\/a>\n<ul><li><a name=\"tex2html71\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node27.html\" id=\"tex2html71\">Downloading from a server (POP3)<\/a>\n<ul><li><a name=\"tex2html72\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node28.html\" id=\"tex2html72\">Getting your mail:<\/a>\n<\/li><li><a name=\"tex2html73\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node29.html\" id=\"tex2html73\">Sending your mail:<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html74\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node30.html\" id=\"tex2html74\">Accessing your Mail at the Server (IMAP)<\/a>\n<ul><li><a name=\"tex2html75\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node31.html\" id=\"tex2html75\">Getting your mail:<\/a>\n<\/li><li><a name=\"tex2html76\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node32.html\" id=\"tex2html76\">Header Caching:<\/a>\n<\/li><li><a name=\"tex2html77\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node33.html\" id=\"tex2html77\">Sending your mail:<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html78\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node34.html\" id=\"tex2html78\">Offline IMAP<\/a>\n<ul><li><a name=\"tex2html79\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node35.html\" id=\"tex2html79\">Getting  your mail:<\/a>\n<\/li><li><a name=\"tex2html80\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node36.html\" id=\"tex2html80\">Sending your mail:<\/a>\n<\/li><\/ul><\/li><\/ul><\/li><li><a name=\"tex2html81\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node37.html\" id=\"tex2html81\">Sending Mail: Sendmail and Friends<\/a>\n<ul><li><a name=\"tex2html82\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node38.html\" id=\"tex2html82\">Sendmail:<\/a>\n<\/li><li><a name=\"tex2html83\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node39.html\" id=\"tex2html83\">Mutt's own SMTP:<\/a>\n<\/li><li><a name=\"tex2html84\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node40.html\" id=\"tex2html84\">SSMTP:<\/a>\n<\/li><li><a name=\"tex2html85\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node41.html\" id=\"tex2html85\">Putmail.py:<\/a>\n<\/li><li><a name=\"tex2html86\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node42.html\" id=\"tex2html86\">Others:<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html87\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node43.html\" id=\"tex2html87\">Your Sent Mail box<\/a>\n<\/li><li><a name=\"tex2html88\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node44.html\" id=\"tex2html88\">Configuring Your Address Book<\/a>\n<ul><li><a name=\"tex2html89\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node45.html\" id=\"tex2html89\">Alias File<\/a>\n<\/li><li><a name=\"tex2html90\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node46.html\" id=\"tex2html90\">Abook<\/a>\n<\/li><li><a name=\"tex2html91\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node47.html\" id=\"tex2html91\">Querying an LDAP Server<\/a>\n<\/li><li><a name=\"tex2html92\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node48.html\" id=\"tex2html92\">The Little Brother Database<\/a>\n<\/li><li><a name=\"tex2html93\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node49.html\" id=\"tex2html93\">Rolo Addressbook<\/a>\n<\/li><\/ul><\/li><li><a name=\"tex2html94\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node50.html\" id=\"tex2html94\">Choosing an Editor<\/a>\n<\/li><li><a name=\"tex2html95\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node51.html\" id=\"tex2html95\">General Settings<\/a>\n<\/li><li><a name=\"tex2html96\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node52.html\" id=\"tex2html96\">Dealing with HTML Messages<\/a>\n<\/li><li><a name=\"tex2html97\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node53.html\" id=\"tex2html97\">Dealing with Attachments<\/a>\n<\/li><li><a name=\"tex2html98\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node54.html\" id=\"tex2html98\">Dealing with Word Doc Attachments<\/a>\n<\/li><li><a name=\"tex2html99\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node55.html\" id=\"tex2html99\">Dealing with URLs in Messages<\/a>\n<\/li><\/ul><br \/><\/li><li><a name=\"tex2html100\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node56.html\" id=\"tex2html100\">Advanced Customization<\/a>\n<ul><li><a name=\"tex2html101\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node57.html\" id=\"tex2html101\">Using Mutt with different configurations<\/a>\n<\/li><li><a name=\"tex2html102\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node58.html\" id=\"tex2html102\">Configuring Colors<\/a>\n<\/li><li><a name=\"tex2html103\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node59.html\" id=\"tex2html103\">Rebinding Keyboard Shortcuts<\/a>\n<\/li><li><a name=\"tex2html104\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node60.html\" id=\"tex2html104\">Writing Macros<\/a>\n<\/li><li><a name=\"tex2html105\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node61.html\" id=\"tex2html105\">Changing the Index View<\/a>\n<\/li><li><a name=\"tex2html106\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node62.html\" id=\"tex2html106\">Changing the Pager View<\/a>\n<\/li><li><a name=\"tex2html107\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node63.html\" id=\"tex2html107\">Dealing with non-ASCII Character Sets<\/a>\n<\/li><li><a name=\"tex2html108\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node64.html\" id=\"tex2html108\">Dealing With Particularly Troublesome Characters<\/a>\n<\/li><li><a name=\"tex2html109\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node65.html\" id=\"tex2html109\">Advanced Printing: Mutt-Print<\/a>\n<\/li><li><a name=\"tex2html110\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node66.html\" id=\"tex2html110\">Hooks for Sending, Accounts, and More<\/a>\n<\/li><\/ul><br \/><\/li><li><a name=\"tex2html111\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node67.html\" id=\"tex2html111\">Other Resources and More Information<\/a>\n<ul><li><a name=\"tex2html112\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node68.html\" id=\"tex2html112\">Information and Tips<\/a>\n<\/li><li><a name=\"tex2html113\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node69.html\" id=\"tex2html113\">Other Software with which Mutt Interfaces<\/a>\n<\/li><\/ul><br \/><\/li><li><a name=\"tex2html114\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node70.html\" id=\"tex2html114\">Appendix: Patterns<\/a>\n<\/li><li><a name=\"tex2html115\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node71.html\" id=\"tex2html115\">Appendix: Functions available for Macros<\/a>\n<ul><li><a name=\"tex2html116\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node72.html\" id=\"tex2html116\">Generic Functions<\/a>\n<\/li><li><a name=\"tex2html117\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node73.html\" id=\"tex2html117\">The Index<\/a>\n<\/li><li><a name=\"tex2html118\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node74.html\" id=\"tex2html118\">The Pager<\/a>\n<\/li><li><a name=\"tex2html119\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node75.html\" id=\"tex2html119\">Aliases<\/a>\n<\/li><li><a name=\"tex2html120\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node76.html\" id=\"tex2html120\">Queries<\/a>\n<\/li><li><a name=\"tex2html121\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node77.html\" id=\"tex2html121\">The Attach menu<\/a>\n<\/li><li><a name=\"tex2html122\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node78.html\" id=\"tex2html122\">The Compose menu<\/a>\n<\/li><li><a name=\"tex2html123\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node79.html\" id=\"tex2html123\">The Postpone menu<\/a>\n<\/li><li><a name=\"tex2html124\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node80.html\" id=\"tex2html124\">Browser<\/a>\n<\/li><li><a name=\"tex2html125\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node81.html\" id=\"tex2html125\">PGP<\/a>\n<\/li><li><a name=\"tex2html126\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node82.html\" id=\"tex2html126\">The Editor<\/a>\n<\/li><\/ul><br \/><\/li><li><a name=\"tex2html127\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node83.html\" id=\"tex2html127\">Appendix: Acknowledgments and License<\/a>\n<\/li><li><a name=\"tex2html128\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node84.html\" id=\"tex2html128\">Appendix: Version History<\/a>\n<\/li><li><a name=\"tex2html129\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/node85.html\" id=\"tex2html129\">About this document ...<\/a>\n<\/li><\/ul><br \/><hr \/><address>\nRandall Wood\n2009-12-02\n<\/address>","created_at":"2018-11-07T16:52:31+0000","updated_at":"2018-11-07T16:52:31+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.therandymon.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":582,"title":"A complete guide to packaging your Electron app","url":"https:\/\/medium.com\/how-to-electron\/a-complete-guide-to-packaging-your-electron-app-1bdc717d739f","content":"<div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"9f61\" class=\"graf graf--h3 graf--leading graf--title\">A complete guide to packaging your Electron\u00a0app<\/h1><p id=\"7bec\" class=\"graf graf--p graf-after--h3\">Packaging an electron app simply means creating a desktop installer (dmg, exe, deb etc). Now if you decide to go around manually packaging your app, you\u2019re gonna have a bad time.<\/p><figure id=\"e299\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*IRLTRUofVIZE6YJiCYhCLg.png\" data-width=\"597\" data-height=\"402\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*IRLTRUofVIZE6YJiCYhCLg.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"194e\" class=\"graf graf--p graf-after--figure\">Luckily there are modules, especially the two mentioned below which make the task easier.<\/p><blockquote id=\"eb70\" class=\"graf graf--blockquote graf-after--p\"><a href=\"https:\/\/github.com\/electron-userland\/electron-builder\" data-href=\"https:\/\/github.com\/electron-userland\/electron-builder\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">electron-builder<\/a><\/blockquote><blockquote id=\"5330\" class=\"graf graf--blockquote graf-after--blockquote\"><a href=\"https:\/\/github.com\/electron-userland\/electron-packager\" data-href=\"https:\/\/github.com\/electron-userland\/electron-packager\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"nofollow noopener\" target=\"_blank\">electron-packager<\/a><\/blockquote><p id=\"0e6e\" class=\"graf graf--p graf-after--blockquote\">We\u2019ll be using electron-builder since it has a built-in support for <a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/wiki\/Code-Signing\" data-href=\"https:\/\/github.com\/electron-userland\/electron-builder\/wiki\/Code-Signing\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Code Signing<\/a>\/<a href=\"https:\/\/github.com\/electron-userland\/electron-builder#auto-update\" data-href=\"https:\/\/github.com\/electron-userland\/electron-builder#auto-update\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Auto Update<\/a> etc.<\/p><h3 id=\"8652\" class=\"graf graf--h3 graf-after--p\">Configure your app to use electron-builder\u00a0:<\/h3><ul class=\"postList\"><li id=\"2ba1\" class=\"graf graf--li graf-after--h3\">Create a directory <a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/wiki\/Options#MetadataDirectories-buildResources\" data-href=\"https:\/\/github.com\/electron-userland\/electron-builder\/wiki\/Options#MetadataDirectories-buildResources\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">build<\/a> in the root of the project and save a <code class=\"markup--code markup--li-code\">background.png<\/code> (macOS DMG background), <code class=\"markup--code markup--li-code\">icon.icns<\/code> (macOS app icon) and <code class=\"markup--code markup--li-code\">icon.ico<\/code> (Windows app icon) into it. The Linux icon set will be generated automatically based on the macOS. Ex. <a href=\"https:\/\/github.com\/zulip\/zulip-electron\/tree\/master\/build\" data-href=\"https:\/\/github.com\/zulip\/zulip-electron\/tree\/master\/build\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">zulip-electron\/build<\/a><\/li><li id=\"0184\" class=\"graf graf--li graf-after--li\">Add electron-builder to your app devDependencies by running:<\/li><\/ul><pre id=\"fe81\" class=\"graf graf--pre graf-after--li\">npm install electron-builder --save-dev<\/pre><ul class=\"postList\"><li id=\"5728\" class=\"graf graf--li graf-after--pre\">In electron-builder you could use two package.json structure as well as single package.json. We\u2019ll be using single package.json structure since it\u2019s easier to maintain.<\/li><\/ul><h4 id=\"354c\" class=\"graf graf--h4 graf-after--li\">Sample One package.json<\/h4><figure id=\"6cce\" class=\"graf graf--figure graf--iframe graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe width=\"700\" height=\"250\" src=\"https:\/\/medium.com\/media\/2bba7919e9d30103c13d888a760fc349?postId=1bdc717d739f\" data-media-id=\"2bba7919e9d30103c13d888a760fc349\" data-thumbnail=\"https:\/\/i.embed.ly\/1\/image?url=https%3A%2F%2Favatars3.githubusercontent.com%2Fu%2F2263909%3Fv%3D3%26s%3D400&amp;key=4fce0568f2ce49e8b54624ef71a8a5bd\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><\/div><\/figure><p id=\"585a\" class=\"graf graf--p graf-after--figure\">On windows build you can also use <em class=\"markup--em markup--p-em\">NSIS <\/em>target which is the default one and recommended by electron-builder.<\/p><pre id=\"f93b\" class=\"graf graf--pre graf-after--p\">\"win\": {<br \/>   \"target\": \"NSIS\"<br \/>}<\/pre><p id=\"1bfc\" class=\"graf graf--p graf-after--pre\">That\u2019s it. You\u2019ve successfully configured electron-builder. Now let\u2019s build the installers. To package your app into an installer use command:<\/p><pre id=\"2806\" class=\"graf graf--pre graf-after--p\">npm run dist<\/pre><p id=\"bd93\" class=\"graf graf--p graf-after--pre\">It will create an installer in\u00a0.<strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">\/dist<\/em><\/strong> folder. By default build command will only generate the installer for current platform and current arch. For an ex. running it on OSX will create \u2014<\/p><pre id=\"6be5\" class=\"graf graf--pre graf-after--p\">$ npm run dist<br \/>    |-- mac<br \/>    |   |-- appname-version-mac.zip<br \/>    |   |-- appname-version.dmg<br \/>    |   |-- appname.app<\/pre><p id=\"5a81\" class=\"graf graf--p graf-after--pre\">On Windows -<\/p><pre id=\"b4e4\" class=\"graf graf--pre graf-after--p\"># When using target \u2014 Squirrel<\/pre><pre id=\"1c75\" class=\"graf graf--pre graf-after--pre\">$ npm run dist<br \/>    |-- win<br \/>    |   |-- REALESES<br \/>    |   |-- appnameSetupversion.exe<br \/>    |   |-- appname-version-full.nupkg<br \/>    |-- win-unpacked<\/pre><pre id=\"8e7a\" class=\"graf graf--pre graf-after--pre\"># When using target \u2014 NSIS<\/pre><pre id=\"6060\" class=\"graf graf--pre graf-after--pre\">$ npm run dist<br \/>    |-- win<br \/>    |   |-- appnameSetupversion.exe<br \/>    |-- win-unpacked<\/pre><p id=\"a02c\" class=\"graf graf--p graf-after--pre\">On Linux -<\/p><pre id=\"610a\" class=\"graf graf--pre graf-after--p\">$ npm run dist<br \/>    |-- appname-version.AppImage<br \/>    |-- appname-version.deb<br \/>    |-- linux-unpacked<\/pre><p id=\"30fa\" class=\"graf graf--p graf-after--pre\">electron-builder is highly configurable. For complete usage check out its <a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/wiki\" data-href=\"https:\/\/github.com\/electron-userland\/electron-builder\/wiki\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">wiki<\/a>. <a href=\"https:\/\/github.com\/zulip\/zulip-electron\" data-href=\"https:\/\/github.com\/zulip\/zulip-electron\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">zulip-electron<\/a> (two package.json) and <a href=\"https:\/\/github.com\/develar\/onshape-desktop-shell\" data-href=\"https:\/\/github.com\/develar\/onshape-desktop-shell\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">onshape-desktop-shell<\/a> (one package.json) are some real world projects that use electron-builder for their packaging.<\/p><p id=\"a2bf\" class=\"graf graf--p graf-after--p graf--trailing\">(<strong class=\"markup--strong markup--p-strong\">Note <\/strong>- This post was originally posted here\u200a\u2014\u200a<a href=\"https:\/\/howtoelectron.app\/a-complete-guide-to-packaging-your-electron-app\/\" data-href=\"https:\/\/howtoelectron.app\/a-complete-guide-to-packaging-your-electron-app\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/howtoelectron.app<\/a>\/)<\/p><\/div><\/div>","created_at":"2018-11-07T16:52:31+0000","updated_at":"2018-11-07T16:52:31+0000","published_at":"2016-12-30T17:24:10+0000","published_by":["Akash Nimare"],"annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*IRLTRUofVIZE6YJiCYhCLg.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":580,"title":"Learn Blockchains by Building One","url":"https:\/\/hackernoon.com\/learn-blockchains-by-building-one-117428612f46?gi=585e3f632ee0","content":"<section class=\"section section--body section--first\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"75dd\" class=\"graf graf--figure graf--layoutFillWidth graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*zutLn_-fZZhy7Ari-x-JWQ.jpeg\" data-width=\"1800\" data-height=\"931\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*zutLn_-fZZhy7Ari-x-JWQ.jpeg\" alt=\"image\" \/><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"e816\" class=\"graf graf--p graf-after--figure\">You\u2019re here because, like me, you\u2019re psyched about the rise of Cryptocurrencies. And you want to know how Blockchains work\u2014the fundamental technology behind them.<\/p><p id=\"bd65\" class=\"graf graf--p graf-after--p\">But understanding Blockchains isn\u2019t easy\u2014or at least wasn\u2019t for me. I trudged through dense videos, followed porous tutorials, and dealt with the amplified frustration of too few examples.<\/p><p id=\"4741\" class=\"graf graf--p graf-after--p\">I like learning by doing. It forces me to deal with the subject matter at a code level, which gets it sticking. If you do the same, at the end of this guide you\u2019ll have a functioning Blockchain with a solid grasp of how they work.<\/p><h3 id=\"6db5\" class=\"graf graf--h3 graf-after--p\">Before you get\u00a0started\u2026<\/h3><p id=\"2c71\" class=\"graf graf--p graf-after--h3\">Remember that a blockchain is an <em class=\"markup--em markup--p-em\">immutable, sequential <\/em>chain of records called Blocks. They can contain transactions, files or any data you like, really. But the important thing is that they\u2019re <em class=\"markup--em markup--p-em\">chained<\/em> together using <em class=\"markup--em markup--p-em\">hashes<\/em>.<\/p><p id=\"d788\" class=\"graf graf--p graf-after--p\">If you aren\u2019t sure what a hash is, <a href=\"https:\/\/learncryptography.com\/hash-functions\/what-are-hash-functions\" data-href=\"https:\/\/learncryptography.com\/hash-functions\/what-are-hash-functions\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">here\u2019s an explanation<\/a>.<\/p><p id=\"3369\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Who is this guide aimed at?<\/em><\/strong> You should be comfy reading and writing some basic Python, as well as have some understanding of how HTTP requests work, since we\u2019ll be talking to our Blockchain over HTTP.<\/p><p id=\"5da1\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">What do I need?<\/em> <\/strong>Make sure that <a href=\"https:\/\/www.python.org\/downloads\/\" data-href=\"https:\/\/www.python.org\/downloads\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Python 3.6<\/a>+ (along with <code class=\"markup--code markup--p-code\">pip<\/code>) is installed. You\u2019ll also need to install Flask and the wonderful Requests library:<\/p><pre class=\"markup--code markup--pre-code\">pip install Flask==0.12.2 requests==2.18.4 <\/pre><p id=\"b587\" class=\"graf graf--p graf-after--pre\">Oh, you\u2019ll also need an HTTP Client, like <a href=\"https:\/\/www.getpostman.com\" data-href=\"https:\/\/www.getpostman.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Postman<\/a> or cURL. But anything will do.<\/p><p id=\"8308\" class=\"graf graf--p graf-after--p graf--trailing\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Where\u2019s the final code? <\/em><\/strong>The source code is <a href=\"https:\/\/github.com\/dvf\/blockchain\" data-href=\"https:\/\/github.com\/dvf\/blockchain\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">available here<\/a>.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"6994\" class=\"graf graf--h3 graf--leading\">Step 1: Building a Blockchain<\/h3><p id=\"0863\" class=\"graf graf--p graf-after--h3\">Open up your favourite text editor or IDE, personally I \u2764\ufe0f <a href=\"https:\/\/www.jetbrains.com\/pycharm\/\" data-href=\"https:\/\/www.jetbrains.com\/pycharm\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">PyCharm<\/a>. Create a new file, called <code class=\"markup--code markup--p-code\">blockchain.py<\/code>. We\u2019ll only use a single file, but if you get lost, you can always refer to the <a href=\"https:\/\/github.com\/dvf\/blockchain\" data-href=\"https:\/\/github.com\/dvf\/blockchain\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">source code<\/a>.<\/p><h4 id=\"6c47\" class=\"graf graf--h4 graf-after--p\">Representing a Blockchain<\/h4><p id=\"48c8\" class=\"graf graf--p graf-after--h4\">We\u2019ll create a <code class=\"markup--code markup--p-code\">Blockchain<\/code> class whose constructor creates an initial empty list (to store our blockchain), and another to store transactions. Here\u2019s the blueprint for our class:<\/p><figure id=\"bb54\" class=\"graf graf--figure graf--iframe graf-after--p\"><figcaption class=\"imageCaption\">Blueprint of our Blockchain Class<\/figcaption><\/figure><p id=\"eb0a\" class=\"graf graf--p graf-after--figure\">Our <code class=\"markup--code markup--p-code\">Blockchain<\/code> class is responsible for managing the chain. It will store transactions and have some helper methods for adding new blocks to the chain. Let\u2019s start fleshing out some methods.<\/p><h4 id=\"1a89\" class=\"graf graf--h4 graf-after--p\">What does a Block look\u00a0like?<\/h4><p id=\"e489\" class=\"graf graf--p graf-after--h4\">Each Block has an <em class=\"markup--em markup--p-em\">index<\/em>, a <em class=\"markup--em markup--p-em\">timestamp<\/em> (in Unix time), a <em class=\"markup--em markup--p-em\">list of transactions<\/em>, a <em class=\"markup--em markup--p-em\">proof<\/em> (more on that later), and the <em class=\"markup--em markup--p-em\">hash of the previous Block<\/em>.<\/p><p id=\"73a8\" class=\"graf graf--p graf-after--p\">Here\u2019s an example of what a single Block looks like:<\/p><figure id=\"f122\" class=\"graf graf--figure graf--iframe graf-after--p\"><figcaption class=\"imageCaption\">Example of a Block in our Blockchain<\/figcaption><\/figure><p id=\"414f\" class=\"graf graf--p graf-after--figure\">At this point, the idea of a <em class=\"markup--em markup--p-em\">chain<\/em> should be apparent\u2014each new block contains within itself, the hash of the previous Block. <strong class=\"markup--strong markup--p-strong\">This is crucial because it\u2019s what gives blockchains immutability:<\/strong> If an attacker corrupted an earlier Block in the chain then <strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">all<\/em><\/strong> subsequent blocks will contain incorrect hashes.<\/p><p id=\"809e\" class=\"graf graf--p graf-after--p\"><em class=\"markup--em markup--p-em\">Does this make sense? If it doesn\u2019t, take some time to let it sink in\u2014it\u2019s the core idea behind blockchains.<\/em><\/p><h4 id=\"04e0\" class=\"graf graf--h4 graf-after--p\">Adding Transactions to a\u00a0Block<\/h4><p id=\"a276\" class=\"graf graf--p graf-after--h4\">We\u2019ll need a way of adding transactions to a Block. Our <code class=\"markup--code markup--p-code\">new_transaction()<\/code> method is responsible for this, and it\u2019s pretty straight-forward:<\/p><figure id=\"a7a8\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"5a22\" class=\"graf graf--p graf-after--figure\">After <code class=\"markup--code markup--p-code\">new_transaction()<\/code> adds a transaction to the list, it returns the <em class=\"markup--em markup--p-em\">index<\/em> of the block which the transaction will be added to\u2014<em class=\"markup--em markup--p-em\">the next one to be mined.<\/em> This will be useful later on, to the user submitting the transaction.<\/p><h4 id=\"0db8\" class=\"graf graf--h4 graf-after--p\">Creating new\u00a0Blocks<\/h4><p id=\"de4b\" class=\"graf graf--p graf-after--h4\">When our <code class=\"markup--code markup--p-code\">Blockchain<\/code> is instantiated we\u2019ll need to seed it with a <em class=\"markup--em markup--p-em\">genesis<\/em> block\u2014a block with no predecessors. We\u2019ll also need to add a <em class=\"markup--em markup--p-em\">\u201cproof\u201d<\/em> to our genesis block which is the result of mining (or proof of work). We\u2019ll talk more about mining later.<\/p><p id=\"8d25\" class=\"graf graf--p graf-after--p\">In addition to creating the <em class=\"markup--em markup--p-em\">genesis<\/em> block in our constructor, we\u2019ll also flesh out the methods for <code class=\"markup--code markup--p-code\">new_block()<\/code>, <code class=\"markup--code markup--p-code\">new_transaction()<\/code> and <code class=\"markup--code markup--p-code\">hash()<\/code>:<\/p><figure id=\"4dc7\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"61d8\" class=\"graf graf--p graf-after--figure\">The above should be straight-forward\u2014I\u2019ve added some comments and <em class=\"markup--em markup--p-em\">docstrings<\/em> to help keep it clear. We\u2019re almost done with representing our blockchain. But at this point, you must be wondering how new blocks are created, forged or mined.<\/p><h4 id=\"4133\" class=\"graf graf--h4 graf-after--p\">Understanding Proof of\u00a0Work<\/h4><p id=\"657e\" class=\"graf graf--p graf-after--h4\">A Proof of Work algorithm (PoW) is how new Blocks are created or <em class=\"markup--em markup--p-em\">mined <\/em>on the blockchain<em class=\"markup--em markup--p-em\">. <\/em>The goal of PoW is to discover a number which solves a problem. The number must be <strong class=\"markup--strong markup--p-strong\">difficult to find<\/strong> <strong class=\"markup--strong markup--p-strong\">but easy to verify<\/strong>\u2014computationally speaking\u2014by anyone on the network. This is the core idea behind Proof of Work.<\/p><p id=\"633b\" class=\"graf graf--p graf-after--p\">We\u2019ll look at a very simple example to help this sink in.<\/p><p id=\"835c\" class=\"graf graf--p graf-after--p\">Let\u2019s decide that the <em class=\"markup--em markup--p-em\">hash<\/em> of some integer <code class=\"markup--code markup--p-code\">x<\/code> multiplied by another <code class=\"markup--code markup--p-code\">y<\/code> must end in <code class=\"markup--code markup--p-code\">0<\/code>. So, <code class=\"markup--code markup--p-code\">hash(x * y) = ac23dc...0<\/code>. And for this simplified example, let\u2019s fix <code class=\"markup--code markup--p-code\">x = 5<\/code>. Implementing this in Python:<\/p><pre id=\"026a\" class=\"graf graf--pre graf-after--p\">from hashlib import sha256<\/pre><pre id=\"faa1\" class=\"graf graf--pre graf-after--pre\">x = 5<br \/>y = 0  # We don't know what y should be yet...<\/pre><pre id=\"55e4\" class=\"graf graf--pre graf-after--pre\">while sha256(f'{x*y}'.encode()).hexdigest()[-1] != \"0\":<br \/>y += 1<\/pre><pre id=\"7453\" class=\"graf graf--pre graf-after--pre\">print(f'The solution is y = {y}')<\/pre><p id=\"bb3f\" class=\"graf graf--p graf-after--pre\">The solution here is <code class=\"markup--code markup--p-code\">y = 21<\/code>. Since, the produced hash ends in <code class=\"markup--code markup--p-code\">0<\/code>:<\/p><pre id=\"02f8\" class=\"graf graf--pre graf-after--p\">hash(5 * 21) = 1253e9373e...5e3600155e860<\/pre><p id=\"45f6\" class=\"graf graf--p graf-after--pre\">In Bitcoin, the Proof of Work algorithm is called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hashcash\" data-href=\"https:\/\/en.wikipedia.org\/wiki\/Hashcash\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Hashcash<\/em><\/a>. And it\u2019s not too different from our basic example above. It\u2019s the algorithm that miners race to solve in order to create a new block. In general, the difficulty is determined by the number of characters searched for in a string. The miners are then rewarded for their solution by receiving a coin\u2014in a transaction.<\/p><p id=\"4ac4\" class=\"graf graf--p graf-after--p\">The network is able to <em class=\"markup--em markup--p-em\">easily <\/em>verify their solution.<\/p><h4 id=\"0cd0\" class=\"graf graf--h4 graf-after--p\">Implementing basic Proof of\u00a0Work<\/h4><p id=\"5e95\" class=\"graf graf--p graf-after--h4\">Let\u2019s implement a similar algorithm for our blockchain. Our rule will be similar to the example above:<\/p><blockquote id=\"5dfc\" class=\"graf graf--blockquote graf-after--p\"><div><em class=\"markup--em markup--blockquote-em\">Find a number <\/em>p<em class=\"markup--em markup--blockquote-em\"> that when hashed with the previous block\u2019s solution a hash with 4 leading <\/em><code class=\"markup--code markup--blockquote-code\"><em class=\"markup--em markup--blockquote-em\">0<\/em><\/code><em class=\"markup--em markup--blockquote-em\">s is produced.<\/em><\/div><\/blockquote><figure id=\"9357\" class=\"graf graf--figure graf--iframe graf-after--blockquote\"><p id=\"33a6\" class=\"graf graf--p graf-after--figure\">To adjust the difficulty of the algorithm, we could modify the number of leading zeroes. But 4 is sufficient. You\u2019ll find out that the addition of a single leading zero makes a mammoth difference to the time required to find a solution.<\/p><p id=\"fa1c\" class=\"graf graf--p graf-after--p graf--trailing\">Our class is almost complete and we\u2019re ready to begin interacting with it using HTTP requests.<\/p><\/figure><\/figure><\/figure><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"40c2\" class=\"graf graf--h3 graf--leading\">Step 2: Our Blockchain as an\u00a0API<\/h3><p id=\"b021\" class=\"graf graf--p graf-after--h3\">We\u2019re going to use the Python Flask Framework. It\u2019s a micro-framework and it makes it easy to map endpoints to Python functions. This allows us talk to our blockchain over the web using HTTP requests.<\/p><p id=\"55fc\" class=\"graf graf--p graf-after--p\">We\u2019ll create three methods:<\/p><ul class=\"postList\"><li id=\"a678\" class=\"graf graf--li graf-after--p\"><code class=\"markup--code markup--li-code\">\/transactions\/new<\/code> to create a new transaction to a block<\/li><li id=\"39e2\" class=\"graf graf--li graf-after--li\"><code class=\"markup--code markup--li-code\">\/mine<\/code> to tell our server to mine a new block.<\/li><li id=\"748e\" class=\"graf graf--li graf-after--li\"><code class=\"markup--code markup--li-code\">\/chain<\/code> to return the full Blockchain.<\/li><\/ul><h4 id=\"fe8e\" class=\"graf graf--h4 graf-after--li\">Setting up\u00a0Flask<\/h4><p id=\"6068\" class=\"graf graf--p graf-after--h4\">Our \u201cserver\u201d will form a single node in our blockchain network. Let\u2019s create some boilerplate code:<\/p><figure id=\"f02e\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"4b5f\" class=\"graf graf--p graf-after--figure\">A brief explanation of what we\u2019ve added above:<\/p><ul class=\"postList\"><li id=\"196c\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Line 15:<\/strong> Instantiates our Node. Read more about Flask <a href=\"http:\/\/flask.pocoo.org\/docs\/0.12\/quickstart\/#a-minimal-application\" data-href=\"http:\/\/flask.pocoo.org\/docs\/0.12\/quickstart\/#a-minimal-application\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">here<\/a>.<\/li><li id=\"d40b\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Line 18:<\/strong> Create a random name for our node.<\/li><li id=\"91a0\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Line 21:<\/strong> Instantiate our <code class=\"markup--code markup--li-code\">Blockchain<\/code> class.<\/li><li id=\"d357\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Line 24\u201326:<\/strong> Create the <code class=\"markup--code markup--li-code\">\/mine<\/code> endpoint, which is a <code class=\"markup--code markup--li-code\">GET<\/code> request.<\/li><li id=\"83c2\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Line 28\u201330:<\/strong> Create the <code class=\"markup--code markup--li-code\">\/transactions\/new<\/code> endpoint, which is a <code class=\"markup--code markup--li-code\">POST<\/code> request, since we\u2019ll be sending data to it.<\/li><li id=\"3244\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Line 32\u201338:<\/strong> Create the <code class=\"markup--code markup--li-code\">\/chain<\/code> endpoint, which returns the full Blockchain.<\/li><li id=\"0276\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Line 40\u201341:<\/strong> Runs the server on port 5000.<\/li><\/ul><h4 id=\"feb2\" class=\"graf graf--h4 graf-after--li\">The Transactions Endpoint<\/h4><p id=\"a52a\" class=\"graf graf--p graf-after--h4\">This is what the request for a transaction will look like. It\u2019s what the user sends to the server:<\/p><pre id=\"0707\" class=\"graf graf--pre graf-after--p\">{<br \/>\"sender\": \"my address\",<br \/>\"recipient\": \"someone else's address\",<br \/>\"amount\": 5<br \/>}<\/pre><p id=\"2909\" class=\"graf graf--p graf-after--pre\">Since we already have our class method for adding transactions to a block, the rest is easy. Let\u2019s write the function for adding transactions:<\/p><figure id=\"4d4e\" class=\"graf graf--figure graf--iframe graf-after--p\"><figcaption class=\"imageCaption\">A method for creating Transactions<\/figcaption><\/figure><h4 id=\"570a\" class=\"graf graf--h4 graf-after--figure\">The Mining\u00a0Endpoint<\/h4><p id=\"3c93\" class=\"graf graf--p graf-after--h4\">Our mining endpoint is where the magic happens, and it\u2019s easy. It has to do three things:<\/p><ol class=\"postList\"><li id=\"bf40\" class=\"graf graf--li graf-after--p\">Calculate the Proof of Work<\/li><li id=\"aa86\" class=\"graf graf--li graf-after--li\">Reward the miner (us) by adding a transaction granting us 1 coin<\/li><li id=\"aed5\" class=\"graf graf--li graf-after--li\">Forge the new Block by adding it to the chain<\/li><\/ol><figure id=\"f042\" class=\"graf graf--figure graf--iframe graf-after--li\"><p id=\"6082\" class=\"graf graf--p graf-after--figure\">Note that the recipient of the mined block is the address of our node. And most of what we\u2019ve done here is just interact with the methods on our Blockchain class. At this point, we\u2019re done, and can start interacting with our blockchain.<\/p><h3 id=\"7101\" class=\"graf graf--h3 graf-after--p\">Step 3: Interacting with our Blockchain<\/h3><p id=\"7139\" class=\"graf graf--p graf-after--h3\">You can use plain old cURL or Postman to interact with our API over a network.<\/p><p id=\"81ca\" class=\"graf graf--p graf-after--p\">Fire up the server:<\/p><pre id=\"3085\" class=\"graf graf--pre graf-after--p\">$ python blockchain.py<\/pre><pre id=\"bbc3\" class=\"graf graf--pre graf-after--pre\">* Running on <a href=\"http:\/\/127.0.0.1:5000\/\" data-href=\"http:\/\/127.0.0.1:5000\/\" class=\"markup--anchor markup--pre-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http:\/\/127.0.0.1:5000\/<\/a> (Press CTRL+C to quit)<\/pre><p id=\"e87a\" class=\"graf graf--p graf-after--pre\">Let\u2019s try mining a block by making a <code class=\"markup--code markup--p-code\">GET<\/code> request to <code class=\"markup--code markup--p-code\">http:\/\/localhost:5000\/mine<\/code>:<\/p><figure id=\"ebf1\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*ufYwRmWgQeA-Jxg0zgYLOA.png\" data-width=\"2140\" data-height=\"1534\" data-action=\"zoom\" data-action-value=\"1*ufYwRmWgQeA-Jxg0zgYLOA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ufYwRmWgQeA-Jxg0zgYLOA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Using Postman to make a GET\u00a0request<\/figcaption><\/figure><p id=\"9ed9\" class=\"graf graf--p graf-after--figure\">Let\u2019s create a new transaction by making a <code class=\"markup--code markup--p-code\">POST<\/code> request to<code class=\"markup--code markup--p-code\">http:\/\/localhost:5000\/transactions\/new<\/code> with a body containing our transaction structure:<\/p><figure id=\"05da\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*O89KNbEWj1vigMZ6VelHAg.png\" data-width=\"2140\" data-height=\"1446\" data-action=\"zoom\" data-action-value=\"1*O89KNbEWj1vigMZ6VelHAg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*O89KNbEWj1vigMZ6VelHAg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Using Postman to make a POST\u00a0request<\/figcaption><\/figure><p id=\"85eb\" class=\"graf graf--p graf-after--figure\">If you aren\u2019t using Postman, then you can make the equivalent request using cURL:<\/p><pre id=\"fbfb\" class=\"graf graf--pre graf-after--p\">$ curl -X POST -H \"Content-Type: application\/json\" -d '{<br \/>\"sender\": \"d4ee26eee15148ee92c6cd394edd974e\",<br \/>\"recipient\": \"someone-other-address\",<br \/>\"amount\": 5<br \/>}' \"<a href=\"http:\/\/localhost:5000\/transactions\/new\" data-href=\"http:\/\/localhost:5000\/transactions\/new\" class=\"markup--anchor markup--pre-anchor\" rel=\"nofollow\" target=\"_blank\">http:\/\/localhost:5000\/transactions\/new<\/a>\"<\/pre><p id=\"311a\" class=\"graf graf--p graf-after--pre\">I restarted my server, and mined two blocks, to give 3 in total. Let\u2019s inspect the full chain by requesting <code class=\"markup--code markup--p-code\"><a href=\"http:\/\/localhost:5000\/chain:\" data-href=\"http:\/\/localhost:5000\/chain:\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">http:\/\/localhost:5000\/chain<\/a><\/code><a href=\"http:\/\/localhost:5000\/chain:\" data-href=\"http:\/\/localhost:5000\/chain:\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">:<\/a><\/p><pre id=\"e8d9\" class=\"graf graf--pre graf-after--p\">{<br \/>\"chain\": [<br \/>{<br \/>\"index\": 1,<br \/>\"previous_hash\": 1,<br \/>\"proof\": 100,<br \/>\"timestamp\": 1506280650.770839,<br \/>\"transactions\": []<br \/>},<br \/>{<br \/>\"index\": 2,<br \/>\"previous_hash\": \"c099bc...bfb7\",<br \/>\"proof\": 35293,<br \/>\"timestamp\": 1506280664.717925,<br \/>\"transactions\": [<br \/>{<br \/>\"amount\": 1,<br \/>\"recipient\": \"8bbcb347e0634905b0cac7955bae152b\",<br \/>\"sender\": \"0\"<br \/>}<br \/>]<br \/>},<br \/>{<br \/>\"index\": 3,<br \/>\"previous_hash\": \"eff91a...10f2\",<br \/>\"proof\": 35089,<br \/>\"timestamp\": 1506280666.1086972,<br \/>\"transactions\": [<br \/>{<br \/>\"amount\": 1,<br \/>\"recipient\": \"8bbcb347e0634905b0cac7955bae152b\",<br \/>\"sender\": \"0\"<br \/>}<br \/>]<br \/>}<br \/>],<br \/>\"length\": 3<br \/>}<\/pre><h3 id=\"849f\" class=\"graf graf--h3 graf-after--pre\">Step 4: Consensus<\/h3><p id=\"c7d9\" class=\"graf graf--p graf-after--h3\">This is very cool. We\u2019ve got a basic Blockchain that accepts transactions and allows us to mine new Blocks. But the whole point of Blockchains is that they should be <em class=\"markup--em markup--p-em\">decentralized<\/em>. And if they\u2019re decentralized, how on earth do we ensure that they all reflect the same chain? This is called the problem of <em class=\"markup--em markup--p-em\">Consensus<\/em>, and we\u2019ll have to implement a Consensus Algorithm if we want more than one node in our network.<\/p><h4 id=\"8e32\" class=\"graf graf--h4 graf-after--p\">Registering new\u00a0Nodes<\/h4><p id=\"461d\" class=\"graf graf--p graf-after--h4\">Before we can implement a Consensus Algorithm, we need a way to let a node know about neighbouring nodes on the network. Each node on our network should keep a registry of other nodes on the network. Thus, we\u2019ll need some more endpoints:<\/p><ol class=\"postList\"><li id=\"5b1c\" class=\"graf graf--li graf-after--p\"><code class=\"markup--code markup--li-code\">\/nodes\/register<\/code> to accept a list of new nodes in the form of URLs.<\/li><li id=\"7e9f\" class=\"graf graf--li graf-after--li\"><code class=\"markup--code markup--li-code\">\/nodes\/resolve<\/code> to implement our Consensus Algorithm, which resolves any conflicts\u2014to ensure a node has the correct chain.<\/li><\/ol><p id=\"d653\" class=\"graf graf--p graf-after--li\">We\u2019ll need to modify our Blockchain\u2019s constructor and provide a method for registering nodes:<\/p><figure id=\"08fb\" class=\"graf graf--figure graf--iframe graf-after--p\"><figcaption class=\"imageCaption\">A method for adding neighbouring nodes to our Network<\/figcaption><\/figure><p id=\"e01e\" class=\"graf graf--p graf-after--figure\">Note that we\u2019ve used a <code class=\"markup--code markup--p-code\">set()<\/code> to hold the list of nodes. This is a cheap way of ensuring that the addition of new nodes is idempotent\u2014meaning that no matter how many times we add a specific node, it appears exactly once.<\/p><h4 id=\"2491\" class=\"graf graf--h4 graf-after--p\">Implementing the Consensus Algorithm<\/h4><p id=\"1921\" class=\"graf graf--p graf-after--h4\">As mentioned, a conflict is when one node has a different chain to another node. To resolve this, we\u2019ll make the rule that <em class=\"markup--em markup--p-em\">the longest valid chain is authoritative.<\/em> In other words, the longest chain on the network is the <em class=\"markup--em markup--p-em\">de-facto<\/em> one. Using this algorithm, we reach <em class=\"markup--em markup--p-em\">Consensus<\/em> amongst the nodes in our network.<\/p><figure id=\"636b\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"4779\" class=\"graf graf--p graf-after--figure\">The first method <code class=\"markup--code markup--p-code\">valid_chain()<\/code> is responsible for checking if a chain is valid by looping through each block and verifying both the hash and the proof.<\/p><p id=\"a99b\" class=\"graf graf--p graf-after--p\"><code class=\"markup--code markup--p-code\">resolve_conflicts()<\/code> is a method which loops through all our neighbouring nodes, <em class=\"markup--em markup--p-em\">downloads<\/em> their chains and verifies them using the above method. <strong class=\"markup--strong markup--p-strong\">If a valid chain is found, whose length is greater than ours, we replace ours.<\/strong><\/p><p id=\"cd9a\" class=\"graf graf--p graf-after--p\">Let\u2019s register the two endpoints to our API, one for adding neighbouring nodes and the another for resolving conflicts:<\/p><figure id=\"9323\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"f01b\" class=\"graf graf--p graf-after--figure\">At this point you can grab a different machine if you like, and spin up different nodes on your network. Or spin up processes using different ports on the same machine. I spun up another node on my machine, on a different port, and registered it with my current node. Thus, I have two nodes: <code class=\"markup--code markup--p-code\"><a href=\"http:\/\/localhost:5000\" data-href=\"http:\/\/localhost:5000\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">http:\/\/localhost:5000<\/a><\/code> and <code class=\"markup--code markup--p-code\">http:\/\/localhost:5001<\/code>.<\/p><figure id=\"6b41\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*Dd78u-gmtwhQWHhPG3qMTQ.png\" data-width=\"2784\" data-height=\"1582\" data-action=\"zoom\" data-action-value=\"1*Dd78u-gmtwhQWHhPG3qMTQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Dd78u-gmtwhQWHhPG3qMTQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Registering a new\u00a0Node<\/figcaption><\/figure><p id=\"75a7\" class=\"graf graf--p graf-after--figure\">I then mined some new Blocks on node 2, to ensure the chain was longer. Afterward, I called <code class=\"markup--code markup--p-code\">GET \/nodes\/resolve<\/code> on node 1, where the chain was replaced by the Consensus Algorithm:<\/p><figure id=\"fe84\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*SGO5MWVf7GguIxfz6S8NVw.png\" data-width=\"2784\" data-height=\"1770\" data-action=\"zoom\" data-action-value=\"1*SGO5MWVf7GguIxfz6S8NVw.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*SGO5MWVf7GguIxfz6S8NVw.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Consensus Algorithm at\u00a0Work<\/figcaption><\/figure><p id=\"88ac\" class=\"graf graf--p graf-after--figure graf--trailing\">And that\u2019s a wrap... Go get some friends together to help test out your Blockchain.<\/p><\/figure><\/figure><\/figure><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"5262\" class=\"graf graf--p graf--leading\">I hope that this has inspired you to create something new. I\u2019m ecstatic about Cryptocurrencies because I believe that Blockchains will rapidly change the way we think about economies, governments and record-keeping.<\/p><p id=\"6163\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Update:<\/strong> I\u2019m planning on following up with a Part 2, where we\u2019ll extend our Blockchain to have a Transaction Validation Mechanism as well as discuss some ways in which you can productionize your Blockchain.<\/p><blockquote id=\"6b32\" class=\"graf graf--pullquote graf-after--p\"><div><em class=\"markup--em markup--pullquote-em\">If you enjoyed this guide, or have any suggestions or questions, let me know in the comments. And if you\u2019ve spotted any errors, feel free to contribute to the code\u00a0<\/em><a href=\"https:\/\/github.com\/dvf\/blockchain\" data-href=\"https:\/\/github.com\/dvf\/blockchain\" class=\"markup--anchor markup--pullquote-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--pullquote-em\">here<\/em><\/a><em class=\"markup--em markup--pullquote-em\">!<\/em><\/div><\/blockquote><figure id=\"d896\" class=\"graf graf--figure graf-after--pullquote graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/bit.ly\/HNshares\" data-href=\"http:\/\/bit.ly\/HNshares\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*5A8U4eQ5xqGBsqA51r4wBw@2x.png\" data-width=\"1400\" data-height=\"750\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*5A8U4eQ5xqGBsqA51r4wBw@2x.png\" alt=\"image\" \/><\/a><\/div><\/figure><\/div><\/div><\/section>","created_at":"2018-11-07T16:52:30+0000","updated_at":"2018-11-07T16:52:30+0000","published_at":"2017-09-24T20:06:20+0000","published_by":["Daniel van Flymen"],"annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"hackernoon.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*zutLn_-fZZhy7Ari-x-JWQ.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":579,"title":"JavaScript Factory Functions with ES6+","url":"https:\/\/medium.com\/javascript-scene\/javascript-factory-functions-with-es6-4d224591a8b1","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"b98b\" class=\"graf graf--h3 graf--leading graf--title\">JavaScript Factory Functions with\u00a0ES6+<\/h1><figure id=\"8064\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*uVpU7iruzXafhU2VLeH4lw.jpeg\" data-width=\"5151\" data-height=\"2346\" data-is-featured=\"true\" data-action=\"zoom\" data-action-value=\"1*uVpU7iruzXafhU2VLeH4lw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*uVpU7iruzXafhU2VLeH4lw.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Smoke Art Cubes to Smoke\u200a\u2014\u200aMattysFlicks\u200a\u2014\u200a(CC BY\u00a02.0)<\/figcaption><\/div><\/figure><blockquote id=\"69e6\" class=\"graf graf--blockquote graf-after--figure\"><div>Note: This is part of the \u201cComposing Software\u201d series on learning functional programming and compositional software techniques in JavaScript ES6+ from the ground up. Stay tuned. There\u2019s a lot more of this to come!<br \/><a href=\"https:\/\/medium.com\/javascript-scene\/functional-mixins-composing-software-ffb66d5e731c\" data-href=\"https:\/\/medium.com\/javascript-scene\/functional-mixins-composing-software-ffb66d5e731c\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">&lt; Previous<\/a> | <a href=\"https:\/\/medium.com\/javascript-scene\/composing-software-an-introduction-27b72500d6ea\" data-href=\"https:\/\/medium.com\/javascript-scene\/composing-software-an-introduction-27b72500d6ea\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">&lt;&lt; Start over at Part 1<\/a> | <a href=\"https:\/\/medium.com\/javascript-scene\/why-composition-is-harder-with-classes-c3e627dcd0aa\" data-href=\"https:\/\/medium.com\/javascript-scene\/why-composition-is-harder-with-classes-c3e627dcd0aa\" class=\"markup--anchor markup--blockquote-anchor\" target=\"_blank\">Next &gt;<\/a><\/div><\/blockquote><p id=\"35dd\" class=\"graf graf--p graf-after--blockquote\">A <strong class=\"markup--strong markup--p-strong\">factory function<\/strong> is any function which is not a class or constructor that returns a (presumably new) object. In JavaScript, any function can return an object. When it does so without the <code class=\"markup--code markup--p-code\">new<\/code> keyword, it\u2019s a factory function.<\/p><p id=\"e8f6\" class=\"graf graf--p graf-after--p\">Factory functions have always been attractive in JavaScript because they offer the ability to easily produce object instances without diving into the complexities of classes and the <code class=\"markup--code markup--p-code\">new<\/code> keyword.<\/p><p id=\"75d7\" class=\"graf graf--p graf-after--p\">JavaScript provides a very handy object literal syntax. It looks something like this:<\/p><pre id=\"1c87\" class=\"graf graf--pre graf-after--p\">const user = {<br \/>  userName: 'echo',<br \/>  avatar: 'echo.png'<br \/>};<\/pre><p id=\"9fa9\" class=\"graf graf--p graf-after--pre\">Like JSON (which is based on JavaScript\u2019s object literal notation), the left side of the\u00a0<code class=\"markup--code markup--p-code\">:<\/code> is the property name, and the right side is the value. You can access props with dot notation:<\/p><pre id=\"94a6\" class=\"graf graf--pre graf-after--p\">console.log(user.userName); \/\/ \"echo\"<\/pre><p id=\"e8da\" class=\"graf graf--p graf-after--pre\">You can access computed property names using square bracket notation:<\/p><pre id=\"37c2\" class=\"graf graf--pre graf-after--p\">const key = 'avatar';<\/pre><pre id=\"3081\" class=\"graf graf--pre graf-after--pre\">console.log( user[key] ); \/\/ \"echo.png\"<\/pre><p id=\"07ed\" class=\"graf graf--p graf-after--pre\">If you have variables in-scope with the same name as your intended property names, you can omit the colon and the value in the object literal creation:<\/p><pre id=\"df5b\" class=\"graf graf--pre graf-after--p\">const userName = 'echo';<br \/>const avatar = 'echo.png';<\/pre><pre id=\"ebb0\" class=\"graf graf--pre graf-after--pre\">const user = {<br \/>  userName,<br \/>  avatar<br \/>};<\/pre><pre id=\"a4a6\" class=\"graf graf--pre graf-after--pre\">console.log(user);<br \/>\/\/ { \"avatar\": \"echo.png\",   \"userName\": \"echo\" }<\/pre><p id=\"f723\" class=\"graf graf--p graf-after--pre\">Object literals support concise method syntax. We can add a\u00a0<code class=\"markup--code markup--p-code\">.setUserName()<\/code> method:<\/p><pre id=\"d73d\" class=\"graf graf--pre graf-after--p\">const userName = 'echo';<br \/>const avatar = 'echo.png';<\/pre><pre id=\"3786\" class=\"graf graf--pre graf-after--pre\">const user = {<br \/>  userName,<br \/>  avatar,<\/pre><pre id=\"7eba\" class=\"graf graf--pre graf-after--pre\">  setUserName (userName) {<br \/>    this.userName = userName;<br \/>    return this;<br \/>  }<br \/>};<\/pre><pre id=\"56b5\" class=\"graf graf--pre graf-after--pre\">console.log(user.setUserName('Foo').userName); \/\/ \"Foo\"<\/pre><p id=\"3629\" class=\"graf graf--p graf-after--pre\">In concise methods, <code class=\"markup--code markup--p-code\">this<\/code> refers to the object which the method is called on. To call a method on an object, simply access the method using object dot notation and invoke it by using parentheses, e.g., <code class=\"markup--code markup--p-code\">game.play()<\/code> would apply\u00a0<code class=\"markup--code markup--p-code\">.play()<\/code> to the <code class=\"markup--code markup--p-code\">game<\/code> object. In order to apply a method using dot notation, that method must be a property of the object in question. You can also apply a method to any other arbitrary object using the function prototype methods,\u00a0<code class=\"markup--code markup--p-code\">.call()<\/code>,\u00a0<code class=\"markup--code markup--p-code\">.apply()<\/code>, or\u00a0<code class=\"markup--code markup--p-code\">.bind()<\/code>.<\/p><p id=\"bb6f\" class=\"graf graf--p graf-after--p\">In this case, <code class=\"markup--code markup--p-code\">user.setUserName('Foo')<\/code> applies\u00a0<code class=\"markup--code markup--p-code\">.setUserName()<\/code> to <code class=\"markup--code markup--p-code\">user<\/code>, so <code class=\"markup--code markup--p-code\">this === user<\/code>. In the\u00a0<code class=\"markup--code markup--p-code\">.setUserName()<\/code> method, we change the\u00a0<code class=\"markup--code markup--p-code\">.userName<\/code> property on the <code class=\"markup--code markup--p-code\">user<\/code> object via its <code class=\"markup--code markup--p-code\">this<\/code> binding, and return the same object instance for method chaining.<\/p><h3 id=\"3dd3\" class=\"graf graf--h3 graf-after--p\">Literals for One, Factories for\u00a0Many<\/h3><p id=\"41d3\" class=\"graf graf--p graf-after--h3\">If you need to create many objects, you\u2019ll want to combine the power of object literals and factory functions.<\/p><p id=\"ca0f\" class=\"graf graf--p graf-after--p\">With a factory function, you can create as many user objects as you want. If you\u2019re building a chat app, for instance, you can have a user object representing the current user, and also a lot of other user objects representing all the other users who are currently signed in and chatting, so you can display their names and avatars, too.<\/p><p id=\"5585\" class=\"graf graf--p graf-after--p\">Let\u2019s turn our <code class=\"markup--code markup--p-code\">user<\/code> object into a <code class=\"markup--code markup--p-code\">createUser()<\/code> factory:<\/p><pre id=\"6b7a\" class=\"graf graf--pre graf-after--p\">const createUser = ({ userName, avatar }) =&gt; ({<br \/>  userName,<br \/>  avatar,<br \/>  setUserName (userName) {<br \/>    this.userName = userName;<br \/>    return this;<br \/>  }<br \/>});<\/pre><pre id=\"f607\" class=\"graf graf--pre graf-after--pre\">console.log(createUser({ userName: 'echo', avatar: 'echo.png' }));<\/pre><pre id=\"174f\" class=\"graf graf--pre graf-after--pre\">\/*<br \/>{<br \/>  \"avatar\": \"echo.png\",<br \/>  \"userName\": \"echo\",<br \/>  \"setUserName\": [Function setUserName]<br \/>}<br \/>*\/<\/pre><h3 id=\"1060\" class=\"graf graf--h3 graf-after--pre\">Returning Objects<\/h3><p id=\"6404\" class=\"graf graf--p graf-after--h3\">Arrow functions (<code class=\"markup--code markup--p-code\">=&gt;<\/code>) have an implicit return feature: if the function body consists of a single expression, you can omit the <code class=\"markup--code markup--p-code\">return<\/code> keyword: <code class=\"markup--code markup--p-code\">() =&gt; 'foo'<\/code> is a function that takes no parameters, and returns the string, <code class=\"markup--code markup--p-code\">\"foo\"<\/code>.<\/p><p id=\"0e64\" class=\"graf graf--p graf-after--p\">Be careful when you return object literals. By default, JavaScript assumes you want to create a function body when you use braces, e.g., <code class=\"markup--code markup--p-code\">{ broken: true }<\/code>. If you want to use an implicit return for an object literal, you\u2019ll need to disambiguate by wrapping the object literal in parentheses:<\/p><pre id=\"bcaf\" class=\"graf graf--pre graf-after--p\">const noop = () =&gt; { foo: 'bar' };<br \/>console.log(noop()); \/\/ undefined<\/pre><pre id=\"7719\" class=\"graf graf--pre graf-after--pre\">const createFoo = () =&gt; ({ foo: 'bar' });<br \/>console.log(createFoo()); \/\/ { foo: \"bar\" }<\/pre><p id=\"74be\" class=\"graf graf--p graf-after--pre\">In the first example, <code class=\"markup--code markup--p-code\">foo:<\/code> is interpreted as a label, and <code class=\"markup--code markup--p-code\">bar<\/code> is interpreted as an expression that doesn\u2019t get assigned or returned. The function returns <code class=\"markup--code markup--p-code\">undefined<\/code>.<\/p><p id=\"131c\" class=\"graf graf--p graf-after--p\">In the <code class=\"markup--code markup--p-code\">createFoo()<\/code> example, the parentheses force the braces to be interpreted as an expression to be evaluated, rather than a function body block.<\/p><h3 id=\"f246\" class=\"graf graf--h3 graf-after--p\">Destructuring<\/h3><p id=\"0a43\" class=\"graf graf--p graf-after--h3\">Pay special attention to the function signature:<\/p><pre id=\"eb85\" class=\"graf graf--pre graf-after--p\">const createUser = ({ userName, avatar }) =&gt; ({<\/pre><p id=\"418d\" class=\"graf graf--p graf-after--pre\">In this line, the braces (<code class=\"markup--code markup--p-code\">{<\/code>, <code class=\"markup--code markup--p-code\">}<\/code>) represent object destructuring. This function takes one argument (an object), but destructures two formal parameters from that single argument, <code class=\"markup--code markup--p-code\">userName<\/code>, and <code class=\"markup--code markup--p-code\">avatar<\/code>. Those parameters can then be used as variables in the function body scope. You can also destructure arrays:<\/p><pre id=\"e49c\" class=\"graf graf--pre graf-after--p\">const swap = ([first, second]) =&gt; [second, first];<\/pre><pre id=\"c52f\" class=\"graf graf--pre graf-after--pre\">console.log( swap([1, 2]) ); \/\/ [2, 1]<\/pre><p id=\"a3fd\" class=\"graf graf--p graf-after--pre\">And you can use the rest and spread syntax (<code class=\"markup--code markup--p-code\">...varName<\/code>) to gather the rest of the values from the array (or a list of arguments), and then spread those array elements back into individual elements:<\/p><pre id=\"d02a\" class=\"graf graf--pre graf-after--p\">const rotate = ([first, ...rest]) =&gt; [...rest, first];<\/pre><pre id=\"e4d2\" class=\"graf graf--pre graf-after--pre\">console.log( rotate([1, 2, 3]) ); \/\/ [2, 3, 1]<\/pre><h3 id=\"1da6\" class=\"graf graf--h3 graf-after--pre\">Computed Property\u00a0Keys<\/h3><p id=\"ffc1\" class=\"graf graf--p graf-after--h3\">Earlier we used square bracket computed property access notation to dynamically determine which object property to access:<\/p><pre id=\"46ad\" class=\"graf graf--pre graf-after--p\">const key = 'avatar';<\/pre><pre id=\"52b0\" class=\"graf graf--pre graf-after--pre\">console.log( user[key] ); \/\/ \"echo.png\"<\/pre><p id=\"fc9d\" class=\"graf graf--p graf-after--pre\">We can also compute the value of keys to assign to:<\/p><pre id=\"fdea\" class=\"graf graf--pre graf-after--p\">const arrToObj = ([key, value]) =&gt; ({ [key]: value });<\/pre><pre id=\"98df\" class=\"graf graf--pre graf-after--pre\">console.log( arrToObj([ 'foo', 'bar' ]) ); \/\/ { \"foo\": \"bar\" }<\/pre><p id=\"0676\" class=\"graf graf--p graf-after--pre\">In this example, <code class=\"markup--code markup--p-code\">arrToObj<\/code> takes an array consisting of a key\/value pair (aka a tuple) and converts it into an object. Since we don\u2019t know the name of the key, we need to compute the property name in order to set the key\/value pair on the object. For that, we borrow the idea of square bracket notation from computed property accessors, and reuse it in the context of building an object literal:<\/p><pre id=\"3a40\" class=\"graf graf--pre graf-after--p\">{ [key]: value }<\/pre><p id=\"e785\" class=\"graf graf--p graf-after--pre\">After the interpolation is done, we end up with the final object:<\/p><pre id=\"0ed6\" class=\"graf graf--pre graf-after--p\">{ \"foo\": \"bar\" }<\/pre><h3 id=\"7799\" class=\"graf graf--h3 graf-after--pre\">Default Parameters<\/h3><p id=\"543e\" class=\"graf graf--p graf-after--h3\">Functions in JavaScript support default parameter values, which have several benefits:<\/p><ol class=\"postList\"><li id=\"7596\" class=\"graf graf--li graf-after--p\">Users are able to omit parameters with suitable defaults.<\/li><li id=\"6937\" class=\"graf graf--li graf-after--li\">The function is more self-documenting because default values supply examples of expected input.<\/li><li id=\"8615\" class=\"graf graf--li graf-after--li\">IDEs and static analysis tools can use default values to infer the type expected for the parameter. For example, a default value of <code class=\"markup--code markup--li-code\">1<\/code> implies that the parameter can take a member of the <code class=\"markup--code markup--li-code\">Number<\/code> type.<\/li><\/ol><p id=\"53cc\" class=\"graf graf--p graf-after--li\">Using default parameters, we can document the expected interface for our <code class=\"markup--code markup--p-code\">createUser<\/code> factory, and automatically fill in <code class=\"markup--code markup--p-code\">'Anonymous'<\/code> details if the user\u2019s info is not supplied:<\/p><pre id=\"8364\" class=\"graf graf--pre graf-after--p\">const createUser = ({<br \/>  userName = 'Anonymous',<br \/>  avatar = 'anon.png'<br \/>} = {}) =&gt; ({<br \/>  userName,<br \/>  avatar<br \/>});<\/pre><pre id=\"5313\" class=\"graf graf--pre graf-after--pre\">console.log(<br \/>  \/\/ { userName: \"echo\", avatar: 'anon.png' }<br \/>  createUser({ userName: 'echo' }),<\/pre><pre id=\"0a16\" class=\"graf graf--pre graf-after--pre\">  \/\/ { userName: \"Anonymous\", avatar: 'anon.png' }<br \/>  createUser()<br \/>);<\/pre><p id=\"1644\" class=\"graf graf--p graf-after--pre\">The last part of the function signature probably looks a little funny:<\/p><pre id=\"a0e2\" class=\"graf graf--pre graf-after--p\">} = {}) =&gt; ({<\/pre><p id=\"9f21\" class=\"graf graf--p graf-after--pre\">The last <code class=\"markup--code markup--p-code\">= {}<\/code> bit just before the parameter signature closes means that if nothing gets passed in for this parameter, we\u2019re going to use an empty object as the default if no argument is passed in. When you try to destructure values from the empty object, the default values for properties will get used automatically, because that\u2019s what default values do: replace <code class=\"markup--code markup--p-code\">undefined<\/code> with some predefined value.<\/p><p id=\"00ee\" class=\"graf graf--p graf-after--p\">Without the <code class=\"markup--code markup--p-code\">= {}<\/code> default value, <code class=\"markup--code markup--p-code\">createUser()<\/code> with no arguments would throw an error because you can\u2019t try to access properties from <code class=\"markup--code markup--p-code\">undefined<\/code>.<\/p><h4 id=\"5152\" class=\"graf graf--h4 graf-after--p\">Type Inference<\/h4><p id=\"ee19\" class=\"graf graf--p graf-after--h4\">JavaScript does not have any native type annotations as of this writing, but several competing formats have sprung up over the years to fill the gaps, including JSDoc (in decline due to the emergence of better options), <a href=\"https:\/\/flow.org\/\" data-href=\"https:\/\/flow.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Facebook\u2019s Flow<\/a>, and <a href=\"https:\/\/www.typescriptlang.org\/\" data-href=\"https:\/\/www.typescriptlang.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Microsoft\u2019s TypeScript<\/a>. I use <a href=\"https:\/\/github.com\/ericelliott\/rtype\" data-href=\"https:\/\/github.com\/ericelliott\/rtype\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">rtype<\/a> for documentation\u200a\u2014\u200aa notation I find much more readable than TypeScript for functional programming.<\/p><p id=\"157c\" class=\"graf graf--p graf-after--p\">At the time of this writing, there is no clear winner for type annotations. None of the alternatives have been blessed by the JavaScript specification, and there seem to be clear shortcomings in all of them.<\/p><p class=\"graf graf--p graf-after--p\">Type inference is the process of inferring types based on the context in which they are used. In JavaScript, it is a very good alternative to type annotations.<\/p><p id=\"380c\" class=\"graf graf--p graf-after--p\">If you provide enough clues for inference in your standard JavaScript function signatures, you\u2019ll get most of the benefits of type annotations with none of the costs or risks.<\/p><p id=\"025c\" class=\"graf graf--p graf-after--p\">Even if you decide to use a tool like TypeScript or Flow, you should do as much as you can with type inference, and save the type annotations for situations where type inference falls short. For example, there\u2019s no native way in JavaScript to specify a shared interface. That\u2019s both easy and useful with TypeScript or rtype.<\/p><p id=\"1763\" class=\"graf graf--p graf-after--p\"><a href=\"http:\/\/ternjs.net\/\" data-href=\"http:\/\/ternjs.net\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Tern.js<\/a> is a popular type inference tool for JavaScript that has plugins for many code editors and IDEs.<\/p><p id=\"9677\" class=\"graf graf--p graf-after--p\">Microsoft\u2019s Visual Studio Code doesn\u2019t need Tern because it brings the type inference capabilities of TypeScript to regular JavaScript code.<\/p><p id=\"1c16\" class=\"graf graf--p graf-after--p\">When you specify default parameters for functions in JavaScript, tools capable of type inference such as Tern.js, TypeScript, and Flow can provide IDE hints to help you use the API you\u2019re working with correctly.<\/p><p id=\"a2ea\" class=\"graf graf--p graf-after--p\">Without defaults, IDEs (and frequently, humans) don\u2019t have enough hints to figure out the expected parameter type.<\/p><figure id=\"a880\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*2sP_9k1e0dkgYqdEPs0G3g.png\" data-width=\"1426\" data-height=\"902\" data-action=\"zoom\" data-action-value=\"1*2sP_9k1e0dkgYqdEPs0G3g.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*2sP_9k1e0dkgYqdEPs0G3g.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Without defaults, the type is unknown for `userName`.<\/figcaption><\/div><\/figure><p id=\"f5a8\" class=\"graf graf--p graf-after--figure\">With defaults, IDEs (and frequently, humans) can infer the types from the examples.<\/p><figure id=\"4e74\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*tFUXCLA8ClAzsPgZXGGR9A.png\" data-width=\"1414\" data-height=\"896\" data-action=\"zoom\" data-action-value=\"1*tFUXCLA8ClAzsPgZXGGR9A.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*tFUXCLA8ClAzsPgZXGGR9A.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">With defaults, the IDE can suggest that `userName` is expecting a\u00a0string.<\/figcaption><\/div><\/figure><p id=\"49c9\" class=\"graf graf--p graf-after--figure\">It doesn\u2019t always make sense to restrict a parameter to a fixed type (that would make generic functions and higher order functions difficult), but when it does make sense, default parameters are often the best way to do it, even if you\u2019re using TypeScript or Flow.<\/p><h3 id=\"fb43\" class=\"graf graf--h3 graf-after--p\">Factory Functions for Mixin Composition<\/h3><p id=\"87eb\" class=\"graf graf--p graf-after--h3\">Factories are great at cranking out objects using a nice calling API. Usually, they\u2019re all you need, but once in a while, you\u2019ll find yourself building similar features into different types of objects, and you\u2019ll want to abstract those features into functional mixins so you can reuse them more easily.<\/p><p id=\"c5e1\" class=\"graf graf--p graf-after--p\">That\u2019s where functional mixins shine. Let\u2019s build a <code class=\"markup--code markup--p-code\">withConstructor<\/code> mixin to add the\u00a0<code class=\"markup--code markup--p-code\">.constructor<\/code> property to all object instances.<\/p><p id=\"fab0\" class=\"graf graf--p graf-after--p\"><code class=\"markup--code markup--p-code\">with-constructor.js<\/code>:<\/p><pre id=\"cb2f\" class=\"graf graf--pre graf-after--p\">const withConstructor = constructor =&gt; o =&gt; ({<br \/>  \/\/ create the delegate [[Prototype]]<br \/>  __proto__: {<br \/>    \/\/ add the constructor prop to the new [[Prototype]]<br \/>    constructor<br \/>  },<br \/>  \/\/ mix all o's props into the new object<br \/>  ...o<br \/>});<\/pre><p id=\"ba72\" class=\"graf graf--p graf-after--pre\">Now you can import it and use it with other mixins:<\/p><pre id=\"ca97\" class=\"graf graf--pre graf-after--p\">import withConstructor from `.\/with-constructor';<\/pre><pre id=\"b7c3\" class=\"graf graf--pre graf-after--pre\">const pipe = (...fns) =&gt; x =&gt; fns.reduce((y, f) =&gt; f(y), x);<br \/>\/\/ or `import pipe from 'lodash\/fp\/flow';`<\/pre><pre id=\"a899\" class=\"graf graf--pre graf-after--pre\">\/\/ Set up some functional mixins<br \/>const withFlying = o =&gt; {<br \/>  let isFlying = false;<\/pre><pre id=\"e15a\" class=\"graf graf--pre graf-after--pre\">  return {<br \/>    ...o,<br \/>    fly () {<br \/>      isFlying = true;<br \/>      return this;<br \/>    },<br \/>    land () {<br \/>      isFlying = false;<br \/>      return this;<br \/>    },<br \/>    isFlying: () =&gt; isFlying<br \/>  }<br \/>};<\/pre><pre id=\"8cb9\" class=\"graf graf--pre graf-after--pre\">const withBattery = ({ capacity }) =&gt; o =&gt; {<br \/>  let percentCharged = 100;<\/pre><pre id=\"c185\" class=\"graf graf--pre graf-after--pre\">  return {<br \/>    ...o,<br \/>    draw (percent) {<br \/>      const remaining = percentCharged - percent;<br \/>      percentCharged = remaining &gt; 0 ? remaining : 0;<br \/>      return this;<br \/>    },<br \/>    getCharge: () =&gt; percentCharged,<br \/>    getCapacity () {<br \/>      return capacity;<br \/>    }<br \/>  };<br \/>};<\/pre><pre id=\"b4b5\" class=\"graf graf--pre graf-after--pre\">const createDrone = ({ capacity = '3000mAh' }) =&gt; pipe(<br \/>  withFlying,<br \/>  withBattery({ capacity }),<br \/>  withConstructor(createDrone)<br \/>)({});<\/pre><pre id=\"9601\" class=\"graf graf--pre graf-after--pre\">const myDrone = createDrone({ capacity: '5500mAh' });<\/pre><pre id=\"67cb\" class=\"graf graf--pre graf-after--pre\">console.log(`<br \/>  can fly:  ${ myDrone.fly().isFlying() === true }<br \/>  can land: ${ myDrone.land().isFlying() === false }<br \/>  battery capacity: ${ myDrone.getCapacity() }<br \/>  battery status: ${ myDrone.draw(50).getCharge() }%<br \/>  battery drained: ${ myDrone.draw(75).getCharge() }%<br \/>`);<\/pre><pre id=\"94a5\" class=\"graf graf--pre graf-after--pre\">console.log(`<br \/>  constructor linked: ${ myDrone.constructor === createDrone }<br \/>`);<\/pre><p id=\"bed4\" class=\"graf graf--p graf-after--pre\">As you can see, the reusable <code class=\"markup--code markup--p-code\">withConstructor()<\/code> mixin is simply dropped into the pipeline with other mixins. <code class=\"markup--code markup--p-code\">withBattery()<\/code> could be used with other kinds of objects, like robots, electric skateboards, or portable device chargers. <code class=\"markup--code markup--p-code\">withFlying()<\/code> could be used to model flying cars, rockets, or air balloons.<\/p><p id=\"3bc9\" class=\"graf graf--p graf-after--p\">Composition is more of a way of thinking than a particular technique in code. You can accomplish it in many ways. Function composition is just the easiest way to build it up from scratch, and factory functions are a simple way to wrap a friendly API around the implementation details.<\/p><h3 id=\"1290\" class=\"graf graf--h3 graf-after--p\">Conclusion<\/h3><p id=\"7a87\" class=\"graf graf--p graf-after--h3\">ES6 provides a convenient syntax for dealing with object creation and factory functions. Most of the time, that\u2019s all you\u2019ll need, but because this is JavaScript, there\u2019s another approach that makes it feel more like Java: the <code class=\"markup--code markup--p-code\">class<\/code> keyword.<\/p><p id=\"ffad\" class=\"graf graf--p graf-after--p\">In JavaScript, classes are more verbose &amp; restrictive than factories, and a bit of a minefield when it comes to refactoring, but they\u2019ve also been embraced by major front-end frameworks like React and Angular, and there are a couple of rare use-cases that make the complexity worthwhile.<\/p><blockquote id=\"ebef\" class=\"graf graf--blockquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cSometimes, the elegant implementation is just a function. Not a method. Not a class. Not a framework. Just a function.\u201d ~ John Carmack<\/div><\/blockquote><p id=\"383a\" class=\"graf graf--p graf-after--blockquote\">Start with the simplest implementation, and move to more complex implementations only as required. When it comes to objects, that progression looks a bit like this:<\/p><p id=\"a9b6\" class=\"graf graf--p graf-after--p\"><code class=\"markup--code markup--p-code\">Pure function -&gt; factory -&gt; functional mixin -&gt; class<\/code><\/p><p id=\"57bb\" class=\"graf graf--p graf-after--p\"><a href=\"https:\/\/medium.com\/javascript-scene\/why-composition-is-harder-with-classes-c3e627dcd0aa\" data-href=\"https:\/\/medium.com\/javascript-scene\/why-composition-is-harder-with-classes-c3e627dcd0aa\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Next: Why Composition is Harder with Classes &gt;<\/em><\/strong><\/a><\/p><h3 id=\"8818\" class=\"graf graf--h3 graf-after--p\">Next Steps<\/h3><p id=\"e44a\" class=\"graf graf--p graf-after--h3\">Want to learn more about object composition with JavaScript?<\/p><p id=\"6a7e\" class=\"graf graf--p graf-after--p\"><a href=\"http:\/\/ericelliottjs.com\/product\/lifetime-access-pass\/\" data-href=\"http:\/\/ericelliottjs.com\/product\/lifetime-access-pass\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener nofollow noopener nofollow noopener nofollow noopener\" target=\"_blank\">Learn JavaScript with Eric Elliott<\/a>. If you\u2019re not a member, you\u2019re missing out!<\/p><figure id=\"fb45\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><a href=\"https:\/\/ericelliottjs.com\/product\/lifetime-access-pass\/\" data-href=\"https:\/\/ericelliottjs.com\/product\/lifetime-access-pass\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\" rel=\"nofollow\"><img class=\"graf-image\" data-image-id=\"1*3njisYUeHOdyLCGZ8czt_w.jpeg\" data-width=\"1024\" data-height=\"329\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*3njisYUeHOdyLCGZ8czt_w.jpeg\" alt=\"image\" \/><\/a><\/div><\/div><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"5e7e\" class=\"graf graf--p graf--leading\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Eric Elliott<\/em><\/strong><em class=\"markup--em markup--p-em\"> is the author of <\/em><a href=\"http:\/\/pjabook.com\" data-href=\"http:\/\/pjabook.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">\u201cProgramming JavaScript Applications\u201d<\/em><\/a><em class=\"markup--em markup--p-em\"> (O\u2019Reilly), and <\/em><a href=\"http:\/\/ericelliottjs.com\/product\/lifetime-access-pass\/\" data-href=\"http:\/\/ericelliottjs.com\/product\/lifetime-access-pass\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener nofollow noopener noopener nofollow noopener nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">\u201cLearn JavaScript with Eric Elliott\u201d<\/em><\/a><em class=\"markup--em markup--p-em\">. He has contributed to software experiences for <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Adobe Systems<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Zumba Fitness<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">The Wall Street Journal<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">ESPN<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">BBC<\/em><\/strong><em class=\"markup--em markup--p-em\">, and top recording artists including <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Usher<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Frank Ocean<\/em><\/strong><em class=\"markup--em markup--p-em\">, <\/em><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Metallica<\/em><\/strong><em class=\"markup--em markup--p-em\">, and many more.<\/em><\/p><p id=\"98ef\" class=\"graf graf--p graf-after--p graf--trailing\"><em class=\"markup--em markup--p-em\">He spends most of his time in the San Francisco Bay Area with the most beautiful woman in the world.<\/em><\/p><\/div><\/div><\/section>","created_at":"2018-11-07T16:52:29+0000","updated_at":"2018-11-07T16:52:29+0000","published_at":"2017-07-20T04:35:40+0000","published_by":["Eric Elliott"],"annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*uVpU7iruzXafhU2VLeH4lw.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":577,"title":"Building a fast, secured and free static site in less than three hours","url":"https:\/\/news.ycombinator.com\/item?id=15226122","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=douglasfshearer\" class=\"hnuser\">douglasfshearer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227756\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nerdponx\" class=\"hnuser\">nerdponx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15230123\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaeloblak\" class=\"hnuser\">michaeloblak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15236642\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=homero\" class=\"hnuser\">homero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15236841\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaeloblak\" class=\"hnuser\">michaeloblak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15236914\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tinalumfoil\" class=\"hnuser\">tinalumfoil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15237149\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaeloblak\" class=\"hnuser\">michaeloblak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15237444\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaeloblak\" class=\"hnuser\">michaeloblak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15237551\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=whorleater\" class=\"hnuser\">whorleater<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15237498\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pasquinelli\" class=\"hnuser\">pasquinelli<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15232887\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alias_neo\" class=\"hnuser\">alias_neo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226617\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pc86\" class=\"hnuser\">pc86<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227026\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=option_greek\" class=\"hnuser\">option_greek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227528\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bluetwo\" class=\"hnuser\">bluetwo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228861\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j_s\" class=\"hnuser\">j_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228939\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dhimes\" class=\"hnuser\">dhimes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229936\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rdiddly\" class=\"hnuser\">rdiddly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15230264\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pc86\" class=\"hnuser\">pc86<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227914\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227261\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pc86\" class=\"hnuser\">pc86<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227291\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomsmeding\" class=\"hnuser\">tomsmeding<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227724\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wolfgang42\" class=\"hnuser\">wolfgang42<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228128\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227896\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dsr_\" class=\"hnuser\">dsr_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228067\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kuschku\" class=\"hnuser\">kuschku<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229315\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=grzm\" class=\"hnuser\">grzm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235357\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kuschku\" class=\"hnuser\">kuschku<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235620\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=grzm\" class=\"hnuser\">grzm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235842\">on Sept 13, 2017<\/a><\/p><br \/><div class=\"comment\">Dangerous territory, here :) To be very clear, this is my speculation on what's going on in other people's minds. It can be useful to think about (as it's an exercise in empathy) but often dangerous to explicitly list everything publicly, as one has to evaluate charitable and uncharitable readings, and in general expressing the uncharitable ones, while realistic, tend to draw the most attention (it's just human nature), and in effective communication one should strive to act on the charitable interpretations while seeking to avoid the uncharitable ones. So, please be gentle ;)<p>The primary issue I see here is that it's an iPhone thread on the day of an announcement. Emotions are going to be even higher than normal and I always take extra care when entering such waters. I remind myself that effective communication takes two parties, and this particular ground is already seeded with land mines.<\/p><p>Questions I'd ask myself, even of facts:<\/p><p>Is this worth saying? Does it add to the discussion? How much does it add to the discussion even if I am able to couch it in the best possible way? Am I correcting someone? How do people react to correction, even if I'm right? What is the reaction to the least charitable reading of what I'm posting? Am I on topic (both in the subthread and the submission)?<\/p><p>Putting on my <i>most uncharitable<\/i> hat here, I'd say this subthread starting with 'xvolter's comment was doomed from the start.<\/p><p><a href=\"https:\/\/news.ycombinator.com\/item?id=15230968\" rel=\"nofollow\">https:\/\/news.ycombinator.com\/item?id=15230968<\/a><\/p><p>It can be read as nitpicky, needlessly negative, and incendiary (likely to evoke a strong, defensive emotional response). The lead-in \"Are they aware\" can be read as very condescending. I haven't read up on the iPhone release, but I suspect Apple used the phrase \"surgical steel\" to describe some component of the iPhone. To take them to task for that is pedantic. It's marketing speak. You have to expect some level of exaggeration. Even if it's technically correct, what is this correction going to contribute to the thread on iPhone X? Is it likely to create constructive discussion? From what I've observed, some HN members will down-vote such comments and those in the subthread <i>even if they're correct<\/i> because it's not constructive to the overall discussion. And flame wars not only are destructive in their own thread, they tend to poison the entire forum as it promotes needless argumentation. Don't get me wrong: it's easy to get caught up in such a thread, thinking \"Look, once I point out this, it'll be resolved. It's just a clarification. I'm just trying to help.\" I know, because I've been there.<\/p><p>Also note that even if this is not exactly what was going on in any one down-voter's head, it's useful in that it gets me thinking about how it might be received and generally how I can improve my comment\u2014which might mean not commenting at all.<\/p><p>To put this even more personally, thoughts that have been going through my mind while composing this post:<\/p><p>- Gosh, this is getting long. I'd really like to tighten it up. Is it worth the effort? How much time should I take reviewing and editing it before posting? (After writing but before posting I'll be sure to review the entire comment, but I know I'm not going to tighten it up much because I'm tired.)<\/p><p>- What I've written can easily be misconstrued, particularly because I've written things that are uncharitable, even though I've put in the disclaimer that's what I'm doing on purpose as an exercise. It's late, and I know I haven't thought through everything or the entire thing very thoroughly. I know that in and of itself opens it up to misunderstanding. Are people going to misread what I've written because I've left something out? Is the possibility of that miscommunication greater than the value the post is contributing?<\/p><p>- I'm particularly concerned about my explicit, uncharitable reading of 'xvolter's comment and the response to the entire subthread. Is it necessary to include this for my comment as a whole to be effective? Can I write the comment while excluding it? How much time would that take? I'm leaving it in because I think it is useful and on balance, really, what does it matter in this single instance? (But it <i>does<\/i> matter, on some level. It's a reflection on me.)<\/p><p>I'll end up clicking \"reply\" because I think you've put yourself out there and sound like you're sincerely looking to improve your writing, and that's worth the effort. I also think that it's clear that my intent is clear and good, and I'm know I myself am still learning: if there's some negative reaction to this, I'll take that and use it to improve my writing next time. I certainly know I've got more to learn :)<\/p><p>In sum: it's a volatile submission, it's an off-topic subthread, and getting involved is fraught with danger.<\/p><p>Wow. Did I really need all of those words if that short summary is all there is to it?\n              <\/p><\/div><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ge96\" class=\"hnuser\">ge96<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233787\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233264\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mikekchar\" class=\"hnuser\">mikekchar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234749\">on Sept 13, 2017<\/a><\/p><br \/><div class=\"comment\">I often use HN to practice writing.  Good writing resonates with people.  A post that is well thought out, well researched, communicates clearly and doesn't assume anything of the reader will often get up votes in my experience.<p>For me, the vast majority of down votes come from me being wrong (that is, I post something that I think is correct; I get down votes; I do more research and discover that I am, indeed, incorrect).  In fact, this is such a good flag for my writing that when I get a down vote I immediately assume I must have made a mistake somewhere and start looking for it.<\/p><p>The next most common source of down votes (for me) is being argumentative.  If I get a down vote, but decide that I am correct, I often discover that the tone of my writing is derisive.  In all honesty, I never intend to have this tone of writing, but I am human and sometimes (either through lack of writing skill, or by unconsciously being an ass) it happens.  My guess is that the person I'm responding too gets pissed off and down votes me.  This has led me to have a kind of rule to try very hard never to piss off the parent poster, even if I think they are wrong.  If I can't think of a way to comment without pissing off the parent, I try not to post.  As a huge bonus, this avoids me getting dragged into &gt;90% of trolling efforts (a statistic that I made up).<\/p><p>The next most common cause of down votes for me is lack of coherency in the comment.  Sometimes people simply don't understand what I'm saying.  It is important to realise that as a writer, you can't assume the readers understand what you are thinking, so this is completely down to you.  Trying to put yourself into the role of a random reader and figuring out what they might think when they read your post is hard.  However, this is what being a good writer is about.  Again, if I write a post and upon reviewing it, I don't think most people will understand, I don't submit it.<\/p><p>Related to the last point, there are lots of times where I have relatively unique ideas.  Very frequently people will not know what to think about them because they have never really explored that way of thinking.  If I write in a way that makes it clear that I assume that the reader will find what I'm saying is obvious, I will get down votes.  In a way it's insulting to say something that is not mainstream and assume that others will instantly understand\/agree.  Sometimes, if I feel the idea is interesting, I'll still post, but I'll try to broach the subject gently\/apologetically.  People are doing <i>me<\/i> a favour by getting out of their comfort zone and considering a different way of thinking.<\/p><p>Finally, when I'm posting, I often pick someone to respond to and try my best to help them.  If I am successful, I find that the post is usually highly regarded by many people, even though my focus is only on helping one person. YMMV!\n              <\/p><\/div><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=grzm\" class=\"hnuser\">grzm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235386\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235422\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"400\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=grzm\" class=\"hnuser\">grzm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235457\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"440\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235536\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235256\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pc86\" class=\"hnuser\">pc86<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227904\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=always_good\" class=\"hnuser\">always_good<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228810\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15232862\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=carroccio\" class=\"hnuser\">carroccio<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233301\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmirate\" class=\"hnuser\">mmirate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233346\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eps\" class=\"hnuser\">eps<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228119\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=curun1r\" class=\"hnuser\">curun1r<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229023\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lord_Zero\" class=\"hnuser\">Lord_Zero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227199\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=clra\" class=\"hnuser\">clra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228034\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=unusximmortalis\" class=\"hnuser\">unusximmortalis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228528\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickjj\" class=\"hnuser\">nickjj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227298\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AdmiralAsshat\" class=\"hnuser\">AdmiralAsshat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228129\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RUG3Y\" class=\"hnuser\">RUG3Y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229459\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanwilson\" class=\"hnuser\">seanwilson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15230389\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrei_says_\" class=\"hnuser\">andrei_says_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234182\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanwilson\" class=\"hnuser\">seanwilson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235196\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mi100hael\" class=\"hnuser\">mi100hael<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228461\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ksejka\" class=\"hnuser\">ksejka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233172\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tbrock\" class=\"hnuser\">tbrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227397\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucideer\" class=\"hnuser\">lucideer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227554\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=buro9\" class=\"hnuser\">buro9<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227773\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Semiapies\" class=\"hnuser\">Semiapies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228383\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KirinDave\" class=\"hnuser\">KirinDave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228161\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=onion2k\" class=\"hnuser\">onion2k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227510\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reaperducer\" class=\"hnuser\">reaperducer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228453\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RUG3Y\" class=\"hnuser\">RUG3Y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229440\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomsmeding\" class=\"hnuser\">tomsmeding<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227688\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=msla\" class=\"hnuser\">msla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228263\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reaperducer\" class=\"hnuser\">reaperducer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228592\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ant6n\" class=\"hnuser\">ant6n<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228630\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mfoy_\" class=\"hnuser\">mfoy_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228857\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ant6n\" class=\"hnuser\">ant6n<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15268444\">on Sept 17, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=msla\" class=\"hnuser\">msla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233483\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=onion2k\" class=\"hnuser\">onion2k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228657\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=msla\" class=\"hnuser\">msla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233487\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=matt_j\" class=\"hnuser\">matt_j<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234100\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eterm\" class=\"hnuser\">eterm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227599\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shakna\" class=\"hnuser\">shakna<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235942\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reificator\" class=\"hnuser\">reificator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231874\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=giancarlostoro\" class=\"hnuser\">giancarlostoro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229372\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jgaa\" class=\"hnuser\">jgaa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227706\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bdcravens\" class=\"hnuser\">bdcravens<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231881\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fapjacks\" class=\"hnuser\">fapjacks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233807\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bdcravens\" class=\"hnuser\">bdcravens<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234222\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fapjacks\" class=\"hnuser\">fapjacks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234729\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=meritt\" class=\"hnuser\">meritt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229015\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RUG3Y\" class=\"hnuser\">RUG3Y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229424\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LolWolf\" class=\"hnuser\">LolWolf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229475\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RUG3Y\" class=\"hnuser\">RUG3Y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15239365\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LolWolf\" class=\"hnuser\">LolWolf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15239804\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=krapp\" class=\"hnuser\">krapp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234306\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tonyztan\" class=\"hnuser\">tonyztan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229602\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwilliams\" class=\"hnuser\">jwilliams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226602\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=elliotec\" class=\"hnuser\">elliotec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228064\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwilliams\" class=\"hnuser\">jwilliams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234396\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=quicksnap\" class=\"hnuser\">quicksnap<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231511\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226557\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=abricot\" class=\"hnuser\">abricot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226809\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dabeeeenster\" class=\"hnuser\">dabeeeenster<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227385\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fazkan\" class=\"hnuser\">fazkan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227005\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=abricot\" class=\"hnuser\">abricot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15237818\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=deepakkarki\" class=\"hnuser\">deepakkarki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227347\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wutwutwutwut\" class=\"hnuser\">wutwutwutwut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227390\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AndrewStephens\" class=\"hnuser\">AndrewStephens<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228536\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mcjiggerlog\" class=\"hnuser\">mcjiggerlog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227389\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=qubyte\" class=\"hnuser\">qubyte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228832\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=falsedan\" class=\"hnuser\">falsedan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226469\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jordanrobinson\" class=\"hnuser\">jordanrobinson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226555\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=czechdeveloper\" class=\"hnuser\">czechdeveloper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226514\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=discreditable\" class=\"hnuser\">discreditable<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228187\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228463\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=desireco42\" class=\"hnuser\">desireco42<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228243\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=discreditable\" class=\"hnuser\">discreditable<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228405\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=IgorPartola\" class=\"hnuser\">IgorPartola<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227723\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j_s\" class=\"hnuser\">j_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229049\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Matrixik\" class=\"hnuser\">Matrixik<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15235920\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=trextrex\" class=\"hnuser\">trextrex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226504\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paulgb\" class=\"hnuser\">paulgb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226542\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=austenallred\" class=\"hnuser\">austenallred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226518\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JorgeGT\" class=\"hnuser\">JorgeGT<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226526\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=trextrex\" class=\"hnuser\">trextrex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227170\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thesmallestcat\" class=\"hnuser\">thesmallestcat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226638\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrift\" class=\"hnuser\">chrift<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227180\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sleepychu\" class=\"hnuser\">sleepychu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227853\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ejcx\" class=\"hnuser\">ejcx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227846\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=orf\" class=\"hnuser\">orf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227197\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dan1234\" class=\"hnuser\">dan1234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226703\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=schrodinger\" class=\"hnuser\">schrodinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226757\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=xxdesmus\" class=\"hnuser\">xxdesmus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226930\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227260\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=schrodinger\" class=\"hnuser\">schrodinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233406\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15236497\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=schrodinger\" class=\"hnuser\">schrodinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15239392\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ic4l\" class=\"hnuser\">ic4l<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226527\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226561\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j_s\" class=\"hnuser\">j_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228870\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ekianjo\" class=\"hnuser\">ekianjo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226532\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=r3bl\" class=\"hnuser\">r3bl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226650\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanwilson\" class=\"hnuser\">seanwilson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226575\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sgallant\" class=\"hnuser\">sgallant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227500\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226598\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cmsd2\" class=\"hnuser\">cmsd2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226627\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226673\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanwilson\" class=\"hnuser\">seanwilson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226636\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aloisdg\" class=\"hnuser\">aloisdg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226926\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jloughry\" class=\"hnuser\">jloughry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233517\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jpz\" class=\"hnuser\">jpz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227398\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=miketery\" class=\"hnuser\">miketery<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15243956\">on Sept 14, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=255kb\" class=\"hnuser\">255kb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226577\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alfonsodev\" class=\"hnuser\">alfonsodev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227808\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=in9\" class=\"hnuser\">in9<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229369\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mizzao\" class=\"hnuser\">mizzao<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226956\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=diggan\" class=\"hnuser\">diggan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226974\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=secminion\" class=\"hnuser\">secminion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227359\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rootlocus\" class=\"hnuser\">rootlocus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227502\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MatthewK\" class=\"hnuser\">MatthewK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226831\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ajanuary\" class=\"hnuser\">ajanuary<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226967\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lord_Zero\" class=\"hnuser\">Lord_Zero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227229\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moduspwnens14\" class=\"hnuser\">moduspwnens14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227520\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dombili\" class=\"hnuser\">dombili<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227009\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kadal\" class=\"hnuser\">kadal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229165\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=azdle\" class=\"hnuser\">azdle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229289\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kadal\" class=\"hnuser\">kadal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229220\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=IamNotAtWork\" class=\"hnuser\">IamNotAtWork<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15226921\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=godman_8\" class=\"hnuser\">godman_8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227139\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ge96\" class=\"hnuser\">ge96<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233803\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=iscoelho\" class=\"hnuser\">iscoelho<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227142\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=justinhj\" class=\"hnuser\">justinhj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227645\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=megawatthours\" class=\"hnuser\">megawatthours<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227711\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KirinDave\" class=\"hnuser\">KirinDave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228156\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomc1985\" class=\"hnuser\">tomc1985<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228955\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CrystalGamma\" class=\"hnuser\">CrystalGamma<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231826\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RobGav\" class=\"hnuser\">RobGav<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15341489\">on Sept 26, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomc1985\" class=\"hnuser\">tomc1985<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231991\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=devopsproject\" class=\"hnuser\">devopsproject<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15230407\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=testloop\" class=\"hnuser\">testloop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227572\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=majewsky\" class=\"hnuser\">majewsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227611\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ArlenBales\" class=\"hnuser\">ArlenBales<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228377\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=testloop\" class=\"hnuser\">testloop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231387\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ArchReaper\" class=\"hnuser\">ArchReaper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228672\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j_s\" class=\"hnuser\">j_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229128\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=majewsky\" class=\"hnuser\">majewsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15257289\">on Sept 15, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=plg\" class=\"hnuser\">plg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228826\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RUG3Y\" class=\"hnuser\">RUG3Y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229494\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Iv\" class=\"hnuser\">Iv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15234701\">on Sept 13, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=free_everybody\" class=\"hnuser\">free_everybody<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231916\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=free_everybody\" class=\"hnuser\">free_everybody<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15231887\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227758\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flangillion\" class=\"hnuser\">flangillion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227935\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228162\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j_s\" class=\"hnuser\">j_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229170\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229278\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sysashi\" class=\"hnuser\">sysashi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228457\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=carlmr\" class=\"hnuser\">carlmr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227468\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomcam\" class=\"hnuser\">tomcam<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15232614\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Spacemolte\" class=\"hnuser\">Spacemolte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227088\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gsempe\" class=\"hnuser\">gsempe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15229426\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=therealmarv\" class=\"hnuser\">therealmarv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15227099\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomc1985\" class=\"hnuser\">tomc1985<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15228942\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eighthnate\" class=\"hnuser\">eighthnate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15233080\">on Sept 12, 2017<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:52:28+0000","updated_at":"2018-11-07T16:52:28+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":578,"title":"electron-userland\/electron-builder","url":"https:\/\/github.com\/electron-userland\/electron-builder\/blob\/master\/README.md","content":"<p>A complete solution to package and build a ready for distribution <a href=\"https:\/\/electronjs.org\" rel=\"nofollow\">Electron<\/a>, <a href=\"https:\/\/proton-native.js.org\/\" rel=\"nofollow\">Proton Native<\/a> or <a href=\"https:\/\/github.com\/brave\/muon\">Muon<\/a> app for macOS, Windows and Linux with \u201cauto update\u201d support out of the box.<\/p>\n<p>See documentation on <a href=\"https:\/\/www.electron.build\" rel=\"nofollow\">electron.build<\/a>.<\/p>\n<ul><li>NPM packages management:\n<ul><li><a href=\"https:\/\/electron.atom.io\/docs\/tutorial\/using-native-node-modules\/\" rel=\"nofollow\">Native application dependencies<\/a> compilation (including <a href=\"http:\/\/yarnpkg.com\/\" rel=\"nofollow\">Yarn<\/a> support).<\/li>\n<li>Development dependencies are never included. You don't need to ignore them explicitly.<\/li>\n<li><a href=\"https:\/\/www.electron.build\/tutorials\/two-package-structure\" rel=\"nofollow\">Two package.json structure<\/a> is supported, but you are not forced to use it even if you have native production dependencies.<\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/www.electron.build\/code-signing\" rel=\"nofollow\">Code Signing<\/a> on a CI server or development machine.<\/li>\n<li><a href=\"https:\/\/www.electron.build\/auto-update\" rel=\"nofollow\">Auto Update<\/a> ready application packaging.<\/li>\n<li>Numerous target formats:\n<ul><li>All platforms: <code>7z<\/code>, <code>zip<\/code>, <code>tar.xz<\/code>, <code>tar.7z<\/code>, <code>tar.lz<\/code>, <code>tar.gz<\/code>, <code>tar.bz2<\/code>, <code>dir<\/code> (unpacked directory).<\/li>\n<li><a href=\"https:\/\/www.electron.build\/configuration\/mac\" rel=\"nofollow\">macOS<\/a>: <code>dmg<\/code>, <code>pkg<\/code>, <code>mas<\/code>.<\/li>\n<li><a href=\"https:\/\/www.electron.build\/configuration\/linux\" rel=\"nofollow\">Linux<\/a>: <a href=\"http:\/\/appimage.org\" rel=\"nofollow\">AppImage<\/a>, <a href=\"http:\/\/snapcraft.io\" rel=\"nofollow\">snap<\/a>, debian package (<code>deb<\/code>), <code>rpm<\/code>, <code>freebsd<\/code>, <code>pacman<\/code>, <code>p5p<\/code>, <code>apk<\/code>.<\/li>\n<li><a href=\"https:\/\/www.electron.build\/configuration\/win\" rel=\"nofollow\">Windows<\/a>: <code>nsis<\/code> (Installer), <code>nsis-web<\/code> (Web installer), <code>portable<\/code> (portable app without installation), AppX (Windows Store), MSI, Squirrel.Windows.<\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/www.electron.build\/configuration\/publish\" rel=\"nofollow\">Publishing artifacts<\/a> to GitHub Releases, Amazon S3, DigitalOcean Spaces and Bintray.<\/li>\n<li>Advanced building:\n<ul><li>Pack in a distributable format <a href=\"https:\/\/www.electron.build\/#pack-only-in-a-distributable-format\" rel=\"nofollow\">already packaged app<\/a>.<\/li>\n<li>Separate <a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/issues\/1102#issuecomment-271845854\">build steps<\/a>.<\/li>\n<li>Build and publish in parallel, using hard links on CI server to reduce IO and disk space usage.<\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/github.com\/electron\/electron-compile\">electron-compile<\/a> support (compile for release-time on the fly on build).<\/li>\n<li><a href=\"https:\/\/www.electron.build\/multi-platform-build#docker\" rel=\"nofollow\">Docker<\/a> images to build Electron app for Linux or Windows on any platform.<\/li>\n<li><a href=\"https:\/\/www.electron.build\/configuration\/configuration\/#proton-native\" rel=\"nofollow\">Proton Native<\/a> and <a href=\"https:\/\/github.com\/brave\/muon\">Muon<\/a> support.<\/li>\n<\/ul><table><thead><tr><th>Question<\/th>\n<th>Answer<\/th>\n<\/tr><\/thead><tbody><tr><td>\u201cI want to configure electron-builder\u201d<\/td>\n<td><a href=\"https:\/\/electron.build\/configuration\/configuration\" rel=\"nofollow\">See options<\/a><\/td>\n<\/tr><tr><td>\u201cI have a question\u201d<\/td>\n<td><a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/issues\">Open an issue<\/a> or <a href=\"https:\/\/slackin.electron.build\" rel=\"nofollow\">join the chat<\/a><\/td>\n<\/tr><tr><td>\u201cI found a bug\u201d<\/td>\n<td><a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/issues\/new\">Open an issue<\/a><\/td>\n<\/tr><tr><td>\u201cI want to support development\u201d<\/td>\n<td><a href=\"https:\/\/www.electron.build\/donate\" rel=\"nofollow\">Donate<\/a><\/td>\n<\/tr><\/tbody><\/table><p>Real project example \u2014 <a href=\"https:\/\/github.com\/develar\/onshape-desktop-shell\">onshape-desktop-shell<\/a>.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p><a href=\"http:\/\/yarnpkg.com\/\" rel=\"nofollow\">Yarn<\/a> is <a href=\"https:\/\/github.com\/electron-userland\/electron-builder\/issues\/1147#issuecomment-276284477\">strongly<\/a> recommended instead of npm.<\/p>\n<p><code>yarn add electron-builder --dev<\/code><\/p>\n<h2><a id=\"user-content-quick-setup-guide\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-setup-guide\"><\/a>Quick Setup Guide<\/h2>\n<p><a href=\"https:\/\/github.com\/electron-userland\/electron-webpack-quick-start\">electron-webpack-quick-start<\/a> is a recommended way to create a new Electron application. See <a href=\"https:\/\/www.electron.build\/#boilerplates\" rel=\"nofollow\">Boilerplates<\/a>.<\/p>\n<ol><li>\n<p>Specify the standard fields in the application <code>package.json<\/code> \u2014 <a href=\"https:\/\/electron.build\/configuration\/configuration#Metadata-name\" rel=\"nofollow\">name<\/a>, <code>description<\/code>, <code>version<\/code> and <a href=\"https:\/\/docs.npmjs.com\/files\/package.json#people-fields-author-contributors\" rel=\"nofollow\">author<\/a>.<\/p>\n<\/li>\n<li>\n<p>Specify the <a href=\"https:\/\/electron.build\/configuration\/configuration#build\" rel=\"nofollow\">build<\/a> configuration in the <code>package.json<\/code> as follows:<\/p>\n<div class=\"highlight highlight-source-json\"><pre>\"build\": {\n  \"appId\": \"your.id\",\n  \"mac\": {\n    \"category\": \"your.app.category.type\"\n  }\n}<\/pre><\/div>\n<p>See <a href=\"https:\/\/www.electron.build\/configuration\/configuration\" rel=\"nofollow\">all options<\/a>. Option <a href=\"https:\/\/www.electron.build\/configuration\/contents#files\" rel=\"nofollow\">files<\/a> to indicate which files should be packed in the final application, including the entry file, maybe required.<\/p>\n<\/li>\n<li>\n<p>Add <a href=\"https:\/\/www.electron.build\/icons\" rel=\"nofollow\">icons<\/a>.<\/p>\n<\/li>\n<li>\n<p>Add the <a href=\"https:\/\/docs.npmjs.com\/cli\/run-script\" rel=\"nofollow\">scripts<\/a> key to the development <code>package.json<\/code>:<\/p>\n<div class=\"highlight highlight-source-json\"><pre>\"scripts\": {\n  \"pack\": \"electron-builder --dir\",\n  \"dist\": \"electron-builder\"\n}<\/pre><\/div>\n<p>Then you can run <code>yarn dist<\/code> (to package in a distributable format (e.g. dmg, windows installer, deb package)) or <code>yarn pack<\/code> (only generates the package directory without really packaging it. This is useful for testing purposes).<\/p>\n<p>To ensure your native dependencies are always matched electron version, simply add script <code>\"postinstall\": \"electron-builder install-app-deps\"<\/code> to your <code>package.json<\/code>.<\/p>\n<\/li>\n<li>\n<p>If you have native addons of your own that are part of the application (not as a dependency), set <a href=\"https:\/\/www.electron.build\/configuration\/configuration#Configuration-nodeGypRebuild\" rel=\"nofollow\">nodeGypRebuild<\/a> to <code>true<\/code>.<\/p>\n<\/li>\n<\/ol><p>Please note that everything is packaged into an asar archive <a href=\"https:\/\/electron.build\/configuration\/configuration#Configuration-asar\" rel=\"nofollow\">by default<\/a>.<\/p>\n<p>For an app that will be shipped to production, you should sign your application. See <a href=\"https:\/\/www.electron.build\/code-signing#where-to-buy-code-signing-certificate\" rel=\"nofollow\">Where to buy code signing certificates<\/a>.<\/p>\n<h2><a id=\"user-content-donate\" class=\"anchor\" aria-hidden=\"true\" href=\"#donate\"><\/a>Donate<\/h2>\n<p>We do this open source work in our free time. If you'd like us to invest more time on it, please <a href=\"https:\/\/www.electron.build\/donate\" rel=\"nofollow\">donate<\/a>. Donation can be used to increase some issue priority.<\/p>\n<h2><a id=\"user-content-sponsors\" class=\"anchor\" aria-hidden=\"true\" href=\"#sponsors\"><\/a>Sponsors<\/h2>\n<p><a href=\"https:\/\/workflowy.com\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f39892b26125d1fa6789f2cdf75abc6d17103cf2\/68747470733a2f2f776f726b666c6f77792e636f6d2f6d656469612f692f77662d69636f6e2d616e64726f69642e706e67\" alt=\"WorkFlowy\" title=\"WorkFlowy\" width=\"75\" height=\"75\" data-canonical-src=\"https:\/\/workflowy.com\/media\/i\/wf-icon-android.png\" \/><\/a>\n<a href=\"https:\/\/tidepool.org\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a22fd6a3f4c9f7582287034b91364d35beb4ffdd\/68747470733a2f2f7777772e656c656374726f6e2e6275696c642f73706f6e736f722d6c6f676f732f54696465706f6f6c5f4c6f676f5f4c696768742e737667\" alt=\"Tidepool\" title=\"Tidepool\" data-canonical-src=\"https:\/\/www.electron.build\/sponsor-logos\/Tidepool_Logo_Light.svg\" \/><\/a><\/p>","created_at":"2018-11-07T16:52:28+0000","updated_at":"2018-11-07T16:52:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/17519142?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":576,"title":"Terminix is a Tiling Terminal Emulator for Linux","url":"https:\/\/www.omgubuntu.co.uk\/2017\/07\/tilix-terminix-terminal-emulator-ubuntu","content":"<p><a href=\"http:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-logo.png\"><img class=\"alignright size-medium wp-image-122777\" src=\"http:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-logo-300x300.png\" alt=\"tilix app icon\" width=\"300\" height=\"300\" srcset=\"https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-logo-300x300.png 300w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-logo-32x32.png 32w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-logo.png 512w\" \/><\/a><strong>Whenever I write about a terminal emulator here on OMG! Ubuntu! I get a lot of comments from readers asking to know why I didn\u2019t mention <em>x-alternative<\/em> or\u00a0<em>y-equivalent<\/em>\u00a0instead.<\/strong><\/p>\n<p>And amongst those comments there\u2019s one terminal app that gets called out time and time again: <strong>Terminix.<\/strong><\/p>\n<p>Now, admittedly, I don\u2019t live at the command line; I regularly use <code>apt<\/code>, <code>snap<\/code> and a few other CLI utilities, and occasionally checkout Git branches.<\/p>\n<p>So, for my needs at least, GNOME Terminal, the default terminal emulator shipped on Ubuntu, has all of the features and functionality I need.<\/p>\n<p>But each of us use our Ubuntu machines differently. GNOME Terminal may not be your cup of tea.<\/p>\n<p>It helps that the Ubuntu archive is packed full of all kinds of terrific terminal emulators, including\u00a0Terminator and Tilda, the drop-down Guake, and the aforementioned Terminix.<\/p>\n<p>Wanting to see what the fuss is all about, I installed Terminix from the Ubuntu Software store.<\/p>\n<h3>Terminix \u2014 The Best Terminal App on Linux?<\/h3>\n<p>Terminix\/Tilix is a <strong>tiling terminal emulator. <\/strong>This\u00a0means it can run more than one terminal in the same window at the same time. You can split terminal sessions and tile them both vertically and horizontally.<\/p>\n<p><a href=\"http:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu.png\"><img class=\"wp-image-122772 size-large aligncenter\" src=\"http:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu-750x431.png\" alt=\"terminix aka tilix on ubuntu\" width=\"750\" height=\"431\" srcset=\"https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu-750x431.png 750w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu-350x200.png 350w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu-300x172.png 300w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu-768x441.png 768w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-on-ubuntu.png 1243w\" \/><\/a><\/p>\n<p>Terminix had to\u00a0<a href=\"https:\/\/gnunn1.github.io\/tilix-web\/2017\/03\/19\/release-1-5-4\" target=\"_blank\" rel=\"noopener\">change its name<\/a> earlier in the year and is now known as <strong>Tilix<\/strong>. The Ubuntu archives has an older build of the app available, which is what I\u2019ve used in this post, and as such still it still uses the Terminix name.<\/p>\n<p>The overall layout and design of the app draws heavily from the\u00a0<a href=\"https:\/\/developer.gnome.org\/hig\/stable\/\">GNOME Human Interface Guidelines<\/a> (HIG). As such, Terminix\/Tilix looks <em>the<\/em> business on a modern GNOME desktop like GNOME Shell and Budgie.<\/p>\n<p>There\u2019s also support for using different profiles in each terminal pane you run; and the ability to run multiple sessions simultaneously within the app and move between them using a thumbnail switcher:<\/p>\n<p><a href=\"http:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/terminix-.jpg\"><img class=\"aligncenter size-large wp-image-122781\" src=\"http:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/terminix--750x422.jpg\" alt=\"terminix terminal emulator running on ubuntu 17.04\" width=\"750\" height=\"422\" srcset=\"https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/terminix--750x422.jpg 750w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/terminix--300x169.jpg 300w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/terminix--768x432.jpg 768w, https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/terminix-.jpg 1200w\" \/><\/a><\/p>\n<p>Just like tearing a tab from your browser you can drag any terminal out of the app to run it independently or drop it on to another terminal to merge it. Each tiled terminal can also be maximized\u00a0to the full height and widget of the main window at any time, and then minimized back to where it was.<\/p>\n<p>The <a href=\"https:\/\/gnunn1.github.io\/tilix-web\/#features\">features page on the Tilix website<\/a> has more info on the features you can find in this terminal emulator, but here\u2019s a concise overview:<\/p>\n<ul><li>Run multiple panes &amp; multiple sessions<\/li>\n<li>Easily arrange terminals using drag and drop<\/li>\n<li>Add custom terminal\/session titles<\/li>\n<li>VTE switcher<\/li>\n<li>Persistent layouts<\/li>\n<li>Synchronised input (type the same command in all terminals)<\/li>\n<li>Terminal themes\/styling (including transparency)<\/li>\n<\/ul><p>Like the sound of it and want to try it?<\/p>\n<h2>Install Terminix on Ubuntu<\/h2>\n<p>If you\u2019re reading this article from Ubuntu 17.04 you can<strong> install Terminix on Ubuntu from the Ubuntu Software app,\u00a0<\/strong>no PPA required:<\/p>\n<p><a class=\"OMGButton OMGButton--usc\" target=\"_blank\" href=\"denied:apt:\/\/terminix\" title=\"Apt URL to install Terminix terminal emulator on Ubuntu\">Install Terminix on Ubuntu<\/a><\/p>\n<p>This is an older build so not all of the features you can read about on the Tilix website are present (but all of the features mentioned in this post, however, are).<\/p>\n<p>Once installed you can find and open Terminix from the Unity Dash, the GNOME Applications Overview, or whatever app menu\/launcher you\u2019re using.<\/p>\n<h4>Install Tilix on Ubuntu<\/h4>\n<p>If you want to <strong>install\u00a0Tilix\u00a0on Ubuntu<\/strong> you\u00a0need to download the latest release \u2014 \u2018Tilix.zip\u2019 \u2014 from the project\u2019s Github releases page below:<\/p>\n<p><a class=\"OMGButton OMGButton--sourceRepo\" target=\"_blank\" href=\"https:\/\/github.com\/gnunn1\/tilix\/releases\" title=\"Download Latest Tilix Release from Github\">Tilix Releases on Github<\/a><\/p>\n<p>Once downloaded just extract the archive to the correct location and compile the schemas by using these two commands:<\/p>\n<pre>sudo unzip tilix.zip -d \/<\/pre>\n<pre>sudo glib-compile-schemas \/usr\/share\/glib-2.0\/schemas\/<\/pre>\n<p>That\u2019s it.<\/p>\n<p>You can go and and launch \u2018Tilix\u2019 from the Unity Dash, the GNOME Applications Overview, or whatever app menu\/launcher you\u2019re using.<\/p>\n<h4>Your thoughts<\/h4>\n<p>If you try the app out (or have done in the past) do let us know what you think about it by placing a hit and run vote or leaving a comment.<\/p>\n<div class=\"apester-media\" data-media-id=\"5968b7987ea93a275889c8fc\">\n<p>\n                              <\/p><div class=\"Post-links\"><a href=\"https:\/\/www.omgubuntu.co.uk\/tag\/cli\" rel=\"tag\">cli<\/a><a href=\"https:\/\/www.omgubuntu.co.uk\/tag\/terminal-apps\" rel=\"tag\">terminal apps<\/a><a href=\"https:\/\/www.omgubuntu.co.uk\/tag\/terminix\" rel=\"tag\">terminix<\/a><a href=\"https:\/\/www.omgubuntu.co.uk\/tag\/tilix\" rel=\"tag\">tilix<\/a><\/div>                  <\/div>","created_at":"2018-11-07T16:52:25+0000","updated_at":"2018-11-07T16:52:25+0000","published_at":"2017-07-14T13:33:15+0000","published_by":["Joey Sneddon"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"www.omgubuntu.co.uk","preview_picture":"https:\/\/www.omgubuntu.co.uk\/wp-content\/uploads\/2017\/07\/tilix-emulator-on-ubuntu.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":573,"title":"What do I look for in a pitch?","url":"https:\/\/news.greylock.com\/what-do-i-look-for-in-a-pitch-866355bddb3?gi=1accf9edfb9e","content":"<figure id=\"4a0b\" class=\"graf graf--figure graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*hEqSyP88GY3w-LFysHHQxA.jpeg\" data-width=\"1200\" data-height=\"700\" data-action=\"zoom\" data-action-value=\"1*hEqSyP88GY3w-LFysHHQxA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*hEqSyP88GY3w-LFysHHQxA.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"59c9\" class=\"graf graf--p graf-after--h3\">Founders will often ask me, \u201cDo you want to go through our deck?\u201d<br \/>\u00a0<br \/>It\u2019s not about the slide deck. We have made the decision to invest in entrepreneurs without slides, whose ambition, passion, depth of understanding of a opportunity and compelling vision come through crystal-clear in conversation.<br \/>\u00a0<br \/>But\u2026that\u2019s a high bar for communication. Most of the time, putting a business plan on paper helps you structure a conversation and spread knowledge without having the same conversation repeatedly. Most importantly, it clarifies your own thinking as a leader, and is a great way to get your team on the same page\u200a\u2014\u200awith a singular vision for your company and a shared plan for execution.<br \/>\u00a0<br \/>After giving this talk to several groups in our network, I\u2019m sharing the slides here in an attempt to demystify one of the most common questions founders have and spread some tribal knowledge\u200a\u2014\u200a<strong class=\"markup--strong markup--p-strong\">what do I look for in a pitch?<\/strong><\/p><figure id=\"b814\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*Bzmj2BoF2L_1ukUpuNJTOQ.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*Bzmj2BoF2L_1ukUpuNJTOQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Bzmj2BoF2L_1ukUpuNJTOQ.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"3ca1\" class=\"graf graf--p graf-after--figure\">Each stage of company building carries different risks, from a maniacal focus on defining an initial product and just shipping it, to the iteration of finding product-market fit. Next comes figuring out how to grow the user or customer base efficiently and consistently\u200a\u2014\u200aand then when you have that growth formula, pushing as fast (often <a href=\"https:\/\/hbr.org\/2016\/04\/blitzscaling\" data-href=\"https:\/\/hbr.org\/2016\/04\/blitzscaling\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">much faster<\/a>) than you can manage\u200a\u2014\u200ain order to achieve market leadership and establish your defensible moats. If the customer base grows, then you suddenly need to build a company that can support that growth\u200a\u2014\u200aan executive team of leaders and an organization that can operate and adapt effectively without the founder driving every decision. Finally, if you\u2019re right, lucky and have executed exceptionally well, you prepare to face the public markets, navigating the IPO process and becoming an enduring, public company.<br \/>\u00a0<br \/>The company-building path rarely looks as up-and-to-the-right as this chart above. I think of it more like climbing mythical Space Mountain\u200a\u2014\u200aone with constant landslides, at log-scale, in 4-dimensional space.<br \/>\u00a0<br \/>We have enormous respect for the founders who attempt this climb. Few other jobs require you to scale like a startup does\u200a\u2014\u200agrowing and remaking yourself as a leader and your organization 2x, 5x, 10x per year, for multiple years on end. Without breaking it completely. We don\u2019t expect people to have all the answers. But in this post, I\u2019m going to focus on what I look for, especially in Series A companies.\u00a0<br \/>\u00a0<br \/>What\u2019s the difference between a great idea, and a early-stage VC-fundable business?<\/p><figure id=\"281f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*jYlxPMLWNabnYCXoqIrM_Q.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*jYlxPMLWNabnYCXoqIrM_Q.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*jYlxPMLWNabnYCXoqIrM_Q.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"412c\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*3skptEpxp4Dx7In6QgcFEA.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*3skptEpxp4Dx7In6QgcFEA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*3skptEpxp4Dx7In6QgcFEA.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"1ed6\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*YArUHANdi6xS9fbH7S8slQ.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*YArUHANdi6xS9fbH7S8slQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*YArUHANdi6xS9fbH7S8slQ.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"252f\" class=\"graf graf--p graf-after--figure\">Before you try to raise venture funding\u200a\u2014\u200aconfirm that option makes sense for you. Success in partnership with investors is about aligned incentives, vision, and personal chemistry. More on that <a href=\"https:\/\/medium.com\/@saranormous\/not-every-startup-should-be-venture-backed-970d728059d\" data-href=\"https:\/\/medium.com\/@saranormous\/not-every-startup-should-be-venture-backed-970d728059d\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">here<\/a>.<\/p><figure id=\"d561\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*PvzpOfGGD86zVtltlkB4dg.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*PvzpOfGGD86zVtltlkB4dg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*PvzpOfGGD86zVtltlkB4dg.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"05ae\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*KqvQirWmJw9NenHdgF4kvg.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*KqvQirWmJw9NenHdgF4kvg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*KqvQirWmJw9NenHdgF4kvg.jpeg\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><em class=\"markup--em markup--figure-em\">This chart from a CBInsights analysis of startup post-mortems is a good reminder of user and customer demand as a north\u00a0star.<\/em><\/figcaption><\/figure><p id=\"bdde\" class=\"graf graf--p graf-after--figure\">Ok, so let\u2019s say you think you have a venture-backable idea. One of the first things I will try to do is understand if there a market need for the product you are building.\u00a0<br \/>\u00a0<br \/>Some entrepreneurs validate customer need through years of having lived the problem. Others go native with users\u200a\u2014\u200asurvey and interview them, show mocks and prototypes, do demand testing by \u201cselling\u201d or advertising a product they haven\u2019t built yet, even get customers to partner and co-develop with them. There\u2019s no replacement for shipping a product and getting adoption, engagement and revenue data. But that is often expensive (especially for an enterprise-quality product), and getting feedback and direction along the way is incredibly valuable. What\u2019s your feedback cycle? How will you convince yourself, and a VC, that the opportunity is real? You don\u2019t need that much money to get useful feedback.<\/p><figure id=\"f4ca\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*3avHcQRE_254npg7u1FsRA.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*3avHcQRE_254npg7u1FsRA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*3avHcQRE_254npg7u1FsRA.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"9164\" class=\"graf graf--p graf-after--figure\">I don\u2019t expect founders to go through each bullet in the above chart at the Series A, but they should have a point of view on how they\u2019ll reach and convince customers to use and buy what they\u2019re offering. \u201cWinning on product\u201d alone is a rare, minority case. At Greylock, obsess about distribution as much as product.<\/p><p id=\"2ded\" class=\"graf graf--p graf-after--p\">The transformative technologies of the past two decades\u200a\u2014\u200athe internet and mobile\u200a\u2014\u200aare new forms of distribution for technology, and huge companies have risen with them. Many of the internet and enterprise giants today are hard to attack because of their distribution and reach to the customer, not because no engineering team can build something better. If you are thoughtful about distribution, you\u2019ll immediately gain my confidence.<\/p><figure id=\"812e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*ltGEPwTHrlLvMKYzrV7e9Q.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*ltGEPwTHrlLvMKYzrV7e9Q.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ltGEPwTHrlLvMKYzrV7e9Q.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"30d9\" class=\"graf graf--p graf-after--figure\">There\u2019s a lot of information out there about how to get a meeting with a VC, so I won\u2019t focus on that here\u200a\u2014\u200asuffice it to say, cold pitches <em class=\"markup--em markup--p-em\">are<\/em> viable. At Greylock I read cold emails, and take referrals, and meet founders serendipitously. Meeting entrepreneurs is a core part of my job, and if opportunities happen to arrive in my inbox without any work on my part\u200a\u2014\u200aall the better.<br \/>\u00a0<br \/>Unfortunately, most cold emails aren\u2019t very compelling. I\u2019m much more likely to read through a 20-slide deck a founder emails me than a 200-page document. Invest in telling a clear story at different lengths. Your intro should either be a short deck, or 2 paragraphs that introduce the team, the idea, and why it\u2019s special.<\/p><p>How do you introduce your company once you get a first meeting? I hesitate to show an outline, because investing isn\u2019t a checklist decision, but I often get asked for one. The most important takeaway from this slide was that there are only a few key questions that make or break an investment. Some of the best pitches I\u2019ve experienced have only lasted 20\u201330 minutes.\u00a0<br \/>\u00a0<br \/>Do you deeply understand the user problem you\u2019re solving? Who are you and why do you care? Can you explain your unique and compelling value proposition? Do you understand the landscape of what\u2019s out there? Do you have some way of distributing that solution, positioning to that customer why they should care, and why it\u2019s better and worth adopting vs. the competition? If and when others come after you, what are your moats?<\/p><p>If you can answer those 6 questions succinctly, ignore my outline and structure your slides however makes sense. Move everything else to backup material.<\/p><figure id=\"2af8\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*52FBY1emh2AgPW3YA6pHhQ.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*52FBY1emh2AgPW3YA6pHhQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*52FBY1emh2AgPW3YA6pHhQ.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"7277\" class=\"graf graf--p graf-after--figure\">I say \u201cnarrative\u201d is important here\u200a\u2014\u200abut another way to think of this is clarity of purpose. If you can tell your story in an emotionally compelling and authentic way\u200a\u2014\u200athat itself is a huge advantage. Invest the time in articulating a tight narrative. Use concrete examples that people can relate to, and tie them to your larger ambition. It will pay dividends, and not just with investors.<\/p><figure id=\"71da\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*MMEo8J8HkSehZYIIOR-N8Q.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*MMEo8J8HkSehZYIIOR-N8Q.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*MMEo8J8HkSehZYIIOR-N8Q.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"8435\" class=\"graf graf--p graf--startsWithDoubleQuote graf-after--figure\">\u201cMarket\u201d is often hugely misunderstood by entrepreneurs.<br \/>\u00a0<br \/>It is often impossible to answer some of these questions (will it be winner-take-all? how much will customers pay for a new product? how many $ billions will the market be worth?) if you are trying to <strong class=\"markup--strong markup--p-strong\">create a new market category<\/strong>, or trying to significantly <strong class=\"markup--strong markup--p-strong\">reshape an existing category<\/strong>. The answers are unknowable at the beginning. You can have only hypotheses.<br \/>\u00a0<br \/>Other questions (who are you targeting? how does this problem\/product fit into your target customer\u2019s life? is it #2 or #40 in terms of their priorities? how many of those customers are there?) you should be able to answer before you build anything.<br \/>\u00a0<br \/>Great entrepreneurs can change and expand markets, even consumer attitudes. Total focus on solving a problem for a customer you understand helps you dominate a \u201csmall market,\u201d and dominance of a tactical, small market gives an agile and ambitious team a wedge to grow from. <strong class=\"markup--strong markup--p-strong\">Many great companies start from a tactical feature in a space full of problems. Initially, they must fight to keep maintain focus and simplicity, and solve a single problem well. That is often their competitive advantage vs. large incumbents.<\/strong> The deepening of product value and expansion of surface area and customer base that happens once you have that foothold is a far more linear task than the black magic of solving that initial problem.<br \/>\u00a0<br \/>The difference then between the companies that get trapped in niches, and those that become strategic, are the teams that have the ambition to keep growing the problem, and the agility and talent to keep attacking it. At the Series A, initial market size is often irrelevant, and eventual market size is hard to see.<br \/>\u00a0<br \/>It would have been easy to write off Airbnb as attacking a niche: creatives who were willing to rent out air mattresses for a few extra bucks. The number of people with underutilized real estate and are interested in sharing their homes and making extra income is larger than anyone could have anticipated. Now it\u2019s one of the biggest hospitality businesses in the world.<\/p><figure id=\"3d4d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*SjDcX3T-f8Mfqv9pljnPZA.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*SjDcX3T-f8Mfqv9pljnPZA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*SjDcX3T-f8Mfqv9pljnPZA.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"f0d3\" class=\"graf graf--p graf-after--figure\">For anyone who hasn\u2019t seen Geoffrey Moore\u2019s classic template, it\u2019s a great place to start in terms of describe your market and your positioning. Try to write this, then test it with your target customer, and other people who know that customer well (advisors, the channel, salespeople).<br \/>\u00a0<br \/>A common red flag is teams choosing too broad an initial target customer. The \u201cglobal 20,000 largest companies\u201d or \u201cconsumers and enterprises\u201d may be in your grand plan, but who will care most? First? How will you find your first 100,000 users or your first 100 business customers in the first year? What defines that group?<br \/>\u00a0<br \/>Because many founders are engineers from Silicon Valley, they often rely on fast-growing Silicon Valley tech companies and other engineers to be their first customers. This is sometimes good, sometimes not. It\u2019s a customer base of early technology adopters they understand\u200a\u2014\u200aso it\u2019s easier and more authentic. At the same time, most b2b companies aren\u2019t technology companies, and most consumers aren\u2019t engineers.\u00a0<br \/>\u00a0<br \/>I ask three questions when I see this. How far can that initial tech customer get you? Do you plan to continue to serve this customer\u200a\u2014\u200aand if so, can you be essential to them? If you plan to expand your customer base beyond that\u200a\u2014\u200ahow will you understand your other customers, and how can you leverage initial adoption into broader adoption?<\/p><figure id=\"3d2e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*gtD94kUq7oytqRUsSO9NSw.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*gtD94kUq7oytqRUsSO9NSw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*gtD94kUq7oytqRUsSO9NSw.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"c5cc\" class=\"graf graf--p graf-after--figure\">We love to talk about \u201cnetwork effects\u201d in Silicon Valley, but many believe that starts and end with the social, viral growth of Facebook. But network effects can build many kinds of moats that defend them against copycats and price erosion.<\/p><figure id=\"f429\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*ahoXgP8uJwmhOwUUSxVBMA.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*ahoXgP8uJwmhOwUUSxVBMA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ahoXgP8uJwmhOwUUSxVBMA.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"cf2a\" class=\"graf graf--p graf-after--figure\">I don\u2019t describe every kind of moat here (e.g., regulatory) and many great companies will have new types of moats\u200a\u2014\u200aones yet to be invented. No moat is absolute or permanent or I\u2019d be out of a job! Still, having a thesis on your defensibility is essential, even at the Series A.<\/p><ul class=\"postList\"><li id=\"967a\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Tech. <\/strong>We continue to believe in new and better fundamental technology as a moat\u200a\u2014\u200anot so much patents or legal protection, but products that are hard to build well or take advantage of a fundamentally new approach. In extreme cases, a technology advantage that translates into a customer benefit can turn into a standard\u200a\u2014\u200aan effective monopoly\u200a\u2014\u200aas in the case of Intel. In most cases, even a great technology moat only gets you far enough to become a market leader, buying you the right to work closely with customers and enabling you to define the market.<\/li><li id=\"8f05\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Ecosystem<\/strong> moats come in many flavors\u200a\u2014\u200afrom becoming the default hosting site for user-generated video content (YouTube), to the rapid, developer-driven free adoption of container technologies and growth of associated tooling (Docker), to the developers that now make &gt;$20B a year through the Apple app store.<\/li><li id=\"21aa\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Networks<\/strong>. Social networks are now pretty well-understood. Marketplaces and on-demand businesses reach a liquidity point (even a local one) and build mindshare with consumers that becomes hard to displace. Often, marketplaces have some advantaged acquisition strategy for at least one side (demand or supply)\u200a\u2014\u200athey\u2019ve built a community, offered a compelling new service, have organic or viral spread.<\/li><li id=\"a36b\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Product leaders<\/strong>. In the enterprise, cloud pioneers like Salesforce, Workday and ServiceNow built complete, core workflow solutions with a new architecture. Their faster-innovating products, (at that time novel) SaaS consumption model, and dominant go-to-market machines have made them category leaders. Each is now expanding their initial moats to become platforms for other applications and workflows.<\/li><li id=\"d92c\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Distribution<\/strong> represents a new generation of user-adopted SaaS products. End-user engagement within organizations is often a weak spot for enterprise software incumbents. Deploying software is a challenge for most organizations, and consumers who are using more software in their own lives expect better choice and usability. Simplicity and quality of design has become a moat, especially when coupled with product features that encourage adoption within an organization, or new layers of value for incremental users within an organization. Is it better for me if my team also uses a tool? What about my manager?<\/li><\/ul><p id=\"2e36\" class=\"graf graf--p graf-after--li\"><a href=\"https:\/\/news.greylock.com\/the-new-moats-53f61aeac2d9\" data-href=\"https:\/\/news.greylock.com\/the-new-moats-53f61aeac2d9\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Capturing a unique dataset is also emerging as a moat<\/a> in the age of AI-powered products. If your workflow product or MVP is useful on its own, and you can use it to collect unique data, you can then learn from that data to build a better and smarter product, improving the user experience\u200a\u2014\u200adriving a new virtuous cycle.<br \/>\u00a0<br \/>Amazon is a great example of building many moats over time\u200a\u2014\u200aa habitual and high-trust relationship with the consumer, economies of scale, systems and culture that allow management of an operationally complex business, pricing power with vendors, a marketplace of sellers, and the bundling of multiple services into a subscription business.<\/p><figure id=\"b5d5\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*tVeDn8idCHO6QnmJ1uImWA.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*tVeDn8idCHO6QnmJ1uImWA.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*tVeDn8idCHO6QnmJ1uImWA.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"fca5\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*nflDxEp64RVyXH_CkxvEOg.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*nflDxEp64RVyXH_CkxvEOg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*nflDxEp64RVyXH_CkxvEOg.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"a550\" class=\"graf graf--p graf-after--figure\">While you may not address this directly in your slides, it\u2019s worth thinking about whether you are a category creator or disruptor. If you are creating a category, focus on market risk\u200a\u2014\u200ado users really feel the pain and want this? Is there budget out there? If you are trying to take over a category, then what are the dynamics, the changes in technology or environment, the 10x better that will allow you to enter that market?<\/p><figure id=\"eff0\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*_LKOJA8zjxT0_S09-UACVg.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*_LKOJA8zjxT0_S09-UACVg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*_LKOJA8zjxT0_S09-UACVg.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"0916\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*YI-8lR5n8r4fAnNsJS7TOw.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*YI-8lR5n8r4fAnNsJS7TOw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*YI-8lR5n8r4fAnNsJS7TOw.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"adf7\" class=\"graf graf--p graf-after--figure\">At the Series A, I\u2019m betting on the early team more than anything else. What\u2019s your story? What do you believe and understand that others don\u2019t? Is there a reason you should win? We are looking for authenticity and attributes\u200a\u2014\u200apassion, focus, vision, urgency\u200a\u2014\u200arather than a specific pedigree.\u00a0<br \/>\u00a0<br \/>I also seek to invest in founders that are magnets for talent. Do customers want them to win? Will people take risks for them? No company is built by a single individual, so the ability to attract and retain a great team is key.<\/p><figure id=\"a5e0\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*-1_uOC9SgWJP-rLhqd8cdg.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*-1_uOC9SgWJP-rLhqd8cdg.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*-1_uOC9SgWJP-rLhqd8cdg.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"ced9\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*RSkM9D_6mlxAlnIwborkww.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*RSkM9D_6mlxAlnIwborkww.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*RSkM9D_6mlxAlnIwborkww.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"9654\" class=\"graf graf--p graf-after--figure\">Product-oriented founders are often at a loss for why they need to get into a financial forecast for a Series A investment, when they often haven\u2019t sold a single customer yet or acquired their first 100,000 users.<br \/>\u00a0<br \/>A financial plan tells us a lot about how you think about the business. What do you think the key drivers of your revenues and costs are? How fast is it possible for you to grow? Do you understand how other companies in your industry operate? How ambitious (and how realistic) are you? How far will you get before you need to raise more capital?<br \/>\u00a0<br \/>How do you know what assumptions to make at the Series A? If you don\u2019t know, it\u2019s ok to say so\u200a\u2014 and you\u2019re just trying to get in the right zone, no need for false precision. A great way to jumpstart your understanding of existing benchmarks for performance is to read the S-1\u2019s of relevant companies\u200a\u2014\u200aand if you\u2019re going to operate very differently, point out those differences. If I don\u2019t understand some of your plan, I may ask what tests you have done (or will do) to validate your assumptions.\u00a0<br \/>\u00a0<br \/>When you bring on a venture investor, you\u2019ll use the financial plan and other operational metrics as the shared understanding of your company\u2019s targets\u200a\u2014 and measure of its performance.<\/p><figure id=\"50be\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*VDvxZuN6KJ764lm2zYTpyQ.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*VDvxZuN6KJ764lm2zYTpyQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*VDvxZuN6KJ764lm2zYTpyQ.jpeg\" alt=\"image\" \/><\/div><\/figure><figure id=\"c3de\" class=\"graf graf--figure graf-after--figure\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*SxSXhIcr8bpyhKHk4ctQkw.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*SxSXhIcr8bpyhKHk4ctQkw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*SxSXhIcr8bpyhKHk4ctQkw.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"7923\" class=\"graf graf--p graf-after--figure\">Intellectual honesty is key\u200a\u2014\u200aboth within your team, and with your investors and partners. If you have that honesty, many of these mistakes you can look out for early and iterate to a solution. Here are some of the ones we\u2019ve learned to look out for.<\/p><figure id=\"4735\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*MXqjrRJF26oYI-BbWflBwA.png\" data-width=\"879\" data-height=\"494\" data-action=\"zoom\" data-action-value=\"1*MXqjrRJF26oYI-BbWflBwA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*MXqjrRJF26oYI-BbWflBwA.png\" alt=\"image\" \/><\/div><\/figure><p id=\"e340\" class=\"graf graf--p graf-after--figure\">I encourage you to use these slides to frame your thinking when you\u2019re drafting an investor pitch, and hope it demystifies the process. But, I would hate for the patterns of the past to cause me to miss the outlier company that matters\u200a\u2014\u200aso please, give feedback and push my thinking.<br \/>\u00a0<br \/>One last thought\u2026<\/p><figure id=\"173b\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*vQ0vliquXppdbeKZxyyTtw.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*vQ0vliquXppdbeKZxyyTtw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*vQ0vliquXppdbeKZxyyTtw.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"7cea\" class=\"graf graf--p graf-after--figure\">A lot of smart investors have been talking about how we\u2019re at the tail end of a secular technology cycle. They complain that we can no longer rely on smartphone growth to drive mobile app adoption, Web 3.0 has matured, new consumer social products can\u2019t break out against seemingly unassailable incumbents, the enterprise shift to SaaS is mostly done, all of commerce belongs to Amazon, and being caught between cloud and open source means few new great infrastructure businesses will get built.<br \/>\u00a0<br \/>I think this is at odds with the accelerating pace of technology-enabled disruption. The average lifespan of a company in the S&amp;P 500 has shrunk from 67 years in the 1920s to 15 years today.<\/p><p id=\"a8ab\" class=\"graf graf--p graf-after--p\">Founders with great ideas will continue to make novel use of our existing key technology platforms. Founders are just beginning to connect capabilities in AI research to real product uses, experiment with emergent interaction modes like AR, and harness consumer electronics supply chains to get to outer space. Entrepreneurs are growing meat in labs, collecting new data in every industry, designing new therapies. They are serving huge communities that think and interact totally differently\u200a\u2014\u200agamers and post-Millenials who are growing up internet-native. We see this every day.<\/p><p id=\"75fa\" class=\"graf graf--p graf-after--p\">The entrepreneurial ecosystem has also never been richer\u200a\u2014\u200aof capital, of on-demand infrastructure resources, of open source code and APIs, of founders and investors offering help and advice, of once-tribal startup knowledge on the web, and most importantly\u200a\u2014\u200aof talent willing to take risks. I think underdog founders have never had a better shot, and I look forward to hearing great ideas from some of you.<\/p><figure id=\"465e\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*2l4iAMuNOl-EsOv4J89-TQ.jpeg\" data-width=\"2800\" data-height=\"1575\" data-action=\"zoom\" data-action-value=\"1*2l4iAMuNOl-EsOv4J89-TQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*2l4iAMuNOl-EsOv4J89-TQ.jpeg\" alt=\"image\" \/><\/div><\/figure>","created_at":"2018-11-07T16:52:24+0000","updated_at":"2018-11-07T16:52:24+0000","published_at":"2017-08-15T15:59:27+0000","published_by":["Sarah Guo"],"annotations":[],"mimetype":"text\/html","reading_time":15,"domain_name":"news.greylock.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*hEqSyP88GY3w-LFysHHQxA.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":574,"title":"Standard Ebooks","url":"https:\/\/standardebooks.org","content":"<section class=\"splash\"><p>Free and liberated ebooks,<br \/>carefully produced for the true book lover.<\/p><\/section><section class=\"different\"><h2>What makes Standard Ebooks different?<\/h2><section class=\"features\"><h2>Modern &amp; consistent typography<\/h2><div><div><p>Other free ebooks don\u2019t put much effort into professional-quality typography: they use \"straight\" quotes instead of \u201ccurly\u201d quotes, they ignore details like em- and en-dashes, and they look more like early-90\u2019s web pages instead of actual books.<\/p><p>The Standard Ebooks project applies a rigorous and modern <a href=\"https:\/\/standardebooks.org\/contribute\/typography\">typography manual<\/a> when developing each and every ebook to ensure they meet a professional-grade and consistent typographical standard. Our ebooks look <em>good<\/em>.<\/p><\/div><figure class=\"stacked\"><img alt=\"An example of bad typography\" class=\"bottom\" src=\"https:\/\/standardebooks.org\/images\/typography-bad.png\" \/><img alt=\"An example of Standard Ebooks typography\" class=\"top\" src=\"https:\/\/standardebooks.org\/images\/typography-good.png\" \/><img alt=\"An arrow pointing from bad typography to good typography\" class=\"arrow\" src=\"https:\/\/standardebooks.org\/images\/arrow-down.png\" \/><\/figure><\/div><\/section><section class=\"features\"><h2>Full proofing with careful corrections<\/h2><div><div><p>Transcriptions from other sources are often filled with typos or suffer from issues like inconsistent spelling, missing accent marks, or missing punctuation. Submitting corrections to such sources can be difficult or impossible, so errors are rarely fixed.<\/p><p>At Standard Ebooks, we do a careful and complete readthrough of each ebook before releasing it, checking it against a scan of the original pages to fix as many typos as possible. Even if we <em>do<\/em> miss something, our ebooks are stored in the hugely popular Git source control system, allowing anyone to easily submit a correction.<\/p><\/div><figure><img alt=\"A text with proofreader's marks\" src=\"https:\/\/standardebooks.org\/images\/proofreading.svg\" \/><\/figure><\/div><\/section><section class=\"features\"><h2>Rich &amp; detailed metadata<\/h2><div><p>Our ebooks include complete, well-researched, and consistent metadata, including original, detailed book blurbs and links to encyclopedia sources. Perfect for machine processing or for extra-curious, technically-minded readers.<\/p><figure><img alt=\"An application dialog displaying ebook metadata\" class=\"top\" src=\"https:\/\/standardebooks.org\/images\/metadata.png\" \/><\/figure><\/div><\/section><section class=\"features\"><h2>State-of-the-art technology<\/h2><div><div><p>Each Standard Ebook takes full advantage of the latest ereader technology, including:<\/p><ul><li>\n\t\t\t\t\t\t\t\t<p>Hyphenation support,<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t<li>\n\t\t\t\t\t\t\t\t<p>Popup footnotes,<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t<li>\n\t\t\t\t\t\t\t\t<p>High-resolution and scalable vector graphics,<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t<li>\n\t\t\t\t\t\t\t\t<p>Ereader-compatible tables of contents,<\/p>\n\t\t\t\t\t\t\t<\/li>\n\t\t\t\t\t\t<\/ul><p>and more. One of our goals is to ensure our ebooks stay up-to-date with the best reading experience technology can provide. Just because it\u2019s a classic doesn\u2019t mean it has to use old technology.<\/p><\/div><figure class=\"stacked\"><img alt=\"An ebookshelf featuring Standard Ebooks covers\" class=\"bottom\" src=\"https:\/\/standardebooks.org\/images\/endnote.png\" \/><img alt=\"An ebookshelf featuring Standard Ebooks covers\" class=\"top\" src=\"https:\/\/standardebooks.org\/images\/toc.png\" \/><\/figure><\/div><\/section><section class=\"features\"><h2>Quality covers<\/h2><div><div><p>Everyone knows a book is judged by its cover, but most free ebooks leave it to your ereader software to generate a drab default cover.<\/p><p>Standard Ebooks draws from a vast collection of public domain fine art to create attractive, unique, appropriate, and consistent covers for each of our ebooks.<\/p><\/div><figure><img alt=\"An ebookshelf featuring Standard Ebooks covers\" src=\"https:\/\/standardebooks.org\/images\/covers.png\" \/><\/figure><\/div><\/section><section class=\"features\"><h2>Clean code &amp; semantic markup<\/h2><div><div><p>Our strict coding standards allow technologists and ebook producers to use Standard Ebooks files as reliable, easy to read, and robust bases for their own work\u2014not to mention as models of what well-crafted ebook files look like. Common code patterns are repeated through different ebooks, so the code never surprises you.<\/p><p>Each ebook is also enhanced with careful standards-based semantic markup that opens the gateway for exciting new kinds of machine processing.<\/p><\/div><figure><img alt=\"A screenshot of the code of a Standard Ebook file\" src=\"https:\/\/standardebooks.org\/images\/code.png\" \/><\/figure><\/div><\/section><section class=\"features\"><h2>Free, open-source, &amp; public domain<\/h2><div><div><p>We use the popular Git source control system to track each and every change made to our ebooks. Anyone can easily see a history of changes, or contribute their own changes with the click of a mouse.<\/p><p>And while all of the ebooks we feature and the cover art we draw from are <em>already<\/em> believed to be in the public domain in the US, Standard Ebooks releases all of the work we put in to each ebook into the public domain too. That makes each and every one of our ebook files not just free, but <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gratis_versus_libre\">libre<\/a> too\u2014because the world deserves more unrestricted culture.<\/p><\/div><figure class=\"oss\"><img alt=\"The Git SCM logo\" src=\"https:\/\/standardebooks.org\/images\/git.svg\" \/><img alt=\"The no-copyright symbol\" src=\"https:\/\/standardebooks.org\/images\/no-copyright.svg\" \/><img alt=\"The no-copyright symbol\" src=\"https:\/\/standardebooks.org\/images\/no-drm.svg\" \/><\/figure><\/div><\/section><\/section>","created_at":"2018-11-07T16:52:24+0000","updated_at":"2018-11-07T16:52:24+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"standardebooks.org","preview_picture":"https:\/\/standardebooks.org\/images\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":575,"title":"Baen Ebooks","url":"https:\/\/www.baen.com\/baenebooks","content":"<div style=\"text-align: right;\"><img src=\"https:\/\/www.baen.com\/images\/ebook-icon.jpg\" alt=\"image\" \/><p style=\"text-align: left;\"><b>Welcome to Baen Ebooks,<\/b> purveyors of DRM-free ebooks.  This means you can enjoy your ebooks on as many eReaders as you choose for no additional cost.<\/p><p>Browse this month's new releases below: <b><a href=\"#eARC\">eARC<\/a>, <a href=\"#Ebook\">Ebook<\/a>, <a href=\"#GVG\">Grantville Gazette<\/a>.<\/b> (Want to know what an eARC is?  <a href=\"https:\/\/www.baen.com\/t-faq?tab=earc\">click here<\/a> to find out.)<\/p><p style=\"text-align: left;\">As of December 16th, 2012, Baen ebooks are now available for sale at other vendors as well as at this site. At the same time, prices for our individual ebooks were raised to meet the rest of the market. Titles available in print as hardcovers only will be $9.99, trade paperbacks $8.99, and mass market paperbacks $6.99. As part of the adjustment, we have also raised ebook royalties to authors by 25%.<\/p><p style=\"text-align: left;\">Available exclusively at Baen.com are <a href=\"https:\/\/www.baen.com\/t-faq?tab=earc\">eARCs<\/a> and <a href=\"https:\/\/www.baen.com\/t-faq?tab=bundles\">serialized monthly bundles<\/a> at our previous prices. There is a significant change from our previous practice in the availability of the bundles: once the books for a given month are printed and the books made available individually at other sellers, these bundles will not be available for orders.<\/p><p style=\"text-align: left;\">Monthly Baen Bundles are usually available for purchase through the first Monday of the month. However, when the first Tuesday falls on the 1st, the cutoff day is the last Monday of the previous month. Individual titles will be released for sale on the first Tuesday of each month.<\/p><p style=\"text-align: left;\">Happy reading!<\/p><\/div><br \/><h3 style=\"text-align: center;\"><img alt=\"Cover\" id=\"_cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483506.jpg\" height=\"240\" style=\"float: right; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"http:\/\/www.baen.com\/uncompromising-honor-earc.html\">Uncompromising Honor - eARC<\/a><\/b><br \/>by David Weber<\/h3><p>HONOR KEEPS HER PROMISE<\/p><p>The Solarian League\u2014for hundreds of years they have borne the banner of human civilization. But the bureaucratic Mandarins who rule today's League are corrupt and looking for scapegoats. They've decided the upstart Star Kingdom of Manticore must be annihilated.<\/p><p>UNCOMPROMISING COURAGE<\/p><p>Honor Harrington has worn the Star Kingdom's uniform for half a century. Very few know war the way Honor Harrington does. So far, hers has been a voice of caution. But now the Mandarins have committed atrocities such as the galaxy has not known in a thousand years. They have finally killed too many of the people Honor Harrington loves.<\/p><p>UNCOMPROMISING VENGEANCE<\/p><p>Now Honor Harrington is coming for the Solarian League. And Hell is riding in her wake.<\/p><p>UNCOMPROMISING HONOR<\/p><p style=\"text-align: center;\"><a href=\"http:\/\/www.baen.com\/uncompromising-honor-earc.html\"><b>Order Now<\/b><\/a><\/p><h3 style=\"text-align: center;\"><img alt=\"Cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483438.jpg\" height=\"240\" style=\"float: left; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"https:\/\/www.baen.com\/black-triumph-earc.html\">Black Triumph - eARC<\/a><\/b><br \/>by Brendan DuBois<\/h3><p><b>The conclusion to Brendan DuBois's Dark Victory alien-invasion trilogy.<\/b> Bestselling and award-winning author Brendan DuBois proves that even in the most desperate of times, there are still men and women alive who will take up weapons to defend what is theirs.<\/p><p>More than ten years ago, the alien Creepers arrived in Earth\u2019s orbit and started a war that was a slaughter for humanity. Nuclear weapons detonated in the atmosphere destroyed all electronic devices, asteroids dropped into oceans and lakes swamped cities with artificial tsunamis, and the nearly invulnerable Creepers arrived on Earth, going forth from their dome bases to attack civilians and military units at will.<\/p><p>Sixteen-year-old Randy Knox, a newly minted lieutenant in the U.S. Army, has been fighting the Creepers since he was twelve. He has seen friends, family members and fellow soldiers killed by the Creepers, and he is tired of war. At one point, it seemed the war was over when the aliens\u2019 orbital battle station had been destroyed.<\/p><p>But a second Creeper orbital battle station has arrived.<\/p><p>The war no longer seems to be over, and while returning to his home unit, Randy\u2019s convoy is ambushed. Separated from his fellow soldiers and his K-9 companion Thor, Randy faces the ultimate horror of every American serviceman: to become a prisoner of war of the aliens.<\/p><p style=\"text-align: center;\"><a href=\"https:\/\/www.baen.com\/black-triumph-earc.html\"><b>Order Now<\/b><\/a><\/p><h3 style=\"text-align: center;\"><img alt=\"Cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483421.jpg\" height=\"240\" style=\"float: right; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"https:\/\/www.baen.com\/her-majesty-american-earc.html\">Her Majesty American - eARC<\/a><\/b><br \/>by Steve White<\/h3><p>In an alternate future where the British empire never crumbled, the space ships of Her Majesty\u2019s Navy work to keep the spaceways safe. Commander Robert Rogers of the Royal Space Navy is a defense intelligence officer known for his unconventional ways. Rogers is American born\u2011and\u2011raised, a loyal citizen of the Viceroyalty of North America. He\u2019s a descendant of the founder of the highly effective, pro\u2011British Rogers\u2019 Rangers during the Colonial unpleasantries of 1776, and he\u2019s inherited his ancestor\u2019s intelligence and drive in spades.<\/p><p>But there are those among the stars who are not so happy being subjects of the British Empire. In the Tau Ceti system, a cauldron of trouble brews as a terrorist faction of the rebellious Sons of Arnold attacks the empire from within, and warships of the theocratic Caliphate enters the system, prepared to do their worse to destroy the hated empire head on. Yet standing against the coming anarchy and tyranny is one intrepid spy prepared to risk all for queen and empire:<\/p><p>COMMANDER ROBERT ROGERS, HER MAJESTY\u2019S AMERICAN<\/p><p style=\"text-align: center;\"><a href=\"https:\/\/www.baen.com\/her-majesty-american-earc.html\"><b>Order Now<\/b><\/a><\/p><h3 style=\"text-align: center;\"><img alt=\"Cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483445.jpg\" height=\"240\" style=\"float: left; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"https:\/\/www.baen.com\/target-rich-environment-earc.html\">Target Rich Environment - eARC<\/a><\/b><br \/>by Larry Correia<\/h3><p>Larry Correia, the <i>New York Times<\/i> best-selling author of the wildly popular Monster Hunter International and Grimnoir Chronicles series of no-holds-barred urban fantasy and the Dead Six trilogy of military adventure novels, now sets his sights on the short story form.<\/p><p>Together for the first time, fourteen action-packed tales of demons, monster, vampires, and cosmic horrors too terrible to name\u2014and the men and women who take them all down. Oh, and toss in an interdimensional insurance salesman for good measure.<\/p><p>You\u2019ll also find: An elven princess from the pages of Monster Hunter International on a mission to redeem her people. A samurai pirate with a blood vendetta against an extremely large sea beast. And a magic-wielding P.I. who walks the mean streets of Detroit.<\/p><p>Journey back to the origins of Monster Hunter International in \u201cBubba Shackleford\u2019s Professional Monster Hunters.\u201d Uncover the origin of one of Dead Six\u2019s central characters in \u201cSweothi City.\u201d And celebrate the holidays with the Grimnoir Chronicle\u2019s own Jake Sullivan in \u201cDetroit Christmas.\u201d<\/p><p>Discover why it always pays to hire the very best when it comes to fighting a war when Correia takes on David Drake\u2019s genre-defining Hammer\u2019s Slammers series. And find out just how much damage a vampire can do when he was a Green Beret before he became one of the undead as Correia tackles Jonathan Maberry\u2019s V Wars universe.<\/p><p>Includes three stories never before in print! \u201cMurder on the Orient Elite\u201d and \u201cThe Adventures of Tom Strange, Interdimensional Insurance Salesman\u201d have previously only been available in audio format (narrated by Adam Baldwin of <i>Firefly<\/i> fame). And \u201cBlood on the Water\u201d makes its first appearance.<\/p><p style=\"text-align: center;\"><a href=\"https:\/\/www.baen.com\/target-rich-environment-earc.html\"><b>Order Now<\/b><\/a><\/p><h3 style=\"text-align: center;\"><img alt=\"Cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483223.jpg\" height=\"240\" style=\"float: right; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"https:\/\/www.baen.com\/avalanche-earc.html\">Avalanche - eARC<\/a><\/b><br \/>by Mercedes Lackey, Cody Martin, Dennis Lee, &amp; Veronica Giguere<\/h3><p>Ultima Thule has been destroyed\u2014but somehow the Thulians mounted an even bigger force to destroy Metis. The Metisians that escaped the carnage and destruction of their secret city now must somehow find somewhere safe to go\u2014without getting snapped up by various world governments. And now the Thulians have changed their tactics to \u201cblitz guerilla warfare,\u201d sending wave after wave of their ships and troops to attack vulnerable targets all over Earth, with no clue for ECHO as to where these things are coming from.<\/p><p>John Murdock and The Seraphym have become a force of nature, but they can\u2019t be everywhere at once, and if they were regarded with suspicion before, now they are regarded by ECHO\u2019s allies with fear as well.<\/p><p>You could say \u201cthings can\u2019t get worse.\u201d<\/p><p>You would be wrong.<\/p><p>Red Djinni\u2019s past is about to catch up with him.<\/p><p>So is Victoria Victrix\u2019s.<\/p><p style=\"text-align: center;\"><a href=\"https:\/\/www.baen.com\/avalanche-earc.html\"><b>Order Now<\/b><\/a><\/p><h3 style=\"text-align: center;\"><img alt=\"Cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483414.jpg\" height=\"240\" style=\"float: left; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"http:\/\/www.baen.com\/worlds-2-earc.html\">Worlds II - eARC<\/a><\/b><br \/>by Eric Flint<\/h3><p>WELCOME TO THE MANY WORLDS OF BEST-SELLING AUTHOR ERIC FLINT<\/p><p>Known for his <i>New York Times<\/i> best-selling alternate history novels, Flint is equally a master of shorter forms, and this second large volume gathers more of Flint\u2019s shorter works. Including several stories and a short novel set in Flint\u2019s celebrated Ring of Fire alternate history series. A commemorative novelette set in Poul Anderson\u2019s <i>Operation Chaos<\/i> universe. A story with Dave Freer, set in his and Flint\u2019s popular Rats, Bats and Vats series. Several humorous short stories addressing Biblical literalism. A long novella in a Martian steampunk setting. And another long novella set in David Weber\u2019s best-selling Honor Harrington universe.<\/p><p>In addition to the fiction, Eric Flint has written an overall introduction, plus an introduction for each story, telling how it came to be written, making this an irresistible book for the many thousands of Eric Flint fans.<\/p><p style=\"text-align: center;\"><a href=\"http:\/\/www.baen.com\/worlds-2-earc.html\"><b>Order Now<\/b><\/a><\/p><h3 style=\"text-align: center;\"><img alt=\"Cover\" src=\"https:\/\/www.baen.com\/images\/product\/medium\/A9781481483407.jpg\" height=\"240\" style=\"float: right; margin-left: 16px; margin-right: 16px;\" \/><b><a href=\"http:\/\/www.baen.com\/alternate-routes-earc.html\">Alternate Routes - eARC<\/a><\/b><br \/>by Tim Powers<\/h3><p>Something weird is happening to the Los Angeles freeways\u2014phantom cars, lanes from nowhere, and sometimes unmarked offramps that give glimpses of a desolate desert highway\u2014and Sebastian Vickery, disgraced ex-Secret Service agent, is a driver for a covert supernatural-evasion car service. But another government agency is using and perhaps causing the freeway anomalies, and their chief is determined to have Vickery killed because of something he learned years ago at a halted Presidential motorcade.<\/p><p>Reluctantly aided by Ingrid Castine, a member of that agency, and a homeless Mexican boy, and a woman who makes her living costumed as Supergirl on the sidewalk in front of the Chinese Theater, Vickery learns what legendary hell it is that the desert highway leads to\u2014and when Castine deliberately drives into it to save him from capture, he must enter it himself to get her out.<\/p><p><i>Alternate Routes<\/i> is a fast-paced supernatural adventure story that sweeps from the sun-blinded streets and labyrinthine freeways of Los Angeles to a horrifying other world out of Greek mythology, and Vickery and Castine must learn to abandon old loyalties and learn loyalty to each other in order to survive as the world goes mad around them.<\/p><p style=\"text-align: center;\"><a href=\"http:\/\/www.baen.com\/alternate-routes-earc.html\"><b>Order Now<\/b><\/a><\/p><br \/><div><a href=\"http:\/\/www.baen.com\/grantville-gazette-volume-78.html\"><img src=\"https:\/\/www.baen.com\/images\/product\/medium\/9781625796592.jpg\" alt=\"image\" \/><\/a><\/div><p><b><a href=\"http:\/\/www.baen.com\/grantville-gazette-volume-78.html\">Grantville Gazette Volume 78<\/a><br \/>&#13;\ncreated by Eric Flint<br \/>&#13;\nedited by Walt Boyes and Bjorn Hasseler<\/b><\/p><div><p>Grantville Gazette 78 opens with \"Goodbye\" by Thomas K. Scot, about a journeyman setting out on his working travels.<\/p><p>\"The Observer\" by Bjorn Hasseler explains how the first down-time profiler got started.<\/p><p>Michael Lockwood's \"The Freeman's Oath\" is set in one of the North American colonies that passed to the control of France. Can a freeman swear allegiance?<\/p><p>\"Cremonensis Faciebat 1632, or Body and Scroll: The Luthier's Tale\" by Tim Sayeau is about a violin maker who hears about Grantville and has to check it out.<\/p><p>\"The Visible Dogs of Grantville\" by Jackie Britton Lopatin picks up from her earlier \"The Invisible Dogs of Grantville\" and addresses the dilemma of what happens when Mom cleans out the house.<\/p><p>In this issue's \"Notes from The Buffer Zone: Conventions in the Sky,\" Kristine Kathryn Rusch remembers writers who have passed on.<\/p><p>A new column debuts: \"Hot Off Eric Flint's Ring of Fire Press!\" Here are the latest 1632 and non-1632 titles coming out.<\/p><p>In the Universe Annex, Mike Watson brings us \"The Enigma of Charlie Peabody.\" Who is this man?<\/p><p style=\"text-align: center;\"><a href=\"http:\/\/www.baen.com\/grantville-gazette-volume-78.html\"><b>Order Now<\/b><\/a><\/p><br \/><\/div>","created_at":"2018-11-07T16:52:24+0000","updated_at":"2018-11-07T16:52:24+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"www.baen.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":572,"title":"11 Open Source Tools for Writers","url":"https:\/\/itsfoss.com\/open-source-tools-writers\/","content":"<p><img class=\"aligncenter size-large wp-image-17696 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers-800x450.jpg?is-pending-load=1\" alt=\"Open Source softwares for writers\" width=\"800\" height=\"450\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers-768x432.jpg 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17696\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers-800x450.jpg\" alt=\"Open Source softwares for writers\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers-768x432.jpg 768w\" \/><\/noscript><p>The open source community produces a large amount of software for different uses. I have already told you about <a href=\"https:\/\/itsfoss.com\/create-interactive-fiction\/\" target=\"_blank\" rel=\"noopener\">open source tools for interactive fictions<\/a>. Here are eleven open source tools to help authors be creative.<\/p>\n<p>Before we begin, I would like to briefly explain why open source is important. When we think of the software we use to write, most people think of programs written by big corporations like Microsoft Word Scrivener. These programs cost money and are built by large teams of programmers. At anytime these companies and these products could go away and not be available anymore.<\/p>\n<p>Open source programs are a little different. The vast majority are free. The code used to create them is freely available, meaning that if the original developer stops work on his project, someone else can take it up. It also means that if you have some coding knowledge, you too can contribute to the project. Open source developers typically respond much quicker to their users than huge multinational organizations.<\/p>\n<p>Now, on with the list.<\/p>\n<h3 id=\"bibisco\">1. Bibisco<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17724 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1-800x200.png?is-pending-load=1\" alt=\"Bibisco open source tool for writers\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17724\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1-800x200.png\" alt=\"Bibisco open source tool for writers\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-1-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"http:\/\/www.bibisco.com\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Bibisco<\/a> is an application designed to help you write stories, mainly novels. Where it shines is in character creation. Bibisco asks you a series of questions about each character in your story. The questions will help you create a solid idea of what your character looks like, what their motives are and what their background is. It also has a place to store images that help you create a mental picture of your characters.<\/p>\n<p>Bibisco also comes with an interesting analysis feature that allows you to see at a glance what characters and what locations appeared in different chapters. It includes a decent look text editor that has basic formatting features<\/p>\n<p>Bibisco is released under GPL. It is available for Linux, Windows, and Mac.<\/p>\n<h3 id=\"manuskript\">2. Manuskript<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17723 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2-800x200.png?is-pending-load=1\" alt=\"open source tool for writers Manuskript\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17723\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2-800x200.png\" alt=\"open source tool for writers Manuskript\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-2-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"http:\/\/www.theologeek.ch\/manuskript\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Manuskript<\/a> is another novel creation tool. This application focuses on outlining. By looking at the detailed outline tool, you can see what stage each chapter is at and what characters are involved. You can also easily rearrange chapters. It uses the <a href=\"http:\/\/www.advancedfictionwriting.com\/articles\/snowflake-method\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">snowflake method<\/a> to help you build your novel.<\/p>\n<p>Manuskript includes a frequency analyzer, so you can see which words or phrases you repeat and how often. They even included a distraction free writing mode.<\/p>\n<p>Manuskript is released under GPL v3. It is available for Linux, Windows, and Mac.<\/p>\n<h3 id=\"ostorybook\">3. oStorybook<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17722 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3-800x200.png?is-pending-load=1\" alt=\"open source tool for writers ostroybook\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17722\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3-800x200.png\" alt=\"open source tool for writers ostroybook\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-3-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"http:\/\/ostorybook.tuxfamily.org\/index.php?lng=en\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">oStorybook<\/a> is yet another tool intended to help you create novels. The goal of this program is to organize the different elements of your story, so you can focus on writing. It includes a hierarchical tree, so you can see how all characters and events are related. The program also features a spell check and a task list. Like the other novel creation apps, you can create reports to see how often characters appear and when.<\/p>\n<p>It is released under GPL. oStorybook is built with Java, so it will run on Linux, Windows, and Mac<\/p>\n<h3 id=\"gitbook\">4. GitBook<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17721 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4-800x200.png?is-pending-load=1\" alt=\"open source tool for writers Gitbook\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17721\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4-800x200.png\" alt=\"open source tool for writers Gitbook\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-4-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"https:\/\/www.gitbook.com\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">GitBook<\/a> is a service mainly used for technical writing, but I don\u2019t see why it would not work for a fiction writer. GitBook makes use of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Git\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">git<\/a> version control system to keep track of changes in the document you are writing. It also enables several users to collaborate on a book.<\/p>\n<p>You can either choose a free account or pay $7 a month. If you choose a free account, all your work will be publicly available as you write. The paid account gives you the ability to create a private book.<\/p>\n<p>The GitBook Editor is an application that allows you to write your documents on your computer using Markdown or Asciidoc. (More on those later.) It is available for Linux, Windows, and Mac.<\/p>\n<h3 id=\"trelby\">5. Trelby<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17720 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5-800x200.png?is-pending-load=1\" alt=\"open source tool for writers trelby\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17720\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5-800x200.png\" alt=\"open source tool for writers trelby\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-5-768x192.png 768w\" \/><\/noscript>\n<p>If screenwriting is more your speed, then you should check out <a href=\"http:\/\/www.trelby.org\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Trelby<\/a>. Trelby is created to \u201cenforces correct script format and pagination\u201d and includes auto-completion, and spell checking. It also features scene, location, character, and dialogue reports. You can also use it to compare your script to see what changed between versions. You can import files from quite a few screenwriting apps including: Final Draft XML (.fdx), Celtx (.celtx), Fountain (.fountain), Adobe Story (.astx) and Fade In Pro (.fadein). You can also export HTML, RTF, Final Draft XML (.fdx) and Fountain (.fountain).<\/p>\n<p>Trelby is available under the GPL license. It runs on Linux and Windows.<\/p>\n<h3 id=\"ghostwriter\">6. ghostwriter<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17719 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6-800x200.png?is-pending-load=1\" alt=\"open source tool for writers ghostwriter\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17719\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6-800x200.png\" alt=\"open source tool for writers ghostwriter\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-6-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"http:\/\/wereturtle.github.io\/ghostwriter\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Ghostwriter<\/a> is a personal favorite of mine. (In fact, I do all my writing in it.) This application allows you to use the Markdown language to create documents. It is distraction-free by design. It can export to HTML, Word, ODT, PDF, Epub, and more. One nice feature is that it will convert headings to chapters if you export to Epub.<\/p>\n<p>Ghostwriter is licensed under GPL v3. It can run on Linux and Windows.<\/p>\n<h3 id=\"scribus\">7. Scribus<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17718 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7-800x200.png?is-pending-load=1\" alt=\"open source tool for writers scribus\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17718\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7-800x200.png\" alt=\"open source tool for writers scribus\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-7-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"https:\/\/www.scribus.net\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Scribus<\/a> is a free and open source desktop publishing application. While it\u2019s not designed to help you write the next big novel, you can use Scribus to layout the finished book. It has support for \u201cpowerful vector drawing tools, support for a huge number of file types via import\/export filters, emulation of color blindness or the rendering of markup languages like LaTeX or Lilypond.\u201d<\/p>\n<p>Scribus can run on Linux, FreeBSD, PC-BSD, NetBSD, OpenBSD, Solaris, OpenIndiana, Debian GNU\/Hurd, Mac OS X, OS\/2 Warp 4, eComStation, Haiku, and Windows. It is released under GPL.<\/p>\n<h3 id=\"markdown\">8. Markdown<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17726 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown-800x200.png?is-pending-load=1\" alt=\"Open Source tools for writers Markdown\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17726\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown-800x200.png\" alt=\"Open Source tools for writers Markdown\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-markdown-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"https:\/\/daringfireball.net\/projects\/markdown\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Markdown<\/a> is more than a tool. It is also a formatting syntax. This means that you can write your stories in a plain text editor and then convert it to any document format. All you have to do is add a <a href=\"https:\/\/daringfireball.net\/projects\/markdown\/basics\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">series of symbols<\/a> through out the document. This is great because it prevents you from being locked into one program or file format only to see it die.<\/p>\n<p>In order to use Markdown, you typically need to use a text editor that supports Markdown. Ghostwrite from above is my favorite. Markdown is released under a <a href=\"https:\/\/daringfireball.net\/projects\/markdown\/license\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">custom license<\/a>.<\/p>\n<h3 id=\"asciidoc\">9. AsciiDoc<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17716 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9-800x200.png?is-pending-load=1\" alt=\"open source tool for writers asciidoc\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17716\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9-800x200.png\" alt=\"open source tool for writers asciidoc\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-9-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"http:\/\/www.methods.co.nz\/asciidoc\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">AsciiDoc<\/a> is another document formatting syntax. While markdown is limited to basic formatting, AsciiDoc has support for more option, such as footnotes, tables, cross references, embedded YouTube videos, and more. It can be used in the creation of notes, documentation, articles, books, ebooks, slideshows, web pages, man pages, and blogs. AsciiDoc files can be converted to HTML, PDF, EPUB, and man pages. <a href=\"http:\/\/asciidocfx.com\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">AsciiDocFX<\/a> is a good place to start.<\/p>\n<p>It is released under GPL v2.<\/p>\n<h3 id=\"fountain\">10. Fountain<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17715 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10-800x200.png?is-pending-load=1\" alt=\"open source tool for writers\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17715\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10-800x200.png\" alt=\"open source tool for writers\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-10-768x192.png 768w\" \/><\/noscript>\n<p>While Markdown and AsciiDoc can be used to create a wide range of documents, <a href=\"https:\/\/fountain.io\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Fountain<\/a> is much more specialized. It is designed for one purpose, to create screenplays. The beauty of Fountain is that it allows you to add the correct formatting to your screenplay as you write without having to take your fingers off the keeps. After a little practice, it can become second nature. <a href=\"https:\/\/fountain.io\/apps\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Here<\/a> is a list of apps that support Fountain.<\/p>\n<p>Fountain is <a href=\"https:\/\/github.com\/nyousefi\/Fountain\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">released<\/a> under the MIT license.<\/p>\n<h3 id=\"latex\">11. LaTeX<\/h3>\n<p><img class=\"aligncenter size-large wp-image-17714 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11-800x200.png?is-pending-load=1\" alt=\"Latex open source tool for writers\" width=\"800\" height=\"200\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11-768x192.png 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-large wp-image-17714\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11-800x200.png\" alt=\"Latex open source tool for writers\" width=\"800\" height=\"200\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11.png 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11-300x75.png 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-writers-tools-11-768x192.png 768w\" \/><\/noscript>\n<p><a href=\"http:\/\/www.latex-project.org\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">LaTeX<\/a> is a human readable document preparation system. While this system was created for scientific papers, it can be used to create beautifully formatted books. You use a series of markup cues to set the structure of your document and also add citations and cross-references. The end product can be converted into a number of file formats. There are several <a href=\"https:\/\/itsfoss.com\/latex-editors-linux\/\" target=\"_blank\" rel=\"noopener\">LaTeX editors for Linux<\/a> and other platforms that you can use. Just to test LaTeX, I suggest give <a href=\"http:\/\/www.lyx.org\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Lyx<\/a> a spin.<\/p>\n<p>LaTeX is released under the <a href=\"http:\/\/www.latex-project.org\/lppl\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">LaTeX Project Public License<\/a>.<\/p>\n<h3 id=\"continue-the-discussion\">Continue the Discussion<\/h3>\n<p>If you are interested you should also read <a href=\"https:\/\/itsfoss.com\/create-ebook-calibre-linux\/\" target=\"_blank\" rel=\"noopener\">how to create ebooks in Linux with Calibre<\/a>.<\/p>\n<p>I have created a <a href=\"https:\/\/www.facebook.com\/groups\/1255256097926288\/\" target=\"_blank\" rel=\"noopener nofollow\" class=\"external\">Facebook group<\/a> to help writers who use open-source tools to meet and share tips. It is also a great place for writers, who don\u2019t know about open source to learn from those who do. If you are a writer, we would be glad to have you join us.<\/p>\n<p>Have you ever used one of the tools on this list? What is your favorite open-source writing tool? Please let us know in the comments below.<\/p>\n<p>If you found this article interesting, please take a minute to share it on social media.<\/p>\n\t\t\t&#13;","created_at":"2018-11-07T16:52:23+0000","updated_at":"2018-11-07T16:52:23+0000","published_at":"2017-08-15T19:15:00+0000","published_by":["John Paul"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"itsfoss.com","preview_picture":"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/open-source-tools-for-writers.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":571,"title":"How Command Line Parameters Are Parsed","url":"http:\/\/daviddeley.com\/autohotkey\/parameters\/parameters.htm","content":"How Command Line Parameters Are Parsed\n<div style=\"text-align: center;\">\n<h3>by<\/h3>\n\u00a9 2009\n<br \/><span>(Updated 2014)<\/span>\n<br \/><span>(Updated 2016)<\/span>\n<p>How to escape the special characters in your parameters\n<\/p><\/div>\n<div class=\"tableofcontents\"><a href=\"#LINUX\">*NIX<\/a><span>(updated Feb. 2011)<a href=\"#R19\" class=\"footnote19\"><sup>19<\/sup><\/a><\/span><ol><li><a href=\"#LINUXCREATE\">How a New Process is Created on *nix (Unix, Linux, Ubuntu, Kubuntu, ...)<\/a><\/li>\n      <li><a href=\"#LINUXPASS\">Passing Parameters to a new process on *nix<\/a>\n      <ol><li><a href=\"#BASH\">Example 1: Bash Shell<\/a><\/li>\n        <li><a href=\"#UBUNTU\">Example 2: Ubuntu Desktop Launcher Icon (GNOME\u2192Nautilus)<\/a><\/li>\n        <li><a href=\"#KUBUNTU\">Example 3: Kubuntu (KDE\u2192Plasma)<\/a><\/li>\n      <\/ol><\/li>\n      <li><a href=\"#LINUXRULES\">The *nix Parameter Parsing Rules<\/a><\/li>\n    <\/ol><br \/><a href=\"#WIN\">WINDOWS<\/a><ol><li><a href=\"#WINCREATE\">How a New Process is Created on Windows<\/a><\/li>\n      <li><a href=\"#WINPASS\">Passing Parameters to a new process on Windows<\/a><\/li>\n      <li><a href=\"#WINARGV\">How does a C\/C++ program on Windows get argv[] ?<\/a><\/li>\n      <li><a href=\"#WINNOSTANDARD\">Everyone Parses Differently<\/a><\/li>\n      <li><a href=\"#WINCRULES\">The C\/C++ Parameter Parsing Rules<\/a>\n      <ol><li><a href=\"#WINCRULESCHANGE\">The rules changed in 2008<\/a><\/li>\n        <li><a href=\"#WINCRULESDOC\">The documented and undocumented rules<\/a><\/li>\n        <li><a href=\"#WINCRULESREPH\">The C\/C++ Parameter Parsing Rules Rephrased<\/a><\/li>\n        <li><a href=\"#WINCRULESSUM\">Summary of rules 5,6,7<\/a><\/li>\n        <li><a href=\"#WINCRULESEX\">Examples<\/a><\/li>\n        <li><a href=\"#WINCRULESCT\">Some Common Tasks<\/a><\/li>\n        <li><a href=\"#WINCRULESMSEX\">The Microsoft Examples Explained<\/a><\/li>\n        <li><a href=\"#WINCRULESDDQ\">Double Double Quote Examples<\/a><\/li>\n        <li><a href=\"#WINCRULESTDQ\">How Triple Double Quotes Are Parsed<\/a><\/li>\n        <li><a href=\"#WINCRULESQDQ\">How Quadruple Double Quotes Are Parsed<\/a><\/li>\n        <li><a href=\"#WINCALGORITHM\">The Microsoft C\/C++ Command Line Parsing Algorithm<\/a><\/li>\n      <\/ol><\/li>\n      <li><a href=\"#WINLINE\">Who Calls CreateProcess?<\/a>\n        <ol><li><a href=\"#WINPA\">ProgA.exe \u2192 CreateProcess()<\/a><\/li>\n          <li><a href=\"#WINCMD\">Command Prompt Window \u2192 CreateProcess()<\/a><\/li>\n        <\/ol><\/li>\n      <li><a href=\"#WINCMDRULES\">The Command Prompt Command Line Parameter Parsing Rules (cmd.exe)<\/a>\n      <br \/>Putting It Together<\/li>\n      <li><a href=\"#PIT\">How parameters are parsed to a C\/C++ Program from the Command Line<\/a>\n      <br \/>How to Pass Parameters to:\n      <ol><li><a href=\"#CPP\">a C\/C++ Program from the Command Line<\/a> <img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/new2a.gif\" width=\"28\" height=\"11\" alt=\"new!\" \/><\/li>\n        <li><a href=\"#BATCH\">a Batch File from the Command Line<\/a><\/li>\n        <li><a href=\"#WSH\">a VBScript, JScript, or WSH script from the Command Line<\/a>\n        <\/li><li><a href=\"#WSHRULES\">The Windows Script Host (WSH) Parameter Parsing Rules<\/a><\/li>\n        <li><a href=\"#WSHALGORITHM\">The WSH Command Line Parameter Parsing Algorithm<\/a><\/li>\n        <li><a href=\"#WSHRULES\">Examples<\/a><\/li>\n        <li><a href=\"#WSHRULES\">Putting it together<\/a><\/li>\n      <\/ol><\/li>\n      <li><a href=\"#PERL\">a Perl script from the Command Line<\/a><\/li>\n      <li><a href=\"#PYTHON\">a Python script from the Command Line<\/a><\/li>\n      <li><a href=\"#REXX\">a REXX script from the Command Line<\/a><\/li>\n      <li><a href=\"#RUBY\">a RUBY script from the Command Line<\/a><\/li>\n      <li><a href=\"#AHK\">an AutoHotkey script from the Command Line<\/a><\/li>\n<\/ol><\/div>\n<div style=\"text-align: center;\"><img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/bara054.gif\" alt=\"bar\" \/><\/div>\n<p>*nix<\/p>\n<p><\/p><div style=\"text-align: center;\">(Unix, Linux, Ubuntu, Kubuntu, ...)<a name=\"S19\" href=\"#R19\" class=\"footnote19\" id=\"S19\"><sup><span>19<\/span><\/sup><\/a><\/div>\n<p><a name=\"LINUXCREATE\" id=\"LINUXCREATE\">1.\u00a0\u00a0How a new process is created on *nix<\/a>\n<br \/>On *nix, you launch a program using the\n<a href=\"http:\/\/linux.about.com\/od\/commands\/l\/blcmdl2_execve.htm\">execve<\/a>\nAPI (or it's cousins execvp, execl, execlp, execle, and execvp).\nThe interesting thing about these APIs is that they allow the <i>caller<\/i>\nto specify each of the command line arguments - the signature for execve is:\n<\/p><p>\u00a0\u00a0\u00a0\u00a0\u00a0int execve(const char *filename, char *const <b>argv []<\/b>, char *const envp[]);<\/p><p><a name=\"LINUXPASS\" id=\"LINUXPASS\">2.\u00a0\u00a0Passing Parameters to a new process on *nix<\/a>\n<br \/>On *nix, the parameters are parsed off by whatever program creates the new process,\n<i>before<\/i> the new process is created.<a name=\"S1\" href=\"#R1\" class=\"footnote\" id=\"S1\"><sup>1<\/sup><\/a>\n<\/p><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table><tr><th colspan=\"6\">*nix<\/th>\n      <\/tr><tr><td class=\"code\">cmdline = \"MyProg\u00a0\u00a0Hello\u00a0\u00a0Goodbye\u00a0\u00a0Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>\n            <table><tr><td class=\"codetop\">argv[] = <\/td>\n                  <td class=\"code\">MyProg.exe\n                  <br \/>Hello\n                  <br \/>Goodbye\n                  <br \/>Friday\n                  <\/td>\n              <\/tr><\/table><\/td>\n          <td class=\"rbdr\">\u00a0<\/td>\n      <\/tr><tr><td class=\"darrow\">\u00a0\u2193 <span>create new process<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>execve(argv[])\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process running MyProg.exe has argv[]<\/td>\n      <\/tr><\/table><\/td><\/tr><\/table><br \/><a name=\"BASH\" id=\"BASH\">2.1\u00a0\u00a0Example 1: Bash Shell<\/a>\n<br \/>If you launch a program using the Bash shell command line,\nThe Bash shell is responsible for turning the string provided by\nthe user into the argv[] argument vector which\nis then passed to the program.\nFor the Bash shell parameter parsing rules, the following links provide some information:\n<ul><li><a href=\"http:\/\/www.gnu.org\/software\/bash\/manual\/bashref.html#Quoting\" target=\"_blank\">http:\/\/www.gnu.org\/software\/bash\/manual\/bashref.html#Quoting<img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/offsite.gif\" width=\"12\" height=\"11\" border=\"0\" alt=\"offsite\" \/><\/a>\n  <\/li><li><a href=\"http:\/\/linux.about.com\/library\/cmd\/blcmdl1_bash.htm\" target=\"_blank\">http:\/\/linux.about.com\/library\/cmd\/blcmdl1_bash.htm<img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/offsite.gif\" width=\"12\" height=\"11\" border=\"0\" alt=\"offsite\" \/><\/a>\n<\/li><\/ul><p><a name=\"UBUNTU\" id=\"UBUNTU\">2.2\u00a0\u00a0Example 2: Ubuntu Desktop Launcher Icon <small>(GNOME\u2192Nautilus)<\/small><\/a>\n<br \/>In Ubuntu a desktop launcher icon can be created by right-clicking on the\ndesktop and selecting \"Create Launcher\". This will generate a GNOME launch file\n~\/Desktop\/launcher-name.desktop\nThis is a text file you can browse (have a look at one). In this text file is\na line starting with Exec= where the command and\nany command line parameters are.<a name=\"S17\" href=\"#R17\" class=\"footnote\" id=\"S17\"><sup>17<\/sup><\/a>\n<\/p><p>So what happens when you double-click on this desktop icon?\n<\/p><p>Ubuntu uses the GNOME Desktop Environment, which in turn uses the Nautilus\nFile Manager. When you double-click on a desktop launcher icon, the Nautilus\nFile Manager parses the Exec= command line in the\n~\/Desktop\/launcher-name.desktop\nfile and calls\nexecve passing it argv[].\nSo to know how your command line parameters are parsed you need to look into\nhow Nautilus parses command lines. There's some good documentation here:\n<\/p><ul><li>The Exec Key<br \/><a href=\"http:\/\/library.gnome.org\/devel\/desktop-entry-spec\/#exec-variables\" target=\"_blank\">http:\/\/library.gnome.org\/devel\/desktop-entry-spec\/#exec-variables<img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/offsite.gif\" width=\"12\" height=\"11\" border=\"0\" alt=\"offsite\" \/><\/a>\n  <\/li><li>The GOption commandline parser<br \/><a href=\"http:\/\/library.gnome.org\/devel\/glib\/stable\/glib-Commandline-option-parser.html\" target=\"_blank\">http:\/\/library.gnome.org\/devel\/glib\/stable\/glib-Commandline-option-parser.html<img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/offsite.gif\" width=\"12\" height=\"11\" border=\"0\" alt=\"offsite\" \/><\/a>\n<\/li><\/ul><p><a name=\"KUBUNTU\" id=\"KUBUNTU\">2.3\u00a0\u00a0Example 3: Kubuntu <small>(KDE\u2192Plasma)<\/small><\/a>\n<br \/>Kubuntu is based on Ubuntu but uses the KDE Plasma Desktop instead of GNOME.\nSo, to determine how your command line is parsed in that environment, you'll\nneed to look into KDE application launchers. This starts to get complicated,\nbecause according to the documentation there are three: Kickoff, classic K\nmenu, or Lancelot Launcher.<a name=\"S18\" href=\"#R18\" class=\"footnote\" id=\"S18\"><sup>18<\/sup><\/a>\n<\/p><p>It gets even more complicated, because it's possible to install both the KDE\nPlasma Desktop (kubuntu-desktop) as well as the GNOME desktop\n(ubuntu-desktop) on the same machine.\n<\/p><p><a name=\"LINUXRULES\" id=\"LINUXRULES\">3.\u00a0\u00a0The *nix Parameter Parsing Rules<\/a>\n<br \/>It all depends on who is processing your command line <i>before<\/i> the new process is created.\nOn *nix, the parameters are parsed off by whatever program creates the new process.\n<\/p><p>This is all I have to say about *nix. The rest is about Windows.\n<\/p><div style=\"text-align: center;\"><img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/bara054.gif\" alt=\"bar\" \/><\/div>\n<p>Windows\u00ae<\/p>\n<br \/>1.\u00a0\u00a0How a new process is created on Windows\n<br \/>On Windows, a new program is launched by calling the\n<a href=\"http:\/\/msdn2.microsoft.com\/en-us\/library\/ms682425.aspx\">CreateProcess()<\/a>\nAPI, which takes the command line as a <i>string<\/i> (the <i>lpComandLine<\/i>\nparameter to CreateProcess):\n<p>\u00a0\u00a0\u00a0\u00a0\u00a0int CreateProcess( ..., <b><i>lpComandLine<\/i><\/b>, ... )<\/p><p>2.\u00a0\u00a0Passing Parameters to a new process on Windows\n<br \/>On Windows, the parameters are parsed off the command line <i>after<\/i> the new process is created,\n<i>by<\/i> the new process.\nIt's considered the responsibility of the newly started application to call the\n<a href=\"http:\/\/msdn2.microsoft.com\/en-us\/library\/ms683156.aspx\">GetCommandLine()<\/a>\nAPI to retrieve the command line string and parse it (possibly using the\n<a href=\"http:\/\/msdn2.microsoft.com\/en-us\/library\/ms647232.aspx\">CommandLineToArgvW()<\/a>\nhelper function).<a name=\"S2\" href=\"#R2\" class=\"footnote\" id=\"S2\"><sup>2<\/sup><\/a>\n<\/p><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table><tr><th colspan=\"6\">Windows<\/th>\n      <\/tr><tr><td class=\"code\">cmdline = \"MyProg.exe\u00a0\u00a0Hello\u00a0\u00a0Goodbye\u00a0\u00a0Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193 <span>create new process<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>CreateProcess(cmdline)\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process running MyProg.exe<\/td>\n      <\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>call<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"code\">GetCommandLine()<\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td>\n          <td>\n            <table><tr><td class=\"codetop\">argv[] = <\/td>\n                  <td class=\"code\">MyProg.exe\n                  <br \/>Hello\n                  <br \/>Goodbye\n                  <br \/>Friday\n                  <\/td>\n              <\/tr><\/table><\/td>\n          \n      <\/tr><\/table><\/td><\/tr><\/table><p>3.\u00a0\u00a0How does a C\/C++ program on Windows get <span>argv[]<\/span>?\n<br \/>The C\/C++ compiler which compiles the program secretly adds extra code to\nthe executable that retrieves and parses the command line to extract\nthe parameters before calling WinMain (or main). Thus, for a C\/C++\nexecutable on Windows, the parameter parsing rules are <b>determined by\nthe C\/C++ compiler that compiled the program.<\/b>\n<\/p><p>4.\u00a0\u00a0Everyone Parses Differently\n<br \/>You'll get different results if you pass a command line to\nShowParams.exe (written in C\/C++), ShowParams.vbs (VBScript),\nor ShowParams.bat (batch file):\n<\/p><p><b id=\"WINSAMP\">Sample Code:<\/b>\n<\/p><a name=\"WINRES\" id=\"WINRES\"><b>Results:<\/b><\/a>\n<ol><li>ShowParams.exe\n  <br \/><table><tr><th>ShowParams.exe (written in c)<\/th>\n        <\/tr><tr><td class=\"cmd\">\n            &gt;ShowParams.exe \"c:\\test a\\\" \"c:\\test b\\\"\n            <br \/>param 0 = ShowParams.exe\n            <br \/>param 1 = c:\\test a\" c:\\test\n            <br \/>param 2 = b\"\n        <\/td><\/tr><\/table><span>( <span>\\\"<\/span> is interpreted as meaning 'a literal double quote.')<\/span>\n      <\/li><li>C#ShowParams.exe\n  <br \/><table><tr><th>C#ShowParams.exe (written in C#)<\/th>\n        <\/tr><tr><td class=\"cmd\">\n            &gt;C#ShowParams.exe \"c:\\test a\\\" \"c:\\test b\\\"\n            <br \/>There are 2 program arguments\n            <br \/>c:\\test a\" c:\\test\n            <br \/>b\"\n        <\/td><\/tr><\/table><span>( <span>\\\"<\/span> is interpreted as meaning 'a literal double quote.')<\/span>\n      <\/li><li>ShowParams.bat\n  <br \/><table><tr><th>ShowParams.bat (a batch file)<\/th>\n        <\/tr><tr><td class=\"cmd\">\n            &gt;ShowParams.bat \"c:\\test a\\\" \"c:\\test b\\\"\n            <br \/>param 1 = \"c:\\test a\\\"\n            <br \/>param 2 = \"c:\\test b\\\"\n        <\/td><\/tr><\/table><span>(the double quotes are included as part of the parameter)<\/span>\n      <\/li><li>ShowParams.vbs\n  <br \/><table><tr><th>ShowParams.vbs (a Visual Basic script)<\/th>\n        <\/tr><tr><td class=\"cmd\">\n            &gt;ShowParams.vbs \"c:\\test a\\\" \"c:\\test b\\\"\n            <br \/>param 0 = c:\\test a\\\n            <br \/>param 1 = c:\\test b\\\n        <\/td><\/tr><\/table><span>(The double quotes are removed from the parameter)<\/span>\n<\/li><\/ol><a name=\"WINDIA\" id=\"WINDIA\"><b>Diagrams:<\/b><\/a>\n<ol><li>ShowParams.exe\n<br \/><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table><tr><th colspan=\"6\">ShowParams.exe\u00a0\u00a0\u00a0\u00a0\u00a0and\u00a0\u00a0\u00a0\u00a0\u00a0C#ShowParams.exe<\/th>\n      <\/tr><tr><td class=\"code\">cmdline = \"ShowParams.exe Hello Goodbye Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>CreateProcess(cmdline)\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process running ShowParams.exe<\/td>\n      <\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>call<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"code\">GetCommandLine()<\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td>\n          <td>\n            <table><tr><td class=\"codetop\">argv[] = <\/td>\n                  <td class=\"code\">ShowParams.exe\n                  <br \/>Hello\n                  <br \/>Goodbye\n                  <br \/>Friday\n                  <\/td>\n              <\/tr><\/table><\/td>\n          \n      <\/tr><\/table><\/td><\/tr><\/table><br \/>The parameter parsing rules are determined\nby <b>the C++ compiler<\/b> that compiled ShowParams.exe\n<\/li><li>ShowParams.bat\n<br \/><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table><tr><th colspan=\"6\">ShowParams.bat<\/th>\n      <\/tr><tr><td class=\"code\">cmdline = \"ShowParams.bat Hello Goodbye Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>CreateProcess(cmdline)\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process runs cmd.exe to process the batch file ShowParams.bat<\/td>\n      <\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>call<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"code\">GetCommandLine()<\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td>\n          <td class=\"code\">%1 = Hello\n            <br \/>%2 = Goodbye\n            <br \/>%3 = Friday\n          <\/td>\n          \n      <\/tr><\/table><\/td><\/tr><\/table><br \/>The parameter parsing rules are determined by <b>cmd.exe<\/b>\nwhich processes the batch file.\n<\/li><li>ShowParams.vbs\n<br \/><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table><tr><th colspan=\"6\">ShowParams.vbs<\/th>\n      <\/tr><tr><td class=\"code\">cmdline = \"ShowParams.vbs Hello Goodbye Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>CreateProcess(cmdline)\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process runs WScript.exe to process the VBScript file ShowParams.vbs<\/td>\n      <\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>call<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"code\">GetCommandLine()<\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td>\n          <td class=\"code\">Wscript.Arguments(0) = Hello\n            <br \/>Wscript.Arguments(1) = Goodbye\n            <br \/>Wscript.Arguments(2) = Friday\n          <\/td>\n          \n      <\/tr><\/table><\/td><\/tr><\/table><br \/>The parameter parsing rules are determined\nby <b>WScript.exe<\/b> which processes the VBScript file.\n<\/li><\/ol><a name=\"WINSUM\" id=\"WINSUM\"><b>Summary:<\/b><\/a>\n<ol><li>Parameters passed to ShowParams.exe are parsed by <b>ShowParams.exe<\/b>\u00a0.\u00a0The parameter parsing rules are determined by the C++ compiler that compiled ShowParams.exe\n<\/li><li>Parameters passed to ShowParams.bat are parsed by <b>cmd.exe<\/b> which is the program that processes batch files.\n<\/li><li>Parameters passed to ShowParams.vbs are parsed by <b>WScript.exe<\/b>\nwhich is the program that processes VBScript files<a name=\"S3\" href=\"#R3\" class=\"footnote\" id=\"S3\"><sup>3<\/sup><\/a>\n<\/li><\/ol>\n(Note: If you only see the first letter of each parameter, the parameters may be in Unicode format.)\n<hr \/><a name=\"WINCRULES\" id=\"WINCRULES\">5.\u00a0\u00a0The C\/C++ Parameter Parsing Rules<\/a>\n<br \/>The documented program parameter parsing rules for Microsoft C\/C++ compilers may be found by searching\n<a href=\"http:\/\/www.msdn.com\" target=\"_blank\">www.msdn.com<\/a>\nfor \"Parsing C++ Command-Line Arguments\".\n<ul><li>There is separate documentation for each version of Microsoft's C\/C++ compiler.\n  <\/li><li>Fortunately the documentation is identical for all of them.\n  <\/li><li>Unfortunately the documentation isn't complete.\n  <\/li><li>The implimentation changed in 2008, which isn't documented.\n  <\/li><li>Fortunately you're reading this.\n<\/li><\/ul><p>\u00a0\u00a0Visual C++ Versions and Corresponding .dll<\/p>\n<table class=\"msvcrt\"><tr><td>msvcr120.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual C++ 2013<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 12.0)<\/td><td>\u00a0\u2014\u00a0<\/td><td>(released on October 17, 2013)<\/td><\/tr><tr><td>msvcr110.dl<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual Studio 2012<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 11.0)<\/td><\/tr><tr><td>msvcr100.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual Studio 2010<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 10.0)<\/td><\/tr><tr><td>msvcr90.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual Studio 2008<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 9.0)<\/td><td>\u00a0\u2014\u00a0<\/td><td>[new command line parsing]<\/td><\/tr><tr><td>msvcr80.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual Studio 2005<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 8.0)<\/td><\/tr><tr><td>msvcr71.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual C++ .NET 2003<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 7.1)<\/td><\/tr><tr><td>msvcr70.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual C++ .NET 2002<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC++ 7.0)<\/td><\/tr><tr><td>msvcrt.dll<\/td><td>\u00a0\u2014\u00a0<\/td><td>Visual C++ 6.0<\/td><td>\u00a0\u2014\u00a0<\/td><td>(VC6)<\/td><\/tr><\/table><p><a href=\"http:\/\/en.wikipedia.org\/wiki\/Visual_C%2B%2B\" target=\"_blank\">http:\/\/en.wikipedia.org\/wiki\/Visual_C%2B%2B<\/a><\/p>\n<p>Redistribution of the shared C runtime component in Visual C++<\/p>\n<p><a href=\"http:\/\/support.microsoft.com\/kb\/326922\" target=\"_blank\">http:\/\/support.microsoft.com\/kb\/326922<\/a><\/p>\n<p>5.1\u00a0\u00a0Here are the documented (and undocumented) rules:<\/p>\n<br \/><ol class=\"verdana\"><li>Arguments are delimited by white space, which is either a space or a tab.<\/li>\n<li>The caret character (^) is not recognized as an escape character or delimiter. The character is handled completely by the command-line parser in the operating system before being passed to the argv array in the program.\n<br \/>They are referring to the scenario discussed in sec. <a href=\"#WINCMD\">6.2<\/a> below, where first the\ncommand line parser (cmd.exe) parses your command,\nhandling such things as the escape character\u00a0^\u00a0, the redirection\ncharacters\u00a0&gt;\u00a0&amp;\u00a0&lt;\u00a0,\nthe pipe character\u00a0|\u00a0, the\u00a0%\u00a0\ncharacter which may identify environment variables that need to be\nexpanded (e.g.\u00a0%PROGRAMFILES%), etc.\nThe rules here describe how your C\/C++ executable will parse the\n<i>lpCommandLine<\/i> that was passed to\u00a0CreateProcess()\nby cmd.exe or whoever calls CreateProcess().\n    <\/li>\n    <li>A string surrounded by double quotation marks (\"<em>string<\/em>\") is interpreted as a single argument, regardless of white space contained within. A quoted string can be embedded in an argument.\n      <br \/>The double quotes don't have to be around the whole parameter. A double quoted part may occur anywhere in the parameter.\n    <\/li>\n    <li><span>A double quotation mark preceded by a backslash (\\\") is interpreted as a literal double quotation mark character (\").<\/span><\/li>\n    <li><span>Backslashes are interpreted literally, unless they immediately precede a double quotation mark.<\/span><\/li>\n    <li><span>If an even number of backslashes is followed by a double quotation mark, one backslash is placed in the argv array for every pair of backslashes, and the double quotation mark is interpreted as a string delimiter.<\/span><\/li>\n    <li><span>If an odd number of backslashes is followed by a double quotation mark, one backslash is placed in the argv array for every pair of backslashes, and the double quotation mark is \"escaped\" by the remaining backslash, causing a literal double quotation mark (\") to be placed in argv.<\/span><\/li>\n    <li><b>The missing undocumented rule has to do with how doubledouble quotes (\"\") are handled:<\/b>\n        <ul><li class=\"pre2008\"><b>Prior to 2008:<\/b>\n              <br \/><span>If a closing \" is followed immediately by another \", the 2nd \" is accepted literally and added to the parameter.<\/span>\n          <\/li>\n          <li><b>After 2008<\/b>\n              <ol><li><span>A double quote encountered outside a double quoted block starts a double quoted block.<\/span><\/li>\n                <li><span>A double quote encountered inside a double quoted block:<\/span>\n                  <ul><li>not followed by another double quote ends the double quoted block.<\/li>\n                    <li>followed immediately by another double quote (e.g. \"\"), a single double quote is added to the output, and the double quoted block continues.<\/li>\n                  <\/ul><\/li>\n              <\/ol><\/li>\n        <\/ul><\/li>\n<\/ol><br \/><a name=\"WINCRULESREPH\" id=\"WINCRULESREPH\">5.2\u00a0\u00a0The Microsoft C\/C++ Parameter Parsing Rules Rephrased<\/a>\n<br \/>These are the rules for parsing a command line passed by CreateProcess() to a program written in C\/C++:\n<ol class=\"verdanam1\"><li>Parameters are <i>always<\/i> separated by a space or tab (multiple spaces\/tabs OK)\n  <br \/><\/li><li>If the parameter does not contain any spaces, tabs, or double quotes, then all the characters in the parameter are accepted as is (there is no need to enclose the parameter in double quotes).\n  <br \/><\/li><li>Enclose spaces and tabs in a double quoted part\n  <br \/><\/li><li>A double quoted part can be anywhere within a parameter\n  <br \/><\/li><li>2n backslashes followed by a \" produce n backslashes + start\/end double quoted part\n  <br \/><\/li><li>2n+1 backslashes followed by a \" produce n backslashes + a literal quotation mark\n  <br \/><\/li><li>n backslashes not followed by a quotation mark produce n backslashes\n  <br \/><\/li><li>undocumented rules regarding double quotes:<br \/><div class=\"pre2008\"><b>Prior to 2008:<\/b><ul><li><span>A \" outside a double quoted block starts a double quoted block<\/span><\/li>\n          <li><span>A \" inside a double quoted block ends the double quoted block<\/span><\/li>\n          <li><span>If a closing \" is followed immediately by another \", the 2nd \" is accepted literally and added to the parameter.<\/span><\/li>\n        <\/ul><\/div>\n<b>Post 2008:<\/b>\n        <ul><li><span>Outside a double quoted block a \" starts a double quoted block.<\/span><\/li>\n          <li><span>Inside a double quoted block a \" followed by a different character (not another \") ends the double quoted block.<\/span><\/li>\n          <li><span>Inside a double quoted block a \" followed immediately by another \" (i.e. \"\") causes a single \" to be added to the output, and the double quoted block continues.<\/span><\/li>\n        <\/ul><\/li>\n<\/ol><br \/><a name=\"WINCRULESSUM\" id=\"WINCRULESSUM\">5.3\u00a0\u00a0Summary of rules 5,6,7:<\/a>\n<br \/><table border=\"0\" class=\"ms12\"><tr><td>Use<\/td><td class=\"code96\">\"<\/td><td>to start\/end a double quoted part<\/td><\/tr><tr><td>Use<\/td><td class=\"code96\">\\\"<\/td><td>to insert a literal \"<\/td><\/tr><tr><td>Use<\/td><td class=\"code96\">\\\\\"<\/td><td>to insert a \\ then start or end a double quoted part<\/td><\/tr><tr><td>Use<\/td><td class=\"code96\">\u00a0\\\\\\\"\u00a0<\/td><td>to insert a literal \\\"<\/td><\/tr><tr><td>Use<\/td><td class=\"code96\">\\<\/td><td>to insert a literal \\<\/td><\/tr><\/table><p><a name=\"WINCRULESEX\" id=\"WINCRULESEX\">5.4\u00a0\u00a0Examples<\/a>\n<br \/><\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><table><tr><th><p>Command-Line<\/p><\/th>\n    <th><p>argv[1]<\/p><\/th>\n    <th><p>Comment<\/p><\/th>\n  <\/tr><tr><td><p>CallMeIshmael<\/p><\/td>\n    <td><p>CallMeIshmael<\/p><\/td>\n    <td><p>a plain parameter can contain any characters except {space} {tab}\u00a0\u00a0\\\u00a0\u00a0\"<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"Call Me Ishmael\"<\/p><\/td>\n    <td><p>\n       <br \/>Call Me Ishmael<\/p><\/td>\n    <td><p>\n       <br \/>spaces enclosed in a double quoted part<\/p><\/td>\n  <\/tr><tr><td><p>\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>Cal\"l Me I\"shmael<\/p><\/td>\n    <td><p>\n       <br \/>Call Me Ishmael<\/p><\/td>\n    <td><p>\n       <br \/>a double quoted part can be anywhere within a parameter<\/p><\/td>\n  <\/tr><tr><td><p>CallMe\\\"Ishmael\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>\u00a0CallMe\"Ishmael\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\\\" \u2192 \"<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"CallMe\\\"Ishmael\"\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\u00a0CallMe\"Ishmael\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\\\" \u2192 \"\u00a0\u00a0(whether or not in a double quoted part)<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"Call Me Ishmael\\\\\"\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>Call Me Ishmael\\\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\\\\\" \u2192 \\ + \" (which may begin or end a double quoted part)\n       <\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"CallMe\\\\\\\"Ishmael\"\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\u00a0CallMe\\\"Ishmael\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\\\\\\\" \u2192 \\\"\n           \u00a0\u00a0\u00a0\u00a0(\\\\ \u2192 \\)\u00a0\u00a0(\\\" \u2192 \")\n           <\/p><\/td>\n  <\/tr><tr><td><p>a\\\\\\b<br \/>\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>a\\\\\\b<br \/>\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>backslashes not followed immediately by a double quotation mark are interpreted literally<\/p><\/td>\n  <\/tr><tr><td><p>\"a\\\\\\b\"<br \/>\u00a0\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>a\\\\\\b<br \/>\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>whether or not the backslashes are in a double quoted part<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<p><a name=\"WINCRULESCT\" id=\"WINCRULESCT\">5.5\u00a0\u00a0Some Common Tasks<\/a>\n<\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><table><tr><th><p>Command-Line<\/p><\/th>\n    <th><p>argv[1]<\/p><\/th>\n    <th><p>Comment<\/p><\/th>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"\\\"Call Me Ishmael\\\"\"\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\"Call Me Ishmael\"\n       <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>\n       <br \/>the parameter includes double quotes<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"C:\\TEST A\\\\\"\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>C:\\TEST A\\\n       <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>\n       <br \/>the parameter includes a trailing slash<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"\\\"C:\\TEST A\\\\\\\"\"\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>\"C:\\TEST A\\\"\n       <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>\n       <br \/>the parameter includes double quotes and a trailing slash<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<p><a name=\"WINCRULESMSEX\" id=\"WINCRULESMSEX\">5.6\u00a0\u00a0The Microsoft Examples Explained<\/a>\n<\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><table><tr><th><p>Command-Line Input<\/p><\/th>\n    <th><p>argv[1]<\/p><\/th>\n    <th><p>argv[2]<\/p><\/th>\n    <th><p>argv[3]<\/p><\/th>\n    <th><p>Comment<\/p><\/th>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2510<br \/>\"a b c\"\u00a0\u00a0d\u00a0\u00a0e<\/p><\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2510<br \/>\u00a0a b c<\/p><\/td>\n    <td><p><br \/>d<\/p><\/td>\n    <td><p><br \/>e<\/p><\/td>\n    <td><p><br \/>spaces enclosed in double quotes<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u00a0\u250c\u2500\u2500\u2510<br \/>\"ab\\\"c\"\u00a0\u00a0\"\\\\\"\u00a0\u00a0d<br \/>\u00a0\u00a0\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2510<br \/>\u00a0ab\"c<br \/>\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>\u250c\u2500\u2510<br \/>\u00a0\\\u00a0<br \/>\u00a0\u2191<\/p><\/td>\n    <td><p><br \/>d<\/p><\/td>\n    <td valign=\"middle\"><p>\\\" \u2192 \"<br \/>\\\\\" \u2192 \\ + begin or end a double quoted part<\/p><\/td>\n  <\/tr><tr><td><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u250c\u2500\u2500\u2500\u2510\u00a0\u2193<br \/>a\\\\\\b d\"e f\"g h<br \/>\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p><br \/>a\\\\\\b<br \/>\u00a0\u2191\u2191\u2191<\/p><\/td>\n    <td><p>\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u2193<br \/>\u00a0de fg<\/p><\/td>\n    <td><p><br \/>h<\/p><\/td>\n    <td><p>backslashes not followed immediately by a double quotation mark are interpreted literally\n                        <br \/><sup>\u00a0\u2193\u00a0\u00a0\u00a0\u2193\u00a0\u00a0<\/sup>parameters are separated by spaces or tabs\n                        <br \/><sup>\u00a0\u250c\u2500\u2500\u2500\u2510\u00a0\u00a0<\/sup>a double quoted part can be anywhere within a parameter\n                        <br \/>the space enclosed in double quotation marks is not a delimiter<\/p><\/td>\n  <\/tr><tr><td><p>a\\\\\\\"b c d<br \/>\u00a0\u2191\u2191\u2191\u2191<\/p><\/td>\n    <td><p>a\\\"b<br \/>\u00a0\u2191\u2191<\/p><\/td>\n    <td><p>c<\/p><\/td>\n    <td><p>d<\/p><\/td>\n    <td><p>2n+1 backslashes before \"\u00a0\u2192 n backslashes + a literal \"<\/p><\/td>\n  <\/tr><tr><td><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2510\u2193\u00a0\u2193<br \/>a\\\\\\\\\"b c\" d e<br \/>\u00a0\u2191\u2191\u2191\u2191\u2191<\/p><\/td>\n    <td><p><br \/>a\\\\b c<\/p><\/td>\n    <td><p><br \/>d<\/p><\/td>\n    <td><p><br \/>e<\/p><\/td>\n    <td><p>2n backslashes followed by a \" produce n backslashes + start\/end double quoted part.\n                        <br \/><sup>\u00a0\u2193\u00a0\u00a0\u00a0\u2193\u00a0\u00a0<\/sup>parameters are separated by spaces or tabs\n                        <br \/><sup>\u00a0\u250c\u2500\u2500\u2500\u2510\u00a0\u00a0<\/sup>a double quoted part can be anywhere within a parameter\n                        <br \/>the space enclosed in double quotation marks is not a delimiter<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<p><a name=\"WINCRULESDDQ\" id=\"WINCRULESDDQ\">5.7\u00a0\u00a0Double Double Quote Examples<\/a>\n<\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><p>(post 2008)<\/p><table style=\"text-align: left; width: 80%;\"><tr><th><p>Command-Line Input<\/p><\/th>\n    <th><p>argv[1]<\/p><\/th>\n    <th><p>argv[2]<\/p><\/th>\n    <th><p>argv[3]<\/p><\/th>\n    <th><p>argv[4]<\/p><\/th>\n    <th><p>argv[5]<\/p><\/th>\n    <th><p>Comment<\/p><\/th>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500<br \/>\"a b c\"\"<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>a b c\"<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    \n    \n    \n    \n    <td>\n      <ul><li>\" Begin double quoted part.<\/li>\n        <li>\"\" while in a double quoted part \u2192 accept 2nd \" literally, double quoted part continues<\/li>\n      <\/ul><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"\"\"CallMeIshmael\"\"\"\u00a0\u00a0b\u00a0\u00a0c\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>\"CallMeIshmael\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p><br \/>b<\/p><\/td>\n    <td><p><br \/>c<\/p><\/td>\n    \n    \n    <td>\n      <ul><li>\" Begin double quoted part.<\/li>\n        <li>\"\" while in a double quoted part \u2192 accept 2nd \" literally, double quoted part continues<\/li>\n        <li>\" not followed by another \" (i.e. not \"\") while in a double quoted part \u2192 ends the double quoted part<\/li>\n        <li>Parameters are delimited by spaces or tabs.<\/li>\n      <\/ul><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"\"\"Call Me Ishmael\"\"\"\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>\"Call Me Ishmael\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    \n    \n    \n    \n    <td>\n      <ul><li>\" Begin double quoted part.<\/li>\n        <li>\"\" while in a double quoted part \u2192 accept 2nd \" literally, double quoted part continues<\/li>\n        <li>\" not followed by another \" (i.e. not \"\") while in a double quoted part \u2192 ends the double quoted part<\/li>\n     <\/ul><\/td>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n       <br \/>\"\"\"\"Call Me Ishmael\"\" b c\n       <br \/>\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>\"Call\n           <br \/>\u2191<\/p><\/td>\n    <td><p><br \/>Me<\/p><\/td>\n    <td><p><br \/>Ishmael<\/p><\/td>\n    <td><p><br \/>b<\/p><\/td>\n    <td><p><br \/>c<\/p><\/td>\n    <td>\n      <ul><li>\" Begin double quoted part.<\/li>\n        <li>\"\" while in a double quoted part \u2192 accept 2nd \" literally, double quoted part continues<\/li>\n        <li>\" not followed by another \" (i.e. not  \"\") in a double quoted part \u2192 ends the double quoted part<\/li>\n        <li>Parameters are delimited by spaces or tabs.<\/li>\n        <li>(note \"\" outside of double quoted block begins and then immediately ends a double quoted part.)<\/li>\n      <\/ul><\/td>\n  <\/tr><\/table><div class=\"pre2008\"><p>(pre 2008)<\/p><table><tr><th><p>Command-Line Input<\/p><\/th>\n    <th><p>argv[1]<\/p><\/th>\n    <th><p>argv[2]<\/p><\/th>\n    <th><p>argv[3]<\/p><\/th>\n    <th><p>Comment<\/p><\/th>\n  <\/tr><tr><td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2510<br \/>\"a b c\"\"<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>a b c\"<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    \n    \n    <td><p>\"\" while in a double quoted part \u2192 end double quoted part and accept 2nd \" literally<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n       <br \/>\"\"\"CallMeIshmael\"\"\"\u00a0\u00a0b\u00a0\u00a0c\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>\"CallMeIshmael\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p><br \/>b<\/p><\/td>\n    <td><p><br \/>c<\/p><\/td>\n    <td><p>\" Begin double quoted part.<br \/>\"\" while in a double quoted part \u2192 end double quoted part and accept 2nd \" literally<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n       <br \/>\"\"\"Call Me Ishmael\"\"\"\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>\"Call\n           <br \/>\u2191<\/p><\/td>\n    <td><p><br \/>Me<\/p><\/td>\n    <td><p><br \/>Ishmael\"<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p>Parameters are delimited by spaces or tabs.<br \/>\" Begin double quoted part.<br \/>\"\" while in a double quoted part \u2192 end double quoted part and accept 2nd \" literally<\/p><\/td>\n  <\/tr><tr><td><p>\u250c\u2510\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       <br \/>\"\"\"\"Call Me Ishmael\"\" b c\n       <br \/>\u00a0\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191<\/p><\/td>\n    <td><p><br \/>\"Call Me Ishmael\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n    <td><p><br \/>b<\/p><\/td>\n    <td><p><br \/>c<\/p><\/td>\n    <td><p>Parameters are delimited by spaces or tabs.<br \/>\" Begin double quoted part.<br \/>\"\" while in a double quoted part \u2192 end double quoted part and accept 2nd \" literally<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div><\/div>\n<p><a name=\"WINCRULESTDQ\" id=\"WINCRULESTDQ\">5.8\u00a0\u00a0Triple Double Quotes<\/a>\n<br \/><\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><p>(post 2008)<\/p><table><tr><th><p>How triple double quotes are parsed (post 2008)<\/p><\/th>\n  <\/tr><tr><td><p>    \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0<span>...\"\"\"Call Me Ishmael\"\"\"...<\/span>\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191\n           <br \/>quote #1: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u251c\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u251c\u2518\u251c\u2518\n           <br \/>quotes #2 &amp; 3: Skip 1st \" take 2nd \" \u2500\u2500\u2500\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u2502\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u2502\n           <br \/>quotes 4 &amp; 5: Skip 1st \" take 2nd \" \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u00a0\u2502\n           <br \/>quote #6: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \"\"\"Call Me Ishmael\"\"\"\n          <br \/>\u00a0param 1 = \"Call Me Ishmael\"\n        <\/p>\n        <p class=\"blk\">an alternative method is\n        <\/p>\n        <p class=\"cmd3\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510<br \/>\u00a0&gt;ShowParams.exe \\\"\"Call Me Ishmael\"\\\"\n          <br \/>\u00a0param 1 = \"Call Me Ishmael\"\n        <\/p>\n        <p class=\"blk\">or\n        <\/p>\n        <p class=\"cmd3\">\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510<br \/>\u00a0&gt;ShowParams.exe \"\\\"Call Me Ishmael\\\"\"\n          <br \/>\u00a0param 1 = \"Call Me Ishmael\"\n        <\/p>\n    <\/td>\n  <\/tr><\/table><div class=\"pre2008\"><p>(pre 2008)<\/p><table><tr><th><p>How triple double quotes were parsed (pre 2008)<\/p><\/th>\n  <\/tr><tr><td><p>    \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0<span>...\"\"\"Call Me Ishmael\"\"\"...<\/span>\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\n           <br \/>quote #1: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\n           <br \/>quote #2: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\n           <br \/>quote #3: and accept this \" literally\u2500\u2500\u2500\u2500\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\n           <br \/>quote #4: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\n           <br \/>quote #5: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\n           <br \/>quote #6: and accept this \" literally\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \"\"\"Call Me Ishmael\"\"\"\n          <br \/>\u00a0param 1 = \"Call\n          <br \/>\u00a0param 2 = Me\n          <br \/>\u00a0param 3 = Ishmael\"\n        <\/p>\n        <p class=\"blk\">an alternative method is\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \\\"Call Me Ishmael\\\"\n          <br \/>\u00a0param 1 = \"Call\n          <br \/>\u00a0param 2 = Me\n          <br \/>\u00a0param 3 = Ishmael\"\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div><\/div><\/div>\n<p><a name=\"WINCRULESQDQ\" id=\"WINCRULESQDQ\">5.9\u00a0\u00a0Quadruple Double Quotes<\/a>\n<\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><p>(post 2008)<\/p><table><tr><th><p>How quadruple double quotes are parsed(post 2008)<\/p><\/th>\n  <\/tr><tr><td><p>    \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0...\"\"\"\"Call\u00a0me\u00a0Ishmael\"\"\"\"...\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191\u2191\n           <br \/>quote #1: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u251c\u2518\u251c\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u251c\u2518\u2502\u2502\n           <br \/>quotes #2 &amp; 3: Skip 1st \" take 2nd \" \u2500\u2500\u2500\u2500\u2518\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u00a0\u2502\u2502\n           <br \/>quote #4: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u00a0\u2502\u2502\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u00a0\u2502\u2502\n           <br \/>quote #5: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u00a0\u2502\u2502\n           <br \/>quotes #6 &amp; 7: Skip 1st \" take 2nd \" \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u00a0\u2502\u2502\n           <br \/>quote #8: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Assuming this isn't another \" \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \"\"\"\"Call Me Ishmael\"\"\"\"\n          <br \/>\u00a0param 1 = \"Call\n          <br \/>\u00a0param 2 = Me\n          <br \/>\u00a0param 3 = Ishmael\"\n        <\/p>\n        <p class=\"blk\">an alternative method is\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \\\"Call Me Ishmael\\\"\n          <br \/>\u00a0param 1 = \"Call\n          <br \/>\u00a0param 2 = Me\n          <br \/>\u00a0param 3 = Ishmael\"\n        <\/p>\n    <\/td>\n  <\/tr><\/table><div class=\"pre2008\"><p>(pre 2008)<\/p><table><tr><th><p>How quadruple double quotes are parsed (pre 2008)<\/p><\/th>\n  <\/tr><tr><td><p>    \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0...\"\"\"\"Call\u00a0me\u00a0Ishmael\"\"\"\"...\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u2191\u2191\n           <br \/>quote #1: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\u2502\u2502\n           <br \/>quote #2: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\u2502\u2502\n           <br \/>quote #3: and accept this \" literally\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\u2502\u2502\n           <br \/>quote #4: Begin another double quoted part\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\u2502\u2502\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u2502\u2502\u2502\u2502\n           <br \/>quote #5: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\u2502\u2502\n           <br \/>quote #6: and accept this \" literally\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\u2502\n           <br \/>quote #7: Begin double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\u2502\n           <br \/>quote #8: End double quoted part\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u2502\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0If this was a double quote we'd accept it literally\u2500\u2500\u2518\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \"\"\"\"Call Me Ishmael\"\"\"\"\n          <br \/>\u00a0param 1 = \"Call Me Ishmael\"\n        <\/p>\n        <p class=\"blk\">Note quotes #7,#8 are not necessary. They contribute nothing.\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \"\"\"\"Call Me Ishmael\"\"\n          <br \/>\u00a0param 1 = \"Call Me Ishmael\"\n        <\/p>\n        <p class=\"blk\">an alternative method is\n        <\/p>\n        <p class=\"cmd3\">\u00a0&gt;ShowParams.exe \"\\\"Call Me Ishmael\\\"\"\n          <br \/>\u00a0param 1 = \"Call Me Ishmael\"\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div><\/div><\/div>\n<p><a name=\"WINCALGORITHM\" id=\"WINCALGORITHM\">5.10\u00a0\u00a0The Microsoft C\/C++ Command Line Parameter Parsing Algorithm<\/a>\n<\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><div class=\"pre2008\"><p>(pre 2008)<\/p>The following algorithm was reverse engineered by disassembling a small C\nprogram compiled using Microsoft Visual C++ and examining the\ndisassembled code.<p>See <a href=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/msvcrtparsing.htm\" target=\"_blank\">msvcrtparsing.htm<\/a> for the actual diasssembled code with annotations.\n<\/p><p>note if odd number of backslashes, we ignore last backslash and begin again.\n<\/p><table border=\"0\"><tr class=\"ms12\"><td valign=\"top\">1.<\/td><td colspan=\"6\">Parse off parameter 0 (the program filename)<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\" align=\"right\">\u2022<\/td><td colspan=\"6\">The entire parameter may be enclosed in double quotes (it handles double quoted parts)<br \/>(Double quotes are necessary if there are any spaces or tabs in the parameter)<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\" align=\"right\">\u2022<\/td><td colspan=\"6\">There is no special processing of backslashes (\\)<\/td><\/tr><tr class=\"ms12\"><\/tr><tr class=\"ms12\"><td valign=\"top\">2.<\/td><td colspan=\"6\">Parse off next parameter:<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\">a.<\/td><td colspan=\"6\">Skip over multiple spaces\/tabs between parameters<\/td><\/tr><tr class=\"ms12\"><td colspan=\"6\">LOOP<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\">b.<\/td><td colspan=\"5\">Count the backslashes (\\). Let m = number of backslashes. (m may be zero.)<\/td><\/tr><tr class=\"ms12\"><td colspan=\"5\">IF even number of backslashes (or zero)<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\">c.<\/td><td colspan=\"5\">IF next character following m backslashes is a double quote:<\/td><\/tr><tr class=\"ms12\"><td colspan=\"4\">Even number of backslashes?<\/td><\/tr><tr class=\"ms12\"><td colspan=\"4\">Yes. If we're not in a double quoted part, \" begins a double quoted part.<\/td><\/tr><tr class=\"ms12\"><td colspan=\"4\">Yes. If we're in a double quoted part, \"\" skip 1st take 2nd \"<\/td><\/tr><tr class=\"ms12\"><td colspan=\"4\">If m is even (or zero)<\/td><\/tr><tr class=\"ms12\"><td colspan=\"3\">if currently in a double quoted part<\/td><\/tr><tr class=\"ms12\"><td colspan=\"2\">IF next character is also a \"<\/td><\/tr><tr class=\"ms12\"><td colspan=\"1\">move to next character (the 2nd \". This character will be added to the parameter.)<\/td><\/tr><tr class=\"ms12\"><td colspan=\"2\">ELSE<\/td><\/tr><tr class=\"ms12\"><td colspan=\"1\">set flag to not add this \" character to the parameter<\/td><\/tr><tr class=\"ms12\"><td colspan=\"2\">toggle double quoted part flag (end double quoted part)<\/td><\/tr><tr class=\"ms12\"><td colspan=\"2\">ENDIF<\/td><\/tr><tr class=\"ms12\"><td colspan=\"3\">else<\/td><\/tr><tr class=\"ms12\"><td colspan=\"2\">set flag to not add this \" character to the parameter<\/td><\/tr><tr class=\"ms12\"><td colspan=\"3\">endif<\/td><\/tr><tr class=\"ms12\"><td colspan=\"4\">Endif<\/td><\/tr><tr class=\"ms12\"><td colspan=\"5\">ENDIF<\/td><\/tr><tr class=\"ms12\"><td colspan=\"5\">ENDIF<\/td><\/tr><tr class=\"ms12\"><td colspan=\"5\">Add backslashes to output<\/td><\/tr><tr class=\"ms12\"><td colspan=\"5\">m = m\/2  (floor divide e.g. 0\/2=0, 1\/2=0, 2\/2=1, 3\/2=1, 4\/2=2, 5\/2=2, etc.)<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\">d.<\/td><td colspan=\"5\">add m backslashes<\/td><\/tr><tr class=\"ms12\"><td valign=\"top\">e.<\/td><td colspan=\"5\">add this character to our parameter<\/td><\/tr><tr class=\"ms12\"><td colspan=\"5\">ENDLOOP<\/td><\/tr><\/table><\/div><p>(post 2008)<\/p><br \/>See<a href=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/msvcr100dparsing.htm\" target=\"_blank\">msvcr100dparsing.htm<\/a>for the actual diasssembled code with annotations.<\/div><\/div><\/div>\n<hr \/><a name=\"WINLINE\" id=\"WINLINE\">6.\u00a0\u00a0Who Calls CreateProcess?<\/a>\n<br \/>So far we've talked about parsing the command line that got passed as the <i>lpCommandLine<\/i>\nparameter to CreateProcess(). But who calls CreateProcess? One possibility is\nyou could write a small program which calls CreateProcess() passing it a command line.\n<p><a name=\"WINPA\" id=\"WINPA\">6.1.\u00a0\u00a0ProgA.exe \u2192 CreateProcess()<\/a>\n<br \/>A small program ProgA.exe calls CreateProcess() passing it a command line:\n<br \/><\/p><p><\/p><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table border=\"0\"><tr><th colspan=\"6\">ProgA.exe \u2192 CreateProcess()<\/th>\n      <\/tr><tr><td class=\"code\">ProgA.exe<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"code\">cmdline = \"ShowParams.exe Hello Goodbye Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193 <span>create new process<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td>CreateProcess(cmdline)\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process running ShowParams.exe<\/td>\n      <\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>call<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"code\">GetCommandLine()<\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td>\n          <td>\n            <table><tr><td class=\"codetop\">argv[] = <\/td>\n                  <td class=\"code\">ShowParams.exe\n                  <br \/>Hello\n                  <br \/>Goodbye\n                  <br \/>Friday\n                  <\/td>\n              <\/tr><\/table><\/td>\n          \n      <\/tr><\/table><\/td><\/tr><\/table>\n<p>Another possibility is you could open a Command Prompt window and type in a command:\n<\/p><p><a name=\"WINCMD\" id=\"WINCMD\">6.2\u00a0\u00a0Command Prompt Window \u2192 CreateProcess()<\/a>\n<br \/>The command prompt window is program cmd.exe (go to START \u2192 RUN, enter \"cmd\").\nProgram cmd.exe displays the command prompt window, reads your command, and parses it handling such things\nas redirection characters (&gt;) &amp; (&lt;),\nthe pipe character (|), the escape character (^),\nidentifying and expanding Environment Variables (e.g. %PROGRAMFILES%), etc.\nThen cmd.exe calls CreateProcess(), passing it a command line.<a name=\"S4\" href=\"#R4\" class=\"footnote\" id=\"S4\"><sup>4<\/sup><\/a>\n<br \/><\/p><p><\/p><table border=\"1\" cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td>\n    <table border=\"0\"><tr><th colspan=\"7\">Command Prompt Window (cmd.exe) \u2192 CreateProcess()<\/th>\n      <\/tr><tr><td colspan=\"2\" class=\"code\">Command Prompt Window (cmd.exe)<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td colspan=\"2\" class=\"darrow\">\u00a0\u2193 <span>get command from user<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td colspan=\"2\" class=\"code\">command = \"ShowParams.exe Hello Goodbye Friday &lt; in.txt &gt; out.txt\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td class=\"darrow\">\u00a0\u2193 <\/td><td class=\"darrow\"><span>parse and process command handling special characters like ^ &lt; &gt; | %<br \/>create a command line<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td colspan=\"2\" class=\"code\">cmdline = \"ShowParams.exe\u00a0\u00a0Hello\u00a0\u00a0Goodbye\u00a0\u00a0Friday\"<\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td colspan=\"2\" class=\"darrow\">\u00a0\u2193 <span>create new process<\/span><\/td><td class=\"rbdr\">\u00a0<\/td><\/tr><tr><td colspan=\"2\">CreateProcess(cmdline)\u00a0\u00a0\u2192<\/td><td class=\"rbdr\">\u00a0<\/td>\n          <td>\u2192<\/td><td>New Process running ShowParams.exe<\/td>\n      <\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>call<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"code\">GetCommandLine()<\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td><td class=\"darrow\">\u00a0\u2193 <span>parse<\/span><\/td><\/tr><tr><td class=\"rbdr\">\u00a0<\/td>\n            <td>\n            <table><tr><td class=\"codetop\">argv[] = <\/td>\n                  <td class=\"code\">ShowParams.exe\n                  <br \/>Hello\n                  <br \/>Goodbye\n                  <br \/>Friday\n                  <\/td>\n              <\/tr><\/table><\/td>\n          \n      <\/tr><\/table><\/td><\/tr><\/table>\n<br \/>Note here your command line is parsed <i>twice<\/i>:\n<ol><li>by the Command Prompt Window (cmd.exe)\n  <\/li><li>by ShowParams.exe\n<\/li><\/ol>\nSo if you want to get a parameter on the command line to ShowParams.exe,\nyou first need to get it through the cmd.exe parser, then through\nthe parser the Microsoft C\/C++ compiler added to ShowParams.exe\n<hr \/><p>7.\u00a0\u00a0The cmd.exe Command Prompt Parsing Rules<\/p>\nThe command prompt window is program cmd.exe<a name=\"S5\" href=\"#R5\" class=\"footnote\" id=\"S5\"><sup>5<\/sup><\/a>\n(go to START \u2192 RUN, enter \"cmd\").\nThe command prompt window.\n<div class=\"ms12\"><b>Escape Character<\/b><ul><li>^ is the escape character for commands passed to cmd.exe<\/li>\n<\/ul><br \/><b>Double Quotes:<\/b><ul><li>An unescaped \" will begin or end a double quoted part, and the double quote will be included as part of the parameter.<\/li>\n<li>^\" outside a double quoted part will insert a \" but will not start a double quoted part.<\/li>\n<li>Environment variables are expanded both inside and outside double quotes.<\/li>\n<\/ul><b>Outside Double Quotes:<\/b><ul><li>Use ^ outside double quotes to escape &lt; &gt; &amp; | ^\n<br \/>\u00a0\u00a0\u00a0to insert &lt; use ^&lt;\n<br \/>\u00a0\u00a0\u00a0to insert &gt; use ^&gt;\n<br \/>\u00a0\u00a0\u00a0to insert | use ^|\n<br \/>\u00a0\u00a0\u00a0to insert ^ use ^^\n<br \/>\u00a0\u00a0\u00a0to insert \" and begin a double quoted part use \"\n<br \/>\u00a0\u00a0\u00a0to insert \" and <b>not<\/b> begin a double quoted part use ^\"\n<\/li>\n<li>sometimes it may be necessary to also escape ( ) @ !<\/li>\n<li>It's OK to escape everything<\/li>\n<\/ul>Note ^ is also the line continuation character.<p><b>Inside Double Quotes:<\/b>\n<\/p><ul><li>Inside Double Quotes, everything is taken literally\n<\/li><li>^\" gives ^\" but does <b>not<\/b> end a double quoted part\n<\/li><li>Environment variable expansion is still performed inside double quotes\n<\/li><\/ul><br \/><b>Percent Sign:<\/b><ul><li>If your command line will be placed in a batch file, double the % character<a name=\"S6\" href=\"#R6\" class=\"footnote\" id=\"S6\"><sup>6<\/sup><\/a>\n<\/li><\/ul><\/div>\n<p>Note: The result is what gets passed as part of the <i>lpCommandLine<\/i>\nargument to the CreateProcess() API. The newly created process still needs to\nretrieve the <i>lpCommandLine<\/i> string and parse off the parameters using\nit's own parsing rules (e.g. if it's a C\/C++ program, it will use the Microsoft C\/C++\nparameter parsing rules to create the argv[] vector.\nSee <a href=\"#PIT\">Putting It Together<\/a> below for more on this).<\/p>\n<br \/><a name=\"CMDEX\" id=\"CMDEX\">7.1\u00a0\u00a0Examples:<\/a>\n<br \/><table><tr><th align=\"left\">Outside of Double Quotes Escape The Essential Characters &lt; &gt; &amp; | ^<\/th>\n        <\/tr><tr><td class=\"cmd4\">&gt;ShowParams.exe\u00a0!\\^\"#$%^&amp;'()*+,-.\/0123456789:;^&lt;=^&gt;?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^^_`abcdefghijklmnopqrstuvwxyz{^|}~\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2534\u2518\u00a0\u00a0\u25b2\u2191_\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191_\u00a0\u2191_\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191_\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191_\n            <p>lpCommandLine = ShowParams.exe !\"#$%&amp;'()*+,-.\/0123456789:;&lt;=&gt;?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\n        <\/p><\/td><\/tr><\/table><table border=\"0\"><tr><td class=\"ms12\">\u2191_\u00a0<\/td>\n          <td>These characters have been escaped with ^<\/td>\n        <\/tr><tr><td class=\"ms12\">\u00a0\u2514\u2534\u2518\u00a0<\/td>\n          <td>The\u00a0\"\u00a0character has been escaped with\u00a0^\u00a0so it doesn't start a double quoted part when cmd.exe parses it.\n              <br \/>Then\u00a0^\"\u00a0is escaped by\u00a0\\\u00a0so ShowParams.exe will see it as a literal \" (Microsoft C\/C++ parsing rules)\n              <br \/>The result is:\n              <br \/>\\^\" \u2192 cmd.exe parses to give \\\" \u2192 ShowParams.exe parses to give \u2192 \"\n          <\/td>\n        <\/tr><tr><td class=\"ms12\">\u00a0\u25b2\u00a0<\/td>\n          <td>Note if this command line is to be put in a batch file, the % will need to be doubled.<\/td>\n        <\/tr><\/table><table><tr><th align=\"left\">Outside of Double Quotes It's OK to Escape Everything<\/th>\n        <\/tr><tr><td class=\"cmd4\">&gt;ShowParams.exe\u00a0^!^\\^\"^#^$^%^&amp;^'^(^)^*^+^,^-^.^\/^0^1^2^3^4^5^6^7^8^9^:^;^&lt;^=^&gt;^?^@^A^B^C^D^E^F^G^H^I^J^K^L^M^N^O^P^Q^R\u00a0\u2192\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2534\u2534\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u25b2\n              <p>\u00a0\u2192\u00a0^S^T^U^V^W^X^Y^Z^[^\\^]^^^_^`^a^b^c^d^e^f^g^h^i^j^k^l^m^n^o^p^q^r^s^t^u^v^w^x^y^z^{^|^}^~\n              <\/p><p>lpCommandLine = ShowParams.exe !\"#$%&amp;'()*+,-.\/0123456789:;&lt;=&gt;?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\n        <\/p><\/td><\/tr><\/table><table border=\"0\"><tr><td class=\"ms12\">\u2192\u00a0<\/td>\n          <td>line continues<\/td>\n        <\/tr><tr><td class=\"ms12\">\u2514\u2534\u2534\u2518\u00a0<\/td>\n          <td>The\u00a0\"\u00a0character has been escaped with\u00a0^\u00a0so it doesn't start a double quoted part when cmd.exe parses it.\n              <br \/>Then\u00a0^\"\u00a0is escaped by\u00a0\\\u00a0so ShowParams.exe will see it as a literal \" (Microsoft C\/C++ parsing rules)\n              <br \/>\\\u00a0itself is escaped by\u00a0^\u00a0just for the heck of it (it's OK to escape everything).\n              <br \/>The result is:\n              <br \/>^\\^\" \u2192 cmd.exe parses to give \\\" \u2192 ShowParams.exe parses to give \u2192 \"\n          <\/td>\n        <\/tr><tr><td class=\"ms12\">\u00a0\u25b2\u00a0<\/td>\n          <td>Note if this command line is to be put in a batch file, the % will need to be doubled.<\/td>\n        <\/tr><\/table><table><tr><th align=\"left\">Inside Double Quotes Escape Nothing<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n            <br \/>&gt;ShowParams.exe\u00a0\"!#$%&amp;'()*+,-.\/0123456789:;&lt;=&gt;?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\"\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u25b2\n              <p>lpCommandLine = ShowParams.exe \"!#$%&amp;'()*+,-.\/0123456789:;&lt;=&gt;?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\"\n        <\/p><\/td><\/tr><\/table><table border=\"0\"><tr><td class=\"ms12\">\u00a0\u00a0\u00a0\"\u00a0<\/td><td>Outside a double quoted part, when not escaped with a\u00a0^, a double quote\u00a0\"\u00a0produces a double quote\u00a0\"\u00a0and begins a double quoted part<\/td><\/tr><tr><td class=\"ms12\">\u00a0\u00a0\u00a0\"\u00a0<\/td><td>Inside a double quoted part, a double quote\u00a0\"\u00a0produces a double quote\u00a0\"\u00a0and ends the double quoted part<\/td><\/tr><tr><td class=\"ms12\">\u00a0\u00a0\u00a0\u00a0\u00a0<\/td><td>The\u00a0\"\u00a0character is always included as part of the parameter<\/td><\/tr><tr><td class=\"ms12\">\u00a0\u00a0\u00a0\u2191\u00a0<\/td><td>Note the\u00a0\"\u00a0character has been left out of the sequence as it would end the double quoted part.\n              (You can't escape a\u00a0\"\u00a0while in a double quoted part.\n               The closest we could get would be to use\u00a0^\"\u00a0which would give us a\u00a0^\"\u00a0and not end the double quoted part.)<\/td><\/tr><tr><td class=\"ms12\">\u25b2\u00a0<\/td>\n          <td>Note if this command line is to be put in a batch file, the % will need to be doubled.<\/td>\n        <\/tr><\/table><p><a name=\"CMDEX2\" id=\"CMDEX2\">7.2\u00a0\u00a0Environment Variables:<\/a>\n<br \/><\/p><div class=\"tablemargin\"><div class=\"ContentArea\"><div class=\"topic\"><table><tr><th><p>Parameter<\/p><\/th>\n    \n    <th><p>Result<\/p><\/th>\n    <th><p>Comment<\/p><\/th>\n  <\/tr><tr><td><p>PROGRAMFILES<\/p><\/td>\n    <td><p>\u2192<\/p><\/td>\n    <td><p>PROGRAMFILES<\/p><\/td>\n    \n  <\/tr><tr><td><p>\"PROGRAMFILES\"<\/p><\/td>\n    <td><p>\u2192<\/p><\/td>\n    <td><p>\"PROGRAMFILES\"<\/p><\/td>\n    \n  <\/tr><tr><td><p>%PROGRAMFILES%<\/p><\/td>\n    <td><p>\u2192<\/p><\/td>\n    <td><p>C:\\Program Files<\/p><\/td>\n    <td><p>an environment variable<\/p><\/td>\n  <\/tr><tr><td><p>\"%PROGRAMFILES%\"<\/p><\/td>\n    <td><p>\u2192<\/p><\/td>\n    <td><p>\"C:\\Program Files\"<\/p><\/td>\n    <td><p>\" + an environment variable + \"<\/p><\/td>\n  <\/tr><tr><td><p>%XYZ%<\/p><\/td>\n    <td><p>\u2192<\/p><\/td>\n    <td><p>%XYZ%<\/p><\/td>\n    <td><p>(if XYZ is not an environment variable)<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<br \/><div class=\"footnotes\"><b>Some brief articles about Command Line parsing:<\/b><ul><li class=\"ddout\"><a href=\"http:\/\/technet.microsoft.com\/en-us\/library\/cc737438(WS.10).aspx?ppud=4\" target=\"_blank\">Command shell overview<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/technet.microsoft.com\/en-us\/library\/cc772622(WS.10).aspx\" target=\"_blank\">Using command redirection operators<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2006\/05\/16\/598893.aspx\" target=\"_blank\">Command line redirection is performed by the command line interpreter<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2006\/05\/18\/600962.aspx\" target=\"_blank\">The redirection operator can occur in the middle of the command line<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2006\/05\/17\/599916.aspx\" target=\"_blank\">Beware of digits before the redirection operator<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2006\/05\/15\/597984.aspx\" target=\"_blank\">The first word on the command line is the program name only by convention<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2006\/08\/23\/714650.aspx\" target=\"_blank\">Environment variable expansion occurs when the command is read<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2005\/01\/28\/362565.aspx\" target=\"_blank\">Capturing the current directory from a batch file<\/a><\/li>\n<li class=\"ddout\"><a href=\"http:\/\/blogs.msdn.com\/oldnewthing\/archive\/2003\/12\/10\/56028.aspx\" target=\"_blank\">What is the command line length limit?<\/a><\/li>\n<\/ul><\/div>\n<hr \/><p>PUTTING IT TOGETHER:<\/p>\n<p>8.<\/p>\n<p>A command line passed to a C\/C++ program passes through <i>two<\/i> parsers:<\/p>\n<ol><li>The cmd.exe command line parser parses your command &amp; parameters<\/li>\n  <li>The C\/C++ program retrieves the resulting command line and parses off the parameters<\/li>\n<\/ol><p>Here are some examples:<\/p>\n      <table><tr><th align=\"left\">Parsing Example 1<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0Bed\u00a0Bath\u00a0^&amp;\u00a0Beyond\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0Bath\u00a0&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0Bath\u00a0&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0argv[1]\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[2]\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[3]\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[4]\n              <p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed\n              <\/p><p>param 2 = Bath\n              <\/p><p>param 3 = &amp;\n              <\/p><p>param 4 = Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets\u00a0^&amp;\u00a0as an escaped\u00a0&amp;\u00a0character so it does not interpret this as \u201cstart a new command.\u201d (It's possible to put more than one DOS command on a command line by separating them with the &amp; character.)<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>C\/C++ command line parser receives this. There are no double quotes for it to process.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>C\/C++ command line parser parses off the command line parameters into the\u00a0argv[]\u00a0array.<\/td><\/tr><\/table><table><tr><th align=\"left\">Parsing Example 2<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0Bed\u00a0\\\"Bath\\\"\u00a0^&amp;\u00a0Beyond\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0\\\"Bath\\\"\u00a0&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0\"Bath\"\u00a0&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0argv[1]\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[2]\u00a0\u2193\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[3]\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[4]\n              <p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed\n              <\/p><p>param 2 = \"Bath\"\n              <\/p><p>param 3 = &amp;\n              <\/p><p>param 4 = Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets the first\u00a0\"\u00a0as \u201cStart a double quoted part\u201d and the second\u00a0\"\u00a0as \u201cEnd the double quoted part.\u201d cmd.exe does <i>not<\/i> remove the double quote characters. (Note that these double quotes are not escaped with the\u00a0^\u00a0character, which is the escape character for cmd.exe. Also note the\u00a0\\\u00a0character is treated as an ordinary character by cmd.exe)<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\"><\/td><td>cmd.exe command line parser interprets\u00a0^&amp;\u00a0as an escaped\u00a0&amp;\u00a0character so it does not interpret this as \u201cstart a new command.\u201d (It's possible to put more than one DOS command on a command line by separating them with the &amp; character.)<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>C\/C++ command line parser receives this. The C\/C++ parser interperts\u00a0\\\"\u00a0as an escaped double quote so neither one starts a double quoted part.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>C\/C++ command line parser parses off the command line parameters into the\u00a0argv[]\u00a0array.<\/td><\/tr><\/table><table><tr><th align=\"left\">Parsing Example 3<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0^\"Bed\u00a0Bath\u00a0^&amp;\u00a0Beyond^\"\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Bed\u00a0Bath\u00a0&amp;\u00a0Beyond\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0Bath\u00a0&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[1]\n              <\/p><p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed Bath &amp; Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\"><\/td><td>cmd.exe command line parser interprets\u00a0^&amp;\u00a0as an escaped\u00a0&amp;\u00a0character so it does not interpret this as \u201cstart a new command.\u201d (It's possible to put more than one DOS command on a command line by separating them with the &amp; character.)<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>C\/C++ command line parser receives this. The C\/C++ parser interperts\u00a0\"\u00a0as \u201cstart or end a double quoted part\u201d and the double quotes are removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>C\/C++ command line parser parses off the command line parameters into the\u00a0argv[]\u00a0array.<\/td><\/tr><\/table><table><tr><th align=\"left\">Parsing Example 4<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0^\"Bed\u00a0\\^\"Bath\\^\"\u00a0^&amp;\u00a0Beyond^\"\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Bed\u00a0\\\"Bath\\\"\u00a0&amp;\u00a0Beyond\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0\"Bath\"\u00a0&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[1]\n              <\/p><p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed \"Bath\" &amp; Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not interpret this as \u201cStart a double quoted part.\u201d<\/td><\/tr><tr><td class=\"ms12\"><\/td><td>cmd.exe command line parser interprets\u00a0^&amp;\u00a0as an escaped\u00a0&amp;\u00a0character so it does not interpret this as \u201cStart a new command.\u201d (It's possible to put more than one DOS command on a command line by separating them with the &amp; character.)<\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>C\/C++ command line parser receives this. The C\/C++ parser interperts\u00a0\\\"\u00a0as an escaped double quote so they do not start a double quoted part.<\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>C\/C++ command line parser parses off the command line parameters.<\/td><\/tr><\/table><p>It gets harder when double quotes are not escaped with ^ as cmd.exe will now interpert them as \u201cStart or end a double quoted part.\u201d\n      <\/p><table><tr><th align=\"left\">Parsing Example 5<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0\"Bed\u00a0Bath\u00a0^&amp;\u00a0Beyond\"\n              <p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Bed\u00a0Bath\u00a0^&amp;\u00a0Beyond\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0Bath\u00a0^&amp;\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[1]\n              <\/p><p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed Bath ^&amp; Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets the first\u00a0\"\u00a0as \u201cStart a double quoted part\u201d and the second\u00a0\"\u00a0as \u201cEnd the double quoted part.\u201d cmd.exe does <i>not<\/i> remove the double quote characters.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\"><\/td><td>Inside the double quoted part cmd.exe interprets all characters literally, including\u00a0^\u00a0and\u00a0&amp;\u00a0<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>C\/C++ command line parser receives this. The parser interperts the first\u00a0\"\u00a0as \u201cStart a double quoted part\u201d and the second\u00a0\"\u00a0as \u201cEnd the double quoted part.\u201d The double quotes are removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>C\/C++ command line parser parses off the command line parameters into the\u00a0argv[]\u00a0array.<\/td><\/tr><\/table><table><tr><th align=\"left\">Parsing Example 6<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0^\"Bed\u00a0Bath\u00a0\\^\"\u00a0Beyond^\"\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Bed\u00a0Bath\u00a0\\\"\u00a0Beyond\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0Bath\u00a0\"\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[1]\n              <\/p><p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed Bath \" Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\"><\/td><td>cmd.exe command line parser interprets\u00a0^&amp;\u00a0as an escaped\u00a0&amp;\u00a0character so it does not interpret this as \u201cstart a new command.\u201d (It's possible to put more than one DOS command on a command line by separating them with the &amp; character.)<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>C\/C++ command line parser receives this. The parser interperts\u00a0\"\u00a0as \u201cstart or end a double quoted part\u201d and the double quotes are removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>C\/C++ command line parser parses off the command line parameters into the\u00a0argv[]\u00a0array.<\/td><\/tr><\/table><table><tr><th align=\"left\">Parsing Example 7<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0^\"Bed\u00a0Bath\u00a0\\\\\\^\"\u00a0Beyond^\"\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Bed\u00a0Bath\u00a0\\\\\\\"\u00a0Beyond\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2534\u2534\u2518\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Bed\u00a0Bath\u00a0\\\"\u00a0Beyond\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[1]\n              <\/p><p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = Bed Bath \\\" Beyond\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part. The\u00a0^\u00a0is removed. The\u00a0\"\u00a0is <i>not<\/i> removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>The C\/C++ command line parser receives this.\n            <br \/>The C\/C++ command line parser interperts the first\u00a0\"\u00a0as \u201cStart a double quoted part\u201d and the double quote is removed.\n            <br \/>The C\/C++ command line parser interperts\u00a0\\\\\\\"\u00a0inside the double quoted part as a literal\u00a0\\\"\u00a0\n            <br \/>The C\/C++ command line parser interperts the ending\u00a0\"\u00a0as \u201cEnd the double quoted part\u201d and the double quote is removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>The C\/C++ command line parser fills the\u00a0argv[]\u00a0array with the result.<\/td><\/tr><\/table><table><tr><th align=\"left\">Parsing Example 8<\/th>\n        <\/tr><tr><td class=\"cmd4\">\n            <br \/>1.\u00a0\u00a0ShowParams.exe\u00a0^\"\\^\"Bed\u00a0Bath\u00a0Beyond\\^\"^\"\n            <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u2514\u2518\n              <p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n            \n              <br \/>2.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"\\\"Bed\u00a0Bath\u00a0Beyond\\\"\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2514\u2518\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>3.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Bed\u00a0Bath\u00a0Beyond\"\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n              <br \/>4.\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[0]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0argv[1]\n              <\/p><p>param 0 = ShowParams.exe\n              <\/p><p>param 1 = \"Bed Bath Beyond\"\n        <\/p><\/td><\/tr><\/table><br \/><table border=\"0\"><tr><td class=\"ms12\">1.\u00a0<\/td><td>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part. The\u00a0^\u00a0is removed. The\u00a0\"\u00a0is <i>not<\/i> removed.\n            <br \/>cmd.exe command line parser interprets\u00a0\\\u00a0as an ordinary character.\n            <br \/>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part. The\u00a0^\u00a0is removed. The\u00a0\"\u00a0is <i>not<\/i> removed.\n            <br \/>cmd.exe command line parser interprets\u00a0Bed Bath Beyond\\\u00a0as ordinary characters.\n            <br \/>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part. The\u00a0^\u00a0is removed. The\u00a0\"\u00a0is <i>not<\/i> removed.\n            <br \/>cmd.exe command line parser interprets\u00a0^\"\u00a0as an escaped\u00a0\"\u00a0character so it does not start or end a double quoted part. The\u00a0^\u00a0is removed. The\u00a0\"\u00a0is <i>not<\/i> removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">2.\u00a0<\/td><td>The C\/C++ command line parser receives this.\n            <br \/>The C\/C++ command line parser interperts the first\u00a0\"\u00a0as \u201cStart a double quoted part\u201d and the double quote is removed.\n            <br \/>The C\/C++ command line parser interperts\u00a0\\\"\u00a0inside the double quoted part as a literal\u00a0\"\u00a0\n            <br \/>The C\/C++ command line parser interperts\u00a0\\\"\u00a0inside the double quoted part as a literal\u00a0\"\u00a0\n            <br \/>The C\/C++ command line parser interperts the ending\u00a0\"\u00a0as \u201cEnd the double quoted part\u201d and the double quote is removed.<\/td><\/tr><tr><td class=\"ms12\"><\/td><\/tr><tr><td class=\"ms12\">3.\u00a0<\/td><td>The C\/C++ command line parser fills the\u00a0argv[]\u00a0array with the result.<\/td><\/tr><\/table><p>8.1\n<\/p><div style=\"text-align: center;\">How To Pass A Parameter to:\n<br \/>a C\/C++ Program from the Command Line<\/div>\n<p><b>A Simplified method:<\/b> <small>(<img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/new2a.gif\" width=\"28\" height=\"11\" alt=\"new!\" \/>2016)<\/small><\/p>\n<p>The trick is to enclose parameters with ^\" instead of just \"<\/p>\n<p>e.g. instead of:<br \/><\/p>\n<p>\"Bed Bath Beyond\"<\/p>\n<p>use<\/p>\n<p>^\"Bed Bath Beyond^\"<\/p>\n<p>The command processor cmd.exe will strip off the ^ and won't start any double quoted parts.<\/p>\n<p><b>Simplified method steps:<\/b><\/p>\n<div><ol><li>Start with the parameter you want<\/li>\n  <li>Escape with ^ all these special characters: ^ &lt; &gt; | &amp; ( ) \"\n<p>i.e. replace\u00a0\u00a0^\u00a0\u00a0with\u00a0\u00a0^^\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&lt;\u00a0\u00a0with\u00a0\u00a0^&lt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&gt;\u00a0\u00a0with\u00a0\u00a0^&gt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0|\u00a0\u00a0with\u00a0\u00a0^|\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&amp;\u00a0\u00a0with\u00a0\u00a0^&amp;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0(\u00a0\u00a0with\u00a0\u00a0^(\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0)\u00a0\u00a0with\u00a0\u00a0^)\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0^\"\n<\/p><\/li>\n<li>Then,\nfor all ^\", double the number of \\ immediately preceeding the ^\" (there may be zero, in which case do nothing).\n<\/li>\n<li>And\nfor all ^\", add a preceeding \\\n<p>Steps 3 &amp; 4 combined are:\n<br \/>replace\u00a0\u00a0^\"\u00a0\u00a0with\u00a0\u00a0\\^\"\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\\^\"\u00a0\u00a0with\u00a0\u00a0\\\\\\^\"\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\\\\^\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\\\\^\"\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\\\\\\^\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\\\\\\\\^\"\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0etc.\n<\/p><\/li>\n<li>Add a leading and trailing ^\"<\/li>\n<\/ol><\/div>\n<p>Example 8.1: Command Line to C\/C++ Program<\/p>\n<div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"2\"><p>Example 8.1:\u00a0\u00a0\u00a0\u00a0\u00a0Bed Bath &amp; Beyond<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td class=\"comb\">Start with the parameter you want (parameter includes spaces)<\/td>\n    <td><p>\u00a0\u00a0Bed Bath &amp; Beyond<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td class=\"comb\">Escape with ^ all special characters: ^ &lt; &gt; | &amp; ( ) \"<\/td><\/tr><tr><td class=\"comms\">\ni.e. replace\u00a0\u00a0^\u00a0\u00a0with\u00a0\u00a0^^\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&lt;\u00a0\u00a0with\u00a0\u00a0^&lt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&gt;\u00a0\u00a0with\u00a0\u00a0^&gt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0|\u00a0\u00a0with\u00a0\u00a0^|\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&amp;\u00a0\u00a0with\u00a0\u00a0^&amp;<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0(\u00a0\u00a0with\u00a0\u00a0^(\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0)\u00a0\u00a0with\u00a0\u00a0^)\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0^\"\n    <\/td>\n    <td><p>\u00a0\u00a0Bed Bath ^&amp; Beyond\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">3.<\/td><td class=\"comb\">For all ^\", double the number of \\ immediately preceeding the ^\"<br \/>(there may be zero)<\/td>\n    <td><p>no \\^\" in parameter\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">4.<\/td><td class=\"comb\">For all ^\", add a leading \\<\/td>\n      <td><p>no ^\" in parameter\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">5.<\/td><td class=\"comb\">Add a leading and trailing ^\"<\/td>\n      <td><p>^\"Bed Bath ^&amp; Beyond^\"\n             <br \/>\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td>6.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>^\"Bed Bath ^&amp; Beyond^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\">\u00a0<\/td><\/tr><tr><td colspan=\"3\" class=\"comb\">Now examine how this parameter gets parsed:<\/td>\n  <\/tr><tr><td class=\"combwt\">A.<\/td><td class=\"com\">cmd.exe treats ^ as escape character and removes them.\n          <br \/>cmd.exe finds no double quoted parts because all \" are escaped with ^<\/td>\n      <td><p>^\"Bed Bath &amp; Beyond^\"\n    <br \/>\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">B.<\/td><td class=\"com\">cmd.exe passes this as parameter to the C\/C++ program\n      <\/td>\n      <td><p>\u00a0\"Bed Bath &amp; Beyond\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">C.<\/td><td class=\"com\">C\/C++ program sees this as a double quoted parameter.\n      <\/td>\n      <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\u00a0\"Bed Bath &amp; Beyond\"\n           <br \/>\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193<\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">D.<\/td><td class=\"com\">C\/C++ program applies the rule\n           2n+1 backslashes followed by a \"<br \/>produce n backslashes + a literal \"\n      <\/td>\n      <td><p>no \\\" in parameter\n      <\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">E.<\/td><td class=\"com\">Result is the desired parameter we started with\n      <\/td>\n      <td><p>\u00a0\u00a0Bed Bath &amp; Beyond\n      <\/p><\/td>\n  <\/tr><\/table><p><b>Sample Test File:<\/b>\n<\/p><table class=\"comtb\"><tr><th colspan=\"2\"><p>Example 8.2:\u00a0\u00a0\u00a0\u00a0\u00a0Bed \"Bath\" &amp; Beyond<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td class=\"comb\">Start with the parameter you want (parameter includes spaces)<\/td>\n    <td><p>\u00a0\u00a0Bed \"Bath\" &amp; Beyond<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td class=\"comb\">Escape with ^ all special characters: ^ &lt; &gt; | &amp; ( ) \"<\/td><\/tr><tr><td class=\"comms\">\ni.e. replace\u00a0\u00a0^\u00a0\u00a0with\u00a0\u00a0^^\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&lt;\u00a0\u00a0with\u00a0\u00a0^&lt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&gt;\u00a0\u00a0with\u00a0\u00a0^&gt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0|\u00a0\u00a0with\u00a0\u00a0^|\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&amp;\u00a0\u00a0with\u00a0\u00a0^&amp;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0(\u00a0\u00a0with\u00a0\u00a0^(\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0)\u00a0\u00a0with\u00a0\u00a0^)\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0^\"\n    <\/td>\n    <td><p>\u00a0\u00a0Bed ^\"Bath^\" ^&amp; Beyond\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191_\u00a0\u00a0\u00a0\u00a0\u2191_\u00a0\u2191_\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">3.<\/td><td class=\"comb\">For all ^\", double the number of \\ immediately preceeding the ^\"<br \/>(there may be zero)<\/td>\n    <td><p>no \\^\" in parameter\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">4.<\/td><td class=\"comb\">For all ^\", add a leading \\<\/td>\n      <td><p>\u00a0\u00a0Bed \\^\"Bath\\^\" ^&amp; Beyond\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191__\u00a0\u00a0\u00a0\u00a0\u2191__\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">5.<\/td><td class=\"comb\">Add a leading and trailing ^\"<\/td>\n      <td><p>^\"Bed \\^\"Bath\\^\" ^&amp; Beyond^\"\n             <br \/>\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td>6.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>^\"Bed \\^\"Bath\\^\" ^&amp; Beyond^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\">\u00a0<\/td><\/tr><tr><td colspan=\"3\" class=\"comb\">Now examine how this parameter gets parsed:<\/td>\n  <\/tr><tr><td class=\"combwt\">A.<\/td><td class=\"com\">cmd.exe treats ^ as escape character and removes them.\n          <br \/>cmd.exe finds no double quoted parts because all \" are escaped with ^\n      <\/td>\n      <td><p>^\"Bed \\^\"Bath\\^\" ^&amp; Beyond^\"\n            <br \/>\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">B.<\/td><td class=\"com\">cmd.exe passes this as parameter to the C\/C++ program\n      <\/td>\n      <td><p>\u00a0\"Bed \\\"Bath\\\" &amp; Beyond\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">C.<\/td><td class=\"com\">C\/C++ program sees this as a double quoted parameter.\n         <br \/>Note that all \" inside the parameter are escaped with \\\n      <\/td>\n      <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\u00a0\"Bed \\\"Bath\\\" &amp; Beyond\"\n           <br \/>\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">D.<\/td><td class=\"com\">C\/C++ program applies the rule 2n+1 backslashes followed by a \"\n           <br \/>produce n backslashes + a literal quotation mark.\n           <br \/>\\\" becomes \" (for n=0)\n      <\/td>\n      <td><p>\u00a0\u00a0Bed \\\"Bath\\\" &amp; Beyond\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">E.<\/td><td class=\"com\">Result is the desired parameter we started with\n      <\/td>\n      <td><p>\u00a0\u00a0Bed \"Bath\" &amp; Beyond\n      <\/p><\/td>\n  <\/tr><\/table><p><b>Sample Test File:<\/b>\n<\/p><table class=\"comtb\" border=\"0\"><tr><th colspan=\"2\"><p>Example 8.3:\u00a0\u00a0\u00a0\u00a0\u00a0Bed Bath &amp; \\\"Beyond\"<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td class=\"comb\">Start with the parameter you want (parameter includes spaces)<\/td>\n    <td><p>\u00a0\u00a0Bed Bath &amp; \\\"Beyond\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td class=\"comb\">Escape with ^ all special characters: ^ &lt; &gt; | &amp; ( ) \"<\/td><\/tr><tr><td class=\"comms\">\ni.e. replace\u00a0\u00a0^\u00a0\u00a0with\u00a0\u00a0^^\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&lt;\u00a0\u00a0with\u00a0\u00a0^&lt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&gt;\u00a0\u00a0with\u00a0\u00a0^&gt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0|\u00a0\u00a0with\u00a0\u00a0^|\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&amp;\u00a0\u00a0with\u00a0\u00a0^&amp;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0(\u00a0\u00a0with\u00a0\u00a0^(\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0)\u00a0\u00a0with\u00a0\u00a0^)\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0^\"\n    <\/td>\n    <td><p>\u00a0\u00a0Bed Bath ^&amp; \\^\"Beyond^\"\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">3.<\/td><td class=\"comb\">For all ^\", double the number of \\ immediately preceeding the ^\"<br \/>(there may be zero)<\/td>\n    <td><p>\u00a0\u00a0Bed Bath ^&amp; \\\\^\"Beyond^\"\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">4.<\/td><td class=\"comb\">For all ^\", add a leading \\<\/td>\n    <td><p>\u00a0\u00a0Bed Bath ^&amp; \\\\\\^\"Beyond\\^\"\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191__\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191__\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">5.<\/td><td class=\"comb\">Add a leading and trailing ^\"<\/td>\n      <td><p>^\"Bed Bath ^&amp; \\\\\\^\"Beyond\\^\"^\"\n             <br \/>\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td>6.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>^\"Bed Bath ^&amp; \\\\\\^\"Beyond\\^\"^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\">\u00a0<\/td><\/tr><tr><td colspan=\"3\" class=\"comb\">Now examine how this parameter gets parsed:<\/td>\n  <\/tr><tr><td class=\"combwt\">A.<\/td><td class=\"com\">cmd.exe treats ^ as escape character and removes them.\n          <br \/>cmd.exe finds no double quoted parts because all \" are escaped with ^<\/td>\n      <td><p>^\"Bed Bath ^&amp; \\\\\\^\"Beyond\\^\"^\"\n    <br \/>\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u2193<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">B.<\/td><td class=\"com\">cmd.exe passes this as parameter to the C\/C++ program\n      <\/td>\n      <td><p>\u00a0\"Bed Bath &amp; \\\\\\\"Beyond\\\"\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">C.<\/td><td class=\"com\">C\/C++ program sees this as a double quoted parameter.\n         <br \/>Note that all \" inside the parameter are escaped with \\\n      <\/td>\n      <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\u00a0\"Bed Bath &amp; \\\\\\\"Beyond\\\"\"\n           <br \/>\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">D.<\/td><td class=\"com\">C\/C++ program applies the rule 2n+1 backslashes followed by a \"\n           <br \/>produce n backslashes + a literal quotation mark\n           <br \/>\\\" becomes \"\n           <br \/>\\\\\\\" becomes \\\"\n      <\/td>\n      <td><p>\u00a0\u00a0Bed Bath &amp; \\\\\\\"Beyond\\\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0----\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0--\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">E.<\/td><td class=\"com\">Result is the desired parameter we started with\n      <\/td>\n      <td><p>\u00a0\u00a0Bed Bath &amp; \\\"Beyond\"\n      <\/p><\/td>\n  <\/tr><\/table><p><b>Sample Test File:<\/b>\n<\/p><table class=\"comtb\" border=\"0\"><tr><th colspan=\"2\"><p>Example 8.4:\u00a0\u00a0\u00a0\u00a0\u00a0(Bed)\u00a0&lt;Bath&gt;\u00a0&amp;\u00a0\"Beyond\"\u00a0|\u00a0\\\"Kohl^s<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td class=\"comb\">Start with the parameter you want (parameter includes spaces)<\/td>\n    <td><p>\u00a0\u00a0(Bed)\u00a0&lt;Bath&gt;\u00a0&amp;\u00a0\"Beyond\"\u00a0|\u00a0\\\"Kohl^s<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td class=\"comb\">Escape with ^ all special characters: ^ &lt; &gt; | &amp; ( ) \"<\/td><\/tr><tr><td class=\"comms\">\ni.e. replace\u00a0\u00a0^\u00a0\u00a0with\u00a0\u00a0^^\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&lt;\u00a0\u00a0with\u00a0\u00a0^&lt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&gt;\u00a0\u00a0with\u00a0\u00a0^&gt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0|\u00a0\u00a0with\u00a0\u00a0^|\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&amp;\u00a0\u00a0with\u00a0\u00a0^&amp;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0(\u00a0\u00a0with\u00a0\u00a0^(\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0)\u00a0\u00a0with\u00a0\u00a0^)\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0^\"\n    <\/td>\n    <td><p>\u00a0\u00a0^(Bed^)\u00a0^&lt;Bath^&gt;\u00a0^&amp;\u00a0^\"Beyond^\"\u00a0^|\u00a0\\^\"Kohl^^s\n          <br \/>\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">3.<\/td><td class=\"comb\">For all ^\", double the number of \\ immediately preceeding the ^\"<br \/>(there may be zero)<\/td>\n    <td><p>\u00a0\u00a0^(Bed^)\u00a0^&lt;Bath^&gt;\u00a0^&amp;\u00a0^\"Beyond^\"\u00a0^|\u00a0\\\\^\"Kohl^^s\n          <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">4.<\/td><td class=\"comb\">For all ^\", add a leading \\<\/td>\n      <td><p>\u00a0\u00a0^(Bed^)\u00a0^&lt;Bath^&gt;\u00a0^&amp;\u00a0\\^\"Beyond\\^\"\u00a0^|\u00a0\\\\\\^\"Kohl^^s\n             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">5.<\/td><td class=\"comb\">Add a leading and trailing ^\"<\/td>\n      <td><p>^\"^(Bed^)\u00a0^&lt;Bath^&gt;\u00a0^&amp;\u00a0\\^\"Beyond\\^\"\u00a0^|\u00a0\\\\\\^\"Kohl^^s^\"\n             <br \/>\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td>6.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>^\"^(Bed^)\u00a0^&lt;Bath^&gt;\u00a0^&amp;\u00a0\\^\"Beyond\\^\"\u00a0^|\u00a0\\\\\\^\"Kohl^^s^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\">\u00a0<\/td><\/tr><tr><td colspan=\"3\" class=\"comb\">Now examine how this parameter gets parsed:<\/td>\n  <\/tr><tr><td class=\"combwt\">A.<\/td><td class=\"com\">cmd.exe treats ^ as escape character and removes them.\n          <br \/>cmd.exe finds no double quoted parts because all \" are escaped with ^<\/td>\n      <td><p>^\"^(Bed^)\u00a0^&lt;Bath^&gt;\u00a0^&amp;\u00a0\\^\"Beyond\\^\"\u00a0^|\u00a0\\\\\\^\"Kohl^^s^\"\n    <br \/>\u2193\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u2193\u00a0\u00a0\u2193<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">B.<\/td><td class=\"com\">cmd.exe passes this as parameter to the C\/C++ program\n      <\/td>\n      <td><p>\u00a0\u00a0\"(Bed)\u00a0&lt;Bath&gt;\u00a0&amp;\u00a0\\\"Beyond\\\"\u00a0|\u00a0\\\\\\\"Kohl^s\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">C.<\/td><td class=\"com\">C\/C++ program sees this as a double quoted parameter.\n           <br \/>Note all \" inside the double quoted parameter are escaped with \\\n      <\/td>\n      <td><p>\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\u00a0\u00a0\"(Bed)\u00a0&lt;Bath&gt;\u00a0&amp;\u00a0\\\"Beyond\\\"\u00a0|\u00a0\\\\\\\"Kohl^s\"\n           <br \/>\u00a0\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193<\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">D.<\/td><td class=\"com\">C\/C++ program applies the rule 2n+1 backslashes followed by a \"\n           <br \/>produce n backslashes + a literal quotation mark\n           <br \/>\\\" becomes \"\n           <br \/>\\\\\\\" becomes \\\"\n      <\/td>\n      <td><p>\u00a0\u00a0\u00a0(Bed)\u00a0&lt;Bath&gt;\u00a0&amp;\u00a0\\\"Beyond\\\"\u00a0|\u00a0\\\\\\\"Kohl^s\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0--\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0--\u00a0\u00a0\u00a0----<\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">E.<\/td><td class=\"com\">Result is the desired parameter we started with\n      <\/td>\n      <td><p>\u00a0\u00a0\u00a0(Bed)\u00a0&lt;Bath&gt;\u00a0&amp;\u00a0\"Beyond\"\u00a0|\u00a0\\\"Kohl^s\n      <\/p><\/td>\n  <\/tr><\/table><p><b>Sample Test File:<\/b>\n<\/p><table class=\"comtb\" border=\"0\"><tr><th colspan=\"2\"><p>Example 8.5:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|()@!\"<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"2\"><p>Start with the parameter you want<br \/>(parameter includes a space)<\/p><\/td>\n    <td><p>\u00a0\u00a0&amp;&lt;&gt;^|()@!\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td class=\"comb\">Escape with ^ all special characters: ^ &lt; &gt;| &amp; ( ) \"<\/td><\/tr><tr><td class=\"comms\">\ni.e. replace\u00a0\u00a0^\u00a0\u00a0with\u00a0\u00a0^^\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&lt;\u00a0\u00a0with\u00a0\u00a0^&lt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&gt;\u00a0\u00a0with\u00a0\u00a0^&gt;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0|\u00a0\u00a0with\u00a0\u00a0^|\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0&amp;\u00a0\u00a0with\u00a0\u00a0^&amp;\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0(\u00a0\u00a0with\u00a0\u00a0^(\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0)\u00a0\u00a0with\u00a0\u00a0^)\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0replace\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0^\"\n    <\/td>\n    <td><p>\u00a0\u00a0^&amp;^&lt;^&gt;^^^|^(^)^@^!^\"\n          <br \/>\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td class=\"comb\">For all ^\", double the number of \\ immediately preceeding the ^\"<br \/>(there may be zero)<\/td>\n    <td><p>no \\^\" in parameter\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">3.<\/td><td class=\"comb\">For all ^\", add a leading \\<\/td>\n      <td><p>\u00a0\u00a0^&amp;^&lt;^&gt;^^^|^(^)^@^!\\^\"\n             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">4.<\/td><td class=\"comb\">Add a leading and trailing ^\"<\/td>\n      <td><p>^\"^&amp;^&lt;^&gt;^^^|^(^)^@^!\\^\"^\"\n             <br \/>\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td>5.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><p>Result: To get desired parameter use this:<\/p><\/td>\n      <td><p>^\"^&amp;^&lt;^&gt;^^^|^(^)^@^!\\^\"^\"\n          <\/p>\n      <\/td>\n  <\/tr><tr><td colspan=\"3\">\u00a0<\/td><\/tr><tr><td colspan=\"3\" class=\"comb\">Now examine how this parameter gets parsed:<\/td>\n  <\/tr><tr><td class=\"combwt\">A.<\/td><td class=\"com\">cmd.exe treats ^ as escape character and removes them.\n          <br \/>cmd.exe finds no double quoted parts because all \" are escaped with ^<\/td>\n      <td><p>^\"^&amp;^&lt;^&gt;^^^|^(^)^@^!\\^\"^\"\n             <br \/>\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u2193\u00a0\u00a0\u2193\u00a0\u2193\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">B.<\/td><td class=\"com\">cmd.exe passes this as parameter to the C\/C++ program\n      <\/td>\n      <td><p>\u00a0\"&amp;&lt;&gt;^|()@!\\\"\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">C.<\/td><td class=\"com\">C\/C++ program sees this as a double quoted parameter.\n           <br \/>Note all \" inside the double quoted parameter are escaped with \\\n      <\/td>\n      <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\u00a0\"&amp;&lt;&gt;^|()@!\\\"\"\n           <br \/>\u00a0\u2193\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2193<\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">D.<\/td><td class=\"com\">C\/C++ program applies the rule 2n+1 backslashes followed by a \"\n           <br \/>produce n backslashes + a literal quotation mark\n           <br \/>\\\" becomes \"\n      <\/td>\n      <td><p>\u00a0\u00a0&amp;&lt;&gt;^|()@!\\\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0--<\/p>\n      <\/td>\n  <\/tr><tr><td class=\"combwt\">E.<\/td><td class=\"com\">Result is the desired parameter we started with\n      <\/td>\n      <td><p>\u00a0\u00a0&amp;&lt;&gt;^|()@!\"\n      <\/p><\/td>\n  <\/tr><\/table><b>Sample Test File:<\/b><\/div><\/div><\/div>\n<p>The following older examples show the harder way of figuring out what to use as parameters.<\/p>\n<p>Example 8.6: Command Line to C\/C++ Program<\/p>\n<div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 8.6a:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|()@ !<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<br \/>(parameter includes a space)<\/p><\/td>\n    <td><p>&amp;&lt;&gt;^|()@ !<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td><p>nothing to replace<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose the whole parameter in double quotes<br \/>(because there's a space in the parameter)<\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"&amp;&lt;&gt;^|()@ !\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n             <br \/>\"&amp;&lt;&gt;^|()@ !\"<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n      <td><p>Nothing to escape because it's all in a double quoted part<br \/>(as seen by cmd.exe)<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"&amp;&lt;&gt;^|()@ !\"<\/p><\/td>\n  <\/tr><tr><th colspan=\"3\"><p>Example 8.6b:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|()@ !<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<br \/>(same as example 1a)<\/p><\/td>\n    <td><p>&amp;&lt;&gt;^|()@ !<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n    <td><p>nothing to replace<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose spaces in double quotes\n        <br \/>(A double quoted part can be anywhere within a parameter)<\/td>\n    <td><p>&amp;&lt;&gt;^|()@\" \"!\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\n            <br \/>&amp;&lt;&gt;^|()@\" \"!\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n      <td><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\n             <br \/>^&amp;^&lt;^&gt;^^^|^(^)^@\" \"^!\n             <br \/>\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>^&amp;^&lt;^&gt;^^^|^(^)^@\" \"^!\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<b>Sample Test File:<\/b>\n<p>Example 8.7: Command Line to C\/C++ Program<\/p>\n<div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 8.7:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|@()!\"&amp;&lt;&gt;^|@() !<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<\/p><\/td>\n    <td><p>&amp;&lt;&gt;^|@()!\"&amp;&lt;&gt;^|@()!<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td><p>&amp;&lt;&gt;^|@()!\\\"&amp;&lt;&gt;^|@()!\n         <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose the whole parameter in double quotes<\/td>\n    <td><p>\"&amp;&lt;&gt;^|@()!\\\"&amp;&lt;&gt;^|@()!\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\n             <br \/>\"&amp;&lt;&gt;^|()@!\\\"&amp;&lt;&gt;^|@()!\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><\/td><td class=\"comt\">we have a problem in that the final\u00a0\"\u00a0is interpreted\n             by cmd.exe as opening a double quoted part. To\n             avoid this, escape that last\u00a0\"\u00a0( the escape character for cmd.exe is ^ )<\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n             <br \/>\"&amp;&lt;&gt;^|()@!\\\"&amp;&lt;&gt;^|@()!^\"\n             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the other special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n             <br \/>\"&amp;&lt;&gt;^|@()!\\\"^&amp;^&lt;^&gt;^^^|@()!^\"\n             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\n          <\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"&amp;&lt;&gt;^|@()!\\\"^&amp;^&lt;^&gt;^^^|@()!^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 1b don't enclose the parameter in double quotes, then\n                         escape all special characters including the double quote\n                         so it doesn't start a double quoted part.<\/p><\/td>\n    <td><p>^&amp;^&lt;^&gt;^^^|@()!\\^\"^&amp;^&lt;^&gt;^^^|@()!\n           <br \/>\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<b>Sample Test File:<\/b>\n<p>\nExample 8.8: Command Line to C\/C++ Program\n<\/p><div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 8.8a:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|@() !\"&amp;&lt;&gt;^|@() !<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want (parameter includes leading and trailing double quotes, plus a double quote inside, and two spaces)<\/p><\/td>\n    <td><p>\"&amp;&lt;&gt;^|@() !\"&amp;&lt;&gt;^|@() !\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td><p>\\\"&amp;&lt;&gt;^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"\n         <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose the whole parameter in double quotes<\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"\\\"&amp;&lt;&gt;^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n    <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u250c\u2500\u2500\u2500\n           <br \/>\"\\\"&amp;&lt;&gt;^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"\"\n        <\/p>\n    <\/td>\n  <\/tr><tr><td colspan=\"2\"><\/td><td class=\"comt\">we have a problem in that the final\u00a0\"\u00a0is interpreted\n             by cmd.exe as opening a double quoted part. To\n             avoid this, escape that last\u00a0\"\u00a0( the escape character for cmd.exe is ^ )<\/td>\n    <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"\\\"&amp;&lt;&gt;^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n    <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"\\\"^&amp;^&lt;^&gt;^^^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"^\"\n           <br \/>\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"\\\"^&amp;^&lt;^&gt;^^^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"^\"\n        <\/p>\n    <\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 2a just escape all special characters including\n                         all double quotes so there are no double quoted parts.<\/p><\/td>\n    <td><p>^\"\\^\"^&amp;^&lt;^&gt;^^^|@() !\\^\"^&amp;^&lt;^&gt;^^^|@() !\\^\"^\"\n           <br \/>\u2191\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><th colspan=\"3\"><p>Example 8.8b:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|@() !\"&amp;&lt;&gt;^|@() !<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want (same as 3a)<\/p><\/td>\n    <td><p>\"&amp;&lt;&gt;^|@() !\"&amp;&lt;&gt;^|@() !\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td><p>\\\"&amp;&lt;&gt;^|@() !\\\"&amp;&lt;&gt;^|@() !\\\"\n         <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose spaces in double quotes\n        <br \/>(A double quoted part can be anywhere within a parameter)<\/td>\n    <td><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\n           <br \/>\\\"&amp;&lt;&gt;^|@()\" \"!\\\"&amp;&lt;&gt;^|@()\" \"!\\\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n    <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u250c\u2500\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\u00a0\u00a0\u250c\u2500\u2500\u2500\n           <br \/>\\\"&amp;&lt;&gt;^|@()\" \"!\\\"&amp;&lt;&gt;^|@()\" \"!\\\"\n        <\/p>\n    <\/td>\n  <\/tr><tr><td colspan=\"2\"><\/td><td class=\"comt\">once again we have a problem in that the final\u00a0\"\u00a0is interpreted\n             by cmd.exe as opening a double quoted part. To\n             avoid this, escape that last\u00a0\"\u00a0( the escape character for cmd.exe is ^ )<\/td>\n    <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u250c\u2500\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\n           <br \/>\\\"&amp;&lt;&gt;^|@()\" \"!\\\"&amp;&lt;&gt;^|@()\" \"!\\^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the other special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n    <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u250c\u2500\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\n           <br \/>\\\"&amp;&lt;&gt;^|@()\" \"!\\\"^&amp;^&lt;^&gt;^^^|@()\" \"!\\^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\\\"&amp;&lt;&gt;^|@()\" \"!\\\"^&amp;^&lt;^&gt;^^^|@()\" \"!\\^\"\n        <\/p>\n    <\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 2a just escape all special characters including\n                         all double quotes so there are no double quoted parts.<\/p><\/td>\n    <td><p>\\^\"^&amp;^&lt;^&gt;^^^|@()^\" ^\"!\\^\"^&amp;^&lt;^&gt;^^^|@()^\" ^\"!\\^\"\n           <br \/>\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<b>Sample Test File:<\/b>\n<p>\nExample 8.9: Command Line to C\/C++ Program\n<\/p><div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 8.9a:\u00a0\u00a0\u00a0\u00a0\u00a0\"C:\\TEST A\\\"<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want (parameter includes double quotes and a space)<\/p><\/td>\n    <td><p>\"C:\\TEST A\\\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td><p>\\\"C:\\TEST A\\\\\\\"\n             <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose the whole parameter in double quotes<br \/>(because there's a space in the parameter)<\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"\\\"C:\\TEST A\\\\\\\"\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n             <br \/>\"\\\"C:\\TEST A\\\\\\\"\"\n          <\/p>\n       <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n      <td><p>No special characters to escape<br \/>(as seen by cmd.exe)<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"\\\"C:\\TEST A\\\\\\\"\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 2a just escape all special characters including\n                         all double quotes so there are no double quoted parts.<\/p><\/td>\n    <td><p>^\"\\^\"C:\\TEST A\\\\\\^\"^\"\n           <br \/>\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><th colspan=\"3\"><p>Example 8.9b:\u00a0\u00a0\u00a0\u00a0\u00a0\"C:\\TEST A\\\"<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want (same as 4a)<\/p><\/td>\n    <td><p>\"C:\\TEST A\\\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td><p>\\\"C:\\TEST A\\\\\\\"\n             <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose spaces in double quotes\n        <br \/>(A double quoted part can be anywhere within a parameter)<\/td>\n    <td class=\"comt\">\n        <p>\\\"C:\\TEST\" \"A\\\\\\\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u250c\u2500\u2500\u2500\u2500\u2510\n             \\\"C:\\TEST\" \"A\\\\\\\"\n          <\/p>\n       <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n      <td><p>It's all in double quoted parts.<br \/>(as seen by cmd.exe)<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>no % in parameter<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\\\"C:\\TEST\" \"A\\\\\\\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 2a just escape all special characters including\n                         all double quotes so there are no double quoted parts.<\/p><\/td>\n    <td><p>\\^\"C:\\TEST^\" ^\"A\\\\\\^\"\n           <br \/>\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<b>Sample Test File:<\/b>\n<p>\nExample 8.10: Command Line to C\/C++ Program\n<\/p><div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 8.10a:\u00a0\u00a0\u00a0\u00a0\u00a0\"C:\\TEST %&amp;^ A\\\"<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<\/p><\/td>\n    <td><p>\"C:\\TEST %&amp;^ A\\\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td>\n        <p>\\\"C:\\TEST %&amp;^ A\\\\\\\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose the whole parameter in double quotes<\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"\\\"C:\\TEST %&amp;^ A\\\\\\\"\"\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n             <br \/>\"\\\"C:\\TEST %&amp;^ A\\\\\\\"\"\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n    <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n           <br \/>\"\\\"C:\\TEST ^%^&amp;^^ A\\\\\\\"\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>\u250c\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2510\n           <br \/>\"\\\"C:\\TEST ^%%^&amp;^^ A\\\\\\\"\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n     <\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"\\\"C:\\TEST ^%%^&amp;^^ A\\\\\\\"\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 2a just escape all special characters including\n                         all double quotes so there are no double quoted parts.<\/p><\/td>\n    <td><p>^\"\\^\"C:\\TEST %^&amp;^^ A\\\\\\^\"^\"\n           <br \/>\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">and if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>^\"\\^\"C:\\TEST %%^&amp;^^ A\\\\\\^\"^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><th colspan=\"3\"><p>Example 8.10b:\u00a0\u00a0\u00a0\u00a0\u00a0\"C:\\TEST %&amp;^ A\\\"<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want (same as 5a)<\/p><\/td>\n    <td><p>\"C:\\TEST %&amp;^ A\\\"<\/p><\/td>\n  <\/tr><tr><td class=\"combwt\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Microsoft C\/C++ parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\"><table class=\"comt\"><tr><td>replace:<\/td>\n      <td>\n            literal\u00a0\u00a0\"\u00a0\u00a0with\u00a0\u00a0\u00a0\\\"\n        <br \/>literal\u00a0\\\"\u00a0\u00a0with\u00a0\u00a0\\\\\\\"\n        <br \/>literal \\\\\" with \\\\\\\\\\\"\n        <\/td><\/tr><\/table><\/td>\n      <td>\n        <p>\\\"C:\\TEST %&amp;^ A\\\\\\\"\n           <br \/>\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">enclose spaces in double quotes\n        <br \/>(A double quoted part can be anywhere within a parameter)<\/td>\n    <td><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\u00a0\u00a0\u00a0\u250c\u2500\u2510\n           <br \/>\\\"C:\\TEST\" \"%&amp;^\" \"A\\\\\\\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u2191\u00a0\u2191\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"combwt\">2.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Prompt parsing rules (cmd.exe)<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts<\/td>\n      <td><p>\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u250c\u2500\u2500\u2500\u2510\u00a0\u250c\u2500\u2500\u2500\u2500\u2510\n             <br \/>\\\"C:\\TEST\" \"%&amp;^\" \"A\\\\\\\"\n        <\/p>\n     <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n      <td class=\"comt\">escape the special characters not in double quoted parts:\n           <br \/>( the escape character for cmd.exe is ^ )\n      <\/td>\n      <td><p>it's all in double quoted parts<br \/>(as seen by cmd.exe)<\/p>\n  <\/td><\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n      <td><p>\\\"C:\\TEST\" \"%%&amp;^\" \"A\\\\\\\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n     <\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"\\\"C:\\TEST ^%%^&amp;^^ A\\\\\\\"\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><hr \/><\/td><\/tr><tr><td colspan=\"3\"><p>Another way to get the same result would be\n                         at step 2a just escape all special characters including\n                         all double quotes so there are no double quoted parts.<\/p><\/td>\n    <td><p>\\^\"C:\\TEST^\" ^\"%^&amp;^^^\" ^\"A\\\\\\^\"\n           <br \/>\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u00a0\u2191\u00a0\u2191\u00a0\u2191\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td class=\"comwt\">c.<\/td>\n    <td class=\"comt\">and if this will be placed in a batch file, double the\u00a0%\u00a0characters<\/td>\n    <td><p>\\^\"C:\\TEST^\" ^\"%%^&amp;^^^\" ^\"A\\\\\\^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div><\/div>\n<br \/><b>Sample Test File:<\/b>\n<hr \/><p>8.2\n<\/p><div style=\"text-align: center;\">How To Pass A Parameter to:\n<br \/>a Batch File from the Command Line<\/div>\n<br \/><p>To get a parameter into a batch file you need to work backwards through the\ntwo parsings it will go through:<\/p>\n<ol><li>The cmd.exe command line parser parses your command line &amp; parameters and then starts running the batch file\n  <\/li><li>The batch file retrieves the parameters, but when you use them, they get parsed again. Here's an example:\n    <ol class=\"ms12la\"><li>a batch file line contains %1\n      <\/li><li>%1 is replaced with your parameter\n      <\/li><li>that line is then sent to the command line parser for execution, where it gets parsed\n    <\/li><\/ol><\/li><\/ol><b>Overview of steps:<\/b>\n<ol><li>Start with the parameter you want\n  <\/li><li>Apply the command line parsing rules that cmd.exe will apply when you enter the command and parameters\u00a0\u00a0[<a href=\"#WINCMDRULES\">sec. 7<\/a>]\n  <\/li><li>Again, apply the command line parsing rules that cmd.exe will apply when you enter the command and parameters\u00a0\u00a0[<a href=\"#WINCMDRULES\">sec. 7<\/a>]\n<\/li><\/ol><b>Step Details:<\/b>\n<br \/>The steps to create your parameter are:\n<div class=\"ms12\"><ol><li>Apply the Command Line Parser (cmd.exe) parsing rules\n  <ol><li>determine what cmd.exe will see as the quoted parts\n    <\/li><li>escape the special characters &lt; &gt; | &amp; ^ not in a double quoted part\n    <\/li><li>again, escape the special characters &lt; &gt; | &amp; ^ not in a double quoted part\n      <ul><li>\u2022the escape character for cmd.exe is ^\n        <\/li><li>\u2022it may also be necessary to escape ( ) @ !\n        <\/li><li>\u2022it's OK to escape everything\n      <\/li><\/ul><\/li><li>If your command line will be placed in a batch file, double the % character<a name=\"S7\" href=\"#R7\" class=\"footnote\" id=\"S7\"><sup>7<\/sup><\/a>\n  <\/li><\/ol><\/li><\/ol><\/div>\n<br \/><b>The Command Line to Batch File Rules:<\/b>\n<div class=\"ms12\">Combining the two steps above gives us the following rules:<ul><li>Use ^^^ outside double quotes to escape &lt; &gt; &amp; | ^\n<br \/>\u00a0\u00a0\u00a0to insert &lt; use ^^^&lt;\n<br \/>\u00a0\u00a0\u00a0to insert &gt; use ^^^&gt;\n<br \/>\u00a0\u00a0\u00a0to insert | use ^^^|\n<br \/>\u00a0\u00a0\u00a0to insert ^ use ^^^^\n<\/li>\n<li>Sometimes it may be necessary to also escape ( ) @ !<\/li>\n<li>It's OK to escape everything<\/li>\n<li>Note ^ is also the line continuation character.<\/li>\n<\/ul><\/div>\n<br \/><b>Space, Tab, Semicolon, Comma:<\/b>\n<br \/>Note that you can't have a space within a parameter without having some\ndouble quotes around it somehow. All spaces must be within a double\nquoted part. Otherwise the space will act as a parameter delimiter. (Same with tabs, semicolons, commas.)\n<ol><li>Command line arguments can be separated by an unquoted semicolon or comma. This affects builtin commands like COPY and batch file parameters.\n<\/li><li>If the command is called by name (no path), the following slash is treated a separator, as in DIR\/P. This affects both builtin commands and external commands.\nNote: If you want to use NT-style slash-separated paths with DOS utilities you have to quote them (as in TYPE \"C:\/BOOT.INI\").<a name=\"S16\" href=\"#R16\" class=\"footnote\" id=\"S16\"><sup>16<\/sup><\/a>\n<\/li><\/ol><br \/><div class=\"ContentAreaB\"><div class=\"topic\"><p>Example 9.1: Command Line to Batch File<\/p><div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\">Example 9.1:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|()@!<\/th>\n    <th>Parameter<\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<\/p><\/td>\n    <td><p>&amp;&lt;&gt;^|()@!<\/p><\/td>\n  <\/tr><tr><td class=\"comb\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Line to Batch File parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts:<\/td>\n      <td><p>it's all unquoted<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td>\n    <td class=\"comt\">escape the special characters not in double quoted parts:<\/td>\n    <td><p>^^^&amp;^^^&lt;^^^&gt;^^^^^^^|()@!\n           <br \/>\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\n        <\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>^^^&amp;^^^&lt;^^^&gt;^^^^^^^|()@!<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div><b>Sample Test File:<\/b><p>Example 9.2: Command Line to Batch File<\/p><div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 9.2:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|()@!<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<\/p><\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"&amp;&lt;&gt;^|()@!\"<\/p><\/td>\n  <\/tr><tr><td class=\"comb\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Line to Batch File parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts:<\/td>\n      <td><p>it's all in a double quoted part<\/p><\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td><td class=\"comt\">escape the special characters not in double quoted parts:<\/td>\n    <td><p>Nothing to escape because it's all in a double quoted part<br \/>(as seen by cmd.exe)<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"&amp;&lt;&gt;^|()@!\"<\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div><b>Sample Test File:<\/b><p>Example 9.3: Command Line to Batch File<\/p><div class=\"tablemargin\"><div class=\"ContentAreaB\"><div class=\"topic\"><table class=\"comtb\"><tr><th colspan=\"3\"><p>Example 9.3:\u00a0\u00a0\u00a0\u00a0\u00a0&amp;&lt;&gt;^|()@!\"&amp;&lt;&gt;^|()@!<\/p><\/th>\n    <th><p>Parameter<\/p><\/th>\n  <\/tr><tr><td colspan=\"3\"><p>Start with the parameter you want<\/p><\/td>\n    <td><p>&amp;&lt;&gt;^|()@!\"&amp;&lt;&gt;^|()@!<\/p><\/td>\n  <\/tr><tr><td class=\"comb\">1.<\/td><td colspan=\"2\" class=\"comb\">Apply the Command Line to Batch File parsing rules<\/td><\/tr><tr><td class=\"comwt\">a.<\/td>\n      <td class=\"comt\">determine what cmd.exe will see as the quoted parts:<\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\n             <br \/>\"&amp;&lt;&gt;^|()@!\"&amp;&lt;&gt;^|()@!\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"2\"><\/td><td><p>we have a problem in that the final\u00a0\"\u00a0is interpreted\n             by cmd.exe as opening a double quoted part. To\n             avoid this, escape that last\u00a0\"\u00a0<\/p><\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n             <br \/>\"&amp;&lt;&gt;^|()@!\"&amp;&lt;&gt;^|()@!^^^\"\n             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\n          <\/p>\n      <\/td>\n  <\/tr><tr><td class=\"comwt\">b.<\/td><td class=\"comt\">escape the special characters not in double quoted parts:<\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"&amp;&lt;&gt;^|()@!\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|()@!^^^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\n        <\/p>\n    <\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Result: To get desired parameter use this:<\/p><\/td>\n    <td><p>\"&amp;&lt;&gt;^|()@!\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|()@!^^^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>Though not necessary, it's OK to escape the rest of the characters:<\/p><\/td>\n    <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n           <br \/>\"&amp;&lt;&gt;^|()@!\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|^^^(^^^)^^^@^^^!^^^\"\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\n        <\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>An easier way to get the same result would be to escape\n                         the\u00a0\"\u00a0within the parameter so it doesn't end the double\n                         quoted part. Then nothing else needs to be escaped since\n                         it's all in a double quoted part.<\/p><\/td>\n      <td><p>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n             <br \/>\"&amp;&lt;&gt;^|()@!\\\"&amp;&lt;&gt;^|()@!\"\n             <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191<\/p><\/td>\n  <\/tr><tr><td colspan=\"3\"><p>\nAnother way to get the same result would be to escape all the special\ncharacters including all the\u00a0\"\u00a0so nothing is in a double quoted part:\n                         <\/p><\/td>\n    \n  <\/tr><tr><td colspan=\"4\"><p>^^^\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|()@!^^^\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|()@!^^^\"<\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><p>As before, though it's not necessary, it is OK to escape the rest of the characters:\n    <\/p><\/td>\n  <\/tr><tr><td colspan=\"4\"><p>^^^\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|^^^(^^^)^^^@^^^!^^^\"^^^&amp;^^^&lt;^^^&gt;^^^^^^^|^^^(^^^)^^^@^^^!^^^\"\n    <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\u00a0\u2191\u2191\u2191\n        <\/p><\/td>\n  <\/tr><\/table><\/div><\/div><\/div><b>Sample Test Files:<\/b><br \/>all 5 test files should produce the same result<\/div><\/div>\n<hr \/><p>10.\n<\/p><div style=\"text-align: center;\">How To Pass A Parameter to:\n<br \/>a VBScript, JScript, or WSH Script<br \/>from the Command Line<\/div>\nVBScript (.vbs), JScript<a name=\"S9\" href=\"#R9\" class=\"footnote\" id=\"S9\"><sup>9<\/sup><\/a> (.js),\nand WSH (.wsh) scripts are run by program wscript.exe,\nwhich is the Microsoft\u00ae Windows\u00ae Script Host (WSH).\nWSH is a language-independent scripting host for Windows Script compatible scripting engines.\nMicrosoft provides both Microsoft Visual Basic Script and JScript scripting engines with WSH.\nWindows Script Host executes scripts that exist outside an HTML or ASP page and that stand on their own as text files.<a name=\"S10\" href=\"#R10\" class=\"footnote\" id=\"S10\"><sup>10<\/sup><\/a>\n<pre>\n     Example:\n      &gt; C:\\WINDOWS\\system32\\wscript.exe ShowParams.vbs hello goodbye Friday\n<\/pre>\nWhen you run ShowParams.vbs you'll notice the window title says, \"Window Script Host\".\n<div style=\"text-align: center;\">\n\n<table cellpadding=\"10\" style=\"width: 75%;\"><tr><td class=\"verdana\">\n\u201cIf you want to get picky, the truth is that you <i>can\u2019t<\/i> read\ncommand-line arguments using VBScript; that\u2019s because VBScript\ndoesn\u2019t know anything about command-line arguments. But\nthat\u2019s all right; after all, VBScript doesn\u2019t <i>have<\/i>\nto know anything about command-line arguments. That\u2019s because\nWindows Script Host takes care of all that stuff.\n<p>\n\u201cAny time you supply a command-line argument to a script that runs under\nWindows Script Host (that includes JScript scripts as well as VBScript\nscripts) those arguments are automatically stored in the Wscript.Arguments\ncollection.\u201d\n<\/p><p>\u00a0\u00a0\u00a0\u00a0\u00a0\u2014<small>Hey, Scripting Guy! 2008 Winter Scripting Games: Retrieving Command-Line Arguments\n<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0<a href=\"http:\/\/www.microsoft.com\/technet\/scriptcenter\/funzone\/games\/tips08\/gtip0104.mspx\" target=\"_blank\">http:\/\/www.microsoft.com\/technet\/scriptcenter\/funzone\/games\/tips08\/gtip0104.mspx<\/a>\n<\/small>\n<\/p><\/td>\n<td><img src=\"http:\/\/daviddeley.com\/autohotkey\/parameters\/drscripto.png\" alt=\"drscripto\" \/><\/td>\n<\/tr><\/table><\/div>\n<p>On startup wscript.exe calls GetCommandLine() to\nget the command line, then calls wscript!SplitCommandLine(),\nwhich parses off the parameters. The rules for WSH command line parameter\nparsing are simple:\n<\/p><p>10.1\u00a0\u00a0The WSH Command Line Parameter Parsing Rules:<\/p>\n<div class=\"ms12\"><ul><li>parameters are separated by a space or tab (multiple spaces\/tabs OK)<\/li>\n<li>\" begins and ends the double quoted part)<\/li>\n<li>all characters within a double quoted part are accepted, including spaces and tabs)<\/li>\n<li>The \" character itself is always stripped from the parameter)\n<br \/>(Note this means you can not pass a double quote as part of a parameter.))<\/li>\n<\/ul><\/div>\n<p>10.2\u00a0\u00a0The Microsoft\u00ae Windows\u00ae Script Host (WSH) Command Line Parameter Parsing Algorithm:<\/p>\nThe following algorithm was reverse engineered by disassembling and examining\nwscript.exe and cscript.exe:\n<p>Algorithm:<\/p>\n<ol><li>call Kernel32.dll!GetCommandLine() to get the command line<\/li>\n<li>make a copy of the command line<\/li>\n<li>call wscript!SplitCommandLine() dry run to determine the number of parameters<\/li>\n<li>allocate space for argv[] array<\/li>\n<li>call wscript!SplitCommandLine() to parse parameters and fill in argv[] array<\/li>\n<\/ol>\nwscript!SplitCommandLine() builds an argv[] array\nof where each parameter begins, and writes a terminating NULL at the end of each parameter.\n<div class=\"ContentArea\"><div class=\"topic\"><table><tr><th><p>Splitting the Command Line to get Parameters<\/p><\/th>\n  <\/tr><tr><td><p>    start with:\u00a0ShowParams.exe hello goodbye Friday\n           <br \/>end with:\u00a0\u00a0\u00a0ShowParams.exe0hello0goodbye0Friday0\n           <br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2191\n           <br \/>\u00a0argv[]\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n           <br \/>param 0: \u2500\u2500\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n           <br \/>param 1: \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n           <br \/>param 2: \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u2502\n           <br \/>param 3: \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n        <\/p>\n    <\/td>\n  <\/tr><\/table><\/div><\/div>\n<pre>\nloop\n   parse off next parameter:\n   skip over spaces, tabs\n   clear \" flag\n   save starting address of this parameter\n   LOOP\n      process this character:\n      If space or tab\n         if \" flag set\n            accept this space or tab as part of the parameter\n         else\n            write a 0 here to terminate this parameter\n            and goto parse off next parameter\n      Else if \"\n         toggle \" flag, strip \" (shift rest of line left 1 char)\n      move to next char\n   ENDLOOP\nendloop\n<\/pre>\n<div class=\"ms12\"><b>Notes:<\/b><ul><li>  \" flag - set if we're currently inside a double quoted part\n<\/li><li>The algorithm only looks for {space}, {tab}, and (\"). All other characters are just characters.\n<\/li><li>You can't pass a double quote (\") as part of a parameter because double quotes are always stripped off.\n<\/li><\/ul><\/div>\n<br \/><a name=\"WSHRULES\" id=\"WSHRULES\">10.3\u00a0\u00a0Putting it together:<\/a>\n<br \/>When you launch a VBScript, JScript, or WSH script from the command line,\nyour command line goes through two parsers:\n<ol><li>First the cmd.exe command line parser parses your command &amp; parameters, using its cmd.exe parsing rules. It builds a command line, and calls CreateProcess() passing it the command line it built.\n  <\/li><li>Then the Windows Script Host retrieves that resulting command line and parses off the parameters, using its WSH parsing rules.\n<\/li><\/ol>\nSo to get a parameter into a VBScript, JScript, or WSH script, you need to\nwork backwards through the two parsers it will go through:\n<ol><li>Start with the parameter you want\n  <\/li><li>Apply the WSH parsing rules that Windows Script Host will apply when parsing the command line it retrieves\n  <\/li><li>Then apply the command line parsing rules that cmd.exe will apply when you enter the command and parameters\n<\/li><\/ol><b>Step Details:<\/b>\n<br \/>The steps to create your parameter are:\n<div class=\"ms12\"><ol><li>Apply the WSH parsing rules:\n  <ol><li>if there are any spaces or tabs within a parameter, add some \" so they are enclosed in a double quoted part\n      <ul><li>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n          <br \/>\u2022you may enclose the entire parameter in double quotes: \"Call Me Ishmael\"\n        <\/li><li>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2510\u00a0\u00a0\u250c\u2500\u2510\n          <br \/>\u2022or you may enclose just the spaces in double quotes: Call\" \"Me\" \"Ishmael\n        <\/li><li>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n          <br \/>\u2022or you can make a mess of it: Ca\"ll Me Is\"hmael\n      <\/li><\/ul><\/li><\/ol><br \/><\/li><li>Apply the Command Line Parser (cmd.exe) parsing rules\n  <ol><li>determine what cmd.exe will see as the double quoted parts<\/li>\n    <li>escape the special characters &lt; &gt; | &amp; ^ not in a double quoted part\n      <ul><li>\u2022the escape character for cmd.exe is ^<\/li>\n        <li>\u2022it may also be necessary to escape ( ) @ !<\/li>\n        <li>\u2022it's OK to escape everything<\/li>\n      <\/ul><\/li>\n    <li>If your command line will be placed in a batch file, double the % character<a name=\"S11\" href=\"#R11\" id=\"S11\"><sup>11<\/sup><\/a><\/li>\n  <\/ol><\/li><\/ol><\/div>\n<p>10.4\u00a0\u00a0Sample Scripts:<\/p>\n<br \/>Sample 'Show Parameters' scripts:\n<p>VBScript:\n<\/p>\nJScript:\n<br \/><hr \/><p>11.<\/p><div style=\"text-align: center;\">How To Pass A Parameter to\n<br \/>a Perl script from the command line<\/div>\nActivePerl<a name=\"S12\" href=\"#R12\" class=\"footnote\" id=\"S12\"><sup>12<\/sup><\/a> and Strawberry Perl<a name=\"S13\" href=\"#R13\" class=\"footnote\" id=\"S13\"><sup>13<\/sup><\/a> both call MSVCRT.dll!__getmainargs\nwhich follows the Microsoft C\/C++ Parameter Parsing Rules.\nTherefore the steps are:\n<ol><li>Start with the parameter you want\n  <\/li><li>Apply the Microsoft C\/C++ parsing rules that ShowParams.exe will apply when parsing the command line it retrieves\u00a0\u00a0[<a href=\"#CPP\">sec. 5<\/a>]\n  <\/li><li>Apply the command line parsing rules that cmd.exe will apply when you enter the command and parameters\u00a0\u00a0[<a href=\"#WINCMDRULES\">sec. 7<\/a>]\n<\/li><\/ol>\nSee sec. 8: <a href=\"#CPP\">How To Pass A Parameter to: a C\/C++ Program from the Command Line<\/a> above.\n<p>Sample 'Show Parameters' Perl script:\n<\/p><br \/><hr \/><p>12.<\/p><div style=\"text-align: center;\">How To Pass A Parameter to\n<br \/>a Python script from the command line<\/div>\n<br \/>Python\u2122<a name=\"S14\" href=\"#R14\" class=\"footnote\" id=\"S14\"><sup>14<\/sup><\/a>\nis written in C and obtains its parameters via the argv[] vector.<a name=\"S15\" href=\"#R15\" class=\"footnote\" id=\"S15\"><sup>15<\/sup><\/a>\nIt thus uses the Microsoft C\/C++ Parameter Parsing Rules.\nTherefore the steps are:\n<ol><li>Start with the parameter you want\n  <\/li><li>Apply the Microsoft C\/C++ parsing rules that ShowParams.exe will apply when parsing the command line it retrieves\u00a0\u00a0[<a href=\"#CPP\">sec. 5<\/a>]\n  <\/li><li>Apply the command line parsing rules that cmd.exe will apply when you enter the command and parameters\u00a0\u00a0[<a href=\"#WINCMDRULES\">sec. 7<\/a>]\n<\/li><\/ol>\nSee sec. 8: <a href=\"#CPP\">How To Pass A Parameter to: a C\/C++ Program from the Command Line<\/a> above.\n<p>Sample 'Show Parameters' Python script for Python version 2:\n<\/p>\nSample 'Show Parameters' Python script for Python version 3:\n<p>Note: Python version 3 is incompatible with Python version 2. Guido van\nRossum wrote \"Python 3.0, also known as Python 3000 or Py3K, is the first\never intentionally backwards incompatible Python release.\"  In\nparticular, print statements that would work with pre-3 versions of Python\nwill not work in 3.x.\nThe change of \"print\" from a statement to a function is identified as a\ncommon stumbling block on this page on what's new in Python 3.0.\n<a href=\"http:\/\/docs.python.org\/3.1\/whatsnew\/3.0.html\" target=\"_blank\">http:\/\/docs.python.org\/3.1\/whatsnew\/3.0.html<\/a><\/p>\n<br \/><hr \/><p>13.<\/p><div style=\"text-align: center;\">How To Pass A Parameter to\n<br \/>a REXX script from the command line<\/div>\n<br \/><span>8\/24\/2009 Under Construction...<\/span>\n\n<p>Sample 'Show Parameters' REXX script:\n<\/p><br \/><hr \/><p>13.<\/p><div style=\"text-align: center;\">How To Pass A Parameter to\n<br \/>a RUBY script from the command line<\/div>\n<br \/>Ruby attempts to mimic the Microsoft C\/C++ parameter parsing rules, but the current\nversions, 1.8.6 &amp; 1.9.1, have a bug in them so they don't work properly.\n<p>(file Win32.c routine rb_w32_cmdvector, the\nmemcpy statement needs to copy one more character to include the trailing NULL.)\n<\/p><p>Sample 'Show Parameters' RUBY script:\n<\/p><br \/><hr \/><p>15.<\/p><div style=\"text-align: center;\">How To Pass A Parameter to\n<br \/>an AutoHotkey script from the command line<\/div>\n<br \/>AutoHotkey is written in C++ and obtains its parameters via the __argv[] vector.\nThus it uses the Microsoft C\/C++ Parameter Parsing Rules.\nTherefore the steps are:\n<ol><li>Start with the parameter you want\n  <\/li><li>Apply the Microsoft C\/C++ parsing rules that ShowParams.exe will apply when parsing the command line it retrieves\u00a0\u00a0[<a href=\"#CPP\">sec. 5<\/a>]\n  <\/li><li>Apply the command line parsing rules that cmd.exe will apply when you enter the command and parameters\u00a0\u00a0[<a href=\"#WINCMDRULES\">sec. 7<\/a>]\n<\/li><\/ol>\nSee sec. 8: <a href=\"#CPP\">How To Pass A Parameter to: a C\/C++ Program from the Command Line<\/a> above.\n<p>Sample 'Show Parameters' AutoHotkey script:\n<\/p><hr style=\"text-align: left;\" \/><div class=\"footnotes\">Footnotes:<br \/><table class=\"footnotes\"><tr><td align=\"right\">\n<a name=\"R1\" href=\"#S1\" id=\"R1\">1.<\/a>\n<\/td><td>Larry Osterman's WebLog <i>The Windows command line is just a string...<\/i>\n<br \/><span><a href=\"http:\/\/blogs.msdn.com\/larryosterman\/archive\/2007\/10\/03\/the-windows-command-line-is-just-a-string.aspx\" target=\"_blank\">http:\/\/blogs.msdn.com\/larryosterman\/archive\/2007\/10\/03\/the-windows-command-line-is-just-a-string.aspx<\/a><\/span>\n\n<p class=\"t\">Thank you to Delan Azabani for the following explanation:<\/p>\n<p>Technically, execl(3) and friends\ndo <i>not<\/i> allow you to specify a command line string<a name=\"Srefi\" href=\"#Rrefi\" class=\"footnote\" id=\"Srefi\"><sup> [i]<\/sup><\/a>,\nbut rather a sequence of previously separated arguments via variadic C arguments.\nexecv(3) and friends are similar, except that they use decayed arrays.<\/p>\n<p>The complete exec family includes exec{l,le,lp,v,ve,vp}, fexecve(3),\nand if you\u2019re using glibc, the execvpe(3) extension.<\/p>\n<p>The family may be <i>invoked<\/i> in a way that allows you to specify a\ncommand line string \u2014 including, for example, together with sh(1):<\/p>\n<p class=\"bq\">execl(\"\/path\/to\/bin\/sh\", \"sh\", \"-c\", command_line, (char *) NULL);<\/p>\n<p>This is roughly how system(3) is specified by POSIX<a name=\"Srefii\" href=\"#Rrefii\" class=\"footnote\" id=\"Srefii\"><sup> [ii]<\/sup><\/a>.<\/p>\n<p><a name=\"Rrefi\" id=\"Rrefi\">[i]<\/a> <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/functions\/exec.html\">http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/functions\/exec.html<\/a><br \/><a name=\"Rrefii\" id=\"Rrefii\">[ii]<\/a> <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/functions\/system.html\">http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/functions\/system.html<\/a><\/p>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R2\" href=\"#S2\" id=\"R2\">2.<\/a>\n<\/td><td><i>ibid.<\/i>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R3\" href=\"#S3\" id=\"R3\">3.<\/a>\n<\/td><td>or possibly CScript.exe; WScript.exe is the Windows version, CScript.exe is the console version. You'll find them in the System32 directory.\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R4\" href=\"#S4\" id=\"R4\">4.<\/a>\n<\/td><td>In some cases cmd.exe may call\nShellExecute() which\neventually calls\nCreateProcess(), passing it a command line.\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R5\" href=\"#S5\" id=\"R5\">5.<\/a>\n<\/td><td>On Windows 95,98,ME the Command Prompt Window program was COMMAND.COM .\nThe parsing rules are the same.\n(Back then it was also known as the DOS Prompt, or MS-DOS Prompt. Technically COMMAND.COM\nis DOS.\nNote there are other Command Prompt Window programs such as JPSoft's \"Take Command\". (<a href=\"http:\/\/jpsoft.com\" target=\"_blank\">http:\/\/jpsoft.com<\/a>)\n<br \/>Note the following CP\/M legacy constructs:\n1. Command line arguments can be separated by an unquoted semicolon or comma. This affects builtin commands like COPY and batch file parameters.\nIn particular, I can think of no way to pass A;B (without quotes) as a paramater to a batch file.\n2. If the command is called by name (no path), the following slash is treated a separator, as in DIR\/P. This affects both builtin commands and external commands.\nNote: If you want to use NT-style slash-separated paths with DOS utilities you have to quote them (as in TYPE \"C:\/BOOT.INI\").\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R6\" href=\"#S6\" id=\"R6\">6.<\/a>\n<\/td><td>It's been documented in the past, such as here:\n<br \/>\"Percent Signs Stripped from Batch File Text\" <a href=\"http:\/\/support.microsoft.com\/kb\/75634\" target=\"_blank\">http:\/\/support.microsoft.com\/kb\/75634<\/a>\n<br \/>and it's mentioned in current documentation here (last sentence of the \"Copy Examples\" section):\n<br \/>\"dtutil Utility\"\n<a href=\"http:\/\/msdn.microsoft.com\/en-us\/library\/ms162820.aspx\" target=\"_blank\">http:\/\/msdn.microsoft.com\/en-us\/library\/ms162820.aspx<\/a>\n<br \/>Note: JPSoft's Take Command command line utility processes % differently. You'll need to quadruple the % char (<a href=\"http:\/\/jpsoft.com\" target=\"_blank\">http:\/\/jpsoft.com<\/a>)\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R7\" href=\"#S7\" id=\"R7\">7.<\/a>\n<\/td><td>See <a href=\"#R6\">footnote 6<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R8\" href=\"#\" id=\"R8\">8.<\/a>\n<\/td><td>See <a href=\"#R6\">footnote 6<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R9\" href=\"#S9\" id=\"R9\">9.<\/a>\n<\/td><td>JScript is Microsoft's own version of JavaScript.\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R10\" href=\"#S10\" id=\"R10\">10.<\/a>\n<\/td><td>\"Windows Script Host Basics\" <a href=\"http:\/\/msdn.microsoft.com\/en-us\/library\/ec0wcxh3(VS.85).aspx\" target=\"_blank\">http:\/\/msdn.microsoft.com\/en-us\/library\/ec0wcxh3(VS.85).aspx<\/a>\n<br \/>\"Description of Windows Script Host (WSH)\" <a href=\"http:\/\/support.microsoft.com\/kb\/188135\" target=\"_blank\">http:\/\/support.microsoft.com\/kb\/188135<\/a>\n<br \/>\"Hosting Environments and Script Engines\" <a href=\"http:\/\/msdn.microsoft.com\/en-us\/library\/s4axe076(VS.85).aspx\" target=\"_blank\">http:\/\/msdn.microsoft.com\/en-us\/library\/s4axe076(VS.85).aspx<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R11\" href=\"#S11\" id=\"R11\">11.<\/a>\n<\/td><td>See <a href=\"#R6\">footnote 6<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R12\" href=\"#S12\" id=\"R12\">12.<\/a>\n<\/td><td><a href=\"http:\/\/www.activestate.com\/activeperl\/\" target=\"_blank\">http:\/\/www.activestate.com\/activeperl\/<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R13\" href=\"#S13\" id=\"R13\">13.<\/a>\n<\/td><td><a href=\"http:\/\/strawberryperl.com\/\" target=\"_blank\">http:\/\/strawberryperl.com\/<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R14\" href=\"#S14\" id=\"R14\">14.<\/a>\n<\/td><td><a href=\"http:\/\/www.python.org\/\" target=\"_blank\">http:\/\/www.python.org\/<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R15\" href=\"#S15\" id=\"R15\">15.<\/a>\n<\/td><td>Technically, WinMain in WinMain.c calls Py_main in main.c passing it __argc and __wargv as argc[] and argv[] respectively.\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R16\" href=\"#S16\" id=\"R16\">16.<\/a>\n<\/td><td>Thanks to Christopher Yeleighton for these two points.\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R17\" href=\"#S17\" id=\"R17\">17.<\/a>\n<\/td><td><a href=\"http:\/\/www.yolinux.com\/TUTORIALS\/GNOME.html#LAUNCHER\" target=\"_blank\">YoLinux.com: GNOME desktop basics \u2192 GNOME Desktop Launcher<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R18\" href=\"#S18\" id=\"R18\">18.<\/a>\n<\/td><td><a href=\"http:\/\/docs.kde.org\/stable\/en\/kdebase-workspace\/kmenuedit\/index.html\" target=\"_blank\">The KDE Menu Editor Handbook<\/a>\n<\/td><\/tr><tr><td align=\"right\">\n<a name=\"R19\" href=\"#S19\" id=\"R19\">19.<\/a>\n<\/td><td>Thanks to Andr\u00e1s Korn for assistance in updating the *nix section (Feb. 2011).\n<\/td><\/tr><\/table><\/div>\n<br \/><hr \/><br \/>David Deley \u00a9 2009\n<br \/><a href=\"http:\/\/daviddeley.com\/\" target=\"_blank\">http:\/\/daviddeley.com<\/a>\n<p>Last update: January 1, 2016<\/p>","created_at":"2018-11-07T16:52:22+0000","updated_at":"2018-11-07T16:52:22+0000","annotations":[],"mimetype":"text\/html","reading_time":59,"domain_name":"daviddeley.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":570,"title":"Electron on iOS \/ Android?","url":"https:\/\/discuss.atom.io\/t\/electron-on-ios-android\/18223\/9?_escaped_fragment_=","content":"<div class=\"post\" itemprop=\"articleBody\"><p>Taking in consideration that the iOS store prohibits the use of a different browser engine (chrome is just a GUI on top of the iOS webkit), I doubt this can be achieved as it is in that platform.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Yeah, that sucks to no end.<br \/>But at least Android should be able to get Atom\u2019s magic, sooner or later.<\/p>\n<p>Has anyone tried JXCore? It <em>should<\/em> be Node compatible, and ready for Android.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Considering we can now side load apps and most people using Atom could probably figure out how to side load a Xcode project, would it be a bad idea to develop it and just let people download it and side load it.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Hey, I actually came in from the other side. I built an iOS and Android app using ionic and have had great success converting them to desktop apps using electron. Things just worked.<br \/><a href=\"https:\/\/github.com\/pliablepixels\/zmNinja\" rel=\"nofollow noopener\">https:\/\/github.com\/pliablepixels\/zmNinja<\/a> (scroll down for desktop app links)<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Can you add a little more detail to this? Did you take your Ionic app as-is and compile it with Electron? Were there any conversion steps at all?<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Happy to.<br \/>The desktop preparation script is here <a href=\"https:\/\/github.com\/pliablepixels\/zmNinja\/blob\/master\/make_desktop.sh\" rel=\"nofollow noopener\">https:\/\/github.com\/pliablepixels\/zmNinja\/blob\/master\/make_desktop.sh<\/a><\/p>\n<p>Overall process:<br \/>a) I downloaded the electron packages from <a href=\"https:\/\/github.com\/electron\/electron\/releases\" rel=\"nofollow noopener\">here<\/a> and unzipped them to a directory<\/p>\n<p>b) As per their instructions, I have a<a href=\"https:\/\/github.com\/pliablepixels\/zmNinja\/tree\/master\/electron_js\" rel=\"nofollow noopener\"> main.js and packages.js<\/a> for my app . The main.js was modified to make it work reliably on my mac - but its the same file for all ports<\/p>\n<p>c) In my main ionic code, I had to find alternatives to cordova plugins (as they don\u2019t exist in desktop land). An example is <a href=\"https:\/\/github.com\/pliablepixels\/zmNinja\/blob\/master\/www\/js\/LogCtrl.js#L89\" rel=\"nofollow noopener\">here<\/a> - where I see if window.cordova exists (device) and do something vs. a diff. action on desktop. Or take <a href=\"https:\/\/github.com\/pliablepixels\/zmNinja\/blob\/master\/www\/js\/MonitorModalCtrl.js#L722-L746\" rel=\"nofollow noopener\">this example<\/a> (a better one), where I am saving an image of a live feed to my disk. If its mobile, I am using a cordova plugin, if its desktop, I\u2019m using a different JS plugin (canvas2plugin)<\/p>\n<p>d) And then the \u201cmake desktop\u201d script just copies the files inside the resource directory of the electron package, and then to keep it pretty, packs the app into an \u201casar\u201d archive - which is not necessary, but I like doing it - it also avoids newbie fiddlers from meddling with the code.<\/p>\n<p>Hope this helps.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Nice, great stuff. Thanks for sharing!<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p><a class=\"mention\" href=\"https:\/\/discuss.atom.io\/u\/leedohm\">@leedohm<\/a><br \/>In two or three weeks Android will officially run on desktops as ChromeOS is getting extended. Official support would be a good opportunity to run Electron apps on there.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Most likely,<a href=\"https:\/\/www.nativescript.org\/\" rel=\"nofollow noopener\"> Native Script<\/a> may help you integrate it with iOS\/Android.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>I actually build my Electron apps in Cordova (targetting the \u2018browser\u2019 platform), and maintain a file in my project that sets a bunch of functions that Cordova fulfils using plugins and Electron fulfils using Node.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>can you tell me more details? im trying to do the same here<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>There are still no plans to make Electron run on iOS or Android.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>If you\u2019re talking particularly about a Cordova\/Electron app, what kind of information are you after?<br \/>My app is Cordova-first, so it\u2019s structured as a standard Cordova app, I build out to the <code>browser<\/code> platform and build my Electron builds from the output (although there is an <code>Electron<\/code> platform which I\u2019ve not used yet).<\/p>\n<p>Then for both Cordova and Electron, I have a JS file that wraps all native functions (either Cordova plugins or Node libraries), so that I can use a single set of functions reliably across both Cordova and Electron.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>There is chrome and node.js on Android.<br \/>Why not support Android first?<br \/>Cordova, ionic, appcan, mui, even Xamarin will all not be the right way to develop mobile app if we have Electron to do this.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>I\u2019m not sure I understand your logic, <a class=\"mention\" href=\"https:\/\/discuss.atom.io\/u\/taoqf\">@taoqf<\/a>.<\/p>\n<p>How do you think an Android version of Electron would differ from Cordova? It\u2019s still a Chromium WebView wrapped in a native shell with native functionality. I can\u2019t see what reason there would be for this to be the \u2018wrong\u2019 way.<\/p>\n<p>Honestly, while you might be able to use Node on Android, I\u2019m not convinced you could use it in the way that you\u2019d want, at least as easily as you\u2019d like, as it\u2019d likely require polyfilling Node modules to fit the Android ecosystem.<\/p>\n<p>At the root of it, it\u2019s out of Electron\u2019s scope and it\u2019s definitely outside of Atom\u2019s.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>With termux <a href=\"https:\/\/termux.com\/\" rel=\"nofollow noopener\">https:\/\/termux.com\/<\/a> we can install nodejs on android <a href=\"https:\/\/www.youtube.com\/watch?v=lnlH_yvYwE0\" rel=\"nofollow noopener\">https:\/\/www.youtube.com\/watch?v=lnlH_yvYwE0<\/a><br \/>Could it be possible to wrap in a unique app, termux, nodejs a server and a webview which comunicates with the server to call the functions of nodejs? In this way we would use the node library instead of cordova plugin and then major portable code written for electron app.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><aside class=\"quote\" data-post=\"1\" data-topic=\"18223\"><div class=\"title\">\n<img alt=\"\" width=\"20\" height=\"20\" src=\"https:\/\/avatars.discourse.org\/v2\/letter\/j\/6f9a4e\/40.png\" class=\"avatar\" \/> jhoxray:<\/div>\n<blockquote>\n<p>bie type of question but i couldn\u2019t find anything definitive on the subject. Are there any plans to make Electron apps run on iOS \/ Android? Thank you!<\/p>\n<\/blockquote>\n<\/aside><p>Would be amazing with the upcoming Samsung Galaxy S8 , Windows Mobile Continuum clone. I mean there is no serious IDE for android.<\/p>\n<p>This would also enable electron \/ the IDEs to run on chromebooks through the android runner<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Intel scrapped crosswalk   ( which is chromium port for android by intel ) .<br \/>Its time to port electron to android.<br \/><\/p><aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"https:\/\/crosswalk-project.org\/assets\/favicon.ico\" class=\"site-icon\" width=\"97\" height=\"97\" alt=\"image\" \/><a href=\"https:\/\/crosswalk-project.org\/blog\/crosswalk-final-release.html\" target=\"_blank\" rel=\"nofollow noopener\">crosswalk-project.org<\/a>\n  <\/header><article class=\"onebox-body\"><img src=\"https:\/\/crosswalk-project.org\/\/assets\/crosswalk-og-banner.jpg\" class=\"thumbnail\" width=\"\" height=\"\" alt=\"image\" \/><h3><a href=\"https:\/\/crosswalk-project.org\/blog\/crosswalk-final-release.html\" target=\"_blank\" rel=\"nofollow noopener\">Crosswalk 23 to be the last Crosswalk release<\/a><\/h3>\n<p>Enable the most advanced web innovations with the Crosswalk Project web runtime to develop powerful Android and Cordova apps.<\/p>\n  <\/article>\n  \n<\/aside><br \/><aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"https:\/\/crosswalk-project.org\/assets\/favicon.ico\" class=\"site-icon\" width=\"97\" height=\"97\" alt=\"image\" \/><a href=\"https:\/\/crosswalk-project.org\/\" target=\"_blank\" rel=\"nofollow noopener\">crosswalk-project.org<\/a>\n  <\/header><article class=\"onebox-body\"><img src=\"https:\/\/crosswalk-project.org\/\/assets\/crosswalk-og-banner.jpg\" class=\"thumbnail\" width=\"\" height=\"\" alt=\"image\" \/><h3><a href=\"https:\/\/crosswalk-project.org\/\" target=\"_blank\" rel=\"nofollow noopener\">Crosswalk<\/a><\/h3>\n<p>Enable the most advanced web innovations with the Crosswalk Project web runtime to develop powerful Android and Cordova apps.<\/p>\n  <\/article>\n  \n<\/aside>\n<p>It have huge market  because:<\/p>\n<ul><li>Developing with cordova is a hell without crosswalk. Many incompatiblity problems across androids versions and device vendor modifed webviews.<\/li>\n<li>not everyone is running latest android version.<\/li>\n<li>android default webview is very unstable up to android 6.0<\/li>\n<li>not everyone does the system update of their rom all the .<\/li>\n<\/ul><\/div><div class=\"post\" itemprop=\"articleBody\"><p>I\u2019m sure the Electron repo would take the pull request <img src=\"https:\/\/discourse-cdn-sjc1.com\/business6\/images\/emoji\/twitter\/wink.png?v=5\" title=\":wink:\" class=\"emoji\" alt=\":wink:\" \/><\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>There are a group of us trying to move this forward here <a href=\"https:\/\/github.com\/gabrielcsapo\/electron-mobile\/issues\/1\" rel=\"nofollow noopener\">https:\/\/github.com\/gabrielcsapo\/electron-mobile\/issues\/1<\/a> If you are interested please come help us out!<\/p><\/div>","created_at":"2018-11-07T16:52:21+0000","updated_at":"2018-11-07T16:52:21+0000","published_at":"2016-05-11T18:31:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"discuss.atom.io","preview_picture":"https:\/\/avatars.discourse.org\/v2\/letter\/p\/cc9497\/200.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":569,"title":"Writing an OS in Rust","url":"https:\/\/os.phil-opp.com\/","content":"<p>In this post we set up the programmable interrupt controller to correctly forward hardware interrupts to the CPU. To handle these interrupts we add new entries to our interrupt descriptor table, just like we did for our exception handlers. We will learn how to get periodic timer interrupts and how to get input from the keyboard.<\/p>","created_at":"2018-11-07T16:52:20+0000","updated_at":"2018-11-07T16:52:20+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"os.phil-opp.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":568,"title":"Mutt - ArchWiki","url":"https:\/\/wiki.archlinux.org\/index.php\/Mutt","content":"<p><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mutt.org\/\">Mutt<\/a> is a text-based mail client renowned for its powerful features. Though over 2 decades old, Mutt remains the mail client of choice for a great number of power-users.\n<\/p><p>Mutt focuses primarily on being a Mail User Agent (MUA), and was originally written to view mail. Later implementations (added for retrieval, sending, and filtering mail) are simplistic compared to other mail applications and, as such, users may wish to use external applications to extend Mutt's capabilities. \n<\/p><p>Nevertheless, the Arch Linux <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=mutt\">mutt<\/a> package is compiled with IMAP, POP3 and SMTP support, removing the necessity for external applications.\n<\/p><p>This article covers using both native IMAP sending and retrieval, and a setup depending on <a href=\"https:\/\/wiki.archlinux.org\/index.php\/OfflineIMAP\" title=\"OfflineIMAP\">OfflineIMAP<\/a> or <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Getmail\" title=\"Getmail\">getmail<\/a> (POP3) to retrieve mail, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Procmail\" title=\"Procmail\">procmail<\/a> to filter it in the case of POP3, and <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Msmtp\" title=\"Msmtp\">msmtp<\/a> to send it.\n<\/p><div id=\"toc\" class=\"toc\"><p><\/p><h2>Contents<\/h2><ul><li class=\"toclevel-1 tocsection-1\"><a href=\"#Installation\">1 Installation<\/a>\n<ul><li class=\"toclevel-2 tocsection-2\"><a href=\"#NeoMutt\">1.1 NeoMutt<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-3\"><a href=\"#Configuration\">2 Configuration<\/a>\n<ul><li class=\"toclevel-2 tocsection-4\"><a href=\"#IMAP\">2.1 IMAP<\/a>\n<ul><li class=\"toclevel-3 tocsection-5\"><a href=\"#Native_IMAP_support\">2.1.1 Native IMAP support<\/a>\n<ul><li class=\"toclevel-4 tocsection-6\"><a href=\"#imap_user\">2.1.1.1 imap_user<\/a><\/li>\n<li class=\"toclevel-4 tocsection-7\"><a href=\"#imap_pass\">2.1.1.2 imap_pass<\/a><\/li>\n<li class=\"toclevel-4 tocsection-8\"><a href=\"#folder\">2.1.1.3 folder<\/a><\/li>\n<li class=\"toclevel-4 tocsection-9\"><a href=\"#spoolfile\">2.1.1.4 spoolfile<\/a><\/li>\n<li class=\"toclevel-4 tocsection-10\"><a href=\"#mailboxes\">2.1.1.5 mailboxes<\/a><\/li>\n<li class=\"toclevel-4 tocsection-11\"><a href=\"#Summary\">2.1.1.6 Summary<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-3 tocsection-12\"><a href=\"#External_IMAP_support\">2.1.2 External IMAP support<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-13\"><a href=\"#POP3\">2.2 POP3<\/a><\/li>\n<li class=\"toclevel-2 tocsection-14\"><a href=\"#Maildir\">2.3 Maildir<\/a><\/li>\n<li class=\"toclevel-2 tocsection-15\"><a href=\"#SMTP\">2.4 SMTP<\/a>\n<ul><li class=\"toclevel-3 tocsection-16\"><a href=\"#Folders\">2.4.1 Folders<\/a><\/li>\n<li class=\"toclevel-3 tocsection-17\"><a href=\"#Native_SMTP_support\">2.4.2 Native SMTP support<\/a><\/li>\n<li class=\"toclevel-3 tocsection-18\"><a href=\"#External_SMTP_support\">2.4.3 External SMTP support<\/a><\/li>\n<li class=\"toclevel-3 tocsection-19\"><a href=\"#Sending_mails_from_Mutt\">2.4.4 Sending mails from Mutt<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-20\"><a href=\"#Multiple_accounts\">2.5 Multiple accounts<\/a><\/li>\n<li class=\"toclevel-2 tocsection-21\"><a href=\"#Passwords_management\">2.6 Passwords management<\/a>\n<ul><li class=\"toclevel-3 tocsection-22\"><a href=\"#Security_concern\">2.6.1 Security concern<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-23\"><a href=\"#Tips_and_tricks\">3 Tips and tricks<\/a>\n<ul><li class=\"toclevel-2 tocsection-24\"><a href=\"#Key_bindings\">3.1 Key bindings<\/a><\/li>\n<li class=\"toclevel-2 tocsection-25\"><a href=\"#Composition\">3.2 Composition<\/a>\n<ul><li class=\"toclevel-3 tocsection-26\"><a href=\"#Encrypt_and_sign_mail_.28GnuPG.29\">3.2.1 Encrypt and sign mail (GnuPG)<\/a><\/li>\n<li class=\"toclevel-3 tocsection-27\"><a href=\"#E-mail_character_encoding\">3.2.2 E-mail character encoding<\/a><\/li>\n<li class=\"toclevel-3 tocsection-28\"><a href=\"#Custom_mail_headers\">3.2.3 Custom mail headers<\/a><\/li>\n<li class=\"toclevel-3 tocsection-29\"><a href=\"#Signature_block\">3.2.4 Signature block<\/a>\n<ul><li class=\"toclevel-4 tocsection-30\"><a href=\"#Random_signature\">3.2.4.1 Random signature<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-3 tocsection-31\"><a href=\"#Compose_and_send_from_command_line\">3.2.5 Compose and send from command line<\/a><\/li>\n<li class=\"toclevel-3 tocsection-32\"><a href=\"#Compose_HTML_e-mails\">3.2.6 Compose HTML e-mails<\/a><\/li>\n<li class=\"toclevel-3 tocsection-33\"><a href=\"#Display_another_email_while_composing\">3.2.7 Display another email while composing<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-34\"><a href=\"#Printing\">3.3 Printing<\/a><\/li>\n<li class=\"toclevel-2 tocsection-35\"><a href=\"#Viewing_content\">3.4 Viewing content<\/a>\n<ul><li class=\"toclevel-3 tocsection-36\"><a href=\"#Viewing_URLs_in_a_web_browser\">3.4.1 Viewing URLs in a web browser<\/a><\/li>\n<li class=\"toclevel-3 tocsection-37\"><a href=\"#Viewing_HTML\">3.4.2 Viewing HTML<\/a><\/li>\n<li class=\"toclevel-3 tocsection-38\"><a href=\"#Filtering_the_message_view\">3.4.3 Filtering the message view<\/a><\/li>\n<li class=\"toclevel-3 tocsection-39\"><a href=\"#Conversation_grouping\">3.4.4 Conversation grouping<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-40\"><a href=\"#Configuring_editors_to_work_with_mutt\">3.5 Configuring editors to work with mutt<\/a>\n<ul><li class=\"toclevel-3 tocsection-41\"><a href=\"#vim\">3.5.1 vim<\/a><\/li>\n<li class=\"toclevel-3 tocsection-42\"><a href=\"#GNU_nano\">3.5.2 GNU nano<\/a><\/li>\n<li class=\"toclevel-3 tocsection-43\"><a href=\"#Emacs\">3.5.3 Emacs<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-44\"><a href=\"#Display_settings\">3.6 Display settings<\/a>\n<ul><li class=\"toclevel-3 tocsection-45\"><a href=\"#Colors\">3.6.1 Colors<\/a><\/li>\n<li class=\"toclevel-3 tocsection-46\"><a href=\"#Index_Format\">3.6.2 Index Format<\/a><\/li>\n<li class=\"toclevel-3 tocsection-47\"><a href=\"#Display_recipient_instead_of_sender_in_.22Sent.22_folder_view\">3.6.3 Display recipient instead of sender in \"Sent\" folder view<\/a><\/li>\n<li class=\"toclevel-3 tocsection-48\"><a href=\"#Variable_column_width\">3.6.4 Variable column width<\/a><\/li>\n<li class=\"toclevel-3 tocsection-49\"><a href=\"#Sidebar\">3.6.5 Sidebar<\/a><\/li>\n<li class=\"toclevel-3 tocsection-50\"><a href=\"#Display_the_index_above_the_pager_view\">3.6.6 Display the index above the pager view<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-51\"><a href=\"#Contact_management\">3.7 Contact management<\/a>\n<ul><li class=\"toclevel-3 tocsection-52\"><a href=\"#Address_aliases\">3.7.1 Address aliases<\/a><\/li>\n<li class=\"toclevel-3 tocsection-53\"><a href=\"#Abook\">3.7.2 Abook<\/a><\/li>\n<li class=\"toclevel-3 tocsection-54\"><a href=\"#Goobook\">3.7.3 Goobook<\/a><\/li>\n<li class=\"toclevel-3 tocsection-55\"><a href=\"#Khard\">3.7.4 Khard<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-2 tocsection-56\"><a href=\"#Manage_multiple_sender_accounts\">3.8 Manage multiple sender accounts<\/a><\/li>\n<li class=\"toclevel-2 tocsection-57\"><a href=\"#Request_IMAP_mail_retrieval_manually\">3.9 Request IMAP mail retrieval manually<\/a><\/li>\n<li class=\"toclevel-2 tocsection-58\"><a href=\"#Avoiding_slow_index_on_large_.28IMAP.29_folders_due_to_coloring\">3.10 Avoiding slow index on large (IMAP) folders due to coloring<\/a><\/li>\n<li class=\"toclevel-2 tocsection-59\"><a href=\"#Speed_up_folders_switch\">3.11 Speed up folders switch<\/a><\/li>\n<li class=\"toclevel-2 tocsection-60\"><a href=\"#Archive_treated_e-mails\">3.12 Archive treated e-mails<\/a><\/li>\n<li class=\"toclevel-2 tocsection-61\"><a href=\"#Migrating_mails_from_one_computer_to_another\">3.13 Migrating mails from one computer to another<\/a><\/li>\n<li class=\"toclevel-2 tocsection-62\"><a href=\"#Default_folder_for_saving_attachments\">3.14 Default folder for saving attachments<\/a><\/li>\n<li class=\"toclevel-2 tocsection-63\"><a href=\"#Pager_behavior\">3.15 Pager behavior<\/a><\/li>\n<li class=\"toclevel-2 tocsection-64\"><a href=\"#Fast_reply\">3.16 Fast reply<\/a><\/li>\n<li class=\"toclevel-2 tocsection-65\"><a href=\"#Ignore_own_e-mail_addresses_from_group-reply\">3.17 Ignore own e-mail addresses from group-reply<\/a><\/li>\n<li class=\"toclevel-2 tocsection-66\"><a href=\"#IMAP_message_cache\">3.18 IMAP message cache<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-67\"><a href=\"#Troubleshooting\">4 Troubleshooting<\/a>\n<ul><li class=\"toclevel-2 tocsection-68\"><a href=\"#Backspace_does_not_work_in_Mutt\">4.1 Backspace does not work in Mutt<\/a><\/li>\n<li class=\"toclevel-2 tocsection-69\"><a href=\"#The_change-folder_function_always_prompt_for_the_same_mailbox\">4.2 The <i>change-folder<\/i> function always prompt for the same mailbox<\/a><\/li>\n<li class=\"toclevel-2 tocsection-70\"><a href=\"#I_cannot_change_folder_when_using_Mutt_read-only_.28Mutt_-R.29\">4.3 I cannot change folder when using Mutt read-only (Mutt -R)<\/a><\/li>\n<li class=\"toclevel-2 tocsection-71\"><a href=\"#Error_sending_message.2C_child_exited_127_.28Exec_error..29.\">4.4 Error sending message, child exited 127 (Exec error.).<\/a><\/li>\n<li class=\"toclevel-2 tocsection-72\"><a href=\"#Character_encoding_problems\">4.5 Character encoding problems<\/a><\/li>\n<li class=\"toclevel-2 tocsection-73\"><a href=\"#Unable_to_login_with_GMail\">4.6 Unable to login with GMail<\/a><\/li>\n<li class=\"toclevel-2 tocsection-74\"><a href=\"#Not_possible_to_open_too_long_URLs_with_urlview\">4.7 Not possible to open too long URLs with urlview<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-75\"><a href=\"#Documentation\">5 Documentation<\/a><\/li>\n<li class=\"toclevel-1 tocsection-76\"><a href=\"#See_also\">6 See also<\/a><\/li>\n<\/ul><\/div><h2>Installation<\/h2><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Install\" class=\"mw-redirect\" title=\"Install\">Install<\/a> the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=mutt\">mutt<\/a> package.  Alternatively consider using the <a href=\"#NeoMutt\">#NeoMutt<\/a> package.\n<\/p><p>Optionally install external helper applications for an IMAP setup, such as <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Isync\" title=\"Isync\">isync<\/a>, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/OfflineIMAP\" title=\"OfflineIMAP\">OfflineIMAP<\/a>, or <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Msmtp\" title=\"Msmtp\">msmtp<\/a>.\n<\/p><p>Or (if using POP3) <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=getmail\">getmail<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=fetchmail\">fetchmail<\/a> or <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=fdm\">fdm<\/a> and <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=procmail\">procmail<\/a>.\n<\/p><div class=\"archwiki-template-box archwiki-template-box-note\"><strong>Note:<\/strong><ul><li>If you just need the authentication methods LOGIN and PLAIN, these are satisfied with the dependency <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=libsasl\">libsasl<\/a><\/li>\n<li>If you want to (or have to) use CRAM-MD5, GSSAPI or DIGEST-MD5, install the package <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=cyrus-sasl-gssapi\">cyrus-sasl-gssapi<\/a><\/li>\n<li>If you are using Gmail as your SMTP server, you may need to install the package <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=cyrus-sasl\">cyrus-sasl<\/a><\/li><\/ul><\/div><h3>NeoMutt<\/h3><div class=\"noprint archwiki-template-message\"><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/File:Tango-go-next.png\" class=\"image\"><img alt=\"Tango-go-next.png\" src=\"https:\/\/wiki.archlinux.org\/images\/f\/f0\/Tango-go-next.png\" width=\"48\" height=\"48\" \/><\/a><b>This article or section is a candidate for moving to <a href=\"https:\/\/wiki.archlinux.org\/index.php?title=NeoMutt&amp;action=edit&amp;redlink=1\" class=\"new\" title=\"NeoMutt (page does not exist)\">NeoMutt<\/a>.<\/b><a href=\"https:\/\/wiki.archlinux.org\/index.php\/File:Tango-go-next.png\" class=\"image\"><img alt=\"Tango-go-next.png\" src=\"https:\/\/wiki.archlinux.org\/images\/f\/f0\/Tango-go-next.png\" width=\"48\" height=\"48\" \/><\/a><\/p><p><b>Notes:<\/b> Is it a drop-in replacement for mutt? Do all configurations on this page apply to it as well? (Discuss in <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/wiki.archlinux.org\/index.php\/Talk:Mutt#Neomutt\">Talk:Mutt#Neomutt<\/a>)<\/p><\/div><p>The <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.neomutt.org\/\">NeoMutt<\/a> project aims to bring together all the patches for Mutt. It adds a large set of <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.neomutt.org\/feature.html\">features<\/a>. Lots of old Mutt patches have been brought up-to-date, tidied and documented.\n<\/p><p>While there are many different packages of mutt in the AUR, each of them providing another set of patches, NeoMutt aims to replace them in the future by implementing appropriate compile options.\n<\/p><p>NeoMutt can be installed with the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=neomutt\">neomutt<\/a> package or alternatively from the AUR at <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/neomutt-git\/\">neomutt-git<\/a><sup><small>AUR<\/small><\/sup> for the development version.\n<\/p><h2>Configuration<\/h2><p>This section covers <a href=\"#IMAP\">#IMAP<\/a>, <a href=\"#POP3\">#POP3<\/a>, <a href=\"#Maildir\">#Maildir<\/a> and <a href=\"#SMTP\">#SMTP<\/a> configuration.\n<\/p><p>Mutt will, by default, search six locations for its configuration file; <code>~\/.muttrc<\/code>, <code>~\/.mutt\/muttrc<\/code>, and <code>$XDG_CONFIG_HOME\/mutt\/muttrc<\/code>, first with <code>-MUTT_VERSION<\/code> appended, then without. Any of these locations will work. In case you decide to put the initialization file somewhere else, use <code>$ mutt -F \/path\/to\/.muttrc<\/code>. You should also know some prerequisite for Mutt configuration. Its syntax is very close to the Bourne Shell. For example, you can get the content of another config file:\n<\/p><pre>source \/path\/to\/other\/config\/file\n<\/pre><p>You can use variables and assign the result of shell commands to them.\n<\/p><pre>set editor=`echo \\$EDITOR`\n<\/pre><p>Here the <code>$<\/code> gets escaped so that it does not get substituted by Mutt before being passed to the shell. Also note the use of the backquotes, as bash syntax <code>$(...)<\/code> does not work. Mutt has a lot of predefined variables, but you can also set your own. User variables <b>must begin with \"my\"!<\/b>\n<\/p><pre>set my_name = \"John Doe\"\n<\/pre><h3>IMAP<\/h3><h4>Native IMAP support<\/h4><p>The <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=mutt\">mutt<\/a> package is compiled with IMAP support. At the very least you need to have four lines in your muttrc file to be able to access your mail.\n<\/p><h5>imap_user<\/h5><pre>set imap_user=USERNAME\n<\/pre><div class=\"archwiki-template-box archwiki-template-box-tip\"><strong>Tip:<\/strong>Continuing with the previous example, remember that Gmail requires your full email address (this is not standard):<pre>set imap_user=your.username@gmail.com<\/pre><\/div><h5>imap_pass<\/h5><p>If unset, the password will be prompted for.\n<\/p><pre>set imap_pass=SECRET\n<\/pre><p><strong>Tip:<\/strong> If you have enabled two-factor authentication in Gmail and you have added an application specific password for Mutt, you will want to use that password here rather than your regular Gmail password.<\/p><h5>folder<\/h5><p>Instead of a local directory which contains all your mail (and directories), use your server (and the highest folder in the hierarchy, if needed).\n<\/p><pre>set folder=imap[s]:\/\/imap.server.domain[:port]\/[folder\/]\n<\/pre><p>You do not have to use a folder, but it might be convenient if you have all your other folders inside your INBOX, for example. Whatever you set here as your folder can be accessed later in Mutt with just an equal sign (=) or a plus sign (+). Example:\n<\/p><pre>set folder=imaps:\/\/imap.gmail.com\/\n<\/pre><p>It should be noted that for several accounts, it is best practice to use different folders -- e.g. for <i>account-hook<\/i>. If you have several Gmail account, use\n<\/p><pre>set folder=imaps:\/\/username@imap.gmail.com\/\n<\/pre><p>instead, where your account is <i>username@gmail.com<\/i>. This way it will be possible to distinguish the different folders. Otherwise it would lead to authentication errors.\n<\/p><h5>spoolfile<\/h5><p>The spoolfile is the folder where your (unfiltered) e-mail arrives. Most e-mail services conventionally name it <i>INBOX<\/i>. You can now use '=' or '+' as a substitution for the full <code>folder<\/code> path that was configured above. For example:\n<\/p><pre>set spoolfile=+INBOX\n<\/pre><h5>mailboxes<\/h5><p>Any IMAP folders that should be checked regularly for new mail should be listed here:\n<\/p><pre>mailboxes =INBOX =family\nmailboxes imaps:\/\/imap.gmail.com\/INBOX imaps:\/\/imap.gmail.com\/family\n<\/pre><p>Alternatively, check for all subscribed IMAP folders (as if all were added with a <code>mailboxes<\/code> line):\n<\/p><pre>set imap_check_subscribed\n<\/pre><p>These two versions are equivalent if you want to subscribe to all folders. So the second method is much more convenient, but the first one gives you more flexibility. Also, newer Mutt versions are configured by default to include a macro bound to the 'y' key which will allow you to change to any of the folders listed under mailboxes.\n<\/p><p>If you do not set this variable, the <i>spoolfile<\/i> will be used by default. This variable is also important for the <a href=\"#Sidebar\">#Sidebar<\/a>.\n<\/p><h5>Summary<\/h5><p>Using these options, you will be able to start Mutt, enter your IMAP password, and start reading your mail. Here is a muttrc snippet (for Gmail) with some other lines you might consider adding for better IMAP support.\n<\/p><pre>set folder      = imaps:\/\/imap.gmail.com\/\nset imap_user   = your.username@gmail.com\nset imap_pass   = your-imap-password\nset spoolfile   = +INBOX\nmailboxes       = +INBOX\n# Store message headers locally to speed things up.\n# If hcache is a folder, Mutt will create sub cache folders for each account which may speeds things up even more.\nset header_cache = ~\/.cache\/mutt\n# Store messages locally to speed things up, like searching message bodies.\n# Can be the same folder as header_cache.\n# This will cost important disk usage according to your e-mail amount.\nset message_cachedir = \"~\/.cache\/mutt\"\n# Specify where to save and\/or look for postponed messages.\nset postponed = +[Gmail]\/Drafts\n# Allow Mutt to open a new IMAP connection automatically.\nunset imap_passive\n# Keep the IMAP connection alive by polling intermittently (time in seconds).\nset imap_keepalive = 300\n# How often to check for new mail (time in seconds).\nset mail_check = 120<\/pre><h4>External IMAP support<\/h4><p>While IMAP support is built into Mutt, it does not download mail for offline use. It is possible to use an external application such as <a href=\"https:\/\/wiki.archlinux.org\/index.php\/OfflineIMAP\" title=\"OfflineIMAP\">OfflineIMAP<\/a> or <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Isync\" title=\"Isync\">isync<\/a> to download your emails to a local folder which can then be processed by Mutt.\n<\/p><p>Consider using applications such as <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=spamassassin\">spamassassin<\/a> or <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/imapfilter\/\">imapfilter<\/a><sup><small>AUR<\/small><\/sup> to sort mail.\n<\/p><h3>POP3<\/h3><p>The <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=mutt\">mutt<\/a> package is compiled with POP3 support, which is configured via the <code>pop_*<\/code> variables as described in <code>muttrc(5)<\/code>.\n<\/p><p>Alternatively, it is possible to use external programs to fetch mail using POP3. One popular option is to use <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Getmail\" title=\"Getmail\">getmail<\/a> for retrieving and <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Procmail\" title=\"Procmail\">procmail<\/a> for filtering the mail.\n<\/p><h3>Maildir<\/h3><p>Maildir is a generic and standardized format. Almost every MUA is able to handle Maildirs and Mutt's support is excellent. There are just a few simple things that you need to do to get Mutt to use them. Open your muttrc and add the following lines:\n<\/p><pre>set mbox_type=Maildir\nset folder=~\/mail\nset spoolfile=+\/\nset header_cache=~\/.cache\/mutt<\/pre><p>This is a minimal Configuration that enables you to access your Maildir and checks for new local Mails in INBOX. This configuration also caches the headers of the eMails to speed up directory-listings. It might not be enabled in your build (but it sure is in the Arch-Package). Note that this does not affect OfflineIMAP in any way. It always syncs all of the directories on a Server. <code>spoolfile<\/code> tells Mutt which local directories to poll for new Mail. You might want to add more Spoolfiles (for example the Directories of Mailing-Lists) and maybe other things. But this is subject to the Mutt manual and beyond the scope of this document.\n<\/p><h3>SMTP<\/h3><p>Whether you use POP or IMAP to receive mail you will probably still send mail using SMTP.\n<\/p><h4>Folders<\/h4><p>There is basically only one important folder here: the one where all your sent e-mails will be saved.\n<\/p><pre>set record = +Sent\n<\/pre><p>Gmail automatically saves sent e-mail to <code>+[Gmail]\/Sent<\/code>, so we do not want duplicates.\n<\/p><pre>unset record\n<\/pre><h4>Native SMTP support<\/h4><p>The <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=mutt\">mutt<\/a> package is compiled with SMTP support.\n<\/p><p>For example:\n<\/p><pre>set my_pass='mysecretpass'\nset my_user=myname\nset realname = 'Your Real Name'\nset from = your-email-address\nset use_from = yes\nset smtp_url=smtps:\/\/$my_user:$my_pass@smtp.domain.tld\nset ssl_force_tls = yes<\/pre><p>Note that if your SMTP credentials are the same as your IMAP credentials, then you can use those variables:\n<\/p><pre>set smtp_url=smtps:\/\/$imap_user:$imap_pass@smtp.domain.tld<\/pre><p>You may need to tweak the security parameters. If you get an error like <code>SSL routines:SSL23_GET_SERVER_HELLO:unknown protocol<\/code>, then your server probably uses the SMTP instead of SMTPS.\n<\/p><pre>set smtp_url=smtp:\/\/$imap_user:$imap_pass@smtp.domain.tld<\/pre><p>There are other variables that you may need to set. For example for use of STARTTLS:\n<\/p><pre>set ssl_starttls = yes<\/pre><p>If your server uses the LOGIN authentication method you might need to specify this explicitly, despite the manual's claim that all methods are tried by default:\n<\/p><pre>set smtp_authenticators = \"login\"<\/pre><p>See <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/muttrc.5\">muttrc(5)<\/a> for more information.\n<\/p><h4>External SMTP support<\/h4><p>An external SMTP agent such as <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Msmtp\" title=\"Msmtp\">msmtp<\/a>, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/SSMTP\" title=\"SSMTP\">SSMTP<\/a> or <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=opensmtpd\">opensmtpd<\/a> can also be used.\n<\/p><p>The <code>sendmail<\/code> variable in your <code>muttrc<\/code> determines the program and arguments used to deliver mail in mutt. Any additional arguments must be interpreted by the program as recipient addresses.\n<\/p><p>For example, if using <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Msmtp\" title=\"Msmtp\">msmtp<\/a>:\n<\/p><pre>muttrc<\/pre><pre>set realname='Disgruntled Kangaroo'\nset sendmail=\"\/usr\/bin\/msmtp\"\nset edit_headers=yes\nset folder=~\/mail\nset mbox=+mbox\nset spoolfile=+inbox\nset record=+sent\nset postponed=+drafts\nset mbox_type=Maildir\nmailboxes +inbox +lovey-dovey +happy-kangaroos<\/pre><h4>Sending mails from Mutt<\/h4><p>Now, startup <code>mutt<\/code>:\n<\/p><p>You should see all the mail in <code>~\/mail\/inbox<\/code>. Press <code>m<\/code> to compose mail; it will use the editor defined by your <code>EDITOR<\/code> environment variable. If this variable is not set, you can fix it before starting Mutt:\n<\/p><pre>$ export EDITOR=your-favorite-editor\n$ mutt\n<\/pre><p>You should store the EDITOR value into your shell resource configuration file (such as <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Bashrc\" class=\"mw-redirect\" title=\"Bashrc\">bashrc<\/a>). You can also set the editor from Mutt's configuration file:\n<\/p><pre>.muttrc<\/pre><pre>set editor=your-favorite-editor<\/pre><p>For testing purposes, address the letter to yourself. After you have written the letter, save and exit the editor. You will return to Mutt, which will now show information about your e-mail. Press <code>y<\/code> to send it.\n<\/p><div class=\"archwiki-template-box archwiki-template-box-warning\"><strong>Warning:<\/strong>If at this point you press<code>q<\/code>by accident, Mutt will ask you<code>Postpone this message? ([yes]\/no)<\/code>. This is really asking whether you want to save the message you just wrote. If you press \"n\" (perhaps because you want to edit the message again) the message will be permanently deleted. When using Mutt, always remember that \"Postpone this message?\" really means \"Save this message?\".<\/div><h3>Multiple accounts<\/h3><p>Now you should have a working configuration for one account at least. You might wonder how to use several accounts, since we put everything into a single file.\n<\/p><p>Well all you need is to write account-specific parameters to their respective files and source them. All the IMAP\/POP3\/SMTP config for each account should go to its respective folder.\n<\/p><p><strong>Warning:<\/strong> When one account is setting a variable that is not specified for other accounts, you <b>must unset<\/b> it for them, otherwise configuration will overlap and you will most certainly experience unexpected behaviour.<\/p><p>Mutt can handle this thanks to one of its most powerful features: hooks. Basically a hook is a command that gets executed before a specific action. There are several hooks available. For multiple accounts, you must use account-hooks <i>and<\/i> folder-hooks.\n<\/p><ul><li>Folder-hooks will run a command before switching folders. This is mostly useful to set the appropriate SMTP parameters when you are in a specific folder. For instance when you are in your work mailbox and you send a e-mail, it will automatically use your work account as sender.<\/li>\n<li>Account-hooks will run a command every time Mutt calls a function related to an account, like IMAP syncing. It does not require you to switch to any folder.<\/li><\/ul><p>Hooks take two parameters:\n<\/p><pre>account-hook [!]regex command\nfolder-hook [!]regex command\n<\/pre><p>The regex is the folder to be matched (or not if preceded by the\u00a0!). The command tells what to do.\n<\/p><p>Let us give a full example:\n<\/p><pre>.muttrc<\/pre><pre>## General options\nset header_cache = \"~\/.cache\/mutt\"\nset imap_check_subscribed\nset imap_keepalive = 300\nunset imap_passive\nset mail_check = 60\nset mbox_type=Maildir\n## ACCOUNT1\nsource \"~\/.mutt\/work\"\n# Here we use the $folder variable that has just been set in the sourced file.\n# We must set it right now otherwise the 'folder' variable will change in the next sourced file.\nfolder-hook $folder 'source ~\/.mutt\/work'\n## ACCOUNT2\nsource \"~\/.mutt\/personal\"\nfolder-hook *user@gmail.com\/ 'source ~\/.mutt\/personal'\nfolder-hook *user@gmail.com\/Family 'set realname=\"Bob\"'\n<\/pre><pre>.mutt\/work<\/pre><pre>## Receive options.\nset imap_user=user@gmail.com\nset imap_pass=****\nset folder = imaps:\/\/user@imap.gmail.com\/\nset spoolfile = +INBOX\nset postponed = +Drafts\nset record = +Sent\n## Send options.\nset smtp_url=smtps:\/\/user:****@smtp.gmail.com\nset realname='User X'\nset from=user@gmail.com\nset hostname=\"gmail.com\"\nset signature=\"John Doe\"\n# Connection options\nset ssl_force_tls = yes\nunset ssl_starttls\n## Hook -- IMPORTANT!\naccount-hook $folder \"set imap_user=user@gmail.com imap_pass=****\"\n<\/pre><p>Finally <code>.mutt\/personal<\/code> should be similar to <code>.mutt\/work<\/code>.\n<\/p><p>Now all your accounts are set, start Mutt. To switch from one account to another, just change the folder (<code>c<\/code> key). Alternatively you can use the <a href=\"#Sidebar\">sidebar<\/a>.\n<\/p><p>To change folder for different mailboxes you have to type the complete address -- for IMAP\/POP3 folders, this may be quite inconvenient -- let us bind some key to it.\n<\/p><pre>## Shortcuts\nmacro index,pager &lt;f2&gt; '&lt;sync-mailbox&gt;&lt;enter-command&gt;source ~\/.mutt\/personal&lt;enter&gt;&lt;change-folder&gt;!&lt;enter&gt;'\nmacro index,pager &lt;f3&gt; '&lt;sync-mailbox&gt;&lt;enter-command&gt;source ~\/.mutt\/work&lt;enter&gt;&lt;change-folder&gt;!&lt;enter&gt;'\n<\/pre><p>With the above shortcuts (or with the sidebar) you will find that changing folders (with <code>c<\/code> by default) is not contextual, <i>i.e.<\/i> it will not list the folders of the current mailbox, but of the one used the last time you changed folders. To make the behaviour more contextual, the trick is to press <i>=<\/i> or <i>+<\/i> for current mailbox. You can automate this with the following macro:\n<\/p><pre>macro index 'c' '&lt;change-folder&gt;?&lt;change-dir&gt;&lt;home&gt;^K=&lt;enter&gt;'\n<\/pre><h3>Passwords management<\/h3><p>Keep in mind that writing your password in <code>.muttrc<\/code> is a security risk. One solution is to always enter the password manually, but this becomes cumbersome.\n<\/p><p>An alternative solution is to encrypt your password with <a href=\"https:\/\/wiki.archlinux.org\/index.php\/GnuPG\" title=\"GnuPG\">GnuPG<\/a> in an encrypted file. <a href=\"https:\/\/wiki.archlinux.org\/index.php\/GnuPG#Create_a_key_pair\" title=\"GnuPG\">Setup your own keypair<\/a> if you have not done so already. <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Create\" class=\"mw-redirect\" title=\"Create\">Create<\/a> a file in a <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Tmpfs\" title=\"Tmpfs\">tmpfs<\/a> with the following contents:\n<\/p><pre>set my_pass = \"<i> password<\/i>\"\n<\/pre><div class=\"archwiki-template-box archwiki-template-box-note\"><strong>Note:<\/strong>Remember that user defined variables<b>must<\/b>start with<code>my_<\/code>.<\/div><p>Then <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Encrypt\" class=\"mw-redirect\" title=\"Encrypt\">encrypt<\/a> this file, setting yourself as the recipient and move it into an accessible location. In this example the encrypted file resides at $HOME\/.my-pwds.gpg.\n<\/p><p>In your mutt configuration file add the following before any account:\n<\/p><pre>source \"gpg -dq $HOME\/.my-pwds.gpg |\"\n<\/pre><p><strong>Note:<\/strong> At the end of the line above, there is no space between the pipe and the double quote.<\/p><p>This decrypts the file quietly and sets the variable <code>my_pass<\/code> in this example. This can be used in any variable after it has been source. For example:\n<\/p><pre>set imap_pass=$my_pass<\/pre><p>If you use external tools like OfflineIMAP and msmtp, you need to set up an agent (e.g. gpg-agent, see <a href=\"https:\/\/wiki.archlinux.org\/index.php\/GnuPG#gpg-agent\" title=\"GnuPG\">GnuPG#gpg-agent<\/a>) to keep the passphrase into cache and thus avoiding those tools always prompting for it.\n<\/p><h4>Security concern<\/h4><p>If <code>enter-command<\/code> is available from the UI, it is possible to see the password unencrypted, which may be undesired if anybody else than you has access to your session while Mutt is running. You may want to disable it for this reason. As a consequence, every command that the user intends to use must be bound to a key in advance, otherwise it will never be accessible.\n<\/p><pre>.muttrc<\/pre><pre> bind generic,alias,attach,browser,editor,index,compose,pager,pgp,postpone ':' noop\n<\/pre><h2>Tips and tricks<\/h2><p>Guides to get you started with using &amp; customizing Mutt\u00a0:\n<\/p><ul><li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/mutt.postle.net\/\">My first Mutt<\/a> (maintained by Bruno Postle)<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.therandymon.com\/woodnotes\/mutt\/using-mutt.html\">The Woodnotes Guide to the Mutt Email Client<\/a> (maintained by Randall Wood)<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/stevelosh.com\/blog\/2012\/10\/the-homely-mutt\">The Homely Mutt<\/a> (by Steve Losh)<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/codesorcery.net\/old\/mutt\/mutt-gnupg-howto\">Everything You Need To Know To Start Using GnuPG with Mutt<\/a> (by Justin R. Miller)<\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/neomutt.org\/contrib\/useful-programs\">A List of useful programs that often are used in combination with neomutt<\/a><\/li><\/ul><p>If you have any Mutt specific questions, feel free to ask in the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/IRC_channel\" class=\"mw-redirect\" title=\"IRC channel\">IRC channel<\/a>.\n<\/p><h3>Key bindings<\/h3><p>The default key bindings are quite far from the more common Emacs-like or Vi-like bindings. You can customize them to your preference. Mutt has a different set of bindings for the pager, the index, the attachment view, etc. Thus you need to specify which <i>map<\/i> you want to modify when you bind a key. You can review the list of commands and key bindings from Mutt's help page (default key: <code>?<\/code>). Example of Vi-like bindings:\n<\/p><pre>muttrc<\/pre><pre>bind pager j next-line\nbind pager k previous-line\nbind attach,index,pager \\CD next-page\nbind attach,index,pager \\CU previous-page\nbind pager g top\nbind pager G bottom\nbind attach,index g first-entry\nbind attach,index G last-entry\n<\/pre><h3>Composition<\/h3><h4>Encrypt and sign mail (GnuPG)<\/h4><p>To start encrypting mail in mutt using <a href=\"https:\/\/wiki.archlinux.org\/index.php\/GnuPG\" title=\"GnuPG\">GnuPG<\/a> copy <code>\/usr\/share\/doc\/mutt\/samples\/gpg.rc<\/code> to your mutt configuration folder (e.g. to <code>~\/.mutt\/gpg.rc<\/code>). Then <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Append\" class=\"mw-redirect\" title=\"Append\">append<\/a> the following to your mutt configuration file (e.g. <code>~\/.mutt\/mutrrc<\/code>):\n<\/p><pre>source ~\/.mutt\/gpg.rc\n<\/pre><p>Most encryption options are then available by pressing <code>p<\/code> in the compose view.\n<\/p><p>See the <code>pgp_*<\/code> and <code>crypt_*<\/code> options in <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/muttrc.5\">muttrc(5)<\/a>.\n<\/p><h4>E-mail character encoding<\/h4><p>When using Mutt there are two levels where the character sets that must be specified:\n<\/p><ul><li>The text editor used to write the e-mail must save it in the desired encoding.<\/li>\n<li>Mutt will then check the e-mail and determine which encoding is the more appropriate according to the priority you specified in the <code>send_charset<\/code> variable. Default: \"us-ascii:iso-8859-1:utf-8\".<\/li><\/ul><p>So if you write an e-mail with characters allowed in ISO-8859-1 (like 'r\u00e9sum\u00e9'), but without characters specific to Unicode, then Mutt will set the encoding to ISO-8859-1.\n<\/p><p>To avoid this behaviour, set the variable in your <code>muttrc<\/code>:\n<\/p><pre>set send_charset=\"us-ascii:utf-8\"\n<\/pre><p>or even\n<\/p><pre>set send_charset=\"utf-8\"\n<\/pre><p>The first compatible charset starting from the left will be used. Since UTF-8 is a superset of US-ASCII it does not harm to leave it in front of UTF-8, it may ensure old MUA will not get confused when seeing the charset in the e-mail header.\n<\/p><h4>Custom mail headers<\/h4><p>One of the greatest thing in Mutt is that you can have full control over your mail header.\n<\/p><p>First, make your headers editable when you write e-mails:\n<\/p><pre>set edit_headers=yes\n<\/pre><p>Mutt also features a special function <code>my_hdr<\/code> for customizing your header. Yes, it is named just like a variable, but in fact it is a function.\n<\/p><p>You can clear it completely, which you <i>should<\/i> do when switching accounts with different headers, otherwise they will overlap:\n<\/p><pre>unmy_hdr *\n<\/pre><p>Other variables have also an impact on the headers, so it is wise to clear them before using <code>my_hdr<\/code>:\n<\/p><pre>unset use_from\nunset use_domain\nunset user_agent\n<\/pre><p>Now, you can add any field you want -- even non-standard one -- to your header using the following syntax:\n<\/p><pre>my_hdr &lt;FIELD&gt;: &lt;VALUE&gt;\n<\/pre><p>Note that &lt;VALUE&gt; can be the result of a command.\n<\/p><p>Example:\n<\/p><pre>## Extra info.\nmy_hdr X-Info: Keep It Simple, Stupid.\n## OS Info.\nmy_hdr X-Operating-System: `uname -s`, kernel `uname -r`\n## This header only appears to MS Outlook users\nmy_hdr X-Message-Flag: WARNING!! Outlook sucks\n## Custom Mail-User-Agent ID.\nmy_hdr User-Agent: Every email client sucks, this one just sucks less.\n<\/pre><h4>Signature block<\/h4><p>Create a .signature in your home directory.  Your signature will be appended at the end of your email. Alternatively you can specify a file in your Mutt configuration:\n<\/p><pre>set signature=\"path\/to\/sig\/file\"\n<\/pre><h5>Random signature<\/h5><p>You can use <i>fortune<\/i> (package <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=fortune-mod\">fortune-mod<\/a>) to add a random signature to Mutt.\n<\/p><p>Create a fortune file and then add the following line to your .muttrc:\n<\/p><pre>set signature=\"fortune pathtofortunefile|\"<\/pre><p>Note the pipe at the end. It tells Mutt that the specified string is not a file, but a command.\n<\/p><h4>Compose and send from command line<\/h4><p>Man pages will show all available commands and how to use them, but here are a couple of examples. You could use Mutt to send alerts, logs or some other system information, triggered by login through <code>.bash_profile<\/code>, or as a regular <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Cron\" title=\"Cron\">cron<\/a> job.\n<\/p><p>Send a message:\n<\/p><pre>mutt -s \"Subject\" somejoeorjane@someserver.com &lt; \/var\/log\/somelog\n<\/pre><p>Send a message with attachment:\n<\/p><pre>mutt -s \"Subject\" somejoeorjane@someserver.com -a somefile &lt; \/tmp\/sometext.txt\n<\/pre><h4>Compose HTML e-mails<\/h4><p>Since Mutt has nothing of a WYSIWIG client, HTML is quite straightforward, and you can do much more than with all WYSIWIG mail clients around since you edit the source code directly.\nSimply write your mail using HTML syntax. For example:\n<\/p><pre>This is normal text&lt;br&gt;\n&lt;b&gt;This is bold text&lt;\/b&gt;\n<\/pre><p>Now before sending the mail, use the <code>edit-type<\/code> command (default shortcut <code>Ctrl+t<\/code>), and replace <code>text\/plain<\/code> by <code>text\/html<\/code>.\n<\/p><p><strong>Note:<\/strong> HTML e-mails are regarded by many people as useless, cumbersome,  and subject to reading issues. Mutt can read HTML mails with a text browser like w3m or lynx, but it has clearly no advantage over a plain-text e-mail. You should avoid writing HTML e-mails when possible.<\/p><h4>Display another email while composing<\/h4><p>A common complaint with Mutt is that when composing a new mail (or reply), you cannot open another mail (i.e. for checking with another correspondent) without closing the current mail (postponing). The following describes a solution:\n<\/p><p>First, fire up Mutt as usual. Then, launch another terminal window. Now start a new Mutt with\n<\/p><pre>mutt -R\n<\/pre><p>This starts Mutt in read-only mode, and you can browse other emails at your convenience. It is strongly recommended to always launch a second Mutt in read-only mode, as conflicts will easily arise otherwise.\n<\/p><div class=\"archwiki-template-box archwiki-template-box-note\"><strong>Note:<\/strong>When changing folders (with<code>c<\/code>or<code>y<\/code>) the read-only mode is not preserved.  Instead<code>Esc c<\/code>has to be used.<\/div><div class=\"archwiki-template-box archwiki-template-box-tip\"><strong>Tip:<\/strong>This solution calls for a bit of typing, so it is suitable to bind the following command to a<a href=\"https:\/\/wiki.archlinux.org\/index.php\/Keyboard_shortcut\" class=\"mw-redirect\" title=\"Keyboard shortcut\">keyboard shortcut<\/a>:<pre>$TERMINAL -e mutt -R\n<\/pre>where<code>$TERMINAL<\/code>is your terminal.<\/div><h3>Printing<\/h3><p>You can install <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/muttprint\/\">muttprint<\/a><sup><small>AUR<\/small><\/sup> for fancier printing quality. In your muttrc file, insert:\n<\/p><pre>set print_command=\"\/usr\/bin\/muttprint\u00a0%s -p {PrinterName}\"\n<\/pre><h3>Viewing content<\/h3><h4>Viewing URLs in a web browser<\/h4><p>Your should start by creating a .mutt directory in $HOME if not done yet.  There, create a file named macros.  Insert the following:\n<\/p><pre> macro pager \\cb &lt;pipe-entry&gt;'urlview'&lt;enter&gt; 'Follow links with urlview'\n<\/pre><p>Then install the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/urlview\/\">urlview<\/a><sup><small>AUR<\/small><\/sup> package.\n<\/p><p>Create a .urlview in $HOME and insert the following:\n<\/p><pre>REGEXP (((http|https|ftp|gopher)|mailto)[.:][^ &gt;\"\\t]*|www\\.[-a-z0-9.]+)[^ .,;\\t&gt;\"&gt;\\):]\nCOMMAND &lt;your-browser&gt;\u00a0%s\n<\/pre><p>When you read an email on the pager, hitting ctrl+b will list all the urls from the email.  Navigate up or down with arrow keys and hit enter on the desired url. Your browser will start and go to the selected site.\n<\/p><p>Some browser will require additional arguments to work properly. For example, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Luakit\" title=\"Luakit\">Luakit<\/a> will close on Mutt exit. You need to fork it to background, using the <code>-n<\/code> parameter:\n<\/p><pre>COMMAND luakit -n\u00a0%s 2&gt;\/dev\/null\n<\/pre><p>The <code>2&gt;\/dev\/null<\/code> is to make it quiet, i.e. to prevent useless message printing where you do not want them to.\n<\/p><p><strong>Note:<\/strong> urlview has a few deficiencies (e.g. the inability to handle certain email encodings) and is fairly feature-bare (e.g. it does not provide context for links it finds). There are a couple alternatives that do better. One, which can handle all email encodings and provides link context, is <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.memoryhole.net\/~kyle\/extract_url\/\">extract_url.pl<\/a>. Another, which can also provide link context but cannot handle all email encodings, is <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/urlscan-git\/\">urlscan-git<\/a><sup><small>AUR<\/small><\/sup>. Both are drop-in replacements for urlview, though extract_url has features which benefit from additional configuration changes.<\/p><h4>Viewing HTML<\/h4><p>It is possible to pass the html body to an external HTML program and then dump it, keeping email viewing uniform and unobtrusive. Three programs are described here: <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=lynx\">lynx<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=w3m\">w3m<\/a> and <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=elinks\">elinks<\/a> (make sure the selected package is <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Pacman\" title=\"Pacman\">installed<\/a>).\n<\/p><p>If <code>~\/.mutt\/mailcap<\/code> does not exist you will need to create it and save the following to it.\n<\/p><pre>text\/html; lynx -assume_charset=%{charset} -display_charset=utf-8 -dump\u00a0%s; nametemplate=%s.html; copiousoutput\n<\/pre><p>or, in case of w3m,\n<\/p><pre>text\/html; w3m -I\u00a0%{charset} -T text\/html; copiousoutput;\n<\/pre><p>or, in case of elinks,\n<\/p><pre>text\/html; elinks -dump\u00a0; copiousoutput;\n<\/pre><p>Edit <code>~\/.muttrc<\/code> and add the following,\n<\/p><pre>set mailcap_path \t= ~\/.mutt\/mailcap\n<\/pre><p>To automatically open HTML messages in lynx, w3m or elinks add this additional line to the muttrc:\n<\/p><pre>auto_view text\/html\n<\/pre><p>The beauty of this is, instead of seeing an html body as source or being opened by a separate program, in this case lynx, you see the formatted content directly, and any url links within the email can be displayed with <code>Ctrl+b<\/code>, assuming you have <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/urlview\/\">urlview<\/a><sup><small>AUR<\/small><\/sup> installed.\n<\/p><p>If you receive many emails with multiple or alternate encodings Mutt may default to treating every email as html.  To avoid this, add the following variable to your ~\/.muttrc to have Mutt default to text when available and use w3m\/lynx only when no text version is availble in the email:\n<\/p><pre>alternative_order text\/plain text\/html\n<\/pre><p>Some HTML mails may not display correctly in a text-based web browser. As a fallback solution, you can bind a key to open a graphical browser in such cases.\nThe following macro will open the HTML mail selected  from the attachment view in the web browser defined in the environment. (Feel free to adapt the <code>~\/.cache\/mutt\/<\/code> folder).\n<\/p><pre>macro attach 'V' \"&lt;pipe-entry&gt;cat &gt;~\/.cache\/mutt\/mail.html &amp;&amp; $BROWSER ~\/.cache\/mutt\/mail.html &amp;&amp; rm ~\/.cache\/mutt\/mail.html&lt;enter&gt;\"\n<\/pre><h4>Filtering the message view<\/h4><p>You can restrict the view to e-mails matching a pattern and specific properties with the <code>limit<\/code> command (default shortcut: <code>l<\/code>).\n<\/p><p>To view all e-mails containing \"foo\" in the header, simply write \"foo\" and you are done. To remove the filter, use the \"all\" keyword.\n<\/p><p>To view all flagged messages, use\n<\/p><pre>~F\n<\/pre><p>To view all unread messages that are either of size \u22651MB or from johndoe, use\n<\/p><pre>~U (~z 1M- | ~f johndoe)\n<\/pre><p>All possible patterns are listed in the <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mutt.org\/doc\/manual\/#patterns\">official manual<\/a>.\n<\/p><h4>Conversation grouping<\/h4><p>The default sort order is by date. Use the <code>sort-mailbox<\/code> command (default key: <code>o<\/code>) to change the sorting option. You can group e-mails by conversation\/thread, in which case you can define how to sort threads and how to sort within a thread.\n<\/p><p>In the following example, threads are sorted according to the date of their last e-mail.\n<\/p><pre>muttrc<\/pre><pre>set sort=threads\nset sort_aux=last-date-received\n<\/pre><h3>Configuring editors to work with mutt<\/h3><h4>vim<\/h4><ul><li>To limit the width of text to 72 characters, edit your .<a href=\"https:\/\/wiki.archlinux.org\/index.php\/Vim\" title=\"Vim\">vimrc<\/a> file and add:<\/li><\/ul><pre>au BufRead \/tmp\/mutt-* set tw=72\n<\/pre><ul><li>Another choice is to use Vim's mail filetype plugin to enable other mail-centric options besides 72 character width. Edit <code>~\/.vim\/filetype.vim<\/code>, creating it if unpresent, and add:<\/li><\/ul><pre>augroup filetypedetect\n  \" Mail\n  autocmd BufRead,BufNewFile *mutt-*              setfiletype mail\naugroup END\n<\/pre><ul><li>To set a different tmp directory, e.g. ~\/.tmp, add a line to your muttrc as follows:<\/li><\/ul><pre>set tmpdir=\"~\/.tmp\"\n<\/pre><ul><li>To reformat a modified text see the Vim context help<\/li><\/ul><pre>:h 10.7\n<\/pre><h4>GNU nano<\/h4><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Nano\" title=\"Nano\">nano<\/a> is another nice console editor to use with Mutt.\n<\/p><p>To limit the width of text to 72 characters, edit your .nanorc file and add:\n<\/p><pre> set fill 72\n<\/pre><p>If you do not want to limit the width of text globally, you can pass the column number as an argument to the hard-wrap option in your muttrc file, e.g.:\n<\/p><pre> set editor=\"nano -r 72\"\n<\/pre><p>Also, in muttrc file, you can specify the line to start editing so that you will skip the mail header:\n<\/p><pre> set editor=\"nano +7\"\n<\/pre><h4>Emacs<\/h4><p>Emacs has a <i>mail<\/i> and a <i>message<\/i> major mode. To switch to mail-mode automatically when Emacs is called from Mutt, you can add the following to your <code>.emacs<\/code>:\n<\/p><pre>.emacs<\/pre><pre>;; Mutt support.\n(setq auto-mode-alist (append '((\"\/tmp\/mutt.*\" . mail-mode)) auto-mode-alist))\n<\/pre><p>If you usually run Emacs daemon, you may want Mutt to connect to it. Add this to your <code>.muttrc<\/code>:\n<\/p><pre>.muttrc<\/pre><pre>set editor=\"emacsclient -a \\\"\\\" -t\"\n<\/pre><h3>Display settings<\/h3><h4>Colors<\/h4><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Append\" class=\"mw-redirect\" title=\"Append\">Append<\/a> the contents of <code>\/usr\/share\/doc\/mutt\/samples\/colors.linux<\/code> to your .muttrc file, or copy and source it. Then adjust to your liking.\n<\/p><p>The actual color each of these settings will produce depends on the colors set in your <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Xresources\" class=\"mw-redirect\" title=\"Xresources\">~\/.Xresources<\/a> file.\n<\/p><p>Alternatively, you can source any file you want containing colors (and thus act as a theme file). See <a rel=\"nofollow\" class=\"external autonumber\" href=\"http:\/\/nongeekshandbook.blogspot.be\/2009\/03\/mutt-color-configuration.html\">[1]<\/a> for a theme example.\n<\/p><h4>Index Format<\/h4><p>Here follows a quick example to put in your <code>.muttrc<\/code> to customize the Index Format, i.e. the columns displayed in the folder view.\n<\/p><pre>set date_format=\"%y-%m-%d %T\"\nset index_format=\"%2C | %Z [%d] %-30.30F (%-4.4c) %s\"\n<\/pre><p>See the <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mutt.org\/doc\/manual\/#index-format\">Mutt Reference<\/a>, <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/strftime.3\">strftime(3)<\/a> and <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/printf.3\">printf(3)<\/a> for more details.\n<\/p><h4>Display recipient instead of sender in \"Sent\" folder view<\/h4><p>By default Mutt uses the <code>%L<\/code> format string in the <code>index_format<\/code> variable, which will display:\n<\/p><ul><li>\"To &lt;list-name&gt;\", if an address in the \"To:\" or \"Cc:\" header field matches an address defined by the user's <code>subscribe<\/code> command.<\/li>\n<li>Otherwise it displays the author name, or recipient name if the message is from you.<\/li><\/ul><p>If you use multiple email addresses in the same mailbox, make sure to configure the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/dev.mutt.org\/trac\/wiki\/UseCases\/MultiAccounts#Settinguptheaddresses:alternates\">alternates variable<\/a>, so that Mutt knows which messages were from you.\n<\/p><h4>Variable column width<\/h4><p>If you resize the window, the subject might get truncated while there is still unused space left for some fields, like for the sender.\nYou can get the maximum number of columns supported by your terminal (i.e. the width) using a shell call to <code>tput cols<\/code>. With this value, you can set a percentage of the width to fields like Sender and Subject.\n<\/p><p>Example using the above folder-hook and a sidebar width of 24:\n<\/p><pre>muttrc<\/pre><pre>## From field gets 30% of remaining space, Subject gets 70%.\n## Remaining space is the total width minus the other fields (35), minus the sidebar (24)\nset my_index_format_pre='set my_col_from = `echo $((30 * ($(tput cols)-35-24) \/ 100))`; set my_col_subject = `echo $((70 * ($(tput cols)-35-24) \/ 100))`; set index_format=\"%2C | %Z [%d] %-$my_col_from.${my_col_from}'\nset my_index_format_post=' (%-4.4c) %?M?&lt;%M&gt; ?%-$my_col_subject.${my_col_subject}s\"'\nfolder-hook .*[sS]ent.* \"$my_index_format_pre\"t\"$my_index_format_post\"\nfolder-hook ! .*[sS]ent.* \"$my_index_format_pre\"F\"$my_index_format_post\"\n<\/pre><p>We <i>must<\/i> set the variables <code>my_col_from<\/code> and <code>my_col_from<\/code> from within the hooks. Otherwise, the column values will not get re-computed.\n<\/p><p>We can add a binding to force re-computing the index format without changing folder:\n<\/p><pre>muttrc<\/pre><pre>macro index,pager \\CL \"&lt;enter-command&gt;$my_index_format_pre\"F\"$my_index_format_post&lt;enter&gt;&lt;redraw-screen&gt;\"\n<\/pre><h4>Sidebar<\/h4><p>Example settings for a sidebar are in <code>\/usr\/share\/doc\/mutt\/samples\/sample.muttrc-sidebar<\/code>, including keybindings. Copy, edit, and source that file in your mutt configuration file. Be sure to change <code>set sidebar_visible = yes<\/code>.\n<\/p><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Append\" class=\"mw-redirect\" title=\"Append\">Append<\/a> the following in order to toggle the sidebar visibility:\n<\/p><pre>bind index,pager B sidebar-toggle-visible\n<\/pre><div class=\"archwiki-template-box archwiki-template-box-note\"><strong>Note:<\/strong>You<i>must<\/i>set the<code>mailboxes<\/code>variables or the<code>imap_check_subscribed<\/code>to tell the sidebar which folder should be displayed. See the<a href=\"#mailboxes\">mailboxes<\/a>section.<\/div><p>Note that with the <code>mailboxes<\/code> option, folders appear in the order they were set to  <code>mailboxes<\/code> if you do not use the <code>sidebar_sort_method<\/code> option.\n<\/p><p><strong>Tip:<\/strong> To add a separator between different mailboxes, add a fake folder to the list of folders For example add:\n mailboxes \"+-- My mailbox -----------\"<\/p><h4>Display the index above the pager view<\/h4><p>Set the following variable in your <code>muttrc<\/code>:\n<\/p><pre>set pager_index_lines=10\n<\/pre><h3>Contact management<\/h3><h4>Address aliases<\/h4><p><i>Aliases<\/i> is the way Mutt manages contacts. An alias is <b>nickname [longname] &lt;address&gt;<\/b>.\n<\/p><ul><li>The <b>nickname<\/b> is what you will type in Mutt to get your contact address. One word only, and should be easy to remember.<\/li>\n<li>The <b>longname<\/b> is optional. It may be several words.<\/li>\n<li>An <b>&lt;address&gt;<\/b> must be in a valid form (i.e. with an <code>@<\/code>).<\/li><\/ul><p>It is quite simple indeed. Add this to <code>.muttrc<\/code>:\n<\/p><pre>set alias_file = \"~\/.mutt\/aliases\"\nset sort_alias = alias\nset reverse_alias = yes\nsource $alias_file<\/pre><p>Explanation:\n<\/p><ul><li><code>alias_file<\/code> is the file where the information is getting stored when you add an alias from within Mutt.<\/li>\n<li><code>sort_alias<\/code> specifies which field to use to sort the alias list when displayed in Mutt. Possible values: alias, address.<\/li>\n<li><code>reverse_alias<\/code> if set to yes mutt will display the \"personal\" name from your aliases in the index menu if it finds an alias that matches the message's sender.<\/li>\n<li><code>source $alias_file<\/code> tells Mutt to read aliases on startup. Needed for auto-completion.<\/li><\/ul><p>Now all you have to do when prompted <code>To:<\/code> is writing the alias instead of the full address. The beauty of it is that you can auto-complete the alias using <code>Tab<\/code>.\nAutocompleting a wrong or an empty string will display the full list. You can select the alias as usual, or by typing its index number.\n<\/p><p>There are two ways to create aliases:\n<\/p><ul><li>From Mutt, press <code>a<\/code> when an e-mail of the targetted person if selected.<\/li>\n<li>Edit the alias_file manually. The syntax is really simple:<\/li><\/ul><pre>alias nickname Long Name &lt;my-friend@domain.tld&gt;\n<\/pre><h4>Abook<\/h4><p><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=abook\">abook<\/a> is a stand-alone program dedicated to contact management. It uses a very simple text-based interface and contacts are stored in a plain text, human-readable database. Besides the desired contact properties are extensible (birthday, address, fax, and so on).\n<\/p><p>Abook is specifically designed to be interfaced with Mutt, so that it can serve as a full, more featured replacement of Mutt internal aliases. If you want to use Abook instead of aliases, remove the aliases configuration in <code>.muttrc<\/code> and add this:\n<\/p><pre>muttrc<\/pre><pre>## Abook\nset query_command= \"abook --mutt-query '%s'\"\nmacro index,pager  a \"&lt;pipe-message&gt;abook --add-email-quiet&lt;return&gt;\" \"Add this sender to Abook\"\nbind editor        &lt;Tab&gt; complete-query\n<\/pre><p>See the man pages <code>abook<\/code> and <code>abookrc<\/code> for more details and a full configuration sample.\n<\/p><h4>Goobook<\/h4><p>Goobook allows you to search your Google contacts from the command line or from within Mutt and can be installed with the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/goobook-git\/\">goobook-git<\/a><sup><small>AUR<\/small><\/sup> package.\n<\/p><p>Before using goobook you must configure <code>~\/.goobookrc<\/code>.  To generate the default template:\n<\/p><pre>$ goobook config-template &gt; ~\/.goobookrc\n<\/pre><p>See <code>~\/.goobookrc<\/code> for configuration options. At a minimum, you will need to enter your <b>email<\/b> and <b>password<\/b>.\n<\/p><p><strong>Note:<\/strong> If you have two-step verification enabled with the Google account, you may need to generate an application password.<\/p><p>If you want to use Goobook instead of aliases, remove any alias configuration in <code>.muttrc<\/code> and add:\n<\/p><pre>muttrc<\/pre><pre>## GooBook\nset query_command=\"goobook query '%s'\"\nmacro index,pager a \"&lt;pipe-message&gt;goobook add&lt;return&gt;\" \"add sender to google contacts\"\nbind editor &lt;Tab&gt; complete-query\n<\/pre><p>When composing an email message within mutt, <code>Tab<\/code> will now search your Google contacts.  While viewing messages <code>a<\/code> will add the sender to Google contacts.\n<\/p><h4>Khard<\/h4><p><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=khard\">khard<\/a> is a command-line addressbook that is able to sync with CardDAV-servers.\n<\/p><h3>Manage multiple sender accounts<\/h3><p>If you use multiple sender accounts, you can automatically associate a specific sender account with a recipient. <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/aur.archlinux.org\/packages\/mutt-vid\/\">mutt-vid<\/a><sup><small>AUR<\/small><\/sup> scans sent emails for the most recent \"From\" details associated with specific recipients, saving these in a file for mutt to source. The next time you email this recipient, mutt will automatically invoke a send-hook with the same email address and real name that you used previously. See mutt-vid's <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/gitlab.com\/protist\/mutt-vid\">homepage<\/a> for more details.\n<\/p><h3>Request IMAP mail retrieval manually<\/h3><p>If you do not want to wait for the next automatic IMAP fetching (or if you did not enable it), you might want to fetch mails manually. There is a mutt command <code>imap-fetch-mail<\/code> for that. Alternatively, you could bind it to a key:\n<\/p><pre>bind index \"^\" imap-fetch-mail\n<\/pre><h3>Avoiding slow index on large (IMAP) folders due to coloring<\/h3><p>Index highlighting by regex is nice, but can lead to slow folder viewing if your regex checks the body of the message.\n<\/p><p>Use folder-hook for only highlighting in for example the inbox (if you manage to empty your mailbox effiently):\n<\/p><pre>folder-hook . 'uncolor index \"~b \\\"Hi Joe\\\" ~R\u00a0!~T\u00a0!~F\u00a0!~p\u00a0!~P\"'\nfolder-hook \"\"!\"\" 'color index brightyellow black \"~b \\\"Hi Joe\\\" ~N\u00a0!~T\u00a0!~F\u00a0!~p\u00a0!~P\"'\n<\/pre><h3>Speed up folders switch<\/h3><p>Add this to your <code>.muttrc<\/code>:\n<\/p><pre>set sleep_time = 0<\/pre><h3>Archive treated e-mails<\/h3><p>When you read an e-mail, you have four choices: Answer it, Flag it, Archive it or Delete it. If you have this in mind, you can keep your inbox slim and fit with this macro (set up for Gmail):\n<\/p><pre>macro index \\' \"&lt;tag-pattern&gt;~R\u00a0!~D\u00a0!~F&lt;enter&gt;\\\n&lt;tag-prefix&gt;&lt;save-message&gt;+[Gmail]\/All &lt;enter&gt;\" \\\n\"Archive\"\n<\/pre><h3>Migrating mails from one computer to another<\/h3><p>In case you are transfering your mails to a new machine (copy&amp;paste), you probably need to delete the header cache (a file or folder like <code>~\/.cache\/mutt<\/code> if you followed the above configuration)  to make Mutt able to read your migrated E-Mails. Otherwise Mutt may freeze.\n<\/p><p>Note that if you had a folder created for you header cache, all mailboxes will have their own cache file, so you can delete caches individually without having to remove everything.\n<\/p><h3>Default folder for saving attachments<\/h3><p>By default Mutt will save attachments to the folder it was started from. If you want to always set the default destination to <code>~\/attachments<\/code>, you can create the following alias, which launches Mutt in this folder:\n<\/p><pre>alias mutt='cd ~\/attachments &amp;&amp; mutt'\n<\/pre><h3>Pager behavior<\/h3><p>Show context lines when going to next page:\n<\/p><pre>set pager_context=3\n<\/pre><p>Stop at the end instead of displaying next mail:\n<\/p><pre>set pager_stop=yes\n<\/pre><h3>Fast reply<\/h3><p>By default Mutt will ask to confirm the recipient and the subject when you reply to an e-mail. It will also ask if you want to include the original mail in your answer. If you assume you will always stick to the default values, you can set up Mutt to skip these questions:\n<\/p><pre>muttrc<\/pre><pre>set fast_reply=yes\nset include=yes\n<\/pre><p>You can still edit the recipient and the subject before sending.\n<\/p><h3>Ignore own e-mail addresses from group-reply<\/h3><p>Mutt will include your e-mail address(es) in the recipient list when you group-reply to a mail you were CC'ed. You can ask Mutt to ignore some addresses with:\n<\/p><pre>alternates mail1@server1|mail2@server2|...\n<\/pre><h3>IMAP message cache<\/h3><p>When using the built-in IMAP support, e-mails are fetched in memory by default. Retrieving a big e-mail several times will download it from your IMAP server that many times.\n<\/p><p>Alternatively, you can ask Mutt to store all fetched messages on disk:\n<\/p><pre>muttrc<\/pre><pre>set message_cachedir=~\/.cache\/mutt\/messages\n<\/pre><p>(The folder must exist.) This will make any future retrieval instantaneous, even with big attachments.\n<\/p><p>If you want to purge the cache from its oldest e-mails exceeding a limit of, say, 50MB, you can use a script like the following:\n<\/p><pre>~\/.mutt\/purgecache.sh<\/pre><pre>#!\/bin\/sh\n## In KB.\nCACHE_LIMIT=51200\ncd \"$1\" 2&gt;\/dev\/null\n[ $? -ne 0 ] &amp;&amp; exit\n[ $(du -s . | cut -f1 -d'\t') -lt $CACHE_LIMIT ] &amp;&amp; exit\nwhile IFS= read -r i; do\n\trm \"$i\"\n\t[ $(du -s . | cut -f1 -d'\t') -lt $CACHE_LIMIT ] &amp;&amp; exit\ndone &lt;&lt;EOF\n$(find . -type f -exec ls -rt1 {} +)\nEOF\n<\/pre><p>and call it on startup:\n<\/p><pre>muttrc<\/pre><pre>set message_cachedir=~\/.cache\/mutt\/messages\nsource \"~\/.mutt\/purgecache.sh '$message_cachedir'|\"\n<\/pre><h2>Troubleshooting<\/h2><h3>Backspace does not work in Mutt<\/h3><p>This is a common problem with some xterm-like terminals. Two solutions:\n<\/p><ul><li>Either rebind the key in <code>.muttrc<\/code><\/li><\/ul><pre>bind index,pager ^? previous-line\n<\/pre><p>Note that <code>^?<\/code> is one single character representing backspace in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Caret_notation\" class=\"extiw\" title=\"wikipedia:Caret notation\">caret notation<\/a>. To type in Emacs, use <code>Ctrl+q Backspace<\/code>, in Vim <code>Ctrl+v Backspace<\/code>.\n<\/p><ul><li>Or fix your terminal:<\/li><\/ul><pre>$ infocmp &gt; termbs.src\n<\/pre><p>Edit <code>termbs.src<\/code> and change <code>kbs=^H<\/code> to <code>kbs=\\177<\/code>, then:\n<\/p><pre>$ tic -x termbs.src\n<\/pre><h3>The <i>change-folder<\/i> function always prompt for the same mailbox<\/h3><p>This is not a bug, this is actually an intended behaviour. See the <a href=\"#Multiple_accounts\">multiple accounts section<\/a> for a workaround.\n<\/p><h3>I cannot change folder when using Mutt read-only (Mutt -R)<\/h3><p>This is certainly because you are using macros like this one:\n<\/p><pre>macro index,pager &lt;f2&gt; '&lt;sync-mailbox&gt;&lt;enter-command&gt;source ~\/.mutt\/personal&lt;enter&gt;&lt;change-folder&gt;!&lt;enter&gt;'\n<\/pre><p>This macro tells Mutt to sync (which is a write operation) before switching.\n<\/p><p>Either use the <a href=\"#Sidebar\">sidebar<\/a> or set another macro:\n<\/p><pre>macro index,pager &lt;f3&gt; '&lt;enter-command&gt;source ~\/.mutt\/personal&lt;enter&gt;&lt;change-folder&gt;!&lt;enter&gt;'\n<\/pre><h3>Error sending message, child exited 127 (Exec error.).<\/h3><p>This is an SMTP error. It means that mutt does not know how to send the message. You can either try installing sendmail and see if that solves your issue, or you can set the  smtp_url variable. If you use gmail, you can add the following to your muttrc to tell mutt to use gmails smtp server.\n<\/p><pre>set smtp_url=smtps:\/\/$imap_user:$imap_pass@smtp.gmail.com\n<\/pre><p>Take note of the smtps protocol, it is important. This should solve the problem.\n<\/p><h3>Character encoding problems<\/h3><p>If you are having problems with character encoding, first read <a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/dev.mutt.org\/trac\/wiki\/MuttFaq\/Charset\">this section<\/a> in the Mutt wiki.\n<\/p><p>If Chinese text is still garbled, it may help to decode with GBK even when GB2312 is specified in the header. You can do this with <code>iconv<\/code> by adding the following to your <code>mailcap<\/code> file:\n<\/p><pre>text\/plain; iconv -f gbk -t utf-8\u00a0%s; test=echo \"%{charset}\" | grep -ic \"gb2312\"; copiousoutput;\n<\/pre><p>and enabling it by adding a line to your <code>.muttrc<\/code>:\n<\/p><pre>auto_view text\/plain\n<\/pre><p>For HTML emails, you can edit the relevant line of your mailcap by replacing <code>%{charset}<\/code> with <code>$(echo %{charset} | sed s\/gb2312\/gbk\/I)<\/code>, for example:\n<\/p><pre>text\/html; w3m -dump -I $(echo\u00a0%{charset} | sed s\/gb2312\/gbk\/I)\u00a0%s; nametemplate=%s.html; copiousoutput\n<\/pre><h3>Unable to login with GMail<\/h3><p>Gmail disables access from apps it considers less secure, including <code>mutt<\/code>. You can enable access by following the instructions <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/support.google.com\/accounts\/answer\/6010255\">here<\/a>\n<\/p><h3>Not possible to open too long URLs with urlview<\/h3><p>Too long URLs are not parsed correctly, because urlview does not decode text (see <a rel=\"nofollow\" class=\"external autonumber\" href=\"https:\/\/bugs.debian.org\/cgi-bin\/bugreport.cgi?bug=127090\">[2]<\/a>). You can let mutt decode the e-mails instead. Replace the line for opening urlview with the following code:\n<\/p><pre>macro index \\cb \"\\\n:set my_tmp_pipe_decode=\\$pipe_decode\\n\\\n:set pipe_decode\\n\\\n|urlview\\n\\\n:set pipe_decode=\\$my_tmp_pipe_decode\\n\\\n:unset my_tmp_pipe_decode\\n\" \\\n'call urlview to extract URLs out of a message'\n<\/pre><h2>Documentation<\/h2><p>Newcomers may find it quite hard to find help for Mutt. Actually most of the topics are covered in the official documentation. We urge you to read it.\n<\/p><ul><li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mutt.org\/doc\/manual\/\">The official manual<\/a>. The stock <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=mutt\">mutt<\/a> package for Arch Linux also installs the HTML and plain text manual at <code>\/usr\/share\/doc\/mutt\/<\/code>.<\/li>\n<li>The <code>mutt<\/code> and <code>muttrc<\/code> man pages.<\/li><\/ul><h2>See also<\/h2><ul><li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mutt.org\/\">The official Mutt website<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mutt.org\/doc\/manual\/\">Official Manual<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/gitlab.com\/muttmua\/mutt\/wikis\/home\/\">The Mutt wiki<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/pbrisbin.com\/posts\/two_accounts_in_mutt\/\">Brisbin's great guide on how to setup different IMAP accounts with Mutt, offlineimap, msmtp<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/srobb.net\/mutt.html\">A Quick Guide to Mutt<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/stevelosh.com\/blog\/2012\/10\/the-homely-mutt\/\">Steve Losh on Mutt, offlineimap, msmtp, notmuch (focused on Gmail)<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.muttrcbuilder.org\/\">muttrc builder<\/a><\/li><\/ul>","created_at":"2018-11-07T16:52:02+0000","updated_at":"2018-11-07T16:52:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":39,"domain_name":"wiki.archlinux.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":567,"title":"Ask HN: How long did it take you to go from side project idea to launch?","url":"https:\/\/news.ycombinator.com\/item?id=15082901","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15082901\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15082901\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=15082901\" class=\"storylink\">Ask HN: How long did it take you to go from side project idea to launch?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        96 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=igammarays\" class=\"hnuser\">igammarays<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15082901\">on Aug 23, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15082901&amp;goto=item%3Fid%3D15082901\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20How%20long%20did%20it%20take%20you%20to%20go%20from%20side%20project%20idea%20to%20launch%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20How%20long%20did%20it%20take%20you%20to%20go%20from%20side%20project%20idea%20to%20launch%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15082901&amp;auth=c2c4b0cef20727f818dde6e0101c0f4d0b4e405d\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15082901\">89\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>I've been working on a SaaS product side project for about 4 months now. Progress feels painfully slow, but I wonder what typical one-man consumer SaaS projects look like. Years of development? 6 months average? Specific examples would be helpful so that I can compare the scope\/size of the idea to mine.<\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:52:01+0000","updated_at":"2018-11-07T16:52:01+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":566,"title":"Prefilling Github Issues","url":"https:\/\/eric.blog\/2016\/01\/08\/prefilling-github-issues\/","content":"<p>Earlier today, I created an internal \u201cCall for testing\u201d post where I essentially asked other Automatticians to break a piece of functionality that I intend to launch.<\/p>\n<p>To help with the testing process, I included a link to a pre-filled Github issue. For the most part, pre-filling the issue with a label, milestone, etc. was pretty easy, but there were a few gotchas.<\/p>\n<p>So, below I\u2019ll break down how to prefill as much of the Github issue form as possible.<\/p>\n<h3>Creating issues<\/h3>\n<p>First of all, you should know that to create an issue for a repository, that you would go to a URL like:<\/p>\n<p>https:\/\/github.com\/{owner]\/{repo_name}\/issues\/new<\/p>\n<p>As long as you\u2019re logged in, when you go to the above URL, you\u2019ll be shown a form to create a new issue. And since the new issue form is in the same place for each repo, you can share this link (and even prefill it) to make life a little bit easier for those that are using and testing your software.<\/p>\n<h3>The secret formula<\/h3>\n<p><a href=\"https:\/\/i1.wp.com\/manofhustle.com\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?strip=all&amp;ssl=1\"><img data-attachment-id=\"3715\" data-permalink=\"https:\/\/eric.blog\/2016\/01\/08\/prefilling-github-issues\/github-new-issue-screen\/\" data-orig-file=\"https:\/\/i0.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?fit=1024%2C640&amp;strip=all&amp;ssl=1\" data-orig-size=\"1024,640\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"github-new-issue-screen\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?fit=300%2C188&amp;strip=all&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?fit=712%2C445&amp;strip=all&amp;ssl=1\" class=\"aligncenter wp-image-3715 size-large\" src=\"https:\/\/i2.wp.com\/manofhustle.com\/wp-content\/uploads\/2016\/01\/github-new-issue-screen-1024x640.jpg?resize=712%2C445&amp;strip=all&amp;ssl=1\" alt=\"github-new-issue-screen\" width=\"712\" height=\"445\" srcset=\"https:\/\/i0.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?w=1024&amp;strip=all&amp;ssl=1 1024w, https:\/\/i0.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?resize=300%2C188&amp;strip=all&amp;ssl=1 300w, https:\/\/i0.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-screen.jpg?resize=768%2C480&amp;strip=all&amp;ssl=1 768w\" data-recalc-dims=\"1\" \/><\/a><\/p>\n<p>In the above screenshot, you can see that there are 5 pieces of information that can be filled out in a new Github issue. These are:<\/p>\n<ul><li>Title<\/li>\n<li>Description<\/li>\n<li>Labels<\/li>\n<li>Milestone<\/li>\n<li>Assignee<\/li>\n<\/ul><p>Now, to prefill these all we need to do is add parameters to the query string. Here\u2019s an example:<\/p>\n<p><a href=\"https:\/\/github.com\/Automattic\/wp-calypso\/issues\/new?labels[]=People%20Management&amp;labels[]=[Type]%20Bug&amp;title=People:&amp;milestone=People%20Management:%20m6&amp;assignee=ebinnion&amp;body=This%20is%20a%20prefilled%20issue\">https:\/\/github.com\/Automattic\/wp-calypso\/issues\/new?labels[]=People%20Management&amp;labels[]=[Type]%20Bug&amp;title=People:&amp;milestone=People%20Management:%20m6&amp;assignee=ebinnion&amp;body=This%20is%20a%20prefilled%20issue<\/a><\/p>\n<p>Note that the values in the URL above are also URL encoded.<\/p>\n<h3>Gotchas<\/h3>\n<p>In the above link note that there is no <code>&amp;amp;description<\/code>. That\u2019s because Github fills the description with whatever is in the <code>&amp;amp;body<\/code> parameter.<\/p>\n<p>Lastly, note the use of <code>labels[]=<\/code> when assigning multiple labels. If you want to just add one label, you should be able to just do <code>labels=<\/code>.<\/p>\n\n<div id=\"jp-relatedposts\" class=\"jp-relatedposts\">\n\t<h3 class=\"jp-relatedposts-headline\"><em>Related<\/em><\/h3>\n<\/div>","created_at":"2018-11-07T16:52:00+0000","updated_at":"2018-11-07T16:52:00+0000","published_at":"2016-01-09T03:11:04+0000","published_by":["View all posts by Eric Binnion"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"eric.blog","preview_picture":"https:\/\/i2.wp.com\/eric.blog\/wp-content\/uploads\/2016\/01\/github-new-issue-feature.jpg?fit=1024%2C640&strip=all&ssl=1","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":564,"title":"C.G. Jennings: Lindenmayer systems","url":"https:\/\/cgjennings.ca\/articles\/l-systems.html","content":"<h3>Natural processes as computation<\/h3><p><strong>Reading time:<\/strong> about 10 minutes<\/p><p class=\"lead\">Lindenmayer systems, or L-systems, can describe how living\nthings develop complex forms by the repeated application of simple rules.\nAristid Lindenmayer (1925\u20131989), the theoretical biologist who proposed L-systems in 1968, \nfirst used them to describe the growth patterns of algae.\nThis page is (mostly) algae-free, but it does let you explore Lindenmayer systems by making\nplant-like structures, classic fractals, and other designs.<\/p><h2>Introduction<\/h2><p>The natural world produces amazingly complex systems from simple starting\npoints. In the biology, computational models have played a critical role in\nhelping scientists understand how these complex systems can develop.\nL-systems are one of the better-known of these models.\nIn L-systems, structures develop through a process of <em>string rewriting<\/em>.\nA string of letters is transformed into a new string of letters using simple rules\ncalled <em>productions<\/em>. The process is repeated indefinitely, each time using\nthe string that was just produced as the source for the next string.\nBecause the strings tend to grow with each rewrite, an L-system can become\narbitrarily complex, but always guided by a simple process dictated by \na fixed set of simple rules.<\/p><p>Before getting into more detail, take a moment\nto play with the app below. Try changing the <strong>Order<\/strong> and\ndrawing the result. Try drawing orders 0, 1, 2, 3, and 4 in order and observe\nthe progression. Then try some other example systems\nto get a feel for what\u2019s possible (there are over 50!). Feel free to play,\njust know that behind the scenes, L-system strings can grow very quickly\nas the order increases. The longer the string, the longer it will take to draw\nthe result! Your browser may even run out of memory, which will\n\u201ccrash\u201d the page\u2014but with no real harm done. Just navigate back to the\npage and choose different settings!<\/p><div class=\"container-fluid\" id=\"applet\"><div class=\"row\"><div class=\"col-sm-6 col-md-4\"><div class=\"bs-component\"><div id=\"system-box\" class=\"panel panel-primary\"><p><\/p><h3 class=\"panel-title\"><label for=\"system\">L-system<\/label><\/h3><div class=\"panel-body\"><p><label for=\"order\">Order <\/label>\n            \u00a0\u00a0\n            <\/p><\/div><\/div><div class=\"form-group\"><div class=\"panel panel-primary\"><p><\/p><h3 class=\"panel-title\">Examples<\/h3><\/div><\/div><\/div><\/div><\/div><\/div><h2>How L-systems work<\/h2><p>L-system repeatedly rewrite a string according to some simple rules:\nYou start from a string called the <em>axiom<\/em>.\nYou can think of it like a seed: it is the starting point that the system grows from.\nRegulating that growth is a set of rules, called <em>productions<\/em>.\nBy applying the produtions to the axiom, you will make new strings that\nrepresent one \u201cstep\u201d of growth.\nAll of the strings you can produce are said to be in the <em>language<\/em> of\nthe L-system. What\u2019s more, you can take any string that you make in this way,\nre-apply the same productions to it, and get a new string that is also\n\u201cin the language\u201d of the L-system.<\/p><p>Let\u2019s look at a simple example:<\/p><pre>axiom Y\nY \u2192 XYX<\/pre><p>This system has only production, the rule <code>Y \u2192 XYX<\/code>. To apply it, you scan a\nstring that is known to be in the system\u2019s language from left to right, and\nanywhere you see the pattern on the left of the <code>\u2192<\/code> symbol, you replace it with\nthe pattern on the right side. Let\u2019s see how that would work.\nTo start with, we only know of one string in the language, the axiom <code>Y<\/code>. But we\nsee that we can produce <code>XYX<\/code> by applying the production <code>Y \u2192 XYX<\/code>. So now we know\nthat <code>XYX<\/code> is also in the language, and we can use it to produce a new string, too.\nThere is no production with <code>X<\/code> on the left side, so those won\u2019t change. But the <code>Y<\/code> in the middle\ncan be changed to <code>XYX<\/code> again, giving the string <code>XXYXX<\/code>. That\u2019s a new string in the\nlanguage, so you can rewrite it, too, and so on.<\/p><h2>Ordered L-systems<\/h2><p>With this system, you can soon see that each time you apply the production you will get exactly one new string\nthat is just like the one you started with but with another <code>X<\/code> added to either\nside of the <code>Y<\/code> in the middle. When an L-system can only make one new string\neach time it is rewritten, we often talk about the system\u2019s \u201corder <em>n<\/em>\u201d\nstring. The order 0 string is the axiom. The order 1 string is made from the\norder 0 string. The order 2 string is made from the order 1 string, and so on.\nThat\u2019s why the app above has a control named <strong>Order<\/strong>: it uses\nthe string of that order to produce the drawing.<\/p><p>Not all L-systems work this way. For example, what would happen if we extend\nthe system above as follows?<\/p><pre>axiom Y\nY \u2192 XYX\nYX \u2192 XY<\/pre><p>This system starts the same, producing <code>Y<\/code> and then <code>XYX<\/code> as before. After that,\nyou run into a problem: two different rules can be applied at the same spot.\nYou can replace the <code>Y<\/code> with <code>XYX<\/code>, or you can replace the <code>Y<\/code> <em>and<\/em> the\nfollowing <code>X<\/code> with <code>XY<\/code>. Which to choose?\nThere are different ways that this can be handled, but what if you allow\n<strong>both<\/strong> possibilities? Then instead of having\none definitive string for each generation, you have sets of strings leading\nto more sets of strings! As you can imagine, systems that allow rules like these\ntend to become very complex, very quickly. On this page we eliminate this\nproblem by restricting the app to a subtype of L-system called a\nd0L-system. A d0L-system has a couple of changes which will guarantee\nthat our systems are nicely ordered:<\/p><ol><li> the left hand side of a production can only be a single letter; and\n  <\/li><li> no two productions can have the same left-hand side.\n<\/li><\/ol><p>The rest of this article deals exclusively with d0L-systems.\nDon\u2019t feel cheated\u2014they still offer lots of possibilities.<\/p><h2>From strings to nature<\/h2><p>So far we\u2019ve seen that L-systems can grow <em>strings<\/em> very well.\nHow does that connect to the natural world? We can get a hint that the connection\nis there by going back to Lindenmayer\u2019s original work with algae growth patterns.\nHere is a system from his original paper:<\/p><pre>axiom B\nA \u2192 AB\nB \u2192 A<\/pre><p>If you write the first 8 orders of this system you\u2019ll get <code>B<\/code>, <code>A<\/code>, <code>AB<\/code>, <code>ABA<\/code>,\n<code>ABAAB<\/code>, <code>ABAABABA<\/code>, <code>ABAABABAABAAB<\/code> and <code>ABAABABAABAABABAABABA<\/code>. Notice a pattern?\nIf not, look closely at the <em>length<\/em> of each of these strings:\n1, 1, 2, 3, 5, 8, 13, 21. Can you see why the strings produced by this system\nare called <em>Fibostrings<\/em>? Quite apart from how algae grow, you\nprobably remember learning that the Fibonacci number sequence shows up\noften in the natural world, most famously in the way that\nleaves and scales are arranged on plants\n(<a href=\"https:\/\/en.wikipedia.org\/wiki\/Phyllotaxis\">phyllotaxis<\/a>).<\/p><h2>Enter the turtle<\/h2><p>Still, as interesting as it is to whip yourself up some Fibostrings, it is\nstill a big leap from that to the images produced by the app above.\nTo get there, we need to give the strings a geometric interpretation.\nThe demo does this using \u201cturtle graphics\u201d, on odd name which comes from the\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Logo_(programming_language)\">Logo programming language<\/a>.\nTurtle graphics work like this: Imagine you\u2019re a turtle on an infinitely large\npiece of paper. Not just any turtle, but one that can understand simple\ninstructions and that has a basic understanding of angles and distances.\nFinally, imagine that you (the unusually bright turtle, that is) also have a\npen in your mouth.<\/p><div class=\"row\"><div class=\"col-sm-6\"><p>Have you got that image in mind? Good.<\/p><p>Now suppose I give you instructions like this:\n  <\/p><ol><li>Touch the pen to the paper.<\/li>\n    <li>Move forward 1 step, turn left 90\u00b0.<\/li>\n    <li>Move forward 1 step, turn left 90\u00b0<\/li>\n    <li>Move forward 1 step, turn left 90\u00b0<\/li>\n    <li>Move forward 1 step.<\/li>\n  <\/ol><p>You\u2019ve just drawn a square. What a good turtle you are.<\/p><\/div><div class=\"col-sm-6\"><div class=\"text-center\"><p><img src=\"https:\/\/cgjennings.ca\/media\/articles\/turtle-graphics.jpg\" width=\"300\" height=\"225\" alt=\"Photo of a turtle robot that has made a design on a sheet of paper\" \/><\/p><p><small>Credit: Valiant Technology Ltd. (<a href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/3.0\">CC BY-SA 3.0<\/a>)<\/small><\/p><\/div><\/div><\/div><p>If we associate symbols in our L-system strings with instructions like the ones\nabove, then a hard-working (virtual) turtle can use those strings to draw pictures. Here is an\nexample: (Notice that <code>F<\/code> produces an empty string. That just means we will\ndelete any <code>F<\/code> symbols when we rewrite the string.)<\/p><pre>axiom FX\nX \u2192 -FX++FY-\nY \u2192 +FX--FY+\nF \u2192<\/pre><p>Now if the turtle interprets the symbol <code>F<\/code> to mean \u201cmove forward\n 1 step with pen down\u201d, the symbol <code>-<\/code> to mean\n\u201cturn right 45\u00b0\u201d and the symbol <code>+<\/code> to mean \u201cturn left 45\u00b0\u201d then the order 8\nstring of this system will make the turtle draw this picture:<\/p><p class=\"text-center\"><img src=\"https:\/\/cgjennings.ca\/media\/articles\/dragon.svg\" width=\"50%\" alt=\"image\" \/><\/p><h2>Trees and branches<\/h2><p>If you are interested in fractals, you may recognize this as a <em>dragon curve<\/em>.\nPretty neat, but it still doesn\u2019t give us much insight into how plants grow!\nFor that we need to extend the turtle\u2019s capabilities a little.\nTo see why, it is useful to consider this rule for drawing a simple tree:<\/p><blockquote><div>\nTo draw a tree, draw a trunk with two shorter trees coming out of the top\nat right angles to each other.\n<\/div><\/blockquote><p>Notice that this definition is self-referential or <em>recursive<\/em>:\npart of a tree is two shorter trees! How can we make any sense of that?\nOne step at a time. Start by drawing a trunk: a straight vertical line.\nNow since a tree has \u201ctwo shorter trees coming out\u201d of the trunk, start\noff those trees by drawing two shorter trunks to make a Y shape.\nBut don\u2019t stop there. Since each branch of the Y is the trunk of a new tree, you\nneed to add two shorter trunks to them as well. And since each of <em>those<\/em>\nnew branches is also the start of a tree, they need trunks\u2026 you can continue\nas many times as you like, each time extending the tree a little in the same way that\nthe next string in an L-system builds on the previous one.\nAfter the first few steps, your tree should have progressed through these stages:<\/p><p class=\"text-center\"><img src=\"https:\/\/cgjennings.ca\/media\/articles\/y-tree-steps.png\" width=\"50%\" alt=\"image\" \/><\/p><p>And if you keep on going you will end up with something like this:<\/p><p class=\"text-center\"><img src=\"https:\/\/cgjennings.ca\/media\/articles\/y-tree.png\" width=\"20%\" alt=\"image\" \/><\/p><p>Which is indeed rather tree-like! So, what new capabilities does our turtle\nneed to recreate this process? The recursion is no problem: in fact, all\nof the L-systems above use recursion! It just means that a symbol on the\nleft-hand side of a production rule also appears somewhere in the that rule's\nright-hand expansion (either immediately or in a later order).\nOne thing that\u2019s missing is changing the turtle\u2019s step\nsize so we can make <em>shorter<\/em> trees on top of our trunks. No problem,\nwe can add a new symbol, <code>@<\/code>, for that.\n(You can optionally follow <code>@<\/code> with a number\nthat you want to mulitply the step size by.)<\/p><p>We will also need a way to give\nthe turtle some kind of <em>memory<\/em>. Now this one is a little trickier!\nWe need a memory in order to create the two branches: after we draw\nthe first branch, we need a way to get back to where we started so we can draw\nthe second one.<\/p><p>The turtle\u2019s memory will be controlled with two new symbols:\n<code>[<\/code> and <code>]<\/code>. The <code>[<\/code> symbol tells the turtle\nto write down its current <em>state<\/em> on a note and put that note on top\nof a special note stack. The <code>]<\/code> symbol tells the turtle to take the\ntop note off the stack and go back to the same state that is written there.\n(The turtle\u2019s state includes its position, direction, step size, and other information:\nanything that would change how it draws a line.)<\/p><p>Now, actual plants don\u2019t have a memory, of course, but we only have one\nturtle to draw the whole tree. On a real plant each branch starts growing\nfrom a bud that is effectively a new turtle just for that branch. In fact, you might \nfind it helpful to think of <code>[<\/code> and <code>]<\/code> as representing\na bud, and everything between those symbols as what grows from it.<\/p><p>Whew! Putting this all together, we can finally recreate our simple tree as an\nL-system:<\/p><pre>axiom FX\nX \u2192 @[+FX][-FX]\n<\/pre><p>Let\u2019s look at each symbol in turn: The axiom is <code>F<\/code> (draw\na line for the trunk) followed by <code>X<\/code> (a placeholder that will\nbe replaced by \u201ctwo shorter trees\u201d). The production will rewrite the <code>X<\/code>\nas <code>@<\/code> (make the step size shorter), <code>[<\/code> (start a new branch),\n<code>+<\/code> (turn left 45\u00b0), <code>FX<\/code> (a new trunk and tree placeholder),\n<code>]<\/code> (finish the branch and go back to where it started), \n<code>[<\/code> (start a second new branch), <code>-<\/code> (turn right 45\u00b0),\n<code>FX<\/code> (the second new tree), <code>]<\/code> (finish the second branch).\nThe turtle instructions are an exact analogue of our tree-drawing rule!<\/p><p class=\"text-center\"><img src=\"https:\/\/cgjennings.ca\/media\/articles\/turtle-animation.gif\" width=\"20%\" alt=\"image\" \/><\/p><p>By the way, the second <code>[<\/code> and <code>]<\/code> in that production\nare not actually necessary. You will get the same series of drawings with\n<code>@[+FX]-FX<\/code>. I leave it an as exercise for you to figure out why.<\/p><h2>Playtime!<\/h2><p>Now you are ready <a href=\"#applet\">to go back to the app<\/a> and explore more examples or try creating systems\nof your own. There are just a couple of things to note: First, when defining a\nproduction, the app uses <code>=<\/code> instead of <code>\u2192<\/code> (it\u2019s\na lot easier to type!). And second, by default, the turtle will turn in 45\u00b0 increments.\nTo change that add a line like this: <code>angle <em>n<\/em><\/code>,\nwhere <em>n<\/em> is the number of turns that should make up a complete circle.\n(There are more commands, but that\u2019s enough to get started. When you are ready,\na complete list of them follows below.)<\/p><p>Happy string rewriting!<\/p><h2>Command reference<\/h2><p>The commands used by the app have been chosen to be\nmostly compatible with the fractal software <em>Fractint<\/em>, though the\ntwo systems are not identical. One difference of note is that this system is\ncase sensitive, while Fractint is not.<\/p><h3>Preamble directives<\/h3><table class=\"table table-striped\"><caption>Commands that affect how systems are drawn<\/caption>\n  <thead><\/thead>\n    \n  <tbody><tr><td><code>; ...<\/code><\/td>\n      <td>Lines that start with a ; are considered comments and are ignored by the computer.<\/td>\n    <\/tr><tr><td><code>angle\u00a0<em>n<\/em><\/code>\u00a0or<br \/><code>turns\u00a0<em>n<\/em><\/code><br \/>Default: 8<\/td>\n      <td>Sets the number of turns that make up a complete circle to <em>n<\/em>. (Each turn will be by 360\u00b0\/<em>n<\/em>.)<\/td>\n    <\/tr><tr><td><code>rotate\u00a0<em>n<\/em><\/code><br \/>Default: 0<\/td>\n      <td>The entire drawing will be rotated by <em>n<\/em>\u00b0. The turtle normally begins pointing to the right.\n          For example, to start with the turtle pointing up, use <code>rotate 90<\/code>.<\/td>\n    <\/tr><tr><td><code>order\u00a0<em>n<\/em><\/code><br \/>Default: 4<\/td>\n      <td>For the included examples, this sets the initial order used when the system is selected from the list.<\/td>      \n    <\/tr><tr><td><code>quality\u00a0<em>low|high<\/em><\/code><br \/>Default: high<\/td>\n      <td>Provides a hint as to whether the drawing should favour quality or speed. (The exact effect depends on the browser, but the low quality setting typically has no antialiasing.)<\/td>\n    <\/tr><\/tbody><\/table><h3>Productions<\/h3><table class=\"table table-striped\"><caption>Commands that determine which d0L-system is produced<\/caption>\n  <thead><\/thead>\n    \n  <tbody><tr><td><code>axiom\u00a0<em>s<\/em><\/code><\/td>\n      <td>The string <em>s<\/em> to use for order 0.<\/td>\n    <\/tr><tr><td><code><em>p<\/em>\u00a0=\u00a0<em>s<\/em><\/code><\/td>\n      <td>When a string is being rewritten, every instance of the letter <em>p<\/em>\n      will be replaced by string <em>s<\/em>.\n      The right hand side can be empty, which removes the left-hand letter from the new string.<\/td>\n    <\/tr><\/tbody><\/table><h3>Drawing<\/h3><table class=\"table table-striped\"><caption>Symbols that have special meaning to the turtle when drawing a string<\/caption>\n  <thead><\/thead>\n    \n  <tbody><tr><td><code>F<\/code><\/td>\n      <td>Moves the turtle forward 1 step with the pen down, drawing a line.<\/td>\n    <\/tr><tr><td><code>G<\/code><\/td>\n      <td>Moves the turtle forward 1 step with the pen up, leaving no mark.<\/td>\n    <\/tr><tr><th colspan=\"2\">Turning<\/th>\n    <\/tr><tr><td><code>+<\/code><\/td>\n      <td>Turns the turtle \u201cleft\u201d one notch without moving. The amount of the turn is determined by <code>angle<\/code>.<\/td>\n    <\/tr><tr><td><code>-<\/code><\/td>\n      <td>Turns the turtle \u201cright\u201d one notch without moving.<\/td>\n    <\/tr><tr><td><code>!<\/code><\/td>\n      <td>Switches the meaning of \u201cleft\u201d and \u201cright\u201d.<\/td>\n    <\/tr><tr><td><code>|<\/code><\/td>\n      <td>Turns the turtle around (as close to 180\u00b0 as the <code>angle<\/code> value allows).<\/td>\n    <\/tr><tr><th colspan=\"2\">Line length<\/th>\n    <\/tr><tr><td><code>@<em>n<\/em><\/code><\/td>\n      <td>Multiplies the turtle\u2019s step length by <em>n<\/em>, changing the distance\n          moved by each <code>F<\/code> or <code>G<\/code> command. If no number is\n          given, the length is multipled by 0.6. The value <em>n<\/em> may be\n          preceded by the letter <code>I<\/code> to use 1\/<em>n<\/em>.\n          It may be preceded by <code>Q<\/code> to use \u221a<em>n<\/em>.<\/td>\n    <\/tr><tr><th colspan=\"2\">Branching<\/th>\n    <\/tr><tr><td><code>[<\/code><\/td>\n      <td>Adds the current turtle state to the stack. The current position, angle,\n          step length, colour, and whether left and right are reversed are all recorded.<\/td>\n    <\/tr><tr><td><code>]<\/code><\/td>\n      <td>Takes the turtle state off the top of the stack, returning the turtle to\n        the same state it was in when the most recent <code>[<\/code> was processed.\n        Note that, like brackets in math, the number of <code>[<\/code> and <code>]<\/code>\n        symbols in a string should match.<\/td>\n    <\/tr><tr><th colspan=\"2\">Pen colours (see below)<\/th>\n    <\/tr><tr><td><code>&gt;<em>n<\/em><\/code><\/td>\n      <td>Changes to the pen colour that is <em>n<\/em> steps after the current colour.\n          If no number follows the symbol, the colour changes by 1 step.<\/td>\n    <\/tr><tr><td><code>&lt;<em>n<\/em><\/code><\/td>\n      <td>Changes to the pen colour that is <em>n<\/em> steps before the current colour.\n          If no number follows the symbol, the colour changes by 1 step.<\/td>\n    <\/tr><tr><td><code>%<em>n<\/em><\/code><\/td>\n      <td>Changes to the pen colour to the <em>n<\/em>th pen in the colour palette.\n          If no number is given, changes to pen 0 (the starting colour).<\/td>\n    <\/tr><\/tbody><\/table><h3>Colour palette<\/h3><p>The turtle has a palette of 256 different coloured pens. The palette has been\nchosen to let you draw a wide variety of designs with nice results. It\nblends smoothly from brown to green, cyan, blue, indigo, violet, red, orange, pale yellow, and\nfinally back to brown. Using <code>&gt;<\/code> and <code>&lt;<\/code>\nyou can move through the palette in either direction; if you move past the\nfirst or last colour, you will wrap back around to the other end of the palette.<\/p><div class=\"palette\"><div class=\"inkwell\"><p>256 (same as 0)<\/p><\/div><\/div><h2>Bibliography<\/h2><ol><li>Aristid Lindenmayer. <strong>Mathematical models for cellular interaction\n      in development.<\/strong> <em>Journal of Theoretical Biology<\/em>, 18:280\u2013315, 1968.<\/li>\n  <li>Prusinkiewicz, Przemys\u0142aw and Aristid Lindenmayer. <a href=\"http:\/\/algorithmicbotany.org\/papers\/#abop\"><em>The Algorithmic Beauty of Plants<\/em><\/a>. Springer-Verlag, 1990.<\/li>\n<\/ol>","created_at":"2018-11-07T16:51:59+0000","updated_at":"2018-11-07T16:51:59+0000","annotations":[],"mimetype":"text\/html","language":"en_CA","reading_time":14,"domain_name":"cgjennings.ca","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":565,"title":"gnupg - How to export my Protonmail private key? - Information Security Stack Exchange","url":"https:\/\/security.stackexchange.com\/questions\/151062\/how-to-export-my-protonmail-private-key\/151954","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>From: <a href=\"https:\/\/github.com\/scastiel\/protonmail-export\" rel=\"noreferrer\">https:\/\/github.com\/scastiel\/protonmail-export<\/a><\/p>\n<blockquote>\n  <p>1) Open the ProtonMail app and log out completely. You should now see\n  the login screen.<\/p>\n  \n  <p>2) Open the dev tools of your browser, and the Network tab to see all\n  network calls.<\/p>\n  \n  <p>3) Enter your username and password and click Login button. You won't\n  need to go further.<\/p>\n  \n  <p>4) In the network calls, find the one to \u201c\/api\/auth\u201d. There may be\n  several ones, be sure to pick the one with \u201cPOST\u201d request method, not\n  the \u201cDELETE\u201d one (which doesn't contain any info anyway).<\/p>\n  \n  <p>5) In this network call, open the Response tab to see raw data\n  returned from the server, find the line beginning with \"PrivateKey\":,\n  and copy the rest of the line, from \"-----BEGIN PGP PRIVATE KEY to the\n  last \", without the trailing comma.<\/p>\n  \n  <p>6) Open the Console tab of the dev tools, type console.log(&lt;PASTE THE COPIED CONTENT HERE&gt;) then press enter.<\/p>\n  \n  <p>7) Copy the result of the command, and put it into a text file, that's\n  it you have your private key!<\/p>\n<\/blockquote>\n<p>But I am stuck at \"6)\". The<\/p>\n<pre>console.log(\"-----BEGIN PGP PRIVATE KEY B........----\")\n<\/pre>\n<p>part doesn't output anything (the \"....\" part is of course censored). <\/p>\n<p>So now.. I can format by hand the private key to look like a private key, I can even import it with \"seahorse\"! I thought I winned the game, but... when I try to: <\/p>\n<pre>gpg --encrypt --sign --armor -r foobar@protonmail.ch message.txt\n<\/pre>\n<p>It asks for a GPG key password. I use a password manager, so I copy\/paste the MAILBOX password that I use, but here is the catch: it doesn't accepts it!<\/p>\n<p><strong>Question:<\/strong> What am I missing? <\/p>\n<p>I also asked scastiel via mail, posting the answer here if he replies, but maybe someone else knows the solution too. <\/p>\n<p>In step \"5)\" I can see that there is a line named \"KeySalt\".. maybe Protonmail introduces salt after the howto from scastiel? <\/p>\n<p>Protonmail support wont help on exporting the private keys. <\/p>\n<p><strong>UPDATE<\/strong>: I wrote \"MAILBOX\" password, that means I still use Protonmail with two passwords: one for login, and one mailbox pw for the \"decryption\". <\/p><\/div><h2 data-answercount=\"5\">&#13;\n                                5 Answers&#13;\n                                5&#13;<\/h2>","created_at":"2018-11-07T16:51:59+0000","updated_at":"2018-11-07T16:51:59+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"security.stackexchange.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/security\/img\/apple-touch-icon@2.png?v=497726d850f9","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":563,"title":"Python For Finance: Algorithmic Trading","url":"https:\/\/medium.com\/datacamp\/python-for-finance-algorithmic-trading-60fdfb9bb20d","content":"<section class=\"section section--body section--first\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"5f67\" class=\"graf graf--h3 graf--leading graf--title\">Python For Finance: Algorithmic Trading<\/h1><p id=\"e905\" class=\"graf graf--p graf-after--h3\"><em class=\"markup--em markup--p-em\">Originally published at <\/em><a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading<\/em><\/a><\/p><p id=\"0936\" class=\"graf graf--p graf-after--p\">Technology has become an asset in finance: financial institutions are now evolving to technology companies rather than just staying occupied with just the financial aspect: besides the fact that technology brings about innovation the speeds and can help to gain a competitive advantage, the speed and frequency of financial transactions, together with the large data volumes, makes that financial institutions\u2019 attention for technology has increased over the years and that technology has indeed become a main enabler in finance.<\/p><p id=\"318e\" class=\"graf graf--p graf-after--p\">Among the hottest programming languages for finance, you\u2019ll find R and Python, alongside languages such as C++, C# and Java. In this tutorial, you\u2019ll learn how to get started with Python for finance. The tutorial will cover the following:<\/p><ul class=\"postList\"><li id=\"4858\" class=\"graf graf--li graf-after--p\">The <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#basics\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#basics\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">basics<\/a> that you need to get started: for those who are new to finance, you\u2019ll first learn more about the stocks and trading strategies, what time series data is and what you need to set up your workspace.<\/li><li id=\"ac03\" class=\"graf graf--li graf-after--li\">An introduction to time series data and some of the most <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#financialanalyses\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#financialanalyses\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">common financial analyses<\/a>, such as moving windows, volatility calculation,\u00a0\u2026 with the Python package Pandas.<\/li><li id=\"5bd6\" class=\"graf graf--li graf-after--li\">The development of a <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#tradingstrategy\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#tradingstrategy\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">simple momentum strategy<\/a>: you\u2019ll first go through the development process step-by-step and start off by formulating and coding up a simple algorithmic trading strategy.<\/li><li id=\"20ad\" class=\"graf graf--li graf-after--li\">Next, you\u2019ll <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#backtesting\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#backtesting\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">backtest<\/a> the formulated trading strategy with Pandas, zipline and Quantopian.<\/li><li id=\"b6ff\" class=\"graf graf--li graf-after--li\">Afterwards, you\u2019ll see how you can do <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#improving\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#improving\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">optimizations<\/a> to your strategy to make it perform better and you\u2019ll eventually <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#evaluating\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading#evaluating\" class=\"markup--anchor markup--li-anchor\" rel=\"nofollow noopener\" target=\"_blank\">evaluate<\/a> your strategy\u2019s performance and robustness.<\/li><\/ul><figure id=\"a496\" class=\"graf graf--figure graf-after--li\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*1Tz2TStuR1fbSCS-.png\" data-width=\"640\" data-height=\"360\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*1Tz2TStuR1fbSCS-.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"f53b\" class=\"graf graf--p graf-after--figure\">Download the Jupyter notebook of this tutorial <a href=\"https:\/\/github.com\/Kacawi\/datacamp-community\" data-href=\"https:\/\/github.com\/Kacawi\/datacamp-community\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><h3 id=\"a211\" class=\"graf graf--h3 graf-after--p\">Getting Started With Python for\u00a0Finance<\/h3><p id=\"03c5\" class=\"graf graf--p graf-after--h3\">Before you go into trading strategies, it\u2019s a good idea to get the hang of the basics first. This first part of the tutorial will focus on explaining the Python basics that you need to get started. This does not mean, however, that you\u2019ll start completely from zero: you should have at least done DataCamp\u2019s free <a href=\"https:\/\/www.datacamp.com\/courses\/intro-to-python-for-data-science\" data-href=\"https:\/\/www.datacamp.com\/courses\/intro-to-python-for-data-science\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Intro to Python for Data Science<\/a> course, in which you learned how to work with Python lists, packages and NumPy. Additionally, it is desired to already know the basics of Pandas, the well-known Python data manipulation package, but this is no requirement. If you do want to already get into Pandas before starting this tutorial, consider taking DataCamp\u2019s <a href=\"https:\/\/www.datacamp.com\/courses\/pandas-foundations\" data-href=\"https:\/\/www.datacamp.com\/courses\/pandas-foundations\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Pandas Foundations<\/a> course.<\/p><p id=\"4d44\" class=\"graf graf--p graf-after--p\">When a company wants to grow and undertake new projects or expand, it can issue stocks to raise capital. A stock represents a share in the ownership of a company and is issued in return for money. Stocks are bought and sold: buyers and sellers trade existing, previously issued shares. The price at which stocks are sold can move independent of the company\u2019s success: the prices instead reflect supply and demand. This means that, whenever a stock is considered as \u2018desirable\u2019, due to a success, popularity,\u00a0\u2026 the stock price will go up.<\/p><p id=\"3c41\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Note<\/strong> that stocks are not exactly the same as bonds, which is when companies raise money through borrowing, either as a loan from a bank or by issuing debt.<\/p><p id=\"2ad7\" class=\"graf graf--p graf-after--p\">As you just read, buying and selling or trading is essential when you\u2019re talking about stocks, but certainly not limited to it: trading is the act of buying or selling <em class=\"markup--em markup--p-em\">an asset<\/em>, which could be financial security, like stock, a bond or a tangible product, such as gold or oil.<\/p><p id=\"704e\" class=\"graf graf--p graf-after--p\">Stock trading is then the process of the cash that is paid for the stocks is converted into a share in the ownership of a company, which can be converted back to cash by selling, and this all hopefully with a profit. Now, to achieve a profitable return, you either go long or short in markets: you either buy shares thinking that the stock price will go up to sell at a higher price in the future, or you sell your stock, expecting that you can buy it back at a lower price and realize a profit. When you follow a fixed plan to go long or short in markets, you have a trading strategy.<\/p><p id=\"fd39\" class=\"graf graf--p graf-after--p\">Developing a trading strategy is something that goes through a couple of phases, just like when you, for example, build machine learning models: you formulate a strategy and specify it in a form that you can test on your computer, you do some preliminary testing or backtesting, you optimize your strategy and lastly, you evaluate the performance and robustness of your strategy.<\/p><p id=\"adeb\" class=\"graf graf--p graf-after--p\">Trading strategies are usually verified by backtesting: you reconstruct, with historical data, trades that would have occurred in the past using the rules that are defined with the strategy that you have developed. This way, you can get an idea of the effectiveness of your strategy and you can use it as a starting point to optimize and improve your strategy before applying it to real markets. Of course, this all relies heavily on the underlying theory or belief that any strategy that has worked out well in the past will likely also work out well in the future, and, that any strategy that has performed poorly in the past will likely also do badly in the future.<\/p><p id=\"da51\" class=\"graf graf--p graf-after--p\">A time series is a sequence of numerical data points taken at successive equally spaced points in time. In investing, a time series tracks the movement of the chosen data points, such as the stock price, over a specified period of time with data points recorded at regular intervals. If you\u2019re still in doubt about what this would exactly look like, take a look at the following example:<\/p><figure id=\"1526\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*dT2NavGUyZY2WlOQ.png\" data-width=\"550\" data-height=\"366\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*dT2NavGUyZY2WlOQ.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"04f2\" class=\"graf graf--p graf-after--figure\">You see that the dates are placed on the x-axis, while the price is featured on the y-axis. The \u201csuccessive equally spaced points in time\u201d in this case means that the days that are featured on the x-axis are 14 days apart: note the difference between 3\/7\/2005 and the next point, 3\/31\/2005, and 4\/5\/2005 and 4\/19\/2005.<\/p><p id=\"bc37\" class=\"graf graf--p graf-after--p\">However, what you\u2019ll often see when you\u2019re working with stock data is not just two columns, that contain period and price observations, but most of the times, you\u2019ll have five columns that contain observations of the period and the opening, high, low and closing prices of that period. This means that, if your period is set at a daily level, the observations for that day will give you an idea of the opening and closing price for that day and the extreme high and low price movement for a particular stock during that day.<\/p><p id=\"e1fc\" class=\"graf graf--p graf-after--p\">For now, you have a basic idea of the basic concepts that you need to know to go through this tutorial. These concepts will come back soon enough and you\u2019ll learn more about them later on in this tutorial.<\/p><p id=\"bc6f\" class=\"graf graf--p graf-after--p\">Getting your workspace ready to go is an easy job: you basically just make sure you have Python and an Integrated Development Environment (IDE) running on your system. However, there are some ways in which you can get started that are maybe a little easier when you\u2019re just starting out.<\/p><p id=\"8d55\" class=\"graf graf--p graf-after--p\">Take for instance <a href=\"https:\/\/anaconda.org\/\" data-href=\"https:\/\/anaconda.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Anaconda<\/a>, a high performance distribution of Python and R and includes over 100 of the most popular Python, R and Scala packages for data science. Additionally, installing Anaconda will give you access to over 720 packages that can easily be installed with conda, our renowned package, dependency and environment manager, that is included in Anaconda. And, besides all that, you\u2019ll get the Jupyter Notebook and Spyder IDE with it.<\/p><p id=\"aacd\" class=\"graf graf--p graf-after--p\">That sounds like a good deal, right?<\/p><p id=\"b74b\" class=\"graf graf--p graf-after--p\">You can install Anaconda from <a href=\"https:\/\/www.continuum.io\/downloads\" data-href=\"https:\/\/www.continuum.io\/downloads\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a> and don\u2019t forget to check out how to set up your Jupyter Notebook in DataCamp\u2019s <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/tutorial-jupyter-notebook\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/tutorial-jupyter-notebook\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Jupyter Notebook Tutorial: The Definitive Guide<\/a>.<\/p><p id=\"a7e1\" class=\"graf graf--p graf-after--p\">Of course, Anaconda is not your only option: you can also check out the <a href=\"https:\/\/www.enthought.com\/products\/canopy\/\" data-href=\"https:\/\/www.enthought.com\/products\/canopy\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Canopy Python distribution<\/a> (which doesn\u2019t come free), or try out the <a href=\"http:\/\/quant-platform.com\/\" data-href=\"http:\/\/quant-platform.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Quant Platform<\/a>.<\/p><p id=\"0c11\" class=\"graf graf--p graf-after--p\">The latter offers you a couple additional advantages over using, for example, Jupyter or the Spyder IDE, since it provides you everything you need specifically to do financial analytics in your browser! With the Quant Platform, you\u2019ll gain access to GUI-based Financial Engineering, interactive and Python-based financial analytics and your own Python-based analytics library. What\u2019s more, you\u2019ll also have access to a forum where you can discuss solutions or questions with peers!<\/p><figure id=\"174f\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*-2aGWFn6d8FXWDB7.png\" data-width=\"800\" data-height=\"246\" data-action=\"zoom\" data-action-value=\"0*-2aGWFn6d8FXWDB7.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*-2aGWFn6d8FXWDB7.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"fc9a\" class=\"graf graf--p graf-after--figure\">When you\u2019re using Python for finance, you\u2019ll often find yourself using the data manipulation package, Pandas. But also other packages such as NumPy, SciPy, Matplotlib,\u2026 will pass by once you start digging deeper.<\/p><p id=\"3b44\" class=\"graf graf--p graf-after--p\">For now, let\u2019s just focus on Pandas and using it to analyze time series data. This section will explain how you can import data, explore and manipulate it with Pandas. On top of all of that, you\u2019ll learn how you can perform common financial analyses on the data that you imported.<\/p><p id=\"f3bf\" class=\"graf graf--p graf-after--p\">The <code class=\"markup--code markup--p-code\">pandas-datareader<\/code> package allows for reading in data from sources such as Google, Yahoo! Finance, World Bank,\u2026 If you want to have an updated list of the data sources that are made available with this function, go to the <a href=\"https:\/\/pandas-datareader.readthedocs.io\/en\/latest\/remote_data.html\" data-href=\"https:\/\/pandas-datareader.readthedocs.io\/en\/latest\/remote_data.html\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">documentation<\/a>. For this tutorial, you will use the package to read in data from Yahoo! Finance.<\/p><pre id=\"d450\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">import pandas_datareader <br \/>as pdr import datetime <br \/>aapl = pdr.get_data_yahoo('AAPL', start=datetime.datetime(2006, 10, 1), end=datetime.datetime(2012, 1, 1))<\/code><\/pre><p id=\"8039\" class=\"graf graf--p graf-after--pre\"><strong class=\"markup--strong markup--p-strong\">Note<\/strong> that the Yahoo API endpoint has recently changed and that, if you want to already start working with the library on your own, you\u2019ll need to install a temporary fix until the patch has been merged into the master brach to start pulling in data from Yahoo! Finance with <code class=\"markup--code markup--p-code\">pandas-datareader<\/code>. Make sure to read up on the issue <a href=\"https:\/\/github.com\/pydata\/pandas-datareader\/issues\/315\" data-href=\"https:\/\/github.com\/pydata\/pandas-datareader\/issues\/315\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a> before you start on your own!<\/p><p id=\"2235\" class=\"graf graf--p graf-after--p\">No worries, though, for this tutorial, the data has been loaded in for you so that you don\u2019t face any issues while learning about finance in Python with Pandas.<\/p><p id=\"a86b\" class=\"graf graf--p graf-after--p\">It\u2019s wise to consider though that, even though <code class=\"markup--code markup--p-code\">pandas-datareader<\/code> offers a lot of options to pull in data into Python, it isn\u2019t the only package that you can use to pull in financial data: you can also make use of libraries such as Quandl, for example, to get data from Google Finance:<\/p><pre id=\"4995\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">import quandl <br \/>aapl = quandl.get(\"WIKI\/AAPL\", start_date=\"2006-10-01\", end_date=\"2012-01-01\")<\/code><\/pre><p id=\"5fe4\" class=\"graf graf--p graf-after--pre\">For more information on how you can use Quandl to get financial data directly into Python, go to <a href=\"https:\/\/www.quandl.com\/tools\/python\" data-href=\"https:\/\/www.quandl.com\/tools\/python\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">this page<\/a>.<\/p><p id=\"b162\" class=\"graf graf--p graf-after--p\">Lastly, if you\u2019ve already been working in finance for a while, you\u2019ll probably know that you most often use Excel also to manipulate your data. In such cases, you should know that you can integrate Python with Excel.<\/p><p id=\"d3b2\" class=\"graf graf--p graf-after--p\">Check out DataCamp\u2019s <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/python-excel-tutorial\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/python-excel-tutorial\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Python Excel Tutorial: The Definitive Guide<\/a> for more information.<\/p><p id=\"e18a\" class=\"graf graf--p graf-after--p\">The first thing that you want to do when you finally have the data in your workspace is getting your hands dirty. However, now that you\u2019re working with time series data, this might not seem as straightforward, since your index now contains DateTime values.<\/p><p id=\"6354\" class=\"graf graf--p graf-after--p\">No worries, though! Let\u2019s start step-by-step and explore the data first with some functions that you\u2019ll might already know if you have some prior programming experience with R or if you\u2019ve already worked with Pandas.<\/p><p id=\"a1e7\" class=\"graf graf--p graf-after--p\">Either way, you\u2019ll see it\u2019s very easy!<\/p><p id=\"6e07\" class=\"graf graf--p graf-after--p\">As you saw in the code chunk above, you have used <code class=\"markup--code markup--p-code\">pandas_datareader<\/code> to import data into your workspace. The resulting object <code class=\"markup--code markup--p-code\">aapl<\/code> is a DataFrame, which is a 2-dimensional labeled data structure with columns of potentially different types. Now, one of the first things that you probably do when you have a regular DataFrame on your hands, is running the <code class=\"markup--code markup--p-code\">head()<\/code> and <code class=\"markup--code markup--p-code\">tail()<\/code> functions to take a peek at the first and the last rows of your DataFrame. Luckily, this doesn\u2019t change when you\u2019re working with time series data!<\/p><p id=\"b68c\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tip<\/strong>: also make sure to use the <code class=\"markup--code markup--p-code\">describe()<\/code> function to get some useful summary statistics about your data.<\/p><p id=\"6e86\" class=\"graf graf--p graf-after--p\">Find the accompanying exercise <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"ff23\" class=\"graf graf--p graf-after--p\">As you have seen in the introduction, this data clearly contains the four columns with the opening and closing price per day and the extreme high and low price movements for the Apple stock for each day. Additionally, you also get two extra columns: <code class=\"markup--code markup--p-code\">Volume<\/code> and <code class=\"markup--code markup--p-code\">Adj Close<\/code>.<\/p><p id=\"cc1a\" class=\"graf graf--p graf-after--p\">The former column is used to register the number of shares that got traded during a single day. The latter, on the other hand, is the adjusted closing price: it\u2019s the closing price of the day that has been slightly adjusted to include any actions that occurred at any time before the next day\u2019s open. You can use this column to examine historical returns or when you\u2019re performing a detailed analysis on historical returns.<\/p><p id=\"67d7\" class=\"graf graf--p graf-after--p\">Note how the index or row labels contain dates, and how your columns or column labels contain numerical values.<\/p><p id=\"54af\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tip<\/strong>: if you now would like to save this data to a csv file with the <code class=\"markup--code markup--p-code\">to_csv()<\/code> function from <code class=\"markup--code markup--p-code\">pandas<\/code> and that you can use the <code class=\"markup--code markup--p-code\">read_csv()<\/code> function to read the data back into Python. This is extremely handy in cases where, for example, the Yahoo API endpoint has changed and you don\u2019t have access to your data any longer\u00a0:)<\/p><pre id=\"d699\" class=\"graf graf--pre graf-after--p\"><code class=\"markup--code markup--pre-code\">import pandas as pd <br \/>aapl.to_csv('data\/aapl_ohlc.csv') <br \/>df = pd.read_csv('data\/aapl_ohlc.csv', header=0, index_col='Date', parse_dates=True)<\/code><\/pre><p id=\"b614\" class=\"graf graf--p graf-after--pre\">Now that you have briefly inspected the first lines of your data and have taken a look at some summary statistics, it\u2019s time to go a little bit deeper.<\/p><p id=\"6261\" class=\"graf graf--p graf-after--p\">One way to do this is by inspecting the index and the columns and by selecting, for example, the last ten rows of a certain column. The latter is called subsetting because you take a small subset of your data. The result of the subsetting is a Series, which is a one-dimensional labeled array that is capable of holding any type.<\/p><p id=\"1304\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Remember<\/strong> that the DataFrame structure was a two-dimensional labeled array with columns that potentially hold different types of data.<\/p><p id=\"b335\" class=\"graf graf--p graf-after--p\">Check all of this out in the exercise below. First, use the <code class=\"markup--code markup--p-code\">index<\/code> and <code class=\"markup--code markup--p-code\">columns<\/code> attributes to take a look at the index and columns of your data. Next, subset the <code class=\"markup--code markup--p-code\">Close<\/code> column by only selecting the last 10 observations of the DataFrame. Make use of the square brackets <code class=\"markup--code markup--p-code\">[]<\/code> to isolate the last ten values. You might already know this way of subsetting from other programming languages, such as R. To conclude, assign the latter to a variable <code class=\"markup--code markup--p-code\">ts<\/code> and then check what type <code class=\"markup--code markup--p-code\">ts<\/code> is by using the <code class=\"markup--code markup--p-code\">type()<\/code> function. You can find the exercise <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"5e59\" class=\"graf graf--p graf-after--p\">The square brackets can be nice to subset your data, but they are maybe not the most idiomatic way to do things with Pandas. That\u2019s why you should also take a look at the <code class=\"markup--code markup--p-code\">loc()<\/code> and <code class=\"markup--code markup--p-code\">iloc()<\/code> functions: you use the former for label-based indexing and the latter for positional indexing.<\/p><p id=\"7996\" class=\"graf graf--p graf-after--p\">In practice, this means that you can pass the label of the row labels, such as <code class=\"markup--code markup--p-code\">2007<\/code> and <code class=\"markup--code markup--p-code\">2006-11-01<\/code>, to the <code class=\"markup--code markup--p-code\">loc()<\/code> function, while you pass integers such as <code class=\"markup--code markup--p-code\">22<\/code> and <code class=\"markup--code markup--p-code\">43<\/code> to the <code class=\"markup--code markup--p-code\">iloc()<\/code> function.<\/p><p id=\"4ffe\" class=\"graf graf--p graf-after--p\">Complete the exercise <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">in the original article<\/a> to understand how both <code class=\"markup--code markup--p-code\">loc()<\/code> and <code class=\"markup--code markup--p-code\">iloc()<\/code> work.<\/p><p id=\"2f26\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tip<\/strong>: if you look closely at the results of the subsetting, you\u2019ll notice that there are certain days missing in the data; If you look more closely at the pattern, you\u2019ll see that it\u2019s usually two or three days that are missing; These days are usually weekend days or public holidays and aren\u2019t part of your data. This is nothing to worry about: it\u2019s completely normal and you don\u2019t have to fill in these missing days.<\/p><p id=\"7125\" class=\"graf graf--p graf-after--p\">Besides indexing, you might also want to explore some other techniques to get to know your data a little bit better. You never know what else will show up. Let\u2019s try to sample some 20 rows from the data set and then let\u2019s resample the data so that <code class=\"markup--code markup--p-code\">aapl<\/code> is now at the monthly level instead of daily. You can make use of the <code class=\"markup--code markup--p-code\">sample()<\/code> and <code class=\"markup--code markup--p-code\">resample()<\/code> functions to do this.<\/p><p id=\"2aa8\" class=\"graf graf--p graf-after--p\">Very straightforward, isn\u2019t it?<\/p><p id=\"ebdd\" class=\"graf graf--p graf-after--p\">The <code class=\"markup--code markup--p-code\">resample()<\/code> function is often used because it provides elaborate control and more flexibility on the frequency conversion of your times series: besides specifying new time intervals yourself and specifying how you want to handle missing data, you also have the option to indicate how you want to resample your data, as you can see in the code example above. This stands in clear contract to the <code class=\"markup--code markup--p-code\">asfreq()<\/code> method, where you only have the first two options.<\/p><p id=\"c911\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tip<\/strong>: try this out for yourself in the IPython console of the above DataCamp Light chunk. Pass in <code class=\"markup--code markup--p-code\">aapl.asfreq(\"M\", method=\"bfill\")<\/code> to see what happens!<\/p><p id=\"b538\" class=\"graf graf--p graf-after--p\">Lastly, before you take your data exploration to the next level and start with visualizing your data and performing some common financial analyses on your data, you might already start to calculate the differences between the opening and closing prices per day. You can easily perform this arithmetic operation with the help of Pandas; Just substract the values in the <code class=\"markup--code markup--p-code\">Open<\/code> column of your <code class=\"markup--code markup--p-code\">aapl<\/code> data from the values of the <code class=\"markup--code markup--p-code\">Close<\/code> column of that same data. Or, in other words, subtract <code class=\"markup--code markup--p-code\">aapl.Close<\/code> from <code class=\"markup--code markup--p-code\">aapl.Open<\/code>. You storethe result in a new column of the <code class=\"markup--code markup--p-code\">aapl<\/code> DataFrame called <code class=\"markup--code markup--p-code\">diff<\/code> and then you delete it again with the help of <code class=\"markup--code markup--p-code\">del.<\/code><\/p><p id=\"3d15\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tip<\/strong>: make sure to comment out the last line of code so that the new column of your <code class=\"markup--code markup--p-code\">aapl<\/code> DataFrame doesn\u2019t get removed and you can check the results of your arithmetic operation!<\/p><p id=\"6e98\" class=\"graf graf--p graf-after--p\">Of course, knowing the gains in absolute terms might already help you to get an idea of whether you\u2019re making a good investment, but as a quant, you might be more interested in a more relative means of measuring your stock\u2019s value, like how much the value of a certain stock has gone up or gone down. A way to do this is by calculating the daily percentage change.<\/p><p id=\"7b20\" class=\"graf graf--p graf-after--p\">This is good to know for now, but don\u2019t worry about it just yet; You\u2019ll go deeper into this in a bit!<\/p><p id=\"60aa\" class=\"graf graf--p graf-after--p\">This section introduced you to some ways to first explore your data before you start performing some prior analyses. However, you can still go a lot further in this; Consider taking our <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/exploratory-data-analysis-python\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/exploratory-data-analysis-python\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Python Exploratory Data Analysis<\/a> if you want to know more.<\/p><p id=\"1cff\" class=\"graf graf--p graf-after--p\">Next to exploring your data by means of <code class=\"markup--code markup--p-code\">head()<\/code>, <code class=\"markup--code markup--p-code\">tail()<\/code>, indexing,\u00a0\u2026 You might also want to visualize your time series data. Thanks to Pandas\u2019 plotting integration with Matplotlib, this task becomes easy; Just use the <code class=\"markup--code markup--p-code\">plot()<\/code> function and pass the relevant arguments to it. Additionally, you can also add the <code class=\"markup--code markup--p-code\">grid<\/code> argument to indicate that the plot should also have a grid in the background.<\/p><p id=\"2e48\" class=\"graf graf--p graf-after--p\">If you run the code in the <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">original article<\/a>, you\u2019ll come to the following plot:<\/p><figure id=\"50ff\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*K2LwTNoHAn1h4elx.png\" data-width=\"378\" data-height=\"242\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*K2LwTNoHAn1h4elx.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"f31e\" class=\"graf graf--p graf-after--figure\">If you want to know more about Matplotlib and how to get started with it, check out DataCamp\u2019s <a href=\"https:\/\/www.datacamp.com\/courses\/intermediate-python-for-data-science\" data-href=\"https:\/\/www.datacamp.com\/courses\/intermediate-python-for-data-science\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Intermediate Python for Data Science<\/a> course.<\/p><h3 id=\"382f\" class=\"graf graf--h3 graf-after--p\">Common Financial Analysis<\/h3><p id=\"1df8\" class=\"graf graf--p graf-after--h3\">Now that you have an idea of your data, what time series data is about and how you can use <code class=\"markup--code markup--p-code\">pandas<\/code> to quickly explore your data, it\u2019s time to dive deeper into some of the common financial analyses that you can do so that you can actually start working towards developing a trading strategy.<\/p><p id=\"aff4\" class=\"graf graf--p graf-after--p\">In the rest of this section, you\u2019ll learn more about the returns, moving windows, volatility calculation and Ordinary Least-Squares Regression (OLS).<\/p><p id=\"e256\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">You can read more and practice these common financial analyses in the original article.<\/strong><\/p><h3 id=\"ab7d\" class=\"graf graf--h3 graf-after--p\"><strong class=\"markup--strong markup--h3-strong\">Creating A Trading\u00a0Strategy<\/strong><\/h3><p id=\"fcad\" class=\"graf graf--p graf-after--h3\">Now that you have done some primary analyses to your data, it\u2019s time to formulate your first trading strategy; But before you go into all of this, why not first get to know some of the most common trading strategies? After a short introduction, you\u2019ll undoubtedly move on more easily your trading strategy.<\/p><p id=\"0910\" class=\"graf graf--p graf-after--p\">From the introduction, you\u2019ll still remember that a trading strategy is a fixed plan to go long or short in markets, but much more information you didn\u2019t really get yet; In general, there are two common trading strategies: the momentum strategy and the reversion strategy.<\/p><p id=\"3e33\" class=\"graf graf--p graf-after--p\">Firstly, the <strong class=\"markup--strong markup--p-strong\">momentum strategy<\/strong> is also called divergence or trend trading. When you follow this strategy, you do so because you believe the movement of a quantity will continue in its current direction. Stated differently, you believe that stocks have momentum, or upward or downward trends, that you can detect and exploit.<\/p><p id=\"2e8f\" class=\"graf graf--p graf-after--p\">Some examples of this strategy are the moving average crossover, the dual moving average crossover and turtle trading:<\/p><ul class=\"postList\"><li id=\"7e7f\" class=\"graf graf--li graf-after--p\">The moving average crossover is when the price of an asset moves from one side of a moving average to the other. This crossover represents a change in momentum and can be used as a point of making the decision to enter or exit the market. You\u2019ll see an example of this strategy, which is the \u201chello world\u201d of quantitative trading later on in this tutorial.<\/li><li id=\"3027\" class=\"graf graf--li graf-after--li\">The dual moving average crossover occurs when a short-term average crosses a long-term average. This signal is used to identify that momentum is shifting in the direction of the short-term average. A buy signal is generated when the short-term average crosses the long-term average and rises above it, while a sell signal is triggered by a short-term average crossing long-term average and falling below it.<\/li><li id=\"7f3c\" class=\"graf graf--li graf-after--li\">Turtle trading is a well-known trend following strategy that was originally taught by Richard Dennis. The basic strategy is to buy futures on a 20-day high and sell on a 20-day low.<\/li><\/ul><p id=\"d891\" class=\"graf graf--p graf-after--li\">Secondly, the <strong class=\"markup--strong markup--p-strong\">reversion strategy<\/strong>, which is also known as convergence or cycle trading. This strategy departs from the belief that the movement of a quantity will eventually reverse. This might seem a little bit abstract, but will not be so any more when you take the example. Take a look at the mean reversion strategy, where you actually believe that stocks return to their mean and that you can exploit when it deviates from that mean.<\/p><p id=\"d648\" class=\"graf graf--p graf-after--p\">That already sounds a whole lot more practical, right?<\/p><p id=\"37d4\" class=\"graf graf--p graf-after--p\">Another example of this strategy, besides the mean reversion strategy, is the pairs trading mean-reversion, which is similar to the mean reversion strategy. Whereas the mean reversion strategy basically stated that stocks return to their mean, the pairs trading strategy extends this and states that if two stocks can be identified that have a relatively high correlation, the change in the difference in price between the two stocks can be used to signal trading events if one of the two moves out of correlation with the other. That means that if the correlation between two stocks has decreased, the stock with the higher price can be considered to be in a short position. It should be sold because the higher-priced stock will return to the mean. The lower-priced stock, on the other hand, will be in a long position becaue the price will rise as the correlation will return to normal.<\/p><p id=\"1194\" class=\"graf graf--p graf-after--p\">Besides these two most frequent strategies, there are also other ones that you might come across once in a while, such as the forecasting strategy, which attempts to predict the direction or value of stock, in this case, in subsequent future time periods based on certain historical factors. There\u2019s also the High Frequency Trading (HFT) strategy, which exploits the sub-millisecond market microstructure.<\/p><p id=\"464d\" class=\"graf graf--p graf-after--p\">That\u2019s all music for the future for now; Let\u2019s focus on developing your first trading strategy for now!<\/p><p id=\"42ac\" class=\"graf graf--p graf-after--p\">As you read above, you\u2019ll start with the \u201chello world\u201d of quantitative trading: the moving average crossover. The strategy that you\u2019ll be developing is simple: you create two separate Simple Moving Averages (SMA) of a time series with differing lookback periods, let\u2019s say, 40 days and 100 days. If the short moving average exceeds the long moving average then you go long, if the long moving average exceeds the short moving average then you exit.<\/p><p id=\"4816\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Remember<\/strong> that when you go long, you think that the stock price will go up and will sell at a higher price in the future (= buy signal); When you go short, you sell your stock, expecting that you can buy it back at a lower price and realize a profit (= sell signal).<\/p><p id=\"7166\" class=\"graf graf--p graf-after--p\">This simple strategy might seem quite complex when you\u2019re just starting out, but let\u2019s take this step by step:<\/p><ul class=\"postList\"><li id=\"3f38\" class=\"graf graf--li graf-after--p\">First define your two different lookback periods: a short window and a long window. You set up two variables and assign one integer per variable. Make sure that the integer that you assign to the short window is shorter than the integer that you assign to the long window variable!<\/li><li id=\"d928\" class=\"graf graf--li graf-after--li\">Next, make an empty <code class=\"markup--code markup--li-code\">signals<\/code> DataFrame, but do make sure to copy the index of your <code class=\"markup--code markup--li-code\">aapl<\/code> data so that you can start calculating the daily buy or sell signal for your <code class=\"markup--code markup--li-code\">aapl<\/code> data.<\/li><li id=\"d3a4\" class=\"graf graf--li graf-after--li\">Create a column in your empty <code class=\"markup--code markup--li-code\">signals<\/code> DataFrame that is named <code class=\"markup--code markup--li-code\">signal<\/code> and initialize it by setting the value for all rows in this column to <code class=\"markup--code markup--li-code\">0.0<\/code>.<\/li><li id=\"c42b\" class=\"graf graf--li graf-after--li\">After the preparatory work, it\u2019s time to create the set of short and long simple moving averages over the respective long and short time windows. Make use of the <code class=\"markup--code markup--li-code\">rolling()<\/code> function to start your rolling window calculcations: within the function, specify the <code class=\"markup--code markup--li-code\">window<\/code> and the <code class=\"markup--code markup--li-code\">min_period<\/code>, and set the <code class=\"markup--code markup--li-code\">center<\/code> argument. In practice, this will result in a <code class=\"markup--code markup--li-code\">rolling()<\/code> function to which you have passed either <code class=\"markup--code markup--li-code\">short_window<\/code> or <code class=\"markup--code markup--li-code\">long_window<\/code>, <code class=\"markup--code markup--li-code\">1<\/code> as the minimum number of observations in the window that are required to have a value, and <code class=\"markup--code markup--li-code\">False<\/code>, so that the labels are not set at the center of the window. Next, don\u2019t forget to also chain the <code class=\"markup--code markup--li-code\">mean()<\/code> function so that you calculate the rolling mean.<\/li><li id=\"73a5\" class=\"graf graf--li graf-after--li\">After you have calculated the mean average of the short and long windows, you should create a signal when the short moving average crosses the long moving average, but only for the period greater than the shortest moving average window. In Python, this will result in a condition: <code class=\"markup--code markup--li-code\">signals['short_mavg'][short_window:] &gt; signals['long_mavg'][short_window:]<\/code>. Note that you add the <code class=\"markup--code markup--li-code\">[short_window:]<\/code> to comply with the condition \u201conly for the period greater than the shortest moving average window\u201d. When the condition is true, the initialized value <code class=\"markup--code markup--li-code\">0.0<\/code> in the <code class=\"markup--code markup--li-code\">signal<\/code> column will be overwitten with <code class=\"markup--code markup--li-code\">1.0<\/code>. A \u201csignal\u201d is created! If the condition is false, the original value of <code class=\"markup--code markup--li-code\">0.0<\/code> will be kept and no signal is generated. You use the NumPy <code class=\"markup--code markup--li-code\">where()<\/code> function to set up this condition. Much the same like you read just now, the variable to which you assign this result is <code class=\"markup--code markup--li-code\">signals['signal'][short_window]<\/code>, because you only want to create signals for the period greater than the shortest moving average window!<\/li><li id=\"cdbe\" class=\"graf graf--li graf-after--li\">Lastly, you take the difference of the signals in order to generate actual trading orders. In other words, in this column of your <code class=\"markup--code markup--li-code\">signals<\/code> DataFrame, you\u2019ll be able to distinguish between long and short positions, whether you\u2019re buying or selling stock.<\/li><\/ul><p id=\"ef34\" class=\"graf graf--p graf-after--li\">See the code <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"3362\" class=\"graf graf--p graf-after--p\">This wasn\u2019t too hard, was it? Print out the <code class=\"markup--code markup--p-code\">signals<\/code> DataFrame and inspect the results. Important to grasp here is what the <code class=\"markup--code markup--p-code\">positions<\/code> and the <code class=\"markup--code markup--p-code\">signal<\/code> columns mean in this DataFrame. You\u2019ll see that it will become very important when you move on!<\/p><p id=\"19b5\" class=\"graf graf--p graf-after--p\">When you have taken the time to understand the results of your trading strategy, quickly plot all of this (the short and long moving averages, together with the buy and sell signals) with Matplotlib:<\/p><figure id=\"7217\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*XduFBw9e7vdRp-54.png\" data-width=\"392\" data-height=\"242\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*XduFBw9e7vdRp-54.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"8317\" class=\"graf graf--p graf-after--figure\">PS. You can find the code for this plot <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"f55d\" class=\"graf graf--p graf-after--p\">The result is pretty cool, isn\u2019t it?<\/p><h3 id=\"210f\" class=\"graf graf--h3 graf-after--p\">Backtesting The Trading\u00a0Strategy<\/h3><p id=\"8b83\" class=\"graf graf--p graf-after--h3\">Now that you\u2019ve got your trading strategy at hand, it\u2019s a good idea to also backtest it and calculate its performance. But right before you go deeper into this, you might want to know just a little bit more about the pitfalls of backtesting, what components are needed in a backtester and what Python tools you can use to backtest your simple algorithm.<\/p><p id=\"3e0b\" class=\"graf graf--p graf-after--p\">If, however, you\u2019re already well up to date, you can simply move on to the implementation of your backtester!<\/p><p id=\"a7ee\" class=\"graf graf--p graf-after--p\">Backtesting is, besides just \u201ctesting a trading strategy\u201d, testing the strategy on relevant historical data to make sure that it\u2019s an actual viable strategy before you start making moves. With backtesting, a trader can simulate and analyze the risk and profitability of trading with a specific strategy over a period of time. However, when you\u2019re backtesting, it\u2019s a good idea to keep in mind that there are some pitfalls, which might not be obvious to you when you\u2019re just starting out.<\/p><p id=\"9820\" class=\"graf graf--p graf-after--p\">For example, there are external events, such as market regime shifts, which are regulatory changes or macroeconomic events, which definitely influence your backtesting. Also liquidity constraints, such as the ban of short sales, could affect your backtesting heavily.<\/p><p id=\"0f47\" class=\"graf graf--p graf-after--p\">Next, there are pitfalls which you might introduce yourself when you, for example, overfit a model (optimization bias), when you ignore strategy rules because you think it\u2019s better like that (interference), or when you accidentally introduce information into past data (lookahead bias).<\/p><p id=\"138a\" class=\"graf graf--p graf-after--p\">These are just a few pitfalls that you need to take into account mainly after this tutorial, when you go and make your own strategies and backtest them.<\/p><p id=\"f3fc\" class=\"graf graf--p graf-after--p\">Besides the pitfalls, it\u2019s good to know that your backtester usually consists of some four essential components, which should usually present in every backtester. As such, a backtester consists of the following:<\/p><ul class=\"postList\"><li id=\"6c78\" class=\"graf graf--li graf-after--p\">A data handler, which is an interface to a set of data,<\/li><li id=\"7782\" class=\"graf graf--li graf-after--li\">A strategy, which generates a signal to go long or go short based on the data,<\/li><li id=\"76dd\" class=\"graf graf--li graf-after--li\">A portfolio, which generates orders and manages Profit &amp; Loss (also known as \u201cPnL\u201d), and<\/li><li id=\"997d\" class=\"graf graf--li graf-after--li\">An execution handler, which sends the order to the broker and receives the \u201cfills\u201d or signals that the stock has been bought or sold.<\/li><\/ul><p id=\"27cf\" class=\"graf graf--p graf-after--li\">Besides these four components, there are many more that you can add to your backtester, depending on the complexity. You can definitely go a lot further than just these four components. However, for this beginner tutorial, you\u2019ll just focus on getting these basic components to work in your code.<\/p><p id=\"8b19\" class=\"graf graf--p graf-after--p\">As you read above, a simple backtester consists of a strategy, a data handler, a portfolio and an execution handler. You have already implemented a strategy above, and you also have access to a data handler, which is the <code class=\"markup--code markup--p-code\">pandas-datareader<\/code> or the Pandas library that you use to get your saved data from Excel into Python. The components that are still left to implement are the execution handler and the portfolio.<\/p><p id=\"d4e0\" class=\"graf graf--p graf-after--p\">However, since you\u2019re just starting out, you\u2019ll not focus on implementing an execution handler just yet. Instead, you\u2019ll see below how you can get started on creating a portfolio which can generate orders and manages the profit and loss:<\/p><ul class=\"postList\"><li id=\"8178\" class=\"graf graf--li graf-after--p\">First off, you\u2019ll create set a variable <code class=\"markup--code markup--li-code\">initial_capital<\/code> to set your initial capital and a new DataFrame <code class=\"markup--code markup--li-code\">positions<\/code>. Once again, you copy the index from another DataFrame; In this case, this is the <code class=\"markup--code markup--li-code\">signals<\/code> DataFrame because you want to consider the time frame for which you have generated the signals.<\/li><li id=\"6ff6\" class=\"graf graf--li graf-after--li\">Next, you create a new column <code class=\"markup--code markup--li-code\">AAPL<\/code> in the DataFrame. On the days that the signal is 1 and the the short moving average crosses the long moving average (for the period greater than the shortest moving average window), you\u2019ll buy a 100 shares. The days on which the signal is 0, the final result will be 0 as a result of the operation <code class=\"markup--code markup--li-code\">100*signals['signal']<\/code>.<\/li><li id=\"edb4\" class=\"graf graf--li graf-after--li\">A new DataFrame <code class=\"markup--code markup--li-code\">portfolio<\/code> is created to store the market value of an open position.<\/li><li id=\"c35b\" class=\"graf graf--li graf-after--li\">Next, you create a DataFrame that stores the differences in positions (or number of stock)<\/li><li id=\"7b9f\" class=\"graf graf--li graf-after--li\">Then the real backtesting begins: you create a new column to the <code class=\"markup--code markup--li-code\">portfolio<\/code> DataFrame with name <code class=\"markup--code markup--li-code\">holdings<\/code>, which stores the value of the positions or shares you have bought, multiplied by the \u2018Adj Close\u2019 price.<\/li><li id=\"1297\" class=\"graf graf--li graf-after--li\">Your <code class=\"markup--code markup--li-code\">portfolio<\/code> also contains a <code class=\"markup--code markup--li-code\">cash<\/code> column, which is the capital that you still have left to spend: it is calculated by taking your <code class=\"markup--code markup--li-code\">initial_capital<\/code> and subtracting your holdings (the price that you paid for buying stock).<\/li><li id=\"3ba6\" class=\"graf graf--li graf-after--li\">You\u2019ll also add a <code class=\"markup--code markup--li-code\">total<\/code> column to your <code class=\"markup--code markup--li-code\">portfolio<\/code> DataFrame, which contains the sum of your cash and the holdings that you own, and<\/li><li id=\"c61d\" class=\"graf graf--li graf-after--li\">Lastly, you also add a <code class=\"markup--code markup--li-code\">returns<\/code> column to your portfolio, in which you\u2019ll store the returns<\/li><\/ul><p id=\"ddeb\" class=\"graf graf--p graf-after--li\">Find the code <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"4663\" class=\"graf graf--p graf-after--p\">As a last exercise for your backtest, visualize the portfolio value or <code class=\"markup--code markup--p-code\">portfolio['total']<\/code> over the years with the help of Matplotlib and the results of your backtest:<\/p><figure id=\"3d29\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*grrZVJFepC8X3YYV.png\" data-width=\"411\" data-height=\"242\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*grrZVJFepC8X3YYV.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"8b81\" class=\"graf graf--p graf-after--figure\">Find the code <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"25e9\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Note<\/strong> that, for this tutorial, the Pandas code for the backtester as well as the trading strategy has been composed in such a way that you can easily walk through it in an interactive way. In a real-life application, you might opt for a more object-oriented design with classes, which contain all the logic. You can find an example of the same moving average crossover strategy, with object-oriented design, <a href=\"https:\/\/www.quantstart.com\/articles\/Backtesting-a-Moving-Average-Crossover-in-Python-with-pandas\" data-href=\"https:\/\/www.quantstart.com\/articles\/Backtesting-a-Moving-Average-Crossover-in-Python-with-pandas\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a> or check out <a href=\"https:\/\/s3.amazonaws.com\/quantstart\/media\/powerpoint\/an-introduction-to-backtesting.pdf\" data-href=\"https:\/\/s3.amazonaws.com\/quantstart\/media\/powerpoint\/an-introduction-to-backtesting.pdf\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">this presentation<\/a>.<\/p><p id=\"5170\" class=\"graf graf--p graf-after--p\">You have seen now how you can implement a backtester with the Python\u2019s popular data manipulation package Pandas. However, you can also see that it\u2019s easy to make mistakes and that this might not be the most fail-safe option to use every time: you need to build most of the components from scratch, even though you already leverage Pandas to get your results.<\/p><p id=\"1f10\" class=\"graf graf--p graf-after--p\">That\u2019s why it\u2019s common to use a backtesting platform, such as Quantopian, for your backtesters. Quantopian is a free, community-centered, hosted platform for building and executing trading strategies. It\u2019s powered by <code class=\"markup--code markup--p-code\">zipline<\/code>, a Python library for algorithmic trading. You can use the library locally, but for the purpose of this beginner tutorial, you\u2019ll use <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/www.quantopian.com\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/www.quantopian.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">Quantopian<\/strong><\/a> to write and backtest your algorithm. Before you can do this, though, make sure that you first sign up and log in.<\/p><p id=\"eac1\" class=\"graf graf--p graf-after--p\">Next, you can get started pretty easily. Click \u201cNew Algorithm\u201d to start writing up your trading algorithm or select one of the examples that has already been coded up for you to get a better feeling of what you\u2019re exactly dealing with\u00a0:)<\/p><figure id=\"d0e2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*pDkHhkEmhF3ETNTW.png\" data-width=\"722\" data-height=\"400\" data-action=\"zoom\" data-action-value=\"0*pDkHhkEmhF3ETNTW.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*pDkHhkEmhF3ETNTW.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"1ecd\" class=\"graf graf--p graf-after--figure\">Let\u2019s start simple and make a new algorithm, but still following our simple example of the moving average crossover, which is the standard example that you find in the <a href=\"http:\/\/www.zipline.io\/\" data-href=\"http:\/\/www.zipline.io\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">zipline Quickstart guide<\/a>. It so happens that this example is very similar to the simple trading strategy that you implemented in the previous section. You see, though, that the structure in the code chunk below and in the screenshot above is somewhat different than what you have seen up until now in this tutorial, namely, you have two definitions that you start working from, namely <code class=\"markup--code markup--p-code\">initialize()<\/code> and <code class=\"markup--code markup--p-code\">handle_data().<\/code><\/p><p id=\"791e\" class=\"graf graf--p graf-after--p\">Find the code <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"0081\" class=\"graf graf--p graf-after--p\">The first function is called when the program is started and performs one-time startup logic. As an argument, the <code class=\"markup--code markup--p-code\">initialize()<\/code> function takes a <code class=\"markup--code markup--p-code\">context<\/code>, which is used to store the state during a backtest or live trading and can be referenced in different parts of the algorithm, as you can see in the code below; You see that <code class=\"markup--code markup--p-code\">context<\/code> comes back, among others, in the definition of the first moving average window. You see that you assign the result of the lookup of a security (stock in this case) by its symbol, (<code class=\"markup--code markup--p-code\">AAPL<\/code> in this case) to <code class=\"markup--code markup--p-code\">context.security<\/code>.<\/p><p id=\"e88a\" class=\"graf graf--p graf-after--p\">The <code class=\"markup--code markup--p-code\">handle_data()<\/code> function is called once per minute during simulation or live-trading to decide what orders, if any, should be placed each minute. The function requires <code class=\"markup--code markup--p-code\">context<\/code> and <code class=\"markup--code markup--p-code\">data<\/code> as input: the <code class=\"markup--code markup--p-code\">context<\/code> is the same as the one that you read about just now, while the data is an object that stores several API functions, such as <code class=\"markup--code markup--p-code\">current()<\/code> to retrieve the most recent value of a given field(s) for a given asset(s) or <code class=\"markup--code markup--p-code\">history()<\/code> to get trailing windows of historical pricing or volume data. These API functions don\u2019t come back in the code below and are not in the scope of this tutorial.<\/p><p id=\"4e03\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Note<\/strong> That the code that you type into the Quantopian console will only work on the platform itself and not in your local Jupyter Notebook, for example!<\/p><p id=\"b02e\" class=\"graf graf--p graf-after--p\">You\u2019ll see that the <code class=\"markup--code markup--p-code\">data<\/code> object allows you to retrieve the <code class=\"markup--code markup--p-code\">price<\/code>, which is the forward-filled, returning last known price, if there is one. If there is none, an <code class=\"markup--code markup--p-code\">NaN<\/code> value will be returned.<\/p><p id=\"9426\" class=\"graf graf--p graf-after--p\">Another object that you see in the code chunk above is the <code class=\"markup--code markup--p-code\">portfolio<\/code>, which stores important information about\u2026. Your portfolio. As you can see in the piece of code <code class=\"markup--code markup--p-code\">context.portfolio.positions<\/code>, this object is stored in the context and is then also accessible in the core functions that <code class=\"markup--code markup--p-code\">context<\/code> has to offer to you as a user. Note that the <code class=\"markup--code markup--p-code\">positions<\/code> that you just read about, store Position objects and include information such as the number of shares and price paid as values. Additionally, you also see that the <code class=\"markup--code markup--p-code\">portfolio<\/code> also has a <code class=\"markup--code markup--p-code\">cash<\/code> property to retrieve the current amount of cash in your portfolio and that the <code class=\"markup--code markup--p-code\">positions<\/code> object also has an <code class=\"markup--code markup--p-code\">amount<\/code> property to explore the whole number of shares in a certain position.<\/p><p id=\"b68e\" class=\"graf graf--p graf-after--p\">The <code class=\"markup--code markup--p-code\">order_target()<\/code> places an order to adjust a position to a target number of shares. If there is no existing position in the asset, an order is placed for the full target number. If there is a position in the asset, an order is placed for the difference between the target number of shares or contracts and the number currently held. Placing a negative target order will result in a short position equal to the negative number specified.<\/p><p id=\"39b3\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Tip<\/strong>: if you have any more questions about the functions or objects, make sure to check the Quantopian <a href=\"https:\/\/www.quantopian.com\/help\" data-href=\"https:\/\/www.quantopian.com\/help\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Help page<\/a>, which contains more information about all (and much more) that you have briefly seen in this tutorial.<\/p><p id=\"09cb\" class=\"graf graf--p graf-after--p\">When you have created your strategy with the <code class=\"markup--code markup--p-code\">initialize()<\/code> and <code class=\"markup--code markup--p-code\">handle_data()<\/code> functions (or copy-pasted the above code) into the console on the left-hand side of your interface, just press the \u201cBuild Algorithm\u201d button to build the code and run a backtest. If you press the \u201cRun Full Backtest\u201d button, a full backtest is run, which is basically the same as the one that you run when you build the algorithm, but you\u2019ll be able to see a lot more in detail. The backtesting, whether \u2018simple\u2019 or full, can take a while; Make sure to keep an eye out on the progress bar on top of the page!<\/p><figure id=\"5534\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*UdA2vAmC5BdqoraW.png\" data-width=\"727\" data-height=\"400\" data-action=\"zoom\" data-action-value=\"0*UdA2vAmC5BdqoraW.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*UdA2vAmC5BdqoraW.png\" alt=\"image\" \/><\/div><\/div><\/figure><p id=\"8f35\" class=\"graf graf--p graf-after--figure\">You can find more information on how to get started with Quantopian <a href=\"https:\/\/www.quantopian.com\/tutorials\/getting-started\" data-href=\"https:\/\/www.quantopian.com\/tutorials\/getting-started\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"b137\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Note<\/strong> that Quantopian is an easy way to get started with zipline, but that you can always move on to using the library locally in, for example, your Jupyter notebook.<\/p><h3 id=\"4148\" class=\"graf graf--h3 graf-after--p\">Improving The Trading\u00a0Strategy<\/h3><p id=\"97f2\" class=\"graf graf--p graf-after--h3\">You have successfully made a simple trading algorithm and performed backtests via Pandas, Zipline and Quantopian. It\u2019s fair to say that you\u2019ve been introduced to trading with Python. However, when you have coded up the trading strategy and backtested it, your work doesn\u2019t stop yet; You might want to improve your strategy. There are one or more algorithms may be used to improve the model on a continuous basis, such as KMeans, k-Nearest Neighbors (KNN), Classification or Regression Trees and the Genetic Algorithm. This will be the topic of a future DataCamp tutorial.<\/p><p id=\"16ac\" class=\"graf graf--p graf-after--p\">Apart from the other algorithms you can use, you saw that you can improve your strategy by working with multi-symbol portfolios. Just incorporating one company or symbol into your strategy often doesn\u2019t really say much. You\u2019ll also see this coming back in the evaluation of your moving average crossover strategy. Other things that you can add or do differently is using a risk management framework or use event-driven backtesting to help mitigate the lookahead bias that you read about earlier. There are still many other ways in which you could improve your strategy, but for now, this is a good basis to start from!<\/p><figure id=\"8909\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"0*Zj4UJG_vuTusun4q.png\" data-width=\"800\" data-height=\"246\" data-action=\"zoom\" data-action-value=\"0*Zj4UJG_vuTusun4q.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*Zj4UJG_vuTusun4q.png\" alt=\"image\" \/><\/div><\/div><\/figure><h3 id=\"024d\" class=\"graf graf--h3 graf-after--figure\">Evaluating Moving Average Crossover Strategy<\/h3><p id=\"1c8d\" class=\"graf graf--p graf-after--h3\">Improving your strategy doesn\u2019t mean that you\u2019re finished just yet! You can easily use Pandas to calculate some metrics to further judge your simple trading strategy. In this section, you\u2019ll learn about the Sharpe ratio, the maximum drawdown and the Compound Annual Growth Rate (CAGR).<\/p><p id=\"c284\" class=\"graf graf--p graf-after--p\">Read and practice more <a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><p id=\"bd69\" class=\"graf graf--p graf-after--p\">Besides these two metrics, there are also many other that you could consider, such as the <strong class=\"markup--strong markup--p-strong\">distribution of returns<\/strong>, <strong class=\"markup--strong markup--p-strong\">trade-level metrics<\/strong>,\u00a0\u2026<\/p><h3 id=\"a7b2\" class=\"graf graf--h3 graf-after--p\">Go Further!<\/h3><p id=\"6550\" class=\"graf graf--p graf-after--h3\">Well done, you\u2019ve made it through this Python Finance introduction tutorial! You\u2019ve covered a lot of ground, but there\u2019s still so much more for you to discover!<\/p><p id=\"9ea0\" class=\"graf graf--p graf-after--p\">Check out Yves Hilpisch\u2019s Python For Finance book, which is a great book for those who already have gathered some background into Finance, but not so much in Python. \u201cMastering Pandas for Data Science\u201d by Michael Heydt is also recommended for those who want to get started with Finance in Python! Also make sure to check out <a href=\"https:\/\/www.quantstart.com\/articles\" data-href=\"https:\/\/www.quantstart.com\/articles\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Quantstart\u2019s articles<\/a> for guided tutorials on algorithmic trading and <a href=\"https:\/\/pythonprogramming.net\/python-programming-creating-automated-trading-strategy\/\" data-href=\"https:\/\/pythonprogramming.net\/python-programming-creating-automated-trading-strategy\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">this complete series on Python programming for finance<\/a>.<\/p><p id=\"799c\" class=\"graf graf--p graf-after--p graf--trailing\">If you\u2019re more interested in continuing your journey into finance with R, consider taking Datacamp\u2019s <a href=\"https:\/\/www.datacamp.com\/tracks\/quantitative-analyst-with-r\" data-href=\"https:\/\/www.datacamp.com\/tracks\/quantitative-analyst-with-r\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Quantitative Analyst with R<\/a> track. And in the meantime, keep posted for our second post on starting finance with Python and check out the <a href=\"https:\/\/github.com\/Kacawi\/datacamp-community\" data-href=\"https:\/\/github.com\/Kacawi\/datacamp-community\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Jupyter notebook of this tutorial<\/a>.<\/p><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"3b90\" class=\"graf graf--p graf--leading graf--trailing\"><em class=\"markup--em markup--p-em\">Originally published at <\/em><a href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" data-href=\"https:\/\/www.datacamp.com\/community\/tutorials\/finance-python-trading\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">www.datacamp.com<\/em><\/a><em class=\"markup--em markup--p-em\">.<\/em><\/p><\/div><\/div><\/section>","created_at":"2018-11-07T16:51:58+0000","updated_at":"2018-11-07T16:51:58+0000","published_at":"2017-06-02T10:00:32+0000","published_by":["Karlijn Willems"],"annotations":[],"mimetype":"text\/html","reading_time":35,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/0*1Tz2TStuR1fbSCS-.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":560,"title":"japaric\/copper","url":"https:\/\/github.com\/japaric\/copper","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>This document has been <strong>DEPRECATED<\/strong> in favor of <a href=\"https:\/\/rust-embedded.github.io\/bookshelf\/book\/index.html\" rel=\"nofollow\">the embedded Rust\nbook<\/a>. Also check out the rest of the <a href=\"https:\/\/github.com\/rust-embedded\/wg\">embedded WG's<\/a> <a href=\"https:\/\/rust-embedded.github.io\/bookshelf\/\" rel=\"nofollow\">bookshelf<\/a>.<\/p>\n<p>-- @japaric, 2018\/09\/26<\/p>\n<p>This document is currently <strong>OUTDATED<\/strong>, and I don't have to update in the short\nterm. In the meantime, you can check out <a href=\"http:\/\/blog.japaric.io\/quickstart\" rel=\"nofollow\">this blog post<\/a> which covers the easy,\nhigh level way of writing Rust application for any ARM Cortex-M microcontroller.<\/p>\n<p>-- @japaric, 2017\/05\/15<\/p>\n<hr \/><p><a href=\"https:\/\/travis-ci.org\/japaric\/copper\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/9d1edc3c4cb30b1eb95bf582f4576a7e41782776\/68747470733a2f2f7472617669732d63692e6f72672f6a6170617269632f636f707065722e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/japaric\/copper.svg?branch=master\" \/><\/a><\/p>\n<blockquote>\n<p>Your guide to programming ARM Cortex-M microcontrollers with Rust<\/p>\n<\/blockquote>\n<p>This is the <a href=\"https:\/\/github.com\/azerupi\/mdBook\">mdbook<\/a> source of the book. You can find the rendered book at\n<a href=\"https:\/\/japaric.github.io\/copper\/\" rel=\"nofollow\">https:\/\/japaric.github.io\/copper\/<\/a><\/p>\n<p>Licensed under either of<\/p>\n<ul><li>\n<p>Apache License, Version 2.0 (<a href=\"https:\/\/github.com\/japaric\/copper\/blob\/master\/LICENSE-APACHE\">LICENSE-APACHE<\/a> or\n<a href=\"http:\/\/www.apache.org\/licenses\/LICENSE-2.0\" rel=\"nofollow\">http:\/\/www.apache.org\/licenses\/LICENSE-2.0<\/a>)<\/p>\n<\/li>\n<li>\n<p>MIT license (<a href=\"https:\/\/github.com\/japaric\/copper\/blob\/master\/LICENSE-MIT\">LICENSE-MIT<\/a> or <a href=\"http:\/\/opensource.org\/licenses\/MIT\" rel=\"nofollow\">http:\/\/opensource.org\/licenses\/MIT<\/a>)<\/p>\n<\/li>\n<\/ul><p>at your option.<\/p>\n<h3><a id=\"user-content-contribution\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribution\"><\/a>Contribution<\/h3>\n<p>Unless you explicitly state otherwise, any contribution intentionally submitted\nfor inclusion in the work by you, as defined in the Apache-2.0 license, shall be\ndual licensed as above, without any additional terms or conditions.<\/p>\n<\/article>","created_at":"2018-11-07T16:51:57+0000","updated_at":"2018-11-07T16:51:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/5018213?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":561,"title":"Git from the inside out","url":"https:\/\/maryrosecook.com\/blog\/post\/git-from-the-inside-out","content":"<p>This essay explains how Git works.  (If you\u2019d rather absorb the same information as a talk, you can watch this <a href=\"https:\/\/www.youtube.com\/watch?v=fCtZWGhQBvo\">video<\/a> instead.)<\/p><p>The essay assumes you understand Git well enough to use it to version control your projects.  It focuses on the graph structure that underpins Git and the way the properties of this graph dictate Git\u2019s behavior. Looking at fundamentals, you build your mental model on the truth rather than on hypotheses constructed from evidence gathered while experimenting with the API. This truer model gives you a better understanding of what Git has done, what it is doing, and what it will do.<\/p><p>The text is structured as a series of Git commands run on a single project. At intervals, there are observations about the graph data structure that Git is built on. These observations illustrate a property of the graph and the behavior that this property produces.<\/p><p>After reading, if you wish to go even deeper into Git, you can look at the <a href=\"http:\/\/gitlet.maryrosecook.com\/docs\/gitlet.html\">heavily annotated source code<\/a> of my implementation of Git in JavaScript.<\/p><h3 id=\"create-the-project\">Create the project<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~ $ mkdir alpha\n~ $ cd alpha\n<\/pre><\/div><p>The user creates <code class=\"highlighter-rouge\">alpha<\/code>, a directory for their project.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ mkdir data\n~\/alpha $ printf 'a' &gt; data\/letter.txt\n<\/pre><\/div><p>They move into the <code class=\"highlighter-rouge\">alpha<\/code> directory and create a directory called <code class=\"highlighter-rouge\">data<\/code>. Inside, they create a file called <code class=\"highlighter-rouge\">letter.txt<\/code> that contains <code class=\"highlighter-rouge\">a<\/code>. The alpha directory looks like this:<\/p><div class=\"highlighter-rouge\"><pre>alpha\n\u2514\u2500\u2500 data\n    \u2514\u2500\u2500 letter.txt\n<\/pre><\/div><h3 id=\"initialize-the-repository\">Initialize the repository<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git init\n          Initialized empty Git repository\n<\/pre><\/div><p><code class=\"highlighter-rouge\">git init<\/code> makes the current directory into a Git repository. To do this, it creates a <code class=\"highlighter-rouge\">.git<\/code> directory and writes some files to it. These files define everything about the Git configuration and the history of the project. They are just ordinary files. No magic in them. The user can read and edit them with a text editor or shell. Which is to say: the user can read and edit the history of their project as easily as their project files.<\/p><p>The <code class=\"highlighter-rouge\">alpha<\/code> directory now looks like this:<\/p><div class=\"highlighter-rouge\"><pre>alpha\n\u251c\u2500\u2500 data\n|   \u2514\u2500\u2500 letter.txt\n\u2514\u2500\u2500 .git\n    \u251c\u2500\u2500 objects\n    etc...\n<\/pre><\/div><p>The <code class=\"highlighter-rouge\">.git<\/code> directory and its contents are Git\u2019s. All the other files are collectively known as the working copy. They are the user\u2019s.<\/p><h3 id=\"add-some-files\">Add some files<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git add data\/letter.txt\n<\/pre><\/div><p>The user runs <code class=\"highlighter-rouge\">git add<\/code> on <code class=\"highlighter-rouge\">data\/letter.txt<\/code>. This has two effects.<\/p><p>First, it creates a new blob file in the <code class=\"highlighter-rouge\">.git\/objects\/<\/code> directory.<\/p><p>This blob file contains the compressed content of <code class=\"highlighter-rouge\">data\/letter.txt<\/code>. Its name is derived by hashing its content. Hashing a piece of text means running a program on it that turns it into a smaller<sup id=\"fnref:1\"><a href=\"#fn:1\" class=\"footnote\">1<\/a><\/sup> piece of text that uniquely<sup id=\"fnref:2\"><a href=\"#fn:2\" class=\"footnote\">2<\/a><\/sup> identifies the original. For example, Git hashes <code class=\"highlighter-rouge\">a<\/code> to <code class=\"highlighter-rouge\">2e65efe2a145dda7ee51d1741299f848e5bf752e<\/code>. The first two characters are used as the name of a directory inside the objects database: <code class=\"highlighter-rouge\">.git\/objects\/2e\/<\/code>. The rest of the hash is used as the name of the blob file that holds the content of the added file: <code class=\"highlighter-rouge\">.git\/objects\/2e\/65efe2a145dda7ee51d1741299f848e5bf752e<\/code>.<\/p><p>Notice how just adding a file to Git saves its content to the <code class=\"highlighter-rouge\">objects<\/code> directory. Its content will still be safe inside Git if the user deletes <code class=\"highlighter-rouge\">data\/letter.txt<\/code> from the working copy.<\/p><p>Second, <code class=\"highlighter-rouge\">git add<\/code> adds the file to the index. The index is a list that contains every file that Git has been told to keep track of. It is stored as a file at <code class=\"highlighter-rouge\">.git\/index<\/code>. Each line of the file maps a tracked file to the hash of its content at the moment it was added. This is the index after the <code class=\"highlighter-rouge\">git add<\/code> command is run:<\/p><div class=\"highlighter-rouge\"><pre>data\/letter.txt 2e65efe2a145dda7ee51d1741299f848e5bf752e\n<\/pre><\/div><p>The user makes a file called <code class=\"highlighter-rouge\">data\/number.txt<\/code> that contains <code class=\"highlighter-rouge\">1234<\/code>.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '1234' &gt; data\/number.txt\n<\/pre><\/div><p>The working copy looks like this:<\/p><div class=\"highlighter-rouge\"><pre>alpha\n\u2514\u2500\u2500 data\n    \u2514\u2500\u2500 letter.txt\n    \u2514\u2500\u2500 number.txt\n<\/pre><\/div><p>The user adds the file to Git.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git add data\n<\/pre><\/div><p>The <code class=\"highlighter-rouge\">git add<\/code> command creates a blob object that contains the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code>. It adds an index entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> that points at the blob. This is the index after the <code class=\"highlighter-rouge\">git add<\/code> command is run a second time:<\/p><div class=\"highlighter-rouge\"><pre>data\/letter.txt 2e65efe2a145dda7ee51d1741299f848e5bf752e\ndata\/number.txt 274c0052dd5408f8ae2bc8440029ff67d79bc5c3\n<\/pre><\/div><p>Notice that only the files in the <code class=\"highlighter-rouge\">data<\/code> directory are listed in the index, though the user ran <code class=\"highlighter-rouge\">git add data<\/code>. The <code class=\"highlighter-rouge\">data<\/code> directory is not listed separately.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '1' &gt; data\/number.txt\n~\/alpha $ git add data\n<\/pre><\/div><p>When the user originally created <code class=\"highlighter-rouge\">data\/number.txt<\/code>, they meant to type <code class=\"highlighter-rouge\">1<\/code>, not <code class=\"highlighter-rouge\">1234<\/code>. They make the correction and add the file to the index again. This command creates a new blob with the new content. And it updates the index entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> to point at the new blob.<\/p><h3 id=\"make-a-commit\">Make a commit<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git commit -m 'a1'\n          [master (root-commit) 774b54a] a1\n<\/pre><\/div><p>The user makes the <code class=\"highlighter-rouge\">a1<\/code> commit. Git prints some data about the commit.  These data will make sense shortly.<\/p><p>The commit command has three steps. It creates a tree graph to represent the content of the version of the project being committed. It creates a commit object. It points the current branch at the new commit object.<\/p><h4 id=\"create-a-tree-graph\">Create a tree graph<\/h4><p>Git records the current state of the project by creating a tree graph from the index. This tree graph records the location and content of every file in the project.<\/p><p>The graph is composed of two types of object: blobs and trees.<\/p><p>Blobs are stored by <code class=\"highlighter-rouge\">git add<\/code>. They represent the content of files.<\/p><p>Trees are stored when a commit is made. A tree represents a directory in the working copy.<\/p><p>Below is the tree object that records the contents of the <code class=\"highlighter-rouge\">data<\/code> directory for the new commit:<\/p><div class=\"highlighter-rouge\"><pre>100664 blob 2e65efe2a145dda7ee51d1741299f848e5bf752e letter.txt\n100664 blob 56a6051ca2b02b04ef92d5150c9ef600403cb1de number.txt\n<\/pre><\/div><p>The first line records everything required to reproduce <code class=\"highlighter-rouge\">data\/letter.txt<\/code>.  The first part states the file\u2019s permissions.  The second part states that the content of this entry is represented by a blob, rather than a tree.  The third part states the hash of the blob.  The fourth part states the file\u2019s name.<\/p><p>The second line records the same for <code class=\"highlighter-rouge\">data\/number.txt<\/code>.<\/p><p>Below is the tree object for <code class=\"highlighter-rouge\">alpha<\/code>, which is the root directory of the project:<\/p><div class=\"highlighter-rouge\"><pre>040000 tree 0eed1217a2947f4930583229987d90fe5e8e0b74 data\n<\/pre><\/div><p>The sole line in this tree points at the <code class=\"highlighter-rouge\">data<\/code> tree.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/1-a1-tree-graph.png\" alt=\"Tree graph for the `a1` commit\" \/><\/p><p>Tree graph for the `a1` commit<\/p><p>In the graph above, the <code class=\"highlighter-rouge\">root<\/code> tree points at the <code class=\"highlighter-rouge\">data<\/code> tree. The <code class=\"highlighter-rouge\">data<\/code> tree points at the blobs for <code class=\"highlighter-rouge\">data\/letter.txt<\/code> and <code class=\"highlighter-rouge\">data\/number.txt<\/code>.<\/p><h4 id=\"create-a-commit-object\">Create a commit object<\/h4><p><code class=\"highlighter-rouge\">git commit<\/code> creates a commit object after creating the tree graph. The commit object is just another text file in <code class=\"highlighter-rouge\">.git\/objects\/<\/code>:<\/p><div class=\"highlighter-rouge\"><pre>tree ffe298c3ce8bb07326f888907996eaa48d266db4\nauthor Mary Rose Cook &lt;mary@maryrosecook.com&gt; 1424798436 -0500\ncommitter Mary Rose Cook &lt;mary@maryrosecook.com&gt; 1424798436 -0500\na1\n<\/pre><\/div><p>The first line points at the tree graph. The hash is for the tree object that represents the root of the working copy.  That is: the <code class=\"highlighter-rouge\">alpha<\/code> directory.  The last line is the commit message.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/2-a1-commit.png\" alt=\"`a1` commit object pointing at its tree graph\" \/><\/p><p>`a1` commit object pointing at its tree graph<\/p><h4 id=\"point-the-current-branch-at-the-new-commit\">Point the current branch at the new commit<\/h4><p>Finally, the commit command points the current branch at the new commit object.<\/p><p>Which is the current branch? Git goes to the <code class=\"highlighter-rouge\">HEAD<\/code> file at <code class=\"highlighter-rouge\">.git\/HEAD<\/code> and finds:<\/p><div class=\"highlighter-rouge\"><pre>ref: refs\/heads\/master\n<\/pre><\/div><p>This says that <code class=\"highlighter-rouge\">HEAD<\/code> is pointing at <code class=\"highlighter-rouge\">master<\/code>.  <code class=\"highlighter-rouge\">master<\/code> is the current branch.<\/p><p><code class=\"highlighter-rouge\">HEAD<\/code> and <code class=\"highlighter-rouge\">master<\/code> are both refs. A ref is a label used by Git or the user to identify a specific commit.<\/p><p>The file that represents the <code class=\"highlighter-rouge\">master<\/code> ref does not exist, because this is the first commit to the repository. Git creates the file at <code class=\"highlighter-rouge\">.git\/refs\/heads\/master<\/code> and sets its content to the hash of the commit object:<\/p><div class=\"highlighter-rouge\"><pre>74ac3ad9cde0b265d2b4f1c778b283a6e2ffbafd\n<\/pre><\/div><p>(If you are typing in these Git commands as you read, the hash of your <code class=\"highlighter-rouge\">a1<\/code> commit will be different from the hash of mine. Content objects like blobs and trees always hash to the same value. Commits do not, because they include dates and the names of their creators.)<\/p><p>Let\u2019s add <code class=\"highlighter-rouge\">HEAD<\/code> and <code class=\"highlighter-rouge\">master<\/code> to the Git graph:<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/3-a1-refs.png\" alt=\"`master` pointing at the `a1` commit\" \/><\/p><p>`HEAD` pointing at `master` and `master` pointing at the `a1` commit<\/p><p><code class=\"highlighter-rouge\">HEAD<\/code> points at <code class=\"highlighter-rouge\">master<\/code>, as it did before the commit. But <code class=\"highlighter-rouge\">master<\/code> now exists and points at the new commit object.<\/p><h3 id=\"make-a-commit-that-is-not-the-first-commit\">Make a commit that is not the first commit<\/h3><p>Below is the Git graph after the <code class=\"highlighter-rouge\">a1<\/code> commit. The working copy and index are included.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/4-a1-wc-and-index.png\" alt=\"`a1` commit shown with the working copy and index\" \/><\/p><p>`a1` commit shown with the working copy and index<\/p><p>Notice that the working copy, index, and <code class=\"highlighter-rouge\">a1<\/code> commit all have the same content for <code class=\"highlighter-rouge\">data\/letter.txt<\/code> and <code class=\"highlighter-rouge\">data\/number.txt<\/code>. The index and <code class=\"highlighter-rouge\">HEAD<\/code> commit both use hashes to refer to blob objects, but the working copy content is stored as text in a different place.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '2' &gt; data\/number.txt\n<\/pre><\/div><p>The user sets the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">2<\/code>. This updates the working copy, but leaves the index and <code class=\"highlighter-rouge\">HEAD<\/code> commit as they are.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/5-a1-wc-number-set-to-2.png\" alt=\"`data\/number.txt` set to `2` in the working copy\" \/><\/p><p>`data\/number.txt` set to `2` in the working copy<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git add data\/number.txt\n<\/pre><\/div><p>The user adds the file to Git. This adds a blob containing <code class=\"highlighter-rouge\">2<\/code> to the <code class=\"highlighter-rouge\">objects<\/code> directory. It points the index entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> at the new blob.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/6-a1-wc-and-index-number-set-to-2.png\" alt=\"`data\/number.txt` set to `2` in the working copy and index\" \/><\/p><p>`data\/number.txt` set to `2` in the working copy and index<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git commit -m 'a2'\n          [master f0af7e6] a2\n<\/pre><\/div><p>The user commits. The steps for the commit are the same as before.<\/p><p>First, a new tree graph is created to represent the content of the index.<\/p><p>The index entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> has changed.  The old <code class=\"highlighter-rouge\">data<\/code> tree no longer reflects the indexed state of the <code class=\"highlighter-rouge\">data<\/code> directory.  A new <code class=\"highlighter-rouge\">data<\/code> tree object must be created:<\/p><div class=\"highlighter-rouge\"><pre>100664 blob 2e65efe2a145dda7ee51d1741299f848e5bf752e letter.txt\n100664 blob d8263ee9860594d2806b0dfd1bfd17528b0ba2a4 number.txt\n<\/pre><\/div><p>The new <code class=\"highlighter-rouge\">data<\/code> tree hashes to a different value from the old <code class=\"highlighter-rouge\">data<\/code> tree.  A new <code class=\"highlighter-rouge\">root<\/code> tree must be created to record this hash:<\/p><div class=\"highlighter-rouge\"><pre>040000 tree 40b0318811470aaacc577485777d7a6780e51f0b data\n<\/pre><\/div><p>Second, a new commit object is created.<\/p><div class=\"highlighter-rouge\"><pre>tree ce72afb5ff229a39f6cce47b00d1b0ed60fe3556\nparent 774b54a193d6cfdd081e581a007d2e11f784b9fe\nauthor Mary Rose Cook &lt;mary@maryrosecook.com&gt; 1424813101 -0500\ncommitter Mary Rose Cook &lt;mary@maryrosecook.com&gt; 1424813101 -0500\na2\n<\/pre><\/div><p>The first line of the commit object points at the new <code class=\"highlighter-rouge\">root<\/code> tree object. The second line points at <code class=\"highlighter-rouge\">a1<\/code>: the commit\u2019s parent. To find the parent commit, Git went to <code class=\"highlighter-rouge\">HEAD<\/code>, followed it to <code class=\"highlighter-rouge\">master<\/code> and found the commit hash of <code class=\"highlighter-rouge\">a1<\/code>.<\/p><p>Third, the content of the <code class=\"highlighter-rouge\">master<\/code> branch file is set to the hash of the new commit.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/7-a2.png\" alt=\"`a2` commit\" \/><\/p><p>`a2` commit<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/8-a2-just-objects-commits-and-refs.png\" alt=\"Git graph without the working copy and index\" \/><\/p><p>Git graph without the working copy and index<\/p><p><strong>Graph property<\/strong>: content is stored as a tree of objects. This means that only diffs are stored in the objects database. Look at the graph above. The <code class=\"highlighter-rouge\">a2<\/code> commit reuses the <code class=\"highlighter-rouge\">a<\/code> blob that was made before the <code class=\"highlighter-rouge\">a1<\/code> commit. Similarly, if a whole directory doesn\u2019t change from commit to commit, its tree and all the blobs and trees below it can be reused. Generally, there are few content changes from commit to commit. This means that Git can store large commit histories in a small amount of space.<\/p><p><strong>Graph property<\/strong>: each commit has a parent. This means that a repository can store the history of a project.<\/p><p><strong>Graph property<\/strong>: refs are entry points to one part of the commit history or another. This means that commits can be given meaningful names. The user organizes their work into lineages that are meaningful to their project with concrete refs like <code class=\"highlighter-rouge\">fix-for-bug-376<\/code>. Git uses symbolic refs like <code class=\"highlighter-rouge\">HEAD<\/code>, <code class=\"highlighter-rouge\">MERGE_HEAD<\/code> and <code class=\"highlighter-rouge\">FETCH_HEAD<\/code> to support commands that manipulate the commit history.<\/p><p><strong>Graph property<\/strong>: the nodes in the <code class=\"highlighter-rouge\">objects\/<\/code> directory are immutable. This means that content is edited, not deleted. Every piece of content ever added and every commit ever made is somewhere in the <code class=\"highlighter-rouge\">objects<\/code> directory<sup id=\"fnref:3\"><a href=\"#fn:3\" class=\"footnote\">3<\/a><\/sup>.<\/p><p><strong>Graph property<\/strong>: refs are mutable. Therefore, the meaning of a ref can change. The commit that <code class=\"highlighter-rouge\">master<\/code> points at might be the best version of a project at the moment, but, soon enough, it will be superseded by a newer and better commit.<\/p><p><strong>Graph property<\/strong>: the working copy and the commits pointed at by refs are readily available, but other commits are not. This means that recent history is easier to recall, but that it also changes more often. Or: Git has a fading memory that must be jogged with increasingly vicious prods.<\/p><p>The working copy is the easiest point in history to recall because it is in the root of the repository. Recalling it doesn\u2019t even require a Git command. It is also the least permanent point in history. The user can make a dozen versions of a file but Git won\u2019t record any of them unless they are added.<\/p><p>The commit that <code class=\"highlighter-rouge\">HEAD<\/code> points at is very easy to recall. It is at the tip of the branch that is checked out. To see its content, the user can just stash<sup id=\"fnref:4\"><a href=\"#fn:4\" class=\"footnote\">4<\/a><\/sup> and then examine the working copy. At the same time, <code class=\"highlighter-rouge\">HEAD<\/code> is the most frequently changing ref.<\/p><p>The commit that a concrete ref points at is easy to recall. The user can simply check out that branch. The tip of a branch changes less often than <code class=\"highlighter-rouge\">HEAD<\/code>, but often enough for the meaning of a branch name to be changeable.<\/p><p>It is difficult to recall a commit that is not pointed at by any ref.  The further the user goes from a ref, the harder it will be for them to construct the meaning of a commit. But the further back they go, the less likely it is that someone will have changed history since they last looked<sup id=\"fnref:5\"><a href=\"#fn:5\" class=\"footnote\">5<\/a><\/sup>.<\/p><h3 id=\"check-out-a-commit\">Check out a commit<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout 37888c2\n          You are in 'detached HEAD' state...\n<\/pre><\/div><p>The user checks out the <code class=\"highlighter-rouge\">a2<\/code> commit using its hash. (If you are running these Git commands, this one won\u2019t work. Use <code class=\"highlighter-rouge\">git log<\/code> to find the hash of your <code class=\"highlighter-rouge\">a2<\/code> commit.)<\/p><p>Checking out has four steps.<\/p><p>First, Git gets the <code class=\"highlighter-rouge\">a2<\/code> commit and gets the tree graph it points at.<\/p><p>Second, it writes the file entries in the tree graph to the working copy. This results in no changes. The working copy already has the content of the tree graph being written to it because <code class=\"highlighter-rouge\">HEAD<\/code> was already pointing via <code class=\"highlighter-rouge\">master<\/code> at the <code class=\"highlighter-rouge\">a2<\/code> commit.<\/p><p>Third, Git writes the file entries in the tree graph to the index. This, too, results in no changes. The index already has the content of the <code class=\"highlighter-rouge\">a2<\/code> commit.<\/p><p>Fourth, the content of <code class=\"highlighter-rouge\">HEAD<\/code> is set to the hash of the <code class=\"highlighter-rouge\">a2<\/code> commit:<\/p><div class=\"highlighter-rouge\"><pre>f0af7e62679e144bb28c627ee3e8f7bdb235eee9\n<\/pre><\/div><p>Setting the content of <code class=\"highlighter-rouge\">HEAD<\/code> to a hash puts the repository in the detached <code class=\"highlighter-rouge\">HEAD<\/code> state. Notice in the graph below that <code class=\"highlighter-rouge\">HEAD<\/code> points directly at the <code class=\"highlighter-rouge\">a2<\/code> commit, rather than pointing at <code class=\"highlighter-rouge\">master<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/9-a2-detached-head.png\" alt=\"Detached `HEAD` on `a2` commit\" \/><\/p><p>Detached `HEAD` on `a2` commit<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '3' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n~\/alpha $ git commit -m 'a3'\n          [detached HEAD 3645a0e] a3\n<\/pre><\/div><p>The user sets the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">3<\/code> and commits the change. Git goes to <code class=\"highlighter-rouge\">HEAD<\/code> to get the parent of the <code class=\"highlighter-rouge\">a3<\/code> commit.  Instead of finding and following a branch ref, it finds and returns the hash of the <code class=\"highlighter-rouge\">a2<\/code> commit.<\/p><p>Git updates <code class=\"highlighter-rouge\">HEAD<\/code> to point directly at the hash of the new <code class=\"highlighter-rouge\">a3<\/code> commit. The repository is still in the detached <code class=\"highlighter-rouge\">HEAD<\/code> state. The commit is not on a branch because no branch ref points at either <code class=\"highlighter-rouge\">a3<\/code> or one of its descendants. This means it is easy to lose.<\/p><p>From now on, trees and blobs will mostly be omitted from the graph diagrams.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/10-a3-detached-head.png\" alt=\"`a3` commit that is not on a branch\" \/><\/p><p>`a3` commit that is not on a branch<\/p><h3 id=\"create-a-branch\">Create a branch<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git branch deputy\n<\/pre><\/div><p>The user creates a new branch called <code class=\"highlighter-rouge\">deputy<\/code>. This just creates a new file at <code class=\"highlighter-rouge\">.git\/refs\/heads\/deputy<\/code> that contains the hash that <code class=\"highlighter-rouge\">HEAD<\/code> is pointing at: the hash of the <code class=\"highlighter-rouge\">a3<\/code> commit.<\/p><p><strong>Graph property<\/strong>: branches are just refs and refs are just files. This means that Git branches are lightweight.<\/p><p>The creation of the <code class=\"highlighter-rouge\">deputy<\/code> branch puts the new <code class=\"highlighter-rouge\">a3<\/code> commit safely on a branch. <code class=\"highlighter-rouge\">HEAD<\/code> is still detached because it still points directly at a commit.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/11-a3-on-deputy.png\" alt=\"`a3` commit now on the `deputy` branch\" \/><\/p><p>`a3` commit now on the `deputy` branch<\/p><h3 id=\"check-out-a-branch\">Check out a branch<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout master\n          Switched to branch 'master'\n<\/pre><\/div><p>The user checks out the <code class=\"highlighter-rouge\">master<\/code> branch.<\/p><p>First, Git gets the <code class=\"highlighter-rouge\">a2<\/code> commit that <code class=\"highlighter-rouge\">master<\/code> points at and gets the tree graph the commit points at.<\/p><p>Second, Git writes the file entries in the tree graph to the files of the working copy. This sets the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">2<\/code>.<\/p><p>Third, Git writes the file entries in the tree graph to the index. This updates the entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> to the hash of the <code class=\"highlighter-rouge\">2<\/code> blob.<\/p><p>Fourth, Git points <code class=\"highlighter-rouge\">HEAD<\/code> at <code class=\"highlighter-rouge\">master<\/code> by changing its content from a hash to:<\/p><div class=\"highlighter-rouge\"><pre>ref: refs\/heads\/master\n<\/pre><\/div><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/12-a3-on-master-on-a2.png\" alt=\"`master` checked out and pointing at the `a2` commit\" \/><\/p><p>`master` checked out and pointing at the `a2` commit<\/p><h3 id=\"check-out-a-branch-that-is-incompatible-with-the-working-copy\">Check out a branch that is incompatible with the working copy<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '789' &gt; data\/number.txt\n~\/alpha $ git checkout deputy\n          Your changes to these files would be overwritten\n          by checkout:\n            data\/number.txt\n          Commit your changes or stash them before you\n          switch branches.\n<\/pre><\/div><p>The user accidentally sets the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">789<\/code>. They try to check out <code class=\"highlighter-rouge\">deputy<\/code>. Git prevents the check out.<\/p><p><code class=\"highlighter-rouge\">HEAD<\/code> points at <code class=\"highlighter-rouge\">master<\/code> which points at <code class=\"highlighter-rouge\">a2<\/code> where <code class=\"highlighter-rouge\">data\/number.txt<\/code> reads <code class=\"highlighter-rouge\">2<\/code>. <code class=\"highlighter-rouge\">deputy<\/code> points at <code class=\"highlighter-rouge\">a3<\/code> where <code class=\"highlighter-rouge\">data\/number.txt<\/code> reads <code class=\"highlighter-rouge\">3<\/code>. The working copy version of <code class=\"highlighter-rouge\">data\/number.txt<\/code> reads <code class=\"highlighter-rouge\">789<\/code>. All these versions are different and the differences must be resolved.<\/p><p>Git could replace the working copy version of <code class=\"highlighter-rouge\">data\/number.txt<\/code> with the version in the commit being checked out.  But it avoids data loss at all costs.<\/p><p>Git could merge the working copy version with the version being checked out.  But this is complicated.<\/p><p>So, Git aborts the check out.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '2' &gt; data\/number.txt\n~\/alpha $ git checkout deputy\n          Switched to branch 'deputy'\n<\/pre><\/div><p>The user notices that they accidentally edited <code class=\"highlighter-rouge\">data\/number.txt<\/code> and sets the content back to <code class=\"highlighter-rouge\">2<\/code>. They check out <code class=\"highlighter-rouge\">deputy<\/code> successfully.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/13-a3ondeputy.png\" alt=\"`deputy` checked out\" \/><\/p><p>`deputy` checked out<\/p><h3 id=\"merge-an-ancestor\">Merge an ancestor<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git merge master\n          Already up-to-date.\n<\/pre><\/div><p>The user merges <code class=\"highlighter-rouge\">master<\/code> into <code class=\"highlighter-rouge\">deputy<\/code>. Merging two branches means merging two commits. The first commit is the one that <code class=\"highlighter-rouge\">deputy<\/code> points at: the receiver. The second commit is the one that <code class=\"highlighter-rouge\">master<\/code> points at: the giver. For this merge, Git does nothing. It reports it is <code class=\"highlighter-rouge\">Already up-to-date.<\/code>.<\/p><p><strong>Graph property<\/strong>: the series of commits in the graph are interpreted as a series of changes made to the content of the repository. This means that, in a merge, if the giver commit is an ancestor of the receiver commit, Git will do nothing. Those changes have already been incorporated.<\/p><h3 id=\"merge-a-descendent\">Merge a descendent<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout master\n          Switched to branch 'master'\n<\/pre><\/div><p>The user checks out <code class=\"highlighter-rouge\">master<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/14-a3-on-master-on-a2.png\" alt=\"`master` checked out and pointing at the `a2` commit\" \/><\/p><p>`master` checked out and pointing at the `a2` commit<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git merge deputy\n          Fast-forward\n<\/pre><\/div><p>They merge <code class=\"highlighter-rouge\">deputy<\/code> into <code class=\"highlighter-rouge\">master<\/code>. Git discovers that the receiver commit, <code class=\"highlighter-rouge\">a2<\/code>, is an ancestor of the giver commit, <code class=\"highlighter-rouge\">a3<\/code>. It can do a fast-forward merge.<\/p><p>It gets the giver commit and gets the tree graph that it points at. It writes the file entries in the tree graph to the working copy and the index. It \u201cfast-forwards\u201d <code class=\"highlighter-rouge\">master<\/code> to point at <code class=\"highlighter-rouge\">a3<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/15-a3-on-master.png\" alt=\"`a3` commit from `deputy` fast-forward merged into `master`\" \/><\/p><p>`a3` commit from `deputy` fast-forward merged into `master`<\/p><p><strong>Graph property<\/strong>: the series of commits in the graph are interpreted as a series of changes made to the content of the repository. This means that, in a merge, if the giver is a descendent of the receiver, history is not changed. There is already a sequence of commits that describe the change to make: the sequence of commits between the receiver and the giver. But, though the Git history doesn\u2019t change, the Git graph does change. The concrete ref that <code class=\"highlighter-rouge\">HEAD<\/code> points at is updated to point at the giver commit.<\/p><h3 id=\"merge-two-commits-from-different-lineages\">Merge two commits from different lineages<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '4' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n~\/alpha $ git commit -m 'a4'\n          [master 7b7bd9a] a4\n<\/pre><\/div><p>The user sets the content of <code class=\"highlighter-rouge\">number.txt<\/code> to <code class=\"highlighter-rouge\">4<\/code> and commits the change to <code class=\"highlighter-rouge\">master<\/code>.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout deputy\n          Switched to branch 'deputy'\n~\/alpha $ printf 'b' &gt; data\/letter.txt\n~\/alpha $ git add data\/letter.txt\n~\/alpha $ git commit -m 'b3'\n          [deputy 982dffb] b3\n<\/pre><\/div><p>The user checks out <code class=\"highlighter-rouge\">deputy<\/code>. They set the content of <code class=\"highlighter-rouge\">data\/letter.txt<\/code> to <code class=\"highlighter-rouge\">b<\/code> and commit the change to <code class=\"highlighter-rouge\">deputy<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/16-a4-b3-on-deputy.png\" alt=\"`a4` committed to `master`, `b3` committed to `deputy` and `deputy` checked out\" \/><\/p><p>`a4` committed to `master`, `b3` committed to `deputy` and `deputy` checked out<\/p><p><strong>Graph property<\/strong>: commits can share parents. This means that new lineages can be created in the commit history.<\/p><p><strong>Graph property<\/strong>: commits can have multiple parents. This means that separate lineages can be joined by a commit with two parents: a merge commit.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git merge master -m 'b4'\n          Merge made by the 'recursive' strategy.\n<\/pre><\/div><p>The user merges <code class=\"highlighter-rouge\">master<\/code> into <code class=\"highlighter-rouge\">deputy<\/code>.<\/p><p>Git discovers that the receiver, <code class=\"highlighter-rouge\">b3<\/code>, and the giver, <code class=\"highlighter-rouge\">a4<\/code>, are in different lineages. It makes a merge commit. This process has eight steps.<\/p><p>First, Git writes the hash of the giver commit to a file at <code class=\"highlighter-rouge\">alpha\/.git\/MERGE_HEAD<\/code>. The presence of this file tells Git it is in the middle of merging.<\/p><p>Second, Git finds the base commit: the most recent ancestor that the receiver and giver commits have in common.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/17-a4-b3-on-deputy.png\" alt=\"`a3`, the base commit of `a4` and `b3`\" \/><\/p><p>`a3`, the base commit of `a4` and `b3`<\/p><p><strong>Graph property<\/strong>: commits have parents. This means that it is possible to find the point at which two lineages diverged. Git traces backwards from <code class=\"highlighter-rouge\">b3<\/code> to find all its ancestors and backwards from <code class=\"highlighter-rouge\">a4<\/code> to find all its ancestors. It finds the most recent ancestor shared by both lineages, <code class=\"highlighter-rouge\">a3<\/code>. This is the base commit.<\/p><p>Third, Git generates the indices for the base, receiver and giver commits from their tree graphs.<\/p><p>Fourth, Git generates a diff that combines the changes made to the base by the receiver commit and the giver commit. This diff is a list of file paths that point to a change: add, remove, modify or conflict.<\/p><p>Git gets the list of all the files that appear in the base, receiver or giver indices. For each one, it compares the index entries to decide the change to make to the file. It writes a corresponding entry to the diff. In this case, the diff has two entries.<\/p><p>The first entry is for <code class=\"highlighter-rouge\">data\/letter.txt<\/code>. The content of this file is <code class=\"highlighter-rouge\">a<\/code> in the base, <code class=\"highlighter-rouge\">b<\/code> in the receiver and <code class=\"highlighter-rouge\">a<\/code> in the giver. The content is different in the base and receiver. But it is the same in the base and giver. Git sees that the content was modified by the receiver, but not the giver. The diff entry for <code class=\"highlighter-rouge\">data\/letter.txt<\/code> is a modification, not a conflict.<\/p><p>The second entry in the diff is for <code class=\"highlighter-rouge\">data\/number.txt<\/code>. In this case, the content is the same in the base and receiver, and different in the giver. The diff entry for <code class=\"highlighter-rouge\">data\/letter.txt<\/code> is also a modification.<\/p><p><strong>Graph property<\/strong>: it is possible to find the base commit of a merge. This means that, if a file has changed from the base in just the receiver or giver, Git can automatically resolve the merge of that file. This reduces the work the user must do.<\/p><p>Fifth, the changes indicated by the entries in the diff are applied to the working copy. The content of <code class=\"highlighter-rouge\">data\/letter.txt<\/code> is set to <code class=\"highlighter-rouge\">b<\/code> and the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> is set to <code class=\"highlighter-rouge\">4<\/code>.<\/p><p>Sixth, the changes indicated by the entries in the diff are applied to the index. The entry for <code class=\"highlighter-rouge\">data\/letter.txt<\/code> is pointed at the <code class=\"highlighter-rouge\">b<\/code> blob and the entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> is pointed at the <code class=\"highlighter-rouge\">4<\/code> blob.<\/p><p>Seventh, the updated index is committed:<\/p><div class=\"highlighter-rouge\"><pre>tree 20294508aea3fb6f05fcc49adaecc2e6d60f7e7d\nparent 982dffb20f8d6a25a8554cc8d765fb9f3ff1333b\nparent 7b7bd9a5253f47360d5787095afc5ba56591bfe7\nauthor Mary Rose Cook &lt;mary@maryrosecook.com&gt; 1425596551 -0500\ncommitter Mary Rose Cook &lt;mary@maryrosecook.com&gt; 1425596551 -0500\nb4\n<\/pre><\/div><p>Notice that the commit has two parents.<\/p><p>Eighth, Git points the current branch, <code class=\"highlighter-rouge\">deputy<\/code>, at the new commit.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/18-b4-on-deputy.png\" alt=\"`b4`, the merge commit resulting from the recursive merge of `a4` into `b3`\" \/><\/p><p>`b4`, the merge commit resulting from the recursive merge of `a4` into `b3`<\/p><h3 id=\"merge-two-commits-from-different-lineages-that-both-modify-the-same-file\">Merge two commits from different lineages that both modify the same file<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout master\n          Switched to branch 'master'\n~\/alpha $ git merge deputy\n          Fast-forward\n<\/pre><\/div><p>The user checks out <code class=\"highlighter-rouge\">master<\/code>. They merge <code class=\"highlighter-rouge\">deputy<\/code> into <code class=\"highlighter-rouge\">master<\/code>. This fast-forwards <code class=\"highlighter-rouge\">master<\/code> to the <code class=\"highlighter-rouge\">b4<\/code> commit. <code class=\"highlighter-rouge\">master<\/code> and <code class=\"highlighter-rouge\">deputy<\/code> now point at the same commit.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/19-b4-master-deputy-on-b4.png\" alt=\"`deputy` merged into `master` to bring `master` up to the latest commit, `b4`\" \/><\/p><p>`deputy` merged into `master` to bring `master` up to the latest commit, `b4`<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout deputy\n          Switched to branch 'deputy'\n~\/alpha $ printf '5' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n~\/alpha $ git commit -m 'b5'\n          [deputy bd797c2] b5\n<\/pre><\/div><p>The user checks out <code class=\"highlighter-rouge\">deputy<\/code>. They set the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">5<\/code> and commit the change to <code class=\"highlighter-rouge\">deputy<\/code>.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git checkout master\n          Switched to branch 'master'\n~\/alpha $ printf '6' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n~\/alpha $ git commit -m 'b6'\n          [master 4c3ce18] b6\n<\/pre><\/div><p>The user checks out <code class=\"highlighter-rouge\">master<\/code>. They set the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">6<\/code> and commit the change to <code class=\"highlighter-rouge\">master<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/20-b5-on-deputy-b6-on-master.png\" alt=\"`b5` commit on `deputy` and `b6` commit on `master`\" \/><\/p><p>`b5` commit on `deputy` and `b6` commit on `master`<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git merge deputy\n          CONFLICT in data\/number.txt\n          Automatic merge failed; fix conflicts and\n          commit the result.\n<\/pre><\/div><p>The user merges <code class=\"highlighter-rouge\">deputy<\/code> into <code class=\"highlighter-rouge\">master<\/code>. There is a conflict and the merge is paused. The process for a conflicted merge follows the same first six steps as the process for an unconflicted merge: set <code class=\"highlighter-rouge\">.git\/MERGE_HEAD<\/code>, find the base commit, generate the indices of the base, receiver and giver commits, create a diff, update the working copy and update the index. Because of the conflict, the seventh commit step and eighth ref update step are never taken. Let\u2019s go through the steps again and see what happens.<\/p><p>First, Git writes the hash of the giver commit to a file at <code class=\"highlighter-rouge\">.git\/MERGE_HEAD<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/21-b6-on-master-with-merge-head.png\" alt=\"`MERGE_HEAD` written during merge of `b5` into `b6`\" \/><\/p><p>`MERGE_HEAD` written during merge of `b5` into `b6`<\/p><p>Second, Git finds the base commit, <code class=\"highlighter-rouge\">b4<\/code>.<\/p><p>Third, Git generates the indices for the base, receiver and giver commits.<\/p><p>Fourth, Git generates a diff that combines the changes made to the base by the receiver commit and the giver commit. This diff is a list of file paths that point to a change: add, remove, modify or conflict.<\/p><p>In this case, the diff contains only one entry: <code class=\"highlighter-rouge\">data\/number.txt<\/code>. The entry is marked as a conflict because the content for <code class=\"highlighter-rouge\">data\/number.txt<\/code> is different in the receiver, giver and base.<\/p><p>Fifth, the changes indicated by the entries in the diff are applied to the working copy. For a conflicted area, Git writes both versions to the file in the working copy. The content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> is set to:<\/p><div class=\"highlighter-rouge\"><pre>&lt;&lt;&lt;&lt;&lt;&lt;&lt; HEAD\n6\n=======\n5\n&gt;&gt;&gt;&gt;&gt;&gt;&gt; deputy\n<\/pre><\/div><p>Sixth, the changes indicated by the entries in the diff are applied to the index. Entries in the index are uniquely identified by a combination of their file path and stage. The entry for an unconflicted file has a stage of <code class=\"highlighter-rouge\">0<\/code>. Before this merge, the index looked like this, where the <code class=\"highlighter-rouge\">0<\/code>s are stage values:<\/p><div class=\"highlighter-rouge\"><pre>0 data\/letter.txt 63d8dbd40c23542e740659a7168a0ce3138ea748\n0 data\/number.txt 62f9457511f879886bb7728c986fe10b0ece6bcb\n<\/pre><\/div><p>After the merge diff is written to the index, the index looks like this:<\/p><div class=\"highlighter-rouge\"><pre>0 data\/letter.txt 63d8dbd40c23542e740659a7168a0ce3138ea748\n1 data\/number.txt bf0d87ab1b2b0ec1a11a3973d2845b42413d9767\n2 data\/number.txt 62f9457511f879886bb7728c986fe10b0ece6bcb\n3 data\/number.txt 7813681f5b41c028345ca62a2be376bae70b7f61\n<\/pre><\/div><p>The entry for <code class=\"highlighter-rouge\">data\/letter.txt<\/code> at stage <code class=\"highlighter-rouge\">0<\/code> is the same as it was before the merge. The entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> at stage <code class=\"highlighter-rouge\">0<\/code> is gone. There are three new entries in its place. The entry for stage <code class=\"highlighter-rouge\">1<\/code> has the hash of the base <code class=\"highlighter-rouge\">data\/number.txt<\/code> content. The entry for stage <code class=\"highlighter-rouge\">2<\/code> has the hash of the receiver <code class=\"highlighter-rouge\">data\/number.txt<\/code> content. The entry for stage <code class=\"highlighter-rouge\">3<\/code> has the hash of the giver <code class=\"highlighter-rouge\">data\/number.txt<\/code> content. The presence of these three entries tells Git that <code class=\"highlighter-rouge\">data\/number.txt<\/code> is in conflict.<\/p><p>The merge pauses.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '11' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n<\/pre><\/div><p>The user integrates the content of the two conflicting versions by setting the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">11<\/code>. They add the file to the index.  Git adds a blob containing <code class=\"highlighter-rouge\">11<\/code>. Adding a conflicted file tells Git that the conflict is resolved. Git removes the <code class=\"highlighter-rouge\">data\/number.txt<\/code> entries for stages <code class=\"highlighter-rouge\">1<\/code>, <code class=\"highlighter-rouge\">2<\/code> and <code class=\"highlighter-rouge\">3<\/code> from the index.  It adds an entry for <code class=\"highlighter-rouge\">data\/number.txt<\/code> at stage <code class=\"highlighter-rouge\">0<\/code> with the hash of the new blob. The index now reads:<\/p><div class=\"highlighter-rouge\"><pre>0 data\/letter.txt 63d8dbd40c23542e740659a7168a0ce3138ea748\n0 data\/number.txt 9d607966b721abde8931ddd052181fae905db503\n<\/pre><\/div><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git commit -m 'b11'\n          [master 251a513] b11\n<\/pre><\/div><p>Seventh, the user commits. Git sees <code class=\"highlighter-rouge\">.git\/MERGE_HEAD<\/code> in the repository, which tells it that a merge is in progress. It checks the index and finds there are no conflicts. It creates a new commit, <code class=\"highlighter-rouge\">b11<\/code>, to record the content of the resolved merge. It deletes the file at <code class=\"highlighter-rouge\">.git\/MERGE_HEAD<\/code>. This completes the merge.<\/p><p>Eighth, Git points the current branch, <code class=\"highlighter-rouge\">master<\/code>, at the new commit.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/22-b11-on-master.png\" alt=\"`b11`, the merge commit resulting from the conflicted, recursive merge of `b5` into `b6`\" \/><\/p><p>`b11`, the merge commit resulting from the conflicted, recursive merge of `b5` into `b6`<\/p><h3 id=\"remove-a-file\">Remove a file<\/h3><p>This diagram of the Git graph includes the commit history, the trees and blobs for the latest commit, and the working copy and index:<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/23-b11-with-objects-wc-and-index.png\" alt=\"The working copy, index, `b11` commit and its tree graph\" \/><\/p><p>The working copy, index, `b11` commit and its tree graph<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git rm data\/letter.txt\n          rm 'data\/letter.txt'\n<\/pre><\/div><p>The user tells Git to remove <code class=\"highlighter-rouge\">data\/letter.txt<\/code>.  The file is deleted from the working copy. The entry is deleted from the index.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/24-b11-letter-removed-from-wc-and-index.png\" alt=\"After `data\/letter.txt` `rm`ed from working copy and index\" \/><\/p><p>After `data\/letter.txt` `rm`ed from working copy and index<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git commit -m '11'\n          [master d14c7d2] 11\n<\/pre><\/div><p>The user commits. As part of the commit, as always, Git builds a tree graph that represents the content of the index. <code class=\"highlighter-rouge\">data\/letter.txt<\/code> is not included in the tree graph because it is not in the index.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/25-11.png\" alt=\"`11` commit made after `data\/letter.txt` `rm`ed\" \/><\/p><p>`11` commit made after `data\/letter.txt` `rm`ed<\/p><h3 id=\"copy-a-repository\">Copy a repository<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ cd ..\n      ~ $ cp -R alpha bravo\n<\/pre><\/div><p>The user copies the contents of the <code class=\"highlighter-rouge\">alpha\/<\/code> repository to the <code class=\"highlighter-rouge\">bravo\/<\/code> directory. This produces the following directory structure:<\/p><div class=\"highlighter-rouge\"><pre>~\n\u251c\u2500\u2500 alpha\n|   \u2514\u2500\u2500 data\n|       \u2514\u2500\u2500 number.txt\n\u2514\u2500\u2500 bravo\n    \u2514\u2500\u2500 data\n        \u2514\u2500\u2500 number.txt\n<\/pre><\/div><p>There is now another Git graph in the <code class=\"highlighter-rouge\">bravo<\/code> directory:<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/26-11-cp-alpha-to-bravo.png\" alt=\"New graph created when `alpha` `cp`ed to `bravo`\" \/><\/p><p>New graph created when `alpha` `cp`ed to `bravo`<\/p><h3 id=\"link-a-repository-to-another-repository\">Link a repository to another repository<\/h3><div class=\"language-bash highlighter-rouge\"><pre>      ~ $ cd alpha\n~\/alpha $ git remote add bravo ..\/bravo\n<\/pre><\/div><p>The user moves back into the <code class=\"highlighter-rouge\">alpha<\/code> repository. They set up <code class=\"highlighter-rouge\">bravo<\/code> as a remote repository on <code class=\"highlighter-rouge\">alpha<\/code>. This adds some lines to the file at <code class=\"highlighter-rouge\">alpha\/.git\/config<\/code>:<\/p><div class=\"highlighter-rouge\"><pre>[remote \"bravo\"]\n\turl = ..\/bravo\/\n<\/pre><\/div><p>These lines specify that there is a remote repository called <code class=\"highlighter-rouge\">bravo<\/code> in the directory at <code class=\"highlighter-rouge\">..\/bravo<\/code>.<\/p><h3 id=\"fetch-a-branch-from-a-remote\">Fetch a branch from a remote<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ cd ..\/bravo\n~\/bravo $ printf '12' &gt; data\/number.txt\n~\/bravo $ git add data\/number.txt\n~\/bravo $ git commit -m '12'\n          [master 94cd04d] 12\n<\/pre><\/div><p>The user goes into the <code class=\"highlighter-rouge\">bravo<\/code> repository. They set the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">12<\/code> and commit the change to <code class=\"highlighter-rouge\">master<\/code> on <code class=\"highlighter-rouge\">bravo<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/27-12-bravo.png\" alt=\"`12` commit on `bravo` repository\" \/><\/p><p>`12` commit on `bravo` repository<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/bravo $ cd ..\/alpha\n~\/alpha $ git fetch bravo master\n          Unpacking objects: 100%\n          From ..\/bravo\n            * branch master -&gt; FETCH_HEAD\n<\/pre><\/div><p>The user goes into the <code class=\"highlighter-rouge\">alpha<\/code> repository. They fetch <code class=\"highlighter-rouge\">master<\/code> from <code class=\"highlighter-rouge\">bravo<\/code> into <code class=\"highlighter-rouge\">alpha<\/code>. This process has four steps.<\/p><p>First, Git gets the hash of the commit that master is pointing at on <code class=\"highlighter-rouge\">bravo<\/code>. This is the hash of the <code class=\"highlighter-rouge\">12<\/code> commit.<\/p><p>Second, Git makes a list of all the objects that the <code class=\"highlighter-rouge\">12<\/code> commit depends on: the commit object itself, the objects in its tree graph, the ancestor commits of the <code class=\"highlighter-rouge\">12<\/code> commit and the objects in their tree graphs. It removes from this list any objects that the <code class=\"highlighter-rouge\">alpha<\/code> object database already has.  It copies the rest to <code class=\"highlighter-rouge\">alpha\/.git\/objects\/<\/code>.<\/p><p>Third, the content of the concrete ref file at <code class=\"highlighter-rouge\">alpha\/.git\/refs\/remotes\/bravo\/master<\/code> is set to the hash of the <code class=\"highlighter-rouge\">12<\/code> commit.<\/p><p>Fourth, the content of <code class=\"highlighter-rouge\">alpha\/.git\/FETCH_HEAD<\/code> is set to:<\/p><div class=\"highlighter-rouge\"><pre>94cd04d93ae88a1f53a4646532b1e8cdfbc0977f branch 'master' of ..\/bravo\n<\/pre><\/div><p>This indicates that the most recent fetch command fetched the <code class=\"highlighter-rouge\">12<\/code> commit of <code class=\"highlighter-rouge\">master<\/code> from <code class=\"highlighter-rouge\">bravo<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/28-12-fetched-to-alpha.png\" alt=\"`alpha` after `bravo\/master` fetched\" \/><\/p><p>`alpha` after `bravo\/master` fetched<\/p><p><strong>Graph property<\/strong>: objects can be copied. This means that history can be shared between repositories.<\/p><p><strong>Graph property<\/strong>: a repository can store remote branch refs like <code class=\"highlighter-rouge\">alpha\/.git\/refs\/remotes\/bravo\/master<\/code>. This means that a repository can record locally the state of a branch on a remote repository. It is correct at the time it is fetched but will go out of date if the remote branch changes.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git merge FETCH_HEAD\n          Updating d14c7d2..94cd04d\n          Fast-forward\n<\/pre><\/div><p>The user merges <code class=\"highlighter-rouge\">FETCH_HEAD<\/code>. <code class=\"highlighter-rouge\">FETCH_HEAD<\/code> is just another ref. It resolves to the <code class=\"highlighter-rouge\">12<\/code> commit, the giver. <code class=\"highlighter-rouge\">HEAD<\/code> points at the <code class=\"highlighter-rouge\">11<\/code> commit, the receiver. Git does a fast-forward merge and points <code class=\"highlighter-rouge\">master<\/code> at the <code class=\"highlighter-rouge\">12<\/code> commit.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/29-12-merged-to-alpha.png\" alt=\"`alpha` after `FETCH_HEAD` merged\" \/><\/p><p>`alpha` after `FETCH_HEAD` merged<\/p><h3 id=\"pull-a-branch-from-a-remote\">Pull a branch from a remote<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git pull bravo master\n          Already up-to-date.\n<\/pre><\/div><p>The user pulls <code class=\"highlighter-rouge\">master<\/code> from <code class=\"highlighter-rouge\">bravo<\/code> into <code class=\"highlighter-rouge\">alpha<\/code>. Pull is shorthand for \u201cfetch and merge <code class=\"highlighter-rouge\">FETCH_HEAD<\/code>\u201d. Git does these two commands and reports that <code class=\"highlighter-rouge\">master<\/code> is <code class=\"highlighter-rouge\">Already up-to-date<\/code>.<\/p><h3 id=\"clone-a-repository\">Clone a repository<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ cd ..\n      ~ $ git clone alpha charlie\n          Cloning into 'charlie'\n<\/pre><\/div><p>The user moves into the directory above. They clone <code class=\"highlighter-rouge\">alpha<\/code> to <code class=\"highlighter-rouge\">charlie<\/code>. Cloning to <code class=\"highlighter-rouge\">charlie<\/code> has similar results to the <code class=\"highlighter-rouge\">cp<\/code> the user did to produce the <code class=\"highlighter-rouge\">bravo<\/code> repository. Git creates a new directory called <code class=\"highlighter-rouge\">charlie<\/code>. It inits <code class=\"highlighter-rouge\">charlie<\/code> as a Git repo, adds <code class=\"highlighter-rouge\">alpha<\/code> as a remote called <code class=\"highlighter-rouge\">origin<\/code>, fetches <code class=\"highlighter-rouge\">origin<\/code> and merges <code class=\"highlighter-rouge\">FETCH_HEAD<\/code>.<\/p><h3 id=\"push-a-branch-to-a-checked-out-branch-on-a-remote\">Push a branch to a checked-out branch on a remote<\/h3><div class=\"language-bash highlighter-rouge\"><pre>      ~ $ cd alpha\n~\/alpha $ printf '13' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n~\/alpha $ git commit -m '13'\n          [master 3238468] 13\n<\/pre><\/div><p>The user goes back into the <code class=\"highlighter-rouge\">alpha<\/code> repository. They set the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">13<\/code> and commit the change to <code class=\"highlighter-rouge\">master<\/code> on <code class=\"highlighter-rouge\">alpha<\/code>.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git remote add charlie ..\/charlie\n<\/pre><\/div><p>They set up <code class=\"highlighter-rouge\">charlie<\/code> as a remote repository on <code class=\"highlighter-rouge\">alpha<\/code>.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git push charlie master\n          Writing objects: 100%\n          remote error: refusing to update checked out\n          branch: refs\/heads\/master because it will make\n          the index and work tree inconsistent\n<\/pre><\/div><p>They push <code class=\"highlighter-rouge\">master<\/code> to <code class=\"highlighter-rouge\">charlie<\/code>.<\/p><p>All the objects required for the <code class=\"highlighter-rouge\">13<\/code> commit are copied to <code class=\"highlighter-rouge\">charlie<\/code>.<\/p><p>At this point, the push process stops. Git, as ever, tells the user what went wrong. It refuses to push to a branch that is checked out on the remote. This makes sense. A push would update the remote index and <code class=\"highlighter-rouge\">HEAD<\/code>. This would cause confusion if someone were editing the working copy on the remote.<\/p><p>At this point, the user could make a new branch, merge the <code class=\"highlighter-rouge\">13<\/code> commit into it and push that branch to <code class=\"highlighter-rouge\">charlie<\/code>. But, really, they want a repository that they can push to whenever they want. They want a central repository that they can push to and pull from, but that no one commits to directly. They want something like a GitHub remote. They want a bare repository.<\/p><h3 id=\"clone-a-bare-repository\">Clone a bare repository<\/h3><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ cd ..\n      ~ $ git clone alpha delta --bare\n          Cloning into bare repository 'delta'\n<\/pre><\/div><p>The user moves into the directory above. They clone <code class=\"highlighter-rouge\">delta<\/code> as a bare repository. This is an ordinary clone with two differences. The <code class=\"highlighter-rouge\">config<\/code> file indicates that the repository is bare. And the files that are normally stored in the <code class=\"highlighter-rouge\">.git<\/code> directory are stored in the root of the repository:<\/p><div class=\"highlighter-rouge\"><pre>delta\n\u251c\u2500\u2500 HEAD\n\u251c\u2500\u2500 config\n\u251c\u2500\u2500 objects\n\u2514\u2500\u2500 refs\n<\/pre><\/div><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/30-13-alpha-cloned-to-delta-bare.png\" alt=\"`alpha` and `delta` graphs after `alpha` cloned to `delta`\" \/><\/p><p>`alpha` and `delta` graphs after `alpha` cloned to `delta`<\/p><h3 id=\"push-a-branch-to-a-bare-repository\">Push a branch to a bare repository<\/h3><div class=\"language-bash highlighter-rouge\"><pre>      ~ $ cd alpha\n~\/alpha $ git remote add delta ..\/delta\n<\/pre><\/div><p>The user goes back into the <code class=\"highlighter-rouge\">alpha<\/code> repository. They set up <code class=\"highlighter-rouge\">delta<\/code> as a remote repository on <code class=\"highlighter-rouge\">alpha<\/code>.<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ printf '14' &gt; data\/number.txt\n~\/alpha $ git add data\/number.txt\n~\/alpha $ git commit -m '14'\n          [master cb51da8] 14\n<\/pre><\/div><p>They set the content of <code class=\"highlighter-rouge\">data\/number.txt<\/code> to <code class=\"highlighter-rouge\">14<\/code> and commit the change to <code class=\"highlighter-rouge\">master<\/code> on <code class=\"highlighter-rouge\">alpha<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/31-14-alpha.png\" alt=\"`14` commit on `alpha`\" \/><\/p><p>`14` commit on `alpha`<\/p><div class=\"language-bash highlighter-rouge\"><pre>~\/alpha $ git push delta master\n          Writing objects: 100%\n          To ..\/delta\n            3238468..cb51da8 master -&gt; master\n<\/pre><\/div><p>They push <code class=\"highlighter-rouge\">master<\/code> to <code class=\"highlighter-rouge\">delta<\/code>. Pushing has three steps.<\/p><p>First, all the objects required for the <code class=\"highlighter-rouge\">14<\/code> commit on the <code class=\"highlighter-rouge\">master<\/code> branch are copied from <code class=\"highlighter-rouge\">alpha\/.git\/objects\/<\/code> to <code class=\"highlighter-rouge\">delta\/objects\/<\/code>.<\/p><p>Second, <code class=\"highlighter-rouge\">delta\/refs\/heads\/master<\/code> is updated to point at the <code class=\"highlighter-rouge\">14<\/code> commit.<\/p><p>Third, <code class=\"highlighter-rouge\">alpha\/.git\/refs\/remotes\/delta\/master<\/code> is set to point at the <code class=\"highlighter-rouge\">14<\/code> commit. <code class=\"highlighter-rouge\">alpha<\/code> has an up-to-date record of the state of <code class=\"highlighter-rouge\">delta<\/code>.<\/p><p><img src=\"https:\/\/maryrosecook.com\/images\/git-from-the-inside-out\/32-14-pushed-to-delta.png\" alt=\"`14` commit pushed from `alpha` to `delta`\" \/><\/p><p>`14` commit pushed from `alpha` to `delta`<\/p><h3 id=\"summary\">Summary<\/h3><p>Git is built on a graph. Almost every Git command manipulates this graph. To understand Git deeply, focus on the properties of this graph, not workflows or commands.<\/p><p>To learn more about Git, investigate the <code class=\"highlighter-rouge\">.git<\/code> directory. It\u2019s not scary. Look inside. Change the content of files and see what happens. Create a commit by hand. Try and see how badly you can mess up a repo. Then repair it.<\/p><div class=\"footnotes\"><ol><li id=\"fn:1\">\n      <p>In this case, the hash is longer than the original content. But, all pieces of content longer than the number of characters in a hash will be expressed more concisely than the original.\u00a0<a href=\"#fnref:1\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:2\">\n      <p>There is a chance that two different pieces of content will hash to the same value. But this chance <a href=\"http:\/\/crypto.stackexchange.com\/a\/2584\">is low<\/a>.\u00a0<a href=\"#fnref:2\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:3\">\n      <p><code class=\"highlighter-rouge\">git prune<\/code> deletes all objects that cannot be reached from a ref. If the user runs this command, they may lose content.\u00a0<a href=\"#fnref:3\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:4\">\n      <p><code class=\"highlighter-rouge\">git stash<\/code> stores all the differences between the working copy and the <code class=\"highlighter-rouge\">HEAD<\/code> commit in a safe place. They can be retrieved later.\u00a0<a href=\"#fnref:4\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:5\">\n      <p>The <code class=\"highlighter-rouge\">rebase<\/code> command can be used to add, edit and delete commits in the history.\u00a0<a href=\"#fnref:5\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>","created_at":"2018-11-07T16:51:57+0000","updated_at":"2018-11-07T16:51:57+0000","annotations":[],"mimetype":"text\/html","reading_time":31,"domain_name":"maryrosecook.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":562,"title":"Algorithmic Art - Tyler Hobbs","url":"http:\/\/www.tylerlhobbs.com\/writings\/watercolor","content":"<div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nI've gotten many requests to explain the technique behind my \"watercolor\"\ngenerative artwork.  My approach is similar to some of the techniques I\ndescribed in my post on <a href=\"http:\/\/www.tylerlhobbs.com\/writings\/creating-softness\">Generating Soft Textures<\/a>,\nso read that first if you haven't already.  The algorithm is not particularly exotic.  Instead, it's\nconceptually simple, but fine tuned.\n            <\/p><\/div><\/div><div class=\"row\"><p><\/p><h2>INSPIRATION<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nI frequently experiment with watercolor paints in my sketchbook.  I\nparticularly love the insanely detailed and varied effects that the paints can\nproduce.  One sketch caught my eye, and that's what I focused on while\ndeveloping this technique:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/inspiration.jpg\" alt=\"WATERCOLOR SKETCH\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nMy goal was not to realistically reproduce everything about the watercolor paint.\nInstead, I just wanted to capture the essence of it that I really enjoy.\n            <\/p><\/div><\/div><div class=\"row\"><p><\/p><h2>POLYGON DEFORMATION<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nAt a basic level, the technique consists of stacking lots of nearly\ntransparent layers on top of each other.  To get areas with a gradual fade\nout, those layers need to have a lot of variation in their shape.  However, in\nareas with sharper boundaries, the layers need little variation.\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nThe core of the watercolor technique is a recursive polygon deformation\nalgorithm.  It's pretty simple, and it goes something like this:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><ol><li>For each line <code>A -&gt; C<\/code> in the polygon, find the midpoint, <code>B<\/code>.  From a <a href=\"http:\/\/www.tylerlhobbs.com\/writings\/probability-distributions-for-artists\">Gaussian distribution<\/a>\n    centered on <code>B<\/code>, pick a new point <code>B'<\/code>.<\/li>\n    <li>Update the polygon, replacing the line <code>A -&gt; C<\/code> with two lines: <code>A -&gt; B'<\/code> and\n    <code>B' -&gt; C<\/code><\/li>\n    <li>If we haven't hit our max recursion depth, repeat from step 1, splitting the child lines.<\/li>\n<\/ol><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nDepending on the variation in your Gaussian distribution and the recursion\ndepth, this will produce a polygon with jagged, detailed edges.\n            <\/p><\/div><\/div><div class=\"row\"><p><\/p><h2>USING THE POLYGON DEFORMATION TECHNIQUE<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nFor this example, I'm going to apply the polygon deformation technique to a simple regular polygon:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/simple-poly.png\" alt=\"STARTING POLYGON\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nThe basic watercolor shape is created through one round of polygon\ndeformation.  The input polygon is taken and\nrun through the deformation function multiple times (somewhere around 7).  The\nresulting polygon becomes the \"base polygon\" for all of the layers we will\ncreate.  It looks something like this:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/base-poly.png\" alt=\"BASE POLYGON\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nFor each layer, start with the \"base polygon\" and run it through the\ndeformation function several more times (maybe 4 or 5).  This will produce a\npolygon that is similar to the base polygon, but differs in all of the\nfine details.  Draw the polygon with low opacity (somewhere around 4%).  Repeat\nthis for 30 to 100 layers.  The result will look like this:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/no-variance.png\" alt=\"SIMPLE WATERCOLOR BLOB\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row\"><p><\/p><h2>REFINEMENT: LOW AND HIGH VARIATION AREAS<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nThe steps so far will produce a shape that has somewhat soft edges all around.  In order\nto give the border sharp edges in some areas and soft edges in others, we can\nassign different levels of \"variance\" to each line segment.  Segments with high\nvariance will undergo large changes in each mutation round, and segments with\nlow variance will undergo small changes.  When a segment is split into two child\nsegments, those children can inherit the parent's variance.  Of course, the variance\nneeds to decrease somewhat, and it's also a good idea to slightly randomize what\nvariance each child gets assigned.  With this change in place, the blobs look more\ninteresting:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/variance1.png\" alt=\"WATERCOLOR BLOB WITH SOFTNESS VARIATION\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/variance2.png\" alt=\"WATERCOLOR BLOB WITH SOFTNESS VARIATION\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row\"><p><\/p><h2>REFINEMENT: TEXTURE MASKING<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nNatural watercolor is not perfectly smooth, and has some variation in the opacity.\nTo capture a little of that, you can use a different texture mask on each of the\n30 to 100 layers.  The mask will make some parts of the layer full transparent,\ncreating variations in the final opacity.  For this example, I'm randomly placing\nabout 1000 small circles on the image for the texture mask.\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><div class=\"highlight\"><pre>; draw the watercolor blob shape\n(with-graphics the-blob-mask\n  (background 0 0 0)\n  (stroke 0 0 layer-alpha)\n  (fill 0 0 layer-alpha)\n  (begin-shape)\n  (doseq [[x y] final-poly]\n    (vertex x y))\n  (end-shape))\n; draw the circles onto the texture mask\n(with-graphics the-texture-mask\n  (background 0 0 0)\n  (no-stroke)\n  (fill 0 0 layer-alpha)\n  (doseq [j (range 900)]\n    (let [x (random 0 (w))\n          y (random 0 (h))\n          len (abs-gauss (w 0.03) (w 0.02))\n          [hue sat bright] (color-fn)]\n      (fill hue sat bright)\n      (ellipse x y len len)))\n  ; Blend the watercolor blob shape layer into the current\n  ; layer, only taking the darkest pixel from each. This\n  ; effectively means we're taking the \"intersection\" of\n  ; the two masks.\n  (blend the-blob-mask 0 0 (w) (h) 0 0 (w) (h) :darkest))\n(with-graphics the-overlay\n  ; make the whole background red\n  (background 5 80 80)\n  ; apply the combination blob\/texture mask\n  (mask-image the-texture-mask))\n; apply the masked layer\n(image the-overlay 0 0)\n<\/pre><\/div><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nThe result looks like this:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/texture.png\" alt=\"WATERCOLOR BLOB WITH TEXTURE MASKING\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row\"><p><\/p><h2>REFINEMENT: COLOR BLENDING<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nBlending multiple colors with this technique works well so long as you interleave\nthe layers.  For example, do five layers of a red blob, then five layers of a yellow\nblob, then five red, etc.  This doesn't really replicate what happens with real\nwatercolor paint, but it looks cool anyway.  When combined with the texture masking\nabove, this looks like so:\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/two-colors.png\" alt=\"BLENDING TWO COLORS\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row\"><p><\/p><h2>EXAMPLE WORKS<\/h2><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nHere are a couple of works that I've created with these techniques so far.\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/isohedral-xi.png\" alt=\"ISOHEDRAL XI\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-8 col-md-offset-2\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/watercolor\/linear.png\" alt=\"LINEAR\" class=\"img-responsive\" \/><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \nThat's all for now!  If you found this interesting, go ahead and sign up for my\nnewsletter in the box below to be notified whenever I write about other aspects\nof generative artwork.  Cheers!\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \n<i>Edit: <a href=\"http:\/\/softdroid.net\">Softdroid<\/a> has kindly published a <a href=\"http:\/\/softdroid.net\/generativnyy-podhod-k-modelirovaniyu\">Russian translation of this article<\/a>.\n    <\/i><\/p><p>A <a href=\"http:\/\/www.opensourceinitiative.net\/edu\/watercolor\">Ukrainian translation<\/a> was also published by <a href=\"http:\/\/www.opensourceinitiative.net\/edu\/\">Open Source Initiative<\/a><\/p><\/div><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><p>\n            \n<i>Edit 2: I went more in-depth on this subject at Strange Loop 2017 with my talk titled \"How to Hack a Painting\". You can watch a video of that talk on <a href=\"https:\/\/www.youtube.com\/watch?v=5R9eywArFTE\">youtube<\/a>.<\/i>\n            <\/p><\/div><\/div><div class=\"row top-buffer\"><p>NEXT: <a href=\"http:\/\/www.tylerlhobbs.com\/writings\/triangle-subdivision\">Aesthetically Pleasing Triangle Subdivision<\/a><\/p><\/div><div class=\"row top-buffer\"><p>PREVIOUS: <a href=\"http:\/\/www.tylerlhobbs.com\/writings\/using-seeds\">Utilizing Random Number Generator Seeds<\/a><\/p><\/div><div class=\"row top-buffer\"><div class=\"col-xs-10 col-xs-offset-1 col-md-12 col-md-offset-0\"><div class=\"row top-buffer\"><div class=\"col-xs-8\"><a href=\"http:\/\/instagram.com\/tylerlhobbs\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/Instagram_Icon_Small.png\" alt=\"image\" \/><\/a>\u00a0 \u00a0<a href=\"https:\/\/twitter.com\/tylhobbs\"><img src=\"http:\/\/www.tylerlhobbs.com\/static\/img\/Twitter_logo_blue_small.png\" alt=\"image\" \/><\/a><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:51:57+0000","updated_at":"2018-11-07T16:51:57+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"www.tylerlhobbs.com","preview_picture":"http:\/\/tylerlhobbs.com\/static\/img\/watercolor\/two-colors.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":559,"title":"My Father The Werewolf","url":"https:\/\/folks.pillpack.com\/my-father-the-werewolf\/","content":"<p>The first time I realized my dad wasn\u2019t like other dads, he sat straight up in bed, wide-eyed, and started screaming: \u201cWHO ARE YOU? WHAT ARE YOU DOING HERE?\u201d<\/p><p>I would later become a smart aleck, but at the time, I was three, maybe four, so I didn\u2019t respond to these questions the way I might now: \u201cHey, if anyone should know, it\u2019s you.\u201d<\/p><p>Besides, my father wasn\u2019t joking. As he shook me by the shoulders, his eyes rolled white in his head, like the eyes of a terrified animal.<\/p><div class=\"wp-caption alignright\"><img class=\"alignright wp-image-10878 size-medium\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-300x300.jpg\" alt=\"\" width=\"300\" height=\"300\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-300x300.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-80x80.jpg 80w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-36x36.jpg 36w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-180x180.jpg 180w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-120x120.jpg 120w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-450x450.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes-470x470.jpg 470w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot6_noeyes.jpg 500w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>A moment before, I had been watching <em>Sesame Street<\/em> on the edge of my parents\u2019 bed. My father was napping. He\u2019d been sick for the last couple of days, so he\u2019d stayed home from the office that day. It must have been late, because my mother, who also worked, was home. I think it was spring or summer, because it was still daylight out.<\/p><p>Or maybe it was a weekend in winter. How can you totally trust a 35-year-old memory? All I know is that when I remember that day, it happens in the evening. In the springtime. And my father is still there, still alive, shaking me by my tiny shoulders and yelling.<\/p><p>\u201cWHO ARE YOU? WHAT ARE YOU DOING HERE?\u201d<\/p><p>Downstairs, my mother hears the commotion. She shouts up the stairs, \u201cBruce? What\u2019s wrong?\u201d<\/p><p>The reedy tremolo of terror in her voice gives my father, deranged, another vector than the small, shivering child he was shaking in his hands. (Did he know I was a child, let alone <em>his<\/em> child? Was he that far gone? Another answer I\u2019ll never know.) He erupts from the bed, hurling me into the corner, and by the time I have picked myself up from the floor, the bedroom is empty.<\/p><p>I follow him into the hallway, sniffling. I feel the overwhelming guilt of the toddler, whose heart crushes itself under the solipsism of his newness. Whatever is wrong with my dad, I must be the cause. Yet I have no real idea of what it is I could have done.<\/p><blockquote><div class=\"page\" title=\"Page 1\"><div class=\"section\"><div class=\"layoutArea\"><div class=\"column\"><p>Whatever is wrong with my Dad, I must be the cause. Yet I have no real idea of what it is I could have done.<\/p><\/div><\/div><\/div><\/div><\/blockquote><div class=\"page\" title=\"Page 1\"><div class=\"section\"><div class=\"layoutArea\"><div class=\"column\"><p>He\u2019s in the hallway now. Sun slants between the bannisters, painting dazzle camouflage against the staircase wall. He strides down it, something primal, senselessly screaming. My mother, also screaming, tries to meet him halfway, but with the effortless grace of an acrobat, he throws her over his shoulder. He has almost a foot of height on her; I peer from the top of the staircase as she is carried, kicking her legs, into the shadowed floor below. I don\u2019t remember what she\u2019s wearing, but her hair is a brown perm, laced with the gold of 1983 sunshine. Her face is toward me.<\/p><\/div><\/div><\/div><\/div><p>\u201cGo upstairs John!\u201d she shouts at me, with frantic eyes, as he carries her down. \u201c<em>Run!<\/em>\u201d<\/p><p>I do run. I run back to the bedroom, where I close the door. <em>Sesame Street<\/em> is still playing. Things have happened so quickly that the very same skit my dad and I were watching together when he suddenly threw me across the room\u2014sculptor Ernie plopping an orange nose onto a clay Bert\u2014is still playing.<\/p><p>There, quaking in shock, I feel the floorboards beneath me vibrate as my parents howl and wail. Then, muffled, they both go quiet, and I hear my father sobbing. It curdles at the edges: a sob of utter existential defeat. There\u2019s no power in it at all.<\/p><p>And that\u2019s when I finally start crying too.<\/p><p>My dad spent the night in the hospital, maybe two. It was explained to me he\u2019d had a reaction to a new medication; weaned off it, my dad was back a few days later, good as new. We never spoke about what happened that day. When I was younger, I thought it would shame him; when he was older, he would have no longer remembered any of it.<\/p><p>But that was how I learned that my dad was mentally ill.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10858 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>As I write this, a third of a century later\u2014my father\u2019s ashes carefully locked in a gold-foil vacuum seal bag and placed in an understated black urn on my desk, so I can look at him as I write\u2014I marvel at how easily my father\u2019s severe depression was slotted into that long-ago toddler\u2019s universe. In the span of a nightmare, I effortlessly swallowed up this new and horrible fact that the author of my being was just a flipped neuron away from being an entirely different person, one who forgot who I was, and even attacked me.<\/p><p>But when I think harder about it, I suppose it\u2019s not so surprising.<\/p><p class=\"p1\">Even before \u00a0his illness was revealed to me in such stark relief, he laid the groundwork preparing me to understand.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10859 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><div id=\"attachment_10750\" class=\"wp-caption alignnone\"><img class=\"size-large wp-image-10750\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family-1030x721.jpg\" alt=\"\" width=\"1030\" height=\"721\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family-1030x721.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family-300x210.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family-768x538.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family-705x494.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family-450x315.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/01_Folks_feature_brownlee-family.jpg 1040w\" \/><p class=\"wp-caption-text\"><strong>The Brownlee Family<\/strong> <em>Clockwise from left: Sally, Bruce, and John.<\/em><\/p><\/div><p>According to family legend, the day my parents took me home from the hospital, they got in their first fight about how I should be raised.<\/p><p>Tired, my mom had asked my father to watch me while she took a nap; when she woke up, my dad and I were cuddled up in the living room\u2019s cathode-ray gloam, watching a lycanthropic Oliver Reed bare his fangs on Channel 56\u2019s Creature Double Feature. As someone who has been on the end of them, I can imagine her reprimanding shrieks. But from that day on, monster movies became a lifetime bond between me and my father.<\/p><p>A few years later, I remember cheap vinyl socks crackling as I sleepily ambled downstairs at night in my Dr. Denton\u2019s, drawn to the flickering, fluorescent orange-blue that filled the living room like will-o\u2019-the-wisps. My dad was in there, drinking a beer\u2014he would stop drinking entirely because of medications a few years later\u2014and watching <em>The Howling<\/em>. He picked me up, sat me on his lap, and we watched the last few minutes together.<\/p><p>I remember a lustrously coiffured Dee Wallace as a nightly news anchorwoman transforming into a gossamer blonde werewolf before being shot to death on-air by her executive-producer boyfriend. I wasn\u2019t scared. I was fascinated. But one thing confused me.<\/p><p>As Dad tucked me back into bed, not even bothering to ask me not to tell Mom about this\u2014we were always complicit on the subject of monsters\u2014I remember asking him: \u201cWhy was the monster lady crying at the end?\u201d<\/p><blockquote><div class=\"page\" title=\"Page 1\"><div class=\"section\"><div class=\"layoutArea\"><div class=\"column\"><p>\u201cIt\u2019s because she\u2019s not a monster. She can\u2019t help it. She\u2019s just cursed.\u201d<\/p><\/div><\/div><\/div><\/div><\/blockquote><p>He pursed his lips in dry mock seriousness, nodded sagaciously, and quipped: \u201cShe must have been having a bad hair day!\u201d (My father and I shared a sense of humor that is best described as equal parts Evelyn Waugh, Groucho Marx, and Al Jaffee\u2019s <em>Snappy Answers to Stupid Questions.<\/em>)<\/p><p>But then, I remember, he became thoughtful. He walked to the door, flicked off my bedroom light, and stood in the doorway. Molten around the edges, his far side illuminated, his profile an unknowable eclipse, he said: \u201cIt\u2019s because she\u2019s not a monster. She can\u2019t help it. She\u2019s just cursed. Love you, John.\u201d<\/p><p>And then he shut the door.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10860 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>So while it\u2019s remarkable that I accepted my father\u2019s depression so readily as a toddler\u2014particularly following such a violent mental episode\u2014it\u2019s not so surprising when you realize that I already knew about werewolves. And what was my father if not a werewolf: the slave to the susurrus of primal tides whispering to him in a voice that only he could hear?<\/p><div id=\"attachment_10794\" class=\"wp-caption alignnone\"><img class=\"wp-image-10794 size-large\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-1030x464.jpg\" alt=\"\" width=\"1030\" height=\"464\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-1030x464.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-300x135.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-768x346.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-1500x676.jpg 1500w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-705x318.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-werewolf-450x203.jpg 450w\" \/><p class=\"wp-caption-text\"><em>Panels from <strong>Tales from the Crypt<\/strong> No. 35, a horror comic John and Bruce used to read together.\u00a0(Published with permission, William M Gaines Agent, Inc.)<\/em><\/p><\/div><p>Not that I believed lycanthropy was what literally had happened to my father that night when he shook me by the throat. I didn\u2019t go into pre-school the next day, and, like Calvin, throw open my hands while dramatically intoning the title of an invisible \u201950s movie marquee: \u201cMy Father Is a Werewolf!\u201d I only mean to suggest that I accepted the fact of his personality-altering mental illness with the same innocence and appreciation of irony with which toddlers accept all fairy-tale curses.<\/p><p>Decades later, the werewolf analogy still helps me make sense of my father. Because, as with a werewolf, it is impossible to peel the man and the depression apart when you\u2019re talking about him. They exist forever co-mingled, not in tooth-and-claw but in temperament.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10861 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><div id=\"attachment_10749\" class=\"wp-caption alignright\"><img class=\"wp-image-10749 size-medium\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/02_Folks_feature_bruce-bird-alt-300x258.jpg\" alt=\"\" width=\"300\" height=\"258\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/02_Folks_feature_bruce-bird-alt-300x258.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/02_Folks_feature_bruce-bird-alt.jpg 500w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/02_Folks_feature_bruce-bird-alt-450x387.jpg 450w\" \/><p class=\"wp-caption-text\"><strong>Bruce Brownlee<\/strong><\/p><\/div><p>My father, Bruce Gordon Brownlee, was born on December 30th, 1950, and died just 66 years and one month later.<\/p><p>As the oldest of three boys, his birthday caused him a great deal of anxiety growing up. He was never sure if there was a conspiracy between his parents to keep a few presents back from Santa to give him on his birthday, thus affecting a karmic sort of gift imbalance between him and his brothers. So, for his only son, he overcompensated. My Christmases growing up were maggoty with presents, and from my father\u2014who never gave a gift that he wouldn\u2019t want to get himself\u2014my childhood was filled with formatively corrupting presents: sci-fi paperbacks, bound collections of horror comics with titles like\u00a0<em>Tales from the Crypt <\/em>and\u00a0<em>The Haunt of Fear,<\/em>\u00a0and home-pirated VHS tapes packed with B-monster movies.<\/p><p>Dad spoiled me growing up, but he was generous with everyone, in quiet, understated ways that taught me a lot about what it meant to be giving. I remember, for example, that every time he went out shopping, he came home with a Hershey\u2019s bar and a romance novel for my mother; unable to tell the covers apart, or remember the titles, he\u2019d find ones she hadn\u2019t read by checking the publication date. This is still, for me, the very definition of romance.<\/p><blockquote><p>My father was generous, but he was also depressed, and the nature of depression is to be selfish.<\/p><\/blockquote><p>I think about my father\u2019s generosity a lot. My father was generous, but he was also depressed, and the nature of depression is to be selfish: to starve those who love you of the best of you, in the relentless feeding of that which can never be nourished. In that, he\u2014the most depressed person I ever met\u2014was also the most selfish. For my entire life, he would give me anything I asked for, as long as it was a movie or a book. But when my mother and I begged him half a dozen times to go see a doctor if he loved us, he wouldn\u2019t lift a finger. How do generosity and selfishness co-exist like that in a person without destroying him?<\/p><p>I don\u2019t know. And, of course, it did eventually destroy him. But that was my father: a lycanthrope of contrasts. Whatever he was, he was also the opposite.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10862 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><div class=\"wp-caption alignnone\"><img class=\"wp-image-10748\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/03_Folks_feature_bruce-john-1030x352.jpg\" alt=\"\" width=\"1060\" height=\"376\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>Dad prided himself on his intelligence.<\/p><p>For a long time, I worshipped him for it. He was undeniably a brilliant man. As long as I knew him, he usually had two or three books going at once, ranging from trashy pulp paperbacks to sweeping overviews of post-Enlightenment culture. A shut-in for the last twenty years of his life, he probably read an average of a book a day. And if you were a fellow reader, to come over to his house meant having an armload of books foisted upon you: obscure Russian sci-fi novels, lectures by Nabokov, thick encyclopedias on film theory, or whatever other subjects your brief conversation might have touched on.<\/p><p>My dad\u2019s recommendations were so good that, later in life, when he had been robbed of most of his memories, I would recommend books <em>back<\/em> to him that he\u2019d made me read years before. He\u2019d then brag to his friends about what great taste in books his son had, never knowing I\u2019d grifted him with his own discernment and taste. (I don\u2019t feel guilty. How many of us can say we\u2019ve had the good fortune to read our favorite books for the first time twice?)<\/p><div class=\"wp-caption alignnone\"><img class=\"wp-image-10751\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/05_Folks_feature_bruce-reading-1030x729.jpg\" alt=\"\" width=\"1059\" height=\"754\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>He knew just as much about film. Even more about music, if you can believe it. But none of his intelligence prevented him from being a colossal dumbass about the important things.<\/p><p>Take his health. The man\u2014a three-to-four-pack-a-day smoker for most of his life\u2014honestly believed that the health concerns around cigarettes were overblown. Since 1985, my parents lived on a steep hill, and the only sunlight he got for the last decade of his life was walking down to the corner to buy his Marlboro Lights 100s. In the last months of his life, it turned out that he had experienced light-headedness, numbness, and chest pains multiple times on this twice-weekly nicotine pilgrimage, but he never once told anyone about it. Instead, he would take so much aspirin his nose would bleed, while wrapping a frozen towel around his body like some sort of strange albino ice yogi, ignoring the tell-tale signs of what were proven later to be multiple heart attacks.<\/p><div class=\"wp-caption alignright\"><img class=\"alignright wp-image-10877 size-medium\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-300x300.jpg\" alt=\"\" width=\"300\" height=\"300\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-300x300.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-80x80.jpg 80w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-36x36.jpg 36w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-180x180.jpg 180w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-120x120.jpg 120w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-450x450.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL-470x470.jpg 470w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot2a_FINAL.jpg 500w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>When he eventually had such a massive cardiac event that his heart practically exploded on the spot, my mother told me she thought he\u2019d known he was having heart attacks but chose to ignore them as a way of killing himself. This, I think, is putting an overly heroic sheen on it, but the interpretation doesn\u2019t surprise me. My mother worshipped my father, and there\u2019s something noble about a slow, plausibly deniable suicide. But an intelligent, well-read man ignoring the obvious signs of his own impending heart attack while his loved ones beg him to see a doctor? That\u2019s just the setup to a bad fucking joke.<\/p><p>The truth is, as I knew him, my father in his later years was a man who could intellectualize his way around anything, up to and including his own heart attacks.\u00a0Intelligence, for him, had become not a tool to be applied to the real world, but something that replaced it: the ultimate nicotine patch.<\/p><p>A self-described \u201cJeffersonian liberal\u201d and \u201cBuckley man,\u201d some of these justifications were political. Incapable of earning a living, my father railed bitterly against social welfare programs designed to help people like him, despite the fact that if it were not for my mother, he could not afford to treat his depression. Likewise, my father was a lifelong defender of science and sweat profusely if the mercury went above 68, but when he died, his reading table was stacked with small press treatises denying climate change.<\/p><p>Yet if anything, I found these rationalizations more understandable for having been political. Such intellectual infections are as subjective as they are universal; not so his other justifications, which served no other purpose than to shield him from the quick of life.<\/p><p>In the last years of his life, flesh-and-blood experiences became things no longer worth having. So though, through his library, he\u2019d practically lived in Europe since he was 18, he never expressed interest in coming to visit me during the decade I lived there; there was nothing there, he insinuated, he hadn\u2019t already experienced. Likewise, if Mozart himself had come back to life, my father\u2013the man who took me to see Miles Davis live when I was three\u2013would have expressed no interest in seeing him perform.<\/p><p>Before my father died, I once thoughtlessly told him that I didn\u2019t care if my own kids grew up to be intelligent as long as they grew up to be kind, daring, and hardworking. To that, he didn\u2019t say anything. He just lit a cigarette. Or he tried to. His hand was shaking so badly, he lit his sleeve on fire.<\/p><p>Intellect, you see, was all the self-worth he had left.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10861 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><div class=\"wp-caption alignnone\"><img class=\"wp-image-10747\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/04_Folks_feature_bruce-john-play-1030x568.jpg\" alt=\"\" width=\"1153\" height=\"627\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>Growing up, my dad was my best friend. We were more than inseparable buddies; he was my hero.<\/p><p>When I was just a toddler, and his depression wasn\u2019t as bad as it would one day become, my dad would take me down to Brookline every Sunday to a movie theater called Off-the-Wall. There, he would buy me a brownie, and we would watch old Buster Keaton shorts, while a tweed-decked septuagenarian accompanied the flickering silver screen on an out-of-tune ragtime piano. This is a quintessential \u201cDad\u201d memory for me: How could anyone have a better date with their father?<\/p><p>When I was four or five, we played the same game every night. It went like this: right after my dad read my bedtime story to me, he would grab my favorite bath toy\u2014a rubber rooster with a built-in squeaker that kind of looked like a down-rent Foghorn Leghorn\u2014and jump under the covers with me. There, we would hide, snorting and giggling, until my mom came in to play her part. \u201cWhere\u2019s Bruce and John?\u201d she\u2019d ask, ever the straight man, while the squirming, human-size lumps under the <em>Return of the Jedi <\/em>duvet responded in chorus: \u201cNobody here but us chickens! <em>Cheep, cheep, cheep<\/em>.\u201d<\/p><p>My dad and I loved to draw together. He was particularly great at it, especially before his meds took away his hand coordination. One of his cartoons, called \u201cMy Pal, the Trashcan,\u201d still sits framed on my desk, 33 years after he drew it. It\u2019s essentially a self-portrait. My dad stands in the background with a fedora and an exclamation point above his head. I\u2019m the pot-bellied kid, his belly button peeking out. My pal\u2019s, quite obviously, the trashcan\u2026not to be confused with the mythical toilet from which my father always insisted, with great solemnity, he had plucked me as a baby and for whom I am named: John.<\/p><div id=\"attachment_10754\" class=\"wp-caption alignright\"><img class=\"size-medium wp-image-10754\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-300x300.jpg\" alt=\"\" width=\"300\" height=\"300\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-300x300.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-80x80.jpg 80w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-36x36.jpg 36w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-180x180.jpg 180w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-120x120.jpg 120w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-450x450.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s.-470x470.jpg 470w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/My_pal__the_trash_can._Drawn_for_me_by_my_dad_back_in_the_early_80s..jpg 500w\" \/><p class=\"wp-caption-text\"><strong>My Pal, the Trashcan<\/strong> <em>Bruce Brownlee<\/em><\/p><\/div><p>No one made me laugh harder. When I went to summer camp, he would send me care packages of comics and non sequitur postcards that were so hilarious that my camp counselors would read them aloud to the entire camp over mail call. One, featuring Humphrey Bogart smoking a cigarette, says on the back: \u201cDear John\u2013Going through some old photographs, we came upon this baby picture of you! Boy, you sure were cute. (Note the pacifier.)\u201d Another, which I\u2019ve sadly lost, alleges to feature him indulging in extracurricular activities during a typical day at summer camp in the \u201950s; the picture on front is Indiana Jones punching out a Nazi.<\/p><p>Years before <em>MST3K,<\/em> he and his brother, my Uncle Bob, initiated me to the joys of movie-riffing, ragging on movies like <em>Plan 9 from Outer Space<\/em> and <em>Robot Monster <\/em>until I was practically hyperventilating with laughter. He introduced me to Harvey Kurtzman, Firesign Theater, Jeeves and Wooster, <em>Mad<\/em> <em>Magazine<\/em>\u2026 influences which shape my sense of humor to this day.<\/p><p>We even traveled together. I remember a family trip we once took to Minnesota, where he kept me busy for three days straight by feeding me horror comics he\u2019d squirreled away under the driver\u2019s seat. One summer, we went down to Washington, D.C., to visit Uncle Bob, and he took me to a John Zorn concert, where the avant-garde saxophonist performed a cover of the \u201960s <em>Batman<\/em> theme as if just for me.<\/p><p>These are the years I remember most vividly about my father.<\/p><blockquote class=\"wolf-full-quote\"><p>The further back I go, the more clearly defined he was. It\u2019s when I remember him in the later years of his life that he grows hazier. The memories themselves don\u2019t go fuzzy; he does.<\/p><\/blockquote><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10864 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5a-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>In 1989, my father\u2014bullied at work for his mental illness\u2014quit his job and became, for all intents and purposes, a shut-in.<\/p><p>As a writer, it makes me sad that, for what would end up being the last three decades of his life, there\u2019s almost nothing to <em>describe<\/em> about this period, except to say it was one of slow-motion entropy. During it, he gradually lost abilities most of us wouldn\u2019t even think to call \u201cskills\u201d: To pick up the phone and call someone, go visit a relative, or even just open the mail.<\/p><p>Within just a few years, my father\u2019s entire life circumscribed an irregular quadrangle just a couple thousand square feet in area. One axis was the local gas station, where he bought his cigarettes; another, his bed, where he slept 16 hours a day. A third was his office, where he kept his computer and which\u2014after his death\u2014was so clogged with depression and ash that you would believe he\u2019d been cremated there on the spot. The last axis was the couch, where a Dad-shaped indentation weighs down the springs to this day. It was here he read books, watched movies, and drank coffee by the pot.<\/p><div class=\"wp-caption alignright\"><img class=\"alignright wp-image-10876 size-medium\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-300x300.jpg\" alt=\"\" width=\"300\" height=\"300\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-300x300.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-80x80.jpg 80w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-36x36.jpg 36w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-180x180.jpg 180w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-120x120.jpg 120w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-450x450.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final-470x470.jpg 470w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolfspot1_Final.jpg 500w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>Over time, he started refusing to see his psychiatrist regularly. This didn\u2019t result in him being cut off from his meds: His psychiatrist continued to prescribe them based on my father\u2019s emails and my mother\u2019s reports on how Dad was doing.<\/p><p>He also stopped seeing other doctors. The result was that we no longer had any idea how many of his shakes and tremors were because of his meds and which were because of his undiagnosed illnesses.<\/p><p>Likewise, he wouldn\u2019t see dentists, so he started losing his teeth. His resulting self-consciousness formed a closed feedback loop with his depression, dead-bolting him in the house. His hair grew long and lanky; he became so pale as to be nearly translucent. At 65, almost nocturnal, he had the look of a toothless white wolf.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10866 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-6-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>Every werewolf story is about a silver bullet.<\/p><p>This was true for my father. But the silver bullet he believed would one day cure him wasn\u2019t literal. It was some magic drug that would one day be fired out of the rotating barrels of a pharmaceutical company\u2019s R&amp;D revolver.<\/p><p>If you suggested that he go for a walk or a run, or go with you to a movie, or volunteer somewhere, or see a therapist, Dad would openly scoff: He\u2019d tried all that, and the only thing that ever worked against his depression was drugs.<\/p><blockquote><p>\u201cNo one really understands why the brain works the way it does.\u201d<\/p><\/blockquote><p>In his youth, he\u2019d apparently abused alcohol and experimented with psychedelics; it was to his great credit that, once I was born, he\u2019d mostly given these up for my mother. \u201cBooze was the only thing that ever really made me happy,\u201d he once told me, wistfully, through the literal rose-colored sunglasses he wore whenever he left the house, even on the cloudiest days.<\/p><p>But until I was about 27 or so, he still held out hope that the next medication he was prescribed would be the one that would finally obliterate his depression.<\/p><p>\u201cNo one really understands why the brain works the way it does,\u201d he would often say, like a priest reciting a holy mystery. According to this logic, then, it was just a roll of the dice before the pharmacologists stumbled upon the mystical alignment of molecules that would make his brain work like it should. Until that moment arrived, there was little point in doing anything but wait; happiness could not be found in anything but chemistry.<\/p><p>Yet even from an early age, I was doubtful my father would recognize the silver bullet he was looking for if and when it ever came for him. If you asked him on any given day how he was feeling, he\u2019d reply: \u201cIt\u2019s the worst day of my life,\u201d no matter whether he was comatose with depression or talking to you on your wedding day. It was almost like my dad had lost his inner compass to tell you how he was feeling: even if he acted like he was feeling better, he\u2019d tell you he was feeling worse.<\/p><p>I once asked my dad what he thought happiness was; he told me it was a gin and tonic. He wasn\u2019t joking. The only way my father could imagine the experience of not being depressed anymore was to equate it with being euphorically drunk or high. But this isn\u2019t what wellness is. Neurotypical humans still feel sad, and hopeless, and anxious, and overwhelmed. It isn\u2019t that they are happy all the time: it\u2019s that they\u2019re capable of happiness in the first place. I realized my dad didn\u2019t understand this, and because he didn\u2019t understand it, all his silver bullets would inevitably wear away to show the brass beneath.<\/p><p>No wonder the medications he thought were \u201cworking\u201d made him manic, and inevitably led to deep crashes, which just made him ever more desperate.<\/p><p>No wonder the silver bullet he eventually turned to came from a can.<\/p><div id=\"attachment_10793\" class=\"wp-caption alignnone\"><img class=\"wp-image-10793 size-large\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-1030x480.jpg\" alt=\"\" width=\"1030\" height=\"480\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-1030x480.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-300x140.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-768x358.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-1500x699.jpg 1500w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-705x329.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-silver-bullet-450x210.jpg 450w\" \/><p class=\"wp-caption-text\"><strong>Tales From The Crypt<\/strong> <em>No. 35\u00a0(Published with permission, William M Gaines Agent, Inc.)<\/em><\/p><\/div><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10858 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-7-first-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>One morning, shortly before dawn, my mother awoke from a nightmare very early to find herself sleeping alone. This wasn\u2019t unusual\u2014my dad was a night owl\u2014but when she went downstairs, she found him in the living room, drinking a six-pack by himself.<\/p><p>It was a bigger deal than it sounds. Drinking while taking the sort of meds he was on could trigger another violent episode like the one that had happened so many years before. Twenty-odd years earlier, in fact, my mom threatened to leave Dad if he ever drank again. From then on, he never drank a drop.<\/p><p>But he was drinking now. It was a sign of how truly desperate things had become.<\/p><p>My mother didn\u2019t leave him. How could she? She worshipped him. But as a last-ditch effort to cure his depression, they went to his psychiatrist and told him\u2014over the doctor\u2019s objections, surprisingly\u2014that they wanted to try something extreme: electro-convulsive treatment, or ECT.<\/p><p>In other words, electroshock.<\/p><blockquote><div class=\"page\" title=\"Page 1\"><div class=\"section\"><div class=\"layoutArea\"><div class=\"column\"><p>In other words, electroshock.<\/p><\/div><\/div><\/div><\/div><\/blockquote><p>I was 26 or 27 and living abroad when my mother called to tell me about this.<\/p><p>\u201cNo one really knows why it works, but it does,\u201d my mother insisted, blindly repeating one of my father\u2019s holy mysteries into the telephone mouthpiece. \u201cIt\u2019s not like in <em>One Flew Over the Cuckoo\u2019s Nest. <\/em>Really, the stigma of this treatment has really prevented a lot of people from being helped!\u201d<\/p><p>Stunned, I said nothing.<\/p><p>More to fill the silence than anything, my mom told me all the possible side effects: confusion, disorientation, memory loss. But those side-effects would be temporary, and if it worked, my dad\u2019s depression might be cured.<\/p><p>When she was finally talked out, I stammered:\u00a0\u201cWhen do you think\u2026 when would this start?\u201d<\/p><p>A long pause.<\/p><p>It had already started.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10859 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-8-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>Initially, I followed the lead of my father\u2019s black humor and treated his ECT like a joke.<\/p><div class=\"wp-caption alignright\"><img class=\"alignright wp-image-10875 size-medium\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-300x300.jpg\" alt=\"\" width=\"300\" height=\"300\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-300x300.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-80x80.jpg 80w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-36x36.jpg 36w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-180x180.jpg 180w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-120x120.jpg 120w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-450x450.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4-470x470.jpg 470w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_spot4.jpg 500w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>From my mother, I heard that my dad had ironically started wearing a T-shirt with Daffy Duck dressed as Napoleon to his ECT sessions. Delighted, I completed his outfit by mailing him a crushed velvet bicorne, replete with a large <em>N<\/em> stitched in French piping across the front. Eventually, I heard from my mother\u2014my dad was always too tired to talk on the phone during this period\u2014that this ensemble was eventually banned from the ECT ward, not because it was insensitive, but because it was making Dad\u2019s fellow depressives laugh too much.<\/p><p>It was only when I came home for Christmas that I realized exactly how invasive the treatment really was.<\/p><p>It was a few months after his treatments began, and he\u2019d had a session before my plane arrived.<\/p><p>\u201cLook, Bruce, it\u2019s John!\u201d my mother said, with a not-quite-convincing air of nonchalance when he came slumping down the stairs as we arrived home from the airport.<\/p><p>My dad covered up his confusion, but I could tell he didn\u2019t recognize me at all.<\/p><p>And suddenly, ECT didn\u2019t seem so funny.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10860 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-1-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>In the end, my father had about 18 ECT sessions over six months. \u201cFar, far too many,\u201d my mom said gravely when I asked her to confirm that number ahead of writing this. (Patients usually receive six to twelve sessions.) \u201cWe should never have kept going.\u201d<\/p><p>Looking back, it\u2019s hard not to agree. But for a while, the changes seemed miraculous.<\/p><p>Yes, my dad\u2019s memory was shot. When it came to me, it was as if he\u2019d experienced a time jump. He remembered me up until my tenth birthday or so, after which, I suddenly aged 17 years overnight.<\/p><p>Those memories never came back; to reminisce with him about something, I\u2019d have to tell him the memory first. He simply had no recollection of how I had grown into the man I had become, and for the rest of his life, viewed me with a combination of love and awe, as if he wondered: How can this normal, functional person be\u00a0<em>my<\/em> son?<\/p><div class=\"wp-caption alignnone\"><img class=\"wp-image-10796\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/06_Folks_feature_bruce-john-staring.jpg\" alt=\"\" width=\"1185\" height=\"834\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>But it had its advantages. For example, I no longer had to worry about whether or not I was buying books he\u2019d already read for Christmas or his birthday. We got to watch <em>Robot Monster <\/em>together and laugh at the Billion Bubble Machine all over again, as if for the first time. And it blotted from my dad\u2019s memory some of the more embarrassing or shitty things he\u2019d caught me doing over the years: when I\u2019d stolen money from his wallet as a teenager, for example, or the time he\u2019d caught me watching some vintage \u201970s porn when I was 15.<\/p><p>I also enjoyed my father\u2019s utter incredulity about some of the more peculiar specifics of our shared history. For example, when I was 19, my dad once woke me up in the middle of the night in a panic. He had somehow infected my mother\u2019s work computer with a virus that had converted the screen of her Windows 95 desktop into what can only be described as a pornographic amalgam of pulsating, 16-bit genitalia. \u201cI don\u2019t know what I did, but you gotta help me get it off before your mom wakes up, John!\u201d he hissed to me through the door.<\/p><p>My dad was in stitches when I told him this; he just couldn\u2019t believe it had happened.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10861 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-2-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>The reason my mother and I were able to overlook my dad\u2019s memory loss was because he suddenly wanted to do things again.<\/p><p>In the middle of his treatments, Mom and Dad took a trip up to Wells Beach in Maine, where I was conceived. There, they walked the beach and talked about their future\u2026something they hadn\u2019t done for years, because depression <em>has<\/em> no future.<\/p><blockquote><p>They talked about their future\u2026 something they hadn\u2019t done for years, because depression has no future.<\/p><\/blockquote><p>And Mom wasn\u2019t the only one who benefited. On my second visit home after his ECT, my dad asked me to take a walk with him, seemingly for the hell of it. We walked for about a half hour, talking about movies. To this day, although I can barely remember where we walked or what we discussed, the emotional memory of that walk makes something swell painfully in my throat.<\/p><p>The truth was that I could live with a father who didn\u2019t remember where I came from, as long as we had a future together.<\/p><p>Right before I moved to Ireland, I remember coming home one night and finding my dad in the living room, sitting quietly. I thought I\u2019d join him, but as I approached the couch, his eyes barely flickered at me before resting back on the silent TV. I must have thought he was engrossed in a movie\u2014maybe one of those old silent movies we used to watch together\u2014but when I sat down, I saw the television was off.<\/p><p>\u201cWatching a \u2018Zero-D\u2019 movie, huh, Dad?\u201d I asked him, recycling one of his favorite jokes\u2013a play on 3-D movies\u2013from when I would hide my eyes while watching scary movies as a kid.<\/p><p>He didn\u2019t laugh. He didn\u2019t smile. His eyes didn\u2019t shift. He just kept smoking, silently. And when he was done, he stubbed out the butt and lit another one, just as quietly.<\/p><p>We sat there for some time, at midnight, watching nothing together. Then I kissed him on the forehead and went to bed, knowing in my heart that the next time I flew back to the States, it might be because he\u2019d spilt his veins into the sink.<\/p><p>Compared to the fatalism of my father\u2019s impending suicide, ECT seemed like a godsend.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10862 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-3-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>My father was the one person who didn\u2019t think the ECT had worked. But we didn\u2019t take him seriously.<\/p><p>\u201cI still hate myself,\u201d he once told me. \u201cJust now, I can\u2019t remember why.\u201d<\/p><p>We didn\u2019t really listen. True, my father wasn\u2019t really a reliable narrator of his own experience: He was the man who called wolf because there was always one hiding right behind his eyes. But that\u2019s not why we didn\u2019t listen. We ignored him because his memory loss seemed inconsequential compared to the fact that after years of neglecting our needs and wants, he was finally giving us new memories to remember him by.<\/p><p>But my dad was right. A year later, maybe two, his depression was worse than ever, except now, he was only a shell of who he once was. For a while, yes, he was disoriented enough to be more compliant with us, to go with the flow. But that compliance was never happiness.<\/p><p>ECT didn\u2019t work, at least for him. It didn\u2019t cure his depression, any more than you can cure a werewolf by ripping out his fangs. For my father, the books he\u2019d read, the movies he\u2019d seen, the albums he\u2019d listened to were his eyeteeth, and by taking them out, all we\u2019d done was succeed in making him toothless. In his compliance, yes, we mistook him for tamer, and in his tameness, we mistook him for being\u2014if not happy\u2014then well. But ECT had robbed him of his sense of self.<\/p><p>He didn\u2019t die immediately after that. In fact, he lived another ten years. But he never tried to get better again.<\/p><div id=\"attachment_10753\" class=\"wp-caption alignnone\"><img class=\"size-large wp-image-10753\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-1030x686.jpg\" alt=\"\" width=\"1030\" height=\"686\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-1030x686.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-300x200.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-768x512.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-1500x1000.jpg 1500w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-705x470.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/07_Folks_feature_john_wedding-450x300.jpg 450w\" \/><p class=\"wp-caption-text\"><strong>John\u2019s Wedding<\/strong><em>: The author\u2019s wife and her parents, then from left to right, the author, Sally, and Bruce<\/em><\/p><\/div><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10863 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-4-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><p>When I was a teenager, I felt like I was the first person whose soul was ever scraped raw by the world. When I met others who felt the same way, we bonded to each other like barnacles. Although I don\u2019t have that much in common with them anymore, these teenage friends are still the ones closest to my heart. It\u2019s probably the same for you.<\/p><p>When my father was a teenager, he had a wickedly clever friend. His name was John.<\/p><p>My father and John found companionship in their love of pulp fiction, weird movies, jazz, rock, and psychedelic drugs. They also connected over their mutual battle with depression.<\/p><p>Then, when he was 24, John murdered his parents. He stabbed his mother to death with a butcher\u2019s knife and killed his father while he was watching TV, smashing the elder\u2019s skull in from behind with a single blow from a sledgehammer.<\/p><p>For his crime, John was institutionalized for life. After that, my dad never seemed quite comfortable making close male friends anymore.<\/p><p>I often wonder about this friendship. What was it like for my father?<\/p><p>For a few years, you have this friend, who loves all the things that you do and shares all the things you\u2019re going through. For the first time, you feel <em>understood<\/em>. But then, this friend, this doppelg\u00e4nger, turns out to be a maniac and has to be locked up.<\/p><p>Do you come to believe the inevitable endpoint of your depression is insanity, then being locked away forever?<\/p><p>If so, perhaps it was a foregone conclusion that once my father attacked me and my mother, he would believe that he should be locked away. And if the outside world wasn\u2019t going to do it, his subconscious would.<\/p><p>After all, every werewolf story also ends up with self-imprisonment\u2014ostensibly to protect others, but really so the werewolf can protect itself from the wounds of the world.<\/p><div class=\"wp-caption alignnone moon-divider\"><img class=\"aligncenter wp-image-10865 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b.jpg\" alt=\"\" width=\"1040\" height=\"42\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b.jpg 1040w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b-300x12.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b-768x31.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b-1030x42.jpg 1030w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b-705x28.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/Moon-5b-450x18.jpg 450w\" \/><p class=\"wp-caption-text\"><em>\u00a0<\/em><\/p><\/div><div id=\"attachment_10755\" class=\"wp-caption alignright\"><img class=\"wp-image-10755 size-medium\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/tales-from-the-crypt-cover-werewolf-211x300.jpg\" alt=\"\" width=\"211\" height=\"300\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/tales-from-the-crypt-cover-werewolf-211x300.jpg 211w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/tales-from-the-crypt-cover-werewolf-496x705.jpg 496w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/tales-from-the-crypt-cover-werewolf-450x639.jpg 450w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/08\/tales-from-the-crypt-cover-werewolf.jpg 600w\" \/><p class=\"wp-caption-text\"><strong>Tales From The Crypt<\/strong> <em>Issue No.\u00a046\u00a0(Published with permission, William M Gaines Agent, Inc.)<\/em><\/p><\/div><p>There\u2019s a theme in literature, closely related in its own way to the duality of the werewolf myth, that I know resonated with my father.<\/p><p>\u201cWhat makes Hornblower a hero is that he secretly believes he\u2019s a coward,\u201d he once told me when crushing one of his favorite C. S. Forester novels into my hand. \u201cIt\u2019s what makes him act, but it\u2019s also what torments him: the insistent belief that if another man were in his place, he would have acted without the same fears and self-doubts.\u201d<\/p><p>What appealed to him in the adventures of Horatio Hornblower was the enormous divide\u2014present even in great men\u2014between our perception of ourselves and the qualities of the person we want to be. It is, I think, a divide that every self-aware adult can identify with; only sociopaths feel there is no divide within themselves at all.<\/p><p>The divide never goes away, but those of us who are able to achieve contentedness in our lives are able to bridge this gap over time, establishing a link of understanding between the reality of our condition and our own expectations of ourselves. A life well-lived is one that adds new boards and nails to that bridge every day.<\/p><blockquote class=\"wolf-full-quote\"><p>As for depression, it\u2019s the chasm that exists between. You build your bridge, don\u2019t look down, and pray it never swallows you, because if it does, you\u2019ll fall forever in that bottomless gulf, and die without ever landing.<\/p><\/blockquote><p>That was my dad. The bridge he tried to build for himself earlier in his life inevitably broke underneath him. And in the gulf into which he fell, he fell forever, stranded within arm\u2019s reach of both the person he was and the person he most wanted to be. We, those who loved him, wanted him to climb out, but the gulf was too wide. Once he fell, the closest we ever saw him come was when his fingers scraped the sides.<\/p><p>My father was my hero, but he was also the person I spent my whole life trying as hard as I knew how not to be. It\u2019s filled me with complicated feelings about him\u2014feelings I don\u2019t know I will ever entirely know how to resolve.<\/p><p>So if, through depression, he was what I\u2019ve called him\u2014a lycanthrope of contrasts\u2014then I\u2019m a werewolf too.<\/p><p>When I remember him, the full moon will ever rise upon my thoughts.\u00a0<img class=\"alignnone wp-image-10795\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/favicon-01-80x80.png\" alt=\"\" width=\"16\" height=\"16\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/favicon-01-80x80.png 80w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/favicon-01-36x36.png 36w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/favicon-01-180x180.png 180w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/favicon-01-120x120.png 120w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/favicon-01.png 256w\" \/><\/p><div id=\"attachment_10792\" class=\"wp-caption alignnone\"><img class=\"wp-image-10792 size-full\" src=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-cover-no39-1.jpg\" alt=\"\" width=\"1000\" height=\"655\" srcset=\"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-cover-no39-1.jpg 1000w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-cover-no39-1-300x197.jpg 300w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-cover-no39-1-768x503.jpg 768w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-cover-no39-1-705x462.jpg 705w, https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/tales-from-the-crypt-cover-no39-1-450x295.jpg 450w\" \/><p class=\"wp-caption-text\"><strong>Tales From The Crypt<\/strong> <em>No. 39 \u00a0(Published with permission, William M Gaines Agent, Inc.)<\/em><\/p><\/div><hr \/><p class=\"artist-credit\"><em><strong>Cover &amp; Illustrations<\/strong> by Skip Sterling<\/em><\/p><div class=\"essb_break_scroll\"><\/div>","created_at":"2018-11-07T16:51:56+0000","updated_at":"2018-11-07T16:51:56+0000","published_at":"2017-09-11T10:03:52+0000","published_by":["John Brownlee"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":33,"domain_name":"folks.pillpack.com","preview_picture":"https:\/\/folks.pillpack.com\/wp-content\/uploads\/2017\/09\/wolf_open_final_mobile.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":558,"title":"Cryptoeconomics 101","url":"https:\/\/thecontrol.co\/cryptoeconomics-101-e5c883e9a8ff?gi=87a8d9f0ab71","content":"<p id=\"e545\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--h3\">Much has been discussed about blockchain-based tokens lately. We hosted the inaugural <a href=\"http:\/\/tokensummit.com\/\" data-href=\"http:\/\/tokensummit.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Token Summit<\/a> in New York on May 25th. Prices of tokens have soared, <a href=\"https:\/\/techcrunch.com\/2017\/06\/01\/brave-ico-35-million-30-seconds-brendan-eich\/\" data-href=\"https:\/\/techcrunch.com\/2017\/06\/01\/brave-ico-35-million-30-seconds-brendan-eich\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">startups are raising millions in token sales within seconds<\/a>, mainstream media outlets like <a href=\"https:\/\/www.bloomberg.com\/view\/articles\/2017-06-01\/this-blockchain-bubble-might-be-a-good-thing\" data-href=\"https:\/\/www.bloomberg.com\/view\/articles\/2017-06-01\/this-blockchain-bubble-might-be-a-good-thing\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Bloomberg<\/a>, <a href=\"https:\/\/www.forbes.com\/sites\/laurashin\/2017\/05\/26\/bubble-so-what-token-summit-marks-cryptocurrencys-revitalization\/#440f965517b5\" data-href=\"https:\/\/www.forbes.com\/sites\/laurashin\/2017\/05\/26\/bubble-so-what-token-summit-marks-cryptocurrencys-revitalization\/#440f965517b5\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Forbes<\/a> and <a href=\"http:\/\/www.cnbc.com\/2017\/05\/25\/bitcoin-nerds-give-way-to-wall-street-suits-at-digital-currency-conference.html\" data-href=\"http:\/\/www.cnbc.com\/2017\/05\/25\/bitcoin-nerds-give-way-to-wall-street-suits-at-digital-currency-conference.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">CNBC <\/a>are talking about token models and every venture fund in the world is now discussing what to make of tokens internally.<\/p><p id=\"b8d8\" class=\"graf graf--p graf-after--p\">In recent discussions with people newly interested in the burgeoning token economy, one thing I\u2019ve noticed is that <strong class=\"markup--strong markup--p-strong\">the concept of cryptoeconomics is largely foreign to people. <\/strong>Cryptoeconomics is the fundamental catalyst for this whole movement, so I think that needs to change.<\/p><blockquote id=\"8019\" class=\"graf graf--pullquote graf-after--p\"><div><strong class=\"markup--strong markup--pullquote-strong\">The goal of this piece is to shine a light on cryptoeconomics and encourage more people to start thinking about it. If you know a young (or old) person interested in math, science, or economics, please share with\u00a0them!<\/strong><\/div><\/blockquote><h4 id=\"cda3\" class=\"graf graf--h4 graf-after--pullquote\">What is cryptoeconomics?<\/h4><p id=\"aa66\" class=\"graf graf--p graf-after--h4\">Cryptoeconomics is the study of economic interaction in adversarial environments. In decentralized P2P systems that do not give control to any third party, one must assume that there will be bad actors looking to disrupt the system. <strong class=\"markup--strong markup--p-strong\">Cryptoeconomic approaches combine cryptography and economics to create robust decentralized P2P networks that thrive over time despite adversaries attempting to disrupt the network.<\/strong> The cryptography underlying these systems is what makes the P2P communication within the networks secure and the economics is what incentivizes all actors to contribute to the network so that it thrives over time.<\/p><figure id=\"5e33\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*EvtW6huoQSgehQCy0KT4_g.png\" data-width=\"402\" data-height=\"611\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*EvtW6huoQSgehQCy0KT4_g.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Cryptoeconomics is about creating secure P2P networks that thrive over\u00a0time.<\/figcaption><\/figure><p id=\"7260\" class=\"graf graf--p graf-after--figure\">Satoshi Nakamoto birthed the field of cryptoeconomics when he created Bitcoin in 2009. Just like Galileo is known as the founding father of physics, Satoshi will forever be known as the founding father of cryptoeconomics. I believe in 50 years, cryptoeconomics may be a discipline as widely studied as physics and Satoshi may be as widely revered as Galileo.<\/p><h4 id=\"97a4\" class=\"graf graf--h4 graf-after--p\">Some history on decentralized systems<\/h4><p id=\"e660\" class=\"graf graf--p graf-after--h4\">Decentralized P2P systems based on cryptography were not new in 2009 (you probably heard of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Kazaa\" data-href=\"https:\/\/en.wikipedia.org\/wiki\/Kazaa\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Kazaa<\/a> and <a href=\"http:\/\/www.bittorrent.com\/\" data-href=\"http:\/\/www.bittorrent.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Bittorrent<\/a> prior). What these earlier decentralized systems lacked was economic incentives, and the lack of baked in economic incentives is arguably what stifled these early P2P systems from persisting and thriving over time.<\/p><p id=\"2a96\" class=\"graf graf--p graf-after--p\">Satoshi added economic incentives to P2P systems when he created Bitcoin in 2009. It was actually previously believed to be impossible to achieve consensus among nodes (the <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/publication\/byzantine-generals-problem\/\" data-href=\"https:\/\/www.microsoft.com\/en-us\/research\/publication\/byzantine-generals-problem\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Byzantine General\u2019s Problem<\/a>) to create a decentralized digital cash system, but Satoshi\u2019s implementation of a proof-of-work consensus mechanism + incentives solved this previously unsolvable problem (see <a href=\"http:\/\/www.mail-archive.com\/cryptography@metzdowd.com\/msg09997.html\" data-href=\"http:\/\/www.mail-archive.com\/cryptography@metzdowd.com\/msg09997.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">here<\/a> for his own explanation). Satoshi\u2019s combination of crypto and incentives resulted in a robust, thriving p2p payment network that today stores over $40B worth of value and processes over $600M worth of transactions daily.<\/p><h4 id=\"36b2\" class=\"graf graf--h4 graf-after--p\">The forefront of cryptoeconomics today<\/h4><p id=\"0d22\" class=\"graf graf--p graf-after--h4\">If Satoshi is the Galileo of cryptoeconomics, <a href=\"https:\/\/twitter.com\/VitalikButerin\" data-href=\"https:\/\/twitter.com\/VitalikButerin\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Vitalik<\/a> may be the Einstein. To understand Vitalik\u2019s thinking about cryptoeconomics, start by watching his intro presentation:<\/p><figure id=\"1b55\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"a346\" class=\"graf graf--p graf-after--figure\">From a cryptoeconomic perspective, Vitalik has designed Ethereum to be different from Bitcoin in a few key ways:<\/p><ul class=\"postList\"><li id=\"1102\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">More expressive scripting language:<\/strong> Satoshi built Bitcoin with a simple scripting language that is not designed to facilitate complex transaction types. Ethereum is built from the ground up to be a general purpose, decentralized computing platform and the team implemented some clever techniques like a flexible fee structure for running more complex contracts (aka <a href=\"https:\/\/ethereum.stackexchange.com\/questions\/3\/what-is-meant-by-the-term-gas\" data-href=\"https:\/\/ethereum.stackexchange.com\/questions\/3\/what-is-meant-by-the-term-gas\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">gas<\/a>).<\/li><li id=\"ca5f\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Proof-of-stake consensus: <\/strong>This is where things get interesting from<strong class=\"markup--strong markup--li-strong\"> <\/strong>a cryptoeconomic perspective. Proof-of-work in Bitcoin is inefficient\u200a\u2014\u200acosting miners hundreds of millions of dollars per year in the hardware and electricity required to secure the network. Proof-of-stake is a consensus mechanism that seeks to solve the inefficiency problem by allowing the stake holders in the token to validate the network. This approach has historically been thought of as insecure (the <a href=\"https:\/\/ethereum.stackexchange.com\/questions\/2402\/what-exactly-is-the-nothing-at-stake-problem\" data-href=\"https:\/\/ethereum.stackexchange.com\/questions\/2402\/what-exactly-is-the-nothing-at-stake-problem\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">nothing at stake problem<\/a> and the <a href=\"https:\/\/blog.ethereum.org\/2014\/05\/15\/long-range-attacks-the-serious-problem-with-adaptive-proof-of-work\/\" data-href=\"https:\/\/blog.ethereum.org\/2014\/05\/15\/long-range-attacks-the-serious-problem-with-adaptive-proof-of-work\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">long-range attack<\/a> have been two of the major unsolved vulnerabilities), but Vitalik and team have been diligently working on solutions to these problems. Ethereum now runs on proof-of-work, but the expectation is to switch to proof-of-stake in the next 12\u201318 months.<\/li><li id=\"746c\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Economics: <\/strong>Bitcoin has a clear monetary policy: 21M BTC will ever be created and the rate at which BTC is created decreases by half every four years. Miners running hardware earn tokens for expending electricity. Ethereum has a more flexible monetary policy that is not yet set in stone at this time. Many cite the fixed supply of BTC and the nebulous supply of ETH as a negative; we\u2019ll see. What is clear is that when the switch to PoS occurs, holders of coins will act as the validators of the network and earn more coins. This is a more inclusive approach to coin distribution than Bitcoin\u2019s current ASIC mining approach.<\/li><\/ul><p id=\"0da7\" class=\"graf graf--p graf-after--li\">Ethereum is such an exciting project to me because Vitalik and his team don\u2019t come from academia or existing industry and they have a new and fresh perspective on building cryptoeconomic systems. Their approach flies directly against what a lot of the traditional academic and crypto communities have thought to be possible (the popular belief is that it\u2019s impossible to build a secure decentralized platform that has a Turing complete scripting language and uses proof-of-stake as the consensus mechanism).<\/p><h4 id=\"4e89\" class=\"graf graf--h4 graf-after--p\">Learn more<\/h4><p id=\"33cd\" class=\"graf graf--p graf-after--h4\">Diversity of perspectives is incredibly important to build robust systems that persist over time, and the hope is that the Ethereum team\u2019s work inspires new minds to enter the field and start solving problems in new ways just like they did. There are many projects beyond Bitcoin and Ethereum that are interesting from a cryptoeconomic perspective as well, including <a href=\"http:\/\/sia.tech\/\" data-href=\"http:\/\/sia.tech\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Sia<\/a>, <a href=\"https:\/\/augur.net\/\" data-href=\"https:\/\/augur.net\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Augur<\/a>, and <a href=\"https:\/\/z.cash\/\" data-href=\"https:\/\/z.cash\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Zcash<\/a>. I expect the number of cryptoeconomic projects to proliferate in the coming years.<\/p><p id=\"8854\" class=\"graf graf--p graf-after--p\">Today, there are probably less than 100 people in the world who are experts in cryptoeconomics. I don\u2019t think there\u2019s a better field for a young person to be studying right now and you don\u2019t need formal education for it either\u200a\u2014\u200aall the tools are at your disposal now, for free (<a href=\"https:\/\/www.reddit.com\/\" data-href=\"https:\/\/www.reddit.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Reddit<\/a>, <a href=\"https:\/\/slack.com\/\" data-href=\"https:\/\/slack.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Slack<\/a>, <a href=\"https:\/\/twitter.com\/\" data-href=\"https:\/\/twitter.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Twitter<\/a>, <a href=\"https:\/\/github.com\/\" data-href=\"https:\/\/github.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Github<\/a>, <a href=\"https:\/\/www.coursera.org\/\" data-href=\"https:\/\/www.coursera.org\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Coursera<\/a> &amp; <a href=\"https:\/\/www.youtube.com\/\" data-href=\"https:\/\/www.youtube.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Youtube<\/a> among others).<\/p><p id=\"0ee8\" class=\"graf graf--p graf-after--p\">If interested in learning more, start by checking out the following:<\/p><ul class=\"postList\"><li id=\"cdc7\" class=\"graf graf--li graf-after--p\"><a href=\"https:\/\/www.coursera.org\/learn\/crypto\" data-href=\"https:\/\/www.coursera.org\/learn\/crypto\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">Cryptography I by Dan Boneh, Stanford<\/a><\/li><li id=\"b366\" class=\"graf graf--li graf-after--li\"><a href=\"http:\/\/slack.1protocol.com\/\" data-href=\"http:\/\/slack.1protocol.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">1protocol Slack<\/a><\/li><li id=\"e444\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.reddit.com\/r\/ethereum\/\" data-href=\"https:\/\/www.reddit.com\/r\/ethereum\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">R\/ethereum subreddit<\/a><\/li><li id=\"efbc\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/www.youtube.com\/watch?v=bSdwqa3Yl0Q\" data-href=\"https:\/\/www.youtube.com\/watch?v=bSdwqa3Yl0Q\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">The Mauve Revolution, by Vitalik Buterin<\/a> (and e<a href=\"http:\/\/vitalik.ca\/\" data-href=\"http:\/\/vitalik.ca\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">verything else that Vitalik writes<\/a>)<\/li><li id=\"3301\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/blog.ethereum.org\/2015\/08\/01\/introducing-casper-friendly-ghost\/\" data-href=\"https:\/\/blog.ethereum.org\/2015\/08\/01\/introducing-casper-friendly-ghost\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">Introducing Casper the Friendly Ghost, by Vlad Zamfir<\/a><\/li><li id=\"c293\" class=\"graf graf--li graf-after--li graf--trailing\">Everything on <a href=\"http:\/\/unenumerated.blogspot.com\/\" data-href=\"http:\/\/unenumerated.blogspot.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">Nick Szabo\u2019s blog Unenumerated<\/a><\/li><\/ul><\/figure>","created_at":"2018-11-07T16:51:54+0000","updated_at":"2018-11-07T16:51:54+0000","published_at":"2017-06-04T18:00:42+0000","published_by":["Nick Tomaino"],"annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"thecontrol.co","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*EvtW6huoQSgehQCy0KT4_g.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":555,"title":"The CTO Journey at a Small Startup","url":"https:\/\/zapier.com\/engineering\/startup-cto\/","content":"<p>As startups grow, we need to make tweaks to the way we work. I\u2019ve found this especially true in engineering. As a co-founder and CTO, my own role has changed a lot over the years. My everyday duties and challenges have shifted, and I\u2019ve had to alter my approach multiple times to help the company reach a new level. <\/p><p>In the early days of Zapier, it was just three guys working nights and weekends, hacking away in the basement on what would become Zapier. We went through three or four versions of Zapier before we came to the one we eventually launched. Six years later we're doing hundreds of millions of API calls and have more than 80+ employees around the entire world, including more than two dozen engineers.<\/p><iframe src=\"https:\/\/www.youtube.com\/embed\/t5Y-kHhT7Qc\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>The growth stage between just the three of us and where we are today was pretty tricky. If you prefer, you can <a href=\"https:\/\/www.youtube.com\/watch?v=t5Y-kHhT7Qc&amp;utm_source=zapier.com&amp;utm_medium=referral&amp;utm_campaign=zapier\">watch my takeaways<\/a> above. Or, read on for the lessons I've learned\u2014from experience and asking others along the way\u2014as I grew as a first time CTO.<\/p><h2>Startup Engineering Upgrade Tree<\/h2><p>Before we go much further, what the heck is a CTO? This is the kind of question that I was asking myself when we were hiring our first couple engineers. Is it like a benevolent dictator for life, which in software communities is someone who watches over the code? Is it some sort of super manager or super hacker?<\/p><p>I asked a lot of people for advice. I reached out to CTOs at Buffer, Unbounce, and other startups that were a little bit bigger than us. And I got an answer: Nobody really knows. It's a fuzzy sort of role that doesn\u2019t fit easily in a box. As I talked to people, I did discover some common patterns for how a small, tech-centric startup grows\u2014and what a CTO needs to do at each stage.<\/p><p><img alt=\"Excuse the nerd: upgrade tree\" src=\"https:\/\/cdn.zapier.com\/storage\/photos\/b36e4ecabe0243113abd8803a6929d88.png\" \/><\/p><p>Since I'm kind of a nerd and like playing video games, it\u2019s made the most sense to me as an upgrade tree. As you get more experience in a game, you level up, and at certain points you can choose different branches.<\/p><p><img alt=\"CTO upgrade tree\" src=\"https:\/\/cdn.zapier.com\/storage\/photos\/8219b8bf4b1f7bb7b40c93b46670840f.png\" \/><\/p><p>Here\u2019s how the startup CTO upgrade tree might look.<\/p><h3>Solo Hacker (1-2 engineers)<\/h3><p>This is the time when we were working in the basement. It\u2019s both a very fun and very fast time. You're building a lot of stuff. You're iterating really quickly. There's really no communication or management overhead because you're all in sync. You're all just trying to figure out what is going to work. <\/p><h3>Small Team of Hackers (2-6 engineers)<\/h3><p>As a startup founder, you want to grow the team, so you start to hire people. Usually this is when you hire friends. Some say not to hire your friends, but I actually think it\u2019s great at this early stage. You know you can work with them, you know how smart they are, and they just fit right in. This small team is when the CTO starts to feel the pain of moving from a hacker to something else. It still feels pretty good because you only have a couple engineers. The communication overhead is pretty small. You're still pretty much in sync. Things are happening that you know about. Nothing's really a surprise.<\/p><h3>Growing the Team (6-12 engineers)<\/h3><p>Then it gets a little weird. You no longer are constantly talking with every single person on the team. You have to start doing this scary thing called management, which for me was totally different. This was not as natural for me as hacking. This is where communication overhead starts to blow right past the time you get to code. You may be trying to figure out how to hire people who aren't your friends. And it gets tricky because things you don\u2019t know about are happening. Things change really quickly in this stage, and you\u2019ll face your first branch on the upgrade tree.<\/p><h3>Organizing the Team (12+ engineers)<\/h3><p>With a team of this size, you likely are working on different areas. You can\u2019t do a good job of both people and code management, so this is where you get a choice.<\/p><p class=\"pullquote\"><strong>VP of Engineering<\/strong>: focus on management<br \/><strong>CTO<\/strong>: focus on hacking and architecture<br \/><\/p><p>The VP of Engineering is usually the person who puts in place process, puts in place tools to make the team more productive, and helps the engineers work through stuff. Or, you can remain CTO, and maintain the hacker aspects of your role. This is someone who knows the system in and out, who knows where all the skeletons are in the closets.<\/p><h2>Should You Still be CTO?<\/h2><p>I wasn\u2019t really prepared for this decision. I didn't even know that you had to choose between the two. I just thought the CTO would be the de facto manager. And that seems to be the case for over half of startups I talked to. Whether you go the manager or hacker route, you have to find someone who will do the opposite.<\/p><table class=\"tg\"><tr><th class=\"tg-031e\">\n    <\/th><th class=\"tg-031e\">engineers<\/th>\n    <th class=\"tg-031e\">time coding<\/th>\n    <th class=\"tg-yw4l\">confidence<\/th>\n  <\/tr><tr><td class=\"tg-z2zr\">solo hacker<\/td>\n    <td class=\"tg-z2zr\">1 or 2<\/td>\n    <td class=\"tg-z2zr\">100%<\/td>\n    <td class=\"tg-j2zy\">100%<\/td>\n  <\/tr><tr><td class=\"tg-031e\">small team hacker<\/td>\n    <td class=\"tg-031e\">2-6<\/td>\n    <td class=\"tg-031e\">~80%<\/td>\n    <td class=\"tg-yw4l\">90%<\/td>\n  <\/tr><tr><td class=\"tg-z2zr\">a bit of everything<\/td>\n    <td class=\"tg-z2zr\">6-12<\/td>\n    <td class=\"tg-z2zr\">&lt; 50%<\/td>\n    <td class=\"tg-j2zy\">10-50%<\/td>\n  <\/tr><tr><td class=\"tg-031e\">VPE<\/td>\n    <td class=\"tg-031e\">12+<\/td>\n    <td class=\"tg-031e\">0%<\/td>\n    <td class=\"tg-yw4l\">90%<\/td>\n  <\/tr><tr><td class=\"tg-z2zr\">CTO<\/td>\n    <td class=\"tg-z2zr\">12+<\/td>\n    <td class=\"tg-z2zr\">~80%<\/td>\n    <td class=\"tg-j2zy\">60%<\/td>\n  <\/tr><\/table><p>To make the decision, it\u2019s good to think back to where you\u2019ve come from. Where was your time spent, how confident did you feel, and which stage is one you\u2019d like to revisit?<\/p><p>In the early days, it's one or two engineers. You spend all your time hacking. It feels great. Everything's wonderful. Whenever you get that small team of up to 6 people, it still feels pretty good. You're spending about 80 percent of your time hacking, maybe 20 percent communicating. And you still feel pretty good. You're feeling like 90 percent confident about this.<\/p><p>And then when it gets to be a bit of everything\u2014that's where it gets much more difficult. You spend less than 50 percent of your time as you get more and more people. There are more and more people to facilitate, and it just doesn't feel as good, because you\u2019re trying to do that while maintaining your hacking schedule. That is when you have to decide whether to go manager or hacker. So, how do you navigate that tree?<\/p><p class=\"pullquote\">You really need to figure out where you get your dopamine hits<br \/><\/p><p>I got this great advice from another founder. Do you feel really good when you help someone do something that they were struggling with and they nail it? Or do you get a hit when you solve some sort of a technical problem? That should give you a little bit of guidance on which way you go.<\/p><h2>Four Lessons from Startup CTOs<\/h2><p>I interviewed around 15 founders and CTOs as I wrestled with my own decision of where to focus. I asked each one about the things they noticed that were difficult along their journey. Four areas kept popping up in these conversations:<\/p><h3>1. Embrace the Breaking Points<\/h3><p>I asked everyone for their biggest mistake\u2013the hardest experience they had and how they fixed it. Then I asked how they would prevent it the next time around. Everybody stopped and thought a bit. Then they would say, \"You know, I probably wouldn't. We needed to go through that phase to really figure out what we were trying to do. I had to learn the opposite side to learn the proper side.\"<\/p><p>To put it in CTO terms, treat this breaking point more like scaling a service. You run into the bottlenecks and then you fix them once you hit them. It's much more efficient than trying to predict them because prediction is very difficult, especially in something like a startup. <\/p><h3>2. No Man's Land<\/h3><p>This is the name that I gave that time between six and 12 engineers. There's a little bit of everything going on and it's kind of confusing. You notice yourself saying things like, \"How come this person didn't know about this? And don't they know about what they should've done?\" These sorts of questions are a hint that you're in this area where it's difficult to either go one way or the other\u2014either hacker or manager\u2014without having someone to pick up the slack on the other side. You're not big enough to really warrant two different roles, but you're not small enough to where you can ignore it. You're kind of stuck in the middle part.<\/p><p>Communication can be a bridge in this stage. When you find yourself saying things like, \"Why did xyz do this?\" you need to start over-communicating. Say things over and over again almost to where everybody when they hear you say it again they roll their eyes\u2013\"Oh, here we go again.\" But that's important because then everyone's on the same page.<\/p><h3>3. Mimic Structure<\/h3><p>So often I would hear these CTOs say, \"Well we saw Spotify do this\" or \"We saw Amazon do that.\" Then they would emulate that sort of structure. As a tech company, if you're innovating, you should not be innovating in management. You should not be innovating in how you organize your company. You should be innovating with your technology, with your code, and with your product.<\/p><p>Pick out things that you think are interesting and try them in your organization. Over time your company becomes unique, with a melting pot of ideas and ways to structure your organization that no one else really has. And maybe you'll come up with a few things on your own\u2014but that really shouldn't be your goal. You should really be trying to just pick up things other people are doing and save yourself the hassle.<\/p><h3>4. Focus Up<\/h3><p>This final lesson comes from personal mistakes. When we started hiring engineers, we thought more people meant we could do more things. Early on, the founders and the first couple employees were juggling so many things, as you do, that we thought it was normal for everyone to be working on two or three projects. At four engineers, that\u2019s eight to 12 projects at a time, right? Nope. That was a bad, bad idea.<\/p><p>More projects means a lot more threads of communication. It gets much, much more difficult to keep track of what's going on and to keep everybody on the same page. In retrospect, when we added more teammates, the right solution was to either hold steady in terms of workload, or even pull back a little bit. <\/p><p>The way we did this is we picked one big project that we all focused on and made it super clear to everyone that it was our top priority. There was some secondary stuff going on, but everything was second to this number one goal.<\/p><hr \/><p>I hope my own journey will be a little bit of a help to anyone out there who's trying to navigate the waters of an early startup, especially as a technical founder. It would be great if I knew all of this before I started, but then maybe I wouldn\u2019t appreciate the lessons. Regardless, it's okay to be a little bit confused, delay some of these decisions, and figure it out as you go. That's pretty much par for the course, because it\u2019s natural for a startup to need different things at different stages.<\/p>","created_at":"2018-11-07T16:51:53+0000","updated_at":"2018-11-07T16:51:53+0000","published_at":"2017-06-22T06:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"zapier.com","preview_picture":"https:\/\/cdn.zapier.com\/storage\/blog\/a726597ab1b4c6e034472899022a9b0a.1800x1000.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":556,"title":"Side Project Checklist","url":"https:\/\/www.sideprojectchecklist.com\/","content":"<main class=\"page-content\" aria-label=\"Content\"><div class=\"wrapper\"><div class=\"home\"><div class=\"panel-wrapper\"><div class=\"panel\"><h3>\n      <a href=\"https:\/\/www.sideprojectchecklist.com\/marketing-checklist\">Get the Marketing Checklist<\/a>\n    <\/h3><div class=\"container\"><p>Give your side project a long-term plan for user acquisition and growth that won't cost you a lot of time or money.<\/p><p><a href=\"https:\/\/www.sideprojectchecklist.com\/marketing-checklist\" class=\"btn btn-large\">Marketing Checklist<\/a><\/p><\/div><\/div><div class=\"panel second\"><h3>\n      <a href=\"https:\/\/www.sideprojectchecklist.com\/sales-checklist\">Get the Sales Checklist<\/a>\n    <\/h3><\/div><\/div><h2>What is This?<\/h2><p>\n    The side project checklists are comprehensive, chronologically ordered lists of marketing and sales tactics you can try with your next side project. They are free, <a href=\"https:\/\/www.sideprojectchecklist.com\/contributing\">open source<\/a>, and designed to be customized to your project's needs.\n  <\/p><h2>Why Checklists?<\/h2><p>\n    I've been building software side projects for years, but <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/why\/\">I've always been terrible at marketing and selling them<\/a>. I come up with some ideas, forget those ideas, do some random stuff, then lose interest. The goal of this project is to fix that by making marketing and sales as simple as a checklist.\n  <\/p><p>Not sure how to use these checklists? <a href=\"https:\/\/www.sideprojectchecklist.com\/2017\/how-to-use\/\">Click here to learn more<\/a>.<\/p><h2>Latest Blog Posts<\/h2><p><a href=\"https:\/\/www.sideprojectchecklist.com\/blog\/\">See All Blog Posts<\/a><\/p><\/div><\/div>\n    <\/main><form id=\"vote-form\" class=\"modal\" action=\"action\">\n  <h3>Recommend this product<\/h3>\n  <p>Enter your email address to vote for this product.<\/p>\n  <p>Email for verification purposes only. We don't share your address.<\/p>\n<\/form><div id=\"thanks-modal\" class=\"modal\"><h3>Thanks!<\/h3><p>Your vote has been cast.<\/p><\/div>","created_at":"2018-11-07T16:51:53+0000","updated_at":"2018-11-07T16:51:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.sideprojectchecklist.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":557,"title":"carnager\/rofi-pass","url":"https:\/\/github.com\/carnager\/rofi-pass\/blob\/master\/README.md","content":"<p>A bash script to handle <a href=\"http:\/\/www.passwordstore.org\/\" rel=\"nofollow\">Simple Password Store<\/a>\nin a convenient way using <a href=\"https:\/\/github.com\/DaveDavenport\/rofi\">rofi<\/a>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/aa05943ca1afbcdf52e7af0330da9cf68683161c\/68747470733a2f2f35333238302e64652f726f66692f726f66692d706173732e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/aa05943ca1afbcdf52e7af0330da9cf68683161c\/68747470733a2f2f35333238302e64652f726f66692f726f66692d706173732e706e67\" alt=\"rofi-pass\" title=\"rofi-pass in action\" data-canonical-src=\"https:\/\/53280.de\/rofi\/rofi-pass.png\" \/><\/a><\/p>\n<h2><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<ul><li>\n<p>Open URLs of entries with hotkey<\/p>\n<\/li>\n<li>\n<p>Add new Entries to Password Storage<\/p>\n<\/li>\n<li>\n<p>Edit existing Entries<\/p>\n<\/li>\n<li>\n<p>Generate new passwords for entries<\/p>\n<\/li>\n<li>\n<p>Inline view, which can copy\/type individual entries<\/p>\n<\/li>\n<li>\n<p>Move\/Delete existing entries<\/p>\n<\/li>\n<li>\n<p>Support for different password stores (roots), e.g. to separate passwords for\nwork from private passwords<\/p>\n<\/li>\n<li>\n<p>Type any field from entry<\/p>\n<\/li>\n<li>\n<p>Auto-typing of user and\/or password fields.\nThe format for password files should look like:<\/p>\n<pre>foobarmysecurepassword\nuser: MyUser\nurl: http:\/\/my.url.foo\n<\/pre>\n<\/li>\n<li>\n<p>Auto-typing of more than one field, using the <code>autotype<\/code> entry:<\/p>\n<pre>foobarmysecurepassword\n---\nuser: MyUser\nSomeField: foobar\nAnotherField: barfoo\nurl: http:\/\/my.url.foo\nautotype: SomeField :tab user :tab AnotherField :tab pass\n<\/pre>\n<p>You can use <code>:tab<\/code>, <code>:enter<\/code>, or <code>:space<\/code> here to type <kbd>Tab<\/kbd>,\n<kbd>Enter<\/kbd>, or <kbd>Space<\/kbd> (useful for toggling checkboxes)\nrespectively.\n<code>:otp<\/code> will generate an OTP, either <code>pass-otp(1)<\/code> style, or according to the\n<code>otp_method:<\/code>, if it is defined.<\/p>\n<\/li>\n<li>\n<p>Generating OTPs.\nThe format for OTPs should either be <code>pass-otp(1)<\/code>-compatible<\/p>\n<pre>[...]\notpauth:\/\/[...]\n<\/pre>\n<p>Or it should define a method for generating OTPs:<\/p>\n<pre>[...]\notp_method: \/opt\/obscure-otp-generator\/oog --some-option some args\n<\/pre>\n<p><code>:delay<\/code> will trigger a delay (2 seconds by default).<\/p>\n<\/li>\n<li>\n<p>All hotkeys are configurable in the config file<\/p>\n<\/li>\n<li>\n<p>The field names for <code>user<\/code>, <code>url<\/code> and <code>autotype<\/code> are configurable<\/p>\n<\/li>\n<li>\n<p>Bookmarks mode (open stored URLs in browser, default: Alt+x)<\/p>\n<\/li>\n<li>\n<p>Share common used passwords between several entries (with different URLs, usernames etc)<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<h3><a id=\"user-content-bsd\" class=\"anchor\" aria-hidden=\"true\" href=\"#bsd\"><\/a>BSD<\/h3>\n<ul><li>gnugrep<\/li>\n<li>gawk<\/li>\n<\/ul><h2><a id=\"user-content-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration\"><\/a>Configuration<\/h2>\n<p>rofi-pass may read its configuration values from different locations in the following order:<\/p>\n<ul><li><code>ROFI_PASS_CONFIG<\/code> (environment variable)<\/li>\n<li><code>$HOME\/.config\/rofi-pass\/config<\/code><\/li>\n<li><code>\/etc\/rofi-pass.conf<\/code>\nrofi-pass only loads the first existing file.\nIn case no config file exists, rofi-pass uses its internal default values.\nYou can set the environment variable like this:<\/li>\n<\/ul><pre>ROFI_PASS_CONFIG=\"$HOME\/path\/to\/config\" rofi-pass\n<\/pre>\n<p>For an example configuration please take a look at the included <code>config.example<\/code> file.<\/p>\n<h2><a id=\"user-content-extras\" class=\"anchor\" aria-hidden=\"true\" href=\"#extras\"><\/a>Extras<\/h2>\n<h3><a id=\"user-content-addpass\" class=\"anchor\" aria-hidden=\"true\" href=\"#addpass\"><\/a>addpass<\/h3>\n<p>rofi-pass comes with a tiny helper script, which makes it easier to create new pass entries.\nJust run it with<\/p>\n<pre>addpass --name \"My new Site\" +user \"zeltak\" +branch \"branch\" +custom \"foobar\" +autotype \"branch :tab user :tab pass\"\n<\/pre>\n<ul><li>First argument <code>--name<\/code> is mandatory. This will be the filename of the new password entry.<\/li>\n<li>Second argument can be <code>--root<\/code> followed by absolute path to your password-store. addpass also uses root config setting from rofi-pass config file. If both are not found, PASSWORD_STORE_DIR variable is checked. If none of the above are found, the default location <code>$HOME\/.password-store<\/code> is used.<\/li>\n<li><code>--root<\/code> can also be a colon separated list of directories, in which case you can navigate between them on the main menu with Shift+Left and Shift+Right.<\/li>\n<li>Fieldnames are defined with <code>+<\/code> and the actual value is defined inside the quotations. You can add as many fields as you like<\/li>\n<\/ul><h3><a id=\"user-content-keepass2-import-script\" class=\"anchor\" aria-hidden=\"true\" href=\"#keepass2-import-script\"><\/a>keepass2 import script<\/h3>\n<p>Also included is an import script for keepass2 databases. It's the same script that can be downloaded from the pass homepage, with some minor modifications to match rofi-pass structure.<\/p>\n<h3><a id=\"user-content-csv-exporter\" class=\"anchor\" aria-hidden=\"true\" href=\"#csv-exporter\"><\/a>csv exporter<\/h3>\n<p>Finally a script to export your pass database to csv is included. The resulting csv was tested in keepassxc.<\/p>\n<h2><a id=\"user-content-sharing-passwords\" class=\"anchor\" aria-hidden=\"true\" href=\"#sharing-passwords\"><\/a>Sharing passwords<\/h2>\n<p>Rofi-pass allows you to easily share common used passwords across multiple entries.\nFor example, if you have an academic account which includes several services (such as a library, Salary, Student portal etc),  all with different URL's, login forms etc. you can share one password across all of them. This is very handy when the passwords change annually.\nTo use this function you need to add the following line instead of the password, referencing a pass file which holds the password.<\/p>\n<pre>#FILE=PATH\/to\/filename\n<\/pre>\n<p>where PATH is relative to your password-store.<\/p>\n<p><em>And yes, you should slap your service provider for forcing you to share passwords across services.<\/em><\/p>\n<h2><a id=\"user-content-user-filename-as-user\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-filename-as-user\"><\/a>User filename as user<\/h2>\n<p>If your password file has no user field you can ask rofi-pass to use the filename instead.\nFor example with this password file path : <code>web\/fsf.org\/rms<\/code> rofi-pass will user <code>rms<\/code> as your username.\nTo get this, you need to set <code>default_user<\/code> to <code>:filename<\/code> in your configuration.<\/p>\n<h2><a id=\"user-content-faq\" class=\"anchor\" aria-hidden=\"true\" href=\"#faq\"><\/a>FAQ<\/h2>\n<h3><a id=\"user-content-rofi-pass-prints-garbage-instead-of-my-actual-passes\" class=\"anchor\" aria-hidden=\"true\" href=\"#rofi-pass-prints-garbage-instead-of-my-actual-passes\"><\/a>rofi pass prints garbage instead of my actual passes<\/h3>\n<p>Make sure to run <code>setxkbmap &lt;language&gt; &lt;variant&gt;<\/code> at the start of your Xorg\nsession.<\/p>\n<h2><a id=\"user-content-alternative\" class=\"anchor\" aria-hidden=\"true\" href=\"#alternative\"><\/a>Alternative<\/h2>\n<p>jreinert has written the roughly compatible tool\n<a href=\"https:\/\/github.com\/jreinert\/autopass\">autopass<\/a>. It has less features, but\ndefinately saner code.\nAlso he provided a nice little script called <code>passed<\/code> to change your\nfieldnames. <a href=\"https:\/\/github.com\/jreinert\/passed\">link<\/a><\/p>","created_at":"2018-11-07T16:51:53+0000","updated_at":"2018-11-07T16:51:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/3321178?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":554,"title":"Get your loved ones off Facebook.","url":"http:\/\/www.salimvirani.com\/facebook\/","content":"<p><em>I originally wrote this for my friends and family in 2015, to explain why the latest Facebook privacy policy is really harmful. It\u2019s since been read by over a million people, and I updated it earlier thise year. External references \u2013 and steps to get off properly \u2013 at the bottom. Oh, and if you\u2019re interested in a privacy-friendly way to stay in touch with friends, I\u2019m looking into making <a href=\"https:\/\/sourceinstitute.typeform.com\/to\/cQ7qWt\">a simple phone app for easy sharing<\/a>.<\/em><\/p><p><em>2017 update: A lot of these concerns I had have proven true. Facebook has been consistent with its pattern of contempt for its users. I\u2019ve updated this with a few more points and links.<\/em><\/p>\n<p>\u201cOh yeah, I\u2019ve been meaning to ask you why you\u2019re getting off Facebook,\u201d is the guilty and reluctant question I\u2019m hearing a lot these days. Like we kinda know Facebook is bad, but don\u2019t really want to know.<\/p>\n<p>I\u2019ve been a big Facebook supporter - one of the first users in my social group who championed what a great way it was to stay in touch, way back in 2006.  I got my mum and brothers on it, and around 20 other people. I\u2019ve even taught Facebook marketing in one of the UK\u2019s biggest tech education projects, Digital Business Academy. I\u2019m a techie and a marketer \u2013 so I can see the implications \u2013 and until now, they hadn\u2019t worried me. I\u2019ve been pretty dismissive towards people who hesitate with privacy concerns.<\/p>\n<h3 id=\"just-checking\">Just checking\u2026<\/h3>\n<p>Over the 2014\/15 holidays, I thought I\u2019d take a few minutes to check on the upcoming privacy policy change, with a cautious \u201cwhat if\u201d attitude. With our financial and location information on top of everything else, there were some concerning possibilities. Turns out what I suspected already happened 2 years ago, starting in 2011! That few minutes of searching turned into a few days of research and reading. I dismissed a lot of claims I found that could be explained as accidents - technically plausible or technically lazy. After all, I\u2019m a startup founder myself and know how hard tech can be.  For example, the excessive permissions that Facebook Android app asks for are a technical issue to do with Android.  But there was still a lot of serious privacy concerns left over, and I considered those facts with techniques that I know to be standard practice in data-driven marketing.<\/p>\n<p>With this latest privacy change on January 30th, 2015 I\u2019m scared.<\/p>\n<p>Facebook has always been slightly worse than all the other tech companies with dodgy privacy records, but now, it\u2019s in its own league.  Getting off isn\u2019t just necessary to protect yourself, it\u2019s necessary to protect your friends and family too.  This could be the point of no return \u2013 but it\u2019s <a href=\"https:\/\/medium.com\/@SaintSal\/you-are-our-first-defence-b86ae9a1fff6\">not too late to take back control.<\/a><\/p>\n<h2 id=\"a-short-list-of-some-facebook-practices\">A short list of some Facebook practices<\/h2>\n<p>It\u2019s not just what Facebook is saying it\u2019ll take from you and do with your information, it\u2019s all the things it\u2019s not saying, and doing anyway because of the loopholes they <a href=\"https:\/\/www.eff.org\/deeplinks\/2010\/04\/facebook-timeline\">create for themselves in their Terms of Service<\/a> and how simply they <a href=\"https:\/\/en.wikipedia.org\/wiki\/Criticism_of_Facebook#cite_ref-informer_33-0\">go back on their word<\/a>. We don\u2019t even need to click \u201cI agree\u201d anymore.  They just change the privacy policy and by staying on Facebook, you agree. Oopsy!<\/p>\n<p>Facebook doesn\u2019t keep any of your data safe or anonymous, no matter how much you lock down your privacy settings. Those are all a decoy. There are very serious privacy breaches, like <a href=\"http:\/\/www.forbes.com\/sites\/anthonykosner\/2013\/01\/21\/facebook-is-recycling-your-likes-to-promote-stories-youve-never-seen-to-all-your-friends\/\">selling your product endorsement to advertisers<\/a> and <a href=\"http:\/\/www.zdnet.com\/article\/is-facebook-damaging-your-reputation-with-sneaky-political-posts\/\">politicians<\/a>, <a href=\"http:\/\/papers.ssrn.com\/sol3\/papers.cfm?abstract_id=1717563\">tracking everything you read on the internet<\/a>, or <a href=\"http:\/\/www.digitaltrends.com\/social-media\/facebook-snitch-on-friends-that-arent-using-real-names\/\">using data from your friends to learn private things about you<\/a> - none of these privacy breaches have an off switch.  Worse yet, Facebook does these things without ever letting you know, or revealing the damage to you if you ask.<\/p>\n<p>Facebook gives your data to \u201cthird-parties\u201d through your use of apps, and then say that\u2019s you doing it, not them. Every time you use an app connected to Facebook, you\u2019re allowing Facebook to escape its own privacy policy with you and with your friends. It\u2019s like when my brother used to make me punch myself and ask, \u201cwhy are you punching yourself?\u201d Then he\u2019d tell my mum it wasn\u2019t his fault.<\/p>\n<p>I dug in. I discovered all the spying Facebook does \u2013 which I double-checked with articles from big reputable news sources and academic studies that were heavily scrutinised.  (Links are all in the Source section at the bottom of this post.)<\/p>\n<p>It sounds nuts when you put it all together!<\/p>\n<ul><li>Facebooks <a href=\"http:\/\/www.forbes.com\/sites\/anthonykosner\/2013\/01\/21\/facebook-is-recycling-your-likes-to-promote-stories-youve-never-seen-to-all-your-friends\/\">creates false endorsements<\/a> for products from you to your friends - and they never reveal this to you.<\/li>\n  <li>When you see a like button on the web, <a href=\"http:\/\/spectrum.ieee.org\/podcast\/telecom\/internet\/stalking-on-facebook-is-easier-than-you-think\">Facebook is tracking that you\u2019re reading that page<\/a>.  It scans the keywords on that page and associates them to you. It knows much time you spend on different sites and topics.<\/li>\n  <li>They <a href=\"http:\/\/www.engadget.com\/2014\/12\/25\/facebook-class-action-privacy-lawsuit\/\">read your private messages <em>and<\/em> the contents of the links you send privately.<\/a><\/li>\n  <li>They\u2019ve introduced <a href=\"http:\/\/www.forbes.com\/sites\/kashmirhill\/2014\/05\/22\/facebook-wants-to-listen-in-on-what-youre-doing\/\">features that turn your phone\u2019s mic on without telling you<\/a>. Based on their <a href=\"http:\/\/www.pcworld.com\/article\/218528\/famous_facebook_flipflops.html\">track-record<\/a> changing <a href=\"http:\/\/mattmckeon.com\/facebook-privacy\/\">privacy settings back without telling you<\/a>, audio surveillance is likely to start happening without your knowledge.<\/li>\n  <li>They can use <a href=\"http:\/\/www.dw.de\/facebook-violates-german-law-hamburg-data-protection-official-says\/a-15290120\">face recognition<\/a>  to <a href=\"http:\/\/www.makeuseof.com\/tag\/exif-photo-data-find-understand\/\">track your location through pictures<\/a> , even those that aren\u2019t on Facebook. (Pictures taken with mobile phones have time, date and GPS data built into them.)<\/li>\n  <li>They <a href=\"https:\/\/www.tripwire.com\/state-of-security\/security-data-protection\/cyber-security\/stalk-location-facebook-messenger\/\">track your location<\/a>, and use that to discover private things about you, like if you\u2019re sick (because you\u2019re at the doctor or a specialist), who you sleep with (who\u2019s beside you at night), where you work, if you\u2019re job-searching (1-hour meetings at competitors offices), etc.<\/li>\n  <li>They\u2019ve used <a href=\"http:\/\/www.digitaltrends.com\/social-media\/facebook-snitch-on-friends-that-arent-using-real-names\/\">snitching campaigns<\/a> to trick people\u2019s friends into revealing information about them that they chose to keep private.<\/li>\n  <li>They use the vast amount of data they have on you, from your likes, things you read, things you type but don\u2019t post, to make <a href=\"http:\/\/www.digitaltrends.com\/social-media\/facebook-snitch-on-friends-that-arent-using-real-names\/\">highly accurate models<\/a>  about who you are \u2013 even if you make it a point of keeping these things secret. There are statistical techniques, which have been used in marketing for decades, that find correlating patterns between someone\u2019s behaviour and their attributes. Even if you never posted anything, they can easily work out your age, gender, sexual orientation and political views. When you post, they work out much more. Then they reveal it to banks, insurance companies, governments, and of course, advertisers.<\/li>\n<\/ul>\n<p>A lot of people aren\u2019t worried about this, feeling they have nothing to hide.  Why would they care about little old me?  Why should I worry about this when I\u2019m not doing anything wrong?<\/p>\n<p>The now famous story of the pregnant teenager being outed by the store Target, after it mined her purchase data \u2013 larger handbags, headache pills, tissues \u2013 and sent her a \u201ccongratulations\u201d message as marketing, which her unknowing father got instead. Oops!<\/p>\n<p>The same is done about you, and revealed to any company without your control.  And this extrapolates into different ways your data can reveal something about you to people you don\u2019t want to know about it.<\/p>\n<p>One of the more obvious problems here is with <a href=\"http:\/\/www.insure.com\/car-insurance\/social-media-future.html\">insurance companies<\/a>. The data they have on you is mined to predict your future. Would you like to be denied health insurance because an algorithm incorrectly predicted you were starting to see a heart specialist?<\/p>\n<p>What about your employer or prospective employer knowing that you might be pregnant?<\/p>\n<p>Would you like your boss to know when you\u2019re not really home sick, or when you\u2019re looking for another job?<\/p>\n<p>Would you like others to know if you\u2019re having trouble paying your mortgage?  If you\u2019re selling your house, buyers will know they have the upper hand.<\/p>\n<blockquote>\n  <p>Don\u2019t confuse privacy with secrecy. I know what you do in the bathroom, but you still close the door. That\u2019s because you want privacy, not secrecy. (From <a href=\"https:\/\/medium.com\/@FabioAEsteves\/i-have-nothing-to-hide-why-should-i-care-about-my-privacy-f488281b8f1d#.qk4pd42ww\">I have nothing to hide. Why should I care about my privacy?<\/a><\/p>\n<\/blockquote>\n<p>While most of us feel we have nothing to hide, we all come across situations where we need certain things to be secret, at least for a while.  But we\u2019re giving that up \u2013 and for what?<\/p>\n<p>From the Terms Of Service (not the Privacy Policy \u2013 see what they did there?):<\/p>\n<blockquote>\n  <p>You give us permission to use your name, profile picture, content and information in connection with commercial, sponsored or related content (such as a brand you like), served or enhanced by us.<\/p>\n<\/blockquote>\n<p>And later:<\/p>\n<blockquote>\n  <p>By \u201cinformation\u201d we mean facts and other information about you, including actions taken by users and non-users who interact with Facebook.<\/p>\n<\/blockquote>\n<p>So this includes everything they\u2019re collecting about you but not telling you. Everything you read online, everything someone ever posts about you, all your private financial transactions.<\/p>\n<p>And, your data starts to get combined with your friends data to make these models more accurate.  It\u2019s not just about you and your data but what gets done with all of it put together.<\/p>\n<p>Even if the issue here isn\u2019t what we have to hide, it\u2019s maintaining an important right to our freedom \u2013 which is the right to privacy.<\/p>\n<p>Article 12 of <a href=\"http:\/\/www.un.org\/en\/universal-declaration-human-rights\/index.html\">UN Universal Declaration of Human Rights<\/a> states:<\/p>\n<blockquote>\n  <p>\u201cNo one must be subjected to arbitrary interference with his privacy, family, home or correspondence, nor to attacks upon his honor and reputation.\u201d<\/p>\n<\/blockquote>\n<p>We have the right to have a say in how information about us is used. But by using Facebook, we\u2019re voluntarily giving it up, not just for us but for our friends and family!<\/p>\n<p>If you\u2019ve ever admitted to something illegal in a private Facebook message, or even mentioned your support for a political cause, this can be used against you in the future, especially by another country\u2019s government. You may find yourself arrested for being at the wrong place at the wrong time, or just pulled aside at the airport one day, now facing jail time because you revealed you did something that government considers illegal 5 years ago. One New York comedian had a SWAT team break into his house based on a joke post. Law enforcement often acts in error, and you\u2019re giving them more power and more chance of error. You\u2019re loading the gun, pointing it at your head, and handing it to every trigger-happy \u201cenforcer\u201d who\u2019s willing to buy your data.<\/p>\n<p>There\u2019s no need to talk hypothetically about government surveillance here. One of the first Facebook investors called Greylock has board connections to a CIA investment firm called In-Q-Tel. According to their website, it \u201cidentifies cutting-edge technologies to help the Central Intelligence Agency and the broader US Intelligence Community to further their missions\u201d. And if you haven\u2019t heard - it was revealed that <a href=\"http:\/\/www.theguardian.com\/world\/2013\/jun\/06\/us-tech-giants-nsa-data\">Facebook data is delivered directly to the PRISM programme<\/a>.<\/p>\n<h3 id=\"commercial-data-brokers\">Commercial data brokers<\/h3>\n<p>And as I\u2019ll explain later, most of this information finds its way into the public anyway. No need for NSA programmes because of  marketing data companies who de-anonymise all your data to sell it again and again. This is done systematically and automatically. There\u2019s an industry around this.  There are <a href=\"http:\/\/www.propublica.org\/article\/everything-we-know-about-what-data-brokers-know-about-you\">marketplaces to buy and sell consumer data<\/a>, originally started around credit agencies and direct mail companies, then growing with the browser toolbar industry when Internet Explorer was big - now they\u2019re filled with more information than ever before.  A recent example is <a href=\"https:\/\/en.wikipedia.org\/wiki\/TowerData\">RapLeaf<\/a> which collected and released personally identifiable information, including Facebook and MySpace IDs. They stopped after serious controversy, but not only was the damage done, there were other companies who escaped the bad PR and kept up the same practice.  It\u2019s not about how marketers target ads to you, it\u2019s that your data is bought and sold to try.<\/p>\n<p>Where might you travel in the future? Do you trust their law enforcement with this information about you? Because they\u2019re buying it.<\/p>\n<p><em>2017 Update:<\/em> <a href=\"https:\/\/www.facebook-studio.com\/news\/item\/new-ways-to-reach-the-right-audience?sf=goyvdok\">Facebook has announced that it\u2019s teaming up with four of the world\u2019s largest corporate data brokers.<\/a><\/p>\n<h2 id=\"intercepting-your-communication\">Intercepting your communication<\/h2>\n<p>The thing is that you don\u2019t need a conspiracy theory to be concerned.  Mark Zuckerburg himself has been public and consistent to his investors about his intentions:<\/p>\n<p>1) To be the middle-man in all personal communication.<\/p>\n<p>That\u2019s why they made Messenger and bought WhatsApp, but don\u2019t forget that they\u2019ve tried worse. When they made Facebook email, they took advantage of users who were syncing their Facebook contacts. They made it so everyone\u2019s @facebook.com address would be the default. Why? So that your friends would email you at your @facebook.com address instead of your normal email address, and they\u2019d be able to read your emails too.<\/p>\n<p>2) To make all personal communication public over time.<\/p>\n<p>That\u2019s why they slowly changed the default privacy settings to public, made privacy settings harder and harder to use, and now are pretending that their privacy helper will change this.<\/p>\n<p><strong>In reality, there a loads of privacy breaches you can\u2019t turn off, like allowing advertisers to use your endorsement to your friends, turning off how Facebook tracks what you read on the internet, or disallowing Facebook from collecting other information about you.  You can\u2019t turn them off!<\/strong><\/p>\n<p>Even if you have nothing to hide, you have to worry about the opposite, what Facebook chooses to hide that you want to be shared. They filter you.<\/p>\n<p>\u201cI\u2019ve been meaning to ask you why you\u2019re getting off\u201d usually comes after something like, \u201cdidn\u2019t you see my post last week?\u201d<\/p>\n<p>If you\u2019ve ever had that conversation, you\u2019ve noticed that there\u2019s a big disconnect between your expectations when you communicate on Facebook and what really happens.  Basically, Facebook filters out your posts based on whether or not people will use Facebook more if they don\u2019t see it.<\/p>\n<p>It feels like Facebook is the only way to stay in touch. Through pictures and comments. It feels like everyone\u2019s on there and you\u2019re getting a good feed on their life.<\/p>\n<p><strong>In reality, lots of your posts are never seen by anyone!<\/strong> And you miss out on their stuff too.  Even if your friends\u2019 stuff gets to you fine, it doesn\u2019t mean your stuff gets to them.<\/p>\n<p>Private messages suck too. How many Facebook messages do you send with no reply?  How many Facebook messages do you think you forget to get back to, or miss altogether?  Is that how you want to treat your friends?<\/p>\n<p>Facebook a really unreliable way to stay in touch.<\/p>\n<p>In the last month, I simply stopped using Facebook.  Something amazing happened.  People phoned me, and we <em>really<\/em> caught up.  My family was more in touch. My brother emailed me with updates.  Friends popped into my place to say hi.<\/p>\n<p>It was, like, social.<\/p>\n<h2 id=\"political-censorship\">Political censorship<\/h2>\n<p>Facebook\u2019s blocks posts based on political content it doesn\u2019t like.  They blocked posts about Fergusson and other political protests. When Zuckerberg allegedly went a bit nuts and banned the word \u201cprivacy\u201d from meetings at Facebook, it was also blocked from any Facebook post. You just got an error message about \u201cinappropriate content\u201d.  Yeah, uh huh. Inappropriate for who?<\/p>\n<p>We shouldn\u2019t be surprised though. Facebook isn\u2019t a neutral platform - we need be aware of the agendas of the people behind it. Zuckerberg\u2019s been public about his intentions.  So has the first board member of Facebook - the politically conservative Peter Thiel. In his younger years, he wrote a book challenging multi-culturalism at Stanford, and now promotes a theory called <a href=\"https:\/\/www.youtube.com\/watch?v=BNkSBy5wWDk\">Memetic Desire<\/a> which, among other positive things, can also use people\u2019s social groups to manipulate their wants and intentions. (I\u2019m a fan of Thiel when it comes to startups - but we often forget that everyone else out there doesn\u2019t know this stuff.)<\/p>\n<p>Facebook goes so far as to let political organisations block your communication.  It just takes a few people to mark the same news article as offensive, and it drops from everyone\u2019s feed. This is often abused.  I can block any article from Facebook by getting a few friends to mark it as offensive. Cheap and easy censorship.<\/p>\n<p><em>2017 update: We\u2019ve seen how this affected the US election. People\u2019s news feeds often had opposing political views filtered, and yet fake news spread easily because those fake headlines reinforce our beliefs and we\u2019re happy to share them.<\/em><\/p>\n<p>All this points to the fact that it\u2019s bad to rely on Facebook to communicate with people who are important to you. Your Facebook habit means other people have to rely on Facebook.<\/p>\n<p>It\u2019s a vicious cycle.<\/p>\n<p>It actually hurts your relationships with a lot of people because you think you\u2019re in touch with them, but you\u2019re not. At best, you\u2019re in touch with a filtered version of your friends. Those relationships fade, while your relationships with people who make \u201cFacebook-friendly\u201d posts take their place.<\/p>\n<p>Not only does Facebook want to read all your communication, it wants to control it too.<\/p>\n<h2 id=\"ratting-on-your-friends\">Ratting on your friends<\/h2>\n<p>Even if you\u2019re sort of okay with this for you, by using Facebook, you\u2019re forcing your friends and family to accept the same. Even the ones who aren\u2019t on Facebook or go as far as to use fake names.<\/p>\n<p>If you\u2019ve ever used Facebook contact sync, or used Facebook on your mobile phone, Facebook took your complete contact list. Real names, phone numbers, addresses, emails, everything.  They then use that to create <a href=\"http:\/\/www.salimvirani.com\/facebook\/spectrum.ieee.org\/podcast\/computing\/networks\/facebook-knows-your-friendseven-if-theyre-not-on-facebook\/\">\u201cshadow profiles\u201d of the people you know who aren\u2019t on Facebook<\/a>. Non Facebook users often see this in action, in the form of emails to them from Facebook, containing their personal information. Facebook users can see this when they upload a picture of a non-Facebook user, and they\u2019re automatically tagged. My friend\u2019s not on Facebook, but since me and a few friends used Facebook on our phones, Facebook has his name and contact information, plus knows who his friends are because it sees him in their address book and calling records. A couple of pictures uploaded with his face, and presto - they can identify him in pictures \u2013 adding location data from the pictures to his shadow profile. Lots of Facebook\u2019s other techniques work on shadow profiles too. On top of all this, they can very accurately infer a lot about him <a href=\"http:\/\/www.ted.com\/talks\/jennifer_golbeck_the_curly_fry_conundrum_why_social_media_likes_say_more_than_you_might_think\">based on statistical similarities to his friends<\/a>.<\/p>\n<p>So basically, we\u2019ve all inadvertently been ratting on our friends who wanted to remain private. Facebook tricked us.<\/p>\n<p>But Facebook\u2019s tricks go further.<\/p>\n<h2 id=\"privacy-doesnt-apply-to-what-facebook-digs-up\">\u201cPrivacy\u201d doesn\u2019t apply to what Facebook digs up<\/h2>\n<p>Like shadow profiles of people, Facebook can <a href=\"http:\/\/www.forbes.com\/sites\/anthonykosner\/2013\/08\/31\/new-facebook-policies-sell-your-face-and-whatever-it-infers\/\">\u201cinfer a like\u201d<\/a> based on other information it has about you, like what you read all over the internet or what you do in apps where you log in with Facebook, or what\u2019s on your credit card bill (more on that later.). Call it a \u201cshadow like.\u201d This allows them to sell you to more advertisers.<\/p>\n<p>It\u2019s already well-documented that Facebook collects this information.  The \u201cshadow like\u201d technique is simply the standard use of statistical techniques in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Database_marketing\">database marketing<\/a>. If you read a lot about a topic, you probably like it. That sort of thing. These techniques have used in marketing since the 80s, and you can hire university statistics students to do them, though of course, Facebook hires the best in the field and are looking to <a href=\"http:\/\/www.wired.com\/2014\/08\/deep-learning-yann-lecun\/\">pioneer state-of-the-art artificial intelligence for this<\/a>. In Europe, Facebook is legally obliged to share exactly what information it has about you - but they refuse. <a href=\"http:\/\/europe-v-facebook.org\">So there\u2019s yet another class action lawsuit against them.<\/a><\/p>\n<h3 id=\"permission\">Permission<\/h3>\n<p>Through its labyrinth of re-definitions of words like \u201cinformation\u201d, \u201ccontent\u201d and \u201cdata\u201d, you\u2019re allowing Facebook to collect all kinds of information about you and expose that to advertisers. With your permission only they say, but the definition of \u201cpermission\u201d includes using apps and who knows what else.<\/p>\n<p>And you thought those Farmville requests were annoying. Every time you saw one, that friend was revealing your information to \u201cthird parties.\u201d See how this works?  You tell Facebook this is for \u201cfriends only\u201d but your friend can reveal it to a third-party. And most apps that they use are third-parties.<\/p>\n<p>So effectively, all that stuff you marked as \u201cfriends only\u201d doesn\u2019t matter so much. By being on Facebook, there\u2019s way more information about you that\u2019s collected, combined, shared, and used.<\/p>\n<p>They say they \u201canonymise\u201d this, but in reality, it\u2019s a simple step to de-anonymise it.  A lot of the anonymous data, like what and when you posted, pictures of you, your location at a given time, is enough for a huge number of companies to tie that anonymous data back to you \u2013 and sell it on.  (That\u2019s why it doesn\u2019t matter if you use a fake name on Facebook, your data is like a fingerprint and will match back to your real name.)<\/p>\n<p>On top of this, they allow all the Facebook apps full access to your information - with your name and everything.  And even if you never use any apps on Facebook, your friends do.  When they use apps, your friends share all your information for you. There\u2019s a whole industry behind this.<\/p>\n<p>Some things DO have off buttons, but keep in mind they are temporary, and <a href=\"http:\/\/mattmckeon.com\/facebook-privacy\/\">as Facebook has done in the past, it will switch them back on without letting you know<\/a>.  When Facebook started (and probably when you joined) it was clearly a safe place to share with <strong>your friends only<\/strong>. That was their big promise. Over time, they switched the default privacy setting to public so that if you still wanted to keep Facebook for friends only, you had to manually find over 100 settings on multiple hidden settings pages.  Then, they started dropping those settings and forced information to be public anyway.<\/p>\n<p>Why are you <em>still<\/em> punching yourself? :)<\/p>\n<h2 id=\"selling-your-endorsement-without-your-consent\">Selling your endorsement without your consent<\/h2>\n<p>You might have noticed Facebook ads with your friends\u2019 endorsement under it.  Basically, Facebook gives advertisers the right to use you as an endorser, but you have no control over it.  It\u2019s not limited to when you\u2019ve actually clicked a like button.  There have been known cases of vegetarians endorsing McDonalds, a long- and happily-married woman endorsing dating sites, and even a young boy endorsing a sex club to his own mother!<\/p>\n<p>Those cases were so embarrassing that the person found out. People called them up.  But in most cases, these are endorsements that don\u2019t get discovered \u2013 people assume them to be true.  That\u2019s even scarier because Facebook is used heavily for political advertising, and product endorsements. People know I raised money for kids with cancer before, so they might not be surprised if they see an ad where I\u2019m endorsing a Christian outreach programme poor kids in Africa.  But I categorically only support programmes that don\u2019t have religious allegiances, since they\u2019re known to bias their support to people who convert. Worse, a lot of people might assume things about my religious beliefs based on these false endorsements. Don\u2019t even get me started on all the hypey startup stuff I don\u2019t condone!<\/p>\n<h3 id=\"abusing-your-friends-trust-in-you\">Abusing your friends\u2019 trust in you<\/h3>\n<p>We can have no idea if our endorsement has been used to sell flakey crap in our name. I don\u2019t want to think about my mom wasting her money on something she thought I was endorsing, or my startup founder clients seeing adverts for useless products with my face under them.<\/p>\n<p>Using Facebook means this happens all the time.  Advertisers can buy your endorsement on Facebook and your information from third-party data brokers. You never get to know about it, and you can\u2019t turn it off.<\/p>\n<h2 id=\"the-latest-privacy-change\">The latest privacy change<\/h2>\n<p>Finally, I want to explain how this latest privacy change makes things way worse, and way more out of your control if you stay on Facebook.<\/p>\n<p>Facebook is demanding to track what you buy, and your financial information like bank account and credit card numbers. You\u2019ve already agreed to it in the new Terms Of Service. It\u2019s already started  <a href=\"http:\/\/www.theage.com.au\/it-pro\/business-it\/mastercard-to-access-facebook-user-data-20141006-10qrqy.html\">sharing data with Mastercard<\/a>. They\u2019ll use the fact that you stayed on Facebook as \u201cpermission\u201d to make deals with all kinds of banks and financial institutions to get your data from them. They\u2019ll call it anonymous, but like they trick your friends to reveal your data to the third-parties with apps, they\u2019ll create loopholes here too.<\/p>\n<p>Facebook is also insisting to track your location via your phone\u2019s GPS, <em>everywhere and all the time<\/em>. It\u2019ll know exactly who you spend your time with. They\u2019ll know your habits, they\u2019ll know when you call in sick at work, but are really out bowling. \u201cSal likes 2pm Bowling at Secret Lanes.\u201d  They\u2019ll know if you join an addict support group, or go to a psychiatrist, or a psychic, or a mistress. They\u2019ll know how many times you\u2019ve been to the doctor or hospital, and be able to share that with prospective insurers or employers.  They\u2019ll know when you\u2019re secretly job hunting, and will sell your endorsement for job sites to your friends and colleagues \u2013 you\u2019ll be revealed.<\/p>\n<p>They\u2019ll know everything that can be revealed by your location, and they\u2019ll use it however they want to make a buck.<\/p>\n<p>And \u2013 it\u2019ll all be done <em>retrospectively<\/em>. If you stay on Facebook past January 30th, there\u2019s nothing stopping all of your past location and financial data to get used. They\u2019ll get your past location data from when your friends checked-in with you, and the GPS data stored in photos of you.  They\u2019ll pull your old financial records - that embarrassing medicine you bought with your credit card 5 years ago will be added to your profile to be used as Facebook chooses. It will be sold again and again, and likely used against you. It will be shared with governments and be freely available from loads of \u201cthird-party\u201d companies who do nothing but sell personal data, and irreversibly eliminate your privacy.<\/p>\n<h3 id=\"this-is-irreversible-now\">This is irreversible now.<\/h3>\n<p>Location and financial data are not just really sensitive, they allow the \u201cthird-parties\u201d <a href=\"http:\/\/www.pcworld.idg.com.au\/article\/565150\/how-three-small-credit-card-transactions-could-reveal-your-identity\/\">de-anonymise information about you.<\/a>  This massively empowers these third-parties to collect all available information about you, including calculated information that you never revealed. This is a situation where even Facebook itself will have trouble maintaining the privacy of its data \u2013 not that they care.<\/p>\n<p>This is unprecedented, and just like you\u2019d never have guessed that Facebook would sell your endorsements when you signed up in 2009, it\u2019s too hard to predict what Facebook and those third-party data sellers will do with this new power.<\/p>\n<p>This is simply a consequence of their business model. <strong>Facebook sells you out, because that\u2019s exactly how they make money.<\/strong> And they\u2019re under heavy pressure from their investors to make more.<\/p>\n<p>What can you do about this? Facebook gives you two options: accept all of this, or get off the Facebook bus.<\/p>\n<p>To be honest, this bus is getting loud, annoying and bit smelly, isn\u2019t it? And the ticket is way too expensive in the first place. You know, I\u2019m not even sure it\u2019s heading in the right direction\u2026<\/p>\n<p>According to the FTC settlement from a few years ago, after Facebook was sued by the US government for its privacy practices, Facebook is \u201crequired to prevent anyone from accessing a user\u2019s material more than 30 days after the user has deleted his or her account;\u201d<\/p>\n<p>There are different interpretations of this.  Some say you need to delete each post separately, others say delete your account, and some say they\u2019ll still keep your data anyway \u2013 that all you can do is stop giving them more data.  Then, there are the data brokers Facebook works with, that now have our data.<\/p>\n<p>So deleting your Facebook account (not deactivating) is necessary to stop this, and then there are a few other steps to try to undo the damage:<\/p>\n<h3 id=\"preparation-optional\">Preparation (optional)<\/h3>\n<ol><li>\n    <p>Use Facebook\u2019s <a href=\"https:\/\/www.facebook.com\/help\/131112897028467\/\">download my archive<\/a> tool<\/p>\n  <\/li>\n  <li>\n    <p>Grab your photos. I used <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.fb.photo\">this Android app<\/a> since the Facebook tool doesn\u2019t give you all your pictures or in full resolution.  ( I also downloaded my friends list page - just by scrolling to the bottom to load everyone, and hitting File -&gt; Save. Honestly, so far I haven\u2019t needed the file yet. Turns out, I don\u2019t need a computer to tell me who my friends are. )<\/p>\n  <\/li>\n  <li>\n    <p>If you want to be extra careful  <a href=\"http:\/\/nickbriz.com\/facebook\/deletePhotos.html\">delete your Facebook posts individually with this script<\/a>.  This is necessary under some interpretations of what Facebook is entitled to keep.<\/p>\n  <\/li>\n  <li>\n    <p>Then there are all the apps you\u2019ve used.  This is one of Facebook\u2019s best loopholes because they say they can\u2019t control what the apps do with your data once you give it to them.  So, I\u2019ve saved the settings page that shows what apps I\u2019ve used onto my hard drive, and removed access to each of them manually.  Each of those apps have their own privacy policy \u2013 most of them are a lost cause, claiming unlimited rights to my data, so I\u2019m just shutting them down and moving on.<\/p>\n  <\/li>\n<\/ol><h2 id=\"deleting-your-account\">Deleting your account<\/h2>\n<ol><li><a href=\"http:\/\/lifehacker.com\/5813506\/how-to-delete-your-facebook-account\">Delete your Facebook account<\/a> Don\u2019t let Facebook trick you into \u201cdeactivating\u201d it instead.<\/li>\n<\/ol><h2 id=\"further-precautions\">Further precautions<\/h2>\n<p>Facebook will still try to track you with a \u201cshadow profile\u201d, but this can be blocked.<\/p>\n<p>To stop Facebook (and others) from tracking what I read on the web, I use Firefox with <a href=\"https:\/\/www.mozilla.org\/en-GB\/firefox\/47.0\/tracking-protection\/start\/\">Tracking Protection<\/a> enabled. (<a href=\"http:\/\/lifehacker.com\/turn-on-tracking-protection-in-firefox-to-make-pages-lo-1706946166\">Here\u2019s how.<\/a> (they do this even if you don\u2019t have a Facebook account)<\/p>\n<p>If you\u2019re not using Firefox, the EFF have a browser plugin called <a href=\"https:\/\/www.eff.org\/privacybadger\">Privacy Badger<\/a>.  (And while you\u2019re at it, the EFF made this great plugin that <a href=\"https:\/\/www.eff.org\/https-everywhere\">automatically chooses the most secure web connection<\/a>, making it more difficult for internet companies and others to snoop your web activity.)<\/p>\n<p>Then there\u2019s the data that\u2019s been leaked to data brokers.  <a href=\"http:\/\/www.eff.org\/deeplinks\/2013\/02\/howto-opt-out-databrokers-showing-your-targeted-advertisements-facebook\">You can request that they remove it<\/a>.<\/p>\n<h2 id=\"what-replaces-facebook\">What replaces Facebook?<\/h2>\n<p><em>2017 update:<\/em> At first, I thought I\u2019d try Facebook alternatives.  I felt a sense of needing to replace Facebook with something similar, like <a href=\"https:\/\/diasporafoundation.org\/about#features\">Diaspora<\/a>  but <em>email and phone have actually been much better!<\/em> After a month off Facebook, I didn\u2019t feel the need for a direct replacement. The phone call was enough- go figure. Everyone already has one, and we forget how super easy and convenient they are to use. I see fewer pictures, but I actually talk to people.  More recently, we all got on a big chat room.  I currently recommend <a href=\"https:\/\/whispersystems.org\/blog\/signal\/\">Signal<\/a> for this. You can encrypted call, chat and share pictures, and very little is stored on their servers.  It\u2019s actually better than Facebook since it\u2019s more immediate and personal. (But since it\u2019s a private company, there\u2019s no reason to trust it in the long term.  It can be bought and fall to the same pressures.)<\/p>\n<p><a href=\"https:\/\/telegram.org\/\">Telegram<\/a> is also a private company, and doesn\u2019t encyrypt messages by default.  But there are tons of people on it, so you\u2019re likely to find enough friends there to start.<\/p>\n<p>In the long run, my big hope is on <a href=\"https:\/\/tox.chat\/\">Tox<\/a> but it\u2019s early days.  When that\u2019s ready for prime time, it\u2019ll be the first messenging platform that\u2019s fully encrypted and peer-to-peer, so no company will be able to mess with it.<\/p>\n<p>I\u2019m thinking of a little free-software phone app that just allows me to snap a pic or make a little comment, and share it with a fixed list of people over email: just close friends and family. All the convenience of Facebook sharing, to my closest friends, without the privacy cost.  What do you think?  Something you\u2019d be up for? If you\u2019d like that app, <a href=\"https:\/\/sourceinstitute.typeform.com\/to\/cQ7qWt\">sign up for updates on the project.<\/a> If you\u2019d like to help, <a href=\"mailto:smile@saintsal.com\">hit me up :) <\/a><\/p>\n<p>If you have any other ideas or advice, please get in touch. This is what I see as a responsible step to prevent myself, my family and my friends from having their freedom taken from them, and their personal relationships made to suffer.<\/p>\n<p>Remember, this isn\u2019t just about the technical stuff. <strong>By staying on Facebook, you\u2019re granting them permission to collect and use information about you, regardless of you even using the Internet.<\/strong> And by staying on, the data they collect on you gets used to create models about your closest friends and family, even the ones who opted out.<\/p>\n<h2 id=\"the-internet-is-free-and-open-but-that-doesnt-mean-we-to-accept-being-spied-on\">The Internet is free and open, but that doesn\u2019t mean we to accept being spied on<\/h2>\n<p>Lastly, the world is full of people who say  \u201cit\u2019ll never happen\u201d, and when it does, they switch to \u201cthere\u2019s nothing we can do.\u201d There is. The Internet was decentralised for 50 years, and is full of options, by design, that allow us to maintain privacy.  We have a say in the world we want to live in \u2013 if we take action ourselves. Plus, we can help everyone understand, and <a href=\"https:\/\/medium.com\/@SaintSal\/you-are-our-first-defence-b86ae9a1fff6\">help them make their own choices more informed.<\/a><\/p>\n<p>This post has been read by 1,000,000 people now. It\u2019s a positive sign that we can inform and educate ourselves!<\/p>\n<p>Please share this with people who are important to you.  But to be honest, even though this post is really popular, it\u2019s clear a lot of people are assuming what\u2019s in it. Sharing a link isn\u2019t as good as talking to someone.<\/p>\n<p>If you got this far and want to share it with someone close, I suggest you do what I did \u2013 pick up the phone.<\/p>\n\n<link href=\"\/\/cdn-images.mailchimp.com\/embedcode\/classic-10_7.css\" rel=\"stylesheet\" type=\"text\/css\" \/>\n\n<h3 id=\"sources\">Sources<\/h3>\n<p>A note on the quality of these sources: I tried to find references from major news outlets, with a range of political biases. These articles are less technically aware, but we can expect they\u2019re more rigorous than blogs at checking their sources.  For the more technical stuff, sources like The Register are known to be more credible, and Techcrunch is notoriously unreliable at fact-checking.  I\u2019ve included some of their articles though, because they\u2019re good at explaining things.<\/p>\n<p>Facebook likes reveal sensitive personal information\n<a href=\"https:\/\/www.eff.org\/deeplinks\/2013\/03\/facebook-likes-reveal-sensitive-personal-information\">eff.org <\/a><\/p>\n<p>Private traits and attributes are predictable from digital records of human behavior \n<a href=\"http:\/\/www.pnas.org\/content\/110\/15\/5802.full\">pnas.org <\/a> <br \/><a href=\"http:\/\/www.pnas.org\/content\/suppl\/2013\/03\/07\/1218772110.DCSupplemental\/st01.pdf\">table of top likes <\/a><\/p>\n<p>New Facebook Policies Sell Your Face And Whatever It Infers\n<a href=\"http:\/\/www.forbes.com\/sites\/anthonykosner\/2013\/08\/31\/new-facebook-policies-sell-your-face-and-whatever-it-infers\/\">forbes.com<\/a><\/p>\n<p>You are what you Facebook Like\n<a href=\"http:\/\/www.washingtonpost.com\/news\/the-intersect\/wp\/2014\/10\/29\/you-are-what-you-facebook-like\/\">washingtonpost.com<\/a><\/p>\n<p>Criticism of Facebook - Wikipedia, the free encyclopedia\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Criticism_of_Facebook\">en.wikipedia.org <\/a><\/p>\n<p>Facebook\n<a href=\"https:\/\/stallman.org\/facebook.html\">stallman.org <\/a><\/p>\n<p>Forcing users onto Messenger\n<a href=\"http:\/\/www.huffingtonpost.com\/sam-fiorella\/the-insidiousness-of-face_b_4365645.html?fb_action_ids=800240053354677&amp;fb_action_types=og.likes&amp;fb_source=other_multiline&amp;action_object_map=[224913967679581]&amp;action_type_map=[%22og.likes%22]&amp;action_ref_map=[]\">huffingtonpost.com <\/a><\/p>\n<p>Permissions\n<a href=\"http:\/\/www.telegraph.co.uk\/technology\/facebook\/11217273\/Facebooks-Mark-Zuckerberg-Why-I-wear-the-same-T-shirt-every-day.html\">telegraph.co.uk <\/a><\/p>\n<p>WhatsApp \n<a href=\"http:\/\/hackread.com\/whatsapp-facebook-privacy-data-share\/\">hackread <\/a><\/p>\n<p>Europe vs Facebook\n<a href=\"http:\/\/www.irishtimes.com\/business\/technology\/more-than-20-000-join-privacy-action-against-facebook-1.1889192\">irishtimes.com<\/a><\/p>\n<p><a href=\"http:\/\/www.slate.com\/blogs\/future_tense\/2014\/05\/01\/a_developer_tool_so_users_can_log_in_with_facebook_but_stay_anonymous.html\">slate.com<\/a><\/p>\n<p><a href=\"http:\/\/europe-v-facebook.org.\/EN\/Objectives\/objectives.html\">europe-v-facebook.org<\/a><\/p>\n<p>Facebook info sharing created Zoosk.com dating profile for married woman \n<a href=\"http:\/\/www.cbc.ca\/1.2844953\">cbc.ca <\/a><\/p>\n<p>@facebook.com e-mail plague chokes phone address books\n<a href=\"http:\/\/arstechnica.com\/business\/2012\/07\/facebook-com-e-mail-plague-chokes-phone-address-books\/\">arstechnica.com <\/a><\/p>\n<p>Facebook Knows Your Friends\u2014Even if They\u2019re Not on Facebook - IEEE Spectrum\n<a href=\"http:\/\/spectrum.ieee.org\/podcast\/computing\/networks\/facebook-knows-your-friendseven-if-theyre-not-on-facebook\/\">spectrum.ieee.org <\/a><\/p>\n<p>Facebook Now Wants To \u2018Spy\u2019 On Android Phone Users!\n<a href=\"http:\/\/www.efytimes.com\/e1\/fullnews.asp?edid=128859\">efytimes.com <\/a><\/p>\n<p>Facebook adds naggy \u201cask\u201d button to profile pages\n<a href=\"http:\/\/arstechnica.com\/business\/2014\/05\/facebook-introduces-naggy-ask-function-into-profile-pages\/\">arstechnica.com <\/a><\/p>\n<p>Facebook users unwittingly revealing intimate secrets, study finds\n<a href=\"http:\/\/www.theguardian.com\/technology\/2013\/mar\/11\/facebook-users-reveal-intimate-secrets\">theguardian.com <\/a><\/p>\n<p>Facebook\u2019s Generation Y nightmare\n<a href=\"http:\/\/www.theguardian.com\/technology\/2012\/sep\/24\/facebook-generation-y\">theguardian.com <\/a><\/p>\n<p>Facebook Knows Your Friends\u2014Even if They\u2019re Not on Facebook - IEEE Spectrum\n<a href=\"http:\/\/spectrum.ieee.org\/podcast\/computing\/networks\/facebook-knows-your-friendseven-if-theyre-not-on-facebook\/\">spectrum.ieee.org <\/a><\/p>\n<p>Facebook\u2019s New Privacy Rules Clear the Way for Payments Push and Location-Based Ads\n<a href=\"http:\/\/recode.net\/2014\/11\/13\/facebooks-new-privacy-rules-clear-the-wear-for-a-payments-push-and-location-based-ads\/\">recode.net <\/a><\/p>\n<p>Dutch Regulator Investigates Facebook\u2019s Privacy Policy\n<a href=\"http:\/\/bits.blogs.nytimes.com\/2014\/12\/16\/dutch-regulator-investigates-facebooks-privacy-policy\/\">bits.blogs.nytimes.com <\/a><\/p>\n<p>Facebook prunes its privacy policy, lays groundwork for location-based ads - GeekWire\n<a href=\"http:\/\/www.geekwire.com\/2014\/facebook-gives-privacy-policy-haircut-adds-provisions-payment-location-based-ads\/\">geekwire.com <\/a><\/p>\n<p>Stalking on Facebook Is Easier Than You Think - IEEE Spectrum\n<a href=\"http:\/\/spectrum.ieee.org\/podcast\/telecom\/internet\/stalking-on-facebook-is-easier-than-you-think\">spectrum.ieee.org <\/a><\/p>\n<p>Millions Will Flow to Privacy Groups Supporting Weak Facebook Settlement | WIRED\n<a href=\"http:\/\/www.wired.com\/2012\/07\/groups-get-facebook-millions\/all\/\">wired.com <\/a><\/p>\n<p>Facebook Is Recycling Your Likes To Promote Stories You\u2019ve Never Seen To All Your Friends\n<a href=\"http:\/\/www.forbes.com\/sites\/anthonykosner\/2013\/01\/21\/facebook-is-recycling-your-likes-to-promote-stories-youve-never-seen-to-all-your-friends\/\">forbes.com <\/a><\/p>\n<p>Is Facebook damaging your reputation with sneaky political posts? | ZDNet\n<a href=\"http:\/\/www.zdnet.com\/article\/is-facebook-damaging-your-reputation-with-sneaky-political-posts\/\">zdnet.com <\/a><\/p>\n<p>Even Google won\u2019t be around for ever, let alone Facebook\n<a href=\"http:\/\/www.theguardian.com\/technology\/2013\/mar\/03\/google-facebook-nothing-lasts-for-ever\">theguardian.com <\/a><\/p>\n<p>Facebook reforms user settings\n<a href=\"http:\/\/dailyemerald.com\/2010\/05\/27\/facebook-reforms-user-settings\/\">dailyemerald.com <\/a><\/p>\n<p>Facebook Privacy: A Bewildering Tangle of Options - Graphic - NYTimes.com\n<a href=\"http:\/\/www.nytimes.com\/interactive\/2010\/05\/12\/business\/facebook-privacy.html?_r=0&amp;ref=personaltech\">nytimes.com <\/a><\/p>\n<p>Corrupt Personalization\n<a href=\"http:\/\/blogs.law.harvard.edu\/niftyc\/archives\/1003\">blogs.law.harvard.edu <\/a><\/p>\n<p>Facebook is not your friend\n<a href=\"http:\/\/www.theguardian.com\/commentisfree\/andrewbrown\/2010\/may\/14\/facebook-not-your-friend\">theguardian.com <\/a><\/p>\n<p>Facebook violates German law, Hamburg data protection official says | Sci-Tech | DW.DE | 02.08.2011\n<a href=\"http:\/\/www.dw.de\/facebook-violates-german-law-hamburg-data-protection-official-says\/a-15290120\">dw.de <\/a><\/p>\n<p>The World from Berlin: \u2018Every User Can Decide Alone What Facebook Knows\u2019 - SPIEGEL ONLINE\n<a href=\"http:\/\/www.spiegel.de\/international\/germany\/critique-of-german-credit-agency-plan-to-mine-facebook-for-data-a-837713.html\">spiegel.de <\/a><\/p>\n<p>Q&amp;A: Facebook privacy changes\n<a href=\"http:\/\/www.bbc.co.uk\/news\/10165573\">bbc.co.uk <\/a><\/p>\n<p>Famous Facebook Flip-Flops\n<a href=\"http:\/\/www.pcworld.com\/article\/218528\/famous_facebook_flipflops.html\">pcworld.com <\/a><\/p>\n<p>No Death, No Taxes - The New Yorker\n<a href=\"http:\/\/www.newyorker.com\/magazine\/2011\/11\/28\/no-death-no-taxes\">newyorker.com <\/a><\/p>\n<p>Facebook censors political satire after complaint from JobCentre Plus\n<a href=\"http:\/\/tompride.wordpress.com\/2013\/04\/25\/facebook-censors-political-satire-after-complaint-from-jobcentre-plus\/\">tompride.wordpress.com <\/a><\/p>\n<p>TED: The curly fry conundrum: Why social media \u201clikes\u201d say more than you might think\n<a href=\"http:\/\/tedtranscripts.blogspot.com\/2014\/10\/the-curly-fry-conundrum-why-social.html\">tedtranscripts.blogspot.com <\/a><\/p>\n<p>With friends like these \u2026 Tom Hodgkinson on the politics of the people behind Facebook\n<a href=\"http:\/\/www.theguardian.com\/technology\/2008\/jan\/14\/facebook\">theguardian.com <\/a><\/p>\n<p>Facebook Must Face Lawsuit Over Scanning of Users\u2019 Messages, Judge Says\n<a href=\"http:\/\/recode.net\/2014\/12\/24\/facebook-must-face-lawsuit-over-scanning-of-users-messages-judge-says\/\">recode.net <\/a><\/p>\n<p>How Target Figured Out A Teen Girl Was Pregnant Before Her Father Did\n<a href=\"http:\/\/www.forbes.com\/sites\/kashmirhill\/2012\/02\/16\/how-target-figured-out-a-teen-girl-was-pregnant-before-her-father-did\/\">forbes.com <\/a><\/p>\n<p>War on General Purpose Computers is the difference between utopia and dystopia\n<a href=\"http:\/\/boingboing.net\/2014\/12\/26\/war-on-general-purpose-compute.html\">boingboing.net <\/a><\/p>\n<p>Don\u2019t Worry About Selling Your Privacy To Facebook. I Already Sold It For You | Just Well Mixed\n<a href=\"http:\/\/jasonlefkowitz.net\/2011\/10\/dont-worry-about-selling-your-privacy-to-facebook-i-already-sold-it-for-you\/\">jasonlefkowitz.net <\/a><\/p>\n<p>Facebook\u2019s Hidden \u201cLike\u201d Isn\u2019t Just Good For Mobile Developers, It\u2019s Good For Facebook \n<a href=\"http:\/\/techcrunch.com\/2012\/06\/26\/facebooks-hidden-like-isnt-just-good-for-mobile-developers-its-good-for-facebook\/\">techcrunch.com<\/a><\/p>\n<p>Mastercard to access Facebook data\n<a href=\"http:\/\/www.theage.com.au\/it-pro\/business-it\/mastercard-to-access-facebook-user-data-20141006-10qrqy.html\">theage.com.au<\/a><\/p>\n<p>Three transactions can reveal your identity\n<a href=\"http:\/\/www.pcworld.idg.com.au\/article\/565150\/how-three-small-credit-card-transactions-could-reveal-your-identity\/\">pcworld.idg.com.au<\/a><\/p>\n<p>Look who\u2019s lurking around your Facebook page: Your insurance company!\n<a href=\"http:\/\/www.insure.com\/car-insurance\/social-media-future.html\">insure.com<\/a><\/p>\n<p>NSA Prism program taps in to user data of Apple, Google and others \n<a href=\"http:\/\/www.theguardian.com\/world\/2013\/jun\/06\/us-tech-giants-nsa-data\">theguardian.com<\/a><\/p>\n<p>Everything We Know About What Data Brokers Know About You\n<a href=\"http:\/\/www.propublica.org\/article\/everything-we-know-about-what-data-brokers-know-about-you\">propublica.org<\/a><\/p>\n<p>How to Stalk Someone\u2019s Location on Facebook Messenger\n<a href=\"https:\/\/www.tripwire.com\/state-of-security\/security-data-protection\/cyber-security\/stalk-location-facebook-messenger\/\">Trip Wire<\/a><\/p>\n<p>What your online friends reveal about where you are\n<a href=\"https:\/\/www.newscientist.com\/article\/mg21328495.500-what-your-online-friends-reveal-about-where-you-are\/\">New Scientist<\/a><\/p>\n<p>Turn on Tracking Protection in Firefox to Make Pages Load 44% Faster\n<a href=\"http:\/\/lifehacker.com\/turn-on-tracking-protection-in-firefox-to-make-pages-lo-1706946166\">Lifehacker<\/a><\/p>","created_at":"2018-11-07T16:51:52+0000","updated_at":"2018-11-07T16:51:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":31,"domain_name":"www.salimvirani.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":553,"title":"So you want to write a type checker (2014)","url":"https:\/\/news.ycombinator.com\/item?id=15583515","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15583515\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15583515\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/languagengine.co\/blog\/so-you-want-to-write-a-type-checker\/\" class=\"storylink\">So you want to write a type checker (2014)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=languagengine.co\">languagengine.co<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        152 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=ChickeNES\" class=\"hnuser\">ChickeNES<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583515\">on Oct 30, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15583515&amp;goto=item%3Fid%3D15583515\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=So%20you%20want%20to%20write%20a%20type%20checker&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=So%20you%20want%20to%20write%20a%20type%20checker\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15583515&amp;auth=e929d75a5a8ab54989ca956770c7997b7fb98814\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15583515\">98\u00a0comments<\/a>              <\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"15583748\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amelius\" class=\"hnuser\">amelius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583748\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583923\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=labbekak\" class=\"hnuser\">labbekak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583923\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584445\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cf\" class=\"hnuser\">cf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584445\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584441\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yorwba\" class=\"hnuser\">yorwba<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584441\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586331\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=psygnisfive\" class=\"hnuser\">psygnisfive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586331\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586532\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=junke\" class=\"hnuser\">junke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586532\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15587208\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=psygnisfive\" class=\"hnuser\">psygnisfive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15587208\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15588455\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tom_mellior\" class=\"hnuser\">tom_mellior<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15588455\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15588799\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=junke\" class=\"hnuser\">junke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15588799\">on Oct 30, 2017<\/a><\/p><br \/><div class=\"comment\">Thanks!<p>Just to expand on my previous example, I can simply modify how bindings are looked up and generalize the associated type when it is not unknown:<\/p><pre>    type(V,Env,T) :-\n        ground(V),\n        member(V:P,Env),\n        generalize(P,T).\n    generalize(P,T) :- nonvar(P), !, copy_term(P,T).\n    generalize(T,T).\n<\/pre>Let's add two rules for tuples and addition, and change the syntax of let for clarity:<pre>    type((A,B),Env,TA * TB) :-\n        type(A,Env,TA),\n        type(B,Env,TB).\n    type(A+B,Env,number) :-\n        type(A,Env,number),\n        type(B,Env,number).\n    type(let(X == XE, E),Env,ET) :-\n        type(XE,Env,XT),\n        type(E,[X:XT|Env],ET).\n<\/pre>Then, typing this:<pre>    let(identity == lambda(v,v),\n        let(increment == lambda(x,x+1),\n            let(compose ==\n                lambda(f,lambda(g,lambda(x,apply(f,apply(g,x))))),\n                (apply(apply(compose,identity),identity),\n                 apply(apply(compose,increment),increment)))))\n<\/pre>... gives:<pre>    (_1775 -&gt; _1775) * (number -&gt; number)\n<\/pre>You could also make use of meta attributes, ie. data stored inside variables, to control what happens during unification.<\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15589861\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=psygnisfive\" class=\"hnuser\">psygnisfive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15589861\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15591642\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tom_mellior\" class=\"hnuser\">tom_mellior<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15591642\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15588863\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=junke\" class=\"hnuser\">junke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15588863\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15587473\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danblick\" class=\"hnuser\">danblick<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15587473\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15588222\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=psygnisfive\" class=\"hnuser\">psygnisfive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15588222\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584477\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gingerb\" class=\"hnuser\">gingerb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584477\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584531\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanmcdirmid\" class=\"hnuser\">seanmcdirmid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584531\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584193\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dchasson\" class=\"hnuser\">dchasson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584193\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15588384\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AceJohnny2\" class=\"hnuser\">AceJohnny2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15588384\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583651\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583651\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585998\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sctb\" class=\"hnuser\">sctb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585998\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586108\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586108\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586085\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dylan16807\" class=\"hnuser\">Dylan16807<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586085\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15589016\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sctb\" class=\"hnuser\">sctb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15589016\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583685\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dunstad\" class=\"hnuser\">dunstad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583685\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583853\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583853\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584051\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=spacetexas\" class=\"hnuser\">spacetexas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584051\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584962\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=baq\" class=\"hnuser\">baq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584962\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584083\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584083\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584450\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Cyph0n\" class=\"hnuser\">Cyph0n<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584450\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585810\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=electrum\" class=\"hnuser\">electrum<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585810\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15589054\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fasquoika\" class=\"hnuser\">fasquoika<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15589054\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15592644\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Cyph0n\" class=\"hnuser\">Cyph0n<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15592644\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585793\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tytytytytytytyt\" class=\"hnuser\">tytytytytytytyt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585793\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584121\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hasenj\" class=\"hnuser\">hasenj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584121\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584679\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=weberc2\" class=\"hnuser\">weberc2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584679\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585912\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585912\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585960\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sillysaurus3\" class=\"hnuser\">sillysaurus3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585960\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586285\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586285\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586971\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sillysaurus3\" class=\"hnuser\">sillysaurus3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586971\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15589082\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fasquoika\" class=\"hnuser\">fasquoika<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15589082\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15588796\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gizmo385\" class=\"hnuser\">gizmo385<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15588796\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586260\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=weberc2\" class=\"hnuser\">weberc2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586260\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584190\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584190\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584297\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonsterling\" class=\"hnuser\">jonsterling<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584297\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585517\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nevir\" class=\"hnuser\">nevir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585517\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585094\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LambdaComplex\" class=\"hnuser\">LambdaComplex<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585094\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585258\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dylan16807\" class=\"hnuser\">Dylan16807<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585258\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585813\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tytytytytytytyt\" class=\"hnuser\">tytytytytytytyt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585813\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586576\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tangent128\" class=\"hnuser\">tangent128<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586576\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15589415\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hinkley\" class=\"hnuser\">hinkley<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15589415\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15623583\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonsterling\" class=\"hnuser\">jonsterling<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15623583\">on Nov 4, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15601656\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tangent128\" class=\"hnuser\">tangent128<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15601656\">on Nov 1, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585966\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dylan16807\" class=\"hnuser\">Dylan16807<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585966\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585661\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Filligree\" class=\"hnuser\">Filligree<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585661\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585926\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dylan16807\" class=\"hnuser\">Dylan16807<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585926\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585866\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucozade\" class=\"hnuser\">lucozade<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585866\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585204\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kreetx\" class=\"hnuser\">kreetx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585204\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583795\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vladimir-y\" class=\"hnuser\">vladimir-y<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583795\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583788\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=madeofpalk\" class=\"hnuser\">madeofpalk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583788\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583835\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583835\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584017\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=monsieurbanana\" class=\"hnuser\">monsieurbanana<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584017\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584457\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584457\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584513\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584513\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585432\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dylan16807\" class=\"hnuser\">Dylan16807<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585432\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15587118\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15587118\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583797\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=colejohnson66\" class=\"hnuser\">colejohnson66<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583797\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585665\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rexpop\" class=\"hnuser\">rexpop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585665\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583652\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mbrock\" class=\"hnuser\">mbrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583652\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583738\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kreetx\" class=\"hnuser\">kreetx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583738\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584401\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dahauns\" class=\"hnuser\">dahauns<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584401\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584459\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584459\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584861\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dahauns\" class=\"hnuser\">dahauns<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584861\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15587099\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15587099\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15587262\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dahauns\" class=\"hnuser\">dahauns<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15587262\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15587481\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15587481\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15592036\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dahauns\" class=\"hnuser\">dahauns<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15592036\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15592638\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"400\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15592638\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584230\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mbrock\" class=\"hnuser\">mbrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584230\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585280\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kreetx\" class=\"hnuser\">kreetx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585280\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585405\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mbrock\" class=\"hnuser\">mbrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585405\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585839\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kreetx\" class=\"hnuser\">kreetx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585839\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586595\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mbrock\" class=\"hnuser\">mbrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586595\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15591760\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kreetx\" class=\"hnuser\">kreetx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15591760\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15593105\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mbrock\" class=\"hnuser\">mbrock<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15593105\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583765\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TeMPOraL\" class=\"hnuser\">TeMPOraL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583765\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584258\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lhnz\" class=\"hnuser\">lhnz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584258\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584367\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584367\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15591623\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15591623\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15590819\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikpukinskis\" class=\"hnuser\">erikpukinskis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15590819\">on Oct 31, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583746\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CapacitorSet\" class=\"hnuser\">CapacitorSet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583746\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583842\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583842\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584037\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=monsieurbanana\" class=\"hnuser\">monsieurbanana<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584037\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584070\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584070\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15585849\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alkonaut\" class=\"hnuser\">alkonaut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585849\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15584062\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Double_a_92\" class=\"hnuser\">Double_a_92<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15584062\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15583800\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=avodonosov\" class=\"hnuser\">avodonosov<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15583800\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr  coll\" id=\"15585346\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=igravious\" class=\"hnuser\">igravious<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15585346\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15586295\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=psygnisfive\" class=\"hnuser\">psygnisfive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15586295\">on Oct 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:51:51+0000","updated_at":"2018-11-07T16:51:51+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":552,"title":"Eclipse GUI broken \/ Applications & Desktop Environments \/ Arch Linux Forums","url":"https:\/\/bbs.archlinux.org\/viewtopic.php?id=200053","content":"<p>you can try this workaround :<\/p><p>edit the file \/usr\/lib\/eclipse\/eclipse.ini,<\/p><p>and add these 2 lines before the line \"--launcher.appendVmargs\" :<\/p><div class=\"codebox\"><pre>--launcher.GTK_version\n2<\/pre><\/div><p>after this modification it's the version \"2\" of gtk launcher who will be used, the GUI will be much better and the background of tooltips will have a correct color<\/p><p class=\"postedit\"><em>Last edited by Potomac (2015-07-21 22:50:20)<\/em><\/p>","created_at":"2018-11-07T16:51:50+0000","updated_at":"2018-11-07T16:51:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"bbs.archlinux.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":550,"title":"Confused About REM and EM?","url":"https:\/\/j.eremy.net\/confused-about-rem-and-em\/","content":"<p>&#13;\n\t\tFebruary 24, 2014<\/p><p class=\"font_large\">REM can be confusing, especially without a solid understanding of its partner EM and their archvillain, the\u00a0PX.<\/p><h2>Relative Units<\/h2><p>Both <strong><em>rem<\/em><\/strong> and <strong><em>em<\/em><\/strong> are relative units, <strong><em>px<\/em><\/strong> is not. Before considering <strong><em>rem<\/em><\/strong>, it\u2019s important to understand the relationship between <strong><em>em<\/em><\/strong>, markup and inheritance.<strong><\/strong><\/p><p class=\"break_medium\">Below, the example demonstrates how each nested child assumes the parent is 1em(100%). Thus children inherit size by scaling in relation to the parent font size.<\/p><div class=\"bleed_left_large_100 media_xs_bleed break_medium\"><div class=\"cols\"><div class=\"col_half\"><p class=\"text_center text_light font_small font_alt\"><em>EM values inherit from their parent<\/em><\/p><p class=\"codepen\" data-height=\"314\" data-theme-id=\"1457\" data-slug-hash=\"rCcIh\" data-default-tab=\"result\" data-animations=\"run\" data-editable=\"\" data-embed-version=\"2\">\nSee the Pen <a href=\"http:\/\/codepen.io\/jeremychurch\/pen\/rCcIh\">rCcIh<\/a> by Jeremy Church (<a href=\"http:\/\/codepen.io\/jeremychurch\">@jeremychurch<\/a>) on <a href=\"http:\/\/codepen.io\">CodePen<\/a>.1457<\/p><\/div><div class=\"col_half\"><p class=\"text_center text_light font_small font_alt\"><em>PX values do not inherit<\/em><\/p><p class=\"codepen\" data-height=\"314\" data-theme-id=\"1457\" data-slug-hash=\"dlyqw\" data-default-tab=\"result\" data-animations=\"run\" data-editable=\"\" data-embed-version=\"2\">\nSee the Pen <a href=\"http:\/\/codepen.io\/jeremychurch\/pen\/dlyqw\">dlyqw<\/a> by Jeremy Church (<a href=\"http:\/\/codepen.io\/jeremychurch\">@jeremychurch<\/a>) on <a href=\"http:\/\/codepen.io\">CodePen<\/a>.1457<\/p><\/div><\/div><\/div><div class=\"aside_left\"><p>Visit <a href=\"http:\/\/pxtoem.com\/\">pxtoem.com<\/a> for conversion tables and calculator.<\/p><\/div><p>While the value remains 0.773em, the actual font size is calculated at 77.3% of its direct parent, which in turn scales from its parent. This continues up the DOM tree whenever a parent has a defined <code>font-size<\/code>.<\/p><p class=\"font_small\">In the example, several elements, each with a font size, are nested for a visual effect. You can see it\u2019s not a good practice, because the compound inheritance creates unwanted results. However, you shouldn\u2019t have to worry about this if your CSS and markup are modular to begin with.<\/p><h2>Save Lives with EM<\/h2><div class=\"aside_right\"><p>You can choose a different ratio\/scale and calculate your own values on <a href=\"http:\/\/type-scale.com\/\">Type Scale<\/a>.<\/p><\/div><p class=\"break_medium\">Well, maybe not lives, but it will save <em>lines<\/em> \u2026 of code. The following examples do the same thing; update font sizes and padding within a media query. The initial values are calculated by incrementing 1em(16px) at a 1:1.2 ratio.<\/p><div class=\"bleed_right_large_100 break_medium\"><div class=\"cols\"><div class=\"col_half\"><p class=\"text_center text_light font_small font_alt\"><em>EM scales by updating one value<\/em><\/p><pre>html { font-size: 1em; }&#13;\nh1 { font-size: 2.074em; }&#13;\nh2 { font-size: 1.728em; }&#13;\nh3 { font-size: 1.44em; }&#13;\nh4 { font-size: 1.2em; }&#13;\nsmall { font-size: 0.833em; }&#13;\n.box { padding: 1.25em; }&#13;\n@media screen and (min-width: 1400px) {&#13;\n\u00a0 html { font-size: 1.25em; }&#13;\n}<\/pre><\/div><div class=\"col_half\"><p class=\"text_center text_light font_small font_alt\"><em>Or, recalculate every PX value<\/em><\/p><pre>html { font-size: 16px; }&#13;\nh1 { font-size: 33px; }&#13;\nh2 { font-size: 28px; }&#13;\nh3 { font-size: 23px; }&#13;\nh4 { font-size: 19px; }&#13;\nsmall { font-size: 13px; }&#13;\n.box { padding: 20px; }&#13;\n@media screen and (min-width: 1400px) {&#13;\n\u00a0 html { font-size: 20px; }&#13;\n\u00a0 h1 { font-size: 41px; }&#13;\n\u00a0 h2 { font-size: 35px; }&#13;\n\u00a0 h3 { font-size: 29px; }&#13;\n\u00a0 h4 { font-size: 24px; }&#13;\n\u00a0 small { font-size: 17px; }&#13;\n  .box { padding: 25px; }&#13;\n}<\/pre><\/div><\/div><\/div><p>There\u2019s only one <strong><em>em<\/em><\/strong> value to overwrite in the media query, because <strong><em>em<\/em><\/strong> inherits and scales relative to its parent (<code>html<\/code> in this case) font size, similar to the way vector paths scale proportionally.<\/p><p>The code is doubled when updating <strong><em>px<\/em><\/strong>, because each value has to be recalculated and defined in the media query.<\/p><h2>REM as in Root EM<\/h2><p>While <strong><em>em<\/em><\/strong> is relative to the font size of its direct or nearest parent, <strong><em>rem<\/em><\/strong> is only relative to the <code>html<\/code> (root) <code>font-size<\/code>. I like to think of it as a reset. If a style sheet is built in a modular fashion, then <strong><em>rem<\/em><\/strong> shouldn\u2019t be needed very often, but it can be handy at times.<\/p><p class=\"break_medium\">For example, if you\u2019re trying to achieve consistent spacing without extra markup, <strong><em>rem<\/em><\/strong> can be used to define the padding and margins.<\/p><div class=\"bleed_left_large_100 media_xs_bleed break_medium\"><div class=\"cols\"><div class=\"col_half\"><p class=\"text_center text_light font_small font_alt\"><em>EM will scale padding and margins too<\/em><\/p><p class=\"codepen\" data-height=\"394\" data-theme-id=\"1457\" data-slug-hash=\"AlxHk\" data-default-tab=\"result\" data-animations=\"run\" data-editable=\"\" data-embed-version=\"2\">\nSee the Pen <a href=\"http:\/\/codepen.io\/jeremychurch\/pen\/AlxHk\">AlxHk<\/a> by Jeremy Church (<a href=\"http:\/\/codepen.io\/jeremychurch\">@jeremychurch<\/a>) on <a href=\"http:\/\/codepen.io\">CodePen<\/a>.1457<\/p><\/div><div class=\"col_half\"><p class=\"text_center text_light font_small font_alt\"><em>Use REM for consistent padding and margins<\/em><\/p><p class=\"codepen\" data-height=\"394\" data-theme-id=\"1457\" data-slug-hash=\"qjEBs\" data-default-tab=\"result\" data-animations=\"run\" data-editable=\"\" data-embed-version=\"2\">\nSee the Pen <a href=\"http:\/\/codepen.io\/jeremychurch\/pen\/qjEBs\">qjEBs<\/a> by Jeremy Church (<a href=\"http:\/\/codepen.io\/jeremychurch\">@jeremychurch<\/a>) on <a href=\"http:\/\/codepen.io\">CodePen<\/a>.1457<\/p><\/div><\/div><\/div><p><strong><em>PX<\/em><\/strong> can be used for consistent padding and margins too, but it doesn\u2019t scale across media queries like <strong><em>em<\/em><\/strong> and <strong><em>rem<\/em><\/strong>.<\/p><h2>Conclusion<\/h2><p>I use <strong><em>em<\/em><\/strong> for nearly everything, <strong><em>rem<\/em><\/strong> occasionally for padding or margins, but only in a pinch, and <strong><em>px<\/em><\/strong> some times for borders.<\/p><p>For the most part, I want children to inherit size. If a sidebar is going to be a smaller font size, then I expect all its children to resize proportionally without having to calculate and redefine individual selectors.<\/p><p class=\"break_medium\">I struggle to think of a good reason to use <strong><em>rem<\/em><\/strong> for <code>font-size<\/code> values, but I\u2019m sure there are exceptions. If I\u2019m trying to reset a font size with <strong><em>rem<\/em><\/strong>, it\u2019s probably a sign my CSS is not very modular and due for a refactor. \u00a0\u220e<\/p><div class=\"font_small\"><h3>Resources<\/h3><p>My favorite <strong><em>em<\/em><\/strong> calculator is <a href=\"http:\/\/pxtoem.com\/\">pxtoem.com<\/a>. In addition to calculating <strong><em>em<\/em><\/strong> or <strong><em>px<\/em><\/strong> values, there\u2019s a handy table of the two which really helps to see the relationship.<\/p><p><strong><em>REM<\/em><\/strong> is the same as <strong><em>em<\/em><\/strong> except it\u2019s always calculated from the html font size. So, just use the same <a href=\"http:\/\/pxtoem.com\/\">pxtoem.com<\/a> calculator to figure out its relation to the root size.<\/p><p>Article in Espa\u00f1ol: <a href=\"http:\/\/www.jesusherman.com\/blog\/te-confunden-rem-y-em\/\">\u00bfTe confunden REM y EM?<\/a><\/p><p>Article in Italian: <a href=\"http:\/\/www.laboratoriocss.it\/sei-confuso-su-rem-ed-em.html\">Sei confuso su REM ed EM?<\/a><\/p><\/div><footer>&#13;\n&#13;\n\t\t&#13;\n&#13;\n    &#13;\n&#13;\n\t\t<hr \/><div class=\"table_padded\"><div class=\"table_cell\"><div class=\"font_small font_alt\"><h3 class=\"remove_margin_top\">&#13;\n            <a href=\"https:\/\/j.eremy.net\/about\">Jeremy Church<\/a>&#13;\n          <\/h3><p>I'm a front-end developer. I built a thing: <a href=\"http:\/\/type-scale.com\/\">type-scale.com<\/a>.&#13;\n          <br \/>I work on the internet in Denver, and motorcycles in\u00a0Brighton.<\/p><ul class=\"inline_list\"><li><a href=\"https:\/\/j.eremy.net\/about\" class=\"remove_border_bottom\">\u00a0 About<\/a><\/li>\n\n\t\t\t\t\t\t<li><a href=\"https:\/\/twitter.com\/jeremybchurch\" class=\"remove_border_bottom\">\u00a0 Twitter<\/a><\/li>\n\n\t\t\t\t\t\t<li><a href=\"https:\/\/j.eremy.net\/feed\" class=\"remove_border_bottom\">\u00a0 Feed<\/a><\/li>\n\t\t\t\t\t<\/ul><\/div><\/div><\/div>&#13;\n\t<\/footer>","created_at":"2018-11-07T16:51:49+0000","updated_at":"2018-11-07T16:51:49+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"j.eremy.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":551,"title":"Lob - APIs to Make the World Programmable","url":"https:\/\/lob.com\/","content":"<div class=\"row\"><p>\n        <\/p><h3>Lob's software platform automates age-old offline business processes in a modern, intelligent, and technology-forward way.<\/h3><\/div><div class=\"row home-products\"><div><img src=\"https:\/\/s3-us-west-2.amazonaws.com\/lob-assets\/nav-letters.svg\" alt=\"image\" \/><h2>Print &amp; Mail<\/h2><p>Programmatically send personalized postcards, letters, and checks to your customers with comprehensive per-piece mail tracking and analytics.<\/p><p><a href=\"https:\/\/lob.com\/products\/print-mail\/postcards\" class=\"analytics-cta\" data-location=\"Homepage Top\">Learn More \u00bb<\/a><\/p><\/div><div><img src=\"https:\/\/s3-us-west-2.amazonaws.com\/lob-assets\/nav-verify.svg\" alt=\"image\" \/><h2>Address Verification<\/h2><p>Use our USPS-certified, global address verification APIs to keep your customer database accurate and ensure that your products and mail reach the right person every time.<\/p><p><a href=\"https:\/\/lob.com\/products\/address-verification\" class=\"analytics-cta\" data-location=\"Homepage Top\">Learn More \u00bb<\/a><\/p><\/div><\/div>","created_at":"2018-11-07T16:51:49+0000","updated_at":"2018-11-07T16:51:49+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"lob.com","preview_picture":"https:\/\/s3-us-west-2.amazonaws.com\/lob-assets\/LobLogoWhite.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":548,"title":"Distill \u2014 Latest articles about machine learning","url":"https:\/\/distill.pub\/","content":"<div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2018\/editorial-update\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Distill Update 2018<\/h2><p>An Update from the Editorial Team<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2018\/differentiable-parameterizations\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Differentiable Image Parameterizations<\/h2><p>A powerful, under-explored tool for neural network visualizations and art.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2018\/feature-wise-transformations\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Feature-wise transformations<\/h2><p>A simple and surprisingly effective family of conditioning mechanisms.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2018\/building-blocks\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>The Building Blocks of Interpretability<\/h2><p>Interpretability techniques are normally studied in isolation. We explore the powerful interfaces that arise when you combine them\u2009\u2014\u2009and the rich structure of this combinatorial space.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2017\/ctc\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Sequence Modeling with CTC<\/h2><p>A visual guide to Connectionist Temporal Classification, an algorithm used to train deep neural networks in speech recognition, handwriting recognition and other sequence problems.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2017\/feature-visualization\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Feature Visualization<\/h2><p>How neural networks build up their understanding of images<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2017\/momentum\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Why Momentum Really Works<\/h2><p>We often think of optimization with momentum as a ball rolling down a hill. This isn\u2019t wrong, but there is much more to the story.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2016\/handwriting\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Experiments in Handwriting with a Neural Network<\/h2><p>Several interactive visualizations of a generative model of handwriting. Some are fun, some are serious.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2016\/deconv-checkerboard\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Deconvolution and Checkerboard Artifacts<\/h2><p>When we look very closely at images generated by neural networks, we often see a strange checkerboard pattern of artifacts.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2016\/misread-tsne\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>How to Use t-SNE Effectively<\/h2><p>Although extremely useful for visualizing high-dimensional data, t-SNE plots can sometimes be mysterious or misleading.<\/p><\/div><\/div><div><div class=\"thumbnail\"><img src=\"https:\/\/distill.pub\/2016\/augmented-rnns\/thumbnail.jpg\" alt=\"image\" \/><\/div>\n        <div class=\"description\"><h2>Attention and Augmented Recurrent Neural Networks<\/h2><p>A visual overview of neural attention, and the powerful extensions of neural networks being built on top of it.<\/p><\/div><\/div>","created_at":"2018-11-07T16:51:48+0000","updated_at":"2018-11-07T16:51:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"distill.pub","preview_picture":"http:\/\/distill.pub\/preview.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":549,"title":"posativ\/isso","url":"https:\/\/github.com\/posativ\/isso","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>Isso \u2013 <em>Ich schrei sonst<\/em> \u2013 is a lightweight commenting server written in\nPython and JavaScript. It aims to be a drop-in replacement for\n<a href=\"http:\/\/disqus.com\" rel=\"nofollow\">Disqus<\/a>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/237cabeb1e1ba756a9dc2577da5b6da5b5fb35bf\/687474703a2f2f706f73617469762e6f72672f253745746d702f6973736f2d73616d706c652e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/237cabeb1e1ba756a9dc2577da5b6da5b5fb35bf\/687474703a2f2f706f73617469762e6f72672f253745746d702f6973736f2d73616d706c652e706e67\" alt=\"Isso in Action\" data-canonical-src=\"http:\/\/posativ.org\/%7Etmp\/isso-sample.png\" \/><\/a><\/p>\n<p>See <a href=\"http:\/\/posativ.org\/isso\/\" rel=\"nofollow\">posativ.org\/isso<\/a> for more details.<\/p>\n<\/article>","created_at":"2018-11-07T16:51:48+0000","updated_at":"2018-11-07T16:51:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/337602?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":546,"title":"Linux Inside \u2013 How the Linux Kernel Works","url":"https:\/\/0xax.gitbooks.io\/linux-insides\/content\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:51:32+0000","updated_at":"2018-11-07T16:51:32+0000","annotations":[],"reading_time":0,"domain_name":"0xax.gitbooks.io","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":547,"title":"My $169 development Chromebook","url":"https:\/\/blog.lessonslearned.org\/building-a-more-secure-development-chromebook\/","content":"<time datetime=\"2017-07-26\">26 Jul 2017<\/time><section class=\"post-content\"><p><em>How Chrome OS, Termux, YubiKey &amp; Duo Mobile make for great usable security<\/em><\/p><p><a href=\"https:\/\/user-images.githubusercontent.com\/326990\/28650249-715b8460-7248-11e7-993f-38e56cc56022.png\"><img title=\"final desktop\" alt=\"final desktop\" src=\"https:\/\/user-images.githubusercontent.com\/326990\/28650249-715b8460-7248-11e7-993f-38e56cc56022.png\" \/><\/a><\/p><hr \/><p>I'll give this a proper update in the coming days, but much of this post is no longer needed. First class support for native VMs and containers have landed in Dev &amp; Canary channels for Chrome OS \u2014 including budget models like the Samsung Chromebook 3 discussed below. Full Wayland support, bi-directional copy\/paste between guest &amp; host, and VM network passthrough to the Chrome browser supported. The default VM is a full-on Debian 9 build. Some highlights in thread and replies here: <a href=\"https:\/\/twitter.com\/kennwhite\/status\/1028505652749062144?s=21\">https:\/\/twitter.com\/kennwhite\/status\/1028505652749062144?s=21<\/a><\/p><h3 id=\"backstory\">Backstory<\/h3><p>In the last year while talking to respected security-focused engineers &amp; developers, I've come to fully appreciate Google's <a href=\"https:\/\/www.chromium.org\/chromium-os\/chromiumos-design-docs\/security-overview\">Chrome OS design<\/a>. The architecture benefited from a modern view of threat modeling and real-world attacks. For example, Trusted Platform Module (TPM) hardware chips are built into every Chromebook and deeply incorporated into the OS. The design documents go into some detail on the specific protections that TPM provides, particularly around <a href=\"https:\/\/www.chromium.org\/developers\/design-documents\/tpm-usage\">critical encryption functions<\/a>.<\/p><p>I also learned that Chromebook is the daily driver for many of Google's own senior developers and security engineers. In short, the combination of the underlying Chromebook hardware with the OS architecture makes for a pretty compelling secure development environment.<\/p><p>Fast forward to March. In the past, I'd never given much thought to bringing my notebook with me while on travel -- it was just something I did as a matter of course. Now, though, with the <a href=\"http:\/\/money.cnn.com\/2017\/03\/21\/technology\/laptop-ban-safety\/index.html\">changing landscape<\/a> of airline restrictions and seemingly ad hoc draconian rules for what can and can't be brought on board, I've had to seriously reconsider my routine. There is <em>zero<\/em> chance I am willing to bag check my beloved Macbook Pro or Purism <a href=\"https:\/\/puri.sm\/products\/\">Librem<\/a>, especially given carriers' long history of indifference to ground crews playing hacky-sack with luggage. Who can forget <a href=\"https:\/\/www.youtube.com\/watch?v=5YGc4zOqozo\">United Breaks Guitars<\/a>? And let's not even get into <a href=\"http:\/\/www.cnn.com\/2015\/04\/13\/us\/airport-luggage-theft\/index.html\">brazen theft<\/a>.<\/p><blockquote>\n  <p><em>The point of this exercise is to retain the hardened posture of the platform <strong>and<\/strong> have a flexible, safe development environment without depending on the crutch of privileged access.<\/em><\/p>\n<\/blockquote><p>It's pretty neat to consider the possibility of pre-travel \"power washing\" (resetting everything clean to factory settings) on an inexpensive Chromebook and later securely restore over the air once at my destination. Since there is a <em>wide<\/em> range in Chromebook prices, the engineering challenge here was to find something powerful enough to comfortably use exclusively for several days of coding, writing, and presenting, but also cheap enough that should it get lost\/stolen\/damaged, I wouldn't lose too much sleep. The threat model here does not include recovery from physical tampering; if the machine were somehow confiscated or otherwise out of my custody, I could treat it as a burner and move on.<\/p><p>After two weeks <a href=\"https:\/\/en.wikipedia.org\/wiki\/Eating_your_own_dog_food\">dogfooding<\/a> this as my exclusive machine, here's what I came up with. And honestly, I'm pretty happy with the results. But it wasn't easy to get here. Nearly every how-to and blog post I've found on \"Chromebooks for developers\" essentially starts with either: \"Boot into Developer Mode\" or \"Install Debian\/Ubuntu as the main OS\". <strong>I'll just say it: This is bad advice.<\/strong> It would be akin to recommending that friends jailbreak their shiny new iPhone.<\/p><p>You're obviously free to do as you wish with your own gear, but recognize that at <em>Step 1<\/em>, you'll have lost most of the core security features of Chromebook, not to mention virtually inviting bad development habits. As far as Debian\/Ubuntu (and crouton), that's fine as far as it goes, but then you don't end up with a Chromebook, just a cheap mini-notebook with flaky drivers. The whole point of this exercise is to retain the hardened posture of the platform <em>and<\/em> have a flexible, safe development environment <em>without<\/em> depending on the crutch of privileged access. <\/p><h3 id=\"goals\">Goals<\/h3><ul><li>A <em>very<\/em> modest budget (ideally &lt;$200 USD, all in)<\/li>\n<li>8-10 hour real-world battery life<\/li>\n<li>A first-class, <em>local\/non-connected<\/em> dev environment for my current toolchain:\n<ul><li>go, c\/c++, node.js, git, curl, nginx, postgres...<\/li><\/ul><\/li>\n<li>Standard network tools (openssl cli, nmap, tcpdump...)<\/li>\n<li>Strong hardware-backed verification &amp; mutual auth<\/li>\n<li>Open source, actively developed software whenever possible<\/li>\n<li>Bonus: secure messaging, voice, video chat, Netflix<\/li>\n<\/ul><h3 id=\"antifeatures\">Anti-Features<\/h3><p>The finished environment should <em>not<\/em>:<\/p><ul><li>bypass core OS security  <\/li>\n<li>disable Verified Boot  <\/li>\n<li>require sudo\/root for dev environment  <\/li>\n<li>rely on Crouton or Ubuntu  <\/li>\n<li>require (or assume) an always-on network  <\/li>\n<\/ul><h3 id=\"supportedgear\">Supported Gear<\/h3><p>Google announced a major development in the Chromebook roadmap recently and it seems to have been overlooked by many in the industry: over the past few months, several models have been updated to work with Play Store apps. In theory, this will open the door to many of the 3+ million Android apps. In practice, there's still a fair amount of work to do to \"optimize\" existing app display layouts and fix other UX quirks. But crucially for our purposes, this opens the way to a killer app called Termux.<\/p><p>The official list of Chromebooks supporting Android apps (and the timeline for those planned in 2017) is here: <br \/><a href=\"https:\/\/sites.google.com\/a\/chromium.org\/dev\/chromium-os\/chrome-os-systems-supporting-android-apps\">https:\/\/sites.google.com\/a\/chromium.org\/dev\/chromium-os\/chrome-os-systems-supporting-android-apps<\/a><\/p><p>While never explicitly stated, I take inclusion on this list as a rough proxy for commitment by manufacturers (and Google product engineering) to support these models.<\/p><p>After reading an absurd number of reviews and forum posts, I eventually settled on the Samsung Chromebook 3. It's lightweight (a little over 2 lbs), comes with 4GB RAM and a 16GB solid state drive (expandable via MicroSD to 256GB), and uses a low-power Intel x86-64 dual core N3060 1.6\/2.4 GHz processor. Other goodies include USB 2\/3 ports, HDMI, 720p webcam, and a water-resistant keyboard.<\/p><p>You would definitely notice the performance boost if you bumped up to 8GB of memory or an m5\/i5 CPU, but prices at that tier <a href=\"https:\/\/www.amazon.com\/gp\/product\/B01EYGSM2I\/ref=as_li_tl?ie=UTF8&amp;tag=amzsearch04b-20&amp;camp=1789&amp;creative=9325&amp;linkCode=as2&amp;creativeASIN=B01EYGSM2I&amp;linkId=c5841ad90c094ba8bb5e6edfc2103fe0\">start at $800<\/a> and rise quickly (see note<a href=\"#amazon-links\">\u00b9<\/a> below on Amazon). There's no shortage of high end Chromebooks available, but in my testing of the \"serviceable, but won't-cry-if-I-lose-it\" category, 4GB of RAM and a non-ARM processor seem to be the sweet spot. If you are tempted to save $40 and opt for the super cheap 2GB models, don't expect to open more than a handful of browser tabs before hitting a tar pit.<\/p><h3 id=\"hardware\">Hardware<\/h3><ul><li>Samsung Chromebook 3 <a href=\"https:\/\/www.amazon.com\/gp\/product\/B01N5P6TJW\/ref=as_li_tl?ie=UTF8&amp;tag=amzsearch04b-20&amp;camp=1789&amp;creative=9325&amp;linkCode=as2&amp;creativeASIN=B01N5P6TJW&amp;linkId=ec09e6b81b11151c1f114e0c8030c2f9\">($169 on Amazon)<\/a><\/li>\n<li>FIDO-certified U2F security key. I recommend the YubiKey U2F blue key <a href=\"https:\/\/www.amazon.com\/gp\/product\/B00NLKA0D8\/ref=as_li_tl?ie=UTF8&amp;tag=amzsearch04b-20&amp;camp=1789&amp;creative=9325&amp;linkCode=as2&amp;creativeASIN=B00NLKA0D8&amp;linkId=3b0326c6bfdbcdb7d70685153fb9a0d8\">($17 on Amazon)<\/a><\/li>\n<\/ul><h3>Prerequisites<\/h3><ul><li>A good 2-Factor Authentication app (installed on your phone, not your Chromebook). Personally, I trust the <a href=\"https:\/\/duo.com\/product\/trusted-users\/two-factor-authentication\/duo-mobile\">Duo Mobile app<\/a>, available for free on iOS and Android. Many people also like Google Authenticator.<\/li>\n<\/ul><h3 id=\"thebuildauthenticationandsecurity\">The Build: Authentication and Security<\/h3><ol><li>If necessary, wipe the Chromebook to factory settings via <a href=\"https:\/\/support.google.com\/chromebook\/answer\/183084?hl=en\">Powerwash<\/a>  <\/li>\n<li>Boot and set up a new Google account to manage your device (this will help segment your data, and if you need it you can still access your existing gmail accounts). Choose: Other Options &gt; Create New Account. You'll likely be asked for a recovery\/SMS phone number, but you can remove it in a few minutes.  <\/li>\n<li>Ensure the latest OS &amp; app updates are installed: <br \/><ul><li>Account photo \/ gear icon \/ Settings dropdown \/ About Chrome OS \/ Check for Updates<\/li><\/ul><\/li>\n<li>Set up mutual authentication with the YubiKey. Open Security settings in the new Google account (<a href=\"https:\/\/myaccount.google.com\/security#signin\">https:\/\/myaccount.google.com\/security#signin<\/a>) and click <em>2-Step Verification<\/em>. You'll be prompted for your password. Follow the prompts for \"Add a Security Key\". Once confirmed by logging out then back in using your YubiKey, continue to the next step below.  <\/li>\n<li>Add a 2FA authentication fallback option. Navigate back to the Settings menu above, then: <br \/><ul><li>2-Step Verification \/ Authenticator App \/ Setup (choose Google Authenticator; the QR codes are universal) \/ Next \/ QR code should display<\/li>\n<li>On your phone, open Duo Mobile then: Begin Setup \/ Scan barcode \/ Next<\/li>\n<li>On your Chromebook, type the code you see in Duo Mobile into the text box \/ Verify \/ Done<\/li><\/ul><\/li>\n<li>Verify that the authenticator is working correctly. The 2FA authenticator verification is an access safety mechanism should you lose the YubiKey: <br \/><ul><li>Log out then log back in. <em>Ignore<\/em> the prompt to insert your security key, and instead click on the link at the bottom of the login page <em>Having Trouble?<\/em>. Choose <em>Get a verification code from the Google Authenticator app<\/em>.<\/li>\n<li>On Duo Mobile app, click the Google account key icon and enter the 2FA code into the browser.<\/li>\n<li>Save or print out the backup codes, should you lose both your mobile phone (or at least the secret TOTP seeds in your 2FA app) <em>and<\/em> your YubiKey.<\/li>\n<li>Note: for both the 2-Step Verification and Authenticator login paths, you may want to consider unchecking <em>Don't ask again on this computer<\/em>.<\/li><\/ul><\/li>\n<li>With two step verification set up, it is safe to remove the phone number you supplied earlier from the account recovery menu<\/li>\n<\/ol><h3 id=\"thebuildprivacycloudsync\">The Build: Privacy &amp; Cloud Sync<\/h3><ol><li>Set privacy and tracking configuration. Depending on your use case, you will probably want to limit the ad data that Google collects. <br \/><ul><li>Navigate to <a href=\"https:\/\/myaccount.google.com\">https:\/\/myaccount.google.com<\/a> (the 3x3 box in the top right corner in Gmail) &gt; Personal Info &amp; Privacy &gt; Your personal info &gt; Ad settings<\/li><\/ul><\/li>\n<li>Set cloud sync preferences. <br \/><ul><li>Chromebook settings &gt; People &gt; Sync<\/li>\n<li>I recommend enabling: <em>Apps<\/em>, <em>Extensions<\/em>, and <em>Settings<\/em>, but not <em>Autofill, History,<\/em> and <em>Open Tabs<\/em>. Also \u2014 crucially \u2014 note that allowing <em>Apps<\/em> and <em>Extension<\/em> sync will likely mirror sensitive content including credentials, keys, API tokens, passwords &amp; password manager vaults, and other secrets stored locally on your notebook. This could be a huge convenience or utterly catastrophic if left unencrypted, depending on your threat model and risk tolerance. You mileage will definitely vary.<\/li><\/ul><\/li>\n<li>If you plan to use Chrome to manage passwords, you should probably set a Sync Passphrase so that Google servers won't store your <a href=\"https:\/\/support.google.com\/chrome\/answer\/1181035\">passwords in plaintext<\/a>.<\/li>\n<\/ol><h3 id=\"androidapps\">Android Apps<\/h3><ol><li>Set Update Channel from <em>Stable<\/em> to <em>Beta<\/em> (some models already officially support Android apps in <em>Stable<\/em>) <br \/><ul><li>Profile picture in lower right &gt; Settings &gt; About Chrome OS (bottom of drop down menu) &gt; Check for Updates.<\/li>\n<li>Settings &gt; About Chrome OS &gt; Detailed Build Information &gt; Channel &gt; Change Channel &gt; Beta &gt; Change Channel &gt; Restart<\/li><\/ul><\/li>\n<li>Enable Chromebook Android apps <br \/><ul><li>Activate Play Store on your Chromebook: Bottom profile photo &gt; gear icon &gt; Settings &gt; Play Store &gt; Turn on <\/li>\n<li>Start the (now) pre-installed \"Play Store (Beta)\" <em>Android app<\/em>, not the Play Store <em>web site<\/em> to accept the terms. Once that's done, your device should be registered. This part is a little wonky -- you can't start in the Play Store web site until this is done, but once it is, you don't have to use the clunky Play Store Beta app.<\/li>\n<li>Navigate to Play Store (web) Settings: <a href=\"https:\/\/play.google.com\/settings\/\">https:\/\/play.google.com\/settings\/<\/a>. Alternatively, from Gmail: top right 3x3 square icon &gt; Play Store &gt; top right gear icon &gt; Settings<\/li>\n<li>Verify your device is registered, e.g.: \"No carrier Samsung Chromebook 3\" and give it a memorable name. <em>If you hit any problems, see the <a href=\"https:\/\/support.google.com\/chromebook\/answer\/7021273?hl=en\">official detailed how-to<\/a> on enabling Android apps on Chromebook<\/em><\/li><\/ul><\/li>\n<\/ol><p><img title=\"google play registration\" alt=\"google play registration\" src=\"https:\/\/user-images.githubusercontent.com\/326990\/28653058-7161bb58-7259-11e7-9c8c-8159db15660b.png\" \/><\/p><h3 id=\"thetermuxsecretsauce\">The Termux Secret Sauce<\/h3><p><img title=\"termux\" alt=\"termux\" src=\"https:\/\/user-images.githubusercontent.com\/326990\/28653560-3261feb0-725c-11e7-915f-702d8d0cff26.png\" \/><\/p><p>This is the killer app that makes flexible dev shell access possible as an unprivileged user. Termux is a terrific project developed &amp; led by Fredrik Fornwall and has been under active development for the past two years (<a href=\"https:\/\/github.com\/termux\/termux-app\">https:\/\/github.com\/termux\/termux-app<\/a>). It runs in non-root userspace and provides a chroot-like option to present a conventional linux filesystem. In most respects it feels like a modern Debian-based distro, with the latest i686 stable versions of widely used development &amp; system tools including:<\/p><ul><li>apache 2.4<\/li>  \n<li>bash 4.4<\/li>  \n<li>clang 4.0.1<\/li>  \n<li>erlang 19.3<\/li>  \n<li>git 2.13<\/li>  \n<li>gnupg 2.1<\/li>  \n<li>golang 1.8.3<\/li>  \n<li>libgcc 4.9<\/li>  \n<li>libsodium 1.0.13<\/li>  \n<li>lighttpd 1.4<\/li>  \n<li>llvm 4.0<\/li>  \n<li>lua 5.3<\/li>  \n<li>mariadb 10.2<\/li>  \n<li>mosh 1.3<\/li>  \n<li>nginx 1.12<\/li>  \n<li>nmap 7.50<\/li>  \n<li>nodejs 6.11<\/li>  \n<li>python 2.7.13 &amp; 3.6.1<\/li>  \n<li>redis 3.2.9<\/li>  \n<li>ruby 2.4.1<\/li>  \n<li>sqlite 3.19<\/li>  \n<li>sslscan 1.11<\/li>  \n<li>tcpdump 4.8<\/li>  \n<li>tor 0.3.0.9<\/li>  \n<li>transmission 2.92<\/li>  \n<\/ul><p>Install the Termux Android app from Play Store: <br \/><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.termux\">https:\/\/play.google.com\/store\/apps\/details?id=com.termux<\/a>.<\/p><p>Start up Termux (see Epilog below for customizing) and install base packages :<\/p><pre>    apt update   \n    apt upgrade\n    apt install coreutils\n    pkg upgrade\n    pkg install termux-tools proot util-linux net-tools \n    pkg install openssh tracepath tree git\n<\/pre><p>If Go or Node are your jam:<\/p><pre>    pkg install golang nodejs \n<\/pre><p>For the full list of current packages (Go, Node, Python, etc):<\/p><pre>    pkg list-all \n<\/pre><p>Create symlinks &amp; permissions for typical Android filesystems <br \/>(Downloads, Audio, Images, SD Card, etc). This can safely be run again if something goes wrong:<\/p><pre>    termux-setup-storage\n<\/pre><h3 id=\"setupshell\">Set up shell<\/h3><ol><li><p>Install a desktop SSH app. I recommend <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/firessh\/mcognlamjmofcihollilalojnckfiajm\">FireSSH<\/a>; GitHub source <a href=\"https:\/\/github.com\/mimecuvalo\/firessh\">here<\/a>, or Google's terminal emulator &amp; SSH client based on NaCl, <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/secure-shell\/pnhechapfaindjhompbnflcldabbghjo\">NaSSH<\/a> app; GitHub source <a href=\"https:\/\/chromium.googlesource.com\/apps\/libapps\/+\/master\/nassh\/doc\/FAQ.md\">here<\/a>.<\/p>\n<p>Invoke chroot environment (do a pgrep -f on chroot to ensure you don't call this multiple times; see Epilog below for setting this to autostart on boot). This is a <strong>critical<\/strong> step:<\/p>\n<pre>termux-chroot\n<\/pre><\/li>\n<li><p>From Termux, create a local ssh keypair (assumes you're in chroot):<\/p>\n<pre>mkdir -p \/storage\/emulated\/0\/Download\/devbox\/ssh   \nln -s \/storage\/emulated\/0\/Download\/devbox \/home\/devbox\nssh-keygen -t rsa -b 4096 -C \"your_email@example.com\"\n<\/pre>\n<p>and follow the prompts:<\/p>\n<pre>Generating public\/private rsa key pair.\nEnter file in which to save the key: \/home\/devbox\/ssh\/id_rsa\nEnter passphrase (empty for no passphrase): \nEnter same passphrase again:\nYour identification has been saved in \/home\/devbox\/ssh\/id_rsa.\nYour public key has been saved in \/home\/devbox\/ssh\/id_rsa.pub.\nThe key fingerprint is:\nSHA256:yzwrsYbtfJGe6b9o9UVXLvAs96HbSjyhWTGwiL3V1YI your_email@example.com\n The key's randomart image is:\n +---[RSA 4096]----+\n |           . . ..|\n |        o . E o o|\n |       . o o B o.|\n |          o . O.o|\n |        So   *.+.|\n |      .oo.. =.o .|\n |     o +== + =o  |\n |    ..= *+  o... |\n |     oo=+.o. ..  |\n +----[SHA256]-----+\n<\/pre><\/li>\n<li><p>Add your public key to the Termux ssh user host file. This is a little unorthodox, but remember the service is listening and on a local NAT interface only, bound to a reserved private IP address; ssh in this context is purely a local tty console. If an attacker compromised your notebook enough to recover your localhost key, you have much bigger problems.<\/p>\n<pre>cat \/home\/devbox\/ssh\/id_rsa.pub &gt;&gt; \/home\/.ssh\/authorized_keys\n<\/pre><\/li>\n<li><p>Find your userid and your Termux private IP number -- <strong>you will need this for any localhost web &amp; ssh servers<\/strong><\/p>\n<pre>$ whoami    \nu0_a49  # &lt;== your value may differ\n$ ifconfig arc0 | awk '\/inet \/{print $2}'\n100.115.92.2  # &lt;== your value may differ \n<\/pre><\/li>\n<li><p>Start the ssh server inside a chroot (it will be running on 8022)<\/p>\n<pre>$ termux-chroot\n$ sshd\n<\/pre><\/li>\n<li><p>Connect with the FireSSH or NaSSH app to the reserved IP address displayed earlier as the current user over port 8022, e.g.: <code>u0_a49@ 100.115.92.2:8022<\/code> using the private key just created at <code>\/Downloads\/devbox\/ssh\/id_rsa<\/code>. The Termux sshd server configs can be found at: <code>\/data\/data\/com.termux\/files\/usr\/etc\/ssh<\/code><\/p><\/li>\n<\/ol><h3 id=\"optionalbuthighlyrecommended\">Optional, but highly recommended<\/h3><ul><li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/ublock-origin\/cjpalhdlnbpafiamejdnhcphjbkeiagm\">uBlock Origin<\/a> high-performance ad blocker extension<\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/cog-system-info-viewer\/difcjdggkffcfgcfconafogflmmaadco?utm_source=chrome-app-launcher-info-dialog\">Cog System Viewer<\/a> app <a href=\"https:\/\/github.com\/beaufortfrancois\/cog-chrome-app\">[GitHub]<\/a><\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/caret\/fljalecfjciodhpcledpamjachpmelml\">Caret<\/a> desktop editor, a solid Sublime clone app  <a href=\"https:\/\/github.com\/thomaswilburn\/Caret\">[GitHub]<\/a><\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/zed-code-editor\/pfmjnmeipppmcebplngmhfkleiinphhp\">Zed<\/a>, a highly configurable desktop code &amp; markdown editor <a href=\"https:\/\/github.com\/zedapp\/zed\">[GitHub]<\/a>. The UI is purposely non-intrusive, but the minimalism can be a bit confusing at first. Check out the docs <a href=\"https:\/\/github.com\/zedapp\/manual\/blob\/master\/commands.md\">here<\/a>.<\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/signal-private-messenger\/bikioccmkafdpakkkcpdbppfkghcmihk\">Signal Desktop<\/a> app, the standard in secure messaging. Fully supports documents &amp; group chats. More info at <a href=\"https:\/\/whispersystems.org\/blog\/signal-desktop\/\">Whisper Systems<\/a><\/li>\n<li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.wire\">Wire Desktop<\/a> Android app, supports secure text, video &amp; voice calls <a href=\"https:\/\/github.com\/wireapp\/wire-android\">[GitHub]<\/a>. I ended up installing Wire because the Signal Desktop app doesn't support voice calls. Note: Consider signing out of the app when not in use, as the background memory use can substantial.<\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/dom-distiller-reading-mod\/oiembdaoobijmdmeobkalaehgifealpl\/related?hl=en\">DOM Distiller<\/a> this is one of my favorite apps. It calls Chrome's reader mode via the Distiller component. Here's <a href=\"https:\/\/twitter.com\/kennwhite\/status\/889876690243026946\">an example<\/a>. GitHub <a href=\"https:\/\/github.com\/metarmask\/dom-distiller-reading-mode\">repo<\/a><\/li>\n<\/ul><h3 id=\"businesssoftware\">Business software<\/h3><ul><li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/hp-print-for-chrome\/cjanmonomjogheabiocdamfpknlpdehm\">HP Print app<\/a> If you have an HP, you'll thank me.<\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/office-editing-for-docs-s\/gbkeegbaiigmenfmjfclcdgdpimamgkj\">Native Microsoft documents<\/a> app by Google to edit &amp; view Word, Excel, Powerpoint files (docx, xlsx, pptx)<\/li>\n<li>Google Docs: native app Option A; more stable, offline by default, but could use a little polish.\n<ul><li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.google.android.apps.docs.editors.docs\">Docs<\/a><\/li>\n<li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.google.android.apps.docs.editors.sheets\">Sheets<\/a><\/li><\/ul><\/li>\n<li>Google Docs: web app Option B; requires connectivity to start, quasi-offline capable. <em>Set config to Open in Window &amp; Pin to Shelf<\/em>.\n<ul><li><a href=\"https:\/\/docs.google.com\/document\/\">Docs<\/a><\/li>\n<li><a href=\"https:\/\/docs.google.com\/spreadsheets\/\">Sheets<\/a><\/li>\n<li><a href=\"https:\/\/docs.google.com\/presentation\/\">Slides<\/a><\/li><\/ul><\/li>\n<\/ul><h3 id=\"dessert\">Dessert<\/h3><p><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/netflix\/deceagebecbceejblnlcjooeohmmeldh\">Netflix<\/a>. Once installed, from Launcher (the <strong>O<\/strong> button on the bottom left menu bar) &gt; Netflix &gt; right click &gt; Open in Window &gt; Pin to Shelf. (This, by the way, is very useful for most Chrome web apps)<\/p><p><img title=\"netflix\" alt=\"netflix\" src=\"https:\/\/user-images.githubusercontent.com\/326990\/28655960-0f5e3364-726d-11e7-82ab-0250de4d4f6f.png\" \/><\/p><p><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/web-quake\/ibkbfanmkmadbbgggonficloplenbefh\">Quake<\/a> (because why not?)<\/p><p><img title=\"quake\" alt=\"quake\" src=\"https:\/\/user-images.githubusercontent.com\/326990\/28488424-b9c50eda-6e77-11e7-9705-c4a9dc5e736e.png\" \/><\/p><h3 id=\"optionalniche\">Optional\/Niche<\/h3><ul><li><a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.brave.browser\">Brave<\/a> Android native app <a href=\"https:\/\/github.com\/brave\/browser-android-tabs\">GitHub<\/a>. Possibly useful to segregate web traffic, but presents as a mobile browser <\/li>\n<li><a href=\"https:\/\/chrome.google.com\/webstore\/detail\/mosh\/ooiklbnjmhbcgemelgfhaeaocllobloj\">Mosh<\/a> persistent mobile SSH shell (useful for working with intermittent connectivity, e.g., remote VMs) <a href=\"https:\/\/github.com\/rpwoodbu\/mosh-chrome\">GitHub<\/a><\/li>\n<\/ul><hr \/><h4 id=\"epilogworkingwithtermux\">Epilog: Working with Termux<\/h4><p><i><br \/>Kill\/restarting shell<\/i><\/p><p>When things get completely borked (which in two weeks of heavy use only happened a couple of times for me), you can kill &amp; restart Termux. There might be a more elegant way to do this, but here's the blunt hammer method:   <\/p><ul><li><p>Slow press on the Termux app &gt; More &gt; Kill Process &gt; OK &gt; close window<\/p><\/li>\n<li><p>Open Launcher (the <strong>O<\/strong> button on the bottom far left menu bar) &gt; right click &gt; App Info &gt; Force Stop &gt; OK<\/p><\/li>\n<li>Relaunch Termux<\/li>\n<\/ul><p><i>termux-chroot (PRoot)<\/i><\/p><p>You will probably want to have chroot &amp; sshd autostart. It's important to call chroot <em>before<\/em> starting sshd so that new sessions will see the more familiar filesystem layout. Termux-chroot itself is a script which spawns a new shell, so if you call it from a script, the remainder of code won't execute until PRoot exits. There is a Termux:Boot add-on <a href=\"https:\/\/github.com\/termux\/termux-app\/issues\/14#issuecomment-308834693\">in the works<\/a>, but in the meantime, here's my simple work around. The logic flow is explained <a href=\"https:\/\/github.com\/termux\/termux-app\/issues\/14#issuecomment-318139827\">here<\/a>.<\/p><pre># Create\/append to: ~\/.bash_profile\nif ! pgrep -f \"termux-chroot\" &gt;\/dev\/null ; then echo \"[Starting chroot...]\" &amp;&amp; termux-chroot; else echo \"[chroot is running]\"; fi\nif ! pgrep \"sshd\" &gt;\/dev\/null ; then echo \"[Starting sshd...]\" &amp;&amp; sshd &amp;&amp; echo \"[OK]\"; else echo \"[ssh is running]\"; fi\n<\/pre><p><i>termux-wake-lock<\/i> <br \/>This is useful option to avoid sleep &amp; network weirdness<\/p><h4 id=\"epilogchromebooktipsquirks\">Epilog: Chromebook tips &amp; quirks<\/h4><ul><li><p>The built-in Chrome developer shell <em>crosh<\/em> is somewhat useful, but is fairly restricted. Worth checking out the <em>help-advanced<\/em> options.<\/p><\/li>\n<li><p>The physical file system location for Downloads outside of Termux is: <code>\/home\/chronos\/user\/Downloads\/<\/code>, or in a browser as: <code>file:\/\/\/home\/chronos\/user\/Downloads\/<\/code><\/p><\/li>\n<li><p>To switch between open apps: Alt + Tab<\/p><\/li>\n<li><p>Ctrl+Alt +\/- resizes the native Termux app fonts<\/p><\/li>\n<li><p>Shift+Ctrl +\/- resizes the Chromebook display (1366x768 is optimal\/native mode on the Samsung)<\/p><\/li>\n<li><p>Shift+Esc is a cool shortcut to the OS task viewer (able to see per-tab browser processes)<\/p><\/li>\n<li><p><code>chrome:\/\/about\/<\/code> is a thing, and if you really want to drill into the inner workings of the system: <code>chrome:\/\/system\/<\/code> (this will take a few seconds to populate -- it's a huge page)<\/p><\/li>\n<li><p>If an Android native app goes full screen (i.e., loses the resize buttons in the top right corner, slow press or right click to get a context menu which includes full-screen toggle off<\/p><\/li>\n<li><p>To share local files easier with other apps, create a transfer folder in Termux to the \/Downloads directory, eg:<\/p>\n<pre>mkdir -p \/storage\/emulated\/0\/Downloads\/code\nln -s \/storage\/emulated\/0\/Downloads\/code \/home\/code\ncd ~\/code\n<\/pre>\n<p>or simply save to an external MicroSD card (the 64GB models are around $25).<\/p>\n<p><strong>Important caveat:<\/strong> You can create and edit files &amp; code under a symlink to Downloads as shown above (e.g.: \/home\/code\/hello.go) or the SD Card (via Caret or Zed accessing the Chrome OS-level fs) but you can't execute from there because of security restrictions. Instead, build and edit where you wish, but save &amp; run executables from your Termux directories (for example: \/home\/localbuild, etc).<br \/><\/p><\/li>\n<li><p>Carefully consider sync settings:\nPicture (bottom right) &gt; Settings &gt; People:Sync &gt; (probably want custom)<\/p><\/li>\n<li><p>Screenshots: Normally you would use the keyboard (ctrl+Desktop icon), but if using an external keyboard, from Chrome: click the menu icon \u22ee &gt; More Tools &gt; Take Screenshot<\/p><\/li>\n<li><p>The bottom bar on the desktop is referred to several ways, sometimes menu, sometimes as task bar, but officially it's called \"the shelf\".<\/p><\/li>\n<li><p>You will make heavy use of the \"Launcher\" - the \u25ef symbol on the far left of the bottom menu bar.<\/p><\/li>\n<li><p>Pressing the magnifying glass icon \ud83d\udd0d above the shift key opens the Launcher, defaulting to the most recently used apps<\/p><\/li>\n<li><p>You can't customize the first screen in the Launcher -- it's always the most recently used apps, plus a few others installed, and then a button to \"All Apps\". You will be clicking that button a lot.<\/p><\/li>\n<li><p>You can customize the shelf several ways. Either right click or two-finger tap the touchpad on each of the apps to see more.\nFix Chrome Apps, Android apps, or even favorite web pages to the shelf by \"pinning\".<\/p><\/li>\n<li><p>For nearly all cases, you will want apps &amp; websites on the Shelf to open \"in a Window\"; besides feeling more like a typical app, the UI is much more svelte.<\/p><\/li>\n<li><p>To bring up a cool map of all keyboard shortcuts, press: Ctrl + Alt + ?  (once it's up, hold down the Shift, Alt, or Ctrl keys -- or combinations of, say, Shift and Ctrl -- to see the different shortcut mappings. You can actually press to invoke the desired behavior). Hit Esc to dismiss the map once you're done exploring.<\/p><\/li>\n<li><p>To quickly lock your Chromebook: \ud83d\udd0d + L<\/p><\/li>\n<li><p>To toggle caps lock (the built-in keyboard doesn't have one): Alt + \ud83d\udd0d<\/p><\/li>\n<li><p>If you really prefer to have a caps lock, you can swap the \ud83d\udd0d key to make it a proper caps lock: Settings &gt; devices &gt; keyboard &gt; Search &gt; (select Caps Lock in drop down)<\/p><\/li>\n<li><p>More info on customizing the Desktop here: <a href=\"https:\/\/support.google.com\/chromebook\/answer\/3113576?hl=en\">https:\/\/support.google.com\/chromebook\/answer\/3113576?hl=en<\/a><\/p><\/li>\n<\/ul><p>That glowing key is your new best friend:<\/p><p><img title=\"security key\" alt=\"security key\" src=\"https:\/\/camo.githubusercontent.com\/984fa560a98e47e0797ec57cb366dea7bdd225cb\/68747470733a2f2f7062732e7477696d672e636f6d2f6d656469612f44445f53613476585941414944502d2e6a7067\" \/><\/p><p><img title=\"yubikey\" alt=\"yubikey\" src=\"https:\/\/user-images.githubusercontent.com\/326990\/28656241-b9ca304a-726e-11e7-95aa-737db55ae666.jpg\" \/><\/p><p><em>Special thanks to Nick, Jonathan, Joe, Wendy, Filippo, Stephan, Tim, and Rob for reviewing an earlier draft of this post.<\/em><\/p><hr \/><p><a id=\"amazon-links\">\u00b9<\/a><em>Disclosure: Be wary of 3rd-party sellers offering \"open box\" and \"refurbished\" products. These often have poor (or no) warranties, and typically have higher return rates. Links to Amazon were carefully researched to ensure the items are as listed in the descriptions. I receive a small referral credit from purchases made through this post.<\/em><\/p><\/section>","created_at":"2018-11-07T16:51:32+0000","updated_at":"2018-11-07T16:51:32+0000","published_at":"2017-07-26T23:22:15+0000","published_by":["Kenneth White"],"annotations":[],"mimetype":"text\/html","reading_time":18,"domain_name":"blog.lessonslearned.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":544,"title":"How To Resize ext3 Partitions Without Losing Data","url":"https:\/\/www.howtoforge.com\/linux_resizing_ext3_partitions","content":"<div id=\"htfContentPage\"><p><a href=\"https:\/\/srv.buysellads.com\/ads\/click\/x\/GTND42QWC6SD653YCWY4YKQMCEYIP2QMCAAITZ3JCWBD42QJF6YIC23KC6BIVKQNCTSDVK3EHJNCLSIZ\" target=\"_blank\" rel=\"nofollow\">Want more customers? Our marketing tools can help you find them.<\/a><\/p><div class=\"contributeEdit\"><div id=\"tocContainer\"><h3>On this page<\/h3><ol class=\"toc\"><li><a href=\"#how-to-resize-ext-partitions-without-losing-data-\">How To Resize ext3 Partitions Without Losing Data<\/a>\n<ol><li><a href=\"#-preliminary-note\">1 Preliminary Note<\/a><\/li>\n<li><a href=\"#-shrinking-an-ext-partition\">2 Shrinking An ext3 Partition<\/a>\n<\/li><\/ol><\/li>\n<\/ol><\/div><\/div><h2 id=\"how-to-resize-ext-partitions-without-losing-data-\">How To Resize ext3 Partitions Without Losing Data <\/h2><p>Version 1.0 <br \/>Author: Falko Timme <br \/><\/p><p>This article is about resizing ext3 partitions without losing data. It shows how to shrink and enlarge existing ext3 partitions and how to merge two ext3 partitions. This can be quite useful if you do not use LVM and you realize that your existing partitioning does not meet your actual needs anymore. <\/p><p>There are many ways of achieving this goal but this is the way I take. I do not issue any guarantee that this will work for you!<\/p><h3 id=\"-preliminary-note\">1 Preliminary Note<\/h3><p>(I run all the commands in this tutorial as the root user, so make sure you're logged in as root. If you are on a Ubuntu system, you can become root like this:<\/p><p class=\"command\">sudo su<\/p><p>) <\/p><p>I have tested this on a Ubuntu Edgy Eft desktop system that has all files in one large partition (around 10 GB, device \/dev\/sda1). The partitioning looks like this:<\/p><p class=\"command\">df -h<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Size\u00a0\u00a0Used\u00a0Avail\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a09.5G\u00a0\u00a04.1G\u00a0\u00a04.9G\u00a0\u00a046%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0132K\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a052K\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a018M\u00a0\u00a0\u00a077M\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p>The partition that is to be resized must be unmounted when we do the resizing; obviously this is not possible if this is the partition that holds all important system files like in this example. Therefore we download a Live Linux-CD such as <a href=\"http:\/\/www.knopper.net\/knoppix-mirrors\/index-en.html\" target=\"_blank\">Knoppix<\/a> from which we boot later on (if you have physical access to the system). If it is a remote system that you don't have physical access to, you need a rescue system on that system (a lot of hosting companies offer dedicated servers with rescue systems nowadays) that you can boot into (instead of Knoppix), and this rescue system must have the following tools: fdisk, umount, fsck, tune2fs, e2fsck, resize2fs.<\/p><p>If the partition that you want to resize doesn't hold any system files (such as \/home partitions, partitions for backups, etc.), you don't need a Knoppix Live-CD or a rescue system, because all steps can be run from the original system.<\/p><p>If you want to resize partitions on production systems, please back up your data before, because it is possible you lose all your data if you don't calculate the size of your new partition correctly (especially when shrinking a partition)! You have been warned! Tutorials about backups can be found here: <a href=\"http:\/\/www.howtoforge.com\/taxonomy_menu\/1\/34\">http:\/\/www.howtoforge.com\/taxonomy_menu\/1\/34<\/a><\/p><p>I'm going to resize \/dev\/sda1 in this tutorial. If your partition is named differently, please replace \/dev\/sda1 with your own device (e.g. \/dev\/hda5, \/dev\/sdb3, etc.). <\/p><h3 id=\"-shrinking-an-ext-partition\">2 Shrinking An ext3 Partition<\/h3><p>This chapter is about shrinking an ext3 partition. I want to shrink \/dev\/sda1 in this example. First we gather some details on our original system:<\/p><p class=\"command\">df<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01K-blocks\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Used\u00a0Available\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a09859036\u00a0\u00a0\u00a04234908\u00a0\u00a0\u00a05123304\u00a0\u00a046%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0132\u00a0\u00a0\u00a0\u00a0\u00a095348\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a010240\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a052\u00a0\u00a0\u00a0\u00a0\u00a010188\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a017580\u00a0\u00a0\u00a0\u00a0\u00a077900\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p class=\"command\">df -B 4k<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a04K-blocks\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Used\u00a0Available\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02464759\u00a0\u00a0\u00a01058727\u00a0\u00a0\u00a01280826\u00a0\u00a046%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a033\u00a0\u00a0\u00a0\u00a0\u00a023837\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02560\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a013\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02547\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a04395\u00a0\u00a0\u00a0\u00a0\u00a019475\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p class=\"command\">df -h <\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Size\u00a0\u00a0Used\u00a0Avail\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a09.5G\u00a0\u00a04.1G\u00a0\u00a04.9G\u00a0\u00a046%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0132K\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a052K\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a018M\u00a0\u00a0\u00a077M\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p class=\"command\">fdisk -l<\/p><p class=\"system\"><br \/>Disk\u00a0\/dev\/sda:\u00a010.7\u00a0GB,\u00a010737418240\u00a0bytes<br \/>255\u00a0heads,\u00a063\u00a0sectors\/track,\u00a01305\u00a0cylinders<br \/>Units\u00a0=\u00a0cylinders\u00a0of\u00a016065\u00a0*\u00a0512\u00a0=\u00a08225280\u00a0bytes<\/p><p>\u00a0\u00a0\u00a0Device\u00a0Boot\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Start\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0End\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Blocks\u00a0\u00a0\u00a0Id\u00a0\u00a0System<br \/>\/dev\/sda1\u00a0\u00a0\u00a0*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01247\u00a0\u00a0\u00a0\u00a010016496\u00a0\u00a0\u00a083\u00a0\u00a0Linux<br \/>\/dev\/sda2\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01248\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01305\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0465885\u00a0\u00a0\u00a0\u00a05\u00a0\u00a0Extended<br \/>\/dev\/sda5\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01248\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01305\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0465853+\u00a0\u00a082\u00a0\u00a0Linux\u00a0swap\u00a0\/\u00a0Solaris<\/p><p class=\"command\">fdisk -s \/dev\/sda1<\/p><p class=\"system\">10016496<\/p><p>Then we shut down the system and boot into our Knoppix Live-CD (or your rescue system) (if the partition you want to resize doesn't hold any system files, you can do everything from the original system; the steps are the same, just omit booting into Knoppix\/your rescue system). <\/p><p class=\"command\">shutdown -r now<\/p><p>After Knoppix has booted, open a terminal and become root by running <\/p><p class=\"command\">su<\/p><p>\/dev\/sda1 should be unmounted by default, but you can run<\/p><p class=\"command\">umount \/dev\/sda1<\/p><p>to go sure. <\/p><p>Then run<\/p><p class=\"command\">fsck -n \/dev\/sda1<\/p><p>The output looks like this:<\/p><p class=\"system\">fsck 1.38 (30-Jun-2005)<br \/>e2fsck 1.38 (30-Jun-2005)<br \/>\/dev\/sda1: clean, 159037\/1254176 files, 1095299\/2504124 blocks <\/p><p>Next we remove the journal from \/dev\/sda1, thus turning it into an ext2 partition:<\/p><p class=\"command\">tune2fs -O ^has_journal \/dev\/sda1<\/p><p>The output looks like this:<\/p><p class=\"system\">tune2fs 1.38 (30-Jun-2005)<\/p><p>Then run<\/p><p class=\"command\">e2fsck -f \/dev\/sda1<\/p><p class=\"system\">e2fsck 1.38 (30-Jun-2005)<br \/>Pass 1: Checking inodes, blocks, and sizes<br \/>Pass 2: Checking directory structure<br \/>Pass 3: Checking directory connectivity<br \/>Pass 4: Checking reference counts<br \/>Pass 5: Checking group summary information<br \/>\/dev\/sda1: 164178\/1254176 files (0.6% non-contiguous), 1051617\/2504124 blocks <\/p><p>Now we resize our file system with resize2fs. resize2fs can resize ext2 file systems, but not ext3 file systems, that's why we had to turn \/dev\/sda1 to ext2. Currently, 4.1GB are used on \/dev\/sda1 (see the df -h output above), So it's safe to shrink it from 10GB to about 6GB (if you make it smaller than 4.1GB, you will lose data!). Therefore we run<\/p><p class=\"command\">resize2fs \/dev\/sda1 6000M<\/p><p>The output is as follows:<\/p><p class=\"system\">resize2fs 1.38 (30-Jun-2005)<br \/>Resizing the filesystem on \/dev\/sda1 to 1536000 (4k) blocks.<br \/>The filesystem on \/dev\/sda1 is now 1536000 blocks long. <\/p><p>Please take note of the amount of blocks (1536000) and their size (4k). We need that soon. <\/p><p>Now we delete our \/dev\/sda1 partition (don't be afraid, no data will be lost) and create a new, smaller one (but still big enough to hold our resized file system!). We can do this with fdisk:<\/p><p class=\"command\">fdisk \/dev\/sda<\/p><p>(Yes, it's \/dev\/sda, not \/dev\/sda1.) <\/p><p class=\"system\">The number of cylinders for this disk is set to 1305.<br \/>There is nothing wrong with that, but this is larger than 1024,<br \/>and could in certain setups cause problems with:<br \/>1) software that runs at boot time (e.g., old versions of LILO)<br \/>2) booting and partitioning software from other OSs<br \/>\u00a0\u00a0\u00a0(e.g., DOS FDISK, OS\/2 FDISK)<\/p><p>Type m to get a list of all commands:<\/p><p class=\"system\">Command\u00a0(m\u00a0for\u00a0help):\u00a0m<br \/>Command\u00a0action<br \/>\u00a0\u00a0\u00a0a\u00a0\u00a0\u00a0toggle\u00a0a\u00a0bootable\u00a0flag<br \/>\u00a0\u00a0\u00a0b\u00a0\u00a0\u00a0edit\u00a0bsd\u00a0disklabel<br \/>\u00a0\u00a0\u00a0c\u00a0\u00a0\u00a0toggle\u00a0the\u00a0dos\u00a0compatibility\u00a0flag<br \/>\u00a0\u00a0\u00a0d\u00a0\u00a0\u00a0delete\u00a0a\u00a0partition<br \/>\u00a0\u00a0\u00a0l\u00a0\u00a0\u00a0list\u00a0known\u00a0partition\u00a0types<br \/>\u00a0\u00a0\u00a0m\u00a0\u00a0\u00a0print\u00a0this\u00a0menu<br \/>\u00a0\u00a0\u00a0n\u00a0\u00a0\u00a0add\u00a0a\u00a0new\u00a0partition<br \/>\u00a0\u00a0\u00a0o\u00a0\u00a0\u00a0create\u00a0a\u00a0new\u00a0empty\u00a0DOS\u00a0partition\u00a0table<br \/>\u00a0\u00a0\u00a0p\u00a0\u00a0\u00a0print\u00a0the\u00a0partition\u00a0table<br \/>\u00a0\u00a0\u00a0q\u00a0\u00a0\u00a0quit\u00a0without\u00a0saving\u00a0changes<br \/>\u00a0\u00a0\u00a0s\u00a0\u00a0\u00a0create\u00a0a\u00a0new\u00a0empty\u00a0Sun\u00a0disklabel<br \/>\u00a0\u00a0\u00a0t\u00a0\u00a0\u00a0change\u00a0a\u00a0partition's\u00a0system\u00a0id<br \/>\u00a0\u00a0\u00a0u\u00a0\u00a0\u00a0change\u00a0display\/entry\u00a0units<br \/>\u00a0\u00a0\u00a0v\u00a0\u00a0\u00a0verify\u00a0the\u00a0partition\u00a0table<br \/>\u00a0\u00a0\u00a0w\u00a0\u00a0\u00a0write\u00a0table\u00a0to\u00a0disk\u00a0and\u00a0exit<br \/>\u00a0\u00a0\u00a0x\u00a0\u00a0\u00a0extra\u00a0functionality\u00a0(experts\u00a0only)<\/p><p>Now we delete partition no. 1 (\/dev\/sda1):<\/p><p class=\"system\">Command (m for help): d<br \/>Partition number (1-5): 1<\/p><p>Next we create a new \/dev\/sda1 partition. It was a primary partition before, so we choose p again, and again it is our partition no. 1:<\/p><p class=\"system\">Command\u00a0(m\u00a0for\u00a0help):\u00a0n<br \/>Command\u00a0action<br \/>\u00a0\u00a0\u00a0l\u00a0\u00a0\u00a0logical\u00a0(5\u00a0or\u00a0over)<br \/>\u00a0\u00a0\u00a0p\u00a0\u00a0\u00a0primary\u00a0partition\u00a0(1-4)<br \/>p<br \/>Partition\u00a0number\u00a0(1-4):\u00a01<\/p><p>Now comes the crucial part - we are asked about the size of the new partition. The first cylinder is no problem, it is the one from the fdisk -l output at the beginning of this chapter (1). <\/p><p>First cylinder (1-1305, default 1): 1<\/p><p>But we don't have a value for the last cylinder of our new partition. Fortunately, we can specify the size in kilobytes (K), so we calculate the size like this: <\/p><p>We multiply the amount of blocks from the resize2fs output (1536000) by the size of a block (4k), and to go sure the partition is big enough, we add 3 to 5% to it (3% was enough for me, but if you want to go sure take 5%):<\/p><p class=\"highlight\">1536000 * 4k * 1.03 = 6328320k<\/p><p>So we prepend that value with a + sign and replace the small k with a capital one (K) and enter it: <\/p><p class=\"system\">Last cylinder or +size or +sizeM or +sizeK (1-1247, default 1247): +6328320K <\/p><p>Our original \/dev\/sda1 had the bootable flag (see the fdisk -l output from the beginning of this chapter), so we must add it to our new \/dev\/sda1 again:<\/p><p class=\"system\">Command (m for help): a<br \/>Partition number (1-5): 1 <\/p><p>Now let's write our new partition table and exit fdisk:<\/p><p class=\"system\">Command (m for help): w<br \/>The partition table has been altered!<\/p><p class=\"system\">Calling ioctl() to re-read partition table.<\/p><p class=\"system\">WARNING: Re-reading the partition table failed with error 16: Device or resource busy.<br \/>The kernel still uses the old table.<br \/>The new table will be used at the next reboot.<br \/>Syncing disks.<\/p><p>Now we reboot the system, and again we boot into our Knoppix system (rescue system; original system if resized partition doesn't hold system files):<\/p><p class=\"command\">shutdown -r now<\/p><p>Become root again (on Knoppix run<\/p><p class=\"command\">su<\/p><p>) <\/p><p>and then run this:<\/p><p class=\"command\">fsck -n \/dev\/sda1<\/p><p>The output should look like this:<\/p><p class=\"system\">fsck 1.38 (30-Jun-2005)<br \/>e2fsck 1.38 (30-Jun-2005)<br \/>\/dev\/sda1: clean, 159036\/765536 files, 1047239\/1536000 blocks<\/p><p>Then we create the journal on our new \/dev\/sda1, thus turning it into an ext3 partition again:<\/p><p class=\"command\">tune2fs -j \/dev\/sda1<\/p><p class=\"system\">tune2fs 1.38 (30-Jun-2005)<br \/>Creating journal inode: done<br \/>This filesystem will be automatically checked every 30 mounts or<br \/>0 days, whichever comes first. Use tune2fs -c or -i to override. <\/p><p>Now we are done. Shut down the system and boot into the original system:<\/p><p class=\"command\">shutdown -r now<\/p><p>If everything goes well, the original system will boot up, and no data has been lost. Now we can gather some details about our new partitioning and compare them with the information we collected at the beginning of this chapter:<\/p><p class=\"command\">df<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01K-blocks\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Used\u00a0Available\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a06047868\u00a0\u00a0\u00a04224140\u00a0\u00a0\u00a01639408\u00a0\u00a073%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0132\u00a0\u00a0\u00a0\u00a0\u00a095348\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a010240\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a052\u00a0\u00a0\u00a0\u00a0\u00a010188\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a095480\u00a0\u00a0\u00a0\u00a0\u00a017580\u00a0\u00a0\u00a0\u00a0\u00a077900\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p class=\"command\">df -B 4k<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a04K-blocks\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Used\u00a0Available\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01511967\u00a0\u00a0\u00a01056035\u00a0\u00a0\u00a0\u00a0409852\u00a0\u00a073%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a033\u00a0\u00a0\u00a0\u00a0\u00a023837\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02560\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a013\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02547\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a023870\u00a0\u00a0\u00a0\u00a0\u00a0\u00a04395\u00a0\u00a0\u00a0\u00a0\u00a019475\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p class=\"command\">df -h<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Size\u00a0\u00a0Used\u00a0Avail\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a05.8G\u00a0\u00a04.1G\u00a0\u00a01.6G\u00a0\u00a073%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0132K\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a052K\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a018M\u00a0\u00a0\u00a077M\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<\/p><p class=\"command\">fdisk -l<\/p><p class=\"system\"><br \/>Disk\u00a0\/dev\/sda:\u00a010.7\u00a0GB,\u00a010737418240\u00a0bytes<br \/>255\u00a0heads,\u00a063\u00a0sectors\/track,\u00a01305\u00a0cylinders<br \/>Units\u00a0=\u00a0cylinders\u00a0of\u00a016065\u00a0*\u00a0512\u00a0=\u00a08225280\u00a0bytes<\/p><p>\u00a0\u00a0\u00a0Device\u00a0Boot\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Start\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0End\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Blocks\u00a0\u00a0\u00a0Id\u00a0\u00a0System<br \/>\/dev\/sda1\u00a0\u00a0\u00a0*\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0789\u00a0\u00a0\u00a0\u00a0\u00a06337611\u00a0\u00a0\u00a083\u00a0\u00a0Linux<br \/>\/dev\/sda2\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01248\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01305\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0465885\u00a0\u00a0\u00a0\u00a05\u00a0\u00a0Extended<br \/>\/dev\/sda5\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01248\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01305\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0465853+\u00a0\u00a082\u00a0\u00a0Linux\u00a0swap\u00a0\/\u00a0Solaris<\/p><p class=\"command\">fdisk -s \/dev\/sda1<\/p><p class=\"system\">6337611<\/p><div><p><img src=\"https:\/\/www.howtoforge.com\/images\/pdficon_small.png\" alt=\"image\" \/><a class=\"pdf-btn\" href=\"denied:javascript:void(0);\">view as pdf<\/a> | <img src=\"https:\/\/www.howtoforge.com\/images\/print.gif\" alt=\"image\" \/><\/p><\/div><div><p><b>Share this page:<\/b><\/p><p>\n<a href=\"https:\/\/www.facebook.com\/sharer.php?u=https%3A%2F%2Fwww.howtoforge.com%2Flinux_resizing_ext3_partitions\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-recommend.png\" height=\"20\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/twitter.com\/intent\/tweet?url=https%3A%2F%2Fwww.howtoforge.com%2Flinux_resizing_ext3_partitions&amp;text=How+To+Resize+ext3+Partitions+Without+Losing+Data&amp;via=howtoforgecom&amp;related=howtoforgecom\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-tweet.png\" height=\"20\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/twitter.com\/howtoforgecom\/\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-follow.png\" height=\"20\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/plus.google.com\/share?url=https%3A%2F%2Fwww.howtoforge.com%2Flinux_resizing_ext3_partitions\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-gplus.png\" height=\"20\" alt=\"image\" \/><\/a>\n<\/p><\/div><\/div><div id=\"htfContentPageComment\"><h2>40 Comment(s)<\/h2><h3>Comments<\/h3><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> dilino<\/p><p><a href=\"#\" data-comment=\"1433\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1434\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1444\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel2\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1451\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel3\"><p><b>By:<\/b> linuxuser<\/p><p><a href=\"#\" data-comment=\"1460\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel4\"><p><b>By:<\/b> Wayan<\/p><p><a href=\"#\" data-comment=\"1465\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel2\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1452\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b><\/p><p><a href=\"#\" data-comment=\"1435\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b><\/p><p><a href=\"#\" data-comment=\"1436\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Austin<\/p><p><a href=\"#\" data-comment=\"1453\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b><\/p><p><a href=\"#\" data-comment=\"1437\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Fish Diesel<\/p><p><a href=\"#\" data-comment=\"1442\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b><\/p><p><a href=\"#\" data-comment=\"1438\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1441\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b><\/p><p><a href=\"#\" data-comment=\"1439\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Thor<\/p><p><a href=\"#\" data-comment=\"1440\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Fedesh03<\/p><p><a href=\"#\" data-comment=\"1443\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1445\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1446\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1450\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Bill Pringlemeir.<\/p><p><a href=\"#\" data-comment=\"1447\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Cuervo79<\/p><p><a href=\"#\" data-comment=\"1448\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> David Greaves<\/p><p><a href=\"#\" data-comment=\"1449\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b><\/p><p><a href=\"#\" data-comment=\"1454\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Arnd<\/p><p><a href=\"#\" data-comment=\"1455\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1456\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> noob<\/p><p><a href=\"#\" data-comment=\"1457\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1466\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1458\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1459\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Andrew B<\/p><p><a href=\"#\" data-comment=\"1461\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div id=\"comment1461\" class=\"commentContainer commentLevel0\"><p>Didn't work for me.\u00a0 Any ideas?<\/p><p>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"5f2d30302b1f2c262736392b6f6e\">[email\u00a0protected]<\/a> ~]# umount \/opt\/iress_ftp<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"fc8e939388bc8f8584959a88cccd\">[email\u00a0protected]<\/a> ~]# df -k<br \/>Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 1K-blocks\u00a0\u00a0\u00a0\u00a0\u00a0 Used Available Use% Mounted on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 6190664\u00a0\u00a0 3231964\u00a0\u00a0 2644232\u00a0 56% \/<br \/>none\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 517268\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 0\u00a0\u00a0\u00a0 517268\u00a0\u00a0 0% \/dev\/shm<br \/>\/dev\/sda2\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 10309828\u00a0\u00a0\u00a0 794716\u00a0\u00a0 8991396\u00a0\u00a0 9% \/opt<br \/>\/dev\/sda3\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 3099292\u00a0\u00a0\u00a0 741312\u00a0\u00a0 2200544\u00a0 26% \/var<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"d9abb6b6ad99aaa0a1b0bfade9e8\">[email\u00a0protected]<\/a> ~]# fsck -n \/dev\/sdb1<br \/>fsck 1.35 (28-Feb-2004)<br \/>e2fsck 1.35 (28-Feb-2004)<br \/>\/iress_ftp: clean, 664\/30277632 files, 53626986\/60548889 blocks<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"74061b1b0034070d0c1d12004445\">[email\u00a0protected]<\/a> ~]# tune2fs -O ^has_journal \/dev\/sdb1<br \/>tune2fs 1.35 (28-Feb-2004)<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"1d6f7272695d6e6465747b692d2c\">[email\u00a0protected]<\/a> ~]# fdisk \/dev\/sdb<\/p><p>The number of cylinders for this disk is set to 45689.<br \/>There is nothing wrong with that, but this is larger than 1024,<br \/>and could in certain setups cause problems with:<br \/>1) software that runs at boot time (e.g., old versions of LILO)<br \/>2) booting and partitioning software from other OSs<br \/>\u00a0\u00a0 (e.g., DOS FDISK, OS\/2 FDISK)<\/p><p>Command (m for help): d<br \/>Selected partition 1<\/p><p>Command (m for help): 1<br \/>1: unknown command<br \/>Command action<br \/>\u00a0\u00a0 a\u00a0\u00a0 toggle a bootable flag<br \/>\u00a0\u00a0 b\u00a0\u00a0 edit bsd disklabel<br \/>\u00a0\u00a0 c\u00a0\u00a0 toggle the dos compatibility flag<br \/>\u00a0\u00a0 d\u00a0\u00a0 delete a partition<br \/>\u00a0\u00a0 l\u00a0\u00a0 list known partition types<br \/>\u00a0\u00a0 m\u00a0\u00a0 print this menu<br \/>\u00a0\u00a0 n\u00a0\u00a0 add a new partition<br \/>\u00a0\u00a0 o\u00a0\u00a0 create a new empty DOS partition table<br \/>\u00a0\u00a0 p\u00a0\u00a0 print the partition table<br \/>\u00a0\u00a0 q\u00a0\u00a0 quit without saving changes<br \/>\u00a0\u00a0 s\u00a0\u00a0 create a new empty Sun disklabel<br \/>\u00a0\u00a0 t\u00a0\u00a0 change a partition's system id<br \/>\u00a0\u00a0 u\u00a0\u00a0 change display\/entry units<br \/>\u00a0\u00a0 v\u00a0\u00a0 verify the partition table<br \/>\u00a0\u00a0 w\u00a0\u00a0 write table to disk and exit<br \/>\u00a0\u00a0 x\u00a0\u00a0 extra functionality (experts only)<\/p><p>Command (m for help): p<\/p><p>Disk \/dev\/sdb: 375.8 GB, 375809638400 bytes<br \/>255 heads, 63 sectors\/track, 45689 cylinders<br \/>Units = cylinders of 16065 * 512 = 8225280 bytes<\/p><p>\u00a0\u00a0 Device Boot\u00a0\u00a0\u00a0\u00a0\u00a0 Start\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 End\u00a0\u00a0\u00a0\u00a0\u00a0 Blocks\u00a0\u00a0 Id\u00a0 System<\/p><p>Command (m for help): n<br \/>Command action<br \/>\u00a0\u00a0 e\u00a0\u00a0 extended<br \/>\u00a0\u00a0 p\u00a0\u00a0 primary partition (1-4)<br \/>p<br \/>Partition number (1-4): 1<br \/>First cylinder (1-45689, default 1): <br \/>Using default value 1<br \/>Last cylinder or +size or +sizeM or +sizeK (1-45689, default 45689): <br \/>Using default value 45689<\/p><p>Command (m for help): w<br \/>The partition table has been altered!<\/p><p>Calling ioctl() to re-read partition table.<br \/>Syncing disks.<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"f0829f9f84b0838988999684c0c1\">[email\u00a0protected]<\/a> ~]# partprobe<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"f0829f9f84b0838988999684c0c1\">[email\u00a0protected]<\/a> ~]# resize2fs \/dev\/sdb1<br \/>resize2fs 1.35 (28-Feb-2004)<br \/>resize2fs: Bad magic number in super-block while trying to open \/dev\/sdb1<br \/>Couldn't find valid filesystem superblock.<br \/>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"cbb9a4a4bf8bb8b2b3a2adbffbfa\">[email\u00a0protected]<\/a> ~]# e2fsck -f \/dev\/sdb1<br \/>e2fsck 1.35 (28-Feb-2004)<br \/>Couldn't find ext2 superblock, trying backup blocks...<br \/>e2fsck: Bad magic number in super-block while trying to open \/dev\/sdb1<\/p><p>The superblock could not be read or does not describe a correct ext2<br \/>filesystem.\u00a0 If the device is valid and it really contains an ext2<br \/>filesystem (and not swap or ufs or something else), then the superblock<br \/>is corrupt, and you might try running e2fsck with an alternate superblock:<br \/>\u00a0\u00a0\u00a0 e2fsck -b 8193 &lt;device&gt;<\/p><p>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"2d5f4242596d5e5455444b591d1c\">[email\u00a0protected]<\/a> ~]# e2fsck -f \/dev\/sdb<br \/>e2fsck 1.35 (28-Feb-2004)<br \/>Couldn't find ext2 superblock, trying backup blocks...<br \/>e2fsck: Bad magic number in super-block while trying to open \/dev\/sdb<\/p><p>The superblock could not be read or does not describe a correct ext2<br \/>filesystem.\u00a0 If the device is valid and it really contains an ext2<br \/>filesystem (and not swap or ufs or something else), then the superblock<br \/>is corrupt, and you might try running e2fsck with an alternate superblock:<br \/>\u00a0\u00a0\u00a0 e2fsck -b 8193 &lt;device&gt;<\/p><p>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"31435e5e45714248495857450100\">[email\u00a0protected]<\/a> ~]# e2fsck -b 8193 \/dev\/sdb1<br \/>e2fsck 1.35 (28-Feb-2004)<br \/>e2fsck: Bad magic number in super-block while trying to open \/dev\/sdb1<\/p><p>The superblock could not be read or does not describe a correct ext2<br \/>filesystem.\u00a0 If the device is valid and it really contains an ext2<br \/>filesystem (and not swap or ufs or something else), then the superblock<br \/>is corrupt, and you might try running e2fsck with an alternate superblock:<br \/>\u00a0\u00a0\u00a0 e2fsck -b 8193 &lt;device&gt;<\/p><p>[<a href=\"https:\/\/www.howtoforge.com\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"beccd1d1cafecdc7c6d7d8ca8e8f\">[email\u00a0protected]<\/a> ~]# dumpe2fs \/dev\/sdb1<br \/>dumpe2fs 1.35 (28-Feb-2004)<br \/>dumpe2fs: Bad magic number in super-block while trying to open \/dev\/sdb1<br \/>Couldn't find valid filesystem superblock<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> gedeef<\/p><p><a href=\"#\" data-comment=\"1462\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> wndrr<\/p><p><a href=\"#\" data-comment=\"1463\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel1\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1464\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> PeLu<\/p><p><a href=\"#\" data-comment=\"1467\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Anonymous<\/p><p><a href=\"#\" data-comment=\"1468\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Peter, Switzerland<\/p><p><a href=\"#\" data-comment=\"1469\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> bryan<\/p><p><a href=\"#\" data-comment=\"17321\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> Ejaz<\/p><p><a href=\"#\" data-comment=\"20020\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><div class=\"commentContainerHead commentLevel0\"><p><b>By:<\/b> K<\/p><p><a href=\"#\" data-comment=\"25585\" class=\"reply-to\">Reply<\/a> \u00a0<\/p><\/div><\/div>","created_at":"2018-11-07T16:51:31+0000","updated_at":"2018-11-07T16:51:31+0000","published_by":["Falko Timme"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":13,"domain_name":"www.howtoforge.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":545,"title":"Roguelike Tutorial Table of Contents \u2022 Jared McFarland","url":"https:\/\/jaredonline.svbtle.com\/roguelike-tutorial-table-of-contents","content":"<p>I decided I want to learn the <a href=\"http:\/\/www.rust-lang.org\/\" rel=\"nofollow\">Rust Programming Language<\/a> and I\u2019ve been looking for an excuse to go through <a href=\"http:\/\/gameprogrammingpatterns.com\/\" rel=\"nofollow\">Game Programming Patterns<\/a>, so I decided to combine the two.<\/p><p>I\u2019m making a <a href=\"http:\/\/en.wikipedia.org\/wiki\/Roguelike\" rel=\"nofollow\">Roguelike<\/a> based on <a href=\"http:\/\/mislav.uniqpath.com\/poignant-guide\/dwemthy\/\" rel=\"nofollow\">Dwemthy\u2019s Array<\/a> in Rust and live-blogging everything I learn along the way.<\/p><figure class=\"postend kudo able clearfix\" id=\"kudo_sJX4x6yfN42zYAcW5ikdes\"><a href=\"#kudo\">\n      \n    <\/a>\n    <p>159<\/p>\n    <p>Kudos<\/p>\n  <\/figure><figure class=\"side kudo able clearfix\" id=\"kudo_side_sJX4x6yfN42zYAcW5ikdes\"><a href=\"#kudo\">\n      \n    <\/a>\n    <p>159<\/p>\n    <p>Kudos<\/p>\n  <\/figure>","created_at":"2018-11-07T16:51:31+0000","updated_at":"2018-11-07T16:51:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"jaredonline.svbtle.com","preview_picture":"https:\/\/jaredonline.svbtle.com\/%5B%5D","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":542,"title":"Key Management","url":"https:\/\/www.gnupg.org\/gph\/en\/manual\/c235.html","content":"<p>Key tampering is a major security weakness with public-key cryptography.\nAn eavesdropper may tamper with a user's keyrings or forge a\nuser's public key and post it for others to download and use.\nFor example, suppose Chloe wants to monitor the messages that Alice\nsends to Blake.\nShe could mount what is called a <i class=\"FIRSTTERM\">man in the \nmiddle<\/i> attack.\nIn this attack, Chloe creates a new public\/private keypair.\nShe replaces Alice's copy of Blake's public key with the new public key.\nShe then intercepts the messages that Alice sends to Blake.\nFor each intercept, she decrypts it using the new private key, reencrypts\nit using Blake's true public key, and forwards the reencrypted\nmessage to Blake.\nAll messages sent from Alice to Blake can now be read by Chloe.<\/p><p>Good key management is crucial in order to ensure not just the integrity\nof your keyrings but the integrity of other users' keyrings as well.\nThe core of key management in GnuPG is the notion of signing keys.\nKey signing has two main purposes: it permits you to detect tampering\non your keyring, and it allows you to certify that a key truly belongs\nto the person named by a user ID on the key.\nKey signatures are also used in a scheme known as the <i class=\"FIRSTTERM\">web of \ntrust<\/i> to extend certification to keys not directly signed by you \nbut signed by others you trust.\nResponsible users who practice good key management can defeat key\ntampering as a practical attack on secure communication with GnuPG.<\/p><div class=\"SECT1\"><p>A keypair has a public key and a private key.\nA public key consists of \nthe public portion of the master signing key,\nthe public portions of the subordinate signing and encryption subkeys, and \na set of user IDs used to associate the public key with a real person.\nEach piece has data about itself.\nFor a key, this data includes its ID, when it was created, when it\nwill expire, etc.\nFor a user ID, this data includes the name of the real person it identifies,\nan optional comment, and an email address.\nThe structure of the private key is similar, except that it contains only\nthe private portions of the keys, and there is no user ID information.<\/p><p>The command-line option \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html\">--edit-key<\/a>\nmay be used to view a keypair.\nFor example,\n<\/p><table border=\"0\" style=\"background-color: #E0E0E0; width: 100%;\"><tr><td><pre class=\"SCREEN\"><code class=\"PROMPT\">chloe%<\/code> <code class=\"USERINPUT\"><b>gpg --edit-key chloe@cyb.org<\/b><\/code>\nSecret key is available.\npub  1024D\/26B6AAE1  created: 1999-06-15 expires: never      trust: -\/u\nsub  2048g\/0CF8CB7A  created: 1999-06-15 expires: never     \nsub  1792G\/08224617  created: 1999-06-15 expires: 2002-06-14\nsub   960D\/B1F423E7  created: 1999-06-15 expires: 2002-06-14\n(1)  Chloe (Jester) &lt;chloe@cyb.org&gt;\n(2)  Chloe (Plebian) &lt;chloe@tel.net&gt;\n<code class=\"PROMPT\">Command&gt;<\/code><\/pre><\/td><\/tr><\/table>The public key is displayed along with an indication of whether\nor not the private key is available.\nInformation about each component of the public key is then listed.\nThe first column indicates the type of the key.\nThe keyword<code class=\"LITERAL\">pub<\/code>identifies the public master signing key,\nand the keyword<code class=\"LITERAL\">sub<\/code>identifies a public subordinate key.\nThe second column indicates the key's bit length, type, and ID.\nThe type is<code class=\"LITERAL\">D<\/code>for a DSA key,<code class=\"LITERAL\">g<\/code>for an \nencryption-only\nElGamal key, and<code class=\"LITERAL\">G<\/code>for an ElGamal key that may be used for\nboth encryption and signing.\nThe creation date and expiration date are given in columns three and four.\nThe user IDs are listed following the keys.<p>More information about the key can be obtained with interactive commands.\nThe command <a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#TOGGLE\"><b class=\"COMMAND\">toggle<\/b><\/a> \nswitches between the public and private\ncomponents of a keypair if indeed both components are available.\n<\/p><table border=\"0\" style=\"background-color: #E0E0E0; width: 100%;\"><tr><td><pre class=\"SCREEN\"><code class=\"PROMPT\">Command&gt;<\/code> <code class=\"USERINPUT\"><b>toggle<\/b><\/code>\n               \nsec  1024D\/26B6AAE1  created: 1999-06-15 expires: never     \nsbb  2048g\/0CF8CB7A  created: 1999-06-15 expires: never     \nsbb  1792G\/08224617  created: 1999-06-15 expires: 2002-06-14\nsbb   960D\/B1F423E7  created: 1999-06-15 expires: 2002-06-14\n(1)  Chloe (Jester) &lt;chloe@cyb.org&gt;\n(2)  Chloe (Plebian) &lt;chloe@tel.net&gt;<\/pre><\/td><\/tr><\/table>The information provided is similar to the listing for the public-key\ncomponent.\nThe keyword<code class=\"LITERAL\">sec<\/code>identifies the private master signing key,\nand the keyword<code class=\"LITERAL\">sbb<\/code>identifies the private subordinates keys.\nThe user IDs from the public key are also listed for convenience.<div class=\"SECT2\"><p>When you distribute your public key, you are distributing the public\ncomponents of your master and subordinate keys as well as the user IDs.\nDistributing this material alone, however, is a security risk since\nit is possible for an attacker to tamper with the key.\nThe public key can be modified by adding or substituting keys, or by\nadding or changing user IDs.\nBy tampering with a user ID, the attacker could change the user ID's email\naddress to have email redirected to himself.\nBy changing one of the encryption keys, the attacker would\nalso be able to decrypt the messages redirected to him.<\/p><p>Using digital signatures is a solution to this problem.\nWhen data is signed by a private key, the corresponding public key\nis bound to the signed data.\nIn other words, only the corresponding public key can be used to\nverify the signature and ensure that the data has not been modified.\nA public key can be protected from tampering by using its corresponding\nprivate master key to sign the public key components and user IDs, thus\nbinding the components to the public master key.\nSigning public key components with the corresponding private master\nsigning key is called <i class=\"FIRSTTERM\">self-signing<\/i>, and a public key that has\nself-signed user IDs bound to it is called a <i class=\"FIRSTTERM\">certificate<\/i>.<\/p><p>As an example, Chloe has two user IDs and three subkeys.\nThe signatures on the user IDs can be checked with the command\n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#CHECK\"><b class=\"COMMAND\">check<\/b><\/a> from the key edit menu.\n<\/p><table border=\"0\" style=\"background-color: #E0E0E0; width: 100%;\"><tr><td><pre class=\"SCREEN\"><code class=\"PROMPT\">chloe%<\/code> <code class=\"USERINPUT\"><b>gpg --edit-key chloe<\/b><\/code>\nSecret key is available.\npub  1024D\/26B6AAE1  created: 1999-06-15 expires: never      trust: -\/u\nsub  2048g\/0CF8CB7A  created: 1999-06-15 expires: never     \nsub  1792G\/08224617  created: 1999-06-15 expires: 2002-06-14\nsub   960D\/B1F423E7  created: 1999-06-15 expires: 2002-06-14\n(1)  Chloe (Jester) &lt;chloe@cyb.org&gt;\n(2)  Chloe (Plebian) &lt;chloe@tel.net&gt;\n<code class=\"PROMPT\">Command&gt;<\/code> <code class=\"USERINPUT\"><b>check<\/b><\/code>\nuid  Chloe (Jester) &lt;chloe@cyb.org&gt;\nsig!       26B6AAE1 1999-06-15   [self-signature]\nuid  Chloe (Plebian) &lt;chloe@tel.net&gt;\nsig!       26B6AAE1 1999-06-15   [self-signature]<\/pre><\/td><\/tr><\/table>As expected, the signing key for each signature is the master signing\nkey with key ID<code class=\"LITERAL\">0x26B6AAE1<\/code>.\nThe self-signatures on the subkeys are present in the public key, but\nthey are not shown by the GnuPG interface.<\/div><div class=\"SECT2\"><p>Both new subkeys and new user IDs may be added to your keypair after\nit has been created.\nA user ID is added using the command \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#ADDUID\"><b class=\"COMMAND\">adduid<\/b><\/a>.\nYou are prompted for a real name, email address, and comment just\nas when you create an initial keypair.\nA subkey is added using the command \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#ADDKEY\"><b class=\"COMMAND\">addkey<\/b><\/a>.\nThe interface is similar to the interface used when creating an initial\nkeypair.\nThe subkey may be a DSA signing key, and encrypt-only ElGamal\nkey, or a sign-and-encrypt ElGamal key.\nWhen a subkey or user ID is generated it is self-signed using your \nmaster signing key, which is why you must supply your passphrase\nwhen the key is generated.<\/p><p>Additional user IDs are useful when you need multiple identities.\nFor example, you may have an identity for your job and an identity\nfor your work as a political activist.\nCoworkers will know you by your work user ID.\nCoactivists will know you by your activist user ID.\nSince those groups of people may not overlap, though, each group\nmay not trust the other user ID.\nBoth user IDs are therefore necessary.<\/p><p>Additional subkeys are also useful.\nThe user IDs associated with your public master key are validated by\nthe people with whom you\ncommunicate, and changing the master key therefore requires recertification.\nThis may be difficult and time consuming if you communicate with\nmany people.\nOn the other hand, it is good to periodically change encryption subkeys.\nIf a key is broken, all the data encrypted with that key will be\nvulnerable.\nBy changing keys, however, only the data encrypted with the one broken\nkey will be revealed.<\/p><p>Subkeys and user IDs may also be deleted.\nTo delete a subkey or user ID you must first select it using the\n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#KEY\"><b class=\"COMMAND\">key<\/b><\/a> or \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#UID\"><b class=\"COMMAND\">uid<\/b><\/a> commands respectively.\nThese commands are toggles.\nFor example, the command <b class=\"COMMAND\">key <code class=\"PARAMETER\"><i>2<\/i><\/code><\/b> \nselects the second subkey,\nand invoking <b class=\"COMMAND\">key <code class=\"PARAMETER\"><i>2<\/i><\/code><\/b> again \ndeselects it.\nIf no extra argument is given, all subkeys or user IDs are deselected.\nOnce the user IDs to be deleted are selected, the command\n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#DELUID\"><b class=\"COMMAND\">deluid<\/b><\/a>\nactually deletes the user IDs from your key.\nSimilarly, the command <a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#DELKEY\"><b class=\"COMMAND\">delkey<\/b><\/a>\ndeletes all selected subkeys from both your public and private keys.<\/p><p>For local keyring management, deleting key components is a good way\nto trim other people's public keys of unnecessary material.\nDeleting user IDs and subkeys on your own key, however, is not always\nwise since it complicates key distribution.\nBy default, when a user imports your updated public key it will be merged\nwith the old copy of your public key on his ring if it exists.\nThe components from both keys are combined in the merge, and this\neffectively restores any components you deleted.\nTo properly update the key, the user must first delete the old version\nof your key and then import the new version.\nThis puts an extra burden on the people with whom you communicate.\nFurthermore, if you send your key to a keyserver, the merge will\nhappen regardless, and anybody who downloads your key from a keyserver\nwill never see your key with components deleted.\nConsequently, for updating your own key it is better to revoke key\ncomponents instead of deleting them.<\/p><\/div><div class=\"SECT2\"><p>To revoke a subkey it must be selected.\nOnce selected it may be revoked with the \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#REVKEY\"><b class=\"COMMAND\">revkey<\/b><\/a> command.\nThe key is revoked by adding a revocation self-signature to the key.\nUnlike the command-line option --gen-revoke, the effect of\nrevoking a subkey is immediate.<\/p><table border=\"0\" style=\"background-color: #E0E0E0; width: 100%;\"><tr><td><pre class=\"SCREEN\"><code class=\"PROMPT\">Command&gt;<\/code> <code class=\"USERINPUT\"><b>revkey<\/b><\/code>\nDo you really want to revoke this key? y\n                                        \nYou need a passphrase to unlock the secret key for\nuser: \"Chloe (Jester) &lt;chloe@cyb.org&gt;\"\n1024-bit DSA key, ID B87DBA93, created 1999-06-28\n                  \npub  1024D\/B87DBA93  created: 1999-06-28 expires: never      trust: -\/u\nsub  2048g\/B7934539  created: 1999-06-28 expires: never     \nsub  1792G\/4E3160AD  created: 1999-06-29 expires: 2000-06-28\nrev! subkey has been revoked: 1999-06-29\nsub   960D\/E1F56448  created: 1999-06-29 expires: 2000-06-28\n(1)  Chloe (Jester) &lt;chloe@cyb.org&gt;\n(2)  Chloe (Plebian) &lt;chloe@tel.net&gt;<\/pre><\/td><\/tr><\/table><p>A user ID is revoked differently.\nNormally, a user ID collects signatures that attest that the user ID\ndescribes the person who actually owns the associated key.\nIn theory, a user ID describes a person forever, since that person will\nnever change.\nIn practice, though, elements of the user ID such as the email address\nand comment may change over time, thus invalidating the user ID.<\/p><p>The OpenPGP\nspecification does not support user ID revocation, but\na user ID can effectively be revoked by revoking the self-signature\non the user ID.\nFor the security reasons described \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/c235.html#AEN266\">previously<\/a>,\ncorrespondents will not trust a user ID with no valid self-signature.<\/p><p>A signature is revoked by using the command \n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#REVSIG\"><b class=\"COMMAND\">revsig<\/b><\/a>.\nSince you may have signed any number of user IDs, the user interface\nprompts you to decide for each signature whether or not to revoke it.<\/p><table border=\"0\" style=\"background-color: #E0E0E0; width: 100%;\"><tr><td><pre class=\"SCREEN\"><code class=\"PROMPT\">Command&gt;<\/code> <code class=\"USERINPUT\"><b>revsig<\/b><\/code>\nYou have signed these user IDs:\n     Chloe (Jester) &lt;chloe@cyb.org&gt;\n   signed by B87DBA93 at 1999-06-28\n     Chloe (Plebian) &lt;chloe@tel.net&gt;\n   signed by B87DBA93 at 1999-06-28\nuser ID: \"Chloe (Jester) &lt;chloe@cyb.org&gt;\"\nsigned with your key B87DBA93 at 1999-06-28\nCreate a revocation certificate for this signature? (y\/N)n\nuser ID: \"Chloe (Plebian) &lt;chloe@tel.net&gt;\"                \nsigned with your key B87DBA93 at 1999-06-28\nCreate a revocation certificate for this signature? (y\/N)y\nYou are about to revoke these signatures:                 \n     Chloe (Plebian) &lt;chloe@tel.net&gt;\n   signed by B87DBA93 at 1999-06-28\nReally create the revocation certificates? (y\/N)y\n                                                 \nYou need a passphrase to unlock the secret key for\nuser: \"Chloe (Jester) &lt;chloe@cyb.org&gt;\"\n1024-bit DSA key, ID B87DBA93, created 1999-06-28\n                  \npub  1024D\/B87DBA93  created: 1999-06-28 expires: never      trust: -\/u\nsub  2048g\/B7934539  created: 1999-06-28 expires: never     \nsub  1792G\/4E3160AD  created: 1999-06-29 expires: 2000-06-28\nrev! subkey has been revoked: 1999-06-29\nsub   960D\/E1F56448  created: 1999-06-29 expires: 2000-06-28\n(1)  Chloe (Jester) &lt;chloe@cyb.org&gt;\n(2)  Chloe (Plebian) &lt;chloe@tel.net&gt;<\/pre><\/td><\/tr><\/table><p>A revoked user ID is indicated by the revocation signature on \nthe ID when the signatures on the key's user IDs are listed.<\/p><table border=\"0\" style=\"background-color: #E0E0E0; width: 100%;\"><tr><td><pre class=\"SCREEN\"><code class=\"PROMPT\">Command&gt;<\/code> <code class=\"USERINPUT\"><b>check<\/b><\/code>\nuid  Chloe (Jester) &lt;chloe@cyb.org&gt;\nsig!       B87DBA93 1999-06-28   [self-signature]\nuid  Chloe (Plebian) &lt;chloe@tel.net&gt;\nrev!       B87DBA93 1999-06-29   [revocation]\nsig!       B87DBA93 1999-06-28   [self-signature]<\/pre><\/td><\/tr><\/table><p>Revoking both subkeys and self-signatures on user IDs adds revocation\nself-signatures to the key.\nSince signatures are being added and no material is deleted, a \nrevocation will always be visible to others when your updated public\nkey is distributed and merged with older copies of it.\nRevocation therefore guarantees that everybody has a consistent\ncopy of your public key.<\/p><\/div><div class=\"SECT2\"><p>The expiration time of a key may be updated with the command\n<a href=\"https:\/\/www.gnupg.org\/gph\/en\/manual\/r899.html#EXPIRE\"><b class=\"COMMAND\">expire<\/b><\/a> from the key edit menu.\nIf no key is selected the expiration time of the primary key\nis updated.\nOtherwise the expiration time of the selected subordinate key\nis updated.<\/p><p>A key's expiration time is associated with the key's self-signature.\nThe expiration time is updated by deleting the old self-signature\nand adding a new self-signature.\nSince correspondents will not have deleted the old self-signature, they\nwill see an additional self-signature on the key when they update\ntheir copy of your key.\nThe latest self-signature takes precedence, however, so all correspondents\nwill unambiguously know the expiration times of your keys.<\/p><\/div><\/div>","created_at":"2018-11-07T16:51:30+0000","updated_at":"2018-11-07T16:51:30+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"www.gnupg.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":543,"title":"Diagnose and Log to Console \u00a0|\u00a0 Tools for Web Developers\n      \u00a0|\u00a0 Google Developers","url":"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-write","content":"<section class=\"wf-byline\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\"><div class=\"attempt-left\">\n    <figure><img itemprop=\"image\" src=\"https:\/\/developers.google.com\/web\/images\/contributors\/pbakaus.jpg\" alt=\"Paul Bakaus\" \/><\/figure><\/div>\n  <section class=\"wf-byline-meta\"><div class=\"wf-byline-desc\">\n        Open Web Developer Advocate at Google \u2022 Tools, Performance, Animation, UX\n    <\/div>\n  <\/section><\/section><section class=\"wf-byline\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\"><div class=\"attempt-left\">\n    <figure><img itemprop=\"image\" src=\"https:\/\/developers.google.com\/web\/images\/contributors\/megginkearney.jpg\" alt=\"Meggin Kearney\" \/><\/figure><\/div>\n  <section class=\"wf-byline-meta\"><div class=\"wf-byline-desc\">\n        Meggin is a Tech Writer\n    <\/div>\n  <\/section><\/section><section class=\"wf-byline\" itemprop=\"author\" itemscope=\"itemscope\" itemtype=\"http:\/\/schema.org\/Person\"><div class=\"attempt-left\">\n    <figure><img itemprop=\"image\" src=\"https:\/\/developers.google.com\/web\/images\/contributors\/flaviocopes.jpg\" alt=\"Flavio Copes\" \/><\/figure><\/div>\n  <section class=\"wf-byline-meta\"><div class=\"wf-byline-desc\">\n        Flavio is a Full Stack Developer\n    <\/div>\n  <\/section><\/section>\nConsole logging is a powerful way to inspect what your page or application does. Let's start with console.log() and explore other advanced usage.\n<h3 class=\"hide-from-toc\" id=\"tldr_hide-from-toc\">TL;DR<\/h3><h2 id=\"writing_to_the_console\">Writing to the console<\/h2>\n<p>Use the <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/debug\/console\/console-reference#consolelogobject--object-\">console.log()<\/a> method for any basic logging to the console. It takes one or more expressions as parameters and writes their current values to the console, concatenating multiple parameters into a space-delimited line.<\/p>\n<p>Executing this line of code in your JavaScript:<\/p>\n<pre>console.log(\"Node count:\", a.childNodes.length, \"and the current time is:\", Date.now());\n<\/pre>\n<p>Will output this in the Console:\n<img alt=\"Log Multiple\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-log-multiple.png\" \/><\/p>\n<h2 id=\"autocomplete\">Autocompleting commands<\/h2>\n<p>When you type in the Console, the Console automatically displays an \nautocomplete dropdown menu of relevant methods that match the text that \nyou have already typed. This includes previous commands that you executed.<\/p>\n<p><img alt=\"example of autocomplete\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/autocomplete.png\" \/><\/p>\n<h2 id=\"organizing\">Organizing Console output<\/h2>\n<h3 id=\"group_messages_together\">Group messages together<\/h3>\n<p>You can group related output together with the group commands. The <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-reference#consolegroupobject-object-\"><code>console.group()<\/code><\/a> command takes a single string parameter to set the name of the group. After calling it in your JavaScript, the console will begin to group all subsequent output together.<\/p>\n<p>To end the grouping you only need to call <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-reference#consolegroupend\"><code>console.groupEnd()<\/code><\/a> when you're done.<\/p>\n<p>Example input:<\/p>\n<pre>var user = \"jsmith\", authenticated = false;\nconsole.group(\"Authentication phase\");\nconsole.log(\"Authenticating user '%s'\", user);\n\/\/ authentication code here...\nif (!authenticated) {\n    console.log(\"User '%s' not authenticated.\", user)\n}\nconsole.groupEnd();\n<\/pre>\n<p>Example output:\n<img alt=\"Simple console group output\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-group.png\" \/><\/p>\n<h4 id=\"nested_groups\">Nested groups<\/h4>\n<p>Log groups may also nest within each other. This is useful to see a large group in smaller pieces at a time.<\/p>\n<p>This example shows a log group for the authentication phase of a login process:<\/p>\n<pre>var user = \"jsmith\", authenticated = true, authorized = true;\n\/\/ Top-level group\nconsole.group(\"Authenticating user '%s'\", user);\nif (authenticated) {\n    console.log(\"User '%s' was authenticated\", user);\n    \/\/ Start nested group\n    console.group(\"Authorizing user '%s'\", user);\n    if (authorized) {\n        console.log(\"User '%s' was authorized.\", user);\n    }\n    \/\/ End nested group\n    console.groupEnd();\n}\n\/\/ End top-level group\nconsole.groupEnd();\nconsole.log(\"A group-less log trace.\");\n<\/pre>\n<p>And here's the nested groups output in the console:\n<img alt=\"Simple console group output\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-nestedgroup.png\" \/><\/p>\n<h4 id=\"auto-collapsing_groups\">Auto-collapsing groups<\/h4>\n<p>When using groups heavily, it can be very useful to not see everything as it happens. For these times you can automatically collapse groups by calling <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-reference#consolegroupcollapsedobject-object-\"><code>console.groupCollapsed()<\/code><\/a> instead of <code>console.group()<\/code>:<\/p>\n<pre>console.groupCollapsed(\"Authenticating user '%s'\", user);\nif (authenticated) {\n    ...\n}\nconsole.groupEnd();\n<\/pre>\n<p>groupCollapsed() output:\n<img alt=\"Initially collapsed group\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-groupcollapsed.png\" \/><\/p>\n<h2 id=\"errors_and_warnings\">Errors and warnings<\/h2>\n<p>Errors and warnings act the same way as normal logging. The only difference is <code>error()<\/code> and <code>warn()<\/code> have styles to bring attention to them.<\/p>\n<h3 id=\"consoleerror\">console.error()<\/h3>\n<p>The <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-reference#consoleerrorobject--object-\"><code>console.error()<\/code><\/a> method displays a red icon along with red message text:<\/p>\n<pre>function connectToServer() {\n    console.error(\"Error: %s (%i)\", \"Server is  not responding\",500);\n}\nconnectToServer();\n<\/pre>\n<p>turns into<\/p>\n<p><img alt=\"Error example output\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-error-server-not-resp.png\" \/><\/p>\n<h3 id=\"consolewarn\">console.warn()<\/h3>\n<p>The <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-reference#consolewarnobject--object-\"><code>console.warn()<\/code><\/a> method displays a yellow warning icon with the message text:<\/p>\n<pre>if(a.childNodes.length &lt; 3 ) {\n    console.warn('Warning! Too few nodes (%d)', a.childNodes.length);\n}\n<\/pre>\n<p>turns into<\/p>\n<p><img alt=\"Warn example\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-warning-too-few-nodes.png\" \/><\/p>\n<h2 id=\"assertions\">Assertions<\/h2>\n<p>The <a href=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/console-reference#consoleassertexpression-object\"><code>console.assert()<\/code><\/a> method conditionally displays an error string (its second parameter) only if its first parameter evaluates to <code>false<\/code>.<\/p>\n<h3 id=\"a_simple_assertion_and_how_it_displays\">A simple assertion and how it displays<\/h3>\n<p>The following code will cause an error message in the console only if the number of child nodes belonging to the <code>list<\/code> element is greater than 500.<\/p>\n<pre>console.assert(list.childNodes.length &lt;= 500, \"Node count is &gt; 500\");\n<\/pre>\n<p>How an assertion failure displays in the console:\n<img alt=\"Assertion failed\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-assert-failed.png\" \/><\/p>\n<h2 id=\"string_substitution_and_formatting\">String substitution and formatting<\/h2>\n<p>The first parameter passed to any of the logging methods may contain one or more format specifiers. A format specifier consists of a <code>%<\/code> symbol followed by a letter that indicates the formatting that applies to the value. The parameters following the string apply to the placeholders in order.<\/p>\n<p>The following example uses the string and digit formatters to insert values into the output string. You will see \"Sam has 100 points\" in the console.<\/p>\n<pre>console.log(\"%s has %d points\", \"Sam\", 100);\n<\/pre>\n<p>The full list of format specifiers is:<\/p>\n<table><thead><tr><th>Specifier<\/th>\n<th align=\"left\">Output<\/th>\n<\/tr><\/thead><tbody><tr><td>%s<\/td>\n<td align=\"left\">Formats the value as a string<\/td>\n<\/tr><tr><td>%i or %d<\/td>\n<td align=\"left\">Formats the value as an integer<\/td>\n<\/tr><tr><td>%f<\/td>\n<td align=\"left\">Formats the value as a floating point value<\/td>\n<\/tr><tr><td>%o<\/td>\n<td align=\"left\">Formats the value as an expandable DOM element. As seen in the Elements panel<\/td>\n<\/tr><tr><td>%O<\/td>\n<td align=\"left\">Formats the value as an expandable JavaScript object<\/td>\n<\/tr><tr><td>%c<\/td>\n<td align=\"left\">Applies CSS style rules to the output string as specified by the second parameter<\/td>\n<\/tr><\/tbody><\/table><p>This example uses the digit specifier to format the value of <code>document.childNodes.length<\/code>. It also uses the floating point specifier to format the value of <code>Date.now()<\/code>.<\/p>\n<p>The code:<\/p>\n<pre>console.log(\"Node count: %d, and the time is %f.\", document.childNodes.length, Date.now());\n<\/pre>\n<p>The output of the previous code sample:\n<img alt=\"Example substitution output\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-log-multiple.png\" \/><\/p>\n<h3 id=\"styling_console_output_with_css\">Styling console output with CSS<\/h3>\n<p>The CSS format specifier allows you to customize the display in the console.\nStart the string with the specifier and give the style you wish to apply as the second parameter.<\/p>\n<p>Try this code:<\/p>\n<pre>console.log(\"%cThis will be formatted with large, blue text\", \"color: blue; font-size: x-large\");\n<\/pre>\n<p>..to make your log output large and blue:<\/p>\n<p><img alt=\"Formatted string\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/console-write-format-string.png\" \/><\/p>\n<h3 id=\"formatting_dom_elements_as_javascript_objects\">Formatting DOM elements as JavaScript objects<\/h3>\n<p>By default, DOM elements are logged into the console as representation of their HTML, but sometimes you want to access the DOM element as JavaScript object and inspect its properties. You can use the <code>%O<\/code> string specifier to do that (see above), or use <code>console.dir<\/code> to achieve the same: <\/p>\n<p><img alt=\"Logging an element using dir()\" src=\"https:\/\/developers.google.com\/web\/tools\/chrome-devtools\/console\/images\/dir-element.png\" \/><\/p>\n<h2 id=\"feedback\">Feedback<\/h2>\n<p>Was this page helpful?<\/p>\n<div class=\"helpful--container\">\n  <section class=\"expandable helpful--section\"><aside class=\"success\">\n      Great! Thank you for the feedback.\n    <\/aside><\/section><section class=\"expandable helpful--section\"><aside class=\"warning\">\n      Sorry to hear that. Please <a href=\"https:\/\/github.com\/google\/WebFundamentals\/issues\/new\">open an\n      issue<\/a> and tell us how we can improve.\n    <\/aside><\/section><\/div>","created_at":"2018-11-07T16:51:30+0000","updated_at":"2018-11-07T16:51:30+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"developers.google.com","preview_picture":"https:\/\/developers.google.com\/web\/images\/social-webfu-16x9.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":541,"title":"HTTP Error 403","url":"http:\/\/www.gridsagegames.com\/blog\/2017\/05\/working-seeds\/","content":"HTTP Error 403\n<p>We're sorry, but we could not fulfill your request for\n\/blog\/2017\/05\/working-seeds\/ on this server.<\/p>\n<p>An invalid request was received from your browser. This may be caused by a malfunctioning proxy server or browser privacy software.<\/p>\n<p>Your technical support key is: <strong>68ec-57d0-1756-6707<\/strong><\/p>\n<p>You can use this key to <a href=\"http:\/\/www.ioerror.us\/bb2-support-key?key=68ec-57d0-1756-6707\">fix this problem yourself<\/a>.<\/p>\n<p>If you are unable to fix the problem yourself, please contact <a href=\"mailto:gridsagegames+nospam@nospam.gmail.com\">gridsagegames at gmail.com<\/a> and be sure to provide the technical support key shown above.<\/p>","created_at":"2018-11-07T16:51:29+0000","updated_at":"2018-11-07T16:51:29+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.gridsagegames.com","http_status":"403"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":538,"title":"A Quick Introduction to Graphviz","url":"https:\/\/www.worthe-it.co.za\/programming\/2017\/09\/19\/quick-introduction-to-graphviz.html","content":"<p>If you studied some form of computer science or software engineering,\n  you would have come across class diagrams. You also would have\n  encountered flow charts. If you then went into the business world, you\n  would have encountered the corporate alternative: boxes with words in\n  them joined arbitrarily by lines, usually on a whiteboard. I\u2019ve been\n  drawing a few of those myself recently.<\/p><p>I want to introduce to you an open source tool named Graphviz. It\n  allows you to create diagrams of the sort that you would have been\n  writing up on the whiteboard. You tell it what your boxes are, and how\n  to connect them. It outputs a picture.<\/p><h2>Graphs<\/h2><p>\u201cGraph\u201d is unfortunately an ambiguous word. When I hear it, I can\u2019t\n  help but think of a chart with an X axis and a Y axis. That is not\n  what I mean by graph in this post.<\/p><p>A graph, in the context of computer science, is a data structure. A\n  graph has a collection of \u201cnodes\u201d which are joined by \u201cedges\u201d.<\/p><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/graph.svg\" alt=\"\/assets\/posts\/graph.svg\" \/><\/p><p>This simple data structure, when drawn out on paper, can convey many\n  different situations. It can be how computers are connected in a\n  network (a network topology diagram). It can be how the steps in an\n  algorithm relate to each other (a flow chart). It can be the the\n  relationships between different classes in an object oriented system\n  (a class diagram). If each node is a concept and the edges are how\n  they relate to each other, then you have a mind map.<\/p><p>Graphs are especially useful as a visualization tool, since many\n  people understand the meaning of boxes connected by lines.<\/p><h2>Minimal documentation<\/h2><p>A picture speaks a thousand words but in the land of software\n  development, in my personal experience at least, we spend very little\n  time actually drawing well considered pictures. Documentation easily\n  descends into hastily written and awkwardly phrased paragraphs. I know\n  that I am personally guilty of writing some pretty nasty prose to\n  explain a feature under time pressure.<\/p><p>Sometimes, just a simple picture of the high level idea can make a big\n  difference to how easy something is to understand.<\/p><h2>Enter Graphviz<\/h2><p>Graphviz is an open source tool where you write a text description of\n  your graph (this is usually quite concise) and have it generate the\n  picture for you. You can then take the picture and embed it in\n  documents, email it around, or really use it however you want.<\/p><p>To get started, you can download Graphviz from <a href=\"http:\/\/graphviz.org\/Download..php\">Graphviz\u2019 web site<\/a>, or\n  from your system\u2019s package manager.<\/p><p>If you just want to try it out without installing anything, there is\n  an <a href=\"http:\/\/viz-js.com\/\">online version available<\/a>.<\/p><p>If you\u2019re an Emacs user, love keeping notes in org-mode like me, and\n  use Babel for embedding and running code blocks in your notes, then\n  you should <a href=\"http:\/\/orgmode.org\/worg\/org-contrib\/babel\/languages\/ob-doc-dot.html\">enable Graphviz in your org-babel config<\/a>.<\/p><h2>Cheat Sheet<\/h2><p>All of the examples below, you put into a text file with the extension\n  <code>.dot<\/code>. The executable that you need to run is also called <code>dot<\/code>.<\/p><p>For example, to create a png out of the file <code>hello.dot<\/code>, you would\n  run<\/p><div class=\"highlight\"><pre>dot -Tpng -ohello.png hello.dot\n<\/pre><\/div><p>I think the best way to give an idea of what you might do with\n  Graphviz is through some examples.<\/p><h3>The Basics<\/h3><p>Graphs can be either undirected<\/p><div class=\"highlight\"><pre>graph {\n  \/\/ C-style comments work\n  A -- B\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/simple-undirected.svg\" alt=\"\/assets\/posts\/simple-undirected.svg\" \/><\/p><p>or directed<\/p><div class=\"highlight\"><pre>digraph {\n  A -&gt; B\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/simple-directed.svg\" alt=\"\/assets\/posts\/simple-directed.svg\" \/><\/p><h3>Layout Direction<\/h3><p>If you don\u2019t want your graph to be laid out from top to bottom, it can\n  be laid out in any direction. My usual choice is left to right.<\/p><div class=\"highlight\"><pre>digraph {\n  \/\/ rankdir is a graph-level attribute\n  rankdir=\"LR\"\n  \/\/ you can declare multiple edges on one line\n  A -&gt; B -&gt; C\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/left_to_right.svg\" alt=\"\/assets\/posts\/left_to_right.svg\" \/><\/p><h3>Attributes<\/h3><p>You can separate node declarations from their logic if they\u2019re more\n  complicated.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"LR\"\n  \/\/ square brackets hold attributes\n  A [ label=\"First Node\" shape=\"circle\" ]\n  B [ label=\"Second Node\" shape=\"square\" ]\n  \/\/ edges have attributes too\n  A -&gt; B [ arrowhead=\"halfopen\" ]\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/nodes-and-edges.svg\" alt=\"\/assets\/posts\/nodes-and-edges.svg\" \/><\/p><h3>Ranks and Subgraphs<\/h3><p>To work out the layout, Graphviz uses a system it calls \u201cranks\u201d. Each\n  node is assigned a higher rank than the highest ranked node that point\n  to it. If your rank direction is set to left to right (<code>rankdir=LR<\/code>),\n  then nodes with a higher rank are placed further to the right.<\/p><p>Notice how all of the arrows on this graph point right.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"LR\"\n  \/\/ Nothing points to A, so its rank is 1\n  \/\/ A points to B, so B's rank is 2\n  A -&gt; B\n  \/\/ B's rank is higher than A's, so C gets a rank\n  \/\/ of 3 to be higher than B's rank\n  A -&gt; C\n  B -&gt; C\n  \/\/ D gets a rank of 4\n  C -&gt; D\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/ranks.svg\" alt=\"\/assets\/posts\/ranks.svg\" \/><\/p><p>You can use subgraphs to logically group nodes. Importantly for\n  layout, you can force all nodes in a subgraph to have the same rank.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"LR\"\n  A -&gt; B\n  A -&gt; C\n  B -&gt; C\n  C -&gt; D\n  subgraph subs {\n    rank=\"same\"\n    B\n    C\n  }\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/subgraphs.svg\" alt=\"\/assets\/posts\/subgraphs.svg\" \/><\/p><h3>Clusters<\/h3><p>If your subgraph name starts with \u2018cluster\u2019, then it actually gets a\n  line around it. You can even put a heading on a cluster.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"LR\"\n  \/\/ the normal ranking algorithm doesn't know what to\n  \/\/ do with clusters with rank=\"same\".\n  \/\/ If you opt in to the 'new' ranking algorithm, it\n  \/\/ works as expected.\n  newrank=\"true\"\n  A -&gt; B\n  A -&gt; C\n  B -&gt; C\n  C -&gt; D\n  subgraph cluster_subs {\n    label=\"Bs and Cs\"\n    rank=\"same\"\n    B\n    C\n  }\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/clusters.svg\" alt=\"\/assets\/posts\/clusters.svg\" \/><\/p><h3>Records<\/h3><p>The \u2018record\u2019 node shape is great for UML class diagrams.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"RL\"\n  \/\/ fields in the label are separated by a |\n  \/\/ ending a line in \\l left aligns it\n  Hero [ \n    shape=\"record\"\n    label=\"Hero|+ health : int\\l|+ save(k : Kingdom) : bool\\l\"\n  ]\n  \/\/ do whatever you want with whitespace\n  Villain [\n    shape=\"record\"\n    label=\"Villain|+ health : int\\l|+ brood() : void\\l\"\n  ]\n  Character [ shape=\"record\" ]\n  Hero -&gt; Character [ arrowhead=\"empty\" ]\n  Villain -&gt; Character [ arrowhead=\"empty\" ]\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/uml.svg\" alt=\"\/assets\/posts\/uml.svg\" \/><\/p><p>If your rankdir is vertical, then you need to use <code>{}<\/code> to change the\n  record type\u2019s direction.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"BT\"\n  Hero [ \n    shape=\"record\"\n    label=\"{Hero|+ health : int\\l|+ save(k : Kingdom) : bool\\l}\"\n  ]\n  Villain [\n    shape=\"record\"\n    label=\"{Villain|+ health : int\\l|+ brood() : void\\l}\"\n  ]\n  Character [ shape=\"record\" ]\n  Hero -&gt; Character [ arrowhead=\"empty\" ]\n  Villain -&gt; Character [ arrowhead=\"empty\" ]\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/uml-vertical.svg\" alt=\"\/assets\/posts\/uml-vertical.svg\" \/><\/p><h2>Further Reference<\/h2><p>For more information on the dot language, and how to make best use of\n  it, check the <a href=\"http:\/\/graphviz.org\/content\/dot-language\">official dot documentation<\/a> and the <a href=\"http:\/\/graphviz.org\/content\/attrs\">official list of all\n  supported attributes<\/a>.<\/p><h2>Bonus: COLOURS<\/h2><p>You can specify all of the colours used in the graph on a per node and\n  edge basis.<\/p><div class=\"highlight\"><pre>digraph {\n  rankdir=\"BT\"\n  bgcolor=\"#222222\"\n  \/\/ defaults for edges and nodes can be specified\n  node [ color=\"#ffffff\" fontcolor=\"#ffffff\" ]\n  edge [ color=\"#ffffff\" ]\n  \n  2 [fillcolor=\"#f22430\" style=filled color=\"#000000\" fontcolor=\"#000000\"]\n  4 [fillcolor=\"#f22430\" style=filled color=\"#000000\" fontcolor=\"#000000\"]\n  5 [fillcolor=\"#f22430\" style=filled color=\"#000000\" fontcolor=\"#000000\"]\n  6 [fillcolor=\"#f22430\" style=filled color=\"#000000\" fontcolor=\"#000000\"]\n  7 [fillcolor=\"#f22430\" style=filled color=\"#000000\" fontcolor=\"#000000\"]\n  \n  1 -&gt; 2\n  1 -&gt; 3\n  1 -&gt; 5\n  2 -&gt; 4 [color=\"#f22430\"]\n  2 -&gt; 6 [color=\"#f22430\"]\n  3 -&gt; 4 \n  3 -&gt; 7 \n  4 -&gt; 8\n  5 -&gt; 6 [color=\"#f22430\"]\n  5 -&gt; 7 [color=\"#f22430\"]\n  6 -&gt; 8\n  7 -&gt; 8\n}\n<\/pre><\/div><p><img src=\"https:\/\/www.worthe-it.co.za\/assets\/posts\/graph-clours.svg\" alt=\"\/assets\/posts\/graph-clours.svg\" \/><\/p>","created_at":"2018-11-07T16:51:26+0000","updated_at":"2018-11-07T16:51:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.worthe-it.co.za","preview_picture":"https:\/\/www.worthe-it.co.za\/assets\/thumbs\/graphs.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":539,"title":"FZF and RipGrep \u2013 Navigate with bash faster than ever before","url":"https:\/\/news.ycombinator.com\/item?id=15514589","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15514589\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15514589\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/owen.cymru\/fzf-ripgrep-navigate-with-bash-faster-than-ever-before\/\" class=\"storylink\">FZF and RipGrep \u2013 Navigate with bash faster than ever before<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=owen.cymru\">owen.cymru<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        171 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=Midasx\" class=\"hnuser\">Midasx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15514589\">on Oct 20, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15514589&amp;goto=item%3Fid%3D15514589\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=FZF%20and%20RipGrep%20%E2%80%93%20Navigate%20with%20bash%20faster%20than%20ever%20before&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=FZF%20and%20RipGrep%20%E2%80%93%20Navigate%20with%20bash%20faster%20than%20ever%20before\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15514589&amp;auth=97f20c73c1701e89421abd1800d7913d99d41ea0\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15514589\">53\u00a0comments<\/a>              <\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"15516885\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cryptonector\" class=\"hnuser\">cryptonector<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516885\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515273\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=guitarbill\" class=\"hnuser\">guitarbill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515273\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515430\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nerdponx\" class=\"hnuser\">nerdponx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515430\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516562\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=oblio\" class=\"hnuser\">oblio<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516562\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516841\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jandrese\" class=\"hnuser\">jandrese<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516841\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15517667\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=okdana\" class=\"hnuser\">okdana<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15517667\">on Oct 20, 2017<\/a><\/p><br \/><div class=\"comment\">&gt; There are good reasons to use Zsh beyond its autocomplete system...<p>zsh also performs much better in practice than bash. Not only does it run the same code faster, it subsumes a great deal of functionality that normally requires external utility calls (which are very slow) into the shell itself.<\/p><p>For example:<\/p><p>* zsh has native floating-point arithmetic, so no need for `awk` or whatever. It can even format (digit-group) the numbers<\/p><p>* Regular-expression matching with =~ can be replaced by extended globs (which offer similar functionality but are significantly faster in most cases)<\/p><p>* `basename`, `dirname`, `readlink -m`, and `readlink -f` can be replaced by parameter-expansion modifiers<\/p><p>* `sort` can be replaced by parameter-expansion flags<\/p><p>* `grep` can be replaced by parameter-expansion flags and extended globs<\/p><p>* `find` (including its `-type` and `-exec` features) can be replaced by globs<\/p><p>* `date` can be replaced by prompt expansion or the `zsh\/datetime` module<\/p><p>* `cp`, `mkdir`, `rm`, and so on can be replaced by the `zsh\/files` module<\/p><p>* `stat` can be replaced by the `zsh\/stat` module<\/p><p>* `column` can be replaced by `print -c` or `print -C`<\/p><p>&gt; there are massive swaths of the program, entire subsystems, that are more or less completely undocumented<\/p><p>Like what? I'm not sure i'd noticed that myself. Sometimes the documentation is <i>vague<\/i>, maybe hard to find (e.g., the completion system's documentation is a bit overwhelming), but it's always been there when i went looking for it.<\/p><p>&gt; Also, the ad-hoc \"plugin\" ecosystem is in a strange and fragmented state.<\/p><p>The plug-in ecosystem (i assume you mean OMZ, zplug, and stuff like that) is entirely unofficial and most of the zsh developers don't seem to care for any of it because it tends to be slow, error-prone, and often just unnecessary. It's also a major support burden for the people on IRC and in the mailing list, which doesn't endear them to it. Might be cool to have something official though.<\/p><p>&gt; It's a shame that more people don't write actual modules in C (yes, Zsh has a native module system), but then again, is the API even documented?<\/p><p>I don't think it is, there's just an example module you can build. bash also supports loadable modules and it's the same way AFAIK.<\/p><p>&gt; One more thing about Zsh: it has emacs syndrome. Do your really need an FTP client and calendar built into your shell?<\/p><p>Most of the weird stuff like the FTP client and calendar and Tetris game are optional modules or even just regular shell scripts. None of them are enabled by default and packagers can omit them (and they often do, especially with static builds).\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15524031\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nerdponx\" class=\"hnuser\">nerdponx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15524031\">on Oct 21, 2017<\/a><\/p><br \/><div class=\"comment\">&gt; zsh also performs much better in practice than bash. Not only does it run the same code faster, it subsumes a great deal of functionality that normally requires external utility calls (which are very slow) into the shell itself.<p>For the uninitiated: Zsh parameters expansions and globbing (called \"filename generation\" in the docs) are beautiful. As I said above, I don't write Bash scripts, I write Zsh scripts, and this kind of thing is why:<\/p><pre>    % cd \/tmp\n    % mkdir example\n    % cd example\n    % touch foo\n    % touch bar\n    % mkdir \"my documents\"\n    % touch \"my documents\/baz\"\n    % tree -N\n    .\n    \u251c\u2500\u2500 bar\n    \u251c\u2500\u2500 foo\n    \u2514\u2500\u2500 my documents\n        \u2514\u2500\u2500 baz\n    1 directory, 3 files\n    % for f in **\/*(.);\n    &gt;   do echo File ${f:t} lives in ${f:A:h}\n    &gt; done\n    File bar lives in \/tmp\/example\n    File foo lives in \/tmp\/example\n    File baz lives in \/tmp\/example\/my documents\n<\/pre>&gt; Like what? I'm not sure i'd noticed that myself. Sometimes the documentation is vague, maybe hard to find (e.g., the completion system's documentation is a bit overwhelming), but it's always been there when i went looking for it.<p>Fair enough. Although IMO \"vague and disorganized\" is just as good as \"nonexistent\". Just try reading the docs for `autoload`\/`typeset`\/`local`, or `zstyle` [0], or `zmodload`. The whole thing badly needs to be tagged and reverse-indexed by functionality (\"how do I accomplish X\"), not by flag (\"what does foo -q mean?\"). I wrote a long comment about this kind of documentation recently [1].<\/p><p>&gt; The plug-in ecosystem (i assume you mean OMZ, zplug, and stuff like that) is entirely unofficial and most of the zsh developers don't seem to care for any of it because it tends to be slow, error-prone, and often just unnecessary. It's also a major support burden for the people on IRC and in the mailing list, which doesn't endear them to it. Might be cool to have something official though.<\/p><p>Yes, this. It seems like everyone and their mother at one point decided to try to write a \"package manager\" for Zsh.<\/p><p>But I think the problem with these unofficial plugins is mainly that sourcing thousands of lines of shell script on every terminal load, and running potentially dozens of functions before every command, is just plain slow. Zsh might be faster than Bash, but it's still really really slow. Just looping over 1..1000000 takes well over a second in Zsh, but less than 1\/2 second in Python (which is already considered a slow language).<\/p><p>What I haven't tried is zcompile-ing all my startup scripts. Would that help?<\/p><p>And of course, actually documenting the C API and build process would be a good first step for the maintainers and devs. The example module is easy enough to follow, but actually writing a module is another matter entirely. This is especially frustrating because AFAICT Zsh builtins are themselves implemented as a statically-linked module called \"zsh\/main\", so it must be a powerful system.<\/p><p>[0]: <a href=\"https:\/\/unix.stackexchange.com\/q\/214657\" rel=\"nofollow\">https:\/\/unix.stackexchange.com\/q\/214657<\/a><\/p><p>[1]: <a href=\"https:\/\/news.ycombinator.com\/item?id=15479273\" rel=\"nofollow\">https:\/\/news.ycombinator.com\/item?id=15479273<\/a>\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516096\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mason55\" class=\"hnuser\">mason55<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516096\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516251\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Midasx\" class=\"hnuser\">Midasx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516251\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15520524\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chewz\" class=\"hnuser\">chewz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15520524\">on Oct 21, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516402\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mason55\" class=\"hnuser\">mason55<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516402\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515967\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=corpMaverick\" class=\"hnuser\">corpMaverick<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515967\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516072\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dsissitka\" class=\"hnuser\">dsissitka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516072\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516050\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=goldfishparty\" class=\"hnuser\">goldfishparty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516050\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15519516\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15519516\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15520080\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=burntsushi\" class=\"hnuser\">burntsushi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15520080\">on Oct 21, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15521404\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15521404\">on Oct 21, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516576\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ssijak\" class=\"hnuser\">ssijak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516576\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516081\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kuwze\" class=\"hnuser\">kuwze<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516081\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15517870\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=perryprog\" class=\"hnuser\">perryprog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15517870\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516422\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=corpMaverick\" class=\"hnuser\">corpMaverick<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516422\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515640\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=graffitici\" class=\"hnuser\">graffitici<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515640\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516020\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=p1xelHer0\" class=\"hnuser\">p1xelHer0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516020\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516962\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jandrese\" class=\"hnuser\">jandrese<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516962\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515675\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=haar\" class=\"hnuser\">haar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515675\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15517038\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gepoch\" class=\"hnuser\">gepoch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15517038\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516490\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=finnh\" class=\"hnuser\">finnh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516490\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515853\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kbd\" class=\"hnuser\">kbd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515853\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515789\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloeki\" class=\"hnuser\">lloeki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515789\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516991\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cryptonector\" class=\"hnuser\">cryptonector<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516991\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515936\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Midasx\" class=\"hnuser\">Midasx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515936\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15518830\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tavianator\" class=\"hnuser\">tavianator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15518830\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515997\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=HurrdurrHodor\" class=\"hnuser\">HurrdurrHodor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515997\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516084\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Midasx\" class=\"hnuser\">Midasx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516084\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515904\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=herrvogel-\" class=\"hnuser\">herrvogel-<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515904\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516994\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmoriarty\" class=\"hnuser\">pmoriarty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516994\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15520490\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chewz\" class=\"hnuser\">chewz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15520490\">on Oct 21, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516179\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=masklinn\" class=\"hnuser\">masklinn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516179\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15519659\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vthriller\" class=\"hnuser\">vthriller<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15519659\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15517207\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jandrese\" class=\"hnuser\">jandrese<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15517207\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15517595\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=masklinn\" class=\"hnuser\">masklinn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15517595\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15518613\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=therein\" class=\"hnuser\">therein<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15518613\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516833\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmoriarty\" class=\"hnuser\">pmoriarty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516833\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515435\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paule89\" class=\"hnuser\">paule89<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515435\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515477\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=burntsushi\" class=\"hnuser\">burntsushi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515477\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515905\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pja\" class=\"hnuser\">pja<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515905\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515718\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloeki\" class=\"hnuser\">lloeki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515718\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515914\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=burntsushi\" class=\"hnuser\">burntsushi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515914\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515617\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaushalmodi\" class=\"hnuser\">kaushalmodi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515617\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15515631\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=burntsushi\" class=\"hnuser\">burntsushi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15515631\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516146\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=steveklabnik\" class=\"hnuser\">steveklabnik<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516146\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15516853\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tycho\" class=\"hnuser\">Tycho<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15516853\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15517108\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fernandotakai\" class=\"hnuser\">fernandotakai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15517108\">on Oct 20, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:51:26+0000","updated_at":"2018-11-07T16:51:26+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":540,"title":"Content Based Intelligent Cropping","url":"http:\/\/engineering.curalate.com\/2017\/04\/13\/content-based-intelligent-cropping.html","content":"<p>Square pegs don\u2019t fit in round holes, but what if you have power tools?<\/p><p>Digital images often don\u2019t fit where we want them: advertisements, social\nnetworks, and printers all require that images be a specific <em>aspect ratio<\/em>\n(i.e., the ratio of the image\u2019s width to height). Take <a href=\"https:\/\/www.facebook.com\/business\/help\/103816146375741\">Facebook\nads<\/a> for example:\ndifferent aspect ratios are required depending on what kind of ad you wish to\nrun. This is a large pain point for marketers: each piece of content must be\nmanually cropped to fit the aspect ratio of the channel. Typically, images are\neither padded with white pixels (thus wasting valuable screen real estate) or\narbitrarily cropped (possibly degrading the content).<\/p><p>But it doesn\u2019t have to be this way! In this post, we present a technique that we\nuse for <em>intelligent cropping<\/em>: a fully automatic method that preserves the\nimage\u2019s content. We\u2019ve included some example code so you can explore on your\nown, and some real-world examples from Curalate\u2019s products.<\/p><p>The following illustrates our approach:<\/p><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/overview.png\" alt=\"\" \/><\/p><ul><li>The input to the algorithm is an image and a desired aspect ratio.<\/li>\n  <li>First, we use a variety of techniques to detect different types of content in\nthe image. Each technique results in a number of <em>content rectangles<\/em> that are\nassigned a value score.<\/li>\n  <li>Second, we select the optimal region of the image as that which contains the\ncontent rectangles with the higest cumulative score.<\/li>\n  <li>Finally, we crop the input image to the optimal region.<\/li>\n<\/ul><p>The result is a cropped image of the desired aspect ratio fully containing the\ncontent in the image.<\/p><ul id=\"markdown-toc\"><li><a href=\"#prerequisites\" id=\"markdown-toc-prerequisites\">Prerequisites<\/a><\/li>\n  <li><a href=\"#identifying-content-in-images\" id=\"markdown-toc-identifying-content-in-images\">Identifying Content in Images<\/a>    <ul><li><a href=\"#face-detection\" id=\"markdown-toc-face-detection\">Face Detection<\/a><\/li>\n      <li><a href=\"#interest-points\" id=\"markdown-toc-interest-points\">Interest Points<\/a><\/li>\n      <li><a href=\"#product-detection\" id=\"markdown-toc-product-detection\">Product Detection<\/a><\/li>\n      <li><a href=\"#content-regions\" id=\"markdown-toc-content-regions\">Content Regions<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#optimal-cropping\" id=\"markdown-toc-optimal-cropping\">Optimal Cropping<\/a>    <ul><li><a href=\"#reducing-to-one-dimension\" id=\"markdown-toc-reducing-to-one-dimension\">Reducing to One Dimension<\/a><\/li>\n      <li><a href=\"#selecting-the-optimal-crop\" id=\"markdown-toc-selecting-the-optimal-crop\">Selecting the Optimal Crop<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#result-gallery\" id=\"markdown-toc-result-gallery\">Result Gallery<\/a><\/li>\n  <li><a href=\"#uses-in-curalate-products\" id=\"markdown-toc-uses-in-curalate-products\">Uses in Curalate Products<\/a><\/li>\n<\/ul><h2 id=\"prerequisites\">Prerequisites<\/h2><p>To run these examples for yourself, you\u2019ll need Python 2 with OpenCV, NumPy, and\nmatplotlib installed. The images used for examples in this post may be\ndownloaded <a href=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/exampleImages.zip\">here<\/a>.\nThis entire post is also available as a <a href=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-Int%0AelligentCropping\/Content%20Based%20Intelligent%20Cropping.ipynb\">python notebook<\/a> if you want to\ntake it for a spin.<\/p><p>To start off, let\u2019s load an image we\u2019d like to use:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">import cv2\nimport urllib\nimport numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline\ndef showImage(img):\n    plt.axis('off')\n    plt.imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))\nimg = cv2.imread(\"input.jpg\")\nshowImage(img)<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_1_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><p>Let\u2019s assume we\u2019re creating a Facebook ad to drive traffic to our website. The\n<a href=\"https:\/\/www.facebook.com\/business\/help\/103816146375741\">recommended resolution<\/a>\nis <code class=\"highlighter-rouge\">1200x628<\/code> for target aspect ratio of <code class=\"highlighter-rouge\">1.91<\/code>.<\/p><p>The naive approach would just crop the center of the image:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">desiredAspectRatio = 1200 \/ float(628)\nnewHeight = img.shape[1] \/ desiredAspectRatio\nstart = img.shape[0] \/ 2 - newHeight \/ 2\nnaiveCrop = img[start:start + newHeight, :]\nshowImage(naiveCrop)<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_3_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><p>Ugh. I wouldn\u2019t click on that. Let\u2019s do something intelligent!<\/p><p>Our first task is to detect different content in the image. Object detection is\nstill an active area of research, though <a href=\"http:\/\/arxiv.org\/abs\/1311.2524\">recent\nadvances<\/a> have started to make it feasible in\nmany applications. Here we explore a few simple techniques that are built into\nOpenCV but you can use any detector you like.<\/p><h2 id=\"face-detection\">Face Detection<\/h2><p>If an image contains a face, it\u2019s likely that the person is a key element in the\nimage. Fortunately, face detection is a common task in computer vision:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)\ncascade = cv2.CascadeClassifier(\"haarcascade_frontalface_default.xml\")\nfaceRegions = cascade.detectMultiScale(gray, minNeighbors=7, scaleFactor=1.1)<\/pre><\/figure><p>The result is a numpy array of rectangles containing the faces:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">def drawRegions(source, res, regions, color=(0, 0, 255), size=4):\n    for (x, y, w, h) in regions:\n        res[y: y + h, x: x + w] = source[y: y + h, x: x + w]\n        cv2.rectangle(res, (x, y), (x + w, y + h), color, size)\n    return res\nfaded = (img * 0.65).astype(np.uint8)\nshowImage(drawRegions(img, faded.copy(), faceRegions))<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_8_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><h2 id=\"interest-points\">Interest Points<\/h2><p>Sometimes, we don\u2019t know what we\u2019re looking for in an image. Low-level image\ncharacteristics, however, often correspond to the interesting area of images.\nThere are many common techniques for identifying interesting areas of an image,\neven ones that estimate visual saliency. Shi-Tomasi\u2019s <a href=\"http:\/\/www.ai.mit.edu\/courses\/6.891\/handouts\/shi94good.pdf\">Good Features To\nTrack<\/a> is one\ntechnique commonly used to indicate interest points in an image. Detecting these\ninterest points is also relatively simple using OpenCv:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">interestPoints = cv2.goodFeaturesToTrack(gray, maxCorners=200, qualityLevel=0.01, minDistance=20).reshape(-1, 2)\ninterestPointRegions = np.concatenate((interestPoints, np.ones(interestPoints.shape)), axis=1).astype(np.int32)\nshowImage(drawRegions(img, faded.copy(), interestPointRegions, (255, 255, 255), size=10))<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_10_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><h2 id=\"product-detection\">Product Detection<\/h2><p>Other times, we know a specific product is in an image and we want to make sure\nwe don\u2019t crop it out. We can achieve this by localizing an image of the product\nin our image of interest.<\/p><p>In our example, the product is:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">productImage = cv2.imread(\"product.jpg\")\nshowImage(productImage)<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_12_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><p>We can locate the product in the image using instance retrieval techniques.\nFirst, we\u2019ll estimate the transformation between the product and the target\nimage:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">flann = cv2.FlannBasedMatcher({'algorithm': 0, 'trees': 8}, {'checks': 100})\ndetector = cv2.SIFT()\nkpts1, descs1 = detector.detectAndCompute(productImage, None)\nkpts2, descs2 = detector.detectAndCompute(img, None)\nmatches = [m for (m, n) in flann.knnMatch(descs1, descs2, k=2) if m.distance &lt; 0.8 * n.distance]\nsourcePoints = np.float32([kpts1[m.queryIdx].pt for m in matches]).reshape(-1, 2)\ndestPoints = np.float32([kpts2[m.trainIdx].pt for m in matches]).reshape(-1, 2)\nM, mask = cv2.findHomography(sourcePoints, destPoints, cv2.RANSAC, 11.0)<\/pre><\/figure><p>The result is a set of correspondence points between the images:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">def drawMatches(img1, kpts1, img2, kpts2, matches):\n    # combine both images\n    out = np.zeros((max([img1.shape[0], img2.shape[0]]), img1.shape[1] + img2.shape[1], 3), dtype='uint8')\n    out[: img1.shape[0], : img1.shape[1]] = img1\n    out[: img2.shape[0], img1.shape[1]:] = img2\n    # draw the lines\n    for match in matches:\n        (x1, y1) = kpts1[match.queryIdx].pt\n        (x2, y2) = kpts2[match.trainIdx].pt\n        cv2.line(out, (int(x1), int(y1)), (int(x2) + img1.shape[1], int(y2)), (0, 0, 255), 4)\n    return out\nshowImage(drawMatches(productImage, kpts1, img, kpts2, np.array(matches)[[np.where(mask.ravel() == 1)[0]]]))<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_16_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><p>We simply take the bounding box around the product\u2019s location:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">pp = destPoints[mask.ravel() == 1]\nxmin = pp[:, 0].min()\nymin = pp[:, 1].min()\nproductRegions = np.array([xmin, ymin, pp[:, 0].max() - xmin, pp[:, 1].max() - ymin]).astype(np.int32).reshape(1, 4)\nshowImage(drawRegions(img, faded.copy(), productRegions, (0, 255, 0)))<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_18_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><h2 id=\"content-regions\">Content Regions<\/h2><p>In summary, we have detected faces, interest points, and products in the image.\nTogether, these form the full set of content regions:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">contentRectangles = np.concatenate((faceRegions, productRegions, interestPointRegions), axis=0)\nvis = faded.copy()\ndrawRegions(img, vis, interestPointRegions, (255, 255, 255), size=10)\ndrawRegions(img, vis, faceRegions)\ndrawRegions(img, vis, productRegions, (0, 255, 0))\nshowImage(vis)<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_20_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><p>Now that we have detected the content regions in the image, we\u2019d like to\nidentify the best way to crop the image to a desired aspect ratio of <code class=\"highlighter-rouge\">1.91<\/code>. The\nstrategy is simple: find the area of the image with the desired aspect ratio\ncontaining the highest sum of the content rectangle scores.<\/p><p>First, let\u2019s assign a score to each content rectangle. For this example, we\u2019ll\njust use the area of each rectangle.<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">contentScores = np.multiply(contentRectangles[:, 2], contentRectangles[:, 3])<\/pre><\/figure><h2 id=\"reducing-to-one-dimension\">Reducing to One Dimension<\/h2><p>Now for the fun part:\nDepending on the input image and desired aspect ratio, the resulting crop will\neither have the same height as the input image and a reduced width, or the same\nwidth as the input image and a reduced height. The <em>principal axis<\/em> is the\ndimension of the input image that needs to be cropped. Let:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">alpha = img.shape[1] \/ float(img.shape[0])<\/pre><\/figure><p>be the aspect ratio of the input image. If <code class=\"highlighter-rouge\">alpha &gt; desiredAspectRatio<\/code>, then\nthe horizontal axis is the principal axis and the system crops the width of the\nimage. Similarly, if <code class=\"highlighter-rouge\">alpha &lt; desiredAspectRatio<\/code>, then the vertical axis is the\nprincipal axis and the system crops the height of the image.<\/p><p>Projecting the content rectangles onto the principal axis simplifies our goal:\nthe optimal crop is simply the window along the principal axis containing the\nhighest sum of content region scores. The length of this window is the size of\nthe final crop along the principal axis.<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">if (alpha &gt; desiredAspectRatio):  # the horizontal axis is the principal axis.\n    finalWindowLength = int(desiredAspectRatio * img.shape[0])\n    projection = np.array([[1, 0, 0, 0], [0, 0, 1, 0]])\nelse:  # the vertical axis is the principal axis.\n    finalWindowLength = int(img.shape[1] \/ desiredAspectRatio)\n    projection = np.array([[0, 1, 0, 0], [0, 0, 0, 1]])\ncontentRegions = np.dot(projection, contentRectangles.T).T<\/pre><\/figure><p>Thus, the content rectangles are reduced from two dimensional rectangles to one\ndimensional regions.<\/p><h2 id=\"selecting-the-optimal-crop\">Selecting the Optimal Crop<\/h2><p>The optimal crop is the window of length <code class=\"highlighter-rouge\">finalWindowLength<\/code> whose\n<code class=\"highlighter-rouge\">contentRegions<\/code>\u2019 scores sum to the largest possible value. We can use a sliding\nwindow approach to quickly and efficiently find such a crop.<\/p><p>First, we\u2019ll define the <em>inflection points<\/em> for the sliding window approach.\nEach inflection point is a location on the number line where the value of the\ncurrent window can change. There are two inflection points for each content\nregion: one that removes the content region\u2019s score when the window passes the\nregion\u2019s starting location, and one that adds a content region\u2019s score when the\nwindow encapsulates it.<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">inflectionPoints = np.concatenate((contentRegions[:, 0], contentRegions[:, 0] + contentRegions[:, 1] - finalWindowLength))\ninflectionDeltas = np.concatenate((-contentScores, contentScores))\ninflections = np.concatenate((inflectionPoints.reshape(-1, 1), inflectionDeltas.reshape(-1, 1)), axis=1)<\/pre><\/figure><p>Next, we\u2019ll sort the inflection points by their locations on the number line,\nand ignore any outside the valid range:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">inflections = inflections[inflections[:, 0].argsort()]  # Sort by location\ninflections = inflections[inflections[:, 0] &gt;= 0]  # drop any outside our range<\/pre><\/figure><p>To implement our sliding window algorithm, we need only accumulate the sum of\nthe inflection points\u2019 values at each location, and then take the maximum:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">inflections[:, 1] = np.cumsum(inflections[:, 1])\noptimalInflectionPoint = max(enumerate(inflections), key=lambda (idx, (s, v)): v)[0]<\/pre><\/figure><p>The <code class=\"highlighter-rouge\">optimalInflectionPoint<\/code> contains a starting location that has the most\nvalue. In fact the range of pixels between that inflection point and the next\none all have that same value. We\u2019ll take the middle of that range for our\nstarting point:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">optimalStartingLocation = (inflections[optimalInflectionPoint, 0] + inflections[optimalInflectionPoint + 1, 0]) \/ 2<\/pre><\/figure><p>Now that we know where the optimal crop begins on the principal axis, we can un-\nproject it to get the final crop:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">if alpha &gt; desiredAspectRatio:\n    optimalCrop = [optimalStartingLocation, 0, finalWindowLength, img.shape[0]]\nelse:\n    optimalCrop = [0, optimalStartingLocation, img.shape[1], finalWindowLength]<\/pre><\/figure><p>Awesome! Now we know where to crop the image! You can see below that the optimal\ncrop indeed includes the product, the face, and a large number of the interest\npoints:<\/p><figure class=\"highlight\"><pre class=\"language-python\" data-lang=\"python\">result = img[optimalCrop[1]: optimalCrop[3] + optimalCrop[1], optimalCrop[0]: optimalCrop[2] + optimalCrop[0]]\nshowImage(result)<\/pre><\/figure><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/files\/content-based-intelligent-cropping_38_0.png\" alt=\"png\" class=\"center-image\" \/><\/p><p>Now that\u2019s a good pic!<\/p><p>Disclaimer: The code above is meant as a demonstration. Optimization, handling\nof edge cases, and parameter tuning are left as an exercise for the reader \ud83d\ude09.<\/p><p>Below are some example results. The desired aspect ratio is listed below the\ninput image.<\/p><p><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/results1.jpg\" alt=\"\" class=\"center-image\" \/><\/p><p>One great place we use intelligent cropping is when displaying our clients\u2019\nimages. Below is a screenshot showing some product images before intelligent\ncropping, and then after. Notice how the models\u2019 faces, the shoe, and the bag\nwere all cropped using the naive method. After intelligent cropping, our\nthumbnails are much more useful representations of the original images.<\/p><table><thead><tr><th>Before Intelligent Cropping<\/th>\n      <th>After Intelligent Cropping<\/th>\n    <\/tr><\/thead><tbody><tr><td><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/curalate-before.png\" alt=\"\" \/><\/td>\n      <td><img src=\"http:\/\/engineering.curalate.com\/assets\/2017-04-13-IntelligentCropping\/curalate-after.png\" alt=\"\" \/><\/td>\n    <\/tr><\/tbody><\/table>","created_at":"2018-11-07T16:51:26+0000","updated_at":"2018-11-07T16:51:26+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"engineering.curalate.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":536,"title":"The Mutt E-Mail Client","url":"http:\/\/www.mutt.org\/doc\/manual\/","content":"<div class=\"sect1\"><div class=\"titlepage\"><div><p><\/p><h2 class=\"title\">1.\u00a0Command-Line Options<\/h2><\/div><\/div><p>\nRunning <code class=\"literal\">mutt<\/code> with no arguments will make Mutt attempt\nto read your spool mailbox.  However, it is possible to read other\nmailboxes and to send messages from the command line as well.\n<\/p><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.1.\u00a0Command line options<\/strong><\/p><p><\/p><table class=\"table\" summary=\"Command line options\" border=\"1\"><colgroup><col \/><col \/><\/colgroup><thead><tr><th>Option<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td>-A<\/td><td>expand an alias<\/td><\/tr><tr><td>-a<\/td><td>attach a file to a message<\/td><\/tr><tr><td>-b<\/td><td>specify a blind carbon-copy (BCC) address<\/td><\/tr><tr><td>-c<\/td><td>specify a carbon-copy (Cc) address<\/td><\/tr><tr><td>-d<\/td><td>log debugging output to ~\/.muttdebug0 if mutt was compiled with +DEBUG; it can range from 1-5 and affects verbosity (a value of 2 is recommended)<\/td><\/tr><tr><td>-D<\/td><td>print the value of all Mutt variables to stdout<\/td><\/tr><tr><td>-E<\/td><td>edit the draft (-H) or include (-i) file<\/td><\/tr><tr><td>-e<\/td><td>specify a config command to be run after initialization files are read<\/td><\/tr><tr><td>-f<\/td><td>specify a mailbox to load<\/td><\/tr><tr><td>-F<\/td><td>specify an alternate file to read initialization commands<\/td><\/tr><tr><td>-h<\/td><td>print help on command line options<\/td><\/tr><tr><td>-H<\/td><td>specify a draft file from which to read a header and body<\/td><\/tr><tr><td>-i<\/td><td>specify a file to include in a message composition<\/td><\/tr><tr><td>-m<\/td><td>specify a default mailbox type<\/td><\/tr><tr><td>-n<\/td><td>do not read the system Muttrc<\/td><\/tr><tr><td>-p<\/td><td>recall a postponed message<\/td><\/tr><tr><td>-Q<\/td><td>query a configuration variable<\/td><\/tr><tr><td>-R<\/td><td>open mailbox in read-only mode<\/td><\/tr><tr><td>-s<\/td><td>specify a subject (enclose in quotes if it contains spaces)<\/td><\/tr><tr><td>-v<\/td><td>show version number and compile-time definitions<\/td><\/tr><tr><td>-x<\/td><td>simulate the mailx(1) compose mode<\/td><\/tr><tr><td>-y<\/td><td>show a menu containing the files specified by the <strong>mailboxes<\/strong> command<\/td><\/tr><tr><td>-z<\/td><td>exit immediately if there are no messages in the mailbox<\/td><\/tr><tr><td>-Z<\/td><td>open the first folder with new message, exit immediately if none<\/td><\/tr><\/tbody><\/table><\/div><br class=\"table-break\" \/><p>\nTo read messages in a mailbox\n<\/p><div class=\"cmdsynopsis\"><p><code class=\"command\">mutt<\/code>  [<code class=\"option\">-nz<\/code>] [<code class=\"option\">-F<\/code>\n<em class=\"replaceable\"><code>muttrc<\/code><\/em>\n] [<code class=\"option\">-m<\/code>\n<em class=\"replaceable\"><code>type<\/code><\/em>\n] [<code class=\"option\">-f<\/code>\n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n]<\/p><\/div><p>\nTo compose a new message\n<\/p><div class=\"cmdsynopsis\"><p><code class=\"command\">mutt<\/code>  [<code class=\"option\">-En<\/code>] [<code class=\"option\">-F<\/code>\n<em class=\"replaceable\"><code>muttrc<\/code><\/em>\n] [<code class=\"option\">-c<\/code>\n<em class=\"replaceable\"><code>address<\/code><\/em>\n] [<code class=\"option\">-Hi<\/code>\n<em class=\"replaceable\"><code>filename<\/code><\/em>\n] [<code class=\"option\">-s<\/code>\n<em class=\"replaceable\"><code>subject<\/code><\/em>\n] [\n<code class=\"option\">-a<\/code>\n<em class=\"replaceable\"><code>file<\/code><\/em>\n [...]\n--\n]   \n<em class=\"replaceable\"><code>address<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>mailto_url<\/code><\/em>\n ... <\/p><\/div><p>\nMutt also supports a \u201cbatch\u201d mode to send prepared\nmessages.  Simply redirect input from the file you wish to send.  For\nexample,\n<\/p><pre class=\"screen\">\nmutt -s \"data set for run #2\" professor@bigschool.edu &lt; ~\/run2.dat<\/pre><p>\nwill send a message to\n<code class=\"literal\">&lt;professor@bigschool.edu&gt;<\/code> with a subject of\n\u201cdata set for run #2\u201d.  In the body of the message will be\nthe contents of the file \u201c~\/run2.dat\u201d.\n<\/p><p>\nAn include file passed with <code class=\"literal\">-i<\/code> will be used as the\nbody of the message.  When combined with <code class=\"literal\">-E<\/code>, the\ninclude file will be directly edited during message composition.  The\nfile will be modified regardless of whether the message is sent or\naborted.\n<\/p><p>\nA draft file passed with <code class=\"literal\">-H<\/code> will be used as the\ninitial header and body for the message.  Multipart messages can be\nused as a draft file.  When combined with <code class=\"literal\">-E<\/code>, the\ndraft file will be updated to the final state of the message after\ncomposition, regardless of whether the message is sent, aborted, or\neven postponed.  Note that if the message is sent encrypted or signed,\nthe draft file will be saved that way too.\n<\/p><p>\nAll files passed with <code class=\"literal\">-a<\/code> <em>file<\/em>\nwill be attached as a MIME part to the message. To attach a single or\nseveral files, use \u201c--\u201d to separate files and recipient\naddresses:\n<\/p><pre class=\"screen\">\nmutt -a image.png -- some@one.org<\/pre><p>\nor\n<\/p><pre class=\"screen\">\nmutt -a *.png -- some@one.org<\/pre><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\nThe <code class=\"literal\">-a<\/code> option must be last in the option list.\n<\/p><\/div><p>\nIn addition to accepting a list of email addresses, Mutt also accepts a URL with\nthe <code class=\"literal\">mailto:<\/code> schema as specified in RFC2368.  This is useful\nwhen configuring a web browser to launch Mutt when clicking on mailto links.\n<\/p><pre class=\"screen\">\nmutt mailto:some@one.org?subject=test&amp;cc=other@one.org<\/pre><\/div><div class=\"sect1\"><div class=\"titlepage\"><div><p><\/p><h2 class=\"title\">2.\u00a0Configuration Commands<\/h2><\/div><\/div><p>\nThe following are the commands understood by Mutt:\n<\/p><div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#account-hook\" title=\"6.\u00a0Managing Multiple Accounts\">account-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n<em class=\"replaceable\"><code>command<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#alias\" title=\"4.\u00a0Defining\/Using Aliases\">alias<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...]  \n<em class=\"replaceable\"><code>key<\/code><\/em>\n   \n<em class=\"replaceable\"><code>address<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>address<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#alias\" title=\"4.\u00a0Defining\/Using Aliases\">unalias<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...] { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>key<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...]  \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">unalternates<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...] { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#alternative-order\" title=\"5.\u00a0MIME Multipart\/Alternative\">alternative_order<\/a><\/code>   \n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#alternative-order\" title=\"5.\u00a0MIME Multipart\/Alternative\">unalternative_order<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#attachments\" title=\"6.\u00a0Attachment Searching and Counting\">attachments<\/a><\/code>   \n<em class=\"replaceable\"><code>{ + | - }disposition<\/code><\/em>\n   \n<em class=\"replaceable\"><code>mime-type<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#attachments\" title=\"6.\u00a0Attachment Searching and Counting\">unattachments<\/a><\/code>   \n<em class=\"replaceable\"><code>{ + | - }disposition<\/code><\/em>\n   \n<em class=\"replaceable\"><code>mime-type<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#auto-view\" title=\"4.\u00a0MIME Autoview\">auto_view<\/a><\/code>   \n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#auto-view\" title=\"4.\u00a0MIME Autoview\">unauto_view<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#bind\" title=\"5.\u00a0Changing the Default Key Bindings\">bind<\/a><\/code>   \n<em class=\"replaceable\"><code>map<\/code><\/em>\n   \n<em class=\"replaceable\"><code>key<\/code><\/em>\n   \n<em class=\"replaceable\"><code>function<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#charset-hook\" title=\"6.\u00a0Defining Aliases for Character Sets\">charset-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>alias<\/code><\/em>\n   \n<em class=\"replaceable\"><code>charset<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#iconv-hook\">iconv-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>charset<\/code><\/em>\n   \n<em class=\"replaceable\"><code>local-charset<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#color\" title=\"9.\u00a0Using Color and Mono Video Attributes\">color<\/a><\/code>   \n<em class=\"replaceable\"><code>object<\/code><\/em>\n   \n<em class=\"replaceable\"><code>foreground<\/code><\/em>\n   \n<em class=\"replaceable\"><code>background<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#color\" title=\"9.\u00a0Using Color and Mono Video Attributes\">color<\/a><\/code>  { \n<code class=\"option\">header<\/code>\n  |   \n<code class=\"option\">body<\/code>\n }  \n<em class=\"replaceable\"><code>foreground<\/code><\/em>\n   \n<em class=\"replaceable\"><code>background<\/code><\/em>\n   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#color\" title=\"9.\u00a0Using Color and Mono Video Attributes\">color<\/a><\/code>   \n<code class=\"option\">index<\/code>\n   \n<em class=\"replaceable\"><code>foreground<\/code><\/em>\n   \n<em class=\"replaceable\"><code>background<\/code><\/em>\n   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#color\" title=\"9.\u00a0Using Color and Mono Video Attributes\">color<\/a><\/code>   \n<code class=\"option\">compose<\/code>\n   \n<em class=\"replaceable\"><code>composeobject<\/code><\/em>\n   \n<em class=\"replaceable\"><code>foreground<\/code><\/em>\n   \n<em class=\"replaceable\"><code>background<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#color\" title=\"9.\u00a0Using Color and Mono Video Attributes\">uncolor<\/a><\/code>  { \n<code class=\"option\">index<\/code>\n  |   \n<code class=\"option\">header<\/code>\n  |   \n<code class=\"option\">body<\/code>\n } { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#crypt-hook\" title=\"21.\u00a0Choosing the Cryptographic Key of the Recipient\">crypt-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n   \n<em class=\"replaceable\"><code>keyid<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#echo\" title=\"17.\u00a0Echoing Text\">echo<\/a><\/code>   \n<em class=\"replaceable\"><code>message<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#exec\" title=\"23.\u00a0Executing Functions\">exec<\/a><\/code>   \n<em class=\"replaceable\"><code>function<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>function<\/code><\/em>\n...]<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#fcc-hook\" title=\"17.\u00a0Specify Default Fcc: Mailbox When Composing\">fcc-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#fcc-save-hook\" title=\"18.\u00a0Specify Default Save Filename and Default Fcc: Mailbox at Once\">fcc-save-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#folder-hook\" title=\"7.\u00a0Setting Variables Based Upon Mailbox\">folder-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]regexp<\/code><\/em>\n   \n<em class=\"replaceable\"><code>command<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#addrgroup\" title=\"3.\u00a0Address Groups\">group<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...] { \n<code class=\"option\">-rx<\/code>\n<em class=\"replaceable\"><code>expr<\/code><\/em>\n...  |   \n<code class=\"option\">-addr<\/code>\n<em class=\"replaceable\"><code>expr<\/code><\/em>\n... }<br \/><code class=\"command\"><a class=\"link\" href=\"#addrgroup\" title=\"3.\u00a0Address Groups\">ungroup<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...] { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<code class=\"option\">-rx<\/code>\n<em class=\"replaceable\"><code>expr<\/code><\/em>\n...  |   \n<code class=\"option\">-addr<\/code>\n<em class=\"replaceable\"><code>expr<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#hdr-order\" title=\"10.3.\u00a0Ordering Displayed Headers\">hdr_order<\/a><\/code>   \n<em class=\"replaceable\"><code>header<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>header<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#hdr-order\" title=\"10.3.\u00a0Ordering Displayed Headers\">unhdr_order<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>header<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#ignore\" title=\"10.2.\u00a0Selecting Headers\">ignore<\/a><\/code>   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#ignore\" title=\"10.2.\u00a0Selecting Headers\">unignore<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#lists\" title=\"12.\u00a0Mailing Lists\">lists<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n]  \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#lists\" title=\"12.\u00a0Mailing Lists\">unlists<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...] { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#macro\" title=\"8.\u00a0Keyboard Macros\">macro<\/a><\/code>   \n<em class=\"replaceable\"><code>menu<\/code><\/em>\n   \n<em class=\"replaceable\"><code>key<\/code><\/em>\n   \n<em class=\"replaceable\"><code>sequence<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>description<\/code><\/em>\n]<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#mailboxes\" title=\"14.\u00a0Monitoring Incoming Mail\">mailboxes<\/a><\/code>   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#mailboxes\" title=\"14.\u00a0Monitoring Incoming Mail\">unmailboxes<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#mailto-allow\" title=\"30.\u00a0Control allowed header fields in a mailto: URL\">mailto_allow<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>header-field<\/code><\/em>\n... }<br \/><code class=\"command\"><a class=\"link\" href=\"#mailto-allow\" title=\"30.\u00a0Control allowed header fields in a mailto: URL\">unmailto_allow<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>header-field<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#mbox-hook\" title=\"13.\u00a0Using Multiple Spool Mailboxes\">mbox-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]regexp<\/code><\/em>\n   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#message-hook\" title=\"20.\u00a0Change Settings Before Formatting a Message\">message-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>command<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#mime-lookup\" title=\"7.\u00a0MIME Lookup\">mime_lookup<\/a><\/code>   \n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#mime-lookup\" title=\"7.\u00a0MIME Lookup\">unmime_lookup<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>mimetype<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#mono\">mono<\/a><\/code>   \n<em class=\"replaceable\"><code>object<\/code><\/em>\n   \n<em class=\"replaceable\"><code>attribute<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#mono\">mono<\/a><\/code>  { \n<code class=\"option\">header<\/code>\n  |   \n<code class=\"option\">body<\/code>\n }  \n<em class=\"replaceable\"><code>attribute<\/code><\/em>\n   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#mono\">mono<\/a><\/code>   \n<code class=\"option\">index<\/code>\n   \n<em class=\"replaceable\"><code>attribute<\/code><\/em>\n   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#mono\">mono<\/a><\/code>   \n<code class=\"option\">compose<\/code>\n   \n<em class=\"replaceable\"><code>composeobject<\/code><\/em>\n   \n<em class=\"replaceable\"><code>attribute<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#mono\">unmono<\/a><\/code>  { \n<code class=\"option\">index<\/code>\n  |   \n<code class=\"option\">header<\/code>\n  |   \n<code class=\"option\">body<\/code>\n } { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#my-hdr\" title=\"15.\u00a0User-Defined Headers\">my_hdr<\/a><\/code>   \n<em class=\"replaceable\"><code>string<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#my-hdr\" title=\"15.\u00a0User-Defined Headers\">unmy_hdr<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>field<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#push\" title=\"22.\u00a0Adding Key Sequences to the Keyboard Buffer\">push<\/a><\/code>   \n<em class=\"replaceable\"><code>string<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#save-hook\" title=\"16.\u00a0Specify Default Save Mailbox\">save-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#score-command\" title=\"24.\u00a0Message Scoring\">score<\/a><\/code>   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>value<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#score-command\" title=\"24.\u00a0Message Scoring\">unscore<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#reply-hook\">reply-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>command<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#send-hook\" title=\"19.\u00a0Change Settings Based Upon Message Recipients\">send-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>command<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#send2-hook\">send2-hook<\/a><\/code>   \n<em class=\"replaceable\"><code>[!]pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>command<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#set\" title=\"26.\u00a0Setting and Querying Variables\">set<\/a><\/code>  { \n[ <code class=\"option\">no<\/code>  |   <code class=\"option\">inv<\/code> ]\n<em class=\"replaceable\"><code>variable<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>variable=value<\/code><\/em>\n } [...]<br \/><code class=\"command\"><a class=\"link\" href=\"#set\" title=\"26.\u00a0Setting and Querying Variables\">toggle<\/a><\/code>   \n<em class=\"replaceable\"><code>variable<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>variable<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#set\" title=\"26.\u00a0Setting and Querying Variables\">unset<\/a><\/code>   \n<em class=\"replaceable\"><code>variable<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>variable<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#set\" title=\"26.\u00a0Setting and Querying Variables\">reset<\/a><\/code>   \n<em class=\"replaceable\"><code>variable<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>variable<\/code><\/em>\n...]<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#setenv\" title=\"7.\u00a0Managing the Environment\">setenv<\/a><\/code>   \n      <em class=\"replaceable\"><code>[?]variable<\/code><\/em>\n      [\n      <em class=\"replaceable\"><code>value<\/code><\/em>\n    ]<br \/><code class=\"command\"><a class=\"link\" href=\"#setenv\" title=\"7.\u00a0Managing the Environment\">unsetenv<\/a><\/code>   \n      <em class=\"replaceable\"><code>variable<\/code><\/em>\n     <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#sidebar-whitelist\">sidebar_whitelist<\/a><\/code>   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#unsidebar-whitelist\">unsidebar_whitelist<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>mailbox<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#source\" title=\"27.\u00a0Reading Initialization Commands From Another File\">source<\/a><\/code>   \n<em class=\"replaceable\"><code>filename<\/code><\/em>\n <\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#spam\" title=\"25.\u00a0Spam Detection\">spam<\/a><\/code>   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>format<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#spam\" title=\"25.\u00a0Spam Detection\">nospam<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#display-munging\" title=\"12.\u00a0Display Munging\">subjectrx<\/a><\/code>   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n   \n<em class=\"replaceable\"><code>replacement<\/code><\/em>\n <br \/><code class=\"command\"><a class=\"link\" href=\"#display-munging\" title=\"12.\u00a0Display Munging\">unsubjectrx<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>pattern<\/code><\/em>\n }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#subscribe\">subscribe<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...]  \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n  [\n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n...]<br \/><code class=\"command\"><a class=\"link\" href=\"#subscribe\">unsubscribe<\/a><\/code>  [\n<code class=\"option\">-group<\/code>\n<em class=\"replaceable\"><code>name<\/code><\/em>\n...] { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>regexp<\/code><\/em>\n... }<\/p><\/div><\/li><li class=\"listitem\"><div class=\"cmdsynopsis\"><p><code class=\"command\"><a class=\"link\" href=\"#unhook\" title=\"28.\u00a0Removing Hooks\">unhook<\/a><\/code>  { \n<em class=\"replaceable\"><code>*<\/code><\/em>\n  |   \n<em class=\"replaceable\"><code>hook-type<\/code><\/em>\n }<\/p><\/div><\/li><\/ul><\/div><\/div><div class=\"sect1\"><div class=\"titlepage\"><div><p><\/p><h2 class=\"title\">3.\u00a0Configuration Variables<\/h2><\/div><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.1.\u00a0abort_noattach<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen the body of the message matches <a class=\"link\" href=\"#abort-noattach-regexp\" title=\"3.2.\u00a0abort_noattach_regexp\">$abort_noattach_regexp<\/a> and\nthere are no attachments, this quadoption controls whether to\nabort sending the message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.2.\u00a0abort_noattach_regexp<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0\u201c<code class=\"literal\">attach<\/code>\u201d<\/p><\/div><p>\nSpecifies a regular expression to match against the body of the\nmessage, to determine if an attachment was mentioned but\nmistakenly forgotten.  If it matches, <a class=\"link\" href=\"#abort-noattach\" title=\"3.1.\u00a0abort_noattach\">$abort_noattach<\/a> will be\nconsulted to determine if message sending will be aborted.\n<\/p><p>\nLike other regular expressions in Mutt, the search is case\nsensitive if the pattern contains at least one upper case letter,\nand case insensitive otherwise.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.3.\u00a0abort_nosubject<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nIf set to <em>yes<\/em>, when composing messages and no subject is given\nat the subject prompt, composition will be aborted.  If set to\n<em>no<\/em>, composing messages with no subject given at the subject\nprompt will never be aborted.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.4.\u00a0abort_unmodified<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf set to <em>yes<\/em>, composition will automatically abort after\nediting the message body if no changes are made to the file (this\ncheck only happens after the <em>first<\/em> edit of the file).  When set\nto <em>no<\/em>, composition will never be aborted.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.5.\u00a0alias_file<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/.muttrc<\/code>\u201d<\/p><\/div><p>\nThe default file in which to save aliases created by the\n<code class=\"literal\"><a class=\"link\" href=\"#create-alias\">&lt;create-alias&gt;<\/a><\/code> function. Entries added to this file are\nencoded in the character set specified by <a class=\"link\" href=\"#config-charset\" title=\"3.35.\u00a0config_charset\">$config_charset<\/a> if it\nis <em>set<\/em> or the current character set otherwise.\n<\/p><p>\n<strong>Note:<\/strong> Mutt will not automatically source this file; you must\nexplicitly use the \u201c<a class=\"link\" href=\"#source\" title=\"27.\u00a0Reading Initialization Commands From Another File\">source<\/a>\u201d command for it to be executed in case\nthis option points to a dedicated alias file.\n<\/p><p>\nThe default for this option is the currently used muttrc file, or\n\u201c~\/.muttrc\u201d if no user muttrc was found.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.6.\u00a0alias_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%4n\u00a0%2f\u00a0%t\u00a0%-10a\u00a0\u00a0\u00a0%r<\/code>\u201d<\/p><\/div><p>\nSpecifies the format of the data displayed for the \u201c<a class=\"link\" href=\"#alias\" title=\"4.\u00a0Defining\/Using Aliases\">alias<\/a>\u201d menu.  The\nfollowing <code class=\"literal\">printf(3)<\/code>-style sequences are available:\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.7.\u00a0allow_8bit<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls whether 8-bit data is converted to 7-bit using either Quoted-\nPrintable or Base64 encoding when sending mail.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.8.\u00a0allow_ansi<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nControls whether ANSI color codes in messages (and color tags in\nrich text messages) are to be interpreted.\nMessages containing these codes are rare, but if this option is <em>set<\/em>,\ntheir text will be colored accordingly. Note that this may override\nyour color choices, and even present a security problem, since a\nmessage could include a line like\n<\/p><pre class=\"screen\">\n[-- PGP output follows ...\n<\/pre><p>\nand give it the same color as your attachment color (see also\n<a class=\"link\" href=\"#crypt-timestamp\" title=\"3.50.\u00a0crypt_timestamp\">$crypt_timestamp<\/a>).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.9.\u00a0arrow_cursor<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, an arrow (\u201c-&gt;\u201d) will be used to indicate the current entry\nin menus instead of highlighting the whole line.  On slow network or modem\nlinks this will make response faster because there is less that has to\nbe redrawn on the screen when moving to the next or previous entries\nin the menu.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.10.\u00a0ascii_chars<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will use plain ASCII characters when displaying thread\nand attachment trees, instead of the default <em>ACS<\/em> characters.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.11.\u00a0askbcc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will prompt you for blind-carbon-copy (Bcc) recipients\nbefore editing an outgoing message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.12.\u00a0askcc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will prompt you for carbon-copy (Cc) recipients before\nediting the body of an outgoing message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.13.\u00a0assumed_charset<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable is a colon-separated list of character encoding\nschemes for messages without character encoding indication.\nHeader field values and message body content without character encoding\nindication would be assumed that they are written in one of this list.\nBy default, all the header fields and message body without any charset\nindication are assumed to be in \u201cus-ascii\u201d.\n<\/p><p>\nFor example, Japanese users might prefer this:\n<\/p><pre class=\"screen\">\nset assumed_charset=\"iso-2022-jp:euc-jp:shift_jis:utf-8\"\n<\/pre><p>\nHowever, only the first content is valid for the message body.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.14.\u00a0attach_charset<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable is a colon-separated list of character encoding\nschemes for text file attachments. Mutt uses this setting to guess\nwhich encoding files being attached are encoded in to convert them to\na proper character set given in <a class=\"link\" href=\"#send-charset\" title=\"3.263.\u00a0send_charset\">$send_charset<\/a>.\n<\/p><p>\nIf <em>unset<\/em>, the value of <a class=\"link\" href=\"#charset\" title=\"3.30.\u00a0charset\">$charset<\/a> will be used instead.\nFor example, the following configuration would work for Japanese\ntext handling:\n<\/p><pre class=\"screen\">\nset attach_charset=\"iso-2022-jp:euc-jp:shift_jis:utf-8\"\n<\/pre><p>\nNote: for Japanese users, \u201ciso-2022-*\u201d must be put at the head\nof the value as shown above if included.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.15.\u00a0attach_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%u%D%I\u00a0%t%4n\u00a0%T%.40d%&gt;\u00a0[%.7m\/%.10M,\u00a0%.6e%?C?,\u00a0%C?,\u00a0%s]\u00a0<\/code>\u201d<\/p><\/div><p>\nThis variable describes the format of the \u201cattachment\u201d menu.  The\nfollowing <code class=\"literal\">printf(3)<\/code>-style sequences are understood:\n<\/p><p>\nFor an explanation of \u201csoft-fill\u201d, see the <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a> documentation.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.16.\u00a0attach_sep<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">\\n<\/code>\u201d<\/p><\/div><p>\nThe separator to add between attachments when operating (saving,\nprinting, piping, etc) on a list of tagged attachments.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.17.\u00a0attach_split<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf this variable is <em>unset<\/em>, when operating (saving, printing, piping,\netc) on a list of tagged attachments, Mutt will concatenate the\nattachments and will operate on them as a single attachment. The\n<a class=\"link\" href=\"#attach-sep\" title=\"3.16.\u00a0attach_sep\">$attach_sep<\/a> separator is added after each attachment. When <em>set<\/em>,\nMutt will operate on the attachments one by one.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.18.\u00a0attribution<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">On\u00a0%d,\u00a0%n\u00a0wrote:<\/code>\u201d<\/p><\/div><p>\nThis is the string that will precede a message which has been included\nin a reply.  For a full listing of defined <code class=\"literal\">printf(3)<\/code>-like sequences see\nthe section on <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.19.\u00a0attribution_locale<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThe locale used by <code class=\"literal\">strftime(3)<\/code> to format dates in the\n<a class=\"link\" href=\"#attribution\" title=\"3.18.\u00a0attribution\">attribution<\/a> string.  Legal values are the strings your system\naccepts for the locale environment variable <code class=\"literal\">$LC_TIME<\/code>.\n<\/p><p>\nThis variable is to allow the attribution date format to be\ncustomized by recipient or folder using hooks.  By default, Mutt\nwill use your locale environment, so there is no need to set\nthis except to override that default.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.20.\u00a0auto_tag<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, functions in the <em>index<\/em> menu which affect a message\nwill be applied to all tagged messages (if there are any).  When\nunset, you must first use the <code class=\"literal\">&lt;tag-prefix&gt;<\/code> function (bound to \u201c;\u201d\nby default) to make the next function apply to all tagged messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.21.\u00a0autoedit<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em> along with <a class=\"link\" href=\"#edit-headers\" title=\"3.64.\u00a0edit_headers\">$edit_headers<\/a>, Mutt will skip the initial\nsend-menu (prompting for subject and recipients) and allow you to\nimmediately begin editing the body of your\nmessage.  The send-menu may still be accessed once you have finished\nediting the body of your message.\n<\/p><p>\n<strong>Note:<\/strong> when this option is <em>set<\/em>, you cannot use send-hooks that depend\non the recipients when composing a new (non-reply) message, as the initial\nlist of recipients is empty.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#fast-reply\" title=\"3.71.\u00a0fast_reply\">$fast_reply<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.22.\u00a0beep<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, mutt will beep when an error occurs.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.23.\u00a0beep_new<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, mutt will beep whenever it prints a message\nnotifying you of new mail.  This is independent of the setting of the\n<a class=\"link\" href=\"#beep\" title=\"3.22.\u00a0beep\">$beep<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.24.\u00a0bounce<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nControls whether you will be asked to confirm bouncing messages.\nIf set to <em>yes<\/em> you don't get asked if you want to bounce a\nmessage. Setting this variable to <em>no<\/em> is not generally useful,\nand thus not recommended, because you are unable to bounce messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.25.\u00a0bounce_delivered<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, mutt will include Delivered-To headers when\nbouncing messages.  Postfix users may wish to <em>unset<\/em> this variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.26.\u00a0braille_friendly<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, mutt will place the cursor at the beginning\nof the current line in menus, even when the <a class=\"link\" href=\"#arrow-cursor\" title=\"3.9.\u00a0arrow_cursor\">$arrow_cursor<\/a> variable\nis <em>unset<\/em>, making it easier for blind persons using Braille displays to\nfollow these menus.  The option is <em>unset<\/em> by default because many\nvisual terminals don't permit making the cursor invisible.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.27.\u00a0browser_abbreviate_mailboxes<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, mutt will abbreviate mailbox\nnames in the browser mailbox list, using '~' and '='\nshortcuts.\n<\/p><p>\nThe default <code class=\"literal\">\"alpha\"<\/code> setting of <a class=\"link\" href=\"#sort-browser\" title=\"3.314.\u00a0sort_browser\">$sort_browser<\/a> uses\nlocale-based sorting (using <code class=\"literal\">strcoll(3)<\/code>), which ignores some\npunctuation.  This can lead to some situations where the order\ndoesn't make intuitive sense.  In those cases, it may be\ndesirable to <em>unset<\/em> this variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.28.\u00a0certificate_file<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/.mutt_certificates<\/code>\u201d<\/p><\/div><p>\nThis variable specifies the file where the certificates you trust\nare saved. When an unknown certificate is encountered, you are asked\nif you accept it or not. If you accept it, the certificate can also\nbe saved in this file and further connections are automatically\naccepted.\n<\/p><p>\nYou can also manually add CA certificates in this file. Any server\ncertificate that is signed with one of these CA certificates is\nalso automatically accepted.\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset certificate_file=~\/.mutt\/certificates\n<\/pre><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.29.\u00a0change_folder_next<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, the <code class=\"literal\">&lt;change-folder&gt;<\/code> function\nmailbox suggestion will start at the next folder in your \u201c<a class=\"link\" href=\"#mailboxes\" title=\"14.\u00a0Monitoring Incoming Mail\">mailboxes<\/a>\u201d\nlist, instead of starting at the first folder in the list.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.30.\u00a0charset<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nCharacter set your terminal uses to display and enter textual data.\nIt is also the fallback for <a class=\"link\" href=\"#send-charset\" title=\"3.263.\u00a0send_charset\">$send_charset<\/a>.\n<\/p><p>\nUpon startup Mutt tries to derive this value from environment variables\nsuch as <code class=\"literal\">$LC_CTYPE<\/code> or <code class=\"literal\">$LANG<\/code>.\n<\/p><p>\n<strong>Note:<\/strong> It should only be set in case Mutt isn't able to determine the\ncharacter set used correctly.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.31.\u00a0check_mbox_size<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, mutt will use file size attribute instead of\naccess time when checking for new mail in mbox and mmdf folders.\n<\/p><p>\nThis variable is <em>unset<\/em> by default and should only be enabled when\nnew mail detection for these folder types is unreliable or doesn't work.\n<\/p><p>\nNote that enabling this variable should happen before any \u201c<a class=\"link\" href=\"#mailboxes\" title=\"14.\u00a0Monitoring Incoming Mail\">mailboxes<\/a>\u201d\ndirectives occur in configuration files regarding mbox or mmdf folders\nbecause mutt needs to determine the initial new mail status of such a\nmailbox by performing a fast mailbox scan when it is defined.\nAfterwards the new mail status is tracked by file size changes.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.32.\u00a0check_new<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\n<strong>Note:<\/strong> this option only affects <em>maildir<\/em> and <em>MH<\/em> style\nmailboxes.\n<\/p><p>\nWhen <em>set<\/em>, Mutt will check for new mail delivered while the\nmailbox is open.  Especially with MH mailboxes, this operation can\ntake quite some time since it involves scanning the directory and\nchecking each file to see if it has already been looked at.  If\nthis variable is <em>unset<\/em>, no check for new mail is performed\nwhile the mailbox is open.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.33.\u00a0collapse_unread<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>unset<\/em>, Mutt will not collapse a thread if it contains any\nunread messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.34.\u00a0compose_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">--\u00a0Mutt:\u00a0Compose\u00a0\u00a0[Approx.\u00a0msg\u00a0size:\u00a0%l\u00a0\u00a0\u00a0Atts:\u00a0%a]%&gt;-<\/code>\u201d<\/p><\/div><p>\nControls the format of the status line displayed in the \u201ccompose\u201d\nmenu.  This string is similar to <a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a>, but has its own\nset of <code class=\"literal\">printf(3)<\/code>-like sequences:\n<\/p><p>\nSee the text describing the <a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a> option for more\ninformation on how to set <a class=\"link\" href=\"#compose-format\" title=\"3.34.\u00a0compose_format\">$compose_format<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.35.\u00a0config_charset<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nWhen defined, Mutt will recode commands in rc files from this\nencoding to the current character set as specified by <a class=\"link\" href=\"#charset\" title=\"3.30.\u00a0charset\">$charset<\/a>\nand aliases written to <a class=\"link\" href=\"#alias-file\" title=\"3.5.\u00a0alias_file\">$alias_file<\/a> from the current character set.\n<\/p><p>\nPlease note that if setting <a class=\"link\" href=\"#charset\" title=\"3.30.\u00a0charset\">$charset<\/a> it must be done before\nsetting <a class=\"link\" href=\"#config-charset\" title=\"3.35.\u00a0config_charset\">$config_charset<\/a>.\n<\/p><p>\nRecoding should be avoided as it may render unconvertable\ncharacters as question marks which can lead to undesired\nside effects (for example in regular expressions).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.36.\u00a0confirmappend<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will prompt for confirmation when appending messages to\nan existing mailbox.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.37.\u00a0confirmcreate<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will prompt for confirmation when saving messages to a\nmailbox which does not yet exist before creating it.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.38.\u00a0connect_timeout<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a030<\/p><\/div><p>\nCauses Mutt to timeout a network connection (for IMAP, POP or SMTP) after this\nmany seconds if the connection is not able to be established.  A negative\nvalue causes Mutt to wait indefinitely for the connection attempt to succeed.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.39.\u00a0content_type<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">text\/plain<\/code>\u201d<\/p><\/div><p>\nSets the default Content-Type for the body of newly composed messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.40.\u00a0copy<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether or not copies of your outgoing messages\nwill be saved for later references.  Also see <a class=\"link\" href=\"#record\" title=\"3.240.\u00a0record\">$record<\/a>,\n<a class=\"link\" href=\"#save-name\" title=\"3.257.\u00a0save_name\">$save_name<\/a>, <a class=\"link\" href=\"#force-name\" title=\"3.78.\u00a0force_name\">$force_name<\/a> and \u201c<a class=\"link\" href=\"#fcc-hook\" title=\"17.\u00a0Specify Default Fcc: Mailbox When Composing\">fcc-hook<\/a>\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.41.\u00a0crypt_autoencrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nSetting this variable will cause Mutt to always attempt to PGP\nencrypt outgoing messages.  This is probably only useful in\nconnection to the \u201c<a class=\"link\" href=\"#send-hook\" title=\"19.\u00a0Change Settings Based Upon Message Recipients\">send-hook<\/a>\u201d command.  It can be overridden\nby use of the pgp menu, when encryption is not required or\nsigning is requested as well.  If <a class=\"link\" href=\"#smime-is-default\" title=\"3.297.\u00a0smime_is_default\">$smime_is_default<\/a> is <em>set<\/em>,\nthen OpenSSL is used instead to create S\/MIME messages and\nsettings can be overridden by use of the smime menu instead.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.42.\u00a0crypt_autopgp<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether or not mutt may automatically enable\nPGP encryption\/signing for messages.  See also <a class=\"link\" href=\"#crypt-autoencrypt\" title=\"3.41.\u00a0crypt_autoencrypt\">$crypt_autoencrypt<\/a>,\n<a class=\"link\" href=\"#crypt-replyencrypt\" title=\"3.47.\u00a0crypt_replyencrypt\">$crypt_replyencrypt<\/a>,\n<a class=\"link\" href=\"#crypt-autosign\" title=\"3.43.\u00a0crypt_autosign\">$crypt_autosign<\/a>, <a class=\"link\" href=\"#crypt-replysign\" title=\"3.48.\u00a0crypt_replysign\">$crypt_replysign<\/a> and <a class=\"link\" href=\"#smime-is-default\" title=\"3.297.\u00a0smime_is_default\">$smime_is_default<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.43.\u00a0crypt_autosign<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nSetting this variable will cause Mutt to always attempt to\ncryptographically sign outgoing messages.  This can be overridden\nby use of the pgp menu, when signing is not required or\nencryption is requested as well. If <a class=\"link\" href=\"#smime-is-default\" title=\"3.297.\u00a0smime_is_default\">$smime_is_default<\/a> is <em>set<\/em>,\nthen OpenSSL is used instead to create S\/MIME messages and settings can\nbe overridden by use of the smime menu instead of the pgp menu.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.44.\u00a0crypt_autosmime<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether or not mutt may automatically enable\nS\/MIME encryption\/signing for messages. See also <a class=\"link\" href=\"#crypt-autoencrypt\" title=\"3.41.\u00a0crypt_autoencrypt\">$crypt_autoencrypt<\/a>,\n<a class=\"link\" href=\"#crypt-replyencrypt\" title=\"3.47.\u00a0crypt_replyencrypt\">$crypt_replyencrypt<\/a>,\n<a class=\"link\" href=\"#crypt-autosign\" title=\"3.43.\u00a0crypt_autosign\">$crypt_autosign<\/a>, <a class=\"link\" href=\"#crypt-replysign\" title=\"3.48.\u00a0crypt_replysign\">$crypt_replysign<\/a> and <a class=\"link\" href=\"#smime-is-default\" title=\"3.297.\u00a0smime_is_default\">$smime_is_default<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.45.\u00a0crypt_confirmhook<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf set, then you will be prompted for confirmation of keys when using\nthe <em>crypt-hook<\/em> command.  If unset, no such confirmation prompt will\nbe presented.  This is generally considered unsafe, especially where\ntypos are concerned.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.46.\u00a0crypt_opportunistic_encrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nSetting this variable will cause Mutt to automatically enable and\ndisable encryption, based on whether all message recipient keys\ncan be located by Mutt.\n<\/p><p>\nWhen this option is enabled, Mutt will enable\/disable encryption\neach time the TO, CC, and BCC lists are edited.  If\n<a class=\"link\" href=\"#edit-headers\" title=\"3.64.\u00a0edit_headers\">$edit_headers<\/a> is set, Mutt will also do so each time the message\nis edited.\n<\/p><p>\nWhile this is set, encryption can't be manually enabled\/disabled.\nThe pgp or smime menus provide a selection to temporarily disable\nthis option for the current message.\n<\/p><p>\nIf <a class=\"link\" href=\"#crypt-autoencrypt\" title=\"3.41.\u00a0crypt_autoencrypt\">$crypt_autoencrypt<\/a> or <a class=\"link\" href=\"#crypt-replyencrypt\" title=\"3.47.\u00a0crypt_replyencrypt\">$crypt_replyencrypt<\/a> enable encryption for\na message, this option will be disabled for that message.  It can\nbe manually re-enabled in the pgp or smime menus.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.47.\u00a0crypt_replyencrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, automatically PGP or OpenSSL encrypt replies to messages which are\nencrypted.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.48.\u00a0crypt_replysign<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, automatically PGP or OpenSSL sign replies to messages which are\nsigned.\n<\/p><p>\n<strong>Note:<\/strong> this does not work on messages that are encrypted\n<em>and<\/em> signed!\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.49.\u00a0crypt_replysignencrypted<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, automatically PGP or OpenSSL sign replies to messages\nwhich are encrypted. This makes sense in combination with\n<a class=\"link\" href=\"#crypt-replyencrypt\" title=\"3.47.\u00a0crypt_replyencrypt\">$crypt_replyencrypt<\/a>, because it allows you to sign all\nmessages which are automatically encrypted.  This works around\nthe problem noted in <a class=\"link\" href=\"#crypt-replysign\" title=\"3.48.\u00a0crypt_replysign\">$crypt_replysign<\/a>, that mutt is not able\nto find out whether an encrypted message is also signed.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.50.\u00a0crypt_timestamp<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, mutt will include a time stamp in the lines surrounding\nPGP or S\/MIME output, so spoofing such lines is more difficult.\nIf you are using colors to mark these lines, and rely on these,\nyou may <em>unset<\/em> this setting.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.51.\u00a0crypt_use_gpgme<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable controls the use of the GPGME-enabled crypto backends.\nIf it is <em>set<\/em> and Mutt was built with gpgme support, the gpgme code for\nS\/MIME and PGP will be used instead of the classic code.  Note that\nyou need to set this option in .muttrc; it won't have any effect when\nused interactively.\n<\/p><p>\nNote that the GPGME backend does not support creating old-style inline\n(traditional) PGP encrypted or signed messages (see <a class=\"link\" href=\"#pgp-autoinline\" title=\"3.177.\u00a0pgp_autoinline\">$pgp_autoinline<\/a>).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.52.\u00a0crypt_use_pka<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nControls whether mutt uses PKA\n(see http:\/\/www.g10code.de\/docs\/pka-intro.de.pdf) during signature\nverification (only supported by the GPGME backend).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.53.\u00a0crypt_verify_sig<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>\u201cyes\u201d<\/em>, always attempt to verify PGP or S\/MIME signatures.\nIf <em>\u201cask-*\u201d<\/em>, ask whether or not to verify the signature.\nIf <em>\u201cno\u201d<\/em>, never attempt to verify cryptographic signatures.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.54.\u00a0date_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">!%a,\u00a0%b\u00a0%d,\u00a0%Y\u00a0at\u00a0%I:%M:%S%p\u00a0%Z<\/code>\u201d<\/p><\/div><p>\nThis variable controls the format of the date printed by the \u201c%d\u201d\nsequence in <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>.  This is passed to the <code class=\"literal\">strftime(3)<\/code>\nfunction to process the date, see the man page for the proper syntax.\n<\/p><p>\nUnless the first character in the string is a bang (\u201c!\u201d), the month\nand week day names are expanded according to the locale.\nIf the first character in the string is a\nbang, the bang is discarded, and the month and week day names in the\nrest of the string are expanded in the <em>C<\/em> locale (that is in US\nEnglish).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.55.\u00a0default_hook<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">~f\u00a0%s\u00a0!~P\u00a0|\u00a0(~P\u00a0~C\u00a0%s)<\/code>\u201d<\/p><\/div><p>\nThis variable controls how \u201c<a class=\"link\" href=\"#message-hook\" title=\"20.\u00a0Change Settings Before Formatting a Message\">message-hook<\/a>\u201d, \u201c<a class=\"link\" href=\"#reply-hook\">reply-hook<\/a>\u201d, \u201c<a class=\"link\" href=\"#send-hook\" title=\"19.\u00a0Change Settings Based Upon Message Recipients\">send-hook<\/a>\u201d,\n\u201c<a class=\"link\" href=\"#send2-hook\">send2-hook<\/a>\u201d, \u201c<a class=\"link\" href=\"#save-hook\" title=\"16.\u00a0Specify Default Save Mailbox\">save-hook<\/a>\u201d, and \u201c<a class=\"link\" href=\"#fcc-hook\" title=\"17.\u00a0Specify Default Fcc: Mailbox When Composing\">fcc-hook<\/a>\u201d will\nbe interpreted if they are specified with only a simple regexp,\ninstead of a matching pattern.  The hooks are expanded when they are\ndeclared, so a hook will be interpreted according to the value of this\nvariable at the time the hook is declared.\n<\/p><p>\nThe default value matches\nif the message is either from a user matching the regular expression\ngiven, or if it is from you (if the from address matches\n\u201c<a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a>\u201d) and is to or cc'ed to a user matching the given\nregular expression.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.56.\u00a0delete<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nControls whether or not messages are really deleted when closing or\nsynchronizing a mailbox.  If set to <em>yes<\/em>, messages marked for\ndeleting will automatically be purged without prompting.  If set to\n<em>no<\/em>, messages marked for deletion will be kept in the mailbox.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.57.\u00a0delete_untag<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf this option is <em>set<\/em>, mutt will untag messages when marking them\nfor deletion.  This applies when you either explicitly delete a message,\nor when you save it to another folder.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.58.\u00a0digest_collapse<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf this option is <em>set<\/em>, mutt's received-attachments menu will not show the subparts of\nindividual messages in a multipart\/digest.  To see these subparts, press \u201cv\u201d on that menu.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.59.\u00a0display_filter<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nWhen set, specifies a command used to filter messages.  When a message\nis viewed it is passed as standard input to <a class=\"link\" href=\"#display-filter\" title=\"3.59.\u00a0display_filter\">$display_filter<\/a>, and the\nfiltered message is read from the standard output.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.60.\u00a0dotlock_program<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">\/usr\/local\/bin\/mutt_dotlock<\/code>\u201d<\/p><\/div><p>\nContains the path of the <code class=\"literal\">mutt_dotlock(8)<\/code> binary to be used by\nmutt.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.61.\u00a0dsn_notify<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable sets the request for when notification is returned.  The\nstring consists of a comma separated list (no spaces!) of one or more\nof the following: <em>never<\/em>, to never request notification,\n<em>failure<\/em>, to request notification on transmission failure,\n<em>delay<\/em>, to be notified of message delays, <em>success<\/em>, to be\nnotified of successful transmission.\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset dsn_notify=\"failure,delay\"\n<\/pre><p>\n<strong>Note:<\/strong> when using <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a> for delivery, you should not enable\nthis unless you are either using Sendmail 8.8.x or greater or a MTA\nproviding a <code class=\"literal\">sendmail(1)<\/code>-compatible interface supporting the <code class=\"literal\">-N<\/code> option\nfor DSN. For SMTP delivery, DSN support is auto-detected so that it\ndepends on the server whether DSN will be used or not.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.62.\u00a0dsn_return<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable controls how much of your message is returned in DSN\nmessages.  It may be set to either <em>hdrs<\/em> to return just the\nmessage header, or <em>full<\/em> to return the full message.\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset dsn_return=hdrs\n<\/pre><p>\n<strong>Note:<\/strong> when using <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a> for delivery, you should not enable\nthis unless you are either using Sendmail 8.8.x or greater or a MTA\nproviding a <code class=\"literal\">sendmail(1)<\/code>-compatible interface supporting the <code class=\"literal\">-R<\/code> option\nfor DSN. For SMTP delivery, DSN support is auto-detected so that it\ndepends on the server whether DSN will be used or not.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.63.\u00a0duplicate_threads<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether mutt, when <a class=\"link\" href=\"#sort\" title=\"3.311.\u00a0sort\">$sort<\/a> is set to <em>threads<\/em>, threads\nmessages with the same Message-Id together.  If it is <em>set<\/em>, it will indicate\nthat it thinks they are duplicates of each other with an equals sign\nin the thread tree.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.64.\u00a0edit_headers<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis option allows you to edit the header of your outgoing messages\nalong with the body of your message.\n<\/p><p>\nAlthough the compose menu may have localized header labels, the\nlabels passed to your editor will be standard RFC 2822 headers,\n(e.g. To:, Cc:, Subject:).  Headers added in your editor must\nalso be RFC 2822 headers, or one of the pseudo headers listed in\n\u201c<a class=\"link\" href=\"#edit-header\" title=\"6.2.\u00a0Editing the Message Header\">edit-header<\/a>\u201d.  Mutt will not understand localized header\nlabels, just as it would not when parsing an actual email.\n<\/p><p>\n<strong>Note<\/strong> that changes made to the References: and Date: headers are\nignored for interoperability reasons.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.65.\u00a0editor<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable specifies which editor is used by mutt.\nIt defaults to the value of the <code class=\"literal\">$VISUAL<\/code>, or <code class=\"literal\">$EDITOR<\/code>, environment\nvariable, or to the string \u201cvi\u201d if neither of those are set.\n<\/p><p>\nThe <code class=\"literal\"><a class=\"link\" href=\"#editor\" title=\"3.65.\u00a0editor\">$editor<\/a><\/code> string may contain a <em>%s<\/em> escape, which will be replaced by the name\nof the file to be edited.  If the <em>%s<\/em> escape does not appear in <code class=\"literal\"><a class=\"link\" href=\"#editor\" title=\"3.65.\u00a0editor\">$editor<\/a><\/code>, a\nspace and the name to be edited are appended.\n<\/p><p>\nThe resulting string is then executed by running\n<\/p><pre class=\"screen\">\nsh -c 'string'\n<\/pre><p>\nwhere <em>string<\/em> is the expansion of <code class=\"literal\"><a class=\"link\" href=\"#editor\" title=\"3.65.\u00a0editor\">$editor<\/a><\/code> described above.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.66.\u00a0encode_from<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will quoted-printable encode messages when\nthey contain the string \u201cFrom \u201d (note the trailing space) in the beginning of a line.\nThis is useful to avoid the tampering certain mail delivery and transport\nagents tend to do with messages (in order to prevent tools from\nmisinterpreting the line as a mbox message separator).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.67.\u00a0entropy_file<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThe file which includes random data that is used to initialize SSL\nlibrary functions.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.68.\u00a0envelope_from_address<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0e-mail\u00a0address<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nManually sets the <em>envelope<\/em> sender for outgoing messages.\nThis value is ignored if <a class=\"link\" href=\"#use-envelope-from\" title=\"3.355.\u00a0use_envelope_from\">$use_envelope_from<\/a> is <em>unset<\/em>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.69.\u00a0error_history<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a030<\/p><\/div><p>\nThis variable controls the size (in number of strings remembered)\nof the error messages displayed by mutt.  These can be shown with\nthe <code class=\"literal\">&lt;error-history&gt;<\/code> function.  The history is cleared each\ntime this variable is set.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.70.\u00a0escape<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">~<\/code>\u201d<\/p><\/div><p>\nEscape character to use for functions in the built-in editor.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.71.\u00a0fast_reply<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, the initial prompt for recipients and subject are skipped\nwhen replying to messages, and the initial prompt for subject is\nskipped when forwarding messages.\n<\/p><p>\n<strong>Note:<\/strong> this variable has no effect when the <a class=\"link\" href=\"#autoedit\" title=\"3.21.\u00a0autoedit\">$autoedit<\/a>\nvariable is <em>set<\/em>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.72.\u00a0fcc_attach<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether or not attachments on outgoing messages\nare saved along with the main body of your message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.73.\u00a0fcc_clear<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, FCCs will be stored unencrypted and\nunsigned, even when the actual message is encrypted and\/or\nsigned.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.74.\u00a0flag_safe<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf set, flagged messages cannot be deleted.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.75.\u00a0folder<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/Mail<\/code>\u201d<\/p><\/div><p>\nSpecifies the default location of your mailboxes.  A \u201c+\u201d or \u201c=\u201d at the\nbeginning of a pathname will be expanded to the value of this\nvariable.  Note that if you change this variable (from the default)\nvalue you need to make sure that the assignment occurs <em>before<\/em>\nyou use \u201c+\u201d or \u201c=\u201d for any other variables since expansion takes place\nwhen handling the \u201c<a class=\"link\" href=\"#mailboxes\" title=\"14.\u00a0Monitoring Incoming Mail\">mailboxes<\/a>\u201d command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.76.\u00a0folder_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%2C\u00a0%t\u00a0%N\u00a0%F\u00a0%2l\u00a0%-8.8u\u00a0%-8.8g\u00a0%8s\u00a0%d\u00a0%f<\/code>\u201d<\/p><\/div><p>\nThis variable allows you to customize the file browser display to your\npersonal taste.  This string is similar to <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>, but has\nits own set of <code class=\"literal\">printf(3)<\/code>-like sequences:\n<\/p><p>\nFor an explanation of \u201csoft-fill\u201d, see the <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a> documentation.\n<\/p><p>\n* = can be optionally printed if nonzero\n<\/p><p>\n%m, %n, and %N only work for monitored mailboxes.\n%m requires <a class=\"link\" href=\"#mail-check-stats\" title=\"3.135.\u00a0mail_check_stats\">$mail_check_stats<\/a> to be set.\n%n requires <a class=\"link\" href=\"#mail-check-stats\" title=\"3.135.\u00a0mail_check_stats\">$mail_check_stats<\/a> to be set (except for IMAP mailboxes).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.77.\u00a0followup_to<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls whether or not the \u201cMail-Followup-To:\u201d header field is\ngenerated when sending mail.  When <em>set<\/em>, Mutt will generate this\nfield when you are replying to a known mailing list, specified with\nthe \u201c<a class=\"link\" href=\"#subscribe\">subscribe<\/a>\u201d or \u201c<a class=\"link\" href=\"#lists\" title=\"12.\u00a0Mailing Lists\">lists<\/a>\u201d commands.\n<\/p><p>\nThis field has two purposes.  First, preventing you from\nreceiving duplicate copies of replies to messages which you send\nto mailing lists, and second, ensuring that you do get a reply\nseparately for any messages sent to known lists to which you are\nnot subscribed.\n<\/p><p>\nThe header will contain only the list's address\nfor subscribed lists, and both the list address and your own\nemail address for unsubscribed lists.  Without this header, a\ngroup reply to your message sent to a subscribed list will be\nsent to both the list and your address, resulting in two copies\nof the same email for you.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.78.\u00a0force_name<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable is similar to <a class=\"link\" href=\"#save-name\" title=\"3.257.\u00a0save_name\">$save_name<\/a>, except that Mutt will\nstore a copy of your outgoing message by the username of the address\nyou are sending to even if that mailbox does not exist.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#record\" title=\"3.240.\u00a0record\">$record<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.79.\u00a0forward_attribution_intro<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">-----\u00a0Forwarded\u00a0message\u00a0from\u00a0%f\u00a0-----<\/code>\u201d<\/p><\/div><p>\nThis is the string that will precede a message which has been forwarded\nin the main body of a message (when <a class=\"link\" href=\"#mime-forward\" title=\"3.160.\u00a0mime_forward\">$mime_forward<\/a> is unset).\nFor a full listing of defined <code class=\"literal\">printf(3)<\/code>-like sequences see\nthe section on <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>.  See also <a class=\"link\" href=\"#attribution-locale\" title=\"3.19.\u00a0attribution_locale\">$attribution_locale<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.80.\u00a0forward_attribution_trailer<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">-----\u00a0End\u00a0forwarded\u00a0message\u00a0-----<\/code>\u201d<\/p><\/div><p>\nThis is the string that will follow a message which has been forwarded\nin the main body of a message (when <a class=\"link\" href=\"#mime-forward\" title=\"3.160.\u00a0mime_forward\">$mime_forward<\/a> is unset).\nFor a full listing of defined <code class=\"literal\">printf(3)<\/code>-like sequences see\nthe section on <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>.  See also <a class=\"link\" href=\"#attribution-locale\" title=\"3.19.\u00a0attribution_locale\">$attribution_locale<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.81.\u00a0forward_decode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls the decoding of complex MIME messages into <code class=\"literal\">text\/plain<\/code> when\nforwarding a message.  The message header is also RFC2047 decoded.\nThis variable is only used, if <a class=\"link\" href=\"#mime-forward\" title=\"3.160.\u00a0mime_forward\">$mime_forward<\/a> is <em>unset<\/em>,\notherwise <a class=\"link\" href=\"#mime-forward-decode\" title=\"3.161.\u00a0mime_forward_decode\">$mime_forward_decode<\/a> is used instead.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.82.\u00a0forward_decrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls the handling of encrypted messages when forwarding a message.\nWhen <em>set<\/em>, the outer layer of encryption is stripped off.  This\nvariable is only used if <a class=\"link\" href=\"#mime-forward\" title=\"3.160.\u00a0mime_forward\">$mime_forward<\/a> is <em>set<\/em> and\n<a class=\"link\" href=\"#mime-forward-decode\" title=\"3.161.\u00a0mime_forward_decode\">$mime_forward_decode<\/a> is <em>unset<\/em>.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.83.\u00a0forward_edit<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis quadoption controls whether or not the user is automatically\nplaced in the editor when forwarding messages.  For those who always want\nto forward with no modification, use a setting of \u201cno\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.84.\u00a0forward_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">[%a:\u00a0%s]<\/code>\u201d<\/p><\/div><p>\nThis variable controls the default subject when forwarding a message.\nIt uses the same format sequences as the <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.85.\u00a0forward_quote<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, forwarded messages included in the main body of the\nmessage (when <a class=\"link\" href=\"#mime-forward\" title=\"3.160.\u00a0mime_forward\">$mime_forward<\/a> is <em>unset<\/em>) will be quoted using\n<a class=\"link\" href=\"#indent-string\" title=\"3.129.\u00a0indent_string\">$indent_string<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.86.\u00a0from<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0e-mail\u00a0address<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nWhen <em>set<\/em>, this variable contains a default from address.  It\ncan be overridden using \u201c<a class=\"link\" href=\"#my-hdr\" title=\"15.\u00a0User-Defined Headers\">my_hdr<\/a>\u201d (including from a \u201c<a class=\"link\" href=\"#send-hook\" title=\"19.\u00a0Change Settings Based Upon Message Recipients\">send-hook<\/a>\u201d) and\n<a class=\"link\" href=\"#reverse-name\" title=\"3.251.\u00a0reverse_name\">$reverse_name<\/a>.  This variable is ignored if <a class=\"link\" href=\"#use-from\" title=\"3.356.\u00a0use_from\">$use_from<\/a> is <em>unset<\/em>.\n<\/p><p>\nThis setting defaults to the contents of the environment variable <code class=\"literal\">$EMAIL<\/code>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.87.\u00a0gecos_mask<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0\u201c<code class=\"literal\">^[^,]*<\/code>\u201d<\/p><\/div><p>\nA regular expression used by mutt to parse the GECOS field of a password\nentry when expanding the alias.  The default value\nwill return the string up to the first \u201c,\u201d encountered.\nIf the GECOS field contains a string like \u201clastname, firstname\u201d then you\nshould set it to \u201c<code class=\"literal\">.*<\/code>\u201d.\n<\/p><p>\nThis can be useful if you see the following behavior: you address an e-mail\nto user ID \u201cstevef\u201d whose full name is \u201cSteve Franklin\u201d.  If mutt expands\n\u201cstevef\u201d to \u201c\"Franklin\" stevef@foo.bar\u201d then you should set the <a class=\"link\" href=\"#gecos-mask\" title=\"3.87.\u00a0gecos_mask\">$gecos_mask<\/a> to\na regular expression that will match the whole name so mutt will expand\n\u201cFranklin\u201d to \u201cFranklin, Steve\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.88.\u00a0hdrs<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>unset<\/em>, the header fields normally added by the \u201c<a class=\"link\" href=\"#my-hdr\" title=\"15.\u00a0User-Defined Headers\">my_hdr<\/a>\u201d\ncommand are not created.  This variable <em>must<\/em> be unset before\ncomposing a new message or replying in order to take effect.  If <em>set<\/em>,\nthe user defined header fields are added to every new message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.89.\u00a0header<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, this variable causes Mutt to include the header\nof the message you are replying to into the edit buffer.\nThe <a class=\"link\" href=\"#weed\" title=\"3.361.\u00a0weed\">$weed<\/a> setting applies.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.90.\u00a0header_cache<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable points to the header cache database.\nIf pointing to a directory Mutt will contain a header cache\ndatabase file per folder, if pointing to a file that file will\nbe a single global header cache. By default it is <em>unset<\/em> so no header\ncaching will be used.\n<\/p><p>\nHeader caching can greatly improve speed when opening POP, IMAP\nMH or Maildir folders, see \u201c<a class=\"link\" href=\"#caching\" title=\"7.\u00a0Local Caching\">caching<\/a>\u201d for details.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.91.\u00a0header_cache_compress<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen mutt is compiled with qdbm, tokyocabinet, or kyotocabinet as header\ncache backend, this option determines whether the database will be compressed.\nCompression results in database files roughly being one fifth\nof the usual diskspace, but the decompression can result in a\nslower opening of cached folder(s) which in general is still\nmuch faster than opening non header cached folders.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.92.\u00a0header_cache_pagesize<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">16384<\/code>\u201d<\/p><\/div><p>\nWhen mutt is compiled with either gdbm or bdb4 as the header cache backend,\nthis option changes the database page size.  Too large or too small\nvalues can waste space, memory, or CPU time. The default should be more\nor less optimal for most use cases.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.93.\u00a0header_color_partial<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, color header regexps behave like color body regexps:\ncolor is applied to the exact text matched by the regexp.  When\n<em>unset<\/em>, color is applied to the entire header.\n<\/p><p>\nOne use of this option might be to apply color to just the header labels.\n<\/p><p>\nSee \u201c<a class=\"link\" href=\"#color\" title=\"9.\u00a0Using Color and Mono Video Attributes\">color<\/a>\u201d for more details.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.94.\u00a0help<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, help lines describing the bindings for the major functions\nprovided by each menu are displayed on the first line of the screen.\n<\/p><p>\n<strong>Note:<\/strong> The binding will not be displayed correctly if the\nfunction is bound to a sequence rather than a single keystroke.  Also,\nthe help line may not be updated if a binding is changed while Mutt is\nrunning.  Since this variable is primarily aimed at new users, neither\nof these should present a major problem.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.95.\u00a0hidden_host<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will skip the host name part of <a class=\"link\" href=\"#hostname\" title=\"3.106.\u00a0hostname\">$hostname<\/a> variable\nwhen adding the domain part to addresses.  This variable does not\naffect the generation of Message-IDs, and it will not lead to the\ncut-off of first-level domains.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.96.\u00a0hide_limited<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will not show the presence of messages that are hidden\nby limiting, in the thread tree.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.97.\u00a0hide_missing<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will not show the presence of missing messages in the\nthread tree.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.98.\u00a0hide_thread_subject<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will not show the subject of messages in the thread\ntree that have the same subject as their parent or closest previously\ndisplayed sibling.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.99.\u00a0hide_top_limited<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will not show the presence of messages that are hidden\nby limiting, at the top of threads in the thread tree.  Note that when\n<a class=\"link\" href=\"#hide-limited\" title=\"3.96.\u00a0hide_limited\">$hide_limited<\/a> is <em>set<\/em>, this option will have no effect.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.100.\u00a0hide_top_missing<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will not show the presence of missing messages at the\ntop of threads in the thread tree.  Note that when <a class=\"link\" href=\"#hide-missing\" title=\"3.97.\u00a0hide_missing\">$hide_missing<\/a> is\n<em>set<\/em>, this option will have no effect.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.101.\u00a0history<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a010<\/p><\/div><p>\nThis variable controls the size (in number of strings remembered) of\nthe string history buffer per category. The buffer is cleared each time the\nvariable is set.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.102.\u00a0history_file<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/.mutthistory<\/code>\u201d<\/p><\/div><p>\nThe file in which Mutt will save its history.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#save-history\" title=\"3.256.\u00a0save_history\">$save_history<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.103.\u00a0history_remove_dups<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, all of the string history will be scanned for duplicates\nwhen a new entry is added.  Duplicate entries in the <a class=\"link\" href=\"#history-file\" title=\"3.102.\u00a0history_file\">$history_file<\/a> will\nalso be removed when it is periodically compacted.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.104.\u00a0honor_disposition<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will not display attachments with a\ndisposition of \u201cattachment\u201d inline even if it could\nrender the part to plain text. These MIME parts can only\nbe viewed from the attachment menu.\n<\/p><p>\nIf <em>unset<\/em>, Mutt will render all MIME parts it can\nproperly transform to plain text.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.105.\u00a0honor_followup_to<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether or not a Mail-Followup-To header is\nhonored when group-replying to a message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.106.\u00a0hostname<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSpecifies the fully-qualified hostname of the system mutt is running on\ncontaining the host's name and the DNS domain it belongs to. It is used\nas the domain part (after \u201c@\u201d) for local email addresses as well as\nMessage-Id headers.\n<\/p><p>\nIts value is determined at startup as follows: the node's\nhostname is first determined by the <code class=\"literal\">uname(3)<\/code> function.  The\ndomain is then looked up using the <code class=\"literal\">gethostname(2)<\/code> and\n<code class=\"literal\">getaddrinfo(3)<\/code> functions.  If those calls are unable to\ndetermine the domain, the full value returned by uname is used.\nOptionally, Mutt can be compiled with a fixed domain name in\nwhich case a detected one is not used.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#use-domain\" title=\"3.354.\u00a0use_domain\">$use_domain<\/a> and <a class=\"link\" href=\"#hidden-host\" title=\"3.95.\u00a0hidden_host\">$hidden_host<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.107.\u00a0idn_decode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will show you international domain names decoded.\nNote: You can use IDNs for addresses even if this is <em>unset<\/em>.\nThis variable only affects decoding. (IDN only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.108.\u00a0idn_encode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will encode international domain names using\nIDN.  Unset this if your SMTP server can handle newer (RFC 6531)\nUTF-8 encoded domains. (IDN only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.109.\u00a0ignore_linear_white_space<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis option replaces linear-white-space between encoded-word\nand text to a single space to prevent the display of MIME-encoded\n\u201cSubject:\u201d field from being divided into multiple lines.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.110.\u00a0ignore_list_reply_to<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nAffects the behavior of the <code class=\"literal\">&lt;reply&gt;<\/code> function when replying to\nmessages from mailing lists (as defined by the \u201c<a class=\"link\" href=\"#subscribe\">subscribe<\/a>\u201d or\n\u201c<a class=\"link\" href=\"#lists\" title=\"12.\u00a0Mailing Lists\">lists<\/a>\u201d commands).  When <em>set<\/em>, if the \u201cReply-To:\u201d field is\nset to the same value as the \u201cTo:\u201d field, Mutt assumes that the\n\u201cReply-To:\u201d field was set by the mailing list to automate responses\nto the list, and will ignore this field.  To direct a response to the\nmailing list when this option is <em>set<\/em>, use the <code class=\"literal\"><a class=\"link\" href=\"#list-reply\">&lt;list-reply&gt;<\/a><\/code>\nfunction; <code class=\"literal\">&lt;group-reply&gt;<\/code> will reply to both the sender and the\nlist.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.111.\u00a0imap_authenticators<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis is a colon-delimited list of authentication methods mutt may\nattempt to use to log in to an IMAP server, in the order mutt should\ntry them.  Authentication methods are either \u201clogin\u201d or the right\nside of an IMAP \u201cAUTH=xxx\u201d capability string, e.g. \u201cdigest-md5\u201d, \u201cgssapi\u201d\nor \u201ccram-md5\u201d. This option is case-insensitive. If it's\n<em>unset<\/em> (the default) mutt will try all available methods,\nin order from most-secure to least-secure.\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset imap_authenticators=\"gssapi:cram-md5:login\"\n<\/pre><p>\n<strong>Note:<\/strong> Mutt will only fall back to other authentication methods if\nthe previous methods are unavailable. If a method is available but\nauthentication fails, mutt will not connect to the IMAP server.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.112.\u00a0imap_check_subscribed<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will fetch the set of subscribed folders from\nyour server on connection, and add them to the set of mailboxes\nit polls for new mail just as if you had issued individual \u201c<a class=\"link\" href=\"#mailboxes\" title=\"14.\u00a0Monitoring Incoming Mail\">mailboxes<\/a>\u201d\ncommands.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.113.\u00a0imap_delim_chars<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">\/.<\/code>\u201d<\/p><\/div><p>\nThis contains the list of characters which you would like to treat\nas folder separators for displaying IMAP paths. In particular it\nhelps in using the \u201c=\u201d shortcut for your <em>folder<\/em> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.114.\u00a0imap_headers<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nMutt requests these header fields in addition to the default headers\n(\u201cDate:\u201d, \u201cFrom:\u201d, \u201cSubject:\u201d, \u201cTo:\u201d, \u201cCc:\u201d, \u201cMessage-Id:\u201d,\n\u201cReferences:\u201d, \u201cContent-Type:\u201d, \u201cContent-Description:\u201d, \u201cIn-Reply-To:\u201d,\n\u201cReply-To:\u201d, \u201cLines:\u201d, \u201cList-Post:\u201d, \u201cX-Label:\u201d) from IMAP\nservers before displaying the index menu. You may want to add more\nheaders for spam detection.\n<\/p><p>\n<strong>Note:<\/strong> This is a space separated list, items should be uppercase\nand not contain the colon, e.g. \u201cX-BOGOSITY X-SPAM-STATUS\u201d for the\n\u201cX-Bogosity:\u201d and \u201cX-Spam-Status:\u201d header fields.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.115.\u00a0imap_idle<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will attempt to use the IMAP IDLE extension\nto check for new mail in the current mailbox. Some servers\n(dovecot was the inspiration for this option) react badly\nto mutt's implementation. If your connection seems to freeze\nup periodically, try unsetting this.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.116.\u00a0imap_keepalive<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a0300<\/p><\/div><p>\nThis variable specifies the maximum amount of time in seconds that mutt\nwill wait before polling open IMAP connections, to prevent the server\nfrom closing them before mutt has finished with them. The default is\nwell within the RFC-specified minimum amount of time (30 minutes) before\na server is allowed to do this, but in practice the RFC does get\nviolated every now and then. Reduce this number if you find yourself\ngetting disconnected from your IMAP server due to inactivity.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.117.\u00a0imap_list_subscribed<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable configures whether IMAP folder browsing will look for\nonly subscribed folders or all folders.  This can be toggled in the\nIMAP browser with the <code class=\"literal\">&lt;toggle-subscribed&gt;<\/code> function.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.118.\u00a0imap_login<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nYour login name on the IMAP server.\n<\/p><p>\nThis variable defaults to the value of <a class=\"link\" href=\"#imap-user\" title=\"3.125.\u00a0imap_user\">$imap_user<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.119.\u00a0imap_pass<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSpecifies the password for your IMAP account.  If <em>unset<\/em>, Mutt will\nprompt you for your password when you invoke the <code class=\"literal\">&lt;imap-fetch-mail&gt;<\/code> function\nor try to open an IMAP folder.\n<\/p><p>\n<strong>Warning<\/strong>: you should only use this option when you are on a\nfairly secure machine, because the superuser can read your muttrc even\nif you are the only one who can read the file.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.120.\u00a0imap_passive<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will not open new IMAP connections to check for new\nmail.  Mutt will only check for new mail over existing IMAP\nconnections.  This is useful if you don't want to be prompted to\nuser\/password pairs on mutt invocation, or if opening the connection\nis slow.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.121.\u00a0imap_peek<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will avoid implicitly marking your mail as read whenever\nyou fetch a message from the server. This is generally a good thing,\nbut can make closing an IMAP folder somewhat slower. This option\nexists to appease speed freaks.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.122.\u00a0imap_pipeline_depth<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a015<\/p><\/div><p>\nControls the number of IMAP commands that may be queued up before they\nare sent to the server. A deeper pipeline reduces the amount of time\nmutt must wait for the server, and can make IMAP servers feel much\nmore responsive. But not all servers correctly handle pipelined commands,\nso if you have problems you might want to try setting this variable to 0.\n<\/p><p>\n<strong>Note:<\/strong> Changes to this variable have no effect on open connections.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.123.\u00a0imap_poll_timeout<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a015<\/p><\/div><p>\nThis variable specifies the maximum amount of time in seconds\nthat mutt will wait for a response when polling IMAP connections\nfor new mail, before timing out and closing the connection.  Set\nto 0 to disable timing out.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.124.\u00a0imap_servernoise<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will display warning messages from the IMAP\nserver as error messages. Since these messages are often\nharmless, or generated due to configuration problems on the\nserver which are out of the users' hands, you may wish to suppress\nthem at some point.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.125.\u00a0imap_user<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThe name of the user whose mail you intend to access on the IMAP\nserver.\n<\/p><p>\nThis variable defaults to your user name on the local machine.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.126.\u00a0implicit_autoview<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf set to \u201cyes\u201d, mutt will look for a mailcap entry with the\n\u201c<code class=\"literal\">copiousoutput<\/code>\u201d flag set for <em>every<\/em> MIME attachment it doesn't have\nan internal viewer defined for.  If such an entry is found, mutt will\nuse the viewer defined in that entry to convert the body part to text\nform.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.127.\u00a0include<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nControls whether or not a copy of the message(s) you are replying to\nis included in your reply.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.128.\u00a0include_onlyfirst<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nControls whether or not Mutt includes only the first attachment\nof the message you are replying.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.129.\u00a0indent_string<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">&gt;\u00a0<\/code>\u201d<\/p><\/div><p>\nSpecifies the string to prepend to each line of text quoted in a\nmessage to which you are replying.  You are strongly encouraged not to\nchange this value, as it tends to agitate the more fanatical netizens.\n<\/p><p>\nThe value of this option is ignored if <a class=\"link\" href=\"#text-flowed\" title=\"3.338.\u00a0text_flowed\">$text_flowed<\/a> is set, because\nthe quoting mechanism is strictly defined for format=flowed.\n<\/p><p>\nThis option is a format string, please see the description of\n<a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a> for supported <code class=\"literal\">printf(3)<\/code>-style sequences.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.130.\u00a0index_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%4C\u00a0%Z\u00a0%{%b\u00a0%d}\u00a0%-15.15L\u00a0(%?l?%4l&amp;%4c?)\u00a0%s<\/code>\u201d<\/p><\/div><p>\nThis variable allows you to customize the message index display to\nyour personal taste.\n<\/p><p>\n\u201cFormat strings\u201d are similar to the strings used in the C\nfunction <code class=\"literal\">printf(3)<\/code> to format output (see the man page for more details).\nFor an explanation of the %? construct, see the <a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a> description.\nThe following sequences are defined in Mutt:\n<\/p><p>\n\u201cSoft-fill\u201d deserves some explanation: Normal right-justification\nwill print everything to the left of the \u201c%&gt;\u201d, displaying padding and\nwhatever lies to the right only if there's room. By contrast,\nsoft-fill gives priority to the right-hand side, guaranteeing space\nto display it and showing padding only if there's still room. If\nnecessary, soft-fill will eat text leftwards to make room for\nrightward text.\n<\/p><p>\nNote that these expandos are supported in\n\u201c<a class=\"link\" href=\"#save-hook\" title=\"16.\u00a0Specify Default Save Mailbox\">save-hook<\/a>\u201d, \u201c<a class=\"link\" href=\"#fcc-hook\" title=\"17.\u00a0Specify Default Fcc: Mailbox When Composing\">fcc-hook<\/a>\u201d and \u201c<a class=\"link\" href=\"#fcc-save-hook\" title=\"18.\u00a0Specify Default Save Filename and Default Fcc: Mailbox at Once\">fcc-save-hook<\/a>\u201d, too.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.131.\u00a0ispell<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">\/usr\/bin\/ispell<\/code>\u201d<\/p><\/div><p>\nHow to invoke ispell (GNU's spell-checking software).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.132.\u00a0keep_flagged<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, read messages marked as flagged will not be moved\nfrom your spool mailbox to your <a class=\"link\" href=\"#mbox\" title=\"3.146.\u00a0mbox\">$mbox<\/a> mailbox, or as a result of\na \u201c<a class=\"link\" href=\"#mbox-hook\" title=\"13.\u00a0Using Multiple Spool Mailboxes\">mbox-hook<\/a>\u201d command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.133.\u00a0mail_check<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a05<\/p><\/div><p>\nThis variable configures how often (in seconds) mutt should look for\nnew mail. Also see the <a class=\"link\" href=\"#timeout\" title=\"3.343.\u00a0timeout\">$timeout<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.134.\u00a0mail_check_recent<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will only notify you about new mail that has been received\nsince the last time you opened the mailbox.  When <em>unset<\/em>, Mutt will notify you\nif any new mail exists in the mailbox, regardless of whether you have visited it\nrecently.\n<\/p><p>\nWhen <em><a class=\"link\" href=\"#mark-old\" title=\"3.143.\u00a0mark_old\">$mark_old<\/a><\/em> is set, Mutt does not consider the mailbox to contain new\nmail if only old messages exist.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.135.\u00a0mail_check_stats<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will periodically calculate message\nstatistics of a mailbox while polling for new mail.  It will\ncheck for unread, flagged, and total message counts.  Because\nthis operation is more performance intensive, it defaults to\n<em>unset<\/em>, and has a separate option, <a class=\"link\" href=\"#mail-check-stats-interval\" title=\"3.136.\u00a0mail_check_stats_interval\">$mail_check_stats_interval<\/a>, to\ncontrol how often to update these counts.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.136.\u00a0mail_check_stats_interval<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a060<\/p><\/div><p>\nWhen <a class=\"link\" href=\"#mail-check-stats\" title=\"3.135.\u00a0mail_check_stats\">$mail_check_stats<\/a> is <em>set<\/em>, this variable configures\nhow often (in seconds) mutt will update message counts.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.137.\u00a0mailcap_path<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable specifies which files to consult when attempting to\ndisplay MIME bodies not directly supported by Mutt.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.138.\u00a0mailcap_sanitize<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, mutt will restrict possible characters in mailcap % expandos\nto a well-defined set of safe characters.  This is the safe setting,\nbut we are not sure it doesn't break some more advanced MIME stuff.\n<\/p><p>\n<strong>DON'T CHANGE THIS SETTING UNLESS YOU ARE REALLY SURE WHAT YOU ARE\nDOING!<\/strong>\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.139.\u00a0maildir_header_cache_verify<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nCheck for Maildir unaware programs other than mutt having modified maildir\nfiles when the header cache is in use.  This incurs one <code class=\"literal\">stat(2)<\/code> per\nmessage every time the folder is opened (which can be very slow for NFS\nfolders).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.140.\u00a0maildir_trash<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, messages marked as deleted will be saved with the maildir\ntrashed flag instead of unlinked.  <strong>Note:<\/strong> this only applies\nto maildir-style mailboxes.  Setting it will have no effect on other\nmailbox types.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.141.\u00a0maildir_check_cur<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, mutt will poll both the new and cur directories of\na maildir folder for new messages.  This might be useful if other\nprograms interacting with the folder (e.g. dovecot) are moving new\nmessages to the cur directory.  Note that setting this option may\nslow down polling for new messages in large folders, since mutt has\nto scan all cur messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.142.\u00a0mark_macro_prefix<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">'<\/code>\u201d<\/p><\/div><p>\nPrefix for macros created using mark-message.  A new macro\nautomatically generated with <em>&lt;mark-message&gt;a<\/em> will be composed\nfrom this prefix and the letter <em>a<\/em>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.143.\u00a0mark_old<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls whether or not mutt marks <em>new<\/em> <strong>unread<\/strong>\nmessages as <em>old<\/em> if you exit a mailbox without reading them.\nWith this option <em>set<\/em>, the next time you start mutt, the messages\nwill show up with an \u201cO\u201d next to them in the index menu,\nindicating that they are old.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.144.\u00a0markers<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls the display of wrapped lines in the internal pager. If set, a\n\u201c+\u201d marker is displayed at the beginning of wrapped lines.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#smart-wrap\" title=\"3.283.\u00a0smart_wrap\">$smart_wrap<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.145.\u00a0mask<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0\u201c<code class=\"literal\">!^\\.[^.]<\/code>\u201d<\/p><\/div><p>\nA regular expression used in the file browser, optionally preceded by\nthe <em>not<\/em> operator \u201c!\u201d.  Only files whose names match this mask\nwill be shown. The match is always case-sensitive.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.146.\u00a0mbox<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/mbox<\/code>\u201d<\/p><\/div><p>\nThis specifies the folder into which read mail in your <a class=\"link\" href=\"#spoolfile\" title=\"3.317.\u00a0spoolfile\">$spoolfile<\/a>\nfolder will be appended.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#move\" title=\"3.167.\u00a0move\">$move<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.147.\u00a0mbox_type<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0folder\u00a0magic<br \/>Default:\u00a0mbox<\/p><\/div><p>\nThe default mailbox type used when creating new folders. May be any of\n\u201cmbox\u201d, \u201cMMDF\u201d, \u201cMH\u201d and \u201cMaildir\u201d. This is overridden by the\n<code class=\"literal\">-m<\/code> command-line option.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.148.\u00a0menu_context<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nThis variable controls the number of lines of context that are given\nwhen scrolling through menus. (Similar to <a class=\"link\" href=\"#pager-context\" title=\"3.172.\u00a0pager_context\">$pager_context<\/a>.)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.149.\u00a0menu_move_off<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>unset<\/em>, the bottom entry of menus will never scroll up past\nthe bottom of the screen, unless there are less entries than lines.\nWhen <em>set<\/em>, the bottom entry may move off the bottom.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.150.\u00a0menu_scroll<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, menus will be scrolled up or down one line when you\nattempt to move across a screen boundary.  If <em>unset<\/em>, the screen\nis cleared and the next or previous page of the menu is displayed\n(useful for slow links to avoid many redraws).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.151.\u00a0message_cache_clean<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, mutt will clean out obsolete entries from the message cache when\nthe mailbox is synchronized. You probably only want to set it\nevery once in a while, since it can be a little slow\n(especially for large folders).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.152.\u00a0message_cachedir<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSet this to a directory and mutt will cache copies of messages from\nyour IMAP and POP servers here. You are free to remove entries at any\ntime.\n<\/p><p>\nWhen setting this variable to a directory, mutt needs to fetch every\nremote message only once and can perform regular expression searches\nas fast as for local folders.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#message-cache-clean\" title=\"3.151.\u00a0message_cache_clean\">$message_cache_clean<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.153.\u00a0message_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%s<\/code>\u201d<\/p><\/div><p>\nThis is the string displayed in the \u201cattachment\u201d menu for\nattachments of type <code class=\"literal\">message\/rfc822<\/code>.  For a full listing of defined\n<code class=\"literal\">printf(3)<\/code>-like sequences see the section on <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.154.\u00a0meta_key<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, forces Mutt to interpret keystrokes with the high bit (bit 8)\nset as if the user had pressed the Esc key and whatever key remains\nafter having the high bit removed.  For example, if the key pressed\nhas an ASCII value of <code class=\"literal\">0xf8<\/code>, then this is treated as if the user had\npressed Esc then \u201cx\u201d.  This is because the result of removing the\nhigh bit from <code class=\"literal\">0xf8<\/code> is <code class=\"literal\">0x78<\/code>, which is the ASCII character\n\u201cx\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.155.\u00a0metoo<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>unset<\/em>, Mutt will remove your address (see the \u201c<a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a>\u201d\ncommand) from the list of recipients when replying to a message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.156.\u00a0mh_purge<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>unset<\/em>, mutt will mimic mh's behavior and rename deleted messages\nto <em>,&lt;old file name&gt;<\/em> in mh folders instead of really deleting\nthem. This leaves the message on disk but makes programs reading the folder\nignore it. If the variable is <em>set<\/em>, the message files will simply be\ndeleted.\n<\/p><p>\nThis option is similar to <a class=\"link\" href=\"#maildir-trash\" title=\"3.140.\u00a0maildir_trash\">$maildir_trash<\/a> for Maildir folders.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.157.\u00a0mh_seq_flagged<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">flagged<\/code>\u201d<\/p><\/div><p>\nThe name of the MH sequence used for flagged messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.158.\u00a0mh_seq_replied<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">replied<\/code>\u201d<\/p><\/div><p>\nThe name of the MH sequence used to tag replied messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.159.\u00a0mh_seq_unseen<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">unseen<\/code>\u201d<\/p><\/div><p>\nThe name of the MH sequence used for unseen messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.160.\u00a0mime_forward<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, the message you are forwarding will be attached as a\nseparate <code class=\"literal\">message\/rfc822<\/code> MIME part instead of included in the main body of the\nmessage.  This is useful for forwarding MIME messages so the receiver\ncan properly view the message as it was delivered to you. If you like\nto switch between MIME and not MIME from mail to mail, set this\nvariable to \u201cask-no\u201d or \u201cask-yes\u201d.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#forward-decode\" title=\"3.81.\u00a0forward_decode\">$forward_decode<\/a> and <a class=\"link\" href=\"#mime-forward-decode\" title=\"3.161.\u00a0mime_forward_decode\">$mime_forward_decode<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.161.\u00a0mime_forward_decode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nControls the decoding of complex MIME messages into <code class=\"literal\">text\/plain<\/code> when\nforwarding a message while <a class=\"link\" href=\"#mime-forward\" title=\"3.160.\u00a0mime_forward\">$mime_forward<\/a> is <em>set<\/em>. Otherwise\n<a class=\"link\" href=\"#forward-decode\" title=\"3.81.\u00a0forward_decode\">$forward_decode<\/a> is used instead.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.162.\u00a0mime_forward_rest<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen forwarding multiple attachments of a MIME message from the attachment\nmenu, attachments which cannot be decoded in a reasonable manner will\nbe attached to the newly composed message if this option is <em>set<\/em>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.163.\u00a0mime_type_query_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis specifies a command to run, to determine the mime type of a\nnew attachment when composing a message.  Unless\n<a class=\"link\" href=\"#mime-type-query-first\" title=\"3.164.\u00a0mime_type_query_first\">$mime_type_query_first<\/a> is set, this will only be run if the\nattachment's extension is not found in the mime.types file.\n<\/p><p>\nThe string may contain a \u201c%s\u201d, which will be substituted with the\nattachment filename.  Mutt will add quotes around the string substituted\nfor \u201c%s\u201d automatically according to shell quoting rules, so you should\navoid adding your own.  If no \u201c%s\u201d is found in the string, Mutt will\nappend the attachment filename to the end of the string.\n<\/p><p>\nThe command should output a single line containing the\nattachment's mime type.\n<\/p><p>\nSuggested values are \u201cxdg-mime query filetype\u201d or\n\u201cfile -bi\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.164.\u00a0mime_type_query_first<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, the <a class=\"link\" href=\"#mime-type-query-command\" title=\"3.163.\u00a0mime_type_query_command\">$mime_type_query_command<\/a> will be run before the\nmime.types lookup.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.165.\u00a0mix_entry_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%4n\u00a0%c\u00a0%-16s\u00a0%a<\/code>\u201d<\/p><\/div><p>\nThis variable describes the format of a remailer line on the mixmaster\nchain selection screen.  The following <code class=\"literal\">printf(3)<\/code>-like sequences are\nsupported:\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.166.\u00a0mixmaster<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">mixmaster<\/code>\u201d<\/p><\/div><p>\nThis variable contains the path to the Mixmaster binary on your\nsystem.  It is used with various sets of parameters to gather the\nlist of known remailers, and to finally send a message through the\nmixmaster chain.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.167.\u00a0move<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0no<\/p><\/div><p>\nControls whether or not Mutt will move read messages\nfrom your spool mailbox to your <a class=\"link\" href=\"#mbox\" title=\"3.146.\u00a0mbox\">$mbox<\/a> mailbox, or as a result of\na \u201c<a class=\"link\" href=\"#mbox-hook\" title=\"13.\u00a0Using Multiple Spool Mailboxes\">mbox-hook<\/a>\u201d command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.168.\u00a0narrow_tree<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable, when <em>set<\/em>, makes the thread tree narrower, allowing\ndeeper threads to fit on the screen.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.169.\u00a0net_inc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a010<\/p><\/div><p>\nOperations that expect to transfer a large amount of data over the\nnetwork will update their progress every <a class=\"link\" href=\"#net-inc\" title=\"3.169.\u00a0net_inc\">$net_inc<\/a> kilobytes.\nIf set to 0, no progress messages will be displayed.\n<\/p><p>\nSee also <a class=\"link\" href=\"#read-inc\" title=\"3.236.\u00a0read_inc\">$read_inc<\/a>, <a class=\"link\" href=\"#write-inc\" title=\"3.367.\u00a0write_inc\">$write_inc<\/a> and <a class=\"link\" href=\"#net-inc\" title=\"3.169.\u00a0net_inc\">$net_inc<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.170.\u00a0new_mail_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will call this command after a new message is received.\nSee the <a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a> documentation for the values that can be formatted\ninto this command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.171.\u00a0pager<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">builtin<\/code>\u201d<\/p><\/div><p>\nThis variable specifies which pager you would like to use to view\nmessages. The value \u201cbuiltin\u201d means to use the built-in pager, otherwise this\nvariable should specify the pathname of the external pager you would\nlike to use.\n<\/p><p>\nUsing an external pager may have some disadvantages: Additional\nkeystrokes are necessary because you can't call mutt functions\ndirectly from the pager, and screen resizes cause lines longer than\nthe screen width to be badly formatted in the help menu.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.172.\u00a0pager_context<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nThis variable controls the number of lines of context that are given\nwhen displaying the next or previous page in the internal pager.  By\ndefault, Mutt will display the line after the last one on the screen\nat the top of the next page (0 lines of context).\n<\/p><p>\nThis variable also specifies the amount of context given for search\nresults. If positive, this many lines will be given before a match,\nif 0, the match will be top-aligned.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.173.\u00a0pager_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">-%Z-\u00a0%C\/%m:\u00a0%-20.20n\u00a0\u00a0\u00a0%s%*\u00a0\u00a0--\u00a0(%P)<\/code>\u201d<\/p><\/div><p>\nThis variable controls the format of the one-line message \u201cstatus\u201d\ndisplayed before each message in either the internal or an external\npager.  The valid sequences are listed in the <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>\nsection.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.174.\u00a0pager_index_lines<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nDetermines the number of lines of a mini-index which is shown when in\nthe pager.  The current message, unless near the top or bottom of the\nfolder, will be roughly one third of the way down this mini-index,\ngiving the reader the context of a few messages before and after the\nmessage.  This is useful, for example, to determine how many messages\nremain to be read in the current thread.  One of the lines is reserved\nfor the status bar from the index, so a setting of 6\nwill only show 5 lines of the actual index.  A value of 0 results in\nno index being shown.  If the number of messages in the current folder\nis less than <a class=\"link\" href=\"#pager-index-lines\" title=\"3.174.\u00a0pager_index_lines\">$pager_index_lines<\/a>, then the index will only use as\nmany lines as it needs.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.175.\u00a0pager_stop<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, the internal-pager will <strong>not<\/strong> move to the next message\nwhen you are at the end of a message and invoke the <code class=\"literal\">&lt;next-page&gt;<\/code>\nfunction.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.176.\u00a0pgp_auto_decode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, mutt will automatically attempt to decrypt traditional PGP\nmessages whenever the user performs an operation which ordinarily would\nresult in the contents of the message being operated on.  For example,\nif the user displays a pgp-traditional message which has not been manually\nchecked with the <code class=\"literal\"><a class=\"link\" href=\"#check-traditional-pgp\">&lt;check-traditional-pgp&gt;<\/a><\/code> function, mutt will automatically\ncheck the message for traditional pgp.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.177.\u00a0pgp_autoinline<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis option controls whether Mutt generates old-style inline\n(traditional) PGP encrypted or signed messages under certain\ncircumstances.  This can be overridden by use of the pgp menu,\nwhen inline is not required.  The GPGME backend does not support\nthis option.\n<\/p><p>\nNote that Mutt might automatically use PGP\/MIME for messages\nwhich consist of more than a single MIME part.  Mutt can be\nconfigured to ask before sending PGP\/MIME messages when inline\n(traditional) would not work.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#pgp-mime-auto\" title=\"3.196.\u00a0pgp_mime_auto\">$pgp_mime_auto<\/a> variable.\n<\/p><p>\nAlso note that using the old-style PGP message format is <strong>strongly<\/strong>\n<strong>deprecated<\/strong>.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.178.\u00a0pgp_check_exit<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, mutt will check the exit code of the PGP subprocess when\nsigning or encrypting.  A non-zero exit code means that the\nsubprocess failed.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.179.\u00a0pgp_check_gpg_decrypt_status_fd<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, mutt will check the status file descriptor output\nof <a class=\"link\" href=\"#pgp-decrypt-command\" title=\"3.182.\u00a0pgp_decrypt_command\">$pgp_decrypt_command<\/a> and <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> for GnuPG status codes\nindicating successful decryption.  This will check for the presence of\nDECRYPTION_OKAY, absence of DECRYPTION_FAILED, and that all\nPLAINTEXT occurs between the BEGIN_DECRYPTION and END_DECRYPTION\nstatus codes.\n<\/p><p>\nIf <em>unset<\/em>, mutt will instead match the status fd output\nagainst <a class=\"link\" href=\"#pgp-decryption-okay\" title=\"3.183.\u00a0pgp_decryption_okay\">$pgp_decryption_okay<\/a>.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.180.\u00a0pgp_clearsign_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis format is used to create an old-style \u201cclearsigned\u201d PGP\nmessage.  Note that the use of this format is <strong>strongly<\/strong>\n<strong>deprecated<\/strong>.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.181.\u00a0pgp_decode_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis format strings specifies a command which is used to decode\napplication\/pgp attachments.\n<\/p><p>\nThe PGP command formats have their own set of <code class=\"literal\">printf(3)<\/code>-like sequences:\n<\/p><p>\nFor examples on how to configure these formats for the various versions\nof PGP which are floating around, see the pgp and gpg sample configuration files in\nthe <code class=\"literal\">samples\/<\/code> subdirectory which has been installed on your system\nalongside the documentation.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.182.\u00a0pgp_decrypt_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to decrypt a PGP encrypted message.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.183.\u00a0pgp_decryption_okay<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf you assign text to this variable, then an encrypted PGP\nmessage is only considered successfully decrypted if the output\nfrom <a class=\"link\" href=\"#pgp-decrypt-command\" title=\"3.182.\u00a0pgp_decrypt_command\">$pgp_decrypt_command<\/a> contains the text.  This is used to\nprotect against a spoofed encrypted message, with multipart\/encrypted\nheaders but containing a block that is not actually encrypted.\n(e.g. simply signed and ascii armored text).\n<\/p><p>\nNote that if <a class=\"link\" href=\"#pgp-check-gpg-decrypt-status-fd\" title=\"3.179.\u00a0pgp_check_gpg_decrypt_status_fd\">$pgp_check_gpg_decrypt_status_fd<\/a> is set, this variable\nis ignored.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.184.\u00a0pgp_default_key<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis is the default key-pair to use for PGP operations.  It will be\nused for encryption (see <a class=\"link\" href=\"#postpone-encrypt\" title=\"3.224.\u00a0postpone_encrypt\">$postpone_encrypt<\/a> and <a class=\"link\" href=\"#pgp-self-encrypt\" title=\"3.199.\u00a0pgp_self_encrypt\">$pgp_self_encrypt<\/a>).\n<\/p><p>\nIt will also be used for signing unless <a class=\"link\" href=\"#pgp-sign-as\" title=\"3.201.\u00a0pgp_sign_as\">$pgp_sign_as<\/a> is set.\n<\/p><p>\nThe (now deprecated) <em>pgp_self_encrypt_as<\/em> is an alias for this\nvariable, and should no longer be used.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.185.\u00a0pgp_encrypt_only_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to encrypt a body part without signing it.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.186.\u00a0pgp_encrypt_sign_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to both sign and encrypt a body part.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.187.\u00a0pgp_entry_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%4n\u00a0%t%f\u00a0%4l\/0x%k\u00a0%-4a\u00a0%2c\u00a0%u<\/code>\u201d<\/p><\/div><p>\nThis variable allows you to customize the PGP key selection menu to\nyour personal taste. This string is similar to <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>, but\nhas its own set of <code class=\"literal\">printf(3)<\/code>-like sequences:\n<\/p><p>\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.188.\u00a0pgp_export_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to export a public key from the user's\nkey ring.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.189.\u00a0pgp_getkeys_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is invoked whenever Mutt needs to fetch the public key associated with\nan email address.  Of the sequences supported by <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a>, %r is\nthe only <code class=\"literal\">printf(3)<\/code>-like sequence used with this format.  Note that\nin this case, %r expands to the email address, not the public key ID (the key ID is\nunknown, which is why Mutt is invoking this command).\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.190.\u00a0pgp_good_sign<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf you assign a text to this variable, then a PGP signature is only\nconsidered verified if the output from <a class=\"link\" href=\"#pgp-verify-command\" title=\"3.207.\u00a0pgp_verify_command\">$pgp_verify_command<\/a> contains\nthe text. Use this variable if the exit code from the command is 0\neven for bad signatures.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.191.\u00a0pgp_ignore_subkeys<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nSetting this variable will cause Mutt to ignore OpenPGP subkeys. Instead,\nthe principal key will inherit the subkeys' capabilities.  <em>Unset<\/em> this\nif you want to play interesting key selection games.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.192.\u00a0pgp_import_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to import a key from a message into\nthe user's public key ring.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.193.\u00a0pgp_list_pubring_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to list the public key ring's contents.  The\noutput format must be analogous to the one used by\n<\/p><pre class=\"screen\">\ngpg --list-keys --with-colons --with-fingerprint\n<\/pre><p>\nThis format is also generated by the <code class=\"literal\">pgpring<\/code> utility which comes\nwith mutt.\n<\/p><p>\nNote: gpg's <code class=\"literal\">fixed-list-mode<\/code> option should not be used.  It\nproduces a different date format which may result in mutt showing\nincorrect key generation dates.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.194.\u00a0pgp_list_secring_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to list the secret key ring's contents.  The\noutput format must be analogous to the one used by:\n<\/p><pre class=\"screen\">\ngpg --list-keys --with-colons --with-fingerprint\n<\/pre><p>\nThis format is also generated by the <code class=\"literal\">pgpring<\/code> utility which comes\nwith mutt.\n<\/p><p>\nNote: gpg's <code class=\"literal\">fixed-list-mode<\/code> option should not be used.  It\nproduces a different date format which may result in mutt showing\nincorrect key generation dates.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.195.\u00a0pgp_long_ids<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, use 64 bit PGP key IDs, if <em>unset<\/em> use the normal 32 bit key IDs.\nNOTE: Internally, Mutt has transitioned to using fingerprints (or long key IDs\nas a fallback).  This option now only controls the display of key IDs\nin the key selection menu and a few other places.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.196.\u00a0pgp_mime_auto<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nThis option controls whether Mutt will prompt you for\nautomatically sending a (signed\/encrypted) message using\nPGP\/MIME when inline (traditional) fails (for any reason).\n<\/p><p>\nAlso note that using the old-style PGP message format is <strong>strongly<\/strong>\n<strong>deprecated<\/strong>.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.197.\u00a0pgp_replyinline<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nSetting this variable will cause Mutt to always attempt to\ncreate an inline (traditional) message when replying to a\nmessage which is PGP encrypted\/signed inline.  This can be\noverridden by use of the pgp menu, when inline is not\nrequired.  This option does not automatically detect if the\n(replied-to) message is inline; instead it relies on Mutt\ninternals for previously checked\/flagged messages.\n<\/p><p>\nNote that Mutt might automatically use PGP\/MIME for messages\nwhich consist of more than a single MIME part.  Mutt can be\nconfigured to ask before sending PGP\/MIME messages when inline\n(traditional) would not work.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#pgp-mime-auto\" title=\"3.196.\u00a0pgp_mime_auto\">$pgp_mime_auto<\/a> variable.\n<\/p><p>\nAlso note that using the old-style PGP message format is <strong>strongly<\/strong>\n<strong>deprecated<\/strong>.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.198.\u00a0pgp_retainable_sigs<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, signed and encrypted messages will consist of nested\n<code class=\"literal\">multipart\/signed<\/code> and <code class=\"literal\">multipart\/encrypted<\/code> body parts.\n<\/p><p>\nThis is useful for applications like encrypted and signed mailing\nlists, where the outer layer (<code class=\"literal\">multipart\/encrypted<\/code>) can be easily\nremoved, while the inner <code class=\"literal\">multipart\/signed<\/code> part is retained.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.199.\u00a0pgp_self_encrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, PGP encrypted messages will also be encrypted\nusing the key in <a class=\"link\" href=\"#pgp-default-key\" title=\"3.184.\u00a0pgp_default_key\">$pgp_default_key<\/a>.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.200.\u00a0pgp_show_unusable<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, mutt will display non-usable keys on the PGP key selection\nmenu.  This includes keys which have been revoked, have expired, or\nhave been marked as \u201cdisabled\u201d by the user.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.201.\u00a0pgp_sign_as<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf you have a different key pair to use for signing, you should\nset this to the signing key.  Most people will only need to set\n<a class=\"link\" href=\"#pgp-default-key\" title=\"3.184.\u00a0pgp_default_key\">$pgp_default_key<\/a>.  It is recommended that you use the keyid form\nto specify your key (e.g. <code class=\"literal\">0x00112233<\/code>).\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.202.\u00a0pgp_sign_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to create the detached PGP signature for a\n<code class=\"literal\">multipart\/signed<\/code> PGP\/MIME body part.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.203.\u00a0pgp_sort_keys<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0sort\u00a0order<br \/>Default:\u00a0address<\/p><\/div><p>\nSpecifies how the entries in the pgp menu are sorted. The\nfollowing are legal values:\n<\/p><p>\nIf you prefer reverse order of the above values, prefix it with\n\u201creverse-\u201d.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.204.\u00a0pgp_strict_enc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will automatically encode PGP\/MIME signed messages as\nquoted-printable.  Please note that unsetting this variable may\nlead to problems with non-verifyable PGP signatures, so only change\nthis if you know what you are doing.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.205.\u00a0pgp_timeout<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a0300<\/p><\/div><p>\nThe number of seconds after which a cached passphrase will expire if\nnot used.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.206.\u00a0pgp_use_gpg_agent<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, mutt will use a possibly-running <code class=\"literal\">gpg-agent(1)<\/code> process.\nNote that as of version 2.1, GnuPG no longer exports GPG_AGENT_INFO, so\nmutt no longer verifies if the agent is running.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.207.\u00a0pgp_verify_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to verify PGP signatures.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.208.\u00a0pgp_verify_key_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to verify key information from the key selection\nmenu.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#pgp-decode-command\" title=\"3.181.\u00a0pgp_decode_command\">$pgp_decode_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(PGP only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.209.\u00a0pipe_decode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nUsed in connection with the <code class=\"literal\">&lt;pipe-message&gt;<\/code> command.  When <em>unset<\/em>,\nMutt will pipe the messages without any preprocessing. When <em>set<\/em>, Mutt\nwill weed headers and will attempt to decode the messages\nfirst.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.210.\u00a0pipe_sep<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">\\n<\/code>\u201d<\/p><\/div><p>\nThe separator to add between messages when piping a list of tagged\nmessages to an external Unix command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.211.\u00a0pipe_split<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nUsed in connection with the <code class=\"literal\">&lt;pipe-message&gt;<\/code> function following\n<code class=\"literal\">&lt;tag-prefix&gt;<\/code>.  If this variable is <em>unset<\/em>, when piping a list of\ntagged messages Mutt will concatenate the messages and will pipe them\nall concatenated.  When <em>set<\/em>, Mutt will pipe the messages one by one.\nIn both cases the messages are piped in the current sorted order,\nand the <a class=\"link\" href=\"#pipe-sep\" title=\"3.210.\u00a0pipe_sep\">$pipe_sep<\/a> separator is added after each message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.212.\u00a0pop_auth_try_all<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will try all available authentication methods.\nWhen <em>unset<\/em>, Mutt will only fall back to other authentication\nmethods if the previous methods are unavailable. If a method is\navailable but authentication fails, Mutt will not connect to the POP server.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.213.\u00a0pop_authenticators<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis is a colon-delimited list of authentication methods mutt may\nattempt to use to log in to an POP server, in the order mutt should\ntry them.  Authentication methods are either \u201cuser\u201d, \u201capop\u201d or any\nSASL mechanism, e.g. \u201cdigest-md5\u201d, \u201cgssapi\u201d or \u201ccram-md5\u201d.\nThis option is case-insensitive. If this option is <em>unset<\/em>\n(the default) mutt will try all available methods, in order from\nmost-secure to least-secure.\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset pop_authenticators=\"digest-md5:apop:user\"\n<\/pre><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.214.\u00a0pop_checkinterval<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a060<\/p><\/div><p>\nThis variable configures how often (in seconds) mutt should look for\nnew mail in the currently selected mailbox if it is a POP mailbox.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.215.\u00a0pop_delete<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-no<\/p><\/div><p>\nIf <em>set<\/em>, Mutt will delete successfully downloaded messages from the POP\nserver when using the <code class=\"literal\"><a class=\"link\" href=\"#fetch-mail\">&lt;fetch-mail&gt;<\/a><\/code> function.  When <em>unset<\/em>, Mutt will\ndownload messages but also leave them on the POP server.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.216.\u00a0pop_host<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThe name of your POP server for the <code class=\"literal\"><a class=\"link\" href=\"#fetch-mail\">&lt;fetch-mail&gt;<\/a><\/code> function.  You\ncan also specify an alternative port, username and password, i.e.:\n<\/p><pre class=\"screen\">\n[pop[s]:\/\/][username[:password]@]popserver[:port]\n<\/pre><p>\nwhere \u201c[...]\u201d denotes an optional part.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.217.\u00a0pop_last<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf this variable is <em>set<\/em>, mutt will try to use the \u201c<code class=\"literal\">LAST<\/code>\u201d POP command\nfor retrieving only unread messages from the POP server when using\nthe <code class=\"literal\"><a class=\"link\" href=\"#fetch-mail\">&lt;fetch-mail&gt;<\/a><\/code> function.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.218.\u00a0pop_pass<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSpecifies the password for your POP account.  If <em>unset<\/em>, Mutt will\nprompt you for your password when you open a POP mailbox.\n<\/p><p>\n<strong>Warning<\/strong>: you should only use this option when you are on a\nfairly secure machine, because the superuser can read your muttrc\neven if you are the only one who can read the file.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.219.\u00a0pop_reconnect<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nControls whether or not Mutt will try to reconnect to the POP server if\nthe connection is lost.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.220.\u00a0pop_user<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nYour login name on the POP server.\n<\/p><p>\nThis variable defaults to your user name on the local machine.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.221.\u00a0post_indent_string<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSimilar to the <a class=\"link\" href=\"#attribution\" title=\"3.18.\u00a0attribution\">$attribution<\/a> variable, Mutt will append this\nstring after the inclusion of a message which is being replied to.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.222.\u00a0postpone<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nControls whether or not messages are saved in the <a class=\"link\" href=\"#postponed\" title=\"3.223.\u00a0postponed\">$postponed<\/a>\nmailbox when you elect not to send immediately.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#recall\" title=\"3.239.\u00a0recall\">$recall<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.223.\u00a0postponed<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/postponed<\/code>\u201d<\/p><\/div><p>\nMutt allows you to indefinitely \u201c<a class=\"link\" href=\"#postpone\" title=\"3.222.\u00a0postpone\">postpone<\/a> sending a message\u201d which\nyou are editing.  When you choose to postpone a message, Mutt saves it\nin the mailbox specified by this variable.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#postpone\" title=\"3.222.\u00a0postpone\">$postpone<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.224.\u00a0postpone_encrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, postponed messages that are marked for encryption will be\nself-encrypted.  Mutt will first try to encrypt using the value specified\nin <a class=\"link\" href=\"#pgp-default-key\" title=\"3.184.\u00a0pgp_default_key\">$pgp_default_key<\/a> or <a class=\"link\" href=\"#smime-default-key\" title=\"3.290.\u00a0smime_default_key\">$smime_default_key<\/a>.  If those are not\nset, it will try the deprecated <a class=\"link\" href=\"#postpone-encrypt-as\" title=\"3.225.\u00a0postpone_encrypt_as\">$postpone_encrypt_as<\/a>.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.225.\u00a0postpone_encrypt_as<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis is a deprecated fall-back variable for <a class=\"link\" href=\"#postpone-encrypt\" title=\"3.224.\u00a0postpone_encrypt\">$postpone_encrypt<\/a>.\nPlease use <a class=\"link\" href=\"#pgp-default-key\" title=\"3.184.\u00a0pgp_default_key\">$pgp_default_key<\/a> or <a class=\"link\" href=\"#smime-default-key\" title=\"3.290.\u00a0smime_default_key\">$smime_default_key<\/a>.\n(Crypto only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.226.\u00a0preconnect<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf <em>set<\/em>, a shell command to be executed if mutt fails to establish\na connection to the server. This is useful for setting up secure\nconnections, e.g. with <code class=\"literal\">ssh(1)<\/code>. If the command returns a  nonzero\nstatus, mutt gives up opening the server. Example:\n<\/p><pre class=\"screen\">\nset preconnect=\"ssh -f -q -L 1234:mailhost.net:143 mailhost.net \\\nsleep 20 &lt; \/dev\/null &gt; \/dev\/null\"\n<\/pre><p>\nMailbox \u201cfoo\u201d on \u201cmailhost.net\u201d can now be reached\nas \u201c{localhost:1234}foo\u201d.\n<\/p><p>\nNote: For this example to work, you must be able to log in to the\nremote machine without having to enter a password.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.227.\u00a0print<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-no<\/p><\/div><p>\nControls whether or not Mutt really prints messages.\nThis is set to \u201cask-no\u201d by default, because some people\naccidentally hit \u201cp\u201d often.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.228.\u00a0print_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">lpr<\/code>\u201d<\/p><\/div><p>\nThis specifies the command pipe that should be used to print messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.229.\u00a0print_decode<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nUsed in connection with the <code class=\"literal\">&lt;print-message&gt;<\/code> command.  If this\noption is <em>set<\/em>, the message is decoded before it is passed to the\nexternal command specified by <a class=\"link\" href=\"#print-command\" title=\"3.228.\u00a0print_command\">$print_command<\/a>.  If this option\nis <em>unset<\/em>, no processing will be applied to the message when\nprinting it.  The latter setting may be useful if you are using\nsome advanced printer filter which is able to properly format\ne-mail messages for printing.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.230.\u00a0print_split<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nUsed in connection with the <code class=\"literal\">&lt;print-message&gt;<\/code> command.  If this option\nis <em>set<\/em>, the command specified by <a class=\"link\" href=\"#print-command\" title=\"3.228.\u00a0print_command\">$print_command<\/a> is executed once for\neach message which is to be printed.  If this option is <em>unset<\/em>,\nthe command specified by <a class=\"link\" href=\"#print-command\" title=\"3.228.\u00a0print_command\">$print_command<\/a> is executed only once, and\nall the messages are concatenated, with a form feed as the message\nseparator.\n<\/p><p>\nThose who use the <code class=\"literal\">enscript<\/code>(1) program's mail-printing mode will\nmost likely want to <em>set<\/em> this option.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.231.\u00a0prompt_after<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf you use an <em>external<\/em> <a class=\"link\" href=\"#pager\" title=\"3.171.\u00a0pager\">$pager<\/a>, setting this variable will\ncause Mutt to prompt you for a command when the pager exits rather\nthan returning to the index menu.  If <em>unset<\/em>, Mutt will return to the\nindex menu when the external pager exits.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.232.\u00a0query_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis specifies the command Mutt will use to make external address\nqueries.  The string may contain a \u201c%s\u201d, which will be substituted\nwith the query string the user types.  Mutt will add quotes around the\nstring substituted for \u201c%s\u201d automatically according to shell quoting\nrules, so you should avoid adding your own.  If no \u201c%s\u201d is found in\nthe string, Mutt will append the user's query to the end of the string.\nSee \u201c<a class=\"link\" href=\"#query\" title=\"8.\u00a0External Address Queries\">query<\/a>\u201d for more information.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.233.\u00a0query_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%4c\u00a0%t\u00a0%-25.25a\u00a0%-25.25n\u00a0%?e?(%e)?<\/code>\u201d<\/p><\/div><p>\nThis variable describes the format of the \u201cquery\u201d menu. The\nfollowing <code class=\"literal\">printf(3)<\/code>-style sequences are understood:\n<\/p><p>\nFor an explanation of \u201csoft-fill\u201d, see the <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a> documentation.\n<\/p><p>\n* = can be optionally printed if nonzero, see the <a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a> documentation.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.234.\u00a0quit<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable controls whether \u201cquit\u201d and \u201cexit\u201d actually quit\nfrom mutt.  If this option is <em>set<\/em>, they do quit, if it is <em>unset<\/em>, they\nhave no effect, and if it is set to <em>ask-yes<\/em> or <em>ask-no<\/em>, you are\nprompted for confirmation when you try to quit.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.235.\u00a0quote_regexp<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0\u201c<code class=\"literal\">^([\u00a0\\t]*[|&gt;:}#])+<\/code>\u201d<\/p><\/div><p>\nA regular expression used in the internal pager to determine quoted\nsections of text in the body of a message. Quoted text may be filtered\nout using the <code class=\"literal\">&lt;toggle-quoted&gt;<\/code> command, or colored according to the\n\u201ccolor quoted\u201d family of directives.\n<\/p><p>\nHigher levels of quoting may be colored differently (\u201ccolor quoted1\u201d,\n\u201ccolor quoted2\u201d, etc.). The quoting level is determined by removing\nthe last character from the matched text and recursively reapplying\nthe regular expression until it fails to produce a match.\n<\/p><p>\nMatch detection may be overridden by the <a class=\"link\" href=\"#smileys\" title=\"3.284.\u00a0smileys\">$smileys<\/a> regular expression.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.236.\u00a0read_inc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a010<\/p><\/div><p>\nIf set to a value greater than 0, Mutt will display which message it\nis currently on when reading a mailbox or when performing search actions\nsuch as search and limit. The message is printed after\nthis many messages have been read or searched (e.g., if set to 25, Mutt will\nprint a message when it is at message 25, and then again when it gets\nto message 50).  This variable is meant to indicate progress when\nreading or searching large mailboxes which may take some time.\nWhen set to 0, only a single message will appear before the reading\nthe mailbox.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#write-inc\" title=\"3.367.\u00a0write_inc\">$write_inc<\/a>, <a class=\"link\" href=\"#net-inc\" title=\"3.169.\u00a0net_inc\">$net_inc<\/a> and <a class=\"link\" href=\"#time-inc\" title=\"3.342.\u00a0time_inc\">$time_inc<\/a> variables and the\n\u201c<a class=\"link\" href=\"#tuning\" title=\"Chapter\u00a08.\u00a0Performance Tuning\">tuning<\/a>\u201d section of the manual for performance considerations.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.237.\u00a0read_only<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, all folders are opened in read-only mode.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.238.\u00a0realname<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable specifies what \u201creal\u201d or \u201cpersonal\u201d name should be used\nwhen sending messages.\n<\/p><p>\nBy default, this is the GECOS field from <code class=\"literal\">\/etc\/passwd<\/code>.  Note that this\nvariable will <em>not<\/em> be used when the user has set a real name\nin the <a class=\"link\" href=\"#from\" title=\"3.86.\u00a0from\">$from<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.239.\u00a0recall<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nControls whether or not Mutt recalls postponed messages\nwhen composing a new message.\n<\/p><p>\nSetting this variable to <em>yes<\/em> is not generally useful, and thus not\nrecommended.  Note that the <code class=\"literal\">&lt;recall-message&gt;<\/code> function can be used\nto manually recall postponed messages.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#postponed\" title=\"3.223.\u00a0postponed\">$postponed<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.240.\u00a0record<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/sent<\/code>\u201d<\/p><\/div><p>\nThis specifies the file into which your outgoing messages should be\nappended.  (This is meant as the primary method for saving a copy of\nyour messages, but another way to do this is using the \u201c<a class=\"link\" href=\"#my-hdr\" title=\"15.\u00a0User-Defined Headers\">my_hdr<\/a>\u201d\ncommand to create a \u201cBcc:\u201d field with your email address in it.)\n<\/p><p>\nThe value of <em><a class=\"link\" href=\"#record\" title=\"3.240.\u00a0record\">$record<\/a><\/em> is overridden by the <a class=\"link\" href=\"#force-name\" title=\"3.78.\u00a0force_name\">$force_name<\/a> and\n<a class=\"link\" href=\"#save-name\" title=\"3.257.\u00a0save_name\">$save_name<\/a> variables, and the \u201c<a class=\"link\" href=\"#fcc-hook\" title=\"17.\u00a0Specify Default Fcc: Mailbox When Composing\">fcc-hook<\/a>\u201d command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.241.\u00a0reflow_space_quotes<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis option controls how quotes from format=flowed messages are displayed\nin the pager and when replying (with <a class=\"link\" href=\"#text-flowed\" title=\"3.338.\u00a0text_flowed\">$text_flowed<\/a> <em>unset<\/em>).\nWhen set, this option adds spaces after each level of quote marks, turning\n\"&gt;&gt;&gt;foo\" into \"&gt; &gt; &gt; foo\".\n<\/p><p>\n<strong>Note:<\/strong> If <a class=\"link\" href=\"#reflow-text\" title=\"3.242.\u00a0reflow_text\">$reflow_text<\/a> is <em>unset<\/em>, this option has no effect.\nAlso, this option does not affect replies when <a class=\"link\" href=\"#text-flowed\" title=\"3.338.\u00a0text_flowed\">$text_flowed<\/a> is <em>set<\/em>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.242.\u00a0reflow_text<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will reformat paragraphs in text\/plain\nparts marked format=flowed.  If <em>unset<\/em>, Mutt will display paragraphs\nunaltered from how they appear in the message body.  See RFC3676 for\ndetails on the <em>format=flowed<\/em> format.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#reflow-wrap\" title=\"3.243.\u00a0reflow_wrap\">$reflow_wrap<\/a>, and <a class=\"link\" href=\"#wrap\" title=\"3.362.\u00a0wrap\">$wrap<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.243.\u00a0reflow_wrap<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a078<\/p><\/div><p>\nThis variable controls the maximum paragraph width when reformatting text\/plain\nparts when <a class=\"link\" href=\"#reflow-text\" title=\"3.242.\u00a0reflow_text\">$reflow_text<\/a> is <em>set<\/em>.  When the value is 0, paragraphs will\nbe wrapped at the terminal's right margin.  A positive value sets the\nparagraph width relative to the left margin.  A negative value set the\nparagraph width relative to the right margin.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#wrap\" title=\"3.362.\u00a0wrap\">$wrap<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.244.\u00a0reply_regexp<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0\u201c<code class=\"literal\">^(re([\\[0-9\\]+])*|aw):[\u00a0\\t]*<\/code>\u201d<\/p><\/div><p>\nA regular expression used to recognize reply messages when threading\nand replying. The default value corresponds to the English \"Re:\" and\nthe German \"Aw:\".\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.245.\u00a0reply_self<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>unset<\/em> and you are replying to a message sent by you, Mutt will\nassume that you want to reply to the recipients of that message rather\nthan to yourself.\n<\/p><p>\nAlso see the \u201c<a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a>\u201d command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.246.\u00a0reply_to<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0ask-yes<\/p><\/div><p>\nIf <em>set<\/em>, when replying to a message, Mutt will use the address listed\nin the Reply-to: header as the recipient of the reply.  If <em>unset<\/em>,\nit will use the address in the From: header field instead.  This\noption is useful for reading a mailing list that sets the Reply-To:\nheader field to the list address and you want to send a private\nmessage to the author of a message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.247.\u00a0resolve<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, the cursor will be automatically advanced to the next\n(possibly undeleted) message whenever a command that modifies the\ncurrent message is executed.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.248.\u00a0resume_draft_files<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, draft files (specified by <code class=\"literal\">-H<\/code> on the command\nline) are processed similarly to when resuming a postponed\nmessage.  Recipients are not prompted for; send-hooks are not\nevaluated; no alias expansion takes place; user-defined headers\nand signatures are not added to the message.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.249.\u00a0resume_edited_draft_files<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, draft files previously edited (via <code class=\"literal\">-E -H<\/code> on\nthe command line) will have <a class=\"link\" href=\"#resume-draft-files\" title=\"3.248.\u00a0resume_draft_files\">$resume_draft_files<\/a> automatically\nset when they are used as a draft file again.\n<\/p><p>\nThe first time a draft file is saved, mutt will add a header,\nX-Mutt-Resume-Draft to the saved file.  The next time the draft\nfile is read in, if mutt sees the header, it will set\n<a class=\"link\" href=\"#resume-draft-files\" title=\"3.248.\u00a0resume_draft_files\">$resume_draft_files<\/a>.\n<\/p><p>\nThis option is designed to prevent multiple signatures,\nuser-defined headers, and other processing effects from being\nmade multiple times to the draft file.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.250.\u00a0reverse_alias<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable controls whether or not Mutt will display the \u201cpersonal\u201d\nname from your aliases in the index menu if it finds an alias that\nmatches the message's sender.  For example, if you have the following\nalias:\n<\/p><pre class=\"screen\">\nalias juser abd30425@somewhere.net (Joe User)\n<\/pre><p>\nand then you receive mail which contains the following header:\n<\/p><pre class=\"screen\">\nFrom: abd30425@somewhere.net\n<\/pre><p>\nIt would be displayed in the index menu as \u201cJoe User\u201d instead of\n\u201cabd30425@somewhere.net.\u201d  This is useful when the person's e-mail\naddress is not human friendly.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.251.\u00a0reverse_name<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIt may sometimes arrive that you receive mail to a certain machine,\nmove the messages to another machine, and reply to some the messages\nfrom there.  If this variable is <em>set<\/em>, the default <em>From:<\/em> line of\nthe reply messages is built using the address where you received the\nmessages you are replying to <strong>if<\/strong> that address matches your\n\u201c<a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a>\u201d.  If the variable is <em>unset<\/em>, or the address that would be\nused doesn't match your \u201c<a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a>\u201d, the <em>From:<\/em> line will use\nyour address on the current machine.\n<\/p><p>\nAlso see the \u201c<a class=\"link\" href=\"#alternates\" title=\"11.\u00a0Alternative Addresses\">alternates<\/a>\u201d command.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.252.\u00a0reverse_realname<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable fine-tunes the behavior of the <a class=\"link\" href=\"#reverse-name\" title=\"3.251.\u00a0reverse_name\">$reverse_name<\/a> feature.\nWhen it is <em>set<\/em>, mutt will use the address from incoming messages as-is,\npossibly including eventual real names.  When it is <em>unset<\/em>, mutt will\noverride any such real names with the setting of the <a class=\"link\" href=\"#realname\" title=\"3.238.\u00a0realname\">$realname<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.253.\u00a0rfc2047_parameters<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen this variable is <em>set<\/em>, Mutt will decode RFC2047-encoded MIME\nparameters. You want to set this variable when mutt suggests you\nto save attachments to files named like:\n<\/p><pre class=\"screen\">\n=?iso-8859-1?Q?file=5F=E4=5F991116=2Ezip?=\n<\/pre><p>\nWhen this variable is <em>set<\/em> interactively, the change won't be\nactive until you change folders.\n<\/p><p>\nNote that this use of RFC2047's encoding is explicitly\nprohibited by the standard, but nevertheless encountered in the\nwild.\n<\/p><p>\nAlso note that setting this parameter will <em>not<\/em> have the effect\nthat mutt <em>generates<\/em> this kind of encoding.  Instead, mutt will\nunconditionally use the encoding specified in RFC2231.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.254.\u00a0save_address<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, mutt will take the sender's full address when choosing a\ndefault folder for saving a mail. If <a class=\"link\" href=\"#save-name\" title=\"3.257.\u00a0save_name\">$save_name<\/a> or <a class=\"link\" href=\"#force-name\" title=\"3.78.\u00a0force_name\">$force_name<\/a>\nis <em>set<\/em> too, the selection of the Fcc folder will be changed as well.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.255.\u00a0save_empty<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>unset<\/em>, mailboxes which contain no saved messages will be removed\nwhen closed (the exception is <a class=\"link\" href=\"#spoolfile\" title=\"3.317.\u00a0spoolfile\">$spoolfile<\/a> which is never removed).\nIf <em>set<\/em>, mailboxes are never removed.\n<\/p><p>\n<strong>Note:<\/strong> This only applies to mbox and MMDF folders, Mutt does not\ndelete MH and Maildir directories.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.256.\u00a0save_history<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nThis variable controls the size of the history (per category) saved in the\n<a class=\"link\" href=\"#history-file\" title=\"3.102.\u00a0history_file\">$history_file<\/a> file.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.257.\u00a0save_name<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable controls how copies of outgoing messages are saved.\nWhen <em>set<\/em>, a check is made to see if a mailbox specified by the\nrecipient address exists (this is done by searching for a mailbox in\nthe <a class=\"link\" href=\"#folder\" title=\"3.75.\u00a0folder\">$folder<\/a> directory with the <em>username<\/em> part of the\nrecipient address).  If the mailbox exists, the outgoing message will\nbe saved to that mailbox, otherwise the message is saved to the\n<a class=\"link\" href=\"#record\" title=\"3.240.\u00a0record\">$record<\/a> mailbox.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#force-name\" title=\"3.78.\u00a0force_name\">$force_name<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.258.\u00a0score<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen this variable is <em>unset<\/em>, scoring is turned off.  This can\nbe useful to selectively disable scoring for certain folders when the\n<a class=\"link\" href=\"#score-threshold-delete\" title=\"3.259.\u00a0score_threshold_delete\">$score_threshold_delete<\/a> variable and related are used.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.259.\u00a0score_threshold_delete<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a0-1<\/p><\/div><p>\nMessages which have been assigned a score equal to or lower than the value\nof this variable are automatically marked for deletion by mutt.  Since\nmutt scores are always greater than or equal to zero, the default setting\nof this variable will never mark a message for deletion.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.260.\u00a0score_threshold_flag<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a09999<\/p><\/div><p>\nMessages which have been assigned a score greater than or equal to this\nvariable's value are automatically marked \"flagged\".\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.261.\u00a0score_threshold_read<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a0-1<\/p><\/div><p>\nMessages which have been assigned a score equal to or lower than the value\nof this variable are automatically marked as read by mutt.  Since\nmutt scores are always greater than or equal to zero, the default setting\nof this variable will never mark a message read.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.262.\u00a0search_context<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nFor the pager, this variable specifies the number of lines shown\nbefore search results. By default, search results will be top-aligned.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.263.\u00a0send_charset<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">us-ascii:iso-8859-1:utf-8<\/code>\u201d<\/p><\/div><p>\nA colon-delimited list of character sets for outgoing messages. Mutt will use the\nfirst character set into which the text can be converted exactly.\nIf your <a class=\"link\" href=\"#charset\" title=\"3.30.\u00a0charset\">$charset<\/a> is not \u201ciso-8859-1\u201d and recipients may not\nunderstand \u201cUTF-8\u201d, it is advisable to include in the list an\nappropriate widely used standard character set (such as\n\u201ciso-8859-2\u201d, \u201ckoi8-r\u201d or \u201ciso-2022-jp\u201d) either instead of or after\n\u201ciso-8859-1\u201d.\n<\/p><p>\nIn case the text cannot be converted into one of these exactly,\nmutt uses <a class=\"link\" href=\"#charset\" title=\"3.30.\u00a0charset\">$charset<\/a> as a fallback.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.264.\u00a0sendmail<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">\/usr\/sbin\/sendmail\u00a0-oem\u00a0-oi<\/code>\u201d<\/p><\/div><p>\nSpecifies the program and arguments used to deliver mail sent by Mutt.\nMutt expects that the specified program interprets additional\narguments as recipient addresses.  Mutt appends all recipients after\nadding a <code class=\"literal\">--<\/code> delimiter (if not already present).  Additional\nflags, such as for <a class=\"link\" href=\"#use-8bitmime\" title=\"3.353.\u00a0use_8bitmime\">$use_8bitmime<\/a>, <a class=\"link\" href=\"#use-envelope-from\" title=\"3.355.\u00a0use_envelope_from\">$use_envelope_from<\/a>,\n<a class=\"link\" href=\"#dsn-notify\" title=\"3.61.\u00a0dsn_notify\">$dsn_notify<\/a>, or <a class=\"link\" href=\"#dsn-return\" title=\"3.62.\u00a0dsn_return\">$dsn_return<\/a> will be added before the delimiter.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.265.\u00a0sendmail_wait<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nSpecifies the number of seconds to wait for the <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a> process\nto finish before giving up and putting delivery in the background.\n<\/p><p>\nMutt interprets the value of this variable as follows:\n<\/p><p>\nNote that if you specify a value other than 0, the output of the child\nprocess will be put in a temporary file.  If there is some error, you\nwill be informed as to where to find the output.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.266.\u00a0shell<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nCommand to use when spawning a subshell.  By default, the user's login\nshell from <code class=\"literal\">\/etc\/passwd<\/code> is used.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.267.\u00a0sidebar_delim_chars<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">\/.<\/code>\u201d<\/p><\/div><p>\nThis contains the list of characters which you would like to treat\nas folder separators for displaying paths in the sidebar.\n<\/p><p>\nLocal mail is often arranged in directories: `dir1\/dir2\/mailbox'.\n<\/p><pre class=\"screen\">\nset sidebar_delim_chars='\/'\n<\/pre><p>\nIMAP mailboxes are often named: `folder1.folder2.mailbox'.\n<\/p><pre class=\"screen\">\nset sidebar_delim_chars='.'\n<\/pre><p>\n<strong>See also:<\/strong> <a class=\"link\" href=\"#sidebar-short-path\" title=\"3.274.\u00a0sidebar_short_path\">$sidebar_short_path<\/a>, <a class=\"link\" href=\"#sidebar-folder-indent\" title=\"3.269.\u00a0sidebar_folder_indent\">$sidebar_folder_indent<\/a>, <a class=\"link\" href=\"#sidebar-indent-string\" title=\"3.271.\u00a0sidebar_indent_string\">$sidebar_indent_string<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.268.\u00a0sidebar_divider_char<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">|<\/code>\u201d<\/p><\/div><p>\nThis specifies the characters to be drawn between the sidebar (when\nvisible) and the other Mutt panels. ASCII and Unicode line-drawing\ncharacters are supported.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.269.\u00a0sidebar_folder_indent<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nSet this to indent mailboxes in the sidebar.\n<\/p><p>\n<strong>See also:<\/strong> <a class=\"link\" href=\"#sidebar-short-path\" title=\"3.274.\u00a0sidebar_short_path\">$sidebar_short_path<\/a>, <a class=\"link\" href=\"#sidebar-indent-string\" title=\"3.271.\u00a0sidebar_indent_string\">$sidebar_indent_string<\/a>, <a class=\"link\" href=\"#sidebar-delim-chars\" title=\"3.267.\u00a0sidebar_delim_chars\">$sidebar_delim_chars<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.270.\u00a0sidebar_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">%B%*\u00a0\u00a0%n<\/code>\u201d<\/p><\/div><p>\nThis variable allows you to customize the sidebar display. This string is\nsimilar to <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>, but has its own set of <code class=\"literal\">printf(3)<\/code>-like\nsequences:\n<\/p><p>\n* = Can be optionally printed if nonzero\n@ = Only applicable to the current folder\n<\/p><p>\nIn order to use %S, %N, %F, and %!, <a class=\"link\" href=\"#mail-check-stats\" title=\"3.135.\u00a0mail_check_stats\">$mail_check_stats<\/a> must\nbe <em>set<\/em>.  When thus set, a suggested value for this option is\n\"%B%?F? [%F]?%* %?N?%N\/?%S\".\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.271.\u00a0sidebar_indent_string<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">\u00a0\u00a0<\/code>\u201d<\/p><\/div><p>\nThis specifies the string that is used to indent mailboxes in the sidebar.\nIt defaults to two spaces.\n<\/p><p>\n<strong>See also:<\/strong> <a class=\"link\" href=\"#sidebar-short-path\" title=\"3.274.\u00a0sidebar_short_path\">$sidebar_short_path<\/a>, <a class=\"link\" href=\"#sidebar-folder-indent\" title=\"3.269.\u00a0sidebar_folder_indent\">$sidebar_folder_indent<\/a>, <a class=\"link\" href=\"#sidebar-delim-chars\" title=\"3.267.\u00a0sidebar_delim_chars\">$sidebar_delim_chars<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.272.\u00a0sidebar_new_mail_only<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen set, the sidebar will only display mailboxes containing new, or\nflagged, mail.\n<\/p><p>\n<strong>See also:<\/strong> <a class=\"link\" href=\"#sidebar-whitelist\">sidebar_whitelist<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.273.\u00a0sidebar_next_new_wrap<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen set, the <code class=\"literal\">&lt;sidebar-next-new&gt;<\/code> command will not stop and the end of\nthe list of mailboxes, but wrap around to the beginning. The\n<code class=\"literal\">&lt;sidebar-prev-new&gt;<\/code> command is similarly affected, wrapping around to\nthe end of the list.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.274.\u00a0sidebar_short_path<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nBy default the sidebar will show the mailbox's path, relative to the\n<a class=\"link\" href=\"#folder\" title=\"3.75.\u00a0folder\">$folder<\/a> variable. Setting <code class=\"literal\">sidebar_shortpath=yes<\/code> will shorten the\nnames relative to the previous name. Here's an example:\n<\/p><p>\n<strong>See also:<\/strong> <a class=\"link\" href=\"#sidebar-delim-chars\" title=\"3.267.\u00a0sidebar_delim_chars\">$sidebar_delim_chars<\/a>, <a class=\"link\" href=\"#sidebar-folder-indent\" title=\"3.269.\u00a0sidebar_folder_indent\">$sidebar_folder_indent<\/a>, <a class=\"link\" href=\"#sidebar-indent-string\" title=\"3.271.\u00a0sidebar_indent_string\">$sidebar_indent_string<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.275.\u00a0sidebar_sort_method<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0sort\u00a0order<br \/>Default:\u00a0order<\/p><\/div><p>\nSpecifies how to sort entries in the file browser.  By default, the\nentries are sorted alphabetically.  Valid values:\n<\/p><div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>alpha (alphabetically)\n<\/p><\/li><li class=\"listitem\"><p>count (all message count)\n<\/p><\/li><li class=\"listitem\"><p>flagged (flagged message count)\n<\/p><\/li><li class=\"listitem\"><p>name (alphabetically)\n<\/p><\/li><li class=\"listitem\"><p>new (unread message count)\n<\/p><\/li><li class=\"listitem\"><p>path (alphabetically)\n<\/p><\/li><li class=\"listitem\"><p>unread (unread message count)\n<\/p><\/li><li class=\"listitem\"><p>unsorted\n<\/p><\/li><\/ul><\/div><p>\nYou may optionally use the \u201creverse-\u201d prefix to specify reverse sorting\norder (example: \u201c<code class=\"literal\">set sort_browser=reverse-date<\/code>\u201d).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.276.\u00a0sidebar_visible<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis specifies whether or not to show sidebar. The sidebar shows a list of\nall your mailboxes.\n<\/p><p>\n<strong>See also:<\/strong> <a class=\"link\" href=\"#sidebar-format\" title=\"3.270.\u00a0sidebar_format\">$sidebar_format<\/a>, <a class=\"link\" href=\"#sidebar-width\" title=\"3.277.\u00a0sidebar_width\">$sidebar_width<\/a>\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.277.\u00a0sidebar_width<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a030<\/p><\/div><p>\nThis controls the width of the sidebar.  It is measured in screen columns.\nFor example: sidebar_width=20 could display 20 ASCII characters, or 10\nChinese characters.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.278.\u00a0sig_dashes<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em>, a line containing \u201c-- \u201d (note the trailing space) will be inserted before your\n<a class=\"link\" href=\"#signature\" title=\"3.280.\u00a0signature\">$signature<\/a>.  It is <strong>strongly<\/strong> recommended that you not <em>unset<\/em>\nthis variable unless your signature contains just your name.  The\nreason for this is because many software packages use \u201c-- \\n\u201d to\ndetect your signature.  For example, Mutt has the ability to highlight\nthe signature in a different color in the built-in pager.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.279.\u00a0sig_on_top<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, the signature will be included before any quoted or forwarded\ntext.  It is <strong>strongly<\/strong> recommended that you do not set this variable\nunless you really know what you are doing, and are prepared to take\nsome heat from netiquette guardians.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.280.\u00a0signature<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0\u201c<code class=\"literal\">~\/.signature<\/code>\u201d<\/p><\/div><p>\nSpecifies the filename of your signature, which is appended to all\noutgoing messages.   If the filename ends with a pipe (\u201c|\u201d), it is\nassumed that filename is a shell command and input should be read from\nits standard output.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.281.\u00a0simple_search<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">~f\u00a0%s\u00a0|\u00a0~s\u00a0%s<\/code>\u201d<\/p><\/div><p>\nSpecifies how Mutt should expand a simple search into a real search\npattern.  A simple search is one that does not contain any of the \u201c~\u201d pattern\noperators.  See \u201c<a class=\"link\" href=\"#patterns\" title=\"3.\u00a0Patterns: Searching, Limiting and Tagging\">patterns<\/a>\u201d for more information on search patterns.\n<\/p><p>\nFor example, if you simply type \u201cjoe\u201d at a search or limit prompt, Mutt\nwill automatically expand it to the value specified by this variable by\nreplacing \u201c%s\u201d with the supplied string.\nFor the default value, \u201cjoe\u201d would be expanded to: \u201c~f joe | ~s joe\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.282.\u00a0sleep_time<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a01<\/p><\/div><p>\nSpecifies time, in seconds, to pause while displaying certain informational\nmessages, while moving from folder to folder and after expunging\nmessages from the current folder.  The default is to pause one second, so\na value of zero for this option suppresses the pause.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.283.\u00a0smart_wrap<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls the display of lines longer than the screen width in the\ninternal pager. If <em>set<\/em>, long lines are wrapped at a word boundary.  If\n<em>unset<\/em>, lines are simply wrapped at the screen edge. Also see the\n<a class=\"link\" href=\"#markers\" title=\"3.144.\u00a0markers\">$markers<\/a> variable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.284.\u00a0smileys<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0regular\u00a0expression<br \/>Default:\u00a0\u201c<code class=\"literal\">(&gt;From\u00a0)|(:[-^]?[][)(&gt;&lt;}{|\/DP])<\/code>\u201d<\/p><\/div><p>\nThe <em>pager<\/em> uses this variable to catch some common false\npositives of <a class=\"link\" href=\"#quote-regexp\" title=\"3.235.\u00a0quote_regexp\">$quote_regexp<\/a>, most notably smileys and not consider\na line quoted text if it also matches <a class=\"link\" href=\"#smileys\" title=\"3.284.\u00a0smileys\">$smileys<\/a>. This mostly\nhappens at the beginning of a line.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.285.\u00a0smime_ask_cert_label<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis flag controls whether you want to be asked to enter a label\nfor a certificate about to be added to the database or not. It is\n<em>set<\/em> by default.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.286.\u00a0smime_ca_location<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable contains the name of either a directory, or a file which\ncontains trusted certificates for use with OpenSSL.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.287.\u00a0smime_certificates<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSince for S\/MIME there is no pubring\/secring as with PGP, mutt has to handle\nstorage and retrieval of keys by itself. This is very basic right\nnow, and keys and certificates are stored in two different\ndirectories, both named as the hash-value retrieved from\nOpenSSL. There is an index file which contains mailbox-address\nkeyid pairs, and which can be manually edited. This option points to\nthe location of the certificates.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.288.\u00a0smime_decrypt_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis format string specifies a command which is used to decrypt\n<code class=\"literal\">application\/x-pkcs7-mime<\/code> attachments.\n<\/p><p>\nThe OpenSSL command formats have their own set of <code class=\"literal\">printf(3)<\/code>-like sequences\nsimilar to PGP's:\n<\/p><p>\nFor examples on how to configure these formats, see the <code class=\"literal\">smime.rc<\/code> in\nthe <code class=\"literal\">samples\/<\/code> subdirectory which has been installed on your system\nalongside the documentation.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.289.\u00a0smime_decrypt_use_default_key<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em> (default) this tells mutt to use the default key for decryption. Otherwise,\nif managing multiple certificate-key-pairs, mutt will try to use the mailbox-address\nto determine the key to use. It will ask you to supply a key, if it can't find one.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.290.\u00a0smime_default_key<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis is the default key-pair to use for S\/MIME operations, and must be\nset to the keyid (the hash-value that OpenSSL generates) to work properly.\n<\/p><p>\nIt will be used for encryption (see <a class=\"link\" href=\"#postpone-encrypt\" title=\"3.224.\u00a0postpone_encrypt\">$postpone_encrypt<\/a> and\n<a class=\"link\" href=\"#smime-self-encrypt\" title=\"3.300.\u00a0smime_self_encrypt\">$smime_self_encrypt<\/a>).\n<\/p><p>\nIt will be used for decryption unless <a class=\"link\" href=\"#smime-decrypt-use-default-key\" title=\"3.289.\u00a0smime_decrypt_use_default_key\">$smime_decrypt_use_default_key<\/a>\nis <em>unset<\/em>.\n<\/p><p>\nIt will also be used for signing unless <a class=\"link\" href=\"#smime-sign-as\" title=\"3.301.\u00a0smime_sign_as\">$smime_sign_as<\/a> is set.\n<\/p><p>\nThe (now deprecated) <em>smime_self_encrypt_as<\/em> is an alias for this\nvariable, and should no longer be used.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.291.\u00a0smime_encrypt_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to create encrypted S\/MIME messages.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.292.\u00a0smime_encrypt_with<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">aes256<\/code>\u201d<\/p><\/div><p>\nThis sets the algorithm that should be used for encryption.\nValid choices are \u201caes128\u201d, \u201caes192\u201d, \u201caes256\u201d, \u201cdes\u201d, \u201cdes3\u201d, \u201crc2-40\u201d, \u201crc2-64\u201d, \u201crc2-128\u201d.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.293.\u00a0smime_get_cert_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to extract X509 certificates from a PKCS7 structure.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.294.\u00a0smime_get_cert_email_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to extract the mail address(es) used for storing\nX509 certificates, and for verification purposes (to check whether the\ncertificate was issued for the sender's mailbox).\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.295.\u00a0smime_get_signer_cert_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to extract only the signers X509 certificate from a S\/MIME\nsignature, so that the certificate's owner may get compared to the\nemail's \u201cFrom:\u201d field.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.296.\u00a0smime_import_cert_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to import a certificate via smime_keys.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.297.\u00a0smime_is_default<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThe default behavior of mutt is to use PGP on all auto-sign\/encryption\noperations. To override and to use OpenSSL instead this must be <em>set<\/em>.\nHowever, this has no effect while replying, since mutt will automatically\nselect the same application that was used to sign\/encrypt the original\nmessage.  (Note that this variable can be overridden by unsetting <a class=\"link\" href=\"#crypt-autosmime\" title=\"3.44.\u00a0crypt_autosmime\">$crypt_autosmime<\/a>.)\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.298.\u00a0smime_keys<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSince for S\/MIME there is no pubring\/secring as with PGP, mutt has to handle\nstorage and retrieval of keys\/certs by itself. This is very basic right now,\nand stores keys and certificates in two different directories, both\nnamed as the hash-value retrieved from OpenSSL. There is an index file\nwhich contains mailbox-address keyid pair, and which can be manually\nedited. This option points to the location of the private keys.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.299.\u00a0smime_pk7out_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to extract PKCS7 structures of S\/MIME signatures,\nin order to extract the public X509 certificate(s).\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.300.\u00a0smime_self_encrypt<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, S\/MIME encrypted messages will also be encrypted\nusing the certificate in <a class=\"link\" href=\"#smime-default-key\" title=\"3.290.\u00a0smime_default_key\">$smime_default_key<\/a>.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.301.\u00a0smime_sign_as<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf you have a separate key to use for signing, you should set this\nto the signing key. Most people will only need to set <a class=\"link\" href=\"#smime-default-key\" title=\"3.290.\u00a0smime_default_key\">$smime_default_key<\/a>.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.302.\u00a0smime_sign_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to created S\/MIME signatures of type\n<code class=\"literal\">multipart\/signed<\/code>, which can be read by all mail clients.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.  NOTE: %c and %k will default\nto <a class=\"link\" href=\"#smime-sign-as\" title=\"3.301.\u00a0smime_sign_as\">$smime_sign_as<\/a> if set, otherwise <a class=\"link\" href=\"#smime-default-key\" title=\"3.290.\u00a0smime_default_key\">$smime_default_key<\/a>.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.303.\u00a0smime_sign_digest_alg<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">sha256<\/code>\u201d<\/p><\/div><p>\nThis sets the algorithm that should be used for the signature message digest.\nValid choices are \u201cmd5\u201d, \u201csha1\u201d, \u201csha224\u201d, \u201csha256\u201d, \u201csha384\u201d, \u201csha512\u201d.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.304.\u00a0smime_sign_opaque_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to created S\/MIME signatures of type\n<code class=\"literal\">application\/x-pkcs7-signature<\/code>, which can only be handled by mail\nclients supporting the S\/MIME extension.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.305.\u00a0smime_timeout<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a0300<\/p><\/div><p>\nThe number of seconds after which a cached passphrase will expire if\nnot used.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.306.\u00a0smime_verify_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to verify S\/MIME signatures of type <code class=\"literal\">multipart\/signed<\/code>.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.307.\u00a0smime_verify_opaque_command<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis command is used to verify S\/MIME signatures of type\n<code class=\"literal\">application\/x-pkcs7-mime<\/code>.\n<\/p><p>\nThis is a format string, see the <a class=\"link\" href=\"#smime-decrypt-command\" title=\"3.288.\u00a0smime_decrypt_command\">$smime_decrypt_command<\/a> command for\npossible <code class=\"literal\">printf(3)<\/code>-like sequences.\n(S\/MIME only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.308.\u00a0smtp_authenticators<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis is a colon-delimited list of authentication methods mutt may\nattempt to use to log in to an SMTP server, in the order mutt should\ntry them.  Authentication methods are any SASL mechanism, e.g.\n\u201cdigest-md5\u201d, \u201cgssapi\u201d or \u201ccram-md5\u201d.\nThis option is case-insensitive. If it is \u201cunset\u201d\n(the default) mutt will try all available methods, in order from\nmost-secure to least-secure.\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset smtp_authenticators=\"digest-md5:cram-md5\"\n<\/pre><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.309.\u00a0smtp_pass<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSpecifies the password for your SMTP account.  If <em>unset<\/em>, Mutt will\nprompt you for your password when you first send mail via SMTP.\nSee <a class=\"link\" href=\"#smtp-url\" title=\"3.310.\u00a0smtp_url\">$smtp_url<\/a> to configure mutt to send mail via SMTP.\n<\/p><p>\n<strong>Warning<\/strong>: you should only use this option when you are on a\nfairly secure machine, because the superuser can read your muttrc even\nif you are the only one who can read the file.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.310.\u00a0smtp_url<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nDefines the SMTP smarthost where sent messages should relayed for\ndelivery. This should take the form of an SMTP URL, e.g.:\n<\/p><pre class=\"screen\">\nsmtp[s]:\/\/[user[:pass]@]host[:port]\n<\/pre><p>\nwhere \u201c[...]\u201d denotes an optional part.\nSetting this variable overrides the value of the <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a>\nvariable.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.311.\u00a0sort<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0sort\u00a0order<br \/>Default:\u00a0date<\/p><\/div><p>\nSpecifies how to sort messages in the \u201cindex\u201d menu.  Valid values\nare:\n<\/p><div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>date or date-sent\n<\/p><\/li><li class=\"listitem\"><p>date-received\n<\/p><\/li><li class=\"listitem\"><p>from\n<\/p><\/li><li class=\"listitem\"><p>mailbox-order (unsorted)\n<\/p><\/li><li class=\"listitem\"><p>score\n<\/p><\/li><li class=\"listitem\"><p>size\n<\/p><\/li><li class=\"listitem\"><p>spam\n<\/p><\/li><li class=\"listitem\"><p>subject\n<\/p><\/li><li class=\"listitem\"><p>threads\n<\/p><\/li><li class=\"listitem\"><p>to\n<\/p><\/li><\/ul><\/div><p>\nYou may optionally use the \u201creverse-\u201d prefix to specify reverse sorting\norder (example: \u201c<code class=\"literal\">set sort=reverse-date-sent<\/code>\u201d).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.312.\u00a0sort_alias<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0sort\u00a0order<br \/>Default:\u00a0alias<\/p><\/div><p>\nSpecifies how the entries in the \u201calias\u201d menu are sorted.  The\nfollowing are legal values:\n<\/p><div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>address (sort alphabetically by email address)\n<\/p><\/li><li class=\"listitem\"><p>alias (sort alphabetically by alias name)\n<\/p><\/li><li class=\"listitem\"><p>unsorted (leave in order specified in .muttrc)\n<\/p><\/li><\/ul><\/div><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.313.\u00a0sort_aux<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0sort\u00a0order<br \/>Default:\u00a0date<\/p><\/div><p>\nWhen sorting by threads, this variable controls how threads are sorted\nin relation to other threads, and how the branches of the thread trees\nare sorted.  This can be set to any value that <a class=\"link\" href=\"#sort\" title=\"3.311.\u00a0sort\">$sort<\/a> can, except\n\u201cthreads\u201d (in that case, mutt will just use \u201cdate-sent\u201d).  You can also\nspecify the \u201clast-\u201d prefix in addition to the \u201creverse-\u201d prefix, but \u201clast-\u201d\nmust come after \u201creverse-\u201d.  The \u201clast-\u201d prefix causes messages to be\nsorted against its siblings by which has the last descendant, using\nthe rest of <a class=\"link\" href=\"#sort-aux\" title=\"3.313.\u00a0sort_aux\">$sort_aux<\/a> as an ordering.  For instance,\n<\/p><pre class=\"screen\">\nset sort_aux=last-date-received\n<\/pre><p>\nwould mean that if a new message is received in a\nthread, that thread becomes the last one displayed (or the first, if\nyou have \u201c<code class=\"literal\">set sort=reverse-threads<\/code>\u201d.)\n<\/p><p>\nNote: For reversed <a class=\"link\" href=\"#sort\" title=\"3.311.\u00a0sort\">$sort<\/a>\norder <a class=\"link\" href=\"#sort-aux\" title=\"3.313.\u00a0sort_aux\">$sort_aux<\/a> is reversed again (which is not the right thing to do,\nbut kept to not break any existing configuration setting).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.314.\u00a0sort_browser<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0sort\u00a0order<br \/>Default:\u00a0alpha<\/p><\/div><p>\nSpecifies how to sort entries in the file browser.  By default, the\nentries are sorted alphabetically.  Valid values:\n<\/p><div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>alpha (alphabetically)\n<\/p><\/li><li class=\"listitem\"><p>count\n<\/p><\/li><li class=\"listitem\"><p>date\n<\/p><\/li><li class=\"listitem\"><p>size\n<\/p><\/li><li class=\"listitem\"><p>unread\n<\/p><\/li><li class=\"listitem\"><p>unsorted\n<\/p><\/li><\/ul><\/div><p>\nYou may optionally use the \u201creverse-\u201d prefix to specify reverse sorting\norder (example: \u201c<code class=\"literal\">set sort_browser=reverse-date<\/code>\u201d).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.315.\u00a0sort_re<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable is only useful when sorting by threads with\n<a class=\"link\" href=\"#strict-threads\" title=\"3.336.\u00a0strict_threads\">$strict_threads<\/a> <em>unset<\/em>.  In that case, it changes the heuristic\nmutt uses to thread messages by subject.  With <a class=\"link\" href=\"#sort-re\" title=\"3.315.\u00a0sort_re\">$sort_re<\/a> <em>set<\/em>, mutt will\nonly attach a message as the child of another message by subject if\nthe subject of the child message starts with a substring matching the\nsetting of <a class=\"link\" href=\"#reply-regexp\" title=\"3.244.\u00a0reply_regexp\">$reply_regexp<\/a>.  With <a class=\"link\" href=\"#sort-re\" title=\"3.315.\u00a0sort_re\">$sort_re<\/a> <em>unset<\/em>, mutt will attach\nthe message whether or not this is the case, as long as the\nnon-<a class=\"link\" href=\"#reply-regexp\" title=\"3.244.\u00a0reply_regexp\">$reply_regexp<\/a> parts of both messages are identical.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.316.\u00a0spam_separator<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">,<\/code>\u201d<\/p><\/div><p>\nThis variable controls what happens when multiple spam headers\nare matched: if <em>unset<\/em>, each successive header will overwrite any\nprevious matches value for the spam label. If <em>set<\/em>, each successive\nmatch will append to the previous, using this variable's value as a\nseparator.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.317.\u00a0spoolfile<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf your spool mailbox is in a non-default place where Mutt cannot find\nit, you can specify its location with this variable.  Mutt will\ninitially set this variable to the value of the environment\nvariable <code class=\"literal\">$MAIL<\/code> or <code class=\"literal\">$MAILDIR<\/code> if either is defined.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.318.\u00a0ssl_ca_certificates_file<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable specifies a file containing trusted CA certificates.\nAny server certificate that is signed with one of these CA\ncertificates is also automatically accepted. (GnuTLS only)\n<\/p><p>\nExample:\n<\/p><pre class=\"screen\">\nset ssl_ca_certificates_file=\/etc\/ssl\/certs\/ca-certificates.crt\n<\/pre><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.319.\u00a0ssl_client_cert<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThe file containing a client certificate and its associated private\nkey.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.320.\u00a0ssl_force_tls<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf this variable is <em>set<\/em>, Mutt will require that all connections\nto remote servers be encrypted. Furthermore it will attempt to\nnegotiate TLS even if the server does not advertise the capability,\nsince it would otherwise have to abort the connection anyway. This\noption supersedes <a class=\"link\" href=\"#ssl-starttls\" title=\"3.322.\u00a0ssl_starttls\">$ssl_starttls<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.321.\u00a0ssl_min_dh_prime_bits<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nThis variable specifies the minimum acceptable prime size (in bits)\nfor use in any Diffie-Hellman key exchange. A value of 0 will use\nthe default from the GNUTLS library. (GnuTLS only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.322.\u00a0ssl_starttls<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0quadoption<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em> (the default), mutt will attempt to use <code class=\"literal\">STARTTLS<\/code> on servers\nadvertising the capability. When <em>unset<\/em>, mutt will not attempt to\nuse <code class=\"literal\">STARTTLS<\/code> regardless of the server's capabilities.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.323.\u00a0ssl_use_sslv2<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable specifies whether to attempt to use SSLv2 in the\nSSL authentication process. Note that SSLv2 and SSLv3 are now\nconsidered fundamentally insecure and are no longer recommended.\n(OpenSSL only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.324.\u00a0ssl_use_sslv3<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis variable specifies whether to attempt to use SSLv3 in the\nSSL authentication process. Note that SSLv2 and SSLv3 are now\nconsidered fundamentally insecure and are no longer recommended.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.325.\u00a0ssl_use_tlsv1<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable specifies whether to attempt to use TLSv1.0 in the\nSSL authentication process.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.326.\u00a0ssl_use_tlsv1_1<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable specifies whether to attempt to use TLSv1.1 in the\nSSL authentication process.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.327.\u00a0ssl_use_tlsv1_2<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nThis variable specifies whether to attempt to use TLSv1.2 in the\nSSL authentication process.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.328.\u00a0ssl_usesystemcerts<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf set to <em>yes<\/em>, mutt will use CA certificates in the\nsystem-wide certificate store when checking if a server certificate\nis signed by a trusted CA. (OpenSSL only)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.329.\u00a0ssl_verify_dates<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em> (the default), mutt will not automatically accept a server\ncertificate that is either not yet valid or already expired. You should\nonly unset this for particular known hosts, using the\n<code class=\"literal\"><a class=\"link\" href=\"#account-hook\" title=\"6.\u00a0Managing Multiple Accounts\">&lt;account-hook&gt;<\/a><\/code> function.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.330.\u00a0ssl_verify_host<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nIf <em>set<\/em> (the default), mutt will not automatically accept a server\ncertificate whose host name does not match the host used in your folder\nURL. You should only unset this for particular known hosts, using\nthe <code class=\"literal\"><a class=\"link\" href=\"#account-hook\" title=\"6.\u00a0Managing Multiple Accounts\">&lt;account-hook&gt;<\/a><\/code> function.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.331.\u00a0ssl_verify_partial_chains<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nThis option should not be changed from the default unless you understand\nwhat you are doing.\n<\/p><p>\nSetting this variable to <em>yes<\/em> will permit verifying partial\ncertification chains, i. e. a certificate chain where not the root,\nbut an intermediate certificate CA, or the host certificate, are\nmarked trusted (in <a class=\"link\" href=\"#certificate-file\" title=\"3.28.\u00a0certificate_file\">$certificate_file<\/a>), without marking the root\nsigning CA as trusted.\n<\/p><p>\n(OpenSSL 1.0.2b and newer only).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.332.\u00a0ssl_ciphers<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nContains a colon-seperated list of ciphers to use when using SSL.\nFor OpenSSL, see ciphers(1) for the syntax of the string.\n<\/p><p>\nFor GnuTLS, this option will be used in place of \"NORMAL\" at the\nstart of the priority string.  See gnutls_priority_init(3) for the\nsyntax and more details. (Note: GnuTLS version 2.1.7 or higher is\nrequired.)\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.333.\u00a0status_chars<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">-*%A<\/code>\u201d<\/p><\/div><p>\nControls the characters used by the \u201c%r\u201d indicator in\n<a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a>. The first character is used when the mailbox is\nunchanged. The second is used when the mailbox has been changed, and\nit needs to be resynchronized. The third is used if the mailbox is in\nread-only mode, or if the mailbox will not be written when exiting\nthat mailbox (You can toggle whether to write changes to a mailbox\nwith the <code class=\"literal\">&lt;toggle-write&gt;<\/code> operation, bound by default to \u201c%\u201d). The fourth\nis used to indicate that the current folder has been opened in attach-\nmessage mode (Certain operations like composing a new mail, replying,\nforwarding, etc. are not permitted in this mode).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.334.\u00a0status_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">-%r-Mutt:\u00a0%f\u00a0[Msgs:%?M?%M\/?%m%?n?\u00a0New:%n?%?o?\u00a0Old:%o?%?d?\u00a0Del:%d?%?F?\u00a0Flag:%F?%?t?\u00a0Tag:%t?%?p?\u00a0Post:%p?%?b?\u00a0Inc:%b?%?l?\u00a0%l?]---(%s\/%S)-%&gt;-(%P)---<\/code>\u201d<\/p><\/div><p>\nControls the format of the status line displayed in the \u201cindex\u201d\nmenu.  This string is similar to <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a>, but has its own\nset of <code class=\"literal\">printf(3)<\/code>-like sequences:\n<\/p><p>\nFor an explanation of \u201csoft-fill\u201d, see the <a class=\"link\" href=\"#index-format\" title=\"3.130.\u00a0index_format\">$index_format<\/a> documentation.\n<\/p><p>\n* = can be optionally printed if nonzero\n<\/p><p>\nSome of the above sequences can be used to optionally print a string\nif their value is nonzero.  For example, you may only want to see the\nnumber of flagged messages if such messages exist, since zero is not\nparticularly meaningful.  To optionally print a string based upon one\nof the above sequences, the following construct is used:\n<\/p><p>\n<code class=\"literal\">%?&lt;sequence_char&gt;?&lt;optional_string&gt;?<\/code>\n<\/p><p>\nwhere <em>sequence_char<\/em> is a character from the table above, and\n<em>optional_string<\/em> is the string you would like printed if\n<em>sequence_char<\/em> is nonzero.  <em>optional_string<\/em> <strong>may<\/strong> contain\nother sequences as well as normal text, but you may <strong>not<\/strong> nest\noptional strings.\n<\/p><p>\nHere is an example illustrating how to optionally print the number of\nnew messages in a mailbox:\n<\/p><p>\n<code class=\"literal\">%?n?%n new messages.?<\/code>\n<\/p><p>\nYou can also switch between two strings using the following construct:\n<\/p><p>\n<code class=\"literal\">%?&lt;sequence_char&gt;?&lt;if_string&gt;&amp;&lt;else_string&gt;?<\/code>\n<\/p><p>\nIf the value of <em>sequence_char<\/em> is non-zero, <em>if_string<\/em> will\nbe expanded, otherwise <em>else_string<\/em> will be expanded.\n<\/p><p>\nYou can force the result of any <code class=\"literal\">printf(3)<\/code>-like sequence to be lowercase\nby prefixing the sequence character with an underscore (\u201c_\u201d) sign.\nFor example, if you want to display the local hostname in lowercase,\nyou would use: \u201c<code class=\"literal\">%_h<\/code>\u201d.\n<\/p><p>\nIf you prefix the sequence character with a colon (\u201c:\u201d) character, mutt\nwill replace any dots in the expansion by underscores. This might be helpful\nwith IMAP folders that don't like dots in folder names.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.335.\u00a0status_on_top<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nSetting this variable causes the \u201cstatus bar\u201d to be displayed on\nthe first line of the screen rather than near the bottom. If <a class=\"link\" href=\"#help\" title=\"3.94.\u00a0help\">$help<\/a>\nis <em>set<\/em>, too it'll be placed at the bottom.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.336.\u00a0strict_threads<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nIf <em>set<\/em>, threading will only make use of the \u201cIn-Reply-To\u201d and\n\u201cReferences:\u201d fields when you <a class=\"link\" href=\"#sort\" title=\"3.311.\u00a0sort\">$sort<\/a> by message threads.  By\ndefault, messages with the same subject are grouped together in\n\u201cpseudo threads.\u201d. This may not always be desirable, such as in a\npersonal mailbox where you might have several unrelated messages with\nthe subjects like \u201chi\u201d which will get grouped together. See also\n<a class=\"link\" href=\"#sort-re\" title=\"3.315.\u00a0sort_re\">$sort_re<\/a> for a less drastic way of controlling this\nbehavior.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.337.\u00a0suspend<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>unset<\/em>, mutt won't stop when the user presses the terminal's\n<em>susp<\/em> key, usually \u201c^Z\u201d. This is useful if you run mutt\ninside an xterm using a command like \u201c<code class=\"literal\">xterm -e mutt<\/code>\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.338.\u00a0text_flowed<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will generate \u201cformat=flowed\u201d bodies with a content type\nof \u201c<code class=\"literal\">text\/plain; format=flowed<\/code>\u201d.\nThis format is easier to handle for some mailing software, and generally\njust looks like ordinary text.  To actually make use of this format's\nfeatures, you'll need support in your editor.\n<\/p><p>\nNote that <a class=\"link\" href=\"#indent-string\" title=\"3.129.\u00a0indent_string\">$indent_string<\/a> is ignored when this option is <em>set<\/em>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.339.\u00a0thorough_search<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nAffects the <code class=\"literal\">~b<\/code> and <code class=\"literal\">~h<\/code> search operations described in\nsection \u201c<a class=\"link\" href=\"#patterns\" title=\"3.\u00a0Patterns: Searching, Limiting and Tagging\">patterns<\/a>\u201d.  If <em>set<\/em>, the headers and body\/attachments of\nmessages to be searched are decoded before searching. If <em>unset<\/em>,\nmessages are searched as they appear in the folder.\n<\/p><p>\nUsers searching attachments or for non-ASCII characters should <em>set<\/em>\nthis value because decoding also includes MIME parsing\/decoding and possible\ncharacter set conversions. Otherwise mutt will attempt to match against the\nraw message received (for example quoted-printable encoded or with encoded\nheaders) which may lead to incorrect search results.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.340.\u00a0thread_received<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt uses the date received rather than the date sent\nto thread messages by subject.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.341.\u00a0tilde<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, the internal-pager will pad blank lines to the bottom of the\nscreen with a tilde (\u201c~\u201d).\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.342.\u00a0time_inc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nAlong with <a class=\"link\" href=\"#read-inc\" title=\"3.236.\u00a0read_inc\">$read_inc<\/a>, <a class=\"link\" href=\"#write-inc\" title=\"3.367.\u00a0write_inc\">$write_inc<\/a>, and <a class=\"link\" href=\"#net-inc\" title=\"3.169.\u00a0net_inc\">$net_inc<\/a>, this\nvariable controls the frequency with which progress updates are\ndisplayed. It suppresses updates less than <a class=\"link\" href=\"#time-inc\" title=\"3.342.\u00a0time_inc\">$time_inc<\/a> milliseconds\napart. This can improve throughput on systems with slow terminals,\nor when running mutt on a remote system.\n<\/p><p>\nAlso see the \u201c<a class=\"link\" href=\"#tuning\" title=\"Chapter\u00a08.\u00a0Performance Tuning\">tuning<\/a>\u201d section of the manual for performance considerations.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.343.\u00a0timeout<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a0600<\/p><\/div><p>\nWhen Mutt is waiting for user input either idling in menus or\nin an interactive prompt, Mutt would block until input is\npresent. Depending on the context, this would prevent certain\noperations from working, like checking for new mail or keeping\nan IMAP connection alive.\n<\/p><p>\nThis variable controls how many seconds Mutt will at most wait\nuntil it aborts waiting for input, performs these operations and\ncontinues to wait for input.\n<\/p><p>\nA value of zero or less will cause Mutt to never time out.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.344.\u00a0tmpdir<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nThis variable allows you to specify where Mutt will place its\ntemporary files needed for displaying and composing messages.  If\nthis variable is not set, the environment variable <code class=\"literal\">$TMPDIR<\/code> is\nused.  If <code class=\"literal\">$TMPDIR<\/code> is not set then \u201c<code class=\"literal\">\/tmp<\/code>\u201d is used.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.345.\u00a0to_chars<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">\u00a0+TCFL<\/code>\u201d<\/p><\/div><p>\nControls the character used to indicate mail addressed to you.  The\nfirst character is the one used when the mail is <em>not<\/em> addressed to your\naddress.  The second is used when you are the only\nrecipient of the message.  The third is when your address\nappears in the \u201cTo:\u201d header field, but you are not the only recipient of\nthe message.  The fourth character is used when your\naddress is specified in the \u201cCc:\u201d header field, but you are not the only\nrecipient.  The fifth character is used to indicate mail that was sent\nby <em>you<\/em>.  The sixth character is used to indicate when a mail\nwas sent to a mailing-list you subscribe to.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.346.\u00a0trash<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nIf set, this variable specifies the path of the trash folder where the\nmails marked for deletion will be moved, instead of being irremediably\npurged.\n<\/p><p>\nNOTE: When you delete a message in the trash folder, it is really\ndeleted, so that you have a way to clean the trash.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.347.\u00a0ts_icon_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">M%?n?AIL&amp;ail?<\/code>\u201d<\/p><\/div><p>\nControls the format of the icon title, as long as \u201c<a class=\"link\" href=\"#ts-enabled\" title=\"3.348.\u00a0ts_enabled\">$ts_enabled<\/a>\u201d is set.\nThis string is identical in formatting to the one used by\n\u201c<a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a>\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.348.\u00a0ts_enabled<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nControls whether mutt tries to set the terminal status line and icon name.\nMost terminal emulators emulate the status line in the window title.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.349.\u00a0ts_status_format<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0\u201c<code class=\"literal\">Mutt\u00a0with\u00a0%?m?%m\u00a0messages&amp;no\u00a0messages?%?n?\u00a0[%n\u00a0NEW]?<\/code>\u201d<\/p><\/div><p>\nControls the format of the terminal status line (or window title),\nprovided that \u201c<a class=\"link\" href=\"#ts-enabled\" title=\"3.348.\u00a0ts_enabled\">$ts_enabled<\/a>\u201d has been set. This string is identical in\nformatting to the one used by \u201c<a class=\"link\" href=\"#status-format\" title=\"3.334.\u00a0status_format\">$status_format<\/a>\u201d.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.350.\u00a0tunnel<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0string<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSetting this variable will cause mutt to open a pipe to a command\ninstead of a raw socket. You may be able to use this to set up\npreauthenticated connections to your IMAP\/POP3\/SMTP server. Example:\n<\/p><pre class=\"screen\">\nset tunnel=\"ssh -q mailhost.net \/usr\/local\/libexec\/imapd\"\n<\/pre><p>\nNote: For this example to work you must be able to log in to the remote\nmachine without having to enter a password.\n<\/p><p>\nWhen set, Mutt uses the tunnel for all remote connections.\nPlease see \u201c<a class=\"link\" href=\"#account-hook\" title=\"6.\u00a0Managing Multiple Accounts\">account-hook<\/a>\u201d in the manual for how to use different\ntunnel commands per connection.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.351.\u00a0uncollapse_jump<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will jump to the next unread message, if any,\nwhen the current thread is <em>un<\/em>collapsed.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.352.\u00a0uncollapse_new<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will automatically uncollapse any collapsed thread\nthat receives a new message. When <em>unset<\/em>, collapsed threads will\nremain collapsed. the presence of the new message will still affect\nindex sorting, though.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.353.\u00a0use_8bitmime<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\n<strong>Warning:<\/strong> do not set this variable unless you are using a version\nof sendmail which supports the <code class=\"literal\">-B8BITMIME<\/code> flag (such as sendmail\n8.8.x) or you may not be able to send mail.\n<\/p><p>\nWhen <em>set<\/em>, Mutt will invoke <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a> with the <code class=\"literal\">-B8BITMIME<\/code>\nflag when sending 8-bit messages to enable ESMTP negotiation.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.354.\u00a0use_domain<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will qualify all local addresses (ones without the\n\u201c@host\u201d portion) with the value of <a class=\"link\" href=\"#hostname\" title=\"3.106.\u00a0hostname\">$hostname<\/a>.  If <em>unset<\/em>, no\naddresses will be qualified.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.355.\u00a0use_envelope_from<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0no<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will set the <em>envelope<\/em> sender of the message.\nIf <a class=\"link\" href=\"#envelope-from-address\" title=\"3.68.\u00a0envelope_from_address\">$envelope_from_address<\/a> is <em>set<\/em>, it will be used as the sender\naddress. If <em>unset<\/em>, mutt will attempt to derive the sender from the\n\u201cFrom:\u201d header.\n<\/p><p>\nNote that this information is passed to sendmail command using the\n<code class=\"literal\">-f<\/code> command line switch. Therefore setting this option is not useful\nif the <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a> variable already contains <code class=\"literal\">-f<\/code> or if the\nexecutable pointed to by <a class=\"link\" href=\"#sendmail\" title=\"3.264.\u00a0sendmail\">$sendmail<\/a> doesn't support the <code class=\"literal\">-f<\/code> switch.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.356.\u00a0use_from<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will generate the \u201cFrom:\u201d header field when\nsending messages.  If <em>unset<\/em>, no \u201cFrom:\u201d header field will be\ngenerated unless the user explicitly sets one using the \u201c<a class=\"link\" href=\"#my-hdr\" title=\"15.\u00a0User-Defined Headers\">my_hdr<\/a>\u201d\ncommand.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.357.\u00a0use_ipv6<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, Mutt will look for IPv6 addresses of hosts it tries to\ncontact.  If this option is <em>unset<\/em>, Mutt will restrict itself to IPv4 addresses.\nNormally, the default should work.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.358.\u00a0user_agent<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will add a \u201cUser-Agent:\u201d header to outgoing\nmessages, indicating which version of mutt was used for composing\nthem.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.359.\u00a0visual<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0path<br \/>Default:\u00a0(empty)<\/p><\/div><p>\nSpecifies the visual editor to invoke when the \u201c<code class=\"literal\">~v<\/code>\u201d command is\ngiven in the built-in editor.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.360.\u00a0wait_key<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls whether Mutt will ask you to press a key after an external command\nhas been invoked by these functions: <code class=\"literal\">&lt;shell-escape&gt;<\/code>,\n<code class=\"literal\">&lt;pipe-message&gt;<\/code>, <code class=\"literal\">&lt;pipe-entry&gt;<\/code>, <code class=\"literal\">&lt;print-message&gt;<\/code>,\nand <code class=\"literal\">&lt;print-entry&gt;<\/code> commands.\n<\/p><p>\nIt is also used when viewing attachments with \u201c<a class=\"link\" href=\"#auto-view\" title=\"4.\u00a0MIME Autoview\">auto_view<\/a>\u201d, provided\nthat the corresponding mailcap entry has a <em>needsterminal<\/em> flag,\nand the external program is interactive.\n<\/p><p>\nWhen <em>set<\/em>, Mutt will always ask for a key. When <em>unset<\/em>, Mutt will wait\nfor a key only if the external command returned a non-zero status.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.361.\u00a0weed<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nWhen <em>set<\/em>, mutt will weed headers when displaying, forwarding,\nprinting, or replying to messages.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.362.\u00a0wrap<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\nWhen set to a positive value, mutt will wrap text at <a class=\"link\" href=\"#wrap\" title=\"3.362.\u00a0wrap\">$wrap<\/a> characters.\nWhen set to a negative value, mutt will wrap text so that there are <a class=\"link\" href=\"#wrap\" title=\"3.362.\u00a0wrap\">$wrap<\/a>\ncharacters of empty space on the right side of the terminal. Setting it\nto zero makes mutt wrap at the terminal width.\n<\/p><p>\nAlso see <a class=\"link\" href=\"#reflow-wrap\" title=\"3.243.\u00a0reflow_wrap\">$reflow_wrap<\/a>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.363.\u00a0wrap_headers<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a078<\/p><\/div><p>\nThis option specifies the number of characters to use for wrapping\nan outgoing message's headers. Allowed values are between 78 and 998\ninclusive.\n<\/p><p>\n<strong>Note:<\/strong> This option usually shouldn't be changed. RFC5233\nrecommends a line length of 78 (the default), so <strong>please only change\nthis setting when you know what you're doing<\/strong>.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.364.\u00a0wrap_search<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls whether searches wrap around the end.\n<\/p><p>\nWhen <em>set<\/em>, searches will wrap around the first (or last) item. When\n<em>unset<\/em>, incremental searches will not wrap.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.365.\u00a0wrapmargin<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a00<\/p><\/div><p>\n(DEPRECATED) Equivalent to setting <a class=\"link\" href=\"#wrap\" title=\"3.362.\u00a0wrap\">$wrap<\/a> with a negative value.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.366.\u00a0write_bcc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0boolean<br \/>Default:\u00a0yes<\/p><\/div><p>\nControls whether mutt writes out the \u201cBcc:\u201d header when preparing\nmessages to be sent.  Exim users may wish to unset this. If mutt\nis set to deliver directly via SMTP (see <a class=\"link\" href=\"#smtp-url\" title=\"3.310.\u00a0smtp_url\">$smtp_url<\/a>), this\noption does nothing: mutt will never write out the \u201cBcc:\u201d header\nin this case.\n<\/p><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">3.367.\u00a0write_inc<\/h3><\/div><\/div><div class=\"literallayout\"><p>Type:\u00a0number<br \/>Default:\u00a010<\/p><\/div><p>\nWhen writing a mailbox, a message will be printed every\n<a class=\"link\" href=\"#write-inc\" title=\"3.367.\u00a0write_inc\">$write_inc<\/a> messages to indicate progress.  If set to 0, only a\nsingle message will be displayed before writing a mailbox.\n<\/p><p>\nAlso see the <a class=\"link\" href=\"#read-inc\" title=\"3.236.\u00a0read_inc\">$read_inc<\/a>, <a class=\"link\" href=\"#net-inc\" title=\"3.169.\u00a0net_inc\">$net_inc<\/a> and <a class=\"link\" href=\"#time-inc\" title=\"3.342.\u00a0time_inc\">$time_inc<\/a> variables and the\n\u201c<a class=\"link\" href=\"#tuning\" title=\"Chapter\u00a08.\u00a0Performance Tuning\">tuning<\/a>\u201d section of the manual for performance considerations.\n<\/p><\/div><\/div><div class=\"sect1\"><div class=\"titlepage\"><div><p><\/p><h2 class=\"title\">4.\u00a0Functions<\/h2><\/div><\/div><p>\nThe following is the list of available functions listed by the mapping\nin which they are available.  The default key setting is given, and an\nexplanation of what the function does.  The key bindings of these\nfunctions can be changed with the <a class=\"link\" href=\"#bind\" title=\"5.\u00a0Changing the Default Key Bindings\">bind<\/a>\ncommand.\n<\/p><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.1.\u00a0Generic Menu<\/h3><\/div><\/div><p>\nThe <em>generic<\/em> menu is not a real menu, but specifies common functions\n(such as movement) available in all menus except for <em>pager<\/em> and\n<em>editor<\/em>.  Changing settings for this menu will affect the default\nbindings for all menus (except as noted).\n<\/p><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.2.\u00a0Default Generic Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Generic Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;top-page&gt;<\/code><\/td><td>H<\/td><td>move to the top of the page<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-entry&gt;<\/code><\/td><td>j<\/td><td>move to the next entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-entry&gt;<\/code><\/td><td>k<\/td><td>move to the previous entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;bottom-page&gt;<\/code><\/td><td>L<\/td><td>move to the bottom of the page<\/td><\/tr><tr><td><code class=\"literal\">&lt;refresh&gt;<\/code><\/td><td>^L<\/td><td>clear and redraw the screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;middle-page&gt;<\/code><\/td><td>M<\/td><td>move to the middle of the page<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-next&gt;<\/code><\/td><td>n<\/td><td>search for next match<\/td><\/tr><tr><td><code class=\"literal\">&lt;exit&gt;<\/code><\/td><td>q<\/td><td>exit this menu<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-entry&gt;<\/code><\/td><td>t<\/td><td>tag the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-page&gt;<\/code><\/td><td>z<\/td><td>move to the next page<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-page&gt;<\/code><\/td><td>Z<\/td><td>move to the previous page<\/td><\/tr><tr><td><code class=\"literal\">&lt;last-entry&gt;<\/code><\/td><td>*<\/td><td>move to the last entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;first-entry&gt;<\/code><\/td><td>=<\/td><td>move to the first entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;enter-command&gt;<\/code><\/td><td>:<\/td><td>enter a muttrc command<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-line&gt;<\/code><\/td><td>&gt;<\/td><td>scroll down one line<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-line&gt;<\/code><\/td><td>&lt;<\/td><td>scroll up one line<\/td><\/tr><tr><td><code class=\"literal\">&lt;half-up&gt;<\/code><\/td><td>[<\/td><td>scroll up 1\/2 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;half-down&gt;<\/code><\/td><td>]<\/td><td>scroll down 1\/2 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;help&gt;<\/code><\/td><td>?<\/td><td>this screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-prefix&gt;<\/code><\/td><td>;<\/td><td>apply next function to tagged messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-prefix-cond&gt;<\/code><\/td><td>\u00a0<\/td><td>apply next function ONLY to tagged messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;end-cond&gt;<\/code><\/td><td>\u00a0<\/td><td>end of conditional execution (noop)<\/td><\/tr><tr><td><code class=\"literal\">&lt;shell-escape&gt;<\/code><\/td><td>!<\/td><td>invoke a command in a subshell<\/td><\/tr><tr><td><code class=\"literal\">&lt;select-entry&gt;<\/code><\/td><td>MUTT_ENTER_S<\/td><td>select the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;search&gt;<\/code><\/td><td>\/<\/td><td>search for a regular expression<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-reverse&gt;<\/code><\/td><td>Esc \/<\/td><td>search backwards for a regular expression<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-opposite&gt;<\/code><\/td><td>\u00a0<\/td><td>search for next match in opposite direction<\/td><\/tr><tr><td><code class=\"literal\">&lt;jump&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to an index number<\/td><\/tr><tr><td><code class=\"literal\">&lt;current-top&gt;<\/code><\/td><td>\u00a0<\/td><td>move entry to top of screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;current-middle&gt;<\/code><\/td><td>\u00a0<\/td><td>move entry to middle of screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;current-bottom&gt;<\/code><\/td><td>\u00a0<\/td><td>move entry to bottom of screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;error-history&gt;<\/code><\/td><td>\u00a0<\/td><td>display recent history of error messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;what-key&gt;<\/code><\/td><td>\u00a0<\/td><td>display the keycode for a key press<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.2.\u00a0Index Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.3.\u00a0Default Index Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Index Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;create-alias&gt;<\/code><\/td><td>a<\/td><td>create an alias from a message sender<\/td><\/tr><tr><td><code class=\"literal\">&lt;bounce-message&gt;<\/code><\/td><td>b<\/td><td>remail a message to another user<\/td><\/tr><tr><td><code class=\"literal\">&lt;break-thread&gt;<\/code><\/td><td>#<\/td><td>break the thread in two<\/td><\/tr><tr><td><code class=\"literal\">&lt;change-folder&gt;<\/code><\/td><td>c<\/td><td>open a different folder<\/td><\/tr><tr><td><code class=\"literal\">&lt;change-folder-readonly&gt;<\/code><\/td><td>Esc c<\/td><td>open a different folder in read only mode<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-unread-mailbox&gt;<\/code><\/td><td>\u00a0<\/td><td>open next mailbox with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;collapse-thread&gt;<\/code><\/td><td>Esc v<\/td><td>collapse\/uncollapse current thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;collapse-all&gt;<\/code><\/td><td>Esc V<\/td><td>collapse\/uncollapse all threads<\/td><\/tr><tr><td><code class=\"literal\">&lt;copy-message&gt;<\/code><\/td><td>C<\/td><td>copy a message to a file\/mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;decode-copy&gt;<\/code><\/td><td>Esc C<\/td><td>make decoded (text\/plain) copy<\/td><\/tr><tr><td><code class=\"literal\">&lt;decode-save&gt;<\/code><\/td><td>Esc s<\/td><td>make decoded copy (text\/plain) and delete<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-message&gt;<\/code><\/td><td>d<\/td><td>delete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-pattern&gt;<\/code><\/td><td>D<\/td><td>delete messages matching a pattern<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-thread&gt;<\/code><\/td><td>^D<\/td><td>delete all messages in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-subthread&gt;<\/code><\/td><td>Esc d<\/td><td>delete all messages in subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit&gt;<\/code><\/td><td>e<\/td><td>edit the raw message<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-label&gt;<\/code><\/td><td>Y<\/td><td>add, change, or delete a message's label<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-type&gt;<\/code><\/td><td>^E<\/td><td>edit attachment content type<\/td><\/tr><tr><td><code class=\"literal\">&lt;forward-message&gt;<\/code><\/td><td>f<\/td><td>forward a message with comments<\/td><\/tr><tr><td><code class=\"literal\">&lt;flag-message&gt;<\/code><\/td><td>F<\/td><td>toggle a message's 'important' flag<\/td><\/tr><tr><td><code class=\"literal\">&lt;group-reply&gt;<\/code><\/td><td>g<\/td><td>reply to all recipients<\/td><\/tr><tr><td><code class=\"literal\">&lt;fetch-mail&gt;<\/code><\/td><td>G<\/td><td>retrieve mail from POP server<\/td><\/tr><tr><td><code class=\"literal\">&lt;imap-fetch-mail&gt;<\/code><\/td><td>\u00a0<\/td><td>force retrieval of mail from IMAP server<\/td><\/tr><tr><td><code class=\"literal\">&lt;imap-logout-all&gt;<\/code><\/td><td>\u00a0<\/td><td>logout from all IMAP servers<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-toggle-weed&gt;<\/code><\/td><td>h<\/td><td>display message and toggle header weeding<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-undeleted&gt;<\/code><\/td><td>j<\/td><td>move to the next undeleted message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-undeleted&gt;<\/code><\/td><td>k<\/td><td>move to the previous undeleted message<\/td><\/tr><tr><td><code class=\"literal\">&lt;limit&gt;<\/code><\/td><td>l<\/td><td>show only messages matching a pattern<\/td><\/tr><tr><td><code class=\"literal\">&lt;link-threads&gt;<\/code><\/td><td>&amp;<\/td><td>link tagged message to the current one<\/td><\/tr><tr><td><code class=\"literal\">&lt;list-reply&gt;<\/code><\/td><td>L<\/td><td>reply to specified mailing list<\/td><\/tr><tr><td><code class=\"literal\">&lt;mail&gt;<\/code><\/td><td>m<\/td><td>compose a new mail message<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-new&gt;<\/code><\/td><td>N<\/td><td>toggle a message's 'new' flag<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-write&gt;<\/code><\/td><td>%<\/td><td>toggle whether the mailbox will be rewritten<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-thread&gt;<\/code><\/td><td>^N<\/td><td>jump to the next thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-subthread&gt;<\/code><\/td><td>Esc n<\/td><td>jump to the next subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;purge-message&gt;<\/code><\/td><td>\u00a0<\/td><td>delete the current entry, bypassing the trash folder<\/td><\/tr><tr><td><code class=\"literal\">&lt;query&gt;<\/code><\/td><td>Q<\/td><td>query external program for addresses<\/td><\/tr><tr><td><code class=\"literal\">&lt;quit&gt;<\/code><\/td><td>q<\/td><td>save changes to mailbox and quit<\/td><\/tr><tr><td><code class=\"literal\">&lt;reply&gt;<\/code><\/td><td>r<\/td><td>reply to a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;show-limit&gt;<\/code><\/td><td>Esc l<\/td><td>show currently active limit pattern<\/td><\/tr><tr><td><code class=\"literal\">&lt;sort-mailbox&gt;<\/code><\/td><td>o<\/td><td>sort messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;sort-reverse&gt;<\/code><\/td><td>O<\/td><td>sort messages in reverse order<\/td><\/tr><tr><td><code class=\"literal\">&lt;print-message&gt;<\/code><\/td><td>p<\/td><td>print the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-thread&gt;<\/code><\/td><td>^P<\/td><td>jump to previous thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-subthread&gt;<\/code><\/td><td>Esc p<\/td><td>jump to previous subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;recall-message&gt;<\/code><\/td><td>R<\/td><td>recall a postponed message<\/td><\/tr><tr><td><code class=\"literal\">&lt;read-thread&gt;<\/code><\/td><td>^R<\/td><td>mark the current thread as read<\/td><\/tr><tr><td><code class=\"literal\">&lt;read-subthread&gt;<\/code><\/td><td>Esc r<\/td><td>mark the current subthread as read<\/td><\/tr><tr><td><code class=\"literal\">&lt;resend-message&gt;<\/code><\/td><td>Esc e<\/td><td>use the current message as a template for a new one<\/td><\/tr><tr><td><code class=\"literal\">&lt;save-message&gt;<\/code><\/td><td>s<\/td><td>save message\/attachment to a mailbox\/file<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-pattern&gt;<\/code><\/td><td>T<\/td><td>tag messages matching a pattern<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-subthread&gt;<\/code><\/td><td>\u00a0<\/td><td>tag the current subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-thread&gt;<\/code><\/td><td>Esc t<\/td><td>tag the current thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;untag-pattern&gt;<\/code><\/td><td>^T<\/td><td>untag messages matching a pattern<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-message&gt;<\/code><\/td><td>u<\/td><td>undelete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-pattern&gt;<\/code><\/td><td>U<\/td><td>undelete messages matching a pattern<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-subthread&gt;<\/code><\/td><td>Esc u<\/td><td>undelete all messages in subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-thread&gt;<\/code><\/td><td>^U<\/td><td>undelete all messages in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-attachments&gt;<\/code><\/td><td>v<\/td><td>show MIME attachments<\/td><\/tr><tr><td><code class=\"literal\">&lt;show-version&gt;<\/code><\/td><td>V<\/td><td>show the Mutt version number and date<\/td><\/tr><tr><td><code class=\"literal\">&lt;set-flag&gt;<\/code><\/td><td>w<\/td><td>set a status flag on a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;clear-flag&gt;<\/code><\/td><td>W<\/td><td>clear a status flag from a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-message&gt;<\/code><\/td><td>MUTT_ENTER_S<\/td><td>display a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;mark-message&gt;<\/code><\/td><td>~<\/td><td>create a hotkey macro for the current message<\/td><\/tr><tr><td><code class=\"literal\">&lt;buffy-list&gt;<\/code><\/td><td>.<\/td><td>list mailboxes with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;sync-mailbox&gt;<\/code><\/td><td>$<\/td><td>save changes to mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-address&gt;<\/code><\/td><td>@<\/td><td>display full address of sender<\/td><\/tr><tr><td><code class=\"literal\">&lt;pipe-message&gt;<\/code><\/td><td>|<\/td><td>pipe message\/attachment to a shell command<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-new&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the next new message<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-new-then-unread&gt;<\/code><\/td><td>&lt;Tab&gt;<\/td><td>jump to the next new or unread message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-new&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the previous new message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-new-then-unread&gt;<\/code><\/td><td>Esc &lt;Tab&gt;<\/td><td>jump to the previous new or unread message<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-unread&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the next unread message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-unread&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the previous unread message<\/td><\/tr><tr><td><code class=\"literal\">&lt;parent-message&gt;<\/code><\/td><td>P<\/td><td>jump to parent message in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;root-message&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to root message in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;extract-keys&gt;<\/code><\/td><td>^K<\/td><td>extract supported public keys<\/td><\/tr><tr><td><code class=\"literal\">&lt;forget-passphrase&gt;<\/code><\/td><td>^F<\/td><td>wipe passphrase(s) from memory<\/td><\/tr><tr><td><code class=\"literal\">&lt;check-traditional-pgp&gt;<\/code><\/td><td>Esc P<\/td><td>check for classic PGP<\/td><\/tr><tr><td><code class=\"literal\">&lt;mail-key&gt;<\/code><\/td><td>Esc k<\/td><td>mail a PGP public key<\/td><\/tr><tr><td><code class=\"literal\">&lt;decrypt-copy&gt;<\/code><\/td><td>\u00a0<\/td><td>make decrypted copy<\/td><\/tr><tr><td><code class=\"literal\">&lt;decrypt-save&gt;<\/code><\/td><td>\u00a0<\/td><td>make decrypted copy and delete<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-next&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to next mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-next-new&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to next mailbox with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-open&gt;<\/code><\/td><td>\u00a0<\/td><td>open highlighted mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-page-down&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll the sidebar down 1 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-page-up&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll the sidebar up 1 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-prev&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to previous mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-prev-new&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to previous mailbox with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-toggle-visible&gt;<\/code><\/td><td>\u00a0<\/td><td>make the sidebar (in)visible<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.3.\u00a0Pager Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.4.\u00a0Default Pager Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Pager Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;break-thread&gt;<\/code><\/td><td>#<\/td><td>break the thread in two<\/td><\/tr><tr><td><code class=\"literal\">&lt;create-alias&gt;<\/code><\/td><td>a<\/td><td>create an alias from a message sender<\/td><\/tr><tr><td><code class=\"literal\">&lt;bounce-message&gt;<\/code><\/td><td>b<\/td><td>remail a message to another user<\/td><\/tr><tr><td><code class=\"literal\">&lt;change-folder&gt;<\/code><\/td><td>c<\/td><td>open a different folder<\/td><\/tr><tr><td><code class=\"literal\">&lt;change-folder-readonly&gt;<\/code><\/td><td>Esc c<\/td><td>open a different folder in read only mode<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-unread-mailbox&gt;<\/code><\/td><td>\u00a0<\/td><td>open next mailbox with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;copy-message&gt;<\/code><\/td><td>C<\/td><td>copy a message to a file\/mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;decode-copy&gt;<\/code><\/td><td>Esc C<\/td><td>make decoded (text\/plain) copy<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-message&gt;<\/code><\/td><td>d<\/td><td>delete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-thread&gt;<\/code><\/td><td>^D<\/td><td>delete all messages in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-subthread&gt;<\/code><\/td><td>Esc d<\/td><td>delete all messages in subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;set-flag&gt;<\/code><\/td><td>w<\/td><td>set a status flag on a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;clear-flag&gt;<\/code><\/td><td>W<\/td><td>clear a status flag from a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit&gt;<\/code><\/td><td>e<\/td><td>edit the raw message<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-label&gt;<\/code><\/td><td>Y<\/td><td>add, change, or delete a message's label<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-type&gt;<\/code><\/td><td>^E<\/td><td>edit attachment content type<\/td><\/tr><tr><td><code class=\"literal\">&lt;forward-message&gt;<\/code><\/td><td>f<\/td><td>forward a message with comments<\/td><\/tr><tr><td><code class=\"literal\">&lt;flag-message&gt;<\/code><\/td><td>F<\/td><td>toggle a message's 'important' flag<\/td><\/tr><tr><td><code class=\"literal\">&lt;group-reply&gt;<\/code><\/td><td>g<\/td><td>reply to all recipients<\/td><\/tr><tr><td><code class=\"literal\">&lt;imap-fetch-mail&gt;<\/code><\/td><td>\u00a0<\/td><td>force retrieval of mail from IMAP server<\/td><\/tr><tr><td><code class=\"literal\">&lt;imap-logout-all&gt;<\/code><\/td><td>\u00a0<\/td><td>logout from all IMAP servers<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-toggle-weed&gt;<\/code><\/td><td>h<\/td><td>display message and toggle header weeding<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-undeleted&gt;<\/code><\/td><td>j<\/td><td>move to the next undeleted message<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-entry&gt;<\/code><\/td><td>J<\/td><td>move to the next entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-undeleted&gt;<\/code><\/td><td>k<\/td><td>move to the previous undeleted message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-entry&gt;<\/code><\/td><td>K<\/td><td>move to the previous entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;link-threads&gt;<\/code><\/td><td>&amp;<\/td><td>link tagged message to the current one<\/td><\/tr><tr><td><code class=\"literal\">&lt;list-reply&gt;<\/code><\/td><td>L<\/td><td>reply to specified mailing list<\/td><\/tr><tr><td><code class=\"literal\">&lt;redraw-screen&gt;<\/code><\/td><td>^L<\/td><td>clear and redraw the screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;mail&gt;<\/code><\/td><td>m<\/td><td>compose a new mail message<\/td><\/tr><tr><td><code class=\"literal\">&lt;mark-as-new&gt;<\/code><\/td><td>N<\/td><td>toggle a message's 'new' flag<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-next&gt;<\/code><\/td><td>n<\/td><td>search for next match<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-thread&gt;<\/code><\/td><td>^N<\/td><td>jump to the next thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-subthread&gt;<\/code><\/td><td>Esc n<\/td><td>jump to the next subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;sort-mailbox&gt;<\/code><\/td><td>o<\/td><td>sort messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;sort-reverse&gt;<\/code><\/td><td>O<\/td><td>sort messages in reverse order<\/td><\/tr><tr><td><code class=\"literal\">&lt;print-message&gt;<\/code><\/td><td>p<\/td><td>print the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-thread&gt;<\/code><\/td><td>^P<\/td><td>jump to previous thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-subthread&gt;<\/code><\/td><td>Esc p<\/td><td>jump to previous subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;purge-message&gt;<\/code><\/td><td>\u00a0<\/td><td>delete the current entry, bypassing the trash folder<\/td><\/tr><tr><td><code class=\"literal\">&lt;quit&gt;<\/code><\/td><td>Q<\/td><td>save changes to mailbox and quit<\/td><\/tr><tr><td><code class=\"literal\">&lt;exit&gt;<\/code><\/td><td>q<\/td><td>exit this menu<\/td><\/tr><tr><td><code class=\"literal\">&lt;reply&gt;<\/code><\/td><td>r<\/td><td>reply to a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;recall-message&gt;<\/code><\/td><td>R<\/td><td>recall a postponed message<\/td><\/tr><tr><td><code class=\"literal\">&lt;read-thread&gt;<\/code><\/td><td>^R<\/td><td>mark the current thread as read<\/td><\/tr><tr><td><code class=\"literal\">&lt;read-subthread&gt;<\/code><\/td><td>Esc r<\/td><td>mark the current subthread as read<\/td><\/tr><tr><td><code class=\"literal\">&lt;resend-message&gt;<\/code><\/td><td>Esc e<\/td><td>use the current message as a template for a new one<\/td><\/tr><tr><td><code class=\"literal\">&lt;save-message&gt;<\/code><\/td><td>s<\/td><td>save message\/attachment to a mailbox\/file<\/td><\/tr><tr><td><code class=\"literal\">&lt;skip-quoted&gt;<\/code><\/td><td>S<\/td><td>skip beyond quoted text<\/td><\/tr><tr><td><code class=\"literal\">&lt;decode-save&gt;<\/code><\/td><td>Esc s<\/td><td>make decoded copy (text\/plain) and delete<\/td><\/tr><tr><td><code class=\"literal\">&lt;tag-message&gt;<\/code><\/td><td>t<\/td><td>tag the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-quoted&gt;<\/code><\/td><td>T<\/td><td>toggle display of quoted text<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-message&gt;<\/code><\/td><td>u<\/td><td>undelete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-subthread&gt;<\/code><\/td><td>Esc u<\/td><td>undelete all messages in subthread<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-thread&gt;<\/code><\/td><td>^U<\/td><td>undelete all messages in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-attachments&gt;<\/code><\/td><td>v<\/td><td>show MIME attachments<\/td><\/tr><tr><td><code class=\"literal\">&lt;show-version&gt;<\/code><\/td><td>V<\/td><td>show the Mutt version number and date<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-toggle&gt;<\/code><\/td><td>\\\\<\/td><td>toggle search pattern coloring<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-address&gt;<\/code><\/td><td>@<\/td><td>display full address of sender<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-new&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the next new message<\/td><\/tr><tr><td><code class=\"literal\">&lt;pipe-message&gt;<\/code><\/td><td>|<\/td><td>pipe message\/attachment to a shell command<\/td><\/tr><tr><td><code class=\"literal\">&lt;help&gt;<\/code><\/td><td>?<\/td><td>this screen<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-page&gt;<\/code><\/td><td>&lt;Space&gt;<\/td><td>move to the next page<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-page&gt;<\/code><\/td><td>-<\/td><td>move to the previous page<\/td><\/tr><tr><td><code class=\"literal\">&lt;top&gt;<\/code><\/td><td>^<\/td><td>jump to the top of the message<\/td><\/tr><tr><td><code class=\"literal\">&lt;sync-mailbox&gt;<\/code><\/td><td>$<\/td><td>save changes to mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;shell-escape&gt;<\/code><\/td><td>!<\/td><td>invoke a command in a subshell<\/td><\/tr><tr><td><code class=\"literal\">&lt;enter-command&gt;<\/code><\/td><td>:<\/td><td>enter a muttrc command<\/td><\/tr><tr><td><code class=\"literal\">&lt;buffy-list&gt;<\/code><\/td><td>.<\/td><td>list mailboxes with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;search&gt;<\/code><\/td><td>\/<\/td><td>search for a regular expression<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-reverse&gt;<\/code><\/td><td>Esc \/<\/td><td>search backwards for a regular expression<\/td><\/tr><tr><td><code class=\"literal\">&lt;search-opposite&gt;<\/code><\/td><td>\u00a0<\/td><td>search for next match in opposite direction<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-line&gt;<\/code><\/td><td>MUTT_ENTER_S<\/td><td>scroll down one line<\/td><\/tr><tr><td><code class=\"literal\">&lt;error-history&gt;<\/code><\/td><td>\u00a0<\/td><td>display recent history of error messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;jump&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to an index number<\/td><\/tr><tr><td><code class=\"literal\">&lt;next-unread&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the next unread message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-new&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the previous new message<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-unread&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the previous unread message<\/td><\/tr><tr><td><code class=\"literal\">&lt;half-up&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll up 1\/2 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;half-down&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll down 1\/2 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;previous-line&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll up one line<\/td><\/tr><tr><td><code class=\"literal\">&lt;bottom&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to the bottom of the message<\/td><\/tr><tr><td><code class=\"literal\">&lt;parent-message&gt;<\/code><\/td><td>P<\/td><td>jump to parent message in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;root-message&gt;<\/code><\/td><td>\u00a0<\/td><td>jump to root message in thread<\/td><\/tr><tr><td><code class=\"literal\">&lt;check-traditional-pgp&gt;<\/code><\/td><td>Esc P<\/td><td>check for classic PGP<\/td><\/tr><tr><td><code class=\"literal\">&lt;mail-key&gt;<\/code><\/td><td>Esc k<\/td><td>mail a PGP public key<\/td><\/tr><tr><td><code class=\"literal\">&lt;extract-keys&gt;<\/code><\/td><td>^K<\/td><td>extract supported public keys<\/td><\/tr><tr><td><code class=\"literal\">&lt;forget-passphrase&gt;<\/code><\/td><td>^F<\/td><td>wipe passphrase(s) from memory<\/td><\/tr><tr><td><code class=\"literal\">&lt;decrypt-copy&gt;<\/code><\/td><td>\u00a0<\/td><td>make decrypted copy<\/td><\/tr><tr><td><code class=\"literal\">&lt;decrypt-save&gt;<\/code><\/td><td>\u00a0<\/td><td>make decrypted copy and delete<\/td><\/tr><tr><td><code class=\"literal\">&lt;what-key&gt;<\/code><\/td><td>\u00a0<\/td><td>display the keycode for a key press<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-next&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to next mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-next-new&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to next mailbox with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-open&gt;<\/code><\/td><td>\u00a0<\/td><td>open highlighted mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-page-down&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll the sidebar down 1 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-page-up&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll the sidebar up 1 page<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-prev&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to previous mailbox<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-prev-new&gt;<\/code><\/td><td>\u00a0<\/td><td>move the highlight to previous mailbox with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;sidebar-toggle-visible&gt;<\/code><\/td><td>\u00a0<\/td><td>make the sidebar (in)visible<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.4.\u00a0Alias Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.5.\u00a0Default Alias Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Alias Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;delete-entry&gt;<\/code><\/td><td>d<\/td><td>delete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-entry&gt;<\/code><\/td><td>u<\/td><td>undelete the current entry<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.5.\u00a0Query Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.6.\u00a0Default Query Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Query Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;create-alias&gt;<\/code><\/td><td>a<\/td><td>create an alias from a message sender<\/td><\/tr><tr><td><code class=\"literal\">&lt;mail&gt;<\/code><\/td><td>m<\/td><td>compose a new mail message<\/td><\/tr><tr><td><code class=\"literal\">&lt;query&gt;<\/code><\/td><td>Q<\/td><td>query external program for addresses<\/td><\/tr><tr><td><code class=\"literal\">&lt;query-append&gt;<\/code><\/td><td>A<\/td><td>append new query results to current results<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.6.\u00a0Attachment Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.7.\u00a0Default Attachment Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Attachment Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;bounce-message&gt;<\/code><\/td><td>b<\/td><td>remail a message to another user<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-toggle-weed&gt;<\/code><\/td><td>h<\/td><td>display message and toggle header weeding<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-type&gt;<\/code><\/td><td>^E<\/td><td>edit attachment content type<\/td><\/tr><tr><td><code class=\"literal\">&lt;print-entry&gt;<\/code><\/td><td>p<\/td><td>print the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;save-entry&gt;<\/code><\/td><td>s<\/td><td>save message\/attachment to a mailbox\/file<\/td><\/tr><tr><td><code class=\"literal\">&lt;pipe-entry&gt;<\/code><\/td><td>|<\/td><td>pipe message\/attachment to a shell command<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-mailcap&gt;<\/code><\/td><td>m<\/td><td>force viewing of attachment using mailcap<\/td><\/tr><tr><td><code class=\"literal\">&lt;reply&gt;<\/code><\/td><td>r<\/td><td>reply to a message<\/td><\/tr><tr><td><code class=\"literal\">&lt;resend-message&gt;<\/code><\/td><td>Esc e<\/td><td>use the current message as a template for a new one<\/td><\/tr><tr><td><code class=\"literal\">&lt;group-reply&gt;<\/code><\/td><td>g<\/td><td>reply to all recipients<\/td><\/tr><tr><td><code class=\"literal\">&lt;list-reply&gt;<\/code><\/td><td>L<\/td><td>reply to specified mailing list<\/td><\/tr><tr><td><code class=\"literal\">&lt;forward-message&gt;<\/code><\/td><td>f<\/td><td>forward a message with comments<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-text&gt;<\/code><\/td><td>T<\/td><td>view attachment as text<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-attach&gt;<\/code><\/td><td>MUTT_ENTER_S<\/td><td>view attachment using mailcap entry if necessary<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-entry&gt;<\/code><\/td><td>d<\/td><td>delete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-entry&gt;<\/code><\/td><td>u<\/td><td>undelete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;collapse-parts&gt;<\/code><\/td><td>v<\/td><td>Toggle display of subparts<\/td><\/tr><tr><td><code class=\"literal\">&lt;check-traditional-pgp&gt;<\/code><\/td><td>Esc P<\/td><td>check for classic PGP<\/td><\/tr><tr><td><code class=\"literal\">&lt;extract-keys&gt;<\/code><\/td><td>^K<\/td><td>extract supported public keys<\/td><\/tr><tr><td><code class=\"literal\">&lt;forget-passphrase&gt;<\/code><\/td><td>^F<\/td><td>wipe passphrase(s) from memory<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.7.\u00a0Compose Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.8.\u00a0Default Compose Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Compose Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;attach-file&gt;<\/code><\/td><td>a<\/td><td>attach file(s) to this message<\/td><\/tr><tr><td><code class=\"literal\">&lt;attach-message&gt;<\/code><\/td><td>A<\/td><td>attach message(s) to this message<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-bcc&gt;<\/code><\/td><td>b<\/td><td>edit the BCC list<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-cc&gt;<\/code><\/td><td>c<\/td><td>edit the CC list<\/td><\/tr><tr><td><code class=\"literal\">&lt;copy-file&gt;<\/code><\/td><td>C<\/td><td>save message\/attachment to a mailbox\/file<\/td><\/tr><tr><td><code class=\"literal\">&lt;detach-file&gt;<\/code><\/td><td>D<\/td><td>delete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-disposition&gt;<\/code><\/td><td>^D<\/td><td>toggle disposition between inline\/attachment<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-description&gt;<\/code><\/td><td>d<\/td><td>edit attachment description<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-message&gt;<\/code><\/td><td>e<\/td><td>edit the message<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-headers&gt;<\/code><\/td><td>E<\/td><td>edit the message with headers<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-file&gt;<\/code><\/td><td>^X e<\/td><td>edit the file to be attached<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-encoding&gt;<\/code><\/td><td>^E<\/td><td>edit attachment transfer-encoding<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-from&gt;<\/code><\/td><td>Esc f<\/td><td>edit the from field<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-fcc&gt;<\/code><\/td><td>f<\/td><td>enter a file to save a copy of this message in<\/td><\/tr><tr><td><code class=\"literal\">&lt;filter-entry&gt;<\/code><\/td><td>F<\/td><td>filter attachment through a shell command<\/td><\/tr><tr><td><code class=\"literal\">&lt;get-attachment&gt;<\/code><\/td><td>G<\/td><td>get a temporary copy of an attachment<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-toggle-weed&gt;<\/code><\/td><td>h<\/td><td>display message and toggle header weeding<\/td><\/tr><tr><td><code class=\"literal\">&lt;ispell&gt;<\/code><\/td><td>i<\/td><td>run ispell on the message<\/td><\/tr><tr><td><code class=\"literal\">&lt;print-entry&gt;<\/code><\/td><td>l<\/td><td>print the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-mime&gt;<\/code><\/td><td>m<\/td><td>edit attachment using mailcap entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;new-mime&gt;<\/code><\/td><td>n<\/td><td>compose new attachment using mailcap entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;postpone-message&gt;<\/code><\/td><td>P<\/td><td>save this message to send later<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-reply-to&gt;<\/code><\/td><td>r<\/td><td>edit the Reply-To field<\/td><\/tr><tr><td><code class=\"literal\">&lt;rename-attachment&gt;<\/code><\/td><td>^O<\/td><td>send attachment with a different name<\/td><\/tr><tr><td><code class=\"literal\">&lt;rename-file&gt;<\/code><\/td><td>R<\/td><td>rename\/move an attached file<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-subject&gt;<\/code><\/td><td>s<\/td><td>edit the subject of this message<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-to&gt;<\/code><\/td><td>t<\/td><td>edit the TO list<\/td><\/tr><tr><td><code class=\"literal\">&lt;edit-type&gt;<\/code><\/td><td>^T<\/td><td>edit attachment content type<\/td><\/tr><tr><td><code class=\"literal\">&lt;write-fcc&gt;<\/code><\/td><td>w<\/td><td>write the message to a folder<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-unlink&gt;<\/code><\/td><td>u<\/td><td>toggle whether to delete file after sending it<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-recode&gt;<\/code><\/td><td>\u00a0<\/td><td>toggle recoding of this attachment<\/td><\/tr><tr><td><code class=\"literal\">&lt;update-encoding&gt;<\/code><\/td><td>U<\/td><td>update an attachment's encoding info<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-attach&gt;<\/code><\/td><td>MUTT_ENTER_S<\/td><td>view attachment using mailcap entry if necessary<\/td><\/tr><tr><td><code class=\"literal\">&lt;send-message&gt;<\/code><\/td><td>y<\/td><td>send the message<\/td><\/tr><tr><td><code class=\"literal\">&lt;pipe-entry&gt;<\/code><\/td><td>|<\/td><td>pipe message\/attachment to a shell command<\/td><\/tr><tr><td><code class=\"literal\">&lt;attach-key&gt;<\/code><\/td><td>Esc k<\/td><td>attach a PGP public key<\/td><\/tr><tr><td><code class=\"literal\">&lt;pgp-menu&gt;<\/code><\/td><td>p<\/td><td>show PGP options<\/td><\/tr><tr><td><code class=\"literal\">&lt;forget-passphrase&gt;<\/code><\/td><td>^F<\/td><td>wipe passphrase(s) from memory<\/td><\/tr><tr><td><code class=\"literal\">&lt;smime-menu&gt;<\/code><\/td><td>S<\/td><td>show S\/MIME options<\/td><\/tr><tr><td><code class=\"literal\">&lt;mix&gt;<\/code><\/td><td>M<\/td><td>send the message through a mixmaster remailer chain<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.8.\u00a0Postpone Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.9.\u00a0Default Postpone Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Postpone Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;delete-entry&gt;<\/code><\/td><td>d<\/td><td>delete the current entry<\/td><\/tr><tr><td><code class=\"literal\">&lt;undelete-entry&gt;<\/code><\/td><td>u<\/td><td>undelete the current entry<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.9.\u00a0Browser Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.10.\u00a0Default Browser Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Browser Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;change-dir&gt;<\/code><\/td><td>c<\/td><td>change directories<\/td><\/tr><tr><td><code class=\"literal\">&lt;display-filename&gt;<\/code><\/td><td>@<\/td><td>display the currently selected file's name<\/td><\/tr><tr><td><code class=\"literal\">&lt;enter-mask&gt;<\/code><\/td><td>m<\/td><td>enter a file mask<\/td><\/tr><tr><td><code class=\"literal\">&lt;sort&gt;<\/code><\/td><td>o<\/td><td>sort messages<\/td><\/tr><tr><td><code class=\"literal\">&lt;sort-reverse&gt;<\/code><\/td><td>O<\/td><td>sort messages in reverse order<\/td><\/tr><tr><td><code class=\"literal\">&lt;select-new&gt;<\/code><\/td><td>N<\/td><td>select a new file in this directory<\/td><\/tr><tr><td><code class=\"literal\">&lt;check-new&gt;<\/code><\/td><td>\u00a0<\/td><td>check mailboxes for new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-mailboxes&gt;<\/code><\/td><td>&lt;Tab&gt;<\/td><td>toggle whether to browse mailboxes or all files<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-file&gt;<\/code><\/td><td>&lt;Space&gt;<\/td><td>view file<\/td><\/tr><tr><td><code class=\"literal\">&lt;buffy-list&gt;<\/code><\/td><td>.<\/td><td>list mailboxes with new mail<\/td><\/tr><tr><td><code class=\"literal\">&lt;create-mailbox&gt;<\/code><\/td><td>C<\/td><td>create a new mailbox (IMAP only)<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-mailbox&gt;<\/code><\/td><td>d<\/td><td>delete the current mailbox (IMAP only)<\/td><\/tr><tr><td><code class=\"literal\">&lt;rename-mailbox&gt;<\/code><\/td><td>r<\/td><td>rename the current mailbox (IMAP only)<\/td><\/tr><tr><td><code class=\"literal\">&lt;subscribe&gt;<\/code><\/td><td>s<\/td><td>subscribe to current mailbox (IMAP only)<\/td><\/tr><tr><td><code class=\"literal\">&lt;unsubscribe&gt;<\/code><\/td><td>u<\/td><td>unsubscribe from current mailbox (IMAP only)<\/td><\/tr><tr><td><code class=\"literal\">&lt;toggle-subscribed&gt;<\/code><\/td><td>T<\/td><td>toggle view all\/subscribed mailboxes (IMAP only)<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.10.\u00a0Pgp Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.11.\u00a0Default Pgp Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Pgp Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;verify-key&gt;<\/code><\/td><td>c<\/td><td>verify a PGP public key<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-name&gt;<\/code><\/td><td>%<\/td><td>view the key's user id<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.11.\u00a0Smime Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.12.\u00a0Default Smime Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Smime Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;verify-key&gt;<\/code><\/td><td>c<\/td><td>verify a PGP public key<\/td><\/tr><tr><td><code class=\"literal\">&lt;view-name&gt;<\/code><\/td><td>%<\/td><td>view the key's user id<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.12.\u00a0Mixmaster Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.13.\u00a0Default Mixmaster Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Mixmaster Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;accept&gt;<\/code><\/td><td>MUTT_ENTER_S<\/td><td>accept the chain constructed<\/td><\/tr><tr><td><code class=\"literal\">&lt;append&gt;<\/code><\/td><td>a<\/td><td>append a remailer to the chain<\/td><\/tr><tr><td><code class=\"literal\">&lt;insert&gt;<\/code><\/td><td>i<\/td><td>insert a remailer into the chain<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete&gt;<\/code><\/td><td>d<\/td><td>delete a remailer from the chain<\/td><\/tr><tr><td><code class=\"literal\">&lt;chain-prev&gt;<\/code><\/td><td>&lt;Left&gt;<\/td><td>select the previous element of the chain<\/td><\/tr><tr><td><code class=\"literal\">&lt;chain-next&gt;<\/code><\/td><td>&lt;Right&gt;<\/td><td>select the next element of the chain<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><div class=\"sect2\"><div class=\"titlepage\"><div><p><\/p><h3 class=\"title\">4.13.\u00a0Editor Menu<\/h3><\/div><\/div><div class=\"table\"><p class=\"title\"><strong>Table\u00a09.14.\u00a0Default Editor Menu Bindings<\/strong><\/p><div class=\"table-contents\"><table class=\"table\" summary=\"Default Editor Menu Bindings\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Function<\/th><th>Default key<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code class=\"literal\">&lt;bol&gt;<\/code><\/td><td>^A<\/td><td>jump to the beginning of the line<\/td><\/tr><tr><td><code class=\"literal\">&lt;backward-char&gt;<\/code><\/td><td>^B<\/td><td>move the cursor one character to the left<\/td><\/tr><tr><td><code class=\"literal\">&lt;backward-word&gt;<\/code><\/td><td>Esc b<\/td><td>move the cursor to the beginning of the word<\/td><\/tr><tr><td><code class=\"literal\">&lt;capitalize-word&gt;<\/code><\/td><td>Esc c<\/td><td>capitalize the word<\/td><\/tr><tr><td><code class=\"literal\">&lt;downcase-word&gt;<\/code><\/td><td>Esc l<\/td><td>convert the word to lower case<\/td><\/tr><tr><td><code class=\"literal\">&lt;upcase-word&gt;<\/code><\/td><td>Esc u<\/td><td>convert the word to upper case<\/td><\/tr><tr><td><code class=\"literal\">&lt;delete-char&gt;<\/code><\/td><td>^D<\/td><td>delete the char under the cursor<\/td><\/tr><tr><td><code class=\"literal\">&lt;eol&gt;<\/code><\/td><td>^E<\/td><td>jump to the end of the line<\/td><\/tr><tr><td><code class=\"literal\">&lt;forward-char&gt;<\/code><\/td><td>^F<\/td><td>move the cursor one character to the right<\/td><\/tr><tr><td><code class=\"literal\">&lt;forward-word&gt;<\/code><\/td><td>Esc f<\/td><td>move the cursor to the end of the word<\/td><\/tr><tr><td><code class=\"literal\">&lt;backspace&gt;<\/code><\/td><td>&lt;Backspace&gt;<\/td><td>delete the char in front of the cursor<\/td><\/tr><tr><td><code class=\"literal\">&lt;kill-eol&gt;<\/code><\/td><td>^K<\/td><td>delete chars from cursor to end of line<\/td><\/tr><tr><td><code class=\"literal\">&lt;kill-eow&gt;<\/code><\/td><td>Esc d<\/td><td>delete chars from the cursor to the end of the word<\/td><\/tr><tr><td><code class=\"literal\">&lt;kill-line&gt;<\/code><\/td><td>^U<\/td><td>delete all chars on the line<\/td><\/tr><tr><td><code class=\"literal\">&lt;quote-char&gt;<\/code><\/td><td>^V<\/td><td>quote the next typed key<\/td><\/tr><tr><td><code class=\"literal\">&lt;kill-word&gt;<\/code><\/td><td>^W<\/td><td>delete the word in front of the cursor<\/td><\/tr><tr><td><code class=\"literal\">&lt;complete&gt;<\/code><\/td><td>&lt;Tab&gt;<\/td><td>complete filename or alias<\/td><\/tr><tr><td><code class=\"literal\">&lt;complete-query&gt;<\/code><\/td><td>^T<\/td><td>complete address with query<\/td><\/tr><tr><td><code class=\"literal\">&lt;buffy-cycle&gt;<\/code><\/td><td>&lt;Space&gt;<\/td><td>cycle among incoming mailboxes<\/td><\/tr><tr><td><code class=\"literal\">&lt;history-up&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll up through the history list<\/td><\/tr><tr><td><code class=\"literal\">&lt;history-down&gt;<\/code><\/td><td>\u00a0<\/td><td>scroll down through the history list<\/td><\/tr><tr><td><code class=\"literal\">&lt;history-search&gt;<\/code><\/td><td>^R<\/td><td>search through the history list<\/td><\/tr><tr><td><code class=\"literal\">&lt;transpose-chars&gt;<\/code><\/td><td>\u00a0<\/td><td>transpose character under cursor with previous<\/td><\/tr><\/tbody><\/table><\/div><\/div><br class=\"table-break\" \/><\/div><\/div>","created_at":"2018-11-07T16:51:23+0000","updated_at":"2018-11-07T16:51:23+0000","annotations":[],"mimetype":"text\/html","reading_time":120,"domain_name":"www.mutt.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":537,"title":"Debugging Rust with gdb","url":"https:\/\/stackoverflow.com\/questions\/24664994\/debugging-rust-with-gdb","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        6&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I'm aware of the <em>debugging Rust<\/em> questions here on StackOverflow and I also used gdb before with Go. However, I'm running into a problem where it seems gdb is unable to locate the debug symbols.<\/p>\n<p>Consider this complex program in <code>main.rs<\/code>?<\/p>\n<pre>pub fn main () {\n    println!(\"run\");\n}\n<\/pre>\n<p>I compile it with debug symbols<\/p>\n<pre>rustc -g main.rs\n<\/pre>\n<p>Then I run gdb<\/p>\n<pre>gdb main\n<\/pre>\n<p>This gives the first clue that something with the loading of debug symbols is not quite right.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/Yy53n.png\" alt=\"enter image description here\" \/><\/p>\n<p>Now when I'm in gdb and type<\/p>\n<pre>list\n<\/pre>\n<p>it leaves me with some C code which isn't what I expect.<\/p>\n<p><img src=\"https:\/\/i.stack.imgur.com\/ONHcO.png\" alt=\"enter image description here\" \/><\/p>\n<p>What am I doing wrong? My gdb version is 7.7 and I'm on OS X 10.9.2 (13C64). My rustc version is rustc 0.11.0-pre (3035d8dfb13077e195eb056568183911c90c1b4b 2014-07-02 21:26:40 +0000)<\/p>\n<p>It may also be helpful to see the output of `gdb --configuration``<\/p>\n<pre>$ gdb --configuration\nThis GDB was configured as follows:\n   configure --host=x86_64-apple-darwin13.1.0 --target=x86_64-apple-darwin13.1.0\n             --with-auto-load-dir=:${prefix}\/share\/auto-load\n             --with-auto-load-safe-path=:${prefix}\/share\/auto-load\n             --with-expat\n             --with-gdb-datadir=\/usr\/local\/share\/gdb (relocatable)\n             --with-jit-reader-dir=\/usr\/local\/lib\/gdb (relocatable)\n             --without-libunwind-ia64\n             --without-lzma\n             --with-python=\/System\/Library\/Frameworks\/Python.framework\/Versions\/2.7\n             --with-separate-debug-dir=\/usr\/local\/lib\/debug (relocatable)\n             --with-zlib\n             --without-babeltrace\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/24664994\/revisions\" title=\"show all edits to this post\">edited Jul 10 '14 at 13:12<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Jul 9 '14 at 23:00&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/288703\/christoph\">Christoph<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            11.2k1567108&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"2\">&#13;\n                                2 Answers&#13;\n                                2&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        4&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Same question, later Rust version (1.0.0-beta), totally different answer:<\/p>\n<p>In GDB, when debugging a Rust executable, <code>break main<\/code> sets a breakpoint in some setup code that is part of the Rust standard library. This isn't what you want.<\/p>\n<p>Instead type: <code>break $YOUR_CRATE::main<\/code>, substituting the name of your program for <code>$YOUR_CRATE<\/code>.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Apr 27 '15 at 17:02&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/94977\/jason-orendorff\">Jason Orendorff<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            30.1k34782&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        3&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Ok, I figured out what was wrong. I have to manually emit the <code>main.o<\/code> file. I thought the <code>-g<\/code> parameter would just cut it.<\/p>\n<p>Now that I run<\/p>\n<pre>rustc -g main.rs --emit=\"obj,link\"\n<\/pre>\n<p>I can run<\/p>\n<pre>gdb main\n<\/pre>\n<p>And everything works like a charme.<\/p>\n<p>I created two aliases for my bash to make things simple:<\/p>\n<pre>alias rd='rustc -g --emit=\"obj,link\"'\ncompile_and_run() {\n     rustc -g --emit=\"obj,link\" $1 &amp;&amp; gdb ${1%.*}\n}\nalias rdr=compile_and_run\n<\/pre>\n<p>Now I can just call <code>rdr main.rs<\/code> and it will start debugging <code>main.rs<\/code> with gdb.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/24677791\/revisions\" title=\"show all edits to this post\">edited Jul 10 '14 at 14:16<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jul 10 '14 at 13:26&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/288703\/christoph\">Christoph<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            11.2k1567108&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-11-07T16:51:23+0000","updated_at":"2018-11-07T16:51:23+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":533,"title":"The Hitchhiker\u2019s Guide to d3.js","url":"https:\/\/medium.com\/@enjalot\/the-hitchhikers-guide-to-d3-js-a8552174733a","content":"<div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"84f2\" class=\"graf graf--h3 graf--leading graf--title\">The Hitchhiker\u2019s Guide to\u00a0d3.js<\/h1><figure id=\"c732\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*2nRwaSWpN28FVF8gBu_uVQ.png\" data-width=\"3144\" data-height=\"2414\" data-action=\"zoom\" data-action-value=\"1*2nRwaSWpN28FVF8gBu_uVQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*2nRwaSWpN28FVF8gBu_uVQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">the d3 learning landscape in all its\u00a0glory<\/figcaption><\/div><\/figure><p id=\"b019\" class=\"graf graf--p graf-after--figure\">The landscape for learning d3 is rich, vast and sometimes perilous. You may be intimidated by the long list of functions in <a href=\"https:\/\/github.com\/d3\/d3\/blob\/master\/API.md\" data-href=\"https:\/\/github.com\/d3\/d3\/blob\/master\/API.md\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">d3\u2019s API documentation<\/a> or paralyzed by choice reviewing the <a href=\"https:\/\/github.com\/d3\/d3\/wiki\/Tutorials\" data-href=\"https:\/\/github.com\/d3\/d3\/wiki\/Tutorials\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">dozens of tutorials<\/a> on the home page. There are over <a href=\"http:\/\/blockbuilder.org\/search\" data-href=\"http:\/\/blockbuilder.org\/search\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">20,000+ d3 examples<\/a> you could learn from, but you never know how approachable any given one will be.<\/p><figure id=\"654e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*C17GW5l4S_W99_52lQMpBQ.png\" data-width=\"2407\" data-height=\"2130\" data-action=\"zoom\" data-action-value=\"1*C17GW5l4S_W99_52lQMpBQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*C17GW5l4S_W99_52lQMpBQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">we are still working on mapping out the whole landscape\u2026<\/figcaption><\/div><\/figure><p id=\"2eb3\" class=\"graf graf--p graf-after--figure\">If all you need is a quick bar or line chart, maybe this article isn\u2019t for you, there are <a href=\"https:\/\/github.com\/wbkd\/awesome-d3#charts\" data-href=\"https:\/\/github.com\/wbkd\/awesome-d3#charts\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">plenty of charting libraries<\/a> out there for that. If you\u2019re into books, check out <a href=\"http:\/\/shop.oreilly.com\/product\/0636920026938.do\" data-href=\"http:\/\/shop.oreilly.com\/product\/0636920026938.do\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Interactive Data Visualization for the Web<\/a> by <a href=\"https:\/\/twitter.com\/alignedleft\" data-href=\"https:\/\/twitter.com\/alignedleft\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Scott Murray<\/a> as a great place to start. <a href=\"https:\/\/www.manning.com\/books\/d3-js-in-action\" data-href=\"https:\/\/www.manning.com\/books\/d3-js-in-action\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">D3.js in Action<\/a> by <a href=\"https:\/\/twitter.com\/Elijah_Meeks\" data-href=\"https:\/\/twitter.com\/Elijah_Meeks\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Elijah Meeks<\/a> is a comprehensive way to go much deeper into some regions of the API.<\/p><p id=\"0336\" class=\"graf graf--p graf-after--p\">This guide is meant to prepare you mentally as well as give you some fruitful directions to pursue. There is a lot to learn besides the d3.js API, both technical knowledge around web standards like HTML, SVG, CSS and JavaScript as well as communication concepts and data visualization principles. Chances are you know something about some of those things, so this guide will attempt to give you good starting points for the things you want to learn more about.<\/p><h3 id=\"902b\" class=\"graf graf--h3 graf-after--p\">Communicate Complex\u00a0Concepts<\/h3><figure id=\"4d45\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*ofBagXi1x0h8TdHF8tc3nA.png\" data-width=\"1838\" data-height=\"812\" data-action=\"zoom\" data-action-value=\"1*ofBagXi1x0h8TdHF8tc3nA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*ofBagXi1x0h8TdHF8tc3nA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/www.r2d3.us\/visual-intro-to-machine-learning-part-1\/\" data-href=\"http:\/\/www.r2d3.us\/visual-intro-to-machine-learning-part-1\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">r2d3.us visual introduction to machine learning<\/a> sets the bar\u00a0high<\/figcaption><\/div><\/figure><p id=\"1dde\" class=\"graf graf--p graf-after--figure\">Before we dive into data visualization principles and technical skills, lets take a second to be aspirational. There are some amazing examples of what\u2019s possible in this medium, whether its a <a href=\"https:\/\/www.google.com\/search?q=new+york+times+d3+interactives&amp;oq=new+york+times+d3+interactives&amp;aqs=chrome..69i57j69i64.6825j0j1&amp;sourceid=chrome&amp;ie=UTF-8\" data-href=\"https:\/\/www.google.com\/search?q=new+york+times+d3+interactives&amp;oq=new+york+times+d3+interactives&amp;aqs=chrome..69i57j69i64.6825j0j1&amp;sourceid=chrome&amp;ie=UTF-8\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">New York Times article<\/a>, <a href=\"http:\/\/www.r2d3.us\/visual-intro-to-machine-learning-part-1\/\" data-href=\"http:\/\/www.r2d3.us\/visual-intro-to-machine-learning-part-1\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">r2d3<\/a>, <a href=\"http:\/\/distill.pub\" data-href=\"http:\/\/distill.pub\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">distill.pub<\/a>, <a href=\"http:\/\/www.datasketch.es\/\" data-href=\"http:\/\/www.datasketch.es\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">datasketch|es<\/a>, <a href=\"https:\/\/pudding.cool\/\" data-href=\"https:\/\/pudding.cool\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">polygraph<\/a>, or <a href=\"http:\/\/ncase.me\/\" data-href=\"http:\/\/ncase.me\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">ncase<\/a>. Be sure to leave a comment with any inspirations I didn\u2019t list here!<\/p><p id=\"62af\" class=\"graf graf--p graf-after--p\">Don\u2019t just look to others though, one of the most important things you can do is set an aspiration for yourself. I found out from <a href=\"https:\/\/medium.com\/@enjalot\/how-do-you-learn-d3-js-ccffc151419b\" data-href=\"https:\/\/medium.com\/@enjalot\/how-do-you-learn-d3-js-ccffc151419b\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">interviewing some of the top data visualization practitioners using d3.js<\/a> that one of the best ways to learn is to set your sights something you really want to build and figure out what you need to build it.<\/p><h3 id=\"a34e\" class=\"graf graf--h3 graf-after--p\">Visual Representations<\/h3><blockquote id=\"7017\" class=\"graf graf--blockquote graf-after--h3\"><div>D3 does not introduce a new visual representation. Unlike <a href=\"http:\/\/processing.org\/\" data-href=\"http:\/\/processing.org\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Processing<\/a>, <a href=\"http:\/\/raphaeljs.com\/\" data-href=\"http:\/\/raphaeljs.com\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Rapha\u00ebl<\/a>, or <a href=\"http:\/\/vis.stanford.edu\/protovis\/\" data-href=\"http:\/\/vis.stanford.edu\/protovis\/\" class=\"markup--anchor markup--blockquote-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Protovis<\/a>, D3\u2019s vocabulary of graphical marks comes directly from web standards: HTML, SVG, and CSS<br \/>\u200a\u2014\u200ahttp:\/\/d3js.org<\/div><\/blockquote><figure id=\"b29a\" class=\"graf graf--figure graf-after--blockquote\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*GaASA7rqnQVDpHQs9BdbqA.png\" data-width=\"921\" data-height=\"293\" data-action=\"zoom\" data-action-value=\"1*GaASA7rqnQVDpHQs9BdbqA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*GaASA7rqnQVDpHQs9BdbqA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Learning d3.js to write charts is like to learning French to write nutrition labels. To be fair, <a href=\"https:\/\/twitter.com\/syntagmatic\" data-href=\"https:\/\/twitter.com\/syntagmatic\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener nofollow\" target=\"_blank\">@syntagmatic<\/a> has made some beautiful <a href=\"http:\/\/bl.ocks.org\/syntagmatic\/2420080\" data-href=\"http:\/\/bl.ocks.org\/syntagmatic\/2420080\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener nofollow\" target=\"_blank\">nutrition visualizations<\/a><\/figcaption><\/div><\/figure><p id=\"154c\" class=\"graf graf--p graf-after--figure\">Charts are just rectangles with shapes inside of them. There are a handful of configurations of those shapes we recognize as common charts or graphs. What d3 provides is a way to define your own visual representations by manipulating graphical marks or creating your own shapes. It makes it easy to add interactions to visuals and declare how your visuals behave. You are here to learn how to express things that aren\u2019t possible to express in any other medium.<\/p><p id=\"077e\" class=\"graf graf--p graf-after--p\">If you want to learn about some of the principles behind the different kinds of marks and graphical representations people use, you can\u2019t go wrong exploring the <a href=\"https:\/\/smile.amazon.com\/Grammar-Graphics-Statistics-Computing\/dp\/0387245448?sa-no-redirect=1\" data-href=\"https:\/\/smile.amazon.com\/Grammar-Graphics-Statistics-Computing\/dp\/0387245448?sa-no-redirect=1\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Grammar of Graphics<\/a>.<\/p><p id=\"254d\" class=\"graf graf--p graf-after--p\">Don\u2019t worry though, there are a ton of creative things you can do with just circles, rectangles and some careful positioning. Start simple, always try to get something to show up on the screen and build up from there.<\/p><h3 id=\"bc62\" class=\"graf graf--h3 graf-after--p\">On the\u00a0web<\/h3><figure id=\"f4ca\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*TwUCVhrN9Xltsj3sDgntUQ.png\" data-width=\"2216\" data-height=\"1056\" data-action=\"zoom\" data-action-value=\"1*TwUCVhrN9Xltsj3sDgntUQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*TwUCVhrN9Xltsj3sDgntUQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/www.visualcinnamon.com\/2016\/04\/svg-beyond-mere-shapes.html\" data-href=\"https:\/\/www.visualcinnamon.com\/2016\/04\/svg-beyond-mere-shapes.html\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">SVG Beyond Mere Shapes<\/a> is an awesome demonstration of web standards graphics manipulation<\/figcaption><\/div><\/figure><p id=\"08b7\" class=\"graf graf--p graf-after--figure\">One of the reasons you want to use d3.js is so you can share your work instantly with anyone who uses a web browser (at least half of the people on Earth!). That means you need to have at least a passing understanding of the medium you\u2019ll be working in, which is HTML5. So before you even start calling d3 API functions you\u2019ll need some basics in SVG, HMTL and CSS. You\u2019ll probably also want to <a href=\"https:\/\/www.w3schools.com\/html\/html5_canvas.asp\" data-href=\"https:\/\/www.w3schools.com\/html\/html5_canvas.asp\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">learn some Canvas<\/a> if you want to be serious about rendering a lot of data (don\u2019t worry, its actually kind of easier to learn than SVG). I recommend <a href=\"https:\/\/medium.freecodecamp.com\/d3-and-canvas-in-3-steps-8505c8b27444\" data-href=\"https:\/\/medium.freecodecamp.com\/d3-and-canvas-in-3-steps-8505c8b27444\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">this is a great intermediate tutorial<\/a> on using d3 with Canvas after you are comfortable with the basics of both.<\/p><p id=\"6bac\" class=\"graf graf--p graf-after--p\">For SVG I recommend starting with this short but sweet <a href=\"http:\/\/alignedleft.com\/tutorials\/d3\/an-svg-primer\" data-href=\"http:\/\/alignedleft.com\/tutorials\/d3\/an-svg-primer\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">SVG primer<\/a> by <a href=\"http:\/\/twitter.com\/alignedleft\" data-href=\"http:\/\/twitter.com\/alignedleft\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Scott Murray<\/a>. Play around with manually creating SVG elements and seeing how they work. Use a tool like <a href=\"http:\/\/blockbuilder.org\" data-href=\"http:\/\/blockbuilder.org\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">BlockBuilder<\/a> to quickly get started without setting up any kind of development environment. You may want to refer to the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/SVG\" data-href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/SVG\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">MDN reference site for SVG<\/a>. Once you\u2019ve mastered the basics, check out <a href=\"https:\/\/www.visualcinnamon.com\/2016\/04\/svg-beyond-mere-shapes.html\" data-href=\"https:\/\/www.visualcinnamon.com\/2016\/04\/svg-beyond-mere-shapes.html\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">SVG beyond mere shapes<\/a> by <a href=\"http:\/\/visualcinnamon.com\" data-href=\"http:\/\/visualcinnamon.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Nadieh Bremer<\/a>.<\/p><figure id=\"fe7e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*PFwvxtqLRRVekNoMufclzw.gif\" data-width=\"1070\" data-height=\"740\" data-action=\"zoom\" data-action-value=\"1*PFwvxtqLRRVekNoMufclzw.gif\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*PFwvxtqLRRVekNoMufclzw.gif\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/blockbuilder.org\" data-href=\"http:\/\/blockbuilder.org\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http:\/\/blockbuilder.org<\/a> makes it easy to play with web standards!<\/figcaption><\/div><\/figure><p id=\"61a0\" class=\"graf graf--p graf-after--figure\">You don\u2019t have to use SVG to make your visualization, it\u2019s relatively common to manipulate HTML elements like <em class=\"markup--em markup--p-em\">&lt;div&gt; <\/em>tags with d3. You will need to be familiar with <a href=\"https:\/\/css-tricks.com\/almanac\/properties\/p\/position\/\" data-href=\"https:\/\/css-tricks.com\/almanac\/properties\/p\/position\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">CSS positioning<\/a> to make it work well. You can even <a href=\"http:\/\/bl.ocks.org\/sxv\/4491174\" data-href=\"http:\/\/bl.ocks.org\/sxv\/4491174\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">mix HTML, SVG and Canvas all at once<\/a>!<\/p><p id=\"5b36\" class=\"graf graf--p graf-after--p\">It can be a bit overwhelming to figure out which rendering system you should use, and even how to use any one of them. I\u2019m going to reiterate the importance of knowing basic HTML, CSS, SVG (and a little Canvas) before getting started with d3.js!<\/p><h3 id=\"efae\" class=\"graf graf--h3 graf-after--p\">Getting started with\u00a0d3.js<\/h3><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"5902\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*KfsnI5vicI0ozs1uP85Pfg.png\" data-width=\"2212\" data-height=\"450\" data-action=\"zoom\" data-action-value=\"1*KfsnI5vicI0ozs1uP85Pfg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*KfsnI5vicI0ozs1uP85Pfg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/d3js.org\" data-href=\"http:\/\/d3js.org\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3js.org<\/a><\/figcaption><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"b1a6\" class=\"graf graf--p graf-after--figure\">How do you build up your visualization from first principles? Piece by piece and a whole lot of utility functions. As you\u2019ve probably seen, <a href=\"https:\/\/github.com\/d3\/d3\/blob\/master\/API.md\" data-href=\"https:\/\/github.com\/d3\/d3\/blob\/master\/API.md\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3\u2019s API<\/a> is massive, so lets call out a few of the utilities that will be particularly helpful early on.<\/p><h4 id=\"175c\" class=\"graf graf--h4 graf-after--p\">d3-scale<\/h4><figure id=\"a3b2\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*c2dJV4ZNJGdWdWNV3AVPxQ.png\" data-width=\"1864\" data-height=\"736\" data-action=\"zoom\" data-action-value=\"1*c2dJV4ZNJGdWdWNV3AVPxQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*c2dJV4ZNJGdWdWNV3AVPxQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/blockbuilder.org\/enjalot\/f1ac6277c9b224ebf4daada75a06294d\" data-href=\"http:\/\/blockbuilder.org\/enjalot\/f1ac6277c9b224ebf4daada75a06294d\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">colors<\/a> are one common use of\u00a0scales!<\/figcaption><\/div><\/figure><p id=\"8455\" class=\"graf graf--p graf-after--figure\">Some of the most fundamental tools in the d3 tool belt are scales. Start with the <a href=\"https:\/\/medium.com\/@mbostock\/introducing-d3-scale-61980c51545f\" data-href=\"https:\/\/medium.com\/@mbostock\/introducing-d3-scale-61980c51545f\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Introducing d3-scale<\/a> post by Mike Bostock to get an overview of what they are and how to use them. No matter what kind of visualization you do, you will likely use at least one kind of scale to make it happen.<\/p><h4 id=\"f4e6\" class=\"graf graf--h4 graf-after--p\">d3-shape<\/h4><figure id=\"1303\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*6HpzsxMbWhLgTbAZNpf9Kg.png\" data-width=\"1914\" data-height=\"806\" data-action=\"zoom\" data-action-value=\"1*6HpzsxMbWhLgTbAZNpf9Kg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*6HpzsxMbWhLgTbAZNpf9Kg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">A <a href=\"http:\/\/bl.ocks.org\/mbostock\/582915\" data-href=\"http:\/\/bl.ocks.org\/mbostock\/582915\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">streamgraph<\/a>, thanks to SVG\u00a0paths!<\/figcaption><\/div><\/figure><p id=\"cf9f\" class=\"graf graf--p graf-after--figure\">SVG paths are pretty intense (<a href=\"https:\/\/css-tricks.com\/svg-path-syntax-illustrated-guide\/\" data-href=\"https:\/\/css-tricks.com\/svg-path-syntax-illustrated-guide\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">see this thorough overview<\/a>), so <a href=\"https:\/\/github.com\/d3\/d3-shape#d3-shape\" data-href=\"https:\/\/github.com\/d3\/d3-shape#d3-shape\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-shape<\/a> includes functions that will make them easier to create and manipulate for some use cases. Read Mike\u2019s <a href=\"https:\/\/medium.com\/@mbostock\/introducing-d3-shape-73f8367e6d12\" data-href=\"https:\/\/medium.com\/@mbostock\/introducing-d3-shape-73f8367e6d12\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Introducing d3-shape<\/a> article to get an overview of what it offers and how to get started using them. d3-shape can also help you render things like lines, areas and arbitrary paths to Canvas with just one extra line of code!<\/p><h4 id=\"4b14\" class=\"graf graf--h4 graf-after--p\">d3-selection<\/h4><figure id=\"e63a\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*-qbObwrIKYG3_bKpZBaUgw.gif\" data-width=\"760\" data-height=\"320\" data-action=\"zoom\" data-action-value=\"1*-qbObwrIKYG3_bKpZBaUgw.gif\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*-qbObwrIKYG3_bKpZBaUgw.gif\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">the <a href=\"https:\/\/bl.ocks.org\/mbostock\/3808234\" data-href=\"https:\/\/bl.ocks.org\/mbostock\/3808234\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">General Update\u00a0Pattern<\/a><\/figcaption><\/div><\/figure><p id=\"9a50\" class=\"graf graf--p graf-after--figure\">One of the hardest parts of d3 to learn is the selection system, also referred to as the <a href=\"https:\/\/bl.ocks.org\/mbostock\/3808218\" data-href=\"https:\/\/bl.ocks.org\/mbostock\/3808218\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">General Update Pattern<\/a>. It took me months of banging my head against my desk to internalize it, but don\u2019t let that scare you! You don\u2019t actually need to be a master of selections to get a ton of cool stuff done. When you are ready to give it a shot, you can start with the <a href=\"https:\/\/github.com\/d3\/d3-selection#d3-selection\" data-href=\"https:\/\/github.com\/d3\/d3-selection#d3-selection\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-selections<\/a> README and be sure to click on the links like <a href=\"https:\/\/bost.ocks.org\/mike\/join\/\" data-href=\"https:\/\/bost.ocks.org\/mike\/join\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Thinking in Joins<\/a>.<\/p><h4 id=\"e673\" class=\"graf graf--h4 graf-after--p\">d3-collection<\/h4><figure id=\"29f5\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*jcnYXBvv1W033jfwFOrS1A.png\" data-width=\"1904\" data-height=\"974\" data-action=\"zoom\" data-action-value=\"1*jcnYXBvv1W033jfwFOrS1A.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*jcnYXBvv1W033jfwFOrS1A.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">d3-nest makes it easy to <a href=\"https:\/\/bl.ocks.org\/mbostock\/9490313\" data-href=\"https:\/\/bl.ocks.org\/mbostock\/9490313\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">group similar things\u00a0together<\/a><\/figcaption><\/div><\/figure><p id=\"5fe9\" class=\"graf graf--p graf-after--figure\">Manipulating data is an incredibly important part of visualizing it. This can often be the hardest part to do depending on how good your data is and how well you understand it. Having more tools that make working with the data by reshaping it, slicing and dicing or aggregating it can really help. To this end I recommend getting familiar with <a href=\"https:\/\/github.com\/d3\/d3-collection\/blob\/master\/README.md#d3-collection\" data-href=\"https:\/\/github.com\/d3\/d3-collection\/blob\/master\/README.md#d3-collection\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-collection<\/a>, particularly the <a href=\"https:\/\/github.com\/d3\/d3-collection\/blob\/master\/README.md#nests\" data-href=\"https:\/\/github.com\/d3\/d3-collection\/blob\/master\/README.md#nests\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">nest<\/a> function.<\/p><h4 id=\"3cda\" class=\"graf graf--h4 graf-after--p\">d3-hierarchy<\/h4><figure id=\"f141\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*1CszAZA3t5oMlTOMFchzSg.png\" data-width=\"1290\" data-height=\"1292\" data-action=\"zoom\" data-action-value=\"1*1CszAZA3t5oMlTOMFchzSg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*1CszAZA3t5oMlTOMFchzSg.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">a <a href=\"https:\/\/bl.ocks.org\/mbostock\/8fe6fa6ed1fa976e5dd76cfa4d816fec\" data-href=\"https:\/\/bl.ocks.org\/mbostock\/8fe6fa6ed1fa976e5dd76cfa4d816fec\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Treemap<\/a> is easy to layout thanks to d3-hierarchy<\/figcaption><\/div><\/figure><p id=\"10ed\" class=\"graf graf--p graf-after--figure\">Continuing the theme of working with data, the key part to many visualizations is laying out your visual representations based on your data\u2019s structure. Some common functions for helping you do this are found in <a href=\"https:\/\/github.com\/d3\/d3-hierarchy#d3-hierarchy\" data-href=\"https:\/\/github.com\/d3\/d3-hierarchy#d3-hierarchy\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-hierarchy<\/a> for making things like trees, treemaps or circle packs.<\/p><h4 id=\"add4\" class=\"graf graf--h4 graf-after--p\">d3-zoom<\/h4><figure id=\"5f35\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*R96BnzJUPFSzTzWoxIBIAw.gif\" data-width=\"958\" data-height=\"482\" data-action=\"zoom\" data-action-value=\"1*R96BnzJUPFSzTzWoxIBIAw.gif\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*R96BnzJUPFSzTzWoxIBIAw.gif\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/bl.ocks.org\/mbostock\/b783fbb2e673561d214e09c7fb5cedee\" data-href=\"https:\/\/bl.ocks.org\/mbostock\/b783fbb2e673561d214e09c7fb5cedee\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">zooming is\u00a0fun<\/a>!<\/figcaption><\/div><\/figure><p id=\"27de\" class=\"graf graf--p graf-after--figure\">A common interaction behavior you may want to add to your visualization is zooming. Mike has <a href=\"http:\/\/blockbuilder.org\/search#text=zoom;user=mbostock;d3version=v4\" data-href=\"http:\/\/blockbuilder.org\/search#text=zoom;user=mbostock;d3version=v4\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">a series of examples<\/a> that show various ways of adding zoom to your visualization with <a href=\"https:\/\/github.com\/d3\/d3-zoom#d3-zoom\" data-href=\"https:\/\/github.com\/d3\/d3-zoom#d3-zoom\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-zoom<\/a>. Check out zoom\u2019s close cousin <a href=\"https:\/\/github.com\/d3\/d3-drag#d3-drag\" data-href=\"https:\/\/github.com\/d3\/d3-drag#d3-drag\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-drag<\/a> for some additional interaction help!<\/p><h4 id=\"a98b\" class=\"graf graf--h4 graf-after--p\">d3-force<\/h4><figure id=\"a0ae\" class=\"graf graf--figure graf-after--h4\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*MiNLItcbqWuZP5y6DHVEsg.gif\" data-width=\"640\" data-height=\"531\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*MiNLItcbqWuZP5y6DHVEsg.gif\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/bl.ocks.org\/syntagmatic\/75c5ca501200b0cf7a5958b4e404f777\" data-href=\"https:\/\/bl.ocks.org\/syntagmatic\/75c5ca501200b0cf7a5958b4e404f777\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">sparse matrices<\/a>, amirite?<\/figcaption><\/div><\/figure><p id=\"44cd\" class=\"graf graf--p graf-after--figure\">One of d3\u2019s capabilities that inspires a lot of people is the force layout. It\u2019s pretty easy to use but definitely difficult to master. Just don\u2019t let it pull you to the dark side! See <a href=\"https:\/\/github.com\/d3\/d3-force#d3-force\" data-href=\"https:\/\/github.com\/d3\/d3-force#d3-force\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">d3-force<\/a> for more details.<\/p><h3 id=\"4afd\" class=\"graf graf--h3 graf-after--p\">Search!<\/h3><p id=\"688f\" class=\"graf graf--p graf-after--h3\">One final tip you can use for any API function is to use <a href=\"http:\/\/blockbuilder.org\/search\" data-href=\"http:\/\/blockbuilder.org\/search\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">BlockBuilder\u2019s search<\/a> capability. You can also limit your searches to examples that use d3 v4!<\/p><figure id=\"27cb\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*AAKuY9w1DNMIq0oBqTe3KA.gif\" data-width=\"984\" data-height=\"428\" data-action=\"zoom\" data-action-value=\"1*AAKuY9w1DNMIq0oBqTe3KA.gif\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*AAKuY9w1DNMIq0oBqTe3KA.gif\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">look at all those\u00a0blocks!<\/figcaption><\/div><\/figure><h3 id=\"36b9\" class=\"graf graf--h3 graf-after--figure\">Community<\/h3><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"0271\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*m3WWkrnN1pr6V-pFZgK8KQ.png\" data-width=\"1412\" data-height=\"765\" data-action=\"zoom\" data-action-value=\"1*m3WWkrnN1pr6V-pFZgK8KQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*m3WWkrnN1pr6V-pFZgK8KQ.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Welcome to the <a href=\"http:\/\/bl.ocks.org\/micahstubbs\/b35f2560f4205570b3328d1b40de0c6c\" data-href=\"http:\/\/bl.ocks.org\/micahstubbs\/b35f2560f4205570b3328d1b40de0c6c\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener nofollow\" target=\"_blank\">Blockiverse<\/a><\/figcaption><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"3af2\" class=\"graf graf--p graf-after--figure\">Connect with like-minded folks! We have a vibrant online community in the <a href=\"https:\/\/d3-slackin.herokuapp.com\/\" data-href=\"https:\/\/d3-slackin.herokuapp.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">d3.js slack channel<\/a>. Find a meetup near you, the SF Bay Area has <a href=\"https:\/\/www.meetup.com\/Bay-Area-d3-User-Group\/\" data-href=\"https:\/\/www.meetup.com\/Bay-Area-d3-User-Group\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">one of the biggest groups<\/a>, but you can probably find some <a href=\"https:\/\/www.meetup.com\/topics\/d3-js\/\" data-href=\"https:\/\/www.meetup.com\/topics\/d3-js\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">groups near you<\/a>. If you\u2019re really dedicated, come to the <a href=\"http:\/\/visfest.com\" data-href=\"http:\/\/visfest.com\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow\" target=\"_blank\">annual d3.unconf<\/a> held in SF every fall!<\/p><\/div><div class=\"section-inner sectionLayout--outsetColumn\"><figure id=\"a685\" class=\"graf graf--figure graf--layoutOutsetCenter graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><img class=\"graf-image\" data-image-id=\"1*gybGRNFfU-ZBYtK3D1qGYA.png\" data-width=\"2550\" data-height=\"3508\" data-action=\"zoom\" data-action-value=\"1*gybGRNFfU-ZBYtK3D1qGYA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*gybGRNFfU-ZBYtK3D1qGYA.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">how I see the d3 community and the many learning curves one encounters on their\u00a0journey<\/figcaption><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"55e6\" class=\"graf graf--p graf-after--figure graf--trailing\"><em class=\"markup--em markup--p-em\">Many thanks to <\/em><a href=\"https:\/\/twitter.com\/erikhazzard\" data-href=\"https:\/\/twitter.com\/erikhazzard\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Erik Hazzard<\/em><\/a><em class=\"markup--em markup--p-em\"> for editing and helping shape this post. Thanks to <\/em><a href=\"https:\/\/twitter.com\/syntagmatic\" data-href=\"https:\/\/twitter.com\/syntagmatic\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Kai Chang<\/em><\/a><em class=\"markup--em markup--p-em\"> for his suggestions in this post but more so for helping shape the d3 community. Thanks to #teaching-d3 in the <\/em><a href=\"http:\/\/d3-slackin.herokuapp.com\" data-href=\"http:\/\/d3-slackin.herokuapp.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">d3js Slack<\/em><\/a><em class=\"markup--em markup--p-em\">, especially <\/em><a href=\"https:\/\/twitter.com\/dashingd3js\" data-href=\"https:\/\/twitter.com\/dashingd3js\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Sebastian<\/em><\/a><em class=\"markup--em markup--p-em\"> and <\/em><a href=\"https:\/\/twitter.com\/JFSIII\" data-href=\"https:\/\/twitter.com\/JFSIII\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">John<\/em><\/a><em class=\"markup--em markup--p-em\"> for their feedback. Of course Infinite thanks to <\/em><a href=\"https:\/\/twitter.com\/mbostock\" data-href=\"https:\/\/twitter.com\/mbostock\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">Mike Bostock<\/em><\/a><em class=\"markup--em markup--p-em\"> for creating the playground we all get to work in!<\/em><\/p><\/div><\/div>","created_at":"2018-11-07T16:51:14+0000","updated_at":"2018-11-07T16:51:14+0000","published_at":"2017-05-12T15:59:31+0000","published_by":["Ian Johnson"],"annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*2nRwaSWpN28FVF8gBu_uVQ.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":534,"title":"Introduction - Slate","url":"https:\/\/docs.slatejs.org\/","content":"<p class=\"_1uc14g3\" data-key=\"a44cdeb26ab94dd0801e4bfe4bcf17e3\">\u200b<a class=\"_1jhzvirl\" href=\"http:\/\/slatejs.org\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"e57377aad180478da2d50e44288935b3\">Slate<\/a> is a <em>completely<\/em> customizable framework for building rich text editors.<\/p><p class=\"_1uc14g3\" data-key=\"1b3d97c03f054ef7a8f412c6bb27df7a\">Slate lets you build rich, intuitive editors like those in <a class=\"_1jhzvirl\" href=\"https:\/\/medium.com\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"ba1dde2563cf46f68f57d2d0314ba481\">Medium<\/a>, <a class=\"_1jhzvirl\" href=\"https:\/\/www.dropbox.com\/paper\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"53ea7573983345b4b4effb898dec88b4\">Dropbox Paper<\/a> or <a class=\"_1jhzvirl\" href=\"https:\/\/www.google.com\/docs\/about\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"bad8934f65794f4a840b5ad60f3c552d\">Google Docs<\/a>\u2014which are becoming table stakes for applications on the web\u2014without your codebase getting mired in complexity.<\/p><p class=\"_1uc14g3\" data-key=\"78091a5787ec481ba50610638c298630\">It can do this because all of its logic is implemented with a series of plugins, so you aren't ever constrained by what <em>is<\/em> or <em>isn't<\/em> in \"core\". You can think of it like a pluggable implementation of <code spellcheck=\"false\" class=\"_146ac1z\">contenteditable<\/code> built on top of <a class=\"_1jhzvirl\" href=\"https:\/\/facebook.github.io\/react\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"8881e4a63ff4465caed2a92cb59e39f1\">React<\/a> and <a class=\"_1jhzvirl\" href=\"https:\/\/facebook.github.io\/immutable-js\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"8b365e9c05bc466daceba502034a7f1a\">Immutable<\/a>. It was inspired by libraries like <a class=\"_1jhzvirl\" href=\"https:\/\/facebook.github.io\/draft-js\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"736a5f663b82414091922283efe5930c\">Draft.js<\/a>, <a class=\"_1jhzvirl\" href=\"http:\/\/prosemirror.net\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"990109e55a324ceaaf09f3b4f9a0f679\">Prosemirror<\/a> and <a class=\"_1jhzvirl\" href=\"http:\/\/quilljs.com\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"f9f74495c4134de0bd68aa885fad37a2\">Quill<\/a>.<\/p><p class=\"_1uc14g3\" data-key=\"e2eca6b483454379924d46bf8581052a\"><strong class=\"_u5j6yl\"><em>Slate is currently in beta<\/em><\/strong><em>. It's useable now, but you might need to pull request a fix or two for advanced use cases.<\/em><\/p><p class=\"_1uc14g3\" data-key=\"d50be950d0344a96b6cdb8325fe5cfbf\">Why create Slate? Well... <em>(Beware: this section has a few of <\/em><a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"2b8e20455d804d11a2467f02269a213c\"><em>my<\/em><\/a><em> opinions!)<\/em><\/p><p class=\"_1uc14g3\" data-key=\"daf4a675493f45f193ca3081e054e2f8\">Before creating Slate, I tried a lot of the other rich text libraries out there\u2014<a class=\"_1jhzvirl\" href=\"https:\/\/facebook.github.io\/draft-js\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"1e376aa0bd824f91b68a0ae0d4229e8e\"><strong class=\"_u5j6yl\">Draft.js<\/strong><\/a>, <a class=\"_1jhzvirl\" href=\"http:\/\/prosemirror.net\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"c4bc957e12fb4b4eab49b077e4f2c892\"><strong class=\"_u5j6yl\">Prosemirror<\/strong><\/a>, <a class=\"_1jhzvirl\" href=\"http:\/\/quilljs.com\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"5cdbe681880649d0880d1b29d53a74dc\"><strong class=\"_u5j6yl\">Quill<\/strong><\/a>, etc. What I found was that while getting simple examples to work was easy enough, once you started trying to build something like <a class=\"_1jhzvirl\" href=\"https:\/\/medium.com\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"1af12ada6aed46349b50d9f06458970b\">Medium<\/a>, <a class=\"_1jhzvirl\" href=\"https:\/\/www.dropbox.com\/paper\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"f03458c6852a4acc93c6a254d9d56e2b\">Dropbox Paper<\/a> or <a class=\"_1jhzvirl\" href=\"https:\/\/www.google.com\/docs\/about\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"aaf02566f3db4bf3b3406b2f35a84da1\">Google Docs<\/a>, you ran into deeper issues...<\/p><ul class=\"_11twlxq\" data-key=\"f95fc5eca58748c5ae211041e483732c\"><li class=\"\"><div data-key=\"176199bf539d41cf828939a58f0a4040\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"57fe8a3023314d9f9f77172db8ef34be\"><strong class=\"_u5j6yl\">The editor's \"schema\" was hardcoded and hard to customize.<\/strong> Things like bold and italic were supported out of the box, but what about comments, or embeds, or even more domain-specific needs?<\/p><\/div><\/li><li class=\"\"><div data-key=\"e023da31c3794af7960ece65427d33f9\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"952fbe6485804477ae94a7dc617866ab\"><strong class=\"_u5j6yl\">Transforming the documents programmatically was very convoluted.<\/strong> Writing as a user may have worked, but making programmatic changes, which is critical for building advanced behaviors, was needlessly complex.<\/p><\/div><\/li><li class=\"\"><div data-key=\"a6db1fe0d7ba4bc49da07e18b9d513dc\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"32b0a379940a4b849787b6e431126a4e\"><strong class=\"_u5j6yl\">Serializing to HTML, Markdown, etc. seemed like an afterthought.<\/strong> Simple things like transforming a document to HTML or Markdown involved writing lots of boilerplate code, for what seemed like very common use cases.<\/p><\/div><\/li><li class=\"\"><div data-key=\"9586501534364536918c29a64a587dff\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"8d6b4397a4614b79bd0c1057188cbb21\"><strong class=\"_u5j6yl\">Re-inventing the view layer seemed inefficient and limiting.<\/strong> Most editors rolled their own views, instead of using existing technologies like React, so you have to learn a whole new system with new \"gotchas\".<\/p><\/div><\/li><li class=\"\"><div data-key=\"6aed801d051e4e9394efcb9da887af22\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"b61179e3d0a94702acedcb17a906be87\"><strong class=\"_u5j6yl\">Collaborative editing wasn't designed for in advance.<\/strong> Often the editor's internal representation of data made it impossible to use to for a realtime, collaborative editing use case without basically rewriting the editor.<\/p><\/div><\/li><li class=\"\"><div data-key=\"59416cd4730b405fb10deb788cc5145d\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"cc1ee1beb62b475ab193bb0b885752a8\"><strong class=\"_u5j6yl\">The repostories were monolithic, not small and reusable.<\/strong> The code bases for many of the editors often didn't expose the internal tooling that could have been re-used by developers, leading to having to reinvent the wheel.<\/p><\/div><\/li><li class=\"\"><div data-key=\"ba928c9018d045fd94caba6b18c23e38\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"b013d2115ab74399b264235fa7bc61b1\"><strong class=\"_u5j6yl\">Building complex, nested documents was impossible.<\/strong> Many editors were designed around simplistic \"flat\" documents, making things like tables, embeds and captions difficult to reason about and sometimes impossible.<\/p><\/div><\/li><\/ul><p class=\"_1uc14g3\" data-key=\"5604a82b35a940269858289f5a257c0f\">Of course not every editor exhibits all of these issues, but if you've tried using another editor you might have run into similar problems. To get around the limitations of their API's and achieve the user experience you're after, you have to resort to very hacky things. And some experiences are just plain impossible to achieve.<\/p><p class=\"_1uc14g3\" data-key=\"85282a533b7044c4b43db49fe042a64e\">If that sounds familiar, you might like Slate.<\/p><p class=\"_1uc14g3\" data-key=\"3d4c302dd0d14c24a5a606712226a2a7\">Which brings me to how Slate solves all of that...<\/p><p class=\"_1uc14g3\" data-key=\"5b12fb99b5e144baa5cccd13a22aef49\">Slate tries to solve the question of \"<a class=\"_1jhzvirl\" data-key=\"d6aef06cfa724f7d89efe84dca406782\" href=\"https:\/\/docs.slatejs.org\/#why\">Why?<\/a>\" with a few principles:<\/p><ol class=\"_11twlxq\" data-key=\"ba3a14fc22124d8ebc465e443bc57aad\"><li class=\"\"><div data-key=\"e3efade090594a959e4793e8d2181af9\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"38286821d3184b4eb9228f90d48c8886\"><strong class=\"_u5j6yl\">First-class plugins.<\/strong> The most important part of Slate is that plugins are first-class entities\u2014the core editor logic is even implemented as its own plugin. That means you can <em>completely<\/em> customize the editing experience, to build complex editors like Medium's or Dropbox's, without having to fight against the library's assumptions.<\/p><\/div><\/li><li class=\"\"><div data-key=\"0fa7e59e50a6483089c4be91f67991af\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"7b4899232b27404dbe7112cc145aa98b\"><strong class=\"_u5j6yl\">Schema-less core.<\/strong> Slate's core logic doesn't assume anything about the schema of the data you'll be editing, which means that there are no assumptions baked into the library that'll trip you up when you need to go beyond the most basic use cases.<\/p><\/div><\/li><li class=\"\"><div data-key=\"39291ecdd31a4ef9be7e8b71dcb80bc2\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"dfda80b3fd614890b6ed835628e84622\"><strong class=\"_u5j6yl\">Nested document model.<\/strong> The document model used for Slate is a nested, recursive tree, just like the DOM itself. This means that creating complex components like tables or nested block quotes are possible for advanced use cases. But it's also easy to keep it simple by only using a single level of hierarchy.<\/p><\/div><\/li><li class=\"\"><div data-key=\"d07bd87a71ac47799679ec90ffcb7080\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"da8cd568b8894484a570bd2ec7027fe3\"><strong class=\"_u5j6yl\">Parallel to the DOM.<\/strong> Slate's data model is based on the DOM\u2014the document is a nested tree, it uses selections and ranges, and it exposes all the standard event handlers. This means that advanced behaviors like tables or nested block quotes are possible. Pretty much anything you can do in the DOM, you can do in Slate.<\/p><\/div><\/li><li class=\"\"><div data-key=\"c2ac31e73fc943618bcd66ca7ddb2306\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"6d2c54d2e273488d8a0cfcb3ef5ca88a\"><strong class=\"_u5j6yl\">Stateless views and immutable data.<\/strong> By using React and Immutable.js, the Slate editor is built in a stateless fashion using immutable data structures, which leads to much easier to reason about code, and a much easier time writing plugins.<\/p><\/div><\/li><li class=\"\"><div data-key=\"213346b3cd184ff79cb6e7c71a47169b\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"122e990a788245f9a9b6789e63384220\"><strong class=\"_u5j6yl\">Intuitive changes.<\/strong> Slate documents are edited using \"changes\", that are designed to be high-level and extremely intuitive to write and read, so that custom functionality is as expressive as possible. This greatly increases your ability to reason about your code.<\/p><\/div><\/li><li class=\"\"><div data-key=\"db58af674d9249eea26cabc95924a48e\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"91cc50b27f1a46f7b0568a49b41884bd\"><strong class=\"_u5j6yl\">Collaboration-ready data model.<\/strong> The data model Slate uses\u2014specifically how changes are applied to the document\u2014has been designed to allow for collaborative editing to be layered on top, so you won't need to rethink everything if you decide to make your editor collaborative.<\/p><\/div><\/li><li class=\"\"><div data-key=\"abb1880584e84a47a12cee01364c4797\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"b4c56726bbe44af0ae6153d567013ba6\"><strong class=\"_u5j6yl\">Clear \"core\" boundaries.<\/strong> With a plugin-first architecture, and a schema-less core, it becomes a lot clearer where the boundary is between \"core\" and \"custom\", which means that the core experience doesn't get bogged down in edge cases.<\/p><\/div><\/li><\/ol><p class=\"_1uc14g3\" data-key=\"a5d07628d43449a1baed78c3342ff27e\">Check out the <a class=\"_1jhzvirl\" href=\"http:\/\/slatejs.org\/\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"06add4f993544d5cbb8e2d867752038c\"><strong class=\"_u5j6yl\">live demo<\/strong><\/a> of all of the examples!<\/p><p class=\"_1uc14g3\" data-key=\"cee2cef32c2d48d5aae3a51f93a0c533\">To get a sense for how you might use Slate, check out a few of the examples:<\/p><ul class=\"_11twlxq\" data-key=\"12453c53e6f84ebdb440e3f60fc52cd2\"><li class=\"\"><div data-key=\"871243acafce45c8ac368e95e0dcdf25\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"e311707053cd4c8a8737ca23d15add4b\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/plain-text\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"ebcfa13d8234454f8d0fab334e0cf5ce\"><strong class=\"_u5j6yl\">Plain text<\/strong><\/a> \u2014 showing the most basic case: a glorified <code spellcheck=\"false\" class=\"_146ac1z\">&lt;textarea&gt;<\/code>.<\/p><\/div><\/li><li class=\"\"><div data-key=\"13d987d41bda41f89a3e54a712b297f8\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"bdb3924d220847b8bb80cff4e79b5f67\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/rich-text\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"bc9ccdce3be942cd9fe4f96449e0e889\"><strong class=\"_u5j6yl\">Rich text<\/strong><\/a> \u2014 showing the features you'd expect from a basic editor.<\/p><\/div><\/li><li class=\"\"><div data-key=\"ea458ef90f8c46bab7523bc5d6a32e2d\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"8e0dbb6f26c44c8ca46c703fdc280fc0\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/markdown-preview\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"6e665b69250847e88c65a001a791ccf6\"><strong class=\"_u5j6yl\">Auto-markdown<\/strong><\/a> \u2014 showing how to add key handlers for Markdown-like shortcuts.<\/p><\/div><\/li><li class=\"\"><div data-key=\"e8d13acc528e422a9b8dc6371bb3995e\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"9d4f491a6ff5496b816428601244d141\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/links\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"5aad7f650df2481a81258b3e7b6c5bee\"><strong class=\"_u5j6yl\">Links<\/strong><\/a> \u2014 showing how wrap text in inline nodes with associated data.<\/p><\/div><\/li><li class=\"\"><div data-key=\"11b5eb4e6e044e86a6cf6851e432dc85\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"8db353c171ca4d3ba95759558477b3b1\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/images\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"7398e28261194435b0d84f919038cc25\"><strong class=\"_u5j6yl\">Images<\/strong><\/a> \u2014 showing how to use void (text-less) nodes to add images.<\/p><\/div><\/li><li class=\"\"><div data-key=\"af652f8a12304ba891cd34b022f5ba3b\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"eefec8d0cc9f4c9b8493edaf83ddf54b\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/hovering-menu\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"4fa9f4f0237e4e249d40d6c132c49f97\"><strong class=\"_u5j6yl\">Hovering menu<\/strong><\/a> \u2014 showing how a contextual hovering menu can be implemented.<\/p><\/div><\/li><li class=\"\"><div data-key=\"4e8b8852d75849fb929154ab64efa364\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"52762c164493499fa0c64511b7e30aa2\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/tables\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"2aad62f7fc534723b1fdb704c601132f\"><strong class=\"_u5j6yl\">Tables<\/strong><\/a> \u2014 showing how to nest blocks to render more advanced components.<\/p><\/div><\/li><li class=\"\"><div data-key=\"045337444c35497e894052187c3bd101\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"c2a162d4d7484c2fbe5150aa2823d8bb\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/paste-html\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"279083f907954c429ce3300ef548133a\"><strong class=\"_u5j6yl\">Paste HTML<\/strong><\/a> \u2014 showing how to use an HTML serializer to handle pasted HTML.<\/p><\/div><\/li><li class=\"\"><div data-key=\"aba85a3d27d646beb8ddb36701b75649\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"1b81e81c63a14eea9e3ea127c14f5a97\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/master\/examples\/code-highlighting\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"873b73e82cc2406f94cd15a00b8f63e9\"><strong class=\"_u5j6yl\">Code Highlighting<\/strong><\/a> \u2014 showing how to use decorations to dynamically mark text.<\/p><\/div><\/li><\/ul><p class=\"_1uc14g3\" data-key=\"9432a1bba2a84f00b342d229cbc64519\">If you have an idea for an example that shows a common use case, pull request it!<\/p><p class=\"_1uc14g3\" data-key=\"3ad05ac43e2444dab88189c30738eef4\">Slate encourages you to write small, reusable modules. Check out the public ones you can use in your project!<\/p><ul class=\"_11twlxq\" data-key=\"b17752ce9a9d40d195d6a3e5fa1b174c\"><li class=\"\"><div data-key=\"c8e0915d98aa4a35876b78bdcc5ad8ef\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"55e4cceb5b0a437a95fb7c7bacc7a747\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate-auto-replace\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"74edf87a8d4848b0b21c625d4ac9f2f5\"><code spellcheck=\"false\" class=\"_146ac1z\">slate-auto-replace<\/code><\/a> auto-replaces text as the user types. Useful for \"smart\" typography!<\/p><\/div><\/li><li class=\"\"><div data-key=\"b8448ea762824a678dc29a0a15c5acd6\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"35fb221efc6b45b4a0a5c8df2ceb1a6a\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate-collapse-on-escape\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"702bbb39856144ef8957fb4ea9949bfe\"><code spellcheck=\"false\" class=\"_146ac1z\">slate-collapse-on-escape<\/code><\/a> simply collapses the selection when <code spellcheck=\"false\" class=\"_146ac1z\">escape<\/code> is pressed.<\/p><\/div><\/li><li class=\"\"><div data-key=\"4511a8ad1016419f888a3eb9ece48182\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"7336b6ea120343cda61f7d13e686371d\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate-paste-linkify\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"675ac2e587804f868da25a8e704c403d\"><code spellcheck=\"false\" class=\"_146ac1z\">slate-paste-linkify<\/code><\/a> wraps the selected text in a link when a URL is pasted from the clipboard.<\/p><\/div><\/li><li class=\"\"><div data-key=\"1a8262987fa242f0933e3351b9b55b9d\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"ce25e063f03745d4979f44c8d99a9d65\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate-soft-break\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"3ca64045268f401190b880c256459c5d\"><code spellcheck=\"false\" class=\"_146ac1z\">slate-soft-break<\/code><\/a> adds a soft break when <code spellcheck=\"false\" class=\"_146ac1z\">enter<\/code> is pressed.<\/p><\/div><\/li><li class=\"\"><div data-key=\"7ea94a1587654c01b3b20d016ed77f05\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"f97f2afb55564690a7b3fd981fc6a5a5\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate-drop-or-paste-images\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"e5649d57094b4e2b9de63934d32d9f08\"><code spellcheck=\"false\" class=\"_146ac1z\">slate-drop-or-paste-images<\/code><\/a> lets users drop or paste images to insert them!<\/p><\/div><\/li><li class=\"\"><div data-key=\"61cd5c9b14ed4f5b819680d3d8aa23f9\" class=\"_17e8c6vg\"><p class=\"_ued81cg\" data-key=\"fdf3a758055d413180f2e848c755217e\">\u200b<a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/docs\/general\/plugins.md\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"5a091445274d47d7ae68e8df0c208453\"><strong class=\"_u5j6yl\">View all plugins...<\/strong><\/a>\u200b<\/p><\/div><\/li><\/ul><p class=\"_1uc14g3\" data-key=\"8eaa53139fc247c89ef0be3ab9918316\">If you're using Slate for the first time, check out the <a class=\"_1jhzvirl\" href=\"http:\/\/docs.slatejs.org\/walkthroughs\/installing-slate\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"ea7008e4c1994d68ab5d963f07806a73\">Getting Started<\/a> walkthroughs and the <a class=\"_1jhzvirl\" href=\"http:\/\/docs.slatejs.org\/guides\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"f4e8b1d6e4434a4ca94909796e3168f5\">Guides<\/a> to familiarize yourself with Slate's architecture and mental models. Once you've gotten familiar with those, you'll probably want to check out the full <a class=\"_1jhzvirl\" href=\"http:\/\/docs.slatejs.org\/slate-core\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"ac827b6064964cbca7deeefabdd72f69\">API Reference<\/a>.<\/p><p class=\"_1uc14g3\" data-key=\"e585ec71c8ad4de2ba0a22ec3e4ca851\">If even that's not enough, you can always <a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/7a71de387cebd17dd5814409b16c994bbb0e53a1\/docs\/src\/README.md\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"afc4f07d30024cd38de30b19fdfbc4ae\">read the source itself<\/a>, which is heavily commented.<\/p><p class=\"_1uc14g3\" data-key=\"fab591ff13c043308f1bd9ae05e53ba7\">There are also translations of the documentation into other languages:<\/p><p class=\"_1uc14g3\" data-key=\"1e5af0d65c1f4953a69b2fa785a2cecc\">If you're maintaining a translation, feel free to pull request it here!<\/p><p class=\"_1uc14g3\" data-key=\"92583c5ced624ac88b1819a2f553cb0a\">All contributions are super welcome! Check out the <a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/7a71de387cebd17dd5814409b16c994bbb0e53a1\/docs\/Contributing.md\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"a8d997b98d604ec6a876465d7721bf5e\">Contributing instructions<\/a> for more info!<\/p><p class=\"_1uc14g3\" data-key=\"71ec9a54bda248e6af35578d5dd0f205\">Slate is <a class=\"_1jhzvirl\" href=\"https:\/\/github.com\/ianstormtaylor\/slate\/tree\/7a71de387cebd17dd5814409b16c994bbb0e53a1\/docs\/License.md\" target=\"_blank\" rel=\"noopener noreferrer\" data-key=\"971c9675fdf7472e867162826cbf1bd9\">MIT-licensed<\/a>.<\/p>","created_at":"2018-11-07T16:51:14+0000","updated_at":"2018-11-07T16:51:14+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"docs.slatejs.org","preview_picture":"https:\/\/www.gitbook.com\/share\/space\/thumbnail\/-Kuki-R9UbW5PP_Bubk1.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":535,"title":"Switching to the Mutt Email Client \u00ab null program","url":"https:\/\/nullprogram.com\/blog\/2017\/06\/15\/","content":"<time datetime=\"2017-06-15\">\n    June 15, 2017\n  <\/time><p>nullprogram.com\/blog\/2017\/06\/15\/<\/p><p><em>Note: The way I manage my email wouldn\u2019t really work for most people,\nso don\u2019t read this as a recommendation. This is just a discussion of\nhow I prefer to use email.<\/em><\/p><p>It was almost four years ago I switched from webmail to a customized\nemail configuration <a href=\"https:\/\/nullprogram.com\/blog\/2013\/09\/03\/\">based on Notmuch and Emacs<\/a>. Notmuch served\nas both as a native back-end that provided indexing and tagging, as well\nas a front-end, written in Emacs Lisp. It dramatically improved my email\nexperience, and I wished I had done it earlier. I\u2019ve really enjoyed\nhaving so much direct control over my email.<\/p><p>However, I\u2019m always fiddling with things \u2014 fiddling feels a lot more\nproductive than it actually is \u2014 and last month I re-invented my email\nsituation, this time switching to a combination of <a href=\"http:\/\/www.mutt.org\/\">Mutt<\/a>,\n<a href=\"http:\/\/www.vim.org\/\">Vim<\/a>, <a href=\"https:\/\/www.djcbsoftware.nl\/code\/mu\/\">mu<\/a>, and <a href=\"https:\/\/tmux.github.io\/\">tmux<\/a>. The entirety of my email\ninterface now resides inside a terminal, and I\u2019m enjoying it even more.\nI feel I\u2019ve \u201cleveled up\u201d again in my email habits.<\/p><p><a href=\"https:\/\/nullprogram.com\/img\/screenshot\/mutt.png\"><img src=\"https:\/\/nullprogram.com\/img\/screenshot\/mutt-thumb.png\" alt=\"\" \/><\/a><\/p><p>On the server-side I also switched from Exim to Postfix and procmail,\nmaking the server configuration a whole lot simpler. Including\nSpamAssassin, it\u2019s just three lines added to the default Debian\nconfiguration. It leaves a lot less room for error, and I could rebuild\nit from scratch with little trouble if there was an emergency. My\nprevious configuration required quite a bit of system configuration,\nsuch as relying on <a href=\"http:\/\/inotify.aiken.cz\/?section=incron&amp;page=about\">incron<\/a> to sort incoming mail, particularly\nspam, but procmail now does this job more cleanly.<\/p><h3 id=\"towards-robustness\">Towards Robustness<\/h3><p>Over the years I\u2019ve gotten less patient when it comes to <a href=\"https:\/\/www.youtube.com\/watch?v=oyLBGkS5ICk\">dealing with\nbreaking changes<\/a> in software, and I\u2019ve gotten more conservative\nabout system stability. Continuously updating my configurations and\nhabits to the latest software changes was an interesting challenge\nearlier in my career, but today there are much better uses of my time.\nDebian Stable, my preferred operating system, runs at pretty much the\nperfect pace for me.<\/p><p>Following these changing preferences, one of the biggest motivations for\nmy recent email change was to make my email setup more robust and\nstable. Until now, email was tied tightly to Emacs, with a configuration\ndrawing directly from <a href=\"https:\/\/melpa.org\/\">MELPA<\/a>, pulling in the bleeding edge\nversion of every package I use. Breaking changes arrive at unexpected\ntimes, and occasionally the current version of a package temporarily\ndoesn\u2019t work. Usually it\u2019s because the developer pushed a bad commit\nright before the latest MELPA build, and so the package is broken for a\nfew hours or days. I\u2019ve been <a href=\"https:\/\/github.com\/skeeto\/elfeed\/issues\/202\">guilty of this myself<\/a>. MELPA\nStable is intended to address these issues, but it seems to break more\noften than normal MELPA. For example, at the time of this writing,\n<a href=\"https:\/\/github.com\/emacs-evil\/evil\">Evil<\/a> is not installable via MELPA Stable due to an unmet\ndependency.<\/p><p>Tying something as vital as email to this Rube Goldberg machine made me\nnervous. Access to my email depended on a number of independent systems\nof various levels of stability to mostly work correctly. My switch to\nMutt cut this down to just a couple of very stable systems.<\/p><h3 id=\"formatflowed\">format=flowed<\/h3><p>I\u2019ve long believed HTML email is an abomination that should never have\nbeen invented. Text is the ideal format for email, and there are a\nnumber of specifications to make it work well across different systems.\nOne of those standards is <a href=\"https:\/\/tools.ietf.org\/html\/rfc3676\">RFC 3676<\/a>, colloquially named\n<a href=\"https:\/\/joeclark.org\/ffaq.html\">format=flowed<\/a>, or just f=f.<\/p><p>Messages encoded with f=f allow mail clients to safely reflow the\nparagraphs to nicely fit the user\u2019s display, whether that display be\nthinner or wider than the sender\u2019s original message. It\u2019s also\ncompletely compatible with mail clients that don\u2019t understand\nformat=flowed, which will display the message as the sender originally\nwrapped it.<\/p><p>The gist of f=f is that messages can have both \u201csoft\u201d and \u201chard\u201d line\nbreaks. If a line ends with a space, then it\u2019s a soft line break. The\nmail client can safely reflow lines separated by a soft line break.\nWithout the trailing space, it\u2019s a hard line break, which prohibits\nflowing with the next line. The last line of a paragraph ends with a\nhard line break. It\u2019s also used for text that shouldn\u2019t reflow, such as\ncode samples.<\/p><p>I\u2019ll illustrate using an underscore in place of a space, so that you can\nsee it:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>This is a message in the format=flowed style, allowing_\nmail clients to flow this message nicely in displays of_\ndifferent widths.\n&gt; This is an example of a quote block in a message,_\n&gt; which is supported by the format=flowed specification.\n&gt;&gt; It also supports nested quote blocks, which means_\n&gt;&gt; this paragraph won't flow into the previous.\n<\/pre><\/div><\/div><p>The RFC covers edge cases that require special \u201cspace-stuffing\u201d rules,\nbut, when editing a text email in an editor, you only need to think\nabout soft and hard line breaks. In my case, Mutt takes care of the rest\nof the details.<\/p><p>Unfortunately <a href=\"https:\/\/www.emacswiki.org\/emacs\/GnusFormatFlowed\">Emacs\u2019s lacks decent support for f=f<\/a>, though\nI\u2019m sure a minor mode could be written to make it work well. On the\nother hand, Vim has been <a href=\"https:\/\/nullprogram.com\/blog\/2017\/04\/01\/\">playing an increasing role in my day-to-day\nediting<\/a>, and it has excellent built-in support for f=f. Since I\u2019m\nnow using Vim to compose all of my email, I get it for free.<\/p><p>First, I tell Mutt that I want to use f=f in my <code class=\"highlighter-rouge\">.muttrc<\/code>:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>set text_flowed\n<\/pre><\/div><\/div><p>Then in Vim, I add the <code class=\"highlighter-rouge\">w<\/code> flag to <code class=\"highlighter-rouge\">formatoptions<\/code>, which tells it to\nwrap paragraphs using soft line breaks.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>set fo+=w\n<\/pre><\/div><\/div><p>If I want to inspect my f=f formatting, I temporarily enable the <code class=\"highlighter-rouge\">list<\/code>\noption, which displays a <code class=\"highlighter-rouge\">$<\/code> for all newlines.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>set list\n<\/pre><\/div><\/div><p>Although few people would notice a difference, I feel a little bad for\n<em>not<\/em> using f=f all these years! A few people may have endured some\nugly, non-flowing emails from me. My only condolance is that at least it\nwasn\u2019t HTML.<\/p><p>It\u2019s not all roses, though. When I reply to a message, Mutt doesn\u2019t\ninsert the quoted text as f=f into my reply, so I have to massage it\ninto f=f myself. Also, just as GitHub doesn\u2019t support Markdown in email\nresponses, neither does it support f=f. When I reply to issues by email,\nGitHub won\u2019t nicely reflow my carefully crafted f=f message, needlessly\nmaking email responses an inferior option.<\/p><h3 id=\"features-unneeded\">Features unneeded<\/h3><p>One reason I didn\u2019t choose this particular email arrangement 4 years ago\nwas that PGP support was one of my prime requirements. Mutt has solid\nPGP support, but, with a Maildir setup (i.e. not IMAP), I\u2019d have to use\nthe key on the server, which was out of the question. Since <a href=\"https:\/\/nullprogram.com\/blog\/2017\/03\/12\/\">I no longer\ncare about PGP<\/a>, my email requirements are more relaxed.<\/p><p>Over the years wasn\u2019t making much use of Notmuch\u2019s tagging system. I\nonly used two tags: \u201cunread\u201d and \u201cinbox\u201d (e.g. read, but still needs\nattention). Otherwise I\u2019d use Notmuch\u2019s powerful search to find what I\nwanted. I still needed to keep track of the tags I was using, so the\nNotmuch index, nearly as large as the email messages themselves, became\npart of my mail backup.<\/p><p>The <a href=\"https:\/\/cr.yp.to\/proto\/maildir.html\">Maildir format<\/a> itself supports some flags: passed (P),\nreplied (R), seen (S), trashed (T), draft (D), and flagged (F). These\nare stored in the message\u2019s filename. In my new configuration, the\n\u201cseen\u201d tag (inversely) takes the place of Notmuch\u2019s \u201cunread\u201d tag. The\n\u201cflagged\u201d tag takes place of the \u201cinbox\u201d tag. Normally in Mutt you\u2019d use\nmailboxes \u2014 i.e. Maildir subdirectories \u2014 for something like this, but I\nprefer all my mail to sit in one big bucket. Search, don\u2019t sort.<\/p><p>Since the two flags are part of the filename, I no longer need to\ninclude a tag database (i.e. the entire Notmuch index) in the backup,\nand my mail backups are much smaller. I could continue to use Notmuch\nfor searching, but I\u2019ve settled on mu instead. When I perform a search,\nmu writes the results to a temporary Maildir using symbolic links, which\nI visit with Mutt. The mu index is transient and doesn\u2019t need to be\nbacked up.<\/p><p>Mu also manages my contacts alias list. It can produce a Mutt-style\nalias file based on the contents of my Maildir:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>mu cfind --format=mutt-alias &gt; aliases\n<\/pre><\/div><\/div><p>It\u2019s been really nice to have all my email sitting around as nothing\nmore than a big pile of files like this. I\u2019ve begun <a href=\"https:\/\/docs.python.org\/3.4\/library\/email.html\">writing little\nscripts to harvest data<\/a> from it, too.<\/p><h3 id=\"configuration-files\">Configuration files<\/h3><p>As with all my personal configuration files, you can <a href=\"https:\/\/github.com\/skeeto\/dotfiles\/blob\/master\/_muttrc\">see my .muttrc\nonline<\/a>. The first few weeks I was tweaking this file hourly,\nbut I\u2019ve now got it basically the way I want.<\/p>","created_at":"2018-11-07T16:51:14+0000","updated_at":"2018-11-07T16:51:14+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"nullprogram.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":529,"title":"Polyglot Programming blog by Piotr Klibert","url":"https:\/\/klibert.pl\/posts\/adding_string_interpolation_to_racket.html","content":"<header><p>\n    <a href=\"https:\/\/en.wikipedia.org\/wiki\/String_interpolation\">String\n    interpolation<\/a> is a convenient language feature which makes creating\n    human-readable strings that contain variables easier. The classical approach\n    to this in languages without string interpolation is something like <code>printf<\/code>\n    - a function which accepts a format string and a variable number of\n    arguments. The format string contains special sequences of characters which\n    get replaced with the values of passed arguments, like this:\n  <\/p><div class=\"code\" data-lang=\"c\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><\/td><td class=\"code\"><div class=\"highlight\"><pre>printf(\"An error occured at %s:%d.\", \"file-name.c\", 42);\n\/\/ Displays: An error occured at file-name.c:42.\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    This has various problems, for example the possibility to supply more (or\n    fewer) arguments than expected. String interpolation allows you to embed\n    variables directly into the format string:\n  <\/p><div class=\"code\" data-lang=\"bash\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><\/td><td class=\"code\"><div class=\"highlight\"><pre>LINE=42\nFILE=\"file-name.c\"\necho \"An error occured at $FILE:$LINE.\"\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    As with many features, some languages provide string interpolation and some\n    don't. In the dark and troubled past of JavaScript, for example, some people\n    switched to CoffeeScript for its string interpolation support (among other\n    things.) More recently, string interpolation came to Python in the form of\n    f-strings (<a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0498\/\">PEP-498<\/a>)\n    and to plain JavaScript in the form\n    of <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Template_literals\">template\n    literals<\/a> (supported\n    by <a href=\"http:\/\/babeljs.io\/docs\/plugins\/transform-es2015-template-literals\/\">Babel<\/a>\n    and other transpilers). While the feature is most often associated with\n    dynamic, dynamically typed languages, some statically typed ones, for\n    example <a href=\"http:\/\/stackoverflow.com\/questions\/9354154\/how-do-i-interpolate-strings#26977877\">C#<\/a>\n    and <a href=\"http:\/\/stackoverflow.com\/questions\/25595542\/string-interpolation-in-swift\">Swift<\/a>,\n    also support it, so it's not exclusive to the former.\n  <\/p><h2>Help! No string interpolation in Racket!<\/h2><aside>\n    Actually, Racket has more than enough tools for printing formatted output.\n    The most powerful of those would be <code>at-exp<\/code> language, part of Scribble but\n    also available in normal Racket. See <a href=\"https:\/\/docs.racket-lang.org\/scribble\/reader-internals.html#%28mod-path._at-exp%29\">the docs<\/a>\n  <\/aside><p>\n    Racket - like JS at one point - lacks the string interpolation support. If\n    Racket was JS, the only options would be to either write a transpiler, or\n    switch to another language. Both choices normally require a lot of work, so\n    developers tend to learn to live without the feature instead and move on.\n  <\/p><p>\n    But Racket <strong>is not<\/strong> JS, it's a Lisp. All Lisps\n    are <i>programmable programming languages<\/i><sup><a class=\"tooltip\" id=\"ft_note_adding_string_interpolation_to_racket_html_ft_1\" data-tooltip-content=\"#ft_note_adding_string_interpolation_to_racket_html_1\">[1]<\/a><\/sup>, but Racket takes it up to eleven\n    with the multidute of tools for language extension (and creation)<sup><a class=\"tooltip\" id=\"ft_note_adding_string_interpolation_to_racket_html_ft_2\" data-tooltip-content=\"#ft_note_adding_string_interpolation_to_racket_html_2\">[2]<\/a><\/sup>. How\n    much work would it be to make Racket interpolate strings?\n  <\/p><h2>How Racket processes code<\/h2><p>\n    To process any code, Racket performs these steps:\n  <\/p><ul class=\"points\"><li>\n      <strong>reading<\/strong> - accomplished by two functions, <code>read<\/code> and\n      <code>read-syntax<\/code>, reading means taking a string and returning an\n      AST<sup><a class=\"tooltip\" id=\"ft_note_adding_string_interpolation_to_racket_html_ft_3\" data-tooltip-content=\"#ft_note_adding_string_interpolation_to_racket_html_3\">[3]<\/a><\/sup> (which\n      happens to be a normal list with normal symbols, strings, numbers, etc.\n      Such a list can be wrapped with a <i>syntax object<\/i>, which adds some\n      meta-data to the raw list. You can always get a raw list out of syntax\n      object with <code>syntax-&gt;datum<\/code>).\n    <\/li>\n    <li>\n      <strong>expansion<\/strong> - all the special forms and macros in the read\n      code are expanded until only the most basic forms (modules, function\n      applications, value literals) are left.\n    <\/li>\n    <li>\n      <strong>compilation<\/strong> - Racket performs byte-code compilation on the expanded\n      code. This is where most of the optimizations are applied.\n    <\/li>\n    <li>\n      <strong>execution<\/strong> - compiled byte-code is handed to the VM for execution. At this\n      stage, on some architectures, further JIT compilation happens, turning\n      (possibly portions of) bytecode into native code.\n    <\/li>\n  <\/ul><p>\n    The first two steps are completely customizable: code expansion\n    with <a href=\"https:\/\/docs.racket-lang.org\/reference\/Macros.html\">macros<\/a>,\n    and code reading via hooks into the\n    Racket <a href=\"https:\/\/docs.racket-lang.org\/guide\/hash-reader.html#%28tech._readtable%29\">parser<\/a>\n    (you can also write a new parser from scratch (or generate it) if you want,\n    but frequently it's not necessary). Symbolically, what happens during\n    reading and expansion can be presented<sup><a class=\"tooltip\" id=\"ft_note_adding_string_interpolation_to_racket_html_ft_4\" data-tooltip-content=\"#ft_note_adding_string_interpolation_to_racket_html_4\">[4]<\/a><\/sup>\n    like this:\n  <\/p><div class=\"post-image\"><a href=\"https:\/\/klibert.pl\/statics\/racket-reader-expander.png?cache_busting_cookie=1\"><img src=\"https:\/\/klibert.pl\/statics\/racket-reader-expander.png?cache_busting_cookie=1\" alt=\"image\" \/><\/a><\/div><h2>Extending the parser<\/h2><p>\n    The Racket parser is a simple, recursive-descent one and it works by\n    utilizing a <code>readtable<\/code> - a mapping from characters to functions responsible\n    for parsing what follows the given character. The <code>read<\/code> and <code>read-syntax<\/code>\n    functions utilize <code>current-readtable<\/code>, which is\n    a <a href=\"http:\/\/docs.racket-lang.org\/reference\/parameters.html\">parameter<\/a>\n    you can override. Adding a new syntax to the reader requires just three\n    steps:\n  <\/p><ol><li>create a function for parsing the new syntax into valid Racket forms<\/li>\n    <li>get the <code>current-readtable<\/code> and add your function to it<\/li>\n    <li>make that expanded readtable from the prev step the new <code>current-readtable<\/code><\/li>\n  <\/ol><h2>Interpolated string syntax<\/h2><p>\n    We'll get to writing the parsing function in a minute, but first we need to\n    know what is it we're going to parse and what the results of the parsing\n    should be. The exact syntax I have in mind would look like this (with <code>#@<\/code>\n    distinguishing interpolated strings from normal ones):\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><\/td><td class=\"code\"><div class=\"highlight\"><pre>#@\"An error occured at @{file}:@{line}.\"\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><h2>Interpolated string compilation<\/h2><p>\n    String interpolation, for example in CoffeeScript, is often compiled to an\n    expression which concatenates the chunks of the string, like this:\n  <\/p><div class=\"code\" data-lang=\"coffee\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><\/td><td class=\"code\"><div class=\"highlight\"><pre>\"An error occured at #{file}:#{line}.\"\n# would be compiled to:\n\"An error occured at \" + file + \":\" + line + \".\"\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    In Racket we have no <code>+<\/code> operator for concatenating the strings and even if\n    we did have it, the prefix notation and resulting nesting wouldn't be\n    pretty. The easier and more natural way is to use a <code>string-append<\/code>\n    procedure, which - conveniently - accepts any number of arguments. The\n    syntax introduced in the previous section would compile down to something\n    like this:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><\/td><td class=\"code\"><div class=\"highlight\"><pre>(string-append \"An error occured at \" file \":\" line \".\")\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><h2>The parsing function<\/h2><p>\n    The parsing function needs to take the string form as an argument and output\n    the code as above. Here's the simplest implementation I could think of:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\"> 1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\"> 2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\"> 3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\"> 4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\"> 5<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-6\"> 6<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-7\"> 7<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-8\"> 8<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-9\"> 9<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-10\">10<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-11\">11<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-12\">12<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-13\">13<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-14\">14<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-15\">15<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-16\">16<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-17\">17<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-18\">18<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-19\">19<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-20\">20<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-21\">21<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-22\">22<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>(define (parse str)\n  ;; Assume that @{} cannot be nested and that the braces are always matched.\n  ;; Obviously, in a real parsing function, these assumptions would need to be validated.\n  (define lst (regexp-split #rx\"@{\" str))\n  ;; After splitting we have a list like this: '(\"An error occured at \" \"file}: \" \"line}.\")\n  ;; We'll go over it, building a list of expressions to be passed to string-append as we go.\n  (define chunks\n    (for\/fold ([result '()])\n              ([chunk (rest lst)])        ; we don't need the first element here\n      (let*\n          ;; convert a string into a port\n          ([is (open-input-string chunk)]\n           ;; call original read to get the expression from inside brackets\n           [form (read is)]\n           ;; read what remains in the port back into a string\n           [after-form (port-&gt;string is)]\n           ;; drop the closing brace\n           [after-brace (substring after-form (add1 (s-index-of after-form \"}\")))])\n        ;; ~a is a generic \"toString\" function in Racket\n        (append result `((~a ,form) ,after-brace)))))\n  ;; chunks now looks like this: ((~a file) \":\" (~a line) \".\")\n  `(string-append ,(first lst) ,@chunks))\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    It's dead simple - just a couple of lines - and it glosses over important\n    details, like the possibility of nesting the <code>@{}<\/code> expressions and syntax\n    errors handling, but it works for the simple cases. We can easily test it,\n    like this:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\">1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\">2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\">3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\">4<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>(parse \"An error occured at @{file}:@{line}.\")\n;; Returns:\n'(string-append \"An error occured at \" (~a file) \":\" (~a line) \".\")\n;; which is exactly what we wanted!\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><h2>The boilerplate code<\/h2><p>\n    The hard part ends here - the rest is just some boilerplate we need to make\n    Racket aware of our extension to its reader. In Racket, this is often done\n    by defining a new language - working with <code>#lang lang-name<\/code> syntax, but in\n    our case we don't need (or want) to create a whole language (it's simpler\n    than it sounds, but it's still a bit more work) - we just change the reader\n    and can reuse everything else from some other language.\n  <\/p><p>\n    For situations like these, Racket supports a <a href=\"https:\/\/docs.racket-lang.org\/guide\/hash-lang_reader.html\">special syntax<\/a> to\n    its <code>#lang<\/code> line, which looks like this:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><\/td><td class=\"code\"><div class=\"highlight\"><pre>#lang reader &lt;reader-implementation-module&gt; &lt;language-the-reader-will-be-used-with&gt;\n;; for example:\n#lang reader \"reader.rkt\" racket\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><h2>Generating a reader module<\/h2><p>\n    Let's first define <code>reader.rkt<\/code>. Writing a whole new reader is of course\n    an option but it's a lot of unnecessary work, as the generic <code>read<\/code> and\n    <code>read-syntax<\/code><sup><a class=\"tooltip\" id=\"ft_note_adding_string_interpolation_to_racket_html_ft_5\" data-tooltip-content=\"#ft_note_adding_string_interpolation_to_racket_html_5\">[5]<\/a><\/sup>  need to handle some more complex things, like\n    module declarations, which we don't care about. Instead, we can use a\n    special language for defining readers, which will generate appropriate\n    <code>read<\/code> and <code>read-syntax<\/code> implementations for us provided we supply a single\n    function to it:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\">1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\">2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\">3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\">4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\">5<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>#lang s-exp syntax\/module-reader\n#:language read\n#:wrapper1 wrap-read\n(require \"ext.rkt\")\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    The <code>#:language<\/code> option takes a function which accepts a string (of what\n    remains after the reader module name on the <code>#lang<\/code> line) and returns a\n    language name to use. A simple <code>read<\/code> happens to work here.\n  <\/p><p>\n    The <code>#:wrapper1<\/code> option takes a function, which takes another function as\n    input and returns the parsed code. The passed in function, when invoked,\n    performs normal read.\n  <\/p><h2>Extending and installing a readtable<\/h2><p>\n    With <code>#:wrapper1<\/code> option, in conjunction with <code>parameterize<\/code>, it's very easy\n    to install our readtable<sup><a class=\"tooltip\" id=\"ft_note_adding_string_interpolation_to_racket_html_ft_6\" data-tooltip-content=\"#ft_note_adding_string_interpolation_to_racket_html_6\">[6]<\/a><\/sup>:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\">1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\">2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\">3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\">4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\">5<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-6\">6<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-7\">7<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-8\">8<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>(define (wrap-read do-read)\n  (define rt (make-readtable (current-readtable)\n                             ;; register our function to be called when #@ is encountered\n                             #\\@ 'dispatch-macro\n                             ;; name of the function to be called\n                             read-str))\n  (parameterize ([current-readtable rt])\n    (do-read)))\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><h2>Wrapping parse so that it works inside readtable<\/h2><p>\n    The only part missing is a <code>read-str<\/code> function, which is a bit more\n    complicated. The function passed to <code>make-readtable<\/code> may be called in two\n    different ways, depending on whether it's used from inside <code>read<\/code> or\n    <code>read-syntax<\/code>. The function distinguishes the cases based on a number of\n    arguments it gets, and returns either syntax object or a simple list\n    depending on the case. Here's its implementation:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\"> 1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\"> 2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\"> 3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\"> 4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\"> 5<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-6\"> 6<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-7\"> 7<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-8\"> 8<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-9\"> 9<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-10\">10<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-11\">11<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-12\">12<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-13\">13<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-14\">14<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-15\">15<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-16\">16<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-17\">17<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>(define read-str\n  (case-lambda\n    [(ch in src line col pos)\n     ;; The caller wants a syntax object. The current position of the `in` port\n     ;; is just after the #@ characters, before the opening quote.\n     (let*\n         ;; Here we read the next expression (which we assume is a string)\n         ([literal-string (read in)])\n       ;; then we parse the contents of the string and return the result,\n       ;; wrapped in a new syntax object. The `in` port is now just after the\n       ;; closing quote and the parser continues from there.\n       (datum-&gt;syntax #f (parse literal-string)))]\n    [(ch in)\n     ;; The caller wants a simple list. Let's parse the input into syntax and\n     ;; transform to a list (this saves as writing the same code twice).\n     (syntax-&gt;datum (read-syntax #f in))]))))\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><h2>It works!<\/h2><p>\n    The reader extension can be used like this:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\">1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\">2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\">3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\">4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\">5<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-6\">6<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>#lang reader \"reader.rkt\" racket\n(define line 42)\n(define file \"file-name.rkt\")\n(displayln #@\"An error occured at @{file}:@{line}.\")\n;; Displays: An error occured at file-name.rkt:42.\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    It also works for more complex expressions out of the box:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\">1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\">2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\">3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\">4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\">5<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>#lang reader \"reader.rkt\" racket\n(require racket\/date)\n(displayln #@\"Current date&amp;time: @{(date-&gt;string (current-date) #t)}\")\n;; Displays: Current date&amp;time: Friday, April 28th, 2017 7:57:11pm\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    This is of course a pretty basic as far as string interpolation features go,\n    but it took thirty (34 to be precise) lines of code (excluding comments) to\n    build it. Adding another twenty lines would be enough to implement error\n    checking, and using one of the parser generators could even shorten the\n    code.\n  <\/p><h2>That's it!<\/h2><p>\n    Aside from how easy it is to extend Racket syntax one thing deserves a\n    mention. That is, such syntax extensions are easily composable (as long as\n    they don't want to install extensions to the same entry in the readtable) -\n    you can load them one by one, each extending the readtable.\n  <\/p><p>\n    The language extension and creation tools of Racket don't end there. Robust\n    module system with support for exporting syntax extensions, powerful\n    <a href=\"https:\/\/docs.racket-lang.org\/syntax\/stxparse.html?q=define-simple-macro\"><code>syntax-parse<\/code><\/a>,\n    traditional <a href=\"https:\/\/docs.racket-lang.org\/reference\/stx-patterns.html\"><code>syntax-rules<\/code> and <code>syntax-case<\/code><\/a>,\n    many tools for generating <a href=\"https:\/\/docs.racket-lang.org\/parser-tools\/LALR_1__Parsers.html\">parsers<\/a>\n    and <a href=\"https:\/\/docs.racket-lang.org\/parser-tools\/Lexers.html\">lexers<\/a> -\n    all these features are geared towards meta-lingustic programming. Racket is\n    a programming language laboratory and allows you to bend it to better suit\n    your needs safely (ie. ensuring that only your code will be affected). Very\n    few languages or transpilers offer that much freedom with that kind of\n    safety guarantees.\n  <\/p><p>\n    It's an ideal environment for people who know what they need from their\n    language - with Racket you're <strong>never<\/strong> stuck, waiting for the\n    next release to support some feature. On the other hand, despite there being\n    a great many helpers and libraries, you need to know at least some basics of\n    programming language creation to do this yourself. The good news is that the\n    extensions written by others are easily installable with a single command of\n    a <a>Racket package manager<\/a>.\n  <\/p><h2>Addendum: doing the same without messing with the reader<\/h2><p>\n    A commenter on Hacker News provided an interesting solution using normal\n    macros only. Turns out there's a special macro, called <code>#%datum<\/code>, which gets\n    wrapped around every literal in the code (when the code is read). By default\n    it's a noop, but we can override it with arbitrary logic. In this solution\n    there's no need for marking strings for interpolation: all strings are\n    interpolate-able by default. Here's the implementation:\n  <\/p><div class=\"code\" data-lang=\"racket\"><table class=\"highlighttable\"><tr><td class=\"linenos\"><div class=\"linenodiv\"><pre><a href=\"#adding_string_interpolation_to_racket.html-1\"> 1<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-2\"> 2<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-3\"> 3<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-4\"> 4<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-5\"> 5<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-6\"> 6<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-7\"> 7<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-8\"> 8<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-9\"> 9<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-10\">10<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-11\">11<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-12\">12<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-13\">13<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-14\">14<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-15\">15<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-16\">16<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-17\">17<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-18\">18<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-19\">19<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-20\">20<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-21\">21<\/a>\n<a href=\"#adding_string_interpolation_to_racket.html-22\">22<\/a><\/pre><\/div><\/td><td class=\"code\"><div class=\"highlight\"><pre>(require\n  (rename-in racket (#%datum core-datum))\n  (for-syntax syntax\/parse))\n(define-syntax (#%datum stx)\n  (syntax-parse stx\n    [(_ . x:str) (interpolate #'x)]\n    [(_ . x)     #'(core-datum . x)]))\n(define-for-syntax (interpolate stx)\n  (define re            #rx\"@{[^}]+}\")\n  (define (trim match)  (substring match 2 (- (string-length match) 1)))\n  (define (to-stx val)  (datum-&gt;syntax stx val))\n  (define (datum val)   (to-stx (cons #'core-datum val)))\n  (define (source text) (to-stx (read (open-input-string text))))\n  (let* ([text     (syntax-&gt;datum stx)]\n         [matches  (regexp-match* re text)]\n         [template (datum (regexp-replace re text \"~a\"))]\n         [values   (map (compose source trim) matches)])\n    (if (null? matches)\n        template\n        (to-stx `(,#'format ,template ,@values)))))\n<\/pre><\/div>\n<\/td><\/tr><\/table><\/div><p>\n    I don't have the time to go over this line by line and explain it right now,\n    but it's possible I'll write another post featuring this solution in the\n    future.\n  <\/p><ul class=\"footnotes\"><li>The quote comes from Paul\n    Graham's <a href=\"http:\/\/www.paulgraham.com\/onlisptext.html\">famous book<\/a> and\n    is attributed to John Foderaro. <a href=\"#ft_note_adding_string_interpolation_to_racket_html_ft_1\">\u21b5<\/a><\/li>\n<li> A good\n    starting point for learning more would be <a href=\"http:\/\/www.ccs.neu.edu\/home\/matthias\/Thoughts\/Racket_is____.html\"><q>Racket\n    is ...<\/q><\/a> article by Matthias Felleisen. At least I assume it's his -\n    it's not signed, but clearly in Matthias home directory ;) <a href=\"#ft_note_adding_string_interpolation_to_racket_html_ft_2\">\u21b5<\/a><\/li>\n<li>Abstract Syntax Tree, see <a href=\"https:\/\/klibert.pl\/posts\/adding_string_interpolation_to_racket.html\">Wikipedia<\/a> <a href=\"#ft_note_adding_string_interpolation_to_racket_html_ft_3\">\u21b5<\/a><\/li>\n<li> Taken from a\n    paper <a href=\"http:\/\/www.ccs.neu.edu\/home\/stchang\/pubs\/ckg-popl2017.pdf\">Type\n    Systems as Macros<\/a> by Stephen Chang, Alex Knauth, Ben Greenman <a href=\"#ft_note_adding_string_interpolation_to_racket_html_ft_4\">\u21b5<\/a><\/li>\n<li> Any module, which exports these two functions, can be used\n    as a reader. <a href=\"#ft_note_adding_string_interpolation_to_racket_html_ft_5\">\u21b5<\/a><\/li>\n<li> In the actual source code the <code>(make-readtable ...)<\/code>\n    part is extracted into separate procedure, which makes it easier to compose\n    this syntax extension with others. <a href=\"#ft_note_adding_string_interpolation_to_racket_html_ft_6\">\u21b5<\/a><\/li>\n<\/ul><a href=\"https:\/\/klibert.pl\/posts\/adding_string_interpolation_to_racket.html\" class=\"title-link\">Comments<\/a><\/header>","created_at":"2018-11-07T16:51:13+0000","updated_at":"2018-11-07T16:51:13+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"klibert.pl","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":530,"title":"Practical Common Lisp","url":"http:\/\/www.gigamonkeys.com\/book\/","content":"<p>My new book,\n      <a href=\"http:\/\/www.codersatwork.com\/\"><i>Coders at Work<\/i><\/a>, a collection\n      of Q&amp;A interviews with fifteen all-time great\n      programmers and computer scientists, is out and available from\n        <a href=\"http:\/\/www.amazon.com\/gp\/product\/1430219483?ie=UTF8&amp;tag=gigamonkeys-20&amp;linkCode=as2&amp;camp=1789&amp;creative=9325&amp;creativeASIN=1430219483\">Amazon<\/a>,\n <a href=\"http:\/\/search.barnesandnoble.com\/booksearch\/isbninquiry.asp?box=1430219483&amp;pos=-1&amp;EAN=9781430219484\">Barnes &amp; Noble<\/a>,\n <a href=\"http:\/\/www.powells.com\/biblio\/71-9781430219484-0?search_avail=1\">Powell's<\/a>,\n other fine booksellers near you and as an <a href=\"http:\/\/www.apress.com\/9781430219484\">eBook from Apress<\/a>.<\/p><p>\n      <a href=\"http:\/\/www.amazon.com\/gp\/product\/1590592395?ie=UTF8&amp;tag=gigamonkeys-20&amp;linkCode=as2&amp;camp=1789&amp;creative=9325&amp;creativeASIN=1590592395\">Amazon<\/a><img src=\"http:\/\/www.assoc-amazon.com\/e\/ir?t=gigamonkeys-20&amp;l=as2&amp;o=1&amp;a=1590592395\" width=\"1\" height=\"1\" border=\"0\" alt=\"\" \/> |\n      <a href=\"http:\/\/www.powells.com\/cgi-bin\/biblio?inkey=62-1590592395-3\">Powells<\/a> |\n      <a href=\"http:\/\/search.barnesandnoble.com\/booksearch\/isbnInquiry.asp?isbn=1590592395\">Barnes &amp; Noble<\/a>\n    <\/p><hr style=\"width: 50%;\" \/><p><b>Download source code: <\/b><a href=\"http:\/\/www.gigamonkeys.com\/book\/practicals-1.0.3.tar.gz\">tar.gz<\/a> | <a href=\"http:\/\/www.gigamonkeys.com\/book\/practicals-1.0.3.zip\">zip<\/a><\/p><hr style=\"width: 50%;\" \/><p>Like what you've read?\n    Then help spread the word. Recommend this book to your friends.\n    Write\n    a <a href=\"http:\/\/www.amazon.com\/gp\/customer-reviews\/write-a-review.html\/104-2515398-4782364?%5Fencoding=UTF8&amp;asin=1590592395&amp;store=books\">review<\/a>\n    on Amazon. Blog about it. Link to this page from your web site.\n    Whatever. Apress took a chance, publishing this book when other\n    publishers thought there was no market for a Lisp book. While\n    it's unlikely that I'll get rich off my royalties, we don't have\n    to sell all that many copies for Apress to turn a profit and show\n    the naysayers that Lisp has legs yet.<\/p>","created_at":"2018-11-07T16:51:13+0000","updated_at":"2018-11-07T16:51:13+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.gigamonkeys.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":531,"title":"Best Themes For Arch Linux, Antergos, Apricity OS And Manjaro","url":"https:\/\/itsfoss.com\/icon-themes-arch-linux\/","content":"<p>Arch Linux is the desire of heart for anyone looking towards extensive customization. I\u2019ve seen customization which, well, could have been a new distro in themselves. You can customize to your content on other distros too. But with Arch, since you are packing everything one by one, you can choose to ditch something, add something extra or even use alternatives so it\u2019s\u2026 Well \u201cYou\u201d.<\/p><p>Another thing that allows you to do your thing on Arch Linux and Arch-based systems is the package manager. <a href=\"https:\/\/wiki.archlinux.org\/index.php\/pacman\" class=\"external\" rel=\"nofollow\" target=\"_blank\">Pacman<\/a> combined with the <a href=\"https:\/\/aur.archlinux.org\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">Arch User Repository<\/a> make sure anything and everything you\u2019d ever want are always at your disposal.<\/p>\n<p>Talking about the themes, the amazing Arch User Repository provides a <a href=\"https:\/\/aur.archlinux.org\/packages.php?K=gtk-theme\" class=\"external\" rel=\"nofollow\" target=\"_blank\">huge number of themes<\/a> for all the desktop environments. In this article, we\u2019ll be having a look on noteworthy of these. The first section will deal with themes and in the second section, let\u2019s browse some cool icon packs. Then we mix and match.<\/p>\n<h2>Best themes for Arch Linux, Manjaro, Antergos and Apricity OS<\/h2>\n<h3>1. Arc<\/h3>\n<p>Let\u2019s start with the most famous (famous for a reason) desktop theme. Arc theme is a highly sophisticated theme which provides a gorgeous view. Arc theme comes in 3 variants. Arc theme modifies the sections of the window into clearly distinguishable parts. It is currently my favorite.<\/p>\n<figure id=\"attachment_15059\" class=\"wp-caption aligncenter\"><img class=\"wp-image-15059 size-full\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-darker.jpg\" alt=\"Arc-darker\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-darker.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-darker-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-darker-768x432.jpg 768w\" \/><figcaption class=\"wp-caption-text\">Arc-darker<\/figcaption><\/figure><figure id=\"attachment_15058\" class=\"wp-caption aligncenter\"><img class=\"wp-image-15058 size-full\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-dark.jpg\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-dark.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-dark-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/arc-dark-768x432.jpg 768w\" alt=\"image\" \/><figcaption class=\"wp-caption-text\">Arc-dark<\/figcaption><\/figure><p>To download and install Arc theme, use the command below.\ufffc\ufffc<\/p>\n<pre>yaourt\u00a0arc-icon-theme-git<\/pre>\n<h3>2. Gnome-Breeze<\/h3>\n<p>Who isn\u2019t impressed by the absolute elegance of <a href=\"https:\/\/neon.kde.org\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">KDE Plasma <\/a>desktop? Gnome-Breeze is the GTK theme which mimics this beauty\u00a0on GTK based desktop environments.<\/p>\n<figure id=\"attachment_15061\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15061\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Gnome-breeze.jpg\" alt=\"\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Gnome-breeze.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Gnome-breeze-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Gnome-breeze-768x432.jpg 768w\" \/><figcaption class=\"wp-caption-text\">Gnome-Breeze<\/figcaption><\/figure><p>To download and install Gnome-Breeze theme, use the command below.<\/p>\n<pre>yaourt gnome-breeze-git<\/pre>\n<h3>3. Zuki<\/h3>\n<p>Zuki is a minimalistic and compact theme that still is quite appealing. Zuki also features some variants. The variants offer slightly different colour schemes. The theme comes with minimal padding and low on other shiny stuff. Still, some people might fancy it. Who am I to judge?<\/p>\n<figure id=\"attachment_15062\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15062\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/zuki.jpg\" alt=\"\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/zuki.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/zuki-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/zuki-768x432.jpg 768w\" \/><figcaption class=\"wp-caption-text\">Zuki<\/figcaption><\/figure><p>To download and install Zuki theme, use the command below.<\/p>\n<pre>yaourt zuki-themes<\/pre>\n<h3>4. Numix<\/h3>\n<p>No theme list can be complete without <a href=\"https:\/\/numixproject.org\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">Numix<\/a>. Numix has some pretty beautiful desktop themes. Well, it has to keep up with its gorgeous icons. \u00a0Numix theme too comes in some variants from which you may choose the one that works for you.<\/p>\n<figure id=\"attachment_15063\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15063\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-frost.jpg\" alt=\"\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-frost.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-frost-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-frost-768x432.jpg 768w\" \/><figcaption class=\"wp-caption-text\">Numix-Frost<\/figcaption><\/figure><p>To download and install Numix\u00a0theme, use the command below.<\/p>\n<pre>yaourt numix-gtk-theme<\/pre>\n<h3>How to change the themes?<\/h3>\n<p>Once you have downloaded the theme on your computer, you can browse and change theme using the below instructions.<\/p>\n<p><strong>Gnome:\u00a0<\/strong><a href=\"https:\/\/itsfoss.com\/install-switch-themes-gnome-shell\/\" target=\"_blank\" rel=\"noopener\">Use the Gnome tweak tool<\/a>. The theme can be changed under appearance.<\/p>\n<p><strong>Cinnamon:\u00a0<\/strong>Search for \u2018Theme\u2019 in the menu and use the entry. Various aspects regarding appearance can be adjusted here.<\/p>\n<p><strong>Mate and Xfce:\u00a0<\/strong>Settings &gt; Appearance<\/p>\n<h2>Best icons for\u00a0Arch Linux, Manjaro, Antergos and Apricity OS<\/h2>\n<p>Now that we have got the theme handled, let\u2019s take a peek at some icon packs to finish off the beautifying.<\/p>\n<h3>1. Numix Circle<\/h3>\n<p>Once you have got a nice theme and a mesmerizing wallpaper set, nothing completes the desktop like Numix Circle icons. Numix icon pack is colorful, vibrant and much more complete and consistent than any other custom icon pack out there. Yes, I said it and I\u2019m ready to put my money on it too.<img class=\"aligncenter size-full wp-image-15064 jetpack-lazy-image\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle.jpg?is-pending-load=1\" alt=\"\" width=\"800\" height=\"450\" data-lazy-srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle-768x432.jpg 768w\" data-lazy-sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-15064\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle.jpg\" alt=\"\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-circle-768x432.jpg 768w\" \/><\/noscript>\n<p>To get Numix circle icons, use the following command<\/p>\n<pre>yaourt numix-circle-icon-theme-git<\/pre>\n<h3>2. Numix square icons<\/h3>\n<p>For those of you who are not really into the whole circle thingy, Numix has got some pretty neat square icons too. Numix has an extensive collection of icons for Linux software.<\/p>\n<figure id=\"attachment_15066\" class=\"wp-caption aligncenter\"><img class=\"wp-image-15066 size-full\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-square.jpg\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-square.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-square-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/numix-square-768x432.jpg 768w\" alt=\"image\" \/><figcaption class=\"wp-caption-text\">Numix-square<\/figcaption><\/figure><p>To get Numix square icons, use the following command<\/p>\n<pre>yaourt\u00a0numix-square-icon-theme-git<\/pre>\n<h3>3. Flattr<\/h3>\n<p>Flattr is a beautiful, very unique set of icons for Linux. They are inspired by the flat icon trend and they look good coupled with a bright wallpaper.<\/p>\n<figure id=\"attachment_15067\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15067\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Flattr_icons.jpeg\" alt=\"\" width=\"800\" height=\"330\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Flattr_icons.jpeg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Flattr_icons-300x124.jpeg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Flattr_icons-768x317.jpeg 768w\" \/><figcaption class=\"wp-caption-text\">Flattr icons<\/figcaption><\/figure><p>To get Flattr icons, use the following command<\/p>\n<pre>yaourt\u00a0flattr-icon-theme<\/pre>\n<h3>4. Paper Icons<\/h3>\n<p>Paper icon set is an exhaustive collection of modern icons for Linux. They are beautiful and look good if coupled with either modern on material wallpapers.<\/p>\n<figure id=\"attachment_15069\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15069\" src=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/paper.jpg\" alt=\"\" width=\"800\" height=\"450\" srcset=\"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/paper.jpg 800w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/paper-300x169.jpg 300w, https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/paper-768x432.jpg 768w\" \/><figcaption class=\"wp-caption-text\">Paper Icons<\/figcaption><\/figure><p>To get Paper icons, use the following command<\/p>\n<pre>yaourt paper-icon-theme-git\ufffc<\/pre>\n<h3>It\u2019s time to let the imagination work<\/h3>\n<p>You can see that I have made the Gnome top panel transparent. I got the idea from Elementary OS and Apricity OS of course. You can get the instructions on <a href=\"https:\/\/youtu.be\/gy2_1qZtWP4\" class=\"external\" rel=\"nofollow\" target=\"_blank\">how to make the Gnome top panel transparent here<\/a>. A transparent top panel blends into the desktop background gracefully.<\/p>\n<p>Well, these are not by any means everything that the Arch User Repository has to offer. There are a<a href=\"https:\/\/aur.archlinux.org\/packages.php?O=0&amp;L=0&amp;C=17&amp;K=icon-theme&amp;SeB=nd&amp;SB=n&amp;SO=a&amp;PP=50&amp;do_Search=Go\" class=\"external\" rel=\"nofollow\" target=\"_blank\"> huge number of icon packs<\/a> in the AUR. But the ones that are impressive and complete to a certain extent are listed here. This is a subjective topic so go ahead and tell us your favorite themes and icon packs in the comments below. Do share the article.<\/p>\n\t\t\t&#13;","created_at":"2018-11-07T16:51:13+0000","updated_at":"2018-11-07T16:51:13+0000","published_at":"2017-03-11T09:00:00+0000","published_by":["Aquil Roshan"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"itsfoss.com","preview_picture":"https:\/\/4bds6hergc-flywheel.netdna-ssl.com\/wp-content\/uploads\/2017\/03\/Arch-Linux-Themes.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":532,"title":"An introduction to bash completion: part 2","url":"https:\/\/debian-administration.org\/article\/317\/An_introduction_to_bash_completion_part_2","content":"<div class=\"tagline\"><p>Posted by <a href=\"https:\/\/debian-administration.org\/users\/Steve\">Steve<\/a> on Sat 24 Dec 2005 at 03:00<\/p><\/div><p>Tags: <a href=\"https:\/\/debian-administration.org\/tag\/bash\" title=\"Find more content with this tag\" rel=\"tag\">bash<\/a>, <a href=\"https:\/\/debian-administration.org\/tag\/command%20completion\" title=\"Find more content with this tag\" rel=\"tag\">command completion<\/a><\/p><p>Previously we showed how to add basic completion to commands, using facilities which were already provided by the bash completion routines. In this second part we'll demonstrate how to add completely new custom completion to commands.<\/p><p>In <a href=\"https:\/\/www.debian-administration.org\/articles\/316\">part one<\/a> we looked at adding hostname completion to arbitrary commands by executing:<\/p><pre>\ncomplete -F _known_hosts xvncviewer\n<\/pre><p>This uses the <code>complete<\/code> command to tell bash that the function <code>_known_hosts<\/code> should be used to handle the completion of arguments to the <code>xvncviewer<\/code>.<\/p><p>If we wish to add custom completion to a command we will instead write our own function, and bind <i>that<\/i> to the command.<\/p><b>A Basic Example<\/b><blockquote>\n<p>As a basic example we'll first look at adding some simple completions to the binary <code>foo<\/code>. This hypothetical command takes three arguments:<\/p>\n<ul><li><code>--help<\/code>\n<ul><li>Shows the help options for <code>foo<\/code>, and exits.<\/li>\n<\/ul><\/li>\n<li><code>--version<\/code>\n<ul><li>Shows the version of the <code>foo<\/code> command, and exits.<\/li>\n<\/ul><\/li>\n<li><code>--verbose<\/code>\n<ul><li>Runs <code>foo<\/code> with extra verbosity<\/li>\n<\/ul><\/li>\n<\/ul><p>To handle these arguments we'll create a new file <code>\/etc\/bash_completion.d\/foo<\/code>. This file will be automatically sourced (or loaded) when the bash completion code is loaded.<\/p>\n<p>Inside that file save the following text:<\/p>\n<pre>\n_foo() \n{\n    local cur prev opts\n    COMPREPLY=()\n    cur=\"${COMP_WORDS[COMP_CWORD]}\"\n    prev=\"${COMP_WORDS[COMP_CWORD-1]}\"\n    opts=\"--help --verbose --version\"\n    if [[ ${cur} == -* ]] ; then\n        COMPREPLY=( $(compgen -W \"${opts}\" -- ${cur}) )\n        return 0\n    fi\n}\ncomplete -F _foo foo\n<\/pre>\n<p>To test it you can now source the file:<\/p>\n<pre>\nskx@lappy:~$ . \/etc\/bash_completion.d\/foo\nskx@lappy:~$ foo --[TAB]\n--help     --verbose  --version  \n<\/pre>\n<p>If you experiment you'll see that it successfully completes the arguments as expected. Type \"<code>foo --h[TAB]<\/code>\" and the <code>--help<\/code> argument is completed. Press [TAB] a few times and you'll see all the options. (In this case it doesn't actually matter if you don't have a binary called <code>foo<\/code> installed upon your system.)<\/p>\n<p>So now that we have something working we should look at how it <i>actually<\/i> works!<\/p>\n<\/blockquote><b>How Completion Works<\/b><blockquote>\n<p>The previous example showed a simple bash function which was invoked to handle completion for a command.<\/p>\n<p>This function starts out by defining some variables <code>cur<\/code> being the current word being typed, <code>prev<\/code> being the previous word typed, and <code>opts<\/code> which is our list of options to complete.<\/p>\n<p>The option completing is then handled by use of the <code>compgen<\/code> command via this line:<\/p>\n<pre>\nCOMPREPLY=( $(compgen -W \"${opts}\" -- ${cur}) )\n<\/pre>\n<p>What this does is set the value of <code>$COMPREPLY<\/code> to the output of running the command:<\/p>\n<pre>\ncompgen -W \"${opts}\" -- ${cur}\n<\/pre>\n<p>If you replace these variables with their contents you'll see how that works:<\/p>\n<pre>\ncompgen -W \"--help --verbose --version\" -- \"userinput\"\n<\/pre>\n<p>This command attempts to return the match of the current word \"<code>${cur}<\/code>\" against the list \"<code>--help --verbose --version<\/code>\". If you run this command in a shell you'll be able to experiment with it and see how it works:<\/p>\n<pre>\nskx@lappy:~$ compgen -W \"--help --verbose --version\" -- --\n--help\n--verbose\n--version\nskx@lappy:~$ compgen -W \"--help --verbose --version\" -- --h \n--help\n<\/pre>\n<p>Here you first see what happens if the user enters just \"<code>--<\/code>\" - all three options match so they are returned. In the second attempt the user enters <code>--h<\/code> and this is enough to specify <code>--help<\/code> unambiguously, so that is returned.<\/p>\n<p>In our function we simply set \"<code>COMPREPLY<\/code>\" to this result, and return. This allows bash to replace our current word with the output. <code>COMPREPLY<\/code> is a special variable which has a particular meaning within bash. Inside completion routines it is used to denote the output of the completion attempt.<\/p>\n<p>From <a href=\"https:\/\/www.gnu.org\/software\/bash\/manual\/bash.html\">the bash reference manual<\/a> we can read the description of <code>COMPREPLY<\/code>:<\/p>\n<dl><dt>COMPREPLY<\/dt>\n<dd>\n<p>An array variable from which Bash reads the possible completions generated by a shell function invoked by the programmable completion facility<\/p>\n<\/dd>\n<dd>\n<p>We can also see how we found the current word using the array <code>COMP_WORDS<\/code> to find both the current and the previous word by looking them up:<\/p>\n<dl><dt>COMP_WORDS<\/dt>\n<dd>\n<p>An array variable consisting of the individual words in the current command line. This variable is available only in shell functions invoked by the programmable completion facilities.<\/p>\n<\/dd>\n<dt>COMP_CWORD<\/dt>\n<dd>\n<p>An index into ${COMP_WORDS} of the word containing the current cursor position. This variable is available only in shell functions invoked by the programmable completion facilities<\/p>\n<\/dd>\n<\/dl><\/dd>\n<\/dl><\/blockquote><b>A Complex Example<\/b><blockquote>\n<p>Many commands are more complicated to fill out, and have numerous options which depend upon their previous ones.<\/p>\n<p>As a relevant example Xen ships with a command called <code>xm<\/code> this has some basic options:<\/p>\n<ul><li><code>xm list<\/code>\n<ul><li>List all running Xen instances<\/li>\n<\/ul><\/li>\n<li><code>xm create ConfigName<\/code>\n<ul><li>Create a new Xen instances using the configuration file in <code>\/etc\/xen<\/code> called <code>ConfigName<\/code>.<\/li>\n<\/ul><\/li>\n<li><code>xm console Name<\/code>\n<ul><li>Connect to the console of the running machine named \"Name\".<\/li>\n<\/ul><\/li>\n<\/ul><p>In general the command is \"<code>xm operation args<\/code>\" where \"<code>args<\/code>\" varies depending upon the initial <code>operation<\/code> selected.<\/p>\n<p>Setting up basic completion of inital operation can be handled in much the same way as our previous example the only difference is that the operations don't start with a \"<code>--<\/code>\" prefix. However completing the arguments requires special handling.<\/p>\n<p>If you recall we have access to the previous token upon the command line, and using that we can take different actions for each operation.<\/p>\n<p>The sample code looks like this:<\/p>\n<pre>\n_xm() \n{\n    local cur prev opts base\n    COMPREPLY=()\n    cur=\"${COMP_WORDS[COMP_CWORD]}\"\n    prev=\"${COMP_WORDS[COMP_CWORD-1]}\"\n    #\n    #  The basic options we'll complete.\n    #\n    opts=\"console create list\"\n    #\n    #  Complete the arguments to some of the basic commands.\n    #\n    case \"${prev}\" in\n        console)\n            local running=$(for x in `xm list --long | grep \\(name | grep -v Domain-0 | awk '{ print $2 }' | tr -d \\)`; do echo ${x} ; done )\n            COMPREPLY=( $(compgen -W \"${running}\" -- ${cur}) )\n            return 0\n            ;;\n        create)\n            local names=$(for x in `ls -1 \/etc\/xen\/*.cfg`; do echo ${x\/\\\/etc\\\/xen\\\/\/} ; done )\n            COMPREPLY=( $(compgen -W \"${names}\" -- ${cur}) )\n            return 0\n            ;;\n        *)\n        ;;\n    esac\n   COMPREPLY=($(compgen -W \"${opts}\" -- ${cur}))  \n   return 0\n}\ncomplete -F _xm xm\n<\/pre>\n<p>Here we've setup the initial completion of the operations and then added special handling for the two operations \"<code>create<\/code>\" and \"<code>console<\/code>\". In both cases we use <code>compgen<\/code> to complete the input based upon the text that is supplied by the user, compared against a dynamically created list.<\/p>\n<p>For the \"<code>console<\/code>\" operation we complete based upon the output of this command:<\/p>\n<pre>\nxm list --long | grep \\(name | grep -v Domain-0 | awk '{ print $2 }' | tr -d \\)\n<\/pre>\n<p>This gives us a list of the running Xen systems.<\/p>\n<p>For the creation operation we complete based upon the output of this command:<\/p>\n<pre>\nfor x in `ls -1 \/etc\/xen\/*.cfg`; do echo ${x\/\\\/etc\\\/xen\\\/\/} ; done\n<\/pre>\n<p>This takes a directory listing of the <code>\/etc\/xen<\/code> directory and outputs the names of any files ending in <code>.cfg<\/code>. For example:<\/p>\n<pre>\nskx@lappy:~$ for x in `ls -1 \/etc\/xen\/*.cfg`; do echo ${x\/\\\/etc\\\/xen\\\/\/}; done\netch.cfg\nroot.cfg\nsarge.cfg\nsteve.cfg\nx.cfg\nskx@lappy:~$ \n<\/pre><\/blockquote><b>Other Completion<\/b><blockquote>\n<p>Using the <code>compgen<\/code> command we've shown how to match user input against particular strings, both by using a fixed set of choices and by using the output of commands.<\/p>\n<p>It is also possible to match directory names, process names, and other things. See the bash manual for a full description by running \"<code>man bash<\/code>\".<\/p>\n<p>The final example demonstrates how to complete files and hostnames in response to two initial options:<\/p>\n<pre>\n#\n#  Completion for foo:\n#\n#  foo file [filename]\n#  foo hostname [hostname]\n#\n_foo() \n{\n    local cur prev opts\n    COMPREPLY=()\n    cur=\"${COMP_WORDS[COMP_CWORD]}\"\n    prev=\"${COMP_WORDS[COMP_CWORD-1]}\"\n    opts=\"file hostname\"\n \n    case \"${prev}\" in\n        file)\n            COMPREPLY=( $(compgen -f ${cur}) )\n            return 0\n            ;;\n        hostname)\n            COMPREPLY=( $(compgen -A hostname ${cur}) )\n            return 0\n            ;;\n        *)\n        ;;\n    esac\n    COMPREPLY=( $(compgen -W \"${opts}\" -- ${cur}) )\n}\ncomplete -F _foo foo\n<\/pre><\/blockquote><p>Using these examples you should now be able to create your own custom completion functions. 95% of the time you only need to complete from a set of available options, the other times you'll need to deal with dynamic argument generation much like we did for the <code>xm<\/code> command.<\/p><p>Breaking the options down into a small collection of pipes and testing them outside the completion environment (just in the shell) is probably the best approach, then once you've got a working command line just paste it into your function.<\/p><p>For reference my <code>xm<\/code> completion code <a href=\"http:\/\/www.cvsrepository.org\/cgi-bin\/trac\/xen-tools\/dir?d=xen-tools\/misc\">can be found online<\/a>.<\/p>","created_at":"2018-11-07T16:51:13+0000","updated_at":"2018-11-07T16:51:13+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"debian-administration.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":528,"title":"Philosophy tool kit","url":"https:\/\/aeon.co\/essays\/with-the-use-of-heuristics-anybody-can-think-like-a-philosopher","content":"<p>Philosophers pride themselves on thinking clearly by seeing what follows from what, exposing sophisms, spotting fallacies, and generally policing our reasoning. Many have spent years honing their skills, often deploying them on arcane topics. But these skills are not the exclusive property of rarefied sages, accessed only with a secret handshake and insider training, as much as some philosophers wish this were so. Instead, some of these skills can be captured by generalisable, all-purpose techniques for the proper conduct of thought, whatever the topic. Many of these are easily taught and learned. As such, they can be utilised by non-philosophers too. At a time when we are bombarded more than ever with specious claims and spurious inferences, clear thinking provides a much-needed safeguard that we should all strive towards.\u00a0<\/p><p>Philosophers place a premium on certain tools for regimenting our thinking, especially logic and probability theory. However, there is a far richer toolbox at our disposal. Over the years, I have observed philosophers repeatedly using various argumentative moves or strategies, which can be encapsulated in rules of thumb that make their tasks easier. These are what might be called <em>philosophical heuristics.<\/em> This should come as no surprise: pretty much every complex activity has its heuristics, which experts teach and beginners learn \u2013 photography, calligraphy, diving, driving, football, foosball, judo, Cluedo, curling, hurling, climbing, rhyming, and so on. Such heuristics are especially well-documented for chess: \u2018castle early and often\u2019, \u2018check every check\u2019, and what have you.<\/p><p>There are also common heuristics for intellectual activities such as mathematics and creative writing. Here\u2019s a good one for mathematics: if you are not making headway on a problem, modify it slightly to make it easier, and solve that one. A good heuristic for creative writing is to juxtapose familiar words and phrases in unfamiliar ways. One might use the \u2018cut-up technique\u2019, popularised by William S Burroughs and by David Bowie, in which written text is cut up and rearranged to create a new text.<\/p><p>Yet philosophy might be thought to be especially unsuitable for such heuristics. The word \u2018philosopher\u2019 comes from the Ancient Greek <em>philosophos<\/em>, meaning \u2018lover of wisdom\u2019. And wisdom, a skeptic might insist, cannot be so easily achieved. Philosophy strives for deep, profound insights, yet heuristics might by their nature be regarded as superficial. I don\u2019t pretend that philosophical heuristics provide shortcuts to profundity \u2013 any more than chess heuristics provide shortcuts to becoming a grandmaster. That said, grandmasters <em>do<\/em> typically castle early and often, and check every check, consciously or not; a chess textbook that ignored these heuristics would be remiss. Likewise, good philosophers <em>do<\/em> use the heuristics I identify, consciously or not, often in the service of deep insights. Indeed, philosophy textbooks have been remiss in ignoring these heuristics.<\/p><p>If we think of logic and probability theory as all-purpose tools for checking for the consistency and coherence of our claims at a high level of abstraction, then the philosophical heuristics collectively form more of a Swiss army knife. Some of these heuristics have a broad application, like an LED light. Others have a narrower application, but are perfect for the occasions on which they apply, like a corkscrew. There is something of a trade-off between how frequently a particular heuristic might be used, and how specific its advice is. Too general, and the heuristic doesn\u2019t provide an applicable strategy \u2013 for example: \u2018Say something insightful!\u2019 Too specific, and it can never be used in another context \u2013 for example, \u2018the reply to <em>Pascal\u2019s wager<\/em> (that you should believe in God because doing so is the best bet) is that it leaves open which God you should believe in\u2019. The best heuristics find \u2018sweet spots\u2019 in this trade-off.<\/p><p>I work in the Western \u2018analytic\u2019 tradition of philosophy. Much of analytic philosophy involves arguing for positions. So some terminology will be needed here. For our purposes, an <em>argument<\/em> is a number of premises followed by a conclusion, where the premises are intended to lend support to the conclusion. A <em>valid<\/em> argument is one in which the support is as strong as can be: the truth of the premises <em>guarantees<\/em> the truth of the conclusion. A <em>sound<\/em> argument is one that is valid and whose premises are true (and so its conclusion is true, too). An <em>unsound<\/em> argument is one that is either invalid or that has at least one false premise.<\/p><p>Let\u2019s begin with a heuristic that is easy to use, but quite fertile. The word \u2018the\u2019 is the most common word in English. A locution of the form \u2018\u2026 the X \u2026\u2019 \u2013 what philosophers call a <em>definite description \u2013<\/em> typically comes with an assumption that there is <em>exactly one<\/em> X. We might be able to challenge that assumption, in two ways: perhaps there is <em>more than one<\/em> X; perhaps there are <em>no<\/em> Xs. So the heuristic here is to see the word \u2018the\u2019 in neon lights, as it were \u2013 by italicising it, underlining it, or otherwise mentally highlighting it \u2013 and to try each challenge.<\/p><p>Here\u2019s an example that is not philosophical, and certainly not profound, but of considerable interest nonetheless. In his Inauguration speech, Donald Trump said: \u2018January 20th 2017 will be remembered as the day the people became the rulers of this nation again.\u2019 There are three occurrences of \u2018the\u2019 here; let\u2019s focus on the first (the other two have plural nouns \u2013 \u2018people\u2019, \u2018rulers\u2019, but even they have a uniqueness presupposition \u2013 a unique set of people, and of rulers). \u2018The day\u2019 presupposes that there is exactly one such day. Some champions of the power of democracy will insist that there are many such days \u2013 namely, every day on which the people vote. Some skeptics of the power of democracy will deny that there is ever such a day, and that includes 20 January 2017. Either way, the definite description faces a challenge.<\/p><p>Turning to a more philosophical example, we often speak of \u2018doing the right thing\u2019. Sometimes, there is exactly one such action. However, there can be different senses of \u2018right\u2019\u2013 for example, what is rational, and what is moral. And even fixing on one of those senses, there might be more than one candidate for the right thing to do: multiple actions that are equally good. Or there might not be any such candidate. Think of moral dilemmas, such as the unspeakable one depicted in William Styron\u2019s novel <em>Sophie\u2019s Choice<\/em> (1979), or Jean-Paul Sartre\u2019s one of a student who is torn between avenging the death of his brother in the Second World War, and looking after his mother.<\/p><p>Relatedly, and more generally, when evaluating some claim, mentally highlight <em>each<\/em> key term, and run through its <em>contrast class<\/em>, the set of relevant alternatives. It is helpful to stress the term and to intone the words \u2018\u2026 as opposed to \u2026\u2019, to bring out that class. For example, one hears claims to the effect that \u2018the human visual system is poor\u2019. Well, let\u2019s see: \u2018the <em>human<\/em> visual system is poor\u2019. Human, as opposed to what? An eagle\u2019s visual system? Yes, the human visual system compares unfavourably. But what about a bear\u2019s visual system? A bat\u2019s? Now the human\u2019s doesn\u2019t seem so bad. Let\u2019s continue: \u2018the human <em>visual<\/em> system is poor\u2019. Our visual system, as opposed to what? Our olfactory system? Surely not \u2013 we are better seers than we are smellers. Our auditory system? Even that doesn\u2019t sound right.<\/p><p class=\"pullquote\">A racial stereotype can all too easily be \u2018confirmed\u2019 if one attends exclusively to instances, not counter-instances, of it\u00a0<\/p><p>Philosophers use contrastive stress to reveal the logical form of various concepts. For example, causation seems to be a two-place relation: smoking a pack of cigarettes a day causes lung cancer \u2013 so far, so good. But consider: smoking <em>a<\/em> pack of cigarettes a day, as opposed to <em>three<\/em> or <em>four<\/em>, causes lung cancer? That doesn\u2019t sound so good. If anything, relative to those alternatives, smoking (only) <em>one<\/em> pack a day seems to help <em>prevent<\/em> lung cancer. So it seems that <a href=\"https:\/\/philpapers.org\/rec\/HITTRO\" target=\"_blank\">causation<\/a> is at least a three-place relation: C causes E <em>relative to C\u2019<\/em>. Similar <a href=\"http:\/\/www.jonathanschaffer.org\/contrastcause.pdf\" target=\"_blank\">reasoning<\/a> suggests that it is even four-place: C <em>rather than C\u2019<\/em> causes E <em>rather than E\u2019<\/em>.<\/p><p>The contrastive-stress heuristic also helps one detect false dichotomies, a favourite strategy among philosophers. It is also a good corrective to certain cognitive biases to which we are prone:\u00a0<\/p><blockquote><div><em>confirmation bias<\/em>, the tendency to look for and to recall evidence that confirms, but not that disconfirms, one\u2019s beliefs and hypotheses; and<br \/><em>congruence bias<\/em>, the tendency to accept a belief or hypothesis without adequately testing alternative hypotheses.<\/div><\/blockquote><p>Indeed, one of the most common fallacies is simply a failure to <a href=\"http:\/\/www.r-5.org\/files\/books\/ethology\/human-mind\/Thomas_Gilovich-How_We_Know_What_Isn%27t_So-EN.pdf\" target=\"_blank\">consider<\/a> contrary cases. For example, a racial stereotype can all too easily be \u2018confirmed\u2019 in one\u2019s mind if one attends exclusively to instances of it, as opposed to counter-instances.<\/p><p>Now let\u2019s turn to a heuristic that is useful in various fields. Start with a potentially hard problem: someone makes a claim that is supposed to cover a wide range of cases, and you want to check whether it has any counterexamples. You might be facing a huge search space. Where should you look first? Here\u2019s an easier sub-problem: check <em>extreme<\/em> cases to see whether any counterexamples lurk there \u2013 the first case, or the last, or the biggest, or the smallest, or the smelliest, or any similar superlative (always being aware of the definite descriptions!) Does the claim still hold there? This should drastically reduce your search space, as it now just involves the \u2018corners\u2019 or \u2018edges\u2019 of the original space.<\/p><p>For example, some philosophers are fond of making grandiose claims, such as: \u2018Every event has a cause.\u2019 Well, is that true? At first, you might be overwhelmed by its grandiosity \u2013 there are lots of events out there! But start by considering an <em>extreme<\/em> event \u2013 the first event, the Big Bang. There was no prior event to cause it, it did not cause itself, and it was not retro-caused by some later event, so we have our counterexample. To be sure, this assumes that there was exactly one Big Bang, but as far as I know, this is a respectable assumption. Or consider the extreme event that is <em>the entire history of the Universe<\/em>. There are many instances of causation <em>within<\/em> this entire history, but arguably <em>it<\/em> was not caused by anything. Any putative cause is just part of the entire history.<\/p><p>Well, perhaps it did have a cause \u2013 namely, God? Hold that thought; we will return to it soon.<\/p><p>Suppose a politician tells you: \u2018You should not follow any advice given to you by a politician.\u2019 What should you do with this advice? <em>Follow<\/em> it? That\u2019s not following the advice, since it was the exact opposite. <em>Not follow<\/em> it? That\u2019s exactly what the advice was, so you would thereby be following it. Self-referential paradoxes have kept philosophers employed since the ancient Greeks. Georg Cantor, Bertrand Russell and Kurt G\u00f6del shook the foundations of mathematics by exploiting self-reference in various ways. We might set our sights rather lower, but still employ self-reference fruitfully.<\/p><p>For example, philosophers perennially debate <em>realism<\/em> about various subject matters \u2013 ethics, aesthetics, mathematical entities, the meanings of our words, the unobservable entities posited by science, and even ordinary macroscopic objects. A popular definition is that realism about Xs is the thesis that Xs exists independently of observers (for instance, realism about electrons is the thesis that electrons exist independently of observers). But wait \u2013 what about realism about <em>observers<\/em>? Observers do not exist independently of observers. How about: \u2018Xs exist independently of minds\u2019? That won\u2019t do either \u2013 what about realism about <em>minds<\/em>? Minds do not exist independently of minds. So the self-referential heuristic here is to give a claim a taste of its own medicine.<\/p><p>Somewhat related is the time-honoured philosopher\u2019s technique of showing that a view (or an argument) faces an <em>infinite regress<\/em> \u2013 its truth (or validity) depends on the truth of some proposition, whose truth in turn depends on the truth of some <em>other<\/em> proposition, whose truth depends on\u2026 The sequence of dependencies has no end.<\/p><p>Think again of the claim that every event has a cause. Focus on some event. According to the claim, it has a cause, which had a cause, which had a cause, which\u2026, <em>ad infinitum<\/em>. This is at least puzzling, and perhaps worse. Another classic example is the regress of justification. In order to have justified belief in something, one must have justified belief in something else; but that requires having justified belief in something <em>else<\/em>; and this chain of justifications never terminates. (There are various replies \u2013 for example, that the chain <em>does<\/em> terminate in some foundational belief.)<\/p><p>An infinite regress is not necessarily absurd \u2013 some regresses are said to be \u2018virtuous\u2019 rather than \u2018vicious\u2019. But some positions lead to the ultimate absurdity: contradiction. These positions must be contradictory themselves, and therefore false. Here\u2019s another good heuristic in mathematics: if you are not sure how to prove some claim, perhaps because it seems so obvious, try <em>reductio ad absurdum<\/em> reasoning. That is, suppose that the claim is false, and show that this leads to a contradiction. This provides a <em>proof<\/em> of the claim, one in which the claim is established conclusively by that reasoning.<\/p><p>Philosophers often employ <em>reductio ad absurdum<\/em> reasoning too. They also employ a related, but less conclusive, technique in order to show that an argument is unsound: \u2018proves too much\u2019. (\u2018Proves\u2019 is tongue-in-cheek.) Start with some argument <em>A<\/em> that you think is unsound, but you cannot pinpoint exactly what is wrong with it \u2013 that\u2019s a hard problem. Parody it with another argument, <em>P<\/em>, that has the same structure as <em>A<\/em>, but whose conclusion is obviously false; thus <em>P<\/em> is obviously unsound. Then argue that since <em>A<\/em> resembles <em>P<\/em> in important respects, it too must be unsound. This is not a proof, but rather an instance of analogical reasoning. The reasoning goes that, by parity of reasoning, <em>A<\/em> must have the same status as <em>P<\/em>; and by <em>parody<\/em> of reasoning \u2013 a turn of phrase, albeit used slightly differently, that I owe to Daniel Dennett\u2019s <em>Intuition Pumps and Other Tools for Thinking<\/em> (2013) \u2013 that status is an unhappy one.<\/p><p>This strategy also resembles the mathematics heuristic that I mentioned earlier, of modifying a hard problem to make it easier. Here, we modify <em>A<\/em> to <em>P<\/em> and, in doing so, modify the hard problem of seeing that <em>A<\/em> is unsound to the easier problem of seeing that <em>P<\/em> is unsound, which it obviously is. However, unlike the mathematics heuristic, the \u2018proves too much\u2019 strategy typically does not involve going <em>back<\/em> to the original argument <em>A<\/em>, and diagnosing exactly what was wrong with <em>it<\/em>. It is tarred with <em>P<\/em>\u2019s brush, and that\u2019s supposed to be that. It\u2019s rather like solving the easier mathematics problem, and resting content. To that extent, the strategy can be unsatisfying.<\/p><p class=\"pullquote\">From Plato\u2019s cave to Singer\u2019s \u2018drowning child\u2019, analogical reasoning pervades philosophical thinking<\/p><p>Perhaps the most famous instance of the \u2018proves too much\u2019 technique is the 11th-century Benedictine monk Gaunilo\u2019s parody of St Anselm\u2019s ontological argument for the existence of God. According to the concept of God, a greater being cannot be conceived. Now, suppose that God does not exist. Then a greater being <em>could<\/em> be conceived \u2013 namely, one with God\u2019s greatness <em>and who does exist.<\/em> But this is a contradiction: a greater being than God is both inconceivable and conceivable. So we must reject the supposition \u2013 that is, we must conclude that God exists. There\u2019s an instance of <em>reductio ad absurdum<\/em> reasoning for you. Gaunilo then parodies it: consider the concept of <em>the perfect island.<\/em> A greater island cannot be conceived. Now, suppose that this island does not exist. Then a greater island <em>could<\/em> be conceived \u2013 namely, one with the island\u2019s greatness <em>and that exists<\/em>. Contradiction. So the island exists. But this is absurd. So we should reject the ontological argument, which employs parallel reasoning \u2013 it \u2018proves too much\u2019.<\/p><p>\u2018Proves too much\u2019 reasoning is a form of analogical reasoning. Now let\u2019s generalise \u2013 itself an exercise in such reasoning. Roughly, such reasoning begins by citing similarities between some entity and another one; moreover, the latter entity has a further feature; one concludes that the former entity also has that feature. Schematically:<\/p><blockquote><div>Entity X has properties F, G, H, \u2026<br \/>Entity Y also has properties F, G, H, and also <em>I.<\/em><br \/>Therefore, (plausibly) entity X also has property <em>I.<\/em><br \/><\/div><\/blockquote><p>The entities might be physical objects, such as planets, or even abstract objects, such as arguments. The properties can be similarly diverse: for example, having water and supporting life, or being unsound. The \u2018therefore\u2019 should be understood to flag an <em>inductive<\/em> inference, one in which the premises are thought to lend support to the conclusion, without guaranteeing it (hence the \u2018plausibly\u2019). In a nutshell: likeness in certain respects supports likeness in further respects.<\/p><p>Analogical reasoning has played an important role in the history of philosophy. Indeed, in <em>Philosophical Essays Concerning Human Understanding<\/em> (1748), David Hume said (perhaps overreaching): \u2018All our reasonings concerning matters of fact are founded on a species of analogy.\u2019 From Plato\u2019s allegory of the cave in the <em>Republic<\/em> to Peter Singer\u2019s \u2018drowning child\u2019 <a href=\"http:\/\/philosophyfaculty.ucsd.edu\/faculty\/rarneson\/Singeressayspring1972.pdf\" target=\"_blank\">argument<\/a>, analogical reasoning pervades philosophical thinking. But the most famous analogical argument of them all is a classic argument for the existence of God.<\/p><p>Philosophers speak of <em>the<\/em> argument from design but, attentive reader that you are, you are questioning the presupposition that there is exactly one. And indeed, there are many such arguments. I will present one, without claiming that it is the best version, but it does showcase the <a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/download?doi=10.1.1.651.1834&amp;rep=rep1&amp;type=pdf\" target=\"_blank\">various<\/a> <a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/download?doi=10.1.1.649.7937&amp;rep=rep1&amp;type=pdf\" target=\"_blank\">heuristics<\/a> that I have presented.<\/p><p>Look at a watch. You see that it is intricate, aesthetically pleasing, and behaves in a regular way. You also know that it had an intelligent designer. Now look at the world. You see that it is intricate, aesthetically pleasing, and behaves in a regular way. By analogical reasoning, you should conclude that (plausibly) it too had an intelligent designer \u2013 namely, God. Or so the argument goes. Its spirit is captured by the old hymn that begins:<\/p><blockquote><div>All things bright and beautiful, <br \/>All creatures great and small, <br \/>All things wise and wonderful, <br \/>The Lord God made them all\u2026<\/div><\/blockquote><p>However, both the argument and the hymn are easily parodied. Look again at the watch. Despite its agreeable properties, it also has flaws \u2013 it keeps time imperfectly, its batteries occasionally need replacing, it is easily scratched. So you should conclude that it had a <em>flawed<\/em> designer. Now look again at the world. It too has its flaws. Monty Python began cataloguing them:<\/p><blockquote><div>All things dull and ugly,<br \/>All creatures short and squat,<br \/>All things rude and nasty,<br \/>The Lord God made the lot \u2026<\/div><\/blockquote><p>You should conclude that the world too had a <em>flawed<\/em> designer, which is not how God is normally conceived. Hang on! The argument from design is in danger of \u2018proving too much\u2019.<\/p><p>In <em>Dialogues Concerning Natural Religion<\/em> (1779), Hume powerfully parodied a version of the argument from design. He also questioned the alleged similarity between human artefacts, of which we have experience, and the Universe; and we have no experience of other universes. Indeed, the Universe might be regarded as an <em>extreme<\/em> case of an entity, and as such quite unlike entities such as watches \u2013 we can question whether it could even enter into causal relations, like being created, at all. Finally, Hume contended that the argument of design involves an infinite regress: the intelligent designer, God, whose existence the argument purports to support, himself demands explanation, requiring a prior intelligent designer. And away we go.<\/p><p>One of the main arguments <em>against<\/em> the existence of God is <em>the problem of evil<\/em>. (Neon lights!) Consider this version of it:<\/p><blockquote><div>1.\u00a0If God existed, he would have created the best of all possible worlds.<br \/>2. Our world is not the best of all possible worlds.<\/div><\/blockquote><p>Therefore,<\/p><blockquote><div>\u00a0Conclusion: God does not exist.<\/div><\/blockquote><p>Here, \u2018worlds\u2019 are entire universes, and \u2018possible worlds\u2019 are ways that a universe could be \u2013 we might think of them as instances of what is sometimes called the \u2018multiverse\u2019.<\/p><p>Premise 1 is meant to be plausible on many of the leading conceptions of God \u2013 in particular, ones that portray him as being omnibenevolent, omnipotent and omniscient. (All three qualities are necessary for the premise to be plausible \u2013 to see why, mentally highlight each of them, and run through the alternatives to which they are opposed.) The premise invites us to imagine various possible worlds, and to imagine God choosing which of these worlds to create. Premise 2 then compares our world to some of these alternatives. It seems we should grant it, since we can easily imagine our world being better \u2013 more happy people, less suffering. (One is reminded of the old joke: an optimist thinks that this is the best of all possible worlds; a pessimist fears that this is true.)<\/p><p class=\"pullquote\">If God existed, he would have created something. Yet perhaps God exists, but didn\u2019t create anything?<\/p><p>There are some problems with this argument that our heuristics help to tease out. With the first premise, I hope you saw \u2018the\u2019 in neon lights. Is there exactly one best of all possible worlds? It seems that there could be many. For example, start with a candidate for the best world, and imagine tweaking it in a way that makes no difference to its goodness \u2013 say, moving one insignificant particle by a nanometer, or mirror-reversing everything. Offhand, the result is equally a candidate for the best world. However, the argument in turn could be tweaked accordingly. Just make this premise \u2018\u2026 he would have created <em>a<\/em> best possible world\u2019 \u2013 <em>one<\/em> of the candidates. And the similarly tweaked premise 2 looks equally secure: this is not (even) <em>a<\/em> best possible world.<\/p><p>But there seems to be more of a problem on the other side: perhaps there is <em>no<\/em> best of all possible worlds. Rather, worlds can keep getting better and better without end \u2013 perhaps just keep adding another happy person, or another happy cow. Never mind the details of how God could create these better worlds. Any limitation on his ability to do so would seem to impugn his omnipotence.<\/p><p>The form of premise 1 is that if God existed, he would have <em>created<\/em> something. But a relevant alternative is that he might not have created anything. Perhaps God exists, but did not create anything?<\/p><p>Yet one might insist that he must have, perhaps regarding that as a part of the meaning of \u2018God\u2019. This brings us to the second premise. Again, note the contrast: <em>our<\/em> world, as opposed to <em>other<\/em> worlds. This prompts a different response: God did not create our world, but he created the best of all possible worlds (instead). This suggests that the argument is invalid: we can imagine premises 1 and 2 being true, without being committed to the conclusion. Or imagine that God <em>did<\/em> create the best of all possible worlds; and the second best; and the third best\u2026 Eventually, we get to our world, which is way down the list, but he created it nonetheless \u2013 perhaps because there is still a net balance of good over evil. Again, this suggests that the argument is invalid \u2013 only \u2018suggests\u2019, mind; perhaps it is impossible for one God to create multiple worlds, for reasons given by David Lewis in <em>On the Plurality of Worlds<\/em> (1986). It presupposes that God faces a world limit.<\/p><p>Where does this leave us? Well, we did not manage to prove the existence of God, nor prove his non-existence. (I hope you are not too disappointed!) But that\u2019s par for the course in philosophy \u2013 it rarely proves <em>anything<\/em> conclusively. Instead, I hope I have given you some sense of what philosophical reasoning is like, and how that reasoning can be stimulated and enhanced by the use of various heuristics. Along the way, we saw some instances of what followed from what (or not), exposed some sophisms, spotted some fallacies, and policed some of our reasoning.<\/p><p>To be sure, the heuristics have their limits. There are many distinct abilities that go into making a good philosopher, and I do not pretend to give heuristics for all that philosophers do, or even a tenth of what they do. In particular, there are no short-cuts to profundity, and I should add that there will always be a role for good judgment and insight \u2013 just as there is in mathematics and chess. That said, heuristics can make difficult reasoning tasks easier, as much in philosophy as in mathematics and chess.<\/p>","created_at":"2018-11-07T16:51:12+0000","updated_at":"2018-11-07T16:51:12+0000","published_at":"2017-04-03T10:00:21+0000","published_by":["Alan H\u00e1jek"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":20,"domain_name":"aeon.co","preview_picture":"https:\/\/epsilon.aeon.co\/images\/0e1f4759-73f7-4fa5-a460-f94b6e5a7a50\/header_essay-final-the-karaoke.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":527,"title":"Penna \u2013 Vintage Typewriter-Inspired Bluetooth Keyboard","url":"https:\/\/news.ycombinator.com\/item?id=13981894","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D13981894\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"13981894\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/www.elretron.com\/penna\" class=\"storylink\">Penna \u2013 Vintage Typewriter-Inspired Bluetooth Keyboard<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=elretron.com\">elretron.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        64 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=tortilla\" class=\"hnuser\">tortilla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13981894\">on Mar 28, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=13981894&amp;goto=item%3Fid%3D13981894\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Penna%20%E2%80%93%20Vintage%20Typewriter-Inspired%20Bluetooth%20Keyboard&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Penna%20%E2%80%93%20Vintage%20Typewriter-Inspired%20Bluetooth%20Keyboard\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=13981894&amp;auth=12ab75bb68ba4a3b22aa5078c68a1ff2fd024bb1\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=13981894\">47\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:51:11+0000","updated_at":"2018-11-07T16:51:11+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":525,"title":"Tails - About","url":"https:\/\/tails.boum.org\/about\/index.en.html","content":"<blockquote><p><strong>amnesia<\/strong>, <em>noun<\/em>:<br \/>forgetfulness; loss of long-term memory.<\/p>\n<p><strong>incognito<\/strong>, <em>adjective &amp; adverb<\/em>:<br \/>(of a person) having one's true identity concealed.<\/p><\/blockquote><p>Tails is a live system that aims to preserve your privacy and anonymity. It\nhelps you to use the Internet anonymously and circumvent censorship almost anywhere you go and on any\ncomputer but leaving no trace unless you ask it to explicitly.<\/p><p>It is a complete operating system designed to be used from a USB stick or a DVD\nindependently of the computer's original operating system. It is <a href=\"https:\/\/tails.boum.org\/doc\/about\/license\/index.en.html\">Free\nSoftware<\/a> and based on <a href=\"https:\/\/www.debian.org\/\">Debian GNU\/Linux<\/a>.<\/p><p>Tails comes with several built-in applications pre-configured with security in\nmind: web browser, instant messaging client, email client, office suite, image\nand sound editor, etc.<\/p><div class=\"toc\"><ol><li class=\"L1\"><a href=\"#index1h1\">Online anonymity and censorship circumvention<\/a>\n\t<ol><li class=\"L2\"><a href=\"#index1h2\">Tor<\/a>\n\t\t<\/li>\n\t<\/ol><\/li>\n\t<li class=\"L1\"><a href=\"#index2h1\">Use anywhere but leave no trace<\/a>\n\t<\/li>\n\t<li class=\"L1\"><a href=\"#index3h1\">State-of-the-art cryptographic tools<\/a>\n\t<\/li>\n\t<li class=\"L1\"><a href=\"#index4h1\">What's next?<\/a>\n\t<\/li>\n\t<li class=\"L1\"><a href=\"#index5h1\">Press and media<\/a>\n\t<\/li>\n\t<li class=\"L1\"><a href=\"#index6h1\">Social Contract<\/a>\n\t<\/li>\n\t<li class=\"L1\"><a href=\"#index7h1\">Acknowledgments and similar projects<\/a>\n\t<\/li>\n\t<li class=\"L1\"><a href=\"#index8h1\">Contact<\/a>\n\t<\/li>\n<\/ol><\/div><p>Tails relies on the Tor anonymity network to protect your privacy\nonline:<\/p><ul><li>all software is configured to connect to the Internet through Tor<\/li>\n<li>if an application tries to connect to the Internet directly, the\nconnection is automatically blocked for security.<\/li>\n<\/ul><p>Tor is an open and distributed network that helps defend\nagainst traffic analysis, a form of\nnetwork surveillance that threatens personal freedom and privacy, confidential\nbusiness activities and relationships, and state security.<\/p><p>Tor protects you by bouncing your communications around a network of\nrelays run by volunteers all around the world: it prevents somebody watching\nyour Internet connection from learning what sites you visit, and it prevents the\nsites you visit from learning your physical location.<\/p><p>Using Tor you can:<\/p><ul><li>be anonymous online by hiding your location,<\/li>\n<li>connect to services that would be censored otherwise;<\/li>\n<li>resist attacks that block the usage of Tor using circumvention tools\nsuch as <a href=\"https:\/\/tails.boum.org\/doc\/first_steps\/startup_options\/bridge_mode\/index.en.html\">bridges<\/a>.<\/li>\n<\/ul><p>To learn more about Tor, see the official <a href=\"https:\/\/www.torproject.org\/\">Tor\nwebsite<\/a>, particularly the following pages:<\/p><ul><li><a href=\"https:\/\/www.torproject.org\/about\/overview.html.en#whyweneedtor\">Tor overview: Why we need\nTor<\/a><\/li>\n<li><a href=\"https:\/\/www.torproject.org\/about\/overview.html.en#thesolution\">Tor overview: How does Tor\nwork<\/a><\/li>\n<li><a href=\"https:\/\/www.torproject.org\/about\/torusers.html.en\">Who uses Tor?<\/a><\/li>\n<li><a href=\"https:\/\/trac.torproject.org\/projects\/tor\/wiki\/doc\/TorALaymansGuide\">Understanding and Using Tor \u2014 An Introduction for the\nLayman<\/a><\/li>\n<\/ul><p>To learn more about how Tails ensures all its network connections use Tor,\nsee our <a href=\"https:\/\/tails.boum.org\/contribute\/design\/Tor_enforcement\/\">design document<\/a>.<\/p><p>Using Tails on a computer doesn't alter or depend on the operating system\ninstalled on it. So you can use it in the same way on your computer, a\nfriend's computer, or one at your local library. After shutting down Tails,\nthe computer will start again with its usual operating system.<\/p><p>Tails is configured with special care to not use the computer's hard-disks,\neven if there is some swap space on them. The only storage space used by Tails is\nin RAM, which is automatically erased when the computer shuts down. So\nyou won't leave any trace on the computer either of the Tails system itself or\nwhat you used it for. That's why we call Tails \"amnesic\".<\/p><p>This allows you to work with sensitive documents on any computer and protects you\nfrom data recovery after shutdown. Of course, you can still explicitly save\nspecific documents to another USB stick or external hard-disk and take them away for\nfuture use.<\/p><p>Tails also comes with a selection of tools to protect your data using strong\nencryption:<\/p><ul><li><p><a href=\"https:\/\/tails.boum.org\/doc\/encryption_and_privacy\/encrypted_volumes\/index.en.html\">Encrypt your USB sticks or external\nhard-disks<\/a> using <a href=\"https:\/\/en.wikipedia.org\/wiki\/Linux%5FUnified%5FKey%5FSetup\">LUKS<\/a>,\nthe Linux standard for disk-encryption.<\/p><\/li>\n<li><p>Automatically use HTTPS to encrypt all your communications to a number of\nmajor websites using <a href=\"https:\/\/www.eff.org\/https-everywhere\">HTTPS\nEverywhere<\/a>, a Firefox extension\ndeveloped by the <a href=\"https:\/\/www.eff.org\">Electronic Frontier Foundation<\/a>.<\/p><\/li>\n<li><p>Encrypt and sign your emails and documents using the <em>de facto<\/em> standard\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Pretty%5FGood%5FPrivacy%23OpenPGP\">OpenPGP<\/a>\neither from Tails email client, text editor or file browser.<\/p><\/li>\n<li><p>Protect your instant messaging conversations using <a href=\"https:\/\/en.wikipedia.org\/wiki\/Off%2Dthe%2DRecord%5FMessaging\">OTR<\/a>, a cryptographic tool that provides encryption,\nauthentication and deniability.<\/p><\/li>\n<li><p><a href=\"https:\/\/tails.boum.org\/doc\/encryption_and_privacy\/secure_deletion\/index.en.html\">Securely delete your files<\/a>\nand clean your diskspace using <a href=\"http:\/\/wipetools.tuxfamily.org\/nautilus-wipe.html\">Nautilus\nWipe<\/a>.<\/p><\/li>\n<\/ul><p><a href=\"https:\/\/tails.boum.org\/doc\/encryption_and_privacy\/index.en.html\">Read more about those tools in the documentation.<\/a><\/p><p>To continue discovering Tails, you can now read:<\/p><ul><li>the <a href=\"https:\/\/tails.boum.org\/doc\/about\/warning\/index.en.html\">warning page<\/a> to better understand the security limitations of Tails and Tor,<\/li>\n<li>more details about the <a href=\"https:\/\/tails.boum.org\/doc\/about\/features\/index.en.html\">features and software<\/a> included in Tails,<\/li>\n<li>our <a href=\"https:\/\/tails.boum.org\/install\/index.en.html\">installation instructions<\/a> to download and install Tails,<\/li>\n<li>our <a href=\"https:\/\/tails.boum.org\/doc\/index.en.html\">documentation<\/a> explaining in detail how to use Tails,<\/li>\n<li>some hints on why <a href=\"https:\/\/tails.boum.org\/doc\/about\/trust\/index.en.html\">you should trust Tails<\/a>,<\/li>\n<li>our <a href=\"https:\/\/tails.boum.org\/contribute\/design\/\">design document<\/a> laying out Tails specification, threat model and implementation,<\/li>\n<li>the <a href=\"https:\/\/tails.boum.org\/contribute\/calendar\/\">calendar<\/a> that holds our release dates, meetings and other events.<\/li>\n<\/ul><p>See <a href=\"https:\/\/tails.boum.org\/press\/index.en.html\">Press and media information<\/a>.<\/p><p>Read our <a href=\"https:\/\/tails.boum.org\/contribute\/working_together\/social_contract\/\">Social Contract<\/a>.<\/p><p>See <a href=\"https:\/\/tails.boum.org\/doc\/about\/acknowledgments_and_similar_projects\/index.en.html\">Acknowledgments and similar projects<\/a>.<\/p><p>See the <a href=\"https:\/\/tails.boum.org\/about\/contact\/index.en.html\">contact page<\/a>.<\/p>","created_at":"2018-11-07T16:50:42+0000","updated_at":"2018-11-07T16:50:42+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"tails.boum.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":526,"title":"BONUS! We Got A Call From An Actual Flat-Earther","url":"https:\/\/www.echoplexmedia.com\/the-plex\/2016\/2\/28\/bonus-flat-earth-interview","content":"<div class=\"sqs-block button-block sqs-block-button\" data-block-type=\"53\" id=\"block-yui_3_17_2_3_1456708800835_16501\"><div class=\"sqs-block-content\">\n<div class=\"sqs-block-button-container--center\" data-alignment=\"center\" data-button-size=\"large\">\n\t<a href=\"http:\/\/198.199.105.67:8000\/stream\" class=\"sqs-block-button-element--large sqs-block-button-element\" target=\"_blank\">CLICK HERE TO LISTEN TO THE PLEX LIVE! SUNDAYS 7PM PACIFIC<\/a>\n<\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-afac29530402ea96c7e9\"><div class=\"sqs-block-content\"><p>There was a great deal of conjecture as to whether he would actually put his money where his mouth is, but sure enough, Flat Earth evangelist and village idiot Wayne Chipman called and spoke with Echoplex Media's own Fuckin' Harrison (<a target=\"_blank\" href=\"http:\/\/www.twitter.com\/effin_harrison\">@effin_harrison<\/a>) about the precise nature of his belief that the Earth is in fact NOT a globe, but a disc. Possibly balanced on the back of a giant turtle, standing on an infinite tower of elephants. Just...trust me, it's gotta be heard to be believed.<\/p><p>EDITOR'S NOTE: Not a scrap of editing was done to this call, from the moment we picked up the phone until the moment we disconnected. You're getting the full context, straight with no chaser. He's really that weird.<\/p><p>Don't forget check out <a target=\"_blank\" href=\"http:\/\/www.echoplexmedia.com\/the-plex\">The Plex<\/a> and <a target=\"_blank\" href=\"http:\/\/www.echoplexmedia.com\/local-love\/\">Local Love<\/a> LIVE every Sunday and Tuesday from 9 - midnight PST! Also, make sure to check out <a target=\"_blank\" href=\"http:\/\/www.echoplexmedia.com\/face-for-radio\/\">Face For Radio<\/a> podcast!<\/p><p>Want to weigh in on what a looney toon Wayne Chipman is? Call the Echoplex Hotline at <strong>408-883-7539<\/strong>\u00a0and leave us a voice mail! We'll play the funniest ones on the next show!<\/p><\/div><\/div>","created_at":"2018-11-07T16:50:42+0000","updated_at":"2018-11-07T16:50:42+0000","published_at":"2016-02-28T00:00:00+0000","published_by":["Echoplex Podcast"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.echoplexmedia.com","preview_picture":"http:\/\/static1.squarespace.com\/static\/5574c851e4b092aac3829b56\/5574ca29e4b0c6a58c759c6a\/56d35cebc6fc087c73ffc970\/1509232128899\/flat-earth-theory.jpg?format=1000w","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":523,"title":"How To Learn Solidity: The Ultimate Ethereum Coding Guide - Blockgeeks","url":"https:\/\/blockgeeks.com\/guides\/how-to-learn-solidity\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:50:41+0000","updated_at":"2018-11-07T16:50:41+0000","annotations":[],"reading_time":0,"domain_name":"blockgeeks.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":524,"title":"Tips for Using Mutt | W. Caleb McDaniel","url":"http:\/\/wcm1.web.rice.edu\/mutt-tips.html","content":"<p>hacks<\/p><p>Posted by W. Caleb McDaniel\n\non July 16, 2012<\/p><p>For about six months now, I\u2019ve been using <a href=\"http:\/\/www.mutt.org\/\" title=\"Mutt\">Mutt<\/a> as my email client. It\u2019s not perfect (as the tagline on the official website admits), but it works for me.<\/p><p>It took a long time of hacking around before Mutt began to save me time, and the initial time investment probably isn\u2019t worth it unless you enjoy the journey as much as the destination. Philosopher and programmer extraordinaire <a href=\"http:\/\/john.macfarlane.usesthis.com\/\" title=\"Interview with John MacFarlane on UsesThis.com\">John MacFarlane<\/a> recently confessed that \u201cI would like to think that the time I\u2019ve spent developing tools to make writing more efficient has been repaid by the time I\u2019ve saved using these tools, but I\u2019m pretty sure it hasn\u2019t.\u201d The same probably applies to Mutt.<\/p><p>That disclaimer means that this essay is less a sales pitch than a running log of tricks I\u2019ve learned about Mutt along the way, aimed at someone who has already imbibed enough of the kool-aid to install the program and give it a try. If there\u2019s nothing here you didn\u2019t already know, you can also check out my <a href=\"http:\/\/pinboard.in\/u:wcaleb\/t:mutt\" title=\"wcaleb's Pinboard links on Mutt\">Pinboard links<\/a> on Mutt or my <a href=\"http:\/\/github.com\/wcaleb\/dotfiles\/blob\/master\/_mutt\/main.muttrc\" title=\"wcaleb's muttrc on Github\">muttrc<\/a>, but the best source of information remains the <a href=\"http:\/\/www.mutt.org\/doc\/manual\/\" title=\"Mutt manual\">manual<\/a>.<\/p><h2 id=\"why-when-i-started-using-mutt\"><del>Why<\/del> When I Started Using Mutt<\/h2><p>This was originally going to be a section on <em>why<\/em> I use Mutt, but as I have already implied, Mutt has its reasons of which Reason knows nothing. Besides, I am an historian by trade, so I tend to think that \u201cwhy\u201d is often a question of <em>when<\/em>.<\/p><p>If I recall correctly, I started using Mutt not long after reading Stephen Ramsay\u2019s <a href=\"http:\/\/stephenramsay.us\/2011\/04\/09\/life-on-the-command-line\/\" title=\"Life on the Command Line by Stephen Ramsay\">Life on the Command Line<\/a>, which included a sales pitch for an even more archaic email program (or set of programs) called <a href=\"http:\/\/www.nongnu.org\/nmh\/\" title=\"nmh - New Mail Handling System\">nmh<\/a>. I was interested enough in his description to read a little more about nmh, but that didn\u2019t take long because there\u2019s not much support documentation for it on the web. Or, at least, there wasn\u2019t enough to read that was intelligible to me. At that time, I had just started learning about Unix, and while I was ready to try some command-line tools, it was clear that Mutt would be a more comfortable place to start than nmh.<\/p><p>What appealed to me about both nmh and Mutt was the ability to use a plain text editor like Vim to edit email. Around the time I read Ramsay\u2019s post, I had begun to experiment with plain text for doing my scholarly writing, largely for the reasons that Lincoln Mullen later detailed in <a href=\"http:\/\/chronicle.com\/blogs\/profhacker\/writing-power-tools-text-editors\/38940\" title=\"Writing Power Tools: Text Editors\">this post<\/a>. I was using the excellent editor <a href=\"http:\/\/www.barebones.com\/products\/TextWrangler\/\" title=\"TextWrangler from Bare Bones Software\">TextWrangler<\/a>, but my brother had also been twisting my arm to try Vim. I was worried about the steep learning curve for Vim, but using it for email seemed like a good way to test it out and see if I wanted to use it exclusively. (Turns out, I did.)<\/p><p>The other thing that appealed to me about Ramsay\u2019s description of nmh was the ability to \u201cpipe\u201d email messages directly to other command-line programs. I had recently developed a <a href=\"http:\/\/wcm1.web.rice.edu\/plain-text-gtd.html\">plain-text system for Getting Things Done<\/a>, and I was also beginning to experiment with a workflow for <a href=\"http:\/\/wcm1.web.rice.edu\/grading-with-my-ipad.html\">grading on my iPad<\/a>. I imagined using a program like nmh or Mutt to take an email with a student paper, press one or two keys, and have the paper converted to PDF, moved to a folder that synchronized with my iPad, and added to my GTD system. Eventually I figured out how to get Mutt to do that, too, but first there were a few essential tweaks I had to implement just to make Mutt work-able for everyday mail.<\/p><h2 id=\"essential-tweaks-and-tips\">Essential Tweaks and Tips<\/h2><h3 id=\"format-flowed\">Format-Flowed<\/h3><p>It took me a couple of months of using Mutt before I realized that all the messages I was sending to other people included weird line breaks. That\u2019s because I didn\u2019t know about <code>format=flowed<\/code>. You can read all about it <a href=\"http:\/\/joeclark.org\/ffaq.html\">here<\/a>, or you can just trust me when I say that you <em>want<\/em> all of your plain text messages to be in this format. Otherwise, they will look ugly and barely readable when your recipients open them on their iPhones or email clients, as they probably will.<\/p><p>To set this up requires small changes both within Mutt and within Vim. In my <code>muttrc<\/code> I had to add this line<\/p><pre>set text_flowed=yes<\/pre><p>Then, within Vim, I edited the filetype file that Vim uses for mail to include this line:<\/p><pre>setlocal fo+=aw<\/pre><p>So far, those two tweaks seem to be working. No more ugly linebreaks!<\/p><h3 id=\"setting-up-contacts\">Setting Up Contacts<\/h3><p>From the beginning, I knew that Mutt wasn\u2019t going to be worth my time if I couldn\u2019t easily import or auto-complete my contacts when writing emails. Fortunately, Mutt has its own syntax for creating <a href=\"http:\/\/wiki.mutt.org\/?MuttGuide\/Aliases\" title=\"Mutt User Manual on aliases\">aliases<\/a>, which basically means creating a long text file with lines that look like this:<\/p><pre>alias mcdaniel-caleb Caleb McDaniel &lt;example@email.com&gt;<\/pre><p>That makes it possible to simply type <code>mcdaniel-caleb<\/code> (or a part of it) when composing a new message and hit \u201ctab\u201d to have it completed from the alias file. Hitting tab a second time shows all of the matches in the alias file, which can also be searched from within Mutt, so it\u2019s never hard to find an alias \u2026 assuming the alias is there.<\/p><p>The practical question I faced was how to build an alias file from scratch, without having to look up all the emails I cared about and convert them into the Mutt alias syntax. I wanted the alias file to be created automatically, so that emails I needed would auto-magically appear in my \u201cto\u201d line, much like they did when I was using Apple Mail.<\/p><p>The solution I found was to use an obscure Mutt setting called <code>set display_filter<\/code>. Typically I think this setting is used to pass a message through an external script that filters out any unwanted text before displaying it. So, for example, if I never wanted to see the words \u201cstrategic dynamism\u201d in an email, I could write a shell script that simply used <code>sed<\/code> or some-such tool to remove or replace the phrase before showing me the message.<\/p><p>Thinking of this setting simply as a \u201cdisplay\u201d filter limits its real utility, though. What it really does is this: <em>it pipes every message you look at through a shell script\u2014any shell script\u2014before showing it to you<\/em>. Once I realized that, I wrote a script that takes a message, finds the email address of the sender, creates a Mutt alias from that address, checks to see if that alias is already in my alias file, and then, if not, appends the new alias to my contact list. Finally, the script shows me the original, unmodified message. The script looks like this:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode bash\">#!\/bin\/sh\nMESSAGE=$(cat)\nNEWALIAS=$(echo \"${MESSAGE}\" | grep ^\"From: \" | sed s\/[\\,\\\"\\']\/\/g | awk '{$1=\"\"; if (NF == 3) {print \"alias\" $0;} else if (NF == 2) {print \"alias\" $0 $0;} else if (NF &gt; 3) {print \"alias\", tolower($(NF-1))\"-\"tolower($2) $0;}}')\nif grep -Fxq \"$NEWALIAS\" $HOME\/.mutt\/aliases.txt; then\n    :\nelse\n    echo \"$NEWALIAS\" &gt;&gt; $HOME\/.mutt\/aliases.txt\nfi\necho \"${MESSAGE}\"<\/pre><\/div><p>All of this happens faster than the blink of an eye, and the result is this: anytime I choose to open a message in Mutt, the sender is automatically added to my alias list.<\/p><p>This method allowed me to rapidly create an alias file from scratch, and it usually means that when I press \u201ctab\u201d to find an alias, I can expect it to already be there. But this isn\u2019t a perfect solution for a couple of reasons. One, there are some services (I\u2019m looking at you, Evite) which send emails with a personal name in the \u201cfrom\u201d field even though the field contains a service-specific reply-to address. Two, this methods makes for a very long alias file that sometimes contains ugly aliases from mailing lists. But the first issue has only been a problem for me once, and it\u2019s easily fixed by deleting the \u201cbad\u201d alias from the alias file. And the second issue is seldom a problem because I rarely open the aliases text file.<\/p><p>Having a lengthy alias file also makes it easy to look up email addresses simply by piping the alias file through <code>grep<\/code>. This can be especially useful when writing an email within Vim. Let\u2019s say I wanted to send Joe Smith\u2019s email address to a third party; while writing that email, I could simply run this command within Vim:<\/p><pre>:r !grep smith-joe \/path\/to\/aliases.txt<\/pre><p>That will (usually) put the relevant line from my aliases file directly into my email message, without my ever having to leave Vim or open up a new window. If this was a frequent use case, I could even make a user-defined command for my <code>vimrc<\/code> that would grab just the email address and use a shorter syntax.<\/p><p>Despite the advantages of using the Mutt alias file, I also sometimes want to be able to look up addresses in my Apple Address Book from within Mutt. I use <a href=\"http:\/\/oif.eafarris.com\/blog\/hooking-os-xs-address-book-to-mutt\">this handy method<\/a> to do that. Using these two methods together, I would estimate that I am able to easily autocomplete email addresses 98% of the time. When composing a message, I just tap the first few letters of the recipient\u2019s last name and press tab once or (if there are multiple matches) twice. If no Mutt-defined aliases comes up, I just finish typing the last name and hit Ctrl-T, and 9 times out of 10 it grabs the email address from my Address Book instead.<\/p><h3 id=\"using-offlineimap-and-notmuch\">Using <code>offlineimap<\/code> and <code>notmuch<\/code><\/h3><p>My current Mutt setup allows me to sign in directly to my university IMAP account and edit messages, tags, and folders directly on the server. Sometimes, however, you want to be able to look at messages while offline. To do that, I use <a href=\"http:\/\/offlineimap.org\/\" title=\"OfflineIMAP homepage\">offlineimap<\/a>, paired with <a href=\"http:\/\/notmuchmail.org\/\" title=\"notmuch homepage\">notmuch<\/a> for searching through all my downloaded messages. I periodically run <code>offlineimap<\/code> from the command line to update my local store of messages, and when I want to browse or search through the full-text of old messages, I type <code>muff<\/code>. Muff is a bash alias I have defined to run<\/p><pre>mutt -F $HOME\/.muffrc<\/pre><p>where <code>$HOME\/.muffrc<\/code> is a small mutt configuration file that points Mutt to my local email store instead of to my IMAP server.<\/p><p>I didn\u2019t know about <code>notmuch<\/code> until recently, and had to do some hacking to get it installed. It\u2019s a great feature, though, because Mutt\u2019s search functions take forever to search the bodies of email messages online. Mutt works great for finding messages if you know the sender, recipient, or a keyword in the subject, but if you want to search full messages or attachments, <code>notmuch<\/code> is the way to go.<\/p><p>To install <code>notmuch<\/code> on Lion, I had to use Macports to install the following dependencies:<\/p><ul><li>Xapian-core 1.2.8<\/li>\n<li>Xapian-bindings 1.2.8<\/li>\n<li>Talloc 2.0.1<\/li>\n<li>Gmime 2.4.23<\/li>\n<\/ul><p>When running the local <code>configure<\/code> command before installing notmuch, I added the <code>--without-emacs<\/code> flag. And before <code>notmuch<\/code> would work, I had to create a <code>.cache<\/code> directory in my $HOME directory.<\/p><p>Once I installed <code>notmuch<\/code>, I used <a href=\"http:\/\/log.or.cz\/?p=228\">these instructions<\/a> to add some keybindings to my <code>$HOME\/.muffrc<\/code> file that allow me to search all of my downloaded messages quickly.<\/p><h3 id=\"attachments-and-the-.mailcap-file\">Attachments and the .mailcap file<\/h3><p>Viewing attachments is perhaps the least intuitive part of Mutt. It requires creating a <code>~\/.mailcap<\/code> file that specifies which applications should be used to open which kinds of attachments. I basically adopted <a href=\"http:\/\/linsec.ca\/Using_mutt_on_OS_X#mailcap\">this system wholesale<\/a>, especially the <code>view_attachment.sh<\/code> script.<\/p><p>For example, these lines in my <code>.mailcap<\/code> file tell Mutt to open Microsoft Word documents in TextEdit, while allowing it to open *.docx files in Microsoft Word. Rich Text files are opened using <a href=\"http:\/\/www.bean-osx.com\/Bean.html\" title=\"Bean homepage\">Bean<\/a>. HTML attachments and messages are piped through <a href=\"http:\/\/johnmacfarlane.net\/pandoc\/\" title=\"Pandoc homepage\">pandoc<\/a> and converted to markdown for easy reading.<\/p><pre>application\/msword; \/Users\/wcm1\/Scripts\/viewattach.sh %s \"-\" '\/Applications\/TextEdit.app' ; print=\/Users\/wcm1\/Scripts\/printattach.sh %s \napplication\/vnd.openxmlformats-officedocument.wordprocessingml.document; \/Users\/wcm1\/Scripts\/viewattach.sh %s \"-\" ; print=\/Users\/wcm1\/Scripts\/printattach.sh %s \napplication\/rtf; \/Users\/wcm1\/Scripts\/viewattach.sh %s \"-\" '\/Applications\/Bean.app' ; print=\/Users\/wcm1\/Scripts\/printattach.sh %s\napplication\/pdf; \/Users\/wcm1\/Scripts\/viewattach.sh %s \"-\" '\/Applications\/Preview.app' ; print=\/Users\/wcm1\/Scripts\/pdfattach.sh %s \ntext\/html; pandoc -f html -t markdown; copiousoutput; compose=vim %s<\/pre><p>For the first four of these lines, the text that follows the second semicolon is not essential for viewing attachments. But using the \u201cprint\u201d setting in a <code>.mailcap<\/code> file has allowed me to further extend Mutt\u2019s ability to work with attachments, as described below.<\/p><h2 id=\"my-peculiar-customizations\">My Peculiar Customizations<\/h2><p>The tips above are essential, I\u2019ve discovered, just for making Mutt workable as an email client. But I originally wanted to use Mutt for peculiar reasons specific to my workflow. In particular, I wanted a way to get the body of email messages into my plain-text GTD system. And because I receive lots of files that I need to convert to PDF, move to my iPad, and annotate or grade, I wanted a way to quickly do that without leaving the keyboard.<\/p><h3 id=\"piping-attachments\">Piping Attachments<\/h3><p>Unfortunately, unlike the nmh system that Ramsay described, Mutt doesn\u2019t have an obvious way to pipe attachments like Microsoft Word documents to other Unix commands. Eventually I discovered a workaround using the \u201cprint\u201d flag in my <code>mailcap<\/code> file. Like the <code>set display_filter<\/code> option, the <code>mailcap<\/code> print flag is intended for a fairly narrow purpose: it allows you to specify what command Mutt will use to send an attachment to the printer, like <code>lpr<\/code>. At some point, however, I realized or read that you could feed any command to the \u201cprint\u201d flag. Printing from the attachment menu in Mutt will basically \u201cpipe\u201d the attachment to the command specified in the <code>mailcap<\/code> entry under \u201cprint.\u201d<\/p><p>That\u2019s why the <code>mailcap<\/code> entries I listed above often include something like this at the end:<\/p><pre>print=\/Users\/wcm1\/Scripts\/printattach.sh %s <\/pre><p>The script <code>printattach.sh<\/code> is just a simple shell script that accomplishes some of the idiosyncratic things I often want to do with my attachments. For example, when presented with a Word file, I usually want to do one of several things: convert it to a PDF and move it to a folder that syncs with my iPad; add a grading rubric if it\u2019s a student paper; and\/or add a note to my GTD system reminding me to read the paper. So I wrote a script that looks like this:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode bash\">#!\/bin\/bash\n#\n# Based on save_attachment.sh by Eric Gebhart\n# Mutt puts everything in \/tmp by default. \n# This gets the basic filename from the full pathname.\nnewfile=`basename $1`\n# Copy the file to our new spot so mutt will not delete it\n# before the script has a chance to print it.\ncp $1 $newfile\n# Prompt user for desired script.\necho \"What do you want to do? (a)nnotate, (r)ubric, (i)nbox or (o)ther folder?\"\nread -n 1 SWITCH\necho \"\"\nif [ $SWITCH = \"a\" ]; then\n    $HOME\/Scripts\/annotate.sh $newfile\nelif [ $SWITCH = \"r\" ]; then\n    $HOME\/Scripts\/any2pdf.sh $newfile\n    pdffile=$(echo $newfile | sed -E 's\/\\.do[cx]+\/\\.pdf\/g')\n    $HOME\/Scripts\/append.rb $pdffile &amp;&amp; mv $pdffile $HOME\/.Trash\/\n    mv $newfile $HOME\/.Trash\/\nelif [ $SWITCH = \"i\" ]; then\n    $HOME\/Scripts\/any2pdf.sh -o $HOME\/Inbox\/ $newfile\n    mv $newfile $HOME\/.Trash\/\nelse\n    echo \"Which folder in home do you want to print PDF to?\"\n    read OTHERPATH\n    $HOME\/Scripts\/any2pdf.sh -o \/Users\/wcm1\/\"$OTHERPATH\" $newfile\n    mv $newfile $HOME\/.Trash\/\nfi<\/pre><\/div><p>Assuming this script is listed in my <code>mailcap<\/code> file under the appropriate entries, here\u2019s what happens when I choose to \u201cprint\u201d an attachment in Mutt: I get a brief dialog asking me what I want to do. Based on my input, the script then re-routes the file to other scripts that accomplish my particular tasks. It\u2019s a somewhat clunky workaround for piping attachments, but it works great and is, by its very nature, extremely customizable.<\/p><h3 id=\"integration-with-gtd\">Integration with GTD<\/h3><p>The other peculiar goal I wanted to accomplish with Mutt was a way to get the body of email messages directly into <a href=\"http:\/\/wcm1.web.rice.edu\/plain-text-gtd.html\">my plain-text GTD system<\/a>. This problem was actually much easier to solve because Mutt does have excellent support for piping <em>messages<\/em> to shell commands. Within Mutt\u2019s pager, all I have to do is hit the vertical bar to pipe the message to any command I choose. So I wrote up a simple script that takes the message, cleans it up, and puts it in a plain text file whose title I specify within Mutt.<\/p><p>In practice, that means if I get a message I need to act on, I hit <code>|<\/code> and then type <code>mqq \"Do something about this\"<\/code>. Behind the scenes, the script will put that \u201cto do\u201d in the folder where I keep my tasks, together with a copy of the message that I can refer to later from within Notational Velocity or any other text editor I choose.<\/p><p>Like many of the other tasks I\u2019ve described here, it was possible to do something similar from within Apple Mail, but using a GUI program would have required me (in this example), to use the mouse or a hotkey to copy the message text, which I then probably would have pasted in Notational Velocity. Those two steps are now reduced to one that does everything without having to leave the keyboard. Whether that\u2019s a savings that justifies using Mutt remains an open question, but for now I\u2019m happy with this setup\u2014happy enough to leave it alone for the foreseeable future!<\/p>","created_at":"2018-11-07T16:50:41+0000","updated_at":"2018-11-07T16:50:41+0000","annotations":[],"mimetype":"text\/html","reading_time":14,"domain_name":"wcm1.web.rice.edu","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":522,"title":"All commands sorted by votes","url":"https:\/\/www.commandlinefu.com\/commands\/browse\/sort-by-votes","content":"<div class=\"box-copy\"><div class=\"toggleable\"><p><strong>commandlinefu.com<\/strong> is the place to record those command-line gems that you return to again and again.\n        That way others can gain from your <abbr title=\"Command-line interface\">CLI<\/abbr> wisdom and you from theirs too.  All commands can be commented on, discussed and \n        voted up or down. <\/p><p class=\"text-center\"><a class=\"btn btn-warning\" href=\"https:\/\/www.commandlinefu.com\/commands\/edit\"><strong><i class=\"glyphicon glyphicon-plus\"> Share Your Commands<\/i><\/strong><\/a><\/p><\/div><\/div><div class=\"box-copy\"><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/1528\/doing-some-floating-point-calculations-with-rounding-e.g.-at-the-3rd-decimal\">Doing some floating point calculations with rounding (e.g. at the 3rd decimal)<\/a><\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/9884\/convert-ascii-string-to-hex\">convert ascii string to hex<\/a><\/p><p>You can use \"decode()\" in a similar manner:\n$ python -c 'print \"68656c6c6f\".decode(\"hex\")'<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/13049\/grabs-a-random-image-from-wallpapers-and-sets-as-the-background\">Grabs a random image from \"~\/wallpapers\" and sets as the background<\/a><\/p><p>don't bother spawning a bc process or counting the number of options, just pick a random one.\n'sort -R' sorts randomly, so pick the top one.<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/13320\/camelcase-to-underscore\">camelcase to underscore<\/a><\/p><p>Convert a camelCase string into snake_case.\nTo complement senorpedro's command.<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/23444\/which-processes-are-listening-on-a-specific-port-e.g.-port-80\">Which processes are listening on a specific port (e.g. port 80)<\/a><\/p><p>swap out \"80\" for your port of interest. Can use port number or named ports e.g. \"http\"<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/290\/see-the-top-10-ip-addresses-in-a-web-access-log\">See the top 10 IP addresses in a web access log<\/a><\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/5599\/ping-mac-address\">ping MAC ADDRESS<\/a><\/p><p># first install arp-scan if not have it\narp-scan 10.1.1.0\/24 .... show ip+mac in localnet\nawk '\/00:1b:11:dc:a9:65\/ {print $1}' .... get ip associated with MAC\n` backtick make do command substitution passing ip to command ping<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/8751\/use-acpi-and-notify-send-to-report-current-temperature-every-five-minutes.\">Use acpi and notify-send to report current temperature every five minutes.<\/a><\/p><p>works best in a shell script run at startup. It will ping localhost once and output to null, after it does that, acpi is called for temperature in fahrenheit and piped through to another loop that feeds notify-send for a tooltip. After waiting five minutes, it will start over.<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/6727\/sudo-for-launching-gui-apps-in-background\">sudo for launching gui apps in background<\/a><\/p><p>Need package: gksu\nNote: \nLaunching gui app in background that needs sudo, won't work great with our old \nfriendly style of launching:\n$ sudo gedit \/etc\/passwd &amp;\nbecause this would put sudo in background !\nUsing gksudo as demonstrated, would popup a gui sudo window.\nMay be this is a common knowledge, but not knowing this\nfrustrated me during my newbie year.<\/p><p><a href=\"https:\/\/www.commandlinefu.com\/commands\/view\/5597\/force-a-rescan-on-a-host-of-scsi-devices-useful-for-adding-partitions-to-vmware-on-the-fly\">force a rescan on a host of scsi devices (useful for adding partitions to vmware on the fly)<\/a><\/p><\/div>","created_at":"2018-11-07T16:50:40+0000","updated_at":"2018-11-07T16:50:40+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.commandlinefu.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":519,"title":"Free Ideas for UI Frameworks, or How To Achieve Polished UI \u2013 ChrisLord.net","url":"http:\/\/chrislord.net\/2017\/05\/27\/free-ideas-for-ui-frameworks-or-how-to-achieve-polished-ui\/","content":"<p>Ever since the original iPhone came out, I\u2019ve had several ideas about how they managed to achieve such fluidity with relatively mediocre hardware. I mean, it was good at the time, but Android still struggles on hardware that makes that look like a 486\u2026 It\u2019s absolutely my fault that none of these have been implemented in any open-source framework I\u2019m aware of, so instead of sitting on these ideas and trotting them out at the pub every few months as we reminisce over what could have been, I\u2019m writing about them here. I\u2019m hoping that either someone takes them and runs with them, or that they get thoroughly debunked and I\u2019m made to look like an idiot. The third option is of course that they\u2019re ignored, which I think would be a shame, but given I\u2019ve not managed to get the opportunity to implement them over the last decade, that would hardly be surprising. I feel I should clarify that these aren\u2019t <em>all<\/em> my ideas, but include a mix of observation of and conjecture about contemporary software. This somewhat follows on from the post I made <a href=\"http:\/\/chrislord.net\/index.php\/2011\/04\/07\/more-lessons-to-learn-from-games-development\/\">6 years ago(!)<\/a> So let\u2019s begin.<\/p>\n<h2>1. No main-thread UI<\/h2>\n<p>The UI should always be able to start drawing when necessary. As careful as you may be, it\u2019s practically impossible to write software that will remain perfectly fluid when the UI can be blocked by arbitrary processing. This seems like an obvious one to me, but I suppose the problem is that legacy makes it very difficult to adopt this at a later date. That said, difficult but not impossible. All the major web browsers have adopted this policy, with caveats here and there. The trick is to switch from the idea of \u2018painting\u2019 to the idea of \u2018assembling\u2019 and then using a compositor to do the painting. Easier said than done of course, most frameworks include the ability to extend painting in a way that would make it impossible to switch to a different thread without breaking things. But as long as it\u2019s possible to block UI, it will inevitably happen.<\/p>\n<h2>2. Contextually-aware compositor<\/h2>\n<p>This follows on from the first point; what\u2019s the use of having non-blocking UI if it can\u2019t respond? Input needs to be handled away from the main thread also, and the compositor (or whatever you want to call the thread that is handling painting) needs to have enough context available that the first response to user input doesn\u2019t need to travel to the main thread. Things like hover states, active states, animations, pinch-to-zoom and scrolling all need to be initiated without interaction on the main thread. Of course, main thread interaction will likely eventually be required to update the view, but that initial response needs to be able to happen without it. This is another seemingly obvious one \u2013 how can you guarantee a response rate unless you have a thread dedicated to responding within that time? Most browsers are doing this, but not going far enough in my opinion. Scrolling and zooming are often catered for, but not hover\/active states, or initialising animations (note; <em>initialising<\/em> animations. Once they\u2019ve been initialised, they are indeed run on the compositor, usually).<\/p>\n<h2>3. Memory bandwidth budget<\/h2>\n<p>This is one of the less obvious ideas and something I\u2019ve really wanted to have a go at implementing, but never had the opportunity. A problem I saw a lot while working on the platform for both Firefox for Android and FirefoxOS is that given the work-load of a web browser (which is not entirely dissimilar to the work-load of any information-heavy UI), it was very easy to saturate memory bandwidth. And once you saturate memory bandwidth, you end up having to block somewhere, and painting gets delayed. We\u2019re assuming UI updates are asynchronous (because of course \u2013 otherwise we\u2019re blocking on the main thread). I suggest that it\u2019s worth tracking frame time, and only allowing large asynchronous transfers (e.g. texture upload, scaling, format transforms) to take a certain amount of time. After that time has expired, it should wait on the next frame to be composited before resuming (assuming there is a composite scheduled). If the composited frame was delayed to the point that it skipped a frame compared to the last unladen composite, the amount of time dedicated to transfers should be reduced, or the transfer should be delayed until some arbitrary time (i.e. it should only be considered ok to skip a frame every X ms).<\/p>\n<p>It\u2019s interesting that you can see something very similar to this happening in early versions of iOS (I don\u2019t know if it still happens or not) \u2013 when scrolling long lists with images that load in dynamically, none of the images will load while the list is animating. The user response was paramount, to the point that it was considered more important to present consistent response than it was to present complete UI. This priority, I think, is a lot of the reason the iPhone feels \u2018magic\u2019 and Android phones felt like junk up until around 4.0 (where it\u2019s better, but still not as good as iOS).<\/p>\n<h2>4. Level-of-detail<\/h2>\n<p>This is something that I did get to partially implement while working on Firefox for Android, though I didn\u2019t do such a great job of it so its current implementation is heavily compromised from how I wanted it to work. This is another idea stolen from game development. There will be times, during certain interactions, where processing time will be necessarily limited. Quite often though, during these times, a user\u2019s view of the UI will be compromised in some fashion. It\u2019s important to understand that you don\u2019t always need to present the full-detail view of a UI. In Firefox for Android, this took the form that when scrolling fast enough that rendering couldn\u2019t keep up, we would render at half the resolution. This let us render more, and faster, giving the impression of a consistent UI even when the hardware wasn\u2019t quite capable of it. I notice Microsoft doing similar things since Windows 8; notice how the quality of image scaling reduces markedly while scrolling or animations are in progress. This idea is very implementation-specific. What can be dropped and what you want to drop will differ between platforms, form-factors, hardware, etc. Generally though, some things you can consider dropping: Sub-pixel anti-aliasing, high-quality image scaling, render resolution, colour-depth, animations. You may also want to consider showing partial UI if you know that it will very quickly be updated. The Android web-browser during the Honeycomb years did this, and I attempted (with limited success, because it\u2019s hard\u2026) to do this with Firefox for Android <a href=\"https:\/\/youtu.be\/2BUJb3LVH8Q\">many years ago<\/a>.<\/p>\n<h2>Pitfalls<\/h2>\n<p>I think it\u2019s easy to read ideas like this and think it boils down to \u201cdo everything asynchronously\u201d. Unfortunately, if you take a na\u00efve approach to that, you just end up with something that can be inexplicably slow sometimes and the only way to fix it is via profiling and micro-optimisations. It\u2019s very hard to guarantee a <em>consistent<\/em> experience if you don\u2019t manage when things happen. Yes, do everything asynchronously, but make sure you do your book-keeping and you manage when it\u2019s done. It\u2019s not only about splitting work up, it\u2019s about making sure it\u2019s done when it\u2019s smart to do so.<\/p>\n<p>You also need to be careful about how you measure these improvements, and to be aware that sometimes results in synthetic tests will even correlate to the opposite of the experience you want. A great example of this, in my opinion, is page-load speed on desktop browsers. All the major desktop browsers concentrate on prioritising the I\/O and computation required to get the page to 100%. For heavy desktop sites, however, this means the browser is often very clunky to use while pages are loading (yes, even with out-of-process tabs \u2013 see the point about bandwidth above). I highlight this specifically on desktop, because you\u2019re quite likely to not only be browsing much heavier sites that trigger this behaviour, but also to have multiple tabs open. So as soon as you load a couple of heavy sites, your entire browsing experience is compromised. I wouldn\u2019t mind the site taking a little longer to load if it didn\u2019t make the whole browser chug while doing so.<\/p>\n<p>Don\u2019t lose sight of your goals. Don\u2019t compromise. Things might take longer to complete, deadlines might be missed\u2026 But polish can\u2019t be overrated. Polish is what people feel and what they remember, and the lack of it can have a devastating effect on someone\u2019s perception. It\u2019s not always conscious or obvious either, even when you\u2019re the developer. Ask yourself \u201cAm I fully satisfied with this\u201d before marking something as complete. You might still be able to ship if the answer is \u201cNo\u201d, but make sure you don\u2019t lose sight of that and make sure it gets the priority it deserves.<\/p>\n<p>One last point I\u2019ll make; I think to really execute on all of this, it requires buy-in from everyone. Not just engineers, not just engineers and managers, but visual designers, user experience, leadership\u2026 Everyone. It\u2019s too easy to do a job that\u2019s good enough and it\u2019s too much responsibility to put it all on one person\u2019s shoulders. You really need to be on the ball to produce the kind of software that Apple does almost routinely, but as much as they\u2019d say otherwise, it isn\u2019t magic.<\/p>\n<div class=\"author-info\">\n<div class=\"author-avatar\">\n<img alt=\"\" src=\"https:\/\/1.gravatar.com\/avatar\/4809f758bdae11a117e078f3e4886519?s=42&amp;d=mm&amp;r=g\" srcset=\"http:\/\/1.gravatar.com\/avatar\/4809f758bdae11a117e078f3e4886519?s=84&amp;d=mm&amp;r=g 2x\" class=\"avatar avatar-42 photo\" height=\"42\" width=\"42\" \/><\/div>\n<div class=\"author-description\">\n<h2 class=\"author-title\">Author: Chris Lord<\/h2>\n<p class=\"author-bio\">\nComputer programmer and banjo player. \n<\/p>\n<\/div>\n<\/div>","created_at":"2018-11-07T16:50:39+0000","updated_at":"2018-11-07T16:50:39+0000","published_by":["View all posts by Chris Lord"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"chrislord.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":520,"title":"stolk\/GPGOAP","url":"https:\/\/github.com\/stolk\/GPGOAP\/blob\/master\/README.md","content":"<h2><a id=\"user-content-introduction\" class=\"anchor\" aria-hidden=\"true\" href=\"#introduction\"><\/a>Introduction<\/h2>\n<p>GOAP, or Goal Oriented Action Planning is a powerful tool to create game AI. For all the details I will refer to <a href=\"http:\/\/web.media.mit.edu\/~jorkin\/goap.html\" rel=\"nofollow\">Jeff Orkin's collection of articles<\/a>. But in short: GOAP will let computer controlled characters (NPCs) make action plans that can achieve desired goals. It will do so in a highly maintainable, easily extendible, highly modular fashion. Naive implementation of AI code will invariably blow up for any non trivial problem. GOAP on the other hand, is robust and is unlikely to buckle under large complexity. This software implements GOAP in the C programming language. It does so in a generic fashion, which makes it suitable for many projects.<\/p>\n<h2><a id=\"user-content-basics-of-goap\" class=\"anchor\" aria-hidden=\"true\" href=\"#basics-of-goap\"><\/a>Basics of GOAP<\/h2>\n<p>Creating a plan for AI controlled entities comes down to the following steps:<\/p>\n<ol><li>Describe the repertoire of actions available.<\/li>\n<li>Describe the current state of the world.<\/li>\n<li>Describe the desired state of the world (goal).<\/li>\n<\/ol><p>To describe the actions, we specify:<\/p>\n<ul><li>The preconditions for the action.<\/li>\n<li>The postconditions (effects) of the action.<\/li>\n<li>The cost of the action.<\/li>\n<\/ul><p>To describe world state, we define a set of world state atoms. Each atom has a tag, and a boolean value.<\/p>\n<p>The planner will then be able to formulate a plan of actions that takes the world to the desired state, provided such a path exists. The plan formulated is guaranteed to be the lowest cost plan.<\/p>\n<h2><a id=\"user-content-example-scenario\" class=\"anchor\" aria-hidden=\"true\" href=\"#example-scenario\"><\/a>Example Scenario<\/h2>\n<p>Let us consider a planner for an AI soldier. Our soldier can perform the following actions: scout, approach, aim, shoot, load, detonatebomb, flee.<\/p>\n<ul><li><strong>scout<\/strong> requires: <em>armedwithgun<\/em>. Effect: <em>enemyvisible<\/em>.<\/li>\n<li><strong>approach<\/strong> requires: <em>enemyvisible<\/em>. Effect: <em>nearenemy<\/em>.<\/li>\n<li><strong>aim<\/strong> requires: <em>enemyvisible<\/em> and <em>weaponloaded<\/em>. Effect: <em>enemylinedup<\/em>.<\/li>\n<li><strong>shoot<\/strong> requires: <em>enemylinedup<\/em>. Effect: <em>!enemyalive<\/em>.<\/li>\n<li><strong>load<\/strong> requires: <em>armedwithgun<\/em>. Effect: <em>weaponloaded<\/em>.<\/li>\n<li><strong>detonatebomb<\/strong> requires: <em>armedwithbomb<\/em> and <em>nearenemy<\/em>. Effect: <em>!alive<\/em> and <em>!enemyalive<\/em>.<\/li>\n<li><strong>flee<\/strong> requires: <em>enemyvisible<\/em>. Effect: <em>!nearenemy<\/em>.<\/li>\n<\/ul><p>Next, we will tell the planner that currently:\nEnemy is not visible, we are armed with gun, our weapon is not loaded, enemy is not lined up, enemy is alive, we are armed with bomb, we are not near enemy, we are alive.<\/p>\n<p>Then we tell our planner what our desired world looks like. We only care about one thing: our enemy is not alive.<\/p>\n<p>With this, the planner can formulate a plan of actions to make this happen. For the moment, let's assume all our actions have the default cost 1 associated with it. This is what the planner will return to us:<\/p>\n<pre>                       ARMEDWITHGUN,enemyvisible,nearenemy,weaponloaded,enemylinedup,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n0: scout               ARMEDWITHGUN,ENEMYVISIBLE,nearenemy,weaponloaded,enemylinedup,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n1: approach            ARMEDWITHGUN,ENEMYVISIBLE,NEARENEMY,weaponloaded,enemylinedup,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n2: detonatebomb        ARMEDWITHGUN,ENEMYVISIBLE,NEARENEMY,weaponloaded,enemylinedup,enemyalive,ARMEDWITHBOMB,alive,\n<\/pre>\n<p>Note: this notation uses lowercase if the condition is false, and uppercase if the condition is true.\nThe first line shows the current world state, and the last line the desired world state, with all intermediate states in between. The plan to execute is: (scout, approach, detonatebomb).<\/p>\n<p>If we follow this plan, we have the unfortunate side effect that not only our enemy dies, but we die as well. This is easily solved by making detonatingbomb a higher cost action. But a more elegant approach would be to change our desired world state, and tell the planner that not only do we want our enemy dead, we want to ourselves alive. The planner will now create:<\/p>\n<pre>                       ARMEDWITHGUN,enemyvisible,nearenemy,weaponloaded,enemylinedup,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n0: scout               ARMEDWITHGUN,ENEMYVISIBLE,nearenemy,weaponloaded,enemylinedup,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n1: load                ARMEDWITHGUN,ENEMYVISIBLE,nearenemy,WEAPONLOADED,enemylinedup,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n2: aim                 ARMEDWITHGUN,ENEMYVISIBLE,nearenemy,WEAPONLOADED,ENEMYLINEDUP,ENEMYALIVE,ARMEDWITHBOMB,ALIVE,\n3: shoot               ARMEDWITHGUN,ENEMYVISIBLE,nearenemy,WEAPONLOADED,ENEMYLINEDUP,enemyalive,ARMEDWITHBOMB,ALIVE,\n<\/pre>\n<h2><a id=\"user-content-example-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#example-code\"><\/a>Example Code<\/h2>\n<p>The entire scenario described above is implemented succinctly in these few lines of code:<\/p>\n<pre>#include \"goap.h\"\n#include \"astar.h\"\nactionplanner_t ap;\ngoap_actionplanner_clear( &amp;ap ); \/\/ initializes action planner\n\/\/ describe repertoire of actions\ngoap_set_pre( &amp;ap, \"scout\", \"armedwithgun\", true );\ngoap_set_pst( &amp;ap, \"scout\", \"enemyvisible\", true );\ngoap_set_pre( &amp;ap, \"approach\", \"enemyvisible\", true );\ngoap_set_pst( &amp;ap, \"approach\", \"nearenemy\", true );\ngoap_set_pre( &amp;ap, \"aim\", \"enemyvisible\", true );\ngoap_set_pre( &amp;ap, \"aim\", \"weaponloaded\", true );\ngoap_set_pst( &amp;ap, \"aim\", \"enemylinedup\", true );\ngoap_set_pre( &amp;ap, \"shoot\", \"enemylinedup\", true );\ngoap_set_pst( &amp;ap, \"shoot\", \"enemyalive\", false );\ngoap_set_pre( &amp;ap, \"load\", \"armedwithgun\", true );\ngoap_set_pst( &amp;ap, \"load\", \"weaponloaded\", true );\ngoap_set_pre( &amp;ap, \"detonatebomb\", \"armedwithbomb\", true );\ngoap_set_pre( &amp;ap, \"detonatebomb\", \"nearenemy\", true );\ngoap_set_pst( &amp;ap, \"detonatebomb\", \"alive\", false );\ngoap_set_pst( &amp;ap, \"detonatebomb\", \"enemyalive\", false );\ngoap_set_pre( &amp;ap, \"flee\", \"enemyvisible\", true );\ngoap_set_pst( &amp;ap, \"flee\", \"nearenemy\", false );\n\/\/ describe current world state.\nworldstate_t fr; \ngoap_worldstate_clear( &amp;fr );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"enemyvisible\", false );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"armedwithgun\", true );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"weaponloaded\", false );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"enemylinedup\", false );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"enemyalive\", true );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"armedwithbomb\", true );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"nearenemy\", false );\ngoap_worldstate_set( &amp;ap, &amp;fr, \"alive\", true );\n\/\/ describe desired world state.\nworldstate_t goal;\ngoap_worldstate_clear( &amp;goal );\ngoap_worldstate_set( &amp;ap, &amp;goal, \"enemyalive\", false );\n\/\/goap_worldstate_set( &amp;ap, &amp;goal, \"alive\", true ); \/\/ add this to avoid suicide actions in the plan.\nworldstate_t states[16];\nconst char* plan[16]; \/\/ The planner will return the action plan in this array.\nint plansz=16; \/\/ Size of our return buffers.\nconst int plancost = astar_plan( &amp;ap, fr, goal, plan, states, &amp;plansz );\n<\/pre>\n<p>And that is it. The plan's size will be returned in <em>plansz<\/em> which is not necessarily the same as the plan cost. You can access the <em>plan<\/em> array elements for the actual steps (actions). For each action in the plan, the resulting world state is also available in the <em>states<\/em> array.<\/p>\n<h2><a id=\"user-content-implementation-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#implementation-notes\"><\/a>Implementation Notes<\/h2>\n<p>To build, use a C99 compliant C compiler. You can invoke with: <code>$ gcc -std=c99 astar.c goap.c main.c<\/code><\/p>\n<p>The strength of GPGOAP is that the API is very generic: both world state and actions are solely described with C strings and booleans. No game specific enums end up in the API. With this power comes responsibility: do not make typos in the action names or atom names. They will end up representing different atoms if you do.<\/p>\n<p>For performance, all the tags for the world state atoms are converted to an entry in a bit field. This bit field is implemented as a 'long long int', which typically is 8 bytes or 64 bits. This means that you cannot use more than 64 atoms to describe the world and the actions to the planner. When using multiple NPCs in your game, I suggest using separate planners for them, so that if two NPCs have different action sets, you don't end up combining their atom name space and exceeding 64 tags.<\/p>\n<p>The only atom type supported is boolean. This means that scalars cannot be used to describe the world. I advice splitting up a scalar value with multiple booleans. E.g. fuelempy, fuellow, fuelfull to encode a single scalar fuel value.<\/p>\n<p>I strongly advice using lowercase names for the world state atoms. I have provided a debug function that will return a string representation of a world state that you can print out. In this representation, lowercase is used for false valued atoms, and uppercase is used for true valued atoms.<\/p>\n<h2><a id=\"user-content-bugs\" class=\"anchor\" aria-hidden=\"true\" href=\"#bugs\"><\/a>Bugs<\/h2>\n<p>No known bugs.<\/p>\n<h2><a id=\"user-content-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#files\"><\/a>Files<\/h2>\n<ul><li><strong>goap.h goap.c<\/strong> implements the planner.<\/li>\n<li><strong>astar.h astar.c<\/strong> implements A* search over the world state space.<\/li>\n<li><strong>main.c<\/strong> sample scenario.<\/li>\n<\/ul><p>goap and astar are codependent unfortunately. Keeping them in separate files makes sense though, as they address two distinct parts of the system.<\/p>\n<h2><a id=\"user-content-thanks\" class=\"anchor\" aria-hidden=\"true\" href=\"#thanks\"><\/a>Thanks<\/h2>\n<ul><li>Thank you Amit Patel for <a href=\"http:\/\/theory.stanford.edu\/~amitp\/GameProgramming\/ImplementationNotes.html\" rel=\"nofollow\">describing<\/a> A* search so succinctly.<\/li>\n<li>Thank you Jeff Orkin for the origins of GOAP.<\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright 2012 <a href=\"http:\/\/stolk.org\" rel=\"nofollow\">Abraham T. Stolk<\/a><\/p>\n<p>Licensed under the Apache License, Version 2.0 (the \"License\");\nyou may not use this file except in compliance with the License.\nYou may obtain a copy of the License at<\/p>\n<p><a href=\"http:\/\/www.apache.org\/licenses\/LICENSE-2.0\" rel=\"nofollow\">http:\/\/www.apache.org\/licenses\/LICENSE-2.0<\/a><\/p>\n<p>Unless required by applicable law or agreed to in writing, software\ndistributed under the License is distributed on an \"AS IS\" BASIS,\nWITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\nSee the License for the specific language governing permissions and\nlimitations under the License.<\/p>","created_at":"2018-11-07T16:50:39+0000","updated_at":"2018-11-07T16:50:39+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/455280?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":521,"title":"git ssb - noffle\/git-ssb-intro@master","url":"https:\/\/git.scuttlebot.io\/%RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc=.sha256","content":"<p>README.md<\/p><blockquote>\n<p>a guide to hacking <em>together<\/em> on the distributed web<\/p>\n<\/blockquote><p>If you're curious about the distributed web, and the prospect of using git\ncollaboratively without a central, closed-source point of origin, you came to\nthe right place!<\/p><p>git-ssb is a great fit for just this. This guide tries to act as a transition\naide for GitHub users into the world of git-ssb. It assumes rudimentary\nknowledge of GitHub. It will walk through:<\/p><ol><li>understanding what \"SSB\" is.<\/li>\n<li>installing scuttlebot -- the SSB peer server -- and joining the network.<\/li>\n<li>installing the <code>git-ssb<\/code> command and the <code>git-ssb-web<\/code> web interface.<\/li>\n<li>a walkthrough of using git-ssb to do common GitHub workflows (creating repos,\nmaking pull requests, merging pull requests, issues, etc).<\/li>\n<\/ol><p><img src=\"https:\/\/git.scuttlebot.io\/%25RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc%3D.sha256\/raw\/37c4aa2dece9133104d7da1691904daf8b9862d4\/git_ssb_activity.png\" alt=\"git-ssb web activity\" \/><\/p><p><em>A view of recent activity among my friends on git-ssb<\/em><\/p><h2 id=\"what-is-ssb-\">What is SSB?<\/h2><p>SSB stands for <a href=\"https:\/\/github.com\/ssbc\/secure-scuttlebutt\">secure scuttlebutt<\/a>,\nthe database\/protocol that powers the <em>peer-to-peer log store<\/em> <a href=\"http:\/\/scuttlebot.io\">scuttlebutt<\/a>. There is a great deal of information to be found by reading these sites, if you're so inclined.<\/p><p>If you're familiar with bitcoin or blockchains, SSB is akin to having your\nown personal blockchain: you can append messages of any kind, with any data\nyou'd like. Unlike bitcoin, there's no money involved: it's just a data\nstructure on your local machine.<\/p><p>Your personal log can only be appended to, and is cryptographically secure: each\nmessage references the hash of the message that came before it. The whole thing\nis then also signed by your public key, making it both tamper-proof, and\nappendable only by the private key holder (you).<\/p><p>The final piece of SSB that makes it all work is the gossip protocol: when you\nrun scuttlebot, you become a peer in a network of other scuttlebot users. You\ncan choose to 'follow' other people's personal logs. As a result, whenever you\nsee that user (or any user that follows that user too) online, you can retrieve\ntheir latest log entries from them. All without any central server or central\nauthentication.<\/p><p><a href=\"http:\/\/git.scuttlebot.io\/%25n92DiQh7ietE%2BR%2BX%2FI403LQoyf2DtR3WQfCkDKlheQU%3D.sha256\">git-ssb<\/a> builds on top of this: things like commits, branches, issues, and pull requests are encoded into log entries on each participant's personal log, while the gossip protocol runs in the background and propagates new content to everyone involved in the git repository.<\/p><h2 id=\"setup\">Setup<\/h2><p>If you're already up and running on Patchwork (or another client), you can skip to step 3.<\/p><h3 id=\"1-install-node-npm\">1. Install node + npm<\/h3><p>You can install node and npm..<\/p><ol><li>from <a href=\"https:\/\/nodejs.org\">https:\/\/nodejs.org<\/a>, or<\/li>\n<li>using <a href=\"https:\/\/github.com\/creationix\/nvm\">nvm<\/a><\/li>\n<\/ol><h3 id=\"2a-install-patchwork\">2a. Install Patchwork<\/h3><p><img src=\"https:\/\/git.scuttlebot.io\/%25RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc%3D.sha256\/raw\/37c4aa2dece9133104d7da1691904daf8b9862d4\/patchwork.jpg\" alt=\"Patchwork\" \/><\/p><p>Patchwork is an easy and user-friendly client that's nice to start your SSB\nexperience with, if you're new. Follow the instructions on the main page at <a href=\"https:\/\/scuttlebutt.nz\">https:\/\/scuttlebutt.nz<\/a><\/p><p>If you'd rather not use Patchwork, follow the next section.<\/p><h3 id=\"2b-install-scuttlebot\">2b. Install scuttlebot<\/h3><p>The <a href=\"http:\/\/www.github.com\/SSBC\/scuttlebot\">scuttlebot<\/a> (or sbot) is an ssb server which will manage replicating data with other peers.<\/p><p>Make sure you <a href=\"https:\/\/github.com\/ssbc\/scuttlebot\/wiki\/Pub-Servers\">get an invite code<\/a> from a pub to increase your ability to find other scuttlers.<\/p><h3 id=\"3-install-ssb-npm\">3. Install ssb-npm<\/h3><p>All of the dependencies for git-ssb <em>and git-ssb itself<\/em> live on the SSB\nnetwork. To access them, we'll need to install <a href=\"http:\/\/git.scuttlebot.io\/%25iqhz%2FsQCZCSp91JYAqfQPzHuDYrjw1geKPf1wJ1CvlA%3D.sha256\">ssb-npm<\/a>:<\/p><p>First, install the ssb-npm-registry plugin:<\/p><pre>sbot plugins.install ssb-npm-registry --from 'http:\/\/localhost:8989\/blobs\/get\/&amp;2afFvk14JEObC047kYmBLioDgMfHe2Eg5\/gndSjPQ1Q=.sha256'\n<\/pre><p>If this doesn't work (or you aren't following enough people) you can fetch the blob from a centralized source too:<\/p><pre>sbot plugins.install ssb-npm-registry --from 'http:\/\/viewer.scuttlebot.io\/&amp;2afFvk14JEObC047kYmBLioDgMfHe2Eg5\/gndSjPQ1Q=.sha256'\n<\/pre><p>then install the ssb-npm command:<\/p><pre>npm install --registry=http:\/\/localhost:8043\/ -g ssb-npm\n<\/pre><p>If you have problems with the above commands you may need to follow different people, so \nthat you can find all the dependencies. <a href=\"http:\/\/git.scuttlebot.io\/%40f%2F6sQ6d2CMxRUhLpspgGIulDxDCwYD7DzFzPNr7u5AU%3D.ed25519\">cel<\/a> is a good bet. <\/p><h3 id=\"4-install-git-ssb\">4. Install git-ssb<\/h3><p><a href=\"http:\/\/git.scuttlebot.io\/%25n92DiQh7ietE%2BR%2BX%2FI403LQoyf2DtR3WQfCkDKlheQU%3D.sha256\">git-ssb<\/a> is primarily the work of <a href=\"http:\/\/git.scuttlebot.io\/%40f%2F6sQ6d2CMxRUhLpspgGIulDxDCwYD7DzFzPNr7u5AU%3D.ed25519\">cel<\/a>, and comes in the form of a command-line program not unlike <code>git<\/code> or GitHub's <a href=\"https:\/\/github.com\/github\/hub\">hub<\/a>.<\/p><p>Let's use <code>npm<\/code> to install <code>git-ssb<\/code> globally so we can use it in the command-line:<\/p><pre>$ ssb-npm install --global git-ssb\n<\/pre><p>You can now run <code>git-ssb<\/code> (or <code>git ssb<\/code>) from the command line and get a feel for the sorts of things it can do.<\/p><p><strong>NOTE<\/strong> - You need to make sure you are running a scuttlebot server when executing git-ssb commands. You can run one standalone, but the simplest way is to just run Patchwork.<\/p><h2 id=\"the-git-ssb-web-interface\">The git-ssb web interface<\/h2><p>If you've already joined the SSB network successfully, you can take a look at\nwhat other hackers on SSB are working on! Run <code>git-ssb web<\/code> and point your web\nbrowser at the URL it outputs. (Remember you need to be running scuttlebot or Patchwork.)<\/p><p><img src=\"https:\/\/git.scuttlebot.io\/%25RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc%3D.sha256\/raw\/37c4aa2dece9133104d7da1691904daf8b9862d4\/git_ssb_repo.png\" alt=\"git-ssb web repo\" \/><\/p><p><em>A repo on git-ssb<\/em><\/p><p>You'll see that it's not dissimilar from GitHub: <a href=\"http:\/\/git.scuttlebot.io\/%25q5d5Du%2B9WkaSdjc8aJPZm%2BjMrqgo0tmfR%2BRcX5ZZ6H4%3D.sha256\">git-ssb-web<\/a><br \/>lets you see commit activity, view repositories, browse commit history, see\nissues and pull requests. All without a centralized authority.<\/p><p>There are at least a few public git-ssb web servers run by the community, for\nbrowsing git-ssb content without running your own scuttlebot:<\/p><h2 id=\"creating-your-first-ssb-git-remote\">Creating your first SSB git remote<\/h2><p>git-ssb makes scuttlebot act like a git remote. Navigate to an existing git\nrepository of yours and run<\/p><pre>$ git-ssb create ssb my-little-repo\n<\/pre><p>This will create a new git repo on SSB. The command will output something that\nlooks like<\/p><pre>Created repo: ssb:\/\/%DEvlJYD+zuudMyNWFBjiiNvZ8DbOyBYhCkE5EVtWSV0=.sha256\n(my-little-repo)\n<\/pre><p>The string <code>%DEvlJYD+zuudMyNWFBjiiNvZ8DbOyBYhCkE5EVtWSV0=.sha256<\/code> is a hash that\nuniquely identifies your git-ssb repository. To see this in action, you can\nnavigate to (for example) <code>~\/projects<\/code> and run<\/p><pre>$ git clone ssb:\/\/%DEvlJYD+zuudMyNWFBjiiNvZ8DbOyBYhCkE5EVtWSV0=.sha256 my-repo\n<\/pre><p>and you'll see that <code>~\/projects\/my-repo<\/code> now exists. git-ssb talked to your local\nscuttlebot server and pulled down the entire git repository.<\/p><p>In fact, you can also view your handiwork by running <code>git-ssb web<\/code> and opening\nyour browser: you'll see your brand new repo near the top of the activity list\nand can browse it like you might've on GitHub.<\/p><h2 id=\"collaboration-forks-and-pull-requests\">Collaboration: forks and pull requests<\/h2><p>If you're reading this article, then you've already found one SSB git\nrepository: this one! Let's create a fork, make some changes, and send a pull\nrequest:<\/p><h3 id=\"clone\">Clone<\/h3><pre>$ git clone ssb:\/\/%RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc=.sha256 git-ssb-intro\n$ cd git-ssb-intro\n<\/pre><h3 id=\"fork\">Fork<\/h3><p>Let's look at the help with <code>git-ssb fork --help<\/code>:<\/p><pre>Usage: git ssb fork [&lt;upstream&gt;] &lt;remote_name&gt;\n  Create a new git-ssb repo as a fork of another repo\n  and add it as a git remote\nArguments:\n  upstream      id, url, or git remote name of the repo to fork.\n                default: 'origin' or 'ssb'\n  remote_name   Name for the new remote\n<\/pre><p>since we just did a fresh git clone, the remote to the original repo is\n<code>origin<\/code>; you can type <code>git remote -v<\/code> and have a look. since <code>origin<\/code> (or\n<code>ssb<\/code>) are the default for forks, you can omit this option.<\/p><p><code>remote_name<\/code> is whatever you'd like to call the remote of your fork. I usually\ncall mine <code>noffle-ssb<\/code>, to indicate that it's a fork remote and that it's my\npersonal one:<\/p><pre>$ git-ssb fork noffle-ssb\nCreated repo: ssb:\/\/%DKWLuN05qtwjdwTG7zCG4aU12muBK2pkpYgaBbeVkTM=.sha256 (git-ssb-intro)\nAdded remote: noffle-ssb\n<\/pre><p>Great! Now we can make some changes and commit them. I forgot to add a license\nto this document, so we can add that, commit, and push:<\/p><pre>$ cat &gt;&gt; README.md\n# license\nISC\n^D\n$ git commit -am 'adds license'\n[master 3a7c5f7] adds license\n 1 file changed, 4 insertions(+)\n$ git push noffle-ssb\nCounting objects: 3, done.\nDelta compression using up to 2 threads.\nCompressing objects: 100% (2\/2), done.\nWriting objects: 100% (3\/3), 2.55 KiB | 0 bytes\/s, done.\nTotal 3 (delta 0), reused 0 (delta 0)\nDecoding objects: 100% (3\/3)\nPushing blobs...pwMeF\/zIvuwFc4\/T3VBwpHr+r6H+dHvWKFyIxCs1+08=.sha256\nPublished %i0Fcr3zRdf50Tin1UdEmYU38eKWthmkO3XgpfMYU1AU=.sha256ha256\nTo ssb:\/\/%DKWLuN05qtwjdwTG7zCG4aU12muBK2pkpYgaBbeVkTM=.sha256\n   6a701c2..3a7c5f7  master -&gt; master\n<\/pre><p>We can see our fork on the web interface, too! navigate to\n<code>http:\/\/localhost:7718<\/code> and you'll see your repo fork.<\/p><p>Let's finally create a pull request with our changes:<\/p><pre>$ git-ssb pull-request\n<\/pre><p>this will launch your text editor and let you type in a message. It's good\npractice to make the first line of your message the subject, and leave two\nnewlines before writing your message. something like<\/p><pre>test pull request\nplease merge me\n<\/pre><p>Sure enough, you can now navigate to the <code>from-github-to-git-ssb<\/code> repo on SSB\nand see your pull request:\n<a href=\"http:\/\/localhost:7718\/%25RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc%3D.sha256\/pulls\">http:\/\/localhost:7718\/%25RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc%3D.sha256\/pulls<\/a><\/p><p>If you were the owner of the git repo, you could merge this pull request from\nthe command line by pulling down the fork remote and merging:<\/p><pre>git fetch ssb:\/\/%RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc=.sha256 noffle-ssb:noffle-ssb\ngit merge noffle-ssb\ngit push ssb\n<\/pre><h2 id=\"alternative-collaboration-without-forks\">Alternative: collaboration without forks<\/h2><p>git-ssb's permissionless model has an interesting consequence: <em>anybody can push\nto anybody else's git repository<\/em>. Coming from a GitHub background this probably\nsounds like madness, but folks on SSB have been using this model with good\nsuccess. A pull request might look like this:<\/p><pre>$ git clone ssb:\/\/%RPKzL382v2fAia5HuDNHD5kkFdlP7bGvXQApSXqOBwc=.sha256 foobar\n$ git checkout -b @noffle\/master\n$ cd foobar\n$ vim index.js\n(edit edit edit)\n$ git commit -am 'wark'\n$ git push ssb\n$ git ssb pull-request\n<\/pre><p>The nomenclature is to treat the <code>ssb<\/code> remote as the \"true\" remote, and work off\nof a branch called <code>@your-username\/master<\/code> as your own master, emulating a\nforked repo. This seems to work well: the SSB network thrives off of being a\ngroup of kind, respectful folks who don't push to each other's <code>master<\/code> branch.\n:)<\/p><h2 id=\"push-conflicts\">Push conflicts<\/h2><p>Since git-ssb allows users to push to the same remote asynchronously, there is a\npossibility that they will conflict. This is different from a git merge\nconflict: in this case, there are two (or more) competing HEADs for e.g.\n<code>master<\/code>.<\/p><p>If this happens, git-ssb will consider the branch as being in a conflict state.\nIt will automatically add a new \"conflict branch\" for each conflicting HEAD. The\nname of each conflict branch will be prefixed by its SSB message ID. To resolve\nthe conflict, one should merge the conflict branches locally as appropriate, and\nthen push the new head and delete the conflict branches. e.g.:<\/p><pre>git checkout master\ngit fetch origin\ngit merge origin\/%msg1\/master [origin\/%msg2\/master...]\n# resolve merge conflict\n...\ngit push origin master -d %msg1\/master [-d %msg2\/master...]\n<\/pre><h2 id=\"go-forth-and-collaborate-\">Go forth and collaborate!<\/h2><p>Well done, and welcome to the scuttleverse and the world of distributed social\ncoding! share git-ssb with your friends, and consider managing an SSB remote of\nyour GitHub work, to encourage further growth of the distributed web.<\/p><p>Thanks for reading! please share feedback or questions as issues on this repo,\nor, if you find any errors in the text, send a pull request.<\/p><ul><li>@noffle: wrote this doc<\/li>\n<li>@cel: answered my many questions, and wrote most of the section on \"push\nconflicts\"<\/li>\n<li>@mixmix: several clarifications and improvements<\/li>\n<\/ul><p>CC0<\/p>","created_at":"2018-11-07T16:50:39+0000","updated_at":"2018-11-07T16:50:39+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"git.scuttlebot.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":518,"title":"Installation guide - ArchWiki","url":"https:\/\/wiki.archlinux.org\/index.php\/Installation_guide","content":"<p>This document is a guide for installing <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Arch_Linux\" title=\"Arch Linux\">Arch Linux<\/a> from the live system booted with the official installation image. Before installing, it would be advised to view the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/FAQ\" class=\"mw-redirect\" title=\"FAQ\">FAQ<\/a>. For conventions used in this document, see <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Help:Reading\" title=\"Help:Reading\">Help:Reading<\/a>. In particular, code examples may contain placeholders (formatted in <code><i>italics<\/i><\/code>) that must be replaced manually.\n<\/p><p>For more detailed instructions, see the respective <a href=\"https:\/\/wiki.archlinux.org\/index.php\/ArchWiki\" class=\"mw-redirect\" title=\"ArchWiki\">ArchWiki<\/a> articles or the various programs' <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Man_page\" title=\"Man page\">man pages<\/a>, both linked from this guide. For interactive help, the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/IRC_channel\" class=\"mw-redirect\" title=\"IRC channel\">IRC channel<\/a> and the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/bbs.archlinux.org\/\">forums<\/a> are also available.\n<\/p><p>Arch Linux should run on any <a href=\"https:\/\/en.wikipedia.org\/wiki\/X86-64\" class=\"extiw\" title=\"wikipedia:X86-64\">x86_64<\/a>-compatible machine with a minimum of 512 MB RAM. A basic installation with all packages from the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/groups\/x86_64\/base\/\">base<\/a> group should take less than 800 MB of disk space. As the installation process needs to retrieve packages from a remote repository, this guide assumes a working internet connection is available.\n<\/p><div id=\"toc\" class=\"toc\"><p><\/p><h2>Contents<\/h2><ul><li class=\"toclevel-1 tocsection-1\"><a href=\"#Pre-installation\">1 Pre-installation<\/a>\n<ul><li class=\"toclevel-2 tocsection-2\"><a href=\"#Set_the_keyboard_layout\">1.1 Set the keyboard layout<\/a><\/li>\n<li class=\"toclevel-2 tocsection-3\"><a href=\"#Verify_the_boot_mode\">1.2 Verify the boot mode<\/a><\/li>\n<li class=\"toclevel-2 tocsection-4\"><a href=\"#Connect_to_the_Internet\">1.3 Connect to the Internet<\/a><\/li>\n<li class=\"toclevel-2 tocsection-5\"><a href=\"#Update_the_system_clock\">1.4 Update the system clock<\/a><\/li>\n<li class=\"toclevel-2 tocsection-6\"><a href=\"#Partition_the_disks\">1.5 Partition the disks<\/a><\/li>\n<li class=\"toclevel-2 tocsection-7\"><a href=\"#Format_the_partitions\">1.6 Format the partitions<\/a><\/li>\n<li class=\"toclevel-2 tocsection-8\"><a href=\"#Mount_the_file_systems\">1.7 Mount the file systems<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-9\"><a href=\"#Installation\">2 Installation<\/a>\n<ul><li class=\"toclevel-2 tocsection-10\"><a href=\"#Select_the_mirrors\">2.1 Select the mirrors<\/a><\/li>\n<li class=\"toclevel-2 tocsection-11\"><a href=\"#Install_the_base_packages\">2.2 Install the base packages<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-12\"><a href=\"#Configure_the_system\">3 Configure the system<\/a>\n<ul><li class=\"toclevel-2 tocsection-13\"><a href=\"#Fstab\">3.1 Fstab<\/a><\/li>\n<li class=\"toclevel-2 tocsection-14\"><a href=\"#Chroot\">3.2 Chroot<\/a><\/li>\n<li class=\"toclevel-2 tocsection-15\"><a href=\"#Time_zone\">3.3 Time zone<\/a><\/li>\n<li class=\"toclevel-2 tocsection-16\"><a href=\"#Localization\">3.4 Localization<\/a><\/li>\n<li class=\"toclevel-2 tocsection-17\"><a href=\"#Network_configuration\">3.5 Network configuration<\/a><\/li>\n<li class=\"toclevel-2 tocsection-18\"><a href=\"#Initramfs\">3.6 Initramfs<\/a><\/li>\n<li class=\"toclevel-2 tocsection-19\"><a href=\"#Root_password\">3.7 Root password<\/a><\/li>\n<li class=\"toclevel-2 tocsection-20\"><a href=\"#Boot_loader\">3.8 Boot loader<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-21\"><a href=\"#Reboot\">4 Reboot<\/a><\/li>\n<li class=\"toclevel-1 tocsection-22\"><a href=\"#Post-installation\">5 Post-installation<\/a><\/li>\n<\/ul><\/div><h2>Pre-installation<\/h2><p>Download and boot the installation medium as explained in <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Getting_and_installing_Arch\" title=\"Getting and installing Arch\">Getting and installing Arch<\/a>. You will be logged in on the first <a href=\"https:\/\/en.wikipedia.org\/wiki\/Virtual_console\" class=\"extiw\" title=\"wikipedia:Virtual console\">virtual console<\/a> as the root user, and presented with a <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Zsh\" title=\"Zsh\">Zsh<\/a> shell prompt.\n<\/p><p>To switch to a different console\u2014for example, to view this guide with <a href=\"https:\/\/wiki.archlinux.org\/index.php\/ELinks\" title=\"ELinks\">ELinks<\/a> alongside the installation\u2014use the <code>Alt+<i>arrow<\/i><\/code> <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Keyboard_shortcuts\" title=\"Keyboard shortcuts\">shortcut<\/a>. To <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Textedit\" class=\"mw-redirect\" title=\"Textedit\">edit<\/a> configuration files, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Nano#Usage\" title=\"Nano\">nano<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/vi\" class=\"extiw\" title=\"wikipedia:vi\">vi<\/a> and <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Vim#Usage\" title=\"Vim\">vim<\/a> are available.\n<\/p><h3>Set the keyboard layout<\/h3><p>The default <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Console_keymap\" class=\"mw-redirect\" title=\"Console keymap\">console keymap<\/a> is <a href=\"https:\/\/en.wikipedia.org\/wiki\/File:KB_United_States-NoAltGr.svg\" class=\"extiw\" title=\"wikipedia:File:KB United States-NoAltGr.svg\">US<\/a>. Available layouts can be listed with:\n<\/p><pre># ls \/usr\/share\/kbd\/keymaps\/**\/*.map.gz\n<\/pre><p>To modify the layout, append a corresponding file name to <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/loadkeys.1\">loadkeys(1)<\/a>, omitting path and file extension. For example, to set a <a href=\"https:\/\/en.wikipedia.org\/wiki\/File:KB_Germany.svg\" class=\"extiw\" title=\"wikipedia:File:KB Germany.svg\">German<\/a> keyboard layout:\n<\/p><pre># loadkeys de-latin1\n<\/pre><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Console_fonts\" class=\"mw-redirect\" title=\"Console fonts\">Console fonts<\/a> are located in <code>\/usr\/share\/kbd\/consolefonts\/<\/code> and can likewise be set with <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/setfont.8\">setfont(8)<\/a>.\n<\/p><h3>Verify the boot mode<\/h3><p>If UEFI mode is enabled on an <a href=\"https:\/\/wiki.archlinux.org\/index.php\/UEFI\" class=\"mw-redirect\" title=\"UEFI\">UEFI<\/a> motherboard, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Archiso\" title=\"Archiso\">Archiso<\/a> will <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Boot\" class=\"mw-redirect\" title=\"Boot\">boot<\/a> Arch Linux accordingly via <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Systemd-boot\" title=\"Systemd-boot\">systemd-boot<\/a>. To verify this, list the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/UEFI#UEFI_variables\" class=\"mw-redirect\" title=\"UEFI\">efivars<\/a> directory:\n<\/p><pre># ls \/sys\/firmware\/efi\/efivars\n<\/pre><p>If the directory does not exist, the system may be booted in <a href=\"https:\/\/en.wikipedia.org\/wiki\/BIOS\" class=\"extiw\" title=\"wikipedia:BIOS\">BIOS<\/a> or CSM mode. Refer to your motherboard's manual for details.\n<\/p><h3>Connect to the Internet<\/h3><p>The installation image enables the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Dhcpcd\" title=\"Dhcpcd\">dhcpcd<\/a> daemon for <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/git.archlinux.org\/archiso.git\/tree\/configs\/releng\/airootfs\/etc\/udev\/rules.d\/81-dhcpcd.rules\">wired network devices<\/a> on boot. The connection may be verified with <a href=\"https:\/\/en.wikipedia.org\/wiki\/ping_(networking_utility)\" class=\"extiw\" title=\"wikipedia:ping (networking utility)\">ping<\/a>:\n<\/p><pre># ping archlinux.org\n<\/pre><p>If no connection is available, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Stop\" class=\"mw-redirect\" title=\"Stop\">stop<\/a> the <i>dhcpcd<\/i> service with <code>systemctl stop dhcpcd@<i>interface<\/i><\/code> where the <code><i>interface<\/i><\/code> name can be <a href=\"https:\/\/en.wikipedia.org\/wiki\/Command-line_completion\" class=\"extiw\" title=\"wikipedia:Command-line completion\">tab-completed<\/a>. Proceed to configure the network as described in <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Network_configuration\" title=\"Network configuration\">Network configuration<\/a>.\n<\/p><h3>Update the system clock<\/h3><p>Use <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/timedatectl.1\">timedatectl(1)<\/a> to ensure the system clock is accurate:\n<\/p><pre># timedatectl set-ntp true\n<\/pre><p>To check the service status, use <code>timedatectl status<\/code>.\n<\/p><h3>Partition the disks<\/h3><p>When recognized by the live system, disks are assigned to a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Device_file#Naming_conventions\" class=\"extiw\" title=\"wikipedia:Device file\">block device<\/a> such as <code>\/dev\/sda<\/code> or <code>\/dev\/nvme0n1<\/code>. To identify these devices, use <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Lsblk\" class=\"mw-redirect\" title=\"Lsblk\">lsblk<\/a> or <i>fdisk<\/i>.\n<\/p><pre># fdisk -l\n<\/pre><p>Results ending in <code>rom<\/code>, <code>loop<\/code> or <code>airoot<\/code> may be ignored.\n<\/p><p>The following <i>partitions<\/i> are <b>required<\/b> for a chosen device:\n<\/p><ul><li>One partition for the root directory <code>\/<\/code>.<\/li>\n<li>If <a href=\"https:\/\/wiki.archlinux.org\/index.php\/UEFI\" class=\"mw-redirect\" title=\"UEFI\">UEFI<\/a> is enabled, an <a href=\"https:\/\/wiki.archlinux.org\/index.php\/EFI_system_partition\" title=\"EFI system partition\">EFI system partition<\/a>.<\/li><\/ul><p><strong>Note:<\/strong> <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Swap\" title=\"Swap\">Swap<\/a> space can be set on a separate partition or a <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Swap_file\" class=\"mw-redirect\" title=\"Swap file\">swap file<\/a>.<\/p><p>To modify <i>partition tables<\/i>, use <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Fdisk\" title=\"Fdisk\">fdisk<\/a> or <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Parted\" title=\"Parted\">parted<\/a>. \n<\/p><pre># fdisk \/dev\/<i>sda<\/i>\n<\/pre><p>See <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Partitioning\" title=\"Partitioning\">Partitioning<\/a> for more information.\n<\/p><p><strong>Note:<\/strong> If you want to create any stacked block devices for <a href=\"https:\/\/wiki.archlinux.org\/index.php\/LVM\" title=\"LVM\">LVM<\/a>, <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Disk_encryption\" title=\"Disk encryption\">disk encryption<\/a> or <a href=\"https:\/\/wiki.archlinux.org\/index.php\/RAID\" title=\"RAID\">RAID<\/a>, do it now.<\/p><h3>Format the partitions<\/h3><p>Once the partitions have been created, each must be formatted with an appropriate <a href=\"https:\/\/wiki.archlinux.org\/index.php\/File_system\" class=\"mw-redirect\" title=\"File system\">file system<\/a>. For example, to format the root partition on <code>\/dev\/<i>sda1<\/i><\/code> with <code><i>ext4<\/i><\/code>, run:\n<\/p><pre># mkfs.<i>ext4<\/i> \/dev\/<i>sda1<\/i>\n<\/pre><p>If you created a partition for swap (for example <code>\/dev\/<i>sda3<\/i><\/code>), initialize it with <i>mkswap<\/i>:\n<\/p><pre># mkswap \/dev\/<i>sda3<\/i>\n# swapon \/dev\/<i>sda3<\/i>\n<\/pre><p>See <a href=\"https:\/\/wiki.archlinux.org\/index.php\/File_systems#Create_a_file_system\" title=\"File systems\">File systems#Create a file system<\/a> for details.\n<\/p><h3>Mount the file systems<\/h3><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Mount\" class=\"mw-redirect\" title=\"Mount\">Mount<\/a> the file system on the root partition to <code>\/mnt<\/code>, for example:\n<\/p><pre># mount \/dev\/<i>sda1<\/i> \/mnt\n<\/pre><p>Create mount points for any remaining partitions and mount them accordingly:\n<\/p><pre># mkdir \/mnt\/<i>boot<\/i>\n# mount \/dev\/<i>sda2<\/i> \/mnt\/<i>boot<\/i>\n<\/pre><p><a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/git.archlinux.org\/arch-install-scripts.git\/tree\/genfstab.in\">genfstab<\/a> will later detect mounted file systems and swap space.\n<\/p><h2>Installation<\/h2><h3>Select the mirrors<\/h3><p>Packages to be installed must be downloaded from <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Mirrors\" title=\"Mirrors\">mirror servers<\/a>, which are defined in <code>\/etc\/pacman.d\/mirrorlist<\/code>. On the live system, all mirrors are enabled, and sorted by their synchronization status and speed at the time the installation image was created.\n<\/p><p>The higher a mirror is placed in the list, the more priority it is given when downloading a package. You may want to edit the file accordingly, and move the geographically closest mirrors to the top of the list, although other criteria should be taken into account.\n<\/p><p>This file will later be copied to the new system by <i>pacstrap<\/i>, so it is worth getting right.\n<\/p><h3>Install the base packages<\/h3><p>Use the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/projects.archlinux.org\/arch-install-scripts.git\/tree\/pacstrap.in\">pacstrap<\/a> script to install the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/groups\/x86_64\/base\/\">base<\/a> package group:\n<\/p><pre># pacstrap \/mnt base\n<\/pre><p>This group does not include all tools from the live installation, such as <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=btrfs-progs\">btrfs-progs<\/a> or specific wireless firmware; see <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/projects.archlinux.org\/archiso.git\/tree\/configs\/releng\/packages.x86_64\">packages.x86_64<\/a> for comparison.\n<\/p><p>To <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Install\" class=\"mw-redirect\" title=\"Install\">install<\/a> packages and other groups such as <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/groups\/x86_64\/base-devel\/\">base-devel<\/a>, append the names to <i>pacstrap<\/i> (space separated) or to individual <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Pacman\" title=\"Pacman\">pacman<\/a> commands after the <a href=\"#Chroot\">#Chroot<\/a> step.\n<\/p><h2>Configure the system<\/h2><h3>Fstab<\/h3><p>Generate an <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Fstab\" title=\"Fstab\">fstab<\/a> file (use <code>-U<\/code> or <code>-L<\/code> to define by <a href=\"https:\/\/wiki.archlinux.org\/index.php\/UUID\" class=\"mw-redirect\" title=\"UUID\">UUID<\/a> or labels, respectively):\n<\/p><pre># genfstab -U \/mnt &gt;&gt; \/mnt\/etc\/fstab\n<\/pre><p>Check the resulting file in <code>\/mnt\/etc\/fstab<\/code> afterwards, and edit it in case of errors.\n<\/p><h3>Chroot<\/h3><p><a href=\"https:\/\/wiki.archlinux.org\/index.php\/Change_root\" class=\"mw-redirect\" title=\"Change root\">Change root<\/a> into the new system:\n<\/p><pre># arch-chroot \/mnt\n<\/pre><h3>Time zone<\/h3><p>Set the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Time_zone\" class=\"mw-redirect\" title=\"Time zone\">time zone<\/a>:\n<\/p><pre># ln -sf \/usr\/share\/zoneinfo\/<i>Region<\/i>\/<i>City<\/i> \/etc\/localtime\n<\/pre><p>Run <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/hwclock.8\">hwclock(8)<\/a> to generate <code>\/etc\/adjtime<\/code>:\n<\/p><pre># hwclock --systohc\n<\/pre><p>This command assumes the hardware clock is set to <a href=\"https:\/\/en.wikipedia.org\/wiki\/UTC\" class=\"extiw\" title=\"wikipedia:UTC\">UTC<\/a>. See <a href=\"https:\/\/wiki.archlinux.org\/index.php\/System_time#Time_standard\" title=\"System time\">System time#Time standard<\/a> for details.\n<\/p><h3>Localization<\/h3><p>Uncomment <code>en_US.UTF-8 UTF-8<\/code> and other needed <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Locale\" title=\"Locale\">locales<\/a> in <code>\/etc\/locale.gen<\/code>, and generate them with:\n<\/p><pre># locale-gen\n<\/pre><p>Set the <code>LANG<\/code> <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Variable\" class=\"mw-redirect\" title=\"Variable\">variable<\/a> in <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/locale.conf.5\">locale.conf(5)<\/a> accordingly, for example:\n<\/p><pre>\/etc\/locale.conf<\/pre><pre>LANG=<i>en_US.UTF-8<\/i><\/pre><p>If you <a href=\"#Set_the_keyboard_layout\">set the keyboard layout<\/a>, make the changes persistent in <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/vconsole.conf.5\">vconsole.conf(5)<\/a>:\n<\/p><pre>\/etc\/vconsole.conf<\/pre><pre>KEYMAP=<i>de-latin1<\/i><\/pre><h3>Network configuration<\/h3><p>Create the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Hostname\" class=\"mw-redirect\" title=\"Hostname\">hostname<\/a> file:\n<\/p><pre>\/etc\/hostname<\/pre><pre><i>myhostname<\/i>\n<\/pre><p>Add matching entries to <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/hosts.5\">hosts(5)<\/a>:\n<\/p><pre>\/etc\/hosts<\/pre><pre>127.0.0.1\tlocalhost\n::1\t\tlocalhost\n127.0.1.1\t<i>myhostname<\/i>.localdomain\t<i>myhostname<\/i>\n<\/pre><p>If the system has a permanent IP address, it should be used instead of <code>127.0.1.1<\/code>.\n<\/p><p>Complete the <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Network_configuration\" title=\"Network configuration\">network configuration<\/a> for the newly installed environment.\n<\/p><h3>Initramfs<\/h3><p>Creating a new <i>initramfs<\/i> is usually not required, because <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Mkinitcpio\" title=\"Mkinitcpio\">mkinitcpio<\/a> was run on installation of the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/www.archlinux.org\/packages\/?name=linux\">linux<\/a> package with <i>pacstrap<\/i>.\n<\/p><p>For special configurations, modify the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/mkinitcpio.conf.5\">mkinitcpio.conf(5)<\/a> file and recreate the initramfs image:\n<\/p><pre># mkinitcpio -p linux\n<\/pre><h3>Root password<\/h3><p>Set the root <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Password\" class=\"mw-redirect\" title=\"password\">password<\/a>:\n<\/p><pre># passwd\n<\/pre><h3>Boot loader<\/h3><p>A Linux-capable boot loader must be installed in order to boot Arch Linux. See <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Category:Boot_loaders\" title=\"Category:Boot loaders\">Category:Boot loaders<\/a> for available choices.\n<\/p><p>If you have an Intel or AMD CPU, enable <a href=\"https:\/\/wiki.archlinux.org\/index.php\/Microcode\" title=\"Microcode\">microcode<\/a> updates.\n<\/p><h2>Reboot<\/h2><p>Exit the chroot environment by typing <code>exit<\/code> or pressing <code>Ctrl+D<\/code>.\n<\/p><p>Optionally manually unmount all the partitions with <code>umount -R \/mnt<\/code>: this allows noticing any \"busy\" partitions, and finding the cause with <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/jlk.fjfi.cvut.cz\/arch\/manpages\/man\/fuser.1\">fuser(1)<\/a>.\n<\/p><p>Finally, restart the machine by typing <code>reboot<\/code>: any partitions still mounted will be automatically unmounted by <i>systemd<\/i>. Remember to remove the installation media and then login into the new system with the root account.\n<\/p><h2>Post-installation<\/h2><p>See <a href=\"https:\/\/wiki.archlinux.org\/index.php\/General_recommendations\" title=\"General recommendations\">General recommendations<\/a> for system management directions and post-installation tutorials (like setting up a graphical user interface, sound or a touchpad).\n<\/p><p>For a list of applications that may be of interest, see <a href=\"https:\/\/wiki.archlinux.org\/index.php\/List_of_applications\" title=\"List of applications\">List of applications<\/a>.\n<\/p>","created_at":"2018-11-07T16:50:38+0000","updated_at":"2018-11-07T16:50:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"wiki.archlinux.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":517,"title":"Ask HN: Which non-technology book has influenced you the most and why?","url":"https:\/\/news.ycombinator.com\/item?id=13967350","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D13967350\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"13967350\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=13967350\" class=\"storylink\">Ask HN: Which non-technology book has influenced you the most and why?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        375 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=fernandohur\" class=\"hnuser\">fernandohur<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967350\">on Mar 27, 2017<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=13967350&amp;goto=item%3Fid%3D13967350\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Which%20non-technology%20book%20has%20influenced%20you%20the%20most%20and%20why%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Which%20non-technology%20book%20has%20influenced%20you%20the%20most%20and%20why%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=13967350&amp;auth=1fe9603ce617028411cbb634765e542e991d6a2f\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=13967350\">332\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Rules:\n1. Only one book.\n2. Can't be a technical book e.g. 'Programming C' doesn't count.<\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"13967829\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=epaga\" class=\"hnuser\">epaga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967829\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968653\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cjf4\" class=\"hnuser\">cjf4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968653\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970302\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=williamle8300\" class=\"hnuser\">williamle8300<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970302\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970225\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kej\" class=\"hnuser\">kej<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970225\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13973331\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lanstin\" class=\"hnuser\">lanstin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13973331\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975707\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975707\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969308\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pklausler\" class=\"hnuser\">pklausler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969308\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971724\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonathansizz\" class=\"hnuser\">jonathansizz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971724\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13976969\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pklausler\" class=\"hnuser\">pklausler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13976969\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13979600\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonathansizz\" class=\"hnuser\">jonathansizz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13979600\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13989935\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13989935\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974276\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nurettin\" class=\"hnuser\">nurettin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974276\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13976140\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonathansizz\" class=\"hnuser\">jonathansizz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13976140\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13989951\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13989951\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"14062002\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nurettin\" class=\"hnuser\">nurettin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=14062002\">on Apr 7, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975711\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975711\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967883\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=randcraw\" class=\"hnuser\">randcraw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967883\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970311\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=XaspR8d\" class=\"hnuser\">XaspR8d<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970311\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970976\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmac0415\" class=\"hnuser\">dmac0415<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970976\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970256\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=genkimind\" class=\"hnuser\">genkimind<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970256\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970911\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=virtualwhys\" class=\"hnuser\">virtualwhys<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970911\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972283\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=willbw\" class=\"hnuser\">willbw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972283\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967923\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0x445442\" class=\"hnuser\">0x445442<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967923\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971198\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=xherberta\" class=\"hnuser\">xherberta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971198\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972089\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0x445442\" class=\"hnuser\">0x445442<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972089\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"14019152\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vram22\" class=\"hnuser\">vram22<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=14019152\">on Apr 2, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969213\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tylerc230\" class=\"hnuser\">tylerc230<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969213\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13989980\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13989980\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967894\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kemiller2002\" class=\"hnuser\">kemiller2002<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967894\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967822\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thedevil\" class=\"hnuser\">thedevil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967822\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968898\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=leobueno\" class=\"hnuser\">leobueno<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968898\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"14019161\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vram22\" class=\"hnuser\">vram22<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=14019161\">on Apr 2, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971183\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ashark\" class=\"hnuser\">ashark<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971183\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968415\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BeetleB\" class=\"hnuser\">BeetleB<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968415\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967982\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kabdib\" class=\"hnuser\">kabdib<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967982\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968126\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nekorosu\" class=\"hnuser\">Nekorosu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968126\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968405\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mtreis86\" class=\"hnuser\">mtreis86<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968405\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970291\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=notsrg\" class=\"hnuser\">notsrg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970291\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971022\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dionidium\" class=\"hnuser\">dionidium<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971022\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13986329\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ajmurmann\" class=\"hnuser\">ajmurmann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13986329\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970108\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=olavk\" class=\"hnuser\">olavk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970108\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970170\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kazinator\" class=\"hnuser\">kazinator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970170\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975726\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975726\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971140\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fripplebubby\" class=\"hnuser\">Fripplebubby<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971140\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969553\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neduma\" class=\"hnuser\">neduma<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969553\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970328\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielbarla\" class=\"hnuser\">danielbarla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970328\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970324\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=notsrg\" class=\"hnuser\">notsrg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970324\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969479\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icey\" class=\"hnuser\">icey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969479\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971228\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=medius\" class=\"hnuser\">medius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971228\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13985732\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yulaow\" class=\"hnuser\">yulaow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13985732\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971614\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bcbrown\" class=\"hnuser\">bcbrown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971614\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969962\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=snikeris\" class=\"hnuser\">snikeris<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969962\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970017\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=s_kilk\" class=\"hnuser\">s_kilk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970017\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970487\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=colebowl\" class=\"hnuser\">colebowl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970487\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13973210\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=meagher\" class=\"hnuser\">meagher<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13973210\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967847\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hollander\" class=\"hnuser\">hollander<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967847\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970092\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cJ0th\" class=\"hnuser\">cJ0th<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970092\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970164\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonmb\" class=\"hnuser\">jonmb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970164\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13979512\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hollander\" class=\"hnuser\">hollander<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13979512\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967671\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=1ba9115454\" class=\"hnuser\">1ba9115454<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967671\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967913\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mywittyname\" class=\"hnuser\">mywittyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967913\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968714\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swah\" class=\"hnuser\">swah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968714\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969251\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mywittyname\" class=\"hnuser\">mywittyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969251\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967990\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andreasklinger\" class=\"hnuser\">andreasklinger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967990\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968055\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walterbell\" class=\"hnuser\">walterbell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968055\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975589\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jgalt212\" class=\"hnuser\">jgalt212<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975589\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972925\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anotherevan\" class=\"hnuser\">anotherevan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972925\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13992941\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13992941\">on Mar 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967683\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tyurok\" class=\"hnuser\">tyurok<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967683\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968469\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skdotdan\" class=\"hnuser\">skdotdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968469\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968889\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=drallison\" class=\"hnuser\">drallison<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968889\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972716\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thomasthomas\" class=\"hnuser\">thomasthomas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972716\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975432\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skdotdan\" class=\"hnuser\">skdotdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975432\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975747\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975747\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"14037026\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RachelF\" class=\"hnuser\">RachelF<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=14037026\">on Apr 4, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969397\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=canes123456\" class=\"hnuser\">canes123456<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969397\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975435\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skdotdan\" class=\"hnuser\">skdotdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975435\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969327\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simonsarris\" class=\"hnuser\">simonsarris<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969327\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975438\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skdotdan\" class=\"hnuser\">skdotdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975438\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975302\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mohanmcgeek\" class=\"hnuser\">mohanmcgeek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975302\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975428\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skdotdan\" class=\"hnuser\">skdotdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975428\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970917\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fernandohur\" class=\"hnuser\">fernandohur<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970917\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967892\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adgasf\" class=\"hnuser\">adgasf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967892\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967987\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0x445442\" class=\"hnuser\">0x445442<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967987\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975632\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=continuational\" class=\"hnuser\">continuational<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975632\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970615\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Retric\" class=\"hnuser\">Retric<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970615\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971063\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=run4yourlives2\" class=\"hnuser\">run4yourlives2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971063\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971435\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Retric\" class=\"hnuser\">Retric<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971435\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968192\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jefurii\" class=\"hnuser\">jefurii<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968192\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967989\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mtreis86\" class=\"hnuser\">mtreis86<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967989\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969045\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rbg246\" class=\"hnuser\">rbg246<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969045\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971261\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=found_reading\" class=\"hnuser\">found_reading<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971261\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13979353\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aml183\" class=\"hnuser\">aml183<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13979353\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968207\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hihat\" class=\"hnuser\">hihat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968207\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969361\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=inerte\" class=\"hnuser\">inerte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969361\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971997\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dakrootie\" class=\"hnuser\">dakrootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971997\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968470\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simonbarker87\" class=\"hnuser\">simonbarker87<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968470\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970109\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rexpop\" class=\"hnuser\">rexpop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970109\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967914\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kemiller2002\" class=\"hnuser\">kemiller2002<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967914\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975647\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bythckr\" class=\"hnuser\">bythckr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975647\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968122\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Spooky23\" class=\"hnuser\">Spooky23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968122\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968740\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nav_Panel\" class=\"hnuser\">Nav_Panel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968740\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974754\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=achompas\" class=\"hnuser\">achompas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974754\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13976388\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nav_Panel\" class=\"hnuser\">Nav_Panel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13976388\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968864\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cponeill\" class=\"hnuser\">cponeill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968864\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968194\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=VLM\" class=\"hnuser\">VLM<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968194\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13973902\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=johngossman\" class=\"hnuser\">johngossman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13973902\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967745\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jMyles\" class=\"hnuser\">jMyles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967745\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968123\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noir_lord\" class=\"hnuser\">noir_lord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968123\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968701\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GedByrne\" class=\"hnuser\">GedByrne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968701\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970567\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rcarmo\" class=\"hnuser\">rcarmo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970567\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970890\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GedByrne\" class=\"hnuser\">GedByrne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970890\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970822\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eli_gottlieb\" class=\"hnuser\">eli_gottlieb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970822\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967772\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tarrosion\" class=\"hnuser\">Tarrosion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967772\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970404\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=d215\" class=\"hnuser\">d215<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970404\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"14016130\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ArielAssaf\" class=\"hnuser\">ArielAssaf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=14016130\">on Apr 2, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975225\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tonyedgecombe\" class=\"hnuser\">tonyedgecombe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975225\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968570\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joatmon-snoo\" class=\"hnuser\">joatmon-snoo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968570\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968273\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thewhitetulip\" class=\"hnuser\">thewhitetulip<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968273\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969463\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aa5a\" class=\"hnuser\">aa5a<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969463\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13973863\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thewhitetulip\" class=\"hnuser\">thewhitetulip<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13973863\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969454\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=canes123456\" class=\"hnuser\">canes123456<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969454\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13973850\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thewhitetulip\" class=\"hnuser\">thewhitetulip<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13973850\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967992\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sharkweek\" class=\"hnuser\">sharkweek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967992\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971232\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fripplebubby\" class=\"hnuser\">Fripplebubby<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971232\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972122\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nappy\" class=\"hnuser\">nappy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972122\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969719\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericskiff\" class=\"hnuser\">ericskiff<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969719\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967690\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LyndsySimon\" class=\"hnuser\">LyndsySimon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967690\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967749\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=inputcoffee\" class=\"hnuser\">inputcoffee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967749\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13978744\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LyndsySimon\" class=\"hnuser\">LyndsySimon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13978744\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13991476\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13991476\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967896\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rahimnathwani\" class=\"hnuser\">rahimnathwani<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967896\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968965\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Findeton\" class=\"hnuser\">Findeton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968965\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969096\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Balgair\" class=\"hnuser\">Balgair<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969096\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970348\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=btschaegg\" class=\"hnuser\">btschaegg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970348\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970706\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Balgair\" class=\"hnuser\">Balgair<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970706\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13978721\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LyndsySimon\" class=\"hnuser\">LyndsySimon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13978721\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13981945\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=libx\" class=\"hnuser\">libx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13981945\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13984223\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Findeton\" class=\"hnuser\">Findeton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13984223\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968025\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=carsongross\" class=\"hnuser\">carsongross<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968025\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967741\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mcone\" class=\"hnuser\">mcone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967741\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972028\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dano\" class=\"hnuser\">dano<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972028\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974640\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ud0\" class=\"hnuser\">ud0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974640\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970598\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rcarmo\" class=\"hnuser\">rcarmo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970598\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970962\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Eyght\" class=\"hnuser\">Eyght<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970962\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969644\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=f1gm3nt\" class=\"hnuser\">f1gm3nt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969644\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13978068\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tonyedgecombe\" class=\"hnuser\">tonyedgecombe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13978068\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972719\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=johngalt\" class=\"hnuser\">johngalt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972719\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13978811\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LyndsySimon\" class=\"hnuser\">LyndsySimon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13978811\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970048\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rexpop\" class=\"hnuser\">rexpop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970048\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972430\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=f1gm3nt\" class=\"hnuser\">f1gm3nt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972430\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13978768\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LyndsySimon\" class=\"hnuser\">LyndsySimon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13978768\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969564\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aakash58\" class=\"hnuser\">aakash58<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969564\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968957\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cponeill\" class=\"hnuser\">cponeill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968957\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974192\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mvindahl\" class=\"hnuser\">mvindahl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974192\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974499\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lazy_gator\" class=\"hnuser\">lazy_gator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974499\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968904\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=arcanus\" class=\"hnuser\">arcanus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968904\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967942\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bshimmin\" class=\"hnuser\">bshimmin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967942\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970535\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=king_panic\" class=\"hnuser\">king_panic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970535\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970825\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bshimmin\" class=\"hnuser\">bshimmin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970825\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13980427\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=king_panic\" class=\"hnuser\">king_panic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13980427\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968776\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=malloryerik\" class=\"hnuser\">malloryerik<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968776\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969709\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalChris\" class=\"hnuser\">CalChris<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969709\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970668\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SwellJoe\" class=\"hnuser\">SwellJoe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970668\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971405\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=scandinavegan\" class=\"hnuser\">scandinavegan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971405\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971848\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SwellJoe\" class=\"hnuser\">SwellJoe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971848\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974856\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ExactoKnight\" class=\"hnuser\">ExactoKnight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974856\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13991572\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13991572\">on Mar 29, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968165\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=emidln\" class=\"hnuser\">emidln<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968165\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968462\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=convexfunction\" class=\"hnuser\">convexfunction<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968462\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969895\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stblack\" class=\"hnuser\">stblack<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969895\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971029\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=germinalphrase\" class=\"hnuser\">germinalphrase<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971029\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968158\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jyriand\" class=\"hnuser\">jyriand<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968158\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971161\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971161\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968369\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=greenhatman\" class=\"hnuser\">greenhatman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968369\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968312\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vinayan3\" class=\"hnuser\">vinayan3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968312\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967922\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=arethuza\" class=\"hnuser\">arethuza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967922\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970288\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kwhitefoot\" class=\"hnuser\">kwhitefoot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970288\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967817\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonstewart\" class=\"hnuser\">jonstewart<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967817\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968468\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=falsedan\" class=\"hnuser\">falsedan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968468\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970344\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=VA3FXP\" class=\"hnuser\">VA3FXP<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970344\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970195\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pokler\" class=\"hnuser\">pokler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970195\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971072\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=socialist_coder\" class=\"hnuser\">socialist_coder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971072\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971758\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neves\" class=\"hnuser\">neves<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971758\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972686\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pokler\" class=\"hnuser\">pokler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972686\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967677\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=macca321\" class=\"hnuser\">macca321<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967677\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967810\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=arethuza\" class=\"hnuser\">arethuza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967810\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974115\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mvindahl\" class=\"hnuser\">mvindahl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974115\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968520\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jackfoxy\" class=\"hnuser\">jackfoxy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968520\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969900\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=panzagl\" class=\"hnuser\">panzagl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969900\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968554\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=johngossman\" class=\"hnuser\">johngossman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968554\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968622\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AngeloAnolin\" class=\"hnuser\">AngeloAnolin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968622\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969138\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MithrilTuxedo\" class=\"hnuser\">MithrilTuxedo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969138\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972091\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikpukinskis\" class=\"hnuser\">erikpukinskis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972091\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968465\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brogrammernot\" class=\"hnuser\">brogrammernot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968465\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968872\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hornbaker\" class=\"hnuser\">hornbaker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968872\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968992\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=costcopizza\" class=\"hnuser\">costcopizza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968992\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13977868\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=selimthegrim\" class=\"hnuser\">selimthegrim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13977868\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967910\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thomasreggi\" class=\"hnuser\">thomasreggi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967910\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968455\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BeetleB\" class=\"hnuser\">BeetleB<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968455\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969166\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brudgers\" class=\"hnuser\">brudgers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969166\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967679\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kevlar1818\" class=\"hnuser\">kevlar1818<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967679\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970043\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=s_kilk\" class=\"hnuser\">s_kilk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970043\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969291\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mattnewton\" class=\"hnuser\">mattnewton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969291\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970274\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=VA3FXP\" class=\"hnuser\">VA3FXP<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970274\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969733\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cickpass_broken\" class=\"hnuser\">cickpass_broken<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969733\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970850\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=calebm\" class=\"hnuser\">calebm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970850\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970383\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=NumberCruncher\" class=\"hnuser\">NumberCruncher<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970383\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968471\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jshawl\" class=\"hnuser\">jshawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968471\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970318\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kwhitefoot\" class=\"hnuser\">kwhitefoot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970318\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13972391\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jononor\" class=\"hnuser\">jononor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13972391\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970929\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eli_gottlieb\" class=\"hnuser\">eli_gottlieb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970929\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974912\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wowsig\" class=\"hnuser\">wowsig<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974912\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968020\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=test6554\" class=\"hnuser\">test6554<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968020\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968211\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tcopeland\" class=\"hnuser\">tcopeland<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968211\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968529\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanumanthan\" class=\"hnuser\">hanumanthan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968529\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969393\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erehweb\" class=\"hnuser\">erehweb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969393\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970233\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pagutierrezn\" class=\"hnuser\">pagutierrezn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970233\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969870\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mindcrime\" class=\"hnuser\">mindcrime<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969870\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967895\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sandworm101\" class=\"hnuser\">sandworm101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967895\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968132\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=magic_beans\" class=\"hnuser\">magic_beans<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968132\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970729\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=socialist_coder\" class=\"hnuser\">socialist_coder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970729\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971691\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=UncleSlacky\" class=\"hnuser\">UncleSlacky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971691\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970732\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j0e1\" class=\"hnuser\">j0e1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970732\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967927\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sixstringtheory\" class=\"hnuser\">sixstringtheory<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967927\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968353\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wazoox\" class=\"hnuser\">wazoox<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968353\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975820\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975820\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968164\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rajesht\" class=\"hnuser\">rajesht<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968164\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13975883\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osullivj\" class=\"hnuser\">osullivj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13975883\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971531\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=xherberta\" class=\"hnuser\">xherberta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971531\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13971644\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=scandinavegan\" class=\"hnuser\">scandinavegan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13971644\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970733\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=queirozfcom\" class=\"hnuser\">queirozfcom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970733\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970498\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=netnichols\" class=\"hnuser\">netnichols<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970498\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970081\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cowpig\" class=\"hnuser\">cowpig<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970081\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969502\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BatFastard\" class=\"hnuser\">BatFastard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969502\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969091\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crazybigdan\" class=\"hnuser\">crazybigdan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969091\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970763\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pge\" class=\"hnuser\">pge<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970763\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970640\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=btschaegg\" class=\"hnuser\">btschaegg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970640\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13967962\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cgrubb\" class=\"hnuser\">cgrubb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13967962\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970711\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ssijak\" class=\"hnuser\">ssijak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970711\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969071\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gentleteblor\" class=\"hnuser\">gentleteblor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969071\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969828\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thunderbong\" class=\"hnuser\">thunderbong<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969828\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970421\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rchen8\" class=\"hnuser\">rchen8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970421\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974274\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mikhail_Edoshin\" class=\"hnuser\">Mikhail_Edoshin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974274\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969623\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=peacetreefrog\" class=\"hnuser\">peacetreefrog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969623\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13992966\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nomentatus\" class=\"hnuser\">Nomentatus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13992966\">on Mar 30, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13970643\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=djkrudy\" class=\"hnuser\">djkrudy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13970643\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13968239\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=notacoward\" class=\"hnuser\">notacoward<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13968239\">on Mar 27, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13969078\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crawfordcomeaux\" class=\"hnuser\">crawfordcomeaux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13969078\">on Mar 27, 2017<\/a><\/p><br \/><div class=\"comment\">\"Nonviolent Communication\" (NVC) by Marshall Rosenberg<p>I was emotionally disconnected from people and myself from ~10~33 years old via information addiction (compulsively doing everything up to the \"doing\" parts of creation).<\/p><p>My disconnect kept me from learning how to communicate my empathy to others, as well as kept me from fully considering\/respecting their needs. Around a year ago, I broke through denial over sexual trauma I experienced as a child and was able to largely come to peace with it in 12 hours thanks to the principles and practices of NVC. That was huge and only the start of the book's impact on me.<\/p><p>Essentially, this book has been teaching me how to human for the past year. I've read it 5 times and may read it again this week. I'm willing to send a used copy from Amazon to anyone who can't afford a copy and thinks they could benefit from it. My contact info is in my profile for those who wish to take me up on the offer. Here are more ways the book has changed my life.<\/p><p>- It taught me how to emphatically connect. The communication framework it teaches is simple, yet highly structured. It was perfect for my systematic way of thinking.<\/p><p>- I gained a new understanding of fundamental human needs and how they relate to emotions we experience.<\/p><p>- I began practicing it with a friend, who I fell in love with through our mutual ability to connect and was briefly engaged to them last year.<\/p><p>-I've introduced the book to my family and the entire dynamic has changed dramatically. Passive aggressive and\/or codependent behaviors are slowly starting to shift.<\/p><p>- The book inspired me to practice abandoning my judgments. All of them. After a few weeks of this, I realized I was enjoying foods I used to hate. And then I started enjoying music I used to hate. And then children became adorable instead of obnoxious.<\/p><p>- I've always known something was wrong with the US justice system, but had no idea how to fix it. I learned about restorative justice through the book and my entire worldview has shifted as a result. I'm now slowly developing ideas for how to shift systems in that direction.<\/p><p>- I developed a model for how humans work I use to program my brain in very intentional ways. NVC principles serve as optimization techniques in it.<\/p><p>- An example of one such optimization was so profound it's worth noting by itself. I realized I can view every moment as meeting some needs of mine. The shittiest of moments can specifically meet a combination of 4 needs: learning to better meet my needs, learning to better empathize with others, gratitude for both of those things, and gratitude for the opportunity to learn. I began processing the world this way on New Year's Eve and it's changed practically everything I do. My reality is not at all what it once was and I love it.<\/p><p>- I have a concern about creating a programming language for human brains and giving it to the world. Specifically, I'm worried about people using it to exploit others. If I don't abandon my worry, I'll first encode NVC practices in it and release it as a sort of mental antivirus.<\/p><p>- In preparing to test an hypothesis my model generated about gender identity, I accidentally spawned a seemingly conscious second voice in my head. NVC is incredibly helpful when it comes to us communicating with each other.\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13973617\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gotrythis\" class=\"hnuser\">gotrythis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13973617\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"13974095\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crawfordcomeaux\" class=\"hnuser\">crawfordcomeaux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13974095\">on Mar 28, 2017<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"morespace\"><\/tr><tr><td><table border=\"0\"><tr><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=13967350&amp;p=2\" class=\"morelink\" rel=\"nofollow\">More<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:50:35+0000","updated_at":"2018-11-07T16:50:35+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":514,"title":"The Rust Programming Language","url":"https:\/\/doc.rust-lang.org\/book\/","content":"The Rust Programming Language\n    \n    \n    \n    <p>\u201cThe Rust Programming Language,\u201d is an introductory book about Rust. There are\nmultiple editions including a printed edition and a free online edition.<\/p>\n<p>The source for all editions lives\n<a href=\"https:\/\/github.com\/rust-lang\/book\">on GitHub<\/a>. Please open issues with any\nquestions, concerns, or tweaks.<\/p>\n<p>The 2018 edition of the book is an online \"living version\" of the book; based\non the second edition. This online book is updated as Rust updates. If you're\nnot sure what edition of the book to read, you should prefer this edition.<\/p>\n<p>You can find the 2018 edition <a href=\"https:\/\/doc.rust-lang.org\/book\/2018-edition\/index.html\">here<\/a>.<\/p>\n<p>The second edition is a complete re-write of the first edition. The second\nedition is based on Rust 1.21 and may not touch upon features found in\nsubsequent versions of Rust.<\/p>\n<p>You can purchase a <a href=\"https:\/\/nostarch.com\/Rust\">dead-tree copy from NoStarch\nPress<\/a>.<\/p>\n<p>This edition of the book is frozen; it contains exactly what's in the printed\nbook.<\/p>\n<p>You can find the second edition <a href=\"https:\/\/doc.rust-lang.org\/book\/second-edition\/index.html\">here<\/a>.<\/p>\n<p>The first edition is completely frozen but is still included in the\ndocumentation for historical reasons.<\/p>\n<p>You can find the first edition <a href=\"https:\/\/doc.rust-lang.org\/book\/first-edition\/index.html\">here<\/a>.<\/p>","created_at":"2018-11-07T16:50:32+0000","updated_at":"2018-11-07T16:50:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"doc.rust-lang.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":515,"title":"intermezzOS: a little OS","url":"https:\/\/intermezzos.github.io\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:50:32+0000","updated_at":"2018-11-07T16:50:32+0000","annotations":[],"reading_time":0,"domain_name":"intermezzos.github.io","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":516,"title":"Character representation of grey scale images","url":"http:\/\/paulbourke.net\/dataformats\/asciiart\/","content":"<div style=\"text-align: center;\">\nWritten by <a href=\"http:\/\/paulbourke.net\/dataformats\/\">Paul Bourke<\/a><br \/>February 1997<\/div>\n<p>\n\"Standard\" character ramp for grey scale pictures, black -&gt; white.\n<\/p><pre>\n   \"$@B%8&amp;WM#*oahkbdpqwmZO0QLCJUYXzcvunxrjft\/\\|()1{}[]?-_+~i!lI;:,\"^`'. \"\n<\/pre>\nA more convincing but shorter sequence for representing 10 levels of\ngrey is\n<pre>\n   \" .:-=+*#%@\"\n<\/pre>\nAn obvious problem is the variation in apparent density \nbetween some typefaces. \n<p>\nAnother issue is the stretching in  \nheight due to the rectangular aspect ratio of characters. \nWhen creating such images by software by sampling \"real\" images it\nis common to sample half as often in the vertical direction. Of course\nif the aspect ratio of the font being used is known then exact vertical\nand horizontal sampling can occur.\n<\/p><p>\nAs an example consider the following grey scale ramp on the left \nand the corresponding ASCII character representation on the right.\n<\/p><table style=\"width: 100%;\"><tr><td style=\"width: 20%;\">\n<img src=\"http:\/\/paulbourke.net\/dataformats\/asciiart\/asciiart1.gif\" alt=\"image\" \/><\/td><td style=\"width: 80%;\">\n<pre><span>\n=======--------:::::::::........                  \n=========--------:::::::::........                \n++=========--------:::::::::........              \n++++========---------:::::::::........            \n++++++========---------:::::::::........          \n++++++++=========--------:::::::::........        \n*+++++++++=========--------:::::::::........      \n***+++++++++=========--------:::::::::........    \n******++++++++=========--------:::::::::........  \n********++++++++=========--------:::::::::........\n#*********++++++++=========--------:::::::::......\n###*********++++++++=========--------:::::::::....\n#####********+++++++++=========--------:::::::::..\n#######********+++++++++=========--------:::::::::\n#########*********++++++++=========--------:::::::\n%%#########********+++++++++========---------:::::\n%%%%#########********+++++++++=========--------:::\n%%%%%%#########********+++++++++=========--------:\n%%%%%%%%#########*********++++++++========--------\n@@%%%%%%%%#########********+++++++++========------\n@@@@%%%%%%%%#########*********++++++++========----\n@@@@@@%%%%%%%%#########********+++++++++=========-\n@@@@@@@@%%%%%%%%#########********+++++++++========\n@@@@@@@@@@%%%%%%%%#########*********++++++++======\n@@@@@@@@@@@@%%%%%%%%#########********+++++++++====\n<\/span><\/pre>\n<\/td><\/tr><\/table><p>\n<b>Other Examples<\/b>\n<\/p><pre>\n                                                         ...                                                    \n                                                       ...::.....                                               \n                                                      ....:=-:--+..                                             \n                                                    ......::-+:.....                                            \n                                                 .......::-+@%=:.:....                                          \n                                              ........:-=%@@@@@#@-......                                        \n                                           ...........::%@@@@@@@#::........                                     \n                                        ....:::+:.:::::--%@@@@@#-::::......::.                                  \n                                     .......:*@#+=-=@@%@@@@@@@@@@%@+@#::--:-::..                                \n                                  ..........::#@@@@@@@@@@@@@@@@@@@@@@@@=@@@@*:..                                \n                         ..................+:-+%@@@@@@@@@@@@@@@@@@@@@@@@@@@*:....                               \n                     ..:.................:+@@%@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=:....                               \n                   ....:=:.....=:......::-+#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=-*:..                              \n                  .....:+-=*::-++-=::::::+@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@*...                              \n                .......::-+@@*@@@%%@%----%@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@--...                              \n             .........:--=%@@@@@@@@@@@@=+@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@+:...                              \n        ..........:+:::-+@@@@@@@@@@@@@@@#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@#-...                               \n   ..............:::%%@%@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@#-....                               \n+++++++*#+++*****#*%@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@#-:.....                               \n     ...........:::=*@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=.....                               \n       ...........:+-:--#@@@@@@@@@@@@@@@%@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@*:...                               \n           ..........::-+#@@@@@@@@@@@@@=+@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@-....                              \n               ........::-#@@%@@@@@@@=---@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@#=...                              \n                 ......:--*#=--+%-*-:::::#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@%*:...                              \n                   ....:=:....:+::....:::=#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@#+@:..                              \n                    ...::...............::=@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@#::...                               \n                        ..................:#-+#@@@@@@@@@@@@@@@@@@@@@@@@@@@%=:....                               \n                                 ...........::#@@@@@@@@@@@@@@@@@@@@@@@@+@@@#-:..                                \n                                    ........:%#@%*-+@@@@@@@@@@@@@@@%@+-:-*:--:..                                \n                                       .....--*#::::+:--=*@@@@@*---:-:.....:-..                                 \n                                          ...........:::*@@@@@@@#::.........                                    \n                                             .........::=@@@@@@@#-.......                                       \n                                                .......:---*@@=:-:....                                          \n                                                   .......::=*::....                                            \n                                                     .....::+:+-+..                                             \n                                                       ...:=....                                                \n                                                        ...:..                                                  \n              .@.                                    .\n              @m@,.                                 .@\n             .@m%nm@,.                            .@m@\n            .@nvv%vnmm@,.                      .@mn%n@\n           .@mnvvv%vvnnmm@,.                .@mmnv%vn@,\n           @mmnnvvv%vvvvvnnmm@,.        .@mmnnvvv%vvnm@\n           @mmnnvvvvv%vvvvvvnnmm@, ;;;@mmnnvvvvv%vvvnm@,\n           `@mmnnvvvvvv%vvvvvnnmmm;;@mmnnvvvvvv%vvvvnmm@\n            `@mmmnnvvvvvv%vvvnnmmm;%mmnnvvvvvv%vvvvnnmm@\n              `@m%v%v%v%v%v;%;%;%;%;%;%;%%%vv%vvvvnnnmm@\n              .,mm@@@@@mm%;;@@m@m@@m@@m@mm;;%%vvvnnnmm@;@,.\n           .,@mmmmmmvv%%;;@@vmvvvvvvvvvmvm@@;;%%vvnnm@;%mmm@,\n        .,@mmnnvvvvv%%;;@@vvvvv%%%%%%%vvvvmm@@;;%%mm@;%%nnnnm@,\n     .,@mnnvv%v%v%v%%;;@mmvvvv%%;*;*;%%vvvvmmm@;;%m;%%v%v%v%vmm@,.\n ,@mnnvv%v%v%v%v%v%v%;;@@vvvv%%;*;*;*;%%vvvvm@@;;m%%%v%v%v%v%v%vnnm@,\n `    `@mnnvv%v%v%v%%;;@mvvvvv%%;;*;;%%vvvmmmm@;;%m;%%v%v%v%vmm@'   '\n         `@mmnnvvvvv%%;;@@mvvvv%%%%%%%vvvvmm@@;;%%mm@;%%nnnnm@'\n            `@mmmmmmvv%%;;@@mvvvvvvvvvvmmm@@;;%%mmnmm@;%mmm@'\n               `mm@@@@@mm%;;@m@@m@m@m@@m@@;;%%vvvvvnmm@;@'\n              ,@m%v%v%v%v%v;%;%;%;%;%;%;%;%vv%vvvvvnnmm@\n            .@mmnnvvvvvvv%vvvvnnmm%mmnnvvvvvvv%vvvvnnmm@\n           .@mmnnvvvvvv%vvvvvvnnmm'`@mmnnvvvvvv%vvvnnmm@\n           @mmnnvvvvv%vvvvvvnnmm@':%::`@mmnnvvvv%vvvnm@'\n           @mmnnvvv%vvvvvnnmm@'`:::%%:::'`@mmnnvv%vvmm@\n           `@mnvvv%vvnnmm@'     `:;%%;:'     `@mvv%vm@'\n            `@mnv%vnnm@'          `;%;'         `@n%n@\n             `@m%mm@'              ;%;.           `@m@\n              @m@'                 `;%;             `@\n              `@'                   ;%;.             '    \n               `                    `;%;          \n   \n                                              _.oo.\n                      _.u[[\/;:,.         .odMMMMMM'\n                   .o888UU[[[\/;:-.  .o@P^    MMM^\n                  oN88888UU[[[\/;::-.        dP^\n                 dNMMNN888UU[[[\/;:--.   .o@P^\n               ,MMMMMMN888UU[[\/;::-. o@^\n                NNMMMNN888UU[[[\/~.o@P^\n                888888888UU[[[\/o@^-..\n               oI8888UU[[[\/o@P^:--..\n            .@^  YUU[[[\/o@^;::---..\n          oMP     ^\/o@P^;:::---..\n       .dMMM    .o@^ ^;::---...\n      dMMMMMMM@^`       `^^^^\n     YMMMUP^\n      ^^\n                 ,:\n               ,?}'\n             =!J |\n           ,*-. ?&amp;\n           WP) Y9P\n          YP   ,W'\n         ,W)  ,WW.'\n         WW) ,WWW)\n         7W),WWWW'\n         `WWWWWW'\n          9---W)\n      ,,--WPL=YXW===\n     (P),CY:,I\/X'F9P\n     WUT===---\/===9)\n     -HP+----Y(C=9W)\n      '9Y3'-'-OWPT-\n       'WWLUIECW\n        (:7L7C7'\n       ,P--=YWFL\n       Y-=:9)UW:L\n       3-'9=WU\/.7\n      ,WP9HTFUW'()\n       9W7W))UF 9)\n       7WYW))PW W\n        7WH)),WC)\n         7L--\/XY)\n         9+-,KY7)\n         W9-Y3+7)\n         W'=9WI7)\n        ,W  '-YY)\n         W    ::W                ,\n        ,T     :X)              ()\n        ()     '9W  'L.         ()         ,-\n        (C     =:9   '9L        ()        ,T\n        ()    ,,-7)    7WL      WW      ,F'\n        ()    , T9)     '9WL    --    ,YF\n        ()    '-\/(W       -==+PE9P7===O)          -,\n        'W, ,  T+\/WX=L-. ,WP+()+3L3,),=WL  --==-T-\n         7)    -,YW '-=9WPL+PT-- ':--L\/=9WP=-'\n         'W-,.-,++W.   WWHP    ,,-\/  .9CP3)\n          W  --':-9:7=9W-T ,-=FT''=++,(TFYW=====---,\n          W    .-='\/.  7W-,WE=--,,=-:9H=9W\"\"~~~~~~'\n          ()   ':'\/Y,  (L-9PXWWW,YWWX,(U3C        \n          9' ,,::\/Y,\/,  7LW+'-'7)()-'(MWW)\n       ,,-\/:',T,'-:',) ,3WWW, .Y=W'.(+WPW)\n      ,F=T:9\/:':C' \/W),WMW9PO),m-+--9+WYW)\n     ,3Y:\/--.'-,',F=FHWWE\/LMWU.'--X3CWW(WL\n     YP:\/:' -\/'-Y-,W-T)9X,WCWWWX=WWWW39\/OW\n     7WF:=,\/:-:P:,P(-'))PWWHYT79WWWHPW0W7W'\n     'WU7C-:=-=-C9'WF,):):H7L   '7CI7WEXP'\n      7L-,Y==3F:::,=,:-\/,'P=.,  ':79UWEW)\n      'WEW9P=\/,)\/ -:,P: \/ L7:'-=,-+YMWWW)\n       'W)+=T,T()\/-,F,,,),)  ',.-+(L=W9WW.\n        '+C\/:I'''',P:''\/ '  ''9.  == '-'7-\n         (W-+'. ,YF )\/:'      ')-. ,-:FX-L\n         'WM\/',\/CP \/,:'    ..:)  ,T','\/: 'W,\n          W--,YXT \/'')   ,P=-\/',P'  '(:'  'W,\n          (WEXWF Y' ,)  ,\/'-,,YT    \/\/\/  ,,'W.\n         ,WWWWT,,' .Y:\/.',,-,=',- ,YY(). +3,W)\n         WFXF:,'P ,,)\/  ,',P',,- ,FI,))) I3'W)\n         -HP,X'',\/ '  ,\/,\/' ,\/',,P3'I(:) W) W)       \/=+=,\n          9WY).,\/'  ,\/'-'   ,-=9-\/'Y'((',W) PW      \/'  '-==L,\n          'WY,'    ,\/,P   ,YP- C\/',',)( (W'(WW.    \/'       '7==L.\n           ()'    \/:\/' ,,WT'  3F',' \/)W (W (K()   \/'   .        '7X\n           ()   ,P,P',)T=:- ,WP'.' ,P,T (W (-9L ,Y)' ,X\/\/, .    Y:P\n          ,F   ,F,',--,\/:' ,+P' '  Y):) (E' YHWLWT)-''-9\/',-' ,,,WF\n         ,P.,P,)-3-- ,-,' ,WF.    ,Y (' (L-WCTWEW30V-\/',:'=\/P+E7WF\n         W- Y,P\/C)',Y',' ,WT      Y) :  (P-=Y:UW9CX)3-=- ,W:9\/PXXW.\n        \/T.\/:P\/)' ,P',' YW-      ,P'',  9M).())WTHW3,C'  9C9='W3WW)\n       ,EPOP\/YR. \/F ,',\/W)       \/'  :  (W)'W979WO0=WC:,..9LPXWWP-\n       3H:WL-R' \/' \/' \/WF       ,) ,,   (U'(HW=WWXO:--:,:'(W=WWF'\n      ,WLWWWI:,F' \/-'3WF '      Y  ) ,  (),T(0)WO9YPL.' ',WP=='\n        --YWX-F  Y',WWT' :':   (' ()7)  (MT: WP)3C)-''  3C'\n            WF  \/' YW--,  ,    Y  W (),YM+C' 9+I3UV:' .YP'\n           (T  3',H3-.. ,..  .,) ,) ()F-=T-. (0,9L,'  \/P'\n           ,W Y' 33P  .  \/    Y  Y) (Y' R,:  7)Y+-),,=W'\n           \/',F.,W)     ,,.' ,) ,W) +)  3),  (WT9XW=3P'\n          \/F:T.:WF.  '..:'   :' (W. 7) '=),  'WT7WWP '\n         ,P,F''WF  . , :-': ,)  YC..\/) 'HY.   WP0WC'\n        ,P:9::YP   '  '('   :   W) .W)  +3)   9TLWC\n        (P\/Y(,P' ... '':, .,)  ,W) :3)  X+.   WFUW)\n        'WW),I','  .., =  ':  ,O+' ,W'  )9,   99U()\n         7W,='.,' ' :.'. . '  ,W)  =3   )+.  ,OH:O)\n         'L,F,: '. :C::' '    (W)  9W   7+    'H,:L\n          7W'++: .. ,':' '    YT   Y).  :-.    XU:W\n          (T':,''','','       3'  ,-)   ,-'    77XW\n          (W),J.-:\/-:))'      P   )9)   :,Y  .  T,9)\n          (WUI:TY:,,,:,      \/' ,- W)   YC:     9\/7)\n          (U),-:-''.'=      (:,F' (W)   ,Y.     3=:L\n           (),:::',)\/'    ,,F9W'  YW)   \/L.    .7=9W,\n           (LUL-L.T-'.' ,WXM(W)   3W)   'U.     ,)-W\n           3X=((:,' ' ,WMWF-(+'   WW'   '=,'    ,ICW\n          ,T)=)K-=':-WPIWP':,:   ,WW    +\/, ..' :+,9)\n          Y):LX:.:=EHR,PU:'\/''   (WW    I:=, .,-9CO)\n          ()-+,,HPT+C:W9= ,)'    \/WT'   T.: --PCXCKF'\n          7LIHTP+OY3LW'3:,L..    WW)    ,,(W('MX'WT'\n           7T,I-:XF:WF(: ,)    ':WWT  ,=PT:T(AY) W\n           (PWW)W3=\/P,P  ,     ,'WW),YP,WH,)Y)TWX9)\n           3)OWRE)-YUY'... '  ..(WWXWW)9W+C)WUP9P3'\n          ,WTHEF:LOP:W ' ,.   ,:(WPY(W,(P::)W(P3+)\n          (P3WF\/:WM:() :.:      (WHY)39HC'U()(W,W)\n          (LW9\/CWY-,E'  ,'     'YPL\/T:WP:,(()3Y W)\n          (TLUEEP=7W+.,:)       -P-:,PWT.:Y()() W+\n          (EP\/30-OAT .'3  . . . .C,P):WP Y)()3C:PW\n          (PWMH:FXW'',-(, '   ',97WMU(7: )LW W .WW\n          (WOWF-7EP)-X3., , ,,WP+WYY+YW' )WW 3 .W7)\n           9W93UOY):.)\/.- :YWCWU-EIMC)E (-WP + ,WW'\n           (XWYUWY.,:'.,,YE3-7WE3WXV(UT,( W),T =P'\n           (PT709),)C:\/FY9)T.(W9YHL\/Y(C T,W)') W.\n           (+UTYH-:-=C-(P(-).WWF3:))3(U))(W)() P'\n           3P7Y3)\/'XP:)WP(J. WXCWKV:)()))(W'U)()\n           7OLY3',H9),YW'F ),W)CT)\/Y((-))'W,U)()\n           7F=T-\/T(=)A3C,)3)(WA()=)TY(CY'YWY(::)\n           W9C=()L\/3,9'\/('Y,YWU(XE\/))()E.YT)3:)L\n           W=P:F:(,)),,'F'\/:WP+3OY':)(R+ \/T,T')W\n        -=WRHX9C9-W'=,),)'A,A)XW779EXWK+.()3W),(,\n      ,W=-'L,,XX)\/)+'I 3)39I(UHE-+LX39TWH\/LUP)(H)\n     ,P:. ,-90\/,(F0'\/:,W \/\/'(YOC':--YY3\/IRW'9LT')\n     3W=:Y:-F.)Y:\/''Y \/O.=:,WL9) ,. ,.79=9PL'9(-W,\n     WWWWWWWPT:,::\/'-WH=9',P=-W3XU3-,W=YL-O3-O)X9WL\n      -7T--''=9W==W=9WWHW====''P======='---=T==F==9)\n<\/pre>\nOr try <a href=\"http:\/\/paulbourke.net\/dataformats\/asciiart\/mandel.c\">this silly piece<\/a> of C code (author unknown).\n<p>\n<b>Software<\/b><\/p><p>\nA Macintosh application has been written, called pict2ascii:,\nthat converts PICT images into an ascii bitmap format. The\nnumerical value for each pixel can be scaled to a number of ranges.\n<br \/>Update 2002, this software is no longer available.\n<\/p>","created_at":"2018-11-07T16:50:32+0000","updated_at":"2018-11-07T16:50:32+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"paulbourke.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":513,"title":"How To Resize ext3 Partitions Without Losing Data - Page 3 - Page 3","url":"https:\/\/www.howtoforge.com\/linux_resizing_ext3_partitions_p3","content":"<p><a href=\"https:\/\/ispprotect.com\/?pk_campaign=htftxt\" target=\"_blank\" rel=\"nofollow\">Scan your Web-Server for Malware with ISPProtect now. Get Free Trial.<\/a><\/p><div class=\"contributeEdit\"><div id=\"tocContainer\"><h3>On this page<\/h3><ol class=\"toc\"><li><a href=\"#-merge-two-ext-partitions\">4 Merge Two ext3 Partitions<\/a><\/li>\n<li><a href=\"#-links\">5 Links<\/a><\/li>\n<\/ol><\/div><\/div><h3 id=\"-merge-two-ext-partitions\">4 Merge Two ext3 Partitions<\/h3><p>In this example I have my system partition \/dev\/sda1 again (about 6GB of size) which is followed directly by the partition \/dev\/sda3 (about 4GB of size) on the hard disk. \/dev\/sda3 is mounted to the \/data directory and doesn't hold files needed by the Linux system, just user data. The current partitioning looks like this:<\/p><p class=\"command\">df -h<\/p><p class=\"system\">Filesystem\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0Size\u00a0\u00a0Used\u00a0Avail\u00a0Use%\u00a0Mounted\u00a0on<br \/>\/dev\/sda1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a05.8G\u00a0\u00a04.1G\u00a0\u00a01.6G\u00a0\u00a073%\u00a0\/<br \/>varrun\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0132K\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a01%\u00a0\/var\/run<br \/>varlock\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/var\/lock<br \/>udev\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a056K\u00a0\u00a0\u00a010M\u00a0\u00a0\u00a01%\u00a0\/dev<br \/>devshm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a0\u00a0\u00a00\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a00%\u00a0\/dev\/shm<br \/>lrm\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a094M\u00a0\u00a0\u00a018M\u00a0\u00a0\u00a077M\u00a0\u00a019%\u00a0\/lib\/modules\/2.6.17-10-generic\/volatile<br \/>\/dev\/sda3\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a03.5G\u00a0\u00a0\u00a072M\u00a0\u00a03.3G\u00a0\u00a0\u00a03%\u00a0\/data<\/p><p>To merge \/dev\/sda1 and \/dev\/sda3, we have to delete \/dev\/sda3 and then enlarge \/dev\/sda1 as described in chapter 3. This means that all data on \/dev\/sda3 will be lost, so if you need it later on please back it up somewhere else and then restore it to the new and bigger \/dev\/sda1 afterwards! <\/p><p>Now we open \/etc\/fstab and remove the line for \/dev\/sda3 there if it exists:<\/p><p class=\"command\">vi \/etc\/fstab <\/p><p>The new file without \/dev\/sda3 could look like this:<\/p><table class=\"\" border=\"1\" cellpadding=\"2\" style=\"margin: auto; background-color: #cccccc; border-color: #000000; border-spacing: 0px; width: 90%;\"><tbody><tr><td class=\"\">\n<pre># \/etc\/fstab: static file system information.&#13;\n#&#13;\n# &#13;\nproc            \/proc           proc    defaults        0       0&#13;\n# \/dev\/sda1&#13;\nUUID=566fd9e9-098f-4aae-9908-51efe171d8ba \/               ext3    defaults,errors=remount-ro 0       1&#13;\n# \/dev\/sda5&#13;\nUUID=82102b65-35db-469a-9532-03d619d8cffb none            swap    sw              0       0&#13;\n\/dev\/hdc        \/media\/cdrom0   udf,iso9660 user,noauto     0       0&#13;\n\/dev\/           \/media\/floppy0  auto    rw,user,noauto  0       0<\/pre>\n<\/td><\/tr><\/tbody><\/table><p>Then we unmount \/dev\/sda3 and run fdisk to delete it. This can still be done on the original system as \/dev\/sda3 doesn't contain system files:<\/p><p class=\"command\">umount \/dev\/sda3<\/p><p class=\"command\">fdisk \/dev\/sda<\/p><p>(Yes, it's \/dev\/sda, not \/dev\/sda1.)<\/p><p class=\"system\">The number of cylinders for this disk is set to 1305.<br \/>There is nothing wrong with that, but this is larger than 1024,<br \/>and could in certain setups cause problems with:<br \/>1) software that runs at boot time (e.g., old versions of LILO)<br \/>2) booting and partitioning software from other OSs<br \/>\u00a0\u00a0\u00a0(e.g., DOS FDISK, OS\/2 FDISK)<\/p><p>Type m to get a list of all commands:<\/p><p class=\"system\">Command\u00a0(m\u00a0for\u00a0help):\u00a0m<br \/>Command\u00a0action<br \/>\u00a0\u00a0\u00a0a\u00a0\u00a0\u00a0toggle\u00a0a\u00a0bootable\u00a0flag<br \/>\u00a0\u00a0\u00a0b\u00a0\u00a0\u00a0edit\u00a0bsd\u00a0disklabel<br \/>\u00a0\u00a0\u00a0c\u00a0\u00a0\u00a0toggle\u00a0the\u00a0dos\u00a0compatibility\u00a0flag<br \/>\u00a0\u00a0\u00a0d\u00a0\u00a0\u00a0delete\u00a0a\u00a0partition<br \/>\u00a0\u00a0\u00a0l\u00a0\u00a0\u00a0list\u00a0known\u00a0partition\u00a0types<br \/>\u00a0\u00a0\u00a0m\u00a0\u00a0\u00a0print\u00a0this\u00a0menu<br \/>\u00a0\u00a0\u00a0n\u00a0\u00a0\u00a0add\u00a0a\u00a0new\u00a0partition<br \/>\u00a0\u00a0\u00a0o\u00a0\u00a0\u00a0create\u00a0a\u00a0new\u00a0empty\u00a0DOS\u00a0partition\u00a0table<br \/>\u00a0\u00a0\u00a0p\u00a0\u00a0\u00a0print\u00a0the\u00a0partition\u00a0table<br \/>\u00a0\u00a0\u00a0q\u00a0\u00a0\u00a0quit\u00a0without\u00a0saving\u00a0changes<br \/>\u00a0\u00a0\u00a0s\u00a0\u00a0\u00a0create\u00a0a\u00a0new\u00a0empty\u00a0Sun\u00a0disklabel<br \/>\u00a0\u00a0\u00a0t\u00a0\u00a0\u00a0change\u00a0a\u00a0partition's\u00a0system\u00a0id<br \/>\u00a0\u00a0\u00a0u\u00a0\u00a0\u00a0change\u00a0display\/entry\u00a0units<br \/>\u00a0\u00a0\u00a0v\u00a0\u00a0\u00a0verify\u00a0the\u00a0partition\u00a0table<br \/>\u00a0\u00a0\u00a0w\u00a0\u00a0\u00a0write\u00a0table\u00a0to\u00a0disk\u00a0and\u00a0exit<br \/>\u00a0\u00a0\u00a0x\u00a0\u00a0\u00a0extra\u00a0functionality\u00a0(experts\u00a0only)<\/p><p>Now let's delete \/dev\/sda3:<\/p><p class=\"system\">Command (m for help): d<br \/>Partition number (1-5): 3<\/p><p>Afterwards we write the new partition table to the disk:<\/p><p class=\"system\">Command (m for help): w<br \/>The partition table has been altered!<\/p><p class=\"system\">Calling ioctl() to re-read partition table.<\/p><p class=\"system\">WARNING: Re-reading the partition table failed with error 16: Device or resource busy.<br \/>The kernel still uses the old table.<br \/>The new table will be used at the next reboot.<br \/>Syncing disks.<\/p><p>Now we shut down the system:<\/p><p class=\"command\">shutdown -r now<\/p><p>and boot into our Knoppix Live-CD (or your rescue system). From here on the steps are identical to chapter 3, beginning with<\/p><p class=\"command\">su<\/p><p class=\"command\">umount \/dev\/sda1<\/p><p>so please refer to that chapter. <\/p><h3 id=\"-links\">5 Links<\/h3><ul><li>Knoppix: <a href=\"http:\/\/www.knopper.net\/knoppix-mirrors\/index-en.html\" target=\"_blank\">http:\/\/www.knopper.net\/knoppix-mirrors\/index-en.html<\/a><\/li>\n<\/ul><div><p><img src=\"https:\/\/www.howtoforge.com\/images\/pdficon_small.png\" alt=\"image\" \/><a class=\"pdf-btn\" href=\"denied:javascript:void(0);\">view as pdf<\/a> | <img src=\"https:\/\/www.howtoforge.com\/images\/print.gif\" alt=\"image\" \/><\/p><\/div><div><p><b>Share this page:<\/b><\/p><p>\n<a href=\"https:\/\/www.facebook.com\/sharer.php?u=https%3A%2F%2Fwww.howtoforge.com%2Flinux_resizing_ext3_partitions_p3\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-recommend.png\" height=\"20\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/twitter.com\/intent\/tweet?url=https%3A%2F%2Fwww.howtoforge.com%2Flinux_resizing_ext3_partitions_p3&amp;text=How+To+Resize+ext3+Partitions+Without+Losing+Data+-+Page+3&amp;via=howtoforgecom&amp;related=howtoforgecom\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-tweet.png\" height=\"20\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/twitter.com\/howtoforgecom\/\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-follow.png\" height=\"20\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/plus.google.com\/share?url=https%3A%2F%2Fwww.howtoforge.com%2Flinux_resizing_ext3_partitions_p3\" target=\"_blank\" rel=\"nofollow\"><img src=\"https:\/\/www.howtoforge.com\/images\/socialmedia\/btn-gplus.png\" height=\"20\" alt=\"image\" \/><\/a>\n<\/p><\/div>","created_at":"2018-11-07T16:50:31+0000","updated_at":"2018-11-07T16:50:31+0000","published_by":["Falko Timme"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"www.howtoforge.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":512,"title":"web interface tools which adhere to the unix philosophy.","url":"https:\/\/www.uzbl.org\/readme.php","content":"<h3>INTRODUCTION<\/h3><p>Any program can only be really useful if it complies with the Unix\nphilosophy. Web browsers (and other tools that work with HTML, such as feed\nreaders) are frequent violators of this principle:<\/p><ul><li>They build in way too much things into one (complex) program, dramatically\ndecreasing the options to do things the way you want.<\/li>\n<li>They store things in way too fancy formats (XML, RDF, SQLite, etc.) which are\nhard to store under version control, reuse in other scripts, and so on.<\/li>\n<\/ul><p>The Uzbl project was started as an attempt to resolve this.<\/p><h3>EDITIONS<\/h3><p>\"Uzbl\" is an umbrella project consisting of different flavors. In the future\nmore things may come, but for now:<\/p><h4>uzbl-core<\/h4><p>Main component meant for integration with other tools and scripts.<\/p><ul><li>Uses WebKitGtk+ for rendering and network interaction (libsoup). CSS,\nJavaScript, and plugin support come for free.<\/li>\n<li>Provides interfaces to get data in (commands\/configuration) and out (events):\nstdin\/stdout\/fifo\/Unix sockets.<\/li>\n<li>You see a WebKit view and (optionally) a status bar which gets populated\nexternally.<\/li>\n<li>No built-in means for URL changing, loading\/saving of bookmarks, saving\nhistory, keybindings, downloads, etc.<\/li>\n<li>Extra functionality: many sample scripts come with it. More are available on\nthe <a href=\"http:\/\/www.uzbl.org\/wiki\/scripts\">Uzbl wiki<\/a> or you can write them\nyourself.<\/li>\n<li>Entire configuration\/state can be changed at runtime.<\/li>\n<li>Uzbl keeps it simple, and puts <strong>you<\/strong> in charge.<\/li>\n<\/ul><h4>uzbl-browser<\/h4><p>A minimal browser which encompasses a default configuration file, event manager\nand a selection of scripts which gives a comparable browsing experience to the\nmore common browsers.<\/p><ul><li>Uses a set of scripts (mostly Python) that will fit most people, so things\nwork out of the box; yet plenty of room for customization.<\/li>\n<li>Brings everything you expect: URL changing, history, downloads, form filling,\nlink navigation, cookies, event management, etc.<\/li>\n<li>Advanced, customizable keyboard interface with support for modes, modkeys,\nmultichars, variables (keywords) etc. (e.g., you can tweak the interface to\nbe Vi-like, Emacs-like or any-other-program-like).<\/li>\n<li>Adequate default configuration.<\/li>\n<li>Focus on plaintext storage for your data and configuration files in simple,\neasy-to-parse formats and adherence to the <a href=\"http:\/\/standards.freedesktop.org\/basedir-spec\/basedir-spec-latest.html\">XDG basedir\nspec<\/a>.<\/li>\n<li>Visually, similar to <code>uzbl-core<\/code> except that the status bar contains useful\ninformation.<\/li>\n<li>One window per page.<\/li>\n<\/ul><h4>uzbl-tabbed<\/h4><p>A flavor of <code>uzbl<\/code> built on top of <code>uzbl-browser<\/code> which adds tabbed browsing.<\/p><ul><li>Spawns one window containing multiple tabs, each tab containing a full\nembedded <code>uzbl-browser<\/code>.<\/li>\n<li>Ideal as a quick and simple solution to manage multiple <code>uzbl-browser<\/code>\ninstances without getting lost.<\/li>\n<\/ul><p>Throughout the documentation, when referring to <code>uzbl<\/code> we mean <code>uzbl-core<\/code>,\nunless otherwise specified.<\/p><h3>CONFIGURATION \/ CONTROL:<\/h3><p>The general idea is that Uzbl by default is very bare bones. You can send it\ncommands to update settings and perform actions, through various\ninterfaces. There is a limited, built-in default configuration. It sets a basic\nstatus bar, title format, and network connection settings. By default, there\nare <em>no<\/em> keybindings defined at all (default keybindings would be\ncounterproductive when you try to customize). For examples of the possibilities\nwhat you can do, please see the sample configurations, and <a href=\"http:\/\/www.uzbl.org\/wiki\/config\">wiki\npage<\/a>.<\/p><p>There are several interfaces to interact with Uzbl:<\/p><ul><li><code>uzbl --config &lt;filename&gt;<\/code>: <code>&lt;filename&gt;<\/code> will be read line by line, and the\ncommands in it will be executed. Useful to configure Uzbl at startup. If you\nhave a file in <code>$XDG_CONFIG_HOME\/uzbl\/config<\/code> (this expands to\n<code>~\/.config\/uzbl\/config<\/code> on most systems), it will be automatically recognized.<\/li>\n<li><code>stdin<\/code>: to write commands into <code>stdin<\/code>, use <code>--config -<\/code> (or <code>-c -<\/code>).<\/li>\n<li><p>Interactive: you can enter commands (and bind them to shortcuts, even at\nruntime). By default, the behaviour is modal (vi-like):<\/p>\n<ul><li>command mode: every keystroke is interpreted to run commands<\/li>\n<li>insert mode: keystrokes are not interpreted so you can enter text into html\nforms<\/li>\n<\/ul><p>There is also support for \"chained\" commands (multiple characters long), and\nkeyworded commands. Also you can trigger incremental matching on commands and\nvariables after pressing return. See the sample configuration file for more\ninformation.<\/p>\n<p>By default, copy and paste works when typing commands:<\/p>\n<ul><li><code>insert<\/code> (paste X cliboard)<\/li>\n<li><code>shift insert<\/code> (paste primary selection buffer)<\/li>\n<\/ul><\/li>\n<li><p>FIFO &amp; socket files: If enabled by setting their paths through one of the\nabove means, you can have socket and FIFO files available which are very\nuseful to control <code>uzbl<\/code> using external programs.<\/p>\n<ul><li>The advantage of the FIFO is you can write directly to it, but it's half\nduplex only (<code>uzbl<\/code> cannot send a response to you).<\/li>\n<li><p>The socket is full duplex but a socket-compatible wrapper such as <code>socat<\/code>\nis needed to work with it. For example:<\/p>\n<p>echo <command label=\"label\" \/> | socat - unix-connect:<\/p><\/li>\n<\/ul><\/li>\n<\/ul><p>When <code>uzbl<\/code> forks a new instance (e.g., \"open in new window\") it will use the\nsame command line arguments (e.g., the same <code>--config &lt;file&gt;<\/code>), except for\n<code>--named<\/code>. If you made changes to the configuration at runtime, these are not\npassed on to the child.<\/p><h4>WebKit1 vs. WebKit2<\/h4><p>WebKit2 changes many things compared to WebKit1, some of which impact <code>uzbl<\/code>\ngreatly. This version of <code>uzbl<\/code> only supports WebKit1 while older version tried\nto support both at the same time.<\/p><p>For a <code>uzbl<\/code> built on WebKit2 check out the\n<a href=\"https:\/\/github.com\/uzbl\/uzbl\/tree\/next\">next<\/a> branch<\/p><h3>COMMAND SYNTAX<\/h3><p><code>Uzbl<\/code> will read commands via standard input, named FIFO pipe (once <code>fifo_dir<\/code>\nis set) and Unix socket (once <code>socket_dir<\/code> is set). For convenience, <code>uzbl<\/code> can\nalso be instructed to read commands from a file on startup by using the\n<code>--config<\/code> option. Indeed, the config file is nothing more than a list of\ncommands with support for comments (using the <code>#<\/code> character).<\/p><p>Each command starts with the name of a command or a <code>uzbl<\/code> variable that expands\nto it. A command is terminated by a newline. Empty lines and lines that start\nwith the hash sign (ignoring leading whitespace) are ignored by the parser.\nCommand names are always written in lowercase.<\/p><p>A list of commands supported by <code>uzbl<\/code> is given here. Optional arguments are\ngiven in square brackets (<code>[]<\/code>) with literal strings options in lowercase\nseparated by a pipe (<code>|<\/code>) character. <code>N<\/code> indicates an integer value and other\nuppercase values are placeholders describing what the value should be. Required\noptions are given in angle brackets (<code>&lt;&gt;<\/code>). Arguments on which argument\nsplitting is not performed are given in curly braces (<code>{}<\/code>). Variable arguments\nare indicated with ellipses (<code>...<\/code>). As an example:<\/p><pre>`example1 &lt;a|b&gt; &lt;N&gt; [URI...]`\n<\/pre><p>means that the command <code>example<\/code> requires its first argument which must either\nbe a literal <code>a<\/code> or <code>b<\/code>, a second argument which is an integer, and any\nremaining arguments are treated as URIs. Another example:<\/p><pre>`example2 &lt;KEY&gt; {VALUE}`\n<\/pre><p>means that the first argument is split off used as a key and the remaining\nstring is used verbatim. The command <code>example2 key many args<\/code> would have <code>key<\/code>\nas <code>KEY<\/code> and <code>many args<\/code> as <code>VALUE<\/code>.<\/p><h4>Navigation<\/h4><ul><li><code>back [N]<\/code>\n<ul><li>Navigate to the Nth (default: 1) previous URI in the instance history.<\/li>\n<\/ul><\/li>\n<li><code>forward [N]<\/code>\n<ul><li>Navigate to the Nth (default: 1) next URI in the instance history.<\/li>\n<\/ul><\/li>\n<li><code>reload [cached|full]<\/code>\n<ul><li>Reload the current page. If <code>cached<\/code> is given (the default), the cache is\nused. If <code>full<\/code> is given, the cache is ignored.<\/li>\n<\/ul><\/li>\n<li><code>stop<\/code>\n<ul><li>Stop loading the current page. This usually means that no new network\nrequests are made, but JavaScript can still run.<\/li>\n<\/ul><\/li>\n<li><code>uri {URI}<\/code>\n<ul><li>Tell <code>uzbl<\/code> to navigate to the given URI.<\/li>\n<\/ul><\/li>\n<li><code>download &lt;URI&gt; [DESTINATION]<\/code>\n<ul><li>Tell WebKit to download a URI.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Page<\/h4><ul><li><code>load &lt;COMMAND&gt;<\/code>\n<ul><li>Load content into the page. Supported subcommands include:<\/li>\n<li><code>html &lt;URI&gt; &lt;CONTENT&gt;<\/code>\n<ul><li>Load content as HTML relative to a URI.<\/li>\n<\/ul><\/li>\n<li><code>text &lt;CONTENT&gt;<\/code>\n<ul><li>Load content as plain text.<\/li>\n<\/ul><\/li>\n<li><code>error_html &lt;URI&gt; &lt;BASE_URI&gt; &lt;CONTENT&gt;<\/code>\n<ul><li>Load content an error page (which does not appear in the forward\/back\nlist) at the given URI with the given<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>frame &lt;list|focus|reload|stop|dump|inject|get|set&gt; [NAME]<\/code>\n<ul><li>Operate on frames in the current page. If not name is given, the <code>_current<\/code>\nframe is used. Supported subcommands include:<\/li>\n<li><code>list<\/code> (Unimplemented)\n<ul><li>Returns a JSON object representing the of frames in the frame.<\/li>\n<\/ul><\/li>\n<li><code>focus<\/code> (Unimplemented)\n<ul><li>Give focus to the frame.<\/li>\n<\/ul><\/li>\n<li><code>reload<\/code>\n<ul><li>Reload the frame.<\/li>\n<\/ul><\/li>\n<li><code>stop<\/code>\n<ul><li>Stop loading the frame.<\/li>\n<\/ul><\/li>\n<li><code>save<\/code> (Unimplemented)\n<ul><li>Save the content frame.<\/li>\n<\/ul><\/li>\n<li><code>load<\/code> (Unimplemented)\n<ul><li>Load content into the frame.<\/li>\n<\/ul><\/li>\n<li><code>get &lt;VARIABLE&gt;<\/code> (Unimplemented)\n<ul><li>Get information about a frame.<\/li>\n<\/ul><\/li>\n<li><code>set &lt;VARIABLE&gt; &lt;VALUE&gt;<\/code> (Unimplemented)\n<ul><li>Set information about a frame.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h4>Cookie<\/h4><ul><li><code>cookie &lt;add|delete|clear&gt;<\/code>\n<ul><li>Manage cookies in <code>uzbl<\/code>. The subcommands work as follows:<\/li>\n<li><code>add &lt;HOST&gt; &lt;PATH&gt; &lt;NAME&gt; &lt;VALUE&gt; &lt;SCHEME&gt; &lt;EXPIRATION&gt;<\/code>\n<ul><li>Manually add a cookie.<\/li>\n<\/ul><\/li>\n<li><code>delete &lt;DOMAIN&gt; &lt;PATH&gt; &lt;NAME&gt; &lt;VALUE&gt;<\/code>\n<ul><li>Delete a cookie from the cookie jar.<\/li>\n<\/ul><\/li>\n<li><code>clear all<\/code>\n<ul><li>Delete all cookies.<\/li>\n<\/ul><\/li>\n<li><code>clear domain [DOMAIN...]<\/code>\n<ul><li>Delete all cookies matching the given domains.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h4>Display<\/h4><ul><li><code>scroll &lt;horizontal|vertical&gt; &lt;VALUE&gt;<\/code>\n<ul><li>Scroll either the horizontal or vertical scrollbar for the page. The value\nmay be one of:<\/li>\n<li><code>begin<\/code>\n<ul><li>Scrolls to the beginning of the page (either left or up).<\/li>\n<\/ul><\/li>\n<li><code>end<\/code>\n<ul><li>Scrolls to the end of the page (either right or down).<\/li>\n<\/ul><\/li>\n<li><code>[-]N<\/code>\n<ul><li>Scrolls N \"units\" as defined by the scrollbar (seems to be pixels).<\/li>\n<\/ul><\/li>\n<li><code>[-]N%<\/code>\n<ul><li>Scrolls N% of a single page.<\/li>\n<\/ul><\/li>\n<li><code>N!<\/code>\n<ul><li>Scrolls to position N on the scrollbar (also in \"units\").<\/li>\n<\/ul><\/li>\n<li><code>N%!<\/code>\n<ul><li>Scrolls to position N% on the scrollbar.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>zoom &lt;in|out|set&gt; [VALUE]<\/code>\n<ul><li>Zoom either in or out by the given amount. When setting, the value is\nrequired. If no value is given, the variable <code>zoom_step<\/code> is used.<\/li>\n<\/ul><\/li>\n<li><code>hardcopy [REGION]<\/code>\n<ul><li>Print the given region. Supported regions include:<\/li>\n<li><code>page<\/code> (the default)\n<ul><li>Prints the entire page.<\/li>\n<\/ul><\/li>\n<li><code>frame [NAME]<\/code>\n<ul><li>Prints the frame with the given name or the current frame if no name is\ngiven.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>geometry &lt;SIZE&gt;<\/code>\n<ul><li>Set the size of the <code>uzbl<\/code> window. This is subject to window manager\npolicy. If the size is <code>maximized<\/code>, the <code>uzbl<\/code> window will try to maximize\nitself. Otherwise, the size is parsed using the\n<code>WIDTHxHEIGHT\u00b1XOFFSET\u00b1YOFFSET<\/code> pattern.<\/li>\n<\/ul><\/li>\n<li><code>snapshot &lt;PATH&gt; &lt;FORMAT&gt; &lt;REGION&gt; [FLAG...]<\/code> (WebKit1 &gt;= 1.9.6)\n<ul><li>Saves the current page as an image to the given path. Currently supported\nformats include: <code>png<\/code>. Acceptable regions include:<\/li>\n<li><code>visible<\/code>\n<ul><li>Only includes the regions of the page which are currently visible.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h4>Content<\/h4><ul><li><code>plugin &lt;COMMAND&gt;<\/code>\n<ul><li>Exposes control of plugin-related information. Supported subcommands include:<\/li>\n<li><code>refresh<\/code>\n<ul><li>Refresh the plugin database.<\/li>\n<\/ul><\/li>\n<li><code>toggle [NAME...]<\/code>\n<ul><li>Toggle whether the given plugins are enabled (or all if none are given).<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>remove_all_db<\/code>\n<ul><li>Remove all web databases.<\/li>\n<\/ul><\/li>\n<li><code>spell &lt;COMMAND&gt;<\/code> (WebKit1 &gt;= 1.5.1)\n<ul><li>Use WebKit's spell check logic. Supported subcommands include:<\/li>\n<li><code>ignore [WORD...]<\/code>\n<ul><li>Add the given words to the list of words to ignore.<\/li>\n<\/ul><\/li>\n<li><code>learn [WORD...]<\/code>\n<ul><li>Teaches that the given words are spelled correctly.<\/li>\n<\/ul><\/li>\n<li><code>autocorrect &lt;WORD&gt;<\/code>\n<ul><li>Returns the word as autocorrected by the checker.<\/li>\n<\/ul><\/li>\n<li><code>guesses &lt;WORD&gt;<\/code>\n<ul><li>Returns the guesses for the word given by the checker as a JSON list.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>favicon &lt;COMMAND&gt;<\/code>\n<ul><li>Controls the favicon database. Supported subcommands include:<\/li>\n<li><code>clear<\/code>\n<ul><li>Clears all cached favicons.<\/li>\n<\/ul><\/li>\n<li><code>uri &lt;URI&gt;<\/code>\n<ul><li>Returns the URI of the favicon for the given URI.<\/li>\n<\/ul><\/li>\n<li><code>save &lt;URI&gt; &lt;PATH&gt;<\/code> (Unimplemented)\n<ul><li>Saves the favicon for the given URI to a path.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>css &lt;COMMAND&gt;<\/code>\n<ul><li>Controls CSS settings in web pages. Supported subcommands include:<\/li>\n<li><code>add &lt;URI&gt;<\/code>\n<ul><li>Adds a CSS file to pages when loaded. Despite the name, subsequent calls\nreplace any previous invocation. Meaning that only one single stylesheet\nmay be activated with this command.<\/li>\n<\/ul><\/li>\n<li><code>clear<\/code>\nClears all user-supplied stylesheets.<\/li>\n<\/ul><\/li>\n<li><code>scheme &lt;SCHEME&gt; {COMMAND}<\/code>\n<ul><li>Registers a custom scheme handler for <code>uzbl<\/code>. The handler should accept a\nsingle argument for the URI to load and return HTML. When run, the output\nis interpreted as content at the URL with a leading line with the MIME\ntype.<\/li>\n<\/ul><\/li>\n<li><code>menu &lt;COMMAND&gt;<\/code>\n<ul><li>Controls the context menu shown in <code>uzbl<\/code>. Supported subcommands include:<\/li>\n<li><code>add &lt;OBJECT&gt; &lt;NAME&gt; &lt;COMMAND&gt;<\/code>\n<ul><li>Appends a new entry to the menu with the given name and runs the given\ncommand when used.<\/li>\n<\/ul><\/li>\n<li><code>add_separator &lt;OBJECT&gt; &lt;NAME&gt;<\/code>\n<ul><li>Appends a new separator to the menu with the given name.<\/li>\n<\/ul><\/li>\n<li><code>remove &lt;NAME&gt;<\/code>\n<ul><li>Remove items which have the given name.<\/li>\n<\/ul><\/li>\n<li><code>query &lt;NAME&gt;<\/code>\n<ul><li>Returns the command for the item with the given name.<\/li>\n<\/ul><\/li>\n<li><code>list<\/code>\n<ul><li>Returns the names of items in the menu as a JSON list.\nThe list of supported objects is:<\/li>\n<\/ul><\/li>\n<li><code>document<\/code><\/li>\n<li><code>link<\/code><\/li>\n<li><code>image<\/code><\/li>\n<li><code>media<\/code><\/li>\n<\/ul><\/li>\n<\/ul><h4>Search<\/h4><ul><li><code>search &lt;COMMAND&gt;<\/code>\n<ul><li>Controls WebKit's search mechanisms. Supported subcommands include:<\/li>\n<li><code>option &lt;OPTION...&gt;<\/code>\n<ul><li>Manage options for searching. Options may be prefixed with <code>+<\/code> to set,\n<code>-<\/code> to unset, <code>!<\/code> to toggle, and <code>~<\/code> to reset to its default. The\ndefault options have <code>wrap<\/code> enabled. Recognized options include:<\/li>\n<li><code>wrap<\/code>\n<ul><li>If set, searches will wrap around when the start or end of the\ndocument is reached.<\/li>\n<\/ul><\/li>\n<li><code>case_insensitive<\/code>\n<ul><li>If set, searches are case-insensitive.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>options<\/code>\n<ul><li>Returns the current options as a JSON list.<\/li>\n<\/ul><\/li>\n<li><code>clear<\/code>\n<ul><li>Turn off searching. The current options and search string are\npreserved.<\/li>\n<\/ul><\/li>\n<li><code>reset<\/code>\n<ul><li>Turn off searching, setting the options back to the default and\nforgetting the search string.<\/li>\n<\/ul><\/li>\n<li><code>find {STRING}<\/code>\n<ul><li>Search forward in the document for the string. If no string is given,\nthe previous search string is used.<\/li>\n<\/ul><\/li>\n<li><code>rfind {STRING}<\/code>\n<ul><li>As <code>find<\/code>, but searches backwards in the document.<\/li>\n<\/ul><\/li>\n<li><code>next<\/code>\n<ul><li>Jumps to the next instance of the search string in the document in the\nsame direction as the last <code>find<\/code> or <code>rfind<\/code> command.<\/li>\n<\/ul><\/li>\n<li><code>prev<\/code>\n<ul><li>As <code>next<\/code>, but searches in the opposite direction from the last <code>find<\/code>\nor <code>rfind<\/code> command.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><p><code>security &lt;SCHEME&gt; &lt;COMMAND&gt; &lt;OPTION&gt;<\/code><\/p>\n<ul><li>Controls security options for a scheme. Recognized commands include:<\/li>\n<li><code>get<\/code>\n<ul><li>Returns <code>true<\/code> if the option is set for the scheme, <code>false<\/code> otherwise.<\/li>\n<\/ul><\/li>\n<li><code>set<\/code>\n<ul><li>Sets the option for the scheme.<\/li>\n<\/ul><\/li>\n<li><code>unset<\/code> (WebKit1 only)\n<ul><li>Unsets the option for the scheme.<\/li>\n<\/ul><\/li>\n<\/ul><p>Supported options include (\"scheme\" here is used as \"content referenced\nwith this scheme\"):<\/p>\n<ul><li><code>local<\/code>\n<ul><li>If set, non-local schemes cannot link or access pages this scheme.<\/li>\n<\/ul><\/li>\n<li><code>no_access<\/code>\n<ul><li>If set, the scheme cannot access other schemes.<\/li>\n<\/ul><\/li>\n<li><code>display_isolated<\/code>\n<ul><li>If set, the scheme can only be displayed in the same scheme.<\/li>\n<\/ul><\/li>\n<li><code>secure<\/code>\n<ul><li>If set, the scheme is considered \"secure\" and will not generate\nwarnings when referenced by an HTTPS page.<\/li>\n<\/ul><\/li>\n<li><code>cors_enabled<\/code>\n<ul><li>If set, the scheme is allowed to make cross-origin resource sharing.<\/li>\n<\/ul><\/li>\n<li><code>empty_document<\/code>\n<ul><li>If set, the scheme is allowed to be committed synchronously.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>inspector &lt;COMMAND&gt;<\/code>\n<ul><li>Control the web inspector. Supported subcommands include:<\/li>\n<li><code>show<\/code>\n<ul><li>Show the web inspector.<\/li>\n<\/ul><\/li>\n<li><code>close<\/code>\n<ul><li>Close the web inspector.<\/li>\n<\/ul><\/li>\n<li><code>coord &lt;X&gt; &lt;Y&gt;<\/code>\n<ul><li>Inspect the page at the given coordinate. The coordinate is relative to\nthe current viewport.<\/li>\n<\/ul><\/li>\n<li><code>node &lt;NODE&gt;<\/code> (WebKit1 &gt;= 1.3.17) (Unimplemented)\n<ul><li>Inspect the node which matches the given CSS selector string. If\nmultiple nodes match, the first one is used.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h4>Execution<\/h4><ul><li><code>js &lt;CONTEXT&gt; &lt;file|string&gt; &lt;VALUE&gt;<\/code>\n<ul><li>Run JavaScript code. If <code>file<\/code> is given, the value is interpreted as a path\nto a file to run, otherwise the value is executed as JavaScript code.\nCurrently supported contexts include:<\/li>\n<li><code>uzbl<\/code>\n<ul><li>Run the code in the <code>uzbl<\/code> context. This context does not (currently)\nhave access to the page's contents.<\/li>\n<\/ul><\/li>\n<li><code>clean<\/code>\n<ul><li>Create a new context just for this command. This is the most secure\noption and should be preferred where possible.<\/li>\n<\/ul><\/li>\n<li><code>frame<\/code>\n<ul><li>Run the code in the current frame's context.<\/li>\n<\/ul><\/li>\n<li><code>page<\/code>\n<ul><li>Run the code in the current page's context. Be careful putting code\ninto this context since anything in here may be hijacked by a malicious\nwebpage. The context is cleared on each page load.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><p><code>spawn &lt;COMMAND&gt; [ARGUMENT...]<\/code> (WILL CHANGE)<\/p>\n<ul><li><p>Spawn a command on the system with the given arguments. This ignore the\nstdout of the command. Will become <code>spawn async<\/code> in the future. The first\nargument command may use the following format to support fallback\ndirectories:<\/p>\n<p>dir1:dir2:dir3:subpath<\/p><\/li>\n<\/ul><p>This will try <code>@{dir1}@{subpath}<\/code> followed by @{dir2}@{subpath}<code>and so on.\nThe first file which exists is used. If none of them exist, it is used as\nthe command verbatim and the<\/code>PATH` environment variable is used.<\/p><\/li>\n<li><code>spawn_sync &lt;COMMAND&gt; [ARGUMENT...]<\/code> (DEPRECATED)\n<ul><li>Spawn a command on the system synchronously with the given arguments and\nreturn its stdout. Will become <code>spawn async<\/code> in the future.<\/li>\n<\/ul><\/li>\n<li><code>spawn_sync_exec &lt;COMMAND&gt; [ARGUMENT...]<\/code> (DEPRECATED)\n<ul><li>Spawn a command synchronously and then execute the output as <code>uzbl<\/code>\ncommands. This will disappear in the future and the command should\ncommunicate to <code>uzbl<\/code> over either the FIFO or the socket.<\/li>\n<\/ul><\/li>\n<li><code>spawn_sh &lt;COMMAND&gt; [ARGUMENT...]<\/code> (DEPRECATED)\n<ul><li>Spawn a command using the default shell. This is deprecated for <code>spawn\n@shell_cmd ...<\/code>.<\/li>\n<\/ul><\/li>\n<li><code>spawn_sh_sync &lt;COMMAND&gt; [ARGUMENT...]<\/code> (DEPRECATED)\n<ul><li>Spawn a command using the default shell. This is deprecated for <code>spawn_sync\n@shell_cmd ...<\/code>.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Uzbl<\/h4><ul><li><code>chain &lt;COMMAND...&gt;<\/code>\n<ul><li>Execute a chain of commands. This is useful for commands which should not\nhave any interruption between them.<\/li>\n<\/ul><\/li>\n<li><code>include {PATH}<\/code>\n<ul><li>Execute a file as a list of uzbl commands.<\/li>\n<\/ul><\/li>\n<li><code>exit<\/code>\n<ul><li>Closes <code>uzbl<\/code>.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Variable<\/h4><ul><li><code>set &lt;NAME&gt; {VALUE}<\/code>\n<ul><li>Set a variable to the given value. Unsetting a variable is not possible\n(currently). Set it to the empty string (behavior is the same).<\/li>\n<\/ul><\/li>\n<li><code>toggle &lt;VARIABLE&gt; [OPTION...]<\/code>\n<ul><li>Toggles a variable. If any options are given, a value is chosen from the\nlist, the option following the current value is used, defaulting to the\nfirst value if the current value is not found. Without any options,\nnon-strings are toggled between 0 and 1 (if it is neither of these values,\n0 is set). User values are set to the empty string unless they are <code>0<\/code> or\n<code>1<\/code> in which case the other is set. If the variable does not exist, it is\nset to <code>1<\/code>.<\/li>\n<\/ul><\/li>\n<li><code>print {STRING}<\/code>\n<ul><li>Performs variable expansion on the string and returns the value.<\/li>\n<\/ul><\/li>\n<li><code>dump_config<\/code>\n<ul><li>Dumps the current config (which may have been changed at runtime) to\nstdout. Uses a format which can be piped into <code>uzbl<\/code> again or saved as a\nconfig file.<\/li>\n<\/ul><\/li>\n<li><code>dump_config_as_events<\/code>\n<ul><li>Dump the current config as a series of <code>VARIABLE_SET<\/code> events, which can be\nhandled by an event manager.<\/li>\n<\/ul><\/li>\n<li><code>event &lt;NAME&gt; [ARGUMENTS...]<\/code>\n<ul><li>Send a custom event.<\/li>\n<\/ul><\/li>\n<li><code>request &lt;NAME&gt; &lt;COOKIE&gt; [ARGUMENTS...]<\/code>\n<ul><li>Send a synchronous request and returns the result of the request. This is\nmeant to be used for synchronous communication between the event manager\nand <code>uzbl<\/code> since <code>spawn_sync<\/code> is not usable to talk to the event manager.<\/li>\n<\/ul><\/li>\n<li><code>choose &lt;NAME&gt; &lt;COOKIE&gt; [ARGUMENTS...]<\/code>\n<ul><li>Send a synchronous choose and returns the result of the request. This is\nmeant to implement the <code>file_chooser_handler<\/code> and <code>color_chooser_handler<\/code>\nwhere a result should be waited upon indefinitely. <strong>This will hang uzbl if\nit is not replied to.<\/strong><\/li>\n<\/ul><\/li>\n<\/ul><h3>VARIABLES AND CONSTANTS<\/h3><p><code>Uzbl<\/code> has a lot of internal variables and constants. You can get the values\n(see the <code>print<\/code> command), and for variables you can also change the value at\nruntime. Some of the values can be passed at start up through command line\narguments, others need to be set by using commands (e.g., in the configuration\nfile).<\/p><ul><li>Some variables have callback functions which will get called after setting\nthe variable to perform some additional logic (see below).<\/li>\n<li>Besides the builtin variables you can also define your own ones and use them\nin the exact same way as the builtin ones.<\/li>\n<\/ul><p>Integer variables may be negative, boolean values should be either <code>0<\/code> or <code>1<\/code>.\nCommands are variables intended to be executed as uzbl commands. All user\nvariables are treated as strings.<\/p><h4>Uzbl<\/h4><ul><li><code>verbose<\/code> (integer) (default: 0)\n<ul><li>Controls the verbosity of <code>uzbl<\/code> as it runs.<\/li>\n<\/ul><\/li>\n<li><code>frozen<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, all network requests and navigation are blocked. This is\nuseful for blocking pages from redirecting or loading any content after the\ncore content is loaded. Network connections which have already started will\nstill complete.<\/li>\n<\/ul><\/li>\n<li><code>print_events<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, events will be printed to stdout.<\/li>\n<\/ul><\/li>\n<li><code>handle_multi_button<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, <code>uzbl<\/code> will intercept all double and triple clicks and the\npage will not see them.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Communication<\/h4><ul><li><code>fifo_dir<\/code> (string) (no default)\n<ul><li>Sets the directory for the FIFO. If set previously, the old FIFO is\nremoved.<\/li>\n<\/ul><\/li>\n<li><code>socket_dir<\/code> (string) (no default)\n<ul><li>Sets the directory for the socket. If set previously, the old socket is\nremoved.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Handler<\/h4><ul><li><code>navigation_handler<\/code> (command) (no default)\n<ul><li>The command to use when determining what to do when navigating to a new\nURI. It is passed the URI as an extra argument. If the command returns a\nstring with the first line containing only the word <code>USED<\/code>, the navigation\nis ignored.<\/li>\n<\/ul><\/li>\n<li><code>request_handler<\/code> (command) (no default)\n<ul><li>The command to use when a new network request is about to be initiated. The\nURI is passed as an argument. If the command returns a non-empty string,\nthe first line of the result is used as the new URI. To cancel a request,\nuse the URI <code>about:blank<\/code>.<\/li>\n<li>NOTE: Do <em>not<\/em> use <code>request<\/code> in WebKit1 as this is called synchronously and\nwill just pause <code>uzbl-core<\/code> until the <code>request<\/code> timeout occurs.<\/li>\n<\/ul><\/li>\n<li><code>download_handler<\/code> (command) (no default) (synchronous in WebKit1)\n<ul><li>The command to use when determining where to save a downloaded file. It is\npassed the URI, suggested filename, content type, and total size as\narguments. If a destination is known, it is passed as well. The result is\nused as the final destination. If it is empty, the download is cancelled.<\/li>\n<li>NOTE: Do <em>not<\/em> use <code>request<\/code> in WebKit1 as this is called synchronously and\nwill just pause <code>uzbl-core<\/code> until the <code>request<\/code> timeout occurs.<\/li>\n<\/ul><\/li>\n<li><code>mime_handler<\/code> (command) (no default) (WebKit1 only)\n<ul><li>The command to use when determining what to do with content based on its\nmime type. It is passed the mime type and disposition as arguments.<\/li>\n<\/ul><\/li>\n<li><code>authentication_handler<\/code> (command) (no default) (WebKit1 only)\n<ul><li>The command to use when requesting authentication.<\/li>\n<\/ul><\/li>\n<li><code>permission_handler<\/code> (command) (no default)\n<ul><li>The command to use when WebKit requesting permission for an action.<\/li>\n<\/ul><\/li>\n<li><code>file_chooser_handler<\/code> (command) (no default) (WebKit &gt;= 1.9.4)\n<ul><li>The command to use when a web page requests a file. By default, the GTK\nwidget file dialog.<\/li>\n<\/ul><\/li>\n<li><code>shell_cmd<\/code> (string) (default: <code>sh -c<\/code>)\n<ul><li>The command to use as a shell. This is used in the <code>spawn_sh<\/code> and <code>sync_sh<\/code>\ncommands as well as <code>@()@<\/code> expansion.<\/li>\n<\/ul><\/li>\n<li><code>enable_builtin_auth<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, WebKit will handle HTTP authentication dialogs.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Window<\/h4><ul><li><code>icon<\/code> (string) (no default)\n<ul><li>The path to an image to use as an icon for <code>uzbl<\/code>. Overrides <code>icon_name<\/code>\nwhen set.<\/li>\n<\/ul><\/li>\n<li><code>icon_name<\/code> (string) (no default)\n<ul><li>The name of an icon use as an icon for <code>uzbl<\/code>. Overrides <code>icon<\/code> when set.<\/li>\n<\/ul><\/li>\n<li><code>window_role<\/code> (string) (no default)\n<ul><li>Sets the role of the window. This is used to hint to the window manager how\nto handle the window.<\/li>\n<\/ul><\/li>\n<li><code>auto_resize_window<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, WebKit will try to honor JavaScript moving the window. May\nstill be blocked by window manager policy.<\/li>\n<\/ul><\/li>\n<\/ul><h4>UI<\/h4><ul><li><code>show_status<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, the status bar is shown.<\/li>\n<\/ul><\/li>\n<li><code>status_top<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, the status bar is placed above the page, otherwise it is\nplaced below.<\/li>\n<\/ul><\/li>\n<li><code>status_format<\/code> (string) (default: <code>&lt;b&gt;@[@TITLE]@&lt;\/b&gt; - @[@uri]@ - &lt;span foreground=\"#bbb\"&gt;@NAME&lt;\/span&gt;<\/code>)\n<ul><li>The format string of the content in the status bar.<\/li>\n<\/ul><\/li>\n<li><code>status_format_right<\/code> (string) (no default)\n<ul><li>The format string of the content in right side of the status bar. The right\nside is elided if the left side overflows.<\/li>\n<\/ul><\/li>\n<li><code>status_background<\/code> (string) (no default)\n<ul><li>The background color of the status bar. Most common formats supported.<\/li>\n<\/ul><\/li>\n<li><code>title_format_long<\/code> (string (default: <code>@keycmd @TITLE - Uzbl browser &lt;@NAME&gt; &gt; @SELECTED_URI<\/code>)\n<ul><li>`The title to use when no status bar is shown.<\/li>\n<\/ul><\/li>\n<li><code>title_format_short<\/code> (string (default: <code>@TITLE - Uzbl browser &lt;@NAME&gt;<\/code>)\n<ul><li>`The title to use when the status bar is shown.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Customization<\/h4><ul><li><code>default_context_menu<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, display the default context menu, ignoring any custom menu\nitems.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Printing<\/h4><ul><li><code>print_backgrounds<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, print backgrounds when printing pages.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Network<\/h4><ul><li><code>proxy_url<\/code> (string) (no default) (WebKit1 only)\n<ul><li>If set, the value is used as a network proxy.<\/li>\n<\/ul><\/li>\n<li><code>max_conns<\/code> (integer) (default: 100) (WebKit1 only)\n<ul><li>The maximum number of connections to initialize total.<\/li>\n<\/ul><\/li>\n<li><code>max_conns_host<\/code> (integer) (default: 6) (WebKit1 only)\n<ul><li>The maximum number of connections to use per host.<\/li>\n<\/ul><\/li>\n<li><code>http_debug<\/code> (enumeration) (default: <code>none<\/code>) (WebKit1 only)\n<ul><li>Debugging level for HTTP connections. Acceptable values include:<\/li>\n<li><code>none<\/code><\/li>\n<li><code>minimal<\/code><\/li>\n<li><code>headers<\/code><\/li>\n<li><code>body<\/code><\/li>\n<\/ul><\/li>\n<li><code>ssl_ca_file<\/code> (string) (no default) (WebKit1 only)\n<ul><li>The path to the CA certificate file in PEM format to use for trusting SSL\ncertificates. By default, the system store is used.<\/li>\n<\/ul><\/li>\n<li><code>ssl_policy<\/code> (enumeration) (default: <code>fail<\/code>)\n<ul><li>The policy to use for handling SSL failures. Acceptable values include:<\/li>\n<li><code>ignore<\/code><\/li>\n<li><code>fail<\/code><\/li>\n<\/ul><\/li>\n<li><code>cache_model<\/code> (enumeration) (default: <code>unknown<\/code>)\n<ul><li>The model to use for caching content. Acceptable values include:<\/li>\n<li><code>document_viewer<\/code>\n<ul><li>Disables all caching.<\/li>\n<\/ul><\/li>\n<li><code>web_browser<\/code>\n<ul><li>Caches heavily to attempt to minimize network usage.<\/li>\n<\/ul><\/li>\n<li><code>document_browser<\/code>\n<ul><li>Caches moderately. This is optimized for navigation of local resources.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h4>Security<\/h4><ul><li><code>enable_private<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, enables WebKit's private browsing mode. Also sets the\n<code>UZBL_PRIVATE<\/code> environment variable for external plugins. DNS prefetching\nis separate from this; see the <code>enable_dns_prefetch<\/code> variable. It is\ncurrently experimental in WebKitGtk itself.<\/li>\n<\/ul><\/li>\n<li><code>permissive<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, permission will be granted by default (if <code>enable_private<\/code> is\n0) if <code>permission_handler<\/code> is either not set or returns an invalid value.<\/li>\n<\/ul><\/li>\n<li><code>enable_universal_file_access<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, content accessed via the <code>file<\/code> scheme will be allowed to\naccess all other content.<\/li>\n<\/ul><\/li>\n<li><code>enable_cross_file_access<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, content accessed via the <code>file<\/code> scheme will be allowed to\naccess other <code>file<\/code> content.<\/li>\n<\/ul><\/li>\n<li><code>enable_hyperlink_auditing<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, <code>&lt;a ping&gt;<\/code> will be allowed in WebKit.<\/li>\n<\/ul><\/li>\n<li><code>cookie_policy<\/code> (enumeration) (default: <code>always<\/code>)\n<ul><li>Sets the cookie policy for WebKit. Acceptable values include:<\/li>\n<li><code>always<\/code><\/li>\n<li><code>never<\/code><\/li>\n<li><code>first_party<\/code>\n<ul><li>Blocks third-party cookies.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>enable_dns_prefetch<\/code> (boolean) (default: 1) (WebKit &gt;= 1.3.13)\n<ul><li>If non-zero, WebKit will prefetch domain names while browsing.<\/li>\n<\/ul><\/li>\n<li><code>display_insecure_content<\/code> (boolean) (default: 1) (WebKit1 &gt;= 1.11.2)\n<ul><li>If non-zero, HTTPS content will be allowed to display content served over\nHTTP.<\/li>\n<\/ul><\/li>\n<li><code>run_insecure_content<\/code> (boolean) (default: 1) (WebKit1 &gt;= 1.11.2)\n<ul><li>If non-zero, HTTPS content will be allowed to run content served over HTTP.<\/li>\n<\/ul><\/li>\n<li><code>maintain_history<\/code> (boolean) (default: 1) (WebKit1 only)\n<ul><li>If non-zero, WebKit will maintain a back\/forward list.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Inspector<\/h4><ul><li><code>profile_js<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, the inspector will profile running JavaScript. Profiles are\navailable as the <code>Console.profiles<\/code> variable in JavaScript.<\/li>\n<\/ul><\/li>\n<li><code>profile_timeline<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, the inspector will profile network activity.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Page<\/h4><ul><li><code>forward_keys<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, <code>uzbl<\/code> will pass keypresses to WebKit.<\/li>\n<\/ul><\/li>\n<li><code>useragent<\/code> (string) (no default)\n<ul><li>The string to send as the <code>User-Agent<\/code> HTTP header. Setting this variable\ndoes not update the useragent in JavaScript as seen by the\n<code>navigator.userAgent<\/code> variable.<\/li>\n<\/ul><\/li>\n<li><code>accept_languages<\/code> (string) (no default)\n<ul><li>The list of languages to send with the <code>Accept-Language<\/code> HTTP header. In\nWebKit1, if set to <code>auto<\/code>, the list will be constructed from the system.<\/li>\n<\/ul><\/li>\n<li><code>zoom_level<\/code> (double) (default: 1.0)\n<ul><li>The current zoom level of the page.<\/li>\n<\/ul><\/li>\n<li><code>zoom_step<\/code> (double) (default: 0.1)\n<ul><li>The amount to step by default with the <code>zoom in<\/code> and <code>zoom out<\/code> commands.\nMust be greater than 0.<\/li>\n<\/ul><\/li>\n<li><code>zoom_text_only<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, only text will be zoomed on a page.<\/li>\n<\/ul><\/li>\n<li><code>caret_browsing<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, pages may be navigated using the arrows to move a cursor\n(similar to a word processor).<\/li>\n<\/ul><\/li>\n<li><code>enable_frame_flattening<\/code> (boolean) (default: 0) (WebKit &gt;= 1.3.5)\n<ul><li>If non-zero, frames will be expanded and collapsed into a single scrollable\narea.<\/li>\n<\/ul><\/li>\n<li><code>enable_smooth_scrolling<\/code> (boolean) (default: 0) (WebKit &gt;= 1.9.0)\n<ul><li>If non-zero, scrolling the page will be smoothed.<\/li>\n<\/ul><\/li>\n<li><code>page_view_mode<\/code> (enumeration) (default: <code>web<\/code>)\n<ul><li>How to render content on a page. Acceptable values include:<\/li>\n<li><code>web<\/code>\n<ul><li>Render content as a web page.<\/li>\n<\/ul><\/li>\n<li><code>source<\/code>\n<ul><li>Display the source of a page as plain text.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><code>transparent<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, the web page will have a transparent background rather than a\nsolid color.<\/li>\n<\/ul><\/li>\n<li><code>window_view_mode<\/code> (enumeration) (default: <code>windowed<\/code>) (WebKit1 &gt;= 1.3.4)\n<ul><li>Sets the value web pages will see for how the content is presented to the\nuser. Acceptable values include:<\/li>\n<li><code>windowed<\/code><\/li>\n<li><code>floating<\/code><\/li>\n<li><code>fullscreen<\/code><\/li>\n<li><code>maximized<\/code><\/li>\n<li><code>minimized<\/code><\/li>\n<\/ul><\/li>\n<li><code>enable_fullscreen<\/code> (boolean) (default: 0) (WebKit &gt;= 1.3.4)\n<ul><li>If non-zero, Mozilla-style JavaScript APIs will be enabled for JavaScript.<\/li>\n<\/ul><\/li>\n<li><code>editable<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, web pages will be editable similar to a WYSIWYG editor.<\/li>\n<\/ul><\/li>\n<\/ul><h4>JavaScript<\/h4><ul><li><code>enable_scripts<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, JavaScript will be allowed to execute on web pages.<\/li>\n<\/ul><\/li>\n<li><code>javascript_windows<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, JavaScript in a page will be allowed to open windows without\nuser intervention.<\/li>\n<\/ul><\/li>\n<li><code>javascript_dom_paste<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, JavaScript in a page will be able to paste content from the\nsystem clipboard.<\/li>\n<\/ul><\/li>\n<li><code>javascript_clipboard<\/code> (boolean) (default: 0) (WebKit &gt;= 1.3.0)\n<ul><li>If non-zero, JavaScript in a page will be able to access content in the\nsystem clipboard directly.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Image<\/h4><ul><li><code>autoload_images<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, images will automatically be loaded.<\/li>\n<\/ul><\/li>\n<li><code>autoshrink_images<\/code> (boolean) (default: 1) (WebKit1 only)\n<ul><li>If non-zero, stand-alone images will be resized to fit within the window.<\/li>\n<\/ul><\/li>\n<li><code>use_image_orientation<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, image orientation tags will be used to reorient images once\nloaded.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Spell checking<\/h4><ul><li><code>enable_spellcheck<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, spell checking will be enabled in text entries.<\/li>\n<\/ul><\/li>\n<li><code>spellcheck_languages<\/code> (string) (no default)\n<ul><li>A comma-separated list of languages to use for spell checking. Must be set\nfor <code>enable_spellcheck<\/code> to actually mean anything.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Form<\/h4><ul><li><code>resizable_text_areas<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, text area form elements will be resizable with a drag handle.<\/li>\n<\/ul><\/li>\n<li><code>enable_spatial_navigation<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, the arrow keys will navigate between form elements.<\/li>\n<\/ul><\/li>\n<li><code>editing_behavior<\/code> (enumeration) (default: <code>unix<\/code>) (WebKit1 only)\n<ul><li>Sets the behavior of text area elements to behave closer to native widgets\non different platforms. Acceptable values include:<\/li>\n<li><code>mac<\/code><\/li>\n<li><code>unix<\/code><\/li>\n<li><code>windows<\/code><\/li>\n<\/ul><\/li>\n<li><code>enable_tab_cycle<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, the tab key will navigate between form elements, otherwise it\nwill always be interpreted as an insertion of a tab character.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Text<\/h4><ul><li><code>default_encoding<\/code> (string) (default: <code>iso-8859-1<\/code>)\n<ul><li>The encoding to assume pages use if not indicated in HTTP headers.<\/li>\n<\/ul><\/li>\n<li><code>custom_encoding<\/code> (string) (no default)\n<ul><li>If non-empty, the value is used to override any other encoding used for the\npage. Causes the page to be reloaded. Set to the empty string to clear any\ncustom encoding.<\/li>\n<\/ul><\/li>\n<li><code>enforce_96_dpi<\/code> (boolean) (default: 0) (WebKit1 only)\n<ul><li>If non-zero, the display will be assumed to be 96-dpi.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Font<\/h4><ul><li><code>default_font_family<\/code> (string) (default: <code>sans-serif<\/code>)\n<ul><li>The font family to use by default for all text.<\/li>\n<\/ul><\/li>\n<li><code>monospace_font_family<\/code> (string) (default: <code>monospace<\/code>)\n<ul><li>The font family to use by default for monospace text.<\/li>\n<\/ul><\/li>\n<li><code>sans_serif_font_family<\/code> (string) (default: <code>sans-serif<\/code>)\n<ul><li>The font family to use by default for sans serif text.<\/li>\n<\/ul><\/li>\n<li><code>serif_font_family<\/code> (string) (default: <code>serif<\/code>)\n<ul><li>The font family to use by default for serif text.<\/li>\n<\/ul><\/li>\n<li><code>cursive_font_family<\/code> (string) (default: <code>serif<\/code>)\n<ul><li>The font family to use by default for any cursive text.<\/li>\n<\/ul><\/li>\n<li><code>fantasy_font_family<\/code> (string) (default: <code>serif<\/code>)\n<ul><li>The font family to use by default for any fantasy text.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Font size<\/h4><p>Note: font sizes are in points.<\/p><ul><li><code>minimum_font_size<\/code> (integer) (default: 5)\n<ul><li>The minimum font size to use for text.<\/li>\n<\/ul><\/li>\n<li><code>minimum_logical_font_size<\/code> (integer) (default: 5)\n<ul><li>The minimum logical font size to use for text. This is the size after any\ntransforms or \"smaller\" CSS specifications.<\/li>\n<\/ul><\/li>\n<li><code>font_size<\/code> (integer) (default: 12)\n<ul><li>The default font size for text.<\/li>\n<\/ul><\/li>\n<li><code>monospace_size<\/code> (integer) (default: 10)\n<ul><li>The default font size for monospace text.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Features<\/h4><ul><li><code>enable_plugins<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, plugins will be enabled on the page.<\/li>\n<\/ul><\/li>\n<li><code>enable_java_applet<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, Java will be enabled for the page. This includes the <code>applet<\/code>\nHTML tag which is not controlled by the <code>enable_plugins<\/code> variable.<\/li>\n<\/ul><\/li>\n<li><code>enable_webgl<\/code> (boolean) (default: 0) (WebKit &gt;= 1.3.14)\n<ul><li>If non-zero, WebGL will be enabled.<\/li>\n<\/ul><\/li>\n<li><code>enable_webaudio<\/code> (boolean) (default: 0) (WebKit &gt;= 1.7.5)\n<ul><li>If non-zero, the WebAudio API will be enabled.<\/li>\n<\/ul><\/li>\n<li><code>enable_3d_acceleration<\/code> (boolean) (default: 0) (WebKit1 &gt;= 1.7.90)\n<ul><li>If non-zero, the GPU will be used to render animations and 3D CSS\ntransforms.<\/li>\n<\/ul><\/li>\n<li><code>enable_inline_media<\/code> (boolean) (default: 1) (WebKit &gt;= 1.9.3)\n<ul><li>If non-zero, media playback within a page will be enabled, otherwise, only\nfullscreen playback is allowed.<\/li>\n<\/ul><\/li>\n<li><code>require_click_to_play<\/code> (boolean) (default: 0) (WebKit &gt;= 1.9.3)\n<ul><li>If non-zero, media may automatically play on a page, otherwise a user\ninteraction is required to even load the media.<\/li>\n<\/ul><\/li>\n<li><code>enable_css_shaders<\/code> (boolean) (default: 0) (WebKit1 &gt;= 1.11.1 and &lt; 2.3.5)\n<ul><li>If non-zero, CSS shaders will be enabled.<\/li>\n<\/ul><\/li>\n<li><code>enable_media_stream<\/code> (boolean) (default: 0) (WebKit1 &gt;= 1.11.1)\n<ul><li>If non-zero, the Media Stream API will be enabled.<\/li>\n<\/ul><\/li>\n<li><code>enable_media_source<\/code> (boolean) (default: 0) (WebKit &gt;= 2.3.3)\n<ul><li>If non-zero, the MediaSource API will be enabled.<\/li>\n<\/ul><\/li>\n<\/ul><h4>HTML5 Database<\/h4><ul><li><code>enable_database<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, the HTML5 client-side SQL API for JavaScript is enabled.<\/li>\n<\/ul><\/li>\n<li><code>enable_local_storage<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, the HTML5 local storage API is enabled.<\/li>\n<\/ul><\/li>\n<li><code>enable_pagecache<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, WebKit will \"freeze\" pages when navigating to optimize\nback\/forward navigation. See\n<a href=\"http:\/\/webkit.org\/blog\/427\/webkit-page-cache-i-the-basics\/\">this article<\/a>\nfor more information.<\/li>\n<\/ul><\/li>\n<li><code>enable_offline_app_cache<\/code> (boolean) (default: 1)\n<ul><li>If non-zero, HTML5 web application cache support will be enabled. This\nallows web applications to run locally without a network connection.<\/li>\n<\/ul><\/li>\n<li><code>app_cache_size<\/code> (size) (default: [no quota]) (WebKit1 &gt;= 1.3.13)\n<ul><li>The maximum size (in bytes) of the local application cache.<\/li>\n<\/ul><\/li>\n<li><code>web_database_directory<\/code> (string) (default: <code>$XDG_DATA_HOME\/webkit\/databases<\/code>) (WebKit1 only)\n<ul><li>Where to store web databases.<\/li>\n<\/ul><\/li>\n<li><code>web_database_quota<\/code> (size) (default: 5242880) (WebKit1 only)\n<ul><li>The maximum size (in bytes) of web databases. (Note: It is unclear if this\nis a total quota or per-database quota.)<\/li>\n<\/ul><\/li>\n<li><code>local_storage_path<\/code> (string) (default: <code>$XDG_DATA_HOME\/webkit\/databases<\/code>) (WebKit1 &gt;= 1.5.2)\n<ul><li>Where to store HTML5 <code>localStorage<\/code> data.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Hacks<\/h4><ul><li><code>enable_site_workarounds<\/code> (boolean) (default: 0)\n<ul><li>If non-zero, WebKit will use site-specific rules to workaround known broken\nbehavior.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Constants<\/h4><p>Constants may not be assigned to. When dumping variables using <code>dump_config<\/code>,\nthey are written as comments.<\/p><ul><li><code>inspected_uri<\/code> (string) (WebKit &gt;= 1.3.17)\n<ul><li>The URI which the inspector is currently focusing on.<\/li>\n<\/ul><\/li>\n<li><code>current_encoding<\/code> (string) (WebKit1 only)\n<ul><li>The encoding of the current page.<\/li>\n<\/ul><\/li>\n<li><code>geometry<\/code> (string)\n<ul><li>A string giving the current window geometry of <code>uzbl<\/code> in\n<code>&lt;width&gt;x&lt;height&gt;+&lt;xoffset&gt;+&lt;yoffset&gt;<\/code> format.<\/li>\n<\/ul><\/li>\n<li><code>plugin_list<\/code> (string) (WebKit1 &gt;= 1.3.8)\n<ul><li>A JSON-formatted list describing loaded plugins.<\/li>\n<\/ul><\/li>\n<li><code>is_online<\/code> (boolean)\n<ul><li>If non-zero, a network is available (not necessarily the Internet).<\/li>\n<\/ul><\/li>\n<li><code>is_playing_audio<\/code> (boolean)\n<ul><li>If non-zero, audio is playing.<\/li>\n<\/ul><\/li>\n<li><code>uri<\/code> (string)\n<ul><li>The current top-level URI of the view.<\/li>\n<\/ul><\/li>\n<li><code>embedded<\/code> (boolean)\n<ul><li>If non-zero, <code>uzbl<\/code> is embedded in another window, otherwise it is\nstand-alone.<\/li>\n<\/ul><\/li>\n<li><code>WEBKIT_MAJOR<\/code> (integer)\n<ul><li>The major version of WebKit at runtime.<\/li>\n<\/ul><\/li>\n<li><code>WEBKIT_MINOR<\/code> (integer)\n<ul><li>The minor version of WebKit at runtime.<\/li>\n<\/ul><\/li>\n<li><code>WEBKIT_MICRO<\/code> (integer)\n<ul><li>The micro version of WebKit at runtime.<\/li>\n<\/ul><\/li>\n<li><code>WEBKIT_UA_MAJOR<\/code> (integer)\n<ul><li>The major useragent version of WebKit at compile time.<\/li>\n<\/ul><\/li>\n<li><code>WEBKIT_UA_MINOR<\/code> (integer)\n<ul><li>The minor useragent version of WebKit at compile time.<\/li>\n<\/ul><\/li>\n<li><code>HAS_WEBKIT2<\/code> (boolean)\n<ul><li>If non-zero, <code>uzbl<\/code> is using WebKit2.<\/li>\n<\/ul><\/li>\n<li><code>ARCH_UZBL<\/code> (string)\n<ul><li>The architecture <code>uzbl<\/code> is compiled for.<\/li>\n<\/ul><\/li>\n<li><code>COMMIT<\/code> (string)\n<ul><li>The commit string of <code>uzbl<\/code> at compiled time.<\/li>\n<\/ul><\/li>\n<li><code>TITLE<\/code> (string)\n<ul><li>The title of the current page.<\/li>\n<\/ul><\/li>\n<li><code>SELECTED_URI<\/code> (string)\n<ul><li>The URI of the link is hovering over; empty if there is no link.<\/li>\n<\/ul><\/li>\n<li><code>NAME<\/code> (string)\n<ul><li>The name of the <code>uzbl<\/code> instance (defaults to the pid).<\/li>\n<\/ul><\/li>\n<li><code>PID<\/code> (integer)\n<ul><li>The process ID of <code>uzbl<\/code>.<\/li>\n<\/ul><\/li>\n<li><code>_<\/code> (string)\n<ul><li>The last result from a command.<\/li>\n<\/ul><\/li>\n<\/ul><h3>VARIABLE EXPANSION<\/h3><p>Variable expansion works similarly to shell interpreters (sh, bash, etc.). This\nmeans you can construct strings with uzbl variables in them and have uzbl\nreplace the variable name with its contents.<\/p><p>In order to let uzbl know what to expand you'll need to prepend @ to the\nvariable name:<\/p><pre>print The variable \\@show_status contains @show_status\n<\/pre><p>The above example demonstrates two things:<\/p><ul><li><code>\\<\/code> is treated as escape character and will use the character immediately\nfollowing it literally this means <code>\\@show_status<\/code> will not expand to the\nvariable content but be rather printed as <code>@show_status<\/code><\/li>\n<li>prepending the variable with <code>@<\/code> will expand to its contents<\/li>\n<\/ul><p>As in the shell you can use <code>@{uzbl_var}<\/code> to denote the beginning\/end of the\nvariable name in cases where it is not obvious what belongs to the name and\nwhat not, e.g., <code>print @{show_status}foobar<\/code><\/p><p>There are multiple constructs for more complex expansion.<\/p><h4>Shell expansion<\/h4><ul><li><code>@(+command arg1 arg2)@<\/code>\n<ul><li>This is equivalent to running <code>spawn_sync command arg1 arg2<\/code> and is\nreplaced by the first line of its stdout.<\/li>\n<\/ul><\/li>\n<li><code>@(command arg1 arg2)@<\/code>\n<ul><li>This is equivalent to running <code>spawn_sh_sync command arg1 arg2<\/code> and is\nreplaced by the first line of its stdout.<\/li>\n<\/ul><\/li>\n<\/ul><h4>Uzbl command expansion<\/h4><ul><li><code>@\/+file\/@<\/code>\n<ul><li>This is equivalent to running <code>include file<\/code> and is replaced by its result.<\/li>\n<\/ul><\/li>\n<li><code>@\/command arg1 arg2\/@<\/code>\n<ul><li>This is equivalent to running <code>command arg1 arg2<\/code> and is replaced by its\nresult.<\/li>\n<\/ul><\/li>\n<\/ul><h4>JavaScript expansion<\/h4><ul><li><code>@*+javascript_file*@<\/code>\n<ul><li>This is equivalent to running <code>js uzbl file javascript_file<\/code> and is\nreplaced by its result.<\/li>\n<\/ul><\/li>\n<li><code>@*javascript*@<\/code>\n<ul><li>This is equivalent to running <code>js uzbl string javascript<\/code> and is replaced\nby its result.<\/li>\n<\/ul><\/li>\n<li><code>@&lt;+javascript_file&gt;@<\/code>\n<ul><li>This is equivalent to running <code>js page file javascript_file<\/code> and is\nreplaced by its result.<\/li>\n<\/ul><\/li>\n<li><code>@&lt;javascript&gt;@<\/code>\n<ul><li>This is equivalent to running <code>js page string javascript<\/code> and is replaced\nby its result.<\/li>\n<\/ul><\/li>\n<li><code>@-+javascript_file-@<\/code>\n<ul><li>This is equivalent to running <code>js clean file javascript_file<\/code> and is\nreplaced by its result.<\/li>\n<\/ul><\/li>\n<li><code>@-javascript-@<\/code>\n<ul><li>This is equivalent to running <code>js clean string javascript<\/code> and is replaced\nby its result.<\/li>\n<\/ul><\/li>\n<\/ul><h4>XML escaping<\/h4><ul><li><code>@[string]@<\/code>\n<ul><li>This escapes <code>string<\/code> for XML entities. This is most useful when using\npage-defined strings in the <code>status_bar<\/code> variable.<\/li>\n<\/ul><\/li>\n<\/ul><h3>JavaScript API<\/h3><p>Uzbl has a JavaScript API available via the <code>js uzbl<\/code> command. There is a\nglobal <code>uzbl<\/code> object can control <code>uzbl<\/code> through two properties: <code>variables<\/code> and\n<code>commands<\/code>. For example, to set the <code>verbose<\/code> variable:<\/p><pre>uzbl.variables.verbose = 1\n<\/pre><p>To call the <code>back<\/code> command:<\/p><pre>uzbl.commands.back()\n<\/pre><p>All arguments to commands be converted to strings, so <code>{}<\/code> will appear to\ncommands as <code>[object Object]<\/code>, not JavaScript objects.<\/p><h4>Accessing the web page<\/h4><p>Currently, access to the webpage is not available through the <code>uzbl<\/code> context.\nThis is currently impossible with WebKit2 (since even <code>uzbl<\/code> does not have\naccess to the page's JavaScript context) and crashes WebKit1.<\/p><h3>TITLE AND STATUS BAR EVALUATION<\/h3><p>The contents of the status bar can be customized by setting the <code>status_format<\/code>\nvariable. The contents of the window title can be customized by setting the\n<code>title_format_short<\/code> variable (which is used when the status bar is displayed)\nand the <code>title_format_long<\/code> variable (which is used when the status bar is not\ndisplayed). Their values can be set using the expansion and substitution\ntechniques described above.<\/p><p>These variables are expanded in multiple stages; once when the variable is set,\nand again every time that the status bar or window title are updated. Expansions\nthat should be evaluated on every update need to be escaped:<\/p><pre>set title_format_short = @(date)@\n# this expansion will be evaluated when the variable is set.\n# the title will stay constant with the date that the variable was set.\nset title_format_short = \\@(date)\\@\n# this expansion will be evaluated when the window title is updated.\n# the date in the title will change when you change pages, for example.\nset title_format_short = \\\\\\@(date)\\\\\\@\n# the title will stay constant as a literal \"@(date)@\"\n<\/pre><p>The <code>status_format<\/code> and <code>status_format_right<\/code> variables can contain\n<a href=\"http:\/\/library.gnome.org\/devel\/pango\/stable\/PangoMarkupFormat.html\">Pango<\/a>\nmarkup . In these variables, expansions that might produce the characters <code>&lt;<\/code>,\n<code>&amp;<\/code> or <code>&gt;<\/code> should be wrapped in <code>@[]@<\/code> substitutions so that they don't\ninterfere with the status bar's markup; see the sample config for examples.<\/p><h3>EXTERNAL SCRIPTS<\/h3><p>You can use external scripts with Uzbl the following ways:<\/p><ul><li>Let <code>uzbl<\/code> call them. These scripts are called \"handlers\" in the <code>uzbl<\/code>\nconfig. These are typically used for policy requests so that behavior can be\ncontrolled with external logic.<\/li>\n<li>Call them yourself from inside <code>uzbl<\/code>. You can bind keys with this. Examples:\nadd new bookmark, load new URL.<\/li>\n<li>You could also use <code>xbindkeys<\/code> or your WM config to trigger scripts if <code>uzbl<\/code>\ndoes not have focus.<\/li>\n<\/ul><p>Scripts called by <code>uzbl<\/code> (the commands starting with <code>spawn<\/code>) have access to\nthe following environment variables:<\/p><ul><li><code>$UZBL_CONFIG<\/code>\n<ul><li>The configuration file loaded by this <code>uzbl<\/code> instance.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_PID<\/code>\n<ul><li>The process ID of this <code>uzbl<\/code> instance.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_XID<\/code>\n<ul><li>The X Windows ID of the process.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_FIFO<\/code>\n<ul><li>The filename of the FIFO being used, if any.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_SOCKET<\/code>\n<ul><li>The filename of the Unix socket being used, if any.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_URI<\/code>\n<ul><li>The URI of the current page.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_TITLE<\/code>\n<ul><li>The current page title.<\/li>\n<\/ul><\/li>\n<li><code>$UZBL_PRIVATE<\/code>\n<ul><li>Set if the <code>enable_private<\/code> variable is non-zero, unset otherwise.<\/li>\n<\/ul><\/li>\n<\/ul><p>Handler scripts (<code>download_handler<\/code>, <code>navigation_handler<\/code>, <code>request_handler<\/code>,\n<code>mime_handler<\/code>, <code>authentication_handler<\/code>, <code>permission_handler<\/code>,\n<code>tls_error_handler<\/code>, <code>file_chooser_handler<\/code>, and <code>color_chooser_handler<\/code>) are\ncalled with special arguments:<\/p><ul><li><p>download handler<\/p>\n<ol><li><code>url<\/code><\/li>\n<\/ol><ul><li>The URL of the item to be downloaded.\n<ol><li><code>suggested_filename<\/code><\/li>\n<\/ol><\/li>\n<li>A filename suggested by the server or based on the URL.\n<ol><li><code>content_type<\/code><\/li>\n<\/ol><\/li>\n<li>The mimetype of the file to be downloaded.\n<ol><li><code>total_size<\/code><\/li>\n<\/ol><\/li>\n<li>The size of the file to be downloaded in bytes. This may be inaccurate.\n<ol><li><code>destination_path<\/code><\/li>\n<\/ol><\/li>\n<li>This is only present if the download was started explicitly using the\n<code>download<\/code> command. If it is present, this is the path that the file\nshould be saved to. A download handler using WebKit's internal downloader\ncan just echo this path and exit when this argument is present.<\/li>\n<\/ul><\/li>\n<li><p>navigation handler<\/p>\n<ol><li><code>uri<\/code><\/li>\n<\/ol><ul><li>The URI of the page to be navigated to.\n<ol><li><code>src_frame<\/code><\/li>\n<\/ol><\/li>\n<li>The name of the frame which is targeted for the navigation.\n<ol><li><code>dest_frame<\/code><\/li>\n<\/ol><\/li>\n<li>The name of the frame which is targeted for the navigation.\n<ol><li><code>type<\/code><\/li>\n<\/ol><\/li>\n<li>The type of navigation that triggered the request. Known values include:\n<ul><li><code>link<\/code><\/li>\n<li><code>form_submission<\/code><\/li>\n<li><code>back_forward<\/code><\/li>\n<li><code>reload<\/code><\/li>\n<li><code>form_resubmission<\/code><\/li>\n<li><code>other<\/code><\/li>\n<li><code>unknown<\/code>\n<ol><li><code>button<\/code><\/li>\n<\/ol><\/li>\n<\/ul><\/li>\n<li>The mouse button used for the navigation (or 0 if a mouse was not used).\n<ol><li><code>modifiers<\/code><\/li>\n<\/ol><\/li>\n<li>The modifiers used for the navigation. This uses the same format as the\n<code>KEY_PRESS<\/code> and <code>KEY_RELEASE<\/code> events.\n<ol><li><code>is_gesture<\/code><\/li>\n<\/ol><\/li>\n<li><code>true<\/code> if the action was triggered by a gesture, <code>false<\/code> otherwise.<\/li>\n<\/ul><\/li>\n<li><p>request handler<\/p>\n<ol><li><code>uri<\/code><\/li>\n<\/ol><ul><li>The URI of the resource which is being requested.\n<ol><li><code>can_display<\/code><\/li>\n<\/ol><\/li>\n<li>One of <code>can_display<\/code>, <code>cant_display<\/code>, or <code>unknown<\/code> to indicate whether\nWebKit knows how to display the data given its MIME type (<code>unknown<\/code> means\nthat <code>uzbl<\/code> was unable to determine the answer).\n<ol><li><code>frame<\/code><\/li>\n<\/ol><\/li>\n<li>The name of the frame requesting the resource.\n<ol><li><code>redirect<\/code><\/li>\n<\/ol><\/li>\n<li>Whether the request is a redirect or not<\/li>\n<\/ul><\/li>\n<li><p>mime handler<\/p>\n<ol><li><code>mime_type<\/code><\/li>\n<\/ol><ul><li>The mime type of the resource.\n<ol><li><code>disposition<\/code><\/li>\n<\/ol><\/li>\n<li>The disposition of the resource. Empty if unknown.<\/li>\n<\/ul><\/li>\n<li><p>authentication handler<\/p>\n<ol><li><code>host<\/code><\/li>\n<\/ol><ul><li>The host requesting authentication.\n<ol><li><code>realm<\/code><\/li>\n<\/ol><\/li>\n<li>The realm requesting authentication.\n<ol><li><code>retry<\/code><\/li>\n<\/ol><\/li>\n<li>Either <code>retrying<\/code> or <code>initial<\/code> depending on whether this request is a\nretrial of a previous request.\n<ol><li><code>scheme<\/code><\/li>\n<\/ol><\/li>\n<li>The authentication scheme being used. Known values include:\n<ul><li><code>default<\/code><\/li>\n<li><code>http_basic<\/code><\/li>\n<li><code>http_digest<\/code><\/li>\n<li><code>html_form<\/code><\/li>\n<li><code>ntlm<\/code><\/li>\n<li><code>negotiate<\/code><\/li>\n<li><code>client_certificate_requested<\/code><\/li>\n<li><code>server_trust_evaluation_requested<\/code>\n<ol><li><code>proxy<\/code><\/li>\n<\/ol><\/li>\n<\/ul><\/li>\n<li>Either <code>proxy<\/code> if a proxy is requesting authentication or <code>origin<\/code> if a\nremote server is.\n<ol><li><code>port<\/code><\/li>\n<\/ol><\/li>\n<li>The port connected to.\n<ol><li><code>save<\/code><\/li>\n<\/ol><\/li>\n<li>Either <code>can_save<\/code> or <code>cant_save<\/code> depending on whether WebKit itself can\nsave the credentials.<\/li>\n<\/ul><\/li>\n<li><p>permission handler<\/p>\n<ol><li><code>uri<\/code><\/li>\n<\/ol><ul><li>The URI requesting permission.\n<ol><li><code>action<\/code><\/li>\n<\/ol><\/li>\n<li>The data or action WebKit is requesting permission for. Known values\ninclude:\n<ul><li><code>geolocation<\/code><\/li>\n<li><code>unknown<\/code>\n<ol><li><code>description<\/code><\/li>\n<\/ol><\/li>\n<\/ul><\/li>\n<li>A description for the request (if available).<\/li>\n<\/ul><\/li>\n<li><p>tls error handler<\/p>\n<ol><li><code>host<\/code><\/li>\n<\/ol><ul><li>The host which failed TLS verification.\n<ol><li><code>flags<\/code><\/li>\n<\/ol><\/li>\n<li>A comma-separated list of the reasons the host failed TLS verification.\nKnown flags include:\n<ul><li><code>unknown_ca<\/code><\/li>\n<li><code>bad_identity<\/code><\/li>\n<li><code>not_activated<\/code><\/li>\n<li><code>expired<\/code><\/li>\n<li><code>revoked<\/code><\/li>\n<li><code>insecure<\/code><\/li>\n<li><code>error<\/code>\n<ol><li><code>cert_info<\/code><\/li>\n<\/ol><\/li>\n<\/ul><\/li>\n<li>A JSON object with information describing the certificate.<\/li>\n<\/ul><\/li>\n<li><p>file chooser handler<\/p>\n<ol><li><code>mode<\/code><\/li>\n<\/ol><ul><li>Either <code>multiple<\/code> or <code>single<\/code> for whether a single file or multiple files\nwere requested.\n<ol><li><code>mimetypes<\/code><\/li>\n<\/ol><\/li>\n<li>A comma-separated list of mimetypes requested.<\/li>\n<\/ul><\/li>\n<li><p>color chooser handler<\/p>\n<ol><li><code>region<\/code><\/li>\n<\/ol><ul><li>The region of the element requesting a color in the <code>WIDTHxHEIGHT+X+Y<\/code>\nformat.<\/li>\n<\/ul><\/li>\n<\/ul><h3>WINDOW MANAGER INTEGRATION<\/h3><p>As mentined before, the contents of the window title can be customized by\nsetting the <code>title_format_short<\/code> variable and the <code>title_format_long<\/code> variable\n(see above to figure out when each of them is used). You can also set <code>icon<\/code>\nvariable to path of the icon file. Some advanced window managers can also take\n<code>WM_WINDOW_ROLE<\/code> in account, which can be set by modifying <code>window_role<\/code>\nvariable.<\/p><p>Uzbl does not automatically use the current page's favicon as its icon. This\ncan, however, be done with external scripts.<\/p><p>Uzbl sets the <code>UZBL_URI<\/code> property on its window which is always the uri of the\npage loaded into uzbl, except for web inspector windows which does not have the\nproperty.<\/p><h3>EVENTS<\/h3><p>Unlike commands, events are not handled in <code>uzbl<\/code> itself, but are dispatched\nasynchronously through a text stream on <code>stdout<\/code> and\/or through a socket.\nYou'll usually use <code>uzbl<\/code> by connecting it to a so-called \"event manager\" (EM).<\/p><p>An EM is a privileged communicator with <code>uzbl<\/code>. EM sockets are given on the\ncommand line via the <code>--connect-socket<\/code> argument. EM sockets receive requests\nin addition to events. All sockets receive event output. Please be aware that\nevents contain sensitive information such as cookie data and keystrokes\n(e.g., passwords).<\/p><p>An example EM is shipped with <code>uzbl<\/code> called <code>uzbl-event-manager<\/code>. See <a href=\"https:\/\/www.uzbl.org\/README.event-manager.md\">its\ndocumentation<\/a> for more.<\/p><p>Many of the features of larger browsers are intended to be implemented via an\nEM. The EM receives all events and requests from <code>uzbl<\/code> and can implement\nhistory, keybindings, cookie management, and more. For more on the EM shipped\nwith <code>uzbl<\/code>, see <a href=\"https:\/\/www.uzbl.org\/README.event-manager.md\">its documentation<\/a><\/p><h4>Format<\/h4><p>Events are reported in the following format.<\/p><pre> EVENT &lt;NAME&gt; &lt;EVENT_NAME&gt; {DETAILS}\n<\/pre><p>Events are line-oriented, so newlines are not supported within events.<\/p><p>Requests use the following format:<\/p><pre>REQUEST-&lt;COOKIE&gt; &lt;REQUEST_NAME&gt; [ARGUMENTS...]\n<\/pre><p>Only EM sockets receive REQUEST lines. Currently, if a reply is not received\nwithin one second from <code>uzbl<\/code> sending a request, <code>uzbl<\/code> will continue without a\nreply. This is because the request blocks the GUI thread. Replies to a request\nuse the following format:<\/p><pre>REPLY-&lt;COOKIE&gt; &lt;REPLY&gt;\n<\/pre><p>If the cookie does not match the cookie from the request, <code>uzbl<\/code> will ignore it.<\/p><h4>Built-in events<\/h4><p>Uzbl will report various events by default. All of these events are part of\nuzbl itself and will be emitted based on what is happening within uzbl-core.<\/p><h5>Navigation<\/h5><ul><li><code>NAVIGATION_STARTING &lt;URI&gt; &lt;SRCFRAME&gt; &lt;DESTFRAME&gt; &lt;TYPE&gt;<\/code>\n<ul><li>Sent when a main navigation is requested for a frame. See <code>navigation_handler<\/code>\nfor the types available.<\/li>\n<\/ul><\/li>\n<li><code>LOAD_START<\/code>\n<ul><li>Sent when a main page navigation is started. This is before WebKit even\nlooks at the URI, so the name, protocol, or port may still be invalid.<\/li>\n<\/ul><\/li>\n<li><code>LOAD_REDIRECTED &lt;URI&gt;<\/code>\n<ul><li>Sent when the main page navigation has been redirected to a new URI.<\/li>\n<\/ul><\/li>\n<li><code>LOAD_COMMIT &lt;URI&gt;<\/code>\n<ul><li>Sent when the main page navigation has been committed and the network\nrequest has been sent to the server.<\/li>\n<\/ul><\/li>\n<li><code>LOAD_PROGRESS &lt;PROGRESS&gt;<\/code>\n<ul><li>Sent when the load percentage changes. The progress is in percentage.<\/li>\n<\/ul><\/li>\n<li><code>LOAD_CANCELLED &lt;URI&gt;<\/code>\n<ul><li>Sent when the user has cancelled a load (e.g., by setting the URI in the\nmiddle of another load).<\/li>\n<\/ul><\/li>\n<li><code>LOAD_ERROR &lt;URI&gt; &lt;CODE&gt; &lt;MESSAGE&gt;<\/code>\n<ul><li>Sent when WebKit has failed to load a page.<\/li>\n<\/ul><\/li>\n<li><code>LOAD_FINISH &lt;URI&gt;<\/code>\n<ul><li>Sent when a page navigation is complete. All of the top-level resources\nhave been retrieved by this point.<\/li>\n<\/ul><\/li>\n<li><code>REQUEST_QUEUED &lt;URI&gt;<\/code> (WebKit1 only)\n<ul><li>Sent when a request is queued for the network.<\/li>\n<\/ul><\/li>\n<li><code>REQUEST_STARTING &lt;URI&gt;<\/code>\n<ul><li>Sent when a request has been sent to the server.<\/li>\n<\/ul><\/li>\n<li><code>REQUEST_FINISHED &lt;URI&gt;<\/code>\n<ul><li>Sent when a request has completed.<\/li>\n<\/ul><\/li>\n<\/ul><h5>Input<\/h5><ul><li><code>KEY_PRESS &lt;MODIFIERS&gt; &lt;KEY&gt;<\/code>\n<ul><li>Sent when a key (e.g., Shift, Ctrl, etc.) is pressed. The key is in UTF-8.\nThis event is also used for mouse clicks which use the format\n<code>&lt;REP&gt;Button&lt;VALUE&gt;<\/code> which would be <code>2Button1<\/code> for a double click of the\nmain mouse button (rep is empty for single clicks). Multiple clicks usually\nonly show as a <code>KEY_RELEASE<\/code> event due to GDK implementation details.<\/li>\n<\/ul><\/li>\n<li><code>KEY_RELEASE &lt;MODIFIERS&gt; &lt;KEY&gt;<\/code>\n<ul><li>Sent when a key (e.g., Shift, Ctrl, etc.) is released. The key is in UTF-8.<\/li>\n<\/ul><\/li>\n<li><code>MOD_PRESS &lt;MODIFIERS&gt; &lt;MODIFIER&gt;<\/code>\n<ul><li>Sent when a modifier (e.g., Shift, Ctrl, etc.) is pressed.<\/li>\n<\/ul><\/li>\n<li><code>MOD_RELEASE &lt;MODIFIERS&gt; &lt;MODIFIER&gt;<\/code>\n<ul><li>Sent when a modifier (e.g., Shift, Ctrl, etc.) is released.<\/li>\n<\/ul><\/li>\n<\/ul><h5>Commands<\/h5><ul><li><code>BUILTINS &lt;COMMANDS&gt;<\/code>\n<ul><li>On startup, <code>uzbl<\/code> will emit this event with a JSON list of the names of\nall commands it understands.<\/li>\n<\/ul><\/li>\n<li><code>COMMAND_ERROR &lt;REASON&gt;<\/code>\n<ul><li>Sent when <code>uzbl<\/code> cannot execute a command.<\/li>\n<\/ul><\/li>\n<li><code>COMMAND_EXECUTED &lt;NAME&gt; [ARGS...]<\/code>\n<ul><li>Sent after a command has executed.<\/li>\n<\/ul><\/li>\n<li><code>FILE_INCLUDED &lt;PATH&gt;<\/code>\n<ul><li>Sent <em>after<\/em> a file is included.<\/li>\n<\/ul><\/li>\n<\/ul><h5>Uzbl<\/h5><ul><li><code>FIFO_SET<\/code>\n<ul><li>Sent <code>uzbl<\/code> opens a communication FIFO.<\/li>\n<\/ul><\/li>\n<li><code>SOCKET_SET &lt;PATH&gt;<\/code>\n<ul><li>Sent <code>uzbl<\/code> opens a communication socket.<\/li>\n<\/ul><\/li>\n<li><code>INSTANCE_START &lt;PID&gt;<\/code>\n<ul><li>Sent on startup.<\/li>\n<\/ul><\/li>\n<li><code>PLUG_CREATED &lt;ID&gt;<\/code>\n<ul><li>Sent if <code>uzbl<\/code> is started in plug mode. The ID is for the Xembed socket.<\/li>\n<\/ul><\/li>\n<li><code>INSTANCE_EXIT &lt;PID&gt;<\/code>\n<ul><li>Sent before <code>uzbl<\/code> quits. When this is sent, <code>uzbl<\/code> has already stopped\nlistening on all sockets.<\/li>\n<\/ul><\/li>\n<li><code>VARIABLE_SET &lt;NAME&gt; &lt;str|int|ull|double&gt; {VALUE}<\/code>\n<ul><li>Sent when a variable has been set. Not all variable changes cause a\n<code>VARIABLE_SET<\/code> event to occur (e.g., any variable managed by WebKit behind\nthe scenes does not trigger this event).<\/li>\n<\/ul><\/li>\n<\/ul><h5>Page<\/h5><ul><li><code>BLUR_ELEMENT &lt;NAME&gt;<\/code> (WebKit1 only)\n<ul><li>Sent when an element loses focus.<\/li>\n<\/ul><\/li>\n<li><code>FOCUS_ELEMENT &lt;NAME&gt;<\/code> (WebKit1 only)\n<ul><li>Sent when an element gains focus.<\/li>\n<\/ul><\/li>\n<li><code>LINK_HOVER &lt;URI&gt; &lt;TITLE&gt;<\/code>\n<ul><li>Sent when a link is hovered over using the mouse.<\/li>\n<\/ul><\/li>\n<li><code>LINK_UNHOVER &lt;URI&gt;<\/code>\n<ul><li>Sent when a link is hovered over using the mouse. The URI is the\npreviously hovered link URI.<\/li>\n<\/ul><\/li>\n<li><code>FORM_ACTIVE &lt;BUTTON&gt;<\/code>\n<ul><li>Sent when a form element has gained focus because of a mouse click.<\/li>\n<\/ul><\/li>\n<li><code>ROOT_ACTIVE &lt;BUTTON&gt;<\/code>\n<ul><li>Sent when the background page has been clicked.<\/li>\n<\/ul><\/li>\n<li><code>SCROLL_HORIZ &lt;VALUE&gt; &lt;MIN&gt; &lt;MAX&gt; &lt;PAGE&gt;<\/code>\n<ul><li>Sent when the page horizontal scroll bar changes. The min and max values\nare the bounds for scrolling, page is the size that fits in the viewport\nand the value is the current position of the scrollbar.<\/li>\n<\/ul><\/li>\n<li><code>SCROLL_VERT &lt;VALUE&gt; &lt;MIN&gt; &lt;MAX&gt; &lt;PAGE&gt;<\/code>\n<ul><li>Similar to <code>SCROLL_HORIZ<\/code>, but for the vertical scrollbar.<\/li>\n<\/ul><\/li>\n<li><code>TITLE_CHANGED &lt;TITLE&gt;<\/code>\n<ul><li>Sent when the page title changes.<\/li>\n<\/ul><\/li>\n<\/ul><h5>Window<\/h5><ul><li><code>GEOMETRY_CHANGED &lt;GEOMETRY&gt;<\/code>\n<ul><li>Sent when the geometry changes.<\/li>\n<\/ul><\/li>\n<li><code>WEBINSPECTOR &lt;open|close&gt;<\/code>\n<ul><li>Sent when the web inspector window is opened or closed.<\/li>\n<\/ul><\/li>\n<li><code>REQ_NEW_WINDOW &lt;URI&gt;<\/code>\n<ul><li>Sent when a new window is being requested.<\/li>\n<\/ul><\/li>\n<li><code>CLOSE_WINDOW<\/code>\n<ul><li>Sent when <code>uzbl<\/code> is closing its window.<\/li>\n<\/ul><\/li>\n<li><code>FOCUS_LOST<\/code>\n<ul><li>Sent when <code>uzbl<\/code> loses the keyboard focus.<\/li>\n<\/ul><\/li>\n<li><code>FOCUS_GAINED<\/code>\n<ul><li>Sent when <code>uzbl<\/code> gains the keyboard focus.<\/li>\n<\/ul><\/li>\n<\/ul><h5>Download<\/h5><ul><li><code>DOWNLOAD_STARTED &lt;DESTINATION&gt;<\/code>\n<ul><li>Sent when a download to the given URI has started.<\/li>\n<\/ul><\/li>\n<li><code>DOWNLOAD_PROGRESS &lt;DESTINATION&gt; &lt;PROGRESS&gt;<\/code>\n<ul><li>Sent when progress for a download has been updated. The progress is a value\nbetween 0 and 1.<\/li>\n<\/ul><\/li>\n<li><code>DOWNLOAD_ERROR &lt;DESTINATION&gt; &lt;REASON&gt; &lt;CODE&gt; &lt;MESSAGE&gt;<\/code>\n<ul><li>Sent when a download has an error.<\/li>\n<\/ul><\/li>\n<li><code>DOWNLOAD_COMPLETE &lt;DESTINATION&gt;<\/code>\n<ul><li>Sent when a download has completed.<\/li>\n<\/ul><\/li>\n<\/ul><h5>Cookie<\/h5><ul><li><code>ADD_COOKIE &lt;DOMAIN&gt; &lt;PATH&gt; &lt;NAME&gt; &lt;VALUE&gt; &lt;SCHEME&gt; &lt;EXPIRATION&gt;<\/code>\n<ul><li>Sent when a cookie is added.<\/li>\n<\/ul><\/li>\n<li><code>DELETE_COOKIE &lt;DOMAIN&gt; &lt;PATH&gt; &lt;NAME&gt; &lt;VALUE&gt; &lt;SCHEME&gt; &lt;EXPIRATION&gt;<\/code>\n<ul><li>Sent when a cookie is deleted.<\/li>\n<\/ul><\/li>\n<\/ul><h3>COMMAND LINE ARGUMENTS<\/h3><p><code>uzbl<\/code> is invoked as<\/p><pre>uzbl-core [ARGUMENTS] [URI]\n<\/pre><p>where <code>arguments<\/code> and <code>uri<\/code> are both optional. <code>arguments<\/code> can be:<\/p><ul><li><code>-v<\/code>, <code>--verbose<\/code>\n<ul><li>Sets <code>verbose<\/code> to be non-zero.<\/li>\n<\/ul><\/li>\n<li><code>-n<\/code>, <code>--named=NAME<\/code>\n<ul><li>Name of the current instance (defaults to Xorg window id or random for\nGtkSocket mode).<\/li>\n<\/ul><\/li>\n<li><code>-c<\/code>, <code>--config=FILE<\/code>\n<ul><li>Path to config file or <code>-<\/code> for stdin.<\/li>\n<\/ul><\/li>\n<li><code>-s<\/code>, <code>--xembed-socket=SOCKETID<\/code>\n<ul><li>Xembed socket ID.<\/li>\n<\/ul><\/li>\n<li><code>--connect-socket=CSOCKET<\/code>\n<ul><li>Connect to server socket for event managing.<\/li>\n<\/ul><\/li>\n<li><code>-p<\/code>, <code>--print-events<\/code>\n<ul><li>Sets <code>print_events<\/code> to be non-zero.<\/li>\n<\/ul><\/li>\n<li><code>-g<\/code>, <code>--geometry=GEOMETRY<\/code>\n<ul><li>Runs <code>geometry GEOMETRY<\/code> with the given value on startup. Ignored if <code>uzbl<\/code> is\nembedded.<\/li>\n<\/ul><\/li>\n<li><code>-V<\/code>, <code>--version<\/code>\n<ul><li>Print the version and exit.<\/li>\n<\/ul><\/li>\n<li><code>--web-extensions-dir<\/code>\n<ul><li>Directory that will be searched for webkit extensions<\/li>\n<\/ul><\/li>\n<li><code>--display=DISPLAY<\/code>\n<ul><li>X display to use.<\/li>\n<\/ul><\/li>\n<li><code>--help<\/code>\n<ul><li>Display help.<\/li>\n<\/ul><\/li>\n<\/ul><p><code>uzbl-core scheme:\/\/address<\/code> will work as you expect. If you don't provide the\n<code>scheme:\/\/<\/code> part, it will check if the argument is an existing file in the\nfilesystem, if it is, it will prepend <code>file:\/\/<\/code>, if not, it will prepend\n<code>http:\/\/<\/code>.<\/p><h3>BUGS<\/h3><p>Please report new issues to <a href=\"https:\/\/github.com\/uzbl\/uzbl\/issues\">Uzbl's issues page on github<\/a><\/p>","created_at":"2018-11-07T16:50:30+0000","updated_at":"2018-11-07T16:50:30+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":44,"domain_name":"www.uzbl.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":511,"title":"Parsing Expressions \u00b7 Crafting Interpreters","url":"http:\/\/www.craftinginterpreters.com\/parsing-expressions.html","content":"<p>6<\/p><blockquote>\n<p>Grammar, which knows how to control even kings.\n<cite>Moli\u00e8re<\/cite><\/p>\n<\/blockquote><p>This chapter marks the first major milestone of the\nbook. Many of us have cobbled together a mishmash of regular expressions and\nsubstring operations to extract some sense out of a pile of text. The code was\nprobably riddled with bugs and a beast to maintain. Writing a <em>real<\/em> parser\u2014one with decent error-handling, a coherent internal structure, and the ability\nto robustly chew through a sophisticated syntax\u2014is considered a rare,\nimpressive skill. In this chapter, you will attain\nit.<\/p><aside><p>\u201cParse\u201d comes to English from the Old French \u201cpars\u201d for \u201cpart of speech\u201d. It\nmeans to take a text and map each word to the grammar of the language. We use it\nhere in the same sense, except that our language is a little more modern than\nOld French.<\/p>\n<\/aside><aside><p>Like many rites of passage, you\u2019ll probably find it looks a little smaller, a\nlittle less daunting when it\u2019s behind you than when it loomed ahead.<\/p>\n<\/aside><p>It\u2019s easier than you think, partially because we front-loaded a lot of the hard\nwork in the <a href=\"http:\/\/www.craftinginterpreters.com\/representing-code.html\">last chapter<\/a>. You already know your way around a formal grammar.\nYou\u2019re familiar with syntax trees, and we have some Java classes to represent\nthem. The only remaining piece is parsing\u2014transmogrifying a sequence of\ntokens into one of those syntax trees.<\/p><p>Some CS textbooks make a big deal out of parsers. In the 60s, computer\nscientists\u2014reasonably fed up with programming in assembly language\u2014started\ndesigning more sophisticated, human-friendly languages\nlike FORTRAN and ALGOL. Alas, they weren\u2019t very <em>machine<\/em>-friendly, for the\nprimitive machines at the time.<\/p><aside><p>Consider how harrowing assembly programming on those old machines must have been\nfor <em>FORTRAN<\/em> to be an improvement.<\/p>\n<\/aside><p>They designed languages that they honestly weren\u2019t even sure how to write\ncompilers for, and then did ground-breaking work inventing parsing and compiling\ntechniques that could handle these new big languages on those old tiny machines.<\/p><p>Classic compiler books read like fawning hagiographies of these pioneers and\ntheir tools. The cover of \u201cCompilers: Principles, Techniques, and Tools\u201d\nliterally has a dragon labeled \u201ccomplexity of compiler design\u201d being slain by a\nknight bearing a sword and shield branded \u201cLALR parser generator\u201d and \u201csyntax\ndirected translation\u201d. They laid it on thick.<\/p><p>A little self-congratulation is well-deserved, but the truth is you don\u2019t need\nto know most of that stuff to bang out a high quality parser for a modern\nmachine. As always, I encourage you to broaden your education and take it in\nlater, but this book omits the trophy case.<\/p><p>In the last chapter, I said you can \u201cplay\u201d a context free grammar like a game in\norder to <em>generate<\/em> strings. Now, we play that game in reverse. Given a\nstring\u2014a series of tokens\u2014we map those tokens to terminals in the grammar\nto figure out which rules could have generated that string.<\/p><p>The \u201ccould have\u201d part is interesting. It\u2019s entirely possible to create a grammar\nthat is <em>ambiguous<\/em>, where different choices of productions can lead to the same\nstring. When you\u2019re using the grammar to <em>generate<\/em> strings, that doesn\u2019t matter\nmuch. Once you have the string, who cares how you got to it?<\/p><p>When parsing, ambiguity means the parser may misunderstand the user\u2019s code.\nHere\u2019s the Lox expression grammar we put together in the last chapter:<\/p><div class=\"codehilite\"><pre>expression \u2192 literal\n           | unary\n           | binary\n           | grouping ;\nliteral    \u2192 NUMBER | STRING | \"false\" | \"true\" | \"nil\" ;\ngrouping   \u2192 \"(\" expression \")\" ;\nunary      \u2192 ( \"-\" | \"!\" ) expression ;\nbinary     \u2192 expression operator expression ;\noperator   \u2192 \"==\" | \"!=\" | \"&lt;\" | \"&lt;=\" | \"&gt;\" | \"&gt;=\"\n           | \"+\"  | \"-\"  | \"*\" | \"\/\" ;\n<\/pre><\/div><p>This is a valid string in that grammar:<\/p><p><img src=\"http:\/\/www.craftinginterpreters.com\/image\/parsing-expressions\/tokens.png\" alt=\"6 \/ 3 - 1\" \/><\/p><p>But there are two ways we could have generated it. One way is:<\/p><ol><li>Starting at <code>expression<\/code>, pick <code>binary<\/code>.<\/li>\n<li>For the left-hand <code>expression<\/code>, pick <code>NUMBER<\/code>, and use <code>6<\/code>.<\/li>\n<li>For the operator, pick <code>\"\/\"<\/code>.<\/li>\n<li>For the right-hand <code>expression<\/code>, pick <code>binary<\/code> again.<\/li>\n<li>In that nested <code>binary<\/code> expression, pick <code>3 - 1<\/code>.<\/li>\n<\/ol><p>Another is:<\/p><ol><li>Starting at <code>expression<\/code>, pick <code>binary<\/code>.<\/li>\n<li>For the left-hand <code>expression<\/code>, pick <code>binary<\/code> again.<\/li>\n<li>In that nested <code>binary<\/code> expression, pick <code>6 \/ 3<\/code>.<\/li>\n<li>Back at the outer <code>binary<\/code>, for the operator, pick <code>\"-\"<\/code>.<\/li>\n<li>For the right-hand <code>expression<\/code>, pick <code>NUMBER<\/code>, and use <code>1<\/code>.<\/li>\n<\/ol><p>Those produce the same <em>strings<\/em>, but not the same <em>syntax trees<\/em>:<\/p><p><img src=\"http:\/\/www.craftinginterpreters.com\/image\/parsing-expressions\/syntax-trees.png\" alt=\"Two valid syntax trees: (6 \/ 3) - 1 and 6 \/ (3 - 1)\" \/><\/p><p>In other words, the grammar allows seeing the expression as <code>(6 \/ 3) - 1<\/code> or <code>6\n\/ (3 - 1)<\/code>. That in turn affects the result of evaluating it. The way\nmathematicians have solved this ambiguity since blackboards were first invented\nis by defining rules for precedence and associativity.<\/p><ul><li>\n<p><strong>Precedence<\/strong> determines which operator\n    is evaluated first in an expression containing a mixture of different\n    operators. Precedence rules tell us that we evaluate the <code>\/<\/code> before the <code>-<\/code>\n    in the above example. Operators with higher precedence are evaluated\n    before operators with lower precedence. Equivalently, higher precedence\n    operators are said to \u201cbind tighter\u201d.<\/p>\n<\/li>\n<li>\n<p><strong>Associativity<\/strong> determines which operator is evaluated first in a series\n    of the <em>same<\/em> operator. When an operator is <strong>left-associative<\/strong> (think\n    \u201cleft-to-right\u201d), operators on the left evaluate before ones of the right.\n    Since <code>-<\/code> is left-associative, this expression:<\/p>\n<div class=\"codehilite\"><pre>5 - 3 - 1\n<\/pre><\/div>\n<p>is equivalent to:<\/p>\n<div class=\"codehilite\"><pre>(5 - 3) - 1\n<\/pre><\/div>\n<p>Assignment, on the other hand, is <strong>right-associative<\/strong>. This:<\/p>\n<div class=\"codehilite\"><pre>a = b = c\n<\/pre><\/div>\n<p>is equivalent to:<\/p>\n<div class=\"codehilite\"><pre>a = (b = c)\n<\/pre><\/div>\n<\/li>\n<\/ul><aside><p>While not common these days, some languages specify that certain pairs of\noperators have <em>no<\/em> relative precedence. That makes it a syntax error to mix\nthose operators in an expression without using explicit grouping.<\/p>\n<p>Likewise, some operators are <strong>non-associative<\/strong>. That means it\u2019s an error to\nuse that operator more than once in a sequence. For example, Perl\u2019s range\noperator isn\u2019t associative, so <code>a .. b<\/code> is OK, but <code>a .. b .. c<\/code> is an error.<\/p>\n<\/aside><p>Without well-defined precedence and associativity, an expression that uses\nmultiple operators is ambiguous\u2014it can be parsed into different syntax trees,\nwhich could in turn evaluate to different results. For Lox, we follow the same\nprecedence rules as C, going from highest to lowest:<\/p><table><thead><tr><td>Name<\/td>\n  <td>Operators<\/td>\n  <td>Associates<\/td>\n<\/tr><\/thead><tbody><tr><td>Unary<\/td>\n  <td><code>!<\/code> <code>-<\/code><\/td>\n  <td>Right<\/td>\n<\/tr><tr><td>Multiplication<\/td>\n  <td><code>\/<\/code> <code>*<\/code><\/td>\n  <td>Left<\/td>\n<\/tr><tr><td>Addition<\/td>\n  <td><code>-<\/code> <code>+<\/code><\/td>\n  <td>Left<\/td>\n<\/tr><tr><td>Comparison<\/td>\n  <td><code>&gt;<\/code> <code>&gt;=<\/code>\n      <code>&lt;<\/code> <code>&lt;=<\/code><\/td>\n  <td>Left<\/td>\n<\/tr><tr><td>Equality<\/td>\n  <td><code>==<\/code> <code>!=<\/code><\/td>\n  <td>Left<\/td>\n<\/tr><\/tbody><\/table><p>How do we stuff these restrictions into our\ncontext-free grammar? Right now, when we have an expression that contains\nsubexpressions, like a binary operator, we allow <em>any<\/em> expression in there:<\/p><aside><p>Instead of baking precedence right into the grammar rules, some parser\ngenerators let you keep the same ambiguous-but-simple grammar and then add in a\nlittle explicit operator precedence metadata on the side in order to\ndisambiguate.<\/p>\n<\/aside><div class=\"codehilite\"><pre>binary \u2192 expression operator expression ;\n<\/pre><\/div><p>The <code>expression<\/code> nonterminal allows us to pick any kind of expression as an\noperand, regardless of the operator we picked. The rules of precedence limit\nthat. For example, an operand of a <code>*<\/code> expression cannot be a <code>+<\/code> expression, since the latter has lower precedence.<\/p><aside><p>Of course, it could be a <em>parenthesized<\/em> addition expression, but that\u2019s because\nthe parentheses themeselves are treated as having the highest precedence.<\/p>\n<\/aside><p>For the multiplication operands, we need a nonterminal that means \u201cany kind of\nexpression of higher precedence than <code>*<\/code>\u201d. Something like:<\/p><div class=\"codehilite\"><pre>multiplication \u2192 higherThanMultiply \"*\" higherThanMultiply ;\n<\/pre><\/div><p>Since <code>*<\/code> and <code>\/<\/code> have the same precedence and the level above them is unary\noperators, a better approximation is:<\/p><div class=\"codehilite\"><pre>multiplication \u2192 unary ( \"*\" | \"\/\" ) unary ;\n<\/pre><\/div><p>Except that\u2019s not <em>quite<\/em> right. We broke associativity. This <code>multiplication<\/code>\nrule doesn\u2019t allow <code>1 * 2 * 3<\/code>. To support associativity, we make one side\npermit expressions at the <em>same<\/em> level. Which side we choose determines if the\noperator is left- or right-associative. Since multiplication and division are left-associative, it\u2019s:<\/p><aside><p>In principle, it doesn\u2019t matter whether you treat multiplication as left- or\nright-associative\u2014you get the same result either way. Alas, in the real world\nwith limited precision, roundoff and overflow mean that associativity can affect\nthe result of a sequence of multiplications. Consider:<\/p>\n<div class=\"codehilite\"><pre>print 0.1 + (0.2 + 0.3);\nprint (0.1 + 0.2) + 0.3;\n<\/pre><\/div>\n<p>In languages like Lox that use <a href=\"https:\/\/en.wikipedia.org\/wiki\/Double-precision_floating-point_format\">IEEE 754<\/a> double-precision floating-point\nnumbers, the first evaluates to <code>0.6<\/code>, while the second yields\n<code>0.6000000000000001<\/code>. Sometimes that tiny difference matters.\n<a href=\"https:\/\/docs.oracle.com\/cd\/E19957-01\/806-3568\/ncg_goldberg.html\">This<\/a> is a good place to learn more.<\/p>\n<\/aside><div class=\"codehilite\"><pre>multiplication \u2192 multiplication ( \"*\" | \"\/\" ) unary ;\n<\/pre><\/div><p>This is correct, but the fact that the first nonterminal in the body of the rule\nis the same as the head of the rule means this production is <strong>left-recursive<\/strong>.\nSome parsing techniques, including the one we\u2019re going to use, have trouble with\nleft recursion. Instead, we\u2019ll use this other style:<\/p><div class=\"codehilite\"><pre>multiplication \u2192 unary ( ( \"\/\" | \"*\" ) unary )* ;\n<\/pre><\/div><p>At the grammar level, this sidesteps left recursion by saying a multiplication\nexpression is a flat <em>sequence<\/em> of multiplications and divisions. This mirrors\nthe code we\u2019ll use to parse a sequence of multiplications.<\/p><p>If we rejigger all of the binary operator rules in the same way, we get:<\/p><div class=\"codehilite\"><pre>expression     \u2192 equality ;\nequality       \u2192 comparison ( ( \"!=\" | \"==\" ) comparison )* ;\ncomparison     \u2192 addition ( ( \"&gt;\" | \"&gt;=\" | \"&lt;\" | \"&lt;=\" ) addition )* ;\naddition       \u2192 multiplication ( ( \"-\" | \"+\" ) multiplication )* ;\nmultiplication \u2192 unary ( ( \"\/\" | \"*\" ) unary )* ;\nunary          \u2192 ( \"!\" | \"-\" ) unary\n               | primary ;\nprimary        \u2192 NUMBER | STRING | \"false\" | \"true\" | \"nil\"\n               | \"(\" expression \")\" ;\n<\/pre><\/div><p>Instead of a single <code>binary<\/code> rule, there are now four separate rules for each binary operator precedence level. The main <code>expression<\/code> rule is no longer a flat series of <code>|<\/code> branches for each kind of expression. Instead, it is simply an alias for the lowest-precedence expression form, <code>equality<\/code>, because that includes all higher-precedence expressions too.<\/p><aside><p>In later chapters, when we expand the grammar to include assignment and logical\noperators, this will change, but equality is the lowest for now.<\/p>\n<\/aside><p>Each binary operator\u2019s operands use the next-higher precedence level. After the\nbinary operators, we go to <code>unary<\/code>, the rule for unary operator expressions,\nsince those bind tighter than binary ones. For <code>unary<\/code>, we <em>do<\/em> use a recursive\nrule because unary operators are right-associative, which means instead of left\nrecursion, we have <strong>right recursion<\/strong>. The <code>unary<\/code> nonterminal is at the end of\nthe body for <code>unary<\/code>, not the beginning, and our parser won\u2019t have any trouble\nwith that.<\/p><p>Finally, the <code>unary<\/code> rule alternately bubbles up to <code>primary<\/code> in cases where it\ndoesn\u2019t match a unary operator. \u201cPrimary\u201d is the time-honored name for the\nhighest level of precedence including the atomic expressions like literals.<\/p><p>Our grammar grew a bit, but it\u2019s unambiguous now. We\u2019re ready to make a parser\nfor it.<\/p><p>There is a whole pack of parsing techniques whose names mostly seem to be\ncombinations of \u201cL\u201d and \u201cR\u201d\u2014<a href=\"https:\/\/en.wikipedia.org\/wiki\/LL_parser\">LL(k)<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/LR_parser\">LR(1)<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/LALR_parser\">LALR<\/a>\u2014along with\nmore exotic beasts like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Parser_combinator\">parser combinators<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Earley_parser\">Earley parsers<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Shunting-yard_algorithm\">the\nshunting yard algorithm<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Parsing_expression_grammar\">packrat parsing<\/a>. For our first interpreter,\none technique is more than sufficient: <strong>recursive descent<\/strong>.<\/p><p>Recursive descent is the simplest way to build a parser, and doesn\u2019t require\nusing complex parser generator tools like Yacc, Bison or ANTLR. All you need is\nstraightforward hand-written code. Don\u2019t be fooled by its simplicity, though.\nRecursive descent parsers are fast, robust, and can support sophisticated\nerror-handling. In fact, GCC, V8 (the JavaScript VM in Chrome), Roslyn (the C#\ncompiler written in C#) and many other heavyweight production language\nimplementations use recursive descent. It kicks ass.<\/p><p>It is considered a <strong>top-down parser<\/strong> because it starts from the top or\noutermost grammar rule (here <code>expression<\/code>) and works its way down into the nested subexpressions before finally\nreaching the leaves of the syntax tree. This is in contrast with bottom-up\nparsers like LR that start with primary expressions and compose them into larger\nand larger chunks of syntax.<\/p><aside><p>It\u2019s called \u201crecursive <em>descent<\/em>\u201d because it walks <em>down<\/em> the grammar.\nConfusingly, we also use direction metaphorically when talking about \u201chigh\u201d and\n\u201clow\u201d precedence, but the orientation is reversed. In a top-down grammar, you\nreach the lowest-precedence expressions first because they may in turn contain\nsubexpressions of higher precedence.<\/p>\n<p><img alt=\"Top-down grammar rules in order of increasing precedence.\" src=\"http:\/\/www.craftinginterpreters.com\/image\/parsing-expressions\/direction.png\" \/><\/p>\n<p>CS people really need to get together and straighten out their metaphors. Don\u2019t\neven get me started on which direction the stack is supposed to grow.<\/p>\n<\/aside><p>A recursive descent parser is a literal translation of the grammar\u2019s rules\nstraight into imperative code. Each rule becomes a function. The body of the\nrule translates to code roughly like:<\/p><table><thead><tr><td>Grammar notation<\/td>\n  <td>Code representation<\/td>\n<\/tr><\/thead><tbody><tr><td>Terminal<\/td><td>Code to match and consume a token<\/td><\/tr><tr><td>Nonterminal<\/td><td>Call to that rule\u2019s function<\/td><\/tr><tr><td><code>|<\/code><\/td><td>If or switch statement<\/td><\/tr><tr><td><code>*<\/code> or <code>+<\/code><\/td><td>While or for loop<\/td><\/tr><tr><td><code>?<\/code><\/td><td>If statement<\/td><\/tr><\/tbody><\/table><p>It\u2019s called \u201c<em>recursive<\/em> descent\u201d because when a grammar rules refers to itself\u2014directly or indirectly\u2014that translates to recursive method calls.<\/p><h3><a href=\"#the-parser-class\" name=\"the-parser-class\"><small>6\u200a.\u200a2\u200a.\u200a1<\/small> The parser class<\/a><\/h3><p>Each grammar rule becomes a method inside this new class:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>create new file<\/p><pre>package com.craftinginterpreters.lox;\nimport java.util.List;                                 \nimport static com.craftinginterpreters.lox.TokenType.*;\nclass Parser {                                         \n  private final List&lt;Token&gt; tokens;                    \n  private int current = 0;                             \n  Parser(List&lt;Token&gt; tokens) {                         \n    this.tokens = tokens;                              \n  }                                                    \n}                                                      \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, create new file<\/p><p>Like the scanner, it consumes a sequence, only now we\u2019re working at the level of\nentire tokens. It takes in a list of tokens and uses <code>current<\/code> to point to the\nnext token eagerly waiting to be used.<\/p><p>We\u2019re going to run straight through the expression grammar now and translate\neach rule to Java code. The first rule, <code>expression<\/code>, simply expands to the\n<code>equality<\/code> rule, so that\u2019s straightforward:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>Parser<\/em>()<\/p><pre>  private Expr expression() {\n    return equality();       \n  }                          \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>Parser<\/em>()<\/p><p>Each method for parsing a grammar rule produces a syntax tree for that rule and\nreturns it to the caller. When the body of the rule contains a nonterminal\u2014a\nreference to another rule\u2014we call that rule\u2019s\nmethod.<\/p><aside><p>This is why left recursion is problematic for recursive descent. The function\nfor a left-recursive rule immediately calls itself, which calls itself again,\nand so on, until the parser hits a stack overflow and dies.<\/p>\n<\/aside><p>The rule for equality is a little more complex:<\/p><div class=\"codehilite\"><pre>equality \u2192 comparison ( ( \"!=\" | \"==\" ) comparison )* ;\n<\/pre><\/div><p>In Java, that becomes:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>expression<\/em>()<\/p><pre>  private Expr equality() {                         \n    Expr expr = comparison();\n    while (match(BANG_EQUAL, EQUAL_EQUAL)) {        \n      Token operator = previous();                  \n      Expr right = comparison();                    \n      expr = new Expr.Binary(expr, operator, right);\n    }                                               \n    return expr;                                    \n  }                                                 \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>expression<\/em>()<\/p><p>Let\u2019s step through it. The left <code>comparison<\/code> nonterminal in the body is\ntranslated to the first call to <code>comparison()<\/code> and we store that in a local\nvariable.<\/p><p>Then, the <code>( ... )*<\/code> loop in the rule is mapped to a while loop. We need to know\nwhen to exit that loop. We can see that inside the rule, we must first find\neither a <code>!=<\/code> or <code>==<\/code> token. So, if we <em>don\u2019t<\/em> see one of those, we must be done\nwith the sequence of equality operators. We express that check using a handy\n<code>match()<\/code> method:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>equality<\/em>()<\/p><pre>  private boolean match(TokenType... types) {\n    for (TokenType type : types) {           \n      if (check(type)) {                     \n        advance();                           \n        return true;                         \n      }                                      \n    }\n    return false;                            \n  }                                          \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>equality<\/em>()<\/p><p>It checks to see if the current token is any of the given types. If so, it\nconsumes it and returns <code>true<\/code>. Otherwise, it returns <code>false<\/code> and leaves the\ntoken where it is.<\/p><p>The <code>match()<\/code> method is defined in terms of two more fundamental operations:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>match<\/em>()<\/p><pre>  private boolean check(TokenType type) {\n    if (isAtEnd()) return false;         \n    return peek().type == type;          \n  }                                      \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>match<\/em>()<\/p><p>This returns <code>true<\/code> if the current token is of the given type. Unlike <code>match()<\/code>,\nit doesn\u2019t consume the token, it only looks at it.<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>check<\/em>()<\/p><pre>  private Token advance() {   \n    if (!isAtEnd()) current++;\n    return previous();        \n  }                           \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>check<\/em>()<\/p><p>This consumes the current token and returns it, similar to how our scanner\u2019s\n<code>advance()<\/code> method did with characters.<\/p><p>These methods bottom out on the last handful of primitive operations:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>advance<\/em>()<\/p><pre>  private boolean isAtEnd() {      \n    return peek().type == EOF;     \n  }\n  private Token peek() {           \n    return tokens.get(current);    \n  }                                \n  private Token previous() {       \n    return tokens.get(current - 1);\n  }                                \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>advance<\/em>()<\/p><p><code>isAtEnd()<\/code> checks if we\u2019ve run out of tokens to parse. <code>peek()<\/code> returns the\ncurrent token we have yet to consume and <code>previous()<\/code> returns the most recently\nconsumed token. The latter makes it easier to use <code>match()<\/code> and then access the\njust-matched token.<\/p><p>That\u2019s most of the parsing infrastructure we need. Where were we? Right, so if\nwe are inside the while loop in <code>equality()<\/code>, then the parser knows it found a\n<code>!=<\/code> or <code>==<\/code> operator and must be parsing an equality expression.<\/p><p>It grabs the token that was matched for the operator so we can track which kind\nof binary expression this is. Then it calls <code>comparison()<\/code> again to parse the\nright-hand operand. It combines the operator and the two operands into a new\n<code>Expr.Binary<\/code> syntax tree node, and then loops around. Each time, it stores the\nexpression back in the same <code>expr<\/code> local variable. As it zips through a sequence\nof equality expressions, that creates a left-associative nested tree of binary\noperator nodes.<\/p><p><img src=\"http:\/\/www.craftinginterpreters.com\/image\/parsing-expressions\/sequence.png\" alt=\"The syntax tree created by parsing 'a == b == c == d == e'\" \/><\/p><aside><p>Parsing <code>a == b == c == d == e<\/code>. Each iteration, we create a new binary\nexpression using the previous one as the left operand.<\/p>\n<\/aside><p>The parser falls out of the loop once it hits a token that\u2019s not an equality\noperator. Finally, it returns the expression. Note that if it doesn\u2019t encounter\na single equality operator, then it never enters the loop. In that case, the\n<code>equality()<\/code> method effectively calls and returns <code>comparison()<\/code>. In that way,\nthis method matches an equality operator <em>or anything of higher precedence<\/em>.<\/p><p>Moving on to the next rule\u2026<\/p><div class=\"codehilite\"><pre>comparison \u2192 addition ( ( \"&gt;\" | \"&gt;=\" | \"&lt;\" | \"&lt;=\" ) addition )* ;\n<\/pre><\/div><p>Translated to Java:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>equality<\/em>()<\/p><pre>  private Expr comparison() {                                \n    Expr expr = addition();\n    while (match(GREATER, GREATER_EQUAL, LESS, LESS_EQUAL)) {\n      Token operator = previous();                           \n      Expr right = addition();                               \n      expr = new Expr.Binary(expr, operator, right);         \n    }                                                        \n    return expr;                                             \n  }                                                          \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>equality<\/em>()<\/p><p>The grammar rule is virtually identical to <code>equality<\/code> and so is the\ncorresponding code. The only differences are the\ntoken types for the operators we match, and the method we call for the operands,\nnow <code>addition()<\/code> instead of <code>comparison()<\/code>. The remaining two binary operator\nrules follow the same pattern:<\/p><aside><p>If you wanted to do some clever Java 8, you could create a helper method for\nparsing a left-associative series of binary operators given a list of token\ntypes and an operand method handle and unify some of this redundant code.<\/p>\n<\/aside><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>comparison<\/em>()<\/p><pre>  private Expr addition() {                         \n    Expr expr = multiplication();\n    while (match(MINUS, PLUS)) {                    \n      Token operator = previous();                  \n      Expr right = multiplication();                \n      expr = new Expr.Binary(expr, operator, right);\n    }                                               \n    return expr;                                    \n  }                                                 \n  private Expr multiplication() {                   \n    Expr expr = unary();                            \n    while (match(SLASH, STAR)) {                    \n      Token operator = previous();                  \n      Expr right = unary();                         \n      expr = new Expr.Binary(expr, operator, right);\n    }                                               \n    return expr;                                    \n  }                                                 \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>comparison<\/em>()<\/p><p>That\u2019s all of the binary operators, parsed with the correct precedence and\nassociativity. We\u2019re crawling up the precedence hierarchy and now we\u2019ve reached\nthe unary operators:<\/p><div class=\"codehilite\"><pre>unary \u2192 ( \"!\" | \"-\" ) unary\n      | primary ;\n<\/pre><\/div><p>The code for this is a little different:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>multiplication<\/em>()<\/p><pre>  private Expr unary() {                     \n    if (match(BANG, MINUS)) {                \n      Token operator = previous();           \n      Expr right = unary();                  \n      return new Expr.Unary(operator, right);\n    }\n    return primary();                        \n  }                                          \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>multiplication<\/em>()<\/p><p>Again, we look at the current token to see how to\nparse. If it\u2019s a <code>!<\/code> or <code>-<\/code>, we must have a unary expression. In that case, we\ngrab the token, and then recursively call <code>unary()<\/code> again to the parse the\noperand. Wrap that all up in a unary expression syntax tree and we\u2019re done.<\/p><aside><p>The fact that the parser looks ahead at upcoming tokens to decide how to parse\nputs recursive descent into the category of <strong>predictive parsers<\/strong>.<\/p>\n<\/aside><p>Otherwise, we must have reached the highest level of precedence, primary\nexpressions.<\/p><div class=\"codehilite\"><pre>primary \u2192 NUMBER | STRING | \"false\" | \"true\" | \"nil\"\n        | \"(\" expression \")\" ;\n<\/pre><\/div><p>Most of the cases for the rule are single terminals, so it\u2019s pretty\nstraightforward:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>unary<\/em>()<\/p><pre><br \/>private Expr primary() {                                 \n    if (match(FALSE)) return new Expr.Literal(false);      \n    if (match(TRUE)) return new Expr.Literal(true);        \n    if (match(NIL)) return new Expr.Literal(null);\n    if (match(NUMBER, STRING)) {                           \n      return new Expr.Literal(previous().literal);         \n    }                                                      \n    if (match(LEFT_PAREN)) {                               \n      Expr expr = expression();                            \n      consume(RIGHT_PAREN, \"Expect ')' after expression.\");\n      return new Expr.Grouping(expr);                      \n    }                                                      \n  }                                                        \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>unary<\/em>()<\/p><p>The interesting branch is the one for handling parentheses. After we match an\nopening <code>(<\/code> and parse the expression inside it, we <em>must<\/em> find a <code>)<\/code> token. If\nwe don\u2019t, that\u2019s an error.<\/p><p>A parser really has two jobs:<\/p><ol><li>\n<p>Given a valid sequence of tokens, produce a corresponding syntax tree.<\/p>\n<\/li>\n<li>\n<p>Given an <em>invalid<\/em> sequence of tokens, detect any errors and tell the\n    user about their mistakes.<\/p>\n<\/li>\n<\/ol><p>Don\u2019t underestimate how important the second job is! In modern IDEs and editors,\nthe parser is constantly reparsing code\u2014often while the user is still editing\nit\u2014in order to syntax highlight and support things like auto-complete. That\nmeans it will encounter code in incomplete, half-wrong states <em>all the time.<\/em><\/p><p>When the user doesn\u2019t realize the syntax is wrong, it is up to the parser to\nhelp guide them back onto the right path. The way it reports errors is a large\npart of your language\u2019s user interface. Good syntax error handling is hard. By\ndefinition, the code isn\u2019t in a well-defined state, so there\u2019s no infallible way\nto know what the user <em>meant<\/em> to write. The parser can\u2019t read your mind.<\/p><aside><p>Not yet at least. With the way things are going in machine learning these days,\nwho knows what the future will bring?<\/p>\n<\/aside><p>There are a couple of hard requirements for when the parser runs into a syntax\nerror:<\/p><ul><li>\n<p><strong>It must detect and report the error.<\/strong> If it doesn\u2019t detect the error and passes the resulting malformed syntax tree on\n    to the interpreter, all manner of horrors may be summoned.<\/p>\n<aside><p>Philosophically speaking, if an error isn\u2019t detected and the interpreter\nruns the code, is it <em>really<\/em> an error?<\/p>\n<\/aside><\/li>\n<li>\n<p><strong>It must not crash or hang.<\/strong> Syntax errors are a fact of life and language\n    tools have to be robust in the face of them. Segfaulting or getting stuck in\n    an infinite loop isn\u2019t allowed. While the source may not be valid <em>code<\/em>,\n    it\u2019s still a valid <em>input to the parser<\/em> because users use the parser to\n    learn what syntax is allowed.<\/p>\n<\/li>\n<\/ul><p>Those are the table stakes if you want to get in the parser game at all, but you\nreally want to raise the ante beyond that. A decent parser should:<\/p><ul><li>\n<p><strong>Be fast.<\/strong> Computers are thousands of times faster than when parser\n    technology was first invented. The days of needing to optimize your parser\n    so that it could get through an entire source file during a coffee break are\n    over. But programmer expectations have risen as quickly, if not faster. They\n    expect their editors to reparse files in milliseconds after every keystroke.<\/p>\n<\/li>\n<li>\n<p><strong>Report as many distinct errors as there are.<\/strong> Aborting after the first\n    error is easy to implement, but it\u2019s annoying for users if every time they\n    fix what they think is the one error in a file, a new one appears. They\n    want to see them all.<\/p>\n<\/li>\n<li>\n<p><strong>Minimize <em>cascaded<\/em> errors.<\/strong> Once a single error is found, the parser no\n    longer really knows what\u2019s going on. It tries to get itself back on track\n    and keep going, but if it gets confused, it may report a slew of ghost\n    errors that don\u2019t indicate other real problems in the code. When the first\n    error is fixed, they disappear, because they merely represent the parser\u2019s\n    own confusion. These are annoying because they can scare the user into\n    thinking their code is in a worse state than it is.<\/p>\n<\/li>\n<\/ul><p>The last two points are in tension. We want to report as many separate errors as\nwe can, but we don\u2019t want to report ones that are merely side effects of an\nearlier one.<\/p><p>The way a parser responds to an error and keeps going to look for later errors\nis called <strong>\u201cerror recovery\u201d<\/strong>. It was a hot research topic in the 60s. Back\nthen, you\u2019d hand a stack of punch cards to the secretary and come back the next\nday to see if the compiler succeeded. With an iteration loop that slow, you\n<em>really<\/em> wanted to find every single error in your code in one pass.<\/p><p>Today, when parsers complete before you\u2019ve even finished typing, it\u2019s less of an\nissue. Simple, fast error recovery is fine.<\/p><h3><a href=\"#panic-mode-error-recovery\" name=\"panic-mode-error-recovery\"><small>6\u200a.\u200a3\u200a.\u200a1<\/small> Panic mode error recovery<\/a><\/h3><aside><p>You know you want to push it.<\/p>\n<p><img alt=\"A big shiny &quot;PANIC&quot; button.\" src=\"http:\/\/www.craftinginterpreters.com\/image\/parsing-expressions\/panic.png\" \/><\/p>\n<\/aside><p>Of all the recovery techniques devised in yesteryear, the one that best stood\nthe test of time is called\u2014somewhat alarmingly\u2014\u201cpanic\nmode\u201d. As soon as the parser detects an error, it enters panic mode. It\nknows at least one token doesn\u2019t make sense given its current state in the\nmiddle of some stack of grammar productions.<\/p><p>Before it can get back to parsing, it needs to get its state and the sequence of\nforthcoming tokens aligned such that the next token does match the rule being\nparsed. This process is called <strong>synchronization<\/strong>.<\/p><p>To do that, we select some rule in the grammar that will mark the\nsynchronization point. The parser fixes its parsing state by jumping out of any\nnested productions until it gets back to that rule. Then it synchronizes the\ntoken stream by discarding tokens until it reaches one that can appear at that\npoint in the rule.<\/p><p>Any additional real syntax errors hiding in those discarded tokens aren\u2019t\nreported, but it also means that any mistaken cascaded errors that are side\neffects of the initial error aren\u2019t <em>falsely<\/em> reported either, which is a decent\ntrade-off.<\/p><p>The traditional place in the grammar to synchronize is between statements. We\ndon\u2019t have those yet, so we won\u2019t actually synchronize in this chapter, but\nwe\u2019ll get the machinery in place for later.<\/p><h3><a href=\"#entering-panic-mode\" name=\"entering-panic-mode\"><small>6\u200a.\u200a3\u200a.\u200a2<\/small> Entering panic mode<\/a><\/h3><p>Back before we went on this side trek about error recovery, we were writing the\ncode to parse a parenthesized expression. After parsing the expression, it\nlooks for the closing <code>)<\/code> by calling <code>consume()<\/code>. Here, finally, is that method:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>match<\/em>()<\/p><pre>  private Token consume(TokenType type, String message) {\n    if (check(type)) return advance();\n    throw error(peek(), message);                        \n  }                                                      \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>match<\/em>()<\/p><p>It\u2019s similar to <code>match()<\/code> in that it checks to see if the next token is of the\nexpected type. If so, it consumes it and everything is groovy. If some other\ntoken is there, then we\u2019ve hit an error. We report it by calling this:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>previous<\/em>()<\/p><pre>  private ParseError error(Token token, String message) {\n    Lox.error(token, message);                           \n    return new ParseError();                             \n  }                                                      \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>previous<\/em>()<\/p><p>First, that shows the error to the user by calling:<\/p><div class=\"codehilite\"><p><em>lox\/Lox.java<\/em><br \/>add after <em>report<\/em>()<\/p><pre>  static void error(Token token, String message) {              \n    if (token.type == TokenType.EOF) {                          \n      report(token.line, \" at end\", message);                   \n    } else {                                                    \n      report(token.line, \" at '\" + token.lexeme + \"'\", message);\n    }                                                           \n  }                                                             \n<\/pre><\/div><p><em>lox\/Lox.java<\/em>, add after <em>report<\/em>()<\/p><p>This reports an error at a given token. It shows the token\u2019s location and the\ntoken itself. This will come in handy later since we use tokens throughout the\ninterpreter to track locations in code.<\/p><p>After this is called, the user knows about the syntax error, but what does the\n<em>parser<\/em> do next? Back in <code>error()<\/code>, it creates and returns a ParseError, an\ninstance of:<\/p><div class=\"codehilite\"><pre class=\"insert-before\">class Parser {                                               \n<\/pre><p><em>lox\/Parser.java<\/em><br \/>nest inside class <em>Parser<\/em><\/p><pre class=\"insert\">  private static class ParseError extends RuntimeException {}\n<br \/><\/pre><pre class=\"insert-after\">  private final List&lt;Token&gt; tokens;                          \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, nest inside class <em>Parser<\/em><\/p><p>This is a simple sentinel class we use to unwind the parser. The <code>error()<\/code>\nmethod <em>returns<\/em> it instead of <em>throwing<\/em> because we want to let the caller\ndecide whether to unwind or not.<\/p><p>Some parse errors occur in places where the parser isn\u2019t likely to get into a\nweird state and we don\u2019t need to synchronize. In\nthose places, we simply report the error and keep on truckin\u2019. For example, Lox\nlimits the number of arguments you can pass to a function. If you pass too many,\nthe parser needs to report that error, but it can and should simply keep on\nparsing the extra arguments instead of freaking out and going into panic mode.<\/p><aside><p>Another way to handle common syntax errors is with <strong>error productions<\/strong>. You\naugment the grammar with a rule that matches the erroneous syntax. The parser\nsafely parses it but then reports it as an error instead of producing a syntax\ntree.<\/p>\n<p>For example, some languages have an unary <code>+<\/code> operator, like <code>+123<\/code>, but Lox\ndoes not. Instead of getting confused when the parser stumbles onto a <code>+<\/code> at the\nbeginning of an expression, we could extend the unary rule to allow it:<\/p>\n<div class=\"codehilite\"><pre>unary \u2192 ( \"!\" | \"-\" | \"+\" ) unary\n        | primary ;\n<\/pre><\/div>\n<p>This lets the parser consume <code>+<\/code> without going into panic mode or leaving the\nparser in a weird state.<\/p>\n<p>Error productions work well because you, the parser author, know <em>how<\/em> the code\nis wrong and what the user was likely trying to do. That means you can give a\nmore helpful message to get the user back on track, like, \u201cUnary \u2018+\u2019 expressions\nare not supported.\u201d Mature parsers tend to accumulate error productions like\nbarnacles since they help users fix common mistakes.<\/p>\n<\/aside><p>In our case, though, the syntax error is nasty enough that we want to panic and\nsynchronize. Discarding tokens is pretty easy, but how do we synchronize the\nparser\u2019s own state?<\/p><h3><a href=\"#synchronizing-a-recursive-descent-parser\" name=\"synchronizing-a-recursive-descent-parser\"><small>6\u200a.\u200a3\u200a.\u200a3<\/small> Synchronizing a recursive descent parser<\/a><\/h3><p>With recursive descent, the parser\u2019s state\u2014which rules it is in the middle of\nrecognizing\u2014is not stored explicitly in fields. Instead, we use Java\u2019s\nown callstack to track what the parser is doing. Each rule in the process of\nbeing parsed is a callframe on the stack. In order to reset that state, we need\nto clear out those callframes.<\/p><p>The natural way to do that in Java is exceptions. When we want to synchronize,\nwe <em>throw<\/em> that ParseError object. Higher up in the method for the grammar rule\nwe are synchronizing to, we\u2019ll catch it. Since we are synchronizing on statement\nboundaries, we\u2019ll catch the exception there. After the exception is caught, the\nparser is in the right state. All that\u2019s left is to synchronize the tokens.<\/p><p>We want to discard tokens until we\u2019re right at the beginning of the next\nstatement. That boundary is pretty easy to spot\u2014it\u2019s one of the main reasons\nwe picked it. <em>After<\/em> a semicolon, we\u2019re probably\nfinished with a statement. Most statements start with a keyword\u2014<code>for<\/code>, <code>if<\/code>,\n<code>return<\/code>, <code>var<\/code>, etc. When the <em>next<\/em> token is any of those, we\u2019re probably\nabout to start a statement.<\/p><aside><p>I say \u201cprobably\u201d because we could hit a semicolon separating clauses in a for\nloop. Our synchronization isn\u2019t perfect, but that\u2019s OK. We\u2019ve already reported\nthe first error precisely, so everything after that is kind of \u201cbest effort\u201d.<\/p>\n<\/aside><p>This method encapsulates that logic:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>error<\/em>()<\/p><pre>  private void synchronize() {                 \n    advance();\n    while (!isAtEnd()) {                       \n      if (previous().type == SEMICOLON) return;\n      switch (peek().type) {                   \n        case CLASS:                            \n        case FUN:                              \n        case VAR:                              \n        case FOR:                              \n        case IF:                               \n        case WHILE:                            \n        case PRINT:                            \n        case RETURN:                           \n          return;                              \n      }                                        \n      advance();                               \n    }                                          \n  }                                            \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>error<\/em>()<\/p><p>It discards tokens until it thinks it found a statement boundary. After catching\na ParseError, we\u2019ll call this and then we are hopefully back in sync. When it\nworks well, we have discarded tokens that would have likely caused cascaded\nerrors anyway and now we can parse the rest of the file starting at the next\nstatement.<\/p><p>Alas, we don\u2019t get to see this method in action, since we don\u2019t have statements\nyet. We\u2019ll get to that <a href=\"http:\/\/www.craftinginterpreters.com\/statements-and-state.html\">in a couple of chapters<\/a>. For now, if an\nerror occurs, we\u2019ll panic and unwind all the way to the top and stop parsing.\nSince we can only parse a single expression anyway, that\u2019s no big loss.<\/p><p>We are mostly done parsing expressions now. There is one other place where we\nneed to add a little error handling. As the parser descends through the parsing\nmethods for each grammar rule, it eventually hits <code>primary()<\/code>. If none of the\ncases in there match, it means we are sitting on a token that can\u2019t start an\nexpression. We need to handle that error too:<\/p><div class=\"codehilite\"><pre class=\"insert-before\">    if (match(LEFT_PAREN)) {                               \n      Expr expr = expression();                            \n      consume(RIGHT_PAREN, \"Expect ')' after expression.\");\n      return new Expr.Grouping(expr);                      \n    }                                                      \n<\/pre><p><em>lox\/Parser.java<\/em><br \/>in <em>primary<\/em>()<\/p><pre class=\"insert\"><br \/>throw error(peek(), \"Expect expression.\");             \n<\/pre><pre class=\"insert-after\">  }                                                        \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, in <em>primary<\/em>()<\/p><p>With that, all that remains in the parser is to define an initial method to kick\nit off. It\u2019s called, naturally enough, <code>parse()<\/code>:<\/p><div class=\"codehilite\"><p><em>lox\/Parser.java<\/em><br \/>add after <em>Parser<\/em>()<\/p><pre>  Expr parse() {                \n    try {                       \n      return expression();      \n    } catch (ParseError error) {\n      return null;              \n    }                           \n  }                             \n<\/pre><\/div><p><em>lox\/Parser.java<\/em>, add after <em>Parser<\/em>()<\/p><p>We\u2019ll revisit this method later when we add statements to the language. For now,\nit parses a single expression and returns it. We also have some temporary code\nto exit out of panic mode. Syntax error recovery is the parser\u2019s job, so we\ndon\u2019t want the ParseError exception to escape into the rest of the interpreter.<\/p><p>When a syntax error does occur, this method returns <code>null<\/code>. That\u2019s OK. The\nparser promises not to crash or hang on invalid syntax, but it doesn\u2019t promise\nto return a <em>usable syntax tree<\/em> if an error is found. As soon as the parser\nreports an error, <code>hadError<\/code> gets set, and subsequent phases are skipped.<\/p><p>Finally, we can hook up our brand new parser to the main Lox class and try it\nout. We still don\u2019t have an interpreter so, for now, we\u2019ll parse to a syntax\ntree and then use the AstPrinter class from the <a href=\"http:\/\/www.craftinginterpreters.com\/representing-code.html#a-(not-very)-pretty-printer\">last chapter<\/a> to\ndisplay it.<\/p><p>Delete the old code to print the scanned tokens and replace it with this:<\/p><div class=\"codehilite\"><pre class=\"insert-before\">    List&lt;Token&gt; tokens = scanner.scanTokens();             \n<\/pre><p><em>lox\/Lox.java<\/em><br \/>in <em>run<\/em>()<br \/>replace 5 lines<\/p><pre class=\"insert\">    Parser parser = new Parser(tokens);                    \n    Expr expression = parser.parse();\n    \/\/ Stop if there was a syntax error.                   \n    if (hadError) return;                                  \n    System.out.println(new AstPrinter().print(expression));\n<\/pre><pre class=\"insert-after\">  }                                                        \n<\/pre><\/div><p><em>lox\/Lox.java<\/em>, in <em>run<\/em>(), replace 5 lines<\/p><p>Congratulations, you have crossed the threshold! That\nreally is all there is to hand-writing a parser. We\u2019ll extend the grammar in\nlater chapters with assignment, statements, and other stuff, but none of that is\nany more complex than the binary operators we tackled here.<\/p><aside><p>It is possible to define a grammar that\u2019s more difficult than Lox\u2019s to parse\nusing recursive descent. Predictive parsing gets tricky when you may need to\nlook ahead a large number of tokens to figure out what you\u2019re sitting on.<\/p>\n<p>In practice, most languages are designed to avoid that. Even in cases where they\naren\u2019t, you can usually hack around it without too much pain. If you can parse\nC++ using recursive descent, you can parse anything.<\/p>\n<\/aside><p>Fire up the interpreter and type in some expressions. See how it handles\nprecedence and associativity correctly? Not bad for less than 200 lines of code.<\/p><div class=\"challenges\"><ol><li>\n<p>In C, a block is a statement form that allows you to pack a series of\n    statements where a single one is expected. The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Comma_operator\">comma operator<\/a> is an\n    analogous syntax for expressions. A comma-separated series of expressions\n    can be given where a single expression is expected (except inside a function\n    call\u2019s argument list). At runtime, the comma operator evaluates the left\n    operand and discards the result. Then it evaluates and returns the right\n    operand.<\/p>\n<p>Add support for comma expressions. Give them the same precedence and\nassociativity as in C. Write the grammar, and then implement the necessary\nparsing code.<\/p>\n<\/li>\n<li>\n<p>Likewise, add support for the C-style conditional or \u201cternary\u201d operator\n    <code>?:<\/code>. What precedence level is allowed between the <code>?<\/code> and <code>:<\/code>? Is the whole\n    operator left-associative or right-associative?<\/p>\n<\/li>\n<li>\n<p>Add error productions to handle each binary operator appearing without a\n    left-hand operand. In other words, detect a binary operator appearing at the\n    beginning of an expression. Report that as an error, but also parse and\n    discard a right-hand operand with the appropriate precedence.<\/p>\n<\/li>\n<\/ol><\/div><div class=\"design-note\"><p>Let\u2019s say we decide to add bitwise <code>&amp;<\/code> and <code>|<\/code> operators to Lox. Where should we\nput them in the precedence hierarchy? C\u2014and most languages that follow in C\u2019s\nfootsteps\u2014place them below <code>==<\/code>. This is widely considered a mistake because\nit means common operations like testing a flag require parentheses:<\/p><div class=\"codehilite\"><pre>if (flags &amp; FLAG_MASK == SOME_FLAG) { ... } \/\/ Wrong.\nif ((flags &amp; FLAG_MASK) == SOME_FLAG) { ... } \/\/ Right.\n<\/pre><\/div><p>Should we fix this for Lox and put bitwise operators higher up the precedence\ntable than C does? There are two strategies we can take.<\/p><p>You almost never want to use the result of an <code>==<\/code> expression as the operand to\na bitwise operator. By making bitwise bind tighter, users don\u2019t need to\nparenthesize as often. So if we do that, and users assume the precedence is\nchosen logically to minimize parentheses, they\u2019re likely to infer it correctly.<\/p><p>This kind of internal consistency makes the language easier to learn because\nthere are fewer edge cases and exceptions users have to stumble into and then\ncorrect. That\u2019s good, because before users can use our language, they have to\nload all of that syntax and semantics into their heads. A simpler, more rational\nlanguage <em>makes sense<\/em>.<\/p><p>But, for many users there is an even faster shortcut to getting our language\u2019s\nideas into their wetware\u2014<em>use concepts they already know.<\/em> Many newcomers to\nour language will be coming from some other language or languages. If our\nlanguage uses some of the same syntax or semantics as those, there is much less\nfor the user to learn (and <em>unlearn<\/em>).<\/p><p>This is particularly helpful with syntax. You may not remember it well today,\nbut way back when you learned your very first programming language, code\nprobably looked alien and unapproachable. Only through painstaking effort did\nyou learn to read and accept it. If you design a novel syntax for your new\nlanguage, you force users to start that process all over again.<\/p><p>Taking advantage of what users already know is one of the most powerful tools\nyou can use to ease adoption of your language. It\u2019s almost impossible to\nunderestimate how useful this is. But it faces you with a nasty problem: What\nhappens when the thing the users all know <em>kind of sucks?<\/em> C\u2019s bitwise operator\nprecedence is a mistake that doesn\u2019t make sense. But it\u2019s a <em>familiar<\/em> mistake\nthat millions have already gotten used to and learned to live with.<\/p><p>Do you stay true to your language\u2019s own internal logic and ignore history? Do\nyou start from a blank slate and first principles? Or do you weave your language\ninto the rich tapestry of programming history and give your users a leg up by\nstarting from something they already know?<\/p><p>There is no perfect answer here, only trade-offs. You and I are obviously biased\ntowards liking novel languages, so our natural inclination is to burn the\nhistory books and start our own story.<\/p><p>In practice, it\u2019s often better to make the most of what users already know.\nGetting them to come to your language requires a big leap. The smaller you can\nmake that chasm, the more people will be willing to cross it. But you can\u2019t\n<em>always<\/em> stick to history, or your language won\u2019t have anything new and\ncompelling to get people to <em>want<\/em> to jump over.<\/p><\/div><footer><a href=\"http:\/\/www.craftinginterpreters.com\/evaluating-expressions.html\" class=\"next\">\n  Next Chapter: \u201cEvaluating Expressions\u201d \u2192\n<\/a>\nHand-crafted by Robert Nystrom\u2002\u2014\u2002<a href=\"https:\/\/github.com\/munificent\/craftinginterpreters\/blob\/master\/LICENSE\" target=\"_blank\">\u00a9 2015&amp;hairsp;\u2013&amp;hairsp;2018<\/a>\n<\/footer>","created_at":"2018-11-07T16:50:29+0000","updated_at":"2018-11-07T16:50:29+0000","annotations":[],"mimetype":"text\/html","reading_time":32,"domain_name":"www.craftinginterpreters.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":509,"title":"Terminal Multiplexers: Screen, Tmux - Hyperpolyglot","url":"http:\/\/hyperpolyglot.org\/multiplexers","content":"<p><a href=\"#cmd-line-opts\">command line options<\/a> | <a href=\"#key-bindings\">key bindings<\/a> | <a href=\"#customization\">customization<\/a> | <a href=\"#terminology\">terminology<\/a> | <a href=\"#doc\">documentation<\/a><\/p><h2 id=\"toc0\">Command Line Options<\/h2><table class=\"wiki-content-table\"><tr><th>screen<\/th>\n<th>tmux<\/th>\n<\/tr><tr><td>create session and attach<\/td>\n<td>$ screen<\/td>\n<td>$ tmux<\/td>\n<\/tr><tr><td>create session <em>foo<\/em> and attach<\/td>\n<td>$ screen -S foo<\/td>\n<td>$ tmux new -s foo<\/td>\n<\/tr><tr><td>create detached session <em>foo<\/em><\/td>\n<td>$ screen -S foo -d -m<\/td>\n<td>$ tmux new -s foo -d<\/td>\n<\/tr><tr><td>list sessions<\/td>\n<td>$ screen -list<\/td>\n<td>$ tmux ls<\/td>\n<\/tr><tr><td>attach<\/td>\n<td>$ screen -r<\/td>\n<td>$ tmux attach<\/td>\n<\/tr><tr><td>attach to session <em>foo<\/em><\/td>\n<td>$ screen -r foo<\/td>\n<td>$ tmux attach -t foo<\/td>\n<\/tr><tr><td>attach to session by pid<\/td>\n<td>$ screen -r <em>pid<\/em><\/td>\n<\/tr><tr><td>kill session <em>foo<\/em><\/td>\n<td>$ screen -r foo -X quit<\/td>\n<td>$ tmux kill-session -t foo<\/td>\n<\/tr><tr><td>send multiplexer command to session <em>foo<\/em><\/td>\n<td>$ screen -r foo -X <em>command<\/em><\/td>\n<td>$ tmux <em>command<\/em> -t foo<\/td>\n<\/tr><tr><td>run ls in session <em>foo<\/em><\/td>\n<td>$ screen -r foo -X stuff \"ls $(echo -ne '\\015')\"<\/td>\n<td>$ tmux send-keys -t foo 'ls' C-m<\/td>\n<\/tr><tr><td>run vi in new window<\/td>\n<td>$ screen vi \/etc\/motd<\/td>\n<td>$ tmux new-window vi \/etc\/motd<\/td>\n<\/tr><\/table><h2 id=\"toc1\">Key Bindings<\/h2><p><a href=\"#sessions\">sessions<\/a> | <a href=\"#windows\">windows<\/a> | <a href=\"#regions\">regions<\/a> | <a href=\"#panes\">panes<\/a> | <a href=\"#paste-buffer\">paste buffer<\/a> | <a href=\"#copy-mode\">copy mode<\/a><\/p><table class=\"wiki-content-table\"><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th>C-a <em>cmd<\/em><\/th>\n<th>command<\/th>\n<th>C-b <em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>help<\/td>\n<td>?<\/td>\n<td>:help<\/td>\n<td>?<\/td>\n<td>:list-keys<\/td>\n<\/tr><tr><td>send prefix to app<\/td>\n<td>a<\/td>\n<td>:meta<\/td>\n<td>C-b<\/td>\n<td>:send-prefix<\/td>\n<\/tr><tr><td>suspend multiplexer<\/td>\n<td>C-z<\/td>\n<td>:suspend<\/td>\n<td>C-z<\/td>\n<td>:suspend-client<\/td>\n<\/tr><tr><td>show previous multiplexer message<\/td>\n<td>m<br \/>C-m<\/td>\n<td>:lastmsg<\/td>\n<td>~<\/td>\n<td>:show-messages<\/td>\n<\/tr><tr><td>source file<\/td>\n<td>:source <em>file<\/em><\/td>\n<td>:source-file <em>file<\/em><\/td>\n<\/tr><tr><td>detach<\/td>\n<td>d<br \/>C-d<\/td>\n<td>:detach<\/td>\n<td>d<\/td>\n<td>:detach-client<\/td>\n<\/tr><tr><th colspan=\"5\">sessions<\/th>\n<\/tr><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th>C-a <em>cmd<\/em><\/th>\n<th>command<\/th>\n<th>C-b <em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>new session<\/td>\n<td>:new<\/td>\n<\/tr><tr><td>new named session<\/td>\n<td>:new -s foo<\/td>\n<\/tr><tr><td>switch session<\/td>\n<td>s<\/td>\n<td>:choose-session<\/td>\n<\/tr><tr><td>rename session<\/td>\n<td>:sessionname foo<\/td>\n<td>$<\/td>\n<td>command-prompt -I #S \"rename-session '%%'\"<\/td>\n<\/tr><tr><td>kill session<\/td>\n<td>C-\\<\/td>\n<td>:quit<\/td>\n<\/tr><tr><th colspan=\"5\">windows<\/th>\n<\/tr><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th>C-a <em>cmd<\/em><\/th>\n<th>command<\/th>\n<th>C-b <em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>create new window<\/td>\n<td>c<br \/>C-c<\/td>\n<td>:screen<\/td>\n<td>c<\/td>\n<td>:new-window<\/td>\n<\/tr><tr><td>switch to next window<\/td>\n<td>n<br \/>C-n<br \/>SPACE<\/td>\n<td>:next<\/td>\n<td>n<\/td>\n<td>:next-window<\/td>\n<\/tr><tr><td>switch to previous window<\/td>\n<td>BACKSPACE<br \/>h<br \/>p<br \/>C-p<\/td>\n<td>:prev<\/td>\n<td>p<\/td>\n<td>:previous-window<\/td>\n<\/tr><tr><td>toggle to last window<\/td>\n<td>C-a<\/td>\n<td>:other<\/td>\n<\/tr><tr><td>select window <em>n<\/em><\/td>\n<td><em>n<\/em><\/td>\n<td>:select-window -t :<em>n<\/em><\/td>\n<\/tr><tr><td>list windows<\/td>\n<td>w<br \/>C-w<\/td>\n<td>:windows<\/td>\n<td>:list-windows<\/td>\n<\/tr><tr><td>show current window number and name<\/td>\n<td>N<\/td>\n<td>:number<\/td>\n<\/tr><tr><td>renumber current window<\/td>\n<td>. <em>position<\/em><\/td>\n<td>:move-window<\/td>\n<\/tr><tr><td>move current window to another session<\/td>\n<td>. <em>sessname<\/em><br \/>. <em>sessnum<\/em>:<em>position<\/em><\/td>\n<\/tr><tr><td>redraw current window<\/td>\n<td>l<br \/>C-l<\/td>\n<td>:redisplay<\/td>\n<td>r<\/td>\n<td>:refresh-client<\/td>\n<\/tr><tr><td>choose window interactively<\/td>\n<td>\"<\/td>\n<td>:windowlist -b<\/td>\n<td>w<\/td>\n<td>:choose-window<\/td>\n<\/tr><tr><td>rename window<\/td>\n<td>A<\/td>\n<td>:title<\/td>\n<td>,<\/td>\n<\/tr><tr><td>select window <em>foo<\/em><\/td>\n<td>'<\/td>\n<td>:select<\/td>\n<td>f <em>foo<\/em><\/td>\n<\/tr><tr><td>close current window<\/td>\n<td>C-k<\/td>\n<td>:kill<\/td>\n<td>&amp;<\/td>\n<td>:kill-window<\/td>\n<\/tr><tr><td>join window 1 to current window<\/td>\n<td>:join-pane -s 1<\/td>\n<\/tr><tr><td>join region 0 of window 1 to current window<\/td>\n<td>:join-pane -s 1.0<\/td>\n<\/tr><tr><th colspan=\"5\">regions<\/th>\n<\/tr><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th>C-a <em>cmd<\/em><\/th>\n<th>command<\/th>\n<th>C-b <em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>split into top and bottom regions<\/td>\n<td>S<\/td>\n<td>:split<\/td>\n<\/tr><tr><td>move down to next region<\/td>\n<td>TAB<\/td>\n<td>:focus<\/td>\n<\/tr><tr><td>make regions same height<\/td>\n<td>:resize =<\/td>\n<\/tr><tr><td>close current region<\/td>\n<td>X<\/td>\n<td>:remove<\/td>\n<\/tr><tr><td>close all but current region<\/td>\n<td>Q<\/td>\n<td>:only<\/td>\n<\/tr><tr><td>clear current region<\/td>\n<td>C<\/td>\n<td>:clear<\/td>\n<\/tr><tr><td>log region to file<\/td>\n<td><em>file is<\/em> screenlog.NN<br \/>:log<\/td>\n<\/tr><tr><td>turn off logging<\/td>\n<td>:log off<\/td>\n<\/tr><tr><td>make current region <em>n<\/em> rows taller\/shorter<\/td>\n<td>:resize +<em>n<\/em><br \/>:resize -<em>n<\/em><\/td>\n<\/tr><tr><td>make current region <em>n<\/em> rows tall<\/td>\n<td>:resize <em>n<\/em><\/td>\n<\/tr><tr><th colspan=\"5\">panes<\/th>\n<\/tr><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th>C-a <em>cmd<\/em><\/th>\n<th>command<\/th>\n<th>C-b <em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>split into left and right panes<\/td>\n<td>%<\/td>\n<td>:split-window -h<\/td>\n<\/tr><tr><td>split into top and bottom panes<\/td>\n<td>\"<\/td>\n<td>:split-window<\/td>\n<\/tr><tr><td>switch to next panes<\/td>\n<td>o<\/td>\n<td>:select-pane<\/td>\n<\/tr><tr><td>rotate panes<\/td>\n<td>C-o<\/td>\n<td>:rotate-window<\/td>\n<\/tr><tr><td>reverse rotate panes<\/td>\n<td>M-o<\/td>\n<td>:rotate-window -D<\/td>\n<\/tr><tr><td>arrange panes side-by-side and same width<\/td>\n<td>M-1<\/td>\n<\/tr><tr><td>arrange panes stacked and same height<\/td>\n<td>M-2<\/td>\n<\/tr><tr><td>swap current and previous pane<\/td>\n<td>{<\/td>\n<td>:swap-pane -U<\/td>\n<\/tr><tr><td>swap current and next pane<\/td>\n<td>}<\/td>\n<td>:swap-pane -D<\/td>\n<\/tr><tr><td>change arrangement of panes<\/td>\n<td>SPACE<\/td>\n<td>:next-layout<\/td>\n<\/tr><tr><td>close current pane<\/td>\n<td>x<\/td>\n<td>:confirm-before kill-pane<\/td>\n<\/tr><tr><td>break current pane into separate window<\/td>\n<td>!<\/td>\n<td>:break-pane<\/td>\n<\/tr><tr><td>list panes<\/td>\n<td>:list-panes<\/td>\n<\/tr><tr><td>display pane numbers<\/td>\n<td>q<\/td>\n<td>:display-panes<\/td>\n<\/tr><tr><td>clear current pane<\/td>\n<td>:clear-history<\/td>\n<\/tr><tr><td>log pane to file<\/td>\n<td>:pipe-pane \"cat &gt; \/tmp\/tmux.log\"<\/td>\n<\/tr><tr><td>turn off logging<\/td>\n<td>:pipe-pane<\/td>\n<\/tr><tr><td>resize pane left\/up <em>n<\/em> cells<\/td>\n<td>:resize-pane -L <em>n<\/em><br \/>:resize-pane -U <em>n<\/em><\/td>\n<\/tr><tr><th colspan=\"5\">paste buffer<\/th>\n<\/tr><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th>C-a <em>cmd<\/em><\/th>\n<th>command<\/th>\n<th>C-b <em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>enter copy mode<\/td>\n<td>[<br \/>C-[<br \/>ESC<\/td>\n<td>:copy<\/td>\n<td>[<\/td>\n<td>:copy-mode<\/td>\n<\/tr><tr><td>paste most recent buffer<\/td>\n<td>]<\/td>\n<td>]<\/td>\n<td>:paste-buffer<\/td>\n<\/tr><tr><td>list buffers<\/td>\n<td><em>only one buffer<\/em><\/td>\n<td>#<\/td>\n<td>:list-buffers<\/td>\n<\/tr><tr><td>choose buffer to paste interactively<\/td>\n<td>=<\/td>\n<td>:choose-buffer<\/td>\n<\/tr><tr><td>write buffer to file<\/td>\n<td><em>writes to<\/em> \/tmp\/screen-exchange:<br \/>&gt;<\/td>\n<td>:writebuf <em>path<\/em><\/td>\n<td>:save-buffer <em>path<\/em><\/td>\n<\/tr><tr><td>copy file to buffer<\/td>\n<td><em>copies from<\/em> \/tmp\/screen-exchange:<br \/>&lt;<\/td>\n<td>:readbuf <em>path<\/em><\/td>\n<td>:load-buffer <em>path<\/em><\/td>\n<\/tr><tr><th colspan=\"5\">copy mode<\/th>\n<\/tr><tr><th colspan=\"2\">screen<\/th>\n<th colspan=\"2\">tmux<\/th>\n<\/tr><tr><th><em>cmd<\/em><\/th>\n<th>command<\/th>\n<th><em>cmd<\/em><\/th>\n<th>command<\/th>\n<\/tr><tr><td>default bindings<\/td>\n<td colspan=\"2\"><em>Vim-style<\/em><\/td>\n<td colspan=\"2\"><em>Emacs-style<\/em><\/td>\n<\/tr><tr><td>set mark<\/td>\n<td>SPACE<\/td>\n<td>C-SPACE<\/td>\n<\/tr><tr><td>copy from mark to point and exit copy mode<\/td>\n<td><em>when mark is set:<\/em><br \/>SPACE<\/td>\n<td>C-w<\/td>\n<\/tr><tr><td>single column movement<\/td>\n<td><em>also left right arrow<\/em><br \/>h l<\/td>\n<td><em>also left right arrow<\/em><br \/>C-b C-h<\/td>\n<\/tr><tr><td>single line movement<\/td>\n<td><em>also down up arrow<\/em><br \/>j k<\/td>\n<td><em>also down up arrow<\/em><br \/>C-n C-p<\/td>\n<\/tr><tr><td>beginning of line<\/td>\n<td>0<\/td>\n<td>C-a<\/td>\n<\/tr><tr><td>end of line<\/td>\n<td>$<\/td>\n<td>C-e<\/td>\n<\/tr><tr><td>forward word<\/td>\n<td>e<\/td>\n<td>M-f<\/td>\n<\/tr><tr><td>backward word<\/td>\n<td>b<\/td>\n<td>M-b<\/td>\n<\/tr><tr><td>page up<\/td>\n<td>C-b<\/td>\n<td>M-v<\/td>\n<\/tr><tr><td>page down<\/td>\n<td>C-f<\/td>\n<td>C-v<\/td>\n<\/tr><tr><td>beginning of buffer<\/td>\n<td>g<\/td>\n<td>M-&lt;<\/td>\n<\/tr><tr><td>end of buffer<\/td>\n<td>G<\/td>\n<td>M-&gt;<\/td>\n<\/tr><tr><td>search backwards<\/td>\n<td>? <em>phrase<\/em><\/td>\n<td>C-r <em>phrase<\/em><\/td>\n<\/tr><tr><td>search forwards<\/td>\n<td>\/ <em>phrase<\/em><\/td>\n<td>C-s <em>phrase<\/em><\/td>\n<\/tr><tr><td>exit copy mode<\/td>\n<td><em>any unbound key also works:<\/em><br \/>ESC<\/td>\n<td>q<\/td>\n<\/tr><\/table><h2 id=\"toc2\">Customization<\/h2><table class=\"wiki-content-table\"><tr><th><a href=\"#screen\">screen<\/a><\/th>\n<th><a href=\"#tmux\">tmux<\/a><\/th>\n<\/tr><tr><td>startup file<\/td>\n<td>~\/.screenrc<\/td>\n<td>~\/.tmux.conf<\/td>\n<\/tr><tr><td>scrollback length<\/td>\n<td>defscrollback 2000<\/td>\n<td>set-option history-limit 2000<\/td>\n<\/tr><tr><td>set prefix<\/td>\n<td><em>how to set prefix to C-b, with a second C-b to send a C-b to the controlling process:<\/em><br \/>escape ^B^B<\/td>\n<td>set-option -g prefix C-a<\/td>\n<\/tr><tr><td>define key binding<\/td>\n<td>bind<\/td>\n<td>bind-key<\/td>\n<\/tr><tr><td>undefine key binding<\/td>\n<td>unbind-key<\/td>\n<\/tr><tr><td>set copy\/scrollback key binding style<\/td>\n<td><em>vi bindings by default.<br \/>When redefining, use vi<br \/>commands on left of equations:<\/em><br \/>markkeys h=^B:l=^F:$=^E<\/td>\n<td><em>emacs by default:<\/em><br \/>setw -g mode-keys vi<\/td>\n<\/tr><tr><td>disable startup message<\/td>\n<td>startup_message off<\/td>\n<\/tr><tr><td>number windows from one<\/td>\n<td><em>0 by default:<\/em><br \/>set -g base-index 1<\/td>\n<\/tr><tr><td>always show status bar<\/td>\n<td><em>splitonly by default:<\/em><br \/>caption always<br \/>caption splitonly<\/td>\n<td><em>on by default:<\/em><br \/>set-option status off<br \/>set-option status on<\/td>\n<\/tr><tr><td>customize caption<\/td>\n<td>caption string \"<em>string<\/em>\"<\/td>\n<td>set-option status-left \"<em>string<\/em>\"<br \/>set-option status-right <em>string<\/em><\/td>\n<\/tr><\/table><table class=\"wiki-content-table\"><tr><th colspan=\"3\">caption escapes<\/th>\n<\/tr><tr><th>screen<\/th>\n<th>tmux<\/th>\n<\/tr><tr><td>shell output<\/td>\n<td><em>first arg is an identifier referenced by the caption string;<br \/>the second and third argument set the refresh in seconds<\/em><p>backtick 1\u00a060\u00a060 <em>cmd<\/em><br \/>caption always \"%1`\"<\/p><\/td>\n<td>#(<em>cmd<\/em>)<\/td>\n<\/tr><tr><td>date (YYYYMMDD)<\/td>\n<td>%Y%m%d<\/td>\n<\/tr><tr><td>month name<\/td>\n<td>%M<\/td>\n<\/tr><tr><td>weekday name<\/td>\n<td>%D<\/td>\n<\/tr><tr><td>24 hr time<\/td>\n<td>%C<\/td>\n<\/tr><tr><td>12 hr time<\/td>\n<td>%c%A<\/td>\n<\/tr><tr><td>color<\/td>\n<td>%{ry}<em>red text, yellow background<\/em>%{dd}foo<\/td>\n<td>#[fg=red,bg=yellow]<em>red text, yellow background<\/em>#[default]<\/td>\n<\/tr><tr><td>fully qualified hostname<\/td>\n<td>#H<\/td>\n<\/tr><tr><td>hostname<\/td>\n<td>%H<\/td>\n<td>#h<\/td>\n<\/tr><tr><td>session name<\/td>\n<td>%S <em>added in 4.1<\/em><\/td>\n<td>#S<\/td>\n<\/tr><tr><td>current window flag<\/td>\n<td>%F<\/td>\n<td>#F<\/td>\n<\/tr><tr><td>current window index<\/td>\n<td>%n<\/td>\n<td>#I<\/td>\n<\/tr><tr><td>current pane index<\/td>\n<td><em>none<\/em><\/td>\n<td>#P<\/td>\n<\/tr><tr><td>current pane title<\/td>\n<td><em>none<\/em><\/td>\n<td>#T<\/td>\n<\/tr><tr><td>window name<\/td>\n<td>%t<\/td>\n<td>#W<\/td>\n<\/tr><tr><td>literal % or #<\/td>\n<td>%%<\/td>\n<td>##<\/td>\n<\/tr><\/table><h2 id=\"toc3\">Terminology<\/h2><p><a href=\"#def-server\">server<\/a> | <a href=\"#def-client\">client<\/a> | <a href=\"#def-session\">session<\/a> | <a href=\"#def-window\">window<\/a> | <a href=\"#def-region\">region<\/a> | <a href=\"#def-pane\">pane<\/a><\/p><p><strong>how ssh works:<\/strong><\/p><p>When a user logs in to a remote host using <code>ssh<\/code>, the <code>ssh<\/code> process contacts an <code>sshd<\/code> process listening on TCP port 22. The <code>sshd<\/code> process opens up a new TCP port and forks off a copy of itself for communicating with the <code>ssh<\/code> process. The new port and child process are for the exclusive use of the connection being established.<\/p><p>The child <code>sshd<\/code> process authenticates the <code>ssh<\/code> process, and if it passes it creates a pseudo-terminal. It then forks the remote user's shell which becomes the controlling process for the pseudo-terminal.<\/p><p>If the network connection is closed, either explicitly by the <code>ssh<\/code> process or because of a loss of network connectivity, the child <code>sshd<\/code> process closes the pseudo-terminal, and this in turn causes the shell to exit.<\/p><p><strong>the SIGHUP problem:<\/strong><\/p><p>If the shell had any process groups running when it exits, they are sent a <code>SIGHUP<\/code> signal followed by a <code>SIGCONT<\/code> signal. By default processes exit when they receive <code>SIGHUP<\/code>. This makes it challenging to run long-running jobs on a remote host when the network connection is unreliable.<\/p><p>A simple solution to the <code>SIGHUP<\/code> problem is to run each job with <code>nohup<\/code>. Optionally, shells such as <code>bash<\/code> and <code>zsh<\/code> have a <code>disown<\/code> built-in which can be used on a process that is already running, should the user have neglected to run it with <code>nohup<\/code>.<\/p><p>The <code>fish<\/code> shell when invoking a process in the background with <code>&amp;<\/code> sets the signal handling state of the process to ignore <code>SIGHUP<\/code>. It will do the same if the process is suspended with <code>^Z<\/code> and then put in the background with <code>bg<\/code>.<\/p><p>Multiplexers offer a solution which protects the shell instead of the job. The user doesn't need to remember to run each job with <code>nohup<\/code>. As an added benefit any state kept by the shell is preserved.<\/p><p><strong>server:<\/strong><\/p><p>The multiplexer server creates pseudo-terminals which are used for running and interacting with programs.<\/p><p><em>Screen<\/em> and <em>Tmux<\/em> servers can create multiple pseudo-terminals. The controlling process for each pseudo-terminal is the user's shell.<\/p><p><strong>client:<\/strong><\/p><p>To see the output of a shell the user must connect to the multiplexer server with a multiplexer client process.<\/p><p>If the multiplexer is being run on a remote machine and the user's connection is lost, the server and its terminals and controlling processes persist, but the client process exits.<\/p><p>When multiple client processes connect to the same server they see the same output. This is a way to share a display across computers.<\/p><p><strong>session:<\/strong><\/p><p>Multiplexers support multiple sessions. Each multiplexer session has its own set of terminals and controlling processes which it is running. The client must choose a session to attach to, and will only be able to see the output of the controlling processes in that session. Sessions can be given names to make it easy for the client to choose the correct session.<\/p><p><em>Screen<\/em> launches a separate server process for each session. <em>Screen<\/em> servers and clients communicate via named pipes.<\/p><p><em>Tmux<\/em> by default will only run one server process per user, and this server process can have multiple sessions. A <em>Tmux<\/em> client and the server communicate via a Unix domain socket in the \/tmp directory. The <code>-L<\/code> option can be used to specify a different socket; a new server is created for each socket.<\/p><p><strong>window:<\/strong><\/p><p>Both <em>Screen<\/em> and <em>Tmux<\/em> have entities which they call windows.<\/p><p>A <em>Screen<\/em> window has a single pseudo-terminal and shell. A <em>Tmux<\/em> window can have multiple pseudo-terminals and shells.<\/p><p><em>Screen<\/em> windows can share the viewport. The <em>Tmux<\/em> viewport can only display one window at a time.<\/p><p>Both <em>Screen<\/em> and <em>Tmux<\/em> windows are numbered starting from zero.<\/p><p><strong>region:<\/strong><\/p><p><em>Screen<\/em> can divide the viewport into multiple regions.<\/p><p><em>Screen<\/em> regions can be empty or they can contain a window. The same window can be displayed in more than one region. When regions share a window their content is identical.<\/p><p><em>Screen<\/em> regions are stacked on top of each other and extend the full width of the window.<\/p><p><strong>pane:<\/strong><\/p><p><em>Tmux<\/em> can divide windows into multiple panes.<\/p><p><em>Tmux<\/em> panes contain a single pseudo-terminal with a shell, and each pseudo-terminal and shell belongs to only one pane.<\/p><p><em>Tmux<\/em> windows can be divided both horizontally and vertically into panes. Each division can be subdivided further.<\/p><p><em>Tmux<\/em> panes are numbered.<\/p><p><em>Tmux<\/em> panes can be moved between windows.<\/p><p><strong>command character (prefix):<\/strong><\/p><p>Multiplexers pass most input on to the shell in the region with focus, but a special command character is used to send commands to the multiplexer.<\/p><p>The default command character in <em>Screen<\/em> is <code>C-a<\/code>. The keystrokes which follow <code>C-a<\/code> are interpreted by <em>Screen<\/em> instead of being passed to the shell.<\/p><p><em>Tmux<\/em> calls the command character the <em>prefix<\/em> and the default value is <code>C-b<\/code>. The keystrokes following the prefix are interpreted by <em>Tmux<\/em> instead of being passed to shell.<\/p><p><strong>scrollback buffer (history):<\/strong><\/p><p><em>Screen<\/em> and <em>Tmux<\/em> keep a history of the output of each shell. The maximum length of the output in lines is configurable.<\/p><p><em>Screen<\/em> calls the history the scrollback buffer.<\/p><p><strong>copy\/scrollback mode (copy mode):<\/strong><\/p><p><em>Screen<\/em> and <em>Tmux<\/em> support two modes for each region. In default mode, input which is not intercepted by the multiplexer is passed to the shell.<\/p><p>When the region is in copy mode the region behaves like a read-only buffer of an editor. The contents are the output of the shell including output that may have scrolled off the top of the region.<\/p><p>The keybindings used by <em>Screen<\/em> in copy\/scrollback mode are Vim-style. It is possible to customize them to be Emacs-style.<\/p><p>The <em>Tmux<\/em> calls copy\/scrollback mode simply copy mode. The keybindings are by default Emacs-style.<\/p><p><strong>paste buffer:<\/strong><\/p><p><em>Screen<\/em> has a single paste buffer.<\/p><p><em>Tmux<\/em> has multiple paste buffers. The <em>Tmux<\/em> paste buffers are numbered; the most recent is number zero. Sessions share a common paste buffer history.<\/p><p><strong>caption (status line):<\/strong><\/p><p>When a <em>Screen<\/em> window is split into multiple regions, a caption line is placed at the bottom of each region. When a window contains a single region, <em>Screen<\/em> by default does not display a caption. The caption, when present, contains information from <em>Screen<\/em>. The information that is displayed can be customized.<\/p><p><em>Tmux<\/em> calls the line at the bottom of a window the status line. By default it is always displayed, though it can be turned off. The status line contains information from <em>Tmux<\/em> which can be customized.<\/p>","created_at":"2018-11-07T16:50:28+0000","updated_at":"2018-11-07T16:50:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"hyperpolyglot.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":510,"title":"Chapter 3. Assessing Community Needs and Resources | Section 14. SWOT Analysis: Strengths, Weaknesses, Opportunities, and Threats | Main Section","url":"https:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/swot-analysis\/main","content":"<table border=\"1px\" style=\"border-color: #CCCCCC;\"><tbody><tr><td>Learn how to conduct a SWOT Analysis to identify situational strengths and weaknesses, as well as opportunities and threats.<\/td>\n    <\/tr><\/tbody><\/table><ul><li>\n<h4>What is a SWOT analysis and why should you use one?<\/h4>\n<\/li>\n<li>\n<h4>When do you use SWOT?<\/h4>\n<\/li>\n<li>\n<h4>What are the elements of a SWOT analysis?<\/h4>\n<\/li>\n<li>\n<h4>How do you create a SWOT analysis?<\/h4>\n<\/li>\n<li>\n<h4>How do you use your SWOT analysis?<\/h4>\n<\/li>\n<\/ul><div style=\"text-align: center;\">\n<p><img alt=\"Photo of SWOT words over map\" class=\"img-responsive\" src=\"http:\/\/ctb.ku.edu\/sites\/default\/files\/chapter_files\/chapter_8_section_14_edit.jpg\" \/><\/p>\n<\/div><p>Change is an inevitable part of community organizing. If you know how to take stock of the strengths, weaknesses, opportunities, and threats, you are more likely to plan and act effectively.<\/p><p>SWOT provides a tool to explore both internal and external factors that may influence your work.<\/p><h2>What is a SWOT analysis and why should you use one?<\/h2><p>SWOT stands for:\u00a0<strong>S<\/strong>trength,\u00a0<strong>W<\/strong>eakness,\u00a0<strong>O<\/strong>pportunity,\u00a0<strong>T<\/strong>hreat. A SWOT analysis guides you to identify your organization\u2019s strengths and weaknesses (S-W), as well as broader opportunities and threats (O-T). Developing a fuller awareness of the situation helps with both strategic planning and decision-making.<\/p><p>The SWOT method was originally developed for business and industry, but it is equally useful in the work of community health and development, education, and even for personal growth.<\/p><p>SWOT is not the only assessment technique you can use. Compare it with\u00a0<a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\">other assessment tools\u00a0in the Community Tool Box<\/a>\u00a0to determine if this is the right approach for your situation.\u00a0The strengths of this method are its simplicity and application to a variety of levels of operation.<\/p><h2>When do you use SWOT?<\/h2><p><strong>A SWOT analysis can offer helpful perspectives at any stage of an effort. You might use it to:<\/strong><\/p><ul><li>Explore possibilities for new efforts or solutions to problems.<\/li>\n<li>Make decisions about the best path for your initiative. Identifying your opportunities for success in context of threats to success can clarify directions and choices.<\/li>\n<li>Determine where change is possible. If you are at a juncture or turning point, an inventory of your strengths and weaknesses can reveal priorities as well as possibilities.<\/li>\n<li>Adjust and refine plans mid-course. A new opportunity might open wider avenues, while a new threat could close a path that once existed.<\/li>\n<\/ul><p>SWOT also offers a simple way of communicating about your initiative or program and an excellent way to organize information you've gathered from studies or surveys.<\/p><h2>What are the elements of a SWOT analysis?<\/h2><p>A SWOT analysis focuses on Strengths, Weaknesses, Opportunities, and Threats.\u00a0<\/p><p>Remember that the purpose of performing a SWOT is to reveal positive forces that work together and potential problems that need to be recognized and possibly addressed.\u00a0<\/p><p>We will discuss the process of creating the analysis below, but first here are a few sample layouts for your SWOT analysis.<\/p><p>Ask participants to answer these simple questions: what are the strengths and weaknesses of your group, community, or effort, and what are the opportunities and threats facing it?<\/p><table border=\"2\" cellpadding=\"5\" style=\"border-color: black; border-spacing: 1px;\"><tbody><tr><td colspan=\"2\" style=\"background-color: #a5a5a5; width: 50%;\"><span><strong>Internal<\/strong><\/span><\/td>\n<td colspan=\"2\" style=\"background-color: #a5a5a5; width: 50%;\"><span><strong>External<\/strong><\/span><\/td>\n<\/tr><tr><td style=\"background-color: #e1e1e1; width: 25%;\">Strengths<\/td>\n<td style=\"background-color: #e1e1e1; width: 25%;\">Weaknesses<\/td>\n<td style=\"background-color: #e1e1e1; width: 25%;\">Opportunities<\/td>\n<td style=\"background-color: #e1e1e1; width: 25%;\">Threats<\/td>\n<\/tr><tr><td colspan=\"2\" style=\"background-color: #f0f0f0;\">\n\n\n\n<\/td>\n<td colspan=\"2\" style=\"background-color: #f0f0f0;\">\n\n\n\n<\/td>\n<\/tr><\/tbody><\/table><p>If a looser structure helps you brainstorm, you can group positives and negatives to think broadly about your organization and its external environment.<\/p><table border=\"2\" cellpadding=\"5\" style=\"border-color: black; border-spacing: 1px;\"><tbody><tr><td style=\"background-color: #a5a5a5;\"><span><strong>Positives<\/strong><\/span><\/td>\n<td style=\"background-color: #a5a5a5;\"><span><strong>Negatives<\/strong><\/span><\/td>\n<\/tr><tr><td style=\"background-color: #f0f0f0;\">\n<ul><li>Strengths<\/li>\n<li>Assets<\/li>\n<li>Resources<\/li>\n<li>Opportunities<\/li>\n<li>Prospects<\/li>\n<\/ul><\/td>\n<td style=\"background-color: #f0f0f0;\">\n<ul><li>Weaknesses<\/li>\n<li>Limitations<\/li>\n<li>Restrictions<\/li>\n<li>Threats<\/li>\n<li>Challenges<\/li>\n<\/ul><\/td>\n<\/tr><\/tbody><\/table><p>Below is a\u00a0third option for structuring your SWOT analysis, which may\u00a0be appropriate for a larger initiative that requires detailed planning. This \"TOWS Matrix\" is adapted from Fred David's <em>Strategic Management<\/em> text.\u00a0<\/p><table border=\"2\" cellpadding=\"5\" style=\"border-color: black; border-spacing: 1px;\"><tbody><tr><td style=\"background-color: #e1e1e1; width: 20%;\"><span><strong>\u00a0<\/strong><\/span><\/td>\n<td style=\"background-color: #a5a5a5; width: 40%;\"><span><strong>STRENGTHS<\/strong><\/span><br \/>1.<br \/>2.<br \/>3.<br \/>4.<\/td>\n<td style=\"background-color: #a5a5a5; width: 40%;\"><span><strong>WEAKNESSES<\/strong><\/span><br \/>1.<br \/>2.<br \/>3.<br \/>4.<\/td>\n<\/tr><tr><td style=\"background-color: #a5a5a5; width: 20%;\"><span><strong>OPPORTUNITIES<\/strong><\/span><br \/>1.<br \/>2.<br \/>3.<br \/>4.<\/td>\n<td style=\"background-color: #f0f0f0; width: 40%;\">Opportunity-Strength (OS) Strategies<br \/>Use the strengths to take advantage of opportunities<br \/>1.<br \/>2.<\/td>\n<td style=\"background-color: #f0f0f0; width: 40%;\">Opportunity-Weakness (OW) Strategies<br \/>Overcome weaknesses by taking advantage of opportunities<br \/>1.<br \/>2.<\/td>\n<\/tr><tr><td style=\"background-color: #a5a5a5; width: 20%;\"><span><strong>THREATS<\/strong><\/span><br \/>1.<br \/>2.<br \/>3.<br \/>4.<\/td>\n<td style=\"background-color: #f0f0f0; width: 40%;\">Threat-Strength (TS) Strategies<br \/>Use strengths to avoid threats<br \/>1.<br \/>2.<\/td>\n<td style=\"background-color: #f0f0f0; width: 40%;\">Threat-Weakness (TW) Strategies<br \/>Minimize weaknesses and avoid threats<br \/>1.<br \/>2.<\/td>\n<\/tr><\/tbody><\/table><p>David gives an example for Campbell Soup Company that stresses financial goals, but it also illustrates how you can pair the items within a SWOT grid to develop strategies. (This version of the chart is abbreviated.)<\/p><table border=\"2\" cellpadding=\"5\" style=\"border-color: black; border-spacing: 1px;\"><tbody><tr><td style=\"background-color: #e1e1e1; width: 30%;\"><span><strong>\u00a0<\/strong><\/span><\/td>\n<td style=\"background-color: #a5a5a5; width: 35%;\"><span><strong>STRENGTHS<\/strong><\/span>\n<ul><li><span>Current profit ratio increased<\/span><\/li>\n<li><span>Employee morale high<\/span><\/li>\n<li><span>Market share has increased<\/span><\/li>\n<\/ul><\/td>\n<td style=\"background-color: #a5a5a5; width: 35%;\"><span><strong>WEAKNESSES<\/strong><\/span>\n<ul><li><span>Legal suits not resolved<\/span><\/li>\n<li><span>Plant capacity has fallen<\/span><\/li>\n<li><span>Lack of strategic management system<\/span><\/li>\n<\/ul><\/td>\n<\/tr><tr><td style=\"background-color: #a5a5a5; width: 30%;\"><span><strong>OPPORTUNITIES<\/strong><\/span>\n<ul><li><span><span>Western European unification<\/span><\/span><\/li>\n<li><span><span>Rising health consciousness in selecting foods<\/span><\/span><\/li>\n<li><span><span>Demand for soups increasing annually<\/span><\/span><\/li>\n<\/ul><\/td>\n<td style=\"background-color: #f0f0f0; width: 35%;\">Opportunity-Strength (OS) Strategies\n<ul><li>Acquire food company in Europe (S1, S3, O1)<\/li>\n<li>Develop new healthy soups (S2, O2)<\/li>\n<\/ul><\/td>\n<td style=\"background-color: #f0f0f0; width: 35%;\">Opportunity-Weakness (OW) Strategies\n<ul><li>Develop new Pepperidge Farm\u00a0products (W1, O2, O3)<\/li>\n<\/ul><\/td>\n<\/tr><tr><td style=\"background-color: #a5a5a5; width: 30%;\"><span><strong>THREATS<\/strong><\/span>\n<ul><li><span><span>Low value of dollar<\/span><\/span><\/li>\n<li><span><span>Tin cans are not biodegradable<\/span><\/span><\/li>\n<\/ul><\/td>\n<td style=\"background-color: #f0f0f0; width: 35%;\">Threat-Strength (TS) Strategies\n<ul><li>Develop new biodegradable soup containers (S1, T2)<\/li>\n<\/ul><\/td>\n<td style=\"background-color: #f0f0f0; width: 35%;\">Threat-Weakness (TW) Strategies\n<ul><li>Close unprofitable European operations (W3, T1)<\/li>\n<\/ul><\/td>\n<\/tr><\/tbody><\/table><p>This example also illustrates how threats can become opportunities (and vice versa). The limitation of tin cans (which aren't biodegradable) creates an opportunity for leadership in developing biodegradable containers. There are several formats you can use to do a SWOT analysis, <a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/swot-analysis\/tools\">including a\u00a0basic SWOT form<\/a> that you can use to prompt analysis, but whatever format you use, don't be surprised if your strengths and weaknesses don't precisely match up to your opportunities and threats. You might need to refine, or you might need to simply look at the facts longer, or from a different angle. Your chart, list or table will certainly reveal patterns.<\/p><h3>Listing Your Internal Factors: Strengths and Weaknesses (S, W)<\/h3><p><strong>Internal factors include your resources and experiences. General areas to consider:<\/strong><\/p><ul><li>Human resources - staff, volunteers, board members, target population<\/li>\n<li>Physical resources - your location, building, equipment\u00a0<\/li>\n<li>Financial - grants, funding agencies, other sources of income<\/li>\n<li>Activities and processes - programs you run, systems you employ<\/li>\n<li>Past experiences - building blocks for learning and success, your reputation in the community<\/li>\n<\/ul><p>Don't be too modest when listing your strengths. If you're having difficulty naming them, start by simply listing your characteristics (e.g.., we're small, we're connected to the neighborhood). Some of these will probably be strengths.<\/p><p>Although the strengths and weakness of your organization are your internal qualities, don't overlook the perspective of people outside your group. Identify strengths and weaknesses from both your own point of view and that of others, including those you\u00a0serve or deal with. Do others see problems--or assets--that you don't?<\/p><p>How do you get information about how outsiders perceive your strengths and weaknesses? You may know already if you've listened to those you serve. If not, this might be the time to gather that type of information. See related sections\u00a0for ideas on conducting <a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/conduct-focus-groups\/main\">focus groups<\/a>, user <a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/conduct-surveys\/main\">surveys<\/a>, and <a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/conduct-public-forums\/main\">listening sessions<\/a>.<\/p><h3>Listing External Factors: Opportunities and Threats (O, T)<\/h3><p>Cast a wide net for the external part of the assessment. No organization, group, program, or neighborhood is immune to outside events and forces. Consider your connectedness, for better and worse, as you compile this part of your SWOT list.<\/p><p><strong>Forces and facts that your group does not control include:<\/strong><\/p><ul><li>Future trends\u00a0in your field or the culture<\/li>\n<li>The economy - local, national, or international<\/li>\n<li>Funding sources - foundations, donors, legislatures<\/li>\n<li>Demographics - changes in the age, race, gender, culture of those you serve or in your area<\/li>\n<li>The physical environment (Is your building in a growing part of town? Is the bus company cutting routes?)<\/li>\n<li>Legislation (Do new federal requirements make your job harder...or easier?)<\/li>\n<li>Local, national or international events<\/li>\n<\/ul><h2>How do you create a SWOT analysis?<\/h2><h3>Who develops the SWOT?<\/h3><p>The most common users of a SWOT analysis are team members and project managers who are responsible for decision-making and strategic planning.<\/p><p>But don't overlook anyone in the creation stage!<\/p><p>An individual or small group can develop a SWOT analysis, but it will be more effective if you take advantage of many stakeholders. Each person or group offers a different perspective on the strengths and weaknesses of your program and has different experiences of both.<\/p><p>Likewise, one staff member, or volunteer or stakeholder may have information about an opportunity or threat that is essential to understanding your position and determining your future.<\/p><h3>When and where do you develop a SWOT analysis?<\/h3><p>A SWOT analysis is often created during a retreat or planning session that allows several hours for brainstorming and analysis. The best results come when the process is collaborative and inclusive.<\/p><p>When creating the analysis, people are asked to pool their individual and shared knowledge and experience. The more relaxed, friendly and constructive the setting, the more truthful, comprehensive, insightful, and useful your analysis will be.<\/p><h3>How do you develop a SWOT analysis?<\/h3><p><strong>Steps for conducting a SWOT analysis:<\/strong><\/p><ul><li>Designate a leader or group facilitator who has good listening and group process skills, and who can keep things moving and on track.<\/li>\n<li>Designate a recorder to back up the leader if your group is large. Use newsprint on a flip chart or a large board to record the analysis and discussion points. You can record later in a more polished fashion to share with stakeholders and to update.<\/li>\n<li>Introduce the SWOT method and its purpose in your organization. This can be as simple as asking, \"Where are we, where can we go?\" If you have time, you could run through a quick example based on a shared experience or well-known public issue.<\/li>\n<li>Depending on the nature of your group and the time available, let all participants introduce themselves. Then divide your stakeholders into smaller groups. If your retreat or meeting draws several groups of stakeholders together, make sure you mix the small groups to get a range of perspectives, and give them a chance to introduce themselves.\n<ul><li>The size of these depends on the size of your entire group \u2013 breakout groups can range from three to ten. If the size gets much larger, some members may not participate.<\/li>\n<\/ul><\/li>\n<li>Have each group designate a recorder, and provide each with newsprint or dry -erase board. Direct them to create a SWOT analysis in the format you choose-a chart, columns, a matrix, or even a page for each quality.\n<ul><li>Give the groups 20-30 minutes to brainstorm and fill out their own strengths, weakness, opportunities and threats chart for your program, initiative or effort. Encourage them not to rule out any ideas at this stage, or the next.<\/li>\n<li>Remind groups that\u00a0the way to have a good idea is to have lots of ideas. Refinement can come later. In this way, the SWOT analysis also supports valuable discussion within your group or organization as you honestly assess.<\/li>\n<li>It helps to generate lots of comments about your organization and your program, and even to put them in multiple categories if that provokes thought.<\/li>\n<li>Once a list has been generated, it helps to refine it to the best 10 or fewer points so that the analysis can be truly helpful.<\/li>\n<\/ul><\/li>\n<li>Reconvene the group at the agreed-upon time to share results. Gather information from the groups, recording on the flip-chart or board. Collect and organize the differing groups' ideas and perceptions.\n<ul><li>Proceed in S-W-O-T order, recording strengths first, weaknesses second, etc.<\/li>\n<li>Or you can begin by calling for the top priorities in each category -the strongest strength, most dangerous weakness, biggest opportunity, worst threat--and continue to work across each category.<\/li>\n<li>Ask one group at a time to report (\"Group A, what do you see as strengths?\") You can vary which group begins the report so a certain group isn't always left \"bringing up the end\" and repeating points made by others. (\"Group B, let's start with you for weaknesses.\")<\/li>\n<li>Or, you can open the floor to all groups (\"What strengths have you noted?\") for each category until all have contributed what they think is needed.<\/li>\n<\/ul><\/li>\n<li>Discuss and record the results. Depending on your time frame and purpose:\n<ul><li>Come to some consensus about the most important items in each category<\/li>\n<li>Relate the analysis to your vision, mission, and goals<\/li>\n<li>Translate the analysis to action plans and strategies<\/li>\n<\/ul><\/li>\n<li>If appropriate, prepare a written summary of the SWOT analysis to share with\u00a0participants for continued use in planning and implementation.<\/li>\n<\/ul><p>More ideas on conducting successful meetings can be found in Community Tool Box resources on\u00a0<a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/conduct-public-forums\/main\">conducting public forums and listening sessions<\/a>, <a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/assessment\/assessing-community-needs-and-resources\/conduct-focus-groups\/main\">conducting focus groups<\/a>, and\u00a0<a href=\"http:\/\/ctb.ku.edu\/en\/table-of-contents\/structure\/training-and-technical-assistance\/retreats\/main\">organizing a retreat<\/a>.<\/p><h2>How do you use your SWOT analysis?<\/h2><p>Better understanding the factors affecting your initiative put you in a better position for action. This understanding helps as you:<\/p><ul><li>Identify the issues or problems you intend to change<\/li>\n<li>Set or reaffirm goals<\/li>\n<li>Create an action plan<\/li>\n<\/ul><p>As you consider your analysis, be open to the possibilities that exist within a weakness or threat. Likewise, recognize that an opportunity can become a threat if everyone else sees the opportunity and plans to take advantage of it as well, thereby increasing your competition.<\/p><p>Finally, during your assessment and planning, you might keep an image in mind to help you make the most of a SWOT analysis: <em>Look for a \"stretch,\" not just a \"fit.\" <\/em>As Radha Balamuralikrishna and John C. Dugger of Iowa State University point out, SWOT usually reflects your current position or situation. Therefore one drawback is that it might not encourage openness to new possibilities. You can use SWOT to justify a course that has already been decided upon, but if your goal is to grow or improve, you will want to keep this in mind.<\/p><h2>In Summary<\/h2><p>A realistic recognition of the weaknesses and threats that exist for your effort is the first step to countering them with a robust set of strategies that build upon strengths and opportunities. A SWOT analysis identifies your strengths, weaknesses, opportunities and threats to assist you in making strategic plans and decisions.<\/p>","created_at":"2018-11-07T16:50:28+0000","updated_at":"2018-11-07T16:50:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"ctb.ku.edu","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":507,"title":"Crypto Tools for DevOps: git-crypt - Tozny","url":"https:\/\/tozny.com\/blog\/crypto-tools-for-devops-git-crypt\/","content":"<img src=\"https:\/\/tozny.com\/wp-content\/uploads\/2017\/04\/Screen-Shot-2017-04-03-at-9.59.38-AM.png\" alt=\"git-crypt cli\" title=\"git-crypt\" class=\"fwI\" \/><p><em>As part of an ongoing series, we\u2019re taking a deep dive into the structure, use, and benefits of various crypto tools for devops.<\/em><\/p><p>Managing credentials securely with end-to-end encryption utilities helps development teams to avoid embarrassing data leaks while sharing passwords among themselves or with deployment environments.<\/p><p>Unfortunately, keeping credentials separate from the application code that uses them can lead to both partition issues (keeping staging credentials separate from production ones) and discovery problems (knowing which credential to use when).<\/p><p>Today\u2019s post won\u2019t address\u00a0<em>all<\/em> of the issues with credentials management, but it will help explain a specific utility that empowers developers to keep credentials and the code that uses them in the same place.<\/p><h2>git-crypt<\/h2><p><a href=\"https:\/\/github.com\/AGWA\/git-crypt\">git-crypt<\/a> is an open source, command line utility that empowers developers to protect specific files within a git repository.<\/p><blockquote><p>Files which you choose to protect are encrypted when committed, and decrypted when checked out. git-crypt lets you freely share a repository containing a mix of public and private content. git-crypt gracefully degrades, so developers without the secret key can still clone and commit to a repository with encrypted files.<\/p><\/blockquote><p>The advantage of using git-crypt is that your protected data stays alongside the code that uses it. You don\u2019t need to configure or manage additional infrastructure (no AWS integration as with <a href=\"https:\/\/tozny.com\/blog\/crypto-tools-for-devops-credstash\/\">Credstash<\/a>).<\/p><p>Unlike Credstash, there is no inherent versioning of credentials supported through git-crypt. Instead, versions are maintained through the very fact that\u00a0<em>all files<\/em> in the repository are versioned natively.<\/p><h2>DevOps Use Case<\/h2><p>As mentioned above, the biggest advantage of git-crypt is that private data and public data can live in the same location. This means any developer can clone your repository, but only\u00a0<em>permitted<\/em> developers can decrypt and utilize the information protected within it.<\/p><p>You can <a href=\"https:\/\/github.com\/AGWA\/git-crypt\/blob\/master\/INSTALL.md\">install git-crypt<\/a> on Mac easily via <a href=\"http:\/\/brewformulas.org\/GitCrypt\">Homebrew<\/a>. You can also easily build the utility on Linux (either Debian or RHEL\/CentOS). This means credentials can be shared both with individual developers on the team\u00a0<em>and<\/em> with individual servers to which application code is deployed.<\/p><p>Servers (and individuals) can either have their own GPG keypairs or use repository-specific symmetric secrets when protecting credentials. If using symmetric keys, you will still need a way to securely share these keys between collaborators (among the development team\u00a0<em>and<\/em> to any deployed servers).<\/p><h2>The Good and the Bad<\/h2><p>git-crypt is a fantastic utility that allows us to keep all of our data in the same spot and automically control who (or what) has access to secrets\u2019 plain text. It uses strong cryptographic keys and leverages existing public key infrastructure for access control.<\/p><p>The biggest downside of working with git-crypt is its requirement to either use GPG itself (requiring additional installed utilities) or to share symmetric keys between developers and systems for access to credentials. If you have a utility with which you can securely distribute symmetric keys, you might as well use the same utility for sharing the credentials themselves.<\/p><h2>Further Reading<\/h2><p>On its own, git-crypt is merely an encrypted storage system for secrets. It\u2019s incredibly useful for keeping secrets alongside the code that uses them. It also enables fine-grained access control through specific GPG keys. git-crypt does not, however manage loading these credentials into the environment for use by the application.<\/p><p>For that, you might want to consider one of the various \u201c<a href=\"https:\/\/www.google.com\/search?q=dotenv\">dotenv<\/a>\u201d packages, available for just about every programming language available.<\/p><p>Next week we\u2019ll move beyond the simple, local\u00a0utilities of Credstash and git-crypt and study a more distributed platform that integrates with server provisioning tools: Hashicorp Vault.<\/p><footer>&#13;\n\t\t\t<\/footer>","created_at":"2018-11-07T16:50:27+0000","updated_at":"2018-11-07T16:50:27+0000","published_at":"2017-04-10T05:00:48+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"tozny.com","preview_picture":"https:\/\/tozny.com\/wp-content\/uploads\/2017\/04\/Screen-Shot-2017-04-03-at-9.59.38-AM.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":508,"title":"Must Have Git Aliases: Advanced Examples - Be Present Now","url":"http:\/\/durdn.com\/blog\/2012\/11\/22\/must-have-git-aliases-advanced-examples\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:50:27+0000","updated_at":"2018-11-07T16:50:27+0000","annotations":[],"reading_time":0,"domain_name":"durdn.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":506,"title":"Rust Bibliography \u00b7 The Rust Programming Language","url":"https:\/\/forge.rust-lang.org\/bibliography.html","content":"<p>This is a reading list of material relevant to Rust. It includes prior\nresearch that has - at one time or another - influenced the design of\nRust, as well as publications about Rust.<\/p><h2 id=\"type-system\">Type system<\/h2><ul><li><a href=\"http:\/\/209.68.42.137\/ucsd-pages\/Courses\/cse227.w03\/handouts\/cyclone-regions.pdf\">Region based memory management in Cyclone<\/a><\/li>\n  <li><a href=\"http:\/\/www.cs.umd.edu\/projects\/PL\/cyclone\/scp.pdf\">Safe manual memory management in Cyclone<\/a><\/li>\n  <li><a href=\"http:\/\/www.ps.uni-sb.de\/courses\/typen-ws99\/class.ps.gz\">Typeclasses: making ad-hoc polymorphism less ad hoc<\/a><\/li>\n  <li><a href=\"https:\/\/www.cs.utah.edu\/plt\/publications\/jfp12-draft-fcdf.pdf\">Macros that work together<\/a><\/li>\n  <li><a href=\"http:\/\/scg.unibe.ch\/archive\/papers\/Scha03aTraits.pdf\">Traits: composable units of behavior<\/a><\/li>\n  <li><a href=\"http:\/\/www.cs.uwm.edu\/faculty\/boyland\/papers\/unique-preprint.ps\">Alias burying<\/a> - We tried something similar and abandoned it.<\/li>\n  <li><a href=\"http:\/\/www.cs.uu.nl\/research\/techreps\/UU-CS-2002-048.html\">External uniqueness is unique enough<\/a><\/li>\n  <li><a href=\"https:\/\/research.microsoft.com\/pubs\/170528\/msr-tr-2012-79.pdf\">Uniqueness and Reference Immutability for Safe Parallelism<\/a><\/li>\n  <li><a href=\"http:\/\/www.cs.ucla.edu\/~palsberg\/tba\/papers\/tofte-talpin-iandc97.pdf\">Region Based Memory Management<\/a><\/li>\n<\/ul><h2 id=\"concurrency\">Concurrency<\/h2><ul><li><a href=\"https:\/\/research.microsoft.com\/pubs\/69431\/osr2007_rethinkingsoftwarestack.pdf\">Singularity: rethinking the software stack<\/a><\/li>\n  <li><a href=\"https:\/\/research.microsoft.com\/pubs\/67482\/singsharp.pdf\">Language support for fast and reliable message passing in singularity OS<\/a><\/li>\n  <li><a href=\"http:\/\/supertech.csail.mit.edu\/papers\/steal.pdf\">Scheduling multithreaded computations by work stealing<\/a><\/li>\n  <li><a href=\"http:\/\/www.eecis.udel.edu\/%7Ecavazos\/cisc879-spring2008\/papers\/arora98thread.pdf\">Thread scheduling for multiprogramming multiprocessors<\/a><\/li>\n  <li><a href=\"http:\/\/www.aladdin.cs.cmu.edu\/papers\/pdfs\/y2000\/locality_spaa00.pdf\">The data locality of work stealing<\/a><\/li>\n  <li><a href=\"http:\/\/citeseerx.ist.psu.edu\/viewdoc\/download?doi=10.1.1.170.1097&amp;rep=rep1&amp;type=pdf\">Dynamic circular work stealing deque<\/a> - The Chase\/Lev deque<\/li>\n  <li><a href=\"http:\/\/www.cs.rice.edu\/%7Eyguo\/pubs\/PID824943.pdf\">Work-first and help-first scheduling policies for async-finish task parallelism<\/a> - More general than fully-strict work stealing<\/li>\n  <li><a href=\"http:\/\/www.coopsoft.com\/ar\/CalamityArticle.html\">A Java fork\/join calamity<\/a> - critique of Java\u2019s fork\/join library, particularly its application of work stealing to non-strict computation<\/li>\n  <li><a href=\"http:\/\/www.stanford.edu\/~ouster\/cgi-bin\/papers\/coscheduling.pdf\">Scheduling techniques for concurrent systems<\/a><\/li>\n  <li><a href=\"http:\/\/www.blagodurov.net\/files\/a8-blagodurov.pdf\">Contention aware scheduling<\/a><\/li>\n  <li><a href=\"http:\/\/www.cse.ohio-state.edu\/hpcs\/WWW\/HTML\/publications\/papers\/TR-12-1.pdf\">Balanced work stealing for time-sharing multicores<\/a><\/li>\n  <li><a href=\"http:\/\/dl.acm.org\/citation.cfm?id=1953616&amp;dl=ACM&amp;coll=DL&amp;CFID=524387192&amp;CFTOKEN=44362705\">Three layer cake for shared-memory programming<\/a><\/li>\n  <li><a href=\"http:\/\/www.cs.bgu.ac.il\/%7Ehendlerd\/papers\/p280-hendler.pdf\">Non-blocking steal-half work queues<\/a><\/li>\n  <li><a href=\"http:\/\/www.mpi-sws.org\/~turon\/reagents.pdf\">Reagents: expressing and composing fine-grained concurrency<\/a><\/li>\n  <li><a href=\"https:\/\/www.cs.rochester.edu\/u\/scott\/papers\/1991_TOCS_synch.pdf\">Algorithms for scalable synchronization of shared-memory multiprocessors<\/a><\/li>\n  <li><a href=\"https:\/\/www.cl.cam.ac.uk\/techreports\/UCAM-CL-TR-579.pdf\">Epoch-based reclamation<\/a>.<\/li>\n<\/ul><h2 id=\"others\">Others<\/h2><ul><li><a href=\"https:\/\/www.usenix.org\/legacy\/events\/hotos03\/tech\/full_papers\/candea\/candea.pdf\">Crash-only software<\/a><\/li>\n  <li><a href=\"http:\/\/people.cs.umass.edu\/~emery\/pubs\/berger-pldi2001.pdf\">Composing High-Performance Memory Allocators<\/a><\/li>\n  <li><a href=\"http:\/\/people.cs.umass.edu\/~emery\/pubs\/berger-oopsla2002.pdf\">Reconsidering Custom Memory Allocation<\/a><\/li>\n<\/ul><ul><li><a href=\"http:\/\/www.cs.indiana.edu\/~eholk\/papers\/hips2013.pdf\">GPU Programming in Rust: Implementing High Level Abstractions in a\nSystems Level\nLanguage<\/a>. Early GPU work by Eric Holk.<\/li>\n  <li><a href=\"https:\/\/www.usenix.org\/conference\/hotpar12\/parallel-closures-new-twist-old-idea\">Parallel closures: a new twist on an old\nidea<\/a>\n    <ul><li>not exactly about Rust, but by nmatsakis<\/li>\n    <\/ul><\/li>\n  <li><a href=\"ftp:\/\/ftp.cs.washington.edu\/tr\/2015\/03\/UW-CSE-15-03-02.pdf\">Patina: A Formalization of the Rust Programming\nLanguage<\/a>. Early\nformalization of a subset of the type system, by Eric Reed.<\/li>\n  <li><a href=\"http:\/\/arxiv.org\/abs\/1505.07383\">Experience Report: Developing the Servo Web Browser Engine using\nRust<\/a>. By Lars Bergstrom.<\/li>\n  <li><a href=\"https:\/\/michaelsproul.github.io\/rust_radix_paper\/rust-radix-sproul.pdf\">Implementing a Generic Radix Trie in\nRust<\/a>. Undergrad\npaper by Michael Sproul.<\/li>\n  <li><a href=\"http:\/\/scialex.github.io\/reenix.pdf\">Reenix: Implementing a Unix-Like Operating System in\nRust<\/a>. Undergrad paper by Alex\nLight.<\/li>\n  <li><a href=\"http:\/\/octarineparrot.com\/assets\/mrfloya-thesis-ba.pdf\">Evaluation of performance and productivity metrics of potential programming languages in the HPC environment<\/a>.\nBachelor\u2019s thesis by Florian Wilkens. Compares C, Go and Rust.<\/li>\n  <li><a href=\"http:\/\/spw15.langsec.org\/papers\/couprie-nom.pdf\">Nom, a byte oriented, streaming, zero copy, parser combinators library\nin Rust<\/a>. By\nGeoffroy Couprie, research for VLC.<\/li>\n  <li><a href=\"http:\/\/compilers.cs.uni-saarland.de\/papers\/lkh15_cgo.pdf\">Graph-Based Higher-Order Intermediate\nRepresentation<\/a>. An\nexperimental IR implemented in Impala, a Rust-like language.<\/li>\n  <li><a href=\"http:\/\/compilers.cs.uni-saarland.de\/papers\/ppl14_web.pdf\">Code Refinement of Stencil\nCodes<\/a>. Another\npaper using Impala.<\/li>\n  <li><a href=\"http:\/\/publications.lib.chalmers.se\/records\/fulltext\/219016\/219016.pdf\">Parallelization in Rust with fork-join and\nfriends<\/a>. Linus\nFarnstrand\u2019s master\u2019s thesis.<\/li>\n  <li><a href=\"http:\/\/munksgaard.me\/papers\/laumann-munksgaard-larsen.pdf\">Session Types for\nRust<\/a>. Philip\nMunksgaard\u2019s master\u2019s thesis. Research for Servo.<\/li>\n  <li><a href=\"http:\/\/amitlevy.com\/papers\/tock-plos2015.pdf\">Ownership is Theft: Experiences Building an Embedded OS in Rust - Amit Levy, et. al.<\/a><\/li>\n  <li><a href=\"https:\/\/raw.githubusercontent.com\/Gankro\/thesis\/master\/thesis.pdf\">You can\u2019t spell trust without Rust<\/a>. Alexis Beingessner\u2019s master\u2019s thesis.<\/li>\n  <li><a href=\"http:\/\/bioinformatics.oxfordjournals.org\/content\/early\/2015\/10\/06\/bioinformatics.btv573\">Rust-Bio: a fast and safe bioinformatics library<\/a>. Johannes K\u00f6ster<\/li>\n  <li><a href=\"https:\/\/octarineparrot.com\/assets\/msci_paper.pdf\">Safe, Correct, and Fast Low-Level Networking<\/a>. Robert Clipsham\u2019s master\u2019s thesis.<\/li>\n  <li><a href=\"http:\/\/hdl.handle.net\/2429\/55609\">Formalizing Rust traits<\/a>. Jonatan Milewski\u2019s master\u2019s thesis.<\/li>\n  <li><a href=\"http:\/\/users.cecs.anu.edu.au\/~steveb\/downloads\/pdf\/rust-ismm-2016.pdf\">Rust as a Language for High Performance GC Implementation<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/Kha\/electrolysis\">Simple Verification of Rust Programs via Functional Purification<\/a>. Sebastian Ullrich\u2019s master\u2019s thesis.<\/li>\n  <li><a href=\"http:\/\/spw17.langsec.org\/papers\/chifflier-parsing-in-2017.pdf\">Writing parsers like it is 2017<\/a> Pierre Chifflier and Geoffroy Couprie for the Langsec Workshop<\/li>\n  <li><a href=\"https:\/\/www.tockos.org\/assets\/papers\/rust-kernel-apsys2017.pdf\">The Case for Writing a Kernel in Rust<\/a><\/li>\n<\/ul>","created_at":"2018-11-07T16:50:26+0000","updated_at":"2018-11-07T16:50:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"forge.rust-lang.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":505,"title":"A Beginners' Guide to Installing Arch Linux","url":"https:\/\/news.softpedia.com\/news\/A-Beginners-Guide-to-Installing-Arch-Linux-352365.shtml","content":"<p class=\"intro\"><strong itemprop=\"description\">The following tutorial will teach any computer user how to install the Arch Linux operating system on their personal computer or laptop.<\/strong><\/p><p class=\"mgbot_20\">Arch Linux, the ultimate frontier! Well, the funny thing is that <a target=\"_blank\" href=\"https:\/\/linux.softpedia.com\/downloadTag\/Arch%20Linux\" title=\"Arch\">Arch<\/a> Linux has become some sort of a scarecrow for many Linux beginners, especially for those who were \u201ctrained\u201d into <a target=\"_blank\" href=\"https:\/\/linux.softpedia.com\/downloadTag\/Ubuntu\" title=\"Ubuntu\">Ubuntu<\/a> or other \u201cLinux distribution for the average Joe or Jane.\u201d And that\u2019s because of the kind of painful installation process, which we\u2019ll try to simplify in this tutorial.<\/p>\n<p class=\"mgbot_20\">However, the truth is that once you install Arch Linux, you\u2019ll get a lightweight and flexible (do-it-yourself) Linux distribution, designed as a Rolling Release, which means that you won\u2019t have to upgrade every 6 months to a different release like some popular Linux OSes do these days. You install once and you update you system on a regular basis.<\/p>\n<p class=\"mgbot_20\">Arch Linux follows the KISS principle: Keep it simple, stupid! It will keep your system up-to-date and on the bleeding edge, so you won\u2019t have to wait six months for a new kernel.<\/p>\n<p class=\"mgbot_20\">What do I need to get started? The wonderful people behind the Arch Linux operating system provide a monthly updated ISO image, just in case any new Linux user wants to try the distribution. <a target=\"_blank\" href=\"http:\/\/linux.softpedia.com\/get\/System\/Operating-Systems\/Linux-Distributions\/Arch-Linux-864.shtml\"><strong>Get the latest version right now from Softpedia<\/strong><\/a>.<\/p>\n<p class=\"mgbot_20\">But, before anything else, you need to partition your hard drive for Arch Linux, so you won\u2019t have to do it manually from within the Live CD. For this, we will use the <a target=\"_blank\" href=\"https:\/\/linux.softpedia.com\/get\/System\/Operating-Systems\/Linux-Distributions\/Parted-Magic-21659.shtml\" title=\"Parted Magic\">Parted Magic<\/a> Live CD.<\/p>\n<p class=\"mgbot_20\">At this moment, you should write both ISO images on two different USB flash drives by following our very popular and useful \u201c<a target=\"_blank\" href=\"https:\/\/news.softpedia.com\/news\/How-to-Run-Linux-from-an-USB-Flash-Drive-93316.shtml\"><strong>How to Run Linux from an USB Flash Drive<\/strong><\/a>\u201d tutorial. In case you don\u2019t have two USB sticks, write the <a target=\"_blank\" href=\"http:\/\/linux.softpedia.com\/get\/System\/Operating-Systems\/Linux-Distributions\/Parted-Magic-21659.shtml\"><strong>Parted Magic<\/strong><\/a> ISO first and overwrite it with the Arch Linux ISO image when you\u2019re done partitioning the hard drive.<\/p>\n<p class=\"mgbot_20\">So, insert the Parted Magic USB stick in a USB port, and reboot the computer in order to boot from the flash drive. Hit the F8, F11 or F12 key (depending on your BIOS) until the boot menu appears, in order to select the USB flash drive as the boot device.<\/p>\n<p class=\"mgbot_20\"><strong>Partition your hard drive:<\/strong><\/p>\n<p class=\"mgbot_20\">Once the Parted Magic CD was loaded into memory, start the Partition Editor tool from the desktop...<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_001.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_001.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">Go to Device -&gt;\u00a0Create Partition Table and click the \"Apply\" button on the warning dialog...<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_002.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_002.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">Right click again on the unallocated entry and in the new window, type the size of your root partition in the \"New size (MiB)\" field (we recommend around 20 GB) and select the \"ext4\" option from the \"Filesystem\" drop down list. Type <strong>\/<\/strong> in the Label field. Click the OK button and, in a few seconds, you'll notice a \"\/\" line with the specified size.<\/p>\n<p class=\"mgbot_20\">Now, right click on the unallocated entry and choose \"New\" to create the  partitions. In the new window, type 2048 in the \"New size (MiB)\" field  and select the \"swap area\" option from the \"Filesystem\" drop down list.  Click the OK button and, in a few seconds, you'll notice a \"swap\" line  with the specified size.<\/p>\n<p class=\"mgbot_20\">Once again, right click on the unallocated entry, type the size of your home partition in the \"New size (MiB)\" field (usually the rest of the hard drive) and select the \"ext4\" option from the \"Filesystem\" drop down list. Type <strong>\/home<\/strong> in the Label field. Click the OK button and, in a few seconds, you'll notice a \"\/home\" line with the specified size.<\/p>\n<p class=\"mgbot_20\">Click the \"Apply\" button to make the changes. Click Apply again on the warning dialog and wait for the partitioning process to finish. Close the dialog.<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_003.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_003.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">Now that your hard drive has been partitioned, reboot your computer and eject the USB drive. As we said before, use the second USB stick or overwrite this one with the Arch Linux ISO image. Do the same as above in order to boot from the flash drive.<\/p>\n<p class=\"mgbot_20\">Wait for the image to load...<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_004.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_004.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">...and select the right architecture for you (64-bit for newer systems or 32-bit for old ones). Wait for the system to load, it will take a few seconds, and all you\u2019ll see now is this...<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_005.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_005.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\"><strong>Editor\u2019s note:<\/strong><em><strong> <\/strong>For those of you who don\u2019t know yet, \/dev\/sda is your first hard drive, in case you have two or more. To see on which drive you want to install, use the fdisk -l command, which will list all your drives, including the USB disk with Arch Linux. Therefore, if you have a single hard drive in your machine, which is recommended for most computers, use \u201csda\u201d from now on during the entire installation process. If not, use \u201csdb\u201d or \u201csdc,\u201d depending on your configuration.<\/em><\/p>\n<p class=\"mgbot_20\">Before we start, let\u2019s execute a simple command, which will tell us if the hard drive has been partitioned correctly or not and if it is the correct one:<\/p>\n<p class=\"mgbot_20\"><em>lsblk \/dev\/sda<\/em><\/p>\n<p class=\"mgbot_20\">...and the output is:<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_006.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_006.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">Now, let\u2019s activate the swap partition with the following command:<\/p>\n<p class=\"mgbot_20\"><em>swapon \/dev\/sda2<\/em><\/p>\n<p class=\"mgbot_20\">Let\u2019s mount the partitions:<\/p>\n<p class=\"mgbot_20\">mount \/dev\/sda1 \/mnt<\/p>\n<p class=\"mgbot_20\">mkdir \/mnt\/home<\/p>\n<p class=\"mgbot_20\">mount \/dev\/sda3 \/mnt\/home<\/p>\n<p class=\"mgbot_20\">Run the following command to install the base system for your future Arch Linux installation:<\/p>\n<p class=\"mgbot_20\"><em>pacstrap \/mnt base base-devel<\/em><\/p>\n<p class=\"mgbot_20\">Hit Enter and wait for the download\/installation process to finish, as it will take a few minutes, depending on your Internet speed. What will happen now? Some packages will be downloaded and installed. You can see the progress in real time. When everything is done, run the following command to configure system configuration file, named fstab, which lists all your partitions:<\/p>\n<p class=\"mgbot_20\"><em>genfstab -U -p \/mnt &gt;&gt; \/mnt\/etc\/fstab<\/em><\/p>\n<p class=\"mgbot_20\">And now, it\u2019s time to configure your Arch Linux system. Let\u2019s get in, shall we? Use the following command:<\/p>\n<p class=\"mgbot_20\"><em>arch-chroot \/mnt<\/em><\/p>\n<p class=\"mgbot_20\">So, we\u2019re in, let\u2019s set up the local language with the following command:<\/p>\n<p class=\"mgbot_20\"><em>nano \/etc\/locale.gen<\/em><\/p>\n<p class=\"mgbot_20\">Scroll the file until you find your language and uncomment the # in front of it. For us it was en_US.UTF-8 UTF-8, which means English United States, Unicode encoding. Save the file using the CTRL+O keyboard shortcut. Exit it using the CTRL+X keyboard shortcut.<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_007.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_007.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">Run the following three commands, one by one, hitting enter after each one:<\/p>\n<p class=\"mgbot_20\"><em>locale-gen<\/em><\/p>\n<p class=\"mgbot_20\">echo LANG=en_US.UTF-8 &gt; \/etc\/locale.conf<\/p>\n<p class=\"mgbot_20\">export LANG=en_US.UTF-8<\/p>\n<p class=\"mgbot_20\"><strong>WARNING:<\/strong> <em>Replace en_US.UTF-8 from the above commands with your selected locale!<\/em><\/p>\n<p class=\"mgbot_20\">Now we need to specify the timezone, so run the following command:<\/p>\n<p class=\"mgbot_20\"><em>ln -s \/usr\/share\/zoneinfo\/<strong>Zone<\/strong>\/<strong>SubZone<\/strong> \/etc\/localtime<\/em><\/p>\n<q class=\"subhead\">Note: Replace <strong>Zone<\/strong> and <strong>SubZone<\/strong> with your country and city. Example: <em>ln -s \/usr\/share\/zoneinfo\/Romania\/Bucharest \/etc\/localtime.<\/em>\nNow let\u2019s set the clock depending on the country and city selected above:\n<em>hwclock --systohc --utc<\/em>\n<a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_008.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_008.png\" \/><\/a>\n<strong>ONLY for 64-bit systems:<\/strong> You need to activate the \u201cmultilib\u201d repository. Execute the following command:\n<em>nano \/etc\/pacman.conf<\/em>\n...and delete the # in front of the following lines:\n<em>[multilib]<\/em>\nInclude = \/etc\/pacman.d\/mirrorlist\nYou know the drill, save the file using the CTRL+O keyboard shortcut. Exit it using the CTRL+X keyboard shortcut.\nLet\u2019s update the software repositories:\n<em>pacman -Sy<\/em>\n<strong>Editor\u2019s note:<\/strong> <em>Please meet pacman, your best friend from now on. pacman will help you install or remove packages from your system, as well as keep your entire machine up to date.<\/em>\nLet\u2019s create a password for the Administrator (also known as root):\n<em>passwd<\/em>\nType the password and hit Enter, twice.\nLet\u2019s add a user now:\n<em>useradd -m -g users -G wheel,storage,power -s \/bin\/bash softpedia<\/em>\n...and set a password for it:\n<em>passwd softpedia<\/em>\nType the password and hit Enter, twice.\nLet\u2019s install the gate keeper... just kidding, only sudo, with the following command:\n<em>pacman -S sudo<\/em>\nAnd now, let\u2019s give some rights to your newly created user. Execute the following command:\n<em>EDITOR=nano visudo<\/em>\nScroll down until you find the <strong>%wheel ALL=(ALL) ALL<\/strong> line. Delete the # in front of it. Save the file using the CTRL+O keyboard shortcut. Exit it using the CTRL+X keyboard shortcut.\n<a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_009.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_009.png\" \/><\/a>\nLet\u2019s install some more useful packages...\n<em>pacman -S bash-completion grub-bios os-prober<\/em>\n...and setup the boot loader...\ngrub-install --target=i386-pc --recheck \/dev\/sda\ncp \/usr\/share\/locale\/en\\@quot\/LC_MESSAGES\/grub.mo \/boot\/grub\/locale\/en.mo\n<em>grub-mkconfig -o \/boot\/grub\/grub.cfg<\/em>\n<strong>Note:<\/strong> Remember to use \/dev\/sda or \/dev\/sdb in the above command, depending on your hard drive!\nThat\u2019s it, let\u2019s exit the system and eject the mounted partitions:\n<em>exit<\/em>\numount \/mnt\/home\numount \/mnt\nreboot\n<a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_010.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_010.png\" \/><\/a>\nWhen you get back, enter your username and password created during the installation. Now we need to configure the Internet, so run the following command:\n<em>ip link<\/em>\n<a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_011.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_011.png\" \/><\/a>\nJust like in the screenshot above, you\u2019ll see the name of your Ethernet connection, so use it in the following commands:\n<em>sudo systemctl enable dhcpcd@&lt;interface&gt;.service<\/em>\nsudo systemctl start dhcpcd@&lt;interface&gt;.service\n<strong>Note:<\/strong> Replace &lt;interface&gt; in the above commands with the name of your interface. For example: <em>sudo systemctl enable dhcpcd@enp2s0.service \/ sudo systemctl start dhcpcd@enp2s0.service<\/em>\nLet\u2019s install the X11 packages and the video drivers:\n<em>sudo pacman -S xorg-server xorg-xinit xorg-server-utils mesa xf86-input-synaptics<\/em>\n<strong>Nvidia users<\/strong> should use the following command:\n<em>sudo pacman -S nvidia (for 32-bit users)<\/em>\nsudo pacman -S nvidia lib32-nvidia-utils (for 64-bit users)\n<strong>ATI users<\/strong> should use the following command:\n<em>sudo pacman -S catalyst-dkms (for 32-bit users)<\/em>\nsudo pacman -S catalyst-dkms lib32-catalyst-utils (for 64-bit users)\n<strong>Intel users<\/strong> should use the following command:\n<em>sudo pacman -S xf86-video-intel<\/em>\nAnd now, it\u2019s time to install a desktop environment. Use the following command for <strong>GNOME<\/strong>:\n<em>sudo pacman -S gnome<\/em>\n...or the following command for <strong>KDE<\/strong>:\n<em>sudo pacman -S kde<\/em>\nThis will download a lot of packages and it will take some time. When everything is done, run the following commands to start the desktop environment:\n<strong>FOR GNOME USERS:<\/strong>\n<em>sudo systemctl enable gdm.service<\/em>\nsudo systemctl start gdm.service\n<strong>FOR KDE USERS:<\/strong><\/q>\n<p class=\"mgbot_20\"><em>sudo systemctl enable kdm.service<\/em><\/p>\n<p class=\"mgbot_20\">sudo systemctl start kdm.service<\/p>\n<div class=\"bbthumbs bbgallery\"><div class=\"fl\"><a title=\"Click to view large image\" href=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/large\/archlinuxinstallation-large_012.jpg\" target=\"_blank\"><img border=\"0\" alt=\"Review image\" src=\"https:\/\/news.softpedia.com\/images\/extra\/LINUX\/small\/archlinuxinstallation-small_012.png\" \/><\/a><\/div><\/div>\n<p class=\"mgbot_20\">That\u2019s it! You should now see your favorite desktop environment. Have fun using Arch Linux, and don\u2019t forget to comment below in case you run into trouble with the installation.<\/p>\n<p class=\"mgbot_20\">Also, don't hesitate to read our \"<a target=\"_blank\" href=\"https:\/\/news.softpedia.com\/news\/How-to-Install-Third-Party-Apps-in-Arch-Linux-356894.shtml\"><strong>How to Install Third-Party Apps in Arch Linux<\/strong><\/a>\" tutorial.<\/p>","created_at":"2018-11-07T16:50:12+0000","updated_at":"2018-11-07T16:50:12+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"news.softpedia.com","preview_picture":"https:\/\/news-cdn.softpedia.com\/images\/news2\/A-Beginners-Guide-to-Installing-Arch-Linux-2.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":501,"title":"What Does the Future of Content Marketing Look Like? (Spoiler: Still Less Content)","url":"https:\/\/thinkgrowth.org\/what-does-the-future-of-content-marketing-look-like-spoiler-still-less-content-7941fcc63acb?gi=23900282fe5f","content":"<div class=\"section-inner sectionLayout--insetColumn\"><figure id=\"9bcf\" class=\"graf graf--figure graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*stka6SaLO8bBy9Yd-zaCgw.jpeg\" data-width=\"1400\" data-height=\"800\" data-action=\"zoom\" data-action-value=\"1*stka6SaLO8bBy9Yd-zaCgw.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*stka6SaLO8bBy9Yd-zaCgw.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"4fdf\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--h3\">T<strong class=\"markup--strong markup--p-strong\">he information age is arriving at its natural conclusion and we\u2019re all drowning in it<\/strong>. In 2015, Gartner\u2019s \u201cHype Cycle\u201d moved content marketing into the \u201ctrough of disillusionment\u201d phase. With good reason\u200a\u2014\u200ait\u2019s getting harder and harder to get results with content marketing.<\/p><figure id=\"ca9e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/blogs.gartner.com\/jake-sorofman\/content-marketing-world-2015-caution-flags-for-content-marketers\/\" data-href=\"https:\/\/blogs.gartner.com\/jake-sorofman\/content-marketing-world-2015-caution-flags-for-content-marketers\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*2kXf9D3CqT7WbbLwij0pvw.png\" data-width=\"684\" data-height=\"484\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*2kXf9D3CqT7WbbLwij0pvw.png\" alt=\"image\" \/><\/a><\/div><\/figure><p id=\"0144\" class=\"graf graf--p graf-after--figure\">In 2015, <a href=\"http:\/\/buzzsumo.com\/blog\/50-of-content-gets-8-shares-or-less-why-content-fails-and-how-to-fix-it\/\" data-href=\"http:\/\/buzzsumo.com\/blog\/50-of-content-gets-8-shares-or-less-why-content-fails-and-how-to-fix-it\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Buzzsumo analyzed performance of one million posts<\/a>. Here\u2019s what they found:<\/p><ul class=\"postList\"><li id=\"3273\" class=\"graf graf--li graf-after--p\">50% of posts receive 8 shares or less<\/li><li id=\"ddbd\" class=\"graf graf--li graf-after--li\">75% of these posts receive 39 shares or less<\/li><\/ul><p id=\"fdae\" class=\"graf graf--p graf-after--li\">And it gets worse.<\/p><p id=\"24f1\" class=\"graf graf--p graf-after--p\">Not only is most content under-performing, it\u2019s also getting harder for people who produce really good content to get attention. Engagement (measured by share count) is steadily dropping:<\/p><figure id=\"003d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*v0mq-VjO2pIEXP5_zsNxGA.png\" data-width=\"1476\" data-height=\"582\" data-action=\"zoom\" data-action-value=\"1*v0mq-VjO2pIEXP5_zsNxGA.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*v0mq-VjO2pIEXP5_zsNxGA.png\" alt=\"image\" \/><\/div><\/figure><p id=\"a35a\" class=\"graf graf--p graf-after--figure\">It\u2019s not that this wasn\u2019t a predictable conclusion. What\u2019s happening in content marketing is really the same thing that happens to every marketing channel\u200a\u2014\u200athey get less effective over time. In <a href=\"https:\/\/medium.com\/@andrewchen\" data-href=\"https:\/\/medium.com\/@andrewchen\" data-anchor-type=\"2\" data-user-id=\"8edc94d7a232\" data-action-value=\"8edc94d7a232\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Andrew Chen<\/a>\u2019s, <a href=\"http:\/\/andrewchen.co\/the-law-of-shitty-clickthroughs\/\" data-href=\"http:\/\/andrewchen.co\/the-law-of-shitty-clickthroughs\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">The Law of Shitty Clickthroughs<\/a>, he explains how this marketing law played out in online advertising. The first banner ever debuted in 1994 with an astonishing CTR of 94%. In 2014, average CTR on Facebook ads was\u00a0.05%. Why? Because the number of internet advertisers skyrocketed, and that\u2019s exactly what\u2019s happening in content marketing land.<\/p><h4 id=\"1e2f\" class=\"graf graf--h4 graf-after--p\">An Antidote: The Content Marketing Power\u00a0Law<\/h4><p id=\"e811\" class=\"graf graf--p graf-after--h4\">I am not suggesting you quit creating content. The days of salespeople having the power are over and if you\u2019re not educating prospects and providing them with resources to make their lives easier, you don\u2019t really have a marketing strategy. <strong class=\"markup--strong markup--p-strong\">Content marketing <em class=\"markup--em markup--p-em\">is <\/em>marketing today.<\/strong><\/p><p id=\"3e53\" class=\"graf graf--p graf-after--p\">So, what\u2019s the solution to getting heard in a distracted world?<\/p><p id=\"ae4c\" class=\"graf graf--p graf-after--p\">Let\u2019s look at some data.<\/p><p id=\"1b89\" class=\"graf graf--p graf-after--p\">I ran this analysis on the <a href=\"https:\/\/medium.com\/@RJMetrics\" data-href=\"https:\/\/medium.com\/@RJMetrics\" data-anchor-type=\"2\" data-user-id=\"40e2c85fbd51\" data-action-value=\"40e2c85fbd51\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">RJMetrics<\/a> blog over a year ago. This chart shows how every post we ever published on the company blog performed in 2015.<\/p><figure id=\"9b4d\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*TvEDkcCIUh4ZvDCxMVy5YQ.png\" data-width=\"800\" data-height=\"499\" data-action=\"zoom\" data-action-value=\"1*TvEDkcCIUh4ZvDCxMVy5YQ.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*TvEDkcCIUh4ZvDCxMVy5YQ.png\" alt=\"image\" \/><\/div><\/figure><p id=\"2347\" class=\"graf graf--p graf-after--figure\">Here is the most important takeway: <strong class=\"markup--strong markup--p-strong\">the top 10% of our posts were responsible for 90% of the results.<\/strong><\/p><figure id=\"5cad\" class=\"graf graf--figure graf--iframe graf-after--p\"><p id=\"3be6\" class=\"graf graf--p graf-after--figure\">It was painful to realize the vast majority of the work my team was doing was resulting in zero meaningful results for the business, but we also weren\u2019t alone. Google \u201ccontent marketing power law\u201d and you\u2019ll find other companies who have seen the exact same pattern play out in their own content.<\/p><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"226a\" class=\"graf graf--p graf-after--figure\">Pageview data can be easily pulled from Google Analytics, do this analysis yourself and let me know what you find.<\/p><p id=\"e0e4\" class=\"graf graf--p graf-after--p\">There are a few reasons why I think this happens:<\/p><ul class=\"postList\"><li id=\"795f\" class=\"graf graf--li graf-after--p\">First, we\u2019re increasingly dependent on algorithms to surface good content.<\/li><li id=\"fa4c\" class=\"graf graf--li graf-after--li\">Second, Search engines have become far more sophisticated in surfacing relevant content.<\/li><li id=\"88f6\" class=\"graf graf--li graf-after--li\">Third, we\u2019re on information overload and if we\u2019re going to share something, it needs to be exceptional.<\/li><\/ul><p id=\"ac77\" class=\"graf graf--p graf-after--li\">I\u2019m sure there are other reasons (share your thoughts in the comments). But what we know is true is that this law exists: <strong class=\"markup--strong markup--p-strong\">Content results are going to a smaller number of high-impact pieces<\/strong>. This has implications on how content marketing is evolving and where it\u2019s going in the future.<\/p><p id=\"cf06\" class=\"graf graf--p graf-after--p\">Let\u2019s talk about what those are.<\/p><h3 id=\"2b06\" class=\"graf graf--h3 graf-after--p\">How Content Marketing is\u00a0Evolving<\/h3><p id=\"5b71\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--h3\">1.<strong class=\"markup--strong markup--p-strong\">Content marketers are acting more like product managers. <\/strong>Product managers manage an organization\u2019s most valuable resource\u200a\u2014\u200aengineering time. They\u2019re highly selective about what gets built. Content marketers are beginning to adapt this same mentality, but not because content marketing time is scarce, rather because attention is scarce.<\/p><p id=\"ad63\" class=\"graf graf--p graf-after--p\">Product managers spend their time figuring out the highest-impact feature to build. They use data to understand the impact of a feature. They project manage. They coordinate a variety of roles to get the job done. They understand how the product, and the exact feature they are building, fits into the competitive landscape. Product managers are highly skilled at prioritization, they have to be\u200a\u2014\u200aengineering time is too valuable to risk spending it on low-impact work.<\/p><blockquote id=\"844f\" class=\"graf graf--pullquote graf-after--p\"><div>Content marketers will also build strong prioritization muscle, we have to\u200a\u2014\u200aattention is too scarce to waste it on low-impact content.<\/div><\/blockquote><p id=\"f56a\" class=\"graf graf--p graf-after--pullquote\">What this means is\u2026<\/p><p id=\"9eaf\" class=\"graf graf--p graf-after--p\">\u2705 Content marketers will get more strategic about what kind of content we create. (<a href=\"https:\/\/medium.com\/@clemnt\" data-href=\"https:\/\/medium.com\/@clemnt\" data-anchor-type=\"2\" data-user-id=\"cb62bf5319ce\" data-action-value=\"cb62bf5319ce\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Clement Vouillon<\/a> has <a href=\"https:\/\/medium.com\/point-nine-news\/content-market-fit-31e83e30f4b2#.duo0bg8nf\" data-href=\"https:\/\/medium.com\/point-nine-news\/content-market-fit-31e83e30f4b2#.duo0bg8nf\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">written about this here<\/a>).<\/p><p id=\"d4e1\" class=\"graf graf--p graf-after--p\">\u2705 Content marketers will become much more sophisticated in what we measure. (<a href=\"https:\/\/medium.com\/@Mallikarjunan\" data-href=\"https:\/\/medium.com\/@Mallikarjunan\" data-anchor-type=\"2\" data-user-id=\"412e47da8455\" data-action-value=\"412e47da8455\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Sam Mallikarjunan<\/a> has <a href=\"https:\/\/thinkgrowth.org\/its-time-to-rethink-the-content-business-ef22f7984d96#.bho1zj2q6\" data-href=\"https:\/\/thinkgrowth.org\/its-time-to-rethink-the-content-business-ef22f7984d96#.bho1zj2q6\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">written about this here<\/a>).<\/p><p id=\"83cd\" class=\"graf graf--p graf-after--p\">\u2705 If a content strategy <em class=\"markup--em markup--p-em\">is<\/em> based on high-volume, it will be assisted by artificial intelligence, not humans (<a href=\"https:\/\/medium.com\/@vedantmisra\" data-href=\"https:\/\/medium.com\/@vedantmisra\" data-anchor-type=\"2\" data-user-id=\"3fbbe05bf3d6\" data-action-value=\"3fbbe05bf3d6\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Vedant Misra<\/a> has <a href=\"https:\/\/thinkgrowth.org\/ai-and-big-data-are-changing-our-attention-spans-959923adad7f\" data-href=\"https:\/\/thinkgrowth.org\/ai-and-big-data-are-changing-our-attention-spans-959923adad7f\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">written about this here<\/a>).<\/p><p id=\"9eef\" class=\"graf graf--p graf-after--p\">There will still be enormous returns in content marketing, but it\u2019s going to take a new approach. One that is powered by better research, smarter prioritization, and emerging technology.<\/p><p id=\"6717\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">2.<strong class=\"markup--strong markup--p-strong\"> Content marketers are becoming \u201cfarmers.\u201d <\/strong><a href=\"https:\/\/medium.com\/@bylymari\" data-href=\"https:\/\/medium.com\/@bylymari\" data-anchor-type=\"2\" data-user-id=\"2a4b6c2b449f\" data-action-value=\"2a4b6c2b449f\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Lymari Morales<\/a> calls this emerging role \u201c<a href=\"https:\/\/medium.com\/digital-trends-index\/need-to-grow-the-audience-for-your-work-bring-in-a-growth-editor-e82dcf687332\" data-href=\"https:\/\/medium.com\/digital-trends-index\/need-to-grow-the-audience-for-your-work-bring-in-a-growth-editor-e82dcf687332\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">growth editor<\/a>.\u201d You might also think of this as a content strategist role. Whatever the word you use, it\u2019s going to begin playing a bigger part in content marketing success. And the basic gist of this: <strong class=\"markup--strong markup--p-strong\">Content marketers will be spending more time optimizing existing content than creating new content<\/strong>.<\/p><p id=\"c2b9\" class=\"graf graf--p graf-after--p\">As always, we can blame Google for this shift.<\/p><p id=\"8d89\" class=\"graf graf--p graf-after--p\">Over the past few years, Google has been steadily changing its algorithm to favor actual experts (vs. content factories). 3\u20135 years ago, the SEO opportunity was in publishing a high number of blog posts that targeted long-tail keywords. Scoop up enough keywords with 100 searches a month, and you could build a nice little following. This doesn\u2019t work so well today.<\/p><p id=\"b764\" class=\"graf graf--p graf-after--p\">As Google\u2019s algorithm gets better at spotting real experts, it\u2019s started to favor \u201c<a href=\"https:\/\/research.hubspot.com\/reports\/topic-clusters-seo\" data-href=\"https:\/\/research.hubspot.com\/reports\/topic-clusters-seo\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">topic clusters<\/a>.\u201d Topic clusters have three parts: pillar content, cluster content, hyperlinks that group them together.<\/p><figure id=\"9de2\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*XChXPndwLDNBQqeopkE01A.png\" data-width=\"696\" data-height=\"390\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*XChXPndwLDNBQqeopkE01A.png\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">This image is from <a href=\"https:\/\/research.hubspot.com\/reports\/topic-clusters-seo\" data-href=\"https:\/\/research.hubspot.com\/reports\/topic-clusters-seo\" class=\"markup--anchor markup--figure-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--figure-em\">Topic Clusters: The Next Evolution of\u00a0SEO<\/em><\/a><\/figcaption><\/figure><p id=\"0f49\" class=\"graf graf--p graf-after--figure\"><a href=\"https:\/\/medium.com\/@emmajs24\" data-href=\"https:\/\/medium.com\/@emmajs24\" data-anchor-type=\"2\" data-user-id=\"879fe0d1ce9\" data-action-value=\"879fe0d1ce9\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--p-user\" target=\"_blank\">Emma Brudner<\/a> is the lead editor on HubSpot\u2019s marketing blog, a blog responsible for 4.5 million monthly visitors. She says this shift has completely changed the way her team produces content:<\/p><blockquote id=\"4e8e\" class=\"graf graf--pullquote graf--hasDropCapModel graf--startsWithDoubleQuote graf-after--p\"><div>\u201cOur topic brainstorms used to be: \u2018Hey let\u2019s write about X because it seems like an interesting topic.\u2019 Today that conversation is: \u2018We should write about X because it\u2019s an organic gap that fits in Y\u00a0cluster.<\/div><\/blockquote><p id=\"c8fc\" class=\"graf graf--p graf-after--pullquote\">In addition, there\u2019s far less focus on new content, and more attention paid to promotion:<\/p><blockquote id=\"bbbd\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cMy editors are spending a lot less time perfecting the written word and a lot more time experimenting with promotion, optimizing promotional channels (email, paid, etc.), and improving existing content.\u201d<\/div><\/blockquote><p id=\"d129\" class=\"graf graf--p graf-after--pullquote\">She struggles with the word \u201cwriter\u201d and \u201ccontent marketer\u201d because, she says, \u201cPeople who are successful in these roles are a lot more strategic and less creative.\u201d Emma\u2019s team is on the leading edge of some of these changes. If you want to read more about what they\u2019re learning, check out these posts:<\/p><p id=\"7e7a\" class=\"graf graf--p graf-after--p\">\u2705 <a href=\"https:\/\/blog.hubspot.com\/marketing\/historical-blog-seo-conversion-optimization\" data-href=\"https:\/\/blog.hubspot.com\/marketing\/historical-blog-seo-conversion-optimization\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">The Blogging Tactic No One Is Talking About: Optimizing the Past<\/a><\/p><p id=\"12cf\" class=\"graf graf--p graf-after--p\">\u2705 <a href=\"https:\/\/blog.hubspot.com\/customers\/double-digit-jumps-in-organic-traffic-with-content-pillars\" data-href=\"https:\/\/blog.hubspot.com\/customers\/double-digit-jumps-in-organic-traffic-with-content-pillars\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">How to Get Double-Digit Jumps in Organic Traffic with Content Pillars<\/a><\/p><p id=\"8cc4\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--p\">3.<strong class=\"markup--strong markup--p-strong\">Content delivery and distribution is playing as big a role in the value of content as the content itself. <\/strong>Everyone knows you need \u201cquality content\u201d to execute a good content strategy. But in a world drowning in content, even really good stuff ends up just adding to the noise. What becomes valuable then is not more content, but helping people find the right content.<\/p><p id=\"1822\" class=\"graf graf--p graf-after--p\">Kevin Kelly, a digital culture expert, has identified <a href=\"https:\/\/thinkgrowth.org\/look-over-here-8-strategies-to-survive-in-the-attention-economy-7a08a9ad39cd\" data-href=\"https:\/\/thinkgrowth.org\/look-over-here-8-strategies-to-survive-in-the-attention-economy-7a08a9ad39cd\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">8 intangibles that make content valuable<\/a>, 7 of these are particularly relevant for marketers. They are:<\/p><ul class=\"postList\"><li id=\"be9d\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Immediacy<\/strong>: providing priority access or immediate delivery<\/li><li id=\"7bd4\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Personalization<\/strong>: content tailored just for you<\/li><li id=\"347f\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Interpretation<\/strong>: providing support, guidance and context in understanding a given area<\/li><li id=\"c60a\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Authenticity<\/strong>: real advice from real people<\/li><li id=\"85cd\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Accessibility<\/strong>: content that can be accessed when and where you want it<\/li><li id=\"402a\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Embodiment<\/strong>: content delivered in the form you want it<\/li><li id=\"4cdd\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Findability<\/strong>: filtering out irrelevant content<\/li><\/ul><p><a href=\"https:\/\/thinkgrowth.org\/look-over-here-8-strategies-to-survive-in-the-attention-economy-7a08a9ad39cd\" data-href=\"https:\/\/thinkgrowth.org\/look-over-here-8-strategies-to-survive-in-the-attention-economy-7a08a9ad39cd\" class=\"markup--anchor markup--mixtapeEmbed-anchor\" title=\"https:\/\/thinkgrowth.org\/look-over-here-8-strategies-to-survive-in-the-attention-economy-7a08a9ad39cd\"><strong class=\"markup--strong markup--mixtapeEmbed-strong\">8 Strategies to Survive in the Attention Economy<\/strong><br \/><em class=\"markup--em markup--mixtapeEmbed-em\">Every business today is in constant competition for today\u2019s most valuable resource\u200a\u2014\u200aattention. And you are competing\u2026<\/em>thinkgrowth.org<\/a><\/p><p id=\"ba1d\" class=\"graf graf--p graf-after--mixtapeEmbed\">Notice that the majority of these, with the exception of authenticity and interpretation, are about how content is discovered. We started to see this trend emerge a few years ago as newsletter roundups starting growing in popularity, people needed a way to sort out what content mattered and good curators delivered.<\/p><p id=\"f517\" class=\"graf graf--p graf-after--p\">Now imagine this trend accelerated by machine learning. Watch out for <a href=\"https:\/\/medium.com\/startup-grind\/ai-will-shape-the-future-of-media-heres-how-768e4f10a00f\" data-href=\"https:\/\/medium.com\/startup-grind\/ai-will-shape-the-future-of-media-heres-how-768e4f10a00f\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">better-than-ever recommendation engines, tools that can predict virality<\/a>, content that is delivered in your preferred format, bots that send you the right content at the right moment, and marketing communications tailored to your precise preferences.<\/p><blockquote id=\"7469\" class=\"graf graf--pullquote graf-after--p\"><div>With AI the marketing dream of \u201cright message to the right person at the right time\u201d could finally become a\u00a0reality.<\/div><\/blockquote><p id=\"9c20\" class=\"graf graf--p graf--hasDropCapModel graf--hasDropCap graf-after--pullquote\">4.<strong class=\"markup--strong markup--p-strong\"> The composition of content marketing teams is changing. <\/strong>All of this means that content marketing is going to become a lot more fun. Instead of content marketing teams being comprised of underpaid college grads churning out listicles and roundup posts we\u2019ll see diverse teams working on really interesting problems.<\/p><p id=\"88bf\" class=\"graf graf--p graf-after--p\">Content teams will be comprised of developers, designers, artificial intelligence engineers, SEOs, videographers, and yes\u2026plenty of writers\u00a0:) This trend too is already in motion:<\/p><ul class=\"postList\"><li id=\"230d\" class=\"graf graf--li graf-after--p\">At <a href=\"https:\/\/medium.com\/@RJMetrics\" data-href=\"https:\/\/medium.com\/@RJMetrics\" data-anchor-type=\"2\" data-user-id=\"40e2c85fbd51\" data-action-value=\"40e2c85fbd51\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--li-user\" target=\"_blank\">RJMetrics<\/a>, we hired a data scientist who spent about 50% of his time doing research for our <a href=\"https:\/\/rjmetrics.com\/resources\" data-href=\"https:\/\/rjmetrics.com\/resources\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">benchmark reports<\/a>.<\/li><li id=\"5246\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/medium.com\/@unbounce\" data-href=\"https:\/\/medium.com\/@unbounce\" data-anchor-type=\"2\" data-user-id=\"eda63d4fd9e4\" data-action-value=\"eda63d4fd9e4\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--li-user\" target=\"_blank\">Unbounce<\/a> has been experimenting with <a href=\"http:\/\/unbounce.com\/campaign-strategy\/why-unbounce-is-creating-on-demand-content\/\" data-href=\"http:\/\/unbounce.com\/campaign-strategy\/why-unbounce-is-creating-on-demand-content\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">on-demand video content<\/a>.<\/li><li id=\"18f3\" class=\"graf graf--li graf-after--li\">Zendesk is building <a href=\"https:\/\/relate.zendesk.com\/\" data-href=\"https:\/\/relate.zendesk.com\/\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\">a compelling editorial brand with Relate<\/a>.<\/li><li id=\"e0e2\" class=\"graf graf--li graf-after--li\"><a href=\"https:\/\/medium.com\/@HubSpot\" data-href=\"https:\/\/medium.com\/@HubSpot\" data-anchor-type=\"2\" data-user-id=\"8732e73183e5\" data-action-value=\"8732e73183e5\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--li-user\" target=\"_blank\">HubSpot<\/a> is doing really cool things with educational content on HubSpot Academy, video content on social, and continuing to up its podcast game.<\/li><\/ul><p id=\"218b\" class=\"graf graf--p graf-after--li\">These are still content projects, but the skill sets have expanded far beyond just writing. These projects are ambitious and fun. They involve creative, talented people working together to solve problems in interesting ways. It means that content strategy will be much more than a calendar of blog posts. It also means that innovating will be harder than just brainstorming new Snapchat ideas, but this is good. <strong class=\"markup--strong markup--p-strong\">Content marketing teams in the future will have access to more resources and talent than we\u2019ve ever had. It\u2019s going to get a lot more fun.<\/strong><\/p><h4 id=\"1b5e\" class=\"graf graf--h4 graf-after--p\">Learned something? Click the \ud83d\udc4f to say \u201cthanks!\u201d and help others find this\u00a0article.<\/h4><p id=\"0305\" class=\"graf graf--p graf-after--h4\"><em class=\"markup--em markup--p-em\">The <\/em><a href=\"https:\/\/thinkgrowth.org\/what-does-the-future-of-content-marketing-look-like-spoiler-far-less-content-812ac1512902\" data-href=\"https:\/\/thinkgrowth.org\/what-does-the-future-of-content-marketing-look-like-spoiler-far-less-content-812ac1512902\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><em class=\"markup--em markup--p-em\">first version<\/em><\/a><em class=\"markup--em markup--p-em\"> of this post was published last year and it struck a nerve. Since then I\u2019ve given a few talks on this topic, had dozens of calls with marketing leaders who are thinking through their content strategy, and updated my own thinking on this just a bit. I\u2019d love to hear your thoughts in the comments.<\/em><\/p><figure id=\"90f4\" class=\"graf graf--figure graf--iframe graf-after--p graf--trailing\"><\/figure><\/div>","created_at":"2018-11-07T16:50:11+0000","updated_at":"2018-11-07T16:50:11+0000","published_at":"2017-09-11T14:32:26+0000","published_by":["Janessa Lantz"],"annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"thinkgrowth.org","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*stka6SaLO8bBy9Yd-zaCgw.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":502,"title":"JavaScript Math.random Normal distribution (Gaussian bell curve)?","url":"https:\/\/stackoverflow.com\/questions\/25582882\/javascript-math-random-normal-distribution-gaussian-bell-curve","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        23&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I want to know if the JavaScript function <code>Math.random<\/code> uses a normal (vs. uniform) distribution or not.<\/p>\n<p>If not, how can I get numbers which use a normal distribution? I haven't found a clear answer on the Internet, for an algorithm to create random normally-distributed numbers.<\/p>\n<p>I want to rebuild a Schmidt-machine (German physicist). The machine produces random numbers of 0 or 1, and they have to be normally-distributed so that I can draw them as a Gaussian bell curve.<\/p>\n<p>For example, the random function produces 120 numbers (0 or 1) and the average (mean) of these summed values has to be near 60.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/25582882\/revisions\" title=\"show all edits to this post\">edited Oct 10 at 7:14<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3672465\/beejor\">Beejor<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            2,7702219&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Aug 30 '14 at 13:30&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3902521\/mangooxx\">Mangooxx<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1821312&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"16\">&#13;\n                                16 Answers&#13;\n                                16&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        -56&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>First of all sorry for not properly fixing the response. I forgot that this page is indexed and the answer MUST be correct as well as being useful to the the person who asked the question. I hope some of you will read revised explanation and remove negative scores or provide comments how this response could be further improved.<\/p>\n<blockquote>\n  <p>I want to know if the JavaScript function Math.random is normal distribution or not<\/p>\n<\/blockquote>\n<p>Javascript Math.random <strong><em>is<\/em> <em>not<\/em><\/strong> a Normal Distribution(Gaussian bell curve). From  <a href=\"https:\/\/tc39.github.io\/ecma262\/#sec-math.random\" rel=\"noreferrer\">ES 2015, 20.2.2.27<\/a> \"Returns a Number value with positive sign, greater than or equal to 0 but less than 1, chosen randomly or pseudo randomly with approximately uniform distribution over that range, using an implementation-dependent algorithm or strategy. This function takes no arguments.\" So the provided collection when n is high enough we will get approximately uniform distribution. All values in the interval will have equal probability of appearance(straight line parallel to the x axis, denoting number between 0.0 and 1.0).<\/p>\n<blockquote>\n  <p>how can I get numbers which are normal distribution<\/p>\n<\/blockquote>\n<p>There are several ways of getting collection of numbers with a normal distribution. As answered by <a href=\"https:\/\/stackoverflow.com\/users\/1222381\/maxwell-collard\">Maxwell Collard<\/a> the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Box%E2%80%93Muller_transform\" rel=\"noreferrer\">Box-Muller transform<\/a>\ndoes transform uniform distribution to normal distribution(the code can be found in <a href=\"https:\/\/stackoverflow.com\/users\/1222381\/maxwell-collard\">Maxwell Collard<\/a> answer).  <\/p>\n<p>An answer to another stackoverflow answer to a <a href=\"https:\/\/stackoverflow.com\/questions\/75677\/converting-a-uniform-distribution-to-a-normal-distribution\">question<\/a> has a reply with other uniform distribution to normal distribution algorithms. Such as:\nZiggurat,\nRatio-of-uniforms,\nInverting the CDF\nBesides one of the answers says that: says: <\/p>\n<blockquote>\n  <p>The Ziggurat algorithm is pretty efficient for this, although the Box-Muller transform is easier to implement from scratch (and not crazy slow).<\/p>\n<\/blockquote>\n<p>And finally <\/p>\n<blockquote>\n  <p>I want to rebuilt a Schmidt-machine (German physicist), the machine produces random numbers of 0 or 1 and they have to be normal distributed so I can draw them in Gaussian bell curve.<\/p>\n<\/blockquote>\n<p>When we have only two values (0 or 1) Gaussian curve looks the same as uniform distribution with 2 possible values. That is why a simple <\/p>\n<pre>function randomZero_One(){\n    return Math.round(Math.random());\n}\n<\/pre>\n<p>would suffice. It would return pseudo-randomly with approximately equal probability values 0 and 1.  <\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/25582948\/revisions\" title=\"show all edits to this post\">edited Jun 5 at 9:49<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 30 '14 at 13:38&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2878846\/kodvin\">kodvin<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        68&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Since this is the first Google result for \"js gaussian random\" in my experience, I feel an obligation to give an actual answer to that query.<\/p>\n<p>The <a href=\"https:\/\/en.wikipedia.org\/wiki\/Box%E2%80%93Muller_transform\" rel=\"noreferrer\">Box-Muller transform<\/a> converts two independent uniform variates on (0, 1) into two standard Gaussian variates (mean 0, variance 1). This probably isn't very performant because of the <code>sqrt<\/code>, <code>log<\/code>, and <code>cos<\/code> calls, but this method is superior to the central limit theorem approaches (summing N uniform variates) because it doesn't restrict the output to the bounded range (-N\/2, N\/2). It's also really simple:<\/p>\n<pre>\/\/ Standard Normal variate using Box-Muller transform.\nfunction randn_bm() {\n    var u = 0, v = 0;\n    while(u === 0) u = Math.random(); \/\/Converting [0,1) to (0,1)\n    while(v === 0) v = Math.random();\n    return Math.sqrt( -2.0 * Math.log( u ) ) * Math.cos( 2.0 * Math.PI * v );\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/36481059\/revisions\" title=\"show all edits to this post\">edited Jul 23 '17 at 11:25<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1298002\/mahdi\">Mahdi<\/a>&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Apr 7 '16 at 15:41&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1222381\/maxwell-collard\">Maxwell Collard<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,0021910&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        32&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I wanted to have approximately gaussian random numbers between 0 and 1, and after <a href=\"http:\/\/jsfiddle.net\/xx9bh8Lz\/1\/\" rel=\"noreferrer\">many tests<\/a> I found this to be the best:<\/p>\n<pre>function gaussianRand() {\n  var rand = 0;\n  for (var i = 0; i &lt; 6; i += 1) {\n    rand += Math.random();\n  }\n  return rand \/ 6;\n}\n<\/pre>\n<hr \/><p>And as a bonus:<\/p>\n<pre>function gaussianRandom(start, end) {\n  return Math.floor(start + gaussianRand() * (end - start + 1));\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Aug 28 '16 at 1:36&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/407213\/dorian\">Dorian<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            12.4k37281&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        20&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Building on Maxwell's Answer, this code uses the <a href=\"https:\/\/i.stack.imgur.com\/nTzY9m.png\" rel=\"noreferrer\">Box\u2013Muller transform<\/a> to give you a normal distribution between 0 and 1 inclusive. It just resamples the values if it's more than 3.6 standard deviations away (less than 0.02% chance).<\/p>\n<pre>function randn_bm() {\n    var u = 0, v = 0;\n    while(u === 0) u = Math.random(); \/\/Converting [0,1) to (0,1)\n    while(v === 0) v = Math.random();\n    let num = Math.sqrt( -2.0 * Math.log( u ) ) * Math.cos( 2.0 * Math.PI * v );\n    num = num \/ 10.0 + 0.5; \/\/ Translate to 0 -&gt; 1\n    if (num &gt; 1 || num &lt; 0) return randn_bm(); \/\/ resample between 0 and 1\n    return num;\n}\n<\/pre>\n<p><a href=\"https:\/\/i.stack.imgur.com\/nTzY9m.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/nTzY9m.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>n = 100<\/strong><\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/JuMCtm.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/JuMCtm.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>n = 10,000<\/strong><\/p>\n<p><a href=\"https:\/\/i.stack.imgur.com\/xfZhkm.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/xfZhkm.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<p><strong>n = 10,000,000<\/strong><\/p>\n<p>This version allows you to give a min, max, and skew factor. See my usage examples at the bottom.<\/p>\n<pre>function randn_bm(min, max, skew) {\n    var u = 0, v = 0;\n    while(u === 0) u = Math.random(); \/\/Converting [0,1) to (0,1)\n    while(v === 0) v = Math.random();\n    let num = Math.sqrt( -2.0 * Math.log( u ) ) * Math.cos( 2.0 * Math.PI * v );\n    num = num \/ 10.0 + 0.5; \/\/ Translate to 0 -&gt; 1\n    if (num &gt; 1 || num &lt; 0) num = randn_bm(min, max, skew); \/\/ resample between 0 and 1 if out of range\n    num = Math.pow(num, skew); \/\/ Skew\n    num *= max - min; \/\/ Stretch to fill range\n    num += min; \/\/ offset to min\n    return num;\n}\n<\/pre>\n<p><a href=\"https:\/\/i.stack.imgur.com\/7jmROm.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/7jmROm.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<pre>randn_bm(-500, 1000, 1);\n<\/pre>\n<p><a href=\"https:\/\/i.stack.imgur.com\/nCexKm.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/nCexKm.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<pre>randn_bm(10, 20, 0.25);\n<\/pre>\n<p><a href=\"https:\/\/i.stack.imgur.com\/QFQMgm.png\" rel=\"noreferrer\"><img src=\"https:\/\/i.stack.imgur.com\/QFQMgm.png\" alt=\"enter image description here\" \/><\/a><\/p>\n<pre>randn_bm(10, 20, 3);\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/49434653\/revisions\" title=\"show all edits to this post\">edited May 16 at 21:44<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Mar 22 at 17:14&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4975772\/joshuakcockrell\">joshuakcockrell<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            964822&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        15&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The Javascript Math.random() pseudorandom function returns variates that are equally distributed between 0 and 1. To get a Gaussian distribution  I use this:<\/p>\n<pre>\/\/ returns a gaussian random function with the given mean and stdev.\nfunction gaussian(mean, stdev) {\n    var y2;\n    var use_last = false;\n    return function() {\n        var y1;\n        if(use_last) {\n           y1 = y2;\n           use_last = false;\n        }\n        else {\n            var x1, x2, w;\n            do {\n                 x1 = 2.0 * Math.random() - 1.0;\n                 x2 = 2.0 * Math.random() - 1.0;\n                 w  = x1 * x1 + x2 * x2;               \n            } while( w &gt;= 1.0);\n            w = Math.sqrt((-2.0 * Math.log(w))\/w);\n            y1 = x1 * w;\n            y2 = x2 * w;\n            use_last = true;\n       }\n       var retval = mean + stdev * y1;\n       if(retval &gt; 0) \n           return retval;\n       return -retval;\n   }\n}\n\/\/ make a standard gaussian variable.     \nvar standard = gaussian(100, 15);\n\/\/ make a bunch of standard variates\nfor(i=0; i&lt;2000; i++) {\n    console.log(standard());\n}\n<\/pre>\n<p>I think I got this from Knuth.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Feb 24 '16 at 10:14&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5973776\/j&#xf3;hann-&#xde;&#xf3;rir-j&#xf3;hannsson\">J\u00f3hann \u00de\u00f3rir J\u00f3hannsson<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        12&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Function that utilises the central limit theorem.<\/p>\n<pre>function normal(mu, sigma, nsamples){\n    if(!nsamples) nsamples = 6\n    if(!sigma) sigma = 1\n    if(!mu) mu=0\n    var run_total = 0\n    for(var i=0 ; i&lt;nsamples ; i++){\n       run_total += Math.random()\n    }\n    return sigma*(run_total - nsamples\/2)\/(nsamples\/2) + mu\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Nov 6 '15 at 13:29&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1972493\/joe\">Joe<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            7231026&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        8&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>You are confusing the output of the function (which is a uniform distribution between 0 and 1) with the need to generate a Gaussian distribution by repeatedly drawing random numbers that are either 0 or 1 - after a large number of trials, their sum will be approximately normally distributed.<\/p>\n<p>You can use the <code>Math.random()<\/code> function, then round the result to an integer: if it's &lt; 0.5, return 0; if its &gt;= 0.5, return 1. Now you have equal probabilities of zero and one, and you can continue with the approach you described in your question.<\/p>\n<p>Just to clarify: I don't think it's possible to have an algorithm that produces either 0's or 1's in a normally distributed way - normal distribution requires a continuous variable.<\/p>\n<p>When you do the above for say 120 numbers, you will on average get 60 1's and 60 0's. The actual distribution you get will be the binomial distribution with a mean of 60 and a standard deviation of<\/p>\n<pre>stdev = sqrt(p(1-p)N) = 5.48\n<\/pre>\n<p>The probability of a particular number <code>k<\/code> when you have <code>n<\/code> samples with probability <code>p<\/code> (which we fixed at 0.5) is<\/p>\n<pre>p = n! \/ ((n-k)! k!) p^k (1-p)^(n-k)\n<\/pre>\n<p>When p = 0.5, you end up with just the binomial coefficients - which approach the normal distribution for n &gt; 30, typically.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/25582921\/revisions\" title=\"show all edits to this post\">edited Aug 30 '14 at 14:06<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 30 '14 at 13:35&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1967396\/floris\">Floris<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            39.7k44587&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        7&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><a href=\"http:\/\/www.ecma-international.org\/ecma-262\/5.1\/#sec-15.8.2.14\" rel=\"nofollow noreferrer\">From the spec:<\/a><\/p>\n<blockquote>\n  <p>15.8.2.14 random ( )<\/p>\n  \n  <p>Returns a Number value with positive sign, greater than or equal to 0\n  but less than 1, chosen randomly or pseudo randomly <strong>with\n  approximately uniform distribution over that range<\/strong>, using an\n  implementation-dependent algorithm or strategy. This function takes no\n  arguments.<\/p>\n<\/blockquote>\n<p>So, it's a <strong>uniform<\/strong> distribution, not normal or Gaussian. That's what you're going to find in just about any standard random number facility in any basic language runtime outside of specialized statistics libraries.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/25582903\/revisions\" title=\"show all edits to this post\">edited Apr 16 at 1:01<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/986862\/andre-figueiredo\">Andre Figueiredo<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            6,94143159&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 30 '14 at 13:32&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/182668\/pointy\">Pointy<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            310k44448510&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        2&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>And a single line example: <\/p>\n<pre>Math.sqrt(-2 * Math.log(Math.random()))*Math.cos((2*Math.PI) * Math.random())\n<\/pre>\n<p>and a Fiddle\n<a href=\"https:\/\/jsfiddle.net\/rszgjqf8\/\" rel=\"nofollow\">https:\/\/jsfiddle.net\/rszgjqf8\/<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Oct 18 '16 at 1:03&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2317607\/unsalted\">unsalted<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            146111&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        2&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>For those interested in generating values of a normal distrubution, I would recommend checking this implementation of the Ziggurat algorithm in JavaScript: <a href=\"https:\/\/www.npmjs.com\/package\/node-ziggurat\" rel=\"nofollow\">https:\/\/www.npmjs.com\/package\/node-ziggurat<\/a><\/p>\n<p>The code of found in the author's page is:<\/p>\n<pre>function Ziggurat(){\nvar jsr = 123456789;\nvar wn = Array(128);\nvar fn = Array(128);\nvar kn = Array(128);\nfunction RNOR(){\n  var hz = SHR3();\n  var iz = hz &amp; 127;\n  return (Math.abs(hz) &lt; kn[iz]) ? hz * wn[iz] : nfix(hz, iz);\n}\nthis.nextGaussian = function(){\n  return RNOR();\n}\nfunction nfix(hz, iz){\n  var r = 3.442619855899;\n  var r1 = 1.0 \/ r;\n  var x;\n  var y;\n  while(true){\n    x = hz * wn[iz];\n    if( iz == 0 ){\n      x = (-Math.log(UNI()) * r1); \n      y = -Math.log(UNI());\n      while( y + y &lt; x * x){\n        x = (-Math.log(UNI()) * r1); \n        y = -Math.log(UNI());\n      }\n      return ( hz &gt; 0 ) ? r+x : -r-x;\n    }\n    if( fn[iz] + UNI() * (fn[iz-1] - fn[iz]) &lt; Math.exp(-0.5 * x * x) ){\n      return x;\n    }\n    hz = SHR3();\n    iz = hz &amp; 127;\n    if( Math.abs(hz) &lt; kn[iz]){\n      return (hz * wn[iz]);\n    }\n  }\n}\nfunction SHR3(){\n  var jz = jsr;\n  var jzr = jsr;\n  jzr ^= (jzr &lt;&lt; 13);\n  jzr ^= (jzr &gt;&gt;&gt; 17);\n  jzr ^= (jzr &lt;&lt; 5);\n  jsr = jzr;\n  return (jz+jzr) | 0;\n}\nfunction UNI(){\n  return 0.5 * (1 + SHR3() \/ -Math.pow(2,31));\n}\nfunction zigset(){\n  \/\/ seed generator based on current time\n  jsr ^= new Date().getTime();\n  var m1 = 2147483648.0;\n  var dn = 3.442619855899;\n  var tn = dn;\n  var vn = 9.91256303526217e-3;\n  var q = vn \/ Math.exp(-0.5 * dn * dn);\n  kn[0] = Math.floor((dn\/q)*m1);\n  kn[1] = 0;\n  wn[0] = q \/ m1;\n  wn[127] = dn \/ m1;\n  fn[0] = 1.0;\n  fn[127] = Math.exp(-0.5 * dn * dn);\n  for(var i = 126; i &gt;= 1; i--){\n    dn = Math.sqrt(-2.0 * Math.log( vn \/ dn + Math.exp( -0.5 * dn * dn)));\n    kn[i+1] = Math.floor((dn\/tn)*m1);\n    tn = dn;\n    fn[i] = Math.exp(-0.5 * dn * dn);\n    wn[i] = dn \/ m1;\n  }\n}\nzigset();\n}\n<\/pre>\n<p>Create a Ziggurat.js file and then:<\/p>\n<pre>var z = new Ziggurat();\nz.nextGaussian();\n<\/pre>\n<p>For me it's working just perfect and as I had read in Wikipedia, this is a more efficient algorithm than the Box-Muller.<\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Box%E2%80%93Muller_transform\" rel=\"nofollow\">enter link description here<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/40044129\/revisions\" title=\"show all edits to this post\">edited Oct 19 '16 at 18:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Oct 14 '16 at 13:14&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5286467\/eduardo-g-r\">Eduardo G.R.<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I found this library that includes lots of useful Random functions.  You can either install it via simjs from npm, or just take the random-node-*.js file out directly for what you need.<\/p>\n<p><a href=\"http:\/\/www.simjs.com\/random.html\" rel=\"nofollow\">http:\/\/www.simjs.com\/random.html<\/a>\n<a href=\"http:\/\/www.simjs.com\/download.html\" rel=\"nofollow\">http:\/\/www.simjs.com\/download.html<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Sep 2 '15 at 14:32&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3250242\/mark\">Mark<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            139110&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>This is my solution to the problem, using the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Marsaglia_polar_method\" rel=\"nofollow noreferrer\">Marsaglia polar method<\/a>. The range depends on the parameters you give, without parameters it almost never generates anything outside of the range.<\/p>\n<p>As it generates two normally distributed numbers per iteration, I declared a variable under window.temp.spareNormal to grab the spare one if it's there. Might not be the best location for it, but hey.<\/p>\n<p>You'd probably have to round the result in order to get what you want.<\/p>\n<pre>window.temp = {\n    spareNormal: undefined\n};\nMath.normal = function (mean, standardDeviation) {\n    let q, u, v, p;\n    mean = mean || 0.5;\n    standardDeviation = standardDeviation || 0.125;\n    if (typeof temp.spareNormal !== 'undefined') {\n        v = mean + standardDeviation * temp.spareNormal;\n        temp.spareNormal = undefined;\n        return v;\n    }\n    do  {\n        u = 2.0 * Math.random() - 1.0;\n        v = 2.0 * Math.random() - 1.0;\n        q = u * u + v * v;\n    } while (q &gt;= 1.0 || q === 0);\n    p = Math.sqrt(-2.0 * Math.log(q) \/ q);\n    temp.spareNormal = v * p;\n    return mean + standardDeviation * u * p;\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jun 1 '17 at 19:44&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1535091\/sjor\">Sjor<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I have tested several functions with the right configuration all work similarly and well.<\/p>\n<p><a href=\"http:\/\/jsfiddle.net\/p3y40gf3\/29\/\" rel=\"nofollow noreferrer\">http:\/\/jsfiddle.net\/p3y40gf3\/29\/<\/a><\/p>\n<p>Central limit is nice, must be with (n=3 for 6) and 12 for 12 to look as others. I configured others also to (6) or 12 or 1\/12 as standard deviation, not sure why 12. <\/p>\n<p>Central limit is a tiny bit less centered than Box\/Muller and Ziggurat.<\/p>\n<p>Box\/Muller and Ziggurat look exactly the same<\/p>\n<p>this variant by Joe(<a href=\"https:\/\/stackoverflow.com\/a\/33567961\/466363\">https:\/\/stackoverflow.com\/a\/33567961\/466363<\/a>) does standard deviation correctly:<\/p>\n<pre>function normal(mu, sigma, nsamples){ \/\/ using central limit\n    if(!nsamples) nsamples = 3\n    if(!sigma) sigma = 1\n    if(!mu) mu=0\n    var run_total = 0\n    for(var i=0 ; i&lt;nsamples ; i++){\n       run_total += Math.random()\n    }\n    return sigma*(run_total - nsamples\/2)\/(nsamples\/2) + mu\n}\n<\/pre>\n<p>Ziggurat is also nice but needs to be adjusted from z score to from 0 to 1 looks like it makes good numbers. <\/p>\n<p>Box\/Muller clipped is good but gives few repeated numbers at clipped edges\nbut it is very similar to others, \nincorrect random numbers should be <strong>discarded<\/strong> not clipped.<\/p>\n<pre>function randn_bm() {\n    var u = 0, v = 0;\n    while(u === 0) u = Math.random(); \/\/Converting [0,1) to (0,1)\n    while(v === 0) v = Math.random();\n    let num = Math.sqrt( -2.0 * Math.log( u ) ) * Math.cos( 2.0 * Math.PI * v );\n    num = num \/ 6.0 + 0.5; \/\/ Translate to 0 -&gt; 1 \/\/ changed here 10 to 6\n    if(num&gt;1||num&lt;0) return randn_bm(); return num; \/\/ bad random numbers should be discared not clipped\n    \/\/return Math.max(Math.min(num, 1), 0); \/\/ cap between 0 and 1\n}\n<\/pre>\n<p>Central limit variant it is called Bates distribution that is average\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Bates_distribution\" rel=\"nofollow noreferrer\">https:\/\/en.wikipedia.org\/wiki\/Bates_distribution<\/a><\/p>\n<p>not confused with Irwin Hall that is a sum\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Irwin%E2%80%93Hall_distribution\" rel=\"nofollow noreferrer\">https:\/\/en.wikipedia.org\/wiki\/Irwin%E2%80%93Hall_distribution<\/a><\/p>\n<p><a href=\"https:\/\/en.wikipedia.org\/wiki\/Normal_distribution#Generating_values_from_normal_distribution\" rel=\"nofollow noreferrer\">https:\/\/en.wikipedia.org\/wiki\/Normal_distribution#Generating_values_from_normal_distribution<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/50365267\/revisions\" title=\"show all edits to this post\">edited May 16 at 11:10<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered May 16 at 7:56&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/466363\/shimon-doodkin\">Shimon Doodkin<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            2,5342424&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>A non verbose function to sample a random value from a Gaussian distribution I wrote some time ago:<\/p>\n<pre>function gaussianRandom(mean, sigma) {\n  let u = Math.random()*0.682;\n  return ((u % 1e-8 &gt; 5e-9 ? 1 : -1) * (Math.sqrt(-Math.log(Math.max(1e-9, u)))-0.618))*1.618 * sigma + mean;\n}\n<\/pre>\n<p>It should work if you clamp the values to the range you want.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/50047525\/revisions\" title=\"show all edits to this post\">edited May 20 at 8:31<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Apr 26 at 16:05&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/942360\/supereggbert\">SuperEggbert<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        -1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<pre>let iset = 0;\nlet gset;\nfunction randn() {\n   let v1, v2, fac, rsq;\n   if (iset == 0) {\n   do {\n     v1 = 2.0*Math.random() - 1.0;\n     v2 = 2.0*Math.random() - 1.0;\n     rsq = v1*v1+v2*v2;\n   } while ((rsq &gt;= 1.0) || (rsq == 0));\n   fac = Math.sqrt(-2.0*Math.log(rsq)\/rsq);\n   gset = v1*fac;\n   iset = 1;\n   return v2*fac;\n } else {\n   iset = 0;\n   return gset;\n }\n}\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jun 17 '17 at 21:52&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/8177276\/luke-czapla\">Luke Czapla<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        -3&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<pre>\/\/This is what I use for a Normal-ish distribution random function.\nfunction normal_random(){\n  var pos = [ Math.random(), Math.random() ];\n  while ( Math.sin( pos[0] * Math.PI ) &gt; pos[1] ){\n    pos = [ Math.random(), Math.random() ];\n  }   \n  return pos[0];\n};\n<\/pre>\n<p>This function returns a value between 0 and 1. Values near 0.5 are returned most often.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Dec 17 '14 at 23:27&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1065762\/bassim-huis\">Bassim Huis<\/a>&#13;\n        &#13;<\/div>","created_at":"2018-11-07T16:50:11+0000","updated_at":"2018-11-07T16:50:11+0000","annotations":[],"mimetype":"text\/html","reading_time":15,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":503,"title":"galaxykate\/tracery","url":"https:\/\/github.com\/galaxykate\/tracery\/blob\/master\/readme.md","content":"<h2><a id=\"user-content-a-text-expansion-library\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-text-expansion-library\"><\/a>A text-expansion library<\/h2>\n<p>There are many new examples of Tracery <a href=\"http:\/\/www.crystalcodepalace.com\/tracery.html\" title=\"Examples\" rel=\"nofollow\">in use<\/a>\nI also have an exciting new <em>interactive<\/em> <a href=\"http:\/\/www.crystalcodepalace.com\/traceryTut.html\" title=\"Tutorial\" rel=\"nofollow\">tutorial<\/a><\/p>\n<p>I strongly recommend using the <a href=\"https:\/\/github.com\/galaxykate\/tracery\/blob\/master\/js\/tracery.min.js\" title=\"Minified\">minified library<\/a><\/p>\n<h3><a id=\"user-content-write-grammar-objects-get-generative-stories\" class=\"anchor\" aria-hidden=\"true\" href=\"#write-grammar-objects-get-generative-stories\"><\/a>Write grammar objects, get generative stories<\/h3>\n<h4><a id=\"user-content-an-example-grammar\" class=\"anchor\" aria-hidden=\"true\" href=\"#an-example-grammar\"><\/a>An example grammar<\/h4>\n<pre>{\n\t\"name\": [\"Arjun\",\"Yuuma\",\"Darcy\",\"Mia\",\"Chiaki\",\"Izzi\",\"Azra\",\"Lina\"],\n\t\"animal\": [\"unicorn\",\"raven\",\"sparrow\",\"scorpion\",\"coyote\",\"eagle\",\"owl\",\"lizard\",\"zebra\",\"duck\",\"kitten\"],\n\t\"mood\": [\"vexed\",\"indignant\",\"impassioned\",\"wistful\",\"astute\",\"courteous\"],\n\t\"story\": [\"#hero# traveled with her pet #heroPet#.  #hero# was never #mood#, for the #heroPet# was always too #mood#.\"],\n\t\"origin\": [\"#[hero:#name#][heroPet:#animal#]story#\"]\n}\n<\/pre>\n<h4><a id=\"user-content-output-of-that-grammar\" class=\"anchor\" aria-hidden=\"true\" href=\"#output-of-that-grammar\"><\/a>Output of that grammar.<\/h4>\n<p>Of course, many grammars are more complex!<\/p>\n<pre>Lina traveled with her pet duck. Lina was never indignant, for the duck was always too indignant.\nYuuma traveled with her pet unicorn. Yuuma was never wistful, for the unicorn was always too indignant.\nAzra traveled with her pet coyote. Azra was never wistful, for the coyote was always too impassioned.\nYuuma traveled with her pet owl. Yuuma was never wistful, for the owl was always too courteous.\nAzra traveled with her pet zebra. Azra was never impassioned, for the zebra was always too astute.\n<\/pre>\n<h3><a id=\"user-content-how-to-use-tracery-as-a-broswer-library\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-use-tracery-as-a-broswer-library\"><\/a>How to use Tracery as a broswer library<\/h3>\n<p>Import tracery\n<code>&lt;script defer src=\"js\/libs\/tracery.js\"&gt;&lt;\/script&gt;<\/code><\/p>\n<p>Use the <code>tracery<\/code> object to create a <code>Grammar<\/code> object from a source object (specification below)\n<code>tracery.createGrammar(spellbook);<\/code><\/p>\n<p>The grammar can create <code>Trace<\/code> objects.  A <code>Trace<\/code> is one possible expansion of a grammar.\n<code>var trace = app.grammar.createTrace();<\/code><\/p>\n<p>The trace can be expanded into a tree structure, step by step, or all at once.\n<code>trace.expand();<\/code>\nOnce expanded, the trace can create a 'flattened' version of itself: a single string of text.\nvar myString = trace.flatten();<\/p>\n<p>Or the grammar can generate a trace and flatten it, all in one step\n<code>var myTitle = app.grammar.createFlattened()<\/code><\/p>\n<p>Traces will start their expansions with the 'origin' symbol by default, but you can also create one from a rule (see \"Rule Syntax\" below), or from a symbol\n<code>var trace = app.grammar.createTrace(\"A story about #character#\");<\/code>\n<code>var trace = app.grammar.createTraceFromSymbol(\"bookTitle\");<\/code><\/p>\n<p>Many traces can be working on a single grammar at the same time, without getting in each others way.<\/p>\n<h3><a id=\"user-content-how-to-use-tracery-as-a-nodejs-library\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-use-tracery-as-a-nodejs-library\"><\/a>How to use Tracery as a Node.js library<\/h3>\n<p>Use this Node library created by George Buckenham: <a href=\"https:\/\/github.com\/v21\/tracery\">https:\/\/github.com\/v21\/tracery<\/a><\/p>\n<h2><a id=\"user-content-input\" class=\"anchor\" aria-hidden=\"true\" href=\"#input\"><\/a>Input<\/h2>\n<h3><a id=\"user-content-syntax-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#syntax-overview\"><\/a>Syntax overview<\/h3>\n<h4><a id=\"user-content-grammar\" class=\"anchor\" aria-hidden=\"true\" href=\"#grammar\"><\/a>Grammar<\/h4>\n<p>A grammar is a key-value storage system for rules.<\/p>\n<h4><a id=\"user-content-rule-syntax\" class=\"anchor\" aria-hidden=\"true\" href=\"#rule-syntax\"><\/a>Rule syntax<\/h4>\n<p>Each symbol should be followed by an array of text strings representing rules<\/p>\n<pre>  \"emotion\" : [\"happy\", \"sad\", \"proud\"],\n<\/pre>\n<p>or, if you're writing a long string of single words, you can use 'split'<\/p>\n<pre>  \"emotion\" : \"happy sad reflective morose proud\".split(\" \"),\n<\/pre>\n<p>Rules can also contain expansion symbols, words surrounded by #'s:<\/p>\n<pre>mainCharacter: [\"Brittany the Wombat\"],\nstory : [\"This is a story about #mainCharacter#\"]\n<\/pre>\n<p>Expansion symbols can have modifiers.  Modifiers can change something about the string expansion of that symbol.\n<code>#animal.capitalize#<\/code> or <code>#animal.a#<\/code> or <code>#animal.s#<\/code><\/p>\n<pre>name: [\"Brittany\"],\nanimal: [\"wombat\"],\nstory : [\"This is a story about #name# the #animal.capitalize#\"]\n<\/pre>","created_at":"2018-11-07T16:50:11+0000","updated_at":"2018-11-07T16:50:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/2497301?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":504,"title":"404 (Page Not Found) Error","url":"http:\/\/mattmik.com\/articles\/ascii\/ascii.html","content":"404 (Page Not Found) Error - Ever feel like you're in the wrong place?\n        <div class=\"body-404\"><div class=\"info-container\"><div class=\"inner-border clear-fix\"><div class=\"site-owner-404\"><h3>If you're the <strong>site owner,<\/strong> one of two things happened:<\/h3><ol><li>\n                                                        1) You entered an incorrect URL into your browser's address bar, or\n                                                <\/li>\n                                                <li>\n                                                        2) You haven't uploaded content.\n                                                <\/li>\n                                        <\/ol><\/div><div class=\"site-visitor-404\"><h3>If you're a <strong>visitor<\/strong> and not sure what happened:<\/h3><ol><li>\n                                                        1) You entered or copied the URL incorrectly or\n                                                <\/li>\n                                                <li>\n                                                        2) The link you used to get here is faulty.\n                                                <\/li>\n                                                <li class=\"last\">\n                                                        (It's an excellent idea to let the link owner know.)\n                                                <\/li>\n                                        <\/ol><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:50:11+0000","updated_at":"2018-11-07T16:50:11+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"mattmik.com","http_status":"403"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":499,"title":"Kevin Kelly","url":"https:\/\/kk.org\/","content":"<h3>Year 2018<\/h3><p><strong>I remain the official Senior Maverick for <a href=\"http:\/\/www.wired.com\/category\/magazine\/\" target=\"_blank\">Wired<\/a>, a magazine I helped co-found 25 years ago. I do one article for Wired per year. My most recent published writings are listed <a href=\"http:\/\/kk.org\/articles\" target=\"_blank\">here<\/a>, in chronological order.<\/strong><\/p><p><strong>My newest book, <a href=\"http:\/\/kk.org\/books\/the-inevitable\/\" target=\"_blank\">The Inevitable<\/a>, a New York Times bestseller, is now available in <a href=\"https:\/\/amzn.to\/2GpZvxc\" target=\"_blank\">paperback<\/a>. The book is also a bestseller in China, and I spend a lot of time visiting China to speak about the inevitable and to see what the Chinese have planned for our future.<\/strong><\/p><p>This is the 20th anniversary of my second book, <a href=\"https:\/\/amzn.to\/2Gwit1p\" target=\"_blank\">The New Rules for the New Economy<\/a>. Despite two decades of fast moving technology, I would not change much in the book. So in celebration of its 20 years, I have been <a href=\"https:\/\/twitter.com\/search?q=%23newrules20th&amp;src=typd&amp;lang=en\" target=\"_blank\">tweeting an axiom per day<\/a> from the book. I tweet several times a week about other things under <a href=\"https:\/\/twitter.com\/kevin2kelly\" target=\"_blank\">@kevin2kelly<\/a>. On Facebook, I am the laughing <a href=\"https:\/\/www.facebook.com\/Kevin2Kelly\" target=\"_blank\">Kevin Kelly<\/a>.<\/p><p>About two years ago I started a weekly newsletter, called <a href=\"http:\/\/recomendo.com\/\" target=\"_blank\">Recomendo<\/a>. Together with Mark Frauenfelder, of Boing Boing fame, and Claudia Dawson, our webmaster and my assistant, we give 6 very short recommendations each week. We briefly share in one page the great stuff we are watching, listening to, reading, or great tools we are enjoying, or great destinations we recommend. Recomendo now has 16,000 subscribers.<\/p><p>My collected wisdom about cool tools was rounded up into a huge oversized book called <a href=\"http:\/\/kk.org\/books\/cool-tools-a-catalog-of-possibiliti\/\" target=\"_blank\">Cool Tools<\/a>. After 5 editions, it is now out of print. This year Amazon created a digital version of Cool Tools for the <a href=\"https:\/\/amzn.to\/2H95h3C\" target=\"_blank\">Kindle<\/a>, or the Kindle app on your phone. It is harder to navigate than the paper book, but much more portable and cheaper. I was skeptical Amazon could successfully make a digital book from this huge sprawling tome, but the results of the Kindle version were better than I expected.<\/p><p>I\u2019m trying to complete visiting all the countries in Asia in order to document their disappearing traditions. From these I am compiling a still-untitled photo book about vanishing Asia. To be published a few years from now.<\/p><p><strong>Five years ago I published my first fiction, <a href=\"http:\/\/silver-cord.net\/\" target=\"_blank\">The Silver Cord<\/a>. I worked on it for 11 years! At 464 oversized pages, this huge book was fan-funded, the result of a successful Kickstarter-funded campaign. It is now available on <a href=\"http:\/\/www.amazon.com\/exec\/obidos\/ASIN\/1940689015\/silvercord-20\" target=\"_blank\">Amazon<\/a>. The Silver Cord contains two volumes of an epic story about angels and robots and a girl hero who must save the humans in between. The craftsmanship of the printing, including a 6-page fold out, is stunning. It has to be seen to be believed. <\/strong><\/p><p> <a href=\"https:\/\/kk.org\/archives\/4330#more-4330\" class=\"more-link\">Read More<\/a><\/p>","created_at":"2018-11-07T16:50:09+0000","updated_at":"2018-11-07T16:50:09+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"kk.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":500,"title":"After updating to 16.04, bluetooth audio A2DP mode stopped working","url":"https:\/\/askubuntu.com\/questions\/775523\/after-updating-to-16-04-bluetooth-audio-a2dp-mode-stopped-working","content":"<p>I have a nice bluetooth audio receiver connect to my HiFi stereo, and I used to be able to connect it in 14.04 as a A2DP sink with nice quality.<\/p><p>After upgrading to 16.04, the device connects only as headset, and when trying to switch it in A2DP mode, I have this error in my syslog: <\/p><pre>May 21 11:35:50 RRyS pulseaudio[6159]: [pulseaudio] module-bluez5-device.c: Refused to switch profile to a2dp_sink: Not connected\n<\/pre><p>and the quality of audio is abismal. How can I fix it? <\/p>","created_at":"2018-11-07T16:50:09+0000","updated_at":"2018-11-07T16:50:09+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"askubuntu.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/askubuntu\/img\/apple-touch-icon@2.png?v=c492c9229955","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":496,"title":"Andr\u00e9 Staltz - An off-grid social network","url":"https:\/\/staltz.com\/an-off-grid-social-network.html","content":"<p><em>Scuttlebutt<\/em> is slang for gossip, particularly among sailors. It is also the name of a peer-to-peer system ideal for social graphs, identity and messaging. <a href=\"http:\/\/scuttlebutt.nz\/\">Scuttlebutt<\/a> was created by Dominic Tarr, a Node.js developer with more than <a href=\"https:\/\/www.npmjs.com\/~dominictarr\">600 modules<\/a> published on npm, who lives on a self-steering sailboat in New Zealand.<\/p><p><a href=\"https:\/\/staltz.com\/img\/dominic2.jpg\"><img src=\"https:\/\/staltz.com\/img\/dominic2.jpg\" alt=\"Dominic Tarr\" \/><\/a><\/p><p>Dominic is often offline, but he\u2019s still able to use a social network to communicate with his friends such as James Halliday (a.k.a. <em>substack<\/em>), who is also often offline. James has also <a href=\"https:\/\/www.npmjs.com\/~substack\">authored hundreds<\/a> of npm modules, such as <a href=\"https:\/\/www.npmjs.com\/package\/browserify\">Browserify<\/a>, and is building a shack with his partner Marina on top of 300-year old lava flows in Hawaii.<\/p><p><a href=\"https:\/\/staltz.com\/img\/substack.jpg\"><img src=\"https:\/\/staltz.com\/img\/substack.jpg\" alt=\"James Halliday\" \/><\/a><\/p><p>Dominic and James are a few key figures in a community of eccentric open source hackers gathering in a social network independent from mainstream internet. The unique properties of Secure Scuttlebutt (SSB) make it possible for digital information to spread easily even in the absence of Internet Service Providers (ISP) and the internet\u2019s backbone. What makes that possible is a <em>decentralized protocol<\/em> based on the mechanics of <em>word of mouth<\/em>.<\/p><p>Scuttlebutt is decentralized in a similar way that <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bitcoin\">Bitcoin<\/a> or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bittorrent\">BitTorrent<\/a> are. Unlike centralized systems like PayPal or Dropbox, there is no single website or server to connect when using decentralized services. Which in turn means there is no single company with control over the network.<\/p><p>However, Scuttlebutt differs from Bitcoin and BitTorrent because there are no \u201csingleton components\u201d in the network. When accessing the BitTorrent network, for instance, you need to connect to a Distributed Hash Table (DHT, think of it as a huge round table where anyone can come and take a seat). However, to get access to the DHT in the first place, you need to connect to a bootstrapping server, such as router.bittorrent.com:6881 or router.utorrent.com:6881. These are very lightweight servers which simply introduce you to the DHT. They still depend on the existence of ISPs and the internet backbone. Also, those systems are concerned about public information. For instance, with Bitcoin, each peer stores the entire log of all transactions ever sent by anyone.<\/p><p>Secure Scuttlebutt is also different to federated social networks like <a href=\"https:\/\/github.com\/tootsuite\/mastodon\">Mastodon<\/a>, <a href=\"https:\/\/joindiaspora.com\/\">Diaspora<\/a>, <a href=\"https:\/\/gnu.io\/social\/\">GNU social<\/a>, OStatus. Those technologies are not peer-to-peer, because each component is either a server or a client, but not both. Federated social networks are slightly better than centralized services like Facebook because they provide some degree of choice where your data should be hosted. However, there is still trust and dependency on third-party servers and ISPs, which makes it possible for admistrators of those to abuse their power, through content policies, privacy violations or censorship.<\/p><p><a href=\"https:\/\/staltz.com\/img\/patchwork.jpg\"><img src=\"https:\/\/staltz.com\/img\/patchwork.jpg\" alt=\"Patchwork\" \/><\/a><\/p><p>In Scuttlebutt, the \u201cmesh\u201d suffices. With simply two computers, a local router, and electricity, you can exchange messages between the computers with minimal effort and no technical skills. Each account in Scuttlebutt is a diary (or \u201clog\u201d) of what a person has publicly and digitally said. As those people move around between different WiFi \/ LAN networks, their log gets copy-pasted to different computers, and so digital information spreads.<\/p><p>What word of mouth is for humans, Scuttlebutt is for social news feeds. It is unstoppable and spreads fast. Once the word is out (just an arbitrary example) that Apple is releasing a new iPhone model, there is no way to restrict that information from spreading. A person may tell that piece of information to any of their friends, and those friends may in turn spread that information onwards.<\/p><p>With typical gossip, however, information deteriorates as it spreads and eventually becomes harmful rumor. Scuttlebutt on the other hand makes word of mouth <em>secure<\/em> with cryptography. Each Scuttlebutt account is comprised of simply two things: an append-only diary and private\/public asymmetric crypto keys. An account\u2019s identity is its public key. There are no unique usernames, because you can\u2019t guarantee two people in separate places from choosing the same username, much like you cannot forbid the name \u201cJohn Smith\u201d to be given to a newborn in Canada if it is already taken by another person in Australia.<\/p><p>All information a person has published is registered in their diary.\nPublic messages (like in Twitter) are the most common type of message in a diary, but you\u2019ll also see \u201cI am friends with that person\u201d type of messages.\nTo \u2018send\u2019 a private message to someone, I simply record a message in my diary, but encrypt it first, so the message isn\u2019t plainly readable by anyone who gets their hands on a copy of the diary. Authenticity of diaries is preserved in that all diary entries reference the message that was written before, and then is signed. This prevents tampering and makes replication easier.<\/p><p><a href=\"https:\/\/staltz.com\/img\/ssb-account.svg\"><img src=\"https:\/\/staltz.com\/img\/ssb-account.svg\" alt=\"ssb-account\" \/><\/a><\/p><p>Every time two Scuttlebutt friends connect to the same WiFi, their computers will synchronize the latest messages in their diaries. Another way of synchronizing information is to connect to a common Scuttlebutt server, known as \u201c<a href=\"https:\/\/github.com\/staltz\/easy-ssb-pub\">pub<\/a>\u201d, set up by any member in the community. Pubs make information spread faster, and globally, but are totally dispensable. It\u2019s even feasible to exchange latest news through <a href=\"https:\/\/en.wikipedia.org\/wiki\/Sneakernet\">sneakernet<\/a>, using e.g. USB sticks.<\/p><p>This architecture is built so that network connections accurately represent the social graph and word of mouth. Typically with social networks like Facebook or Twitter, the network connections are centralized with their servers. The network architecture looks completely different to social architecture. Most users don\u2019t care about this because the network architecture is invisible to them. However, it becomes a real problem once an authoritarian government or even the host company itself takes control over the network architecture in ways that disrupt the social architecture. It is not uncommon for a government to shut down a social network in a country for days\/weeks, affecting how people communicate with each other. This has happened in <a href=\"http:\/\/www.reuters.com\/article\/us-facebook-egypt-idUSKCN0WY3JZ\">Egypt<\/a>, <a href=\"http:\/\/www.bbc.co.uk\/programmes\/p04w0b36\">Cameroon<\/a>, and <a href=\"https:\/\/en.wikipedia.org\/wiki\/WhatsApp#Brazilian_court_orders\">Brazil<\/a>.<\/p><p>With Scuttlebutt, the social graph <em>is<\/em> the network architecture, with peer-to-peer infrastructure accurately matching peer-to-peer interactions. It makes communication and the spread of information highly resilient, bringing improvements to freedom of speech with modern information technologies.<\/p><p>This peer-to-peer system has existed for more than two years and brought unique challenges and possibilities. For instance, unique usernames are impossible without a centralized username registry. On the other hand, this questions the need for a login system in the first place: why do you need to \u201center\u201d into the service? Scuttlebutt will not have a user registration flow, because such thing makes no sense in that world.<\/p><p>So far, the network has received a dedicated <a href=\"https:\/\/github.com\/ssbc\/patchwork\">social network desktop app<\/a>, a <a href=\"https:\/\/github.com\/mmckegg\/ferment\">Soundcloud alternative<\/a>, a <a href=\"http:\/\/viewer.scuttlebot.io\/%b6nlgiAu3ZWkLqKnvkU1T\/9PZCfiqSU\/Ujg1xRmD\/64=.sha256\">Viewer webapp<\/a>, and a <a href=\"https:\/\/github.com\/clehner\/git-ssb\">git layer<\/a> (putting \u201cdistributed\u201d back into \u201cdistributed version control\u201d). These work seemlessly together: a person using the git layer to push a commit will record that on their diary, which is visible also in the social network app, for their friends. Currently, the community is using this to \u201ceat their own dog food\u201d, coordinating team work and contributing code all on the same platform, without any intermediate company. GitHub being down will rarely be a problem for them.<\/p><p>The platform is being improved constantly, in areas such as: mobile support, an NPM alternative, WebRTC support for browser peers, and even legal transactions in New Zealand. It has proved to work as a platform setting the requirements and examples for a human-centered social network, as Dominic <a href=\"https:\/\/www.scuttlebutt.nz\/stories\/scuttlebutt-genesis.html\">well described<\/a>:<\/p><blockquote>\n  <p>I wanted an open platform that anyone could build things on. (\u2026) Also, we couldn\u2019t realistically plan to just sit down and create an app that everyone wants to use, we need many experiments so that one can succeed, therefore we need a decentralized application platform more than we need any given a decentralized application.<\/p>\n<\/blockquote><p>To use Scuttlebutt, I recommend reading the <a href=\"http:\/\/scuttlebutt.nz\/\">ssb handbook<\/a>.<\/p><p>\n      If you liked this article, consider sharing\n      <a href=\"https:\/\/twitter.com\/intent\/tweet?original_referer=https%3A%2F%2Fstaltz.com%2Fan-off-grid-social-network.html&amp;text=An%20off-grid%20social%20network&amp;tw_p=tweetbutton&amp;url=https%3A%2F%2Fstaltz.com%2Fan-off-grid-social-network.html&amp;via=andrestaltz\" rel=\"nofollow\" target=\"_blank\" title=\"tweeting\">(tweeting)\n        \n      <\/a> it to your followers.\n    <\/p><p class=\"verify-in-keybase\">\n      Copyright (C) 2017 Andre 'Staltz' Medeiros, licensed under\n      <a href=\"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/\">Creative Commons BY-NC 4.0<\/a>, translations to other languages allowed. You can make sure that the author wrote this\n      post by copy-pasting\n      <a href=\"https:\/\/raw.githubusercontent.com\/staltz\/staltz.com\/master\/signed_posts\/2017-04-06-an-off-grid-social-network.md.asc\">this signature<\/a> into\n      <a href=\"https:\/\/keybase.io\/verify\">this Keybase page<\/a>.\n    <\/p>","created_at":"2018-11-07T16:50:08+0000","updated_at":"2018-11-07T16:50:08+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"staltz.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":497,"title":"Using Two IMAP Accounts in Mutt","url":"https:\/\/pbrisbin.com\/posts\/two_accounts_in_mutt\/","content":"<p>Mutt can be really great with multiple accounts, but it\u2019s not exactly intuitive to setup. Here I\u2019ll document how I access two Gmail accounts together in one mutt instance.<\/p><div class=\"well\"><p>If you haven\u2019t yet seen my previous mutt post, please go read <a href=\"https:\/\/pbrisbin.com\/posts\/mutt_gmail_offlineimap\">that<\/a> now. I recommend using that post to get a single account setup first before coming back here. Even if you plan to jump right into a multi-account setup, this post assumes you\u2019ve at least read the other one and will focus on the differences and required changes to get from there to here.<\/p><\/div><h2 id=\"offlineimap\">Offlineimap<\/h2><p>To get Offlineimap syncing multiple accounts, we simply need to add additional configuration blocks to sync the second account with another local Maildir.<\/p><p><strong>~\/.offlineimaprc<\/strong><\/p><div class=\"sourceCode\" id=\"cb1\"><pre class=\"sourceCode python\"><a class=\"sourceLine\" id=\"cb1-1\" data-line-number=\"1\">[general]<\/a>\n<a class=\"sourceLine\" id=\"cb1-2\" data-line-number=\"2\">ui = ttyui<\/a>\n<a class=\"sourceLine\" id=\"cb1-3\" data-line-number=\"3\">accounts = Personal,Work<\/a>\n<a class=\"sourceLine\" id=\"cb1-5\" data-line-number=\"5\">[Account Personal]<\/a>\n<a class=\"sourceLine\" id=\"cb1-6\" data-line-number=\"6\">localrepository = Personal-Local<\/a>\n<a class=\"sourceLine\" id=\"cb1-7\" data-line-number=\"7\">remoterepository = Personal-Remote<\/a>\n<a class=\"sourceLine\" id=\"cb1-9\" data-line-number=\"9\">[Account Work]<\/a>\n<a class=\"sourceLine\" id=\"cb1-10\" data-line-number=\"10\">localrepository = Work-Local<\/a>\n<a class=\"sourceLine\" id=\"cb1-11\" data-line-number=\"11\">remoterepository = Work-Remote<\/a>\n<a class=\"sourceLine\" id=\"cb1-13\" data-line-number=\"13\">[Repository Personal-Local]<\/a>\n<a class=\"sourceLine\" id=\"cb1-14\" data-line-number=\"14\">type = Maildir<\/a>\n<a class=\"sourceLine\" id=\"cb1-15\" data-line-number=\"15\">localfolders = ~\/Mail\/Personal<\/a>\n<a class=\"sourceLine\" id=\"cb1-17\" data-line-number=\"17\">[Repository Work-Local]<\/a>\n<a class=\"sourceLine\" id=\"cb1-18\" data-line-number=\"18\">type = Maildir<\/a>\n<a class=\"sourceLine\" id=\"cb1-19\" data-line-number=\"19\">localfolders = ~\/Mail\/Work<\/a>\n<a class=\"sourceLine\" id=\"cb1-21\" data-line-number=\"21\">[Repository Personal-Remote]<\/a>\n<a class=\"sourceLine\" id=\"cb1-22\" data-line-number=\"22\">type = Gmail<\/a>\n<a class=\"sourceLine\" id=\"cb1-23\" data-line-number=\"23\">remoteuser = username@gmail.com<\/a>\n<a class=\"sourceLine\" id=\"cb1-24\" data-line-number=\"24\">remotepass = secret<\/a>\n<a class=\"sourceLine\" id=\"cb1-25\" data-line-number=\"25\">realdelete = no<\/a>\n<a class=\"sourceLine\" id=\"cb1-26\" data-line-number=\"26\">sslcacertfile = \/etc\/ssl\/certs\/ca-certificates.crt<\/a>\n<a class=\"sourceLine\" id=\"cb1-28\" data-line-number=\"28\">[Repository Work-Remote]<\/a>\n<a class=\"sourceLine\" id=\"cb1-29\" data-line-number=\"29\">type = Gmail<\/a>\n<a class=\"sourceLine\" id=\"cb1-30\" data-line-number=\"30\">remoteuser = work-username@gmail.com<\/a>\n<a class=\"sourceLine\" id=\"cb1-31\" data-line-number=\"31\">remotepass = secret<\/a>\n<a class=\"sourceLine\" id=\"cb1-32\" data-line-number=\"32\">realdelete = no<\/a>\n<a class=\"sourceLine\" id=\"cb1-33\" data-line-number=\"33\">sslcacertfile = \/etc\/ssl\/certs\/ca-certificates.crt<\/a><\/pre><\/div><div class=\"well\"><p>Obviously, if either of these accounts weren\u2019t a Gmail server, the configuration blocks would be different.<\/p><\/div><p>You can test your setup by running <code>offlineimap -o<\/code> to sync things once. It could take a while, but once done, you should have a nice folder structure like this:<\/p><pre>Mail\/\n|-- Personal\n|   |-- INBOX\n|   `-- ...\n`-- Work\n    |-- INBOX\n    `-- ...<\/pre><h2 id=\"msmtp\">Msmtp<\/h2><p>Msmtp also handles multiple accounts very elegantly, we just add another account block for the second account.<\/p><p><strong>~\/.msmtprc<\/strong><\/p><pre># shared defaults since both are gmail accounts\ndefaults\nhost smtp.gmail.com\nport 587\nprotocol smtp\nauth on\ntls on\ntls_trust_file \/etc\/ssl\/certs\/ca-certificates.crt\naccount personal\nfrom username@gmail.com\nuser username@gmail.com\npassword secret\naccount work\nfrom work-username@gmail.com\nuser work-username@gmail.com\npassword secret\naccount default : personal<\/pre><p>Now we can simply call <code>msmtp -a personal<\/code> or <code>msmtp -a work<\/code> to use whichever account we want. Omitting the <code>-a<\/code> option will use the default account which we\u2019ve set as personal.<\/p><h2 id=\"mutt\">Mutt<\/h2><p>The goal with mutt is to have certain settings change when we enter certain folders. For example, when we\u2019re viewing <code>+Personal\/INBOX<\/code> we want our <code>from<\/code> setting to be our personal From address and the <code>sendmail<\/code> setting should be <code>msmtp -a personal<\/code>. To provide this functionality, we\u2019re going to do the following:<\/p><ol type=\"1\"><li>Place any account-specific settings in separate files<\/li>\n<li>Use mutt\u2019s <code>folder-hook<\/code> facility to source the proper file and set the proper settings upon entering a folder for a given account.<\/li>\n<\/ol><p>Here are the two account-specific files:<\/p><p><strong>~\/.mutt\/accounts\/personal<\/strong><\/p><pre>set from      = \"username@gmail.com\"\nset sendmail  = \"\/usr\/bin\/msmtp -a personal\"\nset mbox      = \"+Personal\/archive\"\nset postponed = \"+Personal\/drafts\"\ncolor status green default\nmacro index D \\\n    \"&lt;save-message&gt;+Personal\/Trash&lt;enter&gt;\" \\\n    \"move message to the trash\"\nmacro index S \\\n    \"&lt;save-message&gt;+Personal\/Spam&lt;enter&gt;\"  \\\n        \"mark message as spam\"<\/pre><p><strong>~\/.mutt\/accounts\/work<\/strong><\/p><pre>set from      = \"work-username@gmail.com\"\nset sendmail  = \"\/usr\/bin\/msmtp -a work\"\nset mbox      = \"+Work\/archive\"\nset postponed = \"+Work\/drafts\"\ncolor status cyan default\nmacro index D \\\n    \"&lt;save-message&gt;+Work\/Trash&lt;enter&gt;\" \\\n    \"move message to the trash\"\nmacro index S \\\n    \"&lt;save-message&gt;+Work\/Spam&lt;enter&gt;\"  \\\n        \"mark message as spam\"<\/pre><div class=\"well\"><p>Notice the <code>color<\/code> line which changes the status bar depending on what account I\u2019m \u201cin\u201d at any given moment.<\/p><\/div><p>The following settings will tell mutt to source one of these files upon entering a folder matching the given pattern, this will setup all the correct settings when entering a folder for a given account:<\/p><p><strong>~\/.muttrc<\/strong><\/p><pre>set spoolfile = \"+Personal\/INBOX\"\nsource ~\/.mutt\/personal\nfolder-hook Personal\/* source ~\/.mutt\/accounts\/personal\nfolder-hook Work\/*     source ~\/.mutt\/accounts\/work<\/pre><p>The first two lines effectively set Personal as the default account when we open mutt.<\/p><p>Well, that should do it. Open up mutt, change folders, send some mails, and make sure everything\u2019s working as you\u2019d expect.<\/p><p>For reference, my complete and current setup can be found with my <a href=\"https:\/\/github.com\/pbrisbin\/dotfiles\/tree\/v1.0\/tag-mail-recipient\">dotfiles<\/a>.<\/p>","created_at":"2018-11-07T16:50:08+0000","updated_at":"2018-11-07T16:50:08+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"pbrisbin.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":498,"title":"Change extension of file using shell script","url":"https:\/\/stackoverflow.com\/questions\/9590935\/change-extension-of-file-using-shell-script\/40670471","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        3&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>How to change extension of all *.dat files in a directory to *.txt. \nShell script should take the directory name as an argument. Can \ntake multiple directories as arguments. Print the log of command \nresult in appending mode with date and timestamp.<\/p><\/div><div class=\"user-action-time\">&#13;\n        asked Mar 6 '12 at 19:59&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1253131\/mainak\">Mainak<\/a>&#13;\n        &#13;<\/div><h2 data-answercount=\"9\">&#13;\n                                9 Answers&#13;\n                                9&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        16&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><a href=\"http:\/\/lab.artlung.com\/unix-batch-file-rename\/\">Batch File Rename By File Extension in Unix<\/a><\/p>\n<pre># change .htm files to .html\nfor file in *.htm ; do mv $file `echo $file | sed 's\/\\(.*\\.\\)htm\/\\1html\/'` ; done\n# change .html files to .htm\nfor file in *.html ; do mv $file `echo $file | sed 's\/\\(.*\\.\\)html\/\\1htm\/'` ; done\n#change .html files to .shtml\nfor file in *.html ; do mv $file `echo $file | sed 's\/\\(.*\\.\\)html\/\\1shtml\/'` ; done\n#change .html files to php\nfor file in *.html ; do mv $file `echo $file | sed 's\/\\(.*\\.\\)html\/\\1php\/'` ; done\n<\/pre>\n<p>so ==&gt;<\/p>\n<pre># change .dat files to .txt\nfor file in *.dat ; do mv $file `echo $file | sed 's\/\\(.*\\.\\)dat \/\\1txt\/'` ; done\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Mar 6 '12 at 20:04&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1235699\/pben\">Pben<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            903411&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        8&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<pre>#!\/bin\/bash\nfor d in $*; do\n    for f in $(ls $d\/*.dat); do\n        echo $(date) $(mv -v $f ${f%.dat}.txt)\n    done\ndone\n<\/pre>\n<p>Output redirection should be done by the shell when running the script<\/p>\n<p>Leaving out argument validity checks<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Mar 6 '12 at 20:30&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1107999\/alex\">Alex<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            6,49363052&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        8&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Bash can do all of the heavy lifting such as extracting the extension and tagging on a new one. For example:<\/p>\n<pre>for file in $1\/*.dat ; do mv \"$file\" \"${file%.*}.txt\" ; done\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/40670471\/revisions\" title=\"show all edits to this post\">edited Dec 5 '16 at 12:47<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Nov 18 '16 at 6:15&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/7176622\/aki-korhonen\">Aki Korhonen<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        2&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Simple script:<\/p>\n<pre>#!\/bin\/bash\nif [ $# -lt 2 ] then\n    echo \"Usage `basename $0` &lt;any number of directories space separated&gt;\"\n    exit 85              # exit status for wrong number of arguments.\nfi\nfor directories\ndo\n    for files in $(ls $directories\/*.dat); do\n        echo $(date) $(mv -v $files ${files%.dat}.txt)\n    done\ndone\n<\/pre>\n<p>The first for loop by default loops on the <code>$@<\/code> i.e. command-line arguments passed.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered May 19 '12 at 10:46&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1135954\/mtk\">mtk<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            7,148105292&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Follow Pben's solution, if your filename contains blank space, you should use double quotation marks to the variable like the following:<\/p>\n<pre>#remove the space in file name\n#example file name:19-014-0100.mp3 .mp3\n#result file name:19-014-0100.mp3\n$ for file in *.mp3 ; \n    do target=`echo \"$file\" | sed 's\/ \/\/g'`;\n    echo \"$target\"; \n    mv \"$file\" \"$target\"; \ndone;\n#remove the duplicate file extension in file name\n#example file name:19-014-0100.mp3.mp3\n#result file name:19-014-0100.mp3\n$ for file in *.mp3 ; \n    do target=`echo \"$file\" | sed 's\/\\.mp3\\.mp3$\/.mp3\/g'`;\n    echo \"$target\"; \n    mv \"$file\" \"$target\"; \ndone;\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Oct 11 '13 at 4:31&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/807707\/amitabha\">Amitabha<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,06321734&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Script, first finds the names of the given extensions.\nIt removes the extension from names. Then adds backslash()\nfor identification of terminal.<\/p>\n<p>Then the 'mv' command executed.\nHere the '.temp' folder is used to hide the process from user,\nin GUI.<\/p>\n<pre>#!\/bin\/sh\nif [ $# -ne 3 ] \nthen\n    echo \"Usage:  .\/script folder current_extension modify_extension\"\n    exit\nfi\nmkdir .temp\nfind $1 -name \"*.$2\" &gt; .temp\/output_1 &amp;&amp; sed \"s\/$2\/\/\" .temp\/output_1 &gt; .temp\/output_2 &amp;&amp; sed -e \"s\/[ \\t]\/\\\\\\ \/g\" .temp\/output_2 &gt; .temp\/output_3\nwhile read line\ndo\n    mv -v \"$line\"\"$2\" \"$line\"\"$3\"\ndone &lt; .temp\/output_3\nrm -rf .temp\n<\/pre>\n<p>The output files are saved inside the '.temp' folder,later the '.temp' folder is removed.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/20715810\/revisions\" title=\"show all edits to this post\">edited Dec 21 '13 at 5:46<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Dec 21 '13 at 5:27&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3124682\/imran-sri-lanka\">Imran_Sri Lanka<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The top voted answer didn't really work for me. I may have been doing something wrong. My scenario was trying to create a file with the original name, but with the date appended to it, along with changing the extension from .xslx to .csv. This is what worked for me:<\/p>\n<pre>csvname=`echo $xlsx |sed 's\/\\.xlsx\/\/'`\"-$now\"`echo $xlsx | sed 's\/\\(.*\\.\\)xlsx\/\\.csv\/'`\n<\/pre>\n<p>So, for all the .dat files in a directory (without the date addition), you could run something like this:<\/p>\n<pre>for i in *.dat\ndo mv $i `echo $i |sed 's\/\\.dat\/\/'``echo $i | sed 's\/\\(.*\\.\\)dat\/\\.txt\/'`\ndone\n<\/pre>\n<p>From the above, this section of code just removed the extension:<\/p>\n<pre>echo $i |sed 's\/\\.dat\/\/'\n<\/pre>\n<p>And this section changes the .dat to .txt:<\/p>\n<pre>echo $i | sed 's\/\\(.*\\.\\)dat\/\\.txt\/'\n<\/pre>\n<p>And by bumping them next to each other, it concatenates the two outputs into the filename. It's like doing this:<\/p>\n<pre>mv [filename][.dat] [filename] + [.txt]\n<\/pre>\n<p>Though, I did use STDOUT instead of the 'mv' command.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 30 '15 at 13:34&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/4304761\/william-cates\">William Cates<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Following command to change file extention <code>.c<\/code> to <code>.h<\/code><\/p>\n<pre>find . -depth -name \"*.c\" -exec sh -c 'dname=$(dirname {}) &amp;&amp; fname=$(basename {} .c) &amp;&amp; mv {} $dname\/$fname.h' \";\"\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 7 '16 at 11:33&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1521335\/vinoth-kumar\">vinoth kumar<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>To rename (changing extention) all my html files on epub files I use this command line :<\/p>\n<pre>find . -name \"*.html*\" -exec rename -v 's\/\\.html$\/\\.epub\/i' {} \\;\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/34963398\/revisions\" title=\"show all edits to this post\">edited Jan 23 '16 at 12:26<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1743880\/tunaki\">Tunaki<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            86.9k21186256&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jan 23 '16 at 12:23&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5830032\/eugene\">eugene<\/a>&#13;\n        &#13;<\/div>","created_at":"2018-11-07T16:50:08+0000","updated_at":"2018-11-07T16:50:08+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":493,"title":"Add power to your terminal with powerline - Fedora Magazine","url":"https:\/\/fedoramagazine.org\/add-power-terminal-powerline\/","content":"<div class=\"post-content\"><p>A while ago, Fedora Magazine <a href=\"https:\/\/fedoramagazine.org\/major-hayden-how-do-you-fedora\/\">posted this interview<\/a> with <a href=\"http:\/\/rackspace.com\">Rackspace<\/a> architect <a href=\"http:\/\/major.io\">Major Hayden<\/a>\u00a0where he mentioned the <a href=\"https:\/\/powerline.readthedocs.org\/en\/latest\/\">powerline<\/a> utility. If you often use a terminal, you too might find <em>powerline<\/em> useful. It gives you helpful status information, and helps you stay organized.<\/p><h2>For the shell<\/h2><p>By default, the shell plugin\u00a0gives you plenty of helpful data:<\/p><ul><li>Login name<\/li>\n<li>Local time<\/li>\n<li>Current working directory or path. The path is condensed automatically when\u00a0it grows\u00a0longer than the\u00a0terminal width.<\/li>\n<li>The number of active background jobs<\/li>\n<li>The hostname, when\u00a0you connect via SSH to a remote system where powerline is installed<\/li>\n<\/ul><p>This saves you a lot of twiddling with your\u00a0shell environment and complex scripting!\u00a0To install the\u00a0utility, open a terminal and run this command:<\/p><pre>sudo dnf install powerline powerline-fonts<\/pre><p>The rest of these instructions assume you\u2019re using Fedora\u2019s\u00a0standard bash shell. If you\u2019re using a different shell, check out\u00a0the <a href=\"http:\/\/powerline.readthedocs.org\/en\/latest\/usage\/shell-prompts.html\">documentation<\/a> for\u00a0tips.<\/p><p>Next, configure your bash shell to use\u00a0<em>powerline<\/em> by default. Add the following snippet to your\u00a0<em>~\/.bashrc<\/em> file:<\/p><pre>if [ -f `which powerline-daemon` ]; then\n  powerline-daemon -q\n  POWERLINE_BASH_CONTINUATION=1\n  POWERLINE_BASH_SELECT=1\n  . \/usr\/share\/powerline\/bash\/powerline.sh\nfi<\/pre><p>To activate the changes, open a new shell or terminal. You should have a terminal that looks like this:<\/p><p><a href=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-28-10-34-25.png\"><img class=\"alignnone wp-image-9820 size-full\" src=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-28-10-34-25.png\" alt=\"Terminal with powerline running in the bash shell\" width=\"577\" height=\"375\" srcset=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-28-10-34-25.png 577w, https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-28-10-34-25-300x195.png 300w\" \/><\/a><\/p><p>Try changing directories. Watch how the \u201cbreadcrumb\u201d prompt changes to show your current location. Very handy! You\u2019ll also be able to see number of\u00a0pending\u00a0background jobs. And if powerline is installed on a remote system, the prompt includes the hostname when you\u00a0connect via SSH.<\/p><h2>For tmux<\/h2><p>If you\u2019re a command line junkie, you probably also know\u00a0<em>tmux<\/em>. It allows you to split your terminal into many windows and panes, each\u00a0containing its own session. But the\u00a0<em>tmux<\/em> standard status line is not quite as interesting as what\u00a0<em>powerline<\/em>\u00a0provides by default:<\/p><ul><li>Window information<\/li>\n<li>System load<\/li>\n<li>Time and date<\/li>\n<li>Hostname, if you\u2019re connected to a remote system via SSH<\/li>\n<\/ul><p>Therefore, let\u2019s install the plugin:<\/p><pre>sudo dnf install tmux-powerline<\/pre><p>Now add this line to your\u00a0<em>~\/.tmux.conf<\/em> file:<\/p><pre>source \"\/usr\/share\/tmux\/powerline.conf\"<\/pre><p>Next, remove or comment out any lines in your tmux configuration for\u00a0status bar length or content. Examples of these settings\u00a0are <em>status-left, status-right, status-left-length, <\/em>and\u00a0<em>status-right-length.<\/em><\/p><p>Your user configuration is stored in <em>~\/.tmux.conf<\/em>. If you don\u2019t have one, copy an example from the web or <em>\/usr\/share\/tmux<\/em>\u00a0to <em>~\/.tmux.conf,<\/em> and then edit.<\/p><p>When you next start\u00a0<em>tmux<\/em>, you should see the\u00a0powerline status bar:<\/p><p><a href=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-11-38-43.png\"><img class=\"alignnone wp-image-9826\" src=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-11-38-43.png\" alt=\"A tmux session with powerline running the status bar\" width=\"578\" height=\"366\" srcset=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-11-38-43.png 857w, https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-11-38-43-300x190.png 300w, https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-11-38-43-600x380.png 600w\" \/><\/a><\/p><h2>For vim<\/h2><p>If you use the <em>vim<\/em> editor, you\u2019re also in luck.\u00a0There\u2019s a powerful\u00a0plugin for <em>vim,<\/em>\u00a0too. By default, it\u00a0shows:<\/p><ul><li>Operating mode (normal, insert, replace)<\/li>\n<li>Current path and\u00a0file name<\/li>\n<li>Text encodings<\/li>\n<li>Document and line positions<\/li>\n<\/ul><p>To install it, use this command:<\/p><pre>sudo dnf install vim-powerline<\/pre><p>Now add the following lines to your <em>~\/.vimrc<\/em> file:<\/p><pre>python3 from powerline.vim import setup as powerline_setup\npython3 powerline_setup()\npython3 del powerline_setup\nset laststatus=2 \" Always display the statusline in all windows\nset showtabline=2 \" Always display the tabline, even if there is only one tab\nset noshowmode \" Hide the default mode text (e.g. -- INSERT -- below the statusline)\nset t_Co=256<\/pre><p>Now you can start\u00a0<em>vim<\/em> and see a spiffy new status line:<\/p><p><a href=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-13-04-55.png\"><img class=\"alignnone wp-image-9829 size-full\" src=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-13-04-55.png\" alt=\"Vim running with powerline status\" width=\"577\" height=\"375\" srcset=\"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-13-04-55.png 577w, https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/Screenshot-from-2015-08-31-13-04-55-300x195.png 300w\" \/><\/a><\/p><h2>Configuring powerline<\/h2><p>No command line utility is\u00a0complete without configuration options. The configuration in this case\u00a0isn\u2019t exactly simple, though; it\u00a0requires you to\u00a0edit <a href=\"http:\/\/json.org\/\">JSON<\/a> formatted files. But there\u2019s a <a href=\"http:\/\/powerline.readthedocs.org\/en\/latest\/configuration.html\">complete configuration guide<\/a> available in the official\u00a0documentation. And since\u00a0the utility is written in Python, it\u2019s eminently hackable.<\/p><p>When you hack the configuration, it\u2019s usually to add, change, or remove\u00a0<em>segments<\/em>. There are plenty of segments available, such as:<\/p><ul><li>Content of environment variables<\/li>\n<li>Version control system data\u00a0(such as git branch and status!)<\/li>\n<li>Weather<\/li>\n<li>\u2026and many more.<\/li>\n<\/ul><p>To change the status layout in an environment, you create or edit configuration files\u00a0in\u00a0your <em>~\/.config\/powerline\/<\/em> folder. These configurations are\u00a0stored as\u00a0<em>themes<\/em> for each plugin. You can use the\u00a0<em>powerline-lint<\/em> utility to check your configuration\u00a0for parsing errors after making\u00a0changes.<\/p><p>Some changes may require you to reload your session or possibly restart the\u00a0daemon:<\/p><pre>powerline-daemon --replace<\/pre><p>Now you can enjoy more sophisticated status\u00a0data\u00a0in your favorite tools!<\/p><\/div><div class=\"post-author\"><img alt=\"\" src=\"https:\/\/secure.gravatar.com\/avatar\/691b5bb2f0df8d2c07ade695df949319?s=96&amp;d=retro&amp;r=g\" srcset=\"https:\/\/secure.gravatar.com\/avatar\/691b5bb2f0df8d2c07ade695df949319?s=192&amp;d=retro&amp;r=g 2x\" class=\"avatar avatar-96 photo\" height=\"96\" width=\"96\" \/><h4 class=\"title\"><a href=\"https:\/\/fedoramagazine.org\/author\/pfrields\/\">Paul W. Frields<\/a><\/h4><p>Paul W. Frields has been a Linux user and enthusiast since 1997, and joined the Fedora Project in 2003, shortly after launch. He was a founding member of the Fedora Project Board, and has worked on documentation, website publishing, advocacy, toolchain development, and maintaining software. He joined Red Hat as Fedora Project Leader from February 2008 to July 2010, and remains with Red Hat as an engineering manager. He currently lives with his wife and two children in Virginia.<\/p><\/div>","created_at":"2018-11-07T16:50:07+0000","updated_at":"2018-11-07T16:50:07+0000","published_at":"2017-07-04T08:00:12+0000","published_by":["Paul W. Frields"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"fedoramagazine.org","preview_picture":"https:\/\/fedoramagazine.org\/wp-content\/uploads\/2015\/08\/powerline.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":494,"title":"Org Mode tools!","url":"http:\/\/orgmode.org\/worg\/org-tools\/","content":"<ul class=\"org-ul\"><li>Author: Matt Lundin<\/li>\n<li>Code: <a href=\"http:\/\/orgmode.org\/worg\/code\/perl\/org2remind.pl\">org2remind<\/a><\/li>\n<\/ul><p>\nRemind is a command line calendar application for Unix\/Linux that\ncan, among other things, spit out formatted plain text calendars and\nagendas.\n<\/p><p>\n<a href=\"http:\/\/www.roaringpenguin.com\/products\/remind\">http:\/\/www.roaringpenguin.com\/products\/remind<\/a>\n<\/p><pre class=\"example\">\nTo convert org-mode appointments to remind data, simply call the\nscript on one or more org files:\nperl org2remind.pl ~\/org\/*.org\n(Optionally, add the script to your path and make it executable.)\nThe results of this script can saved in a file...\nperl org2remind.pl ~\/org\/*.org &gt;&gt; org.rem \n...or piped directly into a remind command...\nperl org2remind.pl ~\/org\/*.org | remind -\nFor a nice calendar of the current month, use the following: \nperl org2remind.pl ~\/org\/*.org | remind -c -\nThis script is similar to org2rem.el (in the contrib directory of the\norg-mode repository), except that, instead of saving files from within\norg-mode, it allows for easier and faster access to org\/remind data\nfrom the command line.\nCurrently, the script supports appointments (i.e., active timestamps)\nand the org-diary-class sexp.\n<\/pre>","created_at":"2018-11-07T16:50:07+0000","updated_at":"2018-11-07T16:50:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":495,"title":"Semantic Versioning 2.0.0","url":"https:\/\/semver.org\/","content":"<h2 id=\"summary\">Summary<\/h2><p>Given a version number MAJOR.MINOR.PATCH, increment the:<\/p><ol><li>MAJOR version when you make incompatible API changes,<\/li>\n  <li>MINOR version when you add functionality in a backwards-compatible\nmanner, and<\/li>\n  <li>PATCH version when you make backwards-compatible bug fixes.<\/li>\n<\/ol><p>Additional labels for pre-release and build metadata are available as extensions to the MAJOR.MINOR.PATCH format.<\/p><p>In the world of software management there exists a dreaded place called\n\u201cdependency hell.\u201d The bigger your system grows and the more packages you\nintegrate into your software, the more likely you are to find yourself, one\nday, in this pit of despair.<\/p><p>In systems with many dependencies, releasing new package versions can quickly\nbecome a nightmare. If the dependency specifications are too tight, you are in\ndanger of version lock (the inability to upgrade a package without having to\nrelease new versions of every dependent package). If dependencies are\nspecified too loosely, you will inevitably be bitten by version promiscuity\n(assuming compatibility with more future versions than is reasonable).\nDependency hell is where you are when version lock and\/or version promiscuity\nprevent you from easily and safely moving your project forward.<\/p><p>As a solution to this problem, I propose a simple set of rules and\nrequirements that dictate how version numbers are assigned and incremented.\nThese rules are based on but not necessarily limited to pre-existing\nwidespread common practices in use in both closed and open-source software.\nFor this system to work, you first need to declare a public API. This may\nconsist of documentation or be enforced by the code itself. Regardless, it is\nimportant that this API be clear and precise. Once you identify your public\nAPI, you communicate changes to it with specific increments to your version\nnumber. Consider a version format of X.Y.Z (Major.Minor.Patch). Bug fixes not\naffecting the API increment the patch version, backwards compatible API\nadditions\/changes increment the minor version, and backwards incompatible API\nchanges increment the major version.<\/p><p>I call this system \u201cSemantic Versioning.\u201d Under this scheme, version numbers\nand the way they change convey meaning about the underlying code and what has\nbeen modified from one version to the next.<\/p><h2 id=\"semantic-versioning-specification-semver\">Semantic Versioning Specification (SemVer)<\/h2><p>The key words \u201cMUST\u201d, \u201cMUST NOT\u201d, \u201cREQUIRED\u201d, \u201cSHALL\u201d, \u201cSHALL NOT\u201d, \u201cSHOULD\u201d,\n\u201cSHOULD NOT\u201d, \u201cRECOMMENDED\u201d, \u201cMAY\u201d, and \u201cOPTIONAL\u201d in this document are to be\ninterpreted as described in <a href=\"http:\/\/tools.ietf.org\/html\/rfc2119\">RFC 2119<\/a>.<\/p><ol><li>\n    <p>Software using Semantic Versioning MUST declare a public API. This API\ncould be declared in the code itself or exist strictly in documentation.\nHowever it is done, it should be precise and comprehensive.<\/p>\n  <\/li>\n  <li>\n    <p>A normal version number MUST take the form X.Y.Z where X, Y, and Z are\nnon-negative integers, and MUST NOT contain leading zeroes. X is the\nmajor version, Y is the minor version, and Z is the patch version.\nEach element MUST increase numerically. For instance: 1.9.0 -&gt; 1.10.0 -&gt; 1.11.0.<\/p>\n  <\/li>\n  <li>\n    <p>Once a versioned package has been released, the contents of that version\nMUST NOT be modified. Any modifications MUST be released as a new version.<\/p>\n  <\/li>\n  <li>\n    <p>Major version zero (0.y.z) is for initial development. Anything may change\nat any time. The public API should not be considered stable.<\/p>\n  <\/li>\n  <li>\n    <p>Version 1.0.0 defines the public API. The way in which the version number\nis incremented after this release is dependent on this public API and how it\nchanges.<\/p>\n  <\/li>\n  <li>\n    <p>Patch version Z (x.y.Z | x &gt; 0) MUST be incremented if only backwards\ncompatible bug fixes are introduced. A bug fix is defined as an internal\nchange that fixes incorrect behavior.<\/p>\n  <\/li>\n  <li>\n    <p>Minor version Y (x.Y.z | x &gt; 0) MUST be incremented if new, backwards\ncompatible functionality is introduced to the public API. It MUST be\nincremented if any public API functionality is marked as deprecated. It MAY be\nincremented if substantial new functionality or improvements are introduced\nwithin the private code. It MAY include patch level changes. Patch version\nMUST be reset to 0 when minor version is incremented.<\/p>\n  <\/li>\n  <li>\n    <p>Major version X (X.y.z | X &gt; 0) MUST be incremented if any backwards\nincompatible changes are introduced to the public API. It MAY include minor\nand patch level changes. Patch and minor version MUST be reset to 0 when major\nversion is incremented.<\/p>\n  <\/li>\n  <li>\n    <p>A pre-release version MAY be denoted by appending a hyphen and a\nseries of dot separated identifiers immediately following the patch\nversion. Identifiers MUST comprise only ASCII alphanumerics and hyphen\n[0-9A-Za-z-]. Identifiers MUST NOT be empty. Numeric identifiers MUST\nNOT include leading zeroes. Pre-release versions have a lower\nprecedence than the associated normal version. A pre-release version\nindicates that the version is unstable and might not satisfy the\nintended compatibility requirements as denoted by its associated\nnormal version. Examples: 1.0.0-alpha, 1.0.0-alpha.1, 1.0.0-0.3.7,\n1.0.0-x.7.z.92.<\/p>\n  <\/li>\n  <li>\n    <p>Build metadata MAY be denoted by appending a plus sign and a series of dot\nseparated identifiers immediately following the patch or pre-release version.\nIdentifiers MUST comprise only ASCII alphanumerics and hyphen [0-9A-Za-z-].\nIdentifiers MUST NOT be empty. Build metadata SHOULD be ignored when determining\nversion precedence. Thus two versions that differ only in the build metadata,\nhave the same precedence. Examples: 1.0.0-alpha+001, 1.0.0+20130313144700,\n1.0.0-beta+exp.sha.5114f85.<\/p>\n  <\/li>\n  <li>\n    <p>Precedence refers to how versions are compared to each other when ordered.\nPrecedence MUST be calculated by separating the version into major, minor, patch\nand pre-release identifiers in that order (Build metadata does not figure\ninto precedence). Precedence is determined by the first difference when\ncomparing each of these identifiers from left to right as follows: Major, minor,\nand patch versions are always compared numerically. Example: 1.0.0 &lt; 2.0.0 &lt;\n2.1.0 &lt; 2.1.1. When major, minor, and patch are equal, a pre-release version has\nlower precedence than a normal version. Example: 1.0.0-alpha &lt; 1.0.0. Precedence\nfor two pre-release versions with the same major, minor, and patch version MUST\nbe determined by comparing each dot separated identifier from left to right\nuntil a difference is found as follows: identifiers consisting of only digits\nare compared numerically and identifiers with letters or hyphens are compared\nlexically in ASCII sort order. Numeric identifiers always have lower precedence\nthan non-numeric identifiers. A larger set of pre-release fields has a higher\nprecedence than a smaller set, if all of the preceding identifiers are equal.\nExample: 1.0.0-alpha &lt; 1.0.0-alpha.1 &lt; 1.0.0-alpha.beta &lt; 1.0.0-beta &lt;\n1.0.0-beta.2 &lt; 1.0.0-beta.11 &lt; 1.0.0-rc.1 &lt; 1.0.0.<\/p>\n  <\/li>\n<\/ol><h2 id=\"why-use-semantic-versioning\">Why Use Semantic Versioning?<\/h2><p>This is not a new or revolutionary idea. In fact, you probably do something\nclose to this already. The problem is that \u201cclose\u201d isn\u2019t good enough. Without\ncompliance to some sort of formal specification, version numbers are\nessentially useless for dependency management. By giving a name and clear\ndefinition to the above ideas, it becomes easy to communicate your intentions\nto the users of your software. Once these intentions are clear, flexible (but\nnot too flexible) dependency specifications can finally be made.<\/p><p>A simple example will demonstrate how Semantic Versioning can make dependency\nhell a thing of the past. Consider a library called \u201cFiretruck.\u201d It requires a\nSemantically Versioned package named \u201cLadder.\u201d At the time that Firetruck is\ncreated, Ladder is at version 3.1.0. Since Firetruck uses some functionality\nthat was first introduced in 3.1.0, you can safely specify the Ladder\ndependency as greater than or equal to 3.1.0 but less than 4.0.0. Now, when\nLadder version 3.1.1 and 3.2.0 become available, you can release them to your\npackage management system and know that they will be compatible with existing\ndependent software.<\/p><p>As a responsible developer you will, of course, want to verify that any\npackage upgrades function as advertised. The real world is a messy place;\nthere\u2019s nothing we can do about that but be vigilant. What you can do is let\nSemantic Versioning provide you with a sane way to release and upgrade\npackages without having to roll new versions of dependent packages, saving you\ntime and hassle.<\/p><p>If all of this sounds desirable, all you need to do to start using Semantic\nVersioning is to declare that you are doing so and then follow the rules. Link\nto this website from your README so others know the rules and can benefit from\nthem.<\/p><h2 id=\"faq\">FAQ<\/h2><h3 id=\"how-should-i-deal-with-revisions-in-the-0yz-initial-development-phase\">How should I deal with revisions in the 0.y.z initial development phase?<\/h3><p>The simplest thing to do is start your initial development release at 0.1.0\nand then increment the minor version for each subsequent release.<\/p><h3 id=\"how-do-i-know-when-to-release-100\">How do I know when to release 1.0.0?<\/h3><p>If your software is being used in production, it should probably already be\n1.0.0. If you have a stable API on which users have come to depend, you should\nbe 1.0.0. If you\u2019re worrying a lot about backwards compatibility, you should\nprobably already be 1.0.0.<\/p><h3 id=\"doesnt-this-discourage-rapid-development-and-fast-iteration\">Doesn\u2019t this discourage rapid development and fast iteration?<\/h3><p>Major version zero is all about rapid development. If you\u2019re changing the API\nevery day you should either still be in version 0.y.z or on a separate\ndevelopment branch working on the next major version.<\/p><h3 id=\"if-even-the-tiniest-backwards-incompatible-changes-to-the-public-api-require-a-major-version-bump-wont-i-end-up-at-version-4200-very-rapidly\">If even the tiniest backwards incompatible changes to the public API require a major version bump, won\u2019t I end up at version 42.0.0 very rapidly?<\/h3><p>This is a question of responsible development and foresight. Incompatible\nchanges should not be introduced lightly to software that has a lot of\ndependent code. The cost that must be incurred to upgrade can be significant.\nHaving to bump major versions to release incompatible changes means you\u2019ll\nthink through the impact of your changes, and evaluate the cost\/benefit ratio\ninvolved.<\/p><h3 id=\"documenting-the-entire-public-api-is-too-much-work\">Documenting the entire public API is too much work!<\/h3><p>It is your responsibility as a professional developer to properly document\nsoftware that is intended for use by others. Managing software complexity is a\nhugely important part of keeping a project efficient, and that\u2019s hard to do if\nnobody knows how to use your software, or what methods are safe to call. In\nthe long run, Semantic Versioning, and the insistence on a well defined public\nAPI can keep everyone and everything running smoothly.<\/p><h3 id=\"what-do-i-do-if-i-accidentally-release-a-backwards-incompatible-change-as-a-minor-version\">What do I do if I accidentally release a backwards incompatible change as a minor version?<\/h3><p>As soon as you realize that you\u2019ve broken the Semantic Versioning spec, fix\nthe problem and release a new minor version that corrects the problem and\nrestores backwards compatibility. Even under this circumstance, it is\nunacceptable to modify versioned releases. If it\u2019s appropriate,\ndocument the offending version and inform your users of the problem so that\nthey are aware of the offending version.<\/p><p>That would be considered compatible since it does not affect the public API.\nSoftware that explicitly depends on the same dependencies as your package\nshould have their own dependency specifications and the author will notice any\nconflicts. Determining whether the change is a patch level or minor level\nmodification depends on whether you updated your dependencies in order to fix\na bug or introduce new functionality. I would usually expect additional code\nfor the latter instance, in which case it\u2019s obviously a minor level increment.<\/p><p>Use your best judgment. If you have a huge audience that will be drastically\nimpacted by changing the behavior back to what the public API intended, then\nit may be best to perform a major version release, even though the fix could\nstrictly be considered a patch release. Remember, Semantic Versioning is all\nabout conveying meaning by how the version number changes. If these changes\nare important to your users, use the version number to inform them.<\/p><p>Deprecating existing functionality is a normal part of software development and\nis often required to make forward progress. When you deprecate part of your\npublic API, you should do two things: (1) update your documentation to let\nusers know about the change, (2) issue a new minor release with the deprecation\nin place. Before you completely remove the functionality in a new major release\nthere should be at least one minor release that contains the deprecation so\nthat users can smoothly transition to the new API.<\/p><h3 id=\"does-semver-have-a-size-limit-on-the-version-string\">Does semver have a size limit on the version string?<\/h3><p>No, but use good judgment. A 255 character version string is probably overkill,\nfor example. Also, specific systems may impose their own limits on the size of\nthe string.<\/p><p>The Semantic Versioning specification is authored by <a href=\"http:\/\/tom.preston-werner.com\">Tom\nPreston-Werner<\/a>, inventor of Gravatars and\ncofounder of GitHub.<\/p><p>If you\u2019d like to leave feedback, please <a href=\"https:\/\/github.com\/mojombo\/semver\/issues\">open an issue on\nGitHub<\/a>.<\/p><h2 id=\"license\">License<\/h2><p><a href=\"http:\/\/creativecommons.org\/licenses\/by\/3.0\/\">Creative Commons - CC BY 3.0<\/a><\/p>","created_at":"2018-11-07T16:50:07+0000","updated_at":"2018-11-07T16:50:07+0000","published_by":["Tom Preston-Werner"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"semver.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":491,"title":"OpenVPN via NetworkManager timeouts; works through openvpn client","url":"https:\/\/forums.opensuse.org\/showthread.php\/showthread.php\/517728-OpenVPN-via-NetworkManager-timeouts-works-through-openvpn-client?s=17da3d0925e8db29e229965358c54720","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:50:06+0000","updated_at":"2018-11-07T16:50:06+0000","annotations":[],"reading_time":0,"domain_name":"forums.opensuse.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":492,"title":"Tmux and Vim - even better together","url":"https:\/\/blog.bugsnag.com\/tmux-and-vim\/","content":"<p>On their own, <a href=\"https:\/\/tmux.github.io\/\">tmux<\/a> and <a href=\"http:\/\/www.vim.org\/\">vim<\/a> are both great tools; together, they\u2019re even better.<\/p><p>Recently, I wrote about a few <a href=\"https:\/\/blog.bugsnag.com\/benefits-of-using-tmux\/\">benefits of using\ntmux<\/a>. The benefits I\ndiscussed there are independent of one\u2019s choice of text editor. But if you\u2019re a\nvim user, the benefits of tmux reach even further. One of my favorite aspects\nof tmux is that it unlocks the potential for a more powerful vim-based\ndevelopment environment, allowing vim and the shell to feel more like a single\ncohesive tool.<\/p><p>In this post we\u2019ll look at a few ways of customizing tmux and vim to help get\nmore done with less typing and context switching. Configuring tmux and vim to\nget the exact behavior you want is often a time consuming endeavor, but in my\nview it\u2019s worth it. Over the last few years of using tmux and vim together,\nI\u2019ve found a few tips and plugins that have really stood out. This post is a\ncollection of those essential tips that have become indispensable parts of my\ndevelopment workflow.<\/p><p>If you\u2019re just getting started with tmux and vim, see the <a href=\"#related-reading\">related\nreading<\/a> section for links to some resources to get you up to\nspeed.<\/p><h3 id=\"moving-around\">Moving Around<\/h3><p>As a general goal, I want to be able to use vim-style movements and text\nediting patterns whether I\u2019m in vim or some other tmux pane.<\/p><p>In vim, we use <a href=\"http:\/\/blogs.sourceallies.com\/2009\/11\/vim-splits-an-introduction\/\">splits<\/a> to divide up the current view,\nallowing us to edit several files side by side, or even to edit multiple\nregions of a single file without having to scroll around. Similarly,\ntmux\u2019s <a href=\"https:\/\/leanpub.com\/the-tao-of-tmux\/read#panes\">panes<\/a> allow us to divide\nup our window so we can run and view several terminal based commands and programs\nat the same time.<\/p><p>I\u2019m constantly using vim splits, so being able to move between splits\nefficiently is crucial. By default, if you want to move from one split to another,\nvim requires that you hit <code class=\"highlighter-rouge\">ctrl-W<\/code> and then one of the directional keys (i.e. h, j, k, l).\nThis isn\u2019t as efficient as it could be, especially for such a common operation.<\/p><p>To address this, like many vim users, I edited my <code class=\"highlighter-rouge\">.vimrc<\/code> to <a href=\"https:\/\/robots.thoughtbot.com\/vim-splits-move-faster-and-more-naturally#easier-split-navigations\">simplify\nsplit\nnavigation<\/a>,\nso that I can jump between vim splits using <code class=\"highlighter-rouge\">ctrl-j<\/code>, <code class=\"highlighter-rouge\">ctrl-k<\/code>, etc.\nThis is already a great efficiency win. With tmux in the picture, we can\nuse <a href=\"https:\/\/github.com\/christoomey\/vim-tmux-navigator\">vim-tmux-navigator<\/a> to\nnot only set up these vim key bindings, but also to set up similar key bindings for\ntmux pane navigation. This allows us to use <code class=\"highlighter-rouge\">ctrl-&lt;direction&gt;<\/code>\nto move anywhere in our tmux window, whether we\u2019re jumping between vim splits or\ntmux panes.<\/p><p>You may find yourself wanting to customize these navigation key bindings in\nprograms other than vim.<\/p><p>For example, recently I\u2019ve been using the command line fuzzy finder\n<a href=\"https:\/\/github.com\/junegunn\/fzf\">fzf<\/a> quite a lot. Out of the box, fzf\nsupports <code class=\"highlighter-rouge\">ctrl-k<\/code> and <code class=\"highlighter-rouge\">ctrl-j<\/code> for moving up and down its list of search\nmatches. Unfortunately, this won\u2019t work if you\u2019re using vim-tmux-navigator\u2019s\n<a href=\"https:\/\/github.com\/christoomey\/vim-tmux-navigator#add-a-snippet\">suggested<\/a>\nkey bindings in your <code class=\"highlighter-rouge\">.tmux.conf<\/code>, which are as follows:<\/p><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre>is_vim=\"ps -o state= -o comm= -t '#{pane_tty}' \\\n    | grep -iqE '^[^TXZ ]+ +(\\\\S+\\\\\/)?g?(view|n?vim?x?)(diff)?$'\"\nbind-key -n C-h if-shell \"$is_vim\" \"send-keys C-h\"  \"select-pane -L\"\nbind-key -n C-j if-shell \"$is_vim\" \"send-keys C-j\"  \"select-pane -D\"\nbind-key -n C-k if-shell \"$is_vim\" \"send-keys C-k\"  \"select-pane -U\"\nbind-key -n C-l if-shell \"$is_vim\" \"send-keys C-l\"  \"select-pane -R\"\nbind-key -n C-\\ if-shell \"$is_vim\" \"send-keys C-\\\\\" \"select-pane -l\"\n<\/pre><\/div><\/div><p>This bit of configuration works by adding conditional logic to the\n<code class=\"highlighter-rouge\">ctrl-&lt;direction&gt;<\/code> key bindings.  When one of these movement commands is used,\nit checks if the current tmux pane is running vim. If so, the appropriate vim\nsplit navigation command is sent. Otherwise, the appropriate tmux pane\nnavigation command is sent.<\/p><p>Using a slightly different approach, we can add logic so that tmux will treat\nfzf like it treats vim, sending fzf its own internal navigation commands rather\nthan tmux\u2019s pane navigation commands.<\/p><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre>is_vim=\"ps -o state= -o comm= -t '#{pane_tty}' \\\n  | grep -iqE '^[^TXZ ]+ +(\\\\S+\\\\\/)?g?(view|n?vim?x?)(diff)?$'\"\nis_fzf=\"ps -o state= -o comm= -t '#{pane_tty}' \\\n  | grep -iqE '^[^TXZ ]+ +(\\\\S+\\\\\/)?fzf$'\"\nbind -n C-h run \"($is_vim &amp;&amp; tmux send-keys C-h) || \\\n                 tmux select-pane -L\"\nbind -n C-j run \"($is_vim &amp;&amp; tmux send-keys C-j)  || \\\n                 ($is_fzf &amp;&amp; tmux send-keys C-j) || \\\n                 tmux select-pane -D\"\nbind -n C-k run \"($is_vim &amp;&amp; tmux send-keys C-k) || \\\n                 ($is_fzf &amp;&amp; tmux send-keys C-k)  || \\\n                 tmux select-pane -U\"\nbind -n C-l run \"($is_vim &amp;&amp; tmux send-keys C-l) || \\\n                 tmux select-pane -R\"\nbind-key -n C-\\ if-shell \"$is_vim\" \"send-keys C-\\\\\" \"select-pane -l\"\n<\/pre><\/div><\/div><p>Using this general pattern in your <code class=\"highlighter-rouge\">.tmux.conf<\/code>, you can further customize\nvim-tmux-navigator\u2019s behavior to work nicely with any command line utilities\nthat use vim-style navigation.<\/p><p>The following demonstrates how you can move around vim splits, tmux panes, and\nfzf results all using <code class=\"highlighter-rouge\">ctrl-&lt;direction&gt;<\/code>.<\/p><p><img src=\"https:\/\/blog.bugsnag.com\/img\/posts\/tmux-vim-style-nav-with-fzf.gif\" alt=\"\" \/><\/p><p>It\u2019s worth noting that you don\u2019t need to do to this just to use\nvim-tmux-navigator and fzf together. In addition to <code class=\"highlighter-rouge\">ctrl-j<\/code> and <code class=\"highlighter-rouge\">ctrl-k<\/code>, fzf\nsupports <code class=\"highlighter-rouge\">ctrl-n<\/code> and <code class=\"highlighter-rouge\">ctrl-p<\/code> for navigating search results. It even has a\n<code class=\"highlighter-rouge\">--bind<\/code> option for setting custom fzf key bindings. But, if you specifically\nwant to maximize the versatility of the <code class=\"highlighter-rouge\">ctrl-j<\/code>, <code class=\"highlighter-rouge\">ctrl-k<\/code>, etc. bindings\nacross several applications running in tmux, this pattern will help.<\/p><h3 id=\"a-more-vim-like-shell\">A more vim-like shell<\/h3><p>Tmux and bash both have support for vi modes, which can help make your shell\nfeel more like your editor. This is great because it means that when you move from\na vim pane to a shell pane, you don\u2019t have to do as big of a context switch.\nMany of the same patterns for moving around and working with text can still apply.<\/p><p>You can use tmux\u2019s <a href=\"https:\/\/sanctum.geek.nz\/arabesque\/vi-mode-in-tmux\/\">vi\nmode<\/a> to make tmux\u2019s copy\nmode feel more like vim. In this mode you can use familiar vim commands to scroll,\nsearch, select, and copy text.<\/p><p>Similarly, you can try bash\u2019s <a href=\"http:\/\/www.catonmat.net\/blog\/bash-vi-editing-mode-cheat-sheet\/\">vi\nediting-mode<\/a>.\nBeing able to use vi style movements and character matching commands to quickly\nedit shell commands is really nice. Suddenly writing shell commands in bash and\nlines of code in vim begin to feel a lot more similar. Also, if you find\nyourself editing a complex shell command and you want to jump into vim for\nreal, just enter normal mode and hit <code class=\"highlighter-rouge\">v<\/code>. This will drop you into vim, where\nyou can finish editing your command, writing this file will bring you back to\nthe shell and execute the command.<\/p><p>It\u2019s common practice for vim users to remap custom keys to escape, so we can <a href=\"http:\/\/vim.wikia.com\/wiki\/Avoid_the_escape_key\">avoid\nreaching<\/a> so far every time we change from\ninsert to normal mode. For example, I\u2019m using the sequence <code class=\"highlighter-rouge\">kj<\/code> to bring me out of insert\nmode. In order for bash\u2019s vi mode to feel comfortable, it\u2019s important to\nbe able to use your familiar mapping to go from insert to normal mode on the command line.<\/p><p>This can be configured in your <code class=\"highlighter-rouge\">~\/.inputrc<\/code> file. In my case, to make <code class=\"highlighter-rouge\">kj<\/code> work,\nI\u2019ve added the following:<\/p><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre>set editing-mode vi\n# vi settings\n$if mode=vi\n  set keymap vi-insert\n  \"kj\": vi-movement-mode # remap escape\n$endif\n<\/pre><\/div><\/div><h3 id=\"simplify-vim-split-and-tmux-pane-creation\">Simplify Vim Split and Tmux Pane creation<\/h3><p>Splits and panes are both foundational concepts and should be really easy\nto use. We\u2019ve already covered standardizing the way we move between splits and panes,\nbut what about creating them in the first place?<\/p><p>I\u2019m not a big fan of tmux\u2019s defaults for a working with panes for a few\nreasons.  First, if you\u2019re already familiar with vim split terminology, the\nwords horizontal and vertical will mean the opposite of what you\u2019d expect when\nit comes to tmux panes. And second, the default commands (<code class=\"highlighter-rouge\">&lt;Prefix&gt;%<\/code> and\n<code class=\"highlighter-rouge\">&lt;Prefix&gt;\"<\/code>) for creating tmux splits never seemed intuitive to me.<\/p><p>I\u2019m using on the following in my <code class=\"highlighter-rouge\">.tmux.conf<\/code>:<\/p><div class=\"language-bash highlighter-rouge\"><div class=\"highlight\"><pre>bind | split-window -h -c \"#{pane_current_path}\"\nbind - split-window -v -c \"#{pane_current_path}\"\n<\/pre><\/div><\/div><p>The beauty of this is that it allows you to continue thinking about splits and\npanes in terms of the orientation of the divider, the same way you\u2019d think\nabout a vim split. With these bindings, <code class=\"highlighter-rouge\">&lt;Prefix&gt;|<\/code> adds a vertical pane\ndivider (like a vertical split in vim), and <code class=\"highlighter-rouge\">&lt;Prefix&gt;-<\/code> adds a horizontal one.\nIf I remember correctly, I first saw these bindings recommended in the book\n<a href=\"https:\/\/pragprog.com\/book\/bhtmux\/tmux\">Tmux: Productive Mouse-Free\nDevelopment<\/a>, which is full of great\ntmux tips and really worth a read.<\/p><p>In vim, I primarily use vertical splits. So I\u2019ve set the following in my <code class=\"highlighter-rouge\">.vimrc<\/code>\nwhich means that from normal mode, creating a new vertical split is as simple as\nhitting <code class=\"highlighter-rouge\">vv<\/code>.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>\" vv to generate new vertical split\nnnoremap &lt;silent&gt; vv &lt;C-w&gt;v\n<\/pre><\/div><\/div><p>Splits and panes are excellent features, so the most important thing is to find\nkey bindings that make them quick and intuitive for you to use.<\/p><h3 id=\"quickly-run-shell-commands-without-leaving-vim\">Quickly run shell commands without leaving Vim<\/h3><p>Using what we\u2019ve already covered, we now have the ability to quickly create\ntmux panes and move between them, running commands where we wish. Easy as this\nis, it\u2019s sometimes preferable to create a pane and run a command in it without\nleaving vim at all.<\/p><p>I use <a href=\"https:\/\/github.com\/benmills\/vimux\">vimux<\/a> and its associated plugins to\nrun terminal commands from inside vim.<\/p><h4 id=\"vimux-essential-commands\">Vimux essential commands<\/h4><p>One of the basic features of vimux is its ability to run an arbitrary shell\ncommand from within vim. Once vimux is installed, you can access the command\nprompt from within vim by running <code class=\"highlighter-rouge\">:VimuxPromptCommand<\/code>. You can then\nimmediately start typing your shell command. Pressing enter will run the\ncommand in a tmux pane in the current window. If necessary, vimux will create a\nnew pane for the command to run in.<\/p><p>Typing <code class=\"highlighter-rouge\">:VimuxPromptCommand<\/code> every time you want to run a command isn\u2019t the best,\nso it\u2019s recommended to add a mapping to your <code class=\"highlighter-rouge\">.vimrc<\/code> like the following:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>\" Prompt for a command to run\nmap &lt;Leader&gt;vp :VimuxPromptCommand&lt;CR&gt;\n<\/pre><\/div><\/div><p>Now, from normal mode you\u2019ll be able to type <code class=\"highlighter-rouge\">&lt;Leader&gt;vp<\/code> to bring up the\nprompt and issue commands, as shown here.<\/p><p><img src=\"https:\/\/blog.bugsnag.com\/img\/posts\/vimux-prompt-demo.gif\" alt=\"\" \/><\/p><p>That\u2019s already useful, especially for running tests, builds, or data processing\nscripts. In many cases, especially for tests, you\u2019ll find yourself wanting to\nrun a shell command, make some code changes, and then run the shell command\nagain.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>\" Run last command executed by VimuxRunCommand\nmap &lt;Leader&gt;vl :VimuxRunLastCommand&lt;CR&gt;\n<\/pre><\/div><\/div><p>Now, from normal mode <code class=\"highlighter-rouge\">&lt;Leader&gt;vl<\/code> will rerun the most recent vimux command. This\nreally shines when it comes to running tests.<\/p><h4 id=\"on-running-tests\">On running tests<\/h4><p>Vimux is especially helpful when you have a failing test that you are trying to\nfix. The workflow looks like this:<\/p><ol><li>Locate the failing test, read it, and run it<\/li>\n  <li>Make changes to the application code to get the test passing<\/li>\n  <li>Rerun the last test (i.e. that last vimux command) using <code class=\"highlighter-rouge\">&lt;Leader&gt;vl<\/code><\/li>\n  <li>Repeat steps 2 and 3 until the test passes<\/li>\n<\/ol><p>Using this workflow, it\u2019s possible to more efficiently get a failing test\npassing, all without ever leaving vim. Quickly rerunning the last test command\nis especially helpful.<\/p><p>When possible, consider using a vimux\n<a href=\"https:\/\/github.com\/benmills\/vimux#platform-specific-plugins\">platform-specific-plugin<\/a>.\nThese provide support for running the currently focused test in a test file,\nrunning all the tests in the current file, etc.<\/p><h4 id=\"quickly-copying-vimux-output\">Quickly copying vimux output<\/h4><p>Let\u2019s say you want to scroll through or copy a bit of output from the vimux\npane. You could use <code class=\"highlighter-rouge\">ctrl-&lt;direction&gt;<\/code> to move from vim to the vimux pane,\nthen you\u2019d enter tmux\u2019s copy mode using <code class=\"highlighter-rouge\">&lt;Prefix&gt;[<\/code>. That\u2019s a bit involved.\nFortunately, vimux offers a shortcut for this workflow.<\/p><p>Add the following to your <code class=\"highlighter-rouge\">.vimrc<\/code>:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>\" Inspect runner pane\nmap &lt;Leader&gt;vi :VimuxInspectRunner&lt;CR&gt;\n<\/pre><\/div><\/div><p>Now <code class=\"highlighter-rouge\">&lt;Leader&gt;vi<\/code> is all you need to jump from vim into the vimux pane, already in copy mode. From here, you can use <a href=\"https:\/\/sanctum.geek.nz\/arabesque\/vi-mode-in-tmux\/\">vi style movements<\/a> as mentioned earlier, to move around, select text, and copy it to the clipboard.<\/p><h4 id=\"getting-a-better-look-at-vimux-output\">Getting a better look at vimux output<\/h4><p>Sometimes, especially when looking at test output on a laptop screen, it can be\na little cumbersome to read long lines in a small pane. There are all sorts of\nways you can efficiently resize your panes with tmux, but in this case, I\u2019ve\nfound vimux\u2019s zoom feature to be a big help. If you haven\u2019t used zoom before,\nit\u2019s a tmux feature that maximizes the active pane to fill the space of the entire tmux\nwindow. When you\u2019re done zooming, you\u2019ll return to the previous pane arrangement\nyou had before. More info on tmux zoom can be found <a href=\"https:\/\/sanctum.geek.nz\/arabesque\/zooming-tmux-panes\/\">here<\/a>.\nI\u2019ve mapped vz to the vimux command which will zoom in on the tmux runner pane.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>\" Zoom the tmux runner pane\n&lt;Leader&gt;vz :VimuxZoomRunner&lt;CR&gt;\n<\/pre><\/div><\/div><p>Now, you can run your tests with one vimux command, and if you need to, zoom in\non the results with <code class=\"highlighter-rouge\">&lt;Leader&gt;vz<\/code>. Typically, at this point in order to leave\nthe zoomed mode, you\u2019d have to use tmux\u2019s <code class=\"highlighter-rouge\">&lt;Prefix&gt;z<\/code>, and that\u2019ll work just\nfine. But, if you\u2019re using vim-tmux-navigator as mentioned above, you can also\nuse the <code class=\"highlighter-rouge\">ctrl-&lt;direction&gt;<\/code> commands to exit the zoom mode as well.<\/p><h3 id=\"going-faster-with-chords\">Going faster with chords<\/h3><p>I use vimux commands a lot. So much so, that using leader-prefixed commands\nlike I\u2019ve described above (e.g., <code class=\"highlighter-rouge\">&lt;Leader&gt;vl<\/code>) began to feel inefficient. ~3\nkeystrokes is too much for an action that has become so central to my vim\nworkflow.<\/p><p>In practice, I\u2019ve set up <a href=\"https:\/\/github.com\/kana\/vim-arpeggio\">vim chords<\/a> for\ncommon use cases like running a single test, running all the tests in a file,\nrerunning the last command etc. If you haven\u2019t given chords a try, I\u2019d highly\nrecommend it. Anytime the leader key starts feeling more like a speed bump than a\nshortcut, I start asking whether I should use a chord. Vim chords can be mapped\nto whatever functionality you like, but personally I use them almost exclusively\nfor vimux bindings.<\/p><p>Here\u2019s a basic example, showing a simple workflow for a fixing a failing test.\nThis combines several of the elements discussed so far:<\/p><ul><li>The test is originally run using a chord mapped to <a href=\"https:\/\/github.com\/pgr0ss\/vimux-ruby-test\">vimux-ruby\u2019s<\/a> <code class=\"highlighter-rouge\">RunRubyFocusedTest<\/code> test method. The chord is effectively a single keystroke.<\/li>\n  <li>The vim split navigation is done using <code class=\"highlighter-rouge\">ctrl-l<\/code><\/li>\n  <li>Subsequent test runs are done using another chord that reruns the last vimux\ncommand.<\/li>\n<\/ul><p><img src=\"https:\/\/blog.bugsnag.com\/img\/posts\/tmux-test-demo.gif\" alt=\"\" \/><\/p><h3 id=\"final-thoughts\">Final thoughts<\/h3><p>Over the last few years, tmux and vim together with these customizations and\npractices, have formed the basis of a really enjoyable development environment\nfor me. This post covers just a few of the many possible ways to configure tmux\nand vim to get even more out of both. The underlying principle here is about\nseeking efficiency, specifically: when practical, we should reduce the\nkeystrokes and the cognitive load required to do common bits of work. Whether\nyou use vim and tmux or not, it\u2019s worth considering the intent behind these\ntips as it applies to your own development tools. I hope you find the process\nas fun as I have.<\/p><hr \/><p>Many thanks to <a href=\"http:\/\/www.moolenaar.net\/\">Bram Moolenaar<\/a>, <a href=\"https:\/\/github.com\/christoomey\">Chris Toomey<\/a>, <a href=\"https:\/\/github.com\/benmills\">Ben Mills<\/a>, and <a href=\"https:\/\/github.com\/junegunn\">Junegunn\nChoi<\/a> for their thoughtful comments on earlier\nversions of this post and more importantly for their invaluable open source\ncontributions.<\/p><h3 id=\"related-reading\">Related Reading<\/h3><ul><li><a href=\"https:\/\/teamgaslight.com\/blog\/vim-plus-tmux-a-perfect-match\">vim + tmux: A Perfect Match<\/a><\/li>\n  <li><a href=\"https:\/\/leanpub.com\/the-tao-of-tmux\">The Tao of Tmux<\/a><\/li>\n  <li><a href=\"https:\/\/pragprog.com\/book\/bhtmux\/tmux\">Tmux: Productive Mouse-Free Development<\/a><\/li>\n  <li><a href=\"https:\/\/pragprog.com\/book\/bhtmux2\/tmux-2\">Tmux 2: Productive Mouse-Free Development<\/a><\/li>\n  <li><a href=\"https:\/\/thoughtbot.com\/upcase\/tmux\">Thoughtbot\u2019s tmux course<\/a><\/li>\n  <li><a href=\"https:\/\/www.quora.com\/What-are-some-good-resources-for-learning-Vim\/answer\/Abhinay-Dronavalli\">What are some good resources for learning vim?<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/keeganlow\/dotfiles\">My dotfiles<\/a><\/li>\n<\/ul><hr \/><p><em>This blog is part 2 in a series on <a href=\"https:\/\/tmux.github.io\/\">tmux<\/a>. Read part 1 on the <a href=\"https:\/\/blog.bugsnag.com\/benefits-of-using-tmux\/\">Benefits of using tmux - lessons from streamlining a development environment<\/a>.<\/em><\/p><p><em>Keegan is a senior software engineer at Bugsnag. Bugsnag automatically monitors your applications for harmful errors and alerts you to them, giving you visibility into the stability of your software. Take a proactive approach to code quality and fix errors impacting your users.<\/em><\/p>","created_at":"2018-11-07T16:50:06+0000","updated_at":"2018-11-07T16:50:06+0000","published_at":"2017-04-19T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":14,"domain_name":"blog.bugsnag.com","preview_picture":"https:\/\/blog.bugsnag.com\/img\/feature\/tmux.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":490,"title":"japaric\/trust","url":"https:\/\/github.com\/japaric\/trust\/blob\/master\/README.md","content":"<blockquote>\n<p>Travis CI and AppVeyor template to test your Rust crate on 5 architectures\nand publish <a href=\"https:\/\/github.com\/japaric\/trust\/releases\">binary releases<\/a> of it for Linux, macOS and Windows<\/p>\n<\/blockquote>\n<h3><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h3>\n<ul><li>\n<p>CI test your crate (library or binary) on Linux, macOS and Windows and on more\nthan just the x86 architecture.<\/p>\n<\/li>\n<li>\n<p>Cargo artifacts are cached and reused between CI builds.<\/p>\n<\/li>\n<li>\n<p>\"Deploys\": Publish binary releases of your application by just pushing a new\n(Git) tag.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<ul><li>\n<p>Your crate must be hosted on GitHub (free).<\/p>\n<\/li>\n<li>\n<p>A <a href=\"https:\/\/travis-ci.org\/\" rel=\"nofollow\">Travis CI<\/a> account (free).<\/p>\n<\/li>\n<li>\n<p>An <a href=\"https:\/\/www.appveyor.com\/\" rel=\"nofollow\">AppVeyor<\/a> account (free).<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-how-to\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to\"><\/a>How-to<\/h2>\n<h3><a id=\"user-content-use-this-template\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-this-template\"><\/a>Use this template<\/h3>\n<p>Copy the <code>ci<\/code> directory, and the <code>.travis.yml<\/code> and <code>appveyor.yml<\/code> files into the\nrepository where you host your Rust crate.<\/p>\n<p>You'll have to adjust those files to meet your needs. Just look inside those\nfiles for comments that start with the word <code>TODO<\/code>; they'll tell you want needs\nto be changed.<\/p>\n<p>This is an overview of what must \/ can be changed:<\/p>\n<ul><li>\n<p>The GitHub token used for deploys.<\/p>\n<\/li>\n<li>\n<p>The list of test targets. Trim it down to reduce test times.<\/p>\n<\/li>\n<li>\n<p>The Rust channel used for testing \/ deploys.<\/p>\n<\/li>\n<li>\n<p>The \"test phase\". Tweak how your crate is tested.<\/p>\n<\/li>\n<li>\n<p>the \"package phase\". Tweak what goes into the release tarball \/ zipfile.<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-generate-binary-releases\" class=\"anchor\" aria-hidden=\"true\" href=\"#generate-binary-releases\"><\/a>Generate binary releases<\/h3>\n<p>You only need to push an <strong>annotated<\/strong> tag to kick off the build process.<\/p>\n<pre># Optional: Publish a new version of your crate to crates.io\n$ cargo publish\n$ git tag -a $TAG\n$ git push origin $TAG\n<\/pre>\n<h3><a id=\"user-content-use-the-binary-releases-on-travis-ci\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-the-binary-releases-on-travis-ci\"><\/a>Use the binary releases on Travis CI<\/h3>\n<p>There's\nan <a href=\"https:\/\/github.com\/japaric\/trust\/blob\/gh-pages\/install.sh\"><code>install.sh<\/code><\/a>\nscript that you can use to quickly install a binary release produced using this\nCI template.<\/p>\n<pre>$ curl -LSfs https:\/\/japaric.github.io\/trust\/install.sh | \\\n    sh -s -- --git japaric\/cross\n<\/pre>\n<p>For more details about this installation script see <code>install.sh -h<\/code><\/p>\n<h3><a id=\"user-content-how-to-disable-deploys\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-disable-deploys\"><\/a>How to disable deploys?<\/h3>\n<p>If you don't want to generate binary releases at all, perhaps because your Cargo\nproject is a library or you only want to test your project, then you can simply\nchange <code>deploy.on.condition<\/code>, in <code>.travis.yml<\/code>, and <code>deploy.on<\/code>, in\n<code>appveyor.yml<\/code>, to always be false. For example:<\/p>\n<div class=\"highlight highlight-source-yaml\"><pre># .travis.yml\ndeploy:\n  on:\n    condition: $DEPLOY = never<\/pre><\/div>\n<h3><a id=\"user-content-how-to-upgrade-your-ci-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-upgrade-your-ci-configuration\"><\/a>How to upgrade your CI configuration?<\/h3>\n<p>First, figure out which version of the trust template you are using. The version\nis written in the header of the <code>.travis.yml<\/code> and <code>appveyor.yml<\/code> files. If\nthere's no header, that means you are using version <code>v0.1.0<\/code>.<\/p>\n<p>Next, look at the <a href=\"https:\/\/github.com\/japaric\/trust\/blob\/master\/CHANGELOG.md\">change log<\/a> to check if there's a new release\nand to learn, at a high level, how the template has changed: what has been\nfixed, what has been added, etc.<\/p>\n<p>If it makes sense for you to upgrade, you can see the required \"code\" changes by\nlooking at the \"diff\" between the version you are using and the version you are\ngoing to upgrade to. For example:<\/p>\n<p><a href=\"https:\/\/github.com\/japaric\/trust\/compare\/v0.1.0...v0.1.1\">https:\/\/github.com\/japaric\/trust\/compare\/v0.1.0...v0.1.1<\/a><\/p>\n<p>As for the upgrade itself, GitHub can generate a patch from the above diff that\nthen you can apply to your repository with <code>git am<\/code> or similar:<\/p>\n<p><a href=\"https:\/\/github.com\/japaric\/trust\/compare\/v0.1.0...v0.1.1.patch\">https:\/\/github.com\/japaric\/trust\/compare\/v0.1.0...v0.1.1.patch<\/a><\/p>\n<h2><a id=\"user-content-supported-targets\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-targets\"><\/a>Supported targets<\/h2>\n<h3><a id=\"user-content-linux\" class=\"anchor\" aria-hidden=\"true\" href=\"#linux\"><\/a>Linux<\/h3>\n<p>Courtesy of <a href=\"https:\/\/github.com\/japaric\/cross\"><code>cross<\/code><\/a>.<\/p>\n<p>If you run into any problem with any of these targets, report them\nto <a href=\"https:\/\/github.com\/japaric\/cross\/issues\"><code>cross<\/code>'s issue tracker<\/a> but first\ncheck if using a newer release (see <code>ci\/install.sh<\/code>)\nwould\n<a href=\"https:\/\/github.com\/japaric\/cross\/blob\/master\/CHANGELOG.md\">fix your problem<\/a>.<\/p>\n<h3><a id=\"user-content-android\" class=\"anchor\" aria-hidden=\"true\" href=\"#android\"><\/a>Android<\/h3>\n<ul><li>\n<p>aarch64-linux-android<\/p>\n<\/li>\n<li>\n<p>arm-linux-androideabi<\/p>\n<\/li>\n<li>\n<p>armv7-linux-androideabi<\/p>\n<\/li>\n<li>\n<p>i686-linux-android<\/p>\n<\/li>\n<li>\n<p>x86_64-linux-android<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-ios\" class=\"anchor\" aria-hidden=\"true\" href=\"#ios\"><\/a>iOS<\/h3>\n<ul><li>\n<p>aarch64-apple-ios<\/p>\n<\/li>\n<li>\n<p>armv7-apple-ios<\/p>\n<\/li>\n<li>\n<p>armv7s-apple-ios<\/p>\n<\/li>\n<li>\n<p>i386-apple-ios<\/p>\n<\/li>\n<li>\n<p>x86_64-apple-ios<\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-linux-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#linux-1\"><\/a>Linux<\/h3>\n<ul><li>\n<p><code>i686-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>i686-unknown-linux-musl<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-unknown-linux-musl<\/code><\/p>\n<\/li>\n<li>\n<p><code>aarch64-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>arm-unknown-linux-gnueabi<\/code><\/p>\n<\/li>\n<li>\n<p><code>armv7-unknown-linux-gnueabihf<\/code><\/p>\n<\/li>\n<li>\n<p><code>mips-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>mips64-unknown-linux-gnuabi64<\/code><\/p>\n<\/li>\n<li>\n<p><code>mips64el-unknown-linux-gnuabi64<\/code><\/p>\n<\/li>\n<li>\n<p><code>mipsel-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>powerpc-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>powerpc64-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>powerpc64le-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>s390x-unknown-linux-gnu<\/code><\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-macos\" class=\"anchor\" aria-hidden=\"true\" href=\"#macos\"><\/a>macOS<\/h3>\n<ul><li>\n<p><code>i686-apple-darwin<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-apple-darwin<\/code><\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-bsd\" class=\"anchor\" aria-hidden=\"true\" href=\"#bsd\"><\/a>*BSD<\/h3>\n<ul><li>\n<p><code>i686-unknown-freebsd<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-unknown-freebsd<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-unknown-netbsd<\/code><\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-windows-mingw\" class=\"anchor\" aria-hidden=\"true\" href=\"#windows-mingw\"><\/a>Windows (MinGW)<\/h3>\n<ul><li>\n<p><code>i686-pc-windows-gnu<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-pc-windows-gnu<\/code><\/p>\n<\/li>\n<\/ul><h3><a id=\"user-content-windows-msvc\" class=\"anchor\" aria-hidden=\"true\" href=\"#windows-msvc\"><\/a>Windows (MSVC)<\/h3>\n<ul><li>\n<p><code>i686-pc-windows-msvc<\/code><\/p>\n<\/li>\n<li>\n<p><code>x86_64-pc-windows-msvc<\/code><\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Licensed under either of<\/p>\n<ul><li>Apache License, Version 2.0 (<a href=\"https:\/\/github.com\/japaric\/trust\/blob\/master\/LICENSE-APACHE\">LICENSE-APACHE<\/a> or\n<a href=\"https:\/\/www.apache.org\/licenses\/LICENSE-2.0\" rel=\"nofollow\">https:\/\/www.apache.org\/licenses\/LICENSE-2.0<\/a>)<\/li>\n<li>MIT license (<a href=\"https:\/\/github.com\/japaric\/trust\/blob\/master\/LICENSE-MIT\">LICENSE-MIT<\/a> or <a href=\"https:\/\/opensource.org\/licenses\/MIT\" rel=\"nofollow\">https:\/\/opensource.org\/licenses\/MIT<\/a>)<\/li>\n<\/ul><p>at your option.<\/p>\n<h3><a id=\"user-content-contribution\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribution\"><\/a>Contribution<\/h3>\n<p>Unless you explicitly state otherwise, any contribution intentionally submitted\nfor inclusion in the work by you, as defined in the Apache-2.0 license, shall be\ndual licensed as above, without any additional terms or conditions.<\/p>","created_at":"2018-11-07T16:50:04+0000","updated_at":"2018-11-07T16:50:04+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/5018213?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":488,"title":"Axiomatic CSS and Lobotomized Owls","url":"https:\/\/alistapart.com\/article\/axiomatic-css-and-lobotomized-owls","content":"<p>At <a href=\"http:\/\/cssday.nl\/2014\">CSS Day<\/a> last June I introduced, with some trepidation, a peculiar three-character CSS selector. Called the \u201clobotomized owl selector\u201d for its resemblance to an owl\u2019s vacant stare, it proved to be the most popular section of <a href=\"http:\/\/vimeo.com\/101718785\">my talk<\/a>. <\/p>\n\n\n\t\t\t\t\n\n\t\t\t\t<div class=\"utility-side-bar\"> \n<div class=\"banner banner-wide pantheon\">\n  <a href=\"http:\/\/cloudinary.rocks\/ala\">\n    <img class=\"banner-logo\" src=\"https:\/\/alistapart.com\/components\/assets\/img\/cloudinary_vertical_logo_for_white_bg.jpg\" alt=\"Cloudinary Logo\" \/><p>Want the best way to store and serve images and videos? Try Cloudinary. Get your own free account.<\/p>\n  <\/a>\n<\/div>\n<div class=\"banner a-book-apart\"> \n  <a href=\"https:\/\/abookapart.com\">\n    <img class=\"banner-logo\" src=\"https:\/\/alistapart.com\/components\/assets\/img\/logo_a-book-apart-color.png\" alt=\"\" \/><p>A Book Apart:<br \/> Brief books for people who make websites.<\/p>\n  <\/a>\n<\/div>\n<div class=\"banner\">  \n  <a href=\"https:\/\/aneventapart.com\/\">\n    <img class=\"banner-logo\" src=\"https:\/\/alistapart.com\/components\/assets\/img\/AEA_2017_80x80.png\" alt=\"\" \/><p>An Event Apart:<br \/> 3 days of design, code, and content for web &amp; UX designers &amp; devs.<\/p>\n  <\/a>\n<\/div><\/div><p>I couldn\u2019t tell you whether the attendees were applauding the thinking behind the invention or were, instead, nervously laughing at my audacity for including such an odd and seemingly useless construct. Perhaps I was unwittingly speaking to a room full of paid-up owl sanctuary supporters. I don\u2019t know.<\/p>\n\n<p>The lobotomized owl selector looks like this:<\/p>\n\n<pre><code>* + *<\/code><\/pre>\n\n<p>Despite its irreverent name and precarious form, the lobotomized owl selector is no mere thought experiment for me. It is the result of ongoing experimentation into automating the layout of flow content. The owl selector is an \u201caxiomatic\u201d selector with a voracious purview. As such, many will be hesitant to use it, and it will terrify some that I include it in production code. I aim to demonstrate how the selector can reduce bloat, speed up development, and help automate the styling of arbitrary, dynamic content.<\/p>\n\n<h2>Styling by prescription<\/h2>\n\n<p>Almost universally, professional web interface designers (engineers, whatever) have accustomed themselves to styling HTML elements <em>prescriptively<\/em>. We conceive of an interface object, then author styles for the object that are inscribed manually in the markup as \u201chooks.\u201d<\/p>\n\n<p>Despite only pertaining to presentation, not semantic interoperability, the class selector is what we reach for most often. While elements and most attributes are predetermined and standardized, classes are the placeholders that gift us with the freedom of authorship. Classes give us control.<br \/><\/p><pre><code>.my-module {\n\t\/* ... *\/\n}<\/code><\/pre>\n\n<p>CSS frameworks are essentially libraries of non-standard class-based ciphers, intended for forming explicit relationships between styles and their elements. They are vaunted for their ability to help designers produce attractive interfaces quickly, and criticized for the inevitable accessibility shortcomings that result from leading with style (form) rather than content (function).<br \/><\/p><pre><code>\n&lt;a class=\"ui-button\"&gt;press me&lt;\/a&gt;<\/code><\/pre>\n\n<p>Whether you use a framework or your own methodology, the prescriptive styling mode also prohibits non-technical content editors. It requires not just knowledge of presentational markup, but also access to that markup to encode the prescribed styles. WYSIWYG editors and tools like Markdown necessarily lack this complexity so that styling does not impede the editorial process.<\/p>\n\n<h2>Bloat<\/h2>\n\n<p>Regardless of whether you can create and maintain presentational markup, the question of whether you <em>should<\/em> remains. Adding presentational ciphers to your previously terse markup necessarily engorges it, but what\u2019s the tradeoff? Does this allow us to reduce bloat in the stylesheet?<\/p>\n\n<p>By choosing to style entirely in terms of named elements, we make the mistake of asserting that HTML elements exist in a vacuum, not subject to inheritance or commonality. By treating the element as \u201cthis thing that needs to be styled,\u201d we are liable to redundantly set some values for the element in hand that should have already been defined higher in the cascade. Adding new modules to a project invites bloat, which is a hard thing to keep in check.<br \/><\/p><pre><code>.module-new {\n\t\/* So\u2026 what\u2019s actually new here? *\/\n}<\/code><\/pre>\n\n<p>From pre-processors with their addition of variables to object-based CSS methodologies and their application of reusable class \u201cobjects,\u201d we are grappling with sandbags to stem this tide of bloat. It is our industry\u2019s obsession. However, few remedies actually eschew the prescriptive philosophy that invites bloat in the first place. Some interpretations of object-oriented CSS even insist on a flattened hierarchy of styles, citing <a href=\"http:\/\/www.impressivewebs.com\/css-specificity-irrelevant\/\">specificity as a problem to be overcome<\/a>\u2014effectively reducing CSS to SS and denying one of its key features.<\/p>\n\n<p>I am not writing to condemn these approaches and technologies outright, but there are other methods that just may be more effective for certain conditions. Hold onto your hats.<\/p>\n\n<h2>Selector performance<\/h2>\n\n<p>I\u2019m happy to concede that when some of you saw the two asterisks in <code>* + *<\/code> at the beginning of this article, you started shaking your head with vigorous disapproval. There is a precedent for that. <a href=\"http:\/\/meyerweb.com\/eric\/articles\/webrev\/200006a.html\">The universal selector<\/a> is indeed a powerful tool. But it can be good powerful, not just bad powerful. Before we get into that, though, I want to address the perceived performance issue.<\/p>\n\n<p>All the studies I\u2019ve read, including <a href=\"http:\/\/www.stevesouders.com\/blog\/2009\/03\/10\/performance-impact-of-css-selectors\/\">Steve Souders\u2019<\/a> and <a href=\"http:\/\/benfrain.com\/css-performance-revisited-selectors-bloat-expensive-styles\/\">Ben Frain\u2019s<\/a>, have concluded that the comparative performance of different CSS selector types is negligible. In fact, Frain concludes that \u201csweating over the selectors used in modern browsers is futile.\u201d I\u2019ve yet to read any compelling evidence to counter these findings.<\/p>\n\n<p>According to Frain, it is, instead, the quantity of CSS selectors\u2014the bloat\u2014that may cause issues; he mentions unused declarations specifically. In other words, embracing class selectors for their \u201cspeed\u201d is of little use when their proliferation is causing the real performance issue. Well, that and the giant JPEGs and un-subsetted web fonts.<\/p>\n\n<p>Contrariwise, the * selector\u2019s simultaneous control of multiple elements increases brevity, helping to reduce file size and improve performance.<\/p>\n\n<p>The <em>real<\/em> trouble with the universal sector is that it alone doesn\u2019t represent a very compelling axiom\u2014nothing more intelligent than \u201cstyle whatever,\u201d anyway. The trick is in harnessing this basic selector and forming more complex expressions that are context-aware.<\/p>\n\n<h2>Dispensing with margins<\/h2>\n\n<p>The trouble with confining styles to objects is that not everything should be considered a property of an object <em>per se<\/em>. Take margins: margins are something that exist between elements. Simply giving an element a top margin makes no sense, no matter how few or how many times you do it. It\u2019s like applying glue to one side of an object before you\u2019ve determined whether you actually want to stick it to something or what that something might be.<br \/><\/p><pre><code>.module-new {\n\tmargin-bottom: 3em; \/* what, all the time? *\/\n}<\/code><\/pre>\n\n<p>What we need is an expression (a selector) that matches elements only in need of margin. That is, only elements in a contextual relationship with other sibling elements. The <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/Adjacent_sibling_selectors\">adjacent sibling combinator<\/a> does just this: using the form x + n, we can add a top margin to any n where x has come before it.<\/p>\n\n<p>This would, as with standard prescriptive styling, become verbose very quickly if we were to create rules for each different element pairing within the interface. Hence, we adopt the aforementioned universal selector, creating our owl face. The axiom is as follows: \u201cAll elements in the flow of the document that proceed other elements must receive a top margin of one line.\u201d<br \/><\/p><pre><code>* + * {\n\tmargin-top: 1.5em;\n}<\/code><\/pre>\n\n<h3>Completeness<\/h3>\n\n<p>Assuming that your paragraphs\u2019 <var>font-size<\/var> is 1 <code>em<\/code> and its <var>line-height<\/var> is 1.5, we just set a default margin of one line between all successive flow elements of all varieties occurring in any order. Neither we developers nor the folks building content for the project have to worry about any elements being forgotten and not adopting at least a standard margin when rendered one after the other. To achieve this the prescriptive way, we\u2019d have to anticipate specific elements and give them individual margin values. Boring, verbose, and liable to be incomplete.<\/p>\n\n<p>Instead of writing styles, we\u2019ve created a style axiom: an overarching principle for the layout of flow content. It\u2019s highly maintainable, too; if you change the <var>line-height<\/var>, just change this singular <var>margin-top<\/var> value to match.<\/p>\n\n<h3>Contextual awareness<\/h3>\n\n<p>It\u2019s better than that, though. By applying margin between elements only, we don\u2019t generate any redundant margin (exposed glue) destined to combine with the padding of parent elements. Compare solution (a), which adds a top margin to all elements, with solution (b), which uses the owl selector.<br \/><\/p><figure><img src=\"https:\/\/alistapart.com\/d\/406\/owl_1.png\" alt=\"Diagram showing elements with margins, with and without the owl selector.\" \/><figcaption>The diagrams in the left column show margin in dark grey and padding in light gray.<\/figcaption><\/figure><p>Now consider how this behaves in regard to nesting. As illustrated, using the owl selector and just a margin-top value, no first or last element of a set will ever present redundant margin. Whenever you create a subset of these elements, by wrapping them in a nested parent, the same rules that apply to the superset will apply to the subset. No margin, regardless of nesting level, will ever meet padding. With a sort of algorithmic elegance, we protect against compound whitespace throughout our interface.<\/p>\n\n<figure><img src=\"https:\/\/alistapart.com\/d\/406\/owl_2.png\" alt=\"Diagram showing nested elements with margins using the owl selector.\" \/><\/figure><p>This is eminently less verbose and more robust than approaching the problem unaxiomatically and removing the leftover glue <em>after the fact<\/em>, as Chris Coyier reluctantly proposed in \u201c<a href=\"http:\/\/css-tricks.com\/spacing-the-bottom-of-modules\/\">Spacing The Bottom of Modules<\/a>\u201d. It was this article, I should point out, that helped give me the idea for the lobotomized owl.<br \/><\/p><pre><code>.module &gt; *:last-child,\n.module &gt; *:last-child &gt; *:last-child,\n.module &gt; *:last-child &gt; *:last-child &gt; *:last-child {\n\tmargin: 0;\n}<\/code><\/pre>\n\n<p>Note that this only works having defined a \u201cmodule\u201d context (a big ask of a content editor), and requires estimating possible nesting levels. Here, it supports up to three.<\/p>\n\n<h2>Exception-driven design<\/h2>\n\n<p>So far, we\u2019ve not named a single element. We\u2019ve simply written a rule. Now we can take advantage of the owl selector\u2019s <a href=\"http:\/\/css-tricks.com\/specifics-on-css-specificity\/\">low specificity<\/a> and start judiciously building in exceptions, taking advantage of the cascade rather than condemning it as other methods do.<\/p>\n\n<h3>Book-like, justified paragraphs<\/h3>\n\n<pre><code>p {\n\ttext-align: justify;\n}\n\np + p {\nmargin-top: 0;\ntext-indent: 2em;\n}<\/code><\/pre>\n\n<p>Note that only successive paragraphs are indented, which is conventional\u2014another win for the adjacent sibling combinator.<\/p>\n\n<h3>Compact modules<\/h3>\n\n<pre><code>.compact * + * {\n\tmargin-top: 0.75em;\n}<\/code><\/pre>\n\n<p>You can employ a little class-based object orientation if you like, to create a reusable style for more compact modules. In this example, all elements that need margin receive a margin of only half a line.<\/p>\n\n<h3>Widgets with positioning<\/h3>\n\n<pre><code>.margins-off &gt; * {\n\tmargin-top: 0;\n}<\/code><\/pre>\n\n<p>The owl selector is an expressive selector and will affect widgets like maps, where everything is positioned exactly. This is a simple off switch. Increasingly, widgets like these will occur as web components where our margin algorithm will not be inherited anyway. This is thanks to <a href=\"http:\/\/www.html5rocks.com\/en\/tutorials\/webcomponents\/shadowdom-201\/\">the style encapsulation feature of Shadow DOM<\/a>.<\/p>\n\n<h2>The beauty of <code>em<\/code>s<\/h2>\n\n<p>Although a few exceptions are inevitable, by harnessing the <code>em<\/code> unit in our margin value, margins already adjust automatically according to another property: <var>font-size<\/var>. In any instances that we adjust <var>font-size<\/var>, the margin will adapt to it: one-line spaces remain one-line spaces. This is especially helpful when setting an increased or reduced body <var>font-size<\/var> via a <code>@media<\/code> query.<\/p>\n\n<p>When it comes to headings, there\u2019s still more good fortune. Having set heading font sizes in your stylesheet in <code>em<\/code>s, appropriate margin (leading whitespace) for each heading has been set without you writing a single line of additional code.<\/p>\n\n<figure><img src=\"https:\/\/alistapart.com\/d\/406\/owl_3.png\" alt=\"Diagram showing automatically adjusted margins based on font-size.\" \/><\/figure><h2>Phrasing elements<\/h2>\n\n<p>This style declaration is intended to be inherited. That is how it, and CSS in general, is designed to work. However, I appreciate that some will be uncomfortable with just how voracious this selector is, especially after they have become accustomed to avoiding inheritance wherever possible.<\/p>\n\n<p>I have already covered the few exceptions you may wish to employ, but, if it helps further, remember that phrasing elements with a typical display value of <code>inline<\/code> will inherit the top margin but be unaffected in terms of layout. Inline elements only respect horizontal margin, which is as specified and standard behavior across all browsers.<\/p>\n\n<figure><img src=\"https:\/\/alistapart.com\/d\/406\/owl_4.png\" alt=\"Diagram showing inline elements with margin.\" \/><\/figure><p>If you find yourself overriding the owl selector frequently, there may be deeper systemic issues with the design. The owl selector deals with flow content, and flow content should make up the majority of your content. I don\u2019t advise depending heavily on positioned content in most interfaces because they break implicit flow relationships. Even grid systems, with their floated columns, should require no more than a simple <code>.row &gt; *<\/code> selector applying <code>margin-top: 0<\/code> to reset them.<\/p>\n\n<figure><img src=\"https:\/\/alistapart.com\/d\/406\/owl_5.png\" alt=\"Diagram showing floated columns with margins.\" \/><\/figure><h2>Conclusion<\/h2>\n\n<p>I am a very poor mathematician, but I have a great fondness for <a href=\"http:\/\/en.wikipedia.org\/wiki\/Euclidean_geometry#Axioms\">Euclid\u2019s postulates<\/a>: a set of irreducible rules, or axioms, that form the basis for complex and beautiful geometries. Thanks to Euclid, I understand that even the most complex systems must depend on foundational rules, and CSS is no different. Although modularization of a complex interface is a necessary step in its maturation, any interface that does not follow basic governing tenets is going to lack clarity.<\/p>\n\n<p>The owl selector allows you to control flow content, but it is also a way of relinquishing control. By styling elements according to context and circumstance, we accept that the structure of content is\u2014and should be\u2014mutable. Instead of prescribing the appearance of individual items, we build systems to anticipate them. Instead of prescribing the appearance of the interface as a whole, we let the content determine it. We give control back to the people who would make it.<\/p>\n\n<p>When turning off CSS for a webpage altogether, you should notice two things. First, the page is unfalteringly flexible: the content fits the viewport regardless of its dimensions. Second\u2014provided you have written standard, accessible markup\u2014you should see that the content is already styled in a way that is, if not highly attractive, then reasonably traversable. The browser\u2019s user agent styles take care of that, too. <\/p>\n\n<p>Our endeavors to reclaim and enhance the innate device independence offered by user agents are ongoing. It\u2019s time we worked on reinstating content independence as well.<\/p>\n\n\t\t\t\t\n\n\t\t\t\t\n\n\t\t\t\t\n\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t<footer class=\"entry-footer nocontent\"><h1>About the Author<\/h1>\n\n\t\t\t\t<div class=\"byline author vcard\">\n\n\t\t\t\t\t<a class=\"photo\" href=\"https:\/\/alistapart.com\/author\/heydonworks\"><img src=\"https:\/\/alistapart.com\/d\/_made\/pix\/authors\/photos\/photo_131305_120_120_c1.jpg\" alt=\"\" \/><\/a>\n\n\t\t\t\t\t<h2>\n\n\t\t\t\t\tHeydon Pickering is a designer and interface developer from Norwich in the UK. He is lead designer at <a href=\"https:\/\/www.neontribe.co.uk\/\">Neontribe<\/a> and the accessibility editor for Smashing Magazine, where his book <a href=\"https:\/\/shop.smashingmagazine.com\/apps-for-all-coding-accessible-web-applications.html\">Apps For All<\/a> is also available.<\/h2><\/div>\n\n\t\t\t\t\t\n\t\t\t\t\t\t<h3>More from this Author<\/h3>\n\n\t\t\t\t\t\t\t<ul class=\"author-more\"><li><a href=\"https:\/\/alistapart.com\/article\/quantity-queries-for-css\">Quantity Queries for CSS<\/a><\/li>\n\t\t\t\t\t\t<\/ul>\n\n\t\t\t\t \n\n\t\t\t\t<\/footer>","created_at":"2018-11-07T16:50:03+0000","updated_at":"2018-11-07T16:50:03+0000","published_at":"2014-10-21T00:00:00+0000","published_by":["Heydon Pickering"],"annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"alistapart.com","preview_picture":"https:\/\/alistapart.com\/d\/ALA_406_Axiomatic-Owls.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":489,"title":"Scalable and secure access with SSH - Facebook Code","url":"https:\/\/code.fb.com\/production-engineering\/scalable-and-secure-access-with-ssh\/","content":"<p>Consistent security controls and high reliability are common expectations for any systems administrator. How do you deliver both on a network with thousands of servers supporting thousands of engineers? Most off-the-shelf solutions require a compromise in at least one of these areas \u2014 and we refused to accept this. <\/p>\n<p>Most systems administrators use the industry-standard Secure Shell (SSH) for accessing systems, and yet many of its special features are not widely leveraged. At Facebook, we take advantage of those features to use SSH in a way that is reliable, secure, and manageable. SSH, more specifically <a href=\"http:\/\/www.openssh.com\/\">OpenSSH<\/a>, has a great way to provide both the security and reliability we require: signed certificates with principals.<\/p>\n<h2>Signed certificates<\/h2>\n<p>When systems administrators install a Linux server, it's common to create a couple of accounts with passwords; eventually, a few users are given sudo access to escalate privileges. Local account management works well with a few servers, but as a company grows, central authentication, like <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lightweight_Directory_Access_Protocol\">LDAP<\/a> and\/or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Kerberos_(protocol)\">Kerberos<\/a>, is often used to avoid manually managing accounts on every server. With continued growth, systems administrators may come to realize that central authentication is a single, and potentially devastating, point of failure. If it goes down, they will lose access to everything unless they allow root logins to gain access to systems. Getting locked out of your own system is one of the worst things that can happen during an incident. For example, if a service outage brought down your authentication infrastructure, you would not be able to log in and fix it.<\/p>\n<p>In addition to the failure risks of central management, opting for public key authentication over passwords for your users means having to manage their public keys across all your servers. If the administrative headache weren't enough, it's not uncommon to compromise your own security by having unknown public keys in the <code>authorized_keys<\/code> files. Furthermore, <code>authorized_keys<\/code> requires defining trust by individual key pair, which does not scale.<\/p>\n<p>We recognized that authentication with signed certificates provides a single point of trust with no dependency on any third-party infrastructure. We configured our SSH servers to trust our certificate authority (CA) and everything it signs. We borrowed this from a simple and widely used concept that others use across the internet for HTTPS traffic, and we adopted it in the SSH world.<\/p>\n<p>When designing a security system, regardless of purpose or protocol, you need to think of authentication and authorization separately. Authentication securely verifies that the client is who it claims to be but does not grant any permissions. After the successful authentication, authorization decides whether or not the client can perform a specific action. In addition, accounting tracks the entire process, in case you need to investigate later.<\/p>\n<p>When you have a large set of users, you need a central directory to manage them. To avoid a single point of failure, our production systems only have local accounts in <code>\/etc\/passwd<\/code>. The most common account used is <code>root<\/code>. The SSH servers only accept <code>root<\/code> logins if the client certificate has some very specific capabilities given by the CA, and the CA itself determines who can use SSH and where. Using local accounts introduces a problem, though. The standard Unix login accounting (<code>utmp<\/code>, <code>wtmp<\/code>) is of no use here, and the <code>last<\/code> command can only show local accounts that have logged in. However, OpenSSH can give you detailed information on which certificate was used to authenticate, providing enough information for rich accountability.<\/p>\n<p>We also have a centralized syslog infrastructure, collecting logs in real time from the entire fleet. We collect a large quantity of logs and the infrastructure required to support it can quickly become complex. On the aggregators, we have parsers interpreting and transforming logs into tabular data that send those logs to long-term retention databases, like <a href=\"https:\/\/hive.apache.org\/\">Hive<\/a>. This creates a central place for login accounting. Even though the local account we use for login is <code>root<\/code>, for any other local account, the accounting system enables us to easily identify who logged in where and perform statistical analysis on this data.<\/p>\n<p>Logging in as <code>root<\/code> may sound a bit counterintuitive at first. However, with a trusted accounting infrastructure, it actually improves security and reliability because you can better enforce accountability for users' actions. You need to have a robust system that enforces access based on established business rules. We do this by defining security domains according to those rules. <\/p>\n<h2>Security domains<\/h2>\n<p>In typical environments, engineers have direct SSH access to production once they are in some trusted network. In our infrastructure, engineers don't have direct SSH access to production systems, enforced by network firewalls. Instead, we use a bastion host to reach production. Engineers can authenticate to those bastion hosts only from trusted networks. These hosts use centralized LDAP and Kerberos installations to share account information, and they require two-factor authentication to protect against password leakage. Once users are properly authenticated and authorized, a background process contacts our internal CA to request a signed SSH certificate. That certificate contains all principals allowed for that specific engineer. From the bastion host, having obtained the appropriate certificate, engineers can SSH into production machines as either <code>root<\/code> or other lower privileged users. This way, we make sure that no engineer has more access than he or she requires.<\/p>\n<p>We do not place SSH certificates on laptops because it is difficult to control everything that individual employees run on them. Even though they are centrally managed, laptops are more prone to vulnerabilities than the bastion servers. Therefore, we do not trust them with SSH private keys.<\/p>\n<h2>Implementation<\/h2>\n<p>We covered authorization and authentication philosophies. Now let's move on to implementation. What I described above may sound daunting, but it's not that complicated. <\/p>\n<p>Let's walk through a few examples on how to set up a rudimentary certificate authority to sign certificates and hypothetical production servers to trust it, accepting only certain principals.<\/p>\n<p>First, create your own CA, which is essentially just a normal key pair:<\/p>\n<pre>&#13;\n<code>$ umask 77  # you really want to protect this :-)&#13;\n$ mkdir ~\/my-ca &amp;&amp; cd ~\/my-ca&#13;\n&#13;\n$ ssh-keygen -C CA -f ca&#13;\n<\/code>&#13;\n<\/pre>\n<p>You can decide if you want to have a pass phrase to protect the private key.<\/p>\n<p>At this point, you have two files, <code>ca<\/code> (the private key) and <code>ca.pub<\/code> (the public key). You need to distribute <code>ca.pub<\/code> to your entire fleet. Remember, this is meant to be public, so complete access lockdown isn't the goal. For this example, let's place it in <code>\/etc\/ssh\/ca.pub<\/code>. You can change the permissions of <code>ca.pub<\/code> to 0644.<\/p>\n<p>Now configure your SSH servers to trust it with this single line change in <code>\/etc\/ssh\/sshd_config<\/code>:<\/p>\n<pre><code class=\"language-clike\">TrustedUserCAKeys \/etc\/ssh\/ca.pub<\/code><\/pre>\n<p>Now that you have a chain of trust, you can start generating certificates. Ideally, your CA should be a very secure server that only the security team can access. In small places, it may be OK to have a human being signing the certificate requests. For larger deployments, a system that does all that automatically and programmatically is preferred. As you can imagine, we have the latter at Facebook. Under the hood, our complex CA infrastructure simply receives a public key, runs all security checks, invokes <code>ssh-keygen<\/code> to sign it with the CA's private key, and returns the signed certificate back to the client.<\/p>\n<p>One very important security practice is that private keys should never leave the systems where they've been generated, no matter how secure the transport is. <\/p>\n<p>In a user terminal, not on the CA server, let's generate a key for them: <\/p>\n<pre>&#13;\n<code>&#13;\n$ ssh-keygen -t ecdsa  # or -t rsa, up to you&#13;\nGenerating public\/private ecdsa key pair.&#13;\nEnter file in which to save the key (\/home\/mfdutra\/.ssh\/id_ecdsa):&#13;\nEnter passphrase (empty for no passphrase):***&#13;\nEnter same passphrase again:***&#13;\nYour identification has been saved in \/home\/mfdutra\/.ssh\/id_ecdsa.&#13;\nYour public key has been saved in \/home\/mfdutra\/.ssh\/id_ecdsa.pub.&#13;\n...&#13;\n<\/code>&#13;\n<\/pre>\n<p>In your <code>.ssh\/<\/code> directory, you'll see <code>id_ecdsa<\/code> and <code>id_ecdsa.pub<\/code>. Copy the latter to the CA server and get it signed. Because this is public information, the transport isn't important. You can copy and paste, or fax it; just don't copy <code>id_ecdsa<\/code> anywhere.<\/p>\n<p>On the CA server:<\/p>\n<pre>$ ssh-keygen -s ca -I mfdutra -n root -V +1w -z 1 id_ecdsa.pub<\/pre>\n<p>The <code>ssh-keygen<\/code> man page has a great explanation for each argument used. Basically, we're signing <code>id_ecdsa.pub<\/code> with <code>ca<\/code>. The certificate ID will be <code>mfdutra<\/code> and the only principal it has will be <code>root<\/code>. It's valid for one week and has the serial number <code>1<\/code>. You should have <code>id_ecdsa-cert.pub<\/code> now. Copy this back to the user terminal and place it under <code>.ssh\/<\/code>. Remember, this is public information, and this certificate doesn't work without its respective private key.<\/p>\n<p>Since we haven't configured the servers to require a specific set of principals, the default <code>sshd<\/code> configuration will allow this certificate to log in as any user in its principal list. Since I used <code>-n root<\/code> to create the certificate, I can log in as <code>root<\/code> in any system that trusts the CA. If you don't have a special authorization schema in your organization, this is probably enough. However, you may want to have people logging in as another local user and then you can use <code>sudo<\/code> to escalate by putting any other user in the <code>-n<\/code> argument, which takes a comma-separated list of principals.<\/p>\n<p>You can inspect a certificate with this command:<\/p>\n<pre>&#13;\n<code class=\"language-clike\">&#13;\n$ ssh-keygen -Lf id_ecdsa-cert.pub<br \/>id_ecdsa-cert.pub:&#13;\n  Type: ecdsa-sha2-nistp256-cert-v01@openssh.com user certificate&#13;\n  Public key: ECDSA-CERT ...&#13;\n  Signing CA: ECDSA ...&#13;\n  Key ID: \"mfdutra\"&#13;\n  Serial: 1&#13;\n  Valid: from 2016-01-13T15:26:00 to 2016-01-20T15:27:00&#13;\n  Principals:&#13;\n    root&#13;\n  Critical Options: (none)&#13;\n  Extensions:&#13;\n    permit-X11-forwarding&#13;\n    permit-agent-forwarding&#13;\n    permit-port-forwarding&#13;\n    permit-pty&#13;\n    permit-user-rc&#13;\n<\/code>&#13;\n<\/pre>\n<p>Now that you have the certificate along with the private key, you can SSH anywhere as <code>root<\/code>.<\/p>\n<pre>$ ssh root@any-system-that-trusts-my-ca<\/pre>\n<p>Once you're logged in there, take a look at the authentication log. You'll see something like:<\/p>\n<pre>Accepted publickey for root from 1.2.3.4 port 123 ssh2: ECDSA-CERT ID mfdutra (serial 1) CA ECDSA fingerprint...<\/pre>\n<p>You can see that even when logging in directly as <code>root<\/code>, the server could still identify the certificate used to authenticate \u2014 in this case with ID <code>mfdutra<\/code>. This means using a correct <code>-I<\/code> with <code>ssh-keygen<\/code> is very important because it will identify the certificate. Using a unique serial number is also recommended, so you can identify each individual certificate issued. In fact, if you want to use revocation lists, unique serial numbers are a requirement.<\/p>\n<p>Next, we'll define different security zones. For this example, let's create three zones: <code>zone-webservers<\/code>, <code>zone-databases<\/code> and <code>root-everywhere<\/code>.<\/p>\n<p>Now, we'll configure one of our servers to accept only certain principals. Add this line to <code>\/etc\/ssh\/sshd_config<\/code>:<\/p>\n<pre>AuthorizedPrincipalsFile \/etc\/ssh\/auth_principals\/%u<\/pre>\n<p>Populate the principals file:<\/p>\n<pre>$ mkdir \/etc\/ssh\/auth_principals<br \/>$ echo -e 'zone-webservers\\nroot-everywhere' &gt; \/etc\/ssh\/auth_principals\/root<\/pre>\n<p>You can control access to any local user by creating those files under <code>\/etc\/ssh\/auth_principals<\/code>.<\/p>\n<p>Reload the SSH server and try to SSH into it using the certificate you generated before. You'll be denied access, and the following line will show up in the authentication log on the server:<\/p>\n<pre>error: Certificate does not contain an authorized principal<\/pre>\n<p>Why create a zone called <code>root-everywhere<\/code>? Because it's convenient when certain users need to be able to log in everywhere. It's a better option than putting all possible principals in the user certificate, which is difficult to manage long-term. <\/p>\n<p>You can now go back to the CA server and generate a new certificate for this user, with a different principal: <\/p>\n<pre>$ ssh-keygen -s ca -I mfdutra -n zone-webservers -V +1w -z 2 id_ecdsa.pub<\/pre>\n<p>With this new certificate back in the user terminal, you can SSH into the system successfully, because there will be at least one intersection between the list of principals you have in your certificate and the list of principals the server accepts.<\/p>\n<p>Get another server and configure it in the zone <code>zone-databases<\/code>. Play with the certificates to see how it goes. Your imagination is the limit!<\/p>\n<h2>Next steps<\/h2>\n<p>The above covers the approach Facebook has taken to control authorization on production systems. We have invested time in building secure zones and a sophisticated access control management system. Various organizations will need different designs and your design should reflect your organization\u2019s needs; there is no one-size-fits-all model.<\/p>\n<p>Want to learn more? Check out the <code>ssh-keygen<\/code><a href=\"http:\/\/linux.die.net\/man\/1\/ssh-keygen\"> man page<\/a>, especially the section about special options (-O). You can do some pretty cool stuff with certificates, such as having one that only allows running a specific command, or attaching a certificate to a specific host.<\/p>\n<p>You can also use certificates related to host keys. With that, the SSH clients can automatically trust all hosts with a host certificate signed by the CA, eliminating the need to manually accept every new host you SSH into.<\/p>\n<p>A few parting words of advice: When you build your CA, be it a small script or a complex system, make sure you keep track of all certificates you issue. If you find yourself in the unfortunate situation of having a compromised certificate (and its respective private keys) and you don't know how to revoke them, your last resort is to rotate the entire CA. If you end up having a programmatic CA, consider having short-lived certificates, e.g., 24 hours. This shortens the window of opportunity for an attack if you experience a compromise. <\/p>\n<p>Above all, protect your CA private key and consider rotating it regularly. <\/p>\n\n<div id=\"jp-relatedposts\" class=\"jp-relatedposts\">\n\t<h3 class=\"jp-relatedposts-headline\"><em>Related<\/em><\/h3>\n<\/div>","created_at":"2018-11-07T16:50:03+0000","updated_at":"2018-11-07T16:50:03+0000","published_at":"2016-09-12T18:15:25+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":11,"domain_name":"code.fb.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":487,"title":"Things I Learned Managing Site Reliability for Some of the World\u2019s Busiest Gambling Sites","url":"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/","content":"<h2><b>tl;dr<\/b><\/h2>\n<p>For several years I managed the 3rd line site reliability operation for many of the world\u2019s busiest gambling sites, working for a little-known company that built and ran the core backend online software for several businesses that each at peak could take tens of millions of pounds in revenue per hour. I left a couple of years ago, so it\u2019s a good time to reflect on what I learned in the process.<\/p>\n<p>In many ways, what we did was similar to what\u2019s now called an SRE function (I\u2019m going to call us SREs, but the acronym didn\u2019t exist at the time). We were on call, had to respond to incidents, made recommendations for re-engineering, provided robust feedback to developers and customer teams, managed escalations and emergency situations, ran monitoring systems, and so on.<\/p>\n<p>The team I joined was around 5 engineers (all former developers and technical leaders), which grew to around 50 of more mixed experience across multiple locations by the time I left.<\/p>\n<p>I\u2019m going to focus here on process and documentation, since I don\u2019t think they\u2019re talked about usefully enough where I do read about them.<\/p>\n<p>If you want to read something far longer\u00a0<a href=\"https:\/\/landing.google.com\/sre\/book.html\">Google\u2019s SRE book <\/a>is a great resource<strong>.<\/strong><\/p>\n<h2><b>Process<\/b><\/h2>\n<p>Process is essential to running and scaling an SRE operation. It\u2019s the core of everything we achieved. When I joined the team, habits were bad \u2013 there was a ticketing system, but one-journal resolutions were not uncommon (\u2018Site down. Fixed, closing.\u2019).<\/p>\n<p>An SRE operation is basically a factory processing information and should act\u00a0accordingly. You wouldn\u2019t have a factory running without processes to take care of the movement of goods, and by the same token you shouldn\u2019t have a knowledge-intensive SRE operation running without processes to take care of the movement of knowledge.<\/p>\n<p>One frequent objection to process I heard is that it \u2018stifles creativity\u2019. In fact, effective process (bad process implemented poorly can mess anything up!) clears your mind to allow creative thought.<\/p>\n<p>A great book on this subject is <a href=\"http:\/\/atulgawande.com\/book\/the-checklist-manifesto\/\">\u2018The Checklist Manifesto\u2019<\/a>, which inspired many of the changes we made, and was widely read within the team. It cites the examples of the aviation industry\u2019s approach to process, which enables remarkable creativity under stressful conditions by mental automation of routine operations. There\u2019s even a <a href=\"http:\/\/www.imdb.com\/title\/tt3263904\/\">film<\/a> about one incident discussed\u00a0and the pilot himself <a href=\"http:\/\/www.airspacemag.com\/as-interview\/aamps-interview-sullys-tale-53584029\/\">cited checklists and routine as an enabler<\/a> of his fast-thinking creativity and control in that stressful situation. In fact, we used a similar process ourselves: in emergency situations, an experienced engineer would dive into finding a solution, while a more junior one would follow the checklist.<\/p>\n<p>Another critique of process is that process can inhibit effective working and collaboration. It absolutely can if process is treated as an entity justified by its own existence rather than another living asset. The only thing that can guard against this is culture.\u00a0More on that later.<\/p>\n<h3><b>Process \u2013 Tooling<\/b><\/h3>\n<p>The first thing to get right is the ticketing system. Like monitoring solutions, people obsess over which ticketing system is best. And they are wrong to. The ticketing system you use you will generally end up preferring simply due to familiarity. The ticketing system is only bad if it drives or encourages bad processes. What a bad process is depends on the constraints of your business.<\/p>\n<p>It\u2019s far more important to have a ticketing system that functions reliably and supports your processes than the other way round.<\/p>\n<p>Here\u2019s an example. We moved from RT to JIRA during my tenure. JIRA offered many advantages over RT, and I would generally recommend JIRA as a collaborative tool. The biggest problem we had switching, however, was the loss of some functionality we\u2019d\u00a0built into RT, which was critical to us. RT allowed us to get real-time updates on tickets, which meant that collaboration on incidents was somewhere between chat and ticketing. This record was invaluable in post-incident review. RT also allowed us to hide entries from customers, which again was really hard to lose. We got over it, but these things were surprisingly important because they\u2019d\u00a0become embedded in our process and culture.<\/p>\n<p>When choosing or changing your ticketing system, think about what\u2019s really important to operations, not specific features that seem nice when on a list. What\u2019s important to you can vary from how nice it looks (seriously \u2013 your customers might take you more seriously, and your brand might be about good design), to whether the reporting tools are powerful.<\/p>\n<h2><b>Documentation<\/b><\/h2>\n<p>After process, documentation is the most important thing, and the two are intimately related.<\/p>\n<p>There\u2019s a book in documentation, because, again, people focus on the wrong things. The critical thing to understand is that documentation is an asset like any other. Like any business assets, documentation:<\/p>\n<ul class=\"alternate\" type=\"square\"><li>If properly looked after, will return investment many times over<\/li>\n<li>Requires investment to maintain (like the fabric of a factory)<\/li>\n<li>If out of date, costs money simply because it\u2019s there (like out-of-date inventory)<\/li>\n<li>If of poor quality, or not usable is a liability, not an asset<\/li>\n<\/ul><p>But this is not controversial \u2013 few people disagree with the idea that good documentation is useful. The point is: what do you do about it?<\/p>\n<h3><b>Documentation \u2013 Where We Were<\/b><\/h3>\n<p>We were in a situation where documentation provided to us was not useful (eg from devs: \u2018a network partition is not covered here as it is highly unlikely\u2019. Well, guess what happened! And that was documentation they kindly bothered to write\u2026), or we simply relied on previously-journalled investigations (by this time we <em>were<\/em> writing things down) to figure out what to do next time something similar happened.<\/p>\n<p>This was frustrating all of us, and we spent a long time complaining about the documentation fairy not visiting us before we took responsibility for it ourselves.<\/p>\n<h3><b>Documentation \u2013 What I Did<\/b><\/h3>\n<p><img data-attachment-id=\"2752\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/pile_of_papers\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=525\" data-orig-size=\"612,640\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"pile_of_papers\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=525?w=287\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=525?w=525\" class=\"alignnone size-full wp-image-2752\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=525\" alt=\"pile_of_papers\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=525 525w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=143 143w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg?w=287 287w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/pile_of_papers.jpg 612w\" \/><\/p>\n<p>Here\u2019s what I did.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>I took two years\u2019 worth of priority incidents (ie those that triggered \u2013 or would have triggered \u2013 an out of hours call), and listed them. There were over 1700 of them.<\/li>\n<li>Then I categorised them by type of issue.<\/li>\n<li>Then I went through each type of issue and summarised the steps needed to either resolve, or get to a point where escalation was required<\/li>\n<\/ul><p>This took <em>seven months<\/em> of my full-time attention. I was a senior employee and I was costing my company lots of money to sit there and write. And because I had a clueful boss, I never got questioned about whether this was a good use of time. I was trusted (culture, again!). I would say it took four months before any dividends at all were seen from this effort. I remember this four-month period as a nerve-wracking time, as my attention was taken away from operations to what could have been a complete waste of my time and my employer\u2019s money and an embarrassing failure.<\/p>\n<p>Why not give it to an underling to do? For a few reasons. This was so important, and we had not done it before so I needed to know it was being done properly. I knew exactly what was needed, so I knew I could write it in such a way that it would be useful to me at the very least. I was also a relatively experienced writer (arts grad, former journalist), so I liked to think that that would help me write well.<\/p>\n<p>We called these \u2018Incident Models\u2019 as per ITIL, but they can also be called \u2018run books\u2019, \u2018crib sheets\u2019, whatever. It doesn\u2019t matter. What mattered was:<\/p>\n<ul class=\"alternate\" type=\"square\"><li>They were easy to find\/search for<\/li>\n<li>It was easy to identify whether you got a match<\/li>\n<li>They were not duplicated<\/li>\n<li>They could be trusted<\/li>\n<\/ul><p>We put this documentation in plain text within the ticketing system, under a separate JIRA project.<\/p>\n<p>The documentation team got wind of what we were up to and tried to pressure us to use an internal wiki for this. We flat-out refused, and that was critical: the documentation system\u2019s colocation with the ticketing system meant that searching and updating the documentation had no impedance mismatch. Because it was plain text it was fast, simple to update, and uncluttered. We resisted process that jeopardised the utility of what we were doing.<\/p>\n<hr \/><p><strong>If you want to learn more about bash, read\u00a0my book\u00a0<a href=\"https:\/\/leanpub.com\/learnbashthehardway\">Learn Bash the Hard Way<\/a>, available at\u00a0<a href=\"https:\/\/leanpub.com\/learnbashthehardway\">$5<\/a>:<\/strong><\/p>\n<p><a href=\"https:\/\/leanpub.com\/learnbashthehardway\"><img data-attachment-id=\"4192\" data-permalink=\"https:\/\/zwischenzugs.com\/2018\/01\/06\/ten-things-i-wish-id-known-about-bash\/hero\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png\" data-orig-size=\"600,957\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"hero\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=525\" class=\"alignnone size-medium wp-image-4192\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300\" alt=\"hero\" width=\"188\" height=\"300\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300 188w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=376&amp;h=600 376w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=94&amp;h=150 94w\" \/><\/a><\/p>\n<hr \/><h3><b>Documentation and the Criticality of De-cluttering<\/b><\/h3>\n<p>When we started, we designed a schema for these Incident Models which was a thing of beauty, covering every scenario and situation that could crop up.<\/p>\n<p><img data-attachment-id=\"2761\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/bureaucracy\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg?w=525\" data-orig-size=\"435,599\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"bureaucracy\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg?w=525?w=218\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg?w=525?w=435\" class=\"alignnone size-full wp-image-2761\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg?w=525\" alt=\"bureaucracy.jpeg\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg 435w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg?w=109 109w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg?w=218 218w\" \/><\/p>\n<p>In the end it was almost a complete waste of time. What we ended up using was a really dumb structure of:<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Statement of problem<\/li>\n<li>Steps 1-n of what to do<\/li>\n<li>Further\/deeper discussion, related articles<\/li>\n<\/ul><p>That was it. Attempts to structure it more thoroughly all failed as it was either confusing to newcomers, created too much administrative overhead, or didn\u2019t cover enough. Some articles developed their own schema over time that was appropriate to the task, and new categories (eg the \u2018jump-off\u2019 article that told you which article to go to next) evolved over time. We couldn\u2019t design for these things in advance because we didn\u2019t know what would work or what would not.<\/p>\n<p>Call it \u2018agile documentation\u2019 if you want \u2013 agile\u2019s what sells these days (it was ITIL back then). Again, what was critical was that <b>simplicity and utility<\/b> trumped everything else.<\/p>\n\n\n<h3><b>There Is No Documentation Fairy<\/b><\/h3>\n<p>Having spent all this time and effort a couple of other things became clear regarding documentation.<\/p>\n\n<p><img data-attachment-id=\"2753\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/documentation_fairy\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=525\" data-orig-size=\"800,800\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"documentation_fairy\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=525?w=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=525?w=525\" class=\"alignnone size-full wp-image-2753\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=525\" alt=\"documentation_fairy\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=525 525w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=150 150w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=300 300w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg?w=768 768w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/documentation_fairy.jpg 800w\" \/><\/p>\n<p>First, we gave up accepting documentation from other teams. If they commented code, great, if there was something useful on the wiki for us to find, also great. But when it came to handing over projects we stopped \u2018asking for documentation\u2019. Instead we\u2019d arrange sessions with experienced SREs where the design of the project would be discussed.<\/p>\n<p>Invariably (assuming they had no ops experience), the developer would focus on the things they\u2019d built and how it worked \u2013 and these things were often the most thoroughly tested and least likely to fail.<\/p>\n<p>By contrast, the SRE would focus on the weak points, the things that would go wrong. \u2018What happens if the network gets partitioned? What if the database runs out of disk? Can we work out from the logs why the user didn\u2019t get paid?\u2019<\/p>\n<p>We\u2019d then go away and write our own documentation and get the engineer to sign off on it \u2013 the reverse of the traditional flow! They\u2019d often make useful comments and give us added insights in the process.<\/p>\n<p>The second thing we noticed was that our engineers were still reluctant to update the docs that only they were using. There was still a sense that documentation should be given to them. The leadership had to constantly reinforce that this was their documentation, not tablets of stone handed down from on high, and if they didn\u2019t constantly maintain this, they would become useless.<\/p>\n<p><img data-attachment-id=\"2764\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/writer-at-computer\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=525\" data-orig-size=\"640,433\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"writer-at-computer\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=525?w=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=525?w=525\" class=\"alignnone size-full wp-image-2764\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=525\" alt=\"writer-at-computer.jpg\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=525 525w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=150 150w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg?w=300 300w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/writer-at-computer.jpg 640w\" \/><\/p>\n<p>This was a cultural problem and took a long time to undo. Undoing it also required the documentation changes to be reinforced by process.<\/p>\n<p>In the end, I\u2019d say about 10% of the ongoing working time was spent maintaining and writing documentation. After the initial 7-month burst, most of that 10% was spent on maintenance rather than producing new material.<\/p>\n<hr \/><p><strong>If you want to learn more about git, read\u00a0my book\u00a0<a href=\"https:\/\/leanpub.com\/learngitthehardway\">Learn Git the Hard Way<\/a>, available at\u00a0<a href=\"https:\/\/leanpub.com\/learngitthehardway\">$8<\/a>:<\/strong><\/p>\n<p><a href=\"https:\/\/leanpub.com\/learngitthehardway\/\"><img data-attachment-id=\"4278\" data-permalink=\"https:\/\/zwischenzugs.com\/learngitthehardway\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png\" data-orig-size=\"1410,2250\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"learngitthehardway\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png?w=188&amp;h=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png?w=525\" class=\"size-medium wp-image-4278 aligncenter\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png?w=188&amp;h=300\" alt=\"learngitthehardway\" width=\"188\" height=\"300\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png?w=188&amp;h=300 188w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png?w=376&amp;h=600 376w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/03\/learngitthehardway.png?w=94&amp;h=150 94w\" \/><\/a><\/p>\n<hr \/><h3><b>Documentation \u2013 Benefits<\/b><\/h3>\n<p>After getting all this documentation done, we experienced benefits <em>far<\/em> in excess of the 10% ongoing cost. To call out a few:<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Easier onboarding<\/li>\n<\/ul><p>Before this process started we were reluctant to take on less experienced staff. After, onboarding became a breeze. Among other things the training involved following incidents as they happened and shadowing more experienced staff. New staff were tasked with helping maintain docs, which helped them understand what gaps they had in their knowledge.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Better training<\/li>\n<\/ul><p>The docs gave us a resource that allowed us to identify training requirements. This ended up being a curriculum of tools and techniques that any engineer could aim to get a working knowledge of.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Less stress through simpler escalation<\/li>\n<\/ul><p>These was a big one. Before we had the step-by-step incident models, when to escalate was a stressful decision. Some engineers had a reputation for escalating early, and all were insecure about whether they\u2019d \u2018missed something obvious\u2019 before calling a responsible tech lead out of hours. SREs would also get called out for <em>not<\/em> escalating early enough as well!<\/p>\n<p>The incident models removed that problem. Pretty soon, the first question an escalated-to techie asked was \u2018have you followed the incident model\u2019? If so, and there something obvious was missed, then gaps in it became clear and quickly-fixed. Soon, non-SREs were busy updating and maintaining the docs themselves for when they were escalated-to. It became a virtuous circle.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Better discipline<\/li>\n<\/ul><p>The obvious value of documentation to the team helped improve discipline in other respects. Interestingly, SREs previously had the reputation for being the \u2018loudest\u2019 team \u2013 there was often a lot of \u2018lively\u2019 debate, and the team was very social \u2013 which made sense, as we relied on each other as a team to cover a large technical area, dealt with often non-technical customer execs, and sharing knowledge and culture was critical.<\/p>\n<p>As time progressed, the team became quieter and quieter \u2013 partly due to the advent of chatrooms, increased remote working, and international teams, but also due to the fact that so much of the work became routine: follow the incident model, when you\u2019re done, or don\u2019t understand something, escalate to someone more senior<strong>.<\/strong><\/p>\n<ul class=\"alternate\" type=\"square\"><li>Automation<\/li>\n<\/ul><p>Automating the investigations this way meant that the way was clear to further automate them with software.<\/p>\n<p>Having metrics on which tickets were linked to which incident models meant that we knew where best to focus our effort. We wrote scripts to comb through log files in the background, make encoding issues quicker and simpler to figure out, automate responses to customers (\u2018Issue was caused by a change made by app admin user XXX\u2019), and a lot more.<\/p>\n<p>These automations inspired an automation tool we built for ourselves based on pexpect: <a class=\"external-link\" title=\"Follow link\" href=\"http:\/\/ianmiell.github.io\/shutit\/\" rel=\"nofollow\">http:\/\/ianmiell.github.io\/shutit\/<\/a> But that\u2019s another story. Basically, once we got going it was a virtuous circle of continuous improvement.<\/p>\n<h3><b>Back to Process<\/b><\/h3>\n<p>Given you have all these assets, how do you prevent them from degrading in value over time? This is where process is critical.<\/p>\n<p>Two processes were critical in ensuring everything continued smoothly: triage and post-incident review.<\/p>\n<h3><b>Process \u2013 Triage<\/b><\/h3>\n<p><img data-attachment-id=\"2770\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/heres-a-copy-of-our-new-triage-plan-the-order-is-now-walking-wounded-first-the-dying-and-dead-second-lawyers-last\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg?w=525\" data-orig-size=\"400,323\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;Bruce Bolinger via CartoonStock - www.cartoonstock.com\\\/cartoonview.asp?catref=bbon56&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;'Here's a copy of our new triage plan...the order is now walking wounded first, the dying and dead second, lawyers last...'&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;\\u00a9 Bruce Bolinger via CartoonStock - www.cartoonstock.com\\\/cartoonview.asp?catref=bbon56&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;'Here's a copy of our new triage plan...the order is now walking wounded first, the dying and dead second, lawyers last...'&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"\u2018Here\u2019s a copy of our new triage plan\u2026the order is now walking wounded first, the dying and dead second, lawyers last\u2026\u2019\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg?w=525?w=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg?w=525?w=400\" class=\"alignnone size-full wp-image-2770\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg?w=525\" alt=\"'Here's a copy of our new triage plan...the order is now walking wounded first, the dying and dead second, lawyers last...'\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg 400w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg?w=150 150w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/traige.jpg?w=300 300w\" \/><\/p>\n<p>5%-10% of time was spent on the triage process. Again, it took a long time to get the process right, but it resulted in massive savings:<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Reduce the steps to the minimum <em>useful<\/em> steps<\/li>\n<\/ul><p>It\u2019s so tempting to put as much as possible into your triage process, but it\u2019s vital to keep the value in the process over completeness. Any step that is not often useful tends to get skipped over and ignored by the triager.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Focus on saving cost in the process<\/li>\n<\/ul><p>Looking for duplicates, finding the relevant incident model, reverting quickly to the customer, and escalating early all reduced the cost per ticket significantly. It also saved other engineers the context switch of being asked a question while they\u2019re thinking about something else. It\u2019s hard to evaluate the benefits of these items, but we were able to deal with increased volumes of incidents with fewer people and less difficulty. Senior management and customers noticed.<\/p>\n<p>Recording the details of these efforts also saved time, as (for example) engineers given a triaged ticket could see that the triager searched for previous incidents with a string that maybe they could improve on. It also meant that more experienced staff could review the triage quality.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Review triage<\/li>\n<\/ul><p>Experienced staff need to review the triage process regularly to ensure it\u2019s actually being applied effectively.<\/p>\n<p>When I moved to another operations team (in a domain I knew far less about), I cut the incident queue in half in about 3 days, just by applying these techniques properly. The triage process was there, but it wasn\u2019t being followed with any thought or oversight, and was given to a junior member of staff who was not the most capable. Big mistake. Triage must be done \u2013 or overseen \u2013 by someone with a lot of experience, as while it looks routine and mechanical it involves a lot of significant decisions that rest on experience in the field.<\/p>\n<p>And yes, I was the new boss, and I chose to spend my first week doing the \u2018lowly\u2019 task of triage. That\u2019s how important I thought it was.<\/p>\n<ul class=\"alternate\" type=\"square\"><li>Rota the task<\/li>\n<\/ul><p>No-one wants to do triage for long, so we rota\u2019d it per week. This allowed some continuity and consistency, but stopped engineers from going crazy by spending too long doing the same task over and over.<\/p>\n<h3><b>Process \u2013 Post-Incident Review<\/b><\/h3>\n<p><img data-attachment-id=\"2767\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/autopsy\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=525\" data-orig-size=\"590,421\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"autopsy\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=525?w=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=525?w=525\" class=\"alignnone size-full wp-image-2767\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=525\" alt=\"autopsy\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=525 525w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=150 150w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg?w=300 300w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/autopsy.jpg 590w\" \/><\/p>\n<p>The mirror image of triage was the \u2018post incident review\u2019. Every ticket was reviewed by an experienced team member. Again, this was a process that took up about 5% of effort, but was also significant.<\/p>\n<p>A standard form was filled out and any recommendations were added to a list of backlog \u2018improvement\u2019 tasks which could be prioritised. This gave us a number for technical\/process debt that we wanted to look at.<\/p>\n<h3><b>Culture<\/b><\/h3>\n<p><img data-attachment-id=\"2773\" data-permalink=\"https:\/\/zwischenzugs.com\/2017\/04\/04\/things-i-learned-managing-site-reliability-for-some-of-the-worlds-busiest-gambling-sites\/blame\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=525\" data-orig-size=\"4167,4167\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"blame\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=525?w=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=525?w=525\" class=\"alignnone size-full wp-image-2773\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=525\" alt=\"blame.jpg\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=525 525w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=1050 1050w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=150 150w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=300 300w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=768 768w, https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/blame.jpg?w=1024 1024w\" \/><\/p>\n<p>I\u2019ve mentioned culture a few times, and it\u2019s what you always return to if you\u2019re trying to enact any kind of change at all, since culture is at root a set of conceptual frameworks that underlie all our actions.<\/p>\n<p>I\u2019ve also mentioned that people often focus on the \u2018wrong thing\u2019. Time and again I hear people focus on tools and technology rather than culture. Yes, tools and technology are important, but if you\u2019re not using them effectively then they are worse than useless. You can have the best golf clubs in the world, but if you don\u2019t know how to swing and you\u2019re playing baseball then they won\u2019t help much.<\/p>\n<p>Culture requires investment far more than technology does (I invested over half a year just writing documentation, remember). If the culture is right, people will look for the right tools and technology when they need to.<\/p>\n<p>When given a choice about what to spend time and money on, always go for culture first. It cost me a lot of budget, but forcibly removing an \u2018unhelpful\u2019 team member was the best thing I did when I took over another team. The rest of the team flowered once he left, no longer stifled by his aggressive behaviour, and many things got done that didn\u2019t before.<\/p>\n<p>We also built a highly effective team with a budget so small that recruiters would phone me up to yell at me what I was looking for was \u2018impossible\u2019, but by focussing on the right behaviours, investing time in the people we found, and having good processes in place, we got an extremely effective and loyal team that all went on to bigger and better things within and outside the company (but mostly within!).<\/p>\n<h3><b>Politics<\/b><\/h3>\n<p>A quick word on politics. You\u2019ve got to pick your battles. You\u2019re unlikely to get the resources you need, so drop the stuff that wont get done to the floor.<\/p>\n<p>Yes, you need a monitoring solution, better documentation, better trained staff, more testing\u2026 you are not going to get all these things unless you have a money machine, so pick the most important and try and solve that first. If you try and improve all these things at once, you will likely fail.<\/p>\n<p>After process, and documentation, I tried to crack the \u2018reproducible environment\u2019 puzzle. That led me to Docker, and a complete change of career. I talk about these things a little <a href=\"https:\/\/www.youtube.com\/watch?v=Yj5I-yOaDao\">here<\/a>\u00a0and <a href=\"https:\/\/www.youtube.com\/watch?v=-qHwL8C9UoA&amp;t=4s\">here<\/a>.<\/p>\n<h2>Any Questions?<\/h2>\n<p>Reach me on twitter: @ianmiell<\/p>\n<p>Or <a href=\"https:\/\/www.linkedin.com\/in\/ian-miell-694496\">LinkedIn<\/a><\/p>\n<p id=\"76f3\" class=\"graf graf--p graf-after--h3\">My book <a class=\"markup--anchor markup--p-anchor\" href=\"http:\/\/manning.com\/miell\/?a_aid=zwischenzugs&amp;a_bid=e0d48f62\" target=\"_blank\" rel=\"nofollow noopener\">Docker in Practice<\/a>:<\/p>\n<figure id=\"d17e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\">\n<div class=\"aspectRatioPlaceholder-fill\">\n<div class=\"progressiveMedia js-progressiveMedia graf-image is-canvasLoaded is-imageLoaded\"><img class=\"progressiveMedia-image js-progressiveMedia-image\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/03\/5d342-0v9jkbliyi3uzefqq.jpg?w=640\" alt=\"image\" \/><\/div>\n<\/div>\n<\/div><\/figure><p id=\"f56f\" class=\"graf graf--p graf-after--figure graf--trailing\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Get 39% off with the code: 39miell2<\/em><\/strong><\/p>\n<hr \/><p>If you liked this post, you might also like:<\/p>\n<p><a href=\"https:\/\/zwischenzugs.com\/2018\/02\/24\/5-things-i-did-to-change-a-teams-culture\/\" target=\"_blank\" rel=\"noopener\">Five Things I Did to Change a Team\u2019s Culture<\/a><\/p>\n<p><a href=\"https:\/\/zwischenzugs.com\/2018\/01\/06\/ten-things-i-wish-id-known-about-bash\/\" target=\"_blank\" rel=\"noopener\">Ten Things I Wish I\u2019d Known About bash<\/a><\/p>\n<p><a href=\"https:\/\/zwischenzugs.com\/2017\/10\/15\/my-20-year-experience-of-software-development-methodologies\/\" target=\"_blank\" rel=\"noopener\">My 20-Year Experience of Software Development Methodologies<\/a><\/p>\n<p><a href=\"https:\/\/zwischenzugs.com\/2017\/04\/11\/five-books-i-advise-every-devops-engineer-to-read\/\" target=\"_blank\" rel=\"noopener\">Five Books I Advise Every DevOps Engineer to Read<\/a><\/p>\n<hr \/><div class=\"wpcnt\">\n\t\t\t<div class=\"wpa wpmrec\">\n\t\t\t\tAdvertisements\n\t\t\t\t<div class=\"u\">\t\t<div>\n\t\t<div id=\"atatags-208708701-5be317b4b8fa4\">\n\t\t\t\n\t\t<\/div><\/div>\t\t<div>\n\t\t<div id=\"atatags-208708704-5be317b4b8fa7\">\n\t\t\t\n\t\t<\/div><\/div><\/div>\n\t\t\t\t\n\t\t\t<\/div>\n\t\t<\/div>","created_at":"2018-11-07T16:49:57+0000","updated_at":"2018-11-07T16:49:57+0000","published_at":"2017-04-04T08:28:47+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":17,"domain_name":"zwischenzugs.com","preview_picture":"https:\/\/zwischenzugs.files.wordpress.com\/2017\/04\/bureaucracy.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":486,"title":"Bash: Argument Parsing","url":"https:\/\/medium.com\/@Drew_Stokes\/bash-argument-parsing-54f3b81a6a8f","content":"<div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"7ce9\" class=\"graf graf--h3 graf--leading graf--title\">Bash: Argument\u00a0Parsing<\/h1><p id=\"978a\" class=\"graf graf--p graf-after--h3\">One parser to rule them all<\/p><p id=\"30cb\" class=\"graf graf--p graf-after--p\">Bash is a wonderful and terrible language. It can provide extremely elegant solutions to common text processing and system management tasks, but it can also drag you into the depths of convoluted workarounds to accomplish menial jobs.<\/p><p id=\"58e4\" class=\"graf graf--p graf-after--p\">Recently I ran into one of these situations when trying to parse varying inputs in a control script. The standard tools for argument parsing were breaking at every turn as I changed argument order and added flag options for flexibility. The remainder of this article details the problems and the solution I came up with for laying argument parsing woes to rest.<\/p><h3 id=\"ece8\" class=\"graf graf--h3 graf-after--p\">Goals<\/h3><p id=\"2777\" class=\"graf graf--p graf-after--h3\">The ideal argument parser will recognize both short and long option flags, preserve the order of positional arguments, and allow both options and arguments to be specified in any order relative to each other. In other words both of these commands should result in the same parsed arguments:<\/p><pre id=\"0d60\" class=\"graf graf--pre graf-after--p\">$ .\/foo bar -a baz --long thing<br \/>$ .\/foo -a baz bar --long thing<\/pre><h3 id=\"d2d6\" class=\"graf graf--h3 graf-after--pre\">getopt(s)<\/h3><p id=\"ed9d\" class=\"graf graf--p graf-after--h3\">The most widely recognized tools for parsing arguments in bash are getopt and getopts. Though both tools are similar in name, they\u2019re very different. getopt is a GNU library that parses argument strings and supports both short and long form flags. getopts is a bash builtin that also parses argument strings but only supports short form flags. Typically, if you need to write a simple script that only accepts one or two flag options you can do something like:<\/p><pre id=\"a5de\" class=\"graf graf--pre graf-after--p\">while getopts \u201c:a:bc\u201d opt; do<br \/>  case $opt in<br \/>    a) AOPT=$OPTARG ;;<br \/>  esac<br \/>done<\/pre><p id=\"c9bc\" class=\"graf graf--p graf-after--pre\">This works great for simple option parsing, but things start to fall apart if you want to support long options or mixing options and positional arguments together.<\/p><h3 id=\"eaaa\" class=\"graf graf--h3 graf-after--p\">A Better\u00a0Way<\/h3><p id=\"a1ab\" class=\"graf graf--p graf-after--h3\">As it turns out, it\u2019s pretty easy to write your own arg parser once you understand the mechanics of the language. Doing so affords you the ability to cast all manner of spells to bend arguments to your will. Here\u2019s a basic example:<\/p><pre id=\"2891\" class=\"graf graf--pre graf-after--p\">#!\/bin\/bash<\/pre><pre id=\"ca9a\" class=\"graf graf--pre graf-after--pre\">PARAMS=\"\"<\/pre><pre id=\"9505\" class=\"graf graf--pre graf-after--pre\">while (( \"$#\" )); do<br \/>  case \"$1\" in<br \/>    -f|--flag-with-argument)<br \/>      FARG=$2<br \/>      shift 2<br \/>      ;;<br \/>    --) # end argument parsing<br \/>      shift<br \/>      break<br \/>      ;;<br \/>    -*|--*=) # unsupported flags<br \/>      echo \"Error: Unsupported flag $1\" &gt;&amp;2<br \/>      exit 1<br \/>      ;;<br \/>    *) # preserve positional arguments<br \/>      PARAMS=\"$PARAMS $1\"<br \/>      shift<br \/>      ;;<br \/>  esac<br \/>done<\/pre><pre id=\"958f\" class=\"graf graf--pre graf-after--pre\"># set positional arguments in their proper place<br \/>eval set -- \"$PARAMS\"<\/pre><p id=\"908d\" class=\"graf graf--p graf-after--pre\">There\u2019s a lot going on here so let\u2019s break it down. First we set a variable `PARAMS` to save any positional arguments into for later. Next, we create a while loop that evaluates the length of the arguments array and exits when it reaches zero. Inside of the while loop, pass the first element in the arguments array through a case statement looking for either a custom flag or some default flag patterns. If the statement matches a flag, we do something (like the save the value to a variable) and we use the `shift` statement to pop elements off the front of the arguments array before the next iteration of the loop. If the statement matches a regular argument, we save it into a string to be evaluated later. Finally, after all the arguments have been processed, we set the arguments array to the list of positional arguments we saved using the `set` command.<\/p><p id=\"3f35\" class=\"graf graf--p graf-after--p graf--trailing\">With this new knowledge it becomes really easy to write robust and flexible bash utilities that can be deployed to multiple systems without worrying about the order or format of arguments.<\/p><\/div><\/div>","created_at":"2018-11-07T16:49:56+0000","updated_at":"2018-11-07T16:49:56+0000","published_at":"2015-03-24T18:51:30+0000","published_by":["Drew Stokes"],"annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"medium.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":485,"title":"Animista","url":"http:\/\/animista.net\/","content":"<p>All CSS animations generated with Animista.net are free for personal and commercial usage and are licensed under FreeBSD license.<\/p><h3 class=\"content-subtitle\">FreeBSD License<\/h3><p class=\"content-intro\">Copyright 2017 Ana Travas<\/p><p>Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:<\/p><ol><li>Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.<\/li><li>Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and\/or other materials provided with the distribution.<\/li><\/ol><p class=\"content-intro small\">THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS \"AS IS\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.<\/p>","created_at":"2018-11-07T16:49:55+0000","updated_at":"2018-11-07T16:49:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"animista.net","preview_picture":"http:\/\/animista.net\/animista-media-img.gif","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":482,"title":"Using Rust with Ruby, a deep dive with Yehuda Katz | Intercom Engineering","url":"https:\/\/engineering.intercom.com\/yehuda-on-rust-with-ruby\/","content":"<p>In this screencast, <a href=\"https:\/\/twitter.com\/wycats\">Yehuda Katz<\/a> re-implements the <a href=\"http:\/\/api.rubyonrails.org\/classes\/String.html#method-i-blank-3F\">ActiveSupport String#blank?<\/a> method in <a href=\"https:\/\/www.rust-lang.org\/\">Rust<\/a>. Although just a few lines of code, Yehuda brings us all the way down to the metal and explains in detail why Rust is both blazingly fast and safe. He also talks about what the future will look like for Ruby developers who wish to use Rust in their applications.<\/p><p>This is the second in a series of screencasts that we recorded with Yehuda in Intercom. Check out the <a href=\"https:\/\/engineering.intercom.com\/yehuda-on-htmlbars\">first here.<\/a><\/p><iframe width=\"420\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/IqrwPVtSHZI\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe>","created_at":"2018-11-07T16:49:35+0000","updated_at":"2018-11-07T16:49:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"engineering.intercom.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":483,"title":"Changes I would make to Go","url":"http:\/\/sitr.us\/2017\/02\/21\/changes-i-would-make-to-go.html","content":"<p><em>Last updated 2017-03-20<\/em><\/p><p>I have been programming primarily in Go for about six months.\nI find it frustrating.\nThere are two reasons for this:<\/p>\n<p>First, functional programming is particularly difficult in Go.\nIn fact the language <a href=\"https:\/\/en.wikipedia.org\/wiki\/Go_(programming_language)#Conventions_and_code_style\">discourages functional programming<\/a>.\nThis frustrates me because the imperative code that I write requires a lot of\nboilerplate,\nand I think it is more error-prone than it could be if I could use functional\nabstractions.<\/p>\n<p>Second, I see Go as a missed opportunity.\nThere is exciting innovation in programming languages -\nespecially in areas of type checking and type inference to make code safer,\nfaster, and more ergonomic.\nI wish Google had chosen to put their weight behind some of those ideas.<\/p>\n<p>I am not the first person to look at Go this way.\nHere are other posts that echo my feelings:<\/p>\n<ul><li><a href=\"http:\/\/yager.io\/programming\/go.html\">Why Go Is Not Good<\/a><\/li>\n  <li><a href=\"http:\/\/crufter.com\/@crufter\/everyday-hassles-in-go\">Everyday hassles in Go<\/a><\/li>\n  <li><a href=\"https:\/\/www.barrucadu.co.uk\/posts\/etc\/2016-08-25-three-months-of-go.html\">Three Months of Go (from a Haskeller\u2019s perspective)<\/a><\/li>\n  <li><a href=\"http:\/\/journal.stuffwithstuff.com\/2010\/10\/21\/the-language-i-wish-go-was\/\">The Language I Wish Go Was<\/a><\/li>\n<\/ul><p>I will add to those with some of my own opinions.\nTo highlight exactly how I think Go could be better,\nI want to make comparisons to Rust.<\/p>\n\n<p>Work on Go and Rust began at close to the same time:\nGo was first announced to the world in 2009,\nRust in 2010.\nThere is a lot of overlap in their philosophies;\nboth languages:<\/p>\n<ul><li>compile to fast, native binaries<\/li>\n  <li>eschew inheritance in favor of composition<\/li>\n  <li>support imperative programming<\/li>\n  <li>omit exception-catching in favor of explicitly passing error results<\/li>\n  <li>emphasize concurrency<\/li>\n  <li>come with static type-checking<\/li>\n  <li>come with a modern packaging system that promotes modularity<\/li>\n<\/ul><p>Both languages may have been intended to replace C++:\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Go_(programming_language)#History\">Go\u2019s designers have said<\/a> that a primary motivation for Go was their\ndislike of the complexity of C++.\nOne of Mozilla\u2019s prominent applications for Rust is <a href=\"https:\/\/servo.org\/\">Servo<\/a>, a potential\nreplacement for the Gecko HTML rendering engine, which is written in C++.<\/p>\n<p>As I see it the key differences are that\nRust aims for soundness,<sup id=\"fnref:soundness\"><a href=\"#fn:soundness\" rel=\"footnote\">1<\/a><\/sup>, powerful abstractions, and high performance;\nGo aims to be accessible, consistently simple, and fast to compile.<\/p>\n<p>That said, Rust is not necessarily a replacement for Go,\nand I do not mean to say that everyone using Go should switch to Rust.\nRust supports real-time operation,\ncan operate with stack-only allocation if necessary,\nand has a sophisticated type system that can, for example,\ndetect problems due to concurrent access to shared data at compile time.\nThose requirements add to the complexity of Rust programs.\nThe <a href=\"https:\/\/doc.rust-lang.org\/book\/ownership.html\">borrow-checker<\/a> in particular has a reputation for its learning curve.\nMy purpose for making comparisons to Rust is to provide context for specific\nways in which I think Go could be better.\nRust borrows a lot of good ideas from other languages,\nand puts them together in one nice package.\nMost of those ideas are not even tied to borrow-checking.\nI think that Go could be a better language if it adopted some of the same ideas\nthat Rust has adopted.<\/p>\n<p>You can grab the <a href=\"https:\/\/github.com\/hallettj\/comparing-go-and-rust\">code examples<\/a> from this post.\nThe examples come with executable tests to encourage experimentation.<\/p>\n<p>So, here is my view of how Go compares:<\/p>\n<h2 id=\"the-good\">The Good<\/h2>\n<p>I think that Go\u2019s use of interfaces to promote composition is great.<\/p>\n<p>I like the separation of behavior and data:\nstructs store data, methods manipulate data in structs -\nit is clean separation of state (structs), and behavior (methods).\nI think that distinction can become unfortunately blurry in languages that use\ninheritance.<\/p>\n<p>Go is easy to learn.\nGo repurposes object-oriented concepts to make something new in a way that\nmakes it approachable to programmers who are familiar with other\nobject-oriented languages.<\/p>\n<p>There is often a pretty clear \u201cGo way\u201d to solve a problem.\nThis is also an oft-touted virtue of, for example, Python.\nEncouraging consistent idioms via the languages makes it likely that any Go\nprogrammer will be able to understand code written by any other Go programmer.\nThis is part of a philosophy of simplicity that is described in the keynote\n<a href=\"https:\/\/dave.cheney.net\/2015\/03\/08\/simplicity-and-collaboration\">Simplicity and collaboration<\/a>.<\/p>\n<p>There are lots of features in the Go standard libraries that have clearly had\na lot of thought put into them.\nThis is one of my favorites:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\nfiveSeconds := 5 * time.Seconds\n<\/figcaption><\/figure><\/div>\n<p>Goroutines are cheap,\nso programs can be structured in the way that makes most sense algorithmically,\neven if that involves spawning large numbers of goroutines.\n(But this is not unique to Go.\nErlang and Scala also implement lightweight actors.\nRust and other languages have their own solutions for lightweight concurrent\nand parallel programming.)<\/p>\n<p>Since I am using Rust as a point-of-reference I will point out that\nRust has a separation of behavior and data that is a lot like Go\u2019s,\nand Rust also goes for composition-over inheritance.\nInstead of structs and interfaces,\nRust uses structs, enums, and traits.\nRust traits serve the same purpose as interfaces;\nbut they are different enough that they might seem a little weird to\nprogrammers with an object-oriented background.\nRust and Go differ in that Rust prioritizes expressiveness over simplicity,\nand type safety over fast compile times.\n(Fast compile times are a high priority for the Rust team -\njust not the top priority.)<\/p>\n<p>I could go on - there are plenty of nice features in Go.\nBut there are also:<\/p>\n<h2 id=\"the-bad\">The Bad<\/h2>\n<p>These are features of Go that I find especially frustrating.<\/p>\n<h3 id=\"nil\"><code>nil<\/code><\/h3>\n<p>I am disappointed by the decision to include null pointers in a new language\nwhen safer solutions have been in use for decades.\nOr to be more precise: I think it is a bad idea for a language to use <code>nil<\/code> as\na bottom-ish type,\nwhere <code>nil<\/code> is type-compatible with every pass-by-reference type.<\/p>\n<p>I understand that <code>nil<\/code> is not technically a <code>null<\/code> pointer -\nbut its behavior is close enough that criticisms that apply to null pointers\nalso apply to <code>nil<\/code>.\nI have read <a href=\"https:\/\/speakerdeck.com\/campoy\/understanding-nil\">Understanding Nil<\/a>,\nand I understand that it is possible to implement methods where the receiver is\n<code>nil<\/code>,\nand that <code>nil<\/code> can be useful.\nGo does some nice things to make <code>nil<\/code> less bad than it could be.\nBut the fact remains:\n<code>nil<\/code> is type-compatible with every pass-by-reference type,\nwhether or not methods on that type have sensible behavior for <code>nil<\/code> receivers,\nand that leads to lots of opportunities for runtime errors.\nTo me that seems like an opportunity to make language changes so that it is\neasier for the type checker to catch problems at compile time.<\/p>\n<p>Some newer languages have a <code>null<\/code>, but treat it as a distinct type that is\ngenerally not compatible with other types.\n(For example, <a href=\"http:\/\/fantom.org\/\">Fantom<\/a> and <a href=\"https:\/\/flowtype.org\/\">Flow<\/a> do this.)\nIn those languages values are non-nullable by default.\nHere is how one might declare and use a nullable variable in Flow when writing\nReact code:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\nfunction LoginForm(props) {\n  \/\/ The `?` in front of `HTMLInputElement` indicates that `emailInput` might be\n  \/\/ `null`.\n  let emailInput: ?HTMLInputElement\n  \/\/ JSX syntax permits HTML-like tags in code\n  return &lt;form onSubmit={event =&gt; props.onLogin(event, emailInput)}&gt;\n    &lt;input type=\"email\" ref={thisElement =&gt; emailInput = thisElement} \/&gt;\n    &lt;input type=\"submit\" \/&gt;\n  &lt;\/form&gt;\n}\nfunction onLoginTake1(event: Event, emailInput: ?HTMLInputElement) {\n  event.preventDefault()\n  \/\/ Type error! Cannot read property `value` on possibly `null` or `undefined`\n  \/\/ value.\n  dispatch(loginEvent(emailInput.value))\n}\nfunction onLoginTake2(event: Event, emailInput: ?HTMLInputElement) {\n  event.preventDefault()\n  if (emailInput) {\n    \/\/ This is ok because Flow infers that `emailInput` cannot be `null` or\n    \/\/ `undefined` in this block.\n    dispatch(loginEvent(emailInput.value))\n  }\n}\n<\/figcaption><\/figure><\/div>\n<p>Without a concept of nullability,\nuses of <code>nil<\/code> are contradictions of what is stated in\nyour type signatures.\nThis Go signature claims that the argument is a pointer to a <code>User<\/code> struct -\nbut if you take that claim at face value you are likely to get a <code>nil pointer\ndereference<\/code> error:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\nfunc validate(user *User) bool {\n\treturn user.Name != \"\"\n}\n<\/figcaption><\/figure><\/div>\n<p>In Go every variable with a pass-by-reference type comes with the implied\nambiguity,\n\u201c\u2026or it might be <code>nil<\/code>\u201d.\nSupport for non-nullable types makes a language expressive enough to avoid that\nambiguity.<\/p>\n<p>The problem with <code>nil<\/code> in Go is exacerbated by the fact that <code>nil<\/code> checks\nsometimes fail.\nIf an interface value has a concrete type but its value is <code>nil<\/code>,\n<a href=\"http:\/\/devs.cloudimmunity.com\/gotchas-and-common-mistakes-in-go-golang\/index.html#nil_in_nil_in_vals\">a <code>nil<\/code> check will not return <code>true<\/code><\/a>.\nThe purist explanation for this is that the value is not really <code>nil<\/code>:\nit is an interface value that happens to have <code>nil<\/code> in its value slot.\nI don\u2019t find that explanation to be satisfactory.\nWhen a method is dispatched on that not-really-nil value,\nthe receiver value will really be <code>nil<\/code> in the method body.<\/p>\n<p>But what about zero values?\nWhat would the zero value be for a function type or an interface type\nwithout <code>nil<\/code>?\nWell, I think that zero values are also a bad idea.<\/p>\n<p>One of the design decisions in Go is that every type must have\na well-defined default value, which is called the zero value.\nThis can be convenient because it means you do not have to write constructors\nby hand when all you need is a default value.\nBut I suspect that the real reason why zero values are a part of Go is that\nthey provide well-defined behavior for what happens when you use uninitialized\nvariables.\nC and C++ are notorious for <a href=\"http:\/\/blog.llvm.org\/2011\/05\/what-every-c-programmer-should-know.html\">undefined behaviors<\/a>,\nwhich lead to traps for programmers, and to problems making code portable\nbetween compiler implementations.\nUse of uninitialized variables is one notable example of undefined behavior in\nboth languages.<sup id=\"fnref:changing-specs\"><a href=\"#fn:changing-specs\" rel=\"footnote\">2<\/a><\/sup>\nI imagine that the designers of Go learned from C and C++,\nand set out to clearly define as much of Go\u2019s behavior as possible.\nI think that is a great attitude for language designers to adopt!\nBut there is another option that I think leads to better code safety:\nRust, Flow, and other languages use <a href=\"https:\/\/en.wikipedia.org\/wiki\/Data-flow_analysis\">data-flow analysis<\/a> to detect uses of\nuninitialized variables,\nand fail type-checking if any such uses are found.<\/p>\n<p>The zero-value requirement introduces a constraint that <code>nil<\/code> must exist and\nmust be assignable to a variety of types.\nA lot of types have no sensible default value,\nso <code>nil<\/code> is the only choice.\nSo that is one problem.\nAnother problem is that the language does not have enough information to\nproduce sensible default values for domain-specific types.\nThe fact that it tries anyway undermines soundness in code.\nThe zero values for functions and interface values\n(i.e., interface values with no runtime type tag)\nare not going to be useful under any circumstances.\nPointer types can implement methods with <code>nil<\/code> receivers -\nbut that is not useful for types that do not have sensible behavior for\nuninitialized values.\nDefault struct values can be useful in some cases -\nbut in other cases the default value violates invariants that would have been\nenforced by a hand-written constructor.<\/p>\n<p>The author of <a href=\"https:\/\/www.barrucadu.co.uk\/posts\/etc\/2016-08-25-three-months-of-go.html\">Three Months of Go<\/a> called out problems with zero values in\nwork at Pusher:<\/p>\n<blockquote>\n  <p>Zero values caused so many problems over the summer, because everything would appear to be fine, then it suddenly breaks because the zero value wasn\u2019t sensible for its context of use. Perhaps it\u2019s an unrelated change that causes things to break (like a struct getting an extra field).<\/p>\n<\/blockquote>\n<h3 id=\"how-rust-does-it-differently\">How Rust does it differently<\/h3>\n<p>Rust goes a step beyond nullable types:\nit does not have a <code>null<\/code> or <code>nil<\/code> value.\nRust uses enums, which are types whose values come in multiple variants,\nwhere each variant is effectively a distinct struct.\nIf you want to represent an absence of a value you use an enum variant that\nholds no data.\nThe generalized form of this pattern is known as the \u201cOption Pattern\u201d.\nThe definition of the <code>Option<\/code> type from Rust\u2019s standard library looks\napproximately like this:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\npub enum Option&lt;T&gt; {\n    None,    \/\/ Holds no data\n    Some(T), \/\/ Holds a value of the given type\n}\n<\/figcaption><\/figure><\/div>\n<p><code>None<\/code> and <code>Some<\/code> are constructors:\nessentially they are each a function that returns a value of type <code>Option&lt;T&gt;<\/code>.\n<code>Some<\/code> takes one argument, <code>None<\/code> takes zero arguments.\nGiven an <code>Option&lt;T&gt;<\/code> value, you can use pattern matching to determine which\nconstructor was used to create the value;\nand pattern matching is also how you read back any constructor arguments.\n(In the case of a value created by calling <code>Some(x)<\/code>,\npattern matching lets you get access to that <code>x<\/code> value.)<\/p>\n<p>Here is the Option pattern in action\n(<a href=\"http:\/\/rustbyexample.com\/std\/option.html\">source<\/a>):<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\nfn checked_division(dividend: i32, divisor: i32) -&gt; Option&lt;i32&gt; {\n    if divisor == 0 {\n        \/\/ Failure is represented as the `None` variant\n        None\n    } else {\n        \/\/ Result is wrapped in a `Some` variant\n        Some(dividend \/ divisor)\n    }\n}\n#[test]\nfn divides_a_number() {\n    let result = checked_division(12, 4);\n    match result {\n        Some(num) =&gt; assert_eq!(3, num), \/\/ Pattern on the left binds the variable `num`\n        None      =&gt; assert!(false, \"Expected `Some` but got `None`\")\n    };\n}\n<\/figcaption><\/figure><\/div>\n<p>An advantage of the option pattern over nullable types is that you can\ndistinguish between values like <code>None<\/code> versus <code>Some(None)<\/code>.\nIf you are looking up values in a cache (for example),\na <code>None<\/code> result might indicate that there is no entry in the cache for a given\nkey;\nand a <code>Some(None)<\/code> result might indicate that there is an entry,\nand the value of the entry is <code>None<\/code>.<\/p>\n<p>I once advocated for use of the Option pattern at a Java company -\nbut at least one of my coworkers was put off by the idea of allocating an extra\nobject on the heap just to distinguish between a value and an absence of\na value.\nRust is built with the Option pattern in mind,\nand prioritizes zero-cost abstractions:\nIf the type parameter for <code>Option&lt;T&gt;<\/code> is a reference type,\nthen the <code>None<\/code> case can be safely represented at runtime as a null pointer.\nSo the <code>Some<\/code> or <code>None<\/code> wrapper often disappears at compile time.\nIn those cases the code is just as efficient as it would have been if the\nlanguage allowed uses of unsafe <code>null<\/code> values in source code.<\/p>\n<p>In the example above since neither <code>Option&lt;i32&gt;<\/code> nor the <code>i32<\/code> parameter are\nreference types, the compiler allocates contiguous space on the stack for the\nnumeric result, and for a tag to distinguish between <code>Some<\/code> and <code>None<\/code>.\nThere is no extra heap allocation, and no added pointer indirection.<\/p>\n<p>The Rust Book has <a href=\"https:\/\/doc.rust-lang.org\/book\/error-handling.html\">a lot more detail<\/a> on error handling.<\/p>\n<p>It is possible to implement the Option pattern in Go with similar efficiency;\nand one could even get a compile-time check that errors are handled by\nimplementing a <code>match<\/code> method that uses the visitor pattern (<a href=\"https:\/\/github.com\/hallettj\/comparing-go-and-rust\/tree\/master\/go\/option\">example<\/a>).\nBut without generics there would be no type safety for values wrapped in an\n<code>Option<\/code> type.<\/p>\n<h3 id=\"error-handling-boilerplate--lack-of-compile-time-checks-unhandled-errors\">Error-handling boilerplate &amp; lack of compile-time checks unhandled errors<\/h3>\n<p>Error handling in Go has two related problems:\nthere is a lot of boilerplate required;\nand if the programmer neglects to check for an error,\nor makes a small mistake such as checking the wrong error variable,\nthe compiler will not detect the problem.<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\nfunc doStuff() error {\n\t_, err := doThing1()\n\tif err != nil {\n\t\treturn err\n\t}\n\t_, errr := doThing2()  \/\/ Error not propagated due to a bouncy key\n\tif errr != nil {\n\t\treturn err\n\t}\n\treturn nil\n}\n<\/figcaption><\/figure><\/div>\n<p>Rust has a type, <code>Result&lt;T,E&gt;<\/code>,\nthat is quite similar to <code>Option&lt;T&gt;<\/code>.\nThe difference is that the failure variant of the <code>Result&lt;T,E&gt;<\/code> enum is not empty -\nit contains an error value (of type <code>E<\/code>).\nA returned value of type <code>Result&lt;T,E&gt;<\/code> may be either\n<code>Ok(value)<\/code> (on success) or\n<code>Err(err)<\/code> (on error).<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\npub enum Result&lt;T, E&gt; {\n    Ok(T),\n    Err(E),\n}\n<\/figcaption><\/figure><\/div>\n<p>A common complaint about the Option and Result patterns is that unwrapping\nsuccess values is a pain.\nBut a language with support for pattern matching makes unwrapping painless,\nand first-class result values permit combinators that can handle a series of\npotential failures more cleanly and safely than explicit error checks.<\/p>\n<p>Consider this Go function:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\nfunc fetchAllBySameAuthor(postID string) ([]Post, error) {\n\tpost, err := fetchPost(postID)\n\tif err != nil {\n\t\treturn nil, err\n\t}\n\tauthor, err := fetchUser(post.AuthorID)\n\tif err != nil {\n\t\treturn nil, err\n\t}\n\treturn fetchPosts(author.Posts)\n}\n<\/figcaption><\/figure><\/div>\n<p>There are a few ways the <code>fetchAllBySameAuthor<\/code> function this could be\nimplemented in Rust.\nThe pattern matching approach is probably the most accessible to people who do\nnot have prior experience with the Option or Result Pattern:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\nfn fetch_all_by_same_author(post_id: &amp;str) -&gt; Result&lt;Vec&lt;Post&gt;, io::Error&gt; {\n    let post = match fetch_post(post_id) {\n        Ok(p)    =&gt; p,\n        Err(err) =&gt; return Err(err),\n    };\n    let author = match fetch_user(&amp;post.author_id) {\n        Ok(a)    =&gt; a,\n        Err(err) =&gt; return Err(err),\n    };\n    fetch_posts(&amp;author.posts)\n}\n<\/figcaption><\/figure><\/div>\n<p>The <code>match<\/code> keyword introduces a pattern-match block.\nThe block includes a pattern for each possible variant of the type of the\nexpression in the head of the <code>match<\/code> paired with an expression to evaluate if\nthat pattern matches.\nThis is somewhat like the <a href=\"https:\/\/golang.org\/doc\/effective_go.html#type_switch\">type switch<\/a> feature in Go,\nwhere the code that runs depends on the type of the variable in the head of the\n<code>switch<\/code> block.\nBut the Rust version comes with a compile-time check that a pattern is given\nfor every possible variant of the given type,\nwhich avoids potential runtime errors.\nThis is especially helpful when a custom type is updated to add new variants:\nthe compiler will immediately point out any uses of the type that need to be\nupdated.<\/p>\n<p>Anyway, that Rust code is no less verbose than the Go version -\nbut it demonstrates that unwrapping <code>Result&lt;T,E&gt;<\/code> or <code>Option&lt;T&gt;<\/code> values does\nnot have to be any more onerous than <code>nil<\/code> checks;\nand if we had left out a failure check in the Rust version Rust would have\nemitted an error at compile time.<\/p>\n<p>Rust has a macro, <code>try!<\/code>, that abstracts the pattern matches and early returns\nthat we see above.\nSo this is an equivalent function:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\nfn fetch_all_by_same_author(post_id: &amp;str) -&gt; Result&lt;Vec&lt;Post&gt;, io::Error&gt; {\n    let post   = try!(fetch_post(post_id));\n    let author = try!(fetch_user(&amp;post.author_id));\n    fetch_posts(&amp;author.posts)\n}\n<\/figcaption><\/figure><\/div>\n<p><code>try!<\/code> rewrites an expression at compile time.\nFor example, <code>try!(fetch_post(post_id))<\/code> is expanded to put the <code>fetch_post<\/code>\ncall inside a <code>match<\/code>,\nand inserts the boilerplate pattern matches for the <code>Ok<\/code> and <code>Err<\/code> match cases.<\/p>\n<p>The <code>try!<\/code> macro was used so often that Rust\u2019s designers decided to extend the\nlanguage a bit to better support the pattern:\nputting the <code>?<\/code> postfix operator at the end of an expression has the same effect.\nFor example,\nthe line <code>let post = try!(fetch_post(post_id));<\/code> can be equivalently written as\n<code>let post = fetch_post(post_id)?;<\/code>\nAnd type-checking will fail appropriately if you forget the <code>?<\/code>.<\/p>\n<p>But Go does not support macros.\nThankfully, the Result pattern does not require macros to be concise.\nFor the more functionally-inclined,\nhere is another equivalent implementation that uses combinator methods:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\nfn fetch_all_by_same_author(post_id: &amp;str) -&gt; Result&lt;Vec&lt;Post&gt;, io::Error&gt; {\n    let post   = fetch_post(post_id);\n    let author = post.and_then(|p| fetch_user(&amp;p.author_id));\n    author.and_then(|a| fetch_posts(&amp;a.posts))\n}\n<\/figcaption><\/figure><\/div>\n<p><code>and_then<\/code> is a method on <code>Result&lt;T,E&gt;<\/code> values.\nIf the value is a successful result, it runs the given callback,\nwhich should return a new <code>Result&lt;U,E&gt;<\/code> value.\nOr if the value is an error result, <code>and_then<\/code> short-circuits,\nand passes the error result through.\n<code>and_then<\/code> is a lot like the <code>then<\/code> method on Javascript promises.<\/p>\n<p>But wait - what if you want to wrap error results to add context?\nThere is a combinator for that to:\n<code>map_err<\/code> permits arbitrary transformations to error results.<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\nlet post = fetch_post(post_id)\n    .map_err(|e| io::Error::new(io::ErrorKind::NotFound, e));\n<\/figcaption><\/figure><\/div>\n<p>The idea is that failure checks are almost always the same:\ncheck for an error, return the error if it exists, otherwise continue.\nThe <a href=\"http:\/\/deviq.com\/don-t-repeat-yourself\/\">DRY<\/a> thing to do is to abstract the common pattern into a helper method\nor a macro.\nAnd again,\na common theme in all of these Rust implementations is that there is\na compile-time guarantee that every error is handled.\nThat could be with some recovery code,\nor by passing the error up the call stack.<\/p>\n<p><code>Result&lt;T,E&gt;<\/code> does not get the same disappear-at-compile-time optimization that\n<code>Option&lt;T&gt;<\/code> does because both enum variants hold data.\nBut its efficiency compares favorably with Go\u2019s multiple return values.\nGo allocates enough space for every value in a multiple return.\nRust allocates enough space to hold either <code>T<\/code> or <code>E<\/code>\n(i.e., enough space to hold the largest possible value),\nplus a tag to distinguish between an <code>Ok(value)<\/code> value and an <code>Err(err)<\/code> value.<\/p>\n<p>A nice thing about the generality of Rust enums is that if <code>Result&lt;T,E&gt;<\/code> did\nnot exist,\nit would be easy to implement it as a library.\nSo what about using the Result pattern in Go?\nWell, we can\u2019t put methods on Go tuples (a.k.a, multiple return values),\nbecause they are not first-class values.\nIt is not possible to define a function that that accepts a tuple and a callback:\na Go function that takes a tuple cannot accept additional arguments\n(because Go tuples are not first-class values).\nThose constraints make the combinator pattern difficult.\nWe could implement a custom struct type -\nbut without generics it would not be very useful.<\/p>\n<h3 id=\"list-manipulation-is-not-practical\">List manipulation is not practical<\/h3>\n<p>There is a special slap-in-the-face for functional programmers built into Go:\nthere is no good way to write a polymorphic function that can manipulate slices\nwith arbitrary types.\nIn Rust you can write a function with a signature that looks like this:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\nfn map&lt;A, B, F&gt;(callback: F, xs: &amp;[A]) -&gt; Vec&lt;B&gt;\n    where F: Fn(&amp;A) -&gt; B {\n<\/figcaption><\/figure><\/div>\n<p>That is a function that takes a callback and an input slice,\nand returns a new array that is computed by accumulating the results of\napplying the callback to each element in the input array.\nEven better, there are built-in methods in Rust\u2019s iterator types that do\nexactly this.\nThe input slice might hold any type of values;\ntype variables allow the type checker to track how the type of the output array\nrelates to the type of the input slice,\nand also allow the type checker to check that the callback has the appropriate\ninput and output types.<\/p>\n<p>That pattern does not work well in Go.\nWithout type variables the only way to express a type that is polymorphic\nover all slice types is to use the top-type: <code>[]interface{}<\/code>.\nFor example:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\nfunc Map(callback func(x interface{}) interface{}, xs []interface{}) []interface{} {\n\tys := make([]interface{}, len(xs))\n\tfor i, x := range xs {\n\t\tys[i] = f(x)\n\t}\n\treturn ys\n}\n<\/figcaption><\/figure><\/div>\n<p>But that function is not really polymorphic:\na slice type with a more specific type parameter\n(e.g., <code>[]int<\/code>)\nis not type-compatible with <code>[]interface{}<\/code>.\nSo you cannot pass a variable with type <code>[]int<\/code> to that <code>Map<\/code> function.\nYou have to create a new slice of type <code>[]interface{}<\/code> first,\nand copy <code>int<\/code> values to the new slice one-by-one in a <code>for<\/code> loop.\nThen after getting a result from <code>Map<\/code> you have to copy result values into yet\nanother slice to get the proper final slice type.\nThat means two custom loops are required around every invocation of <code>Map<\/code> -\nplus a runtime type assertion or type switch in the callback implementation.<\/p>\n<p>A slice with an arbitrary type parameter <em>is<\/em> type-compatible with the\ntop-type, <code>interface{}<\/code>.\nIf you just use <code>interface{}<\/code> type for every polymorphic argument you get\na signature like this:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\nfunc Map(callback interface{}, xs interface{}) interface{}\n<\/figcaption><\/figure><\/div>\n<p>With that signature you can pass in any slice type and callback type that you want,\nand assign the result to a variable with the proper type.\nBut to make that work it is necessary to use the reflection API to fix up\nruntime type tags for the input slice, the input callback, and the output slice.\nThe process is described in <a href=\"http:\/\/blog.burntsushi.net\/type-parametric-functions-golang\/\">Writing type parametric functions in Go<\/a>.\nThe reflection code is ugly,\nbut can be hidden in implementations of general-purpose functions.\nThe unavoidable downsides are that you lose all compile-time type-safety,\nand there is an order-of-magnitude performance cost.<\/p>\n<p>The same problem applies to other list manipulation functions:\n<code>Filter<\/code>, <code>Take<\/code>, <code>Reduce<\/code>, etc.\nThis is bad because list manipulation is the bread-and-butter of functional\nprogramming.\nThe fact that Go discourages such a basic building block as <code>Map<\/code> means that\nfunctional programming is not likely to thrive in Go,\nand the Go community will not be able to benefit from the\n<a href=\"http:\/\/www.cs.kent.ac.uk\/people\/staff\/dat\/miranda\/whyfp90.pdf\">advantages of functional programming<\/a>.<\/p>\n<p>You might have noticed a pattern here:\nGo does not support generics, and that leads to problems.\nExcept the issue is not just lack of generics.\nDynamic languages like Javascript, Python, and Ruby do not support generics\neither -\nat least not in the sense of compile-time checking.\nBut functional idioms work just fine in those languages.\nFor example,\nin Javascript you can pass any list to a generic list manipulation function,\nand it just works.\nGo occupies an awkward middle ground where it checks types at compile time,\nbut does not provide tools to explain to the compiler how input and output\ntypes relate to each other.<\/p>\n<p>Generics - and type variables in particular -\nare a means to \u201ctalk\u201d about types.\nThey let you use function signatures to make statements like,\n\u201cThis function takes a slice of values of some type, and returns a slice of\nvalues of the same type.\u201d\nWorking in a programming language with no type variables is as frustrating to\nme as talking in a spoken language without the word \u201cthe\u201d.<\/p>\n<p>So Go code must re-implement list abstractions everywhere.\nConsider this Go function:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n\/\/ Take the titles of the first `count` docs that are not archived\nfunc LatestTitles(docs []Document, count int) []string {\n\tvar latest []string\n\tfor _, doc := range docs {\n\t\tif len(latest) &gt;= count {\n\t\t\treturn latest\n\t\t}\n\t\tif !doc.IsArchived {\n\t\t\tlatest = append(latest, doc.Title)\n\t\t}\n\t}\n\treturn latest\n}\n<\/figcaption><\/figure><\/div>\n<p>That function iterates over an input collection, skips over some values, does\nsomething with the values that are not skipped, and returns a collection with\nthe results.\nIn other words, this is a <code>filter<\/code>, <code>map<\/code>, <code>take<\/code> operation.\nHere is an equivalent Rust implementation:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\nfn latest_titles(docs: &amp;[Document], count: usize) -&gt; Vec&lt;&amp;str&gt; {\n    docs.iter()\n        .filter(|doc| !doc.is_archived)\n        .map(|doc| doc.title.as_str())\n        .take(count)\n        .collect()\n}\n<\/figcaption><\/figure><\/div>\n<p>The Rust version lets you say what you mean.\nIn other words, the Rust function is \u201cdeclarative\u201d.\nAnd the difference becomes more pronounced if you want to process values\nconcurrently -\nfor example to make network requests in parallel.\nMore on that in a bit.<\/p>\n<p>I once complained to a coworker about a lack of abstractions in Go.\nMy coworker replied, \u201cWell, maybe you shouldn\u2019t be doing that.\u201d\nSo I want to emphasize that functional abstractions do not necessarily make\ncode less efficient.\nRust typically does \u201clist\u201d manipulation using lazily-evaluated iterators.\n(This is also how the standard data structures in Clojure work.)\nSo you can chain <code>filter<\/code>, <code>map<\/code>, <code>take<\/code>, without intermediate collection\nallocations,\nand without wasting cycles on computing values beyond those that the caller\nrequests.\nThe function above does not run those <code>filter<\/code> and <code>map<\/code> callbacks on every\nelement in the input collection -\nit stops processing elements as soon as it has enough results to satisfy the\n<code>take(count)<\/code> step.\nOn top of that,\n<code>iter<\/code>, <code>filter<\/code>, <code>map<\/code>, <code>take<\/code>, and <code>collect<\/code> are polymorphic methods,\nbut they are dispatched statically thanks to a compile-time monomorphization\nstep.\nAnd the compiler will probably inline the <code>filter<\/code> and <code>map<\/code> callbacks.\nThere are some notes in the Rust Book on\n<a href=\"https:\/\/rust-lang.github.io\/book\/ch13-04-performance.html\">performance of functional abstractions on iterators<\/a>.<\/p>\n<p>It is possible that my coworker was more concerned with cognitive burden\nthan with performance.\nI think that complaints of cognitive burden may be partly a result of looking\nat unfamiliar idioms.\nTo an experienced functional programmer,\na call to, e.g., <code>map<\/code> states a clear intention:\n\u201cThe input collection will be transformed according to this mapping function.\u201d\nWith some practice declarative code is fast to read and understand.\nAnd type checkers are more effective at checking declarative code than at\nchecking custom <code>for<\/code> loops.<\/p>\n<p>Let\u2019s get into the parallel-fetch problem that I mentioned.\nHere is a Go function that I wrote recently to fetch a set of documents in\nparallel:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\nfunc (client docClient) FetchDocuments(ids []int64) ([]models.Document, error) {\n\tdocs := make([]models.Document, len(ids))\n\tvar err error\n\tvar wg sync.WaitGroup\n\twg.Add(len(ids))\n\tfor i, id := range ids {\n\t\tgo func(i int, id int64) {\n\t\t\tdoc, e := client.FetchDocument(id)\n\t\t\tif e != nil {\n\t\t\t\terr = e\n\t\t\t} else {\n\t\t\t\tdocs[i] = *doc\n\t\t\t}\n\t\t\twg.Done()\n\t\t}(i, id)\n\t}\n\twg.Wait()\n\treturn docs, err\n}\n<\/figcaption><\/figure><\/div>\n<p>The use of a <code>WaitGroup<\/code>, the allocation of a new slice, copying result values,\nand explicit error checks are all low-level details that I should not have to\nreimplement every time I want to do things concurrently.<\/p>\n<p>Now that I am looking at this again,\nI realize there might be a problem here with concurrent access to the\n<code>docs<\/code> slice.\nMaybe it would be a good idea to use a mutex around updates to <code>docs<\/code>,\nor to send results from goroutines back to the main thread via a channel.\nBut if I use a channel I will need to implement a custom struct or use two channels,\nbecause I want to capture errors,\nand I cannot send the type <code>(models.Document, error)<\/code> over a channel,\nbecause Go tuples are not first-class values\u2026<\/p>\n<p>Rust gives a compile-time error if a mutable reference to a thread-unsafe data\nstructure is passed to a function that could run in another thread.\nSo I don\u2019t have to worry as much about what is or is not thread-safe when\nwriting Rust code.\nBut that is almost made moot by the fact that Rust can hide hairy concurrency\ndetails in library functions.<\/p>\n<p>Compare the Go code to an equivalent Rust function that uses the <a href=\"https:\/\/aturon.github.io\/blog\/2016\/08\/11\/futures\/\">futures<\/a>\nlibrary:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\nfn fetch_documents(ids: &amp;[i64]) -&gt; Result&lt;Vec&lt;Document&gt;, io::Error&gt; {\n    let results = ids.iter()\n        .map(|&amp;id| fetch_document_future(id));\n    future::join_all(results).wait()\n}\n\/\/ The implementation of `fetch_document_future` is left as an exercise to the reader.\n<\/figcaption><\/figure><\/div>\n<p>The Rust function works the same way that the Go function does:\nif all fetches are successful,\nyou get a collection of data from responses.\nBut if any fetches fail, you get the first failure as an error value.\nThe difference is that in the Rust version the concurrency, mapping, and\nerror-checking details are handled by a general-purpose library.\n(Another difference is that the Rust version returns as soon as any fetch fails,\nbut the Go version always waits for all fetches to finish.)<\/p>\n<p>This example shows that the <code>map<\/code> abstraction is so powerful that concurrent\nversus sequential execution is an implementation detail.\n(I will talk more about more about functional parallel processing in a later\nsection.)<\/p>\n<p>The Rust implementation assumes that <code>fetch_document<\/code> returns a <code>Future<\/code>.\nThe function <code>future::join_all<\/code> also returns a <code>Future<\/code>.\n(Futures work very similarly to promises in Javascript:\nthey represent an eventual result or error.)\nIt would be more idiomatic to return that last future directly instead of using\n<code>wait<\/code> to block on the result.\nBut blocking gives us a function that is logically equivalent to the Go version,\nand shows that using futures in Rust does not lock you into using\ncallback-driven code everywhere.<\/p>\n<p>Using futures, and the related <code>Stream<\/code> type,\nmakes some kinds of network server implementations much simpler than with\nblocking IO.\nIn particular, streaming requests and responses are easy if code is implemented\nin terms of <code>Stream<\/code> values.<\/p>\n<h3 id=\"third-party-libraries-are-second-class-citizens\">Third-party libraries are second-class citizens<\/h3>\n<p>Go has a magical function called <code>make<\/code> that seems to do whatever the standard\nlibrary authors need for a given type.\nIt takes a type as an argument, which is unlike most other Go functions.\nCalled with one argument it initializes a small slice, map, or channel.\n<code>make<\/code> can take one or two integer arguments,\ndepending on the choice of the first argument.\nFor example, when creating a slice you can provide a length and a capacity:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\nmySlice := make([]int, 16, 32)\n<\/figcaption><\/figure><\/div>\n<p>When creating a channel, you can specify the channel\u2019s buffer size with\na second argument to <code>make<\/code>.\nAs far as I know, this is the only way to set a channel\u2019s buffer size.<\/p>\n<p>It seems that standard libraries have a special privilege to overload <code>make<\/code>\nto delegate to custom constructor code when initializing their types.\nThird-party libraries do not get to do this.<\/p>\n<p>We see something like this with the <code>range<\/code> operator also.\n<code>range<\/code> is one of the few constructs that changes its behavior depending on the\nnumber of arguments that are assigned from its output:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\nfor idx, value := range values { \/* ... *\/ }  \/\/ `range` returns indexes and values\nfor idx := range values { \/* ... *\/ }  \/\/ this time it just returns indexes\n<\/figcaption><\/figure><\/div>\n<p>But more importantly:\nonly standard library types get to be range-able.\nThere is no way to make a third-party data type iterable.\nLibrary authors can implement adapters to output a view of their data structure\nas a slice, or to spit out values over a channel.\nBut that puts extra complexity on code that uses third-party data structures,\nand requires programmers to use non-standard idioms.<\/p>\n<p>Yet another privilege is that only standard types can be compared using <code>==<\/code>,\n<code>&gt;<\/code>, etc.<\/p>\n<p>And of course only standard libraries are allowed to define generic types.\nThis is severely limiting to the library ecosystem around Go.\nIt means that, for example,\na third-party functional data structures library,\nor a third-party futures library\nwill never be as usable as the standard library collection types.<\/p>\n<p>Rust supports generics for third-party code;\nand Rust implements iteration, equality, and comparison via traits that any\nthird-party type can implement.\nThird-party Rust types are nearly indistinguishable from standard library\ntypes,\nwhich promotes innovation in Rust\u2019s library ecosystem.<\/p>\n<p>Incidentally,\n<code>make<\/code> and <code>range<\/code> are ad-hoc examples of a pattern that has generalized\nsupport in Rust:\nfunctions that are polymorphic in their return type.\nRust traits are more flexible than typical object-oriented interfaces:\nwhen an interface method is dispatched the choice of the method implementation\nto execute is determined solely by the type of the receiver of the method.\nBut a trait method implementation can be selected based on\nthe type of the receiver,\nthe type of any argument position,\nthe combination of types of multiple argument positions\n(e.g., an equality trait can require that the two polymorphic arguments to an\n<code>equals<\/code> method have the same type),\nor by the expected return type.\nAn implementation of <code>make<\/code> in Rust might look like this:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\n32\n33\n34\n35\n36\n37\n38\n39\n40\n41\n42\n43\n44\n45\n46\n47\n48\n49\n50\n51\n52\n53\n54\n55\n56\n57\n58\n59\n60\nuse std::sync::mpsc::{Receiver, Sender, channel};\n\/\/ A `trait` declares some common behavior\ntrait Make {\n    fn make() -&gt; Self;\n    fn make_with_capacity(capacity: usize) -&gt; Self;\n}\n\/\/ An `impl` provides implementations of trait methods for given concrete types.\n\/\/ Note that we can implement `Make` for the standard `Vec&lt;T&gt;` type right here.\n\/\/ It is not necessary to change the original implementation of `Vec&lt;T&gt;`.\nimpl &lt;T&gt; Make for Vec&lt;T&gt; {\n    fn make() -&gt; Vec&lt;T&gt; {\n        Vec::new()\n    }\n    fn make_with_capacity(capacity: usize) -&gt; Vec&lt;T&gt; {\n        Vec::with_capacity(capacity)\n    }\n}\n\/\/ `Sender` and `Receiver` are Rust's standard channel types.\n\/\/ This trait implementation constructs a connected sender\/receiver pair.\nimpl &lt;T&gt; Make for (Sender&lt;T&gt;, Receiver&lt;T&gt;) {\n    \/\/ Rust channels do not have a fixed buffer size - a channel is either\n    \/\/ non-blocking (with an unlimited buffer), or is blocking. But a non-blocking\n    \/\/ channel has a different type; so that will require a different trait impl.\n    fn make() -&gt; (Sender&lt;T&gt;, Receiver&lt;T&gt;) {\n        channel()\n    }\n    fn make_with_capacity(_: usize) -&gt; (Sender&lt;T&gt;, Receiver&lt;T&gt;) {\n        Make::make()\n    }\n}\n#[test]\nfn makes_a_vec() {\n    \/\/ We specify a type for the variable that holds the output of `make`.\n    \/\/ This instructs the compiler which implementation of `make` to call.\n    let mut v: Vec&lt;&amp;str&gt; = Make::make();\n    v.push(\"some string\");\n    assert_eq!(\"some string\", v[0]);\n}\n#[test]\nfn makes_a_sized_vec() {\n    let v: Vec&lt;isize&gt; = Make::make_with_capacity(4);\n    assert_eq!(4, v.capacity());\n}\n#[test]\nfn makes_a_channel() {\n    \/\/ Or we can just let the compiler infer the type that we expect.\n    let (sender, receiver) = Make::make();\n    let msg    = \"hello\";\n    let _      = sender.send(msg);\n    let result = receiver.recv().expect(\"a successfully received value\");\n    assert_eq!(msg, result);\n}\n<\/figcaption><\/figure><\/div>\n<p>Any type can implement any trait.\nThe only rule is that the code for an implementation must be in the same crate\nas either the type or the trait.\n(A \u201ccrate\u201d is a distributable Rust package.)\nSo if Rust itself or a Rust library implemented <code>make<\/code>,\nthen any third-party library could define their own custom implementations.<\/p>\n<p>I had to implement <code>make<\/code> and <code>make_with_capacity<\/code> as separate methods because\nRust does not support method overloading.\nBut in theory neither does Go.<\/p>\n<h2 id=\"the-ugly\">The Ugly<\/h2>\n<p>These are some features of Go that I dislike on what I think are more\nsubjective grounds than the \u201cbad parts\u201d,\nor that could be worked around if some of the bad parts were fixed.<\/p>\n<h3 id=\"no-tagged-unions-limited-pattern-matching\">No tagged unions, limited pattern matching<\/h3>\n<p>Scala is another language that encourages passing messages over channels.\nScala supports tagged-union types in combination with pattern matching.\nThese features make great companions for channels:\na tagged union describes a fixed set of message types that a channel can accept\nor produce.<\/p>\n<p>Rust supports tagged-union types in the form of enums:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\n32\n33\n34\n35\n36\n37\n38\n39\n40\n41\n42\n43\n44\n45\n46\n47\n48\n49\n50\n51\n52\n53\n54\n55\n56\n57\n58\n59\n60\n61\n62\n63\n64\n65\n66\n67\nuse std::sync::mpsc::{Receiver, Sender, channel};\nuse std::thread;\n\/\/ This is a description of all of the messages that may be sent to a counter.\n\/\/ Sending a value that is not produced by one of these enum constructors is a\n\/\/ compile-time error.\n#[derive(Clone, Copy, Debug)]\npub enum CounterInstruction {\n    Increment(isize),  \/\/ `isize` is an integer type that matches the platform word size\n    Reset,\n    Terminate,\n}\npub type CounterResponse = isize;\nuse self::CounterInstruction::*;\npub fn new_counter() -&gt; (Sender&lt;CounterInstruction&gt;, Receiver&lt;CounterResponse&gt;) {\n    \/\/ Creating a channel produces a connected sender \/ receiver pair\n    let (instr_tx, instr_rx) = channel::&lt;CounterInstruction&gt;();\n    let (resp_tx,  resp_rx)  = channel::&lt;CounterResponse&gt;();\n    \/\/ Run the counter in a background thread\n    thread::spawn(move || {\n        let mut count: isize = 0;\n        \/\/ If the channel has closed then `recv()` will produce an `Err` value\n        \/\/ instead of `Ok`, and the loop will terminate.\n        while let Ok(instr) = instr_rx.recv() {\n            \/\/ We match on messages to pull out the possible distinct values and\n            \/\/ types from channel messages. Because each `enum` type is sealed,\n            \/\/ it is a compile-time error to omit any valid patterns.\n            \/\/ This avoids potential run-time failures if we add message types\n            \/\/ later, but forget to update the match here.\n            match instr {\n                Increment(amount) =&gt; count += amount,\n                Reset             =&gt; count = 0,\n                Terminate         =&gt; return\n            }\n            \/\/ `send` returns a `Result` because sending might fail in some\n            \/\/ circumstances. But for this example we assign the result to `_`\n            \/\/ to ignore it.\n            let _ = resp_tx.send(count);\n        };\n    });\n    \/\/ Return the instruction sender, and the response receiver\n    (instr_tx, resp_rx)\n}\n#[test]\nfn runs_a_counter() {\n    let (tx, rx) = new_counter();\n    let _ = tx.send(Increment(1));\n    let _ = tx.send(Increment(1));\n    let _ = tx.send(Terminate);\n    let mut latest_count = 0;\n    while let Ok(resp) = rx.recv() {\n        latest_count = resp\n    }\n    assert_eq!(2, latest_count);\n}\n<\/figcaption><\/figure><\/div>\n<p>Rust does not require channels to be closed explicitly.\nChannel senders and receivers implement a <code>Drop<\/code> trait;\nany type can implement <code>Drop<\/code> to schedule some cleanup code to run when a value\ngoes out of scope.\nIn this example when the background thread terminates <code>resp_tx<\/code> goes out of\nscope,\nand closes automatically.<\/p>\n<p>This block is the pattern matching code from the example above:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\nmatch instr {\n    Increment(amount) =&gt; count += amount,\n    Reset             =&gt; count = 0,\n    Terminate         =&gt; return\n}\n<\/figcaption><\/figure><\/div>\n<p>The block matches against <code>instr<\/code>,\nwhich has the type <code>CounterInstruction<\/code>.\nThere are three variants for <code>CounterInstruction<\/code>;\neach variant is represented by a pattern in the <code>match<\/code> block with code to run\nin case the pattern matches.<\/p>\n<p>Go has type switches, which are similar to pattern matching.\nComparable Go matching code could look like this:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\nswitch instr := &lt;-instrChan.(type) {\n\tcase Increment:\n\t\tcount += Increment.Amount\n\tcase Reset:\n\t\tcount = 0\n\tcase Terminate:\n\t\treturn\n\tdefault:\n\t\tpanic(\"received an unexpected message!\")\n}\n<\/figcaption><\/figure><\/div>\n<p>The difference is in compile-time type-safety:\nA tagged union describes a <em>fixed<\/em> set of possible messages.\nWhen sending a value to a channel in Rust,\nthe compiler is able to check that the value has a type that the channel\nconsumer knows how to handle.\nAnd what is especially valuable is checking that all pieces of code that send\nor receive on a channel are consistent in the types of values that the they\nproduce or consume.\nIf you make a change to the set of possible messages in a Rust program,\nbut forget to update some critical code to accommodate the change,\nthe type checker will report an error at compile time.<\/p>\n<p>Because Go does not support tagged unions,\nmessages over polymorphic channels are dynamically typed.\nYou can use an interface as the type parameter for a channel,\nwhich does limit the types of messages that are sent over the channel.\nBut a Go interface is not sealed:\nwhen a new type is created that implements the interface\nthere is no compile-time check to ensure that all consumers of the channel are\nupdated to handle the new type.\nUnpacking channel messages with a type switch\n(as opposed to using exclusively interface methods)\ncan lead to bugs that a different type system would have caught.<\/p>\n<h3 id=\"dynamic-typing\">Dynamic typing<\/h3>\n<p>Go makes heavy use of dynamic type-checking.\nAny use of <code>interface{}<\/code>, any type assertion or type switch,\nis dynamic typing.\nThis is good and bad:\nwithout generics, it is often necessary to coerce a value to a different type;\ndynamic type assertions are a safer way to do this than unchecked type casts.\nEither way the program will (probably) crash at runtime if a value ends up\nhaving an unexpected type.\nBut with an unchecked cast the program might corrupt memory or leak information\nto attackers before crashing.<\/p>\n<p>The fact that errors are reported at runtime means that problems are likely to\ngo unnoticed without good test coverage.\nThat does not just mean 100% code coverage -\nthere might be different combinations of conditions that lead to a code path\nand a type error might not manifest under every combination.<\/p>\n<p>A type checker that can detect errors at compile time is like an extra test\nsuite that always tests every combination of conditions.\nDynamic type tests are nearly unheard of in Rust because the compiler resolves\ntypes at compile time.\nPattern matches on enum variants looks a bit like a dynamic type test -\nbut pattern matching is qualitatively different because enums are sealed.\nThere is a compile-time guarantee that pattern matching will not fail.<\/p>\n<p>There are two caveats:<\/p>\n<ul><li><code>unsafe<\/code> blocks, functions, and traits in Rust do make unchecked type casts -\nespecially when interacting with the foreign function interface.\nBut <code>unsafe<\/code> implementations are delimited regions where the normal rules do\nnot apply.\nThe best practice is to keep unsafe code to a minimum;\nand <a href=\"https:\/\/alex-ozdemir.github.io\/rust\/unsafe\/unsafe-in-rust-syntactic-patterns\/#how-many-crates-use-unsafe\">most libraries<\/a> do not use any <code>unsafe<\/code> code.<\/li>\n  <li>Rust supports <a href=\"https:\/\/doc.rust-lang.org\/book\/trait-objects.html\">trait objects<\/a>,\nwhich is are cases where the compiler does not resolve a concrete type for\na value at compile time.\nBut the compiler does verify at compile-time that the value implements a given\ntrait.\nA trait object is a lot like a Go interface value -\nexcept that a trait object cannot have a <code>nil<\/code> value.\nDespite the lack of a compile-time concrete type,\nit is safe to dispatch trait methods on a trait object because there is\ncompile-time verification that whatever value is in the object implements the\ngiven trait and therefore implements the appropriate methods.<\/li>\n<\/ul><h3 id=\"dynamic-dispatch-reinventing-the-wheel\">Dynamic dispatch, reinventing the wheel<\/h3>\n<p>In general Rust resolves every value to a concrete type at compile time.\nThat means that Rust can use static dispatch when invoking trait methods.<\/p>\n<p>(As I mentioned,\nthe exception is <a href=\"https:\/\/doc.rust-lang.org\/book\/trait-objects.html\">trait objects<\/a>,\nwhich do use dynamic dispatch much like Go interface values.\nMy understanding is that idiomatic Rust code uses trait objects sparingly.\nMost uses of traits in Rust use the trait as a bound on a type variable,\nwhich leads to compile-time resolution to a concrete type.\nFor details see the Rust book sections on <a href=\"https:\/\/doc.rust-lang.org\/book\/traits.html\">traits<\/a> and <a href=\"https:\/\/doc.rust-lang.org\/book\/trait-objects.html\">trait objects<\/a>.)<\/p>\n<p>In Go every invocation of an interface method uses dynamic dispatch.\nDynamic dispatch, type assertions, and type switches require runtime reflection,\nwhich adds some runtime overhead.<\/p>\n<p>The technique of resolving concrete types at compile-time is not new to Rust,\nand it has nothing to do with borrow-checking.\nHaskell was doing the same thing at least ten years before work began on Go.\nAnd Haskell has just as much polymorphic flexibility as Rust or Go.\n(Rust traits are an adaptation of Haskell <a href=\"http:\/\/learnyouahaskell.com\/types-and-typeclasses\">type classes<\/a>.)<\/p>\n<p>The designers of Go wanted a model for polymorphism that is simpler and more\nflexible than what other object-oriented languages provide.\nIn particular they wanted composition over inheritance,\nand an ability to implement interfaces after-the-fact\nso that new interfaces can be applied to pre-existing types.\nThis is <em>exactly<\/em> what traits and type classes do.\nThe solution that Go introduced feels to me like a less-capable reinvented wheel.<\/p>\n<h3 id=\"no-first-class-tuple\">No first-class tuple<\/h3>\n<p>Go supports multiple return values.\nOther languages (including Rust) support multiple return values via first-class\ndata types called \u201ctuples\u201d.\nFirst-class values can have methods,\ncan be stored in data structures,\nand can be passed over channels.\nMultiple-values in Go cannot do any of those things.<\/p>\n<p>We already saw tuple return values in the implementation of the <code>Make<\/code> trait,\nand in the <code>new_counter<\/code> example.\nHere is a smaller example:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n\/\/ Returns a tuple\n\/\/ (`isize` is an integer type that matches the platform word size)\nfn min_and_max(xs: &amp;[isize]) -&gt; (isize, isize) {\n    let init = (xs[0], xs[0]);\n    xs.iter()\n        .fold(init, |(min, max), &amp;x| (cmp::min(min, x), cmp::max(max, x)))\n}\n#[test]\nfn consume_tuple() {\n    let xs = vec![1, 2, 3, 4, 5];\n    let (min, max) = min_and_max(&amp;xs); \/\/ unpack tuple with destructuring assignment\n    assert_eq!(1, min);\n    assert_eq!(5, max);\n}\n<\/figcaption><\/figure><\/div>\n<p>A consequence of the lack of first-classness in Go is that there is no obvious\nway to communicate potential failures over a channel.\nIt seems like this should work, but it does not:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n\/\/ Nope\nresults := make(chan (*models.Document, error))\n<\/figcaption><\/figure><\/div>\n<p>As far as I know the best options for doing this are to define a custom\n<code>struct<\/code> type;\nto coerce both success and error values to <code>interface{}<\/code>,\nand use a type switch to distinguish success and error on the other end of the\nchannel;\nor to send success and error values over two parallel channels.<\/p>\n<p>Another consequence is that there is no good way to define methods on multiple\nreturn values,\nor to define a function that accepts multiple return values with a callback.\nThis makes it impractical to define an analog of Rust\u2019s <code>and_then<\/code> combinator\nfor Go return values.<\/p>\n<h3 id=\"shortage-of-high-level-parallelism-and-concurrency-features\">Shortage of high-level parallelism and concurrency features<\/h3>\n<p>The post <a href=\"https:\/\/gist.github.com\/kachayev\/21e7fe149bc5ae0bd878\">Channels Are Not Enough<\/a> makes a detailed argument about problems\nwith Go\u2019s lack of concurrency abstractions.\nIn the section \u201cList manipulation cannot be abstracted\u201d I pointed out that Go\ndoes not provide an easy way to run a set of computations in parallel.\n@kachayev points out that the problem is more general than that.\nReally this is another symptom of lack of generics.<\/p>\n<p>In the parallel fetch example I showed a Rust solution that uses futures,\nwhich are great for concurrent IO.\nThe <a href=\"https:\/\/aturon.github.io\/blog\/2016\/08\/11\/futures\/\">futures<\/a> library relates to a larger async IO library called <a href=\"https:\/\/tokio.rs\/\">Tokio<\/a>,\nwhich seems to be popular.\nBut async IO is not intended to provide parallelism.\nThe excellent book <a href=\"http:\/\/chimera.labs.oreilly.com\/books\/1230000000929\">Parallel and Concurrent Programming in Haskell<\/a> has this\nto say about the distinction between parallel and concurrent code:<\/p>\n<blockquote>\n  <p>A parallel program is one that uses a multiplicity of computational hardware\n(e.g., several processor cores) to perform a computation more quickly. The aim\nis to arrive at the answer earlier, by delegating different parts of the\ncomputation to different processors that execute at the same time.<\/p>\n  <p>By contrast, concurrency is a program-structuring technique in which there\nare multiple threads of control. Conceptually, the threads of control execute\n\u201cat the same time\u201d; that is, the user sees their effects interleaved. Whether\nthey actually execute at the same time or not is an implementation detail;\na concurrent program can execute on a single processor through interleaved\nexecution or on multiple physical processors.<\/p>\n<\/blockquote>\n<p>Go uses goroutines for both parallel and concurrent programming.\nRust libraries offer a selection of tools whose strengths make them especially\nuseful for different kinds of problems.<\/p>\n<p>When you want true multi-core parallel processing,\nthe Map-Reduce pattern offers a powerful solution.\nThis is another application of list abstractions.\nYou can apply Map-Reduce in Rust using parallel iterators from the <a href=\"https:\/\/github.com\/nikomatsakis\/rayon\">Rayon<\/a>\nlibrary:<\/p>\n<div class=\"bogus-wrapper\"><figure class=\"code\"><figcaption>1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n\/\/ This import brings the `par_iter` method into scope.\n\/\/ `par_iter` is a method on a Rayon trait, and Rayon provides an implementation\n\/\/ of that trait for standard slice types.\nuse rayon::prelude::*;\npub fn average_response_time(logs: &amp;[LogEntry]) -&gt; usize {\n    let total = logs.par_iter()\n        .map(|ref entry| entry.end_time - entry.start_time)\n        .sum();\n    total \/ logs.len()\n}\n<\/figcaption><\/figure><\/div>\n<p>Parallel iterators implement a variation of the <code>map<\/code> method that divides work\namong a set of job queues that feed into a worker pool,\nso that invocations of the <code>map<\/code> callback run in parallel.\nAs with goroutines,\nthis is lightweight parallelism that scales to a large number of parallel tasks.\nBut parallel iterators are a high-level abstraction that take care of some\ncomplicated details for you.\nFor example, Rayon transparently splits work into batches,\nwhich provides better performance than queueing up every invocation of the\n<code>map<\/code> callback separately.\n(The default batch size is 5000 items, but that value is tunable.)\nThe <code>sum<\/code> method (which is the Reduce step in this example) is also part of Rayon -\nwhich means that it is optimized for consuming batches of results from worker\nthreads.<\/p>\n<p>It might seem strange to have to use two different libraries for concurrent vs\nparallel code.\nBut those are separate concerns, with different underlying assumptions.\nUsually you do not actually want both at the same time.<\/p>\n<h2>Conclusion<\/h2>\n<p>How do I think Go 2.0 could be better?\nGenerics.\nNearly all of my complaints boil down to lack of support for generics.\nBut I think that support for non-nil-able types,\nand getting rid of zero values would also be useful changes.\nEven Rust-style traits might be workable.\nTraits would require support for receiver-less methods;\nbit it might be possible to make traits work with Go\u2019s implicit implementation\nfeature.<\/p>\n<p>In its current form, I prefer not to use Go.\nIt is not that Go is bad - it is just that there are lots of languages\navailable that I find more enjoyable.\nWhen I work with Go I cannot help thinking about how I could be doing things\ndifferently in another language.<\/p>\n<p>If you are willing to put in the time to understand lifetimes and borrow checking,\nRust makes a fantastic language that does everything that Go does,\nbut has none of the \u201cbad parts\u201d that I called out.<\/p>\n<p>Javascript is (like Go) easy to learn,\nand has wonderful support for concurrency (if not parallelism).\nWhen paired with Flow or Typescript you get more robust type safety than Go\nprovides.\nThe combination of Javascript and Flow in particular has none of the \u201cbad\nparts\u201d from this post.<\/p>\n<p>Erlang and Scala both support lightweight concurrency in the same style that Go\ndoes,\nand they are both great for functional programming.<\/p>\n<p>Clojure is not type-safe - but it does fantastic things!\nMy favorite functional data structure implementations are the ones in Clojure.\nAnd Clojure encourages functional programming.<\/p>\n<p>Haskell has amazing type safety,\nsome of the best concurrency and parallelism features that I have seen in any\nlanguage,\nand is well-suited to network server code.\nHaskell is another language that avoids the \u201cbad parts\u201d from this post.<\/p>\n<p>We have better tools than ever for getting work done.\nEven Go - whether I like it or not - is clearly useful for building cool things.\nBut if you take anything away from this post,\nI hope it is an interest in taking a look beyond the imperative\n\/ object-oriented world.\nI encourage you to pick one of the languages above,\nand take some time to learn about it and to get a good feel for its strengths.\nI think you will be glad that you did.<\/p>\n<h2 id=\"changes\">Changes<\/h2>\n<ul><li><em>2017-03-19:<\/em> Correct inaccuracies regarding <code>nil<\/code>, slice type manipulation, and <code>make<\/code>; add discussion of zero values; lots of edits<\/li>\n  <li><em>2017-03-20:<\/em> Add link to The Language I Wish Go Was<\/li>\n  <li><em>2017-03-20:<\/em> Change one instance of \u201cnullable\u201d to \u201cnon-nullable\u201d; edit to conclusion section<\/li>\n  <li><em>2017-03-20:<\/em> Edits based on reader feedback<\/li>\n<\/ul><div class=\"footnotes\">\n  <ol><li id=\"fn:soundness\">\n      <p>Soundness is a property of a type system where any \u201cclaims\u201d made by program types are guaranteed to hold at runtime. Runtime type errors will not occur if a language is sound.<a href=\"#fnref:soundness\" rel=\"reference\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:changing-specs\">\n      <p>This may have changed over time; the C and C++ language specifications are evolving, and I am not very familiar with the details.<a href=\"#fnref:changing-specs\" rel=\"reference\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>","created_at":"2018-11-07T16:49:35+0000","updated_at":"2018-11-07T16:49:35+0000","published_at":"2017-02-21T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":46,"domain_name":"sitr.us","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":484,"title":"Why Parsing Tools Are Hard","url":"http:\/\/blog.reverberate.org\/2013\/09\/ll-and-lr-in-context-why-parsing-tools.html","content":"<p>In my last blog entry <a href=\"http:\/\/blog.reverberate.org\/2013\/07\/ll-and-lr-parsing-demystified.html\">LL\nand LR Parsing Demystified<\/a>, we explored LL and LR parsers from a\nblack-box perspective. We arrived at a model for these parsers where\nboth their input and output were streams of tokens, with the parser\ninserting rules as appropriate according to Polish and Reverse Polish\nnotation.<\/p><p>In future articles I want to focus in even closer on the details of LL\nand LR algorithms, but I realized that I should first zoom out and\ngive some motivation for why anyone should care about LL or LR to\nbegin with.<\/p>\n<p>As I wrote this article, it turned into an answer to the question \u201cwhy\nis parsing hard?\u201d Or alternatively \u201cwhy doesn\u2019t everybody use parser\ngenerators?\u201d  LL and LR parsing theory is taught in in books like <a href=\"http:\/\/www.amazon.com\/Compilers-Principles-Techniques-Tools-Edition\/dp\/0321486811\">Compilers:\nPrinciples, Techniques, and Tools<\/a> (known as \u201cThe Dragon Book\u201d and\nused in many university compilers courses), but then people graduate\nto find that most parsers in the real world don\u2019t work like this.\nWhat gives? This article is my answer to that question.<\/p>\n<p>The theory of LL and LR parsing is almost 50 years old: Knuth\u2019s paper\n<a href=\"http:\/\/www.cs.dartmouth.edu\/~mckeeman\/cs48\/mxcom\/doc\/knuth65.pdf\">On\nthe Translation of Languages from Left to Right<\/a> that first defined\nLR(k) was published in 1965.  This is only one of an incredible number\nof mathematically-oriented papers about parsing and language theory.\nOver the last 50 years academics have explored the mathematical\ndimensions of parsing with great vigor, but the field is nowhere near\nexhausted; even in the last five years we\u2019ve seen some entirely new\nand important results published.  One of the best surveys of the field\nis the book <a href=\"http:\/\/dickgrune.com\/Books\/PTAPG_2nd_Edition\/\">Parsing\nTechniques: A Practical Guide<\/a>, whose <a href=\"http:\/\/dickgrune.com\/Books\/PTAPG_2nd_Edition\/Additional.html\">bibliography<\/a>\ncontains over 1700 cited papers!<\/p>\n<p>Despite this vast body of theoretical knowledge, few of the parsers\nthat are in production systems today are textbook cases of the theory.\nMany opt for hand-written parsers that are not based on any formalism\nat all.  Language specifications are often defined in terms of a\nformalism like BNF, but it\u2019s almost never the case that real parsers\ncan be generated directly from this formalism.  <a href=\"http:\/\/gcc.gnu.org\/wiki\/New_C_Parser\">GCC moved <em>away<\/em> from\ntheir Bison-based parser to a handwritten recursive descent\nparser<\/a>.  <a href=\"https:\/\/www.reddit.com\/r\/golang\/comments\/46bd5h\/ama_we_are_the_go_contributors_ask_us_anything\/d03zx6f\">So\ndid Go<\/a>.  While some notable language implementations do use Bison\n(like Ruby and PHP), many choose not to.<\/p>\n<p>Why this divergence between theory and practice?  While it is tempting\nto blame ignorance of the literature, that could hardly explain why\nGCC moved <em>away<\/em> from an LR parser.<\/p>\n<p>I think it is safe to say that <em>pure<\/em> LL and LR parsers have proven to\nbe largely inadequate for real-world use cases.  Many grammars that\nyou\u2019d naturally write for real-world use cases are not LL or LR, as we\nwill see.  The two most popular LL and LR-based parsing tools (ANTLR\nand Bison, respectively) both extend the pure LL and LR algorithms in\nvarious ways, adding features such as operator precedence,\nsyntactic\/semantic predicates, optional backtracking, and generalized\nparsing.<\/p>\n<p>But even the evolved tools that are currently available sometimes come\nup short, and are still evolving to address the <a href=\"http:\/\/mortoray.com\/2012\/07\/20\/why-i-dont-use-a-parser-generator\/\">traditional\npain points of parser generators<\/a>.  ANTLR v4 completely reworked\nits parsing algorithm to improve ease-of-use vs ANTLR v3 with a new\nalgorithm it calls <a href=\"http:\/\/antlr.org\/papers\/allstar-techreport.pdf\">ALL(*)<\/a>.\nBison is experimenting with <a href=\"http:\/\/people.cs.clemson.edu\/~malloy\/papers\/sac08\/paper.pdf\">IELR<\/a>,\nan alternative to LALR that was published in 2008 and intended to\nexpand the number of grammars it can accept and parse efficiently.\nSome people have explored alternatives to LL\/LR such as Parsing\nExpression Grammars (PEGs) that attempt to solve these pain points in\na different way entirely.<\/p>\n<p>Does this mean that LL and LR are obsolete?  Far from it.  While\n<em>pure<\/em> LL and LR do indeed come up short in several ways, these\nalgorithms can be extended in ways that preserves their strengths, in\nmuch the same way that a multi-paradigm programming language can offer\nfeatures of imperative, functional, and object-oriented programming\nstyles.  I firmly believe that as parser tools continue to improve\nwith better tooling, better error reporting, better visualization,\nbetter language integration, etc. they will become something you\u2019d\nreach for as readily as you reach for a regex today.  There is a lot\nof room for improvement in this space, and I want to help make that\nhappen (my tabled project <a href=\"http:\/\/www.gazelle-parser.org\/\">Gazelle<\/a> is where I have\ninvested effort so far and I intend to do more).  But I digress.<\/p>\n<p>LL and LR parsers have some indisputable strengths.  They are the most\nefficient parsing algorithms around.  The grammar analysis they\nperform ahead-of-time can tell you important things about your\ngrammar, and properly visualized can help you catch bugs, in much the\nsame way that regex visualizing tools like <a href=\"http:\/\/www.regexper.com\/\">regexper<\/a> can.  They offer some of\nthe earliest and best error reporting of syntax errors at parse time\n(this is separate from the shift\/reduce and reduce\/reduce errors you\nmight get at <em>grammar analysis<\/em> time).<\/p>\n<p>Even if you are not sold on the usefulness of LL and LR, learning\nabout them will help you better understand the tradeoffs that your\nfavorite parsing method makes compared to LL\/LR.  Alternatives to\nLL\/LR are generally forced to give up some at least one of the\nadvantages of LL\/LR.<\/p>\n<p>\u201cLL parser\u201d and \u201cLR parser\u201d are not actually specific algorithms at\nall, but rather generic terms referring to <em>families<\/em> of algorithms.\nYou may have seen names such as LR(k), \u201cfull LL\u201d, LALR(1), SLR, LL(*),\netc; these are specific algorithms (or variants of the same algorithm,\ndepending on how you look at it) that fall under the category of \u201cLL\nparser\u201d or \u201cLR parser.\u201d  These variants have different tradeoffs in\nterms of what grammars they can handle and how big the resulting\nparsing automata are, but they share a common set of characteristics.<\/p>\n<p>LL and LR parsers <em>usually<\/em> (but not always) involve two separate\nsteps: a grammar analysis step that is performed ahead-of-time and the\nactual parser that runs at parse time. The grammar analysis step\nbuilds an automaton if it can, otherwise the grammar is rejected as\nnot LALR\/SLL\/SLR\/whatever. Once the automaton is built, the parsing\nstep is much simpler because the automaton encodes the structure of\nthe grammar such that what to do with each input token is a simple\ndecision.<\/p>\n<p>What then is the distinguishing characteristic that makes a parser an\nLL parser or an LR parser?  We will answer the question with a pair of\ndefinitions.  Don\u2019t worry if these definitions make no sense to you;\nthe entire rest of the article is dedicated to explaining them.  These\ndefinitions are not given in the literature, since they are informal\nterms, but they correspond to the general usage of what is meant if\nyou look at (for example) the Wikipedia pages for \u201cLL Parser\u201d or \u201cLR\nParser.\u201d<\/p>\n<p><em>An <strong>LL parser<\/strong> is a deterministic, canonical top-down parser for\ncontext-free grammars.<\/em><\/p>\n<p><em>An <strong>LR parser<\/strong> is a deterministic, canonical bottom-up parser for\ncontext-free grammars.<\/em><\/p>\n<p>Any parser that meets these definitions is an LL or LR parser.  Both\nthe strengths and weaknesses of LL and LR are encapsulated in these\ndefinitions.<\/p>\n<p>Beware that not every parser with \u201cLR\u201d or \u201cLL\u201d in its name is actually\nan LR or LL parser.  For example, GLR, LR(k,\u221e), and Partitioned LL(k)\nare all examples of parsing algorithms that are not actually LL or LR;\nthey are variations on an LL or LR algorithm, but give up one or more\nof the essential LL\/LR properties.<\/p>\n<p>We will now more deeply explore the key parts of these definitions.<\/p>\n<p>LL and LR parsers use <em>context-free grammars<\/em> as their way of\nspecifying formal languages.  Most programmers have seen context-free\ngrammars in one form or another, possibly in the form of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Backus%E2%80%93Naur_Form\">BNF<\/a>\nor <a href=\"https:\/\/en.wikipedia.org\/wiki\/Extended_Backus%E2%80%93Naur_Form\">EBNF<\/a>.\nA close variant called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Augmented_Backus%E2%80%93Naur_Form\">ABNF<\/a>\nis used in documentation of protocols in RFCs.<\/p>\n<p>On one hand CFGs are really nice, because they match the way that\nprogrammers think about languages.  The fact that RFCs use a CFG-like\nabstraction to write documentation speaks to how readable context-free\ngrammars can be.<\/p>\n<p>Here is the JSON context-free grammar from my previous article:<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">object \u2192 '{' pairs '}'\npairs \u2192 pair pairs_tail | \u03b5\npair \u2192 STRING ':' value\npairs_tail \u2192 ',' pairs | \u03b5\nvalue \u2192 STRING | NUMBER | 'true' | 'false' | 'null' | object | array\narray \u2192 '[' elements ']'\nelements \u2192 value elements_tail | \u03b5\nelements_tail \u2192 ',' elements | \u03b5<\/pre><\/figure><p>This is so intuitive to read that I didn\u2019t even bother to explain it.\nAn object is a bunch of pairs surrounded by curly brackets.  A \u201cpairs\u201d\nis either a pair followed by a pairs_tail or empty.  It reads really\nnicely.<\/p>\n<p>Context-free grammars not only tell us whether a given string is valid\naccording to the language, they also define a tree structure for any\nvalid string.  It\u2019s the tree structure that helps us figure out what\nthe string actually <em>means<\/em>, which is arguably the most important part\nof parsing (a compiler that did nothing but say \u201cyep, this is a valid\nprogram\u201d wouldn\u2019t be very useful).  So writing a CFG really goes a\nlong way in helping us parse and analyze a language.<\/p>\n<p>On the other hand context-free grammars can be frustrating, for two\nrelated reasons:<\/p>\n<ol><li>When writing a CFG intuitively, you often end up with something\nambiguous.<\/li>\n  <li>When writing a CFG intuitively, you often end up with something\nthat is unambiguous but can\u2019t be parsed by LL or LR algorithms.<\/li>\n<\/ol><p>While the second problem is LL\/LR\u2019s \u201cfault\u201d the first is just an\ninherent challenge of designing formal languages.  Let\u2019s talk about\nambiguity first.<\/p>\n<p>If a grammar is ambiguous, it means that there is at least one string\nthat can have multiple valid parse trees.  This is a real problem in\nthe design of a language, because the two valid parse trees almost\ncertainly have different <em>semantic<\/em> meaning.  If both are valid\naccording to your grammar, your users cannot know which meaning to\nexpect.<\/p>\n<p>The simplest and most common example is arithmetic expressions. The\nintuitive way to write a grammar is something like:<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">expr \u2192 expr '+' expr |\n       expr '-' expr |\n       expr '*' expr |\n       expr '\/' expr |\n       expr '^' expr |\n       - expr |\n       NUMBER<\/pre><\/figure><p>But this grammar is highly ambiguous because it doesn\u2019t capture the\nstandard rules of precedence and associativity.  Without these rules\nto disambiguate, a string like <code class=\"highlighter-rouge\">1+2*3-4^5<\/code> has exponentially many\nvalid parse trees, all with different meanings.<\/p>\n<p>It\u2019s possible to rewrite this to capture the rules of precedence and\nassociativity:<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">expr \u2192 expr '+' term |\n       expr '-' term |\n       term\nterm \u2192 term '*' factor |\n       term '\/' factor |\n       factor\nfactor \u2192 '-' factor |\n         prim\nprim \u2192 NUMBER |\n       NUMBER '^' prim<\/pre><\/figure><p>Now we have an unambiguous grammar that encodes the precedence and\nassociativity rules, but it is not at all easy or self-evident what\nthese rules are from reading the grammar.  For example, it\u2019s certainly\nnot obvious at first glance that all of the operators in this grammar\nare left-associative <em>except<\/em> exponentiation (<code class=\"highlighter-rouge\">^<\/code>) which is\nright-associative.  And it\u2019s not easy to write grammars in this style;\nI have a fair amount of experience writing grammars but I still have\nto slow down and be careful (without testing, I\u2019m not even 100%\nconfident I have got it right).<\/p>\n<p>It\u2019s really unfortunate that one of the first and most common use\ncases of text parsing is one that pure context-free grammars are so\nbad at.  No wonder people can get turned off of CFG-based tools when\nsomething that seems like it should be so simple ends up being so\ncomplicated.  It\u2019s especially unfortunate because other non-CFG\nparsing techniques like the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Shunting-yard_algorithm\">Shunting\nYard Algorithm<\/a> are so good at this kind of operator precedence\nparsing. This is clearly one of the most glaring examples where CFGs\nand <em>pure<\/em> LL\/LR let us down.<\/p>\n<p>Another famous example of actual grammar ambiguity is the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Dangling_else\">dangling else\nproblem<\/a>.  For languages that don\u2019t have an \u201cendif\u201d statement, what\ndoes this mean?<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">if a then if b then s else s2\n\/\/ Ambiguity: which of these is meant?\nif a then (if b then s) else s2\nif a then (if b then s else s2)<\/pre><\/figure><p>Unlike with arithmetic expressions, there are no standard\nprecedence\/associativity rules to tell us which of these\ninterpretations is \u201ccorrect.\u201d  The choice here is pretty much\narbitrary.  Any language that has this construct must tell users which\nmeaning is correct. The grammar ambiguity here is a <em>symptom<\/em> of the\nfact that our language has a confusing case.<\/p>\n<p>One final example of ambiguity, this one from C and C++.  This is\nknown as the type\/variable ambiguity.<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">  \/\/ What does this mean?\n  x * y;<\/pre><\/figure><p>The correct answer is that it depends on whether <code class=\"highlighter-rouge\">x<\/code> was previously\ndeclared as a type with <code class=\"highlighter-rouge\">typedef<\/code>.  If <code class=\"highlighter-rouge\">x<\/code> is a type, then this line\ndeclares a pointer-to-<code class=\"highlighter-rouge\">x<\/code> named <code class=\"highlighter-rouge\">y<\/code>. If <code class=\"highlighter-rouge\">x<\/code> is <em>not<\/em> a type, then this\nline multiplies <code class=\"highlighter-rouge\">x<\/code> and <code class=\"highlighter-rouge\">y<\/code> and throws away the result.  The\ntraditional solution to this problem is to give the lexer access to\nthe symbol table so it can lex a type name differently than a regular\nvariable; this is known as <a href=\"https:\/\/en.wikipedia.org\/wiki\/The_lexer_hack\">the lexer\nhack<\/a>.  (While this may seem out of place in an article about\nparsers, the same ambiguity manifests in C++ in a way that cannot so\neasily be confined to the lexer).<\/p>\n<p>In other words, this ambiguity is resolved according to the semantic\ncontext of the statement.  People sometimes refer to this as a\n\u201ccontext-sensitive,\u201d (like the article <a href=\"http:\/\/eli.thegreenplace.net\/2007\/11\/24\/the-context-sensitivity-of-cs-grammar\/\">The\ncontext sensitivity of C\u2019s grammar<\/a>), but <a href=\"http:\/\/en.wikipedia.org\/wiki\/Context-sensitive_grammar\">context-sensitive\ngrammar<\/a> is a very specific term that has a mathematical meaning in\nthe <a href=\"http:\/\/en.wikipedia.org\/wiki\/Chomsky_hierarchy\">Chomsky\nhierarchy<\/a> of languages.  The Chomsky definition refers to\n<em>syntactic<\/em> context sensitivity, which almost never occurs in computer\nlanguages.  Because of this, it\u2019s good to clarify that we are talking\nabout <em>semantic<\/em> context-sensitivity, which is an entirely different\nthing.<\/p>\n<p>A key point about semantic context-sensitivity is that you need a\nTuring-complete language to properly disambiguate between the\nambiguous alternatives.  What this means for parser generator tools is\nthat it\u2019s effectively impossible to parse languages like this\ncorrectly unless you allow the user to write arbitrary disambiguation\ncode snippets in a Turing-complete language. No mathematical formalism\nalone (not CFGs, not PEGs, not operator grammars) can be sufficient to\nexpress these languages. This is one very notable case where theory\nand practice diverge.<\/p>\n<p>In tools such as ANTLR, these disambiguating code snippets are known\nas \u201csemantic predicates.\u201d  For example, to disambiguate the\ntype\/variable ambiguity, you\u2019d need to write code to build\/maintain a\nsymbol table whenever a \u201ctypedef\u201d is seen, and then a predicate to see\nif a symbol is in the table or not.<\/p>\n<p>No matter what parsing strategy is being used, a language designer\nmust be aware of and directly confront any ambiguities in the\nlanguage.  If possible, the best idea is often to change the language\nto avoid having the ambiguity at all.  For example, most languages\nthese days do not have the dangling else problem because \u201cif\u201d\nstatements have an explicit end (either an \u201cendif\u201d keyword or curly\nbrackets that surround the statements in the \u201celse\u201d clause).<\/p>\n<p>If the designer can\u2019t or doesn\u2019t want to remove the ambiguity, they\nmust decide which meaning is intended, implement the ambiguity\nresolution appropriately, and communicate this decision to users.<\/p>\n<p>But to confront ambiguities you must first know about them.\nUnfortunately this is easier said than done.  One of the huge bummers\nabout grammars (both CFGs and other formalisms like PEGs) is that many\nof the useful questions you might want to ask about them are\n<em>undecidable<\/em> (if you haven\u2019t studied Theory of Computation, a rough\napproximation for \u201cundecidable\u201d is \u201cimpossible to compute\u201d).\nDetermining whether a context-free grammar is ambiguous is\nunfortunately one of these undecidable problems.<\/p>\n<p>If it\u2019s impossible to compute whether a grammar is ambiguous <em>a\npriori<\/em>, how can we be aware of ambiguities and address them?<\/p>\n<p>One approach is to use a parsing algorithm that can handle ambiguous\ngrammars (for example GLR).  These algorithms can handle any grammar\nand any input string, and can detect at parse time if the input string\nis ambiguous.  If ambiguity is detected, they can yield all valid\nparse trees and the user can disambiguate between them however they\nsee fit.<\/p>\n<p>But with this strategy you don\u2019t learn about ambiguity until an\nambiguous string is seen in the wild.  You can never be <em>sure<\/em> that\nyour grammar is unambiguous, because it could always be the case that\nyou just haven\u2019t seen the right ambiguous string yet.  You could ship\nyour compiler only to learn, years later, that your grammar has had an\nunknown ambiguity in it all along.  This can actually happen in the\nreal world; it was not discovered that ALGOL 60 had a \u201cdangling else\u201d\nproblem until the language had already been published in a technical\nreport.<\/p>\n<p>Another strategy is to abandon context-free grammars completely and\nuse a formalism like <a href=\"http:\/\/en.wikipedia.org\/wiki\/Parsing_expression_grammar\">parsing\nexpression grammars<\/a> that is unambiguous by definition.  Parsing\nexpression grammars avoid ambiguity by forcing all grammar rules to be\ndefined in terms of <em>prioritized<\/em> choice, so in cases where multiple\ngrammar rules match the input the first one is correct <em>by\ndefinition<\/em>.<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">\/\/ PEG solution of the if\/else ambiguity:\nstmt &lt;- \"if\" cond \"then\" stmt \"else\" stmt \/\n        \"if\" cond \"then\" stmt \/\n        ...<\/pre><\/figure><p>Prioritized choice is a great tool for resolving some ambiguities; it\nworks perfectly for the solving the dangling else problem.  But while\nthis has given us a tool for <em>resolving<\/em> ambiguity, it hasn\u2019t solved\nthe problem of <em>finding<\/em> ambiguities.  Every rule in PEGs is required\nto be defined in terms of prioritized choice, which means that every\nPEG rule could be hiding a \u201cconceptual\u201d ambiguity:<\/p>\n<figure class=\"highlight\"><pre class=\"language-text\" data-lang=\"text\">\/\/ Is this PEG rule equivalent to a &lt;- c \/ b ?\na &lt;- b \/ c\n\/\/ We can't know (it's undecidable in general),\n\/\/ so every rule could be hiding an ambiguity we don't know about.<\/pre><\/figure><p>I call this a \u201cconceptual\u201d ambiguity because even though a PEG-based\ntool does not consider this ambiguous, it still looks ambiguous to a\nuser. Another way of thinking about this is that you have resolved the\nambiguity without ever being aware the ambiguity existed, thus denying\nyou the opportunity to think about it and make a conscious choice\nabout how it should be resolved.  Prioritized choice doesn\u2019t make the\ndangling else problem go away, it just hides it.  Users still see a\nlanguage construct that could plausibly be interpreted in two\ndifferent ways, and users still need to be informed which option the\nparser will choose.<\/p>\n<p>Prioritized choice also requires that an ambiguity is resolved in the\nsame way each time; it can\u2019t accommodate cases like the C\/C++\nvariable\/type ambiguity which are resolved by semantic information.<\/p>\n<p>And unlike GLR, Packrat Parsing (the linear-time algorithm for parsing\nPEGs) doesn\u2019t tell you even at parse time if the input string is\nambiguous.  So with a Packrat-Parsing-based strategy, you are really\nflying blind about whether there are \u201cconceptual\u201d ambiguities in your\ngrammar.  It\u2019s also possible for entire alternatives or rules of a PEG\nto be unreachable (see <a href=\"http:\/\/article.gmane.org\/gmane.comp.parsers.peg.general\/2\">here<\/a>\nfor a bit more discussion of this).  The net result is that with PEGs\nyou know very little about the properties of your grammar.<\/p>\n<p>So far none of the options we\u2019ve discussed can actually help us find\nambiguities up-front. Surely there must be a way of analyzing our\ngrammar ahead of time and proving that it\u2019s not ambiguous, as long as\nthe grammar isn\u2019t too crazy?  Indeed there is, but the answer brings\nus back to where we started: LL and LR parsers.<\/p>\n<p>It turns out that simply trying to construct an LR parser for a\ngrammar is very nearly the most powerful ambiguity test we know of for\nCFGs.  We know it cannot be a perfect test, since we already stated\nthat testing ambiguity is undecidable.  If a grammar is not LR we\ndon\u2019t know whether it is ambiguous or not.  But every grammar that we\n<em>can<\/em> construct an LR parser for is guaranteed to be unambiguous, and\nas a bonus, you also get an efficient linear-time parser for it.<\/p>\n<p>But wait, there\u2019s more.  Let\u2019s review the three types of ambiguity\nwe\u2019ve encountered and the most natural solution for solving each:<\/p>\n<ol><li><em>Arithmetic expressions<\/em>: the ideal solution is to be able to\n<strong>declare precedence\/associativity directly<\/strong>, and <em>not<\/em> have to solve\nit at a grammar level.<\/li>\n  <li><em>Dangling else<\/em>: because this can be resolved by always preferring one\nalternative over another, the ideal solution is <strong>prioritized choice<\/strong>.\nAnother example of this case is C++\u2019s <a href=\"http:\/\/en.wikipedia.org\/wiki\/Most_vexing_parse\">\u201cmost\nvexing parse\u201d<\/a>, which is likewise resolved by simply preferring one\ninterpretation over the other when both are valid.<\/li>\n  <li><em>Type\/variable ambiguity<\/em>: the only real solution is to allow Turing-complete\n<strong>semantic predicates<\/strong> to resolve the ambiguity.  C++ has an even more\nextreme version of this problem since type\/variable disambiguation\ncould require arbitrary amounts of template instantiation, and therefore\njust <em>parsing<\/em> C++ is technically undecidable (!!), unless you limit\ntemplate instantiation depth \u2013 <a href=\"http:\/\/yosefk.com\/c++fqa\/web-vs-c++.html#misfeature-3\">more gory\ndetail and an example here<\/a>.<\/li>\n<\/ol><p>The good news is that all three of these ambiguity-resolution\nstrategies can be incorporated into a LL or LR-based parser generator,\nto some extent.  While <em>pure<\/em> LL or LR only support context-free\ngrammars, it is entirely possible to add operator precedence,\nprioritized choice, <em>and<\/em> semantic predicates to such tools, subject\nto some limitations.  I liken this to multi-paradigm programming\nlanguages.  Just as a language that supports procedural, OO, and\nfunctional styles is more powerful and expressive than a language\noffering just one, so is a CFG+precedence+predicates tool more\npowerful than one that only supports CFGs.<\/p>\n<p>We can sum up all of this information with this venn diagram:<\/p>\n<div style=\"text-align: center;\"><img src=\"https:\/\/docs.google.com\/drawings\/d\/1SATUGvZCcetW91Tnb40NybQ_zag_5WX6zIF41dG8WRM\/pub?w=621&amp;h=442\" alt=\"image\" \/><\/div>\n<p>To more fully understand how LL\/LR grammar analysis can prove grammars\nunambiguous (and how we can add non-CFG features like prioritized\nchoice to them), we will explore the concept of a deterministic\nparser.<\/p>\n<p>Without going into too much detail, a <em>deterministic<\/em> parser is one\nthat works by building a deterministic automaton (this is very much\nrelated to automata theory for regular expressions, except that\nparsing automata also have a stack).  That means that as the parser\nreads tokens left-to-right, it is always in one particular state, and\neach token transitions it into exactly one other state.<\/p>\n<p>More informally, we can say that a deterministic parser is one that\ndoesn\u2019t have to do any guessing or searching.  Terence Parr, author of\nANTLR, often uses the metaphor of parsing as a maze; at every fork in\nthe maze, a deterministic parser always knows which fork to take the\nfirst time.  It might \u201clook ahead\u201d to make the decision, but it never\nmakes a decision and then backs up (a parser that \u201cbacks up\u201d is known\nas a \u201cbacktracking parser\u201d and has exponential worst-case running\ntime).<\/p>\n<p>This determinism is really the defining characteristic that gives\nLL\/LR both their advantages and disadvantages.  They are the fastest\nalgorithms because they are simply transitioning a state machine.\nThey are not only fast, they are <em>predictably<\/em> fast: they have have\n<em>worst-case<\/em> O(n) performance.  Some approaches like NFAs\n(nondeterministic finite automata) or backtracking parsers have good\nperformance in common cases but can degrade severely (even\nexponentially) in degenerate cases.  Many popular regex engines have\nthis issue (see the article <a href=\"https:\/\/swtch.com\/~rsc\/regexp\/regexp1.html\">Regular Expression\nMatching Can Be Simple And Fast<\/a> for an example).  You might think\nthe bad cases are uncommon, but anyone who wants to DoS-attack your\nservice can find and exploit them.<\/p>\n<p>Besides being fast, you can know that an LL\/LR grammar is unambiguous\nbecause an ambiguous grammar won\u2019t allow you to build a deterministic\nautomaton.  To build the automaton you have to be able to prove, for\nevery grammar state and every input token, that there is only one\nvalid path through the grammar that you can take for this token.  A\nBison \u201cshift\/reduce\u201d or \u201creduce\/reduce\u201d conflict is a case where Bison\nwas not able to make the parser deterministic, because two state\ntransitions for the same token were both valid.  But Bison can\u2019t prove\nwhether this is because of grammar ambiguity (in which case both\ntransitions could ultimately lead to a successful parse) or whether\nthis is an unambiguous grammar that just isn\u2019t LR (in which case one\nof the paths would eventually hit a dead-end.<\/p>\n<p>\u201cGeneralized parsing\u201d algorithms like GLR and GLL can handle any\ngrammar, because they just take both paths simultaneously.  If one of\nthe paths hits a dead-end, then we\u2019re back to an unambiguous string.\nBut if multiple paths turn out to all be valid, we have an ambiguous\nstring and the parser can give you <em>all<\/em> of the valid parse trees.<\/p>\n<p>This also gives us a hint about how pure LL\/LR algorithms can be\nextended with extra features.  Any method we can think of for deciding\nwhich path is the \u201cright\u201d one is fair game!  It turns out that\noperator precedence declarations can provide a way of resolving\nshift\/reduce and reduce\/reduce conflicts in LR parsers, and <a href=\"http:\/\/www.gnu.org\/software\/bison\/manual\/html_node\/Precedence.html#Precedence\">Bison\nsupports this with its precedence features<\/a>.  Prioritized choice\ncan also give us enough information in some cases to resolve a\nnondeterminism by deciding that one of the paths is correct because it\nhas a higher priority.  And when all else fails, if we can write our\nown predicates that run at parse time, we can write arbitrary logic\nthat uses whatever other criteria we could possibly want to decide\nwhich path is the correct one.<\/p>\n<p>Given all of this, what is the answer to our original question?  Why\nare parsing tools hard?  I think there are two sets of reasons: some\n<em>inherent<\/em> reasons and some reasons that are just a weakness of\ncurrent parsing tools and could be improved.<\/p>\n<p>The <em>inherent<\/em> reasons parsing tools are hard have to do with\nambiguity.  More specifically:<\/p>\n<ol><li>The input grammar may be ambiguous but we can\u2019t robustly check for\nthis because it\u2019s undecidable.<\/li>\n  <li>We can use a deterministic (LL\/LR) parsing algorithm: this gives\nus a fast parser and a proof that the grammar is unambiguous.\nBut unfortunately no deterministic parsing algorithm can handle\nall unambiguous grammars. So in some cases we\u2019re forced to adapt\nour grammar and debug any nondeterminism.<\/li>\n  <li>We can use a generalized parsing algorithm like GLL or GLR that\ncan handle all grammars (even ambiguous ones), but because of\n(1) we can\u2019t know for sure whether the grammar is ambiguous or\nnot.  With this strategy we have to always be prepared to get\nmultiple valid parse trees at parse time.  If we didn\u2019t know\nabout the ambiguity, we probably won\u2019t know how we should\ndisambiguate.<\/li>\n  <li>We can use a formalism like Parsing Expression Grammars that\ndefines ambiguity away \u2013 this will always give us one unique\nparse tree, but can still hide \u201cconceptual\u201d ambiguities.<\/li>\n  <li>Some real-world ambiguities can\u2019t be resolved at a grammar-level\nbecause they have semantic context-sensitivity. To parse these\nlanguages, there <em>must<\/em> be a way of embedding arbitrary logic\ninto the parser to disambiguate.<\/li>\n<\/ol><p>While this may all seem annoying, ambiguity is a real language design\nissue, and anyone designing a language or implementing a parser\nbenefits from getting early warning about ambiguity.  In other words,\nwhile LL and LR tools are not perfect, some of their pain simply comes\nfrom the fact that <em>parsing<\/em> and <em>language design<\/em> are complicated.\nWhile rolling your own parser will free you from ever getting error\nmessages from your parser generator, it will also keep you from\nlearning about ambiguities you may be inadvertently designing into\nyour language.<\/p>\n<p>The other reasons parsing tools are hard are things that could\nrealistically be improved.  The tools could benefit from greater\nflexibility, more reusable grammars, better ability to compose\nlanguages, and more.  This is where the opportunity lies.<\/p>","created_at":"2018-11-07T16:49:35+0000","updated_at":"2018-11-07T16:49:35+0000","annotations":[],"mimetype":"text\/html","reading_time":23,"domain_name":"blog.reverberate.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":480,"title":"fdehau\/tui-rs","url":"https:\/\/github.com\/fdehau\/tui-rs","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/travis-ci.org\/fdehau\/tui-rs\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5d07853a9247bfc9071ab161b980a763f7a13c31\/68747470733a2f2f7472617669732d63692e6f72672f6664656861752f7475692d72732e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/fdehau\/tui-rs.svg?branch=master\" \/><\/a>\n<a href=\"https:\/\/crates.io\/crates\/tui\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/1524da60b8490a273001c53ed5691253a490d99a\/68747470733a2f2f696d672e736869656c64732e696f2f6372617465732f762f7475692e737667\" alt=\"Crate Status\" data-canonical-src=\"https:\/\/img.shields.io\/crates\/v\/tui.svg\" \/><\/a>\n<a href=\"https:\/\/docs.rs\/crate\/tui\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/590a26ae0157ffbd1b606e155af0e9bc4a05a777\/68747470733a2f2f646f63732e72732f7475692f62616467652e737667\" alt=\"Docs Status\" data-canonical-src=\"https:\/\/docs.rs\/tui\/badge.svg\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/assets\/demo.gif\"><img src=\"https:\/\/github.com\/fdehau\/tui-rs\/raw\/master\/assets\/demo.gif\" alt=\"Demo cast under Linux Termite with Inconsolata font 12pt\" \/><\/a><\/p>\n<p><code>tui-rs<\/code> is a <a href=\"https:\/\/www.rust-lang.org\" rel=\"nofollow\">Rust<\/a> library to build rich terminal\nuser interfaces and dashboards. It is heavily inspired by the <code>Javascript<\/code>\nlibrary <a href=\"https:\/\/github.com\/yaronn\/blessed-contrib\">blessed-contrib<\/a> and the\n<code>Go<\/code> library <a href=\"https:\/\/github.com\/gizak\/termui\">termui<\/a>.<\/p>\n<p>The library itself supports two different backends to draw to the terminal. You\ncan either choose from:<\/p>\n<p>However, some features may only be available in one of the two.<\/p>\n<p>The library is based on the principle of immediate rendering with intermediate\nbuffers. This means that at each new frame you should build all widgets that are\nsupposed to be part of the UI. While providing a great flexibility for rich and\ninteractive UI, this may introduce overhead for highly dynamic content. So, the\nimplementation try to minimize the number of ansi escapes sequences generated to\ndraw the updated UI. In practice, given the speed of <code>Rust<\/code> the overhead rather\ncomes from the terminal emulator than the library itself.<\/p>\n<p>Moreover, the library does not provide any input handling nor any event system and\nyou may rely on the previously cited libraries to achieve such features.<\/p>\n<h3><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a><a href=\"https:\/\/docs.rs\/tui\" rel=\"nofollow\">Documentation<\/a><\/h3>\n<h3><a id=\"user-content-demo\" class=\"anchor\" aria-hidden=\"true\" href=\"#demo\"><\/a>Demo<\/h3>\n<p>The <a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/demo.rs\">source code<\/a> of the demo gif.<\/p>\n<h3><a id=\"user-content-widgets\" class=\"anchor\" aria-hidden=\"true\" href=\"#widgets\"><\/a>Widgets<\/h3>\n<p>The library comes with the following list of widgets:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/block.rs\">Block<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/gauge.rs\">Gauge<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/sparkline.rs\">Sparkline<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/chart.rs\">Chart<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/bar_chart.rs\">BarChart<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/list.rs\">List<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/table.rs\">Table<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/paragraph.rs\">Paragraph<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/canvas.rs\">Canvas (with line, point cloud, map)<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/examples\/tabs.rs\">Tabs<\/a><\/li>\n<\/ul><p>Click on each item to see the source of the example. Run the examples with with\ncargo (e.g. to run the demo <code>cargo run --example demo<\/code>), and quit by pressing <code>q<\/code>.<\/p>\n<h3><a id=\"user-content-third-party-widgets\" class=\"anchor\" aria-hidden=\"true\" href=\"#third-party-widgets\"><\/a>Third-party widgets<\/h3>\n<h3><a id=\"user-content-alternatives\" class=\"anchor\" aria-hidden=\"true\" href=\"#alternatives\"><\/a>Alternatives<\/h3>\n<p>You might want to checkout <a href=\"https:\/\/github.com\/gyscos\/Cursive\">Cursive<\/a> for an\nalternative solution to build text user interfaces in Rust.<\/p>\n<p><a href=\"https:\/\/github.com\/fdehau\/tui-rs\/blob\/master\/LICENSE\">MIT<\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:49:34+0000","updated_at":"2018-11-07T16:49:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/10096255?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":481,"title":"Bash Shell Tricks","url":"https:\/\/dev.to\/bhaskar_vk\/bash-shell-tricks","content":"<p>This was originally published on my <a href=\"https:\/\/www.karambelkar.info\/2007\/06\/bash-shell-tricks\/\">blog<\/a> back in 2007, I am glad to see that this has aged well. I still find personally using many of the tricks mentioned here even today.<\/p><ul><li>  <code>^orig^repl^<\/code> replace 'orig' in previous command with 'repl' and execute the new command.<\/li>\n<\/ul><ul><li>  <code>cp filename{,.bak}<\/code> Copy 'filename' as 'filename.bak.<\/li>\n<li>  <code>mkdir -p dir{1,2,3}\/subdir{1,2}<\/code> Create <code>dir1\/subdir1<\/code>, <code>dir1\/subdir2<\/code>, <code>dir2\/subdir1<\/code>, <code>dir2\/subdir2<\/code>, <code>dir3\/subdir1<\/code>, and <code>dir3\/subdir2<\/code> directories<\/li>\n<\/ul><p>No I am not talking about completing command\/file names. I am talking about completing arguments to various commands, completing filenames based on application. This is one of the touted features of<br \/>'zsh', but unknown to many is the fact that it is also available in bash.<br \/>All you need to do is install <a href=\"http:\/\/www.caliban.org\/bash\/index.shtml#completion\">bash-completion<\/a>.<\/p>\n<p>With bash completion installed you can use the <code>TAB<\/code> key to complete arguments to common commands like <code>rpm, apt, find, grep<\/code> etc. Also bash-completion will complete host-names<br \/>for <code>ssh, scp<\/code>, by looking up hosts inside your <code>$HOME\/.ssh\/authorized_keys<\/code> file. For <code>rpm<\/code> based distros, bash-completion will even lookup package names already installed.<\/p>\n<p>The feature I find most handy is file-name completion based on the context of the command. e.g. if you type <code>tar -zxvf<\/code> and then press the TAB key twice, you will get a list of only files ending in <code>.tar.gz or .tgz<\/code> rather than a list of all files in the directory.<\/p>\n<p>I bet a lot of you haven't heard of this, but bash can indeed perform basic socket communication via <code>\/dev\/tcp, \/dev\/udp<\/code>. These are pseudo devices that bash uses to communicate with network sockets. In fact if you did <code>ls -l \/dev\/tcp \/dev\/udp<\/code> you will get a 'No such file or directory' error message.<\/p>\n<p>So how to use them, we below I present 2 examples.<\/p>\n<p>One to quickly check headers from a HTTP Server. Here is a simple function you can put in your <code>$HOME\/.bashrc<\/code> that will check for headers of HTTP server.<\/p>\n<div class=\"highlight\"><pre>headers() {\n  server=$1; port=${2:-80}\n  exec 5&lt;&gt; \/dev\/tcp\/$server\/$port\n  echo -e \"HEAD \/ HTTP\/1.0\\nHost: ${server}\\n\\n\" &gt;&amp;5;\n  cat 0&lt;&amp;5;\n  exec 5&gt;&amp;-\n}\n<\/pre><\/div>\n<p>Simply invoke it by<\/p>\n<p><code>headers &lt;servername or ip&gt; &lt;port&gt;<\/code><\/p>\n<p>The port number defaults to 80 if not provided.<\/p>\n<p>Second example is a quick way to check if a port is open or not. You can always use netcat or telnet, but I find this easier.<\/p>\n<div class=\"highlight\"><pre>testPort() {\n  server=$1; port=$2; proto=${3:-tcp}\n  exec 5&lt;&gt;\/dev\/$proto\/$server\/$port\n  (( $? == 0 )) &amp;&amp; exec 5&lt;&amp;-\n}\n<\/pre><\/div>\n<p>Again invoke it by<\/p>\n<p><code>testPort &lt;servername or ip&gt; &lt;port&gt; &lt;protocol&gt;<\/code><\/p>\n<p>The protocol can be either tcp or udp and defaults to tcp.<\/p>\n<p>Bash can perform arithmetic evaluations. They are much easier to using <code>expr<\/code> tool. Below are some examples.<em>Note the absense of $ prefix for variable names.<\/em><\/p>\n<div class=\"highlight\"><pre>((count++)) #increment value of variable 'count' by one.\n((total+=current)) # set total = total+current.\n((current&gt;max?max=current:max=max)) # ternary expression.\n<\/pre><\/div>","created_at":"2018-11-07T16:49:34+0000","updated_at":"2018-11-07T16:49:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"dev.to","preview_picture":"https:\/\/res.cloudinary.com\/practicaldev\/image\/fetch\/s--nKNrPpjw--\/c_imagga_scale,f_auto,fl_progressive,h_500,q_auto,w_1000\/https:\/\/static.pexels.com\/photos\/207580\/pexels-photo-207580.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":478,"title":"The Document Which Was Formerly Called The MIT Guide to Lockpicking","url":"http:\/\/www.blurofinsanity.com\/mit\/lockpick.html","content":"The Document Which Was Formerly Called The MIT Guide to Lockpicking\n<p style=\"text-align: CENTER;\"><span>This was \nsnagged by a new member of the Staff known as \"Cal.\" Where he found \nit doesn't matter. Just remember this is just something we found which is only \nintended for professional locksmiths, This site is only linking to this document \nfor educational purposes. Thank you.<\/span><\/p><hr \/><p style=\"text-align: LEFT;\"><b><br \/>MIT Guide \nto Lock Picking<\/b><\/p><p>Ted the Tool <br \/>February 14, 1992<\/p><h2>Distribution<\/h2><p>Copyright \n1987, 1991 Theodore T. Tool. All right reserved.<\/p><p>Permission to reproduce \nthis document on a non-profit basis is granted provided that this copyright and \ndistribution notice is included in full. The information in this booklet is provided \nfor educational purposes only.<\/p><p>August 1991 revision.<\/p><hr style=\"text-align: LEFT;\" \/><h2>Contents<\/h2><dl><dt><b>1<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter1.html\">It's Easy<\/a> <\/dt><dt><b>2<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter2.html\">How \na Key Opens a Lock<\/a> <\/dt><dt><b>3<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter3.html\">The Flatland Model<\/a> \n<\/dt><dt><b>4<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter4.html\">Basic Picking The Binding Defect<\/a> <\/dt><dt><b>5<\/b> \n<a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter5.html\">The Pin Column Model<\/a> <\/dt><dt><b>6<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter6.html\">Basic \nScrubbing<\/a> <\/dt><dt><b>7<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter7.html\">Advanced Lock Picking<\/a> <\/dt><dd><b>7.1<\/b> \nMechanical Skills<br \/><b>7.2<\/b> Zen and the Art of Lock Picking<br \/><b>7.3<\/b> \nAnalytic Thinking<br \/><\/dd><dt><b>8<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter8.html\">Exercises<\/a> <\/dt><dd><b>8.1<\/b> \nExercise 1: Bouncing the pick<br \/><b>8.2<\/b> Exercise 2: Picking Pressure<br \/><b>8.3<\/b> Exercise 3: Picking Torque<br \/><b>8.4<\/b> Exercise 4: Identifying Set \nPins<br \/><b>8.5<\/b> Exercise 5: Projections<br \/><\/dd><dt><b>9<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter9.html\">Recognizing \nand Exploiting Personality Traits<\/a> <\/dt><dd><b>9.1<\/b> Which Way To Turn<br \/><b>9.2<\/b> \nHow Far to Turn<br \/><b>9.3<\/b> Gravity<br \/><b>9.4<\/b> Pins Not Setting<br \/><b>9.5<\/b> \nElastic Deformation<br \/><b>9.6<\/b> Loose Plug<br \/><b>9.7<\/b> Pin Diameter<br \/><b>9.8<\/b> Beveled Holes and Rounded pins<br \/><b>9.9<\/b> Mushroom Driver Pins<br \/><b>9.10<\/b> Which Way To Turn<br \/><b>9.11<\/b> Which Way To Turn<br \/><b>9.12<\/b> \nWhich Way To Turn<br \/><b>9.13<\/b> Disk Tumblers<br \/><\/dd><dt><b>10<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter0.html\">Final \nRemarks<\/a> <\/dt><dt><b>A<\/b> <a href=\"http:\/\/www.blurofinsanity.com\/mit\/appendix.html\">Tools<\/a> <\/dt><dd><b>A.1<\/b> Pick \nShapes<br \/><b>A.2<\/b> Street cleaner bristles<br \/><b>A.3<\/b> Bicycle spokes<br \/><b>A.4<\/b> Brick Strap<br \/><\/dd><dt><b>B<\/b> Legal Issues <\/dt><\/dl><hr style=\"text-align: LEFT;\" \/><p><b>[ \n<a href=\"http:\/\/www.blurofinsanity.com\/mit\/chapter1.html\">Next<\/a> ] - [ <a href=\"http:\/\/www.blurofinsanity.com\/homeofinsanity02.html#Anchor-27505\">Home<\/a> \n]<\/b> <\/p><p><a href=\"http:\/\/www.blurofinsanity.com\/homeofinsanity02.html#Anchor-27505\"><img src=\"http:\/\/www.blurofinsanity.com\/blurimages\/returnhome.jpg\" width=\"212\" height=\"90\" border=\"0\" alt=\"image\" style=\"float: left;\" \/><\/a> \n<\/p>","created_at":"2018-11-07T16:49:33+0000","updated_at":"2018-11-07T16:49:33+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.blurofinsanity.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":479,"title":"4D Toys: a box of four-dimensional toys","url":"http:\/\/marctenbosch.com\/news\/2017\/06\/4d-toys-a-box-of-four-dimensional-toys\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:49:33+0000","updated_at":"2018-11-07T16:49:33+0000","annotations":[],"reading_time":0,"domain_name":"marctenbosch.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":477,"title":"HTTP Error 403","url":"http:\/\/www.gridsagegames.com\/blog\/2014\/06\/procedural-map-generation\/","content":"HTTP Error 403\n<p>We're sorry, but we could not fulfill your request for\n\/blog\/2014\/06\/procedural-map-generation\/ on this server.<\/p>\n<p>An invalid request was received from your browser. This may be caused by a malfunctioning proxy server or browser privacy software.<\/p>\n<p>Your technical support key is: <strong>68ec-57d0-1756-6707<\/strong><\/p>\n<p>You can use this key to <a href=\"http:\/\/www.ioerror.us\/bb2-support-key?key=68ec-57d0-1756-6707\">fix this problem yourself<\/a>.<\/p>\n<p>If you are unable to fix the problem yourself, please contact <a href=\"mailto:gridsagegames+nospam@nospam.gmail.com\">gridsagegames at gmail.com<\/a> and be sure to provide the technical support key shown above.<\/p>","created_at":"2018-11-07T16:49:32+0000","updated_at":"2018-11-07T16:49:32+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.gridsagegames.com","http_status":"403"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":476,"title":"Build Your Own Text Editor","url":"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/","content":"<p>Welcome! This is an instruction booklet that shows you how to build a text\neditor in C.<\/p><p>The text editor is <a href=\"http:\/\/antirez.com\/news\/108\">antirez\u2019s kilo<\/a>, with some\nchanges. It\u2019s about 1000 lines of C in a single file with no dependencies, and\nit implements all the basic features you expect in a minimal editor, as well as\nsyntax highlighting and a search feature.<\/p><p>This booklet walks you through building the editor in <strong>184 steps<\/strong>. Each step,\nyou\u2019ll add, change, or remove a few lines of code. Most steps, you\u2019ll be able\nto <strong>observe the changes<\/strong> you made by compiling and running the program\nimmediately afterwards.<\/p><p>I explain each step along the way, sometimes in a lot of detail. Feel free to\nskim or skip the prose, as the main point of this is that <strong>you are going to\nbuild a text editor from scratch<\/strong>! Anything you learn along the way is bonus,\nand there\u2019s plenty to learn just from typing in the changes to the code and\nobserving the results.<\/p><p>See the <a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/08.appendices.html\">appendices<\/a> for more information on the tutorial\nitself (including what to do if you get stuck, and where to get help).<\/p><p>If you\u2019re ready to begin, then go to <a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/01.setup.html\">chapter 1<\/a>!<\/p><ol><li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/01.setup.html\">Setup<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/02.enteringRawMode.html\">Entering raw mode<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/03.rawInputAndOutput.html\">Raw input and output<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/04.aTextViewer.html\">A text viewer<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/05.aTextEditor.html\">A text editor<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/06.search.html\">Search<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/07.syntaxHighlighting.html\">Syntax highlighting<\/a><\/li>\n<li><a href=\"https:\/\/viewsourcecode.org\/snaptoken\/kilo\/08.appendices.html\">Appendices<\/a><\/li>\n<\/ol>","created_at":"2018-11-07T16:49:30+0000","updated_at":"2018-11-07T16:49:30+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"viewsourcecode.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":474,"title":"Boston Startup Tech Guide","url":"http:\/\/bostontechguide.com\/","content":"<section id=\"home-3\" class=\"parallax parallax-image dark-bg\"><div class=\"container\"><div class=\"parallax-content\"><div class=\"row\"><div class=\"main-logo test\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2014\/02\/download.png\" alt=\"download\" width=\"702\" height=\"154\" class=\"aligncenter size-full wp-image-367\" \/><\/div><div class=\"powered\"><a href=\"http:\/\/nextviewventures.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/created-by-nextview.png\" alt=\"download (1)\" width=\"300\" height=\"46\" class=\"aligncenter size-medium wp-image-368\" \/><\/a><\/div><\/div><\/div><\/div><\/section><section id=\"welcome\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><div class=\"col-md-8\"><h2 class=\"test\">&#13;\n\t\t\t\t\t\t<strong>Boston is home to a thriving web and mobile startup ecosystem.<\/strong>&#13;\n\t\t\t\t\t<\/h2><p>The intellectual power here is staggering, with Harvard, MIT, and dozens of other schools making this the academic capital of the world.<\/p><p>For years, Boston has been at or near the top\u00a0of the list in\u00a0venture capital invested per capita in\u00a0the US. Meanwhile, large groups of tech startups, once\u00a0only found\u00a0out on Route 128 during Web 1.0 days, are now launching\u00a0in centralized areas like Kendall Square, Downtown Crossing, and Seaport, infusing the city with the highest level of entrepreneurial energy in recent memory.<\/p><p>But Boston can feel like an insular town, and with something happening every day in the tech community, it\u2019s only getting tougher to navigate the startup scene.<\/p><p>So whether you\u2019re taking classes, building a business, raising capital, or just trying to get plugged in, let this be your jumping-off point, your front door to the community.<\/p><p><strong>YOUR HITCHHIKER\u2019S GUIDE TO BOSTON TECH. WELCOME!<\/strong><\/p><p><a href=\"http:\/\/bostontechguide.com\/letter-from-the-creators\">Read a letter<\/a> from the creators, or scroll down to keep exploring.<\/p><\/div><div class=\"col-md-4\"><p>&#13;\n\t\t\t\t\t\tCREATED BY&#13;\n\t\t\t\t\t<\/p><div class=\"flex-wrap\"><div class=\"col-md-6 creator\"><img class=\"aligncenter size-thumbnail wp-image-382\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/rob-go.png\" alt=\"Rob Go\" \/>Rob Go<br \/>NextView<br \/><a href=\"https:\/\/twitter.com\/ROBGO\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t\t\t@ROBGO<\/a><\/div><div class=\"col-md-6 creator\"><img class=\"aligncenter size-thumbnail wp-image-382\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jay-acunzo-2015-close-crop-copy-1.jpeg\" alt=\"Jay Acunzo\" \/>Jay Acunzo<br \/>NextView<br \/><a href=\"https:\/\/twitter.com\/JAYACUNZO\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t\t\t@JAYACUNZO<\/a><\/div><div class=\"col-md-6 creator\"><img class=\"aligncenter size-thumbnail wp-image-382\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/keith-frankel.png\" alt=\"Keith Frankel\" \/>Keith Frankel<br \/>Site UX<br \/><a href=\"https:\/\/twitter.com\/thekeithf\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t\t\t@thekeithf<\/a><\/div><div class=\"col-md-6 creator\"><img class=\"aligncenter size-thumbnail wp-image-382\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/ariel-simon.jpg\" alt=\"Ariel Simon\" \/>Ariel Simon<br \/>Site Design<br \/><a href=\"https:\/\/twitter.com\/arielsimonsays\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t\t\t@arielsimonsays<\/a><\/div><\/div><\/div><div class=\"col-md-12 credit\"><p><em>This guide was created\u00a0by <a href=\"http:\/\/nextviewventures.com\/\" target=\"_blank\">NextView<\/a>: high conviction, hands-on seed investors.<\/em><\/p><\/div><\/div><\/div><\/section><section id=\"get-in\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><p><img class=\"size-full wp-image-404 aligncenter\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/Get-In.png\" alt=\"Get+In\" width=\"1000\" height=\"363\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/Get-In.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/Get-In-300x109.png 300w\" \/><img class=\"aligncenter size-full wp-image-407\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/ping.png\" alt=\"ping\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/ping.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/ping-300x12.png 300w\" \/><\/p><p>Concentrated groups of tech startups call these neighborhoods home, making them ideal places to launch new ventures, attend events, or post up at bars and coffee shops.<\/p><p><img class=\"aligncenter size-full wp-image-412\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-map.png\" alt=\"boston+startup+map\" width=\"1000\" height=\"618\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-map.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-map-300x185.png 300w\" \/><\/p><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/inno.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/inno.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>This 1,000 acre area of the South Boston waterfront has recently been transformed into a go-to entrepreneurial neighborhood.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/area2-compressor1.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/area2-compressor1.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>A small group of startups inhabit this neighborhood, which contains Boston\u2019s Little Italy and, on the outskirts, the TD Garden.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-41.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-41.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>Home to City Hall, Faneuil Hall, and plenty of events in the plaza. A small but growing cluster of startups work here, including local tech media BostInno.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-51.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-51.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>Its central location has turned the Financial District into a major hotspot for both entrepreneurs and VCs in recent years.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-91.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-91.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>With similar benefits to the Financial District, this area is home to two dozen tech startups, plus TechStars Boston, Startup Institute, Uber\u2019s local office, and NextView Ventures.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-6.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-6.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>Offices here may go for a premium, but more tech companies are starting to move to this upscale neighborhood (much to the chagrin of their CFOs).<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-7.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-7.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>If there\u2019s a Boston Tech Ground Zero, this is it. Kendall boasts MIT,  the CIC, dozens of startups, and tech giants like Google, Microsoft, and Twitter.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-81.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-81.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>It seems like every year, both Harvard Business School and Harvard College produce a crop of successful young companies. The area\u2019s coffee shops are also frequented by VCs and founders alike.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/area9-compressor.png\" alt=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/area9-compressor.png\" width=\"241\" height=\"113\" class=\"aligncenter size-full\" \/><div class=\"text\"><p>Just two stops from Harvard on the Red Line is Davis Square, a growing focal point of entrepreneurship. Overall, Somerville startups are becoming more common.<\/p><\/div><\/div><\/div><p><img class=\"aligncenter size-full wp-image-459\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/pillar-boston-startups.png\" alt=\"pillar+boston+startups\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/pillar-boston-startups.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/pillar-boston-startups-300x12.png 300w\" \/><\/p><p>These independent companies have established themselves locally, nationally, and even globally through a combination of revenue growth, valuation, and brand awareness. More are joining this list every year.<\/p><div class=\"flex-wrap\"><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/acquia.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/acquia.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBurlington\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/actifio.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/actifio.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tWaltham\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.akamai.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/akamai.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/applause.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/applause.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFramingham\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/bookbub.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bookbub.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/brightcove.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/brightcove.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFinancial District, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/bullhorn.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bullhorn.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tInnovation District, South Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/cargurus.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cargurus.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tEast Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.carbonblack.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/carbon-black.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tLeather District, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.carbonite.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/carbonite.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBack Bay, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/care.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/care-dot-com.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tWaltham\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/gocatalant.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/catalant-logo.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFort Point, South Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/constantcontact.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/constant-contact-1.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tWaltham\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.cyberark.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cyberark.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tNeedham\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/dataxu.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dataxu.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tInnovation District, South Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/demandware.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/demandware.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBurlington\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.digitallumens.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/digital-lumens.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tSeaport, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/draftkings.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/draftkings-1.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tDowntown Crossing, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/drift.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/drift.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tEast Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/drizly.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/drizly.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBack Bay, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/formlabs.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/formlabs2.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tSomerville\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.thegrommet.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/grommet.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tDavis Square, Somerville\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.hopper.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hopper.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/hubspot.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hubspot.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tEast Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/insightsquared.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/insightsquared.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.irobot.com\/us\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/irobot.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBedford\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.jana.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jana.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tDowntown Crossing, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.kensho.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Kensho1.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tHarvard Square, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.localytics.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/localytics.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFinancial District, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/logmein.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/logmein.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tInnovation District, South Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.lolatravel.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/lola.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tSeaport, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/www.loseit.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/lose-it.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tSeaport, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/mimecast.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mimecast.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tWatertown\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/nanigans.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/nanigans.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFinancial District, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/onshape.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/onshape.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tAlewife, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/pillpack.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/pillpack.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tSomerville\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/ruelala.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ruelala.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tInnovation District, South Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.semanticmachines.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/semantic-machines.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tNewton\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/simplisafe.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/simplisafe.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tCambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/simplivity.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/simplivity.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tWestborough\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.tripadvisor.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/tripadvisor.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tNewton\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"https:\/\/upserve.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/upserve.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tProvidence, RI + Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/veracode.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/veracode.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBurlington\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/www.visiblemeasures.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/visible-measures.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tLeather District, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/wayfair.com\/\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wayfair.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBack Bay, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/wistia.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wistia.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tCentral Square, Cambridge\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><a href=\"http:\/\/wordstream.com\" target=\"_blank\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wordstream.png\" width=\"466\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tBack Bay, Boston\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><\/div><p><img class=\"aligncenter size-full wp-image-501\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tech-giants.png\" alt=\"tech+giants\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tech-giants.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tech-giants-300x12.png 300w\" \/><\/p><p>The Boston area\u2019s globally recognized collection of young talent, combined with its resurgent entrepreneurial drive, has attracted top-tier tech companies such as the following:<\/p><div class=\"flex-wrap\"><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amazon.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/apple.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cisco.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t(Multiple Locations)\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dell-emc-1.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/facebook.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ge.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tSeaport, Boston\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/google.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ibm.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tEast Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/microsoft-1.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/oracle.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t(Multiple Locations)\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/paypal.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tFinancial District, Boston\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/twitter.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tKendall Square, Cambridge\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/uber.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tNorth Station, Boston\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/spotify-logo-final.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\t\tDavis Square, Somerville\t\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><p><img class=\"aligncenter size-full wp-image-517\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-media.png\" alt=\"boston+tech+media\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-media.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-media-300x12.png 300w\" \/><\/p><p>As a city, Boston is influenced by its rich history and its modern innovation. This extends to the local media, where both established publications and new, digital-first outlets cover the tech startup world.<\/p><div class=\"flex-wrap\"><div class=\"col-md-6\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/betaboston.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The Boston Globe\u2019s standalone tech blog covers local startups in web, mobile, biotech, robotics, healthcare, and other sectors. The site launched in early 2014.<\/p><a target=\"_blank\" href=\"http:\/\/betaboston.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-6\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostinno.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Operating under parent brand Streetwise Media (and owned by American City Business Journals since its 2012 acquisition), BostInno provides tech news and several pillar events. Their video update, The Beat, is must-watch material for its breaking news and sarcastic, entertaining coverage.<\/p><a target=\"_blank\" href=\"http:\/\/bostinno.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-6\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bbj.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The local arm of American City Business Journals provides regular tech coverage and is perhaps best known for its annual list of the best places to work in Boston.<\/p><a target=\"_blank\" href=\"http:\/\/www.bizjournals.com\/boston\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-6\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/xconomy.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Like the BBJ, this is a local version of a national media outlet. Xconomy\u2019s Boston property also organizes regular events for the tech community.<\/p><a target=\"_blank\" href=\"http:\/\/www.xconomy.com\/boston\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><\/div><p><img class=\"aligncenter size-full wp-image-524\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tech-resources.png\" alt=\"tech+resources\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tech-resources.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tech-resources-300x12.png 300w\" \/><\/p><p>Anyone hoping to break into Boston tech or simply keep up with all the happenings around the ecosystem can rely on these resources.<\/p><div class=\"flex-wrap\"><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bos-startups-guide.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Created by <a href=\"https:\/\/twitter.com\/JayNeely\" target=\"_blank\" rel=\"noopener noreferrer\">Jay Neely<\/a>, this site offers a curated list of Boston startups, along with brief descriptions and job openings.<\/p><a target=\"_blank\" href=\"http:\/\/bostonstartupsguide.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/boston-tweetup.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Run by <a href=\"https:\/\/twitter.com\/JoselinMane\" target=\"_blank\" rel=\"noopener noreferrer\">Joselin Mane<\/a>, the site offers a comprehensive events calendar and other information about networking in Boston. Be sure to follow\u00a0<a href=\"http:\/\/twitter.com\/bostontweetup\" target=\"_blank\" rel=\"noopener noreferrer\">@BostonTweetup<\/a>\u00a0on Twitter.<\/p><a target=\"_blank\" href=\"http:\/\/bostontweetup.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/tcn.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>TCN brings together the area\u2019s largest active community of investors and startup experts to provide founders with education and mentoring to navigate the fundraising process.<\/p><a target=\"_blank\" href=\"http:\/\/www.thecapitalnetwork.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/greenhorn-connect.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>ACF A longstanding resource for Boston entrepreneurs and students which was originally launched by entrepreneur\u00a0<a href=\"http:\/\/twitter.com\/evanish\" target=\"_blank\" rel=\"noopener noreferrer\">Jason Evanish<\/a>.\u00a0The Capital Network recently took over daily operations and provides events, newsletters, and more.<\/p><a target=\"_blank\" href=\"http:\/\/greenhornconnect.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/nv-platform.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>NextView Ventures owns\u00a0a growing library of resources dedicated to challenges facing early-stage startup founders and teams, including guides, advice columns, startup stories, and a podcast.<\/p><a target=\"_blank\" href=\"http:\/\/nextviewventures.com\/platform\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/opv-labs.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>OpenView Venture Partners also runs Labs, where they share advice and resources for software startups and teams, primarily focused on B2B SaaS.<\/p><a target=\"_blank\" href=\"http:\/\/labs.openviewpartners.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/tech-in-bos-medium.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Created by <a href=\"http:\/\/twitter.com\/davegerhardt\" target=\"_blank\" rel=\"noopener noreferrer\">Dave Gerhardt<\/a> as a podcast, Tech in Boston expanded in 2016 to include a curated list of articles written by Boston\u2019s tech community on Medium.<\/p><a target=\"_blank\" href=\"https:\/\/medium.com\/tech-in-boston\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/techscene-bos.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>An interactive map showing the locations of Boston\u2019s startups, as well as funding raised and jobs created.<\/p><a target=\"_blank\" href=\"http:\/\/techscene.at\/boston\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/venturefizz.png\" width=\"932\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Run by\u00a0<a href=\"https:\/\/twitter.com\/kcline6\" target=\"_blank\" rel=\"noopener noreferrer\">Keith Cline<\/a>, a top local recruiter,\u00a0VentureFizz provides both original and syndicated content, as well as several other products like BizzPages, event listings, and a dedicated Boston tech job board. Their\u00a0<a href=\"http:\/\/venturefizz.com\/deals\" target=\"_blank\" rel=\"noopener noreferrer\">running log<\/a>\u00a0of local funding and acquisition announcements is a must-read.<\/p><a target=\"_blank\" href=\"http:\/\/venturefizz.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><\/div><div class=\"col-md-12 podcast\"><p><strong><a id=\"yui_3_17_2_1_1425227896053_6109\" href=\"http:\/\/list.ly\/list\/RxO-online-resources-hitchhikers-guide-to-boston-tech\" target=\"_blank\" rel=\"noopener noreferrer\" data-cke-saved-href=\"http:\/\/list.ly\/list\/RxO-online-resources-hitchhikers-guide-to-boston-tech\">Submit more &amp; view what the community added &gt;&gt;<\/a><\/strong><\/p><p><img class=\"aligncenter size-full wp-image-535\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-podcasts-1.png\" alt=\"boston+podcasts (1)\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-podcasts-1.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-podcasts-1-300x12.png 300w\" \/><\/p><p>In the past few years, multiple podcasts have emerged discussing Boston tech, as well as the broader startup, digital media, and business landscapes.<\/p><div class=\"flex-wrap\"><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-03-at-4.55.31-PM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>HubSpot executives interview others about how they grew their businesses.<\/p><a target=\"_blank\" href=\"http:\/\/www.hubspot.com\/podcast\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-03-at-4.58.07-PM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Spark Capital\u2019s <a href=\"http:\/\/twitter.com\/bijan\" target=\"_blank\" rel=\"noopener noreferrer\">Bijan Sabet<\/a> and <a href=\"http:\/\/twitter.com\/nabeel\" target=\"_blank\" rel=\"noopener noreferrer\">Nabeel Hyatt<\/a> discuss trends in tech startups and VC in their periodic show, Hallway Chats.<\/p><a target=\"_blank\" href=\"https:\/\/itunes.apple.com\/us\/podcast\/bijan\/id579314510?mt=2\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-03-at-4.56.39-PM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>InsightSquared interviews software executives and thought leaders about how they\u2019re using data and technology to accelerate their company growth.<\/p><a target=\"_blank\" href=\"http:\/\/www.insightsquared.com\/business-analytics-podcast\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-07-at-10.21.19-AM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Drift CEO and serial entrepreneur <a href=\"http:\/\/twitter.com\/davidcancel\" target=\"_blank\" rel=\"noopener noreferrer\">David Cancel<\/a>\u00a0and Drift\u2019s <a href=\"http:\/\/twitter.com\/davegerhardt\" target=\"_blank\" rel=\"noopener noreferrer\">Dave Gerhardt<\/a>\u00a0tease out key lessons from Cancel\u2019s experience as a founder and rabid\u00a0learning approach.<\/p><a target=\"_blank\" href=\"https:\/\/soundcloud.com\/seekingwisdom\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-07-at-10.21.29-AM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Hosted by\u00a0<a href=\"https:\/\/twitter.com\/davegerhardt\" target=\"_blank\" rel=\"noopener noreferrer\">Dave Gerhardt<\/a>, who interviews Boston entrepreneurs about product, sales, marketing, and more.<\/p><a target=\"_blank\" href=\"http:\/\/techinboston.co\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-07-at-10.21.42-AM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Hosted by\u00a0<a href=\"http:\/\/twitter.com\/jay_zo\" target=\"_blank\" rel=\"noopener noreferrer\">Jay Acunzo<\/a>, the show shares clever, creative, unusual, and scrappy ways that founders make early progress. The show is produced as a narrative story.<\/p><a target=\"_blank\" href=\"http:\/\/nextviewventures.com\/blog\/category\/podcast\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-14-at-11.10.23-AM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Hosted by Founder Collective\u2019s <a href=\"https:\/\/twitter.com\/micahjay1\" target=\"_blank\" rel=\"noopener noreferrer\">Micah Rosenbloom<\/a>, the show interviews founders and investors about their companies and larger tech trends.<\/p><a target=\"_blank\" href=\"https:\/\/itunes.apple.com\/us\/podcast\/collective-wisdom-for-tech\/id999395945?mt=2\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-14-at-11.14.33-AM.png\" width=\"450\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Innovation Hub looks at how to reinvent our world \u2013 from medicine to education, relationships to time management and more.<\/p><a target=\"_blank\" href=\"https:\/\/itunes.apple.com\/us\/podcast\/innovation-hub\/id517249700?mt=2\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><\/div><\/div><p>&#13;\n\t\t\t\t\t<a target=\"_blank\" href=\"http:\/\/list.ly\/list\/RxV-podcasts-hitchhikers-guide-to-boston-tech\" id=\"yui_3_17_2_1_1425227896053_6480\"><strong>Submit more &amp; view what the community added &gt;&gt;<\/strong><\/a>&#13;<\/p><\/div><\/div><section id=\"get-educated\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><p><img class=\"aligncenter size-full wp-image-542\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-educated.png\" alt=\"get+educated\" width=\"954\" height=\"347\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-educated.png 954w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-educated-300x109.png 300w\" \/><\/p><p><img class=\"aligncenter size-full wp-image-544\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/classes-for-startups.png\" alt=\"classes+for+startups\" width=\"795\" height=\"32\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/classes-for-startups.png 795w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/classes-for-startups-300x12.png 300w\" \/><\/p><p>Education is a huge part of Boston\u2019s DNA, so it\u2019s no surprise that several organizations have launched here to teach entrepreneurs, developers, and other business leaders. Some have stayed local, others have expanded nationally or even globally, and several outside organizations have also moved to the area.<\/p><div class=\"flex-wrap\"><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/si-bos.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Now a global organization, its original location sits in the Leather District. Their 8-week program prepares you to change career paths and\/or work for a startup by teaching both job skills and softer skills.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/startupinstitute.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/intel.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Workshops and classes to train your company\u2019s management or develop your own leadership skills.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/intelligent.ly\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ga.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>With 12 campuses around the world (including Boston\u2019s Fort Point), General Assembly is a global educational institution that provides practical, hands-on learning in technology, business, and design through full-time and part-time training programs.<\/p>\u00a0<a target=\"_blank\" href=\"https:\/\/generalassemb.ly\/boston\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/la.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>This Boston-based program teaches students how to code to launch their careers as developers, and they offer local tech companies a pipeline of qualified tech talent through their career days.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/launchacademy.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/startup-leadership.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Another original chapter of an organization that has since expanded, this program helps train and support startup CEOs.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.startupleadership.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/girldevelopit.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The Boston chapter of this group hosts regular meetups and various courses for women of all backgrounds to learn how to develop software.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.girldevelopit.com\/chapters\/boston\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/girls-code.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>This popular organization and movement offers summer programs in Twitter\u2019s Kendall Square office.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/girlswhocode.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/railsbridge.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Dedicated to teaching women how to program in Ruby and Ruby on Rails.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.railsbridgeboston.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/product.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Several of Boston\u2019s companies offer apprenticeships, mainly in product development, design, and UX. Two popular programs include Fresh Tilled Soil\u2019s <a href=\"http:\/\/www.freshtilledsoil.com\/aux\/\" target=\"_blank\">Apprentice in UX Design<\/a> and Thoughtbot\u2019s <a href=\"http:\/\/www.apprentice.io\/\" target=\"_blank\">Apprentice.io<\/a>.<\/p><\/div><\/div><\/div><\/div><div class=\"col-md-12\"><p><strong id=\"yui_3_17_2_1_1425227896053_7247\"><a id=\"yui_3_17_2_1_1425227896053_7246\" href=\"http:\/\/list.ly\/list\/RxW-classes-for-entrepreneurs-hitchhikers-guide-to-boston-tech\" target=\"_blank\" data-cke-saved-href=\"http:\/\/list.ly\/list\/RxW-classes-for-entrepreneurs-hitchhikers-guide-to-boston-tech\">Submit more &amp; view what the community added &gt;&gt;<\/a><\/strong><\/p><p><img class=\"aligncenter size-full wp-image-562\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/univ-resources.png\" alt=\"univ+resources\" width=\"795\" height=\"32\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/univ-resources.png 795w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/univ-resources-300x12.png 300w\" \/><\/p><p>Aside from student clubs (not shown), several more formal organizations, resources, events, and competitions exist to support student entrepreneurship both on campus and around Boston.<\/p><p><strong>DISCLAIMER:<\/strong><\/p><p>Boston is home to 70+ colleges and universities, so this section alone could be its very own Hitchhiker\u2019s Guide. With that in mind, we\u2019d ask you and others around the community to submit new resources and\/or schools via the link below this section.<\/p><\/div><div class=\"flex-wrap\"><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/campus-recruitment-guide.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The New England Venture Capital Association (NEVCA) offers hiring managers a guide to recruit college students. The goal is to keep Massachusetts-based talent in the state.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.hiremastudents.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/roughdraft.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Rough Draft Ventures is a student partnership that invests in Boston area student entrepreneurs, created by VC firm General Catalyst.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/roughdraft.vc\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/drf.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Boston is one of four regional branches of this student-run venture firm, which invests in other students and is backed by First Round Capital.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/dormroomfund.com\/boston\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/startlabs.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>StartLabs is a non-profit that helps support and grow student entrepreneurship.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.startlabs.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/harvard.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>1.\u00a0<a href=\"http:\/\/i-lab.harvard.edu\/\" target=\"_blank\">Harvard Innovation Lab<\/a><\/p><p>2.\u00a0<a href=\"http:\/\/www.hbs.edu\/entrepreneurship\/mbacurriculum\/rock-accelerator.html\" target=\"_blank\">Harvard\u2019s Rock Accelerator<\/a><\/p><p>3.\u00a0<a href=\"https:\/\/sites.google.com\/site\/startuptribe\/\" target=\"_blank\">HBS Startup Tribe<\/a><\/p><p>4.\u00a0<a href=\"http:\/\/www.hbs.edu\/newventurecompetition\/Pages\/default.aspx\" target=\"_blank\">HBS New Venture Competition<\/a><\/p><p>5.\u00a0<a href=\"http:\/\/www.hackharvard.org\/\" target=\"_blank\">Hack Harvard<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/MIT.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>1.\u00a0<a href=\"https:\/\/entrepreneurship.mit.edu\/\" target=\"_blank\">Martin Trust Center for MIT Entrepreneurship<\/a><\/p><p>2.\u00a0<a href=\"https:\/\/entrepreneurship.mit.edu\/accelerator\/\" target=\"_blank\">MIT Global Founders\u2019 Skill Accelerator<\/a><\/p><p>3.\u00a0<a href=\"http:\/\/www.mitef.org\/s\/1314\/main.aspx?gid=5&amp;pgid=61\" target=\"_blank\">MIT Enterprise Forum<\/a><\/p><p>4.\u00a0<a href=\"http:\/\/www.mit100k.org\/\" target=\"_blank\">MIT $100K Competition<\/a><\/p><p>5.\u00a0<a href=\"http:\/\/vms.mit.edu\/\" target=\"_blank\">MIT Venture Mentoring Service<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/babson.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>1.\u00a0<a href=\"http:\/\/www.babson.edu\/Academics\/centers\/blank-center\/Pages\/home.aspx\" target=\"_blank\">Arthur M. Blank Center for Entrepreneurship<\/a><\/p><p>2.\u00a0<a href=\"http:\/\/www.babson.edu\/Academics\/centers\/blank-center\/venture-accelerator\/beta-challenge\/Pages\/home.aspx\" target=\"_blank\">B.E.T.A. Challenge<\/a><\/p><p>3.\u00a0<a href=\"http:\/\/www.babson.edu\/Academics\/centers\/blank-center\/venture-accelerator\/Pages\/resource-center.aspx\" target=\"_blank\">John E. and Alice L. Butler Venture Accelerator<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download4.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>1.\u00a0<a href=\"http:\/\/www.northeastern.edu\/idea\/\" target=\"_blank\">Northeastern\u2019s IDEA Venture Accelerator<\/a><\/p><p>2.\u00a0<a href=\"http:\/\/www.northeastern.edu\/scout\/\" target=\"_blank\">Scout: Northeastern\u2019s Student-Led Design Studio<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/olin.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p id=\"yui_3_17_2_1_1425312794579_6854\">1.\u00a0<a href=\"http:\/\/olinfoundry.org\/\" target=\"_blank\">The Olin Foundry<\/a><\/p><p>2.\u00a0<a href=\"http:\/\/www.olin.edu\/collaborate\/innovation-lab\/\" target=\"_blank\">Olin Innovation Lab<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/BC.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p><a href=\"http:\/\/bcvc.org\/\" target=\"_blank\">Boston College Venture Competition<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/umass-bos.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p><a href=\"http:\/\/www.umb.edu\/vdc\" target=\"_blank\">Venture Development Center at UMass Boston<\/a><\/p><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/berklee.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p><a href=\"http:\/\/www.berklee.edu\/focused\/ice\" target=\"_blank\">Berklee Institute for Creative Entrepreneurship<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"col-md-12\"><p><a href=\"http:\/\/list.ly\/list\/RxX-university-resources-hitchhikers-guide-to-boston-tech\" target=\"_blank\"><strong>Submit more &amp; view what the community added &gt;&gt;<\/strong><\/a><\/p><\/div><\/div><\/div><\/section><section id=\"get-connected\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><div class=\"col-md-12\"><p><img class=\"aligncenter wp-post-image\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-connected.png\" alt=\"get+connected\" width=\"954\" height=\"347\" \/><\/p><p><img class=\"aligncenter size-full wp-image-570\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/twitter.png\" alt=\"twitter\" width=\"795\" height=\"32\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/twitter.png 795w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/twitter-300x12.png 300w\" \/><\/p><p>Following people from Boston\u2019s tech startup scene is an easy way to learn about the ecosystem from afar \u2026 or dive headfirst into the conversation. It can also be a friendly, human way to network and get offline meetings. The following group of founders, VCs, and community builders are both active and widely followed around Boston (though many more exist that you should get to know).<\/p><p><strong>DISCLAIMER:<\/strong><\/p><h3><strong>In case you were tempted to read into the list as a ranking \u2014 it\u2019s not!<\/strong><\/h3><p>It\u2019s just based on total Twitter followers, and we have a much longer, more complete <a href=\"https:\/\/twitter.com\/NextViewVC\/lists\/hitchhiker-s-guide-to-bos\" target=\"_blank\">Twitter list available to follow here<\/a>. Everybody on the larger list is equally worthy of following and meeting, as they\u2019re all crucial parts of our great ecosystem.<\/p><p><img class=\"aligncenter size-full wp-image-573\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/founders-and-entrep.png\" alt=\"founders+and+entrep\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/founders-and-entrep.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/founders-and-entrep-300x12.png 300w\" \/><\/p><div class=\"flex-wrap photo-grid\"><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/dharmesh\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"512\" height=\"512\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dharmesh.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dharmesh.jpg 512w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dharmesh-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dharmesh-300x300.jpg 300w\" \/><div class=\"wp-caption\">Dharmesh Shah<p>Co-Founder &amp; CTO, HubSpot<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/Wayne\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/waynechang.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/waynechang.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/waynechang-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/waynechang-300x300.png 300w\" \/><div class=\"wp-caption\">Wayne Chang<p>Co-Founder, Crashlytics (acq. by Twitter)<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/DianeHessan\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"500\" height=\"495\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dianehessan.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dianehessan.png 500w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dianehessan-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dianehessan-300x297.png 300w\" \/><div class=\"wp-caption\">Diane Hessan<p>CEO, Startup Institute<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/dcancel\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"346\" height=\"346\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dcancel.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dcancel.jpg 346w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dcancel-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dcancel-300x300.jpg 300w\" \/><div class=\"wp-caption\">David Cancel<p>CEO, Drift<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/lum\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jenniferlum.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jenniferlum.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jenniferlum-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jenniferlum-300x300.jpg 300w\" \/><div class=\"wp-caption\">Jennifer Lum<p>Co-founder, Adelphic Mobile<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/mattlauzon\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"500\" height=\"500\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mattlauzon.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mattlauzon.jpg 500w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mattlauzon-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mattlauzon-300x300.jpg 300w\" \/><div class=\"wp-caption\">Matt Lauzon<p>Former CEO of Gemvara,\u00a0Dunwello<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/angusdav\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"370\" height=\"370\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/angusdav.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/angusdav.png 370w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/angusdav-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/angusdav-300x300.png 300w\" \/><div class=\"wp-caption\">Angus Davis<p>CEO, Upserve (fka Swipely)<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/csavage\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"512\" height=\"512\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/csavage.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/csavage.png 512w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/csavage-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/csavage-300x300.png 300w\" \/><div class=\"wp-caption\">Chris Savage<p>CEO, Wistia<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/jjacobs22\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jjacobs22.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jjacobs22.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jjacobs22-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jjacobs22-300x300.jpg 300w\" \/><div class=\"wp-caption\">Jason Jacobs<p>CEO, RunKeeper<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/smarcelo\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"1426\" height=\"1424\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarcelo.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarcelo.jpg 1426w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarcelo-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarcelo-300x300.jpg 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarcelo-768x767.jpg 768w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarcelo-1024x1024.jpg 1024w\" \/><div class=\"wp-caption\">Sheila Marcelo<p>CEO, Care.com<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/sravish\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sravish.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sravish.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sravish-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sravish-300x300.png 300w\" \/><div class=\"wp-caption\">Sravish Sridhar<p>CEO, Kinvey<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/natalykogan\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/natalykogan.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/natalykogan.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/natalykogan-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/natalykogan-300x300.png 300w\" \/><div class=\"wp-caption\">Nataly Kogan<p>CEO, Happier<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><\/div><p><img class=\"aligncenter size-full wp-image-590\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-investors.png\" alt=\"boston-tech-investors\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-investors.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-investors-300x12.png 300w\" \/><\/p><div class=\"flex-wrap photo-grid\"><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/bijan\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bijan.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bijan.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bijan-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bijan-300x300.jpg 300w\" \/><div class=\"wp-caption\">Bijan Sabet<p>Spark Capital<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/bostonvc\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostonvc.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostonvc.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostonvc-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostonvc-300x300.png 300w\" \/><div class=\"wp-caption\">David Skok<p>Matrix Partners<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/MaiaHeymann\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"300\" height=\"300\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/maia.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/maia.jpg 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/maia-150x150.jpg 150w\" \/><div class=\"wp-caption\">Maia Heymann<p>Converge Venture Partners<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/bussgang\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bussgang.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bussgang.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bussgang-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bussgang-300x300.jpg 300w\" \/><div class=\"wp-caption\">Jeff Bussgang<p>Flybridge Capital<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/robgo\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/robgo.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/robgo.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/robgo-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/robgo-300x300.jpg 300w\" \/><div class=\"wp-caption\">Rob Go<p>NextView; Creator of the Hitchhiker\u2019s Guide to Boston Tech<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/VCMike\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"200\" height=\"200\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mikehirshland-1.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mikehirshland-1.jpg 200w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mikehirshland-1-150x150.jpg 150w\" \/><div class=\"wp-caption\">Mike Hirshland<p>Resolute Ventures<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"http:\/\/twitter.com\/jeanhammond\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Hammond7_smile_400x400.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Hammond7_smile_400x400.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Hammond7_smile_400x400-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Hammond7_smile_400x400-300x300.jpg 300w\" \/><div class=\"wp-caption\">Jean Hammond<p>LearnLaunch<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/larryvc\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/larryvc.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/larryvc.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/larryvc-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/larryvc-300x300.jpg 300w\" \/><div class=\"wp-caption\">Larry Cheng<p>Volition Capital<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/epaley\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/epaley.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/epaley.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/epaley-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/epaley-300x300.jpg 300w\" \/><div class=\"wp-caption\">Eric Paley<p>Founder Collective<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/richminer\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/richminer.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/richminer.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/richminer-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/richminer-300x300.png 300w\" \/><div class=\"wp-caption\">Rich Miner<p>Google Ventures<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/andrewparker\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"900\" height=\"900\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/andrewparker.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/andrewparker.jpg 900w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/andrewparker-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/andrewparker-300x300.jpg 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/andrewparker-768x768.jpg 768w\" \/><div class=\"wp-caption\">Andrew Parker<p>Spark Capital<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/leehower\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/leehower.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/leehower.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/leehower-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/leehower-300x300.jpg 300w\" \/><div class=\"wp-caption\">Lee Hower<p>NextView<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/ktrae\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ktrae.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ktrae.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ktrae-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ktrae-300x300.png 300w\" \/><div class=\"wp-caption\">Katie Rae<p>Project 11<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/davidbeisel\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/davidbeisel.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/davidbeisel.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/davidbeisel-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/davidbeisel-300x300.png 300w\" \/><div class=\"wp-caption\">David Beisel<p>NextView; Creator, Boston Innovators Group<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/jillpreotle\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"376\" height=\"375\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jill-p.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jill-p.jpg 376w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jill-p-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jill-p-300x300.jpg 300w\" \/><div class=\"wp-caption\">Jill Preotle<p>Golden Seeds<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><\/div><div class=\"col-md-12\"><hr \/><h2>Follow the Official Hitchhiker\u2019s Guide Twitter List<\/h2><p>Includes everyone listed above and below, plus many others: entrepreneurs, executives, local media, notable business authors and academics, and community builders, all innovating here in Boston.<\/p><p><a href=\"https:\/\/twitter.com\/NextViewVC\/lists\/hitchhiker-s-guide-to-bos\" target=\"_blank\"><img class=\"aligncenter size-full wp-image-609\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-list.png\" alt=\"boston-tech-list\" width=\"368\" height=\"65\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-list.png 368w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-list-300x53.png 300w\" \/><\/a><\/p><hr \/><p><img class=\"aligncenter wp-image-614 size-full\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download2.png\" alt=\"download\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download2.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download2-300x12.png 300w\" \/><\/p><\/div><div class=\"flex-wrap photo-grid\"><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/danprimack\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"200\" height=\"200\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/danprimack.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/danprimack.jpg 200w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/danprimack-150x150.jpg 150w\" \/><div class=\"wp-caption\">Dan Primack<p>Fortune<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/ScottKirsner\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/scottkirsner.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/scottkirsner.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/scottkirsner-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/scottkirsner-300x300.jpg 300w\" \/><div class=\"wp-caption\">Scott Kirsner<p>Boston Globe<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/gigabarb\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"256\" height=\"256\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/barbdarrow.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/barbdarrow.jpg 256w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/barbdarrow-150x150.jpg 150w\" \/><div class=\"wp-caption\">Barb Darrow<p>Fortune<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/steveannear\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/steveannear.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/steveannear.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/steveannear-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/steveannear-300x300.jpg 300w\" \/><div class=\"wp-caption\">Steve Annear<p>Boston\u00a0Globe<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/galenmoore\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/galenmoore.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/galenmoore.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/galenmoore-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/galenmoore-300x300.png 300w\" \/><div class=\"wp-caption\">Galen Moore<p>BostInno<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/wfrick\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wfrick.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wfrick.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wfrick-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/wfrick-300x300.png 300w\" \/><div class=\"wp-caption\">Walt Frick<p>Harvard Business Review<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/NidhiSubs\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"256\" height=\"256\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/9bd6aa8f2a58125c9502f7c99d43f44b.jpeg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/9bd6aa8f2a58125c9502f7c99d43f44b.jpeg 256w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/9bd6aa8f2a58125c9502f7c99d43f44b-150x150.jpeg 150w\" \/><div class=\"wp-caption\">Nidhi Subbaraman<p>Boston Globe<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/gthuang\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gthuang.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gthuang.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gthuang-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gthuang-300x300.jpg 300w\" \/><div class=\"wp-caption\">Gregory Huang<p>Xconomy<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/curtwoodward\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"512\" height=\"512\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/curt.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/curt.jpg 512w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/curt-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/curt-300x300.jpg 300w\" \/><div class=\"wp-caption\">Curt Woodward<p>Boston Globe<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/GlobeTechLab\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"283\" height=\"284\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hi-1.jpeg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hi-1.jpeg 283w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hi-1-150x150.jpeg 150w\" \/><div class=\"wp-caption\">Hiawatha Bray<p>Boston Globe<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/charlie_acb\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"300\" height=\"300\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amanda.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amanda.jpg 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amanda-150x150.jpg 150w\" \/><div class=\"wp-caption\">Amanda Burke<p>Boston Globe<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/jeffengelxcon\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"240\" height=\"240\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/v2lqrdge.jpeg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/v2lqrdge.jpeg 240w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/v2lqrdge-150x150.jpeg 150w\" \/><div class=\"wp-caption\">Jeff Engel<p>Xconomy<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/DylanLJMartin\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"1128\" height=\"1128\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dylan-martin12.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dylan-martin12.jpg 1128w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dylan-martin12-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dylan-martin12-300x300.jpg 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dylan-martin12-768x768.jpg 768w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dylan-martin12-1024x1024.jpg 1024w\" \/><div class=\"wp-caption\">Dylan Martin<p>BostInno<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><\/div><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-629\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-102.png\" alt=\"download (10)\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-102.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-102-300x12.png 300w\" \/><\/p><\/div><div class=\"flex-wrap photo-grid\"><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/BostonTweet\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostontweet.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostontweet.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostontweet-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bostontweet-300x300.png 300w\" \/><div class=\"wp-caption\">Tom O'Keefe<p>BostonTweet<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/simplebits\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/simplebits.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/simplebits.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/simplebits-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/simplebits-300x300.jpg 300w\" \/><div class=\"wp-caption\">Dan Cederholm<p>Dribbble<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/stevegarfield\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/stevegarfield.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/stevegarfield.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/stevegarfield-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/stevegarfield-300x300.jpg 300w\" \/><div class=\"wp-caption\">Steve Garfield<p>Boston Media Makers<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"http:\/\/twitter.com\/pistachio\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"200\" height=\"200\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/laura-fitton.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/laura-fitton.jpg 200w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/laura-fitton-150x150.jpg 150w\" \/><div class=\"wp-caption\">Laura Fitton<p>HubSpot<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/DonDodge\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dondodge.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dondodge.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dondodge-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dondodge-300x300.png 300w\" \/><div class=\"wp-caption\">Don Dodge<p>Google<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/ca_webb\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"372\" height=\"370\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ca-webb.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ca-webb.png 372w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ca-webb-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/ca-webb-300x298.png 300w\" \/><div class=\"wp-caption\">C.A. Webb<p>Assemble VC<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/JoselinMane\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joselinmane.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joselinmane.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joselinmane-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joselinmane-300x300.png 300w\" \/><div class=\"wp-caption\">Joselin Mane<p>BostonTweetUp<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/HackerChick\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hackerchick.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hackerchick.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hackerchick-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hackerchick-300x300.png 300w\" \/><div class=\"wp-caption\">Abby Fichtner<p>Harvard iLab<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/jayacunzo\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"800\" height=\"806\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jay-acunzo-2015-close-crop-copy-2.jpeg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jay-acunzo-2015-close-crop-copy-2.jpeg 800w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jay-acunzo-2015-close-crop-copy-2-150x150.jpeg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jay-acunzo-2015-close-crop-copy-2-298x300.jpeg 298w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jay-acunzo-2015-close-crop-copy-2-768x774.jpeg 768w\" \/><div class=\"wp-caption\">Jay Acunzo<p>NextView; Creator of the Hitchhiker\u2019s Guide to Boston Tech<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/rowe\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/rowe.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/rowe.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/rowe-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/rowe-300x300.png 300w\" \/><div class=\"wp-caption\">Timothy Rowe<p>CIC<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/katieburkie\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"500\" height=\"500\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/katie-burke.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/katie-burke.jpg 500w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/katie-burke-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/katie-burke-300x300.jpg 300w\" \/><div class=\"wp-caption\">Katie Burke<p>HubSpot<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/changds\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/changds.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/changds.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/changds-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/changds-300x300.png 300w\" \/><div class=\"wp-caption\">David Chang<p>EIR at Harvard Business School<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/BobbieC\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"498\" height=\"475\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-17-at-2.32.07-PM.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-17-at-2.32.07-PM.png 498w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/Screen-Shot-2016-03-17-at-2.32.07-PM-300x286.png 300w\" \/><div class=\"wp-caption\">Bobbie Carlton<p>Mass. Innovation Nights<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/frogandcode\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/frogandcode.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/frogandcode.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/frogandcode-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/frogandcode-300x300.png 300w\" \/><div class=\"wp-caption\">Rich Thornett<p>Dribbble<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/theKeithF\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/keith-f.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/keith-f.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/keith-f-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/keith-f-300x300.png 300w\" \/><div class=\"wp-caption\">Keith Frankel<p>CreativeMornings<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/cortlandt\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cortlandt.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cortlandt.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cortlandt-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cortlandt-300x300.png 300w\" \/><div class=\"wp-caption\">Cort Johnson<p>Accomplice<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/devinemily\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"599\" height=\"598\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/devin.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/devin.jpg 599w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/devin-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/devin-300x300.jpg 300w\" \/><div class=\"wp-caption\">Devin Bramhall<p>Master Slam; Boston Content<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/sarahsherby\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sarah.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sarah.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sarah-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/sarah-300x300.jpg 300w\" \/><div class=\"wp-caption\">Sarah Sherburne<p>TechGen<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><\/div><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-645\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-201.png\" alt=\"download (20)\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-201.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-201-300x12.png 300w\" \/><\/p><\/div><div class=\"flex-wrap photo-grid\"><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/claychristensen\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/claychristensen.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/claychristensen.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/claychristensen-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/claychristensen-300x300.jpg 300w\" \/><div class=\"wp-caption\">Clayton Christensen<p>Author, <em>The Innovator\u2019s Dilemma<\/em>, Professor, Harvard Business School<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/Joi\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joi.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joi.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joi-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/joi-300x300.png 300w\" \/><div class=\"wp-caption\">Joi Ito<p>MIT Media Lab<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/youngmemoon\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"1416\" height=\"1416\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/youngme.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/youngme.jpg 1416w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/youngme-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/youngme-300x300.jpg 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/youngme-768x768.jpg 768w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/youngme-1024x1024.jpg 1024w\" \/><div class=\"wp-caption\">Youngme Moon<p>Author,\u00a0<em>Different<\/em>; Professor, Harvard Business School<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/BillAulet\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"256\" height=\"256\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bill.jpeg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bill.jpeg 256w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bill-150x150.jpeg 150w\" \/><div class=\"wp-caption\">Bill Aulet<p>Martin Trust Center for MIT Entrepreneurship<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/cc_chapman\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cc_chapman.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cc_chapman.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cc_chapman-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/cc_chapman-300x300.png 300w\" \/><div class=\"wp-caption\">C.C. Chapman<p>Author, <em>Content Rules<\/em><\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/edwardboches\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/edwardboches.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/edwardboches.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/edwardboches-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/edwardboches-300x300.png 300w\" \/><div class=\"wp-caption\">Edward Boches<p>Professor, Boston University<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/amcafee\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amcafee.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amcafee.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amcafee-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/amcafee-300x300.png 300w\" \/><div class=\"wp-caption\">Andrew Mcafee<p>Principal Research Scientist, MIT Sloan Co-author, <em>The Second Machine Age<\/em><\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/teisenmann\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/teisenmann.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/teisenmann.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/teisenmann-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/teisenmann-300x300.png 300w\" \/><div class=\"wp-caption\">Tom Eisenmann<p>Professor, Harvard Business School<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/annhandley\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/annhandley.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/annhandley.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/annhandley-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/annhandley-300x300.jpg 300w\" \/><div class=\"wp-caption\">Ann Handley<p>Chief Content Officer at MarketingProfs<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/DrewDavisHere\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"306\" height=\"306\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bra_drewdavisface_306X306.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bra_drewdavisface_306X306.jpg 306w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bra_drewdavisface_306X306-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bra_drewdavisface_306X306-300x300.jpg 300w\" \/><div class=\"wp-caption\">Andrew Davis<p>Author,\u00a0<em>Town Inc<\/em> and\u00a0<em>Brandscaping;\u00a0<\/em>CEO, Monumental Shift<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/perryhewitt\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/perryhewitt.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/perryhewitt.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/perryhewitt-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/perryhewitt-300x300.png 300w\" \/><div class=\"wp-caption\">Perry Hewitt<p>Chief Digital Officer, Harvard<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/gallaugher\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gallaugher.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gallaugher.jpg 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gallaugher-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/gallaugher-300x300.jpg 300w\" \/><div class=\"wp-caption\">John Gallaugher<p>Professor, Boston College<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/nickducoff\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/nickducoff.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/nickducoff.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/nickducoff-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/nickducoff-300x300.png 300w\" \/><div class=\"wp-caption\">Nick Ducoff<p>VP New Ventures, Northeastern<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/jgoldstein\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"401\" height=\"401\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jodi-g.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jodi-g.jpg 401w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jodi-g-150x150.jpg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/jodi-g-300x300.jpg 300w\" \/><div class=\"wp-caption\">Jodi Goldstein<p>Harvard Innovation Lab<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/paulfcostello\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"400\" height=\"400\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/paulfcostello.png\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/paulfcostello.png 400w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/paulfcostello-150x150.png 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/paulfcostello-300x300.png 300w\" \/><div class=\"wp-caption\">Paul Costello<p>Lecturer, Babson<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/susangduffy\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"299\" height=\"299\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/duffy.jpg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/duffy.jpg 299w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/duffy-150x150.jpg 150w\" \/><div class=\"wp-caption\">Susan Duffy<p>Center for Women\u2019s Entrepreneurial Leadership, Babson<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><div class=\"col-md-4\"><a href=\"https:\/\/twitter.com\/Fiona_MIT\" target=\"_blank\">&#13;\n\t\t\t\t\t\t\t\t\t<img width=\"500\" height=\"500\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/fiona.jpeg\" class=\"attachment-full size-full\" alt=\"\" data-sizes=\"auto\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/fiona.jpeg 500w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/fiona-150x150.jpeg 150w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/fiona-300x300.jpeg 300w\" \/><div class=\"wp-caption\">Fiona Murray<p>Professor, MIT<\/p><\/div>&#13;\n\t\t\t\t\t\t\t\t<\/a><\/div><\/div><div class=\"col-md-12\"><p><a href=\"https:\/\/twitter.com\/NextViewVC\/lists\/hitchhiker-s-guide-to-bos\" target=\"_blank\">See the Entire List of Who to Follow &gt;&gt;<\/a><\/p><\/div><\/div><\/div><\/div><\/section><section id=\"get-out-there\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><p><img class=\"aligncenter size-full wp-image-665\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-out-there.png\" alt=\"get+out+there\" width=\"973\" height=\"354\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-out-there.png 973w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-out-there-300x109.png 300w\" \/><\/p><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-667\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/BosTech_v14-3_05-compressor.png\" alt=\"BosTech_v14-3_05-compressor\" width=\"811\" height=\"32\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/BosTech_v14-3_05-compressor.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/BosTech_v14-3_05-compressor-300x12.png 300w\" \/><\/p><p>Boston tech is full of incredibly passionate, exceedingly smart, and fiercely supportive people. This ecosystem prides itself on its pay-it-forward mentality for those seeking advice and support. The events, community groups, and coffee shops below are great ways to join or navigate the local startup scene.<\/p><p><img class=\"aligncenter size-full wp-image-669\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-events.png\" alt=\"boston+tech+events\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-events.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-tech-events-300x12.png 300w\" \/><\/p><p>A sampling of Boston\u2019s ever-growing list of events, starting with some of the larger, pillar conferences.<\/p><div class=\"flex-wrap\"><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/soi.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>BostInno\u2019s flagship annual event attracts over 1,000 business leaders, entrepreneurs, and students. The half-day event covers the state of innovation locally in marketing and advertising, mobile, retail\/e-commerce, big data, transportation, venture capital, and more.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/bostonstateofinnovation.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tedx.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>TEDx conferences are hosted every year in <a href=\"http:\/\/tedxboston.org\/\" target=\"_blank\">Boston<\/a>, <a href=\"http:\/\/www.tedxcambridge.com\/\" target=\"_blank\">Cambridge<\/a>, and\u00a0<a href=\"http:\/\/www.ted.com\/tedx\/events\/8705\" target=\"_blank\">Somerville<\/a>.<\/p><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/sloansports.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The tech community has a big presence at this annual conference, which attracts not only entrepreneurs but pro sports commissioners, owners, management, coaches, and athletes. This event has helped catalyze the massive sports industry into adopting tech and analytics in new and innovative ways across the globe.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.sloansportsconference.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/big-group.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Formerly known as WebInno, BIG is Boston\u2019s biggest and longest-running tech event\u00a0(over 10\u00a0years). This quarterly meetup showcases early-stage startups through various demos and presentations. It attracts over hundreds\u00a0attendees each time.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/bostoninnovatorsgroup.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/inbound.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>INBOUND is HubSpot\u2019s annual conference, which began as a customer meetup and has since expanded to attract 10,000 attendees from the marketing, sales, and startup worlds.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.inbound.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/tugg.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>TUGG (Technology Underwriting the Greater Good) brings together the tech community to fund and support local nonprofits. Their annual Tech Gives Back day sees over 2,000 volunteers taking the day off to donate their time to 50+ sites and organizations.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/tugg.org\/techgivesback\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/nantucket.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>This invite-only weekend event attracts top entrepreneurs and investors from Boston and beyond.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.nantucketconference.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/nevy.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Run by the New England Venture Capital Association, the NEVYs brings over 700 VCs and entrepreneurs together for a night of recognition across multiple categories and sectors.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.newenglandvc.org\/nevy-awards\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/uncon.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>By design, this event has no agenda or planned format. The content is entirely crowdsourced based on what attendees want to discuss.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/masstlcuncon.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/emtech.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>MIT Technology Review\u2019s Emerging Technologies conference attracts senior leaders seeking to understand the future of technology and how it will change the business world.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.technologyreview.com\/emtech\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/cyberposium.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Entering its third decade of operation, Cyberposium is the largest student-run technology event for MBAs in the world. The conference is held at Harvard Business School each year.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.cyberposium.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/bos-techjam.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Boston TechJam is a block party and innovation meetup of the city\u2019s best and brightest tech minds. With live music, interactive challenges, and local food trucks and drinks, it\u2019s one of the more celebratory tech events in town each year.<\/p>\u00a0<a target=\"_blank\" href=\"http:\/\/www.bostontechjam.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-4\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/hubweek.png\" width=\"300\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>HUBweek is a creative festival that celebrates innovation at the intersection of art, science, and technology. Founded by The Boston Globe, Harvard University, MIT and Mass General Hospital, HUBweek brings together the most creative and inventive minds in Boston and\u00a0globally.<\/p>\u00a0<a target=\"_blank\" href=\"https:\/\/hubweek.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><\/div><p><img class=\"aligncenter size-full wp-image-673\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/events-contd.png\" alt=\"events+contd\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/events-contd.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/events-contd-300x12.png 300w\" \/><\/p><p>Even more great ways to network.<\/p><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-94.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-103.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><div class=\"col-md-12\"><p><strong><a href=\"http:\/\/list.ly\/list\/RxZ-events-hitchhikers-guide-to-boston-tech\" target=\"_blank\">Submit more &amp; view what the community added &gt;&gt;<\/a><\/strong><\/p><p><img class=\"aligncenter size-full wp-image-870\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/community1.png\" alt=\"community\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/community1.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/community1-300x12.png 300w\" \/><\/p><p>Whether you join as a member or just drop by an event or meetup, these groups are the glue holding this community together and advancing their various specialty areas here in Boston.<\/p><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-114.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/wistia-office-compressor.jpg\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-707\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/coffee.png\" alt=\"coffee\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/coffee.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/coffee-300x12.png 300w\" \/><\/p><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/voltage.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download2.jpg\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-712\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/venues.png\" alt=\"venues\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/venues.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/venues-300x12.png 300w\" \/><\/p><p>The following are listed based roughly on event frequency and perceived relevancy to the Boston startup community. (In many cases, these venues just \u201cfeel\u201d top of mind to local entrepreneurs.)<\/p><\/div><div class=\"col-md-5\"><ul><li><strong><a href=\"https:\/\/microsoftnewengland.com\/\" target=\"_blank\">Microsoft NERD<\/a><\/strong><\/li>\n<li><a href=\"http:\/\/i-lab.harvard.edu\/\" target=\"_blank\"><strong>Harvard iLab<\/strong><\/a><\/li>\n<li><strong>WeWork: <a href=\"https:\/\/www.wework.com\/locations\/boston\/fort-point\" target=\"_blank\">Fort Point<\/a>, <a href=\"https:\/\/www.wework.com\/locations\/boston\/south-station\" target=\"_blank\">South Station<\/a><\/strong><\/li>\n<li><a href=\"http:\/\/districthallboston.org\/\" target=\"_blank\"><strong>District Hall<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.startupinstitute.com\/\" target=\"_blank\"><strong>Startup Institute office<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/communispace.com\/Home\/\" target=\"_blank\"><strong>Communispace office<\/strong><\/a><\/li>\n<li><strong><a href=\"http:\/\/brooklynboulders.com\/somerville\/\" target=\"_blank\">Brooklyn Boulders (Somerville)<\/a><\/strong><\/li>\n<li><a href=\"https:\/\/www.ibm.com\/developerworks\/community\/groups\/service\/html\/communityview?communityUuid=e902a0a0-d0de-4d45-9df9-230811c7f56f\" target=\"_blank\"><strong>IBM Innovation Center<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.dillonsboston.com\/\" target=\"_blank\"><strong>Dillon\u2019s Restaurant &amp; Bar<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.seaportboston.com\/\" target=\"_blank\"><strong>Seaport Hotel &amp; World Trade Center<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.backbaysocialclub.com\/\" target=\"_blank\"><strong>Back Bay Social Club<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/regencyboston.hyatt.com\/en\/hotel\/home.html\" target=\"_blank\"><strong>Hyatt Regency Boston<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.ritzcarlton.com\/en\/Properties\/BostonCommon\/Default.htm\" target=\"_blank\"><strong>Boston Common Ritz Carlton<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.sonsieboston.com\/home\" target=\"_blank\"><strong>Sonsie<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/www.49social.com\/\" target=\"_blank\"><strong>49 Social<\/strong><\/a><\/li>\n<\/ul><\/div><div class=\"col-md-7\"><p>&#13;\n\t\t\t\t\t\t<img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/district-hall.jpg\" class=\"alignright size-full\" alt=\"image\" \/>District Hall (Innovation District, South Boston)\t\t\t\t\t<\/p><\/div><div class=\"col-md-12\"><p id=\"yui_3_17_2_1_1425323554504_6869\"><a id=\"yui_3_17_2_1_1425323554504_6868\" href=\"http:\/\/list.ly\/list\/Rxb-event-venues-hitchhikers-guide-to-boston-tech\" target=\"_blank\"><strong>Submit more &amp; view what the community added &gt;&gt;<\/strong><\/a><\/p><\/div><\/div><\/div><\/section><section id=\"get-to-work\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><p><img class=\"aligncenter size-full wp-image-778\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-to-work.png\" alt=\"get+to+work\" width=\"973\" height=\"353\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-to-work.png 973w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-to-work-300x109.png 300w\" \/><\/p><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-780\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-jobs.png\" alt=\"boston+startup+jobs\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-jobs.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-jobs-300x12.png 300w\" \/><\/p><p>All those local startups launching regularly around town add up to one great thing for job seekers: opportunity. The following resources can help you find your next career stop:<\/p><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-startup-jobs-1.png\" alt=\"boston+startup+jobs\" class=\"aligncenter size-full\" \/><\/div><div class=\"col-md-6\"><p>&#13;\n\t\t\t\t\t\t<img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download3.jpg\" class=\"aligncenter size-full\" alt=\"image\" \/><\/p><\/div><p>&#13;\n\t\t\t\t\t<\/p><h3><strong><a href=\"http:\/\/venturefizz.com\/jobs\" target=\"_blank\">VentureFizz Job Board<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/bostinno.streetwise.co\/careers\/\" target=\"_blank\">Careers on BostInno<\/a><\/strong><\/h3>\n<h3><strong><a href=\"https:\/\/angel.co\/boston\/jobs\" target=\"_blank\">Boston Startup Jobs on AngelList<\/a><\/strong><\/h3><div class=\"col-md-2\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/handshake.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/div><p>&#13;\n\t\t\t\t\t<\/p><h3><strong><a href=\"http:\/\/boston.startupjobfair.org\/\" target=\"_blank\">Boston Startup Fair Jobs<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/bostonstartupsguide.com\/boston-startup-jobs\/\" target=\"_blank\">BostonStartupsGuide.com Careers<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/www.ventureloop.com\/ventureloop\/home.php\" target=\"_blank\">VentureLoop Startup Job Search<\/a><\/strong><\/h3><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-790\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/incubators.png\" alt=\"incubators\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/incubators.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/incubators-300x12.png 300w\" \/>These organizations have contributed to the growth of Boston\u2019s tech startup ecosystem by providing office space, advice, introductions, support, and even budget to fledgling companies.<\/p><\/div><div class=\"flex-wrap\"><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/techstars.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>TechStars is one of the top accelerators in the world. Its Boston chapter is led by investor Semyon Dukach, who gained early recognition as a student leading the famous MIT blackjack team. TechStars Boston companies (71 as of April 2014) have gone on to raise more than $153 million in funding.<\/p><a target=\"_blank\" href=\"http:\/\/www.techstars.com\/program\/locations\/boston\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/mc.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The largest accelerator in the world, with over 128 startups per program, MassChallenge is a nonprofit that takes no equity in its companies. As of April 2014, MC alumni have raised $550 million, driven $350 million in revenue, and hired over 4,000 employees.<\/p><a target=\"_blank\" href=\"http:\/\/masschallenge.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/summer-at-h.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Local VC firm Highland Capital Partners offers students this 10-week summer program, where they receive office space, mentorship, access to people and organizations, and a stipend of $18,000.<\/p><a target=\"_blank\" href=\"http:\/\/summer.hcp.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/learn-launch.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>The accelerator arm of LearnLaunch, which is focused on education technology.<\/p><a target=\"_blank\" href=\"http:\/\/learnlaunch.com\/accelerator\/program\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/innoloft.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Existing startups focused on helping small businesses can apply to this program, created by Constant Contact. In addition to the usual benefits offered by accelerators, InnoLoft also grants its startups $10,000 in marketing budget.<\/p><a target=\"_blank\" href=\"http:\/\/smbinnoloft.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/launch-bos-1.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Launch creates and incubates ecommerce companies. They focus on a low volume of companies by design, and provide each with access to\u00a0talent, technology and partners.<\/p><a target=\"_blank\" href=\"http:\/\/teamlaunch.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/smarter.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Smarter in the City is the first high-tech accelerator in the Roxbury neighborhood of Boston. They are focused\u00a0on fixing\u00a0the lack of opportunities for computer science and digital media entrepreneurs in lower-income neighborhoods and minority communities.<\/p><a target=\"_blank\" href=\"http:\/\/www.smarterinthecity.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/dat.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>Dat Ventures helps foreign startups break into the US market.<\/p><a target=\"_blank\" href=\"http:\/\/www.datventures.com\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><div class=\"col-md-3\"><div class=\"thumb\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/fintech.png\" class=\"aligncenter size-full\" alt=\"image\" \/><div class=\"text\"><p>FinTech Sandbox is a nonprofit that provides\u00a0free, streamlined access to critical data and resources to FinTech entrepreneurs and startups.<\/p><a target=\"_blank\" href=\"http:\/\/fintechsandbox.org\/\">Visit&gt;&gt;<\/a><\/div><\/div><\/div><\/div><div class=\"col-md-12\"><p><strong><a href=\"http:\/\/list.ly\/list\/Rxc-incubators-and-accelerators-hitchhikers-guide-to-boston-tech\" target=\"_blank\">Submit more &amp; view what the community added &gt;&gt;<\/a><\/strong><\/p><p><img class=\"aligncenter size-full wp-image-804\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/coworking.png\" alt=\"coworking\" width=\"811\" height=\"33\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/coworking.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/coworking-300x12.png 300w\" \/><\/p><p>Coworking spaces have also emerged in seemingly every corner of Boston. Below is a sampling of some of the better-known options, but BostonStartupsGuide.com also created <a href=\"http:\/\/bostonstartupsguide.com\/guide\/2013-boston-coworking-spaces-roundup\/\" target=\"_blank\">a comprehensive list in 2014.<\/a><\/p><\/div><div class=\"col-md-6 sidelist\"><p><strong><a href=\"http:\/\/cic.us\/\" target=\"_blank\" data-cke-saved-href=\"http:\/\/cic.us\/\">CIC<\/a><\/strong><\/p><p><strong><a href=\"https:\/\/www.wework.com\/\" target=\"_blank\" data-cke-saved-href=\"https:\/\/www.wework.com\/\">WEWORK<\/a><\/strong><\/p><p><strong><a href=\"http:\/\/workbar.com\/\" target=\"_blank\" data-cke-saved-href=\"http:\/\/workbar.com\/\">WORKBAR<\/a><\/strong><\/p><p><strong><a href=\"https:\/\/krash.io\/\" target=\"_blank\" data-cke-saved-href=\"https:\/\/krash.io\/\">KRASH<\/a><\/strong><\/p><p><strong><a href=\"http:\/\/www.hackreduce.org\/\" target=\"_blank\" data-cke-saved-href=\"http:\/\/www.hackreduce.org\/\">HACK\/REDUCE<\/a><\/strong><\/p><p><strong><a href=\"https:\/\/cove.is\/index\" target=\"_blank\">COVE<\/a><\/strong><\/p><p><strong><a href=\"https:\/\/www.industriousoffice.com\/locations\/boston-seaport\/\" target=\"_blank\">INDUSTRIOUS<\/a><\/strong><\/p><\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/General-Icons-06.jpg\" class=\"alignleft size-full\" alt=\"image\" \/><br \/>CIC (Kendall Square, Cambridge)<\/div><div class=\"col-md-6\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/General-Icons-06.jpg\" class=\"alignleft size-full\" alt=\"image\" \/><br \/><\/div><p>&#13;\n\t\t\t\t\t<\/p><h3><strong><a href=\"http:\/\/www.koalab.com\/\" target=\"_blank\">KOA LABS<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/greentownlabs.org\/\" target=\"_blank\">GREENTOWN LABS<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/bocoup.com\/loft\/\" target=\"_blank\">BOCOUP LOFT<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/www.nginworkplace.com\/\" target=\"_blank\">NGIN WORKPLACE<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/www.101tremont.com\/\" target=\"_blank\">101 TREMONT<\/a><\/strong><\/h3>\n<h3><strong><a href=\"http:\/\/ideaspaceboston.com\/\" target=\"_blank\">IDEA SPACE<\/a><\/strong><\/h3><div class=\"col-md-12\"><p><a href=\"http:\/\/list.ly\/list\/Rxd-coworking-spaces-hitchhikers-guide-to-boston-tech\" target=\"_blank\"><strong>Submit more &amp; view what the community added &gt;&gt;<\/strong><\/a><\/p><p><img class=\"aligncenter size-full wp-image-815\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/prod-dev-firms.png\" alt=\"prod+dev+firms\" width=\"811\" height=\"32\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/prod-dev-firms.png 811w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/prod-dev-firms-300x12.png 300w\" \/><\/p><p>Have an idea but not sure where to start or how to build it? Just need additional resources while launching? These firms specialize in designing and building apps and products, often working with early-stage startups. Many also go beyond simply building by helping you scope a product and test the market before you launch.<\/p><\/div><div class=\"flex-wrap\"><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"https:\/\/thoughtbot.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/thoughtbot.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tThoughtbot<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.freshtilledsoil.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/freshtilled.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFresh Tilled Soil<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.rocketinsights.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/rocket-insights.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tRocket Insights<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.thebig-studio.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download5.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tThe Big Studio<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.dragoninnovation.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/dragon.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tDragon Innovation<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/extensionengine.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/extension-compressor.jpeg\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tExtension Engine<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.raizlabs.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/raiz-compressor.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tRaizlabs<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.janeirodigital.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/jd.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tJaneiro Digital<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/intrepid.io\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/intrepid.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tIntrepid Pursuits<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/thecreamcreative.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/cream.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tThe Cream Creative<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/www.occomgroup.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/download-115.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tOccom Group<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/firefield.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/FF.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFirefield<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/filamentgroup.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/filmanet-compressor.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tFilament Group<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/pivotallabs.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/pivotal-compressor.jpeg\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tPivotal Labs<\/p><\/div><\/div><div class=\"col-md-2\"><div class=\"thumb\"><a target=\"_blank\" href=\"http:\/\/greenfieldhq.com\/\"><img src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2016\/01\/greenfield.png\" class=\"aligncenter size-full\" alt=\"image\" \/><\/a><p>&#13;\n\t\t\t\t\t\t\t\t\t\t\tGreenfield<\/p><\/div><\/div><\/div><div class=\"col-md-12\"><p><a href=\"http:\/\/list.ly\/list\/Rxe-product-firms-hitchhikers-guide-to-boston-tech\" target=\"_blank\"><strong>Submit more &amp; view what the community added &gt;&gt;<\/strong><\/a><\/p><\/div><\/div><\/div><\/section><section id=\"get-funded\" class=\"page-wrapper light-bg\"><div class=\"container\"><div class=\"row page-content\"><p><img class=\"aligncenter size-full wp-image-841\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-funded.png\" alt=\"get+funded\" width=\"1200\" height=\"439\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-funded.png 1200w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-funded-300x110.png 300w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/get-funded-1024x375.png 1024w\" \/><\/p><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-869\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-vcs.png\" alt=\"boston+vcs\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-vcs.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/boston-vcs-300x12.png 300w\" \/>In 2013, the last full year before this site was created, Boston saw more venture capital invested than any region other than California, with over $3 billion. Some of the most active web and mobile tech VCs are listed below.<\/p><p><img class=\"aligncenter size-full wp-image-871\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/seed-vcs.png\" alt=\"seed+vcs\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/seed-vcs.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/02\/seed-vcs-300x12.png 300w\" \/><\/p><\/div><p><img class=\"aligncenter size-full wp-image-873\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/boston-angels.png\" alt=\"boston+angels\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/boston-angels.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/boston-angels-300x12.png 300w\" \/><\/p><p><img class=\"aligncenter size-full wp-image-887\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/early-stage.png\" alt=\"early+stage\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/early-stage.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/early-stage-300x12.png 300w\" \/><\/p><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-917\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/early-and-growth.png\" alt=\"early+and+growth\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/early-and-growth.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/early-and-growth-300x12.png 300w\" \/><\/p><\/div><div class=\"col-md-12\"><p><img class=\"aligncenter size-full wp-image-921\" src=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/growth-vcs.png\" alt=\"growth+vcs\" width=\"1000\" height=\"40\" srcset=\"http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/growth-vcs.png 1000w, http:\/\/bostontechguide.com\/wp-content\/uploads\/2015\/03\/growth-vcs-300x12.png 300w\" \/><\/p><\/div><p>&#13;\n\t\thttp:\/\/nextviewventures.com\t\t<\/p><h4>&#13;\n\t\t\tLOVE THAT DIRTY WATER\t\t<\/h4>&#13;<\/div><\/div><\/section><\/div><\/section>","created_at":"2018-11-07T16:49:29+0000","updated_at":"2018-11-07T16:49:29+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":19,"domain_name":"bostontechguide.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":475,"title":"MuttGuide\/UseGPG \u2013 Mutt","url":"https:\/\/dev.mutt.org\/trac\/wiki\/MuttGuide\/UseGPG","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:49:29+0000","updated_at":"2018-11-07T16:49:29+0000","annotations":[],"reading_time":0,"domain_name":"dev.mutt.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":473,"title":"Redshift is now part of Morar HPI | Insight for Growing Brands","url":"https:\/\/www.morarhpi.com\/redshift","content":"Redshift services are now part of the Morar HPI offering a comprehensive range of specialist market research and business consultancy services across a wide range of different markets and disciplines. Feel free to<a href=\"https:\/\/www.morarhpi.com\/services\" class=\"link-7\">review our services<\/a>or<a href=\"https:\/\/www.morarhpi.com\/contact\" class=\"link-8\">get in touch<\/a>\u00a0about potential projects.<p>If you have any questions or enquiries concerning past Redshift Research market research work please contact Paul Watts\u00a0<a href=\"mailto:paul.watts@morarhpi.com?subject=Redshift%20Research%20(legacy%20project)\" class=\"link-9\">here<\/a>.<\/p>","created_at":"2018-11-07T16:49:27+0000","updated_at":"2018-11-07T16:49:27+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.morarhpi.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":470,"title":"Using Offlineimap with Gnus","url":"https:\/\/nakkaya.com\/2010\/04\/10\/using-offlineimap-with-gnus\/","content":"<p>\nAs I rely more and more on my <a href=\"http:\/\/en.wikipedia.org\/wiki\/3G\">3G<\/a> connection, its high latency made my\nGnus experience more more unbearable. I wanted to keep using IMAP in\norder to use multiple clients while on the go, so I settled on using\n<a href=\"http:\/\/software.complete.org\/software\/projects\/show\/offlineimap\">OfflineIMAP<\/a>. OfflineIMAP will download all your mail and keep it in\nMaildir format any changes Gnus makes will be synchronized back to\nGMail and as a bonus you get a backup of all your emails. Even though\nboth OfflineIMAP and Gnus has been around for years they still don't\nplay well with each other, Gnus uses its own flags instead of using\nstandard Maildir flags so your changes aren't synchronized back to\nGMail, <a href=\"http:\/\/groups.google.com\/group\/linux.debian.user\/msg\/7594165a2b6d1c49\">someone<\/a> complained about this 6 years ago but sadly nothing has\nchanged. There is still hope though, using the script he posted with\nOfflineIMAP's pre\/post hooks we can sync Gnus flags to Maildir and\nvice versa.\n<\/p><div class=\"org-src-container\"><pre class=\"src src-conf\">[general]\nui = Noninteractive.Basic\naccounts = nakkaya\n[Account nakkaya]\nlocalrepository = nakkayaLocal\nremoterepository = nakkayaRemote\nautorefresh = 1\npresynchook = ~\/Projects\/scripts\/sync_nnmaildir -g\npostsynchook = ~\/Projects\/scripts\/sync_nnmaildir -m\n[Repository nakkayaLocal]\ntype = Maildir\nlocalfolders = ~\/Documents\/mail\/\n[Repository nakkayaRemote]\ntype = Gmail\nremoteuser = nurullah@nakkaya.com\nremotepass = super_secret_password\nholdconnectionopen = true\nkeepalive = 60\n#Setting realdelete = yes will Really Delete email from the server.\n#Otherwise \"deleting\" a message will just remove any labels and \n#retain the message in the All Mail folder.\nrealdelete = no\n<\/pre><\/div><p>\nSave this snippet as \".offlineimaprc\", change remote repository type to\nIMAP if you need to use this with some other IMAP server. Only trick here is\nto sync Gnus flags to Maildir before IMAP sync and sync Maildir flags to\nGnus flags after IMAP sync. Now run <i>offlineimap<\/i> while it\nsynchronizes your messages open your <b>~\/.gnus<\/b> set your select method\nto use nnmaildir,\n<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(setq gnus-select-method \n      '(nnmaildir \"GMail\" \n                  (directory \"~\/Documents\/mail\/\")\n                  (directory-files nnheader-directory-files-safe) \n                  (get-new-mail nil)))\n(define-key gnus-group-mode-map (kbd \"vo\")\n  '(lambda ()\n     (interactive)\n     (shell-command \"offlineimap&amp;\" \"*offlineimap*\" nil)))\n<\/pre><\/div><p>\nNow start Gnus, press ^ (gnus-group-enter-server-mode), open\nnnimap:GMail, for each folder press u\n(gnus-browse-unsubscribe-current-group) to toggle the subscription\nstatus until you\u2019re subscribed to the group. Then it should show up on\nthe group screen (M-x gnus). This also adds a new key binding to the\ngroup buffer, typing \"vo\" will start OfflineIMAP.\n<\/p><p>\nFor sending email using GMail, you need to install gnutls package, then\nadd the following to your .gnus,\n<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(load \"tls\")\n(setq send-mail-function 'smtpmail-send-it\n      message-send-mail-function 'smtpmail-send-it\n      starttls-use-gnutls t\n      starttls-gnutls-program \"\/opt\/local\/bin\/gnutls-cli\"\n      starttls-extra-arguments nil      \n      smtpmail-gnutls-credentials\n      '((\"smtp.gmail.com\" 587 nil nil))\n      ;; smtpmail-auth-credentials\n      ;;'((\"smtp.gmail.com\" 587 \"nurullah@nakkaya.com\" \"pass\" ))\n      smtpmail-starttls-credentials \n      '((\"smtp.gmail.com\" 587 \"nurullah@nakkaya.com\" nil))\n      smtpmail-default-smtp-server \"smtp.gmail.com\"\n      smtpmail-smtp-server \"smtp.gmail.com\"\n      smtpmail-smtp-service 587\n      smtpmail-debug-info t\n      smtpmail-local-domain \"nakkaya.com\")\n<\/pre><\/div><p>\nYou also need a .netrc file in your home directory with the following\ncontent,\n<\/p><pre class=\"example\">\nmachine smtp.gmail.com login nurullah@nakkaya.com password pass port 587\n<\/pre><p>\nWith this setup you should be able to send and receive email. This setup\nisn't perfect, an OS X bug makes OfflineIMAP crash while trying to download\nlarge emails, also it does crash every 10 hours or so but I can live\nwith that, on the plus side it makes Gnus blazing fast.\n<\/p><p>Tags: <a href=\"https:\/\/nakkaya.com\/tags\/#offlineimap\">offlineimap <\/a><a href=\"https:\/\/nakkaya.com\/tags\/#gnus\">gnus <\/a><\/p>","created_at":"2018-11-07T16:49:26+0000","updated_at":"2018-11-07T16:49:26+0000","published_by":["About"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"nakkaya.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":471,"title":"User feedback vs failure to launch: Am I being too cautious?","url":"https:\/\/www.reddit.com\/r\/startups\/comments\/7445i1\/user_feedback_vs_failure_to_launch_am_i_being_too\/","content":"<form action=\"#\" class=\"usertext warn-on-unload\" id=\"form-t3_7445i1qmr\"><div class=\"usertext-body may-blank-within md-container\"><div class=\"md\"><p>My founder and I have been obsessing over data since we came up with our idea and trying to figure out if there is a real need for our product in the market. While I believe, that in ideal circumstances we could scale these user feedbacks (mind you, we don't have a product) to any lengths we want, asking for the smallest details and finding out every single thing we can. However, the problem that arises with that is that more we tell ourselves that we need this and that data to start our business, the more time we take to launch.<\/p>\n<p>So my question would be is when is it enough to not just talk to the users but go out there and launch already? <\/p>\n<p>Is there an arbitrary rule, like, you've asked 100 people and now go launch. What was your experience when opening a startup? When did you feel that I've learnt enough to launch and now I'd rather learn the rest about my customers as I put my product in the market. <\/p>\n<\/div>\n<\/div><\/form>","created_at":"2018-11-07T16:49:26+0000","updated_at":"2018-11-07T16:49:26+0000","published_at":"2017-10-04T00:37:08+0000","published_by":["K-zi"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.reddit.com","preview_picture":"https:\/\/www.redditstatic.com\/new-icon.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":472,"title":"Site not found \u00b7 GitHub Pages","url":"https:\/\/japaric.github.io\/discovery\/","content":"<p><strong>There isn't a GitHub Pages site here.<\/strong><\/p><p>\n        If you're trying to publish one,\n        <a href=\"https:\/\/help.github.com\/pages\/\">read the full documentation<\/a>\n        to learn how to set up <strong>GitHub Pages<\/strong>\n        for your repository, organization, or user account.\n      <\/p><p><a href=\"https:\/\/status.github.com\">GitHub Status<\/a> \u2014\n        <a href=\"https:\/\/twitter.com\/githubstatus\">@githubstatus<\/a><\/p><a href=\"https:\/\/japaric.github.io\/\" class=\"logo logo-img-1x\">\n        <img width=\"32\" height=\"32\" title=\"\" alt=\"\" src=\"denied:data:image\/png;base64,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\" \/><\/a><a href=\"https:\/\/japaric.github.io\/\" class=\"logo logo-img-2x\">\n        <img width=\"32\" height=\"32\" title=\"\" alt=\"\" src=\"denied:data:image\/png;base64,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\" \/><\/a>","created_at":"2018-11-07T16:49:26+0000","updated_at":"2018-11-07T16:49:26+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"japaric.github.io","http_status":"404"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":468,"title":"Browser Extensions","url":"https:\/\/browserext.github.io\/browserext\/","content":"<p>\ufeff\n<\/p>\n    Browser Extensions<section id=\"abstract\"><p>\n            Modern browsers can add new functionality to sites and the browser itself through\n            script-based extensions. This document specifies an API set that allows developers\n            to create interoperable extensions for browsers that support the API set,\n            as well as the packaging format to be used for such extensions.\n        <\/p><\/section><section id=\"sotd\"><p>\n            If you wish to make comments regarding this document, please file issues on\n            <a href=\"https:\/\/github.com\/browserext\/browserext\/issues\/\">github<\/a>.\n            A <a href=\"mailto:public-browserext@w3.org\">public mailing list<\/a>\n            (<a href=\"mailto:public-browserext-request@w3.org?subject=subscribe\">subscribe<\/a>,\n            <a href=\"http:\/\/lists.w3.org\/Archives\/Public\/public-browserext\/\">archives<\/a>)\n            is also available,\n            but is reserved for high level discussions,\n            and is not appropriate for specific comments about this document.\n        <\/p><p>\n            Work on this document is governed by\n            <a href=\"http:\/\/browserext.github.io\/charter\/\">the charter<\/a> of the Browser Extension CG,\n            which includes among other things the <a href=\"http:\/\/browserext.github.io\/charter\/#communication\">Communication<\/a>,\n            <a href=\"http:\/\/browserext.github.io\/charter\/#decision-policy\">Decision<\/a>,\n            and <a href=\"http:\/\/browserext.github.io\/charter\/#contribution-mechanics\">Contribution<\/a> policies\n            of this community group.\n        <\/p><\/section><section class=\"informative\"><h2>Introduction<\/h2><p>\n            Because browsers supported different add-on models, extension authors wound up creating divergent codebases and delivering different browser-specific add-ons. This Community Group's goal is to make browser extension code much more interoperable across browsers by specifying common extension interfaces and well-defined browser behavior. This will allow extension authors to greatly reduce or eliminate the rework necessary to create extensions that target different browsers.\n        <\/p><\/section><section class=\"informative\"><h2>Core APIs<\/h2><p>\n            The goal for these core APIs is to provide functionality that extension authors need for: \n        <\/p><ul><li>page modification for web site content<\/li>\n            <li>UI augmentation for web browsers (host apps)<\/li>\n            <li>network request modification<\/li>\n            <li>passing data between extensions<\/li>\n            <li>passing data from Internet-hosted web pages to extensions<\/li>\n            <li>accessing user settings such as cookies<\/li>\n        <\/ul><p class=\"issue\">Need to resolve ID issue before we can address messaging ('passing data')<br \/>GitHub Issues <a href=\"https:\/\/github.com\/browserext\/browserext\/issues\/14\">14<\/a>\u00a0&amp;\u00a0<a href=\"https:\/\/github.com\/browserext\/browserext\/issues\/17\">17<\/a><\/p><section><h3 id=\"apiavailability\">API Availability<\/h3><section class=\"informative\"><h3>Overview<\/h3><p>\n                    Browser extensions are authored primarily with JavaScript. In addition to the usual <a href=\"https:\/\/www.w3.org\/standards\/webdesign\/script\">JavaScript Web APIs<\/a>, additional APIs are available exclusively to extensions via the <code>browser<\/code> object. However, the specific APIs that are available to an extension from the <code>browser<\/code> object are dependent upon:\n                <\/p><ul><li>Execution context (e.g. background\/event, popup, options, content, extension page)\n                    <\/li><li>Keys declared in the manifest (e.g. browser_action, page_action)\n                    <\/li><li>Permissions declared in the manifest (e.g. contextMenu, tabs, webNavigation, webRequest)\n                    <\/li><li>Content Security Policy (CSP) for the extension and target web page (for content scripts)\n                <\/li><\/ul><\/section><section class=\"informative\"><h4>Execution Context and Manifest Restrictions<\/h4><p>\n                        Browser extension pages and their corresponding scripts are executed in one of the following contexts. The manifest declarations required to create each specific type of page are described below.\n                    <\/p><ul><li>\n                            <b>Browser extension contexts (\"Window\")<\/b> - Most extension APIs are available to this context\n                            <ul><li><b>Background page<\/b> - use the \"background\" manifest key and specify \"persistent\" true <\/li>\n                                <li><b>Event page<\/b> - use the \"background\" manifest key and specify \"persistent\" false <\/li>\n                                <li><b>Popup page<\/b> - use the \"browser_action\" or \"page_action\" manifest key and specify \"default_popup\" <\/li>\n                                <li><b>Options page<\/b> - use the \"options_page\" manifest key <\/li>\n                                <li><b>Custom page<\/b> - such as \"browserext:\/\/&lt;ext_id&gt;\/myCustomPage.html\" <\/li>\n                            <\/ul><\/li>\n                        <li><b>Content script context<\/b> - A limited number of APIs are available to script that extensions use to interact with hosted web site pages\n                        <\/li><li><b>Internet-hosted web page context<\/b> - Messaging APIs can be used to send messages from hosted web site pages such as https:\/\/en.wikipedia.org\/wiki\/Main_Page\/index.html or https:\/\/w3c.github.io\/html\/index.html to extensions\n                    <\/li><\/ul><section><h5 id=\"availability-manifest\">Table: API Availability Summary<\/h5><p>\n                            Some browser extension API require a specific manifest key or specific permission. If the key or permission are not declared or not granted, the object will not be available for scripting. These requirements are summarized in the table below. If manifest keys that are not defined in this specification are specified, other browsers MAY elect to ignore those additional keys.\n                        <\/p><table><tr><th rowspan=\"2\">API object<\/th>\n                                <th colspan=\"3\">Execution Context<\/th>\n                                <th colspan=\"1\">Required Declaration<\/th>\n                            <\/tr><tr><th>Background page,Event page,Popup page,Options page,Custom page<\/th>\n                                <th>Content script page<\/th>\n                                <th>Internet-hosted web page<\/th>\n                                <th>Manifest key<\/th>\n                            <\/tr><tr><th><code>browserAction<\/code><\/th>\n                                <td>Y<sup><a href=\"#note1\">1<\/a><\/sup><\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\"browser_action\"<sup><a href=\"#note1\">1<\/a><\/sup><\/td>\n                            <\/tr><tr><th><code>contextMenus<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\"permissions\":[\"context_menus\"]<\/td>\n                            <\/tr><tr><th><code>extension<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\n                            <\/td><\/tr><tr><th><code>i18n<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\"default_lang\" : \"&lt;lang&gt;\"<sup><a href=\"#note4\">4<\/a><\/sup><\/td>\n                            <\/tr><tr><th><code>pageAction<\/code><\/th>\n                                <td>Y<sup><a href=\"#note1\">1<\/a><\/sup><\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\"page_action\"<sup><a href=\"#note1\">1<\/a><\/sup><\/td>\n                            <\/tr><tr><th><code>runtime<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>Y<\/td>\n                                <td>Y<sup><a href=\"#note3\">3<\/a><\/sup><\/td>\n                                <td>\u201cexternally_connectable\u201d:[...] <sup><a href=\"#note2\">2<\/a><\/sup><\/td>\n                            <\/tr><tr><th><code>tabs<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\"permissions\":[\"tabs\",\"activeTab\"]<\/td>\n                            <\/tr><tr><th><code>webNavigation<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\"permissions\":[\"webNavigation\"]<\/td>\n                            <\/tr><tr><th><code>webRequest<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\"permissions\":[\"webRequest\"]<\/td>\n                            <\/tr><tr><th><code>windows<\/code><\/th>\n                                <td>Y<\/td>\n                                <td>\n                                <\/td><td>\n                                <\/td><td>\n                            <\/td><\/tr><\/table><div id=\"note1\"><sup>1<\/sup>Either<code>browserAction<\/code>or<code>pageAction<\/code>is allowed at the same time, not both<\/div><div id=\"note2\"><sup>2<\/sup>The runtime object is available for use by extensions without declaring<code>\"externally_connectable\"<\/code>in the manifest. However, an extension will only receive messages sent from hosted web pages if this key is declared. The<code>\"externally_connectable\"<\/code>declaration will block messages from other extensions, unless the allowed extension ID is declared in the<code>\"ids\"<\/code>list.<\/div><div id=\"note3\"><sup>3<\/sup>Internet-hosted web pages that use<code>runtime.sendMessage<\/code>to communicate with extensions MUST be explicitly specify the hosting site's domain in the manifest under the \"externally_connectable\" key.<\/div><div id=\"note4\"><sup>4<\/sup>Use of the<code>browser.i18n<\/code>object does not require a manifest permission declaration. However, the manifest does require a value (<code>default_locale<\/code>) that is necessary and cannot be specified by the extension author in script code.<\/div><\/section><\/section><section><h4 id=\"web-idl\">Expressing Availability in WebIDL<\/h4><p>This specification uses WebIDL [[WEBIDL]] to describe the Browser Extensions API.<\/p><section><h4>Attribute: <code>Exposed<\/code><\/h4><p>\n                            To express the aforementioned conditional situations in which certain APIs are made available due to context, the following contexts are used:\n                        <\/p><ul><li><b>Window<\/b> - The primary context for most browser extension APIs. This applies to background, event, popup, options and custom contexts<\/li>\n                            <li><b>ContentScript<\/b> - Applies to content script context<\/li>\n                        <\/ul><\/section><section><h4>Extended Attribute: <code>CheckAnyPermissions<\/code><\/h4><p>\n                            To express the aforementioned conditional situations in which certain APIs are made available due to expressed manifest keys or permissions, the new WebIDL extended attribute <code>\"<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Mozilla\/WebIDL_bindings#CheckAnyPermissions\">CheckAnyPermissions<\/a>\"<\/code> is used.\n                        <\/p><table><tr><td><b>API object<\/b><\/td><td><b>CheckAnyPermissions Value<\/b><\/td><td><b>Manifest Key or Value<\/b><\/td><\/tr><tr><td>browser<\/td><td>\"browserExt\"<\/td><td>N\/A<\/td><\/tr><tr><td>browser.browserAction<\/td><td>\"browserExtBrowserAction\"<\/td><td>\"browserAction\":[...]<\/td><\/tr><tr><td>browser.contextMenu<\/td><td>\"browserExtContextMenus\"<\/td><td>\"permissions\":[\"contextMenus\"]<\/td><\/tr><tr><td>browser.extension<\/td><td>\"browserExt\"<\/td><td><\/td><\/tr><tr><td>i18n<\/td><td>\"browserExt\"<\/td><td>\"default_language\": \"&lt;code&gt;\"<\/td><\/tr><tr><td>pageAction<\/td><td>\"browserExtPageAction\"<\/td><td>\"pageAction\":[...]<\/td><\/tr><tr><td>runtime<\/td><td>\"browserExt\"<\/td><td><\/td><\/tr><tr><td>tabs<\/td><td>\"browserExtTabs\"<\/td><td>\"permissions\":[\"tabs\"] or<br \/>\"permissions\":[\"activeTab\"]<\/td><\/tr><tr><td>webNavigation<\/td><td>\"browserExtWebNavigation\"<\/td><td>\"permissions\":[\"webNavigation\"]<\/td><\/tr><tr><td>webRequest<\/td><td>\"browserExtWebRequest\"<\/td><td>\"permissions\":[\"webRequest\"]<\/td><\/tr><tr><td>windows<\/td><td>\"browserExt\"<\/td><td><\/td><\/tr><\/table><\/section><section class=\"example\"><h4>Example: <code>browser.i18n<\/code><\/h4><p>\n                            In this example, the <code>i18n<\/code> API should be made available to both browser host scripts (background, event, popup, options, etc.) and content scripts. Use of this API does not require a manifest permission declaration. Additionally, however, the manifest does require a value (<code>default_locale<\/code>) that is necessary and cannot be specified by the extension author in script code. This is expressed as a <code>BrowserExtManifest<\/code> dictionary. <\/p><p>NOTE: To avoid naming conflicts with WebIDL definitions elsewhere in this document, the Example below has an underscore (\"_\") appended to end of interface names. In the actual definitions elsewhere in this document, the underscore character is not used.\n                        <\/p><ol><li>\n                                First, declare the methods for the <code>browser.i18n<\/code> object.\n<pre class=\"idl\">\n[NoInterfaceObject]\ninterface BrowserExtI18n_ {\n    DOMString getMessage_(DOMString messageName, sequence&lt;DOMString?&gt; substitutions);\n};\n<\/pre>\n                            <\/li>\n                            <li>\n                                Next, we specify the available contexts (\"Window\" for background\/event\/popup\/options...) and \"ContentScript\" for the content context.\n<pre class=\"idl\">\n[NoInterfaceObject, Exposed=(Window,ContentScript)] \ninterface BrowserExtI18nAPI_ {\n    readonly attribute BrowserExtI18n_ i18n;\n};\n                        <\/pre>\n                            <\/li>\n                            <li>\n                                Connect the <code>i18n<\/code> object to the <code>browser.*<\/code> object.\n<pre class=\"idl\">\nBrowser implements BrowserExtI18nAPI_;\n<\/pre>\n                            <\/li>\n                            <li>\n                                Finally, we express the locale from manifest.json. For example <code>\"default_locale\": \"en\"<\/code>\n<pre class=\"idl\">\npartial dictionary BrowserExtManifest_ {\n    DOMString? defaultLocaleValue;\n};\n<\/pre>\n                            <\/li>\n                        <\/ol><\/section><\/section><section><h4 id=\"availability-csp\"><span>Considerations<\/span>Content Security Policy (CSP) Restrictions<\/h4><p>\n                    Use of some JavaScript APIs such as <code>eval()<\/code> are restricted due to CSP. If an explicit policy is not specified, the default policy is applied. The default CSP is discussed in the specification [[CSP3]] . If the browser extension author specifies a new CSP, it will override the default CSP. Refer to the CSP <a href=\"#csp-overview\">Overview<\/a> section for more information about the availability of these APIs. The <a href=\"#csp-cross-domain\">Cross-domain Privileges<\/a> and <a href=\"#availability-csp-content\">Content and Background Script Communication<\/a> sections provide details about API usage by content scripts.\n                <\/p><ul><li>Script execution via <code>eval()<\/code> and inline script execution are not allowed<\/li>\n                    <li>Only script from the extension's package can be executed<\/li>\n                    <li>Executing script from the web is not allowed<\/li>\n                <\/ul><\/section><\/section><section><h3 id=\"coreapis\">The <code>browser<\/code> object<\/h3><section><h4 id=\"browserAction\">The <code>browserAction<\/code> object<\/h4><\/section><section><section class=\"introductory\"><h4>WebIDL Definition<\/h4><p>\n        A description of the special <a href=\"#web-idl\">WebIDL syntax considerations<\/a> for browser extensions are defined elsewhere in this document.\n    <\/p><pre class=\"def idl\">\nenum BrowserExtContextType { \"all\", \"audio\", \"browser_action\", \"editable\", \"frame\", \"image\", \"link\", \"page\", \"page_action\", \"selection\", \"video\" };\nenum BrowserExtItemType { \"checkbox\",\"normal\",\"radio\",\"separator\" };\ndictionary BrowserExtContextMenuCreateDetails {\n    boolean checked;\n    sequence&lt;ContextType&gt;? contexts;\n    sequence&lt;DOMString&gt;? documentUrlPatterns;\n    boolean enabled;\n    DOMString? id;\n    Function onclick;\n    (long or DOMString?) parentId;\n    sequence&lt;DOMString&gt;? targetUrlPatterns;\n    DOMString? title;\n    DOMString? type;\n};\ndictionary BrowserExtContextMenuUpdateDetails {\n    boolean checked;\n    sequence&lt;BrowserExtContextType&gt;? contexts;\n    sequence&lt;DOMString&gt;? documentUrlPatterns;\n    boolean enabled;\n    Function onclick;\n    long parentId;\n    sequence&lt;DOMString&gt;? targetUrlPatterns;\n    DOMString? title;\n    ItemType? type;\n};\ndictionary BrowserExtContextMenuOnClickedDetails {\n    boolean checked;\n    boolean editable;\n    long frameId;\n    DOMString? frameUrl;\n    DOMString? linkUrl;\n    DOMString? mediaType;\n    (long or DOMString) menuItemId;\n    DOMString? pageUrl;\n    (long or DOMString?) parentMenuItemId;             \n    DOMString? selectionText;\n    DOMString? srcUrl;\n    boolean wasChecked;\n};\ncallback BrowserExtContextMenuOnClickedCallback = void (BrowserExtContextMenuOnClickedDetails details, BrowserExtTabsTab tab);\n[NoInterfaceObject]\ninterface BrowserExtContextMenus {\n    Promise&lt;long&gt; create(BrowserExtContextMenuCreateDetails details);\n    BrowserExtEvent onClicked(BrowserExtContextMenuOnClickedCallback callback);\n\tPromise&lt;void&gt; remove((long or DOMString) itemId);\n    Promise&lt;void&gt;  removeAll();\n    Promise&lt;void&gt;  update((long or DOMString) itemId, BrowserExtContextMenuUpdateDetails details);\n};\n[NoInterfaceObject, Exposed=Window, CheckAnyPermissions_browserExtContextMenus_]\ninterface BrowserExtContextMenusAPI {\n    readonly attribute BrowserExtContextMenus contextMenus; \n};\nBrowser implements BrowserExtContextMenusAPI;\n<\/pre><\/section><\/section><section><h4 id=\"i18n\">The <code>i18n<\/code> object<\/h4><\/section><section><h4 id=\"pageAction\">The <code>pageAction<\/code> object<\/h4><\/section><section><section><h4 id=\"tabs\">The <code>tabs<\/code> object<\/h4><section class=\"introductory\"><h4>WebIDL Definition<\/h4><p>\n                            A description of the special <a href=\"#web-idl\">WebIDL syntax considerations<\/a> for browser extensions are defined elsewhere in this document.\n                        <\/p><pre class=\"def idl\">\nenum BrowserExtTabMutedReasonDetails { \"user\", \"capture\", \"extension\" };\nenum BrowserExtRunAt { \"document_end\", \"document_idle\", \"document_start\" };\nenum BrowserExtTabStatus { \"complete\", \"loading\" };\nenum BrowserExtWindowTypes { \"app\", \"normal\", \"panel\", \"popup\" };\nenum BrowserExtTabsCaptureVisibleTabFormat { \"jpeg\", \"png\" };\ndictionary BrowserExtTabsTab {\n    boolean active;\n    boolean audible;\n    DOMString? favIconUrl;\n    boolean highlighted;\n    long height;\n    long id;\n    boolean incognito;\n    long index;\n    TabMutedDetails? mutedDetails;\n    long openerTabId;\n    boolean pinned;\n    DOMString? sessionId;\n    DOMString? status;\n    DOMString? title;\n    DOMString? url;\n    long width;\n    long windowId;\n};\ndictionary BrowserExtTabMutedDetails {\n    DOMString? extensionId;\n    boolean muted;\n    TabMutedReasonDetails? reason;\n};\ndictionary BrowserExtTabConnectDetails {\n    long frameId;\n    DOMString? name;\n};\ndictionary BrowserExtTabCreateDetails {\n    boolean active;\n    long index;\n    long openerTabId;\n    boolean pinned;\n    DOMString? url;\n    long windowId;\n};\ndictionary BrowserExtTabScriptAndCSSDetails {\n    boolean allFrames;\n    DOMString? code;\n    DOMString? file;\n    long frameId;\n    boolean matchAboutBlank;\n    BrowserExtRunAt? runAt;\n};\ndictionary BrowserExtTabsCaptureVisibleTabDetails {\n    BrowserExtTabsCaptureVisibleTabFormat imageCaptureFormat;\n    long jpegQuality;\n};\ndictionary BrowserExtTabQueryDetails {\n    boolean active;\n    boolean audible;  \n    boolean currentWindow;\n    boolean highlighted;\n    long index;\n    boolean lastFocusedWindow;\n    boolean muted;\n    boolean pinned;\n    TabStatus? status;\n    DOMString? title;\n    (DOMString? or sequence&lt;DOMString&gt;?) url;\n    long windowId;\n    WindowType? windowType;\n};\ndictionary BrowserExtTabReloadDetails {\n    boolean bypassCache;\n};\ndictionary BrowserExtTabSendMessageDetails {\n    long frameId;\n};\ndictionary BrowserExtTabUpdateDetails {\n    boolean active;\n    boolean highlighted;\n    boolean muted;\n    long openerTabId;\n    boolean pinned;\n    DOMString? url;\n};\ndictionary BrowserExtTabIdWindowId {\n    long tabId;\n    long windowId;\n};\ndictionary BrowserExtTabsWindowIdIsWindowClosing {\n    boolean isWindowClosing;\n    long windowId;\n};\ndictionary BrowserExtTabsOnUpdatedChangeDetails {\n    boolean audible;    \n    DOMString? favIconUrl;\n    TabMutedDetails? mutedDetails;\n    boolean pinned;\n    DOMString? status;\n    DOMString? title;\n    DOMString? url;\n};\ncallback BrowserExtTabsOnActivatedCallback = void (BrowserExtTabIdWindowId activeDetails);\ncallback BrowserExtTabsOnCreatedCallback = void (BrowserExtTabTab tab);\ncallback BrowserExtTabsOnRemovedCallback = void (long tabId, BrowserExtTabsWindowIdIsWindowClosing removeDetails);\ncallback BrowserExtTabsOnUpdatedCallback = void (long tabId, BrowserExtTabsOnUpdatedChangeDetails details, BrowserExtTabsTab tab);\n[NoInterfaceObject]\ninterface BrowserExtBrowserTabs {\n    Promise&lt;DOMString&gt; captureVisibleTab(long windowId, BrowserExtTabsCaptureVisibleTabDetails details);\n    RuntimePort connect(long tabId, BrowserExtTabConnectDetails details);\n    Promise&lt;BrowserExtTabsTab&gt; create(BrowserExtTabCreateDetails details);\n    void executeScript(long tabId, BrowserExtTabScriptAndCSSDetails details);\n    Promise&lt;BrowserExtTabsTab&gt; get(long tabId);\n    Promise&lt;BrowserExtTabsTab&gt; getCurrent();\n    void insertCSS(long tabId, BrowserExtTabScriptAndCSSDetails details);\n    void onActivated(BrowserExtTabsOnActivatedCallback callback);\n    void onCreated(BrowserExtTabsOnCreatedCallback callback);\n    void onRemoved(BrowserExtTabsOnRemovedCallback callback);\n    void onUpdated(BrowserExtTabsOnUpdatedCallback callback);\n    Promise&lt;BrowserExtTabsTab&gt; query(BrowserExtTabQueryDetails queryDetails);\n    Promise&lt;void&gt; reload(BrowserExtTabReloadDetails details);\n    Promise&lt;void&gt; remove((long or sequence&lt;long&gt;) tabIds);\n    Promise&lt;any&gt; sendMessage(long tabId, any message, optional BrowserExtTabSendMessageOptions details);\n    Promise&lt;BrowserExtTabsTab&gt; update(optional long tabId, BrowserExtTabUpdateDetails details);\n};\n[NoInterfaceObject, Exposed=Window, CheckAnyPermissions_browserExtTabs_]\ninterface BrowserExtTabsAPI {\nreadonly attribute BrowserExtTabs tabs; \n};\nBrowser implements BrowserExtTabsAPI;\n<\/pre><\/section><\/section><section><section class=\"introductory\"><h4>WebIDL Definition<\/h4><p>\n                            A description of the special <a href=\"#web-idl\">WebIDL syntax considerations<\/a> for browser extensions are defined elsewhere in this document.\n                        <\/p><pre class=\"def idl\">\nenum BrowserExtTransitionType { \"link\", \"typed\" };\ndictionary BrowserExtWebNavigationOnBeforeNavigateDetails {\n    long frameId;\n    long parentFrameId;\n    long tabId;\n    double timeStamp;\n    DOMString url;\n};\ndictionary BrowserExtWebNavigationOnCommittedDetails {\n    long frameId;\n    long processId;\n    long tabId;\n    double timeStamp;\n    TransitionType transitionType;\n    DOMString url;\n};\ndictionary BrowserExtWebNavigationOnCompletedDetails {\n    long frameId;\n    long processId;\n    long tabId;\n    double timeStamp;\n    DOMString url;\n};\ndictionary BrowserExtWebNavigationOnDOMContentLoadedDetails {\n    long frameId;\n    long processId;\n    long tabId;\n    double timeStamp;\n    DOMString url;\n};\ndictionary BrowserExtWebNavigationOnErrorOccurredDetails {\n    DOMString error;\n    long frameId;\n    long tabId;\n    double timeStamp;\n    DOMString url;\n};\ndictionary BrowserExtWebNavigationOnReferenceFragmentUpdatedDetails {\n    long frameId;\n    long processId;\n    long tabId;\n    double timeStamp;                           \n    BrowserExtTransitionType transitionType;\n    DOMString url;\n};\ncallback BrowserExtWebNavigationOnBeforeNavigateCallback = void (BrowserExtWebNavigationOnBeforeNavigateDetails details);\ncallback BrowserExtWebNavigationOnCommittedCallback = void (BrowserExtWebNavigationOnCommittedDetails details);\ncallback BrowserExtWebNavigationOnCompletedCallback = void (BrowserExtWebNavigationOnCompletedDetails details);\ncallback BrowserExtWebNavigationOnDOMContentLoadedCallback = void (BrowserExtWebNavigationOnDOMContentLoadedDetails details);\ncallback BrowserExtWebNavigationOnErrorOccurredCallback = void (BrowserExtWebNavigationOnErrorOccurredDetails details);\ncallback BrowserExtWebNavigationOnReferenceFragmentUpdated = void (BrowserExtWebNavigationOnReferenceFragmentUpdatedDetails details);\n[NoInterfaceObject]\ninterface BrowserExtBrowserWebNavigation {\n    void onBeforeNavigate(BrowserExtWebNavigationOnBeforeNavigateCallback callback);\n    void onCommitted(BrowserExtWebNavigationOnCommittedCallback callback);\n    void onCompleted(BrowserExtWebNavigationOnCompletedCallback callback);\n    void onDOMContentLoaded(BrowserExtWebNavigationOnDOMContentLoadedCallback callback);\n    void onErrorOccurred(BrowserExtWebNavigationOnErrorOccurredCallback callback);\n    void onReferenceFragmentUpdated(BrowserExtWebNavigationOnReferenceFragmentUpdated callback);\n};\n    \n[NoInterfaceObject, Exposed=Window, CheckAnyPermissions_browserExtWebNavigation_]\ninterface BrowserExtWebNavigationAPI {\n    readonly attribute BrowserExtWebNavigation webNavigation; \n};\nBrowser implements BrowserExtWebNavigationAPI;\n<\/pre><\/section><\/section><section><h4 id=\"webRequest\">The <code>webRequest<\/code> object<\/h4><section class=\"introductory\"><h4>WebIDL Definition<\/h4><p>\n                            A description of the special <a href=\"#web-idl\">WebIDL syntax considerations<\/a> for browser extensions are defined elsewhere in this document.\n                        <\/p><pre class=\"def idl\">\nenum BrowserExtResourceType { \"font\", \"image\", \"main_frame\", \"object\", \"other\", \"ping\", \"script\", \"stylesheet\", \"sub_frame\", \"xmlhttprequest\" };\ndictionary BrowserExtWebRequestUploadDetails {\n    any bytes;\n    DOMString? file;\n};\ndictionary BrowserExtWebRequestRequestBody {\n    DOMString? error;\n    object? formData;\n    sequence&lt;BrowserExtWebRequestUploadData&gt;? rawData;\n};\ndictionary BrowserExtWebRequestHttpHeader {\n    DOMString keyName;\n    any value;\n};\ndictionary BrowserExtWebRequestHttpHeaders {\n    sequence&lt;BrowserExtWebRequestHttpHeader&gt; data; \n};\ndictionary BrowserExtWebRequestOnBeforeRequestDetails {\n    long frameId;\n    DOMString method;\n    long parentFrameId;\n    BrowserExtWebRequestRequestBody? requestBody;\n    DOMString requestId;\n    long tabId;\n    double timeStamp;\n    BrowserExtResourceType type;\n    DOMString url;\n};\ndictionary BrowserExtWebRequestOnBeforeSendHeadersDetails {\n    long frameId;\n    DOMString method;\n    long parentFrameId;\n    BrowserExtWebRequestHttpHeaders? requestHeaders;\n    DOMString requestId;\n    long tabId;\n    double timeStamp;\n    BrowserExtResourceType type;\n    DOMString url;\n};\ndictionary BrowserExtWebRequestOnCompletedDetails {\n    long frameId;\n    boolean fromCache;\n    DOMString method;\n    long parentFrameId;\n    DOMString requestId;\n    BrowserExtWebRequestHttpHeaders? responseHeaders;\n    DOMString? serverIP;\n    long statusCode;\n    DOMString statusLine;\n    long tabId;\n    double timeStamp;   \n    BrowserExtResourceType type;\n    DOMString url;\n}; \ndictionary BrowserExtWebRequestOnHeadersReceivedDetails {\n    long frameId;\n    DOMString method;\n    long parentFrameId;\n    DOMString requestId;\n    BrowserExtWebRequestHttpHeaders? responseHeaders;\n    long statusCode;\n    DOMString statusLine;\n    long tabId;\n    double timeStamp;   \n    BrowserExtResourceType type;\n    DOMString url;\n};\ndictionary BrowserExtWebRequestOnSendHeadersDetails {\n    long frameId;\n    DOMString method;\n    long parentFrameId;\n    BrowserExtWebRequestHttpHeaders? requestHeaders;\n    DOMString requestId;\n    long tabId;\n    double timeStamp;   \n    BrowserExtResourceType type;\n    DOMString url;\n};\ncallback BrowserExtWebRequestOnBeforeRequestCallback = void (BrowserExtWebRequestOnBeforeRequestDetails details);\ncallback BrowserExtWebRequestOnBeforeSendHeadersCallback = void (BrowserExtWebRequestOnBeforeSendHeadersDetails details);\ncallback BrowserExtWebRequestOnCompletedCallback = void (BrowserExtWebRequestOnCompletedDetails details);\ncallback BrowserExtWebRequestOnHeadersReceivedCallback = void (BrowserExtWebRequestOnHeadersReceivedDetails details);\ncallback BrowserExtWebRequestOnSendHeadersCallback = void (BrowserExtWebRequestOnSendHeadersDetails details);\n[NoInterfaceObject]\ninterface BrowserExtBrowserWebRequest {\n    void onBeforeRequest(BrowserExtWebRequestOnBeforeRequestCallback callback);\n    void onBeforeSendHeaders(BrowserExtWebRequestOnBeforeSendHeadersCallback callback);\n    void onCompleted(BrowserExtWebRequestOnCompletedCallback callback);\n    void onHeadersReceived(BrowserExtWebRequestOnHeadersReceivedCallback callback);\n    void onSendHeaders(Function callback);\n};\n[NoInterfaceObject, Exposed=Window, CheckAnyPermissions_browserExtWebRequest_]\ninterface BrowserExtWebRequestAPI {\n    readonly attribute BrowserExtWebRequest webRequest; \n};\nBrowser implements BrowserWebRequestAPI;\n<\/pre><\/section><\/section><section><h4 id=\"windows\">The <code>windows<\/code> object<\/h4><section class=\"introductory\"><h4>WebIDL Definition<\/h4><p>\n                            A description of the special <a href=\"#web-idl\">WebIDL syntax considerations<\/a> for browser extensions are defined elsewhere in this document.\n                        <\/p><pre class=\"def idl\">\nenum BrowserExtWindowsWindowType { \"detached_panel\", \"normal\", \"panel\", \"popup\" };\nenum BrowserExtWindowsWindowState { \"docked\", \"fullscreen\", \"maximized\", \"minimized\", \"normal\" };\ndictionary BrowserExtCreateWindowDetails {\n    boolean focused;\n    long height;\n    boolean incognito;\n    long left;\n    BrowserExtWindowsWindowState? state;\n    long tabId;\n    long top;\n    (DOMString? or sequence&lt;DOMString&gt;?) url;\n    long width;\n    BrowserExtWindowsWindowType? windowType;\n};\ndictionary BrowserExtGetWindowDetails {\n    boolean populate;\n    sequence&lt;BrowserExtWindowsWindowType&gt;? windowTypes;\n};\ndictionary BrowserExtWindowUpdateDetails {\n    boolean drawAttention;\n    boolean focused;\n    long height;\n    long left;\n    long top;\n    long width;\n    BrowserExtWindowState? windowState;\n};\ndictionary BrowserExtWindowsWindow {\n    boolean alwaysOnTop;\n    boolean focused;\n    long height;\n    long id;\n    boolean incognito;\n    long left;\n    sequence&lt;BrowserExtTabsTab&gt;? tabs;\n    long top;\n    DOMString? sessionId;\n    BrowserExtWindowsWindowState? state;\n    long width;\n    BrowserExtWindowsWindowType? windowType;\n};\ncallback BrowserExtWindowsOnFocusChangedCallback = void (long windowId);\n[NoInterfaceObject]\ninterface BrowserExtWindows {\n    Promise&lt;BrowserExtWindowsWindow&gt; create(CreateWindowDetails details);\n    Promise&lt;BrowserExtWindowsWindow&gt; get(long windowId, optional BrowserExtGetWindowDetails details);\n    Promise&lt;sequence&lt;BrowserExtWindowsWindow&gt;&gt; getAll(optional BrowserExtGetWindowDetail details);\n    Promise&lt;BrowserExtWindowsWindow&gt; getCurrent(optional BrowserExtGetWindowDetail details);\n    Promise&lt;BrowserExtWindowsWindow&gt; getLastFocused(optional BrowserExtGetWindowDetail details);\n    void onFocusChanged(BrowserExtWindowsOnFocusChangedCallback callback);\n    Promise&lt;BrowserExtWindowsWindow&gt; update(long windowId, optional BrowserExtWindowUpdateDetails details);\n};\n    \n[NoInterfaceObject, Exposed=Window, CheckAnyPermissions_browserExt_]\ninterface BrowserExtWindowsAPI {\nreadonly attribute BrowserExtWindows windows; \n};\nBrowser implements BrowserExtWindowsAPI;\n<\/pre><\/section><\/section><\/section><section class=\"informative\"><h3>Native Messaging<\/h3><p>Native Messaging is covered in another specification. Refer to the Native Messaging for Browser Extensions specification [[browserext-native]] for more information.<\/p><\/section><\/section><section><h3>Events<\/h3><section><h4>Overview<\/h4><p>\n                Extension events are similar to DOM Events [<a href=\"https:\/\/www.w3.org\/TR\/dom\/\">DOM4<\/a>]. However, extension events are dispatched only to objects in the relevant extension scope (background, event, options, content, etc.). Script in the extension scope can observe events by calling <code>addListener()<\/code>.\n            <\/p><pre class=\"def idl\">\n[NoInterfaceObject, Exposed=(Window,ContentScript), CheckAnyPermissions_browserExt_]\ninterface BrowserExtEvent {\n    void    addListener(Function callback);\n    void    removeListener(Function callback);\n    boolean hasListener(Function callback);\n    boolean hasListeners();\n};\n<\/pre><\/section><\/section><section><h2 id=\"message-strings\">Message Strings<\/h2><section><h4>Overview<\/h4><p>Message strings are used in the <code>manifest.json<\/code> file as well <code>browser.i18n.getMessage()<\/code>. <\/p><\/section><section><h4>Predefined messages<\/h4><p>Predefined messages are accessed in the same way as other messages. The syntax for standard messages is:<\/p><code>__MSG_extensionName__<\/code><p>Predefined messages use exactly the same syntax, except with @@ before the message name. For example:<\/p><code>__MSG_@@ui_locale__<\/code><p>The following table shows the different available predefined messages:<\/p><table><tr><td><b>Message name<\/b><\/td><td><b>Description<\/b><\/td><\/tr><tr><td>@@extension_id<\/td><td>The extension or app ID; you might use this string to construct URLs for resources inside the extension. Even unlocalized extensions can use this message. Note that you can't use this message in a manifest file.<\/td><\/tr><tr><td>@@ui_locale<\/td><td>The current locale; you might use this string to construct locale-specific URLs.<\/td><\/tr><tr><td>@@bidi_dir<\/td><td>The text direction for the current locale, either \"ltr\" for left-to-right languages such as English or \"rtl\" for right-to-left languages such as Arabic.<\/td><\/tr><tr><td>@@bidi_reversed_dir<\/td><td>If the @@bidi_dir is \"ltr\", then this is \"rtl\"; otherwise, it's \"ltr\".<\/td><\/tr><tr><td>@@bidi_start_edge<\/td><td>If the @@bidi_dir is \"ltr\", then this is \"left\"; otherwise, it's \"right\".<\/td><\/tr><tr><td>@@bidi_end_edge<\/td><td>If the @@bidi_dir is \"ltr\", then this is \"right\"; otherwise, it's \"left\".<\/td><\/tr><\/table><\/section><\/section><section><h2 id=\"manifest-json\">Manifest Format<\/h2><section><h3>Pre-parsing<\/h3><p>Standards-compliant browsers are expected to ignore \"\/\/\" comments. Disregarding comments, the manifest.json file format for browser extensions is expected to be fully JSON compliant. Malformed JSON files are not supported.  If manifest keys that are not defined in this specification are specified, other browsers MAY ignore those keys.<\/p><\/section><section><h3>List of Keys<\/h3><p>The full list of supported manifest keys is provided below:<\/p><table class=\"simple\"><tr><th style=\"width: 100px;\">Key<\/th><th>Required<\/th><th>Details<\/th><\/tr><tr><td style=\"width: 100px;\"><code>\"name\"<\/code><\/td><td>Required<\/td><td>The name of the extension. <\/td><\/tr><tr><td><code>\"version\"<\/code><\/td><td>Required<\/td><td>A string representation of a number<\/td><\/tr><tr><td><code>\"default_locale\"<\/code><\/td><td>Required if extension uses locales. Not allowed if extension does not use locales<\/td><td>See <a href=\"#message-strings\">Message Strings<\/a> and <a href=\"#i18n\"><code>browser.i18n<\/code><\/a> for more information about the use of locales and localized resources.<\/td><\/tr><tr><td><code>\"description\"<\/code><\/td><td>Optional<\/td><td>Describes the capabilities of the extension.<\/td><\/tr><tr><td><code>\"icons\" : {<br \/>\u00a0\u00a0\u00a0\"&lt;size&gt;\" : \"&lt;filename.jpg&gt;\"<br \/>}<\/code><\/td><td>Optional<\/td><td>A list of pairs of pixel dimensions and JPEG filenames. These are typically used to display visuals for the extension in browser management UI for extensions. Browsers MAY require specific sizes. For example,<br \/><code>\"icons\":{<br \/>\u00a0\u00a0\u00a0\"16\":\"small.jpg\",<br \/>\u00a0\u00a0\u00a0\"32\":\"medium.jpg\"<br \/>}<br \/><\/code> would be used for a small JPEG with height of 16 pixels and width of 16 pixels (small.jpg), and a second JPEG 32x32 pixels in size (medium.jpg).<\/td><\/tr><tr><td><code>\"developer\": {<br \/>\u00a0\u00a0\u00a0\"name\": \"&lt;Company&gt;\",<br \/>\u00a0\u00a0\u00a0\"url\": \"&lt;Website&gt;\"<br \/>}<\/code><\/td><td>Optional<\/td><td>Includes subkeys for <code>\"name\"<\/code> and <code>\"url\"<\/code> that provide information about the entity that authored the extension, typically a company or programmer name. The <code>\"url\"<\/code> subkey SHOULD refer to the company\/programmer site, not a description of the extension.<\/td><\/tr><tr><td><code>\"browser_action\"<\/code><\/td><td>Optional<p>Note that some browsers only permit <code>\"browser_action\"<\/code> or <code>\"page_action\"<\/code>, but not both at the same time<\/p><\/td><td>See <a href=\"#browserAction\">browser.browserAction<\/a> for more information about the subkeys for <code>\"browser_action\"<\/code>.<\/td><\/tr><tr><td><code>\"page_action\"<\/code><\/td><td>Optional<p>Note that some browsers only permit <code>\"browser_action\"<\/code> or <code>\"page_action\"<\/code>, but not both at the same time<\/p><\/td><td>See <a href=\"#pageAction\">browser.pageAction<\/a> for more information about the subkeys for <code>\"page_action\"<\/code>.<\/td><\/tr><tr><td><code>\"browser_specific_settings\":{<br \/>\u00a0\u00a0\u00a0&lt;browser_name&gt;\": {<br \/>\u00a0\u00a0\u00a0\"\u00a0\u00a0\u00a0\"&lt;key&gt;\": \"&lt;value&gt;\"<br \/>\u00a0\u00a0\u00a0}<br \/>}<\/code><\/td><td>Optional<\/td><td>Describes specific values that MAY be considered by a browser. For example, a particular extension may require version 42 or greater of the Gecko browser in order to work properly:<br \/><code>\"browser_specific_settings\":{<br \/>\u00a0\u00a0\u00a0\"gecko\": {<br \/>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"strict_min_version\": \"42.0\"<br \/>\u00a0\u00a0\u00a0}<br \/>}<\/code><br \/><\/td><\/tr><tr><td><code>\"background\" : {<br \/>\u00a0\u00a0\u00a0\"page\": \"&lt;Page URL&gt;\",<br \/>\u00a0\u00a0\u00a0\"scripts\": [...],<br \/>\u00a0\u00a0\u00a0\"persistent\": boolean,<br \/>}\n<\/code><\/td><td>Optional<p>If <code>\"background\"<\/code> is specified, <code>\"persistent\"<\/code> is required.<\/p><p>Other subkeys are optional<\/p><\/td><td>Automatically loads a page or script and begins execution, regardless of the page being viewed in a tab.<\/td><\/tr><tr><td><code>\"content_scripts\" : {<br \/>\u00a0\u00a0\u00a0\"all_frames\": &lt;boolean&gt;,<br \/>\u00a0\u00a0\u00a0\"css\": [...],<br \/>\u00a0\u00a0\u00a0\"exclude_matches\": [...],<br \/>\u00a0\u00a0\u00a0\"js\": [...],<br \/>\u00a0\u00a0\u00a0\"matches\": [...],<br \/>\u00a0\u00a0\u00a0\"run_at\": &lt;see_details&gt;,<br \/>}\n<\/code><\/td><td>Optional<\/td><td>Automatically adds scripts (from the <code>\"js\"<\/code> subkey list) or CSS (from the <code>\"css\"<\/code> subkey list) to each page within a tab.<p>If <code>all_frames<\/code> is <code>false<\/code>, the string patterns from <code>matches<\/code> and <code>exclude_matches<\/code> are used accordingly to determine which page content is modified or excluded.<\/p><p>Valid values for <code>\"run_at\"<\/code> are <code>\"document_start\", \"document_end\"<\/code> and <code>\"document_idle\"<\/code>.<\/p><\/td><\/tr><tr><td><code>\"content_security_policy\"<\/code><\/td><td>Optional<\/td><td>A string representation of the new policy to override the default policy. See <a href=\"#csp-overview\">Content Security Policy (CSP)<\/a> for more information.<\/td><\/tr><tr><td><code>\"options_page\"<\/code><\/td><td>Optional<\/td><td>The page within the extension that allows extension users to modify settings or options for the extension.<\/td><\/tr><tr><td><code>\"manifest_version\"<\/code><\/td><td>Optional<\/td><td>This is not used<\/td><\/tr><tr><td><code>\"required_keys\":[...]<\/code><\/td><td>Optional<\/td><td>See the <a href=\"#required-keys\"><code>required_keys<\/code> key<\/a> section for more information about this key.<\/td><\/tr><tr><td><code>\"web_accessible_resources\":[...]<\/code><\/td><td>Optional<\/td><td>Resources listed here may be accessed by Internet-hosted site pages.<\/td><\/tr><tr><td><code>\"externally_connectable\": [<br \/>\u00a0\u00a0\u00a0\"ids\": [...],<br \/>\u00a0\u00a0\u00a0\"matches\": [...],<br \/>]<\/code><\/td><td>Optional<\/td><td>Other extensions or Internet-hosted site pages may send messages to this extension if their extension ID (<code>\"ids\"<\/code>) or page URL pattern (<code>\"matches\"<\/code>) are specified.<\/td><\/tr><tr><td><code>\"permissions\" : {<br \/>\u00a0\u00a0\u00a0\"activeTab\",<br \/>\u00a0\u00a0\u00a0\"contextMenus\",<br \/>\u00a0\u00a0\u00a0\"tabs\",<br \/>\u00a0\u00a0\u00a0\"webNavigation\",<br \/>\u00a0\u00a0\u00a0\"webRequestBlocking\",<br \/>\u00a0\u00a0\u00a0\"&lt;url-pattern&gt;\",<br \/>}<\/code><\/td><td>Optional <\/td><td>Each subkey must be specified in order to use the associated API.<br \/><ul><li><code>activeTab...<a href=\"#tabs\">browser.tabs<\/a><\/code> (<a href=\"#active-tab\">activeTab<\/a>)<\/li>\n<li><code>contextMenus...<a href=\"#contextMenus\">browser.contextMenus<\/a><\/code><\/li>\n<li><code>tabs...<a href=\"#tabs\">browser.tabs<\/a><\/code><\/li>\n<li><code>webNavigation...<a href=\"#webNavigation\">browser.webNavigation<\/a><\/code><\/li>\n<li><code>webRequest...<a href=\"#webRequest\">browser.webRequest<\/a><\/code><\/li>\n<li><code>webRequestBlocking...<a href=\"#webRequest\">browser.webRequest <\/a><\/code>(blocking)<\/li>\n<\/ul>\nPatterns listed in <code>&lt;url-pattern&gt;<\/code> allow additional permissions for URLs that match those patterns. See <a href=\"#all-urls\">Permissions URL patterns<\/a> for more information. \n                                                          <\/td><\/tr><\/table><\/section><section><h3 id=\"all-urls\">Permissions URL patterns<\/h3><p> The <code>\"permissions\"<\/code> key may include match patterns that follow the structure <code>&lt;scheme&gt;:\/\/&lt;host&gt;\/&lt;path&gt;<\/code>, such as <code>\"*:\/\/*.w3.org\/*\"<\/code>. The special \n                pattern <code>\"&lt;all_urls&gt;\"<\/code> is also permitted. The match patterns identify a group of URLs for which the extension is requesting extra privileges:<\/p><ul><li>XHR access to those origins<\/li>\n                <li>the ability to inject scripts with <code>tabs.executeScript<\/code> into pages from those origins<\/li>\n                <li>the ability to receive events from the <code>webRequest<\/code> API for those origins<\/li>\n            <\/ul><\/section><section><h3 id=\"active-tab\">The <code>activeTab<\/code> permission<\/h3><p> The <code>\"activeTab\"<\/code> permission enables the use of <code>browser.tabs<\/code> API only if the user has actively initiated an action such as the <code>browserAction<\/code>. The <code>browser.tabs<\/code> API can be used until the tab is navigated to another page or closed.<\/p><p>Passive background-only extensions cannot use the <code>browser.tabs<\/code> API unless both the <code>\"tabs\"<\/code> permission and a URL pattern (e.g. <code>\"&lt;all_urls&gt;\"<\/code> or <code>\"http:\/\/*\"<\/code> or <code>\"*:\/\/&amp;.w3.org\/*\"<\/code>) are explicitly declared in <code>manifest.json<\/code>. <\/p><\/section><section><h3 id=\"required-keys\">The <code>required_keys<\/code> key<\/h3><p>This key is optional, but if it is specified, it provides a list of strings representing manifest keys that MUST be supported. If the browser loading the extension does not support the specified key, the browser MUST reject the manifest and not load the extension.<\/p><\/section><section class=\"informative\"><h3>Full Example<\/h3><pre>\n\/\/ Standards-compliant browsers are expected to ignore \"\/\/\" comments. Disregarding comments, \n\/\/ the manifest.json file format for browser extensions is expected to be fully JSON compliant. \n\/\/ Malformed JSON files are not supported.\n\/\/\n\/\/ Other manifest keys that are well-formed JSON but are not listed here MUST be ignored.\n\/\/\n\/\/ Note that some fields marked as Optional here are required by vendor-specific distribution Stores.\n{\n  \"name\": \"The Name of Your Extension\",             \/\/ Required\n  \"version\": \"Your Extension Version\",              \/\/ Required\n  \"default_locale\": \"en\",                           \/\/ Required if locales are used. Otherwise, not allowed\n  \"description\": \"Description for Your Extension\",   \n  \"icons\": {...},                                    \n  \"developer\": {                                    \n       \"name\": \"Your Name or Company\",              \n       \"url\": \"Company Website\"                      \n  }\n  \/\/ Note: Some browsers require an extension to use either browser_action or page_action, but not both simultaneously\n  \"browser_action\": {  \n      \"default_icon\": {...},                        \/\/ Same format as \"icons\"\n      \"default_popup\": \"Page URL\",\n      \"default_title\": \"Title string\"\n  },\n  \"page_action\": {...},                             \/\/ Same format as \"browser_action\"\n  \n  \"browser_specific_settings\": {                    \n       \"&lt;browser_name&gt;\": {                          \/\/ Examples \"gecko\",\"opera\",\"edge\"\n            \"&lt;key&gt;\": \"&lt;value&gt;\"                      \/\/ Examples \"strict_min_version\": \"42.0\", \"id\": \"addon@example.com\"\n       }\n  },\n  \"background\": {                                   \n      \"page\": \"Page URL\",                           \/\/ Only one of \"page\" or \"scripts\" is supported, but not both simultaneously\n      \"scripts\": [],                                \/\/ Only one of \"page\" or \"scripts\" is supported, but not both simultaneously\n      \"persistent\": false                           \/\/ Required if \"background\" is specified\n  },\n  \"content_scripts\": {                                                    \n      \"all_frames\": false,                          \n      \"css\": [],                                   \n      \"exclude_matches\": [],                        \n      \"js\": [],                                     \n      \"matches\": [],                                \n      \"run_at\" : \"document_start\"                   \/\/ Also \"document_end\", \"document_idle\"\n  }\n  \"content_security_policy\": \"&lt;policy-string&gt;\",     \n  \n  \"options_page\": \"Page URL\",                            \/\/ Optional\n  \"manifest_version\": 2,                            \/\/ Not used\n  \"required_keys\": [],                              \/\/ If a browser does not recognize a key in this list, it MUST \n                                                    \/\/ reject the manifest (e.g. \"sidebar_action\")\n                                                    \/\/ FEEDBACK: Rename to \"required_capabilities\"\n  \"permissions\": {                                  \n       \"activeTab\",                                                \n       \"contextMenus\",                              \n       \"tabs\",                                      \n       \"webNavigation\",                            \n       \"webRequest\",                                \n       \"webRequestBlocking\",                        \n       \"&lt;url-pattern&gt;\"                              \/\/ Examples \"http:\/\/*\/*\", \"&lt;all_urls&gt;\"\n  },\n  \"web_accessible_resources\": [...]    \n  \"externally_connectable\" : [\n       \"ids\":[...], \/\/ Identifiers for other extensions that are allowed to send messages\n       \"matches\":[...] \/\/ URL patterns for hosted web pages allowed to send messages \n  ]                             \n}\n<\/pre><\/section><\/section><section><h2>Extension Resources<\/h2><section><h3>The <code>browserext:\/\/<\/code> protocol<\/h3><p>\n                To access resources within the extension, such as <code>browserext:\/\/&lt;ext_id&gt;\/options.html<\/code>, the <code>browserext:\/\/<\/code> protocol is used.\n            <\/p><\/section><section><h3>Extension IDs<\/h3><p>\n                Each extension has an extension ID that follows the <code>browserext:\/\/<\/code> protocol. For example\n            <\/p><p class=\"example\">\n                <code>browserext:\/\/MyExtension_c1wakc4j0nefm\/options.html<\/code>\n                <code>browserext:\/\/dfcijpibodeoenkablikbkiobbdnkfki\/options.html<\/code>\n            <\/p><p>\n                The algorithms that generate these IDs are different for each browser. To access these resources, do not hardcode the ID generated by a particular browser. Instead, use the <code>runtime.getURL()<\/code> method to convert a relative file name or path to the absolute name or path, which includes the extension ID.\n            <\/p><\/section><\/section><section><h2>Content Security Policy (CSP)<\/h2><section><h5 id=\"csp-overview\">Overview<\/h5><p>Content Security Policy (CSP) is intended to prevent inadvertant execution of malicious code. The default CSP for web content is discussed in the specification [[CSP3]] . For extensions, the default CSP is <\/p><pre>\"script-src 'self'; object-src 'self';\"<\/pre><p>This means:<\/p><ul><li>\n                    <b>Only <code>&lt;script&gt;<\/code> and <code>&lt;object&gt;<\/code> resources that are included in the extension package may be accessed<\/b>\n                    <p>If the extension must access additional script libraries such as jQuery, copies of the .js files must be included in the extension package. <\/p><p>Alternatively, an extension may loosen the restriction and allow a remote file to be accessed. In this case, the origin for the hosted .js file may be specified in the <code>\"content_security_policy\"<\/code> key in <code>manifest.json<\/code>. For example:\n                    <\/p><p class=\"example\"><code>\"content_security_policy\" : \"script-src 'self' https:\/\/*.jquery.com; object-src 'self'\"<\/code><\/p>\n                <\/li>\n                <li>\n                    <b>The extension will not evaluate strings as JavaScript<\/b>\n                    <p>The following are restricted by default:<\/p>\n                    <ul><li><code>eval(\"script_goes_here;\");<\/code><\/li>\n                        <li><code>window.setTimeout(\"script_goes_here;\", t);<\/code><\/li>\n                        <li><code>var f = new Function(\"script_goes_here;\");<\/code><\/li>\n                    <\/ul><p>To enable the use of these evaluators, the CSP may be modified to include <code>'unsafe-eval'<\/code> as follows:<\/p>\n                    <p class=\"example\"><code>\"content_security_policy\": \"script-src 'self' 'unsafe-eval'; object-src 'self';\"<\/code><\/p>\n                <\/li>\n                <li>\n                    <b>Inline JavaScript is not executed<\/b>\n                    <p>Inline <code>&lt;script&gt;<\/code> tags and inline event handlers (e.g. <code>&lt;div onclick=\"script_goes_here;\"&gt;<\/code>) are silently ignored. To enable the use of inline script, a SHA256 hash of the entire tag may be added to the <code>\"content_security_policy\"<\/code> key in <code>manifest.json<\/code>. For example, replace <code>&lt;hash_goes_here&gt;<\/code> with the actual hash in the following:<\/p>\n                    <p class=\"example\"><code>\"content_security_policy\" : \"script-src 'self' 'sha256-&lt;hash_goes_here&gt;'; object-src 'self'\"<\/code><\/p>\n                <\/li>\n            <\/ul><\/section><section class=\"informative\"><h5 id=\"csp-cross-domain\">Cross-domain Privileges<\/h5><p> Content scripts get the same cross-domain privileges as the rest of the extension. So if the extension has requested cross-domain access for a domain using the permissions key in <code>manifest.json<\/code>, then its content scripts get access that domain as well.<\/p><\/section><section class=\"informative\"><h5 id=\"availability-csp-content\">Content and Background Script Communication<\/h5><p> Although content scripts can't directly use most of the extension APIs, they can communicate with the extension's background scripts using the messaging APIs, and can therefore indirectly access all the same APIs that the background scripts can.<\/p><\/section><\/section><section><h2 id=\"wd\">Testing Extensions with WebDriver<\/h2><p>\n                Browsers vendors MAY choose to implement WebDriver support in their browsers to allow for the automated testing of browser extensions. To implement this support, the browser MUST support:\n            <\/p><ul><li>The base functionality as defined in the WebDriver specification [[webdriver]] <\/li>\n                <li>Additional functionality required to support browser extensions, defined below in the Browser Extensions specification<\/li>\n            <\/ul><section><h3>Protocol<\/h3><section><h4>List of Endpoints<\/h4><p>In addition to the endpoints defined in the WebDriver specification [[webdriver]], support for browser extensions requires these additional endpoints.<\/p><table class=\"simple\"><tbody><tr><th>Method<\/th>\n                                <th>URI Template<\/th>\n                                <th>Command<\/th>\n                            <\/tr><tr><td>POST<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserext<\/td>\n                                <td>Install Browser Extensions<\/td>\n                            <\/tr><tr><td>DELETE<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserext<var>extensionId<\/var>}<\/td>\n                                <td>Uninstall a Browser Extension<\/td>\n                            <\/tr><tr><td>GET<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserexts<\/td>\n                                <td>Get Browser Extensions<\/td>\n                            <\/tr><tr><td>GET<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserext\/{<var>extensionId<\/var>}\/actions<\/td>\n                                <td>Get Browser Extension Actions<\/td>\n                            <\/tr><tr><td>POST<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserext\/{<var>extensionId<\/var>}\/action\/<\/td>\n                                <td>Take Browser Extension Action<\/td>\n                            <\/tr><tr><td>GET<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/element\/{<var>element id<\/var>}\/browserext\/contextMenuItems<\/td>\n                                <td>Get Browser Extension Context Menu Items<\/td>\n                            <\/tr><tr><td>POST<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/element\/{<var>element id<\/var>}\/browserext\/contextMenuItem\/{<var>menuitemId<\/var>}<\/td>\n                                <td>Select Browser Extension Context Menu Item<\/td>\n                            <\/tr><tr><td>GET<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserext\/permissionPrompt<\/td>\n                                <td>Get Browser Extension Permission Prompt Text<\/td>\n                            <\/tr><tr><td>POST<\/td>\n                                <td>\/session\/{<var>sessionId<\/var>}\/browserext\/permissionPrompt\/{<var>promptId<\/var>}\/action\/{<var>promptAction<\/var>}<\/td>\n                                <td>Take Browser Extension Permission Prompt Action<\/td>\n                            <\/tr><\/tbody><\/table><\/section><section><h4>Handling Errors<\/h4><p>In addition to the codes defined in the Handling Errors section of the WebDriver specification [[webdriver]], support for browser extensions requires support the following additional error codes:<\/p><table class=\"simple\"><tbody><tr><th>\n                                    Error Code\n                                <\/th>\n                                <th>\n                                    HTTP Status\n                                <\/th>\n                                <th>\n                                    JSON Error Code\n                                <\/th>\n                                <th>\n                                    Description\n                                <\/th>\n                            <\/tr><tr><td><dfn id=\"dfn-no-such-extension-path\" data-dfn-type=\"dfn\">no such extension path<\/dfn><\/td>\n                                <td>400<\/td>\n                                <td><code>no such extension path<\/code><\/td>\n                                <td>\n                                    The filepath specified does not point to a valid browser extension on disk.\n                                <\/td>\n                            <\/tr><tr><td><dfn id=\"dfn-extension-not-installable\" data-dfn-type=\"dfn\">extension not installable<\/dfn><\/td>\n                                <td>400<\/td>\n                                <td><code>extension not installable<\/code><\/td>\n                                <td>\n                                    The specified extension could not be installed for some reason, e.g. its manifest.json is malformed or missing\n                                <\/td>\n                            <\/tr><tr><td><dfn id=\"dfn-no-such-browser-extension\" data-dfn-type=\"dfn\">no such browser extension<\/dfn><\/td>\n                                <td>400<\/td>\n                                <td><code>no such browser extension<\/code><\/td>\n                                <td>\n                                    There is no available browser extension on the current page with the given type.\n                                <\/td>\n                            <\/tr><tr><td><dfn id=\"dfn-no-such-browser-extension-action\" data-dfn-type=\"dfn\">no such browser extension action<\/dfn><\/td>\n                                <td>400<\/td>\n                                <td><code>no such browser action<\/code><\/td>\n                                <td>\n                                    A request to take a browser extension action could not be satisfied because\n                                    there is no browser extension action with the given id.\n                                <\/td>\n                            <\/tr><tr><td>\n                                <\/td><td>400<\/td>\n                                <td><code>no such browser extension context menu item<\/code><\/td>\n                                <td>\n                                    A request to select an extension context menu item could not be completed\n                                    because there is no matching extension context menu item for the given\n                                    element.\n                                <\/td>\n                            <\/tr><tr><td><dfn id=\"dfn-no-such-browser-extension-permission-prompt\" data-dfn-type=\"dfn\">no such browser extension permission prompt<\/dfn><\/td>\n                                <td>400<\/td>\n                                <td><code>no such browser extension permission prompt<\/code><\/td>\n                                <td>\n                                    A request to obtain the permission prompt information could not be completed\n                                    because there is no permission prompt at this time or the requested prompt was not available.\n                                <\/td>\n                            <\/tr><tr><td><dfn id=\"dfn-no-such-browser-extension-permission-prompt-action\" data-dfn-type=\"dfn\">no such browser extension permission prompt action<\/dfn><\/td>\n                                <td>400<\/td>\n                                <td><code>no such browser extension permission prompt action<\/code><\/td>\n                                <td>\n                                    A request to take the specified action could not be completed\n                                    because the action was not available.\n                                <\/td>\n                            <\/tr><\/tbody><\/table><\/section><\/section><section><h3>\n                        Capabilities\n                    <\/h3><p>In addition to the Capabilities defined in the WebDriver specification [[webdriver]], support for browser extensions requires the following additional entry on the <i>server capabilities<\/i> JSON object:<\/p><dl><dt>\n                            \"<code>enabledBrowserExtensions<\/code>\"\n                        <\/dt>\n                        <dd>\n                            An array of absolute paths which point to the location of browser extension files. The path <em title=\"MUST\" class=\"rfc2119\">MUST<\/em> point to either a single file appropriate for a specific brower (e.g. .crx) or a folder which contains the extension source files. If a folder is specified, the folder itself <em title=\"MUST\" class=\"rfc2119\">MUST<\/em> contain a <code>manifest.json<\/code> file. Subfolders containing <code>manifest.json<\/code> files will not be searched.\n                        <\/dd>\n                    <\/dl><section><h4>Extension Lifetime<\/h4><p>No browser extension will be installed or pre-loaded during a WebDriver session unless it is explicitly specified in the <code>enabledBrowserExtensions<\/code> array. Browser extensions loaded during the session will not be available to other active instances of the browser that are not part of the WebDriver session. When the session ends, the array of browser extensions are unloaded and uninstalled.<\/p><\/section><\/section><section><h3>\n                        Browser Extensions Commands\n                    <\/h3><p>Users interact with extensions through several means. First there are actions embedded in the UI of the browser, which can do something directly or open a popup. If the extension uses a popup, it can contain any html content. Finally, extensions can add items to the browser's context menu. WebDriver needs to be able to emulate user actions for all three of these controls in order to completely emulate user interaction.<\/p><p>In addition to the commands defined in the WebDriver specification [[webdriver]], support support for browser extensions requires the following additional commands. <\/p><section><h3><code>installBrowserExtensions()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>POST<\/td>\n                                    <td id=\"get-browser-extensions\">\/session\/{<var>sessionId<\/var>}\/browserext<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command installs any number of specified extensions from disk at runtime\n                        <\/p><aside class=\"example\">\n                            <p>\n                                To install an extension located at \"\/ext\/\" in a session with ID <i>0<\/i>, the local end would  \n                                POST to <code>\/session\/0\/browserext\/<\/code> with the request body \n                            <\/p>\n                            <pre class=\"hljs xquery\">{ \"extensionPaths\": [\"\/ext\/\"] }<\/pre>\n                        <\/aside><p>\n                            The value of the <code>extensionPaths<\/code> property MUST be an array of absolute paths which point to the location of browser extension files. The path <em title=\"MUST\" class=\"rfc2119\">MUST<\/em> point to either a single file appropriate for a specific brower (e.g. .crx) or a folder which contains the extension source files. If a folder is specified, the folder itself <em title=\"MUST\" class=\"rfc2119\">MUST<\/em> contain a <code>manifest.json<\/code> file. Subfolders containing <code>manifest.json<\/code> files will not be searched.\n                        <\/p><p>\n                            If any of the absolute paths specified in the <code>extensionPaths<\/code> property are invalid, a\n                            <code><a href=\"#dfn-no-such-extension-path\">no such extension path<\/a><\/code>\n                            status code MUST be returned.\n                        <\/p><p>\n                            If the installation of any extensions specified in the <code>extensionPaths<\/code> property fail, an\n                            <code><a href=\"#dfn-extension-not-installable\">extension not installable<\/a><\/code>\n                            status code MUST be returned.\n                        <\/p><\/section><section><h3><code>uninstallBrowserExtensions()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>DELETE<\/td>\n                                    <td>\/session\/{<var>sessionId<\/var>}\/browserext\/{extensionId}<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command uninstalls an extension in a specified browsing session.\n                        <\/p><dl class=\"parameters\"><dt>DOMString extensionId<\/dt>\n                            <dd>\n                                The unique identifier of the browser extension to be uninstalled from the session.\n                            <\/dd>  \n                            \n                                               \n                                   \n                                    \n                            \n                        <\/dl><\/section><section><h3><code>getBrowserExtensions()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>GET<\/td>\n                                    <td>\/session\/{<var>sessionId<\/var>}\/browserexts<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command returns a list of all browser extensions available to the current browsing session.\n                        <\/p><ul><li><p>Let <var>extension<\/var> be a new JSON Object.<\/p><\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>id<\/code> on <var>extension<\/var> to a\n                                    string which is unique for each browser extension.\n                                <\/p>\n                            <\/li>\n                                <li>\n                                    <p>\n                                        Set the property <code>name<\/code> on <var>extension<\/var> to a\n                                        string which corresponds to the browser extension's <code>\"name\"<\/code> entry in <code>manifest.json<\/code>.\n                                    <\/p>\n                                <\/li>\n                    <\/ul><\/section><section><h3><code>getBrowserExtActions()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>GET<\/td>\n                                    <td id=\"get-browser-extension-actions\">\/session\/{<var>sessionId<\/var>}\/browserext\/{<var>extensionId<\/var>}\/actions<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command returns a list of browser extension actions\n                            available for a particular browser extension. Note that this can change from page to page.\n                            It MUST return a list of objects which describes the available browser extension\n                            actions using the following algorithm for each action:\n                        <\/p><ul><li><p>Let <var>action<\/var> be a new JSON Object.<\/p><\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>type<\/code> on <var>action<\/var> to a\n                                    string which represents the type of browser extension action \n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>actionTitle<\/code> on <var>action<\/var> to the\n                                    mouseover text for the action.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>icon<\/code> on <var>action<\/var> to the\n                                    filename of the icon file or an empty string if none is displayed. If the\n                                    icon is dynamically generated, the current icon MUST be returned as a\n                                    lossless PNG image encoded using Base64.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    If the action supports the <code>badgeText<\/code> property, \n                                    set the property <code>badgeText<\/code> on <var>action<\/var> to the\n                                    badge text displayed over the action or an empty string if none is present.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Return <code>success<\/code> with data <var>action<\/var>\n                                <\/p>\n                            <\/li>\n                        <\/ul><dl class=\"parameters\"><dt>DOMString extensionId<\/dt>\n                            <dd>\n                                The unique identifier of the browser extension whose available actions are to be queried. \n                                Returned by the <code>getBrowserExtensions()<\/code> command.\n                            <\/dd>\n                            \n                                               \n                                   \n                                    \n                                                        \n                        <\/dl><\/section><section><h3><code>takeBrowserExtAction()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>POST<\/td>\n                                    <td id=\"take-browser-extension-action\">\/session\/{<var>sessionId<\/var>}\/browserext\/action\/{<var>browserExtActionId<\/var>}<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command causes a browser extension action to be activated. Depending on the <code>extensionActionType<\/code> specified, this\n                            will have one of several effects:\n                        <\/p><table class=\"simple jsoncommand\"><tbody><tr><th>extensionActionType<\/th>\n                                    <th>Effect<\/th>\n                                <\/tr><tr><td>browserAction<\/td>\n                                    <td>\n                                        This <code>extensionActionType<\/code> causes a browserAction to be activated, \n                                        as if the user had clicked the browser extension's browserAction icon in the browser's chrome. \n                                        Depending on the browser extension, this may cause a popup to be opened. \n                                    <\/td>\n                                <\/tr><tr><td>pageAction<\/td>\n                                    <td>\n                                        This <code>extensionActionType<\/code> causes a pageAction to be activated, \n                                        as if the user had clicked the browser extension's pageAction icon in the browser's address bar. \n                                        Depending on the browser extension, this may cause a popup to be opened. \n                                    <\/td>\n                                <\/tr><tr><td>enable<\/td>\n                                    <td>\n                                        This <code>extensionActionType<\/code> causes the specified extension to be enabled,\n                                        as if the user had turned the extension on from the browser's management UI. If the extension\n                                        is already enabled, this command has no effect. \n                                    <\/td>\n                                <\/tr><tr><td>disable<\/td>\n                                    <td>\n                                        This <code>extensionActionType<\/code> causes the specified extension to be disabled,\n                                        as if the user had turned the extension off from the browser's management UI. If the extension\n                                        is already disabled, this command has no effect. \n                                    <\/td>\n                                <\/tr><\/tbody><\/table><dl class=\"parameters\"><dt>DOMString browserExtId<\/dt>\n                            <dd>\n                                The unique identifier of the browser extension whose available actions are to be invoked. \n                                Returned by the <code>getBrowserExtensions()<\/code> command.\n                            <\/dd>  \n                            \n                                               \n                                   \n                                    \n                                                                              \n                            <dt>DOMString extensionActionType<\/dt>\n                            <dd>\n                                The type of the browser extension action to be taken, as returned\n                                by the <code>getBrowserExtensionActions()<\/code> command. This value\n                                <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> be one of the following: \"<code>pageAction<\/code>\", \n                                \"<code>browserAction<\/code>\", \"<code>enable<\/code>\", or \"<code>disable<\/code>\"\n                            <\/dd>\n                        <\/dl><aside class=\"example\">\n                            <p>\n                                To click a pageAction of an extension with ID <i>1<\/i> in a session with ID <i>0<\/i>, the local end would  \n                                POST to <code>\/session\/0\/browserext\/1\/action\/<\/code> with the request body \n                            <\/p>\n                            <pre class=\"hljs xquery\">{ \"type\": \"pageAction\" }<\/pre>\n                        <\/aside><p>\n                            If the type specified does not correspond to a browser extension action available in the browser\/on the current page, a\n                            <code><a href=\"#dfn-no-such-browser-extension-action\">no such browser extension action<\/a><\/code>\n                            status code MUST be returned.\n                        <\/p><\/section><section><h3><code>getBrowserExtContextMenuItems()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>GET<\/td>\n                                    \n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command opens a context menu for\n                            a given element. It MUST return a list of objects which describes the\n                            available context menu items using the following algorithm for each menu\n                            item:\n                        <\/p><ul><li><p>Let <var>menuitem<\/var> be a new JSON Object.<\/p><\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>id<\/code> on <var>menuitem<\/var> to a\n                                    string which is unique for the current session.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>menuItemTitle<\/code> on <var>menuitem<\/var> to the\n                                    text that is displayed in the UI.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>icon<\/code> on <var>menuitem<\/var> to the\n                                    filename of the icon file or an empty string if none is displayed.\n                                    If the icon is set because of the type of the context menu item (such as\n                                    a radio button or checkbox), the <code>icon<\/code> property MUST be set\n                                    to a string which is consistent for all icons of that type and status\n                                    (e.g. checked), including across sessions.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>parent<\/code> on <var>menuitem<\/var> to the\n                                    id of the item's parent or an empty string if the menuitem does not have\n                                    a parent.\n                                <\/p>\n                            <\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>itemType<\/code> on <var>menuitem<\/var> to a string value identifying the control type.\n                                <\/p>\n                                <table class=\"simple\"><tbody><tr><th>\n                                                String Value\n                                            <\/th>\n                                            <th>\n                                                Description\n                                            <\/th>\n                                        <\/tr><tr><td>\n                                                <code>checkbox<\/code>\n                                            <\/td>\n                                            <td>\n                                                A checkbox element\n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code>normal<\/code>\n                                            <\/td>\n                                            <td>\n                                                A standard menuItem element (text label)\n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code>radio<\/code>\n                                            <\/td>\n                                            <td>\n                                                A radio element. Within a radio group, only one menuItem will have a value of <code>true<\/code>\n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code>separator<\/code>\n                                            <\/td>\n                                            <td>\n                                                A horizontal line separator, typically used between groups of related items such as radio groups\n                                            <\/td>\n                                        <\/tr><\/tbody><\/table><\/li>\n                            <li>\n                                <p>\n                                    Set the property <code>itemChecked<\/code> on <var>menuitem<\/var> to a boolean value indicating\n                                    state of the item according to the following table. This applies only to items of itemType <code>checkbox<\/code> or <code>radio<\/code>. Within a radio group, only one menuItem will have a value of <code>true<\/code>.\n                                <\/p>\n                                <table class=\"simple\"><tbody><tr><th>\n                                                Boolean Value\n                                            <\/th>\n                                            <th>\n                                                Description\n                                            <\/th>\n                                        <\/tr><tr><td>\n                                                <code>true<\/code>\n                                            <\/td>\n                                            <td>\n                                                Selected\n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code>false<\/code>\n                                            <\/td>\n                                            <td>\n                                                Not selected\n                                            <\/td>\n                                        <\/tr><\/tbody><\/table><\/li>\n                            <li>\n                                <p>\n                                    Return <code>success<\/code> with data <var>menuitem<\/var>\n                                <\/p>\n                            <\/li>\n                        <\/ul>This list MUST contain all context menu items added by browser extensions and\n                        MUST NOT contain any other context menu items.<p>\n                            After this command has been run, the context menu MUST remain open until the\n                            next command that dismisses it (e.g. clicking, navigation) has been run.\n                        <\/p><dl class=\"parameters\"><dt>WebElement elementId<\/dt>\n                            <dd>\n                                The id of the element for which the context menu should be opened.\n                            <\/dd>\n                        <\/dl><\/section><section><h3><code>selectBrowserExtContextMenuItem()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>POST<\/td>\n                                    \n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            This command causes a context menu\n                            item to be selected. This MUST have the same effect as a user clicking or\n                            tapping the item in the open context menu.\n                            If the menu item is a parent of other menu items, this command MAY open a\n                            submenu, but MUST NOT have any other effect.\n                        <\/p><p>\n                            This command MUST be executed even if there is no context menu open.\n                        <\/p><dl class=\"parameters\"><dt>String menuitemId<\/dt>\n                            <dd>\n                                The context menu item id, as returned by the\n                                <code>getBrowserExtContextMenuItems()<\/code>\n                                command.\n                            <\/dd>\n                        <\/dl><p>\n                            If no context menu is currently open, a\n                            <code><a href=\"#dfn-no-such-browser-extension-context-menu-item\">no such extension context menu item<\/a><\/code>\n                            status code MUST be returned.\n                        <\/p><\/section><section><h3><code>getBrowserExtPermissionPrompt()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>GET<\/td>\n                                    <td id=\"browser-ext-permission-prompt\">\/session\/{<var>sessionId<\/var>}\/browserext\/permissionPrompt<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            Upon installation of a browser extension, a browser MAY notify the user of potential privacy-related capabilities of the extension. For example, if an extension's <code>manifest.json<\/code> requests the <code>\"tabs\"<\/code> or <code>\"webNavigation\"<\/code> permission, the browser MAY request user confirmation of these permissions. This command returns a JSON object that includes the names of permissions which require user confirmation and other information about the prompt. A similar prompt MAY also occur if an extension has been updated and the newer version requires additional permisisons. If there is no prompt at the time this API is called, the error \"<a href=\"#dfn-no-such-browser-extension-permission-prompt\">no such browser extension permission prompt<\/a>\" is returned.\n                        <\/p><ul><li>Let <var>permissionPrompt<\/var> be a new JSON Object.<\/li>\n                            <li>Set the property <code>promptId<\/code> to a string which is unique for each time a prompt is displayed.<\/li>\n                            <li>Set the property <code>promptMessageText<\/code> to a string which represents the text shown to the user, localized to the default locale specified in <code>manifest.json<\/code>.<\/li>\n                            <li>Set the property <code>promptDefaultLocale<\/code> to the default locale code specified in <code>manifest.json<\/code>. If locales are not used by the extension, the property SHOULD be omitted, but MAY return an empty string (\"\").<\/li>\n                            <li>Set the property <code>promptRequestingUrl<\/code> to a string representing the relevant URL. Strings returned here are the same strings defined in the <code>\"permissions\": [ ]<\/code> section of <code>manisfest.json<\/code>. If a URL is not applicable to the prompt, the property SHOULD be omitted but MAY also return an empty string (\"\").<\/li>\n                            <li>Set the property <code>promptPermissions<\/code> to an array of strings which correspond to requested permissions. Strings returned here are the same strings defined in the <code>\"permissions\": [ ]<\/code> section of <code>manisfest.json<\/code>. If permissions are not applicable to the prompt, the property SHOULD be omitted, but MAY include an empty array ([]) or an array of empty strings ([\"\"]). <\/li>\n                            <li>Set the property <code>promptActions<\/code> to an array of strings which correspond to possible actions the user could take. The actions <code>allow<\/code> and <code>deny<\/code> are required. Other optional values MAY be returned, but the behavior is not prescribed in this specification.<\/li>\n                        <\/ul><table class=\"simple\"><tbody><tr><th>\n                                                promptAction\n                                            <\/th>\n                                            <th>\n                                                Required\n                                            <\/th>\n                                            <th>\n                                                Description\n                                            <\/th>\n                                        <\/tr><tr><td>\n                                                <code>allow<\/code>\n                                            <\/td>\n                                            <td>\n                                                Required\n                                            <\/td>\n                                            <td>\n                                                Allows the extension to use or access the capabilities or sites \n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code>deny<\/code>\n                                            <\/td>\n                                            <td>\n                                                Required\n                                            <\/td>\n                                            <td>\n                                                Denies extension's use of the requested capabilities or sites. The browser implementation of <code>deny<\/code> is not prescribed in this specification, and MAY result in the cancellation of the extension's installation. \n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code>dismiss<\/code>\n                                            <\/td>\n                                            <td>\n                                                Optional\n                                            <\/td>\n                                            <td>\n                                                Some browsers MAY provide the user with an option to close a prompt without taking an <code>accept<\/code> or <code>deny<\/code> action. The browser implementation of <code>dismiss<\/code> is not prescribed in this specification, and MAY result in the cancellation of the extension's installation. \n                                            <\/td>\n                                        <\/tr><tr><td>\n                                                <code><i>{custom}<\/i><\/code>\n                                            <\/td>\n                                            <td>\n                                                Optional\n                                            <\/td>\n                                            <td>\n                                                Some browsers MAY provide additional options not prescribed in this table. For example, <code>alwaysAllow<\/code>, <code>allowOnce<\/code> or <code>allowForThisSiteOnly<\/code> may be presented in a permission prompt. The browser implementation for these custom actions is not prescribed in this specification. \n                                            <\/td>\n                                        <\/tr><\/tbody><\/table><aside class=\"example\">\n                                <p>This extension, which does not use locales, prompts to request broad permissions that do not correspond to a specific URL.<\/p>\n                            <pre class=\"hljs xquery\">\n{ \n\"promptId\": \"1\",\n\"promptMessageText\": \"A new extension Sample1 has been installed. It can access your browsing activity. Do you wish to continue? Yes No Close\",\n\"promptPermissions\": [\"tabs\",\"webNavigation\"],\n\"promptActions\": [\"allow\",\"deny\",\"dismiss\"]\n}<\/pre>\n                        <\/aside><aside class=\"example\">\n                                <p>This extension, which uses locales to provide translated strings, prompts for permissions specific to a banking URL.<\/p>\n                            <pre class=\"hljs xquery\">\n{ \n\"promptId\": \"2\",\n\"promptMessageText\": \"A new extension RegionalBankAssistant1 has been installed. It can read and modify data on regionalbank1.com. Enable Cancel\",\n\"promptDefaultLocale\": \"en\",\n\"promptRequestingUrl\": [\"http:\/\/*.regionalbank1.com\"],\n\"promptActions\": [\"allow\",\"deny\"]\n}<\/pre>\n                        <\/aside><\/section><section><h3><code>takeBrowserExtPermissionPromptAction()<\/code><\/h3><table class=\"simple jsoncommand\"><tbody><tr><th>HTTP Method<\/th>\n                                    <th>Path Template<\/th>\n                                    <th>Notes<\/th>\n                                <\/tr><tr><td>POST<\/td>\n                                    <td id=\"browser-ext-permission-prompt-action\">\/session\/{<var>sessionId<\/var>}\/browserext\/permissionPrompt\/{<var>promptId<\/var>}\/action\/{<var>promptAction<\/var>}<\/td>\n                                    \n                                <\/tr><\/tbody><\/table><p>\n                            Performs the action specified by <code>promptAction<\/code>. If there is no prompt at the time this API is called, or if the requested <code>promptId<\/code> is not available, the error \"<a href=\"#dfn-no-such-browser-extension-permission-prompt\">no such browser extension permission prompt<\/a>\" is returned. If the requested <code>promptAction<\/code> is not available, the error \"<a href=\"#dfn-no-such-browser-extension-permission-prompt-action\">no such browser extension permission prompt action<\/a>\" is returned.\n                        <\/p><\/section><\/section><\/section><section><h2>Packaging<\/h2><section class=\"informative\"><h3>Overview<\/h3><p>\n                    Browser extensions MAY be distributed and subsequently loaded as individual files or via a container, such as a *.zip archive. Regardless of the packaging format, these containers MUST contain the following:\n                <\/p><ul><li>manifest.json<\/li>\n                    <li>Necessary JavaScript files (&lt;filename&gt;.js)<\/li>\n                <\/ul><\/section><section class=\"informative\"><h3>Digital signatures<\/h3><p>\n                    Browser vendors MAY require that extension packages include a digital signature. The signature MAY indicate the source of the extension, such as a distribution store or the extension's publisher. This aspect of packaging varies by browser, and browser implementers MAY enforce different requirements.\n                <\/p><\/section><\/section><section class=\"appendix\"><h2 id=\"acknowledgements\">Acknowledgements<\/h2><p>\n                Many thanks to the members of the <a href=\"https:\/\/www.w3.org\/community\/browserext\/\">Browser Extension Community Group<\/a> for their detailed feedback on this specification. Thanks to the members of the <a href=\"http:\/\/www.w3.org\/testing\/browser\/\">Browser Testing and Tools Working Group<\/a> for their recommendations and feedback on WebDriver support.\n            <\/p><p>Some sections are reproduced from MDN:<\/p><ul><li><a href=\"https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/Internationalization\">https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/Internationalization<\/a><\/li>\n                <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/API\/i18n\/Locale-Specific_Message_reference\">https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/API\/i18n\/Locale-Specific_Message_reference<\/a><\/li>\n                <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/Content_scripts\">https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/Content_scripts<\/a><\/li>\n                <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/Content_Security_Policy\">https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/Content_Security_Policy<\/a><\/li>\n                <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/manifest.json\/content_security_policy\">https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/manifest.json\/content_security_policy<\/a><\/li>\n                <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/API\">https:\/\/developer.mozilla.org\/en-US\/Add-ons\/WebExtensions\/API<\/a> for various parameter\/property\/member descriptions for relevant APIs herein<\/li>\n            <\/ul><\/section><\/section>","created_at":"2018-11-07T16:49:25+0000","updated_at":"2018-11-07T16:49:25+0000","annotations":[],"mimetype":"text\/html","reading_time":35,"domain_name":"browserext.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":469,"title":"On Password Managers","url":"https:\/\/www.tbray.org\/ongoing\/When\/201x\/2017\/07\/16\/On-Password-Managers","content":"<p itemprop=\"description\">It has come to my attention that people are Wrong On The Internet about\npassword managers.  This matters, because almost everybody should be using\none.  Herewith background, opinions, and a description of my own setup,\nwhich is reasonably secure.<\/p><p id=\"p-1\" class=\"p1\">What is a password manager? \u00b7 \nIt\u2019s a piece of software that does the following (although not all of them\ndo all of these):<\/p>\n<ol><li><p>Store your passwords in a safe way, protected by at least a\npassword, which we call the \u201cmaster password\u201d.<\/p>\n<\/li>\n<li><p>Make new passwords for you.  Here\u2019s an example of a generated password:\n<i>QzbaLX}wA8Ad8awk<\/i>.  You\u2019re not expected to remember these.<\/p>\n<\/li>\n<li><p>Make it easy to use passwords.  One way is to copy it out of the\nmanager and paste it into a password field.  Another is to use a\nbrowser plugin that auto-fills login forms.\nOn certain combinations of app and mobile device, you can use your fingerprint\nto open the password manager, which makes everything way faster and\neasier.<\/p>\n<\/li> \n<li><p>Store other stuff too.  I keep various Important Numbers and AWS\ncredentials and recovery phrases and so on in there.<\/p>\n<\/li>\n<li><p>Synchronize between devices.  I have two computers and one phone and I\nneed access to my passwords on all of them.<\/p>\n<\/li>\n<\/ol><p>There\u2019s more, but those are the essentials.  The effect is that you end up\nusing a different password for every site and app, that they\u2019re all strong,\nand that you don\u2019t have to remember very much.<\/p>\n<p>My own manager, which\nI\u2019ve been running for years now, contains 504 items, and I use it a few times\na day, every day.  Granted, many of\nthe 504 are for sites and apps that no longer exist (like the dead people I\ncan\u2019t bear to erase from my contacts).<\/p>\n<p id=\"p-2\" class=\"p1\">How they work \u00b7 \nIt\u2019s pretty straightforward conceptually. They have a little database with\nall the stuff in it, and it\u2019s all encrypted using your password.  So even if\nsomeone steals the database, you\u2019re probably OK because modern crypto makes it\nreally hard to crack the code.<\/p>\n<p>Where it gets interesting is how these things synchronize between devices,\nand how they use the network.<\/p>\n<p>Basically, it comes down to this: Can you get access to your passwords over\nthe Web?  Lots of password managers allow this, but some don\u2019t.  For example,\nI use\n<a href=\"https:\/\/agilebits.com\/downloads\">the 1Password app<\/a>, which has no\nwebsite whatsoever, and has\n<a href=\"https:\/\/support.1password.com\/sync-options\/\">a variety of ways of\nsyncing<\/a> (iCloud, Dropbox, WiFi, local folder) none of which involve\ntalking to a website with a browser. <i>[There are lots of other password\nmanagers, which I\u2019m not gong to write about because I don\u2019t use them.]<\/i>\n<\/p>\n<p id=\"p-3\" class=\"p1\">What\u2019s wrong with a Web site? \u00b7 \nThe problem is that the site has my encrypted\ndata, and at some point, wants me to type in the password.  Thus, in\nprinciple, they can peek and see my passwords.  And hand them over to the\nNSA.  Or to the criminal gang that abducted the CEO\u2019s children.\nThis makes me unhappy.<\/p>\n<p>In principle, this could be OK.  What with modern JavaScript, it\u2019d be\nperfectly practicable to do all the crypto inside my browser, never send the\npassword (or anything unencrypted) over the wire, and have me sleep soundly at\nnight. \nFurthermore, since JavaScript is by definition open-source, I could in\nprinciple look at the code and satisfy myself that it\u2019s wholesome.<\/p>\n<p>In practice, nope.  The JavaScript platform is dynamic to the core and\nhorrifyingly complex even before they start loading massive modern application\nframeworks on it; any teeny little bug or zero-day exploit at any level of the\nstack and I\u2019m\ncooked.  Also, the NSA or a crook only has to make the slightest little\nmod to the code, and take it away a few milliseconds later, and the horse\nwould (silently) be out of the barn.<\/p>\n<p>In the 1Password app\u2019s sync model, however, one assumes they use the\npretty-secure HTTPS-based APIs for each of these products, machine to machine,\nno  JavaScript in the loop.<\/p>\n<p id=\"p-6\" class=\"p1\">Why we\u2019re talking about this \u00b7 \nBecause AgileBits, the company behind\n1Password, is trying to get people to move over to a Web-based thing; that\u2019s\nwhat you find when you go to\n<a href=\"https:\/\/1password.com\/\">1password.com<\/a>.<\/p>\n<p>There\u2019s a\n<a href=\"https:\/\/www.cyberscoop.com\/1password-subscription-no-local-machine-storage\/\">decent summary \nat cyberscoop<\/a> and a longer, more personal\n<a href=\"https:\/\/medium.com\/@kennwhite\/who-moved-my-cheese-1password-6a98a0fc6c56\">narrative from Kenn White<\/a>.<\/p>\n<p>I, like many security-conscious people, am just not gonna use anything\nwhere the same party, who\u2019s not me, gets to see my stored data <em>and<\/em> my\npassword.  Sorry.\nBut I love the 1Password apps and I\u2019d really like to go on using them.\nMore on that later.<\/p>\n<p id=\"p-5\" class=\"p1\">Let\u2019s get serious \u00b7 \nAm I claiming that my app-only approach is 100% safe?  No, because security\njust isn\u2019t binary, ever.  Let\u2019s see:<\/p>\n<ol><li><p>The bad guys could slip a sedative into my coffee at a coffee shop and\ninstall a keylogger on my computer, or<\/p>\n<\/li>\n<li><p>install a camera anywhere I work and focus it on my hands, or<\/p>\n<\/li>\n<li><p>phish me with a super-clever website or poisoned USB key, and get the\nkeylogger in that way, or<\/p>\n<\/li>\n<li><p>point a gun at me and ask me to unlock all my devices (then\nprobably pull the trigger), or<\/p>\n<\/li>\n<li><p>send a National Security Letter to AgileBits and force them to put\nbackdoor code in a future 1Password app release that sends the goodies to the enemies.<\/p>\n<\/li>\n<\/ol><p>And anyhow I\u2019m obviously a lame-ass hypocrite because I use the 1Password\nChrome plugin to fill in forms for me, and this means I type the master\npassword into a browser.  Having said that, I verified that it works when I\nhave the networks turned off, and at the end of the day, the plug-in is no\nmore nor less secure than the app I use all the time.<\/p>\n<p id=\"p-7\" class=\"p1\">Is your setup perfect? \u00b7 \nWell, I only remember four passwords:  For my personal computer,\nfor my work computer, for my AWS account, and the \n1Password master.  And the AWS password is just an accident of history; I\nonly need 3.<\/p>\n<p>Obviously I change them regularly and use password-less ssh access wherever\nI can, and lots of places I go have two-factor, via SMS or hardware token\n(Gemalto, Yubikey) or the Android Authenticator app.<\/p>\n<p>So, on balance I feel pretty secure.  One downside is when I\u2019m setting up a\nnew computer or phone.  The process of typing in long generated passwords on a\nmobile \u201ckeyboard\u201d is so impractical as to be hilarious.<\/p>\n<p>In effect, my security is about as good as my mobile device\u2019s.  Actually a\nbit better, because the 1Password app needs one more\nfingerprint-or-password.<\/p>\n \n<p id=\"p-9\" class=\"p1\">You sync through Dropbox, are you crazy?! \u00b7 \nAfter all, Condi Rice is a board member, which has to worry you.\nBut let\u2019s assume the worst: that Dropbox turns turtle for the Feds, or gets\ntotally pwned by bad guys.  So, congrats, they have my encrypted password\nfile.  It\u2019s not <em>impossible<\/em> that they might crack it.  But it\u2019d\nprobably be easier and cheaper for them to slip a sedative in my coffee,\nor\u2026\u00a0(see above).<\/p>\n<p id=\"p-8\" class=\"p1\">Why is AgileBits doing this? \u00b7 \nFor the same reason that Adobe has been pressuring its customers, for years\nnow, to start subscribing to its products, rather than buying each successive\nversion of each app.\nA subscription business is much nicer to operate than one \nwhere you have to go out and re-convince people to re-buy your software.<\/p>\n<p>I understand, and I support AgileBits wanting to become a subscription\nbiz.  But I still want to keep my data and password away from their servers.\nThis all seems fine to me.  I pay my monthly rent to Adobe and it\u2019s for\nLightroom &amp; Photoshop, not for their unexciting server-side offerings.\n<\/p>\n<p>So AgileBits, why not? \nPlease go ahead and start asking for subscriptions.  But\ndon\u2019t ask paranoid people like me to go anywhere near 1Password.com.<\/p>\n<p>AgileBits has addressed the situation in\n<a href=\"https:\/\/blog.agilebits.com\/2017\/07\/13\/why-we-love-1password-memberships\/\">Why\nWe Love 1Password Memberships<\/a>, but it\u2019s really unsatisfying, totally\nignoring the security concerns.  And (I guess I shouldn\u2019t be surprised)\nfailing to acknowledge the business advantages for them in making this\nmove.<\/p>\n<p class=\"p1\">Am I wrong? \u00b7 \nMaybe there\u2019s something I and the others who are all upset about the\n1Password move are missing; maybe it\u2019s all just OK and there\u2019s really no\nsignificant loss of security.  In which case, AgileBits really needs to\nexplain why.<\/p>\n<hr \/>\n<hr \/>","created_at":"2018-11-07T16:49:25+0000","updated_at":"2018-11-07T16:49:25+0000","published_at":"2017-07-16T00:00:00+0000","annotations":[],"language":"en","reading_time":7,"domain_name":"www.tbray.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":465,"title":"Weightless Books","url":"https:\/\/weightlessbooks.com\/","content":"<div id=\"sidebar\">\n\t\t\t\n\t\t\n\t    \n    <div class=\"eStore_classic_cart_wrapper estore-cart-wrapper-0\"><div class=\"shopping_cart\"><input type=\"image\" src=\"https:\/\/weightlessbooks.com\/wp-content\/plugins\/wp-cart-for-digital-products\/images\/shopping_cart_icon_3.png\" value=\"Shopping Cart\" title=\"Shopping Cart\" \/><h2>Shopping Cart<\/h2><div class=\"eStore_empty_cart_block\">Contains no books!<\/div><\/div><\/div>\t\t\n\t\t\t\t\t\n\t\t<ul><li class=\"menu\" id=\"findbooks\">\n\t\t<h3 class=\"menu_head top\"><a href=\"#\" title=\"Mouseover to Expand\">Find Books<\/a><\/h3>\n\t\t<ul class=\"menu_body\"><li id=\"author\"><a href=\"https:\/\/weightlessbooks.com\/category\/authors\/\">Author<\/a><\/li>\n\t<li id=\"publisher\" class=\"menu\">\n\t\t<h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/\">Publisher<\/a><\/h3>\n\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/ahwa\/\">AHWA<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/anna-yeatts-publisher\/\">Anna Yeatts<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/apex-book-company\/\">Apex Book Company<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/apex-book-company\/apex-publications\/\">Apex Publications<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/aqueduct-press\/\">Aqueduct Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/bad-ducky-industries\/\">Bad Ducky Industries<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/bastion-publications\/\">Bastion Publications<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/blind-eye-books\/\">Blind Eye Books<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/blind-eye-books\/one-block-empire\/\">One Block Empire<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/book-smugglers-publishing\/\">Book Smugglers Publishing<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/bookcyclone\/\">BookCyclone<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/bookcyclone\/scarlet-storm-press\/\">Scarlet Storm Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/bookcyclone\/signal-8-press\/\">Signal 8 Press<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/broad-street-press\/\">Broad Street Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/bull-spec-publisher\/\">Bull Spec<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/burrowing-wombat-press\/\">Burrowing Wombat Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/cheeky-frawg\/\">Cheeky Frawg<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chelsea-station-editions\/\">Chelsea Station Editions<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chizine-publications\/\">ChiZine Publications<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chizine-publications\/chigraphic\/\">ChiGraphic<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chizine-publications\/chiteen\/\">ChiTeen<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chizine-publications\/czp-ebooks\/\">CZP eBooks<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chizine-publications\/kqp\/\">KQP<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/chizine-publications\/undertow\/\">Undertow<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/circlet-press-inc\/\">Circlet Press Inc.<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/clockpunk-studios\/\">Clockpunk Studios<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/counterpoise\/\">counterpoise<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/crossed-genres\/\">Crossed Genres<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/dagan-books-llc\/\">Dagan Books LLC<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/direct-from-author\/\">Direct from Author<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/direct-from-author\/gavin-j-grant-and-kelly-link\/\">Gavin J. Grant and Kelly Link<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/dog-horn-publishing\/\">Dog Horn Publishing<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/dzanc-books\/\">Dzanc Books<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/dzanc-books\/dzanc-books-reprint-series\/\">Dzanc Books rEprint Series<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/epigram-books\/\">Epigram Books<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/escape-artists\/\">Escape Artists<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/fairy-tale-review\/\">Fairy Tale Review<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/fantasy-scroll-press-llc\/\">Fantasy Scroll Press LLC<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/featherproof-books\/\">Featherproof Books<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/fireside-press\/\">Fireside Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/firkin-press\/\">Firkin Press<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/gemma\/\">Gemma<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/gemma\/gemma-open-door\/\">Gemma Open Door<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/hildy-silverman-publisher\/\">Hildy Silverman (Publishe...<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/hydra-house\/\">Hydra House<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/infinity-plus\/\">infinity plus<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/infinity-plus\/infinite-press\/\">infinite press<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/inkless-media\/\">Inkless Media<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/innsmouth-free-press\/\">Innsmouth Free Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/invisible-publishing\/\">Invisible Publishing<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/john-joseph-adams-publisher\/\">John Joseph Adams (Publis...<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/lackington-press\/\">Lackington Press<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/lethe-press\/\">Lethe Press<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/lethe-press\/brazenhead\/\">Brazenhead<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/lethe-press\/tincture\/\">Tincture<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/lethe-press\/unzipped\/\">Unzipped<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/lightspeed-magazine-press\/\">Lightspeed Magazine Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/locus-science-fiction-foundation\/\">Locus Science Fiction Fou...<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/luna-station-press\/\">Luna Station Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/massachusetts-review-inc\/\">Massachusetts Review Inc.<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/math-paper-press\/\">Math Paper Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/metaphysical-circus-press\/\">Metaphysical Circus Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/mithila-review\/\">Mithila Review<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/mythic-delirium-books\/\">Mythic Delirium Books<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/mythic-delirium-books\/antimatter-press\/\">Antimatter Press<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/new-mexico-state-university\/\">New Mexico State Universi...<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/nightmare-magazine-press\/\">Nightmare Magazine Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/papaveria-press\/\">Papaveria Press<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/phoenix-pick\/\">Phoenix Pick<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/phoenix-pick\/hearts-nest-press\/\">Hearts Nest Press<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/plasma-spyglass\/\">Plasma Spyglass<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/\">PM Press<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/arissa-media-group\/\">Arissa Media Group<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/busboys-and-poets-press\/\">Busboys and Poets Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/christie-books\/\">Christie Books<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/kersplebedeb\/\">Kersplebedeb<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/microcosm-publishing\/\">Microcosm Publishing<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/reach-and-teach\/\">Reach and Teach<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/pm-press\/sexto-piso-editorial\/\">Sexto Piso Editorial<\/a><\/li><\/ul><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/prime-books\/\">Prime Books<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/prime-books\/masque-books\/\">Masque Books<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/reckoning-press\/\">Reckoning Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/sator-press\/\">Sator Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/senses-five-press\/\">Senses Five Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/sfwa\/\">SFWA<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/shimmer\/\">Shimmer<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/small-beer-press\/\">Small Beer Press<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/small-beer-press\/big-mouth-house\/\">Big Mouth House<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/spilogale-inc\/\">Spilogale Inc.<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/spilt-milk-press\/\">Spilt Milk Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/stone-skin-press\/\">Stone Skin Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/tachyon-publications\/\">Tachyon Publications<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/the-dark-publisher\/\">TDM Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/temporary-culture\/\">Temporary Culture<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/the-alchemy-press\/\">The Alchemy Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/the-copper-pig-writers-society\/\">The Copper Pig Writers So...<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/the-homeless-moon\/\">The Homeless Moon<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/tta-press\/\">TTA Press<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/twelfth-planet-press\/\">Twelfth Planet Press<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/twelfth-planet-press\/deadlines\/\">Deadlines<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/uncanny-magazine-press\/\">Uncanny Magazine Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/unwrecked-press\/\">UnWrecked Press<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/wildside-press\/\">Wildside Press<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/wildside-press\/alien-ebooks\/\">Alien eBooks<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/wildside-press\/borgo-press\/\">Borgo Press<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/wildside-press\/wildside-press-wildside-press-llc\/\">Wildside Press LLC<\/a><\/li><\/ul><\/li><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/word-horde\/\">Word Horde<\/a><\/li><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/wyrm-publishing\/\">Wyrm Publishing<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/wyrm-publishing\/ministry-of-whimsy\/\">Ministry of Whimsy<\/a><\/li><\/ul><\/li>\n\t\t<\/ul><\/li>\n\t<li id=\"genre\" class=\"menu\">\n\t\t<h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/genre\">Genre<\/a><\/h3>\n\t\t<ul class=\"menu_body\"><li id=\"fiction\" class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\">Fiction<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/fantasy\">Fantasy<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/historical\">Historical<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/horror\">Horror<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/lgbt-fiction\">LGBT<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/literary\">Literary<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/mainstream\">Mainstream<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/mystery-and-detective\">Mystery<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/novel\">Novels<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/romance\">Romance<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/science-fiction\">Science Fiction<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/fiction\/short-stories\">Short Stories<\/a><\/li>\n\t\t\t\t<\/ul><\/li>\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/nonfiction\">Nonfiction<\/a>\n\t\t\t<\/li>\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/category\/genre\/poetry\">Poetry<\/a>\n\t\t\t<\/li>\n\t\t<\/ul><\/li>\n\t<li id=\"subscription\" class=\"menu\">\n\t\t<h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/category\/format\/subscription\/\">Subscription<\/a><\/h3>\n\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\">Apex Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-12-month-subscription\/\">Beneath Ceaseless Skies 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/black-static-annual-subscription\/\">Black Static Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/subscription\/chelsea-station-1-year-subscription\/\">Chelsea Station 1-Year Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/chizine-publications-annual-subscription-2018\/\">ChiZine Publications Annual Subscription 2018<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/clarkesworld-magazine-12-month-subscription\/\">Clarkesworld Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/clarkesworld-magazine-6-month-subscription\/\">Clarkesworld Magazine \u2013 6 Month Subscription<\/a><\/li><li><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/congress-magazine-12-month-subscription\/\">Congress Magazine 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/fantasy-scroll-magazine-12-month-subscription\/\">Fantasy Scroll Magazine \u2013 12 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/flash-fiction-online-12-month-subscription\/\">Flash Fiction Online, 12 month subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/forever-magazine-12-month-subscription\/\">Forever Magazine 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/forever-magazine-6-month-subscription\/\">Forever Magazine 6-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/galaxys-edge-magazine-annual-subscription\/\">Galaxy\u2019s Edge Magazine Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/hearts-kiss-magazine\/\">Heart\u2019s Kiss Magazine Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/interzone-annual-subscription\/\">Interzone Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/lackingtons-12-month-subscription\/\">Lackington\u2019s 4-issue subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/subscription\/lcrw\/lady-churchills-rosebud-wristlet-subscription\/\">Lady Churchill\u2019s Rosebud Wristlet Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/lightspeed-magazine-6-month-subscription\/\">Lightspeed Magazine: 6-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/lightspeed-magazine-12-month-subscription\/\">Lightspeed Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/nonfiction\/locus-6-month-subscription\/\">Locus 6-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/nonfiction\/locus-12-month-subscription\/\">Locus 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/luna-station-quarterly-12-month-subscription\/\">Luna Station Quarterly 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/midnight-echo-magazine-12-month-subscription\/\">Midnight Echo Magazine \u2013 12 month subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/midnight-echo-magazine-24-month-subscription\/\">Midnight Echo Magazine \u2013 24 month subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/mithila-review-yearly-subscription\/\">Mithila Review Yearly Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/nightmare-magazine-6-month-subscription\/\">Nightmare Magazine: 6-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/nightmare-magazine-12-month-subscription\/\">Nightmare Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/fantasy\/new-york-review-of-science-fiction-12-issue-subscription\/\">New York Review of Science Fiction 12-issue Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/on-spec-magazine-1-year-subscription-4-issues\/\">On Spec Magazine 1-Year Subscription (4 issues)<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/plasma-frequency-one-year-subscription\/\">Plasma Frequency One Year Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/plasma-frequency-two-year-subscription\/\">Plasma Frequency Two Year subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/nonfiction\/puerto-del-sol\/\">Puerto del Sol<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/see-the-elephant-annual-subscription\/\">See the Elephant, Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/sherlock-holmes-mystery-magazine-4-issue-subscription\/\">Sherlock Holmes Mystery Magazine 4-Issue Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/space-and-time-magazine-subscription\/\">Space and Time Magazine Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-dark-annual-subscription\/\">The Dark Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-magazine-of-fantasy-and-science-fiction-6-issue-subscription\/\">The Magazine of Fantasy and Science Fiction \u2013 6 Issue Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/the-rifter-subscription\/\">The Rifter Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-sfwa-bulletin-subscription\/\">The SFWA Bulletin Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\">Uncanny Magazine \u2013 12 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/weird-tales-4-issue-subscription\/\">Weird Tales 4-Issue Subscription<\/a><\/li>\n\t\t<\/ul><\/li>\n\t<li id=\"date\" class=\"menu\">\n\t\t<h3 class=\"menu_head\"><a href=\"#\">Date Published<\/a><\/h3>\n\t\t<ul class=\"menu_body\"><li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2018\/\">2018<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2018\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2018\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2017\/\">2017<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2017\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2017\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2016\/\">2016<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2016\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2016\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2015\/\">2015<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2015\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2015\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2014\/\">2014<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2014\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2014\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2013\/\">2013<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2013\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2013\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2012\/\">2012<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2012\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2012\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2011\/\">2011<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2011\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2011\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2010\/\">2010<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2010\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2010\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2009\/\">2009<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2009\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2009\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2008\/\">2008<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2008\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2008\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2007\/\">2007<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2007\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2007\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2006\/\">2006<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2006\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/02\/\">February<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/03\/\">March<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/05\/\">May<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2006\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2005\/\">2005<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2005\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2005\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2005\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2005\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2005\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2005\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2004\/\">2004<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2004\/04\/\">April<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2004\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2004\/10\/\">October<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2003\/\">2003<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2003\/06\/\">June<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2003\/09\/\">September<\/a><\/li>\n\t\t\t\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/2003\/11\/\">November<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2002\/\">2002<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2002\/08\/\">August<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/2001\/\">2001<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/2001\/07\/\">July<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/1970\/\">1970<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/1970\/01\/\">January<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<li class=\"menu\"><h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/1969\/\">1969<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/1969\/12\/\">December<\/a><\/li>\n\t\t\t\n\t\t\t<\/ul><\/li>\n\t\t\n\t\t<\/ul><\/li>\n\t<li id=\"price\" class=\"menu\">\n\t\t<h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/price\/\">Price<\/a><\/h3>\n\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/price\/?low=0&amp;high=0.01&amp;orderby=sales_count&amp;order=DESC&amp;submit=Submit\">Free!<\/a><\/li>\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/price\/?low=0.01&amp;high=1&amp;orderby=sales_count&amp;order=DESC&amp;submit=Submit\">$0.01 - $1<\/a><\/li>\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/price\/?low=1.01&amp;high=5&amp;orderby=sales_count&amp;order=DESC&amp;submit=Submit\">$1.01 - $5<\/a><\/li>\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/price\/?low=5.01&amp;high=10&amp;orderby=sales_count&amp;order=DESC&amp;submit=Submit\">$5.01 - $10<\/a><\/li>\n\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/price\/?low=10.01&amp;high=1000&amp;orderby=sales_count&amp;order=DESC&amp;submit=Submit\">$10+<\/a><\/li>\n\t\t<\/ul><\/li>\n\t\n\t\t<\/ul><\/li>\n\t\n\t\t<li class=\"menu\" id=\"bestsellers\">\n\t\t\t<h3 class=\"menu_head top\"><a href=\"#\" title=\"Click to Expand\">Bestsellers<\/a><\/h3>\n\t\t\t<ul class=\"menu_body\"><li id=\"week\" class=\"menu\"><h3 class=\"menu_head\"><a href=\"#\">This Week<\/a><\/h3>\n\t\t\t\t\t<ul class=\"menu_body\"><\/ul><ol><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/lightspeed-magazine-12-month-subscription\/\">Lightspeed Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\">Uncanny Magazine \u2013 12 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/nonfiction\/locus-october-2018-693\/\">Locus October 2018 (#693)<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-magazine-of-fantasy-and-science-fiction-6-issue-subscription\/\">The Magazine of Fantasy and Science Fiction \u2013 6 Issue Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-24\/\">Uncanny Magazine Issue 24<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-261-tenth-anniversary-month-double-issue-i\/\">Beneath Ceaseless Skies Issue #261, Tenth Anniversary Month Double-Issue I<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\">Apex Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/mithila-review-issue-10\/\">Mithila Review \u2013 Issue 10<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-nameless-dark\/\">The Nameless Dark<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/years-best-ya-speculative-fiction-2014\/\">Year\u2019s Best YA Speculative Fiction 2014<\/a><\/li><\/ol><\/li><\/ul><\/li>\n\t\t\t\t<li id=\"month\" class=\"menu\"><h3 class=\"menu_head\"><a href=\"#\">This Month<\/a><\/h3>\n\t\t\t\t\t<ul class=\"menu_body\"><\/ul><ol><li><a href=\"https:\/\/weightlessbooks.com\/format\/latchkey\/\">Latchkey<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/publisher\/small-beer-press\/big-mouth-house\/archivist-wasp\/\">Archivist Wasp<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-12-month-subscription\/\">Beneath Ceaseless Skies 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-24\/\">Uncanny Magazine Issue 24<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-261-tenth-anniversary-month-double-issue-i\/\">Beneath Ceaseless Skies Issue #261, Tenth Anniversary Month Double-Issue I<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\">Uncanny Magazine \u2013 12 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/flash-fiction-online-12-month-subscription\/\">Flash Fiction Online, 12 month subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/clarkesworld-magazine-12-month-subscription\/\">Clarkesworld Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-magazine-of-fantasy-and-science-fiction-6-issue-subscription\/\">The Magazine of Fantasy and Science Fiction \u2013 6 Issue Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/lightspeed-magazine-12-month-subscription\/\">Lightspeed Magazine: 12-Month Subscription<\/a><\/li><\/ol><\/li><\/ul><h3 class=\"menu_head\"><a href=\"#\">This Year<\/a><\/h3>\n\t\t\t\t\t<ul class=\"menu_body\"><\/ul><ol><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\">Uncanny Magazine \u2013 12 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-12-month-subscription\/\">Beneath Ceaseless Skies 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/the-magazine-of-fantasy-and-science-fiction-6-issue-subscription\/\">The Magazine of Fantasy and Science Fiction \u2013 6 Issue Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/galaxys-edge-magazine-annual-subscription\/\">Galaxy\u2019s Edge Magazine Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/flash-fiction-online-12-month-subscription\/\">Flash Fiction Online, 12 month subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/clarkesworld-magazine-12-month-subscription\/\">Clarkesworld Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/lightspeed-magazine-12-month-subscription\/\">Lightspeed Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\">Apex Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/clarkesworld-magazine-6-month-subscription\/\">Clarkesworld Magazine \u2013 6 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/fantasy\/new-york-review-of-science-fiction-12-issue-subscription\/\">New York Review of Science Fiction 12-issue Subscription<\/a><\/li><\/ol>\n\t\t\t\t<h3 class=\"menu_head\"><a href=\"#\">Overall<\/a><\/h3>\n\t\t\t\t\t<ul class=\"menu_body\"><\/ul><ol><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/clarkesworld-magazine-12-month-subscription\/\">Clarkesworld Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/lightspeed-magazine-12-month-subscription\/\">Lightspeed Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-12-month-subscription\/\">Beneath Ceaseless Skies 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/magazine\/galaxys-edge-magazine-annual-subscription\/\">Galaxy\u2019s Edge Magazine Annual Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\">Uncanny Magazine \u2013 12 Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\">Apex Magazine: 12-Month Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/publisher\/blind-eye-books\/irregulars-preorder\/\">Irregulars<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/the-rifter-subscription\/\">The Rifter Subscription<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/format\/book\/lord-of-the-white-hell-book-one\/\">Lord of the White Hell Book One<\/a><\/li><li><a href=\"https:\/\/weightlessbooks.com\/fiction\/fantasy\/new-york-review-of-science-fiction-12-issue-subscription\/\">New York Review of Science Fiction 12-issue Subscription<\/a><\/li><\/ol>\n\t\t\t\n\t\t\n\t\t\t\t\n\t\t\t\t\t\t<a href=\"https:\/\/weightlessbooks.com\/about\">About<\/a>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/category\/bundles\">Bundles<\/a>\n\t\t\t<h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/affiliates\/\">Affiliate Program<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/affiliates\/\">Affiliate Login<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/widget-generator\/\">Widget Generator<\/a><\/li>\n\t\t\t\t<\/ul>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/contact\">Contact<\/a>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/faq#kindle_emails\">Ebooks Direct to Your Device<\/a>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/mailing-list\">Email List Signup<\/a>\n\t\t\t<h3 class=\"menu_head\"><a href=\"https:\/\/weightlessbooks.com\/faq\">FAQ<\/a><\/h3>\n\t\t\t\t<ul class=\"menu_body\"><li><a href=\"https:\/\/weightlessbooks.com\/faq#howitworks\">How to Buy Ebooks<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/faq#drm\">DRM-Free<\/a><\/li>\t\t\t\t\t\t\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/faq#how-to-read-epub-or-mobi\">How to Read Ebooks<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/faq#privacy\">Privacy<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/faq#refunds\">Refunds<\/a><\/li>\n\t\t\t\t\t<li><a href=\"https:\/\/weightlessbooks.com\/faq#renew\">Renewals<\/a><\/li>\t\t\t\t\t\t\n\t\t\t\t<\/ul>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/for-publishers\">For Publishers<\/a>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/updates\/give-the-gift-of-ebooks\">Gifts<\/a>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/category\/sale\">Sale<\/a>\n\t\t\t<a href=\"https:\/\/weightlessbooks.com\/price\/?low=0.01&amp;high=1&amp;orderby=sales_count&amp;order=DESC&amp;submit=Submit\">$0.99 Ebooks<\/a>\n\t\t\t\t\t\t\n\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/category\/updates\/\">Updates<\/a><\/h3>\t\t\n\t\t\t\t<ul><\/ul>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\n\t\t\n\t\t<ul><\/ul><\/div>\n\n\t\n\t\n\t<div id=\"maincolumn-right\">\n\t\t\n\t\n\t\t<div class=\"featured floatinglist\">\n\t\t\n\t\t\n\t\t\t<ul><li class=\"product\">\n\t\t\t\t<h3>New:<\/h3>\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/fiction\/icefall\/\"><img width=\"100\" height=\"151\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-140x212.jpg\" class=\"attachment-thumb size-thumb wp-post-image\" alt=\"Icefall\" title=\"Icefall\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-140x212.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-200x302.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-768x1161.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-400x605.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover.jpg 1400w\" \/><\/a><div class=\"price\"><b>$4.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9787\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9787\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Icefall Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Icefall Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/icefall-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"4.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"4.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9787\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9787\"><\/div><\/div><\/div>\t\t\t\t\t\t<h2><a href=\"https:\/\/weightlessbooks.com\/fiction\/icefall\/\">Icefall<\/a> Stephanie Gunn<\/h2>\n\t\t\t\t\t\t<p>The Mountain on the planet of Icefall holds the mystery to a lost colony and an irresistible, fatal allure to the climbers of the universe. Maggie is determined to be the first to make the summit. Aisha, injured in a climbing\u2026 <a href=\"https:\/\/weightlessbooks.com\/fiction\/icefall\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t<h3>Featured:<\/h3>\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/fiction\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine\/\"><img width=\"100\" height=\"134\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover-140x187.jpg\" class=\"attachment-thumb size-thumb wp-post-image\" alt=\"The Best of Beneath Ceaseless Skies Online Magazine, Year Nine\" title=\"The Best of Beneath Ceaseless Skies Online Magazine, Year Nine\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover-140x187.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover-200x267.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover-400x533.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover-75x100.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover.jpg 750w\" \/><\/a><div class=\"price\"><b>$3.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9710\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9710\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"The Best of Beneath Ceaseless Skies Online Magazine, Year Nine Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"The Best of Beneath Ceaseless Skies Online Magazine, Year Nine Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine-cover-75x100.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"3.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"3.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9710\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9710\"><\/div><\/div><\/div>\t\t\t\t\t\t<h2><a href=\"https:\/\/weightlessbooks.com\/fiction\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine\/\">The Best of Beneath Ceaseless Skies Online Magazine, Year Nine<\/a> Scott H. Andrews (Editor) et al.<\/h2>\n\t\t\t\t\t\t<p><em>A young artificer and a young inventor fight society to make a place for themselves and their work\u2026<\/em><\/p>\n<p><em>A veteran reunites with a lost comrade in an amnesty city that cocoons anyone who considers violence\u2026<\/em><\/p>\n<p><em>A <\/em>\u2026 <a href=\"https:\/\/weightlessbooks.com\/fiction\/the-best-of-beneath-ceaseless-skies-online-magazine-year-nine\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t<\/li>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<li class=\"product\" id=\"twitter\">\n\t\t\t\t\t\t<h3><a href=\"http:\/\/twitter.com\/WeightlessBooks\">Twitter<\/a><\/h3>\n\t\t\t\t\t\t<ul>   \t\t\n\t\t\t\t\t\t<\/ul><\/li>\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t<\/ul><\/div>\t\n\t\t<div class=\"floatinglist\" id=\"newtitles\">\n\t\t\t\t\n\t\t<ul>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t\t<div class=\"cover\"><div class=\"subscription_label\"> Current Issue:<\/div><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\"><img width=\"75\" height=\"107\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-75x107.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Apex Magazine Issue 114\" title=\"Apex Magazine Issue 114\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-75x107.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-200x286.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-768x1097.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-400x571.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-140x200.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover.jpg 1400w\" \/><\/a><div class=\"price\"><b>$21.95<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"614\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-614\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Apex Magazine: 12-Month Subscription\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Apex Magazine: 12-Month Subscription\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"21.95\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"21.95\" \/><input type=\"hidden\" name=\"price_tmp1_old\" value=\"19.95\" \/><input type=\"hidden\" name=\"item_number\" value=\"614\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-614\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\">Apex Magazine: 12-Month Subscription<\/a> Jason Sizemore<\/h3>\n\t\t\t\t\t\t<p>Apex began publishing the critically-acclaimed <em>Apex Science Fiction and Horror Digest<\/em> in 2005, focusing primarily on the science fiction and horror genres. in 2006, the company expanded its offerings to include \u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-12-month-subscription\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t\t<div class=\"cover\"><div class=\"subscription_label\"> Current Issue:<\/div><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Uncanny Magazine Issue 25\" title=\"Uncanny Magazine Issue 25\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover.jpg 1200w\" \/><\/a><div class=\"price\"><b>$23.88<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"6592\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-6592\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Uncanny Magazine - 12 Month Subscription Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Uncanny Magazine - 12 Month Subscription Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"23.88\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"23.88\" \/><input type=\"hidden\" name=\"price_tmp1_old\" value=\"21.88\" \/><input type=\"hidden\" name=\"item_number\" value=\"6592\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-6592\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\">Uncanny Magazine \u2013 12 Month Subscription<\/a> Lynne M.  Thomas<\/h3>\n\t\t\t\t\t\t<p><strong>About <em><a href=\"http:\/\/uncannymagazine.com\/\">Uncanny Magazine<\/a><\/em><\/strong><\/p>\n<p><em>Uncanny Magazine<\/em> is a bimonthly science fiction and fantasy magazine first published in November 2014. Edited by 2013 and 2014 Hugo award finalists for best semiprozine Lynne M. Thomas and Michael\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-12-month-subscription\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t\t<\/ul><\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"publisher\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<h2><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/twelfth-planet-press\/\" title=\"More titles from Twelfth Planet Press\">Twelfth Planet Press<\/a><\/h2>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/mother-of-invention\/\"><img width=\"75\" height=\"109\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-75x109.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Mother of Invention\" title=\"Mother of Invention\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-75x109.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-200x291.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-768x1119.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-400x583.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-140x204.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover.jpg 1400w\" \/><\/a><div class=\"price\"><b>$11.99AUD<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9704\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9704\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Mother of Invention Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Mother of Invention Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/mother-of-invention-cover-75x109.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"11.99AUD\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"11.99AUD\" \/><input type=\"hidden\" name=\"item_number\" value=\"9704\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9704\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/mother-of-invention\/\">Mother of Invention<\/a> Rivqa Rafael et al.<\/h3>\n\t\t\t\t\t\t<p>Extraordinary short stories about gender, artificial intelligence and the art of building something new. <em>Mother of Invention<\/em> features the work of Seanan McGuire, Ambelin Kwaymullina, Nisi Shawl, John Chu, Justina\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/mother-of-invention\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/fiction\/sharp-edge\/\"><img width=\"75\" height=\"114\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-75x114.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Sharp Edge\" title=\"Sharp Edge\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-75x114.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-200x305.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-768x1172.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-400x611.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-140x214.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover.jpg 1400w\" \/><\/a><div class=\"price\"><b>$6.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9266\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9266\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Sharp Edge Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Sharp Edge Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sharp-edge-cover-75x114.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"6.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"6.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9266\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9266\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/fiction\/sharp-edge\/\">Sharp Edge<\/a> Marianne Delacourt<\/h3>\n\t\t\t\t\t\t<p>Dead bodies, a complicated love life, a conflict between the local drug cartels, and a favour owed to the local bikies have Tara scrambling to stay ahead of the game.\nJust another Tuesday for Tara Sharp.<\/p>\n<p>Garth Wilmot, \u2026 <a href=\"https:\/\/weightlessbooks.com\/fiction\/sharp-edge\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"publisher\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<h2><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/apex-book-company\/apex-publications\/\" title=\"More titles from Apex Publications\">Apex Publications<\/a><\/h2>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-issue-114\/\"><img width=\"75\" height=\"107\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-75x107.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Apex Magazine Issue 114\" title=\"Apex Magazine Issue 114\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-75x107.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-200x286.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-768x1097.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-400x571.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-140x200.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover.jpg 1400w\" \/><\/a><div class=\"price\"><b>$2.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9772\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9772\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Apex Magazine Issue 114 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Apex Magazine Issue 114 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/apex-magazine-issue-114-cover-75x107.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"2.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"2.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9772\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9772\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-issue-114\/\">Apex Magazine Issue 114<\/a> Storm Humbert et al.<\/h3>\n\t\t\t\t\t\t<p><i>Apex Magazine<\/i> is a monthly science fiction, fantasy, and horror magazine featuring original, mind-bending short fiction from many of the top pros of the field. New issues are released on the first Tuesday of every month.\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/apex-magazine-issue-114\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/the-apex-book-of-world-sf-volume-5\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"The Apex Book of World SF: Volume 5\" title=\"The Apex Book of World SF: Volume 5\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-768x1153.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover.jpg 1400w\" \/><\/a><div class=\"price\"><b>$4.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9773\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9773\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"The Apex Book of World SF: Volume 5 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"The Apex Book of World SF: Volume 5 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-apex-book-of-world-sf-volume-5-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"4.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"4.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9773\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9773\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/the-apex-book-of-world-sf-volume-5\/\">The Apex Book of World SF: Volume 5<\/a> Cristina Jurado et al.<\/h3>\n\t\t\t\t\t\t<p>The landmark anthology series of international speculative fiction returns with volume 5 of <em>The Apex Book of World SF<\/em>. Cris Jurado joins series editor Lavie Tidhar to highlight the best speculative fiction from around\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/the-apex-book-of-world-sf-volume-5\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"publisher\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<h2><a href=\"https:\/\/weightlessbooks.com\/category\/publisher\/uncanny-magazine-press\/\" title=\"More titles from Uncanny Magazine Press\">Uncanny Magazine Press<\/a><\/h2>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-25\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Uncanny Magazine Issue 25\" title=\"Uncanny Magazine Issue 25\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-25-cover.jpg 1200w\" \/><\/a><div class=\"price\"><b>$3.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9766\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9766\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Uncanny Magazine Issue 25 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Uncanny Magazine Issue 25 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"3.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"3.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9766\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9766\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-25\/\">Uncanny Magazine Issue 25<\/a> Lynne M.  Thomas et al.<\/h3>\n\t\t\t\t\t\t<p>The November\/December 2018 issue of Hugo Award-winning <em>Uncanny Magazine.<\/em><\/p>\n<p>Featuring new fiction by Isabel Yap, T. Kingfisher, Naomi Kritzer, Monica Valentinelli, and Cassandra Khaw. Reprinted fiction by Sofia \u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-25\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-24\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Uncanny Magazine Issue 24\" title=\"Uncanny Magazine Issue 24\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover.jpg 1200w\" \/><\/a><div class=\"price\"><b>$3.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9679\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9679\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Uncanny Magazine Issue 24 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Uncanny Magazine Issue 24 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/uncanny-magazine-issue-24-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"3.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"3.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9679\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9679\"><\/div><\/div><\/div>\t\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-24\/\">Uncanny Magazine Issue 24<\/a> Lynne M.  Thomas et al.<\/h3>\n\t\t\t\t\t\t<p>The September\/October 2018 issue of Hugo Award-winning <em>Uncanny Magazine.<\/em><\/p>\n<p>Our Disabled People Destroy Science Fiction Special Issue! Guest edited by Dominik Parisen and Elsa Sjunneson-Henry, Nicolette Barischoff,\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/uncanny-magazine-issue-24\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\n\t\t<div class=\"floatinglist\" id=\"oldertitles\">\n\t\t\t<h2><a href=\"https:\/\/weightlessbooks.com\/category\/format\/page\/2\/\">More New Titles<\/a><\/h2>\t\t\n\t\t\t\t\t\t<ul><li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/clarkesworld-magazine-issue-146\/\"><img width=\"75\" height=\"116\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/clarkesworld-magazine-issue-146-cover-75x116.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Clarkesworld Magazine \u2013 Issue 146\" title=\"Clarkesworld Magazine \u2013 Issue 146\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/clarkesworld-magazine-issue-146-cover-75x116.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/clarkesworld-magazine-issue-146-cover-200x309.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/clarkesworld-magazine-issue-146-cover-140x216.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/clarkesworld-magazine-issue-146-cover.jpg 388w\" \/><\/a><div class=\"price\"><b>$3.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9784\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9784\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Clarkesworld Magazine - Issue 146 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Clarkesworld Magazine - Issue 146 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/clarkesworld-magazine-issue-146-cover-75x116.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"3.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"3.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9784\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9784\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/clarkesworld-magazine-issue-146\/\">Clarkesworld Magazine \u2013 Issue 146<\/a> Rich Larson et al.<\/h3>\n\t\t\t\t\t<p><em>Clarkesworld<\/em> is a Hugo and World Fantasy Award-winning science fiction and fantasy magazine. Each month we bring you a mix of fiction (new and classic works), articles, interviews and art.<\/p>\n<p>Our November 2018 issue (#146)\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/clarkesworld-magazine-issue-146\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/forever-magazine-issue-46\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/forever-magazine-issue-46-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Forever Magazine Issue 46\" title=\"Forever Magazine Issue 46\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/forever-magazine-issue-46-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/forever-magazine-issue-46-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/forever-magazine-issue-46-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/forever-magazine-issue-46-cover.jpg 233w\" \/><\/a><div class=\"price\"><b>$2.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9785\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9785\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Forever Magazine Issue 46 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Forever Magazine Issue 46 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/forever-magazine-issue-46-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"2.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"2.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9785\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9785\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/forever-magazine-issue-46\/\">Forever Magazine Issue 46<\/a> Neal Asher et al.<\/h3>\n\t\t\t\t\t<p><em>Forever<\/em> is a monthly science fiction magazine that features previously published stories you might have missed. Each issue will feature a novella, two short stories, and cover art by Ron Guyatt. Edited by the Hugo and\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/forever-magazine-issue-46\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/nonfiction\/locus-november-2018-694\/\"><img width=\"75\" height=\"97\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/locus-november-2018-694-cover-75x97.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Locus November 2018 (#694)\" title=\"Locus November 2018 (#694)\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/locus-november-2018-694-cover-75x97.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/locus-november-2018-694-cover-200x257.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/locus-november-2018-694-cover-140x180.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/locus-november-2018-694-cover.jpg 300w\" \/><\/a><div class=\"price\"><b>$5.50<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9776\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9776\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Locus November 2018 (#694) Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Locus November 2018 (#694) Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"5.50\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"5.50\" \/><input type=\"hidden\" name=\"item_number\" value=\"9776\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9776\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/nonfiction\/locus-november-2018-694\/\">Locus November 2018 (#694)<\/a> Liza Groen Trombi et al.<\/h3>\n\t\t\t\t\t<p>The November 2018 issue of <em>Locus<\/em> magazine has interviews with Andy Duncan and a special feature on SF in Israel with an introduction by Sheldon Teitelbaum, with spotlights on Israeli authors Shimon Adaf, Yael Furman,\u2026 <a href=\"https:\/\/weightlessbooks.com\/nonfiction\/locus-november-2018-694\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/fiction\/shimmer-magazine-issue-46-final-issue\/\"><img width=\"75\" height=\"120\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-75x120.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Shimmer Magazine \u2013 Issue 46 (Final Issue!)\" title=\"Shimmer Magazine \u2013 Issue 46 (Final Issue!)\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-75x120.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-200x320.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-768x1228.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-400x640.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-140x224.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover.jpg 1600w\" \/><\/a><div class=\"price\"><b>$4.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9782\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9782\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Shimmer Magazine - Issue 46 (Final Issue!) Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Shimmer Magazine - Issue 46 (Final Issue!) Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/shimmer-magazine-issue-46-final-issue-cover-75x120.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"4.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"4.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9782\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9782\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/fiction\/shimmer-magazine-issue-46-final-issue\/\">Shimmer Magazine \u2013 Issue 46 (Final Issue!)<\/a> E. Catherine Tobler et al.<\/h3>\n\t\t\t\t\t<p>Come with us, into the woods, where we\u2019ll meet a wolf, a witch, an orchard, a rocking chair, a roller derby, a strip mall, a hungry lake, fairies, cups of tea, a feather, a bit of thistledown, a ghost. The culmination\u2026 <a href=\"https:\/\/weightlessbooks.com\/fiction\/shimmer-magazine-issue-46-final-issue\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/nightmare-magazine-issue-74-november-2018\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Nightmare Magazine, Issue 74 (November 2018)\" title=\"Nightmare Magazine, Issue 74 (November 2018)\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover.jpg 1300w\" \/><\/a><div class=\"price\"><b>$2.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9779\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9779\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Nightmare Magazine, Issue 74 (November 2018) Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Nightmare Magazine, Issue 74 (November 2018) Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/nightmare-magazine-issue-74-november-2018-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"2.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"2.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9779\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9779\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/nightmare-magazine-issue-74-november-2018\/\">Nightmare Magazine, Issue 74 (November 2018)<\/a> John Joseph Adams et al.<\/h3>\n\t\t\t\t\t<p>NIGHTMARE is an online horror and dark fantasy magazine. In NIGHTMARE\u2019s pages, you will find all kinds of horror fiction, from zombie stories and haunted house tales, to visceral psychological horror.<\/p>\n<p>This \u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/nightmare-magazine-issue-74-november-2018\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/lightspeed-magazine-issue-102-november-2018\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Lightspeed Magazine, Issue 102 (November 2018)\" title=\"Lightspeed Magazine, Issue 102 (November 2018)\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/lightspeed-magazine-issue-102-november-2018-cover.jpg 1300w\" \/><\/a><div class=\"price\"><b>$3.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9775\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9775\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Lightspeed Magazine, Issue 102 (November 2018) Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Lightspeed Magazine, Issue 102 (November 2018) Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"3.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"3.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9775\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9775\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/lightspeed-magazine-issue-102-november-2018\/\">Lightspeed Magazine, Issue 102 (November 2018)<\/a> John Joseph Adams et al.<\/h3>\n\t\t\t\t\t<p>LIGHTSPEED is an online science fiction and fantasy magazine. In its pages, you will find science fiction: from near-future, sociological soft SF, to far-future, star-spanning hard SF\u2013and fantasy: from epic\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/lightspeed-magazine-issue-102-november-2018\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/authors\/geoff-ryman-authors\/the-magazine-of-fantasy-and-science-fiction-november-december-2018\/\"><img width=\"75\" height=\"112\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover-75x112.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"The Magazine of Fantasy and Science Fiction \u2013 November\/December 2018\" title=\"The Magazine of Fantasy and Science Fiction \u2013 November\/December 2018\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover-75x112.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover-768x1150.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover-400x599.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover-140x210.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-magazine-of-fantasy-and-science-fiction-november-december-2018-cover.jpg 840w\" \/><\/a><div class=\"price\"><b>$8.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9768\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9768\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"The Magazine of Fantasy and Science Fiction - November\/December 2018 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"The Magazine of Fantasy and Science Fiction - November\/December 2018 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"8.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"8.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9768\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9768\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/authors\/geoff-ryman-authors\/the-magazine-of-fantasy-and-science-fiction-november-december-2018\/\">The Magazine of Fantasy and Science Fiction \u2013 November\/December 2018<\/a> C.C. Finlay et al.<\/h3>\n\t\t\t\t\t<p><i>The Magazine of Fantasy &amp; Science Fiction,<\/i> founded in 1949, is the award-winning SF magazine which is the original publisher of SF classics like Stephen King\u2019s <i>Dark Tower,<\/i> Daniel Keyes\u2019s <i>Flowers <\/i>\u2026 <a href=\"https:\/\/weightlessbooks.com\/authors\/geoff-ryman-authors\/the-magazine-of-fantasy-and-science-fiction-november-december-2018\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/the-dark-issue-42\/\"><img width=\"75\" height=\"116\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-dark-issue-42-cover-75x116.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"The Dark \u2013 Issue 42\" title=\"The Dark \u2013 Issue 42\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-dark-issue-42-cover-75x116.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-dark-issue-42-cover-200x309.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-dark-issue-42-cover-768x1187.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-dark-issue-42-cover-400x618.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/the-dark-issue-42-cover-140x216.jpg 140w\" \/><\/a><div class=\"price\"><b>$2.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9767\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9767\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"The Dark - Issue 42 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"The Dark - Issue 42 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"2.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"2.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9767\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9767\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/the-dark-issue-42\/\">The Dark \u2013 Issue 42<\/a> Sean Wallace<\/h3>\n\t\t\t\t\t<p>Each month <em>The Dark<\/em> brings you the best in dark fantasy and horror! Selected by award-winning editors Silvia Moreno-Garcia and Sean Wallace and published by Prime Books, this issue includes two all-new stories and \u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/the-dark-issue-42\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/flash-fiction-online-issue-62-november-2018\/\"><img width=\"75\" height=\"120\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-75x120.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Flash Fiction Online Issue #62 November 2018\" title=\"Flash Fiction Online Issue #62 November 2018\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-75x120.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-200x320.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-768x1229.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-400x640.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-140x224.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover.jpg 1600w\" \/><\/a><div class=\"price\"><b>$0.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9770\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9770\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Flash Fiction Online Issue #62 November 2018 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Flash Fiction Online Issue #62 November 2018 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/flash-fiction-online-issue-62-november-2018-cover-75x120.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"0.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"0.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9770\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9770\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/flash-fiction-online-issue-62-november-2018\/\">Flash Fiction Online Issue #62 November 2018<\/a> Suzanne Vincent et al.<\/h3>\n\t\t\t\t\t<p>The November 2018 Issue of <em>Flash Fiction Online.<\/em>\u00a0Fantasy, science fiction, horror, and literary short fiction for the modern reader.<\/p>\n<p>A selection of literary short stories from <em>Flash Fiction Online<\/em> curated by Editor-in-Chief\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/flash-fiction-online-issue-62-november-2018\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/galaxys-edge-magazine-issue-35-november-2018\/\"><img width=\"75\" height=\"98\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover-75x98.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Galaxy\u2019s Edge Magazine: Issue 35, November 2018\" title=\"Galaxy\u2019s Edge Magazine: Issue 35, November 2018\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover-75x98.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover-200x260.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover-768x1000.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover-400x521.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover-140x182.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/galaxys-edge-magazine-issue-35-november-2018-cover.jpg 1488w\" \/><\/a><div class=\"price\"><b>$4.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9780\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9780\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Galaxy's Edge Magazine: Issue 35, November 2018  Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Galaxy's Edge Magazine: Issue 35, November 2018  Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"\" \/><input type=\"hidden\" name=\"price\" value=\"4.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"4.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9780\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9780\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/galaxys-edge-magazine-issue-35-november-2018\/\">Galaxy\u2019s Edge Magazine: Issue 35, November 2018<\/a> Mike Resnick et al.<\/h3>\n\t\t\t\t\t<p>A Magazine of Science Fiction and Fantasy<\/p>\n<p>ISSUE 35: November 2018<\/p>\n<p>Mike Resnick, Editor\nTaylor Morris, Copyeditor\nShahid Mahmud, Publisher<\/p>\n<p>Stories by: Brian K. Lowe, Eleanor R. Wood, Harry Turtledove, Larry Hodges,\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/galaxys-edge-magazine-issue-35-november-2018\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-264\/\"><img width=\"75\" height=\"100\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover-75x100.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Beneath Ceaseless Skies Issue #264\" title=\"Beneath Ceaseless Skies Issue #264\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover-75x100.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover-200x267.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover-400x533.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover-140x187.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover.jpg 750w\" \/><\/a><div class=\"price\"><b>$0.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9764\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9764\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Beneath Ceaseless Skies Issue #264 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Beneath Ceaseless Skies Issue #264 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-264-cover-75x100.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"0.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"0.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9764\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9764\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-264\/\">Beneath Ceaseless Skies Issue #264<\/a> Margaret Ronald et al.<\/h3>\n\t\t\t\t\t<p>Issue #264 of <i>Beneath Ceaseless Skies<\/i> online magazine, featuring stories by Margaret Ronald and Jordan Kurella.\u2026 <a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-264\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/publisher\/chizine-publications\/chiteen\/sisters-in-sorrow\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Sisters In Sorrow\" title=\"Sisters In Sorrow\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-140x210.jpg 140w\" \/><\/a><div class=\"price\"><b>$9.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9759\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9759\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Sisters In Sorrow Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Sisters In Sorrow Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/sisters-in-sorrow-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"9.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"9.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9759\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9759\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/publisher\/chizine-publications\/chiteen\/sisters-in-sorrow\/\">Sisters In Sorrow<\/a> Rachael Robie<\/h3>\n\t\t\t\t\t<p>SISTERS IN SORROW is Lord of the Flies on Mars, but with genetic engineering, telepaths, and teenage girls.\u2026 <a href=\"https:\/\/weightlessbooks.com\/publisher\/chizine-publications\/chiteen\/sisters-in-sorrow\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/authors\/tim-major\/machineries-of-mercy\/\"><img width=\"75\" height=\"113\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-75x113.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Machineries of Mercy\" title=\"Machineries of Mercy\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-75x113.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-200x300.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-768x1152.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-400x600.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-140x210.jpg 140w\" \/><\/a><div class=\"price\"><b>$9.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9760\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9760\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Machineries of Mercy Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Machineries of Mercy Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/machineries-of-mercy-cover-75x113.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"9.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"9.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9760\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9760\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/authors\/tim-major\/machineries-of-mercy\/\">Machineries of Mercy<\/a> Tim Major<\/h3>\n\t\t\t\t\t<p><i>Ready Player One<\/i> meets <i>Westworld<\/i> in a sleepy English village.\u2026 <a href=\"https:\/\/weightlessbooks.com\/authors\/tim-major\/machineries-of-mercy\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/new-york-review-of-science-fiction-349\/\"><img width=\"75\" height=\"97\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover-75x97.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"New York Review of Science Fiction #349\" title=\"New York Review of Science Fiction #349\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover-75x97.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover-200x259.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover-400x518.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover-140x181.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover.jpg 555w\" \/><\/a><div class=\"price\"><b>$0<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9762\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9762\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"hidden\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"New York Review of Science Fiction #349 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"New York Review of Science Fiction #349 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/new-york-review-of-science-fiction-349-cover-75x97.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"0\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"0\" \/><input type=\"hidden\" name=\"item_number\" value=\"9762\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9762\"><\/div>\t\t\t\t\n\t\t\t\t<form method=\"post\" class=\"eStore-button-form eStore_mods_download_button\" action=\"\">\n\t\t\t\t<div><input type=\"submit\" value=\"Download\" class=\"eStore_button\" \/><\/div><\/form>\n\t\t\t<\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/new-york-review-of-science-fiction-349\/\">New York Review of Science Fiction #349<\/a> Gardner Dozois et al.<\/h3>\n\t\t\t\t\tSpecial Gardner Dozois Memorial Issue\n\nMemories, anecdotes, appreciations, confessions, and clickbait from Gardner\u2019s friends, fellow travelers, and other assorted fans.\u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/new-york-review-of-science-fiction-349\/\" class=\"read-more\">More<\/a>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/format\/see-the-elephant-issue-4-beyond-death\/\"><img width=\"75\" height=\"117\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-75x117.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"See the Elephant, Issue 4, Beyond Death\" title=\"See the Elephant, Issue 4, Beyond Death\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-75x117.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-200x311.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-768x1196.jpg 768w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-400x623.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-140x218.jpg 140w\" \/><\/a><div class=\"price\"><b>$2.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9750\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9750\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"See the Elephant, Issue 4, Beyond Death Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"See the Elephant, Issue 4, Beyond Death Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/see-the-elephant-issue-4-beyond-death-cover-75x117.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"2.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"2.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9750\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9750\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/format\/see-the-elephant-issue-4-beyond-death\/\">See the Elephant, Issue 4, Beyond Death<\/a> Melanie Lamaga et al.<\/h3>\n\t\t\t\t\t<p>Under an authoritarian regime, artists are blinded and women\u2019s faces are surgically altered into bone-flowers. Aliens arrive on earth and become obsessed with poignant, tragic humans\u2014the only species \u2026 <a href=\"https:\/\/weightlessbooks.com\/format\/see-the-elephant-issue-4-beyond-death\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t\t\t\t\t\t<li class=\"product\">\n\t\t\t\t\t<div class=\"cover\"><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-263\/\"><img width=\"75\" height=\"100\" src=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover-75x100.jpg\" class=\"attachment-sm size-sm wp-post-image\" alt=\"Beneath Ceaseless Skies Issue #263\" title=\"Beneath Ceaseless Skies Issue #263\" srcset=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover-75x100.jpg 75w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover-200x267.jpg 200w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover-400x533.jpg 400w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover-140x187.jpg 140w, https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover.jpg 750w\" \/><\/a><div class=\"price\"><b>$0.99<\/b><br \/><\/div>\n\t\t<div class=\"buy_buttons\" id=\"9748\"><form method=\"post\" action=\"\" class=\"eStore-button-form eStore-button-form-9748\"><div class=\"eStore_variation_top\"><\/div><div>Formats : <\/div><div><br \/><input type=\"hidden\" name=\"add_qty\" value=\"1\" \/><input type=\"submit\" value=\"Add to Cart\" class=\"eStore_button eStore_add_to_cart_button\" \/><input type=\"hidden\" name=\"estore_product_name\" value=\"Beneath Ceaseless Skies Issue #263 Ebook\" \/><input type=\"hidden\" name=\"product_name_tmp1\" value=\"Beneath Ceaseless Skies Issue #263 Ebook\" \/><input type=\"hidden\" name=\"thumbnail_url\" value=\"https:\/\/weightlessbooks.com\/wp-content\/uploads\/images\/beneath-ceaseless-skies-issue-263-cover-75x100.jpg\" \/><input type=\"hidden\" name=\"price\" value=\"0.99\" \/><input type=\"hidden\" name=\"price_tmp1\" value=\"0.99\" \/><input type=\"hidden\" name=\"item_number\" value=\"9748\" \/><input type=\"hidden\" name=\"shipping\" value=\"\" \/><input type=\"hidden\" name=\"tax\" value=\"\" \/><input type=\"hidden\" name=\"addcart_eStore\" value=\"1\" \/><input type=\"hidden\" name=\"cartLink\" value=\"https:\/\/weightlessbooks.com\/\" \/><input type=\"hidden\" name=\"digital_flag\" value=\"1\" \/><\/div><\/form><div class=\"eStore_item_added_msg eStore_item_added_msg-9748\"><\/div><\/div><\/div>\t\t\t\t\t<h3><a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-263\/\">Beneath Ceaseless Skies Issue #263<\/a> Megan Arkenberg et al.<\/h3>\n\t\t\t\t\t<p>Issue #263 of <i>Beneath Ceaseless Skies<\/i> online magazine, featuring stories by Megan Arkenberg and Karen Osborne.\u2026 <a href=\"https:\/\/weightlessbooks.com\/fiction\/beneath-ceaseless-skies-issue-263\/\" class=\"read-more\">More<\/a><\/p>\t\t\t\t\t<div class=\"clear\"><\/div>\n\t\t\t\t<\/li>\n\t\t\t\t\t\t<\/ul><br \/><\/div>\n\t\t\t\t\n\t\n \t\t<div class=\"clear\"><\/div>","created_at":"2018-11-07T16:49:24+0000","updated_at":"2018-11-07T16:49:24+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":9,"domain_name":"weightlessbooks.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":466,"title":"Web Scraping into an E-book with BeautifulSoup and Pandoc \u00ab null program","url":"https:\/\/nullprogram.com\/blog\/2017\/05\/15\/","content":"<time datetime=\"2017-05-15\">\n    May 15, 2017\n  <\/time><p>nullprogram.com\/blog\/2017\/05\/15\/<\/p><p>I recently learned how to use <a href=\"https:\/\/www.crummy.com\/software\/BeautifulSoup\/\">BeautifulSoup<\/a>, a Python library for\nmanipulating HTML and XML parse trees, and it\u2019s been a fantastic\naddition to my virtual toolbelt. In the past when I\u2019ve needed to process\nraw HTML, I\u2019ve tried nasty hacks with Unix pipes, or <a href=\"https:\/\/nullprogram.com\/blog\/2013\/01\/24\/\">routing the\ncontent through a web browser<\/a> so that I could manipulate it via\nthe DOM API. None of that worked very well, but now I finally have\nBeautifulSoup to fill that gap. It\u2019s got a selector interface and,\nexcept for rendering, it\u2019s basically as comfortable with HTML as\nJavaScript.<\/p><p>Today\u2019s problem was that I wanted to read <a href=\"http:\/\/daviddfriedman.blogspot.com\/2017\/05\/something-different-or-maybe-not.html\">a recommended<\/a> online\nbook called <a href=\"https:\/\/banter-latte.com\/portfolio\/interviewing-leather\/\"><em>Interviewing Leather<\/em><\/a>, a story set \u201cin a world where\ncaped heroes fight dastardly villains on an everyday basis.\u201d I say\n\u201conline book\u201d because the 39,403 word story is distributed as a series\nof 14 blog posts. I\u2019d rather not read it on the website in a browser,\ninstead preferring it in e-book form where it\u2019s more comfortable. The\n<a href=\"https:\/\/nullprogram.com\/blog\/2015\/09\/03\/\">last time I did this<\/a>, I manually scraped the entire book into\nMarkdown, spent a couple of weeks editing it for mistakes, and finally\nsent the Markdown to <a href=\"http:\/\/pandoc.org\/\">Pandoc<\/a> to convert into an e-book.<\/p><p>For this book, I just want a quick-and-dirty scrape in order to shift\nformats. I\u2019ve never read it and I may not even like it (<em>update<\/em>: I\nenjoyed it), so I definitely don\u2019t want to spend much time on the\nconversion. Despite <a href=\"https:\/\/nullprogram.com\/blog\/2017\/04\/01\/\">having fun with typing lately<\/a>, I\u2019d also\nprefer to keep all the formating \u2014 italics, etc. \u2014 without re-entering\nit all manually.<\/p><p>Fortunately Pandoc can consume HTML as input, so, in theory, I can feed\nit the original HTML and preserve all of the original markup. The\nchallenge is that the HTML is spread across 14 pages surrounded by all\nthe expected blog cruft. I need some way to extract the book content\nfrom each page, concatenate it together along with chapter headings, and\nsend the result to Pandoc. Enter BeautifulSoup.<\/p><p>First, I need to construct the skeleton HTML document. Rather than code\nmy own HTML, I\u2019m going to build it with BeautifulSoup. I start by\ncreating a completely empty document and adding a doctype to it.<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>from bs4 import BeautifulSoup, Doctype\ndoc = BeautifulSoup()\ndoc.append(Doctype('html'))\n<\/pre><\/div><\/div><p>Next I create the <code class=\"highlighter-rouge\">html<\/code> root element, then add the <code class=\"highlighter-rouge\">head<\/code> and <code class=\"highlighter-rouge\">body<\/code>\nelements. I also add a <code class=\"highlighter-rouge\">title<\/code> element. The original content has fancy\nUnicode markup \u2014 left and right quotation marks, em dash, etc. \u2014 so it\u2019s\nimportant to declare the page as UTF-8, since otherwise these characters\nare likely to be interpreted incorrectly. It always feels odd declaring\nthe encoding within the content being encoded, but that\u2019s just the way\nthings are.<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>html = doc.new_tag('html', lang='en-US')\ndoc.append(html)\nhead = doc.new_tag('head')\nhtml.append(head)\nmeta = doc.new_tag('meta', charset='utf-8')\nhead.append(meta)\ntitle = doc.new_tag('title')\ntitle.string = 'Interviewing Leather'\nhead.append(title)\nbody = doc.new_tag('body')\nhtml.append(body)\n<\/pre><\/div><\/div><p>If I <code class=\"highlighter-rouge\">print(doc.prettify())<\/code> then I see the skeleton I want:<\/p><div class=\"language-html highlighter-rouge\"><div class=\"highlight\"><pre>&lt;!DOCTYPE html&gt;\n&lt;html lang=\"en-US\"&gt;\n &lt;head&gt;\n  &lt;meta charset=\"utf-8\"\/&gt;\n  &lt;title&gt;\n   Interviewing Leather\n  &lt;\/title&gt;\n &lt;\/head&gt;\n &lt;body&gt;\n &lt;\/body&gt;\n&lt;\/html&gt;\n<\/pre><\/div><\/div><p>Next, I assemble a list of the individual blog posts. When I was\nactually writing the script, I first downloaded them locally with <a href=\"https:\/\/nullprogram.com\/blog\/2016\/06\/16\/\">my\nfavorite download tool<\/a>, curl, and ran the script against local\ncopies. I didn\u2019t want to hit the web server each time I tested. (Note:\nI\u2019ve truncated these URLs to fit in this article.)<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>chapters = [\n    \"https:\/\/banter-latte.com\/2007\/06\/26\/...\",\n    \"https:\/\/banter-latte.com\/2007\/07\/03\/...\",\n    \"https:\/\/banter-latte.com\/2007\/07\/10\/...\",\n    \"https:\/\/banter-latte.com\/2007\/07\/17\/...\",\n    \"https:\/\/banter-latte.com\/2007\/07\/24\/...\",\n    \"https:\/\/banter-latte.com\/2007\/07\/31\/...\",\n    \"https:\/\/banter-latte.com\/2007\/08\/07\/...\",\n    \"https:\/\/banter-latte.com\/2007\/08\/14\/...\",\n    \"https:\/\/banter-latte.com\/2007\/08\/21\/...\",\n    \"https:\/\/banter-latte.com\/2007\/08\/28\/...\",\n    \"https:\/\/banter-latte.com\/2007\/09\/04\/...\",\n    \"https:\/\/banter-latte.com\/2007\/09\/20\/...\",\n    \"https:\/\/banter-latte.com\/2007\/09\/25\/...\",\n    \"https:\/\/banter-latte.com\/2007\/10\/02\/...\"\n]\n<\/pre><\/div><\/div><p>I visit a few of these pages in my browser to determine which part of\nthe page I want to extract. I want to look closely enough to see what\nI\u2019m doing, but not <em>too<\/em> closely as to not spoil myself! Right clicking\nthe content in the browser and selecting \u201cInspect Element\u201d (Firefox) or\n\u201cInspect\u201d (Chrome) pops up a pane to structurally navigate the page.\n\u201cView Page Source\u201d would work, too, especially since this is static\ncontent, but I find the developer pane easier to read. Plus it hides\nmost of the content, revealing only the structure.<\/p><p>The content is contained in a <code class=\"highlighter-rouge\">div<\/code> with the class <code class=\"highlighter-rouge\">entry-content<\/code>. I\ncan use a selector to isolate this element and extract its child <code class=\"highlighter-rouge\">p<\/code>\nelements. However, it\u2019s not quite so simple. Each chapter starts with a\nbit of commentary that\u2019s not part of the book, and I don\u2019t want to\ninclude in my extract. It\u2019s separated from the real content by an <code class=\"highlighter-rouge\">hr<\/code>\nelement. There\u2019s also a footer below another <code class=\"highlighter-rouge\">hr<\/code> element, likely put\nthere by someone who wasn\u2019t paying attention to the page structure. It\u2019s\nnot quite the shining example of semantic markup, but it\u2019s regular\nenough I can manage.<\/p><div class=\"language-html highlighter-rouge\"><div class=\"highlight\"><pre>&lt;body&gt;\n  &lt;main class=\"site-main\"&gt;\n    &lt;div class=\"entry-body\"&gt;\n      &lt;div class=\"entry-content\"&gt;\n        &lt;p&gt;A little intro.&lt;\/p&gt;\n        &lt;p&gt;Some more intro.&lt;\/p&gt;\n        &lt;hr\/&gt;\n        &lt;p&gt;Actual book content.&lt;\/p&gt;\n        &lt;p&gt;More content.&lt;\/p&gt;\n        &lt;hr\/&gt;\n        &lt;p&gt;Footer navigation junk.&lt;\/p&gt;\n      &lt;\/div&gt;\n    &lt;\/div&gt;\n  &lt;\/main&gt;\n&lt;\/body&gt;\n<\/pre><\/div><\/div><p>The next step is visiting each of these pages. I use <code class=\"highlighter-rouge\">enumerate<\/code> since I\nwant the chapter numbers when inserting <code class=\"highlighter-rouge\">h1<\/code> chapter elements. Pandoc\nwill use these to build the table of contents.<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>for i, chapter in enumerate(chapters):\n    # Construct h1 for the chapter\n    header = doc.new_tag('h1')\n    header.string = 'Chapter %d' % (i + 1,)\n    body.append(header)\n<\/pre><\/div><\/div><p>Next grab the page content using <code class=\"highlighter-rouge\">urllib<\/code> and parse it with\nBeautifulSoup. I\u2019m using a selector to locate the <code class=\"highlighter-rouge\">div<\/code> with the\nbook content.<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>    # Load chapter content\n    with urllib.request.urlopen(chapter) as url:\n        page = BeautifulSoup(url)\n    content = page.select('.entry-content')[0]\n<\/pre><\/div><\/div><p>Finally I iterate over the child elements of the <code class=\"highlighter-rouge\">div.entry-content<\/code>\nelement. I keep a running count of the <code class=\"highlighter-rouge\">hr<\/code> element and only extract\ncontent when we\u2019ve seen exactly one <code class=\"highlighter-rouge\">hr<\/code> element.<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>    # Append content between hr elements\n    hr_count = 0\n    for child in content.children:\n        if (child.name == 'hr'):\n            hr_count += 1\n        elif (child.name == 'p' and hr_count == 1):\n            child.attrs = {}\n            if (child.string == '#'):\n                body.append(doc.new_tag('hr'))\n            else:\n                body.append(child)\n<\/pre><\/div><\/div><p>If it\u2019s a <code class=\"highlighter-rouge\">p<\/code> element, I copy it into the output document, taking a\nmoment to strip away any attributes present on the <code class=\"highlighter-rouge\">p<\/code> tag, since, for\nsome reason, some of these elements have old-fashioned alignment\nattributes in the original content.<\/p><p>The original content also uses the text \u201c<code class=\"highlighter-rouge\">#<\/code>\u201d by itself in a <code class=\"highlighter-rouge\">p<\/code> to\nseparate sections rather than using the appropriate markup. Despite\nbeing semantically incorrect, I\u2019m thankful for this since more <code class=\"highlighter-rouge\">hr<\/code>\nelements would have complicated matters further. I convert these to the\ncorrect markup for the final document.<\/p><p>Finally I pretty print the result:<\/p><div class=\"language-py highlighter-rouge\"><div class=\"highlight\"><pre>print(doc.prettify())\n<\/pre><\/div><\/div><p>Alternatively I could pipe it through <a href=\"http:\/\/tidy.sourceforge.net\/\">tidy<\/a>.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ python3 extract.py | tidy -indent -utf8 &gt; output.html\n<\/pre><\/div><\/div><p>A brief inspection with a browser indicates that everything seems to\nhave come out correctly. I won\u2019t know for sure, though, until I actually\nread through the whole book. Finally I have Pandoc perform the\nconversion.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ pandoc -t epub3 -o output.epub output.html\n<\/pre><\/div><\/div><p>And that\u2019s it! It\u2019s ready to read offline in my e-book reader of\nchoice. The crude version of my script took around 15\u201320 minutes to\nwrite and test, so I had an e-book conversion in under 30 minutes.\nThat\u2019s about as long as I was willing to spend to get it. Tidying the\nscript up for this article took a lot longer.<\/p><p>I don\u2019t have permission to share the resulting e-book, but I can share\nmy script so that you can generate your own, at least as long as it\u2019s\nhosted at the same place with the same structure.<\/p><ul><li><a href=\"https:\/\/nullprogram.com\/download\/leather\/extract.py\" class=\"download\">extract.py<\/a><\/li>\n<\/ul>","created_at":"2018-11-07T16:49:24+0000","updated_at":"2018-11-07T16:49:24+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"nullprogram.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":467,"title":"Functors, Applicatives, And Monads In Pictures","url":"http:\/\/adit.io\/posts\/2013-04-17-functors,_applicatives,_and_monads_in_pictures.html","content":"<ul id=\"sections\">\n            <li><a href=\"#functors\">Functors<\/a><\/li>\n            <li><a href=\"#just-what-is-a-functor,-really?\">Just What Is A Functor, Really?<\/a><\/li>\n            <li><a href=\"#applicatives\">Applicatives<\/a><\/li>\n            <li><a href=\"#monads\">Monads<\/a><\/li>\n            <li><a href=\"#conclusion\">Conclusion<\/a><\/li>\n            <li><a href=\"#translations\">Translations<\/a><\/li>\n        <\/ul><time>Written April 17, 2013<\/time><p>updated: May 20, 2013<\/p><p>Here's a simple value:\n<img src=\"http:\/\/adit.io\/imgs\/functors\/value.png\" alt=\"\" \/><\/p><p>And we know how to apply a function to this value:\n<img src=\"http:\/\/adit.io\/imgs\/functors\/value_apply.png\" alt=\"\" \/><\/p><p>Simple enough. Lets extend this by saying that any value can be in a context. For now you can think of a context as a box that you can put a value in:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/value_and_context.png\" alt=\"\" \/><\/p><p>Now when you apply a function to this value, you'll get different results <strong>depending on the context<\/strong>. This is the idea that Functors, Applicatives, Monads, Arrows etc are all based on. The <code>Maybe<\/code> data type defines two related contexts:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/context.png\" alt=\"\" \/><\/p><div class=\"highlight\"><pre>data Maybe a = Nothing | Just a\n<\/pre><\/div><p>In a second we'll see how function application is different when something is a <code>Just a<\/code> versus a <code>Nothing<\/code>. First let's talk about Functors!<\/p><h2 id=\"functors\">Functors<\/h2><p>When a value is wrapped in a context, you can't apply a normal function to it:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/no_fmap_ouch.png\" alt=\"\" \/><\/p><p>This is where <code>fmap<\/code> comes in. <code>fmap<\/code> is from the street, <code>fmap<\/code> is hip to contexts. <code>fmap<\/code> knows how to apply functions to values that are wrapped in a context. For example, suppose you want to apply <code>(+3)<\/code> to <code>Just 2<\/code>. Use <code>fmap<\/code>:<\/p><div class=\"highlight\"><pre>&gt; fmap (+3) (Just 2)\nJust 5\n<\/pre><\/div><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap_apply.png\" alt=\"\" \/><\/p><p><strong>Bam!<\/strong> <code>fmap<\/code> shows us how it's done! But how does <code>fmap<\/code> know how to apply the function?<\/p><h2 id=\"just-what-is-a-functor,-really?\">Just what is a Functor, really?<\/h2><p><code>Functor<\/code> is a <a href=\"http:\/\/learnyouahaskell.com\/types-and-typeclasses#typeclasses-101\">typeclass<\/a>. Here's the definition:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/functor_def.png\" alt=\"\" \/><\/p><p>A <code>Functor<\/code> is any data type that defines how <code>fmap<\/code> applies to it. Here's how <code>fmap<\/code> works:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap_def.png\" alt=\"\" \/><\/p><p>So we can do this:<\/p><div class=\"highlight\"><pre>&gt; fmap (+3) (Just 2)\nJust 5\n<\/pre><\/div><p>And <code>fmap<\/code> magically applies this function, because <code>Maybe<\/code> is a Functor. It specifies how <code>fmap<\/code> applies to <code>Just<\/code>s and <code>Nothing<\/code>s:<\/p><div class=\"highlight\"><pre>instance Functor Maybe where\n    fmap func (Just val) = Just (func val)\n    fmap func Nothing = Nothing\n<\/pre><\/div><p>Here's what is happening behind the scenes when we write <code>fmap (+3) (Just 2)<\/code>:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap_just.png\" alt=\"\" \/><\/p><p>So then you're like, alright <code>fmap<\/code>, please apply <code>(+3)<\/code> to a <code>Nothing<\/code>?<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap_nothing.png\" alt=\"\" \/><\/p><div class=\"highlight\"><pre>&gt; fmap (+3) Nothing\nNothing\n<\/pre><\/div><p><img src=\"http:\/\/adit.io\/imgs\/functors\/bill.png\" alt=\"Bill O'Reilly being totally ignorant about the Maybe functor\" \/><\/p><p>Like Morpheus in the Matrix, <code>fmap<\/code> knows just what to do; you start with <code>Nothing<\/code>, and you end up with <code>Nothing<\/code>! <code>fmap<\/code> is zen. Now it makes sense why the <code>Maybe<\/code> data type exists. For example, here's how you work with a database record in a language without <code>Maybe<\/code>:<\/p><div class=\"highlight\"><pre>post = Post.find_by_id(1)\nif post\n  return post.title\nelse\n  return nil\nend\n<\/pre><\/div><p>But in Haskell:<\/p><div class=\"highlight\"><pre>fmap (getPostTitle) (findPost 1)\n<\/pre><\/div><p>If <code>findPost<\/code> returns a post, we will get the title with <code>getPostTitle<\/code>. If it returns <code>Nothing<\/code>, we will return <code>Nothing<\/code>! Pretty neat, huh? <code> is the infix version of <code>fmap<\/code>, so you will often see this instead:<\/code><\/p><div class=\"highlight\"><pre>getPostTitle &lt;$&gt; (findPost 1)\n<\/pre><\/div><p>Here's another example: what happens when you apply a function to a list?<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap_list.png\" alt=\"\" \/><\/p><p>Lists are functors too! Here's the definition:<\/p><div class=\"highlight\"><pre>instance Functor [] where\n    fmap = map\n<\/pre><\/div><p>Okay, okay, one last example: what happens when you apply a function to another function?<\/p><div class=\"highlight\"><pre>fmap (+3) (+1)\n<\/pre><\/div><p>Here's a function:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/function_with_value.png\" alt=\"\" \/><\/p><p>Here's a function applied to another function:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap_function.png\" alt=\"\" \/><\/p><p>The result is just another function!<\/p><div class=\"highlight\"><pre>&gt; import Control.Applicative\n&gt; let foo = fmap (+3) (+2)\n&gt; foo 10\n15\n<\/pre><\/div><p>So functions are Functors too!<\/p><div class=\"highlight\"><pre>instance Functor ((-&gt;) r) where\n    fmap f g = f . g\n<\/pre><\/div><p>When you use fmap on a function, you're just doing function composition!<\/p><h2 id=\"applicatives\">Applicatives<\/h2><p>Applicatives take it to the next level. With an applicative, our values are wrapped in a context, just like Functors:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/value_and_context.png\" alt=\"\" \/><\/p><p>But our functions are wrapped in a context too!<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/function_and_context.png\" alt=\"\" \/><\/p><p>Yeah. Let that sink in. Applicatives don't kid around. <code>Control.Applicative<\/code> defines <code>, which knows how to apply a function <em>wrapped in a context<\/em> to a value <em>wrapped in a context<\/em>:<\/code><\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/applicative_just.png\" alt=\"\" \/><\/p><p>i.e:<\/p><div class=\"highlight\"><pre>Just (+3) &lt;*&gt; Just 2 == Just 5\n<\/pre><\/div><p>Using <code> can lead to some interesting situations. For example:<\/code><\/p><div class=\"highlight\"><pre>&gt; [(*2), (+3)] &lt;*&gt; [1, 2, 3]\n[2, 4, 6, 4, 5, 6]\n<\/pre><\/div><p><img src=\"http:\/\/adit.io\/imgs\/functors\/applicative_list.png\" alt=\"\" \/><\/p><p>Here's something you can do with Applicatives that you can't do with Functors. How do you apply a function that takes two arguments to two wrapped values?<\/p><div class=\"highlight\"><pre>&gt; (+) &lt;$&gt; (Just 5)\nJust (+5)\n&gt; Just (+5) &lt;$&gt; (Just 4)\nERROR ??? WHAT DOES THIS EVEN MEAN WHY IS THE FUNCTION WRAPPED IN A JUST\n<\/pre><\/div><p>Applicatives:<\/p><div class=\"highlight\"><pre>&gt; (+) &lt;$&gt; (Just 5)\nJust (+5)\n&gt; Just (+5) &lt;*&gt; (Just 3)\nJust 8\n<\/pre><\/div><p><code>Applicative<\/code> pushes <code>Functor<\/code> aside. \"Big boys can use functions with any number of arguments,\" it says. \"Armed <code> and <code>, I can take any function that expects any number of unwrapped values. Then I pass it all wrapped values, and I get a wrapped value out! AHAHAHAHAH!\"<\/code><\/code><\/p><div class=\"highlight\"><pre>&gt; (*) &lt;$&gt; Just 5 &lt;*&gt; Just 3\nJust 15\n<\/pre><\/div><p>And hey! There's a function called <code>liftA2<\/code> that does the same thing:<\/p><div class=\"highlight\"><pre>&gt; liftA2 (*) (Just 5) (Just 3)\nJust 15\n<\/pre><\/div><h2 id=\"monads\">Monads<\/h2><p>How to learn about Monads:<\/p><ol><li>Get a PhD in computer science.<\/li>\n<li>Throw it away because you don't need it for this section!<\/li>\n<\/ol><p>Monads add a new twist.<\/p><p>Functors apply a function to a wrapped value:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/fmap.png\" alt=\"\" \/><\/p><p>Applicatives apply a wrapped function to a wrapped value:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/applicative.png\" alt=\"\" \/><\/p><p>Monads apply a function <strong>that returns a wrapped value<\/strong> to a wrapped value. Monads have a function <code>&gt;&gt;=<\/code> (pronounced \"bind\") to do this.<\/p><p>Let's see an example. Good ol' <code>Maybe<\/code> is a monad:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/context.png\" alt=\"Just a monad hanging out\" \/><\/p><p>Suppose <code>half<\/code> is a function that only works on even numbers:<\/p><div class=\"highlight\"><pre>half x = if even x\n           then Just (x `div` 2)\n           else Nothing\n<\/pre><\/div><p><img src=\"http:\/\/adit.io\/imgs\/functors\/half.png\" alt=\"\" \/><\/p><p>What if we feed it a wrapped value?<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/half_ouch.png\" alt=\"\" \/><\/p><p>We need to use <code>&gt;&gt;=<\/code> to shove our wrapped value into the function. Here's a photo of <code>&gt;&gt;=<\/code>:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/plunger.jpg\" alt=\"\" \/><\/p><p>Here's how it works:<\/p><div class=\"highlight\"><pre>&gt; Just 3 &gt;&gt;= half\nNothing\n&gt; Just 4 &gt;&gt;= half\nJust 2\n&gt; Nothing &gt;&gt;= half\nNothing\n<\/pre><\/div><p>What's happening inside? <code>Monad<\/code> is another typeclass. Here's a partial definition:<\/p><div class=\"highlight\"><pre>class Monad m where\n    (&gt;&gt;=) :: m a -&gt; (a -&gt; m b) -&gt; m b\n<\/pre><\/div><p>Where <code>&gt;&gt;=<\/code> is:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/bind_def.png\" alt=\"\" \/><\/p><p>So <code>Maybe<\/code> is a Monad:<\/p><div class=\"highlight\"><pre>instance Monad Maybe where\n    Nothing &gt;&gt;= func = Nothing\n    Just val &gt;&gt;= func  = func val\n<\/pre><\/div><p>Here it is in action with a <code>Just 3<\/code>!<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/monad_just.png\" alt=\"\" \/><\/p><p>And if you pass in a <code>Nothing<\/code> it's even simpler:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/monad_nothing.png\" alt=\"\" \/><\/p><p>You can also chain these calls:<\/p><div class=\"highlight\"><pre>&gt; Just 20 &gt;&gt;= half &gt;&gt;= half &gt;&gt;= half\nNothing\n<\/pre><\/div><p><img src=\"http:\/\/adit.io\/imgs\/functors\/monad_chain.png\" alt=\"\" \/><\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/whoa.png\" alt=\"\" \/><\/p><p>Cool stuff! So now we know that <code>Maybe<\/code> is a <code>Functor<\/code>, an <code>Applicative<\/code>, and a <code>Monad<\/code>.<\/p><p>Now let's mosey on over to another example: the <code>IO<\/code> monad:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/io.png\" alt=\"\" \/><\/p><p>Specifically three functions. <code>getLine<\/code> takes no arguments and gets user input:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/getLine.png\" alt=\"\" \/><\/p><div class=\"highlight\"><pre>getLine :: IO String\n<\/pre><\/div><p><code>readFile<\/code> takes a string (a filename) and returns that file's contents:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/readFile.png\" alt=\"\" \/><\/p><div class=\"highlight\"><pre>readFile :: FilePath -&gt; IO String\n<\/pre><\/div><p><code>putStrLn<\/code> takes a string and prints it:<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/putStrLn.png\" alt=\"\" \/><\/p><div class=\"highlight\"><pre>putStrLn :: String -&gt; IO ()\n<\/pre><\/div><p>All three functions take a regular value (or no value) and return a wrapped value. We can chain all of these using <code>&gt;&gt;=<\/code>!<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/monad_io.png\" alt=\"\" \/><\/p><div class=\"highlight\"><pre>getLine &gt;&gt;= readFile &gt;&gt;= putStrLn\n<\/pre><\/div><p>Aw yeah! Front row seats to the monad show!<\/p><p>Haskell also provides us with some syntactical sugar for monads, called <code>do<\/code> notation:<\/p><div class=\"highlight\"><pre>foo = do\n    filename &lt;- getLine\n    contents &lt;- readFile filename\n    putStrLn contents\n<\/pre><\/div><h2 id=\"conclusion\">Conclusion<\/h2><ol><li>A functor is a data type that implements the <code>Functor<\/code> typeclass.<\/li>\n<li>An applicative is a data type that implements the <code>Applicative<\/code> typeclass.<\/li>\n<li>A monad is a data type that implements the <code>Monad<\/code> typeclass.<\/li>\n<li>A <code>Maybe<\/code> implements all three, so it is a functor, an applicative, <em>and<\/em> a monad.<\/li>\n<\/ol><p>What is the difference between the three?<\/p><p><img src=\"http:\/\/adit.io\/imgs\/functors\/recap.png\" alt=\"\" \/><\/p><ul><li><strong>functors:<\/strong> you apply a function to a wrapped value using <code>fmap<\/code> or <code><\/code><\/li>\n<li><strong>applicatives:<\/strong> you apply a wrapped function to a wrapped value using <code> or <code>liftA<\/code><\/code><\/li>\n<li><strong>monads:<\/strong> you apply a function that returns a wrapped value, to a wrapped value using <code>&gt;&gt;=<\/code> or <code>liftM<\/code><\/li>\n<\/ul><p>So, dear friend (I think we are friends by this point), I think we both agree that monads are easy and a SMART IDEA(tm). Now that you've wet your whistle on this guide, why not pull a Mel Gibson and grab the whole bottle. Check out LYAH's <a href=\"http:\/\/learnyouahaskell.com\/a-fistful-of-monads\">section on Monads<\/a>. There's a lot of things I've glossed over because Miran does a great job going in-depth with this stuff.<\/p><h2 id=\"translations\">Translations<\/h2><p>This post has been translated into:<\/p><p>Human languages:<\/p><ul><li><a href=\"http:\/\/jiyinyiyong.github.io\/monads-in-pictures\/\">Chinese<\/a><\/li>\n<li><a href=\"http:\/\/blog.forec.cn\/2017\/03\/02\/translation-adit-faamip\/\">Another Chinese translation<\/a><\/li>\n<li><a href=\"https:\/\/hltj.me\/kotlin\/2017\/08\/25\/kotlin-functor-applicative-monad-cn.html\">Chinese, Kotlin<\/a><\/li>\n<li><a href=\"http:\/\/www.leonardmeyer.com\/blog\/2014\/06\/functors-applicatives-et-monads-en-images\/\">French<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/madnight\/monad-in-pictures-german\">German<\/a><\/li>\n<li><a href=\"http:\/\/qiita.com\/suin\/items\/0255f0637921dcdfe83b\">Japanese<\/a><\/li>\n<li><a href=\"http:\/\/lazyswamp.tistory.com\/entry\/functorsapplicativesandmonadsinpictures\">Korean<\/a><\/li>\n<li><a href=\"https:\/\/medium.com\/@julianoalves\/functors-applicatives-e-monads-explicados-com-desenhos-2c45d5db7d25#.oxtev31qu\">Portuguese<\/a><\/li>\n<li><a href=\"http:\/\/habrahabr.ru\/post\/183150\/\">Russian<\/a><\/li>\n<li><a href=\"https:\/\/medium.com\/@miguelsaddress\/funtores-aplicativos-y-m%C3%B3nadas-en-im%C3%A1genes-21ab0e60fe23#.azxc90mox\">Spanish<\/a><\/li>\n<li><a href=\"http:\/\/rimbi.github.io\/functors-applicatives-monads-in-pictures.html\">Turkish<\/a><\/li>\n<li><a href=\"http:\/\/zinh.github.io\/haskell\/2015\/09\/16\/functors-applicatives-monads-in-pictures.html\">Vietnamese<\/a><\/li>\n<\/ul><p>Programming languages:<\/p><ul><li><a href=\"https:\/\/medium.com\/@tzehsiang\/javascript-functor-applicative-monads-in-pictures-b567c6415221#.rdwll124i\">Javascript<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/dbrattli\/OSlash\/wiki\/Functors,-Applicatives,-And-Monads-In-Pictures\">Python<\/a><\/li>\n<li><a href=\"http:\/\/www.mokacoding.com\/blog\/functor-applicative-monads-in-pictures\">Swift<\/a><\/li>\n<li><a href=\"https:\/\/hltj.me\/kotlin\/2017\/08\/25\/kotlin-functor-applicative-monad.html\">Kotlin<\/a>. This author also translated this Kotlin version into <a href=\"https:\/\/hltj.me\/kotlin\/2017\/08\/25\/kotlin-functor-applicative-monad-cn.html\">Chinese<\/a>.<\/li>\n<li><a href=\"https:\/\/medium.com\/@aballano\/kotlin-functors-applicatives-and-monads-in-pictures-part-1-3-c47a1b1ce251\">Kotlin (translated from the Swift translation)<\/a><\/li>\n<li><a href=\"https:\/\/medium.com\/@l.mugnaini\/functors-applicatives-and-monads-in-pictures-784c2b5786f7\">Elm<\/a><\/li>\n<\/ul><p>If you translate this post, <a href=\"http:\/\/www.adit.io\">send me an email<\/a> and I'll add it to this list!<\/p><p>For more monads and pictures, check out <a href=\"http:\/\/adit.io\/posts\/2013-06-10-three-useful-monads.html\">three useful monads<\/a>.<\/p><noscript><div>Please enable JavaScript to view the <a href=\"http:\/\/disqus.com\/?ref_noscript\">comments powered by Disqus.<\/a><\/div><\/noscript><a href=\"http:\/\/disqus.com\" class=\"dsq-brlink\">blog comments powered by Disqus<\/a>","created_at":"2018-11-07T16:49:24+0000","updated_at":"2018-11-07T16:49:24+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"adit.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":464,"title":"The Freelancer's Guide To Niching Your Business - Double Your Freelancing","url":"https:\/\/doubleyourfreelancing.com\/freelancers-guide-niching-business\/","content":"<p>\u201cChoosing a niche,\u201d or coming up with strong business positioning, is one of the most difficult exercises for many of the freelancers I\u2019ve worked with over the years.<\/p><p>I think they point to a big disconnect between what niching <strong>actually is<\/strong> and what niching is <strong>thought to be<\/strong>.<\/p><h2>What does it mean to niche?<\/h2><p>Every transaction requires an <em>application<\/em>.<\/p><p>If I\u2019m going to buy a new computer, justifying the purchase requires me to think about how I\u2019ll use the laptop to better my business or my life. Why I end up buying a laptop is different than why you might buy that same laptop.<\/p><p>Likewise, hiring a freelancer (for the sake of argument, let\u2019s say a freelance web designer) requires the business owner to rationalize <em>how<\/em> they\u2019ll use the web designer to achieve the end they have in mind.<\/p><p>Consider for a moment the lifecycle that goes into hiring somebody like you:<\/p><ol><li>A business becomes aware of a problem that they have<\/li>\n<li>They realize that this problem can be solved<\/li>\n<li>They then determine someone like you can solve the problem<\/li>\n<li>They seek out people like you<\/li>\n<\/ol><p>That flow carries a lot of risk.<\/p><p>Companies that could benefit from working with you easily fall out of that funnel.\u00a0Most businesses never realize that\u00a0they even have a problem, or that it can be solved. And the ones who do realize that they have a solvable problem don\u2019t know that finding someone like\u00a0<em>you<\/em> is the best way to solve it.<\/p><p>For every one business that makes it to the end of that funnel, dozens or hundreds self-select out of it.<\/p><p>So to plant your flag and say \u201cI\u2019m a web designer, hire me!\u201d you\u2019re excluding a large part of the available market from ever finding you because they\u2026<\/p><ul><li>Don\u2019t know they have a problem that you can help them solve<\/li>\n<li>If they are aware of the problem, don\u2019t know it can be solved<\/li>\n<li>Don\u2019t know that a provider like you can solve their problem<\/li>\n<\/ul><p>Generalization and not niching means that you\u2019re getting client candidates who know they have a solvable problem, know a way of solving it, and are seeking out \u201cvendors\u201d who can solve their problem.<\/p><p>This immediately commoditizes you: it\u2019s now <strong>you vs. the world<\/strong>. You against every other provider who offers a service similar to you.<\/p><p>But say someone does seek you out and you get to the point of furnishing a proposal.<\/p><p>At this point, you\u2019re forced to niche. You take your skills and your experience and apply it to them.<\/p><p>You explain how you\u2019ll help them, and you explain it in a way that\u2019s specific to them.<\/p><p><strong>Every proposal you write is an example of you niching your services.<\/strong> Most of us just don\u2019t think about it that way.<\/p><h2>Why niching is regularly met with resistance<\/h2><p>If every proposal is an exercise in niching, you\u2019re probably thinking that rarely do you write identical, or even similar, proposals.<\/p><p>Right?<\/p><p>Many generalists often fear niching because they think it requires backing themselves into a corner. They feel forced to select a target market and think that if they choose the wrong audience, they\u2019ll fall flat on their face.<\/p><p>They also feel like they\u2019re intentionally sabotaging their own business.<\/p><p>If you choose one path, one type of customer, or one problem that you decide to go all-in on, you\u2019re bypassing opportunity.<\/p><p>If you (again, as a web designer) are just going to serve dog groomers who aren\u2019t getting enough online customers, you\u2019re missing out on a very wide market of people who could benefit from your ability to create websites.<\/p><p>The above are easy-to-make assumptions that often keep people from ever making an effort to niche their business.<\/p><p>They often think the upside is dubious and the risk is great.<\/p><h2>Niching is more about marketing than the product<\/h2><p>When companies build products, especially products that require a lot of research and development overhead, it\u2019s expensive to come up with versions of their products that target each key demographic that could use the product.<\/p><p>What\u2019s comparatively less expensive is marketing differently per each demographic.<\/p><p>Consider the core demographics that buy MacBooks:<\/p><ul><li><strong>Creatives.<\/strong> Historically, Mac OS has been the friendliest operating system for photographers, videographers, and designers and the available software reflects that.<br \/><strong>Certain types of developers.<\/strong> Because Mac OS is built off UNIX, the OS easily allows developers to write software in an environment that\u2019ll be similar to their production environments.<br \/><strong>Students.<\/strong> Apple has always done a good job at getting their equipment on college campuses and in classrooms.<br \/><strong>Brand snobs.<\/strong> Let\u2019s face it: Apple, as a brand, is a lot sexier than most other hardware manufacturers, and many of us still feel an air of superiority when we\u2019re sitting next to someone sporting a clunky Dell laptop.<\/li>\n<\/ul><p>Each of these groups has different reasons for buying. And Apple does a good job at tailoring the way they <em>communicate<\/em> their products to each of these different groups.<\/p><p>WWDC, the annual Apple conference targeting developers, touts operating system improvements and hardware specs that tell developers that they\u2019ll be using best-in-class equipment.<\/p><p>The famous keynotes that the company gives tout their software, ease-of-use, and 3rd party ecosystem to appeal to creatives and brand snobs.<\/p><p>There\u2019s also distinct marketing in place for students, educators, and other key demographics.<\/p><p>But all of these roads lead back to the same product, the MacBook.<\/p><p><strong>If you recognize that positioning is more about how you communicate what you offer rather than a reflection of what you do, it\u2019s a much easier pill to swallow.<\/strong><\/p><h2>How I niche my own marketing<\/h2><p>My best-selling product is <a href=\"https:\/\/doubleyourfreelancing.com\/rate\/\">Double Your Freelancing Rate<\/a>. Over 8,000 people have bought it.<\/p><p>It\u2019s a self-study course on how to price and pitch your services. It helps any freelancer or agency who sells their services to other businesses.<\/p><p>However, I\u2019m used to hearing the following objections:<\/p><ul><li>\u201cI\u2019m a writer, and most of the case studies listed are from web developers and designers. This can\u2019t work for me.\u201d<\/li>\n<li>\u201cI own an agency. What works for freelancers can\u2019t help us.\u201d<\/li>\n<li>Some permutation of: \u201cI\u2019m X, therefore Y.\u201d<\/li>\n<\/ul><p>By marketing a general course on pricing and proposal creation with general sales and marketing copy, <strong>I\u2019m leaving it up to the prospective customer to figure out if it\u2019s right for them.<\/strong><\/p><p>And any time you\u2019re leaving it up to someone to determine if hiring or buying from you will help them, you\u2019re <em>guaranteed<\/em> to be losing out on sales.<\/p><p>I could niche the product by creating a bunch of new courses: \u201cDouble Your Freelance Design Rate\u201d, \u201cDouble Your Freelance Web Development Rate, \u201cDouble Your Agency Design Rate\u201d, and so on.<\/p><p>Being that specific is actually a really good way to create your first product\u00a0because you can laser focus both your <strong>marketing<\/strong> and your <strong>product<\/strong> to a specific need that someone has.<\/p><p>But it wouldn\u2019t work for me\u00a0because my course is designed to help so many different types of freelancers and agencies.<\/p><p>So what did I do?<\/p><p>I simply changed the sales page and the communication that leads people to that sales page depending on who someone is.<\/p><p>If I know you\u2019re a designer and I know that you run an agency, the headline, body copy, and soon the testimonials will change to speak directly to people like you.<\/p><p><img class=\"aligncenter size-large wp-image-30207\" src=\"https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.09-PM-1024x439.jpg\" alt=\"DYFRate Headline For Designers\" width=\"1024\" height=\"439\" srcset=\"https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.09-PM-1024x439.jpg 1024w, https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.09-PM-600x257.jpg 600w, https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.09-PM-300x129.jpg 300w, https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.09-PM.jpg 1106w\" \/><\/p><p>And if you\u2019re going through my email course, <a href=\"https:\/\/doubleyourfreelancing.com\/free-pricing-course\/\"><em>Charge What You\u2019re Worth<\/em><\/a>, and tell me that you\u2019re struggling to close more proposals (as opposed to figuring out what to charge or upping your prices), I\u2019m going to <em>position<\/em> how I sell the course to laser focus on how it helps with proposal writing.<\/p><p><img class=\"aligncenter size-large wp-image-30208\" src=\"https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.11-PM-1024x277.jpg\" alt=\"Charge What You're Worth on the fly niching example\" width=\"1024\" height=\"277\" srcset=\"https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.11-PM-1024x277.jpg 1024w, https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.11-PM-600x162.jpg 600w, https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.11-PM-300x81.jpg 300w, https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.11-PM.jpg 1032w\" \/><\/p><p>The course stays the same, but the way I describe the course changes.<\/p><p>That\u2019s all you need to do when it comes to positioning your own business. You don\u2019t need to necessarily go all-in on one type of customer, <strong>but you do need to start with one.<\/strong><\/p><p>Instead of thinking \u201cthis is the only type of customer I\u2019ll work with,\u201d think: \u201cHow can I reach people who fit this demographic and likely have this problem? And then how can I speak directly to them?\u201d<\/p><p>And then setup a funnel for that group. Develop an acquisition strategy (online or offline), put in place an automated nurturing sequence or website\/subsite that speaks directly to them, and create a compelling pitch that explains how you can help them.<\/p><p>This doesn\u2019t need to replace everything you do. Rather, this can be just a branch. And over time, you can create more branches that all lead to you (or cull your branches to better focus in on the channels that are working well.)<\/p><p>Can you go all-in and exclusively work with one type of customer?<\/p><p>Sure. You\u2019ll definitely want to start considering doing this as you begin developing more repeatable, more productized services that allow you to focus your attention on really owning that space.<\/p><p>But it\u2019s by no means a requirement. And because people think that they need to give up their existing portfolio of clients in order to niche correctly, they often flounder.<\/p><h2>Niching is a framework, not a phase<\/h2><p>As you consider \u201cchoosing a niche,\u201d think less about squeezing into some arbitrary amount of time the activities that go into niching, and more about a framework that you can establish that lets you:<\/p><ul><li>Identify unfair advantages that you have\u2014combined expertise (like knowing how to write sales copy and write code), experience (e.g. working with a few clients in the natural gas space), and other interests you might have.<\/li>\n<li>Extrapolate how you can apply this expertise and experience to actual business problems that a demographic has. The single best way to do this is to talk to people. Keep your ear to the floor and listen to what people are saying, and use <a href=\"https:\/\/doubleyourfreelancing.com\/socratically-question-new-project-leads\/\">Socratic Questioning<\/a> to determine how your skill set can satisfy their needs.<\/li>\n<li>Normalize what you learn doing this and put together a marketing funnel, which is a fancy way of saying a set of processes, both internal and external (a sales page, email course, Facebook ad, etc.), that lead someone from first contact to the end goal of becoming a client.<\/li>\n<li>Make a habit of looking at how this funnel is performing, noting what clients and prospects are telling you, and ruthlessly test and optimize this funnel.<\/li>\n<\/ul><p><strong>The funnel is your niche.<\/strong> It\u2019s taking what happens at the time of writing a proposal and putting it in front of someone\u2014possibly even before they\u2019ve realized they need someone like you.<\/p><p>The best business owners are those who know how to capture, analyze, and act on data.<\/p><p>They aren\u2019t looking for a magic pill.<\/p><p>They aren\u2019t waiting for someone to tell them what they should niche in, what they should say in their marketing, and so on.<\/p><p>Instead, they listen.<\/p><p>They determine what people are saying (what \u201cthe market needs\u201d, to channel Adam Smith.)<\/p><p>They look at what they\u2019re able to offer\u00a0and come up with a way to advertise that to people who are the type who could benefit from what they have to offer\u00a0but aren\u2019t necessarily seeking out someone who can give them what they need.<\/p><p>They develop a conveyor belt\u2026 a marketing funnel, whose sole purpose is to get people to be aware of the problem they have, realize it can be solved, teach them how they could solve it, and then present a way of solving it.<\/p><p>Unlike freelancers who haven\u2019t gone beyond \u201cI\u2019m an X for hire,\u201d they\u2019ve realized that they\u2019re missing out on a lot of opportunities by only helping people who have already figured out how to solve the problems they have.<\/p><h2>Closing thoughts<\/h2><p>Remember the typical lifecycle of a client that I mapped out earlier in this article?<\/p><ol><li>A business becomes aware of a problem that they have<\/li>\n<li>They realize that this problem can be solved<\/li>\n<li>They then determine someone like you can solve the problem<\/li>\n<li>They seek out people like you<\/li>\n<\/ol><p>When you\u2019ve created a niched funnel (or a series of niched funnels), the lifecycle starts to look like this:<\/p><ol><li>You make first contact with a business who has a problem (via an advertisement, blog article, a talk you give at an event, etc.)<\/li>\n<li>You offer to teach them more about how their problem can be solved, and why it\u2019s important that they solve it.<\/li>\n<li>You then present a service that exactly solves their problem.<\/li>\n<\/ol><p>Instead of waiting for someone to make it through the gauntlet of doubts and hesitations that come when walking the path of identifying a problem, realizing it\u2019s solvability, and learning about how it can be solved, you should be casting a wider net and looking for the type of businesses who fit the description of who your niched marketing funnel serves.<\/p><p>Because if you think telling the world that you\u2019re a web designer is going to help, you\u2019re right\u2026 but you\u2019re only going to help companies who already know they need a web designer.<\/p><p>But if you decide to share with the world that you specialize in helping online stores increase sales by creating human-centered designs, you\u2019re now speaking to anyone who runs an online store. (And you can then educate them about how design, done right, can help increase sales.)<\/p><p>That\u2019s all niching is.<\/p><p>It doesn\u2019t require you to cut off the way you\u2019re getting work now. <strong>Niching adds, it never subtracts.<\/strong><\/p><p>Many freelancers, however, end up becoming super successful in a particular niche and then decide to <em>only<\/em> work with those types of clients. That\u2019s admirable, smart, but it\u2019s their prerogative. It\u2019s by no means a requirement.<\/p><p>Take a few minutes now to ask yourself:<\/p><ul><li>\u201cWhat unfair advantages do I have?\u201d<\/li>\n<li>\u201cWhat kind of companies can benefit from what I\u2019m able to offer?\u201d<\/li>\n<li>\u201cWhat frame of mind are they typically in before they go down the path that ends with working with someone like me?\u201d<\/li>\n<li>\u201cHow can I reliably reach these people and help them down that path?\u201d<\/li>\n<\/ul><p>Much of what goes into automating your funnel is beyond the scope of this article\u00a0but is the subject of what I\u2019ll be writing about over the next few weeks. (Topics like building an audience, educating them at scale, leading them to a service offering, etc.)<\/p><p>However, niching can start with you meeting someone who suffers from X, emailing them with a plan of how you could help solve X, learning more about what X means for their business, and then slowly building up a portfolio of material, like a sales page, that you can leverage the next time you meet someone suffering from X.<\/p><p>So if you\u2019ve ever resisted or feared niching, you should now know you have <strong>zero<\/strong> to worry about.<\/p><p>Remember: niching doesn\u2019t mean that you need to change or eliminate anything that you\u2019re doing now. It\u2019s not a restriction. It\u2019s not a limitation. It\u2019s simply a way of speaking to a group of people who all share a problem, and leading them (in your case) to you.<\/p>","created_at":"2018-11-07T16:49:23+0000","updated_at":"2018-11-07T16:49:23+0000","published_at":"2017-03-14T20:15:43+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":12,"domain_name":"doubleyourfreelancing.com","preview_picture":"https:\/\/doubleyourfreelancing.com\/wp-content\/uploads\/2017-03-14-at-4.09-PM-1024x439.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":462,"title":"How To Make Homemade Bitters | Kitchn","url":"https:\/\/www.thekitchn.com\/how-to-make-homemade-bitters-cooking-lessons-from-the-kitchn-197883","content":"<section class=\"ContentBundle ContentBundle--post\"><div class=\"SitePost\"><div><div class=\"PostBundle PostBundle--with-rail\"><article class=\"Post\"><p>by Emily HanUpdated Sep 6, 2018<\/p><p>We support our readers with carefully chosen product recommendations to improve life at home. You support us through our independently chosen links, many of which earn us a commission.<\/p><div class=\"Post__item Post__image\"><figure class=\"Figure\"><div class=\"Figure__image\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,e_pixelate:100,w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cfd880c4ef9c97db4daa12161bccf05052df2a8c.jpeg\" alt=\"(placeholder)\" class=\"Figure__image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">(Image credit: <a href=\"https:\/\/www.thekitchn.com\/authors\/emilyhan\">Emily Han<\/a>)<\/figcaption><\/figure><\/div><div class=\"Post__item Post__html\"><div><p> From the medicine cabinet to the bar, bitters have a long history of curing ailments and flavoring drinks. Though they may seem mysterious, at heart, bitters are simply bitter and aromatic herbs and spices infused or tinctured in spirits. Combining the various flavors is where the fun part comes in. Here\u2019s a primer to making your own bitters at home, which you can use in <a href=\"https:\/\/www.thekitchn.com\/cocktail-basics-all-about-bitt-64948\">cocktails<\/a>, <a href=\"https:\/\/www.thekitchn.com\/soda-with-bitters-simple-refre-125082\">sodas<\/a>, cooking, and even as <a href=\"https:\/\/www.thekitchn.com\/urban-moonshine-188101\">digestive aids<\/a>. <\/p><\/div><\/div><div class=\"Post__item Post__image\"><figure class=\"Figure\"><div class=\"Figure__image\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,e_pixelate:100,w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/f8b87f876cc336148bb729ad9e6384b1267f1fab.jpeg\" alt=\"(placeholder)\" class=\"Figure__image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">(Image credit: <a href=\"https:\/\/www.thekitchn.com\/authors\/emilyhan\">Emily Han<\/a>)<\/figcaption><\/figure><\/div><div class=\"Post__item Post__html\"><p><\/p><h3>What Botanicals to Use?<\/h3><\/div><div class=\"Post__item Post__html\"><div><p> Bitters are comprised of bitter-tasting roots, barks, or leaves, plus other botanicals that provide aroma and flavor (and medicinal properties). In general you should use whole ingredients rather than ground, as they are easier to strain out. You can chop ingredients up or coarsely crack them to expose more surface area for infusing. <\/p><\/div><\/div><div class=\"Post__item Post__html\"><div><p> <strong>Bittering agents<\/strong> usually make up 10 to 50% of the blend and may include plants like angelica root, artichoke leaf, barberry root, black walnut leaf, burdock root, calamus root, cinchona bark, citrus peel, dandelion root and leaf, devil\u2019s club root, gentian root, horehound, licorice root, mugwort, Oregon grape root, orris root, quassia bark, sarsaparilla, wild cherry bark, and wormwood. <\/p><\/div><\/div><div class=\"Post__item Post__html\"><div><p> <strong>Aromatic and flavor agents<\/strong> round out the bitters and may include just about any herb, spice, flower, fruit, or nut. Use your imagination! Also use organic ingredients when possible, especially when it comes to fruit peels. Some examples: <\/p><\/div><\/div><div class=\"Post__item Post__html\"><div><ul><li><strong>Spices<\/strong> \u2013 allspice, aniseed, caraway, cardamom, cassia, celery seed, chiles, cinnamon, cloves, coriander, fennel, ginger, juniper berries, nutmeg, peppercorns, star anise, vanilla beans<\/li> \n <li><strong>Herbs &amp; Flowers<\/strong> \u2013 chamomile, hibiscus, hops, lavender, lemongrass, mint, rose, rosemary, sage, thyme, yarrow<\/li> \n <li><strong>Fruits<\/strong> \u2013 fresh or dried citrus peel (lemon, lime, orange, grapefruit), dried fruit (apples, cherries, figs, raisins)<\/li> \n <li><strong>Nuts<\/strong> \u2013 toasted almonds, pecans, walnuts, etc.<\/li> \n <li><strong>Beans<\/strong> \u2013 cacao beans, cocoa nibs, coffee beans<\/li> \n<\/ul><\/div><\/div><div class=\"Post__item Post__html\"><div><p> To learn more about botanicals, I highly recommend the book <a href=\"https:\/\/www.thekitchn.com\/the-drunken-botanist-by-amy-stewart-new-cookbook-184184\"><em>The Drunken Botanist<\/em><\/a> by Amy Stewart. <\/p><\/div><\/div><div class=\"Post__item Post__image\"><figure class=\"Figure\"><div class=\"Figure__image\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,e_pixelate:100,w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/638daed5c52cc93fa34d5e6cef8338be5db07861.jpeg\" alt=\"(placeholder)\" class=\"Figure__image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">(Image credit: <a href=\"https:\/\/www.thekitchn.com\/authors\/emilyhan\">Emily Han<\/a>)<\/figcaption><\/figure><\/div><div class=\"Post__item Post__html\"><p><\/p><h3>What Alcohol to Use?<\/h3><\/div><div class=\"Post__item Post__html\"><div><p> For maximum flavor extraction and preservation, use a <strong>high-proof liquor \u2014 at least 100 proof or 50% alcohol by volume (ABV)<\/strong>. You\u2019ll get the most neutral flavor using grain alcohol (such as Everclear) or vodka (Absolut and Smirnoff are fairly accessible brands of 100-proof). You can also experiment with other spirits such as 101-proof bourbon and rye and 151-proof rum. <\/p><\/div><\/div><div class=\"Post__item Post__html\"><p><\/p><h3>Additional Ingredients<\/h3><\/div><div class=\"Post__item Post__html\"><div><p> In addition, bitters may be lightly sweetened with simple syrup, caramel, molasses, honey, or other sweeteners. They may also be diluted with distilled water, bringing the final product no lower than 80 to 90 proof or 40 to 45% ABV. <\/p><\/div><\/div><div class=\"Post__item Post__html\"><p><\/p><h3>Which Infusing Method? <\/h3><\/div><div class=\"Post__item Post__html\"><div><p> There are two main ways to make bitters. One method is to combine all of your botanicals and infuse them in liquor together. The other method (and the one shown here) is to make a separate infusion or tincture of each botanical and then blend them to taste. I prefer this method because different ingredients infuse at different rates. Tincturing them separately gives you more control over the outcome. However, if you have a good recipe, the first method might work just fine. <\/p><\/div><\/div><div class=\"Post__item Post__html\"><p><\/p><h3>How Long to Infuse? <\/h3><\/div><div class=\"Post__item Post__html\"><div><p> Depending on the botanical, infusing time may range from a day to several weeks. Regularly smell and sample each tincture or infusion; it will be ready when it strongly conveys the ingredient. To smell, put a couple drops of the infusion in your palms, rub them together, and hold your hands up to your nose. To taste, put a couple drops in a glass of still or sparkling water. If you taste it straight bear in mind it will be rather intense! <\/p><\/div><\/div><div class=\"Post__item Post__html\"><p><\/p><h3>Sources for Ingredients and Supplies<\/h3><\/div><div class=\"Post__item Post__html\"><div><p> Bitters ingredients may be found at grocery stores, spice shops, medicinal herb shops, in gardens, growing wild, and online. <\/p><\/div><\/div><div class=\"Post__item Post__html\"><div><ul><li><strong>Botanicals from <a href=\"https:\/\/www.mountainroseherbs.com\/\">Mountain Rose Herbs<\/a><\/strong><\/li> \n  <li><strong>Bottles from <a href=\"http:\/\/www.specialtybottle.com\/\">Specialty Bottle<\/a><\/strong><\/li> \n  <li><strong>Martha Stewart Crafts Labels from Michaels<\/strong><\/li> \n <\/ul><\/div><\/div><div class=\"Post__item Post__gallery\"><div class=\"Gallery Gallery--collapsed\"><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-0 Gallery__slide--selected\" role=\"button\" tabindex=\"0\" data-image-slide=\"0\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"1\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cbbd353ce4fa1a7ce200d379eef907081d0798ac.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cbbd353ce4fa1a7ce200d379eef907081d0798ac.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cbbd353ce4fa1a7ce200d379eef907081d0798ac.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cbbd353ce4fa1a7ce200d379eef907081d0798ac.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cbbd353ce4fa1a7ce200d379eef907081d0798ac.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cbbd353ce4fa1a7ce200d379eef907081d0798ac.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Place botanicals in separate jars<\/figcaption><p>1 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-1\" role=\"button\" tabindex=\"0\" data-image-slide=\"1\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"2\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/21ca360f1b30648ba81c746baed27677ccb84627.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/21ca360f1b30648ba81c746baed27677ccb84627.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/21ca360f1b30648ba81c746baed27677ccb84627.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/21ca360f1b30648ba81c746baed27677ccb84627.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/21ca360f1b30648ba81c746baed27677ccb84627.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/21ca360f1b30648ba81c746baed27677ccb84627.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Cover with liquor<\/figcaption><p>2 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-2\" role=\"button\" tabindex=\"0\" data-image-slide=\"2\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"3\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/64ea179d071e2c054bf5774d85514d67e5ddc9b0.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/64ea179d071e2c054bf5774d85514d67e5ddc9b0.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/64ea179d071e2c054bf5774d85514d67e5ddc9b0.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/64ea179d071e2c054bf5774d85514d67e5ddc9b0.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/64ea179d071e2c054bf5774d85514d67e5ddc9b0.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/64ea179d071e2c054bf5774d85514d67e5ddc9b0.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Label the jars<\/figcaption><p>3 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-3\" role=\"button\" tabindex=\"0\" data-image-slide=\"3\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"4\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/fb7461fb3665a2e3d36efd4a3ca3bcff04d4e037.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/fb7461fb3665a2e3d36efd4a3ca3bcff04d4e037.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/fb7461fb3665a2e3d36efd4a3ca3bcff04d4e037.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/fb7461fb3665a2e3d36efd4a3ca3bcff04d4e037.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/fb7461fb3665a2e3d36efd4a3ca3bcff04d4e037.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/fb7461fb3665a2e3d36efd4a3ca3bcff04d4e037.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Shake the jars once a day<\/figcaption><p>4 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-4\" role=\"button\" tabindex=\"0\" data-image-slide=\"4\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"6\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/1ed637236cc8d93ec9b8ccbe47c05431cd513ea1.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/1ed637236cc8d93ec9b8ccbe47c05431cd513ea1.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/1ed637236cc8d93ec9b8ccbe47c05431cd513ea1.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/1ed637236cc8d93ec9b8ccbe47c05431cd513ea1.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/1ed637236cc8d93ec9b8ccbe47c05431cd513ea1.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/1ed637236cc8d93ec9b8ccbe47c05431cd513ea1.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Wait for botanicals to infuse<\/figcaption><p>5 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-5\" role=\"button\" tabindex=\"0\" data-image-slide=\"5\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"7\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/db0f4838c5e1107e8f6567d2effcb5258c04fbb3.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/db0f4838c5e1107e8f6567d2effcb5258c04fbb3.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/db0f4838c5e1107e8f6567d2effcb5258c04fbb3.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/db0f4838c5e1107e8f6567d2effcb5258c04fbb3.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/db0f4838c5e1107e8f6567d2effcb5258c04fbb3.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/db0f4838c5e1107e8f6567d2effcb5258c04fbb3.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Strain<\/figcaption><p>6 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-6\" role=\"button\" tabindex=\"0\" data-image-slide=\"6\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"8\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/41a88d23673db96a31bb17633f79649c80be4f4f.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/41a88d23673db96a31bb17633f79649c80be4f4f.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/41a88d23673db96a31bb17633f79649c80be4f4f.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/41a88d23673db96a31bb17633f79649c80be4f4f.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/41a88d23673db96a31bb17633f79649c80be4f4f.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/41a88d23673db96a31bb17633f79649c80be4f4f.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Blend<\/figcaption><p>7 of 8<\/p><\/div><div class=\"Gallery__slide Gallery__slide--image Gallery__slide--image-7\" role=\"button\" tabindex=\"0\" data-image-slide=\"7\"><div class=\"Gallery__slide-image\" data-gtm-gallery-type=\"1\" data-gtm-slide-type=\"content\" data-gtm-slide-idx=\"9\"><img src=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png,c_fill,g_auto,e_pixelate:100\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/74542d7911cb42ea8ec948f31744910b5c132c55.jpeg\" alt=\"\" srcset=\"https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_730\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/74542d7911cb42ea8ec948f31744910b5c132c55.jpeg 730w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_630\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/74542d7911cb42ea8ec948f31744910b5c132c55.jpeg 630w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_530\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/74542d7911cb42ea8ec948f31744910b5c132c55.jpeg 530w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_430\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/74542d7911cb42ea8ec948f31744910b5c132c55.jpeg 430w, https:\/\/cdn.apartmenttherapy.info\/image\/fetch\/f_png, c_fill, g_auto, denied:e_pixelate:100, w_330\/https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/74542d7911cb42ea8ec948f31744910b5c132c55.jpeg 330w\" class=\"Gallery__slide-image--placeholder\" \/><\/div><figcaption class=\"ImageCaption\">Bottle<\/figcaption><p>8 of 8<\/p><\/div><\/div><\/div><div class=\"Post__item Post__recipe\"><div class=\"SiteRecipe\"><div class=\"SiteRecipe__interactions\"><div class=\"Checkbox\"><div class=\"Checkbox__input\" data-gtm-checkbox-action=\"make\"><label class=\"Checkbox__label\" for=\"i-made-this-11315\" id=\"i-made-this-11315--label\">I made this!<\/label><\/div><\/div><div class=\"RecipeRating\">12Ratings<\/div><\/div><div class=\"Recipe\" itemtype=\"http:\/\/schema.org\/Recipe\" itemscope=\"itemscope\"><h3>Ingredients<\/h3><div><div class=\"Recipe__ingredient-group\"><ul class=\"Recipe__ingredients\"><li class=\"Recipe__ingredient\" itemprop=\"recipeIngredient\"><p>Bittering agents<\/p><\/li><li class=\"Recipe__ingredient\" itemprop=\"recipeIngredient\"><p>Aromatic\/Flavor agents<\/p><\/li><li class=\"Recipe__ingredient\" itemprop=\"recipeIngredient\"><p>High-proof liquor (at least 100 proof or higher)<\/p><\/li><li class=\"Recipe__ingredient\" itemprop=\"recipeIngredient\"><p>Distilled water for diluting (optional)<\/p><\/li><li class=\"Recipe__ingredient\" itemprop=\"recipeIngredient\"><p>Sweetener (optional)<\/p><\/li><\/ul><\/div><\/div><div><div class=\"Recipe__supplies-group\"><h4 class=\"Recipe__supplies-group-name\">Equipment<\/h4><ul class=\"Recipe__supplies\"><li class=\"Recipe__supply-item\"><p>Cutting board, knife, vegetable peeler, and\/or zester for cutting and peeling fruit<\/p><\/li><li class=\"Recipe__supply-item\"><p>Knife and\/or mortar and pestle for cracking spices<\/p><\/li><li class=\"Recipe__supply-item\"><p>Measuring tools (cups, spoons, scale)<\/p><\/li><li class=\"Recipe__supply-item\"><p>Clean mason jars or other containers for making tinctures<\/p><\/li><li class=\"Recipe__supply-item\"><p>Labels for tincture jars (a Sharpie and masking tape or painter's tape works well)<\/p><\/li><li class=\"Recipe__supply-item\"><p>Fine-mesh strainer, cheesecloth, and\/or coffee filters for straining<\/p><\/li><li class=\"Recipe__supply-item\"><p>Dropper, pipette, or syringe for blending<\/p><\/li><li class=\"Recipe__supply-item\"><p>Small funnel<\/p><\/li><li class=\"Recipe__supply-item\"><p>Clean bottles for finished bitters<\/p><\/li><\/ul><\/div><\/div><h3>Instructions<\/h3><div><div class=\"Recipe__instruction-group Recipe__instruction-group--step\"><ol class=\"Recipe__instructions\"><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Place botanicals in separate jars:<\/strong> Place each botanical (bittering agents or aromatic\/flavor agent) in a separate jar. You may wish to chop up or crack ingredients to expose more surface area for better and faster infusion.<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Cover with liquor:<\/strong> Add liquor to each jar, making sure the botanicals are completely submerged. Cover the jar tightly.<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Label the jars:<\/strong> Don't forget to label the jar with the contents and date. You may also wish to write down the measurements you used, either on the label or in separate notes.<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Shake the jars once a day:<\/strong> Give each jar a good shake, and continue to shake them once a day.<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Wait for botanicals to infuse:<\/strong> Depending on the botanical, infusing time may range from a day to several weeks. Regularly smell and sample each tincture; it will be ready when it strongly conveys the ingredient. To smell, put a couple drops of the infusion in your palms, rub them together, and hold your hands up to your nose. To taste, put a couple drops in a glass of still or sparkling water \u2014 or taste it straight but bear in mind it will be rather intense!<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Strain:<\/strong> When each tincture is ready, strain out the solids. For finer straining, use a coffee filter.<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Blend:<\/strong> Now comes the creative part. Using a dropper, pipette, or syringe, start blending the different tinctures together in a small glass or clean bitters bottle. You might use anywhere from a couple drops to a full ounce or two of each tincture. Optionally, you can dilute it with distilled water and\/or lightly sweeten it. (Remember to take notes so you can reproduce your bitters recipe if it turns out well!) Try your blend using the smelling and tasting tips in Step 5. Give it a few days or weeks for the flavors to really marry.<\/p><\/li><li class=\"Recipe__instruction-step\" itemprop=\"recipeInstructions\"><p><strong>Bottle:<\/strong> If you didn't blend your bitters right in the bottle, transfer the mixture to a clean container. It can last for years!<\/p><\/li><\/ol><\/div><\/div><div><h3>Recipe Notes<\/h3><div class=\"Recipe__notes\"><h3>Simple Bitters Blends to Try<\/h3><p>I encourage you to experiment with your own bitters blends, but if you want to get started with something fairly simple, here are some blends that use mostly common ingredients and six tinctures or less. Taste your blends as you go along and adjust depending on the strength and nature of your particular infusions.<\/p><ul><li><strong>Orange Bitters:<\/strong> 12 parts orange peel, 2 parts gentian, 2 parts cardamom, 2 parts coriander, 1 part allspice, 1 part cloves<\/li><li><strong>Lavender Bitters:<\/strong> 20 parts lavender, 6 parts orange, 2 parts vanilla, 1 part ginger<\/li><li><strong>Coffee Bitters:<\/strong> 10 parts coffee bean, 3 parts cocoa nib, 2 parts wormwood, 1 part orange, 1 part cinnamon; sweeten to taste with molasses<\/li><\/ul><p><strong>Quantities and measurements:<\/strong> If you're new to making bitters, I recommend starting with 1 to 2 teaspoons of dried botanicals per 4 ounces of liquor (the size of a small mason jar). You can make as many separate tinctures as you like, and once you get the bug you'll probably want to experiment with many. In the beginning I suggest starting with at least 6 to 12, including 1 or more bittering agents.<\/p><p>If you want to get more specific, you can make your tinctures using the classic herbalists' weight to volume ratio of 1 part dried botanical to 5 parts liquor, or 1 part fresh botanical to 2 parts liquor.<\/p><\/div><\/div><\/div><\/div><\/div><p>Originally published Dec 5, 2013<\/p><div class=\"Post__recirc\"><p>Loading...<\/p><\/div><\/article><\/div><\/div><p>Loading...<\/p><\/div><\/section><section class=\"ContentBundle ContentBundle--spread-expandedGrid\"><section class=\"ContentBundle ContentBundle--spread-mini\"><\/section><\/section>","created_at":"2018-11-07T16:48:56+0000","updated_at":"2018-11-07T16:48:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"www.thekitchn.com","preview_picture":"https:\/\/storage.googleapis.com\/gen-atmedia\/3\/2013\/12\/cfd880c4ef9c97db4daa12161bccf05052df2a8c.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":463,"title":"Little Things I Like to Do with Git \u2013 CSS Wizardry \u2013 CSS Architecture, Web Performance Optimisation, and more, by Harry Roberts","url":"https:\/\/csswizardry.com\/2017\/05\/little-things-i-like-to-do-with-git\/","content":"<p>I was chatting with <a href=\"https:\/\/twitter.com\/tkadlec\">Tim<\/a> just the other day about\nhow much I love Git\u2014it\u2019s such a powerful, elegant tool and it\u2019s one I use the\nmost often. I thought I would note down some useful little Git snippets that\nI use the most frequently.<\/p><h2 id=\"leaderboards\">Leaderboards<\/h2><p>Whether you think gamification or competitiveness in your work is a good or\na bad thing is a different discussion for a different time, but if you\u2019re\ninterested in seeing who has committed how much to your project, <code class=\"highlighter-rouge\">shortlog<\/code> can\ngive us the answer:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git shortlog -sn --all --no-merges\n    80  Harry Roberts\n    34  Samantha Peters\n     3  Tom Smith\n<\/pre><\/div><\/div><p>The <code class=\"highlighter-rouge\">shortlog<\/code> gives us summaries of <code class=\"highlighter-rouge\">git log<\/code>; the <code class=\"highlighter-rouge\">-s<\/code> flag will <q>suppress\ncommit description and provide a commit count summary only<\/q>, and the <code class=\"highlighter-rouge\">-n<\/code>\nflag will <q>sort output according to the number of commits per author instead\nof author alphabetic order.<\/q> The <code class=\"highlighter-rouge\">--all<\/code> flag logs all branches, and\n<code class=\"highlighter-rouge\">--no-merges<\/code> ensures that merge commits aren\u2019t being counted.<\/p><p>The above shows all commits for the lifetime of the project, but if you want to\nsee how much people have done during a specified timeframe, you can use the\n<code class=\"highlighter-rouge\">--since<\/code> and <code class=\"highlighter-rouge\">--until<\/code> flags:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git shortlog -sn --since='10 weeks' --until='2 weeks'\n    59  Harry Roberts\n    24  Samantha Peters\n<\/pre><\/div><\/div><p>I have this aliased to <code class=\"highlighter-rouge\">$ git stats<\/code>.<\/p><h2 id=\"praise-people\">Praise People<\/h2><p>Git has a very useful <code class=\"highlighter-rouge\">blame<\/code> feature which allows us to check which developer\nwas responsible for changing a particular piece of code:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre># See who last changed lines 5 through 10 of the buttons\u2019 CSS:\n$ git blame -L5,10 _components.buttons.scss\n<\/pre><\/div><\/div><p>The issue here is the wording. It\u2019s pretty loaded and implies that the developer\nwe\u2019re looking for did something wrong, and this might not always be the\ncase\u2014they may have done something particularly clever or impressive and we might\nwant to find out who we should ask about it (<q>Whoa! I haven\u2019t seen this\nfeature before\u2014I wonder who did that?!<\/q>)<\/p><p>Taking the lead from SVN, I alias <code class=\"highlighter-rouge\">praise<\/code> onto <code class=\"highlighter-rouge\">blame<\/code>, so I can use either\ndepending on my context:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git config --global alias.praise blame\n<\/pre><\/div><\/div><p>This now means I can do this:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre># Find out who implemented Resource Hints and buy them a coffee:\n$ git praise -L18,23 _includes\/head.html\n<\/pre><\/div><\/div><p>Small change, but it\u2019s pretty nice.<\/p><p>When you <code class=\"highlighter-rouge\">diff<\/code> or <code class=\"highlighter-rouge\">show<\/code> an object with a lot of whitespace changes, we\u2019re left\nlooking at a lot of visual noise which can make it hard to see anything more\nimportant.<\/p><p>Thankfully, removing this noise is pretty trivial by using the <code class=\"highlighter-rouge\">-w<\/code> flag which\ncan be used alongside <code class=\"highlighter-rouge\">git diff<\/code> and <code class=\"highlighter-rouge\">git show<\/code>. For example, before:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre> a {\n   color: $color-links;\n-&amp;:hover {\n-  color: $color-links-hover;\n-}\n+  &amp;:hover {\n+    color: $color-links-hover;\n+    text-decoration: underline;\n+  }\n}\n<\/pre><\/div><\/div><p>\u2026and after:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre> a {\n   color: $color-links;\n   &amp;:hover {\n     color: $color-links-hover;\n+    text-decoration: underline;\n   }\n}\n<\/pre><\/div><\/div><p>Now it\u2019s easy to see that the only <em>meaningful<\/em> change was the addition of\n<code class=\"highlighter-rouge\">text-decoration: underline;<\/code>, and the rest of the diff was somewhat misleading\nnoise.<\/p><p>When editing prose, as opposed to code, it can often be much more useful to see\nchanged words rather than whole changed lines; this is particularly helpful when\nwriting markdown, like I am right now.<\/p><p>Thankfully, we can show only changes words by using the <code class=\"highlighter-rouge\">--word-diff<\/code> flag:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git diff --word-diff\n<\/pre><\/div><\/div><p>Running a diff without the <code class=\"highlighter-rouge\">--word-diff<\/code> flag shows quite a large difference:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>-My friend Tom recently gave an excellent talk\n+My good friend Tom gave an excellent talk\n<\/pre><\/div><\/div><p>\u2026but rerunning the diff with <code class=\"highlighter-rouge\">--word-diff<\/code> enabled gives us a much more\ndigestible and helpful overview:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>My {+good+} friend Tom [-recently-] gave an excellent talk\n<\/pre><\/div><\/div><p>Note how only the changed text is highlighted (within <code class=\"highlighter-rouge\">{+ +}<\/code> and <code class=\"highlighter-rouge\">[- -]<\/code>).<\/p><h2 id=\"see-which-branches-you-recently-worked-on\">See Which Branches You Recently Worked On<\/h2><p>It\u2019s not uncommon for me to jump between lots of different branches on any given\nproject, and keeping track of them can be pretty tricky. We can get Git to help\nus work this out:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git for-each-ref --count=10 --sort=-committerdate refs\/heads\/ --format=\"%(refname:short)\"\n<\/pre><\/div><\/div><p>This will show us the last 10 (<code class=\"highlighter-rouge\">--count=10<\/code>) branches that we worked on, sorted\nby the time that we were last working there. It only shows us local branches\n(<code class=\"highlighter-rouge\">refs\/heads\/<\/code>) and in a much nicer <code class=\"highlighter-rouge\">--format<\/code>.<\/p><p>This is a bit verbose to commit to memory, so I have it aliased to <code class=\"highlighter-rouge\">$ git\nrecent<\/code>.<\/p><h2 id=\"see-what-everyones-been-getting-up-to\">See What Everyone\u2019s Been Getting Up To<\/h2><p>Sometimes\u2014especially for team leads\u2014it\u2019s useful to get a general idea of what\neveryone has been getting up to across all branches. Once again, Git makes that\nreally easy for us:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --all --oneline --no-merges\n<\/pre><\/div><\/div><p>This will give us a simplified log of everybody\u2019s work across <code class=\"highlighter-rouge\">--all<\/code> branches\n(with <code class=\"highlighter-rouge\">--no-merges<\/code>).<\/p><p>We can also limit the number of commits we return by complementing the command\nwith a <code class=\"highlighter-rouge\">--since<\/code>:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --all --since='2 weeks' --oneline --no-merges\n<\/pre><\/div><\/div><p>Now we can see what everyone did in the last two weeks (since the start of the\nsprint, for example).<\/p><p>I have this aliased to <code class=\"highlighter-rouge\">$ git overview<\/code><\/p><h2 id=\"remind-yourself-what-youve-been-up-to\">Remind Yourself What You\u2019ve Been Up To<\/h2><p>You move back onto an old project, or come back to the office after a prolonged\nbreak, and you\u2019re not sure what the last things you worked on are\u2014it happens. We\ncan ask Git to give us a quick recap of our work on a project quite easily:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --all --oneline --no-merges --author=&lt;your email address&gt;\n<\/pre><\/div><\/div><p>This is really, really similar to the above, only we\u2019re limiting the log only to\nour own commits. We can also augment it with <code class=\"highlighter-rouge\">--since<\/code> limits.<\/p><p>I have this aliased to <code class=\"highlighter-rouge\">$ git recap<\/code>.<\/p><h2 id=\"todays-work\">Today\u2019s Work<\/h2><p>Again, I\u2019m not here to discuss how developers\u2019 productivity should be measured,\nbut I frequently find it useful (mandatory, even) to let my clients know what\nI\u2019ve been up to on any given day. Rather than keeping a detailed list of tasks\nthat I\u2019ve completed, I can just ask Git to pull up all of that information for\nme:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --since=00:00:00 --all --no-merges --oneline --author=&lt;your email address&gt;\n<\/pre><\/div><\/div><p>This will <code class=\"highlighter-rouge\">log<\/code> <code class=\"highlighter-rouge\">--all<\/code> branches showing you the work that the specified\n<code class=\"highlighter-rouge\">--author<\/code> has done <code class=\"highlighter-rouge\">--since<\/code> midnight that day (but <code class=\"highlighter-rouge\">--no-merges<\/code>), and present\na simplified <code class=\"highlighter-rouge\">--oneline<\/code> overview.<\/p><p>I have this aliased to <code class=\"highlighter-rouge\">$ git today<\/code>.<\/p><h2 id=\"generate-a-changelog\">Generate a Changelog<\/h2><p>Writing a CHANGELOG can be a little tedious; we have to look over all of the\nwork we\u2019ve done since our last release and then pull out all of the useful bits.\nThankfully we can use Git to give us a head start:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --oneline --no-merges &lt;last tag&gt;..HEAD\n<\/pre><\/div><\/div><p><strong>N.B.<\/strong> <code class=\"highlighter-rouge\">HEAD<\/code> is optional here\u2014if you omit it (i.e. <code class=\"highlighter-rouge\">... --no-merges &lt;last\ntag&gt;..<\/code>) then <code class=\"highlighter-rouge\">HEAD<\/code> is implied. That saves you a couple of keystrokes.<\/p><p>This will create a simplified log showing all commits (excluding merge commits)\nthat took place between your last release and <code class=\"highlighter-rouge\">HEAD<\/code>.<\/p><p>For example:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --oneline --no-merges 1.0.0..\n1257b95 [refs #00019] Bump version\n2b9b28e [refs #00019] Add auto width class\n17b8eb1 [refs #00015] Tidy up README.md\nbbe7d05 [refs #00012] Rename Supercell main mixin\n<\/pre><\/div><\/div><p>This tells me that since my last release (<code class=\"highlighter-rouge\">1.0.0<\/code>) and the current state of my\nproject (<code class=\"highlighter-rouge\">HEAD<\/code>), I\u2019ve done these bits of work. That\u2019s quite a nice basis for\na CHANGELOG.<\/p><p><strong>N.B.<\/strong> You\u2019re not limited to just tags: you can use commit hashes.<\/p><p>If you haven\u2019t worked on a project for a little while, you might want to check\nwhat\u2019s happened upstream before you pull all of those changes down into your\nlocal branch.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --oneline --no-merges HEAD..&lt;remote&gt;\/&lt;branch&gt;\n<\/pre><\/div><\/div><p><strong>N.B.<\/strong> Again, <code class=\"highlighter-rouge\">HEAD<\/code> is optional here, and omitting it will leave it implied.<\/p><p>For example, let\u2019s see what someone has been doing in a particular feature\nbranch whilst you were on holiday:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git checkout feature\/fonts\n$ git fetch\n$ git log --oneline --no-merges ..origin\/feature\/fonts\n<\/pre><\/div><\/div><p>I have this aliased to <code class=\"highlighter-rouge\">$ git upstream<\/code>.<\/p><p>Hopefully you commit and push often, but if\u2014for whatever reason\u2014you find\nyourself with a large amount of local commits that are yet to be pushed, it\u2019s\nprobably wise to quickly review what they all are.<\/p><p>In order to do this, we\u2019re effectively just inverting the previous command:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --oneline --no-merges &lt;remote&gt;\/&lt;branch&gt;..HEAD\n<\/pre><\/div><\/div><p>For example:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git fetch\n$ git log --oneline --no-merges origin\/feature\/fonts..HEAD\n<\/pre><\/div><\/div><p><strong>N.B.<\/strong> Again, <code class=\"highlighter-rouge\">HEAD<\/code> is optional here, and omitting it will leave it implied.<\/p><p>This logs the commits that <code class=\"highlighter-rouge\">&lt;remote&gt;\/&lt;branch&gt;<\/code> needs before it resembles <code class=\"highlighter-rouge\">HEAD<\/code>.<\/p><p>I have this aliased to <code class=\"highlighter-rouge\">$ git local<\/code>.<\/p><h2 id=\"view-complex-logs\">View Complex Logs<\/h2><p>Pretty much every example above uses simplified logs because I just want to get\na quick idea of what\u2019s going on. For more forensic details I use a <code class=\"highlighter-rouge\">--graph<\/code> log\nwith some extra options:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>$ git log --graph --all --decorate --stat --date=iso\n<\/pre><\/div><\/div><p>This will give me <code class=\"highlighter-rouge\">--graph<\/code>-based logs for <code class=\"highlighter-rouge\">--all<\/code> branches showing commit\n<code class=\"highlighter-rouge\">--stat<\/code>s (additions, deletions). <code class=\"highlighter-rouge\">--decorate<\/code> will give me information about\nwhich branch a commit was made on where applicable, and I also get a much\nstricter date format.<\/p><p>I have this aliased to <code class=\"highlighter-rouge\">$ git graph<\/code>.<\/p><h2 id=\"and-you\">And you\u2026?<\/h2><p>I\u2019m sure you\u2019ve all got loads of your own tips and trick: share them on this\npost\u2019s <a href=\"https:\/\/github.com\/csswizardry\/csswizardry.github.com\/issues\/66\">GitHub\nissue<\/a>.<\/p><hr \/><p>\n<a href=\"https:\/\/csswizardry.com\/services\/\" class=\"btn  btn--full\">Did you enjoy this? <strong>Hire me!<\/strong><\/a>\n<\/p>","created_at":"2018-11-07T16:48:56+0000","updated_at":"2018-11-07T16:48:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"csswizardry.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":461,"title":"How to build your own VPN if you\u2019re (rightfully) wary of commercial options","url":"https:\/\/arstechnica.com\/gadgets\/2017\/05\/how-to-build-your-own-vpn-if-youre-rightfully-wary-of-commercial-options\/","content":"<figure class=\"intro-image intro-left\"><img src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/05\/vpn-privacy-800x450.jpg\" alt=\"How to build your own VPN if you\u2019re (rightfully) wary of commercial options\" \/><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/05\/vpn-privacy.jpg\" class=\"enlarge-link\" data-height=\"450\" data-width=\"800\">Enlarge<\/a>Aurich \/ Thinkstock<\/figcaption><\/figure><p>In the wake of this spring's <a href=\"https:\/\/arstechnica.com\/information-technology\/2017\/03\/how-isps-can-sell-your-web-history-and-how-to-stop-them\/\">Senate ruling<\/a> nixing FCC privacy regulations imposed on ISPs, you may be (even more) worried about how your data is used, misused, and abused. There have been a lot of opinions on this topic since, ranging from \"the sky is falling\" to \"<a href=\"https:\/\/cei.org\/blog\/six-reasons-fcc-rules-aren%E2%80%99t-needed-protect-privacy\">move along, citizen, nothing to see here<\/a>.\" The fact is, ISPs tend to be pretty unscrupulous, sometimes even ruthless, about how they gather and use their customers' data. You may not be sure how it's a problem if your ISP gives advertisers more info to serve ads you'd like to see\u2014but what about when your ISP literally <a href=\"https:\/\/arstechnica.com\/tech-policy\/2014\/09\/why-comcasts-javascript-ad-injections-threaten-security-net-neutrality\/\"><em>edits<\/em><\/a> your HTTP traffic, <a href=\"https:\/\/arstechnica.com\/tech-policy\/2014\/09\/meet-the-tech-company-performing-ad-injections-for-big-cable\/\">inserting <em>more<\/em> ads<\/a> and possibly breaking webpages?<\/p>\n<p>With a Congress that has demonstrated its lack of interest in protecting you from your ISP, and ISPs that have repeatedly demonstrated a \"whatever-we-can-get-away-with\" attitude toward customers' data privacy and integrity, it may be time to look into how to get your data out from under your ISP's prying eyes and grubby fingers intact. To do that, you'll need a VPN.<\/p>\n<h2>The scope of the problem (and of the solution)<\/h2>\n<p>Before you can fix this problem, you need to understand it. That means knowing what your ISP can (and cannot) detect (and modify) in your traffic. HTTPS traffic is already relatively secure\u2014or, at least, its <em>content<\/em> is. Your ISP can't actually read the encrypted traffic that goes between you and an HTTPS website (at least, they can't unless they convince you to install a MITM certificate, like <a href=\"https:\/\/arstechnica.com\/security\/2015\/02\/lenovo-pcs-ship-with-man-in-the-middle-adware-that-breaks-https-connections\/\">Lenovo did to unsuspecting users of its consumer laptops<\/a> in 2015). However, ISPs <em>do<\/em> know that you visited that website, when you visited it, how long you stayed there, and how much data went back and forth.<\/p>\n<p>They know this a couple of ways. First, if your website uses <a href=\"https:\/\/en.wikipedia.org\/wiki\/Server_Name_Indication\">Server Name Indication<\/a> (SNI) to allow multiple HTTPS sites to be served from a single IP address, the hostname is sent in the clear so that the server knows which certificate to use for the connection. Second, and more importantly, your DNS traffic gives you away. Whether you're going to Amazon.com or BobsEmporiumOfDiscountFurryMemorabilia.com, your computer needs to resolve that domain name to an IP address. That's done in the clear, meaning it's easily intercepted (and even changeable in flight!) by your ISP (or any other <a href=\"https:\/\/en.wikipedia.org\/wiki\/Man-in-the-middle_attack\">MITM<\/a>) whether you're actually using your ISP's DNS servers or not.<\/p>\n<p>This is already enough to build a valuable profile on you for advertising purposes. Depending on your level of paranoia, it's also enough to build a profile on you for blackmail purposes or to completely compromise your Web traffic if you aren't incredibly careful and observant. Imagine an attacker <a href=\"https:\/\/arstechnica.com\/security\/2016\/09\/firefox-ready-to-block-certificate-authority-that-threatened-web-security\/\">has the use of a Certificate Authority to generate their own (valid!) certificates<\/a>; with both that and DNS, they can easily redirect you to a server of their own choosing, which uses a certificate your browser trusts to set up an invisible proxy between you and the site you're trying to securely access. Even without the use of a rogue CA, control of your DNS makes it easier for an attacker to use <a href=\"https:\/\/arstechnica.com\/security\/2017\/04\/chrome-firefox-and-opera-users-beware-this-isnt-the-apple-com-you-want\/\">punycode domain names<\/a> and similar tricks to slide under your radar.<\/p>\n<p>Beyond that, any unencrypted traffic\u2014including but not limited to HTTP (plain old port 80 Web traffic), much peer-to-peer traffic, and more\u2014can be simply edited on-the-fly directly. Which, may I remind you, ISPs have repeatedly <a href=\"https:\/\/arstechnica.com\/tech-policy\/2014\/09\/meet-the-tech-company-performing-ad-injections-for-big-cable\/\">demonstrated themselves<\/a> as perfectly willing to do.<\/p>\n<p>You can't protect yourself from <em>all<\/em> potential attackers. Unfortunately, an awful lot of the critical infrastructure of your access to the Web <em>is<\/em> unencrypted and really <em>cannot<\/em> be secured. As a person with limited resources who can't afford to consider personal security more than a part-time job, you (and I) are unfortunately closer to Secret Squirrel than to James Bond. You can, however, move your vulnerable, unencrypted transmissions out of your <em>ISP's<\/em> reach. So that's what we'll aim to do here.<\/p>\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2015\/06\/freewifi.jpg\" class=\"enlarge\" data-height=\"1661\" data-width=\"2500\"><img alt=\"If you think this will be secure, please read on.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2015\/06\/freewifi-640x425.jpg\" width=\"640\" height=\"425\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2015\/06\/freewifi-1280x850.jpg 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2015\/06\/freewifi.jpg\" class=\"enlarge-link\" data-height=\"1661\" data-width=\"2500\">Enlarge<\/a> \/ If you think this will be secure, please read on.<a rel=\"nofollow\" class=\"caption-link\" href=\"https:\/\/www.flickr.com\/photos\/khawkins04\/6170218244\/\">Ken Hawkins<\/a><\/figcaption><\/figure><h2>A problem of trust<\/h2>\n<p>We've already established\u2014actually, your <em>ISP<\/em> has already established\u2014that your ISP cannot be trusted. The obvious solution, then, is a VPN, a Virtual Private Network, that tunnels all of your vulnerable, unencrypted data outside the ISP's reach. The problem is, that data will be just as vulnerable when it exits the endpoint. Essentially, you've traded one set of vulnerabilities for another, hopefully less-troublesome set.<\/p>\n<p>Imagine you've got a particularly pesky evil genius of a little brother who has learned how to tap the network traffic coming out of your room and who delights in embarrassing you at school with the gossip you shared privately (or so you thought) with your friends. One of those friends says, \"Hey, how about you just setup a VPN between your house and mine? Then everything coming out of your room will be encrypted, and your little brother can't mess with it.\" So far, so good, but then it turns out that friend has a pesky evil genius of a little <em>sister<\/em>, and now <em>she's<\/em> reading your IMs and passing the juicy bits back to your little brother. In the end, you're no better than you started off.<\/p>\n<p>This is the situation you're in today <a href=\"https:\/\/arstechnica.com\/security\/2016\/06\/aiming-for-anonymity-ars-assesses-the-state-of-vpns-in-2016\/\">when you start looking at VPN providers<\/a>: perhaps they're trustworthy, perhaps they're not. Unfortunately, the very same characteristics that make them attractive (claims not to log your traffic, no good relationships with authorities, presence in a lot of countries, inexpensive plans) make them... dubious. How do you <em>know<\/em> the company with the offices in Moldova that is charging you $2\/mo for all the bandwidth you can eat isn't actually monetizing your data just like your ISP did, or maybe even worse? Has anybody audited <em>their<\/em> facilities to independently verify any claims of zero logging? Probably not.<\/p>\n<p>Now, a VPN called Private Internet Access recently made a few\u00a0waves by standing up to an FBI subpoena\u00a0to some degree. In a case surrounding a possible bomb threat hoax (<a href=\"https:\/\/www.scribd.com\/doc\/303226103\/Fake-bomb-threat-arrest\">PDF<\/a>), Private Internet Access appears to <a href=\"https:\/\/torrentfreak.com\/vpn-providers-no-logging-claims-tested-in-fbi-case-160312\/\">have made good on its no logging claims<\/a>. According to the <a href=\"https:\/\/www.unitedstatescourts.org\/federal\/flsd\/480077\/1-0.html\">criminal complaint<\/a>, \"a subpoena was sent to London Trust Media and the only information they could provide is that the cluster of IP addresses being used was from the east coast of the United States.\"<\/p>\n<p>Does this mean everyone\u00a0can trust VPNs in general? Of course not. And as for Private Internet Access specifically, one public success doesn't necessarily remove all doubt\u2014prosecutors in the case didn't push any further given they had plenty of other evidence to support their argument. So if you can't trust your ISP and you can't trust a VPN provider either, what's the plan <em>then?<\/em> Well, you're left\u00a0with an option possibly not suitable for your average Internet user: roll your own VPN at an inexpensive cloud hosting provider like Linode or Digital Ocean.<\/p>\n<p>There are no absolute guarantees with this avenue, either, But while you probably can't avoid your local ISP (few of us have more than two choices, if that), the Internet is full of hosting providers. It's a much bigger deal if one of them generates a lot of customer anger for messing around with customer data. These companies are also less likely to roll over quickly for improperly tendered law enforcement requests than a typical ISP (although, again, there are no guarantees). Getting your data safely away from a predatory ISP is one thing; getting it away from a nation-state adversary or <a href=\"https:\/\/en.wikipedia.org\/wiki\/Advanced_persistent_threat\">APT<\/a> that truly <em>wants<\/em> it is <a href=\"https:\/\/xkcd.com\/538\/\">something else entirely<\/a> and probably beyond our scope.<\/p>\n\n      <div itemprop=\"articleBody\" class=\"article-content post-page\">\n        \n\n\n\n\n\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2014\/12\/happydancensa.jpg\" class=\"enlarge\" data-height=\"1510\" data-width=\"2005\"><img alt=\"Actual slide from the Snowden leaks: dramatization of what nation-states do when they come across users not even trying.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2014\/12\/happydancensa-640x482.jpg\" width=\"640\" height=\"482\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2014\/12\/happydancensa-1280x964.jpg 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2014\/12\/happydancensa.jpg\" class=\"enlarge-link\" data-height=\"1510\" data-width=\"2005\">Enlarge<\/a> \/ Actual slide from the Snowden leaks: dramatization of what nation-states do when they come across users not even trying.<\/figcaption><\/figure><\/div><h2>Grab an SSH key and get started<\/h2>\n<p>When it comes to hosting, there are plenty of great providers out there, and in particular I've used both Digital Ocean and Linode extensively. Both offer instances starting at $5\/mo with multiple data centers and great management features. For this project, I chose Digital Ocean, but either (or any of several others) will do.<\/p>\n<p>Before we start setting up a VM, though, let's make sure we can access it securely. If you've already got an SSH key and know how to use it, great. If not, you'll need to make one. On Linux or a Mac, get to the shell and type in <strong>ssh-keygen -t rsa<\/strong>. This will create an SSH keypair\u00a0and give you the option of adding a challenge phrase, which I highly recommend. Without the challenge phrase, anybody who has the files comprising the keypair can log in as you immediately; <em>with<\/em> the challenge phrase, even an attacker with possession of the key can't get in unless they already know your challenge as well (which is never stored anywhere).<\/p>\n<p>On Windows, it's a bit more complicated. <a href=\"http:\/\/www.chiark.greenend.org.uk\/~sgtatham\/putty\/latest.html\">Download PuTTY<\/a> if you don't already have it, then follow <a href=\"https:\/\/www.digitalocean.com\/community\/tutorials\/how-to-create-ssh-keys-with-putty-to-connect-to-a-vps\">DigitalOcean's guide to generating SSH keys with puttygen<\/a>. (The guide is accurate whether you're choosing DigitalOcean as a provider or not.)<\/p>\n<p>Now that you've got a key, back it up and put the challenge phrase in your password manager. Once you've done that, we're finally ready to spin up a VM.<\/p>\n<h2>Spinning up a VM<\/h2>\n<p>We'll briefly walk through the exact steps to set up a VM on DigitalOcean, because this part really isn't hard. And if you decided to use Linode or some other provider instead, you shouldn't have much of a challenge figuring it out.<\/p>\n<p>Once you've browsed to DigitalOcean and logged in (creating a user account first, if necessary), click the big green \"Create Droplet\" button on the upper right. This provides you with a range of choices of operating system, droplet size, and data center. We will be using Ubuntu 16.04 and a $5 droplet (offering 1 CPU, 512 MB RAM, 20 GB storage, and 1TB\/mo of bandwidth). You should probably select the data center physically nearest you\u2014unless you're specifically worried about your government, in which case you can pick one overseas. Under \"additional options,\" I'd recommend Monitoring (which doesn't cost extra). Under \"Add your SSH keys,\" if you haven't already added yours to your account, click \"New SSH key,\" paste in your <em>public<\/em> key (id_rsa.pub) from the pair you generated earlier, give it a useful, human-readable friendly name, and click \"Add SSH key.\" Now make sure that key's checkbox is ticked on the Create Droplet form (after all, it's your way into the VM).<\/p>\n<p>Next, choose a hostname for your droplet (something memorable and easy to identify, preferably) and click \"Create.\" About 15 seconds later, your VPN will be ready to log in, and its IP address will be visible on the main Droplets page. From Linux or a Mac, ssh root@your.new.ip.address. From Windows, enter and save root@your.new.ip.address as a destination, then connect to it, and you're ready to go. Once you've entered in your challenge phrase (if you used one), you'll be staring at something along the lines of <strong>root@ars-vpn-test:~#<\/strong>.<\/p>\n      <div class=\"gallery shortcode-gallery gallery-wide\">\n      <ul><li data-thumb=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-1-150x150.png\" data-src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-1.png\" data-responsive=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-1-980x404.png 1080, https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-1.png 2560\" data-sub-html=\"#caption-1082237\">\n            <figure><div class=\"image\">\n                              <figcaption id=\"caption-1082237\">\n                      The main Droplets page. See the big green button at the upper right? \n\nPUSH THE BUTTON, FRANK.                    <\/figcaption><\/div>\n                                                    <\/figure><\/li>\n                  <li data-thumb=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-2-150x150.png\" data-src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-2.png\" data-responsive=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-2-980x748.png 1080, https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-2.png 2560\" data-sub-html=\"#caption-1082239\">\n            <figure><div class=\"image\">\n                              <figcaption id=\"caption-1082239\">\n                      Ubuntu 16.04 is already the default distribution, but make sure you change the Droplet size to $5\/mo.  The default, $20\/mo, is more VM than you need for this job.                    <\/figcaption><\/div>\n                                                    <\/figure><\/li>\n                  <li data-thumb=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-3-150x150.png\" data-src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-3.png\" data-responsive=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-3.png 1080, https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/digital-ocean-create-droplet-3.png 2560\" data-sub-html=\"#caption-1082241\">\n            <figure><div class=\"image\">\n                              <figcaption id=\"caption-1082241\">\n                      Don't forget to select an appropriate data center, enable Monitoring if you like (it won't cost you anything), and add and select your SSH key. Be sure to name the Droplet something helpful, too. You don't want to be deciphering \"ubuntu-512mb-nyc2-01\" through \"ubuntu-512mb-nyc2-99\" at some point later, if you really catch the \"roll your own servers\" bug.                    <\/figcaption><\/div>\n                                                    <\/figure><\/li>\n              <\/ul><\/div>\n  \n\n<h2>Automatic security upgrades<\/h2>\n<p>Given that the whole point of this exercise is to <em>increase<\/em> your security, not to screw it up. We want to make sure your new VM gets automatic security upgrades as they become available.<\/p>\n<div>\n<pre> root@ars-vpn-test:~$ apt update ; apt dist-upgrade -y&#13;\n root@ars-vpn-test:~$ apt install unattended-upgrades&#13;\n root@ars-vpn-test:~$ dpkg-reconfigure unattended-upgrades<\/pre>\n<\/div>\n<p>What we did here, in order, was to update ourselves with a fresh list of what's available in package repositories, then upgrade everything <em>already<\/em> installed to the newest version, then install the <em>unattended-upgrades<\/em> package, and, finally, reconfigure that package. Speaking of which, you'll be looking at an ncurses text-mode dialog right now asking you to do just that; all you need to do is make sure <strong>&lt;Yes&gt; <\/strong>is highlighted in red (it should be by default), then hit enter twice (once to accept that you want to turn on automatic upgrades and the second to accept the default pattern for automatically installed upgrades, which boils down to \"security related stuff only, not new shiny features\").<\/p>\n\n      <div itemprop=\"articleBody\" class=\"article-content post-page\">\n        \n\n\n\n\n\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/unattended-upgrades-1.png\" class=\"enlarge\" data-height=\"438\" data-width=\"732\"><img alt=\"The first step to improved security is making sure it &lt;em&gt;stays&lt;\/em&gt; improved.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/unattended-upgrades-1-640x383.png\" width=\"640\" height=\"383\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/unattended-upgrades-1.png 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/unattended-upgrades-1.png\" class=\"enlarge-link\" data-height=\"438\" data-width=\"732\">Enlarge<\/a> \/ The first step to improved security is making sure it <em>stays<\/em> improved.<\/figcaption><\/figure><\/div><h2>Installing and configuring OpenVPN server<\/h2>\n<p>OK, now it's time for the chewy part. The packages you'll need are <strong>openvpn<\/strong> and <strong>easy-rsa<\/strong>.<\/p>\n<div>\n<pre> root@ars-vpn-test:~$ apt install openvpn easy-rsa<\/pre>\n<\/div>\n<p>Seriously, that was all there was to the <em>installation,<\/em> but now it's time to start configuring. First of all, we need to enable ipv4 packet forwarding:<\/p>\n<div>\n<pre> root@ars-vpn-test:~$ nano \/etc\/sysctl.conf<\/pre>\n<\/div>\n<p>All you need to do in here is find the line that says <strong>#net.ipv4.ip_forward = 1<\/strong>\u00a0and remove the leading pound sign (\"hashtag\", for you young'uns). This uncomments the line, enabling it to take effect. Press ctrl-X to exit and Y to say Yes, you'd like to save it, and you're back at the command prompt again.<\/p>\n<div>\n<pre> root@ars-vpn-test:~$ sysctl -p<\/pre>\n<\/div>\n<p>The system will inform you that you've set the value <strong>net.ipv4.ip_forward = 1<\/strong>. Congratulations, you're ready to route. Next step is getting the helpful scripts from the <strong>easy-rsa<\/strong> package in place and using them to generate your own private Certificate Authority.<\/p>\n<div>\n<pre> root@ars-vpn-test:~$ cp -a \/usr\/share\/easy-rsa\/* \/etc\/openvpn\/&#13;\n root@ars-vpn-test:~$ cd \/etc\/openvpn&#13;\n root@ars-vpn-test:\/etc\/openvpn$ nano vars<\/pre>\n<\/div>\n<p>The <strong>\/etc\/openvpn\/vars<\/strong> file contains three really important settings - <strong>KEY_SIZE<\/strong>, which controls just what it sounds like and should be set to a <em>minimum<\/em> of 2048 (I use 4096); and <strong>CA_EXPIRE<\/strong> and <strong>KEY_EXPIRE<\/strong>, which control how long your keys are good for, with a default 10-year expiration date. Expired keys don't work anymore, so this is basically a drop-dead date for your VPN configuration. After these dates come and go, you'll have to regenerate and re-distribute your keys again.<\/p>\n<p>You should generally change the rest of the values\u2014<strong>KEY_COUNTRY<\/strong>, <strong>KEY_PROVINCE<\/strong>, and so forth\u2014to match what makes sense for you, but these are basically human-readable fluff, and the computer doesn't particularly care what's in there. Anything, including the default values, will work.<\/p>\n<p>Once you've CTRL-X'ed and Y'ed your way out of <strong>vars <\/strong>and saved it, it's time to get back to work at the shell.<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ source .\/vars&#13;\n root@ars-vpn-test:\/etc\/openvpn$ .\/clean-all<\/pre>\n<\/div>\n<p>You'll get some dire warnings about deleting all your keys here\u2014ignore it. Running a preparatory .\/clean-all is a <strong>mandatory <\/strong>step on modern OpenVPN installations. Just don't run it <em>again<\/em> unless you want to lose all your keys and certificates.<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ .\/build-ca&#13;\n root@ars-vpn-test:\/etc\/openvpn$ .\/build-key-server server<\/pre>\n<\/div>\n<p>There will be quite a lot of pressing [enter] through these steps to confirm the default values you set in <strong>vars<\/strong> earlier. You can just enter your way through, but be sure to actually press Y twice at the end to first sign the certificate, then commit it. Our next step will be building the Diffie-Hellman key, with the <strong>.\/build-dh<\/strong> command. You need to specify a keysize argument on the command line for this one. More is generally better, so I went with a 4096 bit key... which took quite a while to generate; 26 minutes and 35 seconds, to be exact. Once you get this process started, you might want to go watch a TV show or something. It <em>will<\/em> be a while.<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ .\/build-dh 4096<\/pre>\n<\/div>\n<p>Now you're ready to create your actual certificates and keys. Remember how many times you had to press [enter] when you created your CA certificate and your server key a couple of steps above? If you found that annoying, you might want to do some purely optional prep work here. The <strong>build-key<\/strong> and <strong>build-key-pass<\/strong> tools are both human-readable scripts that invoke <strong>pkitool. <\/strong>Both of them can be edited to remove the <strong>--interact<\/strong> argument that they pass it by default, which results in that obnoxious ten-in-a-row [enter] sequence. Instead of editing that line directly (and maybe screwing it up and not knowing how to get it back to the original), copy the line, paste it into a new one, then comment out the first line with a pound sign. Once that's done, remove the <strong>--interact<\/strong> argument in the second line you pasted in. For <strong>build-key<\/strong>, you should wind up with this:<\/p>\n<div>\n<pre> #!\/bin\/sh&#13;\n &#13;\n # Make a certificate\/private key pair using a locally generated&#13;\n # root certificate.&#13;\n &#13;\n export EASY_RSA=\"${EASY_RSA:-.}\"&#13;\n #\"$EASY_RSA\/pkitool\" --interact $*&#13;\n \"$EASY_RSA\/pkitool\" $*<\/pre>\n<\/div>\n<p>... and for <strong>build-key-pass<\/strong>, you'll wind up with this:<\/p>\n<div>\n<pre> #!\/bin\/sh&#13;\n &#13;\n # Similar to build-key, but protect the private key&#13;\n # with a password.&#13;\n &#13;\n export EASY_RSA=\"${EASY_RSA:-.}\"&#13;\n #\"$EASY_RSA\/pkitool\" --interact --pass $*&#13;\n \"$EASY_RSA\/pkitool\" --pass $*<\/pre>\n<\/div>\n<p>Once you've edited and saved both of these (or ignored this step, if you don't mind wearing the paint off your Enter key), you're ready to actually generate some keys and certificates to authenticate with. I generated three keypairs\u2014<strong>client-no-pass<\/strong>, <strong>client-with-pass<\/strong>, and <strong>R8000<\/strong>\u2014the latter being intended for use with my handy-dandy Netgear router with DD-WRT firmware.<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ .\/build-key R8000&#13;\n root@ars-vpn-test:\/etc\/openvpn$ .\/build-key client-no-pass&#13;\n root@ars-vpn-test:\/etc\/openvpn$ .\/build-key-pass client-with-pass<\/pre>\n<\/div>\n<p>Much like when we generated an SSH keypair earlier, the difference between keys generated with <strong>build-key<\/strong> and <strong>build-key-pass<\/strong> is that the first work \"naked,\" but the second will require a challenge phrase to be entered in when the OpenVPN tunnel is started each time. In general, keypairs with a challenge are more appropriate for credentials you'll use directly at your computer or phone and start and stop manually, while keypairs <em>without<\/em> a challenge phrase are more appropriate for unattended use in which devices might automatically start and stop tunnels without user intervention. (DD-WRT can actually use keypairs with challenge phrases, but in doing so it subverts most of the point: you have to save the challenge phrase <em>in<\/em> the configs; so presumably an attacker that got hold of the keypair would be able to get hold of the challenge phrase just as easily.)<\/p>\n<p>Now that you've got your credentials generated, it's time to configure the server itself. First, let's grab a reference copy of a typical-ish OpenVPN server config from where the packages left it:<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ cp \/usr\/share\/doc\/openvpn\/examples\/sample-config-files\/server.conf.gz \/etc\/openvpn\/&#13;\n root@ars-vpn-test:\/etc\/openvpn$ gunzip .\/server.conf.gz&#13;\n root@ars-vpn-test:\/etc\/openvpn$ mv server.conf server.conf.dist<\/pre>\n<\/div>\n<p>It can be pretty overwhelming wading through the mass of comments in that default config file, so we're going to rename it something safe\u2014<strong>server.conf.<em>dist<\/em><\/strong>, so that the system won't try to execute it\u2014and just start over with a clean file. That will look like this:<\/p>\n<div>\n<pre> # \/etc\/openvpn\/server.conf - Ars Technica Edition&#13;\n port 1194&#13;\n proto udp&#13;\n dev tun&#13;\n &#13;\n ca \/etc\/openvpn\/keys\/ca.crt&#13;\n cert \/etc\/openvpn\/keys\/server.crt&#13;\n key \/etc\/openvpn\/keys\/server.key # This file should be kept secret&#13;\n dh \/etc\/openvpn\/keys\/dh4096.pem&#13;\n &#13;\n cipher AES-256-CBC&#13;\n auth SHA512&#13;\n &#13;\n server 10.8.0.0 255.255.255.0&#13;\n push \"redirect-gateway def1 bypass-dhcp\"&#13;\n push \"dhcp-option DNS 8.8.8.8\"&#13;\n push \"dhcp-option DNS 8.8.4.4\"&#13;\n &#13;\n ifconfig-pool-persist ipp.txt&#13;\n keepalive 10 120&#13;\n &#13;\n comp-lzo&#13;\n &#13;\n persist-key&#13;\n persist-tun&#13;\n &#13;\n status openvpn-status.log&#13;\n verb 3<\/pre>\n<\/div>\n<p>The settings<strong> port 1194<\/strong>,<strong> proto udp<\/strong>, and<strong> dev tun<\/strong> set us on the IANA reserved port for OpenVPN, using the UDP protocol (highly recommended; OpenVPN can use TCP, but your performance will suffer <em>significantly<\/em> for it if you do) and the <strong>tun<\/strong> style of virtual network adapter rather than <strong>tap<\/strong>. The differences between tun and tap are fairly arcane, but at the end of the day <strong>tun<\/strong> both provides more security for individual client devices from other client devices and, in my experience, it's also less likely to crash for no apparent reason, so <strong>tun<\/strong> it is.<\/p>\n<p>The next block of four settings should be pretty self-explanatory, pointing OpenVPN to the location of the credentials it needs to run. It gets more interesting with <strong>cipher<\/strong> and <strong>auth. <\/strong>These are the cipher used to encrypt your data and the digest used to handle authentication, respectively. The options I've chosen here are err on the paranoid side, which won't hurt your performance any on your own computer or at the VM you're running the server on (even my Celeron J1900 Homebrew can push &gt;200 Mbps with these settings), but it may be a bit much for a consumer router if you're setting up a router-based network-wide VPN tunnel like we will be.<\/p>\n<p>If you're interested in twiddling the performance-vs-security slider, I've got your back: my Netgear R8000 test router managed about 25 Mbps throughput configured with <strong>AES-256-CBC\/SHA512<\/strong> as shown here. <strong>AES-256-CBC\/SHA1, <\/strong>which is still reasonably secure, got a more respectable 37.24 Mbps. That's about as far down as you can go and still call the result \"a VPN\" with a straight face, though. If you want better performance than that, you should probably consider either running the VPN client directly on your computer itself or on a more powerful router.<\/p>\n      <div class=\"gallery shortcode-gallery gallery-wide\">\n      <ul><li data-thumb=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/R8000-speedtest-2-150x150.png\" data-src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/R8000-speedtest-2.png\" data-responsive=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/R8000-speedtest-2.png 1080, https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/R8000-speedtest-2.png 2560\" data-sub-html=\"#caption-1082725\">\n            <figure><div class=\"image\">\n                              <figcaption id=\"caption-1082725\">\n                      This is pretty representative of the Netgear R8000's throughput on my most paranoid OpenVPN settings. Technically, this is *double* VPN'ed\u2014the Nighthawk built its tunnel *through* the tunnel the Homebrew is already running upstream of it. However, the real constraint here (on the download side, at least) is the R8000's own performance\u2014it can't hit anywhere near the nominal 100 Mbps provided by Spectrum or even whatever's left of that after the Netflix streaming and whatever else my wife, kids, and my other computers were doing while the tests ran.                    <\/figcaption><\/div>\n                                                    <\/figure><\/li>\n                  <li data-thumb=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2-150x150.png\" data-src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2.png\" data-responsive=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2.png 1080, https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2.png 2560\" data-sub-html=\"#caption-1082723\">\n            <figure><div class=\"image\">\n                              <figcaption id=\"caption-1082723\">\n                      This is a speedtest.net run on the Homebrew, powered by a Celeron J1900, using the nice-and-paranoid combination of AES-256-CBC\/SHA512. The results you're seeing\u201480 Mbps down and 5-8 Mbps up\u2014are the limits of my Internet connection and\/or the stuff my wife, kids, and various computers were doing while I ran the test. When not constrained by the actual network connection, the Homebrew will push &gt; 200 Mbps of OpenVPN traffic.                    <\/figcaption><\/div>\n                                                    <\/figure><\/li>\n              <\/ul><\/div>\n  \n\n<p>Moving on, <strong>server 10.8.0.0 255.255.255.0 <\/strong>defines the IP address range used by the server and its clients. As configured, the server itself will occupy 10.8.0.1 (and several more addresses), and each client will have its own address. Actually, since we're using the <strong>tun<\/strong> adapter, each client will use <em>four<\/em> IP addresses in its own <a href=\"http:\/\/www.subnet-calculator.com\/cidr.php\">\/30 subnet<\/a>\u2014up to 10.8.0.255. This means we'll be limited to about 60 total clients, so if you're setting this VPN up for your friends and your friends' friends and maybe some people you don't even really like all that much, you may need to consider a bigger subnet.<\/p>\n<p>The next line, <strong>push \"redirect-gateway def1 bypass-dhcp\"<\/strong>, is optional but recommended for our purposes. It instructs any connecting client to route all of its traffic across the VPN, ignoring any settings it might have to the contrary from its local DHCP server. It's also possible to set this directly in the <em>client<\/em> configuration, which we'll cover next. The <strong>push \"dhcp-option DNS\" <\/strong>lines tell OpenVPN to force the client to use Google's multicast DNS servers instead of whatever it was using previously, which might have been an ISP-controlled DNS server or servers. If you don't like Google either, you can choose Level 3's DNS servers at 4.2.2.4 and 4.2.2.2, or OpenDNS at 208.67.222.222 and 208.67.220.220. (Warning: although they do it \"for good reasons,\" OpenDNS edits the results returned from DNS lookups done against their servers. <em>Caveat Emptor<\/em>.)<\/p>\n<p>The <strong>ifconfig-pool-persist<\/strong> directive has OpenVPN keep track of what tunnel IP addresses it has handed out to which clients in the past and try to maintain some consistency as it hands out new ones in the future. The <strong>keepalive<\/strong> directive sends pings down the tunnel and restarts it if they don't get back within a certain amount of time. The <strong>comp-lzo <\/strong>directive uses LZO compression on the tunnel's contents, which is typically a win. Much of what you send won't be further compressible, but <strong>LZO<\/strong> is so cheap as to be nearly free in terms of CPU time, so I prefer to use it when possible. The <strong>persist-key<\/strong> and <strong>persist-tun<\/strong> directives try to reuse existing pieces of the setup wherever possible when restarting the tunnel, <strong>status<\/strong> keeps a log file, and <strong>verb<\/strong> sets how verbose the logging is.<\/p>\n<p>If you change any of these settings from what they're shown here, make a note of it. You'll need to make the same changes on your client configs later. Aside from that, you're ready to start your OpenVPN server for the first time!<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ systemctl enable openvpn&#13;\n root@ars-vpn-test:\/etc\/openvpn$ systemctl start openvpn<\/pre>\n<\/div>\n<p>Now check to make sure it's actually running:<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn$ ps wwaux | grep vpn | grep -v grep<\/pre>\n<\/div>\n<p>This should return a process ID and a command mutex for OpenVPN, like this:<\/p>\n<div>\n<pre> root@ars-vpn-test:\/etc\/openvpn# ps wwaux | grep vpn | grep -v grep&#13;\n root 5493 0.0 1.1 35792 5576 ? Ss 18:01 0:01 \/usr\/sbin\/openvpn --daemon ovpn-server --status &#13;\n \/run\/openvpn\/server.status 10 --cd \/etc\/openvpn --script-security 2 --config \/etc\/openvpn\/server.conf&#13;\n --writepid \/run\/openvpn\/server.pid<\/pre>\n<\/div>\n<p>If you get no output at all from this command, you might try starting the openvpn server in interactive mode to see if its output helps you troubleshoot. If you never get to \"Initialization Sequence Completed,\" you've got a problem you'll need to resolve before moving on.<\/p>\n<p>Now that you've got your OpenVPN server running, you'll need to configure NAT from the tunnel to the outside world. \u00a0For that, you'll need to place a simple script in\u00a0<strong>\/etc\/network\/if-pre-up.d\/masquerade:<\/strong><\/p>\n<div>\n<pre>#!\/bin\/sh&#13;\n\/sbin\/iptables -t nat -A POSTROUTING -s 10.8.0.0\/24 -o eth0 -j MASQUERADE<\/pre>\n<\/div>\n<p>Once you've created\u00a0<strong>\/etc\/network\/if-pre-up.d\/masquerade<\/strong>, you'll need to\u00a0<strong>chmod 755 \/etc\/network\/if-pre-up.d\/masquerade<\/strong>, and then run it directly to bring it up for the first time.<\/p>\n<div>\n<pre>root@ars-vpn-test:\/etc\/openvpn$ \/etc\/network\/if-pre-up.d\/masquerade<\/pre>\n<\/div>\n<p>That's it - you'll NAT all traffic coming from the tunnel now before it's passed on to the internet at large; and the NAT config will come up again automatically any time a network interface is raised in the future, since it's in <strong>\/etc\/network\/if-pre-up.d<\/strong>.<\/p>\n\n      <div itemprop=\"articleBody\" class=\"article-content post-page\">\n        \n\n\n\n\n\n<h2>Client credentials and configs<\/h2>\n<p>For each client you want to connect, you need three credential files: the Certificate Authority cert (this is the same for all clients) at <strong>\/etc\/openvpn\/keys\/ca.crt<\/strong>, the client cert at <strong>\/etc\/openvpn\/keys\/clientname.crt<\/strong>, and the client's private key at <strong>\/etc\/openvpn\/keys\/clientname.key<\/strong>. (If you want to be sure that even an attacker who gets root on your server can't connect to the server later using your own VPN infrastructure, you can delete the <strong>clientname.key<\/strong> from the server once you have it safely available on the client.)<\/p>\n<p>If you're on Linux or Mac, you can use the <strong>scp<\/strong> tool to grab these files:<\/p>\n<div>\n<pre> you@linuxormac:~$ scp root@yourserver:\/etc\/openvpn\/keys\/ca.crt .\/&#13;\n you@linuxormac:~$ scp root@yourserver:\/etc\/openvpn\/keys\/clientname.crt .\/&#13;\n you@linuxormac:~$ scp root@yourserver:\/etc\/openvpn\/keys\/clientname.key .\/<\/pre>\n<\/div>\n<p>If you're a Windows user, it's usually easiest to just <strong>cat \/etc\/openvpn\/keys\/ca.crt<\/strong> inside your PuTTY window, then highlight the text with your mouse. It's automatically copied into the clipboard just from highlighting it; you don't need to Ctrl-C or anything. From there, you can open up a Notepad instance, paste into there, then Save As ca.crt (be sure to change the file type to \"All Files\" first so Windows doesn't \"helpfully\" stick a .txt on the end of it). Do the same for the other two files as well.<\/p>\n<p>Now, you'll need to install OpenVPN\u2014<strong>sudo apt install openvpn<\/strong> on an Ubuntu machine. On a Mac, you can either use Homebrew to install a command-line version and treat it like Linux, or you can download and install <a href=\"https:\/\/tunnelblick.net\/\">Tunnelblick<\/a> (which you're on your own configuring). On Windows, you'll need to <a href=\"https:\/\/openvpn.net\/index.php\/open-source\/downloads.html\">download and install OpenVPN<\/a> from its website. On Linux or (non-Tunnelblick) Mac, you'll need to move the files to <strong>\/etc\/openvpn\/keys<\/strong> (which you may need to create if it doesn't already exist). On Windows, they need to go in <strong>C:\\Program Files\\OpenVPN\\config\\keys<\/strong> (the config directory will be there already, but you'll need to create the keys directory).<\/p>\n<p>Next, you're ready to create the last piece of the puzzle: your OpenVPN <em>client<\/em> configuration file, which we will cleverly name <strong>clientname.conf<\/strong> on a Linux or Mac machine, or <strong>clientname.ovpn<\/strong> on a Windows machine (again being careful to avoid accidentally getting an extra<strong> .txt<\/strong> stuck on the end). This goes in <strong>\/etc\/openvpn\/clientname.conf<\/strong> on Linux or (non-Tunnelblick) Mac or in <strong>C:\\Program Files\\OpenVPN\\config\\clientname.ovpn<\/strong>\u00a0(again, be sure to avoid the tricky .txt Notepad will try to stick on the end) on Windows.<\/p>\n<div>\n<pre> # OpenVPN clientname.ovpn - Ars Technica Edition&#13;\n ca keys\/ca.crt&#13;\n cert keys\/clientname.crt&#13;\n key keys\/clientname.key&#13;\n &#13;\n remote your.server.ip.address 1194&#13;\n comp-lzo&#13;\n client&#13;\n dev tun&#13;\n redirect-gateway def1&#13;\n &#13;\n remote-cert-tls server&#13;\n cipher AES-256-CBC&#13;\n auth SHA512&#13;\n &#13;\n proto udp&#13;\n resolv-retry infinite&#13;\n nobind&#13;\n &#13;\n # Try to preserve some state across restarts.&#13;\n persist-key&#13;\n persist-tun&#13;\n &#13;\n # Set log file verbosity.&#13;\n verb 3&#13;\n mute 20<\/pre>\n<\/div>\n<p>We're not going to go through this one line by line, since it mostly just matches the server config file we already looked at. However, be sure that if you changed anything in the server config from our template here, you also change it to match in the client\u2014particularly the <strong>comp-lzo<\/strong>, <strong>cipher<\/strong>, <strong>proto<\/strong>, and <strong>auth<\/strong> directives. The <strong>redirect-gateway def1<\/strong> option isn't strictly necessary if you left the matching option in the server's config file. It does the same thing here, and specified in either location it will have the same effect. It doesn't hurt to have it on both ends for a sort of belt-and-suspenders that assures you you're redirecting all your traffic down the tunnel no matter which config you decide to look at, though.<\/p>\n<p>Once you've got this file in place and ready to go, you can fire up your VPN manually\u2014<strong>openvpn \/etc\/openvpn\/clientname.conf <\/strong>on Ubuntu or (non-Tunnelblick) Mac; or right-click the .ovpn file and select <strong>Start OpenVPN on this config file<\/strong> under Windows. (Windows also installed an OpenVPN GUI icon in your system tray, which you can interact with there if you like.) If it connects and gets all the way to \"Initialization Sequence Completed,\" you should be good to go. Test that your traffic goes through by doing <strong>traceroute -n 8.8.8.8 <\/strong>on Linux or Mac (you may need to install the <strong>traceroute<\/strong> package first) or <strong>tracert -d 8.8.8.8 <\/strong>on Windows, making sure the route goes through your VPN IP addresses.<\/p>\n<p>Once you're satisfied that your VPN is up and working properly, if you'd like to make it an everyday thing, enable it as a service. On Ubuntu, <strong>systemctl enable openvpn ; systemctl start openvpn<\/strong> will do the trick. On Windows, go into the Services applet from the Control Panel, set the OpenVPN Service to \"automatic\", then start it. On a Mac using Homebrew, <a href=\"http:\/\/www.aandcp.com\/launchdaemons-and-mac-os-x-openvpn-as-an-example\">set up a LaunchDaemon<\/a>, or if using TunnelBlick, <a href=\"https:\/\/help.my-private-network.co.uk\/support\/solutions\/articles\/6000044022-os-x-tunnelblick-openvpn-auto-start\">set it to connect automatically<\/a>.<\/p>\n<h2>Um... now Netflix won't work.<\/h2>\n<p>Yeah, that's a problem you're going to have. Netflix is under some pretty odious obligations to region-lock a lot of their content, since they're only licensed to show some UK content to UK users, US content to US users, and so forth. The content providers <em>do<\/em> complain if they perceive that their region locks aren't being honored (as do Netflix's <em>competitors<\/em> in other regions, such as SKY). As a result, Netflix tends to block known data centers, proxy servers, and vpn providers as much as possible.<\/p>\n<p>If you're running your OpenVPN client directly on your computer, you'll likely have to drop the VPN connection whenever you want to use Netflix (or other media service that blocks your data center). Luckily, if you want to run your OpenVPN directly on the router that your whole network goes through, you've got a better option: policy-based routing.<\/p>\n<h2>OpenVPN on a Homebrew router<\/h2>\n<p>If you're running <a href=\"https:\/\/arstechnica.com\/gadgets\/2016\/04\/the-ars-guide-to-building-a-linux-router-from-scratch\/\">a Homebrew router<\/a> like mine, getting the whole network behind your new VPN is almost embarrassingly easy. Following the Ubuntu instructions above, you just <strong>apt update ; apt install openvpn<\/strong>, put the <strong><em>clientname<\/em>.conf<\/strong> in <strong>\/etc\/openvpn <\/strong>and the <strong>ca.crt<\/strong>, <em><strong>clientname<\/strong><\/em><strong>.key<\/strong> and <em><strong>clientname<\/strong><\/em><strong>.crt<\/strong> in <strong>\/etc\/openvpn\/keys<\/strong>, <strong>systemctl enable openvpn ; systemctl start openvpn<\/strong> and poof, you're connected. The only thing left to do after that is to allow forwarding with masquerade across the VPN tunnel as well as across the WAN, in <strong>\/etc\/network\/if-pre-up.d\/firewall<\/strong>.<\/p>\n<p>You can find a full set of sample configs for a Homebrew router (including dhcpd, bind9, firewall, openvpn, and more configs) at my <a href=\"https:\/\/github.com\/jimsalterjrs\/homebrew-router\">Github repository<\/a>, and it would probably be a good idea to reference them there if you really want to do it for yourself. We really only added two lines to our firewall: <strong>VPNOUT=tun0<\/strong> to define an interface for our outbound VPN connection and <strong>-A POSTROUTING -o $VPNOUT -j MASQUERADE<\/strong> to allow outbound, NATted traffic through the tunnel.<\/p>\n<p>Once through that, everything seemed to work swimmingly on my own home network... until I got a plaintive cry from upstairs. My daughter Jane stumbled on the same problem we mentioned above, \"Daddy, do you know why Netflix says we're using a blocker?\" Oops.<\/p>\n<p>As such, the next task for my Homebrew router was <em>policy-based routing.<\/em> I needed to tell it <em>not<\/em> to send any traffic from my Rokus through the tunnel, instead allowing them direct access through the WAN interface. The first step was finding their MAC addresses (visible from their Network Settings screens) and using these to add host definitions to my Homebrew's <strong>\/etc\/dhcp\/dhcpd.conf<\/strong> file:<\/p>\n<div>\n<pre> subnet 192.168.0.0 netmask 255.255.255.0 {&#13;\n range 192.168.0.100 192.168.0.199;&#13;\n option routers 192.168.0.1;&#13;\n option domain-name-servers 192.168.0.1;&#13;\n option broadcast-address 192.168.0.255;&#13;\n }&#13;\n &#13;\n host Downstairs-Roku {&#13;\n hardware ethernet B0:A7:1A:97:3D:5E;&#13;\n fixed-address 192.168.0.50;&#13;\n }&#13;\n &#13;\n host Upstairs-Roku {&#13;\n hardware ethernet b8:3e:4e:d4:5b:62;&#13;\n fixed-address 192.168.0.51;&#13;\n }<\/pre>\n<\/div>\n<p>Once <strong>\/etc\/dhcp\/dhcpd.conf <\/strong>was reconfigured to add the host leases, <strong>systemctl restart isc-dhcp-server<\/strong> got them applied. At the Rokus themselves, at the same Settings -&gt; Network Settings screen I'd gotten their MAC addresses from, a simple \"Update network settings\" released and renewed their DHCP leases and then I was ready to turn them into special, non-VPN'ed snowflakes. All this required was adding three quick lines to the WAN interface settings in <strong>\/etc\/network\/interfaces<\/strong>:<\/p>\n<div>\n<pre> # This file describes the network interfaces available on your system&#13;\n # and how to activate them. For more information, see interfaces(5).&#13;\n &#13;\n source \/etc\/network\/interfaces.d\/*&#13;\n &#13;\n # The loopback network interface&#13;\n auto lo&#13;\n iface lo inet loopback&#13;\n &#13;\n # the LAN interface&#13;\n auto enp2s0&#13;\n iface enp2s0 inet static&#13;\n address 192.168.0.1&#13;\n netmask 255.255.255.0&#13;\n dns-nameservers 8.8.8.8 8.8.4.4&#13;\n &#13;\n # The WAN interface&#13;\n auto enp1s0&#13;\n iface enp1s0 inet dhcp &#13;\n # Rokus on .50 and .51 need their traffic to bypass&#13;\n # the outbound VPN, so that Netflix won't block it.&#13;\n #&#13;\n # we accomplish that by adding its traffic to a &#13;\n # route table that we're implicitly creating here&#13;\n # by referencing it with a number.&#13;\n #&#13;\n post-up ip rule add from 192.168.0.50 lookup 100&#13;\n post-up ip rule add from 192.168.0.51 lookup 100&#13;\n post-up ip route add default dev enp1s0 table 100&#13;\n<\/pre>\n<\/div>\n<p>Those last three lines\u2014the post-up ip rule and post-up ip route directives\u2014are the only things that actually changed from the original config I was using. The settings in <strong>\/etc\/network\/interfaces<\/strong> don't actually take effect until after a reboot, so I jumped the gun a bit by running them directly at the command line (just as they're shown in the config file, only without the \"post-up\" bit at the start). Presto, the Rokus work again; Netflix no longer accuses me of using \"an unblocker,\" which means no more plaintive daughter calling downstairs.<\/p>\n<p>At this point, the entire house is getting its traffic routed safely out from under the ISP's nose to an endpoint at Digital Ocean\u2014except for the Rokus, of course, which are happily going out directly through my residential ISP where Netflix won't complain about them. It has actually been running this way for several days now. When I told my wife last night that all her stuff was going out through a VPN, her only response was to blink at me and say \"Huh.\"<\/p>\n\n      <\/div><div itemprop=\"articleBody\" class=\"article-content post-page\">\n        \n\n\n\n\n\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2.png\" class=\"enlarge\" data-height=\"650\" data-width=\"730\"><img alt=\"This is a speedtest.net run on the Homebrew, powered by a Celeron J1900, using the nice-and-paranoid combination of AES-256-CBC\/SHA512. The results you're seeing\u201480 Mbps down and 5-8 Mbps up\u2014are the limits of my Internet connection and\/or the stuff my wife, kids, and various computers were doing while I ran the test. When not constrained by the actual network connection, the Homebrew will push &amp;gt; 200 Mbps of OpenVPN traffic.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2-640x570.png\" width=\"640\" height=\"570\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2.png 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/homebrew-speedtest-2.png\" class=\"enlarge-link\" data-height=\"650\" data-width=\"730\">Enlarge<\/a> \/ This is a speedtest.net run on the Homebrew, powered by a Celeron J1900, using the nice-and-paranoid combination of AES-256-CBC\/SHA512. The results you're seeing\u201480 Mbps down and 5-8 Mbps up\u2014are the limits of my Internet connection and\/or the stuff my wife, kids, and various computers were doing while I ran the test. When not constrained by the actual network connection, the Homebrew will push &gt; 200 Mbps of OpenVPN traffic.<\/figcaption><\/figure><\/div><h2>OpenVPN on a (Netgear) consumer router<\/h2>\n<p>You can also run OpenVPN on all sorts of consumer routers, running either OpenWRT or DD-WRT. I'm going to just talk very specifically about getting it running on a Netgear Nighthawk here for good reason: Netgear directly runs <a href=\"https:\/\/www.myopenrouter.com\/\">myopenrouter.com<\/a>, where they actually <em>collaborate<\/em> with open source developers who are adapting builds of open source firmware for installation on Netgear routers. This is extremely cool, not least because it means that you can install firmware from myopenrouter directly onto a supported Netgear router <em>using the router's own Web-based interface<\/em>.<\/p>\n<p>It's certainly possible to install DD-WRT or OpenWRT on a non-Netgear consumer router, but it's generally a giant pain in the ass and a good way to potentially brick your router. Typically, the vendor doesn't support it, doesn't <em>want<\/em> you to do it, and you have to put the router into a TFTP firmware flash mode and\/or use some sort of\u00a0stack-smashing hack to break out of the OEM firmware in the first place. Thousands of people do exactly that, of course. A disturbing number of them brick a few routers along the way, though, so the heck with that: Netgear is actually <em>supporting<\/em> open source, so here's me supporting them right back.<\/p>\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2016\/09\/17-nighthawk-x6.jpg\" class=\"enlarge\" data-height=\"3456\" data-width=\"4608\"><img alt=\"It's a facehugger. It's a router. It's a Nighthawk!\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2016\/09\/17-nighthawk-x6-640x480.jpg\" width=\"640\" height=\"480\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2016\/09\/17-nighthawk-x6-1280x960.jpg 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2016\/09\/17-nighthawk-x6.jpg\" class=\"enlarge-link\" data-height=\"3456\" data-width=\"4608\">Enlarge<\/a> \/ It's a facehugger. It's a router. It's a Nighthawk!Jim Salter<\/figcaption><\/figure><p>I had a Netgear Nighthawk X6 on hand already (the very same one I used in <a href=\"https:\/\/arstechnica.com\/gadgets\/2016\/01\/numbers-dont-lie-its-time-to-build-your-own-router\/\">the first Homebrew router article<\/a>). I fired it up, browsed to <a href=\"https:\/\/www.myopenrouter.com\/\">myopenrouter.com<\/a>, and logged in. (You do need to create a free-as-in-beer account before you can download any of the firmware from the site.) From there, I clicked Downloads, changed the Search Downloads combo box to \"R8000\" (the internal, non-market-y codename of my Nighthawk X6), and downloaded the latest \"DD-WRT Kong Mod\" for the router. That was a ZIP file, which I extracted into my Downloads directory. With that done, I logged into the Nighthawk's Web interface and went to Advanced -&gt; Security -&gt; Router Update, browsed to the .CHK file I'd extracted from the ZIP, and that was that. (You might get a warning that the firmware you've selected is older than your current firmware, but that's just a \"misfeature.\" It compares version numbers, but the version number history for Kong's DD-WRT mods is completely separate from the version number history for Netgear's OEM firmware.)<\/p>\n<p>That's really all there is to installing Kong's DD-WRT build on a Nighthawk; after ignoring the version number mismatch, it installs and reboots the router automagically in just a minute or two, and you're ready to rock, with the router on 192.168.1.1 and handing out addresses in the 192.168.1.0\/24 range on your LAN. The first thing you'll notice after grabbing an IP address, browsing to http:\/\/192.168.1.1, and logging back into your newly DD-WRT'ed Nighthawk is that it tells you that you have to change your username and password, which are currently insecure default. Do that, and I would suggest you pick the username \"root.\" I'm not sure it actually pays attention to the username field here, even though you're allowed to change it.<\/p>\n<p>You should already be set up reasonably well for most residential or small business Internet connections at this point, with a local subnet of 192.168.1.0\/24, DHCP running on the WAN, and dnsmasq providing DHCP for the LAN. One thing you absolutely <strong>need<\/strong> to change, though, is the DNS setting in the \"home page\" of DD-WRT's interface. By default it's 0.0.0.0 in all three blocks and will be filled by the ISP's settings pushed to you via DHCP. You do <strong>not<\/strong> want that, since our whole point here is camouflaging things from your ISP. So instead, set them to 8.8.8.8, 8.8.4.4, and 4.2.2.4 (Google, Google, and Layer3 anycast DNS server addresses). If you leave any of them blank, they will get filled with your ISP's servers, defeating most of our purpose.<\/p>\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-basic-setup.png\" class=\"enlarge\" data-height=\"909\" data-width=\"807\"><img alt=\"Note that the DNS settings are COMPLETELY filled in here. If you leave one of the three blanks at 0.0.0.0, it will be overwritten by one of your ISP's DNS servers, potentially leaking information to them (or opening up vulnerabilities for manipulation) that you don't want.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-basic-setup-640x721.png\" width=\"640\" height=\"721\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-basic-setup.png 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-basic-setup.png\" class=\"enlarge-link\" data-height=\"909\" data-width=\"807\">Enlarge<\/a> \/ Note that the DNS settings are COMPLETELY filled in here. If you leave one of the three blanks at 0.0.0.0, it will be overwritten by one of your ISP's DNS servers, potentially leaking information to them (or opening up vulnerabilities for manipulation) that you don't want.<\/figcaption><\/figure><p>Setting up the VPN itself is pretty easy if you know what you're doing. Navigate to Services, then VPN, and look for \"Client.\" Tick the radio box to enable Client and then the one that opens up the \"Advanced\" options for the client. Find your ca.crt, and the client .crt and .key (in my case, I generated R8000.crt and R8000.key earlier), and paste them into the appropriate boxes. Set \"Encryption Cipher\" to AES-256-CBC and \"Hash Algorithm\" to SHA512 (or whatever you set them to in your OpenVPN server's config file), set NAT to \"enabled,\" and check the box for \"nsCertType verification.\"<\/p>\n<p>Finally, to avoid the dreaded Netflix problem, we take advantage of DD-WRT's built-in policy-based routing feature. This is considerably wimpier than the one I used under Ubuntu for the Homebrew, but it's much, <em>much<\/em> easier than getting the job done from DD-WRT's command line. (You can shell into DD-WRT with SSH, if you need to and if you enable that option. That's... kinda cowboy, though, so since we <em>can<\/em> avoid it, we're <em>going<\/em> to.)<\/p>\n<figure class=\"image shortcode-img center full\"><img alt=\"The DD-WRT VPN dialog, with most of the things you'll need to change from default values highlighted. Note that the actual CA certificate and client certificate and key aren't shown here\u2014they're just off the bottom of the screen, with the &quot;Save&quot; and &quot;Apply&quot; buttons underneath them a little further.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-services-vpn.png\" width=\"594\" height=\"880\" \/><figcaption class=\"caption\">The DD-WRT VPN dialog, with most of the things you'll need to change from default values highlighted. Note that the actual CA certificate and client certificate and key aren't shown here\u2014they're just off the bottom of the screen, with the \"Save\" and \"Apply\" buttons underneath them a little further.<\/figcaption><\/figure><p>Under DD-WRT, any address or subnet you enter into the \"policy-based routing\" text box will be <em>routed<\/em> through the VPN, not excluded from it. You do need to be careful, though. If you get frisky and put the wrong thing in here\u2014like 192.168.1.0\/24, the entire local subnet\u2014you can pretty easily lock your router up tighter than a drum. If you do that, it's not going to talk to you again until it has been rebooted... and won't talk to you <em>then<\/em>, either, if it manages to reconnect its VPN. So if you've managed to lock it up this way, you'll want to unplug its WAN cable, <em>then<\/em> reboot it, which will allow you back into the interface to fix your screw-up. Dire warnings done, what you'll want to enter in here are three single lines: 192.168.1.64\/26, 192.168.1.128\/26, and 192.168.1\/192\/26. These add up to mean that anything from 192.168.1.64-192.168.1.254 will get routed out through your VPN, leaving anything from 192.168.1.1 (the router, which <em>must<\/em> be allowed direct access) through 192.168.1.63 able to just hit the 'Net directly.<\/p>\n<p>That's it. Scroll down to the bottom of the page, click Save first, then click Apply. Your VPN connection will fire up and should finish connecting in about 10 to 20 seconds. You can then check on it under Status -&gt; OpenVPN\u2014once it hits \"Initialization Sequence Completed,\" all is well.<\/p>\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-status-openvpn.png\" class=\"enlarge\" data-height=\"879\" data-width=\"814\"><img alt=\"You get lots of chewy dialog here when the VPN client attempts to connect to the server. The money line here is &quot;Initialization Sequence Completed.&quot; If you see that, all is almost certainly well. If you don't, something absolutely is not, and you'll need to chase the problem down.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-status-openvpn-640x691.png\" width=\"640\" height=\"691\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-status-openvpn.png 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-status-openvpn.png\" class=\"enlarge-link\" data-height=\"879\" data-width=\"814\">Enlarge<\/a> \/ You get lots of chewy dialog here when the VPN client attempts to connect to the server. The money line here is \"Initialization Sequence Completed.\" If you see that, all is almost certainly well. If you don't, something absolutely is not, and you'll need to chase the problem down.<\/figcaption><\/figure><p>Finally, you may need to set some static leases for your media center devices. My Rokus don't allow static configuration within themselves as an option, so doing it on the router is a necessity. Hit the Services tab; under the \"Services\" sub-tab (the \"home screen\" of Services, where you should already be), you'll see a section for adding static leases, under \"DHCP Server.\" The UI here is a little funky: clicking the \"Add\" button makes a new blank row show up; it doesn't add something you've already entered in. Now that you've got a blank row, put the MAC address of your Roku or other dumb, needs-to-get-directly-to-the-Internet device in, specify an IP address for it that's below .63... and <em>don't<\/em> hit Add again. Instead, scroll all the way to the bottom of the page, where you'll find buttons to click to first Save, then Apply.<\/p>\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-static-lease.png\" class=\"enlarge\" data-height=\"659\" data-width=\"808\"><img alt=\"In the policy-based routing section, we told DD-WRT to send clients at 192.168.1.64\/26, 192.168.1.128\/26, and 192.168.1.192\/26 through the VPN instead of directly out to the Internet. Staticking my test server monolith here at 192.168.1.10\u2014within that 0-63 range we didn't cover in the policy routes\u2014means that monolith won't use the VPN at all. In the real world, you'd use this for Rokus or other media client devices that need to avoid Netflix\/Hulu\/SKY\/whatever anti-VPN restrictions.\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-static-lease-640x522.png\" width=\"640\" height=\"522\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-static-lease.png 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/ddwrt-static-lease.png\" class=\"enlarge-link\" data-height=\"659\" data-width=\"808\">Enlarge<\/a> \/ In the policy-based routing section, we told DD-WRT to send clients at 192.168.1.64\/26, 192.168.1.128\/26, and 192.168.1.192\/26 through the VPN instead of directly out to the Internet. Staticking my test server monolith here at 192.168.1.10\u2014within that 0-63 range we didn't cover in the policy routes\u2014means that monolith won't use the VPN at all. In the real world, you'd use this for Rokus or other media client devices that need to avoid Netflix\/Hulu\/SKY\/whatever anti-VPN restrictions.<\/figcaption><\/figure><p>Most of the steps above will be the same (or at least very similar) for <em>any<\/em> consumer router you've managed to shoehorn DD-WRT (or a DD-WRT variant, like Tomato) onto.<\/p>\n<h2>On consumer routers, price, and performance<\/h2>\n<p>If you're considering buying a consumer router specifically for running a VPN out of your network, the R8000 might or might not be the best fit for you. I'm going to stick to recommending Netgear Nighthawks no matter what for two reasons: Netgear actually <em>supports<\/em> the process of you replacing their OEM firmware with DD-WRT builds, and Netgear uses relatively high-powered ARM A9 multi-core CPUs in their Nighthawk series, where many consumer routers are using much, much weaker MIPS CPUs. This makes a <em>huge<\/em> impact on VPN throughput.<\/p>\n<p>The lowest-powered choice in this line is the Netgear R6400, but that model has a slightly slower CPU than the one in the R8000 I tested here (which could handle 25 Mbps throughput on AES-256-CBC\/SHA512\u00a0and 35+ Mbps on AES-256-CBC\/SHA1). The R8000 uses a 1 GHz ARM A9 CPU, where the R6400 uses an 800MHz ARM A9. I haven't directly tested it, but I would assume you're looking at roughly 80 percent of the performance of the higher-end part, maybe less. I don't recommend the R6400 currently even if you can live with the performance drop, because the <strong>R6700<\/strong> is the same price on Amazon ($110) and features the higher-performance 1 GHz CPU\u2014the exact same part as the R8000 I tested here.<\/p>\n<p>The R6700 and the R8000 share the same CPU and will perform equivalently for VPNs, but the R8000 is a tri-band router while the R6700 is only dual-band. Is the addition of a second 5GHz radio worth the additional $150, with the R8000 currently running $260 on Amazon? That depends on how many Wi-Fi devices you have in the house. If you've got a couple of TVs, several phones, a couple of laptops, and a tablet or two with several people who might be using a bunch of them simultaneously, the answer is probably a resounding \"yes.\" On the other hand, if you've got a high-quality <a href=\"https:\/\/arstechnica.com\/gadgets\/2017\/04\/send-wi-fi-companies-floor-plans-receive-the-ultimate-mesh-networking-test\/\">mesh network kit<\/a> like Orbi or Plume handling Wi-Fi duties, you don't need the R8000's radios <em>at all<\/em>, so you can get the R6700 to handle routing and VPN duties and let your mesh kit keep handling the Wi-Fi. (And before you ask, no, you can't just run DD-WRT on the mesh kit itself.)<\/p>\n<figure class=\"image shortcode-img center large\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/Amazon-R9000.jpg\" class=\"enlarge\" data-height=\"923\" data-width=\"1350\"><img alt=\"The Nighthawk R9000 offers twice the VPN throughput of the R6700... but at $450, it does so at just over four times the cost. (And our Homebrew still smacks its butt and sends it running home to mommy.)\" src=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/Amazon-R9000-640x438.jpg\" width=\"640\" height=\"438\" srcset=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/Amazon-R9000-1280x875.jpg 2x\" \/><\/a><figcaption class=\"caption\"><a href=\"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/04\/Amazon-R9000.jpg\" class=\"enlarge-link\" data-height=\"923\" data-width=\"1350\">Enlarge<\/a> \/ The Nighthawk R9000 offers twice the VPN throughput of the R6700... but at $450, it does so at just over four times the cost. (And our Homebrew still smacks its butt and sends it running home to mommy.)<a rel=\"nofollow\" class=\"caption-link\" href=\"https:\/\/www.amazon.com\/dp\/B01M12RE4A\/?tag=arstech20-20\">Amazon.com<\/a><\/figcaption><\/figure><p>Finally, if the sky's the limit, Netgear also offers a Nighthawk X10 (R9000) with a whopping 1.7 GHz <em>quad<\/em>-core ARM A9 CPU. I have not tested one yet, but if I had to hazard a guess, I'd expect nearly double the OpenVPN throughput that the R8000's 1 GHz dual-core ARM A9 managed (figure an estimated 45+ Mbps throughput on AES-256-CBC\/SHA512, or 65+ Mbps on AES-256-CBC\/SHA1). You'll also get MU-MIMO, quad-stream 2.4 GHz and 5 GHz radios, and where the R8000 had a second 5 GHz radio, the R9000 has an 802.11ad 60 GHz \"wigig\" radio instead. (You almost certainly don't have a wigig <em>client<\/em> device to connect to it with, but hey, you'll have the radio on your router.) All of this comes at a pretty whopping price tag, though: Amazon is currently listing the R9000 at $450.<\/p>\n<h2>On (in)security and performance<\/h2>\n<p>A final option, which I will go ahead and discuss but flat-out tell you <strong>we do not recommend<\/strong> is to sacrifice security for performance almost entirely. Weakening the encryption protocol and dropping the authentication protocol entirely\u2014AES-128-CBC\/None\u2014resulted in 51.25 Mbps throughput on my R8000. Finally, dropping both authentication <em>and<\/em> encryption (at which point, yes, you've <em>tunneled<\/em> your data but somebody who actually cared could still pick it apart) to None\/None will go wire speed on just about any Internet connection you throw it at.<\/p>\n<p>File this under \"I'd rather talk with you kids about it than have you learn it on the playground.\" Yes, these options are faster. And if all you're concerned about\u2014we truly mean <em>all<\/em> you're concerned about\u2014is throwing a monkey wrench into your ISP's very casual predatory snooping, they'll <em>probably<\/em> do the trick. An actual attacker will absolutely be able to unravel these \"fast\" tunnels and view or modify the data running down them, though, so really... <em>don't<\/em> do it. Or, at the very least, don't claim nobody told you it was a bad idea.<\/p>\n<h2>Conclusions<\/h2>\n<p>It's really <em>not<\/em> that hard to roll your own, personally hosted VPN service to get your data away from prying eyes at your ISP (or at the coffee shop; we didn't cover the minutiae of installation here, but you can use OpenVPN credentials on Android and iOS phones and tablets, too). <em>Extremely<\/em> heavy data users might have problems with bandwidth overage costs from their VPS provider, but the 1TB\/month allotment from DigitalOcean will easily cover my household usage. Whether you want to set the whole thing up on a router to blanket-cover your whole network, on individual devices, or both at once, you can get it done.<\/p>\n<h3>The Good:<\/h3>\n<ul><li>OpenVPN can be configured extremely securely, is free as in speech and as in beer, and can be run on just about any device you can think of: Windows, Macs, Linux or BSD machines, phones, tablets, and even (some) consumer routers.<\/li>\n<\/ul><h3>The Bad:<\/h3>\n<ul><li>At the end of the day, your insecure traffic is still <em>insecure\u2014<\/em>you've just moved your point of vulnerability, not eliminated it. You (understandably) didn't trust your ISP, so you moved it out of their reach. You (understandably) didn't trust VPN providers, so you didn't use them. But you're still trusting your hosting provider... <em>and<\/em> everybody <em>they're<\/em> downstream of.<\/li>\n<\/ul><h3>The Ugly:<\/h3>\n<ul><li>You now have one more machine to maintain. Your Ubuntu 16.04 LTS VM will automatically apply security upgrades, and it's supported through April 2021 (after which it will need an upgrade to a newer LTS version, which can generally be done relatively painlessly and in place for simple systems like this), but that's not forever. There's no guarantee that some new crypto breakthrough won't force you to reevaluate your cipher\/digest choices before then, either.<\/li>\n<\/ul>","created_at":"2018-11-07T16:48:54+0000","updated_at":"2018-11-07T16:48:54+0000","published_by":["Jim Salter"],"annotations":[],"mimetype":"text\/html","reading_time":46,"domain_name":"arstechnica.com","preview_picture":"https:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2017\/05\/vpn-privacy-760x380.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":460,"title":"Better Git configuration","url":"https:\/\/blog.scottnonnenberg.com\/better-git-configuration\/","content":"<p>I like <a href=\"https:\/\/git-scm.com\/\">Git<\/a>. I use it <a href=\"https:\/\/github.com\/scottnonnenberg\/thoughts-system#other-files\">all the time<\/a>. As <a href=\"https:\/\/blog.scottnonnenberg.com\/eslint-part-1-exploration\/#making-it-my-own\">I sometimes do<\/a>, I recently took some time to really dig in, read through documentation, and review my global Git configuration. Welcome to my fourth <a href=\"https:\/\/blog.scottnonnenberg.com\/tags\/stack-improvements\/\">stack improvements<\/a> post!<\/p><h2 id=\"its-all-git\">It\u2019s all Git<\/h2><p>I started coding in the <a href=\"https:\/\/blog.scottnonnenberg.com\/2017-twenty-years-online\/\">bad old days<\/a> of plain filesystem copies and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Microsoft_Visual_SourceSafe\">Visual SourceSafe<\/a> with its exclusive locks on checkout. Even so, back then the concept of source control was so amazing to me I wished I had access to it when coding at home.<\/p><p>Later, at <a href=\"http:\/\/www.calpoly.edu\/\">Cal Poly<\/a>, I was exposed to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Concurrent_Versions_System\">Concurrent Versions System (CVS)<\/a> as I worked on group projects. There were only so many group projects, so I never got very good at it.<\/p><p>During my <a href=\"https:\/\/www.microsoft.com\/\">Microsoft<\/a> years, I used <a href=\"https:\/\/en.wikipedia.org\/wiki\/Team_Foundation_Server\">Team Foundation Server<\/a> for source control during what we called \u201cApp Week,\u201d where non-developers in Visual Studio would spend a whole week <a href=\"https:\/\/en.wikipedia.org\/wiki\/Eating_your_own_dog_food\">dogfooding<\/a> the product to make sure it was ready to ship. But all of my personal programming during that time was with <a href=\"https:\/\/en.wikipedia.org\/wiki\/Apache_Subversion\">Subversion<\/a>. It was free and easy to run locally. I used it to keep track of all my local changes over time.<\/p><p>Fall of 2010 was when things changed. I learned <a href=\"http:\/\/rubyonrails.org\/\">Ruby on Rails<\/a> to develop <a href=\"https:\/\/scottnonnenberg.com\/work\/#stark-raving-bits-2010-q-3-to-2011-q-2\">a side project<\/a>, and the comprehensive tutorial I went through covered <a href=\"https:\/\/www.heroku.com\/\">Heroku<\/a> and a new source control system: <strong>Git<\/strong>. It was amazing - I could treat it like it was hosted locally, but also interact with others. No exclusive locking, full productivity when offline, and intelligent merging. I was hooked.<\/p><p>Since then <a href=\"https:\/\/rhodecode.com\/insights\/version-control-systems-2016\">Git has truly taken off<\/a>. It is the de facto standard for open source. It is supported by every major player in cloud source code hosting. It is supported by many GUI tools - dedicated source control tools, as well as code editors.<\/p><p>It\u2019s important to know, and to know well.<\/p><h2 id=\"global-config\">Global config<\/h2><p>Whether you know it or not, you\u2019ve got global Git configuration. A <code>.gitconfig<\/code> file in your home directory. Most <code>.gitconfig<\/code> files have your name and email, as established by your average <a href=\"https:\/\/help.github.com\/articles\/set-up-git\/\">\u2018getting started with Git\u2019 tutorial<\/a>. But there\u2019s <a href=\"https:\/\/git-scm.com\/docs\/git-config#_variables\">so much more configuration available<\/a> for that file!<\/p><p>My <a href=\"https:\/\/gist.github.com\/scottnonnenberg\/fefa3f65fdb3715d25882f3023b31c29\">entire global <code>.gitconfig<\/code> is available via a gist<\/a> with inline comments. Jump directly there if you like, but here we\u2019ll talk about each of the sections in a bit more detail.<\/p><h3 id=\"alias\">Alias<\/h3><p>The bread and butter of a custom <code>.gitconfig<\/code> is the <a href=\"https:\/\/git-scm.com\/docs\/git-config#git-config-alias\"><code>[alias]<\/code> section<\/a>, where you can create your own commands. Feel like something is missing? Add it here! Something doesn\u2019t quite work the way you want? Add your own version of it!<\/p><ul><li><code>prune = fetch --prune<\/code> - When working on a project where others push branches to the main repository, I end up with lots of random local branches. <a href=\"https:\/\/git-scm.com\/docs\/git-fetch#git-fetch--p\">Prune<\/a> deletes any local branch which has been deleted from the remote. It\u2019s here because I always forget about it.<\/li>\n<li><code>undo = reset --soft HEAD^<\/code> - If I\u2019ve made a mistake in making a commit, <a href=\"https:\/\/git-scm.com\/docs\/git-reset\">this command<\/a> sets things up the way they were before the commit. To be fair, usually I just <a href=\"https:\/\/git-scm.com\/docs\/git-commit#git-commit---amend\">amend<\/a> the existing commit in this situation, since that keeps the commit message.<\/li>\n<li><code>stash-all = stash save --include-untracked<\/code> - <a href=\"https:\/\/git-scm.com\/docs\/git-stash\">Stash<\/a> is extremely useful when someone randomly asks you to check out another branch, but you\u2019re right in the middle of something. This command ensures that when you stash, you catch the new files you haven\u2019t caught with a <code>git add<\/code> yet.<\/li>\n<li><code>glog = log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)&lt;%an&gt;%Creset'<\/code> - Commit history shown via the <a href=\"https:\/\/git-scm.com\/docs\/git-log\">default <code>git log<\/code><\/a> is space-inefficient, and doesn\u2019t really focus on the most important information. This colorful, graphical display is easier to parse, especially when branching gets complex.<\/li>\n<\/ul><h3 id=\"merge\">Merge<\/h3><p><code>ff = only<\/code> ensures that you get an error unless every merge is <a href=\"https:\/\/sandofsky.com\/images\/fast_forward.pdf\">fast-forward<\/a>. No merge commits, no joining together of two histories, just a smooth progression from commit to commit.<\/p><p>You might be wondering how you\u2019d get real work done doing this! The answer is <a href=\"https:\/\/git-scm.com\/docs\/git-rebase\"><code>git rebase<\/code><\/a>, a way to take a branch of the commit tree and attach it to a different part, giving it a new \u2018base.\u2019 <a href=\"https:\/\/nathanleclaire.com\/blog\/2014\/09\/14\/dont-be-scared-of-git-rebase\/\">It\u2019s an extremely useful tool<\/a>.<\/p><p>It\u2019s how I update a pull request when it has conflicts with <code>master<\/code>. Essentially I get a chance to re-apply all of my commits in a future world, where others have made changes. I think it\u2019s a better metaphor than \u201cmerge everyone else\u2019s changes into my changes,\u201d which is what you get if you merge the latest <code>master<\/code> into your branch.<\/p><p>This option is great because I will never accidentally create a merge commit. If I intend to create a merge commit I can force that behavior with an explicit <a href=\"https:\/\/git-scm.com\/docs\/git-merge#git-merge---ff\"><code>git merge --ff<\/code><\/a>.<\/p><p><code>conflictstyle = diff3<\/code> gives you a little more information when a merge conflict happens. Normally you get two sections - the intended changes from the \u2018left\u2019 and the intended changes from the \u2018right.\u2019 With <a href=\"https:\/\/git-scm.com\/docs\/git-config#git-config-mergeconflictStyle\">this option<\/a> you get a third section, the original changes before \u2018left\u2019 and \u2018right\u2019 tried to change it.<\/p><h3 id=\"commit\">Commit<\/h3><p><code>gpgSign = true<\/code> ensures that all of your <a href=\"https:\/\/git-scm.com\/docs\/git-commit#git-commit--Sltkeyidgt\">commits are signed<\/a> by your <a href=\"https:\/\/www.gnupg.org\/\">GPG<\/a> key. This is generally a good idea because there\u2019s no verification of the <a href=\"https:\/\/git-scm.com\/docs\/git-config#git-config-useremail\">user information in your <code>.gitconfig<\/code> file<\/a>, which means that commits that look like they are from you could easily show up in someone else\u2019s pull requests.<\/p><p>In fact, for a contract I once had to use someone else\u2019s credentials because account and machine provisioning were taking too long. My pull requests were submitted through someone else\u2019s account, but all the commits inside had my real name on them.<\/p><p>You can remove all question about where a commit came from by <a href=\"https:\/\/help.github.com\/articles\/adding-a-new-gpg-key-to-your-github-account\/\">adding your public GPG key to your GitHub account<\/a>. Your signed commits will have <a href=\"https:\/\/github.com\/scottnonnenberg\/eslint-config-thehelp\/pull\/5\/commits\">a little \u201cVerified\u201d badge on them<\/a>.<\/p><p><em>Note:<\/em><\/p><ul><li>If you have more than one GPG key, you can specify which you\u2019d like to use with the <a href=\"https:\/\/git-scm.com\/docs\/git-config#git-config-usersigningKey\"><code>user.signingKey<\/code> option<\/a>.<\/li>\n<li>GUI apps won\u2019t necessarily sign your commits even with this option set. You\u2019ll need to explore the application\u2019s options.<\/li>\n<li><code>gpg-agent<\/code> can save your passphrase for a time, making things a little easier. Use it!<\/li>\n<\/ul><h3 id=\"push\">Push<\/h3><p><code>default = simple<\/code> is <a href=\"https:\/\/git-scm.com\/docs\/git-config#git-config-pushdefault\">an option<\/a> you probably already have set. It makes it easier to push your local branch to a branch named the same thing in your target remote.<\/p><p><code>followTags = true<\/code> is quite simple. Instead of manually pushing new tags with <code>--follow-tags<\/code>, you always send your local tags up along with a <code>git push<\/code>. I don\u2019t know about you, but once I have tags created locally, I want them included when I push.<\/p><h3 id=\"status\">Status<\/h3><p><code>showUntrackedFiles = all<\/code> Normally, when you\u2019ve added a new directory, but haven\u2019t staged it yet with <code>git add<\/code>, your <code>git status<\/code> will just show the directory name. This has tripped me up quite a few times, since a new, large directory shows up as just one line. <a href=\"https:\/\/git-scm.com\/docs\/git-config#git-config-statusshowUntrackedFiles\">This option<\/a> shows you all of the files underneath that new directory during a <code>git status<\/code>.<\/p><p><em>Note:<\/em> This can make things slower in very large repositories.<\/p><h3 id=\"transfer\">Transfer<\/h3><p><code>fsckobjects = true<\/code> tells Git that you\u2019d like it to do some extra checks when receiving or sending changes. Why do extra checks? You want to be alerted to evidence of data corruption sooner rather than later, right?<\/p><p><em>Note:<\/em> This can make transfers a bit slower.<\/p><p>In addition to the built-in <code>git diff<\/code> command, Git allows you to specify <a href=\"https:\/\/git-scm.com\/docs\/git-difftool\">an external tool to visualize your diffs<\/a>. This collection of entries sets Git up to use <a href=\"https:\/\/github.com\/jeffkaufman\/icdiff\"><code>icdiff<\/code><\/a> to display the differences between two states of your repository:<\/p><pre class=\"language-text\">[diff]\n  tool = icdiff\n[difftool]\n  prompt = false\n[difftool \"icdiff\"]\n  cmd = \/usr\/local\/bin\/icdiff --line-numbers $LOCAL $REMOTE\n<\/pre><p>You can use it just like normal: <code>git difftool master branch<\/code><\/p><p><code>icdiff<\/code> is interesting in that it attempts to replicate colorful, GitHub-style, split diffs right in your console. <a href=\"https:\/\/blog.scottnonnenberg.com\/top-ten-pull-request-review-mistakes\/#3-unified-diffs\">A little easier to read<\/a> than the normal chunk-based diff style.<\/p><p><em>Note:<\/em><\/p><ul><li>You might have some difficulty installing <code>icdiff<\/code>. Happily, <a href=\"https:\/\/github.com\/jeffkaufman\/icdiff\/issues\/72\">there is a simple workaround<\/a>.<\/li>\n<li>Keep <code>git diff<\/code> in your back pocket - <code>icdiff<\/code> doesn\u2019t seem to handle comparisons against <code>\/dev\/null<\/code>. For example, try <code>git difftool --cached<\/code> after you\u2019ve staged a new file.<\/li>\n<\/ul><h2 id=\"bonus-more-revisions\">Bonus: More revisions!<\/h2><p>You type <code>git checkout master<\/code> all the time, right? In that command, <code>master<\/code> is an example of a <em><a href=\"https:\/\/git-scm.com\/docs\/gitrevisions\">revision<\/a><\/em>, specifically a shorthand referring to the latest commit in the <code>master<\/code> branch. These are the common <em>revision<\/em> formats:<\/p><pre class=\"language-bash\"># Check out a branch\ngit checkout branchname\n# Check out a remote branch\ngit checkout remotes\/origin\/branchname\n# Check out a specific commit\ngit checkout 158e4ef8409a7f115250309e1234567a44341404\n# Check out most recent commit for current branch\ngit checkout HEAD\n<\/pre><p>But it turns out that there\u2019s a whole language to specify <em>revisions<\/em>. All of these operators apply to any of the nouns used above:<\/p><pre class=\"language-bash\"># ^ means 'first parent commit,' therefore the second-most recent commit in the master branch\ngit checkout master^\n# If it's a merge commit, with more than one parent, this gets the second parent\ngit checkout master^2\n# Same thing as three ^ characters - three 'first-parent' steps up the tree\ngit checkout master~3\n# The first commit prior to a day ago\ngit checkout master@{yesterday}\n# The first commit prior to 5 minutes ago\ngit checkout master@{5.minutes.ago}\n<\/pre><p>You can find the whole set of supported <em>revision<\/em> formats here: <a href=\"https:\/\/git-scm.com\/docs\/gitrevisions\">https:\/\/git-scm.com\/docs\/gitrevisions<\/a>. I was surprised how comprehensive it is!<\/p><p>Remember, you can use a <em>revision<\/em> with most Git commands. Try this: <code>git glog master@{10.days.ago}..master<\/code><\/p><h2 id=\"go-forth\">Go forth!<\/h2><p>Start with my <code>.gitconfig<\/code> or just use it as an inspiration for your own customizations. You could even do a deep-dive like I did - the <a href=\"https:\/\/git-scm.com\/docs\">command list<\/a> and the <a href=\"https:\/\/git-scm.com\/docs\/git-config#_variables\">list of options<\/a> are extremely educational.<\/p><p>Make Git maximally useful to you. Comfort with Git leaves more room for the hard bugs and the actual merge conflicts. Dig in!<\/p><hr \/><p><em>Some good references:<\/em><\/p><ul><li>Many <a href=\"https:\/\/git-scm.com\/docs\/git-config#_variables\">configuration options<\/a> are better-suited to a specific project. Use the same format inside of your project-specific configuration file: <code>project-dir\/.git\/config<\/code><\/li>\n<li>Get up to date on all the <code>git log<\/code> options: <a href=\"https:\/\/git-scm.com\/docs\/git-log#_options\">https:\/\/git-scm.com\/docs\/git-log#_options<\/a><\/li>\n<li>Dig into the various Git revision formats: <a href=\"https:\/\/git-scm.com\/docs\/gitrevisions\">https:\/\/git-scm.com\/docs\/gitrevisions<\/a><\/li>\n<li>Learn about signing commits: <a href=\"https:\/\/mikegerwitz.com\/papers\/git-horror-story\">https:\/\/mikegerwitz.com\/papers\/git-horror-story<\/a><\/li>\n<li>Customize aliases for fun and profit: <a href=\"https:\/\/hackernoon.com\/lesser-known-git-commands-151a1918a60\">https:\/\/hackernoon.com\/lesser-known-git-commands-151a1918a60<\/a><\/li>\n<li>More about git repository corruption:\n<ul><li><a href=\"https:\/\/groups.google.com\/forum\/#!topic\/binary-transparency\/f-BI4o8HZW0\">https:\/\/groups.google.com\/forum\/#!topic\/binary-transparency\/f-BI4o8HZW0<\/a><\/li>\n<li><a href=\"http:\/\/git.661346.n2.nabble.com\/propagating-repo-corruption-across-clone-td7580504i40.html\">http:\/\/git.661346.n2.nabble.com\/propagating-repo-corruption-across-clone-td7580504i40.html<\/a><\/li>\n<\/ul><\/li>\n<li>Frustrated with Git\u2019s complexity? You\u2019re not alone: <a href=\"https:\/\/stevebennett.me\/2012\/02\/24\/10-things-i-hate-about-git\/\">https:\/\/stevebennett.me\/2012\/02\/24\/10-things-i-hate-about-git\/<\/a><\/li>\n<\/ul>","created_at":"2018-11-07T16:48:51+0000","updated_at":"2018-11-07T16:48:51+0000","published_at":"2017-04-05T20:25:28+0000","published_by":["Scott Nonnenberg"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"blog.scottnonnenberg.com","preview_picture":"https:\/\/static.sinap.ps\/img\/profile\/me_2012_square_300px_full.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":459,"title":"r\/gamedev - Resources for Procedural Interactive Fiction Generation?","url":"https:\/\/www.reddit.com\/r\/gamedev\/comments\/11bs45\/resources_for_procedural_interactive_fiction\/","content":"<h2>Posting Guidelines v3.5<\/h2><p><a href=\"http:\/\/reddit.com\/r\/gamedev\">\/r\/gamedev<\/a> is a game development community for developer-oriented content. We hope to promote discussion and a sense of community among game developers on reddit.<\/p><h2>Off Topic<\/h2><p><strong>Feedback requests<\/strong> \/ <strong>\"Play my game\"<\/strong><br \/>Post an <a href=\"https:\/\/www.reddit.com\/r\/gamedev\/wiki\/good_posts\">article<\/a> about your game or use the <a href=\"https:\/\/www.reddit.com\/r\/gamedev\/comments\/7fpqqu\/weekly_threads_101_making_good_use_of_rgamedev\/\">weekly threads<\/a> to trade feedback<\/p><p><strong>Job Offers, Recruiting, and related activities<\/strong><br \/>Use <a href=\"https:\/\/www.reddit.com\/r\/gamedevclassifieds\">\/r\/gamedevclassifieds<\/a> and <a href=\"https:\/\/www.reddit.com\/r\/INAT\">\/r\/INAT<\/a> for that<\/p><p><strong>Promotion<\/strong><br \/>Once-per-game feedback requests or release threads are OK, but a considerable history of participation on <a href=\"https:\/\/www.reddit.com\/r\/gamedev\">\/r\/gamedev<\/a> is required. Must be a Text Post.<\/p><p><strong>Devlogs<\/strong> that do not have a focus on being useful to other developers. Do not talk about what advancements occurred on your game this week. That is what Feedback Friday, Screenshot Saturday, <a href=\"https:\/\/www.reddit.com\/r\/devblogs\">\/r\/devblogs<\/a>, and <a href=\"https:\/\/www.reddit.com\/r\/gamedevscreens\">\/r\/gamedevscreens<\/a> are for.<\/p><p><strong>Memes<\/strong><\/p><h2>Explicitly On Topic<\/h2><p><strong>Gamedev-relevant Articles, Videos, and Announcements<\/strong><\/p><p><strong>Free Assets<\/strong> (please specify license)<\/p><p>09\/2017: Posting on-sale assets is no longer allowed<\/p><p><strong>Language\/Framework discussions<\/strong><br \/>Be sure to check the <a href=\"https:\/\/www.reddit.com\/r\/gamedev\/wiki\/faq\">FAQ<\/a>.<\/p><p><strong>AMAs<\/strong>\nIf you have a unique perspective on something, we'd love to hear it. This is not a way to promote your game\/kickstarter\/etc, please see above. The focus should be on providing info to the community, not promoting yourself. Be sure to include your education and years of experience to provide some context.<\/p><h2>Restrictions<\/h2><p><strong>Do not use [tags]<\/strong>, assign flair to your post after it's created.<\/p><p><strong>Question posts...<\/strong><br \/>should include what you've already tried and why it was inadequate. Be sure to check the <a href=\"https:\/\/www.reddit.com\/r\/gamedev\/wiki\/faq\">FAQ<\/a>.<\/p><p><strong>Minimum Text Submission Length<\/strong><br \/>40 words or so. That's about two tweets.<\/p><p><strong>Surveys and polls...<\/strong><br \/>should have their results shared with us.<\/p><p><strong>Shared Assets...<\/strong><br \/>should have a proper license included in the post itself.<br \/><em>Please include images\/samples in your post!<\/em><\/p><p><strong>\"Share Your Stuff\" threads...<\/strong><br \/>should have the OP posting in the comments alongside everyone else, if they post.<\/p>","created_at":"2018-11-07T16:48:49+0000","updated_at":"2018-11-07T16:48:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.reddit.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":458,"title":"Install Arch Linux","url":"https:\/\/www.wikihow.com\/Install-Arch-Linux","content":"<div id=\"intro\" class=\"section  sticky\">\n<h1 class=\"firstHeading\"><a href=\"https:\/\/www.wikihow.com\/Install-Arch-Linux\">How to Install Arch Linux<\/a><\/h1>\n<p id=\"method_toc\" class=\"sp_method_toc\">In this Article:<a id=\"summary_toc\" href=\"#\">Article Summary<\/a><a href=\"#Booting_into_the_Installer_sub\">Booting into the Installer<\/a><a href=\"#Creating_Partitions_sub\">Creating Partitions<\/a><a href=\"#Installing_sub\">Installing<\/a><a id=\"qa_toc\" href=\"#Questions_and_Answers_sub\">Community Q&amp;A<\/a><\/p>\n<p>This wikiHow teaches you how to replace your computer's current operating system with Arch Linux, which is an advanced version of Linux. You can install Arch Linux on both Windows and Mac computers.<\/p>\n<div class=\"clearall\">\n\n<\/div>\n<h2 class=\"hidden\">Steps<\/h2>\n<div class=\"section steps   sticky  steps_first\">\n<h3>\nPart 1\n<\/h3><\/div>\nBooting into the Installer\n\n<div id=\"steps_1\" class=\"section_text\">\n<ol class=\"steps_list_2\"><li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-1-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-1-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 1\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_8d713b939b\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/a\/ae\/Install-Arch-Linux-Step-1-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-1-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 1\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/a\/ae\/Install-Arch-Linux-Step-1-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-1-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript>\n<\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 1\">1<\/div>\n<div class=\"step\">\n<b class=\"whb\"><a href=\"https:\/\/www.wikihow.com\/Back-Up-Data\" title=\"Back Up Data\">Back up your computer<\/a> to an external hard drive.<\/b> You'll be erasing your current operating system, so make sure that you have a backup of anything that you want to keep before continuing.\n\n\n<\/div>\n<div class=\"clearall\">\n<div class=\"wh_ad_inner step_ad\" data-service=\"adsense\" data-loaded=\"1\" data-lateload=\"0\" data-adsensewidth=\"728\" data-adsenseheight=\"90\" data-slot=\"5875012246\" data-channels=\"8752286177\" data-viewablerefresh=\"1\" data-aps-timeout=\"2000\">\n<ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-9543332082073187\" data-ad-slot=\"5875012246\">\n<\/ins><\/div>\n<div class=\"clearall adclear\">\n<\/div><\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-2-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-2-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 2\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_dd04b700b8\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/b5\/Install-Arch-Linux-Step-2-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-2-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 2\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/b5\/Install-Arch-Linux-Step-2-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-2-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 2\">2<\/div>\n<div class=\"step\">\n<b class=\"whb\">Download the Arch installation image.<\/b> Arch Linux can be downloaded in ISO format, from which point you can burn it onto a blank DVD in order to run the image on your computer. To download this image:\n\n<ul><li>Make sure that you have either BitTorrent or <a href=\"https:\/\/www.wikihow.com\/Install-uTorrent\" title=\"Install uTorrent\">uTorrent<\/a> installed.<\/li>\n<li>Go to <a rel=\"nofollow\" class=\"external free\" href=\"https:\/\/www.archlinux.org\/download\/\">https:\/\/www.archlinux.org\/download\/<\/a> in your browser.<\/li>\n<li>Click the <b>Torrent<\/b> link below the \"BitTorrent\" heading.<\/li>\n<li>Open the downloaded torrent in BitTorrent or uTorrent.<\/li>\n<li>Wait for the Arch Linux torrent file to finish downloading.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-3-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-3-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 3\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_0e4c06ac36\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/b9\/Install-Arch-Linux-Step-3-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-3-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 3\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/b9\/Install-Arch-Linux-Step-3-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-3-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 3\">3<\/div>\n<div class=\"step\">\n<b class=\"whb\"><a href=\"https:\/\/www.wikihow.com\/Burn-ISO-Files-to-DVD\" title=\"Burn ISO Files to DVD\">Burn the image onto a blank DVD<\/a>.<\/b> Once your Arch Linux ISO file finishes downloading via your torrent client, you'll burn it onto a blank DVD using your computer's DVD drive. Once the burning is complete, leave the DVD in your computer.\n\n<ul><li>If your computer doesn't have a DVD drive, you'll need to buy an external DVD drive and attach it to your computer via USB cable.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-4-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-4-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 4\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_217377ecce\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/3c\/Install-Arch-Linux-Step-4-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-4-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 4\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/3c\/Install-Arch-Linux-Step-4-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-4-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 4\">4<\/div>\n<div class=\"step\">\n<b class=\"whb\">Restart your computer.<\/b> Click <b>Start<\/b>\n<div class=\"mwimg  t techicon floatright\">\n<noscript><img alt=\"Image titled Windowsstart.png\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/0\/07\/Windowsstart.png\/30px-Windowsstart.png\" width=\"30\" height=\"30\" class=\"whcdn\" \/><\/noscript>\n<\/div>\n, click <b>Power<\/b>\n<div class=\"mwimg  t techicon floatright\">\n<noscript><img alt=\"Image titled Windowspower.png\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/5\/5d\/Windowspower.png\/30px-Windowspower.png\" width=\"30\" height=\"30\" class=\"whcdn\" \/><\/noscript>\n<\/div>\n, and click <b>Restart<\/b> in the menu.\n\n<ul><li>On a Mac, you'll instead click the <b>Apple menu<\/b>\n<div class=\"mwimg  t techicon floatright\">\n<noscript><img alt=\"Image titled Macapple1.png\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/ff\/Macapple1.png\/29px-Macapple1.png\" width=\"29\" height=\"30\" class=\"whcdn\" \/><\/noscript>\n<\/div>\n, click <b>System Preferences...<\/b>, click <b>Startup<\/b>, click <b>External Drive<\/b>, and then restart by clicking <b>Restart...<\/b> in the Apple menu and clicking <b>Restart<\/b> when prompted.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-5-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-5-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 5\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_d7240ebbbe\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/5\/59\/Install-Arch-Linux-Step-5-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-5-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 5\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/5\/59\/Install-Arch-Linux-Step-5-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-5-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 5\">5<\/div>\n<div class=\"step\">\n<b class=\"whb\">Press the key that allows you to change the boot order.<\/b> On most newer computers, this is <kbd class=\"keyboard-key nowrap\">F12<\/kbd>, though the exact key should be displayed on the screen during boot up. If there is no key to change boot order, press the BIOS Setup key (usually <kbd class=\"keyboard-key nowrap\">F1<\/kbd>, <kbd class=\"keyboard-key nowrap\">F2<\/kbd>, <kbd class=\"keyboard-key nowrap\">F10<\/kbd>, or <kbd class=\"keyboard-key nowrap\">Del<\/kbd>).\n\n<ul><li>Skip this step on a Mac.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-6-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-6-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 6\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_e6a4bf4f2c\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4c\/Install-Arch-Linux-Step-6-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-6-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 6\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4c\/Install-Arch-Linux-Step-6-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-6-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 6\">6<\/div>\n<div class=\"step\">\n<b class=\"whb\">Select your installation drive as the primary boot drive.<\/b> Set the drive (e.g., \"DVD Drive\" or \"Disk Drive\") that contains your Arch Linux DVD as the primary drive by selecting it and pressing the <kbd class=\"keyboard-key nowrap\">+<\/kbd> key until it's at the top of the menu.\n\n<ul><li>Skip this step on a Mac.<\/li>\n<li>On some PCs, you may first have to open an \"Advanced\" tab or select a \"Boot Options\" section.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-7-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-7-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 7\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_c7e84ff29a\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/94\/Install-Arch-Linux-Step-7-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-7-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 7\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/94\/Install-Arch-Linux-Step-7-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-7-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 7\">7<\/div>\n<div class=\"step\">\n<b class=\"whb\">Save and exit the \"Boot Options\" screen.<\/b> You should see a key at the bottom or in the bottom-right corner of the screen that you can press to save and exit. After doing so, your computer will resume restarting.\n\n<ul><li>Skip this step on a Mac.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-8-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-8-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 8\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_4f53b0e14c\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/94\/Install-Arch-Linux-Step-8-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-8-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 8\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/94\/Install-Arch-Linux-Step-8-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-8-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 8\">8<\/div>\n<div class=\"step\">\n<b class=\"whb\">Select Boot Arch Linux and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/b> This will launch the Arch Linux installer, at which point you can proceed with partitioning your hard drive.\n\n<\/div>\n<div class=\"clearall\">\n<div class=\"wh_ad_inner step_ad\" data-service=\"dfp\" data-adtargetid=\"methodad\" data-loaded=\"0\" data-lateload=\"0\" data-adsensewidth=\"728\" data-adsenseheight=\"90\" data-channels=\"8752286177\" data-viewablerefresh=\"1\" data-apsload=\"1\" data-aps-timeout=\"2000\">\n<div id=\"methodad\">\n\n<\/div>\n<div class=\"clearall adclear\">\n\n<\/div><\/div><\/div><\/li>\n<\/ol><div class=\"clearall\">\n<\/div>\n<\/div>\n<div class=\"section steps   sticky\">\n<h3>\nPart 2\n<\/h3><\/div>\nCreating Partitions\n\n<div id=\"steps_2\" class=\"section_text\">\n<ol class=\"steps_list_2\"><li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-9-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-9-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 9\" width=\"728\" height=\"547\" class=\"whcdn content-fill\" id=\"img_2a3b2b4148\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/c\/c8\/Install-Arch-Linux-Step-9-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-9-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 9\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/c\/c8\/Install-Arch-Linux-Step-9-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-9-Version-4.jpg\" width=\"728\" height=\"547\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 1\">1<\/div>\n<div class=\"step\">\n<b class=\"whb\">Check your existing drives.<\/b> You'll have at least two drives available: your computer's hard drive, and the Arch Linux installation media. To check your current drives:\n\n<ul><li>Type in <kbd>fdisk -l<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Find the name of the largest hard drive on the results screen. The name will be something like \"\/dev\/sda\" and can be found to the right of the \"Disk\" heading.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-10-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-10-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 10\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_24764c3f50\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/65\/Install-Arch-Linux-Step-10-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-10-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 10\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/65\/Install-Arch-Linux-Step-10-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-10-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 2\">2<\/div>\n<div class=\"step\">\n<b class=\"whb\">Bring up the partition page.<\/b> Type in <kbd>cfdisk [drive name]<\/kbd>, making sure to replace <b>[drive name]<\/b> with your hard drive's name, and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>, then select <b>DOS<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> again.\n\n<ul><li>For example: if the drive is named \"\/dev\/sda\", you would type <kbd>cfdisk \/dev\/sda<\/kbd> into the terminal.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-11-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-11-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 11\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_3326b27c88\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/bc\/Install-Arch-Linux-Step-11-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-11-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 11\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/bc\/Install-Arch-Linux-Step-11-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-11-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 3\">3<\/div>\n<div class=\"step\">\n<b class=\"whb\">Delete the hard drive's contents.<\/b> Select a partition in the middle of the screen, select <b>Delete<\/b> at the bottom of the screen, press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>, and repeat for any other partitions in the middle of the screen. You should ultimately be left with a line entitled <kbd>Pri\/Log Free Space<\/kbd>.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-12-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-12-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 12\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_b8888b0ddb\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/67\/Install-Arch-Linux-Step-12-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-12-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 12\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/67\/Install-Arch-Linux-Step-12-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-12-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 4\">4<\/div>\n<div class=\"step\">\n<b class=\"whb\">Create a \"swap\" partition.<\/b> This partition will function as Arch Linux's RAM. To do so:\n\n<ul><li>Select <b>New<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Select <b>Primary<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in a number of megabytes (e.g., 1024 for one gigabyte) and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>. As a rule of thumb, you should make the swap partition 2 or 3 times your RAM (e.g., if you have 4 gigabytes of RAM, you should make the swap partition 8192 or 12288 megabytes).<\/li>\n<li>Select <b>End<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-13-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-13-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 13\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_9b319029df\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/2\/26\/Install-Arch-Linux-Step-13-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-13-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 13\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/2\/26\/Install-Arch-Linux-Step-13-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-13-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 5\">5<\/div>\n<div class=\"step\">\n<b class=\"whb\">Create your main hard drive partition.<\/b> This is the partition on which your Arch Linux operating system, files, and other information will be stored. To do this:\n\n<ul><li>Make sure that the <b>Pri\/Log Free Space<\/b> partition is selected.<\/li>\n<li>Select <b>New<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Select <b>Primary<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Make sure that the number next to the \"Size (in MB)\" heading is correct.<\/li>\n<li>Press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Select the primary partition again.<\/li>\n<li>Select <b>Bootable<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-14-Version-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-14-Version-4.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 14\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_e6677708ea\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/7\/79\/Install-Arch-Linux-Step-14-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-14-Version-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 14\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/7\/79\/Install-Arch-Linux-Step-14-Version-4.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-14-Version-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 6\">6<\/div>\n<div class=\"step\">\n<b class=\"whb\">Label the \"swap\" partition.<\/b> This will set it as your system's RAM:\n\n<ul><li>Select the \"swap\" partition.<\/li>\n<li>Select <b>Type<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in <kbd>82<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>With the \"swap\" partition still selected, select <b>Write<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in <kbd>yes<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-15-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-15-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 15\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_0fc0cd59eb\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/96\/Install-Arch-Linux-Step-15-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-15-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 15\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/96\/Install-Arch-Linux-Step-15-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-15-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 7\">7<\/div>\n<div class=\"step\">\n<b class=\"whb\">Note the names of your partitions.<\/b> In the \"Name\" column on the far-left side of the screen, you should see a name (e.g., \"sda1\") next to your \"swap\" partition, and a similar name (e.g., \"sda2\") next to the primary partition. You'll need both of these names in order to format your partitions.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-16-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-16-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 16\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_84ff201fd6\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/0\/0d\/Install-Arch-Linux-Step-16-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-16-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 16\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/0\/0d\/Install-Arch-Linux-Step-16-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-16-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 8\">8<\/div>\n<div class=\"step\">\n<b class=\"whb\">Exit the \"cfdisk\" utility.<\/b> Select <b>Quit<\/b> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> to do so.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-17-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-17-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 17\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_8aa7e08c4e\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/1\/18\/Install-Arch-Linux-Step-17-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-17-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 17\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/1\/18\/Install-Arch-Linux-Step-17-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-17-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 9\">9<\/div>\n<div class=\"step\">\n<b class=\"whb\">Format your primary partition.<\/b> This will make it usable with your operating system. To do this, type in <kbd>mkfs.ext4 \/dev\/[primary partition name]<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.\n\n<ul><li>For a partition named \"sda2\", you'd enter <kbd>mkfs.ext4 \/dev\/sda2<\/kbd> here.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-18-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-18-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 18\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_e59f6416b9\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/9b\/Install-Arch-Linux-Step-18-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-18-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 18\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/9b\/Install-Arch-Linux-Step-18-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-18-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 10\">10<\/div>\n<div class=\"step\">\n<b class=\"whb\">Mount the formatted partition.<\/b> Type in <kbd>mount \/dev\/[partition name] \/mnt<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>. This will make the partition into a usable drive.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-19-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-19-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 19\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_857bf7aefd\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/2\/27\/Install-Arch-Linux-Step-19-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-19-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 19\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/2\/27\/Install-Arch-Linux-Step-19-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-19-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 11\">11<\/div>\n<div class=\"step\">\n<b class=\"whb\">Add a swap file to the \"swap\" partition.<\/b> Type in <kbd>mkswap \/dev\/[partition name]<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>, then type in <kbd>swapon \/dev\/sda1<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> again. After completing this step, you can proceed with installing Arch Linux.\n\n<ul><li>For a \"swap\" partition named \"sda1\", for example, you'd enter <kbd>mkswap \/dev\/sda1<\/kbd> and then <kbd>swapon \/dev\/sda1<\/kbd> here.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<div class=\"wh_ad_inner step_ad\" data-service=\"adsense\" data-loaded=\"1\" data-lateload=\"0\" data-adsensewidth=\"728\" data-adsenseheight=\"90\" data-slot=\"5875012246\" data-channels=\"8752286177\" data-viewablerefresh=\"1\" data-aps-timeout=\"2000\">\n<ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-9543332082073187\" data-ad-slot=\"5875012246\">\n<\/ins><\/div>\n<div class=\"clearall adclear\">\n<\/div><\/div><\/li>\n<\/ol><div class=\"clearall\">\n<\/div>\n<\/div>\n<div class=\"section steps   sticky\">\n<h3>\nPart 3\n<\/h3><\/div>\nInstalling\n\n<div id=\"steps_3\" class=\"section_text\">\n<ol class=\"steps_list_2\"><li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-20-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-20-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 20\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_4bf13723ab\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/88\/Install-Arch-Linux-Step-20-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-20-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 20\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/88\/Install-Arch-Linux-Step-20-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-20-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 1\">1<\/div>\n<div class=\"step\">\n<b class=\"whb\">Set up a Wi-Fi connection.<\/b> If your computer is plugged into your router via ethernet, you can skip this step. <a href=\"https:\/\/www.wikihow.com\/Set-up-Ethernet\" title=\"Set up Ethernet\">Using ethernet<\/a> is preferable to using Wi-Fi.<sup id=\"_ref-1\" class=\"reference\" aria-label=\"Link to Reference 1\"><a href=\"#_note-1\">[1]<\/a><\/sup><ul><li>Type <kbd>ip link<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> to determine the interface name of your network adapter.<\/li>\n<li>Type <kbd>pacman -S iw wpa_supplicant<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> to install the necessary software.<\/li>\n<li>Type <kbd>pacman -S dialog<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> to install the Wi-Fi menu.<\/li>\n<li>Type <kbd>pacman -S wpa_actiond<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> to install the software that allows you to automatically connect to known networks.<\/li>\n<li>Type <kbd>systemctl enable netctl-auto@<var>interfacename<\/var>.service<\/kbd> to turn on the auto-connection service for your wireless adapter.<\/li>\n<li>Next time you reboot, type <kbd>wifi-menu <var>interfacename<\/var><\/kbd> to access the wireless menu for your adapter. After you connect to the network for the first time, you will be automatically connected for subsequent boots. <i>Do not<\/i> enter this now, or you will lose access to your network.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-21-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-21-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 21\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_8268bac894\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/38\/Install-Arch-Linux-Step-21-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-21-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 21\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/38\/Install-Arch-Linux-Step-21-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-21-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 2\">2<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install the base system.<\/b> Type in <kbd>pacstrap \/mnt base base-devel<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>. The system will begin installing onto your computer.\n\n<ul><li>This process will usually take around 15-30 minutes depending on your Internet connection speed.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-22-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-22-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 22\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_ec1a54a79a\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4e\/Install-Arch-Linux-Step-22-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-22-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 22\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4e\/Install-Arch-Linux-Step-22-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-22-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 3\">3<\/div>\n<div class=\"step\">\n<b class=\"whb\">Open \"chroot\" access.<\/b> Type in <kbd>arch-chroot \/mnt<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>. This will allow you to change aspects of the root directory, including the password.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-23-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-23-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 23\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_240401135b\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/f5\/Install-Arch-Linux-Step-23-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-23-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 23\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/f5\/Install-Arch-Linux-Step-23-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-23-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 4\">4<\/div>\n<div class=\"step\">\n<b class=\"whb\">Set a password.<\/b> This is the password that you'll use to log into the root account. To do so:\n\n<ul><li>Type in <kbd>passwd<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in a password and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Retype your password and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-24-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-24-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 24\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_7535b595ce\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/5\/56\/Install-Arch-Linux-Step-24-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-24-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 24\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/5\/56\/Install-Arch-Linux-Step-24-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-24-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 5\">5<\/div>\n<div class=\"step\">\n<b class=\"whb\">Set a language.<\/b> To do so:\n\n<ul><li>Type in <kbd>nano \/etc\/locale.gen<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Scroll down to select your preferred language.<\/li>\n<li>Select the letter directly in front of the \"#\" symbol behind your language and press <kbd class=\"keyboard-key nowrap\">Del<\/kbd>.<\/li>\n<li>Delete the \"#\" symbol from any other versions of your language (e.g., all versions of \"en_US\").<\/li>\n<li>Press <kbd class=\"keyboard-key nowrap\">Ctrl<\/kbd>+<kbd class=\"keyboard-key nowrap\">O<\/kbd> (or <kbd class=\"keyboard-key nowrap\">\u2318 Command<\/kbd>+<kbd class=\"keyboard-key nowrap\">O<\/kbd> on a Mac), then press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Exit by pressing <kbd class=\"keyboard-key nowrap\">Ctrl<\/kbd>+<kbd class=\"keyboard-key nowrap\">X<\/kbd> or <kbd class=\"keyboard-key nowrap\">\u2318 Command<\/kbd>+<kbd class=\"keyboard-key nowrap\">X<\/kbd>.<\/li>\n<li>Type in <kbd>locale-gen<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd> to finish setting up your language.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-25-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-25-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 25\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_185258feab\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/e\/e6\/Install-Arch-Linux-Step-25-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-25-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 25\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/e\/e6\/Install-Arch-Linux-Step-25-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-25-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 6\">6<\/div>\n<div class=\"step\">\n<b class=\"whb\">Select a time zone.<\/b> To do so:\n\n<ul><li>Type in <kbd>cd usr\/share\/zoneinfo<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in <kbd>ls<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Find your country or region, then type in <kbd>cd usr\/share\/zoneinfo\/<i>country<\/i><\/kbd> (e.g., <kbd>America<\/kbd>) and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in <kbd>ls<\/kbd> again and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Find your preferred time zone, then type in <kbd>ln -s \/usr\/share\/zoneinfo\/<i>country<\/i>\/<i>timezone<\/i> \/etc\/localtime<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-26-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-26-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 26\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_74393338df\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/35\/Install-Arch-Linux-Step-26-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-26-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 26\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/35\/Install-Arch-Linux-Step-26-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-26-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 7\">7<\/div>\n<div class=\"step\">\n<b class=\"whb\">Set a host name for your computer.<\/b> To do so, type in <kbd>echo <i>name<\/i> &gt; \/etc\/hostname<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.\n\n<ul><li>To name your computer \"Panda\", for example, you'd enter <kbd>echo Panda &gt; \/etc\/hostname<\/kbd> here.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-27-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-27-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 27\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_39b61a2f49\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/bb\/Install-Arch-Linux-Step-27-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-27-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 27\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/b\/bb\/Install-Arch-Linux-Step-27-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-27-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 8\">8<\/div>\n<div class=\"step\">\n<b class=\"whb\">Download the GRUB bootloader.<\/b> This is the program that will install your Arch Linux for you. To do so:\n\n<ul><li>Type in <kbd>pacman -S grub-bios<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Type in <kbd>y<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<li>Wait for GRUB to finish downloading.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-28-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-28-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 28\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_3b760028a5\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/0\/0d\/Install-Arch-Linux-Step-28-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-28-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 28\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/0\/0d\/Install-Arch-Linux-Step-28-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-28-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 9\">9<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install GRUB.<\/b> Make sure that when you do this, you install it on your actual hard drive (e.g., \"sda\"), not the partition (e.g., \"sda1\"). To install GRUB:\n\n<ul><li>Type in <kbd>grub-install \/dev\/<i>drive name<\/i><\/kbd> (e.g., <kbd>grub-install \/dev\/sda<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-29-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-29-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 29\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_682705c3ac\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4b\/Install-Arch-Linux-Step-29-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-29-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 29\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4b\/Install-Arch-Linux-Step-29-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-29-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 10\">10<\/div>\n<div class=\"step\">\n<b class=\"whb\">Create an \"init\" file.<\/b> This file stores information about your computer's hardware, making it available for utilization by Linux. To do so, type in <kbd>mkinitcpio -p linux<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-30-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-30-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 30\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_4c7a4664d2\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/c\/c9\/Install-Arch-Linux-Step-30-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-30-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 30\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/c\/c9\/Install-Arch-Linux-Step-30-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-30-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 11\">11<\/div>\n<div class=\"step\">\n<b class=\"whb\">Create a configuration file for GRUB.<\/b> To do so, type in <kbd>grub-mkconfig -o \/boot\/grub\/grub.cfg<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-31-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-31-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 31\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_669fa62172\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/8f\/Install-Arch-Linux-Step-31-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-31-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 31\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/8f\/Install-Arch-Linux-Step-31-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-31-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 12\">12<\/div>\n<div class=\"step\">\n<b class=\"whb\">Create an \"fstab\" file.<\/b> Type in <kbd>genfstab \/mnt &gt;&gt; \/mnt\/etc\/fstab<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>. Doing so allows Arch Linux to identify your partition's file systems.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-32-Version-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-32-Version-3.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 32\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_8d3d410b49\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/83\/Install-Arch-Linux-Step-32-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-32-Version-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 32\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/83\/Install-Arch-Linux-Step-32-Version-3.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-32-Version-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 13\">13<\/div>\n<div class=\"step\">\n<b class=\"whb\">Reboot your computer.<\/b> To do so, type in <kbd>umount \/mnt<\/kbd> and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>, type in <kbd>reboot<\/kbd>, press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>, remove your installation media, and wait for your system to finish rebooting.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage final_li\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Install-Arch-Linux-Step-33.jpg\" class=\"image lightbox\" data-href=\"\/Image:Install-Arch-Linux-Step-33.jpg?ajax=true&amp;aid=581742\"><div class=\"content-spacer\">\n<img alt=\"Image titled Install Arch Linux Step 33\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_968d9d69ad\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/3a\/Install-Arch-Linux-Step-33.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-33.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Install Arch Linux Step 33\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/3a\/Install-Arch-Linux-Step-33.jpg\/aid581742-v4-728px-Install-Arch-Linux-Step-33.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 14\">14<\/div>\n<div class=\"step\">\n<b class=\"whb\">Log into your account.<\/b> Type <kbd>root<\/kbd> into the \"login\" field and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>, then type your password in and press <kbd class=\"keyboard-key nowrap\">\u21b5 Enter<\/kbd>. You have successfully installed and opened Arch Linux on your computer.\n\n<ul><li>If you want to install a graphical user interface (GUI) that allows you to use the mouse, <a href=\"https:\/\/www.wikihow.com\/Install-Gnome-on-Arch-Linux\" title=\"Install Gnome on Arch Linux\">try installing GNOME<\/a> on your computer.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<div class=\"wh_ad_inner step_ad\" data-service=\"adsense\" data-loaded=\"1\" data-lateload=\"0\" data-adsensewidth=\"728\" data-adsenseheight=\"90\" data-slot=\"5875012246\" data-channels=\"8752286177\" data-viewablerefresh=\"1\" data-aps-timeout=\"2000\">\n<ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-9543332082073187\" data-ad-slot=\"5875012246\">\n<\/ins><\/div>\n<div class=\"clearall adclear\">\n<\/div><\/div><\/li>\n<\/ol><div class=\"clearall\">\n<\/div>\n<\/div>\n<div id=\"donate_section\" data-image_count=\"5\" data-image_number=\"\" data-non_profit_name=\"\">\n\n<div class=\"qa section sticky qa_show_unanswered_questions\" data-few_contribs=\"1\" data-search_enabled=\"0\" data-aid=\"581742\">\n\n\t<h2>\n\t\t\n\t\tCommunity Q&amp;A\n\t<\/h2>\n\t<div id=\"qa\" class=\"section_text\">\n\t\t<div id=\"qa_answered_questions_container\">\n\t\t\t<div id=\"qa_aq_search_container\">\n\t\t\t\t<div id=\"qa_aq_search_heading\" class=\"qa_section_heading\">Search<\/div>\n\t\t\t\t<input class=\"qa_aq_search\" placeholder=\"Type your query here\" \/><\/div>\n\t\t\t<a id=\"qa_add_curated_question\">Add New Question<\/a>\n\t\t\t<ul id=\"qa_article_questions\" class=\"qa_questions qa_aqs\"><li class=\"qa_aq qa_block\" data-aqid=\"154200\" data-cqid=\"154217\" data-caid=\"155073\" data-sqid=\"796695\" data-qa_inactive=\"0\" data-votes_up=\"3\" data-votes_down=\"2\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">When booting from USB, Arch does a kernel panic when I try to boot as a UEFI, but I can boot as legacy. The system boots Windows as a UEFI. I want to dual boot the system. How should I partition the disk?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">The Legacy\/UEFI problem has nothing to do with how you partitioned the disk. If it works in legacy mode, then double check your bios settings. Set Boot Mode to EFI. Disable Secure Boot. Check if your stick actually has an EFI folder. If not, try a greater update of Arch and\/or of the software you use to put Arch on the stick.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 2<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 3<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"511032\" data-cqid=\"511049\" data-caid=\"511905\" data-sqid=\"2246783\" data-qa_inactive=\"0\" data-votes_up=\"0\" data-votes_down=\"0\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">I had to exit out of chroot in order to follow step 12. Is there a reason for this?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">The answer is twofold. While you're chrooted into the Arch partition, you're effectively using the system. The system as you currently have it does not have the command genfstab in it, and therefore cannot use it. The path that the tutorial uses, \/mnt\/etc\/fstab, is only valid when you are not chrooted into the system. Remember that when you chroot, you chroot into \/mnt, and therefore effectively make that the root (\/) directory. Thus, if you tried to use the same command, you'd pretty much be saying \"put the fstab file in \/mnt\/mnt\/etc\/fstab\" when there is no such directory. <\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 0<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 0<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t<\/ul><\/div>\n\t\t<div class=\"qa_divider qa_no_unpatrolled_questions\">\n\t\t<div class=\"qa_divider\">\n\n\t\t\t\t<div id=\"qa_submitted_questions_container\">\n\t\t\t\t\t<div class=\"qa_section_heading\">Unanswered Questions<\/div>\n\t\t\t\t\t<ul id=\"qa_submitted_questions\" class=\"qa_questions\"><li class=\"qa_sq\" data-sqid=\"2190024\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">Why update this? This guide was just fine a couple of weeks ago, the way it was before. In fact, I think you now have 1 or 2 steps out of order.<\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t\t\t<li class=\"qa_sq\" data-sqid=\"677668\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">What do I do if I have an error when installing? <\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t<\/ul><div id=\"qa_submitted_spinner\" class=\"qa_spinner\">\n\t\t\t\t\t\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"qa_divider\">\n\n\t\t<div class=\"qa_section_heading\" id=\"qa_ask_heading\">Ask a Question<\/div>\n\t\t<div id=\"qa_submitted_question_form\">\n\t\t\t<input id=\"qa_asked_question\" type=\"text\" maxlength=\"200\" placeholder=\"What do you need to know? We'll do our best to find the answer.\" \/><div id=\"qa_asked_count\">200 characters left<\/div>\n\t\t\t<input id=\"qa_email\" type=\"text\" placeholder=\"Email address (optional)\" \/><div id=\"qa_email_prompt\">Include your email address to get a message when this question is answered.<\/div>\n\t\t\t<a id=\"qa_submit_button\" class=\"button primary\">Submit<\/a>\n\t\t<\/div>\n\n\t\t\t\n\t\t\t<div id=\"qa_flag_options\" data-type=\"\">\n\t\t\t<div class=\"hint_box\">\n\t\t\t\t<ul><li><a href=\"#\" data-option=\"answered\">Already answered<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#\" data-option=\"not_question\">Not a question<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#\" data-option=\"bad_question\">Bad question<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#\" data-option=\"other\">Other<\/a><\/li>\n\t\t\t\t<\/ul><\/div>\n\t\t<\/div>\n\t\t<div id=\"qa_answer_flag_options\" data-type=\"\">\n\t\t\t<div class=\"hint_box\">\n\t\t\t\t<ul><\/ul><\/div>\n\t\t<\/div>\n\n\t\t\n\t\t\n<\/div>\n<\/div>\n<div id=\"qa_box\">\n\n<div class=\"section tips  sticky\">\n<h2>Tips<\/h2>\n<div id=\"tips\" class=\"section_text\">\n<ul><li>Once the installation is complete, you will have a minimal system which uses only the command line. There are many desktop managers or other window management programs which can be installed on your system.<\/li>\n<\/ul><div class=\"clearall\">\n<\/div>\n<\/div>\n<div class=\"section warnings  sticky\">\n<h2>Warnings<\/h2>\n<div id=\"warnings\" class=\"section_text\">\n<ul><li>Ensure that you pick a root password that is both memorable and secure. If you forget it, you will not be able to properly administer your system.<\/li>\n<\/ul><div class=\"clearall\">\n<\/div>\n<\/div>\n\n\n<div class=\"section sourcesandcitations  sticky\">\n<h2>Sources and Citations<\/h2>\n<div id=\"sourcesandcitations\" class=\"section_text\">\n<div class=\"references-small\">\n<ol class=\"references sources\"><li id=\"_note-1\">\n<a href=\"#_ref-1\" target=\"_blank\">\u2191<\/a> <a rel=\"nofollow\" class=\"external free\" href=\"https:\/\/www.youtube.com\/watch?v=Wqh9AQt3nho\" target=\"_blank\">https:\/\/www.youtube.com\/watch?v=Wqh9AQt3nho<\/a>\n<\/li>\n<\/ol><\/div>\n<br \/><div class=\"clearall\">\n<\/div>\n<\/div>\n\n\n\n<div id=\"summary_wrapper\"><div id=\"summary_text\">\n<h2>Article Summary<a href=\"#\" id=\"summary_close\">X<\/a>\n<\/h2>\n<p>1. Back up your PC to an external drive.<br \/>\n2. Download the installation image from <b>archlinux.com\/download<\/b>.<br \/>\n3. Burn the image to a DVD.<br \/>\n4. Restart the computer.<br \/>\n5. Boot from the DVD.<br \/>\n6. Select <b>Boot Arch Linux<\/b> and press <b>Enter<\/b>.<br \/>\n7. Create main and swap partitions.<br \/>\n8. Use \u2033pacstrap \/mnt base base-devel\u2033 to install.<br \/><\/p>\n<div class=\"clearall\">\n<div class=\"s-help-wrap\" data-type=\"summarytexthelp\">\nDid this summary help you?<button class=\"s-help-response\" data-value=\"1\" data-type=\"summarytexthelp\" data-prompt-text=\"We are so glad to have helped!\" data-finish-prompt=\"Thank you! &lt;a href='#' class='firststeplink'&gt;Return to the article to learn more...&lt;\/a&gt;\" data-textarea-prompt=\"Please describe what was helpful in the video...\" type=\"submit\">Yes<\/button><button class=\"s-help-response\" data-value=\"0\" data-type=\"summarytexthelp\" data-prompt-text=\"Sorry that the video wasn't helpful.\" data-finish-prompt=\"Sorry the summary wasn't helpful. &lt;a href='#' class='firststeplink'&gt;Return to the article to learn more...&lt;\/a&gt;\" data-textarea-prompt=\"Please tell us what you would have liked to see in the video...\" type=\"submit\">No<\/button><div class=\"s-help-feedback-wrap\">\n<textarea class=\"s-help-textarea\" rows=\"10\" cols=\"50\"><\/textarea><\/div>\n<\/div>\n<\/div><\/div>\n\n\n\n\n\n<\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:48:47+0000","updated_at":"2018-11-07T16:48:47+0000","published_at":"2009-04-09T00:00:00+0000","published_by":["wikiHow"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"www.wikihow.com","preview_picture":"https:\/\/www.wikihow.com\/images\/3\/3a\/Install-Arch-Linux-Step-33.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":457,"title":"Who says engineers can\u2019t become entrepreneurs?","url":"https:\/\/medium.freecodecamp.org\/who-says-engineers-cant-be-entrepreneurs-8c7f7a6834da?gi=efb054d80e1e","content":"<figure id=\"f121\" class=\"graf graf--figure graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*yyfvd_Hgg7KLMXgRVokqYQ.jpeg\" data-width=\"717\" data-height=\"501\" data-action=\"zoom\" data-action-value=\"1*yyfvd_Hgg7KLMXgRVokqYQ.jpeg\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*yyfvd_Hgg7KLMXgRVokqYQ.jpeg\" alt=\"image\" \/><\/div><\/figure><p id=\"8138\" class=\"graf graf--p graf-after--figure\">A lot of people warned me not to <a href=\"https:\/\/medium.com\/swlh\/why-i-left-the-best-job-in-the-world-3689a5a4649a\" data-href=\"https:\/\/medium.com\/swlh\/why-i-left-the-best-job-in-the-world-3689a5a4649a\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">walk away from my great position at Andreessen Horowitz<\/a> to pursue software engineering.<\/p><p id=\"86b9\" class=\"graf graf--p graf-after--p\">\u2026And now that I\u2019m leaving my <a href=\"https:\/\/medium.com\/swlh\/what-happened-after-i-left-the-best-job-in-the-world-to-become-an-engineer-ee06caca7db2#.ingil61hq\" data-href=\"https:\/\/medium.com\/swlh\/what-happened-after-i-left-the-best-job-in-the-world-to-become-an-engineer-ee06caca7db2#.ingil61hq\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">dream job engineering at Coinbase<\/a> to co-found a startup, you can bet that \u201ca lot of people\u201d have been reminding me how many startups fail.<\/p><p id=\"8529\" class=\"graf graf--p graf-after--p\">It was my Mom\u2019s reaction, though, that stuck with me:<\/p><blockquote id=\"0c57\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cWhy would you leave the best job in the world\u2026\u00a0again?\u201d<\/div><\/blockquote><p id=\"c63b\" class=\"graf graf--p graf-after--pullquote\">It\u2019s a good question\u200a\u2014\u200aand I\u2019m writing this post to answer it.<\/p><p id=\"1ab3\" class=\"graf graf--p graf-after--p\">As those of you who\u2019ve been following along know, the past year and half has been a wild one for me, from falling passionately in love with programming, to leaving an amazing job to pursue this passion, to enduring Hack Reactor\u2019s intensive coding program, to tackling complex problems as an engineer at a major cryptocurrency startup.<\/p><p id=\"a061\" class=\"graf graf--p graf-after--p\">Long story short, it was an exhilarating journey with countless late nights, early mornings, and moments of self-doubt.<\/p><p id=\"888f\" class=\"graf graf--p graf-after--p\">Today, I\u2019ll pick up <a href=\"https:\/\/medium.com\/swlh\/what-happened-after-i-left-the-best-job-in-the-world-to-become-an-engineer-ee06caca7db2#.kxlberhb2\" data-href=\"https:\/\/medium.com\/swlh\/what-happened-after-i-left-the-best-job-in-the-world-to-become-an-engineer-ee06caca7db2#.kxlberhb2\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">where I left off<\/a>, sharing some of my ups and downs as an engineer at Coinbase\u2026 and reveal some of the exciting projects I\u2019m diving into now!<\/p><h3 id=\"6aa2\" class=\"graf graf--h3 graf-after--p\">My experience as a full-time software\u00a0engineer<\/h3><p id=\"bdda\" class=\"graf graf--p graf-after--h3\">It was April of 2016 when Coinbase offered me a software engineering role on their Brokerage team. The team was brilliant, the industry was cutting edge. Needless to say, I accepted immediately!\u00a0<br \/>\u00a0<br \/>The days before I started were nerve-wracking. (If you\u2019ve ever waited to start a new job before, you know how fear and anxiety can make you second-guess everything.) I haven\u2019t been so nervous to walk into a building since the first day of middle school. \ud83d\ude33<\/p><p id=\"4262\" class=\"graf graf--p graf-after--p\">My internal monologue as I rode the elevator up their towering office building in San Francisco\u2019s Financial District went something like this: \u201cI\u2019ve never been a full-time software engineer at a company before. Do I even know what I\u2019m doing? What if I get asked to build a micro-service within 48 hours? Or to rebuild the entire website within a week? I\u2019m surely doomed\u2026\u201d<br \/>\u00a0<br \/>I stepped out of the elevator, took a deep breath\u200a\u2014\u200aand stood there frozen for two minutes, trying to build up enough courage to ring the buzzer. \u201cBuzzzzz\u201d\u00a0<br \/>\u00a0<br \/>Naturally, my nervousness was misplaced. The receptionist was warm and kind. The recruiting coordinator was more worried about how I liked my workspace than what I would be doing at it. I\u2019d barely put down my backpack before the CEO, Brian Armstrong, came by to say \u201cWelcome! Glad to have you on board!\u201d<br \/>\u00a0<br \/>\u201cPhew, no micro-services just yet,\u201d I thought.<br \/>\u00a0<br \/>By then, my nervousness turned into excitement\u200a\u2014\u200aI couldn\u2019t wait to get hacking. Perhaps I\u2019m not ready to build a micro-service on my first day, but I sure was ready to build <em class=\"markup--em markup--p-em\">something<\/em>!\u00a0<br \/>\u00a0<br \/>The next few weeks were a steady build of responsibility and familiarity\u200a\u2014\u200agetting comfortable with the code bases, merging my first few pull requests, and integrating with the workflow, team, and culture. The challenge of diving into a completely new code base, gaining momentum in fixing bugs and implementing features, and building relationships with new friends kept me engaged and excited.<br \/>\u00a0<br \/>Fast-forward a couple months later, and I was firmly settled in. By then, I had two major projects under my belt\u200a\u2014\u200aone where I A\/B tested and implemented a new fee structure, and another where I built a new React\/Redux front-end for supporting Ethereum buys and sells. I was ready to start on my third and largest project, which was to rewrite the existing monolith Rails front-end in React. Four months later, we <a href=\"https:\/\/twitter.com\/iam_preethi\/status\/796074288432357376\" data-href=\"https:\/\/twitter.com\/iam_preethi\/status\/796074288432357376\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener noopener\" target=\"_blank\">shipped v1<\/a> of that massive rewrite. Phew!\u00a0<br \/>\u00a0<br \/>By then it had already been six months and there were lots of things that I was really enjoying about working at Coinbase:<\/p><h4 id=\"3733\" class=\"graf graf--h4 graf-after--p\">Team<\/h4><p id=\"1df3\" class=\"graf graf--p graf-after--h4\">My gut instinct about the team was right. I couldn\u2019t have asked for a smarter bunch of engineers to work with everyday.<\/p><p id=\"ac44\" class=\"graf graf--p graf-after--p\">From infrastructure, data science, fraud, payments, web and mobile, I had infinite opportunity to learn and grow from engineers across different engineering domains.<\/p><p id=\"60aa\" class=\"graf graf--p graf-after--p\">Beyond their strong technically chops, I admired that they were also well-rounded. (Not to mention mature thoughtful and independent thinkers.)<\/p><h4 id=\"ffee\" class=\"graf graf--h4 graf-after--p\">Ownership<\/h4><p id=\"f9fa\" class=\"graf graf--p graf-after--h4\">I loved that Coinbase gives engineers a high degree of autonomy and ownership over the implementation details of the features they build, rather than a top-down regimented system.<\/p><p id=\"4118\" class=\"graf graf--p graf-after--p\">Being invested in my projects and tackling problems independently helped me learn and grow more quickly as an engineer.<\/p><h4 id=\"bf81\" class=\"graf graf--h4 graf-after--p\">Growth mindset<\/h4><p id=\"325e\" class=\"graf graf--p graf-after--h4\">Another thing that I appreciated was that the company deeply cares about continuous feedback and learning. Nothing we did was ever perfect\u200a\u2014\u200ain fact, disorder was the norm.<\/p><p id=\"a306\" class=\"graf graf--p graf-after--p\">Regardless, the team would make sure to have retrospectives after every project to discuss what did and didn\u2019t go well.<\/p><p id=\"c605\" class=\"graf graf--p graf-after--p\">\u2026And let me tell you, we didn\u2019t hold back. We criticized and laughed at ourselves, and then thought carefully about how we could avoid those mistakes the next time around. Personal and team progress is hard but important work.<\/p><h4 id=\"fc39\" class=\"graf graf--h4 graf-after--p\">Oh, so many\u00a0perks<\/h4><p id=\"4de4\" class=\"graf graf--p graf-after--h4\">I don\u2019t really have to explain this one\u200a\u2014\u200awho can resist the fully stocked kitchen, lovely office staff, pleasantly decorated bathrooms, and catered meals\u200a\u2014\u200aall with a 360-degree view of beautiful San Francisco?<\/p><h3 id=\"044d\" class=\"graf graf--h3 graf-after--p\">The downsides<\/h3><p id=\"85e1\" class=\"graf graf--p graf-after--h3\">As with everything, every pro comes with cons. Here are a few things that I personally didn\u2019t like about plugging into the corporate engineering structure:<\/p><h4 id=\"757c\" class=\"graf graf--h4 graf-after--p\"><strong class=\"markup--strong markup--h4-strong\">Working in\u00a0Rails<\/strong><\/h4><p id=\"08d4\" class=\"graf graf--p graf-after--h4\">While half of my time was spent in the new JavaScript and React front-end, the other half inevitably involved fixing bugs and implementing features in our monolithic Rails application. Rails is a phenomenal framework for many reasons, but personally, I was struggling to fall in love with it. It eventually became the least favorite part of my job.\u00a0<br \/>\u00a0<br \/>As regular readers know, the JavaScript community and ecosystem is a <em class=\"markup--em markup--p-em\">huge<\/em> passion area for me. So naturally, I wanted to contribute my time and energy to mastering JavaScript, functional programming, and the web. Instead, I was fixing uninteresting bugs and implementing unsexy features in what <em class=\"markup--em markup--p-em\">felt<\/em> like the wrong framework.<\/p><p id=\"f815\" class=\"graf graf--p graf-after--p\">Unfortunately, the reality is that maintaining old monolithic software is a core part of what we, as software engineers, must do when we\u2019re working in a large engineering organization. Those are just the facts of life!<\/p><h4 id=\"5814\" class=\"graf graf--h4 graf-after--p\">Open office work\u00a0space<\/h4><p id=\"d090\" class=\"graf graf--p graf-after--h4\">Open offices are <a href=\"http:\/\/www.npr.org\/sections\/money\/2016\/06\/03\/480625378\/episode-704-open-office\" data-href=\"http:\/\/www.npr.org\/sections\/money\/2016\/06\/03\/480625378\/episode-704-open-office\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener noopener\" target=\"_blank\">wildly popular<\/a> in the tech world\u200a\u2014\u200ain spite of the the many studies showing how they can kill creativity and make focus impossible.<\/p><p id=\"eaf5\" class=\"graf graf--p graf-after--p\">The constant human traffic and frequent interruptions made it feel like I was coding on live television. I found it difficult to be productive and eventually found myself working from home more and more frequently\u200a\u2014\u200ait was the only place I could focus and get deep work done.<\/p><h4 id=\"7ccf\" class=\"graf graf--h4 graf-after--p\"><strong class=\"markup--strong markup--h4-strong\">The product and projects I wasn\u2019t passionate about<\/strong><\/h4><p id=\"d996\" class=\"graf graf--p graf-after--h4\">Alright, enough complaining Preethi! Time to get honest about why I felt restless\u200a\u2014\u200athe truth is, <em class=\"markup--em markup--p-em\">my unhappiness had nothing to do with Coinbase as an organization.<\/em><br \/>\u00a0<br \/>At the end of the day, it was because <em class=\"markup--em markup--p-em\">I wasn\u2019t passionate about the product Coinbase offers<\/em>.<\/p><p id=\"f1f1\" class=\"graf graf--p graf-after--p\">First, it wasn\u2019t something that fired me up to get out of bed in the morning. It was difficult to care deeply about the features we were building, when financial products and services wasn\u2019t a passion area for me personally.<br \/>\u00a0<br \/>Second, I realized that being at a company with separate roles for product management, project management, engineering management, design, and engineering inevitably meant that the role I served was purely engineering\u200a\u2014\u200ato implement and engineer solutions and fix bugs.<\/p><p id=\"6717\" class=\"graf graf--p graf-after--p\">Those of you who know me can probably imagine how neglecting the entrepreneurial part of my mind would leave me unsatisfied. I wanted to do everything from identifying the business purpose, to building the product, to maintaining and advancing the codebase.<\/p><p id=\"53c8\" class=\"graf graf--p graf-after--p\">I needed to be both the entrepreneur <em class=\"markup--em markup--p-em\">and<\/em> the programmer.<\/p><h3 id=\"2633\" class=\"graf graf--h3 graf-after--p\">Doubting the corporate engineering trajectory<\/h3><p id=\"defc\" class=\"graf graf--p graf-after--h3\">At this point, I knew that the career trajectory I was on needed to change course. Maybe not as drastically as leaving finance for engineering\u200a\u2014\u200abut something that engaged me as both an engineer and an entrepreneur needed to happen. Soon.\u00a0<br \/>\u00a0<br \/>Nights and weekends became my \u201centrepreneur-programmer\u201d time, and I meticulously planned my schedule so I\u2019d have 1\u20132 hours every day to explore, learn, hack, write, and have fun. Soon, that 1\u20132 hours grew into 2\u20134 hours a day. At a certain point, side projects like writing and hacking on app ideas began to seem more \u201creal\u201d than my 9\u20135.\u00a0<br \/>\u00a0<br \/>Did I feel guilty for wanting to leave a perfect team to work on another crazy dream? <em class=\"markup--em markup--p-em\">Of course.<\/em><\/p><p id=\"4bc1\" class=\"graf graf--p graf-after--p\">Rather than make a spur of the moment decision, I decided to take some time off to reflect on what I was feeling. It\u2019s tricky to figure out what truly makes you tick. But I needed to figure it out so I could answer the burning question: w<em class=\"markup--em markup--p-em\">here can I contribute to this world in the most meaningful way?<\/em><\/p><p id=\"b3dd\" class=\"graf graf--p graf-after--p\">I treated this time like a \u201ctest drive\u201d to see what it\u2019s like to wake up every morning and do whatever my heart and mind desired. Was this just a phase, or could I turn my ideas into a full-time reality?\u00a0<br \/>\u00a0<br \/>Two weeks turned into a month. And then three months. And then to this moment, writing this post: the conviction that I\u2019m ready to leave Coinbase and pursue new, more personal coding challenges and entrepreneurial endeavors.<\/p><h3 id=\"433c\" class=\"graf graf--h3 graf-after--p\">Is this the the \u201cright time\u201d to jump\u00a0ship?<\/h3><p id=\"d4ac\" class=\"graf graf--p graf-after--h3\">One of the biggest things we struggle with as entrepreneurs is knowing when \u201cthe right time\u201d is\u200a\u2014\u200athe right time to pursue that passion, the right time to shun reason in the name of fulfilling our dreams.<\/p><p id=\"7529\" class=\"graf graf--p graf-after--p\">But I\u2019ve come to realize that there is no rule book that will give us the answer. That the \u201cright time\u201d never comes\u200a\u2014\u200ayou can spend your whole life getting your checks and balances in order and it\u2019ll never feel \u201cjust right.\u201d There\u2019s <em class=\"markup--em markup--p-em\">always<\/em> more you can do to prepare, get more experience, and minimize risk.<\/p><p id=\"58a3\" class=\"graf graf--p graf-after--p\">\u2026But that doesn\u2019t mean we <em class=\"markup--em markup--p-em\">should<\/em>. Our minds will always rationalize the fear of the unknown.\u00a0<br \/>\u00a0<br \/>So is now the right time for me to leave Coinbase? I honestly don\u2019t know for sure. In fact, a few of my close friends and mentors said that waiting until I get \u201c2 years of experience\u201d seems more appropriate. Or that \u201ctesting the waters\u201d by working at a small startup would be a better way to ease into it.<\/p><p id=\"7e83\" class=\"graf graf--p graf-after--p\">Is it really, though? Or would that just be an excuse to escape the fear and uncertainty of doing it <em class=\"markup--em markup--p-em\">now<\/em>. And here\u2019s my conclusion:<\/p><p id=\"d496\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">I don\u2019t believe another six months or two years will make the answer any more clear.<\/strong><\/p><p id=\"0475\" class=\"graf graf--p graf-after--p\">What I <em class=\"markup--em markup--p-em\">do<\/em> know is that I feel more confident than ever in my skills as a software engineer. I\u2019m proud of everything I\u2019ve been able to learn and improve my skills at thanks to Coinbase, such as:<\/p><ul class=\"postList\"><li id=\"55e5\" class=\"graf graf--li graf-after--p\">Version control (git)<\/li><li id=\"8615\" class=\"graf graf--li graf-after--li\">Deployment (Docker, circleCI)<\/li><li id=\"0ded\" class=\"graf graf--li graf-after--li\">Monitoring (New Relic)<\/li><li id=\"6761\" class=\"graf graf--li graf-after--li\">Error reporting (Bugsnag)<\/li><li id=\"a07c\" class=\"graf graf--li graf-after--li\">How to refactor legacy code into a new framework (Rails \u2192 React)<\/li><li id=\"2e1b\" class=\"graf graf--li graf-after--li\">How to refactor from one state management library to another (jQuery \u2192 Redux \u2192 Mobx)<\/li><li id=\"1d98\" class=\"graf graf--li graf-after--li\">How to give thoughtful, constructive code reviews<\/li><li id=\"3428\" class=\"graf graf--li graf-after--li\">How to efficiently and effectively collaborate with designers in going from idea \u2192 mockup \u2192 wireframe \u2192 code<\/li><li id=\"17ef\" class=\"graf graf--li graf-after--li\">How to test code using unit tests and integration tests<\/li><li id=\"b33b\" class=\"graf graf--li graf-after--li\">Knowing when to choose between the \u201ccorrect and thoughtful\u201d solution that takes longer to implement but is more robust and scalable vs. a solution that is quick and \u201cdoes the job\u201d but isn\u2019t robust or scalable<\/li><li id=\"bb7c\" class=\"graf graf--li graf-after--li\">How to handle situations where you deploy a really terrible bug that causes the entire application to crash in the middle of a big product launch \ud83d\ude43<\/li><li id=\"fab5\" class=\"graf graf--li graf-after--li\">How to collaborate with a team of engineers to build a production-grade application that scales to millions of users<\/li><li id=\"da0d\" class=\"graf graf--li graf-after--li\">\u2026 and so much more<\/li><\/ul><p id=\"3206\" class=\"graf graf--p graf-after--li\">And of course, there are tons of things I haven\u2019t learned yet\u200a\u2014\u200aand all the tantalizing things I\u2019ll probably <em class=\"markup--em markup--p-em\">never<\/em> learn about programming.<\/p><p id=\"3de7\" class=\"graf graf--p graf-after--p\">Ultimately, it\u2019s okay, because the most important skill I\u2019ve learned over the last two years is <em class=\"markup--em markup--p-em\">how to learn<\/em>\u200a\u2014\u200athe master key to learning and conquering any tool, framework or language I need.<\/p><p id=\"9a66\" class=\"graf graf--p graf-after--p\">This is all I ever wanted when I set out to go \u201clearn to code,\u201d and now I have it at my finger tips.\u00a0<br \/>\u00a0<br \/>Why wouldn\u2019t I use these skills to build my own products full time? The answer is simple: <em class=\"markup--em markup--p-em\">\u201cwhy not.\u201d<\/em><\/p><p id=\"aacb\" class=\"graf graf--p graf-after--p\">Life is too short not to experiment\u200a\u2014\u200aand the time to see what value I can create through those experiments is now, while I\u2019m young, driven, and the worst case scenarios aren\u2019t really that bad.<\/p><p id=\"b3bb\" class=\"graf graf--p graf-after--p\">As Leo Tolstoy puts it so eloquently:<\/p><blockquote id=\"5d55\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201cOnce we\u2019ve thrown off our habitual paths, we think all is lost; but it\u2019s only here that the new and the good\u00a0begins.\u201d<\/div><\/blockquote><h3 id=\"77f3\" class=\"graf graf--h3 graf-after--pullquote\">So what\u2019s\u00a0next?<\/h3><h4 id=\"fb11\" class=\"graf graf--h4 graf-after--h3\"><strong class=\"markup--strong markup--h4-strong\">Hacking<\/strong><\/h4><p id=\"2a93\" class=\"graf graf--p graf-after--h4\">There\u2019s so much in the world of software that I\u2019ve been exploring and hacking on \u2014and I\u2019m particularly interested in a few problems I\u2019d love to tackle. I couldn\u2019t be more excited to have the opportunity to go deeper into these problems with my entrepreneurial hat on.<\/p><h4 id=\"4587\" class=\"graf graf--h4 graf-after--p\">Teaching and\u00a0Writing<\/h4><p id=\"9957\" class=\"graf graf--p graf-after--h4\">In addition to hacking away, I\u2019ll continue to spend a portion of my time writing, teaching, speaking, and advocating JavaScript, the web, and functional programming.<\/p><p id=\"f6ba\" class=\"graf graf--p graf-after--p\">If leveling up your Functional JavaScript knowledge is on the to-do list, you can sign up for my free email course on <a href=\"https:\/\/preethikasireddy.typeform.com\/to\/yC9qQr\" data-href=\"https:\/\/preethikasireddy.typeform.com\/to\/yC9qQr\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">The Fundamentals of Functional Programming<\/a>. I\u2019ll also be creating a long-form video course called \u201c<a href=\"https:\/\/preethikasireddy.typeform.com\/to\/scykJM\" data-href=\"https:\/\/preethikasireddy.typeform.com\/to\/scykJM\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Mastering Functional JavaScript<\/a>\u201d later this year.<\/p><h3 id=\"66ba\" class=\"graf graf--h3 graf-after--p\">What are the\u00a0risks?<\/h3><p id=\"430f\" class=\"graf graf--p graf-after--h3\">Any experiment comes with the risk of failure\u200a\u2014\u200aparticularly high when it comes to startups since, after all, <a href=\"http:\/\/fortune.com\/2014\/09\/25\/why-startups-fail-according-to-their-founders\/\" data-href=\"http:\/\/fortune.com\/2014\/09\/25\/why-startups-fail-according-to-their-founders\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">nine out of ten startups fail.<\/a><\/p><p id=\"b715\" class=\"graf graf--p graf-after--p\">That said, I believe that the <em class=\"markup--em markup--p-em\">personal and career <\/em>\u201crisks\u201d of taking time to build a startup idea are way, way lower than we\u2019re lead to believe.<\/p><p id=\"c43f\" class=\"graf graf--p graf-after--p\">Particularly as an engineer with no major family obligations, the worst case scenario for me is that I crash, burn, and take a new engineering job a few months later. Starting salaries are higher than most other industries, and there are always more engineering jobs available than qualified candidates to fill them.<\/p><p id=\"be20\" class=\"graf graf--p graf-after--p\">So yes, my startup ideas could totally fail. Or I could end up losing interest in my passions. Or <em class=\"markup--em markup--p-em\">I could get hit by a bus tomorrow.<\/em> The list goes on.<\/p><p id=\"016b\" class=\"graf graf--p graf-after--p\">Do I have control over any of these events from happening? If I try my hardest, are any of them my fault? Probably not.<\/p><p id=\"37c9\" class=\"graf graf--p graf-after--p\">If there\u2019s one thing I\u2019ve learned through the major transitions in my life\u200a\u2014\u200afrom moving across continents, countries, and cities, to becoming financially independent at 18, to hustling my way into Silicon Valley, to conquering my dream of becoming a programmer\u200a\u2014\u200ait\u2019s that uncertainty motivates me to succeed. In a way, it\u2019s almost comfortable.<\/p><p id=\"d242\" class=\"graf graf--p graf-after--p\">In my eyes, the only true failure would be not starting in the first place.<\/p><h3 id=\"60b0\" class=\"graf graf--h3 graf-after--p\">Looking forward<\/h3><p id=\"f2ec\" class=\"graf graf--p graf-after--h3\">I spent the last two years building a strong foundation as an engineer and more importantly, finding myself\u200a\u2014\u200amy passions, interests, strengths and weaknesses. The next stage of my growth will be all about taking that experience and creating something meaningful for the world out of it.<\/p><p id=\"0718\" class=\"graf graf--p graf-after--p\">I love the saying by Sidney A. Friendman:<\/p><blockquote id=\"ba7e\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div><em class=\"markup--em markup--pullquote-em\">\u201cYou can achieve anything you want in life if you have the courage to dream it, the intelligence to make a realistic plan, and the will to see that plan through to the\u00a0end.\u201d<\/em><\/div><\/blockquote><p id=\"b330\" class=\"graf graf--p graf-after--pullquote\">I\u2019ve dreamt it, sketched out a realistic plan, and now it\u2019s time I see it through to the end.<\/p><p id=\"1807\" class=\"graf graf--p graf-after--p\">Now it\u2019s your turn: <em class=\"markup--em markup--p-em\">tell me about your journey<\/em>.<\/p><p id=\"5088\" class=\"graf graf--p graf-after--p\">Have you ever felt unhappy with your day-to-day? What doubts keep you from chasing your dream\u200a\u2014\u200aor what steps have you taken to overcome them?<\/p><p id=\"7cb2\" class=\"graf graf--p graf-after--p\">Your story is just as important as mine. I can\u2019t wait to hear it!<\/p><p id=\"3497\" class=\"graf graf--p graf-after--p graf--trailing\">Want to stay up to date on my journey and work? Just click \u201cfollow\u201d on <a href=\"https:\/\/medium.com\/@preethikasireddy\" data-href=\"https:\/\/medium.com\/@preethikasireddy\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Medium<\/a>, Twitter (<a href=\"https:\/\/twitter.com\/iam_preethi\" data-href=\"https:\/\/twitter.com\/iam_preethi\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow nofollow nofollow nofollow noopener\" target=\"_blank\">@iam_preethi<\/a>) or <a href=\"http:\/\/preethikasireddy.me\" data-href=\"http:\/\/preethikasireddy.me\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener nofollow nofollow nofollow nofollow noopener\" target=\"_blank\">my site<\/a>.<\/p>","created_at":"2018-11-07T16:48:46+0000","updated_at":"2018-11-07T16:48:46+0000","published_at":"2017-03-17T14:09:25+0000","published_by":["Preethi Kasireddy"],"annotations":[],"mimetype":"text\/html","reading_time":13,"domain_name":"medium.freecodecamp.org","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*yyfvd_Hgg7KLMXgRVokqYQ.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":454,"title":"What is your most productive shortcut with Vim?","url":"https:\/\/stackoverflow.com\/questions\/1218390\/what-is-your-most-productive-shortcut-with-vim\/1220118","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1127&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I've heard a lot about <a href=\"http:\/\/www.vim.org\/\" rel=\"noreferrer\">Vim<\/a>, both pros and cons.\nIt really seems you should be (as a developer) faster with Vim than with any other editor.\nI'm using Vim to do some basic stuff and I'm at best 10 times <em>less productive<\/em> with Vim.<\/p>\n<p>The only two things you should care about when you talk about speed (you may not care enough about them, but you should) are:<\/p>\n<ol><li>Using alternatively left and right\nhands is the <strong>fastest<\/strong> way to use the\nkeyboard. <\/li>\n<li>Never touching the mouse is the\nsecond way to be as fast as possible.\nIt takes ages for you to move your hand,\ngrab the mouse, move it, and bring it\nback to the keyboard (and you often have\nto look at the keyboard to be sure you\nreturned your hand properly to the right place)<\/li>\n<\/ol><p>Here are two examples demonstrating why I'm far less productive with Vim.<\/p>\n<p><strong>Copy\/Cut &amp; paste.<\/strong> I do it all the time. With all the contemporary editors you press <kbd>Shift<\/kbd>  with the left hand, and you move the cursor with your right hand to select text. Then <kbd>Ctrl<\/kbd>+<kbd>C<\/kbd> copies, you move the cursor and <kbd>Ctrl<\/kbd>+<kbd>V<\/kbd> pastes.<\/p>\n<p>With Vim it's horrible:<\/p>\n<ul><li><strong><code>yy<\/code><\/strong> to copy one line (you almost never want the whole line!)<\/li>\n<li><strong><code>[number xx]yy<\/code><\/strong> to copy <code>xx<\/code> lines into the buffer. But you never know exactly if you've selected what you wanted. I often have to do <strong><code>[number xx]dd<\/code><\/strong> then <strong><code>u<\/code><\/strong> to undo!<\/li>\n<\/ul><p>Another example? <strong>Search &amp; replace.<\/strong><\/p>\n<ul><li>In <a href=\"http:\/\/en.wikipedia.org\/wiki\/PSPad\" rel=\"noreferrer\">PSPad<\/a>: <kbd>Ctrl<\/kbd>+<kbd>f<\/kbd> then type what you want you search for, then press <kbd>Enter<\/kbd>.<\/li>\n<li>In Vim: <strong><code>\/<\/code><\/strong>, then type what you want to search for, then if there are some special characters put <strong><code>\\<\/code><\/strong> before <em>each<\/em> special character, then press <kbd>Enter<\/kbd>.<\/li>\n<\/ul><p>And everything with Vim is like that: it seems I don't know how to handle it the right way.<\/p>\n<p>NB : <strong>I've already read the Vim <a href=\"http:\/\/www.viemu.com\/vi-vim-cheat-sheet.gif\" rel=\"noreferrer\">cheat<\/a> <a href=\"http:\/\/www.fprintf.net\/vimCheatSheet.html\" rel=\"noreferrer\">sheet<\/a><\/strong> :)<\/p>\n<p>My question is:<\/p>\n<p>What is the way you use Vim that makes you more productive than with a contemporary editor?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218390\/revisions\" title=\"show all edits to this post\">edited Aug 16 '17 at 10:58<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218390\" href=\"https:\/\/stackoverflow.com\/posts\/1218390\/revisions\" title=\"show revision history for this post\">&#13;\n        12 revs, 9 users 36%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/106140\">Olivier Pons<\/a>&#13;\n        &#13;<\/div><h2 data-answercount=\"50\">&#13;\n                                50 Answers&#13;\n                                50&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        2862&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<h3>Your problem with <strong>Vim<\/strong> is that  you don't grok <strong>vi<\/strong>.<\/h3>\n<p>You mention cutting with <code>yy<\/code> and complain that you almost never want to cut whole lines.  In fact programmers, editing source code, very often want to work on whole lines, ranges of lines and blocks of code.  However, <code>yy<\/code> is only one of many way to yank text into the anonymous copy buffer (or \"register\" as it's called in <strong>vi<\/strong>).<\/p>\n<p>The \"Zen\" of <strong>vi<\/strong> is that you're speaking a language.  The initial <code>y<\/code> is a verb.  The statement <code>yy<\/code> is a synonym for <code>y_<\/code>. The <code>y<\/code> is doubled up to make it easier to type, since it is such a common operation.<\/p>\n<p>This can also be expressed as <code>dd<\/code> <code>P<\/code> (delete the current line and paste a copy back into place; leaving a copy in the anonymous register as a side effect).  The <code>y<\/code> and <code>d<\/code> \"verbs\" take any movement as their \"subject.\"  Thus <code>yW<\/code> is \"yank from here (the cursor) to the end of the current\/next (big) word\" and <code>y'a<\/code> is \"yank from here to the line containing the mark named '<em>a<\/em>'.\"<\/p>\n<p>If you only understand basic up, down, left, and right cursor movements then <strong>vi<\/strong> will be no more productive than a copy of \"notepad\" for you.  (Okay, you'll still have syntax highlighting and the ability to handle files larger than a piddling ~45KB or so; but work with me here).<\/p>\n<p><strong>vi<\/strong> has 26 \"marks\" and 26 \"registers.\"  A mark is set to any cursor location using the <code>m<\/code> command.  Each mark is designated by a single lower case letter.  Thus <code>ma<\/code> sets the '<em>a<\/em>' mark to the current location, and <code>mz<\/code> sets the '<em>z<\/em>' mark.  You can move to the line containing a mark using the <code>'<\/code> (single quote) command.  Thus <code>'a<\/code> moves to the beginning of the line containing the '<em>a<\/em>' mark.  You can move to the precise location of any mark using the <code>`<\/code> (backquote) command.  Thus <code>`z<\/code> will move directly to the exact location of the '<em>z<\/em>' mark.<\/p>\n<p>Because these are \"movements\" they can also be used as subjects for other \"statements.\"<\/p>\n<p>So, one way to cut an arbitrary selection of text would be to drop a mark (I usually use '<em>a<\/em>' as my \"first\" mark, '<em>z<\/em>' as my next mark, '<em>b<\/em>' as another, and '<em>e<\/em>' as yet another (I don't recall ever having interactively used more than four marks in 15 years of using <strong>vi<\/strong>; one creates one's own conventions regarding how marks and registers are used by macros that don't disturb one's interactive context).  Then we go to the other end of our desired text; we can start at either end, it doesn't matter.  Then we can simply use <code>d`a<\/code> to cut or <code>y`a<\/code> to copy.  Thus the whole process has a 5 keystrokes overhead (six if we started in \"insert\" mode and needed to <kbd>Esc<\/kbd> out command mode).  Once we've cut or copied then pasting in a copy is a single keystroke: <code>p<\/code>.<\/p>\n<p>I say that this is one way to cut or copy text.  However, it is only one of many.  Frequently we can more succinctly describe the range of text without moving our cursor around and dropping a mark.  For example if I'm in a paragraph of text I can use <code>{<\/code> and <code>}<\/code> movements to the beginning or end of the paragraph respectively.  So, to move a paragraph of text I cut it using <code>{<\/code> <code>d}<\/code> (3 keystrokes).  (If I happen to already be on the first or last line of the paragraph I can then simply use <code>d}<\/code> or <code>d{<\/code> respectively.  <\/p>\n<p>The notion of \"paragraph\" defaults to something which is usually intuitively reasonable.  Thus it often works for code as well as prose.<\/p>\n<p>Frequently we know some pattern (regular expression) that marks one end or the other of the text in which we're interested.  Searching forwards or backwards are movements in <strong>vi<\/strong>.  Thus they can also be used as \"subjects\" in our \"statements.\"  So I can use <code>d\/foo<\/code> to cut from the current line to the next line containing the string \"foo\" and <code>y?bar<\/code> to copy from the current line to the most recent (previous) line containing \"bar.\"  If I don't want whole lines I can still use the search movements (as statements of their own), drop my mark(s) and use the <code>`x<\/code> commands as described previously.<\/p>\n<p>In addition to \"verbs\" and \"subjects\" <strong>vi<\/strong> also has \"objects\" (in the grammatical sense of the term).  So far I've only described the use of the anonymous register.  However, I can use any of the 26 \"named\" registers by <em>prefixing<\/em> the \"object\" reference with <code>\"<\/code> (the double quote modifier).  Thus if I use <code>\"add<\/code> I'm cutting the current line into the '<em>a<\/em>' register and if I use <code>\"by\/foo<\/code> then I'm yanking a copy of the text from here to the next line containing \"foo\" into the '<em>b<\/em>' register.  To paste from a register I simply prefix the paste with the same modifier sequence: <code>\"ap<\/code> pastes a copy of the '<em>a<\/em>' register's contents into the text after the cursor and <code>\"bP<\/code> pastes a copy from '<em>b<\/em>' to before the current line.<\/p>\n<p>This notion of \"prefixes\" also adds the analogs of grammatical \"adjectives\" and \"adverbs'  to our text manipulation \"language.\"  Most commands (verbs) and movement (verbs or objects, depending on context) can also take numeric prefixes. Thus <code>3J<\/code> means \"join the next three lines\" and <code>d5}<\/code> means \"delete from the current line through the end of the fifth paragraph down from here.\"<\/p>\n<p>This is all intermediate level <strong>vi<\/strong>.  None of it is <strong>Vim<\/strong> specific and there are far more advanced tricks in <strong>vi<\/strong> if you're ready to learn them.  If you were to master just these intermediate concepts then you'd probably find that you rarely need to write any macros because the text manipulation language is sufficiently concise and expressive to do most things easily enough using the editor's \"native\" language.<\/p>\n<hr \/><h3>A sampling of more advanced tricks:<\/h3>\n<p>There are a number of <code>:<\/code> commands, most notably the <code>:% s\/foo\/bar\/g<\/code> global substitution technique.  (That's not advanced but other <code>:<\/code> commands can be).  The whole <code>:<\/code> set of commands was historically inherited by <strong>vi<\/strong>'s previous incarnations as the <strong>ed<\/strong> (line editor) and later the <strong>ex<\/strong> (extended line editor) utilities.  In fact <strong>vi<\/strong> is so named because it's the visual interface to <strong>ex<\/strong>.<\/p>\n<p><code>:<\/code> commands normally operate over lines of text.  <strong>ed<\/strong> and <strong>ex<\/strong> were written in an era when terminal screens were uncommon and many terminals were \"teletype\" (TTY) devices.  So it was common to work from printed copies of the text, using commands through an extremely terse interface (common connection speeds were 110 baud, or, roughly, 11 characters per second -- which is slower than a fast typist; lags were common on multi-user interactive sessions; additionally there was often some motivation to conserve paper).<\/p>\n<p>So the syntax of most <code>:<\/code> commands includes an address or range of addresses (line number) followed by a command.  Naturally one could use literal line numbers: <code>:127,215 s\/foo\/bar<\/code> to change the first occurrence of \"foo\" into \"bar\" on each line between 127 and 215.  One could also use some abbreviations such as <code>.<\/code> or <code>$<\/code> for current and last lines respectively.  One could also use relative prefixes <code>+<\/code> and <code>-<\/code> to refer to offsets after or before the curent line, respectively.  Thus: <code>:.,$j<\/code> meaning \"from the current line to the last line, join them all into one line\".  <code>:%<\/code> is synonymous with <code>:1,$<\/code> (all the lines).<\/p>\n<p>The <code>:... g<\/code> and <code>:... v<\/code> commands bear some explanation as they are incredibly powerful.  <code>:... g<\/code> is a prefix for \"globally\" applying a subsequent command to all lines which match a pattern (regular expression) while <code>:... v<\/code> applies such a command to all lines which do NOT match the given pattern (\"v\" from \"conVerse\").  As with other <strong>ex<\/strong> commands these can be prefixed by addressing\/range references.  Thus <code>:.,+21g\/foo\/d<\/code> means \"delete any lines containing the string \"foo\" from the current one through the next 21 lines\" while <code>:.,$v\/bar\/d<\/code> means \"from here to the end of the file, delete any lines which DON'T contain the string \"bar.\"<\/p>\n<p>It's interesting that the common Unix command <strong>grep<\/strong> was actually inspired by this <strong>ex<\/strong> command (and is named after the way in which it was documented).  The <strong>ex<\/strong> command <code>:g\/re\/p<\/code> (grep) was the way they documented how to \"globally\" \"print\" lines containing a \"regular expression\" (re).  When <strong>ed<\/strong> and <strong>ex<\/strong> were used, the <code>:p<\/code> command was one of the first that anyone learned and often the first one used when editing any file.  It was how you printed the current contents (usually just one page full at a time using <code>:.,+25p<\/code> or some such).<\/p>\n<p>Note that <code>:% g\/...\/d<\/code> or (its reVerse\/conVerse counterpart: <code>:% v\/...\/d<\/code> are the most common usage patterns.  However there are  couple of other <code>ex<\/code> commands which are worth remembering:<\/p>\n<p>We can use <code>m<\/code> to move lines around, and <code>j<\/code> to join lines.  For example if you have a list and you want to separate all the stuff matching (or conversely NOT matching some pattern) without deleting them, then you can use something like: <code>:% g\/foo\/m$<\/code> ... and all the \"foo\" lines will have been moved to the end of the file.  (Note the other tip about using the end of your file as a scratch space).  This will have preserved the relative order of all the \"foo\" lines while having extracted them from the rest of the list.  (This would be equivalent to doing something like: <code>1G!GGmap!Ggrep foo&lt;ENTER&gt;1G:1,'a g\/foo'\/d<\/code> (copy the file to its own tail, filter the tail through <code>grep<\/code>, and delete all the stuff from the head).<\/p>\n<p>To join lines usually I can find a pattern for all the lines which need to be joined to their predecessor (all the lines which start with \"^   \" rather than \"^ * \" in some bullet list, for example).  For that case I'd use: <code>:% g\/^   \/-1j<\/code> (for every matching line, go up one line and join them). (BTW: for bullet lists trying to search for the bullet lines and join to the next doesn't work for a couple reasons ... it can join one bullet line to another, and it won't join any bullet line to <em>all<\/em> of its continuations; it'll only work pairwise on the matches).<\/p>\n<p>Almost needless to mention you can use our old friend <code>s<\/code> (substitute) with the <code>g<\/code> and <code>v<\/code> (global\/converse-global) commands.  Usually you don't need to do so.  However, consider some case where you want to perform a substitution only on lines matching some other pattern.  Often you can use a complicated pattern with captures and use back references to preserve the portions of the lines that you DON'T want to change.  However, it will often be easier to separate the match from the substitution: <code>:% g\/foo\/s\/bar\/zzz\/g<\/code> -- for every line containing \"foo\" substitute all \"bar\" with \"zzz.\"  (Something like <code>:% s\/\\(.*foo.*\\)bar\\(.*\\)\/\\1zzz\\2\/g<\/code> would only work for the cases those instances of \"bar\" which were PRECEDED by \"foo\" on the same line; it's ungainly enough already, and would have to be mangled further to catch all the cases where \"bar\" preceded \"foo\")<\/p>\n<p>The point is that there are more than just <code>p<\/code>, <code>s<\/code>, and <code>d<\/code> lines in the <code>ex<\/code> command set. <\/p>\n<p>The <code>:<\/code> addresses can also refer to marks.  Thus you can use: <code>:'a,'bg\/foo\/j<\/code> to join any line containing the string foo to its subsequent line, if it lies between the lines between the '<em>a<\/em>' and '<em>b<\/em>' marks.  (Yes, all of the preceding <code>ex<\/code> command examples can be limited to subsets of the file's lines by prefixing with these sorts of addressing expressions).<\/p>\n<p>That's pretty obscure (I've only used something like that a few times in the last 15 years). However, I'll freely admit that I've often done things iteratively and interactively that could probably have been done more efficiently if I'd taken the time to think out the correct incantation.<\/p>\n<p>Another very useful <strong>vi<\/strong> or <strong>ex<\/strong> command is <code>:r<\/code> to read in the contents of another file.  Thus: <code>:r foo<\/code> inserts the contents of the file named \"foo\" at the current line.  <\/p>\n<p>More powerful is the <code>:r!<\/code> command.  This reads the results of a command.  It's the same as suspending the <strong>vi<\/strong> session, running a command, redirecting its output to a temporary file, resuming your <strong>vi<\/strong> session, and reading in the contents from the temp. file.<\/p>\n<p>Even more powerful are the <code>!<\/code> (bang) and <code>:... !<\/code> (<strong>ex<\/strong> bang) commands.  These also execute external commands and read the results into the current text. However, they also filter selections of our text through the command!  This we can sort all the lines in our file using <code>1G!Gsort<\/code> (<code>G<\/code> is the <strong>vi<\/strong> \"goto\" command; it defaults to going to the last line of the file, but can be prefixed by a line number, such as 1, the first line).  This is equivalent to the <strong>ex<\/strong> variant <code>:1,$!sort<\/code>.  Writers often use <code>!<\/code> with the Unix <strong>fmt<\/strong> or <strong>fold<\/strong> utilities for reformating or \"word wrapping\" selections of text.  A very common macro is <code>{!}fmt<\/code> (reformat the current paragraph).  Programmers sometimes use it to run their code, or just portions of it, through <strong>indent<\/strong> or other code reformatting tools.<\/p>\n<p>Using the <code>:r!<\/code> and <code>!<\/code> commands means that any external utility or filter can be treated as an extension of our editor.  I have occasionally used these with scripts that pulled data from a database, or with <strong>wget<\/strong> or <strong>lynx<\/strong> commands that pulled data off a website, or <strong>ssh<\/strong> commands that pulled data from remote systems.<\/p>\n<p>Another useful <strong>ex<\/strong> command is <code>:so<\/code> (short for <code>:source<\/code>).  This reads the contents of a file as a series of commands.  When you start <strong>vi<\/strong> it normally, implicitly, performs a <code>:source<\/code> on <code>~\/.exinitrc<\/code> file (and <strong>Vim<\/strong> usually does this on <code>~\/.vimrc<\/code>, naturally enough).  The use of this is that you can change your editor profile on the fly by simply sourcing in a new set of macros, abbreviations, and editor settings.  If you're sneaky you can even use this as a trick for storing sequences of <strong>ex<\/strong> editing commands to apply to files on demand.<\/p>\n<p>For example I have a seven line file (36 characters) which runs a file through <strong>wc<\/strong>, and inserts a C-style comment at the top of the file containing that word count data.  I can apply that \"macro\" to a file by using a command like: <code>vim +'so mymacro.ex' .\/mytarget<\/code><\/p>\n<p>(The <code>+<\/code> command line option to <strong>vi<\/strong> and <strong>Vim<\/strong> is normally used to start the editing session at a given line number.  However it's a little known fact that one can follow the <code>+<\/code> by any valid <strong>ex<\/strong> command\/expression, such as a \"source\" command as I've done here; for a simple example I have scripts which invoke: <code>vi +'\/foo\/d|wq!' ~\/.ssh\/known_hosts<\/code> to remove an entry from my SSH known hosts file non-interactively while I'm re-imaging a set of servers).<\/p>\n<p>Usually it's far easier to write such \"macros\" using Perl, AWK, <strong>sed<\/strong> (which is, in fact, like <strong>grep<\/strong> a utility inspired by the <strong>ed<\/strong> command).<\/p>\n<p>The <code>@<\/code> command is probably the most obscure <strong>vi<\/strong> command.  In occasionally teaching advanced systems administration courses for close to a decade I've met very few people who've ever used it.  <code>@<\/code> executes the contents of a register as if it were a <strong>vi<\/strong> or <strong>ex<\/strong> command.<br \/>Example: I often use: <code>:r!locate ...<\/code> to find some file on my system and read its name into my document.  From there I delete any extraneous hits, leaving only the full path to the file I'm interested in.  Rather than laboriously <kbd>Tab<\/kbd>-ing through each component of the path (or worse, if I happen to be stuck on a machine without Tab completion support in its copy of <strong>vi<\/strong>) I just use:<\/p>\n<ol><li><code>0i:r<\/code> (to turn the current line into a valid <strong>:r<\/strong> command),<\/li>\n<li><code>\"cdd<\/code> (to delete the line into the \"c\" register) and<\/li>\n<li><code>@c<\/code> execute that command.<\/li>\n<\/ol><p>That's only 10 keystrokes (and the expression <code>\"cdd<\/code> <code>@c<\/code> is effectively a finger macro for me, so I can type it almost as quickly as any common six letter word).<\/p>\n<hr \/><h3>A sobering thought<\/h3>\n<p>I've only scratched to surface of <strong>vi<\/strong>'s power and none of what I've described here is even part of the \"improvements\" for which <strong>vim<\/strong> is named!  All of what I've described here should work on any old copy of <strong>vi<\/strong> from 20 or 30 years ago.<\/p>\n<p>There are people who have used considerably more of <strong>vi<\/strong>'s power than I ever will. <\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1220118\/revisions\" title=\"show all edits to this post\">edited Dec 22 '11 at 11:43<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1220118\" href=\"https:\/\/stackoverflow.com\/posts\/1220118\/revisions\" title=\"show revision history for this post\">&#13;\n        18 revs, 16 users 64%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/149076\">Jim Dennis<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        505&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>You are talking about text selecting and copying, I think that you should give a look to the <a href=\"http:\/\/www.vim.org\/htmldoc\/visual.html\" rel=\"nofollow noreferrer\">Vim Visual Mode<\/a>.<\/p>\n<p>In the visual mode, you are able to select text using Vim commands, then you can do whatever you want with the selection.<\/p>\n<p>Consider the following common scenarios:<\/p>\n<p>You need to select to the next matching parenthesis.<\/p>\n<p>You could do:<\/p>\n<ul><li><strong><code>v%<\/code><\/strong> if the cursor is on the starting\/ending parenthesis<\/li>\n<li><strong><code>vib<\/code><\/strong> if the cursor is inside the parenthesis block<\/li>\n<\/ul><p><img src=\"https:\/\/dl.dropbox.com\/u\/35146\/stackoverflow\/Vim\/parens.gif\" alt=\"image\" \/><\/p>\n<p>You want to select text between quotes:<\/p>\n<ul><li><strong>vi\"<\/strong> for double quotes <\/li>\n<li><strong>vi'<\/strong> for single quotes<\/li>\n<\/ul><p><img src=\"https:\/\/dl.dropbox.com\/u\/35146\/stackoverflow\/Vim\/quotes.gif\" alt=\"image\" \/><\/p>\n<p>You want to select a curly brace block (very common on C-style languages):<\/p>\n<ul><li><strong><code>viB<\/code><\/strong><\/li>\n<li><strong><code>vi{<\/code><\/strong><\/li>\n<\/ul><p><img src=\"https:\/\/dl.dropbox.com\/u\/35146\/stackoverflow\/Vim\/viB.gif\" alt=\"image\" \/><\/p>\n<p>You want to select the entire file:<\/p>\n<p><a href=\"http:\/\/vimdoc.sourceforge.net\/htmldoc\/visual.html#CTRL-V\" rel=\"nofollow noreferrer\">Visual block selection<\/a> is another really useful feature, it allows you to select a rectangular area of text, you just have to press <kbd>Ctrl<\/kbd>-<kbd>V<\/kbd> to start it, and then select the text block you want and perform any type of operation such as yank, delete, paste, edit, etc. It's great to edit <em>column oriented<\/em> text.<\/p>\n<p><img src=\"https:\/\/dl.dropbox.com\/u\/35146\/stackoverflow\/Vim\/blockselection.gif\" alt=\"image\" \/><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218429\/revisions\" title=\"show all edits to this post\">edited Feb 8 '17 at 14:14<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Aug 2 '09 at 8:27&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/5445\/cms\">CMS<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            579k158833808&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        207&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Some productivity tips:<\/p>\n<p><strong>Smart movements<\/strong><\/p>\n<ul><li><code>*<\/code> and  <code>#<\/code> search for the word under the cursor forward\/backward.<\/li>\n<li><code>w<\/code> to the next word<\/li>\n<li><code>W<\/code> to the next space-separated word<\/li>\n<li><code>b<\/code> \/ <code>e<\/code> to the begin\/end of the current word. (<code>B<\/code> \/ <code>E<\/code> for space separated only)<\/li>\n<li><code>gg<\/code> \/ <code>G<\/code> jump to the begin\/end of the file.<\/li>\n<li><code>%<\/code> jump to the matching { .. } or ( .. ), etc..<\/li>\n<li><code>{<\/code> \/ <code>}<\/code> jump to next paragraph.<\/li>\n<li><code>'.<\/code> jump back to last edited line.<\/li>\n<li><code>g;<\/code> jump back to last edited position.<\/li>\n<\/ul><p><strong>Quick editing commands<\/strong><\/p>\n<ul><li><code>I<\/code> insert at the begin.<\/li>\n<li><code>A<\/code> append to end.<\/li>\n<li><code>o<\/code> \/ <code>O<\/code> open a new line after\/before the current.<\/li>\n<li><code>v<\/code> \/ <code>V<\/code> \/ <code>Ctrl+V<\/code> visual mode (to select text!)<\/li>\n<li><code>Shift+R<\/code> replace text<\/li>\n<li><code>C<\/code> change remaining part of line.<\/li>\n<\/ul><p><strong>Combining commands<\/strong><\/p>\n<p>Most commands accept a amount and direction, for example:<\/p>\n<ul><li><code>cW<\/code> = change till end of word<\/li>\n<li><code>3cW<\/code> = change 3 words<\/li>\n<li><code>BcW<\/code> = to begin of full word, change full word<\/li>\n<li><code>ciW<\/code> = change inner word.<\/li>\n<li><code>ci\"<\/code> = change inner between \"..\"<\/li>\n<li><code>ci(<\/code> = change text between ( .. )<\/li>\n<li><code>ci&lt;<\/code> = change text between &lt; .. &gt;  (needs <code>set matchpairs+=&lt;:&gt;<\/code> in vimrc)<\/li>\n<li><code>4dd<\/code> = delete 4 lines<\/li>\n<li><code>3x<\/code> = delete 3 characters.<\/li>\n<li><code>3s<\/code> = substitute 3 characters.<\/li>\n<\/ul><p><strong>Useful programmer commands<\/strong><\/p>\n<ul><li><code>r<\/code> replace one character (e.g. <code>rd<\/code> replaces the current char with <code>d<\/code>).<\/li>\n<li><code>~<\/code> changes case.<\/li>\n<li><code>J<\/code> joins two lines<\/li>\n<li>Ctrl+A \/ Ctrl+X increments\/decrements a number.<\/li>\n<li><code>.<\/code> repeat last command (a simple macro)<\/li>\n<li><code>==<\/code> fix line indent<\/li>\n<li><code>&gt;<\/code> indent block (in visual mode)<\/li>\n<li><code>&lt;<\/code> unindent block (in visual mode)<\/li>\n<\/ul><p><strong>Macro recording<\/strong><\/p>\n<ul><li>Press <code>q[ key ]<\/code> to start recording.<\/li>\n<li>Then hit <code>q<\/code> to stop recording.<\/li>\n<li>The macro can be played with <code>@[ key ]<\/code>.<\/li>\n<\/ul><p>By using very specific commands and movements, VIM can replay those exact actions for the next lines. (e.g. A for append-to-end, <code>b<\/code> \/ <code>e<\/code> to move the cursor to the begin or end of a word respectively)<\/p>\n<p><strong>Example of well built settings<\/strong><\/p>\n<pre># reset to vim-defaults\nif &amp;compatible          # only if not set before:\n  set nocompatible      # use vim-defaults instead of vi-defaults (easier, more user friendly)\nendif\n# display settings\nset background=dark     # enable for dark terminals\nset nowrap              # dont wrap lines\nset scrolloff=2         # 2 lines above\/below cursor when scrolling\nset number              # show line numbers\nset showmatch           # show matching bracket (briefly jump)\nset showmode            # show mode in status bar (insert\/replace\/...)\nset showcmd             # show typed command in status bar\nset ruler               # show cursor position in status bar\nset title               # show file in titlebar\nset wildmenu            # completion with menu\nset wildignore=*.o,*.obj,*.bak,*.exe,*.py[co],*.swp,*~,*.pyc,.svn\nset laststatus=2        # use 2 lines for the status bar\nset matchtime=2         # show matching bracket for 0.2 seconds\nset matchpairs+=&lt;:&gt;     # specially for html\n# editor settings\nset esckeys             # map missed escape sequences (enables keypad keys)\nset ignorecase          # case insensitive searching\nset smartcase           # but become case sensitive if you type uppercase characters\nset smartindent         # smart auto indenting\nset smarttab            # smart tab handling for indenting\nset magic               # change the way backslashes are used in search patterns\nset bs=indent,eol,start # Allow backspacing over everything in insert mode\nset tabstop=4           # number of spaces a tab counts for\nset shiftwidth=4        # spaces for autoindents\n#set expandtab           # turn a tabs into spaces\nset fileformat=unix     # file mode is unix\n#set fileformats=unix,dos    # only detect unix file format, displays that ^M with dos files\n# system settings\nset lazyredraw          # no redraws in macros\nset confirm             # get a dialog when :q, :w, or :wq fails\nset nobackup            # no backup~ files.\nset viminfo='20,\\\"500   # remember copy registers after quitting in the .viminfo file -- 20 jump links, regs up to 500 lines'\nset hidden              # remember undo after quitting\nset history=50          # keep 50 lines of command history\nset mouse=v             # use mouse in visual mode (not normal,insert,command,help mode\n# color settings (if terminal\/gui supports it)\nif &amp;t_Co &gt; 2 || has(\"gui_running\")\n  syntax on          # enable colors\n  set hlsearch       # highlight search (very useful!)\n  set incsearch      # search incremently (search while typing)\nendif\n# paste mode toggle (needed when using autoindent\/smartindent)\nmap &lt;F10&gt; :set paste&lt;CR&gt;\nmap &lt;F11&gt; :set nopaste&lt;CR&gt;\nimap &lt;F10&gt; &lt;C-O&gt;:set paste&lt;CR&gt;\nimap &lt;F11&gt; &lt;nop&gt;\nset pastetoggle=&lt;F11&gt;\n# Use of the filetype plugins, auto completion and indentation support\nfiletype plugin indent on\n# file type specific settings\nif has(\"autocmd\")\n  # For debugging\n  #set verbose=9\n  # if bash is sh.\n  let bash_is_sh=1\n  # change to directory of current file automatically\n  autocmd BufEnter * lcd %:p:h\n  # Put these in an autocmd group, so that we can delete them easily.\n  augroup mysettings\n    au FileType xslt,xml,css,html,xhtml,javascript,sh,config,c,cpp,docbook set smartindent shiftwidth=2 softtabstop=2 expandtab\n    au FileType tex set wrap shiftwidth=2 softtabstop=2 expandtab\n    # Confirm to PEP8\n    au FileType python set tabstop=4 softtabstop=4 expandtab shiftwidth=4 cinwords=if,elif,else,for,while,try,except,finally,def,class\n  augroup END\n  augroup perl\n    # reset (disable previous 'augroup perl' settings)\n    au!  \n    au BufReadPre,BufNewFile\n    \\ *.pl,*.pm\n    \\ set formatoptions=croq smartindent shiftwidth=2 softtabstop=2 cindent cinkeys='0{,0},!^F,o,O,e' \" tags=.\/tags,tags,~\/devel\/tags,~\/devel\/C\n    # formatoption:\n    #   t - wrap text using textwidth\n    #   c - wrap comments using textwidth (and auto insert comment leader)\n    #   r - auto insert comment leader when pressing &lt;return&gt; in insert mode\n    #   o - auto insert comment leader when pressing 'o' or 'O'.\n    #   q - allow formatting of comments with \"gq\"\n    #   a - auto formatting for paragraphs\n    #   n - auto wrap numbered lists\n    #   \n  augroup END\n  # Always jump to the last known cursor position. \n  # Don't do it when the position is invalid or when inside\n  # an event handler (happens when dropping a file on gvim). \n  autocmd BufReadPost * \n    \\ if line(\"'\\\"\") &gt; 0 &amp;&amp; line(\"'\\\"\") &lt;= line(\"$\") | \n    \\   exe \"normal g`\\\"\" | \n    \\ endif \nendif # has(\"autocmd\")\n<\/pre>\n<p>The settings can be stored in <code>~\/.vimrc<\/code>, or system-wide in <code>\/etc\/vimrc.local<\/code> and then by read from the <code>\/etc\/vimrc<\/code> file using:<\/p>\n<pre>source \/etc\/vimrc.local\n<\/pre>\n<p>(you'll have to replace the <code>#<\/code> comment character with <code>\"<\/code> to make it work in VIM, I wanted to give proper syntax highlighting here).<\/p>\n<p>The commands I've listed here are pretty basic, and the main ones I use so far. They already make me quite more productive, without having to know all the fancy stuff.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2559262\/revisions\" title=\"show all edits to this post\">edited Oct 29 '12 at 18:51<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2559262\" href=\"https:\/\/stackoverflow.com\/posts\/2559262\/revisions\" title=\"show revision history for this post\">&#13;\n        12 revs, 3 users 99%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/146289\">vdboor<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        98&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>The <kbd>Control<\/kbd>+<kbd>R<\/kbd> mechanism is very useful :-)  In either <strong>insert mode<\/strong> or <strong>command mode<\/strong> (i.e. on the <code>:<\/code> line when typing commands), continue with a numbered or named register:<br \/><\/p>\n<ul><li><kbd>a<\/kbd> - <kbd>z<\/kbd> the named registers<\/li>\n<li><kbd>\"<\/kbd> the unnamed register, containing the text of the last delete or yank<br \/><\/li>\n<li><kbd>%<\/kbd> the current file name<br \/><\/li>\n<li><kbd>#<\/kbd> the alternate file name<br \/><\/li>\n<li><kbd>*<\/kbd> the clipboard contents (X11: primary selection)<br \/><\/li>\n<li><kbd>+<\/kbd> the clipboard contents<br \/><\/li>\n<li><kbd>\/<\/kbd> the last search pattern<br \/><\/li>\n<li><kbd>:<\/kbd> the last command-line<br \/><\/li>\n<li><kbd>.<\/kbd> the last inserted text<br \/><\/li>\n<li><kbd>-<\/kbd> the last small (less than a line) delete<br \/><\/li>\n<li><kbd>=5*5<\/kbd> insert 25 into text (mini-calculator)<\/li>\n<\/ul><p>See <code>:help i_CTRL-R<\/code> and <code>:help c_CTRL-R<\/code> for more details, and snoop around nearby for more CTRL-R goodness.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2960630\/revisions\" title=\"show all edits to this post\">edited Apr 12 '12 at 7:46<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2960630\" href=\"https:\/\/stackoverflow.com\/posts\/2960630\/revisions\" title=\"show revision history for this post\">&#13;\n        5 revs, 4 users 53%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/348785\">kev<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        91&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>There are a lot of good answers here, and one amazing one about the zen of vi.  One thing I don't see mentioned is that vim is extremely extensible via plugins.  There are scripts and plugins to make it do all kinds of crazy things the original author never considered.  Here are a few examples of incredibly handy vim plugins: <\/p>\n<p>Rails.vim is a plugin written by tpope. It's an incredible tool for people doing rails development.  It does magical context-sensitive things that allow you to easily jump from a method in a controller to the associated view, over to a model, and down to unit tests for that model.  It has saved dozens if not hundreds of hours as a rails developer. <\/p>\n<p>This plugin allows you to select a region of text in visual mode and type a quick command to post it to <a href=\"http:\/\/gist.github.com\" rel=\"noreferrer\">gist.github.com<\/a>.  This allows for easy pastebin access, which is incredibly handy if you're collaborating with someone over IRC or IM.  <\/p>\n<p>This plugin provides special functionality to the spacebar.  It turns the spacebar into something analogous to the period, but instead of repeating actions it repeats motions.  This can be very handy for moving quickly through a file in a way you define on the fly. <\/p>\n<p>This plugin gives you the ability to work with text that is delimited in some fashion.  It gives you objects which denote things inside of parens, things inside of quotes, etc.  It can come in handy for manipulating delimited text. <\/p>\n<p>This script brings fancy tab completion functionality to vim.  The autocomplete stuff is already there in the core of vim, but this brings it to a quick tab rather than multiple different multikey shortcuts.  Very handy, and incredibly fun to use.  While it's not VS's intellisense, it's a great step and brings a great deal of the functionality you'd like to expect from a tab completion tool.<\/p>\n<p>This tool brings external syntax checking commands into vim.  I haven't used it personally, but I've heard great things about it and the concept is hard to beat.  Checking syntax without having to do it manually is a great time saver and can help you catch syntactic bugs as you introduce them rather than when you finally stop to test. <\/p>\n<p>Direct access to git from inside of vim.  Again, I haven't used this plugin, but I can see the utility.  Unfortunately I'm in a culture where svn is considered \"new\", so I won't likely see git at work for quite some time. <\/p>\n<p>A tree browser for vim.  I started using this recently, and it's really handy.  It lets you put a treeview in a vertical split and open files easily.  This is great for a project with a lot of source files you frequently jump between.  <\/p>\n<p>This is an unmaintained plugin, but still incredibly useful.  It provides the ability to open files using a \"fuzzy\" descriptive syntax.  It means that in a sparse tree of files you need only type enough characters to disambiguate the files you're interested in from the rest of the cruft.  <\/p>\n<h2>Conclusion<\/h2>\n<p>There are a lot of incredible tools available for vim.  I'm sure I've only scratched the surface here, and it's well worth searching for tools applicable to your domain.  The combination of traditional vi's powerful toolset, vim's improvements on it, and plugins which extend vim even further, it's one of the most powerful ways to edit text ever conceived.  Vim is easily as powerful as emacs, eclipse, visual studio, and textmate.  <\/p>\n<h2>Thanks<\/h2>\n<p>Thanks to <a href=\"http:\/\/twitter.com\/duwanis\" rel=\"noreferrer\">duwanis<\/a> for his <a href=\"http:\/\/github.com\/duwanis\/vim_config\" rel=\"noreferrer\">vim configs<\/a> from which I have learned much and borrowed most of the plugins listed here. <\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/2558011\/revisions\" title=\"show all edits to this post\">Apr 1 '10 at 3:44<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2558011\" href=\"https:\/\/stackoverflow.com\/posts\/2558011\/revisions\" title=\"show revision history for this post\">&#13;\n        Benson&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        53&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><code>.<\/code> Repeat last text-changing command<\/p>\n<p>I save a lot of time with this one.<\/p>\n<p>Visual mode was mentioned previously, but block visual mode has saved me a lot of time when editing fixed size columns in text file.  (accessed with Ctrl-V).<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2347142\/revisions\" title=\"show all edits to this post\">edited Mar 31 '10 at 23:01<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2347142\" href=\"https:\/\/stackoverflow.com\/posts\/2347142\/revisions\" title=\"show revision history for this post\">&#13;\n        4 revs, 2 users 89%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/227736\">Cooper6581<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        48&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong>gi<\/strong><\/p>\n<p>Go to last edited location (very useful if you performed some searching and than want go back to edit)<\/p>\n<p><strong>^P and ^N<\/strong><\/p>\n<p>Complete previous (^P) or next (^N) text. <\/p>\n<p><strong>^O and ^I<\/strong><\/p>\n<p>Go to previous (<code>^O<\/code> - <code>\"O\"<\/code> for old) location or to the next (<code>^I<\/code> - <code>\"I\"<\/code> just near to <code>\"O\"<\/code>).\nWhen you perform searches, edit files etc., you can navigate through these \"jumps\" forward and back.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218586\/revisions\" title=\"show all edits to this post\">edited Dec 24 '12 at 14:50<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218586\" href=\"https:\/\/stackoverflow.com\/posts\/1218586\/revisions\" title=\"show revision history for this post\">&#13;\n        3 revs, 3 users 87%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/135960\">dimba<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        46&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I recently (got) discovered this site: <a href=\"http:\/\/vimcasts.org\/\" rel=\"noreferrer\">http:\/\/vimcasts.org\/<\/a><\/p>\n<p>It's pretty new and really really good. The guy who is running the site switched from textmate to vim and hosts very good and concise casts on specific vim topics. Check it out!<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/2555940\/revisions\" title=\"show all edits to this post\">Mar 31 '10 at 19:37<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2555940\" href=\"https:\/\/stackoverflow.com\/posts\/2555940\/revisions\" title=\"show revision history for this post\">&#13;\n        Ronny Brendel&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        42&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p><strong><kbd>CTRL<\/kbd> + <kbd>A<\/kbd><\/strong> increments the number you are standing on.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218957\/revisions\" title=\"show all edits to this post\">edited Feb 27 '10 at 11:20<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218957\" href=\"https:\/\/stackoverflow.com\/posts\/1218957\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 67%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/17916\">hcs42<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        38&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>All in Normal mode:<\/p>\n<p><strong>f&lt;char&gt;<\/strong> to move to the next instance of a particular character on the current line, and <strong>;<\/strong> to repeat.<\/p>\n<p><strong>F&lt;char&gt;<\/strong> to move to the previous instance of a particular character on the current line and <strong>;<\/strong> to repeat.<\/p>\n<p>If used intelligently, the above two can make you killer-quick moving around in a line.<\/p>\n<p><strong>*<\/strong> on a word to search for the next instance.<\/p>\n<p><strong>#<\/strong> on a word to search for the previous instance.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1219591\/revisions\" title=\"show all edits to this post\">edited Aug 28 '09 at 15:23<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1219591\" href=\"https:\/\/stackoverflow.com\/posts\/1219591\/revisions\" title=\"show revision history for this post\">&#13;\n        3 revs<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/26054\">Eric Smith<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        34&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<h2>Session<\/h2>\n<p>a. save session<\/p>\n<blockquote>\n  <p>:mks <em>sessionname<\/em><\/p>\n<\/blockquote>\n<p>b. force save session<\/p>\n<blockquote>\n  <p>:mks! <em>sessionname<\/em><\/p>\n<\/blockquote>\n<p>c. load session<\/p>\n<blockquote>\n  <p>gvim or vim -S <em>sessionname<\/em><\/p>\n<\/blockquote>\n<hr \/><h2>Adding and Subtracting<\/h2>\n<p>a. Adding and Subtracting<\/p>\n<blockquote>\n  <p>CTRL-A        ;Add [count] to the number or alphabetic character at\n               or after the cursor.  {not in Vi<\/p>\n  \n  <p>CTRL-X        ;Subtract [count] from the number or alphabetic\n              character at or after the cursor.  {not in Vi}<\/p>\n<\/blockquote>\n<p>b. Window key unmapping<\/p>\n<blockquote>\n  <p>In window,\n  Ctrl-A already mapped for whole file selection you need to unmap in rc file.\n  mark mswin.vim CTRL-A mapping part as comment \n  or \n  add your rc file with  unmap <\/p>\n<\/blockquote>\n<p>c. With Macro<\/p>\n<blockquote>\n  <p>The CTRL-A command is very useful in a macro.  Example: \n  Use the following steps to make a numbered list.<\/p>\n  \n  <ol><li>Create the first list entry, make sure it starts with a number.<\/li>\n  <li>qa      - start recording into buffer 'a'<\/li>\n  <li>Y      - yank the entry<\/li>\n  <li>p      - put a copy of the entry below the first one<\/li>\n  <li>CTRL-A    - increment the number<\/li>\n  <li>q      - stop recording<\/li>\n  <li>@a - repeat the yank, put and increment  times<\/li>\n  <\/ol><\/blockquote><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/3519848\/revisions\" title=\"show all edits to this post\">Aug 19 '10 at 8:08<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-3519848\" href=\"https:\/\/stackoverflow.com\/posts\/3519848\/revisions\" title=\"show revision history for this post\">&#13;\n        agfe2&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        32&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Last week at work our project inherited a lot of Python code from another project.  Unfortunately the code did not fit into our existing architecture - it was all done with global variables and functions, which would not work in a multi-threaded environment.  <\/p>\n<p>We had ~80 files that needed to be reworked to be object oriented - all the functions moved into classes, parameters changed, import statements added, etc.  We had a list of about 20 types of fix that needed to be done to each file.  I would estimate that doing it by hand one person could do maybe 2-4 per day.<\/p>\n<p>So I did the first one by hand and then wrote a vim script to automate the changes.  Most of it was a list of vim commands e.g.<\/p>\n<pre>\" delete an un-needed function \"\ng\/someFunction(\/ d\n\" add wibble parameter to function foo \"\n%s\/foo(\/foo( wibble,\/\n\" convert all function calls bar(thing) into method calls thing.bar() \"\ng\/bar(\/ normal nmaf(ldi(`aPa.\n<\/pre>\n<p>The last one deserves a bit of explanation:<\/p>\n<pre>g\/bar(\/  executes the following command on every line that contains \"bar(\"\nnormal   execute the following text as if it was typed in in normal mode\nn        goes to the next match of \"bar(\" (since the :g command leaves the cursor position at the start of the line)\nma       saves the cursor position in mark a\nf(       moves forward to the next opening bracket\nl        moves right one character, so the cursor is now inside the brackets\ndi(      delete all the text inside the brackets\n`a       go back to the position saved as mark a (i.e. the first character of \"bar\")\nP        paste the deleted text before the current cursor position\na.       go into insert mode and add a \".\" \n<\/pre>\n<p>For a couple of more complex transformations such as generating all the import statements I embedded some python into the vim script.<\/p>\n<p>After a few hours of working on it I had a script that will do at least 95% of the conversion.  I just open a file in vim then run <code>:source fixit.vim<\/code> and the file is transformed in a blink of the eye.<\/p>\n<p>We still have the work of changing the remaining 5% that was not worth automating and of testing the results, but by spending a day writing this script I estimate we have saved weeks of work.<\/p>\n<p>Of course it would have been possible to automate this with a scripting language like Python or Ruby, but it would have taken far longer to write and would be less flexible - the last example would have been difficult since regex alone would not be able to handle nested brackets, e.g. to convert <code>bar(foo(xxx))<\/code> to <code>foo(xxx).bar()<\/code>.  Vim was perfect for the task.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2347243\/revisions\" title=\"show all edits to this post\">edited Aug 18 '12 at 21:44<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2347243\" href=\"https:\/\/stackoverflow.com\/posts\/2347243\/revisions\" title=\"show revision history for this post\">&#13;\n        8 revs, 2 users 98%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/206417\">Dave Kirby<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        28&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Use the builtin file explorer! The command is <code>:Explore<\/code> and it allows you to navigate through your source code very very fast. I have these mapping in my <code>.vimrc<\/code>:<\/p>\n<pre>map &lt;silent&gt; &lt;F8&gt;   :Explore&lt;CR&gt;\nmap &lt;silent&gt; &lt;S-F8&gt; :sp +Explore&lt;CR&gt;\n<\/pre>\n<p>The explorer allows you to make file modifications, too. I'll post some of my favorite keys, pressing <code>&lt;F1&gt;<\/code> will give you the full list:<\/p>\n<ul><li><strong>-<\/strong>: The most useful: Change to upper directory (<code>cd ..<\/code>)<\/li>\n<li><strong>mf<\/strong>: Mark a file<\/li>\n<li><strong>D<\/strong>: Delete marked files or the file the cursor is on, if nothing ismarked.<\/li>\n<li><strong>R<\/strong>: Rename the file the cursor is on.<\/li>\n<li><strong>d<\/strong>: Create a new directory in the current directory<\/li>\n<li><strong>%<\/strong>: Create a new file in the current directory<\/li>\n<\/ul><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218442\/revisions\" title=\"show all edits to this post\">edited Aug 2 '09 at 11:17<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218442\" href=\"https:\/\/stackoverflow.com\/posts\/1218442\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/44562\">soulmerge<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        25&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I am a member of the American Cryptogram Association.  The bimonthly magazine includes over 100 cryptograms of various sorts.  Roughly 15 of these are \"cryptarithms\" - various types of arithmetic problems with letters substituted for the digits.  Two or three of these are sudokus, except with letters instead of numbers.  When the grid is completed, the nine distinct letters will spell out a word or words, on some line, diagonal, spiral, etc., somewhere in the grid.<\/p>\n<p>Rather than working with pencil, or typing the problems in by hand, I download the problems from the members area of their website.<\/p>\n<p>When working with these sudokus, I use vi, simply because I'm using facilities that vi has that few other editors have.  Mostly in converting the lettered grid into a numbered grid, because I find it easier to solve, and then the completed numbered grid back into the lettered grid to find the solution word or words.<\/p>\n<p>The problem is formatted as nine groups of nine letters, with <code>-<\/code>s representing the blanks, written in two lines.  The first step is to format these into nine lines of nine characters each.  There's nothing special about this, just inserting eight linebreaks in the appropriate places.<\/p>\n<p>The result will look like this:<\/p>\n<pre>T-O-----C\n-E-----S-\n--AT--N-L\n---NASO--\n---E-T---\n--SPCL---\nE-T--OS--\n-A-----P-\nS-----C-T\n<\/pre>\n<p>So, first step in converting this into numbers is to make a list of the distinct letters.  First, I make a copy of the block.  I position the cursor at the top of the block, then type <code>:y}}p<\/code>.  <code>:<\/code> puts me in command mode, <code>y<\/code> yanks the next movement command.  Since <code>}<\/code> is a move to the end of the next paragraph, <code>y}<\/code> yanks the paragraph.  <code>}<\/code> then moves the cursor to the end of the paragraph, and <code>p<\/code> pastes what we had yanked just after the cursor.  So <code>y}}p<\/code> creates a copy of the next paragraph, and ends up with the cursor between the two copies.<\/p>\n<p>Next, I to turn one of those copies into a list of distinct letters.  That command is a bit more complex:<\/p>\n<pre>:!}tr -cd A-Z | sed 's\/\\(.\\)\/\\1\\n\/g' | sort -u | tr -d '\\n'\n<\/pre>\n<p><code>:<\/code> again puts me in command mode.  <code>!<\/code> indicates that the content of the next yank should be piped through a command line.  <code>}<\/code> yanks the next paragraph, and the command line then uses the <code>tr<\/code> command to strip out everything except for upper-case letters, the <code>sed<\/code> command to print each letter on a single line, and the <code>sort<\/code> command to sort those lines, removing duplicates, and then <code>tr<\/code> strips out the newlines, leaving the nine distinct letters in a single line, replacing the nine lines that had made up the paragraph originally.  In this case, the letters are: <code>ACELNOPST<\/code>.<\/p>\n<p>Next step is to make another copy of the grid.  And then to use the letters I've just identified to replace each of those letters with a digit from 1 to 9.  That's simple: <code>:!}tr ACELNOPST 0-9<\/code>.  The result is:<\/p>\n<pre>8-5-----1\n-2-----7-\n--08--4-3\n---4075--\n---2-8---\n--7613---\n2-8--57--\n-0-----6-\n7-----1-8\n<\/pre>\n<p>This can then be solved in the usual way, or entered into any sudoku solver you might prefer.  The completed solution can then be converted back into letters with <code>:!}tr 1-9 ACELNOPST<\/code>.<\/p>\n<p>There is power in vi that is matched by very few others.  The biggest problem is that only a very few of the vi tutorial books, websites, help-files, etc., do more than barely touch the surface of what is possible.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2565107\/revisions\" title=\"show all edits to this post\">edited Jun 15 '11 at 13:39<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2565107\" href=\"https:\/\/stackoverflow.com\/posts\/2565107\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 92%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/243563\">Jeff Dege<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        23&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Bulk text manipulations!<\/p>\n<p>Either through macros:<\/p>\n<ul><li>Start with recording: <code>qq<\/code><\/li>\n<li>Do stuff<\/li>\n<li>Stop recording: <code>q<\/code><\/li>\n<li>Repeat: <code>@q<\/code> (the first time), <code>@@<\/code> after that.<\/li>\n<li>Repeat 20 times: <code>20@@<\/code><\/li>\n<\/ul><p>Or through regular expressions:<\/p>\n<ul><li>Replace stuff: <code>:%s\/[fo]+\/bar\/g<\/code><\/li>\n<\/ul><p>(But be warned: if you do the latter, you'll have 2 problems :).)<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/1218478\/revisions\" title=\"show all edits to this post\">Aug 2 '09 at 8:59<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218478\" href=\"https:\/\/stackoverflow.com\/posts\/1218478\/revisions\" title=\"show revision history for this post\">&#13;\n        jqno&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        23&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I recently discovered <code>q:<\/code>.  It opens the \"command window\" and shows your most recent ex-mode (command-mode) commands.  You can move as usual within the window, and pressing <code>&lt;CR&gt;<\/code> executes the command.  You can edit, etc. too.  Priceless when you're messing around with some complex command or regex and you don't want to retype the whole thing, or if the complex thing you want to do was 3 commands back.  It's almost like bash's <code>set -o vi<\/code>, but for vim itself (heh!).<\/p>\n<p>See <code>:help q:<\/code> for more interesting bits for going back and forth.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/9972713\/revisions\" title=\"show all edits to this post\">Apr 2 '12 at 7:56<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-9972713\" href=\"https:\/\/stackoverflow.com\/posts\/9972713\/revisions\" title=\"show revision history for this post\">&#13;\n        David Pope&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        22&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I just discovered Vim's omnicompletion the other day, and while I'll admit I'm a bit hazy on what does which, I've had surprisingly good results just mashing either <kbd>Ctrl<\/kbd> + <kbd>x<\/kbd> <kbd>Ctrl<\/kbd> + <kbd>u<\/kbd> or \n<kbd>Ctrl<\/kbd> + <kbd>n<\/kbd>\/<kbd>Ctrl<\/kbd> +<kbd>p<\/kbd> in insert mode. It's not quite <a href=\"http:\/\/en.wikipedia.org\/wiki\/IntelliSense\" rel=\"noreferrer\">IntelliSense<\/a>, but I'm still learning it.<\/p>\n<p>Try it out! <code>:help ins-completion<\/code><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218468\/revisions\" title=\"show all edits to this post\">edited Feb 27 '10 at 11:29<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218468\" href=\"https:\/\/stackoverflow.com\/posts\/1218468\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 56%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/63550\">Peter Mortensen<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        20&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>These are not shortcuts, but they are related:<\/p>\n<ol><li>Make capslock an additional ESC (or Ctrl)<\/li>\n<li>map leader to \",\" (comma), with this command: let mapleader=\",\"<\/li>\n<\/ol><p>They boost my productivity.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/2443508\/revisions\" title=\"show all edits to this post\">Mar 14 '10 at 19:49<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2443508\" href=\"https:\/\/stackoverflow.com\/posts\/2443508\/revisions\" title=\"show revision history for this post\">&#13;\n        tfmoraes&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        19&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Another useful vi \"shortcut\" I frequently use is 'xp'. This will swap the character under the cursor with the next character.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/3750570\/revisions\" title=\"show all edits to this post\">Sep 20 '10 at 10:34<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-3750570\" href=\"https:\/\/stackoverflow.com\/posts\/3750570\/revisions\" title=\"show revision history for this post\">&#13;\n        Costyn&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        18&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>&lt;Ctrl&gt; + W, V to split the screen vertically <br \/>&lt;Ctrl&gt; + W, W to shift between the windows<\/p>\n<p>!python % [args] to run the script I am editing in this window<\/p>\n<p>ZF in visual mode to fold arbitrary lines<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/1218548\/revisions\" title=\"show all edits to this post\">Aug 2 '09 at 9:47<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218548\" href=\"https:\/\/stackoverflow.com\/posts\/1218548\/revisions\" title=\"show revision history for this post\">&#13;\n        Peter Ellis&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        18&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Visual Mode<\/p>\n<p>As several other people have said, visual mode is the answer to your copy\/cut &amp; paste problem.  Vim gives you 'v', 'V', and C-v.  Lower case 'v' in vim is essentially the same as the shift key in notepad.  The nice thing is that you don't have to hold it down.  You can use any movement technique to navigate efficiently to the starting (or ending) point of your selection.  Then hit 'v', and use efficient movement techniques again to navigate to the other end of your selection.  Then 'd' or 'y' allows you to cut or copy that selection.<\/p>\n<p>The advantage vim's visual mode has over Jim Dennis's description of cut\/copy\/paste in vi is that you don't have to get the location exactly right.  Sometimes it's more efficient to use a quick movement to get to the general vicinity of where you want to go and then refine that with other movements than to think up a more complex single movement command that gets you exactly where you want to go.<\/p>\n<p>The downside to using visual mode extensively in this manner is that it can become a crutch that you use all the time which prevents you from learning new vi(m) commands that might allow you to do things more efficiently.  However, if you are very proactive about learning new aspects of vi(m), then this probably won't affect you much.<\/p>\n<p>I'll also re-emphasize that the visual line and visual block modes give you variations on this same theme that can be very powerful...especially the visual block mode.<\/p>\n<p>On Efficient Use of the Keyboard<\/p>\n<p>I also disagree with your assertion that alternating hands is the fastest way to use the keyboard.  It has an element of truth in it.  Speaking very generally, repeated use of the same thing is slow.  This most significant example of this principle is that consecutive keystrokes typed with the same finger are very slow.  Your assertion probably stems from the natural tendency to use the s\/finger\/hand\/ transformation on this pattern.  To some extent it's correct, but at the extremely high end of the efficiency spectrum it's incorrect.<\/p>\n<p>Just ask any pianist.  Ask them whether it's faster to play a succession of a few notes alternating hands or using consecutive fingers of a single hand in sequence.  The fastest way to type 4 keystrokes is not to alternate hands, but to type them with 4 fingers of the same hand in either ascending or descending order (call this a \"run\").  This should be self-evident once you've considered this possibility.<\/p>\n<p>The more difficult problem is optimizing for this.  It's pretty easy to optimize for absolute distance on the keyboard.  Vim does that.  It's much harder to optimize at the \"run\" level, but vi(m) with it's modal editing gives you a better chance at being able to do it than any non-modal approach (ahem, emacs) ever could.<\/p>\n<p>On Emacs<\/p>\n<p>Lest the emacs zealots completely disregard my whole post on account of that last parenthetical comment, I feel I must describe the root of the difference between the emacs and vim religions.  I've never spoken up in the editor wars and I probably won't do it again, but I've never heard anyone describe the differences this way, so here it goes.  The difference is the following tradeoff:<\/p>\n<p>Vim gives you unmatched raw text editing efficiency\nEmacs gives you unmatched ability to customize and program the editor<\/p>\n<p>The blind vim zealots will claim that vim has a scripting language.  But it's an obscure, ad-hoc language that was designed to serve the editor.  Emacs has Lisp!  Enough said.  If you don't appreciate the significance of those last two sentences or have a desire to learn enough about functional programming and Lisp to develop that appreciation, then you should use vim.<\/p>\n<p>The emacs zealots will claim that emacs has viper mode, and so it is a superset of vim.  But viper mode isn't standard.  My understanding is that viper mode is not used by the majority of emacs users.  Since it's not the default, most emacs users probably don't develop a true appreciation for the benefits of the modal paradigm.<\/p>\n<p>In my opinion these differences are orthogonal.  I believe the benefits of vim and emacs as I have stated them are both valid.  This means that the ultimate editor doesn't exist yet.  It's probably true that emacs would be the easiest platform on which to base the ultimate editor.  But modal editing is not entrenched in the emacs mindset.  The emacs community could move that way in the future, but that doesn't seem very likely.<\/p>\n<p>So if you want raw editing efficiency, use vim.  If you want the ultimate environment for scripting and programming your editor use emacs.  If you want some of both with an emphasis on programmability, use emacs with viper mode (or program your own mode).  If you want the best of both worlds, you're out of luck for now.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2562090\/revisions\" title=\"show all edits to this post\">edited Apr 1 '10 at 17:00<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2562090\" href=\"https:\/\/stackoverflow.com\/posts\/2562090\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/307058\">user307058<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        17&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Spend 30 mins doing the vim tutorial (run vimtutor instead of vim in terminal). You will learn the basic movements, and some keystrokes, this will make you at least as productive with vim as with the text editor you used before. After that, well, read Jim Dennis' answer again :)<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/2557010\/revisions\" title=\"show all edits to this post\">Mar 31 '10 at 22:44<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2557010\" href=\"https:\/\/stackoverflow.com\/posts\/2557010\/revisions\" title=\"show revision history for this post\">&#13;\n        konryd&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        16&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<blockquote>\n  <p>What is the way you use Vim that makes\n  you more productive than with a\n  contemporary editor?<\/p>\n<\/blockquote>\n<p>Being able to execute complex, repetitive edits with very few keystrokes (often using <a href=\"http:\/\/vimdoc.sourceforge.net\/htmldoc\/repeat.html#complex-repeat\" rel=\"noreferrer\">macros<\/a>). Take a look at <a href=\"http:\/\/www.vimgolf.com\/\" rel=\"noreferrer\">VimGolf<\/a> to witness the power of Vim!<\/p>\n<p>After over ten years of almost daily usage, it's hard to imagine using any other editor.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/4674946\/revisions\" title=\"show all edits to this post\">Jan 12 '11 at 22:52<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-4674946\" href=\"https:\/\/stackoverflow.com\/posts\/4674946\/revisions\" title=\"show revision history for this post\">&#13;\n        Johnsyweb&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        16&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Use <code>\\c<\/code> anywhere in a search to ignore case (overriding your ignorecase or smartcase settings).\nE.g. <code>\/\\cfoo<\/code> or <code>\/foo\\c<\/code> will match <code>foo<\/code>, <code>Foo<\/code>, <code>fOO<\/code>, <code>FOO<\/code>, etc.<\/p>\n<p>Use <code>\\C<\/code> anywhere in a search to force case matching.\nE.g. <code>\/\\Cfoo<\/code> or <code>\/foo\\C<\/code> will only match foo.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/3004973\/revisions\" title=\"show all edits to this post\">edited Jun 15 '11 at 13:42<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-3004973\" href=\"https:\/\/stackoverflow.com\/posts\/3004973\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 67%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/348785\">kev<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        16&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I was surprised to find no one mention the <code>t<\/code> movement. I frequently use it with parameter lists in the form of <code>dt,<\/code> or <code>yt,<\/code><\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/2560001\/revisions\" title=\"show all edits to this post\">edited Jun 15 '11 at 13:44<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-2560001\" href=\"https:\/\/stackoverflow.com\/posts\/2560001\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 67%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/306815\">David Corbin<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        16&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Odd nobody's mentioned ctags. Download \"exuberant ctags\" and put it ahead of the crappy preinstalled version you already have in your search path. Cd to the root of whatever you're working on; for example the Android kernel distribution. Type \"ctags -R .\" to build an index of  source files anywhere beneath that dir in a file named \"tags\". This contains all tags, nomatter the language nor where in the dir, in one file, so cross-language work is easy.<\/p>\n<p>Then open vim in that folder and read :help ctags for some commands. A few I use often:<\/p>\n<ul><li><p>Put cursor on a method call and type CTRL-] to go to the method\ndefinition. <\/p><\/li>\n<li><p>Type :ta name to go to the definition of name.<\/p><\/li>\n<\/ul><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/9145364\/revisions\" title=\"show all edits to this post\">edited May 6 '12 at 20:50<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-9145364\" href=\"https:\/\/stackoverflow.com\/posts\/9145364\/revisions\" title=\"show revision history for this post\">&#13;\n        3 revs<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/563868\">Bradjcox<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        14&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Automatic indentation:<\/p>\n<p><code>gg<\/code> (go to start of document)<br \/><code>=<\/code> (indent time!)<br \/><code>shift-g<\/code> (go to end of document)<br \/><\/p>\n<p>You'll need 'filetype plugin indent on' in your <code>.vimrc<\/code> file, and probably appropriate 'shiftwidth' and 'expandtab' settings.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1236603\/revisions\" title=\"show all edits to this post\">edited Feb 27 '10 at 11:19<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1236603\" href=\"https:\/\/stackoverflow.com\/posts\/1236603\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 67%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/150807\">David Claridge<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        14&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>You asked about productive shortcuts, but I think your real question is: Is vim worth it? The answer to this stackoverflow question is -&gt; \"Yes\"<\/p>\n<p>You must have noticed two things. Vim is powerful, and vim is hard to learn. Much of it's power lies in it's expandability and endless combination of commands. Don't feel overwhelmed. Go slow. One command, one plugin at a time. Don't overdo it. <\/p>\n<p>All that investment you put into vim will pay back a thousand fold. You're going to be inside a text editor for many, many hours before you die. Vim will be your companion.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/3323973\/revisions\" title=\"show all edits to this post\">Jul 24 '10 at 5:41<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-3323973\" href=\"https:\/\/stackoverflow.com\/posts\/3323973\/revisions\" title=\"show revision history for this post\">&#13;\n        autodidakto&#13;\n        <\/a>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        13&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Multiple buffers, and in particular fast jumping between them to compare two files with <code>:bp<\/code> and <code>:bn<\/code> (properly remapped to a \nsingle <kbd>Shift<\/kbd> + <kbd>p<\/kbd> or <kbd>Shift<\/kbd> + <kbd>n<\/kbd>)<\/p>\n<p><code>vimdiff<\/code> mode (splits in two vertical buffers, with colors to show the differences)<\/p>\n<p>Area-copy with <kbd>Ctrl<\/kbd> + <kbd>v<\/kbd><\/p>\n<p>And finally, tab completion of identifiers (search for \"mosh_tab_or_complete\"). That's a life changer.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/1218514\/revisions\" title=\"show all edits to this post\">edited Feb 27 '10 at 11:23<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1218514\" href=\"https:\/\/stackoverflow.com\/posts\/1218514\/revisions\" title=\"show revision history for this post\">&#13;\n        2 revs, 2 users 67%<br \/><\/a><a href=\"https:\/\/stackoverflow.com\/users\/78374\">Stefano Borini<\/a>&#13;\n&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        12&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Agreed with the top poster - the <strong>:r!<\/strong> command is <em>very<\/em> useful.<\/p>\n<p>Most often I use it to \"paste\" things:<\/p>\n<pre>:r!cat\n**Ctrl-V to paste from the OS clipboard**\n^D\n<\/pre>\n<p>This way I don't have to fiddle with <code>:set paste<\/code>.<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered <a href=\"https:\/\/stackoverflow.com\/posts\/1347985\/revisions\" title=\"show all edits to this post\">Aug 28 '09 at 16:07<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n            community wiki&#13;\n        &#13;<\/div><div class=\"user-details\">&#13;\n        <a id=\"history-1347985\" href=\"https:\/\/stackoverflow.com\/posts\/1347985\/revisions\" title=\"show revision history for this post\">&#13;\n        David Wolever&#13;\n        <\/a>&#13;<\/div>","created_at":"2018-11-07T16:48:45+0000","updated_at":"2018-11-07T16:48:45+0000","annotations":[],"mimetype":"text\/html","reading_time":47,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":455,"title":"Roughing It Dev Style: Coding Without a Computer","url":"https:\/\/dev.to\/chznbaum\/roughing-it-dev-style-coding-without-a-computer","content":"<p>A terrible, horrible, no good, very bad thing happened to me this month: the Surface Pro that was my exclusive personal and work computer died. Even though I found an inexpensive replacement to order, I couldn't imagine not coding for about a week. I was totally freaking out.<\/p><p>After a few unsatisfying days of merely reading about web development topics, I set to work on figuring out how I could get some real work done, armed only with my iPhone.<\/p>\n<h2>Trial and Error<\/h2>\n<p>Not wanting to hunt down a new solution if it was unnecessary, I first turned to tools I've used before:<\/p>\n<h3>Codepen<\/h3>\n<p>Codepen wasn't terrible in a pinch. However, the layout and on-screen keyboard really wasn't optimized for coding on a phone. Making minor changes took a lot of time, and it was easy to end up with really buggy code.<\/p>\n<h3>Cloud9 IDE<\/h3>\n<p>Cloud9 had the same problems but with an already more cluttered interface. It explicitly <a href=\"https:\/\/community.c9.io\/t\/supported-browsers\/1988\/4\">doesn't support mobile browsers<\/a> and it has been made clear multiple times that making Cloud9 mobile-friendly or developing a native mobile app for it is not anywhere on the roadmap.<\/p>\n<h3>CodeAnywhere<\/h3>\n<p>When exploring the Cloud9 issue, I saw mention of an alternative cross-platform cloud IDE called CodeAnywhere. However, the iOS app appeared to have not been updated since 2014 and after four unsuccessful attempts to so much as create an account and login, I figured there had to be something better.<\/p>\n<h2>What Finally Worked<\/h2>\n<p>After some frustrating experiences, I came across tools and techniques that worked beyond my expectations, allowing me to code productively with just a phone and that can allow you to do it, too.<\/p>\n<h3>Editor: <a href=\"http:\/\/buffereditor.com\/\">Buffer Editor<\/a>\n<\/h3>\n<p><a href=\"http:\/\/buffereditor.com\/\">Buffer Editor<\/a> was a hidden gem in the App Store. It didn't have a presence from advertising or word of mouth or even enough reviews to warrant a rating. It would be easy to underestimate Buffer.<\/p>\n<p>When actually downloaded, though, it's clear that Buffer Editor is a real, powerful editor truly optimized for a mobile experience. Features that won me over included:<\/p>\n<ul><li>An extended keyboard that makes typing quick (for one-handed) and intentional.<\/li>\n<li>A well-designed built in terminal.<\/li>\n<li>Full-screen editing, a necessity on such a small screen.<\/li>\n<li>Syntax highlighting and autocomplete for 40+ languages and technologies.<\/li>\n<li>The ability to connect via SSH, FTP, SFTP, GitHub, Dropbox and iCloud.<\/li>\n<li>Sending files by email.<\/li>\n<\/ul><p><a href=\"https:\/\/res.cloudinary.com\/practicaldev\/image\/fetch\/s--SfY5Q6qy--\/c_limit%2Cf_auto%2Cfl_progressive%2Cq_auto%2Cw_880\/https:\/\/thepracticaldev.s3.amazonaws.com\/i\/91wdnmj76i0r0uqyd2ha.jpg\" class=\"article-body-image-wrapper\"><img src=\"https:\/\/res.cloudinary.com\/practicaldev\/image\/fetch\/s--SfY5Q6qy--\/c_limit%2Cf_auto%2Cfl_progressive%2Cq_auto%2Cw_880\/https:\/\/thepracticaldev.s3.amazonaws.com\/i\/91wdnmj76i0r0uqyd2ha.jpg\" alt=\"Buffer Editor Extended Keyboard\" \/><\/a><\/p>\n<p>It also included useful <a href=\"http:\/\/buffereditor.com\/features\/\">features<\/a> that I didn't utilize such as Vim support and support for bluetooth keyboards.<\/p>\n<p>While it isn't free, the 4.99 USD price I got it for was well worth it, especially compared to the popular Coda at almost 25 USD.<\/p>\n<h3>Server: <a href=\"https:\/\/www.digitalocean.com\/\">Digital Ocean<\/a>\n<\/h3>\n<p>Being confined to a phone gave me the push I needed to start building more websites and applications directly in a VPS solution like <a href=\"https:\/\/www.digitalocean.com\/solutions\/\">Digital Ocean<\/a> rather than starting out solely on a local machine. Using a Virtual Private Server (Digital Ocean calls them droplets) allows setting up a development environment and managing files\/servers without needing a local environment to work with, since it's on another computer in the cloud. Another benefit of VPS is that from the start, a project isn't tied down to one local machine, even if it hasn't been checked in to version control yet.<\/p>\n<p>Digital Ocean has proven to be easy to use (and inexpensive), and there is a wealth of documentation and guides for folks new to running servers.<\/p>\n<p>Accessing the droplet remotely was easy in Buffer Editor, simply involving adding a new SSH connection and filling out the relevant settings with information from the droplet.<\/p>\n<p><a href=\"https:\/\/res.cloudinary.com\/practicaldev\/image\/fetch\/s--GexO45y5--\/c_limit%2Cf_auto%2Cfl_progressive%2Cq_auto%2Cw_880\/https:\/\/thepracticaldev.s3.amazonaws.com\/i\/gbb7k3irbt7uz0y2juyw.jpg\" class=\"article-body-image-wrapper\"><img src=\"https:\/\/res.cloudinary.com\/practicaldev\/image\/fetch\/s--GexO45y5--\/c_limit%2Cf_auto%2Cfl_progressive%2Cq_auto%2Cw_880\/https:\/\/thepracticaldev.s3.amazonaws.com\/i\/gbb7k3irbt7uz0y2juyw.jpg\" alt=\"Buffer Editor New Connection Form\" \/><\/a><\/p>\n<h3>Browser Access: Local Server Binding<\/h3>\n<p>The wonderful and problematic thing about web development is, of course, the primary thing you need is a browser. Using a VPS to develop, you have access to a terminal console, but not a Graphical User Interface or a traditional web browser.<\/p>\n<p>When developing a web app, you often need to be able to start a local server and access <code>localhost<\/code> in the web browser, since that's where the development server displays the app. Unfortunately <code>localhost<\/code> means <em>this computer<\/em>, and so it's inaccessible outside of the VPS hosting the app files.<\/p>\n<p>The workaround here involved learning more about servers and requests:<\/p>\n<p>When starting a server, it sends what is called a bind request to indicate that it is ready to receive requests associated with an IP address. Local servers typically bind to <code>127.0.0.1<\/code> because that IP address is used to loop back to the requesting computer; each computer can only request <code>127.0.0.1<\/code> from itself, which is usually convenient for developing.<\/p>\n<p>That's clearly not an option. You need to have the server bind request indicate its publicly accessible IP address instead so that it's accessible from a browser outside of that computer. This IP address is easily obtained from Digital Ocean, and the workaround involved adding a <code>--binding<\/code> flag like this, for starting a Rails server:<\/p>\n<div class=\"highlight\"><pre>rails s --binding=XX.XX.XX.XX\n<\/pre><\/div>\n<p>Now, instead of typing <code>localhost:3000<\/code> into the browser's address bar to view the app, you would type the publicly accessible IP address, like <code>XX.XX.XX.XX:3000<\/code>. As long as the local server is running with the binding flag, the app will be accessible remotely from the server's IP address.<\/p>\n<p>Buffer Editor is especially convenient by keeping the server running after you back out of the terminal (so you can open it again and continue to develop). This can be confusing at first, but to stop the server, lock and unlock your phone.<\/p>\n<h3>Get Back to Coding Productivity<\/h3>\n<p>From here, you can utilize git like usual, get to work and debug your code wherever you are and regardless of WiFi\/Ethernet availability. In fact, the day before my new computer arrived, I was able to make 6 commits and push them to GitHub, all from my iPhone.<\/p>\n<p>While my day to day coding will be on a traditional desktop or laptop when it's available (since two hands are faster than one), I continue to utilize these tools to work on my projects when I can't bring my laptop somewhere or when mobile internet is all I'll have access to, such as riding in a car.<\/p>\n<p>I'm actually thankful for the experience of losing access to the computer, because it forced me to find a solution that now allows me to code virtually anywhere.<\/p>","created_at":"2018-11-07T16:48:45+0000","updated_at":"2018-11-07T16:48:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"dev.to","preview_picture":"https:\/\/res.cloudinary.com\/practicaldev\/image\/fetch\/s--N5cVGl_a--\/c_imagga_scale,f_auto,fl_progressive,h_500,q_auto,w_1000\/https:\/\/thepracticaldev.s3.amazonaws.com\/i\/lkla9cooi9y21v7ymdsz.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":456,"title":"Spacemacs - Configuring the Solarized Theme","url":"http:\/\/philipdaniels.com\/blog\/2017\/02\/spacemacs---configuring-the-solarized-theme\/","content":"<div class=\"reading-time\">929 words, estimated reading time 4 minutes at 300 w.p.m.<\/div><p>There are at least 3 different Solarized theme packages for Emacs:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/bbatsov\/solarized-emacs\">BBatsov\u2019s Solarized<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/sellout\/emacs-color-theme-solarized\">Greg Pfeil\u2019s (aka Sellout) Solarized<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/purcell\/color-theme-sanityinc-solarized\">Steve Purcell\u2019s (aka SanityInc) Solarized<\/a><\/li>\n<\/ul><p>Steve\u2019s is based on Greg\u2019s. Of the three, I have always used Greg\u2019s in the past,\nfinding it to have a comprehensive set of faces. So naturally I wanted to get it\nworking in Spacemacs. That was a bit trickier than expected, because it uses a\nnon-standard loading mechanism, however, I found an issue on Github that\nexplained how to do it. I reproduce it here.<\/p>\n<p>The package you want is called <code>color-theme-solarized<\/code>. But you will get a load\nerror if you try and add it to the <code>dotspacemacs-themes<\/code> variable in your\n<code>.spacemacs<\/code>. Instead, add it here:<\/p>\ndotspacemacs-additional-packages '(color-theme-solarized)\n<p>Then, in your <code>dotspacemacs\/user-config<\/code> function, you need to manually load the\ntheme:<\/p>\n(set-terminal-parameter nil 'background-mode 'dark)\n(set-frame-parameter nil 'background-mode 'dark)\n(spacemacs\/load-theme 'solarized)\n<p>Thanks to <a href=\"https:\/\/github.com\/cpaulik\">Christoph Paulik<\/a> who posted this\nsolution at\n<a href=\"https:\/\/github.com\/syl20bnr\/spacemacs\/issues\/1269#issuecomment-198309213\">https:\/\/github.com\/syl20bnr\/spacemacs\/issues\/1269#issuecomment-198309213<\/a><\/p>\n<p>Things look pretty good out of the box, you even get rainbow delimiters:<\/p>\n<p><img src=\"http:\/\/philipdaniels.com\/blog\/img\/2017\/spacemacs-solarized\/default-solarized.gif\" alt=\"Solarized Dark in action\" \/><\/p>\n<p>I find the provided modeline to be a bit bland, I prefer something little bolder\nand that functions as a sort of \u201con\/off\u201d modeline where the modeline for\ninactive buffers is the inverse of the active one. To accomplish this I adjusted\nsome faces after the theme has loaded. Spacemacs has a layer\ncalled\n<a href=\"https:\/\/github.com\/syl20bnr\/spacemacs\/tree\/master\/layers\/%2Bthemes\/theming\">theming<\/a>\nwhich helps with this.<\/p>\n<p>First ensure you bring in the <code>theming<\/code> package in the usual way, by adding it to your\n<code>dotspacemacs-configuration-layers<\/code> variable in the <code>.spacemacs<\/code> file. Then add\nthe following code <em>just before<\/em> the lines above that call <code>spacemacs\/load-theme<\/code>.<\/p>\n;; Get color-theme-solarized working. It is specified as an additional package\n;; above. First we setup some theme modifications - we must do this *before*\n;; we load the theme. Note that the color-theme-solarized package appears in\n;; the list of themes as plain old 'solarized'.\n(setq theming-modifications\n      '((solarized\n         ;; Provide a sort of \"on-off\" modeline whereby the current buffer has a nice\n         ;; bright blue background, and all the others are in cream.\n         ;; TODO: Change to use variables here. However, got error:\n         ;; (Spacemacs) Error in dotspacemacs\/user-config: Wrong type argument: stringp, pd-blue\n         (mode-line :foreground \"#e9e2cb\" :background \"#2075c7\" :inverse-video nil)\n         (powerline-active1 :foreground \"#e9e2cb\" :background \"#2075c7\" :inverse-video nil)\n         (powerline-active2 :foreground \"#e9e2cb\" :background \"#2075c7\" :inverse-video nil)\n         (mode-line-inactive :foreground \"#2075c7\" :background \"#e9e2cb\" :inverse-video nil)\n         (powerline-inactive1 :foreground \"#2075c7\" :background \"#e9e2cb\" :inverse-video nil)\n         (powerline-inactive2 :foreground \"#2075c7\" :background \"#e9e2cb\" :inverse-video nil)\n         ;; Make a really prominent helm selection line.\n         (helm-selection :foreground \"white\" :background \"red\" :inverse-video nil)\n         ;; See comment above about dotspacemacs-colorize-cursor-according-to-state.\n         (cursor :background \"#b58900\")\n       )))\n(set-terminal-parameter nil 'background-mode 'dark)\n(set-frame-parameter nil 'background-mode 'dark)\n(spacemacs\/load-theme 'solarized)\n<p>Now, these modifications will be applied after solarized is loaded. <code>#e9e2cb<\/code> is\nthe creamy-white color from solarized, and <code>#2075c7<\/code> is the blue colour. The\nfinal result looks like this:<\/p>\n<p><img src=\"http:\/\/philipdaniels.com\/blog\/img\/2017\/spacemacs-solarized\/solarized-adjusted.gif\" alt=\"Solarized with tweaked modeline\" \/><\/p>\n<p>You can take this opportunity to customize other faces too. I changed the cursor\nto the yellow color from solarized, but to get this to work you must set the\nvariable <code>dotspacemacs-colorize-cursor-according-to-state<\/code> to nil, otherwise\nSpacemacs will change the cursor back to what it thinks it should be! That took\na while to figure out, it was an unrelated posting on\nthe <a href=\"https:\/\/gitter.im\/syl20bnr\/spacemacs\">Gitter Channel<\/a> that tipped me off to\nthat one.<\/p>\n<p>Oh, and my helm selection line looks like this - I like there to be no ambiguity\nabout what I am selecting!<\/p>\n<p><img src=\"http:\/\/philipdaniels.com\/blog\/img\/2017\/spacemacs-solarized\/helm-selection.gif\" alt=\"Prominent helm-selection line\" \/><\/p>\n<p>My last theming tweak was to adjust the parentheses highlighting to something\nmore to my taste. As I noted\nabove, <a href=\"https:\/\/github.com\/Fanael\/rainbow-delimiters\">rainbow delimiters<\/a> are on\nout of the box, but dynamic highlighting\nby <a href=\"https:\/\/github.com\/tsdh\/highlight-parentheses.el\">highlight-parentheses<\/a> is\nalso enabled. I think <a href=\"https:\/\/github.com\/Fuco1\/smartparens\">smart-parens<\/a> also\ngets in on the act.<\/p>\n<p>Anyway, I found the light green highlighting on top of rainbow delimiters to be\na little too much\u2026it made it <em>harder<\/em> to see the matching parentheses.<\/p>\n<p>I like to use a package\ncalled <a href=\"https:\/\/www.emacswiki.org\/emacs\/mic-paren.el\">mic-paren<\/a> to do my\nhighlighting. I found it on\nthe <a href=\"https:\/\/www.emacswiki.org\/emacs\/SiteMap\">Emacs Wiki<\/a> but it is available\non <a href=\"https:\/\/melpa.org\/#\/\">MELPA<\/a>.<\/p>\n<p>To setup mic-paren, first let\u2019s ban <code>highlight-parentheses<\/code> by adding it to the variable\n<code>dotspacemacs-excluded-packages<\/code> in the <code>.spacemacs<\/code> file:<\/p>\n<p>\ndotspacemacs-excluded-packages '(highlight-parentheses)\n<\/p>\n<p>Finally, in\nmy <a href=\"http:\/\/philipdaniels.com\/blog\/2017\/02\/spacemacs---first-steps\">pd layer\u2019s<\/a>\n<code>packages.el<\/code> file I added <code>mic-paren<\/code> to the list of packages I needed, and\ninitialized it as follows:<\/p>\n<p>\n(defun pd\/init-mic-paren ()\n  \"Customize mic-paren. The standard package highlight-parentheses is banned in my\n  .spacemacs file, I much prefer this.\"\n  (use-package mic-paren\n    :config\n    (setq blink-matching-paren nil)\n    (paren-activate)\n    (setq paren-match-face 'mode-line)\n  ))\n<\/p>\n<p>You can pick any face you want, I like to use the same as my active modeline.<\/p>\n<p>I also needed to stop <code>smart-parens<\/code> from interfering by adding these two lines to my\n<code>dotspacemacs\/user-config<\/code> function:<\/p>\n<p>\n;; Turn this off to stop it interfering with mic-paren.\n(set-face-attribute 'sp-show-pair-match-face nil :foreground 'unspecified :background 'unspecified)\n(set-face-attribute 'sp-show-pair-mismatch-face nil :foreground 'unspecified :background 'unspecified)\n<\/p>\n<p>The final result looks like this:<\/p>\n<p><img src=\"http:\/\/philipdaniels.com\/blog\/img\/2017\/spacemacs-solarized\/mic-paren.gif\" alt=\"mic-paren highlighting\" \/><\/p>","created_at":"2018-11-07T16:48:45+0000","updated_at":"2018-11-07T16:48:45+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"philipdaniels.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":451,"title":"Concise Electronics for Geeks (2010)","url":"https:\/\/news.ycombinator.com\/item?id=15703472","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15703472\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/lcamtuf.coredump.cx\/electronics\/\" class=\"storylink\">Concise Electronics for Geeks (2010)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=coredump.cx\">coredump.cx<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        481 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=nbmh\" class=\"hnuser\">nbmh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15703472\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15703472&amp;goto=item%3Fid%3D15703472\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Concise%20Electronics%20for%20Geeks&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Concise%20Electronics%20for%20Geeks\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15703472&amp;auth=32cd41762505cb10dbfd82c07ee5bd8c07f03a5f\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15703472\">49\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:48:44+0000","updated_at":"2018-11-07T16:48:44+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":452,"title":"Concise electronics for geeks","url":"http:\/\/lcamtuf.coredump.cx\/electronics\/","content":"Concise electronics for geeks\nCopyright (C) 2010 by Michal Zalewski &lt;<a href=\"mailto:lcamtuf@coredump.cx\">lcamtuf@coredump.cx<\/a>&gt;\n<p>\nThere are quite a few primers on electronics on the Internet; sadly, almost all of the top hits resort to gross oversimplifications (e.g., \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Hydraulic_analogy\">hydraulic analogies<\/a>), or convenient omission, when covering subtle but\nincredibly important topics such as the real-world behavior of semiconductors. There are some exceptions - but they often suffer\nfrom another malady: regressions into mundane, academic rigor, complete with differential equations and complex number algebra in transient \nanalysis - a trait that is highly unlikely to be accessible, or even useful, to hobbyists.\n<\/p><p>\nThe goal of this guide is to bridge this gap; it should give you an anatomically correct insight into the underlying physical phenomena needed to \naccurately understand the behavior of semiconductors, capacitors, or inductors - but should be far more readable and way shorter than a typical academic\ntextbook. The target audience is people who want to meaningfully tinker with more complex \nelectronic circuits, or perhaps understand how computers really work - but for whom getting there is not meant to be a full-time job.\n<\/p><h2>The physics of conduction <a name=\"1\" href=\"#1\" class=\"slink\" id=\"1\">[link]<\/a><\/h2>\nAs you probably recall from your school years, the dated but still useful\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Bohr_model_of_the_atom\">Bohr model of the atom<\/a> explains that atoms consist of a dense center (nucleus) with\na variable number of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Proton\">protons<\/a> and <a href=\"http:\/\/en.wikipedia.org\/wiki\/Neutron\">neutrons<\/a>. This\ncenter is surrounded by a distant cloud of small, elementary particles called <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electron\">electrons<\/a> occupying\na number of discrete energy states around it.\n<p>\nThe two most important forces governing the atomic structure are: the powerful but short-range \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Residual_strong_force\">residual strong force<\/a> that binds the protons and neutrons together; and a significantly \nweaker but far-reaching  <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electromagnetic_force\">electromagnetic force<\/a> that causes subatomic particles carrying\nelementary charges to create <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electric_field\">electric fields<\/a>, repelling similar charges, and attracting\ndissimilar ones. Charges traveling in relation to each other are also mutually subjected to\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Magnetic_field\">magnetic fields<\/a>.\n<\/p><p>\nProtons have an inherent, elementary positive charge; electrons have an identical charge with the opposite sign; and both produce a\ngradually vanishing (but infinite) electric field around them. This field is customarily visualized by observing how it would affect a\npositive charge in its vicinity: field lines are shown to be radially emanating away from positive charges, or toward negatives ones.\n<\/p><p>\nNeutrons don't generate and do not respond to electric fields - but are affected by magnetic fields created by charges in motion; this is an artifact of\ntheir <a href=\"http:\/\/en.wikipedia.org\/wiki\/Quark\">quark<\/a> makeup.\n<\/p><p>\nThe positive electric charge (Q) of <code>6.24 * 10<sup>18<\/sup><\/code> protons is known as one <a href=\"http:\/\/en.wikipedia.org\/wiki\/Coulomb\">coulomb<\/a> (C); \nyou don't need to remember this number: its only significance is that it corresponds neatly to other SI units used to measure \nrelated physical phenomena.\n<\/p><p>\nThe electromagnetic force causes protons to attract electrons, and forces the captive electrons\nto orbit the nucleus; it also causes protons to repel each other - but\nthis effect is easily overcome by the strong nuclear force that holds the center of the atom together. Given the opportunity, any sufficiently stable atomic\nnucleus will generally seek to accumulate a number of electrons matching the number of protons, in order to form an electrically neutral particle.\n<\/p><p>\nThe strongly bound and heavy nuclei of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Stable_isotope\">stable isotopes<\/a> do not undergo any structural\nchanges under everyday circumstances. The electrons occupying the energetic, outer (<a href=\"http:\/\/en.wikipedia.org\/wiki\/Valence_electron\">valence<\/a>) \norbitals of an atom are a very different story; they are usually attracted to the nucleus fairly weakly, and can engage in two rather interesting behaviors:\n<\/p><ul><li> Some valence electron configurations are energetically preferred over others, due to the complex proton-proton and electron-electron interactions\nand quantum probability distributions that affect the geometry of the nuclear electromagnetic field. Because of this, in favorable conditions, atoms may \nbe willing to share some of their electrons with neighbors, forming electrically neutral <a href=\"http:\/\/en.wikipedia.org\/wiki\/Molecule\">molecules<\/a>\nheld together by the electromagnetic force - and bringing us chemistry as we know it.\n<p>\nThree primary types of intramolecular bonds are recognized, depending on the exact nature of this exchange:\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Covalent_bond\">covalent<\/a>, <a href=\"http:\/\/en.wikipedia.org\/wiki\/Ionic_bond\">ionic<\/a>, and\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Metallic_bond\">metallic<\/a>. This last category is of particular interest: shared valence electrons in metals are so\nweakly bound, that they enjoy largely unhindered mobility through the crystal lattice, and are seemingly not tied to any particular atom - forming\nwhat is called an <a href=\"http:\/\/en.wikipedia.org\/wiki\/Free_electron_model\">electron gas<\/a> distributed through the material.\n<\/p><\/li><li> Certain atoms or molecules may be not particularly strongly opposed to acquiring some additional electrons, or losing some of their own, without\nfirmly attaching to any other single molecule - gaining a net negative or positive electrical charge in the process. Such charged particles are known as \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Ion\">ions<\/a> (anions and cations, respectively).\n<p>\nIon formation may be driven by certain chemical reactions, by the action of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Solvation\">polar solvents<\/a> on\nsome substances, by certain types of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Ionization\">energetic collisions<\/a> - or even by simple\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Triboelectric_effect\">mechanical action<\/a>: when materials of different affinity to electrons are brought into\ncontact and then pulled apart, a significant charge imbalance may form.\n<\/p><\/li><\/ul>\nIt is worth noting that with sufficient energy, almost any substance will eventually form free ions; that said, \nmaterials in which ions are not formed easily under normal operating conditions are known as \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Insulator_(electrical)\">insulators<\/a>.\n<p>\nIn dissociated solutions or in <a href=\"http:\/\/en.wikipedia.org\/wiki\/Plasma_%28physics%29\">plasma<\/a>, ions themselves are fairly mobile, moving\naround in response to electric fields and other processes; <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electrochemistry\">electrochemistry<\/a> studies these behaviors in \ngreat detail, with <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electrochemical_cell\">batteries<\/a> being one of the most important applications of this principle.\nIn solids, however, ions typically do not enjoy appreciable mobility. Therefore, many materials where ions can form with ease are still non-conductive: their surfaces may\naccept localized charges easily, but these charges do not move from one point to another on their own. Such materials may be useful for energy storage,\nbut are poorly suited for signal distribution and processing.\n<\/p><p>\nAs hinted earlier, <a href=\"http:\/\/en.wikipedia.org\/wiki\/Metal\">metals<\/a> are a special case in the world of solids, however: their\ncrystal lattice may be considered to consist of (still immobile) metal cations, surrounded by mostly-free-roaming electrons. These electrons are not really bound \nto any specific atom; and if any localized charge imbalance is created by depositing or removing electrons, this imbalance gets equalized through the \nconductor at a significant fraction of the speed of light. It is important to note that the electrons themselves don't move through the medium nearly \nas fast - in fact, their traverse speed in a metallic conductor is usually measured in \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electric_current#Drift_speed\">millimeters per second<\/a>; think of pressure equalization in a fluid, instead.\nThis effect in response to an externally applied, electromotive force (say, a chemical reaction that forcibly\ntakes electrons from one place and shoves them elsewhere) is known as the flow of electricity.\n<\/p><p>\nIn solid-state conductors, negatively-charged electrons are the only real, mobile charge carriers. Their positive counterpart - the \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Positron\">positron<\/a> - is a form of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Antimatter\">antimatter<\/a>, and\nis not commonly encountered on Earth (although it can certainly be spotted <a href=\"http:\/\/en.wikipedia.org\/wiki\/Pair_production\">at times<\/a>).\n<\/p><h2>The curious case of semiconductors <a name=\"2\" href=\"#2\" class=\"slink\" id=\"2\">[link]<\/a><\/h2>\nIn metals, there is an abundance of mobile electrons in the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electronic_band_structure\">conduction\nenergy band<\/a> at any given time because there is no <a href=\"http:\/\/en.wikipedia.org\/wiki\/Band_gap\">energetic band gap<\/a> between the valence and \nconduction bands - i.e., valence electrons already have the range of energies needed to move around in the material at will. Conversely, in insulators, \nthere are no virtually no mobile charge carriers, because the band gap is large enough not to be casually crossed by valence electrons under \nnormal operating conditions.\n<p>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Semiconductor\">Semiconductors<\/a> are a special, intermediate case: similar to insulators, they have a band gap -\nbut the gap is small enough to allow electrons to be periodically \"knocked out\" of their usual orbit and into the mobile conduction band, simply\ndue to the inherent thermal vibration of atoms in the crystal lattice. In absence of an external electric field, these electrons quickly return to their\noriginal positions, and the process repeats over and over again.\n<\/p><p>\nWhen a sufficiently strong electric field is applied, the picture changes slightly: the temporarily liberated electrons will immediately dart off \ntoward the more positive region (a direction said to go \"against\" the direction of the electric field)  - leaving \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electron_hole\">holes<\/a> in the valence bands of their original hosts. Nearby electrons closer to the more negative\nregion are also subject to the same \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electromotive_force\">electromotive force<\/a>,\nand will be eager to jump into that energetically favorable hole, in leaving a new hole at their original location.\nOver time, this will seemingly cause holes to drift in the crystal lattice \"along\" the field. The process ends with the liberated electron arriving\nat one edge of the semiconductor, ready to continue its journey elsewhere should any opportunity present itself; \nand with the hole bubbling up toward the other end, willing to accept any\nexternally donated electrons. This form of charge transfer permits conduction, albeit only to a very limited extent (only a limited number of\nexcited charge carriers is available at any given time, and due to spontaneous recombination, their lifetime is short). Of course, poor conductivity is not\na significant feat by itself - about the only interesting\nproperty of this material is that more electrons get knocked into the conduction band when the material is exposed to light (with photons impacting\nthe surface and transferring their energy to it) - giving birth to <a href=\"http:\/\/en.wikipedia.org\/wiki\/Photoresistor\">photoresistors<\/a> and related\nlight-sensitive electronic devices.\n<\/p><p>\nThe situation gets slightly more remarkable once the material is <a href=\"http:\/\/en.wikipedia.org\/wiki\/Doping_(semiconductors)\">doped<\/a> with certain\nother, closely related atoms. In n-type semiconductors, the dopant is a substance \neager to donate its own weakly bound electrons to plug holes - before the temporarily excited electrons\nrightfully belonging to the doped material have a chance to return to their original state. This creates an abundance of long-lived negative charge carriers\nin the conduction band, in a stable material that has no net electrical charge (every dopant cation is offset by a free electron). These electrons\nwill happily  drift against any externally applied electric field with little effort.\n<\/p><p>\nIn p-type semiconductors, on the other hand, the dopants are quick to snatch and trap any excited electrons with more force than the atoms from which\nthe electrons originally came from. This achieves the opposite effect: an\nabundance of long-lived holes that want to accept any externally supplied wandering electrons into their valence bands. In an external field, this sea\nof holes will allow other electrons to surf from one atom to another without too much work (although jumping between holes requires a bit more\nenergy than just drifting in the conduction band).\n<\/p><p>\nBecause of the continuous availability of long-lived charge carriers, both of these materials are highly conductive - but by themselves, perhaps not \nparticularly exciting. The really interesting bit is when a <a href=\"http:\/\/en.wikipedia.org\/wiki\/P-N_junction\">p-n junction<\/a> is formed by bringing \nthese two types of semiconductors into contact, however: initially, some of the idle electrons from the n-type material will enter the p-type semiconductor to \nrecombine with nearby holes; and p-type semiconductor holes will seemingly \"propagate\" across the junction when some nearby, temporarily excited valence\nelectrons from the n-type material are captured on the p-type side, leaving holes in their original locations. This leads to the formation of a thin \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Depletion_layer\">depletion layer<\/a>, with very few mobile charge carriers available (and therefore, with rather poor \nconductivity). The formation of this layer is a self-limiting effect, though: with no external supply of charges, the recombination leaves the dopant ions \nin the vicinity of the junction unpaired, and the resulting electrostatic field (positive on the n-type side, and negative on the p-type side) eventually \nbecomes strong enough to capture and hold any wandering electrons and holes, and prevent them from getting across the junction.\n<\/p><p>\nApplying an external field to the junction in the direction identical to this intrinsic field - a process called reverse biasing - only serves\nto pull charge carriers further away, and therefore widen the depletion layer; appreciable conduction will not occur under these circumstances, at least \nuntil the field is strong enough to trigger <a href=\"http:\/\/en.wikipedia.org\/wiki\/Avalanche_breakdown\">avalanche breakdown<\/a> or a similar\nphenomenon.\n<\/p><p>\nForward biasing the junction, on the other hand, will push the remaining charge carriers on both sides toward the junction, overcoming the internal field.\nIf an external supply of charges is maintained, this allows continuous and efficient hole-electron recombination in the junction area, and the flow of \ncurrent: n-type semiconductor accepts electrons to offset unbalanced cation charges, and the p-type semiconductor gives them up and frees up holes to avoid \ndeveloping an unbalanced negative charge of its own. As a bonus, recombination is sometimes accompanied by \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/LED\">photon emission<\/a>, as excited conduction band\nelectrons finally drop to a lower energy level after finding a spot in nearby valence shells, and then continue the ride hole-jumping in the valence band.\n<\/p><p>\nThis field-controlled, unidirectional conduction of semiconductor junctions is one of the more important discoveries in the history of electronics\n- allowing a large number of solid-state <a href=\"http:\/\/en.wikipedia.org\/wiki\/Nonlinear_element\">nonlinear<\/a> or \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Active_component\">active<\/a> components to be built; we will discuss them in more detail later on.\n<\/p><h2>Core concepts in electrical circuits <a name=\"3\" href=\"#3\" class=\"slink\" id=\"3\">[link]<\/a><\/h2>\n<h3>Current (I) <a name=\"4\" href=\"#4\" class=\"slink\" id=\"4\">[link]<\/a><\/h3>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electric_current\">Current<\/a> is the measure of the\nflow of electric charges through a selected section of the circuit. The \nunit of current - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Ampere\">ampere<\/a> (A) - is one of the base SI units, defined as the flow of some \n<code>6.24 * 10<sup>18<\/sup><\/code> elementary charges (1 coulomb) per second.\n<p>\nThe flow of current is the mechanism by which electrical circuits can perform <a href=\"http:\/\/en.wikipedia.org\/wiki\/Work_%28physics%29\">useful work<\/a>. \nNaturally, unbalanced charges deposited in a variety of materials may generate electromagnetic fields even when no current is actively flowing, and\nthese fields may in turn affect nearby objects, change the distribution of other charges in conductors, or control the behavior of semiconductor\njunctions; that said, depositing these charges in the first place requires the flow of current.\n<\/p><p>\nCurrent through a conductor is generally expressed as a non-negative, absolute value, no matter which way it is flowing. The direction of the flow - for historical reasons, marked\nin the direction <u>opposite<\/u> to the actual direction of electron travel, i.e. from a positive pole of the supply to the negative one\n- is usually conveyed when describing the voltage instead (see later on).\n<\/p><p>\nRudimentary <a href=\"http:\/\/en.wikipedia.org\/wiki\/Ammeter\">ammeters<\/a> can be designed to measure currents by detecting the magnetic field generated by \ncharges in motion, for example by placing a permanent magnet on a spring near a coiled conductor; \nideal ammeters would not impede the flow of current in any way, although in practice, some power is always dissipated.\n<\/p><h3>Resistance (R) <a name=\"5\" href=\"#5\" class=\"slink\" id=\"5\">[link]<\/a><\/h3>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electrical_resistance\">Resistance<\/a> is the measure of opposition to the <u>steady<\/u> flow of an electric current.\nWith the exception of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Superconductor\">superconductors<\/a>, all conductive materials impede the flow of electrons to\nsome extent - generally through a linear, <a href=\"http:\/\/en.wikipedia.org\/wiki\/Time-invariant\">time-invariant<\/a> effect comparable to \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Drag_%28physics%29\">aerodynamic drag<\/a>; the energy wasted to overcome the drag excites the medium, and is eventually \ndissipated as heat.\n<p>\nThe unit of resistance - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Ohm\">ohm<\/a> (\u2126) - is a derived SI unit, defined as the resistance of a conductor that,\nwhen subjected to a steady current of one ampere, dissipates one <a href=\"http:\/\/en.wikipedia.org\/wiki\/Watt\">watt<\/a> (W) of heat - that's one \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Joule\">joule<\/a> (J) per every coulomb of charge moved around.\n<\/p><p>\nIn any uniform metallic conductor, the resistance is roughly equal to its length, times material-specific resistivity constant, divided by the cross-section\nof said conductor; some temperature dependency is also present. In practical circuits, conductors are generally selected so that the effects of their \nresistance are negligible; a specialized class of components, known as <a href=\"http:\/\/en.wikipedia.org\/wiki\/Resistor\">resistors<\/a>, is used to introduce \npredictable, linear resistance into the circuit, instead. Resistors with a very significant temperature dependency are known as\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Thermistor\">thermistors<\/a>, but are used rarely.\n<\/p><p>\nSome materials may impede the flow of currents in a time-dependent manner, usually associated with the creation of magnetic or electric fields; this \neffect is <u>not<\/u> considered to be resistance, and is studied separately.\nOther materials may exhibit resistance patterns that are time-invariant, but vary in relation to some other parameter (e.g., in semiconductor junctions \nthat respond to the magnitude and direction of external electromotive forces); this effect is still resistance, but such devices are not called resistors.\n<\/p><h3>Voltage (V) <a name=\"6\" href=\"#6\" class=\"slink\" id=\"6\">[link]<\/a><\/h3>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Voltage\">Voltage<\/a> across any two points in the circuit has several definitions, but most intuitively, can be \nunderstood as the measure of an electromotive force that would drive a current through a hypothetical conductor connected across these two points.\n<p>\nThis force does not depend merely on the difference in the number of accumulated elementary charges - but also on the \"pressure\" these charges are under \nto leave their current spots, due to the influence of resulting electric and magnetic fields. As a simple illustration, consider depositing 100 electrons \non two conductive, reasonably distant plates - one of them large, and one small. Electrons deposited on the smaller plate will be packed more densely, and \ntherefore, electric repulsion between them will be far more violent than in the other material.\nWhen the two elements are bridged with a conductive path, some current will briefly flow as charge densities (and not counts) equalize.\n<\/p><p>\nThe unit of voltage - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Volt\">volt<\/a> (V) - is a derived SI unit defined as the amount of electromotive force needed \nto induce the current of one ampere across a conductor with the resistance of one ohm.\n<\/p><p>\nIn any electrical circuit, voltage measurements are meaningful only if relative to a clearly defined reference point; many \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Direct_current\">direct current<\/a> circuits use a common rail connected to the negative pole of a two-pole\nbattery as a \"ground\", customarily defined as 0 volts - in which case, all measurements are implicitly understood to relate to this point. That said, \nthe voltage between this rail and any other external, insulated object is essentially undefined: one man's 0V may be somebody else's -20,000V, even if \nsimply due to electrostatic <a href=\"http:\/\/abstrusegoose.com\/strips\/today_we_celebrate_our_INDEPENDENCE_DAY.PNG\">charge buildup<\/a>. \nA well-designed physical connection to soil may provide a common 0V reference, but this\ncomes with its own perils, and therefore, is used only when absolutely necessary.\n<\/p><p>\nIn almost all signal-processing circuits, useful information is \nencoded using changing voltage levels, rather than current magnitudes; exceptions happen, but are fairly rare.\n<\/p><p>\nSpecifying steady voltages is not rocket science. but some ambiguity exists when it comes to constantly changing \nvoltages in <a href=\"http:\/\/en.wikipedia.org\/wiki\/Alternating_current\">alternating current<\/a> circuits, where the direction of current flow \nperiodically reverses. In these settings, three different values may be used:\n<\/p><ol><li>Peak voltage (V<sub>peak<\/sub>): measures voltage amplitude of the waveform in reference to its center point.\nThis approach is commonly employed when dealing with small-signal AC circuits.<\/li><li>Peak-to-peak voltage (V<sub>p-p<\/sub>): measures voltage difference between the lowest and highest voltage observed in the waveform \n(equal to <code>2 * V<sub>peak<\/sub><\/code>). This is conceptually similar to how signal voltage is measured in DC circuits.<\/li><li><a href=\"http:\/\/en.wikipedia.org\/wiki\/Root_mean_square\">Root mean square<\/a> voltage (V<sub>RMS<\/sub>): uses the value of an equivalent DC voltage\nthat would deliver the same power to the load (about <code>0.71 * V<sub>peak<\/sub><\/code> for \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Sine_wave\">sine wave<\/a> currents). This approach is particularly common in electrical power distribution\ncalculations.\n<\/li><\/ol>\nThis graph demonstrates the difference more \nsuccinctly:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/v_measurement.jpg\" alt=\"image\" \/><\/p><p>\nAlways be aware which measurement method is being used; for example, \"110V\" AC mains voltage is actually given as RMS; this means that the actual sine\nwave swing is 154V, and peak-to-peak voltage is close to 310V (ouch!).\n<\/p><h3>Ohm's law <a name=\"7\" href=\"#7\" class=\"slink\" id=\"7\">[link]<\/a><\/h3>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Ohm's_law\">Ohm's law<\/a> is one of the most important rules describing the behavior of electrical circuits by \nformalizing the relationship between currents, voltages, and resistances - and a pretty straightforward one, too. The law states that the steady\ncurrent flowing through a conductor is directly proportional to the voltage applied to it, and inversely proportional to its resistance:\n<p> \n<code>I = V\/R<\/code>\n<\/p><p>\nThe same equation tells us that the voltage measured across a conductor must be equal to the steady current applied to it, times resistance:\n<\/p><p>\n<code>V = IR<\/code><\/p><p>\nAnd lastly, that the resistance of a conductor can be derived from the observed current and the voltage across its terminals:<\/p><p>\n<code>R = V\/I<\/code>\n<\/p><p>\nThis law is what makes resistors such useful components: they can be used to limit the current flowing through parts of the circuit to a specific,\ndesired value; or, by creating voltage differences across resistor terminals, to create lower voltages for a variety of purposes.\n<\/p><p>\nOhm's law also makes it easy to turn ammeters into <a href=\"http:\/\/en.wikipedia.org\/wiki\/Voltmeter\">voltmeters<\/a> and \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Ohmmeter\">ohmmeters<\/a>. Voltmeters can simply measure the very low current flowing through when a large resistor of \nknown value is placed across the test points (this resistor needs to be large enough not to disrupt the circuit in an appreciable way) - and then \ncompute the voltage according to the <code>V = IR<\/code> rule; ohmmeters may be constructed by applying a known voltage to an unknown resistive load, \nmeasuring the resulting current, and then solving the equation for R.\n<\/p><h3>Joule's law <a name=\"8\" href=\"#8\" class=\"slink\" id=\"8\">[link]<\/a><\/h3>\nPower (P) is a measure of the rate at which work is performed, or at which energy is converted into another form. In electrical circuits, \na form of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Joule%27s_law\">Joule's law<\/a> gives us the following formula:\n<p>\n<code>P = IV<\/code>\n<\/p><p>\n...where <code>I<\/code> is the current through a component, and <code>V<\/code> is the observed voltage drop. This relationship is actually intuitive: if \nthere is no current flowing, obviously no useful work is being done by a circuit; and if no voltage drop occurs, all the current must be passed as-is,\nrather than utilized in any way.\n<\/p><p>\nElectrical energy may be dissipated as heat or radiation, converted to motion, stored in electromagnetic fields, in electrochemical reactions, and\nso forth; in all cases, the equation gives you an accurate picture of the sum of all the processes at work, based on the observed or predicted\nvoltage and current alone.\n<\/p><p>\nHeat dissipation is of particular interest in electronics, as it must be kept in check to avoid destroying resistive loads. By combining Joule's law\nand Ohm's law, it can be shown that the heat dissipated by a resistor is given by any of these two equations:\n<\/p><p><code>P = I\u00b2R<\/code><br \/><code>P = V\u00b2\/R<\/code><\/p><p>\nThis is pretty useful: you can use these rules to show that it's impossible to exceed the power rating of a 1\/4 watt, 150 \u2126 resistor working \nin a 5V circuit; and that a 15 \u2126 resistor will light up the night\nsky in that same setting, unless the current through it is limited by other means.\n<\/p><h3>Intermission: working with decibels <a name=\"9\" href=\"#9\" class=\"slink\" id=\"9\">[link]<\/a><\/h3>\nIn component datasheets and some calculations, engineers often opt to express the ratio between power quantities in terms of a somewhat arbitrary but \nconvenient logarithmic measure - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Decibel\">decibels<\/a> (dB):\n<p>\n<code>L<sub>dB<\/sub> = 10 * log<sub>10<\/sub>(P<sub>1<\/sub> \/ P<sub>2<\/sub>)<\/code>\n<\/p><p>\nFor example, a 25% increase in the power of an acoustic wave is equal to about +1.0 dB; a two-fold increase will be equal to about +3 dB; and a 100-fold \nincrease will be +20 dB. \n<\/p><p>\nIn electronics, it is usually more desirable to compare voltages and signal amplitudes, rather than to directly measure power. Since the relation between \nvoltage and power is usually quadratic - in these uses, decibels are calculated using a slightly different formula:\n<\/p><p>\n<code>L<sub>dB<\/sub> = 10 * log<sub>10<\/sub>(V<sub>1<\/sub>\u00b2 \/ V<sub>2<\/sub>\u00b2) = 20 * log<sub>10<\/sub>(V<sub>1<\/sub> \/ V<sub>2<\/sub>)<\/code>\n<\/p><p>\nTherefore, a 50% increase in voltage is +1.0 dB, a two-fold increase is about +6 dB, and so forth. There is nothing profoundly significant about\ndecibels, but they will crop up every now and then - so it's best to get this over with right away.\n<\/p><h2>Time-dependent currents <a name=\"10\" href=\"#10\" class=\"slink\" id=\"10\">[link]<\/a><\/h2>\nSome types of linear electronic components may exhibit time-dependent (and therefore, signal frequency dependent) current characteristics. This pattern \nis generally associated with the (reversible) storage of some of the supplied energy, most commonly in the form of electromagnetic fields.\n<h3>Capacitance (C) <a name=\"11\" href=\"#11\" class=\"slink\" id=\"11\">[link]<\/a><\/h3>\nWhen two conductive plates are brought close to each other, it becomes unusually easy to store opposing electrostatic charges on them: the strong \nelectromagnetic field caused by electrons drawn away from one of the plates will be largely offset by a symmetrical field caused by charges pushed \nonto the other plate. The closer the plates are together, the larger surface area they have, and the better the medium between them is at guiding and \namplifying the resulting electric field (due to the field-induced alignment of \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Dipole\">dipoles<\/a>) - the greater \nnumber of elementary charges can be deposited on one end, and removed on the other, before the electromotive force opposing this process becomes as \nstrong as the voltage driving it. Note that this works only if the charge transfer is symmetrical, though: if charge carriers are not efficiently \nremoved from one of the plates, they can't be efficiently added on the other one, not without having to fight a much stronger, unbalanced\nelectric field.\n<p>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Capacitance\">Capacitance<\/a> is the measure of this phenomenon: the ratio of the number of charges deposited \n(in coloumbs), versus the electromotive force needed to deposit them (in volts):\n<\/p><p>\n<code>C = Q\/V<\/code>\n<\/p><p>Another way to look at it is that one <a href=\"http:\/\/en.wikipedia.org\/wiki\/Farad\">farad<\/a> (F) - the unit of capacitance - corresponds to the \nsteady current of one ampere being supplied for one second, and resulting in one volt potential difference forming across the plates:\n<\/p><p>\n<code>C = I*t\/V<\/code>\n<\/p><p>\nMost electronic components are designed to keep their inherent capacitance at a negligible level; some parasitic capacitance may develop between \nparallel wires or circuit board traces, but this is a concern only in certain specialized applications. Intentional capacitances are usually introduced where\nneeded with dedicated components - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Capacitor\">capacitors<\/a>; these devices usually consist of two long ribbons of a \nmetallic conductor, separated by <a href=\"http:\/\/en.wikipedia.org\/wiki\/Dielectric_medium\">dielectric<\/a> film.\n<\/p><p>\nThe interesting property of capacitors is that when subjected to a constant external voltage, they initially allow the flow of current - even though \nelectrons do not actually pass through the gap, they can perform useful work in other parts of the circuit on their way onto, or away from, the plates. \nThis current will gradually drop to zero as the voltage across the plates approaches the supply voltage, though - and eventually, \nall apparent conduction will cease:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/cap-charging.jpg\" alt=\"image\" \/><\/p><p>\nWhen the capacitor is then disconnected from the charging source, it will retain the charge imbalance (in theory, indefinitely); the stored\ncharges will create a transient current as soon as a conductive path is created across the capacitor terminals, a process that will return the stored\nenergy, and return the capacitor to its discharged state. The speed at which this happens is given by a differential equation, but it's usually \nsufficient to know the following (R is resistance through which the device is charged or discharged):\n<\/p><p>\n<code>t = 1\/3RC \u2192 I = 75%, V = 25%<\/code><br \/><code>t = 1\/2RC \u2192 I = 60%, V = 40%<\/code><br \/><code>t = 2\/3RC \u2192 I = 52%, V = 48%<\/code><br \/><code>t = 1RC \u2192 I = 37%, V = 63%<\/code><br \/><code>t = 2RC \u2192 I = 14%, V = 86%<\/code><br \/><code>t = 3RC \u2192 I = 5%, V = 95%<\/code> - practically fully charged<br \/><code>t = 4RC \u2192 I = 2%, V = 98%<\/code><br \/><code>t = 5RC \u2192 I = 0.7%, V = 99.3%<\/code>\n<\/p><p>\nCapacitors often serve as a simple energy-storage device - but their other interesting property is that while they largely block conduction for \nlow-frequency currents (spare for the brief time when they are getting charged), they allow apparent conduction at higher frequencies: if the current is reversed before the charging is complete, there is\na continuous reciprocating motion of charges through the surrounding circuit. This effect is capacitance-, current-, frequency-, and waveform-dependent - \nbut in essence, any given capacitor charged and discharged with a well-controlled alternating current acts as a simple \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/High-pass_filter\">high-pass filter<\/a>; average current \nthrough one particular capacitor in function of the frequency of the driving sine wave is shown in this graph:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/cap-highpass.jpg\" alt=\"image\" \/><\/p><p>\nWe will discuss frequency filters in more detail later on.\n<\/p><h3>Inductance (L) <a name=\"12\" href=\"#12\" class=\"slink\" id=\"12\">[link]<\/a><\/h3>\nAs mentioned earlier, charged particles tend to interact with other particles moving in relation to them. On a macroscopic scale, this manifests itself\nas a magnetic force between materials with aligned electron spins; and more interestingly, as an electromotive force experienced by conductors subjected \nto magnetic field gradients in space or in time. In essence, any varying magnetic field, perhaps created by a varying current in a conductor or by \nmoving a permanent magnet around, may induce transient voltage in any other conductor subjected to this influence; this process is the subject of \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Faraday's_law_of_induction\">Faraday's law of induction<\/a>.\n<p>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Inductor\">Inductors<\/a> are a class of devices consisting of conductors arranged to create strong, coherent\nmagnetic fields; these fields will then induce voltages in the conductor in response to changing currents, resisting these changes - a sort of an\ninertial effect. When connected to a supply, such a component begins in a non-conductive state, and then gradually ramps up current as the field becomes\nsaturated. Later, when the supply is cut off, the energy of the collapsing magnetic field is returned into the circuit, sustaining the flow of current for a \nbrief while (or, if this is not possible, causing a substantial, if very short-lived, voltage to build up before all the energy is dissipated internally).\n<\/p><p>\nThe unit of inductance - one <a href=\"http:\/\/en.wikipedia.org\/wiki\/Henry_%28unit%29\">henry<\/a> - is defined as the behavior of a conductor that, when \nsubjected to a linear change of current by one ampere per second, opposes the change by developing a voltage of one volt across its terminals:\n<\/p><p>\n<code>L = V*t\/I<\/code>\n<\/p><p>\nLike with capacitors, determining the voltage and current through its terminals at any given time requires solving a differential equation; but this\ntable is often good enough:\n<\/p><p>\n<code>t = 1\/3L\/R \u2192 I = 25%, V = 75%<\/code><br \/><code>t = 1\/2L\/R \u2192 I = 40%, V = 60%<\/code><br \/><code>t = 2\/3L\/R \u2192 I = 48%, V = 52%<\/code><br \/><code>t = 1L\/R \u2192 I = 63%, V = 37%<\/code><br \/><code>t = 2L\/R \u2192 I = 86%, V = 14%<\/code><br \/><code>t = 3L\/R \u2192 I = 95%, V = 5%<\/code> - practically fully conducting<br \/><code>t = 4L\/R \u2192 I = 98%, V = 2%<\/code><br \/><code>t = 5L\/R \u2192 I = 99.3%, V = 0.7%<\/code>\n<\/p><p>\nNote the relation of the definition of inductance to that of capacitance; and the relationship between the time characteristics. \nInductors behave in a manner conceptually opposite to capacitors, permitting low-frequency or static currents to flow almost freely (after\na brief ramp-up period), but creating a \nsignificant, current-dependent voltage drop at high frequencies (a low-pass filter), as all the energy is constantly consumed \non reversing the magnetic field. Combining the two in clever ways is also a convenient way to build oscillators; more about this soon.\n<\/p><h3>Reactance (X) and impedance (Z) <a name=\"13\" href=\"#13\" class=\"slink\" id=\"13\">[link]<\/a><\/h3>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electrical_reactance\">Reactance<\/a> is a parameter describing the <i>apparent<\/i> resistance caused by\ncapacitive and inductive behaviors. For sine wave alternating currents of a given frequency (f), reactance can be computed as:\n<p>\n<code>X<sub>C<\/sub> = -1 \/ (2\u03c0fC)<\/code><br \/><code>X<sub>L<\/sub> = 2\u03c0fL<\/code><br \/><code>X<sub>total<\/sub> = X<sub>C<\/sub> + X<sub>L<\/sub><\/code>\n<\/p><p>\nIt is important to underscore that these equations apply <u>only<\/u> to steady, symmetrical, sine wave alternating currents, though.\n<\/p><p>\nReactance and resistance are distinct phenomena, with the former being associated with reversible energy storage, and the latter, with one-way losses;\nmixing them up is usually not desirable - but neither is considering them completely separate; \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electrical_impedance\">impedance<\/a> is a convenient way to analyze resistance and reactance together in the form\nof a complex number (<code>j = \u221a-1<\/code>):\n<\/p><p>\n<code>Z = R + j*X<\/code>\n<\/p><p>\nComplex impedance can be used in place of resistance to extend Ohm's law and many other fundamental principles in electronics; but in many cases,\nits meaning is more prosaic: when the behavior of a circuit is dominated by one of the characteristics (e.g., just resistance), \"impedance\" may be\njust a shorthand for that particular property.\n<\/p><h2>Basic electronic components <a name=\"14\" href=\"#14\" class=\"slink\" id=\"14\">[link]<\/a><\/h2>\nThis section gives a basic overview of common, discrete components used in electronic circuits, along with their schematic representations. Some of them\nmay have multiple symbols due to random differences between the United States and the rest of the world; in such cases,\nyou should almost certainly memorize all variants.\n<h3>Voltage supplies <a name=\"15\" href=\"#15\" class=\"slink\" id=\"15\">[link]<\/a><\/h3>\nIdeal voltage sources maintain a constant voltage across their terminals (or a constant voltage amplitude, in case of AC circuits), regardless of the \ncurrent flowing through a connected load. In real life, the behavior of voltage sources is approximated to some extent by a variety of \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Power_supply\">power supplies<\/a>, including batteries - but with two caveats:\n<ul><li>Only a finite current can be supplied; once the source-specific maximum is reached, the device may reduce output voltage, shut down, or simply\ncatch fire.<\/li><li>Some dependency between the supplied current and observed terminal voltage is commonly seen. This dependency may be nearly linear\nin <a href=\"http:\/\/www.linear.com\/pc\/images\/product\/fullsize\/111567_1.jpg\">unregulated sources<\/a> or\n<a href=\"http:\/\/1.bp.blogspot.com\/_bcej-id-XqQ\/THF35jTHKeI\/AAAAAAAAAu4\/Nf9eVvKO3a4\/s400\/battIV.gif\">batteries<\/a>; or steeply non-linear in\n<a href=\"http:\/\/pubpages.unh.edu\/~mwidholm\/nanosat\/IVhi.gif\">solar cells<\/a> and regulated supplies.\n<\/li><\/ul>\nMost of the supplies used in electronic circuits are between 3 and 12V, and can supply between 100 mA and several amps. Typical schematic symbols include:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/voltage.jpg\" alt=\"image\" \/><\/p><p>\nThe behavior of non-ideal voltage sources is sometimes coarsely accounted for by defining the hypothetical internal impedance of the supply - a resistor that,\nwhen introduced in series with an ideal voltage supply, would produce a similar current-dependent voltage drop. This model is seldom perfect, but can be a \nuseful analytic tool. For example, one of the important rules in circuit design is that high impedance (low current capacity) signal\nsources should not be driving low impedance (power-hungry) loads, to prevent overloading the source and distorting the input signal;\nthis also applies to voltage supplies.\n<\/p><h3>Current sources <a name=\"16\" href=\"#16\" class=\"slink\" id=\"16\">[link]<\/a><\/h3>\nPure current sources are expected to maintain a constant current flowing through the connected load, by adjusting the force driving it - that is, the\nvoltage - as necessary. This implies that open-circuit voltages should rise with no upper bound, until the force is significant enough to create\na conductive path through free space; that may require thousands or millions of volts in everyday cases, and therefore, is not a very realistic scenario. \nReal-world current supplies usually have a modest maximum voltage they can produce in an effort to maintain a particular rate of charge flow.\n<p>\nThese types of sources are not commonly encountered in nature, but constructing them from active components is useful in certain cases (for example,\nfor charging capacitors at a linear rate). Their standard symbol is:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/current.jpg\" alt=\"image\" \/><\/p><p>\nThis symbols is rarely seen, though - current supplies are usually integral to the circuit that needs them, with the circuit itself driven by\na voltage source.\n<\/p><h3>Common passive components <a name=\"17\" href=\"#17\" class=\"slink\" id=\"17\">[link]<\/a><\/h3>\nPassive components are named so for their inability to amplify signals; this is in contrast with devices such as the now largely obsolete\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Vacuum_tube\">vacuum tubes<\/a>, or their contemporary\nsuccessors - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Transistor\">transistors<\/a>.\nThe boundaries of this subset are somewhat fuzzy, to be sure - power supplies, optoelectronics, and electromechanical components are commonly excluded, for \nexample; so tend to be all semiconductors, even though some of them are not meant to be amplifiers. In any case, some of the\nuncontested passive elements commonly used in low-voltage electronic circuits include:\n<p>\n<b>Resistors:<\/b> most of the general-purpose resistors consist of nickel-chromium metal films, and are manufactured with a power rating of 0.125 or 0.25W,\naccuracy of 5% or 1%, and a temperature dependence under 0.01% per kelvin (thermistors are designed with a much higher temperature coefficient); \ndifferent tolerances, power ratings, and designs are available at a premium, but seldom necessary. It's definitely better to stock a full range of\ncommon resistances using the lowest-cost resistors (around $0.02 a piece), rather than cherry-pick more expensive ones for every individual circuit.\n<\/p><p>\nStandard (\"preferred\") values are selected to make most sense given the expected tolerances, with\n<a href=\"http:\/\/www.logwell.com\/tech\/components\/resistor_values.html\">E6 scale<\/a> being used most commonly. These values are usually encoded\nusing <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electronic_color_code\">color codes<\/a> on through-hole resistors, or as a two-digit value and a single-digit\nexponent on surface mount devices (e.g., \"473\" means 47 * 10<sup>3<\/sup> \u2126, or 47 k\u2126).\n<\/p><p>\nIn schematics, the ohm symbol is often omitted - or sometimes substituted with the letter \"R\" if no unit prefix is present. For example, 1 k\u2126 may \nbecome \"1k\", while 100 \u2126 may become \"100\" or \"100R\". To further shorten the notation, \"R\" or the unit prefix is moved in place of the decimal dot -\ne.g., 1.2 k\u2126 may be written as \"1k2\". Commonly used resistances span from 1 \u2126 to 10 M\u2126; stocking anything outside these limits\nis probably not very useful in everyday work.\n<\/p><p>\nThe usual resistor symbols are:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/resistor.jpg\" alt=\"image\" \/><\/p><p>\n<b>Capacitors:<\/b> low-cost, general-purpose capacitors use ceramics, polyester, or metal oxides as a dielectric; and aluminum, tantalum,\nor an electrolyte solution to form the plates themselves; other compositions are available where extreme stability or high voltage capability are required,\nbut tend to be more expensive - and again, make sense only in specialized applications.\n<\/p><p>\nElectrolytic capacitors enjoy some of the highest capacitances in proportion to their cost and size - but need to be polarized, work well only for \nfairly low voltages, have some leakage current, and tend to exhibit non-trivial resistance (denoted as <a href=\"http:\/\/en.wikipedia.org\/wiki\/Equivalent_series_resistance\">ESR<\/a>,\nand limiting their ability to deal with high-frequency signals);\nso avoiding electrolytics as long as possible is generally a good idea (cheap multi-layer ceramic capacitors - MLCC - are available up to at least 10 \u00b5F).\n<\/p><p>\nCommonly used capacitances span from 1 pF to 1 mF in everyday applications, and go up to several hundred farads for supercapacitors used in energy storage.\nWhen shopping, keep in mind that in some markets, unit prefixes \"nano\" and \"milli\" are not used when indicating capacitance; it's a silly practice -\n47 nF is more readable than 0.047 \u00b5F or 47,000 pF - but you may have to live with it.\n<\/p><p>\nSmall capacitors use the same three-digit exponential notation as used for resistors (and similar preferred\nvalues) to express capacitance in picofarads (e.g., \"224\" = 22 * 10<sup>4<\/sup> pF = 220 nF); larger capacitors usually display \nthe capacitance and unit prefix as-is. The following symbols are used for capacitors:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/capacitor.jpg\" alt=\"image\" \/><\/p><p>\n<b>Inductors:<\/b> constructed by making a coil out of a piece of a very long but low-resistance wire, so that the concentric magnetic\nfield lines around the conductor add up inside the coil to form a powerful, coherent field \n(<a href=\"http:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/11\/Basic_Inductor_with_B-field.svg\/400px-Basic_Inductor_with_B-field.svg.png\">image<\/a>).\nBecause of the length of the wire needed, many small inductors will have a noticeable resistance - usually between 0.5 and 20 \u2126.\n<\/p><p>\nCommon inductor values in electronic circuits range from 1 \u00b5H to 470 mH or so. The usual symbol of an inductor is:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/inductor.jpg\" alt=\"image\" \/><\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Transformer\">Transformers<\/a>:<\/b> constructed by pairing two inductors, typically\nwrapped around a common ferromagnetic core that guides and contains the electromagnetic field to improve performance.\nThe interesting property of this arrangement is that changes in the current flowing through one of the \ninductors will generate a back electromotive force in the other, coupled coil. Coil turn ratio determines the ratio between the driving and the\ninduced voltage, making transformers extremely useful for converting voltages (the law of \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Conservation_of_energy\">conservation of energy<\/a> ensures that transformers are not a free energy scheme:\nthe power (V*I) must remain the same, and every increase in voltage reduces current capability proportionately). Another common use is\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Galvanic_isolation\">galvanically isolating<\/a> circuits for safety reasons - i.e., only magnetic coupling, but no\ndirect charge flow pathways, would be maintained.\n<\/p><p>\nTheir primary caveat is that they do not exhibit their trademark behavior when fed steady DC signals (once the core is saturated and peak magnetic\nfield is attained, a significant winding current is admitted and largely goes to waste); and that optimum efficiency (often 95%+)\nis generally maintained over\na fairly narrow range of AC frequencies (depending on the construction of the transformer). Low-frequency transformers, such as the ones used for 50-60 Hz\nmains signals, need large and heavy, slowly saturating cores; while high-frequency transformers can be much smaller.\n<\/p><p>\nThe usual symbol is:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/transformer.jpg\" alt=\"image\" \/><\/p><p>\nSometimes, windings may have multiple taps, or additional windings can be provided as a feedback mechanism for building high\nvoltage <a href=\"http:\/\/en.wikipedia.org\/wiki\/Flyback_transformer\">flyback transformers<\/a> that exploit resonant frequencies of the ferrite core.\n<\/p><h3>Electromechanical and optoelectronic devices <a name=\"18\" href=\"#18\" class=\"slink\" id=\"18\">[link]<\/a><\/h3>\nThere is a multitude of components that allow electronic circuits to interface with the outside world - from \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Pyroelectricity\">heat<\/a> or\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Geiger-Mueller_tube\">radiation<\/a> detectors, to \nvarious <a href=\"http:\/\/en.wikipedia.org\/wiki\/Vacuum_fluorescent_display\">exotic display technologies<\/a>, to\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Thermoelectric_effect#Peltier_effect\">thermal transfer devices<\/a>. Listing them all is a futile task -\nbut here's an overview of some of the most common components you may encounter:\n<p>\n<b>Switches:<\/b> one of the most rudimentary electromechanical devices, switches interrupt and close circuits in response to an externally applied force. \nThese devices come in a wide variety of designs (rotary, paddle \/ toggle, rocker, pushbutton, slide, etc), and an even greater variety of sizes. \nOther than the power rating,\ntheir most important parameters are the number of actual switch positions, the number of \"throws\" (signal outputs the switch can alternate between -\nthis must obviously be equal or smaller than the number of positions allowed), the number of \"poles\" (separate switching pathways put in a single \npackage), and the type of switching action (sustained \/ latching or momentary; with momentary switches further divided between normally closed or \nnormally open designs - \"NC\" and \"NO\").\n<\/p><p>\nSymbol for a single pole, single throw switch is shown on the left; a momentary, normally open switch is in the middle; and\na single pole double throw is shown on the right:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/switch.jpg\" alt=\"image\" \/><\/p><p>\nMechanical switches may also be operated by magnetic fields (<a href=\"http:\/\/en.wikipedia.org\/wiki\/Reed_switch\">Reed switch<\/a>),\ngravity (mercury and ball-type <a href=\"http:\/\/en.wikipedia.org\/wiki\/Tilt_switch\">tilt switches<\/a>), liquid levels\n(<a href=\"http:\/\/en.wikipedia.org\/wiki\/Float_switch\">float switches<\/a>), and much more.\n<\/p><p>\n<b>Simple actuators - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Electromagnet\">electromagnets<\/a>,\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Loudspeaker\">speakers<\/a>, \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Solenoid\">solenoids<\/a>, \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Relays\">relays<\/a><\/b>: as noted previously, electric currents flowing through properly oriented wires\ncan induce strong, coherent magnetic fields; these magnetic fields may, in turn, interact with materials where electron spins are coherently aligned:\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Magnet\">permanent magnets<\/a> and \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Ferromagnetism\">ferromagnetic<\/a> \/ \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Paramagnetism\">paramagnetic<\/a> metals. An\ninductor designed with this in mind is called an electromagnet; the three most basic uses of electromagnets are speakers (where a diaphragm is being \nmoved back and forth to create audible sound waves), solenoids (pushing or pulling a plunger - e.g., in car door locks or valves), and relays\n(mechanical switches operated by an electromagnet).\n<\/p><p>\nGeneric electromagnets and solenoids are usually drawn as inductors, and merely annotated to indicate their intended function. The symbol of a loudspeaker\nis:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/speaker.jpg\" alt=\"image\" \/><\/p><p>\n...and for relays, we tend to use:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/relay.jpg\" alt=\"image\" \/><\/p><p>\nAmusingly, relays and solenoids can be used as crude speakers when driven by audio-frequency signals.\n<\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Electric_motor\">Motors<\/a> and \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electrical_generator\">generators<\/a>:<\/b> motors exploit the interaction of electromagnets and magnetic\nmaterials to put their output shaft in\ncontinuous rotary motion. This can be achieved by driving the coils directly with alternating sine wave currents \n(<a href=\"http:\/\/en.wikipedia.org\/wiki\/Synchronous_motor\">synchronous motors<\/a>), by providing carefully timed voltage pulses \n(<a href=\"http:\/\/en.wikipedia.org\/wiki\/Brushless_motor\">brushless<\/a> and <a href=\"http:\/\/en.wikipedia.org\/wiki\/Stepper_motor\">stepper motors<\/a>), or \nwith steady direct currents, mechanically switched by commutators (<a href=\"http:\/\/en.wikipedia.org\/wiki\/Brushed_motor\">brushed motors<\/a>).\n<\/p><p>\nThis operating principle is reversible: when rotated by an external force, motors serve as generators, producing a voltage across their terminals.\nA striking demonstration of this is connecting the terminals of two similar, brushed DC motors together: rotating one of them will induce a voltage \nsufficient to turn the other.\n<\/p><p>\nThe symbols used for motors are:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/motor.jpg\" alt=\"image\" \/><\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Piezoelectricity\">Piezoelectric<\/a> crystals:<\/b> certain materials tend to generate voltages in response\nto mechanical strain, and vice versa - contract or expand in response to applied fields or the currents flowing through them; in some settings,\nthis action may in turn affect the applied voltage or admitted current, resulting in oscillating action.\n<\/p><p>\nThese principles are used in a number of piezoelectric devices, the most common of which are \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Crystal_oscillator\">crystal oscillators<\/a> - where the crystal vibrates at a \nresonant frequency when subjected to an external current, similar to some capacitor-inductor arrangements we will discuss later on. \nOther applications use exotic, precision \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Piezoelectric_motor\">motors \/ actuators<\/a>, piezoelectric \ntransducers used to sense or generate soundwaves, pressure sensors, and so forth.\n<\/p><p>\nThe usual symbol of a crystal oscillator is:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/oscillator.jpg\" alt=\"image\" \/><\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Potentiometer\">Potentiometers<\/a>:<\/b> manually adjustable resistors; can be constructed by placing a conductive \nwiper across a resistive substrate. Available as miniature board-mounted \"trim pots\" that need to be operated with a screwdriver, and are meant to be\nadjusted at assembly or servicing time; larger panel-mount devices with a knob or a slider used for interfacing with users; more expensive but precise \nmulti-turn units of both flavors; and potentiometer-based position-, deflection-, or pressure-sensing components\nthat provide the circuit with feedback about the position of monitored mechanical assemblies. In all cases, the usual symbol is something close to:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/potentiometer.jpg\" alt=\"image\" \/><\/p><p>\nPotentiometers are usually not meant to conduct significant currents (and dissipate the resulting heat); therefore, using them requires some care.\n<\/p><p>\n<b>Trimmer caps:<\/b> manually adjustable capacitors; available only for small capacitances, usually controlled by gradually moving capacitor plates apart\nwhen a knob or a screw is rotated. Panel-mount devices were once very popular in radio frequency circuits to control the frequency of\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Heterodyne\">heterodyne receivers<\/a> - but are now largely displaced by digital frequency generation that can be\ncontrolled with mechanically simpler input devices.\n<\/p><p>\nThe symbol is analogous to that of a potentiometer - except that the \"wiper\" is not actually a terminal of the device:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/variable_cap.jpg\" alt=\"image\" \/><\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Incandescent_light_bulb\">Incandescent lamps<\/a>:<\/b> simple, well-known components that exploit the heat-driven \nlight emission of a resistive metal wire placed under vacuum or in a low-pressure, inert gas (to prevent oxidation and minimize convection-based heat transfer). As the wire\nheats up, its resistance tends to rise - resulting in a self-limiting current flow that prevents the device from overheating and being destroyed within the \n(usually very generous) range of voltages it is designed for. Their primary downside is that they radiate most of their energy in the infrared range, which \nis not very desirable; and that they are slow to turn on and switch off, making them unsuitable for optical data signaling.\n<\/p><p>\nLightbulbs are fairly unremarkable electrically, other than their interesting non-linear current characteristics: while the wire is cold, a \nsignificant current can flow, but only for a limited time. This sometimes results in non-conventional uses of lightbulbs in high-quality sine \nwave oscillators, such as the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Wien_bridge_oscillator\">Wien bridge<\/a>.\n<\/p><p>\nIncandescent lamp symbols:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/lightbulb.jpg\" alt=\"image\" \/><\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/LED\">Light-emitting diodes<\/a>:<\/b> \ncertain types of semiconductor p-n junctions tend to emit light as electrons and holes recombine; LEDs are designed to exploit this phenomenon, instead\nof making use of the electrical characteristics of the junction itself. Unlike incandescent light bulbs, LEDs dissipate most\nof the energy as light, not heat - spare for some resistive losses in the semiconductor itself; and can be started and stopped very quickly.\n<\/p><p>\nThese devices are more difficult to operate than lightbulbs, because they are very sensitive to the applied voltage - it must be enough to overcome the \njunction bias, but will result in a destructive current if off by as little as 0.2V. This is because diodes are only weakly conductive up until the \npotential of the junction is overcome - and shortly past that point, begin conducting like crazy. Current-limiting resistors or constant current \nsupplies are commonly employed to prevent trouble.\n<\/p><p>\nJunction voltage drop is commonly in the 1.1 - 1.4V range for infrared diodes, 1.6 - 2.0V for red LEDs, 2.0 - 2.3V for yellow and orange, 2.2 - 2.8V\nfor green, and 3.2 - 3.8V for blue, violet, white, and UV diodes. Commodity indicator LEDs are designed for currents between \n5 and 30 mA. These parameters will vary from one device to another, and should always be confirmed with the datasheet.\n<\/p><p>\nIn all other aspects, LEDs behave very much like a regular semiconductor diode, discussed in the next section. Their symbol is (arrow pointing in the\ndirection of current flow - toward the negative pole of the supply):\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/led.jpg\" alt=\"image\" \/><\/p><p>\n<b>Light sensors - <a href=\"http:\/\/en.wikipedia.org\/wiki\/Photovoltaic_cell\">photovoltaic cells<\/a>, \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Photoresistor\">photoresistors<\/a>, \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Photodiode\">photodiodes<\/a>:<\/b> three additional uses of semiconductor materials where we don't really\ncare about their electrical characteristics as such. Photovoltaic cells rely on a p-n junction to generate free charge carriers in response to photon \nexcitation, and also to separate charges - so that an electromotive force is created across their terminals, forming a voltage source. Photodiodes work \nin a very similar way, but are used with an external voltage to achieve a current dependent on the amount of light shining on a reverse-biased\njunction. Finally, photoresistors are monolithic semiconductors where a higher number of photon-induced charge carriers promotes conduction; compared\nto photodiodes, they are much slower, but have other favorable characteristics, such as a very broad sensitivity range.\n<\/p><p>\nTheir respective symbols:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/photosensitive.jpg\" alt=\"image\" \/><\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Microphone\">Microphones<\/a>:<\/b> a variety of sound-sensitive devices (generally variable resistors,\ncapacitors, or electromagnets) used to sample audio signals. The most common variety today are \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Electret_microphone\">electret microphones<\/a> that utilize an internal, permanently charged membrane to form\nan interesting type of a capacitor; but even a regular speaker and some vanilla capacitors exhibit some audio sensitivity.\n<\/p><p>\nThe usual symbol of a microphone is:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/microphone.jpg\" alt=\"image\" \/><\/p><p>\n<b>Fuses:<\/b> a variety of components designed to either irreversibly blow up (<a href=\"http:\/\/en.wikipedia.org\/wiki\/Fuse_%28electrical%29\">traditional\nfuses<\/a>), toggle a mechanical switch, or just temporarily stop conducting (<a href=\"http:\/\/en.wikipedia.org\/wiki\/Resettable_fuse\">PPTC fuses<\/a>)\nwhen the current flowing through them exceeds the desired value. Fuses are meant to protect more expensive or harder-to-service components from\nbeing affected when something goes wrong (e.g., a stray metal item shorts traces on the circuit board); and to prevent the circuit from overheating\nand starting a fire if the problem goes uncorrected. \n<\/p><p>\nThe use of fuses in low-voltage, low-power consumer electronics is often a matter of a judgment call; but if the power supply can source significant\ncurrents, enough to blow a hole in the circuit board, adding a fuse may be a good idea.\n<\/p><h3>\"Proper\" discrete semiconductors <a name=\"19\" href=\"#19\" class=\"slink\" id=\"19\">[link]<\/a><\/h3>\n<p>\n<b>Diodes:<\/b> a simple p-n junction that begins conducting well, but only in one direction, when an internal junction voltage (usually somewhere \naround 0.6 - 0.7V) is exceeded. It will not conduct in the opposite direction unless a much higher breakdown voltage is applied (this figure \ncommonly lies between 20 and 1000V, although diodes with much smaller breakdown voltages can be manufactured).\nTwo special types of diodes can be seen in electronic circuits along the \"normal\" ones: fast, low-threshold \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Schottky_diode\">Schottky diodes<\/a> used for small signal switching (with junction threshold of merely 0.2 - 0.4V);\nand <a href=\"http:\/\/en.wikipedia.org\/wiki\/Zener_diode\">Zener diodes<\/a> with very precisely defined reverse breakdown voltage, often used as \nvoltage references or \nvoltage-limiting <a href=\"http:\/\/en.wikipedia.org\/wiki\/Shunt_%28electrical%29\">shunts<\/a> that protect sensitive components. \n<\/p><p>\nThe current-voltage curve for a typical diode may look roughly like this:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/diode_iv.jpg\" alt=\"image\" \/><\/p><p>\nPast a certain forward or reverse voltage, any diode will happily attempt to conduct practically arbitrary currents. The interesting property of a diode\n(or any other p-n junction) is that while in this mode, the device will always maintain a potential close to that threshold voltage across its terminals: \nthis electric field is necessary to keep the junction conductive, and the diode will develop an apparent resistance needed to maintain it. \nWith higher currents, the measured voltage across the will increase subtly due to the resistance of the semiconductor material itself - \nbut in most cases, this effect is not very pronounced (some Schottky diodes are an exception).\n<\/p><p>\nThe symbol of a diode is an arrow pointing in the direction of current flow during forward-biased operation:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/diode.jpg\" alt=\"image\" \/><\/p><p>\nSome people use separate symbols for Zener \n(<a href=\"http:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/7\/7f\/Zener_diode_symbol.svg\/200px-Zener_diode_symbol.svg.png\">image<\/a>)\nand Schottky (<a href=\"http:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c9\/Schottky_diode_symbol.svg\/200px-Schottky_diode_symbol.svg.png\">image<\/a>)\ndiodes, but you should certainly not bank on this.\n<\/p><p>\nSome of the most common general-purpose diodes are\n<a href=\"http:\/\/www.nxp.com\/documents\/data_sheet\/1N4148_1N4448.pdf\">1N4148<\/a> (rated for 1 A, 100V reverse voltage) and\n<a href=\"http:\/\/www.fairchildsemi.com\/ds\/1N%2F1N4001.pdf\">1N400x series<\/a> (1 A, assorted voltages from 50 to 1000V);\n<a href=\"http:\/\/lcamtuf.coredump.cx\/electronics\/www.datasheetcatalog.org\/datasheet\/vishay\/85662.pdf\">BAT46<\/a> is a popular Schottky diode rated for 100 mA,\nwith a low-current voltage drop of merely 0.25V; while \n<a href=\"http:\/\/www.ing.unlp.edu.ar\/electrotecnia\/circelec\/1N4738.pdf\">1N47xx<\/a>\nis a family of popular Zener diodes with reverse breakdown voltages ranging from 3.3V to 200V.\n<\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Field-effect_transistor\">Field-effect transistors<\/a>:<\/b> transistors are a very important class of \nsemiconductor devices that, generally speaking, allow \"weak\" (low amplitude or high impedance) signals to control much greater currents flowing through\nthe device - a sort of an electronically controlled potentiometer. They are primarily used as signal amplifiers, impedance converters, and \ndigital switches.\n<\/p><p>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/MOSFET\">MOSFETs<\/a> - metal-oxide-semiconductor field effect transistors - are perhaps the most intuitive, if\nrecent, variety. Three-terminal, enhancement mode, n-channel MOSFETs (the most common type) consist of a nominally non-conductive n-p-n junction.\nOne of the terminals - called \"drain\" (D) - is connected to one of the n-type regions; \"source\" (S) terminal is connected to the p-type and the other n-type\nregion; and \"gate\" (G) is placed across the p-type substrate, isolated with a very thin layer of glass; interestingly, this layer of glass is \nso thin that it is easily destroyed by electrostatic discharge - requiring all MOSFETs to be handled with care.\n<\/p><p>\nThis arrangement of connections - shown below on the left - results in a normal p-n diode that allows conduction from source to drain - but no\nconduction the other way round; MOSFET transistors are operated with this junction reverse-biased - i.e., drain more positive than source in case of \nn-p-n devices.\n<\/p><p>\nThe interesting part is that the isolated gate-source pathway through the p-type semiconductor acts as a small capacitor: applying a voltage higher \nthan the source to the gate region will deposit a positive charge on the terminal, which pushes aways holes in the p-type semiconductor - and pulls \nfree electrons from n-type terminals in, forming a junction-free conduction path (shown on the right):\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/mosfet_operation.jpg\" alt=\"image\" \/><\/p><p>\nThere is some minimal gate voltage needed to cause a sufficient charge separation; this threshold voltage - V<sub>GS(TH)<\/sub> - \ndepends on the geometry of the transistor, and\ngenerally ranges from 1 to 2V (but may be as high as 3-4V for high power devices).\n<\/p><p>\nMOSFET transistors allow very significant currents to be controlled by applying small, very high impedance signals to the gate terminal \n(there is virtually no gate-source current observed, and this \"capacitor\" retains the charge even after the source is disconnected); the \nresistance of the created conductive channel is proportional to the applied voltage, resulting in a transistor-specific two- or three-digit \namplification factor (you shouldn't depend on its exact value, though).\n<\/p><p>\nOther common types of field-effect transistors are p-channel MOSFETs, which use p-n-p junctions, and switch on with a gate voltage is <u>lower<\/u> than the\nsource voltage (and have slightly inferior electrical characteristics); four-terminal MOSFETs with no internal connection between the source and\nthe middle semiconductor layer (the fourth terminal connected to this area is referred to as \"body\"), useful in some switching applications; less common \ndepletion mode MOSFETs, which have a reversed operation, and \nare normally conductive until a field is applied to disrupt the channel; and somewhat simpler, depletion-only\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/JFET\">JFETs<\/a>, which do not feature a glass insulator, and exhibit higher\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Transconductance\">transconductance<\/a>.\n<\/p><p>\nSymbols for n-channel and p-channel enhancement MOSFETs, and their on-by-default depletion mode equivalents, are shown below:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/mosfet.jpg\" alt=\"image\" \/><\/p><p>\nInteresting enhancement mode MOSFET transistors include\n<a href=\"http:\/\/www.fairchildsemi.com\/ds\/2N\/2N7000.pdf\">2N7000<\/a> (n-channel, 200 mA),\n<a href=\"http:\/\/www.onsemi.com\/pub_link\/Collateral\/NTD4858N.PDF\">NTD4858N<\/a> (n-channel, 73A), or\n<a href=\"http:\/\/www.st.com\/stonline\/books\/pdf\/docs\/11236.pdf\">STP12PF06<\/a> (p-channel, 12A).\n<\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Bipolar_junction_transistor\">Bipolar junction transistors<\/a>:<\/b> BJTs are a bit more messy predecessor\nto FETs, allowing the current between their two terminals - \"collector\" (C) and \"emitter\" (E) - to be controlled by a smaller current flowing from\nthe third terminal - \"base\" (B) - to the emitter.\n<\/p><p>\nBJTs consist of a nominally non-conductive junction, most commonly n-p-n - with the outer layers connected to the collector and emitter terminals,\nand a very thin p-type layer sandwiched in between connected to the base terminal (see image below, left).\n<\/p><p>\nIn normal operation, collector is connected to a more positive region, and emitter is more negative; in this configuration, the B-C junction\nis reverse biased and does not conduct. When a small current between the base and the collector is applied, however, the B-E junction depletion\nlayer collapses; and because of how thin the base layer is, the depletion layer on the B-C junction is also compromised. High-velocity electrons that\nare accelerated by the B-E voltage and inject from the emitter into the p-type semiconductor will easily make it to the collector layer - and a \nsignificant current will flow (a device-specific two- or three-digit current amplification will be seen, the exact value of which is - \nagain - not to be relied upon):\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/bjt_operation.jpg\" alt=\"image\" \/><\/p><p>\nBipolar junction transistors with n-p-n arrangements are, somewhat unimaginatively, called NPN. Their p-n-p counterparts (PNP) operate with\nthe emitter being more positive, and conduct when a current flows from the emitter to the base. The symbols of NPN and PNP transistors are:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/bjt.jpg\" alt=\"image\" \/><\/p><p>\nAll bipolar transistors are more hairy than MOSFETs, for a couple of reasons. Firstly, they can't be driven by extremely high impedance signals, because\nsome small base current must flow to enable conduction. Secondly, their diode junctions do not disappear fully when the transistor is\nconducting, resulting in an unavoidable voltage drop (V<sub>CE<\/sub>) across the device - usually between 0.1V and 0.4V when the transistor is\nsaturated (i.e., further increases of base current have no effect). Similarly to MOSFETs, there is also a minimum base-emitter voltage required to \novercome the potential of the BE junction (V<sub>BE<\/sub>), usually in the 0.6V - 0.8V range.\n<\/p><p>\nThe three most common BJTs are \n<a href=\"http:\/\/www.fairchildsemi.com\/ds\/2N\/2N3904.pdf\">2N3904<\/a> (NPN, 200 mA),\n<a href=\"http:\/\/www.fairchildsemi.com\/ds\/PN\/PN2222A.pdf\">PN2222A<\/a> (NPN, 1A), and\n<a href=\"http:\/\/www.fairchildsemi.com\/ds\/2N\/2N3906.pdf\">2N3906<\/a> (PNP, 200 mA). Higher-rated variants, such as\n<a href=\"http:\/\/www.classiccmp.org\/rtellason\/transdata\/fjn965.pdf\">FJN965<\/a> and\n<a href=\"http:\/\/www.datasheetcatalog.org\/datasheet\/fairchild\/FJC1386.pdf\">FJC1386<\/a> (NPN and PNP, 5A) can also be found, but MOSFETs\nare now being used almost exclusively in power applications.\n<\/p><p>\n<b>\"Derived\" transistors:<\/b> a variety of more exotic transistor arrangements exists. <a href=\"http:\/\/en.wikipedia.org\/wiki\/Thyristor\">Thyristors<\/a>\n(also known as silicon controlled rectifiers, SCRs) are four-layer semiconductors that behave like latching transistors - i.e., keep conducting even after\nthe base current stops; this effect can be also approximated by two discrete transistors \n(<a href=\"http:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/c\/c5\/Thyristor.svg\/500px-Thyristor.svg.png\">image<\/a>).\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Darlington_transistor\">Darlington transistor<\/a> is a pair of two NPN or PNP transistors stacked to achieve\na much higher current gain; and <a href=\"http:\/\/en.wikipedia.org\/wiki\/Sziklai_pair\">Sziklai pair<\/a> is its heterogeneous (NPN+PNP) counterpart. \nField effect transistors with multiple \"competing\" gate electrodes can also be seen, and are useful in signal mixing.\n<\/p><p>\nWith the operation of transistors explained, it's good to additionally mention\n<a href=\"http:\/\/lcamtuf.coredump.cx\/electronics\/en.wikipedia.org\/wiki\/Phototransistor\">phototransistors<\/a> - optoelectronic devices that\noperate very much like photodiodes, but feature internal gain; and optointerrupers - LED-phototransistor pairs mounted in a common enclosure,\nused for reflective or transmissive (slot) sensing in a variety of mechanical applications.\n<\/p><h2>Analog circuits <a name=\"20\" href=\"#20\" class=\"slink\" id=\"20\">[link]<\/a><\/h2>\nNow that we have the basic characteristics of electronic circuits, and the common components, sorted out, it's time to see what it's good for -\nstarting with traditional, analog circuitry.\n<h3>Components in parallel and in series <a name=\"21\" href=\"#21\" class=\"slink\" id=\"21\">[link]<\/a><\/h3>\nWhen considering the behavior of real-world circuits, it is often essential to understand the actual resistance, capacitance, or inductance of\nsignal pathways that may consist of more than one discrete component. Thankfully, this is fairly simple.\n<p>\nAs should be fairly obvious, several resistors placed in series will have an equivalent resistance equal to the sum of all individual resistances:\n<\/p><p>\n<code>R<sub>equiv<\/sub> = R<sub>1<\/sub> + R<sub>2<\/sub> + R<sub>3<\/sub> + ...<\/code>\n<\/p><p>\nResistors placed in parallel, on the other hand, have an equivalent resistance equal to the inverse of the sum of their inverse resistances:\n<\/p><p>\n<code>R<sub>equiv<\/sub> = 1 \/ (1\/R<sub>1<\/sub> + 1\/R<sub>2<\/sub> + 1\/R<sub>3<\/sub> + ...)<\/code>\n<\/p><p>\nThis latter calculation is somewhat cumbersome, but for simplicity, it's good to remember that any number of identical resistors in parallel will have an\nequivalent resistance equal to <code>R\/count<\/code>, and the power rating will increase accordingly; and that if and if one of the resistors in parallel \nhas a resistance several orders of magnitude lower than the rest, the resistances and power ratings of the remaining resistors may often be safely \nignored altogether.\n<\/p><p>\nExactly the same rules apply to the inductance and power ratings of inductors; and the voltage and current capability of\nvoltage supplies.\n<\/p><p>\nCapacitors, on the other hand, behave differently; in parallel, their capacitance increases - and in series, drops. This is easy to understand by\nconsidering that several identical capacitors in series are essentially equivalent to a single capacitor with a thicker layer of a dielectric (the\nfact that these is a piece of conductor in the middle is of no significance for its operation) -\nand therefore, a lower ability to cancel out the electric field generated by unbalanced charges deposited on the plates; consequently, several\ncapacitors in parallel resemble one capacitor with a larger plate surface area.\n<\/p><p>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Th%C3%A9venin%27s_theorem\">Thevenin's theorem<\/a> is a sometimes useful analytic tool, too: it gives a method\nfor replacing any network of resistors, voltage sources, and current sources, with a single voltage source and a single resistor. You do not have\nto memorize it, but it's good to be aware of this possibility.\n<\/p><h3>Resistors as current limiters <a name=\"22\" href=\"#22\" class=\"slink\" id=\"22\">[link]<\/a><\/h3>\nOhm's law states that when a particular voltage is applied across resistor terminals, only a current proportional to this voltage, and\ninversely proportional to resistance, will be allowed to flow: <code>I = V\/R<\/code>. This brings us to the simplest application of resistors:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/res_current.jpg\" alt=\"image\" \/><\/p><p>\nIn this circuit, a 9V battery is supposed to light up a red LED. From the datasheet, we\nknow that the LED, when supplied with 1.9V, sinks exactly 20 mA. Equipped with this knowledge, we can \ncalculate that at this exact point, its internal resistance is 95 \u2126. Unfortunately, diodes have a highly non-linear I-V \ncurve - so when we connect 9V, its resistance will suddenly drop, several amps will be allowed flow through the junction; and the whole thing\ncould - nay, will - catch fire.\n<\/p><p>\nThankfully, we can prevent this current from flowing in a very simple way. First, we need to find an equivalent resistor that, when placed across the \nterminals of our 9V battery, would conduct just 20 mA. Then, we can subtract 95 \u2126 - and get the resistance that needs to be placed in series with \nthe diode to achieve the same result:\n<\/p><p>\n<code>R<sub>equiv<\/sub> = 9V \/ 0.02A = 450 \u2126<\/code><br \/><code>R = R<sub>equiv<\/sub> - 95 \u2126 = 355 \u2126<\/code>\n<\/p><p>\nThe voltage drop across this resistor would be <code>0.02A * 355 \u2126 = 7.1V<\/code> - leaving exactly 1.9V for the diode.\n<\/p><p>\nTo build this circuit in practice, we can safely use use the nearest \"preferred\" resistor value, 330 \u2126 (which would actually allow \n21.5 mA to flow through the device - well within the safety limits of most components).\n<\/p><p>\nIn this particular use, the power dissipated by the resistor (<code>P = I\u00b2R<\/code>) will also be negligible - around 1\/8 watt; for higher\ncurrents, the amount of dissipated heat goes through the roof pretty quickly, though - and therefore, resistor-based current limiters are \nuseful only in low-power uses. This device is also not a perfect current limiter: if the apparent resistance of the driven device drops\nduring normal operation, a more significant current will be allowed to flow. In both these cases, a more sophisticated, active circuit -\nimplementing a \"real\" current source - is necessary instead.\n<\/p><p>\nIt is also important not to treat a current-limiting resistor as a way to reduce voltage: note that with the load (LED) disconnected, no current\nwill flow through the resistor - and therefore, there will be no voltage drop across its terminals (<code>V = IR<\/code>); the output voltage \nof the circuit will be identical to that of the supply - and will drop quickly, but thanks to parasitic capacitances not instantaneously, when the \nload is connected again.\n<\/p><h3>Resistors as voltage dividers <a name=\"23\" href=\"#23\" class=\"slink\" id=\"23\">[link]<\/a><\/h3>\nIn the following circuit, with the lightbulb disconnected, the current flowing through the circuit will obviously\nbe equal to <code>V<sub>supply<\/sub> \/ (R<sub>1<\/sub> + R<sub>2<\/sub>)<\/code>:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/res_voltage.jpg\" alt=\"image\" \/><\/p><p>\nFrom this, we can calculate the voltage measurable across the terminals of R2:\n<\/p><p>\n<code>V<sub>BA<\/sub> = IR<sub>2<\/sub> = V<sub>supply<\/sub> * (R<sub>2<\/sub> \/ (R<sub>1<\/sub> + R<sub>2<\/sub>))<\/code>\n<\/p><p>\nIf R1 is equal to R2, the voltage across B and A will be 1\/2 V<sub>supply<\/sub>; if R2 is twice the R1, the B-A voltage will be 2\/3 V<sub>supply<\/sub>; \nand so forth - it's a simple matter of the ratio of resistances, independent of the absolute values of the resistors (and the resulting current flow).\nThe circuit simply divides supply voltage by a desired factor - and in this, lies its utility.\n<\/p><p>\nThis divider is not a perfect voltage source, however: when you connect any resistive load (R3) between A and B - for example, the lightbulb shown on the\nschematic - it will introduce a new resistance parallel with R2. You may remember that this is equivalent to replacing R2 with a new resistor:\n<\/p><p>\n<code>R<sub>new<\/sub> = 1 \/ (1\/R<sub>2<\/sub> + 1\/R<sub>3<\/sub>)<\/code>\n<\/p><p>\nWhen the expected range of R3 values to be seen by the circuit is much higher than the value of R2, the effect is negligible - but when R2 and R3\nare in the same league, the resulting voltage drop may become significant. Therefore, the resistors need to be picked with the expected loads - and the\nacceptable voltage swings - in mind. In many cases, this is not a big deal; but when driving power-hungry devices, R1 and R2 may have to be so\nlow, that the resulting quiescent current through them would render the arrangement completely impractical.\n<\/p><p>\nIn other words, practical resistor-based voltage dividers can be considered medium to high impedance sources, and should not be driving low \nimpedance loads.\n<\/p><h3>Capacitors for energy storage <a name=\"24\" href=\"#24\" class=\"slink\" id=\"24\">[link]<\/a><\/h3>\nConsider the following circuit:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/cap_storage.jpg\" alt=\"image\" \/><\/p><p>\nWhen SW1 is toggled, the capacitor will gradually charge through a resistor. Without this resistor, the time it takes for it to charge would be \ntheoretically zero - and in practice, would be limited by poorly predictable factors such as the internal resistance of the device, or impedance of the\npower supply; but with a resistor, At <code>t = 5RC<\/code>,\nthe capacitor will be more than 99% charged, with almost no current flowing - and a voltage equivalent to that of the power source \nwill be present across its terminals.\n<\/p><p>\nWhat happens if we open SW1 and close SW2 at that point? A current will be allowed to flow through the lightbulb, discharging the capacitor\nin a familiar manner, with discharge time governed by capacitance and the equivalent resistance of the driven device. The lightbulb will\nbe getting more and more dim as the voltage across the capacitor drops; depending on its exact characteristics, somewhere around\n<code>t = 2R<sub>bulb<\/sub>C<\/code> (15% of supply voltage), it will probably go completely dark.\n<\/p><p>\nA more interesting case is what happens when both switches are switched on at the same time: assuming the non-linearity of the lightbulb is negligible\n(which is fair if the capacitor is large enough), initially, the capacitor will offer a low-impedance path, with very little current flowing through the\nbulb; but as the charge builds up, the voltage across its terminals will begin to rise - and more current will flow through the bulb. The capacitor will \nnever charge past the voltage across the terminals of the lightbulb (which can be calculated from the lightbulb's resistance, as per Ohm's law) - and will \ndischarge through it when SW1 is opened. In effect, this circuit gradually fades in and fades out the light when the switch is toggled.\n<\/p><p>\nThe action of a capacitor in parallel with a load appears to be reminiscent to the behavior of a series inductor. There is one key difference inductors\noppose changes in current for a longer time when the current is higher; whereas a capacitor has a more pronounced effect when the current that\ncharges or discharges it is lower; were we to eliminate the upstream resistance, the capacitor would charge instantaneously.\n<\/p><h3>Passive resistor-capacitor (RC) signal filters <a name=\"25\" href=\"#25\" class=\"slink\" id=\"25\">[link]<\/a><\/h3>\nA variety of resistor-capacitor-inductor arrangements can be used to form circuits with interesting frequency response characteristics. Perhaps the\nthe simplest case are the following two RC designs:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_low_high.jpg\" alt=\"image\" \/><\/p><p>\nIn the circuit shown on the left, the capacitor will be charged or discharged by any (sufficiently low-impedance) input signal at a rate controlled solely\nby the resistor; with the capacitor discharged, the output voltage will start at zero, and will begin approaching the input voltage only if the signal is \napplied for long enough (3RC or so). This has an interesting consequence: when the signal is a sine wave of a very low frequency, much lower than \nthe capacitor charge time - the output voltage will, with a slight lag (phase shift), simply follow the input. As the frequency increases, however, the \nsignal will get more and more attenuated - and the phase shift will become greater - as the source can't charge or discharge the \ncapacitor quickly enough. This circuit is, therefore, a simple, passive low-pass filter. \n<\/p><p>\nAn example of a linear response curve for this filter may look like this:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_low_lin.jpg\" alt=\"image\" \/><\/p><p>\nA logarithmic plot over a larger scale of frequencies is perhaps more informative, though:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_low_log.jpg\" alt=\"image\" \/><\/p><p>\nWhat about the second circuit shown on the earlier image? As it turns out, it exhibits exactly the opposite behavior: the capacitor prevents the flow of \nany DC currents - or the transfer of DC voltages - but when a negative voltage is applied to one of the plates, electrons will be pushed away from the \nother, and vice versa. If the\nresulting flow of charges is slow enough, they will simply drain through the resistor; but if the rate of change is very high, the resistor will\nlimit the current, and develop a substantial, momentary voltage difference across its terminals. This circuit passes higher frequency sine wave \nsignals largely unaltered, but attenuates slow-changing sine waves - a high-pass filter.\n<\/p><p>\nIt is customary to describe these circuits in terms of their cutoff frequency - a sine wave frequency at which the signal is attenuated to half its\npower (or, if you recall from the section on decibels, 0.7 of its input voltage). This point can be calculated as:\n<\/p><p>\n<code>f<sub>cutoff<\/sub> = 1 \/ (2 * \u03c0 * RC)<\/code>\n<\/p><p>\nFor any sine wave frequency f, amplitude multiplier (g) for the transferred signal is ideally:\n<\/p><p>\n<code>g<sub>lowpass<\/sub> = 1 \/ \u221a(1 + (f \/ f<sub>cutoff<\/sub>)\u00b2)<\/code><br \/><code>g<sub>highpass<\/sub> = (f \/ f<sub>cutoff<\/sub>) \/ \u221a(1 + (f \/ f<sub>cutoff<\/sub>)\u00b2)<\/code>\n<\/p><p>\nAs should be evident, there are many resistor-capacitor combinations that yield the same R*C product, and therefore, the same frequency response. The \ndifference is that the smaller the resistor, and the larger the capacitor, the lower output impedance the filter will have - but the more power it \nwill keep wasting, and the lower the impedance of the driving signal source would need to be to avoid distortion. The usual procedure, then,\nis to select R one or two orders of magnitude smaller than the impedance of the driven load - and calculate the right C for the required cutoff frequency.\n<\/p><p>\nNote that these circuits work as expected only for sine waves (and in such a case, will output a sine wave, too). When fed with other waveforms, including \nsquare, triangle, or sawtooth signals, they will behave differently - and introduce interesting\nartifacts of their own. For example, for square waves (shown in\ngreen), the output of a low-pass filter below its cutoff frequency (i.e., transmissive state) may be:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_low_distort1.jpg\" alt=\"image\" \/><\/p><p>\n...and well over the cutoff frequency, it may begin to resemble a sawtooth wave, with an amplitude decreasing in function of frequency:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_low_distort2.jpg\" alt=\"image\" \/><\/p><p>\nThe output for a low-pass filter is essentially proportional to how long the signal stayed in a particular state; for this reason, it is sometimes\ncalled an <a href=\"http:\/\/en.wikipedia.org\/wiki\/Integral\">integrator<\/a>.\n<\/p><p>\nThe behavior of a high-pass filter fed with a square wave is perhaps even more interesting - with a voltage proportional to the rate of change of the\ninput signal (hence the nickname: <a href=\"http:\/\/en.wikipedia.org\/wiki\/Differential_%28mathematics%29\">differentiator<\/a>). Above the cut-off \nfrequency (i.e., in its transmissive state), the output may look like this:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_high_distort1.jpg\" alt=\"image\" \/><\/p><p>\n...while well below the cut-off frequency, it may begin to resemble:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rc_high_distort2.jpg\" alt=\"image\" \/><\/p><p>\nNote that in both cases, the high-pass filter peak output amplitude is nearly twice the input amplitude; only the average power of the signal (RMS) \nis affected when the input frequency changes.\n<\/p><p>\nThe ability to spot these waveforms is actually extremely useful when debugging digital circuits, where square wave signals are used extensively.\nAn unexpected low-pass filter distortion seen in a digital signal is usually indicative of excessive capacitance of the signal path, perhaps\nbecause the connection is too long, or runs too close to others; while a high-pass pattern may indicate a broken trace or cut wire, forming an unwelcome \ncapacitor in series with the source of the signal.\n<\/p><p>\nLow-pass and high-pass filters can be cascaded to form band-pass or band-stop filters; identical filters can also be stacked\nto achieve a steeper response curve (n-th order filters, with g<sup>n<\/sup> frequency transmission function). That said, this process \nhas its limits: every stage needs to have an impedance much lower than the \nsubsequent stage it is driving, which quickly leads to impractical or very inefficient arrangements; and every stage attenuates the desired signal to\nsome extent, inevitably reducing signal-to-noise ratio.\n<\/p><h3>Inductors in signal filters (RL and RLC circuits) <a name=\"26\" href=\"#26\" class=\"slink\" id=\"26\">[link]<\/a><\/h3>\nInductors can replace capacitors in RC filters, reversing their operation: an RL filter with the same topology as an RC low-pass filter will act as a \nhigh-pass filter, and vice versa; the reason for this should be fairly evident. The cutoff frequency for RL filters is:\n<p>\n<code>f<sub>cutoff<\/sub> = R \/ (2 * \u03c0 * L)<\/code>\n<\/p><p>\nIn other aspects, the signal response of RL filters is essentially identical to that of their RC counterparts.\n<\/p><p>\nA more interesting class of circuits are RLC arrangements where the capacitor is \"fighting\" the inductor, and the resistor moderates the maximum\ncurrent supplied to them:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rlc_filters.jpg\" alt=\"image\" \/><\/p><p>\nThe circuit on the left is, essentially, a band-pass filter: the capacitor needs the signal to change slowly enough to charge it up to an appreciable\nlevel - and above this frequency, serves as a shunt; but when the current is not changing fast enough, the inductor will begin conducting and will\ndischarge the capacitor. This can be thought of as a combination of an RC low-pass and an RL high-pass filter, with a peak bandpass frequency of:\n<\/p><p>\n<code>f<sub>peak<\/sub> = 1 \/ (2 * \u03c0 * \u221a(LC))<\/code>\n<\/p><p>\n...and a bandwidth between 3dB attenuation points of:\n<\/p><p>\n<code>BW = R \/ (2 * \u03c0 * L)<\/code>\n<\/p><p>\nThe effect of adjusting bandwidth by changing the value of R is shown below:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rlc_pass_gentle.jpg\" alt=\"image\" \/><\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rlc_pass_sharp.jpg\" alt=\"image\" \/><\/p><p>\nIn practice, the optimal way to design an RLC band-pass filter is to pick R to achieve the desired output impedance; continue with L for the desired\nbandwidth; and finally, C to set the center frequency.\n<\/p><p>\nAs to the other circuit shown earlier, on the right - it is a band-stop filter: the coil conducts well only below a certain frequency, and the capacitor \nconducts well only above another; the signal will be shunted to the ground most efficiently in between the frequencies. The only difference is that\nthe blocked bandwidth is given as:\n<\/p><p>\n<code>BW = 1 \/ (2 * \u03c0 * RC)<\/code>\n<\/p><p>\nMany other RLC circuits can be designed, although these two approaches are most useful when easily understood dependence on source and load impedances\nis required. As with RC and RL filters, the gotcha with RLC circuits is that in signal processing, the impedance of the driven load must be \nsignificantly higher than that of the signal source and the series resistor in the circuit - or else, distortion will appear.\n<\/p><h3>Transistors as switches <a name=\"27\" href=\"#27\" class=\"slink\" id=\"27\">[link]<\/a><\/h3>\nTransistors differ from passive electronic circuits because of their ability to introduce gain - that is, turn low-amplitude voltage signals into\nhigh-amplitude ones, or high-impedance signals into low-impedance ones. For now, let's neglect their more complex response characteristics, \nand focus on a very simple, binary switching application - using a potentially low voltage or high impedance signal to control a power-hungry or \nhigh-voltage load:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/transistor_sw.jpg\" alt=\"image\" \/><\/p><p>\nThe circuits in the A column show common, proper transistor switch arrangements for NPN, PNP, and MOSFET n-channel enhancement mode transistors -\nknown as \"common emitter\" (BJT) or \"common drain\" (FET). In\nNPN and PNP circuits, note the use of a resistor to limit the base-emitter current:\nthe current flowing through this path <u>must be controlled<\/u>, because the corresponding junction is essentially a normal, forward-biased diode -\nand will conduct as much current as you supply, possibly destroying the transistor in the process (and certainly making it misbehave).\n<\/p><p>\nBipolar transistors have a device-specific current gain ratio, usually labeled as <code>h<sub>FE<\/sub><\/code> in the datasheets.\nRelying on this intrinsic gain is usually not a good idea for production-grade circuits, as \nthe parameter can change significantly at a whim - but by looking at its general order of magnitude, you can see what input impedance would be \nappropriate. For example, if h<sub>FE<\/sub> = 100, and the load requires a\ncurrent of 500 mA, at least 5 mA of base current should be supplied (supplying more will be wasteful and may lower switching speed, but the margins\nhere are pretty wide). Base-collector junction voltage is probably around 0.7V, so with a 5V supply, a voltage of 4.3V would be present across the \nresistor; from Ohm's law, and with an ample safety margin, a 680 \u2126 resistor would be a good bet.\n<\/p><p>\nBecause of the need to limit current, the NPN arrangement shown in column B is <u>universally<\/u> a bad idea: it will likely destroy the transistor;\nits PNP equivalent is equally disastrous.\n<\/p><p>\nThe MOSFET transistor shown in column A generally does not require a resistor, at least at low signal frequencies, because it does not allow any \nappreciably long-lived current \nto flow through the gate (at very high frequencies, the small but non-zero gate-source capacitance becomes a factor, though). It does, however, \nneed bipolar switching: if you simply apply a positive voltage to the gate, and then disconnect it - the gate-source \"capacitor\" will stay \ncharged, and the transistor will continue conducting for a longer while (dependent on humidity, handling, etc); even after this charge disappears,\na new one can be easily accumulated due to further handling, parasitic coupling, and so forth. This is why the MOSFET circuit shown in column B is\nalways a bad idea, too. It can be alternatively fixed by adding a large pull-down resistor from the gate to the ground, to dissipate the deposited charge\nwhen the switch is opened, but at the expense of lowering input impedance.\n<\/p><p>\nColumn C shows another arrangement that is not universally problematic, but should be avoided in switching where possible - and is all-too-common in\nhobbyist work: loading the emitter (BJT) or drain (MOSFET) - a configuration known as \"common collector\" or \"common source\". The problem with this is \nthat all real-world loads will develop some voltage across them in normal operation; this raises the emitter or source voltage accordingly - perhaps \nclose to, or even above, the driving base \/ gate voltage. This causes the switching to be far more erratic, especially with inductive loads. There\nare uses where this arrangement makes sense - but switching is not one of them.\n<\/p><p>\nIn all cases, the driving voltage applied to the base (or gate) must be high enough to trigger the transistor; that is, at least 0.6V in BJT, and at least\n1-2V for most MOSFETs. At this point, the transistor just barely begins conducting - so ideally, in switching applications, V<sub>in<\/sub> should\nbe comfortably higher. In BJTs, around 1V (\"base saturation voltage\") is sufficient to achieve a full range of meaningful base currents; in MOSFETs,\nresistance becomes negligible usually around 3.8V.\n<\/p><p>\nThe drawback of the transistor switches shown earlier is that they work in a manner similar to a single-pole switch: they can connect the load to \nthe ground (NPN, n-channel MOSFET) or to the supply rail (PNP, p-channel MOSFET), or simply leave it in open circuit state. When turning on a \nlightbulb or a LED, this is not a problem - but when driving inductive or capacitive loads (including the gates of MOSFET transistors), or encoding\ninformation as voltage levels, it may \nbe more desirable to offer two-pole operation, where the output can be switched between two low-impedance rails. The simplest way to do so is a\npush-pull circuit that uses complementary transistors to do the job:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/transistor_pushpull.jpg\" alt=\"image\" \/><\/p><p>\nWhen the input voltage is close to 0V, the bottom transistor will not conduct, but the upper one will. When the voltage is\nclose to the positive rail, the situation will reverse. When the voltage is somewhere in between, though, both transistors may end up conducting, shorting\nthe circuit - so caution must be exercised; this problem can be controlled by carefully biasing bases \/ gates using resistor-based voltage\ndividers, but it may affect switching performance.\n<\/p><p>\nAlso note that the upper transistor, connected to the positive rail, is PNP or p-type MOSFET; and the bottom one, grounded, is NPN or n-type MOSFET. \nThis ensures that both transistors are operating in common emitter \/ common source mode, suitable for switching. Reversing the order of these transistors\nmay lead to trouble for the reasons discussed above; the same goes for trying to use two identical transistors, just hooked up in opposite ways.\n<\/p><p>\nWhat else? Ah, inductive loads, such as motors, deserve some special attention: when a current supplied to an inductor is suddenly cut off, the energy stored\nin the collapsing magnetic field will try to keep the charge carriers in motion; and if this proves to be impossible, a very brief\n(nanoseconds) but significant (up to kilovolts) electromotive force will be created. This phenomenon is responsible for the sparking \ncommonly seen in low-voltage brushed motors or toggle switches.\n<\/p><p>\nIn some cases, this transient voltage may damage sensitive transistors in the vicinity - for example, compromising the gate insulator in MOSFETs,\nor simply overheating and vaporizing a part of the junction. To prevent this, a reverse-biased, normally non-conductive diode with a sufficient power rating \nis sometimes placed across such loads:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/transistor_motor.jpg\" alt=\"image\" \/><\/p><p>\nThe function of this diode is to begin conducting, due to the breakdown effect, when the terminal voltage gets dangerously high.\nTo be fair: in most cases, such a diode is not necessary - because switching is never perfectly instantaneous, inductors \nare never ideal, and the parasitic inductances, capacitances, and resistances of the circuit will dissipate the energy before it gets near a fragile \nsemiconductor - but these factors are not necessarily predictable, so especially with more significant inductances, it is simply a good engineering\npractice... after all, a suitable diode usually costs just several cents.\n<\/p><h3>Transistors as voltage followers and amplifiers <a name=\"28\" href=\"#28\" class=\"slink\" id=\"28\">[link]<\/a><\/h3>\nThe previous section discusses the use of transistors as binary switches operated in their saturation region - that is, the point where the resistance\nis minimal, and the admitted current is at its peak. Another important and more subtle use of these semiconductor devices is amplification, however \n- modulating output signals in relation to input voltage or current.\n<p>\nPerhaps the most rudimentary, but surprisingly not the most interesting application of a transistor as an amplifier\nis the previously discussed common emitter \/ common source setup:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/transistor_camp.jpg\" alt=\"image\" \/><\/p><p>\nThe current supplied to the load remains in some clear relation to the input current (BJT) or voltage (MOSFET), for as long as the input remains in the\n\"linear\" range for the device. Alas, the exact input-output ratio (h<sub>fe<\/sub>) is hard to predict - the parameter changes from one batch of \ntransistors to another, and is also highly influenced by factors such as voltage or temperature.\n<\/p><p>\nYou can use this arrangement to adjust the speed of a motor with a small potentiometer, or perform other non-critical tasks of this nature; but for anything\nthat requires precision and repeatability, this is just poor engineering: people should be able to substitute transistors used in your circuit with\ncomparable alternatives, or use a 5% accuracy input resistor, and still have it work.\n<\/p><p>\nSo, instead, let's have a look at a more useful, if modest, type of a simple transistor circuit: a voltage follower, also known as a buffer. \nIn this \"common collector\" \/ \"common drain\" device, output voltage simply follows the input voltage 1:1 (a setup known as \"unity gain\"); its utility is \nthat a high-impedance input can be \"converted\" in a low-impedance, high current capability output:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/v_followers.jpg\" alt=\"image\" \/><\/p><p>\nTo understand why the circuit shown in the first column works as advertised (and why no base resistor is needed), it is important to \nrecall that an NPN transistor will only admit any base-emitter current (and therefore, collector-emitter current) when the voltage difference between\nthe base and the emitter is sufficient to \novercome the potential of the internal p-n diode junction (circa 0.6V). In other words, the transistor will only try to conduct just enough current \n- bulk of it through the collector-emitter pathway - to keep \nthe voltage across R1 (and therefore, at V<sub>out<\/sub>) at exactly V<sub>in<\/sub> - 0.6.\n<\/p><p>\nOhm's law states that the current needed to develop a particular voltage across the resistor will be proportional to the desired voltage, and\ninversely proportional to resistance; if R1 is reasonable, so is the collector-emitter current. The base-emitter voltage is kept near the diode\nconduction threshold, so the current flowing through this path is also very low and self-limiting (and naturally, avoided entirely in MOSFETs).\n<\/p><p>\nThis circuit has some drawbacks, though. For starters, there is a voltage offset present between the input and the output; for inputs between 0 and 0.6V,\nthe output will be simply clipped at 0V. Connecting the emitter to a negative voltage at least 0.6V higher than the lowest signal voltage is a \npotential solution to the clipping problem - but it may be impractical in some settings. Another simple (but imperfect) way to prevent this problem begins \nwith AC coupling - the introduction of a series capacitor that prevents any steady DC currents or voltage levels from propagating through it, but allows \ntime-varying\nsignals to induce a voltage across it. Once this is done, a simple resistor-based voltage divider can be used to add an arbitrary bias to the resulting AC \nsignal, for example to add +3V. This circuit is shown in column B above. Its key flaw is evident once we recognize the capacitor and the resistors\nas a high-pass filter: this filter inevitably attenuates sufficiently low frequencies and DC voltage drifts. In many cases, this is not a big deal -\nthe capacitor and the resistors can be selected with the interesting range of frequencies in mind; but fundamentally, the follower is no longer\nmaintaining direct relationship between input and output voltages - and merely between their rates of change.\n<\/p><p>\nBias and clipping aside, another problem with simple voltage followers that they can only source, but not drain, significant currents. This is more of\na deal here with voltage followers than it is with switches: consider driving a capacitor as a load, connected across the \"out\" node and the ground.\nWhen V<sub>in<\/sub> is 6V, the capacitor will very quickly charge to 5.4V, as a significant current flows through the transistor. Alas, \nwhen V<sub>in<\/sub> later drops to 2V, V<sub>out<\/sub> will stay at 5.4V - and because emitter voltage is now higher than base voltage, the\ntransistor will not conduct. The capacitor will slowly discharge through R1 - but R1 can't be arbitrarily low (the quiescent current flowing at all\ntimes through the follower would be rather impractical if we go too far, and input loading would increase, too). Until then, the follower will not\nbe following the input voltage at all.\n<\/p><p>\nA simple way to fix it is to have the input signal drive two complementary transistors at the same time, placed on both sides of the \"out\" node;\nthis is known as a push-pull amplifier. This arrangement still suffers from the 0.6V bias - but this time, it can be solved more neatly than with\nAC coupling: by biasing both transistors into symmetrical, slight conduction when V<sub>in<\/sub> is at the mid-point - and simply relying on the input \nsignal to swing the ratio. This bias-compensated, DC-coupled voltage follower is shown in column C above; try to think for a moment about how it works \nin practice - and observe that the arrangement of transistors is opposite to that in a push-pull switch.\n<\/p><p>\nVoltage followers may sound boring, but they are immensely useful in electronic circuits; for example, they can be used to bridge a significant\nnumber of RLC band-pass or band-stop filters, or to convert very faint high-impedance currents from external sensors into outputs suitable for \nother circuitry. Still, you are probably wondering about a different task more commonly associated with\namplifiers: increasing the amplitude of a weak signal. When faced with this problem, almost all hobbyists immediately think of the intuitive, \nh<sub>fe<\/sub>-dependent circuit discussed initially in this section - but there is a better way:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/transistor_camp2.jpg\" alt=\"image\" \/><\/p><p>\nLet's see how this works: the transistor will seek to supply such a current to R2 to achieve a voltage matching V<sub>in<\/sub> - 0.6V across this resistor\n(assuming for a while that R1 permits so);\nas per Ohm's law, <code>I = (V<sub>in<\/sub> - 0.6) \/ R<sub>2<\/sub><\/code>. This flow of current, however, will also create a voltage\nacross R1: <code>V = IR<sub>1<\/sub> = (V<sub>in<\/sub> - 0.6) * R<sub>1<\/sub> \/ R<sub>2<\/sub><\/code>. In other words, V<sub>out<\/sub> will be amplified\nby <sub>R1\/R2<\/sub> - with no dependency on h<sub>fe<\/sub>! Note that in this particular circuit, the output signal is inverted (i.e., is at supply \nvoltage when V<sub>in<\/sub> is 0V); that problem is obviously easy to correct if necessary - for example by adding a second stage with R<sub>1<\/sub> = \nR<sub>2<\/sub>.\n<\/p><p>\nAn interesting extension of this simple concept is a differential amplifier - a device that compares two voltages, and outputs the difference between\nthem, multiplied by some factor:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/amplifier_diff.jpg\" alt=\"image\" \/><\/p><p>\nThis circuit essentially consists of two conjoined voltage amplifiers from the previous example. When their base voltages are identical, they agree on\nwhat the voltage of their emitters should be; each will contribute 50% of the current that needs to flow through R2 to produce that result.\nWhen V<sub>in2<\/sub> is higher than V<sub>in1<\/sub>, the right transistor will insist on getting the emitter voltage to a point where the left\none no longer conducts - and so, the current flowing through the right R1 (and the associated voltage drop) will increase.\n<\/p><p>\nAs to the point of it all: differential amplifiers are extremely useful for subtracting RF interference or removing or adding bias voltages from \ninput signals without the need to use AC coupling.\n<\/p><p>\nNaturally, a wide range of refinements to these basic designs is known today - say, to improve linearity, bandwidth, or to compensate for\ntemperature variations. You will also probably stumble upon many other, more exotic transistor circuits not covered here - such as, for example, \n<a href=\"http:\/\/www.ecircuitcenter.com\/Circuits_Audio_Amp\/Diff_Amp_w_Mirror\/image004.gif\">current mirrors<\/a>\nor <a href=\"http:\/\/en.wikipedia.org\/wiki\/Common_base\">common base<\/a>\n\/ <a href=\"http:\/\/en.wikipedia.org\/wiki\/Common_gate\">common gate<\/a> amplifiers - but you can probably figure them out by now.\n<\/p><h3>Diodes as shunts and regulators <a name=\"29\" href=\"#29\" class=\"slink\" id=\"29\">[link]<\/a><\/h3>\nThe tendency for diodes to maintain a precise voltage drop across their terminals is often exploited to provide precise voltage references in a\ncircuit. Forward-biased diodes, with their 0.2 - 0.7V voltage drops, are usually less useful; but reverse-biased Zener diodes are a wholly different story.\n<p>\nThe most unsophisticated use of a diode is a shunt; shunts may be placed across the terminals of high-impedance, low-precision signal sources to trim\nthe voltage to a particular level; or to suppress transient voltage spikes caused by inductors, electrostatic discharge, or even\nlightning strikes (using a specialized <a href=\"http:\/\/en.wikipedia.org\/wiki\/Varistor\">varistor<\/a> diode). The usual \"dumb\" shunt circuit is:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/diode_shunt.jpg\" alt=\"image\" \/><\/p><p>\nWhile very useful for controlling high-impedance signals, the diode simply serves as a \"crowbar\" across the supply terminals - and therefore, for\ninput voltage sources that can source a significant current, this arrangement gets dangerously inefficient; a resistor can be used to limit\nsupply current, of course - but this simply takes you back to the high-impedance scenario - not very useful for, say, driving motors.\n<\/p><p>\nA simple modification of this approach is to combine a voltage follower with a reverse-biased diode; the follower will then always attempt to output\na voltage equivalent to diode breakdown voltage, minus the inherent drop across the follower itself:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/diode_shunt2.jpg\" alt=\"image\" \/><\/p><p>\nThe resistor on diode side should be selected to limit the current through the diode to a reasonable value, so that it has an opportunity to develop\na voltage across its terminals in the first place, but so that it does not waste too much power. The spec for the diode usually gives a \n\"rated reverse current\" or I<sub>ZT<\/sub> figure, corresponding to the rated reverse voltage; in most cases, several milliamps is ideal.\n<\/p><p>\nThe previously discussed modifications of the voltage follower circuit - to allow adjustable voltage amplification and push-pull operation - are\njust as applicable to this regulator. The key advantage of this design is that unlike a resistor-based voltage divider, it does not waste\ntons of energy - and achieves much better regulation over a wide range of possible loads.\n<\/p><p>\nDiodes are also commonly used to build constant-current sources, such as \n<a href=\"http:\/\/upload.wikimedia.org\/wikipedia\/commons\/7\/74\/Const_cur_src_111.gif\">this circuit<\/a>: this arrangement\nwill admit only as much current as needed to create a particular voltage across the constant \"sense\" resistor, R2, regardless of the potentially\nvariable voltage drop seen across the connected load.\n<\/p><h3>Diodes as rectifiers, DC restorers, and voltage multipliers <a name=\"30\" href=\"#30\" class=\"slink\" id=\"30\">[link]<\/a><\/h3>\nForward-biased diodes serve a number of useful functions of their own. Let's begin with basic rectifiers - devices used to extract the absolute value\nof a voltage that swings both ways from the 0V rail:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/rectifiers.jpg\" alt=\"image\" \/><\/p><p>\nThe behavior of the first circuit - known as a half-wave rectifier - should be fairly clear: the diode conducts, and therefore creates a voltage across\nthe resistor (a dummy load), only if the first input is more positive than the other; in this circuit, diode breakdown voltage is selected high enough\nnot to interfere with this process. The second circuit - a bridge or full-wave rectifier -\nis a bit more clever, but also easy to follow: opposing pairs of diodes are used to select the\nmore positive or negative out of two input leads, and always produce a particular output polarity.\n<\/p><p>\nWhy is this useful? Well, the input AC signal, which (if symmetrical and steady) has an average DC voltage of 0V; if you connect a large capacitor\nto it, its average charge will also be 0V, and all the energy used for charging will be wasted.\nThe output waveform from a rectifier has non-zero average DC voltage, though - corresponding to the amplitude \nof the input signal (shown below in green in the graph below). This output can be then used to construct rudimentary approximations of\nDC voltages using just a single cap (red is the voltage across the load):\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/dc_filter.jpg\" alt=\"image\" \/><\/p><p>\nThis circuit, common in mains-powered supplies, should be fairly easy to understand: a low impedance AC signal is rectified, and then used to instantaneously\ncharge a capacitor; the capacitor will only be charged if its terminal voltage is below that of the rectified waveform, too - at all other times, the diodes\nare reverse-biased and do not conduct.\n<\/p><p>\nAs long as the rate at which this capacitor is discharged by a load is much lower than the rate at which it can be recharged, the output signal is a close \napproximation of a DC supply; the larger the capacitor, the smoother the DC output is.\n<\/p><p>\nA variation of this circuit is called an envelope follower: in this case, the resistor and the capacitor are matched to form a low-pass filter that \n\"masks out\" a high-frequency carrier, but passes through a lower frequency envelope (in this use, no significant output loading is present). Envelope \nfollowers are commonly used to separate modulated audio signals from high-frequency carrier signals used in radio transmission:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/envelope_follower.jpg\" alt=\"image\" \/><\/p><p>\nAnother interesting use of a diode is DC restoration - a process where a symmetrical AC signal is offset so that the flow of current becomes unidirectional\n- but that the waveform is not distorted otherwise:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/dc_restorer.jpg\" alt=\"image\" \/><\/p><p>\nThis arrangement may be a bit tricky to understand: the key point is that when a negative voltage -X is applied to a reasonably large capacitor by \na low-impedance source, the internal electric field will induce a similar voltage on its other terminal (i.e., any external electromotive force will \npropagate through the dielectric medium when no existing charge is deposited on its plates) - and this will prompt the diode to begin conducting, allowing\nthe current to flow, charging the capacitor to -X volts, and keeping the potential between output terminals close to 0V.\n<\/p><p>\nWhen the input voltage then swings to +X some time later, the resulting electromagnetic field propagates a 2X voltage swing across the capacitor; but\nthis time, the reverse-biased diode will not conduct, and will prevent the already deposited charge from exiting the capacitor (except through any\nexternally connected load). The subsequent -2X swing will bring the output voltage close to 0V, then back to 2X, and so on.\n<\/p><p>\nNaturally, to avoid surprises, the capacitor must be large enough not to form such a high-pass filter with the load (or\nthe always weakly conductive diode) as to substantially attenuate the input AC signal.\n<\/p><p>\nA really interesting combination of the previous two circuits is a diode-based voltage multiplier: the output of a DC restorer can be used to gradually charge\na capacitor to the peak output voltage, which is then used as a ground reference for another DC restorer. The behavior is illustrated below, with a 5V\namplitude AC signal as the input:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/diode_multiplier.jpg\" alt=\"image\" \/><\/p><p>\nVoila - we turned an AC signal with peak-to-peak voltage of 10V into a DC signal with peak voltage of 20V. The multiplier can be stacked, although diode \nand capacitor leakage currents and other losses put some constraints on its scalability.\n<\/p><h3>Transistor-based oscillators <a name=\"31\" href=\"#31\" class=\"slink\" id=\"31\">[link]<\/a><\/h3>\nResistors, inductors, capacitors, and piezoelectric crystals may be arranged into\nmany types of resonant circuits that will generate alternating voltages; one of the simplest\nexamples is just a capacitor and an inductor in a closed, series loop: a previously charged capacitor will discharge through the inductor; the resulting \nmagnetic field will then collapse and induce a current that re-charges the capacitor - and so forth (the resonant frequency is <code>1 \/ \n(2 * \u03c0 * \u221a(LC))<\/code>).\n<p>\nThe only problem with such circuits is that they lose energy and stop oscillating appreciably after, at best, several dozen swings: some of the \nmagnetic field will be inevitably dissipated, and so will some of the current due to real-world resistances and capacitor leakage. \nTransistors come to the rescue, however: by detecting a particular state of the oscillator and supplying additional energy - a well-timed\npush - the oscillation can be sustained indefinitely. Simple transistor LC circuits, such as\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Hartley_oscillator\">Hartley<\/a>, \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Armstrong_oscillator\">Armstrong<\/a>, or\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Clapp_oscillator\">Clapp<\/a> oscillators make use of this operating principle.\n<\/p><p>\nTo illustrate the basic operation of many types of oscillators, let's have a look at this simple,\nRC-only square wave oscillator (aka \"multivibrator\"):\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/multivibrator.jpg\" alt=\"image\" \/><\/p><p>\nThe circuit consists of two identical MOSFET switches driving small resistive loads equipped with an output tap; this tap will be at V<sub>supply<\/sub> \nwhen the MOSFET is off, and close to 0V when fully conducting. This output voltage is then AC coupled through a capacitor (C = 1 \u00b5F), biased \nto \u00bd V<sub>supply<\/sub> using a voltage divider consisting of two large (R = 100k) resistors - and finally employed to drive the gate of a\ncomplementary MOSFET.\n<\/p><p>\nWhen the circuit is initially turned on, the MOSFET on the right will turn on first, as its gate is driven through a slightly smaller resistor. This will\ncause it to pull the output down, and create a negative voltage across the capacitor driving the left transistor. This negative voltage will\nprevent that MOSFET from turning on until the charge is dissipated through the voltage divider (which takes time proportional to the capacitances\nand resistances involved) - and the gate is positive enough again.\n<\/p><p>\nAt this point, the left transistor will turn on, and create a negative voltage on the gate of the other MOSFET, turning it off for a while. The cycle\nwill repeat indefinitely, generating a square wave (the frequency of which is approximately <code>0.72 \/ RC<\/code> - or about 7 Hz in this particular case).\n<\/p><h3>Analog integrated circuits <a name=\"32\" href=\"#32\" class=\"slink\" id=\"32\">[link]<\/a><\/h3>\nIn addition to discrete components, a wide selection of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Integrated_circuit\">integrated circuits<\/a> is available today.\nThese highly miniaturized, modular building blocks consist from anywhere from several to hundreds of millions individual components packed on a single die,\nusually produced by the means of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Photolithography\">photolithography<\/a>. Some of the most popular varieties of\nanalog ICs are:\n<p>\n<b>Transistor drivers:<\/b> one of the most rudimentary classes of integrated circuits, these chips consists of a modest number of individual\ntransistors, often with protective resistors and diodes, to be used as line drivers (i.e., impedance converters). One of the most common devices \nin this category is <a href=\"http:\/\/focus.ti.com\/lit\/ds\/symlink\/uln2003a.pdf\">ULN2003A<\/a>, a set of six Darlington transistors; more complex devices\nwith push-pull control or low-distortion signal switching, also exist, and are just as easy to follow.\n<\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Operational_amplifier\">Operational amplifiers<\/a>:<\/b> advanced, differential push-pull voltage amplifiers with very\nhigh impedances (sometimes using JFET or MOSFET transistors), variable gain, temperature and supply voltage compensation, internal biasing,\nvoltage drop compensation, and so forth - essentially designed to approximate a perfect amplifier within a fairly wide range of operating conditions. \nOp-amps usually accept two inputs (non-inverting, marked as \"+\"; and \ninverting, marked as \"-\") and have one output. When the input voltages are identical, the output voltage is exactly 50% of the supply range; and when\nthe voltages differ, the difference will be amplified and output as a swing toward the negative or the positive supply rail.\nThe \"default\" amplification of an op-amp is very high and not necessarily well-defined - but can be controlled very precisely by adding a negative \nfeedback pathway between the output and the inverting input.\n<\/p><p>\nThe most common use of an op-amp is probably building a \"normal\" single-input amplifier; the amplification is then equal to the ratio of the \nfeedback resistor (R1) to the impedance of the other signal supplied the inverting input (R2, generally in the 10 - 100 k\u2126, depending on the \nop-amp). The two most common circuits here are:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/opamps.jpg\" alt=\"image\" \/><\/p><p>\nOther uses of op-amps include voltage followers (R1 = R2), voltage comparators (both inputs untied for maximum amplification),\nlinear drivers for power-hungry loads such as audio speakers; and in general, the approximation of the ideal behavior\nof a variety of other electronic circuits in amplifiers, oscillators, rectifiers, filters, and so forth;\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Operational_amplifier_applications\">Wikipedia<\/a> has a nice catalog.\n<\/p><p>\nMany traditional op-amps require the input voltages to keep a certain distance from the supply voltages, requiring careful biasing or a dual voltage\nsupply in certain uses. Many modern chips used in signal processing accept rail-to-rail inputs and have rail-to-rail outputs, however, allowing them to be \nmore easily operated from a single supply.\n<\/p><p>\nOne of the more popular chips suitable for single-supply operation\nare the <a href=\"http:\/\/www.national.com\/ds\/LM\/LM124.pdf\">LM124<\/a> family (including LM224, LM324, LM2902) and the\n<a href=\"http:\/\/www.national.com\/ds\/LM\/LM158.pdf\">LM158<\/a> family (LM258, LM358, LM2904) - all of which vary slightly, but are interchangeable\nin almost all uses. For extremely high input impedances,\n<a href=\"http:\/\/www.datasheetcatalog.org\/datasheet\/texasinstruments\/tlc27l4.pdf\">TLC27L4<\/a> is a reasonably priced family of MOSFET-based chips; and\nfor low-noise applications (e.g. audio processing),\n<a href=\"http:\/\/www.onsemi.com\/pub\/Collateral\/NE5534-D.PDF\">NE5534<\/a> is an affordable choice.\nOh, a relatively cheap high current op-amp is \n<a href=\"http:\/\/www.st.com\/stonline\/books\/pdf\/docs\/1317.pdf\">L165V<\/a>, capable of delivering over 3A.\n<\/p><p>\n<b><a href=\"http:\/\/en.wikipedia.org\/wiki\/Voltage_comparator\">Voltage comparators<\/a>:<\/b> \nspecialized op-amp like chips designed for very high amplification and fast recovery from saturation, used specifically to\ncompare voltages and output discrete voltage levels corresponding to the result. They are used in accurate sensing circuitry, in power management,\nin some types of oscillators, etc. Common examples include the <a href=\"http:\/\/www.national.com\/ds\/LM\/LM393.pdf\">LM193<\/a> family\n(LM293, LM393, LM2903); voltage threshold detectors (with internal, diode-based voltage references) are also available.\n<\/p><p>\n<b>Oscillators:<\/b> many types of ICs are used as timers (one-shot operation) or oscillators (continuous, astable operation); perhaps the most famous\nmember of this family is <a href=\"http:\/\/www.national.com\/ds\/LM\/LM555.pdf\">555<\/a>, a very versatile pulse and square \/ sawtooth wave generator that\ncan be controlled over a wide range of frequencies with an external capacitor and two resistors. The chip is fairly dated and quirky, but it\nhas a number of modern, power-efficient FET successors, such as <a href=\"http:\/\/www.intersil.com\/data\/fn\/fn2867.pdf\">ICM7555<\/a>. A good\noverview of possible 555 circuits can be found on <a href=\"http:\/\/www.kpsec.freeuk.com\/555timer.htm\">this webpage<\/a>.\n<\/p><p>\n<b>Rectifiers:<\/b> fairly self-explanatory: pre-packaged diodes in half-wave or full-wave arrangements.\n<\/p><p>\n<b>Regulators:<\/b> high current capability, transistor-based voltage or current limiters, often adjustable with an external resistor. We covered the\noperating principle of linear regulators earlier in this guide; some of the more advanced and power-efficient switched mode devices use more\nsophisticated circuitry to pulse inductors or capacitors - and sometimes, are even capable of generating voltages higher\nthan the input signal - although their operation is electrically more noisy.\n<\/p><p>\n<b>Sensor chips:<\/b> miniature optoelectronic or electromechanical devices are being increasingly more common; some of the popular examples\ninclude <a href=\"http:\/\/en.wikipedia.org\/wiki\/Opto-isolator\">optocouplers<\/a> (LED and a phototransistor in a single package, to relay optical signals\nwhile maintaining galvanic separation between circuits to minimize noise or ensure user safety), a variety of internally amplified sensors (such as solid-state\ncompasses based on <a href=\"http:\/\/en.wikipedia.org\/wiki\/Hall_effect\">Hall effect<\/a>), <a href=\"http:\/\/en.wikipedia.org\/wiki\/Accelerometer\">accelerometers<\/a>,\nsingle-chip <a href=\"http:\/\/en.wikipedia.org\/wiki\/Gyroscope\">gyroscopes<\/a>, and so on.\n<\/p><h2>Digital circuits <a name=\"33\" href=\"#33\" class=\"slink\" id=\"33\">[link]<\/a><\/h2>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Digital_electronics\">Digital electronics<\/a> are a class of easy to understand circuits that use discrete voltage\nranges and square waveforms to transmit and\nprocess data - most commonly, representing <a href=\"http:\/\/en.wikipedia.org\/wiki\/Binary_numeral_system\">binary numbers<\/a> for use in \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Boolean_algebra_%28logic%29\">Boolean algebra<\/a>: a signal close to 0V is meant to signify \"0\", and a signal close\nto V<sub>cc<\/sub> signifies \"1\". They are designed to allow one digital component to be interfaced to any other without having to carefully consider the\neffects related to impedance, attenuation, noise, and other \"analog\" phenomena (well, within certain common-sense boundaries) -\ngreatly simplifying large-scale circuit design.\n<p>\nThis section gives a quick overview of the most fundamental building blocks in digital signal processing and computing. While many types of highly \nintegrated digital circuits, such as <a href=\"http:\/\/en.wikipedia.org\/wiki\/Microcontroller\">microcontrollers<\/a>, are extremely complex, their \nfunctionality is generally implemented using these basic devices - and can be easily analyzed as such. In fact, a simple general-purpose computer can be \nassembled from scratch in a couple of weeks without any exceptional skill or resources at your disposal; it will be much slower and equipped with only \nrudimentary memory and I\/O - but will otherwise function perfectly.\n<\/p><h3>Logic gates <a name=\"34\" href=\"#34\" class=\"slink\" id=\"34\">[link]<\/a><\/h3>\nBoolean algebra is a fairly simple and intuitive binary number (aka logic value)\ncalculus that, by the virtue of its completeness, can be used to implement any other algebraic system, including \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Arithmetic\">decimal arithmetics<\/a>. The system is commonly explained in terms of the following\noperators (often illustrated using <a href=\"http:\/\/en.wikipedia.org\/wiki\/Truth_table\">truth tables<\/a>):\n<ul><li><b>Nullary<\/b>:\n<ul><li>Constant 0 output\n<\/li><li>Constant 1 output\n<\/li><\/ul><\/li><li><b>Unary<\/b>:\n<ul><li>Identity (buffer) - output identical to the input value\n<\/li><li>Negation (NOT) - output inverse to the input value\n<\/li><\/ul><\/li><li><b>Binary<\/b>:\n<ul><li>Conjunction (AND) - output is 1 only if both inputs are 1\n<\/li><li>Disjunction (OR) - output is 1 if any of the inputs is 1\n<\/li><li>Exclusive-or (XOR) - output is 1 only if input values are different\n<\/li><li>Sheffer stroke (NAND) - inverted-output AND\n<\/li><li>Pierce arrow (NOR) - inverted-output OR\n<\/li><li>Equivalence (EQV \/ XNOR) - inverted-output XOR\n<\/li><\/ul><\/li><\/ul>\nCircuit symbols of gates implementing these operators are:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/gate_symbols.jpg\" alt=\"image\" \/><\/p><p>\nOther derived, if less useful, binary operators are permitted (a total of 16); so are other arities - but all these variants can be constructed from\nthe basic set of operators outlined above. In fact, even this set is redundant - any operator can be constructed from NAND or NOR alone, for example:\n<\/p><p>\n<code>NOT(a) \u2192 NAND(a, a)<\/code><br \/><code>AND(a, b) \u2192 NOT(NAND(a, b))<\/code><br \/><code>OR(a, b) \u2192 NAND(NOT(a), NOT(b))<\/code><br \/><\/p><p>\nSeveral transistor-based approaches to implementing voltage-based Boolean algebra were pursued in the past; most notably, in the 80s and 90s,\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Transistor-transistor_logic\">transistor-to-transistor logic<\/a> (TTL), a scheme that employed exotic dual-emitter\nBJT transistors, dominated the marketplace - but today, simpler and more power-efficient \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/CMOS\">complementary MOSFET<\/a> (CMOS) gates are used almost exclusively in all modern chips. At their core,\nthese devices employ a simple, low-power MOSFET push-pull switch you may remember from earlier sections of this guide:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/mosfet_pushpull.jpg\" alt=\"image\" \/><\/p><p>\nThis switch, by itself, is a NOT gate: it pulls the output to V<sub>cc<\/sub> if input is 0V, and vice versa. To construct a NAND gate, this arrangement is \nextended the following way (note that Q1 is a four-terminal transistor, so that gate operates in reference to circuit ground, rather than to source\nterminal voltage):\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/cmos_nand.jpg\" alt=\"image\" \/><\/p><p>\nLet's see how this works: when input A and B are low, Q3 and Q4 will conduct, but Q1 and Q2 will not - and this will pull the output high. When both inputs \nare high, only Q1 and Q2 will conduct - and pull the output low. With only A high, Q1 will attempt to conduct, but this effort will be thwarted by\nQ2 in off state;\nin the meantime, Q3 will pull the output high. Finally, with only B input in high state, Q1 will prevent Q2 from doing anything useful, but Q4 will\nconduct - again, resulting in \"1\". This behavior coincides with the truth table for the NAND gate - amazing!\n<\/p><p>\nBecause gates are not expected to directly drive significant loads (several milliamps is usually the limit), CMOS transistors are selected so that their \nresistances and voltage response \ncharacteristics eliminate the risk of short-circuit damage when input voltage is somewhere between ground and V<sub>cc<\/sub>; that said, output voltage\nlevels and other gate characteristics are guaranteed to be sane only within certain specific ranges of \"0\" and \"1\" inputs. For example, for 5V CMOS\nchips, \"0\" should be under approximately 1.3V, and \"1\" should be over 3.7V - in which case, the output is guaranteed to stay under 0.2V for \n\"0\" and over 4.7V for \"1\", respectively. As should be evident, low output impedance gates with such output levels can be chained indefinitely\nwith no signal deterioration - which is the key premise of all digital circuitry.\n<\/p><p>\nNow, older TTL chips may expect under 0.8V and over 2.0V as inputs, and promise output under 0.35V and over 3.3V, respectively; this has some consequences\nwhen attempting to interface TTL and CMOS circuits in a reliable way - sometimes requiring pull-up or pull-down resistors on the input lines; but that's\na separate story.\n<\/p><p>\nOne of the most popular families of individual CMOS logic gates available today is\nundoubtedly <a href=\"http:\/\/www.mouser.com\/Semiconductors\/Integrated-Circuits-ICs\/Logic-ICs\/_\/N-6j77xZscv7?P=1z0y3hmZ1z0z63x\">74HC00<\/a>;\nfollowed by now somewhat less common\n<a href=\"http:\/\/www.mouser.com\/Semiconductors\/Integrated-Circuits-ICs\/Logic-ICs\/_\/N-6j77xZscv7?P=1yzxt7kZ1z0z63x\">4000 series<\/a> chips.\n<\/p><h3>Adder circuits <a name=\"35\" href=\"#35\" class=\"slink\" id=\"35\">[link]<\/a><\/h3>\nTo understand the power of Boolean algebra, it is useful to take a stab at building a simple adder - a device that computes a sum of two integers \nrepresented as binary numbers, and one of the fundamental components of <a href=\"http:\/\/en.wikipedia.org\/wiki\/Arithmetic_logic_unit\">arithmetic\nlogic units<\/a> in all computers.\n<p>\nBinary addition works in a fairly straightforward way, analogous to what we are accustomed to in the decimal system. When adding individual digits of\ntwo numbers, right to left, the following rules apply:\n<\/p><p>\n<code>0 + 0 = 0<\/code><br \/><code>0 + 1 = 1<\/code><br \/><code>1 + 0 = 1<\/code><br \/><code>1 + 1 = 0, 1 carries over<\/code>\n<\/p><p>\nIt is easy to notice that the immediate result is equivalent to XOR of the input values; while the carry bit is determined by the AND operation.\nA simple circuit that implements this logic - called \"half adder\" - is shown below:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/adders.jpg\" alt=\"image\" \/><\/p><p>\nThe other circuit - a full adder - can additionally accept a third input, the carry-over value from the previous column; this value is fed into\nan second adder stage after the immediate inputs are summed; additionally, if any of the stages contributes a carry bit, this value is propagated to\nthe output:\n<\/p><p>\n<code>0 + 0 + carry 0 = 0 + 0 = 0<\/code><br \/><code>0 + 1 + carry 0 = 1 + 0 = 0<\/code><br \/><code>1 + 0 + carry 0 = 1 + 0 = 0<\/code><br \/><code>1 + 1 + carry 0 = 0 + 0 = 0, 1 carries over<\/code><\/p><p>\n<code>0 + 0 + carry 1 = 0 + 1 = 1<\/code><br \/><code>0 + 1 + carry 1 = 1 + 1 = 0, 1 carries over<\/code><br \/><code>1 + 0 + carry 1 = 1 + 1 = 0, 1 carries over<\/code><br \/><code>1 + 1 + carry 1 = 0 + 1 = 1, 1 carries over<\/code>\n<\/p><p>\nThis full adder circuit can be stacked to implement full addition of arbitrarily large binary integers; for 4 bits, this may look the\nfollowing way:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/4bit_adder.jpg\" alt=\"image\" \/><\/p><p>\nOther basic arithmetic operations, including incremental counting (a simplified case of always adding +1 to the operand), \ncan be implemented in similar ways - enabling you to build a rudimentary calculator.\n<\/p><h3>Memory <a name=\"36\" href=\"#36\" class=\"slink\" id=\"36\">[link]<\/a><\/h3>\nThe essential difference between computers and dumb calculators is the ability to execute programs in a manner first generalized by \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Universal_Turing_machine\">Alan Turing<\/a>. To accomplish this goal, the machine needs not only the ability to\nperform calculations, but also to select the type of calculation to be made (i.e., enable the appropriate ALU block) in accordance to a stored program\n- and to save the output of this calculation for later reuse.\n<p>\nSeveral types of digital memory are used today, but the most rudimentary variety is an arrangement of NOR or NAND gates with internal feedback, known\nas a flip-flop or a <a href=\"http:\/\/en.wikipedia.org\/wiki\/Latch_%28electronics%29\">latch<\/a>:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/memory.jpg\" alt=\"image\" \/><\/p><p>\nLet's consider the behavior of the circuit on the left (known as an SR latch):\n<\/p><p>\n<code>SET=0 RESET=0 OUT<sub>t-1<\/sub>=0 \u2192 bottom gate outputs 1, top gate outputs 0, OUT<sub>t<\/sub>=0 (stable)<\/code><br \/><code>SET=0 RESET=0 OUT<sub>t-1<\/sub>=1 \u2192 bottom gate outputs 0, top gate outputs 1, OUT<sub>t<\/sub>=1 (stable)<\/code><\/p><p>\n<code>SET=0 RESET=1 OUT<sub>t-1<\/sub>=0 \u2192 bottom gate outputs 1, top gate outputs 1, OUT<sub>t<\/sub>=1 (evolves into the next case)<\/code><br \/><code>SET=0 RESET=1 OUT<sub>t-1<\/sub>=1 \u2192 bottom gate outputs 0, top gate outputs 0, OUT<sub>t<\/sub>=0 (stable)<\/code><\/p><p>\n<code>SET=1 RESET=0 OUT<sub>t-1<\/sub>=0 \u2192 bottom gate outputs 1, top gate outputs 1, OUT<sub>t<\/sub>=1 (stable)<\/code><br \/><code>SET=1 RESET=0 OUT<sub>t-1<\/sub>=1 \u2192 bottom gate outputs 0, top gate outputs 0, OUT<sub>t<\/sub>=0 (evolves into the previous case)<\/code><\/p><p>\n<code>SET=1 RESET=1 OUT<sub>t-1<\/sub>=X \u2192 unstable oscillation<\/code>\n<\/p><p>\nThe important observation is that when SET and RESET are both low, this arrangement of gates retains the previous output value indefinitely; when\nRESET is high and SET low, it eventually settles on \"0\" output regardless of the starting state; and that with SET high and RESET low, it always\nsettles on \"1\". This circuit can store one bit of data.\n<\/p><p>\nIn practice, to minimize the number of inputs and avoid unstable states, the arrangement shown on the right - known as D or gated latch - may be used, \ninstead. In this case, any input value is loaded into the flip-flop when ENABLE is high; and is retained when ENABLE is low. The ENABLE line can\nbe shared across any number of simultaneously accessed flip-flops.\n<\/p><h3>Line decoders, multiplexers, demultiplexers <a name=\"37\" href=\"#37\" class=\"slink\" id=\"37\">[link]<\/a><\/h3>\nIn some cases, it is necessary to convert binary numbers to \"one out of X\" signals; for example, to efficiently address 4 bit memory, \na 2-bit binary integer provided by the CPU may be used to decide which flip-flop needs to receive the ENABLE signal (with locations identified\nas 00, 01, 10, 11). In these cases, a circuit known as a line decoder is commonly used:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/decoder.jpg\" alt=\"image\" \/><\/p><p>\nA slight variation of this circuit is a <a href=\"http:\/\/en.wikipedia.org\/wiki\/Demultiplexer#Digital_demultiplexers\">demultiplexer<\/a> - a device \nthat takes an arbitrary digital input value, and routes it to a specified output pin:\n<\/p><p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/demux.jpg\" alt=\"image\" \/><\/p><p>\nDemultiplexers are particularly useful in communications, where - when controlled by a synchronous binary counter - they can be used to convert\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Serial_communications\">serial data streams<\/a> to <a href=\"http:\/\/en.wikipedia.org\/wiki\/Parallel_communication\">parallel\noutput<\/a> stored in memory.\n<\/p><p>\nIts counterpart, used to serialize parallel data, or to switch between multiple input signals, is known as \na <a href=\"http:\/\/en.wikipedia.org\/wiki\/Multiplexer\">multiplexer<\/a>; the circuit is very similar, except for multiple independent data inputs, and\none multi-input OR gate providing the output (<a href=\"http:\/\/en.wikipedia.org\/wiki\/File:Multiplexer_Example01.svg\">schematic<\/a>).\n<\/p><h3>Digital-to-analog and analog-to-digital conversions <a name=\"38\" href=\"#38\" class=\"slink\" id=\"38\">[link]<\/a><\/h3>\nDigital circuits can, with minimal interfacing, accept most types of binary inputs - and with the use of external power transistors or relays, can\nalso switch arbitrarily large loads. It is more complicated, however, to allow them to generate or sample continuously variable, analog signals.\nA specialized type of a circuit, <a href=\"http:\/\/en.wikipedia.org\/wiki\/Digital-to-analog_converter\">digital-to-analog converter<\/a> (DAC) is used to\naccomplish the former task. The most rudimentary type of a DAC is simply a resistor bridge:\n<p>\n<img src=\"http:\/\/lcamtuf.coredump.cx\/electronics\/dac.jpg\" alt=\"image\" \/><\/p><p>\nBy selecting every resistor to have a different value, it is possible to achieve a distinct output voltage level for any possible input combination\n- allowing simple conversion of binary numbers to nearly continuous, if slightly quantized (16 values in this case) analog signals.\n<\/p><p>\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Analog_to_digital_converter\">Analog-to-digital converters<\/a> (ADC) may be built from such a DAC and a voltage\ncomparator: the circuit may attempt to compare different generated voltage levels to the reference signal, and use the digital input that caused \nthe difference to fall within a threshold as the result of the conversion.\n<\/p><h3>Clocks in digital circuits <a name=\"39\" href=\"#39\" class=\"slink\" id=\"39\">[link]<\/a><\/h3>\nAll types of digital circuits require some time to respond to changes in input conditions and reach a stable state - a problem especially evident in\nthe case of flip-flops, where a single oscillation will occur for some state transitions before the gates settle in a stable configuration. In addition,\nmany digital operations, such as multiplexing, demultiplexing, or analog-to-digital conversions, or iterative arithmetics (e.g., naive multiplication) \nrequire synchronous timing signals to properly coordinate multi-step tasks, enabling and disabling subsequent processing stages as necessary.\n<p>\nBecause of this, almost all modern computers use global clock signals to ensure that all operations are properly synchronized, and all circuits\nare allowed enough time to reach the expected output state. For example, AND gates may be employed to make sure that the input or output stage is active only\nwhen the clock signal is high; <a href=\"http:\/\/en.wikipedia.org\/wiki\/File:Edge_triggered_D_flip-flop.png\">modified flip-flops<\/a>\nmay be used to detect clock signal state transitions and initiate actions on rising or falling edges; \nand binary counters may be used to ensure that a particular number of clock cycles passes between any two operations, or to select the appropriate\npart of a circuit to use when executing a sequence of calculations.\n<\/p><p>\nClock signals are sometimes divided (with the use of simple synchronous counters) to drive slower output buses while allowing the core processor to run at\nhigh speed - anywhere from 100 kHz to several gigahertz is common.\n<\/p><p>\nWhile experimental computers not using clocks at all, or using individual non-synchronized clocks for various subsystems, were devised in the past,\nthe extra complexity of doing so made them, at least so far, rather impractical in real-world applications. The same fate met all attempts to build\nnon-binary computers (e.g., <a href=\"http:\/\/en.wikipedia.org\/wiki\/Ternary_computer\">ternary<\/a>).\n<\/p>\nWell, that's all I have for you! A couple extra pointers:\n<ul><li> If you want to learn more about electronics, you should probably consider\n<a href=\"http:\/\/www.amazon.com\/Art-Electronics-Paul-Horowitz\/dp\/0521370957\">\"The Art of Electronics\"<\/a> by Horowitz and Hill.\nA more accessible, but <a href=\"http:\/\/www.eg.bucknell.edu\/physics\/ph235\/errata.pdf\">somewhat less accurate<\/a> book,\n<a href=\"http:\/\/www.amazon.com\/gp\/product\/0071452818\">\"Practical Electronics for Inventors\"<\/a> by Scherz, is probably also a good pick.\n<\/li><li> To experiment with circuit behavior, you should install <a href=\"http:\/\/www.linear.com\/designtools\/software\/ltspice.jsp\">LTspice<\/a> or other \n<a href=\"http:\/\/en.wikipedia.org\/wiki\/SPICE\">SPICE<\/a> UI.<\/li><li> To order parts, you should check out <a href=\"http:\/\/www.mouser.com\/\">Mouser<\/a>.<\/li><li> For prototyping, a <a href=\"http:\/\/www.allelectronics.com\/make-a-store\/item\/PB-1680\/BREADBOARD-1680-CONTACTS\/\/1.html\">simple breadboard<\/a>\nwill come handy.<\/li><li> For soldering, you should get a decent soldering station, such as \n<a href=\"http:\/\/www.amazon.com\/Weller-WES51-Analog-Soldering-Station\/dp\/B000BRC2XU\/\">Weller WES51<\/a>.<\/li><li> For troubleshooting issues, an oscilloscope is indispensable, but can be expensive; some low-cost options\ninclude \n<a href=\"http:\/\/www.rigolna.com\/products\/digital-oscilloscopes\/ds1000e\/\">Rigol DS1000E<\/a> and \n<a href=\"http:\/\/www.testequipmentdepot.com\/bk-precision\/oscilloscopes\/2530.htm\">BK Precision 2530<\/a>. In\nabsence of this, you will need at least a good <a href=\"http:\/\/www.amazon.com\/Extech-EX330-Autoranging-Multimeter-Thermometer\/dp\/B000EX0AE4\">multimeter<\/a>,\nbut you will be working in the dark.\n<\/li><\/ul><h2>Credits and contact info <a name=\"41\" href=\"#41\" class=\"slink\" id=\"41\">[link]<\/a><\/h2>\nMany thanks to Damien Miller, Evan Zalys-Geller, James Kehl,\nand other fine folks who took their time to review and comment on this doc.\n<p>\nSuggestions, corrections, and other feedback welcome - you can contact the author at <a href=\"mailto:lcamtuf@coredump.cx\">lcamtuf@coredump.cx<\/a>.\n<\/p><p>\n<span>Your lucky number: 18908978\n<\/span><\/p>","created_at":"2018-11-07T16:48:44+0000","updated_at":"2018-11-07T16:48:44+0000","annotations":[],"mimetype":"text\/html","reading_time":97,"domain_name":"lcamtuf.coredump.cx","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":453,"title":"How to Get Startup Ideas","url":"http:\/\/paulgraham.com\/startupideas.html","content":"<img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2238_514\" width=\"206\" height=\"18\" border=\"0\" alt=\"How to Get Startup Ideas\" style=\"margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><p><span><img src=\"http:\/\/www.virtumundo.com\/images\/spacer.gif\" height=\"15\" width=\"1\" alt=\"image\" \/><span>\n<b>Want to start a startup?<\/b>  Get funded by\n<a href=\"http:\/\/ycombinator.com\/apply.html\">Y Combinator<\/a>.\n<\/span>\n<br \/><img src=\"http:\/\/www.virtumundo.com\/images\/spacer.gif\" height=\"5\" width=\"1\" alt=\"image\" \/>\nNovember 2012<\/span><\/p><p>The way to get startup ideas is not to try to think of startup\nideas.  It's to look for problems, preferably problems you have\nyourself.<\/p><p>The very best startup ideas tend to have three things in common:\nthey're something the founders themselves want, that they themselves\ncan build, and that few others realize are worth doing.  Microsoft,\nApple, Yahoo, Google, and Facebook all began this way.<\/p><p><b>Problems<\/b><\/p><p>Why is it so important to work on a problem you have?  Among other\nthings, it ensures the problem really exists.  It sounds obvious\nto say you should only work on problems that exist.  And yet by far\nthe most common mistake startups make is to solve problems no one\nhas.<\/p><p>I made it myself.  In 1995 I started a company to put art galleries\nonline.  But galleries didn't want to be online.  It's not how the\nart business works.  So why did I spend 6 months working on this\nstupid idea?  Because I didn't pay attention to users.  I invented\na model of the world that didn't correspond to reality, and worked\nfrom that.  I didn't notice my model was wrong until I tried\nto convince users to pay for what we'd built.  Even then I took\nembarrassingly long to catch on.  I was attached to my model of the\nworld, and I'd spent a lot of time on the software.  They had to\nwant it!<\/p><p>Why do so many founders build things no one wants?  Because they\nbegin by trying to think of startup ideas.  That m.o. is doubly\ndangerous: it doesn't merely yield few good ideas; it yields bad\nideas that sound plausible enough to fool you into working on them.<\/p><p>At YC we call these \"made-up\" or \"sitcom\" startup ideas.  Imagine\none of the characters on a TV show was starting a startup.  The\nwriters would have to invent something for it to do.  But coming\nup with good startup ideas is hard.  It's not something you can do\nfor the asking.  So (unless they got amazingly lucky) the writers\nwould come up with an idea that sounded plausible, but was actually\nbad.<\/p><p>For example, a social network for pet owners.  It doesn't sound\nobviously mistaken.  Millions of people have pets.  Often they care\na lot about their pets and spend a lot of money on them.  Surely\nmany of these people would like a site where they could talk to\nother pet owners.  Not all of them perhaps, but if just 2 or 3\npercent were regular visitors, you could have millions of users.\nYou could serve them targeted offers, and maybe charge for premium\nfeatures. \n<span>[<a href=\"#f1n\"><span>1<\/span><\/a>]<\/span><\/p><p>The danger of an idea like this is that when you run it by your\nfriends with pets, they don't say \"I would <i>never<\/i> use this.\" They\nsay \"Yeah, maybe I could see using something like that.\" Even when\nthe startup launches, it will sound plausible to a lot of people.\nThey don't want to use it themselves, at least not right now, but\nthey could imagine other people wanting it.  Sum that reaction\nacross the entire population, and you have zero users. \n<span>[<a href=\"#f2n\"><span>2<\/span><\/a>]<\/span><\/p><p><b>Well<\/b><\/p><p>When a startup launches, there have to be at least some users who\nreally need what they're making \u2014 not just people who could see\nthemselves using it one day, but who want it urgently.  Usually\nthis initial group of users is small, for the simple reason that\nif there were something that large numbers of people urgently needed\nand that could be built with the amount of effort a startup usually\nputs into a version one, it would probably already exist.  Which\nmeans you have to compromise on one dimension: you can either build\nsomething a large number of people want a small amount, or something\na small number of people want a large amount.  Choose the latter.\nNot all ideas of that type are good startup ideas, but nearly all\ngood startup ideas are of that type.<\/p><p>Imagine a graph whose x axis represents all the people who might\nwant what you're making and whose y axis represents how much they\nwant it.  If you invert the scale on the y axis, you can envision\ncompanies as holes.  Google is an immense crater: hundreds of\nmillions of people use it, and they need it a lot.  A startup just\nstarting out can't expect to excavate that much volume.  So you\nhave two choices about the shape of hole you start with.  You can\neither dig a hole that's broad but shallow, or one that's narrow\nand deep, like a well.<\/p><p>Made-up startup ideas are usually of the first type.  Lots of people\nare mildly interested in a social network for pet owners.<\/p><p>Nearly all good startup ideas are of the second type.  Microsoft\nwas a well when they made Altair Basic.  There were only a couple\nthousand Altair owners, but without this software they were programming\nin machine language.  Thirty years later Facebook had the same\nshape.  Their first site was exclusively for Harvard students, of\nwhich there are only a few thousand, but those few thousand users\nwanted it a lot.<\/p><p>When you have an idea for a startup, ask yourself: who wants this\nright now?  Who wants this so much that they'll use it even when\nit's a crappy version one made by a two-person startup they've never\nheard of?  If you can't answer that, the idea is probably bad. \n<span>[<a href=\"#f3n\"><span>3<\/span><\/a>]<\/span><\/p><p>You don't need the narrowness of the well per se.  It's depth you\nneed; you get narrowness as a byproduct of optimizing for depth\n(and speed).  But you almost always do get it.  In practice the\nlink between depth and narrowness is so strong that it's a good\nsign when you know that an idea will appeal strongly to a specific\ngroup or type of user.<\/p><p>But while demand shaped like a well is almost a necessary condition\nfor a good startup idea, it's not a sufficient one.  If Mark\nZuckerberg had built something that could only ever have appealed\nto Harvard students, it would not have been a good startup idea.\nFacebook was a good idea because it started with a small market\nthere was a fast path out of.  Colleges are similar enough that if\nyou build a facebook that works at Harvard, it will work at any\ncollege. So you spread rapidly through all the colleges.  Once you\nhave all the college students, you get everyone else simply by\nletting them in.<\/p><p>Similarly for Microsoft: Basic for the Altair; Basic for other\nmachines; other languages besides Basic; operating systems;\napplications; IPO.<\/p><p><b>Self<\/b><\/p><p>How do you tell whether there's a path out of an idea?  How do you\ntell whether something is the germ of a giant company, or just a\nniche product?  Often you can't. The founders of Airbnb didn't\nrealize at first how big a market they were tapping.  Initially\nthey had a much narrower idea.  They were going to let hosts rent\nout space on their floors during conventions.  They didn't foresee\nthe expansion of this idea; it forced itself upon them gradually.\nAll they knew at first is that they were onto something.  That's\nprobably as much as Bill Gates or Mark Zuckerberg knew at first.<\/p><p>Occasionally it's obvious from the beginning when there's a path\nout of the initial niche.  And sometimes I can see a path that's\nnot immediately obvious; that's one of our specialties at YC.  But\nthere are limits to how well this can be done, no matter how much\nexperience you have.  The most important thing to understand about\npaths out of the initial idea is the meta-fact that these are hard\nto see.<\/p><p>So if you can't predict whether there's a path out of an idea, how\ndo you choose between ideas?  The truth is disappointing but\ninteresting: if you're the right sort of person, you have the right\nsort of hunches.  If you're at the leading edge of a field that's\nchanging fast, when you have a hunch that something is worth doing,\nyou're more likely to be right.<\/p><p>In <i>Zen and the Art of Motorcycle Maintenance<\/i>, Robert Pirsig says:\n<\/p><blockquote><div>\n  You want to know how to paint a perfect painting? It's easy.  Make\n  yourself perfect and then just paint naturally.\n<\/div><\/blockquote>\nI've wondered about that passage since I read it in high school.\nI'm not sure how useful his advice is for painting specifically,\nbut it fits this situation well.  Empirically, the way to have good\nstartup ideas is to become the sort of person who has them.<p>Being at the leading edge of a field doesn't mean you have to be\none of the people pushing it forward.  You can also be at the leading\nedge as a user.  It was not so much because he was a programmer\nthat Facebook seemed a good idea to Mark Zuckerberg as because he\nused computers so much.  If you'd asked most 40 year olds in 2004\nwhether they'd like to publish their lives semi-publicly on the\nInternet, they'd have been horrified at the idea.  But Mark already\nlived online; to him it seemed natural.<\/p><p>Paul Buchheit says that people at the leading edge of a rapidly\nchanging field \"live in the future.\"  Combine that with Pirsig and\nyou get:\n<\/p><blockquote><div>\n  Live in the future, then build what's missing.\n<\/div><\/blockquote>\nThat describes the way many if not most of the biggest startups got\nstarted.  Neither Apple nor Yahoo nor Google nor Facebook were even\nsupposed to be companies at first.  They grew out of things their\nfounders built because there seemed a gap in the world.<p>If you look at the way successful founders have had their ideas,\nit's generally the result of some external stimulus hitting a\nprepared mind.  Bill Gates and Paul Allen hear about the Altair and\nthink \"I bet we could write a Basic interpreter for it.\" Drew Houston\nrealizes he's forgotten his USB stick and thinks \"I really need to\nmake my files live online.\" Lots of people heard about the Altair.\nLots forgot USB sticks.  The reason those stimuli caused those\nfounders to start companies was that their experiences had prepared\nthem to notice the opportunities they represented.<\/p><p>The verb you want to be using with respect to startup ideas is not\n\"think up\" but \"notice.\" At YC we call ideas that grow naturally\nout of the founders' own experiences \"organic\" startup ideas.  The\nmost successful startups almost all begin this way.<\/p><p>That may not have been what you wanted to hear.  You may have\nexpected recipes for coming up with startup ideas, and instead I'm\ntelling you that the key is to have a mind that's prepared in the\nright way.  But disappointing though it may be, this is the truth.\nAnd it is a recipe of a sort, just one that in the worst case takes\na year rather than a weekend.<\/p><p>If you're not at the leading edge of some rapidly changing field,\nyou can get to one.  For example, anyone reasonably smart can\nprobably get to an edge of programming (e.g. building mobile apps)\nin a year.  Since a successful startup will consume at least 3-5\nyears of your life, a year's preparation would be a reasonable\ninvestment.  Especially if you're also looking for a cofounder.\n<span>[<a href=\"#f4n\"><span>4<\/span><\/a>]<\/span><\/p><p>You don't have to learn programming to be at the leading edge of a\ndomain that's changing fast.  Other domains change fast.  But while\nlearning to hack is not necessary, it is for the forseeable future\nsufficient. As Marc Andreessen put it, software is eating the world,\nand this trend has decades left to run.<\/p><p>Knowing how to hack also means that when you have ideas, you'll be\nable to implement them.  That's not absolutely necessary (Jeff Bezos\ncouldn't) but it's an advantage.  It's a big advantage, when you're\nconsidering an idea like putting a college facebook online, if\ninstead of merely thinking \"That's an interesting idea,\" you can\nthink instead \"That's an interesting idea.  I'll try building an\ninitial version tonight.\"  It's even better when you're both a\nprogrammer and the target user, because then the cycle of generating\nnew versions and testing them on users can happen inside one head.<\/p><p><b>Noticing<\/b><\/p><p>Once you're living in the future in some respect, the way to notice\nstartup ideas is to look for things that seem to be missing.  If\nyou're really at the leading edge of a rapidly changing field, there\nwill be things that are obviously missing.  What won't be obvious\nis that they're startup ideas.  So if you want to find startup\nideas, don't merely turn on the filter \"What's missing?\" Also turn\noff every other filter, particularly \"Could this be a big company?\"\nThere's plenty of time to apply that test later.  But if you're\nthinking about that initially, it may not only filter out lots\nof good ideas, but also cause you to focus on bad ones.<\/p><p>Most things that are missing will take some time to see.  You almost\nhave to trick yourself into seeing the ideas around you.<\/p><p>But you <i>know<\/i> the ideas are out there.  This is not one of those\nproblems where there might not be an answer.  It's impossibly\nunlikely that this is the exact moment when technological progress\nstops.  You can be sure people are going to build things in the\nnext few years that will make you think \"What did I do before x?\"<\/p><p>And when these problems get solved, they will probably seem flamingly\nobvious in retrospect.  What you need to do is turn off the filters\nthat usually prevent you from seeing them.  The most powerful is\nsimply taking the current state of the world for granted.  Even the\nmost radically open-minded of us mostly do that.  You couldn't get\nfrom your bed to the front door if you stopped to question everything.<\/p><p>But if you're looking for startup ideas you can sacrifice some of\nthe efficiency of taking the status quo for granted and start to\nquestion things.  Why is your inbox overflowing?  Because you get\na lot of email, or because it's hard to get email out of your inbox?\nWhy do you get so much email?  What problems are people trying to\nsolve by sending you email?  Are there better ways to solve them?\nAnd why is it hard to get emails out of your inbox?  Why do you\nkeep emails around after you've read them?  Is an inbox the optimal\ntool for that?<\/p><p>Pay particular attention to things that chafe you.  The advantage\nof taking the status quo for granted is not just that it makes life\n(locally) more efficient, but also that it makes life more tolerable.\nIf you knew about all the things we'll get in the next 50 years but\ndon't have yet, you'd find present day life pretty constraining,\njust as someone from the present would if they were sent back 50\nyears in a time machine.  When something annoys you, it could be\nbecause you're living in the future.<\/p><p>When you find the right sort of problem, you should probably be\nable to describe it as <i>obvious<\/i>, at least to you.  When we started\nViaweb, all the online stores were built by hand, by web designers\nmaking individual HTML pages.  It was obvious to us as programmers\nthat these sites would have to be generated by software.\n<span>[<a href=\"#f5n\"><span>5<\/span><\/a>]<\/span><\/p><p>Which means, strangely enough, that coming up with startup ideas\nis a question of seeing the obvious.  That suggests how weird this\nprocess is: you're trying to see things that are obvious, and yet\nthat you hadn't seen.<\/p><p>Since what you need to do here is loosen up your own mind, it may\nbe best not to make too much of a direct frontal attack on the\nproblem \u2014 i.e. to sit down and try to think of ideas.  The best\nplan may be just to keep a background process running, looking for\nthings that seem to be missing.  Work on hard problems, driven\nmainly by curiosity, but have a second self watching over your\nshoulder, taking note of gaps and anomalies.  \n<span>[<a href=\"#f6n\"><span>6<\/span><\/a>]<\/span><\/p><p>Give yourself some time.  You have a lot of control over the rate\nat which you turn yours into a prepared mind, but you have less\ncontrol over the stimuli that spark ideas when they hit it.  If\nBill Gates and Paul Allen had constrained themselves to come up\nwith a startup idea in one month, what if they'd chosen a month\nbefore the Altair appeared?  They probably would have worked on a\nless promising idea.  Drew Houston did work on a less promising\nidea before Dropbox: an SAT prep startup.  But Dropbox was a much\nbetter idea, both in the absolute sense and also as a match for his\nskills.\n<span>[<a href=\"#f7n\"><span>7<\/span><\/a>]<\/span><\/p><p>A good way to trick yourself into noticing ideas is to work on\nprojects that seem like they'd be cool.  If you do that, you'll\nnaturally tend to build things that are missing.  It wouldn't seem\nas interesting to build something that already existed.<\/p><p>Just as trying to think up startup ideas tends to produce bad ones,\nworking on things that could be dismissed as \"toys\" often produces\ngood ones.  When something is described as a toy, that means it has\neverything an idea needs except being important.  It's cool; users\nlove it; it just doesn't matter.  But if you're living in the future\nand you build something cool that users love, it may matter more\nthan outsiders think.  Microcomputers seemed like toys when Apple\nand Microsoft started working on them.  I'm old enough to remember\nthat era; the usual term for people with their own microcomputers\nwas \"hobbyists.\"  BackRub seemed like an inconsequential science\nproject.  The Facebook was just a way for undergrads to stalk one\nanother.<\/p><p>At YC we're excited when we meet startups working on things that\nwe could imagine know-it-alls on forums dismissing as toys.  To us\nthat's positive evidence an idea is good.<\/p><p>If you can afford to take a long view (and arguably you can't afford\nnot to), you can turn \"Live in the future and build what's missing\"\ninto something even better:\n<\/p><blockquote><div>\n  Live in the future and build what seems interesting.\n<\/div><\/blockquote><p><b>School<\/b><\/p><p>That's what I'd advise college students to do, rather than trying\nto learn about \"entrepreneurship.\"  \"Entrepreneurship\" is something\nyou learn best by doing it.  The examples of the most successful\nfounders make that clear.  What you should be spending your time\non in college is ratcheting yourself into the future.  College is\nan incomparable opportunity to do that.  What a waste to sacrifice\nan opportunity to solve the hard part of starting a startup \u2014 becoming \nthe sort of person who can have organic startup ideas \u2014 by \nspending time learning about the easy part.  Especially since\nyou won't even really learn about it, any more than you'd learn\nabout sex in a class.  All you'll learn is the words for things.<\/p><p>The clash of domains is a particularly fruitful source of ideas.\nIf you know a lot about programming and you start learning about\nsome other field, you'll probably see problems that software could\nsolve.  In fact, you're doubly likely to find good problems in\nanother domain: (a) the inhabitants of that domain are not as likely\nas software people to have already solved their problems with\nsoftware, and (b) since you come into the new domain totally ignorant,\nyou don't even know what the status quo is to take it for granted.<\/p><p>So if you're a CS major and you want to start a startup, instead\nof taking a class on entrepreneurship you're better off taking a\nclass on, say, genetics.  Or better still, go work for a biotech\ncompany.  CS majors normally get summer jobs at computer hardware\nor software companies.  But if you want to find startup ideas, you\nmight do better to get a summer job in some unrelated field. \n<span>[<a href=\"#f8n\"><span>8<\/span><\/a>]<\/span><\/p><p>Or don't take any extra classes, and just build things.  It's no\ncoincidence that Microsoft and Facebook both got started in January.\nAt Harvard that is (or was) Reading Period, when students have no\nclasses to attend because they're supposed to be studying for finals.\n<span>[<a href=\"#f9n\"><span>9<\/span><\/a>]<\/span><\/p><p>But don't feel like you have to build things that will become startups.  That's\npremature optimization. Just build things.  Preferably with other\nstudents.  It's not just the classes that make a university such a\ngood place to crank oneself into the future.  You're also surrounded\nby other people trying to do the same thing.  If you work together\nwith them on projects, you'll end up producing not just organic\nideas, but organic ideas with organic founding teams \u2014 and that,\nempirically, is the best combination.<\/p><p>Beware of research.  If an undergrad writes something all his friends\nstart using, it's quite likely to represent a good startup idea.\nWhereas a PhD dissertation is extremely unlikely to.  For some\nreason, the more a project has to count as research, the less likely\nit is to be something that could be turned into a startup.\n<span>[<a href=\"#f10n\"><span>10<\/span><\/a>]<\/span>\nI think the reason is that the subset of ideas that count as research\nis so narrow that it's unlikely that a project that satisfied that\nconstraint would also satisfy the orthogonal constraint of solving\nusers' problems.  Whereas when students (or professors) build\nsomething as a side-project, they automatically gravitate toward\nsolving users' problems \u2014 perhaps even with an additional energy\nthat comes from being freed from the constraints of research.<\/p><p><b>Competition<\/b><\/p><p>Because a good idea should seem obvious, when you have one you'll\ntend to feel that you're late.  Don't let that deter you.  Worrying\nthat you're late is one of the signs of a good idea.  Ten minutes\nof searching the web will usually settle the question.  Even if you\nfind someone else working on the same thing, you're probably not\ntoo late.  It's exceptionally rare for startups to be killed by\ncompetitors \u2014 so rare that you can almost discount the possibility.\nSo unless you discover a competitor with the sort of lock-in that\nwould prevent users from choosing you, don't discard the idea.<\/p><p>If you're uncertain, ask users.  The question of whether you're too\nlate is subsumed by the question of whether anyone urgently needs\nwhat you plan to make.  If you have something that no competitor\ndoes and that some subset of users urgently need, you have a\nbeachhead.  \n<span>[<a href=\"#f11n\"><span>11<\/span><\/a>]<\/span><\/p><p>The question then is whether that beachhead is big enough. Or more\nimportantly, who's in it: if the beachhead consists of people doing\nsomething lots more people will be doing in the future, then it's\nprobably big enough no matter how small it is.  For example, if\nyou're building something differentiated from competitors by the\nfact that it works on phones, but it only works on the newest phones,\nthat's probably a big enough beachhead.<\/p><p>Err on the side of doing things where you'll face competitors.\nInexperienced founders usually give competitors more credit than\nthey deserve.  Whether you succeed depends far more on you than on\nyour competitors.  So better a good idea with competitors than a\nbad one without.<\/p><p>You don't need to worry about entering a \"crowded market\" so long\nas you have a thesis about what everyone else in it is overlooking.\nIn fact that's a very promising starting point.  Google was that\ntype of idea.  Your thesis has to be more precise than \"we're going\nto make an x that doesn't suck\" though. You have to be able to\nphrase it in terms of something the incumbents are overlooking.\nBest of all is when you can say that they didn't have the courage\nof their convictions, and that your plan is what they'd have done\nif they'd followed through on their own insights.  Google was that\ntype of idea too.  The search engines that preceded them shied away\nfrom the most radical implications of what they were doing \u2014 particularly \nthat the better a job they did, the faster users would\nleave.<\/p><p>A crowded market is actually a good sign, because it means both\nthat there's demand and that none of the existing solutions are\ngood enough.  A startup can't hope to enter a market that's obviously\nbig and yet in which they have no competitors.  So any startup that\nsucceeds is either going to be entering a market with existing\ncompetitors, but armed with some secret weapon that will get them\nall the users (like Google), or entering a market that looks small\nbut which will turn out to be big (like Microsoft).  \n<span>[<a href=\"#f12n\"><span>12<\/span><\/a>]<\/span><\/p><p><b>Filters<\/b><\/p><p>There are two more filters you'll need to turn off if you want to\nnotice startup ideas: the unsexy filter and the schlep filter.<\/p><p>Most programmers wish they could start a startup by just writing\nsome brilliant code, pushing it to a server, and having users pay\nthem lots of money.  They'd prefer not to deal with tedious problems\nor get involved in messy ways with the real world.  Which is a\nreasonable preference, because such things slow you down.  But this\npreference is so widespread that the space of convenient startup\nideas has been stripped pretty clean.  If you let your mind wander\na few blocks down the street to the messy, tedious ideas, you'll\nfind valuable ones just sitting there waiting to be implemented.<\/p><p>The schlep filter is so dangerous that I wrote a separate essay\nabout the condition it induces, which I called \n<a href=\"http:\/\/paulgraham.com\/schlep.html\">schlep blindness<\/a>.\nI gave Stripe as an example of a startup that benefited from turning\noff this filter, and a pretty striking example it is.  Thousands\nof programmers were in a position to see this idea; thousands of\nprogrammers knew how painful it was to process payments before\nStripe.  But when they looked for startup ideas they didn't see\nthis one, because unconsciously they shrank from having to deal\nwith payments.  And dealing with payments is a schlep for Stripe,\nbut not an intolerable one.  In fact they might have had net less\npain; because the fear of dealing with payments kept most people\naway from this idea, Stripe has had comparatively smooth sailing\nin other areas that are sometimes painful, like user acquisition.\nThey didn't have to try very hard to make themselves heard by users,\nbecause users were desperately waiting for what they were building.<\/p><p>The unsexy filter is similar to the schlep filter, except it keeps\nyou from working on problems you despise rather than ones you fear.\nWe overcame this one to work on Viaweb. There were interesting\nthings about the architecture of our software, but we weren't\ninterested in ecommerce per se.  We could see the problem was one\nthat needed to be solved though.<\/p><p>Turning off the schlep filter is more important than turning off\nthe unsexy filter, because the schlep filter is more likely to be\nan illusion.  And even to the degree it isn't, it's a worse form\nof self-indulgence.  Starting a successful startup is going to be\nfairly laborious no matter what.  Even if the product doesn't entail\na lot of schleps, you'll still have plenty dealing with investors,\nhiring and firing people, and so on.  So if there's some idea you\nthink would be cool but you're kept away from by fear of the schleps\ninvolved, don't worry: any sufficiently good idea will have as many.<\/p><p>The unsexy filter, while still a source of error, is not as entirely\nuseless as the schlep filter.  If you're at the leading edge of a\nfield that's changing rapidly, your ideas about what's sexy will\nbe somewhat correlated with what's valuable in practice.  Particularly\nas you get older and more experienced.  Plus if you find an idea\nsexy, you'll work on it more enthusiastically. \n<span>[<a href=\"#f13n\"><span>13<\/span><\/a>]<\/span><\/p><p><b>Recipes<\/b><\/p><p>While the best way to discover startup ideas is to become the sort\nof person who has them and then build whatever interests you,\nsometimes you don't have that luxury.  Sometimes you need an idea\nnow.  For example, if you're working on a startup and your initial\nidea turns out to be bad.<\/p><p>For the rest of this essay I'll talk about tricks for coming up\nwith startup ideas on demand.  Although empirically you're better\noff using the organic strategy, you could succeed this way. You\njust have to be more disciplined.  When you use the organic method,\nyou don't even notice an idea unless it's evidence that something\nis truly missing.  But when you make a conscious effort to think\nof startup ideas, you have to replace this natural constraint with\nself-discipline.  You'll see a lot more ideas, most of them bad,\nso you need to be able to filter them.<\/p><p>One of the biggest dangers of not using the organic method is the\nexample of the organic method.  Organic ideas feel like inspirations.\nThere are a lot of stories about successful startups that began\nwhen the founders had what seemed a crazy idea but \"just knew\" it\nwas promising.  When you feel that about an idea you've had while\ntrying to come up with startup ideas, you're probably mistaken.<\/p><p>When searching for ideas, look in areas where you have some expertise.\nIf you're a database expert, don't build a chat app for teenagers\n(unless you're also a teenager).  Maybe it's a good idea, but you\ncan't trust your judgment about that, so ignore it.  There have to\nbe other ideas that involve databases, and whose quality you can\njudge.  Do you find it hard to come up with good ideas involving\ndatabases?  That's because your expertise raises your standards.\nYour ideas about chat apps are just as bad, but you're giving\nyourself a Dunning-Kruger pass in that domain.<\/p><p>The place to start looking for ideas is things you need.  There\n<i>must<\/i> be things you need.\n<span>[<a href=\"#f14n\"><span>14<\/span><\/a>]<\/span><\/p><p>One good trick is to ask yourself whether in your previous job you\never found yourself saying \"Why doesn't someone make x?  If someone\nmade x we'd buy it in a second.\" If you can think of any x people\nsaid that about, you probably have an idea.  You know there's demand,\nand people don't say that about things that are impossible to build.<\/p><p>More generally, try asking yourself whether there's something unusual\nabout you that makes your needs different from most other people's.\nYou're probably not the only one.  It's especially good if you're\ndifferent in a way people will increasingly be.<\/p><p>If you're changing ideas, one unusual thing about you is the idea\nyou'd previously been working on.  Did you discover any needs while\nworking on it?  Several well-known startups began this way.  Hotmail\nbegan as something its founders wrote to talk about their previous\nstartup idea while they were working at their day jobs. \n<span>[<a href=\"#f15n\"><span>15<\/span><\/a>]<\/span><\/p><p>A particularly promising way to be unusual is to be young.  Some\nof the most valuable new ideas take root first among people in their\nteens and early twenties.  And while young founders are at a\ndisadvantage in some respects, they're the only ones who really\nunderstand their peers.  It would have been very hard for someone\nwho wasn't a college student to start Facebook.  So if you're a\nyoung founder (under 23 say), are there things you and your friends\nwould like to do that current technology won't let you?<\/p><p>The next best thing to an unmet need of your own is an unmet need\nof someone else.  Try talking to everyone you can about the gaps\nthey find in the world.  What's missing?  What would they like to\ndo that they can't?  What's tedious or annoying, particularly in\ntheir work?  Let the conversation get general; don't be trying too\nhard to find startup ideas.  You're just looking for something to\nspark a thought.  Maybe you'll notice a problem they didn't consciously\nrealize they had, because you know how to solve it.<\/p><p>When you find an unmet need that isn't your own, it may be somewhat\nblurry at first.  The person who needs something may not know exactly\nwhat they need.  In that case I often recommend that founders act\nlike consultants \u2014 that they do what they'd do if they'd been\nretained to solve the problems of this one user.  People's problems\nare similar enough that nearly all the code you write this way will\nbe reusable, and whatever isn't will be a small price to start out\ncertain that you've reached the bottom of the well.\n<span>[<a href=\"#f16n\"><span>16<\/span><\/a>]<\/span><\/p><p>One way to ensure you do a good job solving other people's problems\nis to make them your own.  When Rajat Suri of E la Carte decided\nto write software for restaurants, he got a job as a waiter to learn\nhow restaurants worked.  That may seem like taking things to extremes,\nbut startups are extreme.  We love it when founders do such things.<\/p><p>In fact, one strategy I recommend to people who need a new idea is\nnot merely to turn off their schlep and unsexy filters, but to seek\nout ideas that are unsexy or involve schleps.  Don't try to start\nTwitter.  Those ideas are so rare that you can't find them by looking\nfor them.  Make something unsexy that people will pay you for.<\/p><p>A good trick for bypassing the schlep and to some extent the unsexy\nfilter is to ask what you wish someone else would build, so that\nyou could use it.  What would you pay for right now?<\/p><p>Since startups often garbage-collect broken companies and industries,\nit can be a good trick to look for those that are dying, or deserve\nto, and try to imagine what kind of company would profit from their\ndemise.  For example, journalism is in free fall at the moment.\nBut there may still be money to be made from something like journalism.\nWhat sort of company might cause people in the future to say \"this\nreplaced journalism\" on some axis?<\/p><p>But imagine asking that in the future, not now.  When one company\nor industry replaces another, it usually comes in from the side.\nSo don't look for a replacement for x; look for something that\npeople will later say turned out to be a replacement for x.  And\nbe imaginative about the axis along which the replacement occurs.\nTraditional journalism, for example, is a way for readers to get\ninformation and to kill time, a way for writers to make money and\nto get attention, and a vehicle for several different types of\nadvertising.  It could be replaced on any of these axes (it has\nalready started to be on most).<\/p><p>When startups consume incumbents, they usually start by serving\nsome small but important market that the big players ignore.  It's\nparticularly good if there's an admixture of disdain in the big\nplayers' attitude, because that often misleads them.  For example,\nafter Steve Wozniak built the computer that became the Apple I, he\nfelt obliged to give his then-employer Hewlett-Packard the option\nto produce it.  Fortunately for him, they turned it down, and one\nof the reasons they did was that it used a TV for a monitor, which\nseemed intolerably d\u00e9class\u00e9 to a high-end hardware company like HP\nwas at the time. \n<span>[<a href=\"#f17n\"><span>17<\/span><\/a>]<\/span><\/p><p>Are there groups of \n<a href=\"http:\/\/paulgraham.com\/marginal.html\">scruffy<\/a> \nbut sophisticated users like the early\nmicrocomputer \"hobbyists\" that are currently being ignored by the\nbig players?  A startup with its sights set on bigger things can\noften capture a small market easily by expending an effort that\nwouldn't be justified by that market alone.<\/p><p>Similarly, since the most successful startups generally ride some\nwave bigger than themselves, it could be a good trick to look for\nwaves and ask how one could benefit from them.  The prices of gene\nsequencing and 3D printing are both experiencing Moore's Law-like\ndeclines.  What new things will we be able to do in the new world\nwe'll have in a few years?  What are we unconsciously ruling out\nas impossible that will soon be possible?<\/p><p><b>Organic<\/b><\/p><p>But talking about looking explicitly for waves makes it clear that\nsuch recipes are plan B for getting startup ideas.  Looking for\nwaves is essentially a way to simulate the organic method.  If\nyou're at the leading edge of some rapidly changing field, you don't\nhave to look for waves; you are the wave.<\/p><p>Finding startup ideas is a subtle business, and that's why most\npeople who try fail so miserably.  It doesn't work well simply to\ntry to think of startup ideas.  If you do that, you get bad ones\nthat sound dangerously plausible.  The best approach is more indirect:\nif you have the right sort of background, good startup ideas will\nseem obvious to you.  But even then, not immediately.  It takes\ntime to come across situations where you notice something missing.\nAnd often these gaps won't seem to be ideas for companies, just\nthings that would be interesting to build.  Which is why it's good\nto have the time and the inclination to build things just because\nthey're interesting.<\/p><p>Live in the future and build what seems interesting.  Strange as\nit sounds, that's the real recipe.<\/p><p><b>Notes<\/b><\/p><p>[<a name=\"f1n\" id=\"f1n\"><span>1<\/span><\/a>]\nThis form of bad idea has been around as long as the web.  It\nwas common in the 1990s, except then people who had it used to say\nthey were going to create a portal for x instead of a social network\nfor x.  Structurally the idea is stone soup: you post a sign saying\n\"this is the place for people interested in x,\" and all those people\nshow up and you make money from them.  What lures founders into\nthis sort of idea are statistics about the millions of people who\nmight be interested in each type of x.  What they forget is that\nany given person might have 20 affinities by this standard, and no\none is going to visit 20 different communities regularly.<\/p><p>[<a name=\"f2n\" id=\"f2n\"><span>2<\/span><\/a>]\nI'm not saying, incidentally, that I know for sure a social\nnetwork for pet owners is a bad idea.  I know it's a bad idea the\nway I know randomly generated DNA would not produce a viable organism.\nThe set of plausible sounding startup ideas is many times larger\nthan the set of good ones, and many of the good ones don't even\nsound that plausible.  So if all you know about a startup idea is\nthat it sounds plausible, you have to assume it's bad.<\/p><p>[<a name=\"f3n\" id=\"f3n\"><span>3<\/span><\/a>]\nMore precisely, the users' need has to give them sufficient\nactivation energy to start using whatever you make, which can vary\na lot.  For example, the activation energy for enterprise software\nsold through traditional channels is very high, so you'd have to\nbe a <i>lot<\/i> better to get users to switch.  Whereas the activation\nenergy required to switch to a new search engine is low.  Which in\nturn is why search engines are so much better than enterprise\nsoftware.<\/p><p>[<a name=\"f4n\" id=\"f4n\"><span>4<\/span><\/a>]\nThis gets harder as you get older.  While the space of ideas\ndoesn't have dangerous local maxima, the space of careers does.\nThere are fairly high walls between most of the paths people take\nthrough life, and the older you get, the higher the walls become.<\/p><p>[<a name=\"f5n\" id=\"f5n\"><span>5<\/span><\/a>]\nIt was also obvious to us that the web was going to be a big\ndeal.  Few non-programmers grasped that in 1995, but the programmers\nhad seen what GUIs had done for desktop computers.<\/p><p>[<a name=\"f6n\" id=\"f6n\"><span>6<\/span><\/a>]\nMaybe it would work to have this second self keep a journal,\nand each night to make a brief entry listing the gaps and anomalies\nyou'd noticed that day.  Not startup ideas, just the raw gaps and\nanomalies.<\/p><p>[<a name=\"f7n\" id=\"f7n\"><span>7<\/span><\/a>]\nSam Altman points out that taking time to come up with an\nidea is not merely a better strategy in an absolute sense, but also\nlike an undervalued stock in that so few founders do it.<\/p><p>There's comparatively little competition for the best ideas, because\nfew founders are willing to put in the time required to notice them.\nWhereas there is a great deal of competition for mediocre ideas,\nbecause when people make up startup ideas, they tend to make up the\nsame ones.<\/p><p>[<a name=\"f8n\" id=\"f8n\"><span>8<\/span><\/a>]\nFor the computer hardware and software companies, summer jobs\nare the first phase of the recruiting funnel.  But if you're good\nyou can skip the first phase.  If you're good you'll have no trouble\ngetting hired by these companies when you graduate, regardless of\nhow you spent your summers.<\/p><p>[<a name=\"f9n\" id=\"f9n\"><span>9<\/span><\/a>]\nThe empirical evidence suggests that if colleges want to help\ntheir students start startups, the best thing they can do is leave\nthem alone in the right way.<\/p><p>[<a name=\"f10n\" id=\"f10n\"><span>10<\/span><\/a>]\nI'm speaking here of IT startups; in biotech things are different.<\/p><p>[<a name=\"f11n\" id=\"f11n\"><span>11<\/span><\/a>]\nThis is an instance of a more general rule: focus on users,\nnot competitors.  The most important information about competitors\nis what you learn via users anyway.<\/p><p>[<a name=\"f12n\" id=\"f12n\"><span>12<\/span><\/a>]\nIn practice most successful startups have elements of both.\nAnd you can describe each strategy in terms of the other by adjusting\nthe boundaries of what you call the market.  But it's useful to\nconsider these two ideas separately.<\/p><p>[<a name=\"f13n\" id=\"f13n\"><span>13<\/span><\/a>]\nI almost hesitate to raise that point though.  Startups are\nbusinesses; the point of a business is to make money; and with that\nadditional constraint, you can't expect you'll be able to spend all\nyour time working on what interests you most.<\/p><p>[<a name=\"f14n\" id=\"f14n\"><span>14<\/span><\/a>]\nThe need has to be a strong one.  You can retroactively\ndescribe any made-up idea as something you need.  But do you really\nneed that recipe site or local event aggregator as much as Drew\nHouston needed Dropbox, or Brian Chesky and Joe Gebbia needed Airbnb?<\/p><p>Quite often at YC I find myself asking founders \"Would you use this\nthing yourself, if you hadn't written it?\" and you'd be surprised\nhow often the answer is no.<\/p><p>[<a name=\"f15n\" id=\"f15n\"><span>15<\/span><\/a>]\nPaul Buchheit points out that trying to sell something bad\ncan be a source of better ideas:<\/p><p>\"The best technique I've found for dealing with YC companies that\nhave bad ideas is to tell them to go sell the product ASAP (before\nwasting time building it). Not only do they learn that nobody\nwants what they are building, they very often come back with a\nreal idea that they discovered in the process of trying to sell\nthe bad idea.\"<\/p><p>[<a name=\"f16n\" id=\"f16n\"><span>16<\/span><\/a>]\nHere's a recipe that might produce the next Facebook, if\nyou're college students.  If you have a connection to one of the\nmore powerful sororities at your school, approach the queen bees\nthereof and offer to be their personal IT consultants, building\nanything they could imagine needing in their social lives that\ndidn't already exist.  Anything that got built this way would be\nvery promising, because such users are not just the most demanding\nbut also the perfect point to spread from.<\/p><p>I have no idea whether this would work.<\/p><p>[<a name=\"f17n\" id=\"f17n\"><span>17<\/span><\/a>]\nAnd the reason it used a TV for a monitor is that Steve Wozniak\nstarted out by solving his own problems.  He, like most of his\npeers, couldn't afford a monitor.<\/p><p><b>Thanks<\/b> to Sam Altman, Mike Arrington, Paul Buchheit, John Collison,\nPatrick Collison, Garry Tan, and Harj Taggar for reading drafts of\nthis, and Marc Andreessen, Joe Gebbia, Reid Hoffman, Shel Kaphan,\nMike Moritz and Kevin Systrom for answering my questions about\nstartup history.<\/p>","created_at":"2018-11-07T16:48:44+0000","updated_at":"2018-11-07T16:48:44+0000","annotations":[],"mimetype":"text\/html","reading_time":36,"domain_name":"paulgraham.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":449,"title":"High Noon Drifter brings Doom to the Weird West","url":"https:\/\/www.rockpapershotgun.com\/2017\/11\/12\/high-noon-drifter-doom-mod\/","content":"<div><main><p>One of the best things about Doom is that there\u2019s a version of it for everybody. If you like your weapons a little more old-school and your Westerns thoroughly Weird, then High Noon Drifter might just be what you\u2019re looking for in a retro FPS.<\/p>\n<p>It\u2019s High Noon (or was when I was writing this), so dust off your hat and slip into the well-worn cowboy boots of ghostly drifter Corzo for your regularly scheduled fix of demon-slaughter across a thousand worlds with an arsenal of chunky revolvers, lever-action shotguns and a vicious whip.<\/p>\n<p>While not the freshest of mods now, the version 1.2 update this week is a major upgrade, officially adding support for Hexen, as well as claymation freeware FPS\u00a0<a href=\"http:\/\/rabotik.nl\/harmony.htm\">Harmony<\/a>, plus your usual slew of new features, tweaks and improvements. You can see the full list in <a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=43&amp;t=57098\">the impressively long patch-notes here<\/a>, or get the abridged version in the slickly edited trailer below.<\/p>\n<p>The success of High Noon Drifter is more of a surprise to its creator (prolific modder TerminusEst13) than anyone. Originally planned to be a learning exercise so he could teach himself the new ZScript system, a new modding language in recent iterations of ZDoom. The project has evolved well past its original release plans, and is now one of the more popular character mods available for its flexibility and overall stylish nature. Who doesn\u2019t want to play as a whip-cracking ghost cowboy, after all?<\/p>\n<p>While a pretty straightforward (and thus easier to balance) character mod, High Noon Drifter does have a few tricks up its sleeve. Best of all has to be the Throw Hat button. When the full HUD is visible, you can see the brim of your hat just at the top of your field of view. Throwing it distracts enemies, forcing them to shoot at it briefly, a bit like the cigarette button in Vanquish. You can even keep your own hat aloft longer by shooting it, and pick it up \u2013 good as new \u2013 once it lands. Just wouldn\u2019t be proper to leave your hat behind, after all.<\/p>\n<p>There\u2019s some other goodies, too. The ultimate weapon of the mod, Basilissa\u2019s Mask, allows you to either throw beefy fireballs or \u2013 far more interestingly \u2013 tag Corzo out for a while and let his co-protagonist Basilissa, Baroness Of Hell come out to play. Corzo himself stays behind in a nice protective circle while you go on a time-limited demonic rampage, and when unsummoned, your perspective shifts back to where you were previously standing. It\u2019s a fun power trip, but not <em>wildly<\/em> more powerful than the good old BFG.<\/p>\n<p>High Noon Drifter is purely a character &amp; weapon modification, but it\u2019s designed specifically to play nice with replacement enemies, and even other Doom-engine games such as Heretic or even Chex Quest. The<a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=43&amp;t=57098\"> ZDoom forum thread for the mod contains<\/a> an official list of recommended levels and mods to mix it with. One of the more interesting being Kriegsland, a Wolfenstein: The New Order-inspired thing replacing Doom\u2019s demons with a legion of pulp sci-fi Nazis. Because Doom can be about a ghost-cowboy fighting <a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=43&amp;t=57053\">space-Nazis<\/a> in <a href=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/megawads\/epic2\">ancient Egypt<\/a> and still feel like Doom, somehow.<\/p>\n<p>And yes, it is compatible with Brutal Doom. The <a href=\"http:\/\/www.moddb.com\/mods\/brutal-doom\/addons\/brutal-doom-monsters-only\">monsters-only variant<\/a>, at least.<\/p>\n<p>You can <a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=43&amp;t=57098\">get High Noon Drifter here<\/a>. You\u2019ll need <a href=\"https:\/\/zdoom.org\/downloads\">a newer version of GZDoom<\/a> to run it\u00a0plus Doom 2, although you can stack up almost any other Doom-engine WAD on top of all of that to get yourself another world to mosey on through.<\/p>\n\t\t\t\t\t\t\t\t\t<\/main><\/div><div class=\"container-690\"><p class=\"disclaimer\">\n\t\t\t\t\tIf you click our links to online stores and make a purchase we may receive a few pennies. Find more information <a href=\"https:\/\/www.rockpapershotgun.com\/affiliate-linking-policy\/\">here<\/a>.\t\t\t\t<\/p><a href=\"https:\/\/www.rockpapershotgun.com\/support-us?utm_source=article-promo&amp;utm_medium=web&amp;utm_campaign=supporter\" class=\"support-us-promo\" rel=\"nofollow\">\n\t\t\n\t\t\t<img src=\"https:\/\/www.rockpapershotgun.com\/wp-content\/themes\/rps-2018\/images\/logo-inverted.svg\" alt=\"Rock Paper Shotgun\" \/>\n\t\t\n\t\t\n\t\t\t\n\t\t\t\tSupport RPS and get an ad-free site, extra articles, and free stuff!\t\t\t\n\t\t\tTell me more\n\t\t\n\t\t\n\t<\/a><\/div>","created_at":"2018-11-07T16:48:42+0000","updated_at":"2018-11-07T16:48:42+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/assets.rockpapershotgun.com\/images\/\/2017\/11\/highnoon-620x300.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":450,"title":"Concise Electronics for Geeks (2010)","url":"https:\/\/news.ycombinator.com\/item?id=15703472","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15703472\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15703472\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/lcamtuf.coredump.cx\/electronics\/\" class=\"storylink\">Concise Electronics for Geeks (2010)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=coredump.cx\">coredump.cx<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        481 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=nbmh\" class=\"hnuser\">nbmh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15703472\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15703472&amp;goto=item%3Fid%3D15703472\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Concise%20Electronics%20for%20Geeks&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Concise%20Electronics%20for%20Geeks\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15703472&amp;auth=32cd41762505cb10dbfd82c07ee5bd8c07f03a5f\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15703472\">49\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:48:42+0000","updated_at":"2018-11-07T16:48:42+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":447,"title":"6 great Doom mods not called \u201cBrutal Doom.\u201d","url":"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/","content":"<p class=\"western\">I\u2019m not a big fan of list articles. At best, you could find interesting stuff that might intrigue you and maybe share to your friends on Facebook. At worst, you find terrible click-bait articles that seem to be written more for a paycheck than any informative value. It\u2019s something I\u2019ve refrained from doing here, as\u00a0I prefer writing interesting long form stuff instead.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1032\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png\"><img data-attachment-id=\"1032\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/crackedinanutshell\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png\" data-orig-size=\"680,800\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=255\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=680\" class=\"wp-image-1032\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=320&amp;h=376\" alt=\"Seriously, this is what Cracked is now. I weep for our future.\" width=\"320\" height=\"376\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=320&amp;h=376 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=640&amp;h=752 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=128&amp;h=150 128w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/crackedinanutshell.png?w=255&amp;h=300 255w\" \/><\/a><figcaption class=\"wp-caption-text\">Seriously, this is what Cracked is now.\u00a0Remember when they actually wrote parody articles? Probably better than \u201c11 Amazing Things You Didn\u2019t Know About Your Pants,\u201d anyway.<\/figcaption><\/figure><p class=\"western\">One particular list article by Cracked irked me considerably. A recent list, \u201c<a href=\"http:\/\/www.cracked.com\/article_20987_6-awesome-hacks-that-did-mind-blowing-things-with-old-games.html\">6 Awesome Hacks That Did Mind-Blowing Things With Old Games<\/a>\u201d featured some cool stuff like Iron Man or the Incredible Hulk in <i>Grand Theft Auto IV<\/i>, or the entirety of <i>The Elder Scrolls: Morrowind<\/i> in <i>Oblivion<\/i>\u2018s engine. Even <em>Just Cause 2: Multiplayer Mod<\/em>, where you could go crazy in the world of <i>Just Cause 2<\/i> with hundreds of people, made the list. But what was number one? Brutal Doom.<\/p>\n<p class=\"western\">I won\u2019t go too much into Brutal Doom as it\u2019s made the internet rounds everywhere over the past couple of years, but it\u2019s championed as the \u201cdefinitive way to play Doom,\u201d with more gore, violence, <em>Mortal Kombat<\/em>-style\u00a0fatalities, even a key that\u2019s simply dedicated to flipping off enemies. It\u2019s amazing at first, but it outstays its welcome pretty quick.<\/p>\n<p class=\"western\">I\u2019m gonna be honest: I think Brutal Doom is an overrated, mediocre mod. The only thing it has going for it is the ridiculous macho\u00a0factor, and that seems to be championed by every average dude who always writes about the Doom mod scene. Brutal Doom is usually mentioned as \u201cthe way Doom was meant to be,\u201d but it really isn\u2019t. Brutal Doom isn\u2019t the way Doom was meant to be played, it\u2019s Doom\u00a0if it was\u00a0a terrible caricature of itself. It\u2019s the <a href=\"http:\/\/doomwiki.org\/wiki\/Doom_comic\">Doom comic<\/a> in game form.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1047\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png\"><img data-attachment-id=\"1047\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_105708\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_105708\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=809\" class=\"wp-image-1047\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=320&amp;h=240\" alt=\"Seriously, I wonder if there's people who love this sort of stuff.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_105708.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">All this does is make Doomguy look like a god damn psychopath who should be in a mental asylum, not fighting monsters.<\/figcaption><\/figure><p class=\"western\">So, as a response to the article (as well as breaking my own personal rule of no lists), I\u2019m making a list of 6 awesome <em>Doom<\/em> mods that aren\u2019t called Brutal Doom. These are ones that the Doom community swear by, and are more worth your time than Brutal Doom any day of the week.<\/p>\n<p class=\"western\">As always, these require <em>Doom<\/em> to run. You can get <a href=\"http:\/\/store.steampowered.com\/app\/2300\/\"><em>Doom II<\/em> on Steam<\/a> for the low low price of $5. After that, you\u2019ll need a source port to play these. I recommend <a href=\"http:\/\/forum.drdteam.org\/viewforum.php?f=23\">GZDoom<\/a> (for Singleplayer) and <a href=\"http:\/\/zandronum.com\/\">Zandronum<\/a> (for Multiplayer). While these mods will work perfectly fine with the default <em>Doom II<\/em> levels, I do recommend\u00a0playing these with\u00a0custom PWADs, which I\u2019ll link to as well. Unless stated otherwise, these are meant to be played in single player.<\/p>\n<p class=\"western\"><b>Police Brutality: Wildweasel presents Terrorists!<\/b><\/p>\n<p class=\"western\">(<a href=\"http:\/\/www.doomworld.com\/idgames\/index.php?file=combos\/ww-trror.zip\">idgames link<\/a>)<\/p>\n<p class=\"western\">Ever wanted to be an action hero? Terrorists! will live out your dreams of being the next Chuck Norris. Armed with only a pistol and the ability to kick dudes in the face, this mod adds real life weapons and enemies for you to fight in.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1063\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png\"><img data-attachment-id=\"1063\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140607_191011\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140607_191011\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=809\" class=\"wp-image-1063\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=320&amp;h=240\" alt=\"Stopping crime the only way possible: with a six shooter and a bunch of bullets.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1910111.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Stopping crime the only way possible: with a six shooter and a bunch of bullets.<\/figcaption><\/figure><p class=\"western\">In addition to the weapons and enemies, there\u2019s also a new feature where you level up your guns by killing enemies with them. Upgrades start out simple, like faster fire rate, but as you level them up, they get more crazy, like a Beretta that converts to burst fire, or a revolver that becomes a long-range rifle. Even your melee and grenades can get upgrades, from electric grenades to explosive roundhouse kicks.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1050\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png\"><img data-attachment-id=\"1050\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_182539\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_182539\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=809\" class=\"wp-image-1050\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=320&amp;h=240\" alt=\"Our hero, kicking robots like it ain't no big deal.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_182539.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Our hero, kicking robots like it ain\u2019t no big deal.<\/figcaption><\/figure><p class=\"western\">Wildweasel\u2019s made some other great mods, like the WWII-inspired <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=29815\">Nazis!<\/a> (which goes great with the Egyptian themed <a href=\"http:\/\/www.doomworld.com\/idgames\/index.php?file=levels\/doom2\/megawads\/epic2.zip\">EPIC 2<\/a>), and the action packed <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=27102\">Diaz<\/a>. Terrorists became one of my favorites, only because of the weapon upgrades. The three I just mentioned are all pretty good mods for Doom, and give enough gameplay changes to make it just as fun. For those who want to live out their action movie dreams without actually getting hurt, Terrorists! will do the job nicely.<\/p>\n<p class=\"western\"><i>(Disclaimer: I might have a slight bias as I am friends with the guy who made this mod. It\u2019s still high quality, though!)<\/i><\/p>\n<p class=\"western\"><b>Samsara<\/b><\/p>\n<p class=\"western\">(<a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=33219\">ZDoom forum link<\/a>)<\/p>\n<p class=\"western\">Ever wanted to play through Doom campaigns with characters besides Doomguy? Well, Samsara adds characters from many old games of the era, from Duke Nukem, to\u00a0B.J. Blazkowicz, even the heroes from <em>Chex Quest<\/em> and <em>Marathon<\/em> make an appearance here. Now playing as each character will allow you to use only that character\u2019s weapons, so you can\u2019t run around with 4-5 weapons from different games, sadly. However, that\u2019s a compromise I can deal with considering the variety of classic characters involved.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1060\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png\"><img data-attachment-id=\"1060\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_193005\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_193005\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=809\" class=\"wp-image-1060\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=320&amp;h=240\" alt=\"Yeah. I'm the Ranger, a friend of mine is the Chex Quest guy, and we're fighting a Baron of Hell. Welcome to Samsara.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_193005.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Yeah. I\u2019m the Ranger, a friend of mine is the Chex Quest guy, and we\u2019re fighting a Baron of Hell. Welcome to Samsara.<\/figcaption><\/figure><p class=\"western\">I\u2019ve always loved crossovers between different game series, official and non-official. Samsara scratches that crossover itch. Playing as the various characters gives a much different take on Doom. Nothing\u2019s more fun than ripping through Chex Quest as Duke Nukem, or playing custom levels like Community Chest 4 with Ranger, or bringing firearms to the world of <em>Heretic<\/em>.\u00a0There\u2019s even mods that add the enemies from those games, giving us an unusual mix that you normally don\u2019t see in games like this.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1056\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png\"><img data-attachment-id=\"1056\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_194447\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_194447\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=809\" class=\"wp-image-1056\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=320&amp;h=240\" alt=\"B.J. Blazkowicz, Duke Nukem, and the guy from Marathon, all fighting on the same level. What's not to like?\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_194447.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">B.J. Blazkowicz, Duke Nukem, and the guy from Marathon, all fighting on the same level. What\u2019s not to like?<\/figcaption><\/figure><p class=\"western\">Samsara is\u00a0meant to be played online, either fragging with friends in deathmatch or working together in co-op, with each person choosing different characters for each situation. Though you can play this single player in GZDoom, it\u2019s not the recommended way to play this. Get some buddies together in survival co-op and rip through as many levels as possible. Just make sure you put it on random character for the ideal\u00a0Samsara\u00a0experience.<\/p>\n<p class=\"western\">\n<\/p><p class=\"western\"><b>Reelism<\/b><\/p>\n<p class=\"western\">(<a href=\"http:\/\/www.thekinsie.com\/reelism\/\">Official website<\/a>)<\/p>\n<p class=\"western\">Let me put it this way: Reelism is one bizarre mod. It\u2019s a wave-based survival mod where you score\u00a0points by killing dudes, but as opposed to other Horde Mode-style games, this one\u2019s different. Every wave starts a slot machine which determines the the rules for the wave. Could be anything from fighting <em>Wolfenstein 3D<\/em> enemies with explosive weapons and the ability to fly, to fighting enemies from <em>Strife<\/em> in a freakin\u2019 tank. This, coupled with the levels of a bridge, a hell planet and a generic JRPG world, makes for a\u00a0ridiculous\u00a0experience that you gotta\u00a0play\u00a0to believe.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1053\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png\"><img data-attachment-id=\"1053\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140604_183706\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140604_183706\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=809\" class=\"wp-image-1053\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=320&amp;h=240\" alt=\"I wasn't kidding about the tank part.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140604_1837061.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">I wasn\u2019t kidding about the tank part.<\/figcaption><\/figure><p class=\"western\">Some of the mods make it so you\u2019re using weapons from a certain game and enemies usually from a different game, but sometimes they throw tricks in, like activating\u00a0<em>Rise of the Triad<\/em><em>\u2018s\u00a0<\/em>Dog Mode for a Round, or throwing in harmless bronies. Others make it more challenging, like removing friction to Nintendo 64-like fog to exploding enemies. Some are just useless, like touch screen mode. All of this plays into Reelism\u2019s absurdity.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1044\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png\"><img data-attachment-id=\"1044\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_095004\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_095004\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=320&amp;h=240\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=809\" class=\"wp-image-1044\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=320&amp;h=240\" alt=\"I'm in a futuristic cyberworld fighting Doom enemies with Hexen weapons and can fly around at will. Don't try to understand it, it'll just hurt your brain.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_095004.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">I\u2019m in a futuristic cyber world fighting <em>Doom<\/em> enemies with <em>Hexen<\/em> weapons and can fly around at will. Don\u2019t try to understand it, you\u2019ll just strain\u00a0your brain.<\/figcaption><\/figure><p class=\"western\">Basically Reelism is the god damn melting pot of Doom mods. The items, enemies and weapons come from dozens of different old-school shooters, while also adding a few enemies of its own. All the levels are inspired by other games, and it has no consistent style. Though in this case, the randomness of it works in Reelism\u2019s favor, making sure the experience is different every time you play.<\/p>\n<p class=\"western\">However,\u00a0the mod is not easy. Most of the time I was lucky to get past three rounds, and that meant games that were no longer than 2-3 minutes at a time. Despite that, it\u2019s a great mod for when you got a few minutes of time to kill.<\/p>\n<p class=\"western\"><b>Psychic<\/b><\/p>\n<p class=\"western\">(<a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=21681\">ZDoom forum thread<\/a>)<\/p>\n<p class=\"western\">Psychic is a bit different from the other mods shown here. Where most of these just replaced weapons or threw in everything but the kitchen sink, Psychic adds elemental magic to its arsenal.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1054\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png\"><img data-attachment-id=\"1054\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140607_143935\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140607_143935\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=809\" class=\"wp-image-1054\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=320&amp;h=240\" alt=\"The shop. Want to heal instantly or get weapons immediately? Here you go. Just make sure you got enough bloody cash.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_143935.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">The shop. Want to heal instantly or get weapons immediately? Here you go. Just make sure you got enough bloody cash.<\/figcaption><\/figure><p class=\"western\">When you kill enemies, you pick up Bloodmoney which you use to buy items in the shop. The things you can buy range from\u00a0buying\/upgrading\u00a0weapons and magic, health and armor pickups, even brief powerups like enemy damage giving you health back or enemies being scared of you. Bigger enemies give more Bloodmoney, especially if killed with your elemental powers.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1030\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png\"><img data-attachment-id=\"1030\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140607_154557\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=809\" class=\"wp-image-1030\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=320&amp;h=240\" alt=\"\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_154557.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Learn lightning, and strike these fools down in an instant. They\u2019ll never know what hit \u2019em.<\/figcaption><\/figure><p class=\"western\">This gives a wholly new <em>Doom<\/em> experience, even with level packs like <a href=\"http:\/\/www.doomworld.com\/idgames\/index.php?file=levels\/doom2\/megawads\/requiem.zip\">Requiem<\/a>. Granted, this doesn\u2019t make Doom itself any easier, in fact it\u2019s a bit more challenging than the other mods mostly\u00a0due to the limited weapons available. Don\u2019t expect this mod to make <em>Doom<\/em> a cakewalk, even with powerful magic and guns.<\/p>\n<p class=\"western\">At times, Psychic\u2019s design feels like cyberpunk with a pinch of Gothic flavor, which makes it stand out a bit differently than the other mods shown here, and it\u2019s pretty cool. It reminds me of 2000s-era action\/horror movies, which was kind of a popular thing during the era. It\u2019s quite a shame the idea of guns and magic has never been touched upon\u00a0much before this, because Psychic has an interesting idea here.<\/p>\n<p class=\"western\"><b>Beautiful Doom <\/b><\/p>\n<p class=\"western\">(<a href=\"http:\/\/www.moddb.com\/games\/doom-ii\/addons\/beautiful-doom-v53\">Mod DB page<\/a>)<\/p>\n<p class=\"western\">Up to this point, I\u2019ve covered mods that drastically changed the gameplay. Usually by adding new weapons and abilities. But what about the more cosmetic side of Doom modding? Where it\u2019s more about looks than drastically changing the gameplay mechanics? I\u2019ve got your answer right here with Beautiful Doom.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1052\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png\"><img data-attachment-id=\"1052\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140607_161555-2\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140607_161555\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=809\" class=\"wp-image-1052\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=320&amp;h=240\" alt=\"Screenshot_Doom_20140607_161555\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140607_1615551.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">The happy medium: Gory violence, while still\u00a0playing like the Doom we all grew up with.<\/figcaption><\/figure><p class=\"western\">I always recommend that new <em>Doom<\/em> players try to play <em>Doom<\/em> the closest to the original as possible (that means no mouse look, jumping, or crouching!) because using mods or source ports that add features can basically break the game and give players the wrong impression on how <em>Doom<\/em> was played. Beautiful Doom doesn\u2019t do that. The core gameplay is the same as it was back in \u201994, but now features flashier graphics, more blood and wonderful particle effects. It takes those kind of aesthetics that Brutal Doom enforced, but without the drastic gameplay changes alongside it.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1051\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png\"><img data-attachment-id=\"1051\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_183701\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_183701\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=809\" class=\"wp-image-1051\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=320&amp;h=240\" alt=\"All the action of Brutal Doom but without the bitter aftertaste.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_183701.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">All the violence of Brutal Doom but without the bitter aftertaste.<\/figcaption><\/figure><p class=\"western\">If you liked the pretty graphical effects of Brutal Doom, but didn\u2019t care for the ridiculous dismemberment, I recommend Beautiful Doom\u00a0instead. Same cool effects, but with the classic Doom gameplay experience intact. No\u00a0ridiculous finishing moves and taunt buttons here, and it\u2019s better as a result.<\/p>\n<p class=\"western\"><b>Russian Overkill<\/b><\/p>\n<p class=\"western\">(<a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?t=29915\">ZDoom forum page<\/a>)<\/p>\n<p class=\"western\">I saved the best for last. Get ready to praise glorious Mother Russia, comrade.<\/p>\n<p class=\"western\">Russian Overkill lives up to its name. It has ridiculous weapons that pretty much destroy anything in their path. Dozens of shotguns, machine guns, and rocket launchers are at your dispersal. There\u2019s also Black Hole launchers, explosive plasma launchers that can obliterate large groups of enemies, shotguns where the alternate fire launches spent shell casings\u2026 this mod is god damn nuts.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1041\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png\"><img data-attachment-id=\"1041\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_021515\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_021515\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=809\" class=\"wp-image-1041\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=320&amp;h=240\" alt=\"Because who doesn't want to kill their enemies with loads of spent shotgun shells?\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_021515.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Because who doesn\u2019t want to kill their enemies with loads of spent shotgun shells?<\/figcaption><\/figure><p class=\"western\">This mod is clearly meant for those who just wanna kill everything in sight with the most ridiculous, overpowered guns without any worry of dying. Even through later levels of some of the hardest mods, I had loads of ammo and plenty of health and armor to spare. I was an unstoppable killing machine, the way Stalin would\u2019ve intended. Provided he was around and supported crazy\u00a0weaponry like this.<\/p>\n<p class=\"western\">Seriously, this is meant to be played with levels where there\u2019s boatloads of enemies at every turn. Throw on something like\u00a0<a href=\"http:\/\/www.doomworld.com\/idgames\/index.php?file=levels\/doom2\/Ports\/megawads\/scythe2.zip\">Scythe II<\/a> and let \u2018er rip with these weapons. The weapons are so overpowered that you\u2019ll feel like a real badass even if you\u2019re not playing on the highest difficulty.<\/p>\n<p class=\"western\">Even the difficulty levels are ridiculous. One of the difficulty levels just replaces all the default weapons with the most powerful, devastating, area-clearing weapons. It\u2019s great if you wanna clear entire levels with these guns, but bad if you value your survival at all.<\/p>\n<figure data-shortcode=\"caption\" id=\"attachment_1048\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png\"><img data-attachment-id=\"1048\" data-permalink=\"https:\/\/youfoundasecret.wordpress.com\/2014\/06\/08\/6-great-doom-mods-not-called-brutal-doom\/screenshot_doom_20140608_111901\/\" data-orig-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png\" data-orig-size=\"1024,768\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screenshot_Doom_20140608_111901\" data-image-description=\"\" data-medium-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=300\" data-large-file=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=809\" class=\"wp-image-1048\" src=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=320&amp;h=240\" alt=\"Shortly after I used this, I killed pretty much everything in the room. Including myself.\" width=\"320\" height=\"240\" srcset=\"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=320&amp;h=240 320w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=640&amp;h=480 640w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=150&amp;h=113 150w, https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=300&amp;h=225 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Shortly after I used this, I killed pretty much everything. Including myself.<\/figcaption><\/figure><p class=\"western\">If there\u2019s any <em>Doom<\/em> mod from the list that should be tried at least, it\u2019s this one. It\u2019s overpowered and hardly balanced, but it doesn\u2019t really need to be. You\u2019ll want to wreck everything in sight, and you don\u2019t care what you use to get there. It truly is overkill.<\/p>\n<p class=\"western\"><b>Conclusion<\/b><\/p>\n<p class=\"western\">Granted, there\u2019s a lot of <em>Doom<\/em> mods that have come out for the past 20+ years. More than likely I missed a few here and there, so if there\u2019s a Doom mod that\u2019s a personal favorite, leave it in the comments.<\/p>\n<p class=\"western\">Keep in mind that I don\u2019t hate Brutal Doom per se, stylistically it\u2019s a well-done mod. It\u2019s the excess gore, the over-the-top elements, and its ridiculous \u201chardcore\u201d attitude that rubs me the wrong way. It\u2019s not a bad mod, just highly overrated. All the other mods I mentioned are much better, more fun experiences. For anyone who wants to get into what the Doom community\u2019s been doing over the past few years, these are much, much better examples than what Cracked thought deserved the #1 spot in their crappy mod list. (I would\u2019ve replaced the <em>Baldur\u2019s Gate<\/em> mod with something like\u00a0<em>Black Mesa<\/em> instead. Bunch of guys take ten years to release a remake of\u00a0<em>Half-Life,\u00a0<\/em>now that\u2019s a story.)<\/p>\n<p class=\"western\">Other <em>Doom<\/em>\u00a0mods I\u2019d highly recommend include <a href=\"http:\/\/action.mancubus.net\/actiondoom\/\">Action Doom<\/a> and its sequel <a href=\"http:\/\/action.mancubus.net\/\">Urban Brawl<\/a>, <a href=\"http:\/\/doomwiki.org\/wiki\/Zen_Dynamics\">Zen Dynamics<\/a>, and <a href=\"http:\/\/www.doomworld.com\/idgames\/?file=combos\/zharky_x.zip\">Zharkov Goes to the Store<\/a>. I could easily cover a whole bunch of these, but I only wanted to cover some of my\u00a0favorites. There\u2019s a whole boatload of great Doom mods, this is just a small sampling of them. Hell, I\u2019d recommend checking out the\u00a0<a href=\"http:\/\/www.doomworld.com\/cacowards\/\">Cacowards<\/a>\u00a0at Doomworld, there\u2019s\u00a0a bunch of great mods there too.<\/p>\n<p class=\"western\">This is why <em>Doom<\/em> is one of the best games of all time. You won\u2019t find these absurd mods in a modern shooter, that\u2019s for certain. And if they did, they\u2019d probably sell it to you for $15 a piece instead.<\/p>\n<blockquote><p>Hi there! If you somehow stumbled upon this in a web search, welcome! I have many other cool interesting posts, most of them not really about Doom. Check em out!<\/p>\n<p>I do this as a labor of love, and I don\u2019t get paid for it. If you like this kind of stuff, I have a Ko-Fi that you can donate to. You don\u2019t <em>have\u00a0<\/em>to, I\u2019m plugging it since this is my most popular post on the blog. <a href=\"https:\/\/www.ko-fi.com\/tonicbh\">You can check it out here<\/a>. Thanks!<\/p><\/blockquote>\n<div id=\"atatags-370373-5be31768cf07e\">\n        \n    <\/div>\t\t<div class=\"wpcnt\">\n\t\t\t<div class=\"wpa wpmrec\">\n\t\t\t\tAdvertisements\n\t\t\t\t<div class=\"u\">\t\t<div>\n\t\t<div id=\"atatags-26942-5be31768cf0ab\">\n\t\t\t\n\t\t<\/div><\/div>\t\t<div>\n\t\t<div id=\"atatags-114160-5be31768cf0ae\">\n\t\t\t\n\t\t<\/div><\/div><\/div>\n\t\t\t\t\n\t\t\t<\/div>\n\t\t<\/div>","created_at":"2018-11-07T16:48:41+0000","updated_at":"2018-11-07T16:48:41+0000","published_at":"2014-06-09T03:31:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"youfoundasecret.wordpress.com","preview_picture":"https:\/\/youfoundasecret.files.wordpress.com\/2014\/06\/screenshot_doom_20140608_111901.png?w=640","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":448,"title":"ZDoom \u2022 View topic - High Noon Drifter v1.2","url":"https:\/\/forum.zdoom.org\/viewtopic.php?f=43&t=57098","content":"<div class=\"content\"><iframe title=\"YouTube video player\" width=\"659\" height=\"371\" src=\"https:\/\/www.youtube.com\/embed\/mox_kEhhPuo\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><iframe title=\"YouTube video player\" width=\"659\" height=\"371\" src=\"https:\/\/www.youtube.com\/embed\/2c7nARhF_X8\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>There's been a drifter that's been makin' the rounds. He seems a simple enough fellow. Quiet. Polite. But I saw a peek of him, with a burnt cloak that hid a rock of a man, and a hat shadowing his face. I could see his eyes underneath, and they spooked me--shining with a light that wasn't of this world. He tipped his hat at me and wished me good day, but in that moment I felt like I'd glimpsed at eternity.<br \/>I'd asked around town, and the folk had some tall tales to say about him. Some say that he's a demon, rebelling from his life in Hell and fighting against his former masters. Others say that he's a vampire, living forever off the blood of those he slays. Others say that he's a priest, doin' God's work on these forsaken lands. And yet others say that he's just a luckier and tougher son of a bitch than the rest. But whoever he is, they all say he's not so much driftin' through towns as he is...worlds. Goin' from one reality to the next, leavin' only hot casings and cold bodies.<\/p><p>He won't be here long. He never is. Guess I can't blame him, not really any place on Earth that could fit a man like him. He only stays long enough to head to a bar, talk to some folks, get some food, and grab a bottle of Corzo white for the road. But such is the life of a drifter.<br \/>Who knows where he'll head next? There's some real dark corners in the world out there...<\/p><p>-----------------------------------------------------------------------<\/p><p><a href=\"https:\/\/drive.google.com\/file\/d\/1hbaH0VvLaspBOsjW6faM7D-jhH7LCfiH\/view?usp=sharing\" class=\"postlink\">DOWNLOAD HERE.<\/a> (24.3 mb)<br \/><a href=\"http:\/\/the-powerhouse.net\/HighNoonDrifter-v1.2c.pk3\" class=\"postlink\">Hosting mirror provided by Powerhouse.<\/a><\/p><p><a href=\"https:\/\/drive.google.com\/file\/d\/0B4p2zWkJEFFtMXRBVl9JTG82NHM\/view?usp=sharing\" class=\"postlink\">v1.1b<\/a> (22.4 mb)<br \/><a href=\"http:\/\/the-powerhouse.net\/HighNoonDrifter-v1.1b.pk3\" class=\"postlink\">v1.1b mirror<\/a><\/p><p><a href=\"https:\/\/drive.google.com\/file\/d\/0B4p2zWkJEFFtY0ttNVUzOHI0Z00\/view?usp=sharing\" class=\"postlink\">v1.0<\/a> (19.0 mb)<br \/><a href=\"http:\/\/the-powerhouse.net\/HighNoonDrifter-v1.pk3\" class=\"postlink\">v1.0 mirror<\/a><\/p><p><a href=\"https:\/\/drive.google.com\/file\/d\/0B4p2zWkJEFFtak5pSGkxMmpOcVk\/view?usp=sharing\" class=\"postlink\">Compatibility<\/a> with <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?t=49369\" class=\"postlink\">Strange Aeons<\/a> (Optional patch)<br \/><a href=\"https:\/\/drive.google.com\/file\/d\/11pk4NPAmRwtf87KrkSe_QJkMPgphwSN7\/view\" class=\"postlink\">Compatibility<\/a> with <a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?t=32674\" class=\"postlink\">Pirate Doom<\/a> (Optional patch, by SiFi270)<\/p><p>CHANGELOG:<br \/><\/p><blockquote class=\"Splr\"><div><cite class=\"SplrA\"> Spoiler:<\/cite><\/div><\/blockquote><p>-----------------------------------------------------------------------<\/p><p>High Noon Drifter is a GZDoom weaponset where you play as a gunslinger that's stuck around for a little longer than the end of the wild west era. The weapons and overall tone is relatively low-tech, complimenting more natural-style mapsets for roaming around castles, ruins, and other assorted gothic setpieces. You make use of some ordinary guns that any joe could scrounge up, a handful of antiquated pieces, and a couple dark artifacts in a quest to dispense frontier justice to an ethereal threat.<br \/>If you've been craving a \"Weird West\" sort of weaponset, or even more of an adventurer-type rather than a soldier using military tech, hopefully this'll be up your alley.<\/p><p>High Noon Drifter is also a simple and gimmick-less weaponset, probably the most straightforward mod I've made yet--you pick up the guns, hit the fire button, they go bang and kill the demons.<br \/>A lot of gameplay mods nowadays come with a sort of a big crazy gimmick or a whole bunch of over-the-top weapons with lots of intricate functions. And that ain't a bad thing! But with this, I wanted to re-study the basics of what makes a gun fun to use, sticking with just the basic weapon types and then making them N E A T.<br \/>This also provided me with an opportunity to familiarize myself with the intricacies of ZScript and Anonymous Functions before using them on a certain bigger project, because Lord knows both of those provide a lot of fun and laughs.<\/p><p>FEATURES:<br \/>- Works in Doom, Doom 2, TNT: Evilution, Plutonia, Heretic, and Hexen! Well, and Chex Quest and Harmony.<br \/>- An idiot's first attempts at ZScript, with slowly more integration and usage over time!<br \/>- A complete weaponset with 10 (technically 11) different weapons to play with!<br \/>- A whole bunch of new and original graphics!<br \/>- All Doom\/Doom 2 items and weapons have been replaced!<br \/>- Compatibility with most Boom and ZDoom mapsets!<br \/>- Monsters are un-touched, allowing people to load with their monster packs or gore systems of choice, or for custom monsters in maps to not break horribly!<br \/>- A bunch of cvars to customize your gameplay experience how you want!<br \/>- A unique, never-before-seen-in-Doom \"Hat System\" that keeps track of whether you're wearing your hat or not and adjusts accordingly!<br \/>- Compatible in multiplayer! I think? I mean, I haven't tried it, but some friends tell me it works, so...<br \/>- A dedicated \"crash the game\" button, just hit Mission Objectives!<\/p><p>Screenshots:<br \/><img src=\"http:\/\/i.imgur.com\/qLDqKdyh.png\" alt=\"image\" \/><\/p><p><img src=\"http:\/\/i.imgur.com\/0VT8Ycgh.png\" alt=\"image\" \/><\/p><p><img src=\"http:\/\/i.imgur.com\/4o1H609h.png\" alt=\"image\" \/><\/p><p><img src=\"http:\/\/i.imgur.com\/e5J4dTBh.png\" alt=\"image\" \/><\/p><p><img src=\"http:\/\/i.imgur.com\/RQ7M624h.png\" alt=\"image\" \/><\/p><p><img src=\"http:\/\/i.imgur.com\/pm9QmyUh.png\" alt=\"image\" \/><\/p><p><img src=\"http:\/\/i.imgur.com\/qcDnrWGh.png\" alt=\"image\" \/><\/p><p>CREDITS:<br \/><\/p><blockquote class=\"Splr\"><div><cite class=\"SplrA\"> Spoiler:<\/cite><\/div><\/blockquote><p>WHAT SHOULD I PLAY THIS WITH?:<br \/><\/p><blockquote class=\"Splr\"><div><cite class=\"SplrA\"> Spoiler:<\/cite><\/div><\/blockquote><\/div><p>Last edited by <a href=\"https:\/\/forum.zdoom.org\/memberlist.php?mode=viewprofile&amp;u=4255&amp;sid=426e47299f310173cd32390f99945f0e\">TerminusEst13<\/a> on Mon Jul 02, 2018 7:43 pm, edited 27 times in total.<\/p>","created_at":"2018-11-07T16:48:41+0000","updated_at":"2018-11-07T16:48:41+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"forum.zdoom.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":446,"title":"Why I Love: Doom WADs","url":"https:\/\/www.pcgamer.com\/why-i-love-doom-wads\/","content":"<p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-320-80.png\" alt=\"The Adventures of Square\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600.png\" \/><\/p><div class=\"fancy-box\"><h5>Why I Love<\/h5><div class=\"fancy_box_body\"><p>In <a href=\"https:\/\/www.pcgamer.com\/why-i-love\/\">Why I Love<\/a>, PC Gamer writers pick an aspect of PC gaming that they love and write about why it's brilliant. Today, Phil explains why the Doom modding community is still the best.<\/p><\/div><\/div><p>What's you favourite game for mods? Maybe it's Skyrim, which offers robust fixes and improvements, new quests and items, and a safe haven for perverts. Maybe it's Half-Life 2, which provides such flexibility and variety via the Source SDK that a myriad different tones, styles and genres are available. If you're anything like me\u2014or, if you're anything like me in regard to a very specific appreciation of an aspect of PC gaming\u2014it's Doom.<\/p><p>If you're unaware of the scope and scale of Doom modding, you're probably thinking about <a href=\"http:\/\/www.moddb.com\/mods\/brutal-doom\" target=\"_blank\">Brutal Doom<\/a> right now. That's totally understandable. With its ultra-absurd violence, it makes Id's 21-year-old shooter (or Id's 20-year-old shooter if you're playing it with Doom 2) feel over-the-top and subversive again. But there's more to Doom modding than blood dripping from the ceilings. The community has been active for all of its lifespan, and they're still going strong.<\/p><p>Want a small sampler of what's out there? <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?t=32674\" target=\"_blank\">Pirate Doom<\/a> is like a cross between The Secret of Monkey Island and, well, Doom. It's brilliant: jolly music clashing against the disintegrating corpse of a flintlock'd pirate.<\/p><p>For something more Doom-y, there's <a href=\"http:\/\/doomwiki.org\/wiki\/Doom_the_Way_id_Did\" target=\"_blank\">Doom The Way Id Did<\/a>; a collaborative community compilation. The idea was to create new maps that mimicked the style and design philosophies of the original game's creators. It produced an instant familiarity, but one that was nestled inside something new.<\/p><p>Maybe you need something more modern for your retro FPS. <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=25836\" target=\"_blank\">Project MSX<\/a> 'upgrades' Doom to feature modern weapons, NPC barks and, of course, regenerating shields. It's a joke made playable\u2014a satirical bullet fired from the gun of the game that helped first popularise the genre.<\/p><figure><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7-320-80.png\" alt=\"Pirate Doom. It's Doom, but with pirates.\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7.png\" \/><\/p><figcaption>Pirate Doom. It's Doom, but with pirates.<\/figcaption><\/figure><p>There are thousands of these. A two-decade strong community has kept making and improving their output\u2014changing the game in every way imaginable, and then refining on that process. It's not like they had little to work with to begin with. Doom is still an incredible game. I'd argue it holds up better than most early-'90s shooters. It's fast, but not unbearably so, with a precision to its movement that helps you to navigate the labyrinthine layouts of rooms and corridors.<\/p><p>It's weapons are great, too. Sure, they're not fetishistically detailed\u2014every nut, bolt and cranny meticulously recreated. There's a more primal enjoyment than that. Put simply: they're a hoot to shoot. Each feels direct and snappy in a digital way that rejects realism. Point a gun in a man's direction and you will probably hit that man. It's a game that doesn't want you to worry about the small details. The action isn't a precise technical challenge, as much as its the lubricant that greases the exploration and discovery.<\/p><p>That's an extremely strong base from which to base a mod scene on, and it's little surprise the community has been as enduring as it has.<\/p><p>If you want a direct route to the best Doom maps and mods, the place to go is the <a href=\"http:\/\/www.doomworld.com\/cacowards\/\" target=\"_blank\">Cacowards<\/a>. Every year, this Doom World event recognises the ten best mods, as well as a variety of other, special categories. Through it, you can discover gems like <a href=\"http:\/\/www.doomworld.com\/vb\/wads-mods\/70749-wip-demonsteele\/\" target=\"_blank\">DemonSteele<\/a>\u2014a frantic weapons mod that gives you a sword, a health system and a lot more enemies.<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd-320-80.png\" alt=\"DemonSteele\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd.png\" \/><\/p><p>The Cacowards are a playlist of the finest works from the last 12 months, covering a variety of looks and styles. <a href=\"http:\/\/adventuresofsquare.com\/index.php\" target=\"_blank\">The Adventures of Square<\/a>, for instance, isn't recognisably Doom. It's the Adult Swim version of Doom, in which a purple rectangle grabs his trusty paint gun and sets out to put an end to authoritarian circles. What makes TAoS great, though, isn't just the way it changes Doom's look, but also how Doom plays. It's as much a platformer as a shooter\u2014filled with wide gaps and jumping puzzles.<\/p><p>Other mods are less visually striking, and more technically amazing. <a href=\"http:\/\/www.doomworld.com\/idgames\/?file=levels\/doom2\/Ports\/megawads\/gd.zip\" target=\"_blank\">Going Down<\/a> is a 32 map megawad\u2014an epic achievement in level design that packs in more variety than should be possible from a single high-rise building. These aren't just good Doom levels, they're good game levels. They toy with and tease you in a way that would be unthinkable for anyone who hasn't been following what's possible in these games.<\/p><p>I'll leave you with the best, most surprising, most consistently enjoyable experience I had with Doom last year\u2014direct from the Cacowards 'Play Thy Role' award. It's in recognition of the perfect mod combination. This year, it went to combined might of <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=33292\" target=\"_blank\">Doom RPG Mod<\/a> and <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=37044\" target=\"_blank\">DoomRL Arsenal<\/a>. One brings stats, levelling and currency to the game, while the other adds new weapons, items and player classes. Together they combine into an absurd overhaul, and a great showcase of why the Doom modding community still has plenty of life left in it.<\/p><p><em>To play these mods, you'll need either Doom or Doom 2 (or both), which are available through Steam. You'll also need GZDoom. Got both of those?<\/em><em> <\/em><a href=\"http:\/\/doom.wikia.com\/wiki\/How_to_play_or_start_wads\" target=\"_blank\"><em>Follow these instructions<\/em><\/a><em>.<\/em><\/p>","created_at":"2018-11-07T16:48:40+0000","updated_at":"2018-11-07T16:48:40+0000","published_by":["Phil Savage"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.pcgamer.com","preview_picture":"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-1200-80.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":445,"title":"Why I Love: Doom WADs","url":"https:\/\/www.pcgamer.com\/why-i-love-doom-wads\/","content":"<p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-320-80.png\" alt=\"The Adventures of Square\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600.png\" \/><\/p><div class=\"fancy-box\"><h5>Why I Love<\/h5><div class=\"fancy_box_body\"><p>In <a href=\"https:\/\/www.pcgamer.com\/why-i-love\/\">Why I Love<\/a>, PC Gamer writers pick an aspect of PC gaming that they love and write about why it's brilliant. Today, Phil explains why the Doom modding community is still the best.<\/p><\/div><\/div><p>What's you favourite game for mods? Maybe it's Skyrim, which offers robust fixes and improvements, new quests and items, and a safe haven for perverts. Maybe it's Half-Life 2, which provides such flexibility and variety via the Source SDK that a myriad different tones, styles and genres are available. If you're anything like me\u2014or, if you're anything like me in regard to a very specific appreciation of an aspect of PC gaming\u2014it's Doom.<\/p><p>If you're unaware of the scope and scale of Doom modding, you're probably thinking about <a href=\"http:\/\/www.moddb.com\/mods\/brutal-doom\" target=\"_blank\">Brutal Doom<\/a> right now. That's totally understandable. With its ultra-absurd violence, it makes Id's 21-year-old shooter (or Id's 20-year-old shooter if you're playing it with Doom 2) feel over-the-top and subversive again. But there's more to Doom modding than blood dripping from the ceilings. The community has been active for all of its lifespan, and they're still going strong.<\/p><p>Want a small sampler of what's out there? <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?t=32674\" target=\"_blank\">Pirate Doom<\/a> is like a cross between The Secret of Monkey Island and, well, Doom. It's brilliant: jolly music clashing against the disintegrating corpse of a flintlock'd pirate.<\/p><p>For something more Doom-y, there's <a href=\"http:\/\/doomwiki.org\/wiki\/Doom_the_Way_id_Did\" target=\"_blank\">Doom The Way Id Did<\/a>; a collaborative community compilation. The idea was to create new maps that mimicked the style and design philosophies of the original game's creators. It produced an instant familiarity, but one that was nestled inside something new.<\/p><p>Maybe you need something more modern for your retro FPS. <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=25836\" target=\"_blank\">Project MSX<\/a> 'upgrades' Doom to feature modern weapons, NPC barks and, of course, regenerating shields. It's a joke made playable\u2014a satirical bullet fired from the gun of the game that helped first popularise the genre.<\/p><figure><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7-320-80.png\" alt=\"Pirate Doom. It's Doom, but with pirates.\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/7d202f288ced12d69253ef2079cb95b7.png\" \/><\/p><figcaption>Pirate Doom. It's Doom, but with pirates.<\/figcaption><\/figure><p>There are thousands of these. A two-decade strong community has kept making and improving their output\u2014changing the game in every way imaginable, and then refining on that process. It's not like they had little to work with to begin with. Doom is still an incredible game. I'd argue it holds up better than most early-'90s shooters. It's fast, but not unbearably so, with a precision to its movement that helps you to navigate the labyrinthine layouts of rooms and corridors.<\/p><p>It's weapons are great, too. Sure, they're not fetishistically detailed\u2014every nut, bolt and cranny meticulously recreated. There's a more primal enjoyment than that. Put simply: they're a hoot to shoot. Each feels direct and snappy in a digital way that rejects realism. Point a gun in a man's direction and you will probably hit that man. It's a game that doesn't want you to worry about the small details. The action isn't a precise technical challenge, as much as its the lubricant that greases the exploration and discovery.<\/p><p>That's an extremely strong base from which to base a mod scene on, and it's little surprise the community has been as enduring as it has.<\/p><p>If you want a direct route to the best Doom maps and mods, the place to go is the <a href=\"http:\/\/www.doomworld.com\/cacowards\/\" target=\"_blank\">Cacowards<\/a>. Every year, this Doom World event recognises the ten best mods, as well as a variety of other, special categories. Through it, you can discover gems like <a href=\"http:\/\/www.doomworld.com\/vb\/wads-mods\/70749-wip-demonsteele\/\" target=\"_blank\">DemonSteele<\/a>\u2014a frantic weapons mod that gives you a sword, a health system and a lot more enemies.<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd-320-80.png\" alt=\"DemonSteele\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/64996f9cc1f5c3b670664145a828cebd.png\" \/><\/p><p>The Cacowards are a playlist of the finest works from the last 12 months, covering a variety of looks and styles. <a href=\"http:\/\/adventuresofsquare.com\/index.php\" target=\"_blank\">The Adventures of Square<\/a>, for instance, isn't recognisably Doom. It's the Adult Swim version of Doom, in which a purple rectangle grabs his trusty paint gun and sets out to put an end to authoritarian circles. What makes TAoS great, though, isn't just the way it changes Doom's look, but also how Doom plays. It's as much a platformer as a shooter\u2014filled with wide gaps and jumping puzzles.<\/p><p>Other mods are less visually striking, and more technically amazing. <a href=\"http:\/\/www.doomworld.com\/idgames\/?file=levels\/doom2\/Ports\/megawads\/gd.zip\" target=\"_blank\">Going Down<\/a> is a 32 map megawad\u2014an epic achievement in level design that packs in more variety than should be possible from a single high-rise building. These aren't just good Doom levels, they're good game levels. They toy with and tease you in a way that would be unthinkable for anyone who hasn't been following what's possible in these games.<\/p><p>I'll leave you with the best, most surprising, most consistently enjoyable experience I had with Doom last year\u2014direct from the Cacowards 'Play Thy Role' award. It's in recognition of the perfect mod combination. This year, it went to combined might of <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=33292\" target=\"_blank\">Doom RPG Mod<\/a> and <a href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=37044\" target=\"_blank\">DoomRL Arsenal<\/a>. One brings stats, levelling and currency to the game, while the other adds new weapons, items and player classes. Together they combine into an absurd overhaul, and a great showcase of why the Doom modding community still has plenty of life left in it.<\/p><p><em>To play these mods, you'll need either Doom or Doom 2 (or both), which are available through Steam. You'll also need GZDoom. Got both of those?<\/em><em> <\/em><a href=\"http:\/\/doom.wikia.com\/wiki\/How_to_play_or_start_wads\" target=\"_blank\"><em>Follow these instructions<\/em><\/a><em>.<\/em><\/p>","created_at":"2018-11-07T16:48:39+0000","updated_at":"2018-11-07T16:48:39+0000","published_by":["Phil Savage"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.pcgamer.com","preview_picture":"https:\/\/cdn.mos.cms.futurecdn.net\/21fbbe3438c988ddea26749ac5723600-1200-80.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":443,"title":"Strange Aeons","url":"https:\/\/mikestoybox.net\/2017\/01\/04\/strange-aeons\/","content":"<p>Note: Recommended resolution s 800\u00d7600 or better.<\/p>\n<p><a href=\"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png\"><img data-attachment-id=\"1283\" data-permalink=\"https:\/\/mikestoybox.net\/doom-maps\/strangeaeons-shot\/\" data-orig-file=\"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png\" data-orig-size=\"800,503\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"strangeaeons shot\" data-image-description=\"\" data-medium-file=\"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png?w=300&amp;h=189\" data-large-file=\"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png?w=627\" class=\"aligncenter size-medium wp-image-1283\" src=\"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png?w=300&amp;h=189\" alt=\"strangeaeons shot\" width=\"300\" height=\"189\" srcset=\"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png?w=300&amp;h=189 300w, https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png?w=600&amp;h=378 600w, https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png?w=150&amp;h=94 150w\" \/><\/a><\/p>\n<p><strong>STRANGE AEONS<\/strong><\/p>\n<p><strong><a href=\"https:\/\/www.impie.drdteam.org\/SA%20Download.zip\">DOWNLOAD IT!<\/a><\/strong><\/p>\n<p>v5.7<\/p>\n<p>5-Episode Partial Conversion for Doom<\/p>\n<p>Ultimate Doom Recommended; Doom 2 Compatible<\/p>\n<p>(Requires ZDoom or GZdoom to run)<\/p>\n<p><a href=\"http:\/\/www.doomworld.com\/23years\/index.php\">Runner-up for the 13th Annual Cacowards!<\/a><\/p>\n<p>This is an atmospheric map set (45 levels) based on the works of H P Lovecraft, and inspired by a Doom clone I designed when I was a kid.\u00a0 Reunite with your dead son by traveling through your dreams and conquering four nightmarish realms: a shattered planet floating in aether, a dilapitated fortress of the future, the badlands of prehistoric earth, and the frozen Plateau of Leng. Now includes a fifth bonus episode, Out of the Aeons, set in the Underworld and the City of R\u2019lyeh!<\/p>\n<p><strong>GUNS N\u2019 MONSTERS<\/strong><\/p>\n<p><strong><a href=\"https:\/\/www.impie.drdteam.org\/strangeaeonsaddon.pk3\">DOWNLOAD IT!<\/a><\/strong><\/p>\n<p>A mod that allows you to use SA\u2019s guns and monsters in other map sets.<\/p>\n<p><strong>RYLAYEH MUSIC PATCH<\/strong><\/p>\n<p><strong><a href=\"https:\/\/www.impie.drdteam.org\/rylayeh%20music%20patch.wad\">DOWNLOAD IT!<\/a><\/strong><\/p>\n<p>Combine Guns N\u2019 Monsters with this music patch and the excellent <a href=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh\">Rylayeh wad<\/a>. It now plays like an unofficial sixth episode!<\/p>\n<p><iframe class=\"youtube-player\" width=\"627\" height=\"353\" src=\"https:\/\/www.youtube.com\/embed\/2WLzXIVzQcw?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>\n<p><iframe class=\"youtube-player\" width=\"627\" height=\"353\" src=\"https:\/\/www.youtube.com\/embed\/U0Pvhee9Ue4?version=3&amp;rel=1&amp;fs=1&amp;autohide=2&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>","created_at":"2018-11-07T16:48:38+0000","updated_at":"2018-11-07T16:48:38+0000","published_at":"2017-01-04T07:44:40+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"mikestoybox.net","preview_picture":"https:\/\/deadendsolutions.files.wordpress.com\/2014\/06\/strangeaeons-shot.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":444,"title":"Rocky Jaiswal | Get started with Compojure & ClojureScript","url":"https:\/\/rockyj.in\/2015\/11\/01\/clojurescript_compojure.html","content":"<p>Clojure is probably the programming language that has fascinated and intimidated me for the longest time. I tried to pick it up many times and gave up an equal number of times. However, it was getting harder and harder to ignore ClojureScript's rising popularity and its awesome integration with React using <a href=\"https:\/\/github.com\/omcljs\/om\">Om<\/a> \/ <a href=\"http:\/\/reagent-project.github.io\/\">Reagent<\/a>. So I tried once again and instead of doing a lot over a single weekend, I tried to take it one step at a time. Some resources that I highly recommend for Clojure beginners are -<\/p><ul><li><a href=\"http:\/\/www.braveclojure.com\/\">http:\/\/www.braveclojure.com\/<\/a><\/li>\n              <li><a href=\"http:\/\/4clojure.com\">http:\/\/4clojure.com<\/a><\/li>\n              <li><a href=\"https:\/\/github.com\/functional-koans\/clojure-koans\">https:\/\/github.com\/functional-koans\/clojure-koans<\/a><\/li>\n              <li><a href=\"https:\/\/github.com\/gigasquid\/wonderland-clojure-katas\">https:\/\/github.com\/gigasquid\/wonderland-clojure-katas<\/a><\/li>\n              <li><a href=\"http:\/\/yogthos.github.io\/ClojureDistilled.html\">http:\/\/yogthos.github.io\/ClojureDistilled.html<\/a><\/li>\n            <\/ul><p>Although I am just a noob right now and a little knowledge can be a dangerous thing, this post might help people like me get started with ClojureScript. First, we will create a simple project using <a href=\"http:\/\/leiningen.org\/\">lein<\/a>. To get the setup and dependencies right, our <strong>project.clj<\/strong> may look like -<\/p><pre>(defproject simple-app \"0.1.0-SNAPSHOT\"\n  :description \"a simple app with ring+compojure+cljs\"\n  :url \"http:\/\/rockyj.in\"\n  :min-lein-version \"2.0.0\"\n  :source-paths [\"src\/clj\"]\n  :dependencies [[org.clojure\/clojure \"1.7.0\"]\n                 [compojure \"1.4.0\"]\n                 [ring\/ring-defaults \"0.1.5\"]\n                 [ring\/ring-json \"0.4.0\"]\n                 ;;ClojureScript\n                 [org.clojure\/clojurescript \"1.7.122\"]\n                 [cljs-ajax \"0.5.1\"]\n                 [prismatic\/dommy \"1.1.0\"]\n                 ]\n  :plugins [[lein-ring \"0.8.13\"]\n            [lein-figwheel \"0.4.1\"]\n            ]\n  :ring {:handler simple-app.handler\/app}\n  :cljsbuild {\n              :builds [ { :id \"simple-app\"\n                         :source-paths [\"src\/cljs\"]\n                         :figwheel true\n                         :compiler {:main \"simple-app.app\"\n                                    :asset-path \"js\/out\"\n                                    :output-to \"resources\/public\/js\/app.js\"\n                                    :output-dir \"resources\/public\/js\/out\"} } ]\n              }\n  :profiles\n  {:dev {:dependencies [[javax.servlet\/servlet-api \"2.5\"]\n                        [ring\/ring-mock \"0.3.0\"]\n                        ]}})\n<\/pre><p>This will setup for us -<\/p><ul><li><a href=\"https:\/\/github.com\/weavejester\/compojure\">Compojure<\/a><\/li>\n              <li>Ring with JSON support<\/li>\n              <li>ClojureScript<\/li>\n              <li><a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\">Figwheel<\/a><\/li>\n              <li>The ability to build our ClojureScript code into JavaScript<\/li>\n              <li>And other ClojureScript libraries which we will discuss later<\/li>\n            <\/ul><p>For this simple application -<\/p><ul><li>We will serve a simple static HTML5 page using Compojure.<\/li>\n              <li>We will also have a route than can take in a JSON param and send back a JSON response.<\/li>\n              <li>On the client side we will make a XHR call to this endpoint and manipulate some DOM with the results.<\/li>\n            <\/ul><p>This is all very simple but good enough to get us started.<\/p><p>On the server side the code looks like <strong>handler.clj<\/strong> -<\/p><pre>(ns simple-app.handler\n  (:require [compojure.core :refer :all]\n            [compojure.route :as route]\n            [ring.middleware.json :as middleware]\n            [ring.middleware.keyword-params :refer [wrap-keyword-params]]\n            [ring.util.response :as response]\n            ))\n(defroutes app-routes\n  (GET \"\/\" [] (response\/resource-response \"index.html\" {:root \"public\"}))\n  (route\/resources \"\/\")\n  (POST \"\/hello\" request\n        (response\/response {:greeting\n                            (get-in request [:params :user])}))\n  (route\/not-found \"Not Found\"))\n(def app\n  (-&gt; app-routes\n      wrap-keyword-params\n      middleware\/wrap-json-params\n      middleware\/wrap-json-response))\n<\/pre><p><strong>index.html<\/strong>:<\/p><pre>&lt;!DOCTYPE html&gt;\n&lt;html lang=\"en\"&gt;\n  &lt;head&gt;\n    &lt;meta charset=\"utf-8\"&gt;\n    &lt;meta http-equiv=\"X-UA-Compatible\" content=\"IE=edge\"&gt;\n    &lt;meta name=\"viewport\" content=\"width=device-width, initial-scale=1\"&gt;\n    &lt;title&gt;Welcome to Clojure!&lt;\/title&gt;\n  &lt;\/head&gt;\n  &lt;body&gt;\n    &lt;h1&gt;Hello World!&lt;\/h1&gt;\n    &lt;h1 class=\"fromServer\"&gt;&lt;\/h1&gt;\n    &lt;script src=\"js\/app.js\"&gt;&lt;\/script&gt;\n  &lt;\/body&gt;\n&lt;\/html&gt;\n<\/pre><p>On the client side, <strong>app.cljs<\/strong> -<\/p><pre>(ns simple-app.app\n  (:require [simple-app.greet :as greet]))\n(greet\/say-hello)\n(.log js\/console \"Hello Cljs!\")\n<\/pre><p>and <strong>greet.cljs<\/strong> -<\/p><pre>(ns simple-app.greet\n  (:require [ajax.core :refer [GET POST]]\n            [dommy.core :as dommy :refer-macros [sel sel1]]))\n(defn handler [response]\n  (.log js\/console \"server responded...\")\n  (-&gt; (sel1 :.fromServer)\n      (dommy\/set-text!\n       (str \"Hello \" (aget (clj-&gt;js response) \"greeting\")))))\n(defn error-handler [{:keys [status status-text]}]\n  (.log js\/console (str \"something bad happened: \" status \" \" status-text)))\n(defn say-hello []\n  (POST \"\/hello\"\n        {:params {:user \"from Server!\"}\n         :handler handler\n         :error-handler error-handler\n         :format :json\n         :response-format :json}))\n<\/pre><p>Two of the most common things we do in JavaScript code are -<\/p><ul><li>DOM manipulation<\/li>\n              <li>XHR handling<\/li>\n            <\/ul><p>As we see with the code above, this is made easy in ClojureScript with <a href=\"https:\/\/github.com\/Prismatic\/dommy\">dommy<\/a> and <a href=\"https:\/\/github.com\/JulianBirch\/cljs-ajax\">cljs-ajax<\/a>.<\/p><p>The project layout looks like -<\/p><pre>.\n\u251c\u2500\u2500 dev-resources\n\u251c\u2500\u2500 project.clj\n\u251c\u2500\u2500 README.md\n\u251c\u2500\u2500 resources\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 public\n\u2502\u00a0\u00a0     \u251c\u2500\u2500 index.html\n\u2502\u00a0\u00a0     \u2514\u2500\u2500 js\n\u2502\u00a0\u00a0         \u2514\u2500\u2500 app.js\n\u251c\u2500\u2500 src\n\u2502\u00a0\u00a0 \u251c\u2500\u2500 clj\n\u2502\u00a0\u00a0 \u2502\u00a0\u00a0 \u2514\u2500\u2500 simple_app\n\u2502\u00a0\u00a0 \u2502\u00a0\u00a0     \u2514\u2500\u2500 handler.clj\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 cljs\n\u2502\u00a0\u00a0     \u2514\u2500\u2500 simple_app\n\u2502\u00a0\u00a0         \u251c\u2500\u2500 app.cljs\n\u2502\u00a0\u00a0         \u2514\u2500\u2500 greet.cljs\n\u2514\u2500\u2500 test\n    \u2514\u2500\u2500 simple_app\n        \u2514\u2500\u2500 handler_test.clj\n<\/pre><p>Finally, the commands - <strong>lein ring server<\/strong> and <strong>lein figwheel<\/strong> can be run to get the application working.<\/p><p>Even in this small demo application we can see a few advantages of this stack -<\/p><ol><li>We programmed in a single language using the same principles of code organization for the client and server side.<\/li>\n              <li>We managed all external dependencies of the project in a single file.<\/li>\n              <li>The build process is simple and clean.<\/li>\n              <li>We have a REPL for both client side and server side development.<\/li>\n            <\/ol><p>So hope this post will help you to get started with ClojureScript, in the future I will try and look at Om \/ Reagent.<\/p>","created_at":"2018-11-07T16:48:38+0000","updated_at":"2018-11-07T16:48:38+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"rockyj.in","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":440,"title":"Eternal DOOM IV: Return from Oblivion","url":"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1","content":"<div class=\"dir\"><img src=\"https:\/\/www.doomworld.com\/idgames\/icons\/back.gif\" alt=\"\" \/><a href=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/\">Back to Directory<\/a><\/div>\n<br \/><table style=\"margin: auto; width: 720px;\"><tr><td valign=\"top\" style=\"width: 550px;\">\n<table class=\"filelist\"><tr><td class=\"filelist_label\">Title:<\/td><td class=\"filelist_field1\">\nEternal DOOM IV: Return from Oblivion<\/td><\/tr><tr><td class=\"filelist_label\">Filename:<\/td><td class=\"filelist_field2\">\nthemes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip<\/td><\/tr><tr><td class=\"filelist_label\">Size:<\/td><td class=\"filelist_field1\">\n15.43 MB<\/td><\/tr><tr><td class=\"filelist_label\">Date:<\/td><td class=\"filelist_field2\">\n01\/11\/08<\/td><\/tr><tr><td class=\"filelist_label\">Author:<\/td><td class=\"filelist_field1\">\nTeamTNT<\/td><\/tr><tr><td class=\"filelist_label\">Description:<\/td><td class=\"filelist_field2\">\nAll maps are: (i) medieval-themed, and generally very large compared to typical DooM maps; (ii) non-linear and feature multiple paths; (iii) feature castle and fortress architecture, fights on an epic scale, and magnificent vistas. Maps 08-10 can be played in any order within the hub, and powerups &amp; enemies are different in any given map depending on whether you play it first or play it after entering from another map. This provides additional replay value.<p>This is a 7-map demo. TeamTNT intends to create a full megawad next.<\/p><p>Tested with ZDooM and GZDooM<\/p><\/td><\/tr><tr><td class=\"filelist_label\">Credits:<\/td><td class=\"filelist_field1\">\nBjorn 'Vader' Ostmann for thoroughly play-testing all the maps and for providing excellent suggestions Scott (Birdman) Bardy Paul Fleschute Daniel 'Tormentor' Gimmer Gary Gosson Gothic Team Dale 'Cadman' Harris Don Howard Sverre Kvernmo Rich Nagel Russell Pearson Paul Schmitz Dia Westerteicher<p>The authors of the ZDooM Monster Resource Wad Randy Heit for ZDooM id Software<\/p><\/td><\/tr><tr><td class=\"filelist_label\">Base:<\/td><td class=\"filelist_field2\">\nMaps08-10 &amp; 27 are from scratch Map 11 is based on a map in TeamTNT's Eternal DooM III Maps 12 and 13 have a few reused and redone areas from TeamTNT's Eternal DooM III. Music is from TeamTNT's Eternal DooM III The Monster Resource Wad (ZDooM community effort) Textures are from TeamTNT's Eternal DooM III; additional textures by Chris Couleur and Esa Repo Sounds are from TeamTNT's Eternal DooM III; additional sounds by Esa Repo<\/td><\/tr><tr><td class=\"filelist_label\">Build time:<\/td><td class=\"filelist_field1\">\n3 years 9 months<\/td><\/tr><tr><td class=\"filelist_label\">Editor(s) used:<\/td><td class=\"filelist_field2\">\nWadAuthor and Doom Builder for maps, XWE for resource and wad management, DooM Builder to merge maps into resource wad<\/td><\/tr><tr><td class=\"filelist_label\">Bugs:<\/td><td class=\"filelist_field1\">\nNone<\/td><\/tr><tr><td class=\"filelist_label\">Rating:<\/td><td class=\"filelist_field2\">\n<img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/emptystar.gif\" alt=\"\" \/> (124 votes)\n<\/td><\/tr><\/table><\/td>\n<td valign=\"top\" align=\"right\" style=\"width: 170px;\">\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/download.png\" width=\"160\" height=\"60\" alt=\"Download here\" title=\"IF ANYONE ASKS HOW TO DOWNLOAD FILES ANYMORE I WILL MURDER THEM\" \/><\/div>\n<br \/><table class=\"download\"><tr><td valign=\"top\">\n<strong>Download mirrors:<\/strong>\n<ul class=\"square\"><li><a href=\"ftp:\/\/ftp.fu-berlin.de\/pc\/games\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">Germany<\/a> <img title=\"Primary download mirror\" src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"ftp:\/\/mirrors.syringanetworks.net\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">Idaho<\/a> <img title=\"Primary download mirror\" src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"ftp:\/\/ftp.ntua.gr\/pub\/vendors\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">Greece<\/a> (<a href=\"http:\/\/ftp.ntua.gr\/pub\/vendors\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">HTTP<\/a>)<br \/><\/li><li><a href=\"http:\/\/ftp.mancubus.net\/pub\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">Texas<\/a><br \/><\/li><li><a href=\"https:\/\/www.quaddicted.com\/files\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">Germany (TLS)<\/a><br \/><\/li><li><a href=\"http:\/\/youfailit.net\/pub\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">New York<\/a> <img title=\"Does not mirror \/themes\/x-rated\" src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"http:\/\/www.gamers.org\/pub\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.zip\">Virginia<\/a>\n<\/li><\/ul><strong>\/idgames protocol:<\/strong>\n<ul class=\"idgamesprotocol\"><li><a href=\"denied:idgames:\/\/15193\">idgames:\/\/15193<\/a><\/li>\n<\/ul><\/td>\n<\/tr><\/table><br \/><\/td>\n<\/tr><\/table><br \/><div style=\"text-align: center;\">\n<div style=\"text-align: center;\"><a target=\"_new\" href=\"http:\/\/www.gamers.org\/pub\/idgames\/themes\/TeamTNT\/ed4_rfo\/ed4_rfo1.txt\">View ed4_rfo1.txt<\/a><\/div>\n<textarea rows=\"20\" cols=\"88\">\n===========================================================================&#13;\nAdvanced engine needed  : ZDooM v2.1.7 or later&#13;\nPrimary purpose         : Single + coop play&#13;\n===========================================================================&#13;\nTitle                   : Eternal DOOM IV: Return from Oblivion&#13;\nFilename                : ed4_rfo1.wad&#13;\nOther Files Included    : ed4_rfo1.txt&#13;\n                          coop.txt&#13;\n                          hubnotes.txt&#13;\nRelease date            : January 11, 2008&#13;\nAuthor                  : TeamTNT&#13;\nEmail Address           : rfo@teamtnt.com&#13;\nMisc. Author Info       : Ty Halderman: TeamTNT head honcho, webmaster, and&#13;\n                            resource manager&#13;\n                          Chris Couleur: Maps 11-13, project manager,&#13;\n                            textures, play-testing&#13;\n                          Esa 'Espi' Repo: Map 27, textures, sounds,&#13;\n                            play-testing&#13;\n                          Rex Claussen: Maps 08-10, resource management,&#13;\n                            project management support, play-testing&#13;\n&#13;\nDescription             : All maps are: &#13;\n                          (i) medieval-themed, and generally very large &#13;\n                            compared to typical DooM maps; &#13;\n                          (ii) non-linear and feature multiple paths; &#13;\n                          (iii) feature castle and fortress architecture, &#13;\n                            fights on an epic scale, and magnificent vistas. &#13;\n                          Maps 08-10 can be played in any order within the &#13;\n                            hub, and powerups &amp; enemies are different in any&#13;\n                            given map depending on whether you play it first&#13;\n                            or play it after entering from another map. This &#13;\n                            provides additional replay value.&#13;\n                          &#13;\n                          This is a 7-map demo. TeamTNT intends to create&#13;\n                          a full megawad next.&#13;\n                          &#13;\n                          Tested with ZDooM and GZDooM&#13;\n&#13;\nIMPORTANT INFORMATION   : Maps 08-10 constitute a hub, and MUST be started &#13;\n                            from Map 08. &#13;\n&#13;\n                          To ensure proper gameplay do not use the IDCLEV&#13;\n                            cheat code to warp to any of the maps in the hub. &#13;\n                            (i.e. *DO NOT* begin the game by warping to Map 09&#13;\n                            or 10). Doing so will mess up the scripts and will &#13;\n                            leave you stuck in the map to which you warp. [You &#13;\n                            can warp to a map just to look around if you wish]&#13;\n&#13;\n                          To ensure proper gameplay do not use the IDCLIP &#13;\n                            cheat code to leave the intro area or the hub area&#13;\n                            in Map 08. Doing so will mess up the scripts and &#13;\n                            will leave you stuck in the map.&#13;\n&#13;\n                          To ensure proper gameplay do not use the IDKFA cheat &#13;\n                            code to get weapons in the hub. Doing so could mess &#13;\n                            up the scripts and will leave youstuck in the map &#13;\n                            in which you used the cheat code.  [This is &#13;\n                            particularly true of the chainsaw, as once you've &#13;\n                            gotten it through the cheat code you cannot pick it&#13;\n                            up in the map. This could prevent you from being&#13;\n                            able to leave the map.]&#13;\n&#13;\n                          For co-operative play read the text file coop.txt &#13;\n                            before playing.&#13;\n&#13;\nAdditional Credits to   : Bjorn 'Vader' Ostmann for thoroughly play-testing all&#13;\n                            the maps and for providing excellent suggestions&#13;\n                          Scott (Birdman) Bardy&#13;\n                          Paul Fleschute&#13;\n                          Daniel 'Tormentor' Gimmer&#13;\n                          Gary Gosson&#13;\n                          Gothic Team&#13;\n                          Dale 'Cadman' Harris&#13;\n                          Don Howard&#13;\n                          Sverre Kvernmo&#13;\n                          Rich Nagel&#13;\n                          Russell Pearson&#13;\n                          Paul Schmitz&#13;\n                          Dia Westerteicher&#13;\n                          &#13;\n                          The authors of the ZDooM Monster Resource Wad&#13;\n                          Randy Heit for ZDooM&#13;\n                          id Software&#13;\n===========================================================================&#13;\n* What is included *&#13;\n&#13;\nNew levels              : 7&#13;\nSounds                  : Yes&#13;\nMusic                   : Yes&#13;\nGraphics                : Yes&#13;\nDehacked\/BEX Patch      : No&#13;\nDemos                   : No&#13;\nOther                   : MAPINFO, Scripts&#13;\nOther files required    : None&#13;\n&#13;\n&#13;\n* Play Information *&#13;\n&#13;\nGame                    : Doom II v1.9 &amp; ZDooM v2.1.7 or higher&#13;\nMap #                   : Map08-13, Map27&#13;\nSingle Player           : Designed for&#13;\nCooperative 2-4 Player  : Designed for&#13;\nDeathmatch 2-4 Player   : No&#13;\nOther game styles       :  (co-op starts only in Maps 11-13, but not&#13;\n                          tested)&#13;\nDifficulty Settings     : Yes&#13;\n&#13;\n&#13;\n* Construction *&#13;\n&#13;\nBase                    : Maps08-10 &amp; 27 are from scratch&#13;\n                          Map 11 is based on a map in TeamTNT's &#13;\n                            Eternal DooM III&#13;\n                          Maps 12 and 13 have a few reused and redone &#13;\n                            areas from TeamTNT's Eternal DooM III.&#13;\n                          Music is from TeamTNT's Eternal DooM III&#13;\n                          The Monster Resource Wad (ZDooM community effort)&#13;\n                          Textures are from TeamTNT's Eternal DooM III; &#13;\n                            additional textures by Chris Couleur and Esa Repo&#13;\n                          Sounds are from TeamTNT's Eternal DooM III; &#13;\n                            additional sounds by Esa Repo&#13;\n&#13;\nBuild Time              : 3 years 9 months&#13;\nEditor(s) used          : WadAuthor and Doom Builder for maps, XWE for&#13;\n                            resource and wad management, DooM Builder to&#13;\n                            merge maps into resource wad&#13;\nKnown Bugs              : None&#13;\nMay Not Run With...     : The Pack (cheesy reference to the classic Bad&#13;\n                            Company album)&#13;\n&#13;\n&#13;\n* Copyright \/ Permissions *&#13;\n&#13;\nAuthors may NOT use the contents of this file as a base for modification or&#13;\nreuse.  Permissions have been obtained from original authors for any of&#13;\ntheir resources modified or included in this file.&#13;\n&#13;\nYou MAY distribute this file, provided you include this text file, with no&#13;\nmodifications.  You may distribute this file in any electronic format (BBS,&#13;\nDiskette, CD, etc) as long as you include this file intact.  I have&#13;\nreceived permission from the original authors of any modified or included&#13;\ncontent in this file to allow further distribution.&#13;\n&#13;\n* Where to get the file that this text file describes *&#13;\n&#13;\nThe Usual: ftp:\/\/archives.3dgamers.com\/pub\/idgames\/ and mirrors&#13;\n&#13;\n<\/textarea><\/div>\n<p><a href=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/ed4_rfo\/admin\/edit.php?id=15193\" target=\"another\" rel=\"nofollow\">mong<\/a><\/p>\n<form action=\"problem.php?id=15193\" method=\"post\"><table border=\"0\" style=\"margin: auto;\"><tr><td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n  <\/td><\/tr><\/table><\/form>","created_at":"2018-11-07T16:48:13+0000","updated_at":"2018-11-07T16:48:13+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"www.doomworld.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":441,"title":"Eternal Doom - Doomworld \/idgames database frontend","url":"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/eternal","content":"<div class=\"dir\"><img src=\"https:\/\/www.doomworld.com\/idgames\/icons\/back.gif\" alt=\"\" \/><a href=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/\">Back to Directory<\/a><\/div>\n<br \/><table style=\"margin: auto; width: 720px;\"><tr><td valign=\"top\" style=\"width: 550px;\">\n<table class=\"filelist\"><tr><td class=\"filelist_label\">Title:<\/td><td class=\"filelist_field1\">\nEternal Doom<\/td><\/tr><tr><td class=\"filelist_label\">Filename:<\/td><td class=\"filelist_field2\">\nthemes\/TeamTNT\/eternal\/eternal.zip<\/td><\/tr><tr><td class=\"filelist_label\">Size:<\/td><td class=\"filelist_field1\">\n9.03 MB<\/td><\/tr><tr><td class=\"filelist_label\">Date:<\/td><td class=\"filelist_field2\">\n11\/14\/97<\/td><\/tr><tr><td class=\"filelist_label\">Author:<\/td><td class=\"filelist_field1\">\nTeam Eternal<\/td><\/tr><tr><td class=\"filelist_label\">Description:<\/td><td class=\"filelist_field2\">\nEternal is complete. 32 single&#13;\nplayer levels including six new&#13;\nlevels and two bonus levels.&#13;\nIncludes a powerful shell. Final&#13;\nDOOM compatible. Download this,&#13;\nunzip it in a new directory and type&#13;\nETERNAL. Full instructions are in&#13;\nINSTALL.TXT inside. By Team Eternal.&#13;\nhttp:\/\/www.teamtnt.com&#13;\n<\/td><\/tr><tr><td class=\"filelist_label\">Credits:<\/td><td class=\"filelist_field1\">\n<\/td><\/tr><tr><td class=\"filelist_label\">Base:<\/td><td class=\"filelist_field2\">\n<\/td><\/tr><tr><td class=\"filelist_label\">Build time:<\/td><td class=\"filelist_field1\">\n<\/td><\/tr><tr><td class=\"filelist_label\">Editor(s) used:<\/td><td class=\"filelist_field2\">\n<\/td><\/tr><tr><td class=\"filelist_label\">Bugs:<\/td><td class=\"filelist_field1\">\n<\/td><\/tr><tr><td class=\"filelist_label\">Rating:<\/td><td class=\"filelist_field2\">\n<img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/emptystar.gif\" alt=\"\" \/> (340 votes)\n<\/td><\/tr><\/table><\/td>\n<td valign=\"top\" align=\"right\" style=\"width: 170px;\">\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/download.png\" width=\"160\" height=\"60\" alt=\"Download here\" title=\"IF ANYONE ASKS HOW TO DOWNLOAD FILES ANYMORE I WILL MURDER THEM\" \/><\/div>\n<br \/><table class=\"download\"><tr><td valign=\"top\">\n<strong>Download mirrors:<\/strong>\n<ul class=\"square\"><li><a href=\"ftp:\/\/ftp.fu-berlin.de\/pc\/games\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">Germany<\/a> <img title=\"Primary download mirror\" src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"ftp:\/\/mirrors.syringanetworks.net\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">Idaho<\/a> <img title=\"Primary download mirror\" src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"ftp:\/\/ftp.ntua.gr\/pub\/vendors\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">Greece<\/a> (<a href=\"http:\/\/ftp.ntua.gr\/pub\/vendors\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">HTTP<\/a>)<br \/><\/li><li><a href=\"http:\/\/ftp.mancubus.net\/pub\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">Texas<\/a><br \/><\/li><li><a href=\"https:\/\/www.quaddicted.com\/files\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">Germany (TLS)<\/a><br \/><\/li><li><a href=\"http:\/\/youfailit.net\/pub\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">New York<\/a> <img title=\"Does not mirror \/themes\/x-rated\" src=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"http:\/\/www.gamers.org\/pub\/idgames\/themes\/TeamTNT\/eternal\/eternal.zip\">Virginia<\/a>\n<\/li><\/ul><strong>\/idgames protocol:<\/strong>\n<ul class=\"idgamesprotocol\"><li><a href=\"denied:idgames:\/\/8688\">idgames:\/\/8688<\/a><\/li>\n<\/ul><\/td>\n<\/tr><\/table><br \/><\/td>\n<\/tr><\/table><br \/><div style=\"text-align: center;\">\n<div style=\"text-align: center;\"><a target=\"_new\" href=\"http:\/\/www.gamers.org\/pub\/idgames\/themes\/TeamTNT\/eternal\/eternal.txt\">View eternal.txt<\/a><\/div>\n<textarea rows=\"20\" cols=\"88\">\nETERNAL.ZIP - Eternal I, II, and III&#13;\n&#13;\nEternal is complete. 32 single&#13;\nplayer levels including six new&#13;\nlevels and two bonus levels.&#13;\nIncludes a powerful shell. Final&#13;\nDOOM compatible. Download this,&#13;\nunzip it in a new directory and type&#13;\nETERNAL. Full instructions are in&#13;\nINSTALL.TXT inside. By Team Eternal.&#13;\nhttp:\/\/www.teamtnt.com&#13;\n<\/textarea><\/div>\n<p><a href=\"https:\/\/www.doomworld.com\/idgames\/themes\/TeamTNT\/eternal\/admin\/edit.php?id=8688\" target=\"another\" rel=\"nofollow\">mong<\/a><\/p>\n<form action=\"problem.php?id=8688\" method=\"post\"><table border=\"0\" style=\"margin: auto;\"><tr><td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n  <\/td><\/tr><\/table><\/form>","created_at":"2018-11-07T16:48:13+0000","updated_at":"2018-11-07T16:48:13+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.doomworld.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":442,"title":"RYLAYEH - Doomworld \/idgames database frontend","url":"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh","content":"<div class=\"dir\"><img src=\"https:\/\/www.doomworld.com\/idgames\/icons\/back.gif\" alt=\"\" \/><a href=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/\">Back to Directory<\/a><\/div>\n<br \/><table style=\"margin: auto; width: 720px;\"><tr><td valign=\"top\" style=\"width: 550px;\">\n<table class=\"filelist\"><tr><td class=\"filelist_label\">Title:<\/td><td class=\"filelist_field1\">\nRYLAYEH<\/td><\/tr><tr><td class=\"filelist_label\">Filename:<\/td><td class=\"filelist_field2\">\nlevels\/doom2\/Ports\/p-r\/rylayeh.zip<\/td><\/tr><tr><td class=\"filelist_label\">Size:<\/td><td class=\"filelist_field1\">\n1.78 MB<\/td><\/tr><tr><td class=\"filelist_label\">Date:<\/td><td class=\"filelist_field2\">\n03\/29\/06<\/td><\/tr><tr><td class=\"filelist_label\">Author:<\/td><td class=\"filelist_field1\">\nJ S Graham<\/td><\/tr><tr><td class=\"filelist_label\">Description:<\/td><td class=\"filelist_field2\">\nTen levels based on the Lost City of R'Lyeh.<\/td><\/tr><tr><td class=\"filelist_label\">Credits:<\/td><td class=\"filelist_field1\">\nDeathmatch\/Multiplayer adaptation and Dehacked by Doorhenge, (J. Otey) Thanks for much help and many great ideas! (doorhenge@ya-nospam-hoo.com) <p>Danimetal, for playtesting\/checking<\/p><p>Art credits listed below<\/p><\/td><\/tr><tr><td class=\"filelist_label\">Base:<\/td><td class=\"filelist_field2\">\nfrom scratch<\/td><\/tr><tr><td class=\"filelist_label\">Build time:<\/td><td class=\"filelist_field1\">\n???<\/td><\/tr><tr><td class=\"filelist_label\">Editor(s) used:<\/td><td class=\"filelist_field2\">\nWadAuthor, DoomBuilder, XWE, Wintex, ZenNode.<\/td><\/tr><tr><td class=\"filelist_label\">Bugs:<\/td><td class=\"filelist_field1\">\nNone<\/td><\/tr><tr><td class=\"filelist_label\">Rating:<\/td><td class=\"filelist_field2\">\n<img src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/star.gif\" alt=\"\" \/><img src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/emptystar.gif\" alt=\"\" \/> (36 votes)\n<\/td><\/tr><\/table><\/td>\n<td valign=\"top\" align=\"right\" style=\"width: 170px;\">\n<div style=\"text-align: center;\">\n<img src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/download.png\" width=\"160\" height=\"60\" alt=\"Download here\" title=\"IF ANYONE ASKS HOW TO DOWNLOAD FILES ANYMORE I WILL MURDER THEM\" \/><\/div>\n<br \/><table class=\"download\"><tr><td valign=\"top\">\n<strong>Download mirrors:<\/strong>\n<ul class=\"square\"><li><a href=\"ftp:\/\/ftp.fu-berlin.de\/pc\/games\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">Germany<\/a> <img title=\"Primary download mirror\" src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"ftp:\/\/mirrors.syringanetworks.net\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">Idaho<\/a> <img title=\"Primary download mirror\" src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"ftp:\/\/ftp.ntua.gr\/pub\/vendors\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">Greece<\/a> (<a href=\"http:\/\/ftp.ntua.gr\/pub\/vendors\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">HTTP<\/a>)<br \/><\/li><li><a href=\"http:\/\/ftp.mancubus.net\/pub\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">Texas<\/a><br \/><\/li><li><a href=\"https:\/\/www.quaddicted.com\/files\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">Germany (TLS)<\/a><br \/><\/li><li><a href=\"http:\/\/youfailit.net\/pub\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">New York<\/a> <img title=\"Does not mirror \/themes\/x-rated\" src=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/images\/info.gif\" alt=\"image\" \/><br \/><\/li><li><a href=\"http:\/\/www.gamers.org\/pub\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.zip\">Virginia<\/a>\n<\/li><\/ul><strong>\/idgames protocol:<\/strong>\n<ul class=\"idgamesprotocol\"><li><a href=\"denied:idgames:\/\/14278\">idgames:\/\/14278<\/a><\/li>\n<\/ul><\/td>\n<\/tr><\/table><br \/><\/td>\n<\/tr><\/table><br \/><div style=\"text-align: center;\">\n<div style=\"text-align: center;\"><a target=\"_new\" href=\"http:\/\/www.gamers.org\/pub\/idgames\/levels\/doom2\/Ports\/p-r\/rylayeh.txt\">View rylayeh.txt<\/a><\/div>\n<textarea rows=\"20\" cols=\"88\">\n===========================================================================&#13;\nAdvanced engine needed  : Limit removing source-port&#13;\nPrimary purpose         : Fun&#13;\n===========================================================================&#13;\nTitle                   : RYLAYEH&#13;\nFilename                : RYLAYEH.wad&#13;\nRelease date            : March 29th, 2006&#13;\nAuthor                  : J S Graham &#13;\nEmail Address           : psyren@easy.com&#13;\nOther Files By Author   : Azagthoth, Crimson Canyon&#13;\nMisc. Author Info       :&#13;\n&#13;\nDescription             : Ten levels based on the Lost City of R'Lyeh.&#13;\n&#13;\nAdditional Credits to   : Deathmatch\/Multiplayer adaptation and Dehacked &#13;\n\t\t\t  by Doorhenge, (J. Otey)&#13;\n\t\t  \t  Thanks for much help and many great ideas!  &#13;\n  \t\t\t  (doorhenge@ya-nospam-hoo.com)  &#13;\n\t\t\t  &#13;\n\t\t\t  Danimetal, for playtesting\/checking&#13;\n&#13;\n\t\t\t  Art credits listed below&#13;\n===========================================================================&#13;\n* What is included *&#13;\n&#13;\nNew levels              : 10&#13;\nSounds                  : No&#13;\nMusic                   : Yes&#13;\nGraphics                : yes&#13;\nDehacked\/BEX Patch      : Yes - see below&#13;\nDemos                   : No&#13;\nOther                   : No&#13;\nOther files required    : No&#13;\n&#13;\nLimit removing source port required.  Some secret areas are accessible only by jumping.  Tested in GZDoom, JDoom, Risen3D, ZDoomGL and Legacy.&#13;\n&#13;\nPlease \"disable decoration models\" if playing with JDoom or Risen3D. (otherwise you will see small trees and electrical columns in very strange places)&#13;\n&#13;\n- The .DEH file only changes level names in the auto-map, the text between map06 and map07, and the sounds for the SS Soldier.  It has no effect on anything else. &#13;\n&#13;\n&#13;\n* Play Information *&#13;\n&#13;\nGame                    : DOOM2&#13;\nMap #                   : Maps 01-10&#13;\nSingle Player           : Yes&#13;\nCooperative 2-4 Player  : Yes&#13;\nDeathmatch 2-4 Player   : Yes&#13;\nOther game styles       : No&#13;\nDifficulty Settings     : Yes&#13;\n&#13;\n&#13;\n* Construction *&#13;\n&#13;\nBase                    : from scratch&#13;\nBuild Time              : ???&#13;\nEditor(s) used          : WadAuthor, DoomBuilder, XWE, Wintex, ZenNode.&#13;\nKnown Bugs              : None&#13;\nMay Not Run With...     : Doom2.&#13;\n&#13;\n===========================================================================&#13;\n* Art Credits *&#13;\n&#13;\nSky modified from Final Doom, \"Plutonia Experiment\" by Team TNT.&#13;\n&#13;\nEDEM1_5, EDEM1_6, EDOOR_1, EFC_1, EFC_2, EFC_3, EFC_4, EFLOR7_2, EFLOR7_8, EWALL_2, EWALL_3, EWEED, SLADRIP1-3, modified from original Doom2 textures by Id Software.  &#13;\n&#13;\nDecals for EWALL_5, EWALL_6, EWEED, and EGIGER_2 textures modified from &#13;\n\"The Brotherhood of Ruin\"  (ruinbros.wad) by Kristian Aro.&#13;\n&#13;\nEGIGER modified from \"Biomechanical Landscape I\" by H.R.Giger.&#13;\n&#13;\nParts of EGIGER_2 modified from \"Li II\" and \"New York City XI\" by H.R.Giger.&#13;\n&#13;\nSS Soldier Sprites are from \"The ZDoom Monster Resource Wad\".  &#13;\n&#13;\n2 KEEN Sprites modified from the original \"Blood\" game, by Monolith Productions.&#13;\n&#13;\nFCAN sprites modified from Hexen by Id Software.&#13;\n&#13;\nSMIT and TRE1 sprites from Doom64 by Midway.&#13;\n&#13;\nSprites recolored using ColorSwap. http:\/\/www.plexityhide.com\/ColorSwap.htm&#13;\n&#13;\nBigger blood splats by Chaoscorp.&#13;\n&#13;\nTITLEPIC modified from \"Homage to A.Boecklin\" by H.R.Giger.&#13;\n&#13;\nINTERPIC modified from \"Landscape XXX\" by H.R.Giger.&#13;\n&#13;\nMusic came from http:\/\/www.ravensdarkside.com\/index-21.html and http:\/\/canvaswerks.com\/spookymidi.htm.  I have no idea what the titles are or who made them, except for maps 03 and 10.  Map03 is the theme from the Disney film, \"The Black Hole.\"  Map10 is a version of \"Call of Cthulu\" by Metallica.&#13;\n&#13;\n===========================================================================&#13;\n* Copyright \/ Permissions *&#13;\n&#13;\nAuthors MAY use the contents of this file as a base for modification or reuse.  &#13;\nPermissions have not been obtained from original authors for any of their resources &#13;\nmodified or included in this file.&#13;\n&#13;\nYou MAY distribute this file in any electronic format.  I have not received permission from the original authors of any modified or included content in this file to allow further distribution.&#13;\n&#13;\n&#13;\n* Where to get the file that this text file describes *&#13;\n&#13;\nThe Usual: ftp:\/\/archives.3dgamers.com\/pub\/idgames\/ and mirrors&#13;\nWeb sites:&#13;\nFTP sites:&#13;\n&#13;\n&#13;\n===========================================================================&#13;\n<\/textarea><\/div>\n<p><a href=\"https:\/\/www.doomworld.com\/idgames\/levels\/doom2\/Ports\/p-r\/admin\/edit.php?id=14278\" target=\"another\" rel=\"nofollow\">mong<\/a><\/p>\n<form action=\"problem.php?id=14278\" method=\"post\"><table border=\"0\" style=\"margin: auto;\"><tr><td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n<\/td>\n<td valign=\"top\">\n  <\/td><\/tr><\/table><\/form>","created_at":"2018-11-07T16:48:13+0000","updated_at":"2018-11-07T16:48:13+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"www.doomworld.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":436,"title":"Introducing Scrivener 3 | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=15760559","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15760559\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/www.literatureandlatte.com\/introducing-scrivener-3\" class=\"storylink\">Introducing Scrivener 3<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=literatureandlatte.com\">literatureandlatte.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        214 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=Sujan\" class=\"hnuser\">Sujan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15760559\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15760559&amp;goto=item%3Fid%3D15760559\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Introducing%20Scrivener%203&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Introducing%20Scrivener%203\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15760559&amp;auth=a996e8502b31d8e09d390f6d8d45597253d67218\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15760559\">102\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:48:12+0000","updated_at":"2018-11-07T16:48:12+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":437,"title":"Clojure and 64-bit lwjgl (lein 2)","url":"https:\/\/stackoverflow.com\/questions\/14697783\/clojure-and-64-bit-lwjgl-lein-2","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        10&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>On a 64-bit JVM, in Linux, whatever does the native library loading insists on trying to load the 32-bit library.<\/p>\n<p>I'm asking this as a clojure and (even moreso) java n00b.<\/p>\n<p>The answers I've found so far (<a href=\"https:\/\/stackoverflow.com\/questions\/10558795\/using-lwjgl-in-leiningen-clojure\">Using lwjgl in Leiningen\/Clojure<\/a> seems to be the most prominent in search results) seem to be for older versions of lein, with issues swirling around the :native-dependencies or LD_LIBRARY_PATH in project.clj.<\/p>\n<p>I'm re-inventing the wheel and translating the basic jME tutorials into clojure as a little personal homework assignment toward learning them both. In all the tutorials and examples I've found so far of people showing how they did this, this part seems to \"just work.\"<\/p>\n<p>(It worked fine for me under Windows, for that matter).<\/p>\n<p>I'm using lein to build a new, empty project. I've set up dependencies on various versions of the jMonkeyEngine libraries on clojars. After 'lein deps', liblwjgl64.so and libopenal64.so wind up in the root of my project directory.<\/p>\n<p>When I try to 'lein run', it shows the monkey splash screen for the settings, then it throws an exception when it tries to actually run because it's trying to load liblwjgl.so.<\/p>\n<p>That file is present under target\/native\/linux and target\/native\/linux64 (though, oddly enough, not under target\/native\/linux32).<\/p>\n<p>If I copy the file it wants into my project's root, the error changes to \"wrong ELF class: ELFCLASS32 (Possible cause: architecture word width mismatch)\" which is the other set of discussions that google keeps giving me. The solutions mentioned there all seem to amount to \"Switch to a 32-bit JVM to get Minecraft working\", but I'd much rather get an actual clue about what's going on.<\/p>\n<p>This problem is very consistent across every set of dependency libraries I could get my hands on in clojars (back to version 2, anyway...those had bigger issues that didn't seem to be worth digging into). The charles-stain set \"Just Worked\" for me on windows.<\/p>\n<p>The best hypothesis I've been able to come up with so far runs something along these lines:<\/p>\n<p>I suspect that a version must get specified at some point. <a href=\"http:\/\/docs.oracle.com\/javase\/7\/docs\/technotes\/guides\/javaws\/developersguide\/syntax.html#resources\" rel=\"nofollow noreferrer\">http:\/\/docs.oracle.com\/javase\/7\/docs\/technotes\/guides\/javaws\/developersguide\/syntax.html#resources<\/a> mentions a resources property that looks like a perfect fit, but this doesn't seem to have anything to do with JNLP. I'm wondering if there's a manifest file sitting around somewhere in a .jar that I haven't been able to find (here's where my n00bishness comes into play...I don't really have a clue what I'm looking for).<\/p>\n<p>So, does anyone have any pointers about where I should look or who I should ask? I don't know enough to even know where to go from here. Would it make more sense to pester the jMonkeyEngine forums, #clojure, or the lwjgl mailing list (or whatever they use...I haven't actually looked into their side of things at all...should I?)<\/p>\n<p>The next thing I'm going to try is bundling up the jME libraries into my own repository. That seems like a big, daunting enough task that I decided to ask on here while I'm tackling that angle.<\/p>\n<p>I know this is vague, and I apologize for that. My google-fu has failed me. I'd appreciate any suggestions anyone can come up with.<\/p>\n<p>Thanks in advance!<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/14697783\/revisions\" title=\"show all edits to this post\">edited May 23 '17 at 12:19<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Feb 4 '13 at 23:41&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/114334\/james\">James<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            155112&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"3\">&#13;\n                                3 Answers&#13;\n                                3&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        1&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>I know this is an old question, but I found it upon deciding to work on a Clojure + jMonkeyEngine project.  I spent some time fighting with the packages in clojars, then looking into what it would take to build my own maven package, and I found a page on the jMonkeyEngine wiki that had this stern warning:<\/p>\n<blockquote>\n  <p>Note that using jME3 with maven is neither recommended nor supported\n  by the core jME team and the libraries in the maven repository might\n  be outdated!<\/p>\n<\/blockquote>\n<p>Later I found that I was able to successfully get a minimal JME app running by putting their jars directly in my project, and including them on the classpath using leiningen's :resource-paths config.<\/p>\n<p>Find the jars by downloading the JME3 SDK, and looking in jmonkeyplatform\/jmonkeyplatform\/libs.  Copy all of the jars into a \"lib\" folder inside of your project (yes you can remove the jars you won't use).  Then configure your lein project like so:<\/p>\n<pre>(defproject my-jme-project \"0.1.0-SNAPSHOT\"\n  :dependencies [[org.clojure\/clojure \"1.5.1\"]]\n  :resource-paths [\"resources\" \"lib\/*\"])\n<\/pre>\n<p>Run lein-repl, and the following should launch a minimal JME3 app:<\/p>\n<pre>(import 'com.jme3.app.SimpleApplication)\n(def app (proxy [SimpleApplication] [] (simpleInitApp [] nil)))\n(.start app)\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/19968016\/revisions\" title=\"show all edits to this post\">edited Nov 20 '13 at 19:40<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Nov 14 '13 at 1:48&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2117302\/michael-willis\">Michael Willis<\/a>&#13;\n        &#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Using native libraries with Java programs can be a pain at first, but it is not so very complicated in fact.<\/p>\n<p>First of all, your error, <code>wrong ELF class: ELFCLASS32<\/code> means that certain <em>64-bit program<\/em> tried to load <em>32-bit shared library<\/em>. This happens, obviously, when there is 32-bit library with the required name is present on the path where the program looks for the libraries, but the program itself is 64-bit. That's pretty much all for the explanation of the error.<\/p>\n<p>Now, about the main part of the question. There are indeed several ways to use native libraries with Java applications, some of them involving even environment variables. But IMO the most easiest way is to specify Java <em>system property<\/em> <code>java.library.path<\/code>, which is intended exactly for managing native libraries. This property should point to the directory where all native libraries for the program are present. JVM will look into that directory for all required shared objects.<\/p>\n<p>When you run JAR files directly, it is possible to specify system properties in command line, like this:<\/p>\n<pre>java -Djava.library.path=natives\/linux64 -jar yourprogram.jar\n<\/pre>\n<p>Here, we're specifying a path to the libraries relative to the current directory. As far as I concerned, this is a preferred way to set properties for standalone applications, which usually started by scripts.<\/p>\n<p>As for REPL, it seems it is possible to set JVM options for it inside <code>project.clj<\/code>, like this:<\/p>\n<pre>(defproject project \"version\"\n  ...\n  :jvm-opts [\"-Djava.library.path=target\/natives\/linux64\"])\n<\/pre>\n<p>I don't remember what is default working directory for the REPL, so this may need a little experimenting with the exact path, but you've got the point.<\/p>\n<p>Right now I'm developing lwjgl program myself (though not in Clojure but in plain Java, using Maven), and all what was really required is extract native libraries from jars (<code>maven-native-plugin<\/code> does this) and set <code>java.library.path<\/code> for them.<\/p>\n<p>BTW, it looks like that leiningen supports another important option in <code>project.clj<\/code>, <code>:native-path<\/code>, which seems to enable something like <code>maven-native-plugin<\/code> functionality: it specifies a directory to which all native dependencies should be extracted. Combining this with <code>:jvm-opts<\/code> should give you correct platform-independent REPL :)<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/14700660\/revisions\" title=\"show all edits to this post\">edited Feb 5 '13 at 5:25<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        &#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Feb 5 '13 at 5:18&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/788207\/vladimir-matveev\">Vladimir Matveev<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            66.6k14161203&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Still old question, but I got same problem, so I was able to setup jME3 on Closure by adding jCenter repository on <code>project.clj<\/code>file:<\/p>\n<pre>(defproject example-project \"0.1.0-SNAPSHOT\"\n  :description \"FIXME: write description\"\n  :url \"http:\/\/example.com\/FIXME\"\n  :license {:name \"Eclipse Public License\"\n            :url \"http:\/\/www.eclipse.org\/legal\/epl-v10.html\"}\n  :repositories [[\"jcenter\" \"http:\/\/jcenter.bintray.com\"]]\n  :dependencies [[org.clojure\/clojure \"1.6.0\"]\n                 [org.jmonkeyengine\/jme3-core \"3.1.0-beta1\"]\n                 [org.jmonkeyengine\/jme3-desktop \"3.1.0-beta1\"]\n                 [org.jmonkeyengine\/jme3-lwjgl \"3.1.0-beta1\"]])\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Oct 12 '16 at 15:16&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1697276\/afonso-lage\">Afonso Lage<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            130315&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-11-07T16:48:12+0000","updated_at":"2018-11-07T16:48:12+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":438,"title":"arcadia-unity\/Arcadia","url":"https:\/\/github.com\/arcadia-unity\/Arcadia\/blob\/develop\/USAGE.md","content":"<p><em>This document aims to capture the current state of Arcadia. The project is still pre-alpha and has not had an official release yet, so the ideas here remain in flux. That said, the API and architecture have stabilized enough that it merits recording for those willing to take the early plunge. The <a href=\"https:\/\/gitter.im\/arcadia-unity\/Arcadia\" rel=\"nofollow\">community Gitter<\/a> is the best place to ask questions.<\/em><\/p>\n<p><em>Most older videos and code snippets are no longer valid, particularly anything to do with <code>defcomponent<\/code>, as we underwent a major rearchitecting a year ago.<\/em><\/p>\n<p><em>This document assumes familiarity with <a href=\"http:\/\/clojure.org\/\" rel=\"nofollow\">Clojure<\/a> and <a href=\"http:\/\/unity3d.com\/\" rel=\"nofollow\">Unity<\/a>.<\/em><\/p>\n<ul><li>\n<p><a href=\"#arcadia\">Introduction<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#starting-arcadia\">Starting Arcadia from Scratch<\/a><\/p>\n<ul><li>\n<p><a href=\"#unity-projects\">Unity Projects<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#cloning-arcadia\">Cloning Arcadia<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#unity-configuration\">Unity Configuration<\/a><\/p>\n<ul><li><a href=\"#mono-runtime\">Mono Runtime<\/a><\/li>\n<li><a href=\"#running-in-the-background\">Running in the Background<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li>\n<p><a href=\"#arcadia-configuration\">Arcadia Configuration<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#improving-startup-times\">Improving Startup Times<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#livecoding-and-the-repl\">Livecoding and the Repl<\/a><\/p>\n<ul><li>\n<p><a href=\"#under-the-hood\">Under the Hood<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#editor-integration\">Editor Integration<\/a><\/p>\n<ul><li><a href=\"#command-line\">Command Line<\/a><\/li>\n<li><a href=\"#emacs\">Emacs<\/a><\/li>\n<li><a href=\"#sublimetext\">SublimeText<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li>\n<p><a href=\"#programming-in-arcadia\">Programming in Arcadia<\/a><\/p>\n<ul><li>\n<p><a href=\"#clojure-clr\">Clojure CLR<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#unity-interop\">Unity Interop<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#hooks\">Hooks<\/a><\/p>\n<ul><li><a href=\"#why-use-vars-rather-than-functions\">Why use Vars rather than functions?<\/a><\/li>\n<li><a href=\"#hooks-serialization-and-namespaces\">Hooks, Serialization and Namespaces<\/a><\/li>\n<li><a href=\"#entry-point\">Entry Point<\/a><\/li>\n<li><a href=\"#workflow\">Workflow<\/a><\/li>\n<\/ul><\/li>\n<li>\n<p><a href=\"#state\">State<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#multithreading\">Multithreading<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#namespace-roots\">Namespace Roots<\/a><\/p>\n<\/li>\n<li>\n<p><a href=\"#vm-restarting\">VM Restarting<\/a><\/p>\n<\/li>\n<\/ul><\/li>\n<li>\n<p><a href=\"#packages\">Packages<\/a><\/p>\n<ul><li><a href=\"#leiningen-projects\">Leiningen Projects<\/a><\/li>\n<li><a href=\"#other-projects\">Other Projects<\/a><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-arcadia\" class=\"anchor\" aria-hidden=\"true\" href=\"#arcadia\"><\/a>Arcadia<\/h2>\n<p>Arcadia is the integration of the <a href=\"http:\/\/clojure.org\/\" rel=\"nofollow\">Clojure programming language<\/a> with the <a href=\"http:\/\/unity3d.com\/\" rel=\"nofollow\">Unity 3D game engine<\/a>. The goal is to combine a modern, expressive programming language with the industry standard cross-platform engine for interactive media to transform the way we make creative work. Arcadia is free and open source and always will be.<\/p>\n<h2><a id=\"user-content-using-arcadia\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-arcadia\"><\/a>Using Arcadia<\/h2>\n<p>The basic Arcadia workflow is:<\/p>\n<ol><li><a href=\"#unity-projects\">Start a Unity project<\/a><\/li>\n<li><a href=\"#unity-configuration\">Configure Unity to work with Arcadia<\/a><\/li>\n<li><a href=\"#cloning-arcadia\">Clone the Arcadia repository into project<\/a><\/li>\n<li><a href=\"#livecoding-and-the-repl\">Connect to the REPL<\/a><\/li>\n<li><a href=\"#programming-in-arcadia\">Write a game, library, etc<\/a><\/li>\n<li>Export or publish(#leiningen) your project<\/li>\n<\/ol><h3><a id=\"user-content-unity-projects\" class=\"anchor\" aria-hidden=\"true\" href=\"#unity-projects\"><\/a>Unity Projects<\/h3>\n<p><em>These steps are not Arcadia specific.<\/em><\/p>\n<p>Arcadia is installed on a per-project basis. It is developed and tested Unity 2017.1.0f3<\/p>\n<p>Launch Unity and select <em>NEW<\/em> to start a new project<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b8e216221ffb6b28aba87331336b7a2662ab65cd\/687474703a2f2f696d6775722e636f6d2f4b38394c355a302e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/b8e216221ffb6b28aba87331336b7a2662ab65cd\/687474703a2f2f696d6775722e636f6d2f4b38394c355a302e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/K89L5Z0.png\" \/><\/a><\/p>\n<p>Fill out the <em>Project name<\/em> and <em>Location<\/em> fields. These determine where on your hard drive the files associated with the project will be created: a subdirectory with named <em>Project name<\/em> in the folder <em>Location<\/em>. In the following example, the project files will be created in <code>\/Users\/nasser\/Projects\/ArcadiaDocumentation<\/code>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/fa9cb514317ff67a6c339f905858cb4e09dff1d9\/687474703a2f2f696d6775722e636f6d2f7a784d6e41424d2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/fa9cb514317ff67a6c339f905858cb4e09dff1d9\/687474703a2f2f696d6775722e636f6d2f7a784d6e41424d2e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/zxMnABM.png\" \/><\/a><\/p>\n<p>Unity creates four folders to manage each project: Temp, Assets, Library, ProjectSettings. The folder that contains these four folders is the project itself. All of your own game assets (e.g. code, sound files, images, and 3D models) go in Assets. Everything else is essentially Unity's business. When version controlling a Unity project do not commit Temp or Library as their contents are often changed by Unity.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/eb1a53b5d7e0df692422bbdeb43b760a1863dda3\/687474703a2f2f696d6775722e636f6d2f3756593135735a2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/eb1a53b5d7e0df692422bbdeb43b760a1863dda3\/687474703a2f2f696d6775722e636f6d2f3756593135735a2e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/7VY15sZ.png\" \/><\/a><\/p>\n<p>You now have a new Unity project, almost ready for Arcadia.<\/p>\n<h3><a id=\"user-content-unity-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#unity-configuration\"><\/a>Unity Configuration<\/h3>\n<p><em>We are trying to automate this step, and it may disappear in the future<\/em><\/p>\n<p>By default, Unity is configured in a way that does not allow Arcadia to run well.<\/p>\n<h4><a id=\"user-content-mono-runtime\" class=\"anchor\" aria-hidden=\"true\" href=\"#mono-runtime\"><\/a>Mono Runtime<\/h4>\n<p>Unity uses a subset of the .NET 2.0 framework by default, which prevents any Clojure code from running. To change this to the full framework, select from the main menu Edit \u2192 Project Settings \u2192 Player. This will open the Player inspector in your inspector tab. In the Player inspector, locate Other Settings \u2192 Optimization \u2192 Api Compatibility Level and set that to .NET 2.0<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/52fcf22c96cf428071d738e878c7f3eefda29bf7\/687474703a2f2f696d6775722e636f6d2f697862613133592e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/52fcf22c96cf428071d738e878c7f3eefda29bf7\/687474703a2f2f696d6775722e636f6d2f697862613133592e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/ixba13Y.png\" \/><\/a><\/p>\n<h4><a id=\"user-content-running-in-the-background\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-in-the-background\"><\/a>Running in the Background<\/h4>\n<p>Unity will stop executing code if its application does not have focus. This is an optimization that gets in the way of live coding, where having Unity in the background while you send code to it from your editor is typical. To change this, in the Player inspector locate Resolution and Presentation \u2192 Run in Background and make sure the box is checked.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/aa7bd6934aa5815aa3c49d89595d70e4529a2876\/687474703a2f2f696d6775722e636f6d2f586d6a476b41652e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/aa7bd6934aa5815aa3c49d89595d70e4529a2876\/687474703a2f2f696d6775722e636f6d2f586d6a476b41652e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/XmjGkAe.png\" \/><\/a><\/p>\n<p>Unity is now ready for Arcadia.<\/p>\n<h3><a id=\"user-content-cloning-arcadia\" class=\"anchor\" aria-hidden=\"true\" href=\"#cloning-arcadia\"><\/a>Cloning Arcadia<\/h3>\n<p>As Arcadia has not been released, the only way to use it at the moment is by cloning the git repository. In the future, Arcadia builds will be available from the Unity Asset store, but the repository cloning path will remain useful for developers who need the most bleeding edge features and fixes.<\/p>\n<p>Git clients work differently, but the basic task is to<\/p>\n<ol><li>Clone the Arcadia repository at <code>https:\/\/github.com\/arcadia-unity\/Arcadia.git<\/code> into a folder called Arcadia in your Assets folder<\/li>\n<li>Switch to the <code>develop<\/code> branch<\/li>\n<\/ol><p>On OSX\/Linux machine, the shell commands are<\/p>\n<pre>$ cd ~\/Projects\/ArcadiaDocumentation\/Assets\n$ git clone https:\/\/github.com\/arcadia-unity\/Arcadia.git\n$ cd Arcadia\/\n$ git checkout develop\n<\/pre>\n<p>On Windows or using a GUI git client you will have to perform the equivalent operations.<\/p>\n<p>Switch back into Unity. It is normal for it to take a moment to load all the new code.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/380756fe7c0690f70953cc90d8acc694b4f2527b\/687474703a2f2f696d6775722e636f6d2f43644833714d692e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/380756fe7c0690f70953cc90d8acc694b4f2527b\/687474703a2f2f696d6775722e636f6d2f43644833714d692e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/CdH3qMi.png\" \/><\/a><\/p>\n<p>To confirm Arcadia's installation, check your Unity console. If it is not open, select from the main menu Window \u2192 Console. There may be errors and warnings, but if it ends with <code>Starting REPL<\/code> and <code>Arcadia Started!<\/code> you should be set.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b169b8af60fca0309ab6e9771fab119c27b0b666\/687474703a2f2f696d6775722e636f6d2f5072576f6330502e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/b169b8af60fca0309ab6e9771fab119c27b0b666\/687474703a2f2f696d6775722e636f6d2f5072576f6330502e706e67\" alt=\"\" data-canonical-src=\"http:\/\/imgur.com\/PrWoc0P.png\" \/><\/a><\/p>\n<h3><a id=\"user-content-arcadia-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#arcadia-configuration\"><\/a>Arcadia Configuration<\/h3>\n<p>Configuration for Arcadia itself is done via a user-supplied <code>configuration.edn<\/code> file that should be placed under <code>Assets<\/code> in your Unity project. Options to <code>configuration.edn<\/code> are given in the format of a Clojure map literal, and are documented in <a href=\"https:\/\/github.com\/arcadia-unity\/Arcadia\/blob\/develop\/Arcadia\/configuration.edn\"><code>Arcadia\/configuration.edn<\/code><\/a>, which is also the default configuration file.<\/p>\n<p>To determine the final set of configuration settings, Arcadia merges the user-supplied <code>configuration.edn<\/code> under <code>Assets<\/code> with the built-in <code>Arcadia\/configuration.edn<\/code>; in other words, user settings overwrite those specified in <code>Arcadia\/configuration.edn<\/code>. Please do not modify the default configuration file itself, as it provides the default behavior of Arcadia!<\/p>\n<p>Noteworthy configuration options include:<\/p>\n<ul><li>\n<p><code>:dependencies<\/code><\/p>\n<p>Specify project dependencies, using basically the same format as Leiningen; see <a href=\"#packages\">Packages<\/a> below.<\/p>\n<\/li>\n<li>\n<p><code>:arcadia.compiler\/loadpaths<\/code><\/p>\n<p>Specify paths to Clojure source Arcadia wouldn't pick up on otherwise.<\/p>\n<\/li>\n<li>\n<p><code>:reactive<\/code><\/p>\n<p>Specify whether Arcadia should automatically respond to changes in the file system.<\/p>\n<\/li>\n<li>\n<p><code>:reload-on-change<\/code><\/p>\n<p>Specify whether Arcadia should automatically reload Clojure files corresponding to already-loaded namespaces when they are saved, similar to Clojurescript's <a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\">Figwheel<\/a> library.<\/p>\n<\/li>\n<li>\n<p><code>:repl\/injections<\/code><\/p>\n<p>Specify forms to evaluate along with every form entered into the repl. Useful for things like always having the <code>clojure.repl<\/code> namespace available (ie, <code>:repl\/injections (use 'clojure.repl)<\/code>).<\/p>\n<\/li>\n<\/ul><p>See <a href=\"https:\/\/github.com\/arcadia-unity\/Arcadia\/blob\/develop\/configuration.edn\"><code>Arcadia\/configuration.edn<\/code><\/a> for more details.<\/p>\n<h3><a id=\"user-content-improving-startup-times\" class=\"anchor\" aria-hidden=\"true\" href=\"#improving-startup-times\"><\/a>Improving Startup Times<\/h3>\n<p>Arcadia's startup time can be improved by clicking <code>Arcadia \u2192 Build \u2192 Internal Namespaces<\/code> menu item. You only need to do this once per project, and each time you update Arcadia from git.<\/p>\n<p>This will compile all of Arcadia's internal namespaces (including Clojure's standard library) into <code>.clj.dll<\/code> files (containing MSIL bytecode) ahead of time. These files can be loaded faster than the <code>.clj<\/code> source code files we currently ship. Otherwise, internal namespaces are compiled every time <a href=\"#vm-restarting\">the VM is restarted<\/a>, which can slow down your workflow considerably.<\/p>\n<h3><a id=\"user-content-livecoding-and-the-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#livecoding-and-the-repl\"><\/a>Livecoding and the REPL<\/h3>\n<p>One of Arcadia's most exciting features is it's ability to livecode Unity. This is enabled by Clojure which, as a member of the Lisp family of programming languages, is designed with a Read Evaluate Print Loop (or REPL) in mind. The function of a REPL is to<\/p>\n<ol><li><strong>Read incoming source code<\/strong>, turning text into expressions<\/li>\n<li><strong>Evaluate new expressions<\/strong>, producing values, creating new functions, or effecting the state of the application<\/li>\n<li><strong>Print the results<\/strong> back to the programmer<\/li>\n<li><strong>Loop<\/strong> back to the first step, waiting for more code<\/li>\n<\/ol><p>The evaluated code can be as large as a whole file, or as small as an individual expression. The REPL is used to incrementally sculpt your game while developing it, to query the state of things while debugging, to put on livecoding performances, and much more.<\/p>\n<h4><a id=\"user-content-under-the-hood\" class=\"anchor\" aria-hidden=\"true\" href=\"#under-the-hood\"><\/a>Under the Hood<\/h4>\n<p>Arcadia uses a custom UDP socket REPL listening on port 11211. <em>It does not use nREPL<\/em>, though it may move to the built in Clojure socket REPL at some point in the future.<\/p>\n<p>The current protocol aims to be straightforward:<\/p>\n<ol><li>The REPL runs inside of Unity on a separate thread<\/li>\n<li>It expects strings of valid Clojure code over UDP on port 11211<\/li>\n<li>When new code is received, it is placed in a queue<\/li>\n<li>Meanwhile, on a regular interval, the contents of the queue are evaluated on the <em>main thread<\/em><\/li>\n<li>The results of the evaluation are sent back over the socket<\/li>\n<\/ol><h4><a id=\"user-content-editor-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#editor-integration\"><\/a>Editor Integration<\/h4>\n<p>Arcadia becomes most powerful when you can integrate the REPL into your coding editor. This is an area we are actively working on and has not settled down completely. We currently have a number of solutions, and you find what works best for you, your workflow, your editor, and your OS.<\/p>\n<h5><a id=\"user-content-command-line\" class=\"anchor\" aria-hidden=\"true\" href=\"#command-line\"><\/a>Command Line<\/h5>\n<p>We include two REPL clients <code>Editor\/repl-client.rb<\/code> and <code>Editor\/repl-client.javascript<\/code>. Running these from the command line via <code>ruby<\/code> or <code>node<\/code> will connect you to a listening Arcadia REPL and present you with a prompt at where you can evaluate code. Any editor that has facilities for interactive command line script integration can use these, too. You will need <a href=\"https:\/\/nodejs.org\" rel=\"nofollow\">Node.js<\/a> or <a href=\"https:\/\/www.ruby-lang.org\" rel=\"nofollow\">Ruby<\/a> installed to use these scripts (Ruby is installed by default on OSX).<\/p>\n<p>They both do the same thing, though the node REPL is somewhat more robust. Use whichever works best for you.<\/p>\n<h5><a id=\"user-content-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#emacs\"><\/a>Emacs<\/h5>\n<p><a href=\"https:\/\/github.com\/arcadia-unity\/arcadia-dot-el\">Emacs support<\/a> uses the included Ruby and Node REPL clients.<\/p>\n<h5><a id=\"user-content-sublimetext\" class=\"anchor\" aria-hidden=\"true\" href=\"#sublimetext\"><\/a>SublimeText<\/h5>\n<p>Our <a href=\"https:\/\/github.com\/arcadia-unity\/repl-sublimetext\">current SumblimeText support<\/a> builds on <a href=\"http:\/\/sublimerepl.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">SublimeREPL<\/a>, which you will have to install first. It uses copies of the Ruby and Node REPL clients.<\/p>\n<p>Work has started on an <a href=\"https:\/\/github.com\/nasser\/pipe\">alternative SumblimeText REPL integration<\/a> that does not depend on the script clients.<\/p>\n<h3><a id=\"user-content-programming-in-arcadia\" class=\"anchor\" aria-hidden=\"true\" href=\"#programming-in-arcadia\"><\/a>Programming in Arcadia<\/h3>\n<p>Arcadia is different from both Unity and Clojure in important ways. Knowledge of both is important, but so is understanding how Arcadia itself works is<\/p>\n<h4><a id=\"user-content-clojure-clr\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure-clr\"><\/a>Clojure CLR<\/h4>\n<p>Most Clojure programmers are familiar with the JVM-based version of the language, but Arcadia does not use that. Instead, it is built on the <a href=\"https:\/\/github.com\/clojure\/clojure-clr\">official port to the Common Language Runtime<\/a> that <a href=\"https:\/\/github.com\/dmiller\">David Miller<\/a> maintains. We maintain <a href=\"https:\/\/github.com\/arcadia-unity\/clojure-clr\">our own fork<\/a> of the compiler so that we can introduce Unity specific fixes.<\/p>\n<p>As an Arcadia programmer you should be aware of the differences between ClojureCLR and ClojureJVM, and the <a href=\"https:\/\/github.com\/clojure\/clojure-clr\/wiki\">ClojureCLR wiki<\/a> is a good place to start, in particular the pages on <a href=\"https:\/\/github.com\/clojure\/clojure-clr\/wiki\/Basic-CLR-interop\">interop<\/a> and <a href=\"https:\/\/github.com\/clojure\/clojure-clr\/wiki\/Specifying-types\">types<\/a>.<\/p>\n<h4><a id=\"user-content-unity-interop\" class=\"anchor\" aria-hidden=\"true\" href=\"#unity-interop\"><\/a>Unity Interop<\/h4>\n<p>Arcadia does not go out of its way to \"wrap\" the Unity API in Clojure functions. Instead, a lot of Arcadia programming bottoms out in interoperating directly with Unity. For a function to point the camera at a point in space could look like<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn point-camera [p]\n  (.. Camera\/main transform (LookAt p)))<\/pre><\/div>\n<p>This uses Clojure's <a href=\"http:\/\/clojure.org\/reference\/java_interop#_the_dot_special_form\" rel=\"nofollow\">dot special form<\/a> to access the static <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/Camera-main.html\" rel=\"nofollow\"><code>main<\/code><\/a> field of the <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/Camera.html\" rel=\"nofollow\"><code>Camera<\/code><\/a> class, which is a <code>Camera<\/code> instance and has a <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/Component-transform.html\" rel=\"nofollow\"><code>transform<\/code><\/a> property of type <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/Transform.html\" rel=\"nofollow\"><code>Transform<\/code><\/a> that has a <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/Transform.LookAt.html\" rel=\"nofollow\"><code>LookAt<\/code><\/a> method that takes a <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/Vector3.html\" rel=\"nofollow\"><code>Vector3<\/code><\/a>. It is the equivalent of <code>Camera.main.transform.LookAt(p)<\/code> in C#.<\/p>\n<p>Unity is a highly mutable, stateful system. The above function will mutate the main camera's rotation to look at the new point. Furthermore, the reference to <code>Camera\/main<\/code> could be changed by some other bit of code. Unity's API is single-threaded by design, so memory corruption is avoided. Your own Clojure code can still be be as functional and multi-threaded as you like, but keep in mind that talking to Unity side effecting and impure.<\/p>\n<p>There are parts of the Unity API that we have wrapped in Clojure functions, however. These are usually very commonly used methods that would be clumsy to use without wrapping, or would benefit from protocolization. The scope of what does and does not get wrapped is an on going design exercise of the framework, but in general we plan to be conservative about how much of our own ideas we impose on top of the Unity API.<\/p>\n<h4><a id=\"user-content-hooks\" class=\"anchor\" aria-hidden=\"true\" href=\"#hooks\"><\/a>Hooks<\/h4>\n<p>Unity will notify GameObjects when particular events occur, such as collisions with other objects or interaction from the user. Methods that listen for these events are called \"Messages\", and most are listed <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/MonoBehaviour.html\" rel=\"nofollow\">here<\/a>.<\/p>\n<p>In C# Unity development, you specify how to respond to messages by implementing <a href=\"https:\/\/docs.unity3d.com\/Manual\/CreatingComponents.html\" rel=\"nofollow\">Component<\/a> classes implementing message methods, and attaching instances of these classes to GameObjects.<\/p>\n<p>Arcadia provides a simple and consistent associative (key-based) interface into Unity's message system. Users can process the scene graph as data, transparently converting messages and state to Clojure persistent maps and back. Rather than defining static Component classes, Arcadia users associate callback functions (referred to as <em>hooks<\/em>) with GameObjects and Unity message types.<\/p>\n<p>This comprises a system of related functions:<\/p>\n<table><thead><tr><th>function<\/th>\n<th>description<\/th>\n<\/tr><\/thead><tbody><tr><td><code>(hook obj, message-kw, key)<\/code><\/td>\n<td>Retrieves a message callback (hook) from GameObject obj on a key<\/td>\n<\/tr><tr><td><code>(hook+ obj, message-kw, key, IFn)<\/code><\/td>\n<td>Sets a message callback on a key<\/td>\n<\/tr><tr><td><code>(hook- obj, message-kw, key)<\/code><\/td>\n<td>Removes a message callback on a key<\/td>\n<\/tr><tr><td><code>(state obj, key)<\/code><\/td>\n<td>Retrieves a piece of state from obj on a key<\/td>\n<\/tr><tr><td><code>(state+ obj, key)<\/code><\/td>\n<td>Sets a piece of state on a key<\/td>\n<\/tr><tr><td><code>(state- obj, key)<\/code><\/td>\n<td>Removes a piece of state on a key<\/td>\n<\/tr><tr><td><code>(update-state obj, key, f &amp; args)<\/code><\/td>\n<td>Updates the state on a key by applying function <code>f<\/code> to it<\/td>\n<\/tr><tr><td><code>(role obj, key)<\/code><\/td>\n<td>Retrieves a map containing all hooks and state on a key<\/td>\n<\/tr><tr><td><code>(role+ obj, key, role-map)<\/code><\/td>\n<td>Sets state and multiple hooks on a key<\/td>\n<\/tr><tr><td><code>(role- obj, key)<\/code><\/td>\n<td>Removes state and all hooks on a key<\/td>\n<\/tr><tr><td><code>(roles obj)<\/code><\/td>\n<td>Retrieves mapping from all keys on obj to the role-map for a given key<\/td>\n<\/tr><tr><td><code>(roles+ obj, roles-map)<\/code><\/td>\n<td>Sets multiple roles at once (may remove hooks or state)<\/td>\n<\/tr><\/tbody><\/table><p>The hook system is complemented by a type-definition mechanism for high-performance mutable state:<\/p>\n<table><thead><tr><th>form<\/th>\n<th>description<\/th>\n<\/tr><\/thead><tbody><tr><td><code>(defmutable OrbitState [^float radius])<\/code><\/td>\n<td>Defines mutable type with unboxed primitive field <code>radius<\/code><\/td>\n<\/tr><tr><td><code>(snapshot orbit-state)<\/code><\/td>\n<td>Retrieves persistent representation of <code>defmutable<\/code>-type instance <code>orbit-state<\/code><\/td>\n<\/tr><tr><td><code>(mutable orbit-state-map)<\/code><\/td>\n<td>Constructs <code>defmutable<\/code>-type instance from persistent representation <code>orbit-state-map<\/code><\/td>\n<\/tr><\/tbody><\/table><p><code>roles<\/code> will snapshot any <code>defmutable<\/code>-type instances in a GameObject's state, and <code>role+<\/code> will convert any <code>defmutable<\/code> snapshot data to a corresponding <code>defmutable<\/code> instance. Users can thereby process the scene graph as immutable data without sacrificing imperative performance.<\/p>\n<p>The hook system is fully optional and can coexist with other approaches. If you want to roll your own treatment of the scene graph Arcadia will not get in your way.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(use 'arcadia.core 'arcadia.linear)\n(import '[UnityEngine GameObject Time Mathf Transform])\n(defn orbit [^GameObject obj, k]         ; Takes the GameObject and the key this function was attached with\n  (let [{:keys [:radius]} (state obj k)] ; Looks up the piece of state corresponding to the key `k`\n    (with-cmpt obj [tr Transform]\n      (set! (. tr position)\n        (v3 (* radius (Mathf\/Cos Time\/realtimeSinceStartup))\n            0\n            (* radius (Mathf\/Sin Time\/realtimeSinceStartup)))))))\n(let [gobj (create-primitive :cube \"Orbiter\")]\n  (state+ gobj :orbit {:radius 5})          ; set up state\n  (hook+ gobj :fixed-update :orbit #'orbit) ; set up message callback (hook)\n  )<\/pre><\/div>\n<p>The role system can abbreviate this and treat it as data:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(let [gobj (create-primitive :cube \"Orbiter\")]\n  (role+ gobj :orbit\n    {:state {:radius 5}\n     :fixed-update #'orbit}))<\/pre><\/div>\n<p><code>role+<\/code> is intended to emulate <code>assoc<\/code> with the GameObject in the place of a map, and will therefore <em>replace<\/em> any hooks and state already on the GameObject at the specified key.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn orbit-collide [obj1 obj2 k]\n  (UnityEngine.Debug\/Log \"a collision!\"))\n;; We'll start by setting up a role.\n(role+ (object-named \"Orbiter\") :orbit\n  {:state {:radius 5},\n   :fixed-update #'orbit\n   :on-collision-enter #'orbit-collide})\n(role (object-named \"Orbiter\") :orbit)\n;; =&gt;\n;; {:state {:radius 5},\n;;  :fixed-update #'orbit\n;;  :on-collision-enter #'orbit-collide}\n(defn other-orbit-collide [obj1 obj2 k]\n  (UnityEngine.Debug\/Log \"another collision!\"))\n;; If we call `role+` again with the same key but a different map of\n;; hooks and state, it will:\n;; - reset hooks and state to the specified values\n;; - remove any existing hooks or state not specified\n(role+ (object-named \"Orbiter\") :orbit\n  {:state {:radius 8},\n   :on-collision-enter #'other-orbit-collide})\n(role (object-named \"Orbiter\") :orbit)\n;; =&gt;\n;; {:state {:radius 8},\n;;  :on-collision-enter #'other-orbit-collide}\n;; Note that the :fixed-update hook has been removed.<\/pre><\/div>\n<p>Attached state and roles can then be retrieved as persistent maps using <code>role<\/code>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>  (role (object-named \"Orbiter\") :orbit)\n  ;; =&gt;\n  ;; {:state {:radius 5},\n  ;;  :fixed-update #'user\/orbit}<\/pre><\/div>\n<p>Multiple callbacks and pieces of state may be attached in this manner.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>;; ...\n(defn bobble [^GameObject obj, k] ; Takes the GameObject and the key this function was attached with\n  (let [{:keys [:amplitude]} (state obj k)] ; Looks up the piece of state corresponding to the key `k`\n    (with-cmpt obj [tr Transform]\n      (set! (. tr position)\n        (v3 (.. tr position x)\n            (* amplitude (Mathf\/Cos Time\/realtimeSinceStartup))\n            (.. tr position y))))))\n(let [gobj (create-primitive :cube \"Orbiter\")]\n  (state+ gobj :orbit {:radius 5})          ; set up state\n  (hook+ gobj :fixed-update :orbit #'orbit) ; set up the hook\n  (state+ gobj :bobble {:amplitude 5})      ; set up another piece of state\n  (hook+ gobj :fixed-update #'bobble)       ; set up another hook\n  )<\/pre><\/div>\n<p>All keys can be accessed and managed at once using <code>roles<\/code> and <code>roles+<\/code>.<\/p>\n<p><code>roles<\/code> bundles up all the keys, hooks, and state into a map with the hook or state keys as its keys, and maps suitable for use in <code>role+<\/code> as its values.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>  (roles (object-named \"Orbiter\"))\n  ;; =&gt;\n  ;; {:orbit {:state {:radius 5},\n  ;;          :fixed-update #'orbit},\n  ;;  :bobble {:state {:amplitude 5},\n  ;;           :fixed-update #'bobble}}<\/pre><\/div>\n<p><code>roles+<\/code> attaches a bundle of keys, hooks, and state, such as might be returned by <code>roles<\/code>, to the scene graph.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>  (roles+ (object-named \"Orbiter\")\n    {:orbit {:state {:radius 5},\n             :fixed-update #'orbit},\n     :bobble {:state {:amplitude 5},\n              :fixed-update #'bobble}})<\/pre><\/div>\n<h5><a id=\"user-content-why-use-vars-rather-than-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-use-vars-rather-than-functions\"><\/a>Why use Vars rather than functions?<\/h5>\n<p>While anonymous <code>fn<\/code> functions can be used as hook callbacks, Vars are greatly preferred because they can be serialized. Any data that would be closed over by an anonymous function instance can be stored in the state corresponding to a callback Var.<\/p>\n<p>For example, consider the following function that takes a GameObject and a numeric rate of spin, and starts the GameObject spinning at that rate.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn start-spinning [obj, rate]\n  (let [f (fn [obj, _]\n            (with-cmpt obj [tr Transform]\n              (set! (.. tr rotation)\n                (qq* (.. tr rotation)\n                     (aa rate 0 1 0)))))] ; uses `rate` to set degrees spun per frame\n    (role+ obj ::gyrate\n      {:update f})))<\/pre><\/div>\n<p>Here we attach an anonymous <code>fn<\/code> to <code>obj<\/code> as the <code>:update<\/code> hook. Anonymous <code>fn<\/code>'s cannot serialize, therefore this GameObject cannot serialize, greatly limiting its role in our scene.<\/p>\n<p>We can solve this problem by boosting the <code>:update<\/code> hook into a top-level named function, and storing <code>rate<\/code> in the state:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn spin [obj, k]\n  (let [{:keys [rate]} (state obj k)] ; retrieve rate of spinning from state\n    (with-cmpt obj [tr Transform]\n      (set! (.. tr rotation)\n        (qq* (.. tr rotation)\n             (aa rate 0 1 0))))))\n(defn start-spinning [obj, rate]\n (role+ obj ::gyrate\n   {:update #'spin        ; attach the Var #'spin as an update hook\n    :state {:rate rate}}) ; store rate of spinning in state\n )<\/pre><\/div>\n<p>This arrangement will serialize correctly.<\/p>\n<h5><a id=\"user-content-hooks-serialization-and-namespaces\" class=\"anchor\" aria-hidden=\"true\" href=\"#hooks-serialization-and-namespaces\"><\/a>Hooks, Serialization and Namespaces<\/h5>\n<p>Hooks attached to a GameObject are stored and managed by instances of <a href=\"https:\/\/github.com\/arcadia-unity\/Arcadia\/blob\/develop\/Components\/ArcadiaBehaviour.cs\">ArcadiaBehaviour<\/a> components.<\/p>\n<p>When the ArcadiaBehaviour instance is <a href=\"https:\/\/docs.unity3d.com\/Manual\/script-Serialization.html\" rel=\"nofollow\">serialized<\/a>, the hooks and their keys are converted into an <a href=\"https:\/\/github.com\/edn-format\/edn\">edn<\/a> string. Upon deserialization, any Vars attached as hooks are <a href=\"https:\/\/clojuredocs.org\/clojure.core\/intern\" rel=\"nofollow\">interned<\/a>, but their corresponding namespace is not immediately <a href=\"https:\/\/clojuredocs.org\/clojure.core\/require\" rel=\"nofollow\"><code>require<\/code>'d<\/a>. If a referenced user namespace includes definitions or logic that assume the existence of certain elements in the scene graph, for example, or use methods forbidden during deserialization such as <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/GameObject.Find.html\" rel=\"nofollow\"><code>UnityEngine.Object\/Find<\/code><\/a>, we want to delay loading them. The situation is similar to that motivating <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/GlobalEventHandlers\/onload\" rel=\"nofollow\"><code>window.onload<\/code><\/a> in frontend Javascript.<\/p>\n<p>The namespace corresponding to a deserialized Var is <code>required<\/code> once for whichever of the following happens first:<\/p>\n<ul><li>The <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/MonoBehaviour.Awake.html\" rel=\"nofollow\"><code>Awake<\/code><\/a> method of the <code>ArcadiaBehaviour<\/code> instance is called.<\/li>\n<li>The particular message method of the <code>ArcadiaBehaviour<\/code> instance is called -- ie, for an instance of the derived <a href=\"https:\/\/github.com\/arcadia-unity\/Arcadia\/blob\/develop\/Components\/FixedUpdateHook.cs\"><code>FixedUpdateHook.cs<\/code><\/a> class, when the <code>FixedUpdate<\/code> method is first called.<\/li>\n<\/ul><p>Remember <code>require<\/code> in Clojure by default will <em>not<\/em> run the code, including definitions, in a given namespace if that namespace has already been <code>require<\/code>'d. It is therefore safe to call <code>require<\/code> multiple times on the same namespace.<\/p>\n<h5><a id=\"user-content-entry-point\" class=\"anchor\" aria-hidden=\"true\" href=\"#entry-point\"><\/a>Entry Point<\/h5>\n<p>Clojure developers often look for a \"main\" function or an \"entry point\" that they can use to kick off their code. Unity does not have such a thing. <em>All<\/em> code at runtime is triggered through Messages, meaning <em>all code at runtime is triggered by Components attached to GameObjects<\/em>. If your game logic would benefit from an entry point, you will have to add an empty GameObject to the scene with a <code>:start<\/code> hook associated with the function you want to run at startup.<\/p>\n<h5><a id=\"user-content-workflow\" class=\"anchor\" aria-hidden=\"true\" href=\"#workflow\"><\/a>Workflow<\/h5>\n<p>For hooks that serialize and persist, writing code to a file is crucial. At the same time, the UI for adding hooks is not finished yet, so the REPL is still needed. A good workflow is the following:<\/p>\n<ol><li>Start a Clojure file in <code>Assets<\/code> following Clojure folder naming conventions (e.g. <code>Assets\/game\/core.clj<\/code>)<\/li>\n<li>Give the file a namespace form as usual with any <code>require<\/code>s you need (e.g. <code>(ns game.core (:require arcadia.core))<\/code>)<\/li>\n<li>Define a function in that file that you want to use as a hook (e.g. <code>(defn rotate [gameobject] (.. gameobject transform (Rotate 0 1 0)))<\/code>)<\/li>\n<li>Save the file<\/li>\n<li>In the REPL, require the namespace and use <code>arcadia.core<\/code> (e.g. <code>(require 'game.core) (use 'arcadia.core)<\/code>)<\/li>\n<li>In the REPL, attach the var refering to the function to a GameObject as a hook (e.g. <code>(hook+ (object-named \"Main Camera\") :update #'game.core\/rotate)<\/code>)<\/li>\n<li>Save the scene<\/li>\n<li>The GameObject will <code>require<\/code> the namespace associated with the var and lookup the function so that it can be invoked whenever unity sends its message. Hooks require the namespace of their associated vars once, either when their message is first called or on <code>Awake<\/code>, whichever comes first.<\/li>\n<\/ol><h4><a id=\"user-content-multithreading\" class=\"anchor\" aria-hidden=\"true\" href=\"#multithreading\"><\/a>Multithreading<\/h4>\n<p>While the Unity scene graph API is ostensibly single threaded, the Mono VM it runs on is not. This means you can write multithreaded Clojure code in all its functional glory, provided that you do not call Unity scene graph methods off the main thread (they will throw an exception).<\/p>\n<p>Note that not all types in the UnityEngine namespace need to be used from the main thread. Value types such as Vector3 are usable anywhere, for example.<\/p>\n<p>To trigger behavior restricted to the main thread from other threads, consider implementing a callback queue driven by the <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/MonoBehaviour.Update.html\" rel=\"nofollow\">Update Monobehaviour method<\/a> of a component in the scene graph, or an <a href=\"https:\/\/docs.unity3d.com\/ScriptReference\/EditorApplication-update.html\" rel=\"nofollow\">EditorApplication.update<\/a> delegate for edit-mode operations. An example of the latter can be found in the (internal, unstable) Arcadia namespace <a href=\"https:\/\/github.com\/arcadia-unity\/Arcadia\/blob\/develop\/Source\/arcadia\/internal\/editor_callbacks.clj\">arcadia.internal.editor-callbacks<\/a>.<\/p>\n<h4><a id=\"user-content-namespace-roots\" class=\"anchor\" aria-hidden=\"true\" href=\"#namespace-roots\"><\/a>Namespace Roots<\/h4>\n<p>The <code>Assets<\/code> folder is the root of your namespaces. So a file at <code>Assets\/game\/logic\/hard_ai.clj<\/code> should correspond to the namespace <code>game.logic.hard-ai<\/code>. Arcadia internally manages other namespace roots as well for its own operation, but <code>Assets<\/code> is where your own logic should go.<\/p>\n<h4><a id=\"user-content-vm-restarting\" class=\"anchor\" aria-hidden=\"true\" href=\"#vm-restarting\"><\/a>VM Restarting<\/h4>\n<p>Unity will restart the Mono VM at specific times<\/p>\n<ul><li>Whenever the editor compiles a file (e.g. when a <code>.cs<\/code> file under <code>Assets<\/code> is updated or a new <code>.dll<\/code> file is added)<\/li>\n<li>Whenever the editor enters play mode<\/li>\n<\/ul><p>When this happens, any state you had set up in the REPL will be lost and you will have to re-<code>require<\/code> any namespaces you were working with. This is a deep part of Unity's architecture and not something we can meaningfully affect.<\/p>\n<h2><a id=\"user-content-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#packages\"><\/a>Packages<\/h2>\n<p>Arcadia ships with a built-in package manager compatible with <a href=\"https:\/\/leiningen.org\/\" rel=\"nofollow\">Leiningen<\/a>.<\/p>\n<p>Arcadia supports developing (and publishing, via Leiningen) multiple Clojure libraries in a single Unity project. Dependencies can be specified either in <code>Assets\/configuration.edn<\/code> or in the <code>project.clj<\/code> file of any Leiningen project in <code>Assets<\/code>.<\/p>\n<p>By default, Arcadia will automatically pull in any newly-specified dependencies from Maven when one of these files changes. This happens on a separate thread, and will not block the normal operation of Unity or Arcadia.<\/p>\n<h3><a id=\"user-content-specifying-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#specifying-dependencies\"><\/a>Specifying Dependencies<\/h3>\n<p>Dependencies may be specified in the user-supplied <code>Assets\/configuration.edn<\/code> with a vector keyed to <code>:dependencies<\/code>, with the same vector-of-vectors format used by Leiningen.<\/p>\n<p>Arcadia will merge the dependencies declared in <code>Assets\/configuration.edn<\/code> with those declared in all Leiningen <code>project.clj<\/code>s to compute the final set of dependencies and pull them in from Maven.<\/p>\n<h3><a id=\"user-content-leiningen-projects\" class=\"anchor\" aria-hidden=\"true\" href=\"#leiningen-projects\"><\/a>Leiningen Projects<\/h3>\n<p>Leiningen projects should work correctly in Arcadia. Any directory in the <code>Assets<\/code> folder that contains a properly-formatted Leiningen <code>project.clj<\/code> will be considered a Leiningen project. Arcadia will treat the roots of these directories the same way Leiningen does, using <code>:source-paths<\/code> if it is defined and <code>src<\/code> otherwise.<\/p>\n<p>For example, say one has a Leiningen project <code>some_lein_project<\/code> under <code>Assets<\/code> with the following structure:<\/p>\n<pre>Assets\n|- some_lein_project\n   |- project.clj\n   |- src\n      |- top_namespace\n         |- inner_namespace\n            |- core.clj\n<\/pre>\n<div class=\"highlight highlight-source-clojure\"><pre>(require 'top-namespace.inner-namespace.core)<\/pre><\/div>\n<p>In this case, Arcadia would look in <code>Assets\/some_lein_project\/src\/top_namespace\/inner_namespace\/core.clj<\/code> for the corresponding file (and find it).<\/p>\n<p>Arcadia will extract the declared dependencies of a Leiningen project's <code>project.clj<\/code> and pull them in from Maven, just like running <code>lein deps<\/code>.<\/p>\n<p>Arcadia libraries can be published the same way you would publish a Leiningen library. Once published, a Leiningen library can be declared as a dependency either in <code>Assets\/configuration.edn<\/code> or the <code>project.clj<\/code> of a Leiningen project directly under <code>Assets<\/code>.<\/p>\n<p>There is no guarantee, of course, that a Clojure library which works fine on the JVM will work in Arcadia. Porting JVM libraries to Arcadia is usually fairly easy, however. Libraries with no host interop can usually be ported directly, and C# is similar enough to Java that even porting libraries with heavy interop is often straightforward.<\/p>\n<h3><a id=\"user-content-other-projects\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-projects\"><\/a>Other Projects<\/h3>\n<p>If you need to set the root of a project somewhere else, you can specify it in the <code>Assets\/configuration.edn<\/code> file. Place all paths you wish to consider roots for Clojure namespaces in a vector keyed to <code>:arcadia.compiler\/loadpaths<\/code>, this works similarly to Leiningen <code>:source-paths<\/code>.<\/p>\n<p>This option is especially useful for cloning repositories into Arcadia that don't have a Leiningen-style structure, and that don't assume their containing directory should be their Clojure namespace root.<\/p>","created_at":"2018-11-07T16:48:12+0000","updated_at":"2018-11-07T16:48:12+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":21,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/8031566?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":439,"title":"finding closest craigslist site for given address","url":"https:\/\/stackoverflow.com\/questions\/7845542\/finding-closest-craigslist-site-for-given-address","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        2&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>how to map closest craigslist site for a given address(zip)? or I was thinking of making that mapping myself by looking at the list of craigslist sites city-wise\/state-wise.<\/p><\/div><div class=\"user-action-time\">&#13;\n        asked Oct 21 '11 at 6:03&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/477403\/amol-pujari\">Amol Pujari<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1,2441235&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"1\">&#13;\n                                1 Answer&#13;\n                                1&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Indeed, I believe the best solution is to get a list of Craigslist sites\/cities, prebuild a mapping of cities to coordinates using some geocoding API (<a href=\"http:\/\/tinygeocoder.com\/\" rel=\"nofollow\">Tiny Geocoder<\/a> comes to mind but Yahoo and Bing offer good solutions too), and then storing it in a way that makes sort-by-nearest easy (say, <a href=\"http:\/\/dev.mysql.com\/doc\/refman\/5.0\/en\/spatial-extensions.html\" rel=\"nofollow\">MySQL's geospatial extensions<\/a> or <a href=\"http:\/\/www.mongodb.org\/display\/DOCS\/Geospatial+Indexing\" rel=\"nofollow\">MongoDB's<\/a>).<\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Oct 21 '11 at 7:56&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/179125\/jordan-running\">Jordan Running<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            73.6k10128129&#13;\n        <\/div>&#13;<\/div>","created_at":"2018-11-07T16:48:12+0000","updated_at":"2018-11-07T16:48:12+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":435,"title":"Introducing Scrivener 3 | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=15760559","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15760559\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15760559\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/www.literatureandlatte.com\/introducing-scrivener-3\" class=\"storylink\">Introducing Scrivener 3<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=literatureandlatte.com\">literatureandlatte.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        214 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=Sujan\" class=\"hnuser\">Sujan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15760559\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15760559&amp;goto=item%3Fid%3D15760559\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Introducing%20Scrivener%203&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Introducing%20Scrivener%203\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15760559&amp;auth=a996e8502b31d8e09d390f6d8d45597253d67218\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15760559\">102\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:48:10+0000","updated_at":"2018-11-07T16:48:10+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":434,"title":"Clojure Design Patterns","url":"http:\/\/mishadoff.com\/blog\/clojure-design-patterns\/","content":"<div id=\"page\"><p><em>Quick<\/em> overview of the classic <a href=\"http:\/\/en.wikipedia.org\/wiki\/Design_Patterns\">Design Patterns<\/a> in Clojure<\/p><p><strong>Disclaimer:<\/strong> <em>Most patterns are easy to implement because\nwe use dynamic typing, functional programming and, of course,\nClojure. Some of them look wrong and ugly. It\u2019s okay.\nAll characters are fake, coincidences are accidental.<\/em><\/p><h3 id=\"index\">Index<\/h3><ul><li><a href=\"#intro\">Intro<\/a><\/li>\n  <li><a href=\"#episode-1-command\">Episode  1. Command<\/a><\/li>\n  <li><a href=\"#episode-2-strategy\">Episode  2. Strategy<\/a><\/li>\n  <li><a href=\"#episode-3-state\">Episode  3. State<\/a><\/li>\n  <li><a href=\"#episode-4-visitor\">Episode  4. Visitor<\/a><\/li>\n  <li><a href=\"#episode-5-template-method\">Episode  5. Template Method<\/a><\/li>\n  <li><a href=\"#episode-6-iterator\">Episode  6. Iterator<\/a><\/li>\n  <li><a href=\"#episode-7-memento\">Episode  7. Memento<\/a><\/li>\n  <li><a href=\"#episode-8-prototype\">Episode  8. Prototype<\/a><\/li>\n  <li><a href=\"#episode-9-mediator\">Episode  9. Mediator<\/a><\/li>\n  <li><a href=\"#episode-10-observer\">Episode 10. Observer<\/a><\/li>\n  <li><a href=\"#episode-11-interpreter\">Episode 11. Interpreter<\/a><\/li>\n  <li><a href=\"#episode-12-flyweight\">Episode 12. Flyweight<\/a><\/li>\n  <li><a href=\"#episode-13-builder\">Episode 13. Builder<\/a><\/li>\n  <li><a href=\"#episode-14-facade\">Episode 14. Facade<\/a><\/li>\n  <li><a href=\"#episode-15-singleton\">Episode 15. Singleton<\/a><\/li>\n  <li><a href=\"#episode-16-chain-of-responsibility\">Episode 16. Chain of Responsibility<\/a><\/li>\n  <li><a href=\"#episode-17-composite\">Episode 17. Composite<\/a><\/li>\n  <li><a href=\"#episode-18-factory-method\">Episode 18. Factory Method<\/a><\/li>\n  <li><a href=\"#episode-19-abstract-factory\">Episode 19. Abstract Factory<\/a><\/li>\n  <li><a href=\"#episode-20-adapter\">Episode 20. Adapter<\/a><\/li>\n  <li><a href=\"#episode-21-decorator\">Episode 21. Decorator<\/a><\/li>\n  <li><a href=\"#episode-22-proxy\">Episode 22. Proxy<\/a><\/li>\n  <li><a href=\"#episode-23-bridge\">Episode 23. Bridge<\/a><\/li>\n  <li><a href=\"#cheatsheet\">Cheatsheet<\/a><\/li>\n  <li><a href=\"#cast\">Cast<\/a><\/li>\n<\/ul><blockquote>\n  <p>Our programming language is fucked up.<br \/>That\u2019s why we need design patterns.<\/p>\n  <p>\u2013 Anonymous<\/p>\n<\/blockquote><p>Two modest programmers <strong>Pedro Veel<\/strong> and <strong>Eve Dopler<\/strong>\nare solving common software engineering problems and\napplying design patterns.<\/p><h3 id=\"episode-1-command\">Episode 1. Command<\/h3><blockquote>\n  <p>Leading IT service provider <strong>\u201cSerpent Hill &amp; R.E.E\u201d<\/strong>\nacquired new project for USA customer.\nFirst delivery is a register, login and logout functionality\nfor their brand new site.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Oh, that\u2019s easy. You just need a Command interface\u2026<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>interface Command {\n  void execute();\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Every action should implement this interface and define specific <code class=\"highlighter-rouge\">execute<\/code> behaviour.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class LoginCommand implements Command {\n  private String user;\n  private String password;\n  public LoginCommand(String user, String password) {\n    this.user = user;\n    this.password = password;\n  }\n  @Override\n  public void execute() {\n    DB.login(user, password);\n  }\n}\n<\/pre><\/div><\/div><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class LogoutCommand implements Command {\n  private String user;\n  public LogoutCommand(String user) {\n    this.user = user;\n  }\n  @Override\n  public void execute() {\n    DB.logout(user);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Usage is simple as well.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>(new LoginCommand(\"django\", \"unCh@1ned\")).execute();\n(new LogoutCommand(\"django\")).execute();\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> What do you think, Eve?<br \/><strong>Eve:<\/strong> Why are you using redundant wrapping into <code class=\"highlighter-rouge\">LoginCommand<\/code> and just don\u2019t call <code class=\"highlighter-rouge\">DB.login<\/code>?<br \/><strong>Pedro:<\/strong> It\u2019s important to wrap here, because now we can operate on generic <code class=\"highlighter-rouge\">Command<\/code> objects.<br \/><strong>Eve:<\/strong> For what purpose?<br \/><strong>Pedro:<\/strong> Delayed call, logging, history tracking, caching, plenty of usages.<br \/><strong>Eve:<\/strong> Ok, how about that?<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn execute [command]\n  (command))\n(execute #(db\/login \"django\" \"unCh@1ned\"))\n(execute #(db\/logout \"django\"))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> What the hell this hash sign?<br \/><strong>Eve:<\/strong> A shortcut for javaish<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>new SomeInterfaceWithOneMethod() {\n  @Override\n  public void execute() {\n    \/\/ do\n  }\n};\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Just like <code class=\"highlighter-rouge\">Command<\/code> interface\u2026<br \/><strong>Eve:<\/strong> Or if you want - no-hash solution.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn execute [command &amp; args]\n  (apply command args))\n(execute db\/login \"django\" \"unCh@1ned\")\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And how do you save function for delayed call in that case?<br \/><strong>Eve:<\/strong> Answer yourself. What do you need to call a function?<br \/><strong>Pedro:<\/strong> Its name\u2026<br \/><strong>Eve:<\/strong> And?<br \/><strong>Pedro:<\/strong> \u2026arguments.<br \/><strong>Eve:<\/strong> Bingo. All you do is saving a pair (<em>function-name<\/em>, <em>arguments<\/em>) and call it whenever you want using\n<code class=\"highlighter-rouge\">(apply function-name arguments)<\/code><br \/><strong>Pedro:<\/strong> Hmm\u2026 Looks simple.<br \/><strong>Eve:<\/strong> Definitely, <strong>Command is just a function.<\/strong><\/p><h3 id=\"episode-2-strategy\">Episode 2. Strategy<\/h3><blockquote>\n  <p><strong>Sven Tori<\/strong> pays a lot of money\nto see a page with list of users.\nBut users must be sorted by name and\nusers with subscription must appear <em>before<\/em>\nall other users.\nObviously, because they pay.\nReverse sorting should keep subscripted users on top.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Ha, just call <code class=\"highlighter-rouge\">Collections.sort(users, comparator)<\/code>\nwith custom comparator.<br \/><strong>Eve:<\/strong> How would you implement <em>custom comparator<\/em>?<br \/><strong>Pedro:<\/strong> You need to take <code class=\"highlighter-rouge\">Comparator<\/code> interface\nand provide implementation for <code class=\"highlighter-rouge\">compare(Object o1, Object o2)<\/code> method.\nAlso you need another implementation for <code class=\"highlighter-rouge\">ReverseComparator<\/code><br \/><strong>Eve:<\/strong> Stop talking, show me the code!<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class SubsComparator implements Comparator&lt;User&gt; {\n  @Override\n  public int compare(User u1, User u2) {\n    if (u1.isSubscription() == u2.isSubscription()) {\n      return u1.getName().compareTo(u2.getName());\n    } else if (u1.isSubscription()) {\n      return -1;\n    } else {\n      return 1;\n    }\n  }\n}\nclass ReverseSubsComparator implements Comparator&lt;User&gt; {\n  @Override\n  public int compare(User u1, User u2) {\n    if (u1.isSubscription() == u2.isSubscription()) {\n      return u2.getName().compareTo(u1.getName());\n    } else if (u1.isSubscription()) {\n      return -1;\n    } else {\n      return 1;\n    }\n  }\n}\n\/\/ forward sort\nCollections.sort(users, new SubsComparator());\n\/\/ reverse sort\nCollections.sort(users, new ReverseSubsComparator());\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Could you do the same?<br \/><strong>Eve:<\/strong> Yeah, something like that<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(sort (comparator \n       (fn [u1 u2]\n         (cond\n          (= (:subscription u1)\n             (:subscription u2)) (neg? (compare (:name u1)\n                                                (:name u2)))\n             (:subscription u1) true\n             :else false))) users)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Pretty similar.<br \/><strong>Eve:<\/strong> But we can do it better<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; forward sort\n(sort-by (juxt (complement :subscription) :name) users)\n;; reverse sort\n(sort-by (juxt :subscription :name) #(compare %2 %1) users)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Oh my gut! Monstrous oneliners.<br \/><strong>Eve:<\/strong> Functions, you know.<br \/><strong>Pedro:<\/strong> Whatever, it\u2019s very hard to understand what\u2019s happening there.<\/p><p><em>Eve explains juxt, complement and sort-by<\/em><br \/><em>10 minutes later<\/em><\/p><p><strong>Pedro:<\/strong> Very doubtful approach to pass strategy.<br \/><strong>Eve:<\/strong> I don\u2019t care, because Strategy is just a <strong>function passed to another function.<\/strong><\/p><h3 id=\"episode-3-state\">Episode 3. State<\/h3><blockquote>\n  <p>Sales person <strong>Karmen Git<\/strong>\ninvestigated the market and decided\nto provide user-specific functionality.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Smooth requirements.<br \/><strong>Eve:<\/strong> Let\u2019s clarify them.<\/p><ul><li><em>If user has subscription show him all news in a feed<\/em><\/li>\n  <li><em>Otherwise, show him only recent 10 news<\/em><\/li>\n  <li><em>If he pays money, add the amount to his account balance<\/em><\/li>\n  <li><em>If user doesn\u2019t have subscription and there is enough money\nto buy subscription, change his state to\u2026<\/em><\/li>\n<\/ul><p><strong>Pedro:<\/strong> State! Awesome pattern. First we make a user state enum<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public enum UserState {\n  SUBSCRIPTION(Integer.MAX_VALUE),\n  NO_SUBSCRIPTION(10);\n  private int newsLimit;\n  UserState(int newsLimit) {\n    this.newsLimit = newsLimit;\n  }\n  public int getNewsLimit() {\n    return newsLimit;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> User logic is following<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class User {\n  private int money = 0;\n  private UserState state = UserState.NO_SUBSCRIPTION;\n  private final static int SUBSCRIPTION_COST = 30;\n  public List&lt;News&gt; newsFeed() {\n    return DB.getNews(state.getNewsLimit());\n  }\n  public void pay(int money) {\n    this.money += money;\n    if (state == UserState.NO_SUBSCRIPTION\n\t    &amp;&amp; this.money &gt;= SUBSCRIPTION_COST) {\n      \/\/ buy subscription\n      state = UserState.SUBSCRIPTION;\n      this.money -= SUBSCRIPTION_COST;\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Lets call it<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>User user = new User(); \/\/ create default user\nuser.newsFeed(); \/\/ show him top 10 news\nuser.pay(10); \/\/ balance changed, not enough for subs\nuser.newsFeed(); \/\/ still top 10\nuser.pay(25); \/\/ balance enough to apply subscription\nuser.newsFeed(); \/\/ show him all news\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> You just hide value that affects behaviour inside <code class=\"highlighter-rouge\">User<\/code> object. We could use strategy to pass it directly <code class=\"highlighter-rouge\">user.newsFeed(subscriptionType)<\/code>.<br \/><strong>Pedro:<\/strong> Agreed, State is very close to the Strategy.\nThey even have the same UML diagrams. but we encapsulate balance and bind it to user.<br \/><strong>Eve:<\/strong> I think it achieves the same goal using another mechanism. Instead of providing strategy explicitly, it depends on some state. From clojure perspective it can be implemented\nthe same way as strategy pattern.<br \/><strong>Pedro:<\/strong> But successive calls can change object\u2019s state.<br \/><strong>Eve:<\/strong> Correct, but it has nothing to do with <code class=\"highlighter-rouge\">Strategy<\/code> it is just implementation detail.<br \/><strong>Pedro:<\/strong> What about \u201canother mechanism\u201d?<br \/><strong>Eve:<\/strong> Multimethods.<br \/><strong>Pedro:<\/strong> Multi <em>what<\/em>?<br \/><strong>Eve:<\/strong> Look at this<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defmulti news-feed :user-state)\n(defmethod news-feed :subscription [user]\n  (db\/news-feed))\n(defmethod news-feed :no-subscription [user]\n  (take 10 (db\/news-feed)))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And <code class=\"highlighter-rouge\">pay<\/code> function it\u2019s just a plain function, which changes state of object. We don\u2019t like state too much in clojure, but if you wish.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def user (atom {:name \"Jackie Brown\"\n                 :balance 0\n                 :user-state :no-subscription}))\n\t\t\t\t \n(def ^:const SUBSCRIPTION_COST 30)\n(defn pay [user amount]\n  (swap! user update-in [:balance] + amount)\n  (when (and (&gt;= (:balance @user) SUBSCRIPTION_COST)\n             (= :no-subscription (:user-state @user)))\n    (swap! user assoc :user-state :subscription)\n    (swap! user update-in [:balance] - SUBSCRIPTION_COST)))\n(news-feed @user) ;; top 10\n(pay user 10)\n(news-feed @user) ;; top 10\n(pay user 25)\n(news-feed @user) ;; all news\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Is dispatching by multimethods better than dispatching by enum?<br \/><strong>Eve:<\/strong> No, in this particlular case, but in general yes.<br \/><strong>Pedro:<\/strong> Explain, please<br \/><strong>Eve:<\/strong> Do you know what <em>double dispatch<\/em> is?<br \/><strong>Pedro:<\/strong> Not sure.<br \/><strong>Eve:<\/strong> Well, it is topic for <code class=\"highlighter-rouge\">Visitor<\/code> pattern.<\/p><h3 id=\"episode-4-visitor\">Episode 4. Visitor<\/h3><blockquote>\n  <p><strong>Natanius S. Selbys<\/strong> suggested to implement functionality\nwhich allows users export their messages, activities\nand achievements in different formats.<\/p>\n<\/blockquote><p><strong>Eve:<\/strong> So, how do you plan to do it?<br \/><strong>Pedro:<\/strong> We have one hierarchy for item types\n(Message, Activity) and another\nfor file formats (PDF, XML)<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>abstract class Format { }\nclass PDF extends Format { }\nclass XML extends Format { }\npublic abstract class Item {\n  void export(Format f) {\n    throw new UnknownFormatException(f);\n  }\n  abstract void export(PDF pdf);\n  abstract void export(XML xml);\n}\nclass Message extends Item {\n  @Override\n  void export(PDF f) {\n    PDFExporter.export(this);\n  }\n  @Override\n  void export(XML xml) {\n    XMLExporter.export(this);\n  }\n}\nclass Activity extends Item {\n  @Override\n  void export(PDF pdf) {\n    PDFExporter.export(this);\n  }\n  @Override\n  void export(XML xml) {\n    XMLExporter.export(this);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> That\u2019s all.<br \/><strong>Eve:<\/strong> Nice, but how do you dispatch on argument type?<br \/><strong>Pedro:<\/strong> What the problem?<br \/><strong>Eve:<\/strong> Consider this snippet<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Item i = new Activity();\nFormat f = new PDF();\ni.export(f);\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Nothing suspicious here.<br \/><strong>Eve:<\/strong> Actually, if you run this code you get <code class=\"highlighter-rouge\">UnknownFormatException<\/code><br \/><strong>Pedro:<\/strong> Wait\u2026Really?<br \/><strong>Eve:<\/strong> In java you can use only <em>single dispatch<\/em>. That means if you call <code class=\"highlighter-rouge\">i.export(f)<\/code>\nyou dispatches on the actual type of <code class=\"highlighter-rouge\">i<\/code>, not <code class=\"highlighter-rouge\">f<\/code>.<br \/><strong>Pedro:<\/strong> I\u2019m surprised. So, there is no dispatch on argument type?<br \/><strong>Eve:<\/strong> That\u2019s what visitor hack for. After you got a dispatch on <code class=\"highlighter-rouge\">i<\/code> type, you\nadditionally call <code class=\"highlighter-rouge\">f.someMethod(i)<\/code> and dispatched on <code class=\"highlighter-rouge\">f<\/code> type.<br \/><strong>Pedro:<\/strong> How that looks in code?<br \/><strong>Eve:<\/strong> You separately define export operations for all types as a <code class=\"highlighter-rouge\">Visitor<\/code><\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public interface Visitor {\n  void visit(Activity a);\n  void visit(Message m);\n}\npublic class PDFVisitor implements Visitor {\n  @Override\n  public void visit(Activity a) {\n    PDFExporter.export(a);\n  }\n  @Override\n  public void visit(Message m) {\n    PDFExporter.export(m);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Your items change signature to accept different visitors.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public abstract class Item {\n  abstract void accept(Visitor v);\n}\nclass Message extends Item {\n  @Override\n  void accept(Visitor v) {\n    v.visit(this);\n  }\n}\nclass Activity extends Item {\n  @Override\n  void accept(Visitor v) {\n    v.visit(this);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> To use it you may call<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Item i = new Message();\nVisitor v = new PDFVisitor(); \ni.accept(v);\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And everything works fine.\nMoreover, you can add new operations for\nactivities and messages\nby just defining new visitors and\nwithout changing their code.<br \/><strong>Pedro:<\/strong> That\u2019s really useful. But implementation is tough,\nit is the same for clojure?<br \/><strong>Eve:<\/strong> Not really, clojure supports it natively via multimethods<br \/><strong>Pedro:<\/strong> Multi <em>what<\/em>?<br \/><strong>Eve:<\/strong> Just follow the code\u2026\nFirst we define dispatcher <em>function<\/em><\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defmulti export\n  (fn [item format] [(:type item) format]))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> It accepts <code class=\"highlighter-rouge\">item<\/code> and <code class=\"highlighter-rouge\">format<\/code> to be exported. Examples:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; Message\n{:type :message :content \"Say what again!\"}\n;; Activity\n{:type :activity :content \"Quoting Ezekiel 25:17\"}\n;; Formats\n:pdf, :xml\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And now you just provide a functions\nfor different combinatations, and dispatcher decide which one to call.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defmethod export [:activity :pdf] [item format]\n  (exporter\/activity-&gt;pdf item))\n(defmethod export [:activity :xml] [item format]\n  (exporter\/activity-&gt;xml item))\n(defmethod export [:message :pdf] [item format]\n  (exporter\/message-&gt;pdf item))\n(defmethod export [:message :xml] [item format]\n  (exporter\/message-&gt;xml item))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> What if unknown format passed?<br \/><strong>Eve:<\/strong> We could specify default dipatcher function.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defmethod export :default [item format]\n  (throw (IllegalArgumentException. \"not supported\")))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Ok, but there is no hierarchy for <code class=\"highlighter-rouge\">:pdf<\/code> and <code class=\"highlighter-rouge\">:xml<\/code>.\nThey are just keywords?<br \/><strong>Eve:<\/strong> Correct, simple problem - simple solution.\nIf you need advanced features, you could use <em>adhoc hierarchies<\/em>\nor dispatch by <code class=\"highlighter-rouge\">class<\/code>.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(derive ::pdf ::format)\n(derive ::xml ::format)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Quadrocolons?!<br \/><strong>Eve:<\/strong> Assume they are just keywords.<br \/><strong>Pedro:<\/strong> Ok.<br \/><strong>Eve:<\/strong> Then you add functions for every dispatch type\n<code class=\"highlighter-rouge\">::pdf<\/code>, <code class=\"highlighter-rouge\">::xml<\/code> and <code class=\"highlighter-rouge\">::format<\/code><\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defmethod export [:activity ::pdf])\n(defmethod export [:activity ::xml])\n(defmethod export [:activity ::format])\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> If some new format (i.e. csv) appears in the system<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(derive ::csv ::format)\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> It will be dispatched to <code class=\"highlighter-rouge\">::format<\/code> function,\nuntil you add a separate <code class=\"highlighter-rouge\">::csv<\/code> function.<br \/><strong>Pedro:<\/strong> Seems good.<br \/><strong>Eve:<\/strong> Definitely, much easier.<br \/><strong>Pedro:<\/strong> So, basically, <strong>if a language support multiple dispatch,\nyou don\u2019t need Visitor pattern<\/strong>?<br \/><strong>Eve:<\/strong> Exactly.<\/p><h3 id=\"episode-5-template-method\">Episode 5. Template Method<\/h3><blockquote>\n  <p>MMORPG <strong>Mech Dominore Fight Saga<\/strong> requested to implement\na game bot for their VIP users. Not fair.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> First, we must decide what actions \nshould be automated with bot.<br \/><strong>Eve:<\/strong> Have you ever played RPG?<br \/><strong>Pedro:<\/strong> Fortunately, no<br \/><strong>Eve:<\/strong> Oh my\u2026 Let\u2019s go, I\u2019ll show you\u2026<\/p><p><em>2 weeks later<\/em><\/p><p><strong>Pedro:<\/strong> \u2026fantastic, I found epic sword, which has +100 attack.<br \/><strong>Eve:<\/strong> Unbelievable. But now, it\u2019s time for bot.<br \/><strong>Pedro:<\/strong> Easy-peasy. We could select following events<\/p><ul><li>Battle<\/li>\n  <li>Quest<\/li>\n  <li>Open Chest<\/li>\n<\/ul><p><strong>Pedro:<\/strong> Characters behave differently in\ndifferent events, for example mages cast spells in battle,\nbut rogues prefer silent melee combat; locked chests are skipped\nby most characters, but rogues can unlock them, etc.<br \/><strong>Eve:<\/strong> Looks like ideal candidate for <code class=\"highlighter-rouge\">Template Method<\/code>?<br \/><strong>Pedro:<\/strong> Yes. We define abstract algorithm, and then specify\ndifferences in subclasses.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public abstract class Character {\n  void moveTo(Location loc) {\n    if (loc.isQuestAvailable()) {\n      Journal.addQuest(loc.getQuest());\n    } else if (loc.containsChest()) {\n      handleChest(loc.getChest());\n    } else if (loc.hasEnemies()) {\n      attack(loc.getEnemies());\n    }\n    moveTo(loc.getNextLocation());\n  }\n  private void handleChest(Chest chest) {\n    if (!chest.isLocked()) {\n      chest.open();\n    } else {\n      handleLockedChest(chest);\n    }\n  }\n  abstract void handleLockedChest(Chest chest);\n  abstract void attack(List&lt;Enemy&gt; enemies);\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> We\u2019ve separated to <code class=\"highlighter-rouge\">Character<\/code> class everything common to all characters.\nNow we can create subclasses, that define how character should behave in specific situation.\nIn out case: <em>handling locked chests<\/em> and <em>attacking enemies<\/em>.<br \/><strong>Eve:<\/strong> Let\u2019s start with a Mage class.<br \/><strong>Pedro:<\/strong> Mage? Okay.\nHe can\u2019t open locked chest, so implementation is just <em>do nothing<\/em>.\nAnd if he is attacking enemies, if there are more than 10 enemies,\nfreeze them, and cast teleport to run away. If there are 10 enemies or less\ncast fireball on each of them.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class MageCharacter extends Character {\n  @Override\n  void handleLockedChest(Chest chest) {\n    \/\/ do nothing\n  }\n  @Override\n  void attack(List&lt;Enemy&gt; enemies) {\n    if (enemies.size() &gt; 10) {\n      castSpell(\"Freeze Nova\");\n      castSpell(\"Teleport\");\n    } else {\n      for (Enemy e : enemies) {\n        castSpell(\"Fireball\", e);\n      }\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Excellent, what about Rogue class?<br \/><strong>Pedro:<\/strong> Easy as well, rogues can unlock chests and\nprefer silent combat, handle enemies one by one.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class RogueCharacter extends Character {\n  @Override\n  void handleLockedChest(Chest chest) {\n    chest.unlock();\n  }\n  @Override\n  void attack(List&lt;Enemy&gt; enemies) {\n    for (Enemy e : enemies) {\n      invisibility();\n\t  attack(\"backstab\", e);\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Excellent. But how this approach is differrent from Strategy?<br \/><strong>Pedro:<\/strong> What?<br \/><strong>Eve:<\/strong> I mean, you redefined behaviour by using subclasses,\nbut in Strategy pattern you did the same:\nredefined behaviour by using functions.<br \/><strong>Pedro:<\/strong> Well, another approach.<br \/><strong>Eve:<\/strong> State was handled with another approach as well.<br \/><strong>Pedro:<\/strong> What are you trying to say?<br \/><strong>Eve:<\/strong> You are solving the same kind of problem,\nbut change the approach to it.<br \/><strong>Pedro:<\/strong> How do yo solve this problem using strategy in clojure?<br \/><strong>Eve:<\/strong> Just pass a set of specific functions for each character. For example, your abstract move may look like:<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn move-to [character location]\n  (cond\n   (quest? location)\n   (journal\/add-quest (:quest location))\n   (chest? location)\n   (handle-chest (:chest location))\n   (enemies? location)\n   (attack (:enemies location)))\n  (move-to character (:next-location location)))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> To add character-specific implementation\nof methods <code class=\"highlighter-rouge\">handle-chest<\/code> and <code class=\"highlighter-rouge\">attack<\/code>, implement them\nand pass as an argument.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; Mage-specific actions\n(defn mage-handle-chest [chest])\n(defn mage-attack [enemies]\n  (if (&gt; (count enemies) 10)\n    (do (cast-spell \"Freeze Nova\")\n        (cast-spell \"Teleport\"))\n    ;; otherwise\n    (doseq [e enemies]\n      (cast-spell \"Fireball\" e))))\n;; Signature of move-to will change to\n(defn move-to [character location\n               &amp; {:keys [handle-chest attack]\n                  :or {handle-chest (fn [chest])\n                       attack (fn [enemies] (run-away))}}]\n  ;; previous implementation\n)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> OMG, what\u2019s happening there?<br \/><strong>Eve:<\/strong> We changed signature of <code class=\"highlighter-rouge\">move-to<\/code> to accept\n<code class=\"highlighter-rouge\">handle-chest<\/code> and <code class=\"highlighter-rouge\">attack<\/code> functions.<br \/>Think of them like optional parameters.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(move-to character location\n  :handle-chest mage-handle-chest\n  :attack       mage-attack)\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Keep in mind that if these functions are not provided\nwe use default behavior: do nothing for <code class=\"highlighter-rouge\">handle-chest<\/code> and\nrun away from enemies in <code class=\"highlighter-rouge\">attack<\/code><br \/><strong>Pedro:<\/strong> Fine, but is this better than approach by subclassing? Seems that we have a lot of redundant information in\n<code class=\"highlighter-rouge\">move-to<\/code> call.<br \/><strong>Eve:<\/strong> It\u2019s fixable, just define this call once, and give it\nalias<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn mage-move [character location]\n  (move-to character location\n    :handle-chest mage-handle-chest\n    :attack       mage-attack))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Or use multimethods, it\u2019s even better.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defmulti move \n  (fn [character location] (:class character)))\n(defmethod move :mage [character location]\n  (move-to character location\n    :handle-chest mage-handle-chest\n    :attack       mage-attack))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> I understand. But why do you think pass as argument is better than subclassing?<br \/><strong>Eve:<\/strong> You can change behaviour dynamically. Assume your mage\nhas no mana, so instead of trying to cast fireballs, he can just teleport and run away, you just provide new function.<br \/><strong>Pedro:<\/strong> Makes sense. <strong>Functions everywhere<\/strong>.<\/p><h3 id=\"episode-6-iterator\">Episode 6. Iterator<\/h3><blockquote>\n  <p>Technical consultant <strong>Kent Podiololis<\/strong>\ncomplains for C-style loops usage.<\/p>\n  <p>\u201cAre we in 1980 or what?\u201d \u2013 Kent<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> We definitely should use pattern Iterator from java.<br \/><strong>Eve:<\/strong> Don\u2019t be fool, nobody\u2019s using <code class=\"highlighter-rouge\">java.util.Iterator<\/code><br \/><strong>Pedro:<\/strong> Everybody use it implicitly in <code class=\"highlighter-rouge\">for-each<\/code> loop. It\u2019s a good way to traverse a container.<br \/><strong>Eve:<\/strong> What does it mean <em>\u201cto traverse a container\u201d<\/em>?<br \/><strong>Pedro:<\/strong> Formally, the container should provide two methods for you:<br \/><code class=\"highlighter-rouge\">next()<\/code> to return next element and <code class=\"highlighter-rouge\">hasNext()<\/code> to return true if container has more elements.<br \/><strong>Eve:<\/strong> Ok. Do you know what linked list is?<br \/><strong>Pedro:<\/strong> Singly linked list?<br \/><strong>Eve:<\/strong> Singly linked list.<br \/><strong>Pedro:<\/strong> Sure. It is a container consists of nodes. \nEach node has a data value and reference to the next node.\nAnd <code class=\"highlighter-rouge\">null<\/code> value if there is no next node.<br \/><strong>Eve:<\/strong> Correct. Now tell me how traversing such list is differ\nfrom traversing via iterator?<br \/><strong>Pedro:<\/strong> Emmm\u2026<\/p><p>Pedro wrote two traversing snippets:<\/p><ul><li>Traversing using iterator<\/li>\n<\/ul><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Iterator i;\nwhile (i.hasNext()) {\n  i.next();\n}\n<\/pre><\/div><\/div><ul><li>Traversing using linked list<\/li>\n<\/ul><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Node next = root;\nwhile (next != null) {\n  next = next.next;\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> They are pretty similar\u2026What is analogue of <code class=\"highlighter-rouge\">Iterator<\/code> in clojure?<br \/><strong>Eve:<\/strong> <code class=\"highlighter-rouge\">seq<\/code> function.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(seq [1 2 3])       =&gt; (1 2 3)\n(seq (list 4 5 6))  =&gt; (4 5 6)\n(seq #{7 8 9})      =&gt; (7 8 9)\n(seq (int-array 3)) =&gt; (0 0 0)\n(seq \"abc\")         =&gt; (\\a \\b \\c)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> It returns a list\u2026<br \/><strong>Eve:<\/strong> Sequence, because <strong>Iterator is just a sequence<\/strong><br \/><strong>Pedro:<\/strong> Is it possible to make <code class=\"highlighter-rouge\">seq<\/code> works on custom datastructures?<br \/><strong>Eve:<\/strong> Implement <code class=\"highlighter-rouge\">clojure.lang.Seqable<\/code> interface<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(deftype RedGreenBlackTree [&amp; elems]\n  clojure.lang.Seqable\n  (seq [self]\n\t;; traverse element in needed order\n\t))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Fine then. But I\u2019ve heard iterator is often used to achive <em>laziness<\/em>, for example\nto calculate value only during <code class=\"highlighter-rouge\">getNext()<\/code> call, how list handle that?<br \/><strong>Eve:<\/strong> List can be lazy as well, clojure calls such list <em>\u201clazy sequence\u201d<\/em>.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def natural-numbers (iterate inc 1))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> We defined thing to represent <em>ALL<\/em> natural numbers, but we haven\u2019t got <code class=\"highlighter-rouge\">OutOfMemory<\/code> yet,\nbecause we haven\u2019t requested any value. It\u2019s lazy.<br \/><strong>Pedro:<\/strong> Could you explain more?<br \/><strong>Eve:<\/strong> Unfortunately, I am too lazy for that.<br \/><strong>Pedro:<\/strong> I will remember that!<\/p><h3 id=\"episode-7-memento\">Episode 7: Memento<\/h3><blockquote>\n  <p>User <strong>Chad Bogue<\/strong> lost the message he was writing\nfor two days. Implement save button for him.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> I don\u2019t believe there are people who\ncan type in textbox for two days. Two. Days.<br \/><strong>Eve:<\/strong> Let\u2019s <em>save<\/em> him.<br \/><strong>Pedro:<\/strong> I <em>googled<\/em> this problem. Most popular approach\nto implement save button is Memento pattern.\nYou need <em>originator<\/em>, <em>caretaker<\/em> and <em>memento<\/em> objects.<br \/><strong>Eve:<\/strong> What\u2019s that?<br \/><strong>Pedro:<\/strong> <em>Originator<\/em> is just an object or state that we want to preserve.\n(text inside a textbox), <em>caretaker<\/em> is responsible to save state (save button)\nand <em>memento<\/em> is just an object to encapsulate state.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class TextBox {\n  \/\/ state for memento\n  private String text = \"\";\n  \/\/ state not handled by memento\n  private int width = 100;\n  private Color textColor = Color.BLACK;\n  public void type(String s) {\n    text += s;\n  }\n  public Memento save() {\n    return new Memento(text);\n  }\n  public void restore(Memento m) {\n    this.text = m.getText();\n  }\n  @Override\n  public String toString() {\n    return \"[\" + text + \"]\";\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Memento is just an immutable object<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public final class Memento {\n  private final String text;\n  public Memento(String text) {\n    this.text = text;\n  }\n  public String getText() {\n    return text;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And caretaker is a demo<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ open browser, init empty textbox\nTextBox textbox = new TextBox();\n\/\/ type something into it\ntextbox.type(\"Dear, Madonna\\n\");\ntextbox.type(\"Let me tell you what \");\n\/\/ press button save\nMemento checkpoint1 = textbox.save();\n\/\/ type again\ntextbox.type(\"song 'Like A Virgin' is about. \");\ntextbox.type(\"It's all about a girl...\");\n\/\/ suddenly browser crashed, restart it, reinit textbox\ntextbox = new TextBox();\n\/\/ but it's empty! All work is gone!\n\/\/ not really, you rollback to last checkpoint\ntextbox.restore(checkpoint1);\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Just a note if you want a multiple checkpoints, save memento\u2019s to the list.<br \/><strong>Eve:<\/strong> <em>Originator<\/em>, <em>caretaker<\/em>, <em>memento<\/em> - looks as a bunch of nouns, but actually it\u2019s all about two functions <code class=\"highlighter-rouge\">save<\/code> and <code class=\"highlighter-rouge\">restore<\/code>.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def textbox (atom {}))\n(defn init-textbox [] \n (reset! textbox {:text \"\"\n                  :color :BLACK\n                  :width 100}))\n(def memento (atom nil))\n(defn type-text [text]\n  (swap! textbox\n    (fn [m]\n      (update-in m [:text] (fn [s] (str s text))))))\n(defn save []\n  (reset! memento (:text @textbox)))\n(defn restore []\n  (swap! textbox assoc :text @memento))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And demo as well.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(init-textbox)\n(type-text \"'Like A Virgin' \")\n(type-text \"it's not about this sensitive girl \")\n(save)\n(type-text \"who meets nice fella\")\n;; crash\n(init-textbox)\n(restore)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> It\u2019s pretty the same code.<br \/><strong>Eve:<\/strong> Yes, but you must care about memento immutability<br \/><strong>Pedro:<\/strong> What does it mean?<br \/><strong>Eve:<\/strong> You are lucky, that you got <code class=\"highlighter-rouge\">String<\/code> object in this example,\n<code class=\"highlighter-rouge\">String<\/code> is immutable. But if you have something, that\nmay change its internal state, you need to perform deep copy of this object for memento.<br \/><strong>Pedro:<\/strong> Oh, right. It\u2019s just a recursive <code class=\"highlighter-rouge\">clone()<\/code> calls to\nobtain prototype.<br \/><strong>Eve:<\/strong> We will talk about Prototype in a minute, but just\nremember that <code class=\"highlighter-rouge\">Memento<\/code> is not about <em>caretaker<\/em> and <em>originator<\/em>, it is about <strong>save<\/strong> and <strong>restore<\/strong>.<\/p><h3 id=\"episode-8-prototype\">Episode 8: Prototype<\/h3><blockquote>\n  <p><strong>Dex Ringeus<\/strong> detected that users feel uncomfortable\nwith registration form. Make it more usable.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> So, what\u2019s the problem with the registration?<br \/><strong>Eve:<\/strong> There are lot of fields users bored to type in.<br \/><strong>Pedro:<\/strong> For example?<br \/><strong>Eve:<\/strong> For example, <code class=\"highlighter-rouge\">weight<\/code>. Having such field scares\n90% of female users.<br \/><strong>Pedro:<\/strong> But this field is important for our analytics\nsystem, we make food and clothes recomendations based on that field.<br \/><strong>Eve:<\/strong> Then, make it optional, and if it is not provided, take\nsome default value.<br \/><strong>Pedro:<\/strong> <code class=\"highlighter-rouge\">60 kg<\/code> is ok?<br \/><strong>Eve:<\/strong> I think so.<br \/><strong>Pedro:<\/strong> Ok, give me two minutes.<\/p><p><em>2 hours later<\/em><\/p><p><strong>Pedro:<\/strong> I suggest to use some registration <em>prototype<\/em>\nwhich has all fields are filled with default values. \nAfter user completes the form we modify filled values.<br \/><strong>Eve:<\/strong> Sounds great.<br \/><strong>Pedro:<\/strong> Here it is our standard registration form, with prototype in <code class=\"highlighter-rouge\">clone()<\/code> method.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class RegistrationForm implements Cloneable {\n  private String name = \"Zed\";\n  private String email = \"zzzed@gmail.com\";\n  private Date dateOfBirth = new Date(1970, 1, 1);\n  private int weight = 60;\n  private Gender gender = Gender.MALE;\n  private Status status = Status.SINGLE;\n  private List&lt;Child&gt; children = Arrays.asList(new Child(Gender.FEMALE));\n  private double monthSalary = 1000;\n  private List&lt;Brand&gt; favouriteBrands = Arrays.asList(\"Adidas\", \"GAP\");\n  \/\/ few hundreds more properties\n  @Override\n  protected RegistrationForm clone() throws CloneNotSupportedException {\n    RegistrationForm prototyped = new RegistrationForm();\n      prototyped.name = name;\n      prototyped.email = email;\n      prototyped.dateOfBirth = (Date)dateOfBirth.clone();\n      prototyped.weight = weight;\n      prototyped.status = status;\n      List&lt;Child&gt; childrenCopy = new ArrayList&lt;Child&gt;();\n      for (Child c : children) {\n        childrenCopy.add(c.clone());\n      }\n      prototyped.children = childrenCopy;\n      prototyped.monthSalary = monthSalary;\n      List&lt;String&gt; brandsCopy = new ArrayList&lt;String&gt;();\n      for (String s : favouriteBrands) {\n        brandsCopy.add(s);\n      }\n      prototyped.favouriteBrands = brandsCopy;\n    return  prototyped;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Every time we create a user, call <code class=\"highlighter-rouge\">clone()<\/code> and then override needed properties.<br \/><strong>Eve:<\/strong> Awful! In mutable world <code class=\"highlighter-rouge\">clone()<\/code> is needed to create\nnew object with the same properties.\nThe hard part is the copy must be deep, i.e. instead of copying\nreference you need recursively <code class=\"highlighter-rouge\">clone()<\/code> other objects,\nand what if one of them doesn\u2019t have <code class=\"highlighter-rouge\">clone()<\/code>\u2026<br \/><strong>Pedro:<\/strong> That\u2019s the problem and this pattern solves it.<br \/><strong>Eve:<\/strong> I don\u2019t think it is a solution if you need to implement\nclone every time you adding new object.<br \/><strong>Pedro:<\/strong> How clojure avoid this?<br \/><strong>Eve:<\/strong> Clojure has immutable data structures. That\u2019s all.<br \/><strong>Pedro:<\/strong> How does it solve prototype problem?<br \/><strong>Eve:<\/strong> Every time you modify object, you get\na fresh new immutable copy of your data, and old one is not changed. <strong>Prototype is not needed in immutable world<\/strong><\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def registration-prototype\n\t {:name          \"Zed\"\n\t  :email         \"zzzed@gmail.com\"\n\t  :date-of-birth \"1970-01-01\"\n\t  :weight        60\n\t  :gender        :male\n\t  :status        :single\n\t  :children      [{:gender :female}]\n\t  :month-salary  1000\n\t  :brands        [\"Adidas\" \"GAP\"]})\n;; return new object\n(assoc registration-prototype \n\t :name \"Mia Vallace\"\n\t :email \"tomato@gmail.com\"\n\t :weight 52\n\t :gender :female\n\t :month-salary 0)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Great! But how this affects performance?<br \/>Copying million rows each time you adding new value seems\ntime consuming operation.<br \/><strong>Eve:<\/strong> No, it is not. Go to Google and search for <em>persistent data structures<\/em> and <em>structural sharing<\/em><br \/><strong>Pedro:<\/strong> Thanks a lot.<\/p><h3 id=\"episode-9-mediator\">Episode 9: Mediator<\/h3><blockquote>\n  <p>Recently performed external code review\nshows a lot of issues with current codebase.\n<strong>Veerco Wierde<\/strong> emphasizes tight coupling \nin chat application.<\/p>\n<\/blockquote><p><strong>Eve:<\/strong> What is tight coupling?<br \/><strong>Pedro:<\/strong> It\u2019s the problem when objects\nknow too much about each other.<br \/><strong>Eve:<\/strong> Could you be more specific?<br \/><strong>Pedro:<\/strong> Look at the current chat implementation<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class User {\n  private String name;\n  List&lt;User&gt; users = new ArrayList&lt;User&gt;();\n  public User(String name) {\n    this.name = name;\n  }\n  public void addUser(User u) {\n    users.add(u);\n  }\n  void sendMessage(String message) {\n    String text = String.format(\"%s: %s\\n\", name, message);\n    for (User u : users) {\n      u.receive(text);\n    }\n  }\n  private void receive(String message) {\n    \/\/ process message\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> The problem here is the user knows\neverything about other users. It is very hard to use and maintain such code.\nWhen new user connects to the chat, you must add a reference to him via <code class=\"highlighter-rouge\">addUser<\/code>\nfor every existing user.<br \/><strong>Eve:<\/strong> So, we just move one piece of responsibility to another class?<br \/><strong>Pedro:<\/strong> Yes, kind of. We create <em>mega-aware<\/em> class, called mediator, that binds\nall parts together. Obviously, each part knows only about mediator.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class User {\n  String name;\n  private Mediator m;\n  public User(String name, Mediator m) {\n    this.name = name;\n    this.m = m;\n  }\n  public void sendMessage(String text) {\n    m.sendMessage(this, text);\n  }\n  public void receive(String text) {\n    \/\/ process message\n  }\n}\npublic class Mediator {\n  List&lt;User&gt; users = new ArrayList&lt;User&gt;();\n  public void addUser(User u) {\n    users.add(u);\n  }\n  public void sendMessage(User u, String text) {\n    for (User user : users) {\n      u.receive(text);\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> That seems like a simple refactoring problem.<br \/><strong>Pedro:<\/strong> Profit may be underestimated, but\nif you have hundreds of mutually connected components\n(UI for example) mediator is really a savior.<br \/><strong>Eve:<\/strong> Agreed.<br \/><strong>Pedro:<\/strong> Now the clojure turn.<br \/><strong>Eve:<\/strong> Ok\u2026let\u2019s look\u2026your mediator is responsible\nfor <em>saving users<\/em> and <em>sending messages<\/em><\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def mediator\n  (atom {:users []\n         :send (fn [users text]\n\t\t\t     (map #(receive % text) users))}))\n(defn add-user [u]\n\t(swap! mediator \n\t  (fn [m]\n\t    (update-in m [:users] conj u))))\n(defn send-message [u text]\n\t(let [send-fn (:send @mediator)\n\t      users (:users @mediator)]\n\t  (send-fn users (format \"%s: %s\\n\" (:name u) text))))\n(add-user {:name \"Mister White\"})\n(add-user {:name \"Mister Pink\"})\n(send-message {:name \"Joe\"} \"Toby?\")\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Good enough.<br \/><strong>Eve:<\/strong> Nothing interesing here, because it is just a one approach <strong>to reduce coupling<\/strong>.<\/p><h3 id=\"episode-10-observer\">Episode 10: Observer<\/h3><blockquote>\n  <p>Independent security commision\ndetected hacker <strong>Dartee Hebl<\/strong> has over a billion\ndollars balance on his account.\nTrack big transactions to the accounts.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Are we <em>Holmses<\/em>?<br \/><strong>Eve:<\/strong> No, but there is a lack of logging in the system,\nneed to find a way to track all changes to balance.<br \/><strong>Pedro:<\/strong> We could add observers. Every time we modify\nbalance, if change is <em>big enough<\/em>, notify about this and track the reason. We need <code class=\"highlighter-rouge\">Observer<\/code> interface<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public interface Observer {\n  void notify(User u);\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And two specific observers<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class MailObserver implements Observer {\n  @Override\n  public void notify(User user) {\n    MailService.sendToFBI(user);\n  }\n}\nclass BlockObserver implements Observer {\n  @Override\n  public void notify(User u) {\n    DB.blockUser(u);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> <code class=\"highlighter-rouge\">Tracker<\/code> class would be responsible\nfor managing observers.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class Tracker {\n  private Set&lt;Observer&gt; observers = new HashSet&lt;Observer&gt;();\n  public void add(Observer o) {\n    observers.add(o);\n  }\n  public void update(User u) {\n    for (Observer o : observers) {\n      o.notify(u);\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And the last part: init tracker with the user and\nmodify its <code class=\"highlighter-rouge\">addMoney<\/code> method. If transcation amount is greaterthan <code class=\"highlighter-rouge\">100$<\/code>, notify FBI and block this user.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class User {\n  String name;\n  double balance;\n  Tracker tracker;\n  public User() {\n    initTracker();\n  }\n  private void initTracker() {\n    tracker = new Tracker();\n    tracker.add(new MailObserver());\n    tracker.add(new BlockObserver());\n  }\n  public void addMoney(double amount) {\n    balance += amount;\n    if (amount &gt; 100) {\n      tracker.update(this);\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Why are you created two separate observers?\nYou could use it in a one.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class MailAndBlock implements Observer {\n  @Override\n  public void notify(User u) {\n    MailService.sendToFBI(u);\n    DB.blockUser(u);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Single responsibility principle.<br \/><strong>Eve:<\/strong> Oh, yeah.<br \/><strong>Pedro:<\/strong> And you can compose observer functionality\ndynamically.<br \/><strong>Eve:<\/strong> I see your point.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; Tracker\n(def observers (atom #{}))\n(defn add [observer]\n  (swap! observers conj observer))\n(defn notify [user]\n  (map #(apply % user) @observers))\n;; Fill Observers\n(add (fn [u] (mail-service\/send-to-fbi u)))\n(add (fn [u] (db\/block-user u)))\n;; User\n(defn add-money [user amount]\n  (swap! user\n    (fn [m]\n      (update-in m [:balance] + amount)))\n  ;; tracking\n  (if (&gt; amount 100) (notify)))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> It\u2019s a pretty the same way?<br \/><strong>Eve:<\/strong> Yeah, in fact observer is just a way to register <strong>function, which will be called after another function<\/strong>.<br \/><strong>Pedro:<\/strong> It is still the pattern.<br \/><strong>Eve<\/strong> Sure, but we can improve solution a bit using clojure watches.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(add-watch\n  user\n  :money-tracker \n  (fn [k r os ns] \n    (if (&lt; 100 (- (:balance ns) (:balance os)))\n      (notify))))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Why is that better?<br \/><strong>Eve:<\/strong> First of all, our <code class=\"highlighter-rouge\">add-money<\/code> function is clean, it just adds money. Also, watcher tracks <em>every<\/em> change to the state, not the ones we handle in mutator functions, like <code class=\"highlighter-rouge\">add-money<\/code><br \/><strong>Pedro:<\/strong> Explain please.<br \/><strong>Eve:<\/strong> If there is provided another secred method <code class=\"highlighter-rouge\">secret-add-money<\/code> for changing balance, watchers will handle that as well.<br \/><strong>Pedro:<\/strong> That\u2019s awesome!<\/p><h3 id=\"episode-11-interpreter\">Episode 11: Interpreter<\/h3><blockquote>\n  <p><strong>Bertie Prayc<\/strong> stole important data from our server\nand shared it via BitTorrent system.\nCreate a fake account for <em>Bertie<\/em> to discredit him.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> BitTorrent system based on <code class=\"highlighter-rouge\">.torrent<\/code> files.\nWe need to write <a href=\"http:\/\/en.wikipedia.org\/wiki\/Bencode\">Bencode<\/a> encoder.<br \/><strong>Eve:<\/strong> Yes, but first let\u2019s agree on the format spec<\/p><p>Bencode encoding rules:<\/p><ul><li>Two datatypes are supported:\n    <ul><li>Integer <code class=\"highlighter-rouge\">N<\/code> is encoded as <code class=\"highlighter-rouge\">i&lt;N&gt;e<\/code>. (42 = i42e)<\/li>\n      <li>String <code class=\"highlighter-rouge\">S<\/code> is encoded as <code class=\"highlighter-rouge\">&lt;length&gt;:&lt;contents&gt;<\/code> (hello = 5:hello)<\/li>\n    <\/ul><\/li>\n  <li>Two containers are supported:\n    <ul><li>List of values is encoded as <code class=\"highlighter-rouge\">l&lt;contents&gt;e<\/code> ([1, \u201cBye\u201d] = li1e3:Byee)<\/li>\n      <li>Map of values is encoded as <code class=\"highlighter-rouge\">d&lt;contents&gt;e<\/code> ({\u201cR\u201d 2, \u201cD\u201d 2} = d1:Ri2e1:Di2ee)\n        <ul><li>Keys are just strings, Values are any bencode element<\/li>\n        <\/ul><\/li>\n    <\/ul><\/li>\n<\/ul><p><strong>Pedro:<\/strong> Seems easy.<br \/><strong>Eve:<\/strong> Maybe, but take into account that values may be nested, list inside list, etc.<br \/><strong>Pedro:<\/strong> Sure. I think we can use <em>Interpreter<\/em> pattern for bencode encoding.<br \/><strong>Eve:<\/strong> Try it.<br \/><strong>Pedro:<\/strong> We start from interface for all bencode elements<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>interface BencodeElement {\n  String interpret();\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Then we provide implementation for each datatype and datacontainer<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class IntegerElement implements BencodeElement {\n  private int value;\n  public IntegerElement(int value) {\n    this.value = value;\n  }\n  @Override\n  public String interpret() {\n    return \"i\" + value + \"e\";\n  }\n}\nclass StringElement implements BencodeElement {\n  private String value;\n  StringElement(String value) {\n    this.value = value;\n  }\n  @Override\n  public String interpret() {\n    return value.length() + \":\" + value;\n  }\n}\nclass ListElement implements BencodeElement {\n  private List&lt;? extends BencodeElement&gt; list;\n  ListElement(List&lt;? extends BencodeElement&gt; list) {\n    this.list = list;\n  }\n  @Override\n  public String interpret() {\n    String content = \"\";\n    for (BencodeElement e : list) {\n      content += e.interpret();\n    }\n    return \"l\" + content + \"e\";\n  }\n}\nclass DictionaryElement implements BencodeElement {\n  private Map&lt;StringElement, BencodeElement&gt; map;\n  DictionaryElement(Map&lt;StringElement, BencodeElement&gt; map) {\n    this.map = map;\n  }\n  @Override\n  public String interpret() {\n    String content = \"\";\n    for (Map.Entry&lt;StringElement, BencodeElement&gt; kv : map.entrySet()) {\n      content += kv.getKey().interpret() + kv.getValue().interpret();\n    }\n    return \"d\" + content + \"e\";\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And finally, our bencoded string can be constructed from common datastructures programmatically<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>\/\/ discredit user\nMap&lt;StringElement, BencodeElement&gt; mainStructure = new HashMap&lt;StringElement, BencodeElement&gt;();\n\/\/ our victim\nmainStructure.put(new StringElement(\"user\"), new StringElement(\"Bertie\"));\n\/\/ just downloads files\nmainStructure.put(new StringElement(\"number_of_downloaded_torrents\"), new IntegerElement(623));\n\/\/ and nothing uploads\nmainStructure.put(new StringElement(\"number_of_uploaded_torrents\"), new IntegerElement(0));\n\/\/ and nothing donates\nmainStructure.put(new StringElement(\"donation_in_dollars\"), new IntegerElement(0));\n\/\/ prefer dirty categories\nmainStructure.put(new StringElement(\"preffered_categories\"),\n                      new ListElement(Arrays.asList(\n                          new StringElement(\"porn\"),\n                          new StringElement(\"murder\"),\n                          new StringElement(\"scala\"),\n                          new StringElement(\"pokemons\")\n                      )));\nBencodeElement top = new DictionaryElement(mainStructure);\n\/\/ let's totally discredit him\nString bencodedString = top.interpret();\nBitTorrent.send(bencodedString);\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Interesting, but that is ton of code!<br \/><strong>Pedro:<\/strong> We pay readability for capabilities.<br \/><strong>Eve:<\/strong> I suppose you\u2019ve heard concept <strong>Code is Data<\/strong>, that\u2019s a lot easier in clojure<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; multimethod to handle bencode structure\n(defmulti interpret class)\n;; implementation of bencode handler for each type\n(defmethod interpret java.lang.Long [n]\n  (str \"i\" n \"e\"))\n(defmethod interpret java.lang.String [s]\n  (str (count s) \":\" s))\n(defmethod interpret clojure.lang.PersistentVector [v]\n  (str \"l\" \n       (apply str (map interpret v))\n\t   \"e\"))\n(defmethod interpret clojure.lang.PersistentArrayMap [m]\n  (str \"d\" \n       (apply str (map (fn [[k v]] \n\t                     (str (interpret k)\n                              (interpret v))) m))\n       \"e\"))\n;; usage\n(interpret {\"user\" \"Bertie\"\n            \"number_of_downloaded_torrents\" 623\n\t\t\t\"number_of_uploaded_torrent\" 0\n\t\t\t\"donation_in_dollars\" 0\n\t\t\t\"preffered_categories\" [\"porn\"\n                                    \"murder\"\n\t\t\t                        \"scala\"\n\t\t\t\t\t\t\t\t\t\"pokemons\"]})\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> You see how it is much easier define specific data?<br \/><strong>Pedro:<\/strong> Sure, and <code class=\"highlighter-rouge\">interpret<\/code> it\u2019s just a <strong>function<\/strong> per bencode type, instead of separate class.<br \/><strong>Eve:<\/strong> Correct, interpreter is nothing but a <strong>set of functions to process a tree<\/strong>.<\/p><h3 id=\"episode-12-flyweight\">Episode 12: Flyweight<\/h3><blockquote>\n  <p>Administrators of the lawyer firm <strong>Cristopher, Matton &amp; Pharts<\/strong>\ndetected that reporting system consumes a lot of memory\nand garbage collector continiously hangs system for seconds. Fix that.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> I\u2019ve seen this issue before.<br \/><strong>Eve:<\/strong> What\u2019s wrong?<br \/><strong>Pedro:<\/strong> They have realtime charts with a lot of different points.\nIt\u2019s a huge amount of memory. As a result garbage collector stops the system.<br \/><strong>Eve:<\/strong> Hmmm, what we can do?<br \/><strong>Pedro:<\/strong> Not much, caching does not help us because points are\u2026<br \/><strong>Eve:<\/strong> Wait!<br \/><strong>Pedro:<\/strong> What?<br \/><strong>Eve:<\/strong> They use age values for points, why not precompute these points for most common ages?\nSay for age [0, 100]<br \/><strong>Pedro:<\/strong> You mean use <em>Flyweight<\/em> pattern?<br \/><strong>Eve:<\/strong> I mean reuse objects.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class Point {\n  int x;\n  int y;\n  \/* some other properties*\/\n  \/\/ precompute 10000 point values at class loading time\n  private static Point[][] CACHED;\n  static {\n    CACHED = new Point[100][];\n    for (int i = 0; i &lt; 100; i++) {\n      CACHED[i] = new Point[100];\n      for (int j = 0; j &lt; 100; j++) {\n        CACHED[i][j] = new Point(i, j);\n      }\n    }\n  }\n  Point(int x, int y) {\n    this.x = x;\n    this.y = y;\n  }\n  static Point makePoint(int x, int y) {\n    if (x &gt;= 0 &amp;&amp; x &lt; 100 &amp;&amp;\n        y &gt;= 0 &amp;&amp; y &lt; 100) {\n      return CACHED[x][y];\n    } else {\n      return new Point(x, y);\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> For this pattern we need two things: precompute\nmost used points at a startup time, and use static factory\nmethod instead of constructor to return cached object.<br \/><strong>Eve:<\/strong> Have you tested it?<br \/><strong>Pedro:<\/strong> Sure, the system works like a clock.<br \/><strong>Eve:<\/strong> Excellent, here is my version<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn make-point [x y]\n  [x y {:some \"Important Properties\"}])\n(def CACHE\n  (let [cache-keys (for [i (range 100) j (range 100)] [i j])]\n\t  (zipmap cache-keys (map #(apply make-point %) cache-keys))))\n(defn make-point-cached [x y]\n  (let [result (get CACHE [x y])]\n    (if result\n\t  result\n\t  (make-point x y))))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> It creates a flat map with pair <code class=\"highlighter-rouge\">[x, y]<\/code> as a key, instead of two-dimensional array.<br \/><strong>Pedro:<\/strong> Pretty the same.<br \/><strong>Eve:<\/strong> No, it is much flexible, you can\u2019t use two-dimensional array if you need to cache three points or non-integer values.<br \/><strong>Pedro:<\/strong> Oh, got it.<br \/><strong>Eve:<\/strong> Even better, in clojure you can just use <code class=\"highlighter-rouge\">memoize<\/code> function to cache calls to factory function <code class=\"highlighter-rouge\">make-point<\/code><\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def make-point-memoize (memoize make-point))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Every call (except first one) with the same parameters\nreturn cached value.<br \/><strong>Pedro:<\/strong> That\u2019s awesome!<br \/><strong>Eve:<\/strong> Of course, but remember <strong>if your function has side-effects, memoization is bad idea<\/strong>.<\/p><h3 id=\"episode-13-builder\">Episode 13: Builder<\/h3><blockquote>\n  <p><strong>Tuck Brass<\/strong> complains that his old automatic\ncoffee-making system is very slow in usage.\nCustomers can\u2019t wait long enough and going away.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Need to understand what\u2019s the problem exactly?<br \/><strong>Eve:<\/strong> I\u2019ve researched, system is old, written in <em>COBOL<\/em> and built around expert system <em>question-answer<\/em>. They were very popular long time ago.<br \/><strong>Pedro:<\/strong> What do you mean by <em>\u201cquestion-answer\u201d<\/em><br \/><strong>Eve:<\/strong> There is operator in front of terminal. System asks: <em>\u201cDo yo want to add water?\u201d<\/em>, operator answers <em>\u201cYes\u201d<\/em>. Then system asks again: <em>\u201cDo you want to add coffee\u201d<\/em>, operator answers <em>\u201cYes\u201d<\/em> and so forth.<br \/><strong>Pedro:<\/strong> It\u2019s a nightmare, I just want coffee with milk. Why they don\u2019t use predefined options: coffee with milk, coffee with sugar and so on.<br \/><strong>Eve:<\/strong> Because it is the raisin of the system: customer can make coffee with any set of ingridients <em>by themselves<\/em><br \/><strong>Pedro:<\/strong> Okay, let\u2019s fix it with Builder pattern.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class Coffee {\n  private String coffeeName; \/\/ required\n  private double amountOfCoffee; \/\/ required\n  private double water; \/\/ required\n  private double milk; \/\/ optional\n  private double sugar; \/\/ optional\n  private double cinnamon; \/\/ optional\n  private Coffee() { }\n  public static class Builder {\n    private String builderCoffeeName;\n    private double builderAmountOfCoffee; \/\/ required\n    private double builderWater; \/\/ required\n    private double builderMilk; \/\/ optional\n    private double builderSugar; \/\/ optional\n    private double builderCinnamon; \/\/ optional\n    public Builder() { }\n    public Builder setCoffeeName(String name) {\n      this.builderCoffeeName = name;\n      return this;\n    }\n    public Builder setCoffee(double coffee) {\n      this.builderAmountOfCoffee = coffee;\n      return this;\n    }\n    public Builder setWater(double water) {\n      this.builderWater = water;\n      return this;\n    }\n    public Builder setMilk(double milk) {\n      this.builderMilk = milk;\n      return this;\n    }\n    public Builder setSugar(double sugar) {\n      this.builderSugar = sugar;\n      return this;\n    }\n    public Builder setCinnamon(double cinnamon) {\n      this.builderCinnamon = cinnamon;\n      return this;\n    }\n    public Coffee make() {\n      Coffee c = new Coffee();\n        c.coffeeName = builderCoffeeName;\n        c.amountOfCoffee = builderAmountOfCoffee;\n        c.water = builderWater;\n        c.milk = builderMilk;\n        c.sugar = builderSugar;\n        c.cinnamon = builderCinnamon;\n      \/\/ check required parameters and invariants\n      if (c.coffeeName == null || c.coffeeName.equals(\"\") ||\n          c.amountOfCoffee &lt;= 0 || c.water &lt;= 0) {\n        throw new IllegalArgumentException(\"Provide required parameters\");\n      }\n      return c;\n    }\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> As you see, you can\u2019t instantiate <code class=\"highlighter-rouge\">Coffee<\/code> class easily, you need to set parameters with nested <code class=\"highlighter-rouge\">Builder<\/code> class<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Coffee c = new Coffee.Builder()\n        .setCoffeeName(\"Royale Coffee\")\n        .setCoffee(15)\n        .setWater(100)\n        .setMilk(10)\n        .setCinnamon(3)\n        .make();\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Calling to method <code class=\"highlighter-rouge\">make<\/code> checks all required parameters, and could validate and throw an exception if object is in inconsistent state.<br \/><strong>Eve:<\/strong> Awesome functionality, but why so verbose?<br \/><strong>Pedro:<\/strong> Beat it.<br \/><strong>Eve:<\/strong> A piece of cake, clojure supports <strong>optional arguments<\/strong>, everything what builder pattern is about.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn make-coffee [name amount water\n                   &amp; {:keys [milk sugar cinnamon]\n                      :or {milk 0 sugar 0 cinnamon 0}}]\n  ;; definition goes here\n  )\n(make-coffee \"Royale Coffee\" 15 100\n             :milk 10\n             :cinnamon 3)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Aha, you have three required parameters and three optionals, but required parameters still without names.<br \/><strong>Eve:<\/strong> What do you mean?<br \/><strong>Pedro:<\/strong> From the client call I see number <code class=\"highlighter-rouge\">15<\/code> but I have no idea what it might be.<br \/><strong>Eve:<\/strong> Agreed. Then, let\u2019s make all parameters are named and add precondition for required, the same way you do with the builder.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn make-coffee\n  [&amp; {:keys [name amount water milk sugar cinnamon]\n      :or {name \"\" amount 0 water 0 milk 0 sugar 0 cinnamon 0}}]\n  {:pre [(not (empty? name))\n         (&gt; amount 0)\n         (&gt; water 0)]}\n  ;; definition goes here\t\t \n  )\n(make-coffee :name \"Royale Coffee\"\n             :amount 15\n             :water 100\n             :milk 10\n             :cinnamon 3)\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> As you see all parameters are named and all required params are checked in <code class=\"highlighter-rouge\">:pre<\/code> constraint. If constraints are violated <code class=\"highlighter-rouge\">AssertionError<\/code> is thrown.<br \/><strong>Pedro:<\/strong> Interesting, <code class=\"highlighter-rouge\">:pre<\/code> is a part of a language?<br \/><strong>Eve:<\/strong> Sure, it\u2019s just a simple assertion. There is also <code class=\"highlighter-rouge\">:post<\/code> constraint, with the similar effect.<br \/><strong>Pedro:<\/strong> Hm, okay. But as you know <code class=\"highlighter-rouge\">Builder<\/code> pattern often used as a mutable datastucture, <code class=\"highlighter-rouge\">StringBuilder<\/code> for example.<br \/><strong>Eve:<\/strong> It\u2019s not a part of clojure philosophy to use mutables, but if you <em>really<\/em> want, no problem.\nJust create a new class with <code class=\"highlighter-rouge\">deftype<\/code> and do not forget to use <code class=\"highlighter-rouge\">volatile-mutable<\/code> on the properties you want to mutate.<br \/><strong>Pedro:<\/strong> Where is the code?<br \/><strong>Eve:<\/strong> Here is example of custom implementation of mutable <code class=\"highlighter-rouge\">StringBuilder<\/code> in clojure. It has a lot of drawbacks and limitations but you\u2019ve got the idea.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; interface\n(defprotocol IStringBuilder\n\t(append [this s])\n\t(to-string [this]))\n;; implementation\n(deftype ClojureStringBuilder [charray ^:volatile-mutable last-pos]\n\tIStringBuilder\n\t(append [this s] \n\t  (let [cs (char-array s)]\n\t    (doseq [i (range (count cs))]\n\t      (aset charray (+ last-pos i) (aget cs i))))\n\t  (set! last-pos (+ last-pos (count s))))\n\t(to-string [this] (apply str (take last-pos charray))))\n;; clojure binding\n(defn new-string-builder []\n  (ClojureStringBuilder. (char-array 100) 0))\n;; usage\n(def sb (new-string-builder))\n(append sb \"Toby Wong\")\n(to-string sb) =&gt; \"Toby Wong\"\n(append sb \" \")\n(append sb \"Toby Chung\") =&gt; \"Toby Wang Toby Chung\"\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Not as hard as I thought.<\/p><h3 id=\"episode-14-facade\">Episode 14: Facade<\/h3><blockquote>\n  <p>Our new member <strong>Eugenio Reinn Jr.<\/strong> commited\nfile with diff in 134 lines to processing servlet.\nBut actual work is just to process request.\nAll other code injection and imports.\nIt MUST be one line commit.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Who cares how many lines are committed?<br \/><strong>Eve:<\/strong> Someone cares.<br \/><strong>Pedro:<\/strong> Let\u2019s see where is the problem<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class OldServlet {\n  @Autowired\n  RequestExtractorService requestExtractorService;\n  @Autowired\n  RequestValidatorService requestValidatorService;\n  @Autowired\n  TransformerService transformerService;\n  @Autowired\n  ResponseBuilderService responseBuilderService;\n  public Response service(Request request) {\n    RequestRaw rawRequest = requestExtractorService.extract(request);\n    RequestRaw validated = requestValidatorService.validate(rawRequest);\n    RequestRaw transformed = transformerService.transform(validated);\n    Response response = responseBuilderService.buildResponse(transformed);\n    return response;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Oh shi\u2026<br \/><strong>Pedro:<\/strong> That\u2019s our internal API for developers, every time they need to process request, inject\n4 services, include all imports, and write this code.<br \/><strong>Eve:<\/strong> Let\u2019s refactor it with\u2026<br \/><strong>Pedro:<\/strong> \u2026Facade pattern. We resolve all dependencies to a <strong>single point of access<\/strong> and simplify API usage.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class FacadeService {\n  @Autowired\n  RequestExtractorService requestExtractorService;\n  @Autowired\n  RequestValidatorService requestValidatorService;\n  @Autowired\n  TransformerService transformerService;\n  @Autowired\n  ResponseBuilderService responseBuilderService;\n  RequestRaw extractRequest(Request req) {\n    return requestExtractorService.extract(req);\n  } \n  \n  RequestRaw validateRequest(RequestRaw raw) {\n    return requestValidatorService.validate(raw);\n  }\n  \n  RequestRaw transformRequest(RequestRaw raw) {\n    return transformerService.transform(raw);\n  }\n  \n  Response buildResponse(RequestRaw raw) {\n    return responseBuilderService.buildResponse(raw);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Then if you need any service or set of services in the code\nyou just injecting facade to your code<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class NewServlet {\n  @Autowired\n  FacadeService facadeService;\n  Response service(Request request) {\n    RequestRaw rawRequest = facadeService.extractRequest(request);\n    RequestRaw validated = facadeService.validateRequest(rawRequest);\n    RequestRaw transformed = facadeService.transformRequest(validated);\n    Response response = facadeService.buildResponse(transformed);\n    return response;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Wait, you\u2019ve just moved all dependencies to one and everytime using this one, correct?<br \/><strong>Pedro:<\/strong> Yes, now everytime some functionality is needed, use <code class=\"highlighter-rouge\">FacadeService<\/code>. Dependency is already there.<br \/><strong>Eve:<\/strong> But we did the same in <a href=\"#mediator\">Mediator<\/a> pattern?<br \/><strong>Pedro:<\/strong> Mediator is behavioral pattern. We resolved all dependency to Mediator and added <em>new behavior<\/em> to it.<br \/><strong>Eve:<\/strong> And facade?<br \/><strong>Pedro:<\/strong> Facade is structural, we don\u2019t add new functionality, we just <em>expose existing functionality<\/em> with facade.<br \/><strong>Eve:<\/strong> Got it. But seems that pattern very loud word for such little tweak.<br \/><strong>Pedro:<\/strong> Maybe.<br \/><strong>Eve:<\/strong> Here is clojure version using structure by namespaces<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(ns application.old-servlet\n  (:require [application.request-extractor :as re])\n  (:require [application.request-validator :as rv])\n  (:require [application.transformer :as t])\n  (:require [application.response-builder :as rb]))\n(defn service [request]\n  (-&gt; request\n      (re\/extract)\n      (rv\/validate)\n\t  (t\/transform)\n\t  (rb\/build)))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Exposing all services via facade.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(ns application.facade\n  (:require [application.request-extractor :as re])\n  (:require [application.request-validator :as rv])\n  (:require [application.transformer :as t])\n  (:require [application.response-builder :as rb]))\n(defn request-extract [request]\n  (re\/extract request))\n(defn request-validate [request]\n  (rv\/validate request))\n(defn request-transform [request]\n  (t\/transform request))\n(defn response-build [request]\n  (rb\/build request))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And use it.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(ns application.old-servlet\n  (:use [application.facade]))\n(defn service [request]\n  (-&gt; request\n      (request-extract)\n      (request-validate)\n\t  (request-transform)\n\t  (request-build)))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> What the difference between <code class=\"highlighter-rouge\">:use<\/code> and <code class=\"highlighter-rouge\">:require<\/code>?<br \/><strong>Eve:<\/strong> They are almost similar, but with <code class=\"highlighter-rouge\">:require<\/code> you expose functionality via namespace qualificator\n<code class=\"highlighter-rouge\">(namespace\/function)<\/code> where with <code class=\"highlighter-rouge\">:use<\/code> you can refer to it directly <code class=\"highlighter-rouge\">(function)<\/code><br \/><strong>Pedro:<\/strong> So, <code class=\"highlighter-rouge\">:use<\/code> is better.<br \/><strong>Eve:<\/strong> No, be aware of <code class=\"highlighter-rouge\">:use<\/code> because it can conflict with existing names in your namespace.<br \/><strong>Pedro:<\/strong> Oh, I see your point. And every time you call <code class=\"highlighter-rouge\">(:use [application.facade])<\/code> in some namespace\nall existing functionality from facade is available?<br \/><strong>Eve:<\/strong> Yes.<br \/><strong>Pedro:<\/strong> Pretty the same.<\/p><h3 id=\"episode-15-singleton\">Episode 15: Singleton<\/h3><blockquote>\n  <p><strong>Feverro O\u2019Neal<\/strong> complains that we have a lot of different styles for UI.<br \/>Force one per application UI configuration.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> But wait, there was requirement to save UI style per user.<br \/><strong>Eve:<\/strong> Probably it was changed.<br \/><strong>Pedro:<\/strong> Ok, then we should just save configuration to <code class=\"highlighter-rouge\">Singleton<\/code> and use it from all the places.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public final class UIConfiguration {\n  public static final UIConfiguration INSTANCE = new UIConfiguration(\"ui.config\");\n  private String backgroundStyle;\n  private String fontStyle;\n  \/* other UI properties *\/\n  private UIConfiguration(String configFile) {\n    loadConfig(configFile);\n  }\n  private static void loadConfig(String file) {\n    \/\/ process file and fill UI properties\n\tINSTANCE.backgroundStyle = \"black\";\n    INSTANCE.fontStyle = \"Arial\";\n  }\n  public String getBackgroundStyle() {\n    return backgroundStyle;\n  }\n  public String getFontStyle() {\n    return fontStyle;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> That way all configuration will be shared across the UIs.<br \/><strong>Eve:<\/strong> Yes, but\u2026why so much code?<br \/><strong>Pedro:<\/strong> We guarantee that only one instance of <code class=\"highlighter-rouge\">UIConfiguration<\/code> will exist.<br \/><strong>Eve:<\/strong> Let me ask you: what\u2019s the difference between singleton and global varaible.<br \/><strong>Pedro:<\/strong> What?<br \/><strong>Eve:<\/strong> \u2026the difference between singleton and global variable.<br \/><strong>Pedro:<\/strong> Java does not support global variables.<br \/><strong>Eve:<\/strong> But <code class=\"highlighter-rouge\">UIConfiguration.INSTANCE<\/code> is global variable.<br \/><strong>Pedro:<\/strong> Well, sort of.<br \/><strong>Eve:<\/strong> Your code is just simple <code class=\"highlighter-rouge\">def<\/code> in clojure.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def ui-config (load-config \"ui.config\"))\n(defn load-config [config-file]\n  ;; process config file and return map with configuratios\n  {:bg-style \"black\" :font-style \"Arial\"})\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> But, how do you change the style?<br \/><strong>Eve:<\/strong> The same way you will change it in your code.<br \/><strong>Pedro:<\/strong> Uhm\u2026 Ok, we need a simple tweak. Make <code class=\"highlighter-rouge\">UIConfiguration.loadConfig<\/code> is public and\ncall it when the configuration changes.<br \/><strong>Eve:<\/strong> Then we make <code class=\"highlighter-rouge\">ui-config<\/code> an atom and call <code class=\"highlighter-rouge\">swap!<\/code> when configuration changes.<br \/><strong>Pedro:<\/strong> But atoms are useful only in concurrent environment.<br \/><strong>Eve:<\/strong> First, yes, they useful, but NOT only in concurrent environment.\nSecond, atom read is not as slow as you think. Third, it changes the state of\nUI configuration <em>atomically<\/em><br \/><strong>Pedro:<\/strong> It is redundant for such simple example.<br \/><strong>Eve:<\/strong> No, it is not. There is a posibility that UI configuration changes\nand some renders read new <code class=\"highlighter-rouge\">backgroundStyle<\/code>, but old <code class=\"highlighter-rouge\">fontStyle<\/code><br \/><strong>Pedro:<\/strong> Ok, use <code class=\"highlighter-rouge\">synchronized<\/code> for <code class=\"highlighter-rouge\">loadConfig<\/code><br \/><strong>Eve:<\/strong> Then you must use <code class=\"highlighter-rouge\">synchonized<\/code> on getters as well, it is slow.<br \/><strong>Pedro:<\/strong> There is still <em>Double-Checked Locking<\/em> idiom<br \/><strong>Eve:<\/strong> Double-checked locking is clever but <a href=\"http:\/\/www.javaworld.com\/article\/2074979\/java-concurrency\/double-checked-locking--clever--but-broken.html\">broken<\/a><br \/><strong>Pedro:<\/strong> Ok, I give up, you won.<\/p><h3 id=\"episode-16-chain-of-responsibility\">Episode 16: Chain Of Responsibility<\/h3><blockquote>\n  <p>New York marketing organization <strong>\u201cA Profit NY\u201d<\/strong> opened request to\nfilter profanity words from their public chat system.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Fuck, they don\u2019t like the word \u201c<em>fuck<\/em>\u201d?<br \/><strong>Eve:<\/strong> It is profit organization, they lose money if someone use\nprofanity words in public chat.<br \/><strong>Pedro:<\/strong> Who defined profanity words list?<br \/><strong>Eve:<\/strong> <a href=\"https:\/\/www.youtube.com\/watch?v=5ssJtD08vCc\">George Carlin<\/a><\/p><p><em>Watching and laughing<\/em><\/p><p><strong>Pedro:<\/strong> Ok, so let\u2019s just add a filter to replace these rude words with the asterisks.<br \/><strong>Eve:<\/strong> Make sure your solution is extendable, other filters could be applied.<br \/><strong>Pedro:<\/strong> Chain of Responisibility seems like a good pattern candidate for that.\nFirst of all we make some abstract filter.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public abstract class Filter {\n  protected Filter nextFilter;\n  abstract void process(String message);\n  public void setNextFilter(Filter nextFilter) {\n    this.nextFilter = nextFilter;\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Then, provide implementation for each specific filter you want to apply<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class LogFilter extends Filter {\n  @Override\n  void process(String message) {\n    Logger.info(message);\n    if (nextFilter != null) nextFilter.process(message);\n  }\n}\nclass ProfanityFilter extends Filter {\n  @Override\n  void process(String message) {\n    String newMessage = message.replaceAll(\"fuck\", \"f*ck\");\n    if (nextFilter != null) nextFilter.process(newMessage);\n  }\n}\nclass RejectFilter extends Filter {\n  @Override\n  void process(String message) {\n    System.out.println(\"RejectFilter\");\n    if (message.startsWith(\"[A PROFIT NY]\")) {\n      if (nextFilter != null) nextFilter.process(message);\n    } else {\n      \/\/ reject message - do not propagate processing\n    }\n  }\n}\nclass StatisticsFilter extends Filter {\n  @Override\n  void process(String message) {\n    Statistics.addUsedChars(message.length());\n    if (nextFilter != null) nextFilter.process(message);\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And finally build a chain of filters which defines an order how message will be processed.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Filter rejectFilter = new RejectFilter();\nFilter logFilter = new LogFilter();\nFilter profanityFilter = new ProfanityFilter();\nFilter statsFilter = new StatisticsFilter();\nrejectFilter.setNextFilter(logFilter);\nlogFilter.setNextFilter(profanityFilter);\nprofanityFilter.setNextFilter(statsFilter);\nString message = \"[A PROFIT NY] What the fuck?\";\nrejectFilter.process(message);\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Ok, now clojure turn. Just define each filter as a function.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; define filters\n(defn log-filter [message]\n  (logger\/log message)\n  message)\n(defn stats-filter [message]\n  (stats\/add-used-chars (count message))\n  message)\n(defn profanity-filter [message]\n  (clojure.string\/replace message \"fuck\" \"f*ck\"))\n(defn reject-filter [message]\n  (if (.startsWith message \"[A Profit NY]\")\n    message))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And use <code class=\"highlighter-rouge\">some-&gt;<\/code> macro to chain filters<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn chain [message]\n  (some-&gt; message\n          reject-filter\n          log-filter\n          stats-filter\n          profanity-filter))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> You see how much it is easier, you don\u2019t need every-time call\n<code class=\"highlighter-rouge\">if (nextFilter != null) nextFilter.process()<\/code>, because it\u2019s natural.\nThe next filter defined at the <code class=\"highlighter-rouge\">some-&gt;<\/code> level naturally, instead of calling manuall <code class=\"highlighter-rouge\">setNext<\/code>.<br \/><strong>Pedro:<\/strong> That\u2019s definitely better for composability, but why did you use <code class=\"highlighter-rouge\">some-&gt;<\/code> instead of <code class=\"highlighter-rouge\">-&gt;<\/code>?<br \/><strong>Eve:<\/strong> Just for <code class=\"highlighter-rouge\">reject-filter<\/code>. It could stop further processing, so <code class=\"highlighter-rouge\">some-&gt;<\/code> returns <code class=\"highlighter-rouge\">nil<\/code> as soon\nas <code class=\"highlighter-rouge\">nil<\/code> encountered as a filter<br \/><strong>Pedro:<\/strong> Could you explain more?<br \/><strong>Eve:<\/strong> Look at the usage<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(chain \"fuck\") =&gt; nil\n(chain \"[A Profit NY] fuck\") =&gt; \"f*ck\"\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Understood.<br \/><strong>Eve:<\/strong> <em>Chain of Responsibility<\/em> just an approach to <strong>function composition<\/strong><\/p><h3 id=\"episode-17-composite\">Episode 17: Composite<\/h3><blockquote>\n  <p>Actress <strong>Bella Hock<\/strong> don\u2019t see user avatars in our social network on her computer.<\/p>\n  <p>\u201cEverything is black. Is it black holes?\u201d<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> It is black squares.<br \/><strong>Eve:<\/strong> Hmm, the same problem on our side.<br \/><strong>Pedro:<\/strong> Seems, that latest feature broke user avatars.<br \/><strong>Eve:<\/strong> Strange, because avatars rendered the same way as other elements, but they are visible.<br \/><strong>Pedro:<\/strong> Are you sure it rendered the same way?<br \/><strong>Eve:<\/strong> Well\u2026no<\/p><p><em>Digging code<\/em><\/p><p><strong>Pedro:<\/strong> What the hell is going on here?<br \/><strong>Eve:<\/strong> Someone copypasted code, but forgot to reflect changes in avatars.<br \/><strong>Pedro:<\/strong> Let\u2019s blame him, <code class=\"highlighter-rouge\">git-blame<\/code><br \/><strong>Eve:<\/strong> <em>Blame<\/em> is good, but we need to fix the problem<br \/><strong>Pedro:<\/strong> It\u2019s simple just additional line here.<br \/><strong>Eve:<\/strong> I mean, really solve the problem. Why do we need two similar snippets of code to\nprocess the same blocks?<br \/><strong>Pedro:<\/strong> Correct, I think we could use Composite pattern to handle rendering of the whole page.\nOur smallest element to render is block.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public interface Block {\n  void addBlock(Block b);\n  List&lt;Block&gt; getChildren();\n  \n  void render();\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Obviously blocks may contain other blocks, that\u2019s the point of Composite pattern.\nWe may create some specific blocks.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class Page implements Block { }\npublic class Header implements Block { }\npublic class Body implements Block { }\npublic class HeaderTitle implements Block { }\npublic class UserAvatar implements Block { }\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And treat every specific element as a <code class=\"highlighter-rouge\">Block<\/code><\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Block page = new Page();\nBlock header = new Header();\nBlock body = new Body();\nBlock title = new HeaderTitle();\nBlock avatar = new UserAvatar();\npage.addBlock(header);\npage.addBlock(body);\nheader.addBlock(title);\nheader.addBlock(avatar);\npage.render();\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> This is a structural pattern, a good way to <em>compose<\/em> objects. That\u2019s why it is called <em>composite<\/em><br \/><strong>Eve:<\/strong> Hey, <strong>composite is a simple tree<\/strong> structure.<br \/><strong>Pedro:<\/strong> Yes.<br \/><strong>Eve:<\/strong> There is pattern for every datastructure?<br \/><strong>Pedro:<\/strong> No, just for list and tree.<br \/><strong>Eve:<\/strong> Actually, tree can be represented as a list<br \/><strong>Pedro:<\/strong> How?<br \/><strong>Eve:<\/strong> First element of a list is a node value, next elements are children, each of them is\u2026<br \/><strong>Pedro:<\/strong> I understand.<br \/><strong>Eve:<\/strong> To be specific, here is the tree<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>        A\n     \/  |  \\\n    B   C   D\n    |   |  \/ \\\n\tE   H J   K\n   \/ \\       \/|\\\n  F   G     L M N\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> And here is the list represents this tree<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def tree\n  '(A (B (E (F) (G))) (C (H)) (D (J) (K (L) (M) (N)))))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> It\u2019s a plenty of parentheses!<br \/><strong>Eve:<\/strong> They define structure, you know<br \/><strong>Pedro:<\/strong> But it\u2019s hard to understand<br \/><strong>Eve:<\/strong> It\u2019s easy for machine, there is awesome <code class=\"highlighter-rouge\">tree-seq<\/code> function to process the tree.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(map first (tree-seq next rest tree)) =&gt; (A B E F G C H D J K L M N)\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> If you need more advanced traversals, use <a href=\"https:\/\/clojure.github.io\/clojure\/clojure.walk-api.html\">clojure.walk<\/a><br \/><strong>Pedro:<\/strong> I don\u2019t know, everything seems just a bit harder.<br \/><strong>Eve:<\/strong> No, you define the whole tree with one datastructure and use one function to operate on it.<br \/><strong>Pedro:<\/strong> What this function will do?<br \/><strong>Eve:<\/strong> It traverses the tree and applies to every node, so in our case it can render each component.<br \/><strong>Pedro:<\/strong> I don\u2019t know, maybe I am too young for the trees, let\u2019s move forward.<\/p><h3 id=\"episode-18-factory-method\">Episode 18. Factory Method<\/h3><blockquote>\n  <p><strong>Sir Dry Bang<\/strong> suggest to create new levels for their popular\ngame. More levels - more money.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> How can we create new levels?<br \/><strong>Eve:<\/strong> Just change assets and add new blocks: paper, wood, iron\u2026<br \/><strong>Pedro:<\/strong> It\u2019s too silly, isn\u2019t it?<br \/><strong>Eve:<\/strong> The whole game is too silly. If user pay for color hats for their characters, then they will pay for wooden blocks as well.<br \/><strong>Pedro:<\/strong> I think it\u2019s a crap, but anyway, let\u2019s make <code class=\"highlighter-rouge\">MazeBuilder<\/code> generic and add specific builder\nfor each type of the block. It\u2019s a Factory Method pattern.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class Maze { }\nclass WoodMaze extends Maze { }\nclass IronMaze extends Maze { }\ninterface MazeBuilder {\n  Maze build();\n}\nclass WoodMazeBuilder {\n  @Override\n  Maze build() {\n    return new WoodMaze();\n  }\n}\nclass IronMazeBuilder {\n  @Override\n  Maze build() {\n    return new IronMaze();\n  }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Isn\u2019t it obvious that IronMazeBuilder will return IronMazes?<br \/><strong>Pedro:<\/strong> Not for program. But see, to make a maze from another blocks we just change implementation responsible for block creation.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>MazeBuilder builder = new WoodMazeBuilder();\nMaze maze = builder.build();\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> I\u2019ve seen something similar before.<br \/><strong>Pedro:<\/strong> What exactly?<br \/><strong>Eve:<\/strong> For me it seems like a strategy or state pattern.<br \/><strong>Pedro:<\/strong> No way! Strategy is about performing specific operations and factory is for creating specific object.<br \/><strong>Eve:<\/strong> But create is an operation as well.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn maze-builder [maze-fn])\n(defn make-wood-maze [])\n(defn make-iron-maze [])\n(def wood-maze-builder (partial maze-builder make-wood-maze))\n(def iron-maze-builder (partial maze-builder make-iron-maze))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Hm, seems similar.<br \/><strong>Eve:<\/strong> Think about it.<br \/><strong>Pedro:<\/strong> Any usage examples?<br \/><strong>Eve:<\/strong> No, everything is obvious here, just re-read <a href=\"#strategy\">Strategy<\/a>, <a href=\"#state\">State<\/a> or <a href=\"#template_method\">TemplateMethod<\/a>\nepisodes.<\/p><h3 id=\"episode-19-abstract-factory\">Episode 19: Abstract Factory<\/h3><blockquote>\n  <p>Users are not buying new levels in the game.\n<strong>Saimank Gerr<\/strong> build a complains cloud and the most popular\nnegative feedback words are: \u201cugly\u201d, \u201ccrap\u201d and \u201cshit\u201d.<\/p>\n  <p>Improve levels-building system.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> I said this is crap.<br \/><strong>Eve:<\/strong> Sure, snow background with wooden walls, space invaders with wooden walls every setting with wooden walls.<br \/><strong>Pedro:<\/strong> Then we must separate game worlds and build a set of specific objects for particular world.<br \/><strong>Eve:<\/strong> Explain.<br \/><strong>Pedro:<\/strong> Instead of using Factory Method for building specific blocks, we use Abstract Factory\nto build a set of related objects, to make a level look less crappy.<br \/><strong>Eve:<\/strong> Example would be good.<br \/><strong>Pedro:<\/strong> Code is my example. First we define <em>abstract<\/em> behaviour of level factory<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public interface LevelFactory {\n  Wall buildWall();\n  Back buildBack();\n  Enemy buildEnemy();\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Then we have a hierarchy of objects that level consists of<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class Wall {}\nclass PlasmaWall extends Wall {}\nclass StoneWall extends Wall {}\nclass Back {}\nclass StarsBack extends Back {}\nclass EarthBack extends Back {}\nclass Enemy {}\nclass UFOSoldier extends Enemy {}\nclass WormScout extends Enemy {}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> See? We have a specific object for each level, let\u2019s create factory for them.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class SpaceLevelFactory implements LevelFactory {\n  @Override\n  public Wall buildWall() {\n    return new PlasmaWall();\n  }\n  @Override\n  public Back buildBack() {\n    return new StarsBack();\n  }\n  @Override\n  public Enemy buildEnemy() {\n    return new UFOSoldier();\n  }\n}\nclass UndergroundLevelFactory implements LevelFactory {\n  @Override\n  public Wall buildWall() {\n    return new StoneWall();\n  }\n  @Override\n  public Back buildBack() {\n    return new EarthBack();\n  }\n  @Override\n  public Enemy buildEnemy() {\n    return new WormScout();\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Each implementation of level factory creates related objects for level.\nLevels will look less crappy for sure.<br \/><strong>Eve:<\/strong> Let me understand. I really can\u2019t spot the difference.<br \/><strong>Pedro:<\/strong> Factory Method <em>defers object creation to a subclasses<\/em>,\nAbstract Factory do the same but for a <em>set of related object<\/em>.<br \/><strong>Eve:<\/strong> Aha, that means I need to pass <strong>set of related functions<\/strong> to abstract builder<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(defn level-factory [wall-fn back-fn enemy-fn])\n(defn make-stone-wall [])\n(defn make-plasma-wall [])\n(defn make-earth-back [])\n(defn make-stars-back [])\n(defn make-worm-scout [])\n(defn make-ufo-soldier [])\n(def underground-level-factory\n  (partial level-factory\n           make-stone-wall\n           make-earth-back\n           make-worm-scout))\n(def space-level-factory\n  (partial level-factory\n           make-plasma-wall\n           make-stars-back\n           make-ufo-soldier))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> I knew.<br \/><strong>Eve:<\/strong> Everything is fair. Your lovely \u201cset of related Xs\u201d, where X is a function<br \/><strong>Pedro:<\/strong> Yes, clarify, what <code class=\"highlighter-rouge\">partial<\/code> is.<br \/><strong>Eve:<\/strong> Provide some parameters for function. So, <code class=\"highlighter-rouge\">underground-level-factory<\/code> knows how to construct walls, backs and enemies. Everything other <em>inherited<\/em> from abstract <code class=\"highlighter-rouge\">level-factory<\/code> function.<br \/><strong>Pedro:<\/strong> Handy.<\/p><h3 id=\"episode-20-adapter\">Episode 20: Adapter<\/h3><blockquote>\n  <p><strong>Deam Evil<\/strong> conducts a medieval tournament for knights.\nThe prize is $100.000<\/p>\n  <p>I\u2019ll pay you the half if you break the system and allow\nmy armed commando to take part in competition.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Finally we\u2019ve got interesting work.<br \/><strong>Eve:<\/strong> Funny to see the competition. Especially, M16 vs Iron Sword part.<br \/><strong>Pedro:<\/strong> Knights have a good armor.<br \/><strong>Eve:<\/strong> F1 grenade <em>does not care<\/em> about armor.<br \/><strong>Pedro:<\/strong> Nevermind, we do the work, we get the money.<br \/><strong>Eve:<\/strong> Fifty grands - nice compensation<br \/><strong>Pedro:<\/strong> Yes, look at this, I\u2019ve stolen sources of the competition system,\nthough it is not possible to modify their sources, we can find some vulneravility.<br \/><strong>Eve:<\/strong> Here it is<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public interface Tournament {\n  void accept(Knight knight);\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Aha! System validates only incoming types via <code class=\"highlighter-rouge\">Knight<\/code> interface.\nAll we need to do is <em>to adapt<\/em> commando to be a knight. Let\u2019s see how knight look like<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>interface Knight {\n  void attackWithSword();\n  void attackWithBow();\n  void blockWithShield();\n}\nclass Galahad implements Knight {\n  @Override\n  public void blockWithShield() {\n    winkToQueen();\n    take(shield);\n    block();\n  }\n  @Override\n  public void attackWithBow() {\n    winkToQueen();\n    take(bow);\n    attack();\n  }\n  @Override\n  public void attackWithSword() {\n    winkToQueen();\n    take(sword);\n    attack();\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> To accept the commando let\u2019s take an old implementation<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class Commando {\n    void throwGrenade(String grenade) { }\n    shot(String rifleType) { }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> And adapt it.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class Commando implements Knight {\n  @Override\n  public void blockWithShield() {\n    \/\/ commando don't block\n  }\n  @Override\n  public void attackWithBow() {\n    throwGrenade(\"F1\");\n  }\n  @Override\n  public void attackWithSword() {\n    shotWithRifle(\"M16\");\n  }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> That\u2019s it.<br \/><strong>Eve:<\/strong> It\u2019s simpler in clojure.<br \/><strong>Pedro:<\/strong> Really?<br \/><strong>Eve:<\/strong> We don\u2019t love types so their validation won\u2019t work at all<br \/><strong>Pedro:<\/strong> So, how do you replace knight with commando?<br \/><strong>Eve:<\/strong> Basically, what knight is? It\u2019s a map, consists of data and behaviour<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>{:name \"Lancelot\"\n :speed 1.0\n :attack-bow-fn attack-with-bow\n :attack-sword-fn attack-with-sword\n :block-fn block-with-shield}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> To adapt commando, just pass his functions instead original ones<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>{:name \"Commando\"\n :speed 5.0\n :attack-bow-fn (partial throw-grenade \"F1\")\n :attack-sword-fn (partial shot \"M16\")\n :block-fn nil}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> How did we share money?<br \/><strong>Eve:<\/strong> 50\/50<br \/><strong>Pedro:<\/strong> I wrote more code, I want 70<br \/><strong>Eve:<\/strong> Ok, 70\/70<br \/><strong>Pedro:<\/strong> Deal.<\/p><h3 id=\"episode-21-decorator\">Episode 21: Decorator<\/h3><blockquote>\n  <p><strong>Podrea Vesper<\/strong> caught us on cheating for the tournament.\nWe have a choice: to be busted by police or to help\nhis super knight to take part in competition<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> I don\u2019t wanna go to prison.<br \/><strong>Eve:<\/strong> Me either.<br \/><strong>Pedro:<\/strong> Let\u2019s cheat for him one more time.<br \/><strong>Eve:<\/strong> This is the same solution, isn\u2019t it?<br \/><strong>Pedro<\/strong> Similar, but not the same.\nCommando was a soldier and they are not allowed on the tournament. We <em>adapted<\/em> it. But knight is allowed for tournament we don\u2019t need to adapt it, we <em>must<\/em> add functionality to existing object.<br \/><strong>Eve:<\/strong> Inheritance or composition?<br \/><strong>Pedro:<\/strong> Composition, the main goal of decorator to change behaviour at a runtime<br \/><strong>Eve:<\/strong> So how we deal with this super knight?<br \/><strong>Pedro:<\/strong> They plan to use <code class=\"highlighter-rouge\">Galahad<\/code> knight and <em>decorate<\/em> it with <em>more HP<\/em> and <em>Power Armor<\/em><br \/><strong>Eve:<\/strong> Heh, funny that cops are playing Fallout<br \/><strong>Pedro:<\/strong> Yeah, let\u2019s make knight an abstract class<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class Knight {\n    protected int hp;\n    private Knight decorated;\n    public Knight() { }\n    public Knight(Knight decorated) {\n        this.decorated = decorated;\n    }\n    public void attackWithSword() {\n        if (decorated != null) decorated.attackWithSword();\n    }\n    public void attackWithBow() {\n        if (decorated != null) decorated.attackWithBow();\n    }\n    public void blockWithShield() {\n        if (decorated != null) decorated.blockWithShield();\n    }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> So what we improved there?<br \/><strong>Pedro:<\/strong> First af all we make class <code class=\"highlighter-rouge\">Knight<\/code> instead of interface to have access to hit points. Then we provide two diferent constructors, default for standard behaviour, and decorated, which delegates call to decorated object.<br \/><strong>Eve:<\/strong> Is it fair to use abstract class instead of interface?<br \/><strong>Pedro:<\/strong> No, but we avoid two classes with similar behaviour and fill each decorated object with default implementation, instead of forcing to implement each of the methods.<br \/><strong>Eve:<\/strong> Ok, what\u2019s about Power Armor?<br \/><strong>Pedro:<\/strong> Easy as well<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>public class KnightWithPowerArmor extends Knight {\n    public KnightWithPowerArmor(Knight decorated) {\n        super(decorated);\n    }\n    @Override\n    public void blockWithShield() {\n        super.blockWithShield();\n        Armor armor = new PowerArmor();\n        armor.block();\n    }\n}\npublic class KnightWithAdditionalHP extends Knight {\n    public KnightWithAdditionalHP(Knight decorated) {\n        super(decorated);\n        this.hp += 50;\n    }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Two decorators that fulfils FBI requirements, and we able\nto create super knight, with behaviour like Galahad, but super armor and 50 more hit points.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Knight superKnight =\n     new KnightWithAdditionalHP(\n     new KnightWithPowerArmor(\n\t new Galahad()));\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Nice trick!<br \/><strong>Pedro:<\/strong> You are welcome to show the similar behaviour in clojure<br \/><strong>Eve:<\/strong> Here it is<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(def galahad {:name \"Galahad\"\n              :speed 1.0\n              :hp 100\n              :attack-bow-fn attack-with-bow\n              :attack-sword-fn attack-with-sword\n              :block-fn block-with-shield})\n(defn make-knight-with-more-hp [knight]\n  (update-in knight [:hp] + 50))\n(defn make-knight-with-power-armor [knight]\n  (update-in knight [:block-fn]\n             (fn [block-fn]\n               (fn []\n                 (block-fn)\n                 (block-with-power-armor)))))\n;; create the knight\n(def superknight (-&gt; galahad\n                     make-knight-with-power-armor\n                     make-knight-with-more-hp)\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> The same functionality.<br \/><strong>Eve:<\/strong> Yes, just pay attention to power armor decorator.<\/p><h3 id=\"episode-22-proxy\">Episode 22: Proxy<\/h3><blockquote>\n  <p><strong>Deren Bart<\/strong> manages the system for making mixed drinks.\nIt is rigid system, because after making a drink <em>Bart<\/em> must manually subtract used ingredients\nfrom the bar. Do this automatically.<\/p>\n<\/blockquote><p><strong>Pedro:<\/strong> Can we get access to his codebase?<br \/><strong>Eve:<\/strong> No, but he sent some APIs.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>interface IBar {\n    void makeDrink(Drink drink);\n}\ninterface Drink {\n    List&lt;Ingredient&gt; getIngredients();\n}\n    \ninterface Ingredient {\n    String getName();\n    double getAmount();\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Bart doesn\u2019t want us to modify sources, instead we need to provide some additional\nimplementation for <code class=\"highlighter-rouge\">IBar<\/code> interface with autosubtracting used ingredients.<br \/><strong>Eve:<\/strong> And what are we responsible for?<br \/><strong>Pedro:<\/strong> Implementing <em>Proxy Pattern<\/em>, I\u2019ve read about it some times ago.<br \/><strong>Eve:<\/strong> I\u2019m all listening.<br \/><strong>Pedro:<\/strong> Basically we delegate all existing functionality to standard <code class=\"highlighter-rouge\">IBar<\/code> implementation\nand provide new functionality inside <code class=\"highlighter-rouge\">ProxiedBar<\/code><\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class ProxiedBar implements IBar {\n    BarDatabase bar;\n    IBar standardBar;\n    public void makeDrink(Drink drink) {\n       standardBar.makeDrink(drink);\n       for (Ingredient i : drink.getIngredients()) {\n           bar.subtract(i);\n       }\n    }\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> They need to replace <code class=\"highlighter-rouge\">StandardBar<\/code> implementation with our <code class=\"highlighter-rouge\">ProxiedBar<\/code>.<br \/><strong>Eve:<\/strong> Seems super easy.<br \/><strong>Pedro:<\/strong> Yes, additional plus that we don\u2019t break existing functionality.<br \/><strong>Eve:<\/strong> Are you sure? We didn\u2019t run regression tests.<br \/><strong>Pedro:<\/strong> Everything we do is delegating functionality to already tested <code class=\"highlighter-rouge\">StandardBar<\/code><br \/><strong>Eve:<\/strong> But you also substracts used ingredients from <code class=\"highlighter-rouge\">BarDatabase<\/code><br \/><strong>Pedro:<\/strong> We assume, they are <em>decoupled<\/em>.<br \/><strong>Eve:<\/strong> Oh\u2026<br \/><strong>Pedro:<\/strong> Does clojure have some alternative?<br \/><strong>Eve:<\/strong> Well, I don\u2019t know. What I see here you are using function composition.<br \/><strong>Pedro:<\/strong> Explain.<br \/><strong>Eve:<\/strong> <code class=\"highlighter-rouge\">IBar<\/code> implementation is a set of functions, and another <code class=\"highlighter-rouge\">IBar<\/code> is another set of functions.\nEverything you talking about additional implementation could be covered by function composition. It\u2019s like\n<code class=\"highlighter-rouge\">make-drink<\/code> and after that <code class=\"highlighter-rouge\">subtract-ingredients<\/code> from bar.<br \/><strong>Pedro:<\/strong> Maybe code is more clear?<br \/><strong>Eve:<\/strong> Yes, but I don\u2019t think something special here<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; interface\n(defprotocol IBar\n  (make-drink [this drink]))\n;; Bart's implementation\n(deftype StandardBar []\n  IBar\n  (make-drink [this drink]\n    (println \"Making drink \" drink)\n    :ok))\n;; our implementation\n(deftype ProxiedBar [db ibar]\n  IBar\n  (make-drink [this drink]\n    (make-drink ibar drink)\n    (subtract-ingredients db drink)))\n;; this how it was before\n(make-drink (StandardBar.)\n\t{:name \"Manhattan\"\n\t :ingredients [[\"Bourbon\" 75] [\"Sweet Vermouth\" 25] [\"Angostura\" 5]]})\n;; this how it becomes now\n(make-drink (ProxiedBar. {:db 1} (StandardBar.))\n\t{:name \"Manhattan\"\n     :ingredients [[\"Bourbon\" 75] [\"Sweet Vermouth\" 25] [\"Angostura\" 5]]})\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> We could leverage protocol and types to group set of functions as a single object.<br \/><strong>Pedro:<\/strong> Looks like clojure has object-oriented capabilities as well.<br \/><strong>Eve:<\/strong> Correct, moreover it has a <code class=\"highlighter-rouge\">reify<\/code> function, which allow you to create proxies in a runtime<br \/><strong>Pedro:<\/strong> Like class in a runtime?<br \/><strong>Eve:<\/strong> Sort of.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>(reify IBar\n  (make-drink [this drink]\n    ;; implementation goes here\n  ))\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Looks handy.<br \/><strong>Eve:<\/strong> Yes, but I still don\u2019t understand how it differs from Decorator.<br \/><strong>Pedro:<\/strong> They are completely different.<br \/><strong>Eve:<\/strong> Decorator adds functionality to the same interface, and so does Proxy.<br \/><strong>Pedro:<\/strong> Well, but Proxy is\u2026<br \/><strong>Eve:<\/strong> Even more, Adapter is not very different as well.<br \/><strong>Pedro:<\/strong> It uses another interface.<br \/><strong>Eve:<\/strong> But from implementation perspective all these pattern are the same,\nwrap something and delegate calls to wrapper. <em>\u201cWrapper\u201d<\/em> could be a good name for these patterns.<\/p><h3 id=\"episode-23-bridge\">Episode 23: Bridge<\/h3><blockquote>\n  <p>Girls from HR agency \u201cHurece\u2019s Sour Man\u201d trying to identify candidates\nto their open job positions. The problem is jobs often created by customers, but requirements to the jobs\nare developed by HR. Provide them with a flexible way of collaborating.<\/p>\n<\/blockquote><p><strong>Eve:<\/strong> I don\u2019t understand the problem.<br \/><strong>Pedro:<\/strong> I have a bit of background. They have a very strange system which defines job requirements as an interface.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>interface JobRequirement {\n    boolean accept(Candidate c);\n}\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Every specific requirement implemented as a new subclass of this.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>class JavaRequirement implements JobRequirement {\n    public boolean accept(Candidate c) {\n        return c.hasSkill(\"Java\");\n    }\n}\nclass Experience10YearsRequirement implements JobRequirement {\n    public boolean accept(Candidate c) {\n        return c.getExperience() &gt;= 10;\n    }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> I\u2019ve got an idea.<br \/><strong>Pedro:<\/strong> Take into account, this requirement hierarchy is designed by HR Department.<br \/><strong>Eve:<\/strong> Ok.<br \/><strong>Pedro:<\/strong> And they have a <code class=\"highlighter-rouge\">Job<\/code> hierarchy, where each specific job is subclass as well.<br \/><strong>Eve:<\/strong> Why do they need a class for each job? It should be an object.<br \/><strong>Pedro:<\/strong> The system was designed when classes were more popular than objects, so live with this.<br \/><strong>Eve:<\/strong> Class were popular than objects?!<br \/><strong>Pedro:<\/strong> Yes, listen and don\u2019t interrupt me. Jobs with requirements is completely separate hierarchy,\nand it is developed by customers. We introduce pattern <code class=\"highlighter-rouge\">Bridge<\/code> to separate these two hierarchies and allow\nthem live independently.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>abstract class Job {\n    protected List&lt;? extends JobRequirement&gt; requirements;\n    public Job(List&lt;? extends JobRequirement&gt; requirements) {\n        this.requirements = requirements;\n    }\n\tprotected boolean accept(Candidate c) {\n        for (JobRequirement j : requirements) {\n            if (!j.accept(c)) {\n                return false;\n\t\t\t}\n        }\n        return true;\n    }\n}\nclass CognitectClojureDeveloper extends Job {\n    public CognitectClojureDeveloper() {\n        super(Arrays.asList(\n                  new ClojureJobRequirement(),\n                  new Experience10YearsRequirement()\n        ));\n    }\n}\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> So where is the bridge?<br \/><strong>Pedro:<\/strong> <code class=\"highlighter-rouge\">JobRequirement<\/code>, <code class=\"highlighter-rouge\">JavaRequirement<\/code>, <code class=\"highlighter-rouge\">ExperienceRequirement<\/code> is one hierarchy, yes?<br \/><strong>Eve:<\/strong> Yes.<br \/><strong>Pedro:<\/strong> <code class=\"highlighter-rouge\">Job<\/code>, <code class=\"highlighter-rouge\">CongnitectClojureDeveloperJob<\/code>, <code class=\"highlighter-rouge\">OracleJavaDeveloperJob<\/code> is another hierarchy.<br \/><strong>Eve:<\/strong> Oh, now I see. A link from Job to JobRequirement is a bridge.<br \/><strong>Pedro:<\/strong> Exactly! This is how HR can use this system to find candidate matches.<\/p><div class=\"language-java highlighter-rouge\"><div class=\"highlight\"><pre>Candidate joshuaBloch = new Candidate();\n(new CognitectClojureDeveloper()).accept(joshuaBloch);\n(new OracleSeniorJavaDeveloper()).accept(joshuaBloch);\n<\/pre><\/div><\/div><p><strong>Pedro:<\/strong> Here is the point. Customers use <code class=\"highlighter-rouge\">Job<\/code> as abstraction and <code class=\"highlighter-rouge\">JobRequirement<\/code> as an implementation.\nThey just create a job with descriptions or so, and HR is responsible to convert these descriptions into specific set\nof <code class=\"highlighter-rouge\">JobRequirement<\/code> objects.<br \/><strong>Eve:<\/strong> Got it.<br \/><strong>Pedro:<\/strong> So as far as I understand, clojure could mimic this pattern by using <code class=\"highlighter-rouge\">defprotocol<\/code> and <code class=\"highlighter-rouge\">defrecord<\/code>?<br \/><strong>Eve:<\/strong> Yes, but I want to revisit the problem.<br \/><strong>Pedro:<\/strong> What\u2019s wrong?<br \/><strong>Eve:<\/strong> Here is we have a static flow: customer create job positions, human resources convert job position to a set of requirements\nand run a script against their candidates database to find a matches.<br \/><strong>Pedro:<\/strong> Correct.<br \/><strong>Eve:<\/strong> So there is already dependency, HR can\u2019t do anything without open job positions.<br \/><strong>Pedro:<\/strong> Well, yes. But they can develop a set of structured requirements without knowing what open positions might be.<br \/><strong>Eve:<\/strong> For what purpose?<br \/><strong>Pedro:<\/strong> Later this can be reused by Job creators, so HR avoid doing the same work twice.<br \/><strong>Eve:<\/strong> Okay, got it, but this problem is artificial. Basically what we need is a way to colaborate between abstraction and implementation.<br \/><strong>Pedro:<\/strong> Maybe, but I want to see your clojure way solution for that specific problem using bridge pattern.<br \/><strong>Eve:<\/strong> Easy. Let\u2019s use adhoc hierarchies.<br \/><strong>Pedro:<\/strong> For abstractions?<br \/><strong>Eve:<\/strong> Yes, jobs hierarchy is <em>abstraction<\/em>, and people need just to enhance hierarchy.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; abstraction\n(derive ::clojure-job ::job)\n(derive ::java-job ::job)\n(derive ::senior-clojure-job ::clojure-job)\n(derive ::senior-java-job    ::java-job)\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> HR department are like <em>developers<\/em>, they provide implementation for this abstraction.<\/p><div class=\"language-clojure highlighter-rouge\"><div class=\"highlight\"><pre>;; implementation\n(defmulti accept :job)\n(defmethod accept :java [candidate]\n  (and (some #{:java} (:skills candidate))\n       (&gt; (:experience candidate) 1)))\n<\/pre><\/div><\/div><p><strong>Eve:<\/strong> Later, when new jobs are created, but requirements are not yet developed and no <code class=\"highlighter-rouge\">accept<\/code> method implementation for this type of job, we fallback using adhoc hierarchy.<br \/><strong>Pedro:<\/strong> Hm?<br \/><strong>Eve:<\/strong> Assume someone created a new <code class=\"highlighter-rouge\">::senior-java<\/code> as a child of <code class=\"highlighter-rouge\">::java<\/code> job.<br \/><strong>Pedro:<\/strong> Oh, and if HR not provided <code class=\"highlighter-rouge\">accept<\/code> implementation for dispatch value <code class=\"highlighter-rouge\">::senior-java<\/code>, method with dispatch value <code class=\"highlighter-rouge\">::java<\/code> will be called, yes?<br \/><strong>Eve:<\/strong> You learning so fast.<br \/><strong>Pedro:<\/strong> But is it real bridge pattern?<br \/><strong>Eve:<\/strong> There is no <em>bridge<\/em> here, but abstraction and implementation can live independently.<\/p><h3 id=\"cheatsheet\">Cheatsheet<\/h3><p>It is very confusing to understand patterns, which often presented in object-oriented way\nwith bunch of UML diagrams, fancy nouns and exist to solve language-specific problems,\nso here is a revisited mini cheatsheet, which helps you to understand patterns by analogy.<\/p><ul><li><strong>Command<\/strong> - <em>function<\/em><\/li>\n  <li><strong>Strategy<\/strong> - <em>function, which accepts function<\/em><\/li>\n  <li><strong>State<\/strong> - <em>strategy, depends on state<\/em><\/li>\n  <li><strong>Visitor<\/strong> - <em>multiple dispatch<\/em><\/li>\n  <li><strong>Template Method<\/strong> - <em>strategy with defaults<\/em><\/li>\n  <li><strong>Iterator<\/strong> - <em>sequence<\/em><\/li>\n  <li><strong>Memento<\/strong> - <em>save and restore<\/em><\/li>\n  <li><strong>Prototype<\/strong> - <em>immutability<\/em><\/li>\n  <li><strong>Mediator<\/strong> - <em>reduce coupling<\/em><\/li>\n  <li><strong>Observer<\/strong> - <em>function, which calls after another function<\/em><\/li>\n  <li><strong>Interpreter<\/strong> - <em>set of functions to process a tree<\/em><\/li>\n  <li><strong>Flyweight<\/strong> - <em>cache<\/em><\/li>\n  <li><strong>Builder<\/strong> - <em>optional arguments<\/em><\/li>\n  <li><strong>Facade<\/strong> - <em>single point of access<\/em><\/li>\n  <li><strong>Singleton<\/strong> - <em>global variable<\/em><\/li>\n  <li><strong>Chain of Responsibility<\/strong> - <em>function composition<\/em><\/li>\n  <li><strong>Composite<\/strong> - <em>tree<\/em><\/li>\n  <li><strong>Factory Method<\/strong> - <em>strategy for creating objects<\/em><\/li>\n  <li><strong>Abstract Factory<\/strong> - <em>strategy for creating set of related objects<\/em><\/li>\n  <li><strong>Adapter<\/strong> - <em>wrapper, same functionality, different type<\/em><\/li>\n  <li><strong>Decorator<\/strong> - <em>wrapper, same type, new functionality<\/em><\/li>\n  <li><strong>Proxy<\/strong> - <em>wrapper, function composition<\/em><\/li>\n  <li><strong>Bridge<\/strong> - <em>separate abstraction and implementation<\/em><\/li>\n<\/ul><h3 id=\"cast\">Cast<\/h3><blockquote>\n  <p>A long time ago in a galaxy far, far away\u2026<\/p>\n<\/blockquote><p>With the lack of imagination all characters\nand names are just anagrams.<\/p><p><strong>Pedro Veel<\/strong> - Developer<br \/><strong>Eve Dopler<\/strong> - Developer<br \/><strong>Serpent Hill &amp; R.E.E.<\/strong> - Enterprise Hell<br \/><strong>Sven Tori<\/strong> - Investor<br \/><strong>Karmen Git<\/strong> - Marketing<br \/><strong>Natanius S. Selbys<\/strong> - Business Analyst<br \/><strong>Mech Dominore Fight Saga<\/strong> - Heroes of Might and Magic<br \/><strong>Kent Podiololis<\/strong> - I don\u2019t like loops<br \/><strong>Chad Bogue<\/strong> - Douchebag<br \/><strong>Dex Ringeus<\/strong> - UX Designer<br \/><strong>Veerco Wierde<\/strong> - Code Review<br \/><strong>Dartee Hebl<\/strong> - Heartbleed<br \/><strong>Bertie Prayc<\/strong> - Cyber Pirate<br \/><strong>Cristopher, Matton &amp; Pharts<\/strong> - Important Charts &amp; Reports<br \/><strong>Tuck Brass<\/strong> - Starbucks<br \/><strong>Eugenio Reinn Jr.<\/strong> - Junior Engineer<br \/><strong>Feverro O\u2019Neal<\/strong> - Forever Alone<br \/><strong>A Profit NY<\/strong> - Profanity<br \/><strong>Bella Hock<\/strong> - Black Hole<br \/><strong>Sir Dry Bang<\/strong> - Angry Birds<br \/><strong>Saimank Gerr<\/strong> - Risk Manager<br \/><strong>Deam Evil<\/strong> - Medieval<br \/><strong>Podrea Vesper<\/strong> - Eavesdropper<br \/><strong>Deren Bart<\/strong> - Bartender<br \/><strong>Hurece\u2019s Sour Man<\/strong> - Human Resources<\/p><p><strong>P.S.<\/strong> I\u2019ve started to write this article <em>more than<\/em> 2 years ago. Time has passed, things have changed and even Java 8 was released.<\/p><ul class=\"tag_box inline\"><li><a href=\"http:\/\/mishadoff.com\/categories\/#clojure-ref\">\n    \t\tclojure\n    \t<\/a><\/li>\n     \n    \t<li><a href=\"http:\/\/mishadoff.com\/categories\/#programming-ref\">\n    \t\tprogramming\n    \t<\/a><\/li>\n     \n    \t<li><a href=\"http:\/\/mishadoff.com\/categories\/#java-ref\">\n    \t\tjava\n    \t<\/a><\/li>\n     \n    \t<li><a href=\"http:\/\/mishadoff.com\/categories\/#story-ref\">\n    \t\tstory\n    \t<\/a><\/li>\n     \n    \t<li><a href=\"http:\/\/mishadoff.com\/categories\/#patterns-ref\">\n    \t\tpatterns\n    \t<\/a><\/li>\n    \n  \n  <\/ul><address class=\"signature\">\n     \n    18 December 2015\n    \n  <\/address><p><a href=\"http:\/\/mishadoff.com\/blog\/code-golf-new-year-tree\/\" class=\"next\" title=\"Code Golf: New Year Tree\">Code Golf: New Year Tree \u2192<\/a>\n  \n  \n    <a href=\"http:\/\/mishadoff.com\/blog\/littlepot-autofilled-cache\/\" class=\"prev\" title=\"Littlepot: Autofilled Cache\">\u2190 Littlepot: Autofilled Cache<\/a><\/p><\/div><address>\n  \t\t\n  \t\t\trandom thoughts about programming, math and life | <a href=\"http:\/\/mishadoff.com\/about\/\">mishadoff<\/a> \n  \t\t\n  \t\t\nDesign by \n<a href=\"http:\/\/mark.reid.name\/\">Mark Reid<\/a> | Powered by <a href=\"http:\/\/github.com\/mojombo\/jekyll\/\" title=\"A static site generator\">Jekyll<\/a>\n  \t\t\n  \t<\/address>","created_at":"2018-11-07T16:48:09+0000","updated_at":"2018-11-07T16:48:09+0000","annotations":[],"mimetype":"text\/html","reading_time":58,"domain_name":"mishadoff.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":433,"title":"A minimalist guide to tmux","url":"https:\/\/news.ycombinator.com\/item?id=15776995","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15776995\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15776995\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/medium.com\/@peterxjang\/a-minimalist-guide-to-tmux-13675fb160fa\" class=\"storylink\">A minimalist guide to tmux<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=medium.com\">medium.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        267 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=peterxjang\" class=\"hnuser\">peterxjang<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15776995\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15776995&amp;goto=item%3Fid%3D15776995\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=A%20minimalist%20guide%20to%20tmux&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=A%20minimalist%20guide%20to%20tmux\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15776995&amp;auth=7f66e0eb2263447e3a87bd2c095c8d2151792c1e\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15776995\">96\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:48:08+0000","updated_at":"2018-11-07T16:48:08+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":432,"title":"Ask HN: If you could automate anything, what would you automate?","url":"https:\/\/news.ycombinator.com\/item?id=15790562","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15790562\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15790562\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=15790562\" class=\"storylink\">Ask HN: If you could automate anything, what would you automate?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        31 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=juancampa\" class=\"hnuser\">juancampa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15790562\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15790562&amp;goto=item%3Fid%3D15790562\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20If%20you%20could%20automate%20anything%2C%20what%20would%20you%20automate%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20If%20you%20could%20automate%20anything%2C%20what%20would%20you%20automate%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15790562&amp;auth=9acd55c7467f4e3e6de8083e4ab35ab203190e31\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15790562\">85\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Try to ignore the cost of creating it. You can assume automation is free, money and time wise.<\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:48:07+0000","updated_at":"2018-11-07T16:48:07+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":430,"title":"Library of Blabber by nothke","url":"https:\/\/nothke.itch.io\/library-of-blabber","content":"<p>Explore the Library of Blabber, an infinite structure filled with endless bookshelves of eternal nonsense!<\/p><p>..Or so they say..<\/p><p>Can any sense be found?<br \/>Do answers lie somewhere in there?<br \/>How long will you stay sane looking for them?<br \/>Well, you'll have to flip the pages to find out..<\/p><p>Inspired by Jorge Luis Borges' story The Library of Babel, and Jonathan Basile's <a href=\"https:\/\/libraryofbabel.info\/\" target=\"_blank\" rel=\"nofollow noopener\">libraryofbabel.info<\/a> project, this \"little\" game is focused on exploration and the fun of meaningful discoveries in the noise of random data. Although it does not posses the mathematical correctness and \"purity\" of the original story, it still does contain enough pages of nonsense worthy of several hundred lifetimes..<br \/><\/p><p>Made for #procjam2015<br \/>by <a href=\"https:\/\/twitter.com\/Nothke\" rel=\"nofollow noopener\">Ivan Notaros<br \/>@Nothke<\/a><\/p>#procjam2016 update [016]<ul><li>New book types, including Markov chain stories and poems<\/li><li>Performance improvements, much much faster now<\/li><li>Page flipping is nicer<\/li><li>Improved wall textures, stone wall blocks now make much more sense<\/li><li>Books no longer pass through shelves when held<\/li><li>Many more tiny fixes<\/li><\/ul>","created_at":"2018-11-07T16:48:06+0000","updated_at":"2018-11-07T16:48:06+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"nothke.itch.io","preview_picture":"https:\/\/img.itch.zone\/aW1hZ2UvNDQ2NDMvMTkyNjgwLnBuZw==\/original\/ors4aU.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":431,"title":"Tour a virtual gallery in The Zium Museum","url":"https:\/\/www.rockpapershotgun.com\/2017\/11\/27\/the-zium-museum-is-a-virtual-gallery\/","content":"<div><main><p>Have you ever been wandering around a museum, turned a corner, and been baptised in a never-ending stream of confetti and balloons? I hadn\u2019t either until I took a stroll around The Zium Museum, <a href=\"https:\/\/theziumsociety.itch.io\/the-zium-museum\">a virtual gallery<\/a> \u201ccomposed of eclectic and wonderful things\u201d. There\u2019s more to see than confetti and balloons, promise.<\/p>\n<p>The gallery was curated by Michael Berto, who\u2019s used the work of 37 developers and artists to construct a virtual museum. I\u2019ve only walked around it for ten minutes, but you can bet I\u2019ll be heading back this evening for a more thorough investigation. <\/p>\n<p>In a neat touch, the entrance to the museum lets you toggle various ambient sounds on or off that can accompany your wanderings. I opted for the quiet bustling of a library, though was tempted by the trickling of a riverbed.<\/p>\n<p><img src=\"https:\/\/assets.rockpapershotgun.com\/images\/\/2017\/11\/Ziummuseum1-620x349.jpg\" alt=\"Ziummuseum1\" width=\"620\" height=\"349\" class=\"alignnone size-medium wp-image-497429\" srcset=\"https:\/\/assets.rockpapershotgun.com\/images\/2017\/11\/Ziummuseum1-620x349.jpg 620w, https:\/\/assets.rockpapershotgun.com\/images\/2017\/11\/Ziummuseum1-768x432.jpg 768w\" \/><\/p>\n<p>One exhibit consists of three halls with various artworks set up on plinths at the side. Some of those you can pick up and throw on the floor, in case the art isn\u2019t avant-garde enough for your tastes. Other exhibits transport you elsewhere, into cramped rooms with towering ceilings or onto a bridge that starts off in a blank void and then gets bathed in waves of different coloured light. Someone, somewhere, is going to explore the gallery while they\u2019re high, and I can only imagine the kind of things they\u2019ll see.<\/p>\n<p>The museum itself is an impressive virtual space, where you don\u2019t know what you\u2019re going to see around the next corner. It\u2019s made me remember that I actually quite like walking around real-life museums, though I\u2019m inevitably going to be disappointed when I next visit one and the exhibits and hallways don\u2019t transcend three dimensional space.<\/p>\n<p><img src=\"https:\/\/assets.rockpapershotgun.com\/images\/\/2017\/11\/Ziummuseum2-620x349.jpg\" alt=\"Ziummuseum2\" width=\"620\" height=\"349\" class=\"alignnone size-medium wp-image-497431\" srcset=\"https:\/\/assets.rockpapershotgun.com\/images\/2017\/11\/Ziummuseum2-620x349.jpg 620w, https:\/\/assets.rockpapershotgun.com\/images\/2017\/11\/Ziummuseum2-768x432.jpg 768w\" \/><\/p>\n<p>The Museum features work from many folks you might know. Cosmo D made surreal adventure game <a href=\"https:\/\/www.rockpapershotgun.com\/tag\/the-norwood-suite\/\">The Norwood Suite<\/a>. There\u2019s also a contribution from Fi Silva, who recently made a <a href=\"https:\/\/www.rockpapershotgun.com\/2017\/10\/02\/a-bright-light-in-the-middle-of-the-ocean-is-a-lovely-lighthouse-em-up\/\">lovely little lighthouse simulator<\/a>, as well as one from George Buckenham \u2013 who <a href=\"https:\/\/www.rockpapershotgun.com\/2015\/10\/02\/george-buckenham-interview\">Hannah Nicklin interviewed<\/a> in her excellent series on <a href=\"https:\/\/www.rockpapershotgun.com\/tag\/a-psychogeography-of-games\/\">the psychogeography of games<\/a>.<\/p>\n<p>There are plenty more big names where those came from. You\u2019ll find something from <a href=\"https:\/\/www.rockpapershotgun.com\/tag\/oleomingus\/\">Oleomingus<\/a>, a two-person team that\u2019s enthralled Alice with bits and pieces of their upcoming game <a href=\"https:\/\/www.rockpapershotgun.com\/tag\/somewhere\/\">Somewhere<\/a>. There\u2019s an offering from <a href=\"https:\/\/www.rockpapershotgun.com\/tag\/pippin-barr\/\">Pippin Barr<\/a>, who <a href=\"https:\/\/www.rockpapershotgun.com\/2017\/04\/03\/v-r-3-digital-water\">previously made<\/a> a virtual museum of his own displaying video game water. <a href=\"https:\/\/www.rockpapershotgun.com\/tag\/pol-clarissou\/\">Pol Clarissou<\/a> has also chipped in, who\u2019s previously unsettled us with <a href=\"https:\/\/www.rockpapershotgun.com\/2016\/11\/16\/the-nightmare-of-a-city-i-thought-i-knew\">nightmarish depictions of pixelated cities<\/a> and settled us back down again with his object twiddling game, <a href=\"https:\/\/www.rockpapershotgun.com\/2017\/03\/09\/vignettes\">Vignettes<\/a>. <\/p>\n<p>You can download The Zium Museum for free from <a href=\"https:\/\/theziumsociety.itch.io\/the-zium-museum\">from Itch.io<\/a>.<\/p>\n\t\t\t\t\t\t\t\t\t<\/main><\/div><div class=\"container-690\"><p class=\"disclaimer\">\n\t\t\t\t\tIf you click our links to online stores and make a purchase we may receive a few pennies. Find more information <a href=\"https:\/\/www.rockpapershotgun.com\/affiliate-linking-policy\/\">here<\/a>.\t\t\t\t<\/p><a href=\"https:\/\/www.rockpapershotgun.com\/support-us?utm_source=article-promo&amp;utm_medium=web&amp;utm_campaign=supporter\" class=\"support-us-promo\" rel=\"nofollow\">\n\t\t\n\t\t\t<img src=\"https:\/\/www.rockpapershotgun.com\/wp-content\/themes\/rps-2018\/images\/logo-inverted.svg\" alt=\"Rock Paper Shotgun\" \/>\n\t\t\n\t\t\n\t\t\t\n\t\t\t\tSupport RPS and get an ad-free site, extra articles, and free stuff!\t\t\t\n\t\t\tTell me more\n\t\t\n\t\t\n\t<\/a><\/div>","created_at":"2018-11-07T16:48:06+0000","updated_at":"2018-11-07T16:48:06+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/assets.rockpapershotgun.com\/images\/\/2017\/11\/Ziummuseumheader-620x300.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":428,"title":"16 Lessons Learned Bootstrapping Kinsta From $0 to 7-Figures in Revenue","url":"https:\/\/kinsta.com\/blog\/bootstrapping-startup\/","content":"<p>We\u2019ve been asked a lot by our clients over the years why Kinsta was born, how we started it, and how we got to the level at which we are now. Each company and startup has its own unique story, and personally, I love to read them because I always find a few hidden gems, tips, and how-tos we can use and apply to our own business.<\/p><p>This article is about the <strong>lessons we learned during our bootstrapping journey<\/strong>, and how we turned our startup into a 7-figure business.\u00a0Some of our decisions turned out to be great, while others cost us a lot of money and time. But I want to show you that it\u2019s possible to build a profitable SaaS business even if you don\u2019t have any connections, resources, or money.<\/p><p>For those of you who have successfully bootstrapped a business from the ground up, some parts of this post will sound familiar and you\u2019ll probably relate to what I\u2019m talking about. For those of you who are just looking to start your first business, I\u2019m going to try and show you the pros and cons of bootstrapping, what it looks like, and what worked for us. Hopefully, you\u2019ll find something that is applicable to your business.<\/p><div class=\"cta-mini\"><h4 class=\"cta-mini--title\">Still looking for that perfect WordPress host?<\/h4><p>At Kinsta, we do hosting different than the rest. We\u2019ve combined the fastest platform in the world (Google Cloud) with the best support team in the industry. Ready to take your site the next level?<\/p><p><a class=\"cta-mini--link\" href=\"https:\/\/kinsta.com\/plans\/\">Check out our plans<\/a><\/p><\/div><p>&#13;\n\t\t\t&#13;\n\t\t\tIt's possible to build a profitable #SaaS business without any connections, resources, or money. \ud83e\udd18&#13;\n\t\t\t&#13;\n\t\t\tClick to Tweet&#13;<\/p><p>To be clear, don\u2019t expect any secret growth hacks that nobody has ever used before or silver bullets that will earn you instant big media mentions and thousands of signups. <strong>We\u2019re not smarter<\/strong> or more clever than you. Your success really all depends on you and how much you want to succeed.<\/p><p>Achieving millions of annual recurring revenue (ARR) from zero feels good. In fact, if feels great.\u00a0? But once you reach this you\u2019ll quickly move on and start thinking about your next steps, your next product, service, add-on or marketing channel that will bring you a buttload of new leads and customers. Your <strong>next goal suddenly becomes the 8-figure revenue mark<\/strong> and how to get that.<\/p><p>But it\u2019s important to note that <strong>running a business is not just about the money<\/strong> or X figures of revenue. It\u2019s about building a team, solving problems continuously day after day, driving innovation, hiring and firing, responsibility, self-learning, self-management and so much more. It\u2019s something that cannot be learned in school or taught by anyone. You need to do it, practice, solve problems, make mistakes, learn from them and improve.<\/p><p>I hope this post will shed some light on bootstrapping and will inspire more of you to start a business.<\/p><h2>Alright, Shall We Begin?<\/h2><p>First, let\u2019s start with a little background that might add context to this story.<\/p><p>Kinsta was founded by four of us; Mark, Anita, Peter and myself. Back in 2013, we provided web development and online marketing services for small and medium-sized businesses. It was fine but wasn\u2019t fun and most importantly we discovered that business model wasn\u2019t scalable. We had to find new clients each month to maintain the revenue, but due to the limited resources, we couldn\u2019t start more projects which were required for growth. We knew we had to change this.<\/p><p>At the time, we were working exclusively with the <a href=\"https:\/\/kinsta.com\/blog\/why-use-wordpress\/\" target=\"_blank\" rel=\"noopener\">WordPress<\/a> content management system. Why? Because it\u2019s free, has tens of thousands of extensions (<a href=\"https:\/\/wordpress.org\/plugins\/\" target=\"_blank\" rel=\"noopener\">plugins and themes<\/a>), and you can relatively quickly build any type of website using it. However, after we finished a project the client usually asked the following:<\/p><blockquote><p>Can you provide hosting for my website? I\u2019m not a sysadmin and don\u2019t want to deal with it.<\/p><\/blockquote><p>And we replied no, we don\u2019t provide web hosting. 20 similar cases later we clearly saw that this was a business opportunity and after several discussions between us, we finally decided to give it a try.<\/p><p>Now, 4 years later, <a href=\"https:\/\/kinsta.com\" target=\"_blank\" rel=\"noopener\">Kinsta<\/a> provides web hosting for thousands of WordPress sites. Our clients range from small businesses and blogs up to large Inc. 500\/5000 enterprises. It\u2019s a <a href=\"https:\/\/kinsta.com\/blog\/managed-wordpress-hosting\" target=\"_blank\" rel=\"noopener\">managed solution<\/a>, meaning we take care of the infrastructure, server updates, protection of your site, analytics, make sure they load blazingly fast, and provide 24\u00d77 support.<\/p><p>What I can tell you is that we never ever imagined how hard it would be, how many issues we would face on a weekly or monthly basis, and how much sacrifice would be needed to make it work. We\u2019re a <strong>100% bootstrapped<\/strong> company. When we decided to give Kinsta a try back in 2013 we didn\u2019t have any money in the bank, just what was in our pockets.<\/p><p>Previously we\u2019ve written about the <a href=\"https:\/\/kinsta.com\/blog\/vc-route\/\" target=\"_blank\" rel=\"noopener\">benefits of bootstrapping and not going the VC route<\/a>. Since Kinsta was founded, many VCs have reached out (and keep reaching out) for a potential partnership. But in the first few years, <strong>we wanted to own 100% of our business<\/strong> and were absolutely fine with slower growth that\u2019s more controlled. We\u2019re not against VCs at all! They\u2019re great and necessary in a lot of cases. Many businesses in the IT sector (and many others) wouldn\u2019t even exist without them.<\/p><p>But when it comes to raising money, you have to <strong>weigh the pros and cons<\/strong> and prioritize things that are important to you. We stuck to our original idea and so we\u2019re a self-funded (I like the term <em>client funded<\/em>) and bootstrapped business.<\/p><p>Did you know that <a href=\"http:\/\/fortune.com\/2014\/09\/25\/why-startups-fail-according-to-their-founders\/\" target=\"_blank\" rel=\"noopener\">nine out of ten<\/a> startups fail? This is the trend and perhaps the nature of the business. If you\u2019re a bootstrapper many of the lessons below will sound familiar to you (hello soulmates). If you\u2019re just planning to launch your new project, keep reading and learn how to help your startup survive and grow. Below are the most important lessons we\u2019ve learned bootstrapping Kinsta from a startup into a business.<\/p><h2>1. Build a Brand From Day One<\/h2><p>I\u2019m sure you\u2019ve heard the term blue and red ocean strategy or blue and red ocean companies. Red oceans are the industries in existence today where companies try to outperform each other and grab a greater share of the existing market. Cutthroat competition turns this ocean bloody red \u2013 similar to a nice playground of sharks. But you know what? A crowded market with lots of competitors means people need that service or product and they are willing to pay money for it.<\/p><div id=\"attachment_15000\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15000\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/red-ocean-strategy-vs-blue-ocean-strategy.jpg\" alt=\"Red ocean strategy vs blue ocean strategy\" width=\"1398\" height=\"787\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/red-ocean-strategy-vs-blue-ocean-strategy.jpg 1398w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/red-ocean-strategy-vs-blue-ocean-strategy-300x169.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/red-ocean-strategy-vs-blue-ocean-strategy-768x432.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/red-ocean-strategy-vs-blue-ocean-strategy-1024x576.jpg 1024w\" \/><p class=\"wp-caption-text\">Red ocean strategy vs blue ocean strategy<\/p><\/div><p>Then you have blue ocean companies, which operate in an untapped market space and create demand while having the opportunity for rapid growth. The competition is low or non-existent.<\/p><p>If your new venture is a blue ocean company that\u2019s great; you can be an industry leader and dominate your niche. However, blue ocean often means the customer pool is low and sometimes it\u2019s a completely new field. There might not even be a need for your service or product yet.<\/p><p>Chances are a lot higher that you\u2019ll find yourself in a red ocean environment among established companies with years of advantage on their side. They have revenue, a large client base, established social\u00a0media presence, and big marketing budgets. So how can you compete with them in the long-term? The only way is if you have a brand and you consciously build it from the beginning. A brand that your potential user base can recognize and remember.<\/p><p><strong>You can only differentiate yourself from your competitors if you have a brand<\/strong>. I see so many companies out there using non-brandable names (general terms\/location+keywords), such as michigandesigner.com, denverdesignagency.com or bestwphosting.com.<\/p><p>Even if these generic-branded companies provide a great service, I probably wouldn\u2019t spend money with them as there is a lack of trust. If you do business, do it as a brand and stand out from the crowd by having your own brandable name.<\/p><p>I remember we spent several weeks coming up with the final name \u201cKinsta.\u201d We weren\u2019t sitting 24\/7 staring each other\u2019s eyes, we were just thinking about it while working on other tasks. You need to find something unique, short, and brandable.<\/p><h2>2. Getting Your First 100 Customers Is the Hardest Journey of Your Life<\/h2><p>Your website is ready (actually it\u2019s never ready, but good enough to start) and you\u2019re waiting for your first clients to come and pay for your service. The sad reality is that they won\u2019t come unless you make sure they know about you. Your business is brand new, you don\u2019t have any clients, testimonials, or company history, so why would they trust you? I\u2019m sure you have heard of Moz, they have some cool SEO tools. Their founder and content marketing wizard, Rand Fishkin, said the following in one of his <a href=\"https:\/\/moz.com\/blog\/the-greatest-misconception-in-content-marketing-whiteboard-friday\" target=\"_blank\" rel=\"noopener\">Whiteboard Friday presentations<\/a>:<\/p><div id=\"attachment_15001\" class=\"wp-caption aligncenter\"><img class=\"wp-image-15001 size-full\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/moz-whiteboard-friday-e1511470370428.png\" alt=\"Moz whiteboard Fridy\" width=\"798\" height=\"357\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/moz-whiteboard-friday-e1511470370428.png 798w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/moz-whiteboard-friday-e1511470370428-300x134.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/moz-whiteboard-friday-e1511470370428-768x344.png 768w\" \/><p class=\"wp-caption-text\">Moz Whiteboard Friday<\/p><\/div><p><strong>It takes them 7.5 visits to convert someone into a free trial.<\/strong> Not into a paid customer, just sign them up for a free trial. And Moz has a crazy strong brand and is a well-known player in the SEO industry.<\/p><p>We faced the exact same situation in our first few months in late 2013 and early 2014. Leads were coming in each day asking the following:<\/p><blockquote><p>Could you show me some of your current clients, or any clients from XYZ industry?<\/p><\/blockquote><p>It was hard and frustrating. Even if you\u2019re a very well connected guy with hundreds of friends and business contacts, it\u2019s hard to get your first paying customers. Crazy hard. Here are a few things we did to go about getting our first customers:<\/p><ul><li><strong>Give away a lot of value, like free content and show that you are an expert<\/strong> \u2013 Kickstart your blog section and provide some badass content. Blog often and make sure that what you write is top-notch and useful for your target audience. <a href=\"https:\/\/kinsta.com\/learn\/content-marketing\/\" target=\"_blank\" rel=\"noopener\">Content marketing really works and converts<\/a> when it\u2019s done right. See more about this in the content marketing section below.<\/li>\n<li><strong>Give away some free accounts<\/strong> \u2013 We did this and it worked out pretty good. We were able to partner up with some big publishers and in exchange for free hosting, showcase them on our <a href=\"https:\/\/kinsta.com\/clients\/\" target=\"_blank\" rel=\"noopener\">client\u2019s page<\/a> along with their awesome testimonials. Plus, being early users, they provided us with some invaluable product feedback!<\/li>\n<li><strong>Email is free, don\u2019t be afraid to use it<\/strong> \u2013 In the early days, I was responsible for our marketing efforts. I\u2019ve built a lot of relationships with other WordPress business owners and bloggers just by reaching out with a cold email to start a conversation. I connected with a lot of people and am still doing this today. I remember we got many of our first clients this way. For example, I reached out to <a href=\"https:\/\/www.matthewbarby.com\/\" target=\"_blank\" rel=\"noopener\">Matthew Barby<\/a>, a very well-known SEO expert. He\u2019s now the head of Growth at Hubspot.com. I found one of his case studies extremely helpful and reached out to him just to thank him for the post. It turned out he was looking for a new hosting provider as he had outgrown his current environment.<\/li>\n<\/ul><p>And 3+ years later he\u2019s still our honored customer. See? It all started with a simple email.<\/p><p>The point is that even without any resources, you have options and there are always some creative ways to make some waves and get that ball rolling. As soon as you have your very first paying customer it gets easier to get new ones on board. But getting your first 10 clients takes the same amount of time as getting your first 100!<\/p><h2>3. Find the Best Talent Globally (and Even Save Some Money)<\/h2><p>We have a distributed team of awesome people all around the world. We have team members in the US, Europe, Australia, India, Indonesia, Nicaragua and so on. Below are a few reasons why we went with a distributed team for Kinsta.<\/p><p>Due to the nature of our service, we have to <strong>provide 24\/7 support 365 days a year<\/strong>. Not only that, we have to be very responsive! We have clients from all over the world, meaning we need to cover all time zones. If you only work with a local team and need to provide 24\/7 global coverage, it could be difficult to manage. However, if you don\u2019t need to provide this level of support, searching globally for your next employees can still be a good idea.<\/p><p>Let\u2019s say you are looking for a developer, DevOps, content marketer or sales agent. The reality is that you probably won\u2019t find the most talented people in your local hometown or even in your own country. And even if you found one or two, there probably aren\u2019t many of them. This tweet below pretty much sums up the benefits of having a partially remote team.<\/p><p>Many jobs in the IT industry can be done from home and <a href=\"https:\/\/kinsta.com\/blog\/working-remotely\/\" target=\"_blank\" rel=\"noopener\">working remotely is now very common<\/a>. <strong>Talented people can manage their time<\/strong> and can be very productive working from a home office. You can also save a lot of money by not renting a huge office downtown.<\/p><p>And did I mention the savings on salaries? A good back-end developer from the West Coast in the United States would cost you 2x or even more than someone from Central Europe. Salaries are the biggest part of business expenses. If you add up these amounts you\u2019ll quickly realize how much money you can save on a yearly basis and invest that money back into hiring other people, or spend more on marketing and product development. Hiring remote workers can have a <strong>huge financial implication<\/strong>.<\/p><p>Another piece of advice is that you should\u00a0<strong>hire people smarter than you<\/strong> if you want to succeed. You can be good in coding, sales, and hiring, but you\u2019ll never be excellent. Don\u2019t be a control freak, let someone who can do it better than you, take it to the next level.<\/p><p>&#13;\n\t\t\t&#13;\n\t\t\tYou should and need to hire people smarter than you \ud83e\udd13 if you want to succeed.&#13;\n\t\t\t&#13;\n\t\t\tClick to Tweet&#13;<\/p><h2>4. Build MVP, Launch and Market Early<\/h2><p>MVP (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Minimum_viable_product\" target=\"_blank\" rel=\"noopener\">minimum viable product<\/a>) is a common term when it comes to launching your product. Essentially it is a product with<strong> just enough features<\/strong> to satisfy early customers. If you\u2019re a perfectionist, you probably hate this idea, but let\u2019s move on and get familiar with it. Otherwise, you\u2019ll never launch, and will continue to have more delays and other deadlines. You\u2019ll also miss some early invaluable customer feedback, ability to gauge the market, revenue, and early customers.<\/p><div id=\"attachment_15004\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15004\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/minimum-viable-product.png\" alt=\"Minimum viable product vs product\" width=\"996\" height=\"404\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/minimum-viable-product.png 996w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/minimum-viable-product-300x122.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/minimum-viable-product-768x312.png 768w\" \/><p class=\"wp-caption-text\">Minimum viable product vs product<\/p><\/div><p>We were faced with the same issue and first, it was hard to accept that our product would never be perfect or completely ready. Or that we wouldn\u2019t be able to deliver all the features that we wanted in the first 6 months.<\/p><p>So we built an MVP, which was lacking many features compared to our competitors, but it was good enough to start and market. We didn\u2019t have a <a href=\"https:\/\/kinsta.com\/knowledgebase\/staging-environment\/\" target=\"_blank\" rel=\"noopener\">staging area<\/a>, automated SSL, 1 global <a href=\"https:\/\/kinsta.com\/knowledgebase\/google-cloud-data-center-locations\/\" target=\"_blank\" rel=\"noopener\">data center locations<\/a>, premium DNS integration, or a detailed analytics breakdown in the admin dashboard. But step by step we added all of these. And while we were working on these new must-have features, new clients were signing up for the service each day. We\u2019re really happy that we decided to go this route and launched early. If we had to do it over again, we would approach it the same way.<\/p><p>My advice is to <strong>start marketing early<\/strong>. Test the market. Get early feedback and learn from it. Adapt, change, and make improvements. As Steve Blank <a href=\"https:\/\/hackernoon.com\/an-mvp-is-not-a-cheaper-product-it-s-about-smart-learning-77eed770f60c\" target=\"_blank\" rel=\"noopener\">said<\/a>, MVP is not a cheaper product, it\u2019s about smart learning. Also, this advice from\u00a0<a href=\"https:\/\/producthabits.com\/why-trello-failed-to-build-a-1-billion-business\/\">Hiten Shah<\/a>\u00a0below is spot on. In SaaS, you win by continually <strong>solving the problem better<\/strong>.<\/p><div id=\"attachment_15005\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15005\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/saas-quote.jpeg\" alt=\"SaaS quote from Hiten Shah\" width=\"1024\" height=\"512\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/saas-quote.jpeg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/saas-quote-300x150.jpeg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/saas-quote-768x384.jpeg 768w\" \/><p class=\"wp-caption-text\">SaaS quote from Hiten Shah<\/p><\/div><p>Remember, you\u2019re in the long-term game. So launch early and keep solving the problem your users are paying for.<\/p><h2>5. Follow the \u201cHire Slow Fire Fast\u201d Principle<\/h2><p>This one sounds obvious, but not many people follow it. <strong>Your employees are the heart of your company<\/strong>. Without them, you couldn\u2019t have the best app or best software. They interact with your customers, they represent your brand, and are responsible for the continued growth of your business. Make sure you treat them well.<\/p><p>Whenever we are about to hire a new team member, we always craft a very detailed job description, post it to several portals, online groups, forums and then wait. Wait to receive 50, 100 or even more applications. Make sure you have a big pool of candidates because that\u2019s how you\u2019ll find your next dream employee.<\/p><p>After that, we start narrowing down the list to the top 5 or 10 candidates. We have multiple rounds of job interviews, usually 2 or 3. Thanks to this we are able to narrow down our list to the top 3 candidates and make the final decision. Yes, this process takes time, usually weeks. If you\u2019re not in a hurry you should follow this slow method, trust me, it\u2019s worth the time.<\/p><p>Like most companies, we\u2019ve also had a few bad hires, but we know why. We didn\u2019t follow the hiring process mentioned above. We were short on time and had to hire someone quickly and couldn\u2019t wait weeks or months to find the perfect candidate. If you\u2019re in a hurry you are more likely to make a mistake.<\/p><p><img class=\"aligncenter wp-image-15053\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/bad-employee.gif\" alt=\"bad employee\" width=\"515\" height=\"277\" \/><\/p><p>Once you find out that particular person doesn\u2019t fit into your team and company culture, it\u2019s better to <strong>say goodbye as soon as possible<\/strong> and move on. When I say ASAP, I mean within a few days after you realized it\u2019s not going to work. First of all, you don\u2019t have much money and you don\u2019t want to spend it on an employee who doesn\u2019t fit into your team. Second, it doesn\u2019t make any sense to talk about things over and over, trying to change their attitude or personality. Say goodbye and move on.<\/p><h2>6. Know Your Numbers and Focus on Profitability<\/h2><p>This step is really crucial. If you\u2019re self-funded then you must always know your most important financial metrics. You can only spend the money you made last month, as you don\u2019t have any outside money to invest back in. You\u2019ll quickly <strong>learn the importance of cash<\/strong>. Spending money that depends on revenue you generate is a lot different than spending someone else\u2019s money (like VC dollars).<\/p><p>Or think of it like this:\u00a0<em>\u201cAs a kid, you spent your summer job earnings more wisely than the \u201ceasy\u201d money your grandma gave you, didn\u2019t you?<\/em> \u201c<\/p><p>As a bootstrapper you need to generate revenue, otherwise, <strong>you won\u2019t survive<\/strong>. You\u2019re naturally more careful. If you keep operating from your hard-earned revenue, in this case, it also means that you\u2019re profitable. And now that you have profit, you can start thinking about it how to distribute it.<\/p><p>In our case, we re-invest as much as possible. We keep hiring new team members all the time and spend more and more on content creation + promotion and paid customer acquisition. The great news is that what you do with the extra money is up to you.<\/p><div id=\"attachment_15006\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15006\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/mrr.png\" alt=\"Kinsta MRR (monthly recurring revenue)\" width=\"1244\" height=\"317\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/mrr.png 1244w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/mrr-300x76.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/mrr-768x196.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/mrr-1024x261.png 1024w\" \/><p class=\"wp-caption-text\">Kinsta MRR (monthly recurring revenue)<\/p><\/div><p>As our MRR grows we can re-invest more and more. Make sure you know the basics of simple\u00a0<a href=\"https:\/\/baremetrics.com\/academy\/glossary\" target=\"_blank\" rel=\"noopener\">financial terms<\/a>. If someone asks you how much your churn rate is, you shouldn\u2019t answer. Instead, you should ask them <em>\u201cDid you mean revenue churn or user churn?\u201d<\/em> You\u2019d be surprised how many founders and entrepreneurs <strong>don\u2019t know the difference between these two metrics<\/strong>.<\/p><p>At Kinsta we use <a href=\"https:\/\/baremetrics.com\/\" target=\"_blank\" rel=\"noopener\">Baremetrics<\/a>. These awesome guys provide analytics for Stripe. They built a very intuitive dashboard for processing all of your data from Stripe (and from other sources) and turn them into charts and metrics you can understand, and then make business decisions based on the data. <a href=\"https:\/\/demo.baremetrics.com\/\" target=\"_blank\" rel=\"noopener\">Here\u2019s a demo<\/a>, make sure you take a look. Another tool I use and recommend is <a href=\"https:\/\/www.profitwell.com\/\" target=\"_blank\" rel=\"noopener\">Profitwell<\/a>, created by Patrick Campbell. It\u2019s completely free!<\/p><p>I carefully watch and try to <strong>keep our expenses as low as possible.<\/strong> It\u2019s one of the keys to making things work. We use a lot of <a href=\"https:\/\/kinsta.com\/blog\/saas-products\/\" target=\"_blank\" rel=\"noopener\">SaaS products<\/a> to power our business. From time to time we review them and make sure we pay only for those we actually use and need to run the business. Regardless of your size, I\u2019d advise you do the same, cancel everything you don\u2019t use or that doesn\u2019t provide any value for your business.<\/p><p>When you\u2019re just starting your business you\u2019ll have a very tight budget. If you really need a service to run your business, don\u2019t forget to ask for discounts when possible. You shouldn\u2019t be ashamed, it\u2019s just a question and you\u2019d be surprised how many times companies say yes. Remember, it\u2019s business and everything is <a href=\"https:\/\/www.groovehq.com\/blog\/startup-negotiation\" target=\"_blank\" rel=\"noopener\">negotiable<\/a>. If you cannot pay for it, you can offer them some level of co-promotion or a partnership that\u2019s beneficial for both parties.<\/p><p>Having a limited budget will make you use your money smart, and\u00a0<strong>develop the ability to solve issues in a creative way<\/strong>. It\u2019ll force you to operate profitably from day one. Focus on profitability, otherwise, why are you building a company?<\/p><p>Also, we have a detailed post on pricing make sure you take a look:\u00a0<a href=\"https:\/\/kinsta.com\/blog\/race-to-the-bottom\/\" target=\"_blank\" rel=\"noopener\">How a Race to the Bottom Hurts Your Business\u2019s <\/a>Bottom Line.<\/p><h2>7. \u201cI Can\u2019t Solve This\u201d Is Not an Option<\/h2><p>I\u2019d suggest you to delete the phrases \u201cI can\u2019t,\u201d \u201cI don\u2019t know,\u201d and \u201cI give up\u201d from your dictionary. It\u2019s not an option. No matter how hard and unsolvable a situation you are facing seems, you\u2019ll find a solution. Sometimes it takes weeks, months, or a half year to get it fixed, but you\u2019ll come up with a solution. After you put countless hours of hard work into your company, failing is not an option.<\/p><p>How many times were we faced with a hard situation? Many many times. And you\u2019ll find yourself in the same situation no matter what you think.<\/p><p>Just to mention one example. We were already operating when one day without any notice our Stripe account was suspended due to some fraudulent activity which was out of our control. Stripe\u2019s support replied to our case with a scripted response saying our account was closed permanently and there is no way to reactivate it. You\u2019ve got to be kidding me, right?<\/p><p>We weren\u2019t able to charge credit cards or collect our revenue. We quickly started looking for a solution, however, integrating a new payment processor into your existing system is not a 5-minute job. Despite that, we started to work on it and in the meantime, we crafted this <a href=\"https:\/\/kinsta.com\/blog\/startups-avoid-stripe\/\" target=\"_blank\" rel=\"noopener\">blog post telling our story<\/a> to the startup community. It was soon trending on HN and Reddit. Stripe then reached out to us to <a href=\"https:\/\/kinsta.com\/blog\/how-we-got-our-stripe-account-back\/\" target=\"_blank\" rel=\"noopener\">solve the case<\/a>, because no one likes bad press. You see, first, it looked there is no way to get it solved and within 48 hours we were back in business.<\/p><h2>8. It\u2019s Not a 9 to 5 Job. Prepare to Work Crazy Hours<\/h2><p>Don\u2019t have any misconceptions, success doesn\u2019t come easy. You\u2019ll have to work your a!?*^\u00a3%^ off to get results. It often means sitting 16 hours a day in front of your computer, replying to support tickets, talking to leads, and closing deals. At the beginning, it\u2019s better if you forget about weekends\u00a0or holidays. I remember back in December 2015, one of our server providers experienced a huge DDoS attack resulting 2-3 days of downtime for our entire data centers. And it happened during the Christmas rush.<\/p><div id=\"attachment_15008\" class=\"wp-caption aligncenter\"><img class=\"wp-image-15008 size-full\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/ddos.png\" alt=\"DDoS\" width=\"775\" height=\"244\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/ddos.png 775w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/ddos-300x94.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/ddos-768x242.png 768w\" \/><p class=\"wp-caption-text\">DDoS (Img src: <a href=\"https:\/\/status.linode.com\/incidents\/tm2b0tl5yc9g\" target=\"_blank\" rel=\"noopener\">Linode status<\/a>)<\/p><\/div><p>Best Christmas ever\u2026<\/p><p>We were sitting in front of our computers 24\u00d77 replying to customers and moving sites all day long to other providers to get the sites back online as soon as possible. You can imagine this wasn\u2019t the Christmas gift we were hoping for, but we\u2019ll never forget it that\u2019s for sure. And most importantly we learned a valuable lesson, and it confirmed our decision that we needed to <a href=\"https:\/\/kinsta.com\/blog\/managed-wordpress-hosting-in-the-google-cloud\/\" target=\"_blank\" rel=\"noopener\">move our service to the cloud<\/a>.<\/p><p>For bootstrappers the first few years are tough. Countless entrepreneurs are out there stating that you can build a business from scratch just by working a few hours a day or a week. BS, I don\u2019t believe them and you shouldn\u2019t either, it\u2019s an illusion. We personally know how hard it is to build your business on your own and I\u2019m sure many of you can attest to this.<\/p><p><img class=\"aligncenter wp-image-15007\" title=\"Work long hours\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/work-long-hours.gif\" alt=\"work long hours\" width=\"437\" height=\"317\" \/><\/p><p>When I head to the bed at 12 a.m. or later I still have emails I should have replied to, 5+ tabs open with industry news, articles, and tutorials I didn\u2019t have time to read that day, and Trello cards to contribute to. And there is always something else I should have taken care of.<\/p><p>It doesn\u2019t mean my productivity is low or I cannot manage myself, this is the result when you wear several hats at the same time and you don\u2019t have enough time to finish everything by the end of your workday. But that\u2019s fine and let\u2019s say fun as it keeps challenging you each and every day.<\/p><p>If you\u2019re not willing to work crazy hours at the beginning, well maybe you shouldn\u2019t start a bootstrapped business. If you prefer a 9 to 5 day that\u2019s fine, I completely understand and respect that. It has its own pros. But when it comes to bootstrapping and building something from the ground up you have to go the extra mile and work twice as much and twice as hard as anyone else.<\/p><h2>9. Dedicate All Your Available Time to Your Business<\/h2><p>My advice would be, if you\u2019ve found a good market fit and you truly believe in your idea, <strong>dedicate all your attention and resources<\/strong> to it. Don\u2019t treat it as a side project or weekend hobby. What could happen if you fail? Nothing, but at least you won\u2019t regret in the future not giving it an honest try.<\/p><p>Back in 2014, Nathan Barry, the founder of <a href=\"https:\/\/convertkit.com\/\">ConvertKit<\/a> also <a href=\"http:\/\/nathanbarry.com\/5k\/\" target=\"_blank\" rel=\"noopener\">had to make this decision<\/a>. ConvertKit is an email marketing solution for bloggers. Nathan had some early success, he had paying customers and some early MRR but ConvertKit didn\u2019t deliver the results he was hoping for. He treated it as a side project, which wasn\u2019t an option anymore. It was decision time. He had two choices, shut it down or focus and work 100% on it.<\/p><div id=\"attachment_15009\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15009\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/convertkit-decision-time.png\" alt=\"ConvertKit - decision time\" width=\"656\" height=\"322\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/convertkit-decision-time.png 656w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/convertkit-decision-time-300x147.png 300w\" \/><p class=\"wp-caption-text\">ConvertKit \u2013 decision time<\/p><\/div><p>He decided to give it a try and he dedicated all his attention to ConvertKit. Every other project became secondary and he started working full time on the business. What happened, did he succeed? Well, today they are growing at a very good rate and having a whopping <a href=\"https:\/\/convertkit.baremetrics.com\/stats\/mrr\" target=\"_blank\" rel=\"noopener\">$800k MRR<\/a>. Not bad.<\/p><h2>10. Invest Heavily In Content Marketing and SEO<\/h2><p>If you\u2019re a bootstrapped company you probably don\u2019t have any marketing budget at all, unless you belong to one of the lucky ones (congratulations btw!). Either way, your budget is probably very limited.<\/p><p>Starting a PPC campaign and targeting your money keywords or your competitor\u2019s keywords with a 10k+ monthly budget is not an option. Neither is spending thousands on Facebook ads or purchasing a few million banner impressions on industry-leading blogs. But it doesn\u2019t mean you can\u2019t spread the word about your products, services, and knowledge.<\/p><p>There is this thing called <strong>content marketing<\/strong>. I\u2019m sure you\u2019ve heard this at least 3x today and probably 95 times this week, but the reality is that this stuff truly works and converts when it\u2019s done the way it should be.<\/p><p>Since day one we took our <a href=\"https:\/\/kinsta.com\/blog\" target=\"_blank\" rel=\"noopener\">blog section<\/a> very seriously, as we didn\u2019t have any other options. Here is a screenshot from our Google Analytics showing how the traffic grew over the years. As you can, see it takes some time to build up your \u201ccontent reputation.\u201d\u00a0And building an audience takes even more time and commitment. But after the first 12-16 months, the traffic will grow at a faster rate. Getting good organic traffic takes the longest but pays off the most over time.<\/p><div id=\"attachment_15010\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15010\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/organic-traffic-growth.png\" alt=\"Organic traffic growth\" width=\"1011\" height=\"171\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/organic-traffic-growth.png 1011w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/organic-traffic-growth-300x51.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/organic-traffic-growth-768x130.png 768w\" \/><p class=\"wp-caption-text\">Organic traffic growth<\/p><\/div><p>The sooner you begin the sooner you\u2019ll see results. We love to publish in-depth, very detailed guides because we believe in quality over quantity.\u00a0<strong>2 million blog posts are published every day<\/strong> but the reality is that most of them are never read, shared, or linked to by anyone. They are just too general, short, and low-quality and don\u2019t provide any unique value. These posts are a waste of time and resources.<\/p><p>So what should you do? Instead of writing 5\u00d7500 word posts a week which take up a lot of your time, research the topic and write a 5000-word\u00a0bad-ass guide that\u2019s share-worthy and provides real value for your readers. Or publish just 2 posts per month but make sure they\u2019re really in-depth and the best resource in your industry. By following this principle we were able to <a href=\"https:\/\/kinsta.com\/blog\/wordpress-seo-checklist\/\" target=\"_blank\" rel=\"noopener\">improve our organic traffic by 571% in 13 months<\/a>!<\/p><p>To show you what I mean here are 3 well-performing posts on our blog:<\/p><p><a href=\"https:\/\/kinsta.com\/blog\/wordpress-seo-checklist\/\" target=\"_blank\" rel=\"noopener\">WordPress SEO Checklist \u2013 45 Tips for 2017<\/a>, 10,007 words<\/p><div id=\"attachment_15011\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15011\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-1.png\" alt=\"Blog post example 1\" width=\"1202\" height=\"192\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-1.png 1202w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-1-300x48.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-1-768x123.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-1-1024x164.png 1024w\" \/><p class=\"wp-caption-text\">Blog post example 1<\/p><\/div><p><a href=\"https:\/\/kinsta.com\/blog\/google-cloud-hosting\/\" target=\"_blank\" rel=\"noopener\">Top 7 Advantages of Choosing Google Cloud Hosting<\/a>, 3620 words<\/p><div id=\"attachment_15012\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15012\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-2.png\" alt=\"Blog post example 2\" width=\"1197\" height=\"175\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-2.png 1197w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-2-300x44.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-2-768x112.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-2-1024x150.png 1024w\" \/><p class=\"wp-caption-text\">Blog post example 2<\/p><\/div><p><a href=\"https:\/\/kinsta.com\/blog\/google-pagespeed-insights\/\" target=\"_blank\" rel=\"noopener\">How to Score 100\/100 in Google PageSpeed Insights<\/a>, 4842 words<\/p><div id=\"attachment_15013\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15013\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-3.png\" alt=\"Blog post example 3\" width=\"1208\" height=\"172\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-3.png 1208w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-3-300x43.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-3-768x109.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/blog-post-3-1024x146.png 1024w\" \/><p class=\"wp-caption-text\">Blog post example 3<\/p><\/div><p>You can see these guides get a lot of social shares, acquire links, and drive a lot of organic traffic each day as they rank for dozens of keywords. These posts are very detailed with a lot of visual elements, they are easy to understand, and you can implement the tips immediately. They provide real value for the readers. This is what makes you an authority in your industry. And most importantly this traffic converts readers into customers, as well as newsletter and <a href=\"https:\/\/kinsta.com\/blog\/wordpress-push-notifications\/\" target=\"_blank\" rel=\"noopener\">push notification<\/a> subscribers.<\/p><p>If you don\u2019t give up your content marketing efforts too early your Google Analytics live overview will look like this throughout most of the day:<\/p><div class=\"in-post-container\"><div class=\"vertical newsletter-promo flip-container\"><div class=\"flipper\"><div class=\"front\"><div><div class=\"default\"><div><p>We grew our traffic 1,187% with WordPress.We\u2019ll show you how.<\/p><p>Join 20,000+ others who get our weekly newsletter with insider WordPress tips!<\/p><\/div>Subscribe Now<\/div><\/div><\/div><div class=\"back\"><div><p>Subscribe to the Kinsta newsletter<\/p><\/div><\/div><\/div><\/div><\/div><div id=\"attachment_15014\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15014\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/google-analytics-real-time-visitors.png\" alt=\"Google Analytics - real-time visitors\" width=\"1006\" height=\"297\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/google-analytics-real-time-visitors.png 1006w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/google-analytics-real-time-visitors-300x89.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/google-analytics-real-time-visitors-768x227.png 768w\" \/><p class=\"wp-caption-text\">Google Analytics \u2013 real-time visitors<\/p><\/div><p>Content marketing is a long-term game but it can drive you a\u00a0crazy amount of quality, targeted, and free traffic which converts into paying customers.<\/p><p>If you have money, ads are great as they can drive you instant targeted traffic, but these are one-time clicks and you have to keep your wallet open to maintain the level of the traffic. Having a great blog or resources section which is up to date and you add new content to on a regular basis can have a huge impact on your business. Buffer, Copyblogger, and Moz have all built an empire based on nothing but content. If you would like to learn more on how to get started, take a look at\u00a0<a href=\"http:\/\/billwidmer.com\/white-hat-seo-case-study\/\" target=\"_blank\" rel=\"noopener\">Bill Widmer\u2019s Seo case study<\/a>.<\/p><h2>11. Sit Down And Get Sh&gt;?! Done. Focus on Important Things<\/h2><p>The best way to get things done is to simply sit down and start working on them. There will be always excuses to do something else or you tell yourself that you\u2019ll start working on it tomorrow. But the truth is that no one will do your work for you. Often it\u2019s boring as hell when I need to spend months to figure out our new pricing model, server costs, profit margins, but it\u2019s my job to get it done as soon as I can.<\/p><p>I would advise you to <strong>focus on things that really matter<\/strong>\u00a0as well as applying the <a href=\"https:\/\/www.entrepreneurs-journey.com\/397\/80-20-rule-pareto-principle\/\" target=\"_blank\" rel=\"noopener\">80\/20 rule<\/a>. Let\u2019s take an example, your website design. In the early days don\u2019t waste any time to change the Buy Now button colors on your pricing page from green to pale green, changing your background image, or spending hours to add 3 more icons to your homepage.<\/p><p>Why? Because NO ONE CARES. We did this several times and nothing changed. And we know why.<\/p><p>When your site gets 100 visitors a day making changes like this won\u2019t have any impact. Why? Your sample is too small, tiny, and insignificant. You\u2019re better spending your time driving 1000 daily visitors to your homepage and make changes when it makes sense. When your sample becomes big enough, then that\u2019s the right time to make some CRO tests.<\/p><h2>12. Your Weakness Can be Your Biggest Opportunity \u2013 Adapt and Move Fast<\/h2><p>As an early stage startup with a small client base and lack of resources, you might think that you\u2019re weak and your competitors will eat you alive.<\/p><p>Wrong. You being small compared to them can be a huge advantage. What do I mean? Your competitors have thousands of customers and a robust infrastructure. But they also have a lot of corporate regulations and steps to follow before making any changes.<\/p><div id=\"attachment_15015\" class=\"wp-caption aligncenter\"><img class=\"wp-image-15015 size-full\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/adapt.jpeg\" alt=\"Adapt and move fast\" width=\"854\" height=\"480\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/adapt.jpeg 854w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/adapt-300x169.jpeg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/adapt-768x432.jpeg 768w\" \/><p class=\"wp-caption-text\">Adapt and move fast (Img src: <a href=\"https:\/\/www.vg247.com\/\" target=\"_blank\" rel=\"noopener\">VG 24\/7<\/a>)<\/p><\/div><p>Usually, it takes them months if not more to implement a feature request, make a UI change or implement an add-on which is crucial for many of their customers. But your hands are not tied by corporate policies or large infrastructure, meaning you can move fast and roll out changes quickly.<\/p><p>When the first production-ready version of PHP 7 was out we made it available to our clients within days! The industry news was all about the benefits of upgrading to PHP 7 and it was already available on our platform. For most of the big guys in our industry, it took 1+ year to make it available for their clients (many still haven\u2019t added it or have made it available only for higher enterprise plans which is nonsense).<\/p><p>I can tell you that by implementing this little change, where you can switch your site from PHP 5.6 to 7 by <a href=\"https:\/\/kinsta.com\/knowledgebase\/wordpress-php-version\/\" target=\"_blank\" rel=\"noopener\">clicking on a single button<\/a>, generated us a lot of new business.<\/p><div id=\"attachment_15016\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15016\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/php-version.gif\" alt=\"Change PHP version\" width=\"837\" height=\"420\" \/><p class=\"wp-caption-text\">Change PHP version<\/p><\/div><p>This small feature drives a lot of business. Web agencies with big portfolios switched from our competitors to us. Why? Because we were able to move fast and made this feature available. Long story short,\u00a0<strong>being small has also benefits, you just need to look at this in a different way<\/strong>.<\/p><h2>13. Network. Make Friends. Help Others. And Hustle.<\/h2><p>No matter which industry you operate in, I bet there are a lot of experts, influencers, and bloggers you should, want\u00a0and need to connect with. You need to be part of this community and make as many friends as you can. You can discuss the latest industry changes, help to solve problems that perhaps one of you has already faced, recommend tools, marketing techniques, services or people to work with and so on.<\/p><p>Unfortunately, we don\u2019t help others as much as we could. We all have personal goals and busy schedules filling out every minute of the day. But building relationships by helping others is very important. Share your knowledge and expertise, give advice on topics you are familiar and comfortable with, share your resources and data if possible, share others content on social media or include them in your newsletter. These little good deeds will come back to you tenfold.<\/p><p><strong>Every man is the architect of his own fortune<\/strong>. Perhaps you\u2019ve planned to send 20 super customized and targeted cold emails this week? Send 30. Send 40 next week. Connect to +1 influencer on social media. Improve your overall productivity just by 1% each week and at the end of the year, it will have a huge impact. Be active on Twitter, and reach out to all relevant blogs in your industry and make sure they know you exist.<\/p><p>Gary Vaynerchuk could have retired many years ago but he\u2019s doing the opposite. He is super active on social media, runs a daily YouTube vlog and podcast, actively gives out advice, tweets you back, leads his media company and other X ventures. And <a href=\"https:\/\/twitter.com\/garyvee\/status\/858878581547642881?lang=en\">hustling on a late Sunday night<\/a> is not a problem for him. So what\u2019s your excuse?<\/p><p>Should I add any comments to his tweet? I don\u2019t think so.<\/p><h2>14. Automate, Automate, And Automate Some More<\/h2><p>When it comes to scaling your business to the next level one of the key points is how you can handle the increased interest in your service and support requests. The good thing is that there are a lot of great tools, apps, and services you can integrate that makes it possible to manage a huge user base with a relatively small team.<\/p><p>At Kinsta we use a lot of <a href=\"https:\/\/kinsta.com\/blog\/saas-products\/\">SaaS tools<\/a> to make our lives easier, and we\u2019ve integrated many third-party services to handle different tasks which help free up the human workload. Machines are perfect when it comes to monotone tasks, because they rarely make a mistake and can complete tasks in a much faster and more cost-effective way. But first you need to solve these tasks and issues manually to get crucial insights about how they should be solved.<\/p><p>Over the years, we\u2019ve automated so many things that it would be hard to mention all of them in this section, but here\u2019s a few of them:<\/p><ul><li>Thanks to Intercom we automated most of our welcome and onboarding messages<\/li>\n<li>We send automatic emails when your payment failed<\/li>\n<li>You get an email when your credit card is about to expire<\/li>\n<li>You\u2019re automatically notified when your annual plan is about to renew<\/li>\n<li>Our system checks the volume of outgoing transactional emails so if something is going on and it notifies our sysadmins<\/li>\n<li>Server load is checked automatically to maintain maximum performance<\/li>\n<li>Automated and simplified the most support heavy tasks, like website staging creation, staging push to live, SSL creation and renewal, querying error logs on the dashboard, etc.<\/li>\n<li>And 42 other tasks just in 2017<\/li>\n<\/ul><p>Users love to interact with real humans, especially if they pay money for your service. So that also means you shouldn\u2019t automate everything. But you should automate all the tasks that can be completed by a code-based or programmable solution. If you don\u2019t, your user base will grow faster than you can hire new engineers and support people to maintain the quality of your service.<\/p><h2>15. Shiny Headlines Don\u2019t Reflect The Reality<\/h2><p>When you follow Techcrunch, VentureBeat or TNW you might think that building a business is easy. Countless headlines stating \u201cX\u201d raised millions, \u201cZ\u201d has been acquired by \u201cY,\u201d and XZ doubled their revenue\/userbase\/customer base in the last 12 months.<\/p><p>We love reading and sharing success stories. But let\u2019s not forget that these stories usually only focus on the results, the great numbers achieved. They almost never mention the years of hard work behind that success. The countless number of working hours, all the profit that has been reinvested, personal sacrifices, sleepless nights, burnouts, health issues and so on.<\/p><p>What I want to say here is that <strong>media is often one-sided<\/strong> and we could get the impression that building a business from scratch is easier than ever. It\u2019s true for starting a company but turning it into a business is where the work starts. You will have ups and downs all the time, but hard work pays off and gets rewarded. This image below perfectly illustrates how the life of an entrepreneur looks like:<\/p><div id=\"attachment_15017\" class=\"wp-caption aligncenter\"><img class=\"size-full wp-image-15017\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/entrepreneurs-life.png\" alt=\"Entrepreneurs life\" width=\"847\" height=\"260\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/entrepreneurs-life.png 847w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/entrepreneurs-life-300x92.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/entrepreneurs-life-768x236.png 768w\" \/><p class=\"wp-caption-text\">Entrepreneurs life (Img src: <a href=\"https:\/\/baremetrics.com\/blog\/long-slow-saas-ramp-of-death\" target=\"_blank\" rel=\"noopener\">Baremetrics<\/a>)<\/p><\/div><p>The road to success isn\u2019t straight. This is another great tweet by a well-known marketer and online entrepreneur Noah Kagan. The point here is not the get rich section, rather the years of hard work needed.<\/p><h2>16. Get a Hobby. Seriously.<\/h2><p>If you spend the majority of your day in front of a computer you can burn out pretty quickly. Just a few long months later you will find yourself unmotivated and less effective. If you\u2019re passionate about your business you know the feeling because you\u2019re thinking about your next move, feature, marketing campaign 24\/7. You\u2019re excited and don\u2019t be surprised when your startup is a frequent performer in your dreams.<\/p><p>But don\u2019t forget that occasionally you need to turn off this type of thinking and spend time with a totally different activity. <strong>Your brain needs downtime<\/strong>.<\/p><p><img class=\"aligncenter size-full wp-image-15129\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/downtime.gif\" alt=\"downtime\" width=\"500\" height=\"313\" \/><\/p><p>It will help your brain recharge, stay sharp, and the best ideas usually come from a different environment (yeah, <a href=\"https:\/\/blog.bufferapp.com\/why-we-have-our-best-ideas-in-the-shower-the-science-of-creativity\" target=\"_blank\" rel=\"noopener\">the shower effect<\/a>).<\/p><p>If you already have a hobby that\u2019s awesome! If you don\u2019t, I\u2019d advise you to find one. It doesn\u2019t matter what kind of hobby is it, just make sure you can practice it on a regular basis, it\u2019s fun and can help you disconnect from the business. Mental breaks increase productivity, attention, and encourage creativity.<\/p><p>For example, I have collected antique books since the age of 16. I know a lot about them, I\u2019m in touch with a lot of collectors and it\u2019s pretty challenging to find a 120-year-old book which has only been printed in 100 copies. But it\u2019s fun and a really good feeling when you can build up your collection with a rare volume.<\/p><h2>Big Thanks<\/h2><p>We\u2019d like to say a <strong>big thank you to all of our customers<\/strong>!\u00a0?\u00a0 You\u2019re all awesome and without your support, Kinsta wouldn\u2019t be here.<\/p><p>And to the creators of WordPress and members of the community; without you, it wouldn\u2019t be possible to create outstanding websites. And to our primary partner, <a href=\"https:\/\/kinsta.com\/blog\/google-cloud-hosting\/\" target=\"_blank\" rel=\"noopener\">Google Cloud Platform<\/a>, for the awesome global platform you have built over the years and keep improving day by day. Every new data center and feature you release immediately gets passed down to our clients.<\/p><p>We also want to give a shout out to our friends at <a href=\"https:\/\/stripe.com\/\" target=\"_blank\" rel=\"noopener\">Stripe<\/a> for collecting our recurring revenue and providing an awesome payment processor, it\u2019s a dream for our devs working with your code!<\/p><p>And thanks to the folks at <a href=\"https:\/\/www.intercom.com\/\" target=\"_blank\" rel=\"noopener\">Intercom<\/a> which has powered our support\/live chat system since day one. You\u2019re the direct link between us and our customers 24\u00d77.<\/p><p>Finally, I\u2019d personally like to say a huge thank you to all of our team members. You\u2019re the best team! Our clients can always count on you and you\u2019re the ones who turn Kinsta into a service that users love.<\/p><h2>What Would We Do Differently?<\/h2><p>I\u2019d only mention one thing. We should have launched our <strong>affiliate program way sooner.<\/strong> Having a referral program could have a huge impact on your growth. Our program was in beta over the past year and was invite-only. We were working closely with top bloggers to gather as much feedback as possible.<\/p><p>The reason we didn\u2019t launch it publicly, in the beginning, is simple, we didn\u2019t have the time to develop it. We had to focus our efforts on building our infrastructure and features.<\/p><p>If you don\u2019t have any special requirements you can use any 3rd party affiliate solutions and set it up quickly. <a href=\"https:\/\/kinsta.com\/blog\/affiliate-system\/\" target=\"_blank\" rel=\"noopener\">We do have a special case<\/a> because besides the one-time affiliate commission\u00a0we also provide a lifetime\u00a0<a href=\"https:\/\/kinsta.com\/affiliates\/\" target=\"_blank\" rel=\"noopener\">recurring commission<\/a> as long as the referred client stays a client. This is something that none of our competitors do! And finally, the program is live.<\/p><h2>Conclusion<\/h2><p>If you\u2019re still with me, I\u2019d like to say a big thank you for reading this post, as it was 7,000+ words.\u00a0?\u00a0Bootstrapping will teach you and force you to operate profitably and innovate constantly. You\u2019ll be faced with countless issues during your journey, but with hard work, consistency, and creativity nothing is impossible.<\/p><p>Have you faced any similar situations to ours? If so, how did you solve them? What has been the hardest thing you have experienced while building your business? I would love to hear your thoughts in the comments below!<\/p>","created_at":"2018-11-07T16:48:05+0000","updated_at":"2018-11-07T16:48:05+0000","published_at":"2017-11-28T07:19:39+0000","published_by":["Tom Zsomborgi"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":37,"domain_name":"kinsta.com","preview_picture":"https:\/\/kinsta.com\/wp-content\/uploads\/2017\/11\/bootstrapping-startup-1.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":429,"title":"Library of Babel","url":"https:\/\/libraryofbabel.info\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:48:05+0000","updated_at":"2018-11-07T16:48:05+0000","annotations":[],"reading_time":0,"domain_name":"libraryofbabel.info","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":425,"title":"The Spriters Resource","url":"https:\/\/www.spriters-resource.com\/","content":"Hello everyone! I'm back for what feels like the first time in a while to bring you this week's update! But first, I'd like to highlight a change made to the site's navigation that should hopefully make things a little less cluttered and more relevant.<p>Over the last week, I realized that on tSR in particular, the navigation menu was getting a bit too long as more and more consoles met the minimum to be shown outside of the Other Systems section. At first, I just increased that minimum but decided that didn't really make sense so instead, I started tracking views on each console and changed the navigation to show the top 25 most popular consoles at all times with everything else shown under Other Systems.<\/p><p>I'm toying with the idea of resetting those view counts monthly to make sure the navigation always stays relevant but I'm not sure how necessary that is. If you have any thoughts on the matter, please feel free to share them in the comments below!<\/p><p>But now on to what you're all actually here for - the new sheets! As always, this week brings us plenty of great new content. In no particular order, we have WarioWare Gold, some new Paper Mario customs, Oracle of Ages and Seasons (two very underrated but very fun portable Zelda games), and unsurprisingly, some initial rips from DELTARUNE. And of course, I'd be remiss if I didn't mention the continued appearance of THE iDOLM@STER (though in a different game this time)! That's only a small taste though so have a look through everything and enjoy your weekend!\n  <\/p>","created_at":"2018-11-07T16:48:02+0000","updated_at":"2018-11-07T16:48:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.spriters-resource.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":426,"title":"PC \/ Computer - Wolfenstein 3D","url":"https:\/\/www.spriters-resource.com\/pc_computer\/wolfenstein3d\/","content":"<p><a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/0.html\">#<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/A.html\">A<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/B.html\">B<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/C.html\">C<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/D.html\">D<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/E.html\">E<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/F.html\">F<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/G.html\">G<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/H.html\">H<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/I.html\">I<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/J.html\">J<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/K.html\">K<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/L.html\">L<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/M.html\">M<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/N.html\">N<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/O.html\">O<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/P.html\">P<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/Q.html\">Q<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/R.html\">R<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/S.html\">S<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/T.html\">T<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/U.html\">U<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/V.html\">V<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/W.html\" class=\"active\">W<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/X.html\">X<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/Y.html\">Y<\/a> - <a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/Z.html\">Z<\/a><\/p><div class=\"section\"><p>Playable Character<\/p><img src=\"https:\/\/www.spriters-resource.com\/images\/light\/buttons\/nav-open.png\" alt=\"\" \/>[3]<\/div><div class=\"section\"><p>Enemies<\/p><img src=\"https:\/\/www.spriters-resource.com\/images\/light\/buttons\/nav-open.png\" alt=\"\" \/>[12]<\/div><div class=\"section\"><p>Miscellaneous<\/p><img src=\"https:\/\/www.spriters-resource.com\/images\/light\/buttons\/nav-open.png\" alt=\"\" \/>[6]<\/div><table class=\"display\"><tr class=\"altrow0\"><td><a href=\"https:\/\/www.spriters-resource.com\/pc_computer\/\">PC \/ Computer<\/a><\/td>\n    <td><a href=\"https:\/\/www.spriters-resource.com\/genre\/shooter\/\">Shooter<\/a><\/td>\n    \n  <\/tr><\/table><p>You must be <a href=\"https:\/\/www.spriters-resource.com\/login\/\">logged in<\/a> with an active forum account to post comments.<\/p>","created_at":"2018-11-07T16:48:02+0000","updated_at":"2018-11-07T16:48:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.spriters-resource.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":427,"title":"20+ SaaS Products We Use to Grow Our Web Hosting Company","url":"https:\/\/kinsta.com\/blog\/saas-products\/","content":"<p>When it comes to running a successful company, sometimes you are only as good as your tools! Of course, you need talented people behind those tools, but choosing the right SaaS products can help boost productivity for your team, generate more leads, and overall, put you on the fast track to success. We thought it would be interesting to share with you all the different SaaS products we use here at Kinsta to <a href=\"https:\/\/kinsta.com\/blog\/bootstrapping-startup\/\" target=\"_blank\" rel=\"noopener\">grow our managed WordPress hosting company<\/a>.<\/p><h2>SaaS Products We Use<\/h2><p>First off, what exactly do we mean by SaaS? <a href=\"https:\/\/en.wikipedia.org\/wiki\/Software_as_a_service\" target=\"_blank\" rel=\"noopener noreferrer\">SaaS<\/a> is short for <strong>software as a service<\/strong>.\u00a0This is a software distribution method which typically involves paying on a subscription model for a product, tool, or service. In most cases, SaaS products\u00a0are now hosted in the cloud and are accessible from anywhere via your web browser or even mobile device.<\/p><p>If you are a brand new startup, we don\u2019t recommend simply going out and subscribing to all these tools below. A lot of companies have had great success bootstrapping their companies from the ground up and adding additional\u00a0tools as they grow their revenue. This is exactly what we did here at Kinsta. Tools and services like these can add up fast, so do your research, be smart, and determine which ones you might need first to help you accomplish your priority goals faster.<\/p><div class=\"cta-mini\"><h4 class=\"cta-mini--title\">Still looking for that perfect WordPress host?<\/h4><p>At Kinsta, we do hosting different than the rest. We\u2019ve combined the fastest platform in the world (Google Cloud) with the best support team in the industry. Ready to take your site the next level?<\/p><p><a class=\"cta-mini--link\" href=\"https:\/\/kinsta.com\/plans\/\">Check out our plans<\/a><\/p><\/div><h3>1. Stripe<\/h3><p>The very first thing most companies have to decide is what payment processor they are going to use to accept payments from their customers. There a lot out there to choose from such as PayPal, Stripe, Flint, Authorize.net, Braintree and Adyen.<\/p><div id=\"attachment_6431\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/stripe.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6431 size-full\" title=\"Stripe\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/stripe-payment-processor.jpg\" alt=\"stripe payment processor\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/stripe-payment-processor.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/stripe-payment-processor-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/stripe-payment-processor-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/stripe-payment-processor-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Stripe<\/p><\/div><p>Here at Kinsta we decided to go with <a href=\"https:\/\/stripe.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Stripe<\/a> because it is very developer friendly, has low credit card processing transaction\u00a0fees, and was easy to integrate into our back-end dashboard. For those of you who have been following us for a while know that our <a href=\"https:\/\/kinsta.com\/blog\/how-we-got-our-stripe-account-back\/\" target=\"_blank\" rel=\"noopener noreferrer\">Stripe\u00a0relationship had a rocky start<\/a>\u00a0back in 2014. But all companies make\u00a0mistakes at some point, and Stripe did everything in their power to resolve the\u00a0situation we encountered. Fast forward two years and we can say we still love Stripe!\u00a0We really think it is the\u00a0best thing that has happened to online payments since PayPal, and are happy to call them a business partner. The takeaway is that <strong>customer support is more important than ever<\/strong>!<\/p><p>Recommended for further reading:\u00a0<a href=\"https:\/\/kinsta.com\/blog\/credit-card-fraud-stripe\/\" target=\"_blank\" rel=\"noopener\">How to Prevent and Reduce Credit Card Fraud by 98% Using Stripe Radar<\/a><\/p><h3>2. Baremetrics<\/h3><p><a href=\"https:\/\/baremetrics.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Baremetrics<\/a> is just like their brand name says, it is all about the metrics! They offer analytics and insights for payment processors such as Braintree, Stripe, and Recurly.\u00a0When it comes to choosing some SaaS products for your business, this is a must-have.<\/p><div id=\"attachment_6432\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/baremetrics.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6432 size-full\" title=\"Baremetrics\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics.jpg\" alt=\"baremetrics\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Baremetrics<\/p><\/div><p>Why do you need\u00a0Baremetrics? Well, basically they simplify the process of understanding your business\u2019s revenue income stream and help make data-driven decisions on where you could improve. You can instantly see juicy bits of information such as\u00a0monthly recurring revenue, average revenue per user, refunds, upgrades, lifetime customer values, user churn, and probably the two most important ones, your net revenue and revenue growth rate. You could calculate all this information yourself in a spreadsheet, but then immediately you should be weighing the cost of your time for doing this.<\/p><p>Check out their <a href=\"https:\/\/demo.baremetrics.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">dashboard demo<\/a> (seen below).\u00a0We love using Baremetrics\u00a0at Kinsta, along with the Stripe integration, as this allows us to instantly see how our company is performing.<\/p><div id=\"attachment_6453\" class=\"wp-caption alignnone\"><img class=\"wp-image-6453 size-full\" title=\"Baremetrics dashboard analytics\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-dashboard.jpg\" alt=\"baremetrics dashboard\" width=\"2061\" height=\"1542\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-dashboard.jpg 2061w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-dashboard-300x224.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-dashboard-768x575.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-dashboard-1024x766.jpg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/baremetrics-dashboard-409x306.jpg 409w\" \/><p class=\"wp-caption-text\">Baremetrics dashboard analytics<\/p><\/div><h3>3. Intercom<\/h3><p class=\"t__h3 hero__h3\"><a href=\"https:\/\/www.intercom.io\/\" target=\"_blank\" rel=\"noopener noreferrer\">Intercom<\/a> is a customer platform with a suite of products for live chat, marketing, feedback, and support. Remember, <a href=\"https:\/\/kinsta.com\/knowledgebase\/scope-of-support\/\" target=\"_blank\" rel=\"noopener\">support is key<\/a> to any successful company. So if you want our advice, make sure you have a good support system in place, whether this is chat, tickets, etc.<\/p><div id=\"attachment_6433\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.intercom.io\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6433 size-full\" title=\"Intercom\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom.jpg\" alt=\"intercom\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Intercom<\/p><\/div><p>We have used Intercom since day one here at Kinsta and it has worked very well for us. We pride ourselves on providing support from team members who are actually\u00a0professional developers and work with WordPress every day, contribute to WordPress Core and other open source projects. As a result, the expertise of our support department is second to none. Nothing is worse than launching a chat session and feeling like you know more than the person on the other end. This will never happen at Kinsta!<\/p><p>Intercom has provided an easy way for us to integrate chat and support tickets (seen below) it into our dashboard so customers know we are always one click away, 24\u00d77.\u00a0Our average response time in Q1 2018 was 2 minutes! This is one of our favorite SaaS products.<\/p><div id=\"attachment_6454\" class=\"wp-caption aligncenter\"><img class=\"wp-image-6454 size-full\" title=\"Intercom chat\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-chat.jpg\" alt=\"intercom chat\" width=\"1199\" height=\"932\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-chat.jpg 1199w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-chat-300x233.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-chat-768x597.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-chat-1024x796.jpg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/intercom-chat-394x306.jpg 394w\" \/><p class=\"wp-caption-text\">Intercom chat<\/p><\/div><h3>4. Slack<\/h3><p>You have probably heard of <a href=\"https:\/\/slack.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Slack<\/a> by now! If not, you should immediately go give it a try. Slack is a messaging\u00a0app for teams which is designed to make communicating with your colleagues fast and easy, no matter what platform they are on.<\/p><div id=\"attachment_6434\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/slack.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6434 size-full\" title=\"Slack\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack.jpg\" alt=\"slack\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Slack<\/p><\/div><p>Every member on the team at Kinsta uses Slack on a daily, or probably hourly basis. It helps us do everything from chat about industry news, collaborate\u00a0on team projects, and of course, share an <a href=\"https:\/\/kinsta.com\/blog\/wordpress-gifs\/\" target=\"_blank\" rel=\"noopener noreferrer\">occasional Giphy<\/a>.\u00a0We actually use the paid version but the free version works great as well if the few limitations do not bother you. Slack is great for any company just starting out, especially for remote teams. Kinsta\u2019s team is spread across multiple continents\u00a0and so we depend on Slack for communication across time-zones and platforms. When it comes to SaaS products, this is a must-have.<\/p><div id=\"attachment_6480\" class=\"wp-caption aligncenter\"><img class=\"wp-image-6480 size-full\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-sass-products-must-have.jpg\" alt=\"slack sass products must have\" width=\"1220\" height=\"820\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-sass-products-must-have.jpg 1220w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-sass-products-must-have-300x202.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-sass-products-must-have-768x516.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-sass-products-must-have-1024x688.jpg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/slack-sass-products-must-have-455x306.jpg 455w\" \/><p class=\"wp-caption-text\">src:\u00a0<a href=\"https:\/\/brandfolder.com\/slack\" target=\"_blank\" rel=\"noopener noreferrer\">Slack Branding<\/a><\/p><\/div><h3>5. Mention<\/h3><p>Keeping up with monitoring your brand on social media is a never-ending process and can sometimes be overwhelming and time-consuming. <a href=\"https:\/\/mention.com\/en\/\" target=\"_blank\" rel=\"noopener noreferrer\">Mention<\/a> has the ability to monitor billions of sources in over 40 languages so you won\u2019t miss anything being said about\u00a0your brand.<\/p><div id=\"attachment_6435\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/mention.com\/en\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6435 size-full\" title=\"Mention\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention.jpg\" alt=\"mention\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Mention<\/p><\/div><p>We use Mention to keep track of social mentions of Kinsta around the web. One great thing too is that they have an <a href=\"https:\/\/mention.com\/blog\/mention-slack-integration\/\" target=\"_blank\" rel=\"noopener noreferrer\">integration with Slack<\/a>\u00a0(seen below). This allows us to easily see Tweets and mentions from right within a private slack Channel, along with an influencer score. This way we never have to waste time logging into a separate dashboard or get yet another email.<\/p><div id=\"attachment_6456\" class=\"wp-caption aligncenter\"><img class=\"wp-image-6456 size-full\" title=\"Mention Slack integration\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration.jpg\" alt=\"mention slack integration\" width=\"1601\" height=\"600\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration.jpg 1601w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration-300x112.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration-768x288.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration-1024x384.jpg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration-460x172.jpg 460w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mention-slack-integration-1600x600.jpg 1600w\" \/><p class=\"wp-caption-text\">Mention Slack integration<\/p><\/div><h3>6. Close.io<\/h3><p>Close.io\u2019s slogan on their homepage pretty much says it all, \u201cClose More Details, Make More Sales.\u201d The software was created by sales people for salespeople. Nobody likes manual entry and so Close.io tries to automate as much as possible, from calls, emails, and activity so you can spend more time closing deals instead of entering data.<\/p><div id=\"attachment_6436\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/close.io\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6436 size-full\" title=\"Close.io\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/close.io_.jpg\" alt=\"close.io\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/close.io_.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/close.io_-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/close.io_-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/close.io_-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Close.io<\/p><\/div><p>Our sales team\u00a0at Kinsta uses <a href=\"http:\/\/close.io\/\" target=\"_blank\" rel=\"noopener noreferrer\">Close.io<\/a>\u00a0for lead management, along with the powerful <a href=\"https:\/\/zapier.com\/zapbook\/closeio\/\" target=\"_blank\" rel=\"noopener noreferrer\">Zapier integrations<\/a> to automate it even more. And of course, some of these automation\u2019s include Slack integrations. You can probably see a trend starting here. Take advantage of Slack channels to simply your workflows.<\/p><h3>7. New Relic<\/h3><p><a href=\"https:\/\/newrelic.com\" target=\"_blank\" rel=\"noopener noreferrer\">New Relic <\/a>offers all sorts of services, from\u00a0application performance monitoring, server monitoring, mobile monitoring, and real-time user insights.<\/p><div id=\"attachment_6437\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/newrelic.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6437 size-full\" title=\"New Relic\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic.jpg\" alt=\"new relic\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">New Relic<\/p><\/div><p>We use two of New Relic\u2019s products here at Kinsta, and that is <a href=\"https:\/\/newrelic.com\/application-monitoring\" target=\"_blank\" rel=\"noopener noreferrer\">application monitoring<\/a>\u00a0and <a href=\"https:\/\/newrelic.com\/synthetics\" target=\"_blank\" rel=\"noopener noreferrer\">synthetics<\/a>. We use New Relic on\u00a0all WordPress sites hosted with Kinsta to monitor uptime and also performance (btw. <a href=\"https:\/\/kinsta.com\/knowledgebase\/custom-new-relic-tracking\/\" target=\"_blank\" rel=\"noopener noreferrer\">with your own license key you can set up monitoring as well<\/a>). If your website starts to slow down, we can easily see this on our end and then take action to see what might be causing the issue. New Relic has been a great tool for us to ensure we provide the best support for our WordPress customers.<\/p><div id=\"attachment_6465\" class=\"wp-caption aligncenter\"><img class=\"wp-image-6465 size-full\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic.png\" alt=\"New Relic APM\" width=\"1018\" height=\"429\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic.png 1018w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic-300x126.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic-768x324.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/new-relic-460x194.png 460w\" \/><p class=\"wp-caption-text\">New Relic \u2013 APM tool for SaaS products<\/p><\/div><p>Check out our in-depth post on <a href=\"https:\/\/kinsta.com\/blog\/wordpress-performance-new-relic\/\" target=\"_blank\" rel=\"noopener noreferrer\">how we use New Relic<\/a> at Kinsta to troubleshoot performance issues on client\u2019s sites.<\/p><h3>8. CloudFlare<\/h3><p><a href=\"https:\/\/www.cloudflare.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">CloudFlare<\/a> is a free global CDN (or rather fully proxy service), DNS, <a href=\"https:\/\/kinsta.com\/blog\/ddos-attack\/\" target=\"_blank\" rel=\"noopener\">DDoS protection<\/a> &amp; web security provider that can speed up and protect\u00a0your website.<\/p><div id=\"attachment_6438\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.cloudflare.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6438 size-full\" title=\"Cloudflare\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/cloudflare.jpg\" alt=\"cloudflare\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/cloudflare.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/cloudflare-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/cloudflare-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/cloudflare-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Cloudflare<\/p><\/div><p>Kinsta uses CloudFlare for customers who are looking for a free and or premium solution to speed up their websites around the globe, while also offering DDoS protection. Another advantage our customers get is that they can <a href=\"https:\/\/kinsta.com\/knowledgebase\/cloudflare-bandwidth-usage\/\" target=\"_blank\" rel=\"noopener noreferrer\">save a lot on bandwidth<\/a>. Typically a CDN will\u00a0handle the delivery of 60-70% of a website\u2019s traffic load. So by offloading assets to a CDN, this can ensure you don\u2019t go over your bandwidth limit with Kinsta. Make sure to check out our <a href=\"https:\/\/kinsta.com\/knowledgebase\/cloudflare\/\" target=\"_blank\" rel=\"noopener noreferrer\">CloudFlare integration guide<\/a>, as well as our <a href=\"https:\/\/kinsta.com\/knowledgebase\/how-do-i-add-keycdn-to-my-site\/\" target=\"_blank\" rel=\"noopener noreferrer\">KeyCDN integration guide<\/a>.<\/p><h3>9. Crazy Egg<\/h3><p><a href=\"https:\/\/www.crazyegg.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Crazy Egg<\/a> is a heat map and website data analysis tool which helps you dive into how people are interacting with your website.<\/p><div id=\"attachment_6439\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.crazyegg.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6439 size-full\" title=\"Crazy Egg\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/crazy-egg.jpg\" alt=\"crazy egg\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/crazy-egg.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/crazy-egg-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/crazy-egg-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/crazy-egg-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Crazy Egg<\/p><\/div><p>One of the main features we use at Kinsta from Crazy Egg is the awesome heatmap feature! You can just setup an experiment, let it run, and come back to see exactly where people are clicking on your website. See how many people scroll down the page, where they stop, and even what sources of traffic are clicking on what. Crazy Egg is great for <a href=\"https:\/\/kinsta.com\/blog\/conversion-rate-optimization-tips\/\" target=\"_blank\" rel=\"noopener\">improving your CRO<\/a> and testing out new landing pages.<\/p><h3>10. Github<\/h3><p>As many of you probably know, <a href=\"https:\/\/github.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Github<\/a> is a free open source repository that is used by people, companies, and developers to help\u00a0build software. GitHub allows teams to contribute and fork to over 38 million projects, including <a href=\"https:\/\/github.com\/WordPress\/WordPress\" target=\"_blank\" rel=\"noopener noreferrer\">WordPress<\/a> itself.<\/p><div id=\"attachment_6441\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/github.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6441 size-full\" title=\"GitHub\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/github.jpg\" alt=\"github\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/github.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/github-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/github-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/github-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">GitHub<\/p><\/div><p>We don\u2019t do anything that is publicly available to people on GitHub, but the team at Kinsta uses GitHub to manage our internal development for our website and other projects. If you work in a team with any type of code, then GitHub is awesome because it offers a version control system which easily allows multiple developers to contribute and commit\u00a0updates.<\/p><h3>11. MailChimp<\/h3><p><a href=\"http:\/\/mailchimp.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">MailChimp<\/a> is used by over 12 million businesses and individuals to send emails to and newsletters. As they say, \u201cThe money is in the list.\u201d <a href=\"https:\/\/kinsta.com\/blog\/how-to-build-an-email-list\/\" target=\"_blank\" rel=\"noopener\">Building and maintaining a high-quality email list<\/a> can be a great way for a company to generate more sales and keep their current customers informed. MailChimp makes this process silly easy and has been perfecting their software for many years.<\/p><div id=\"attachment_6442\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/mailchimp.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6442 size-full\" title=\"MailChimp\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mailchimp.jpg\" alt=\"mailchimp\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mailchimp.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mailchimp-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mailchimp-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/mailchimp-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">MailChimp<\/p><\/div><p>We use MailChimp here at Kinsta and we are on a paid plan, but they have a free plan which allows you up to 2,000 subscribers. We are creating a lot of content that we know you will love and so MailChimp is a great way to allow us to share it with you each week. For startups, MailChimp\u2019s free plan get be a great solution to get up and going while you start to build your email list.<\/p><h3>12. G Suite (Email + Collaboration Tools)<\/h3><p>In 2018, nobody likes getting emails\u00a0and it seems that everyone is constantly striving and failing to reach that inbox zero goal. But every company needs a way to communicate through email. When it comes to email hosting there are a lot to choose from, such\u00a0as Zoho, Office 365, Rack Space, and <a href=\"https:\/\/gsuite.google.com\/\" target=\"_blank\" rel=\"noopener\">G Suite<\/a>.<\/p><div id=\"attachment_18941\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/gsuite.google.com\/\" target=\"_blank\" rel=\"noopener\"><img class=\"wp-image-18941 size-full\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/g-suite-saas-tools.png\" alt=\"G Suite\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/g-suite-saas-tools.png 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/g-suite-saas-tools-300x89.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/g-suite-saas-tools-768x227.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/g-suite-saas-tools-610x180.png 610w\" \/><\/a><p class=\"wp-caption-text\">G Suite<\/p><\/div><p>Kinsta\u2019s managed WordPress hosting is powered by the <a href=\"https:\/\/kinsta.com\/blog\/google-cloud-hosting\/\" target=\"_blank\" rel=\"noopener\">Google Cloud Platform<\/a>, so obviously we\u2019re big fans of their infrastructure and apps. We use G Suite for email hosting, which allows every employee to have their own @kinsta.com email address and access it via Gmail or their favorite email client. According to <a href=\"https:\/\/emailclientmarketshare.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Email Market Share<\/a>, as of February 2018, Gmail accounts for 26% of email clients usage based on the number of opens.\u00a0When it comes to Gmail, you can\u2019t really go wrong and it is priced reasonably at $5 per user per month. This also gives employees 30 GB of online storage for their Google drive and access to Google Docs, which we use a lot here at Kinsta. On more tips how and why we use it please check our detailed guide: <a href=\"https:\/\/kinsta.com\/blog\/g-suite\/\" target=\"_blank\" rel=\"noopener\">The Benefits of G Suite for Your Business<\/a>.<\/p><h3>13. Accuranker<\/h3><p>If you do any kind of inbound marketing or <a href=\"https:\/\/kinsta.com\/blog\/wordpress-seo-checklist\/\" target=\"_blank\" rel=\"noopener\">SEO on your website<\/a>, you should really be keeping track of your keyword rankings. SEO is still alive and well in 2018 and keywords are still important. Even though things have shifted focus more to \u201ctopic\u201d writing, keywords will always exist as Google\u2019s algorithm is after all,\u00a0a computer algorithm.<\/p><div id=\"attachment_6444\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.accuranker.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6444 size-full\" title=\"Accuranker\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/accuranker.jpg\" alt=\"accuranker\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/accuranker.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/accuranker-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/accuranker-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/accuranker-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Accuranker<\/p><\/div><p>We use <a href=\"https:\/\/www.accuranker.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">AccuRanker<\/a> at Kinsta to keep track of our important keywords that we are focusing on ranking. This helps us know whether or not we need to give an article a little more love. There are a lot of on-page SEO tricks you can do to rank an article higher in SERPs. We share these with you in <a href=\"https:\/\/kinsta.com\/blog\/wordpress-seo-checklist\/\" target=\"_blank\" rel=\"noopener\">this post<\/a>. AccuRanker\u00a0allows us to track the keywords around a post, add a note of what we optimized, and then later return to see how our rankings were affected by the changes. After you learn what works, it is basically a game of rinse and repeat.<\/p><h3>14. Dropbox<\/h3><p>500 million people around the globe, along with over 200,000 businesses, use <a href=\"https:\/\/www.dropbox.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Dropbox<\/a> on a regular basis to access and share documents and files on the go.<\/p><div id=\"attachment_6445\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.dropbox.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6445 size-full\" title=\"Dropbox\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/dropbox.jpg\" alt=\"dropbox\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/dropbox.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/dropbox-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/dropbox-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/dropbox-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Dropbox<\/p><\/div><p>We use Dropbox at Kinsta to share large files and enable them to be more easily accessible by remote team members from anywhere, no matter what platform they are working on.<\/p><h3>15. Wistia<\/h3><p><a href=\"https:\/\/wistia.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Wistia<\/a> is a high-end\u00a0video hosting and analytics platform. We like to think of Wistia as the Roll\u2019s Royce of video hosting, and are used by companies such as Moz, MailChimp, and Hubspot.<\/p><div class=\"in-post-container\"><div class=\"vertical newsletter-promo flip-container\"><div class=\"flipper\"><div class=\"front\"><div><div class=\"default\"><div><p>We grew our traffic 1,187% with WordPress.We\u2019ll show you how.<\/p><p>Join 20,000+ others who get our weekly newsletter with insider WordPress tips!<\/p><\/div>Subscribe Now<\/div><\/div><\/div><div class=\"back\"><div><p>Subscribe to the Kinsta newsletter<\/p><\/div><\/div><\/div><\/div><\/div><div id=\"attachment_6446\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/wistia.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6446 size-full\" title=\"Wistia\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/wistia.jpg\" alt=\"wistia saas products\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/wistia.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/wistia-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/wistia-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/wistia-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Wistia<\/p><\/div><p>We use Wistia at Kinsta to showcase our customer testimonial videos.\u00a0While there are other similar platforms we could use such as YouTube and Vimeo, we chose Wistia because of the seamless integration they provide with our website. If you want an elegant looking HTML5 video player integration, than Wistia is definitely one you should check out. And if you are OK with their branding on your player, they even offer up to 5 videos hosted for free. This can be a great alternative to YouTube for companies just starting out.<\/p><h3>16. Amazon\u00a0Route 53<\/h3><p><a href=\"https:\/\/aws.amazon.com\/route53\/\" target=\"_blank\" rel=\"noopener noreferrer\">Amazon Route 53<\/a> is a highly available and scalable cloud DNS service. It is designed to give developers and businesses an extremely reliable and cost-effective way to route end users to Internet applications.<\/p><div id=\"attachment_6447\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/aws.amazon.com\/route53\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6447 size-full\" title=\"Amazon Route 53\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/amazon-route-53-saas-products.jpg\" alt=\"amazon route 53 saas products\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/amazon-route-53-saas-products.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/amazon-route-53-saas-products-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/amazon-route-53-saas-products-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/amazon-route-53-saas-products-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Amazon Route 53<\/p><\/div><p>Here at Kinsta, we don\u2019t settle for mediocre when it comes to web performance and reliability. DNS is just one area that we take very seriously. We include Amazon Route 53 <a href=\"https:\/\/kinsta.com\/blog\/premium-dns\/\" target=\"_blank\" rel=\"noopener\">premium DNS<\/a> services for every customer that hosts their WordPress website with us. Amazon has an enormous infrastructure and a long history of providing additional DNS features such as latency-based routing, Geo DNS, traffic flow, and most important, DNS failover. Kinsta customers can even\u00a0<a href=\"https:\/\/kinsta.com\/knowledgebase\/dns\/\" target=\"_blank\" rel=\"noopener noreferrer\">modify their DNS records<\/a> right from their MyKinsta dashboard.<\/p><h3>17. SendGrid<\/h3><p><a href=\"https:\/\/sendgrid.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">SendGrid<\/a> is another email marketing platform and service which focuses on transactional emails.\u00a0Some of their clients include Uber, Spotify, Airbnb, and Yelp.<\/p><div id=\"attachment_6448\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/sendgrid.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6448 size-full\" title=\"SendGrid\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/sendgrid.jpg\" alt=\"sendgrid\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/sendgrid.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/sendgrid-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/sendgrid-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/sendgrid-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">SendGrid<\/p><\/div><p>While MailChimp works great for our inbound and <a href=\"https:\/\/kinsta.com\/learn\/content-marketing\/\" target=\"_blank\" rel=\"noopener\">content marketing<\/a>, we use SendGrid at Kinsta for the transactional side of things. This includes emails between servers and applications. Like many other companies, we previously used Mandrill (owned by MailChimp) until it was recently shut down. For transactional based emails, SendGrid is a much more cost-effective solution. If you send a lot of emails from your software or applications, SendGrid does a good job.<\/p><p>However, we don\u2019t necessarily recommend them for typical marketers. There are little things such as the UTM parameters having to be manually updated for each campaign which can be a little frustrating. Their WYSIWYG builder has improved tremendously over the past couple years though and they are working hard to provide similar functionality as services such as MailChimp. But remember, their focus is on transactional emails, not marketing.<\/p><h3>18.\u00a0Iconfinder<\/h3><p>Every good website needs good visuals, and that is where <a href=\"https:\/\/www.iconfinder.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Iconfinder<\/a> thrives. They currently have a database of over 1.1 million icons, consisting of over 24,000 sets.<\/p><div id=\"attachment_6449\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.iconfinder.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6449 size-full\" title=\"Iconfinder\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder.jpg\" alt=\"iconfinder\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Iconfinder<\/p><\/div><p>We use Iconfinder for icons\u00a0on both our website and the Kinsta blog. Most of the featured images you see here on the blog and in our knowledge base articles are sourced from Iconfinder. One of the best features of Iconfinder is that they provide you with all 4 formats, SVG, PNG, AI, and CSH. We typically prefer working with <a href=\"https:\/\/kinsta.com\/blog\/wordpress-svg\/\" target=\"_blank\" rel=\"noopener\">SVG<\/a> because they are scalable to any resolution without losing any image quality.<\/p><div id=\"attachment_6481\" class=\"wp-caption aligncenter\"><img class=\"wp-image-6481 size-full\" title=\"Iconfinder search\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-wordpress-results.jpg\" alt=\"iconfinder wordpress results\" width=\"2050\" height=\"1477\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-wordpress-results.jpg 2050w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-wordpress-results-300x216.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-wordpress-results-768x553.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-wordpress-results-1024x738.jpg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/iconfinder-wordpress-results-425x306.jpg 425w\" \/><p class=\"wp-caption-text\">Iconfinder search<\/p><\/div><h3>19.\u00a0ClearBooks<\/h3><p>Unfortunately, the not so fun part of any business is having to run the books, payroll, and accounting. But it is one of those things that has to be done. <a href=\"https:\/\/www.clearbooks.co.uk\/\" target=\"_blank\" rel=\"noopener noreferrer\">ClearBooks<\/a> is an online accounting software that tries to simplify the process both for startups, small businesses, and even accountants.<\/p><div id=\"attachment_6450\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.clearbooks.co.uk\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6450 size-full\" title=\"ClearBooks\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/clear-books.jpg\" alt=\"clear books\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/clear-books.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/clear-books-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/clear-books-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/clear-books-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">ClearBooks<\/p><\/div><p>We use ClearBooks at Kinsta partially because\u00a0they are a UK company, but also because they have a lot of experience with simplifying\u00a0the online accounting process. None of us here at Kinsta are accountants after all. Having launched in 2008, they are not new to the financial market and are\u00a0accredited by ICB, ICAEW and ACCA. If you are in the UK, ClearBooks might be one you want to check out.<\/p><h3>20. SEMrush<\/h3><p><a href=\"https:\/\/www.semrush.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">SEMrush<\/a> is an all-in-one competitive data and analytics tool for digital marketing professionals. Keyword research for one is very important if you are doing inbound marketing for your company, as well as getting a feel for where your competitors stand.<\/p><div id=\"attachment_6451\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.semrush.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6451 size-full\" title=\"SEMrush\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush.jpg\" alt=\"semrush\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">SEMrush<\/p><\/div><p>We use SEMrush at Kinsta to research new blog post ideas and strategically make data-driven decisions for our inbound marketing. This tool allows us to easily see things such as search volume for a keyword or topic as well as how competitive that keyword is. As some of you may know, Google\u00a0recently changed their Keyword Planner Tool and now <a href=\"https:\/\/woorkup.com\/close-variant-keywords\/\" target=\"_blank\" rel=\"noopener noreferrer\">close variant keywords<\/a> show up as having the same search volume. SEMrush collects their search volume from a number of different sources, as well as their own database, and so this allows them to still provide you with accurate search volume. When it comes to keyword research, <strong>never assume people are searching. Write natural, but write smart.<\/strong><\/p><div id=\"attachment_6482\" class=\"wp-caption aligncenter\"><img class=\"wp-image-6482 size-full\" title=\"SEMRush keyword research\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-keyword-research.jpg\" alt=\"semrush keyword research\" width=\"1634\" height=\"783\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-keyword-research.jpg 1634w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-keyword-research-300x144.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-keyword-research-768x368.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-keyword-research-1024x491.jpg 1024w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/semrush-keyword-research-460x220.jpg 460w\" \/><p class=\"wp-caption-text\">SEMRush keyword research<\/p><\/div><h3>21. Trello<\/h3><p><a href=\"https:\/\/trello.com\/\" target=\"_blank\" rel=\"noopener noreferrer\">Trello<\/a> is an online collaborative tool which allows you to easily create boards, lists, and cards to organize your projects. Trello is used by businesses all over the world, some of which include Adobe, Pixar, Google, PayPal, and National Geographic.<\/p><div id=\"attachment_6452\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/trello.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-6452 size-full\" title=\"Trello\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/trello.jpg\" alt=\"trello\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/trello.jpg 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/trello-300x89.jpg 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/trello-768x227.jpg 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/trello-460x136.jpg 460w\" \/><\/a><p class=\"wp-caption-text\">Trello<\/p><\/div><p>We love Trello here at Kinsta! Trello allows us to easily create tasks and keep track of who is working on what. Even for our blog and inbound marketing calendar, we have a Trello board to keep track of new topics, ideas, and for setting deadlines. Sharing a board is as simple as one click. We use the free plan, which makes this another must have tool for startups! Btw. Asana is another great project management tool worth to take a look.\u00a0To compare the features make sure you check out this detailed <a href=\"https:\/\/www.beewits.com\/trello-vs-asana\/\" target=\"_blank\" rel=\"noopener noreferrer\">Trello vs. Asana guide<\/a>.<\/p><h3>22.\u00a0OneSignal<\/h3><p>Push notifications are something that have become very popular recently. The major benefits are that they provide another avenue of communication for your WordPress site. Lots of people skim over emails nowadays, or simply don\u2019t even read them. Push notifications can help bring customers and visitors back to your site by placing your message right in front of them.\u00a0\u00a0Some people, such as Jon over on the Fat Stacks Entrepreneur blog, are <a href=\"http:\/\/fatstacksblog.com\/pushcrew-review\/\" target=\"_blank\" rel=\"noopener noreferrer\">getting between 18%-25% CTR<\/a> on their push notifications, which is very high!<\/p><div id=\"attachment_10264\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/onesignal.com\/\" target=\"_blank\" rel=\"noopener noreferrer\"><img class=\"wp-image-10264 size-full\" title=\"OneSignal\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/onesignal.png\" alt=\"onesignal\" width=\"845\" height=\"250\" srcset=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/onesignal.png 845w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/onesignal-300x89.png 300w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/onesignal-768x227.png 768w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/onesignal-610x180.png 610w, https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/onesignal-460x136.png 460w\" \/><\/a><p class=\"wp-caption-text\">OneSignal<\/p><\/div><p>We use OneSignal here at Kinsta for our blog. Whenever we publish a new post,\u00a0we simply check the \u201cPush desktop notification\u201d box and then subscribers to our push notifications instantly get notified! Make sure to also check out our official <a href=\"https:\/\/onesignal.com\/case-studies\/kinsta\" target=\"_blank\" rel=\"noopener noreferrer\">case study with OneSignal<\/a>. In less than 3 months of using OneSignal, it is already our 3rd highest organic traffic source.<\/p><div class=\"wp-caption aligncenter\"><img title=\"OneSignal dashboard\" src=\"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/12\/onesignal-push-notification-dashboard-1.jpg\" alt=\"OneSignal dashboard\" width=\"2073\" height=\"1768\" \/><p class=\"wp-caption-text\">OneSignal dashboard<\/p><\/div><p>To get started using push notifications, check out our post on <a href=\"https:\/\/kinsta.com\/blog\/wordpress-push-notifications\/\" target=\"_blank\" rel=\"noopener noreferrer\">How to Setup Free WordPress Push Notifications<\/a>.<\/p><h2>Summary<\/h2><p>As you can see, we utilize quite a few different SaaS products here at Kinsta,\u00a0in all sorts of different areas. Each one of them though serves a specific purpose and they help us to be more productive and focus on what is really important, which of course is providing you with the best possible WordPress hosting experience possible!<\/p><p>Do you have any awesome SaaS products you currently use that you think we might like? If so, feel free to share them below.<\/p>","created_at":"2018-11-07T16:48:02+0000","updated_at":"2018-11-07T16:48:02+0000","published_at":"2016-09-07T06:22:44+0000","published_by":["Brian Jackson"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":17,"domain_name":"kinsta.com","preview_picture":"https:\/\/kinsta.com\/wp-content\/uploads\/2016\/09\/saas-products.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":424,"title":"The Technology Behind Procedural FPS","url":"https:\/\/80.lv\/articles\/gibhard-indie-shooter-hardcore\/","content":"<section class=\"index-section\"><div class=\"wrap\">\n        <div class=\"container_margin_fix\">\n            <div class=\"container_for_grid\">\n                <div id=\"container_top_column\" class=\"container-item width_x_3 for_mw-m container-item-normal-b column_element stick clear_b\">\n                    <figure id=\"mainBanner\" class=\"sh through_b\"><\/figure><\/div>\n\n                 \n<div class=\"container_right_column_x2 for_mw-m\">\n    <div class=\"container_right_column\">\n        <div class=\"container-item container-item-normal-b\">\n            <figure id=\"pos_top\" class=\"sh\"><\/figure><\/div>\n    <\/div>\n\n    <div class=\"container_right_column container_left_for_xl_column\">\n        <div class=\"container-item-for-title-block container-item-h\">Vendors<\/div><div class=\"container-item with_bg_block green2\">\n                              <div class=\"container-item-inn\">\n                                <div class=\"container-item-h\"><a href=\"https:\/\/80.lv\/vendors\/unity-technologies-3\/\">Unity Technologies<\/a><\/div>\n                                <img src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2014\/11\/unity-master-black.jpg\" class=\"attachment-medium size-medium wp-post-image\" alt=\"\" \/><div class=\"line_sep\">\n                                <div class=\"about-block-info\">\n                                  <div class=\"about-block-info-row\">\n                                    <div class=\"about-block-info-elem\">\n                                      <div class=\"icon-div-container\">\n                                        <div class=\"icon-div followers\">\n                                      <\/div>\n                                      <div class=\"icon-text\">33 followers<\/div>\n                                    <\/div>\n                                    <div class=\"about-block-info-elem about-block-info-elem-without-margin-right\">\n                                      <div class=\"icon-div-container\">\n                                        <div class=\"icon-div posts\">\n                                      <\/div>\n                                      <div class=\"icon-text\"><a href=\"https:\/\/80.lv\/vendors\/unity-technologies-3\/#button_select_block_elem_2\">468 posts<\/a><\/div>\n                                    <\/div>\n                                  <\/div>\n                                  <div class=\"about-block-info-row\">\n                                    <div class=\"about-block-info-elem\">\n                                      <div class=\"icon-div-container\">\n                                        \n                                      <\/div>\n                                      <div class=\"icon-text\"><a href=\"https:\/\/80.lv\/vendors\/unity-technologies-3\/#button_select_block_elem_1\">0 questions answered<\/a><\/div>\n                                    <\/div>\n                                  <\/div>\n                                <\/div>\n                                <div class=\"tags\"><a class=\"tag\">Marketing<\/a><a class=\"tag\">Static<\/a><a class=\"tag\">Tools<\/a><\/div>\n                              <\/div>\n                            <\/div>        \n                \n    <\/div>\n\n    <div class=\"container_right_column\">\n        <div class=\"subscribe-block-wrapper container-item\">\n    <div class=\"container-item-inn\">\n        <div class=\"container-item-h\">Weekly Newsletter<\/div>\n        <input type=\"email\" name=\"subscribe-email\" placeholder=\"email address\" \/><div class=\"error-msg subscribe-error\">\n        Thank you for subscription. You'll be getting weekly digest every Friday.\n        <div class=\"subscribe-block-btn green-submit\">\n            Subscribe\n        <\/div>\n    <\/div>\n<\/div><div class=\"container-item v-right orange for_mw-m\">\n    <div class=\"container-item-inn\">\n        <div class=\"container-item-h\">Events<\/div>\n        <div class=\"calendar-events-list bitext sm\">\n            <div class=\"calendar-events-list-sub\">\n                Subscribe:\u00a0\n                <a href=\"denied:webcal:\/\/80.lv\/events.ics\" class=\"light\">iCal<\/a>\n                \u00a0|\u00a0\n                <a class=\"light\" target=\"_blank\" href=\"http:\/\/www.google.com\/calendar\/hosted\/80.lv\/render?cid=https:\/\/80.lv\/events.ics&amp;pli=1\">Google Calendar<\/a>\n            <\/div>\n                                <div class=\"calendar-events-list-elem\">\n                        <div class=\"calendar-events-list-elem-header\"><a class=\"title_elem\" href=\"https:\/\/80.lv\/events\/casual-connect-asia-2018\">Casual Connect Asia 2018<\/a><\/div>\n                        <div class=\"calendar-events-list-elem-body\">\n                            <div class=\"table\">\n                                <div class=\"tr\">\n                                                                        <div class=\"td\">\n                                                                                    <div class=\"flag_block flag-icon flag-icon-background flag-icon-cn\">\n                                                                            <\/div>\n                                    <div class=\"td\">\n                                                                                    Shenzhen CN \u00a0\n                                                                                12, Nov \u2014 15, Nov                                    <\/div>\n                                <\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                                        <div class=\"calendar-events-list-elem\">\n                        <div class=\"calendar-events-list-elem-header\"><a class=\"title_elem\" href=\"https:\/\/80.lv\/events\/g-star-2018\">G-STAR 2018<\/a><\/div>\n                        <div class=\"calendar-events-list-elem-body\">\n                            <div class=\"table\">\n                                <div class=\"tr\">\n                                                                        <div class=\"td\">\n                                                                                    <div class=\"flag_block flag-icon flag-icon-background flag-icon-kr\">\n                                                                            <\/div>\n                                    <div class=\"td\">\n                                                                                    Busan KR \u00a0\n                                                                                15, Nov \u2014 19, Nov                                    <\/div>\n                                <\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                                        <div class=\"calendar-events-list-elem\">\n                        <div class=\"calendar-events-list-elem-header\"><a class=\"title_elem\" href=\"https:\/\/80.lv\/events\/devgamm-minsk-2018\">DevGAMM Minsk 2018<\/a><\/div>\n                        <div class=\"calendar-events-list-elem-body\">\n                            <div class=\"table\">\n                                <div class=\"tr\">\n                                                                        <div class=\"td\">\n                                                                                    <div class=\"flag_block flag-icon flag-icon-background flag-icon-by\">\n                                                                            <\/div>\n                                    <div class=\"td\">\n                                                                                    Minsk BY \u00a0\n                                                                                15, Nov \u2014 17, Nov                                    <\/div>\n                                <\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                                        <div class=\"calendar-events-list-elem\">\n                        <div class=\"calendar-events-list-elem-header\"><a class=\"title_elem\" href=\"https:\/\/80.lv\/events\/gamenode-hackathon\">GameNode Hackathon<\/a><\/div>\n                        <div class=\"calendar-events-list-elem-body\">\n                            <div class=\"table\">\n                                <div class=\"tr\">\n                                                                        <div class=\"td\">\n                                                                                    <div class=\"flag_block flag-icon flag-icon-background flag-icon-by\">\n                                                                            <\/div>\n                                    <div class=\"td\">\n                                                                                    Minsk BY \u00a0\n                                                                                16, Nov \u2014 19, Nov                                    <\/div>\n                                <\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                                        <div class=\"calendar-events-list-elem\">\n                        <div class=\"calendar-events-list-elem-header\"><a class=\"title_elem\" href=\"https:\/\/80.lv\/events\/pax-unplugged-2018\">PAX Unplugged 2018<\/a><\/div>\n                        <div class=\"calendar-events-list-elem-body\">\n                            <div class=\"table\">\n                                <div class=\"tr\">\n                                                                        <div class=\"td\">\n                                                                                    <div class=\"flag_block flag-icon flag-icon-background flag-icon-us\">\n                                                                            <\/div>\n                                    <div class=\"td\">\n                                                                                    Philadelphia US \u00a0\n                                                                                30, Nov \u2014 3, Dec                                    <\/div>\n                                <\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                            <\/div>\n        <div class=\"form_block\">\n            <div class=\"form_block_row form_block_row_last\"><a class=\"btn btn-orange full_width\" href=\"https:\/\/80.lv\/events\/\">\n                    <div>More events<\/div><\/a><\/div>\n        <\/div>\n    <\/div>\n<\/div>        <div class=\"container-item container-item-normal-b\">\n            <figure id=\"pos_middle\" class=\"sh\"><\/figure><\/div>\n        <div class=\"container-item v-right green2 for_mw-m\">\n    <div class=\"container-item-inn\">\n                <div class=\"container-item-h\">Latest comments<\/div>\n        <div class=\"last-comment-list\">\n                            <div class=\"last-comment-list-elem bitext sm\">\n                    <div class=\"last-comment-list-elem-header\">\n                        \n                    <\/div>\n                    <div class=\"last-comment-list-elem-body\">\n                        <p>\"..A whole interior out of the existing assets.\" So it's a compositing tool? But I love that bit! It's UVing that needs AI - Give robots the jobs we don't want.<\/p>\n                    <\/div>\n                <\/div>\n                                <div class=\"last-comment-list-elem bitext sm\">\n                    <div class=\"last-comment-list-elem-header\">\n                        \n                    <\/div>\n                    <div class=\"last-comment-list-elem-body\">\n                        <p>Amazing tutorial and documentation! This piece is truly inspiring :D<\/p>\n                    <\/div>\n                <\/div>\n                                <div class=\"last-comment-list-elem bitext sm\">\n                    <div class=\"last-comment-list-elem-header\">\n                        \n                    <\/div>\n                    <div class=\"last-comment-list-elem-body\">\n                        <p>Love seeing people just play with stuff to push it into new territory!<\/p>\n                    <\/div>\n                <\/div>\n                        <\/div>\n    <\/div>\n<\/div>\n        <div class=\"container-item container-item-normal-b\">\n            <figure id=\"pos_bottom\" class=\"sh\"><\/figure><\/div>\n    <\/div>\n\n    \n    <div class=\"fix_right_block_marker\">\n    <div class=\"fix_right_block container_right_column\">\n        <div class=\"fix_right_block_inn\">\n                    <\/div>\n    <\/div><\/div>\n\n\n                <div class=\"container_left_column container_left_column_max_x3 block_for_inn_content\">\n                    <div class=\"container-item v2 green2 width_full_with_r_column change-size block_for_fixed_soc_icons single_block\">\n                        <div class=\"container-item-inn\">\n                                                        <div class=\"container-item-h container-item-h-v2\">Gibhard: The Technology Behind Procedural FPS<\/div>\n                            <div class=\"container-item-h-c\">\n                                <div class=\"container-item-h-c-elem\">\n                                    <div class=\"container-item-data\">29 November, 2016<\/div>\n                                <\/div>\n                                <div class=\"container-item-h-c-elem\"><div class=\"tags tags-v3\"><div class=\"tag\">Interview<\/div><\/div><\/div>                            <\/div>\n                            <div class=\"share_content_help\" title=\"https:\/\/80.lv\/articles\/gibhard-indie-shooter-hardcore\/\">\n                            <div class=\"container-item-h-share-block-real share-block__top target_for_header_social\">\n                                \n<div class=\"social share-block\">\n    <div class=\"top-part\">\n        <div class=\"container-item-h-share-block-real-all\">\n            0shares\n        <\/div>\n        <div class=\"container-item-h-share-block-real-mez\">\/<\/div>\n    <\/div>\n    <div class=\"bottom-part\">\n        Share on\n        <div class=\"social-btn fb\">\n                    \n               <\/div><div class=\"social-btn tw\">\n                    \n                <\/div><div class=\"social-btn lk\">\n                    \n                <\/div><div class=\"social-btn rd\">\n                    \n                <\/div>        <a data-flip-widget=\"flipit\" href=\"https:\/\/flipboard.com\">Add to Flipboard Magazine.<\/a>\n    <\/div>\n<\/div>\n\n\n\n                            <\/div>\n\n                            <div class=\"for_fixed_soc_icons_out\">\n                                <div class=\"for_fixed_soc_icons\">\n                                    \n<div class=\"social share-block\">\n    <div class=\"top-part\">\n        <div class=\"container-item-h-share-block-real-all\">\n            0shares\n        <\/div>\n        <div class=\"container-item-h-share-block-real-mez\">\/<\/div>\n    <\/div>\n    <div class=\"bottom-part\">\n        Share on\n        <div class=\"social-btn fb\">\n                    \n               <\/div><div class=\"social-btn tw\">\n                    \n                <\/div><div class=\"social-btn lk\">\n                    \n                <\/div><div class=\"social-btn rd\">\n                    \n                <\/div>        <a data-flip-widget=\"flipit\" href=\"https:\/\/flipboard.com\">Add to Flipboard Magazine.<\/a>\n    <\/div>\n<\/div>\n\n\n\n                                <\/div>\n                            <\/div>\n\n                            <div class=\"old-cont post\">\n                                <div class=\"container-item-grey bitext container-item-text-v3\">\n                                    <div class=\"container-item-grey bitext container-item-text-v3\">\n                                        <h5><a href=\"http:\/\/www.brameulaers.net\/\" target=\"_blank\">Bram Eulaers<\/a>\u00a0is an experienced game developer and an environment artist, who worked on a number of big-budget games, but in his free time he\u2019s creating an awesome procedural shooter\u00a0<a href=\"http:\/\/www.gibhard.com\/\" target=\"_blank\">Gibhard<\/a>. We\u2019ve talked about the technology that fuels this project and discussed the beauty of retro FPS.<\/h5>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_01.jpg\"><img class=\"aligncenter size-large wp-image-47825\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_01-1024x641.jpg\" alt=\"80level_01\" width=\"1024\" height=\"641\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_01-1024x641.jpg 1024w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_01-300x188.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_01-768x481.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_01.jpg 1378w\" \/><\/a><\/p>\n<h3>Introduction<\/h3>\n<p>Hi! My name is <a href=\"http:\/\/www.brameulaers.net\/\" target=\"_blank\">Bram Eulaers<\/a>, I grew up in Belgium, and currently live in sunny Southern California. I\u2019ve worked in the game industry for over 10 years, and have worked on various games such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bionic_Commando_(2009_video_game)\" target=\"_blank\">Bionic Commando<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Terminator_Salvation_(video_game)\" target=\"_blank\">Terminator Salvation<\/a> and most recently on <a href=\"https:\/\/playoverwatch.com\/en-us\/\" target=\"_blank\">Overwatch<\/a> at <a href=\"http:\/\/eu.blizzard.com\/en-gb\/\" target=\"_blank\">Blizzard Entertainment<\/a>, where I currently work as a senior environment artist.<\/p>\n<div class=\"double_image_table double_image_table_2_3\">\n<div class=\"double_image_table_tr double_image_table_tr_2_3\">\n<div class=\"double_image_table_td double_image_table_td_40 double_image_table_td_left bitext\">\n<div class=\"wp-video\">\n<a href=\"http:\/\/i.imgur.com\/VJRq4fA.webm\">http:\/\/i.imgur.com\/VJRq4fA.webm<\/a><\/div>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_02.jpg\"><img class=\"alignnone size-full wp-image-47824\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_02.jpg\" alt=\"80level_02\" width=\"1378\" height=\"863\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_02.jpg 1378w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_02-300x188.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_02-768x481.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_02-1024x641.jpg 1024w\" \/><\/a><\/p>\n<\/div>\n<div class=\"double_image_table_td double_image_table_td_60 double_image_table_td_right bitext\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_03.jpg\"><img class=\"alignnone size-full wp-image-47823\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_03.jpg\" alt=\"80level_03\" width=\"1398\" height=\"872\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03.jpg 1398w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03-768x479.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03-1024x639.jpg 1024w\" \/><\/a><\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh.jpg\"><img class=\"alignnone size-full wp-image-47811\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh.jpg\" alt=\"ctwi0bsuaaawqxh\" width=\"1136\" height=\"700\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh.jpg 1136w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-300x185.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-768x473.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-1024x631.jpg 1024w\" \/><\/a><\/p>\n<\/div>\n<\/div>\n<\/div>\n<h3>Gibhard<\/h3>\n<p><a href=\"http:\/\/www.gibhard.com\/\" target=\"_blank\">Gibhard<\/a> actually started out as a 2D prototype for procedurally generating good level layouts. Just for fun I threw in some crude FPS controls with simple enemies, and it basically grew from there.<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_03.jpg\"><img class=\"aligncenter size-large wp-image-47823\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_03-1024x639.jpg\" alt=\"80level_03\" width=\"1024\" height=\"639\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03-1024x639.jpg 1024w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03-768x479.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_03.jpg 1398w\" \/><\/a><\/p>\n<p>Big influences for me are of course some of the old-school FPS games I grew up with, like Doom, Duke Nukem 3D and Quake. A more recent game that\u2019s been a big influence is Spelunky, which has been a big inspiration for the overall structure of the game. Visually, the artwork by The Bitmap Brothers on games like \u201cZ\u201d and The Chaos Engine has been very inspirational for me.<\/p>\n<div class=\"photo_slick\"><div class=\"photo_slider_main\"><ul id=\"gallery-1\" class=\"slider-for gallery galleryid-46606 gallery-columns-4 gallery-size-full\"><li class=\"gallery-item\"><a class=\"fancybox\" rel=\"gallery3367\" href=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS.jpg\"><img width=\"700\" height=\"437\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS.jpg\" class=\"attachment-gallery_image size-gallery_image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS-1024x640.jpg 1024w\" \/><\/a>\n\t\t\t\t<div class=\"wp-caption-text caption_empty gallery-caption\">\n\t\t\t\t<\/div><\/li><li class=\"gallery-item\"><a class=\"fancybox\" rel=\"gallery3367\" href=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj.jpg\"><img width=\"700\" height=\"437\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj.jpg\" class=\"attachment-gallery_image size-gallery_image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj-1024x640.jpg 1024w\" \/><\/a>\n\t\t\t\t<div class=\"wp-caption-text caption_empty gallery-caption\">\n\t\t\t\t<\/div><\/li><li class=\"gallery-item\"><a class=\"fancybox\" rel=\"gallery3367\" href=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3.jpg\"><img width=\"700\" height=\"437\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3.jpg\" class=\"attachment-gallery_image size-gallery_image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3-1024x640.jpg 1024w\" \/><\/a>\n\t\t\t\t<div class=\"wp-caption-text caption_empty gallery-caption\">\n\t\t\t\t<\/div><\/li><li class=\"gallery-item\"><a class=\"fancybox\" rel=\"gallery3367\" href=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG.jpg\"><img width=\"700\" height=\"437\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG.jpg\" class=\"attachment-gallery_image size-gallery_image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG-1024x640.jpg 1024w\" \/><\/a>\n\t\t\t\t<div class=\"wp-caption-text caption_empty gallery-caption\">\n\t\t\t\t<\/div><\/li><\/ul><\/div><div class=\"photo_slider_pager\"><ul class=\"slider-nav\"><li class=\"item\"><img width=\"144\" height=\"90\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS.jpg\" class=\"attachment-gallery_thumb size-gallery_thumb\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUQAAH0BS-1024x640.jpg 1024w\" \/><\/li><li class=\"item\"><img width=\"144\" height=\"90\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj.jpg\" class=\"attachment-gallery_thumb size-gallery_thumb\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9FUUAAneEj-1024x640.jpg 1024w\" \/><\/li><li class=\"item\"><img width=\"144\" height=\"90\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3.jpg\" class=\"attachment-gallery_thumb size-gallery_thumb\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GUUAEfJf3-1024x640.jpg 1024w\" \/><\/li><li class=\"item\"><img width=\"144\" height=\"90\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG.jpg\" class=\"attachment-gallery_thumb size-gallery_thumb\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG.jpg 1199w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG-768x480.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CyLji9GVQAAgubG-1024x640.jpg 1024w\" \/><\/li><\/ul><\/div>\n        <\/div>\n<h3>Art Style<\/h3>\n<p>Well, first of all, GIBHARD is my attempt to make a game by myself in my free time, so I had to figure out ways to control the scope and generate artwork fast. Going for a retro aesthetic helps in cutting corners while simultaneously allowing me to make something visually pleasing.<\/p>\n<div class=\"two_column_text\">\n<div class=\"two_column_text_elem two_column_text_elem_left\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot12.jpg\"><img class=\"aligncenter wp-image-46611 size-full\" title=\"gibhard, gamedev, game development, fps, indiedev, unity, hardcore, game, games, steam, early access\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot12.jpg\" alt=\"webshot12\" width=\"703\" height=\"461\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot12.jpg 703w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot12-300x197.jpg 300w\" \/><\/a><\/p>\n<\/div>\n<div class=\"two_column_text_elem two_column_text_elem_right\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot13.jpg\"><img class=\"aligncenter wp-image-46610 size-full\" title=\"gibhard, gamedev, game development, fps, indiedev, unity, hardcore, game, games, steam, early access\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot13.jpg\" alt=\"webshot13\" width=\"703\" height=\"460\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot13.jpg 703w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot13-300x196.jpg 300w\" \/><\/a><\/p>\n<\/div>\n<\/div>\n<p>Basically it allows me to work on the big picture without having to waste time on the tiny details current gen games require. Going with camera facing 2D sprites for a lot of the content has been a huge time saver for me, it allows me to create a lot of content in little time. Sprites don\u2019t need to be modeled, unwrapped, textured, rigged, etc! Especially for creating the enemies this approach helped me a lot. I actually took inspiration from how iD Software made characters on the original doom, by basically creating these awesome clay and latex models, and photographing them from different angles. It took this exact same approach for my characters, but did it all digitally with modern tools, sculpting them in ZBrush and \u201cphotographing\u201d them in 3D software!<\/p>\n<p>Techniques like this, of course, drives the style into a very retro direction and getting content out fast like this allows me to focus on what I consider a very big pillar of the game\u2019s visual style: Great visual readability through color and lighting!<\/p>\n<h3>Mixing Old School Style With Modern Tricks<\/h3>\n<p>What I\u2019m doing in GIBHARD is not treating the old school elements as old school. The game can very much be considered current gen as I\u2019m making full use of realtime lighting and shaders. For example almost all sprites have normal maps attached to them, which actually light them as if they were a 3D object and makes them feel natural in the environment! There was of course a lot of tweaking involved to get the balance just right but I think I got a good grasp on it by now. In the end pixels are pixels, whether it\u2019s next-gen or retro. If you author the content correctly, you can light it correctly and create whatever effects you want. It\u2019s been a lot of fun trying out modern techniques like normal maps and volumetric noise shaders, and applying them on a retro aesthetic!<\/p>\n<div class=\"two_column_text\">\n<div class=\"two_column_text_elem two_column_text_elem_left\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/CtWIzKkUkAAoExe.jpg\"><img class=\"aligncenter size-large wp-image-47812\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/CtWIzKkUkAAoExe-1024x631.jpg\" alt=\"ctwizkkukaaoexe\" width=\"1024\" height=\"631\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWIzKkUkAAoExe-1024x631.jpg 1024w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWIzKkUkAAoExe-300x185.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWIzKkUkAAoExe-768x473.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWIzKkUkAAoExe.jpg 1136w\" \/><\/a><\/p>\n<\/div>\n<div class=\"two_column_text_elem two_column_text_elem_right\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh.jpg\"><img class=\"aligncenter size-large wp-image-47811\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-1024x631.jpg\" alt=\"ctwi0bsuaaawqxh\" width=\"1024\" height=\"631\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-1024x631.jpg 1024w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-300x185.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh-768x473.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/CtWI0bsUAAAWQXh.jpg 1136w\" \/><\/a><\/p>\n<\/div>\n<\/div>\n<h3>Lighting<\/h3>\n<p>For the overall ambient lighting in, I actually generate a skylight texture, which gets blurred and is essentially mapped on the whole map using XZ world coordinates for the UV. Using custom shaders and this texture I\u2019m able to create soft lighting that has nice transitions between indoor and outdoor areas. I actually use this same technique to create glowing surfaces like for example lava, it\u2019s essentially a skylight in reverse. The skylight texture is basically a simplified lightmap, that gets generated in real time when the map level is created.<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot10.jpg\"><img class=\"aligncenter wp-image-46609 size-large\" title=\"gibhard, gamedev, game development, fps, indiedev, unity, hardcore, game, games, steam, early access\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot10-1024x671.jpg\" alt=\"webshot10\" width=\"1024\" height=\"671\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot10-1024x671.jpg 1024w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot10-300x197.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot10-768x503.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot10.jpg 1401w\" \/><\/a><\/p>\n<p>While this largely handles the ambient lighting, I still make use of point lights a lot to make things emit light. I\u2019m running GIBHARD with a deferred renderer which helps a lot in getting good performance with lots of point lights.<\/p>\n<div class=\"double_image_table double_image_table_3_2\">\n<div class=\"double_image_table_tr double_image_table_tr_3_2\">\n<div class=\"double_image_table_td  double_image_table_td_60 double_image_table_td_left bitext\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U.jpg\"><img class=\"alignnone size-full wp-image-47817\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U.jpg\" alt=\"cs3paurviaarb4u\" width=\"1200\" height=\"747\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U.jpg 1200w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-768x478.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-1024x637.jpg 1024w\" \/><\/a><\/p>\n<\/div>\n<div class=\"double_image_table_td double_image_table_td_40 double_image_table_td_right bitext\">\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_ambientlighting_02.jpg\"><img class=\"alignnone size-full wp-image-47822\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_ambientlighting_02.jpg\" alt=\"80level_ambientlighting_02\" width=\"1372\" height=\"1678\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_ambientlighting_02.jpg 1372w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_ambientlighting_02-245x300.jpg 245w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_ambientlighting_02-768x939.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_ambientlighting_02-837x1024.jpg 837w\" \/><\/a><\/p>\n<\/div>\n<\/div>\n<\/div>\n<p>Lastly, I\u2019m also getting a lot of use out of emissive materials and glowy sprites. There\u2019s a lot of complex looking lighting effects you can create for an almost free render cost using emissive textures!<\/p>\n<h3>Level Generation Algorithm<\/h3>\n<p>There\u2019s a lot of elements to the level generation, but I\u2019ll try to keep it simple. The maps in GIBHARD are essentially 2D layouts, which are tile based. Layouts basically get generated on this grid in a sort of tree like structure, with algorithms \u201cgrowing\u201d rooms and hallways. So there are really no pre made rooms, however because it\u2019s tile based, most of the level geometry does consist out of pre made modular wall, floor and ceiling tiles.<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot11.jpg\"><img class=\"aligncenter wp-image-46608 size-full\" title=\"gibhard, gamedev, game development, fps, indiedev, unity, hardcore, game, games, steam, early access\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot11.jpg\" alt=\"webshot11\" width=\"703\" height=\"460\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot11.jpg 703w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot11-300x196.jpg 300w\" \/><\/a><\/p>\n<p>I would not say it\u2019s very similar to No Man\u2019s Sky, which uses true randomness and noise to create geometry, what I\u2019m attempting to do with GIBHARD is create random levels that feel properly designed. It\u2019s more of a procedural designer than a random generator. Each level is still very designed, as I have tight control over a whole lot of parameters that decide what goes into a level, from what items get placed to shapes and sizes of rooms. It\u2019s been a very interesting challenge balancing design and randomness, and translating my own sense for level design into code!<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_modular_geometry.jpg\"><img class=\"aligncenter size-full wp-image-47819\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_modular_geometry.jpg\" alt=\"80level_modular_geometry\" width=\"894\" height=\"876\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_modular_geometry.jpg 894w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_modular_geometry-300x294.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_modular_geometry-768x753.jpg 768w\" \/><\/a><\/p>\n<blockquote class=\"twitter-video\" data-lang=\"en\">\n<p dir=\"ltr\" lang=\"en\">Get him boys <a href=\"https:\/\/twitter.com\/hashtag\/GIBHARD?src=hash\">#GIBHARD<\/a> <a href=\"https:\/\/twitter.com\/hashtag\/gamedev?src=hash\">#gamedev<\/a> <a href=\"https:\/\/t.co\/KL5DemyWVj\">pic.twitter.com\/KL5DemyWVj<\/a><\/p>\n<p>\u2014 Bram Eulaers (@Leukbaars) <a href=\"https:\/\/twitter.com\/Leukbaars\/status\/778868784115232768\">September 22, 2016<\/a><\/p><\/blockquote>\n<p>\n<\/p><h3>Speed<\/h3>\n<p>The easy answer of course is, keep it simple :). Keeping everything very low poly and low detail helps a whole lot. Next to that I like to optimise as I go\u2026 I think there\u2019s a really enjoyable challenge in creating more with less. For example, I do a lot of effects on just a single quad with shaders using scrolling or alpha erosion. It\u2019s amazing what kind of complex looking effects you can create with smart shader setups. Another thing I like to do is spread calculations out over several frames. The enemy movement with collision detection for example is only calculated at about 10 fps, however the enemy graphic sprites are interpolated and move at the full frame rate, so it all looks smooth. This plus the fact that they are rendered as simple sprites, means I can have a huge amount of enemies on screen for little performance cost!<\/p>\n<div class=\"wp-video\"><a href=\"http:\/\/i.imgur.com\/lBVcBq6.webm\">http:\/\/i.imgur.com\/lBVcBq6.webm<\/a><\/div>\n<h3>Weapons<\/h3>\n<p>I actually started out modeling them in 3D, but it didn\u2019t feel right with the enemies being sprites. So I basically baked down my 3D models on 2D textures. These older versions you can still see in some older screenshots of the game. For the final version of the weapons, I did so much paintovers and adjustments that there\u2019s little left of the original bakes in them. I found out I got the best results by straight up just painting them as much as possible in photoshop, and actually painting normal maps by hand by picking normal direction colors from a sphere, a technique I use on a lot of other assets as well. This makes sure they always light nicely when you are running through the environment.<\/p>\n<p>Animation wise it\u2019s all done through code, interpolating their positions and using sine waves to create weapon bops. Some guns have glow effects on them as well that are simply done using emissive textures and code driven intensities. Muzzle flash animations are simply done using an alpha erosion style shader effect!<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U.jpg\"><img class=\"aligncenter size-large wp-image-47817\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-1024x637.jpg\" alt=\"cs3paurviaarb4u\" width=\"1024\" height=\"637\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-1024x637.jpg 1024w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-300x187.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U-768x478.jpg 768w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Cs3PAurVIAArB4U.jpg 1200w\" \/><\/a><\/p>\n<h3>Hardcore Difficulty<\/h3>\n<p>I think hard difficulty is fine as long as it\u2019s fun. It should be fun to die and you should always feel like it\u2019s your own fault you died, or at least it should be very clear to you what happened! It helps a lot too that you also never get to play the same level twice. You can\u2019t memorize exactly where the enemies will be appearing and where all the secrets are located, so even after dying over and over again, it will stay fresh. The game will by default be pretty hard but there are different play styles that can make it easier or harder for you. Trying to speedrun for the shortest completion time will be very hard, whereas a very cautious player who prefers to take cover and shoot from behind corners will have it a lot easier.<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_character_bake.jpg\"><img class=\"aligncenter size-full wp-image-47821\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/80level_character_bake.jpg\" alt=\"80level_character_bake\" width=\"702\" height=\"705\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_character_bake.jpg 702w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_character_bake-150x150.jpg 150w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/80level_character_bake-300x300.jpg 300w\" \/><\/a><\/p>\n<p>My take on difficulty in games? I think a lot of games are a bit easy and often not very challenging, however I really like recent games like Devil Daggers or the Dark Souls series that are very hard but also very enjoyable. Really love this wave of games that are hard, but fair!<\/p>\n<p><a href=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot08.jpg\"><img class=\"aligncenter wp-image-46607 size-full\" title=\"gibhard, gamedev, game development, fps, indiedev, unity, hardcore, game, games, steam, early access\" src=\"https:\/\/80lv-cdn.akamaized.net\/80.lv\/uploads\/2016\/11\/webshot08.jpg\" alt=\"webshot08\" width=\"703\" height=\"462\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot08.jpg 703w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/webshot08-300x197.jpg 300w\" \/><\/a><\/p>\n<p>I\u2019m hoping to put the game up on Steam for early access before the end of this year! I actually have most of the content in the game already, and I\u2019m currently basically ironing out bugs and polishing up corners. There\u2019s still a lot of balance work to do as well, which will be an ongoing process during early access. I\u2019m very much looking forward to having people try it out and give me feedback!<\/p>\n<h3><a href=\"http:\/\/www.brameulaers.net\/\" target=\"_blank\">Bram Eulaers<\/a>, Senior Environment Artist<\/h3>\n<h5 class=\"m_-451044497439043302m_6037022412980483350gmail-p1\">Interview conducted by <a href=\"https:\/\/ua.linkedin.com\/in\/kikatok\" target=\"_blank\">Kirill Tokarev<\/a>.\u00a0<\/h5>\n<h2>Follow 80.lv\u00a0on <a href=\"https:\/\/www.facebook.com\/LevelEighty\/\" target=\"_blank\">Facebook<\/a>, <a href=\"https:\/\/twitter.com\/EightyLevel\" target=\"_blank\">Twitter<\/a>\u00a0and <a href=\"https:\/\/www.instagram.com\/80.lv\/\" target=\"_blank\">Instagram<\/a><\/h2>\n                                    <\/div>\n                                <\/div>\n                            <\/div>\n\n                            \n                            <div class=\"container-item-h-share-block-real\">\n                                \n<div class=\"social share-block\">\n    <div class=\"top-part\">\n        <div class=\"container-item-h-share-block-real-all\">\n            0shares\n        <\/div>\n        <div class=\"container-item-h-share-block-real-mez\">\/<\/div>\n    <\/div>\n    <div class=\"bottom-part\">\n        Share on\n        <div class=\"social-btn fb\">\n                    \n               <\/div><div class=\"social-btn tw\">\n                    \n                <\/div><div class=\"social-btn lk\">\n                    \n                <\/div><div class=\"social-btn rd\">\n                    \n                <\/div>        <a data-flip-widget=\"flipit\" href=\"https:\/\/flipboard.com\">Add to Flipboard Magazine.<\/a>\n    <\/div>\n<\/div>\n\n\n\n                            <\/div>\n\n                            <div class=\"follow-block\">\n    <div class=\"follow-title\">Follow 80.lv on<\/div>\n    <a href=\"https:\/\/www.facebook.com\/LevelEighty\/\" target=\"_blank\">\n        <img src=\"https:\/\/cdn.80.lv\/80.lv\/img\/social\/icon-follow-fb.svg\" alt=\"image\" \/><\/a>\n    <a href=\"https:\/\/telegram.me\/level80\" target=\"_blank\">\n        <img src=\"https:\/\/cdn.80.lv\/80.lv\/img\/social\/icon-follow-telegram.svg\" alt=\"image\" \/><\/a>\n    <a href=\"https:\/\/api.kik.com\/v1\/code\/215b2347351eb43350e0f622b8365dc33b13ed02?c=8\" target=\"_blank\">\n        <img src=\"https:\/\/cdn.80.lv\/80.lv\/img\/social\/icon-follow-kik.svg\" alt=\"image\" \/><\/a>\n    <a href=\"https:\/\/join.skype.com\/bot\/7f97d37e-4cdd-4f57-aad6-6228dfd6a630\" target=\"_blank\">\n        <img src=\"https:\/\/cdn.80.lv\/80.lv\/img\/social\/icon-follow-skype.svg\" alt=\"image\" \/><\/a>\n    <a href=\"https:\/\/news360.com\/web\/sources\/806957255\" target=\"_blank\">\n        <img src=\"https:\/\/cdn.80.lv\/80.lv\/img\/social\/icon-follow-news360.svg\" alt=\"image\" \/><\/a>\n<\/div>\n                        <\/div>\n                    <\/div>\n                <\/div>\n                <div class=\"container_left_column container_left_column_max_x3 container_left_column_min_x3\">\n        <div class=\"container-item-for-title-block container-item-for-title-block-v2 container-item-h\">Related <\/div>\n        <div class=\"this-iso\">\n            <div class=\"container-item container-item-half-width this-iso-elem\">\n                <div class=\"container-item-inn\">\n                    <div class=\"relaited_articles\">\n                        <div class=\"relaited_articles_tr\">\n                            <div class=\"relaited_articles_td relaited_articles_img\"><a href=\"https:\/\/80.lv\/articles\/10-rules-of-building-great-games-and-levels\/\"><img width=\"150\" height=\"150\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/05\/Hourences-level-design-secrets-80lv_featured-150x150.jpg\" class=\"attachment-thumbnail size-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/05\/Hourences-level-design-secrets-80lv_featured-150x150.jpg 150w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/05\/Hourences-level-design-secrets-80lv_featured-300x300.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/05\/Hourences-level-design-secrets-80lv_featured.jpg 500w\" \/><\/a><\/div>\n                            <div class=\"relaited_articles_td relaited_articles_text\">\n                                <div class=\"container-item-h\"><a href=\"https:\/\/80.lv\/articles\/10-rules-of-building-great-games-and-levels\/\">10 Rules of Building Great Games &amp; Levels from Hourences<\/a><\/div>\n                                <div class=\"tags\"><div class=\"tag\">Opinion<\/div><div class=\"tag\">Research<\/div><\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                <\/div>\n            <\/div><div class=\"container-item container-item-half-width this-iso-elem\">\n                <div class=\"container-item-inn\">\n                    <div class=\"relaited_articles\">\n                        <div class=\"relaited_articles_tr\">\n                            <div class=\"relaited_articles_td relaited_articles_img\"><a href=\"https:\/\/80.lv\/articles\/escape-from-tarkov-takes-unity-to-the-next-level\/\"><img width=\"150\" height=\"150\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/11\/escape-from-takkov-80lv-Copy-150x150.jpg\" class=\"attachment-thumbnail size-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/11\/escape-from-takkov-80lv-Copy-150x150.jpg 150w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/11\/escape-from-takkov-80lv-Copy-300x300.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/11\/escape-from-takkov-80lv-Copy.jpg 660w\" \/><\/a><\/div>\n                            <div class=\"relaited_articles_td relaited_articles_text\">\n                                <div class=\"container-item-h\"><a href=\"https:\/\/80.lv\/articles\/escape-from-tarkov-takes-unity-to-the-next-level\/\">\u201cEscape from Tarkov\u201d Takes Unity To the Next Level<\/a><\/div>\n                                <div class=\"tags\"><div class=\"tag\">News<\/div><\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                <\/div>\n            <\/div><div class=\"container-item container-item-half-width this-iso-elem\">\n                <div class=\"container-item-inn\">\n                    <div class=\"relaited_articles\">\n                        <div class=\"relaited_articles_tr\">\n                            <div class=\"relaited_articles_td relaited_articles_img\"><a href=\"https:\/\/80.lv\/articles\/10-tips-on-getting-a-job-on-a-aaa-game\/\"><img width=\"150\" height=\"150\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/05\/leandro-amaral-leandro-amaral-0035-layer-2-copy-36-150x150.jpg\" class=\"attachment-thumbnail size-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/05\/leandro-amaral-leandro-amaral-0035-layer-2-copy-36-150x150.jpg 150w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/05\/leandro-amaral-leandro-amaral-0035-layer-2-copy-36-300x300.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/05\/leandro-amaral-leandro-amaral-0035-layer-2-copy-36.jpg 1024w\" \/><\/a><\/div>\n                            <div class=\"relaited_articles_td relaited_articles_text\">\n                                <div class=\"container-item-h\"><a href=\"https:\/\/80.lv\/articles\/10-tips-on-getting-a-job-on-a-aaa-game\/\">10 Tips on Getting a Job on a AAA-game<\/a><\/div>\n                                <div class=\"tags\"><div class=\"tag\">Opinion<\/div><\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                <\/div>\n            <\/div><div class=\"container-item container-item-half-width this-iso-elem\">\n                <div class=\"container-item-inn\">\n                    <div class=\"relaited_articles\">\n                        <div class=\"relaited_articles_tr\">\n                            <div class=\"relaited_articles_td relaited_articles_img\"><a href=\"https:\/\/80.lv\/articles\/a-light-in-chorus-interview\/\"><img width=\"150\" height=\"150\" src=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/04\/ALIC_featured-150x150.jpg\" class=\"attachment-thumbnail size-thumbnail wp-post-image\" alt=\"ALIC_featured\" srcset=\"https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/04\/ALIC_featured-150x150.jpg 150w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/04\/ALIC_featured-300x300.jpg 300w, https:\/\/cdn.80.lv\/80.lv\/uploads\/2015\/04\/ALIC_featured.jpg 900w\" \/><\/a><\/div>\n                            <div class=\"relaited_articles_td relaited_articles_text\">\n                                <div class=\"container-item-h\"><a href=\"https:\/\/80.lv\/articles\/a-light-in-chorus-interview\/\">A Light In Chorus: Making Games Out of Light<\/a><\/div>\n                                <div class=\"tags\"><div class=\"tag\">Interview<\/div><\/div>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                <\/div>\n            <\/div>\n        <\/div>\n    <\/div>                <div id=\"zergnet-widget-46272\">\n            <\/div>\n        <\/div>\n    <\/div>\n<\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/section><link type=\"text\/css\" rel=\"stylesheet\" href=\"https:\/\/80.lv\/wp-content\/plugins\/xsolla-gallery-video\/style.css\" \/>","created_at":"2018-11-07T16:48:00+0000","updated_at":"2018-11-07T16:48:00+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"80.lv","preview_picture":"https:\/\/cdn.80.lv\/80.lv\/uploads\/2016\/11\/Collage_Fotor-3-1024x512.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":423,"title":"ZDoom \u2022 View topic - Resource Bookmarks","url":"https:\/\/forum.zdoom.org\/viewtopic.php?f=37&t=24346","content":"Post links to any resource sites (textures, sounds, sprites, models, etc) that you have bookmarked. Be sure to check that the links still work.<p>If everyone were to contribute to this thread, we'd quickly have a huge list of resources.<\/p><p>Newer links will be added to the bottom of the list.<\/p><p>General resources<\/p><p><a class=\"postlink\" href=\"http:\/\/opengameart.org\/\">http:\/\/opengameart.org\/<\/a><br \/>Free, legal art for open source game projects - textures, sprites and models available<\/p><p>Textures<\/p><p><a class=\"postlink\" href=\"http:\/\/www.cgtextures.com\/\">http:\/\/www.cgtextures.com\/<\/a><br \/>THE place to go for high-res photos and textures. Most (if not all) require manual editing to make them tile-able.<\/p><p><a class=\"postlink\" href=\"http:\/\/stilgar.bplaced.de\/\">http:\/\/stilgar.bplaced.de\/<\/a><br \/>Some nice textures here<\/p><p><a class=\"postlink\" href=\"http:\/\/www.sharecg.com\/b\/6\/Textures?PSID=96e7b510ad87a12d3f9998877f311a8d\">http:\/\/www.sharecg.com\/b\/6\/Textures?PSI ... 877f311a8d<\/a><br \/>Textures on ShareCG<\/p><p><a class=\"postlink\" href=\"http:\/\/www.davegh.com\/index.php\">http:\/\/www.davegh.com\/index.php<\/a><br \/>David Gurrea's textures<\/p><p><a class=\"postlink\" href=\"http:\/\/archivetextures.net\/\">http:\/\/archivetextures.net\/<\/a><br \/>More textures<\/p><p><a class=\"postlink\" href=\"http:\/\/www.mayang.com\/textures\/\">http:\/\/www.mayang.com\/textures\/<\/a><br \/>Mayang's Free Textures<\/p><p><a class=\"postlink\" href=\"http:\/\/www.bittbox.com\/freebies\/24-free-high-res-textures\">http:\/\/www.bittbox.com\/freebies\/24-free ... s-textures<\/a><br \/>24 free high res textures<\/p><p><a class=\"postlink\" href=\"http:\/\/www.tofurious.com\/2009\/02\/grunge-textures\/\">http:\/\/www.tofurious.com\/2009\/02\/grunge-textures\/<\/a><br \/>Grunge textures... the site is mainly about wedding photos but scroll to the bottom and you'll find 50 rusty\/grungey textures.<\/p><p><a class=\"postlink\" href=\"http:\/\/www.art.eonworks.com\/gallery\/texture\/texture_gallery_1.html\">http:\/\/www.art.eonworks.com\/gallery\/tex ... ery_1.html<\/a><br \/>Dawid Michalczyk's textures<\/p><p><a class=\"postlink\" href=\"http:\/\/gamebanana.com\/textures\">http:\/\/gamebanana.com\/textures<\/a><br \/>GameBanana is a site dedicated to providing resources for FPS games. Lots of content here.<\/p><p><a class=\"postlink\" href=\"http:\/\/local.wasp.uwa.edu.au\/~pbourke\/texture_colour\/\">http:\/\/local.wasp.uwa.edu.au\/~pbourke\/texture_colour\/<\/a> (submitted by Enjay)<br \/>Paul Bourke's textures... a pretty impressive choice of tiling textures can be found here<\/p><p><a class=\"postlink\" href=\"http:\/\/www.grsites.com\/textures\/\">http:\/\/www.grsites.com\/textures\/<\/a> (submitted by NeuralStunner)<br \/>Absolute Background Textures Archive. Free, tiling textures in various dimensions. Many GIF but some JPEG as well, but those are usually too detailed to notice the artifcts much.<\/p><p><a class=\"postlink-local\" href=\"http:\/\/forum.zdoom.org\/viewtopic.php?f=37&amp;t=25274\">viewtopic.php?f=37&amp;t=25274<\/a> (submitted by Ceeb)<br \/>Ceeb's stuff - an assortment of rips from various games.<\/p><p><a class=\"postlink\" href=\"http:\/\/scmapdb.com\/textures\">http:\/\/scmapdb.com\/textures<\/a><br \/>A collection of textures for Half-Life. You'd need a Half-Life WAD viewer (like Wally or our very own <a href=\"http:\/\/slade.mancubus.net\/\" class=\"postlink\">SLADE 3<\/a>!) before you can get your hands on the textures. Plenty of stuff here!<\/p><p><a class=\"postlink\" href=\"http:\/\/www.texturemate.com\/\">http:\/\/www.texturemate.com\/<\/a> (submitted by ETTiNGRiNDER) New!<br \/>Textures, stock photos for texturing, brushes, etc. More permissive with the terms of use than many similar sites.<\/p><p>Sprites<\/p><p><a class=\"postlink\" href=\"http:\/\/sdb.drshnaps.com\/\">http:\/\/sdb.drshnaps.com\/<\/a> (submitted by Xim)<br \/>A sprite sheet resource. Tons of sprite sheets from various games.<\/p><p><a class=\"postlink\" href=\"http:\/\/mmimageworld.dugtrio17.com\/\">http:\/\/mmimageworld.dugtrio17.com\/<\/a> (submitted by Enjay)<br \/>Wolf3D-related sprite edits... some are cool, some are crap. Useful for Wolf3-like mods.<\/p><p><a class=\"postlink\" href=\"http:\/\/spriters-resource.com\/\">http:\/\/spriters-resource.com\/<\/a> (submitted by Enjay)<br \/>The Spriter's Resource is a huge collection of sprite rips from various games.<\/p><p>Skyboxes<\/p><p><a class=\"postlink\" href=\"http:\/\/kell.quaddicted.com\/skyboxes.html\">http:\/\/kell.quaddicted.com\/skyboxes.html<\/a> (submitted by Enjay)<br \/>A collection of game-ready skyboxes generated using Terragen.<\/p><p><a class=\"postlink\" href=\"http:\/\/lvlworld.com\/review.php?id=2023\">http:\/\/lvlworld.com\/review.php?id=2023<\/a><br \/>Quake 3 Arena skybox collection. Someone put together a massive collection of skyboxes for Q3A. Get it while you can! Lots of skyboxes in there!<\/p><p><a class=\"postlink\" href=\"http:\/\/alexcpeterson.com\/spacescape\">http:\/\/alexcpeterson.com\/spacescape<\/a> (submitted by zrrion the insect) New!<br \/>Spacescape is a tool for creating space skyboxes with stars and nebulas.<\/p><p>Models<\/p><p><a class=\"postlink\" href=\"http:\/\/www.md2.sitters-electronics.nl\/\">http:\/\/www.md2.sitters-electronics.nl\/<\/a> (submitted by Enjay)<br \/>sitters is a talented modeler. You may have seen his stuff in some model packs and especially his Risen3D mods, which heavily use models. All of his models are low-poly and Doom-ready.<\/p><p><span style=\"text-decoration: line-through;\">http:\/\/www.blackrayne.net\/<\/span> (submitted by Enjay) (Domain has expired. Anyone have alternatives?)<br \/>Blackrayne has put up various low-poly models that he's created. They are all game-ready and easy to bring into GZDoom.<\/p><p>Sounds<\/p><p><a class=\"postlink\" href=\"http:\/\/www.freesound.org\/\">http:\/\/www.freesound.org\/<\/a> (submitted by Wildweasel)<br \/>The FreeSound Project, home of hundreds of thousands of Creative Commons licensed sound effects.<\/p><p><a class=\"postlink\" href=\"http:\/\/www.grsites.com\/sounds\/\">http:\/\/www.grsites.com\/sounds\/<\/a> (submitted by NeuralStunner)<br \/>Absolute Sound Effects Archive<\/p><p><a class=\"postlink\" href=\"http:\/\/noproblo.dayjo.org\/ZeldaSounds\/\">http:\/\/noproblo.dayjo.org\/ZeldaSounds\/<\/a> (submitted by NeuralStunner)<br \/>Legend of Zelda sound effects by HelpTheWretched<\/p><p>Music<\/p><p><a class=\"postlink\" href=\"http:\/\/www.vgmusic.com\/\">http:\/\/www.vgmusic.com\/<\/a> (submitted by NeuralStunner)<br \/>Video Game Music Archive. For those looking into doing TCs, or who just enjoy a particular game. Some awesome remixes too.<\/p><p><a class=\"postlink\" href=\"http:\/\/www.freemidi.org\/\">http:\/\/www.freemidi.org\/<\/a> (submitted by NeuralStunner)<br \/>An archive of MIDIs.<\/p><p><a class=\"postlink\" href=\"http:\/\/www.mididb.com\/\">http:\/\/www.mididb.com\/<\/a> (submitted by NeuralStunner)<br \/>Another archive of MIDIs.<\/p><p><a class=\"postlink\" href=\"http:\/\/mirsoft.info\/\">http:\/\/mirsoft.info\/<\/a> (submitted by Kinsie)<br \/>MIDIs, MODs and related<\/p><p><a class=\"postlink\" href=\"http:\/\/snesmusic.org\/v2\/\">http:\/\/snesmusic.org\/v2\/<\/a> (submitted by Kinsie)<br \/>SNES SPCs<\/p><p><a class=\"postlink\" href=\"http:\/\/project2612.org\/\">http:\/\/project2612.org\/<\/a> (submitted by Kinsie)<br \/>Genesis VGMs<\/p><p><a class=\"postlink\" href=\"http:\/\/www.thealmightyguru.com\/VGMPF\/Index.php\">http:\/\/www.thealmightyguru.com\/VGMPF\/Index.php<\/a> (submitted by Enjay)<br \/>OGG versions of otherwise hard-to-listen-to game tunes.<\/p><p><a class=\"postlink\" href=\"http:\/\/modarchive.org\/\">http:\/\/modarchive.org\/<\/a> (submitted by Enjay)<br \/>A huge resource of tracker music.<\/p><p><a class=\"postlink\" href=\"http:\/\/davidorr.newgrounds.com\/\">http:\/\/davidorr.newgrounds.com\/<\/a> (submitted by DoomRater)<br \/>David Orr's music archive<\/p><p><a class=\"postlink\" href=\"http:\/\/bifi.msxnet.org\/kss\/\">http:\/\/bifi.msxnet.org\/kss\/<\/a> (submitted by FireSeraphim) New!<br \/>An archive full of .KSS music files<\/p><p><a class=\"postlink\" href=\"http:\/\/midimelody.ru\/\">http:\/\/midimelody.ru\/<\/a> (submitted by zrrion the insect) New!<br \/>A collection of MIDIs of heavy metal bands, among other things<\/p><p>Fonts<\/p><p><a class=\"postlink\" href=\"http:\/\/www.grsites.com\/archive\/fonts\/\">http:\/\/www.grsites.com\/archive\/fonts\/<\/a> (submitted by NeuralStunner)<br \/>Lots of fonts available here... convert these for use in ZDoom by using <a href=\"http:\/\/baghead.drdteam.org\/tools.xml\" class=\"postlink\">Bagheadspidey's ZDoom Font Generator<\/a><\/p><p><a class=\"postlink\" href=\"http:\/\/bmf.wz.cz\/\">http:\/\/bmf.wz.cz\/<\/a> (submitted by Ceeb)<br \/>BMF fonts collection<\/p><p>Tools New!<\/p><p><a class=\"postlink\" href=\"http:\/\/info.sonicretro.org\/RotSprite\">http:\/\/info.sonicretro.org\/RotSprite<\/a> (submitted by wildweasel\/Blox) New!<br \/>Sprite rotation and scaling utility<\/p><p><a class=\"postlink\" href=\"http:\/\/www.positech.co.uk\/content\/explosion\/explosiongenerator.html\">http:\/\/www.positech.co.uk\/content\/explo ... rator.html<\/a> (submitted by Enjay) New!<br \/>Automatically generates a sprite sheet of explosions<\/p><p><a class=\"postlink\" href=\"http:\/\/www.saschawillems.de\/?page_id=253\">http:\/\/www.saschawillems.de\/?page_id=253<\/a> (submitted by Blox) New!<br \/>Another explosion generator<\/p>","created_at":"2018-11-07T16:47:59+0000","updated_at":"2018-11-07T16:47:59+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"forum.zdoom.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":422,"title":"Monkee's Image World","url":"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/index.cgi?category=guard&page=1","content":"<table border=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td align=\"left\"><\/td><td align=\"right\"><span>Total entries: 94<br \/>Goto page  1, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=2\">2<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=3\">3<\/a>,  ... <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=8\">8<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=9\">9<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=10\">10<\/a>  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=2\">Next<\/a><\/span><\/td><\/tr><\/table><hr \/><b>Preview<\/b>\n\t<b>Description<\/b>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/wsjafrika.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Afrika Korps Pack<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=wsj&amp;page=1\">WSJ<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=wsjafrika\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>421kb (Downloaded 1438 times)<\/span>\n\t\t<p>Here's another enemy sprites pack from Wolf Skevos-Jones. This time, the guards are meant to look like the Afrika Korps or other Nazi desert troops. They're pretty similar to the ones from Wolf's earlier \"Heer Sprites Pack\" but they have tan uniforms.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/half.gif\" alt=\"image\" \/>\n4.0 (36 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\" id=\"vote\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=male&amp;page=1\">Males<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-01-20&amp;page=1\">2005-01-20<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/soldierAK47.png\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>AK-47 Soldier<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=mr%20wolf&amp;page=1\">Mr Wolf<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=soldierAK47\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>57kb (Downloaded 1235 times)<\/span>\n\t\t<p>Uses a recolored Blake Stone Security guard with an officer head and the SS death frames. Done to replace the SS. Currently does not appear in any projects. Give credit to Mr Wolf if you use this character.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/half.gif\" alt=\"image\" \/>\n3.6 (32 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=male&amp;page=1\">Males<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-01-01&amp;page=1\">2005-01-01<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/allgemeiness.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Allgemeine SS Pack<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=wsj&amp;page=1\">WSJ<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=allgemeiness\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>402kb (Downloaded 1532 times)<\/span>\n\t\t<p>Here's a third enemy sprites pack from Wolf Skevos-Jones. This time, the guards are meant to look like the Allgemeine-SS, and as such, they have black uniforms with red Nazi armbands. Please give credit to WSJ if you use these guards in your project.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/half.gif\" alt=\"image\" \/>\n3.7 (29 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=male&amp;page=1\">Males<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-01-26&amp;page=1\">2005-01-26<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/ami.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Ami<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=ian%20summerfield&amp;page=1\">Ian Summerfield<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=ami\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>61kb (Downloaded 645 times)<\/span>\n\t\t<p>This character is hand drawn, and is based on the character by the same name from the \"Hi Hi puffy Ami Yumi\" TV show and Pop Rock Band. She is designed to replace guards of just about any type from Wolf3D. Please give credit to Ian Summerfield if you use this character in your project.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/>\n2.4 (18 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=cartoon&amp;page=1\">Cartoons<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=fem&amp;page=1\">Females<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-06-02&amp;page=1\">2005-06-02<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/amp.png\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Amp<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=ian%20summerfield&amp;page=1\">Ian Summerfield<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=amp\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>53kb (Downloaded 972 times)<\/span>\n\t\t<p>This creature is based on the Imp enemy from Doom, with the arms of one of the gun toting zombies, also from Doom. Please give credit to Ian Summerfield if you use this character in your project. Currently appears in no projects.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/>\n3.1 (34 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=doom&amp;page=1\">Doom<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-01-01&amp;page=1\">2005-01-01<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/archangel.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Archangel<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=majik%20monkee&amp;page=1\">Majik Monkee<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=archangel\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>10kb (Downloaded 662 times)<\/span>\n\t\t<p>This character is based on the Alien Controller enemy from \"Half Life\", and uses images taken directly from the game, scaled down and recolored. I originally used it as a SOD type ghost, but it could be used to replace a projectile only boss type enemy as well. I originally converted this character for \"Countdown to Chaos\", though I'm not sure I'll complete the project now.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/>\n2.5 (27 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=alien&amp;page=1\">Aliens and Monsters<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-02-28&amp;page=1\">2005-02-28<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/barney.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Barney Calhoun<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=tobixy&amp;page=1\">TobiXY<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=barney\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>58kb (Downloaded 889 times)<\/span>\n\t\t<p>Submitted by TobiXY, this character uses the brown guard from Wolf3D as a base, with some editing to make his uniform resemble that of Barney Calhoun (the security guard) from \"Half Life\". Please give credit to TobiXY if you use this character in your project.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/>\n3.2 (17 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=male&amp;page=1\">Males<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-02-25&amp;page=1\">2005-02-25<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/bjguard.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>BJ Guard<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=mason&amp;page=1\">Mason<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=bjguard\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>85kb (Downloaded 1237 times)<\/span>\n\t\t<p>Submitted by Mason, this is a set of BJ sprites that are based on recolored guards from the original Wolfenstein games.  Suitable for replacing most guard types from Wolf and SOD, though the images are not currently in BMP format.  Please give credit to Mason if you use this character in your project<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/half.gif\" alt=\"image\" \/>\n3.6 (29 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=male&amp;page=1\">Males<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-10-21&amp;page=1\">2005-10-21<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/bloo.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Bloo<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=joshua%20moore&amp;page=1\">Joshua Moore<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=bloo\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>28kb (Downloaded 524 times)<\/span>\n\t\t<p>Submitted by Joshua Moore, this character is entirely hand drawn, and is based on the character from the cartoon \"Foster's Home for Imaginary Friends\". It's designed to replace any guard from Wolfenstein or SOD. Please give credit to Craig McCracken for the character idea and Joshua for the Wolf3D version if you use this character in your project.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/>\n2.1 (28 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=cartoon&amp;page=1\">Cartoons<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-02-16&amp;page=1\">2005-02-16<\/a><\/span><br \/>\n<hr \/>\n<div style=\"text-align: center;\"><img src=\"http:\/\/mmimageworld.dugtrio17.com\/images\/briohulk2.gif\" title=\"Click to enlarge\" alt=\"image\" \/><\/div>\t\n\t\n\t\t<span><b>Brio Hulk 2<\/b><\/span> by <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?author=ian%20summerfield&amp;page=1\">Ian Summerfield<\/a>\n\t\t\t  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?download=briohulk2\"><img border=\"0\" src=\"http:\/\/mmimageworld.dugtrio17.com\/disk.png\" alt=\"image\" \/><\/a>\n\t\t\t<span>60kb (Downloaded 538 times)<\/span>\n\t\t<p>This character is hand drawn, and is based on the character by the same name from the \"Crash Bandicoot\" games.  He's designed to replace Mancubus from Doom II, though he can also be used as a replacement for Wolf3D and SOD guards. Please give credit to Ian Summerfield if you use this character in your project.<br \/><\/p><table border=\"0\" style=\"width: 100%;\"><tr><td style=\"width: 300px;\">User rating: <img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/whole.gif\" alt=\"image\" \/><img src=\"http:\/\/mmimageworld.dugtrio17.com\/stars\/half.gif\" alt=\"image\" \/>\n2.5 (21 votes)\t<\/td><td>\n\t<form method=\"get\" action=\"cgi-bin\/index.cgi\">\n\t\t<\/form>\n<\/td><\/tr><\/table><span>Categories: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=cartoon&amp;page=1\">Cartoons<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=crash&amp;page=1\">Crash Bandicoot<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=1\">Guards<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=male&amp;page=1\">Males<\/a><\/span><br \/><span>Date added: <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?date=2005-06-02&amp;page=1\">2005-06-02<\/a><\/span><br \/>\n<hr \/><span>Goto page  1, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=2\">2<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=3\">3<\/a>,  ... <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=8\">8<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=9\">9<\/a>, <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=10\">10<\/a>  <a href=\"http:\/\/mmimageworld.dugtrio17.com\/cgi-bin\/cgi-bin\/index.cgi?category=guard&amp;page=2\">Next<\/a><\/span><br \/><span>\tSubmissions and comments: <a href=\"mailto:dhw_majik_monkee@yahoo.com\">Majik Monkee<\/a><br \/>Site design by: <a href=\"mailto:winwolf3d@fastmail.us\">Adam Biser<\/a><br \/>Version 1.12.4\n<\/span>\n<br \/><a href=\"http:\/\/t.extreme-dm.com\/?login=mmiw\" target=\"_top\"><img src=\"http:\/\/t1.extreme-dm.com\/i.gif\" border=\"0\" height=\"38\" width=\"41\" alt=\"eXTReMe Tracker\" id=\"EXim\" \/><\/a>\n<noscript><img height=\"1\" width=\"1\" alt=\"\" src=\"http:\/\/e0.extreme-dm.com\/s9.g?login=mmiw&amp;j=n&amp;jv=n\" \/><\/noscript>","created_at":"2018-11-07T16:47:58+0000","updated_at":"2018-11-07T16:47:58+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"mmimageworld.dugtrio17.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":420,"title":"The 50 best free games on PC","url":"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/","content":"<div><main><p>The best free games are on PC, and if you want to know what the best 50 are then you\u2019ve come to the right place.<\/p>\n<p>Below you\u2019ll find an <em>entirely objective list<\/em> of best games you can download and play right now. You\u2019ll find links to each game, pointers towards similar games, and links to where you can read more about them on this very website.<\/p>\n<p>This list was published previously, but is being re-published now in a non-gallery format. The list of games remains unchanged for now \u2013 they\u2019re all still great.<\/p>\n<p>Before we continue, let\u2019s explain the rules.<\/p>\n<p>When we say free, we mean free. Free-to-play games aren\u2019t allowed, because no matter how generous a game might be, it\u2019s not free if there are items, characters or hats waiting to be bought. Games that have an alternate paid version like a remake or sequel, or which were once free before being let loose, are fine and you\u2019ll find a few in the list below. The rule is simple though: all of the games here are completely, one hundred percent free. <\/p>\n<p>Lastly, unless it has a standalone version, no modifications are allowed. If you need to buy a game in order to play the free game, then it\u2019s not free. Again: when we say free, we mean free. And if you\u2019re looking for the top ten in video form, we have you covered\u2026<\/p>\n<p>Can\u2019t find a game you love in the list? That must be because you are objectively wrong, but no matter. Hop to the comments and write an entry of your own, explaining why you adore the game, and drop in a link so other people can share in your objectively wrong love. For those unsure what to comment, I\u2019ll save you some time: Cave Story isn\u2019t on the list.<\/p>\n<p>Onwards towards the games! Hit the page links below to skip around.<\/p>\n\t\t\t\t\t<p>Page 1<a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/2\/\">Page 2<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/3\/\">Page 3<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/4\/\">Page 4<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/5\/\">Page 5<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/6\/\">Page 6<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/7\/\">Page 7<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/8\/\">Page 8<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/9\/\">Page 9<\/a><\/p>\t\t\t\t<\/main><\/div><div class=\"container-690\"><p class=\"disclaimer\">\n\t\t\t\t\tIf you click our links to online stores and make a purchase we may receive a few pennies. Find more information <a href=\"https:\/\/www.rockpapershotgun.com\/affiliate-linking-policy\/\">here<\/a>.\t\t\t\t<\/p><a href=\"https:\/\/www.rockpapershotgun.com\/support-us?utm_source=article-promo&amp;utm_medium=web&amp;utm_campaign=supporter\" class=\"support-us-promo\" rel=\"nofollow\">\n\t\t\n\t\t\t<img src=\"https:\/\/www.rockpapershotgun.com\/wp-content\/themes\/rps-2018\/images\/logo-inverted.svg\" alt=\"Rock Paper Shotgun\" \/>\n\t\t\n\t\t\n\t\t\t\n\t\t\t\tSupport RPS and get an ad-free site, extra articles, and free stuff!\t\t\t\n\t\t\tTell me more\n\t\t\n\t\t\n\t<\/a><\/div>","created_at":"2018-11-07T16:47:29+0000","updated_at":"2018-11-07T16:47:29+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/www.rockpapershotgun.com\/images\/14\/dec\/awardsheader_small.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":421,"title":"hackmud by Drizzly Bear","url":"https:\/\/drizzlybear.itch.io\/hackmud","content":"<p>&#13;\n\th a c k m u d is a incredible tangle of features, caveats, bugs and ruins --and I am but one solo developer. This makes this game &#13;\n\t<i>unbelievably complex<\/i> to maintain. HOWEVER-- As long as there are servers running the game, there will be balance patches and fixes.&#13;\n<\/p><p>&#13;\n\tRead AS MUCH AS YOU CAN about the game and player's experiences before making your purchase decision.&#13;\n<\/p><p>&#13;\n\tExcited to have you in game making messes with us. :D&#13;\n<\/p><p>&#13;\n\tHappy Hacking!&#13;\n<\/p><p>&#13;\n\tSean&#13;\n<\/p><p>&#13;\n\t<img src=\"https:\/\/www.dropbox.com\/s\/adjciismfz14rue\/trust%20communication%20greenlight.png?dl=1\" alt=\"image\" \/><br \/>&#13;\nh a c k m u d is a cyberpunk themed text-based hacking simulator for intel-compatible personal home computers. Get lost in a deep and complex world full of challenging puzzles as you engage in a rich narrative riddled with intrigue and deception.&#13;\n<\/p><ul><li>Learn who you can trust as you make friends and enemies exploring the depths of abandoned future internet cyber-structure<\/li>\n\t<li>Prove your worth to cyber-society by navigating the trials laid out for you by the Multi-User Domain's governing super-intelligence, 'The Trust'<\/li>\n\t<li>'Patch into the hardline' to access other users' protected systems and steal their digital goods<\/li>\n\t<li>Demonstrate hunger for power, flexibility of thought, and keen strategy to convince The Trust of your value and escape the virtual L.A.N.<\/li>\n\t<li>Let the text-based interface transport you back to the early internet-era when computing was full of discovery<\/li>\n\t<li>Create and destroy digital empires while exploring the constantly evolving user-created multiplayer world<\/li>\n\t<li>Use real-world hacking skills (social engineering, cracking, research, scripting...) for power and influence in the M.U.D. at large<\/li>\n<\/ul><h3>FEATURES<\/h3><ul><li>A deep, nuanced and engaging storyline full of discovery that will keep you guessing<\/li>\n\t<li>Encourages deep exploration and testing the limits of the system through rich side-content<\/li>\n\t<li>An intuitive retro-futuristic text-based computing interface that escalates to advanced with glitches, hidden commands, and arcane syntax<\/li>\n\t<li>90s film style 'hacking mode,' for when you need to get some 'serious' hacking done<\/li>\n\t<li>An original electronic soundtrack which will chill you out while exploring or amp you up when the pressure's on<\/li>\n\t<li>Persistent and constantly-evolving game universe powered by an almost endless stream of user created tools, puzzles, traps and much more<\/li>\n\t<li>A fluid and organic class system determined by the currently loaded upgrades on your system<\/li>\n\t<li>A fully scriptable in-game world with a global player-based economy of code and in-game botting<\/li>\n\t<li>Regular balance patches to keep the game fun and engaging for players of all styles and at all levels of experience<\/li>\n<\/ul><p>&#13;\n\t<strong>NOTICE: This game is online-only and <\/strong><strong>requires a steam account to play<\/strong><strong>.<\/strong>&#13;\n<\/p><p>&#13;\n\t(itch.io authentication is planned for the future)&#13;\n\t&#13;\n<\/p>","created_at":"2018-11-07T16:47:29+0000","updated_at":"2018-11-07T16:47:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"drizzlybear.itch.io","preview_picture":"https:\/\/img.itch.zone\/aW1hZ2UvNjE1OTMvMjc4MjE1LnBuZw==\/original\/JvchtA.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":417,"title":"Webcam Support for all browsers? \u00b7 Issue #97 \u00b7 auduno\/clmtrackr","url":"https:\/\/github.com\/auduno\/clmtrackr\/issues\/97","content":"Webcam Support for all browsers? \u00b7 Issue #97 \u00b7 auduno\/clmtrackr \u00b7 GitHub\n    \n  \n  \n    \n  \n        \n  <div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n    \n  \n  \n  <details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark\" open=\"open\"><summary aria-haspopup=\"dialog\" aria-label=\"Close dialog\">\n      \n    <\/summary><\/details>\n<p>Press h to open a hovercard with more details.<\/p>","created_at":"2018-11-07T16:47:28+0000","updated_at":"2018-11-07T16:47:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/10137?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":418,"title":"Plain Text Accounting, a guide to Ledger and friends","url":"https:\/\/plaintextaccounting.org\/","content":"<h3 id=\"frequently-asked-questions\">frequently asked questions<\/h3><p><strong>Who is this for?<\/strong><br \/>Those who are comfortable on the command line and who understand the value of storing information in plain text. If you need a complete GUI providing lots of guidance, you may prefer to use something else.<\/p><p><strong>Must I edit text and type cryptic commands?<\/strong><br \/>Not entirely! \"Plain Text Accounting\" is a broad description, referring mainly to the data format. We welcome optional <strong><a href=\"#ui-console\">GUIs<\/a><\/strong>, and they are <a href=\"#ui-console\">coming<\/a>.<\/p><p><strong>Who is using this, and how?<\/strong><br \/>See <strong><a href=\"https:\/\/github.com\/ledger\/ledger\/wiki\/Who's-using-ledger%3F\">Who's using Ledger?<\/a><\/strong> for some stories.<\/p><p><strong>What are the alternatives?<\/strong><br \/><strong><a href=\"https:\/\/en.wikipedia.org\/wiki\/Free_and_open-source_software\">FOSS<\/a><\/strong> tools like GNUCash, Grisbi, KMyMoney. Proprietary tools like Quicken\/Quickbooks, You Need A Budget. Online tools like Xero, FreeAgent. (There are <strong><a href=\"https:\/\/en.wikipedia.org\/wiki\/Comparison_of_accounting_software\">many<\/a><\/strong>). <strong><a href=\"https:\/\/backchannel.com\/a-spreadsheet-way-of-knowledge-8de60af7146e#.2hr7pi9pb\">Spreadsheets<\/a><\/strong>. Paper accounting. A bookkeeper.<\/p><p><strong>Why is this better than QuickBooks?<\/strong><br \/>Your data remains accessible. No yearly fees. Free software you can fix and port. Cross platform. Scriptable. Efficient.<\/p><p><strong>How do I communicate with my accountant?<\/strong><br \/>Clean up text reports by hand, print them as PDF, export CSV reports to a spreadsheet..<\/p><p><strong>What if my accountant only uses QuickBooks?<\/strong><br \/>You'll need a new accountant, or a duplicate set of books in QuickBooks, or to pioneer *ledger -&gt; QuickBooks exporting.<\/p><p><strong>Can I use this to do my taxes?<\/strong><br \/>You can use this to track and report the data needed for tax reporting. Fill out and submit tax forms with another tool, or by hand.<\/p><p><strong>How do I do budgeting?<\/strong><br \/>See <a href=\"#budgeting\">budgeting<\/a> below. I emulate YNAB-ish envelope budgetting (see third link).<\/p><p><strong>Double entry accounting? Where are the debits and credits?<\/strong><br \/>Most (not all) plain text accounting implementations use signed amounts instead of debits and credits. This makes them \"double entry light\" perhaps, but it has been a rather successful simplification, intuitive to most newcomers.<\/p><p><strong>Isn't personal accounting a waste of time?<\/strong><br \/>People have very different needs and practise personal accounting for many different reasons. There is of course a point of diminishing returns; tailor your accounting practices to your needs. Needs change over time. Some of us would benefit from doing more (or better) accounting, some less (I would guess this second group is smaller). In <a href=\"https:\/\/en.wikipedia.org\/wiki\/The_Millionaire_Next_Door\">The Millionaire Next Door<\/a> (highly recommended), one research finding was that above-average wealth accumulators spend more time on financial planning, which for many of us requires accounting as a foundation. \"Minimal time dedicated to financial planning is a leading indicator of a UAW [Under Accumulator of Wealth]\".<\/p><p><strong>Do you really enter every little transaction?<\/strong><br \/>Yes! Many folks in our community do it. Mahatma Gandhi reconciled to the penny every night. J.D. Rockefeller was famous for his ledgers. It's not required. I started doing it as a temporary learning exercise, and still like it. It makes troubleshooting and reconciling easier.<\/p><p><strong>How is that possible?<\/strong><br \/>Practice, and a process\/toolset that suits you. Some folks import most of the data from their banks, so little manual data entry is required. A few prefer to manually enter everything, for the increased awareness and insight. \"Manual\" data entry is usually assisted in some way: interactive console tools (hledger add and similar), web-based tools (hledger-web and similar), GUI tools (ledgerhelpers), smart editors (eg emacs &amp; ledger-mode), recurring transaction scripts. I currently use a mixture of bank CSV import and rapid copy\/paste in emacs. I spend 15 minutes a day on average, and for me that's currently a good investment.<\/p><p><strong>How do I use the transaction data in my bank's web or mobile app?<\/strong><br \/>If you can export it as CSV, you can import it and run queries against it. There are also some tools for converting OFX, QIF etc.<\/p><p><strong>So I've got a huge list of transactions recorded, duplicating my bank statements. How does that help?<\/strong><br \/>Accounting is modelling flows of money (or other value). Such a model aggregates information from many sources, in one trusted place. With it you can efficiently generate reports, forecast things (cashflow!), answer questions, try experiments. Some people need a very simple model, others benefit from a more detailed one, and we don't know up front what we might need in future. The most fundamental accounting data is a simple list of transactions (the journal). Once you have captured this, you can mine it for anything you may want later on. Plain text accounting provides nice open data format(s), tools and practices for doing this, and could be a good foundation for more powerful tools.<\/p><p><strong>Isn't a command-line tool too limited for real-world accounting needs?<\/strong><br \/><em>\"I am sure for a simple expense\/budget ledger it will work OK, but when it comes to recurring journals, multiple reconciliation accounts, inter company transfers, control account tracing etc., give me a nice GUI any day...\"<\/em><br \/>Understandable. The current plain text accounting tools provide a very generic double entry accounting system with which you can model such things, and script them. There are a number of generic GUIs available (hledger has curses and web interfaces, and there are web\/curses\/GTK interfaces for Ledger and beancount). But there are not yet a lot of rich task-specific GUIs. There's no reason they can't be built, though.<\/p><p><strong>Isn't a plain text format too limited for large organizations?<\/strong><br \/><em>\"it's pretty obvious that plain-text files don't scale to a multinational, with hundreds of accountants of various types all trying to work with the same files. Even with proper use of Git I bet that would get old fast. You would instead want a real database, with a schema, and some data validation and some programs\/webpages to smooth out the data entry and querying and whatnot.\"<\/em><br \/>I'm not sure. Current plain text accounting tools can do some schema definition and data validation, and will do more in future. The plain text storage format is open, human-readable, future-proof (useful even without the software), scales smoothly from simple to complex needs, and taps a huge ecosystem of highly useful tooling, such as version control systems. And, despite the name, there's no reason these tools can't support other kinds of storage, such as a database.<\/p><p><strong>Where can I see a comparison of hledger, Ledger, beancount, and the rest?<\/strong><br \/>Glad you asked! See below, and also <a href=\"#comparisons\">comparisons<\/a>. hledger's FAQ discusses differences from Ledger, Beancount docs probably do too.<\/p>","created_at":"2018-11-07T16:47:28+0000","updated_at":"2018-11-07T16:47:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"plaintextaccounting.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":419,"title":"simonmichael\/hledger","url":"https:\/\/github.com\/simonmichael\/hledger\/wiki\/hledger-basics-tutorial","content":"<p><strong>Contents<\/strong><\/p>\n<ol><li><a href=\"#check-your-hledger-installation\">Check your hledger installation<\/a><\/li>\n<li><a href=\"#locate-your-journal-file-with-hledger-stats\">Locate your journal file with \"hledger stats\"<\/a><\/li>\n<li><a href=\"#record-a-transaction-with-hledger-add\">Record a transaction with \"hledger add\"<\/a><\/li>\n<li><a href=\"#show-transactions-with-hledger-print\">Show transactions with \"hledger print\"<\/a><\/li>\n<li><a href=\"#examine-your-journal-file\">Examine your journal file<\/a><\/li>\n<li><a href=\"#a-convenience-inferred-amounts\">A convenience: inferred amounts<\/a><\/li>\n<li><a href=\"#edit-the-journal-file\">Edit the journal file<\/a><\/li>\n<li><a href=\"#two-spaces\">Two spaces<\/a><\/li>\n<li><a href=\"#unbalanced-transactions\">Unbalanced transactions<\/a><\/li>\n<li><a href=\"#record-a-transaction-by-editing\">Record a transaction by editing<\/a><\/li>\n<li><a href=\"#show-postings-and-a-running-total-with-hledger-register\">Show postings and a running total with \"hledger register\"<\/a><\/li>\n<li><a href=\"#show-a-per-account-register-report\">Show a per-account register report<\/a><\/li>\n<li><a href=\"#set-initial-account-balances\">Set initial account balances<\/a><\/li>\n<li><a href=\"#query-expressions\">Query expressions<\/a><\/li>\n<li><a href=\"#show-accounts-and-their-balances-with-hledger-balance\">Show accounts and their balances with \"hledger balance\"<\/a><\/li>\n<li><a href=\"#balance-shows-the-sum-of-matched-posting-amounts\">balance shows the sum of matched posting amounts<\/a><\/li>\n<li><a href=\"#review\">Review<\/a><\/li>\n<\/ol><p>Here you can learn hledger (and a little double-entry bookkeeping)\nby practicing, in small hands-on steps. You'll need:<\/p>\n<ol><li>\n<p>A little familiarity with\nthe <a href=\"http:\/\/tutorial.djangogirls.org\/en\/intro_to_command_line\/index.html\" rel=\"nofollow\">command<\/a>\n<a href=\"http:\/\/write.flossmanuals.net\/command-line\/introduction\" rel=\"nofollow\">line<\/a>.\nOr, the ability to ask for help on IRC.<\/p>\n<\/li>\n<li>\n<p>hledger. <a href=\"http:\/\/hledger.org\/download.html\" rel=\"nofollow\">Download<\/a> and install it if you haven't already.<\/p>\n<\/li>\n<\/ol><p>You'll learn the most if you master each step, in order.  Most\nsteps give you a task with a clear goal; if not, your task is to run\nthe examples and understand them.<\/p>\n<p>If you get stuck, find a problem, or have feedback, please report it on the\n<a href=\"http:\/\/hledger.org\/docs#helpfeedback\" rel=\"nofollow\">IRC channel or mail list<\/a>,\nor send a pull request for this page.<\/p>\n<p>Let's get started!<\/p>\n<h2>\n<a id=\"user-content-check-your-hledger-installation\" class=\"anchor\" href=\"#check-your-hledger-installation\" aria-hidden=\"true\"><\/a>Check your hledger installation<\/h2>\n<p>Get a command prompt, and run hledger to check the version. It should be reasonably <a href=\"http:\/\/hledger.org\/release-notes.html\" rel=\"nofollow\">up to date<\/a>. These exercises were last tested with this version:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger --version\nhledger 1.9<\/pre><\/div>\n<h2>\n<a id=\"user-content-locate-your-journal-file-with-hledger-stats\" class=\"anchor\" href=\"#locate-your-journal-file-with-hledger-stats\" aria-hidden=\"true\"><\/a>Locate your journal file with \"hledger stats\"<\/h2>\n<p>hledger reads financial transactions from a \"journal file\" (so named because it represents a <a href=\"http:\/\/en.wikipedia.org\/wiki\/General_Journal\" rel=\"nofollow\">General Journal<\/a>).\nThe default journal file is in your home directory; check its path using the <a href=\"http:\/\/hledger.org\/manual.html#stats\" rel=\"nofollow\">stats<\/a> command.\nYou should see something like:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger stats\nThe hledger journal file \"\/home\/YOU\/.hledger.journal\" was not found.\nPlease create it first, eg with \"hledger add\" or a text editor.\nOr, specify an existing journal file with -f or LEDGER_FILE.<\/pre><\/div>\n<p>Most hledger commands read this file but can not change it; the <code>add<\/code> and <code>web<\/code> commands can also write it.<\/p>\n<p>(If <code>stats<\/code> reports that the file exists, eg because you previously created it, move it out of the way temporarily for these exercises.)<\/p>\n<h2>\n<a id=\"user-content-record-a-transaction-with-hledger-add\" class=\"anchor\" href=\"#record-a-transaction-with-hledger-add\" aria-hidden=\"true\"><\/a>Record a transaction with \"hledger add\"<\/h2>\n<p>Follow the help and use the <a href=\"http:\/\/hledger.org\/manual.html#add\" rel=\"nofollow\">add<\/a> command to record your first transaction,\nan imaginary purchase at the supermarket.\nWe'll go through this in detail. Later you'll learn other ways to enter data.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger add\nCreating hledger journal file \"\/home\/YOU\/.hledger.journal\".\nAdding transactions to journal file \/home\/YOU\/.hledger.journal\nAny command line arguments will be used as defaults.\nUse tab key to complete, readline keys to edit, enter to accept defaults.\nAn optional (CODE) may follow transaction dates.\nAn optional ; COMMENT may follow descriptions or amounts.\nIf you make a mistake, enter &lt; at any prompt to restart the transaction.\nTo end a transaction, enter . when prompted.\nTo quit, enter . at a date prompt or press control-d or control-c.\nDate [2015\/05\/25]:<\/pre><\/div>\n<p><code>add<\/code> prompts for each transaction field. The first is the date.\nThe value in square brackets is the suggested default (today's date). Press enter to accept it.<\/p>\n<pre>Description: trip to the supermarket\n<\/pre>\n<p>Transactions have an optional description (a single line of text) to help you understand them.\nYou can describe the transaction here, or put a payee name, or leave it blank.\nType <code>trip to the supermarket<\/code> and press enter.<\/p>\n<pre>Account 1: expenses\n<\/pre>\n<p>Transactions have two or more accounts. Keep it simple; just enter <code>expenses<\/code> for the first one.<\/p>\n<p>If you're thinking \"expenses sounds more like a category\": it is, but double entry accounting calls those \"accounts\", too.\nA purchase is a transfer of money from an asset account to an expense account.\nAn asset is something you own, like some money in a bank account or in your pocket.\nOnce the money has been \"moved\" to an expense, you no longer own it, but the increasing balance in the expense account reminds you where it went.<\/p>\n<pre>Amount  1: $10\n<\/pre>\n<p>The amount being \"moved\" to <code>expenses<\/code>. In this case 10 US dollars.<\/p>\n<pre>Account 2: assets\n<\/pre>\n<p>Next, specify which account the money comes from. Just say <code>assets<\/code>.<\/p>\n<pre>Amount  2 ? [$-10.0]: \n<\/pre>\n<p>Now you're asked for the amount to \"move\" to or from the <code>assets<\/code> account.\nAs the default, hledger offers the amount required to \"balance\" the postings entered so far.\nThe minus sign indicates the money is moving from this account.\n(hledger uses the positive and negative sign instead of accounting's traditional \"debit\" and \"credit\" terminology.)\nIn a balanced transaction, the sum of posted amounts is zero, in other words no money disappears into thin air.\nPress enter to accept the default. It has an extra decimal place, but never mind.<\/p>\n<pre>Account 3 (or . to finish this transaction): .\n<\/pre>\n<p>Type <code>.<\/code> (period) and press enter.<\/p>\n<pre>2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets          $-10.0\nSave this transaction to the journal ? [y]:\n<\/pre>\n<p>You are given a chance to review the transaction just entered.\nHere you see hledger's plain text data format for journal entries:\na non-indented YYYY\/MM\/DD date, space, and description,\nfollowed by two or more indented posting lines, each containing an account name,\ntwo or more spaces, and an amount.\n(Account names can contain spaces, so at least two spaces are needed to separate them from the amount.)\nPress enter.<\/p>\n<pre>Saved.\nStarting the next transaction (. or ctrl-D\/ctrl-C to quit)\nDate [2015\/05\/25]: &lt;CTRL-D&gt;\n<\/pre>\n<p>hledger has saved it to the journal file and is ready for the next\nentry.  Press control-d (on Windows, control-c) once to exit.<\/p>\n<p><code>stats<\/code> should now report that your journal exists and contains one transaction:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger stats\nMain journal file        : \/home\/YOU\/.hledger.journal\nIncluded journal files   : \nTransactions span        : 2015-05-25 to 2015-05-26 (1 days)\nLast transaction         : 2015-05-25 (0 days ago)\nTransactions             : 1 (1.0 per day)\nTransactions last 30 days: 1 (0.0 per day)\nTransactions last 7 days : 1 (0.1 per day)\nPayees\/descriptions      : 1\nAccounts                 : 2 (depth 1)\nCommodities              : 1 ($)<\/pre><\/div>\n<h2>\n<a id=\"user-content-show-transactions-with-hledger-print\" class=\"anchor\" href=\"#show-transactions-with-hledger-print\" aria-hidden=\"true\"><\/a>Show transactions with \"hledger print\"<\/h2>\n<p>The <a href=\"http:\/\/hledger.org\/manual.html#print\" rel=\"nofollow\">print<\/a> command shows a tidied-up view of the transaction entries in your journal.\nSince there's just one so far, you should see:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger print\n2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets            $-10\n<\/pre><\/div>\n<h2>\n<a id=\"user-content-examine-your-journal-file\" class=\"anchor\" href=\"#examine-your-journal-file\" aria-hidden=\"true\"><\/a>Examine your journal file<\/h2>\n<p>List and print the journal file (on Windows, use <code>dir<\/code> and <code>type<\/code> and the file path from <code>hledger stats<\/code>):<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ ls -l ~\/.hledger.journal\n-rw-r--r--  1 YOU  YOU  114 May 25 16:55 \/home\/YOU\/.hledger.journal\n$ cat ~\/.hledger.journal\n; journal created 2015-05-25 by hledger\n2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets<\/pre><\/div>\n<h2>\n<a id=\"user-content-a-convenience-inferred-amounts\" class=\"anchor\" href=\"#a-convenience-inferred-amounts\" aria-hidden=\"true\"><\/a>A convenience: inferred amounts<\/h2>\n<p>Why is the amount missing from the assets posting above ?\nAs a convenience to make manual data entry easier, if one amount is missing\nhledger infers it so as to balance the transaction ($-10 in this case).\nOnly one missing amount is allowed in each transaction.\n<code>add<\/code> uses the same convention when it writes an entry.\n(To see all such inferred amounts in full, you can use <code>hledger print -x<\/code>.)<\/p>\n<h2>\n<a id=\"user-content-edit-the-journal-file\" class=\"anchor\" href=\"#edit-the-journal-file\" aria-hidden=\"true\"><\/a>Edit the journal file<\/h2>\n<p>Since the journal file is plain text, you can edit it directly with any text editor.\nEdit the file and change it to test whether two missing amounts is reported as an error. Eg:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ emacs ~\/.hledger.journal<\/pre><\/div>\n<p>Remove the expenses amount and save the file. It now looks like this:<\/p>\n<pre>2015\/05\/25 trip to the supermarket\n    expenses           \n    assets\n<\/pre>\n<p>Running <code>print<\/code> again, you should see:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>hledger: could not balance this transaction (can't have more than one missing amount; remember to put 2 or more spaces before amounts)\n2015\/05\/25 trip to the supermarket\n    expenses\n    assets\n<\/pre><\/div>\n<p>All hledger commands expect the journal to be well-formed, and will report an error and exit otherwise.<\/p>\n<h2>\n<a id=\"user-content-two-spaces\" class=\"anchor\" href=\"#two-spaces\" aria-hidden=\"true\"><\/a>Two spaces<\/h2>\n<p>Notice the last part of that error message: \"<code>... remember to put 2 or more spaces before amounts)<\/code>\".\nAnother cause of this error is forgetting to put two spaces before the\namount, like this:<\/p>\n<pre>2015\/05\/25 trip to the supermarket\n    expenses $10  ; &lt;- problem: only one space between expenses and $10\n    assets\n<\/pre>\n<p>Since account names may contain spaces, hledger thinks the first\nposting is to an account named \"<code>expenses $10<\/code>\", with a missing\namount.  So remember: two or more spaces.<\/p>\n<h2>\n<a id=\"user-content-unbalanced-transactions\" class=\"anchor\" href=\"#unbalanced-transactions\" aria-hidden=\"true\"><\/a>Unbalanced transactions<\/h2>\n<p>Edit the file to look like this:<\/p>\n<pre>2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets             $10  ; &lt;- deliberate problem: both amounts are positive\n<\/pre>\n<p>Here, we wrote both posting amounts but got the sign wrong on one of them, so they don't add up to zero.\nhledger should detect this mistake. Verify it by running some command, eg <code>print<\/code>. You should see:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger print\nhledger: could not balance this transaction (real postings are off by $20)\n2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets             $10\n<\/pre><\/div>\n<p>That makes sense. (It calls them \"real\" postings because there are some other kinds of posting you haven't learned about yet; they aren't important.)<\/p>\n<p>Correct the mistake by adding the minus sign, or just removing the assets amount entirely, and verify\nthat <code>print<\/code> works again:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger print\n2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets            $-10\n<\/pre><\/div>\n<h2>\n<a id=\"user-content-record-a-transaction-by-editing\" class=\"anchor\" href=\"#record-a-transaction-by-editing\" aria-hidden=\"true\"><\/a>Record a transaction by editing<\/h2>\n<p>Edit the file again and manually add a second purchase transaction.\nIt's often quickest to copy &amp; paste a similar entry, then change it.\nMake the file look like this:<\/p>\n<pre>2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets            $-10\n2015\/05\/26 forgot the bread\n    expenses            $5\n    assets\n<\/pre>\n<p>The blank line between transactions is customary, though not required.\nTest your work with <code>print<\/code>. You should see:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger print\n2015\/05\/25 trip to the supermarket\n    expenses           $10\n    assets            $-10\n2015\/05\/26 forgot the bread\n    expenses            $5\n    assets             $-5\n<\/pre><\/div>\n<h2>\n<a id=\"user-content-show-postings-and-a-running-total-with-hledger-register\" class=\"anchor\" href=\"#show-postings-and-a-running-total-with-hledger-register\" aria-hidden=\"true\"><\/a>Show postings and a running total with \"hledger register\"<\/h2>\n<p>The <a href=\"http:\/\/hledger.org\/manual.html#register\" rel=\"nofollow\">register<\/a> command shows transactions in a different format. More precisely, it shows postings.\nRemember, a posting is an increase or decrease of some account by some amount, and a transaction contains two or more of them.\nRun <code>register<\/code> and compare with the output of <code>print<\/code> above. You should see:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger register\n2015\/05\/25 trip to the supermarket  expenses                          $10           $10\n                                    assets                           $-10             0\n2015\/05\/26 forgot the bread         expenses                           $5            $5\n                                    assets                            $-5             0<\/pre><\/div>\n<p>Postings are displayed one per line.\nThe transaction's date and description is displayed only for the first posting in each transaction.\nNext we see the posted account's name and the amount posted.\nThe final column is a running total of the posted amounts.<\/p>\n<h2>\n<a id=\"user-content-show-a-per-account-register-report\" class=\"anchor\" href=\"#show-a-per-account-register-report\" aria-hidden=\"true\"><\/a>Show a per-account register report<\/h2>\n<p>Notice how the running total above keeps resetting to 0.\nThis makes sense (since we know each transaction's postings add up to zero) but isn't very useful.\nThe register report is more useful when we restrict it to a subset of postings -\nsay, only the postings within a single account.\nYou can do this by specifying the account name as a command line argument.<\/p>\n<p>Run a register report for the <code>expenses<\/code> account. You should see something like the below.\n(On POSIX platforms, this command uses the terminal width so the output may look slightly different.\nYou can force it to look like the below by running <code>export COLUMNS=80<\/code> first:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger register expenses\n2015\/05\/25 trip to the super..  expenses                       $10           $10\n2015\/05\/26 forgot the bread     expenses                        $5           $15<\/pre><\/div>\n<p>Now it's clear that your <code>expenses<\/code> balance - ie, the total amount spent - has increased to $15.<\/p>\n<p>Your <code>assets<\/code> balance should have dropped accordingly. Check it:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger register assets\n2015\/05\/25 trip to the super..  assets                        $-10          $-10\n2015\/05\/26 forgot the bread     assets                         $-5          $-15<\/pre><\/div>\n<h2>\n<a id=\"user-content-set-initial-account-balances\" class=\"anchor\" href=\"#set-initial-account-balances\" aria-hidden=\"true\"><\/a>Set initial account balances<\/h2>\n<p>hledger assumes every account starts with a zero balance,\nso in the previous example, we see the withdrawals producing a negative running balance.\nLet's assume <code>assets<\/code> represents a real-world asset, like your bank checking account, and you want to start tracking it from 2015\/05\/01 onward, and on that day it contained exactly $500.\nTo show the real-world account balance, edit your journal file and add this transaction at the top:<\/p>\n<pre>2015\/05\/01 set initial assets balance\n    assets                              $500\n    equity:opening balances\n<\/pre>\n<p>The other account name doesn't matter too much; <code>equity:opening balances<\/code> is conventional.\n(You could also use an <a href=\"http:\/\/hledger.org\/manual.html#virtual-postings\" rel=\"nofollow\">unbalanced transaction<\/a> for this if you prefer.)\nNow the report looks like this, with an accurate running balance on each date (hledger calls this a <em>historical balance<\/em>):<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger register assets\n2015\/05\/01 set initial asset..  assets                        $500          $500\n2015\/05\/25 trip to the super..  assets                        $-10          $490\n2015\/05\/26 forgot the bread     assets                         $-5          $485<\/pre><\/div>\n<h2>\n<a id=\"user-content-query-expressions\" class=\"anchor\" href=\"#query-expressions\" aria-hidden=\"true\"><\/a>Query expressions<\/h2>\n<p>The account name argument above is an example of a\n<a href=\"http:\/\/hledger.org\/manual.html#queries\" rel=\"nofollow\">query expression<\/a>, a search pattern which restricts a report to a subset of the data.\nIn this way  you can make very precise queries.<\/p>\n<p>Note that it is a case-insensitive regular expression which matches anywhere inside the account name.\nSo \"<code>e<\/code>\" would match both <code>expenses<\/code> and <code>assets<\/code>.<\/p>\n<p>And if you had an account named <code>other assets<\/code>, \"<code>assets<\/code>\" would also match that, so to match only the <code>assets<\/code>\naccount you'd need a more precise pattern like \"<code>^assets$<\/code>\".\n(In a regular expression <code>^<\/code> means \"match at the beginning\" and <code>$<\/code> means \"match at the end\".)\nIf this doesn't make sense, read a little about <a href=\"http:\/\/hledger.org\/manual.html#regular-expressions\" rel=\"nofollow\">regular expressions<\/a>.<\/p>\n<p>Multiple query arguments are ANDed and ORed together in a fixed way - follow the link for details.\nBasically queries on the same field are ORed, and queries on different fields are ANDed.<\/p>\n<p>Run the following examples and make sure they make sense, consulting the manual as needed.<\/p>\n<p>Show only transactions whose description ends with \"bread\":<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger print desc:bread$\n2015\/05\/26 forgot the bread\n    expenses            $5\n    assets             $-5\n<\/pre><\/div>\n<p>Show only postings on or after a certain date to an account whose name contains \"exp\":<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger register date:2015\/5\/26- exp\n2015\/05\/26 forgot the bread     expenses                        $5            $5<\/pre><\/div>\n<h2>\n<a id=\"user-content-show-accounts-and-their-balances-with-hledger-balance\" class=\"anchor\" href=\"#show-accounts-and-their-balances-with-hledger-balance\" aria-hidden=\"true\"><\/a>Show accounts and their balances with \"hledger balance\"<\/h2>\n<p>The third of hledger's three core reporting commands is <a href=\"http:\/\/hledger.org\/manual.html#balance\" rel=\"nofollow\">balance<\/a>.\nUse it to list all the accounts posted to, and their ending balance.\nYou should see account balances agreeing with the final running total in the register reports above:<\/p>\n<pre>$ hledger balance\n                $-15  assets\n                 $15  expenses\n--------------------\n                   0\n<\/pre>\n<p>The overall total of these balances is also shown. As with other reports, you can use a query expression to select a subset of the data to report on.\nEg:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger balance assets\n                $-15  assets\n--------------------\n                $-15<\/pre><\/div>\n<h2>\n<a id=\"user-content-balance-shows-the-sum-of-matched-posting-amounts\" class=\"anchor\" href=\"#balance-shows-the-sum-of-matched-posting-amounts\" aria-hidden=\"true\"><\/a>balance shows the sum of matched posting amounts<\/h2>\n<p>Here's a balance report based only on the postings dated 2015\/5\/26:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger balance date:2015\/5\/26\n                 $-5  assets\n                  $5  expenses\n--------------------\n                   0<\/pre><\/div>\n<p>As you can see from this, <code>balance<\/code> does not necessarily report\nreal-world account balances; rather, it shows the sum of the postings\nyou have selected.  If you're not sure what those are, run a\n<code>register<\/code> report with the same arguments to see them:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ hledger register date:2015\/5\/26\n2015\/05\/26 forgot the bread     expenses                        $5            $5\n                                assets                         $-5             0<\/pre><\/div>\n<h2>\n<a id=\"user-content-review\" class=\"anchor\" href=\"#review\" aria-hidden=\"true\"><\/a>Review<\/h2>\n<p>You have learned:<\/p>\n<ul><li>\n<p>a simple plain text notation for recording financial transactions, used by hledger, Ledger and others<\/p>\n<\/li>\n<li>\n<p>what is the journal file, where it is, and how to get statistics on it with <code>hledger stats<\/code><\/p>\n<\/li>\n<li>\n<p>how to record new transactions using <code>hledger add<\/code><\/p>\n<\/li>\n<li>\n<p>how to record transactions by editing the journal file<\/p>\n<\/li>\n<li>\n<p>what the journal entry for a purchase looks like<\/p>\n<\/li>\n<li>\n<p>how to detect some common errors, by eye or with hledger<\/p>\n<\/li>\n<li>\n<p>how hledger selects data to report on, and how to select by account, description, or date<\/p>\n<\/li>\n<li>\n<p>how to list transactions with <code>hledger print<\/code><\/p>\n<\/li>\n<li>\n<p>how to list postings and see an account's balance over time with <code>hledger register<\/code><\/p>\n<\/li>\n<li>\n<p>how to list accounts and their current balance, or the sum of their postings in some period, with <code>hledger balance<\/code><\/p>\n<\/li>\n<\/ul>","created_at":"2018-11-07T16:47:28+0000","updated_at":"2018-11-07T16:47:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/10144?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":415,"title":"auduno\/clmtrackr","url":"https:\/\/github.com\/auduno\/clmtrackr\/blob\/dev\/examples\/clm_emotiondetection.html","content":"&lt;!doctype html&gt;\n      \n        &lt;html lang=\"en\"&gt;\n      \n        \t&lt;head&gt;\n      \n        \t\t&lt;title&gt;Face tracker&lt;\/title&gt;\n      \n        \t\t&lt;meta charset=\"utf-8\"&gt;\n      \n        \t\t&lt;link href=\".\/styles\/bootstrap.min.css\" rel=\"stylesheet\" type=\"text\/css\"&gt;\n      \n        \t\t&lt;style&gt;\n      \n        \t\t\t@import url(https:\/\/fonts.googleapis.com\/css?family=Lato:300italic,700italic,300,700);\n      \n        \n      \n        \t\t\tbody {\n      \n        \t\t\t\tfont-family: 'Lato';\n      \n        \t\t\t\tbackground-color: #f0f0f0;\n      \n        \t\t\t\tmargin: 0px auto;\n      \n        \t\t\t\tmax-width: 1150px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#overlay {\n      \n        \t\t\t\tposition: absolute;\n      \n        \t\t\t\ttop: 0px;\n      \n        \t\t\t\tleft: 0px;\n      \n        \t\t\t\t-o-transform : scaleX(-1);\n      \n        \t\t\t\t-webkit-transform : scaleX(-1);\n      \n        \t\t\t\ttransform : scaleX(-1);\n      \n        \t\t\t\t-ms-filter : fliph; \/*IE*\/\n      \n        \t\t\t\tfilter : fliph; \/*IE*\/\n      \n        \n      \n        \t\t\t\t\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#videoel {\n      \n        \t\t\t\t-o-transform : scaleX(-1);\n      \n        \t\t\t\t-webkit-transform : scaleX(-1);\n      \n        \t\t\t\ttransform : scaleX(-1);\n      \n        \t\t\t\t-ms-filter : fliph; \/*IE*\/\n      \n        \t\t\t\tfilter : fliph; \/*IE*\/\n      \n        \n      \n        \t\t\t\t\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#container {\n      \n        \t\t\t\tposition : relative;\n      \n        \t\t\t\twidth : 370px;\n      \n        \t\t\t\t\/*margin : 0px auto;*\/\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#content {\n      \n        \t\t\t\tmargin-top : 50px;\n      \n        \t\t\t\tmargin-left : auto;\n      \n        \t\t\t\tmargin-right : auto;\n      \n        \t\t\t\tmax-width: 600px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\th2 {\n      \n        \t\t\t\tfont-weight : 400;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t.btn {\n      \n        \t\t\t\tfont-family: 'Lato';\n      \n        \t\t\t\tfont-size: 16px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#controls {\n      \n        \t\t\t\ttext-align : center;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#emotion_container {\n      \n        \t\t\t\twidth: 600px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#emotion_icons {\n      \n        \t\t\t\theight: 50px;\n      \n        \t\t\t\tpadding-left: 40px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t.emotion_icon {\n      \n        \t\t\t\twidth : 40px;\n      \n        \t\t\t\theight : 40px;\n      \n        \t\t\t\tmargin-top: 5px;\n      \n        \t\t\t\t\/*margin-left : 13px;*\/\n      \n        \t\t\t\tmargin-left : 35px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#emotion_chart, #emotion_icons {\n      \n        \t\t\t\tmargin: 0 auto;\n      \n        \t\t\t\twidth : 400px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#icon1, #icon2, #icon3, #icon4, #icon5, #icon6 {\n      \n        \t\t\t\tvisibility : hidden;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t\/* d3 *\/\n      \n        \t\t\t.bar {\n      \n        \t\t\t\tfill : steelblue;\n      \n        \t\t\t\tfill-opacity : .9;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t&lt;\/style&gt;\n      \n        \t\t&lt;script&gt;\n      \n        \t\t\t\/\/ getUserMedia only works over https in Chrome 47+, so we redirect to https. Also notify user if running from file.\n      \n        \t\t\tif (window.location.protocol == \"file:\") {\n      \n        \t\t\t\talert(\"You seem to be running this example directly from a file. Note that these examples only work when served from a server or localhost due to canvas cross-domain restrictions.\");\n      \n        \t\t\t} else if (window.location.hostname !== \"localhost\" &amp;&amp; window.location.protocol !== \"https:\"){\n      \n        \t\t\t\twindow.location.protocol = \"https\";\n      \n        \t\t\t}\n      \n        \t\t&lt;\/script&gt;\n      \n        \t\t&lt;script type=\"text\/javascript\"&gt;\n      \n        \n      \n        \t\t\tvar _gaq = _gaq || [];\n      \n        \t\t\t_gaq.push(['_setAccount', 'UA-32642923-1']);\n      \n        \t\t\t_gaq.push(['_trackPageview']);\n      \n        \n      \n        \t\t\t(function() {\n      \n        \t\t\t\tvar ga = document.createElement('script'); ga.type = 'text\/javascript'; ga.async = true;\n      \n        \t\t\t\tga.src = ('https:' == document.location.protocol ? 'https:\/\/ssl' : 'http:\/\/www') + '.google-analytics.com\/ga.js';\n      \n        \t\t\t\tvar s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);\n      \n        \t\t\t})();\n      \n        \n      \n        \t\t&lt;\/script&gt;\n      \n        \t&lt;\/head&gt;\n      \n        \t&lt;body&gt;\n      \n        \t\t&lt;script src=\".\/js\/libs\/utils.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\"..\/build\/clmtrackr.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\"..\/models\/model_pca_20_svm.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/libs\/Stats.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/libs\/d3.min.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/emotion_classifier.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/emotionmodel.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;div id=\"content\"&gt;\n      \n        \t\t\t&lt;h2&gt;Emotion detection example&lt;\/h2&gt;\n      \n        \t\t\t&lt;p&gt;This a test of emotion detection based on parameter output from &lt;em&gt;clmtrackr&lt;\/em&gt;.&lt;\/p&gt;\n      \n        \t\t\t&lt;div id=\"container\"&gt;\n      \n        \t\t\t\t&lt;video id=\"videoel\" width=\"400\" height=\"300\" preload=\"auto\" loop playsinline autoplay&gt;\n      \n        \t\t\t\t&lt;\/video&gt;\n      \n        \t\t\t\t&lt;canvas id=\"overlay\" width=\"400\" height=\"300\"&gt;&lt;\/canvas&gt;\n      \n        \t\t\t&lt;\/div&gt;\n      \n        \t\t\t&lt;div id=\"emotion_container\"&gt;\n      \n        \t\t\t\t&lt;div id=\"emotion_icons\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon1\" src=\".\/media\/icon_angry.png\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon2\" src=\".\/media\/icon_sad.png\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon3\" src=\".\/media\/icon_surprised.png\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon4\" src=\".\/media\/icon_happy.png\"&gt;\n      \n        \t\t\t\t&lt;\/div&gt;\n      \n        \t\t\t\t&lt;div id='emotion_chart'&gt;&lt;\/div&gt;\n      \n        \t\t\t&lt;\/div&gt;\n      \n        \t\t\t&lt;div id=\"controls\"&gt;\n      \n        \t\t\t\t&lt;input class=\"btn\" type=\"button\" value=\"wait, loading video\" disabled=\"disabled\" onclick=\"startVideo()\" id=\"startbutton\"&gt;&lt;\/input&gt;\n      \n        \t\t\t&lt;\/div&gt;\n      \n        \t\t\t&lt;a href=\"https:\/\/github.com\/auduno\/clmtrackr\"&gt;&lt;img style=\"position: absolute; top: 0; left: 0; border: 0;\" src=\"https:\/\/s3.amazonaws.com\/github\/ribbons\/forkme_left_green_007200.png\" alt=\"Fork me on GitHub\"&gt;&lt;\/a&gt;\n      \n        \t\t\t&lt;script&gt;\n      \n        \t\t\t\tvar vid = document.getElementById('videoel');\n      \n        \t\t\t\tvar vid_width = vid.width;\n      \n        \t\t\t\tvar vid_height = vid.height;\n      \n        \t\t\t\tvar overlay = document.getElementById('overlay');\n      \n        \t\t\t\tvar overlayCC = overlay.getContext('2d');\n      \n        \n      \n        \t\t\t\t\/********** check and set up video\/webcam **********\/\n      \n        \n      \n        \t\t\t\tfunction enablestart() {\n      \n        \t\t\t\t\tvar startbutton = document.getElementById('startbutton');\n      \n        \t\t\t\t\tstartbutton.value = \"start\";\n      \n        \t\t\t\t\tstartbutton.disabled = null;\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction adjustVideoProportions() {\n      \n        \t\t\t\t\t\/\/ resize overlay and video if proportions are different\n      \n        \t\t\t\t\t\/\/ keep same height, just change width\n      \n        \t\t\t\t\tvar proportion = vid.videoWidth\/vid.videoHeight;\n      \n        \t\t\t\t\tvid_width = Math.round(vid_height * proportion);\n      \n        \t\t\t\t\tvid.width = vid_width;\n      \n        \t\t\t\t\toverlay.width = vid_width;\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction gumSuccess( stream ) {\n      \n        \t\t\t\t\t\/\/ add camera stream if getUserMedia succeeded\n      \n        \t\t\t\t\tif (\"srcObject\" in vid) {\n      \n        \t\t\t\t\t\tvid.srcObject = stream;\n      \n        \t\t\t\t\t} else {\n      \n        \t\t\t\t\t\tvid.src = (window.URL &amp;&amp; window.URL.createObjectURL(stream));\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t\tvid.onloadedmetadata = function() {\n      \n        \t\t\t\t\t\tadjustVideoProportions();\n      \n        \t\t\t\t\t\tvid.play();\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t\tvid.onresize = function() {\n      \n        \t\t\t\t\t\tadjustVideoProportions();\n      \n        \t\t\t\t\t\tif (trackingStarted) {\n      \n        \t\t\t\t\t\t\tctrack.stop();\n      \n        \t\t\t\t\t\t\tctrack.reset();\n      \n        \t\t\t\t\t\t\tctrack.start(vid);\n      \n        \t\t\t\t\t\t}\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction gumFail() {\n      \n        \t\t\t\t\talert(\"There was some problem trying to fetch video from your webcam. If you have a webcam, please make sure to accept when the browser asks for access to your webcam.\");\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tnavigator.getUserMedia = navigator.getUserMedia || navigator.webkitGetUserMedia || navigator.mozGetUserMedia || navigator.msGetUserMedia;\n      \n        \t\t\t\twindow.URL = window.URL || window.webkitURL || window.msURL || window.mozURL;\n      \n        \n      \n        \t\t\t\t\/\/ check for camerasupport\n      \n        \t\t\t\tif (navigator.mediaDevices) {\n      \n        \t\t\t\t\tnavigator.mediaDevices.getUserMedia({video : true}).then(gumSuccess).catch(gumFail);\n      \n        \t\t\t\t} else if (navigator.getUserMedia) {\n      \n        \t\t\t\t\tnavigator.getUserMedia({video : true}, gumSuccess, gumFail);\n      \n        \t\t\t\t} else {\n      \n        \t\t\t\t\talert(\"This demo depends on getUserMedia, which your browser does not seem to support. :(\");\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tvid.addEventListener('canplay', enablestart, false);\n      \n        \n      \n        \t\t\t\t\/*********** setup of emotion detection *************\/\n      \n        \n      \n        \t\t\t\t\/\/ set eigenvector 9 and 11 to not be regularized. This is to better detect motion of the eyebrows\n      \n        \t\t\t\tpModel.shapeModel.nonRegularizedVectors.push(9);\n      \n        \t\t\t\tpModel.shapeModel.nonRegularizedVectors.push(11);\n      \n        \n      \n        \t\t\t\tvar ctrack = new clm.tracker({useWebGL : true});\n      \n        \t\t\t\tctrack.init(pModel);\n      \n        \t\t\t\tvar trackingStarted = false;\n      \n        \n      \n        \t\t\t\tfunction startVideo() {\n      \n        \t\t\t\t\t\/\/ start video\n      \n        \t\t\t\t\tvid.play();\n      \n        \t\t\t\t\t\/\/ start tracking\n      \n        \t\t\t\t\tctrack.start(vid);\n      \n        \t\t\t\t\ttrackingStarted = true;\n      \n        \t\t\t\t\t\/\/ start loop to draw face\n      \n        \t\t\t\t\tdrawLoop();\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction drawLoop() {\n      \n        \t\t\t\t\trequestAnimFrame(drawLoop);\n      \n        \t\t\t\t\toverlayCC.clearRect(0, 0, vid_width, vid_height);\n      \n        \t\t\t\t\t\/\/psrElement.innerHTML = \"score :\" + ctrack.getScore().toFixed(4);\n      \n        \t\t\t\t\tif (ctrack.getCurrentPosition()) {\n      \n        \t\t\t\t\t\tctrack.draw(overlay);\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t\tvar cp = ctrack.getCurrentParameters();\n      \n        \n      \n        \t\t\t\t\tvar er = ec.meanPredict(cp);\n      \n        \t\t\t\t\tif (er) {\n      \n        \t\t\t\t\t\tupdateData(er);\n      \n        \t\t\t\t\t\tfor (var i = 0;i &lt; er.length;i++) {\n      \n        \t\t\t\t\t\t\tif (er[i].value &gt; 0.4) {\n      \n        \t\t\t\t\t\t\t\tdocument.getElementById('icon'+(i+1)).style.visibility = 'visible';\n      \n        \t\t\t\t\t\t\t} else {\n      \n        \t\t\t\t\t\t\t\tdocument.getElementById('icon'+(i+1)).style.visibility = 'hidden';\n      \n        \t\t\t\t\t\t\t}\n      \n        \t\t\t\t\t\t}\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tdelete emotionModel['disgusted'];\n      \n        \t\t\t\tdelete emotionModel['fear'];\n      \n        \t\t\t\tvar ec = new emotionClassifier();\n      \n        \t\t\t\tec.init(emotionModel);\n      \n        \t\t\t\tvar emotionData = ec.getBlank();\n      \n        \n      \n        \t\t\t\t\/************ d3 code for barchart *****************\/\n      \n        \n      \n        \t\t\t\tvar margin = {top : 20, right : 20, bottom : 10, left : 40},\n      \n        \t\t\t\t\twidth = 400 - margin.left - margin.right,\n      \n        \t\t\t\t\theight = 100 - margin.top - margin.bottom;\n      \n        \n      \n        \t\t\t\tvar barWidth = 30;\n      \n        \n      \n        \t\t\t\tvar formatPercent = d3.format(\".0%\");\n      \n        \n      \n        \t\t\t\tvar x = d3.scale.linear()\n      \n        \t\t\t\t\t.domain([0, ec.getEmotions().length]).range([margin.left, width+margin.left]);\n      \n        \n      \n        \t\t\t\tvar y = d3.scale.linear()\n      \n        \t\t\t\t\t.domain([0,1]).range([0, height]);\n      \n        \n      \n        \t\t\t\tvar svg = d3.select(\"#emotion_chart\").append(\"svg\")\n      \n        \t\t\t\t\t.attr(\"width\", width + margin.left + margin.right)\n      \n        \t\t\t\t\t.attr(\"height\", height + margin.top + margin.bottom)\n      \n        \n      \n        \t\t\t\tsvg.selectAll(\"rect\").\n      \n        \t\t\t\t\tdata(emotionData).\n      \n        \t\t\t\t\tenter().\n      \n        \t\t\t\t\tappend(\"svg:rect\").\n      \n        \t\t\t\t\tattr(\"x\", function(datum, index) { return x(index); }).\n      \n        \t\t\t\t\tattr(\"y\", function(datum) { return height - y(datum.value); }).\n      \n        \t\t\t\t\tattr(\"height\", function(datum) { return y(datum.value); }).\n      \n        \t\t\t\t\tattr(\"width\", barWidth).\n      \n        \t\t\t\t\tattr(\"fill\", \"#2d578b\");\n      \n        \n      \n        \t\t\t\tsvg.selectAll(\"text.labels\").\n      \n        \t\t\t\t\tdata(emotionData).\n      \n        \t\t\t\t\tenter().\n      \n        \t\t\t\t\tappend(\"svg:text\").\n      \n        \t\t\t\t\tattr(\"x\", function(datum, index) { return x(index) + barWidth; }).\n      \n        \t\t\t\t\tattr(\"y\", function(datum) { return height - y(datum.value); }).\n      \n        \t\t\t\t\tattr(\"dx\", -barWidth\/2).\n      \n        \t\t\t\t\tattr(\"dy\", \"1.2em\").\n      \n        \t\t\t\t\tattr(\"text-anchor\", \"middle\").\n      \n        \t\t\t\t\ttext(function(datum) { return datum.value;}).\n      \n        \t\t\t\t\tattr(\"fill\", \"white\").\n      \n        \t\t\t\t\tattr(\"class\", \"labels\");\n      \n        \n      \n        \t\t\t\tsvg.selectAll(\"text.yAxis\").\n      \n        \t\t\t\t\tdata(emotionData).\n      \n        \t\t\t\t\tenter().append(\"svg:text\").\n      \n        \t\t\t\t\tattr(\"x\", function(datum, index) { return x(index) + barWidth; }).\n      \n        \t\t\t\t\tattr(\"y\", height).\n      \n        \t\t\t\t\tattr(\"dx\", -barWidth\/2).\n      \n        \t\t\t\t\tattr(\"text-anchor\", \"middle\").\n      \n        \t\t\t\t\tattr(\"style\", \"font-size: 12\").\n      \n        \t\t\t\t\ttext(function(datum) { return datum.emotion;}).\n      \n        \t\t\t\t\tattr(\"transform\", \"translate(0, 18)\").\n      \n        \t\t\t\t\tattr(\"class\", \"yAxis\");\n      \n        \n      \n        \t\t\t\tfunction updateData(data) {\n      \n        \t\t\t\t\t\/\/ update\n      \n        \t\t\t\t\tvar rects = svg.selectAll(\"rect\")\n      \n        \t\t\t\t\t\t.data(data)\n      \n        \t\t\t\t\t\t.attr(\"y\", function(datum) { return height - y(datum.value); })\n      \n        \t\t\t\t\t\t.attr(\"height\", function(datum) { return y(datum.value); });\n      \n        \t\t\t\t\tvar texts = svg.selectAll(\"text.labels\")\n      \n        \t\t\t\t\t\t.data(data)\n      \n        \t\t\t\t\t\t.attr(\"y\", function(datum) { return height - y(datum.value); })\n      \n        \t\t\t\t\t\t.text(function(datum) { return datum.value.toFixed(1);});\n      \n        \n      \n        \t\t\t\t\t\/\/ enter\n      \n        \t\t\t\t\trects.enter().append(\"svg:rect\");\n      \n        \t\t\t\t\ttexts.enter().append(\"svg:text\");\n      \n        \n      \n        \t\t\t\t\t\/\/ exit\n      \n        \t\t\t\t\trects.exit().remove();\n      \n        \t\t\t\t\ttexts.exit().remove();\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\t\/******** stats ********\/\n      \n        \n      \n        \t\t\t\tstats = new Stats();\n      \n        \t\t\t\tstats.domElement.style.position = 'absolute';\n      \n        \t\t\t\tstats.domElement.style.top = '0px';\n      \n        \t\t\t\tdocument.getElementById('container').appendChild( stats.domElement );\n      \n        \n      \n        \t\t\t\t\/\/ update stats on every iteration\n      \n        \t\t\t\tdocument.addEventListener('clmtrackrIteration', function(event) {\n      \n        \t\t\t\t\tstats.update();\n      \n        \t\t\t\t}, false);\n      \n        \n      \n        \t\t\t&lt;\/script&gt;\n      \n        \t\t&lt;\/div&gt;\n      \n        \t&lt;\/body&gt;\n      \n        &lt;\/html&gt;","created_at":"2018-11-07T16:47:27+0000","updated_at":"2018-11-07T16:47:27+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1588066?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":416,"title":"Face tracker","url":"https:\/\/www.auduno.com\/clmtrackr\/examples\/clm_emotiondetection.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:47:27+0000","updated_at":"2018-11-07T16:47:27+0000","annotations":[],"reading_time":0,"domain_name":"www.auduno.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":413,"title":"Wizard of Legend Kickstarter Demo by contingent99","url":"https:\/\/contingent99.itch.io\/wizard-of-legend-kickstarter-demo","content":"<div class=\"left_col column\"><div class=\"formatted_description user_formatted\"><p>For the purposes of better illustrating the game mechanics and lore in Wizard of Legend, we've created a game... about the game just for the Kickstarter Campaign! The game doesn't represent the final product, but will introduce you to the world of Wizard of Legend in a fun and interactive way.<\/p><p>This demo takes place after the timeline of Wizard of Legend and is shown from a more modern perspective where the events of the game are now studied as an important part of history. In the Museum of Lanova's Chaos Trials exhibit, you can experience a few of the spells and challenges that will be present in Wizard of Legend.<\/p><p><b>Disclaimer: <\/b>Though this demo features the gameplay and combat mechanics of Wizard of Legend, it is a completely separate experience from the game. This demo was created from a work in progress and everything is subject to change. That being said, have fun!<\/p><\/div><div class=\"download_instructions formatted_description user_formatted\"><h2 id=\"instructions\">Install instructions<\/h2><p>The demo requires 32MB download to download and 235MB of free space to extract.<\/p><\/div><div class=\"uploads\"><h2 id=\"download\">Download<\/h2><\/div><\/div><div class=\"right_col column\"><div class=\"video_embed\"><iframe frameborder=\"0\" allowfullscreen=\"allowfullscreen\" src=\"https:\/\/www.youtube.com\/embed\/HNv_aC7H8pY\"> <\/iframe><\/div><div class=\"screenshot_list\"><a href=\"https:\/\/img.itch.zone\/aW1hZ2UvNzI2NzYvMzM0MTczLnBuZw==\/original\/9DOEj7.png\" target=\"_blank\" data-image_lightbox=\"true\"><img data-screenshot_id=\"334173\" srcset=\"https:\/\/img.itch.zone\/aW1hZ2UvNzI2NzYvMzM0MTczLnBuZw==\/347x500\/O5BdMZ.png 1x, https:\/\/img.itch.zone\/aW1hZ2UvNzI2NzYvMzM0MTczLnBuZw==\/794x1000\/a59F2s.png 2x\" class=\"screenshot\" src=\"https:\/\/img.itch.zone\/aW1hZ2UvNzI2NzYvMzM0MTczLnBuZw==\/347x500\/O5BdMZ.png\" alt=\"image\" \/><\/a><\/div><\/div>","created_at":"2018-11-07T16:47:26+0000","updated_at":"2018-11-07T16:47:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"contingent99.itch.io","preview_picture":"https:\/\/img.itch.zone\/aW1hZ2UvNzI2NzYvMzM0MTcxLnBuZw==\/original\/R2OgcI.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":414,"title":"auduno\/clmtrackr","url":"https:\/\/github.com\/auduno\/clmtrackr\/blob\/dev\/examples\/clm_emotiondetection.html","content":"&lt;!doctype html&gt;\n      \n        &lt;html lang=\"en\"&gt;\n      \n        \t&lt;head&gt;\n      \n        \t\t&lt;title&gt;Face tracker&lt;\/title&gt;\n      \n        \t\t&lt;meta charset=\"utf-8\"&gt;\n      \n        \t\t&lt;link href=\".\/styles\/bootstrap.min.css\" rel=\"stylesheet\" type=\"text\/css\"&gt;\n      \n        \t\t&lt;style&gt;\n      \n        \t\t\t@import url(https:\/\/fonts.googleapis.com\/css?family=Lato:300italic,700italic,300,700);\n      \n        \n      \n        \t\t\tbody {\n      \n        \t\t\t\tfont-family: 'Lato';\n      \n        \t\t\t\tbackground-color: #f0f0f0;\n      \n        \t\t\t\tmargin: 0px auto;\n      \n        \t\t\t\tmax-width: 1150px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#overlay {\n      \n        \t\t\t\tposition: absolute;\n      \n        \t\t\t\ttop: 0px;\n      \n        \t\t\t\tleft: 0px;\n      \n        \t\t\t\t-o-transform : scaleX(-1);\n      \n        \t\t\t\t-webkit-transform : scaleX(-1);\n      \n        \t\t\t\ttransform : scaleX(-1);\n      \n        \t\t\t\t-ms-filter : fliph; \/*IE*\/\n      \n        \t\t\t\tfilter : fliph; \/*IE*\/\n      \n        \n      \n        \t\t\t\t\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#videoel {\n      \n        \t\t\t\t-o-transform : scaleX(-1);\n      \n        \t\t\t\t-webkit-transform : scaleX(-1);\n      \n        \t\t\t\ttransform : scaleX(-1);\n      \n        \t\t\t\t-ms-filter : fliph; \/*IE*\/\n      \n        \t\t\t\tfilter : fliph; \/*IE*\/\n      \n        \n      \n        \t\t\t\t\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#container {\n      \n        \t\t\t\tposition : relative;\n      \n        \t\t\t\twidth : 370px;\n      \n        \t\t\t\t\/*margin : 0px auto;*\/\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#content {\n      \n        \t\t\t\tmargin-top : 50px;\n      \n        \t\t\t\tmargin-left : auto;\n      \n        \t\t\t\tmargin-right : auto;\n      \n        \t\t\t\tmax-width: 600px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\th2 {\n      \n        \t\t\t\tfont-weight : 400;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t.btn {\n      \n        \t\t\t\tfont-family: 'Lato';\n      \n        \t\t\t\tfont-size: 16px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#controls {\n      \n        \t\t\t\ttext-align : center;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#emotion_container {\n      \n        \t\t\t\twidth: 600px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#emotion_icons {\n      \n        \t\t\t\theight: 50px;\n      \n        \t\t\t\tpadding-left: 40px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t.emotion_icon {\n      \n        \t\t\t\twidth : 40px;\n      \n        \t\t\t\theight : 40px;\n      \n        \t\t\t\tmargin-top: 5px;\n      \n        \t\t\t\t\/*margin-left : 13px;*\/\n      \n        \t\t\t\tmargin-left : 35px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#emotion_chart, #emotion_icons {\n      \n        \t\t\t\tmargin: 0 auto;\n      \n        \t\t\t\twidth : 400px;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t#icon1, #icon2, #icon3, #icon4, #icon5, #icon6 {\n      \n        \t\t\t\tvisibility : hidden;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t\t\/* d3 *\/\n      \n        \t\t\t.bar {\n      \n        \t\t\t\tfill : steelblue;\n      \n        \t\t\t\tfill-opacity : .9;\n      \n        \t\t\t}\n      \n        \n      \n        \t\t&lt;\/style&gt;\n      \n        \t\t&lt;script&gt;\n      \n        \t\t\t\/\/ getUserMedia only works over https in Chrome 47+, so we redirect to https. Also notify user if running from file.\n      \n        \t\t\tif (window.location.protocol == \"file:\") {\n      \n        \t\t\t\talert(\"You seem to be running this example directly from a file. Note that these examples only work when served from a server or localhost due to canvas cross-domain restrictions.\");\n      \n        \t\t\t} else if (window.location.hostname !== \"localhost\" &amp;&amp; window.location.protocol !== \"https:\"){\n      \n        \t\t\t\twindow.location.protocol = \"https\";\n      \n        \t\t\t}\n      \n        \t\t&lt;\/script&gt;\n      \n        \t\t&lt;script type=\"text\/javascript\"&gt;\n      \n        \n      \n        \t\t\tvar _gaq = _gaq || [];\n      \n        \t\t\t_gaq.push(['_setAccount', 'UA-32642923-1']);\n      \n        \t\t\t_gaq.push(['_trackPageview']);\n      \n        \n      \n        \t\t\t(function() {\n      \n        \t\t\t\tvar ga = document.createElement('script'); ga.type = 'text\/javascript'; ga.async = true;\n      \n        \t\t\t\tga.src = ('https:' == document.location.protocol ? 'https:\/\/ssl' : 'http:\/\/www') + '.google-analytics.com\/ga.js';\n      \n        \t\t\t\tvar s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);\n      \n        \t\t\t})();\n      \n        \n      \n        \t\t&lt;\/script&gt;\n      \n        \t&lt;\/head&gt;\n      \n        \t&lt;body&gt;\n      \n        \t\t&lt;script src=\".\/js\/libs\/utils.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\"..\/build\/clmtrackr.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\"..\/models\/model_pca_20_svm.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/libs\/Stats.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/libs\/d3.min.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/emotion_classifier.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;script src=\".\/js\/emotionmodel.js\"&gt;&lt;\/script&gt;\n      \n        \t\t&lt;div id=\"content\"&gt;\n      \n        \t\t\t&lt;h2&gt;Emotion detection example&lt;\/h2&gt;\n      \n        \t\t\t&lt;p&gt;This a test of emotion detection based on parameter output from &lt;em&gt;clmtrackr&lt;\/em&gt;.&lt;\/p&gt;\n      \n        \t\t\t&lt;div id=\"container\"&gt;\n      \n        \t\t\t\t&lt;video id=\"videoel\" width=\"400\" height=\"300\" preload=\"auto\" loop playsinline autoplay&gt;\n      \n        \t\t\t\t&lt;\/video&gt;\n      \n        \t\t\t\t&lt;canvas id=\"overlay\" width=\"400\" height=\"300\"&gt;&lt;\/canvas&gt;\n      \n        \t\t\t&lt;\/div&gt;\n      \n        \t\t\t&lt;div id=\"emotion_container\"&gt;\n      \n        \t\t\t\t&lt;div id=\"emotion_icons\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon1\" src=\".\/media\/icon_angry.png\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon2\" src=\".\/media\/icon_sad.png\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon3\" src=\".\/media\/icon_surprised.png\"&gt;\n      \n        \t\t\t\t\t&lt;img class=\"emotion_icon\" id=\"icon4\" src=\".\/media\/icon_happy.png\"&gt;\n      \n        \t\t\t\t&lt;\/div&gt;\n      \n        \t\t\t\t&lt;div id='emotion_chart'&gt;&lt;\/div&gt;\n      \n        \t\t\t&lt;\/div&gt;\n      \n        \t\t\t&lt;div id=\"controls\"&gt;\n      \n        \t\t\t\t&lt;input class=\"btn\" type=\"button\" value=\"wait, loading video\" disabled=\"disabled\" onclick=\"startVideo()\" id=\"startbutton\"&gt;&lt;\/input&gt;\n      \n        \t\t\t&lt;\/div&gt;\n      \n        \t\t\t&lt;a href=\"https:\/\/github.com\/auduno\/clmtrackr\"&gt;&lt;img style=\"position: absolute; top: 0; left: 0; border: 0;\" src=\"https:\/\/s3.amazonaws.com\/github\/ribbons\/forkme_left_green_007200.png\" alt=\"Fork me on GitHub\"&gt;&lt;\/a&gt;\n      \n        \t\t\t&lt;script&gt;\n      \n        \t\t\t\tvar vid = document.getElementById('videoel');\n      \n        \t\t\t\tvar vid_width = vid.width;\n      \n        \t\t\t\tvar vid_height = vid.height;\n      \n        \t\t\t\tvar overlay = document.getElementById('overlay');\n      \n        \t\t\t\tvar overlayCC = overlay.getContext('2d');\n      \n        \n      \n        \t\t\t\t\/********** check and set up video\/webcam **********\/\n      \n        \n      \n        \t\t\t\tfunction enablestart() {\n      \n        \t\t\t\t\tvar startbutton = document.getElementById('startbutton');\n      \n        \t\t\t\t\tstartbutton.value = \"start\";\n      \n        \t\t\t\t\tstartbutton.disabled = null;\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction adjustVideoProportions() {\n      \n        \t\t\t\t\t\/\/ resize overlay and video if proportions are different\n      \n        \t\t\t\t\t\/\/ keep same height, just change width\n      \n        \t\t\t\t\tvar proportion = vid.videoWidth\/vid.videoHeight;\n      \n        \t\t\t\t\tvid_width = Math.round(vid_height * proportion);\n      \n        \t\t\t\t\tvid.width = vid_width;\n      \n        \t\t\t\t\toverlay.width = vid_width;\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction gumSuccess( stream ) {\n      \n        \t\t\t\t\t\/\/ add camera stream if getUserMedia succeeded\n      \n        \t\t\t\t\tif (\"srcObject\" in vid) {\n      \n        \t\t\t\t\t\tvid.srcObject = stream;\n      \n        \t\t\t\t\t} else {\n      \n        \t\t\t\t\t\tvid.src = (window.URL &amp;&amp; window.URL.createObjectURL(stream));\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t\tvid.onloadedmetadata = function() {\n      \n        \t\t\t\t\t\tadjustVideoProportions();\n      \n        \t\t\t\t\t\tvid.play();\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t\tvid.onresize = function() {\n      \n        \t\t\t\t\t\tadjustVideoProportions();\n      \n        \t\t\t\t\t\tif (trackingStarted) {\n      \n        \t\t\t\t\t\t\tctrack.stop();\n      \n        \t\t\t\t\t\t\tctrack.reset();\n      \n        \t\t\t\t\t\t\tctrack.start(vid);\n      \n        \t\t\t\t\t\t}\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction gumFail() {\n      \n        \t\t\t\t\talert(\"There was some problem trying to fetch video from your webcam. If you have a webcam, please make sure to accept when the browser asks for access to your webcam.\");\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tnavigator.getUserMedia = navigator.getUserMedia || navigator.webkitGetUserMedia || navigator.mozGetUserMedia || navigator.msGetUserMedia;\n      \n        \t\t\t\twindow.URL = window.URL || window.webkitURL || window.msURL || window.mozURL;\n      \n        \n      \n        \t\t\t\t\/\/ check for camerasupport\n      \n        \t\t\t\tif (navigator.mediaDevices) {\n      \n        \t\t\t\t\tnavigator.mediaDevices.getUserMedia({video : true}).then(gumSuccess).catch(gumFail);\n      \n        \t\t\t\t} else if (navigator.getUserMedia) {\n      \n        \t\t\t\t\tnavigator.getUserMedia({video : true}, gumSuccess, gumFail);\n      \n        \t\t\t\t} else {\n      \n        \t\t\t\t\talert(\"This demo depends on getUserMedia, which your browser does not seem to support. :(\");\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tvid.addEventListener('canplay', enablestart, false);\n      \n        \n      \n        \t\t\t\t\/*********** setup of emotion detection *************\/\n      \n        \n      \n        \t\t\t\t\/\/ set eigenvector 9 and 11 to not be regularized. This is to better detect motion of the eyebrows\n      \n        \t\t\t\tpModel.shapeModel.nonRegularizedVectors.push(9);\n      \n        \t\t\t\tpModel.shapeModel.nonRegularizedVectors.push(11);\n      \n        \n      \n        \t\t\t\tvar ctrack = new clm.tracker({useWebGL : true});\n      \n        \t\t\t\tctrack.init(pModel);\n      \n        \t\t\t\tvar trackingStarted = false;\n      \n        \n      \n        \t\t\t\tfunction startVideo() {\n      \n        \t\t\t\t\t\/\/ start video\n      \n        \t\t\t\t\tvid.play();\n      \n        \t\t\t\t\t\/\/ start tracking\n      \n        \t\t\t\t\tctrack.start(vid);\n      \n        \t\t\t\t\ttrackingStarted = true;\n      \n        \t\t\t\t\t\/\/ start loop to draw face\n      \n        \t\t\t\t\tdrawLoop();\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tfunction drawLoop() {\n      \n        \t\t\t\t\trequestAnimFrame(drawLoop);\n      \n        \t\t\t\t\toverlayCC.clearRect(0, 0, vid_width, vid_height);\n      \n        \t\t\t\t\t\/\/psrElement.innerHTML = \"score :\" + ctrack.getScore().toFixed(4);\n      \n        \t\t\t\t\tif (ctrack.getCurrentPosition()) {\n      \n        \t\t\t\t\t\tctrack.draw(overlay);\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t\tvar cp = ctrack.getCurrentParameters();\n      \n        \n      \n        \t\t\t\t\tvar er = ec.meanPredict(cp);\n      \n        \t\t\t\t\tif (er) {\n      \n        \t\t\t\t\t\tupdateData(er);\n      \n        \t\t\t\t\t\tfor (var i = 0;i &lt; er.length;i++) {\n      \n        \t\t\t\t\t\t\tif (er[i].value &gt; 0.4) {\n      \n        \t\t\t\t\t\t\t\tdocument.getElementById('icon'+(i+1)).style.visibility = 'visible';\n      \n        \t\t\t\t\t\t\t} else {\n      \n        \t\t\t\t\t\t\t\tdocument.getElementById('icon'+(i+1)).style.visibility = 'hidden';\n      \n        \t\t\t\t\t\t\t}\n      \n        \t\t\t\t\t\t}\n      \n        \t\t\t\t\t}\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\tdelete emotionModel['disgusted'];\n      \n        \t\t\t\tdelete emotionModel['fear'];\n      \n        \t\t\t\tvar ec = new emotionClassifier();\n      \n        \t\t\t\tec.init(emotionModel);\n      \n        \t\t\t\tvar emotionData = ec.getBlank();\n      \n        \n      \n        \t\t\t\t\/************ d3 code for barchart *****************\/\n      \n        \n      \n        \t\t\t\tvar margin = {top : 20, right : 20, bottom : 10, left : 40},\n      \n        \t\t\t\t\twidth = 400 - margin.left - margin.right,\n      \n        \t\t\t\t\theight = 100 - margin.top - margin.bottom;\n      \n        \n      \n        \t\t\t\tvar barWidth = 30;\n      \n        \n      \n        \t\t\t\tvar formatPercent = d3.format(\".0%\");\n      \n        \n      \n        \t\t\t\tvar x = d3.scale.linear()\n      \n        \t\t\t\t\t.domain([0, ec.getEmotions().length]).range([margin.left, width+margin.left]);\n      \n        \n      \n        \t\t\t\tvar y = d3.scale.linear()\n      \n        \t\t\t\t\t.domain([0,1]).range([0, height]);\n      \n        \n      \n        \t\t\t\tvar svg = d3.select(\"#emotion_chart\").append(\"svg\")\n      \n        \t\t\t\t\t.attr(\"width\", width + margin.left + margin.right)\n      \n        \t\t\t\t\t.attr(\"height\", height + margin.top + margin.bottom)\n      \n        \n      \n        \t\t\t\tsvg.selectAll(\"rect\").\n      \n        \t\t\t\t\tdata(emotionData).\n      \n        \t\t\t\t\tenter().\n      \n        \t\t\t\t\tappend(\"svg:rect\").\n      \n        \t\t\t\t\tattr(\"x\", function(datum, index) { return x(index); }).\n      \n        \t\t\t\t\tattr(\"y\", function(datum) { return height - y(datum.value); }).\n      \n        \t\t\t\t\tattr(\"height\", function(datum) { return y(datum.value); }).\n      \n        \t\t\t\t\tattr(\"width\", barWidth).\n      \n        \t\t\t\t\tattr(\"fill\", \"#2d578b\");\n      \n        \n      \n        \t\t\t\tsvg.selectAll(\"text.labels\").\n      \n        \t\t\t\t\tdata(emotionData).\n      \n        \t\t\t\t\tenter().\n      \n        \t\t\t\t\tappend(\"svg:text\").\n      \n        \t\t\t\t\tattr(\"x\", function(datum, index) { return x(index) + barWidth; }).\n      \n        \t\t\t\t\tattr(\"y\", function(datum) { return height - y(datum.value); }).\n      \n        \t\t\t\t\tattr(\"dx\", -barWidth\/2).\n      \n        \t\t\t\t\tattr(\"dy\", \"1.2em\").\n      \n        \t\t\t\t\tattr(\"text-anchor\", \"middle\").\n      \n        \t\t\t\t\ttext(function(datum) { return datum.value;}).\n      \n        \t\t\t\t\tattr(\"fill\", \"white\").\n      \n        \t\t\t\t\tattr(\"class\", \"labels\");\n      \n        \n      \n        \t\t\t\tsvg.selectAll(\"text.yAxis\").\n      \n        \t\t\t\t\tdata(emotionData).\n      \n        \t\t\t\t\tenter().append(\"svg:text\").\n      \n        \t\t\t\t\tattr(\"x\", function(datum, index) { return x(index) + barWidth; }).\n      \n        \t\t\t\t\tattr(\"y\", height).\n      \n        \t\t\t\t\tattr(\"dx\", -barWidth\/2).\n      \n        \t\t\t\t\tattr(\"text-anchor\", \"middle\").\n      \n        \t\t\t\t\tattr(\"style\", \"font-size: 12\").\n      \n        \t\t\t\t\ttext(function(datum) { return datum.emotion;}).\n      \n        \t\t\t\t\tattr(\"transform\", \"translate(0, 18)\").\n      \n        \t\t\t\t\tattr(\"class\", \"yAxis\");\n      \n        \n      \n        \t\t\t\tfunction updateData(data) {\n      \n        \t\t\t\t\t\/\/ update\n      \n        \t\t\t\t\tvar rects = svg.selectAll(\"rect\")\n      \n        \t\t\t\t\t\t.data(data)\n      \n        \t\t\t\t\t\t.attr(\"y\", function(datum) { return height - y(datum.value); })\n      \n        \t\t\t\t\t\t.attr(\"height\", function(datum) { return y(datum.value); });\n      \n        \t\t\t\t\tvar texts = svg.selectAll(\"text.labels\")\n      \n        \t\t\t\t\t\t.data(data)\n      \n        \t\t\t\t\t\t.attr(\"y\", function(datum) { return height - y(datum.value); })\n      \n        \t\t\t\t\t\t.text(function(datum) { return datum.value.toFixed(1);});\n      \n        \n      \n        \t\t\t\t\t\/\/ enter\n      \n        \t\t\t\t\trects.enter().append(\"svg:rect\");\n      \n        \t\t\t\t\ttexts.enter().append(\"svg:text\");\n      \n        \n      \n        \t\t\t\t\t\/\/ exit\n      \n        \t\t\t\t\trects.exit().remove();\n      \n        \t\t\t\t\ttexts.exit().remove();\n      \n        \t\t\t\t}\n      \n        \n      \n        \t\t\t\t\/******** stats ********\/\n      \n        \n      \n        \t\t\t\tstats = new Stats();\n      \n        \t\t\t\tstats.domElement.style.position = 'absolute';\n      \n        \t\t\t\tstats.domElement.style.top = '0px';\n      \n        \t\t\t\tdocument.getElementById('container').appendChild( stats.domElement );\n      \n        \n      \n        \t\t\t\t\/\/ update stats on every iteration\n      \n        \t\t\t\tdocument.addEventListener('clmtrackrIteration', function(event) {\n      \n        \t\t\t\t\tstats.update();\n      \n        \t\t\t\t}, false);\n      \n        \n      \n        \t\t\t&lt;\/script&gt;\n      \n        \t\t&lt;\/div&gt;\n      \n        \t&lt;\/body&gt;\n      \n        &lt;\/html&gt;","created_at":"2018-11-07T16:47:26+0000","updated_at":"2018-11-07T16:47:26+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1588066?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":410,"title":"Luciano Fiandesio - VIM configuration for happy Java coding","url":"http:\/\/www.lucianofiandesio.com\/vim-configuration-for-happy-java-coding","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:47:25+0000","updated_at":"2018-11-07T16:47:25+0000","annotations":[],"reading_time":0,"domain_name":"www.lucianofiandesio.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":411,"title":"clojure_docs.md","url":"https:\/\/gist.github.com\/oakes\/8db57ac808bf6ec144d627fd83a89da3","content":"<h2><a id=\"user-content-how-clojures-documentation-can-leapfrog-other-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-clojures-documentation-can-leapfrog-other-languages\"><\/a>How Clojure's documentation can leapfrog other languages<\/h2>\n<h3><a id=\"user-content-summary\" class=\"anchor\" aria-hidden=\"true\" href=\"#summary\"><\/a>Summary<\/h3>\n<p>I made a documentation generator that cashes in on Clojure's dynamism. See <a href=\"https:\/\/oakes.github.io\/play-cljs\/\" rel=\"nofollow\">the play-cljs docs<\/a> (a ClojureScript game library) for an example of its output.<\/p>\n<h3><a id=\"user-content-the-problem\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-problem\"><\/a>The Problem<\/h3>\n<p>Like many of you, I've often wondered what my final regret will be on my deathbed. My best guess came to me in a dream recently. I was walking across the charred earth of an apocalyptic future world, maneuvering around the remains of the less fortunate. I was startled to find a young girl, barely holding onto her life. She murmured something to me. I asked her to repeat it, and she said more loudly: \"I...wish your Clojure projects didn't have such crappy documentation.\"<\/p>\n<p>If you take criticism the way I do, you probably had the same first thought: Let the girl die. But in truth she had a valid point. Dynamic languages have a bad reputation when it comes to documentation. On the surface, it makes sense: Documentation generators work by statically analyzing source code, and since static languages have more information <em>available<\/em> statically (duh), their docs will tend to be better. Perhaps the only winning move for dynamic languages is <a href=\"https:\/\/www.youtube.com\/watch?v=6DGNZnfKYnU\" rel=\"nofollow\">not to play<\/a>.<\/p>\n<h3><a id=\"user-content-static-vs-dynamic-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#static-vs-dynamic-documentation\"><\/a>Static vs. Dynamic Documentation<\/h3>\n<p>Try to forget this notion that documentation is primarily a static thing that library authors put on the web for users and search engines to passively consume. Instead, imagine if end users created docs on-the-fly for every dependency in their project, using their doc generator as a real-time developer tool. That is the idea behind my new project called <a href=\"https:\/\/github.com\/oakes\/Dynadoc\">Dynadoc<\/a>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/0045b7cbd9a29b9276a2b18efd8f4ddc5067d5f8\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f636f72652e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/0045b7cbd9a29b9276a2b18efd8f4ddc5067d5f8\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f636f72652e706e67\" alt=\"clojure.core\" data-canonical-src=\"https:\/\/gist.github.com\/oakes\/8db57ac808bf6ec144d627fd83a89da3\/raw\/84da0d8584ab55b9af16a3151deeb690ac94aad1\/core.png\" \/><\/a><\/p>\n<p>Dynadoc runs <em>inside<\/em> your Clojure(Script) project, spinning up a server on localhost to show its UI. It finds all the namespaces and generates searchable docs instantly. This is possible because in Clojure, everything is available at runtime, including docstrings and source code. You can even define a new function in your project, eval it, refresh Dynadoc and see it there.<\/p>\n<h3><a id=\"user-content-interactive-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#interactive-examples\"><\/a>Interactive Examples<\/h3>\n<p>Since we're running inside your project, we can do some pretty crazy stuff. Dynadoc allows you to define code examples for anything, and it will display them in the correct page in a fully interactive way. Here's the venerable <code>conj<\/code> function with interactive examples:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/e2af513fef37c9efee93e30c013faff9a4eb763f\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f64656d6f2d636c6a2e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/e2af513fef37c9efee93e30c013faff9a4eb763f\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f64656d6f2d636c6a2e676966\" alt=\"demo-clj\" data-canonical-src=\"https:\/\/gist.github.com\/oakes\/8db57ac808bf6ec144d627fd83a89da3\/raw\/84da0d8584ab55b9af16a3151deeb690ac94aad1\/demo-clj.gif\" \/><\/a><\/p>\n<p>Some people <a href=\"https:\/\/youtu.be\/nrpsMB2gYI0?t=14m30s\" rel=\"nofollow\">get fixated on docstrings<\/a> as the source of our documentation woes, but I disagree completely. You could write an entire novel about <code>conj<\/code> and it still won't be as good as a few interactive examples. And for non-English speakers, interactive examples beat docstrings by a long shot.<\/p>\n<p>To define these examples, you just need to use a <a href=\"https:\/\/github.com\/oakes\/defexample\">special macro<\/a> I made. The above examples were created thusly:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defexamples clojure.core\/conj\n  [\"Add a name to a vector\"\n   (conj [\"Alice\" \"Bob\"] \"Charlie\")]\n  [\"Add a number to a list\"\n   (conj '(2 3) 1)]\n  [\"Add a key-val pair to a hash map\"\n   (conj {:name \"Alice\"} [:age 30])])<\/pre><\/div>\n<p>The best part is, you don't need any buy-in from the core team or library authors. We can start writing examples <em>now<\/em>, for anything in Clojure, and they will be usable for any tool that wants to do something with them.<\/p>\n<p>ClojureScript examples are even wilder. They can optionally ask for a DOM element, so the examples can do something visual:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/2e79b12f60fd3197c0a8e807a0e9b6483d56b1b7\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f64656d6f2d636c6a732e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/2e79b12f60fd3197c0a8e807a0e9b6483d56b1b7\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f64656d6f2d636c6a732e676966\" alt=\"demo-cljs\" data-canonical-src=\"https:\/\/gist.github.com\/oakes\/8db57ac808bf6ec144d627fd83a89da3\/raw\/84da0d8584ab55b9af16a3151deeb690ac94aad1\/demo-cljs.gif\" \/><\/a><\/p>\n<p>That's a <a href=\"https:\/\/github.com\/reagent-project\/reagent\">Reagent<\/a> component, with a completely interactive example inside its own documentation page! You can <a href=\"https:\/\/oakes.github.io\/dynadoc-examples\/\" rel=\"nofollow\">try it live<\/a>. Here's the code:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn clicks\n  \"Shows the number of times the user clicked the button.\"\n  [button-text]\n  (let [state (r\/atom {:clicks 0})]\n    (fn []\n      [:div\n       [:p \"You clicked \" (:clicks @state) \" times\"]\n       [:button {:on-click (fn []\n                             (swap! state update :clicks inc))}\n        button-text]])))\n\n(defexample clicks\n  {:with-card card\n   :with-focus [focus [clicks \"Click me!\"]]}\n  (reagent.core\/unmount-component-at-node card)\n  (reagent.core\/render-component focus card)\n  nil)<\/pre><\/div>\n<p>This example is a bit weirder. The <code>:with-card<\/code> option takes a symbol that will be associated with a DOM element in your example. The <code>:with-focus<\/code> option takes a binding form where you can define a small chunk of code that you want to be exclusively visible in the example. That way, the distracting setup code isn't viewable or editable. In the body of the example, we make sure to unmount any previous component and mount a new one, <code>focus<\/code>, on the DOM element, <code>card<\/code>.<\/p>\n<p>Through the magic of Clojure macros, the code is rewritten and evaluated in your browser. Every time the user edits the code in the browser-based editor, the example is re-executed, so the component is unmounted\/remounted with the new data. The great thing about this is that it works with absolutely anything. It gives you a DOM element and you take care of the rest. Here's an example I made for my game library, <a href=\"https:\/\/github.com\/oakes\/play-cljs\">play-cljs<\/a>:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/4bc6b03e4e80d03bb3187fa6d880c1723c922ed3\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f64656d6f2d636c6a73322e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/4bc6b03e4e80d03bb3187fa6d880c1723c922ed3\/68747470733a2f2f676973742e6769746875622e636f6d2f6f616b65732f38646235376163383038626636656331343464363237666438336138396461332f7261772f383464613064383538346162353562396166313661333135316465656236393061633934616164312f64656d6f2d636c6a73322e676966\" alt=\"demo-cljs2\" data-canonical-src=\"https:\/\/gist.github.com\/oakes\/8db57ac808bf6ec144d627fd83a89da3\/raw\/84da0d8584ab55b9af16a3151deeb690ac94aad1\/demo-cljs2.gif\" \/><\/a><\/p>\n<p>Check out the full <a href=\"https:\/\/oakes.github.io\/play-cljs\/\" rel=\"nofollow\">play-cljs dynadocs<\/a>. The fact that they are being served on Github Pages reflects the last feature...good ol' fashioned static HTML export. Dynadoc supports it via the \"Export\" link. The Clojure examples won't be interactive (they need a JVM to be present), but the ClojureScript examples can be - because it can compile itself in the browser. <em>Drops mic.<\/em><\/p>\n<h3><a id=\"user-content-conclusion\" class=\"anchor\" aria-hidden=\"true\" href=\"#conclusion\"><\/a>Conclusion<\/h3>\n<p>Clojure is dope, and if you use it your life will be dope too.<\/p>","created_at":"2018-11-07T16:47:25+0000","updated_at":"2018-11-07T16:47:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"gist.github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/216147?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":412,"title":"LWJGL\/lwjgl3-wiki","url":"https:\/\/github.com\/LWJGL\/lwjgl3-wiki\/wiki\/1.3.-Memory-FAQ","content":"<blockquote>\n<p>This page is a must read for all LWJGL users.<\/p>\n<\/blockquote>\n<h2>\n<a id=\"user-content-why-does-lwjgl-use-buffers-so-much\" class=\"anchor\" href=\"#why-does-lwjgl-use-buffers-so-much\" aria-hidden=\"true\"><\/a>Why does LWJGL use buffers so much?<\/h2>\n<p>LWJGL requires the use of off-heap memory when passing data to native libraries. Likewise, any\nbuffers returned from native libraries are always backed by off-heap memory. This is not an LWJGL\nlimitation. There are two issues with Java objects and arrays that live on the JVM heap:<\/p>\n<ul><li>It is not possible to control the layout of Java objects. Different JVMs and different JVM\nsettings produce very different field layouts. Native libraries on the other hand expect data with\nvery precisely defined layouts.<\/li>\n<li>Any Java object or array may be moved by the GC at any time, concurrently with the execution of\nof a native method call. All JNI methods are executed at a safepoint so, by definition, must not\naccess heap data.<\/li>\n<\/ul><p>The standard approach is:<\/p>\n<ol><li>Using JNI functions to access Java objects, which is painfully slow.<\/li>\n<li>Using JNI functions to \"pin\" Java arrays (<code>Get\/ReleasePrimitiveArrayCritical<\/code> or Hotspot\nCritical Natives) which is also inefficient for several reasons.<\/li>\n<\/ol><p>LWJGL on the other hand is designed to be used with direct (off-heap) <code>java.nio<\/code> buffer classes\nfor passing data to and from native code. <code>ByteBuffer<\/code> and the other classes is not the best\npossible abstraction for off-heap data and their API is not ideal, but it is the only officially\nsupported way to access off-heap data in Java.<\/p>\n<p>The easiest way to think of <code>ByteBuffer<\/code> is as a wrapper over a native C pointer, plus the array\nlength (the <code>buffer.capacity()<\/code>). LWJGL maps C primitive types to the corresponding class in\n<code>java.nio<\/code>. Arrays of pointers are mapped to the <code>org.lwjgl.PointerBuffer<\/code> class. Pointers to\nstructs are mapped to the corresponding struct class. Pointers to struct arrays are mapped to the\ncorresponding <code>&lt;StructClass&gt;.Buffer<\/code> class. <code>PointerBuffer<\/code> and the struct <code>Buffer<\/code> classes have\nan API very similar to <code>java.nio<\/code> buffers.<\/p>\n<h2>\n<a id=\"user-content-what-javaniobyteorder-should-be-used\" class=\"anchor\" href=\"#what-javaniobyteorder-should-be-used\" aria-hidden=\"true\"><\/a>What <code>java.nio.ByteOrder<\/code> should be used?<\/h2>\n<p>The buffer byte order must be set to <code>ByteOrder.nativeOrder()<\/code>. It is basically required for\ncorrect closs-platform behavior. It also results in the best performance.<\/p>\n<p>All buffer instances created by LWJGL are always set to the <em>native byte order<\/em>.<\/p>\n<h2>\n<a id=\"user-content-how-do-i-allocate-and-deallocate-buffers\" class=\"anchor\" href=\"#how-do-i-allocate-and-deallocate-buffers\" aria-hidden=\"true\"><\/a>How do I allocate and deallocate buffers?<\/h2>\n<p>After getting familiar with the above mappings, the next step is learning how to handle allocation\nof such buffers. This is a critical issue and LWJGL offers several options. The options are listed\nbelow ordered from more-to-less efficient. Every time you make a decision on how to handle an\nallocation, you should consider the first option. If that's not applicable, consider the second\noption, and so on.<\/p>\n<h3>\n<a id=\"user-content-1-stack-allocation\" class=\"anchor\" href=\"#1-stack-allocation\" aria-hidden=\"true\"><\/a>1. Stack allocation<\/h3>\n<p>Java does not support explicit stack allocation of Java objects and obviously does not support\noff-heap stack allocation either. In C it's very simple: you declare a variable inside a function\nand it's stack allocated. When the function returns, the variable's memory is reclaimed\nautomatically (and without overhead). There's no such equivalent in Java.<\/p>\n<blockquote>\n<p>Similarly, it is not possible in Java to call a native function that expects or returns a\n<strong>struct by value<\/strong>. Such functions in LWJGL bindings are wrapped and exposed with a\n<strong>pointer-to-struct<\/strong> parameter or return value.<\/p>\n<\/blockquote>\n<p>This is a problem because it's very common to need small, short-lived allocations when calling\nnative functions. For example, creating a vertex buffer object in OpenGL in C:<\/p>\n<div class=\"highlight highlight-source-c\"><pre>GLuint vbo;\nglGenBuffers(1, &amp;vbo); \/\/ very simple<\/pre><\/div>\n<p>and with LWJGL:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>IntBuffer ip = ...; \/\/ need a 4-byte buffer here\nglGenBuffers(ip);\nint vbo = ip.get(0);<\/pre><\/div>\n<p>A real <code>IntBuffer<\/code> allocation in the above example, regardless of the implementation, would be\nvastly more inefficient than the stack pointer in the equivalent C code.<\/p>\n<p>The usual answer to this problem, in LWJGL 2 and other Java libraries, is to allocate the buffer\nonce, cache it and reuse it in many method calls. This is an incredibly unsatisfying solution:<\/p>\n<ul><li>It leads to ugly code and wastes memory.<\/li>\n<li>To avoid wasting memory, static buffers are usually used.<\/li>\n<li>Using static buffers leads to either concurrency bugs or less than ideal performance (due to\nsynchronization).<\/li>\n<\/ul><p>The LWJGL 3 answer is the <code>org.lwjgl.system.MemoryStack<\/code> API. It's been designed to be used with\nstatic imports and try-with-resources blocks. The above example becomes:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>int vbo;\ntry (MemoryStack stack = stackPush()) {\n    IntBuffer ip = stack.callocInt(1);\n    glGenBuffers(ip);\n    vbo = ip.get(0);\n} \/\/ stack automatically popped, ip memory automatically reclaimed<\/pre><\/div>\n<p>It is obviously more verbose, but has the following advantages:<\/p>\n<ul><li>More than one allocation is usually required, but the try-with-resources boilerplate remains the\nsame.<\/li>\n<li>The semantics of the above code perfectly match the requirements. The stack memory is\nthread-local, just like a real C thread stack.<\/li>\n<li>Performance is ideal. The stack push and pop are simple bumps of a pointer and the <code>IntBuffer<\/code>\ninstance allocation is either eliminated with escape analysis or handled by the next minor\/eden GC\ncycle (super efficiently).<\/li>\n<\/ul><blockquote>\n<p>Note 1: The default stack size is <code>64kb<\/code>. It can be changed with <code>-Dorg.lwjgl.system.stackSize<\/code> or\n<code>Configuration.STACK_SIZE<\/code>.<\/p>\n<\/blockquote>\n<blockquote>\n<p>Note 2: Structs and struct buffers can also be allocated on the <code>MemoryStack<\/code>.<\/p>\n<\/blockquote>\n<blockquote>\n<p>Note 3: The static, thread-local <code>MemoryStack<\/code> API is just a convenience. There's additional API that\nlets you create and\/or use <code>MemoryStack<\/code> instances as you see fit.<\/p>\n<\/blockquote>\n<h3>\n<a id=\"user-content-2-memoryutil-mallocfree\" class=\"anchor\" href=\"#2-memoryutil-mallocfree\" aria-hidden=\"true\"><\/a>2. MemoryUtil (malloc\/free)<\/h3>\n<p>Sometimes stack allocation cannot be used. The memory that must be allocated is too big or the\nallocation is long lived. In such cases, the next best option is explicit memory management.\nEither via the <code>org.lwjgl.system.MemoryUtil<\/code> API or a specific memory allocator (currently\navailable in LWJGL: <code>stdlib<\/code>, <code>jemalloc<\/code>). Example:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>ByteBuffer buffer = memAlloc(2 * 1024 * 1024); \/\/ 2MB\n\/\/ use buffer...\nmemFree(buffer); \/\/ free when no longer needed<\/pre><\/div>\n<blockquote>\n<p>Note 1: Just like in C, the user is responsible for deallocating memory allocated with <code>malloc<\/code>\nusing <code>free<\/code>.<\/p>\n<\/blockquote>\n<blockquote>\n<p>Note 2: API for the standard functions <code>calloc<\/code>, <code>realloc<\/code> and <code>aligned_alloc<\/code> is also\navailable.<\/p>\n<\/blockquote>\n<blockquote>\n<p>Note 3: The Java objects allocated with the explicit memory management functions are also\nsubject to escape analysis.<\/p>\n<\/blockquote>\n<h3>\n<a id=\"user-content-3-bufferutils-bytebufferallocatedirect\" class=\"anchor\" href=\"#3-bufferutils-bytebufferallocatedirect\" aria-hidden=\"true\"><\/a>3. BufferUtils (ByteBuffer.allocateDirect)<\/h3>\n<p>Sometimes the explicit memory management API cannot be used either. Maybe a particular allocation\nis hard to track without complicating the code, or it might not be possible to know exactly when\nit is no longer required. Such cases are legitimate candidates for using <code>org.lwjgl.BufferUtils<\/code>.\nThis class existed in older LWJGL versions with the same API. It uses <code>ByteBuffer.allocateDirect<\/code>\nto do the allocations which has one major advantage: the user does not need to deallocate the\noff-heap memory explicitly, it is done automatically by the GC.<\/p>\n<p>On the other hand, it has the following disadvantages:<\/p>\n<ul><li>It is slow, much slower than the raw <code>malloc<\/code> call. A lot of overhead on top of a function that\nis already slow.<\/li>\n<li>It scales badly with concurrent allocations.<\/li>\n<li>It arbitrarily limits the amount of allocated memory (<code>-XX:MaxDirectMemorySize<\/code>).<\/li>\n<li>Like Java arrays, the allocated memory is always zeroed-out. This is not necessarily bad, but\nhaving the option would be better.<\/li>\n<li>There's no way to deallocate the allocated memory on demand (without JDK-specific reflection\nhacks). Instead, a reference queue is used that usually requires two GC cycles to free the native\nmemory. This may lead to OOM errors under pressure.<\/li>\n<\/ul><blockquote>\n<p>An example of LWJGL using <code>BufferUtils<\/code> internally is for allocating the memory that backs the\nthread-local <code>MemoryStack<\/code> instances. It is a long lived allocation that must be deallocated\nwhen the thread dies, so we let the GC take care of it.<\/p>\n<\/blockquote>\n<h3>\n<a id=\"user-content-tldr\" class=\"anchor\" href=\"#tldr\" aria-hidden=\"true\"><\/a>tl;dr<\/h3>\n<ol><li>Use <code>org.lwjgl.system.MemoryStack<\/code> and if not possible...<\/li>\n<li>Use <code>org.lwjgl.system.MemoryUtil<\/code> and if not possible...<\/li>\n<li>Use <code>org.lwjgl.BufferUtil<\/code>\n<\/li>\n<\/ol><h2>\n<a id=\"user-content-i-want-to-know-more-got-something-for-me\" class=\"anchor\" href=\"#i-want-to-know-more-got-something-for-me\" aria-hidden=\"true\"><\/a>I want to know more, got something for me?<\/h2>\n<p>Yes, read the <a href=\"https:\/\/blog.lwjgl.org\/memory-management-in-lwjgl-3\/\" rel=\"nofollow\">Memory Management in LWJGL 3<\/a> blog post.<\/p>","created_at":"2018-11-07T16:47:25+0000","updated_at":"2018-11-07T16:47:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/2757344?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":407,"title":"A simple DIY Heroku replacement to keep your hosting costs down","url":"https:\/\/news.ycombinator.com\/item?id=15892617","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D15892617\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"15892617\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/github.com\/githubsaturn\/captainduckduck\" class=\"storylink\">A simple DIY Heroku replacement to keep your hosting costs down<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=github.com\">github.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        209 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=pritambarhate\" class=\"hnuser\">pritambarhate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15892617\">11 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=15892617&amp;goto=item%3Fid%3D15892617\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=A%20simple%20DIY%20Heroku%20replacement%20to%20keep%20your%20hosting%20costs%20down&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=A%20simple%20DIY%20Heroku%20replacement%20to%20keep%20your%20hosting%20costs%20down\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=15892617&amp;auth=a721d19d0c3bbb4d9fa47b1c5a07464be5591dcd\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=15892617\">74\u00a0comments<\/a>              <\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"15894615\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelbuckbee\" class=\"hnuser\">michaelbuckbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894615\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894876\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=timewarrior\" class=\"hnuser\">timewarrior<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894876\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896124\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=glenngillen\" class=\"hnuser\">glenngillen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896124\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898011\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reaperducer\" class=\"hnuser\">reaperducer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898011\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898603\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simplify\" class=\"hnuser\">simplify<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898603\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898718\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alpha_squared\" class=\"hnuser\">alpha_squared<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898718\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15899712\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simplify\" class=\"hnuser\">simplify<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15899712\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894936\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=malyk\" class=\"hnuser\">malyk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894936\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895128\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelbuckbee\" class=\"hnuser\">michaelbuckbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895128\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15902239\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=timewarrior\" class=\"hnuser\">timewarrior<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15902239\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894986\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=acjohnson55\" class=\"hnuser\">acjohnson55<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894986\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898326\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rdoherty\" class=\"hnuser\">rdoherty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898326\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895831\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895831\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894918\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tedmiston\" class=\"hnuser\">tedmiston<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894918\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894233\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gitgud\" class=\"hnuser\">gitgud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894233\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894327\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894327\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894449\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894449\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896619\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmacvicar\" class=\"hnuser\">dmacvicar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896619\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896728\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896728\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896707\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmacvicar\" class=\"hnuser\">dmacvicar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896707\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894487\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894487\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895073\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianburrell\" class=\"hnuser\">ianburrell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895073\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894510\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894510\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894539\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894539\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894668\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894668\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15897531\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ryanl-ee\" class=\"hnuser\">ryanl-ee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15897531\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15897906\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yebyen\" class=\"hnuser\">yebyen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15897906\">11 months ago<\/a><\/p><br \/><div class=\"comment\">I use Deis on a single server.  It works great.<p>There are a few things you need to be aware of, but the tl;dr is that it works great for us.  We'd be using it more, except that our org hasn't adopted containers in any meaningful way yet, and Deis is large and complicated, and a little too much for them to bite off and swallow all at once.  They want to do a pilot on ECS so that we can say we're using more AWS services (yay lock-in, we love vendor lock-in.)<\/p><p>So, those things you need to be aware of:<\/p><p>#1 - In the default mode, Minio is used instead of a real Object Storage backend like S3 or Swift, and Minio is configured to keep all objects in RAM by default.  Use S3 instead.  This is an option to configure in values.yaml.<\/p><p>#2 - builder really needs to be on an SSD, but if you're not backing your Kube node with SSD, you can at least prevent a lot of builder crashes by making sure there isn't much swap to fill and bog down the disk with swapped pages.  This may be counter-intuitive (don't things crash when they run out of RAM?  Isn't that why you have swap...) until you realize that Builder has to respond to health checks by default once per second.  So if the node that builder runs on is getting even remotely bogged down by swappiness and lots of swappy pages, it will be killed and restarted, which takes some time and is noticeable and annoying.<\/p><p>#3 - The remaining cloud services that Deis can optionally depend on IMHO are take-or-leave, depending on your cloud provider.  Deis can do everything in-cluster.  I actually use this configuration with Minikube and Deis, and it works great.  I let Deis spin up its Postgres database on-cluster and restore from S3 WAL backups each day, it just works.<\/p><p>I'd be using Deis more heavily, except that my app development has a production database attached.  (Our infosec department is leary about letting me put backups of sensitive data into S3.  So we really have to use RDS for our app, even though Deis is creating a Postgres database that nothing else is stopping you from using with your app.  It's for Deis controller operation though.)<\/p><p>#4 - This one may be a dealbreaker if your company is very serious, Deis is now \"End-of-Life.\"  This means that Microsoft has committed to provide break fixes for Deis until March only, and by middle of next year, any break fixes will have to come from the community.  There is AFAIK nowhere to buy Deis support, although there is a new group that has formed to commit to support Deis into the future, it's not clear what form that support will take, or whether you'll be able to pay money for an SLA, or anything like that.  It is open-source.<\/p><p>This may be exactly what you'd prefer, or it may be a deal breaker.  I like Open Source, but my group likes support.<\/p><p>I do not work for Deis or any other cloud PaaS provider, although I have evaluated a lot of them including OpenShift (and I consider Deis to be technologically superior in some ways to OpenShift.)  Because of my ECS requirement, I'm looking at Convox now.  If you have to use ECS, so far I would definitely recommend Convox.  If you're like the rest of the world and using K8S, I can't think of any really good reason not to recommend Deis.\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898183\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898183\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898643\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yebyen\" class=\"hnuser\">yebyen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898643\">11 months ago<\/a><\/p><br \/><div class=\"comment\">I will tell you this: there is a group of people who are committed to keeping Deis alive, called Team Hephy and the Deis slack is still active.  I've used OpenShift and it is certainly more active, but I also have the very strong feeling that Deis is \"done\" where many parts of OpenShift look like they are still evolving.  (And I mean \"done\" in a good way.)<p>If you are not sure, feel free to drop in and ask any and all questions: <a href=\"https:\/\/slack.deis.io\/\" rel=\"nofollow\">https:\/\/slack.deis.io\/<\/a><\/p><p>The announcement of EOL came in July, and the announcement that Team Hephy would take over was almost one month to the day later, in August.  I have a lot of faith in the Deis codebase and I'm on Team Hephy now.<\/p><p>We have yet to put out a fully e2e acceptance test passing and providence-verified signed, versioned release of our own, but we have and know of a number of production users that still use Deis in Production.<\/p><p>It definitely does not feel like a ghost town on the Deis slack.  Questions get answered.  Neither does OpenShift (it is definitely not deserted, and it does not seem to suffer from \"abandoned k8s fork\" syndrome, they do keep up with the release cadence, because they have to.)<\/p><p>I would go with the one that solves your problem better.  I'm personally still recommending Deis, as I like the built-in support for Heroku buildpacks that OpenShift lacks.  (Convox also has support for this, on ECS.)  I'm not aware of any K8S-native Deis competitor (other than Hephy) at least for now.<\/p><p>Edit:<\/p><p>The really nice thing about working in Kubernetes is that you have the conformance suite[1].  You shouldn't worry about breaking something with the next Ubuntu update, because if Ubuntu update has broken Kubernetes in a way that will cause Deis to fail, then it will not be a Kubernetes-conformant distro anymore (and there is a large list[2] of K8S-conformant distros and platform to choose from, so you should pick a conformant one[3] instead.)<\/p><p>So, your question should be whether Deis will continue to work on Kubernetes distributions that are conformant, and with some minimal research, I think you should be able to say that with a high degree of certainty it will, regardless of ongoing vendor support from either side.<\/p><p>Then, you can ask whether your K8S distro is likely to remain conformant.  (This should be a no-brainer.)  Finally, is there anything that can break within Deis internally (not related to how it interacts with Kubernetes) that would mean you'd have to stop using it?<\/p><p>For me, the big reveal \/ final test will be, can we upgrade from Cedar stack to Heroku 16, which is on Team Hephy's planned roadmap for the future.<\/p><p>If that goes smoothly, then I'm not sure what else there is to worry about.<\/p><p>[1]: <a href=\"https:\/\/www.cncf.io\/certification\/software-conformance\/\" rel=\"nofollow\">https:\/\/www.cncf.io\/certification\/software-conformance\/<\/a><\/p><p>[2]: <a href=\"https:\/\/www.cncf.io\/certification\/software-conformance\/#logos\" rel=\"nofollow\">https:\/\/www.cncf.io\/certification\/software-conformance\/#logo...<\/a><\/p><p>[3]: <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1LxSqBzjOxfGx3cmtZ4EbB_BGCxT_wlxW_xgHVVa23es\/edit#gid=0\" rel=\"nofollow\">https:\/\/docs.google.com\/spreadsheets\/d\/1LxSqBzjOxfGx3cmtZ4Eb...<\/a>\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15899271\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15899271\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15899755\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yebyen\" class=\"hnuser\">yebyen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15899755\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898749\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898749\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896412\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dosuken123\" class=\"hnuser\">dosuken123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896412\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896448\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896448\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898381\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dosuken123\" class=\"hnuser\">dosuken123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898381\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896959\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896959\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894422\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Sevrene\" class=\"hnuser\">Sevrene<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894422\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894492\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894492\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895610\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Chickenosaurus\" class=\"hnuser\">Chickenosaurus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895610\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896454\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896454\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895776\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=friendly_chap\" class=\"hnuser\">friendly_chap<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895776\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895631\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joshmn\" class=\"hnuser\">joshmn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895631\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15897569\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joshmanders\" class=\"hnuser\">joshmanders<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15897569\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896855\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevekintor\" class=\"hnuser\">stevekintor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896855\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894613\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vfulco\" class=\"hnuser\">vfulco<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894613\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896133\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=samwillis\" class=\"hnuser\">samwillis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896133\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895380\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ufmace\" class=\"hnuser\">ufmace<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895380\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894252\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nkkollaw\" class=\"hnuser\">nkkollaw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894252\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894477\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cies\" class=\"hnuser\">cies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894477\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894571\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=phantom_oracle\" class=\"hnuser\">phantom_oracle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894571\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896076\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=y0ghur7_xxx\" class=\"hnuser\">y0ghur7_xxx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896076\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896121\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ytcs\" class=\"hnuser\">ytcs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896121\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15897945\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maxehmookau\" class=\"hnuser\">maxehmookau<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15897945\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15899027\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=diggan\" class=\"hnuser\">diggan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15899027\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15918148\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maxehmookau\" class=\"hnuser\">maxehmookau<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15918148\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896636\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanwilson\" class=\"hnuser\">seanwilson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896636\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15892628\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pritambarhate\" class=\"hnuser\">pritambarhate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15892628\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894740\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nbrempel\" class=\"hnuser\">nbrempel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894740\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894131\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894131\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894178\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=whack\" class=\"hnuser\">whack<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894178\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894436\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894436\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894931\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tedmiston\" class=\"hnuser\">tedmiston<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894931\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896766\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eggie5\" class=\"hnuser\">eggie5<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896766\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896811\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yebyen\" class=\"hnuser\">yebyen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896811\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898526\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898526\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15898942\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yebyen\" class=\"hnuser\">yebyen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15898942\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15899162\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15899162\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15899500\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yebyen\" class=\"hnuser\">yebyen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15899500\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15900290\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacques_chester\" class=\"hnuser\">jacques_chester<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15900290\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894320\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pbreit\" class=\"hnuser\">pbreit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894320\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15894570\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mikewhy\" class=\"hnuser\">mikewhy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15894570\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15896255\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cutler\" class=\"hnuser\">cutler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15896255\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895004\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=squid3\" class=\"hnuser\">squid3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895004\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"15895764\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=steventhedev\" class=\"hnuser\">steventhedev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15895764\">11 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:47:24+0000","updated_at":"2018-11-07T16:47:24+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":408,"title":"Shader can't be compiled","url":"https:\/\/stackoverflow.com\/questions\/20931528\/shader-cant-be-compiled","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        8&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>I'm following book \"OpenGL Programming Guide 8th Edition\". I just want to run the first program introduced in the book on my Mac.<\/p>\n<p>It's Mavericks + Xcode 4.6.1 + Intel HD graphics 4000. So the problem is, the shader can't be compiled.<\/p>\n<p>Shader codes:<\/p>\n<pre>#version 410 core\nlayout(location = 0) in vec4 vPosition;\nvoid\nmain()\n{\n    gl_Position = vPosition;\n}\n<\/pre>\n<p>And the error message is:<\/p>\n<pre>Shader compilation failed: ERROR: 0:1: '' :  version '410' is not supported\nERROR: 0:1: '' : syntax error #version\nERROR: 0:3: 'layout' : syntax error syntax error\n<\/pre>\n<p>I tried version 420\/400\/330, none of them works.<\/p>\n<p>By the way, the program uses latest glew 1.10(<a href=\"http:\/\/glew.sourceforge.net\" rel=\"noreferrer\">http:\/\/glew.sourceforge.net<\/a>), and I found that I have to set \"glewExperimental = GL_TRUE;\" before calling glewInit. Otherwise \"glGenVertexArray\" is a NULL pointer. So I'm wondering maybe glew doesn't support Mavericks?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/20931528\/revisions\" title=\"show all edits to this post\">edited Jan 5 '14 at 10:34<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1960455\/t-niese\">t.niese<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            20.5k63563&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked Jan 5 '14 at 8:21&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/444994\/mark-zhang\">Mark Zhang<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            1482411&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"2\">&#13;\n                                2 Answers&#13;\n                                2&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        22&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>MacOS uses <a href=\"https:\/\/developer.apple.com\/graphicsimaging\/opengl\/capabilities\/GLInfo_1090.html\"><strong>Legacy Profile<\/strong><\/a>  as default for all created OpenGL context. Therefor by default only OpenGL up to 2.1 and GLSL up to 1.20 is supported.<\/p>\n<p>To use OpenGL 3.2+ you need to switch to the <a href=\"https:\/\/developer.apple.com\/graphicsimaging\/opengl\/capabilities\/\"><strong>Core Profile<\/strong><\/a>. The naming there is a little bit confusing because it stats only 3.2Core profile, but actually this 3.2 or later (every OpenGL profile that is supported by the system\/driver that is backwards compatible to 3.2)<\/p>\n<p>For glut (depends on the version of glut if it works) the command on MacOS is:<\/p>\n<pre>glutInitDisplayMode(GLUT_3_2_CORE_PROFILE | ...  )\n<\/pre>\n<p>Where <code>| ...<\/code> would be the other options you want to pass to <code>glutInitDisplayMode<\/code>.<\/p>\n<p>About <code>glew<\/code>, normally you don't require <code>glew<\/code> on MacOS due the way how the OpenGL layer is implemented in MacOS. You are restricted to the OpenGL features MacOS provides\/exposes. So either the features are available via the headers of MacOS or not. There header would be <code>#include &lt;OpenGL\/gl3.h&gt;<\/code> where also the naming is missleading, it does not mean only OpenGL 3, it is the same like with the context. <\/p>\n<p>I would recommend to use <a href=\"http:\/\/www.glfw.org\/\"><strong>GLFW<\/strong><\/a> it is a great cross platform library similar to <code>GLUT<\/code> but as I think better to use.\nThere you would switch the context like this:<\/p>\n<pre> glfwWindowHint(GLFW_CONTEXT_VERSION_MAJOR, 3);\n glfwWindowHint(GLFW_CONTEXT_VERSION_MINOR, 2);\n glfwWindowHint(GLFW_OPENGL_PROFILE, GLFW_OPENGL_CORE_PROFILE);\n glfwWindowHint(GLFW_OPENGL_FORWARD_COMPAT, GL_TRUE);\n<\/pre><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 5 '14 at 11:06&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1960455\/t-niese\">t.niese<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            20.5k63563&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        5&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>It can be that your driver does not support the GLSL version you require. I had this problem on a laptop with Intel HD card on fresh install of Ubuntu half a year ago.<\/p>\n<p>Use glGetString to figure out which version is available to you:\n<a href=\"http:\/\/www.opengl.org\/sdk\/docs\/man\/xhtml\/glGetString.xml\" rel=\"noreferrer\">http:\/\/www.opengl.org\/sdk\/docs\/man\/xhtml\/glGetString.xml<\/a>.\nFor example,<\/p>\n<pre>printf(\"Supported GLSL version is %s.\\n\", (char *)glGetString(GL_SHADING_LANGUAGE_VERSION));\n<\/pre>\n<p>Here you can find about glewExperimental:\n<a href=\"http:\/\/www.opengl.org\/wiki\/Extension_Loading_Library\" rel=\"noreferrer\">http:\/\/www.opengl.org\/wiki\/Extension_Loading_Library<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Jan 5 '14 at 8:56&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3127029\/user3127029\">user3127029<\/a>&#13;\n        &#13;<\/div>","created_at":"2018-11-07T16:47:24+0000","updated_at":"2018-11-07T16:47:24+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":409,"title":"The Git Parable","url":"http:\/\/tom.preston-werner.com\/2009\/05\/19\/the-git-parable.html","content":"<p>Git is a simple, but extremely powerful system. Most people try to teach Git by\ndemonstrating a few dozen commands and then yelling \u201ctadaaaaa.\u201d I believe this\nmethod is flawed. Such a treatment may leave you with the ability to use Git to\nperform simple tasks, but the Git commands will still feel like magical\nincantations. Doing anything out of the ordinary will be terrifying. Until you\nunderstand the concepts upon which Git is built, you\u2019ll feel like a stranger in\na foreign land.<\/p><p>The following parable will take you on a journey through the creation of a\nGit-like system from the ground up. Understanding the concepts presented here\nwill be the most valuable thing you can do to prepare yourself to harness the\nfull power of Git. The concepts themselves are quite simple, but allow for an\namazing wealth of functionality to spring into existence. Read this parable all\nthe way through and you should have very little trouble mastering the various\nGit commands and wielding the awesome power that Git makes available to you.<\/p><h2 id=\"the-parable\">The Parable<\/h2><p>Imagine that you have a computer that has nothing on it but a text editor and a\nfew file system commands. Now imagine that you have decided to write a large\nsoftware program on this system. Because you\u2019re a responsible software\ndeveloper, you decide that you need to invent some sort of method for keeping\ntrack of versions of your software so that you can retrieve code that you\npreviously changed or deleted. What follows is a story about how you might\ndesign one such version control system (VCS) and the reasoning behind those\ndesign choices.<\/p><h2 id=\"snapshots\">Snapshots<\/h2><p>Alfred is a friend of yours that works down at the mall as a photographer in one\nof those \u201cSpecial Moments\u201d photo boutiques. All day long he takes photos of\nlittle kids posing awkwardly in front of jungle or ocean backdrops. During one\nof your frequent lunches at the pretzel stand, Alfred tells you a story about a\nwoman named Hazel who brings her daughter in for a portrait every year on the\nsame day. \u201cShe brings the photos from all the past years with her,\u201d Alfred tells\nyou. \u201cShe likes to remember what her daughter was like at each different stage,\nas if the snapshots really let her move back and forth in time to those saved\nmemories.\u201d<\/p><p>Like some sort of formulaic plot device, Alfred\u2019s innocent statement acts as a\ncatalyst for you to see the ideal solution to your version control dilemma.\nSnapshots, like save points in a video game, are really what you care about when\nyou need to interact with a VCS. What if you could take snapshots of your\ncodebase at any time and resurrect that code on demand? Alfred reads the dawning\nrealization spreading across your face and knows you\u2019re about to leave him\nwithout another word to go back and implement whatever genius idea he just\ncaused you to have. You do not disappoint him.<\/p><p>You start your project in a directory named <code>working<\/code>. As you code,\nyou try to write one feature at a time. When you complete a self-contained\nportion of a feature, you make sure that all your files are saved and then make\na copy of the entire working directory, giving it the name\n<code>snapshot-0<\/code>. After you perform this copy operation, you make sure to\nnever again change the code files in the new directory. After the next chunk of\nwork, you perform another copy, only this time the new directory gets the name\n<code>snapshot-1<\/code>, and so on.<\/p><p>To make it easy to remember what changes you made in each snapshot, you add a\nspecial file named <code>message<\/code> to each snapshot directory that contains\na summary of the work that you did and the date of completion. By printing the\ncontents of each message, it becomes easy to find a specific change that you\nmade in the past, in case you need to resurrect some old code.<\/p><h2 id=\"branches\">Branches<\/h2><p>After a bit of time on the project, a candidate for release begins to emerge.\nLate nights at the keyboard finally yield <code>snapshot-99<\/code>, the nascent\nform of what will become Release Version 1.0. It comes to pass that this\nsnapshot is packaged and distributed to the eagerly awaiting masses. Stoked by\nexcellent response to your software, you push forward, determined to make the\nnext version an even bigger success.<\/p><p>Your VCS has so far been a faithful companion. Old versions of your code are\nthere when you need them and can be accessed with ease. But not long after the\nrelease, bug reports start to come in. Nobody\u2019s perfect, you reassure yourself,\nand <code>snapshot-99<\/code> is readily retrievable, glad to be brought back to\nlife for the purposes of applying bug fixes.<\/p><p>Since the release, you\u2019ve created 10 new snapshots. This new work must not be\nincluded in the 1.0.1 bug fix version you now need to create. To solve this, you\ncopy <code>snapshot-99<\/code> to <code>working<\/code> so that your working\ndirectory is at exactly the point where Version 1.0 was released. A few swift\nlines of code and the bug is fixed in the working directory.<\/p><p>It is here that a problem becomes apparent. The VCS deals very well with linear\ndevelopment, but for the first time ever, you need to create a new snapshot that\nis not a direct descendent of the preceding snapshot. If you create a\n<code>snapshot-110<\/code> (remember that you created 10 snapshots since the\nrelease), then you\u2019ll be interrupting the linear flow and will have no way of\ndetermining the ancestry of any given snapshot. Clearly, you need something more\npowerful than a linear system.<\/p><p>Studies show that even short exposures to nature can help recharge the mind\u2019s\ncreative potential. You\u2019ve been sitting behind the artificially polarized light\nof your monitor for days on end. A walk through the woods in the brisk Autumn\nair will do you some good and with any luck, will help you arrive at an ideal\nsolution to your problem.<\/p><p>The great oaks that line the trail have always appealed to you. They seem to\nstand stark and proud against the perfectly blue sky. Half the ruddy leaves have\ndeparted from their branches, leaving an intricate pattern of branches in their\nwake. Fixating on one of the thousands of branch tips you idly try to follow it\nback to the solitary trunk. This organically produced structure allows for such\ngreat complexity, but the rules for finding your way back to the trunk are so\nsimple, and perfect for keeping track of multiple lines of development! It turns\nout that what they say about nature and creativity are true.<\/p><p>By looking at your code history as a tree, solving the problem of ancestry\nbecomes trivial. All you need to do is include the name of the parent snapshot\nin the <code>message<\/code> file you write for each snapshot. Adding just a\nsingle upstream pointer will enable you to easily and accurately trace the\nhistory of any given snapshot all the way back to the root.<\/p><h2 id=\"branch-names\">Branch Names<\/h2><p>Your code history is now a tree. Instead of having a single latest snapshot, you\nhave two: one for each branch. With a linear system, your sequential numbering\nsystem let you easily identify the latest snapshot. Now, that ability is lost.<\/p><p>Creating new development branches has become so simple that you\u2019ll want to take\nadvantage of it all the time. You\u2019ll be creating branches for fixes to old\nreleases, for experiments that may not pan out; indeed it becomes possible to\ncreate a new branch for every feature you begin!<\/p><p>But like everything good in life, there is a price to be paid. Each time you\ncreate a new snapshot, you must remember that the new snapshot becomes the\nlatest on its branch. Without this information, switching to a new branch would\nbecome a laborious process indeed.<\/p><p>Every time you create a new branch you probably give it a name in your head.\n\u201cThis will be the Version 1.0 Maintenance Branch,\u201d you might say. Perhaps you\nrefer to the former linear branch as the \u201cmaster\u201d branch.<\/p><p>Think about this a little further, though. From the perspective of a tree, what\ndoes it mean to name a branch? Naming every snapshot that appears in the history\nof a branch would do the trick, but requires the storage of a potentially large\namount of data. Additionally, it still wouldn\u2019t help you efficiently locate the\nlatest snapshot on a branch.<\/p><p>The least amount of information necessary to identify a branch is the location\nof the latest snapshot on that branch. If you need to know the list of snapshots\nthat are part of the branch you can easily trace the parentage.<\/p><p>Storing the branch names is trivial. In a file named <code>branches<\/code>,\nstored outside of any specific snapshot, you simply list the name\/snapshot pairs\nthat represent the tips of branches. To switch to a named branch you need only\nlook up the snapshot for the corresponding name from this file.<\/p><p>Because you\u2019re only storing the latest snapshot on each branch, creating a new\nsnapshot now contains an additional step. If the new snapshot is being created\nas part of a branch, the <code>branches<\/code> file must be updated so that the\nname of the branch becomes associated with the new snapshot. A small price to\npay for the benefit.<\/p><p>After using branches for a while you notice that they can serve two purposes.\nFirst, they can act as movable pointers to snapshots so that you can keep track\nof the branch tips. Second, they can be pointed at a single snapshot and never\nmove.<\/p><p>The first use case allows you to keep track of ongoing development, things like\n\u201cRelease Maintenance\u201d. The second case is useful for labeling points of\ninterest, like \u201cVersion 1.0\u201d and \u201cVersion 1.0.1\u201d.<\/p><p>Mixing both of these uses into a single file feels messy. Both types are\npointers to snapshots, but one moves and one doesn\u2019t. For the sake of clarity\nand elegance, you decide to create another file called <code>tags<\/code> to\ncontain pointers of the second type.<\/p><p>Keeping these two inherently different pointers in separate files will help you\nfrom accidentally treating a branch as a tag or vice versa.<\/p><h2 id=\"distributed\">Distributed<\/h2><p>Working on your own gets pretty lonely. Wouldn\u2019t it be nice if you could invite\na friend to work on your project with you? Well, you\u2019re in luck. Your friend Zoe\nhas a computer setup just like yours and wants to help with the project. Because\nyou\u2019ve created such a great version control system, you tell her all about it\nand send her a copy of all your snapshots, branches, and tags so she can enjoy\nthe same benefits of the code history.<\/p><p>It\u2019s great to have Zoe on the team but she has a habit of taking long trips to\nfar away places without internet access. As soon as she has the source code, she\ncatches a flight to Patagonia and you don\u2019t hear from her for a week. In the\nmeantime you both code up a storm. When she finally gets back, you discover a\ncritical flaw in your VCS. Because you\u2019ve both been using the same numbering\nsystem, you each have directories named \u2018snapshot-114\u2019, \u2018snapshot-115\u2019, and so\non, but with different contents!<\/p><p>To make matters worse, you don\u2019t even know who authored the changes in those new\nsnapshots. Together, you devise a plan for dealing with these problems. First,\nsnapshot messages will henceforth contain author name and email. Second,\nsnapshots will no longer be named with simple numbers. Instead, you\u2019ll use the\ncontents of the message file to produce a hash. This hash will be guaranteed to\nbe unique to the snapshot since no two messages will ever have the same date,\nmessage, parent, and author. To make sure everything goes smoothly, you both\nagree to use the SHA1 hash algorithm that takes the contents of a file and\nproduces a 40 character hexadecimal string. You both update your histories with\nthe new technique and instead of clashing \u2018snapshot-114\u2019 directories, you now\nhave distinct directories named \u20188ba3441b6b89cad23387ee875f2ae55069291f4b\u2019 and\n\u2018db9ecb5b5a6294a8733503ab57577db96ff2249e\u2019.<\/p><p>With the updated naming scheme, it becomes trivial for you to fetch all the new\nsnapshots from Zoe\u2019s computer and place them next to your existing snapshots.\nBecause every snapshot specifies its parent, and identical messages (and\ntherefor identical snapshots) have identical names no matter where they are\ncreated, the history of the codebase can still be drawn as a tree. Only now, the\ntree is comprised of snapshots authored by both Zoe and you.<\/p><p>This point is important enough to warrant repeating. A snapshot is identified by\na SHA1 that uniquely identifies it (and its parent). These snapshots can be\ncreated and moved around between computers without losing their identity or\nwhere they belong in the history tree of a project. What\u2019s more, snapshots can\nbe shared or kept private as you see fit. If you have some experimental\nsnapshots that you want to keep to yourself, you can do so quite easily. Just\ndon\u2019t make them available to Zoe!<\/p><h2 id=\"offline\">Offline<\/h2><p>Zoe\u2019s travel habits cause her to spend countless hours on airplanes and boats.\nMost of the places she visits have no readily available internet access. At the\nend of the day, she spends more time offline than online.<\/p><p>It\u2019s no surprise, then, that Zoe raves about your VCS. All of the day to day\noperations that she needs to do can be done locally. The only time she needs a\nnetwork connection is when she\u2019s ready to share her snapshots with you.<\/p><h2 id=\"merges\">Merges<\/h2><p>Before Zoe left on her trip, you had asked her to start working off of the\nbranch named \u2018math\u2019 and to implement a function that generated prime numbers.\nMeanwhile, you were also developing off of the \u2018math\u2019 branch, only you were\nwriting a function to generate magic numbers. Now that Zoe has returned, you are\nfaced with the task of merging these two separate branches of development into a\nsingle snapshot. Since you both worked on separate tasks, the merge is simple.\nWhile constructing the snapshot message for the merge, you realize that this\nsnapshot is special. Instead of just a single parent, this merge snapshot has\ntwo parents! The first parent is your latest on the \u2018math\u2019 branch and the second\nparent is Zoe\u2019s latest on her \u2018math\u2019 branch. The merge snapshot doesn\u2019t contain\nany changes beyond those necessary to merge the two disparate parents into a\nsingle codebase.<\/p><p>Once you complete the merge, Zoe fetches all the snapshots that you have that\nshe does not, which include your development on the \u2018math\u2019 branch and your merge\nsnapshot. Once she does this, both of your histories match exactly!<\/p><h2 id=\"rewriting-history\">Rewriting History<\/h2><p>Like many software developers you have a compulsion to keep your code clean and\nvery well organized. This carries over into a desire to keep your code history\nwell groomed. Last night you came home after having a few too many pints of\nGuinness at the local brewpub and started coding, producing a handful of\nsnapshots along the way. This morning, a review of the code you wrote last night\nmakes you cringe a little bit. The code is good overall, but you made a lot of\nmistakes early on that you corrected in later snapshots.<\/p><p>Let\u2019s say the branch on which you did your drunken development is called \u2018drunk\u2019\nand you made three snapshots after you got home from the bar. If the name\n\u2018drunk\u2019 points at the latest snapshot on that branch, then you can use a useful\nnotation to refer to the parent of that snapshot. The notation \u2018drunk^\u2019 means\nthe parent of the snapshot pointed to by the branch name \u2018drunk\u2019. Similarly\n\u2018drunk^^\u2019 means the grandparent of the \u2018drunk\u2019 snapshot. So the three snapshots\nin chronological order are \u2018drunk^^\u2019, \u2018drunk^\u2019, and \u2018drunk\u2019.<\/p><p>You\u2019d really like those three lousy snapshots to be two clean snapshots. One\nthat changes an existing function, and one that adds a new file. To accomplish\nthis revision of history you copy \u2018drunk\u2019 to \u2018working\u2019 and delete the file that\nis new in the series. Now \u2018working\u2019 represents the correct modifications to the\nexisting function. You create a new snapshot from \u2018working\u2019 and write the\nmessage to be appropriate to the changes. For the parent you specify the SHA1 of\nthe \u2018drunk^^^\u2019 snapshot, essentially creating a new branch off of the same\nsnapshot as last night. Now you can copy \u2018drunk\u2019 to \u2018working\u2019 and roll a\nsnapshot with the new file addition. As the parent you specify that snapshot you\ncreated just before this one.<\/p><p>As the last step, you change the branch name \u2018drunk\u2019 to point to the last\nsnapshot you just made.<\/p><p>The history of the \u2018drunk\u2019 branch now represents a nicer version of what you did\nlast night. The other snapshots that you\u2019ve replaced are no longer needed so you\ncan delete them or just leave them around for posterity. No branch names are\ncurrently pointing at them so it will be hard to find them later on, but if you\ndon\u2019t delete them, they\u2019ll stick around.<\/p><h2 id=\"staging-area\">Staging Area<\/h2><p>As much as you try to keep your new modifications related to a single feature or\nlogical chunk, you sometimes get sidetracked and start hacking on something\ntotally unrelated. Only half-way into this do you realize that your working\ndirectory now contains what should really be separated as two discrete\nsnapshots.<\/p><p>To help you with this annoying situation, the concept of a staging directory is\nuseful. This area acts as an intermediate step between your working directory\nand a final snapshot. Each time you finish a snapshot, you also copy that to a\n<code>staging<\/code> directory. Now, every time you finish an edit to a new\nfile, create a new file, or remove a file, you can decide whether that change\nshould be part of your next snapshot. If it belongs, you mimic the change inside\n<code>staging<\/code>. If it doesn\u2019t, you can leave it in <code>working<\/code>\nand make it part of a later snapshot. From now on, snapshots are created\ndirectly from the staging directory.<\/p><p>This separation of coding and preparing the stage makes it easy to specify what\nis and is not included in the next snapshot. You no longer have to worry too\nmuch about making an accidental, unrelated change in your working directory.<\/p><p>You have to be a bit careful, though. Consider a file named <code>README<\/code>.\nYou make an edit to this file and then mimic that in <code>staging<\/code>. You\ngo on about your business, editing other files. After a bit, you make another\nchange to <code>README<\/code>. Now you have made two changes to that file, but\nonly one is in the staging area! Were you to create a snapshot now, your second\nchange would be absent.<\/p><p>The lesson is this: every new edit must be added to the staging area if it is to\nbe part of the next snapshot.<\/p><h2 id=\"diffs\">Diffs<\/h2><p>With a working directory, a staging area, and loads of snapshots laying around,\nit starts to get confusing as to what the specific code changes are between\nthese directories. A snapshot message only gives you a summary of what changed,\nnot exactly what lines were changed between two files.<\/p><p>Using a diffing algorithm, you can implement a small program that shows you the\ndifferences in two codebases. As you develop and copy things from your working\ndirectory to the staging area, you\u2019ll want to easily see what is different\nbetween the two, so that you can determine what else needs to be staged. It\u2019s\nalso important to see how the staging area is different from the last snapshot,\nsince these changes are what will become part of the next snapshot you produce.<\/p><p>There are many other diffs you might want to see. The differences between a\nspecific snapshot and its parent would show you the \u201cchangeset\u201d that was\nintroduced by that snapshot. The diff between two branches would be helpful for\nmaking sure your development doesn\u2019t wander too far away from the mainline.<\/p><h2 id=\"eliminating-duplication\">Eliminating Duplication<\/h2><p>After a few more trips to Namibia, Istanbul, and Galapagos, Zoe starts to\ncomplain that her hard drive is filling up with hundreds of nearly identical\ncopies of the software. You too have been feeling like all the file duplication\nis wasteful. After a bit of thinking, you come up with something very clever.<\/p><p>You remember that the SHA1 hash produces a short string that is unique for a\ngiven file contents. Starting with the very first snapshot in the project\nhistory, you start a conversion process. First, you create a directory named\n<code>objects<\/code> outside of the code history. Next, you find the most deeply\nnested directory in the snapshot. Additionally, you open up a temporary file for\nwriting. For each file in this directory you perform three steps. Step 1:\nCalculate the SHA1 of the contents. Step 2: Add an entry into the temp file that\ncontains the word \u2018blob\u2019 (binary large object), the SHA1 from the first step,\nand the filename. Step 3: Copy the file to the objects directory and rename it\nto the SHA1 from step 1. Once finished with all the files, find the SHA1 of the\ntemp file contents and use that to name the temp file, also placing it in the\nobjects directory.<\/p><p>If at any time the objects directory already contains a file with a given name,\nthen you have already stored that file\u2019s contents and there is no need to do so\nagain.<\/p><p>Now, move up one directory and start over. Only this time, when you get to the\nentry for the directory that you just processed, enter the word \u2018tree\u2019, the SHA1\nof the temp file from last time, and the directory\u2019s name into the new temp\nfile. In this fashion you can build up a tree of directory object files that\ncontain the SHA1s and names of the files and directory objects that they\ncontain.<\/p><p>Once this has been accomplished for every directory and file in the snapshot,\nyou have a single root directory object file and its corresponding SHA1. Since\nnothing contains the root directory, you must record the root tree\u2019s SHA1\nsomewhere. An ideal place to store it is in the snapshot message file. This way,\nthe uniqueness of the SHA1 of the message also depends on the entire contents of\nthe snapshot, and you can guarantee with absolute certainty that two identical\nsnapshot message SHA1s contain the same files!<\/p><p>It\u2019s also convenient to create an object from the snapshot message in the same\nway that you do for blobs and trees. Since you\u2019re maintaining a list of branch\nand tag names that point to message SHA1s you don\u2019t have to worry about losing\ntrack of which snapshots are important to you.<\/p><p>With all of this information stored in the objects directory, you can safely\ndelete the snapshot directory that you used as the source of this operation. If\nyou want to reconstitute the snapshot at a later date it\u2019s simply a matter of\nfollowing the SHA1 of the root tree stored in the message file and extracting\neach tree and blob into their corresponding directory and file.<\/p><p>For a single snapshot, this transformation process doesn\u2019t get you much. You\u2019ve\nbasically just converted one filesystem into another and created a lot of work\nin the process. The real benefits of this system arise from reuse of trees and\nblobs across snapshots. Imagine two sequential snapshots in which only a single\nfile in the root directory has changed. If the snapshots both contain 10\ndirectories and 100 files, the transformation process will create 10 trees and\n100 blobs from the first snapshot but only one new blob and one new tree from\nthe second snapshot!<\/p><p>By converting every snapshot directory in the old system to object files in the\nnew system, you can drastically reduce the number of files that are stored on\ndisk. Now, instead of storing perhaps 50 identical copies of a rarely changed\nfile, you only need to keep one.<\/p><h2 id=\"compressing-blobs\">Compressing Blobs<\/h2><p>Eliminating blob and tree duplication significantly reduces the total storage\nsize of your project history, but that\u2019s not the only thing you can do to save\nspace. Source code is just text. Text can be very efficiently compressed using\nsomething like the LZW or DEFLATE compression algorithms. If you compress every\nblob before computing its SHA1 and saving it to disk you can reduce the total\nstorage size of the project history by another very admirable quantity.<\/p><h2 id=\"the-true-git\">The True Git<\/h2><p>The VCS you have constructed is now a reasonable facsimile of Git. The main\ndifference is that Git gives you very nice command lines tools to handle such\nthings as creating new snapshots and switching to old ones (Git uses the term\n\u201ccommit\u201d instead of \u201csnapshot\u201d), tracing history, keeping branch tips\nup-to-date, fetching changes from other people, merging and diffing branches,\nand hundreds of other common (and not-so-common tasks).<\/p><p>As you continue to learn Git, keep this parable in mind. Git is really very\nsimple underneath, and it is this simplicity that makes it so flexible and\npowerful. One last thing before you run off to learn all the Git commands:\nremember that it is almost impossible to lose work that has been committed. Even\nwhen you delete a branch, all that\u2019s really happened is that the pointer to that\ncommit has been removed. All of the snapshots are still in the objects\ndirectory, you just need to dig up the commit SHA. In these cases, look up\n<code>git reflog<\/code>. It contains a history of what each branch pointed to\nand in times of crisis, it will save the day.<\/p><p>Here are some resources that you should follow as your next step. Now, go, and\nbecome a Git master!<\/p><ul><li><a href=\"http:\/\/learn.github.com\/\">Learn Git<\/a><\/li>\n  <li><a href=\"http:\/\/book.git-scm.com\/\">Git Community Book<\/a><\/li>\n  <li><a href=\"http:\/\/www-cs-students.stanford.edu\/~blynn\/gitmagic\/\">Git Magic<\/a><\/li>\n<\/ul><p>\u2013<\/p><p><a href=\"http:\/\/news.ycombinator.com\/item?id=615308\">Discuss this post on Hacker News<\/a><\/p><p><a rel=\"license\" href=\"http:\/\/creativecommons.org\/licenses\/by-nc-sa\/3.0\/us\/\"><img alt=\"Creative Commons License\" src=\"http:\/\/i.creativecommons.org\/l\/by-nc-sa\/3.0\/us\/80x15.png\" \/><\/a><\/p>","created_at":"2018-11-07T16:47:24+0000","updated_at":"2018-11-07T16:47:24+0000","annotations":[],"mimetype":"text\/html","reading_time":21,"domain_name":"tom.preston-werner.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":406,"title":"About - Project Euler","url":"https:\/\/projecteuler.net\/","content":"<img src=\"https:\/\/projecteuler.net\/images\/euler_portrait.png\" alt=\"Leonhard Euler (1707-1783)\" \/><h3>What is Project Euler?<\/h3><p>Project Euler is a series of challenging mathematical\/computer programming problems that will require more than just mathematical insights to solve. Although mathematics will help you arrive at elegant and efficient methods, the use of a computer and programming skills will be required to solve most problems.<\/p><p>The motivation for starting Project Euler, and its continuation, is to provide a platform for the inquiring mind to delve into unfamiliar areas and learn new concepts in a fun and recreational context.<\/p><br \/><h3>Who are the problems aimed at?<\/h3><p>The intended audience include students for whom the basic curriculum is not feeding their hunger to learn, adults whose background was not primarily mathematics but had an interest in things mathematical, and professionals who want to keep their problem solving and mathematics on the cutting edge.<\/p><br \/><h3>Can anyone solve the problems?<\/h3><p>The problems range in difficulty and for many the experience is inductive chain learning. That is, by solving one problem it will expose you to a new concept that allows you to undertake a previously inaccessible problem. So the determined participant will slowly but surely work his\/her way through every problem.<\/p><br \/><h3>What next?<\/h3><p>In order to track your progress it is necessary to setup an account and have Cookies enabled. If you already have an account then <a href=\"https:\/\/projecteuler.net\/sign_in\">Login<\/a>, otherwise please <a href=\"https:\/\/projecteuler.net\/register\">Register<\/a> \u2013 it's completely free!<\/p><p>However, as the problems are challenging then you may wish to view the <a href=\"https:\/\/projecteuler.net\/problems\">Problems<\/a> before registering.<\/p><p>\"Project Euler exists to encourage, challenge, and develop the skills and enjoyment of anyone with an interest in the fascinating world of mathematics.\"<\/p>","created_at":"2018-11-07T16:47:23+0000","updated_at":"2018-11-07T16:47:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"projecteuler.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":405,"title":"Orgmode for GTD","url":"https:\/\/emacs.cafe\/emacs\/orgmode\/gtd\/2017\/06\/30\/orgmode-gtd.html","content":"<p>I\u2019ve been using <a href=\"http:\/\/orgmode.org\/\">Orgmode<\/a> to implement\nthe <a href=\"https:\/\/en.wikipedia.org\/wiki\/Getting_Things_Done\">GTD methodology<\/a> for the\nlast 4 years.<\/p><p>Rather than explaining the GTD methodology itself or how Orgmode works, in this\npost I\u2019ll detail how I use Orgmode to implement GTD.<\/p>\n<p>If you don\u2019t know Orgmode and are curious about it, you should head\nto <a href=\"http:\/\/orgmode.org\/\">its website<\/a> first.<\/p>\n<blockquote class=\"twitter-tweet\" data-lang=\"en\"><p lang=\"en\" dir=\"ltr\">After 4 years of using <a href=\"https:\/\/twitter.com\/hashtag\/gtd?src=hash\">#gtd<\/a>  daily with <a href=\"https:\/\/twitter.com\/hashtag\/orgmode?src=hash\">#orgmode<\/a>, I now have 39,175 lines of orgmode in 4 files + archives. <a href=\"https:\/\/twitter.com\/hashtag\/emacs?src=hash\">#emacs<\/a><\/p><div>\u2014 Nicolas Petton (@NicolasPetton) <\/div><a href=\"https:\/\/twitter.com\/NicolasPetton\/status\/879673555616714752\">June 27, 2017<\/a><\/blockquote>\n<h2 id=\"the-orgmode-files\">The orgmode files<\/h2>\n<p>I split my GTD in four separate files:<\/p>\n<ul><li><code class=\"highlighter-rouge\">inbox.org<\/code>: where I collect everything;<\/li>\n  <li><code class=\"highlighter-rouge\">gtd.org<\/code>: where I put all my projects;<\/li>\n  <li><code class=\"highlighter-rouge\">someday.org<\/code>: All inactive tasks that I might do at some point in the future,\nbut don\u2019t want to see all the time;<\/li>\n  <li><code class=\"highlighter-rouge\">tickler.org<\/code>: I put entries in this file with\na <a href=\"http:\/\/orgmode.org\/manual\/Timestamps.html\">timestamp<\/a> to get reminded at\nthe right moment.<\/li>\n<\/ul><p>It\u2019s important to add these files to the agenda file (more on the agenda later),\nlike the following:<\/p>\n<div class=\"language-emacs-lisp highlighter-rouge\"><pre>(setq org-agenda-files '(\"~\/gtd\/inbox.org\"\n                         \"~\/gtd\/gtd.org\"\n                         \"~\/gtd\/tickler.org\"))\n<\/pre>\n<\/div>\n<h2 id=\"1-the-gtd-inbox\">1. The GTD inbox<\/h2>\n<p>One of the most important aspects of GTD is the inbox. Every thought should be\ncollected there, and processed later on.<\/p>\n<p><img src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/1\/1b\/GTDcanonical.png\/1280px-GTDcanonical.png\" alt=\"inbox\" \/><\/p>\n<p>Orgmode has a great feature that fits really well with this\nconcept: <a href=\"http:\/\/orgmode.org\/manual\/Capture.html\">org-capture<\/a>.<\/p>\n<p>Capturing a thought is one key press away: simply Press <code class=\"highlighter-rouge\">C-c c<\/code>, and a capture\npopup will appear in Emacs. Once you\u2019re done capturing, <code class=\"highlighter-rouge\">C-c C-c<\/code> and it will\nget stored in the inbox.<\/p>\n<p><img src=\"https:\/\/emacs.cafe\/img\/gtd-capture.png\" alt=\"capture\" \/><\/p>\n<p>Here\u2019s how I set it up:<\/p>\n<div class=\"language-emacs-lisp highlighter-rouge\"><pre>(setq org-capture-templates '((\"t\" \"Todo [inbox]\" entry\n                               (file+headline \"~\/gtd\/inbox.org\" \"Tasks\")\n                               \"* TODO %i%?\")\n                              (\"T\" \"Tickler\" entry\n                               (file+headline \"~\/gtd\/tickler.org\" \"Tickler\")\n                               \"* %i%? \\n %U\")))\n<\/pre>\n<\/div>\n<p>The syntax of capture templates is\nexplained\n<a href=\"http:\/\/orgmode.org\/manual\/Capture-templates.html#Capture-templates\">here<\/a>. It\noffers lots of customization options.<\/p>\n<p>I press <code class=\"highlighter-rouge\">C-c c t<\/code> to add an entry to my inbox, and <code class=\"highlighter-rouge\">C-c c T<\/code> to add an entry to\nthe tickler (more on that later).<\/p>\n<p>Here\u2019s how my inbox looks like:<\/p>\n<p><img src=\"https:\/\/emacs.cafe\/img\/gtd-inbox.png\" alt=\"inbox\" \/><\/p>\n<p>My inbox is then processed and emptied daily. When processing the inbox,\nI <a href=\"http:\/\/orgmode.org\/manual\/Refile-and-copy.html#Refile-and-copy\">refile<\/a> each\nentry that is actionable and belongs to a project using <code class=\"highlighter-rouge\">C-c C-w<\/code>, moving the\nentry to the appropriate place. If need be, I create a new project out of it.<\/p>\n<p>I have set up the refile targets as follows:<\/p>\n<div class=\"language-emacs-lisp highlighter-rouge\"><pre>(setq org-refile-targets '((\"~\/gtd\/gtd.org\" :maxlevel . 3)\n                           (\"~\/gtd\/someday.org\" :level . 1)\n                           (\"~\/gtd\/tickler.org\" :maxlevel . 2)))\n<\/pre>\n<\/div>\n<p>So that <code class=\"highlighter-rouge\">C-c C-w<\/code> prompts me either for a project, the tickler, or someday\/maybe\nlist.<\/p>\n<h2 id=\"2-the-projects-file\">2. The projects file<\/h2>\n<p>My main file is <code class=\"highlighter-rouge\">gtd.org<\/code>. That\u2019s where I keep all my active projects. I usually\nhave around 30 active projects at the same time.<\/p>\n<p>Each project contains actions to be performed. The first action of each project\nis called its \u201cnext action\u201d, and that\u2019s always the one I will do when working on\na project. Once a task is done, I mark it as such using the <code class=\"highlighter-rouge\">DONE<\/code> todo keyword.<\/p>\n<p>Here\u2019s an example project:<\/p>\n<p><img src=\"https:\/\/emacs.cafe\/img\/gtd-project-example.png\" alt=\"project example\" \/><\/p>\n<p>The completion percentage you see on the screenshot\nis <a href=\"http:\/\/orgmode.org\/manual\/Checkboxes.html\">yet another neat feature<\/a> of\nOrgmode :)<\/p>\n<h3 id=\"tags\">Tags<\/h3>\n<p>Tagging is done using <code class=\"highlighter-rouge\">C-c C-c<\/code> on a headline, whether it\u2019s a project or\naction. I use tags for several purposes:<\/p>\n<ul><li>Regular categories, like <code class=\"highlighter-rouge\">:emacs:<\/code> or <code class=\"highlighter-rouge\">:writing:<\/code>;<\/li>\n  <li>Tags that link to people, like <code class=\"highlighter-rouge\">:daniel:<\/code>;<\/li>\n  <li>GTD contexts.<\/li>\n<\/ul><p>GTD contexts are just regular tags, starting with <code class=\"highlighter-rouge\">@<\/code>. I make heavy use of them\nin\n<a href=\"http:\/\/orgmode.org\/worg\/org-tutorials\/org-custom-agenda-commands.html\">custom Org Agenda commands<\/a>.<\/p>\n<p>My contexts tend to change over time, but I always have at least <code class=\"highlighter-rouge\">@home<\/code>,\n<code class=\"highlighter-rouge\">@office<\/code>, <code class=\"highlighter-rouge\">@travelling<\/code>, <code class=\"highlighter-rouge\">@phone<\/code>, <code class=\"highlighter-rouge\">@email<\/code>, <code class=\"highlighter-rouge\">@errands<\/code> to filter out next\nactions based on my current location for instance.<\/p>\n<h3 id=\"todo-keywords\">TODO keywords<\/h3>\n<p>I put a todo keyword in all project entries. I think I use fairly regular todo\nkeywords: <code class=\"highlighter-rouge\">TODO<\/code>, <code class=\"highlighter-rouge\">WAITING<\/code>, <code class=\"highlighter-rouge\">DONE<\/code> and <code class=\"highlighter-rouge\">CANCELLED<\/code> mostly. The first two for\nare used for incomplete states, and the last two for completed states.<\/p>\n<div class=\"language-emacs-lisp highlighter-rouge\"><pre>(setq org-todo-keywords '((sequence \"TODO(t)\" \"WAITING(w)\" \"|\" \"DONE(d)\" \"CANCELLED(c)\")))\n<\/pre>\n<\/div>\n<p>When on a headline, press <code class=\"highlighter-rouge\">C-c C-t<\/code> to set the TODO keyword.<\/p>\n<h3 id=\"timestamps-scheduling--deadlines\">Timestamps, scheduling &amp; deadlines<\/h3>\n<p>I tend to avoid using timestamps in my projects as much as possible. The reason\nis simple: unless an entry is an appointment (to the dentist for instance) or\nhas a fixed deadline (a release scheduled with a customer), I should decide what\nto work on depending on the current context (among other things). This also\nkeeps my agenda clean, free of any fake or self-imposed deadline or schedule.<\/p>\n<p>But scheduling sometimes makes sense. To do that, press <code class=\"highlighter-rouge\">C-c C-s<\/code> on an entry,\nand enter the date and\/or time. To add a deadline, press <code class=\"highlighter-rouge\">C-c C-d<\/code>. Note that\nOrgmode is quite smart about how you can enter a date, if you don\u2019t know about\nit, refer to\nthe <a href=\"http:\/\/orgmode.org\/manual\/Deadlines-and-scheduling.html\">manual entry<\/a>.<\/p>\n<h3 id=\"filtering-projects--actions\">Filtering projects &amp; actions<\/h3>\n<p>When deciding what to work on, I use\neither <a href=\"http:\/\/orgmode.org\/manual\/Sparse-trees.html\">sparse trees<\/a> \u2013 which\nmakes it easy to filter the content of my GTD projects by tag, search term,\netc., or I\nuse\n<a href=\"http:\/\/orgmode.org\/worg\/org-tutorials\/org-custom-agenda-commands.html\">custom agenda commands<\/a>. When\ndiscovering Orgmode, most people think that its agenda is just a regular\nagenda. Sure, it does daily\/weekly agendas, but it offers much more than\nthat. Quoting the manual:<\/p>\n<blockquote>\n  <p>Org-mode\u2019s built-in agenda commands are powerful tools for searching your\nnotes and for gathering, sorting, filtering, and displaying your tasks.<\/p>\n<\/blockquote>\n<p>I use custom agenda commands mostly to get an overview of actions by context or\ntag. Here\u2019s an example custom agenda command that will display all actions for\nthe <code class=\"highlighter-rouge\">@office<\/code> context:<\/p>\n<div class=\"language-emacs-lisp highlighter-rouge\"><pre>(setq org-agenda-custom-commands \n      '((\"o\" \"At the office\" tags-todo \"@office\"\n         ((org-agenda-overriding-header \"Office\")))))\n<\/pre>\n<\/div>\n<p>Following the GTD principle, what I usually want is to only show the <em>first\naction to be done<\/em> (or next action) for each project with the <code class=\"highlighter-rouge\">@office<\/code> tag.<\/p>\n<p>That can be achieved using a skipping condition:<\/p>\n<div class=\"language-emacs-lisp highlighter-rouge\"><pre>(setq org-agenda-custom-commands \n      '((\"o\" \"At the office\" tags-todo \"@office\"\n         ((org-agenda-overriding-header \"Office\")\n          (org-agenda-skip-function #'my-org-agenda-skip-all-siblings-but-first)))))\n(defun my-org-agenda-skip-all-siblings-but-first ()\n  \"Skip all but the first non-done entry.\"\n  (let (should-skip-entry)\n    (unless (org-current-is-todo)\n      (setq should-skip-entry t))\n    (save-excursion\n      (while (and (not should-skip-entry) (org-goto-sibling t))\n        (when (org-current-is-todo)\n          (setq should-skip-entry t))))\n    (when should-skip-entry\n      (or (outline-next-heading)\n          (goto-char (point-max))))))\n\t\t  \n(defun org-current-is-todo ()\n  (string= \"TODO\" (org-get-todo-state)))\n<\/pre>\n<\/div>\n<p>Creating custom agenda commands can be a bit tricky at first, one easy way is to\ncustomize them via <code class=\"highlighter-rouge\">M-x customize-variable RET org-agenda-custom-commands<\/code>.<\/p>\n<p>To select an agenda command to execute, press <code class=\"highlighter-rouge\">C-c a<\/code>.<\/p>\n<h2 id=\"3-the-somedaymaybe-list\">3. The \u201cSomeday\/Maybe\u201d list<\/h2>\n<p>Did you notice that <code class=\"highlighter-rouge\">someday.org<\/code> is not part of the agenda files set in\n<code class=\"highlighter-rouge\">org-agenda-files<\/code>?<\/p>\n<p>That\u2019s because I do not want to see any entry from this file appearing in my\nagenda buffers, unless I\u2019m doing my weekly review. That\u2019s exactly the purpose of\nthe \u201cSomeday\/Maybe\u201d list.<\/p>\n<p>This file should be reviewed once a week as part of the weekly review (which I\ndo this on Sunday evenings).<\/p>\n<p>During each weekly review, I move projects back and forth between the \u201cactive\u201d\nstate (in <code class=\"highlighter-rouge\">gtd.org<\/code>) and \u201clater\u201d state (in <code class=\"highlighter-rouge\">someday.org<\/code>).<\/p>\n<p>For instance, if a project has moved forward to a certain point, but I know that\nit will stall for a few weeks for some reason, I move it to\n<code class=\"highlighter-rouge\">someday.org<\/code>. During a later weekly review, I\u2019ll move it back to <code class=\"highlighter-rouge\">gtd.org<\/code> when\nit will become active again.<\/p>\n<p>To move projects around, I also use refiling.<\/p>\n<h2 id=\"4-the-tickler\">4. The tickler<\/h2>\n<p>The tickler is one of the best concepts of GTD in my opinion.<\/p>\n<p>Let\u2019s say you will have to pay a bill in a month. You need to write it down in\nyour GTD if you don\u2019t want to miss the deadline. But you also don\u2019t want to be\nreminded of that each and every time you browse through your GTD projects: now\nis not the time to pay it.<\/p>\n<p>That\u2019s where the tickler kicks in: Add an entry to your tickler file with a\ntimestamp, and forget about it!<\/p>\n<p>When time will come, the action will appear in your Org Agenda, reminding you of\nthe bill you have to pay, and all you will have to do is moving it to your\ninbox. Until then, you can just focus on something else and free your mind from\nthis task.<\/p>\n<h2 id=\"references\">References<\/h2>\n<p>All reference documents are put in a <code class=\"highlighter-rouge\">references<\/code> folder next to my Orgmode\nfiles. They are then linked (using <code class=\"highlighter-rouge\">org-store-link<\/code>) from Dired buffers into my\nprojects for quick access.<\/p>\n<p>I also link emails (which I happen to read within Emacs) using <code class=\"highlighter-rouge\">org-store-link<\/code>.<\/p>\n<h2 id=\"archiving\">Archiving<\/h2>\n<p>During my weekly reviews, I archive done projects using <code class=\"highlighter-rouge\">C-c C-x C-a<\/code>\n(<code class=\"highlighter-rouge\">org-archive-subtree-default<\/code>), which moved the entry at point to an archive\nfile.<\/p>\n<p>This way my GTD files remain uncluttered and I never delete any data.<\/p>\n<h2 id=\"conclusion\">Conclusion<\/h2>\n<p>This is a very wide topic, so obviously I haven\u2019t covered everything, but I hope\nthat explains the basics of how I use Orgmode to implement GTD.<\/p>\n<p>This is really only my way of doing it.  Orgmode is such a moldable tool that I\ndon\u2019t think that there are 2 exactly identical setups.<\/p>","created_at":"2018-11-07T16:47:22+0000","updated_at":"2018-11-07T16:47:22+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"emacs.cafe","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":403,"title":"How to get your first 100k active users","url":"https:\/\/news.ycombinator.com\/item?id=15909395","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ryanwaggoner\" class=\"hnuser\">ryanwaggoner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909776\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909911\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=onion2k\" class=\"hnuser\">onion2k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910119\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=firloop\" class=\"hnuser\">firloop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910165\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jjeaff\" class=\"hnuser\">jjeaff<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911431\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=murukesh_s\" class=\"hnuser\">murukesh_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911784\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saagarjha\" class=\"hnuser\">saagarjha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912306\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ch4s3\" class=\"hnuser\">ch4s3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912371\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saagarjha\" class=\"hnuser\">saagarjha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912409\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=WA\" class=\"hnuser\">WA<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912509\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=murukesh_s\" class=\"hnuser\">murukesh_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15921860\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=murukesh_s\" class=\"hnuser\">murukesh_s<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15921848\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saagarjha\" class=\"hnuser\">saagarjha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15926710\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=endless1234\" class=\"hnuser\">endless1234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914926\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saagarjha\" class=\"hnuser\">saagarjha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15916429\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dgregd\" class=\"hnuser\">dgregd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913669\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911363\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=awalton\" class=\"hnuser\">awalton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910499\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jjeaff\" class=\"hnuser\">jjeaff<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911439\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ajdlinux\" class=\"hnuser\">ajdlinux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911599\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TrueGeek\" class=\"hnuser\">TrueGeek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911941\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=onion2k\" class=\"hnuser\">onion2k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913306\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ythn\" class=\"hnuser\">ythn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910009\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910132\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911373\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgeecollins\" class=\"hnuser\">georgeecollins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915494\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910086\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=samschooler\" class=\"hnuser\">samschooler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910259\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jeremyjh\" class=\"hnuser\">jeremyjh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911043\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amitmathew\" class=\"hnuser\">amitmathew<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910261\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GuiA\" class=\"hnuser\">GuiA<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910572\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amitmathew\" class=\"hnuser\">amitmathew<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910797\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jcadam\" class=\"hnuser\">jcadam<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915544\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GuiA\" class=\"hnuser\">GuiA<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15916150\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jcadam\" class=\"hnuser\">jcadam<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15917236\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910384\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amitmathew\" class=\"hnuser\">amitmathew<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910430\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jcun4128\" class=\"hnuser\">jcun4128<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913154\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dhimes\" class=\"hnuser\">dhimes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915009\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=drharby\" class=\"hnuser\">drharby<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15918852\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dhimes\" class=\"hnuser\">dhimes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15922374\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amelius\" class=\"hnuser\">amelius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913698\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LoSboccacc\" class=\"hnuser\">LoSboccacc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913778\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amitmathew\" class=\"hnuser\">amitmathew<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914077\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fbr\" class=\"hnuser\">fbr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914045\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amelius\" class=\"hnuser\">amelius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914207\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910091\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910120\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910218\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=quickthrower2\" class=\"hnuser\">quickthrower2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910243\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adventured\" class=\"hnuser\">adventured<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910175\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910416\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Vinnl\" class=\"hnuser\">Vinnl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912893\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912943\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonas21\" class=\"hnuser\">jonas21<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910411\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910475\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BukhariH\" class=\"hnuser\">BukhariH<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910617\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=warent\" class=\"hnuser\">warent<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910709\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913246\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913001\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=s3cur3\" class=\"hnuser\">s3cur3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912205\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912973\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nemuri\" class=\"hnuser\">nemuri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911266\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912938\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ohtwenty\" class=\"hnuser\">ohtwenty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913057\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jeffmess\" class=\"hnuser\">jeffmess<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913127\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914508\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nudpiedo\" class=\"hnuser\">nudpiedo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913865\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danielskogly\" class=\"hnuser\">danielskogly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914219\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aagha\" class=\"hnuser\">aagha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15940114\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vineet\" class=\"hnuser\">vineet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910632\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pascalxus\" class=\"hnuser\">pascalxus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910527\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=soared\" class=\"hnuser\">soared<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911840\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=inimino\" class=\"hnuser\">inimino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912712\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway46786\" class=\"hnuser\">throwaway46786<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911455\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=notaboutdave\" class=\"hnuser\">notaboutdave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910204\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adventured\" class=\"hnuser\">adventured<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910560\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910703\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wahnfrieden\" class=\"hnuser\">wahnfrieden<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914951\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Caligula\" class=\"hnuser\">Caligula<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911199\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fragsworth\" class=\"hnuser\">fragsworth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911557\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=folli\" class=\"hnuser\">folli<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913265\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wahnfrieden\" class=\"hnuser\">wahnfrieden<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914943\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15916000\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911737\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=meteor333\" class=\"hnuser\">meteor333<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915505\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=giarc\" class=\"hnuser\">giarc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911541\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909521\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jqbx_jason\" class=\"hnuser\">jqbx_jason<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909906\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909938\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jqbx_jason\" class=\"hnuser\">jqbx_jason<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909966\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910074\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911592\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=forkLding\" class=\"hnuser\">forkLding<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909728\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909986\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=p0larboy\" class=\"hnuser\">p0larboy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912430\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jesse_\" class=\"hnuser\">jesse_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913136\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=boyce\" class=\"hnuser\">boyce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914430\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jesse_\" class=\"hnuser\">jesse_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915002\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=boyce\" class=\"hnuser\">boyce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15930539\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15916006\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dna_polymerase\" class=\"hnuser\">dna_polymerase<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910047\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=liquidise\" class=\"hnuser\">liquidise<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910130\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fpisfun\" class=\"hnuser\">fpisfun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910488\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910724\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=samastur\" class=\"hnuser\">samastur<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914187\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910107\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=forkLding\" class=\"hnuser\">forkLding<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909698\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909996\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kashprime\" class=\"hnuser\">kashprime<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911849\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=username223\" class=\"hnuser\">username223<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909866\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dboreham\" class=\"hnuser\">dboreham<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909863\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=samschooler\" class=\"hnuser\">samschooler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910316\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910040\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=evanspa\" class=\"hnuser\">evanspa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915476\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Ologn\" class=\"hnuser\">Ologn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911075\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AWebOfBrown\" class=\"hnuser\">AWebOfBrown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911207\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yeukhon\" class=\"hnuser\">yeukhon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910501\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=abraae\" class=\"hnuser\">abraae<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910599\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yeukhon\" class=\"hnuser\">yeukhon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910806\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fabianuribe\" class=\"hnuser\">fabianuribe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909981\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sebleon\" class=\"hnuser\">sebleon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912574\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=greggarious\" class=\"hnuser\">greggarious<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910017\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tmatthewj\" class=\"hnuser\">tmatthewj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912594\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mkalygin\" class=\"hnuser\">mkalygin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910112\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910180\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mherdeg\" class=\"hnuser\">mherdeg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912629\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914370\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mherdeg\" class=\"hnuser\">mherdeg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15917041\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mkalygin\" class=\"hnuser\">mkalygin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910227\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yakshaving_jgt\" class=\"hnuser\">yakshaving_jgt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913459\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15916032\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=miradu\" class=\"hnuser\">miradu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15909894\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=myf01d\" class=\"hnuser\">myf01d<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910615\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910006\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=40acres\" class=\"hnuser\">40acres<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910490\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=annekate\" class=\"hnuser\">annekate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15910746\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skeet\" class=\"hnuser\">skeet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15915163\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=subirm\" class=\"hnuser\">subirm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15913070\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=intrasight\" class=\"hnuser\">intrasight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911963\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wellboy\" class=\"hnuser\">wellboy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914400\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=agentofoblivion\" class=\"hnuser\">agentofoblivion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15914535\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hrykora\" class=\"hnuser\">hrykora<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911709\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=giarc\" class=\"hnuser\">giarc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912122\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hrykora\" class=\"hnuser\">hrykora<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911694\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=giarc\" class=\"hnuser\">giarc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912132\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sshumaker\" class=\"hnuser\">sshumaker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15911732\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=subirm\" class=\"hnuser\">subirm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912986\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=subirm\" class=\"hnuser\">subirm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=15912719\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:47:21+0000","updated_at":"2018-11-07T16:47:21+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":404,"title":"TheClimateCorporation\/prng","url":"https:\/\/github.com\/TheClimateCorporation\/prng","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Pseudorandom number generators in Clojure.<\/p>\n<p><strong>This project is in pre-release development and subject to change.<\/strong><\/p>\n<h2><a id=\"user-content-concepts\" class=\"anchor\" aria-hidden=\"true\" href=\"#concepts\"><\/a>Concepts<\/h2>\n<p>A pseudorandom number generator (PRNG) is a deterministic algorithm that\ngenerates a sequence of numbers that appear to be random.<\/p>\n<p>A PRNG has the following components:<\/p>\n<ul><li>State space, S: a finite set of possible states.<\/li>\n<li>Output space, U: a finite set of possible outputs (pseudorandom numbers)\nthat can be generated.<\/li>\n<li>Transition function, f: a pure function that takes a state as its\nargument and returns a (usually different) state.<\/li>\n<li>Output function, g: a pure function that maps a state to an element in\nthe output space. Different output functions may be used to generate\ndifferent numeric types (e.g. longs or doubles) from the same\nunderlying sequence of states.<\/li>\n<\/ul><p>The state of a PRNG evolves according to the recurrence relationship\n(LaTeX math markup):<\/p>\n<pre>s_{i} = f(s_{i - 1})\nu_{i} = g(s{i})\n<\/pre>\n<p>where<\/p>\n<pre>f: S \\rightarrow S\n<\/pre>\n<p>is the state transition function and<\/p>\n<pre>g: S \\rightarrow U\n<\/pre>\n<p>is the output function.<\/p>\n<p>The initial state of a PRNG is typically constructed by provided a\n\"seed\" (a source of bits) to a factory function. When initialized with\nthe same seed, a PRNG is guaranteed to subsequently generate exactly the\nsame sequence of numbers.<\/p>\n<p>Most widely used PRNG libraries represent PRNGs as stateful objects. A\ntypical example is the <code>java.util.Random<\/code> class in Java, which\nencapsulates a hidden state. Calling a method to generate the next\npseudorandom number in the sequence has the side effect of mutating this\nhidden state:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def r (java.util.Random. 1234567))\n(.nextDouble r)\n;=&gt; 0.24283348968807417\n; nextDouble is impure! the second call returns a different result!\n(.nextDouble r)\n;=&gt; 0.7520183689538884<\/pre><\/div>\n<p>Such statefulness makes it difficult to reason about concurrency or\nreproducibility. Also, as shown in this example, stateful PRNG libraries\ntypically conflate the transition function and the output function in\na single method.<\/p>\n<p>The <code>prng<\/code> library takes a different approach to PRNGs that is more\nfunctional and Clojure-like. Instead of implicitly mutating a stateful\nobject, the state of the PRNG is directly evolved by calling a\ntransition function that returns the next state, and a pseudorandom\nnumber is directly generated by calling an output function on a state.\nStates are values that can be passed to functions that produce other\nvalues. The transition function and the output function are separated\nconcerns. This functional design is intended to make it easy to reason\nabout concurrency and reproducibility, especially when using PRNGs in\nlarge distributed systems in which coordination is costly.<\/p>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>This is a quick demo that you can follow in a Clojure REPL to get\nacquainted with the <code>prng<\/code> library.<\/p>\n<p>The namespace <code>com.climate.prng.core<\/code> defines a <code>PRNGState<\/code> protocol\nwhose methods can be invoked on all implementing types.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[com.climate.prng.core :as prng])<\/pre><\/div>\n<p>Specific PRNG implementations are found in namespaces under the prefix\n<code>com.climate.prng.generators<\/code>. Currently, two implementations are\navailable:<\/p>\n<ul><li>A 64-bit version of the well-known MT19937 <a href=\"http:\/\/www.math.sci.hiroshima-u.ac.jp\/~m-mat\/MT\/emt.html\" rel=\"nofollow\">Mersenne\nTwister<\/a>\nPRNG<\/li>\n<li>A counter-based PRNG that uses the <a href=\"https:\/\/www.schneier.com\/threefish.html\" rel=\"nofollow\">Threefish block\ncipher<\/a> as the source of\npseudorandomness<\/li>\n<\/ul><p>Try using the Mersenne Twister:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[com.climate.prng.generators.mersenne-twister :as mt])<\/pre><\/div>\n<p>To instantiate a Mersenne Twister state, call a factory function with a\n<code>long[]<\/code> seed argument.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def seed (long-array [0 1 2 3]))\n(def initial-state (mt\/seed-state seed)))<\/pre><\/div>\n<p>The <code>seed-state<\/code> factory function returns a state value that satisfies\nthe <code>PRNGState<\/code> protocol.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(satisfies? prng\/PRNGState initial-state)\n;=&gt; true<\/pre><\/div>\n<p>For a given seed argument, the same return value (\"same\" in the sense of\n<code>equals()<\/code>) is returned:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(= initial-state (mt\/seed-state seed))\n;=&gt; true<\/pre><\/div>\n<p>The <code>PRNGState<\/code> protocol methods are pure functions that take a state\nvalue argument and return other values, without mutating the input state\nvalue.<\/p>\n<p>To sample a long from this PRNG, simply call the <code>-&gt;long<\/code> protocol\nmethod on the state. <code>-&gt;long<\/code> is a pure function; for a given state, it\nalways returns the same value, and it does not mutate its state argument.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(prng\/-&gt;long initial-state)\n;=&gt; -3722652993627521228\n; Multiple calls return the same value.\n(prng\/-&gt;long initial-state)\n;=&gt; -3722652993627521228<\/pre><\/div>\n<p>Similarly, the <code>-&gt;double<\/code> protocol method returns a double:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(prng\/-&gt;double initial-state)\n;=&gt; 0.4049867569919572<\/pre><\/div>\n<p>The <code>next-state<\/code> protocol method takes a state argument and returns a\nnew state that is one step ahead in the state sequence of the PRNG. The\ninput state is not mutated.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def new-state (prng\/next-state initial-state))\n(= new-state initial-state)\n;=&gt; false\n; Get a long from this new state\n(prng\/-&gt;long new-state)\n;=&gt; -909579606142348662\n; Get a long from the old state (compare to previous calls)\n(prng\/-&gt;long initial-state)\n;=&gt; -3722652993627521228<\/pre><\/div>\n<p>The <code>jump-state<\/code> protocol method takes two arguments, a state and a jump\nsize, and jumps ahead by the provided jump size. <code>jump-state<\/code> returns\nthe same state value as iterative composition of <code>next-state<\/code>, but it is\nmuch faster.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def jump-size 1000000)\n(def state-after-jump (prng\/jump-state initial-state jump-size))\n(= state-after-jump ((apply comp (repeat jump-size prng\/next-state)) initial-state))\n;=&gt; true<\/pre><\/div>\n<h2><a id=\"user-content-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#tests\"><\/a>Tests<\/h2>\n<p><code>lein test<\/code> runs unit tests.<\/p>\n<p><code>lein test :benchmark<\/code> runs tests with the <code>:benchmark<\/code> test selector,\nwhich are a few performance benchmarks. They may take a while to\ncomplete.<\/p>\n<h2><a id=\"user-content-known-issues\" class=\"anchor\" aria-hidden=\"true\" href=\"#known-issues\"><\/a>Known issues<\/h2>\n<h3><a id=\"user-content-concurrency\" class=\"anchor\" aria-hidden=\"true\" href=\"#concurrency\"><\/a>Concurrency<\/h3>\n<p>The Java classes <code>MersenneTwisterState<\/code> and <code>ThreefishState<\/code> have public\nmethods that are untested for thread safety and are probably unsafe.\nAlthough the <code>PRNGState<\/code> protocol is designed for concurrency, it\npossible to bypass the protocol and directly invoke the unsafe methods\nof the implementation classes. To ensure thread safety, these\nimplementation classes ought to be re-implemented with Clojure-style\npersistent data structures.<\/p>\n<h3><a id=\"user-content-speed\" class=\"anchor\" aria-hidden=\"true\" href=\"#speed\"><\/a>Speed<\/h3>\n<p>For the <code>MersenneTwisterState<\/code> implementation class, the <code>jump-state<\/code>\nprotocol method uses a naive algorithm to compute jump-ahead from the\nremainder polynomial. Maromoto and colleauges <a href=\"http:\/\/dx.doi.org\/10.1007\/978-3-540-85912-3_26\" rel=\"nofollow\">demonstrated two\nalternative algorithms<\/a>\nto compute jump-ahead more quickly. The <code>jump-state<\/code> method ought to be\nreimplemented with one of these faster algorithms.<\/p>\n<p>The implementation of the <code>jump-state<\/code> method for <code>ThreefishState<\/code> uses\nkludgy <code>BigInteger<\/code> and hexadecimal string manipulations to perform\nbinary arithmetic. Performance can almost certainly be improved by\nthoughtful bit twiddling.<\/p>\n<h3><a id=\"user-content-pseudorandomness\" class=\"anchor\" aria-hidden=\"true\" href=\"#pseudorandomness\"><\/a>Pseudorandomness<\/h3>\n<p>The Mersenne Twister is a well-known, high-quality PRNG that has been\nthoroughly tested for pseudorandomness. The implementation in this\nlibrary is a bit-exact match to the reference implementation in C.<\/p>\n<p>The counter-based Threefish PRNG has not been as well tested, and its\npseudorandomness ought to be teststed with a suite of statistical tests.<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2014 The Climate Corporation<\/p>\n<p>Distributed under the Apache License, Version 2.0<\/p>\n<\/article>","created_at":"2018-11-07T16:47:21+0000","updated_at":"2018-11-07T16:47:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1661603?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":401,"title":"Emacsair! Magit, the magical Git interface","url":"https:\/\/emacsair.me\/2017\/09\/01\/the-magical-git-interface\/","content":"Emacsair! Magit, the magical Git interface\n    <header>\n      <h4>Collecting frequent-committer miles<\/h4>\n    <\/header><div id=\"prelude\"><h3 id=\"what-is-magit\">What is Magit?<\/h3><p><a href=\"https:\/\/magit.vc\">Magit<\/a> is a text-based <a href=\"https:\/\/git-scm.com\">Git<\/a> user interface that puts an unmatched\nfocus on streamlining workflows.  Commands are invoked using short\nmnemonic key sequences that take the cursor\u2019s position in the highly\nactionable interface into account to provide context-sensitive\nbehavior.<\/p><p>With Magit you can do nearly everything that you can do when using Git\non the command-line, but at greater speed and while taking advantage\nof advanced features that previously seemed too daunting to use on a\ndaily basis.  Many users will find that by using Magit they can become\nmore effective Git user.<\/p><h3 id=\"the-crowd-funding-campaign\">The crowd-funding campaign<\/h3><p>Please consider supporting the\nongoing <a href=\"https:\/\/www.kickstarter.com\/projects\/1681258897\/its-magit-the-magical-git-client?ref=a4ifug\">crowd-funding campaign<\/a>.  If it succeeds, then I\ncan work full-time on Magit for a whole year.  I am still overflowing\nwith ideas, and depend on your support to realize them, and to address\nthe remaining long-standing issues.  Thank you!<\/p><h3 id=\"related-articles\">Related articles<\/h3><p>On the occasion of the campaign I have written several articles\noutlining what separates Magit from other Git clients and what my\nplans for its future are.  You are currently reading one of those\narticles, but maybe <a href=\"https:\/\/emacsair.me\/2017\/09\/01\/campaign-articles#start\">one of the others<\/a> is a more suitable\nstarting point for you.<\/p><\/div>\n<p><a name=\"start\" id=\"start\">\n<br \/><\/a><a title=\"visit Magit's homepage\" href=\"https:\/\/magit.vc\">\n  <img id=\"magit-logo\" src=\"https:\/\/magit.vc\/assets\/magit-168x200px.png\" alt=\"image\" \/><\/a><\/p>\n<h3 id=\"basic-similarities-and-differences-to-other-interfaces\">Basic similarities and differences to other interfaces<\/h3>\n<p>Magit is a text-based user interface to Git.  It looks similar to the\noutput of <code class=\"highlighter-rouge\">git<\/code> itself.  Unlike when using <code class=\"highlighter-rouge\">git<\/code> directly, a Magit\nuser does not execute commands by typing them into a shell.  The main\nentry point to Magit isn\u2019t the command-line but its status buffer,\nwhich looks like this:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status.png\" alt=\"\" \/><\/p>\n<p>Unlike the read-only output of <code class=\"highlighter-rouge\">git<\/code>, everything that you can see in\nthe status buffer, or any other Magit buffer, is actionable.  Commands\nare invoked, not by typing them out, but by pressing short mnemonic\nkey sequences.  Many commands act on the thing the cursor is on,\neither by showing more detailed or alternative information about it in\na separate buffer or by performing an action with side-effects on it.<\/p>\n<p>Git GUIs are similar to Magit in that many things you can see in a GUI\ncan also be acted on directly.  Of course in a GUI you would normally\nclick on something to act on it, while Magit is primarily<sup id=\"fnref:1\"><a href=\"#fn:1\" class=\"footnote\">1<\/a><\/sup> driven\nusing the keyboard.  But the differences between Magit and Git GUIs go\nmuch further than just looks and input methods.<\/p>\n<p>Current GUIs only expose a limited subset of Git\u2019s features, while\nMagit is an essentially complete interface.  And many of the actions\nthat GUIs do expose tend to only by available in some view or dialog\nbut not another, even though the object that the feature would act on\nis visible in both.  In Magit, on the other hand, you can do virtually\neverything with some object that can be done with objects of that type\nin any other contexts.<\/p>\n<p>So one could say that Magit combines the completeness of Git with the\ninteractive nature of a GUI, while not limiting the availability of\nfeatures to certain specialized views.  The result is an interface\nthat, with a little practice, is both faster and more intuitive than\neither the command line or a GUI.<\/p>\n<h3 id=\"what-separates-magit-from-other-git-interfaces\">What separates Magit from other Git interfaces?<\/h3>\n<p>Magit heavily relies on a few simple, yet powerful interface concepts.\nWhat makes their application in Magit powerful isn\u2019t necessarily (or\nonly) some innovative approach, but the fact that these concepts are\napplied very consistently across all features.  Let\u2019s look at the most\nimportant of these concepts in more detail.<\/p>\n<h4 id=\"in-magit-all-output-is-actionable\">In Magit all output is actionable<\/h4>\n<p>Magit looks quite similar to the Git command line, prettier maybe, but\nMagit remains a text based interface.  It is worth repeating that the\ncrucial difference between Git and Magit is that Magit\u2019s output is not\n\u201cread-only\u201d.<\/p>\n<p>Instead all status information is automatically kept up-to-date and\nrelevant to the task at hand.  And most importantly everything that\ncan be seen is actionable.  Furthermore, unlike in GUI interfaces,\nevery action you might reasonably (based on your knowledge of what Git\nitself can do) want to perform on some visible (or invisible) thing\nactually is available, using only a single or very few mnemonic key\npresses in most cases.<\/p>\n<h4 id=\"magit-is-essentially-complete\">Magit is essentially complete<\/h4>\n<p>By now, Magit is an essentially complete Git interface.  That does not\nonly mean that it supports most Git features \u2014 most features\nthat are supported are also improved upon.<\/p>\n<p>The reason that there still exist some gaps is that fully integrating\nsupported features into the interface, by taking advantage of Magit\u2019s\ninterface concepts whenever possible, is more important than trying to\nachieve completeness for its own sake.<\/p>\n<p>If I had to put a number on it, I would say that Magit supports about\n90% of what can be done using <code class=\"highlighter-rouge\">git<\/code>, while other third-party\ninterfaces are stuck somewhere around 60%.  But what is more important\nthan the relative completeness at any given time is that support for\nall of Git features could be added to Magit without any changes to its\nuser interface concepts.<\/p>\n<p>Magit enables even advanced Git users to perform nearly all of their\ndaily version control tasks with a fraction of the effort that would\nbe required on the command line, without having to drop to that\ncommand line for \u201cedge cases\u201d as would be necessary with many other\ninterfaces.<\/p>\n<h4 id=\"magit-will-grow-with-you\">Magit will grow with you<\/h4>\n<p>Magit does more than make it possible to use the commands that you\nalready know more quickly than on the command line.  By offering\nspecialized variants of common commands, optimized for certain common\nwork-flow variations, it also gently encourages its users to learn\nmore advanced concepts.<\/p>\n<p>Like Git and unlike many graphical interfaces to it, Magit grows with\nyou as you move beyond the basics and start performing more involved\nversion control tasks.  Until you do, the more advanced features do\nnot get in the way.<\/p>\n<p>Magit encourages (but never forces) its users to use Git features that\nmany users of the command line would deem too advanced and\/or too\ncumbersome to be used except in special circumstances.  Many of these\nfeatures simply are not available when using other GUI and TUI\ninterfaces.  Magit makes advanced features just as accessible as the\nbasic features, which even the less than enthusiastic Git user cannot\nget around learning eventually.<\/p>\n<p>With GUI interfaces many users at some point reach a level of Git\nproficiency that forces them to drop back to the command line on a\nregular basis even for things they do every day.  With Magit this will\nonly happen very rarely and only for things that even advanced users\nonly do irregularly.<\/p>\n<p>More advanced actions are not harder to perform than the trivial ones.\nWith Git\u2019s porcelain the user not only has to learn a new concept, but\noften s\/he also has to deal with a more complex interface that comes\nwith the advanced task.  With Magit the user of course also has to\nunderstand the concept and learn to deal with the more complicated\ncases, but the interface for an advanced task is seldom more complex\nthan that of a trivial one.<\/p>\n<h4 id=\"magit-has-co-evolved-with-its-users\">Magit has co-evolved with its users<\/h4>\n<p>Initially, users new to Magit will likely only use a few basic\nfeatures and work-flows that they are familiar to them based on prior\nexposure to <code class=\"highlighter-rouge\">git<\/code> on the command line and to other GUIs and TUIs.<\/p>\n<p>However, for reasons given before, most users will eventually start\nusing features that previously seemed too daunting or cumbersome to\nuse.  Another reason why doing so comes naturally is that Magit has\nco-evolved with its users.<\/p>\n<p>Magit has been around for many years, and historically many features\nthat are now being taken for granted did not exist.  Many features,\nespecially those concerning streamlined work-flows, were originally\nproposed by users who enjoyed and had grown accustomed to some\nexisting feature, only to find themselves wondering why some other\nfeature was not as convenient to use.<\/p>\n<p>This has been going on for many years now, so if you start using Magit\nnow, then chances are good that most work-flow improvements you might\ncome to desire have already been implemented.  And because those\nfeatures had seemed like obvious intentions at the time they were\nfirst implemented, it might even be the case that they work exactly as\nyou would have expected.<\/p>\n<p>This process has progressed very far, but it is not over.  If you are\nmissing some feature, then just ask.  If it is a simple feature, then\nchances are good that that in a few days it won\u2019t be missing anymore.\nI hope that by encouraging non-Emacs users to use Magit, we will also\nget some fresh ideas.<\/p>\n<p>Despite all this, it would be wrong to assume that Magit started out\nwith a predefined set of interface concepts and abstractions, and all\nthat was left to be done was incrementally filling in the gaps.<\/p>\n<h4 id=\"magit-streamlines-work-flows\">Magit streamlines work-flows<\/h4>\n<p>Magit does not \u2014 and maybe this is what ultimately separates it from\nthe competition \u2014 restrict users to a certain set of simple features\nand tasks.  Magit supports and streamlines the use of Git features\nthat many users and developers of other Git clients apparently thought\ncould not be reasonably mapped to a non-command-line interface.<\/p>\n<p>While Git offers many commands and arguments that perform a wide\nvariety of tasks, Magit goes one step further by not only preserving\nthat flexibility but by also providing higher-level commands that are\nnot so much bound by the (wonderful) underlying data-model and whose\ndesign is instead primarily oriented on the actual tasks that users\noften have to perform.<\/p>\n<p>The output of Git commands can guide the user in their work, but on\nthe command-line users have to extract the relevant information\nthemselves, mentally and by manually feeding it to the next command.\nMagit commands automates that manual step by taking advantage of the\ninterface context and the state of the repository to guide their\noperation, which not only speeds up the process but also frees mental\ncapacities.<\/p>\n<p>Additionally Magit provides many \u201cvariants\u201d of often used commands\nthat are optimized for particular use-cases, as well as dwim (do what\nI mean) variants that very reliably guess the user\u2019s intention based\non the cursor\u2019s position.  (In cases when the user wishes to act on\nsomething other than the thing under the cursor or if s\/he wants to\nfine-tune the behavior using certain <code class=\"highlighter-rouge\">git<\/code> arguments explicitly, then\nthat is of course also possible.)<\/p>\n<div class=\"footnotes\"><ol><li id=\"fn:1\">\n      <p>Better mouse support, including drop-down menus, is one of the\n  planned features.\u00a0<a href=\"#fnref:1\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>\n    \n    <noscript><p><img src=\"https:\/\/tarsius.piwikpro.com\/piwik.php?idsite=5\" alt=\"\" \/><\/p><\/noscript>","created_at":"2018-11-07T16:47:18+0000","updated_at":"2018-11-07T16:47:18+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"emacsair.me","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":402,"title":"discourse\/discourse","url":"https:\/\/github.com\/discourse\/discourse","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"http:\/\/www.discourse.org\/\" rel=\"nofollow\"><img src=\"https:\/\/github.com\/discourse\/discourse\/raw\/master\/images\/discourse.png\" alt=\"Logo\" \/><\/a><\/p>\n<p>Discourse is the 100% open source discussion platform built for the next decade of the Internet. Use it as a:<\/p>\n<ul><li>mailing list<\/li>\n<li>discussion forum<\/li>\n<li>long-form chat room<\/li>\n<\/ul><p>To learn more about the philosophy and goals of the project, <a href=\"http:\/\/www.discourse.org\" rel=\"nofollow\">visit <strong>discourse.org<\/strong><\/a>.<\/p>\n<h2><a id=\"user-content-screenshots\" class=\"anchor\" aria-hidden=\"true\" href=\"#screenshots\"><\/a>Screenshots<\/h2>\n<p><a href=\"https:\/\/bbs.boingboing.net\" rel=\"nofollow\"><img alt=\"Boing Boing\" src=\"https:\/\/cloud.githubusercontent.com\/assets\/1385470\/25397876\/3fe6cdac-29c0-11e7-8a41-9d0c0279f5a3.png\" width=\"720px\" \/><\/a>\n<a href=\"https:\/\/twittercommunity.com\/\" rel=\"nofollow\"><img src=\"https:\/\/cloud.githubusercontent.com\/assets\/1385470\/25397920\/71b24e4c-29c0-11e7-8bcf-7a47b888412e.png\" width=\"720px\" alt=\"image\" \/><\/a>\n<a href=\"http:\/\/discuss.howtogeek.com\" rel=\"nofollow\"><img src=\"https:\/\/cloud.githubusercontent.com\/assets\/1385470\/25398049\/f0995962-29c0-11e7-99d7-a3b9c4f0b357.png\" width=\"720px\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/talk.turtlerockstudios.com\/\" rel=\"nofollow\"><img src=\"https:\/\/cloud.githubusercontent.com\/assets\/1385470\/25398115\/2d560d96-29c1-11e7-9a96-b0134a4fedff.png\" width=\"720px\" alt=\"image\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/2c40bee0087e6374f61df8defd60102d1e86578d\/68747470733a2f2f7777772e646973636f757273652e6f72672f612f696d672f61626f75742f6d6f62696c652d646576696365732d32782e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/2c40bee0087e6374f61df8defd60102d1e86578d\/68747470733a2f2f7777772e646973636f757273652e6f72672f612f696d672f61626f75742f6d6f62696c652d646576696365732d32782e6a7067\" alt=\"Mobile\" width=\"414\" data-canonical-src=\"https:\/\/www.discourse.org\/a\/img\/about\/mobile-devices-2x.jpg\" \/><\/a><\/p>\n<p>Browse <a href=\"https:\/\/www.discourse.org\/customers\" rel=\"nofollow\">lots more notable Discourse instances<\/a>.<\/p>\n<h2><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h2>\n<p>To get your environment setup, follow the community setup guide for your operating system.<\/p>\n<ol><li>If you're on macOS, try the <a href=\"https:\/\/meta.discourse.org\/t\/beginners-guide-to-install-discourse-on-macos-for-development\/15772\" rel=\"nofollow\">macOS development guide<\/a>.<\/li>\n<li>If you're on Ubuntu, try the <a href=\"https:\/\/meta.discourse.org\/t\/beginners-guide-to-install-discourse-on-ubuntu-for-development\/14727\" rel=\"nofollow\">Ubuntu development guide<\/a>.<\/li>\n<li>If you're on Windows, try the <a href=\"https:\/\/meta.discourse.org\/t\/beginners-guide-to-install-discourse-on-windows-10-for-development\/75149\" rel=\"nofollow\">Windows 10 development guide<\/a>.<\/li>\n<\/ol><p>If you're familiar with how Rails works and are comfortable setting up your own environment, you can also try out the <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/docs\/DEVELOPER-ADVANCED.md\"><strong>Discourse Advanced Developer Guide<\/strong><\/a>, which is aimed primarily at Ubuntu and macOS environments.<\/p>\n<p>Before you get started, ensure you have the following minimum versions: <a href=\"http:\/\/www.ruby-lang.org\/en\/downloads\/\" rel=\"nofollow\">Ruby 2.4+<\/a>, <a href=\"http:\/\/www.postgresql.org\/download\/\" rel=\"nofollow\">PostgreSQL 10+<\/a>, <a href=\"http:\/\/redis.io\/download\" rel=\"nofollow\">Redis 2.6+<\/a>. If you're having trouble, please see our <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/docs\/TROUBLESHOOTING.md\"><strong>TROUBLESHOOTING GUIDE<\/strong><\/a> first!<\/p>\n<h2><a id=\"user-content-setting-up-discourse\" class=\"anchor\" aria-hidden=\"true\" href=\"#setting-up-discourse\"><\/a>Setting up Discourse<\/h2>\n<p>If you want to set up a Discourse forum for production use, see our <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/docs\/INSTALL.md\"><strong>Discourse Install Guide<\/strong><\/a>.<\/p>\n<p>If you're looking for business class hosting, see <a href=\"https:\/\/www.discourse.org\/buy\/\" rel=\"nofollow\">discourse.org\/buy<\/a>.<\/p>\n<h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<p>Discourse is built for the <em>next<\/em> 10 years of the Internet, so our requirements are high:<\/p>\n<table><thead><tr><th>Browsers<\/th>\n<th>Tablets<\/th>\n<th>Phones<\/th>\n<\/tr><\/thead><tbody><tr><td>Safari 6.1+<\/td>\n<td>iPad 3+<\/td>\n<td>iOS 8+<\/td>\n<\/tr><tr><td>Google Chrome 32+<\/td>\n<td>Android 4.3+<\/td>\n<td>Android 4.3+<\/td>\n<\/tr><tr><td>Internet Explorer 11+<\/td>\n<\/tr><tr><td>Firefox 27+<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-built-with\" class=\"anchor\" aria-hidden=\"true\" href=\"#built-with\"><\/a>Built With<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/rails\/rails\">Ruby on Rails<\/a> \u2014 Our back end API is a Rails app. It responds to requests RESTfully in JSON.<\/li>\n<li><a href=\"https:\/\/github.com\/emberjs\/ember.js\">Ember.js<\/a> \u2014 Our front end is an Ember.js app that communicates with the Rails API.<\/li>\n<li><a href=\"http:\/\/www.postgresql.org\/\" rel=\"nofollow\">PostgreSQL<\/a> \u2014 Our main data store is in Postgres.<\/li>\n<li><a href=\"http:\/\/redis.io\/\" rel=\"nofollow\">Redis<\/a> \u2014 We use Redis as a cache and for transient data.<\/li>\n<\/ul><p>Plus <em>lots<\/em> of Ruby Gems, a complete list of which is at <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/Gemfile\">\/master\/Gemfile<\/a>.<\/p>\n<h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p><a href=\"https:\/\/travis-ci.org\/discourse\/discourse\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/4147ebf5cecef41dc085af01a2783a88b630da07\/68747470733a2f2f6170692e7472617669732d63692e6f72672f646973636f757273652f646973636f757273652e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/api.travis-ci.org\/discourse\/discourse.svg?branch=master\" \/><\/a><\/p>\n<p>Discourse is <strong>100% free<\/strong> and <strong>open source<\/strong>. We encourage and support an active, healthy community that\naccepts contributions from the public \u2013 including you!<\/p>\n<p>Before contributing to Discourse:<\/p>\n<ol><li>Please read the complete mission statements on <a href=\"http:\/\/www.discourse.org\" rel=\"nofollow\"><strong>discourse.org<\/strong><\/a>. Yes we actually believe this stuff; you should too.<\/li>\n<li>Read and sign the <a href=\"http:\/\/discourse.org\/cla\" rel=\"nofollow\"><strong>Electronic Discourse Forums Contribution License Agreement<\/strong><\/a>.<\/li>\n<li>Dig into <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/CONTRIBUTING.md\"><strong>CONTRIBUTING.MD<\/strong><\/a>, which covers submitting bugs, requesting new features, preparing your code for a pull request, etc.<\/li>\n<li>Always strive to collaborate <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/docs\/code-of-conduct.md\">with mutual respect<\/a>.<\/li>\n<li>Not sure what to work on? <a href=\"http:\/\/meta.discourse.org\/t\/so-you-want-to-help-out-with-discourse\/3823\" rel=\"nofollow\"><strong>We've got some ideas.<\/strong><\/a><\/li>\n<\/ol><p>We look forward to seeing your pull requests!<\/p>\n<h2><a id=\"user-content-security\" class=\"anchor\" aria-hidden=\"true\" href=\"#security\"><\/a>Security<\/h2>\n<p>We take security very seriously at Discourse; all our code is 100% open source and peer reviewed. Please read <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/docs\/SECURITY.md\">our security guide<\/a> for an overview of security measures in Discourse, or if you wish to report a security issue.<\/p>\n<h2><a id=\"user-content-the-discourse-team\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-discourse-team\"><\/a>The Discourse Team<\/h2>\n<p>The original Discourse code contributors can be found in <a href=\"https:\/\/github.com\/discourse\/discourse\/blob\/master\/docs\/AUTHORS.md\"><strong>AUTHORS.MD<\/strong><\/a>. For a complete list of the many individuals that contributed to the design and implementation of Discourse, please refer to <a href=\"http:\/\/blog.discourse.org\/2013\/02\/the-discourse-team\/\" rel=\"nofollow\">the official Discourse blog<\/a> and <a href=\"https:\/\/github.com\/discourse\/discourse\/contributors\">GitHub's list of contributors<\/a>.<\/p>\n<h2><a id=\"user-content-copyright--license\" class=\"anchor\" aria-hidden=\"true\" href=\"#copyright--license\"><\/a>Copyright \/ License<\/h2>\n<p>Copyright 2014 - 2018 Civilized Discourse Construction Kit, Inc.<\/p>\n<p>Licensed under the GNU General Public License Version 2.0 (or later);\nyou may not use this work except in compliance with the License.\nYou may obtain a copy of the License in the LICENSE file, or at:<\/p>\n<p><a href=\"http:\/\/www.gnu.org\/licenses\/old-licenses\/gpl-2.0.txt\" rel=\"nofollow\">http:\/\/www.gnu.org\/licenses\/old-licenses\/gpl-2.0.txt<\/a><\/p>\n<p>Unless required by applicable law or agreed to in writing, software\ndistributed under the License is distributed on an \"AS IS\" BASIS,\nWITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\nSee the License for the specific language governing permissions and\nlimitations under the License.<\/p>\n<p>Discourse logo and \u201cDiscourse Forum\u201d \u00ae, Civilized Discourse Construction Kit, Inc.<\/p>\n<h2><a id=\"user-content-dedication\" class=\"anchor\" aria-hidden=\"true\" href=\"#dedication\"><\/a>Dedication<\/h2>\n<p>Discourse is built with <a href=\"http:\/\/www.youtube.com\/watch?v=Xe1TZaElTAs\" rel=\"nofollow\">love, Internet style.<\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:47:18+0000","updated_at":"2018-11-07T16:47:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/3220138?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":399,"title":"Cider locks up Emacs regularly after cider-connect \u00b7 Issue #1758 \u00b7 clojure-emacs\/cider","url":"https:\/\/github.com\/clojure-emacs\/cider\/issues\/1758","content":"Cider locks up Emacs regularly after cider-connect \u00b7 Issue #1758 \u00b7 clojure-emacs\/cider \u00b7 GitHub\n    \n  \n  \n    \n  \n        \n  <div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n    \n  \n  \n  <details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark\" open=\"open\"><summary aria-haspopup=\"dialog\" aria-label=\"Close dialog\">\n      \n    <\/summary><\/details>\n<p>Press h to open a hovercard with more details.<\/p>","created_at":"2018-11-07T16:46:55+0000","updated_at":"2018-11-07T16:46:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/3483575?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":400,"title":"Emacsair! A walk through the Magit interface","url":"https:\/\/emacsair.me\/2017\/09\/01\/magit-walk-through\/","content":"Emacsair! A walk through the Magit interface\n    <header>\n      <h4>Collecting frequent-committer miles<\/h4>\n    <\/header><div id=\"prelude\"><h3 id=\"what-is-magit\">What is Magit?<\/h3><p><a href=\"https:\/\/magit.vc\">Magit<\/a> is a text-based <a href=\"https:\/\/git-scm.com\">Git<\/a> user interface that puts an unmatched\nfocus on streamlining workflows.  Commands are invoked using short\nmnemonic key sequences that take the cursor\u2019s position in the highly\nactionable interface into account to provide context-sensitive\nbehavior.<\/p><p>With Magit you can do nearly everything that you can do when using Git\non the command-line, but at greater speed and while taking advantage\nof advanced features that previously seemed too daunting to use on a\ndaily basis.  Many users will find that by using Magit they can become\nmore effective Git user.<\/p><h3 id=\"the-crowd-funding-campaign\">The crowd-funding campaign<\/h3><p>Please consider supporting the\nongoing <a href=\"https:\/\/www.kickstarter.com\/projects\/1681258897\/its-magit-the-magical-git-client?ref=a4ifug\">crowd-funding campaign<\/a>.  If it succeeds, then I\ncan work full-time on Magit for a whole year.  I am still overflowing\nwith ideas, and depend on your support to realize them, and to address\nthe remaining long-standing issues.  Thank you!<\/p><h3 id=\"related-articles\">Related articles<\/h3><p>On the occasion of the campaign I have written several articles\noutlining what separates Magit from other Git clients and what my\nplans for its future are.  You are currently reading one of those\narticles, but maybe <a href=\"https:\/\/emacsair.me\/2017\/09\/01\/campaign-articles#start\">one of the others<\/a> is a more suitable\nstarting point for you.<\/p><\/div>\n<p><a name=\"start\" id=\"start\">\n<br \/><\/a><a title=\"visit Magit's homepage\" href=\"https:\/\/magit.vc\">\n  <img id=\"magit-logo\" src=\"https:\/\/magit.vc\/assets\/magit-168x200px.png\" alt=\"image\" \/><\/a><\/p>\n<p>This article demonstrates some of Magit\u2019s most essential features in\norder to give you an impression of how the interface works.  It also\nhints at some of the design principals behind that interface.<\/p>\n<p>But this is only the tip of the iceberg.  Magit is a complete\ninterface to Git, which does not limit itself to the \u201cmost essential\nfeatures everyone needs\u201d.  I hope that this article succeeds in\ndemonstrating how Magit\u2019s focus on work-flows allows its users to\nbecome more effective Git users.  Here we concentrate on some\nessential work-flows, but note that more advanced features and\nwork-flows have been optimized with the same attention to detail.<\/p>\n<p>If you would rather concentrate on the big picture, then read the\narticle <a href=\"https:\/\/emacsair.me\/2017\/09\/01\/the-magical-git-interface#start\">Magit the magical Git interface<\/a> instead (or\nafterwards).<\/p>\n<h2 id=\"in-magit-everything-is-actionable\">In Magit everything is actionable<\/h2>\n<p>Almost everything you see in a Magit buffer can be acted on by\npressing some key, but that\u2019s not obvious from just seeing how Magit\nlooks.  The screenshots below are accompanied by text explaining how\nwhat you see can be used to perform a variety of actions.<\/p>\n<p>Regardless of where in a Magit buffer you are, you can always show\nmore details about (or an alternative view of) the thing at point\n<em>without<\/em> having to type or copy-paste any information, as you often\nwould have to do on the command line.<\/p>\n<p>The status buffer, which can be shown while inside a Git repository by\ntyping <i>C-x g<\/i> (<i>Control+x<\/i> followed by <i>g<\/i> <sup id=\"fnref:1\"><a href=\"#fn:1\" class=\"footnote\">1<\/a><\/sup>), is\nMagit\u2019s equivalent of typing <code class=\"highlighter-rouge\">git status<\/code> in a shell.  It shows a\nquick overview of the current repository.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status.png\" alt=\"\" \/><\/p>\n<p>As you can see, Magit shows more information than Git.  On the command\nline, you would have to also use <code class=\"highlighter-rouge\">git diff<\/code>, <code class=\"highlighter-rouge\">git diff --cached<\/code>, <code class=\"highlighter-rouge\">git\nlog --oneline origin\/master..<\/code>, <code class=\"highlighter-rouge\">git log --oneline ..origin\/master<\/code>,\nand a few other commands to get the same information.<\/p>\n<h3 id=\"other-emacs-themes\">Other Emacs themes<\/h3>\n<p>Many themes exist for Emacs.  Throughout this guide we use\nthe <a href=\"https:\/\/github.com\/bbatsov\/solarized-emacs\">Solarized light<\/a> theme, but here is how the status\nbuffer looks when using some other popular themes:<\/p>\n<h4 id=\"solarized-dark\">Solarized dark<\/h4>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status-solarized-dark.png\" alt=\"\" \/><a href=\"https:\/\/github.com\/bbatsov\/solarized-emacs\">Homepage<\/a><\/p>\n<h4 id=\"zenburn\">Zenburn<\/h4>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status-zenburn.png\" alt=\"\" \/><a href=\"https:\/\/github.com\/bbatsov\/zenburn-emacs\">Homepage<\/a><\/p>\n<h4 id=\"light-emacs-default\">Light Emacs default<\/h4>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status-light-default.png\" alt=\"\" \/><\/p>\n<h4 id=\"dark-emacs-default\">Dark Emacs default<\/h4>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status-dark-default.png\" alt=\"\" \/><\/p>\n<h2 id=\"hiding-details\">Hiding details<\/h2>\n<p>You might have noticed that some diff hunks, those for <code class=\"highlighter-rouge\">README.md<\/code>,\nare not shown.  That\u2019s because the respective sections have been\ncollapsed.  You can collapse diff sections and any other section\n(which are larger than a single line) by moving the cursor into it and\nthen pressing <i>TAB<\/i>.<\/p>\n<p>Note that the current section is highlighted with a different\nbackground color so that you always know the extent of the current\nsection.  If you moved up one line, \u201cUnpulled from origin\/master\u201d\nwould become the current section.  The current line as well as all the\ncommits that follow would be highlighted<\/p>\n<p>If you don\u2019t currently need to know any details about the unpulled\ncommits, then you could now type <i>TAB<\/i>.  After doing so you still\nsee the name of the upstream branch and how many commits it is ahead\nof <code class=\"highlighter-rouge\">master<\/code>, but the details (the individual commits in this case) are\nhidden.<\/p>\n<p>If you want to see them, just type <i>TAB<\/i> again.  On the command\nline you would have to type something like <code class=\"highlighter-rouge\">git log --oneline\n..@{upstream}<\/code>.<\/p>\n<p>In Magit\u2019s status buffer you can always show as much or little as\ncurrently appropriate without having to remember both the command that\nshows little or much information.  Instead you can temporally hide or\nshow details as needed by pressing a single key.<\/p>\n<p>After hiding as much as possible, a status buffer looks like this:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status-hidden.png\" alt=\"\" \/><\/p>\n<p>Here is the same buffer, showing a little more than that:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/status-more.png\" alt=\"\" \/><\/p>\n<p><em>If you use Emacs to edit and save a file which is located inside the\nrepository or if you use Magit to perform some action such as staging\na change, then the status buffer is automatically updated.<\/em><\/p>\n<p>And if you change something outside of Emacs, then you can press\n<i>g<\/i> to refresh the current status buffer or <i>G<\/i> to refresh\nall Magit buffers.<\/p>\n<h2 id=\"acting-on-what-you-see\">Acting on what you see<\/h2>\n<p><em>A major advantage Magit has over Git on the command line is that\nnearly everything you see in a Magit buffer can be acted on.  Hiding\nand showing a section is just one example of that.<\/em><\/p>\n<p>Besides <i>TAB<\/i>, another key that works nearly everywhere is\n<i>RET<\/i> (<i>return<\/i>).  It shows a more detailed view of the\nthing at point, be that a commit, branch, diff, hunk, stash, \u2026<\/p>\n<h2 id=\"commit-details\">Commit details<\/h2>\n<p>For example, if you type <i>RET<\/i> while the cursor is on a commit,\nthen the message and diff for that commit are shown.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/commit-top.png\" alt=\"\" \/><\/p>\n<p>In this case the diff does not completely fit into the window.  You\ncould hide some sections \u2014 the metadata, commit, and diffstat at\nthe top \u2014 to bring the complete diff into view, or you could\nmove down a few sections to view the rest of the diff.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/commit-bottom.png\" alt=\"\" \/><\/p>\n<p>The basic keys movement keys are <i>C-p<\/i> to move up one section,\nand <i>C-n<\/i> to move down one section.  <i>M-p<\/i>\n(i.e. <i>Meta+p<\/i>, also commonly known as <i>Alt+p<\/i>) and\n<i>M-n<\/i> would move to the previous or next section on the same\nlevel as the current section.  The latter two commands can be used to\nnavigate Magit buffers more quickly by skipping child sections.  For\nexample, you could jump from one file to another without having to\nstep through the hunks of the first.<\/p>\n<h2 id=\"beyond-the-default-action\">Beyond the default action<\/h2>\n<p>If you were only interested in the diff but not the commit message in\nthe first place, you could have typed <i>d d<\/i> (<i>d<\/i> by itself\nand then <i>d<\/i> again) in the status buffer instead of <i>RET<\/i>.<\/p>\n<p>But you already pressed <i>RET<\/i>, now what?  Type <i>q<\/i> to quit\n(i.e. hide) the current Magit buffer, in this case the commit buffer,\nto go back to the previous buffer (more precisely the previous window\nconfiguration).  In this case that would take you back to the status\nbuffer.<\/p>\n<p>So while <i>RET<\/i> performs the most likely action for the thing at\npoint, there are many other keys which perform other, equally useful\nactions.  There is <i>d<\/i> to show a diff, <i>l<\/i> to show a log,\nand dozens more, some of which will be mentioned below.<\/p>\n<p>If the cursor is on the thing you want to act on, then it is very\nconvenient that many commands default to acting on that thing.  But\nsometimes you <em>do<\/em> want to act on some other thing. It might be\ninconvenient to first move to its representation in the current\nbuffer, or the thing you want to act on might not even be present in\nthe current buffer.  Many commands therefore also allow you to act on\nsome other thing.<\/p>\n<p>Above we pressed <i>d<\/i> twice to show the diff for the commit at\npoint.  Pressing <i>d<\/i> just once actually does something by itself:\nit shows the \u201cdiff popup\u201d, which you can see below.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-diff.png\" alt=\"\" \/><\/p>\n<p>The window is split into two panes.  (Note that in Emacs, which\npredates window systems, one would say the <em>frame<\/em> is split into two\n<em>windows<\/em>.) The pane at the top still shows the status buffer, while\nthe pane at the bottom shows the diff popup.<\/p>\n<p>The diff popup shows some actions at the bottom and some arguments,\nwhich you know from your command line use of Git, at the top.<\/p>\n<p>Previously we pressed <i>d<\/i> again.  That causes the popup\nbuffer\/window to disappear and the \u201cdwim\u201d (do what I mean) diff\nvariant to be invoked.  As with most dwim variants, this particular\none acts on the thing at point.  Alternatively, you could press any\nletter show under \u201cActions\u201d (such as <i>s<\/i> to show the index).<\/p>\n<p>Or you might want to show the diff for an arbitrary range, in which\ncase you would press <i>r<\/i>.  After doing so, you can type the range\nwith completion:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/diff-range.png\" alt=\"\" \/><\/p>\n<p>Most popups also allow users to set arguments which will then be used\nwhen the invoked action calls Git.  In the diff popup, for example,\none can enable the use of the <code class=\"highlighter-rouge\">--function-context<\/code> argument by typing\n<i>- f<\/i> before invoking an action.  When an argument takes a value,\nthen that is read from the user, using completion when appropriate.<\/p>\n<p>Usually arguments are just used once.  When you later enter the same\npopup again, then it is not enabled anymore.  However, you can save\nthe arguments you want to be used by default.  To do so, enter the\npopup, set the arguments as desired, and then instead of invoking an\naction, type <i>C-c C-c<\/i> to save them as the default for the\ncurrent popup.  These settings persist between Emacs sessions.<\/p>\n<p><i>C-c C-c<\/i> is only one of the commands that is available in all\npopups.  To show them all, type <i>C-t<\/i>.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-common.png\" alt=\"\" \/><\/p>\n<p>There are many popups.  For the most commonly used ones, you will\nquickly learn the keys you have to press because they are mnemonic.\nUntil then, or when you need to do something you rarely do, then the\n\u201cpopup of popups\u201d comes to the rescue.  Show it by typing <i>?<\/i>.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-popup.png\" alt=\"\" \/><\/p>\n<p>Many Magit commands are invoked by first showing the available\nvariants and then picking one of those variants.  (For many popups, it\nis possible to press the same key twice to invoke the most common\nvariant, often a dwim variant.)  But there are also many commands\nwhich are not implemented using such a popup.  We have seen the\nhide\/show command (<i>TAB<\/i>) and the visit command (<i>RET<\/i>) so\nfar.  Other such commands are staging (<i>s<\/i>) and unstaging\n(<i>u<\/i>), which always act on the file or hunk at point.  These and\nsimilar commands will be discussed below.<\/p>\n<h2 id=\"staging-changes\">Staging changes<\/h2>\n<p>Magit makes it very simple to stage and later commit only some\nchanges, while leaving other changes in the working tree to be\ncommitted separately.<\/p>\n<p>On the command line you have to invoke special staging and unstaging\nsessions using <code class=\"highlighter-rouge\">git add --patch<\/code> or <code class=\"highlighter-rouge\">git reset --patch<\/code>, which is\nquite cumbersome as you have to go through all the available hunks one\nby one and in order.<\/p>\n<p>In Magit, just act on what you see.  To stage only the <code class=\"highlighter-rouge\">-59,7 +60,7<\/code>\nhunk you would move there and simply press <i>s<\/i>, without having to\nfirst tell Git that you don\u2019t want to stage <code class=\"highlighter-rouge\">-45,13 +45,14<\/code> and then\nalso that you don\u2019t want to stage <code class=\"highlighter-rouge\">-69,7 +70,7<\/code>.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/stage-before.png\" alt=\"\" \/><\/p>\n<p>After you have done so, the buffer automatically updates and the\ncursor is moved to the next hunk. Therefore you can stage multiple\nhunks by pressing <i>s<\/i> multiple times, if that\u2019s what you want.\nIf not, then Magit won\u2019t bother you with the other hunks like Git\ndoes.  The staged hunk now appears inside the \u201cStaged changes\u201d\nsection.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/stage-after.png\" alt=\"\" \/><\/p>\n<p>Unstaging of course works the same way.  Move to a staged change and\npress <i>u<\/i>.  You can also stage or unstage everything at once by\npressing <i>S<\/i> or <i>U<\/i>.<\/p>\n<h2 id=\"acting-on-the-selection\">Acting on the selection<\/h2>\n<p>You can also stage multiple files or hunks at once. To do so, mark\nthese sibling sections using the region. (The region is an Emacs term\nfor \u201cthe selected text\u201d.)<\/p>\n<p>First press <i>C-SPC<\/i> to mark one end of the region, and then move\nthe cursor until you reach what should be the other end of it.\nAlternatively you could use the mouse.  If the region constitutes a\nvalid \u201cMagit selection\u201d which can be acted on as a unit, then it looks\nlike the example below, where both <code class=\"highlighter-rouge\">CONTRIBUTING.md<\/code> and <code class=\"highlighter-rouge\">README.md<\/code>\nare selected but <code class=\"highlighter-rouge\">lisp\/magit.el<\/code> is not.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/sibling-selection.png\" alt=\"\" \/><\/p>\n<p>When such a selection is active, many commands which would normally\nact on the current section instead act on the selection (i.e. all the\nmarked sections).  Staging (<i>s<\/i>) is such a command.<\/p>\n<p>If a region is active, but it does not constitute a valid selection\nbecause the sections that fall into it are not siblings of one\nanother, then the region looks as it usually does in Emacs.  That\nmakes it trivial to see if you are about to act on the current section\nonly or on the selection.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/no-selection.png\" alt=\"\" \/><\/p>\n<h2 id=\"staging-parts-of-a-hunk\">Staging parts of a hunk<\/h2>\n<p>Besides sibling selections, Magit supports a second selection, the\n\u201chunk internal region\u201d.  You can mark just part of a hunk using the\nregion and then only stage (or unstage or otherwise apply) just that\npart of the hunk.<\/p>\n<p>Here you can see what it looks like when only the part of the hunk\nwhich removes the word \u201colder\u201d is selected using the region:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/stage-region-before.png\" alt=\"\" \/><\/p>\n<p>If you pressed <i>s<\/i> now, then only that part would be staged,\nresulting in:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/stage-region-after.png\" alt=\"\" \/><\/p>\n<p>(Note how conveniently the cursor is placed.  We take pride in such\ndetails.  This particular detail might not be all that important when\nlooked at in isolation, but we go that little extra step in many\nplaces, and that does make a difference.)<\/p>\n<h2 id=\"changing-diff-arguments\">Changing diff arguments<\/h2>\n<p>Instead of using a hunk internal region to just stage parts of the\nhunk, you could also have told Git to make the hunks smaller using\nthe <code class=\"highlighter-rouge\">-U<\/code> argument.  To do so, press <i>-<\/i> until the hunk in question\nbreaks up into two.  Use <i>+<\/i> to do the opposite.<\/p>\n<p>This works in the status buffer as well as all other buffers that may\ncontain diffs.  So far we have seen plain diff buffers and commit\nbuffers.<\/p>\n<p>As we have seen above, you can specify several arguments when using the\ndiff popup to create the diff buffer.  But when showing the status\nbuffer by pressing <i>C-x g<\/i> or when showing the commit buffer by\npressing <i>RET<\/i>, then that is not possible.  You can however change\nthe diff arguments used in the current buffer.  Press <i>D<\/i> to bring\nthe diff arguments popup.  It is very similar to the diff popup in\nthat it offers the same arguments, but instead of actions which show\nsome diff in another buffer, it offers actions which affect the\ncurrent buffer.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-diff-arguments.png\" alt=\"\" \/><\/p>\n<p>Press <i>g<\/i> to use the arguments that are currently selected in the\npopup in the current buffer.  <i>s<\/i> sets these arguments as the\ndefault for buffers of the same type, so if you did that in the status\nbuffer of some repository, then the same arguments would be used for\nstatus buffers of other repositories that you subsequently create.\nBut this only lasts until you restart Emacs; to permanently save new\ndefaults use <i>w<\/i>.<\/p>\n<p>A similar \u201clog arguments\u201d popup exists on <i>L<\/i>.<\/p>\n<h2 id=\"applying-changes\">Applying changes<\/h2>\n<p>Some other graphical tools approach Magit when it comes to the staging\nfeatures described above, but I don\u2019t think any one of them quite\nmakes it.  One more thing that sets Magit apart from these tools,\nhowever, is that these features are not only available for staging and\nunstaging, but also when \u201cotherwise applying changes\u201d.<\/p>\n<p>With Magit you can also discard, reverse, or apply, the file, files,\nhunk, hunks, or region at point using the exact same interface as\ndescribed above.  For more information about these apply variants\n<a href=\"https:\/\/emacsair.me\/manual\/magit\/Applying.html\">consult<\/a> the manual.<\/p>\n<p>If you noticed that the diff popup lacked your favorite argument, then\nfear not.  It is easy to add arguments to an existing popup,\nas <a href=\"https:\/\/emacsair.me\/manual\/magit-popup\/Customizing-existing-popups.html\">described<\/a> in\nthe manual.<\/p>\n<p>Some arguments are missing because they are not actually required.\n<code class=\"highlighter-rouge\">--cached<\/code> falls into that category; it\u2019s not needed because you\nsimply use the \u201cstaged changes\u201d diff variant (<i>s<\/i>) to show the\ncached changes (those in the index) .<\/p>\n<p>Other arguments, e.g. <code class=\"highlighter-rouge\">--irreversible-delete<\/code>, are missing because we\nhave to strike a balance between making every argument available that\nanyone might ever need and not overwhelming beginners with arguments\nthat \u201cnobody will ever need\u201d.<\/p>\n<p>(While it is fairly easy for a user to add additional arguments to a\npopup, I intend to make it trivial.  It will be possible to temporally\nshow all supported arguments and then select which ones should be\nshown by default, using a checkbox or similar.  In other words, users\nwill no longer have to \u201cadd additional arguments\u201d.  Instead they can\nchoose the ones they need.)<\/p>\n<h2 id=\"log-buffers\">Log buffers<\/h2>\n<p>Like it is possible to show a diff in a separate buffer, that is of\ncourse also possible for logs.  To do so press <i>l<\/i> to show the\nlog popup, which looks like this:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-log.png\" alt=\"\" \/><\/p>\n<p>After choosing what log you want to see, that log is shown in the\nactual log buffer, which looks quite similar to what you would get on\nthe command line.  Here is an example of the \u201cshow all branches\u201d\nvariant (notice the arguments displayed at the top):<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/log.png\" alt=\"\" \/><\/p>\n<h2 id=\"moving-through-logs\">Moving through logs<\/h2>\n<p>Things get more interesting once you press <i>RET<\/i> to show the\ncommit at point in a separate buffer.  Actually in this case it is\nbetter to press <i>SPC<\/i>, which displays the same buffer but the\ncurrent log buffer remains the current buffer.  If you now move\nthrough the log buffer using <i>C-p<\/i> and <i>C-n<\/i>, the commit\nbuffer is automatically updated to show the commit which the cursor is\non in the log buffer.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/log-and-commit.png\" alt=\"\" \/><\/p>\n<h2 id=\"innovative-variants-of-many-commands\">Innovative variants of many commands<\/h2>\n<p>As we have seen, Magit provides commands that allow invoking certain\nvariant of Git commands without having to remember and type many\narguments.  The \u201cshow diff for staged\u201d changes, for example, is such\ncommand that saves you from having to remember and type <code class=\"highlighter-rouge\">--cached<\/code>.\nThat, while useful, isn\u2019t all that innovative by itself.  Also many\nexperienced users of the Git command line have probably created\naliases for that sort of thing by now.<\/p>\n<p>But Magit also provides many truly innovative commands which go beyond\nwhat you could do with aliases.  These include, among others, commands\nfor committing, rebasing, stashing, and branching.<\/p>\n<h2 id=\"committing\">Committing<\/h2>\n<p>Press <i>c<\/i> to show the committing popup and then <i>c<\/i> to\ncreate a regular new commit.  This calls <code class=\"highlighter-rouge\">git commit<\/code> and arranges for\nthe current Emacs session to be used as the <code class=\"highlighter-rouge\">$GIT_EDITOR<\/code>.  The commit\nmessage buffer is opened in one buffer, and another buffer shows the\nchanges about to be committed.  When done, press <i>C-c C-c<\/i>.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/committing.png\" alt=\"\" \/><\/p>\n<p>You can also amend to <code class=\"highlighter-rouge\">HEAD<\/code> by pressing <i>a<\/i>.  This is just like\n<code class=\"highlighter-rouge\">git commit --amend<\/code>.  While editing the message while amending, press\n<i>C-c C-d<\/i> to toggle between showing only the changes that are\nbeing added to HEAD and all changes that will make it into the amended\ncommit.<\/p>\n<p>There are also fixup and squash variants.  And then there are some\nvariants that don\u2019t exist in Git.  Because the terms used here do not\ncorrespond to terms you might be familiar with from Git, this might be\na good time to show you how to get more information about a popup\naction or argument.  To show information about the \u201creword\u201d variant,\ntype <i>?<\/i> and then <i>w<\/i>, which gives you this:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-help.png\" alt=\"\" \/><\/p>\n<p>So \u201creword\u201d lets you change <code class=\"highlighter-rouge\">HEAD<\/code>\u2019s message without adding any\nchanges to it.  Similarly \u201cextend\u201d adds the staged changes without\nforcing you to review the message, which is quite useful for simple\ntypo fixes.<\/p>\n<p>\u201cInstant fixup and squash\u201d let you pick a commit to be modified with\nthe staged changes, but then they immediately initiate a rebase.\n\u201cInstand fixup\u201d is mostly like \u201cextend\u201d but on commits other than\n<code class=\"highlighter-rouge\">HEAD<\/code>.<\/p>\n<p>The commit to be modified is picked using a specialized log buffer.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/log-squash.png\" alt=\"\" \/><\/p>\n<h2 id=\"rebasing\">Rebasing<\/h2>\n<p>The rebasing popup also provides a variety of commands that are unique\nto Magit.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-rebase.png\" alt=\"\" \/><\/p>\n<p>There is one variant which lets you edit a single commit, without\nhaving to go through a list of commits to be rebased.  Simply go\nto the commit you want to rebase in any log (of course including the\nlog of unpushed commits in the status buffer), press <i>r<\/i> to enter\nthe rebase popup, and then <i>m<\/i> to modify that commit.  Likewise\nthere is a variant which allows only editing the message of a single\ncommit.<\/p>\n<p>When performing an interactive rebase (which is supposed to affect more\nthan a single commit), then this is done as usual by editing the file\n<code class=\"highlighter-rouge\">git-rebase-todo<\/code>, but Magit makes a few additional key bindings and\nother convenience features available:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/rebase-edit.png\" alt=\"\" \/><\/p>\n<p>When a rebase sequence is in progress, then a log-like section with\ninformation about the rebase is shown in the status buffer.  Like in a\nregular log, you can of course perform various actions on the listed\ncommits, like viewing them.  And of course conflicts are shown too.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/rebase-status.png\" alt=\"\" \/><\/p>\n<p>You can resolve the conflict by visiting the file.  After\npressing <i>RET<\/i> on a conflict hunk you would see something like\nthis:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/smerge.png\" alt=\"\" \/><\/p>\n<p>There you can use the Smerge package to do so in style, or you can\njust edit the file.  You can also use the Ediff package, which shows\nthe two sides and optionally the common ancestor in separate windows.\nMagit wraps the features provided by these packages, but since they\nare not actually part of Magit, we skip looking at them in detail.<\/p>\n<p>Of course you could also abort the rebase by pressing <i>r<\/i> again and\nthen <i>a<\/i>:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-rebase-abort.png\" alt=\"\" \/><\/p>\n<h2 id=\"merging\">Merging<\/h2>\n<p>The merging popup is quite simple.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-merge.png\" alt=\"\" \/><\/p>\n<p>When a merge is in progress, then the status buffer shows information\nabout that:<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/merge-status.png\" alt=\"\" \/><\/p>\n<h2 id=\"fetching-pulling-and-pushing\">Fetching, pulling and pushing<\/h2>\n<p>Without going into any details, here are the popups for fetching,\npulling and pushing.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-fetch.png\" alt=\"\" \/><\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-pull.png\" alt=\"\" \/><\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-push.png\" alt=\"\" \/><\/p>\n<p>One thing worth noting though is that all of these popups feature a\n\u201c<i>p<\/i>ush-remote\u201d, an \u201c<i>u<\/i>pstream\u201d, and an \u201c<i>e<\/i>lsewhere\u201d\nvariant.  A branch\u2019s \u201cpush-remote\u201d is somewhat similar to the\n\u201cupstream\u201d, but it usually is a different branch.  The \u201cpush-remote\u201d\nis actually a Git feature, but not many Git users know about it because\nit is hardly documented.  To learn more about the \u201cpush-remote\u201d and\nhow it is different from the \u201cupstream\u201d,\nsee <a href=\"https:\/\/magit.vc\/manual\/magit\/The-two-remotes.html\">this<\/a><\/p>\n<h2 id=\"branching\">Branching<\/h2>\n<p>The branching popup is used to edit, create, and\/or checkout a branch.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-branch.png\" alt=\"\" \/><\/p>\n<p>The \u201ccreate new spin-off\u201d variant is particularly interesting.  It\ncreates and checks out a new branch whose upstream is the previously\nchecked out branch.  Then it rewinds that branch to its upstream.\nThis is useful when you began working on some new feature directly on\n<code class=\"highlighter-rouge\">master<\/code> and then realize that you should be using a feature branch.<\/p>\n<p>The branching popup is also one of the few popups that have a\nsub-popup.  Press <i>C<\/i> to show the branch configuration popup.<\/p>\n<p><img src=\"https:\/\/magit.vc\/screenshots\/popup-branch-config.png\" alt=\"\" \/><\/p>\n<p>In that popup you can see the values of some important Git variables\nconcerning the current branch.  The values obviously can be changed\nusing the shown keys.<\/p>\n<p>Footnotes:<\/p>\n<div class=\"footnotes\"><ol><li id=\"fn:1\">\n      <p><i>C-x g<\/i> is actually only the <em>recommended<\/em> binding for the\ncommand <code class=\"highlighter-rouge\">magit-status<\/code>.  To use this binding add <code class=\"highlighter-rouge\">(global-set-key\n(kbd \"C-x g\") 'magit-status)<\/code> to the init file <code class=\"highlighter-rouge\">~\/.emacs.d\/init.el<\/code>.\u00a0<a href=\"#fnref:1\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>\n    \n    <noscript><p><img src=\"https:\/\/tarsius.piwikpro.com\/piwik.php?idsite=5\" alt=\"\" \/><\/p><\/noscript>","created_at":"2018-11-07T16:46:55+0000","updated_at":"2018-11-07T16:46:55+0000","annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"emacsair.me","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":397,"title":"AlesTsurko\/microscale","url":"https:\/\/github.com\/AlesTsurko\/microscale","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Generated in real-time from random Wikipedia articles, <strong>microscale<\/strong> is a web-based, generative album, also available as a rendered physical CDR and digital download. Each article functions a step sequencer \u2013 where the letters are the sequencer steps and the track titles are regular expressions that switch the steps on and off. The web version is generated in real-time and is also hackable \u2013 the regular expressions, samples, tempo and other parameters can be edited or replaced \u2013 enabling <strong>microscale<\/strong> to be used as an instrument and sequencer for the listener\u2019s own music.<\/p>\n<p>The concept behind <strong>microscale<\/strong> is to show that through transforming one media (text) into another media (music), the meaning can be transformed \u2013 the article has its own meaning, but the music generated from the article has a completely different meaning.<\/p>\n<p><strong>microscale<\/strong> is available from <a href=\"http:\/\/alestsurko.by\/microscale\/\" rel=\"nofollow\">http:\/\/alestsurko.by\/microscale\/<\/a>. The physical version is available as individually rendered CDRs with every disc totally unique.<\/p>\n<\/article>","created_at":"2018-11-07T16:46:53+0000","updated_at":"2018-11-07T16:46:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/8686568?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":398,"title":"How to globally disable eldoc? - Emacs Stack Exchange","url":"https:\/\/emacs.stackexchange.com\/questions\/31414\/how-to-globally-disable-eldoc\/31415","content":"<div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n&#13;\n<p>Using Rust\/Racer\/Eldoc is causing performance problems while navigating around the file (hanging while <code>eldoc<\/code> is fetching the docs).<\/p>\n<p>Running <code>eldoc-mode<\/code> reports that it's disabled but still runs Racer's <code>eldoc<\/code> function.<\/p>\n<p>Is there a way to globally disable <code>eldoc<\/code>?<\/p><\/div><h2 data-answercount=\"1\">&#13;\n                                1 Answer&#13;\n                                1&#13;<\/h2><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Since Emacs 25.1 a new global minor mode <code>global-eldoc-mode<\/code> exists, and is enabled by default.<\/p>\n<p>To disable it, simply add this to your init file:<\/p>\n<pre>(global-eldoc-mode -1)\n<\/pre>\n<p>You can alternatively <code>(setq-local eldoc-documentation-function #'ignore)<\/code> in buffers for which you wish eldoc to have no effect, without disabling it elsewhere.<\/p><\/div>","created_at":"2018-11-07T16:46:53+0000","updated_at":"2018-11-07T16:46:53+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"emacs.stackexchange.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/emacs\/img\/apple-touch-icon@2.png?v=dbead6deb8b0","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":394,"title":"I want to believe (in Lisp vs C performance)","url":"https:\/\/news.ycombinator.com\/item?id=591897","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D591897\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"591897\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/www.imagine27.com\/articles\/2009-05-03-195227_i_want_to_believe_in_lisp_performance.html\" class=\"storylink\">I want to believe (in Lisp vs C performance)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=imagine27.com\">imagine27.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        20 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=jgrant27\" class=\"hnuser\">jgrant27<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=591897\">on May 4, 2009<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=591897&amp;goto=item%3Fid%3D591897\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=I%20want%20to%20believe%20(in%20Lisp%20vs%20C%20performance)&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=I%20want%20to%20believe%20(in%20Lisp%20vs%20C%20performance)\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=591897&amp;auth=593848b1071fd0a44526900252eb2071fc733371\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=591897\">19\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:46:52+0000","updated_at":"2018-11-07T16:46:52+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":395,"title":"Basic Synthesis: Part 3 \u2013 Envelopes","url":"https:\/\/www.innovativesynthesis.com\/basic-synthesis-part-3-\u2013-envelopes\/","content":"<p class=\"MsoNormal\">Envelopes are the key to the articulation of your sound. Without them your patch will immediately start off at full blast, and stay there, and then disappear all of a sudden when you let go of the key. Envelopes, although difficult to understand at first allow you to change that, so you can create expressive and dynamic sounds with your synthesizer.<\/p><p class=\"MsoNormal\">The standard envelope is in 4 main stages, described below:<\/p><ul type=\"disc\"><li class=\"MsoNormal\">Attack \u2013 the sound rising up to its maximum level. If it\u2019s set to nothing, the sound plays at full blast straight away, whereas if you set it quite high then the sound gradually fades (good for string sounds).<\/li>\n<li class=\"MsoNormal\">Decay \u2013 this is how long the sound stays at the level the attack brings it up to. If it\u2019s set as high as it will go, it will stay at the maximum level forever (rendering the sustain stage useless).<\/li>\n<li class=\"MsoNormal\">Sustain \u2013 this is the level that the sound stays at after the decay stage has passed. Some synthesizers also have a dedicated \u2018sustain time\u2019 setting, which decays the sustain stage after an adjustable amount of time too.<\/li>\n<li class=\"MsoNormal\">Release \u2013 a bit like reverb at the end of your sound \u2013 it is how long the sustain level takes to die down to silence. Set the release to nothing and you won\u2019t get that effect \u2013 it will be instant.<\/li>\n<\/ul><p class=\"MsoNormal\">The key to programming envelopes on your synthesizer is to practice and visual the envelope. Think of it like a graph with 4 stages, and you are plotting points higher or lower on the domain as the sound changes through the stages of the envelope. The sound rises through the attack, it dies down through he decay, it stays at the selected level in the sustain, and when you let go it disappears through the release.<\/p><p class=\"MsoNormal\">Filter envelopes are just that \u2013 they articulate the filter using an envelope. To do this, there is usually a knob dedicated to the filter frequency (cut-off point) just for the filter envelope. Turn the normal filter cut-off down, and turn the filter envelope cut-off higher, and then program the filter envelope like a normal envelope. Takes a while to figure out, but its worth it \u2013 for example, turn the sustain on the filter envelope off, forget about the release and attack (set them to zero), and make the decay short \u2013 then turn up the release and wow! The thing to remember here is that its not the volume (amplitude) that your changing, it\u2019s the cut-off frequency with the envelope, so you can create great effects with clever programming \u2013 such as emulating brass sounds by increasing the attack on the filter envelope.<\/p>","created_at":"2018-11-07T16:46:52+0000","updated_at":"2018-11-07T16:46:52+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"www.innovativesynthesis.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":396,"title":"Your tracks on wax: How to press your own record","url":"https:\/\/mixmag.net\/feature\/how-to-press-your-own-record","content":"<p>People are coming back to vinyl in record numbers and it's never been easier to get your own music onto wax. <\/p><p>A rejuvenation of record pressing plants has resulted in <a href=\"https:\/\/mixmag.net\/feature\/our-favourite-vinyl-only-labels\" target=\"_blank\">a fertile underground independent label scene<\/a> and we all know that the true mark of credibility is no longer in MP3s, SoundCloud plays and streaming services \u2013 it's whether you've gone to the effort of pressing some records.<\/p><p>So for those thinking about striking out on their own and wishing to see their music on more than just a hard drive, we've put together some guidelines and what to expect while going through the process.<\/p>","created_at":"2018-11-07T16:46:52+0000","updated_at":"2018-11-07T16:46:52+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"mixmag.net","preview_picture":"https:\/\/mixmag.net\/assets\/uploads\/images\/_facebook\/Cutting-Machine-Top.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":393,"title":"Table of Contents \u2014 LLVM 8 documentation","url":"https:\/\/llvm.org\/docs\/tutorial\/","content":"LLVM Tutorial: Table of Contents \u2014 LLVM 8 documentation\n<div class=\"logo\"><a href=\"https:\/\/llvm.org\/docs\/index.html\">\n    <img src=\"https:\/\/llvm.org\/docs\/_static\/logo.png\" alt=\"LLVM Logo\" width=\"250\" height=\"88\" \/><\/a><\/div>\n    \n    <div class=\"document\"><div class=\"documentwrapper\"><div class=\"body\" role=\"main\"><div class=\"section\" id=\"llvm-tutorial-table-of-contents\"><div class=\"section\" id=\"kaleidoscope-implementing-a-language-with-llvm\"><h2>Kaleidoscope: Implementing a Language with LLVM<\/h2><div class=\"toctree-wrapper compound\"><ul><li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl01.html\">1. Kaleidoscope: Tutorial Introduction and the Lexer<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl02.html\">2. Kaleidoscope: Implementing a Parser and AST<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl03.html\">3. Kaleidoscope: Code generation to LLVM IR<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl04.html\">4. Kaleidoscope: Adding JIT and Optimizer Support<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl05.html\">5. Kaleidoscope: Extending the Language: Control Flow<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl06.html\">6. Kaleidoscope: Extending the Language: User-defined Operators<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl07.html\">7. Kaleidoscope: Extending the Language: Mutable Variables<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl08.html\">8. Kaleidoscope: Compiling to Object Code<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl09.html\">9. Kaleidoscope: Adding Debug Information<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/LangImpl10.html\">10. Kaleidoscope: Conclusion and other useful LLVM tidbits<\/a><\/li>\n<\/ul><\/div><\/div><div class=\"section\" id=\"kaleidoscope-implementing-a-language-with-llvm-in-objective-caml\"><h2>Kaleidoscope: Implementing a Language with LLVM in Objective Caml<\/h2><div class=\"toctree-wrapper compound\"><ul><li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl1.html\">1. Kaleidoscope: Tutorial Introduction and the Lexer<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl2.html\">2. Kaleidoscope: Implementing a Parser and AST<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl3.html\">3. Kaleidoscope: Code generation to LLVM IR<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl4.html\">4. Kaleidoscope: Adding JIT and Optimizer Support<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl5.html\">5. Kaleidoscope: Extending the Language: Control Flow<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl6.html\">6. Kaleidoscope: Extending the Language: User-defined Operators<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl7.html\">7. Kaleidoscope: Extending the Language: Mutable Variables<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/OCamlLangImpl8.html\">8. Kaleidoscope: Conclusion and other useful LLVM tidbits<\/a><\/li>\n<\/ul><\/div><\/div><div class=\"section\" id=\"building-a-jit-in-llvm\"><h2>Building a JIT in LLVM<\/h2><div class=\"toctree-wrapper compound\"><ul><li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/BuildingAJIT1.html\">1. Building a JIT: Starting out with KaleidoscopeJIT<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/BuildingAJIT2.html\">2. Building a JIT: Adding Optimizations \u2013 An introduction to ORC Layers<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/BuildingAJIT3.html\">3. Building a JIT: Per-function Lazy Compilation<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/BuildingAJIT4.html\">4. Building a JIT: Extreme Laziness - Using Compile Callbacks to JIT from ASTs<\/a><\/li>\n<li class=\"toctree-l1\"><a class=\"reference internal\" href=\"https:\/\/llvm.org\/docs\/tutorial\/BuildingAJIT5.html\">5. Building a JIT: Remote-JITing \u2013 Process Isolation and Laziness at a Distance<\/a><\/li>\n<\/ul><\/div><\/div><div class=\"section\" id=\"external-tutorials\"><h2>External Tutorials<\/h2><dl class=\"docutils\"><dt><a class=\"reference external\" href=\"http:\/\/jonathan2251.github.com\/lbd\/\">Tutorial: Creating an LLVM Backend for the Cpu0 Architecture<\/a><\/dt>\n<dd>A step-by-step tutorial for developing an LLVM backend. Under\nactive development at <a class=\"reference external\" href=\"https:\/\/github.com\/Jonathan2251\/lbd\">https:\/\/github.com\/Jonathan2251\/lbd<\/a> (please\ncontribute!).<\/dd>\n<dt><a class=\"reference external\" href=\"http:\/\/www.embecosm.com\/appnotes\/ean10\/ean10-howto-llvmas-1.0.html\">Howto: Implementing LLVM Integrated Assembler<\/a><\/dt>\n<dd>A simple guide for how to implement an LLVM integrated assembler for an\narchitecture.<\/dd>\n<\/dl><\/div><\/div><\/div><\/div><\/div>\n    \n    <p>\u00a9 Copyright 2003-2018, LLVM Project.\n      Last updated on 2018-11-07.\n      Created using <a href=\"http:\/\/sphinx-doc.org\/\">Sphinx<\/a> 1.7.6.<\/p>","created_at":"2018-11-07T16:46:51+0000","updated_at":"2018-11-07T16:46:51+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"llvm.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":390,"title":"syl20bnr\/spacemacs","url":"https:\/\/github.com\/syl20bnr\/spacemacs\/wiki\/Debugging","content":"<div class=\"markdown-body\">\n          <p>This page explains some concepts needed to debug Spacemacs and to report issues\n  in an useful way.<\/p>\n<ul><li><a href=\"#how-to-make-a-great-bug-report\">How to make a great bug report<\/a><\/li>\n  <li><a href=\"#get-the-description-of-your-installation\">Get the description of your installation<\/a><\/li>\n  <li><a href=\"#make-a-step-by-step-reproduction-guide\">Make a step-by-step reproduction guide<\/a><\/li>\n  <li>\n<a href=\"#enable-debug\">Enable debug<\/a>\n    <ul><li><a href=\"#enable-debug-at-initialization\">Enable debug at initialization<\/a><\/li>\n      <li><a href=\"#if-load-time-is-a-problem\">If load time is a problem<\/a><\/li>\n      <li><a href=\"#debug-a-freezing-instance\">Debug a freezing instance<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"#get-the-backtrace\">Get the backtrace<\/a><\/li>\n  <li><a href=\"#debugging-segmentation-faults-with-gdb\">Debugging segmentation faults with gdb<\/a><\/li>\n  <li><a href=\"#dummy-example-of-a-good-bug-report\">Dummy example of a good bug report<\/a><\/li>\n<\/ul><p>A great bug report <img class=\"emoji\" title=\":metal:\" alt=\":metal:\" src=\"https:\/\/camo.githubusercontent.com\/e531c6ea506f115c719d80d4c4c275ae3075dad3\/68747470733a2f2f6173736574732d63646e2e6769746875622e636f6d2f696d616765732f69636f6e732f656d6f6a692f756e69636f64652f31663931382e706e67\" height=\"20\" width=\"20\" data-canonical-src=\"https:\/\/assets-cdn.github.com\/images\/icons\/emoji\/unicode\/1f918.png\" style=\"text-align: absmiddle;\" \/> should contain:<\/p>\n<ul><li>A description of your problem<\/li>\n  <li>The description of your system:\n    <ul><li><a href=\"#get-the-description-of-your-installation\">Get the description of your installation<\/a><\/li>\n    <\/ul><\/li>\n  <li>A step-by-step reproduction guide:\n    <ul><li><a href=\"#make-a-step-by-step-reproduction-guide\">Make a step-by-step reproduction guide<\/a><\/li>\n      <li>Describe the observed and expected behaviours<\/li>\n    <\/ul><\/li>\n  <li>A backtrace (if relevant):\n    <ul><li><a href=\"#enable-debug\">Enable debug<\/a><\/li>\n      <li><a href=\"#get-the-backtrace\">Get the backtrace<\/a><\/li>\n    <\/ul><\/li>\n<\/ul><p>This is a template you can use for reporting issues on Spacemacs:<\/p>\n<p><i>Note: Triple backquotes are not showing up because of some wiki parsing issues.\n  Don\u2019t forget to add them around the backtrace.<\/i><\/p>\n<pre lang=\"text\">\n&lt;PLEASE REPLACE ALL UPPERCASE EXPRESSIONS AND DELETE THIS LINE&gt;\n#### Description\nDESCRIPTION OF THE PROBLEM\n#### Reproduction guide\n- Start Emacs\n- REPRODUCTION STEPS\n*Observed behaviour:*\nDESCRIPTION OF THE OBSERVED BEHAVIOUR\n*Expected behaviour:*\nDESCRIPTION OF THE EXPECTED BEHAVIOUR\n#### System Info\nREPLACE THIS SECTION BY THE OUTPUT OF `SPC h d s`\n#### Backtrace\nBACKTRACE IF RELEVANT; PUT IT BETWEEN TRIPLE BACKQUOTES\n<\/pre>\n<p>You can see an example of it <a href=\"#dummy-example-of-a-good-bug-report\">at the end of this page<\/a>.<\/p>\n<p>The versions of software your are using can be obtained by pressing <code>SPC h d s<\/code>\n  or <code>M-m h d s<\/code> (bind to <code>spacemacs\/describe-system-info<\/code>). This will copy to the\n  system clipboard some important information, such as your activated layers, or\n  the Spacemacs and Emacs versions you are using.<\/p>\n<p>The copied text is in markdown format, so you can paste it directly in <code>gitter<\/code>\n  or in issues, <b>without<\/b> quoting it.<\/p>\n<p>Spacemacs contributors are more efficient if they have a way to reproduce the\n  bugs on their side. For this, nothing is better than a detailed step-by-step\n  reproduction guide. It should describe a minimal workflow that make the problem\n  appears. It should look like a bullet-points list, beginning with \u201cStart emacs\u201d.<\/p>\n<p>This is a good example:<\/p>\n<ul><li>Start Emacs<\/li>\n  <li>Open the `.spacemacs` file with <code>SPC f e d<\/code>\n<\/li>\n  <li>Search for the \u201cdotspacemacs\u201d with <code>\/ dotspacemacs RET<\/code>\n<\/li>\n  <li>Notice <code>n<\/code> and <code>N<\/code> are not working<\/li>\n<\/ul><p>On `develop` you can also activate the <code>command-log<\/code> layer. That will offer you\n  the possibility to toggle this mode with <code>SPC a L<\/code> or <code>M-m a L<\/code>. All the key you\n  will press will be recorded with their associated functions on a side buffer.\n  You can then use the content of this buffer as a starting point of a\n  reproduction guide.<\/p>\n<p>To turn on the <code>debug-on-error<\/code> mode, what will give a backtrace of an error\n  popping up at run-time, you can use <code>SPC t D<\/code> or <code>M-m t D<\/code>. This keybinding\n  toggle the mode, so it can also be used to deactivate it.<\/p>\n<p>After having enabled the <code>debug-on-error<\/code> mode, you can redo the action making\n  the error to appear, then read the <a href=\"#reporting-the-backtrace\">Reporting the backtrace<\/a> section on how to get\n  and report the problematic backtrace.<\/p>\n<h2>\n<a id=\"user-content-enable-debug-at-initialization\" class=\"anchor\" href=\"#enable-debug-at-initialization\" aria-hidden=\"true\"><\/a>Enable debug at initialization<\/h2>\n<p>In order to get a backtrace of an error appearing at initialization time, you\n  need to run emacs with the <code>--debug-init<\/code> flag, as in:<\/p>\n<pre lang=\"text\">\nemacs --debug-init\n<\/pre>\n<p>The <code>debug-on-error<\/code> mode stays activated after the initialization, so no need\n  to use <code>SPC t D<\/code> afterward to debug errors.<\/p>\n<h2>\n<a id=\"user-content-if-load-time-is-a-problem\" class=\"anchor\" href=\"#if-load-time-is-a-problem\" aria-hidden=\"true\"><\/a>If load time is a problem<\/h2>\n<p>If part of the issue is the amount of time spacemacs takes to load, you can\n  launch emacs with the command<\/p>\n<pre lang=\"text\">\nemacs --debug-init --timed-requires\n<\/pre>\n<p>which will add a timer to all <code>require<\/code> and <code>load<\/code> statements. Any file that\n  takes longer than 150ms (not configurable at the moment) to load will be logged\n  in a buffer called <code>*load-times*<\/code>, which may help you narrow down the cause of\n  the slow startup.<\/p>\n<h2>\n<a id=\"user-content-debug-a-freezing-instance\" class=\"anchor\" href=\"#debug-a-freezing-instance\" aria-hidden=\"true\"><\/a>Debug a freezing instance<\/h2>\n<p>When an instance of emacs is freezing, it is difficult to debug and to get a\n  backtrace. Sometimes, even pressing `C-g` is not enough to escape the\n  problematic behaviour.<\/p>\n<p>Emacs handle the system <code>SIGUSR2<\/code> signal for such cases. This tries, by acting\n  at a lower level, to stop the problematic behaviour and enable the debugger.<\/p>\n<p>According to this <a href=\"http:\/\/emacs.1067599.n5.nabble.com\/PATCH-Unconditional-quit-on-SIGUSR2-td220976.html\" rel=\"nofollow\">mailing list thread<\/a>, here is what this does:<\/p>\n<blockquote>\n  <p>When we receive it, we set debug-on-quit to t, inhibit-quit to nil, and\n    quit-flag to t. These operations will force Emacs to stop at the next possible\n    opportunity and display a backtrace. From here, users can save their work, and\n    developers can figure out what caused the hang.<\/p>\n<\/blockquote>\n<p>Here are the general steps to debug a freeze:<\/p>\n<ol><li>Start emacs.<\/li>\n  <li>In a terminal, find the PID of the emacs process: <code>ps aux | grep emacs | grep -v grep<\/code>\n<\/li>\n  <li>Go to emacs and trigger the freeze.<\/li>\n  <li>Return to the terminal and execute <code>kill -USR2 12345<\/code> where <code>12345<\/code> must be replaced with the PID of the emacs process from step 2.<\/li>\n  <li>Retrun to emacs, you should see a window open with a backtrace.<\/li>\n  <li>Copy the backtrace in an issue, between two lines that contain: <code>```<\/code> to display it as a block or in a bug report to the emacs mailing list.<\/li>\n<\/ol><blockquote>\n  <p>NOTE: The backtrace might contain null characters (seen as <code>^@<\/code> in emacs). In that case you might not be able to copy paste it entirely. Save the backtrace in a file and upload it in the issue or email.<\/p>\n<\/blockquote>\n<p>After having enabled the <code>debug-on-error<\/code> mode, if an error occurs its backtrace\n  will be accessible in the <code>*Backtrace*<\/code> buffer. This buffer shows-up after an\n  error. If it is not the case you can access it with <code>SPC b b<\/code> or <code>M-m b b<\/code>.<\/p>\n<p>In order to get this buffer content, you can use <code>M-m b Y<\/code>, what will copy the\n  buffers content in the system clipboard. You can paste it directly on Github\n  issues between triple backquotes <code>```<\/code>. If the content is too long, or for\n  sharing it on <code>gitter<\/code>, it\u2019s recommended to use a pastebin service.<\/p>\n<p>Note: If you have the `github` layer enabled, you can use <code>M-m g g b<\/code> to\n  directly send the content of the buffer to a gist and get back it\u2019s URL. The URL\n  is copied to the system clipboard so you can directly use it.<\/p>\n<p>Here are the general steps to debug segfaults:<\/p>\n<ol><li>Start emacs.<\/li>\n  <li>In a terminal, find the PID of the emacs process: <code>ps aux | grep emacs | grep -v grep<\/code>\n<\/li>\n  <li>Start <code>gdb<\/code> with: <code>gdb attach 12345<\/code> where <code>12345<\/code> must be replaced with the PID of the emacs process from step 2. You might need to use <code>sudo<\/code> if <code>gdb<\/code> complains about ptrace.<\/li>\n  <li>While inside <code>gdb<\/code> press <code>c<\/code> (short from continue) and then enter (RET).<\/li>\n  <li>Go to emacs and trigger the crash.<\/li>\n  <li>Return to the <code>gdb<\/code> terminal and it should tell you that it received <code>SIGSEGV<\/code>. Type <code>bt<\/code> (short from backtrace) and copy paste it in an issue, between two lines that contain: <code>```<\/code> to display it as a block or in a bug report to the emacs mailing list.<\/li>\n<\/ol><p>You might get a backtrace that looks like this:\n  \n<\/p><pre>#0  0x00000000004ae1f0 in  ()\n#1  0x00000000004b2472 in  ()\n#2  0x000000000041cd3a in  ()\n...<\/pre>\n<p>This is not really usefull since it does not contain function names. You can get them by installing emacs from source:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>git clone https:\/\/github.com\/emacs-mirror\/emacs.git ~\/emacs\ncd ~\/emacs\ngit checkout emacs-25.2\n.\/autogen.sh\n# The `--program-transform-name` flag ensures you don't screw your current ctags package.\nCFLAGS='-O2 -g3' .\/configure --program-transform-name='s\/^ctags$\/ctags.emacs\/'\nmake\nsudo make install<\/pre><\/div>\n<p>You might also need a few extra packages installed, like <code>gcc<\/code>, <code>autoconf<\/code>, and <code>texinfo<\/code>. The <code>.\/autogen.sh<\/code> and <code>make<\/code> commands will give you hints about what is missing in their output.<\/p>\n<p>Make sure you do a <code>git checkout<\/code> with the same emacs version you currently have. You can also try to building master and check if the problems is reproducible there, maybe a fix was comitted but not released.<\/p>\n<p>The interesting part is: \u201cCFLAGS=\u2019-O2 -g3\u2019\u201d. You need the <code>-g3<\/code> flag for the debug symbols. The <code>-O2<\/code> flag is for performance, you can try compiling with <code>-O0<\/code> if you want to make your life easier while inspecting variable contents in <code>gdb<\/code> but emacs will move slowly.<\/p>\n<p><i>Note: Triple backquotes are not showing up because of some wiki parsing issues.\n  Don\u2019t forget to add them around the backtrace.<\/i><\/p>\n<pre lang=\"text\">\n#### Description\nSearching next\/previous elements in vim-mode with `n` and `N` are not working.\n#### Reproduction guide\n- Start Emacs\n- Be sure to use `vim` editing mode\n- Open the `.spacemacs` file with `SPC f e d`\n- Search for the \"dotspacemacs\" expression with `\/ dotspacemacs RET`\n*Observed behaviour:*\nI observe `n` and `N` are not having any effect.\n*Expected behaviour:*\nI expect `n` to search the next occurence of an expression, and `N` to search the previous one.\n#### System Info\n- OS: gnu\/linux\n- Emacs: 24.5.1\n- Spacemacs: 0.105.0\n- Spacemacs branch: contrib\/bepo (rev. cc4b324)\n- Distribution: spacemacs\n- Layers:\n(asciidoc auto-completion better-defaults c-c++ clojure colors command-log deft dockerfile\n              (elfeed :variables rmh-elfeed-org-files\n                      (list \"~\/org\/elfeed.org\"))\n              emacs-lisp emoji erc ess evil-cleverparens extra-langs finance git github haskell html ipython-notebook javascript latex markdown mu4e no-dots ocaml org pandoc python ranger restclient ruby semantic shell shell-scripts speed-reading\n              (spell-checking :variables spell-checking-enable-auto-dictionary t)\n              syntax-checking themes-megapack theming typography\n              (version-control :variables version-control-diff-tool 'diff-hl)\n              yaml)\n#### Backtrace\nDebugger entered--Lisp error: (error \"Don't know where `evil-lisp-state-major-mode-map-prefix' is defined\")\n  signal(error (\"Don't know where `evil-lisp-state-major-mode-map-prefix' is defined\"))\n  error(\"Don't know where `%s' is defined\" evil-lisp-state-major-mode-map-prefix)\n  find-function-search-for-symbol(evil-lisp-state-major-mode-map-prefix nil nil)\n  find-function-noselect(evil-lisp-state-major-mode-map-prefix)\n  find-definition-noselect(evil-lisp-state-major-mode-map-prefix nil)\n  find-function-do-it(evil-lisp-state-major-mode-map-prefix nil switch-to-buffer)\n  find-function(evil-lisp-state-major-mode-map-prefix)\n  helm-descbinds-action:find-func((\"SPC k\" . evil-lisp-state-major-mode-map-prefix))\n  helm-execute-selection-action-1()\n  helm-execute-selection-action()\n<\/pre>\n        <\/div><div id=\"wiki-footer\" class=\"mt-5 mb-0 wiki-footer gollum-markdown-content\">\n          <div class=\"Box Box--condensed bg-gray box-shadow\">\n            <div class=\"Box-body  markdown-body\">\n              <p><em>From Spacemacs, with <img class=\"emoji\" title=\":heart:\" alt=\":heart:\" src=\"https:\/\/camo.githubusercontent.com\/a71ac48290b2d5856b481ef9f4643099aacf756f\/68747470733a2f2f6173736574732d63646e2e6769746875622e636f6d2f696d616765732f69636f6e732f656d6f6a692f756e69636f64652f323736342e706e67\" height=\"20\" width=\"20\" data-canonical-src=\"https:\/\/assets-cdn.github.com\/images\/icons\/emoji\/unicode\/2764.png\" style=\"text-align: absmiddle;\" \/><\/em><\/p>\n            <\/div>\n          <\/div>\n        <\/div>","created_at":"2018-11-07T16:46:50+0000","updated_at":"2018-11-07T16:46:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/1243537?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":391,"title":"Better Email with mu4e","url":"https:\/\/notanumber.io\/2016-10-03\/better-email-with-mu4e\/","content":"October 3, 2016<p>I\u2019ve never been a big fan of email, but at this point, it\u2019s so ingrained into the day to day of every profession that the only thing one can do is conform. I used gmail ever since it was introduced by Google in 2004, and I have to admit that out of all the free email providers, it\u2019s pretty great. Sure its UI has had its ups and downs, and it has been the centre of some controversy with the scanning of emails for advertisements, but the feature set it provides is hard to argue against, but it also has its shortcomings.<\/p><p>When the amount of email you receive starts to encroach on unmanageable territory, you start realizing that:<\/p><ul><li>the UI becomes very cumbersome<\/li>\n<li>bulk actions takes a lot of mouse movement<\/li>\n<li>as fast as it is, seeing that \u201cLoading\u2026\u201d message at the top of the screen is just <em>ugh<\/em><\/li>\n<li>and most importantly\u2026<\/li>\n<\/ul><p>Checking email in a browser is a <strong>huge<\/strong> distraction. I spend most of my day staring at emacs, and when email notifications ding, I feel compelled to check; switching to the completely browser pulls me out of my flow. Once you\u2019re in the browser, you realize that you are on a slippery slope\u2026 It starts tempting you to check <a href=\"https:\/\/news.ycombinator.com\">hacker news<\/a> or <a href=\"https:\/\/reddit.com\">reddit<\/a> (at least for me), and often times it\u2019s impossible to not succumb to it; we all know how easily time goes missing \u201cjust checking if anything new popped up on the front-page\u201d.<\/p><p>Addressing these annoyances is where <a href=\"https:\/\/www.djcbsoftware.nl\/code\/mu\/mu4e.html\">mu4e<\/a> comes into play. mu4e is an email frontend for emacs that built on top of mu, which is an email indexer\/searcher. My reasons for why mu4e has greatly improved my email experience are:<\/p><ul><li>super fast and optimized UI\n<ul><li>keyboard driven actions<\/li>\n<li>bulk selection\/actions can leverage all of emacs\u2019 search\/marking functionality<\/li>\n<li>emacs text editing for writing email<\/li>\n<\/ul><\/li>\n<li>fast search supporting complex querying<\/li>\n<li>extensible because emacs :fireworks:<\/li>\n<li>keeps me in my workflow since I\u2019m already in emacs<\/li>\n<li>no distractions<\/li>\n<\/ul><p><em>Note: there are a number of other text based email clients which are similar to mu4e (another client I\u2019ve tried for a couple months was mutt), but I feel like mu4e is much better overall; it so much more customizable and fits really nicely with gmail\u2019s IMAP implementation (more on this later).<\/em><\/p><p>Getting mu4e up and running is not a difficult task, but it\u2019s definitely more work than you are probably used to, if you haven\u2019t tried other text based email clients before. The following describes my configuration and setup that hopefully you makes it a little easier for you to jump in.<\/p><p>My email setup consists of 2 gmail accounts, one personal and one for work, both of which are on Google Apps and secured via two-factor auth. The steps below are also tailored to macOS. If your setup is different, much of the following should still be relevant with a few tweaks.<\/p><p>mu4e\/mu works off of a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Maildir\">maildir<\/a>, so the first step is to actually get your email accounts setup to sync with your local maildir. It feels a little weird to be storing copies of your email locally in this day and age, but one often overlooked advantage to this is that you always have full access to your historical email (eg. when you don\u2019t have internet access). I also use my phone to check my email, so being able to sync email state between devices is a must; this rules out using POP and leaves me with IMAP. There are a number of applications that can be used for this: <a href=\"http:\/\/www.offlineimap.org\/\">offlineimap<\/a>, <a href=\"http:\/\/isync.sourceforge.net\/\">mbsync<\/a>, and others. I personally use offlineimap because it seems to offer the most documentation and exaples, and it seems to be the only one that supports XOAUTH2 which is the preferred way to access gmail accounts.<\/p><h2 id=\"gmail---special-snowflake\">gmail - Special Snowflake<\/h2><p>Before going into the configuration, you should be aware that the way gmail does IMAP is non-standard. In regular IMAP, each email can only exist once across all folders; gmail has labels instead of folders, and each email can be assigned multiple labels, so this uniqueness constraint clearly doesn\u2019t work. Instead, gmail has a notion of an \u201cAll Mail\u201d folder, which includes a copy of every single one of your emails (this includes your received, sent, drafts, trashed, etc.; the only exception is your spam which does not appear here). Your inbox and all your labels get translated into \u201cfolders\u201d and as a result you are going to have multiple copies of the same email. This wrecks havoc on regular IMAP clients, and as a result, special considerations need to be taken into account.<\/p><p>mu4e, has a special configuration option to hide duplicate copies of emails, but when actions are taken on an email (eg. archiving, trashing), only that copy gets affected, and as a result you run into weird consistency issues (this inconsistency persists until you perform two syncs; the first sync informs gmail of the changes so that it can resolve the inconsistency, and the second sync pulls the resolved state down). I\u2019ve ended up settling on an approach detailed by <a href=\"https:\/\/groups.google.com\/forum\/#!topic\/mu-discuss\/BpGtwVHMd2E\">Magnus Therning<\/a>, where I completely ignore the label \u201cfolders\u201d and rely almost completely on the \u201cAll Mail\u201d folder. This works extremely well because mu is essentially just an email indexer and searcher, and since all the gmail labels are also included as part of a header (as tags) in the email messages themselves, it\u2019s possible to reconstruct them by building custom search queries for viewing email as well as adding and removing these tags to \u201cchange\u201d the labels for email. This cleverly solves the issue of duplicated emails and also cuts down the amount of email that needs to be synced.<\/p><p>The following assumes this approach.<\/p><h2 id=\"offlineimap\">offlineimap<\/h2><p>First, install offlineimap<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">brew install offlineimap<\/pre><\/figure><p>The next step is to configure offlineimap. offlineimap also supports loading a python environment file so that configuration can be evaluated using python; any <code class=\"highlighter-rouge\">import<\/code>s and any other things such as custom function definitions should all live in that environment file. I take advantage of this feature to <a href=\"https:\/\/developer.apple.com\/legacy\/library\/documentation\/Darwin\/Reference\/ManPages\/man1\/security.1.html\">pull my oauth credentials from the macOS keychain<\/a> so that they don\u2019t need to be stored in plaintext in the config file. My python environment file lives at <code class=\"highlighter-rouge\">~\/.offlineimap.py<\/code> and contains:<\/p><figure class=\"highlight\"><table class=\"rouge-table\"><tbody><tr><td class=\"gutter gl\"><pre class=\"lineno\">1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n<\/pre><\/td><td class=\"code\"><pre>import json\nimport subprocess\ndef secure_string_for(account, service, value):\n    # this relies on the macOS `security` tool\n    # https:\/\/developer.apple.com\/legacy\/library\/documentation\/Darwin\/Reference\/ManPages\/man1\/security.1.html\n    return json.loads(subprocess.check_output([\"security\",\n                                               \"find-generic-password\",\n                                               \"-a\", account,\n                                               \"-s\", service,\n                                               \"-w\"]).strip())[value]<\/pre><\/td><\/tr><\/tbody><\/table><\/figure><p>offlineimap looks for your configuration at <code class=\"highlighter-rouge\">~\/.offlineimaprc<\/code>; my configuration is as follows and includes annotated comments (for complete documentation of the configuration options see <a href=\"https:\/\/github.com\/OfflineIMAP\/offlineimap\/blob\/master\/offlineimap.conf\">the example config<\/a>):<\/p><figure class=\"highlight\"><table class=\"rouge-table\"><tbody><tr><td class=\"gutter gl\"><pre class=\"lineno\">1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\n32\n33\n34\n35\n36\n37\n38\n39\n40\n41\n42\n43\n44\n45\n46\n47\n48\n49\n50\n51\n52\n53\n54\n55\n56\n57\n58\n59\n60\n61\n62\n63\n64\n65\n66\n67\n68\n69\n70\n71\n72\n73\n74\n75\n76\n77\n78\n79\n80\n81\n82\n83\n84\n85\n86\n87\n88\n89\n90\n91\n92\n93\n94\n95\n<\/pre><\/td><td class=\"code\"><pre>[general]\n# This lists the accounts that I will be configuring.\naccounts = nine27, do\n# offlineimap by default will fsync() to try and reduce chances for data\n# inconsistency. In most cases disabling this is fine, and also reduces \n# the number write cycles.\nfsync = false\n# This is the path to the aforementioned python environment that will be used.\npythonfile = ~\/.offlineimap.py\n# This helps prevent offlineimap hanging when network connection is lost.\nsocktimeout = 60\n# This is the account definition for my first account, \n# note the name matching that in the accounts list in [general].\n[Account nine27]\n# These repository names can be anything as they are only used to link the \n# local and remote repository config to the account.\nlocalrepository = nine27-local\nremoterepository = nine27-remote\n# Use sqlite instead of a flat file for better performance.\nstatus_backend = sqlite\n# This enables syncing the labels on the emails: \\Inbox, \\Flagged, \\Trashed, etc.\nsynclabels = yes\n# This is the header name that the labels are stored at, gmail uses X-Keywords.\nlabelsheader = X-Keywords\n# Here you can configure a custom script to be run each time a sync completes.\npostsynchook = ~\/.offlineimap\/hooks\/postsync-nine27\n# This is the local repository config (local meaning the local maildir).\n[Repository nine27-local]\n# Since this is a gmail account (special quirks), the type is GmailMaildir; \n# regular IMAP accounts should use Maildir.\ntype = GmailMaildir\n# This is the path to the maildir.\nlocalfolders = ~\/.mail\/nine27\n# This is the remote repository config (imap server)\n[Repository nine27-remote]\n# Since this is a gmail account, the type is Gmail; regular imap accounts should\n# use IMAP.\ntype = Gmail\n# gmail no longer supports regular LOGIN for auth, you either need to use OAuth, setup \n# access for insecure apps (https:\/\/support.google.com\/accounts\/answer\/6010255?hl=en),\n# or generate an app password (https:\/\/support.google.com\/accounts\/answer\/185833).\n# Accounts that use two-factor auth will need to use OAuth or an app password.\n# The following example is for OAuth, for instructions on setting this up see \n# https:\/\/github.com\/OfflineIMAP\/offlineimap\/blob\/master\/offlineimap.conf#L809-L865.\n# When using an insecure login or an app password, this should be LOGIN.\nauth_mechanisms = XOAUTH2\nremoteuser = <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"a0cec1cee0cec9cec592978ec3cfcd\">[email\u00a0protected]<\/a>\n# The following are the config options for OAuth login, note the secure_string_for\n# python method that I've defined in ~\/.offlineimap.py to securely fetch credentials from\n# the macOS keychain.\noauth2_client_id_eval = secure_string_for(\"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"87e9e6e9c7e9eee9e2b5b0a9e4e8ea\">[email\u00a0protected]<\/a>\", \"mail.nine27.com\", \"client_id\")\noauth2_client_secret_eval = secure_string_for(\"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"38565956785651565d0a0f165b5755\">[email\u00a0protected]<\/a>\", \"mail.nine27.com\", \"client_secret\")\noauth2_refresh_token_eval = secure_string_for(\"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"6f010e012f0106010a5d58410c0002\">[email\u00a0protected]<\/a>\", \"mail.nine27.com\", \"refresh_token\")\n# Location of the ssl ca certificate.\nsslcacertfile = \/usr\/local\/etc\/openssl\/cert.pem\n# Normally deleting an email would actually delete it, however gmail works differently\n# in that deleting an email just adds a \/Trashed label to it. So if you want to use the Trash\n# functionality of gmail, this should be set to no.\nrealdelete = no\n# This is evaluated python that determines which folders to sync. The key here is to ignore\n# the label \"folders\" and rely on only the All Mail folder and the other folders that map onto\n# mu4e's flag actions.\nfolderfilter = lambda folder: folder.startswith('[Gmail]\/') and \\\n    folder[8:] in ['All Mail', 'Trash', 'Drafts', 'Spam'] \ntrashfolder = [Gmail]\/Trash \n# The following is essentially the identical configuration, but for my work email.\n[Account do]\nlocalrepository = do-local\nremoterepository = do-remote\nstatus_backend = sqlite\nsynclabels = yes\nlabelsheader = X-Keywords\npostsynchook = ~\/.offlineimap\/hooks\/postsync-do\n[Repository do-local]\ntype = GmailMaildir\nlocalfolders = ~\/.mail\/do\n[Repository do-remote]\ntype = Gmail\nauth_mechanisms = XOAUTH2\nremoteuser = <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"9ff1e5f7f0f1f8dffbf6f8f6ebfef3f0fcfafef1b1fcf0f2\">[email\u00a0protected]<\/a>\noauth2_client_id_eval = secure_string_for(\"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"0f6175676061684f6b6668667b6e63606c6a6e61216c6062\">[email\u00a0protected]<\/a>\", \"mail.digitalocean.com\", \"client_id\")\noauth2_client_secret_eval = secure_string_for(\"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"3b55415354555c7b5f525c524f5a5754585e5a5515585456\">[email\u00a0protected]<\/a>\", \"mail.digitalocean.com\", \"client_secret\")\noauth2_refresh_token_eval = secure_string_for(\"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"bfd1c5d7d0d1d8ffdbd6d8d6cbded3d0dcdaded191dcd0d2\">[email\u00a0protected]<\/a>\", \"mail.digitalocean.com\", \"refresh_token\")\nssl=true\nsslcacertfile = \/usr\/local\/etc\/openssl\/cert.pem\nrealdelete = no\nfolderfilter = lambda folder: folder.startswith('[Gmail]\/') and \\\n    folder[8:] in ['All Mail', 'Trash', 'Drafts', 'Spam'] \ntrashfolder = [Gmail]\/Trash <\/pre><\/td><\/tr><\/tbody><\/table><\/figure><p>Once your configuration is in place, running<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">offlineimap -o<\/pre><\/figure><p>will kick off a sync and pull down all your email. This might take a while depending on how much email you have.<\/p><p>For those of you not familiar with syncing mail between a local maildir and a remote server, you should be aware that when you take any actions on your email (mark as read, archiving, trashing, etc.), the changes will not be propagated to the remote server until your next sync. This means that you need to run your sync periodically to keep your mail up to date. There are a number of approaches to doing this, but the most common approaches are:<\/p><ul><li>run offlineimap without the <code class=\"highlighter-rouge\">-o<\/code> flag and configuring it to sync periodically via the <code class=\"highlighter-rouge\">autorefresh<\/code> config option<\/li>\n<li>use something else like <code class=\"highlighter-rouge\">launchd<\/code> or <code class=\"highlighter-rouge\">cron<\/code> to periodically sync<\/li>\n<\/ul><p>I don\u2019t recommend using <code class=\"highlighter-rouge\">autorefresh<\/code> mostly because offlineimap has a tendency to randomly hang; it happens especially often when you lose\/regain internet connectivity (which happens all the time if you are on a laptop). This is pretty annoying because you are left thinking you have no new mail, when instead it\u2019s offlineimap not behaving. Running it periodically and specifying a <code class=\"highlighter-rouge\">socktimeout<\/code> value more or less relieved this issue for me.<\/p><p>Some other useful command line flags for offlineimap are:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">offlineimap -a &lt;account name&gt;<\/pre><\/figure><p>to sync a specific account, and<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">offlineimap -q<\/pre><\/figure><p>to do a quick sync (skip syncing the flags on emails).<\/p><p>In my configuration I\u2019ve set up postsync hooks for my accounts so that I can be notified of any new email that has been synced. For example, the postsync script for my personal account looks like:<\/p><figure class=\"highlight\"><table class=\"rouge-table\"><tbody><tr><td class=\"gutter gl\"><pre class=\"lineno\">1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n<\/pre><\/td><td class=\"code\"><pre>#!\/bin\/sh\n# Count new mail for every maildir\nmaildirnew=\"$HOME\/.mail\/nine27\/*\/new\"\nnew=\"$(find $maildirnew -type f | grep -v Gmail | wc -l | sed 's\/ \/\/g')\"\n# Notify me via a notification\nif [ $new -gt 0 ]\nthen\n    \/usr\/local\/bin\/reattach-to-user-namespace \/usr\/local\/bin\/terminal-notifier -title 'offlineimap' -subtitle '<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"9bf5faf5dbf5f2f5fea9acb5f8f4f6\">[email\u00a0protected]<\/a>' -message \"You have $new new email(s)!\" -group offlineimap-nine27\nfi<\/pre><\/td><\/tr><\/tbody><\/table><\/figure><p>Now that you are able to receive email, we can move on to configuring the sending of email. There are many tools for this similar to syncing, but the one I use and have seen the most configuration examples for is msmtp.<\/p><p>You can install it with:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">brew install msmtp<\/pre><\/figure><p>msmtp looks for its configuration at <code class=\"highlighter-rouge\">~\/.msmtprc<\/code>, and it is pretty straightforward to configure. My configuration with annotated comments where things are not self-explanatory is as follows:<\/p><figure class=\"highlight\"><table class=\"rouge-table\"><tbody><tr><td class=\"gutter gl\"><pre class=\"lineno\">1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\n<\/pre><\/td><td class=\"code\"><pre>account nine27\nhost smtp.gmail.com\nport 587\nprotocol smtp\nauth on\nfrom <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"4e202f200e2027202b7c79602d2123\">[email\u00a0protected]<\/a>\nuser <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"ed838c83ad83848388dfdac38e8280\">[email\u00a0protected]<\/a>\n# msmtp does not support OAuth, so instead you will have to use user\/pass. If you don't have\n# two factor auth enabled, enabling insecure apps to access your email should be sufficient and\n# you can use your normal password, but if you do have it enabled you will need to generate an\n# app password. See the comments in the offlineimap config for details.\n# Once again I'm using `security` to pull the password from my macOS keychain.\npasswordeval security find-internet-password -g -a <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"f09e919eb09e999e95c2c7de939f9d\">[email\u00a0protected]<\/a> -s smtp.gmail.com -w\ntls on\ntls_starttls on\n# If you don't have this cert, you can generate it by `brew install curl` and then running\n# `\/usr\/local\/Cellar\/curl\/&lt;curl version&gt;\/libexec\/mk-ca-bundle.pl`\ntls_trust_file ~\/.msmtp\/ca-bundle.crt\naccount do\nhost smtp.gmail.com\nport 587\nprotocol smtp\nauth on\nfrom <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"f39d899b9c9d94b3979a949a87929f9c9096929ddd909c9e\">[email\u00a0protected]<\/a>\nuser <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"4d23372522232a0d29242a24392c21222e282c23632e2220\">[email\u00a0protected]<\/a>\npasswordeval security find-internet-password -g -a <a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"95fbeffdfafbf2d5f1fcf2fce1f4f9faf6f0f4fbbbf6faf8\">[email\u00a0protected]<\/a> -s smtp.gmail.com -w\ntls on\ntls_trust_file ~\/.msmtp\/ca-bundle.crt\naccount default : nine27<\/pre><\/td><\/tr><\/tbody><\/table><\/figure><p>And that\u2019s it! A nice thing about accessing your mail via a local maildir and sending mail using a SMTP client, is that you can write mail and queue it up to be sent when you don\u2019t have an internet connection, and once you are connected the client will send it out.<\/p><p>Now that both receiving and sending email has been configured, we can put it all together using mu\/mu4e. You can install it with:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">brew install mu --with-emacs --HEAD<\/pre><\/figure><p><em>Note: The reason for building mu from HEAD is because the <code class=\"highlighter-rouge\">'mu4e-mark-execute-pre-hook<\/code> was not available in the latest release at the time of writing. This may no longer be the case.<\/em><\/p><p>The first thing you will want to do is index your email<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">mu index -m ~\/.mail<\/pre><\/figure><p>mu is pretty fast at doing this, but this might still take some time depending on how much email you have.<\/p><p>Now all that is left is to configure your emacs. A quick disclaimer, much of this config uses pieces that have been written by others, and I definitely don\u2019t claim to be using mu4e anywhere close to its full potential. My configuration only scratches the surface of the possible customizability (it\u2019s pretty incredible that even at this state the user experience is so great), but the following is my configuration with annotations where things are not self-explanatory:<\/p><figure class=\"highlight\"><table class=\"rouge-table\"><tbody><tr><td class=\"gutter gl\"><pre class=\"lineno\">1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\n32\n33\n34\n35\n36\n37\n38\n39\n40\n41\n42\n43\n44\n45\n46\n47\n48\n49\n50\n51\n52\n53\n54\n55\n56\n57\n58\n59\n60\n61\n62\n63\n64\n65\n66\n67\n68\n69\n70\n71\n72\n73\n74\n75\n76\n77\n78\n79\n80\n81\n82\n83\n84\n85\n86\n87\n88\n89\n90\n91\n92\n93\n94\n95\n96\n97\n98\n99\n100\n101\n102\n103\n104\n105\n106\n107\n108\n109\n110\n111\n112\n113\n114\n115\n<\/pre><\/td><td class=\"code\"><pre>;;; mu.el --- mu email config\n(use-package mu4e\n  :config\n  ;; This is a helper to help determine which account context I am in based \n  ;; on the folder in my maildir the email (eg. ~\/.mail\/nine27) is located in.\n  (defun mu4e-message-maildir-matches (msg rx)\n    (when rx\n      (if (listp rx)\n\t  ;; If rx is a list, try each one for a match\n\t  (or (mu4e-message-maildir-matches msg (car rx))\n\t      (mu4e-message-maildir-matches msg (cdr rx)))\n\t;; Not a list, check rx\n\t(string-match rx (mu4e-message-field msg :maildir)))))\n  ;; Choose account label to feed msmtp -a option based on From header\n  ;; in Message buffer; This function must be added to\n  ;; message-send-mail-hook for on-the-fly change of From address before\n  ;; sending message since message-send-mail-hook is processed right\n  ;; before sending message.\n  (defun choose-msmtp-account ()\n    (if (message-mail-p)\n\t(save-excursion\n\t  (let*\n\t      ((from (save-restriction\n\t\t       (message-narrow-to-headers)\n\t\t       (message-fetch-field \"from\")))\n\t       (account\n\t\t(cond\n\t\t ((string-match \"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"f39d899b9c9d94b3979a949a87929f9c9096929ddd909c9e\">[email\u00a0protected]<\/a>\" from) \"do\")\n\t\t ((string-match \"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"650b040b250b0c0b0057524b060a08\">[email\u00a0protected]<\/a>\" from) \"nine27\"))))\n\t    (setq message-sendmail-extra-arguments (list '\"-a\" account))))))\n  (setq mail-user-agent 'mu4e-user-agent)\n  (setq mu4e-mu-binary \"\/usr\/local\/bin\/mu\")\n  (setq mu4e-maildir \"~\/.mail\")\n  (setq mu4e-get-mail-command \"offlineimap -o\")\n  (setq mu4e-update-interval 300)\n  (setq mu4e-view-show-images t)\n  (setq mu4e-html2text-command \"w3m -dump -T text\/html\")\n  ;; This enables unicode chars to be used for things like flags in the message index screens.\n  ;; I've disabled it because the font I am using doesn't support this very well. With this\n  ;; disabled, regular ascii characters are used instead.\n  ;(setq mu4e-use-fancy-chars t)\n  ;; This enabled the thread like viewing of email similar to gmail's UI.\n  (setq mu4e-headers-include-related t)\n  (setq mu4e-attachment-dir  \"~\/Downloads\")\n  ;; This prevents saving the email to the Sent folder since gmail will do this for us on their end.\n  (setq mu4e-sent-messages-behavior 'delete)\n  (setq message-kill-buffer-on-exit t)\n  ;; Enable inline images.\n  (setq mu4e-view-show-images t)\n  ;; Use imagemagick, if available.\n  (when (fboundp 'imagemagick-register-types)\n    (imagemagick-register-types))\n  ;; Sometimes html email is just not readable in a text based client, this lets me open the\n  ;; email in my browser.\n  (add-to-list 'mu4e-view-actions '(\"View in browser\" . mu4e-action-view-in-browser) t)\n  ;; Spell checking ftw.\n  (add-hook 'mu4e-compose-mode-hook 'flyspell-mode)\n  ;; This hook correctly modifies the \\Inbox and \\Starred flags on email when they are marked.\n  ;; Without it refiling (archiving) and flagging (starring) email won't properly result in\n  ;; the corresponding gmail action.\n  (add-hook 'mu4e-mark-execute-pre-hook\n\t    (lambda (mark msg)\n\t      (cond ((member mark '(refile trash)) (mu4e-action-retag-message msg \"-\\\\Inbox\"))\n\t\t    ((equal mark 'flag) (mu4e-action-retag-message msg \"\\\\Starred\"))\n\t\t    ((equal mark 'unflag) (mu4e-action-retag-message msg \"-\\\\Starred\")))))\n  ;; This sets up my two different context for my personal and work emails.\n  (setq mu4e-contexts\n\t`( ,(make-mu4e-context\n\t     :name \"nine27\"\n\t     :enter-func (lambda () (mu4e-message \"Switch to the nine27 context\"))\n\t     :match-func (lambda (msg)\n\t\t\t   (when msg\n\t\t\t     (mu4e-message-maildir-matches msg \"^\/nine27\")))\n\t     :leave-func (lambda () (mu4e-clear-caches))\n\t     :vars '((user-mail-address     . \"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"48262926082621262d7a7f662b2725\">[email\u00a0protected]<\/a>\")\n\t\t     (user-full-name        . \"Nan Zhong\")\n\t\t     (mu4e-sent-folder      . \"\/nine27\/[Gmail].All Mail\")\n\t\t     (mu4e-drafts-folder    . \"\/nine27\/[Gmail].Drafts\")\n\t\t     (mu4e-trash-folder     . \"\/nine27\/[Gmail].Trash\")\n\t\t     (mu4e-refile-folder    . \"\/nine27\/[Gmail].All Mail\")))\n\t   ,(make-mu4e-context\n\t     :name \"do\"\n\t     :enter-func (lambda () (mu4e-message \"Switch to the do context\"))\n\t     :match-func (lambda (msg)\n\t\t\t   (when msg\n\t\t\t     (mu4e-message-maildir-matches msg \"^\/do\")))\n\t     :leave-func (lambda () (mu4e-clear-caches))\n\t     :vars '((user-mail-address     . \"<a href=\"https:\/\/notanumber.io\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"37594d5f58595077535e505e43565b58545256591954585a\">[email\u00a0protected]<\/a>\")\n\t\t     (user-full-name        . \"Nan Zhong\")\n\t\t     (mu4e-sent-folder      . \"\/do\/[Gmail].All Mail\")\n\t\t     (mu4e-drafts-folder    . \"\/do\/[Gmail].Drafts\")\n\t\t     (mu4e-trash-folder     . \"\/do\/[Gmail].Trash\")\n\t\t     (mu4e-refile-folder    . \"\/do\/[Gmail].All Mail\")))))\n  ;; Configure sending mail.\n  (setq message-send-mail-function 'message-send-mail-with-sendmail\n\tsendmail-program \"\/usr\/local\/bin\/msmtp\"\n\tuser-full-name \"Nan Zhong\")\n  ;; Use the correct account context when sending mail based on the from header.\n  (setq message-sendmail-envelope-from 'header)\n  (add-hook 'message-send-mail-hook 'choose-msmtp-account)\n  ;; Bookmarks for common searches that I use.\n  (setq mu4e-bookmarks '((\"\\\\\\\\Inbox\" \"Inbox\" ?i)\n\t\t\t (\"flag:unread\" \"Unread messages\" ?u)\n\t\t\t (\"date:today..now\" \"Today's messages\" ?t)\n\t\t\t (\"date:7d..now\" \"Last 7 days\" ?w)\n\t\t\t (\"mime:image\/*\" \"Messages with images\" ?p))))<\/pre><\/td><\/tr><\/tbody><\/table><\/figure><p>You\u2019re all done! <code class=\"highlighter-rouge\">M-x mu4e<\/code> and off you go! I won\u2019t go into how to actually use mu4e because there is simply too much to cover and the <a href=\"https:\/\/www.djcbsoftware.nl\/code\/mu\/mu4e\/\">official manual<\/a> does a much better job.<\/p><p>Hopefully this has been helpful to you. :)<\/p><p><img src=\"https:\/\/notanumber.io\/assets\/mu4e.png\" alt=\"mu4e\" \/><\/p><noscript><div>Please enable JavaScript to view the <a href=\"https:\/\/disqus.com\/?ref_noscript\">comments powered by Disqus.<\/a><\/div><\/noscript>","created_at":"2018-11-07T16:46:50+0000","updated_at":"2018-11-07T16:46:50+0000","published_at":"2016-10-03T21:51:54+0000","published_by":["Nan Zhong"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"notanumber.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":392,"title":"Blog | Rifle-Oriented Programming with Clojure","url":"http:\/\/thinkrelevance.com\/blog\/2009\/08\/12\/rifle-oriented-programming-with-clojure-2","content":"<h3 class=\"post-title\"><a href=\"http:\/\/thinkrelevance.com\/blog\/2009\/08\/12\/rifle-oriented-programming-with-clojure-2\">Rifle-Oriented Programming with Clojure<\/a><\/h3><p>Any comparison of hot JVM languages is likely to note that \u201cClojure is not object-oriented.\u201d This is true, but it may lead you to the wrong conclusions. It\u2019s a little like saying that a rifle is not arrow-oriented. In this article, you will see some of the ways that Clojure addresses the key concerns of OO: encapsulation, polymorphism, and inheritance.<\/p><p>This is a whirlwind tour, and we won't have time to cover the full details of all the Clojure code you will see. When we are done, I hope you will decide to explore for yourself. You can download and start using Clojure by following the instructions on the <a href=\"http:\/\/clojure.org\/getting_started\">getting started<\/a> page.<\/p><h2>Just Enough Clojure Syntax<\/h2><p>Clojure has <em>vectors<\/em>, which are accessed by integer indexes:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  [1 2 3 4]\n  -&gt; [1 2 3 4]\n  (get [:a :b :c :d :e] 2)\n  -&gt; :c<\/pre><\/div><\/div><p>In the preceding example, the initial `[1 2 3 4]` is input that you enter at the Read-Eval-Print Loop (REPL). The `-&gt;` indicates the response from the REPL.<\/p><p>Clojure has <em>maps<\/em>, which are key\/value collections:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  {:fname \"Stu\", :lname \"Halloway\"}\n  -&gt; {:fname \"Stu\", :lname \"Halloway\"}<\/pre><\/div><\/div><p><em>Sets<\/em> contain a set of values, and their literal form is preceded with a hash. Here is the set of English vowels, using backslash to introduce a character literal:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  #{\\a \\e \\i \\o \\u}\n  -&gt; #{\\a \\e \\i \\o \\u}<\/pre><\/div><\/div><p>Lists are singly-linked lists, and are enclosed with parentheses. Lists are special: Not only are they data, they also act as the syntax for invoking functions. The list below invokes the plus (`+`) function:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (+ 1 2 3 4 5)\n  -&gt; 15<\/pre><\/div><\/div><p>Collections themselves act as functions. They take an argument which is the key\/index to look up:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  ([:a :b :c :d :e] 2)\n  -&gt; :c\n  ({:name \"Stu\" :ext 101} :name)\n  -&gt; \"Stu\"<\/pre><\/div><\/div><p>Enough syntax, let's get started.<\/p><h2>Encapsulation<\/h2><p>Encapsulation is the hiding of implementation details so that clients of your code do not accidentally become dependent on them. In object-oriented languages, this is usually done at the class level. A class has <em>public<\/em> methods, <em>private<\/em> implementation details, and various other scopes in between.<\/p><p>Clojure accomplishes the purposes of encapsulation in three ways: closures, namespaces, and immutability.<\/p><h3 id=\"closures\">Closures<\/h3><p>A <em>closure<\/em> closes over (remembers) the environment at the time it was created. For example, the function `make-counter` below closes over the initial value passed via `init-val`:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defn make-counter [init-val] \n    (let [c (atom init-val)] #(swap! c inc)))<\/pre><\/div><\/div><p>Let\u2019s break this down:<\/p><ul><li>`defn` defines a new function, named `make-counter`, that takes a single argument `init-val`.<\/li>\n<li>The `let` binds the name `c` to a new `atom`.<\/li>\n<li>The `atom` creates a threadsafe, deadlock-proof mutable reference to a value.<\/li>\n<li>The octothorpe (`#`) prefix introduces an anonymous function<\/li>\n<li>The call to `swap!` updates the value referenced by `c` by calling `inc` on it.<\/li>\n<li>The value of the let is the value of its last expression. This `let` returns a function that increments a counter, which is then the return value of `make-counter`.<\/li>\n<\/ul><p>The atom `c` is private to the function returned by `make-counter`. The only <em>public<\/em> thing you can do is increment it by one:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (def c (make-counter))\n  -&gt; #'user\/c\n  (c)\n  -&gt; 1\n  (c)\n  -&gt; 2\n  (c)\n  -&gt; 3<\/pre><\/div><\/div><p>The counter example returned a single function, but nothing stops you from returning multiple functions. These multiple functions can then share private state. The new version of `make-counter` below returns two functions: one to increment the counter, and one to reset it.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defn make-counter [init-val] \n    (let [c (atom init-val)] \n      {:next #(swap! c inc)\n       :reset #(reset! c init-val)}))<\/pre><\/div><\/div><p>This new `make-counter` returns a map whose `:next` value increments the counter, and whose `:reset` value resets it:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (def c (make-counter 10))\n  -&gt; #'user\/c\n  ((c :next))\n  -&gt; 11\n  ((c :next))\n  -&gt; 12\n  ((c :reset))\n  -&gt; 10<\/pre><\/div><\/div><p>Why the double parentheses above? Two functions calls: The inner function call looks up the appropriate function, and the outer one calls it.<\/p><p>Closing over data is far more general than the simplistic model offered by private, protected, public, friend, et al. in OO languages. By combining multiple lets and multiple return values from a function, you can create arbitrary encapsulation strategies.<\/p><p>Similar encapsulation possibilities are available in any language that supports closures. Douglas Crockford describes a <a href=\"http:\/\/www.crockford.com\/javascript\/private.html\">similar idiom<\/a> in JavaScript.<\/p><h3 id=\"namespaces\">Namespaces<\/h3><p>A Clojure namespace groups a set of related data and functions. Inside a namespace, a Clojure <em>var<\/em> can refer to a function or to data, and can be public or private.<\/p><p>For example, Chris Houser\u2019s error-kit library implements a condition\/restart system for Clojure.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (with-handler\n    (vec (map int-half [2 4 5 8]))\n      (handle *number-error* [n]\n        (continue-with 0))) <\/pre><\/div><\/div><p>In the code above, `with-handler`, `handle`, and `continue-with` are public vars of the `clojure.contrib.error-kit` namespace. The `int-half` is a demo function that blows up on odd inputs. When a `*number-error*` occurs, the handler causes execution to continue with the value 0. (Note how this is more flexible than try\/catch exception handling, which cannot recover back into the middle of some operation.)<\/p><p>Internally, error-kit keeps track of available handlers and continues using these private vars:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defvar- *handler-stack* () \n    \"Stack of bound handler symbols\")\n  (defvar- *continues* {} \n    \"Map of currently available continue forms\")<\/pre><\/div><\/div><p>The trailing minus sign on the end of `defvar-` marks the vars as private. These vars are implementation details, and are invisible to code outside the `clojure.contrib.error-kit` namespace.<\/p><h3 id=\"immutability\">Immutability<\/h3><p>In OO languages, another purpose of encapsulation is to prevent object A from modifying or corrupting the private data used by object B.<\/p><p>In Clojure, this problem does not exist. Data structures are immutable. They cannot possibly be corrupted, or changed in any way, period. You can write query functions that return \u201cprivate\u201d state, without any fear of data corruption. <\/p><h2 id=\"polymorphism\">Polymorphism<\/h2><p>For our purposes here, polymorphism is the ability to choose a different method implementation based on the type of the caller. So for example:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  Flyer a = new Airplane();\n  Flyer b = new Bird();\n  a.fly();\n  b.fly();<\/pre><\/div><\/div><p>`a.fly()` and `b.fly()` do different things because they are called on different concrete types.<\/p><p>Clojure provides a generalization of polymorphism called multimethods. A multimethod definition begins with `defmulti`, and then has a name, plus a <em>dispatch function<\/em> that is used to select the actual implementation: To mimic polymorphism, simply dispatch on the `class` of the argument:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defmulti fly class)<\/pre><\/div><\/div><p>Individual methods of a multimethod begin with `defmethod`, then the multimethod name, then the object that must match the dispatch function.  Finally, you get the argument list in a vector, followed by the implementation of the method. For example:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defmethod fly Bird [b] (flap-wings b))\n  (defmethod fly Airplane [a] (turn-propeller a))<\/pre><\/div><\/div><p>Unlike polymorphism, multimethods do not limit you to dispatching on class. You can dispatch based on any arbitrary function of the method arguments. So for example, a bank account might have a `:type` entry that is used to determine the interest rate:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defmulti interest :type)\n  (defmethod interest :checking [a] 0)\n  (defmethod interest :savings [a] 0.05M)<\/pre><\/div><\/div><p>The `:type` attribute is a convention, but nothing prevents you from dispatching on a different attribute, or even dispatching on more than one at the same time! For example, the `service-charge` multimethod below dispatches on two different facets of the same object: the object\u2019s `account-level` (`::Basic` or `::Premium`) and its `:tag:` (`::Checking` or `::Savings`)<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defmulti service-charge \n    (fn [acct] [(account-level acct) (:tag acct)]))\n  (defmethod service-charge [::Basic ::Checking]   [_] 25)\n  (defmethod service-charge [::Basic ::Savings]    [_] 10)\n  (defmethod service-charge [::Premium ::Checking] [_] 0)\n  (defmethod service-charge [::Premium ::Savings]  [_] 0)<\/pre><\/div><\/div><p>The `_` is a legal name, and is used idiomatically to indicate that an argument will be ignored. (There is no need to even look at the argument, since all the work has been done in choosing which method to dispatch to!) This example also demonstrates two other concepts:<\/p><ul><li>The double-colon prefix resolves a keyword in a namespace. This prevents name collisions among <em>keywords<\/em>, just as object-oriented langauges use namespaces to prevent name collisions between <em>type names<\/em>.<\/li>\n<li>`account-level` is a function (not shown here), not a simple key lookup. It returns `::Premium` or `::Basic` based on the the account type and the current balance. Thus an account can dynamically change its account level as its balance changes.<\/li>\n<\/ul><p>As you can see, multimethods are far more general than polymorphism. Instead of being limited to type-based dispatch, multimethods can dispatch on any arbitrary function of an argument list. This allows programming models that more closely resemble reality: after all, what real-world entities are limited to a single type hierarchy, and forbidden to change types over time?<\/p><h3 id=\"inheritance\">Inheritance<\/h3><p>In OO languages, inheritance allows you to create a <em>derived<\/em> type that reuses the behavior of a <em>base<\/em> type. For example:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  class Person {\n    String fullName() { \/* impl details *\/ }\n  }\n  class Employee extends Person {\n    AddressBookItem companyDirectoryEntry() { \/* impl details *\/ }\n  }<\/pre><\/div><\/div><p>This kind of reuse is so natural in Clojure that it doesn\u2019t even have a name. For example, here is a function that returns the full name of a person, based on first and last names:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defn full-name [p]\n    (str (:first-name p) \" \" (:last-name p)))<\/pre><\/div><\/div><p>Employees are like people, but have other properties and behaviors, such as a telephone extension. The `company-directory-entry` returns a vector of an employee's full name and telephone extension, like this:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defn company-directory-entry [p]\n    [(full-name p) (:extension p)])<\/pre><\/div><\/div><p>Notice that `company-directory-entry` \u201creuses\u201d the person-ness of its argument `p` by calling `full-name` on it. There is no special inheritance ceremony required to set this up, you just call functions when you need them.<\/p><p>You can pass either a person or an employee to `full-name`. For `company-directory-entry`, though, you must have an employee. Or, more accurately, you must have something that <em>resembles<\/em> an employee, to the extent of having a `:first-name`, `:last-name`, and `:extension`. This is an example of <em>duck typing<\/em>: if it walks like a duck and quacks like a duck, we assume it <em>is<\/em> a duck, without asking it to present its `IDuck` papers.<\/p><h3>Many Functions, Few Types<\/h3><p>The example above demonstrates another negative consequence of idiomatic OO style: the over-specification of data types. The return value of `companyDirectoryEntry` is given its own unique type, `AddressBookItem`. Each new data type like `AddressBookItem` requires its own life-support system: constructors, accessors, `equals`, `hashCode`, and so on.<\/p><p>In Clojure, an address book item would simply be a vector or a map. No new types, and no life support system required. Moreover, an address book item can be manipulated with any of the large arsenal of functions in Clojure's <em>sequence library<\/em>.<\/p><p>To see the problem with overspecifying types, consider this method from the Apache Commons:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  \/\/ From Apache Commons Lang, http:\/\/commons.apache.org\/lang\/\n  public static int indexOfAny(String str, char[] searchChars) {\n      if (isEmpty(str) || ArrayUtils.isEmpty(searchChars)) {\n    return -1;\n      }\n      for (int i = 0; i &lt; str.length(); i++) {\n    char ch = str.charAt(i);\n    for (int j = 0; j &lt; searchChars.length; j++) {\n        if (searchChars[j] == ch) {\n        return i;\n        }\n    }\n      }\n      return -1;\n  }<\/pre><\/div><\/div><p>The purpose of `indexOfAny` is to find the index of the first occurrence of one of the `searchChars` that appears in `str`. Note the unnecessary specificity of types: it works only with strings and character arrays.<\/p><p>Here's the Clojure version, using the sequence library's `map`, `iterate`, and `for` forms:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (defn indexed [coll] (map vector (iterate inc 0) coll))\n  (defn index-filter [pred coll]\n    (when pred \n      (for [[idx elt] (indexed coll) :when (pred elt)] idx)))<\/pre><\/div><\/div><p>Here is an example calling `index-filter`:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (index-filter #{\\a \\e \\i \\o \\u} \"Lts f cnsnts nd n vwel\")\n  -&gt; (20)<\/pre><\/div><\/div><p>The expression above finds the index of the first vowel in the string \"Lts f cnsnts nd n vwel\", that is, 20. But `index-filter` is more general than the Commons version in several ways:<\/p><p>1. `index-filter` returns all the matches, not just one.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (index-filter #{\\a \\e \\i \\o \\o} \"The quick brown fox\")\n  -&gt; (2 6 12 17)<\/pre><\/div><\/div><p>2. `index-filter` works with any sequence, not just a string of characters. For example, the call below works against a range of integers:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (index-filter #{2 3 5 7} (range 6))\n  -&gt; (2 3 5)<\/pre><\/div><\/div><p>3. `index-filter` works with any predicate, not just a test against a character array. In the example below, the predicate is an anonymous function that tests for strings longer than three characters:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>  (index-filter #(&gt; (.length %) 3) [\"The\" \"quick\" \"brown\" \"fox\"])\n  -&gt; (1 2)<\/pre><\/div><\/div><p>That is a lot of extra power, especially given that the function is shorter, easier to write, and easier to read (given some Clojure experience, of course) than the Commons version.<\/p><h2>Conclusion<\/h2><p>Clojure solves the same problems that OO solves, but it solves them in different ways. Instead of encapsulation, polymorphism, and inheritance, you have closures, namespaces, pure functions, immutable data, and multimethods. Idiomatic OO gives you a bloated type system with duplicated code hidden away behind encapsulation boundaries and little hope for thread safety. Clojure offers a radical alternative: a lean type system, a rich function library, and language-level concurrency support that is usable by mere mortals.<\/p><p>There is a lot more to Clojure than we have covered here: lazy and infinite sequences, destructuring, macros, software transactional memory, agents, seamless Java interop, and more. But those are topics for another day.<\/p><p>[This article was originally published in the May 2009 issue of <a href=\"http:\/\/www.nofluffjuststuff.com\/home\/magazine_subscribe\">NFJS, the Magazine<\/a>. I will be speaking about Clojure at several upcoming NFJS events, <a href=\"http:\/\/www.nofluffjuststuff.com\">come join the fun<\/a>.]<\/p>","created_at":"2018-11-07T16:46:50+0000","updated_at":"2018-11-07T16:46:50+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"thinkrelevance.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":388,"title":"A Handcrafted Feel: \u2018Unexplored\u2019 Explores Cyclic Dungeon Generation","url":"http:\/\/ctrl500.com\/tech\/handcrafted-feel-dungeon-generation-unexplored-explores-cyclic-dungeon-generation\/","content":"<p><strong>Stop creating branching paths, start using cyclic dungeon generation. Your levels will feel so much more hand crafted.<\/strong><\/p><p>The most common strategy to generate dungeons in roguelike games is to keep adding generated or pre-designed dungeon pieces to the map. This way a dungeon is grown from a starting point much like a tree. However, trees end in branches and this creates a lot of dead-ends. To get around this problem most dungeon generators look for places to randomly reconnect the branches so that the player might actually wander in circles instead of having to backtrack all the time.<\/p>\n<p>The dungeon crawler <a href=\"http:\/\/store.steampowered.com\/app\/506870\/\" target=\"_blank\" rel=\"noopener noreferrer\"><em>Unexplored<\/em><\/a> changes this fundamentally (see figure 1). Instead of taking linear paths as its most basic structure, it generates everything in cycles from the get go. The differences are striking: where good cycles and interesting cycles might randomly appear using the old technique, in <a href=\"http:\/\/store.steampowered.com\/app\/506870\/\" target=\"_blank\" rel=\"noopener noreferrer\">Unexplored<\/a> they a planned feature of the generator\u2019s output.<\/p>\n<p><a href=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png\"><img data-attachment-id=\"3368\" data-permalink=\"https:\/\/ctrl500.com\/tech\/handcrafted-feel-dungeon-generation-unexplored-explores-cyclic-dungeon-generation\/attachment\/basic-cyclic-explantion\/\" data-orig-file=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?fit=2446%2C2094&amp;ssl=1\" data-orig-size=\"2446,2094\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Basic cyclic explantion\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?fit=300%2C257&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?fit=750%2C642&amp;ssl=1\" class=\"aligncenter wp-image-3368 size-large\" src=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion-1024x877.png?resize=750%2C642\" alt=\"Basic cyclic explantion\" srcset=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?resize=1024%2C877&amp;ssl=1 1024w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?resize=300%2C257&amp;ssl=1 300w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?resize=768%2C657&amp;ssl=1 768w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?w=1500&amp;ssl=1 1500w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/basic-cyclic-explantion.png?w=2250&amp;ssl=1 2250w\" data-recalc-dims=\"1\" \/><\/a><\/p>\n<p><em>Figure 1 \u2013 Cycles versus trees.<\/em><\/p>\n<p>So how does this work? When <a href=\"http:\/\/store.steampowered.com\/app\/506870\/\" target=\"_blank\" rel=\"noopener noreferrer\">Unexplored<\/a> starts generating a dungeon it starts from an arbitrary starting point somewhere on the map. But instead of simply creating a branch to a target point it creates a cycle consisting of two paths connecting to the entrance to the goal. These cycles accommodate basic level design patterns far better than linear paths do.<\/p>\n<p>Each cycle has a path \u2018a\u2019 that leads from the entry point towards the goal, and a path \u2018b\u2019 that leads back. Depending on the absolute and relative lengths of \u2018a\u2019 and \u2018b\u2019 you can change each cycle into a design pattern (see figure 2). For example when \u2018a\u2019 is short and \u2018b\u2019 is long, you can place a locked door at the cycles goal, and place the key at the end of \u2018b\u2019 just before you can cut back to the main path. This creates a level where players first encounter the locked door, and when they do find the right key they do not have to go far to reach the door it unlocks. Obviously, this requires that the player cannot reach \u2018b\u2019 from the cycle\u2019s entrance straight away, although it might be interesting if they can already see \u2018b\u2019 and the key.<\/p>\n<p>Another example that works well with a long \u2018a\u2019 and a short \u2018b\u2019 is to put a secret door between the entrance and \u2018b\u2019, creating a hidden shortcut. Or maybe \u2018b\u2019 is simply a more dangerous route then \u2018a\u2019, taking the player past a dangerous foe or requiring the player to navigate a trap infested room. There are many possibilities, figure 2 lists a couple more.<\/p>\n<p><a href=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png\"><img data-attachment-id=\"3369\" data-permalink=\"https:\/\/ctrl500.com\/tech\/handcrafted-feel-dungeon-generation-unexplored-explores-cyclic-dungeon-generation\/attachment\/cyclic_patterns\/\" data-orig-file=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?fit=2534%2C2086&amp;ssl=1\" data-orig-size=\"2534,2086\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Cyclic patterns\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?fit=300%2C247&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?fit=750%2C617&amp;ssl=1\" class=\"aligncenter wp-image-3369 size-large\" src=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns-1024x843.png?resize=750%2C617\" alt=\"Cyclic patterns\" srcset=\"https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?resize=1024%2C843&amp;ssl=1 1024w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?resize=300%2C247&amp;ssl=1 300w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?resize=768%2C632&amp;ssl=1 768w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?w=1500&amp;ssl=1 1500w, https:\/\/i0.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/cyclic_patterns.png?w=2250&amp;ssl=1 2250w\" data-recalc-dims=\"1\" \/><\/a><\/p>\n<p><em>Figure 2 \u2013 a couple of cyclic design patterns.<\/em><\/p>\n<p>The results are awesome. Figure 3 shows a level generated by <a href=\"http:\/\/store.steampowered.com\/app\/506870\/\" target=\"_blank\" rel=\"noopener noreferrer\">Unexplored<\/a>. It consists of a large lock and key cycle with\u00a0the cycle filled\u00a0by a large chasm. A smaller gambit cycle was added to create more spice. The construction is simple but effective. In fact, we\u2019ve come to realize that just a couple of cycles often is enough for most levels; with just a couple of cycles the levels end up with a distinct shape and character.<\/p>\n<p><a href=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png\"><img data-attachment-id=\"3370\" data-permalink=\"https:\/\/ctrl500.com\/tech\/handcrafted-feel-dungeon-generation-unexplored-explores-cyclic-dungeon-generation\/attachment\/the-ring-level\/\" data-orig-file=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?fit=1061%2C644&amp;ssl=1\" data-orig-size=\"1061,644\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"the ring level\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?fit=300%2C182&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?fit=750%2C456&amp;ssl=1\" class=\"aligncenter wp-image-3370 size-large\" src=\"https:\/\/i2.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level-1024x622.png?resize=750%2C456\" alt=\"the ring level\" srcset=\"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?resize=1024%2C622&amp;ssl=1 1024w, https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?resize=300%2C182&amp;ssl=1 300w, https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?resize=768%2C466&amp;ssl=1 768w, https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/the-ring-level.png?w=1061&amp;ssl=1 1061w\" data-recalc-dims=\"1\" \/><\/a><\/p>\n<p><em>Figure 3 \u2013 a level walkthrough.<\/em><\/p>\n<p><em><strong>Acknowledgements: The idea of cyclic dungeon creation arose during a research workshop at the <a href=\"https:\/\/www.banffcentre.ca\/visual-and-digital-arts\" target=\"_blank\" rel=\"noopener noreferrer\">Banff Center<\/a> in Canada. I cannot take sole credit for this idea, instead I share it with everyone who was in that particular workgroup.<\/strong><\/em><\/p>","created_at":"2018-11-07T16:46:49+0000","updated_at":"2018-11-07T16:46:49+0000","published_at":"2016-07-04T10:32:00+0000","published_by":["John Sweeney (Naughty Dog)","Joris Dormans","Roy Van De Mortel (Digital Dreams)","Richard Boeser (Sparpweed)","Jan Willem Nijman (Vlambeer)","Ard Jacobs","Eric Bartelson","Tim Baijens","Joris Dormans","David Jimenez (2Awesome Studio)"],"annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"ctrl500.com","preview_picture":"https:\/\/i1.wp.com\/ctrl500.com\/wp-content\/uploads\/2016\/07\/UNEXPLORED-LEVEL-BT.png?fit=1537%2C865&ssl=1","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":389,"title":"Cyclic Dungeon Generation explained in 47 seconds","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=_wvkTT-6P3Q","content":"<iframe width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/_wvkTT-6P3Q?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:46:49+0000","updated_at":"2018-11-07T16:46:49+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/_wvkTT-6P3Q\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":387,"title":"Joris Dormans - Cyclic Dungeon Generation","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=mA6PacEZX9M","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/mA6PacEZX9M?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:46:40+0000","updated_at":"2018-11-07T16:46:40+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/mA6PacEZX9M\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":386,"title":"Effective_Engineer.md","url":"https:\/\/gist.github.com\/rondy\/af1dee1d28c02e9a225ae55da2674a6f","content":"<p><em>FWIW: I didn't produce the content presented here (the outline from Edmond Lau's book). I've just copy-pasted it from somewhere over the Internet, but I cannot remember what exactly the original source is. I was also not able to find the author's name, so I cannot give him\/her the proper credits.<\/em><\/p>\n<hr \/><h1><a id=\"user-content-effective-engineer---notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#effective-engineer---notes\"><\/a>Effective Engineer - Notes<\/h1>\n<ul><li>By Edmond Lau<\/li>\n<li>Highly Recommended \ud83d\udc4d<\/li>\n<li><a href=\"http:\/\/www.theeffectiveengineer.com\/\" rel=\"nofollow\">http:\/\/www.theeffectiveengineer.com\/<\/a><\/li>\n<\/ul><h3><a id=\"user-content-whats-an-effective-engineer\" class=\"anchor\" aria-hidden=\"true\" href=\"#whats-an-effective-engineer\"><\/a>What's an Effective Engineer?<\/h3>\n<ul><li>They are the people who get things done. Effective Engineers produce results.<\/li>\n<\/ul><h2><a id=\"user-content-adopt-the-right-mindsets\" class=\"anchor\" aria-hidden=\"true\" href=\"#adopt-the-right-mindsets\"><\/a>Adopt the Right Mindsets<\/h2>\n<h3><a id=\"user-content-focus-on-high-leverage-activities\" class=\"anchor\" aria-hidden=\"true\" href=\"#focus-on-high-leverage-activities\"><\/a>Focus on High Leverage Activities<\/h3>\n<ul><li>Leverage = Impact Produced \/ Time Invested<\/li>\n<li>Use Leverage as Your Yardstick for Effectiveness<\/li>\n<li>80% of the impact comes from 20% of the work.<\/li>\n<li>Focus on high leverage and not just easy wins.<\/li>\n<\/ul><h3><a id=\"user-content-optimize-for-learning\" class=\"anchor\" aria-hidden=\"true\" href=\"#optimize-for-learning\"><\/a>Optimize for Learning<\/h3>\n<ul><li>Change jobs if you have to.<\/li>\n<li>Optimizing for learning is high leverage.<\/li>\n<li>Adopt a growth mindset.\n<ul><li>Talk to people. Become good at telling stories. It gets better with time.<\/li>\n<li>Those with a growth mindset believe that they can cultivate and grow their intelligence and skills through effort.<\/li>\n<li>Own your story.<\/li>\n<\/ul><\/li>\n<li>Invest in the rate of learning\n<ul><li>Learning compounds. Compounding leads to exponential growth. Earlier the compounding starts, the better.<\/li>\n<li>Working on unchallenging tasks is a huge opportunity cost. You missed out on compounded learning.<\/li>\n<li>Prioritize learning over profitability.<\/li>\n<li>Invest your time in activities with the highest learning rate.<\/li>\n<\/ul><\/li>\n<li>Seek Work Environments Conducive to Learning\n<ul><li>Fast Growth: Companies where #problems &gt;&gt; #resources. Opportunity to choose high impact work.<\/li>\n<li>Make sure you are working on high priority projects.<\/li>\n<li>Openness: Look for culture with curiosity, where everyone is encouraged to ask questions.<\/li>\n<li>Fast Paced.<\/li>\n<li>People smarter than you.<\/li>\n<li>Autonomy: Freedom to choose what to work on. Smaller companies =&gt; More autonomy.<\/li>\n<\/ul><\/li>\n<li>While on Job\n<ul><li>Make a daily habit of acquiring new skills.<\/li>\n<li>Read code written by brilliant engineers.<\/li>\n<li>Jump fearlessly into code you don't know.<\/li>\n<li>Always be learning. Invest in skills that are in high demand.<\/li>\n<li>Read Books. Attend Conferences.<\/li>\n<li>Build and maintain strong relationships.<\/li>\n<\/ul><\/li>\n<\/ul><h3><a id=\"user-content-prioritize-regularly\" class=\"anchor\" aria-hidden=\"true\" href=\"#prioritize-regularly\"><\/a>Prioritize Regularly<\/h3>\n<ul><li>Opportunity cost of working on wrong ideas can set back growth by years.<\/li>\n<li>Prioritize tasks based on ROI.<\/li>\n<li>Regular prioritization is high leverage activity.<\/li>\n<li>On TODO Lists:\n<ul><li>Maintain a 'single' todo lists where all tasks are listed.<\/li>\n<li>Don't try to remember stuff. Brain is bad at remembering. It's rather good at processing.<\/li>\n<\/ul><\/li>\n<li>Ask yourself regularly: Is this the most important thing I should be working on?<\/li>\n<li>Focus on what directly produces value.<\/li>\n<li>Learn to say no.<\/li>\n<li>Focus on the important and non-urgent.<\/li>\n<li>Find ways to get into flow. \u201cA state of effortless concentration so deep that they lose their sense of time, of themselves, of their problems.\u201d<\/li>\n<li>When possible, preserve larger blocks of focused time in your schedule.<\/li>\n<li>Limit the amount of Work in Progress.\n<ul><li>Cost of context switching is high.<\/li>\n<\/ul><\/li>\n<li>Prioritizing is difficult.<\/li>\n<li>Prioritization is high leverage. It has huge impact on your ability to get right things done.<\/li>\n<\/ul><h3><a id=\"user-content-invest-in-iteration-speed\" class=\"anchor\" aria-hidden=\"true\" href=\"#invest-in-iteration-speed\"><\/a>Invest in Iteration Speed<\/h3>\n<ul><li>Continuous Deployment is high leverage.\n<ul><li>Will save a lot of time in manual deployment of code. They are the people who get things done. Effective Engineers produce results.<\/li>\n<\/ul><\/li>\n<li>Move fast to learn fast.\n<ul><li>Move fast and break things.<\/li>\n<li>Moving fast enables us to build more things and learn at faster rate.<\/li>\n<\/ul><\/li>\n<li>Invest in time saving tools.\n<ul><li>If you have to do something more than twice, write a tool the third time.<\/li>\n<li>Tools are multipliers that allow your to scale your impact beyond the confines of a day.<\/li>\n<li>Faster tools get used more often.<\/li>\n<li>Faster tools can enable new workflows that previously weren't possible.<\/li>\n<li>Productivity skyrockets with tools.<\/li>\n<li>Time saving property of tools also scale with team adoption.<\/li>\n<\/ul><\/li>\n<li>Shorten your debugging and validation Loops.\n<ul><li>Extra time spent in optimizing debugging workflow can help you fix annoying bugs with less headache.<\/li>\n<li>Debugging is hard. It's time consuming. Upfront investments to shorten debugging loops are worth it.<\/li>\n<\/ul><\/li>\n<li>High test coverage to reduce build and site breakages.<\/li>\n<li>Fast unit tests to encourage people to run them.<\/li>\n<li>Fast and incremental compiles and reloads to reduce development time.<\/li>\n<li>Master you programming environment.\n<ul><li>One editor. One high level language. Shell. Keyboard &gt; Mouse. Automate manual workflows. Use interactive shell. Make running specific tests easy.<\/li>\n<\/ul><\/li>\n<li><strong>Faster you can iterate, faster you can learn.<\/strong><\/li>\n<\/ul><h3><a id=\"user-content-measure-what-you-want-to-improve\" class=\"anchor\" aria-hidden=\"true\" href=\"#measure-what-you-want-to-improve\"><\/a>Measure what you want to Improve<\/h3>\n<ul><li>Use metric to drive progress.\n<ul><li>If you can't measure it, you can't improve it.<\/li>\n<li>Good metric.\n<ul><li>Helps you focus on right things.<\/li>\n<li>Drives forward progress.<\/li>\n<li>Helps you guard against future regressions.<\/li>\n<li><strong>Performance ratcheting<\/strong>: Any change should strictly improve the metric.<\/li>\n<li>Bad metric can lead to unwanted behavior.<\/li>\n<li>Examples:\n<ul><li>#hours worked &lt; productivity.<\/li>\n<li>click through rates &lt; long click through rates.<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li>Metric you choose influences your decisions and behavior.<\/li>\n<li>Look for metric that, when optimized, maximizes impact for the team.<\/li>\n<li>Actionable metric - Whose movement can be casually explained by team's effort.<\/li>\n<li>Responsive metric - Updates quickly to give back feedback whether a given change was =ve or -ive.<\/li>\n<li>Choosing a metric is high leverage.<\/li>\n<li>Dedicate time to pick right metric.<\/li>\n<\/ul><\/li>\n<li>Instrument everything to understand what's going on.\n<ul><li>Measure anything, measure everything.<\/li>\n<li>Graphite, statsd. A single line of code lets you define a new counter or timer on the fly.<\/li>\n<li>Measuring goals you want to achieve is high leverage.<\/li>\n<\/ul><\/li>\n<li>Internalize useful numbers.\n<ul><li>Knowledge of useful numbers provide a valuable shortcut for knowing where to invest efforts to maximize gains.<\/li>\n<li>Need upfront work. Need not be accurate, ballpark idea suffices.<\/li>\n<li>Knowing useful numbers enables you to do back of the envelope calculations to quickly estimate the performance properties of a design without actually building it.<\/li>\n<li>Internalizing useful number help you spot anomalies.\nBe skeptical about data integrity.<\/li>\n<\/ul><\/li>\n<li>Log data liberally.<\/li>\n<li>Build tools to iterate on data accuracy sooner.<\/li>\n<li>Examine data sooner.<\/li>\n<li>When numbers look off, dig in to it sooner.<\/li>\n<\/ul><p>\u2714\ufe0f Measure your progress. Carefully choose your top-level metric. Instrument your system. Know your numbers. Prioritize data integrity.<\/p>\n<h3><a id=\"user-content-validate-your-ideas-early-and-often\" class=\"anchor\" aria-hidden=\"true\" href=\"#validate-your-ideas-early-and-often\"><\/a>Validate your ideas early and often.<\/h3>\n<ul><li>Not validating early leads to wasted efforts.<\/li>\n<li>Don't delay get feedback.<\/li>\n<li>Find low effort ways to validate work.<\/li>\n<li>Power of small batches. Helps you avoid making a big mistake by stopping the flow.<\/li>\n<li>Approach problem iteratively.<\/li>\n<li>No large implementations.<\/li>\n<li>Working solo? Be wary. Be extra vocal and get feedback.<\/li>\n<\/ul><h3><a id=\"user-content-improve-project-estimation-skills\" class=\"anchor\" aria-hidden=\"true\" href=\"#improve-project-estimation-skills\"><\/a>Improve project estimation skills.<\/h3>\n<ul><li>Beware of mythical man month. Communication overhead is significant.<\/li>\n<li>Reduce risk early.<\/li>\n<li>Rewrite projects - almost always fail.<\/li>\n<li>Additional hours hurt productivity. Causes burnout.<\/li>\n<li>Do the riskiest task first.<\/li>\n<li>Allow buffer room for the unknown.<\/li>\n<\/ul><h3><a id=\"user-content-balance-quality-with-pragmatism\" class=\"anchor\" aria-hidden=\"true\" href=\"#balance-quality-with-pragmatism\"><\/a>Balance Quality with Pragmatism<\/h3>\n<ul><li>High code quality. Code readability.<\/li>\n<li>Establish sustainable code review process.<\/li>\n<li>Code reviews help:\n<ul><li>Catch bugs and design problems early.<\/li>\n<li>Sharing working knowledge of the codebase.<\/li>\n<li>Increases long term agility. Easier to understand, quicker to modify.<\/li>\n<\/ul><\/li>\n<\/ul><h4><a id=\"user-content-manage-complexity-through-abstraction\" class=\"anchor\" aria-hidden=\"true\" href=\"#manage-complexity-through-abstraction\"><\/a>Manage complexity through Abstraction<\/h4>\n<ul><li>Example: MapReduce.<\/li>\n<li>Right abstractions make huge difference.<\/li>\n<li>\u201cPick the right ones, and programming will flow naturally from design; modules will have small and simple interfaces; and new functionality will more likely fit in without extensive reorganization,\u201d<\/li>\n<li>\u201cPick the wrong ones, and programming will be a series of nasty surprises: interfaces will become baroque and clumsy as they are forced to accommodate unanticipated interactions, and even the simplest of changes will be hard to make.\u201d<\/li>\n<li>The right abstraction can increase engineering productivity by an order of magnitude.<\/li>\n<li>Simple abstractions avoid interweaving multiple concepts, so that you can reason about them independently rather than being forced to consider them together.<\/li>\n<li>Designing good abstractions take work.<\/li>\n<li>An abstraction's usage and popularity provides a reasonable proxy for its quality.<\/li>\n<\/ul><h4><a id=\"user-content-automate-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#automate-testing\"><\/a>Automate Testing<\/h4>\n<ul><li>Unit test cases and some integration testing provide a scalable way of managing growing codebase.<\/li>\n<li>A suite of extensive and automated tests can reduce overall error rates by validating the quality and by safeguarding against regressions.<\/li>\n<li>Tests also allow engineers to make changes, especially large refactorings, with significantly higher confidence.<\/li>\n<li>Despite its benefits, it can be difficult to foster a culture of automated testing.<\/li>\n<li>Focus on high leverage tests.<\/li>\n<li>Writing more tests, creating a virtuous feedback cycle and saving more development time.<\/li>\n<\/ul><h4><a id=\"user-content-repay-technical-debt\" class=\"anchor\" aria-hidden=\"true\" href=\"#repay-technical-debt\"><\/a>Repay Technical Debt<\/h4>\n<ul><li>Technical debt refers to all the deferred work that\u2019s necessary to improve the health and quality of the codebase and that would slow us down if left unaddressed.<\/li>\n<li>Accumulating technical debt is fine as far as it is repaid within time.<\/li>\n<li>Refactor often.<\/li>\n<\/ul><h3><a id=\"user-content-reduce-operational-complexity\" class=\"anchor\" aria-hidden=\"true\" href=\"#reduce-operational-complexity\"><\/a>Reduce Operational Complexity<\/h3>\n<ul><li>Keep no. of technologies low. Don\u2019t sway towards shiny new technologies.<\/li>\n<li>Every additional technology you add is is guaranteed to go wrong eventually. Will need your time.<\/li>\n<li>Do the simple thing first.<\/li>\n<li>Embrace operational simplicity.<\/li>\n<li>The first solution that comes to mind is generally complex. Don't stop. Keep peeling off the layers of onion.<\/li>\n<li>Simplify the architecture to reduce their operational burden.<\/li>\n<li>\u201cWhat\u2019s the simplest solution that can get the job done while also reducing our future operational burden?\u201d<\/li>\n<li>Discipline to focus on simplicity is high leverage.<\/li>\n<\/ul><h3><a id=\"user-content-fail-fast\" class=\"anchor\" aria-hidden=\"true\" href=\"#fail-fast\"><\/a>Fail Fast<\/h3>\n<ul><li>Fail immediately and visibly.<\/li>\n<li>Doesn\u2019t necessarily mean crashing your programs for users.<\/li>\n<li>fail-fast to surface issues immediately.<\/li>\n<li>Failing fast is high leverage as it saves debugging time.<\/li>\n<\/ul><h3><a id=\"user-content-relentlessly-automate\" class=\"anchor\" aria-hidden=\"true\" href=\"#relentlessly-automate\"><\/a>Relentlessly Automate<\/h3>\n<ul><li>Automating mechanics is good.<\/li>\n<li>Automating decision making - no.<\/li>\n<li>Hone your ability to respond and recover quickly.\n<ul><li>Leverage recovering quickly &gt; Leverage preventing failures.<\/li>\n<\/ul><\/li>\n<li>\u201cscript for success,\u201d practice failure scenarios, and work on our ability to recover quickly.<\/li>\n<li>Make batch process idempotent<\/li>\n<li>Make processes retryable, i.e., not leaving any global state.<\/li>\n<\/ul><h3><a id=\"user-content-invest-in-your-teams-growth\" class=\"anchor\" aria-hidden=\"true\" href=\"#invest-in-your-teams-growth\"><\/a>Invest in your team's Growth<\/h3>\n<ul><li>Invest in onboarding.<\/li>\n<li>The higher you climb up the engineering ladder, the more your effectiveness will be measured not by your individual contributions but by your impact on the people around you.<\/li>\n<li><strong>\"You\u2019re a staff engineer if you\u2019re making a whole team better than it would be otherwise. You\u2019re a principal engineer if you\u2019re making the whole company better than it would be otherwise. And you\u2019re distinguished if you\u2019re improving the industry.\u201d<\/strong>\n\ufffc- Focus primarily on making everyone around you succeed.<\/li>\n<li>Your career depends on your team's success.<\/li>\n<li>Make hiring everyone's responsibility.<\/li>\n<li>Shared ownership of code.\n<ul><li>Keep bus factor more than one.<\/li>\n<li>Shared ownership removes isolated silos of information.<\/li>\n<\/ul><\/li>\n<li>Build collective wisdom through post mortems.<\/li>\n<li>Invest in automated testing.\n<ul><li>Automated test cases lead to higher confidence when refactoring.<\/li>\n<li>Write test cases when the code is fresh in mind.<\/li>\n<li>Don\u2019t be dogmatic about 100% code coverage.<\/li>\n<li>Value of tests increases over time and cost to write goes down.<\/li>\n<\/ul><\/li>\n<li>Hire the best.<\/li>\n<li>Surround yourself with great advisors<\/li>\n<\/ul><p>\u2600\ufe0f<strong>\u201cLeverage is the lens through which effective engineers view their activities. \u201d<\/strong> \u2600\ufe0f<\/p>\n<h3><a id=\"user-content-10-books-to-read\" class=\"anchor\" aria-hidden=\"true\" href=\"#10-books-to-read\"><\/a>10 Books to read:<\/h3>\n<ul><li>Peopleware Productive projects and Teams. Amazon. My Summary.<\/li>\n<li>Team Geek: A Software Developer\u2019s Guide to Working Well with Others. (Debugging Teams) Amazon. My Summary.<\/li>\n<li>High Output Management<\/li>\n<li>Getting Things Done: The Art of Stress-Free Productivity<\/li>\n<li>The 4-Hour Workweek: Escape 9-5, Live Anywhere, and Join the New Rich<\/li>\n<li>The 7 Habits of Highly Effective People: Powerful Lessons in Personal Change<\/li>\n<li>Conscious Business: How to Build Value Through Values<\/li>\n<li>Your Brain at Work<\/li>\n<li>Flow: The Psychology of Optimal Experience<\/li>\n<li>Succeed: How We Can Reach Our Goals<\/li>\n<\/ul><h3><a id=\"user-content-blogs\" class=\"anchor\" aria-hidden=\"true\" href=\"#blogs\"><\/a>Blogs:<\/h3>\n<p>Recommended Blogs To Follow:<\/p>\n<ul><li><a href=\"http:\/\/www.theeffectiveengineer.com\/\" rel=\"nofollow\">http:\/\/www.theeffectiveengineer.com\/<\/a> - The Effective Engineer is my personal blog, where I write about engineering habits, productivity tips, leadership, and culture.<\/li>\n<li><a href=\"http:\/\/www.kalzumeus.com\/\" rel=\"nofollow\">http:\/\/www.kalzumeus.com\/<\/a> - Patrick McKenzie runs his own software business and has written many excellent long-form articles on career advice, consulting, SEO, and software sales.<\/li>\n<li><a href=\"http:\/\/katemats.com\/\" rel=\"nofollow\">http:\/\/katemats.com\/<\/a> - Kate Matsudaira, who has worked at large companies like Microsoft and Amazon as well as at startups, shares advice about tech, leadership, and life on her blog.<\/li>\n<li><a href=\"http:\/\/randsinrepose.com\/\" rel=\"nofollow\">http:\/\/randsinrepose.com\/<\/a> - Michael Lopp has worked for many years in leadership positions at Netscape, Apple, Palantir, and Pinterest, and writes about tech life and engineering management.<\/li>\n<li><a href=\"http:\/\/softwareleadweekly.com\/\" rel=\"nofollow\">http:\/\/softwareleadweekly.com\/<\/a> - Oren Ellenbogen curates a high-quality weekly newsletter on engineering leadership and culture.<\/li>\n<li><a href=\"http:\/\/calnewport.com\/\" rel=\"nofollow\">http:\/\/calnewport.com\/<\/a> - Cal Newport, an assistant professor of computer science at Georgetown, focuses on evidence-based advice for building a successful and fulfilling life.<\/li>\n<li><a href=\"http:\/\/www.joelonsoftware.com\/\" rel=\"nofollow\">http:\/\/www.joelonsoftware.com\/<\/a> - Joel Spolsky, the co-founder of Stack Exchange, provides all sorts of programming pearls of wisdom on his blog.<\/li>\n<li><a href=\"http:\/\/martinfowler.com\/\" rel=\"nofollow\">http:\/\/martinfowler.com\/<\/a> - Martin Fowler, author of the book Refactoring, writes about how to maximize the productivity of software teams and provides detailed write-ups of common programming patterns.<\/li>\n<li><a href=\"http:\/\/pgbovine.net\/\" rel=\"nofollow\">http:\/\/pgbovine.net\/<\/a> - Philip Guo, a computer science professor, has written extensively and openly about his graduate school and work experiences.<\/li>\n<\/ul>","created_at":"2018-11-07T16:46:39+0000","updated_at":"2018-11-07T16:46:39+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"gist.github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/6609?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":385,"title":"PurelyFunctional.tv Newsletter 200: Patterns, Bell Labs, Aliens - PurelyFunctional.tv","url":"https:\/\/purelyfunctional.tv\/issues\/purelyfunctional-tv-newsletter-200-patterns-bell-labs-aliens\/","content":"<table border=\"0\" cellpadding=\"0\" class=\"force-row\" style=\"margin: auto; border-spacing: 0px; width: 600px;\"><tbody><tr><td align=\"center\" class=\"col\" valign=\"top\">Issue 200 \u2013 November 21, 2016<\/td><\/tr><\/tbody><\/table><p><em>This email newsletter used to be called the <em>Clojure Gazette<\/em>. Thanks for reading and for your patience during this quick transition period.<\/em><\/p><p>Hi Functionalistas,<\/p><p>Can you believe that this is the 200th issue of this newsletter? I can\u2019t. I started it back in 2012, focusing on the Clojure community, thinking that it was a small group of optimistic Lispers. Since then, Functional Programming has become cool, Clojure and six other functional languages are on people\u2019s TO LEARN lists, and other languages and technologies are referencing Clojure as inspiration. It has been quite a ride and there appears to be more coming.<\/p><p>Back when I started, I had had to struggle with my interest in FP. I would bring it up at work and people would balk at Lisp, though they knew nothing of it. The insults and jeers were like recess in grade school. Now that there\u2019s enough momentum, even the most skeptical are curious.<\/p><p>The work building the future is not done yet. Now is the time to remember what brought us here. Many people did a lot of work in the past that led us to where we are. We don\u2019t have the good habit of knowing the history of our field. We grab at every new framework, tool, or language, looking for the answer to the software crisis. But the software crisis is as old as software. It won\u2019t be solved so easily.<\/p><p>What we need to do is to read more. To learn what people have already done. It\u2019s not that old ideas are better than new ones. It\u2019s that we need to learn history so that we can learn <em>from<\/em> it and avoid repeating the mistakes. Unfortunately, the internet fosters a \u201cpop culture\u201d. The cure is to read printed material\u2013books and papers.<\/p><p>Unfortunately, this newsletter is part of that pop culture. But I\u2019m now dedicating two links per issue to papers and two to books. This is going to be a challenge because, well, it\u2019s harder to read a book than a blog post, but the challenge will be good for me. Life is too short to not read books. Please share with me the influential books and papers that you think are worth reading.<\/p><p>Please share this issue with your friends and enemies. Happy 200th issue, everybody!<\/p><p>Rock on!<br \/><a href=\"http:\/\/twitter.com\/ericnormand\">Eric Normand<\/a> &lt;<a href=\"mailto:eric@lispcast.com\">eric@lispcast.com<\/a>&gt;<\/p><p>PS Want to get this in your email? <a href=\"http:\/\/www.clojuregazette.com\/\">Subscribe<\/a>!<\/p><hr \/><p>I\u2019ve been really enjoying this classic book of essays by Lisp legend Richard Gabriel.<\/p><hr \/><p>Alan Perlis was the first recipient of the <em>Turing Award<\/em>. His lecture was oddly prescient about the current state of JavaScript with current versions compiling down to older versions. Also, he talks a lot about the importance of homoiconicity and the strange stability of the Algol syntax.<\/p><hr \/><p>The Euroclojure videos have started going up and this one by David Nolen is great. It chronicles the rocketship trajectory of ClojureScript and discusses how we can get more involved. If you\u2019re experiencing <em>JavaScript Fatigue<\/em>, watch this talk.<\/p><hr \/><p>Marvin Minsky postulates two principles that all intelligences must have evolved to deal with, and hence we would share these experiences. They would give us a basis for communication. I\u2019ve been reading up on these things since Alan Kay has been promoting the analagous idea of software from different systems, without any human coordination, being able to communicate.<\/p><hr \/><p>The <a href=\"http:\/\/pftv.wpengine.com\/clojure-in-one-hour\/\" title=\"\"><em>Clojure in One Hour<\/em><\/a> series is a set of small projects you can do in less than an hour. This week I recorded myself creating a run tracker from scratch. The only thing I didn\u2019t get to was the backend for saving the runs. But it does track my speed, pace, total time, and total distance.<\/p><hr \/><p>Mike Pearson does an excellent job talking about the design decisions that went into communicating a difficult subject.<\/p><hr \/><p>Srihari Sriraman has developed a model of Carnatic music, which comes from southern India. He uses Overtone.<\/p><hr \/><p>This book by Jon Gertner chronicles the many inventions, researchers, and leaders who made Bell Labs so productive as an research lab. I\u2019m halfway through this book and I love it.<\/p><hr \/><p>I\u2019ve been <a href=\"http:\/\/pftv.wpengine.com\/speaker-interview\/\" title=\"\">interviewing the speakers<\/a> of the upcoming Clojure\/conj to help attendees prepare for the mindbending talks. This one is by Jon Smock who will be talking about SAT solvers.<\/p>","created_at":"2018-11-07T16:46:38+0000","updated_at":"2018-11-07T16:46:38+0000","published_at":"2016-11-21T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":3,"domain_name":"purelyfunctional.tv","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":383,"title":"EmacsWiki: Gnus MSMTP","url":"https:\/\/www.emacswiki.org\/emacs\/GnusMSMTP","content":"<p>Don\u2019t confuse the program <code>msmtp<\/code> with the ability to use <a class=\"local\" href=\"https:\/\/www.emacswiki.org\/emacs\/MultipleSMTPAccounts\">MultipleSMTPAccounts<\/a> in Gnus.<\/p><h2>Basic Setup<\/h2><p>Here\u2019s a quick and short description of setting Gnus and <code>msmtp<\/code> to work with Gmail. This should work with the following software on Windows and *nix platform. The examples here are shown for the Microsoft Windows versions of the software but should work with just the path changes on *nix platforms.<\/p><pre>No Gnus v0.6 or higher\nGNU Emacs 21.3 or higher\nmsmtp version 1.4.9 or higher\n<\/pre><p>The details above can be got by  <\/p><pre>1.\tM-x gnus-version in Emacs \n2.\tM-x emacs-version in emacs \n3.\tmsmtp --version at the command prompt \n<\/pre><p>Your typical <code>msmtprc.txt<\/code> file would look like this. Note <code>msmtp --version<\/code> will also tell you where <code>msmtp<\/code> looks for the configuration file.<\/p><pre>########################\naccount anderson\nhost smtp.gmail.com\ntls on\nauth on\nuser sivaram.notthere@gmail.com\npassword XXXXXXX\nport 587\nlogfile c:\\msmtplog.txt\n########################\naccount neo\nhost smtp.gmail.com\ntls on\nauth on\nuser sivaram.anotherguy@gmail.com\npassword XXXXXXX\nport 587\nlogfile c:\\msmtplog.txt\n########################\n<\/pre><p>Attention: As in version 1.4.31 of msmtp, there is no need to fiddle with <code>\u2018message-sendmail-extra-arguments\u2019<\/code>. Inside your msmtprc file, add the <code>\u2018from\u2019<\/code> command to each account. msmtp will match the <code>\u2018from\u2019<\/code> value against the `From: \u2018 field of the e-mail header (only the address).<\/p><p>I have found an alternate way to allow msmtp to determine from the email message which SMTP account to use. The following settings in my .gnus.el works for msmtp version 1.4.26: <\/p><pre>;; This is needed to allow msmtp to do its magic:\n(setq message-sendmail-f-is-evil 't)\n;;need to tell msmtp which account we're using\n(setq message-sendmail-extra-arguments '(\"--read-envelope-from\"))\n<\/pre><p>A close reading of the msmtp documentation and the definition of message-send-mail-with-sendmail in message.el will reveal how this works. --taltman<\/p><p>Now, we\u2019d like to use Gnus to send email through <code>msmtp<\/code>. Add the following lines to the <code>.gnus.el<\/code> file.<\/p><pre>;; with Emacs 23.1, you have to set this explicitly (in MS Windows)\n;; otherwise it tries to send through OS associated mail client\n(setq message-send-mail-function 'message-send-mail-with-sendmail)\n;; we substitute sendmail with msmtp\n(setq sendmail-program \"c:\/gnu\/bin\/msmtp.exe\")\n;;need to tell msmtp which account we're using\n(setq message-sendmail-extra-arguments '(\"-a\" \"anderson\"))\n;; you might want to set the following too\n(setq mail-host-address \"gmail.com\")\n(setq user-full-name \"Anderson Guy\")\n(setq user-mail-address \"sivaram.notthere@gmail.com\")\n<\/pre><p>Now compose a mail in Gnus to see whether the headers come up correctly and then send the mail message by hitting <code>C-c C-c<\/code> After a brief wait, the mail should be sent to the address specified. If there are any errors, <code>msmtp<\/code> error codes are displayed in the *echo* area as well as the <code>msmtplog.txt<\/code> file.<\/p><p>Or you can see what\u2019s wrong here by looking up the error code in the man page<\/p><p><a class=\"url http\" href=\"http:\/\/freebsd-man.page2go2.com\/man3\/sysexits_3.html\">http:\/\/freebsd-man.page2go2.com\/man3\/sysexits_3.html<\/a><\/p><p>In the example above, we have used only 1 of the email accounts. Gnus allows you to post using different accounts, which can be accomplished with <a class=\"local\" href=\"https:\/\/www.emacswiki.org\/emacs\/PostingStyles\">PostingStyles<\/a>.<\/p><p>To use <a class=\"local\" href=\"https:\/\/www.emacswiki.org\/emacs\/PostingStyles\">PostingStyles<\/a> in Gnus with <code>msmtp<\/code>, here\u2019s an example of how it\u2019s done<\/p><pre>(setq gnus-parameters\n  ;;Use notthere id for all gmane news group postings\n  '((\".*news\\\\.gmane\\\\.org.*\"\n     (posting-style\n      (address \"sivaram.notthere@gmail.com\")\n      (name \"Sivaram N\")\n      (body \"\\n\\n\\n sivaram\\n -- \")\n      (eval (setq message-sendmail-extra-arguments '(\"-a\" \"anderson\")))\n      (user-mail-address \"sivaram.notthere@gmail.com\")))\n      ;;use anotherguy id for all normal mails\n    (\".*mail\\\\.misc\"\n     (posting-style\n      (address \"sivaram.anotherguy@gmail.com\")\n      (name \"Sivaram Another\")\n      (body \"\\n\\n\\n Sivaram A\\n -- \\n\")\n      (eval (setq message-sendmail-extra-arguments '(\"-a\" \"neo\")))\n      (user-mail-address \"sivaram.anotherguy@gmail.com\")))))\n<\/pre><p>The variable <code>message-sendmail-extra-arguments<\/code> decides which <code>msmtp<\/code> account is called depending in which folder one is currently in Gnus.<\/p><p>Thanks to ssSslang on <b>gnu.emacs.gnus<\/b> who requested for <code>message-sendmail-extra-arguments<\/code> to handle multiple accounts of <code> msmtp <\/code><\/p><p>\u2013 Updated: 4\/18\/2008 (blbrown): To the original post, I couldn\u2019t get msmtp and gnus to work properly. It kept returning an exit value 78 error. I went to the msmtp documentation and made sure I could send a message directly from the command line with msmtp. I would make sure that works first. Also, here is my ~\/<em>DOT<\/em>msmtprc file. <\/p><pre>##################\ndefaults\ntls on\ntls_starttls on\naccount default\nhost smtp.gmail.com\ndomain smtp.gmail.com\ntls on\ntls_starttls on\nauth on\nuser USER@gmail.com\nfrom USER@gmail.com\npassword PASSWORD\nport 587\nlogfile \/home\/USER\/msmtplog.txt\n##################\n<\/pre><p>msmtp complains about the \u2018default\u2019 account settings, so if you receive a generic error while running with emacs; go to the command-line and if see if you receive more verbose error messages. Also, recheck your rc file.<\/p><pre>Above configuration tested with:\nOS: Ubuntu Linux 7.10\nEmacs: 22.1.1 x86\nmsmtp version 1.4.14\nTLS\/SSL library: OpenSSL\n<\/pre><p><b>Note that tls_cert_check = off makes you susceptible to man-in-the-middle attacks<\/b>. The correct thing to do is to download the appropriate SSL certificates and configure them as describe in the msmtp man page.<\/p><h2>Yet Another Gnus\/MSMTP Setting for Easy Multi-SMTP Sending<\/h2><p>Multi-SMTP email sending in Gnus tortured me for long time. I searched for a sound solution for long time and can not find one. And I think using Posting Styles is too tedious, boring and prone to shooting myself. ILast Saturday night, I spent several hours, and found a sound and safe solution, and its setting is quite simple.<\/p><pre>;; Choose account label to feed msmtp -a option based on From header in Message buffer;\n;; This function must be added to message-send-mail-hook for on-the-fly change of From address\n;; before sending message since message-send-mail-hook is processed right before sending message.\n(defun cg-feed-msmtp ()\n  (if (message-mail-p)\n          (save-excursion\n                (let* ((from\n                               (save-restriction\n                                         (message-narrow-to-headers)\n                                         (message-fetch-field \"from\")))\n                           (account\n                                   (cond\n                                            ;; I use email address as account label in ~\/.msmtprc\n                                        ((string-match \"myemail1@server1.com\" from)\"myemail1@server1.com\")\n                                ;; Add more string-match lines for your email accounts\n                                ((string-match \"myemail2@server2.com\" from)\"myemail2@server2.com\"))))\n                  (setq message-sendmail-extra-arguments (list '\"-a\" account)))))) ; the original form of this script did not have the ' before \"a\" which causes a very difficult to track bug --frozencemetery\n(setq message-sendmail-envelope-from 'header)\n(add-hook 'message-send-mail-hook 'cg-feed-msmtp)\n<\/pre><p>Though I use email address as account label, it\u2019s not necessary. You can use any style you like. Just keep labels the same as in .msmtprc and code above.<\/p><p>Anyone who can read Chinese, and if you are interested, you can visit original article at my site  - <a class=\"url http outside\" href=\"http:\/\/cyberhut.org\/2007\/10\/gnusmsmtp.html\">\/ Gnus+MSMTP<\/a> or the same article in Chinese Emacs wiki - <a class=\"url http outside\" href=\"http:\/\/www.emacs.cn\/Gnus\/Msmtp\">\/ Gnus+MSMTP<\/a>.<\/p><div class=\"color one level0\"><p> Q: Is there any way to protect the password? I don\u2019t like the idea of leaving it in plain text in the file, and for some reason I do not get prompted in gnus if I leave it out (although it works fine from the command line).<\/p><\/div><div class=\"color two level0\"><p> A1: MSMTP can use gnome keyring to store your password. If you specify auth but no pass, it will interrogate gnome-keyring-daemon for a password. This mitigates the risk of plain text password storage but is difficult to configure in a non-desktop environment.<\/p><p>A2: MSMTP takes the argument <b>--passwordeval=&lt;cmd&gt;<\/b>. In order to retrieve the password for the specified account, it will then run <b>&lt;cmd&gt;<\/b>. Since <b>&lt;cmd&gt;<\/b> can be anything, in particular <b>gpg<\/b>, you need not store your password unencrypted, just add <b>\"--passwordeval=gpg --quiet --batch -d name-of-file\"<\/b> to <b>message-sendmail-extra-arguments<\/b>, where <b>name-of-file<\/b> should be a gpg-encrypted text file containing nothing but the password. Note that if you miss <b>--batch<\/b> or <b>--quiet<\/b>, you will get errors which are hard to trace; the two disable unnecessary output to stdout and stderr, both of which emacs would capture.<\/p><\/div><h2>Using msmtpQ to queue your mail<\/h2><p>msmtp comes with a simple script which adds the ability to queue mail. I use it in gnus with these settings: <\/p><pre>(setq message-send-mail-function 'message-send-mail-with-sendmail)\n(setq sendmail-program \"\/usr\/local\/bin\/msmtpQ\"\n      mail-specify-envelope-from t\n;; needed for debians message.el cf. README.Debian.gz \n      message-sendmail-f-is-evil nil                \n      mail-envelope-from 'header\n      message-sendmail-envelope-from 'header)\n<\/pre><p>msmtpQ sends the mail if there is a connection or queues it when there is none. It is part of the debian package: (\/usr\/share\/doc\/msmtp\/examples\/msmtpq\/), but afaics is also in included in upstream. I put it into \u201c\/usr\/local\/bin\/\u201d. Don\u2019t forget to set 3 vars in the script and make it +x. To send your mails, you flush your queue with \u201cmsmtpq -r\u201d (Yes, small q).<\/p><hr \/><p><a class=\"local\" href=\"https:\/\/www.emacswiki.org\/emacs\/CategoryGnus\">CategoryGnus<\/a><\/p>","created_at":"2018-11-07T16:46:37+0000","updated_at":"2018-11-07T16:46:37+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"www.emacswiki.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":384,"title":"A Comprehensive Beginner's Guide to Aperture, Shutter Speed, and ISO","url":"https:\/\/petapixel.com\/2016\/06\/25\/comprehensive-beginners-guide-aperture-shutter-speed-iso\/","content":"<p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/pyramid-800x420.jpg\" alt=\"pyramid\" width=\"800\" height=\"420\" class=\"aligncenter size-large wp-image-224887\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/pyramid-800x420.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/pyramid-320x168.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/pyramid.jpg 1200w\" \/><\/p><p>This guide to photographic exposure aims to help you take full control of your camera. I often tell my students that I want them to move away from the idea \u201ctaking a photograph\u201d and towards the idea of \u201cmaking a photograph.\u201d I teach them how to take the camera off auto mode and take full control of the settings themselves in order to create the photograph they want. <\/p><p>Why let the camera decide these things for you? Do you let your mother choose your clothes? Maybe some of you do, I don\u2019t know. Frankly, I don\u2019t want to know.<\/p><p>I hope to do the same for the readers of this tutorial. I want you take control of your camera. In order to do this, it\u2019s essential to understand the 3 components of what we call \u201cThe Exposure Triangle\u201d. These are: aperture, shutter speed and ISO. By the end of this tutorial, you should understand what these 3 components are and how they affect the final photograph. You will also learn how to use the 3 main shooting modes on your camera: aperture priority, shutter priority and manual. Finally, I\u2019ll explain how to decide which settings to choose as you prepare to shoot a scene.<\/p><h3>What is exposure?<\/h3><p>First of all we need to define what we mean by exposure. Exposure refers to the amount of light that enters the camera and hits the digital sensor. Basically, it is a measure of how dark or bright a photograph is.<\/p><p>If the image is too bright, it is overexposed. Too much light has been allowed to hit the sensor. If it is too dark, it is underexposed. Not enough light has been allowed to hit the sensor. We can control how much light reaches the sensor by changing the aperture, shutter speed and ISO settings.<\/p><p>Exposure is measured in \u2018stops\u2019. For example, if you find that your photo has turned out too dark (underexposed), you may increase your exposure by a \u2018stop\u2019 or two to make it brighter. Conversely, if the image is overexposed, you may need to decrease the exposure by a stop or two. There is no such thing as the \u2018perfect\u2019 exposure, only the right exposure for the photograph you are creating. Some photos such as night shots are supposed to be dark while photos taken in the snow for example are supposed to be bright.<\/p><h3>Measuring exposure using a histogram<\/h3><p>All digital cameras allow you to see a visual representation of exposure using the <a href=\"http:\/\/digital-photography-school.com\/how-to-read-and-use-histograms\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">histogram<\/a>. Check your camera\u2019s manual to find out how to turn on the histogram feature. There was a member of my photography club who would tell all new members to RTFM. This stood for \u2018Read the Manual\u2019. I\u2019ll let you figure out what the \u2018f\u2019 stood for yourself.<\/p><p>The histogram is a graph that represents the spread of tones in a photograph, from the shadows, to the mid tones to the highlights. It allows you to check if the photograph has any shadows that are too dark or \u2018clipped\u2019 and to see if you have any highlights that are too bright or \u2018blown out\u2019.<\/p><p>Clipped shadows are areas of pure black and contain no detail. Blown out highlights are areas of pure white and also contain no detail. Very generally speaking, you will want to avoid both of these. That said, I personally don\u2019t mind a little clipping in the shadows as it adds punch to the image.<\/p><p>If you look at the histogram below, you will see that some of the graph is right up against the left hand axis of the graph. This means that some of the shadows are clipped. If you look at the right, you will see that a very tiny amount of highlights have been blown out as a very small part of the graph is up against the right hand edge. Sometimes this is unavoidable for example with street lights or if the sun in the frame. Remember, that the histogram is only a guide.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/histogram-800x356.jpg\" alt=\"histogram\" width=\"800\" height=\"356\" class=\"aligncenter size-large wp-image-224855\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/histogram-800x356.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/histogram-320x143.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/histogram.jpg 1062w\" \/><\/p><h3>Examples of underexposed and overexposed photographs<\/h3><p>Below we have examples of an underexposed photo, an overexposed photo and a correctly exposed photo.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/under-800x380.jpg\" alt=\"under\" width=\"800\" height=\"380\" class=\"aligncenter size-large wp-image-224856\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/under-800x380.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/under-320x152.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/under.jpg 900w\" \/><\/p><p><b>Underexposed photograph<\/b>: This photograph is underexposed by about 3 stops. You can see that the histogram is completely bunched up to the left as a result. There are lots of clipped shadows on the underside of the gondolas.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/over-800x380.jpg\" alt=\"over\" width=\"800\" height=\"380\" class=\"aligncenter size-large wp-image-224857\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/over-800x380.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/over-320x152.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/over.jpg 900w\" \/><\/p><p><b>Overexposed photograph<\/b>: This image is overexposed by about 3 stops and as you can see, the histogram is bunched up to the right as a result. There are a lot of clipped highlights is this photo. In fact, the entire sky is pure white and contains no detail whatsoever.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/correct-800x380.jpg\" alt=\"correct\" width=\"800\" height=\"380\" class=\"aligncenter size-large wp-image-224858\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/correct-800x380.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/correct-320x152.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/correct.jpg 900w\" \/><\/p><p><b>Correctly exposed photograph<\/b>: The photo above has the right exposure for the scene in question. You can see on the histogram that there is a good spread of shadows, mid-tones and highlights. It\u2019s quite a bright image as you can see from the fact that the graph spikes on the right of the graph. <\/p><p>There is a little clipping in the shadows which I don\u2019t mind as it adds some punch to the shot. As you can see from the right hand side of the graph, there are some very bright areas but the highlights are not blown out.<\/p><h3>Using the highlights warning feature on your camera<\/h3><p>It\u2019s always a good idea to check the histogram after you\u2019ve taken a shot in order to prevent too many clipped shadows and blown out highlights. Most digital cameras also have a \u2018highlight warning\u2019 feature. <\/p><p>This makes areas of the image that have blown highlights flash on your screen. It\u2019s an incredibly useful feature and I keep it turned on all the time. Below, you can see how the highlight warning looks on the overexposed gondola photo. A huge amount of the photo is flashing because so many of the highlights have been blown out.<\/p><h3>What is aperture?<\/h3><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/lens-aperture.jpg\" alt=\"lens-aperture\" width=\"300\" height=\"243\" class=\"alignright size-full wp-image-224859\" \/><\/p><p>The aperture refers to the size of the opening in the lens through which the light enters the camera. The size of this opening can be adjusted and the aperture size is measured in f-stops. The image on the right shows you exactly what the aperture on a lens looks like.<\/p><p>When you change the f-stop value, you change the size of the opening. Here\u2019s the weird thing though. The higher the f-stop, the smaller the opening. <\/p><p>Take a look at the chart below to see what different apertures look like at different f-stops. On the far left, you can see that setting an aperture of f16 will result in a small opening. Choosing an aperture of f1.4 will result in a very wide opening.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/aperture-chart-800x145.jpg\" alt=\"aperture-chart\" width=\"800\" height=\"145\" class=\"alignright size-large wp-image-224861\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/aperture-chart-800x145.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/aperture-chart-320x58.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/aperture-chart.jpg 1024w\" \/><\/p><h3>How does your choice of aperture affect the photograph?<\/h3><p>The most noticeable effect your choice of aperture has on the photograph is the depth of field. What do we mean by this exactly? In very simple terms, depth of field refers to the amount of the image that is sharp. What does this mean in practice?<\/p><p>If you use a wide aperture, the depth of field will be shallow. Only part of the image is sharp and the rest will be out of focus or blurred. Look at the picture on the left below. The cat is perfectly sharp but the background is blurred. Using a wide aperture works well for portrait style photographs as it makes the subject of the shot really stand out against the blurred background.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram-800x534.jpg\" alt=\"depth-of-field-diagram\" width=\"800\" height=\"534\" class=\"alignnone size-large wp-image-224866\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram-800x534.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram-320x214.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram-450x300.jpg 450w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram-300x200.jpg 300w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram-120x80.jpg 120w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/depth-of-field-diagram.jpg 1092w\" \/><\/p><p>In this case, the depth of field extends from about the tip of the cat\u2019s nose to just behind its head, no more than a few centimeters (from point A to point B in the diagram). Anything not in this range, either in front of it or behind will not be sharp. For this shot, I used a wide aperture of f\/3.5.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat-800x533.jpg\" alt=\"OLYMPUS DIGITAL CAMERA\" width=\"800\" height=\"533\" class=\"alignnone size-large wp-image-224867\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat-800x533.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat-320x213.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat-450x300.jpg 450w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat-300x200.jpg 300w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat-120x80.jpg 120w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/curious-cat.jpg 900w\" \/><\/p><p>When you use a narrow aperture, the depth of field is deep. When the depth of field is deep, all of the photograph from foreground to background is sharp. <\/p><p>Take a look at the photo below taken in the <a href=\"http:\/\/www.dublindocklands.ie\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">Dublin Docklands<\/a>. Everything from the dock cleat in the foreground to the bridge in the background is sharp. In this case the depth of field is several hundred metres, extending right from the foreground to the background of the scene. In this case, I used a narrower aperture of f\/11.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin-800x533.jpg\" alt=\"samuel-beckett-bridge-at-sunset-dublin\" width=\"800\" height=\"533\" class=\"alignnone size-large wp-image-224869\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin-800x533.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin-320x213.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin-450x300.jpg 450w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin-300x200.jpg 300w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin-120x80.jpg 120w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/samuel-beckett-bridge-at-sunset-dublin.jpg 900w\" \/><\/p><p>Most of the time, we want to achieve a deep depth of field when shooting landscapes. We want all of the image to be pin sharp.<\/p><p>The mid range apertures (around f\/8) are good for shooting handheld for example when doing <a href=\"http:\/\/erickimphotography.com\/blog\/the-ultimate-beginners-guide-for-street-photography\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">street photography<\/a>. You get a good balance between having enough depth of field and fast enough shutter speeds to shoot hand held. We\u2019ll discuss shutter speeds in more detail later.<\/p><p>The chart below gives you a good idea how different apertures will affect the depth of field in your photographs. You can see that as the aperture gets wider, the pyramid in the background becomes more blurred.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/aperture-effect-chart-2-710x700.jpg\" alt=\"aperture-effect-chart-2\" width=\"710\" height=\"700\" class=\"aligncenter size-large wp-image-224870\" \/><\/p><h3>What is shutter speed?<\/h3><p>The shutter speed refers to the length of time the opening in the lens remains open to let light into the camera and onto the sensor. The shutter speed can be as fast as 1\/10,000 of a second or as slow as several minutes.<\/p><h3>How does your choice of shutter speed affect the photograph?<\/h3><p>Fast shutter speeds have the effect of freezing motion in the scene you are photographing. Conversely, slow shutter speeds will blur motion in a scene. Both of these can be used to great creative effect.<\/p><p>The shutter speed settings on your camera provide a great way to experiment with capturing motion in your landscape photography. This is especially the case with moving water.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1-800x533.jpg\" alt=\"OLYMPUS DIGITAL CAMERA\" width=\"800\" height=\"533\" class=\"alignnone size-large wp-image-224871\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1-800x533.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1-320x213.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1-450x300.jpg 450w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1-300x200.jpg 300w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1-120x80.jpg 120w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/autumn-waterfall-ireland-1.jpg 1200w\" \/><\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia-800x533.jpg\" alt=\"fishing-boat-at-sunset-tunisia\" width=\"800\" height=\"533\" class=\"alignnone size-large wp-image-224872\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia-800x533.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia-320x213.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia-450x300.jpg 450w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia-300x200.jpg 300w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia-120x80.jpg 120w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/fishing-boat-at-sunset-tunisia.jpg 1200w\" \/><\/p><p>By using a slow shutter speed (1\/2 second), we can blur the water in a waterfall for example and create a sense of motion even though it\u2019s a still image. You can see this in this photo of a waterfall in Ireland above. When working with slow shutter speeds, it is essential to use a tripod otherwise camera shake will result in a completely blurred photo.<\/p><p>In the second photograph taken in Tunisia, I used an extremely long shutter speed of 160 seconds. To achieve this, I used a <a href=\"http:\/\/www.martinlawrencephotography.com\/articles-tips-locations\/articles\/how-to-use-a-10-stop-neutral-density-filter\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">10 stop neutral density filter<\/a>. This reduced the light entering the camera down to 1\/1000th of what it would be without the filter. This, in turn, allowed me to set such a long exposure time. <\/p><p>As you can see, the clouds moved across the sky during the almost 3 minutes it took to take the photo resulting in the blurred effect.<\/p><p>You can also use fast shutter speeds to freeze motion like in this black and white seascape below.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/atlantic-coast-donegal-ireland-800x405.jpg\" alt=\"atlantic-coast-donegal-ireland\" width=\"800\" height=\"405\" class=\"alignnone size-large wp-image-224873\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/atlantic-coast-donegal-ireland-800x405.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/atlantic-coast-donegal-ireland-320x162.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/atlantic-coast-donegal-ireland.jpg 1200w\" \/><\/p><p>For this photograph, I wanted to freeze the motion of the waves crashing against the shore. A fast shutter speed of 1\/320th of a second ensured that the wave seems to \u2018freeze\u2019 in time. Landscapes that include moving water afford great opportunities to experiment with different exposure times.<\/p><p>The chart below shows how different shutter speeds would effect the sense of motion if you were photographing a person running. Fast shutter speeds will freeze the motion. This technique is often used in sports photography. The slower the shutter speed becomes, the more blurred the person running becomes in the photograph.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/shutter-speed-effect-chart-790x700.jpg\" alt=\"shutter-speed-effect-chart\" width=\"790\" height=\"700\" class=\"alignnone size-large wp-image-224874\" \/><\/p><h3>How do you know if your shutter speed is fast enough to shoot handheld?<\/h3><p>There is a very simple trick to check if your shutter speed is fast enough to shoot hand held. Simply look at the focal length you have zoomed in to on the lens. <\/p><p>On the lens below, the focal length is set at about 30mm. In this case I simply multiply the focal length by 2 and divide it into 1 to get the minimum shutter speed required to shoot hand held. So, 30 x 2 is 60 therefore the minimum shutter speed required to shoot hand held is 1\/60 of a second.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/camera-lens-focal-length-556x700.jpg\" alt=\"camera-lens-focal-length\" width=\"556\" height=\"700\" class=\"aligncenter size-large wp-image-224875\" \/><\/p><p>This means that you can get away with using slower shutter speeds when the angle is wider. It\u2019s obviously harder to keep the camera steady when zooming in. Think of how difficult it is to keep your sights on an object when using binoculars. It\u2019s the same principle.<\/p><p>If you find that the light is low and you can\u2019t get a fast enough shutter speed, you can increase the ISO. In the next section, I\u2019ll explain what ISO is and how it effects the photograph.<\/p><h3>What is ISO?<\/h3><p>The ISO refers to how sensitive the digital sensor in your camera is to light. The lower the ISO number, the less sensitive it is to light. Setting a higher ISO number increases the sensitivity of your camera sensor to light. Most cameras have ISOs ranging from about 50 or 100 ISO right up to 16,000 ISO or higher.<\/p><h3>How does your choice of ISO affect the photograph?<\/h3><p>As you increase the ISO value, your camera sensor becomes more sensitive to light. This means that you can achieve higher shutter speeds. This can be extremely useful when shooting in low light without a tripod. You may find that shooting at 100 ISO results in shutter speeds that are too slow to hand hold without camera shake. By increasing the ISO to 800 ISO for example, you may find that your shutter speed is now fast enough to hand hold.<\/p><p>You may be wondering: why not just use a really high ISO every time to ensure a sharp photo? The problem is that there is a trade off when it comes to image quality. The higher the ISO used, the more digital noise will be present in the image. Digital noise results in a graininess that can have a negative effect on image quality. Take a look at the labels of this bottle of wine shot at different ISOs.<\/p><p>The first one was shot at 100 ISO.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/iso100-700x700.jpg\" alt=\"iso100\" width=\"700\" height=\"700\" class=\"aligncenter size-large wp-image-224876\" \/><\/p><p>The second photo was shot at a very high ISO 3200. You can see that the graininess has degraded the image quality quite a lot.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/iso3200-700x700.jpg\" alt=\"iso3200\" width=\"700\" height=\"700\" class=\"aligncenter size-large wp-image-224877\" \/><\/p><p>When I finished taking these shots of the bottle of wine, I of course sampled the contents. I\u2026 eh\u2026 wanted to learn about French culture. Funnily enough, after I finished the bottle, the image quality from my own eyes degraded somewhat.<\/p><p>The chart below illustrates the effect of ISO on image quality.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/iso-effect-chart-1-800x169.jpg\" alt=\"iso-effect-chart-1\" width=\"800\" height=\"169\" class=\"alignnone size-large wp-image-224878\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/iso-effect-chart-1-800x169.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/iso-effect-chart-1-320x68.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/iso-effect-chart-1.jpg 1327w\" \/><\/p><p>This does not mean that you should not increase your ISO when the need arises. The example of ISO 3200 above is quite extreme. <a href=\"https:\/\/www.digitalphotomentor.com\/how-to-overcome-the-fear-of-high-iso-and-take-sharper-photos\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">Most of the newer cameras actually handle higher ISOs very well<\/a> and retain high image quality. I know that in low light conditions, I prefer to increase my ISO a little to avoid camera shake even it it means a little graininess. It\u2019s usually not enough to seriously degrade the image quality though.<\/p><p>Next, we\u2019re going to take a look at how to actually set the aperture and shutter speed in your camera. There are 3 modes you can use: aperture priority, shutter priority and manual.<\/p><h3>How to use Aperture Priority Mode on your camera<\/h3><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/aperture-priority.jpg\" alt=\"aperture-priority\" width=\"260\" height=\"180\" class=\"alignright size-full wp-image-224879\" \/><\/p><p>Aperture priority mode is a semi-manual mode. When using this mode, you choose the aperture you want and the camera chooses an appropriate shutter speed in order to achieve a correctly exposed photo. To switch your camera to aperture priority, turn the dial on top of your camera to \u2018A\u2019.<\/p><p>This is actually the shooting mode I use 90% of the time when shooting urban landscapes. I usually choose an aperture of around f16 to ensure maximum depth of field and then let the camera choose the correct shutter speed. As I usually use a tripod, I am generally not too concerned about the shutter speed being too slow.<\/p><p>If I am shooting hand held, I always keep an eye on the shutter speed the camera has chosen just to make sure it isn\u2019t too slow. If it is too slow, I use a wider aperture which will give a faster shutter speed as the opening is larger and lets the light in faster. I also have the option of increasing the ISO to get a faster shutter speed.<\/p><h3>How to use Shutter Priority Mode on your camera<\/h3><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/shutter-priority.jpg\" alt=\"shutter-priority\" width=\"260\" height=\"180\" class=\"alignright size-full wp-image-224880\" \/><\/p><p>Shutter priority is basically the opposite to aperture priority. You set the shutter speed you want and the camera sets the aperture. To switch your camera to shutter priority, turn the dial on top of your camera to \u2018S\u2019. On Canon models, this mode is actually called \u201cTv\u201d mode which stands for \u201ctime value\u201d.<\/p><p>I personally don\u2019t use this mode too often. It can be useful if you need to set a minimum shutter speed in order to avoid camera shake. You may also want a specific longer shutter speed in order to create motion blur. I tend to use manual mode in this case as it gives me greater control over the shutter speed and aperture together. More on manual mode later.<\/p><h3>How to use the Exposure Compensation feature on your camera<\/h3><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation-button.jpg\" alt=\"exposure-compensation-button\" width=\"300\" height=\"300\" class=\"alignright size-full wp-image-224881\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation-button.jpg 300w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation-button-100x100.jpg 100w\" \/><\/p><p>Sometimes when you use aperture or shutter priority modes, you may find that your images are too bright or too dark. Sometimes the lighting conditions may confuse the camera and it results in the image being underexposed or overexposed. Thankfully, there is a way of fixing this. It\u2019s called exposure compensation. To switch this on, press the button with the plus\/minus symbols.<\/p><p>This will bring up a chart that goes from -5 to +5. Sometimes these numbers are different and may only range from -3 to +3 depending on the camera. This chart represents the exposure of your photograph. So how does it work?<\/p><p>When you are using aperture priority mode for example, the camera will set a shutter speed that makes the camera expose at the \u201c0\u201d point of this chart, right in the middle. In theory, this should be the correct exposure. In reality though, this is not always the case. As we said, some photos are supposed to be bright and others are supposed to be dark.<\/p><p>If you find that your photo is too bright or overexposed, you simply dial down the exposure by a stop or whatever you think is needed. When you turn the dial to the left (RTFM to see which dial), you can set the exposure at -1 for example. This will make the photograph 1 stop darker. When you turn the dial to the right, you can make the photo brighter. You may need to experiment a little to get the exposure you want.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation-800x241.jpg\" alt=\"exposure-compensation\" width=\"800\" height=\"241\" class=\"alignnone size-large wp-image-224882\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation-800x241.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation-320x96.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-compensation.jpg 1506w\" \/><\/p><p>How does exposure compensation work exactly? If you are using aperture priority mode and dial the exposure down 2 stops for example, the aperture will stay the same but the shutter speed will change to a faster speed so that less light enters the camera and the picture is made darker.<\/p><p>The opposite happens when you dial up the exposure. The aperture stays the same but the shutter speed will get longer to let more light in and make the image brighter. As already mentioned, keep an eye on the shutter speed if you are shooting hand held. Don\u2019t allow it to become too slow in order to avoid camera shake.<\/p><p>Exposure compensation works in the same way when using shutter priority mode except that the shutter speed will stay the same and the aperture will be changed by the camera accordingly.<\/p><h3>How to use Manual Mode on your camera<\/h3><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/manual-mode.jpg\" alt=\"manual-mode\" width=\"260\" height=\"180\" class=\"alignright size-full wp-image-224884\" \/><\/p><p>Here comes the scary one: manual mode! When you set the camera to manual mode, you set both the aperture and shutter speed. How do you know what combination to use to ensure the right exposure? It\u2019s actually quite easy. When you switch to manual mode on the dial (M), you again see an exposure chart that is exactly the same as the exposure compensation chart.<\/p><p>You then turn the aperture and shutter speed dials until the exposure is set to 0. Check your manual to see which dials to use.<\/p><p>Here is an example of how I might use manual mode when shooting a landscape:<\/p><ol><li>I decide what aperture I want to use. If it\u2019s a landscape, I might pick an aperture of about f\/16 to ensure plenty of depth of field. After all, I want everything to be sharp from the foreground to the background. I turn the aperture dial until, the aperture is set to f\/16.<\/li>\n<li>I then turn the shutter speed dial until the marker on the exposure chart is at zero. This in theory should mean that I now have the correct combination of aperture and shutter speed to ensure the right exposure.<\/li>\n<li>I then check that I am happy with both the aperture and shutter speed and make some adjustments if necessary.<\/li>\n<li>If I find that the shot is too bright or too dark I retake it after moving the dial to either minus a stop or plus a stop (or more as the case may be). The \u2018right\u2019 exposure may not always be at the \u201c0\u201d point in the middle. As I have said a few times now, some photos are supposed to be bright or dark.<\/li>\n<\/ol><h3>How do I decide which settings to use in manual mode?<\/h3><p>This is where your own creativity comes in to play. I usually decide which is the most important element in the photo and set this first. <\/p><p>As I mainly shoot urban landscape photos, this means I usually set the aperture first as ensuring plenty of depth of field is my biggest concern. I then set the shutter speed. It\u2019s basically a balancing act and with practice you will gain an intuition for what settings you need to achieve the vision you have for a particular photograph.<\/p><h3>What if the highlights are blown out or the shadows are clipped no matter what settings I use?<\/h3><p>Sometimes the contrast in a scene is simply too much for your camera to handle no matter which combination of aperture and shutter speed you use. In this case, bracketing can be used to solve the problem.<\/p><p>When I bracket a photo, I usually take 3 photos of the same scene, one with the exposure set to \u201c0\u201d, another deliberately underexposed by 2 stops and a final one deliberately overexposed by 2 stops. I can then combine these these in post-processing to get the \u2018perfect\u2019 exposure. There are several methods of doing this which I will cover in a future tutorial.<\/p><p>In the example below, I took 3 exposures of the Charles Bridge in Prague and blended them in post production to produce a single photograph with plenty of detail in all areas of the frame.<\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposurebracketing-800x203.jpg\" alt=\"exposurebracketing\" width=\"800\" height=\"203\" class=\"aligncenter size-large wp-image-224885\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposurebracketing-800x203.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposurebracketing-320x81.jpg 320w\" \/><\/p><p><img src=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-blending-example-800x431.jpg\" alt=\"exposure-blending-example\" width=\"800\" height=\"431\" class=\"aligncenter size-large wp-image-224886\" srcset=\"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-blending-example-800x431.jpg 800w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-blending-example-320x172.jpg 320w, https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/exposure-blending-example.jpg 1200w\" \/><\/p><p>The final photograph is a blend of all 3 images, leading to plenty of detail throughout the image. There are also no clipped shadows or blown out highlights. As you can see, I also cropped the final image to create a better composition.<\/p><hr \/><p>I hope that after reading this tutorial that you will be confident to take your camera off auto mode and take control of the settings yourself. In this way, you can move from merely taking a photography to making a photograph. Don\u2019t be afraid to experiment with all of the settings you have just learnt about. Over time, you won\u2019t even have to think too much about the settings.<\/p><p>I often advise students to go on a photo shoot where they specifically experiment with different apertures, another to experiment with shutter speed and so on.<\/p><hr \/><p><i><b>About the author<\/b>: Barry O Carroll is a Dublin, Ireland-based photographer specializing in landscape photography with a particular emphasis on urban landscapes, street scenes and architecture photography. You can find more of his work on <a href=\"http:\/\/www.bocphotography.com\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">his website<\/a> or by following him on <a href=\"https:\/\/www.facebook.com\/bocphotographyinfo\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">Facebook<\/a> and <a href=\"https:\/\/twitter.com\/bocphotos\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">Twitter<\/a>. This article was also published <a href=\"http:\/\/www.bocphotography.com\/guide-aperture-shutter-speed-iso-exposure-triangle\/\" data-wpel-link=\"external\" target=\"_blank\" rel=\"follow external noopener noreferrer\">here<\/a>.<\/i><\/p>","created_at":"2018-11-07T16:46:37+0000","updated_at":"2018-11-07T16:46:37+0000","published_at":"2016-06-25T16:06:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":20,"domain_name":"petapixel.com","preview_picture":"https:\/\/petapixel.com\/assets\/uploads\/2016\/06\/pyramid.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":380,"title":"Two Ways to Access Properties in ClojureScript","url":"http:\/\/squirrel.pl\/blog\/2013\/03\/28\/two-ways-to-access-properties-in-clojurescript\/","content":"<p>There are two pairs of complementary functions to set properties on objects in ClojureScript. One is <code>aset<\/code> and <code>aget<\/code>, another is <code>set!<\/code> and <code>.-propname<\/code>:<\/p>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(def scope (js-obj))&#13;\n(aset scope \"var1\" \"Value\")&#13;\n(aget scope \"var1\")&#13;\n<\/pre>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(def scope (js-obj))&#13;\n(set! (.-var2 scope) \"Value\")&#13;\n(.-var2 scope)&#13;\n<\/pre>\n<p>Are they equivalent? Is syntax the only difference?<\/p>\n<p>For a demo, let\u2019s say we set and print two fields on an object, like this:<\/p>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(def scope (js-obj))&#13;\n(aset scope \"var1\" \"Value\")&#13;\n(set! (.-var2 scope) \"Value\")&#13;\n&#13;\n(.log js\/console (str \"(aget scope \\\"var1\\\") -&gt;\" (aget scope \"var1\")))&#13;\n(.log js\/console (str \"(.-var2 scope) -&gt;\" (.-var2 scope)))&#13;\n(.log js\/console (str \"(.-var1 scope) -&gt;\" (.-var1 scope)))&#13;\n(.log js\/console (str \"(aget scope \\\"var2\\\") -&gt;\" (aget scope \"var2\")))&#13;\n<\/pre>\n<p>The resulting \u201csetting\u201d instructions are deceptively similar:<\/p>\n<pre class=\"brush: jscript; title: ; notranslate\" title=\"\">&#13;\nmynamespace.scope = {};&#13;\nmynamespace.scope[\"var1\"] = \"Value\";&#13;\nmynamespace.scope.var2 = \"Value\";&#13;\n<\/pre>\n<p>And of course all combinations print out correctly \u2013 I can see this on the console:<\/p>\n<pre class=\"brush: plain; title: ; notranslate\" title=\"\">&#13;\n(aget scope \"var1\") -&gt;Value&#13;\n(.-var2 scope) -&gt;Value&#13;\n(.-var1 scope) -&gt;Value&#13;\n(aget scope \"var2\") -&gt;Value&#13;\n<\/pre>\n<p>So far so good.<\/p>\n<p>Now, if you\u2019re at least a little bit serious about using ClojureScript, you have to consider advanced optimizations. What happens in this mode?<\/p>\n<p>My code got compiled to:<\/p>\n<pre class=\"brush: jscript; title: ; notranslate\" title=\"\">&#13;\nvar Xm;&#13;\nXm = {var1:\"Value\", lc:\"Value\"};&#13;\n<\/pre>\n<p>And on the console all I see is:<\/p>\n<pre class=\"brush: plain; title: ; notranslate\" title=\"\">&#13;\n(aget scope \"var1\") -&gt;Value&#13;\n(.-var2 scope) -&gt;Value&#13;\n(.-var1 scope) -&gt;&#13;\n(aget scope \"var2\") -&gt;&#13;\n<\/pre>\n<p>What happened to <code>var2<\/code>? Why is <code>set!<\/code> suddenly incompatible with <code>aget<\/code> and <code>aset<\/code> with <code>.-field<\/code>?<\/p>\n<p>If you look at compiler output in both cases, you should see that in one case the value is bound by name and in the other by symbol. In my understanding, it has a few consequences:<\/p>\n<ul><li>Advanced compiler renames symbols, but leaves strings intact.<\/li>\n<li><code>aset<\/code> and <code>aget<\/code> operate on names. <code>set!<\/code> and <code>.-var<\/code> operate on symbols. If you set a variable using \u201csymbolic reference\u201d, the compiler will rename the symbol so you won\u2019t be able to get it by name anymore. If you set it by name with <code>aset<\/code>, reading by symbol as <code>.-var<\/code> will not work either because the symbol is renamed.<\/li>\n<li>If you deal only with ClojureScript, you should probably use the symbolic references. They look more idiomatic and will benefit from advanced compiler features such as minimization. On the other hand, in this case you often may want to leverage the idiomatic data bindings \u2013 vars, atoms and such.<\/li>\n<li>As soon as you want your names to be accessible from outside, or you want to access names set outside, use <strong>only<\/strong> <code>aset<\/code>\/<code>aget<\/code>. It obviously applies to interop with JavaScript libraries. Less obvious case (for me) was libraries that operate on DOM, like Angular or Knockout. If you want that <code>{{user.name}}<\/code> to work, you cannot use the symbolic version.<\/li>\n<\/ul><p>It all makes sense now, but I haven\u2019t seen it documented. Well, that\u2019s still true for most ClojureScript.<\/p>","created_at":"2018-11-07T16:46:35+0000","updated_at":"2018-11-07T16:46:35+0000","published_by":["Konrad Garus"],"annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"squirrel.pl","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":381,"title":"No Support VPS Hosting - Portal Home","url":"http:\/\/nosupportvpshosting.com\/index.php","content":"<div class=\"reasons\"><h2>Reliable<\/h2><ul><li>RAID10 SAS<\/li>\n              <li>Server Grade Hardware<\/li>\n              <li>Dual UPS Protected<\/li>\n            <\/ul><\/div><div class=\"reasons\"><h2>Cheap<\/h2><ul><li>$15 per month<\/li>\n              <li>100GB disk space<\/li>\n              <li>2GB RAM &amp; 8 CPU<\/li>\n            <\/ul><\/div>No Support VPS Hosting provides reliable hosting services<a href=\"http:\/\/nosupportvpshosting.com\/cart.php?a=confproduct&amp;i=0\">Register Today &gt;&gt;<\/a>","created_at":"2018-11-07T16:46:35+0000","updated_at":"2018-11-07T16:46:35+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"nosupportvpshosting.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":382,"title":"No Support Linux Hosting - cPanel Powered Linux Web Hosting","url":"http:\/\/www.nosupportlinuxhosting.com\/","content":"<h3>Experts Host Sites Here for $1\/month<\/h3><p>Do you like paying extra so other people can ask <a href=\"http:\/\/www.nosupportlinuxhosting.com\/AmateurQuestions.html\"><strong>amateur questions<\/strong><\/a>?  That's how it is at other hosting companies where beginners and experts pay the same price.  Beginners drive up the cost by asking a lot of novice support questions while the experts don't contact support.  That is great for amateurs, and unfair to the experts like you.<\/p><p>No Support Linux Hosting has a completely different business model.  We ignore the support questions, and pass the savings on to you!  If you are an expert who does not want to pay extra for help with amateur support issues, then you can host with us and save big money.<\/p><p>Experts like you can <a href=\"https:\/\/www.nosupportlinuxhosting.com\/signup.aspx\"><strong>sign up now for free<\/strong><\/a>.  We charge $1\/month per website, and there is no limit to the number of websites you can host in your account.  This is the best deal in the web hosting industry, as long as you are the type of person who can find his or her own answers.<\/p>","created_at":"2018-11-07T16:46:35+0000","updated_at":"2018-11-07T16:46:35+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.nosupportlinuxhosting.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":379,"title":"Random with care | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=16060438","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16060438\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16060438\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/eev.ee\/blog\/2018\/01\/02\/random-with-care\/\" class=\"storylink\">Random with care<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=eev.ee\">eev.ee<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        89 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=dhotson\" class=\"hnuser\">dhotson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16060438\">10 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16060438&amp;goto=item%3Fid%3D16060438\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Random%20with%20care&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Random%20with%20care\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16060438&amp;auth=12926fe00839e49dec2d1b5a1d6aaa383f18e4aa\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16060438\">36\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:46:15+0000","updated_at":"2018-11-07T16:46:15+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":378,"title":"2,875,000+ free and premium vector icons. SVG, PNG, AI, CSH and PNG format.","url":"https:\/\/www.iconfinder.com\/","content":"2,875,000+ free and premium vector icons. SVG, PNG, AI, CSH and PNG format.<noscript>\n    \n    \n    \n    \n    \n        \n<header class=\"bg-dark text-light text-center pt-3 pb-5 d-none d-lg-block\" data-toggle=\"navigation-sticky\"><div class=\"container\"><div class=\"row justify-content-center\"><div class=\"col-lg-8 col-md-10\"><img src=\"https:\/\/cdn0.iconfinder.com\/static\/img\/logo\/white.svg?d0036af2c3\" alt=\"Iconfinder logo in white\" height=\"70\" \/><\/div><\/div><\/div>\n    <\/header><section class=\"d-none d-md-block bg-white py-4\"><div class=\"container\"><div class=\"row\"><div class=\"col-lg-7\"><h5>Get started with premium icons<\/h5><p class=\"m-0\">Read more about premium icons and how to purchase them. <\/p><\/div><div class=\"col-lg-5\"><p><a href=\"https:\/\/www.iconfinder.com\/buy-icons\" class=\"btn btn-outline-secondary px-md-4\" data-toggle=\"popover\" data-trigger=\"hover\" data-placement=\"bottom\" data-title=\"Prepayment\" data-content=\"A prepayment allows you to deposit money to your account balance on Iconfinder, which you can then spend on buying icons without having to enter your credit card details every time that you wish to make a purchase.\">Make prepayment<\/a>\n               <a href=\"https:\/\/www.iconfinder.com\/buy-icons\" class=\"btn btn-primary btn-blue px-md-4\" data-toggle=\"popover\" data-trigger=\"hover\" data-placement=\"bottom\" data-title=\"Iconfinder Pro subscription\" data-content=\"An Iconfinder Pro subscription offers you a convenient way of downloading and is ideal if you need icons often. You will save money and not have to complete a purchase every time you need premium icons.\">Subscribe to Pro<\/a><\/p><\/div><\/div><\/div><\/section><section class=\"bg-light border-top py-md-4\"><div class=\"container\"><div class=\"row my-3\"><p><\/p><h2 class=\"h5 m-0\">Featured premium icon sets<\/h2><div class=\"col-md-3\"><div class=\"text-md-right text-muted pt-1\"><a href=\"https:\/\/www.iconfinder.com\/icon-sets\/featured\/premium\" class=\"btn btn-sm btn-ghost btn-outline-secondary\">View more<\/a><a href=\"https:\/\/www.iconfinder.com\/iconsets\/feed\/rss\/premium\" class=\"btn btn-sm btn-ghost btn-outline-secondary\" title=\"RSS feed\" data-toggle=\"tooltip\">\n                                \n    <img src=\"https:\/\/icongr.am\/material\/rss.svg?size=16&amp;color=6c757d\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><\/div><\/div><div class=\"row\"><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/waste-sorting\">\n                    <img src=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/waste-sorting.png\" srcset=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/waste-sorting.png 1x, https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/waste-sorting.png 2x\" alt=\"Waste Sorting\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/waste-sorting\" class=\"d-block text-dark text-truncate\" title=\"Waste Sorting\">\n                                    Waste Sorting\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118432\" data-loading-title=\"Adding to cart...\">\n                                            $9.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                40 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118432\" data-type=\"iconset\" data-id=\"118432\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/prosymbols\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/9f097c29f08fc997d8de4a40b143d844?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            ProSymbols\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/marketing-and-digital-marketing\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/marketing-and-digital-marketing.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/marketing-and-digital-marketing.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/marketing-and-digital-marketing.png 2x\" alt=\"Marketing and Digital Marketing\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/marketing-and-digital-marketing\" class=\"d-block text-dark text-truncate\" title=\"Marketing and Digital Marketing\">\n                                    Marketing and Digital Marketing\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118364\" data-loading-title=\"Adding to cart...\">\n                                            $15.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                66 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118364\" data-type=\"iconset\" data-id=\"118364\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/KOLOdesign\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/3596f1155d7026a0f425b7467970c54e?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            KOLO design\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/green-energy-outline-1\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/green-energy-outline-1.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/green-energy-outline-1.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/green-energy-outline-1.png 2x\" alt=\"Green Energy (OUTLINE)\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/green-energy-outline-1\" class=\"d-block text-dark text-truncate\" title=\"Green Energy (OUTLINE)\">\n                                    Green Energy (OUTLINE)\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118344\" data-loading-title=\"Adding to cart...\">\n                                            $9.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                16 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118344\" data-type=\"iconset\" data-id=\"118344\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/andinur\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/ce3655de30e9aa43dbcb0f0f543e0143?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Andi Nur Abdillah\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/fast-food-outline-6\">\n                    <img src=\"https:\/\/cdn1.iconfinder.com\/data\/iconsets\/previews\/medium\/fast-food-outline-6.png\" srcset=\"https:\/\/cdn1.iconfinder.com\/data\/iconsets\/previews\/medium\/fast-food-outline-6.png 1x, https:\/\/cdn1.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/fast-food-outline-6.png 2x\" alt=\"Fast food Outline\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/fast-food-outline-6\" class=\"d-block text-dark text-truncate\" title=\"Fast food Outline\">\n                                    Fast food Outline\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118146\" data-loading-title=\"Adding to cart...\">\n                                            $25.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                36 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118146\" data-type=\"iconset\" data-id=\"118146\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/turkkub\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/6cefe733891eaef6713bce6a17d83911?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Turkkub .\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/professional-seo-v1-black-line\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/professional-seo-v1-black-line.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/professional-seo-v1-black-line.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/professional-seo-v1-black-line.png 2x\" alt=\"Professional SEO v1 (Black Line)\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/professional-seo-v1-black-line\" class=\"d-block text-dark text-truncate\" title=\"Professional SEO v1 (Black Line)\">\n                                    Professional SEO v1 (Black Line)\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118229\" data-loading-title=\"Adding to cart...\">\n                                            $12.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                25 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118229\" data-type=\"iconset\" data-id=\"118229\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Binpodo\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/4ad592ddcd26ba75365a5924fb9e79c4?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Binpodo Studio\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/black-friday-shopping-1\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/black-friday-shopping-1.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/black-friday-shopping-1.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/black-friday-shopping-1.png 2x\" alt=\"Black Friday Shopping\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/black-friday-shopping-1\" class=\"d-block text-dark text-truncate\" title=\"Black Friday Shopping\">\n                                    Black Friday Shopping\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118174\" data-loading-title=\"Adding to cart...\">\n                                            $25.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                22 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118174\" data-type=\"iconset\" data-id=\"118174\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/roundicons\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/ea4fcbd3c7139b205e27a03a315049a7?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            roundicons.com\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/traffic-jam\">\n                    <img src=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/traffic-jam.png\" srcset=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/traffic-jam.png 1x, https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/traffic-jam.png 2x\" alt=\"Traffic jam\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/traffic-jam\" class=\"d-block text-dark text-truncate\" title=\"Traffic jam\">\n                                    Traffic jam\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"118138\" data-loading-title=\"Adding to cart...\">\n                                            $29.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                20 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118138\" data-type=\"iconset\" data-id=\"118138\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/antto\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/213c23bb67ccb7bd547d0ca0420fd957?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Siwat V\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/items-outline\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/items-outline.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/items-outline.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/items-outline.png 2x\" alt=\"Items \/ outline\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/items-outline\" class=\"d-block text-dark text-truncate\" title=\"Items \/ outline\">\n                                    Items \/ outline\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"117264\" data-loading-title=\"Adding to cart...\">\n                                            $8.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                36 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117264\" data-type=\"iconset\" data-id=\"117264\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/e.c\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/c03b55ad5353aae1066b04737896b970?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Elisabetta Calabritto\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/computer-technology-21\">\n                    <img src=\"https:\/\/cdn1.iconfinder.com\/data\/iconsets\/previews\/medium\/computer-technology-21.png\" srcset=\"https:\/\/cdn1.iconfinder.com\/data\/iconsets\/previews\/medium\/computer-technology-21.png 1x, https:\/\/cdn1.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/computer-technology-21.png 2x\" alt=\"Computer Technology\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/computer-technology-21\" class=\"d-block text-dark text-truncate\" title=\"Computer Technology\">\n                                    Computer Technology\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"117476\" data-loading-title=\"Adding to cart...\">\n                                            $20.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                36 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117476\" data-type=\"iconset\" data-id=\"117476\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Studio03\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/777f14d4f7290e14011c4a2d80f4cca9?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            NhorPhai\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/black-friday-outline-2\">\n                    <img src=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/black-friday-outline-2.png\" srcset=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/black-friday-outline-2.png 1x, https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/black-friday-outline-2.png 2x\" alt=\"Black Friday OutLine\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/black-friday-outline-2\" class=\"d-block text-dark text-truncate\" title=\"Black Friday OutLine\">\n                                    Black Friday OutLine\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"117732\" data-loading-title=\"Adding to cart...\">\n                                            $18.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                50 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117732\" data-type=\"iconset\" data-id=\"117732\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/kerismaker\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/332c40af85eb319cc51d81904e80b395?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            kerismaker\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/black-friday-outline-filled\">\n                    <img src=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/black-friday-outline-filled.png\" srcset=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/black-friday-outline-filled.png 1x, https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/black-friday-outline-filled.png 2x\" alt=\"Black Friday Outline Filled\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/black-friday-outline-filled\" class=\"d-block text-dark text-truncate\" title=\"Black Friday Outline Filled\">\n                                    Black Friday Outline Filled\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"117733\" data-loading-title=\"Adding to cart...\">\n                                            $18.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                50 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117733\" data-type=\"iconset\" data-id=\"117733\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/kerismaker\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/332c40af85eb319cc51d81904e80b395?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            kerismaker\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/shopping-252\">\n                    <img src=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/shopping-252.png\" srcset=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/shopping-252.png 1x, https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/shopping-252.png 2x\" alt=\"Shopping\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/shopping-252\" class=\"d-block text-dark text-truncate\" title=\"Shopping\">\n                                    Shopping\n                                <\/a>\n                            <\/h4><p><a href=\"#\" data-cart=\"add\" data-remote=\"\/commerce\/ajax\/cart\/\" data-method=\"post\" data-reference_identifier=\"package\" data-reference_primary_key=\"117663\" data-loading-title=\"Adding to cart...\">\n                                            $11.00\n                                        <\/a><\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                60 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\"><a href=\"https:\/\/www.iconfinder.com\/licenses\/basic\" class=\"license-helper\" target=\"_blank\">Basic license<\/a><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon nor its edited version<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital templates<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon as (part of) a logo<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117663\" data-type=\"iconset\" data-id=\"117663\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/vectorsmarket\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/84a8170fe27723433a4a94f6296c22a2?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Vectors Market\n        <\/a><\/div><\/div><\/div><\/div><\/div><\/section><section class=\"bg-light border-top py-md-4\"><div class=\"container\"><div class=\"row my-3\"><p><\/p><h2 class=\"h5 m-0\">Featured free icon sets<\/h2><div class=\"col-md-3\"><div class=\"text-md-right text-muted pt-1\"><a href=\"https:\/\/www.iconfinder.com\/icon-sets\/featured\/free\" class=\"btn btn-sm btn-ghost btn-outline-secondary\">View more<\/a><a href=\"https:\/\/www.iconfinder.com\/iconsets\/feed\/rss\/free\" class=\"btn btn-sm btn-ghost btn-outline-secondary\" title=\"RSS feed\" data-toggle=\"tooltip\">\n                                \n    <img src=\"https:\/\/icongr.am\/material\/rss.svg?size=16&amp;color=6c757d\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><\/div><\/div><div class=\"row\"><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/business-and-finance-colorful-free-hand-drawn-set\">\n                    <img src=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/business-and-finance-colorful-free-hand-drawn-set.png\" srcset=\"https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium\/business-and-finance-colorful-free-hand-drawn-set.png 1x, https:\/\/cdn3.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/business-and-finance-colorful-free-hand-drawn-set.png 2x\" alt=\"Business and Finance Colorful Free Hand-drawn Set\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/business-and-finance-colorful-free-hand-drawn-set\" class=\"d-block text-dark text-truncate\" title=\"Business and Finance Colorful Free Hand-drawn Set\">\n                                    Business and Finance Colorful Free Hand-drawn Set\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                15 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/118573\" data-type=\"iconset\" data-id=\"118573\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/iconka\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/e4bb438844025dc2941d3ef4186c5324?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Iconka\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/jumpicon-basic-ui-line-1\">\n                    <img src=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/jumpicon-basic-ui-line-1.png\" srcset=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/jumpicon-basic-ui-line-1.png 1x, https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/jumpicon-basic-ui-line-1.png 2x\" alt=\"Jumpicon - Basic UI (Line)\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/jumpicon-basic-ui-line-1\" class=\"d-block text-dark text-truncate\" title=\"Jumpicon - Basic UI (Line)\">\n                                    Jumpicon - Basic UI (Line)\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                30 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by-nc\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution-Noncommercial 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117160\" data-type=\"iconset\" data-id=\"117160\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Ayub_Irawan\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/8dd328b1a7e3b7cfa4d4dfa135c3a234?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Ayub Irawan\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/jumpicon-basic-ui-glyph-1\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/jumpicon-basic-ui-glyph-1.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/jumpicon-basic-ui-glyph-1.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/jumpicon-basic-ui-glyph-1.png 2x\" alt=\"Jumpicon - Basic UI (Glyph)\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/jumpicon-basic-ui-glyph-1\" class=\"d-block text-dark text-truncate\" title=\"Jumpicon - Basic UI (Glyph)\">\n                                    Jumpicon - Basic UI (Glyph)\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                30 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by-nc\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution-Noncommercial 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/117159\" data-type=\"iconset\" data-id=\"117159\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Ayub_Irawan\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/8dd328b1a7e3b7cfa4d4dfa135c3a234?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Ayub Irawan\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/free-halloween-1\">\n                    <img src=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/free-halloween-1.png\" srcset=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/free-halloween-1.png 1x, https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/free-halloween-1.png 2x\" alt=\"Free Halloween\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/free-halloween-1\" class=\"d-block text-dark text-truncate\" title=\"Free Halloween\">\n                                    Free Halloween\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                17 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\">Free for commercial use (Include link to authors website)<h6>You must:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/alert.svg?size=16&amp;color=fdba50\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Include a link to the author's website<\/td>\n                    <\/tr><\/table><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital goods<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Sell the icon<\/td>\n                    <\/tr><\/table><\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/116137\" data-type=\"iconset\" data-id=\"116137\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/seanauicons\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/151405c5999de8ca82f5e754d8fb223b?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Iconsmama\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/halloween-01\">\n                    <img src=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/halloween-01.png\" srcset=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/halloween-01.png 1x, https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/halloween-01.png 2x\" alt=\"Halloween 01\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/halloween-01\" class=\"d-block text-dark text-truncate\" title=\"Halloween 01\">\n                                    Halloween 01\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                20 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\">Free for commercial use (Include link to authors website)<h6>You must:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/alert.svg?size=16&amp;color=fdba50\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Include a link to the author's website<\/td>\n                    <\/tr><\/table><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital goods<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Sell the icon<\/td>\n                    <\/tr><\/table><\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/116102\" data-type=\"iconset\" data-id=\"116102\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/josyanimator\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/b97bc60c17aac0566d39b3b1399efe03?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Josy Dom Alexis\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/halloween-free-1\">\n                    <img src=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/halloween-free-1.png\" srcset=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/halloween-free-1.png 1x, https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/halloween-free-1.png 2x\" alt=\"Halloween Free\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/halloween-free-1\" class=\"d-block text-dark text-truncate\" title=\"Halloween Free\">\n                                    Halloween Free\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                25 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\">Free for commercial use<h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital goods<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Sell the icon<\/td>\n                    <\/tr><\/table><p>Note: It is nice to attribute the author, but not mandatory for this license type.<\/p>\n        \n    <\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/82590\" data-type=\"iconset\" data-id=\"82590\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Nadiinko\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/efdcca9cd61a3a96a39edcce385280da?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Nadiinko\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/materia-flat-halloween-free\">\n                    <img src=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/materia-flat-halloween-free.png\" srcset=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/materia-flat-halloween-free.png 1x, https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/materia-flat-halloween-free.png 2x\" alt=\"Materia Flat Halloween free\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/materia-flat-halloween-free\" class=\"d-block text-dark text-truncate\" title=\"Materia Flat Halloween free\">\n                                    Materia Flat Halloween free\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                41 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\">Free for commercial use (Include link to authors website)<h6>You must:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/alert.svg?size=16&amp;color=fdba50\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Include a link to the author's website<\/td>\n                    <\/tr><\/table><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital goods<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Sell the icon<\/td>\n                    <\/tr><\/table><\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/115620\" data-type=\"iconset\" data-id=\"115620\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Icojam\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/dad057c7893f4c5dd7154165714ea405?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Icojam\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/materia-halloween-free\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/materia-halloween-free.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/materia-halloween-free.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/materia-halloween-free.png 2x\" alt=\"Materia Halloween free\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/materia-halloween-free\" class=\"d-block text-dark text-truncate\" title=\"Materia Halloween free\">\n                                    Materia Halloween free\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                41 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><div class=\"small text-truncate text-secondary\">Free for commercial use (Include link to authors website)<h6>You must:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/alert.svg?size=16&amp;color=fdba50\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Include a link to the author's website<\/td>\n                    <\/tr><\/table><h6>You are free to:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon for commercial purposes<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Share the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Edit the icon<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on merchandise for sale (T-shirts, mugs etc.)<\/td>\n                    <\/tr><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/check.svg?size=16&amp;color=339266\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Use the icon on mass distributed digital goods<\/td>\n                    <\/tr><\/table><h6>You cannot:<\/h6>\n            <table class=\"mb-2\"><tr><td class=\"pb-1 pr-1 align-top\">\n    <img src=\"https:\/\/icongr.am\/material\/close.svg?size=16&amp;color=b03d3e\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/><\/td>\n                        <td class=\"pb-1 align-top\">Sell the icon<\/td>\n                    <\/tr><\/table><\/div><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/115619\" data-type=\"iconset\" data-id=\"115619\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Icojam\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/dad057c7893f4c5dd7154165714ea405?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Icojam\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/space-82\">\n                    <img src=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/space-82.png\" srcset=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/space-82.png 1x, https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/space-82.png 2x\" alt=\"Space\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/space-82\" class=\"d-block text-dark text-truncate\" title=\"Space\">\n                                    Space\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                78 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/92330\" data-type=\"iconset\" data-id=\"92330\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Field5\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/3744109a1df131c1fd79525d95868581?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Good Stuff No Nonsense\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/seo-and-web-33\">\n                    <img src=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/seo-and-web-33.png\" srcset=\"https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium\/seo-and-web-33.png 1x, https:\/\/cdn2.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/seo-and-web-33.png 2x\" alt=\"SEO and WEB\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/seo-and-web-33\" class=\"d-block text-dark text-truncate\" title=\"SEO and WEB\">\n                                    SEO and WEB\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                36 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by-nc\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution-Noncommercial 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/115117\" data-type=\"iconset\" data-id=\"115117\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/photo3idea\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/e1e7addde5a0637af81b8cd8cbe10624?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Photo3idea studio\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/customer-support-24\">\n                    <img src=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/customer-support-24.png\" srcset=\"https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium\/customer-support-24.png 1x, https:\/\/cdn0.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/customer-support-24.png 2x\" alt=\"Customer Support\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/customer-support-24\" class=\"d-block text-dark text-truncate\" title=\"Customer Support\">\n                                    Customer Support\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                30 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/113180\" data-type=\"iconset\" data-id=\"113180\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/ratch0013\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/d3c6e0e5450fb8f04371f75d2ff8633d?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Eucalyp Studio\n        <\/a><\/div><\/div><\/div><div class=\"col-xl-3 col-lg-4 col-sm-6\"><div class=\"iconset-preview mb-4\"><div class=\"card\"><a href=\"https:\/\/www.iconfinder.com\/iconsets\/halloween-line-terror-night\">\n                    <img src=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/halloween-line-terror-night.png\" srcset=\"https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium\/halloween-line-terror-night.png 1x, https:\/\/cdn4.iconfinder.com\/data\/iconsets\/previews\/medium-2x\/halloween-line-terror-night.png 2x\" alt=\"Halloween Line - Terror night\" class=\"card-img-top\" \/><\/a><div class=\"card-body py-2 px-3\"><div class=\"row\"><p><\/p><h4 class=\"h6 m-0\">\n                                <a href=\"https:\/\/www.iconfinder.com\/iconsets\/halloween-line-terror-night\" class=\"d-block text-dark text-truncate\" title=\"Halloween Line - Terror night\">\n                                    Halloween Line - Terror night\n                                <\/a>\n                            <\/h4><p>Free<\/p><\/div><div class=\"row\"><p><small class=\"d-block text-muted\">\n                                25 icons\n                            <\/small><\/p><div class=\"col-7 text-right\"><p><a href=\"http:\/\/creativecommons.org\/licenses\/by\/3.0\/\" class=\"\" target=\"_blank\">Creative Commons (Attribution 3.0 Unported)<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"p-1\"><div class=\"float-right\"><a href=\"#\" data-toggle=\"tooltip\" title=\"Add to collection\" data-requires=\"user\" data-modal=\"collections\/_add\" data-remote=\"\/collections\/ajax\/list\/iconset\/113854\" data-type=\"iconset\" data-id=\"113854\">\n                            \n    <img src=\"https:\/\/icongr.am\/material\/bookmark.svg?size=14&amp;color=6c757d\" alt=\"\" width=\"14\" height=\"14\" class=\"ui-icon align-text-bottom\" \/><\/a><\/div><a href=\"https:\/\/www.iconfinder.com\/Chanut-is\" class=\"d-inline-block text-secondary small\">\n            <img src=\"https:\/\/www.gravatar.com\/avatar\/fde53f5b34b91e323bacebbc8145a9ef?d=https%3A\/\/www.iconfinder.com\/static\/img\/default-avatar-user.png&amp;s=32\" height=\"16\" width=\"16\" class=\"align-text-bottom rounded-circle\" alt=\"image\" \/>\n            Chanut is Industries\n        <\/a><\/div><\/div><\/div><\/div><\/div><\/section><section class=\"bg-light border-top py-md-4\"><div class=\"container\"><h2 class=\"h5 my-3\">Categories<\/h2><div class=\"row\"><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/abstract-icons\" class=\"text-secondary\">\n                        Abstract\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/agriculture-farming-gardening-icons\" class=\"text-secondary\">\n                        Agriculture, farming, &amp; gardening\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/animal-icons\" class=\"text-secondary\">\n                        Animals\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/arrow-icons\" class=\"text-secondary\">\n                        Arrows\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/avatars-smiley-icons\" class=\"text-secondary\">\n                        Avatars &amp; smileys\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/business-finance-icons\" class=\"text-secondary\">\n                        Business &amp; finance\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/christmas-icons\" class=\"text-secondary\">\n                        Christmas\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/clothes-accessory-icons\" class=\"text-secondary\">\n                        Clothes &amp; accessories\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/computer-hardware-icons\" class=\"text-secondary\">\n                        Computer &amp; hardware\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/culture-communities-icons\" class=\"text-secondary\">\n                        Culture &amp; communities\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/desktop-app-icons\" class=\"text-secondary\">\n                        Desktop apps\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/easter-icons\" class=\"text-secondary\">\n                        Easter\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/ecology-environmentalism-icons\" class=\"text-secondary\">\n                        Ecology &amp; environmentalism\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/education-science-icons\" class=\"text-secondary\">\n                        Education &amp; science\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/electronics-appliances-icons\" class=\"text-secondary\">\n                        Electronics &amp; appliances\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/events-and-entertainment-icons\" class=\"text-secondary\">\n                        Events &amp; Entertainment\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/fall-icons\" class=\"text-secondary\">\n                        Fall\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/family-home-icons\" class=\"text-secondary\">\n                        Family &amp; home\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/file-folder-icons\" class=\"text-secondary\">\n                        Files &amp; folders\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/flag-icons\" class=\"text-secondary\">\n                        Flags\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/food-drinks-icons\" class=\"text-secondary\">\n                        Food &amp; drinks\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/furniture-home-decorations-icons\" class=\"text-secondary\">\n                        Furniture &amp; home decorations\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/gaming-gambling-icons\" class=\"text-secondary\">\n                        Gaming &amp; gambling\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/halloween-icons\" class=\"text-secondary\">\n                        Halloween\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/health-beauty-and-fashion-icons\" class=\"text-secondary\">\n                        Health, Beauty, &amp; Fashion\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/healthcare-medical-icons\" class=\"text-secondary\">\n                        Healthcare &amp; medical\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/holidays-icons\" class=\"text-secondary\">\n                        Holidays\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/interior-building-icons\" class=\"text-secondary\">\n                        Interior &amp; buildings\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/ios7-optimized-icons\" class=\"text-secondary\">\n                        iOS7 optimized\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/maps-navigation-icons\" class=\"text-secondary\">\n                        Maps &amp; navigation\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/military-law-enforcement-icons\" class=\"text-secondary\">\n                        Military &amp; law enforcement\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/mixed-icons\" class=\"text-secondary\">\n                        Mixed\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/mobile-app-icons\" class=\"text-secondary\">\n                        Mobile apps\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/music-multimedia-icons\" class=\"text-secondary\">\n                        Music &amp; multimedia\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/nature-outdoor-icons\" class=\"text-secondary\">\n                        Nature &amp; outdoors\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/network-communication-icons\" class=\"text-secondary\">\n                        Network &amp; communications\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/nsfw-icons\" class=\"text-secondary\">\n                        NSFW\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/photography-graphic-design-icons\" class=\"text-secondary\">\n                        Photography &amp; graphic design\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/real-estate-icons\" class=\"text-secondary\">\n                        Real Estate\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/recreation-hobby-icons\" class=\"text-secondary\">\n                        Recreation &amp; hobbies\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/religion-philosophy-icons\" class=\"text-secondary\">\n                        Religion &amp; philosophy\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/romance-icons\" class=\"text-secondary\">\n                        Romance\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/seasons-icons\" class=\"text-secondary\">\n                        Seasons\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/security-icons\" class=\"text-secondary\">\n                        Security\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/seo-web-icons\" class=\"text-secondary\">\n                        SEO &amp; web\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/shipping-delivery-and-fulfillment-icons\" class=\"text-secondary\">\n                        Shipping, Delivery, &amp; Fulfillment\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/shopping-ecommerce-icons\" class=\"text-secondary\">\n                        Shopping &amp; e-commerce\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/sign-symbol-icons\" class=\"text-secondary\">\n                        Signs &amp; symbols\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/social-media-icons\" class=\"text-secondary\">\n                        Social media\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/sports-award-icons\" class=\"text-secondary\">\n                        Sports &amp; awards\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/spring-icons\" class=\"text-secondary\">\n                        Spring\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/summer-icons\" class=\"text-secondary\">\n                        Summer\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/tools-construction-icons\" class=\"text-secondary\">\n                        Tools &amp; construction\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/touch-gesture-icons\" class=\"text-secondary\">\n                        Touch gestures\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/toys-childhood-icons\" class=\"text-secondary\">\n                        Toys &amp; childhood\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/transportation-icons\" class=\"text-secondary\">\n                        Transportation\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/travel-hotel-icons\" class=\"text-secondary\">\n                        Travel &amp; hotels\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/ui-icons\" class=\"text-secondary\">\n                        UI\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/weather-icons\" class=\"text-secondary\">\n                        Weather\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/winter-icons\" class=\"text-secondary\">\n                        Winter\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><\/div><h2 class=\"h5 my-3\">Styles<\/h2><div class=\"row\"><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/glyph-style-icons\" class=\"text-secondary\">\n                        Glyph\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/outline-style-icons\" class=\"text-secondary\">\n                        Outline\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/flat-style-icons\" class=\"text-secondary\">\n                        Flat\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/filled_outline-style-icons\" class=\"text-secondary\">\n                        Filled outline\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/photorealistic-style-icons\" class=\"text-secondary\">\n                        Photorealistic\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/3d-style-icons\" class=\"text-secondary\">\n                        3D\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/handdrawn-style-icons\" class=\"text-secondary\">\n                        Handdrawn\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/cartoon-style-icons\" class=\"text-secondary\">\n                        Cartoon\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/pixel-style-icons\" class=\"text-secondary\">\n                        Pixel\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled py-md-2\"><li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/smooth-style-icons\" class=\"text-secondary\">\n                        Smooth\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/long-shadow-style-icons\" class=\"text-secondary\">\n                        Long Shadow\n                    <\/a>\n                <\/li>\n                <li class=\"pb-2\">\n                    <a href=\"https:\/\/www.iconfinder.com\/badge-style-icons\" class=\"text-secondary\">\n                        Badge\n                    <\/a>\n                <\/li>\n        <\/ul><\/div><\/div><\/div><\/section><footer class=\"py-4 py-md-5 bg-white border-top\"><div class=\"container\"><p class=\"mb-md-5\">\n            <a href=\"https:\/\/www.iconfinder.com\/\">\n                <img src=\"https:\/\/cdn0.iconfinder.com\/static\/img\/logo\/black.svg?32b294b019\" alt=\"Iconfinder\" height=\"30\" class=\"opacity-30\" \/><\/a>\n        <\/p><div class=\"row\"><div class=\"col-sm-6 col-lg-3\"><ul class=\"list-unstyled\"><li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/buy-icons\" class=\"text-secondary\">Buy icons<\/a><\/li>\n                    <li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/custom-icon-design\" class=\"text-secondary\">Custom icon design<\/a><\/li>\n                    <li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/icon-editor\" class=\"text-secondary\">Icon editor<\/a><\/li>\n                    <li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/api-solution\" class=\"text-secondary\">API<\/a><\/li>\n                    <li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/tags\" class=\"text-secondary\">Tags<\/a><\/li>\n                    <li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/free_icons\" class=\"text-secondary\">Free icons<\/a><\/li>\n                    <li class=\"pb-2\"><a href=\"https:\/\/www.iconfinder.com\/about\" class=\"text-secondary\">About Iconfinder<\/a><\/li>\n                <\/ul><\/div><div class=\"col-lg-6\"><p class=\"text-secondary\">\n                    Made in Copenhagen \u00b7\n                    Copyright 2018 \u00b7\n                    All rights reserved.\n                <\/p><p class=\"text-secondary\">\n                    <a href=\"https:\/\/www.facebook.com\/iconfinder\" class=\"text-secondary\" target=\"_blank\">Facebook<\/a> \u00b7\n                    <a href=\"https:\/\/www.twitter.com\/iconfinder\" class=\"text-secondary\" target=\"_blank\">Twitter<\/a> \u00b7\n                    <a href=\"https:\/\/www.dribbble.com\/team-iconfinder\" class=\"text-secondary\" target=\"_blank\">Dribbble<\/a> \u00b7\n                    <a href=\"https:\/\/www.instagram.com\/iconfinder\" class=\"text-secondary\" target=\"_blank\">Instagram<\/a>\n                <\/p><p class=\"creditcard-logos\">\n                    <a href=\"http:\/\/support.iconfinder.com\/purchasing-and-downloading\/supported-payment-methods\" data-toggle=\"tooltip\" title=\"We accept VISA\">\n                        <img src=\"https:\/\/cdn0.iconfinder.com\/static\/img\/cards\/visa.svg?0ebb52d964\" width=\"32\" alt=\"image\" \/><\/a>\n                    <a href=\"http:\/\/support.iconfinder.com\/purchasing-and-downloading\/supported-payment-methods\" data-toggle=\"tooltip\" title=\"We accept MasterCard\">\n                        <img src=\"https:\/\/cdn0.iconfinder.com\/static\/img\/cards\/mastercard.svg?063b13c064\" width=\"32\" alt=\"image\" \/><\/a>\n                    <a href=\"http:\/\/support.iconfinder.com\/purchasing-and-downloading\/supported-payment-methods\" data-toggle=\"tooltip\" title=\"We accept American Express\">\n                        <img src=\"https:\/\/cdn0.iconfinder.com\/static\/img\/cards\/amex.svg?444731bb95\" width=\"32\" alt=\"image\" \/><\/a>\n                    <a href=\"http:\/\/support.iconfinder.com\/purchasing-and-downloading\/supported-payment-methods\" data-toggle=\"tooltip\" title=\"We accept PayPal\">\n                        <img src=\"https:\/\/cdn0.iconfinder.com\/static\/img\/cards\/paypal.svg?3bd17ec425\" width=\"32\" alt=\"image\" \/><\/a>\n                    <a href=\"http:\/\/support.iconfinder.com\/purchasing-and-downloading\/is-credit-card-data-sent-using-a-secure-connection\" class=\"text-secondary ml-3\" target=\"_blank\">\n                        \n    <img src=\"https:\/\/icongr.am\/material\/lock.svg?size=16&amp;color=6c757d\" alt=\"\" width=\"16\" height=\"16\" class=\"ui-icon align-text-bottom\" \/>\n                        Secure connection\n                    <\/a>\n                <\/p><\/div><\/div><\/div>\n<\/footer>\n<\/noscript>","created_at":"2018-11-07T16:46:13+0000","updated_at":"2018-11-07T16:46:13+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"www.iconfinder.com","preview_picture":"https:\/\/cdn0.iconfinder.com\/static\/img\/meta-generic.png?01f5058368","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":376,"title":"Harvesting credit card numbers and passwords from websites","url":"https:\/\/news.ycombinator.com\/item?id=16084575","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=btown\" class=\"hnuser\">btown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086248\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=djsumdog\" class=\"hnuser\">djsumdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086597\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087869\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JonathonW\" class=\"hnuser\">JonathonW<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088539\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tomte\" class=\"hnuser\">Tomte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16116630\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=asaph\" class=\"hnuser\">asaph<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088145\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=altstar\" class=\"hnuser\">altstar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16095644\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jarofgreen\" class=\"hnuser\">jarofgreen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086596\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nikanj\" class=\"hnuser\">nikanj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086603\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jarofgreen\" class=\"hnuser\">jarofgreen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086703\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ehsankia\" class=\"hnuser\">ehsankia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090233\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087908\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bubble_boi\" class=\"hnuser\">bubble_boi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088257\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088304\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rahimnathwani\" class=\"hnuser\">rahimnathwani<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086665\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danschumann\" class=\"hnuser\">danschumann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086776\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nallerooth\" class=\"hnuser\">nallerooth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16095411\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jmadsen\" class=\"hnuser\">jmadsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087639\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ufmace\" class=\"hnuser\">ufmace<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16091973\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alceta\" class=\"hnuser\">alceta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16113439\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=milesvp\" class=\"hnuser\">milesvp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087696\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swyx\" class=\"hnuser\">swyx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090165\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swyx\" class=\"hnuser\">swyx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090354\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chmln\" class=\"hnuser\">chmln<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16089181\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jmadsen\" class=\"hnuser\">jmadsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087788\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Gaelan\" class=\"hnuser\">Gaelan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16089061\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jmadsen\" class=\"hnuser\">jmadsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16095304\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=drdrey\" class=\"hnuser\">drdrey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087761\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jmadsen\" class=\"hnuser\">jmadsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087795\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=IshKebab\" class=\"hnuser\">IshKebab<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088297\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jmadsen\" class=\"hnuser\">jmadsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088499\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fps_doug\" class=\"hnuser\">fps_doug<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16095585\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=IshKebab\" class=\"hnuser\">IshKebab<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16092820\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rdiddly\" class=\"hnuser\">rdiddly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088138\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=downandout\" class=\"hnuser\">downandout<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087657\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dbattaglia\" class=\"hnuser\">dbattaglia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088305\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mundungus_a\" class=\"hnuser\">mundungus_a<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16092541\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=agumonkey\" class=\"hnuser\">agumonkey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16092446\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=spiznnx\" class=\"hnuser\">spiznnx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085163\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=deckard1\" class=\"hnuser\">deckard1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088008\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ehsankia\" class=\"hnuser\">ehsankia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090242\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fny\" class=\"hnuser\">fny<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086600\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saryant\" class=\"hnuser\">saryant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086721\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cesarb\" class=\"hnuser\">cesarb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087724\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=martin-adams\" class=\"hnuser\">martin-adams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16092211\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bsilvereagle\" class=\"hnuser\">bsilvereagle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087939\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bubble_boi\" class=\"hnuser\">bubble_boi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088272\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=monokrome\" class=\"hnuser\">monokrome<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16095407\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=awq123\" class=\"hnuser\">awq123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16096886\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aristeol2000\" class=\"hnuser\">aristeol2000<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16097803\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brown9-2\" class=\"hnuser\">brown9-2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086926\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=epmatsw\" class=\"hnuser\">epmatsw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086952\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hrrsn\" class=\"hnuser\">hrrsn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087970\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Shoothe\" class=\"hnuser\">Shoothe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085445\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dkersten\" class=\"hnuser\">dkersten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085706\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Silhouette\" class=\"hnuser\">Silhouette<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087016\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Shoothe\" class=\"hnuser\">Shoothe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087513\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rmrfrmrf\" class=\"hnuser\">rmrfrmrf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088731\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087810\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Silhouette\" class=\"hnuser\">Silhouette<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087996\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmitriid\" class=\"hnuser\">dmitriid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090069\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bo1024\" class=\"hnuser\">bo1024<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088252\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=martin-adams\" class=\"hnuser\">martin-adams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16092309\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danschumann\" class=\"hnuser\">danschumann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086821\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bubble_boi\" class=\"hnuser\">bubble_boi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088249\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tlrobinson\" class=\"hnuser\">tlrobinson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090505\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tlrobinson\" class=\"hnuser\">tlrobinson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090553\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=danschumann\" class=\"hnuser\">danschumann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086866\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vjeux\" class=\"hnuser\">vjeux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086942\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saryant\" class=\"hnuser\">saryant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086867\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=true_tuna\" class=\"hnuser\">true_tuna<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088302\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tyingq\" class=\"hnuser\">tyingq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087477\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mwarkentin\" class=\"hnuser\">mwarkentin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087502\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=beams_of_light\" class=\"hnuser\">beams_of_light<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086676\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gcornut\" class=\"hnuser\">gcornut<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16097795\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moreless\" class=\"hnuser\">moreless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16104713\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmlnr\" class=\"hnuser\">pmlnr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16122796\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nailer\" class=\"hnuser\">nailer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086748\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bdcravens\" class=\"hnuser\">bdcravens<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086978\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nailer\" class=\"hnuser\">nailer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087825\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xmohit\" class=\"hnuser\">0xmohit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086778\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SuperNinKenDo\" class=\"hnuser\">SuperNinKenDo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088293\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noxecanexx\" class=\"hnuser\">noxecanexx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088987\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=systematical\" class=\"hnuser\">systematical<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16086655\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mkopinsky\" class=\"hnuser\">mkopinsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16093841\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=slayer_birden\" class=\"hnuser\">slayer_birden<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16157722\">9 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sam_goody\" class=\"hnuser\">sam_goody<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16181928\">9 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=drdrey\" class=\"hnuser\">drdrey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087778\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=macawfish\" class=\"hnuser\">macawfish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087997\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rmrfrmrf\" class=\"hnuser\">rmrfrmrf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088743\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=z3t4\" class=\"hnuser\">z3t4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087425\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087929\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=z3t4\" class=\"hnuser\">z3t4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088327\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16089580\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=z3t4\" class=\"hnuser\">z3t4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16089926\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rmrfrmrf\" class=\"hnuser\">rmrfrmrf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088776\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16089585\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=z3t4\" class=\"hnuser\">z3t4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16089903\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=FLUX-YOU\" class=\"hnuser\">FLUX-YOU<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087653\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shaneos\" class=\"hnuser\">shaneos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088059\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Qerub\" class=\"hnuser\">Qerub<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088212\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shaneos\" class=\"hnuser\">shaneos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088397\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adbachman\" class=\"hnuser\">adbachman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16091107\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wahnfrieden\" class=\"hnuser\">wahnfrieden<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16090691\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=antsar\" class=\"hnuser\">antsar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16091001\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=program_whiz\" class=\"hnuser\">program_whiz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16097395\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmlnr\" class=\"hnuser\">pmlnr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16122808\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bmease\" class=\"hnuser\">bmease<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087848\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MollyR\" class=\"hnuser\">MollyR<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088027\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rollcat\" class=\"hnuser\">rollcat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088208\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joneil\" class=\"hnuser\">joneil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16093711\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MollyR\" class=\"hnuser\">MollyR<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16088415\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kw71\" class=\"hnuser\">kw71<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085073\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdamodharan\" class=\"hnuser\">sdamodharan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085815\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GoToRO\" class=\"hnuser\">GoToRO<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085896\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jbg_\" class=\"hnuser\">jbg_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085849\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mulmen\" class=\"hnuser\">mulmen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087960\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codedokode\" class=\"hnuser\">codedokode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085618\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=prh8\" class=\"hnuser\">prh8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16087124\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dkersten\" class=\"hnuser\">dkersten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085674\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway2048\" class=\"hnuser\">throwaway2048<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16085647\">10 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:46:12+0000","updated_at":"2018-11-07T16:46:12+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":377,"title":"Ten Things I Wish I\u2019d Known About bash","url":"https:\/\/zwischenzugs.com\/2018\/01\/06\/ten-things-i-wish-id-known-about-bash\/","content":"<h2>Intro<\/h2>\n<p>Recently\u00a0I wanted to deepen my understanding of bash by researching as much of it as possible. Because I felt bash is an often-used (and under-understood) technology, I ended up writing a <a href=\"https:\/\/leanpub.com\/learnbashthehardway\">book on it<\/a>.<\/p>\n<p>A preview is available\u00a0<a title=\"learnbashthehardway-sample-preview\" href=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/learnbashthehardway-sample-preview.pdf\">here<\/a>.<\/p>\n<p>You don\u2019t have to look hard on the internet to find plenty of useful one-liners in bash, or scripts. And there are guides to bash that seem somewhat intimidating through either their thoroughness or their focus on esoteric detail.<\/p>\n<p>Here I\u2019ve focussed on the things that either confused me or increased my power and productivity in bash significantly, and tried to communicate them (as in my book) in a way that emphasises getting the understanding right.<\/p>\n<p>Enjoy!<\/p>\n<p><a href=\"https:\/\/leanpub.com\/learnbashthehardway\"><img class=\"alignnone size-thumbnail wp-image-4192\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=94&amp;h=150\" alt=\"hero\" width=\"94\" height=\"150\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=94&amp;h=150 94w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300 188w\" \/><\/a><\/p>\n<h2>1)\u00a0 <code>``<\/code>\u00a0vs\u00a0<code>$()<\/code><\/h2>\n<p>These two operators do the same thing. Compare these two lines:<\/p>\n<pre>$ echo `ls`\n$ echo $(ls)<\/pre>\n<p>Why these two forms existed confused me for a long time.<\/p>\n<p>If you don\u2019t know, both forms substitute the output of the command contained within it into the command.<\/p>\n<p>The principal difference is that nesting is simpler.<\/p>\n<p>Which of these is easier to read (and write)?<\/p>\n<pre class=\"p1\">\u00a0 \u00a0 $ echo `echo \\`echo \\\\\\`echo inside\\\\\\`\\``\n<\/pre>\n<p>or:<\/p>\n<pre class=\"p2\">\u00a0 \u00a0 $ echo $(echo $(echo $(echo inside)))<\/pre>\n<p>If you\u2019re interested in going deeper, see <a href=\"http:\/\/mywiki.wooledge.org\/BashFAQ\/082\">here<\/a> or <a href=\"https:\/\/stackoverflow.com\/questions\/4708549\/what-is-the-difference-between-command-and-command-in-shell-programming\">here<\/a>.<\/p>\n<h2>2) globbing vs regexps<\/h2>\n<p>Another one that can confuse if never thought about or researched.<\/p>\n<p>While globs and regexps can look similar, they are not the same.<\/p>\n<p>Consider this command:<\/p>\n<pre class=\"p1\">$ rename -n 's\/(.*)\/new$1\/' *<\/pre>\n<p>The two asterisks are interpreted in different ways.<\/p>\n<p>The first is ignored by the shell (because it is in quotes), and is interpreted as \u20180 or more characters\u2019 by the rename application. So it\u2019s interpreted as a regular expression.<\/p>\n<p>The second is interpreted by the shell (because it is not in quotes), and gets replaced by a list of all the files in the current working folder. It is interpreted as a glob.<\/p>\n<p>So by looking at\u00a0<code>man bash<\/code>\u00a0can you figure out why these two commands produce different output?<\/p>\n<pre>$ ls *\n$ ls .*<\/pre>\n<p>The second looks even more like a regular expression. But it isn\u2019t!<\/p>\n<h2>3) Exit Codes<\/h2>\n<p>Not everyone knows that every time you run a shell command in bash, an \u2018exit code\u2019 is returned to bash.<\/p>\n<p>Generally, if a command \u2018succeeds\u2019 you get an error code of <code>0<\/code>. If it doesn\u2019t succeed, you get a non-zero code. <code>1<\/code>\u00a0is a \u2018general error\u2019, and others can give you more information (eg which signal killed it, for example).<\/p>\n<p>But these rules don\u2019t always hold:<\/p>\n<pre>$ grep not_there \/dev\/null\n$ echo $?<\/pre>\n<p><code>$?<\/code>\u00a0is a special bash variable that\u2019s set to the exit code of each command after it runs.<\/p>\n<p>Grep uses exit codes to indicate whether it matched or not. I have to look up every time which way round it goes: does finding a match or not return\u00a0<code>0<\/code>?<\/p>\n<p>Grok this and a lot will click into place in what follows.<\/p>\n<h2>4) <code>if<\/code> statements, <code>[<\/code>\u00a0and <code>[[<\/code><\/h2>\n<p>Here\u2019s another \u2018spot the difference\u2019 similar to the backticks one above.<\/p>\n<p>What will this output?<\/p>\n<pre>if grep not_there \/dev\/null\nthen\n    echo hi\nelse\n    echo lo\nfi<\/pre>\n<p>grep\u2019s return code makes code like this work more intuitively as a side effect of its use of exit codes.<\/p>\n<p>Now what will this output?<\/p>\n<p>a) <code>hihi<\/code><br \/>\nb) <code>lolo<\/code><br \/>\nc) something else<\/p>\n<pre>if [ $(grep not_there \/dev\/null) = '' ]\nthen\n\u00a0 \u00a0 echo -n hi\nelse\n\u00a0 \u00a0 echo -n lo\nfi\nif [[ $(grep not_there \/dev\/null) = '' ]]\nthen\n\u00a0 \u00a0 echo -n hi\nelse\n\u00a0 \u00a0 echo -n lo\nfi<\/pre>\n<p>The difference between <code>[<\/code>\u00a0and <code>[[<\/code>\u00a0was another thing I never really understood. <code>[<\/code>\u00a0is the original form for tests, and then <code>[[<\/code>\u00a0was introduced, which is more flexible and intuitive. In the first <code>if<\/code> block above, the if statement barfs because the <code>$(grep not_there \/dev\/null)<\/code>\u00a0is evaluated to nothing, resulting in this comparison:<\/p>\n<p><code>[ = '' ]<\/code><\/p>\n<p>which makes no sense. The double bracket form handles this for you.<\/p>\n<p>This is why you occasionally see comparisons like this in bash scripts:<\/p>\n<p><code>if [ x$(grep not_there \/dev\/null) = 'x' ]<\/code><\/p>\n<p>so that if the command returns nothing it still runs. There\u2019s no need for it, but that\u2019s why it exists.<\/p>\n<h2>5) <code>set<\/code>s<\/h2>\n<p>Bash has configurable options which can be set on the fly. I use two of these all the time:<\/p>\n<pre>set -e<\/pre>\n<p>exits from a script if any command returned a non-zero exit code (see above).<\/p>\n<p>This outputs the commands that get run as they run:<\/p>\n<pre>set -x<\/pre>\n<p>So a script might start like this:<\/p>\n<pre>#!\/bin\/bash\nset -e\nset -x\ngrep not_there \/dev\/null\necho $?<\/pre>\n<p>What would that script output?<\/p>\n<h2>6) \u200b\u200b<code>&lt;()<\/code><\/h2>\n<p>This is my favourite. It\u2019s so under-used, perhaps because it can be initially baffling, but I use it all the time.<\/p>\n<p>It\u2019s similar to <code>$()<\/code>\u00a0in that the output of the command inside is re-used.<\/p>\n<p>In this case, though, the output is treated as a file. This file can be used as an argument to commands that take files as an argument.<\/p>\n<p>Confused? Here\u2019s an example.<\/p>\n<p>Have you ever done something like this?<\/p>\n<pre>$ grep somestring file1 &gt; \/tmp\/a\n$ grep somestring file2 &gt; \/tmp\/b\n$ diff \/tmp\/a \/tmp\/b<\/pre>\n<p>That works, but instead you can write:<\/p>\n<pre>diff &lt;(grep somestring file1) &lt;(grep somestring file2)<\/pre>\n<p>Isn\u2019t that neater?<\/p>\n<h2>7) Quoting<\/h2>\n<p>Quoting\u2019s a knotty subject in bash, as it is in many software contexts.<\/p>\n<p>Firstly, variables in quotes:<\/p>\n<pre class=\"p1\">A='123' \u00a0\necho \"$A\"\necho '$A'<\/pre>\n<p>Pretty simple \u2013 double quotes dereference variables, while single quotes go literal.<\/p>\n<p>So what will this output?<\/p>\n<pre>mkdir -p tmp\ncd tmp\ntouch a\necho \"*\"\necho '*'<\/pre>\n<p>Surprised? I was.<\/p>\n<h2>8) Top three shortcuts<\/h2>\n<p>There are plenty of shortcuts listed in <code>man bash<\/code>, and it\u2019s not hard to find comprehensive lists. This list consists of the ones I use most often, in order of how often I use them.<\/p>\n<p>Rather than trying to memorize them all, I recommend picking one, and trying to remember to use it until it becomes unconscious. Then take the next one. I\u2019ll skip over the most obvious ones (eg <code>!!<\/code>\u00a0\u2013 repeat last command, and <code>~<\/code>\u00a0\u2013 your home directory).<\/p>\n<p><code><strong>!$<\/strong><\/code><\/p>\n<p>I use this dozens of times a day. It repeats the last argument of the last command. If you\u2019re working on a file, and can\u2019t be bothered to re-type it command after command it can save a lot of work:<\/p>\n<pre>grep somestring \/long\/path\/to\/some\/file\/or\/other.txt\nvi !$<\/pre>\n\n<p>\u200b\u200b<strong><code>!:1-$<\/code><\/strong><\/p>\n<p>This bit of magic takes this further. It takes all the arguments to the previous command and drops them in. So:<\/p>\n<pre>grep isthere\u00a0\/long\/path\/to\/some\/file\/or\/other.txt\negrep !:1-$\nfgrep !:1-$<\/pre>\n<p>The <code>!<\/code>\u00a0means \u2018look at the previous command\u2019, the <code>:<\/code>\u00a0is a separator, and the <code>1<\/code>\u00a0means \u2018take the first word\u2019, the <code>-<\/code>\u00a0means \u2018until\u2019 and the <code>$<\/code>\u00a0means \u2018the last word\u2019.<\/p>\n<p><strong>Note:<\/strong> you can achieve the same thing with <code>!*<\/code>. Knowing the above gives you the control to limit to a specific contiguous subset of arguments, eg with <code>!:2-3<\/code>.<\/p>\n<p><code><strong>:h<\/strong><\/code><\/p>\n<p>I use this one a lot too. If you put it after a filename, it will change that filename to remove everything up to the folder. Like this:<\/p>\n<pre>grep isthere\u00a0\/long\/path\/to\/some\/file\/or\/other.txt\ncd !$:h<\/pre>\n<p>which can save a lot of work in the course of the day.<\/p>\n<h2>9) startup order<\/h2>\n<p>The order in which bash runs startup scripts can cause a lot of head-scratching. I keep this diagram handy (from <a href=\"https:\/\/blog.flowblok.id.au\/2013-02\/shell-startup-scripts.html\">this<\/a> great page):<\/p>\n<p><img data-attachment-id=\"4194\" data-permalink=\"https:\/\/zwischenzugs.com\/2018\/01\/06\/ten-things-i-wish-id-known-about-bash\/shell-startup-actual\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=525\" data-orig-size=\"1558,955\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"shell-startup-actual\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=525?w=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=525?w=525\" class=\"alignnone size-full wp-image-4194\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=525\" alt=\"shell-startup-actual\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=525 525w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=1050 1050w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=150 150w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=300 300w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=768 768w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/shell-startup-actual.png?w=1024 1024w\" \/><\/p>\n<p>It shows which scripts bash decides to run from the top, based on decisions made about the context bash is running in (which decides the colour to follow).<\/p>\n<p>So if you are in a local (non-remote), non-login, interactive shell (eg when you run bash itself from the command line), you are on the \u2018green\u2019 line, and these are the order of files read:<\/p>\n<pre>\/etc\/bash.bashrc\n~\/.bashrc\n[bash runs, then terminates]\n~\/.bash_logout<\/pre>\n<p>This can save you a hell of a lot of time debugging.<\/p>\n<h2>10) getopts (cheapci)<\/h2>\n<p>If you go deep with bash, you might end up writing chunky utilities in it. If you do, then getting to grips with <code>getopts<\/code>\u00a0can pay large dividends.<\/p>\n<p>For fun, I once wrote a <a href=\"https:\/\/github.com\/ianmiell\/cheapci\/blob\/master\/cheapci\">script<\/a> called <code>cheapci<\/code>\u00a0which I used to work like a Jenkins job.<\/p>\n<p>The code <a href=\"https:\/\/github.com\/ianmiell\/cheapci\/blob\/master\/cheapci#L70-L96\">here<\/a>\u00a0implements the reading of the\u00a0<a href=\"https:\/\/github.com\/ianmiell\/cheapci\/blob\/master\/cheapci#L33-L51\">two required, and 14 non-required arguments<\/a>. Better to learn this than to build up a bunch of bespoke code that can get very messy pretty quickly as your utility grows.<\/p>\n<hr \/><p><strong>This is based on some of the contents of my book<\/strong> <a href=\"https:\/\/leanpub.com\/learnbashthehardway\">Learn Bash the Hard Way<\/a>, available at <a href=\"https:\/\/leanpub.com\/learnbashthehardway\">$7<\/a>:<\/p>\n<p><a href=\"https:\/\/leanpub.com\/learnbashthehardway\"><img data-attachment-id=\"4192\" data-permalink=\"https:\/\/zwischenzugs.com\/2018\/01\/06\/ten-things-i-wish-id-known-about-bash\/hero\/\" data-orig-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png\" data-orig-size=\"600,957\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"hero\" data-image-description=\"\" data-medium-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300\" data-large-file=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=525\" class=\"alignnone size-medium wp-image-4192\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300\" alt=\"hero\" width=\"188\" height=\"300\" srcset=\"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=188&amp;h=300 188w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=376&amp;h=600 376w, https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png?w=94&amp;h=150 94w\" \/><\/a><\/p>\n<p>Preview available\u00a0here.<\/p>\n<hr \/><p>I also wrote\u00a0<a href=\"https:\/\/www.manning.com\/books\/docker-in-practice-second-edition?a_aid=zwischenzugs&amp;a_bid=550032fc\">Docker in Practice\u00a0<\/a><\/p>\n<p><a href=\"https:\/\/www.manning.com\/books\/docker-in-practice-second-edition?a_aid=zwischenzugs&amp;a_bid=550032fc\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Get 39% off with the code: 39miell2<\/em><\/strong><\/a><\/p>\n<p><img class=\"progressiveMedia-image js-progressiveMedia-image\" src=\"https:\/\/zwischenzugs.files.wordpress.com\/2017\/03\/5d342-0v9jkbliyi3uzefqq.jpg?w=640\" alt=\"image\" \/><\/p>\n\n\t\t<div class=\"wpcnt\">\n\t\t\t<div class=\"wpa wpmrec\">\n\t\t\t\tAdvertisements\n\t\t\t\t<div class=\"u\">\t\t<div>\n\t\t<div id=\"atatags-208708701-5be316d4775ea\">\n\t\t\t\n\t\t<\/div><\/div>\t\t<div>\n\t\t<div id=\"atatags-208708704-5be316d4775ee\">\n\t\t\t\n\t\t<\/div><\/div><\/div>\n\t\t\t\t\n\t\t\t<\/div>\n\t\t<\/div>","created_at":"2018-11-07T16:46:12+0000","updated_at":"2018-11-07T16:46:12+0000","published_at":"2018-01-06T11:06:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"zwischenzugs.com","preview_picture":"https:\/\/zwischenzugs.files.wordpress.com\/2018\/01\/hero.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":375,"title":"spring, by ioflow","url":"https:\/\/ioflow.bandcamp.com\/","content":"<p id=\"band-name-location\">\n        ioflow\n        San Diego, California\n    <\/p><div class=\"signed-out-artists-bio-text\"><p id=\"bio-text\">experiments. moments. images. play. beautiful things to reflect the joy and peace of the Creator.<\/p><\/div><ol id=\"band-links\"><li><a target=\"_blank\" href=\"https:\/\/soundcloud.com\/ioflow\" itemprop=\"url\">SoundCloud<\/a><\/li>\n            \n                <li><a target=\"_blank\" href=\"https:\/\/instagram.com\/ioflow\" itemprop=\"url\">Instagram<\/a><\/li>\n            \n                <li><a target=\"_blank\" href=\"https:\/\/twitter.com\/ioflow\" itemprop=\"url\">Twitter<\/a><\/li>\n            \n                <li><a target=\"_blank\" href=\"https:\/\/vimeo.com\/ioflow\" itemprop=\"url\">vimeo.com<\/a><\/li>\n            \n                <li><a target=\"_blank\" href=\"https:\/\/www.youtube.com\/channel\/UC8r3gvnXtvyOKaTWLlJlUHw\" itemprop=\"url\">YouTube<\/a><\/li>\n            \n        <\/ol>","created_at":"2018-11-07T16:46:10+0000","updated_at":"2018-11-07T16:46:10+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"ioflow.bandcamp.com","preview_picture":"https:\/\/f4.bcbits.com\/img\/a0642114919_5.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":373,"title":"Storyseeker by Miles","url":"https:\/\/qwertyprophecy.itch.io\/storyseeker","content":"<div class=\"left_col column\"><div class=\"formatted_description user_formatted\"><p>Storyseeker\u00a0is a minimalist narrative experiment driven by your curiosity. Follow the trails of weasels, talk to ghosts, cross the ocean. Find out what happened to this world.<\/p><p>Music by Joel J\u00e4rventausta.<\/p><\/div><\/div><div class=\"right_col column\"><div class=\"video_embed\"><iframe frameborder=\"0\" allowfullscreen=\"allowfullscreen\" src=\"https:\/\/www.youtube.com\/embed\/qQG27d4mdH0\"> <\/iframe><\/div><\/div>","created_at":"2018-11-07T16:46:09+0000","updated_at":"2018-11-07T16:46:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"qwertyprophecy.itch.io","preview_picture":"https:\/\/img.itch.zone\/aW1hZ2UvMjA0MDM1Lzk3Mzc0My5wbmc=\/original\/XKsZFL.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":374,"title":"The 9 Best Portfolio Website Builders Ranked & Reviewed | Nov 18","url":"https:\/\/www.websitebuilderexpert.com\/best-portfolio-website-builders\/","content":"<h3>Showcase your work and stand out from the crowd<\/h3><p><strong>Website builders make it incredibly easy to create your own stunning portfolio website and show off your skills to the world \u2013 saving time and effort you could better spend pursuing your creative passions.<\/strong><\/p><p>Let\u2019s imagine a gallery full of your work and no-one else\u2019s, that was entirely dedicated to showing off your skills and that anyone could visit at any time of time of day. This would give you more exposure, less direct competition, and the chance to show off your work in the best way possible \u2013 that sounds pretty great right?<\/p><p>Now what if you could completely design and build that gallery yourself \u2013 even if you didn\u2019t know the first thing about construction? Even better, no?!<\/p><p><strong>The best website builders for portfolios provide lots of templates specifically designed to help you show off your work in the best way possible.<\/strong><\/p><p>You\u2019re free to edit and amend the demo content as you see fit \u2013 so if you like a certain feature, say a mosaic-style gallery \u2013 then simply replace the images in the frames with your own photos or content!<\/p><p><strong>This not only puts your creative work front and centre for users, but is incredibly easy and time-saving to do.<\/strong><\/p><p>There\u2019s no mucking about building things from scratch, just a laser-focus on making your body of work look absolutely epic to anyone visiting your site.<\/p><p><strong>As templates are pre-designed with portfolio stores in mind (but are still totally editable too), visitors have no trouble finding the content they\u2019re looking for. Navigation has already been optimized to be simple + straight-forward \u2013 whether it\u2019s a gallery or images, links to writing examples, or even a contact page to get in touch with you.<\/strong><\/p><p>You can even browse app markets (most website builders have one in some form or another) for additional display options or extra features to show-off your work and make it stand out in a crowded marketplace \u2013 It\u2019s essentially the same as downloading apps for your smartphone or tablet!<\/p><p>Whether you are an artist, a photographer, a model, or any other kind of creative \u2013 you need a platform to showcase and show-off your catalog of work. Portfolio website builders provide you with wide-ranging selections of eye catching templates that help you stand out from the crowd, lots of gallery and display options to make your work pop, apps to add attention-grabbing features and encourage simple navigation so visitors can quickly see everything you have to offer.<\/p><h3>Get found and gain an audience<\/h3><p><strong>The best portfolio website builders can help boost your online visibility, make it easy to connect your website with social media accounts for cross-platform promotion, and provide marketing tools to help you get the word out to fans + potential clients!<\/strong><\/p><p>For all the criticisms they receive, website builders actually offer pretty good SEO (search engine optimization) tools that can improve your site to help it rank higher in search engine results pages (SERPs).<\/p><p>It\u2019s easy to add \u2018alt tags\u2019 to images and \u2018meta data\u2019 to webpages \u2013 both of which act like a description of your portfolio website that helps search engines sort and categorize it, similar to a blurb on the back of a book.<\/p><p>Having \u2018alt tags\u2019 and \u2018meta data\u2019 is essential in order to rank higher and get found by people online \u2013 but more on that when we take you through the best portfolio website builders below.<\/p><p><strong>You can also add a <a href=\"https:\/\/www.websitebuilderexpert.com\/how-to-start-a-blog\/\">blog<\/a> very easily to your website through most website builders, which can be a great way to show off your knowledge and promote your creative work.<\/strong><\/p><p>This is known as content marketing and can also help you rank higher in SERPS because Google scans the content on your page and picks out the words relevant to what someone is searching for.<\/p><p>For example, if you are a photographer and write a lot of blog posts about the types of film you use when shooting landscapes, then you\u2019ll stand a better chance of ranking highly when someone searches for \u2018best camera film for landscape photography\u2019. At the very least you\u2019ll definitely rank above someone who doesn\u2019t write about this!<\/p><p>Maintaining a blog is an easy \u2013 yet important \u2013 way of creating regular new content for your site and visitors.<\/p><p><strong>There are also marketing tools available to use to further promote your portfolio and spread the word about your creative talents.<\/strong><\/p><p>Most website builders offer some form of email marketing tool, such as Wix\u2019s \u2018Wix Shout\u2019, which are ideal for sending out news updates or for letting your fans know about your next exhibition or show.<\/p><p>You can support these email campaigns with additional marketing apps, such as \u2018Traffic Generator\u2019 that can create Google Adwords campaigns for you and drive traffic to your portfolio \u2013 a great way to be found by your target market.<\/p><p>You could even download an app to connect your social media feeds with your website.<\/p><p><strong>Imagine syncing your portfolio updates with your instagram account for super-simple, cross-platform marketing that can really build you a dedicated audience!<\/strong><\/p><p>Platforms like Wix and Squarespace provide several marketing tools to help you push your portfolio to the next level. Create specialized email marketing campaigns to notify subscribers of updates to your portfolio, integrate with social media and show off your work across multiple channels, or even add a blog to show off your knowledge and expertise. People can\u2019t appreciate your work if they don\u2019t know it exists, so take advantage of the marketing tools provided by website builders to promote yourself to fans and potential clients.<\/p><h3>Generate industry contacts and new business<\/h3><p><strong>Our recommended portfolio website builders are the best way to create a professional-looking website and improve your website\u2019s visibility in search engines so that you make the best impression possible on potential clients.<\/strong><\/p><p>As this will be the first time most prospective clients are seeing your work it\u2019s super important that you make a great first impression \u2013 just like if it was a job interview.<\/p><p>I know that I (and many friends and colleagues, and I\u2019m sure you too!) have always had more success when making a strong first impression and the same is true for your website.<\/p><p>Whether it\u2019s a job interview or your portfolio website, it pays to create an impression that makes an impact. The portfolio website builders we recommend below are ideal for getting yourself noticed in the best way possible.<\/p><p><strong>Having lots of templates to choose from is a big plus point for anyone building a portfolio website because it gives you a lot of options to choose from when deciding on how best to make a first impression.<\/strong><\/p><p><strong>User-friendly editors keep website design simple and uncomplicated<\/strong>, so it\u2019s painless to create a specific contact page or add a contact box so that people can easily submit enquiries or reach out to you.<\/p><p>Remember, people like convenience, so if it\u2019s hard to get in touch with you then you could lose out on potentially lucrative business opportunities!<\/p><p><strong>Just imagine how time-saving it would be to do away with cold emails and cold calls in order to try and drum up business \u2013 Just add a contact form and let people get in touch with you directly!<\/strong><\/p><p>Website builders make it easy to add contact pages and forms to your portfolio website so that prospective clients can contact you directly. This does away with the need for you to send cold emails or make stressful cold calls asking for business and so is a huge time-saving opportunity. Add to this SEO boosting apps and features to help your website rank higher and get found in Search Engine Results Pages (SERPs) and you\u2019ll be on your way to generating lots of new business!<\/p><h3>Save money (by being affordable and scalable)<\/h3><p><strong>Using a website builder is a far more affordable option than hiring a professional designer that might set you back thousands of dollars.<\/strong><\/p><p>Nearly all website builders offer a free plan, which is great if your needs aren\u2019t too demanding.<\/p><p><strong>However, if you want more storage, more templates, or a custom domain (like a personalized website address) then it\u2019s worth considering a paid premium plan.<\/strong><\/p><p>For the price of a Starbucks Grande Caffe Latte each month (around $5-10) you\u2019ll not only be able to build your own portfolio website, but <strong>you won\u2019t even need to worry about things like hosting, storage, and site security as they will all be taken care of!<\/strong><\/p><p><strong>This is a huge time, effort and money saver as you\u2019d otherwise need to spend time researching, comparing, and paying for it out of your own pocket if working with a designer (who only creates the website, typically nothing else).<\/strong><\/p><p><strong>Paid plans also give you more storage space and bandwidth, which are both important in helping your website to grow.<\/strong><\/p><p>Having more storage lets you add more content to your portfolio to keep users engaged, while bandwidth is the name for the data used to power your website when people visit and is provided by your website builder of choice.<\/p><p>Having more bandwidth means you can support more visitors and higher levels of traffic.<\/p><p><strong>It would take years for you to rack up a bill on a website builder that would match the cost of <a href=\"https:\/\/www.websitebuilderexpert.com\/why-you-should-not-hire-a-website-designer\/\">hiring a website designer<\/a>, so to my mind it makes much more sense to remove as much financial stress as possible if your budget is tight.<\/strong><\/p><p>Our recommended portfolio website builders are not only easy to use, they keep things within budget too. Paid monthly premium plans are affordable and work out much cheaper than the on-going costs you\u2019d be stuck with if hiring a costly web designer or developer to create your site from scratch. Paid plans also include hosting, storage and bandwidth \u2013 all of which would prove increasingly expensive if you had to sort them all out separately by yourself.<\/p><h3>Save time and effort (by being easy to pick up and use)<\/h3><p><strong>The drag and drop editors used by website builders makes it incredibly easy for even the biggest technophobe to create stunning, professional-looking website in no time at all<\/strong><\/p><p>Nearly all major <a href=\"https:\/\/www.websitebuilderexpert.com\/diy-website-builder\/\">website builders<\/a> provide an easy-to-use drag and drop editor that pretty much does what it says on the tin! You select an element such as text or image and then you drag it to where you want it on the page and drop it there \u2013 simple.<\/p><p><strong>Drag and drop editors are also known as \u2018What You See Is What You Get\u2019 editors (WYSIWYG), which means that any changes you make in the editor will be reflected on your live website.<\/strong><\/p><p>It\u2019s sort of like when you create a custom birthday card on a site like Moonpig \u2013 the card you receive will look exactly like the one you designed.<\/p><p>Drag and drop website building is incredibly easy and doesn\u2019t require you to know anything about coding at all, which does away with fiddly code tweaking and lets you focus on creating a stunning portfolio website.<\/p><p>This is a very time-saving way of working because you\u2019re essentially editing in real-time, so you know exactly how your portfolio will look to visitors.<\/p><p>But for some readers, time may be at even more of a premium and you might not be able to afford a few hours to dedicate to building a portfolio website. In this case, just work with the pre-designed templates and edit as much or as little as you want, then publish it to the web.<\/p><p><strong>If you really wanted to, you could have a website set up and published to the web within 20 minutes<\/strong> \u2013 saving you a whole lot of time you could dedicate to an unfinished painting or a photoshoot that needs touching up.<\/p><p><strong>For added peace of mind, website builders provide lots of help to keep your building experience running smoothly and ensure you don\u2019t run into trouble.<\/strong><\/p><p>Among the available support options are Wix\u2019s in-editor tooltips that explain what tools do if you hover over the little question marks, it includes Weebly\u2019s live help chat for efficient problem solving, as well as Squarespace\u2019s in-depth tutorial videos and wide-ranging support guides.<\/p><p>It\u2019s important to have support available when you need it as it helps to solve issues quickly, so you can carry on working and don\u2019t waste time waiting around for an answer.<\/p><p>When you can simply drag and drop your photos or artwork wherever you want, it makes designing a unique portfolio incredibly easy. It also means you don\u2019t need to know how to work with HTML, which means you can create your website quicker. And because your portfolio will look like it does in the editor, it means you can fine-tune your design and know exactly how it will look to your visitors. If you really need to save time \u2013 then there\u2019s pre-designed templates for you to edit and publish, simple!<\/p>","created_at":"2018-11-07T16:46:09+0000","updated_at":"2018-11-07T16:46:09+0000","published_at":"2017-11-22T06:57:51+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":11,"domain_name":"www.websitebuilderexpert.com","preview_picture":"https:\/\/www.websitebuilderexpert.com\/wp-content\/uploads\/2017\/11\/best-portfolio-website-builders.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":370,"title":"Using Transducers with Core.async in ClojureScript","url":"https:\/\/blog.venanti.us\/using-transducers-with-core-async-clojurescript\/","content":"<p>I was working on a pretty big post last year when I was sidelined with a <a href=\"https:\/\/www.airbnb.com\/\">new job<\/a> and started writing a book, <a href=\"http:\/\/www.amazon.com\/Learning-ClojureScript-Rafik-Naccache-ebook\/dp\/B016WKHQ1C\/ref=sr_1_1?ie=UTF8&amp;qid=1453276003&amp;sr=8-1&amp;keywords=learning+clojurescript\">Learning ClojureScript<\/a>. Working on the book has given me a lot of opportunities to explore areas of the  <a href=\"https:\/\/github.com\/clojure\/clojurescript\">ClojureScript<\/a> ecosystem that I wouldn't otherwise have explored, and some of the learnings that have resulted have been pretty cool.<\/p><p>Specifically, two of the things I've been learning about lately are <a href=\"https:\/\/github.com\/clojure\/core.async\">core.async<\/a> and <a href=\"http:\/\/clojure.org\/reference\/transducers\">transducers<\/a>. Core.async is a highly popular Clojure(Script) library for embedding CSP-style concurrent program design in libraries and applications; transducers are composable functions useful for performing transformations on data.<\/p><p>There are a number of articles about core.async already floating around on the internet, but transducers appear to remain largely misunderstood. Speaking personally, I've have had quite a bit of trouble wrapping my mind around them, and one of the things that has helped me tighten up my understanding has been to get a better understanding of their specific use cases.<\/p><p>In general, I think the fundamental insight for understanding transducers is to realize that their application is not so different from other higher-order functions, but that they're applied to slightly different (though no less valuable) contexts. Thus far, the most awesome examples I've seen have been ones in which transducers were used in combination with core.async to build data transformations into asynchronous application message queues, so that's what I'll focus on here.<\/p><p>Let's get into it.<\/p><h4 id=\"errorhandlingpatterns\">Error Handling Patterns<\/h4><p>One of the advantages of core.async is that it encourages programs with a clear separation of concerns. Data is produced in one part of your application, and consumed and processed in another.<\/p><p>This separation of concerns has a drawback, however. If we have code that expects to be able to read from a channel, what happens when errors occur on the input side of the channel? We could catch and handle errors on the input side, but there are a number of circumstances in which that might be undesirable. For instance, what if there are many possible places in the codebase in which we put value into this channel? It is not unreasonable to imagine that we might want a central pipeline for handling errors.<\/p><p>There are a few different ways of going about this. One possibility would be to create a separate channel for error messages and to have an independent listener for that channel (a pattern discussed on the ClojureScript mailing list <a href=\"https:\/\/groups.google.com\/forum\/#!topic\/clojurescript\/9EP2w4OSsyo\">here<\/a>).<\/p><p>Let\u2019s assume we have an application where we're regularly polling the kitten factory to make sure production is running smoothly. If things are running smoothly, we'll create a report, and if they go badly, we'll log that to our exception tracking service.<\/p><p>Side note: <em>What even happens at the kitten factory? Is it a factory that makes kittens, or do kittens run the factory? If they\u2019re running the factory, what are they making \u2013 balls of yarn? I have trouble imagining this is a very efficient factory.<\/em><\/p><p>Such code might look like the following:<\/p><pre class=\"language-clojure\">(ns app.factory.kitten\n  \"Logic for handling messages from the kitten factory.\"\n  (:require [ajax.core :refer [GET]]\n            [app.exceptions :as exc]\n            [app.report :as report]\n            [cljs.core.async :as async])\n  (:require-macros\n    [cljs.core.async.macros :as async-macros]))\n;; define our channels\n(def channel (async\/chan 5))\n(def error-channel (async\/chan 5))\n;; simple helper\n(defn enqueue-val\n  \"Enqueue a new value into our channel.\"\n  [c v]\n  (async-macros\/go (async\/&gt;! c v)))\n(defn poll-kitten-factory\n  \"Poll the kitten factory endpoint for its current status.\"\n  []\n  (GET \"\/kitten-factory\"\n       {:handler (fn [r] (enqueue-val channel r))\n        :error-handler (fn [r] (enqueue-val error-channel r))}))\n(defn listen\n  []\n  \"Listen for the latest message from the kitten factory channels.\"\n  (async-macros\/go\n    (while true\n      (let [[v ch] (async\/alts! [channel error-channel])]\n        (case ch\n          channel (report\/send-success-report-to-cat-hq v)\n          error-channel (exc\/report-exception v))))))\n<\/pre><p>An alternative pattern that we could embrace would be one in which we passed the errors themselves into the channel and checked to see if a message was an error when consuming. This is a pattern that David Nolen has written about previously on his blog <a href=\"http:\/\/swannodette.github.io\/2013\/08\/31\/asynchronous-error-handling\/\">here<\/a>.<\/p><p>If you read his post, he refers to a <code>&lt;?<\/code> macro, the source code for which is available on his blog's GitHub repo. The <code>&lt;?<\/code> macro just checks to see if the value that\u2019s been pulled off of the channel is an error, and if it is, throws. Without this, we\u2019d just be passing an error around by value and the compiler would have no reason to actually throw the error (thereby preventing us from catching it, as well).<\/p><p>We'll copy the relevant code from his blog's GitHub repo into our app, modulo a few changes:<\/p><pre class=\"language-clojure\">(ns app.helpers)\n(defn throw-err\n  [x]\n  (if (instance? js\/Error x)\n    (throw x)\n    x))\n<\/pre><pre class=\"language-clojure\">(ns app.macros)\n(defmacro &lt;?\n  \"Actively throw an exception if something goes wrong when waiting on a channel message.\"\n  [expr]\n  `(app.helpers\/throw-err (cljs.core.async\/&lt;! ~expr)))\n<\/pre><p>Now that we have our helpers in place, let's change a few parts of our namespace.<\/p><pre class=\"language-clojure\">;; we only need the one channel now, not two.\n(def channel (async\/chan 5))\n;; we've changed the error-handler to cast the response\n;; to a js\/Error before putting it on the channel\n(defn poll-kitten-factory\n  \"Poll the kitten factory endpoint for its current status.\"\n  []\n  (GET \"\/kitten-factory\"\n       {:handler (fn [r] (enqueue-val channel r))\n        :error-handler (fn [r] (enqueue-val channel (js\/Error. r)))}))\n(defn listen\n  []\n  \"Listen for the latest message from the kitten factory channel.\n  If message is an error, throw and catch.\"\n  (async-macros\/go\n    (while true\n      (try\n        ;; note the use of `&lt;?`\n        (let [v (&lt;? channel)]\n          (report\/send-success-report-to-cat-hq v))\n        (catch js\/Error e\n          (exc\/report-exception e))))))\n<\/pre><p>This is a perfectly good solution, but do we really need that <code>&lt;?<\/code> macro? This is a great chance to use a pure function (a transducer!) instead. Transducers can be passed as an optional argument to the <code>chan<\/code> constructor function, and they'll perform the given transformation on any values passed through the channel.<\/p><p>This means that we have an opportunity to write a transducer for our channel that just maps our helpful <code>throw-err<\/code> function over values being taken from the channel. Let's see what that looks like:<\/p><pre class=\"language-clojure\">;; Initialize the channel with a transducer\n(def channel (async\/chan 5 (map h\/throw-err)))\n;; note: no change to poll-kitten factory from our \n;; last example; we're still casting errors to \n;; js\/Error before enqueuing\n(defn listen\n  []\n  \"Listen for the latest message from the kitten factory channel.\"\n  (async-macros\/go\n    (while true\n      (try\n        ;; back to normal `&lt;!` macro\n        (let [v (async\/&lt;! channel)]\n          (report\/send-success-report-to-cat-hq v))\n        (catch js\/Error e\n          (exc\/report-exception e)))))))\n<\/pre><p>Wow - simple!<\/p><p>Let's take a look at another example - this one courtesy of my friend Allen (<a href=\"https:\/\/www.twitter.com\/arohner\">@arohner<\/a>), with a few modifications.<\/p><h4 id=\"batchingdata\">Batching Data<\/h4><p>Let's say we've got a similar setup to earlier, but instead we want to batch up at least 10 events before we send a report back to cat hq. If we were going to design this without transducers, we'd probably do something like the following:<\/p><pre class=\"language-clojure\">;; two channels. `single` is being populated by\n;; polling logic, `batch` is being populated here.\n(def single (async\/chan 100))\n;;  taking from batch returns:\n;; [kitten kitten kitten kitten...]\n(def batch (async\/chan 50))\n(defn batch-kittens\n  \"Batch up kittens into vectors of length 10 and put\n   them into the batch channel.\"\n  []\n  (let [batch-size 10\n        state (atom [])]\n    (async\/go\n      (while true\n        (let [v (async\/&lt;! single)]\n          (swap! state conj v)\n          (when-let [vs (first (partition batch-size @state))]\n            (reset! state [])\n            (async\/&gt;! batch vs)))))))\n;; let's not worry about error handling for now\n(defn listen\n  \"Listen for batch jobs and send them.\"\n  []\n  (async\/go\n    (while true\n      (let [v (async\/&lt;! batch)]\n        (report\/send-batch-report-to-cat-hq v)))))\n<\/pre><p>This could probably be cleaned up some more, but at the very least it feels like we're doing more work than we should be. By contrast, if we were to use a transducer to batch these up, it would look like the following:<\/p><pre class=\"language-clojure\">;; we no longer need two channels. `channel`\n;; is populated upstream and we pull from it here\n(def channel (async\/chan 50 (partition-all 10)))\n;; we don't need a batch-kittens function anymore either.\n(defn listen\n  \"Listen for batch jobs and send them.\"\n  []\n  (async\/go\n    (while true\n       (let [v (async\/&lt;! channel)]\n          (report\/send-batch-report-to-cat-hq v))))))\n<\/pre><p>This is unquestionably much cleaner than the version without transducers. As Allen put it: <em>\"You put one kitten in, and take ten out. You can't explain that.\"<\/em><\/p><h4 id=\"conclusion\">Conclusion<\/h4><p>The thing that stands out to me about transducers is that they empower us in a way that is simple. As a language feature, it's easy to imagine an implementation where performing data transformations on asynchronous streams could have been over-engineered and a nightmare to write code for. Instead, Clojure(Script)'s transducers are easy to read, write and reason about.<\/p><p>There are plenty of other applications for transducers that I haven't talked about in this post, but for me, the core.async tie-in is where their value was first made clear. If you think transducers are n.b.d. and can explain why I can't find that image of Rich Hickey pasted onto the guy from Ancient Aliens on Google, feel free to let me know.<\/p><p>~ V<\/p><p><em>Discuss this post on <a href=\"https:\/\/news.ycombinator.com\/item?id=10946849\">Hacker News<\/a> or on Twitter (<a href=\"https:\/\/www.twitter.com\/venantius\">@venantius<\/a>)<\/em><\/p><p>Thanks to Keith Ballinger (<a href=\"https:\/\/www.twitter.com\/keithba\">@keithba<\/a>), Bill Cauchois (<a href=\"https:\/\/www.twitter.com\/wcauchois\">@wcauchois<\/a>), Katherine Geenberg (<a href=\"https:\/\/www.twitter.com\/mywolfymaw\">@MyWolfyMaw<\/a>), and Allen Rohner (<a href=\"https:\/\/www.twitter.com\/arohner\">@arohner<\/a>) for not only reading drafts of this post, but for adding to it.<\/p><p>Finally, a shameless plug: if this is the sort of thing that interests you and you're interested in learning more, <a href=\"http:\/\/www.amazon.com\/Learning-ClojureScript-Rafik-Naccache-ebook\/dp\/B016WKHQ1C\/ref=sr_1_1?ie=UTF8&amp;qid=1453276003&amp;sr=8-1&amp;keywords=learning+clojurescript\">pre-order<\/a> a copy of the book!<\/p>","created_at":"2018-11-07T16:46:08+0000","updated_at":"2018-11-07T16:46:08+0000","published_at":"2016-01-21T17:41:49+0000","published_by":["Venantius"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"blog.venanti.us","preview_picture":"https:\/\/blog.venanti.us\/content\/images\/2016\/01\/photo-1444362408440-274ecb6fc730.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":371,"title":"clojure\/clojurescript","url":"https:\/\/github.com\/clojure\/clojurescript\/wiki\/Advanced-Compilation","content":"<blockquote>\n<p><strong>DEPRECATION NOTICE<\/strong>: Please do not edit this wiki. Instead submit pull requests to\n<a href=\"https:\/\/github.com\/clojure\/clojurescript-site\">https:\/\/github.com\/clojure\/clojurescript-site<\/a><\/p>\n<p>The up to date version of this page can be found at <a href=\"https:\/\/clojurescript.org\/reference\/advanced-compilation\" rel=\"nofollow\">https:\/\/clojurescript.org\/reference\/advanced-compilation<\/a><\/p>\n<\/blockquote>\n<p>If you are targeting traditional JavaScript clients (web browsers) it's important to think about advanced compilation from the very beginning. Otherwise you will inevitably find yourself going through issues that could have been easily avoided with a little bit of up front preparation. In general do not wait to try an advanced build, you should always periodically generate a production build to catch issues sooner.<\/p>\n<h2>\n<a id=\"user-content-avoid-foreign-libraries\" class=\"anchor\" href=\"#avoid-foreign-libraries\" aria-hidden=\"true\"><\/a>Avoid Foreign Libraries<\/h2>\n<p>It's best to simply avoid foreign libraries if a solution exists either in Google Closure Library or in an existing ClojureScript library. Foreign libraries must supply externs for advanced compilation to work consistently.<\/p>\n<p>Of course, in some cases foreign libraries cannot be avoided.<\/p>\n<h2>\n<a id=\"user-content-using-foreign-libraries\" class=\"anchor\" href=\"#using-foreign-libraries\" aria-hidden=\"true\"><\/a>Using Foreign Libraries<\/h2>\n<p>If you must use a foreign library use a curated one like those provided by <a href=\"http:\/\/cljsjs.github.io\" rel=\"nofollow\">CLJSJS<\/a>. These come packaged with externs so that you do not have to supply them yourself.<\/p>\n<p>Occasionally you may find yourself needing a foreign library which has not been prepackaged.<\/p>\n<h2>\n<a id=\"user-content-providing-externs\" class=\"anchor\" href=\"#providing-externs\" aria-hidden=\"true\"><\/a>Providing externs<\/h2>\n<p>When working with a foreign library which does not supply externs take the time to write an externs file for the API you intend to use. Externs files are surprisingly simple to provide. For example if the foreign library has some property <code>Foo.bar<\/code> that you wish to access your externs file should have the following entry:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Foo.bar;<\/pre><\/div>\n<p>If the foreign library has some method <code>Foo.baz<\/code> that you wish to invoke your externs file should have the following entry:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Foo.baz = function() {};<\/pre><\/div>\n<p>Sometimes there will not be a top level API but rather some method naming convention, that is, an ad-hoc Interface \/ Protocol. In these cases define your externs using Object:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Object.foo = function() {};\nObject.bar = function() {};<\/pre><\/div>\n<p>Of course sometimes you will miss an extern entry and the the production file will produce a cryptic error. Thanks to a couple of Closure compiler options these issues are no longer difficult to debug.<\/p>\n<h2>\n<a id=\"user-content-fixing-advanced-compilation-issues\" class=\"anchor\" href=\"#fixing-advanced-compilation-issues\" aria-hidden=\"true\"><\/a>Fixing Advanced Compilation Issues<\/h2>\n<p>Change your production build to use two additional options <code>:pseudo-names true<\/code> and <code>:pretty-print true<\/code>. Now your error will show a name that corresponds to the name in the original source. Add an externs entry for this missed case.<\/p>\n<p>For more information about the specifics of <code>:foreign-libs<\/code> compiler option syntax consult <a class=\"internal present\" href=\"https:\/\/github.com\/clojure\/clojurescript\/wiki\/Dependencies\">Dependencies<\/a>.<\/p>","created_at":"2018-11-07T16:46:08+0000","updated_at":"2018-11-07T16:46:08+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/317875?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":372,"title":"clojure\/clojurescript","url":"https:\/\/github.com\/clojure\/clojurescript\/wiki\/Advanced-Compilation","content":"<blockquote>\n<p><strong>DEPRECATION NOTICE<\/strong>: Please do not edit this wiki. Instead submit pull requests to\n<a href=\"https:\/\/github.com\/clojure\/clojurescript-site\">https:\/\/github.com\/clojure\/clojurescript-site<\/a><\/p>\n<p>The up to date version of this page can be found at <a href=\"https:\/\/clojurescript.org\/reference\/advanced-compilation\" rel=\"nofollow\">https:\/\/clojurescript.org\/reference\/advanced-compilation<\/a><\/p>\n<\/blockquote>\n<p>If you are targeting traditional JavaScript clients (web browsers) it's important to think about advanced compilation from the very beginning. Otherwise you will inevitably find yourself going through issues that could have been easily avoided with a little bit of up front preparation. In general do not wait to try an advanced build, you should always periodically generate a production build to catch issues sooner.<\/p>\n<h2>\n<a class=\"anchor\" href=\"#avoid-foreign-libraries\" aria-hidden=\"true\"><\/a>Avoid Foreign Libraries<\/h2>\n<p>It's best to simply avoid foreign libraries if a solution exists either in Google Closure Library or in an existing ClojureScript library. Foreign libraries must supply externs for advanced compilation to work consistently.<\/p>\n<p>Of course, in some cases foreign libraries cannot be avoided.<\/p>\n<h2>\n<a class=\"anchor\" href=\"#using-foreign-libraries\" aria-hidden=\"true\"><\/a>Using Foreign Libraries<\/h2>\n<p>If you must use a foreign library use a curated one like those provided by <a href=\"http:\/\/cljsjs.github.io\" rel=\"nofollow\">CLJSJS<\/a>. These come packaged with externs so that you do not have to supply them yourself.<\/p>\n<p>Occasionally you may find yourself needing a foreign library which has not been prepackaged.<\/p>\n<h2>\n<a class=\"anchor\" href=\"#providing-externs\" aria-hidden=\"true\"><\/a>Providing externs<\/h2>\n<p>When working with a foreign library which does not supply externs take the time to write an externs file for the API you intend to use. Externs files are surprisingly simple to provide. For example if the foreign library has some property <code>Foo.bar<\/code> that you wish to access your externs file should have the following entry:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Foo.bar;<\/pre><\/div>\n<p>If the foreign library has some method <code>Foo.baz<\/code> that you wish to invoke your externs file should have the following entry:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Foo.baz = function() {};<\/pre><\/div>\n<p>Sometimes there will not be a top level API but rather some method naming convention, that is, an ad-hoc Interface \/ Protocol. In these cases define your externs using Object:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>Object.foo = function() {};\nObject.bar = function() {};<\/pre><\/div>\n<p>Of course sometimes you will miss an extern entry and the the production file will produce a cryptic error. Thanks to a couple of Closure compiler options these issues are no longer difficult to debug.<\/p>\n<h2>\n<a class=\"anchor\" href=\"#fixing-advanced-compilation-issues\" aria-hidden=\"true\"><\/a>Fixing Advanced Compilation Issues<\/h2>\n<p>Change your production build to use two additional options <code>:pseudo-names true<\/code> and <code>:pretty-print true<\/code>. Now your error will show a name that corresponds to the name in the original source. Add an externs entry for this missed case.<\/p>\n<p>For more information about the specifics of <code>:foreign-libs<\/code> compiler option syntax consult <a class=\"internal present\" href=\"https:\/\/github.com\/clojure\/clojurescript\/wiki\/Dependencies\">Dependencies<\/a>.<\/p>","created_at":"2018-11-07T16:46:08+0000","updated_at":"2018-11-07T16:46:08+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/317875?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":369,"title":"Using a Yubikey for GPG and SSH","url":"https:\/\/0day.work\/using-a-yubikey-for-gpg-and-ssh\/","content":"<header class=\"col-xs-12\">&#13;\n&#13;\n&#13;\n        <\/header><section class=\"col-xs-12\"><p>It's been a long time since my last blogpost, but I'm back with a post about how to use your Yubikey 4 for GPG and SSH keys.<\/p><p>The rather small <a href=\"http:\/\/yubico.com\/\">Yubikeys are sold by Yubico<\/a> and I obtained two as part of a student offer last year. You can find a detailed <a href=\"https:\/\/www.yubico.com\/products\/yubikey-hardware\/yubikey4\/\">product description<\/a> on their website.<\/p><p><img src=\"https:\/\/0day.work\/content\/images\/2018\/01\/yubikey.jpeg\" alt=\"yubikey\" \/><\/p><p>My goal was to use them as two-factor devices, because I do not like the mobile- or tan-based two-factor authentication approach. The reasoning behind this is, that while travelling through certain countries, receiving text messages can cost up to multiple euros and one does not travel with his tan list...<\/p><p>A small piece of hardware that I can attach to my key ring should solve the problems in my opinion :) However, it turned out, that a Yubikey can also be used as a GPG smartcard, which in turn can be used to authenticate against SSH servers. Fancy! ;)<\/p><p>There are several good tutorials that helped with my setup and you should read them and\/or google a bit. Especially <a href=\"https:\/\/ocramius.github.io\/blog\/yubikey-for-ssh-gpg-git-and-local-login\/\">this one<\/a> has some invaluable comments and hints, that you shouldn't miss! Simply because this won't be a detailed step-by-step guide!<\/p><p>A Yubikey can act as a GPG smartcard allowing us to safely store our private GPG keys on it. I won't go into detail on how to create GPG keys, but I will assume that you have <a href=\"https:\/\/incenp.org\/notes\/2015\/using-an-offline-gnupg-master-key.html\">a masterkey and three subkeys<\/a>:<\/p><ul><li>One for signing [S] (e.g. Emails)<\/li>\n<li>One for encryption [E] (e.g. Emails)<\/li>\n<li>One for authentication [A] (e.g. SSH)<\/li>\n<\/ul><p>Plug your Yubikey into a free USB port and make sure that <code>gpg --card-status<\/code> shows it:<\/p><pre class=\"language-bash\">$&gt; gpg --card-status\n[...]\nVersion ..........: 2.1\nManufacturer .....: Yubico\n[...]\n<\/pre><p>Edit your GPG key you want to store on the Yubikey with <code>gpg --edit-key AAAABBBB<\/code>. Use the <code>keytocard<\/code> command to send the private keys to the Yubikey. <em><strong>Note that this step is destructive and it will delete the private keys from your computer! Make sure to have a backup of your master keys!<\/strong><\/em><\/p><p>If that step succeeded, the <code>gpg --card-status<\/code> command should show the different subkeys in the overview:<\/p><pre class=\"language-bash\">$&gt; gpg --card-status | grep key\nURL of public key : [...]\nSignature key ....: [...] XXXX YYYY\nEncryption key....: [...] ZZZZ VVVV\nAuthentication key: [...] AAAA BBBB\nGeneral key info..: sub  rsa4096\/QQQQQQ &lt;foobar@domain.tld&gt;\n<\/pre><p>To check if everything works as intended, we can create an encrypted message and decrypt it:<\/p><pre class=\"language-bash\">$&gt; echo \"Hello World\" &gt; \/tmp\/message.txt  \n$&gt; gpg --encrypt  --output \/tmp\/message.txt.enc -r contact@0day.work \/tmp\/message.txt\n$&gt; gpg --decrypt \/tmp\/message.txt.enc\ngpg: verschl\u00fcsselt mit 4096-Bit RSA Schl\u00fcssel, ID XXXXXX, erzeugt 2017-12-07\n      \"Sebastian Neef &lt;contact@0day.work&gt;\"\nHello World\n<\/pre><p>Do not forget to set the PINs for the admin\/non-admin commands of <code>gpg --card-edit<\/code>!<\/p><p>The SSH key is derived from the private key on your Yubikey. If you haven't followed the steps of the previous section, you should definitely do so.<\/p><p>First, make sure that the Yubikey is plugged into an USB port and it has an authentication key stored.<\/p><pre class=\"language-bash\">$&gt; gpg --card-status\nReader ...........: 1050:0407:0006481979:0\nApplication ID ...: D2760001XXXX....\nVersion ..........: 2.1\nManufacturer .....: Yubico\nSerial number ....: XXXX\n[...]\nAuthentication key: [....] AAAA BBBB\n      created ....: 2017-12-07 14:53:26\n[...]\n<\/pre><p>You can then export your SSH public key using the following command:<\/p><pre class=\"language-bash\">$&gt; gpg --export-ssh-key 0xAAAABBBB\nssh-rsa AAAAB3Nza[...]openpgp:0xAAAABBBB\n<\/pre><p>Upload this public key to  your servers or wherever you need to authenticate with the SSH key.<\/p><p>To use the key, you have to configure the GPG agent to enable SSH support and act as a SSH agent:<\/p><pre class=\"language-bash\">echo 'enable-ssh-support' &gt;&gt; ~\/.gnupg\/gpg-agent.conf       \n<\/pre><p>Stop and restart all runnig GPG\/SSH agents:<\/p><pre class=\"language-bash\">killall gpg-agent\nkillall ssh-agent\ngpg-agent --daemon\n<\/pre><p>Furthermore, we need to unset the <code>SSH_AUTH_SOCK<\/code>, so add the following lines into your <code>~\/.bashrc<\/code>:<\/p><pre class=\"language-plain\">export SSH_AGENT_PID=\"\"\nexport SSH_AUTH_SOCK=$(gpgconf --list-dirs agent-ssh-socket)\n<\/pre><p>With your Yubikey still plugged in, you should see your SSH key when running the ssh-add command:<\/p><pre class=\"language-bash\">$&gt;  ssh-add -l\n4096 SHA256:XXXX cardno:0006064XXXX (RSA)\n<\/pre><p>That's all! SSH will now use the SSH key from your Yubikey, so don't forget to plug it in, before running <code>ssh server<\/code>.<\/p><p>-=-<\/p><\/section>","created_at":"2018-11-07T16:46:07+0000","updated_at":"2018-11-07T16:46:07+0000","published_at":"2018-01-13T22:04:54+0000","published_by":["Sebastian Neef"],"annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"0day.work","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":367,"title":"Webhooks - Developers - Dropbox","url":"https:\/\/www.dropbox.com\/developers\/reference\/webhooks","content":"<p>Webhooks are a way for web apps to get real-time notifications when users' files change in Dropbox.<\/p><p>Once you register a URI to receive webhooks, Dropbox will send an HTTP request to that URI every time there's a change for any of your app's registered users.<\/p><p>In this tutorial, we'll walk through a very simple example app that uses webhooks to convert Markdown files to HTML. Thanks to webhooks, the app will convert files as soon as they're added to Dropbox. The sample app is running at <a href=\"http:\/\/mdwebhook.herokuapp.com\/\">mdwebhook.herokuapp.com<\/a>, and the full source code is available <a href=\"https:\/\/github.com\/dropbox\/mdwebhook\">on GitHub<\/a>.<\/p><h2 class=\"toc-el\" id=\"verification\" title=\"Verification request\">The verification request<\/h2><p>To set up a new webhook, find your app in the <a href=\"https:\/\/www.dropbox.com\/developers\/apps\">App Console<\/a>, and add the full URI for your webhook (e.g. <code>https:\/\/www.example.com\/dropbox-webhook<\/code>) in the \"Webhooks\" section. Note that the URI needs to be one publicly accessible over the internet. For example, <code>127.0.0.1<\/code> and <code>localhost<\/code> URIs will not work, since Dropbox's servers will not be able to contact your local computer.<\/p><p>Once you enter your webhook URI, an initial \"<a href=\"https:\/\/www.dropbox.com\/developers\/webhooks\/docs#verification\">verification request<\/a>\" will be made to that URI. This verification is an HTTP <code>GET<\/code> request with a query parameter called <code>challenge<\/code>. Your app needs to respond by echoing back that <code>challenge<\/code> parameter. In order to avoid introducing a reflected XSS vulnerability, set the following headers in your response to the verification request:<\/p><pre class=\"dbprettyprint literal-block prettyprint lang-None\">'Content-Type: text\/plain'\n'X-Content-Type-Options: nosniff'\n<\/pre><p>The purpose of this verification request is to demonstrate that your app really does want to receive notifications at that URI. If you accidentally entered the wrong URI (or if someone maliciously entered your server as their webhook), your app would fail to respond correctly to the challenge request, and Dropbox would not send any notifications to that URI.<\/p><p>The following is Python code (using the Flask framework) that responds correctly to a verification request:<\/p><pre class=\"dbprettyprint literal-block prettyprint lang-None\">@app.route('\/webhook', methods=['GET'])\ndef verify():\n    '''Respond to the webhook verification (GET request) by echoing back the challenge parameter.'''\n    resp = Response(request.args.get('challenge'))\n    resp.headers['Content-Type'] = 'text\/plain'\n    resp.headers['X-Content-Type-Options'] = 'nosniff'\n    return resp\n<\/pre><p>If your app responds correctly to the challenge request, Dropbox will start sending notifications to your webhook URI every time one of your users adds, removes, or changes a file. If your app fails to respond correctly, you'll see an error message in the App Console telling you about the problem.<\/p><h2 class=\"toc-el\" id=\"notifications\" title=\"Notifications\">Receiving notifications<\/h2><p>Once your webhook URI is added, your app will start receiving \"<a href=\"https:\/\/www.dropbox.com\/developers\/webhooks\/docs#notification\">notification requests<\/a>\" every time a user's files change. A notification request is an HTTP <code>POST<\/code> request with a JSON body. The JSON has the following format:<\/p><pre class=\"dbprettyprint literal-block prettyprint lang-None\">{\n    \"list_folder\": {\n        \"accounts\": [\n            \"dbid:AAH4f99T0taONIb-OurWxbNQ6ywGRopQngc\",\n            ...\n        ]\n    },\n    \"delta\": {\n        \"users\": [\n            12345678,\n            23456789,\n            ...\n        ]\n    }\n}\n<\/pre><p>Note that the payload of the notification request does not include the actual file changes. It only informs your app of which users have changes. You will typically want to call <a href=\"https:\/\/www.dropbox.com\/developers\/documentation\/http#documentation-files-list_folder-continue\">\/files\/list_folder\/continue<\/a> to get the latest changes for each user in the notification, keeping track of the latest cursor for each user as you go.<\/p><p>Every notification request will include a header called <code>X-Dropbox-Signature<\/code> that includes an HMAC-SHA256 signature of the request body, using your app secret as the signing key. This lets your app verify that the notification really came from Dropbox. It's a good idea (but not required) that you check the validity of the signature before processing the notification.<\/p><p>Below is Python code (again using Flask) that validates a notification request and then calls a function <code>process_user<\/code> once for each user ID in the payload:<\/p><pre class=\"dbprettyprint literal-block prettyprint lang-None\">from hashlib import sha256\nimport hmac\nimport threading\n@app.route('\/webhook', methods=['POST'])\ndef webhook():\n    '''Receive a list of changed user IDs from Dropbox and process each.'''\n    # Make sure this is a valid request from Dropbox\n    signature = request.headers.get('X-Dropbox-Signature')\n    if not hmac.compare_digest(signature, hmac.new(APP_SECRET, request.data, sha256).hexdigest()):\n        abort(403)\n    for account in json.loads(request.data)['list_folder']['accounts']:\n        # We need to respond quickly to the webhook request, so we do the\n        # actual work in a separate thread. For more robustness, it's a\n        # good idea to add the work to a reliable queue and process the queue\n        # in a worker process.\n        threading.Thread(target=process_user, args=(account,)).start()\n    return ''\n<\/pre><p>Typically, the code you run in response to a notification will make a call to <a href=\"https:\/\/www.dropbox.com\/developers\/documentation\/http#documentation-files-list_folder-continue\">\/files\/list_folder\/continue<\/a> to get the latest changes for a user. In our sample Markdown-to-HTML converter, we're keeping track of each user's OAuth access token and their latest cursor in <a href=\"http:\/\/redis.io\">Redis<\/a> (a key-value store). This is the implementation of <code>process_user<\/code> for our Markdown-to-HTML converter sample app:<\/p><pre class=\"dbprettyprint literal-block prettyprint lang-None\">def process_user(account):\n    '''Call \/files\/list_folder for the given user ID and process any changes.'''\n    # OAuth token for the user\n    token = redis_client.hget('tokens', account)\n    # cursor for the user (None the first time)\n    cursor = redis_client.hget('cursors', account)\n    dbx = Dropbox(token)\n    has_more = True\n    while has_more:\n        if cursor is None:\n            result = dbx.files_list_folder(path='')\n        else:\n            result = dbx.files_list_folder_continue(cursor)\n        for entry in result.entries:\n            # Ignore deleted files, folders, and non-markdown files\n            if (isinstance(entry, DeletedMetadata) or\n                isinstance(entry, FolderMetadata) or\n                not entry.path_lower.endswith('.md')):\n                continue\n            # Convert to Markdown and store as &lt;basename&gt;.html\n            _, resp = dbx.files_download(entry.path_lower)\n            html = markdown(resp.content)\n            dbx.files_upload(html, entry.path_lower[:-3] + '.html', mode=WriteMode('overwrite'))\n        # Update cursor\n        cursor = result.cursor\n        redis_client.hset('cursors', account, cursor)\n        # Repeat only if there's more to do\n        has_more = result.has_more\n<\/pre><h2 class=\"toc-el\" id=\"best-practices\" title=\"Best practices\">Best practices<\/h2><h3>Always respond to webhooks quickly<\/h3><p>Your app only has ten seconds to respond to webhook requests. For the verification request, this is never really an issue, since your app doesn't need to do any real work to respond. For notification requests, however, your app will usually do something that takes time in response to the request. For example, an app processing file changes will call <a href=\"https:\/\/www.dropbox.com\/developers\/documentation\/http#documentation-files-list_folder-continue\">\/files\/list_folder\/continue<\/a> and then process the changed files. (In our Markdown example, we needed to download each Markdown file, convert it to HTML, and then upload the result.) It's important to keep in mind that list_folder payloads can sometimes be <em>very<\/em> large and require multiple round-trips to the Dropbox API, particularly when a new user first links your app and has a lot of files.<\/p><p>To make sure you can always respond within ten seconds, you should always do your work on a separate thread (as in the simple example above) or asynchronously using a queue.<\/p><h3>Manage concurrency<\/h3><p>When a user makes a number of changes in rapid succession, your app is likely to receive multiple notifications for the same user at roughly the same time. If you're not careful about how you manage concurrency, your app can end up processing the same changes for the same user more than once.<\/p><p>For some applications, this is not a serious issue. In our Markdown-to-HTML conversion app, if the same changes are processed more than once, a file just ends up getting overwritten with the same content, so no harm is done. Work that can be repeated without changing the outcome is called <strong>idempotent<\/strong>. If your app's actions are always idempotent, you don't need to worry much about concurrency.<\/p><p>Unfortunately, not every app can be made idempotent easily. For example, suppose you have an app that sends email every time a certain file is changed. To avoid sending duplicate emails, you need to make sure that your app never processes the same user on multiple threads\/processes at the same time. The simplest solution is to use <strong>leases<\/strong>. When a thread or process starts processing a certain user, it will first acquire a lease on that user, giving it exclusive access.<\/p>","created_at":"2018-11-07T16:46:01+0000","updated_at":"2018-11-07T16:46:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"www.dropbox.com","preview_picture":"https:\/\/cfl.dropboxstatic.com\/static\/images\/logo_catalog\/glyph_m1@2x-vflA6lTFZ.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":368,"title":"Goal-Oriented Action Planning (GOAP)","url":"http:\/\/alumni.media.mit.edu\/~jorkin\/goap.html","content":"<table cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td valign=\"top\">\n      <img border=\"0\" src=\"http:\/\/alumni.media.mit.edu\/~jorkin\/medialab.gif\" alt=\"image\" \/><br \/><b><span>J\u00a0e\u00a0f\u00a0f\u00a0\u00a0\u00a0 O\u00a0r\u00a0k\u00a0i\u00a0n<\/span><\/b><br \/><a href=\"http:\/\/www.media.mit.edu\">MIT Media Lab<\/a> | <a href=\"http:\/\/www.media.mit.edu\/cogmac\">Cognitive Machines<\/a><br \/><a href=\"mailto:jorkin@alum.mit.edu\">jorkin@alum.mit.edu<\/a><br \/><a href=\"http:\/\/www.jorkin.com\">http:\/\/www.jorkin.com<\/a><\/td>\n      <td align=\"right\" style=\"width: 78%;\">\n      <img border=\"0\" src=\"http:\/\/alumni.media.mit.edu\/~jorkin\/jorkin.jpg\" alt=\"image\" \/><br \/><\/td>\n      <\/tr><\/table>\n\n<span>\nGoal-Oriented Action Planning (GOAP)\n<\/span>\n   \n\n   <table cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n         \nGoal-Oriented Action Planning (aka GOAP, rhymes with soap) refers to a simplfied STRIPS-like planning architecture specifically designed for real-time control of autonomous character behavior in games.  \nI originally implemented GOAP for <a href=\"http:\/\/en.wikipedia.org\/wiki\/F.E.A.R.\">F.E.A.R.<\/a> while working at <a href=\"http:\/\/www.lith.com\">Monolith Productions<\/a>. \nThis A.I. architecture simultaneously powered Monolith's <a href=\"http:\/\/www.gamerankings.com\/htmlpages2\/926309.asp?q=condemned\">Condemned: Criminal Origins<\/a>. \n(Brian Legge was responsible for the A.I. in Condemned). \nMy GOAP implementation was inspired by conversations within the A.I. Interface Standards Committee's (<a href=\"http:\/\/www.igda.org\/ai\/\">AIISC<\/a>) GOAP Working Group, as well as ideas from the Synthetic Characters Group's <a href=\"http:\/\/characters.media.mit.edu\/Papers\/gdc01.pdf\">C4<\/a> agent architecture at the <a href=\"http:\/\/media.mit.edu\">MIT Media Lab<\/a>, and Nils Nilsson's description of STRIPS planning in his <a href=\"http:\/\/www.amazon.com\/gp\/product\/1558604677?ie=UTF8&amp;tag=jefork-20&amp;linkCode=as2&amp;camp=1789&amp;creative=9325&amp;creativeASIN=1558604677\">AI book<\/a>.\n            \n         <\/td>\n      <\/tr><tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n            <\/td><\/tr>             \n                        \n                                           \n                                    \n            \n         \n      <tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n            <\/td><\/tr>           \n                      \n                    \n                                 \n                       \n                        \n                          \n                      \n                             \n                 \n            \n         \n      <tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n\t    <\/td><\/tr>        \n\t                 \n\t                  \n\t         \n            \n         \n      <\/table>\n\n   <table cellpadding=\"0\" style=\"border-spacing: 0px;\"><tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n            <\/td><\/tr>      \n                   \n                                 \n                      \n                  \n                    \n                  \n                   \n                     \n                  \n                  \n                    \n                     \n                  \n                    \n\t         \n            \n         \n      <tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n            <\/td><\/tr>                  \n                   \n                         \n                     \n                             \n                              \n                            \n                 \n            \n         \n      <tr><td style=\"background-color: 000000;\">\n            <\/td>\n      <\/tr><tr><td>\n            <\/td><\/tr>                     \n                                 \n                                    \n                                \n                                  \n            \n         \n      <\/table>","created_at":"2018-11-07T16:46:01+0000","updated_at":"2018-11-07T16:46:01+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"alumni.media.mit.edu","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":365,"title":"Document ClojureScript setup \u00b7 Issue #495 \u00b7 clojure-emacs\/cider","url":"https:\/\/github.com\/clojure-emacs\/cider\/issues\/495","content":"Document ClojureScript setup \u00b7 Issue #495 \u00b7 clojure-emacs\/cider \u00b7 GitHub\n    \n  \n  \n    \n  \n        \n  <div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n    \n  \n  \n  <details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark\" open=\"open\"><summary aria-haspopup=\"dialog\" aria-label=\"Close dialog\">\n      \n    <\/summary><\/details>\n<p>Press h to open a hovercard with more details.<\/p>","created_at":"2018-11-07T16:46:00+0000","updated_at":"2018-11-07T16:46:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/103882?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":366,"title":"Start Your Own ISP","url":"https:\/\/startyourownisp.com\/","content":"<p>This site is dedicated to helping you start your own Internet Service Provider. Specifically this guide is about building a Wireless ISP (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Wireless_Internet_service_provider\">WISP<\/a>).<\/p><p>This guide is focused on the very earliest stages of starting a WISP - determining feasibility up through connecting the first few customers. There are many challenges that will come up at 100, 1,000 or 10,000 customers that are not (yet) covered in this guide.<\/p><p><a href=\"https:\/\/riot.im\/app\/#\/room\/#startyourownisp:matrix.org\"><strong>Join the discussion!<\/strong><\/a> Chat with me (the author) and others interested in this kind of thing here: <a href=\"https:\/\/riot.im\/app\/#\/room\/#startyourownisp:matrix.org\">#startyourownisp:matrix.org<\/a>.<\/p><div class=\"admonition note\"><p class=\"admonition-title\">This site is a work in progress!<\/p><p>Only some of the content is up so far and there\u2019s still some bugs in the interface. Check out the <a href=\"https:\/\/startyourownisp.com\/posts\/changelog\/\">changelog<\/a> to see what\u2019s new.<\/p><\/div><h2 id=\"getting-started\">Getting Started<\/h2><p><a href=\"https:\/\/startyourownisp.com\/posts\/introduction\/\"><strong>What is a WISP?<\/strong><\/a>\nAnd why might you want to build one? Also defines some terminology.<\/p><p><a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1jjUYOQMuZ4cRyTv1M5X8HGtrKFiW96_gNlfG8BBzRN0\/edit?usp=sharing\"><strong>Costs<\/strong><\/a>\nWhat does it cost to build a wireless Internet Service Provider? (Link to a Google Sheet that you can copy and customize.)<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/about-me\/\"><strong>About Me<\/strong><\/a>\nWho am I? Why am I doing this?<\/p><h2 id=\"step-by-step-guide\">Step by Step Guide<\/h2><p><a href=\"https:\/\/startyourownisp.com\/posts\/location-location-location\/\"><strong>Step 1: Evaluate an Area<\/strong><\/a>: Make sure your area is a good candidate for a Wireless Internet network.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/fiber-provider\/\"><strong>Step 2: Find a Fiber Provider<\/strong><\/a>: Find a building where you can purchase a fiber connection and use the rooftop to start your wireless network.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/relay-sites\/\"><strong>Step 3: Find Relay Sites<\/strong><\/a>:\nExtend your network wirelessly toward your customers.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/hardware-platform\/\"><strong>Step 4: Pick a Hardware Platform<\/strong><\/a>:\nEvaluate available options for wireless hardware.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/billing-customer-management\/\"><strong>Step 5: Billing and Customer Management<\/strong><\/a>:\nMake sure you\u2019re able to get paid and support your customers.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/network-topology\/\"><strong>Step 6: Network Topology<\/strong><\/a>:\nDesign your network topology to make your network reliable and scalable. Routers, switches, IP addresses, VLANs, etc.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/build-your-infrastructure\/\"><strong>Step 7: Build your Infrastructure<\/strong><\/a>:\nInstall hardware for your fiber connection and your relay sites.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/install-a-customer\/\"><strong>Step 8: Install a Customer<\/strong><\/a>:\nGet your first customer online!<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/marketing\/\"><strong>Step 9: Marketing<\/strong><\/a>:\nLet people know about your service so they can experience a better Internet connection!<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/maintenance\/\"><strong>Step 10: Maintenance<\/strong><\/a>:\nKeep your network running smoothly, even in bad weather.<\/p><h2 id=\"miscellaneous\">Miscellaneous<\/h2><p><a href=\"https:\/\/startyourownisp.com\/posts\/form477\/\"><strong>Form 477: How to prepare and file with the FCC<\/strong><\/a>\nForm 477 is used by the FCC to determine which providers are servicing which areas. ISPs must file this form twice a year.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/tool-list\/\"><strong>Tools you\u2019ll want to have<\/strong><\/a>\nA list of the tools you\u2019ll need to install relays sites and customers.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/aim-backhaul\/\"><strong>Aim a Backhaul<\/strong><\/a>\nA guide describing the proper techniques for aiming backhauls. Designed to be printed out and taken to the site for reference.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/backhaul-picker\/\"><strong>Backhaul List<\/strong><\/a>\nIf you just need to get a solid wireless connection from Point A to Point B then use this list to pick the right equipment and get it set up.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/channel-planning\/\"><strong>RF Basics and Channel Planning<\/strong><\/a>\nAvoid self interference by carefully choosing channels for your access points and backhauls.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/mdus\/\"><strong>MDUs (Multiple Dwelling Units)<\/strong><\/a>\nBest practices for providing service to apartment buildings, condos, attached townhomes, etc.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/guide-to-google-earth\/\"><strong>Guide to Google Earth<\/strong><\/a>\nSome tips and tricks for using Google Earth to plan and build your network.<\/p><p><a href=\"https:\/\/startyourownisp.com\/posts\/roof-and-ladder-safety\/\"><strong>Roof and Ladder Safety<\/strong><\/a>\nStay safe out there!<\/p><aside class=\"copyright\" role=\"note\"><div style=\"text-align: center;\">\n\t\t\t\t\tSponsored by <a href=\"https:\/\/broadbandnow.com\/\">BroadbandNow<\/a>.<br \/>Looking to donate? <a href=\"https:\/\/www.patreon.com\/startyourownisp\">Patreon<\/a> or BTC 1JKa1Kdrp3r4xPSXBRJ6nPC6YYdLcqQ4Bp. Many thanks!\n\t\t\t\t\t<br \/>\u00a9 2018 startyourownisp.com \u2013\n\t\t\t\t\t\n\t\t\t\t\tDocumentation built with\n\t\t\t\t\t<a href=\"https:\/\/www.gohugo.io\" target=\"_blank\">Hugo<\/a>\n\t\t\t\t\tusing the\n\t\t\t\t\t<a href=\"http:\/\/github.com\/digitalcraftsman\/hugo-material-docs\" target=\"_blank\">Material<\/a> theme.\n\t\t\t\t<\/div>\n\t\t\t<\/aside>","created_at":"2018-11-07T16:46:00+0000","updated_at":"2018-11-08T02:22:02+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"startyourownisp.com","preview_picture":"https:\/\/startyourownisp.com\/images\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":364,"title":"cichli\/cider","url":"https:\/\/github.com\/cichli\/cider","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"http:\/\/www.gnu.org\/licenses\/gpl-3.0.txt\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/1ba3411d21f0105e0df26c7e29aa214578d377b5\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d47504c5f332d677265656e2e737667\" alt=\"License GPL 3\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/license-GPL_3-green.svg\" \/><\/a>\n<a href=\"http:\/\/melpa.org\/#\/cider\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/1b0fdf055ff3a1ca9f822217fa3098e4b416f25f\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f63696465722d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"http:\/\/melpa.org\/packages\/cider-badge.svg\" \/><\/a>\n<a href=\"http:\/\/stable.melpa.org\/#\/cider\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a0eda14a92f95440f17799b1f53b2c591c8e40f0\/687474703a2f2f737461626c652e6d656c70612e6f72672f7061636b616765732f63696465722d62616467652e737667\" alt=\"MELPA Stable\" data-canonical-src=\"http:\/\/stable.melpa.org\/packages\/cider-badge.svg\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/clojure-emacs\/cider\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5526d265aed5b9e33a1dd6aa52e25afa5cdb00ee\/68747470733a2f2f7472617669732d63692e6f72672f636c6f6a7572652d656d6163732f63696465722e706e673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/clojure-emacs\/cider.png?branch=master\" \/><\/a>\n<a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=GRQKNBM6P8VRQ\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e14c85b542e06215f7e56c0763333ef1e9b9f9b7\/68747470733a2f2f7777772e70617970616c6f626a656374732e636f6d2f656e5f55532f692f62746e2f62746e5f646f6e6174655f534d2e676966\" alt=\"Paypal\" data-canonical-src=\"https:\/\/www.paypalobjects.com\/en_US\/i\/btn\/btn_donate_SM.gif\" \/><\/a><\/p>\n<p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/7cea4fbf053aa9c4f9eafd6167065c828928a56c\/68747470733a2f2f7261772e6769746875622e636f6d2f636c6f6a7572652d656d6163732f63696465722f6d61737465722f6c6f676f2f63696465722d6c6f676f2d773634302e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/7cea4fbf053aa9c4f9eafd6167065c828928a56c\/68747470733a2f2f7261772e6769746875622e636f6d2f636c6f6a7572652d656d6163732f63696465722f6d61737465722f6c6f676f2f63696465722d6c6f676f2d773634302e706e67\" alt=\"CIDER Logo\" data-canonical-src=\"https:\/\/raw.github.com\/clojure-emacs\/cider\/master\/logo\/cider-logo-w640.png\" \/><\/a>\n<\/p>\n<p>CIDER (formerly <em>nrepl.el<\/em>) is the <strong>C<\/strong>lojure(Script) <strong>I<\/strong>nteractive\n<strong>D<\/strong>evelopment <strong>E<\/strong>nvironment that <strong>R<\/strong>ocks!<\/p>\n<p>CIDER extends Emacs with support for interactive programming in Clojure. The\nfeatures are centered around <code>cider-mode<\/code>, an Emacs minor-mode that complements\n<a href=\"https:\/\/github.com\/clojure-emacs\/clojure-mode\">clojure-mode<\/a>. While <code>clojure-mode<\/code> supports editing Clojure source files,\n<code>cider-mode<\/code> adds support for interacting with a running Clojure process for\ncompilation, debugging, definition and documentation lookup, running tests and\nso on.<\/p>\n<p>CIDER is the successor to the now deprecated combination of using <a href=\"https:\/\/github.com\/slime\/slime\">SLIME<\/a> +\n<a href=\"https:\/\/github.com\/technomancy\/swank-clojure\">swank-clojure<\/a> for Clojure development.<\/p>\n<p>If you like the project, please consider <a href=\"#donations\">supporting its ongoing development<\/a>.<\/p>\n<p><a href=\"https:\/\/gitter.im\/clojure-emacs\/cider?utm_source=badge&amp;utm_medium=badge&amp;utm_campaign=pr-badge&amp;utm_content=badge\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/da2edb525cde1455a622c58c0effc3a90b9a181c\/68747470733a2f2f6261646765732e6769747465722e696d2f4a6f696e253230436861742e737667\" alt=\"Gitter\" data-canonical-src=\"https:\/\/badges.gitter.im\/Join%20Chat.svg\" \/><\/a><\/p>\n<p><strong>This documentation tracks the <code>master<\/code> branch of CIDER. Some of\nthe features and settings discussed here might not be available in\nolder releases (including the current stable release). Please, consult\nthe relevant git tag (e.g. v0.11.0) if you need documentation for a\nspecific CIDER release.<\/strong><\/p>\n<hr \/><ul><li><a href=\"#overview\">Overview<\/a><\/li>\n<li><a href=\"#installation\">Installation<\/a>\n<ul><li><a href=\"#prerequisites\">Prerequisites<\/a><\/li>\n<li><a href=\"#installation-via-packageel\">Installation via package.el<\/a><\/li>\n<li><a href=\"#ciders-nrepl-middleware\">CIDER's nREPL middleware<\/a>\n<ul><li><a href=\"#setting-up-a-standalone-repl\">Setting up a standalone REPL<\/a><\/li>\n<li><a href=\"#using-embedded-nrepl-server\">Using embedded nREPL server<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#basic-usage\">Basic Usage<\/a>\n<ul><li><a href=\"#setting-up-a-leiningen-or-boot-project-optional\">Setting up a Leiningen or Boot project (optional)<\/a><\/li>\n<li><a href=\"#launch-a-nrepl-server-and-client-from-emacs\">Launch a nREPL server and client from Emacs<\/a><\/li>\n<li><a href=\"#connect-to-a-running-nrepl-server\">Connect to a running nREPL server<\/a><\/li>\n<li><a href=\"#using-cider-mode\">Using cider-mode<\/a><\/li>\n<li><a href=\"#using-the-repl\">Using the REPL<\/a>\n<ul><li><a href=\"#repl-configuration\">REPL Configuration<\/a>\n<ul><li><a href=\"#repl-history\">REPL history<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#clojurescript-usage\">ClojureScript usage<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#extended-workflow\">Extended Workflow<\/a>\n<ul><li><a href=\"#macroexpansion\">Macroexpansion<\/a><\/li>\n<li><a href=\"#value-inspection\">Value Inspection<\/a><\/li>\n<li><a href=\"#running-tests\">Running Tests<\/a><\/li>\n<li><a href=\"#navigating-stacktraces\">Navigating Stacktraces<\/a><\/li>\n<li><a href=\"#enlighten-display-local-values\">Enlighten (display local values)<\/a><\/li>\n<li><a href=\"#debugging\">Debugging<\/a><\/li>\n<li><a href=\"#code-reloading\">Code reloading<\/a><\/li>\n<li><a href=\"#managing-multiple-connections\">Managing multiple connections<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#configuration\">Configuration<\/a>\n<ul><li><a href=\"#basic-configuration\">Basic configuration<\/a><\/li>\n<li><a href=\"#overlays\">Overlays<\/a><\/li>\n<li><a href=\"#specifying-indentation\">Specifying indentation<\/a><\/li>\n<li><a href=\"#minibuffer-completion\">Minibuffer completion<\/a><\/li>\n<li><a href=\"#auto-completion\">Auto-completion<\/a><\/li>\n<li><a href=\"#pretty-printing\">Pretty-printing<\/a><\/li>\n<li><a href=\"#integration-with-other-modes\">Integration with other modes<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#quick-reference\">Quick reference<\/a><\/li>\n<li><a href=\"#caveats\">Caveats<\/a>\n<ul><li><a href=\"#var-metadata\">Var Metadata<\/a><\/li>\n<li><a href=\"#clojurescript-limitations\">ClojureScript limitations<\/a><\/li>\n<li><a href=\"#microsoft-windows\">Microsoft Windows<\/a><\/li>\n<li><a href=\"#powershell-el\">powershell.el<\/a><\/li>\n<li><a href=\"#clojureclr-support\">ClojureCLR support<\/a><\/li>\n<li><a href=\"#injecting-dependencies-and-leiningen-pedantic-abort-mode\">Injecting dependencies and Leiningen pedantic: abort mode<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#troubleshooting\">Troubleshooting<\/a><\/li>\n<li><a href=\"#presentations\">Presentations<\/a><\/li>\n<li><a href=\"#changelog\">Changelog<\/a><\/li>\n<li><a href=\"#team\">Team<\/a><\/li>\n<li><a href=\"#release-policy\">Release policy<\/a><\/li>\n<li><a href=\"#logo\">Logo<\/a><\/li>\n<li><a href=\"#contributing\">Contributing<\/a>\n<ul><li><a href=\"#discussion\">Discussion<\/a><\/li>\n<li><a href=\"#issues\">Issues<\/a><\/li>\n<li><a href=\"#patches\">Patches<\/a><\/li>\n<li><a href=\"#documentation\">Documentation<\/a><\/li>\n<li><a href=\"#donations\">Donations<\/a><\/li>\n<li><a href=\"#running-the-tests-in-batch-mode\">Running the tests in batch mode<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2>\n<p>CIDER aims to provide an interactive development experience similar to the one\nyou'd get when programming in Emacs Lisp, Common Lisp (with <a href=\"https:\/\/github.com\/slime\/slime\">SLIME<\/a> or <a href=\"https:\/\/github.com\/capitaomorte\/sly\">Sly<\/a>),\nScheme (with <a href=\"https:\/\/github.com\/jaor\/geiser\">Geiser<\/a>) and Smalltalk.<\/p>\n<p>Programmers are expected to program in a very dynamic and incremental manner,\nconstantly re-evaluating existing Clojure definitions and adding new ones to\ntheir running applications. You never stop\/start a Clojure application while\nusing CIDER - you're constantly interacting with it and changing it.<\/p>\n<p>You can find more details about the typical CIDER workflow in the <code>Basic Usage<\/code>\nsection. While we're a bit short on video tutorials, you can check out this\n<a href=\"https:\/\/www.youtube.com\/watch?v=_B_4vhsmRRI\" rel=\"nofollow\">tutorial about SLIME<\/a> to get a\nfeel about what do we mean by an \"Interactive Development Environment\".\nThere are plenty of differences between CIDER and SLIME, but the core ideas are\npretty much the same (and SLIME served as the principle inspiration for CIDER).<\/p>\n<p>CIDER's built on top of <a href=\"https:\/\/github.com\/clojure\/tools.nrepl\">nREPL<\/a>, the Clojure networked REPL server.<\/p>\n<p>CIDER's basic architecture looks something like this:<\/p>\n<p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/doc\/images\/cider_architecture.png\"><img src=\"https:\/\/github.com\/cichli\/cider\/raw\/master\/doc\/images\/cider_architecture.png\" width=\"600\" alt=\"image\" \/><\/a>\n<\/p>\n<p>Clojure code gets executed by an nREPL server. CIDER sends requests to the\nserver and processes its responses. The server's functionality is augmented by\nadditional nREPL middleware, designed specifically to address the needs of an\ninteractive development environment like CIDER. Much of the middleware we\ndeveloped for CIDER is editor-agnostic and is being used by other Clojure\ndevelopment environments as well (e.g. <a href=\"https:\/\/github.com\/tpope\/vim-fireplace\">vim-fireplace<\/a> &amp; <a href=\"https:\/\/github.com\/laurentpetit\/ccw\">CCW<\/a>).<\/p>\n<p>CIDER packs plenty of features. Here are some of them (in no particular order):<\/p>\n<ul><li><a href=\"#using-the-repl\">Powerful REPL<\/a><\/li>\n<li><a href=\"#using-cider-mode\">Interactive code evaluation<\/a><\/li>\n<li>Compilation notes (error and warning highlighting)<\/li>\n<li><a href=\"#navigating-stacktraces\">Human-friendly stacktraces<\/a><\/li>\n<li><a href=\"#auto-completion\">Smart code completion<\/a><\/li>\n<li>Definition lookup<\/li>\n<li>Documentation lookup<\/li>\n<li>Resource lookup<\/li>\n<li>Apropos<\/li>\n<li><a href=\"#debugging\">Debugger<\/a><\/li>\n<li><a href=\"#value-inspection\">Value inspector<\/a><\/li>\n<li>Function tracing<\/li>\n<li><a href=\"#macroexpansion\">Interactive macroexpansion<\/a><\/li>\n<li>Enhanced Clojure font-locking and indentation<\/li>\n<li><a href=\"http:\/\/conj.io\/\" rel=\"nofollow\">Grimoire<\/a> integration<\/li>\n<li><a href=\"#running-tests\"><code>clojure.test<\/code> integration<\/a><\/li>\n<li><a href=\"#code-reloading\">Smart code reloading<\/a><\/li>\n<li><a href=\"#pretty-printing\">Pretty-printing of results<\/a><\/li>\n<li>Classpath browser<\/li>\n<li>Namespace browser<\/li>\n<li>nREPL session management<\/li>\n<li>Scratchpad<\/li>\n<li>Minibuffer code evaluation<\/li>\n<li>Integration with <a href=\"http:\/\/company-mode.github.io\/\" rel=\"nofollow\">company-mode<\/a> and <a href=\"https:\/\/github.com\/clojure-emacs\/ac-cider\">auto-complete-mode<\/a><\/li>\n<li><a href=\"#managing-multiple-connections\">Support for working with multiple simultaneous nREPL connections<\/a><\/li>\n<\/ul><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/screenshots\/cider-overview.png\"><img src=\"https:\/\/github.com\/cichli\/cider\/raw\/master\/screenshots\/cider-overview.png\" alt=\"CIDER Screenshot\" \/><\/a><\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>The canonical way to install CIDER is via <code>package.el<\/code> (Emacs's built-in package\nmanager), but plenty of other options exist. (see the\n<a href=\"https:\/\/github.com\/clojure-emacs\/cider\/wiki\/Installation\">wiki<\/a>).<\/p>\n<h3><a id=\"user-content-prerequisites\" class=\"anchor\" aria-hidden=\"true\" href=\"#prerequisites\"><\/a>Prerequisites<\/h3>\n<p>You'll need to have Emacs installed (preferably the latest stable\nrelease). If you're new to Emacs you might want to read\n<a href=\"http:\/\/clojure-doc.org\/articles\/tutorials\/emacs.html\" rel=\"nofollow\">this tutorial<\/a>,\ndedicated to setting up Emacs for Clojure development first.<\/p>\n<p>CIDER officially supports Emacs 24.3+, Java 7+ and Clojure(Script) 1.7+.\nCIDER 0.10 was the final release which supported Java 6 and Clojure 1.5 and 1.6.<\/p>\n<p>You'll also need a recent version of your favorite build tool (Leiningen, Boot\nor Gradle) to be able to start CIDER via <code>cider-jack-in<\/code>. Generally it's a good\nidea to use their latest stable versions.<\/p>\n<p><strong>CIDER does not support ClojureCLR.<\/strong><\/p>\n<h3><a id=\"user-content-installation-via-packageel\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation-via-packageel\"><\/a>Installation via package.el<\/h3>\n<p><code>package.el<\/code> is the built-in package manager in Emacs.<\/p>\n<p>CIDER is available on the two major <code>package.el<\/code> community\nmaintained repos -\n<a href=\"http:\/\/stable.melpa.org\" rel=\"nofollow\">MELPA Stable<\/a>\nand <a href=\"http:\/\/melpa.org\" rel=\"nofollow\">MELPA<\/a>.<\/p>\n<p>You can install CIDER with the following command:<\/p>\n<p><kbd>M-x package-install [RET] cider [RET]<\/kbd><\/p>\n<p>or by adding this bit of Emacs Lisp code to your Emacs initialization file\n(<code>.emacs<\/code> or <code>init.el<\/code>):<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(unless (package-installed-p 'cider)\n  (package-install 'cider))<\/pre><\/div>\n<p>If the installation doesn't work try refreshing the package list:<\/p>\n<p><kbd>M-x package-refresh-contents [RET]<\/kbd><\/p>\n<p>Keep in mind that MELPA packages are built automatically from\nthe <code>master<\/code> branch, meaning bugs might creep in there from time to\ntime. Never-the-less, installing from MELPA is a reasonable way of\nobtaining CIDER, as the <code>master<\/code> branch is normally quite stable\nand serious regressions there are usually fixed pretty quickly.<\/p>\n<p>Generally, users of the non-adventurous kind are advised to stick\nwith the stable releases, available from MELPA Stable.\nIn Emacs 24.4+, you can pin CIDER to always use MELPA\nStable by adding this to your Emacs initialization:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'package-pinned-packages '(cider . \"melpa-stable\") t)<\/pre><\/div>\n<p><strong>CIDER has dependencies (e.g. <code>queue<\/code> &amp; <code>seq<\/code>) that are only available in the\n<a href=\"https:\/\/elpa.gnu.org\/\" rel=\"nofollow\">GNU ELPA repository<\/a>. It's the only package repository\nenabled by default in Emacs and you should not disable it!<\/strong><\/p>\n<h3><a id=\"user-content-ciders-nrepl-middleware\" class=\"anchor\" aria-hidden=\"true\" href=\"#ciders-nrepl-middleware\"><\/a>CIDER's nREPL middleware<\/h3>\n<p>Much of CIDER's functionality depends on the presence of CIDER's own\n<a href=\"https:\/\/github.com\/clojure-emacs\/cider-nrepl\">nREPL middleware<\/a>. Starting with version 0.11, When <code>cider-jack-in<\/code> (<kbd>C-c M-j<\/kbd>) is\nused, CIDER takes care of injecting it and its other dependencies.<\/p>\n<p><strong><code>profiles.clj<\/code> or <code>profile.boot<\/code> don't need to be modified anymore for the above usecase!<\/strong><\/p>\n<p>If you don't want <code>cider-jack-in<\/code> to inject dependencies automatically, set\n<code>cider-inject-dependencies-at-jack-in<\/code> to nil. Note that you'll have to setup\nthe dependencies yourself (see the section below), just as in CIDER 0.10 and older.<\/p>\n<p>If a standalone REPL is preferred, you need to invoke <code>cider-connect<\/code> (instead\nof <code>cider-jack-in<\/code>) and you'll need to manually add the dependencies to your\nClojure project (explained in the following section).<\/p>\n<h4><a id=\"user-content-setting-up-a-standalone-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#setting-up-a-standalone-repl\"><\/a>Setting up a standalone REPL<\/h4>\n<h5><a id=\"user-content-using-leiningen\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-leiningen\"><\/a>Using Leiningen<\/h5>\n<p>Use the convenient plugin for defaults, either in your project's\n<code>project.clj<\/code> file or in the :repl profile in <code>~\/.lein\/profiles.clj<\/code>.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:plugins [[cider\/cider-nrepl \"x.y.z\"]]<\/pre><\/div>\n<p>A minimal <code>profiles.clj<\/code> for CIDER would be:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:repl {:plugins [[cider\/cider-nrepl \"0.11.0\"]]}}<\/pre><\/div>\n<p><strong>Be careful not to place this in the <code>:user<\/code> profile, as this way CIDER's\nmiddleware will always get loaded, causing <code>lein<\/code> to start slower.  You really\nneed it just for <code>lein repl<\/code> and this is what the <code>:repl<\/code> profile is for.<\/strong><\/p>\n<h5><a id=\"user-content-using-boot\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-boot\"><\/a>Using Boot<\/h5>\n<p>Boot users can configure the tool to include the middleware automatically in\nall of their projects using a <code>~\/.boot\/profile.boot<\/code> file like so:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require 'boot.repl)\n(swap! boot.repl\/*default-dependencies*\n       concat '[[cider\/cider-nrepl \"0.11.0\"]])\n(swap! boot.repl\/*default-middleware*\n       conj 'cider.nrepl\/cider-middleware)<\/pre><\/div>\n<p>For more information visit <a href=\"https:\/\/github.com\/boot-clj\/boot\/wiki\/Cider-REPL\">boot-clj wiki<\/a>.<\/p>\n<h4><a id=\"user-content-using-embedded-nrepl-server\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-embedded-nrepl-server\"><\/a>Using embedded nREPL server<\/h4>\n<p>If you're embedding nREPL in your application you'll have to start the\nserver with CIDER's own nREPL handler.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns my-app\n  (:require [clojure.tools.nrepl.server :as nrepl-server]\n            [cider.nrepl :refer (cider-nrepl-handler)]))\n(defn -main\n  []\n  (nrepl-server\/start-server :port 7888 :handler cider-nrepl-handler))<\/pre><\/div>\n<p>It goes without saying that your project should depend on <code>cider-nrepl<\/code>.<\/p>\n<hr \/><p><code>x.y.z<\/code> should match the version of CIDER you're currently using (say <code>0.7.1<\/code>).\nFor snapshot releases of CIDER you should use the snapshot of the plugin as well\n(say <code>0.7.1-SNAPSHOT<\/code>).<\/p>\n<p><strong>Note that you need to use at least CIDER 0.7 for the nREPL middleware to work\nproperly.  Don't use cider-nrepl with CIDER 0.6.<\/strong><\/p>\n<h2><a id=\"user-content-basic-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-usage\"><\/a>Basic Usage<\/h2>\n<p>The only requirement to use CIDER is to have a nREPL server to\nwhich it may connect. Many Clojurians favour the use of the Leiningen or Boot tools\nto start an nREPL server, but the use of Leiningen or Boot is not a prerequisite to use\nCIDER (however, it <em>is<\/em> required if you want to use the <code>cider-jack-in<\/code> command).<\/p>\n<h3><a id=\"user-content-setting-up-a-leiningen-or-boot-project-optional\" class=\"anchor\" aria-hidden=\"true\" href=\"#setting-up-a-leiningen-or-boot-project-optional\"><\/a>Setting up a Leiningen or Boot project (optional)<\/h3>\n<p><a href=\"http:\/\/leiningen.org\/\" rel=\"nofollow\">Leiningen<\/a> is the de facto standard build\/project\nmanagement tool for Clojure. <a href=\"http:\/\/boot-clj.com\/\" rel=\"nofollow\">Boot<\/a> is a newer build tool\noffering abstractions and libraries to construct more complex build\nscenarios. Both have a similar scope to the Maven build tool favoured by Java\ndevelopers (and they actually reuse many things from the Maven ecosystem).<\/p>\n<p>CIDER features a command called <code>cider-jack-in<\/code> that will start an nREPL server\nfor a particular Leiningen or Boot project and connect to it automatically.\nThis functionality depends on Leiningen 2.5.2+ or Boot\n2.0.0+. Older versions are not supported. For Leiningen, follow the installation\ninstructions on its web site to get it up and running and afterwards create a\nproject like this:<\/p>\n<pre>$ lein new demo\n<\/pre>\n<p>The two main ways to obtain an nREPL connection are discussed in the following sections of the manual.<\/p>\n<h3><a id=\"user-content-launch-a-nrepl-server-and-client-from-emacs\" class=\"anchor\" aria-hidden=\"true\" href=\"#launch-a-nrepl-server-and-client-from-emacs\"><\/a>Launch a nREPL server and client from Emacs<\/h3>\n<p>Simply open in Emacs a file belonging to your <code>lein<\/code> or <code>boot<\/code> project (like\n<code>foo.clj<\/code>) and type <kbd>M-x cider-jack-in<\/kbd>. This will start a nREPL server with\nall the project dependencies loaded in and CIDER will automatically connect to it.<\/p>\n<p>Alternatively you can use <kbd>C-u M-x cider-jack-in<\/kbd> to specify the name of\na <code>lein<\/code> or <code>boot<\/code> project, without having to visit any file in it.<\/p>\n<p>In Clojure(Script) buffers the command <code>cider-jack-in<\/code> is bound to <kbd>C-c M-j<\/kbd>.<\/p>\n<h3><a id=\"user-content-connect-to-a-running-nrepl-server\" class=\"anchor\" aria-hidden=\"true\" href=\"#connect-to-a-running-nrepl-server\"><\/a>Connect to a running nREPL server<\/h3>\n<p>You can go to your project's directory in a terminal and type there\n(assuming you're using Leiningen that is):<\/p>\n<pre>$ lein repl\n<\/pre>\n<p>Or with Boot:<\/p>\n<pre>$ boot repl wait\n<\/pre>\n<p>Alternatively you can start nREPL either manually or by the facilities provided by your\nproject's build tool (Maven, etc).<\/p>\n<p>After you get your nREPL server running go back to Emacs.\nTyping there <kbd>M-x cider-connect<\/kbd> will allow you to connect to the running nREPL server.<\/p>\n<p>In Clojure(Script) buffers the command <code>cider-connect<\/code> is bound to <kbd>C-c M-c<\/kbd>.<\/p>\n<h3><a id=\"user-content-using-cider-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-cider-mode\"><\/a>Using cider-mode<\/h3>\n<p>CIDER comes with a handy minor mode called <code>cider-mode<\/code> (complementing\n<code>clojure-mode<\/code>) that allows you to evaluate code in your Clojure source\nfiles and load it directly in the REPL. <code>cider-mode<\/code> is the primary\nway you're supposed to be interacting with your REPL process. If you\nwant to get productive with CIDER, you'll have to get intimately familiar\nwith it.<\/p>\n<p>Here's a list of <code>cider-mode<\/code>'s keybindings:<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>C-x C-e<\/kbd> <kbd>C-c C-e<\/kbd><\/td>\n<td>Evaluate the form preceding point and display the result in the echo area and\/or in an buffer overlay (according to <code>cider-use-overlays<\/code>).  If invoked with a prefix argument, insert the result into the current buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-w<\/kbd><\/td>\n<td>Evaluate the form preceding point and replace it with its result.<\/td>\n<\/tr><tr><td><kbd>C-c M-e<\/kbd><\/td>\n<td>Evaluate the form preceding point and output it result to the REPL buffer.  If invoked with a prefix argument, takes you to the REPL buffer after being invoked.<\/td>\n<\/tr><tr><td><kbd>C-c M-p<\/kbd><\/td>\n<td>Load the form preceding point in the REPL buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-p<\/kbd><\/td>\n<td>Evaluate the form preceding point and pretty-print the result in a popup buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-f<\/kbd><\/td>\n<td>Evaluate the top level form under point and pretty-print the result in a popup buffer.<\/td>\n<\/tr><tr><td><kbd>C-M-x<\/kbd> <kbd>C-c C-c<\/kbd><\/td>\n<td>Evaluate the top level form under point and display the result in the echo area.<\/td>\n<\/tr><tr><td><kbd>C-u C-M-x<\/kbd> <kbd>C-u C-c C-c<\/kbd><\/td>\n<td>Debug the top level form under point and walk through its evaluation<\/td>\n<\/tr><tr><td><kbd>C-c C-r<\/kbd><\/td>\n<td>Evaluate the region and display the result in the echo area.<\/td>\n<\/tr><tr><td><kbd>C-c C-b<\/kbd><\/td>\n<td>Interrupt any pending evaluations.<\/td>\n<\/tr><tr><td><kbd>C-c C-m<\/kbd><\/td>\n<td>Invoke <code>macroexpand-1<\/code> on the form at point and display the result in a macroexpansion buffer.  If invoked with a prefix argument, <code>macroexpand<\/code> is used instead of <code>macroexpand-1<\/code>.<\/td>\n<\/tr><tr><td><kbd>C-c M-m<\/kbd><\/td>\n<td>Invoke <code>clojure.walk\/macroexpand-all<\/code> on the form at point and display the result in a macroexpansion buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-n<\/kbd><\/td>\n<td>Eval the ns form.<\/td>\n<\/tr><tr><td><kbd>C-c M-n<\/kbd><\/td>\n<td>Switch the namespace of the REPL buffer to the namespace of the current buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-z<\/kbd><\/td>\n<td>Switch to the relevant REPL buffer. Use a prefix argument to change the namespace of the REPL buffer to match the currently visited source file.<\/td>\n<\/tr><tr><td><kbd>C-u C-u C-c C-z<\/kbd><\/td>\n<td>Switch to the REPL buffer based on a user prompt for a directory.<\/td>\n<\/tr><tr><td><kbd>C-c M-z<\/kbd><\/td>\n<td>Load (eval) the current buffer and switch to the relevant REPL buffer. Use a prefix argument to change the namespace of the REPL buffer to match the currently visited source file.<\/td>\n<\/tr><tr><td><kbd>C-c M-d<\/kbd><\/td>\n<td>Display default REPL connection details, including project directory name, buffer namespace, host and port.<\/td>\n<\/tr><tr><td><kbd>C-c M-r<\/kbd><\/td>\n<td>Rotate and display the default nREPL connection.<\/td>\n<\/tr><tr><td><kbd>C-c C-o<\/kbd><\/td>\n<td>Clear the last output in the REPL buffer. With a prefix argument it will clear the entire REPL buffer, leaving only a prompt. Useful if you're running the REPL buffer in a side by side buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-k<\/kbd><\/td>\n<td>Load (eval) the current buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-l<\/kbd><\/td>\n<td>Load (eval) a Clojure file.<\/td>\n<\/tr><tr><td><kbd>C-c C-x<\/kbd><\/td>\n<td>Reload all modified files on the classpath. If invoked with a prefix argument, reload all files on the classpath. If invoked with a double prefix argument, clear the state of the namespace tracker before reloading.<\/td>\n<\/tr><tr><td><kbd>C-c C-d d<\/kbd><\/td>\n<td>Display doc string for the symbol at point.  If invoked with a prefix argument, or no symbol is found at point, prompt for a symbol.<\/td>\n<\/tr><tr><td><kbd>C-c C-d j<\/kbd><\/td>\n<td>Display JavaDoc (in your default browser) for the symbol at point.  If invoked with a prefix argument, or no symbol is found at point, prompt for a symbol.<\/td>\n<\/tr><tr><td><kbd>C-c M-i<\/kbd><\/td>\n<td>Inspect expression. Will act on expression at point if present.<\/td>\n<\/tr><tr><td><kbd>C-c M-t v<\/kbd><\/td>\n<td>Toggle var tracing.<\/td>\n<\/tr><tr><td><kbd>C-c M-t n<\/kbd><\/td>\n<td>Toggle namespace tracing.<\/td>\n<\/tr><tr><td><kbd>C-c C-u<\/kbd><\/td>\n<td>Undefine a symbol. If invoked with a prefix argument, or no symbol is found at point, prompt for a symbol.<\/td>\n<\/tr><tr><td><kbd>C-c C-t t<\/kbd><\/td>\n<td>Run test at point.<\/td>\n<\/tr><tr><td><kbd>C-c C-t n<\/kbd><\/td>\n<td>Run tests for current namespace.<\/td>\n<\/tr><tr><td><kbd>C-c C-t l<\/kbd><\/td>\n<td>Run tests for all loaded namespaces.<\/td>\n<\/tr><tr><td><kbd>C-c C-t p<\/kbd><\/td>\n<td>Run tests for all project namespaces. This loads the additional namespaces.<\/td>\n<\/tr><tr><td><kbd>C-c C-t r<\/kbd><\/td>\n<td>Re-run test failures\/errors.<\/td>\n<\/tr><tr><td><kbd>C-c C-t b<\/kbd><\/td>\n<td>Show the test report buffer.<\/td>\n<\/tr><tr><td><kbd>M-.<\/kbd><\/td>\n<td>Jump to the definition of a symbol.  If invoked with a prefix argument, or no symbol is found at point, prompt for a symbol.<\/td>\n<\/tr><tr><td><kbd>C-c M-.<\/kbd><\/td>\n<td>Jump to the resource referenced by the string at point.<\/td>\n<\/tr><tr><td><kbd>C-c C-.<\/kbd><\/td>\n<td>Jump to some namespace on the classpath.<\/td>\n<\/tr><tr><td><kbd>M-,<\/kbd><\/td>\n<td>Return to your pre-jump location.<\/td>\n<\/tr><tr><td><kbd>M-TAB<\/kbd><\/td>\n<td>Complete the symbol at point.<\/td>\n<\/tr><tr><td><kbd>C-c C-d r<\/kbd><\/td>\n<td>Lookup symbol in Grimoire.<\/td>\n<\/tr><tr><td><kbd>C-c C-d a<\/kbd><\/td>\n<td>Apropos search for functions\/vars.<\/td>\n<\/tr><tr><td><kbd>C-c C-d A<\/kbd><\/td>\n<td>Apropos search for documentation.<\/td>\n<\/tr><tr><td><kbd>C-c C-q<\/kbd><\/td>\n<td>Quit the current nREPL connection. With a prefix argument it will quit all connections.<\/td>\n<\/tr><\/tbody><\/table><p>There's no need to memorize this list. In any Clojure buffer with <code>cider-mode<\/code>\nactive you'll have a CIDER menu available, which lists all the most important\ncommands and their keybindings. You can also invoke <code>C-h f RET cider-mode<\/code> to\nget a list of the keybindings for <code>cider-mode<\/code>.<\/p>\n<h3><a id=\"user-content-using-the-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-the-repl\"><\/a>Using the REPL<\/h3>\n<p>CIDER comes with a powerful REPL, which is quite handy when you want to\nexperiment with the code you're working on or just explore some stuff (e.g. a\nlibrary you're playing with).  The REPL offers a number of advanced features:<\/p>\n<ul><li>auto-completion<\/li>\n<li>font-locking (the same as in <code>clojure-mode<\/code>)<\/li>\n<li>quick access to many CIDER commands (e.g. definition and documentation lookup, tracing, etc)<\/li>\n<\/ul><p>Here's a list of the keybindings that are available in CIDER's REPL:<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>RET<\/kbd><\/td>\n<td>Evaluate the current input in Clojure if it is complete. If incomplete, open a new line and indent. If invoked with a prefix argument is given then the input is evaluated without checking for completeness.<\/td>\n<\/tr><tr><td><kbd>C-RET<\/kbd><\/td>\n<td>Close any unmatched parenthesis and then evaluate the current input in Clojure.<\/td>\n<\/tr><tr><td><kbd>C-j<\/kbd><\/td>\n<td>Open a new line and indent.<\/td>\n<\/tr><tr><td><kbd>C-c C-o<\/kbd><\/td>\n<td>Remove the output of the previous evaluation from the REPL buffer. With a prefix argument it will clear the entire REPL buffer, leaving only a prompt.<\/td>\n<\/tr><tr><td><kbd>C-c M-o<\/kbd><\/td>\n<td>Switch between the Clojure and ClojureScript REPLs for the current project.<\/td>\n<\/tr><tr><td><kbd>C-c C-u<\/kbd><\/td>\n<td>Kill all text from the prompt to the current point.<\/td>\n<\/tr><tr><td><kbd>C-c C-b<\/kbd> <kbd>C-c C-c<\/kbd><\/td>\n<td>Interrupt any pending evaluations.<\/td>\n<\/tr><tr><td><kbd>C-up<\/kbd> <kbd>C-down<\/kbd><\/td>\n<td>Goto to previous\/next input in history.<\/td>\n<\/tr><tr><td><kbd>M-p<\/kbd> <kbd>M-n<\/kbd><\/td>\n<td>Search the previous\/next item in history using the current input as search pattern. If <kbd>M-p\/M-n<\/kbd> is typed two times in a row, the second invocation uses the same search pattern (even if the current input has changed).<\/td>\n<\/tr><tr><td><kbd>M-s<\/kbd> <kbd>M-r<\/kbd><\/td>\n<td>Search forward\/reverse through command history with regex.<\/td>\n<\/tr><tr><td><kbd>C-c C-n<\/kbd> <kbd>C-c C-p<\/kbd><\/td>\n<td>Move between the current and previous prompts in the REPL buffer. Pressing <kbd>RET<\/kbd> on a line with old input copies that line to the newest prompt.<\/td>\n<\/tr><tr><td><kbd>C-c C-x<\/kbd><\/td>\n<td>Reload all modified files on the classpath.<\/td>\n<\/tr><tr><td><kbd>C-u C-c C-x<\/kbd><\/td>\n<td>Reload all files on the classpath.<\/td>\n<\/tr><tr><td><kbd>TAB<\/kbd><\/td>\n<td>Complete symbol at point.<\/td>\n<\/tr><tr><td><kbd>C-c C-d d<\/kbd><\/td>\n<td>Display doc string for the symbol at point.  If invoked with a prefix argument, or no symbol is found at point, prompt for a symbol<\/td>\n<\/tr><tr><td><kbd>C-c C-d j<\/kbd><\/td>\n<td>Display JavaDoc (in your default browser) for the symbol at point.  If invoked with a prefix argument, or no symbol is found at point, prompt for a symbol.<\/td>\n<\/tr><tr><td><kbd>C-c C-d r<\/kbd><\/td>\n<td>Lookup symbol in Grimoire.<\/td>\n<\/tr><tr><td><kbd>C-c C-d a<\/kbd><\/td>\n<td>Apropos search for functions\/vars.<\/td>\n<\/tr><tr><td><kbd>C-c C-d A<\/kbd><\/td>\n<td>Apropos search for documentation.<\/td>\n<\/tr><tr><td><kbd>C-c C-z<\/kbd><\/td>\n<td>Switch to the previous Clojure buffer. This complements <kbd>C-c C-z<\/kbd> used in cider-mode.<\/td>\n<\/tr><tr><td><kbd>C-c M-i<\/kbd><\/td>\n<td>Inspect expression. Will act on expression at point if present.<\/td>\n<\/tr><tr><td><kbd>C-c M-n<\/kbd><\/td>\n<td>Select a namespace and switch to it.<\/td>\n<\/tr><tr><td><kbd>C-c C-.<\/kbd><\/td>\n<td>Jump to some namespace on the classpath.<\/td>\n<\/tr><tr><td><kbd>C-c M-t v<\/kbd><\/td>\n<td>Toggle var tracing.<\/td>\n<\/tr><tr><td><kbd>C-c M-t n<\/kbd><\/td>\n<td>Toggle namespace tracing.<\/td>\n<\/tr><tr><td><kbd>C-c C-t t<\/kbd><\/td>\n<td>Run test at point.<\/td>\n<\/tr><tr><td><kbd>C-c C-t n<\/kbd><\/td>\n<td>Run tests for current namespace.<\/td>\n<\/tr><tr><td><kbd>C-c C-t l<\/kbd><\/td>\n<td>Run tests for all loaded namespaces.<\/td>\n<\/tr><tr><td><kbd>C-c C-t p<\/kbd><\/td>\n<td>Run tests for all project namespaces. This loads the additional namespaces.<\/td>\n<\/tr><tr><td><kbd>C-c C-t r<\/kbd><\/td>\n<td>Re-run test failures\/errors.<\/td>\n<\/tr><tr><td><kbd>C-c C-t b<\/kbd><\/td>\n<td>Show the test report buffer.<\/td>\n<\/tr><tr><td><kbd>C-c C-q<\/kbd><\/td>\n<td>Quit the current nREPL connection. With a prefix argument it will quit all connections.<\/td>\n<\/tr><\/tbody><\/table><p>There's no need to memorize this list. In any REPL buffer you'll have a <code>REPL<\/code>\nmenu available, which lists all the most important commands and their\nkeybindings. You can also invoke <code>C-h f RET cider-repl-mode<\/code> to get a list of the\nkeybindings for <code>cider-repl-mode<\/code>.<\/p>\n<p>In the REPL you can also use \"shortcut commands\" by pressing <code>,<\/code> at the\nbeginning of a REPL line. You'll be presented with a list of commands you can\nquickly run (like quitting, displaying some info, clearing the REPL, etc). The\ncharacter used to trigger the shortcuts is configurable via\n<code>cider-repl-shortcut-dispatch-char<\/code>. Here's how you can change it to <code>:<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-shortcut-dispatch-char ?\\:)<\/pre><\/div>\n<h4><a id=\"user-content-repl-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#repl-configuration\"><\/a>REPL Configuration<\/h4>\n<ul><li>Enable <code>eldoc<\/code> in REPL buffers:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-repl-mode-hook #'eldoc-mode)<\/pre><\/div>\n<ul><li>\n<p>You can customize the prompt in REPL buffer. To do that you can customize\n<code>cider-repl-prompt-function<\/code> and set it to a function that takes one argument,\na namespace name. For convenience, three functions are already provided:\n<code>cider-repl-prompt-lastname<\/code>, <code>cider-repl-prompt-abbreviated<\/code>,\n<code>cider-repl-prompt-default<\/code> and by default the last one is being used.\nPrompt for each of them for namespace <code>leiningen.core.ssl<\/code>:<\/p>\n<ul><li><code>cider-repl-prompt-lastname<\/code>:<\/li>\n<\/ul><pre>ssl&gt;\n<\/pre>\n<ul><li><code>cider-repl-prompt-abbreviated<\/code>:<\/li>\n<\/ul><pre>l.c.ssl&gt;\n<\/pre>\n<ul><li><code>cider-repl-prompt-default<\/code>:<\/li>\n<\/ul><pre>leiningen.core.ssl&gt;\n<\/pre>\n<p>You may, of course, write your own function. For example, in <code>leiningen<\/code> there\nare two namespaces with similar names - <code>leiningen.classpath<\/code> and\n<code>leiningen.core.classpath<\/code>. To make them easily recognizable you can either\nuse the default value or you can opt to show only two segments of the\nnamespace and still be able to know which is the REPL's current\nnamespace. Here is an example function that will do exactly that:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun cider-repl-prompt-show-two (namespace)\n  \"Return a prompt string with the last name in NAMESPACE.\"\n  (let* ((names (reverse (-take 2 (reverse (split-string namespace \"\\\\.\"))))))\n    (concat (car names) \".\" (cadr names) \"&gt; \")))<\/pre><\/div>\n<\/li>\n<li>\n<p>You can control the <kbd>TAB<\/kbd> key behavior in the REPL via the\n<code>cider-repl-tab-command<\/code> variable.  While the default command\n<code>cider-repl-indent-and-complete-symbol<\/code> should be an adequate choice for\nmost users, it's very easy to switch to another command if you wish\nto. For instance if you'd like <kbd>TAB<\/kbd> to only indent (maybe\nbecause you're used to completing with <kbd>M-TAB<\/kbd>) use the\nfollowing snippet:<\/p>\n<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-tab-command #'indent-for-tab-command)<\/pre><\/div>\n<ul><li>Change the result prefix for REPL evaluation (by default there's no prefix):<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-result-prefix \";; =&gt; \")<\/pre><\/div>\n<p>And here's the result of that change:<\/p>\n<pre>user&gt; (+ 1 2)\n;; =&gt; 3\n<\/pre>\n<ul><li>The REPL buffer name has the format <code>*cider-repl project-name*<\/code>.\nChange the separator from space to something else by overriding <code>nrepl-buffer-name-separator<\/code>.<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq nrepl-buffer-name-separator \"-\")<\/pre><\/div>\n<ul><li>The REPL buffer name can also display the port on which the nREPL server is running.\nBuffer name will look like <code>*cider-repl project-name:port*<\/code>.<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq nrepl-buffer-name-show-port t)<\/pre><\/div>\n<ul><li>Normally code in the REPL is font-locked the same way as in\n<code>clojure-mode<\/code>. Before CIDER 0.10 by default REPL input was font-locked with\n<code>cider-repl-input-face<\/code> (after you press <code>RET<\/code>) and results were font-locked with\n<code>cider-repl-result-face<\/code>. If you want to restore the old behaviour use:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-use-clojure-font-lock nil)<\/pre><\/div>\n<h5><a id=\"user-content-pretty-printing-in-the-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#pretty-printing-in-the-repl\"><\/a>Pretty printing in the REPL<\/h5>\n<p>Make the REPL always pretty-print the results of your commands.<\/p>\n<p><kbd>M-x cider-repl-toggle-pretty-printing<\/kbd><\/p>\n<h5><a id=\"user-content-limiting-printed-output-in-the-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#limiting-printed-output-in-the-repl\"><\/a>Limiting printed output in the REPL<\/h5>\n<p>Accidentally printing large objects can be detrimental to your\nproductivity. Clojure provides the <code>*print-length*<\/code> var which, if set,\ncontrols how many items of each collection the printer will print. You\ncan supply a default value for REPL sessions via the <code>repl-options<\/code>\nsection of your Leiningen project's configuration.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:repl-options {:init (set! *print-length* 50)}<\/pre><\/div>\n<h5><a id=\"user-content-repl-history\" class=\"anchor\" aria-hidden=\"true\" href=\"#repl-history\"><\/a>REPL history<\/h5>\n<ul><li>To make the REPL history wrap around when its end is reached:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-wrap-history t)<\/pre><\/div>\n<ul><li>To adjust the maximum number of items kept in the REPL history:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-history-size 1000) ; the default is 500<\/pre><\/div>\n<ul><li>To store the REPL history in a file:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-history-file \"path\/to\/file\")<\/pre><\/div>\n<p>Note that the history is written to the file when you kill the REPL\nbuffer (which includes invoking <code>cider-quit<\/code>) or you quit Emacs.<\/p>\n<h3><a id=\"user-content-clojurescript-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojurescript-usage\"><\/a>ClojureScript usage<\/h3>\n<p>ClojureScript support relies on the\n<a href=\"https:\/\/github.com\/cemerick\/piggieback\">piggieback<\/a> nREPL middleware being\npresent in your REPL session.<\/p>\n<ol><li>\n<p>Add the following dependencies to your <code>project.clj<\/code><\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>[com.cemerick\/piggieback \"0.2.1\"]\n[org.clojure\/clojure \"1.7.0\"]<\/pre><\/div>\n<p>as well as the following option:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:repl-options {:nrepl-middleware [cemerick.piggieback\/wrap-cljs-repl]}<\/pre><\/div>\n<\/li>\n<li>\n<p>Issue <kbd>M-x<\/kbd> <code>customize-variable<\/code> <kbd>RET<\/kbd> <code>cider-cljs-lein-repl<\/code> if\nyou'd like to change the REPL used (the default is <code>rhino<\/code>).<\/p>\n<\/li>\n<li>\n<p>Open a file in your project and issue <kbd>M-x<\/kbd>\n<code>cider-jack-in-clojurescript<\/code>. This will start up the nREPL server, and then create\ntwo REPL buffers for you, one in Clojure and one in ClojureScript. All usual\nCIDER commands will be automatically directed to the appropriate REPL,\ndepending on whether you're visiting a <code>.clj<\/code> or a <code>.cljs<\/code> file.<\/p>\n<\/li>\n<\/ol><h4><a id=\"user-content-browser-connected-clojurescript-repl\" class=\"anchor\" aria-hidden=\"true\" href=\"#browser-connected-clojurescript-repl\"><\/a>Browser-connected ClojureScript REPL<\/h4>\n<p>Using Weasel, you can also have a browser-connected REPL.<\/p>\n<ol><li>\n<p>Add <code>[weasel \"0.7.0\"]<\/code> to your project's <code>:dependencies<\/code>.<\/p>\n<\/li>\n<li>\n<p>Issue <kbd>M-x<\/kbd> <code>customize-variable<\/code> <kbd>RET<\/kbd> <code>cider-cljs-lein-repl<\/code>\nand choose the <code>Weasel<\/code> option.<\/p>\n<\/li>\n<li>\n<p>Add this to your ClojureScript code:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(ns my.cljs.core\n  (:require [weasel.repl :as repl]))\n(repl\/connect \"ws:\/\/localhost:9001\")<\/pre><\/div>\n<\/li>\n<li>\n<p>Open a file in your project and issue <code>M-x cider-jack-in-clojurescript<\/code>.<\/p>\n<\/li>\n<\/ol><p>Provided that a Piggieback-enabled ClojureScript environment is active in your\nREPL session, code loading and evaluation will work seamlessly regardless of the\npresence of the <code>cider-nrepl<\/code> middleware. If the middleware is present then most\nother features of CIDER will also be enabled (including code completion,\ndocumentation lookup, the namespace browser, and macroexpansion).<\/p>\n<h4><a id=\"user-content-browser-connected-clojurescript-repl-in-boot-project\" class=\"anchor\" aria-hidden=\"true\" href=\"#browser-connected-clojurescript-repl-in-boot-project\"><\/a>Browser-connected ClojureScript REPL in Boot project<\/h4>\n<ol><li>Add this to your dependencies in <code>build.boot<\/code>:<\/li>\n<\/ol><div class=\"highlight highlight-source-clojure\"><pre>[adzerk\/boot-cljs-repl   \"0.3.0\"]\n[com.cemerick\/piggieback \"0.2.1\"  :scope \"test\"]\n[weasel                  \"0.7.0\"  :scope \"test\"]\n[org.clojure\/tools.nrepl \"0.2.12\" :scope \"test\"]<\/pre><\/div>\n<ol start=\"2\"><li>\n<p>Start <code>boot dev<\/code> in a terminal.<\/p>\n<\/li>\n<li>\n<p><code>M-x cider-connect<\/code> to localhost and select the repl process.<\/p>\n<\/li>\n<li>\n<p>Execute <code>(start-repl)<\/code> at the prompt: <code>boot.user&gt; (start-repl)<\/code>.<\/p>\n<\/li>\n<li>\n<p>Connect to the running server with your browser.<\/p>\n<\/li>\n<\/ol><p>For more information visit <a href=\"https:\/\/github.com\/adzerk-oss\/boot-cljs-repl\">boot-cljs-repl<\/a>.<\/p>\n<h4><a id=\"user-content-using-the-figwheel-repl-leiningen-only\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-the-figwheel-repl-leiningen-only\"><\/a>Using the Figwheel REPL (Leiningen-only)<\/h4>\n<p>You can also use <a href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/Figwheel\">https:\/\/github.com\/bhauman\/lein-figwheel<\/a> with CIDER.<\/p>\n<ol><li>\n<p>Set up Figwheel as normal, but make sure <code>:cljsbuild<\/code> and <code>:figwheel<\/code> settings are\nin the root of your Leiningen project definition.<\/p>\n<\/li>\n<li>\n<p>Add these to your dev <code>:dependencies<\/code>:<\/p>\n<\/li>\n<\/ol><p>[com.cemerick\/piggieback \"0.2.1\"]\n[figwheel-sidecar \"0.5.0-2\"]<\/p>\n<pre>\n3. Add this to your dev `:repl-options`:\n```clojure\n:nrepl-middleware [cemerick.piggieback\/wrap-cljs-repl]\n<\/pre>\n<ol start=\"4\"><li>Instruct CIDER to use Figwheel in your Emacs config:<\/li>\n<\/ol><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-cljs-lein-repl \"(do (use 'figwheel-sidecar.repl-api) (start-figwheel!) (cljs-repl))\")<\/pre><\/div>\n<ol start=\"5\"><li>\n<p>Start the REPL with <code>cider-jack-in-clojurescript<\/code> (<kbd>C-c M-J<\/kbd>)<\/p>\n<\/li>\n<li>\n<p>Open a browser to the Figwheel URL so that it can connect to your application.<\/p>\n<\/li>\n<\/ol><p>You now have two nREPL connections, one for Clojure and one for ClojureScript.\nCIDER will determine which to use based on the type of file you're editing.<\/p>\n<p>You should also check out\n<a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\/wiki\/Using-the-Figwheel-REPL-within-NRepl\">Figwheel<\/a>'s\nwiki.<\/p>\n<h2><a id=\"user-content-extended-workflow\" class=\"anchor\" aria-hidden=\"true\" href=\"#extended-workflow\"><\/a>Extended workflow<\/h2>\n<p>CIDER packs a ton of extra functionality, besides basic Clojure code\nevaluation. Much of the functionality is centered around additional major modes,\nwhich provide you with convenient ways to get something done or inspect\nsomething.<\/p>\n<h3><a id=\"user-content-evaluating-clojure-code-in-the-minibuffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#evaluating-clojure-code-in-the-minibuffer\"><\/a>Evaluating Clojure code in the minibuffer<\/h3>\n<p>You can evaluate Clojure code in the minibuffer from pretty much everywhere\nby using <code>M-x cider-read-and-eval<\/code> (bound in <code>cider-mode<\/code> buffers to <code>C-c C-:<\/code>).\n<code>TAB<\/code> completion will work in the minibuffer, just as in a REPL\/source buffer.<\/p>\n<p>You can also enable <code>eldoc-mode<\/code> in the minibuffer by adding the following to your\nconfig:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'eval-expression-minibuffer-setup-hook #'eldoc-mode)<\/pre><\/div>\n<p>You can also enable <code>paredit<\/code> or <code>smartparens<\/code> for minibuffer evaluations:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'eval-expression-minibuffer-setup-hook #'paredit-mode)<\/pre><\/div>\n<h3><a id=\"user-content-macroexpansion\" class=\"anchor\" aria-hidden=\"true\" href=\"#macroexpansion\"><\/a>Macroexpansion<\/h3>\n<p>Pressing <kbd>C-c C-m<\/kbd> after some form in a source buffer or the REPL will\nresult in a new buffer, showing the macroexpansion of form in question. You'll\nhave access to additional keybindings in the macroexpansion buffer (which is\ninternally using <code>cider-macroexpansion-mode<\/code>):<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>C-c C-m<\/kbd><\/td>\n<td>Invoke <code>macroexpand-1<\/code> on the form at point and replace the original form with its expansion.  If invoked with a prefix argument, <code>macroexpand<\/code> is used instead of <code>macroexpand-1<\/code>.<\/td>\n<\/tr><tr><td><kbd>C-c M-m<\/kbd><\/td>\n<td>Invoke <code>clojure.walk\/macroexpand-all<\/code> on the form at point and replace the original form with its expansion.<\/td>\n<\/tr><tr><td><kbd>g<\/kbd><\/td>\n<td>The prior macroexpansion is performed again and the current contents of the macroexpansion buffer are replaced with the new expansion.<\/td>\n<\/tr><tr><td><kbd>C-\/<\/kbd> <kbd>C-x u<\/kbd><\/td>\n<td>Undo the last inplace expansion performed in the macroexpansion buffer.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-value-inspection\" class=\"anchor\" aria-hidden=\"true\" href=\"#value-inspection\"><\/a>Value inspection<\/h3>\n<p>Pressing <kbd>C-c C-i<\/kbd> after some form in a source buffer or the REPL will\nresult in a new buffer, showing the structure of the result of the form in question. You'll\nhave access to additional keybindings in the inspector buffer (which is\ninternally using <code>cider-inspector-mode<\/code>):<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>Tab<\/kbd> and <kbd>Shift-Tab<\/kbd><\/td>\n<td>Navigate inspectable sub-objects<\/td>\n<\/tr><tr><td><kbd>Return<\/kbd><\/td>\n<td>Inspect sub-objects<\/td>\n<\/tr><tr><td><kbd>l<\/kbd><\/td>\n<td>Pop to the parent object<\/td>\n<\/tr><tr><td><kbd>g<\/kbd><\/td>\n<td>Refresh the inspector (e.g. if viewing an atom\/ref\/agent)<\/td>\n<\/tr><tr><td><kbd>SPC<\/kbd><\/td>\n<td>Jump to next page in paginated view<\/td>\n<\/tr><tr><td><kbd>M-SPC<\/kbd><\/td>\n<td>Jump to previous page in paginated view<\/td>\n<\/tr><tr><td><kbd>s<\/kbd><\/td>\n<td>Set a new page size in paginated view<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-running-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-tests\"><\/a>Running tests<\/h3>\n<p>You can run <code>clojure.test<\/code> tests pretty quickly in CIDER. Pressing <kbd>C-c C-t\nn<\/kbd> in a source buffer or a REPL buffer will run the tests for the namespace\nyou're currently in. CIDER is smart enough to figure out the namespace\ncontaining the tests. You can also run all loaded tests with <kbd>C-c C-t\nl<\/kbd> and all tests within a project with <kbd>C-c C-t p<\/kbd> (note that this\nwill loaded <strong>all<\/strong> namespaces in your project). Using <kbd>C-c C-t t<\/kbd>, you\ncan execute only the test a point.<\/p>\n<p>All test commands are available in REPL buffers as well. There you can also use\n<kbd>,<\/kbd> there to invoke some of the testing commands.<\/p>\n<p>In the buffer displaying the test execution results (<code>*cider-test-results*<\/code>)\nyou'll have a bit of additional functionality at your disposal.<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>C-c C-t t<\/kbd><\/td>\n<td>Run test at point.<\/td>\n<\/tr><tr><td><kbd>C-c C-t n<\/kbd><\/td>\n<td>Run tests for current namespace.<\/td>\n<\/tr><tr><td><kbd>C-c C-t l<\/kbd><\/td>\n<td>Run tests for all loaded namespaces.<\/td>\n<\/tr><tr><td><kbd>C-c C-t p<\/kbd><\/td>\n<td>Run tests for all project namespaces. This loads the additional namespaces.<\/td>\n<\/tr><tr><td><kbd>C-c C-t r<\/kbd><\/td>\n<td>Re-run test failures\/errors.<\/td>\n<\/tr><tr><td><kbd>M-p<\/kbd><\/td>\n<td>Move point to previous test.<\/td>\n<\/tr><tr><td><kbd>M-n<\/kbd><\/td>\n<td>Move point to next test.<\/td>\n<\/tr><tr><td><kbd>t<\/kbd> and <kbd>M-.<\/kbd><\/td>\n<td>Jump to test definition.<\/td>\n<\/tr><tr><td><kbd>d<\/kbd><\/td>\n<td>Display diff of actual vs expected.<\/td>\n<\/tr><tr><td><kbd>e<\/kbd><\/td>\n<td>Display test error cause and stacktrace info.<\/td>\n<\/tr><\/tbody><\/table><p>Certain aspects of the test execution behavior are configurable:<\/p>\n<ul><li>\n<p>If your tests are not following the <code>some.ns-test<\/code> naming convention you can\ncustomize the variable <code>cider-test-infer-test-ns<\/code>. It should be bound to a\nfunction that takes the current ns and returns the matching test ns (which may\nbe the same as the current ns).<\/p>\n<\/li>\n<li>\n<p>If you want to view the test report regardless of whether the tests have\npassed or failed:<\/p>\n<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-test-show-report-on-success t)<\/pre><\/div>\n<h3><a id=\"user-content-navigating-stacktraces\" class=\"anchor\" aria-hidden=\"true\" href=\"#navigating-stacktraces\"><\/a>Navigating stacktraces<\/h3>\n<p>CIDER comes with a powerful solution to the problem of verbose Clojure\nstacktraces.  Stacktraces are presented in a special major mode\n(<code>cider-stacktrace-mode<\/code>), which gives you the possibility to filter out certain\nstack frames and some handy ways to navigate causes.  You'll also be able to go\nto the code in question with a single keystroke.<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>M-p<\/kbd><\/td>\n<td>Move point to previous cause<\/td>\n<\/tr><tr><td><kbd>M-n<\/kbd><\/td>\n<td>Move point to next cause<\/td>\n<\/tr><tr><td><kbd>M-.<\/kbd> and <kbd>Return<\/kbd><\/td>\n<td>Navigate to the source location (if available) for the stacktrace frame<\/td>\n<\/tr><tr><td><kbd>Tab<\/kbd><\/td>\n<td>Cycle current cause detail<\/td>\n<\/tr><tr><td><kbd>0<\/kbd> and <kbd>S-Tab<\/kbd><\/td>\n<td>Cycle all cause detail<\/td>\n<\/tr><tr><td><kbd>1<\/kbd><\/td>\n<td>Cycle cause #1 detail<\/td>\n<\/tr><tr><td><kbd>2<\/kbd><\/td>\n<td>Cycle cause #2 detail<\/td>\n<\/tr><tr><td><kbd>3<\/kbd><\/td>\n<td>Cycle cause #3 detail<\/td>\n<\/tr><tr><td><kbd>4<\/kbd><\/td>\n<td>Cycle cause #4 detail<\/td>\n<\/tr><tr><td><kbd>5<\/kbd><\/td>\n<td>Cycle cause #5 detail<\/td>\n<\/tr><tr><td><kbd>j<\/kbd><\/td>\n<td>Toggle display of java frames<\/td>\n<\/tr><tr><td><kbd>c<\/kbd><\/td>\n<td>Toggle display of clj frames<\/td>\n<\/tr><tr><td><kbd>r<\/kbd><\/td>\n<td>Toggle display of repl frames<\/td>\n<\/tr><tr><td><kbd>t<\/kbd><\/td>\n<td>Toggle display of tooling frames (e.g. compiler, nREPL middleware)<\/td>\n<\/tr><tr><td><kbd>d<\/kbd><\/td>\n<td>Toggle display of duplicate frames<\/td>\n<\/tr><tr><td><kbd>a<\/kbd><\/td>\n<td>Toggle display of all frames<\/td>\n<\/tr><\/tbody><\/table><ul><li>Error buffer stacktraces may be filtered by default. Valid filter types\ninclude <code>java<\/code>, <code>clj<\/code>, <code>repl<\/code>, <code>tooling<\/code>, and <code>dup<\/code>. Setting this to <code>nil<\/code> will\nshow all stacktrace frames.<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-stacktrace-default-filters '(tooling dup))<\/pre><\/div>\n<ul><li>Error messages may be wrapped for readability. If this value is nil, messages\nwill not be wrapped; if it is truthy but non-numeric, the default <code>fill-column<\/code>\nwill be used.<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-stacktrace-fill-column 80)<\/pre><\/div>\n<h3><a id=\"user-content-enlighten-display-local-values\" class=\"anchor\" aria-hidden=\"true\" href=\"#enlighten-display-local-values\"><\/a>Enlighten (display local values)<\/h3>\n<p>This feature displays the value of locals in realtime, as your code is being\nexecuted. This is somewhat akin to one of the features of the Light Table\neditor.<\/p>\n<ul><li>To turn it on, issue <code>M-x cider-enlighten-mode<\/code>.<\/li>\n<li>To use it, evaluate your functions one at a time (e.g., use <code>C-M-x<\/code> or <code>C-x C-e<\/code>, because <code>C-c C-k<\/code> won't work).<\/li>\n<\/ul><p>That's it! Once your code executes, the regular old buffer on the left will turn\ninto the brilliant show of lights on the right.<\/p>\n<p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/doc\/images\/enlighten-off.png\"><img src=\"https:\/\/github.com\/cichli\/cider\/raw\/master\/doc\/images\/enlighten-off.png\" alt=\"image\" \/><\/a>\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/doc\/images\/enlighten-on.png\"><img src=\"https:\/\/github.com\/cichli\/cider\/raw\/master\/doc\/images\/enlighten-on.png\" alt=\"image\" \/><\/a>\n<\/p>\n<p>To stop displaying the locals you'll have to disable <code>cider-enlighten-mode<\/code>\nand reevaluate the definitions you had instrumented previously.<\/p>\n<p>You can also trigger this on specific functions (without having to turn on the\nminor mode) by writing <code>#light<\/code> before the <code>(def<\/code> and reevaluating it.<\/p>\n<h3><a id=\"user-content-debugging\" class=\"anchor\" aria-hidden=\"true\" href=\"#debugging\"><\/a>Debugging<\/h3>\n<p>The debugger can be invoked in several ways, the simplest one is to type\n<kbd>C-u C-M-x<\/kbd>. This will take the current top-level form, place as many\nbreakpoints inside it as possible (instrument it), and then evaluate it a\nnormal. Whenever a breakpoint is reached, you'll be shown the value and asked\nfor input (see below). Note that if the current form is a <code>defn<\/code>, it will stay\ninstrumented, so the debugger will be triggered every time the function is\ncalled. To uninstrument <code>defn<\/code> (or similar forms), you just have to evaluate it\nagain as you'd normally do (e.g. with <kbd>C-M-x<\/kbd>).<\/p>\n<p>Another way to trigger the debugger is by placing breakpoints yourself. Just\nwrite <code>#break<\/code> before a form, and the debugger will popup every time that form is\nevaluated. For instance, if you hit <kbd>C-M-x<\/kbd> on the following, a\nbreakpoint is triggered every time <code>(inspector msg)<\/code> is evaluated.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn eval-msg [{:keys [inspect] :as msg}]\n  (if inspect\n    #break (inspector msg)\n    msg))<\/pre><\/div>\n<p>Instead of <code>#break<\/code> you can also write <code>#dbg<\/code> before a form, this will not only\nbreakpoint the form but also everything inside it. In the example above, this\nplaces a breakpoint around <code>(inspector msg)<\/code> and another around <code>msg<\/code>. If you've\nbeen paying attention, you may have noticed that the first option (<kbd>C-u\nC-M-x<\/kbd>) is a quick way of evaluating the current top-level form with <code>#dbg<\/code>\nin front.<\/p>\n<p>At any point, you can bring up a list of all currently instrumented <code>def<\/code>s with\nthe command <code>cider-browse-instrumented-defs<\/code>. Protocols and types can be\ninstrumented as well, but they will not be listed by this command.<\/p>\n<h4><a id=\"user-content-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#keys\"><\/a>Keys<\/h4>\n<p><code>cider-debug<\/code> tries to be consistent with\n<a href=\"http:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Edebug.html\" rel=\"nofollow\">Edebug<\/a>. So\nit makes available the following bindings while stepping through code.<\/p>\n<table><thead><tr><th>Keyboard shortcut<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>n<\/kbd><\/td>\n<td>Next step<\/td>\n<\/tr><tr><td><kbd>c<\/kbd><\/td>\n<td>Continue without stopping<\/td>\n<\/tr><tr><td><kbd>o<\/kbd><\/td>\n<td>Move out of the current sexp (like <code>up-list<\/code>)<\/td>\n<\/tr><tr><td><kbd>i<\/kbd><\/td>\n<td>Inject a value into running code<\/td>\n<\/tr><tr><td><kbd>e<\/kbd><\/td>\n<td>Eval code in current context<\/td>\n<\/tr><tr><td><kbd>l<\/kbd><\/td>\n<td>Inspect local variables<\/td>\n<\/tr><tr><td><kbd>s<\/kbd><\/td>\n<td>Show the current stack<\/td>\n<\/tr><tr><td><kbd>h<\/kbd><\/td>\n<td>Skip all sexps up to \u201chere\u201d (current position). Move the cursor before doing this.<\/td>\n<\/tr><tr><td><kbd>q<\/kbd><\/td>\n<td>Quit execution<\/td>\n<\/tr><\/tbody><\/table><p>In addition, all the usual evaluation commands (such as <kbd>C-x C-e<\/kbd> or\n<kbd>C-c M-:<\/kbd>) will use the current lexical context (local variables) while\nthe debugger is active.<\/p>\n<h3><a id=\"user-content-code-reloading\" class=\"anchor\" aria-hidden=\"true\" href=\"#code-reloading\"><\/a>Code reloading<\/h3>\n<p><code>cider-refresh<\/code> wraps\n<a href=\"https:\/\/github.com\/clojure\/tools.namespace\">clojure.tools.namespace<\/a>, and as\nsuch the same\n<a href=\"https:\/\/github.com\/clojure\/tools.namespace#reloading-code-motivation\">benefits<\/a>\nand\n<a href=\"https:\/\/github.com\/clojure\/tools.namespace#reloading-code-preparing-your-application\">caveats<\/a>\nregarding writing reloadable code also apply.<\/p>\n<p>Calling <code>cider-refresh<\/code> will cause all modified Clojure files on the classpath\nto be reloaded. You can also provide a single prefix argument to reload all\nClojure files on the classpath unconditionally, or a double prefix argument to\nfirst clear the state of the namespace tracker before reloading.<\/p>\n<p>The above three operations are analogous to\n<a href=\"http:\/\/clojure.github.io\/tools.namespace\/#clojure.tools.namespace.repl\/refresh\" rel=\"nofollow\"><code>clojure.tools.namespace.repl\/refresh<\/code><\/a>,\n<a href=\"http:\/\/clojure.github.io\/tools.namespace\/#clojure.tools.namespace.repl\/refresh-all\" rel=\"nofollow\"><code>clojure.tools.namespace.repl\/refresh-all<\/code><\/a>\nand\n<a href=\"http:\/\/clojure.github.io\/tools.namespace\/#clojure.tools.namespace.repl\/clear\" rel=\"nofollow\"><code>clojure.tools.namespace.repl\/clear<\/code><\/a>\n(followed by a normal refresh), respectively.<\/p>\n<ul><li>You can define Clojure functions to be called before reloading, and after a\nsuccessful reload, when using <code>cider-refresh<\/code>:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-refresh-before-fn \"user\/stop-system!\"\n      cider-refresh-after-fn \"user\/start-system!\")<\/pre><\/div>\n<ul><li>\n<p>These must be set to the namespace-qualified names of vars bound to functions\nof no arguments. The functions must be synchronous (blocking), and are\nexpected to be side-effecting - they will always be executed serially, without\nretries.<\/p>\n<\/li>\n<li>\n<p>By default, messages regarding the status of the in-progress reload will be\ndisplayed in the echo area after you call <code>cider-refresh<\/code>. The same\ninformation will also be recorded in the <code>*cider-refresh-log*<\/code> buffer, along\nwith anything printed to <code>*out*<\/code> or <code>*err*<\/code> by <code>cider-refresh-before-fn<\/code> and\n<code>cider-refresh-start-fn<\/code>.<\/p>\n<\/li>\n<li>\n<p>You can make the <code>*cider-refresh-log*<\/code> buffer display automatically after you\ncall <code>cider-refresh<\/code> by setting the <code>cider-refresh-show-log-buffer<\/code> variable\nto a non-nil value (this will also prevent any related messages from also\nbeing displayed in the echo area):<\/p>\n<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-refresh-show-log-buffer t)<\/pre><\/div>\n<h3><a id=\"user-content-managing-multiple-connections\" class=\"anchor\" aria-hidden=\"true\" href=\"#managing-multiple-connections\"><\/a>Managing multiple connections<\/h3>\n<p>You can connect to multiple nREPL servers using <kbd>M-x cider-jack-in<\/kbd> (or\n<code>cider-connect<\/code>) multiple times.  To close the current nREPL connection, use\n<kbd>M-x cider-quit<\/kbd>.<\/p>\n<p>CIDER maintains a list of nREPL connections and a single 'default'\nconnection. When you execute CIDER commands in a Clojure editing buffer such as\nto compile a namespace, these commands are executed against a specific\nconnection. This is controlled by the variable <code>cider-request-dispatch<\/code> - when\nit's set to <code>'dynamic<\/code> (the default), CIDER will try to infer which connection\nto use from the current project and currently visited file; when <code>'static<\/code>\ndispatch is used all requests will always be routed to the default connection\n(this was the default behavior in CIDER before 0.10).<\/p>\n<p>You can display the current nREPL connection using <kbd>C-c M-d<\/kbd>\nand rotate the default connection using <kbd>C-c M-r<\/kbd>. Another\noption for setting the default connection is to execute the command\n<kbd>M-x cider-make-connection-default<\/kbd> in the appropriate\nREPL buffer.<\/p>\n<p>To switch to the relevant REPL buffer based on the Clojure namespace\nin the current Clojure buffer, use: <kbd>C-c C-z<\/kbd>. You can then\nuse the same key combination to switch back to the Clojure buffer you\ncame from.<\/p>\n<p>The single prefix <kbd>C-u C-c C-z<\/kbd>, will switch you to the\nrelevant REPL buffer and set the namespace in that buffer based on\nnamespace in the current Clojure buffer.<\/p>\n<p>To change the designation used for CIDER buffers use <kbd>M-x\ncider-change-buffers-designation<\/kbd>. This changes the CIDER REPL\nbuffer, nREPL connection buffer and nREPL server buffer. For example\nusing <code>cider-change-buffers-designation<\/code> with the string \"foo\" would\nchange <code>*cider-repl localhost*<\/code> to <code>*cider-repl foo*<\/code>.<\/p>\n<h2><a id=\"user-content-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration\"><\/a>Configuration<\/h2>\n<p>You can certainly use CIDER without configuring it any further,\nbut here are some ways other folks are adjusting their CIDER\nexperience.<\/p>\n<h3><a id=\"user-content-basic-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-configuration\"><\/a>Basic configuration<\/h3>\n<ul><li>Enable <code>eldoc<\/code> in Clojure buffers:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-mode-hook #'eldoc-mode)<\/pre><\/div>\n<ul><li>Suppress auto-enabling of <code>cider-mode<\/code> in <code>clojure-mode<\/code> buffers, when starting\nCIDER:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-auto-mode nil)<\/pre><\/div>\n<p>By default CIDER will enable <code>cider-mode<\/code> in all <code>clojure-mode<\/code> buffers when the\nfirst CIDER connection is established. It will also add a <code>clojure-mode<\/code> hook to\nenable it on newly created <code>clojure-mode<\/code> buffers. The configuration snippet\nabove allows you to override this (somewhat non-standard) behavior.<\/p>\n<ul><li>By default, interactive commands that require a symbol (e.g. <code>cider-doc<\/code>) will\nprompt for the symbol, with the prompt defaulting to the symbol at point. You\ncan set <code>cider-prompt-for-symbol<\/code> to nil to instead try the command with the\nsymbol at point first, and only prompt if that fails (this used to be the\ndefault behavior in older CIDER releases).<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-prompt-for-symbol nil)<\/pre><\/div>\n<ul><li>Don't log communication with the nREPL server:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq nrepl-log-messages nil)<\/pre><\/div>\n<p>Basically, this will dispose of buffers like <code>*nrepl-messages connection-name*<\/code>. The communication log is invaluable for debugging CIDER\nissues, so you're generally advised to keep it around.<\/p>\n<ul><li>You can hide the <code>*nrepl-connection*<\/code> and <code>*nrepl-server*<\/code> buffers\nfrom appearing in some buffer switching commands like\n<code>switch-to-buffer<\/code>(<kbd>C-x b<\/kbd>) like this:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq nrepl-hide-special-buffers t)<\/pre><\/div>\n<p>When using <code>switch-to-buffer<\/code>, pressing <kbd>SPC<\/kbd> after the command will\nmake the hidden buffers visible. They'll always be visible in\n<code>list-buffers<\/code> (<kbd>C-x C-b<\/kbd>).<\/p>\n<ul><li>To prefer local resources to remote (tramp) ones when both are available:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-prefer-local-resources t)<\/pre><\/div>\n<ul><li>Prevent the auto-display of the REPL buffer in a separate window\nafter connection is established:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-pop-to-buffer-on-connect nil)<\/pre><\/div>\n<ul><li>\n<p>Configure whether the error buffer with stacktraces should be automatically\nshown on error:<\/p>\n<ul><li>\n<p>Don't show on error:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-show-error-buffer nil)<\/pre><\/div>\n<\/li>\n<\/ul><p>Independently of the value of <code>cider-show-error-buffer<\/code>, the error buffer is\nalways generated in the background. Use <code>cider-visit-error-buffer<\/code> to visit\nthis buffer.<\/p>\n<ul><li>\n<p>Selective strategies:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-show-error-buffer 'except-in-repl) ; or\n(setq cider-show-error-buffer 'only-in-repl)<\/pre><\/div>\n<\/li>\n<\/ul><\/li>\n<li>\n<p>To disable auto-selection of the error buffer when it's displayed:<\/p>\n<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-auto-select-error-buffer nil)<\/pre><\/div>\n<ul><li>Make <kbd>C-c C-z<\/kbd> switch to the CIDER REPL buffer in the current window:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-repl-display-in-current-window t)<\/pre><\/div>\n<ul><li>Prevent <kbd>C-c C-k<\/kbd> from prompting to save the file corresponding to\nthe buffer being loaded, if it's modified:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Don't prompt and don't save\n(setq cider-prompt-save-file-on-load nil)\n;; Just save without prompting\n(setq cider-prompt-save-file-on-load 'always-save)<\/pre><\/div>\n<ul><li>Change the result prefix for interactive evaluation (by default it's <code>=&gt; <\/code>):<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-interactive-eval-result-prefix \";; =&gt; \")<\/pre><\/div>\n<p>To remove the prefix altogether just set it to an empty string(<code>\"\"<\/code>).<\/p>\n<ul><li>CIDER can syntax highlight symbols that are known to be defined. By default,\nthis is done on symbols from the <code>clojure.core<\/code> namespace as well as macros\nfrom any namespace. If you'd like CIDER to also colorize usages of functions\nand variables from any namespace, do:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-font-lock-dynamically '(macro core function var))<\/pre><\/div>\n<ul><li>You can configure known endpoints used by the cider command offered via a\ncompleting read. This is useful if you have a list of common host\/ports you\nwant to establish remote nREPL connections to. Using an optional label is\nhelpful for identifying each host.<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-known-endpoints '((\"host-a\" \"10.10.10.1\" \"7888\") (\"host-b\" \"7888\")))<\/pre><\/div>\n<ul><li>If you are targeting the JVM and prefer a local copy of the JDK API\ndocumentation over Oracle's official copy (e.g., for\n<a href=\"http:\/\/docs.oracle.com\/javase\/8\/docs\/api\/\" rel=\"nofollow\">JavaSE 8<\/a>), per nREPL's\n<a href=\"http:\/\/docs.oracle.com\/javase\/8\/docs\/api\/\" rel=\"nofollow\"><code>javadoc-info<\/code> logic (accurate as of 29 Dec 2014)<\/a>,\nyou can arrange your project to include the <strong>root<\/strong> path of the local API doc\n(i.e., where the <code>index.html<\/code> is located) to be included on your classpath\n(i.e., where the doc HTML files can be located by\n<code>clojure.java.io\/resource<\/code>). For example, for Leiningen, with the local API\npath being <code>\/usr\/share\/doc\/java\/api\/<\/code>, put the following line in\n<code>project.clj<\/code>:<\/li>\n<\/ul><div class=\"highlight highlight-source-clojure\"><pre>:dev {:resource-paths [\"\/usr\/share\/doc\/java\/api\/\"]}<\/pre><\/div>\n<p><strong>or<\/strong> the following line in <code>$HOME\/.lein\/profiles.clj<\/code>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:user {:resource-paths [\"\/usr\/share\/doc\/java\/api\/\"]}<\/pre><\/div>\n<p>More details can be found <a href=\"https:\/\/github.com\/clojure-emacs\/cider\/issues\/930\">here<\/a>.<\/p>\n<h3><a id=\"user-content-overlays\" class=\"anchor\" aria-hidden=\"true\" href=\"#overlays\"><\/a>Overlays<\/h3>\n<p>When you evaluate code in Clojure files, the result is displayed in the buffer\nitself, in an overlay right after the evaluated code.  If you want this overlay\nto be font-locked (syntax-highlighted) like Clojure code, set the following\nvariable.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-overlays-use-font-lock t)<\/pre><\/div>\n<p>You can disable overlays entirely (and display results in the echo-area at the\nbottom) with the <code>cider-use-overlays<\/code> variable.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-use-overlays nil)<\/pre><\/div>\n<h3><a id=\"user-content-specifying-indentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#specifying-indentation\"><\/a>Specifying indentation<\/h3>\n<p>It is common for macros to require special indentation mechanisms. This is most\ncommon in macros that start with <code>do<\/code>, <code>def<\/code>, or <code>with-<\/code>.  CIDER has some\nheuristics to detect these macros, but it also lets you explicitly specify how\na macro should be indented.<\/p>\n<p>Here's a simple example of how someone would specify the indent spec for a macro\nthey've written (using an example in core):<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defmacro with-in-str\n  \"[DOCSTRING]\"\n  {:style\/indent 1}\n  [s &amp; body]\n  ...cut for brevity...)<\/pre><\/div>\n<p>And here's a more complex one:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defmacro letfn\n  \"[DOCSTRING]\"\n  {:style\/indent [1 [[:defn]] :form]}\n  [fnspecs &amp; body]\n  ...cut for brevity...)<\/pre><\/div>\n<p>Don't worry if this looks intimidating. For most macros the indent spec should\nbe either just a number, or one of the keywords <code>:defn<\/code> or <code>:form<\/code>. A full\ndescription of the spec is provided in <a href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/doc\/Indent-Spec.md\">Indent-Spec.md<\/a>.<\/p>\n<p>If you <em>don't<\/em> want to use this feature, you can disable it by setting\n<code>cider-dynamic-indentation<\/code> to <code>nil<\/code> in your Emacs init file.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-dynamic-indentation nil)<\/pre><\/div>\n<h3><a id=\"user-content-minibuffer-completion\" class=\"anchor\" aria-hidden=\"true\" href=\"#minibuffer-completion\"><\/a>Minibuffer completion<\/h3>\n<p>Out-of-the box CIDER uses the standard <code>completing-read<\/code> Emacs mechanism. While\nit's not fancy it certainly gets the job done (just press <code>TAB<\/code>). There are,\nhowever, ways to improve upon the standard completion if you wish to.<\/p>\n<h4><a id=\"user-content-icomplete\" class=\"anchor\" aria-hidden=\"true\" href=\"#icomplete\"><\/a>icomplete<\/h4>\n<p><code>icomplete<\/code> is bundled with Emacs and enhances the default minibuffer completion:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'icomplete)<\/pre><\/div>\n<h4><a id=\"user-content-ido\" class=\"anchor\" aria-hidden=\"true\" href=\"#ido\"><\/a>ido<\/h4>\n<p><code>ido<\/code> is also bundled with Emacs and offers more features than <code>icomplete<\/code>.\nIf you are using <code>ido<\/code>, be sure to use both <code>ido-everywhere<\/code>\nand <a href=\"https:\/\/github.com\/DarwinAwardWinner\/ido-ubiquitous\"><code>ido-ubiquitous<\/code><\/a>.\nYou might also want to install <a href=\"https:\/\/github.com\/lewang\/flx\"><code>ido-flex<\/code><\/a>.<\/p>\n<h3><a id=\"user-content-auto-completion\" class=\"anchor\" aria-hidden=\"true\" href=\"#auto-completion\"><\/a>Auto-completion<\/h3>\n<p>CIDER users are advised to use <a href=\"http:\/\/company-mode.github.io\/\" rel=\"nofollow\"><code>company-mode<\/code><\/a>\nto enable auto-completion inside of source code and REPL buffers.  To install\n<code>company-mode<\/code> do:<\/p>\n<p><code>M-x package-install &lt;RET&gt; company &lt;RET&gt;<\/code><\/p>\n<p>After installation, company can be turned on  globally, like so --<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(global-company-mode)<\/pre><\/div>\n<p>-- or through mode-specific hooks:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-repl-mode-hook #'company-mode)\n(add-hook 'cider-mode-hook #'company-mode)<\/pre><\/div>\n<p>When <code>company-mode<\/code> is thus enabled, it will receive completion information\nfrom <code>cider-complete-at-point<\/code>, and requires no additional setup or plugins.<\/p>\n<p>If you'd prefer to trigger completions manually you can add this to you config:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq company-idle-delay nil) ; never start completions automatically\n(global-set-key (kbd \"M-TAB\") #'company-complete) ; use M-TAB, a.k.a. C-M-i, as manual trigger<\/pre><\/div>\n<p>To make <code>TAB<\/code> complete, without losing the ability to manually indent, you can\nadd this to your config:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(global-set-key (kbd \"TAB\") #'company-indent-or-complete-common)<\/pre><\/div>\n<p><code>company-indent-or-complete-common<\/code> is available only in <code>company-mode<\/code> 0.9+ (at\nthe time of this writing it's still in development).<\/p>\n<h4><a id=\"user-content-migrating-from-auto-complete-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#migrating-from-auto-complete-mode\"><\/a>Migrating from <code>auto-complete-mode<\/code><\/h4>\n<ul><li>\n<p>Disable <code>ac-cider-setup<\/code> or <code>ac-nrepl-setup<\/code> from running on CIDER hooks<\/p>\n<\/li>\n<li>\n<p>Remove <code>cider-mode<\/code> and <code>cider-repl-mode<\/code> from the <code>ac-modes<\/code> list<\/p>\n<\/li>\n<\/ul><h4><a id=\"user-content-completion-annotations\" class=\"anchor\" aria-hidden=\"true\" href=\"#completion-annotations\"><\/a>Completion annotations<\/h4>\n<p>Completion candidates will be annotated by default with an abbreviation\ncorresponding to their type, and (contextually) their ns. The function used to\nformat the annotation can be configured by <code>cider-annotate-completion-function.<\/code>\nThe abbreviations used are configured by <code>cider-completion-annotations-alist<\/code>\nand the context in which their namespace is included is configured by\n<code>cider-completion-annotations-include-ns.<\/code><\/p>\n<p>Completion annotations can be disabled by setting\n<code>cider-annotate-completion-candidates<\/code> to nil.<\/p>\n<p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/screenshots\/completion-annotations.png\"><img src=\"https:\/\/github.com\/cichli\/cider\/raw\/master\/screenshots\/completion-annotations.png\" width=\"400\" alt=\"image\" \/><\/a>\n<\/p>\n<h3><a id=\"user-content-pretty-printing\" class=\"anchor\" aria-hidden=\"true\" href=\"#pretty-printing\"><\/a>Pretty-printing<\/h3>\n<p>You can configure the function used by CIDER for pretty-printing evaluation\nresults and other data using the <code>cider-pprint-fn<\/code> option.<\/p>\n<p>This can be one of three values (defaults to <code>pprint<\/code>):<\/p>\n<ul><li>\n<p><code>pprint<\/code> to use the built-in <code>clojure.pprint\/pprint<\/code>.<\/p>\n<\/li>\n<li>\n<p><code>fipp<\/code> to use the\n<a href=\"https:\/\/github.com\/brandonbloom\/fipp\">Fast Idiomatic Pretty-Printer<\/a>. This is\napproximately 5-10x faster than <code>clojure.core\/pprint<\/code>.<\/p>\n<\/li>\n<li>\n<p><code>puget<\/code> to use <a href=\"https:\/\/github.com\/greglook\/puget\">Puget<\/a>, which builds on\nFipp to provide a\n<a href=\"https:\/\/github.com\/greglook\/puget#canonical-representation\">canonical serialization<\/a>\nof data, at a slight performance cost.<\/p>\n<\/li>\n<\/ul><p>Alternatively, <code>cider-pprint-fn<\/code> can be set to the namespace-qualified name of a\nClojure function that takes a single argument and will pretty-print the value of\nsaid argument to <code>*out*<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq cider-pprint-fn \"user\/my-pprint\")<\/pre><\/div>\n<p>This function must be resolvable by CIDER at the time it is called (i.e. its\ncontaining namespace must have already been required).<\/p>\n<p>CIDER will bind <code>*print-length*<\/code>, <code>*print-level*<\/code>, <code>*print-meta*<\/code>, and\n<code>clojure.pprint\/*print-right-margin*<\/code> when calling the pretty-printing\nfunction - the function you provide is expected to respect these options.<\/p>\n<h3><a id=\"user-content-integration-with-other-modes\" class=\"anchor\" aria-hidden=\"true\" href=\"#integration-with-other-modes\"><\/a>Integration with other modes<\/h3>\n<ul><li>Enabling <code>CamelCase<\/code> support for editing commands(like\n<code>forward-word<\/code>, <code>backward-word<\/code>, etc) in the REPL is quite useful since\nwe often have to deal with Java class and method names. The built-in\nEmacs minor mode <code>subword-mode<\/code> provides such functionality:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-repl-mode-hook #'subword-mode)<\/pre><\/div>\n<ul><li>The use of <a href=\"http:\/\/mumble.net\/~campbell\/emacs\/paredit.html\" rel=\"nofollow\">paredit<\/a>\nwhen editing Clojure (or any other Lisp) code is highly\nrecommended.  You're probably using it already in your <code>clojure-mode<\/code>\nbuffers (if you're not you probably should). You might also want to\nenable <code>paredit<\/code> in the REPL buffer as well:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-repl-mode-hook #'paredit-mode)<\/pre><\/div>\n<ul><li><a href=\"https:\/\/github.com\/Fuco1\/smartparens\">smartparens<\/a> is an excellent\nalternative to paredit. Many Clojure hackers have adopted it\nrecently and you might want to give it a try as well. To enable\n<code>smartparens<\/code> in the REPL buffer use the following code:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-repl-mode-hook #'smartparens-strict-mode)<\/pre><\/div>\n<ul><li><a href=\"https:\/\/github.com\/Fanael\/rainbow-delimiters\">RainbowDelimiters<\/a> is a\nminor mode which highlights parentheses, brackets, and braces\naccording to their depth. Each successive level is highlighted in a\ndifferent color. This makes it easy to spot matching delimiters,\norient yourself in the code, and tell which statements are at a\ngiven depth. Assuming you've already installed RainbowDelimiters you can\nenable it in the REPL like this:<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'cider-repl-mode-hook #'rainbow-delimiters-mode)<\/pre><\/div>\n<ul><li>\n<p><a href=\"http:\/\/cx4a.org\/software\/auto-complete\/\" rel=\"nofollow\">auto-complete<\/a> is a popular Emacs\ninteractive auto-completion framework. <a href=\"https:\/\/github.com\/clojure-emacs\/ac-cider\">ac-cider<\/a>\nprovides a completion source for auto-complete-mode, including, where CIDER provides it,\npop-up documentation for completed symbols.<\/p>\n<\/li>\n<li>\n<p><a href=\"https:\/\/github.com\/hchbaw\/eval-sexp-fu.el\">eval-sexp-fu<\/a> provides some visual\nfeedback when evaluating expressions. <a href=\"https:\/\/github.com\/clojure-emacs\/cider-eval-sexp-fu\">cider-eval-sexp-fu<\/a> provides\nCIDER integration for <code>eval-sexp-fu<\/code>.<\/p>\n<\/li>\n<\/ul><div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'cider-eval-sexp-fu)<\/pre><\/div>\n<h2><a id=\"user-content-quick-reference\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-reference\"><\/a>Quick reference<\/h2>\n<p>A single-page quick reference PDF for CIDER commands is available\n<a href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/doc\/cider-refcard.pdf\">here<\/a>. This PDF can be created manually by running\n<code>pdflatex<\/code> on <a href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/doc\/cider-refcard.tex\">the CIDER refcard LaTeX file<\/a>.<\/p>\n<h2><a id=\"user-content-caveats\" class=\"anchor\" aria-hidden=\"true\" href=\"#caveats\"><\/a>Caveats<\/h2>\n<h3><a id=\"user-content-clojurescript-limitations\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojurescript-limitations\"><\/a>ClojureScript limitations<\/h3>\n<p>Currently, the following features are not supported for ClojureScript\ndevelopment:<\/p>\n<ul><li>Grimoire lookup<\/li>\n<li>Reloading<\/li>\n<li>Running tests<\/li>\n<li>Tracing<\/li>\n<li>Debugging (check out <a href=\"https:\/\/github.com\/clojure-emacs\/cider\/issues\/1416\">this ticket<\/a> dedicated to porting the debugger to ClojureScript)<\/li>\n<li>Enlighten<\/li>\n<\/ul><p>There is currently no support for both Clojure and ClojureScript evaluation in\nthe same nREPL session. If Piggieback is active, code evaluation and all\nfeatures will assume ClojureScript.<\/p>\n<h4><a id=\"user-content-var-metadata\" class=\"anchor\" aria-hidden=\"true\" href=\"#var-metadata\"><\/a>Var Metadata<\/h4>\n<p>Currently var metadata about the location of the var's definition within the\nClojureScript source code (file, line &amp; column) is set only when evaluating the\nentire source buffer (<kbd>C-c C-k<\/kbd>). All other interactive code evaluation\ncommands (e.g. <kbd>C-c C-e<\/kbd>) don't set this metadata and you won't be able\nto use commands like <code>find-var<\/code> on such vars.  This is a limitation of nREPL and\npiggieback, that's beyond CIDER. You can find some discussions on the subject\n<a href=\"http:\/\/dev.clojure.org\/jira\/browse\/NREPL-59\" rel=\"nofollow\">here<\/a> and\n<a href=\"https:\/\/github.com\/clojure-emacs\/cider\/issues\/830\">here<\/a>.<\/p>\n<h3><a id=\"user-content-microsoft-windows\" class=\"anchor\" aria-hidden=\"true\" href=\"#microsoft-windows\"><\/a>Microsoft Windows<\/h3>\n<h4><a id=\"user-content-line-separators\" class=\"anchor\" aria-hidden=\"true\" href=\"#line-separators\"><\/a>Line separators<\/h4>\n<p>On Microsoft Windows the JVM default line separator string is <code>\\r\\n<\/code>\nwhich can appear in Emacs as <code>^M<\/code> characters at the end of lines\nprinted out by the JVM. One option is to set the\n<code>buffer-display-table<\/code> to not show these characters as detailed\n<a href=\"http:\/\/stackoverflow.com\/questions\/10098925\/m-character-showing-in-clojure-slime-repl\/11787550#11787550\" rel=\"nofollow\">here<\/a>\n(changing <code>slime-repl-mode-hook<\/code> to\n<code>cider-repl-mode-hook<\/code>). Alternatively, setting the system property\n<code>line.separator<\/code> to <code>\\n<\/code> at JVM startup will stop the carriage return\nfrom being printed and will fix output in all cider buffers. To do so\nadd <code>\"-Dline.separator=\\\"\\n\\\"\"<\/code> to <code>:jvm-opts<\/code> in\n<code>~\/.lein\/profiles.clj<\/code>.<\/p>\n<h4><a id=\"user-content-definition-lookup-in-jar-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#definition-lookup-in-jar-files\"><\/a>Definition lookup in jar files<\/h4>\n<p>In order for source lookup commands to work with <code>.jar<\/code> files you'll need to\ninstall either <a href=\"http:\/\/www.7-zip.org\/\" rel=\"nofollow\">7zip<\/a> or <code>pkunzip<\/code> and add its\ninstallation folder to Emacs's <code>exec-path<\/code>. Here's an example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'exec-path \"C:\/Program Files\/7-Zip\")<\/pre><\/div>\n<h3><a id=\"user-content-powershellel\" class=\"anchor\" aria-hidden=\"true\" href=\"#powershellel\"><\/a>powershell.el<\/h3>\n<p>The powershell inferior shell mode truncates CIDER's REPL output when\nloaded. As a workaround remove<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'powershell)<\/pre><\/div>\n<p>from your Emacs config.<\/p>\n<h3><a id=\"user-content-clojureclr-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojureclr-support\"><\/a>ClojureCLR Support<\/h3>\n<p>CIDER currently doesn't support ClojureCLR. The reasons for this are the following:<\/p>\n<ul><li>nREPL itself runs only on the JVM (because it leverages Java APIs\ninternally). There's an\n<a href=\"https:\/\/github.com\/clojure\/clr.tools.nrepl\">nREPL port for ClojureCLR<\/a>, but\nit's not actively maintained and it doesn't behave like the Clojure nREPL.<\/li>\n<li><code>cider-nrepl<\/code> uses a lot of Java code internally itself.<\/li>\n<\/ul><p>Those issues are not insurmountable, but are beyond the scope of our current roadmap.\nIf someone would like to tackle them, we'd be happy to provide assitance.<\/p>\n<h3><a id=\"user-content-injecting-dependencies-and-leiningen-pedantic-abort-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#injecting-dependencies-and-leiningen-pedantic-abort-mode\"><\/a>Injecting dependencies and Leiningen pedantic: abort mode<\/h3>\n<p>Because injection currently creates an override of <code>tools.nrepl<\/code> dependency that\nLeingingen also pulls in starting up the REPL will fail if <code>:pedantic? :abort<\/code>\nis set. Either remove the <code>:pedantic? :abort<\/code> setting or switch off injecting\nthe dependencies with setting <code>cider-inject-dependencies-at-jack-in<\/code> to nil and\nprovide the dependencies by editing your <code>~\/.lein\/profiles.clj<\/code> as described in\nthe <a href=\"#setting-up-a-standalone-repl\">standalone REPL<\/a> section.<\/p>\n<h2><a id=\"user-content-troubleshooting\" class=\"anchor\" aria-hidden=\"true\" href=\"#troubleshooting\"><\/a>Troubleshooting<\/h2>\n<p>In case you run into issues here are a few tips that can help you diagnose the\nproblem.<\/p>\n<p>Generally, it's not a bad idea to configure Emacs to spit the backtrace on error\n(instead of just logging the error in the <code>*Messages*<\/code> buffer. You can toggle\nthis behavior by using <code>M-x toggle-debug-on-error<\/code>.<\/p>\n<h3><a id=\"user-content-repl-not-starting\" class=\"anchor\" aria-hidden=\"true\" href=\"#repl-not-starting\"><\/a>REPL not starting<\/h3>\n<p>Make sure that your CIDER version matches your <code>cider-nrepl<\/code> version. Check\nthe contents of the <code>*Messages*<\/code> buffer for CIDER-related errors. You should\nalso check the nREPL messages passed between CIDER and nREPL in\n<code>*nrepl-messages*<\/code>. If you don't see anything useful there it's time to bring\nout the big guns.<\/p>\n<h4><a id=\"user-content-debugging-the-repl-init\" class=\"anchor\" aria-hidden=\"true\" href=\"#debugging-the-repl-init\"><\/a>Debugging the REPL init<\/h4>\n<p>Emacs features a super powerful built-in\n<a href=\"http:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Edebug.html\" rel=\"nofollow\">Emacs Lisp debugger<\/a>\nand using it is the best way to diagnose problems of any kind. To debug CIDER's\nREPL initialization it's a good idea to hook into one of its entry points. Add a\nbreakpoint to <code>cider-make-repl<\/code> (<code>C-u C-M-x<\/code>, while in its body). Next time you\nstart CIDER you'll be dropped in the debugger and you can step forward until you\nfind the problem.<\/p>\n<p>Here's a <a href=\"https:\/\/www.youtube.com\/watch?v=odkYXXYOxpo\" rel=\"nofollow\">great crash course<\/a> on\nusing the debugger.<\/p>\n<h3><a id=\"user-content-missing-nrepl-messages-buffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#missing-nrepl-messages-buffer\"><\/a>Missing <code>*nrepl-messages*<\/code> buffer<\/h3>\n<p>Check the value of <code>nrepl-log-messages<\/code>. It should be non-nil.<\/p>\n<h3><a id=\"user-content-cider-debug-complains-that-it-failed-to-instrument-\" class=\"anchor\" aria-hidden=\"true\" href=\"#cider-debug-complains-that-it-failed-to-instrument-\"><\/a><code>cider-debug<\/code> complains that it \u201cfailed to instrument ...\u201d<\/h3>\n<p>In the REPL buffer, issue the following.<\/p>\n<pre>your.namespace&gt; (ns cider.nrepl.middleware.util.instrument)\ncider.nrepl.middleware.util.instrument&gt; (def verbose-debug true)\n<\/pre>\n<p>This will cause cider to print extensive information on the REPL buffer when you\ntry to debug an expression (e.g., with <code>C-u C-M-x<\/code>). <a href=\"https:\/\/github.com\/clojure-emacs\/cider-repl\/issues\/new\">File an issue<\/a>\nand copy this information.<\/p>\n<h3><a id=\"user-content-warning-saying-you-have-to-use-nrepl-0212\" class=\"anchor\" aria-hidden=\"true\" href=\"#warning-saying-you-have-to-use-nrepl-0212\"><\/a>Warning saying you have to use nREPL 0.2.12+<\/h3>\n<p>CIDER currently requires at least nREPL 0.2.12 to work properly (there were some\nnasty bugs in older version and no support tracking where some var was defined\nin the source code). Leiningen users can add this to their <code>profiles.clj<\/code> to\nforce the proper dependency:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:repl {:dependencies [[org.clojure\/tools.nrepl \"0.2.12\"]]}}<\/pre><\/div>\n<p>Make sure you add the newer nREPL dependency to the <code>:dependencies<\/code> key instead\nof <code>:plugins<\/code> (where <code>cider-nrepl<\/code> Lein plugin resides). That's a pretty common\nmistake.<\/p>\n<p>Generally you're advised to use the newest nREPL with CIDER, as bugs get fixed\nin pretty much every release.<\/p>\n<p>Note, that running <code>cider-jack-in<\/code> from outside the scope of a project will\nresult in the <strong>older (0.2.6) nREPL dependency being used<\/strong> (at least on Leiningen\n2.5.1). This is likely a Leiningen bug.<\/p>\n<h3><a id=\"user-content-missing-clojure--function-after-cider-update\" class=\"anchor\" aria-hidden=\"true\" href=\"#missing-clojure--function-after-cider-update\"><\/a>Missing clojure-... function after CIDER update<\/h3>\n<p>Most likely you've updated CIDER, without updating <code>clojure-mode<\/code> as well.<\/p>\n<p>CIDER depends on <code>clojure-mode<\/code> and you should always update them together, as\nthe latest CIDER version might depend on functionality present only in the latest\n<code>clojure-mode<\/code> version.<\/p>\n<h3><a id=\"user-content-i-upgraded-cider-using-packageel-and-it-broke\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-upgraded-cider-using-packageel-and-it-broke\"><\/a>I upgraded CIDER using <code>package.el<\/code> and it broke<\/h3>\n<p>The built-in package manager isn't perfect and sometimes it messes up.  If you\njust updated and encountered an error you should try the following before\nopening an issue: Go into the <code>.emacs.d\/elpa<\/code> directory, delete any folders\nrelated to CIDER, restart Emacs and then re-install the missing packages.  Note\nthat the order here matters.<\/p>\n<h3><a id=\"user-content-i-upgraded-cider-using-packageel-and-nothing-changed\" class=\"anchor\" aria-hidden=\"true\" href=\"#i-upgraded-cider-using-packageel-and-nothing-changed\"><\/a>I upgraded CIDER using <code>package.el<\/code> and nothing changed<\/h3>\n<p>Emacs doesn't load the new files, it only installs them on disk.  To see the\neffect of changes you have to restart Emacs.<\/p>\n<h3><a id=\"user-content-cider-should-inject-its-dependencies-but-i-still-get-warnings-when-i-use-cider-jack-in\" class=\"anchor\" aria-hidden=\"true\" href=\"#cider-should-inject-its-dependencies-but-i-still-get-warnings-when-i-use-cider-jack-in\"><\/a>CIDER should inject its dependencies but I still get warnings when I use <code>cider-jack-in<\/code><\/h3>\n<p>Injecting the depencies should override the old settings in the <code>~\/lein\/profiles.clj<\/code> file however only if those settings were done in the <code>:repl<\/code> profile. If the <code>:user<\/code> profile is used overriding does not happen. To fix this issue stop your REPL, remove CIDER and <code>tools.nrepl<\/code> dependencies from your <code>~\/.lein\/profiles.clj<\/code> and start the REPL again.<\/p>\n<h2><a id=\"user-content-presentations\" class=\"anchor\" aria-hidden=\"true\" href=\"#presentations\"><\/a>Presentations<\/h2>\n<p>If you're interested in more details about CIDER's history and architecture you\ncan check out the Clojure\/conj presentation\n<a href=\"https:\/\/www.youtube.com\/watch?v=4X-1fJm25Ww&amp;list=PLZdCLR02grLoc322bYirANEso3mmzvCiI&amp;index=6\" rel=\"nofollow\">The Evolution of the Emacs tooling for Clojure<\/a>\nand the <a href=\"http:\/\/blog.cognitect.com\/cognicast\/080\" rel=\"nofollow\">Cognicast's episode on CIDER<\/a>.\nThere's also a\n<a href=\"https:\/\/skillsmatter.com\/skillscasts\/7225-cider-the-journey-so-far-and-the-road-ahead\" rel=\"nofollow\">ClojureX 2015 presentation<\/a>\ndedicated to CIDER 0.9 and 0.10 and the future of the project.<\/p>\n<h2><a id=\"user-content-changelog\" class=\"anchor\" aria-hidden=\"true\" href=\"#changelog\"><\/a>Changelog<\/h2>\n<p>An extensive changelog is available <a href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/CHANGELOG.md\">here<\/a>.<\/p>\n<h2><a id=\"user-content-team\" class=\"anchor\" aria-hidden=\"true\" href=\"#team\"><\/a>Team<\/h2>\n<h2><a id=\"user-content-release-policy\" class=\"anchor\" aria-hidden=\"true\" href=\"#release-policy\"><\/a>Release policy<\/h2>\n<p>We\u2019re following <a href=\"http:\/\/semver.org\/\" rel=\"nofollow\">SemVer<\/a> (as much as one can be\nfollowing it when the major version is 0). At this point bumps of the\nminor (second) version number are considered major releases and always\ninclude new features or significant changes to existing features. API\ncompatibility between major releases is not a (big) concern (although we try\nto break the API rarely and only for a good reason).<\/p>\n<p>The development cycle for the next major\nrelease starts immediately after the previous one has been\nshipped. Bugfix\/point releases (if any) address only serious bugs and\nnever contain new features.<\/p>\n<p>The versions of CIDER and <code>cider-nrepl<\/code> are always kept in sync. If you're\ntracking the <code>master<\/code> branch of CIDER, you should also be tracking the <code>master<\/code>\nbranch of <code>cider-nrepl<\/code>.<\/p>\n<h2><a id=\"user-content-logo\" class=\"anchor\" aria-hidden=\"true\" href=\"#logo\"><\/a>Logo<\/h2>\n<p>CIDER's logo was created by <a href=\"https:\/\/github.com\/ndr-qef\">@ndr-qef<\/a>. You can find\nthe logo in various formats\n<a href=\"https:\/\/github.com\/clojure-emacs\/cider\/tree\/master\/logo\">here<\/a>.<\/p>\n<p>The logo is licensed under a\n<a href=\"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/deed.en_GB\" rel=\"nofollow\">Creative Commons Attribution-NonCommercial 4.0 International License<\/a>.<\/p>\n<h2><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<h3><a id=\"user-content-discussion\" class=\"anchor\" aria-hidden=\"true\" href=\"#discussion\"><\/a>Discussion<\/h3>\n<p>For questions, suggestions and support refer to our\n<a href=\"https:\/\/groups.google.com\/forum\/#!forum\/cider-emacs\" rel=\"nofollow\">official mailing list<\/a> ,\nthe Freenode channel <code>#clojure-emacs<\/code>, <code>#cider<\/code> on\n<a href=\"https:\/\/clojurians.slack.com\/\" rel=\"nofollow\">slack<\/a> or our\n<a href=\"https:\/\/gitter.im\/clojure-emacs\/cider\" rel=\"nofollow\">gitter channel<\/a>. StackOverflow users\nshould use the <a href=\"http:\/\/stackoverflow.com\/questions\/tagged\/cider\" rel=\"nofollow\">cider<\/a> tag\n(ideally combined with the tags <code>emacs<\/code> and <code>clojure<\/code>).<\/p>\n<p>Please, don't report\nissues there, as this makes them harder to track.<\/p>\n<h3><a id=\"user-content-issues\" class=\"anchor\" aria-hidden=\"true\" href=\"#issues\"><\/a>Issues<\/h3>\n<p>Report issues and suggest features and improvements on the\n<a href=\"https:\/\/github.com\/clojure-emacs\/cider\/issues\">GitHub issue tracker<\/a>. Don't ask\nquestions on the issue tracker - the mailing list and the IRC channel are the\nplaces for questions.<\/p>\n<p>If you want to file a bug, please clone this repo and provide sufficient details\nto reproduce the issue. Start by running <code>make run-cider<\/code>.  This will bring up\nEmacs with only the latest version of CIDER loaded. By starting fresh, with the\nlatest code, we can ensure that the problem at hand isn't already fixed or\ncaused by interactions with other packages.<\/p>\n<h3><a id=\"user-content-patches\" class=\"anchor\" aria-hidden=\"true\" href=\"#patches\"><\/a>Patches<\/h3>\n<p>Patches in any form are always welcome! GitHub pull requests are even better! :-)<\/p>\n<p>Before submitting a patch or a pull request make sure all tests are\npassing and that your patch is in line with the <a href=\"https:\/\/github.com\/cichli\/cider\/blob\/master\/.github\/CONTRIBUTING.md\">contribution\nguidelines<\/a>.<\/p>\n<h3><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h3>\n<p>Consider improving and extending the <a href=\"https:\/\/github.com\/clojure-emacs\/cider\/wiki\">community wiki<\/a>.<\/p>\n<h3><a id=\"user-content-donations\" class=\"anchor\" aria-hidden=\"true\" href=\"#donations\"><\/a>Donations<\/h3>\n<p>You can support the development of CIDER, clojure-mode and inf-clojure via\n<a href=\"https:\/\/salt.bountysource.com\/teams\/cider\" rel=\"nofollow\">Salt<\/a>,\n<a href=\"https:\/\/www.gratipay.com\/cider\" rel=\"nofollow\">Gratipay<\/a> and PayPal.<\/p>\n<p><a href=\"https:\/\/gratipay.com\/cider\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/29139a6ba3f8f34cd6b8b8216403ab97c5ce3fe7\/68747470733a2f2f63646e2e7261776769742e636f6d2f67726174697061792f67726174697061792d62616467652f322e312e332f646973742f67726174697061792e706e67\" alt=\"Support via Gratipay\" data-canonical-src=\"https:\/\/cdn.rawgit.com\/gratipay\/gratipay-badge\/2.1.3\/dist\/gratipay.png\" \/><\/a><\/p>\n<p><a href=\"https:\/\/www.paypal.com\/cgi-bin\/webscr?cmd=_s-xclick&amp;hosted_button_id=GRQKNBM6P8VRQ\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e14c85b542e06215f7e56c0763333ef1e9b9f9b7\/68747470733a2f2f7777772e70617970616c6f626a656374732e636f6d2f656e5f55532f692f62746e2f62746e5f646f6e6174655f534d2e676966\" alt=\"Paypal\" data-canonical-src=\"https:\/\/www.paypalobjects.com\/en_US\/i\/btn\/btn_donate_SM.gif\" \/><\/a><\/p>\n<h3><a id=\"user-content-running-the-tests-in-batch-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-the-tests-in-batch-mode\"><\/a>Running the tests in batch mode<\/h3>\n<p>Install <a href=\"https:\/\/github.com\/cask\/cask\">cask<\/a> if you haven't\nalready, then:<\/p>\n<pre>$ cd \/path\/to\/cider\n$ cask\n<\/pre>\n<p>Run all tests with:<\/p>\n<pre>$ make test\n<\/pre>\n<p>(Note: tests may not run correctly inside Emacs' <code>shell-mode<\/code> buffers. Running\nthem in a terminal is recommended.)<\/p>\n<p>You can also check for the presence of byte-compilation warnings in batch mode:<\/p>\n<pre>$ make test-bytecomp\n<\/pre>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2012-2016 Tim King, Phil Hagelberg, Bozhidar Batsov, Hugo\nDuncan, Steve Purcell and\n<a href=\"https:\/\/github.com\/clojure-emacs\/cider\/contributors\">contributors<\/a>.<\/p>\n<p>Distributed under the GNU General Public License, version 3<\/p>\n<\/article>","created_at":"2018-11-07T16:45:57+0000","updated_at":"2018-11-07T16:45:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":50,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1759291?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":363,"title":"BIND Open Source DNS Server","url":"https:\/\/www.isc.org\/downloads\/bind\/","content":"<p>BIND implements the DNS protocols. The DNS protocols are part of the core Internet standards. They specify the process by which one computer can find another computer on the basis of its name. The BIND\u00a0software distribution contains all of the software necessary for asking and answering name service questions.<\/p><p>The BIND software distribution has three parts:<\/p><h4>1. Domain Name Resolver<\/h4><p>A\u00a0<strong>resolver<\/strong>\u00a0is a program that resolves questions about names by sending those questions to appropriate servers and responding appropriately to the servers\u2019 replies. In the most common application, a web browser uses a local stub resolver library on the same computer to look up names in the DNS. That stub resolver is part of the operating system. (Many operating system distributions use the BIND resolver library.) The stub resolver usually will forward queries to a caching resolver, a server or group of servers on the network dedicated to DNS services. Those resolvers will send\u00a0queries to one or multiple authoritative servers in order to find the IP address for that DNS name.<\/p><h4>2. Domain Name Authority server<\/h4><p>An\u00a0<strong>authoritative<\/strong>\u00a0DNS server\u00a0answers requests from resolvers, using information about the domain names it is authoritative for. \u00a0You can provide DNS services on the Internet by installing this software on a server and giving it information about your domain names.<\/p><h4>3. Tools<\/h4><p>We include a number of diagnostic and operational tools. Some of them, such as the popular DIG\u00a0tool, are not specific to BIND and can be used with any DNS server.<\/p>","created_at":"2018-11-07T16:45:56+0000","updated_at":"2018-11-07T16:45:56+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.isc.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":362,"title":"Chapter 2 DNS Concepts","url":"http:\/\/zytrax.com\/books\/dns\/ch2\/","content":"<p>If you already understand what DNS is and does and how it fits into the greater scheme of things - skip this chapter.<\/p><ol class=\"o-n\"><li><a href=\"#history\" class=\"t-db\">2.1 A brief History of Name Servers<\/a><\/li>\n<li><a href=\"#concepts\" class=\"t-db\">2.2 DNS Concepts &amp; Implementation<\/a>\n  <ol class=\"o-n\"><li><a href=\"#overview\" class=\"t-db\">2.2.1 DNS Overview<\/a><\/li>\n    <li><a href=\"#domains\" class=\"t-db\">2.2.2 Domains and Delegation<\/a><\/li>\n    <li><a href=\"#structure\" class=\"t-db\">2.2.3 DNS Organization and Structure<\/a><\/li>\n    <li><a href=\"#components\" class=\"t-db\">2.2.4 DNS System Components<\/a><\/li>\n    <li><a href=\"#zones\" class=\"t-db\">2.2.5 Zones and Zone Files<\/a><\/li>\n    <li><a href=\"#queries\" class=\"t-db\">2.2.6 DNS Queries<\/a>\n      <ol class=\"o-n\"><li><a href=\"#recursive\" class=\"t-db\">2.2.6.1 Recursive Queries<\/a><\/li>\n      <li><a href=\"#iterative\" class=\"t-db\">2.2.6.2 Iterative Queries<\/a><\/li>\n\t\t  <li><a href=\"#reverse\" class=\"t-db\">2.2.6.3 Inverse Queries<\/a><\/li>\n\t\t\t<\/ol><\/li>\n    <li><a href=\"#update\" class=\"t-db\">2.2.7 Zone Updates<\/a>\n\t\t  <ol class=\"o-n\"><li><a href=\"#axfr-update\" class=\"t-db\">2.2.7.1 Full Zone Transfer (AXFR)<\/a><\/li>\n\t\t  <li><a href=\"#ixfr-update\" class=\"t-db\">2.2.7.2 Incremental Zone Transfer (IXFR)<\/a><\/li>\n\t\t  <li><a href=\"#notify\" class=\"t-db\">2.2.7.3 Notify (NOTIFY)<\/a><\/li>\n\t\t  <li><a href=\"#dyn-update\" class=\"t-db\">2.2.7.4 Dynamic Zone Updates<\/a><\/li>\n\t\t  <li><a href=\"#dyn-alt\" class=\"t-db\">2.2.7.5 Alternative Dynamic DNS Approaches<\/a><\/li>\n\t\t  <\/ol><\/li>\n\t\t<\/ol><\/li>\n<li><a href=\"#security\" class=\"t-db\">2.3 DNS Security Overview<\/a>\n  <ol class=\"o-n\"><li><a href=\"#security-threats\" class=\"t-db\">2.3.1 Security Threats<\/a><\/li>\n  <li><a href=\"#security-types\" class=\"t-db\">2.3.2 Security Types<\/a><\/li>\n\t<li><a href=\"#security-local\" class=\"t-db\">2.3.3 Local Security<\/a><\/li>\n\t<li><a href=\"#security-server\" class=\"t-db\">2.3.4 Server-Server (TSIG Transactions)<\/a><\/li>\n\t<li><a href=\"#security-clients\" class=\"t-db\">2.3.5 Server-Client (DNSSEC)<\/a><\/li>\n\t<\/ol><\/li>\n<\/ol><h3 id=\"history\">2.1 A brief History of Name Servers<\/h3><h4> .. or why do we have DNS servers <\/h4><p>Without a Name Service there would simply not be a viable Internet. To understand why, we need to look at what DNS does and how and why it evolved.<\/p><ol><li><p>A DNS translates (or maps) the name of a resource to its physical IP address - typically referred to as forward mapping<\/p><\/li>\n    <li><p>A DNS can also translate the physical IP address to the name of a resource - typically called reverse mapping.<\/p><\/li>\n  <\/ol><p>Big deal.<\/p><p>Remember that the Internet (or any network for that matter) works by allocating every point (host, server, router, interface etc.) a physical IP address (which may be locally unique or globally unique).<\/p><h4>Separation of Church and State.....<\/h4><p>Without DNS every host (PC) which wanted to access a resource on the network (Internet), say a simple web page, for example, www.thing.com, would need to know its <b>physical IP address<\/b>. With 100 of millions of hosts and billions of web pages it is an impossible task - it's also pretty daunting even with just a handful of hosts and resources.<\/p><p>To solve this problem the concept of Name Servers was created in the mid 70's to enable certain attributes (properties) of a <b>named resource<\/b> to be maintained in a known location - the Name Server.<\/p><p>With a <b>Name Server<\/b> present in the network any host only needs to know the <b>physical address of a Name Server<\/b> and the <b>name<\/b> of the resource it wishes to access. Using this data it can find the address (or any other stored attribute or property) of the resource by interrogating (<b>querying<\/b>) the Name Server. Resources can be added, moved, changed or deleted at a single location - the Name Server. At a stroke network management was simplified and made more dynamic.<\/p><h4>If it's broke....<\/h4><p>We now have a new problem with our newly created Name Server concept. If our Name Server is not working our host cannot access any resource on the network. We have made the Name Server a critical resource. So we had better have more than one Name Server in case of failure.<\/p><p>To fix this problem the concept of Primary and Secondary Name Servers (many systems allow tertiary or more Name Servers) was born. If the Primary Name Server does not respond a host can use the Secondary (or tertiary etc.).<\/p><h4>Man, we got more names than Webster....<\/h4><p>As our network grows we start to build up a serious number of Names in our Name Server (database). This gives rise to three new problems.<\/p><ol><li><p>Finding any entry in the database of names becomes increasingly slow as we power through many millions of names looking for the one we want. We need a way to index or organize the names.<\/p><\/li>\n\t\t<li><p>If every host is accessing our Name Servers the load becomes very high. Maybe we need a way to spread the load across a number of servers.<\/p><\/li>\n    <li><p>With many Name (resource) records in our database the management problem becomes increasingly difficult as everyone tries to update all the records at the same time. Maybe we need a way to separate (or <b>delegate<\/b>) the administration of these Name (<b>resource<\/b>) records.<\/p><\/li>\n\t\t<\/ol><p>Which leads us nicely into the characteristics of the Internet's Domain Name System (DNS).<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><p>The Internet's Domain Name System (DNS) is just a specific implementation of the Name Server concept optimized for the prevailing conditions on the Internet.<\/p><h3 id=\"overview\">2.2.1 DNS Overview<\/h3><p>From our brief history of Name Servers we saw how three needs emerged:<\/p><ol><li>The need for a hierarchy of names<\/li>\n    <li><p>The need to spread the operational loads on our name servers<\/p><\/li>\n    <li>The need to delegate the administration of our Name servers<\/li>\n  <\/ol><p>The Internet Domain Name System elegantly solves all these problems at the single stroke of a pen (well actually the whole of RFC 1034 to be precise).<\/p><h3 id=\"domains\">2.2.2 Domains and Delegation<\/h3><p>The Domain Name System uses a tree (or hierarchical) name structure. At the top of the tree is the root\nfollowed by the Top Level Domains (TLDs) then the domain-name and any number of lower levels each separated with a dot. <\/p><p><b>NOTE:<\/b> The root of the tree is represented most of the time as a silent dot ('.') but there are times as we shall see later when it very important.<\/p><p>Top Level Domains (TLDs) were split into two types:\n<\/p><ol><li><p>Generic Top Level Domains (gTLD), for example, .com, .edu, .net, .org, .mil etc.<\/p><\/li>\n<li><p>Country Code Top Level Domain (ccTLD), for example .us, .ca, .tv , .uk etc.<\/p><\/li>\n<\/ol><p><b>Note:<\/b> Country Code TLDs (ccTLDs) use a standard two letter sequence defined by ISO 3166.<\/p><p>In 2004 a sub-category of the gTLDs known as sTLDs (Sponsored TLDs) which implies they may have limited registration was created. Examples of early sTLDs included .aero, .museum, .travel, and .jobs. The historic gTLDs offered a loose categorization of users who could register under the gTLD but in practice many had open registration requirements but this notably excluded .mil, .edu, .gov and .int all of which has (and still have) limited registration.<\/p><p>Finally, since 2011 the TLD policy is essentially unrestricted, if you pay enough money and adopt the operating procedures laid down anyone can register a sponsored TLD. Look forward to a whole set of new TLDs like .singles, .kitchen and .construction arriving.<\/p><p>Figure 1-1 shows this diagrammatically.<\/p><img class=\"center\" src=\"http:\/\/zytrax.com\/books\/dns\/ch2\/fwd-delegation.png\" alt=\"Root and Domain structure\" \/><p class=\"t-c\"><b>Figure 1-1 Domain Structure and Delegation<\/b><\/p><p>What is commonly called a <b>Domain Name<\/b> is actually a combination of a domain-name and a TLD and is written from LEFT to RIGHT with the lowest level in the hierarchy on the left and the highest level on the right.<\/p><p class=\"g-c-n\">domain-name.tld # example.com<\/p><p>In the case of the gTLDs, such as .com, .net etc., the user part of the delegated name - the name the user registered - is a Second Level Domain (SLD). It is the second level in the hierarchy. The user part is therefore frequently simply referred to as the SLD. So the the Domain Name in the example above can be re-defined to consist of:<\/p><pre class=\"g-c-n\">\nsld.tld # example.com\n<\/pre><p>The term Second Level Domain (SLD) is much less useful with ccTLDs where the user registered part is typically the Third Level Domain, for example:<\/p><pre class=\"g-c-n\">\nexample.co.uk\nexample.com.br\n<\/pre><p>The term Second Level Domain (SLD) provides technical precision but can be confusing when applied to a generic concept like a user domain - unless the precision is required we will continue to use the generic term Domain Name or simply Domain to describe the whole name, for instance, what this guide calls a Domain Name would be <b>example.com<\/b> or <b>example.co.uk<\/b>.<\/p><h4>Authority and Delegation<\/h4><p>The concepts of <b>Delegation<\/b> and <b>Authority<\/b> lie at the core of the domain name system hierarchy. The <b>Authority<\/b> for the root domain lies with <a href=\"http:\/\/www.icann.org\" class=\"t-db\">Internet Corporation for Assigned Numbers and Names (ICANN)<\/a>. Since 1998 ICANN, a non-profit organisation, has assumed this responsibility from the US government.<\/p><p>The gTLDs are <b>authoritatively<\/b> administered by ICANN and delegated to a series of accredited registrars. The ccTLDs are delegated to the individual countries for administration purposes. Figure 1.0 above shows how any authority may in turn delegate to lower levels in the hierarchy, in other words it may delegate anything for which it is <b>authoritative<\/b>. Each layer in the hierarchy may <b>delegate<\/b> the <b>authoritative<\/b> control to the next lower level.<\/p><p>In the case of ccTLDs countries like Canada (ccTLD .ca) and the US (ccTLD .us) and others with federated governments decided that they would administer at the national level and delegate to each province (Canada) or state (US) a two character province\/state code, for example, .qc = Quebec, .ny = New York, md = Maryland etc.. Thus <b>mycompany.md.us<\/b> would be the Domain Name of <b>mycompany<\/b> which was delegated from the state of MaryLand in the US. This was the delegation model until around 2006 when both countries changed their registration policies and adopted an essentially flat delegation model. Thus, today you can register mycompany.us or mycompany.ca (as long as they are available). The old delegation models are still valid and you still see domains such as quebec.qc.ca as well as numerous other examples of the multi-layer delegation model.<\/p><p>Countries with more centralized governments, like the UK, Brazil and Spain and others, opted for functional segmentation in their delegation models, for example, .co = company, .ac = academic etc.. Thus <b>mycompany.co.uk<\/b> is the <b>Domain Name<\/b> of <b>mycompany<\/b> registered as a company from the UK registration authority.<\/p><p>Delegation within any domain may be almost limitless and is <b>decided by the delegated authority<\/b>, for example, the US and Canada both delegated city within province\/state domains thus the address (or URL) <b>tennisshoes.nb.us<\/b> is the town of Tennis Shoes in the State of Nebraska in the United States and we could even have <b>mycompany.tennisshoes.nb.us<\/b>.<\/p><p>By reading a domain name from RIGHT to LEFT you can track its delegation. This unit of delegation can also be referred to as a <b>zone<\/b> in standards documentation.<\/p><h4>So What is www.example.com<\/h4><p>From our reading above we can see that www.example.com is built up from <b>www<\/b> and <b>example.com<\/b>. The <b>Domain-Name<\/b> example.com part was delegated from a gTLD registrar which in turn was delegated from ICANN.<\/p><p>The <b>www<\/b> part was chosen by the owner of the domain since they are now the delegated authority for the <b>example.com<\/b> name. They own EVERYTHING to the LEFT of the delegated <b>Domain Name<\/b>.<\/p><p>The leftmost part, <b>www<\/b> in this case, is called a host name. By convention (but only convention) web sites have the 'host' name of www (for world wide web) but you can have a web site whose name is fred.example.com - no-one may think of typing this into their browser but that does not stop you doing it! Equally you may have a web site whose access address (URL) is www.example.com running on a server whose real name is mary.example.com. Again this is perfectly permissable. In short the host part may refer to a real host name or a service name such as <b>www<\/b>. Since the domain owner controls this process it's all allowed.<\/p><p>Every computer, or service, that is addressable (has a URL) via the Internet or an internal network has a host name part, here are some more illustrative examples:<\/p><pre class=\"g-c-n\">\nwww.example.com - the company web service\nftp.example.com - the company file transfer protocol server\npc17.example.com  - a normal PC or host\naccounting.example.com - an accounting system\n<\/pre><p>A host name part must be unique within the <b>Domain Name<\/b> but can be anything the owner of <b>example.com<\/b> wants.<\/p><p>Finally lets look at this name:<\/p><p class=\"g-c-n\">www.us.example.com<\/p><p>From our previous reading we figure its <b>Domain Name<\/b> is example.com, <b>www<\/b> probably indicates a web site, which leaves the <b>us<\/b> part.<\/p><p>The <b>us<\/b> part was allocated by the owner of <b>example.com<\/b> (they are authoritative) and is called a sub-domain. In this case the <b>delegated authority<\/b> for example.com has decided that their company organization is best served by a country based sub-domain structure. They could have <b>delegated<\/b> the responsibility internally to the US subsidiary for administration of this sub-domain, which may in turn have created a plant based structure, such as, www.cleveland.us.example.com which could indicate the web site of the Cleveland plant in the US organisation of <b>example.com<\/b>.<\/p><p>To summarise the OWNER can delegate, IN ANY WAY THEY WANT, ANYTHING to the LEFT of the <b>Domain Name<\/b> they own (were delegated). The owner is also RESPONSIBLE for administering this delegation which means running, or delegating the task of running, a DNS containing <b>Authoritative<\/b> information (or records) for their Domain Name (or <b>zone<\/b>).<\/p><p class=\"p-m-s\"><b>Note:<\/b> Names such as www.example.com and www.us.example.com are commonly - but erroneously - referred to as Fully Qualified Domain Names (FQDN). Technically an FQDN unambiguously defines a name from <i>any starting point<\/i> to the root and as such must contain the normally silent dot at the end. To illustrate \"www.example.com.\" is an FQDN \"www.example.com\" is not.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h3 id=\"structure\">2.2.3 DNS Organization and Structure<\/h3><p>The Internet's DNS exactly maps the 'Domain Name' delegation structure described above. There is a DNS server running at each level in the delegated hierarchy and the responsibility for running the DNS lies with the AUTHORITATIVE control at that level.<\/p><p>Figure 1-2 shows this diagrammatically.<\/p><img class=\"center\" src=\"http:\/\/zytrax.com\/books\/dns\/ch2\/dns-servers.png\" alt=\"Root and Domain structure\" \/><p class=\"t-c\"><b>Figure 1-2 DNS mapped to Domain Delegation<\/b><\/p><p>The Root Servers (Root DNS) are the responsibility of ICANN but operated by a consortium under a delegation agreement. ICANN created the <a href=\"http:\/\/www.icann.org\/committees\/dns-root\/\" class=\"t-db\">Root Servers Systems Advisory Committee (RSSAC)<\/a> to provide advice and guidance as to the operation and development of this critical resource. The IETF was requested by the RSSAC to develop the engineering standards for operation of the Root-Servers. This request resulted in the publication of RFC 2870.<\/p><p>There are currently (mid 2003) <a href=\"http:\/\/www.root-servers.org\" class=\"t-db\">13 root-servers world-wide<\/a>. The Root-Servers are known to every public DNS server in the world and are the starting point for every name lookup operation (or query). To create additional resilience each root-server typically has multiple <b>instances<\/b> (copies) spread throughout the world. Each instance has the same IP address but data is sent to the closest instance using a process called <b>anycasting<\/b>.<\/p><p>The TLD servers (ccTLD and gTLD) are operated by a variety of agencies and organizations (under a fairly complex set of agreements) called <b>Registry Operators<\/b>.<\/p><p>The Authority and therefore the responsibility for the User (or <b>Domain Name<\/b>) DNS servers lies with the owner of the domain. In many cases this responsibility is delegated by the owner of the Domain to an ISP, Web Hosting company or increasingly a registrar. Many companies, however, elect to run their own DNS servers and even delegate the Authority and responsibility for sub-domain DNS servers to separate parts of their organization.<\/p><p>When any DNS cannot answer (resolve) a request (a <b>query<\/b>) for a domain name from a client, for instance, example.com, the query is passed to a <b>root-server<\/b> which will direct (<b>refer<\/b>) the query to the appropriate TLD DNS server (for .com) which will in turn direct (<b>refer<\/b>) it to the appropriate Domain (User) DNS server.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h3 id=\"components\">2.2.4 DNS System Components<\/h3><p>A Domain Name System (DNS) as defined by RFC 1034 includes three parts:<\/p><ol><li>Data which describes the domain(s)<\/li>\n<li>One or more Name Server programs.<\/li>\n<li>A resolver program or library.<\/li>\n<\/ol><p>A single DNS server may support many domains. The data for each domain describes global properties of the domain and its hosts (or services). This data is defined in the form of textual <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/resource.html\" class=\"t-db\">Resource Records<\/a> organized in <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/zones.html\" class=\"t-db\">Zone Files<\/a>. The format of Zone files is defined in RFC 1035 and is supported by most DNS software.<\/p><p>The Name Server program typically does three things:<\/p><ol><li>It will read a configuration file which defines the zones for which it is responsible.<\/li>\n\t\t<li>Depending on the Name Servers <a href=\"http:\/\/zytrax.com\/books\/dns\/ch4\/\" class=\"t-db\">functionality<\/a> a configuration file may describe various behaviours, for instance, to cache or not. Some DNS servers are very specialized and do not provide this level of control.<\/li>\n\t\t<li>Respond to questions (<b>queries<\/b>) from local or remote hosts.<\/li>\n\t\t<\/ol><p><a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/resolver.html\" class=\"t-db\">The resolver program or library <\/a>is located on each host and provides a means of translating a users request for, say, www.thing.com into one or more queries to DNS servers using UDP (or TCP) protocols.<\/p><p class=\"p-m-s\"><b>Note:<\/b> The resolver on all Windows systems and the majority of *nix systems is actually a <b>stub<\/b> resolver - a minimal resolver that can only work with a DNS that supports <a href=\"#recursive\" class=\"t-db\">recursive<\/a> queries. The caching resolver on MS Windows 2K and XP is a <b>stub<\/b> resolver with a cache to speed up responses and reduce network usage.<\/p><p>While <a href=\"http:\/\/www.isc.org\/products\/bind\/\" class=\"t-db\">BIND<\/a> is the best known of the DNS servers and much of this guide documents BIND features, it is by no means the only solution or for that matter the only Open Source solution. <a href=\"http:\/\/zytrax.com\/books\/dns\/apc\/\" class=\"t-db\">Appendix C:<\/a> lists many alternate solutions. The zone file formats which constitute the majority of the work (depending on how many sites you operate) is standard (defined by RFC 1035) and is typically supported by all DNS suppliers. Where a feature is unique to BIND we indicate it clearly in the text so you can keep your options <b>open<\/b>!<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h3 id=\"zones\">2.2.5 Zones and Zone Files<\/h3><p>Zone files contain <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/resource.html\" class=\"t-db\">Resource Records<\/a> that describe a <a href=\"http:\/\/zytrax.com\/books\/dns\/ch1\/index.html#terms\" class=\"t-db\">domain or sub-domain<\/a>. The format of zone files is an IETF standard defined by RFC 1035. Almost any sensible DNS software should be able to read zone files. A zone file will consist of the following types of data:<\/p><ol><li>Data that indicates the top of the zone and some of its general properties (a <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA Record)<\/a>.<\/li>\n<li>Authoritative data for all nodes or hosts within the zone (typically <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/a.html\" class=\"t-db\">A (IPv4) or AAAA (IPv6) Records)<\/a>.<\/li>\n<li>Data that describes global information for the zone (including mail <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/mx.html\" class=\"t-db\">MX Records<\/a> and Name Server <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/ns.html\" class=\"t-db\">NS Records<\/a>).<\/li>\n<li>In the case of <a href=\"http:\/\/zytrax.com\/books\/dns\/ch9\/subdomain.html\" class=\"t-db\">sub-domain delegation<\/a> the name servers responsible for this sub-domain (one or more <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/ns.html\" class=\"t-db\">NS Records)<\/a>.<\/li>\n<li>In the case of <a href=\"http:\/\/zytrax.com\/books\/dns\/ch9\/subdomain.html\" class=\"t-db\">sub-domain delegation<\/a> one or more <b>glue<\/b> records that allows a name server to reach the sub-domain (typically one or more <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/a.html\" class=\"t-db\">A or AAAA Records)<\/a> for the sub-domain name servers.<\/li>\n<\/ol><p>The individual <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/\" class=\"t-db\">Resource Records<\/a> are described and numerous <a href=\"http:\/\/zytrax.com\/books\/dns\/ch6\/\" class=\"t-db\">sample configuration files<\/a> are provided and documented.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h3 id=\"queries\">2.2.6 DNS Queries<\/h3><p>The major task carried out by a DNS server is to respond to queries (questions) from <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/resolver.html\" class=\"t-db\">a local or remote resolver<\/a> or other DNS acting on behalf of a resolver. A query would be something like 'what is the IP address of fred.example.com'.<\/p><p>A DNS server may receive such a query for any domain. DNS servers may be <a href=\"http:\/\/zytrax.com\/books\/dns\/ch4\/\" class=\"t-db\">configured<\/a> to be authoritative for some domains, slaves for others, forward queries or other combinations.<\/p><p>Most of the queries that a DNS server will receive will be for domains for which it has no knowledge, that is, for which it has no local <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/zones.html\" class=\"t-db\">zone files<\/a>. DNS software typically allows the name server to respond in different ways to queries about which it has no knowledge.<\/p><p>There are three types of queries defined for DNS:<\/p><ol><li><p>A recursive query - the complete answer to the question is always returned. DNS servers are not required to support recursive queries.<\/p><\/li>\n<li><p>An Iterative (or non-recursive) query - where the complete answer MAY be returned or a <b>referral<\/b> provided to another DNS. All DNS servers must support Iterative queries.<\/p><\/li>\n<li><p>An Inverse query - where the user wants to know the domain name given a <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/resource.html\" class=\"t-db\">resource record<\/a>. Reverse queries were poorly supported, very infrequent and are now obsolete (RFC 3425).<\/p><\/li>\n<\/ol><p><b>Note:<\/b> The process called <a href=\"http:\/\/zytrax.com\/books\/dns\/ch3\/\" class=\"t-db\">Reverse Mapping<\/a> (returns a host name given an IP address) does not use Inverse queries but instead uses Recursive and Iterative (non-recursive) queries using the special domain name IN-ADDR.ARPA.<\/p><p>Historically reverse IPv4 mapping was not mandatory. Many systems however now use reverse mapping for security and simple authentication schemes (especially mail servers) so proper implementation and maintenance is now practically essential. IPv6 originally mandated reverse mapping but, like a lot of the original IPv6 mandates, has now been rolled-back.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"recursive\">2.2.6.1 Recursive Queries<\/h4><p>A recursive query is one where the DNS server will fully answer the query (or give an error). DNS servers are not required to support recursive queries and both the <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/resolver.html\" class=\"t-db\">resolver<\/a> (or another DNS acting recursively on behalf of another resolver) negotiate use of recursive service using a bit (RD) in the query header.<\/p><p>There are three possible responses to a recursive query:<\/p><ol><li>The answer to the query accompanied by any <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/cname.html\" class=\"t-db\">CNAME records<\/a> (aliases) that may be useful. The response will indicate whether the data is authoritative or cached.<\/li>\n<li>An error indicating the domain or host does not exist (NXDOMAIN). This response may also contain CNAME records that pointed to the non-existing host.<\/li>\n<li>An temporary error indication - for instance, can't access other DNS's due to network error etc..<\/li>\n<\/ol><p>In a recursive query a DNS Resolver will, on behalf of the client (stub-resolver), chase the trail of DNS system across the universe to get the real answer to the question. The journey of a simple query such as 'what is the IP address of www.example.com' to a DNS Resolver which supports recursive queries but is not authoritative for example.com is shown in Diagram 1-3 below:<\/p><img class=\"center\" src=\"http:\/\/zytrax.com\/books\/dns\/ch2\/recursive-query.png\" alt=\"recursive queries\" \/><p class=\"t-c\">Diagram 1-3 Recursive Query Processing<\/p><ol><li><p>The user types www.example.com into their browser address bar. The browser issues a standard function library call (1) to the local stub-resolver.<\/p><\/li>\n<li><p>The stub-resolver sends a query (2) 'what is the IP address of www.example.com' to locally configured DNS resolver (aka recursive name server). This is a standard DNS query requesting recursive services (<a href=\"http:\/\/zytrax.com\/books\/dns\/ch15\/#header\" class=\"t-db\">RD (Recursion Desired) = 1<\/a>).<\/p><\/li>\n<li><p>The DNS Resolver looks up the address of www.example.com in its local tables (its <b>cache<\/b>) and does not find it. (If it were found it would be returned immediately to the Stub-resolver in an answer message and the transaction would be complete.)<\/p><\/li>\n<li><p>The DNS resolver sends a query (3) to a root-server (every DNS resolver is configured with a file that tells it the names and IP addresses of the root servers) for the IP of www.example.com. (Root-servers, TLD servers and correctly configured user name servers do not, a matter of policy, support recursive queries so the Resolver will, typically, not set Recursion Desired (RD = 0) - this query is, in fact, an Iterative query.)<\/p><\/li>\n<li><p>The root-server knows nothing about example.com, let alone the www part, but it does know about the next level in the hierarchy, in this case, the .com part so it replies (answers) with a <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/referral.html\" class=\"t-db\">referral<\/a> (3) pointing at the TLD servers for .com.<\/p><\/li>\n<li><p>The DNS Resolver sends a new query (4) 'what is the IP address of www.example.com' to one of the .com TLD servers. Again it will use, typically, an Iterative query.<\/p><\/li>\n<li><p>The TLD server knows about example.com, but knows nothing about www so, since it cannot supply a complete response to the query, it replies (4) with a <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/referral.html\" class=\"t-db\">referral<\/a> to the name servers for example.com.<\/p><\/li>\n<li><p>The DNS Resolver sends yet another query (5) 'what is the IP address www.example.com' to one of the name servers for example.com. Once again it will use, typically, an Iterative query.<\/p><\/li>\n<li><p>The example.com zone file defines a <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/a.html\" class=\"t-db\">A (IPv4 address) record<\/a> so the authoritative server for example.com returns (5) the A record for www.example.com (it fully answers the question).<\/p><\/li>\n<li><p>The DNS Resolver sends the response (answer) www.example.com=x.x.x.x to the client's stub-resolver (2) and then places this information in its cache.<\/p><\/li>\n<li><p>The stub-resolver places the information www.example.com=x.x.x.x in its cache (since around 2003 most stub-resolvers have been caching stub-resolvers) and responds to the original standard library function call (1) with www.example.com = x.x.x.x.<\/p><\/li>\n<li><p>The browser receives the response to its standard function call, places the information in its cache (really) and initiates an HTTP session to the address x.x.x.x. DNS transaction complete. Quite simple really, not much could possibly go wrong.<\/p><\/li>\n<\/ol><p>In summary, the stub-resolver demands recursive services from the DNS Resolver. The DNS Resolver provides a recursive service but uses, typically, Iterative queries to achieve it.<\/p><p><b>Note:<\/b> The resolver on Windows and most *nix systems is a <b>stub-resolver<\/b> (in point of fact, in most modern systems it is a <b>Caching stub-Resolver<\/b>) - which is defined in the standards to be a minimal resolver which cannot follow <b>referrals<\/b>. If you reconfigure your local PC or Workstation to point to a DNS server that only supports Iterative queries - it will not work. Period.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"iterative\">2.2.6.2 Iterative (non-recursive) Queries<\/h4><p>A Iterative (or non-recursive) query is one where the DNS server may provide an answer or a partial answer (a referral) to the query (or give an error). All DNS servers must support non-recursive (Iterative) queries. An Iterative query is technically simply a normal DNS query that does not request Recursive Services.<\/p><p>There are four possible responses to a non-recursive query:<\/p><ol><li>The answer to the query accompanied by any <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/cname.html\" class=\"t-db\">CNAME records<\/a> (aliases) that may be useful (in a Iterative Query this will ONLY occur if the requested data is already available in the cache). The response will indicate whether the data is authoritative or cached.<\/li>\n<li>An error indicating the domain or host does not exist (NXDOMAIN). This response may also contain CNAME records that pointed to the non-existing host.<\/li>\n<li>An temporary error indication, for instance, can't access other DNS's due to network error etc..<\/li>\n<li>A <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/referral.html\" class=\"t-db\"><b>referral<\/b><\/a>: If the requested data is not available in the cache then the name and IP addess(es) of one or more name server(s) that are closer to the requested domain name (in all cases this is the next lower level in the DNS hierarchy) will be returned. This referral may, or may not be, to the authoritative name server for the target domain.<\/li>\n<\/ol><p>In Diagram 1-3 above the transactions (3), (4) and (5) are normally all Iterative queries. Even if the DNS server requested Recursion (RD=1) it would be denied and a normal referral (or answer) returned. Why use Iterative queries? They are much faster, the DNS server receiving the query either already has the answer in its cache, in which case it sends it, or not, in which case it sends a referral. No messing around. Iterative queries give the requestor greater control. A referral typically contains a list of name servers for the next level in the DNS hierarchy. The requestor may have additional information about one or more of these name servers in its cache (including which is the fastest) from which it can make a better decision about which name server to use. Iterative queries are also extremely useful in diagnostic situations.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"reverse\">2.2.6.3 Inverse Queries<\/h4><p>Historically, an Inverse query mapped a resource record to a domain. An example Inverse query would be 'what is the domain name for this MX record'. Inverse query support was optional and it was permitted for the DNS server to return a response <b>Not Implemented<\/b>.<\/p><p>Inverse queries are NOT used to find a host name given an IP address. This process is called <a href=\"http:\/\/zytrax.com\/books\/dns\/ch3\/\" class=\"t-db\">Reverse Mapping (Look-up)<\/a> uses recursive and Iterative (non-recursive) queries with the special domain name IN-ADDR.ARPA.<\/p><p>Inverse queries went the way of all \"seemed like a good idea at the time\" concepts when they were finally obsoleted by RFC 3425.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><p>The initial design of DNS allowed for changes to be propagated using Zone Transfer (AXFR) but the world of the Internet was simpler and more sedate in those days (1987). The desire to speed up the process of zone update propagation while minimising resources used has resulted in a number of changes to this aspect of DNS design and implementation from simple - but effective - tinkering such as <b>Incremental Zone Transfer (IXFR)<\/b> and <b>NOTIFY<\/b> messages to the concept of <b>Dynamic Updates<\/b> which have significant security consequences if not properly implemented. Diagram 1-4 show zone transfer capabilities.<\/p><p><b>Warning:<\/b> While zone transfers are generally essential for the operation of DNS systems they are also a source of threat. A A slave Name Server can become <b>poisoned<\/b> if it accepts zone updates from a malicious source. Care should be taken during configuration to ensure that, as a minimum, the 'slave' will only accept transfers from known sources. The <a href=\"http:\/\/zytrax.com\/books\/dns\/ch6\/\" class=\"t-db\">example configurations<\/a> provide these minimum precautions. <a href=\"#security\" class=\"t-db\">Security Overview<\/a> outlines some of the potential threats involved.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><p>The original DNS specifications (<a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc1034.txt\" target=\"_blank\" class=\"t-db\">RFC 1034<\/a> &amp; <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc1035.txt\" target=\"_blank\" class=\"t-db\">RFC 1035<\/a>) envisaged that Slave (or secondary) Name Servers would 'poll' the Domain (or zone) Master. The time between such 'polling' is determined by the <b>refresh<\/b> value on the domain's <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA Resource Record<\/a><\/p><p>The polling process is accomplished by the Slave sending a query to the Master requesting its current SOA resource record (RR). If the <b>serial number<\/b> of this RR is higher than the current one maintained by the Slave, a zone transfer (AXFR) is requested. This is why it is vital to be very disciplined about updating the SOA serial number every time anything changes in ANY of the zone records.<\/p><p>Zone transfers are always carried out using TCP on port 53 whereas normal DNS query operations use UDP on port 53.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><p>Transferring very large zone files can take a long time and waste bandwidth and other resources. This is especially wasteful if only a single RR has been changed! <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc1995.txt\" target=\"_blank\" class=\"t-db\">RFC 1995<\/a> introduced Incremental Zone Transfers (IXFR) which as the name suggests allows the Slave and Master to transfer only those records that have changed.<\/p><p>The process works as for AXFR. The Slave sends a query for the domain's <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA RR<\/a> every <b>refresh<\/b> interval. If the <b>serial number<\/b> of the SOA record is higher than the current one maintained by the Slave it requests a zone transfer and indicates whether or not it is capable of accepting an Incremental Transfer (IXFR). If both Master and slave support the feature an Incremental Transfer (IXFR) takes place otherwise a Full Transfer (AXFR) takes place. Incremental Zone transfers use TCP on port 53, whereas normal DNS queries operations use UDP on port 53.<\/p><p>The default mode for BIND when acting as a Slave is to request IXFR unless it is configured not to using the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#request-ixfr\" class=\"t-db\"><b>request-ixfr<\/b><\/a> parameter in the <b>server<\/b> or <b>options<\/b> clause of the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/\" class=\"t-db\">named.conf<\/a> file.<\/p><p>The default mode for BIND when acting as a Master is to use IXFR only when the zone is <a href=\"#dyn-update\" class=\"t-db\">dynamic<\/a>. The use of IXFR is controlled using the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#provide-ixfr\" class=\"t-db\"><b>provide-ixfr<\/b><\/a> parameter in the <b>server<\/b> or <b>options<\/b> clause of the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/\" class=\"t-db\">named.conf<\/a> file.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"notify\">2.2.7.3 Notify (NOTIFY)<\/h4><p><a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc1912.txt\" target=\"_blank\" class=\"t-db\">RFC 1912<\/a> recommends a REFRESH interval of up to 12 hours on the REFRESH interval of an <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA Resource Record<\/a>. This means that, in the worst case, changes to the Master Name Server may not be visible at the Slave Name Server(s) for up to 12 hours. In a dynamic environment this may be unacceptable.<\/p><p><a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc1996.txt\" target=\"_blank\" class=\"t-db\">RFC 1996<\/a> introduced a scheme whereby the <b>Master<\/b> will send a NOTIFY message to the <b>Slave<\/b> Name Server(s) that a change MAY have occurred in the domain records. The Slave(s) on receipt of the NOTIFY will request the latest <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA Resource Record<\/a> and if the <b>serial number<\/b> of the SOA RR is greater than its current value it will initiate a zone transfer using either a Full Zone Transfer (AXFR) or an Incremental Zone Transfer (IXFR).<\/p><p>NOTIFY behaviour in BIND is controlled by <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#notify\" class=\"t-db\"><b>notify<\/b><\/a>, <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#also-notify\" class=\"t-db\"><b>also-notify<\/b><\/a> and <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#notify-source\" class=\"t-db\"><b>notify-source<\/b><\/a> parameters in the <b>zone<\/b> or <b>options<\/b> statements of the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/\" class=\"t-db\">named.conf<\/a> file.<\/p><img class=\"center\" src=\"http:\/\/zytrax.com\/books\/dns\/ch2\/master-slave-zone-transfer.png\" alt=\"\" \/><p class=\"t-c\">Diagram 1-4 Master - Slave Interaction and Zone Transfer<\/p><h4>Zone Transfer Process Description<\/h4><p>The time taken to propagate zone changes throughout the Internet is determined by two major factors. First, the time taken to update all the Domain's Name servers when any zone change occurs. This, in turn, is determined by the method used to initiate zone transfers to all Slave Name Servers which may be passive (the Slave will periodically poll the Master) or Active (the Master will send a NOTIFY to its configured Slave(s)). Both methods are described below. Second, the current <a href=\"http:\/\/zytrax.com\/books\/dns\/apa\/ttl.html\" class=\"t-db\">TTL<\/a> value (prior to its change) on any changed zone record will determine when Resolvers will refresh their caches by interrogating the Authoritative Name Server.<\/p><ol><li><p>If the Master has been configured to support NOTIFY messages then whenever the status of the Master's zone file (1) changes it will send a NOTIFY message (2) to each configured Slave. A NOTIFY message does not necessarily indicate that the zone file has changed, for example, if the Master or the zone is reloaded then a NOTIFY message is triggered even if no changes have occured. When the Slave receives a NOTIFY message it follows the procedure defined in Step 3 below.<\/p><\/li>\n<li><p>Irrespective of whether the Master has been configured to support NOTIFY messages or not the Slave will always use the passive or 'polling' process described in this step. (While on its face this seem superflous in cases where the Master has been configured to use NOTIFY, however, it does provide protection again lost NOTIFY messages due to mal-configuration or malicious attack.) When a Slave server is loaded it will read any current saved zone file (see <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/zone.html#file\" class=\"t-db\">file statement<\/a>) or immediately intitate a zone transfer if there is no saved zone file. It then starts a timer using the <b>refresh<\/b> value in the zone's <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA RR<\/a>. When this timer expires the Slave follows the procedure defined in Step 3 below.<\/p><\/li>\n<li><p>If the Slave's <b>refresh<\/b> timer expires OR it receives a NOTIFY message the Slave will immediately issue a query for the zone Master's SOA RR (3).<\/p><\/li>\n<li><p>When the answer arrives (4) the Slave compares the <b>serial number<\/b> of its current <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA RR<\/a> with that of the answer (the Master's SOA RR). If the value of the Master SOA RR <b>serial number<\/b> is greater than the current <b>serial number<\/b> in the Slave's SOA copy then a zone transfer (5) is intiated by the Slave. (The gruesome details of the serial number arithmetic is defined in <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc1982.txt\" class=\"t-db\">RFC 1982<\/a> and clarified in <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc2181.txt\" class=\"t-db\">RFC 2181<\/a>, the date based convention used for serial numbers is <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">defined here<\/a>). If the slave fails to read the Master's SOA RR (or fails to intiate the zone transfer) then it will try again after the <b>retry<\/b> time defined in the zone's <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA RR<\/a> but will continue answering Authoritatively for the Domain (or zone). The retry procedure will be repeated (every <b>retry<\/b> interval) either until it succeeds (in which case the process continues at step 5 below) or until the <b>expiry<\/b> timer of the zone's <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA RR<\/a> is reached, at which point the Slave will stop answering queries for the Domain.<\/p><\/li>\n<li><p>The Slave always intitiates (5) a zone transfer operation (using AXFR or IXFR) using TCP on Port 53 (this can be configured using the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#transfer-source\" class=\"t-db\">transfer-source<\/a> statement).<\/p><\/li>\n<li><p>The Master will transfer the requested zone file (6) to the slave. On completion the Slave will reset its <b>refresh<\/b> and <b>expiry<\/b> timers.<\/p><\/li>\n<\/ol><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><p>The classic method of updating Zone <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/\" class=\"t-db\">Resource Records<\/a> is to manually edit the zone file and then stop and start the name server to propagate the changes. When the volume of changes reaches a certain level this can become operationally unacceptable - especially considering that in organisations which handle large numbers of Zone Files, such as service providers, BIND itself can take a long time to restart at it plows through very large numbers of zone statements.<\/p><p>The 'holy grail' of DNS is to provide a method of dynamically changing the DNS records while DNS continues to service requests.<\/p><p>There are two architectural approaches to solving this problem:<\/p><ol><li>Allow 'run-time' updating of the Zone Records from an external source\/application.<\/li>\n<li>Directly feed the DNS from a database which can be dynamically updated. BIND provides two APIs to enable this, other DNS implemenations provide embedded database access.<\/li>\n<\/ol><p><a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc2136.txt\" target=\"_blank\" class=\"t-db\">RFC 2136<\/a> takes the first approach and defines a process where zone records can be updated from an external source. The key limitation in this specification is that a new domain cannot be added dynamically. All other records within an existing zone can be added, changed or deleted. This limitation is also true for both of BIND's APIs as well.<\/p><p>As part of RFC 2136 the term <b>Primary Master<\/b> was coined to describe the Name Server defined in the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch8\/soa.html\" class=\"t-db\">SOA Resource Record<\/a> for the zone. The significance of this term is that when dynamically updating records it is essential to update only one server even though there may be multiple <b>master<\/b> servers for the zone. In order to solve this problem a 'boss' server must be selected, this 'boss' server (the <b>Primary Master<\/b>) has no special characteristics other than it is defined as the Name Server in the SOA record and may appear in an <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#allow-update\" class=\"t-db\"><b>allow-update<\/b><\/a> clause to control the update process.<\/p><p>DDNS is normally associated with Secure DNS features such as TSIG - <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc2845.txt\" target=\"_blank\" class=\"t-db\">RFC 2845<\/a> &amp; TKEY - <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc2930.txt\" target=\"_blank\" class=\"t-db\">RFC 2930<\/a>. Dynamic DNS (<b>DDNS<\/b>) does not REQUIRE TSIG\/TKEY. However by enabling Dynamic DNS you are also opening up the possibility of <b>master<\/b> zone file corruption or poisoning. Simple IP address protection (acl) can be configured into BIND but this provides - at best - limited protection. For that reason, serious users of Dynamic DNS will always use TSIG\/TKEY procedures to authenticate incoming update requests.<\/p><p>In BIND DDNS is defaulted to <b>deny from all hosts<\/b>. Control of Dynamic Update is provided by the BIND <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#allow-update\" class=\"t-db\"><b>allow-update<\/b><\/a> (usable with and without TSIG\/TKEY) and <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/xfer.html#update-policy\" class=\"t-db\"><b>update-policy<\/b><\/a> (only usable with TSIG\/TKEY) clauses in the <b>zone<\/b> or <b>options<\/b> statements of the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/\" class=\"t-db\">named.conf<\/a> file.<\/p><p>There are a number of Open Source tools which will initiate Dynamic DNS updates these include <a href=\"http:\/\/dnsupdate.sourceforge.net\/\" class=\"t-db\"><b>dnsupdate<\/b><\/a> (not the same as DNSUpdate) and <b>nsupdate<\/b> which is distributed with <b>bind-utils<\/b>.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"dyn-alt\">2.2.7.5 Alternative Dynamic DNS Approaches<\/h4><p>As noted above the major limitation in the standard Dynamic DNS (RFC 2136) approach is that new domains cannot be created dynamically.<\/p><p><a href=\"http:\/\/bind-dlz.sourceforge.net\/\" class=\"t-db\">BIND-DLZ<\/a> takes a much more radical approach and, using a serious patch to BIND, allows replacement of all zone files with a single zone file which defines a database entry. The database support, which includes most of the major databases (MySQL, PostgreSQL, BDB and LDAP among others) allows the addition of new domains as well as changes to pre-existing domains without the need to stop and start BIND. As with all things in life there is a trade-off and performance can drop precipitously. Current work being carried (early 2004) out with a High performance Berkeley DB (BDB) is showing excellent results approaching raw BIND performance.<\/p><p><a href=\"http:\/\/www.powerdns.com\/\" class=\"t-db\">PowerDNS<\/a> an authoritative only name server takes a similar approach with its own (non-BIND) code base by referring all queries to the database back-end and thereby allow new domains to be added dynamically.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h3 id=\"security\">2.3 Security Overview<\/h3><p>DNS Security is a huge and complex topic. It is made worse by the fact that almost all the documentation dives right in and you fail to see the forest for all the d@!#*d trees.<\/p><p>The critical point is to first understand what you want to secure - or rather what threat level you want to secure against. This will be very different if you run a root server rather than running a modest in-house DNS serving a couple of low volume web sites.<\/p><p>The term DNSSEC is thrown around as a blanket term in a lot of documentation. This is not correct. There are at least three types of DNS security, two of which are - relatively - painless and DNSSEC which is - relatively - painful.<\/p><p>Security is always an injudicious blend of real threat and paranoia - but remember just because you are naturally paranoid does not mean that <b>they<\/b> are not after you!<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"security-threats\">2.3.1 Security Threats<\/h4><p>In order to be able to assess both the potential threats and the possible counter-measures it is first and foremost necessary to understand the normal data flows in a DNS system. Diagram 1-5 below shows this flow.<\/p><img class=\"center\" src=\"http:\/\/zytrax.com\/books\/dns\/ch2\/dns-security-1.png\" alt=\"DNS Data\/Threat flow\" \/><p class=\"t-c\"><b>Diagram 1-5 DNS Data Flow<\/b><\/p><p>Every data flow (each RED line above) is a potential source of threat! Using the numbers from the above diagram here is what can happen at each flow - beware, you may not sleep soundly tonight:<\/p><table class=\"p-m-s\"><tr class=\"g-h-ss\"><td>Number<\/td>\n<td>Area<\/td>\n<td>Threat<\/td>\n<\/tr><tr><td>(1)<\/td>\n<td>Zone Files<\/td>\n<td>File Corruption (malicious or accidental). Local threat.<\/td>\n<\/tr><tr><td>(2)<\/td>\n<td>Dynamic Updates<\/td>\n<td>Unauthorized Updates, IP address spoofing (impersonating update source). Server to Server (TSIG Transaction) threat.<\/td>\n<\/tr><tr><td>(3)<\/td>\n<td>Zone Transfers<\/td>\n<td>IP address spoofing (impersonating update source). Server to Server (TSIG Transaction) threat.<\/td>\n<\/tr><tr><td>(4)<\/td>\n<td>Remote Queries<\/td>\n<td>Cache Poisoning by IP spoofing, data interception, or a subverted Master or Slave. Server to Client (DNSSEC) threat.<\/td>\n<\/tr><tr><td>(5)<\/td>\n<td>Resolver Queries<\/td>\n<td>Data interception, Poisoned Cache, subverted Master or Slave, local IP spoofing. Remote Client-client (DNSSEC) threat.<\/td>\n<\/tr><\/table><p>The first phase of getting a handle on the problem is to figure (audit) what threats are applicable and how seriously do YOU rate them or do they even apply. As an example; if you don't do Dynamic Updates (BIND's default mode) - there is no Dynamic Update threat! Finally in this section a warning: <b>the further you go from the Master the more complicated the solution and implementation<\/b>. Unless there is a very good reason for not doing so, we would always recommend that you start from the Master and work out.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"security-types\">2.3.2 Security Types<\/h4><p>We classify each threat type below. This classification simply allows us select appropriate remedies and strategies for avoiding or securing our system. The numbering used below relates to diagram 1-3.<\/p><p>(1) The primary source of Zone data is normally the Zone Files (and don't forget the <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/\" class=\"t-db\">named.conf<\/a> file which contains lots of interesting data as well). This data should be secure and securely backed up. This threat is classified as <b>Local<\/b> and is typically handled by good system administration.<\/p><p>(2) The BIND default is to <b>deny<\/b> Dynamic Zone Updates. If you have enabled this service or require to it poses a serious threat to the integrity of your Zone files and should be protected. This is classified as a  <b>Server-Server (Transaction)<\/b> threat.<\/p><p>(3) If you run slave servers you will do zone transfers. <b>Note:<\/b> You do NOT have to run with slave servers, you can run with multiple masters and eliminate the transfer threat entirely. This is classified as a <b>Server-Server (Transaction)<\/b> threat.<\/p><p>(4) The possibility of Remote Cache Poisoning due to IP spoofing, data interception and other hacks is a judgement call if you are running a simple web site. If the site is high profile, open to competitive threat or is a high revenue earner you have probably implemented solutions already. This is classified as a  <b>Server-Client<\/b> threat.<\/p><p>(5) The current DNSSEC standards define a security aware resolver and this concept is under active development by an number of groups round the world. This is classified as a  <b>Server-Client<\/b> threat.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"security-local\">2.3.3 Security - Local<\/h4><p>Normal system administration practices such as ensuring that files (configuration and zone files) are securely backed-up, proper read and write permissions applied and sensible physical access control to servers may be sufficient.<\/p><p>Implementing a <a href=\"http:\/\/zytrax.com\/books\/dns\/ch4\/index.html#stealth\" class=\"t-db\">Stealth (or Split)<\/a> DNS server provides a more serious solution depending on available resources.<\/p><p>Finally you can run BIND (named) in a <b>chroot jail<\/b>.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"security-server\">2.3.4 Server-Server (TSIG Transactions)<\/h4><p><b>Zone transfers<\/b>. If you have slave servers you will do zone transfers. BIND provides <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/acl.html\" class=\"t-db\"><b>Access Control Lists (ACLs)<\/b><\/a> which allow simple IP address protection. While IP based <b>ACLs<\/b> are relatively easy to subvert using IP addreess spoofing they are a <b>lot<\/b> better than nothing and require very little work. You can run with multiple masters (no slaves) and eliminate the threat entirely. You will have to manually synchronise zone file updates but this may be a simpler solution if changes are not frequent.<\/p><p><b>Dynamic Updates<\/b>. If you must run with this service it should be secured. BIND provides <a href=\"http:\/\/zytrax.com\/books\/dns\/ch7\/acl.html\" class=\"t-db\"><b>Access Control Lists (ACLs)<\/b><\/a> which allow simple IP address protection but this is probably not adequate unless you can secure the IP addresses, that is, all systems are behind a firewall\/DMZ\/NAT or the updating hosts are using private IP addresses.<\/p><p><b>TSIG\/TKEY<\/b> If all other solutions fail DNS specifications (<a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc2845.txt\" target=\"_blank\" class=\"t-db\">RFC 2845<\/a> - TSIG and <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc2930.txt\" target=\"_blank\" class=\"t-db\">RFC 2930<\/a> - TKEY) provide authentication protocol enhancements to secure these Server-Server transactions.<\/p><p><b>TSIG<\/b> and <b>TKEY<\/b> implementations are messy but not too complicated simply because of the scope of the problem. With <b>Server-Server<\/b> transactions there is a finite and normally small number of hosts involved. The protocols depend on a <b>shared secret<\/b> between the master and the slave(s) or updater(s). It is further assumed that you can get the <b>shared secret<\/b> securely to the peer server by some means not covered in the protocol itself. This process, known as <b>key exchange<\/b>, may not be trivial (typically long random strings of base64 characters are involved) but you can use the telephone(!), mail, fax or PGP email among other methods.<\/p><p>The <b>shared-secret<\/b> is open to <b>brute-force<\/b> attacks so frequent (monthly or more) changing of <b>shared secrets<\/b> will become a fact of life. <b>TKEY<\/b> allows automation of <b>key-exchange<\/b> using a Diffie-Hellman algorithm but starts with a <b>shared secret<\/b>! TKEY appears to have very limited, if any, usage.<\/p><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><h4 id=\"security-clients\">2.3.5 Server-Client (DNSSEC)<\/h4><p>The classic Remote Poisoned cache problem is not trivial to solve simply because there may be an infinitely large number of Remote Caches involved. It is not reasonable to assume that you can use a <b>shared secret<\/b>.<\/p><p>Instead <b>DNSSEC<\/b> relies on <b>public\/private key authentication<\/b>. The DNSSEC specifications (<a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc4033.txt\" target=\"_blank\" class=\"t-db\">RFC 4033<\/a>, <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc4034.txt\" target=\"_blank\" class=\"t-db\">RFC 4034<\/a> and <a href=\"http:\/\/zytrax.com\/books\/dns\/apd\/rfc4035.txt\" target=\"_blank\" class=\"t-db\">RFC 4035<\/a> augmented with others) attempt to answer three questions:<\/p><ol><li>Authentication - the DNS responding really is the DNS that the request was sent to.<\/li>\n<li>Integrity - the response is complete and nothing is missing or changed.<\/li>\n<li>Proof of non-existance - if the DNS returns a status that the name does not exist (NXDOMAIN) this response can be proven to have come from the authoritative server.<\/li>\n<\/ol><p><a href=\"#contents\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_up.gif\" alt=\"up icon\" \/><\/a><\/p><p><a href=\"http:\/\/zytrax.com\/books\/dns\/ch3\" class=\"t-db\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_to.gif\" alt=\"next icon\" \/>Chapter 3 - Reverse Mapping<\/a><\/p><p><a href=\"http:\/\/zytrax.com\/books\/dns\/\" class=\"t-db\"><img class=\"w-27\" src=\"http:\/\/zytrax.com\/images\/go_back.gif\" alt=\"back icon\" \/>Main Contents<\/a><\/p><hr \/><p class=\"p-m-n\"><br \/>Problems, comments, suggestions, corrections (including broken links) or something to add? Please take the time from a busy life to 'mail us' (at top of screen), the webmaster (below) or <a href=\"denied:javascript:mailus('info-support','zytrax.com','Support%20Issue')\" class=\"t-db\">info-support at zytrax<\/a>. You will have a warm inner glow for the rest of the day.<\/p>","created_at":"2018-11-07T16:45:55+0000","updated_at":"2018-11-07T16:45:55+0000","annotations":[],"mimetype":"text\/html","reading_time":38,"domain_name":"zytrax.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":361,"title":"How To Create Vanity or Branded Nameservers with DigitalOcean Cloud Servers","url":"https:\/\/www.digitalocean.com\/community\/tutorials\/how-to-create-vanity-or-branded-nameservers-with-digitalocean-cloud-servers","content":"<div class=\"container tutorial-header\"><p>We hope you find this tutorial helpful. In addition to guides like this one, we provide simple cloud infrastructure for developers. <a href=\"https:\/\/www.digitalocean.com\/?utm_medium=community&amp;utm_source=community_gencta&amp;utm_campaign=needaserver&amp;utm_content=how-to-create-vanity-or-branded-nameservers-with-digitalocean-cloud-servers\">\n        Learn more \u2192\n<\/a><\/p>PostedJuly 10, 2013\n    145k views<a class=\"tag\" href=\"https:\/\/www.digitalocean.com\/community\/tags\/dns?type=tutorials\">DNS<\/a><a class=\"tag\" href=\"https:\/\/www.digitalocean.com\/community\/tags\/digitalocean?type=tutorials\">DigitalOcean<\/a><\/div><div class=\"content-body tutorial-content tutorial-content-legacy\" data-growable-markdown=\"\"><h3>Introduction<\/h3><p>Of particular interest to hosting providers or resellers, having branded or&#13;\nvanity nameservers provides a more professional look to clients. It&#13;\neliminates the need of asking your clients to point their domains to another&#13;\ncompany's nameservers. This tutorial will outline two approaches to creating&#13;\ncustom nameservers: (i) Vanity and (ii) Branded.<\/p><h2>Types<\/h2><p><strong>Vanity nameservers<\/strong> allow you to use your own domain name, without having to setup complicated zone files; you can do this using DigitalOcean's &#13;\nnameservers and DNS Manager. This is accomplished by&#13;\nmapping your custom nameservers to DigitalOcean's IPs.<\/p><p><strong>Branded Nameservers<\/strong> require a little more configuration, but allow you to exert complete control over DNS for your domain. The added control, however, carries with it the burden of having to self-manage your DNS. You'll need to deploy at least two VPS, with specialized software such as BIND, PowerDNS or NSD&#13;\n(for \"name server daemon\"). Wikipedia publishes a nice <a target=\"_blank\" href=\"http:\/\/en.wikipedia.org\/wiki\/Comparison_of_DNS_server_software\">comparison of DNS server software<\/a>.<\/p><h2>Naming<\/h2>You can use any naming scheme you want. If you're unsure, the most common\nschemes are<code>ns1.yourdomain.com<\/code>or<code>a.ns.yourdomain.com<\/code>.<h2>Prerequisites<\/h2><h3>Ingredients for Both Vanity &amp; Branded Nameservers:<\/h3><p>1. Registered domain name from an established registrar, e.g. GoDaddy; NameCheap; 1&amp;1; NetworkSolutions; Register.com etc. (at this time, DigitalOcean does not offer domain registration services.)<\/p><p>2. <a href=\"http:\/\/en.wikipedia.org\/wiki\/Glue_records#Circular_dependencies_and_glue_records\">Glue Records<\/a>: Ascertain your domain registrar's procedure for creating glue&#13;\n    records. Different registrar's refer to glue records by different names,&#13;\n    such as GoDaddy whom refers to them as host names. Other providers may refer&#13;\n    to the process as \"registering a nameserver\" or \"creating a host record.\"&#13;\n    Glue records tell the rest of the world where to find your nameservers and are&#13;\n    needed to prevent circular references. Circular references exist where the&#13;\n    nameservers for a domain can't be resolved without resolving the domain&#13;\n    they're responsible for. <strong>If you are not able to determine how to&#13;\n    create Glue Records at your particular domain registrar (that is, how to&#13;\n    \"register a nameserver or host name\"), then you need to contact your registrar&#13;\n    directly and let them know that you need to register a nameserver.<\/strong><\/p><h3>For Vanity Nameservers Only<\/h3><p>DigitalOcean's current IP addresses for its nameservers (which can be obtained by clicking on the respective hyperlinks, below; or, via nslookup; dig; or ping commands):<\/p><p><a target=\"_blank\" href=\"http:\/\/reports.internic.net\/cgi\/whois?whois_nic=ns1.digitalocean.com&amp;type=nameserver\">ns1.digitalocean.com<\/a><\/p><p><a target=\"_blank\" href=\"http:\/\/reports.internic.net\/cgi\/whois?whois_nic=ns2.digitalocean.com&amp;type=nameserver\">ns2.digitalocean.com<\/a><\/p><p><a target=\"_blank\" href=\"http:\/\/reports.internic.net\/cgi\/whois?whois_nic=ns3.digitalocean.com&amp;type=nameserver\">ns3.digitalocean.com<\/a><\/p><h3>Additional Requirements if You'd Like to Maximize Control Over Your Domain's DNS, with Branded Nameservers:<\/h3><p>Create or identify at least two VPS that you control that will act as&#13;\n    Primary and Secondary Nameservers.<\/p><p>NOTE: It's technically possible to have only one VPS act as both the Primary&#13;\n    and Secondary Nameserver. This approach, however, is not recommended because&#13;\n    it sacrifices the safety that redundancy provides (i.e., fault tolerance).&#13;\n    Keep in mind, however, that there's no hard limit of only two nameservers for&#13;\n    your domain. You're only limited by the number of nameservers that your&#13;\n    domain registrar allows you to register.<\/p><p>Deploy a DNS Server on your Primary and Secondary Nameservers. <i>See<\/i>&#13;\n    <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-setup-dns-slave-auto-configuration-using-virtualmin-webmin-on-ubuntu\">&#13;\n       How to Setup DNS Slave Auto Configuration Using Virtualmin\/Webmin on Ubuntu<\/a>;&#13;\n    <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-install-the-bind-dns-server-on-centos-6\">&#13;\n       How to Install the BIND DNS Server on CentOS 6<\/a>; or&#13;\n    <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-install-powerdns-on-centos-6-3-x64\">&#13;\n       How To Install PowerDNS on CentOS 6.3 x64<\/a><\/p><h3>The Quick &amp; Easy Recipe: Vanity Nameservers:<\/h3><p>1. First, login to your <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/the-digitalocean-control-panel\">DigitalOcean Control Panel<\/a> and add your domain name to&#13;\n    the <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-set-up-a-host-name-with-digitalocean\">DigitalOcean DNS Manager<\/a>.<\/p><p>2. Then, create A Records for your vanity nameservers and point them to&#13;\n    DigitalOcean's IPs for ns1.digitalocean.com; ns2.digitalocean.com; ns3.digitalocean.com.<\/p><p>To accomplish this, create a new host A-Record with <strong>ns1.yourdomain.com.<\/strong> (do <strong>NOT<\/strong> forget to end the hostname with a period) in the&#13;\n    hostname field. The IP address to use for ns1.yourdomain.com. is the IP&#13;\n    address you discovered for ns1.digitalocean.com (above). Repeat these steps for&#13;\n    <strong>ns2.yourdomain.com.<\/strong> and <strong>ns3.yourdomain.com.<\/strong><\/p><p>For example:<\/p><p><strong>(Do not forget the trailing dots)<\/strong><\/p><pre>A   ns1.yourdomain.com<strong>.<\/strong>     [IP address for ns1.digitalocean.com]<\/pre><pre>A   ns2.yourdomain.com<strong>.<\/strong>     [IP address for ns2.digitalocean.com]<\/pre><pre>A   ns3.yourdomain.com<strong>.<\/strong>     [IP address for ns3.digitalocean.com]<\/pre><p>3. Next, you need to replace DigitalOcean's NS Records with each of&#13;\nyour vanity nameservers in&#13;\n    the <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-set-up-a-host-name-with-digitalocean\">DigitalOcean DNS Manager<\/a>.<\/p><p><strong>(Do not forget the trailing dots)<\/strong><\/p><pre>NS      ns1.yourdomain.com<strong>.<\/strong><\/pre><pre>NS      ns2.yourdomain.com<strong>.<\/strong><\/pre><pre>NS      ns3.yourdomain.com<strong>.<\/strong><\/pre><p>4. This next step will vary, depending on your domain name's registrar: Login&#13;\n    to your domain name registrar's control panel and register the IPs of your&#13;\n    nameservers by creating Glue Records. In another words, associate (or map)       DigitalOcean's nameserver IPs with your vanity nameservers' hostnames.<\/p><p>With GoDaddy, for example, simply login to your Domain Name Control Panel&#13;\n    and look for the area where you can list Host Names. There, click on Manage&#13;\n    =&gt; Add Hostname and enter NS1 for the Hostname and ns1.digitalocean.com's IP&#13;\n    address; click Add Hostname again and enter NS2 for the Hostname and&#13;\n    ns2.digitalocean.com's IP Address. Click Add Hostname yet a third time and add&#13;\n    NS3 for the Hostname and ns3.digitalocean.com's IP Address.<\/p><p>5. Almost done! Skip down to the DNS Testing section.<\/p><h3>Recipe for Maximum Control, with Branded Nameservers:<\/h3><p>The simplest way to configure DNS is to have someone else do it. For that reason,&#13;\nyou should consider using DigitalOcean's <a target=\"_blank\" href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-set-up-a-host-name-with-digitalocean\">DNS Manager<\/a>.<\/p><p>If you really want to manage your domain's DNS yourself, however, you next&#13;\nneed to deploy a DNS server such as BIND. A complete zone-file configuration is&#13;\nbeyond the scope of this tutorial. However, you need to ensure that you apply the&#13;\nsame principals described above:<\/p><p>1. Create both A &amp; NS Records for ns1.yourdomain.com. and ns2.yourdomain.com.&#13;\n(with BIND, especially, do not forget the trailing periods).<\/p><p>2. Ultimately, your zone file will contain the following entries:<\/p><pre>ns1.yourdomain.com. IN  A   1.2.3.4<\/pre><pre>ns2.yourdomain.com. IN  A   1.2.3.5<\/pre><pre>yourdomain.com.     IN  NS  ns1.yourdomain.com.<\/pre><pre>yourdomain.com.     IN  NS  ns2.yourdomain.com.<\/pre><p>3. Remember, the IP addresses for your ns1 and ns2 A Records (and for your Glue Records) come from you--in that you have to set up <strong>at least<\/strong> two VPS to run your name servers.<\/p><p>4. Login to your domain name registrar's control panel and create Glue Records&#13;\n   for as many nameservers you wish to deploy. Just make sure that you are&#13;\n   using the IP addresses of servers under your control (and not the addresses of&#13;\n   DigitalOcean's nameservers).<\/p><h2>DNS Testing<\/h2><p>To make sure you configured everything correctly, you can run the&#13;\n<a target=\"_blank\" href=\"http:\/\/www.webdnstools.com\/dnstools\/domain_check\">Check&#13;\n   Domain Configuration<\/a> tool. Keep in mind, however, that, depending on your&#13;\n   registrar, nameserver changes can take up to 72 hours to properly propagate&#13;\n   throughout the Internet.<\/p><p>Article Submitted by: <a href=\"http:\/\/vdevices.com\">Pablo Carranza<\/a><\/p><\/div>","created_at":"2018-11-07T16:45:52+0000","updated_at":"2018-11-07T16:45:52+0000","published_at":"2013-07-10T17:14:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.digitalocean.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":360,"title":"Photon Design System","url":"http:\/\/design.firefox.com\/photon\/welcome.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:45:51+0000","updated_at":"2018-11-07T16:45:51+0000","annotations":[],"reading_time":0,"domain_name":"design.firefox.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":359,"title":"Buttons \u00b7 An Introduction to Elm","url":"https:\/\/guide.elm-lang.org\/architecture\/user_input\/buttons.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:45:50+0000","updated_at":"2018-11-07T16:45:50+0000","annotations":[],"reading_time":0,"domain_name":"guide.elm-lang.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":358,"title":"ClojureScript Unraveled","url":"https:\/\/funcool.github.io\/clojurescript-unraveled\/","content":"<pre data-lang=\"clojure\">(def a (my-atom 0))\n;; =&gt; #&lt;MyAtom 0&gt;\n(add-watch a :log (fn [key a oldval newval]\n                    (println {:key key\n                              :old oldval\n                              :new newval})))\n;; =&gt; #&lt;MyAtom 0&gt;\n(swap! a inc)\n;; {:key :log, :old 0, :new 1}\n;; =&gt; 1\n(swap! a + 2)\n;; {:key :log, :old 1, :new 3}\n;; =&gt; 3\n(swap! a - 2)\n;; {:key :log, :old 3, :new 1}\n;; =&gt; 1\n(swap! a + 2 3)\n;; {:key :log, :old 1, :new 6}\n;; =&gt; 6\n(swap! a + 4 5 6)\n;; {:key :log, :old 6, :new 21}\n;; =&gt; 21\n(swap! a * 2)\n;; {:key :log, :old 21, :new 42}\n;; =&gt; 42\n(remove-watch a :log)\n;; =&gt; #&lt;MyAtom 42&gt;<\/pre>","created_at":"2018-11-07T16:45:26+0000","updated_at":"2018-11-07T16:45:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"funcool.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":353,"title":"The Quad textured - LWJGL","url":"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html","content":"<div id=\"toc\" class=\"toc\"><p><\/p><h2>Contents<\/h2><ul><li class=\"toclevel-1 tocsection-1\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Introduction\">1 Introduction<\/a><\/li>\n<li class=\"toclevel-1 tocsection-2\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Resources\">2 Resources<\/a><\/li>\n<li class=\"toclevel-1 tocsection-3\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Texture_Units\">3 Texture Units<\/a><\/li>\n<li class=\"toclevel-1 tocsection-4\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#.28Texture.29_Coordinates\">4 (Texture) Coordinates<\/a><\/li>\n<li class=\"toclevel-1 tocsection-5\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Using_the_attribute_lists\">5 Using the attribute lists<\/a><\/li>\n<li class=\"toclevel-1 tocsection-6\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Texture_Shaders\">6 Texture Shaders<\/a>\n<ul><li class=\"toclevel-2 tocsection-7\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Vertex_Shader\">6.1 Vertex Shader<\/a><\/li>\n<li class=\"toclevel-2 tocsection-8\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Fragment_Shader\">6.2 Fragment Shader<\/a><\/li>\n<li class=\"toclevel-2 tocsection-9\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Binding_the_shader_variables\">6.3 Binding the shader variables<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-10\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Loading_textures\">7 Loading textures<\/a><\/li>\n<li class=\"toclevel-1 tocsection-11\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#The_result\">8 The result<\/a><\/li>\n<li class=\"toclevel-1 tocsection-12\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Complete_source_code\">9 Complete source code<\/a>\n<ul><li class=\"toclevel-2 tocsection-13\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Vertex_shader_2\">9.1 Vertex shader<\/a><\/li>\n<li class=\"toclevel-2 tocsection-14\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Fragment_shader_2\">9.2 Fragment shader<\/a><\/li>\n<li class=\"toclevel-2 tocsection-15\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#TexturedVertex_class\">9.3 TexturedVertex class<\/a><\/li>\n<li class=\"toclevel-2 tocsection-16\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Application\">9.4 Application<\/a><\/li>\n<\/ul><\/li>\n<li class=\"toclevel-1 tocsection-17\"><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/The_Quad_textured.html#Credit\">10 Credit<\/a><\/li>\n<\/ul><\/div><h2>Introduction<\/h2><p>This tutorial will cover how to apply textures to a model, as with the other \u201cquad\u201d tutorials we\u2019ll build on the former knowledge about VBO\u2019s and Shaders. We\u2019ll talk about what to do to make sure textures are rendered on the screen. If you are a unsure how to create an interleaved VBO or how to setup a shader read the previous Quad tutorials first. As always you can find the complete source code at the end of the tutorial. Let\u2019s begin.\n<\/p><h2>Resources<\/h2><p>For this tutorial we\u2019ll use an extra library to load our images, and we\u2019ll also use a set of images of course. The image loader we use is TWL\u2019s PNG decoder tool which you can find here: <a rel=\"nofollow\" class=\"external autonumber\" href=\"http:\/\/twl.l33tlabs.org\/#downloads\">[1]<\/a>. It\u2019s a JAR file that you must add to your project\u2019s libraries. \n<\/p><p>Finally the textures we\u2019ll use are a set of \u201cUV grids\u201d. We\u2019ll talk about UV\u2019s a bit later. We\u2019re going to use 2 textures, these are PNG images but the original came from this site: <a rel=\"nofollow\" class=\"external autonumber\" href=\"http:\/\/www.pixelcg.com\/blog\/?p=146\">[2]<\/a>. Pick 2 of them and convert them into PNG\u2019s with your image editor of choice.\n<\/p><h2>Texture Units<\/h2><p>The first thing we must know is that OpenGL uses a list of texture units. Just like with attribute lists of a VBO there are only a set amount of texture units available. The number of texture units varies, the minimum that must be available is 2. We can get these texture units by calling \u201cGL_TEXTUREX\u201d where X is the unit of the texture. The 2 units that must be supported are GL_TEXTURE0 and GL_TEXTURE1.\n<\/p><p>Because a model can use many different textures (diffuse, specular, normal, baked shadow, \u2026) we must tell OpenGL what texture to place where. In this tutorial we\u2019ll only use GL_TEXTURE0, because our shader only looks at the diffuse texture. A diffuse texture is just the base color of the object. Up until now we\u2019ve assigned colors to our vertices and OpenGL interpolated between the vertices creating our colored Quad. By using textures we can map an image on to this quad and have easy control over each pixels color.\n<\/p><h2>(Texture) Coordinates<\/h2><p>Most of the time we want to create a 3D render of something so our vertices have 3 components (X, Y, Z). While our quad has no depth, the vertices still have a Z value (0). Texture coordinates however only have 2 components (S, T). 3D modelers know another name for these coordinates namely \u201cUV\u2019. Here\u2019s a graphical presentation of the coordinates that OpenGL uses:\n<\/p><p><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/File:Coordinates.png.html\" class=\"image\"><img alt=\"Coordinates.png\" src=\"http:\/\/wiki.lwjgl.org\/images\/5\/51\/Coordinates.png\" width=\"512\" height=\"318\" \/><\/a>\n<\/p><p>The XYZ components have nothing to do with the ST components, as the process of going from XYZ to ST is more often called unwrapping where an artist can define each individual vertex it\u2019s ST position independent of the XYZ position.\nThe graphical example are the exact numbers we are using in this tutorial. We\u2019ve updated the custom Vertex so we can assign ST values to it. We define our vertices like this:\n<\/p><pre class=\"brush:java\">\n\/\/ We'll define our quad using 4 vertices of the custom 'TexturedVertex' class\nTexturedVertex v0 = new TexturedVertex(); \nv0.setXYZ(-0.5f, 0.5f, 0); v0.setRGB(1, 0, 0); v0.setST(0, 0);\nTexturedVertex v1 = new TexturedVertex(); \nv1.setXYZ(-0.5f, -0.5f, 0); v1.setRGB(0, 1, 0); v1.setST(0, 1);\nTexturedVertex v2 = new TexturedVertex(); \nv2.setXYZ(0.5f, -0.5f, 0); v2.setRGB(0, 0, 1); v2.setST(1, 1);\nTexturedVertex v3 = new TexturedVertex(); \nv3.setXYZ(0.5f, 0.5f, 0); v3.setRGB(1, 1, 1); v3.setST(1, 0);\n<\/pre><h2>Using the attribute lists<\/h2><p>If you are unfamiliar with the workings of our custom vertex class see the code at the end or read the tutorial about \u201cinterleaving\u201d. We are still interleaving our vertex data, so let\u2019s do that now with the extra ST data:\n<\/p><pre class=\"brush:java\">\n\/\/ Create a new Vertex Buffer Object in memory and select it (bind)\nvboId = GL15.glGenBuffers();\nGL15.glBindBuffer(GL15.GL_ARRAY_BUFFER, vboId);\nGL15.glBufferData(GL15.GL_ARRAY_BUFFER, verticesBuffer, GL15.GL_STATIC_DRAW);\n\/\/ Put the position coordinates in attribute list 0\nGL20.glVertexAttribPointer(0, TexturedVertex.positionElementCount, GL11.GL_FLOAT, \n\t\tfalse, TexturedVertex.stride, TexturedVertex.positionByteOffset);\n\/\/ Put the color components in attribute list 1\nGL20.glVertexAttribPointer(1, TexturedVertex.colorElementCount, GL11.GL_FLOAT, \n\t\tfalse, TexturedVertex.stride, TexturedVertex.colorByteOffset);\n\/\/ Put the texture coordinates in attribute list 2\nGL20.glVertexAttribPointer(2, TexturedVertex.textureElementCount, GL11.GL_FLOAT, \n\t\tfalse, TexturedVertex.stride, TexturedVertex.textureByteOffset);\nGL15.glBindBuffer(GL15.GL_ARRAY_BUFFER, 0);\n<\/pre><h2>Texture Shaders<\/h2><p>In OpenGL 3.x and up we have to rely on shaders to do all of the graphical work. Therefore we must update our shaders so it can read and process the ST data. Mapping textures to the model is done in the \u201cfragment\u201d shaders. Because we are defining individual pixels and that\u2019s the job of the fragment shader. \n<\/p><h4>Vertex Shader<\/h4><p>But it all starts with the vertex shader, so we have to define a new \u201cin\u201d variable in our vertex shader and simply pass the values to our fragment shader:\n<\/p><pre class=\"brush:c\">\n#version 150 core\nin vec4 in_Position;\nin vec4 in_Color;\nin vec2 in_TextureCoord;\nout vec4 pass_Color;\nout vec2 pass_TextureCoord;\nvoid main(void) {\n\tgl_Position = in_Position;\n\t\n\tpass_Color = in_Color;\n\tpass_TextureCoord = in_TextureCoord;\n}\n<\/pre><h4>Fragment Shader<\/h4><p>And now our fragment shader can use the texture coordinates to sample our texture and apply color to our pixels:\n<\/p><pre class=\"brush:c\">\n#version 150 core\nuniform sampler2D texture_diffuse;\nin vec4 pass_Color;\nin vec2 pass_TextureCoord;\nout vec4 out_Color;\nvoid main(void) {\n\tout_Color = pass_Color;\n\t\/\/ Override out_Color with our texture pixel\n\tout_Color = texture(texture_diffuse, pass_TextureCoord);\n}\n<\/pre><p>Notice that the fragment shader \u201csamples\u201d our texture, which it can access by our uniform variable \u201ctexture_diffuse\u201d (of the type sampler2D, remember this for later). It must sample the texture because our screen pixels will usually not be a 1 to 1 copy of the texture pixels. If we zoom in our zoom out we\u2019ll need more or less pixels. If we need 3 pixels where there originally only were 2, the sampler will create a middle pixel by for example averaging the neighboring pixels.\n<\/p><h4>Binding the shader variables<\/h4><p>In any case, we now have 3 kinds of data for each vertex: an XYZ coordinate, a color and an ST coordinate. We must bind our texture coordinate variable of the vertex shader to an attribute list of our VBO:\n<\/p><pre class=\"brush:java\">\n\/\/ Position information will be attribute 0\nGL20.glBindAttribLocation(pId, 0, \"in_Position\");\n\/\/ Color information will be attribute 1\nGL20.glBindAttribLocation(pId, 1, \"in_Color\");\n\/\/ Textute information will be attribute 2\nGL20.glBindAttribLocation(pId, 2, \"in_TextureCoord\");\n<\/pre><p>Remember to use the exact same variable names when binding attributes to an attribute list. \n<\/p><h2>Loading textures<\/h2><p>We\u2019ve got everything hooked up, all we need is to load the actual texture. To import the images we\u2019re going to use an external library called the TWL PNG decoder. How this is done is of no importance, what is important is that OpenGL requires a buffer of bytes that resemble the pixel color components. Which is what we get from the PNG decoder:\n<\/p><pre class=\"brush:java\">\nByteBuffer buf = null;\nint tWidth = 0;\nint tHeight = 0;\ntry {\n\t\/\/ Open the PNG file as an InputStream\n\tInputStream in = new FileInputStream(filename);\n\t\/\/ Link the PNG decoder to this stream\n\tPNGDecoder decoder = new PNGDecoder(in);\n\t\n\t\/\/ Get the width and height of the texture\n\ttWidth = decoder.getWidth();\n\ttHeight = decoder.getHeight();\n\t\n\t\n\t\/\/ Decode the PNG file in a ByteBuffer\n\tbuf = ByteBuffer.allocateDirect(\n\t\t\t4 * decoder.getWidth() * decoder.getHeight());\n\tdecoder.decode(buf, decoder.getWidth() * 4, Format.RGBA);\n\tbuf.flip();\n\t\n\tin.close();\n} catch (IOException e) {\n\te.printStackTrace();\n\tSystem.exit(-1);\n}\n<\/pre><p>After we run this code, we get a buffer filled with pixel data in the RGBA format. We must also tell OpenGL what format the buffer is filled with, thanks so the PNG decoder we can simply pick one that fits our situation. RGBA is a very common format, another common format is BRG(A) which DirectX uses.\n<\/p><p>Next we\u2019ll create our texture object in memory, the \u201ctextureUnit\u201d variable is the Texture Unit as explained in the beginning of this tutorial. We must of course bind this texture object to apply change, the first change is to push the byte buffer to memory:\n<\/p><pre class=\"brush:java\">\n\/\/ Create a new texture object in memory and bind it\nint texId = GL11.glGenTextures();\nGL13.glActiveTexture(textureUnit);\nGL11.glBindTexture(GL11.GL_TEXTURE_2D, texId);\n\/\/ All RGB bytes are aligned to each other and each component is 1 byte\nGL11.glPixelStorei(GL11.GL_UNPACK_ALIGNMENT, 1);\n<\/pre><p>Remember that our shader variable is a sampler2D, the 2D part tells OpenGL what type of texture it is (a 3D type  also exists for example). When uploading the buffer we must use the appropriate texture type:\n<\/p><pre class=\"brush:java\">\n\/\/ Upload the texture data and generate mip maps (for scaling)\nGL11.glTexImage2D(GL11.GL_TEXTURE_2D, 0, GL11.GL_RGB, tWidth, tHeight, 0, \n\t\tGL11.GL_RGBA, GL11.GL_UNSIGNED_BYTE, buf);\nGL30.glGenerateMipmap(GL11.GL_TEXTURE_2D);\n<\/pre><p>A good explanation of how the different kinds of \u201cwraps\u201d work can be found here <a rel=\"nofollow\" class=\"external autonumber\" href=\"http:\/\/www.flipcode.com\/archives\/Advanced_OpenGL_Texture_Mapping.shtml\">[3]<\/a>.\n<\/p><p>Done, we\u2019ve now uploaded our texture to OpenGL (memory). In this example we\u2019ll use 2 textures and switch between them by using the keyboard, we\u2019ll use an integer array to hold the ID\u2019s:\n<\/p><pre class=\"brush:java\">\ntexIds[0] = this.loadPNGTexture(\"assets\/stGrid1.png\", GL13.GL_TEXTURE0);\ntexIds[1] = this.loadPNGTexture(\"assets\/stGrid2.png\", GL13.GL_TEXTURE0);\n<\/pre><p>Notice that we both upload them to texture unit 0. If we have multiple smapler variables in our shader we can assign textures to the different variables by using a different texture unit. We select what texture we\u2019re going to display by pressing 2 keys on the keyboard:\n<\/p><pre class=\"brush:java\">\nwhile(Keyboard.next()) {\n\t\/\/ Only listen to events where the key was pressed (down event)\n\tif (!Keyboard.getEventKeyState()) continue;\n\t\n\t\/\/ Switch textures depending on the key released\n\tswitch (Keyboard.getEventKey()) {\n\tcase Keyboard.KEY_1:\n\t\ttextureSelector = 0;\n\t\tbreak;\n\tcase Keyboard.KEY_2:\n\t\ttextureSelector = 1;\n\t\tbreak;\n\t}\n}\n<\/pre><p>Of course we\u2019ll have to bind the texture in our rendering loop:\n<\/p><pre class=\"brush:java\">\nGL20.glUseProgram(pId);\n\/\/ Bind the texture\nGL13.glActiveTexture(GL13.GL_TEXTURE0);\nGL11.glBindTexture(GL11.GL_TEXTURE_2D, texIds[textureSelector]);\n<\/pre><p>And destroy them when the application closes:\n<\/p><pre class=\"brush:java\">\n\/\/ Delete the texture\nGL11.glDeleteTextures(texIds[0]);\nGL11.glDeleteTextures(texIds[1]);\n<\/pre><h2>The result<\/h2><p>Here\u2019s how the program looks, we can switch between the 2 textures by pressing the 1 and 2 keys at the top of our keyboard:\n<\/p><p><a href=\"http:\/\/wiki.lwjgl.org\/wiki\/File:Result_textured.png.html\" class=\"image\"><img alt=\"Result textured.png\" src=\"http:\/\/wiki.lwjgl.org\/images\/4\/48\/Result_textured.png\" width=\"672\" height=\"358\" \/><\/a>\n<\/p><h2>Complete source code<\/h2><h4>Vertex shader<\/h4><pre class=\"brush:c\">\n#version 150 core\nin vec4 in_Position;\nin vec4 in_Color;\nin vec2 in_TextureCoord;\nout vec4 pass_Color;\nout vec2 pass_TextureCoord;\nvoid main(void) {\n\tgl_Position = in_Position;\n\t\n\tpass_Color = in_Color;\n\tpass_TextureCoord = in_TextureCoord;\n}\n<\/pre><h4>Fragment shader<\/h4><pre class=\"brush:c\">\n#version 150 core\nuniform sampler2D texture_diffuse;\nin vec4 pass_Color;\nin vec2 pass_TextureCoord;\nout vec4 out_Color;\nvoid main(void) {\n\tout_Color = pass_Color;\n\t\/\/ Override out_Color with our texture pixel\n\tout_Color = texture(texture_diffuse, pass_TextureCoord);\n}\n<\/pre><h4>TexturedVertex class<\/h4><pre class=\"brush:java\">\npublic class TexturedVertex {\n\t\/\/ Vertex data\n\tprivate float[] xyzw = new float[] {0f, 0f, 0f, 1f};\n\tprivate float[] rgba = new float[] {1f, 1f, 1f, 1f};\n\tprivate float[] st = new float[] {0f, 0f};\n\t\n\t\/\/ The amount of bytes an element has\n\tpublic static final int elementBytes = 4;\n\t\n\t\/\/ Elements per parameter\n\tpublic static final int positionElementCount = 4;\n\tpublic static final int colorElementCount = 4;\n\tpublic static final int textureElementCount = 2;\n\t\n\t\/\/ Bytes per parameter\n\tpublic static final int positionBytesCount = positionElementCount * elementBytes;\n\tpublic static final int colorByteCount = colorElementCount * elementBytes;\n\tpublic static final int textureByteCount = textureElementCount * elementBytes;\n\t\n\t\/\/ Byte offsets per parameter\n\tpublic static final int positionByteOffset = 0;\n\tpublic static final int colorByteOffset = positionByteOffset + positionBytesCount;\n\tpublic static final int textureByteOffset = colorByteOffset + colorByteCount;\n\t\n\t\/\/ The amount of elements that a vertex has\n\tpublic static final int elementCount = positionElementCount + \n\t\t\tcolorElementCount + textureElementCount;\t\n\t\/\/ The size of a vertex in bytes, like in C\/C++: sizeof(Vertex)\n\tpublic static final int stride = positionBytesCount + colorByteCount + \n\t\t\ttextureByteCount;\n\t\n\t\/\/ Setters\n\tpublic void setXYZ(float x, float y, float z) {\n\t\tthis.setXYZW(x, y, z, 1f);\n\t}\n\t\n\tpublic void setRGB(float r, float g, float b) {\n\t\tthis.setRGBA(r, g, b, 1f);\n\t}\n\t\n\tpublic void setST(float s, float t) {\n\t\tthis.st = new float[] {s, t};\n\t}\n\t\n\tpublic void setXYZW(float x, float y, float z, float w) {\n\t\tthis.xyzw = new float[] {x, y, z, w};\n\t}\n\t\n\tpublic void setRGBA(float r, float g, float b, float a) {\n\t\tthis.rgba = new float[] {r, g, b, 1f};\n\t}\n\t\n\t\/\/ Getters\t\n\tpublic float[] getElements() {\n\t\tfloat[] out = new float[TexturedVertex.elementCount];\n\t\tint i = 0;\n\t\t\n\t\t\/\/ Insert XYZW elements\n\t\tout[i++] = this.xyzw[0];\n\t\tout[i++] = this.xyzw[1];\n\t\tout[i++] = this.xyzw[2];\n\t\tout[i++] = this.xyzw[3];\n\t\t\/\/ Insert RGBA elements\n\t\tout[i++] = this.rgba[0];\n\t\tout[i++] = this.rgba[1];\n\t\tout[i++] = this.rgba[2];\n\t\tout[i++] = this.rgba[3];\n\t\t\/\/ Insert ST elements\n\t\tout[i++] = this.st[0];\n\t\tout[i++] = this.st[1];\n\t\t\n\t\treturn out;\n\t}\n\t\n\tpublic float[] getXYZW() {\n\t\treturn new float[] {this.xyzw[0], this.xyzw[1], this.xyzw[2], this.xyzw[3]};\n\t}\n\t\n\tpublic float[] getRGBA() {\n\t\treturn new float[] {this.rgba[0], this.rgba[1], this.rgba[2], this.rgba[3]};\n\t}\n\t\n\tpublic float[] getST() {\n\t\treturn new float[] {this.st[0], this.st[1]};\n\t}\n}\n<\/pre><h4>Application<\/h4><pre class=\"brush:java\">\nimport java.io.BufferedReader;\nimport java.io.FileInputStream;\nimport java.io.FileReader;\nimport java.io.IOException;\nimport java.io.InputStream;\nimport java.nio.ByteBuffer;\nimport java.nio.FloatBuffer;\nimport org.lwjgl.BufferUtils;\nimport org.lwjgl.LWJGLException;\nimport org.lwjgl.input.Keyboard;\nimport org.lwjgl.opengl.ContextAttribs;\nimport org.lwjgl.opengl.Display;\nimport org.lwjgl.opengl.DisplayMode;\nimport org.lwjgl.opengl.GL11;\nimport org.lwjgl.opengl.GL13;\nimport org.lwjgl.opengl.GL15;\nimport org.lwjgl.opengl.GL20;\nimport org.lwjgl.opengl.GL30;\nimport org.lwjgl.opengl.PixelFormat;\nimport org.lwjgl.util.glu.GLU;\nimport de.matthiasmann.twl.utils.PNGDecoder;\nimport de.matthiasmann.twl.utils.PNGDecoder.Format;\npublic class TheQuadExampleTextured {\n\t\/\/ Entry point for the application\n\tpublic static void main(String[] args) {\n\t\tnew TheQuadExampleTextured();\n\t}\n\t\n\t\/\/ Setup variables\n\tprivate final String WINDOW_TITLE = \"The Quad: Textured\";\n\tprivate final int WIDTH = 320;\n\tprivate final int HEIGHT = 320;\n\t\/\/ Quad variables\n\tprivate int vaoId = 0;\n\tprivate int vboId = 0;\n\tprivate int vboiId = 0;\n\tprivate int indicesCount = 0;\n\t\/\/ Shader variables\n\tprivate int vsId = 0;\n\tprivate int fsId = 0;\n\tprivate int pId = 0;\n\t\/\/ Texture variables\n\tprivate int[] texIds = new int[] {0, 0};\n\tprivate int textureSelector = 0;\n\t\n\tpublic TheQuadExampleTextured() {\n\t\t\/\/ Initialize OpenGL (Display)\n\t\tthis.setupOpenGL();\n\t\t\n\t\tthis.setupQuad();\n\t\tthis.setupShaders();\n\t\tthis.setupTextures();\n\t\t\n\t\twhile (!Display.isCloseRequested()) {\n\t\t\t\/\/ Do a single loop (logic\/render)\n\t\t\tthis.loopCycle();\n\t\t\t\n\t\t\t\/\/ Force a maximum FPS of about 60\n\t\t\tDisplay.sync(60);\n\t\t\t\/\/ Let the CPU synchronize with the GPU if GPU is tagging behind\n\t\t\tDisplay.update();\n\t\t}\n\t\t\n\t\t\/\/ Destroy OpenGL (Display)\n\t\tthis.destroyOpenGL();\n\t}\n\tprivate void setupTextures() {\n\t\ttexIds[0] = this.loadPNGTexture(\"assets\/stGrid1.png\", GL13.GL_TEXTURE0);\n\t\ttexIds[1] = this.loadPNGTexture(\"assets\/stGrid2.png\", GL13.GL_TEXTURE0);\n\t\t\n\t\tthis.exitOnGLError(\"setupTexture\");\n\t}\n\tprivate void setupOpenGL() {\n\t\t\/\/ Setup an OpenGL context with API version 3.2\n\t\ttry {\n\t\t\tPixelFormat pixelFormat = new PixelFormat();\n\t\t\tContextAttribs contextAtrributes = new ContextAttribs(3, 2)\n\t\t\t\t.withForwardCompatible(true)\n\t\t\t\t.withProfileCore(true);\n\t\t\t\n\t\t\tDisplay.setDisplayMode(new DisplayMode(WIDTH, HEIGHT));\n\t\t\tDisplay.setTitle(WINDOW_TITLE);\n\t\t\tDisplay.create(pixelFormat, contextAtrributes);\n\t\t\t\n\t\t\tGL11.glViewport(0, 0, WIDTH, HEIGHT);\n\t\t} catch (LWJGLException e) {\n\t\t\te.printStackTrace();\n\t\t\tSystem.exit(-1);\n\t\t}\n\t\t\n\t\t\/\/ Setup an XNA like background color\n\t\tGL11.glClearColor(0.4f, 0.6f, 0.9f, 0f);\n\t\t\n\t\t\/\/ Map the internal OpenGL coordinate system to the entire screen\n\t\tGL11.glViewport(0, 0, WIDTH, HEIGHT);\n\t\t\n\t\tthis.exitOnGLError(\"setupOpenGL\");\n\t}\n\t\n\tprivate void setupQuad() {\n\t\t\/\/ We'll define our quad using 4 vertices of the custom 'TexturedVertex' class\n\t\tTexturedVertex v0 = new TexturedVertex(); \n\t\tv0.setXYZ(-0.5f, 0.5f, 0); v0.setRGB(1, 0, 0); v0.setST(0, 0);\n\t\tTexturedVertex v1 = new TexturedVertex(); \n\t\tv1.setXYZ(-0.5f, -0.5f, 0); v1.setRGB(0, 1, 0); v1.setST(0, 1);\n\t\tTexturedVertex v2 = new TexturedVertex(); \n\t\tv2.setXYZ(0.5f, -0.5f, 0); v2.setRGB(0, 0, 1); v2.setST(1, 1);\n\t\tTexturedVertex v3 = new TexturedVertex(); \n\t\tv3.setXYZ(0.5f, 0.5f, 0); v3.setRGB(1, 1, 1); v3.setST(1, 0);\n\t\t\n\t\tTexturedVertex[] vertices = new TexturedVertex[] {v0, v1, v2, v3};\n\t\t\/\/ Put each 'Vertex' in one FloatBuffer\n\t\tFloatBuffer verticesBuffer = BufferUtils.createFloatBuffer(vertices.length *\n\t\t\t\tTexturedVertex.elementCount);\n\t\tfor (int i = 0; i &lt; vertices.length; i++) {\n\t\t\t\/\/ Add position, color and texture floats to the buffer\n\t\t\tverticesBuffer.put(vertices[i].getElements());\n\t\t}\n\t\tverticesBuffer.flip();\t\n\t\t\/\/ OpenGL expects to draw vertices in counter clockwise order by default\n\t\tbyte[] indices = {\n\t\t\t\t0, 1, 2,\n\t\t\t\t2, 3, 0\n\t\t};\n\t\tindicesCount = indices.length;\n\t\tByteBuffer indicesBuffer = BufferUtils.createByteBuffer(indicesCount);\n\t\tindicesBuffer.put(indices);\n\t\tindicesBuffer.flip();\n\t\t\n\t\t\/\/ Create a new Vertex Array Object in memory and select it (bind)\n\t\tvaoId = GL30.glGenVertexArrays();\n\t\tGL30.glBindVertexArray(vaoId);\n\t\t\n\t\t\/\/ Create a new Vertex Buffer Object in memory and select it (bind)\n\t\tvboId = GL15.glGenBuffers();\n\t\tGL15.glBindBuffer(GL15.GL_ARRAY_BUFFER, vboId);\n\t\tGL15.glBufferData(GL15.GL_ARRAY_BUFFER, verticesBuffer, GL15.GL_STATIC_DRAW);\n\t\t\n\t\t\/\/ Put the position coordinates in attribute list 0\n\t\tGL20.glVertexAttribPointer(0, TexturedVertex.positionElementCount, GL11.GL_FLOAT, \n\t\t\t\tfalse, TexturedVertex.stride, TexturedVertex.positionByteOffset);\n\t\t\/\/ Put the color components in attribute list 1\n\t\tGL20.glVertexAttribPointer(1, TexturedVertex.colorElementCount, GL11.GL_FLOAT, \n\t\t\t\tfalse, TexturedVertex.stride, TexturedVertex.colorByteOffset);\n\t\t\/\/ Put the texture coordinates in attribute list 2\n\t\tGL20.glVertexAttribPointer(2, TexturedVertex.textureElementCount, GL11.GL_FLOAT, \n\t\t\t\tfalse, TexturedVertex.stride, TexturedVertex.textureByteOffset);\n\t\t\n\t\tGL15.glBindBuffer(GL15.GL_ARRAY_BUFFER, 0);\n\t\t\n\t\t\/\/ Deselect (bind to 0) the VAO\n\t\tGL30.glBindVertexArray(0);\n\t\t\n\t\t\/\/ Create a new VBO for the indices and select it (bind) - INDICES\n\t\tvboiId = GL15.glGenBuffers();\n\t\tGL15.glBindBuffer(GL15.GL_ELEMENT_ARRAY_BUFFER, vboiId);\n\t\tGL15.glBufferData(GL15.GL_ELEMENT_ARRAY_BUFFER, indicesBuffer, GL15.GL_STATIC_DRAW);\n\t\tGL15.glBindBuffer(GL15.GL_ELEMENT_ARRAY_BUFFER, 0);\n\t\t\n\t\tthis.exitOnGLError(\"setupQuad\");\n\t}\n\t\n\tprivate void setupShaders() {\t\t\n\t\t\/\/ Load the vertex shader\n\t\tvsId = this.loadShader(\"src\/thequad\/vertex_textured.glsl\", GL20.GL_VERTEX_SHADER);\n\t\t\/\/ Load the fragment shader\n\t\tfsId = this.loadShader(\"src\/thequad\/fragment_textured.glsl\", GL20.GL_FRAGMENT_SHADER);\n\t\t\n\t\t\/\/ Create a new shader program that links both shaders\n\t\tpId = GL20.glCreateProgram();\n\t\tGL20.glAttachShader(pId, vsId);\n\t\tGL20.glAttachShader(pId, fsId);\n\t\t\/\/ Position information will be attribute 0\n\t\tGL20.glBindAttribLocation(pId, 0, \"in_Position\");\n\t\t\/\/ Color information will be attribute 1\n\t\tGL20.glBindAttribLocation(pId, 1, \"in_Color\");\n\t\t\/\/ Textute information will be attribute 2\n\t\tGL20.glBindAttribLocation(pId, 2, \"in_TextureCoord\");\n\t\t\n\t\tGL20.glLinkProgram(pId);\n\t\tGL20.glValidateProgram(pId);\n\t\t\n\t\tthis.exitOnGLError(\"setupShaders\");\n\t}\n\t\n\tprivate void loopCycle() {\n\t\t\/\/ Logic\n\t\twhile(Keyboard.next()) {\n\t\t\t\/\/ Only listen to events where the key was pressed (down event)\n\t\t\tif (!Keyboard.getEventKeyState()) continue;\n\t\t\t\n\t\t\t\/\/ Switch textures depending on the key released\n\t\t\tswitch (Keyboard.getEventKey()) {\n\t\t\tcase Keyboard.KEY_1:\n\t\t\t\ttextureSelector = 0;\n\t\t\t\tbreak;\n\t\t\tcase Keyboard.KEY_2:\n\t\t\t\ttextureSelector = 1;\n\t\t\t\tbreak;\n\t\t\t}\n\t\t}\n\t\t\n\t\t\/\/ Render\n\t\tGL11.glClear(GL11.GL_COLOR_BUFFER_BIT);\n\t\t\n\t\tGL20.glUseProgram(pId);\n\t\t\n\t\t\/\/ Bind the texture\n\t\tGL13.glActiveTexture(GL13.GL_TEXTURE0);\n\t\tGL11.glBindTexture(GL11.GL_TEXTURE_2D, texIds[textureSelector]);\n\t\t\n\t\t\/\/ Bind to the VAO that has all the information about the vertices\n\t\tGL30.glBindVertexArray(vaoId);\n\t\tGL20.glEnableVertexAttribArray(0);\n\t\tGL20.glEnableVertexAttribArray(1);\n\t\tGL20.glEnableVertexAttribArray(2);\n\t\t\n\t\t\/\/ Bind to the index VBO that has all the information about the order of the vertices\n\t\tGL15.glBindBuffer(GL15.GL_ELEMENT_ARRAY_BUFFER, vboiId);\n\t\t\n\t\t\/\/ Draw the vertices\n\t\tGL11.glDrawElements(GL11.GL_TRIANGLES, indicesCount, GL11.GL_UNSIGNED_BYTE, 0);\n\t\t\n\t\t\/\/ Put everything back to default (deselect)\n\t\tGL15.glBindBuffer(GL15.GL_ELEMENT_ARRAY_BUFFER, 0);\n\t\tGL20.glDisableVertexAttribArray(0);\n\t\tGL20.glDisableVertexAttribArray(1);\n\t\tGL20.glDisableVertexAttribArray(2);\n\t\tGL30.glBindVertexArray(0);\n\t\t\n\t\tGL20.glUseProgram(0);\n\t\t\n\t\tthis.exitOnGLError(\"loopCycle\");\n\t}\n\t\n\tprivate void destroyOpenGL() {\t\n\t\t\/\/ Delete the texture\n\t\tGL11.glDeleteTextures(texIds[0]);\n\t\tGL11.glDeleteTextures(texIds[1]);\n\t\t\n\t\t\/\/ Delete the shaders\n\t\tGL20.glUseProgram(0);\n\t\tGL20.glDetachShader(pId, vsId);\n\t\tGL20.glDetachShader(pId, fsId);\n\t\t\n\t\tGL20.glDeleteShader(vsId);\n\t\tGL20.glDeleteShader(fsId);\n\t\tGL20.glDeleteProgram(pId);\n\t\t\n\t\t\/\/ Select the VAO\n\t\tGL30.glBindVertexArray(vaoId);\n\t\t\n\t\t\/\/ Disable the VBO index from the VAO attributes list\n\t\tGL20.glDisableVertexAttribArray(0);\n\t\tGL20.glDisableVertexAttribArray(1);\n\t\t\n\t\t\/\/ Delete the vertex VBO\n\t\tGL15.glBindBuffer(GL15.GL_ARRAY_BUFFER, 0);\n\t\tGL15.glDeleteBuffers(vboId);\n\t\t\n\t\t\/\/ Delete the index VBO\n\t\tGL15.glBindBuffer(GL15.GL_ELEMENT_ARRAY_BUFFER, 0);\n\t\tGL15.glDeleteBuffers(vboiId);\n\t\t\n\t\t\/\/ Delete the VAO\n\t\tGL30.glBindVertexArray(0);\n\t\tGL30.glDeleteVertexArrays(vaoId);\n\t\t\n\t\tthis.exitOnGLError(\"destroyOpenGL\");\n\t\t\n\t\tDisplay.destroy();\n\t}\n\t\n\tprivate int loadShader(String filename, int type) {\n\t\tStringBuilder shaderSource = new StringBuilder();\n\t\tint shaderID = 0;\n\t\t\n\t\ttry {\n\t\t\tBufferedReader reader = new BufferedReader(new FileReader(filename));\n\t\t\tString line;\n\t\t\twhile ((line = reader.readLine())\u00a0!= null) {\n\t\t\t\tshaderSource.append(line).append(\"\\n\");\n\t\t\t}\n\t\t\treader.close();\n\t\t} catch (IOException e) {\n\t\t\tSystem.err.println(\"Could not read file.\");\n\t\t\te.printStackTrace();\n\t\t\tSystem.exit(-1);\n\t\t}\n\t\t\n\t\tshaderID = GL20.glCreateShader(type);\n\t\tGL20.glShaderSource(shaderID, shaderSource);\n\t\tGL20.glCompileShader(shaderID);\n\t\t\n\t\tif (GL20.glGetShader(shaderID, GL20.GL_COMPILE_STATUS) == GL11.GL_FALSE) {\n\t\t\tSystem.err.println(\"Could not compile shader.\");\n\t\t\tSystem.exit(-1);\n\t\t}\n\t\t\n\t\tthis.exitOnGLError(\"loadShader\");\n\t\t\n\t\treturn shaderID;\n\t}\n\t\n\tprivate int loadPNGTexture(String filename, int textureUnit) {\n\t\tByteBuffer buf = null;\n\t\tint tWidth = 0;\n\t\tint tHeight = 0;\n\t\t\n\t\ttry {\n\t\t\t\/\/ Open the PNG file as an InputStream\n\t\t\tInputStream in = new FileInputStream(filename);\n\t\t\t\/\/ Link the PNG decoder to this stream\n\t\t\tPNGDecoder decoder = new PNGDecoder(in);\n\t\t\t\n\t\t\t\/\/ Get the width and height of the texture\n\t\t\ttWidth = decoder.getWidth();\n\t\t\ttHeight = decoder.getHeight();\n\t\t\t\n\t\t\t\n\t\t\t\/\/ Decode the PNG file in a ByteBuffer\n\t\t\tbuf = ByteBuffer.allocateDirect(\n\t\t\t\t\t4 * decoder.getWidth() * decoder.getHeight());\n\t\t\tdecoder.decode(buf, decoder.getWidth() * 4, Format.RGBA);\n\t\t\tbuf.flip();\n\t\t\t\n\t\t\tin.close();\n\t\t} catch (IOException e) {\n\t\t\te.printStackTrace();\n\t\t\tSystem.exit(-1);\n\t\t}\n\t\t\n\t\t\/\/ Create a new texture object in memory and bind it\n\t\tint texId = GL11.glGenTextures();\n\t\tGL13.glActiveTexture(textureUnit);\n\t\tGL11.glBindTexture(GL11.GL_TEXTURE_2D, texId);\n\t\t\n\t\t\/\/ All RGB bytes are aligned to each other and each component is 1 byte\n\t\tGL11.glPixelStorei(GL11.GL_UNPACK_ALIGNMENT, 1);\n\t\t\n\t\t\/\/ Upload the texture data and generate mip maps (for scaling)\n\t\tGL11.glTexImage2D(GL11.GL_TEXTURE_2D, 0, GL11.GL_RGB, tWidth, tHeight, 0, \n\t\t\t\tGL11.GL_RGBA, GL11.GL_UNSIGNED_BYTE, buf);\n\t\tGL30.glGenerateMipmap(GL11.GL_TEXTURE_2D);\n\t\t\n\t\t\/\/ Setup the ST coordinate system\n\t\tGL11.glTexParameteri(GL11.GL_TEXTURE_2D, GL11.GL_TEXTURE_WRAP_S, GL11.GL_REPEAT);\n\t\tGL11.glTexParameteri(GL11.GL_TEXTURE_2D, GL11.GL_TEXTURE_WRAP_T, GL11.GL_REPEAT);\n\t\t\n\t\t\/\/ Setup what to do when the texture has to be scaled\n\t\tGL11.glTexParameteri(GL11.GL_TEXTURE_2D, GL11.GL_TEXTURE_MAG_FILTER, \n\t\t\t\tGL11.GL_NEAREST);\n\t\tGL11.glTexParameteri(GL11.GL_TEXTURE_2D, GL11.GL_TEXTURE_MIN_FILTER, \n\t\t\t\tGL11.GL_LINEAR_MIPMAP_LINEAR);\n\t\t\n\t\tthis.exitOnGLError(\"loadPNGTexture\");\n\t\t\n\t\treturn texId;\n\t}\n\t\n\tprivate void exitOnGLError(String errorMessage) {\n\t\tint errorValue = GL11.glGetError();\n\t\t\n\t\tif (errorValue\u00a0!= GL11.GL_NO_ERROR) {\n\t\t\tString errorString = GLU.gluErrorString(errorValue);\n\t\t\tSystem.err.println(\"ERROR - \" + errorMessage + \": \" + errorString);\n\t\t\t\n\t\t\tif (Display.isCreated()) Display.destroy();\n\t\t\tSystem.exit(-1);\n\t\t}\n\t}\n}\n<\/pre><h2>Credit<\/h2><p><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.mathiasverboven.net\">Mathias Verboven (moci)<\/a>\n<\/p>","created_at":"2018-11-07T16:45:23+0000","updated_at":"2018-11-07T16:45:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"wiki.lwjgl.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":354,"title":"Blog | My Clojure Workflow, Reloaded","url":"http:\/\/thinkrelevance.com\/blog\/2013\/06\/04\/clojure-workflow-reloaded","content":"<h3 class=\"post-title\"><a href=\"http:\/\/thinkrelevance.com\/blog\/2013\/06\/04\/clojure-workflow-reloaded\">My Clojure Workflow, Reloaded<\/a><\/h3><p>This is a companion post to my discussion with <a href=\"http:\/\/thinkrelevance.com\/team\/members\/craig-andera\">Craig Andera<\/a> on <a href=\"http:\/\/thinkrelevance.com\/blog\/2013\/05\/29\/stuart-sierra-episode-032\">Relevance Podcast Episode 32<\/a> and my Clojure\/West talk <a href=\"http:\/\/www.infoq.com\/presentations\/Clojure-Large-scale-patterns-techniques\">Clojure in the Large<\/a>. I've talked about various bits and pieces of this workflow at other times, too, but I'll try to bring it all together here in the hopes that others will find it useful.<\/p><p>One of the great pleasures of working with a dynamic language is being able to build a system while simultaneously interacting with it. To make this possible, first you need the ability to redefine parts of the program while it is running: Clojure provides this capability admirably. However, some aspects of Clojure's runtime are not quite as <em>late-binding<\/em> as one might wish for interactive development. For example, the effect of a changed macro definition will not be seen until code which uses the macro has been recompiled. Changes to methods of a <code>defrecord<\/code> or <code>deftype<\/code> will not have any effect on existing instances of that type.<\/p><p>The facilities that Clojure provides for loading code from files are not sufficient to deal with these issues. I wrote the second version of <a href=\"https:\/\/github.com\/clojure\/tools.namespace\">tools.namespace<\/a> to make a \"smarter\" <code>require<\/code> that recognizes dependencies between namespaces and reloads them appropriately.<\/p><p>But tools.namespace is only part of the story. To really get the benefit of interactive development, I want to ensure that the version of the application I am currently interacting with is congruent with the source files I'm editing. That means not only that the application must be running the most up-to-date version of the code, but also that any <strong>state<\/strong> in the application was <strong>produced<\/strong> by that same code. It is dangerously easy, when changing and reloading code at the REPL, to get an application into a state which <strong>could not have been reached<\/strong> by the code it is currently running.<\/p><p>Therefore, after every significant code change, I want to restart the application from scratch. But I don't want to restart the JVM and reload all my Clojure code in order to do it: that takes too long and is too disruptive to my workflow. Instead, I want to design my application in such a way that I can quickly shut it down, discard any transient state it might have built up, start it again, and return to a similar state. And when I say <em>quickly<\/em>, I mean that the whole process should take less than a second.<\/p><p>To achieve this goal, I make the application itself into a transient object. Instead of the application being a singleton tied to a JVM process, I write code to construct <strong>instances<\/strong> of my application, possibly many of them within one JVM. Each time I make a change, I discard the old instance and construct a new one. The technique is similar to dealing with virtual machines in a cloud environment: rather than try to transition a VM from an old state to a new state, we simply discard the old one and spin up a new one.<\/p><p>Designing applications this way requires discipline. First and foremost, all state must be local. Any global state, anywhere, breaks the whole model. Second, all resources acquired by the application instance must be carefully managed so that they can be released when the instance is destroyed.<\/p><p>Enough talk. Here's how it works.<\/p><p>In some \"main\" namespace, I provide a constructor function for the application. I usually call it <code>system<\/code> because it represents the whole system I am working on.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In src\/com\/example\/my_project\/system.clj\n(ns com.example.my-project.system)\n(defn system\n  \"Returns a new instance of the whole application.\"\n  []\n  ...)<\/pre><\/div><\/div><p>The system constructor can optionally take parameters which specify its configuration.<\/p><p>Creating a system is not the same as starting it and should not have side effects. Usually the system constructor will create instances of other components it depends on and return a data structure such as a map or <code>defrecord<\/code> which contains them. My system <strong>instance<\/strong> might look something like this:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>{:db        {:uri \"datomic:mem:\/\/dev\"}\n :scheduler #&lt;ScheduledThreadPoolExecutorService ...&gt;\n :cache     #&lt;Atom {}&gt;\n :handler   #&lt;Fn ...&gt;\n :server    #&lt;Jetty ...&gt;}<\/pre><\/div><\/div><p>Sometimes I have different versions of the constructor that produce different systems for interactive development, testing, and production.<\/p><p>Notice that some things which are \"global\" from the point of view of the application, such as my web server and scheduled thread pool, become \"local\" instances in this data structure. Any function which needs one of these components has to take it as a parameter. This isn't as burdensome as it might seem: each function gets, at most, one extra argument providing the \"context\" in which it operates. That context could be the entire system object, but more often will be some subset. With judicious use of lexical closures, the extra arguments disappear from most code. In addition to enabling more interactive development, this approach makes testing easier. See my post <a href=\"http:\/\/stuartsierra.com\/2013\/03\/29\/perils-of-dynamic-scope\">On the Perils of Dynamic Scope<\/a> for more background, as well as the Clojure\/West talk when it comes out.<\/p><p>Next, I have functions to start and stop the system. Ideally, these behave like real functions, in that they return a new value representing the \"started\" or \"stopped\" system, but they also have to perform side effects along the way, such as opening a connection to a database or starting a web server.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In src\/com\/example\/my_project\/system.clj\n(defn start\n  \"Performs side effects to initialize the system, acquire resources,\n  and start it running. Returns an updated instance of the system.\"\n  [system]\n  ...)\n(defn stop\n  \"Performs side effects to shut down the system and release its\n  resources. Returns an updated instance of the system.\"\n  [system]\n  ...)<\/pre><\/div><\/div><p>These functions can call similar start\/stop functions on sub-systems in turn. In the past, I've talked about a \"Lifecycle\" protocol containing <code>start<\/code> and <code>stop<\/code> methods. It's not necessary, but is sometimes useful to ensure that all components of the system can be started and stopped in a consistent way.<\/p><p>There's usually a bit of trial-and-error while I get the start\/stop functions working correctly. If something in start\/stop throws an exception, I could easily end up in a state where a sub-system has acquired a resource \u2014 such as a socket connection \u2014 but I do not have any handle on that sub-system with which to shut it down and release the resource. In that situation, there's nothing for it but to restart the JVM.<\/p><p>You probably know that the Clojure REPL starts by default in the <code>user<\/code> namespace. In addition, if there is a file named <code>user.clj<\/code> at the root of the Java classpath, Clojure will load that file automatically when it starts.<\/p><p>You probably don't want <code>user.clj<\/code> to be loaded in a deployed production app or library release, but by using Leiningen 2 <strong>profiles<\/strong> we can ensure that it is only loaded during development.<\/p><p>In my Leiningen <code>project.clj<\/code> file, I create a <code>:dev<\/code> profile with an extra <code>:source-paths<\/code> directory, plus whatever dependencies I want to use during development. tools.namespace has to be there, and I frequently add testing\/development tools such as <a href=\"https:\/\/github.com\/clojure\/java.classpath\">java.classpath<\/a> or <a href=\"https:\/\/github.com\/hugoduncan\/criterium\">Criterium<\/a>.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In project.clj:\n(defproject com.example\/my-project \"0.1.0-SNAPSHOT\"\n  :dependencies [[org.clojure\/clojure \"1.5.1\"]]\n  :profiles {:dev {:source-paths [\"dev\"]\n                   :dependencies [[org.clojure\/tools.namespace \"0.2.3\"]\n                                  [org.clojure\/java.classpath \"0.2.0\"]]}})<\/pre><\/div><\/div><p>Leiningen will automatically merge the <code>:dev<\/code> profile into the project configuration for the <code>repl<\/code>, <code>test<\/code>, and <code>run<\/code> tasks, but not <code>jar<\/code> or <code>uberjar<\/code>. That means if I deploy my application (or release a library) as a JAR file, the files in <code>dev<\/code> will be excluded.<\/p><p>I create a <code>user.clj<\/code> file in the <code>dev<\/code> directory which defines a normal namespace called <code>user<\/code> and refers a bunch of symbols I commonly use during development, as well as the symbols to construct, start, and stop the system.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In dev\/user.clj\n(ns user\n  (:require [clojure.java.io :as io]\n            [clojure.string :as str]\n            [clojure.pprint :refer (pprint)]\n            [clojure.repl :refer :all]\n            [clojure.test :as test]\n            [clojure.tools.namespace.repl :refer (refresh refresh-all)]\n            [com.example.my-project.system :as system]))<\/pre><\/div><\/div><p>Also in <code>user.clj<\/code>, I have a few things that I will only use during development, starting with a global Var to hold the system itself:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In dev\/user.clj\n(def system nil)<\/pre><\/div><\/div><p><em>Now wait a minute<\/em>, you might say, <em>isn't that the global state you told us to avoid?<\/em> It would be, if it were part of the application. Instead it's a container in which I can put the current instance of the application. I'm only going to use it for interactive development.<\/p><p>The <code>system<\/code> Var is manipulated by the following functions:<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In dev\/user.clj\n(defn init\n  \"Constructs the current development system.\"\n  []\n  (alter-var-root #'system\n    (constantly (system\/system))))\n(defn start\n  \"Starts the current development system.\"\n  []\n  (alter-var-root #'system system\/start))\n(defn stop\n  \"Shuts down and destroys the current development system.\"\n  []\n  (alter-var-root #'system\n    (fn [s] (when s (system\/stop s)))))\n(defn go\n  \"Initializes the current development system and starts it running.\"\n  []\n  (init)\n  (start))<\/pre><\/div><\/div><p>The exact division of these functions isn't important. Sometimes I omit <code>init<\/code> and <code>start<\/code> and just have <code>go<\/code>. The important thing is to have one function that creates and starts the system, and another function that tears it down.<\/p><p>Finally, the heart of my workflow: the <code>reset<\/code> function. This is one function which I can call at the REPL to 1) stop the current application instance; 2) reload any source files that have changed; and 3) create and start a new application instance.<\/p><div class=\"CodeRay\"><div class=\"code\"><pre>;; In dev\/user.clj\n(defn reset []\n  (stop)\n  (refresh :after 'user\/go))<\/pre><\/div><\/div><p>The real work of reloading files is handled by the <code>clojure.tools.namespace.repl\/refresh<\/code> function. It takes my <code>go<\/code> function as an argument, but <code>go<\/code> has to be passed as a namespace-qualified symbol so that it can be resolved <strong>after<\/strong> the <code>user<\/code> namespace has been reloaded. (This is a trick that <code>refresh<\/code> knows how to do.)<\/p><p>I do all my Clojure development in <a href=\"http:\/\/www.gnu.org\/software\/emacs\/\">Emacs<\/a> using <a href=\"https:\/\/github.com\/kingtim\/nrepl.el\">nREPL.el<\/a>, but nothing about this workflow is Emacs-specific. It should work with any environment that provides a REPL, as long as it doesn't try to do any code-reloading of its own. (For example, the reload-on-every-request functionality of <a href=\"https:\/\/github.com\/ring-clojure\/ring\">ring-devel<\/a> is incompatible with tools.namespace.) The fact that I use Emacs as my REPL is one reason I use <code>user.clj<\/code> instead of <code>:repl-options<\/code> in Leiningen's <code>project.clj<\/code>: those options have no effect on remote nREPL sessions.<\/p><p>The first thing I do when I start work is launch an <a href=\"https:\/\/github.com\/clojure\/tools.nrepl\">nREPL<\/a> session and call <code>reset<\/code>. Now my application is running and I can start working on it. Every time I make a change, I save the file and call <code>reset<\/code> at the REPL. (I have an <a href=\"https:\/\/github.com\/stuartsierra\/dotfiles\/blob\/d0d1c46ccc4fdd8d2add363615e625cc29d035b0\/.emacs#L307-L312\">Elisp helper function<\/a> that I can bind to a keystroke.) Presto! My application is running again in a clean state with all the new code.<\/p><p>Rather than switching the REPL among several namespaces, I generally stay in <code>user<\/code>, where I have all my development tools like <code>clojure.pprint<\/code> and <code>clojure.repl<\/code>. I use the REPL itself for examining the application's state and testing individual functions. I frequently define little helper functions to examine the state of the application as I work on it, all of which are accessible by navigating the <code>system<\/code> object.<\/p><p>Anything I want to hang on to, such as a snippet of test data, I define in the <code>user.clj<\/code> file, because tools.namespace will destroy any Vars I created with <code>def<\/code> at the REPL.<\/p><p>This process isn't perfect by any means. One of the more irritating aspects is that any syntactic errors in a source file prevent <strong>all<\/strong> the code from being loaded, including <code>user.clj<\/code>. If a file fails to compile during the tools.namespace reloading process, any namespaces which depend on it no longer exist. So the <code>reset<\/code> function isn't available to call, nor are any of my aliases or referred symbols in the <code>user<\/code> namespace.<\/p><p>As a work-around, in tools.namespace 0.2.3 I added a feature to recover aliases and referred symbols in the current REPL namespace after a failed reload. This isn't perfect: the <code>reset<\/code> function still doesn't exist. But at least I can call the <code>refresh<\/code> function from tools.namespace without typing out its fully-qualified name <code>clojure.tools.namespace.repl\/refresh<\/code>. Once I have successfully reloaded all the source files with <code>refresh<\/code>, I can call <code>reset<\/code> again to start the app.<\/p><p>A slightly worse problem occurs when starting a new REPL process: if there are any compilation errors in something loaded by <code>user.clj<\/code>, then the REPL will not start at all. I try to avoid this by starting the REPL from a known working commit, then only changing code after it's running. I also try not to commit any code which does not compile, but sometimes it happens. :)<\/p><p>Occasionally I do get my application into a state that I cannot recover from. Usually this happens when something in a <code>start<\/code> or <code>stop<\/code> function throws an exception. At that point, some component of the application may be in a broken state but I don't have a reference to it that I can use to shut it down. If that component acquired external resources which I need to release before restarting it, e.g. socket connections, then there's basically nothing I can do but restart the JVM. Fortunately, these situations usually only occur while I'm writing the start\/stop functions themselves, so after a few development cycles to get them working I don't have to worry about it.<\/p><p>The central thrust of this approach is to design your application so that you can construct multiple <strong>instances<\/strong> of it within a single JVM process. That's ideal for development, but what about production?<\/p><p>If you control the entry point to your application process, it's easy. Just write a <code>-main<\/code> function that creates a single instance of your application and starts it.<\/p><p>But often we deploy apps to environments where we do <strong>not<\/strong> control the <code>-main<\/code> function. For example, <a href=\"https:\/\/github.com\/ring-clojure\/ring\">Ring<\/a> web apps deployed to a Servlet container have no <code>-main<\/code>. Furthermore, they expect a static reference to a Var which contains the root web handler function. If that handler is meant to be a closure over some contextual state, there's no place to construct it.<\/p><p>There are a couple of ways to work around this. One is to have a separate namespace in a \"production\" profile that constructs a single instance of the application and assigns it to a global Var. Alternatively, if the framework provides an \"initialization\" hook (as <a href=\"https:\/\/github.com\/weavejester\/lein-ring\/blob\/ec519295c6f234e90b8304ae1d58d48a4b5fb5b1\/README.md#general-options\">lein-ring does<\/a>), you can use that to create the application instance and store it in a global Var. The root web handler function, created exclusively for production deployment, can pass the system object to functions that need it.<\/p><p>I'm continually tweaking this process, looking for improvements, but overall I'm pretty happy with it. It has enabled me to work rapidly on some fairly large applications. Best of all, it's agnostic with regard to development tools. You can adapt this workflow to any build tool that can substitute different CLASSPATHs for different circumstances.<\/p><p>Some of my Relevance coworkers like this approach, others find it too constraining. The <a href=\"http:\/\/pedestal.io\/\">Pedestal<\/a> team uses pieces of this technique, such as the <code>:dev<\/code> profile, but without tools.namespace. They were annoyed that compiler errors prevented them from starting a new REPL, so they came up with a variation that uses a <a href=\"https:\/\/github.com\/pedestal\/pedestal\/blob\/239d2c7899ac3466a32ab124ab3850e56abeb0c5\/app-template\/src\/leiningen\/new\/pedestal_app\/annotated\/dev\/user.clj\">function<\/a> in <code>user.clj<\/code> to load another file called <a href=\"https:\/\/github.com\/pedestal\/pedestal\/blob\/239d2c7899ac3466a32ab124ab3850e56abeb0c5\/app-template\/src\/leiningen\/new\/pedestal_app\/annotated\/dev\/dev.clj\">dev.clj<\/a>.<\/p><p>I wanted to write about this to clarify and expand upon things I've presented elsewhere. Check out the <a href=\"http:\/\/thinkrelevance.com\/blog\/2013\/05\/29\/stuart-sierra-episode-032\">podcast<\/a> to hear me talk more about this approach specifically, and the <a href=\"http:\/\/www.infoq.com\/presentations\/Clojure-Large-scale-patterns-techniques\">Clojure\/West video<\/a> for more background.<\/p><p><strong>UPDATES:<\/strong><\/p><ul><li><p>Fixed name of clojure.string in example code (June 4)<\/p><\/li>\n<li><p>More <a href=\"https:\/\/news.ycombinator.com\/item?id=5819487\">comments on Hacker News<\/a> (June 4)<\/p><\/li>\n<li><p>Fixed description of which Leiningen tasks include <code>:dev<\/code> profile, per Phil's comment (June 5)<\/p><\/li>\n<li><p>Added links to released Clojure\/West video (July 2)<\/p><\/li>\n<li><p>I created a <a href=\"https:\/\/github.com\/stuartsierra\/reloaded\">Leiningen template<\/a> based on this article (August 5)<\/p><\/li>\n<\/ul>","created_at":"2018-11-07T16:45:23+0000","updated_at":"2018-11-07T16:45:23+0000","annotations":[],"mimetype":"text\/html","reading_time":13,"domain_name":"thinkrelevance.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":355,"title":"Who Should Read This Handbook? | Theme Developer Handbook","url":"https:\/\/developer.wordpress.org\/themes\/getting-started\/who-should-read-this-handbook\/","content":"<div id=\"secondary\" class=\"widget-area\" role=\"complementary\">\n\t\t\n\t\t<div id=\"secondary-content\">\n\t\t\t\t\t\t<aside id=\"search-3\" class=\"widget widget_search\"><div class=\"search-section section clear hide-if-js\">\n\n\n\t\n\t<form role=\"search\" method=\"get\" class=\"searchform searchform-handbook\" action=\"https:\/\/developer.wordpress.org\/themes\/\">\n\t\t<div><label for=\"search-field\" class=\"screen-reader-text\">Search for:<\/label>\n\t\t<input type=\"text\" id=\"search-field\" class=\"search-field\" placeholder=\"Search \u2026\" value=\"\" name=\"s\" \/><button class=\"button button-primary button-search\"><i class=\"dashicons dashicons-search\">Search plugins<\/i><\/button>\n\t\n\t<\/div><\/form>\n\n<\/div>\n<\/aside><aside id=\"nav_menu-2\" class=\"widget widget_nav_menu\"><h2 class=\"widget-title\">Chapters<\/h2><div class=\"menu-theme-table-of-contents-container\"><ul id=\"menu-theme-table-of-contents\" class=\"menu\"><li id=\"menu-item-5797\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook current-theme-handbook-ancestor current-menu-ancestor current-menu-parent current-theme-handbook-parent menu-item-has-children menu-item-5797\"><a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/\">Getting Started<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-5799\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook current-menu-item menu-item-5799\"><a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/who-should-read-this-handbook\/\">Who Should Read This Handbook?<\/a><\/li>\n\t<li id=\"menu-item-5802\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-5802\"><a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/what-is-a-theme\/\">What is a Theme?<\/a><\/li>\n\t<li id=\"menu-item-5805\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-5805\"><a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/wordpress-licensing-the-gpl\/\">WordPress Licensing &amp; the GPL<\/a><\/li>\n\t<li id=\"menu-item-5808\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-5808\"><a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/setting-up-a-development-environment\/\">Setting up a Development Environment<\/a><\/li>\n\t<li id=\"menu-item-5814\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-5814\"><a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/theme-development-examples\/\">Theme Development Examples<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-5828\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-5828\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/\">Theme Basics<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11989\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-11989\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/template-files\/\">Template Files<\/a>\n\t<ul class=\"sub-menu\"><li id=\"menu-item-24946\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-24946\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/main-stylesheet-style-css\/\">Main Stylesheet (style.css)<\/a><\/li>\n\t<\/ul><\/li>\n\t<li id=\"menu-item-11344\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11344\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/post-types\/\">Post Types<\/a><\/li>\n\t<li id=\"menu-item-5837\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-5837\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/organizing-theme-files\/\">Organizing Theme Files<\/a><\/li>\n\t<li id=\"menu-item-11293\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11293\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/template-hierarchy\/\">Template Hierarchy<\/a><\/li>\n\t<li id=\"menu-item-11292\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11292\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/template-tags\/\">Template Tags<\/a><\/li>\n\t<li id=\"menu-item-11291\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11291\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/the-loop\/\">The Loop<\/a><\/li>\n\t<li id=\"menu-item-11313\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11313\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/theme-functions\/\">Theme Functions<\/a><\/li>\n\t<li id=\"menu-item-11312\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11312\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/linking-theme-files-directories\/\">Linking Theme Files &amp; Directories<\/a><\/li>\n\t<li id=\"menu-item-11311\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11311\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/including-css-javascript\/\">Including CSS &amp; JavaScript<\/a><\/li>\n\t<li id=\"menu-item-11310\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11310\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/conditional-tags\/\">Conditional Tags<\/a><\/li>\n\t<li id=\"menu-item-11347\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11347\"><a href=\"https:\/\/developer.wordpress.org\/themes\/basics\/categories-tags-custom-taxonomies\/\">Categories, Tags, &amp; Custom Taxonomies<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-17411\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-17411\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/\">Template Files Section<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-17416\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-17416\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/post-template-files\/\">Post Templates<\/a><\/li>\n\t<li id=\"menu-item-11294\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11294\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/page-template-files\/\">Page Templates<\/a><\/li>\n\t<li id=\"menu-item-17414\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-17414\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/attachment-template-files\/\">Attachment Template Files<\/a><\/li>\n\t<li id=\"menu-item-17413\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-17413\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/custom-post-type-template-files\/\">Custom Post Type Template Files<\/a><\/li>\n\t<li id=\"menu-item-17412\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-17412\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/partial-and-miscellaneous-template-files\/\">Partial and Miscellaneous Template Files<\/a>\n\t<ul class=\"sub-menu\"><li id=\"menu-item-11366\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11366\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/partial-and-miscellaneous-template-files\/comments\/\">Comment Template<\/a><\/li>\n\t<\/ul><\/li>\n\t<li id=\"menu-item-11343\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11343\"><a href=\"https:\/\/developer.wordpress.org\/themes\/template-files-section\/taxonomy-templates\/\">Taxonomy Templates<\/a><\/li>\n\t<li id=\"menu-item-25204\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25204\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/404-pages\/\">404 Pages<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-5827\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-5827\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/\">Theme Functionality<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-17365\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-17365\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/core-supported\/\">Core-Supported Features<\/a>\n\t<ul class=\"sub-menu\"><li id=\"menu-item-25203\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25203\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/administration-menus\/\">Administration Menus<\/a><\/li>\n\t\t<li id=\"menu-item-11327\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11327\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/custom-headers\/\">Custom Headers<\/a><\/li>\n\t\t<li id=\"menu-item-25160\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25160\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/custom-logo\/\">Custom Logo<\/a><\/li>\n\t\t<li id=\"menu-item-11376\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11376\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/post-formats\/\">Post Formats<\/a><\/li>\n\t\t<li id=\"menu-item-25202\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25202\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/sticky-posts\/\">Sticky Posts<\/a><\/li>\n\t<\/ul><\/li>\n\t<li id=\"menu-item-11326\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11326\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/sidebars\/\">Sidebars<\/a><\/li>\n\t<li id=\"menu-item-11325\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11325\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/widgets\/\">Widgets<\/a><\/li>\n\t<li id=\"menu-item-11324\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11324\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/navigation-menus\/\">Navigation Menus<\/a><\/li>\n\t<li id=\"menu-item-11367\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11367\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/pagination\/\">Pagination<\/a><\/li>\n\t<li id=\"menu-item-11365\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-11365\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/media\/\">Media<\/a>\n\t<ul class=\"sub-menu\"><li id=\"menu-item-25197\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25197\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/media\/audio\/\">Audio<\/a><\/li>\n\t\t<li id=\"menu-item-25144\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25144\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/media\/images\/\">Images<\/a><\/li>\n\t\t<li id=\"menu-item-25158\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25158\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/media\/galleries\/\">Galleries<\/a><\/li>\n\t\t<li id=\"menu-item-25157\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25157\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/media\/video\/\">Video<\/a><\/li>\n\t<\/ul><\/li>\n\t<li id=\"menu-item-11364\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11364\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/featured-images-post-thumbnails\/\">Featured Images &amp; Post Thumbnails<\/a><\/li>\n\t<li id=\"menu-item-11375\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11375\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/internationalization\/\">Internationalization<\/a><\/li>\n\t<li id=\"menu-item-11374\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11374\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/localization\/\">Localization<\/a><\/li>\n\t<li id=\"menu-item-11386\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11386\"><a href=\"https:\/\/developer.wordpress.org\/themes\/functionality\/accessibility\/\">Accessibility<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11419\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-11419\"><a href=\"https:\/\/developer.wordpress.org\/themes\/customize-api\/\">Theme Options \u2013 The Customizer API<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-25020\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25020\"><a href=\"https:\/\/developer.wordpress.org\/themes\/customize-api\/customizer-objects\/\">Customizer Objects<\/a><\/li>\n\t<li id=\"menu-item-25023\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25023\"><a href=\"https:\/\/developer.wordpress.org\/themes\/customize-api\/tools-for-improved-user-experience\/\">Tools for Improved User Experience<\/a><\/li>\n\t<li id=\"menu-item-25022\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25022\"><a href=\"https:\/\/developer.wordpress.org\/themes\/customize-api\/the-customizer-javascript-api\/\">The Customizer JavaScript API<\/a><\/li>\n\t<li id=\"menu-item-25021\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25021\"><a href=\"https:\/\/developer.wordpress.org\/themes\/customize-api\/javascriptunderscore-js-rendered-custom-controls\/\">JavaScript\/Underscore.js-Rendered Custom Controls<\/a><\/li>\n\t<li id=\"menu-item-25019\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25019\"><a href=\"https:\/\/developer.wordpress.org\/themes\/customize-api\/advanced-usage\/\">Advanced Usage<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-25095\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-25095\"><a href=\"https:\/\/developer.wordpress.org\/themes\/theme-security\/\">Theme Security<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-25093\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25093\"><a href=\"https:\/\/developer.wordpress.org\/themes\/theme-security\/data-sanitization-escaping\/\">Data Sanitization\/Escaping<\/a><\/li>\n\t<li id=\"menu-item-25094\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25094\"><a href=\"https:\/\/developer.wordpress.org\/themes\/theme-security\/data-validation\/\">Data Validation<\/a><\/li>\n\t<li id=\"menu-item-25096\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25096\"><a href=\"https:\/\/developer.wordpress.org\/themes\/theme-security\/using-nonces\/\">Using Nonces<\/a><\/li>\n\t<li id=\"menu-item-25185\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25185\"><a href=\"https:\/\/developer.wordpress.org\/themes\/theme-security\/common-vulnerabilities\/\">Common Vulnerabilities<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-37686\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-37686\"><a href=\"https:\/\/developer.wordpress.org\/themes\/theme-privacy\/\">Theme Privacy<\/a><\/li>\n<li id=\"menu-item-5826\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-5826\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/\">Advanced Theme Topics<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11420\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11420\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/child-themes\/\">Child Themes<\/a><\/li>\n\t<li id=\"menu-item-11415\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11415\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/ui-best-practices\/\">UI Best Practices<\/a><\/li>\n\t<li id=\"menu-item-11414\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11414\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/javascript-best-practices\/\">JavaScript Best Practices<\/a><\/li>\n\t<li id=\"menu-item-11413\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11413\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/theme-testing\/\">Theme Testing<\/a><\/li>\n\t<li id=\"menu-item-11412\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11412\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/validating-your-theme\/\">Validating Your Theme<\/a><\/li>\n\t<li id=\"menu-item-11411\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11411\"><a href=\"https:\/\/developer.wordpress.org\/themes\/advanced-topics\/plugin-api-hooks\/\">Plugin API Hooks<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-5825\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-5825\"><a href=\"https:\/\/developer.wordpress.org\/themes\/release\/\">Releasing Your Theme<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11440\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11440\"><a href=\"https:\/\/developer.wordpress.org\/themes\/release\/required-theme-files\/\">Required Theme Files<\/a><\/li>\n\t<li id=\"menu-item-11439\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11439\"><a href=\"https:\/\/developer.wordpress.org\/themes\/release\/testing\/\">Testing<\/a><\/li>\n\t<li id=\"menu-item-11437\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11437\"><a href=\"https:\/\/developer.wordpress.org\/themes\/release\/theme-review-guidelines\/\">Theme Review Guidelines<\/a><\/li>\n\t<li id=\"menu-item-11436\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11436\"><a href=\"https:\/\/developer.wordpress.org\/themes\/release\/writing-documentation\/\">Writing Documentation<\/a><\/li>\n\t<li id=\"menu-item-11435\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-11435\"><a href=\"https:\/\/developer.wordpress.org\/themes\/release\/submitting-your-theme-to-wordpress-org\/\">Submitting Your Theme to WordPress.org<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-25055\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-25055\"><a href=\"https:\/\/developer.wordpress.org\/themes\/references\/\">References<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-25056\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25056\"><a href=\"https:\/\/developer.wordpress.org\/themes\/references\/list-of-template-tags\/\">List of Template Tags<\/a><\/li>\n\t<li id=\"menu-item-25069\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25069\"><a href=\"https:\/\/developer.wordpress.org\/themes\/references\/list-of-conditional-tags\/\">List of Conditional Tags<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11272\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-has-children menu-item-11272\"><a href=\"https:\/\/developer.wordpress.org\/themes\/credits\/\">Credits<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-25184\" class=\"menu-item menu-item-type-post_type menu-item-object-theme-handbook menu-item-25184\"><a href=\"https:\/\/developer.wordpress.org\/themes\/credits\/feedback\/\">Feedback<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/div><\/aside><\/div>\n\t<\/div>\n\n\n\t\t<main id=\"primary\" class=\"site-main post-5796 theme-handbook type-theme-handbook status-publish hentry type-handbook\" role=\"main\"><div class=\"breadcrumb-trail breadcrumbs\" itemprop=\"breadcrumb\">\n\t\t\tBrowse: \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\" title=\"WordPress Developer Resources\" rel=\"home\">Home<\/a> \/ \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\/themes\/\">Theme Handbook<\/a> \/ \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\/themes\/getting-started\/\" title=\"Getting Started\">Getting Started<\/a> \/ Who Should Read This Handbook?\n\t\t<\/div>\n\t\t\n\t\t\t\n<h1>Who Should Read This Handbook?<\/h1>\n\n\n<div class=\"table-of-contents\"><h2>Topics<\/h2><ul class=\"items\"><li><a href=\"#your-skill-level\">Your skill level<\/a><\/li>\n<li><a href=\"#what-this-handbook-will-cover\">What this handbook will cover<\/a><\/li><\/ul><\/div>\n<p>The Theme Developer Handbook is a self-contained resource to help you learn the basics of WordPress theme development. You should read this handbook to accomplish a number of things:<\/p>\n<ul><li>develop a child theme meant to work with a parent theme;<\/li>\n<li>create a new theme based on an existing one;<\/li>\n<li>understand the inner workings of themes in general; and<\/li>\n<li>develop a theme based on nothing but your imagination.<\/li>\n<\/ul><h2 class=\"toc-heading\" id=\"your-skill-level\" tabindex=\"-1\">Your skill level <a href=\"#your-skill-level\" class=\"anchor\">#Your skill level<\/a><\/h2>\n<p>To get the most out of this handbook, you should have\u00a0an understanding and some experience with web technologies, such as HTML, CSS, and PHP. You should also be comfortable with setting up and configuring websites using WordPress.<\/p>\n<div class=\"callout callout-info\"><p>Note: An understanding of how MySQL databases work, as well as server technologies in general, is helpful but not a requirement for developing WordPress themes. Likewise, while JavaScript knowledge can be useful, it is not required. <\/p>\n<\/div>\n<p class=\"toc-jump\"><a href=\"#top\">Top \u2191<\/a><\/p><h2 class=\"toc-heading\" id=\"what-this-handbook-will-cover\" tabindex=\"-1\">What this handbook will cover <a href=\"#what-this-handbook-will-cover\" class=\"anchor\">#What this handbook will cover<\/a><\/h2>\n<p>This handbook provides the <strong>basic information required to develop WordPress themes<\/strong>, including in-depth coverage of essential template tags and functions.<\/p>\n<p>WordPress is a vast subject. Covering every possible tag, function, and scenario is beyond the scope of this handbook.\u00a0If you need help with specific functions, you can find that information in the <a href=\"https:\/\/developer.wordpress.org\/reference\">code reference<\/a>.<\/p>\n<p>The\u00a0purpose of this handbook is to give you a solid foundation for WordPress theme development, to provide\u00a0step-by-step instruction in building\u00a0basic themes,\u00a0and to provide tips and resources useful in furthering your skills.<\/p>\n\n\n<div class=\"bottom-of-entry\">\u00a0<\/div>\n\n\t\t\t\n\t\t<\/main>","created_at":"2018-11-07T16:45:23+0000","updated_at":"2018-11-07T16:45:23+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"developer.wordpress.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":356,"title":"Better, Simpler Grid Systems \u2014 Solved by Flexbox \u2014 Cleaner, hack-free CSS","url":"https:\/\/philipwalton.github.io\/solved-by-flexbox\/demos\/grids\/","content":"<p>Most grid systems today use one of two layout methods: <code>float<\/code> or <code>inline-block<\/code>. But neither of these methods were really intended to be used for layout and as a result have pretty significant problems and limitations.<\/p><p>Using floats requires clearing them which has a whole host of layout issues, most notoriously that clearing an element sometimes forces it below an unrelated part of the page (take this <a href=\"https:\/\/github.com\/twbs\/bootstrap\/issues\/295#issuecomment-2282969\">Bootstrap issue<\/a> for example). In addition, clearing floats usually requires using both before and after pseudo-elements, preventing you from using them for something else.<\/p><p>Inline block layouts must address the problem of <a href=\"http:\/\/css-tricks.com\/fighting-the-space-between-inline-block-elements\/\">white-space between inline-block items<\/a>, and all of the <a href=\"http:\/\/davidwalsh.name\/remove-whitespace-inline-block\">solutions<\/a> to that problem are <a href=\"https:\/\/github.com\/suitcss\/components-grid\/blob\/master\/lib\/grid.css#L30\">hacky<\/a> and <a href=\"https:\/\/twitter.com\/thierrykoblentz\/status\/305152267374428160\">annoying<\/a>.<\/p><p>Flexbox not only eliminates these problems, it opens up an entirely new world of possibilities.<\/p><h2>Features of a Flexbox Grid System<\/h2><p>Grid systems usually come with a myriad of sizing options, but the vast majority of the time you just want two or three elements side-by-side. Given this, why should we be required to put sizing classes on every single cell?<\/p><p>Listed below are some of my criteria for an ideal grid system. Fortunately, with Flexbox we get most of these features for free.<\/p><ul><li>By default, each grid cell is the same width and height as every other cell in the row. Basically they all size to fit by default.<\/li><li>For finer control, you can add sizing classes to individual cells. Without these classes, the cells simply divide up the available space as usual.<\/li><li>For responsive grids, you can add media query-specific classes to the cells.<\/li><li>Individual cells can be aligned vertically to the top, bottom, or middle.<\/li><li>When you want all of the cells in a grid to have the same sizing, media, or alignment values, you should be able to just add a single class to the container to avoid unnecessary repetition.<\/li><li>Grids can be nested as many levels deep as needed.<\/li><\/ul><h3>Basic Grids<\/h3><p>The grid cells below do not specify any widths, they just naturally space themselves equally and expand to fit the entire row. They\u2019re also equal height by default.<\/p><div class=\"Grid Grid--gutters Grid--flexCells\"><div class=\"Grid-cell\"><p>Full-height, even when my content doesn't fill the space.<\/p><\/div><div class=\"Grid-cell\"><p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vestibulum mollis velit non gravida venenatis. Praesent consequat lectus purus, ut scelerisque velit condimentum eu. Maecenas sagittis ante ut turpis varius interdum. Quisque tellus ipsum, eleifend non ipsum id, suscipit ultricies neque.<\/p><\/div><\/div><h3>Individual Sizing<\/h3><p>When equal widths aren\u2019t what you want, you can add sizing classes to individual cells. Cells without sizing classes simply divide up the remaining space as normal.<\/p><p>The cells below labeled \u201cauto\u201d do not have sizing classes specified.<\/p><h3>Responsive<\/h3><p>Responsive Grids work by adding media classes to the Grid cells or containers. When those media values are met, the grids automatically adjust accordingly.<\/p><p>The cells below should be full width by default and scaled to fit above <code>48em<\/code>. Resize your browser to see them in action.<\/p><div class=\"Grid Grid--gutters Grid--full large-Grid--fit u-textCenter\"><div class=\"Grid-cell\"><p>Full \/ Halves<\/p><\/div><div class=\"Grid-cell\"><p>Full \/ Halves<\/p><\/div><\/div><div class=\"Grid Grid--gutters Grid--full large-Grid--fit u-textCenter\"><div class=\"Grid-cell\"><p>Full \/ Thirds<\/p><\/div><div class=\"Grid-cell\"><p>Full \/ Thirds<\/p><\/div><div class=\"Grid-cell\"><p>Full \/ Thirds<\/p><\/div><\/div><h3>Grid-ception<\/h3><p>Grid components are infinitely nestable inside of other grid components.<\/p><h2>Alignment Features<\/h2><h3>Top-aligned Grid Cells<\/h3><div class=\"Grid Grid--gutters Grid--top\"><div class=\"Grid-cell\"><p>This cell should be top-aligned.<\/p><\/div><div class=\"Grid-cell u-1of2\"><p>Pellentesque sagittis vel erat ac laoreet. Phasellus ac aliquet enim, eu aliquet sem. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed pulvinar porta leo, eu ultricies nunc sollicitudin vitae. Curabitur pulvinar dolor lectus, quis porta turpis ullamcorper nec. Quisque eget varius turpis, quis iaculis nibh.<\/p><\/div><div class=\"Grid-cell\"><p>This cell should be top-aligned.<\/p><\/div><\/div><h3>Bottom-aligned Grid Cells<\/h3><h3>Vertically Centered Grid Cells<\/h3><div class=\"Grid Grid--gutters Grid--center\"><div class=\"Grid-cell\"><p>This cell should be vertically-centered with the cell to its right.<\/p><\/div><div class=\"Grid-cell\"><p>Curabitur pulvinar dolor lectus, quis porta turpis ullamcorper nec. Quisque eget varius turpis, quis iaculis nibh. Ut interdum ligula id metus hendrerit cursus. Integer eu leo felis. Aenean commodo ultrices nunc, sit amet blandit elit gravida in. Sed est ligula, ornare ac nisi adipiscing, iaculis facilisis tellus. Nullam vel facilisis libero. Duis semper lobortis elit, vitae dictum erat.<\/p><\/div><\/div><h3>Mixed Vertical Alignment<\/h3><div class=\"Grid Grid--gutters\"><div class=\"Grid-cell Grid-cell--top\"><p>This cell should be top aligned.<\/p><\/div><div class=\"Grid-cell\"><p>Curabitur pulvinar dolor lectus, quis porta turpis ullamcorper nec. Quisque eget varius turpis, quis iaculis nibh. Ut interdum ligula id metus hendrerit cursus. Integer eu leo felis. Aenean commodo ultrices nunc, sit amet blandit elit gravida in. Sed est ligula, ornare ac nisi adipiscing, iaculis facilisis tellus.<\/p><\/div><div class=\"Grid-cell Grid-cell--center\"><p>This cell should be center-aligned.<\/p><\/div><\/div><h2>The HTML<\/h2><pre class=\"language-html\">&lt;div class=\"Grid\"&gt;\n  &lt;div class=\"Grid-cell\"&gt;\u2026&lt;\/div&gt;\n  &lt;div class=\"Grid-cell\"&gt;\u2026&lt;\/div&gt;\n  &lt;div class=\"Grid-cell\"&gt;\u2026&lt;\/div&gt;\n&lt;\/div&gt;\n<\/pre><h2>The CSS<\/h2><h3>Basic Grid Styles<\/h3><pre class=\"language-css\">.Grid {\n  display: flex;\n}\n.Grid-cell {\n  flex: 1;\n}\n<\/pre><h3>Grid Style Modifiers<\/h3><pre class=\"language-css\">\/* With gutters *\/\n.Grid--gutters {\n  margin: -1em 0 0 -1em;\n}\n.Grid--gutters &gt; .Grid-cell {\n  padding: 1em 0 0 1em;\n}\n\/* Alignment per row *\/\n.Grid--top {\n  align-items: flex-start;\n}\n.Grid--bottom {\n  align-items: flex-end;\n}\n.Grid--center {\n  align-items: center;\n}\n\/* Alignment per cell *\/\n.Grid-cell--top {\n  align-self: flex-start;\n}\n.Grid-cell--bottom {\n  align-self: flex-end;\n}\n.Grid-cell--center {\n  align-self: center;\n}\n<\/pre><h3>Responsive Modifiers (a mobile-first approach)<\/h3><pre class=\"language-css\">\/* Base classes for all media *\/\n.Grid--fit &gt; .Grid-cell {\n  flex: 1;\n}\n.Grid--full &gt; .Grid-cell {\n  flex: 0 0 100%;\n}\n.Grid--1of2 &gt; .Grid-cell {\n  flex: 0 0 50%\n}\n.Grid--1of3 &gt; .Grid-cell {\n  flex: 0 0 33.3333%\n}\n.Grid--1of4 &gt; .Grid-cell {\n  flex: 0 0 25%\n}\n\/* Small to medium screens *\/\n@media (min-width: 24em) {\n  .small-Grid--fit &gt; .Grid-cell {\n    flex: 1;\n  }\n  .small-Grid--full &gt; .Grid-cell {\n    flex: 0 0 100%;\n  }\n  .small-Grid--1of2 &gt; .Grid-cell {\n    flex: 0 0 50%\n  }\n  .small-Grid--1of3 &gt; .Grid-cell {\n    flex: 0 0 33.3333%\n  }\n  .small-Grid--1of4 &gt; .Grid-cell {\n    flex: 0 0 25%\n  }\n}\n\/* Large screens *\/\n@media (min-width: 48em) {\n  .large-Grid--fit &gt; .Grid-cell {\n    flex: 1;\n  }\n  .large-Grid--full &gt; .Grid-cell {\n    flex: 0 0 100%;\n  }\n  .large-Grid--1of2 &gt; .Grid-cell {\n    flex: 0 0 50%\n  }\n  .large-Grid--1of3 &gt; .Grid-cell {\n    flex: 0 0 33.3333%\n  }\n  .large-Grid--1of4 &gt; .Grid-cell {\n    flex: 0 0 25%\n  }\n}\n<\/pre><p>View the full <a href=\"https:\/\/github.com\/philipwalton\/solved-by-flexbox\/blob\/master\/assets\/css\/components\/grid.css\">source<\/a> for the <code>Grid<\/code> component used in this demo on Github.<\/p>","created_at":"2018-11-07T16:45:23+0000","updated_at":"2018-11-07T16:45:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"philipwalton.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":357,"title":"Objective-See","url":"https:\/\/objective-see.com\/products\/lulu.html","content":"<hr class=\"gradient\" \/>In today's connected world, it is rare to find an application or piece of malware that<i>doesn't<\/i>talk to a remote server. Let's control this!<p>LuLu is the free, shared-source macOS firewall that aims to block unknown outgoing connections, unless explicitly approved by the user.<\/p><div>compatibility: OS X 10.12+<br \/>current version: 1.1.2 (<b><a href=\"https:\/\/objective-see.com\/products\/changelogs\/LuLu.txt\" class=\"inlineLink\">change log<\/a><\/b>)<br \/>zip's sha-1: E0A03769D32AF9F010FDBBF24B01E7D7CF2C6CA9<p>source code: <b><a href=\"https:\/\/github.com\/objective-see\/LuLu\" class=\"inlineLink\">LuLu<\/a><\/b><\/p><\/div><br \/><img src=\"https:\/\/objective-see.com\/images\/LL\/lulu.png\" width=\"820px;\" alt=\"image\" \/><p>Note, as with any security tool, direct or proactive attempts to specifically bypass LuLu's protections will likely succeed. By design, LuLu (currently) implements only limited 'self-defense' mechanisms. If you're interested in this topic, I'll be giving a talk, <a href=\"https:\/\/www.virusbulletin.com\/conference\/vb2018\/abstracts\/fire-ice-making-and-breaking-macos-firewalls\" class=\"inlineLink\">\"Fire &amp; Ice: Making and Breaking MacOS Firewalls\"<\/a> at VirusBulletin 2018!<\/p><br \/>LuLu is the free,<a href=\"https:\/\/github.com\/objective-see\/LuLu\" class=\"inlineLink\">shared-source<\/a>firewall for macOS. It's goal is simple; block any unknown outgoing connections, until approved by the user. While it was designed to generically detect malware by flagging unauthorized networking connections, LuLu can also be used to block OS components or 3rd-party applications from transmitting information to remote servers.<div>What's to like about LuLu? Lots!<img src=\"https:\/\/objective-see.com\/images\/LL\/heart.png\" height=\"32\" alt=\"image\" \/><\/div><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/free.png\" height=\"32\" alt=\"image\" \/>100% free<\/div><p>As in no ads, no time trials, no missing features. Because why not!?\n\t\t<br \/>And no, it doesn't track, monitor, or spy on you - as that'd just be pure evil!<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/code.png\" height=\"32\" alt=\"image\" \/>shared source<\/div><p>The full source code for LuLu is available on <a href=\"https:\/\/github.com\/objective-see\/LuLu\" class=\"inlineLink\">GitHub<\/a>. Such transparency allows anybody to audit its code, or understand exactly what is going on.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/firewall.png\" height=\"32\" alt=\"image\" \/>protects<\/div><p>LuLu aims to alert you whenever an unauthorized network connection is attempted. As such, it can generically detect malware, or be used to block legitimate applications that may be transmitting private data to remote servers.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/easy.png\" height=\"32\" alt=\"image\" \/>simple<\/div><p>\"Do one thing, do it well!\" LuLu is designed as simply as possible. Sure this means complex features may not be available, but it also means it's easier to use and has a smaller attack surface!<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/chat.png\" height=\"32\" alt=\"image\" \/>enterprise friendly<\/div><p>Want to know what network events are being detected? Or rules your users have added? LuLu provides simple mechanisms to subscribe to such events, and stores data such as rules in an open, easily digestible manner.<\/p><div>Want to support LuLu? ...you can via my<a class=\"inlineLink\" href=\"https:\/\/www.patreon.com\/objective_see\">patreon<\/a>page! Mahalo \u2661<p><a class=\"inlineLink\" href=\"https:\/\/www.patreon.com\/objective_see\">\n\t\t\t<img src=\"https:\/\/objective-see.com\/images\/blog\/blog_0x1D\/patreon.png\" width=\"700\" alt=\"image\" \/><\/a>\n\t\t<br \/><\/p><\/div><p>It's also important to understand LuLu's limitations! Some of these will be addressed as the software matures, while others are design decisions (mostly with the goal of keeping things simple).\n\t\t<br \/><\/p><ul><li>Network Monitoring<br \/>By design, LuLu only monitors for outgoing network connections. Apple's built in firewall does a great job blocking unauthorized incoming connections. \n\t\t<\/li>\n\t\t<li>Rules<br \/>Currently, LuLu only supports rules at the 'process level', meaning a process (or application) is either allowed to connect to the network or not. As is the case with other firewalls, this also means that if a legitimate (allowed) process is abused by malicious code to perform network actions, this will be allowed.\n\t\t<\/li>\n\t\t<li>Single User<br \/>For now, LuLu can only be installed for a single user. Future versions will likely allow it to be installed by multiple users on the same system.\n\t\t<\/li>\n\t\t<li>Self-Defense<br \/>Legitimate attackers\/security professionals know that any security tool can be trivially bypassed if specifically targeted - even if the tool employs advanced self-defense mechanisms. Such self-defense mechanisms are often complex to implement and in the end, almost always futile. As such, by design LuLu (currently) implements few self-defense mechanisms. For example, an attacker could enumerate all running processes to find the LuLu component responsible for displaying alerts and terminate it (via a sigkill).\n\t\t<\/li>\n\t\t<\/ul><div><img src=\"https:\/\/objective-see.com\/images\/LL\/install.png\" height=\"32\" alt=\"image\" \/><b>Installing LuLu<\/b><\/div><br \/>To install LuLu, first<a class=\"inlineLink\" href=\"https:\/\/bitbucket.org\/objective-see\/deploy\/downloads\/LuLu_1.1.2.zip\">download<\/a>the zip archive containing the application. Depending on your browser, you may need to manually unzip the application by double-clicking on the zipped archive.<p>Then, simply double-click on 'LuLu Installer.app'. Click 'Install' to install the tool: \n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/installApp.png\" width=\"500\" alt=\"image\" \/><br \/>During installation, LuLu will perform the following:\n\t\t<br \/><\/p><ul><li>install all LuLu components (kernel extension, launch daemon, login item, etc.)<\/li>\n\t\t\t<li>\n\t\t\tenumerate all (pre)installed applications:\n\t\t\t<p><img src=\"https:\/\/objective-see.com\/images\/LL\/enumApps.png\" width=\"500\" alt=\"image\" \/><br \/><\/p><\/li>\n\t\t\t<li>\n\t\t\trebuild the kernel cache:\n\t\t\t<p><img src=\"https:\/\/objective-see.com\/images\/LL\/kernelCache.png\" width=\"500\" alt=\"image\" \/><br \/><\/p><\/li>\n\t\t<\/ul><br \/>Note that these last two steps may take a few minutes, so please be patient!<br \/>Also in order to complete the install, you must reboot your Mac:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/restart.png\" width=\"500\" alt=\"image\" \/><\/p><p>Upon reboot, LuLu will display a 'welcome' window with various click-thru screens. For example, these allow one to configure the firewall:\n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/configure.png\" width=\"820px;\" alt=\"image\" \/><\/p><p>On versions of macOS High Sierra (10.13)+, in order to load the firewall's kernel extension, user assistance is required. \n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/loadKext.png\" width=\"820px;\" alt=\"image\" \/><br \/>Click on the 'Open System Preferences' button. This will load the System Preferences application, and then open the 'General' tab under 'Security and Privacy' pane. At the bottom, click the 'allow' button to allow the Objective-See LuLu kernel extension to load. (For more details on \"User-Approved Kernel Extension Loading\" see Apple's <a class=\"inlineLink\" href=\"https:\/\/developer.apple.com\/library\/content\/technotes\/tn2459\/_index.html\"> documentation<\/a>).\n\t\t\n\t\t<\/p><p>Once LuLu is installed, it will be running and is set to automatically start each time you log in. Unless configured to run without a status-bar icon, it will appear in the status bar:\n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/statusBarMenu.png\" alt=\"image\" \/><br \/><\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/uninstall.png\" height=\"32\" alt=\"image\" \/><b>Uninstalling LuLu<\/b><\/div>To uninstall LuLu, simply re-run the 'LuLu Installer.app'. Click 'Uninstall' to completer remove the tool:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/uninstallApp.png\" width=\"500\" alt=\"image\" \/><br \/>Note that this also requires a reboot to complete.\n\t\t<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/icon.png\" height=\"32\" alt=\"image\" \/><b>Using LuLu (Alerts)<\/b><\/div>Once LuLu is installed, it aims to alert you anytime an new or unauthorized process attempts to create an outgoing network connection. Here's a LuLul alert that's displayed, when the 'Russian' (APT28) malware<a href=\"https:\/\/download.bitdefender.com\/resources\/files\/News\/CaseStudies\/study\/143\/Bitdefender-Whitepaper-APT-Mac-A4-en-EN-web.pdf\" class=\"inlineLink\">'XAgent'<\/a>attempts to connect out to its command and control server for tasking:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/lulu.png\" width=\"820px;\" alt=\"image\" \/><br \/>The alert is designed to be fairly self-explanatory, but let's discuss some of its elements:\n\t\t<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/process.png\" height=\"32\" alt=\"image\" \/>process icon<\/div><p>The icon of the process is displayed in the top right of the alert window. If the process does not have an icon (i.e. its a command-line utility or a background daemon) a default system icon will be displayed.<\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/lock.png\" height=\"32\" alt=\"image\" \/>signing status<\/div><div>The 'signing status' of the process that is attempting to create a remote connection is also displayed in the LuLu alert window. The lock icon can be one of the following three images:<ul><li>\n\t\t\t<div><img src=\"https:\/\/objective-see.com\/images\/LL\/signedApple.png\" height=\"32\" alt=\"image\" \/>signed by Apple proper (i.e. core OS X\/macOS binary)<\/div>\n\t\t\t<\/li>\n\t\t\t<li>\n\t\t\t<div><img src=\"https:\/\/objective-see.com\/images\/LL\/signed3rdParty.png\" height=\"32\" alt=\"image\" \/>signed via a developer ID, or ad-hoc<\/div>\n\t\t\t<\/li>\n\t\t\t<li>\n\t\t\t<div><img src=\"https:\/\/objective-see.com\/images\/LL\/unsigned.png\" height=\"32\" alt=\"image\" \/>not signed (\"code object is not signed at all\")<\/div>\n\t\t\t<\/li>\n\t\t<\/ul>To view more information about the code-signing status of the process, click on the icon:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/signingInfo.png\" width=\"820px;\" alt=\"image\" \/><br \/><\/p><\/div><div><img src=\"https:\/\/objective-see.com\/images\/LL\/virusTotal.png\" width=\"32\" height=\"32\" alt=\"image\" \/>virus total information<\/div><div><a href=\"https:\/\/www.virustotal.com\/#\/home\/upload\" class=\"inlineLink\">VirusTotal<\/a>is cloud service that, given a file hash, will return the number of anti-virus engines that have flagged the file as malicious. Clicking the 'virus total' button in LuLu's alert window, will reveal a popover that contains this detection ratio for the process that is attempting to create a remote connection:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/virusTotalResults.png\" width=\"820px;\" alt=\"image\" \/><br \/>Click the 'details' link in the popup, to open the VirusTotal report in a browser.\n\t\t<br \/><\/p><\/div><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/hierarchy.png\" height=\"30\" alt=\"image\" \/>\u00a0process hierarchy<\/div><p>Click the 'process hierarchy' button in the LuLu alert to view the hierarchy for the process that is attempting to create a remote connection.\n\t\t<br \/><\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/processInfo.png\" height=\"32\" alt=\"image\" \/>process information (pid &amp; path)<\/div><p>The LuLu alert window also contains the process id (pid) and full path of the process that is attempting to create a remote connection. \n\t\t<br \/><\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/connect.png\" height=\"32\" alt=\"image\" \/>(attempted) connection information<\/div><p>The remote endpoint information, specifically the ip address, port &amp; and protocol that the process that is attempting connect to, are also displayed in the LuLu alert window.\n\t\t<br \/><\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/click.png\" height=\"32\" alt=\"image\" \/>block or allow<\/div><div><br \/>Clicking the 'block' button:<ul><li>prevents the process from establishing the outgoing connection<\/li>\n\t\t\t<li>creates a rule for the process, disallowing it from establishing any network connections<\/li>\n\t\t<\/ul><br \/>Clicking the 'allow' button:<br \/><ul><li>allows the process to establishing the outgoing connection<\/li>\n\t\t\t<li>creates a rule for the process, allowing it to establishing any network connections<\/li>\n\t\t<\/ul><\/div><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/rules.png\" height=\"32\" alt=\"image\" \/><b>Using LuLu (Rules)<\/b><\/div>Process are either allowed to access the network, or blocked, based on LuLu's rules. (Of course for those that LuLu doesn't have a rule for, a connection alert is displayed).<p>The 'rules' window displays these rules, as well as allows one to manually create or delete rules:\n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/rulesWindow.png\" width=\"820px;\" alt=\"image\" \/><br \/>This window can be access either by launching LuLu's application (\/Applications\/LuLu.app), or by clicking on 'Rules' in LuLu's status bar menu. \n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/showRules.png\" alt=\"image\" \/><br \/>There are five tabs in the rules window:\n\t\t<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/rulesAll.png\" height=\"32\" alt=\"image\" \/>All Rules<\/div><p>The first tab shows all of LuLu's rules. In other words, it is a combination of the default, apple, baseline, and user rules.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/rulesDefault.png\" height=\"32\" alt=\"image\" \/>Default Rules<\/div><p>The second tab shows LuLu's default or system rules. These rules (which cannot be deleted via the UI), are for Apple\/macOS processes that must be allowed communicate with the network in order to preserve system functionality.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/rulesApple.png\" height=\"32\" alt=\"image\" \/>Apple Rules<\/div><p>When the 'Allow Apple Programs' option has been selected (either in the welcome configuration screen, or LuLu's preferences), any process that is signed by Apple proper will be automatically allowed to connect to the network. Also, an 'allow' rule will be created, and will show up under here, under 'Apple Rules'.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/rulesBaseline.png\" height=\"32\" alt=\"image\" \/>Baseline Rules<\/div><p>When the 'Allow Installed Applications' option has been selected (either in the welcome configuration screen, or LuLu's preferences), any applications (and their components) that were (pre)installed will be automatically allowed to connect to the network. Also, an 'allow' rule will be created, and will show up under here, under 'Baseline Rules'.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/rulesUser.png\" height=\"32\" alt=\"image\" \/>User Rules<\/div><p>The fifth and final tab shows rules the user has created, either manually via the 'add rule' button, or by clicking 'block' or 'allow' in a LuLu connection alert window.<\/p><br \/>To manually add a rule, click on the 'add rule' button at the bottom of the rules window. This will bring up an 'Add Rule' dialogue box:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/addRule.png\" width=\"500px;\" alt=\"image\" \/><br \/>In this dialog box, enter the path to the target application or process (or click 'browse' to open a file chooser window). Then, select 'block' or 'allow', and finally click 'add' to add the rule. The new rule will be added as a 'user rule':\n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/newRule.png\" width=\"820px;\" alt=\"image\" \/><br \/>Note that if a rule already exists for the process or application, that 'add rule' will fail. In other word, the existing rule has to be deleted first.\n\t\t<\/p><p>To delete a rule, simply click the 'x' button on the right hand side of the rule, in the rules window. If the 'x' button is disabled, it means the rule cannot be deleted via the UI (i.e. default\/system rules).\n\t\t<\/p><p>Also, one can right or control click on a selected rule, and click on the 'delete' rule:\n\t\t<\/p><p><img src=\"https:\/\/objective-see.com\/images\/LL\/deleteRule.png\" width=\"820px;\" alt=\"image\" \/><\/p><p>LuLu's rules are stored in \/Library\/Objective-See\/LuLu\/rules.plist. If one has root privileges, by design, the rules can be directly read, and\/or modified:\n\t\t<\/p><div>$ cat \/Library\/Objective-See\/LuLu\/rules.plist<p>&lt;?xml version=\"1.0\" encoding=\"UTF-8\"?&gt;<br \/>&lt;!DOCTYPE plist PUBLIC \"-\/\/Apple\/\/DTD PLIST 1.0\/\/EN\" ...&gt;<br \/>&lt;plist version=\"1.0\"&gt;<br \/>&lt;dict&gt;<br \/>&lt;key&gt;\/Applications\/App Store.app&lt;\/key&gt;<br \/>&lt;dict&gt;<br \/>\u00a0\u00a0\u00a0&lt;key&gt;action&lt;\/key&gt;<br \/>\u00a0\u00a0\u00a0&lt;integer&gt;1&lt;\/integer&gt;<br \/>\u00a0\u00a0\u00a0&lt;key&gt;type&lt;\/key&gt;<br \/>\u00a0\u00a0\u00a0&lt;integer&gt;0&lt;\/integer&gt;<br \/>\u00a0\u00a0\u00a0&lt;key&gt;user&lt;\/key&gt;<br \/>\u00a0\u00a0\u00a0&lt;integer&gt;0&lt;\/integer&gt;<br \/>&lt;\/dict&gt;<br \/>...<br \/><\/p><\/div><p>Rules can also be imported or exported via the UI:\n\t\t<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/importRules.png\" height=\"32\" alt=\"image\" \/>Import Rules<\/div><p>To import a new set of rules, simply click the 'import' button at the bottom left of the Rules window. In the file selection panel, choose the file that contains the rules to import. Note that importing a rules is 'global' - it will fully replace all existing rules!<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/exportRules.png\" height=\"32\" alt=\"image\" \/>Export Rules<\/div><p>To export, or save, the existing rules, simply click the 'export' button at the bottom left of the Rules window. In the 'save' panel, choose the location where you'd like to save the rules.<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/preferences.png\" height=\"32\" alt=\"image\" \/><b>Using LuLu (Preferences)<\/b><\/div>LuLu can be configured via it's preferences pane. To open this pane, either in the main LuLu application (\/Applications\/LuLu.app), or via LuLu's status bar menu, click on 'Preferences'<p><img src=\"https:\/\/objective-see.com\/images\/LL\/showPrefs.png\" alt=\"image\" \/>\n\t\t\n\t\tThe preference pane has three tabs.\n\t\t<\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/prefsRules.png\" height=\"32\" alt=\"image\" \/>Rules<\/div>The 'rules' tab, allows one to configure LuLu how to (automatically) generate rules:<p><img src=\"https:\/\/objective-see.com\/images\/LL\/prefsRulesPane.png\" width=\"820px;\" alt=\"image\" \/><br \/><\/p><ul><li>'Allow Apple Programs'\n\t\t\t<br \/>When this option is selected any process that is signed by Apple proper will be automatically allowed to connect to the network. Also, an 'allow' rule will be created, and will show up in the Rules window, under 'Apple Rules'.\n\t\t\t<\/li>\n\t\t\t<li>'Allow Installed Applications'\n\t\t\t<br \/>When this option is selected any applications (and their components) that were (pre)installed will be automatically allowed to connect to the network. Also, an 'allow' rule will be created, and will show up in the Rules window, under 'Baseline Rules'.\n\t\t\t<\/li>\n\t\t<\/ul><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/prefsVisuals.png\" height=\"32\" alt=\"image\" \/>Visuals<\/div>The 'visual' tab, allows one to configure LuLu to run in passive mode (no alerts, new connections allowed), or in an 'icon-less' mode (no icon in the status bar).<p><img src=\"https:\/\/objective-see.com\/images\/LL\/prefsVisualPane.png\" width=\"820px;\" alt=\"image\" \/><\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/prefsUpdate.png\" height=\"32\" alt=\"image\" \/>Update<\/div>The 'update' tab, allows one to check for new versions, as well as disable the automatic check for new versions of LuLu.<p><img src=\"https:\/\/objective-see.com\/images\/LL\/prefsUpdatePane.png\" width=\"820px;\" alt=\"image\" \/><\/p><div><img src=\"https:\/\/objective-see.com\/images\/LL\/questions.png\" height=\"32\" alt=\"image\" \/>FAQs<\/div><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/zero.png\" height=\"32\" alt=\"image\" \/>Why is LuLu called LuLu?<\/div><p>In Hawaiian, the word 'LuLu' means protection, shield, or peace. As this tool aims to instill peace, by providing a protective shield, it seemed the fitting name.  And as LuLu, (along with all of Objective-See's tools) are coded with aloha on the lovely island of Maui, it's the perfect name!<\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/one.png\" height=\"32\" alt=\"image\" \/>Do I need LuLu if I've turned on the built-in macOS firewall?<\/div><p>Yes! Apple's built-in firewall only blocks incoming connections. LuLu is designed to detect and block outgoing connections, such as those generated by malware when the malware attempts to connect to it's command &amp; control server for tasking, or exfiltrates data.<\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/two.png\" height=\"32\" alt=\"image\" \/>Does LuLu conflict with other (paid) macOS firewalls or security products?<\/div><p>Although at this point testing has been limited, LuLu appears to play nice with other tools :)<\/p><br \/><div><img src=\"https:\/\/objective-see.com\/images\/LL\/three.png\" height=\"32\" alt=\"image\" \/>I found a bug (or issue) with LuLu. Can you fix it?<\/div><p>For sure! \n\t\t<br \/>If you encounter any, please shoot me an email at <a class=\"inlineLink\" href=\"mailto:bugs@objective-see.com\">bugs@objective-see.com<\/a>, or create an issue on <a class=\"inlineLink\" href=\"https:\/\/github.com\/objective-see\/LuLu\/issues\">GitHub<\/a>.<\/p><br \/>","created_at":"2018-11-07T16:45:23+0000","updated_at":"2018-11-07T16:45:23+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"objective-see.com","preview_picture":"https:\/\/objective-see.com\/images\/LL\/lulu.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":349,"title":"LWJGL Textures and Strings","url":"https:\/\/stackoverflow.com\/questions\/10801016\/lwjgl-textures-and-strings","content":"<div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        14&#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n&#13;\n<p>Is it possible to load PNG Textures and draw Strings in <code>LWJGL<\/code> WITHOUT using the Slick Framework? <\/p>\n<p>Everytime I google <em>\"how to load png images in lwjgl\"<\/em> I get answers like this -&gt; <em>\"hey just use the textureloader from the slick framework\"<\/em>.<br \/>Same for <em>\"how to draw strings in lwjgl\"<\/em> -&gt; <em>\"just use the TTFFont Class from the slick framework\"<\/em><\/p>\n<p>But I don't want to use this half-way-crossframework design. Because I don't think that this is the best way.<\/p>\n<p>Are there any Libraries or Extensions for <code>LWJGL<\/code> that are only made for Textures or Strings?<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/10801016\/revisions\" title=\"show all edits to this post\">edited Jan 21 '14 at 21:26<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/2390075\/benka\">benka<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            4,282113958&#13;\n        <\/div>&#13;<\/div><div class=\"user-action-time\">&#13;\n        asked May 29 '12 at 14:18&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1407823\/michael-malura\">Michael Malura<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            5771630&#13;\n        <\/div>&#13;<\/div><h2 data-answercount=\"2\">&#13;\n                                2 Answers&#13;\n                                2&#13;<\/h2><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        23&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n        accepted&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>Basically, you take a <code>BufferedImage<\/code>, use <a href=\"http:\/\/docs.oracle.com\/javase\/1.4.2\/docs\/api\/java\/awt\/image\/BufferedImage.html#getRGB%28int,%20int,%20int,%20int,%20int%5B%5D,%20int,%20int%29\" rel=\"nofollow noreferrer\">getRBG()<\/a> to get the RGB of each pixel, take that data and put it into a <a href=\"http:\/\/docs.oracle.com\/javase\/1.5.0\/docs\/api\/java\/nio\/ByteBuffer.html\" rel=\"nofollow noreferrer\">ByteBuffer<\/a> (the data type used to input image data to OpenGL), set some texture data, and create the <code>GL_TEXTURE_2D<\/code>.  <\/p>\n<p>This code by <a href=\"https:\/\/stackoverflow.com\/users\/3214889\">Krythic<\/a> does it:<\/p>\n<pre>import java.awt.image.BufferedImage;\nimport java.io.IOException;\nimport java.nio.ByteBuffer;\nimport javax.imageio.ImageIO;\nimport org.lwjgl.BufferUtils;\nimport org.lwjgl.opengl.GL12;\nimport static org.lwjgl.opengl.GL11.*;\npublic class TextureLoader {\n    private static final int BYTES_PER_PIXEL = 4;\/\/3 for RGB, 4 for RGBA\n       public static int loadTexture(BufferedImage image){\n          int[] pixels = new int[image.getWidth() * image.getHeight()];\n            image.getRGB(0, 0, image.getWidth(), image.getHeight(), pixels, 0, image.getWidth());\n            ByteBuffer buffer = BufferUtils.createByteBuffer(image.getWidth() * image.getHeight() * BYTES_PER_PIXEL); \/\/4 for RGBA, 3 for RGB\n            for(int y = 0; y &lt; image.getHeight(); y++){\n                for(int x = 0; x &lt; image.getWidth(); x++){\n                    int pixel = pixels[y * image.getWidth() + x];\n                    buffer.put((byte) ((pixel &gt;&gt; 16) &amp; 0xFF));     \/\/ Red component\n                    buffer.put((byte) ((pixel &gt;&gt; 8) &amp; 0xFF));      \/\/ Green component\n                    buffer.put((byte) (pixel &amp; 0xFF));               \/\/ Blue component\n                    buffer.put((byte) ((pixel &gt;&gt; 24) &amp; 0xFF));    \/\/ Alpha component. Only for RGBA\n                }\n            }\n            buffer.flip(); \/\/FOR THE LOVE OF GOD DO NOT FORGET THIS\n            \/\/ You now have a ByteBuffer filled with the color data of each pixel.\n            \/\/ Now just create a texture ID and bind it. Then you can load it using \n            \/\/ whatever OpenGL method you want, for example:\n          int textureID = glGenTextures(); \/\/Generate texture ID\n            glBindTexture(GL_TEXTURE_2D, textureID); \/\/Bind texture ID\n            \/\/Setup wrap mode\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL12.GL_CLAMP_TO_EDGE);\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL12.GL_CLAMP_TO_EDGE);\n            \/\/Setup texture scaling filtering\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);\n            \/\/Send texel data to OpenGL\n            glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, image.getWidth(), image.getHeight(), 0, GL_RGBA, GL_UNSIGNED_BYTE, buffer);\n            \/\/Return the texture ID so we can bind it later again\n          return textureID;\n       }\n       public static BufferedImage loadImage(String loc)\n       {\n            try {\n               return ImageIO.read(MainClass.class.getResource(loc));\n            } catch (IOException e) {\n                \/\/Error Handling Here\n            }\n           return null;\n       }\n}\n<\/pre>\n<p>To use this code, do something like this:<\/p>\n<pre>BufferedImage image = TextureLoader.loadImage(\"\/res\/test.png\");\/\/The path is inside the jar file\nint textureID = TextureLoader.loadTexture(image);\n<\/pre>\n<p>You can either save the textureID as a <code>final<\/code> variable(if the texture never changes), or unload the texture after each render using <code>GL11.glDeleteTextures(textureID);<\/code><\/p>\n<p>To do text, just create a <code>BufferedImage<\/code> manually, and use <a href=\"http:\/\/docs.oracle.com\/javase\/6\/docs\/api\/java\/awt\/image\/BufferedImage.html#createGraphics%28%29\" rel=\"nofollow noreferrer\">createGraphics()<\/a> to get a <a href=\"http:\/\/docs.oracle.com\/javase\/1.4.2\/docs\/api\/java\/awt\/Graphics2D.html\" rel=\"nofollow noreferrer\">graphics2D()<\/a> instance for the image.  Then, use <code>drawString()<\/code> to draw onto the <code>BufferedImage<\/code>, load it into the <code>TextureLoader<\/code>, render it onscreen, and unload the texture using the method above.<\/p><\/div><div class=\"user-action-time\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/posts\/10872080\/revisions\" title=\"show all edits to this post\">edited May 23 '17 at 11:46<\/a>&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/-1\/community\">Community<\/a>\u2666&#13;\n        &#13;<\/div><div class=\"user-action-time\">&#13;\n        answered Jun 3 '12 at 16:40&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/1203881\/flafla2\">Flafla2<\/a>&#13;\n        <div class=\"-flair\">&#13;\n            490512&#13;\n        <\/div>&#13;<\/div><div class=\"vote\"><h3>Vote count:&#13;\n        &#13;\n        0&#13;\n        &#13;\n&#13;\n&#13;\n&#13;\n&#13;\n&#13;\n<\/h3><\/div><div class=\"post-text\" itemprop=\"text\">&#13;\n<p>LWJGL now includes STB bindings, which is the preferred way of loading images and fonts, without having to use Slick or even AWT.<\/p>\n<p>To load a PNG :<\/p>\n<pre>import static org.lwjgl.opengl.GL11.GL_REPEAT;\nimport static org.lwjgl.opengl.GL11.GL_LINEAR;\nimport static org.lwjgl.opengl.GL11.GL_RGBA;\nimport static org.lwjgl.opengl.GL11.GL_TEXTURE_2D;\nimport static org.lwjgl.opengl.GL11.GL_TEXTURE_MAG_FILTER;\nimport static org.lwjgl.opengl.GL11.GL_TEXTURE_MIN_FILTER;\nimport static org.lwjgl.opengl.GL11.GL_TEXTURE_WRAP_S;\nimport static org.lwjgl.opengl.GL11.GL_TEXTURE_WRAP_T;\nimport static org.lwjgl.opengl.GL11.GL_UNPACK_ALIGNMENT;\nimport static org.lwjgl.opengl.GL11.GL_UNSIGNED_BYTE;\nimport static org.lwjgl.opengl.GL11.glBindTexture;\nimport static org.lwjgl.opengl.GL11.glGenTextures;\nimport static org.lwjgl.opengl.GL11.glPixelStorei;\nimport static org.lwjgl.opengl.GL11.glTexImage2D;\nimport static org.lwjgl.opengl.GL11.glTexParameteri;\nimport static org.lwjgl.opengl.GL30.glGenerateMipmap;\nimport static org.lwjgl.stb.STBImage.stbi_load_from_memory;\nimport static org.lwjgl.system.MemoryStack.stackPush;\nimport static org.lwjgl.demo.util.IOUtils.ioResourceToByteBuffer;\nimport java.nio.ByteBuffer;\nimport java.nio.IntBuffer;\nimport org.lwjgl.system.MemoryStack;\npublic class Texture{\n    private int width;\n    private int height;\n    private int id;\n    public Texture(String imagePath) {\n        ByteBuffer imageData = ioResourceToByteBuffer(imagePath, 1024);\n        try (MemoryStack stack = stackPush()) {\n            IntBuffer w = stack.mallocInt(1);\n            IntBuffer h = stack.mallocInt(1);\n            IntBuffer components = stack.mallocInt(1);\n            \/\/ Decode texture image into a byte buffer\n            ByteBuffer decodedImage = stbi_load_from_memory(imageData, w, h, components, 4);\n            this.width = w.get();\n            this.height = h.get();\n            \/\/ Create a new OpenGL texture \n            this.id = glGenTextures();\n            \/\/ Bind the texture\n            glBindTexture(GL_TEXTURE_2D, this.id);\n            \/\/ Tell OpenGL how to unpack the RGBA bytes. Each component is 1 byte size\n            glPixelStorei(GL_UNPACK_ALIGNMENT, 1);\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT);\n            glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT);\n            \/\/ Upload the texture data\n            glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA, this.width, this.height, 0, GL_RGBA, GL_UNSIGNED_BYTE, decodedImage);\n            \/\/ Generate Mip Map\n            glGenerateMipmap(GL_TEXTURE_2D);\n        }\n    }\n}\n<\/pre>\n<p>More complete examples for image loading, and text printing can be found in  LWJGL source code :<\/p>\n<p><a href=\"https:\/\/github.com\/LWJGL\/lwjgl3\/tree\/master\/modules\/core\/src\/test\/java\/org\/lwjgl\/demo\/stb\" rel=\"nofollow noreferrer\">org.lwjgl.demo.stb<\/a><\/p><\/div><div class=\"user-action-time\">&#13;\n        answered Oct 27 '17 at 11:15&#13;<\/div><div class=\"user-details\">&#13;\n        <a href=\"https:\/\/stackoverflow.com\/users\/3149379\/atom\">aTom<\/a>&#13;\n        &#13;<\/div>","created_at":"2018-11-07T16:45:22+0000","updated_at":"2018-11-07T16:45:22+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"stackoverflow.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/stackoverflow\/img\/apple-touch-icon@2.png?v=73d79a89bded","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":350,"title":"How To Run Your Own Mail Server","url":"https:\/\/www.c0ffee.net\/blog\/mail-server-guide\/","content":"<header>\n        <p class=\"subtitle\">A guide to self-hosting your email on FreeBSD using Postfix, Dovecot, Rspamd, and LDAP.\n      <\/p><\/header><p>Getting off GMail is one of the best ways to <a href=\"https:\/\/www.c0ffee.net\/blog\/mail-server-guide\/self-hosting-guide\">take back your data<\/a>\n      in the face of dragnet surveillance. Unfortunately, it's also one of the most difficult. If\n      you search for \"how to run your own mail server,\" many of the results will tell you it's\n      simply too difficult to reliably do it yourself.\n      <\/p><p>Luckily, running your own mail server is not as daunting as many would have you believe.\n      After all, that is how email is <em>designed to work<\/em>. Email is perhaps the most successful\n      federated, decentralized protocol to ever exist. It's a shame we've allowed a centralized,\n      monolithic advertising company to obtain a near monopoly on such a great technology. Luckily,\n      I've spent the last few years tweaking my mail server setup, and I'm willing to enable your\n      laziness in the spirit of a more free and open internet.\n      <\/p><p>Before we dive in, an important caveat: You will become a sysadmin of your own mail server.\n      <em>If you screw something up, you might not be able to send or receive email.<\/em> Sure, if\n      all else fails, you can admit defeat and point your MX records at a Google Apps account and\n      carry on. However, you run the risk of losing emails in the interim. I'm going to try to make\n      this guide as foolproof as possible, but if you're not comfortable assuming this responsibility,\n      it might be best to leave it to the experts.\n      <\/p><p>With that out of the way, here's what this guide will get you:\n      <\/p><ul><li>Opportunistic TLS for sending\/receiving email with <a href=\"http:\/\/www.postfix.org\">Postfix<\/a><\/li>\n        <li>Access to your email anywhere using IMAP with <a href=\"https:\/\/dovecot.org\">Dovecot<\/a><\/li>\n        <li>Fast, modern spam filtering with <a href=\"https:\/\/rspamd.com\">Rspamd<\/a><\/li>\n        <li>Full-text search of your mailbox with <a href=\"https:\/\/lucene.apache.org\/solr\/\">Apache Solr<\/a><\/li>\n        <li>DKIM, SPF, and DMARC records to ensure your outgoing mail is not spam-filtered<\/li>\n        <li>Statistical training of your spam filter based on what you move in\/out of the Junk folder<\/li>\n        <li>Custom email sorting\/filtering using <a href=\"http:\/\/sieve.info\">sieve<\/a><\/li>\n      <\/ul><p>After running this setup for year or so, I have found many aspects to be <em>better<\/em>\n      than GMail. I think the spam filtering works better, and being able to write a sieve script\n      to sort your inbox is a godsend compared to GMail's convoluted filter system. And as Google\n      occasionally seems to <a href=\"http:\/\/www.washingtontimes.com\/news\/2017\/aug\/1\/jordan-b-petersons-youtube-account-locked-during-b\/\">suspend the accounts of those with opinions they don't like<\/a>,\n      you'll have peace of mind with all your mail on a system <em>you<\/em> control.\n      <\/p><p>A final note: I run everything on <a href=\"https:\/\/www.freebsd.org\/\">FreeBSD<\/a>, so you\n      may have to massage some of the commands to work on your Linux flavor of choice, if that's\n      your thing. All the software used is open source, and should run equally well on any Unix-like\n      OS.\n      <\/p><hr \/><ol><li><a href=\"#overview\">Overview<\/a>\n        <\/li><li><a href=\"#dns\">DNS Records<\/a>\n        <\/li><li><a href=\"#postfix\">Postfix: The Mail Transfer Agent<\/a>\n        <\/li><li><a href=\"#dovecot\">Dovecot: The Mail Delivery Agent<\/a>\n        <\/li><li><a href=\"#solr\">Solr: Full-Text Search<\/a>\n        <\/li><li><a href=\"#rspamd\">Rspamd: Spam Filtering<\/a>\n        <\/li><li><a href=\"#dkim\">DKIM: Validation for Your Outgoing Mail<\/a>\n        <\/li><li><a href=\"#sieve\">Sieve: Scripting Your Mailbox<\/a>\n        <\/li><li><a href=\"#conclusion\">Conclusion<\/a>\n      <\/li><\/ol><hr \/><section><h2>Overview<\/h2><p>Now, take a deep breath, and feast your eyes on this convoluted Rube Goldberg machine of\n        an email system which you will soon bring lumbering to life:\n        <img alt=\"architecture diagram\" src=\"https:\/\/www.c0ffee.net\/img\/mail_server_diagram.svg\" \/><\/p><p>Let's walk through the architecture. All email (whether sent or received) first goes\n        through <a href=\"http:\/\/www.postfix.org\">Postfix<\/a> - the Mail Transfer Agent (MTA).\n        Postfix is responsible for transferring email to wherever it's supposed to go. If the\n        recipient domain's mailbox is on our own server, Postfix will accept the message for\n        delivery. Otherwise, Postfix will relay the message to the recipient's mail server.\n        <\/p><p>If you are <em>sending<\/em> an email to someone, you connect to Postfix over SMTP with\n        mandatory STARTTLS encryption on port 587. After successfully validating your username and\n        password, Postfix will accept your message to relay to the recipient's mail server over SMTP\n        using opportunistic TLS. (\"Opportunistic\" means that we will use encryption if the other\n        side supports it, and fall back to plaintext if they don't.)\n        <\/p><p>If you are <em>receiving<\/em> an email from someone, the sender's mail server will\n        connect to Postfix over SMTP with opportunistic TLS on port 25. After doing some basic\n        security and spam checks (like prohibiting unauthorized relaying and ensuring the other side\n        has valid DNS records in place), Postfix will accept the message for delivery.\n        <\/p><p>Whether a message is relayed (you sent an email to someone) or accepted for delivery\n        (someone sent an email to you), it is first filtered through the <a href=\"https:\/\/rspamd.com\">rspamd<\/a>\n        milter. (<em><a href=\"http:\/\/www.postfix.org\/MILTER_README.html\">milter<\/a><\/em> is just a\n        term for a Postfix mail filter.) For incoming mail, rspamd does some spam checking and may\n        flag the message as spam or reject it altogether. For outgoing mail, rspamd just takes care\n        of the <a href=\"http:\/\/www.dkim.org\">DKIM<\/a> signing (discussed in the <a href=\"#dkim\">DKIM section<\/a>).\n        <\/p><p>If an accepted email makes it through the rspamd milter, it is delivered to <a href=\"https:\/\/www.dovecot.org\">Dovecot<\/a>\n        - the Mail Delivery Agent (MDA) - using a local unix domain socket. Dovecot takes the emails\n        from Postfix and stores them in the user's mailbox. The mailbox can be stored on the\n        filesystem in a few different formats - with Maildir and mbox being commonly used.\n        Dovecot provides remote access to your mailbox using the POP3 and IMAP protocols. In\n        addition, it supports automatic mail filtering using custom <a href=\"http:\/\/sieve.info\">sieve<\/a>\n        scripts.\n        <\/p><p>When Dovecot receives a message, it first passes it to the <a href=\"http:\/\/lucene.apache.org\/solr\/\">Solr<\/a>\n        server, which indexes it for future full-text searches. Then, a sieve script will run which\n        automatically places mail with the X-SPAM header into the\n        user's Junk folder. In addition, when a user moves a message in or out of their Junk folder,\n        Dovecot will inform rspamd to re-train a message as spam or ham.  POP3 is ancient history,\n        so in our setup we'll only expose our emails over SSL-secured IMAP.\n        <\/p><p>I run a local <a href=\"https:\/\/www.ldap.com\/getting-started-with-ldap\">LDAP server<\/a>,\n        which allows me to use a single username and password for all my self-hosted apps. In this\n        guide, I'll describe how to store your mail credentials and email aliases in LDAP. However,\n        it's just as easy to store this information in a relational database or flat files.\n        <\/p><p>A quick note: Dovecot supports <a href=\"https:\/\/tools.ietf.org\/html\/rfc2177\">IMAP IDLE<\/a>,\n        which provides a pseudo-push functionality to get notified of new mail \"instantly.\" While\n        this works for most desktop mail clients, such as <a href=\"http:\/\/www.mutt.org\">mutt<\/a>\n        and OS X's Mail.app, it isn't supported by mobile clients (with Android's <a href=\"https:\/\/k9mail.github.io\">K-9 Mail<\/a>\n        being a notable exception). There is an open-source implementation of the Exchange\n        ActiveSync protocol called <a href=\"http:\/\/z-push.org\">Z-Push<\/a>, which integrates with\n        your IMAP server to provide push notifications to iOS and Android devices.\n        <\/p><p>I ran Z-Push for a few months, but there were enough small bugs (along with the hassle\n        of running PHP in production) to make me uninstall it. Setting my iPhone to check for new\n        mail every 15 minutes is plenty fast for me, and I still have about 40% battery left at the\n        end of the day. However, if you are dead set on having <em>instant<\/em> email notifications,\n        it might be something to look into. It works well enough, I just didn't find it worth the\n        trouble.\n        <\/p><p><em><strong>EDIT (20 Sep 2017): I've got native push notifications working on iOS.<\/strong> See\n        <a href=\"https:\/\/www.c0ffee.net\/blog\/mail-server-guide\/dovecot-push-notifications\">this blog post<\/a> for a step-by-step guide.<\/em>\n        <\/p><p>Finally, all the examples in this guide will use the example.com\n        domain and non-routable IP addresses. Obviously, you'll need to acquire your own domain\n        name and public IP address for your mail server, and modify my examples accordingly. All\n        commands in this guide should be run as root unless otherwise specified.\n      <\/p><\/section><section><h2 id=\"dns\">DNS Records<\/h2><p>First, you will need to set some DNS records so that other mail servers know where to send\n        your email. If you use a <a href=\"https:\/\/www.c0ffee.net\/blog\/mail-server-guide\/dns-hidden-master\">hidden master server to host your DNS<\/a>\n        like I do, you can just edit your zone file and reload\n        BIND.  Otherwise, you can set these records in your DNS provider's management portal.\n        <\/p><p>You probably want to set a low TTL (around 60 seconds) until you've verified everything\n        is working correctly. It's very annoying to realize you typed the wrong IP address and then\n        have to wait 3 hours for the TTL to expire in order to test again.\n        <\/p><p>The first thing you'll need is an MX record. MX records tell other mail servers which\n        hosts are running mail servers for a given domain. You can specify multiple MX records, each\n        with their own priority, to use as fallback mail servers when your primary one is down. For\n        my example.com domain, I only have a single mail server,\n        mail.example.com, so in my zone file I have a single MX\n        record. Of course, you will need to substitute your own IP address and domain name.\n<\/p><pre>\/usr\/local\/etc\/named\/master\/example.com.db$TTL    10800\n$ORIGIN example.com.\n@  1D  IN  SOA ns1.example.com.  root.example.com. (\n            2017080802 ; serial\n            1d ; refresh\n            3m ; retry\n            1w ; expire\n            3h ; minimum\n           )\n       IN  MX  10 mail.example.com.\n; host definitions\n@      IN  A      203.0.113.41\n@      IN  AAAA   2001:db8::2\nmail   IN  A      203.0.113.42\nmail   IN  AAAA   2001:db8::3\n<\/pre><p>It is <strong>critically important<\/strong> that you have <strong>reverse<\/strong> DNS\n        records for your mail server host. If your reverse DNS records do not point back to your\n        mail server, other mail servers may reject your email as spam. Normally, you can set reverse\n        DNS records in your hosting provider's web portal. Make sure to set reverse DNS for both the\n        IPv4 and IPv6 address of your mail server.\n        <\/p><p>Your mail server's reverse DNS and fully qualified domain name don't have to match\n        exactly, but your hostname in reverse DNS should probably resolve to the same IP as your\n        MX record. Your best bet is to have it point straight back to your mail server's hostname.\n        You can verify your reverse DNS record using the dig command:\n<\/p><pre>dig +short -x 203.0.113.42\n# mail.example.com.<\/pre><p>However, just having any reverse DNS record in place is enough for most mail\n        servers to accept your mail.\n        <\/p><p>You will also need to set an <a href=\"http:\/\/www.openspf.org\">SPF record<\/a> for your\n        mail server. While an SPF record is not technically required for your mail server to work,\n        having one in place significantly reduces the likelihood of your mail being rejected as\n        spam. An SPF record tells other mail servers which hosts are allowed to send mail on behalf\n        of your domain. In my case, I have a single mail server, and I want others to accept\n        mail from that server only. So my SPF record looks like this:\n<\/p><pre>\/usr\/local\/etc\/named\/master\/example.com.db@  IN  TXT  \"v=spf1 mx -all\"<\/pre><p>This indicates that only my MX records (mail.example.com)\n        are allowed to send mail for example.com, and mail coming\n        from anywhere else should be assumed to be spam and rejected. You can add your SPF definition as a\n        TXT record in your DNS provider's web portal (without the quotes).\n      <\/p><\/section><section><h2 id=\"postfix\">Postfix: The Mail Transfer Agent<\/h2><p>Now we're ready to install and configure some software. I install my packages from FreeBSD\n        ports, so substitute the appropriate commands for your package manager of choice. First,\n        bring everything up to date:\n<\/p><pre>portsnap fetch &amp;&amp; portsnap update\nportmaster -aBd<\/pre><p>Postfix will be responsible for sending and receiving mail for our domain over SMTP.\n        Install it from ports:\n<\/p><pre>cd \/usr\/ports\/mail\/postfix\nmake install clean<\/pre><p>For my setup, I use the following build options: LDAP PCRE PGSQL TLS.\n        <\/p><p>Once you've watched the magic text scroll for a few minutes, postfix should be installed.\n        Navigate to it's configuration directory and edit the config file. (On FreeBSD, software\n        packages from ports are always installed under the \/usr\/local\n        prefix. It should be under \/etc\/postfix if you're on Linux.)\n<\/p><pre>cd \/usr\/local\/etc\/postfix\nvim main.cf<\/pre><p>Below is what I have in my main.cf. I've given you some\n        brief commentary on what all the config options do, but you can always check\n        <a href=\"http:\/\/www.postfix.org\/postconf.5.html\">the man page<\/a> if you need more details\n        on something. This is a pretty standard <em>virtual users<\/em> setup. Rather than host mail\n        for actual unix accounts on the server, postfix will use an external database (LDAP, in my\n        case) to determine which email addresses it should accept mail for.\n        <\/p><p>When a message is accepted for delivery, postfix hands it off to Dovecot. Postfix will\n        also use Dovecot's auth mechanism to authenticate SMTP\n        users. We'll configure the LDAP authentication in the <a href=\"#dovecot\">Dovecot section<\/a>\n        and let postfix just re-use it.\n        <\/p><p>We'll set some standard spam-proofing and security options here as well, but we'll be\n        coming back to this file when we set up rspamd in the <a href=\"#rspamd\">section below<\/a>.\n<\/p><pre>\/usr\/local\/etc\/postfix\/main.cf# \"2\" is current for postfix 3.2 configs\ncompatibility_level = 2\n# disable \"new mail\" notifications for local unix users\nbiff = no\n# directory to store mail for local unix users\nmail_spool_directory = \/var\/mail\/local\n# Name of this mail server, used in the SMTP HELO for outgoing mail. Make\n# sure this resolves to the same IP as your reverse DNS hostname.\nmyhostname = mail.example.com\n# Domains for which postfix will deliver local mail. Does not apply to\n# virtual domains, which are configured below. Make sure to specify the FQDN\n# of your sever, as well as localhost.\n# Note: NEVER specify any virtual domains here!!! Those come later.\nmydestination = awesomebox.example.com, localhost.example.com, localhost\n# Domain appended to mail sent locally from this machine - such as mail sent\n# via the `sendmail` command.\nmyorigin = awesomebox.example.com\n# prevent spammers from searching for valid users\ndisable_vrfy_command = yes\n# require properly formatted email addresses - prevents a lot of spam\nstrict_rfc821_envelopes = yes\n# don't give any helpful info when a mailbox doesn't exist\nshow_user_unknown_table_name = no\n# limit maximum e-mail size to 50MB. mailbox size must be at least as big as\n# the message size for the mail to be accepted, but has no meaning after\n# that since we are using Dovecot for delivery.\nmessage_size_limit = 51200000\nmailbox_size_limit = 51200000\n# require addresses of the form \"user@domain.tld\"\nallow_percent_hack = no\nswap_bangpath = no\n# allow plus-aliasing: \"user+tag@domain.tld\" delivers to \"user\" mailbox\nrecipient_delimiter = +\n# path to the SSL certificate for the mail server\nsmtpd_tls_cert_file = \/usr\/local\/etc\/ssl\/certs\/mail.example.com.rsa.crt\nsmtpd_tls_key_file = \/usr\/local\/etc\/ssl\/certs\/mail.example.com.rsa.key\n# I have two certificates - one is RSA, the other uses the newer ECC. ECC is\n# faster and arguably more secure, but many mail servers don't yet support\n# it. I enable both types in postfix, but you most likely only have a single\n# RSA cert, and don't need to include these three lines.\nsmtpd_tls_eccert_file = \/usr\/local\/etc\/ssl\/certs\/mail.example.com.ecc.crt\nsmtpd_tls_eckey_file = \/usr\/local\/etc\/ssl\/certs\/mail.example.com.ecc.key\nsmtpd_tls_eecdh_grade = ultra\n# Path to your trusted certificates file. Usually provided by a\n# ca-certificates package or similar.\nsmtp_tls_CAfile=\/etc\/ssl\/cert.pem\n# These two lines define how postfix will connect to other mail servers.\n# DANE is a stronger form of opportunistic TLS. You can read about it here:\n# <a href=\"http:\/\/www.postfix.org\/TLS_README.html#client_tls_dane\">http:\/\/www.postfix.org\/TLS_README.html#client_tls_dane<\/a>\nsmtp_tls_security_level = dane\nsmtp_dns_support_level = dnssec\n# DANE requires a DNSSEC capable resolver. If your DNS resolver doesn't\n# support DNSSEC, remove the above two lines and uncomment the below:\n#   smtp_tls_security_level = may\n# IP address used by postfix to send outgoing mail. You only need this if\n# your machine has multiple IP addresses - set it to your MX address to\n# satisfy your SPF record.\nsmtp_bind_address = 203.0.113.42\nsmtp_bind_address6 = 2001:db8::3\n# Here we define the options for \"mandatory\" TLS. In our setup, TLS is only\n# \"mandatory\" for authenticating users. I got these settings from Mozilla's\n# SSL reccomentations page.\n#\n# NOTE: do not attempt to make TLS mandatory for all incoming\/outgoing\n# connections. Do not attempt to change the default cipherlist for non-\n# mandatory connections either. There are still a lot of mail servers out\n# there that do not use TLS, and many that do only support old ciphers.\n# Forcing TLS for everyone *will* cause you to lose mail.\nsmtpd_tls_mandatory_protocols = !SSLv2, !SSLv3, !TLSv1, !TLSv1.1, TLSv1.2\nsmtpd_tls_mandatory_ciphers = high\ntls_high_cipherlist = ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA256\n# allow other mail servers to connect using TLS, but don't require it\nsmtpd_tls_security_level = may\n# tickets and compression have known vulnerabilities\ntls_ssl_options = no_ticket, no_compression\n# it's more secure to generate your own DH params\nsmtpd_tls_dh512_param_file  = \/usr\/local\/etc\/ssl\/dh512.pem\nsmtpd_tls_dh1024_param_file = \/usr\/local\/etc\/ssl\/dh2048.pem\n# cache incoming and outgoing TLS sessions\nsmtpd_tls_session_cache_database = btree:${data_directory}\/smtpd_tlscache\nsmtp_tls_session_cache_database  = btree:${data_directory}\/smtp_tlscache\n# enable SMTPD auth. Dovecot will place an `auth` socket in postfix's\n# runtime directory that we will use for authentication.\nsmtpd_sasl_auth_enable = yes\nsmtpd_sasl_path = private\/auth\nsmtpd_sasl_type = dovecot\n# only allow authentication over TLS\nsmtpd_tls_auth_only = yes\n# don't allow plaintext auth methods on unencrypted connections\nsmtpd_sasl_security_options = noanonymous, noplaintext\n# but plaintext auth is fine when using TLS\nsmtpd_sasl_tls_security_options = noanonymous\n# add a message header when email was recieved over TLS\nsmtpd_tls_received_header = yes\n# require that connecting mail servers identify themselves - this greatly\n# reduces spam\nsmtpd_helo_required = yes\n# The following block specifies some security restrictions for incoming\n# mail. The gist of it is, authenticated users and connections from\n# localhost can do anything they want. Random people connecting over the\n# internet are treated with more suspicion: they must have a reverse DNS\n# entry and present a valid, FQDN HELO hostname. In addition, they can only\n# send mail to valid mailboxes on the server, and the sender's domain must\n# actually exist.\nsmtpd_client_restrictions =\n  permit_mynetworks,\n  permit_sasl_authenticated,\n  reject_unknown_reverse_client_hostname,\n  # you might want to consider:\n  #  reject_unknown_client_hostname,\n  # here. This will reject all incoming connections without a reverse DNS\n  # entry that resolves back to the client's IP address. This is a very\n  # restrictive check and may reject legitimate mail.\n  reject_unauth_pipelining\nsmtpd_helo_restrictions =\n  permit_mynetworks,\n  permit_sasl_authenticated,\n  reject_invalid_helo_hostname,\n  reject_non_fqdn_helo_hostname,\n  # you might want to consider:\n  #  reject_unknown_helo_hostname,\n  # here. This will reject all incoming mail without a HELO hostname that\n  # properly resolves in DNS. This is a somewhat restrictive check and may\n  # reject legitimate mail.\n  reject_unauth_pipelining\nsmtpd_sender_restrictions =\n  permit_mynetworks,\n  permit_sasl_authenticated,\n  reject_non_fqdn_sender,\n  reject_unknown_sender_domain,\n  reject_unauth_pipelining\nsmtpd_relay_restrictions =\n  permit_mynetworks,\n  permit_sasl_authenticated,\n  # !!! THIS SETTING PREVENTS YOU FROM BEING AN OPEN RELAY !!!\n  reject_unauth_destination\n  # !!!      DO NOT REMOVE IT UNDER ANY CIRCUMSTANCES      !!!\nsmtpd_recipient_restrictions =\n  permit_mynetworks,\n  permit_sasl_authenticated,\n  reject_non_fqdn_recipient,\n  reject_unknown_recipient_domain,\n  reject_unauth_pipelining,\n  reject_unverified_recipient\nsmtpd_data_restrictions =\n  permit_mynetworks,\n  permit_sasl_authenticated,\n  reject_multi_recipient_bounce,\n  reject_unauth_pipelining\n# deliver mail for virtual users to Dovecot's LMTP socket\nvirtual_transport = lmtp:unix:private\/dovecot-lmtp\n# LDAP query to find which domains we accept mail for\nvirtual_mailbox_domains = ldap:\/usr\/local\/etc\/postfix\/ldap-virtual-mailbox-domains.cf\n# LDAP query to find which email addresses we accept mail for\nvirtual_mailbox_maps = ldap:\/usr\/local\/etc\/postfix\/ldap-virtual-mailbox-maps.cf, hash:\/usr\/local\/etc\/postfix\/system-virtual-mailboxes\n# LDAP query to find a user's email aliases\nvirtual_alias_maps = ldap:\/usr\/local\/etc\/postfix\/ldap-virtual-alias-maps.cf\n# We'll uncomment these when we set up rspamd later:\n# milter_protocol = 6\n# milter_default_action = accept\n# smtpd_milters = unix:\/var\/run\/rspamd\/milter.sock\n# milter_mail_macros = i {mail_addr} {client_addr} {client_name} {auth_authen}\n<\/pre><p>We'll need to create some of the files and directories we specified in this file:\n<\/p><pre># create the local mail spool\nmkdir \/var\/mail\/local\nchown root:mail \/var\/mail\/local\nchmod 775 \/var\/mail\/local\n# create the dhparams\nopenssl dhparam -out \/usr\/local\/etc\/ssl\/dh512.pem 512\nopenssl dhparam -out \/usr\/local\/etc\/ssl\/dh2048.pem 2048\nchmod 644 \/usr\/local\/etc\/ssl\/dh{512,2048}.pem\n<\/pre><p>In master.cf, you can set service-specific overrides for the\n      options you set in main.cf. We'll use this file to make\n      encryption mandatory on the submission port (587). You can just change the lines below the\n      smtp and submission declarations\n      - no need to change anything else in this file.\n<\/p><pre>\/usr\/local\/etc\/postfix\/master.cf# disable authentication on port 25. This port should only be used by other\n# mail servers\nsmtp      inet  n       -       n       -       -       smtpd\n  -o smtpd_sasl_auth_enable=no\n# on the authenticated submission port, force TLS and use our more secure\n# cipher list\nsubmission inet n       -       n       -       -       smtpd\n  -o smtpd_tls_security_level=encrypt\n  -o tls_preempt_cipherlist=yes\n<\/pre><p>Now we need to tell postfix which users and domains we want to accept mail for.\n        I use LDAP, so the files below just tell postfix which queries to run to get the neccessary\n        information. You can check out how I setup OpenLDAP <a class=\"disabled\">here<\/a> (another blog post I\n        need to write). If you want to use flat files instead, you can use hash:\n        instead of ldap: in the last three lines of main.cf\n        shown above. You can then enter your users, domains, and aliases in plain text directly in\n        those files, and \"compile\" them with the postmap command. Or\n        you could put the information in a PostgreSQL database - there's plenty of online tutorials\n        for that as well. You can read more details on the <a href=\"http:\/\/www.postfix.org\/VIRTUAL_README.html\">postfix virtual readme<\/a>.\n        <\/p><p>Anyway, if you're using LDAP, read on. I haven't spent much time perfecting my LDAP\n        schema\u2014I got a personal \"single sign on\" system working with it, and left it at that. If\n        you have a more robust LDAP setup, you'll probably need to tweak these queries. The first\n        file specifies which domains this server will accept mail for:\n<\/p><pre>\/usr\/local\/etc\/postfix\/ldap-virtual-mailbox-domains.cfserver_host = ldapi:\/\/%2fvar%2frun%2fopenldap%2fldapi\/\nsearch_base = ou=domains,dc=example,dc=com\nversion = 3\nbind = no\nquery_filter = (&amp;(ObjectClass=dNSDomain)(dc=%s))\nresult_attribute = dc\n<\/pre><p>The next file specifies the mailboxes (email addresses) this server will accept mail for:\n<\/p><pre>\/usr\/local\/etc\/postfix\/ldap-virtual-mailbox-maps.cfserver_host = ldapi:\/\/%2fvar%2frun%2fopenldap%2fldapi\/\nsearch_base = ou=users,dc=example,dc=com\nversion = 3\nbind = no\nquery_filter = (&amp;(objectclass=inetLocalMailRecipient)(mail=%s))\nresult_attribute = mail\n<\/pre><p>The last file specifies the aliases for those email addresses:\n<\/p><pre>\/usr\/local\/etc\/postfix\/ldap-virtual-mailbox-alias-maps.cfserver_host = ldapi:\/\/%2fvar%2frun%2fopenldap%2fldapi\/\nsearch_base = ou=users,dc=example,dc=com\nversion = 3\nbind = no\nquery_filter = (&amp;(objectclass=inetLocalMailRecipient)(mailLocalAddress=%s))\nresult_attribute = mail\n<\/pre><p>The easiest way to set up your users and domains in LDAP is to set up an SSH tunnel to\n        your server and run <a href=\"http:\/\/directory.apache.org\/studio\/\">Apache Directory Studio<\/a>.\n        You can see how I have my domains set up in the query below:\n<\/p><pre><strong>$ ldapsearch -b dc=example.com,ou=domains,dc=example,dc=com<\/strong>\n# extended LDIF\n#\n# LDAPv3\n# base &lt;dc=example.com,ou=domains,dc=example,dc=com&gt; with scope subtree\n# filter: (objectclass=*)\n# requesting: ALL\n#\n# example.com, domains, example.com\ndn: dc=example.com,ou=domains,dc=example,dc=com\ndc: example.com\nobjectClass: dNSDomain\nobjectClass: top\n# search result\nsearch: 2\nresult: 0 Success\n# numResponses: 2\n# numEntries: 1<\/pre><p>The next query shows how I have users configured. The username is specified by the\n        uid parameter. mail contains\n        the user's primary mailbox, and each mailLocalAddress entry\n        specifies an email alias for the user.\n<\/p><pre><strong>$ ldapsearch -b uid=alphonsus,ou=users,dc=example,dc=com<\/strong>\n# extended LDIF\n#\n# LDAPv3\n# base &lt;uid=alphonsus,ou=users,dc=example,dc=com&gt; with scope subtree\n# filter: (objectclass=*)\n# requesting: ALL\n#\n# alphonsus, users, example.com\ndn: uid=alphonsus,ou=users,dc=example,dc=com\ncn: Alphonsus Liguori\ngivenName: Alphonsus\nmail: alphonsus@example.com\nmailLocalAddress: administrator@example.com\nmailLocalAddress: root@example.com\nmailLocalAddress: postmaster@example.com\no: The Church Triumphant\nobjectClass: person\nobjectClass: organizationalPerson\nobjectClass: inetOrgPerson\nobjectClass: inetLocalMailRecipient\nsn: Liguori\nuid: alphonsus\ndisplayName: Alphonsus Liguori\nuserPassword: e0HEfj4uiQyYSQxMCRTbHVTYXJshYf4aS4udX...\n# search result\nsearch: 2\nresult: 0 Success\n# numResponses: 2\n# numEntries: 1<\/pre><p>With those files in place, you can now start postfix! On FreeBSD, you will need to\n        disable the sendmail daemon included by default in base and\n        enable postfix to start on boot:\n<\/p><pre>\/etc\/rc.confsendmail_enable=\"NONE\"\npostfix_enable=\"YES\"<\/pre><p>Now, you can kill sendmail and start postfix:\n<\/p><pre>service sendmail onestop\nservice postfix start<\/pre><p>Check the output of \/var\/log\/maillog for any errors. Also,\n        be sure to enable TCP traffic to ports 25 and 587 in your firewall.\n        <\/p><p>You should now have almost everything you need you need for a working postfix\n        installation. We just need to get Dovecot running so postfix has somewhere to deliver the\n        mail. Luckily, Dovecot is a little less arcane to set up.\n      <\/p><\/section><section><h2 id=\"dovecot\">Dovecot: The Mail Delivery Agent<\/h2><p>Dovecot will provide IMAP access to your mailbox, so you can access your email remotely\n        from multiple devices. It can be easily installed from ports on FreeBSD:\n<\/p><pre>cd \/usr\/ports\/mail\/dovecot\nmake install clean<\/pre><p>I use the following build options on my server: KQUEUE LZ4 LDAP PGSQL ICU SOLR\n        <\/p><p>Be sure to install dovecot-piegeonhole as well in order to\n        get sieve filtering.\n<\/p><pre>cd \/usr\/ports\/mail\/dovecot-pigeonhole\nmake install clean<\/pre><p>I use the LDAP and MANAGESIEVE\n        options on my server. The managesieve protocol allows you to\n        edit your sieve scripts from your email client. Unfortunately, the only client I know of\n        that supports this functionality is <a href=\"https:\/\/www.kde.org\/applications\/internet\/kmail\/\">KMail<\/a>,\n        so I haven't messed with it yet.\n        <\/p><p>After Dovecot is installed, you will need to create a vmail\n        user to handle mail for the virtual users. You will also need to create the directories\n        to store the virtual mail. We will use a separate attachments directory, so Dovecot will\n        never waste disk space by storing the same attachment twice. Pretty neat!\n<\/p><pre>pw useradd -u 145 -n vmail -c \"Virtual Mail Handler\" -d \/var\/mail\/vhosts -s \/usr\/sbin\/nologin\nmkdir \/var\/mail\/{vhosts,attachments}\nchown vmail:vmail \/var\/mail\/{vhosts,attachments}\nchmod 770 \/var\/mail\/{vhosts,attachments}<\/pre><p>Then, you can navigate to Dovecot's configuration directory.\n<\/p><pre>cd \/usr\/local\/etc\/dovecot<\/pre><p>Here you should find a README and an example-config\n        directory. Feel free to read over them, or you can just dive into my configs. First,\n        create some additional directories:\n<\/p><pre>mkdir \/usr\/local\/etc\/dovecot\/{conf.d,sieve,sieve-before.d,sieve-after.d}\nchown root:dovecot \/usr\/local\/etc\/dovecot\/{sieve,sieve-before.d,sieve-after.d}<\/pre><p>Open dovecot.conf first. Here, we will set some generic\n        options and include the rest of the config files.\n<\/p><pre>\/usr\/local\/etc\/dovecot\/dovecot.conf# IMAP for remote access, LMTP for local delivery\nprotocols = imap lmtp\n# set these to the uid of your `vmail` user\nfirst_valid_uid = 145\nlast_valid_uid = 145\n!include conf.d\/*.conf\n!include_try local.conf<\/pre><p>The remaining config files live in the conf.d directory.\n        Each file from my configuration is reproduced in its entirety below, with some helpful\n        comments where appropriate.\n<\/p><pre>\/usr\/local\/etc\/dovecot\/conf.d\/10-auth.conf# cache all authentication results for one hour\nauth_cache_size = 10M\nauth_cache_ttl = 1 hour\nauth_cache_negative_ttl = 1 hour\n# only use plain username\/password auth - OK since everything is over TLS\nauth_mechanisms = plain\n# passdb specifies how users are authenticated - LDAP in my case\npassdb {\n  driver = ldap\n  args = \/usr\/local\/etc\/dovecot\/ldap.conf.ext\n}\n# userdb specifies the location of users' \"home\" directories - where their\n# mail is stored. e.g. \/var\/mail\/vhosts\/exmaple.com\/user\n# %d = domain, %n = user\nuserdb {\n  driver = static\n  args = uid=vmail gid=vmail home=\/var\/mail\/vhosts\/%d\/%n\n}\n<\/pre><p>The file specified by passdb provides the LDAP query\n      information necessary to authenticate usernames and passwords. Note that in my setup,\n      I authenticate with the <em>user<\/em> part only. Some other mail providers like to use\n      the entire email address as the username\u2014not so in this case.\n<\/p><pre>\/usr\/local\/etc\/dovecot\/ldap.conf.exthosts = localhost\nauth_bind = no\nldap_version = 3\nbase = ou=users,dc=example,dc=com\nderef = never\nscope = subtree\nuser_attrs =\nuser_filter = (&amp;(objectclass=inetOrgPerson)(uid=%n)\npass_attrs = mail=user,userPassword=password\npass_filter = (&amp;(objectclass=inetOrgPerson)(uid=%n))\ndefault_pass_scheme = CRYPT\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/10-mail.conf# default home directory location for all users\nmail_home = \/var\/mail\/vhosts\/%d\/%n\n# directory to store mail. The tilda makes it relative to the *dovecot*\n# virtual home directory.\n#\n# I use mdbox - this is Dovecot's own high-performance mail store format.\n# There are other slower, more \"traditional\" formats you can choose from.\n# Read about them here: <a href=\"https:\/\/wiki2.dovecot.org\/MailboxFormat\">https:\/\/wiki2.dovecot.org\/MailboxFormat<\/a>\nmail_location = mdbox:~\/mdbox\n# nothing fancy - just a standard default namespace with '\/' as the\n# hierarchy separator\nnamespace inbox {\n  separator = \/\n  inbox = yes\n}\n# set this to the group that owns your vmail directory.\nmail_privileged_group = vmail\n# these lines enable attachment deduplication. Attachments must be somewhat\n# large (64k) to store them separately from the mail store.\nmail_attachment_dir = \/var\/mail\/attachments\nmail_attachment_min_size = 64k\n# we'll uncomment this after we set up Solr in the following section:\n# mail_plugins = $mail_plugins fts fts_solr\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/10-master.conf# to improve performance, disable fsync globally - we will enable it for\n# some specific services later on\nmail_fsync = never\nservice imap-login {\n  # plain-text IMAP should only be accessible from localhost\n  inet_listener imap {\n    address = 127.0.0.1, ::1\n  }\n  # enable high-performance mode, described here:\n  # <a href=\"https:\/\/wiki.dovecot.org\/LoginProcess\">https:\/\/wiki.dovecot.org\/LoginProcess<\/a>\n  service_count = 0\n  # set to the number of CPU cores on your server\n  process_min_avail = 3\n  vsz_limit = 1G\n}\n# disable POP3 altogether\nservice pop3-login {\n  inet_listener pop3 {\n    port = 0\n  }\n  inet_listener pop3s {\n    port = 0\n  }\n}\n# enable semi-long-lived IMAP processes to improve performance\nservice imap {\n  service_count = 256\n  # set to the number of CPU cores on your server\n  process_min_avail = 3\n}\n# expose an LMTP socket for postfix to deliver mail\nservice lmtp {\n unix_listener \/var\/spool\/postfix\/private\/dovecot-lmtp {\n   group = postfix\n   mode = 0600\n   user = postfix\n  }\n}\n# expose an auth socket for postfix to authenticate users\nservice auth {\n  unix_listener \/var\/spool\/postfix\/private\/auth {\n    mode = 0660\n    user = postfix\n    group = postfix\n  }\n}\n# no need to run this as root\nservice auth-worker {\n  user = vmail\n}\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/10-ssl.conf# require SSL for all non-localhost connections\nssl = required\nssl_cert = &lt;\/usr\/local\/etc\/ssl\/certs\/mail.example.com.ecc.crt\nssl_key = &lt;\/usr\/local\/etc\/ssl\/certs\/mail.example.com.ecc.key\n# require modern crypto - taken from Mozilla's SSL recommendations page\nssl_dh_parameters_length = 2048\nssl_protocols = !SSLv2 !SSLv3 !TLSv1 !TLSv1.1 TLSv1.2\nssl_cipher_list = ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA256\nssl_prefer_server_ciphers = yes\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/15-lda.conf# configuration for mail delivered by the `dovecot-lda` command. Shouldn't\n# be needed since we are using LMTP, but kept for backwards compatibility.\nprotocol lda {\n  # use fsync for write-safety - this deals with delivering actual mail\n  mail_fsync = optimized\n  mail_plugins = $mail_plugins sieve\n}\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/15-mailboxes.conf# define any special IMAP folders here. You can force them to be created or\n# created+subscribed automatically used the `auto` option.\nnamespace inbox {\n  mailbox Drafts {\n    auto = subscribe\n    special_use = \\Drafts\n  }\n  mailbox Junk {\n    auto = create\n    special_use = \\Junk\n  }\n  mailbox Trash {\n    auto = create\n    special_use = \\Trash\n  }\n  mailbox Archive {\n    auto = subscribe\n    special_use = \\Archive\n  }\n  mailbox Sent {\n    auto = subscribe\n    special_use = \\Sent\n  }\n}\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/20-imap.conf# Use a longer IDLE interval to reduce network chatter and save battery\n# life. Max is 30 minutes.\nimap_idle_notify_interval = 29 mins\nprotocol imap {\n  # max IMAP connections per IP address\n  mail_max_userip_connections = 50\n  # imap_sieve will be used for spam training by rspamd\n  mail_plugins = $mail_plugins imap_sieve\n}\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/20-lmtp.confprotocol lmtp {\n  # use fsync for write-safety - this deals with delivering actual mail\n  mail_fsync = optimized\n  mail_plugins = $mail_plugins sieve\n}\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/20-managesieve.confprotocols = $protocols sieve\n# uncomment if you want remote managesieve functionality - unfortunately\n# almost no mail clients support it :(\n#\n# service managesieve-login {\n#   inet_listener sieve_deprecated {\n#     port = 0\n#   }\n# \n#   service_count = 0\n#   process_min_avail = 3\n#   vsz_limit = 1G\n# }\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/90-fts.conf# We'll uncomment these when we set up Solr in the next section:\n# plugin {\n#   fts_autoindex = yes\n#   fts = solr\n#   fts_solr = url=http:\/\/127.0.0.1:8983\/solr\/dovecot\/\n# }\n<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/90-sieve.confplugin {\n  # location of users' sieve directory and their \"active\" sieve script\n  sieve = file:~\/sieve;active=~\/.dovecot.sieve\n  # directory of global sieve scripts to run before and after processing ALL\n  # incoming mail\n  sieve_before = \/usr\/local\/etc\/dovecot\/sieve-before.d\n  sieve_after  = \/usr\/local\/etc\/dovecot\/sieve-after.d\n  # make sieve aware of user+tag@domain.tld aliases\n  recipient_delimiter = +\n  # maximum size of all user's sieve scripts\n  sieve_quota_max_storage = 50M\n}\n<\/pre><p>You should now have a fully functional IMAP server. Enable Dovecot to start on boot:\n<\/p><pre>\/etc\/rc.confdovecot_enable=\"YES\"<\/pre><p>Then you can start Dovecot:\n<\/p><pre>service dovecot start<\/pre><p>Make sure to check \/var\/log\/maillog for any errors.\n        Congratulations! You should now have a fully functional mail server! Make sure TCP traffic\n        to port 993 is enabled in your firewall.\n        <\/p><p>The remaining sections\u2014full-text search, spam filtering, etc\u2014are just extras. Now is a\n        good time to stop and make sure everything works before you go any further.\n        <\/p><section><h3 id=\"mail-test\">Testing Your Mail Server<\/h3><p>First, get a separate terminal open and tail your mail log while you are testing\n          everything:\n<\/p><pre>tail -f \/var\/log\/maillog<\/pre><p>Start by checking your MX records in DNS using the dig\n          command. If your DNS is configured correctly, you should see results like the following:\n<\/p><pre>dig example.com MX +short\n# 10 mail.example.com.\ndig mail.example.com +short\n# 203.0.113.42\n<\/pre><p>Now we can test your ability to send and receive mail. Don't worry if you can't connect\n          to your server over port 25\u2014almost all residential ISPs block all traffic on this port to\n          deter spammers. From your local machine, you can use an email client to connect to your\n          mail server. Your connectivity details will be the following:\n          <\/p><ul><li>Incoming Mail: IMAP<\/li>\n            <li><strong>username:<\/strong> ldap username<\/li>\n              <li><strong>password:<\/strong> ldap password<\/li>\n              <li><strong>server:<\/strong> mail.example.com<\/li>\n              <li><strong>port:<\/strong> 993<\/li>\n              <li><strong>SSL:<\/strong> yes<\/li>\n            <\/ul>Outgoing Mail: SMTP\n            <ul><li><strong>username:<\/strong> ldap username<\/li>\n              <li><strong>password:<\/strong> ldap password<\/li>\n              <li><strong>email address:<\/strong> username@example.com<\/li>\n              <li><strong>server:<\/strong> mail.example.com<\/li>\n              <li><strong>port:<\/strong> 587<\/li>\n              <li><strong>SSL:<\/strong> STARTTLS<\/li>\n            <\/ul><p>If you can log in, you should see an empty mailbox with default folders (Inbox, Drafts,\n          Junk, etc). If not, you'll need to check your mail log and troubleshoot Dovecot.\n          <\/p><p>Next, you'll need to test outgoing mail. Hopefully you have a spare GMail account or\n          another working email address you can test with. When you send an email through your mail\n          client, you should see something like this in your mail log:\n<\/p><pre>postfix\/smtpd[39231]: connect from SOME-IP-ADDRESS\npostfix\/smtpd[39231]: AA752DDA7D8: client=SOME-IP_ADDRESS, sasl_method=PLAIN, sasl_username=user@example.com\npostfix\/cleanup[39248]: AA752DDA7D8: message-id=&lt;EE3CDE5B-9A46-405D-A74A-E409769F791D@example.com&gt;\npostfix\/qmgr[39210]: AA752DDA7D8: from=&lt;user@example.com&gt;, size=762, nrcpt=1 (queue active)\npostfix\/smtp[39251]: AA752DDA7D8: to=&lt;YOUR-OTHER-EMAIL@gmail.com&gt;, relay=gmail-smtp-in.l.google.com[74.125.130.27]:25, delay=8.9, delays=4.9\/0.03\/3\/1, dsn=2.0.0, status=sent (250 2.0.0 OK 1502808287 y15si6196590pli.453 - gsmtp)\npostfix\/qmgr[39210]: AA752DDA7D8: removed\npostfix\/smtpd[39231]: disconnect from SOME-IP-ADDRESS ehlo=2 starttls=1 auth=1 mail=1 rcpt=1 data=1 quit=1 commands=8<\/pre><p>If you don't see any errors in the log, you should find the email you just sent in your\n          other email account's inbox. If it's not there, be sure to check your spam folder. (If it\n          got flagged as spam, the testing steps in the <a href=\"#dkim\">DKIM section<\/a> should help\n          determine why.)\n          <\/p><p>Now, you can test incoming mail. In your other email account, reply to the message you\n          just sent yourself. This should be the first email ever sent to your new, personal email\n          server. Fingers crossed! If the other side successfully finds your mail server and\n          delivers the message, you should see something like this in your mail log:\n<\/p><pre>postfix\/smtpd[39448]: connect from mail-yw0-x22b.google.com[2607:f8b0:4002:c05::22b]\npostfix\/verify[39459]: cache btree:\/var\/db\/postfix\/verify_cache full cleanup: retained=9 dropped=0 entries\npostfix\/smtpd[39448]: 98DA9DDA7D8: client=mail-yw0-x22b.google.com[2607:f8b0:4002:c05::22b]\npostfix\/cleanup[39460]: 98DA9DDA7D8: message-id=&lt;CALM7EhMmCkzpex5pKu0TXh0z3+iYeLCGpw5VO9Sv_RanqgohYQ@mail.gmail.com&gt;\npostfix\/qmgr[39210]: 98DA9DDA7D8: from=&lt;YOUR-OTHER-EMAIL@gmail.com&gt;, size=3449, nrcpt=1 (queue active)\ndovecot: lmtp(39462): Connect from local\ndovecot: lmtp(user@example.com): MMb\/MfoPk1kmmgAA\/gk8FQ: sieve: msgid=&lt;CALM7EhMmCkzpex5pKu0TXh0z3+iYeLCGpw5VO9Sv_RanqgohYQ@mail.gmail.com&gt;: stored mail into mailbox 'INBOX'\npostfix\/lmtp[39461]: 98DA9DDA7D8: to=&lt;user@example.com&gt;, relay=mail.example.com[private\/dovecot-lmtp], delay=0.26, delays=0.21\/0.01\/0.01\/0.03, dsn=2.0.0, status=sent (250 2.0.0 &lt;user@example.com&gt; MMb\/MfoPk1kmmgAA\/gk8FQ Saved)\ndovecot: lmtp(39462): Disconnect from local: Successful quit\npostfix\/qmgr[39210]: 98DA9DDA7D8: removed\npostfix\/smtpd[39448]: disconnect from mail-yw0-x22b.google.com[2607:f8b0:4002:c05::22b] ehlo=2 starttls=1 mail=1 rcpt=1 data=1 quit=1 commands=7\npostfix\/smtpd[39471]: connect from unknown[181.49.241.194]\npostfix\/smtpd[39471]: disconnect from unknown[181.49.241.194] helo=1 auth=0\/1 quit=1 commands=2\/3<\/pre><p>If there were no errors, you should see the very first message in your new IMAP inbox!\n          If there was a failure logged, you'll have to investigate postfix. Most likely there was\n          problem finding your virtual mailbox, or postfix had some issue handing the mail off to\n          Dovecot.\n          <\/p><p>If you didn't receive the email, and you don't see anything logged <em>at all<\/em>,\n          then you probably have a firewall issue or some problem with your MX records in DNS. Also,\n          make sure your hosting provider allows traffic on port 25\u2013some of them are paranoid.\n          <\/p><p>Once you're able to send and receive emails from your new mail server, you're ready to\n          get the rest of the goodies like full-text searching and spam filtering. Read on.\n        <\/p><\/section><\/section><section><h2 id=\"solr\">Solr: Full-Text Search<\/h2><p>Solr is a standalone full-text search platform from Apache the Apache Foundation. It's\n        heavyweight and based on Java, but currently it's your best bet if you want open source\n        full-text search with Dovecot. It's easily installed from ports on FreeBSD:\n<\/p><pre>cd \/usr\/ports\/textproc\/apache-solr\nmake install clean<\/pre><p>This will pull in Java as a dependency, so you may want to grab a coffee or something\n        while the JVM compiles. The defaults in \/usr\/local\/etc\/solr.in.sh\n        are fine. Enable solr to start on boot:\n<\/p><pre>\/etc\/rc.confsolr_enable=\"YES\"<\/pre><p>Finally, start the solr server:\n<\/p><pre>service solr start<\/pre><p>Now we need to create a collection for Dovecot to use.\n<\/p><pre>su -m solr -c \"\/usr\/local\/solr\/bin\/solr create -c dovecot -n dovecot\"<\/pre><p>This will create a Solr collection for Dovecot in \/var\/db\/solr\/dovecot.\n        Navigate to the configuration directory\u2014we'll need to include some magic incantations I\n        found on <a href=\"http:\/\/mor-pah.net\/2016\/08\/15\/dovecot-2-2-with-solr-6-or-5\/\">this guy's blog<\/a>\n        to get Solr working with Dovecot.\n<\/p><pre>cd \/var\/db\/solr\/dovecot\/conf<\/pre><p>Delete the managed-schema file:\n<\/p><pre>rm \/var\/db\/solr\/dovecot\/conf\/managed-schema<\/pre><p>Create a file called schema.xml with the contents <a href=\"https:\/\/www.c0ffee.net\/files\/schema.xml\">here<\/a>:\n<\/p><pre>curl -o \/var\/db\/solr\/dovecot\/conf\/schema.xml https:\/\/www.c0ffee.net\/files\/schema.xml<\/pre><p>Then, replace the entire contents of solrconfig.xml\n        with the version <a href=\"https:\/\/www.c0ffee.net\/files\/solrconfig.xml\">here<\/a>:\n<\/p><pre>curl -o \/var\/db\/solr\/dovecot\/conf\/solrconfig.xml https:\/\/www.c0ffee.net\/files\/solrconfig.xml<\/pre><p>Finally, make sure the solr user still owns those files:\n<\/p><pre>chown solr:solr \/var\/db\/solr\/dovecot\/conf\/{schema,solrconfig}.xml<\/pre><p>Solr should now be ready to index email for Dovecot. Restart the server for your changes\n        to take effect.\n<\/p><pre>service solr restart<\/pre><p>Now you just need to configure Dovecot to use the Solr server. All you have to do is\n        uncomment some lines in the files from <a href=\"#dovecot\">the previous section<\/a>.\n<\/p><pre>\/usr\/local\/etc\/dovecot\/conf.d\/10-mail.confmail_plugins = $mail_plugins fts fts_solr<\/pre><pre>\/usr\/local\/etc\/dovecot\/conf.d\/90-fts.confplugin {\n  fts_autoindex = yes\n  fts = solr\n  fts_solr = url=http:\/\/127.0.0.1:8983\/solr\/dovecot\/\n}<\/pre><p>Reload Dovecot for your changes to take effect:\n<\/p><pre>service dovecot reload<\/pre><p>Now, whenever you receive an email or move a message from one IMAP folder to another,\n        you should see something like this in your mail log:\n<\/p><pre>dovecot: indexer-worker(user@example.com): Indexed 1 messages in INBOX<\/pre><p>Indexing only occurs when a new message arrives in a folder. To index the entire inbox\n        of all your mail users, run the following command:\n<\/p><pre>doveadm index -A inbox<\/pre><p>Finally, according to <a href=\"https:\/\/wiki.dovecot.org\/Plugins\/FTS\/Solr\">the Dovecot documentation<\/a>,\n        the Solr database needs to be issued a commit and optimize\n        command every so often. This is easily accomplished by adding the following cron jobs (make\n        sure you get the path to curl correct if you're using Linux):\n<\/p><pre>\/etc\/crontab15 2 * * *  solr \/usr\/local\/bin\/curl -s 'http:\/\/localhost:8983\/solr\/dovecot\/update?optimize=true' &gt; \/dev\/null\n*\/5 * * * * solr \/usr\/local\/bin\/curl -s 'http:\/\/localhost:8983\/solr\/dovecot\/update?commit=true' &gt; \/dev\/null<\/pre><p>You should now have full-text search of your mailbox over IMAP. You can test it out by\n        searching for some text on your mobile IMAP client (often desktop mail clients will perform\n        searches using local cached data). Or, you can test it yourself from your server using\n        telnet. Type the following lines (including the preceding\n        numbers) into the prompt:\n<\/p><pre>telnet localhost imap\n* OK [CAPABILITY IMAP4rev1 LITERAL+ SASL-IR...] I am ready.\n1 login your_username your_password\n2 select Inbox\n3 SEARCH text \"test\"<\/pre><p>If you encounter any issues, Solr provides a web interface on localhost port 8983. You\n        can set up an SSH tunnel to your server and forward the port to your local machine to check\n        its status. You can also view the logs in \/var\/log\/solr.\n      <\/p><\/section><section><h2 id=\"rspamd\">Rspamd: Spam Filtering<\/h2><p><a href=\"https:\/\/rspamd.com\">Rspamd<\/a> is a fast, open source spam filtering system\n        which utilizes multiple techniques to prevent spam from reaching your mailbox. I was a happy\n        user of <a href=\"https:\/\/sourceforge.net\/projects\/dspam\/\">dspam<\/a> for years, but it has\n        been <a href=\"https:\/\/sourceforge.net\/p\/dspam\/mailman\/message\/32585111\/\">offically abandoned since 2014<\/a>.\n        <\/p><p>Rspamd is actively developed and appears to be a much more modern project.\n        Written in C, it is quite fast and integrates directly into postfix as a milter. Like dspam,\n        it uses bayesian filtering to learn what users consider to be spam and ham. However, it also\n        incorporates many other anti-spam measures: <a href=\"https:\/\/rspamd.com\/doc\/modules\/rbl.html\">RBL checks<\/a>,\n        <a href=\"https:\/\/rspamd.com\/doc\/modules\/spf.html\">SPF<\/a>\/<a href=\"https:\/\/rspamd.com\/doc\/modules\/dkim.html\">DKIM<\/a>\/<a href=\"https:\/\/dmarc.org\">DMARC<\/a>\n        validation, <a href=\"https:\/\/www.rhyolite.com\/dcc\/\">DCC bulk mail checks<\/a>, and <a href=\"https:\/\/rspamd.com\/doc\/modules\/greylisting.html\">greylisting<\/a>,\n        to name a few. In addition, it has some other useful features like <a href=\"https:\/\/rspamd.com\/doc\/modules\/dkim_signing.html\">DKIM signing of outgoing mail<\/a>,\n        <a href=\"https:\/\/rspamd.com\/doc\/modules\/replies.html\">automatic whitelisting when you reply to someone<\/a>,\n        and a <a href=\"https:\/\/www.rspamd.com\/webui\/\">web interface<\/a> where you can see spam\n        checking results in real time.\n        <\/p><p>Sold yet? Let's get rspamd installed. It's available in FreeBSD ports:\n<\/p><pre>cd \/usr\/ports\/mail\/rspamd\nmake install clean<\/pre><p>I use the following build options: GD HYPERSCAN. The\n        <a href=\"https:\/\/01.org\/hyperscan\">hyperscan<\/a> option is especially cool\u2014it enables high\n        performance regex matching.\n        <\/p><p>Enable rspamd to start at boot:\n<\/p><pre>\/etc\/rc.confrspamd_enable=\"YES\"<\/pre><p>Finally, start rspamd:\n<\/p><pre>service rspamd start<\/pre><p>Many rspamd modules require a <a href=\"https:\/\/redis.io\">redis<\/a>\n        instance to work, so install that as well. The default build options are fine.\n<\/p><pre>cd \/usr\/ports\/databases\/redis\nmake install clean<\/pre><p>You'll want to set some redis options recommended by the\n        <a href=\"https:\/\/rspamd.com\/doc\/quickstart.html#caching-setup\">official rspamd tutorial<\/a>:\n<\/p><pre>\/usr\/local\/etc\/redis.conf# listen only on localhost\nbind 127.0.0.1 ::1\n# limit the max amount of memory used - appropriate value will depend on\n# your email volume\nmaxmemory 512mb\nmaxmemory-policy volatile-lru<\/pre><p>Enable redis to start at boot:\n<\/p><pre>\/etc\/rc.confredis_enable=\"YES\"<\/pre><p>Finally, start the redis daemon:\n<\/p><pre>service redis start<\/pre><p>One of the rspamd modules we will use depends on the <a href=\"https:\/\/www.dcc-servers.net\/dcc\/\">DCC daemon<\/a>\n        to check mails for bulkiness. You can install it from ports as well. The default build\n        options are fine.\n<\/p><pre>cd \/usr\/ports\/mail\/dcc-dccd\nmake install clean<\/pre><p>Uncomment the following options in \/usr\/local\/dcc\/dcc_conf.\n        We'll be using the DCC service to detect spam, but won't be actively reporting anything.\n<\/p><pre>\/usr\/local\/dcc\/dcc_confDCCM_LOG_AT=NEVER\nDCCM_REJECT_AT=MANY\nDCCIFD_ENABLE=on<\/pre><p>DCC is rather arcane and doesn't seem to log to syslog like modern software. I haven't\n        figured out a nice way of rotating its logs with newsyslogd\n        yet, so for now I just have a cronjob to delete them every day:\n        <\/p><pre>\/etc\/crontab0 2 * * * \/usr\/bin\/find \/usr\/local\/dcc\/log\/ -not -newermt '1 days ago' -delete<\/pre><p>Enable DCC to start at boot:\n<\/p><pre>\/etc\/rc.confdccifd_enable=\"YES\"<\/pre><p>Finally, start the DCC daemon:\n<\/p><pre>service dccifd start<\/pre><p>Now, your're ready to configure rspamd itself. Rspamd's default configuration files live\n        in \/usr\/local\/etc\/rspamd. Don't edit these files directly,\n        as they will be overwritten when the rspamd package is updated. Instead, create a local\n        configuration directory. All your custom configuration will go here:\n<\/p><pre>mkdir \/usr\/local\/etc\/rspamd\/local.d<\/pre><p>On the rspamd website, you'll find the <a href=\"https:\/\/www.rspamd.com\/doc\/quickstart.html\">rspamd quick start guide<\/a>,\n          along with the <a href=\"https:\/\/www.rspamd.com\/doc\/modules\/\">documentation for all the modules<\/a>.\n          Feel free design your own setup, or you can just copy what I've done in the files below.\n          I've enabled most of the useful modules and provided some commentary where appropriate.\n<\/p><pre>\/usr\/local\/etc\/rspamd\/local.d\/worker-controller.inc# generate a password hash using the `rspamadm pw` command and put it here\n# this one is the hash for 'hunter2'\npassword = \"$2$or9n9ffj4qsogh7i8d9qi5u1hxt53q6o$ntp4kj...\";\n# dovecot will use this socket to communicate with rspamd\n# note to self - tighten up these permissions\nbind_socket = \"\/var\/run\/rspamd\/rspamd.sock mode=0666 owner=nobody\";\n# you can comment this out if you don't need the web interface\nbind_socket = \"127.0.0.1:11334\";<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/worker-normal.inc# we're not running rspamd in a distributed setup, so this can be disabled\n# the proxy worker will handle all the spam filtering\nenabled = false;\n<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/worker-proxy.inc# this worker will be used as postfix milter\nmilter = yes;\n# note to self - tighten up these permissions\nbind_socket = \"\/var\/run\/rspamd\/milter.sock mode=0666 owner=nobody\";\n# the following specifies self-scan mode, for when rspamd is on the same\n# machine as postfix\ntimeout = 120s;\nupstream \"local\" {\n  default = yes;\n  self_scan = yes;\n}<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/redis.conf# just specifying a server enables redis for all modules that can use it\nservers = \"127.0.0.1\";<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/classifier-bayes.confautolearn = true;\nbackend = \"redis\";<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/dcc.conf# path to dcc socket\nhost = \"\/usr\/local\/dcc\/dccifd\";\ntimeout = 5.0;<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/dkim_signing.conf# enable dkim signing - we will set this up in the DKIM section later\npath = \"\/var\/db\/rspamd\/dkim\/$domain.$selector.key\";\nselector = \"dkim\";<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/mx_check.conf# checks if sender's domain has at least one connectable MX record\nenabled = true;<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/phishing.conf# check messages against some anti-phishing databases\nopenphish_enabled = true;\nphishtank_enabled = true;<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/replies.conf# whitelist messages from threads that have been replied to\naction = \"no action\";<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/surbl.conf# follow redirects when checking URLs in emails for spaminess\nredirector_hosts_map = \"\/usr\/local\/etc\/rspamd\/redirectors.inc\";<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/url_reputation.conf# check URLs within messages for spaminess\nenabled = true;<\/pre><pre>\/usr\/local\/etc\/rspamd\/local.d\/url_tags.conf# cache some URL tags in redis\nenabled = true;<\/pre><p>Restart rspamd for all these changes to take effect:\n<\/p><pre>service rspamd restart<\/pre><p>Now, you'll need to configure postfix to use rspamd as a milter. Open \/usr\/local\/etc\/postfix\/main.cf\n        and uncomment the lines you added in <a href=\"#postfix\">the postfix section<\/a>.\n<\/p><pre>\/usr\/local\/etc\/postfix\/main.cfmilter_protocol = 6\n# if rspamd is down, don't reject mail\nmilter_default_action = accept\nsmtpd_milters = unix:\/var\/run\/rspamd\/milter.sock\nmilter_mail_macros = i {mail_addr} {client_addr} {client_name} {auth_authen}<\/pre><p>Reload postfix for your changes to take effect:\n<\/p><pre>service postfix reload<\/pre><p>At this point, rspamd should be filtering your incoming email as it arrives through\n        postfix. To verify, send yourself a test email and check the output of your mail log, along\n        with \/var\/log\/rspamd\/rspamd.log.  You also can view real-time\n        filtering stats in the rspamd web interface. Just set up an SSH tunnel to your server,\n        forward port 11334 to localhost, and point your web browser there.\n        <\/p><p>Once you've verified rspamd and postfix are working together, all that's left is to\n        configure Dovecot to train rspamd when you move messages in and out of your Junk folder.\n        The steps for achieving this were derived from <a href=\"https:\/\/wiki2.dovecot.org\/HowTo\/AntispamWithSieve\">this Dovecot guide<\/a>.\n        Add a new file, 90-imapsieve.conf, under Dovecot's conf.d\n        directory with the following contents:\n<\/p><pre>\/usr\/local\/etc\/dovecot\/conf.d\/90-imapsieve.confplugin {\n  sieve_plugins = sieve_imapsieve sieve_extprograms\n  # From elsewhere to Junk folder\n  imapsieve_mailbox1_name = Junk\n  imapsieve_mailbox1_causes = COPY\n  imapsieve_mailbox1_before = file:\/usr\/local\/etc\/dovecot\/sieve\/report-spam.sieve\n  # From Junk folder to elsewhere\n  imapsieve_mailbox2_name = *\n  imapsieve_mailbox2_from = Junk\n  imapsieve_mailbox2_causes = COPY\n  imapsieve_mailbox2_before = file:\/usr\/local\/etc\/dovecot\/sieve\/report-ham.sieve\n  sieve_pipe_bin_dir = \/usr\/local\/etc\/dovecot\/sieve\n  sieve_global_extensions = +vnd.dovecot.pipe\n}<\/pre><p>Restart Dovecot to load the necessary sieve plugins:\n<\/p><pre>service dovecot restart<\/pre><p>Create the following sieve scripts in the \/usr\/local\/etc\/dovecot\/sieve\n        directory:\n<\/p><pre>\/usr\/local\/etc\/dovecot\/sieve\/report-spam.sieverequire [\"vnd.dovecot.pipe\", \"copy\", \"imapsieve\", \"environment\", \"variables\"];\nif environment :matches \"imap.email\" \"*\" {\n  set \"email\" \"${1}\";\n}\npipe :copy \"train-spam.sh\" [ \"${email}\" ];<\/pre><pre>\/usr\/local\/etc\/dovecot\/sieve\/report-ham.sieverequire [\"vnd.dovecot.pipe\", \"copy\", \"imapsieve\", \"environment\", \"variables\"];\nif environment :matches \"imap.mailbox\" \"*\" {\n  set \"mailbox\" \"${1}\";\n}\nif string \"${mailbox}\" \"Trash\" {\n  stop;\n}\nif environment :matches \"imap.email\" \"*\" {\n  set \"email\" \"${1}\";\n}\npipe :copy \"train-ham.sh\" [ \"${email}\" ];<\/pre><p>These sieve scripts will be triggered whenever email gets moved in or out of your Junk\n        folder. You will need to manually compile them, since Dovecot doesn't have write permission\n        to this directory.\n<\/p><pre>cd \/usr\/local\/etc\/dovecot\/sieve\nsievec report-spam.sieve\nsievec report-ham.sieve<\/pre><p>Now you need to create the shell scripts that the sieve scripts will execute. Create\n        train-spam.sh and train-ham.sh\n        in the same directory:\n<\/p><pre>\/usr\/local\/etc\/dovecot\/sieve\/train-spam.shexec \/usr\/local\/bin\/rspamc -h \/var\/run\/rspamd\/rspamd.sock learn_spam<\/pre><pre>\/usr\/local\/etc\/dovecot\/sieve\/train-ham.shexec \/usr\/local\/bin\/rspamc -h \/var\/run\/rspamd\/rspamd.sock learn_ham<\/pre><p>Make sure they are executable.\n<\/p><pre>chmod +x \/usr\/local\/etc\/dovecot\/sieve\/train-{spam,ham}.sh<\/pre><p>Finally, we need to tell Dovecot to automatically move emails that rspamd tagged as spam to\n        your Junk folder. We do this using a global filter on the X-SPAM header.\n<\/p><pre>\/usr\/local\/etc\/dovecot\/sieve-before.d\/10-rspamd.sieverequire [\"fileinto\"];\nif header :is \"X-Spam\" \"Yes\" {\n  fileinto \"Junk\";\n}<\/pre><p>You'll need to pre-compile this script:\n<\/p><pre>cd \/usr\/local\/etc\/dovecot\/sieve-before.d\nsievec 10-rspamd.sieve<\/pre><p>Reload Dovecot for your changes to take effect.\n<\/p><pre>service dovecot reload<\/pre><p>Dovecot should now inform the rspamd daemon to retrain a messages as spam or ham when\n        you move it in or out of your Junk folder. To test, move a message in and out of Junk from\n        mail client, and check your mail and rspamd logs. You should a message in \/var\/log\/rspamd\/rspamd.log\n        whenever a message is retrained.\n      <\/p><\/section><section><h2 id=\"dkim\">DKIM: Validation for Your Outgoing Mail<\/h2><a href=\"http:\/\/www.dkim.org\">DKIM<\/a>, or DomainKeys Identified Mail, is a standard by\n        which mail servers can cryptographically verify that an email claiming to originate from\n        a given domain did indeed originate from a server within that domain's control. The general\n        idea is that the sender's mail server signs outgoing messages with a private key, and the\n        recipient's mail server retrieves the public key from a TXT record hosted in the sender's\n        DNS zone to verify the authenticity of the message. Using DKIM, a recipient can be certain\n        that a message really did come from the sender's authorized mail server, and that it wasn't\n        tampered with in transit.<p>Luckily, rspamd provides <a href=\"https:\/\/rspamd.com\/doc\/modules\/dkim_signing.html\">built-in support for DKIM signing<\/a>\n        using the postfix milter we already configured in the <a href=\"#rspamd\">previous section<\/a>.\n        That means there's no additional software to configure! All you need to do is generate your\n        private key, and create a DKIM record containing your public key in DNS.\n        <\/p><p>Recall the following file we created in the <a href=\"#rspamd\">rspamd section<\/a>:\n<\/p><pre>\/usr\/local\/etc\/rspamd\/local.d\/dkim_signing.confpath = \"\/var\/db\/rspamd\/dkim\/$domain.$selector.key\";\nselector = \"dkim\";<\/pre><p>The $domain variable does what you'd expect: an outgoing\n        email from an @example.com email address will be signed using\n        the corresponding example.com private key file. The $selector\n        variable bears further explanation: The DKIM specification allows you to use multiple keys\n        for a single domain, and distinguish between them using a <em>selector<\/em> value in the\n        DKIM header of the message. In our case, we'll keep it simple and just use a single selector\n        called \"dkim\" for everything.\n        <\/p><p>You can generate a DKIM keypair using the rspamadm\n        command:\n<\/p><pre>rspamadm dkim_keygen -k \/var\/db\/rspamd\/dkim\/example.com.dkim.key -b 2048 -s dkim -d example.com<\/pre><p>This will save a private key file for example.com in\n        rspamd's DKIM directory, and print a DNS record for the DKIM public key to standard output.\n        You should see something like this:\n<\/p><pre>dkim._domainkey IN TXT ( \"v=DKIM1; k=rsa; \"\n        \"p=MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAzGdxkFW0tIDYdNrGyj\/J2Hff7N\/9BEWE2qxMw6PBW5FhJRullZT9WNZOVrrXk1TsiBHRq8YQrSS1TfLbNV9PE7sE0vGx0eLgkiqnqLMwTy5Y9+jEbiNrddNR6v+TGHuMckYJO3JMjiROhMi\/86Lv6P\/rv2R\/lxFldCeYQxa41\/8LH+b3ZXWTLYRM6y2\/2UpGz\/wtknvA+DtO0rn+Y\"\n        \"uLuPrh+ftzmJb6i3g01XFgAO8ZzMLMdO\/k7UJDX\/Q6himKxVv2t3vSvS1MGqiWThXiU3WxhQED0zZUlkC5Lfx4BCo1h0v7fwZeMdu2NPOzlDBMDq5HRYgbwuFXTAmxSM7WRqQIDAQAB\"\n) ;<\/pre><p>If you are running your own DNS server, you can just paste that directly into your zone\n        file. If you are using a DNS provider's web interface, there's a small caveat: the 2048-bit\n        public key is too large to fit in a single DNS record. In a zone file, the parenthetical\n        syntax above causes your DNS server to split the key into separate records. When you add it\n        to your provder's web interface, you'll need to create a TXT record at\n        dkim._domainkey.example.com and generate the value by contatenating all\n        three lines together (removing the quotes).\n        <\/p><p>At this point, you should have both SPF and DKIM properly configured. However, you\n        haven't informed other mail servers what to <em>do<\/em> if either of those safeguards should\n        fail. That's where <a href=\"https:\/\/dmarc.org\">DMARC<\/a> comes in. DMARC uses a special DNS\n        record to declare how other mail servers should treat mail from your domain when validations\n        fail.\n        <\/p><p>Create a TXT record at _dmarc.example.com like the\n        following:\n<\/p><pre>\/usr\/local\/etc\/named\/master\/example.com.db_dmarc  IN  TXT  \"v=DMARC1; p=reject; adkim=r; aspf=r; sp=reject\"<\/pre><p>The p=reject field indicates that any emails from our\n        domain which do not pass SPF and DKIM checks should be assumed to be spam and rejected.\n        The adkim=r and aspf=r fields\n        indicate that we are using relaxed validation, which is the\n        default. Finally, the sp=reject field indicates that this\n        policy should apply to all subdomains of example.com.\n        <\/p><section><h3 id=\"testing-dkim\">Testing DKIM, DMARC, and SPF<\/h3><p>Once your public key and DMARC records are in DNS, you're ready to test DKIM signing. You\n          can check whether your records have propagated using dig:\n<\/p><pre>dig +short TXT dkim._domainkey.example.com\n# \"v=DKIM1\\; k=rsa\\; \" \"p=MIIBIjANBgkqhkiG9w0BAQEFAAOC...\ndig +short TXT _dmarc.example.com\n# \"v=DMARC1\\; p=reject\\; adkim=r\\; aspf=r\\; sp=reject\"\n<\/pre><p>There are lots of online tools you can use to test whether or not DKIM signing is\n          working for your domain. My favorite is <a href=\"http:\/\/dkimvalidator.com\">DKIMValidator.com<\/a>.\n          The site provides you a randomized email address. Simply send a test email from your mail\n          server to the provided email address, wait a few seconds, and hit \"View Results.\" You will\n          be presented with a report describing whether or not DKIM and SPF validation passed for your\n          domain.\n          <\/p><p>You can also test DKIM, DMARC, and SPF by sending a test email to a GMail account. In\n          your GMail inbox, click the arrow next to the reply button and choose \"Show Original.\"\n          You will be presented with a table containing pass\/fail results for SPF, DKIM, and DMARC.\n          <\/p><p>If any of the validations fail, check the message headers for clues, as well as your rspamd\n          and mail logs for errors. If they all passed, then congratulations\u2014your mail server is\n          more professionally configured than half the mail I get from \"real\" companies.\n        <\/p><\/section><\/section><section><h2 id=\"sieve\">Sieve: Scripting Your Mailbox<\/h2><p>Once of the best things about running your own mail server is being able to script your\n        inbox. Instead of creating filters using a kludgy synax through a web interface, you can\n        literally write programs to determine how to sort your mail. With sieve, regular\n        expressions, if\/else control structures, and full message metadata (including headers) are\n        at your disposal.\n        <\/p><p>Recall in <a href=\"#dovecot\">the Dovecot section<\/a> that sieve scripts live in the\n        <em>virtual home directory<\/em> of each IMAP user. Dovecot's sieve functionality allows\n        a user to have multiple sieve scripts stored on the server, but only one of them can be\n        \"active\" at any given time. In our configuration, the active sieve script for the user\n        alphonsus would be located at \/var\/mail\/vhosts\/example.com\/alphonsus\/.dovecot.sieve.\n        The .dovecot.sieve file should be a symlink to a script in the\n        sieve directory:\n<\/p><pre><strong>root@awesomebox:\/var\/mail\/vhosts\/example.com\/alphonsus # ls -l<\/strong>\ntotal 12\nlrwxr-xr-x  1 vmail  vmail   18 Mar  4 07:35 .dovecot.sieve@ -&gt; sieve\/my_cool_script.sieve\n-rw-rw-r--  1 vmail  vmail  472 Mar  4 07:38 .dovecot.svbin\ndrwx------  4 vmail  vmail  512 Aug 15 04:45 mdbox\/\ndrwxrwxr-x  2 vmail  vmail  512 Mar  4 07:35 sieve\/<\/pre><p>If you use a mail client with ManageSieve support (like KMail), these details are handled\n        for you automatically. You can also use <a href=\"https:\/\/github.com\/philpennock\/sieve-connect\">sieve-connect<\/a>,\n        a command-line ManageSieve client. I rarely modify my sieve script, so when I need to make a\n        change I just SSH to my server and edit it in vim.\n        <\/p><p>Let's make a basic sieve script. Open up a root shell and navigate to your IMAP user's\n        virtual home directory under \/var\/mail\/vhosts.\n<\/p><pre>cd \/var\/mail\/vhosts\/example.com\/user<\/pre><p>Create the sieve directory if it doesn't already exist:\n<\/p><pre>mkdir sieve<\/pre><p>Then, create a sieve script. You can name this file whatever you like:\n<\/p><pre>vim sieve\/my_awesome_script.sieve<\/pre><p>Now you're ready to write your script. You can check out some example sieve scripts on\n        the Dovecot wiki <a href=\"https:\/\/wiki2.dovecot.org\/Pigeonhole\/Sieve\/Examples\">here<\/a>.\n        We'll keep it simple for now.\n<\/p><pre>\/var\/mail\/vhosts\/example.com\/user\/sieve\/my_awesome_script.sieverequire [\"regex\", \"fileinto\", \"imap4flags\"];\n\/* put all of FreeBSD's `periodic` output into Logs folder *\/\nif allof (address :is \"from\" \"root@awesomebox.example.com\", header :contains \"subject\" \"run output\") {\n    fileinto \"Logs\";\n    stop;\n}\n\/* put online shopping\/advertising into Shopping folder *\/\nif anyof (address :is         \"from\" \"auto-confirm@amazon.com\",\n          address :is :domain \"from\" \"earthfare.com\") {\n   fileinto \"Shopping\";\n   stop;\n}<\/pre><p>I don't believe the sieve script will create any new IMAP folders for you, so be sure to\n        create any folders you need in your mail client first. Now you need to create a symlink so\n        Dovecot knows that this is your \"active\" script:\n<\/p><pre>ln -s sieve\/my_awesome_script.sieve .dovecot.sieve<\/pre><p>Make sure your script actually compiles:\n<\/p><pre>sievec .dovecot.sieve<\/pre><p>Finally, make sure everything is still owned by the vmail\n        user:\n<\/p><pre>chown -R vmail:vmail .dovecot.sieve .dovecot.svbin sieve<\/pre><p>Check your mail log the next time you receive an email to verify that Dovecot didn't have\n        any trouble executing your script. You can get as convoluted as you want with sieve\u2014I've\n        seen some cool setups where people call external shell scripts when certain emails are\n        receieved.\n      <\/p><\/section><section><h2 id=\"conclusion\">Conclusion<\/h2><p>If you made it this far, you should have a pretty decent personal email server. I've been\n        hosting my family's email on a VPS with this setup for a year or so with great results. I've\n        yet to have a spammer make it past rspamd, and sync and search work great with all of our\n        iDevices. Also, despite the warnings of various online naysayers, I've never had an issue\n        with others not receieving my mail, and tech recruiters, for better or worse, seem to have\n        no trouble reaching my inbox.\n        <\/p><p>One thing I didn't cover in this guide was a webmail setup\u2014mostly because I've never felt\n        the need for one. Mutt, Mail.app, and the iOS mail client have served me quite well, and I\n        honestly hate dealing with the additional attack vector of some PHP\/Ruby-off-the-Rails\n        web application. However, if that's something you need, I have heard good things about\n        <a href=\"https:\/\/roundcube.net\">Roundcube<\/a>.\n        <\/p><p>Also, and I can't stress this enough, <strong>back up your mail directory regularly!!<\/strong>\n        <\/p><p>No doubt, it's a lot of work to get a mail server up and running. But once you get it\n        working, it's very set-and-forget. Just keep your system updated, rotate your SSL\n        certificates before they expire, and use good DevOps practices, and it will serve you well\n        as long as you keep paying your hosting provider.\n        <\/p><p>I'm sure many will argue that this is too much work for too little gain, because the NSA can probably crack all our\n        encryption, most mail goes through GMail anyway, tons of email is still plain text, etc.\n        However, my <a href=\"https:\/\/www.c0ffee.net\/blog\/self-hosting-guide\">reasons for self-hosting things<\/a>\n        are more philosophical than technical. The net neutrality debate is moot if all of our\n        online communication is controlled by a handful of monolothic advertising companies. The\n        internet's greatest strength has always been in the free exchange of ideas and information.\n        Let's make sure it stays that way.\n      <\/p><\/section>","created_at":"2018-11-07T16:45:22+0000","updated_at":"2018-11-07T16:45:22+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":52,"domain_name":"www.c0ffee.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":351,"title":"How To Cleanly Integrate Java and Clojure In The Same Package","url":"https:\/\/joelholder.com\/2015\/10\/23\/how-to-cleanly-integrate-java-and-clojure-in-the-same-package\/","content":"<div class=\"figure\">\n<p><a href=\"https:\/\/uberpwn.files.wordpress.com\/2015\/10\/wpid-emacs-my-app.png\"><img class=\"alignnone\" src=\"https:\/\/uberpwn.files.wordpress.com\/2015\/10\/wpid-emacs-my-app.png?w=1920&amp;h=1040\" alt=\"emacs-my-app.png\" width=\"1920\" height=\"1040\" \/><\/a><\/p>\n<\/div>\n<h2>A hybrid Java\/Clojure library designed to demonstrate how to setup Java interop using Maven<\/h2>\n<div id=\"outline-container-orgheadline1\" class=\"outline-2\">\n<div id=\"text-orgheadline1\" class=\"outline-text-2\">\n<p>This is a complete Maven-first Clojure\/Java interop application. It details how to create a Maven application, enrich it with clojure code, call into clojure from Java, and hook up the entry points for both Java and Clojure within the same project.<\/p>\n<p>Further, it contains my starter examples of using the fantastic <a href=\"http:\/\/incanter.org\/\">Incanter Statistical and Graphics Computing Library<\/a> in clojure. I include both a pom.xml and a project.clj showing how to pull in the dependencies.<\/p>\n<p>The outcome is a consistent maven-archetyped project, wherein maven and leiningen play nicely together. This allows the best of both ways to be applied together. For the emacs user, I include support for cider and swank. NRepl by itself is present for general purpose use as well.<\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline35\" class=\"outline-2\">\n<h2 id=\"orgheadline35\">Starting a project<\/h2>\n<div id=\"text-orgheadline35\" class=\"outline-text-2\">\n<div id=\"outline-container-orgheadline34\" class=\"outline-3\">\n<h3 id=\"orgheadline34\">Maven first<\/h3>\n<div id=\"text-orgheadline34\" class=\"outline-text-3\">\n<div id=\"outline-container-orgheadline2\" class=\"outline-4\">\n<h4 id=\"orgheadline2\">Create Maven project<\/h4>\n<div id=\"text-orgheadline2\" class=\"outline-text-4\">\n<p>follow these steps<\/p>\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn archetype:generate -DgroupId=com.mycompany.app -DartifactId=my-app -DarchetypeArtifactId=maven-archetype-quickstart -DinteractiveMode=false&#13;\n&#13;\ncd my-app&#13;\n&#13;\nmvn package&#13;\n&#13;\njava -cp target\/my-app-1.0-SNAPSHOT.jar com.mycompany.app.App&#13;\n<\/pre>\n<pre class=\"example\">Hello World&#13;\n<\/pre>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline3\" class=\"outline-4\">\n<h4 id=\"orgheadline3\">Add Clojure code<\/h4>\n<div id=\"text-orgheadline3\" class=\"outline-text-4\">\n<p>Create a clojure core file<\/p>\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmkdir -p src\/main\/clojure\/com\/mycompany\/app&#13;\n&#13;\ntouch src\/main\/clojure\/com\/mycompany\/app\/core.clj&#13;\n<\/pre>\n<p>Give it some goodness\u2026<\/p>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n  (ns com.mycompany.app.core&#13;\n  (:gen-class)&#13;\n  (:use (incanter core stats charts)))&#13;\n&#13;\n(defn -main [&amp; args]&#13;\n  (println \"Hello Clojure!\")&#13;\n  (println \"Java main called clojure function with args: \"&#13;\n           (apply str (interpose \" \" args))))&#13;\n&#13;\n&#13;\n(defn run []&#13;\n  (view (histogram (sample-normal 1000))))&#13;\n<\/pre>\n<p>Notice that we\u2019ve added in the <a href=\"http:\/\/incanter.org\/\">Incanter Library<\/a> and made a run function to pop up a histogram of sample data<\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline4\" class=\"outline-4\">\n<h4 id=\"orgheadline4\">Add dependencies to your pom.xml<\/h4>\n<div id=\"text-orgheadline4\" class=\"outline-text-4\">\n<pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n  &lt;dependencies&gt;&#13;\n    &lt;dependency&gt;&#13;\n      &lt;groupId&gt;org.clojure&lt;\/groupId&gt;&#13;\n      &lt;artifactId&gt;clojure&lt;\/artifactId&gt;&#13;\n      &lt;version&gt;1.7.0&lt;\/version&gt;&#13;\n    &lt;\/dependency&gt;&#13;\n    &lt;dependency&gt;&#13;\n      &lt;groupId&gt;org.clojure&lt;\/groupId&gt;&#13;\n      &lt;artifactId&gt;clojure-contrib&lt;\/artifactId&gt;&#13;\n      &lt;version&gt;1.2.0&lt;\/version&gt;&#13;\n    &lt;\/dependency&gt;&#13;\n    &lt;dependency&gt;&#13;\n      &lt;groupId&gt;incanter&lt;\/groupId&gt;&#13;\n      &lt;artifactId&gt;incanter&lt;\/artifactId&gt;&#13;\n      &lt;version&gt;1.9.0&lt;\/version&gt;&#13;\n    &lt;\/dependency&gt;&#13;\n    &lt;dependency&gt;&#13;\n      &lt;groupId&gt;org.clojure&lt;\/groupId&gt;&#13;\n      &lt;artifactId&gt;tools.nrepl&lt;\/artifactId&gt;&#13;\n      &lt;version&gt;0.2.10&lt;\/version&gt;&#13;\n    &lt;\/dependency&gt;&#13;\n   &lt;!-- pick your poison swank or cider. just make sure the version of nRepl matches. --&gt;&#13;\n    &lt;dependency&gt;&#13;\n      &lt;groupId&gt;cider&lt;\/groupId&gt;&#13;\n      &lt;artifactId&gt;cider-nrepl&lt;\/artifactId&gt;&#13;\n      &lt;version&gt;0.10.0-SNAPSHOT&lt;\/version&gt;&#13;\n    &lt;\/dependency&gt;&#13;\n    &lt;dependency&gt;&#13;\n      &lt;groupId&gt;swank-clojure&lt;\/groupId&gt;&#13;\n      &lt;artifactId&gt;swank-clojure&lt;\/artifactId&gt;&#13;\n      &lt;version&gt;1.4.3&lt;\/version&gt;&#13;\n    &lt;\/dependency&gt;&#13;\n  &lt;\/dependencies&gt;&#13;\n<\/pre>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline5\" class=\"outline-4\">\n<h4 id=\"orgheadline5\">Java main class<\/h4>\n<div id=\"text-orgheadline5\" class=\"outline-text-4\">\n<p>Modify your java main to call your clojure main like in the following:<\/p>\n<pre class=\"brush: java; title: ; notranslate\" title=\"\">&#13;\npackage com.mycompany.app;&#13;\n&#13;\n\/\/ for clojure's api&#13;\nimport clojure.lang.IFn;&#13;\nimport clojure.java.api.Clojure;&#13;\n&#13;\n\/\/ for my api&#13;\nimport clojure.lang.RT;&#13;\n&#13;\npublic class App&#13;\n{&#13;\n  public static void main( String[] args )&#13;\n  {&#13;\n&#13;\n    System.out.println(\"Hello Java!\" );&#13;\n&#13;\n    try {&#13;\n&#13;\n      \/\/ running my clojure code&#13;\n      RT.loadResourceScript(\"com\/mycompany\/app\/core.clj\");&#13;\n      IFn main = RT.var(\"com.mycompany.app.core\", \"main\");&#13;\n      main.invoke(args);&#13;\n&#13;\n      \/\/ running the clojure api&#13;\n      IFn plus = Clojure.var(\"clojure.core\", \"+\");&#13;\n      System.out.println(plus.invoke(1, 2).toString());&#13;\n&#13;\n    } catch(Exception e) {&#13;\n      e.printStackTrace();&#13;\n    }&#13;\n&#13;\n  }&#13;\n}&#13;\n<\/pre>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline6\" class=\"outline-4\">\n<h4 id=\"orgheadline6\">Maven plugins for building<\/h4>\n<div id=\"text-orgheadline6\" class=\"outline-text-4\">\n<p>You should add in these plugins to your pom.xml<\/p>\n<\/div>\n<ul class=\"org-ul\"><li>Add the maven-assembly-plugin\n<div id=\"text-orgheadline7\" class=\"outline-text-5\">\n<p>Create an Ubarjar<\/p>\n<p>Bind the maven-assembly-plugin to the package phase this will create a jar file without the dependencies suitable for deployment to a container with deps present.<\/p>\n<pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n  &lt;plugin&gt;&#13;\n    &lt;artifactId&gt;maven-assembly-plugin&lt;\/artifactId&gt;&#13;\n    &lt;configuration&gt;&#13;\n      &lt;descriptorRefs&gt;&#13;\n        &lt;descriptorRef&gt;jar-with-dependencies&lt;\/descriptorRef&gt;&#13;\n      &lt;\/descriptorRefs&gt;&#13;\n      &lt;archive&gt;&#13;\n        &lt;manifest&gt;&#13;\n&#13;\n          &lt;!-- use clojure main --&gt;&#13;\n          &lt;!-- &lt;mainClass&gt;com.mycompany.app.core&lt;\/mainClass&gt; --&gt;&#13;\n&#13;\n          &lt;!-- use java main --&gt;&#13;\n          &lt;mainClass&gt;com.mycompany.app.App&lt;\/mainClass&gt;&#13;\n&#13;\n        &lt;\/manifest&gt;&#13;\n      &lt;\/archive&gt;&#13;\n    &lt;\/configuration&gt;&#13;\n    &lt;executions&gt;&#13;\n      &lt;execution&gt;&#13;\n        &lt;id&gt;make-assembly&lt;\/id&gt;&#13;\n        &lt;phase&gt;package&lt;\/phase&gt;&#13;\n        &lt;goals&gt;&#13;\n          &lt;goal&gt;single&lt;\/goal&gt;&#13;\n        &lt;\/goals&gt;&#13;\n      &lt;\/execution&gt;&#13;\n    &lt;\/executions&gt;&#13;\n  &lt;\/plugin&gt;&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Add the clojure-maven-plugin\n<div id=\"text-orgheadline8\" class=\"outline-text-5\">\n<p>Add this plugin to give your project the mvn: clojure:\u2026 commands<\/p>\n<p>A full list of these is posted later in this article.<\/p>\n<pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n  &lt;plugin&gt;&#13;\n    &lt;groupId&gt;com.theoryinpractise&lt;\/groupId&gt;&#13;\n    &lt;artifactId&gt;clojure-maven-plugin&lt;\/artifactId&gt;&#13;\n    &lt;version&gt;1.7.1&lt;\/version&gt;&#13;\n    &lt;configuration&gt;&#13;\n      &lt;mainClass&gt;com.mycompany.app.core&lt;\/mainClass&gt;&#13;\n    &lt;\/configuration&gt;&#13;\n    &lt;executions&gt;&#13;\n      &lt;execution&gt;&#13;\n        &lt;id&gt;compile-clojure&lt;\/id&gt;&#13;\n        &lt;phase&gt;compile&lt;\/phase&gt;&#13;\n        &lt;goals&gt;&#13;\n          &lt;goal&gt;compile&lt;\/goal&gt;&#13;\n        &lt;\/goals&gt;&#13;\n      &lt;\/execution&gt;&#13;\n      &lt;execution&gt;&#13;\n        &lt;id&gt;test-clojure&lt;\/id&gt;&#13;\n        &lt;phase&gt;test&lt;\/phase&gt;&#13;\n        &lt;goals&gt;&#13;\n          &lt;goal&gt;test&lt;\/goal&gt;&#13;\n        &lt;\/goals&gt;&#13;\n      &lt;\/execution&gt;&#13;\n    &lt;\/executions&gt;&#13;\n  &lt;\/plugin&gt;&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Add the maven-compiler-plugin\n<div id=\"text-orgheadline9\" class=\"outline-text-5\">\n<p>Add Java version targeting<\/p>\n<p>This is always good to have if you are working against multiple versions of Java.<\/p>\n<pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n  &lt;plugin&gt;&#13;\n    &lt;groupId&gt;org.apache.maven.plugins&lt;\/groupId&gt;&#13;\n    &lt;artifactId&gt;maven-compiler-plugin&lt;\/artifactId&gt;&#13;\n    &lt;version&gt;3.3&lt;\/version&gt;&#13;\n    &lt;configuration&gt;&#13;\n      &lt;source&gt;1.8&lt;\/source&gt;&#13;\n      &lt;target&gt;1.8&lt;\/target&gt;&#13;\n    &lt;\/configuration&gt;&#13;\n  &lt;\/plugin&gt;&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Add the maven-exec-plugin\n<div id=\"text-orgheadline10\" class=\"outline-text-5\">\n<p>Add this plugin to give your project the mvn exec:\u2026 commands<\/p>\n<p>The maven-exec-plugin is nice for running your project from the commandline, build scripts, or from inside an IDE.<\/p>\n<pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n  &lt;plugin&gt;&#13;\n    &lt;groupId&gt;org.codehaus.mojo&lt;\/groupId&gt;&#13;\n    &lt;artifactId&gt;exec-maven-plugin&lt;\/artifactId&gt;&#13;\n    &lt;version&gt;1.4.0&lt;\/version&gt;&#13;\n    &lt;executions&gt;&#13;\n      &lt;execution&gt;&#13;\n        &lt;goals&gt;&#13;\n          &lt;goal&gt;exec&lt;\/goal&gt;&#13;\n        &lt;\/goals&gt;&#13;\n      &lt;\/execution&gt;&#13;\n    &lt;\/executions&gt;&#13;\n    &lt;configuration&gt;&#13;\n      &lt;mainClass&gt;com.mycompany.app.App&lt;\/mainClass&gt;&#13;\n    &lt;\/configuration&gt;&#13;\n  &lt;\/plugin&gt;&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Add the maven-jar-plugin\n<div id=\"text-orgheadline11\" class=\"outline-text-5\">\n<p>With this plugin you can manipulate the manifest of your default package. In this case, I\u2019m not adding a main, because I\u2019m using the uberjar above with all the dependencies for that. However, I included this section for cases, where the use case is for a non-stand-alone assembly.<\/p>\n<pre class=\"brush: xml; title: ; notranslate\" title=\"\">&#13;\n  &lt;plugin&gt;&#13;\n    &lt;groupId&gt;org.apache.maven.plugins&lt;\/groupId&gt;&#13;\n    &lt;artifactId&gt;maven-jar-plugin&lt;\/artifactId&gt;&#13;\n    &lt;version&gt;2.6&lt;\/version&gt;&#13;\n    &lt;configuration&gt;&#13;\n      &lt;archive&gt;&#13;\n        &lt;manifest&gt;&#13;\n&#13;\n          &lt;!-- use clojure main --&gt;&#13;\n          &lt;!-- &lt;mainClass&gt;com.mycompany.app.core&lt;\/mainClass&gt; --&gt;&#13;\n&#13;\n          &lt;!-- use java main --&gt;&#13;\n          &lt;!-- &lt;mainClass&gt;com.mycompany.app.App&lt;\/mainClass&gt; --&gt;&#13;\n&#13;\n        &lt;\/manifest&gt;&#13;\n      &lt;\/archive&gt;&#13;\n    &lt;\/configuration&gt;&#13;\n  &lt;\/plugin&gt;&#13;\n<\/pre>\n<\/div>\n<\/li>\n<\/ul><\/div>\n<div id=\"outline-container-orgheadline29\" class=\"outline-4\">\n<h4 id=\"orgheadline29\">Using Maven<\/h4>\n<div id=\"text-orgheadline29\" class=\"outline-text-4\">\n<ul class=\"org-ul\"><li>building\n<div id=\"text-orgheadline12\" class=\"outline-text-5\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn package&#13;\n<\/pre>\n<\/div>\n<ul class=\"org-ul\"><li>Run from cli with\n<ul class=\"org-ul\"><li>run from java entry point:\n<div id=\"text-orgheadline13\" class=\"outline-text-7\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\njava -cp target\/my-app-1.0-SNAPSHOT-jar-with-dependencies.jar com.mycompany.app.App&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Run from Clojure entry point:\n<div id=\"text-orgheadline14\" class=\"outline-text-7\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\njava -cp target\/my-app-1.0-SNAPSHOT-jar-with-dependencies.jar com.mycompany.app.core&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Run with entry point specified in uberjar MANIFEST.MF:\n<div id=\"text-orgheadline15\" class=\"outline-text-7\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\njava -jar target\/my-app-1.0-SNAPSHOT-jar-with-dependencies.jar&#13;\n<\/pre>\n<\/div>\n<\/li>\n<\/ul><\/li>\n<li>Run from maven-exec-plugin\n<ul class=\"org-ul\"><li>With plugin specified entry point:\n<div id=\"text-orgheadline17\" class=\"outline-text-7\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn exec:java&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Specify your own entry point:\n<ul class=\"org-ul\"><li>Java main\n<div id=\"text-orgheadline18\" class=\"outline-text-8\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn exec:java -Dexec.mainClass=\"com.mycompany.app.App\"&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Clojure main\n<div id=\"text-orgheadline19\" class=\"outline-text-8\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn exec:java -Dexec.mainClass=\"com.mycompany.app.core\"&#13;\n<\/pre>\n<\/div>\n<\/li>\n<\/ul><\/li>\n<li>Feed args with this directive\n<div id=\"text-orgheadline21\" class=\"outline-text-7\">\n<pre class=\"brush: plain; title: ; notranslate\" title=\"\">&#13;\n-Dexec.args=\"foo\"&#13;\n<\/pre>\n<\/div>\n<\/li>\n<\/ul><\/li>\n<li>Run with maven-clojure-plugin\n<ul class=\"org-ul\"><li>Clojure main\n<div id=\"text-orgheadline23\" class=\"outline-text-7\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn clojure:run&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Clojure test\n<ul class=\"org-ul\"><li>Add a test\n<div id=\"text-orgheadline24\" class=\"outline-text-8\">\n<p>In order to be consistent with the test location convention in maven, create a path and clojure test file like this:<\/p>\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmkdir src\/test\/clojure\/com\/mycompany\/app&#13;\n&#13;\ntouch src\/test\/clojure\/com\/mycompany\/app\/core_test.clj&#13;\n<\/pre>\n<p>Add the following content:<\/p>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(ns com.mycompany.app.core-test&#13;\n  (:require [clojure.test :refer :all]&#13;\n            [com.mycompany.app.core :refer :all]))&#13;\n&#13;\n(deftest a-test&#13;\n  (testing \"Rigourous Test :-)\"&#13;\n    (is (= 0 0))))&#13;\n<\/pre>\n<\/div>\n<\/li>\n<\/ul><\/li>\n<li>Testing\n<div id=\"text-orgheadline26\" class=\"outline-text-7\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn clojure:test&#13;\n<\/pre>\n<p>Or<\/p>\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nmvn clojure:test-with-junit&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Available Maven clojure:\u2026 commands\n<div id=\"text-orgheadline27\" class=\"outline-text-7\">\n<p>Here is the full set of options available from the clojure-maven-plugin:<\/p>\n<pre class=\"brush: plain; title: ; notranslate\" title=\"\">&#13;\nmvn ...&#13;\n&#13;\nclojure:add-source&#13;\nclojure:add-test-source&#13;\nclojure:compile&#13;\nclojure:test&#13;\nclojure:test-with-junit&#13;\nclojure:run&#13;\nclojure:repl&#13;\nclojure:nrepl&#13;\nclojure:swank&#13;\nclojure:nailgun&#13;\nclojure:gendoc&#13;\nclojure:autodoc&#13;\nclojure:marginalia&#13;\n<\/pre>\n<p>See documentation:<\/p>\n<p><a href=\"https:\/\/github.com\/talios\/clojure-maven-plugin\">https:\/\/github.com\/talios\/clojure-maven-plugin<\/a><\/p>\n<\/div>\n<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><\/div>\n<div id=\"outline-container-orgheadline33\" class=\"outline-4\">\n<h4 id=\"orgheadline33\">Add Leiningen support<\/h4>\n<div id=\"text-orgheadline33\" class=\"outline-text-4\">\n<ul class=\"org-ul\"><li>Create project.clj\n<div id=\"text-orgheadline30\" class=\"outline-text-5\">\n<p>Next to your pom.xml, create the Clojure project file<\/p>\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\ntouch project.clj&#13;\n<\/pre>\n<p>Add this content<\/p>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(defproject my-sandbox \"1.0-SNAPSHOT\"&#13;\n :description \"My Encanter Project\"&#13;\n :url \"http:\/\/joelholder.com\"&#13;\n :license {:name \"Eclipse Public License\"&#13;\n           :url \"http:\/\/www.eclipse.org\/legal\/epl-v10.html\"}&#13;\n :dependencies [[org.clojure\/clojure \"1.7.0\"]&#13;\n                [incanter \"1.9.0\"]]&#13;\n :main com.mycompany.app.core&#13;\n :source-paths [\"src\/main\/clojure\"]&#13;\n :java-source-paths [\"src\/main\/java\"]&#13;\n :test-paths [\"src\/test\/clojure\"]&#13;\n :resource-paths [\"resources\"]&#13;\n :aot :all)&#13;\n<\/pre>\n<p>Note that we\u2019ve set the source code and test paths for both java and clojure to match the maven-way of doing this.<\/p>\n<p>This gives us a consistent way of hooking the code from both <code>lein<\/code> and <code>mvn<\/code>. Additionally, I\u2019ve added the incanter library here. The dependency should be expressed in the project file, because when we run nRepl from this directory, we want it to be available in our namespace, i.e. <code>com.mycompany.app.core<\/code><\/p>\n<\/div>\n<\/li>\n<li>Run with Leiningen\n<div id=\"text-orgheadline31\" class=\"outline-text-5\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nlein run&#13;\n<\/pre>\n<\/div>\n<\/li>\n<li>Test with Leiningen\n<div id=\"text-orgheadline32\" class=\"outline-text-5\">\n<pre class=\"brush: bash; title: ; notranslate\" title=\"\">&#13;\nlein test&#13;\n<\/pre>\n<\/div>\n<\/li>\n<\/ul><\/div>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline36\" class=\"outline-2\">\n<h2 id=\"orgheadline36\">Running with org-babel<\/h2>\n<div id=\"text-orgheadline36\" class=\"outline-text-2\">\n<p>This blog entry was exported to html from the README.org of this project.  It sits in the base directory of the project.  By using it to describe the project and include executable blocks of code from the project itself, we\u2019re able to provide working examples of how to use the library in it\u2019s documentation.  People can simply clone our project and try out the library by executing it\u2019s documentation.  Very nice..<\/p>\n<p>Make sure you jack-in to cider first:<\/p>\n<p>M-x cider-jack-in (Have it mapped to F9 in my emacs)<\/p>\n<\/div>\n<div id=\"outline-container-orgheadline37\" class=\"outline-3\">\n<h3 id=\"orgheadline37\">Clojure code<\/h3>\n<div id=\"text-orgheadline37\" class=\"outline-text-3\">\n<p>The Clojure code block<\/p>\n<pre class=\"example\">&#13;\n#+begin_src clojure :tangle .\/src\/main\/clojure\/com\/mycompany\/app\/core.clj :results output &#13;\n  (-main)&#13;\n  (run)&#13;\n#+end_src&#13;\n<\/pre>\n<p>Blocks are run in org-mode with C-c C-c <\/p>\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(-main)&#13;\n(run)&#13;\n<\/pre>\n<pre class=\"example\">Hello Clojure!&#13;\nJava main called clojure function with args:&#13;\n<\/pre>\n<p>Note that we ran both our main and run functions here. -main prints out the text shown above. The run function actually opens the incanter java image viewer and shows us a picture of our graph.<\/p>\n<div class=\"figure\">\n<p><img src=\"https:\/\/uberpwn.files.wordpress.com\/2015\/10\/wpid-run.png?w=840\" alt=\"run.png\" \/><\/p>\n<\/div>\n<p>I have purposefully not invested in styling these graphs in order to keep the code examples simple and focussed, however incanter makes really beautiful output. Here\u2019s a link to get you started:<\/p>\n<p><a href=\"http:\/\/incanter.org\/\">http:\/\/incanter.org\/<\/a><\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline38\" class=\"outline-3\">\n<h3 id=\"orgheadline38\">Playing with Incanter<\/h3>\n<div id=\"text-orgheadline38\" class=\"outline-text-3\">\n<pre class=\"brush: clojure; title: ; notranslate\" title=\"\">&#13;\n(use '(incanter core charts pdf))&#13;\n;;; Create the x and y data:&#13;\n(def x-data [0.0 1.0 2.0 3.0 4.0 5.0])&#13;\n(def y-data [2.3 9.0 2.6 3.1 8.1 4.5])&#13;\n(def xy-line (xy-plot x-data y-data))&#13;\n(view xy-line)&#13;\n(save-pdf xy-line \"img\/incanter-xy-line.pdf\")&#13;\n(save xy-line \"img\/incanter-xy-line.png\")&#13;\n<\/pre>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline39\" class=\"outline-3\">\n<h3 id=\"orgheadline39\">PNG<\/h3>\n<div id=\"text-orgheadline39\" class=\"outline-text-3\">\n<div class=\"figure\">\n<p><img src=\"https:\/\/uberpwn.files.wordpress.com\/2015\/10\/wpid-incanter-xy-line.png?w=840\" alt=\"incanter-xy-line.png\" \/><\/p>\n<\/div>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline40\" class=\"outline-3\">\n<h3 id=\"orgheadline40\">PDF<\/h3>\n<div id=\"text-orgheadline40\" class=\"outline-text-3\">\n<p><a href=\"https:\/\/uberpwn.files.wordpress.com\/2015\/10\/wpid-incanter-xy-line.pdf\">img\/incanter-xy-line.pdf<\/a><\/p>\n<\/div>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline41\" class=\"outline-2\">\n<h2 id=\"orgheadline41\">Resources<\/h2>\n<div id=\"text-orgheadline41\" class=\"outline-text-2\">\n<p>Finally here are some resources to move you along the journey. I drew on the links cited below along with a night of hacking to arrive a nice clean interop skeleton. Feel free to use my code available here:<\/p>\n<p><a href=\"https:\/\/github.com\/jclosure\/my-app\">https:\/\/github.com\/jclosure\/my-app<\/a><\/p>\n<p>For the eager, here is a link to my full pom:<\/p>\n<p><a href=\"https:\/\/github.com\/jclosure\/my-app\/blob\/master\/pom.xml\">https:\/\/github.com\/jclosure\/my-app\/blob\/master\/pom.xml<\/a><\/p>\n<\/div>\n<div id=\"outline-container-orgheadline42\" class=\"outline-3\">\n<h3 id=\"orgheadline42\">Org-babel clojure<\/h3>\n<div id=\"text-orgheadline42\" class=\"outline-text-3\">\n<p><a href=\"http:\/\/orgmode.org\/worg\/org-contrib\/babel\/languages\/ob-doc-clojure.html\">http:\/\/orgmode.org\/worg\/org-contrib\/babel\/languages\/ob-doc-clojure.html<\/a><\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline43\" class=\"outline-3\">\n<h3 id=\"orgheadline43\">Org-scraps<\/h3>\n<div id=\"text-orgheadline43\" class=\"outline-text-3\">\n<p><a href=\"https:\/\/eschulte.github.io\/org-scraps\/\">https:\/\/eschulte.github.io\/org-scraps\/<\/a><\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline44\" class=\"outline-3\">\n<h3 id=\"orgheadline44\">Project setup<\/h3>\n<div id=\"text-orgheadline44\" class=\"outline-text-3\">\n<p><a href=\"http:\/\/data-sorcery.org\/2009\/11\/20\/leiningen-clojars\/\">http:\/\/data-sorcery.org\/2009\/11\/20\/leiningen-clojars\/<\/a><\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline45\" class=\"outline-3\">\n<h3 id=\"orgheadline45\">Working with Apache Storm (multilang)<\/h3>\n<div id=\"text-orgheadline45\" class=\"outline-text-3\">\n<p>Starter project:<\/p>\n<p>This incubator project from the Apache Foundation demos drinking from the twitter hose with twitter4j and fishing in the streams with Java, Clojure, Python, and Ruby. Very cool and very powerful..<\/p>\n<p><a href=\"https:\/\/github.com\/apache\/storm\/tree\/master\/examples\/storm-starter\">https:\/\/github.com\/apache\/storm\/tree\/master\/examples\/storm-starter<\/a><\/p>\n<p>Testing Storm Topologies in Clojure:<\/p>\n<p><a href=\"http:\/\/www.pixelmachine.org\/2011\/12\/17\/Testing-Storm-Topologies.html\">http:\/\/www.pixelmachine.org\/2011\/12\/17\/Testing-Storm-Topologies.html<\/a><\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline46\" class=\"outline-3\">\n<h3 id=\"orgheadline46\">Vinyasa<\/h3>\n<div id=\"text-orgheadline46\" class=\"outline-text-3\">\n<p>READ this to give your clojure workflow more flow<\/p>\n<p><a href=\"https:\/\/github.com\/zcaudate\/vinyasa\">https:\/\/github.com\/zcaudate\/vinyasa<\/a><\/p>\n<\/div>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline47\" class=\"outline-2\">\n<h2 id=\"orgheadline47\">Wrapping up<\/h2>\n<div id=\"text-orgheadline47\" class=\"outline-text-2\">\n<p>Clojure and Java are siblings on the JVM; they should play nicely together. Maven enables them to be easily mixed together in the same project or between projects.  For a more indepth example of creating and consuming libraries written in Clojure, see Michael Richards\u2019 article detailing how to use Clojure to implement interfaces defined in Java. He uses a FactoryMethod to abstract the mechanics of getting the implementation back into Java, which make\u2019s the Clojure code virtually invisible from an API perspective. Very nice. Here\u2019s the link:<\/p>\n<p><a href=\"http:\/\/michaelrkytch.github.io\/programming\/clojure\/interop\/2015\/05\/26\/clj-interop-require.html\">http:\/\/michaelrkytch.github.io\/programming\/clojure\/interop\/2015\/05\/26\/clj-interop-require.html<\/a><\/p>\n<p>Happy hacking!..<\/p>\n<\/div>\n<\/div>\n\t<\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:45:22+0000","updated_at":"2018-11-07T16:45:22+0000","published_at":"2015-10-24T04:52:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"joelholder.com","preview_picture":"https:\/\/uberpwn.files.wordpress.com\/2015\/10\/wpid-incanter-xy-line.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":352,"title":"How to detect OS in Java \u2013 Mkyong.com","url":"http:\/\/www.mkyong.com\/java\/how-to-detect-os-in-java-systemgetpropertyosname\/","content":"<div id=\"post-content\"><div class=\"pic\"><img src=\"http:\/\/www.mkyong.com\/wp-content\/uploads\/2009\/05\/OS_thumb.jpg\" alt=\"\" title=\"OS_thumb\" width=\"300\" height=\"300\" \/><\/div><p>Here is a handy Java class that use <code>System.getProperty(\"os.name\")<\/code> to detect which type of operating system (OS) you are using now.<\/p><p>This code can detect \u201cWindows\u201d, \u201cMac\u201d, \u201cUnix\u201d and \u201cSolaris\u201d.<\/p><pre class=\"language-java\">&#13;\npackage com.mkyong.core;&#13;\n&#13;\npublic class OSValidator {&#13;\n&#13;\n\tprivate static String OS = System.getProperty(\"os.name\").toLowerCase();&#13;\n\t&#13;\n\tpublic static void main(String[] args) {&#13;\n\t\t&#13;\n\t\tSystem.out.println(OS);&#13;\n\t\t&#13;\n\t\tif (isWindows()) {&#13;\n\t\t\tSystem.out.println(\"This is Windows\");&#13;\n\t\t} else if (isMac()) {&#13;\n\t\t\tSystem.out.println(\"This is Mac\");&#13;\n\t\t} else if (isUnix()) {&#13;\n\t\t\tSystem.out.println(\"This is Unix or Linux\");&#13;\n\t\t} else if (isSolaris()) {&#13;\n\t\t\tSystem.out.println(\"This is Solaris\");&#13;\n\t\t} else {&#13;\n\t\t\tSystem.out.println(\"Your OS is not support!!\");&#13;\n\t\t}&#13;\n\t}&#13;\n&#13;\n\tpublic static boolean isWindows() {&#13;\n&#13;\n\t\treturn (OS.indexOf(\"win\") &gt;= 0);&#13;\n&#13;\n\t}&#13;\n&#13;\n\tpublic static boolean isMac() {&#13;\n&#13;\n\t\treturn (OS.indexOf(\"mac\") &gt;= 0);&#13;\n&#13;\n\t}&#13;\n&#13;\n\tpublic static boolean isUnix() {&#13;\n&#13;\n\t\treturn (OS.indexOf(\"nix\") &gt;= 0 || OS.indexOf(\"nux\") &gt;= 0 || OS.indexOf(\"aix\") &gt; 0 );&#13;\n\t\t&#13;\n\t}&#13;\n&#13;\n\tpublic static boolean isSolaris() {&#13;\n&#13;\n\t\treturn (OS.indexOf(\"sunos\") &gt;= 0);&#13;\n&#13;\n\t}&#13;\n&#13;\n}&#13;\n<\/pre><ol><li>os.name value \u2013 <a href=\"http:\/\/lopica.sourceforge.net\/os.html\" target=\"_blank\">http:\/\/lopica.sourceforge.net\/os.html<\/a><\/li>\n<li>System.getProperty() \u2013 <a href=\"http:\/\/java.sun.com\/j2se\/1.4.2\/docs\/api\/java\/lang\/System.html\" target=\"_blank\">http:\/\/java.sun.com\/j2se\/1.4.2\/docs\/api\/java\/lang\/System.html<\/a><\/li>\n<\/ol><\/div><h4>About the Author<\/h4><h4>Comments<\/h4>","created_at":"2018-11-07T16:45:22+0000","updated_at":"2018-11-07T16:45:22+0000","published_by":["mkyong"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.mkyong.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":346,"title":"'Red Oceans': How to Find Profitable Startup Ideas","url":"https:\/\/capitalandgrowth.org\/articles\/770\/how-to-find-profitable-business-ideas.html","content":"<img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/247-header.jpg\" alt=\"image\" \/><p><em><strong>The reason why I wrote the piece is because most people in the entrepreneurial\/startup world seem to focus specifically on chasing \"unicorns\" and the next big \"billion dollar idea\". One thing that I have noticed (in my business as an M&amp;A advisor who represents business owners selling their lower\/mid-market businesses) is that nearly all of the business owners I deal with have gotten rich by executing well on Red Ocean ideas. They are not hitting grand slams, but they are hitting singles and doubles. And the odds of hitting a base hit or a double look a hell of a lot better than the odds of hitting a grand slam. And to me, selling a $5 to $15 million dollar business is a pretty damn good way to eventually win the game.<\/strong><br \/><\/em>\n               <br \/>In an era of tech startups, angel investors, venture capital funds and unicorn hunting, an often-overlooked avenue for entrepreneurs to create successful businesses with great long-term potential is to discover viable Red Ocean business opportunities within established markets. \n               <\/p><p>Generally, there are two ways that entrepreneurs can create a successful business by meeting a market\u2019s unmet needs \u2013 \n               <a href=\"https:\/\/capitalandgrowth.org\/articles\/720\/part-3-six-paths-to-discovering-a-blue-ocean-idea.html\">Blue Oceans<\/a> and Red Oceans. \n               <\/p><p>The first method, discovering Blue Oceans, is sexy and generates tons of press. It involves identifying unmet customer needs within a market, and then solving those unmet needs with a new, customized business solution. \n               <\/p><p>Phrases like \u201cBlue Ocean Strategy\u201d and \u201cbusiness model discovery\u201d are tossed around liberally to describe this approach. If executed correctly, this method can yield entrepreneurs billions of dollars. Examples of successful strategy type have been Google\u2019s search engine, Apple\u2019s iPhone and the Ford Motor Company\u2019s Model T. W. Chan Kim and Ren\u00e9e Mauborgne refer to this type of business idea as a \n               <a href=\"https:\/\/capitalandgrowth.org\/articles\/720\/part-3-six-paths-to-discovering-a-blue-ocean-idea.html\">Blue Ocean Strategy<\/a>. \n               <\/p><p>The second method, finding a good Red Ocean business opportunity, is not particularly sexy, but it is tried and true. Entrepreneurs can identify if a market problem is being underserved by the current providers of solutions for that particular market problem. \n               <\/p><p>If so, then entrepreneurs can replicate an existing business solution for that particular market to address that unmet market demand. An example of this would be an entrepreneur opening a casual, family-style restaurant in a small town with sufficient demand for that type of restaurant. \n               <\/p><p>Casual, family style restaurants have been in existence for decades, if not centuries. The entrepreneur is not creating a new, customized business solution to address the market\u2019s demand. Instead, the entrepreneur identified a \u201cgap\u201d in a market that already exists and is simply filling that gap with a readily available business solution. W. Chan Kim and Ren\u00e9e Mauborgne refer to this type of business idea as a \n               <a href=\"https:\/\/www.blueoceanstrategy.com\/tools\/\"><strong>Red Ocean strategy<\/strong><\/a>.\n               <sup><sup>[1]<\/sup><\/sup><\/p><p>We have already discussed \n               <a href=\"https:\/\/capitalandgrowth.org\/articles\/694\/blue-ocean-strategy-finding-and-exploiting-unconte.html\">Blue Ocean Strategy<\/a> in detail here on Capital &amp; Growth. Now we will focus in on often overlooked ways to discover Red Ocean business ideas. Red Ocean business ideas can be uncovered through using either of both of the two following processes: \n               <\/p><p>1.The Internal Approach \n               <\/p><p>2.The External Approach \n               <br \/><\/p><ul><li>Observation<\/li>\n                <li>Focus Groups <\/li>\n                <li>Reverse Brainstorming <\/li>\n                <li>Market Growth <\/li>\n                <li>Matrix Charting for Insights <\/li>\n                <li>The \u201cSlice of Life\u201d Approach <\/li>\n                <li>The Market-Area Saturation Approach <\/li>\n                <li>The Competitive Matrix Approach <\/li>\n               <\/ul><br \/>We will flesh out these two approaches in detail and demonstrate that successful entrepreneurs often don\u2019t need to discover a world changing business model, they often just need great insight into the customer demand within a particular market.<h2><strong>The Internal Approach <\/strong><\/h2>Entrepreneurs often approach the task of coming up with a business idea by looking at their own background, skills, passion and abilities. In using this approach, the entrepreneur basically determines that they are the primary factor in whether they will succeed or fail in business.<h4><strong>Let your Skills and Experience Guide You <\/strong><\/h4>An entrepreneur\u2019s prior industry experience is the most common single source of entrepreneurs\u2019 business ideas. According to Inc. magazine, 43 percent of entrepreneurs reported they got their business idea from the experience they gained while working in the same industry or profession.<sup><sup>[2]<\/sup><\/sup><p>The experience the entrepreneurs gained from working within their industry helped them identify gaps in the industry\u2019s market or provided the knowledge and insight to create a company that could serve that industry\u2019s market more efficiently. \n               <\/p><p>Some entrepreneurs start businesses based upon their hobbies or avocational interests. They turned their passion for a particular activity or subject into a business venture. Some entrepreneurs start comic book or model train stores, small vineyards, or bed and breakfasts. Often these businesses are not started to reap considerable profits, but instead to pursue a lifestyle that brings joy and personal satisfaction to the entrepreneur. \n               <\/p><h4><strong>Entrepreneurship by Braille <\/strong><\/h4>Stephen Harper, in his excellent book<em>Extraordinary Entrepreneurship [link]<\/em>, argues that the hobby or personal interest approach to starting a business is \u201centrepreneurship by Braille.\u201d<sup><sup>[3]<\/sup><\/sup><p>This could also be starting a business solely based on where you live or where you want to live. Essentially, this is starting a business in search of customers. It is a very different proposition than creating a business to serve readily defined customers who already exist. If what you like is what the market objectively wants, then pursuing the business idea makes sense. \n               <\/p><p>But if the market for what you like, what interests you or where you live is too small or diffused to support your business idea, then creating that business is a fool\u2019s errand (unless you have money to burn). \n               <\/p><h4><strong>Developing \u201cOrganic\u201d Ideas<\/strong><\/h4>In his fantastic essay, \u201c<a href=\"http:\/\/www.paulgraham.com\/startupideas.html\">How to Generate Startup Ideas <\/a>,\u201d Paul Graham advises that entrepreneurs<em>should<\/em>conduct \u201centrepreneurship by braille\u201d \u2013 uncovering business ideas that fall within the experience and expertise of the entrepreneur.<sup><sup>[4]<\/sup><\/sup><p>He claims that entrepreneurs \u201cnotice\u201d startup ideas when they perceive solutions to urgent market problems to which adequate solutions do not currently exist. Graham says that noticing a market opportunity is essentially a passive process, and it is usually not achieved by actively trying to create a great business idea. He remarks, \u201cIf you look at the way successful founders have had their ideas, it's generally the result of some external stimulus hitting a prepared mind.\u201d \n               <\/p><p>Graham advises that entrepreneurs should focus on startup ideas in areas where they have expertise, first-hand experience (which ensures the problem actually exists) or a deep understanding of potential customer pain points. He claims that, \u201cthe best approach is more indirect: if you have the right sort of background, good startup ideas will seem obvious to you.\u201d He also states, \u201cideas\u2026 that grow naturally out of the founders' own experiences are organic startup ideas. The most successful startups almost all begin this way.\u201d \n               <\/p><h2><strong>The External Approach<\/strong><\/h2><br \/>The External Approach (or the \u201cOutside-In approach\u201d) to discovering viable business ideas looks first to the external market (verses the skills, knowledge and tastes of the entrepreneur) and tries to methodically discover market gaps that already exist.<sup><sup>[5]<\/sup><\/sup><p>The central idea of this approach \u2013 \n               <em>opportunity recognition<\/em> \u2013 is that a business can succeed only if it responds to, or creates, a need in the market. In plain English, the External Approach is about finding a gap in a market. A \u201cgap\u201d in the market is where the market, at present, does not contain a solution to a customer problem. \n               <\/p><p>These market gaps are also called \u201cwhite space.\u201d Market gaps are everywhere. They are in a constant state of flux, morphing and changing as technology, demographics, regulation, culture and global influences shift. \n               <\/p><p>It is up to entrepreneurs to determine how big these market opportunities are and why they exist. The \u201cwhite space\u201d opportunity may be fleeting \u2013 a very short-term prospect \u2013 or the white space market may not be big enough to sustain a profitable business. But then again, it may just be a defensible, exploitable gap in the market, overlooked and waiting to be exploited. \n               <\/p><p>Entrepreneurs must be able to scan their surrounding environment and recognize opportunities (market gaps). This requires the entrepreneur to develop what Harper calls an \u201copportunity antenna.\u201d \n               <\/p><p>Recognizing opportunities always seems easy and obvious in hindsight, but in the present \u2013 and that\u2019s what counts \u2013 it is a difficult and often infuriating process. \n               <\/p><p>A wildly successful example of recognizing a large, sustainable market gap was Chipotle. Chipotle filled a gap in the market between family style restaurants (good Mexican food with fresh ingredients at modest prices) but in a convenient, informal fast food venue. \n               <\/p><p><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/244-internal-1.png\" alt=\"image\" \/><br \/>The outside-in approach is generally better than the inside-out approach because it attempts to analyze market opportunities in an objective fashion. But the best approach is when it is possible to combine both approaches \u2013 the entrepreneur has a deep knowledge and expertise in an attractive market. \n               <\/p><p>We will discuss and analyze the following outside-in techniques for discovering market gaps: \n               <\/p><ul><li>Observing the Market <\/li>\n                <li>Focus Groups <\/li>\n                <li>Reverse Brainstorming <\/li>\n                <li>Market Growth <\/li>\n                <li>Matrix Charting for Insights <\/li>\n                <li>The \u201cSlice of Life\u201d Approach <\/li>\n                <li>The Market-Area Saturation Approach <\/li>\n                <li>The Competitive Matrix Approach <\/li>\n               <\/ul><br \/>None of these approaches is a silver bullet. Its best to use a combination of the methods described below. The more of these approaches that an entrepreneur uses to evaluate a business idea, the more likely accurate the entrepreneur\u2019s analysis is going to be.<h4><strong>Observing the Market <\/strong><\/h4>This technique is deceptively simple and requires an entrepreneur to experience a \u201cblinding flash of the obvious.\u201d<sup><sup>[6]<\/sup><\/sup><p>The potential entrepreneur should look around and listen to what people are complaining about. The entrepreneur should use their complaints and gripes to identify market gaps that need solutions. An entrepreneur could listen to people in public places, observe complaints on social media (Facebook, Twitter, LinkedIn\u2026 etc.) or traditional media and try to formulate possible solutions to the problems those complaints identify. \n               <\/p><h4><strong>Focus Groups <\/strong><\/h4>Focus groups are where an entrepreneur gets a sample of people together, lets say 5 to 20, and poses open-ended questions to them to uncover \u201cgaps\u201d in their personal and professional lives.<br \/>By providing a guided discussion, the entrepreneur can focus in on the participants\u2019 problems.<p>A list of problem areas can be recorded by the entrepreneur during the focus group. After the session, the entrepreneur can attempt to create solutions for the identified problems. The focus group does not have to be conducted face-to-face. An entrepreneur can use social media, such as Facebook, to pose open ended qualitative questions to a particular circle of friends or acquaintances. \n               <\/p><h4><strong>Reverse Brainstorming <\/strong><\/h4>Reverse brainstorming is where a focus group is asked to consider a particular problem with a product or service (such as smartphones) and identify possible solutions for it. The idea is to basically crowdsource an innovative idea by soliciting a group of people for a creative insight on how to fill in a gap in the market.<p>Once again, social media networks can be leveraged to solicit a group of friends or acquaintances. Questions that can be asked to help solve pain points are: \n               <\/p><p>1) Can it be made smaller, larger, better, or differently? \n               <br \/>2) Are there other users, uses or ways to increase the usage rate? \n               <br \/>3) Can it be combined with something else? \n               <br \/>4) Can it be done more quickly? \n               <br \/>5) Can it be made more convenient, available, attractive, quieter or last longer? \n               <\/p><h4><strong>Matrix Charting <\/strong><\/h4>Matrix charting is a way to systematically search for opportunities by listing important elements for each product or service along the left-hand vertical column of a grid and questions regarding each of these elements along the top horizontal row of the same grid.<sup><sup>[7]<\/sup><\/sup><p>The entrepreneur then tries to fill in the grid boxes with answers for each attribute question. An entrepreneur can use this technique by herself or with a focus group. Questions may include: \n               <\/p><ul><li>What can it be used for? <\/li>\n                <li>Where can it be used? <\/li>\n                <li>Who can use it? <\/li>\n                <li>When can it be used? <\/li>\n                <li>How can it be used?<\/li>\n               <\/ul><p><strong>Below is an example of a Matrix Chart for a Laundry Service: <\/strong>\n               <br \/><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/245-internal-2.png\" alt=\"image\" \/><\/p><p><strong>Industry Growth Rates<\/strong>\n               <br \/>At any given time, there are dozens of industries operating within new and underserved markets that are experiencing rapid growth. Some of these industries are growing at up to ten or twenty times faster than the overall economy. \n               <\/p><p>These industries are usually in their early growth phase and the companies within them are highly profitable. Furthermore, these industries often have low barriers to entry. This almost always represents a possible entrepreneurial opportunity. \n               <\/p><p>To find out which industries are experiencing rapid growth, its often useful to look at secondary research sources. Three great secondary resources to gauge industry growth are IBIS Industry Reports, Dunn &amp; Bradstreet and SageWorks [insert links for all three] \n               <\/p><p>For example, IBIS offers comprehensive market reports on over 700 industries. IBIS\u2019s media center on their website is a great resource to find out what industries or industry segments are experiencing rapid growth and high profitability. \n               <\/p><p>In 2018, IBIS identifies the following as five of the \n               <a href=\"https:\/\/www.ibisworld.com\/media\/2017\/12\/15\/life-in-the-fast-lane-industries-expected-to-rise-in-2018\/\">fastest growing industries <\/a>:\n               <\/p><p>Medical and Recreational Marijuana Retail Stores: 40.2% \n               <br \/>Medical and Recreational Marijuana Growers: 43.6% \n               <br \/>Wind Turbines: 48.8% \n               <br \/>Wireless Communications Carrier Industry: 8.6% \n               <br \/>Armored Vehicles: 17.4% \n               <\/p><p>SageWorks is a real-time database of private company financial information. It offers its information services to lending institutions to help those companies gauge lending risk to their borrowers. \n               <\/p><p>It too publishes a ranking of the fastest growing industries in the United States every year. The following are its forecast of the top 10 growing industries in 2016. \n               <\/p><p><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/246-internal-3.png\" alt=\"image\" \/><\/p><h4><strong>The \u201cSlice of Life\u201d Approach<\/strong><\/h4>Richard White, in<em>The Entrepreneur\u2019s Manual<\/em>, [insert link to resource] creating a decision tree that divides life by time and experiences exactly in half. For example, White recommends a person initially divide life between work and leisure.<p>If you choose work, then you could divide work between productive work time and unproductive work time; if you choose unproductive work time, you can then choose between unproductive work time at work before lunch or after lunch. \n               <\/p><p>Once a \u201cslice of life\u201d has been sufficiently broken down into a discrete slice, the object is to analyze that particular slice of life for problems people have during that particular slice of life. \n               <\/p><p><strong>The \u201cSlice of Life\u201d Decision Tree: <br \/><\/strong>\n               <img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/248-internal-4.png\" alt=\"image\" \/><br \/>There are potentially hundreds of discrete problems people have with this particular slice of life. For example, people could have trouble focusing on work because they are sleepy after lunch; they may not be productive because they do not have access to small amounts of downtime to let their brain rest; they may not be able to concentrate on work because they are hungry and need a snack; they may be unproductive in the afternoon because they are stiff and sore from sitting at a desk for hours by that point. \n               <\/p><p>There are hundreds of problems that exist within this slice of life. Life can be sliced up according to geography, time, activity, objects, experiences or people. \n               <\/p><p>After you identify the problems, begin searching for solutions. You can do this by coming up with solutions yourself or asking a focus group. For example, a possible solution for \u201ccrashing\u201d after lunch may be a simulative energy drink (like coffee, Red Bull, Five Hour Energy) that is natural and won\u2019t harm your body. \n               <\/p><p>Each of these problems offer scores of solutions, but they these problems may already be addressed by other firms in the market or the solutions may not be profitable. It is up to the entrepreneur to sift through these potential solutions and look for one that satisfies two conditions: \n               <br \/>1)the solution must be potentially profitable and \n               <br \/>2) the market currently does not have a solution addressing that problem (or at least one not addressing it well). \n               \n               <br \/><strong>Note:<\/strong> This technique is derived from a fantastic book by Stephen C. Harper called \n               <a href=\"https:\/\/www.amazon.com\/Extraordinary-Entrepreneurship-Professionals-Exceptional-Enterprise\/dp\/0470087277\"><em>Extraordinary Entrepreneurship <\/em><\/a>. I highly recommend this book if you want to explore this opportunity identification method further. Harper derived the Slice of Life approach from Richard White\u2019s classic book, \n               <a href=\"https:\/\/www.amazon.com\/Entrepreneurs-Manual-Start-Ups-Innovative-Management\/dp\/1626548714\/ref=sr_1_1?s=books&amp;ie=UTF8&amp;qid=1514155359&amp;sr=1-1&amp;keywords=The+Entrepreneur%E2%80%99s+Manual\"><em>The Entrepreneur\u2019s Manual <\/em><\/a>. \n               <br \/><\/p><h4><strong>The Market-Area Saturation Approach<\/strong><\/h4>The Market-Area Saturation Approach is used to find potential market gaps within a specific geographic area. The Market-Area Saturation Approach \u201cis based upon the premise that for a town with X number of people, there should be Y number of businesses of Z type.\u201d<sup><sup>[1]<\/sup><\/sup><p>Basically, entrepreneurs can analyze particular population centers and then try to identify if the current brick-and-mortar retail or service enterprises are underserving those markets. \n               <\/p><p>This could be determined by taking the number of specific establishments (say pizza parlors) in your state and divide it by your state\u2019s population. Then, multiply that fraction by the population of the geographic area you are analyzing (say 100,000 people). \n               <\/p><p>That product would be the average number of establishments, in your state, that a particular geographic area could support. \n               <\/p><p><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/249-internal-5.png\" alt=\"image\" \/><br \/>For example, a certain town (or a collection of towns within a given geographic radius) you are analyzing may have roughly 100,000 people. \n               <\/p><p>After conducting some research, you discover that in your state, you may roughly expect that a population of 100,000 people would contain, on average, 11.3 pizza restaurants, 2.0 florist shops, 1.6 tanning salons, 4.0 gyms and 1.0 bowling alley. A great resource for conducting this research is \n               <a href=\"http:\/\/simplyanalytics.com\/\">SimplyAnalytics<\/a>. Another resource is the Business &amp; Industry section of the U.S. Census Bureau at \n               <a href=\"http:\/\/www.census.gov\/econ\/index.html\">http:\/\/www.census.gov\/econ\/index.html<\/a>. \n               <\/p><p>After completing this task, you would then survey a specific geographic area with a similar population profile (100,000 for instance) and count the number of establishments serving customers within this area. \n               <\/p><p>A resource you could use for this is \n               <a href=\"https:\/\/www.google.com\/maps\">Google Maps <\/a>. Your search may turn up that your target geographic area contains nine pizza parlors, three florist shops, two tanning salons, one gym and one bowling alley. \n               <\/p><p><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/250-internal-6.png\" alt=\"image\" \/><br \/>According to this analysis, it would appear that the particular geographic area and population you analyzed has significant unmet demand for gyms and modestly unmet demand for both pizza parlors and tanning salons. \n               <\/p><p>Your analysis also determined that bowling alley demand is currently being met while there is probably an oversupply of florist shops. Based upon these findings, it would probably be a good bet that there is enough demand within this particular geography and population to justify opening a gym. \n               <\/p><p>The market-area saturation approach should be seen as a rule of thumb, not a precise scientific instrument. The size, location, competitive strength and qualitative factors of each establishment are not considered with this analysis. But this broad brushstroke approach may find potential business opportunities that deserve closer scrutiny. \n               <\/p><h4><strong>Create a Competitive Matrix:<\/strong><\/h4>If a prospective entrepreneur has already decided what general product or service she wants to sell within a niche market space (geographic and otherwise), then the Competitive Matrix may help identify potential gaps in the market.<p>The Competitive Matrix is an attempt to quantify the existence of a potential market gap, but the value of that quantification relies on the quality of the entrepreneur\u2019s analysis of the market. \n               <\/p><p>An entrepreneur should divide the market that they want to enter into corresponding market segments. These market segments can be characterized by geography, demographic and psychographic characteristics, and behavioral patterns. \n               <\/p><p>For each of these market segments, the entrepreneur will create a matrix. The top horizontal row will identify the current businesses that currently provide value proposition(s) (goods and services the entrepreneur seeks to provide) to the market segment. \n               <\/p><p>The left vertical column of the matrix will list key factors important to the business\u2019 value proposition(s). A blank customer segment competitive matrix example is shown below. \n               <\/p><p><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/251-internal-7.png\" alt=\"image\" \/><br \/>The entrepreneur will need to conduct first hand research (by using polling research and\/or focus groups of the market segment\u2019s customers) to determine: \n               <\/p><p>1. \n               <strong>Competitive Factor Importance <\/strong> \u2013 The level of importance that members of a market segment place on each competitive factor (value proposition). This number is measured as a whole number from 1 to 5, with 1 being the worst and 5 being the best. This number is placed within the upper left-hand box in each Competitive Factor\/Business # cell. \n               <\/p><p>2. \n               <strong>Competitive Factor Satisfaction <\/strong> \u2013 How well each current business meets its customers\u2019 needs for each competitive factor (value proposition) to the customers within a particular market segment. This number is measured as a whole number from 1 to 5, with 1 being the worst and 5 being the best. This number is placed within the upper right-hand box in each Competitive Factor\/Business # cell. \n               <\/p><p>A \n               <strong><em>market gap <\/em><\/strong> exists within a market segment when existing businesses that cater to it are not supplying the type or quantity of value that the market segment demands. In other words, market gaps exist where a business is underserving a competitive factor to their customers relative to the level of importance customers in that market segment place on that competitive factor. \n               <\/p><p>For example, if quality is important to a particular market segment (5), but the current businesses catering to that market segment are providing goods with mediocre quality (3), there is then a gap in the market related to this competitive factor. \n               <\/p><p>For each competitive factor, gaps of only one point of separation between what the segment considers important but existing businesses do not provide the level of value demanded by the market segment indicates a small gap in the market. But existing businesses can usually adapt to cover this gap without much effort. \n               <\/p><p>For each competitive factor, gaps of at least 2 points of separation between what the segment considers important but existing businesses do not provide that level of value to the segment will require those businesses to make significant changes in the existing business\u2019 strategy. This is a market gap ripe for exploitation. \n               <\/p><p><strong>Items of Note: <\/strong>\n               <br \/><\/p><ul><li>If the difference between the Competitive Factor Importance and the Competitive Factor Satisfaction is a gap of one point, it will be signified within each bottom right box of its competitive matrix cell by a *. <\/li>\n                <li>If the difference between the Competitive Factor Importance and the Competitive Factor Satisfaction is a gap of two or three points, it will be signified within each bottom right box of its competitive matrix cell with either ** or ***. <\/li>\n                <li>The combined score for each competitive factor is the product of the business\u2019s Competitive Factor Importance and the business\u2019s Competitive Factor Satisfaction. For example: 3 x 4 = 12. The number signifies how well each current business satisfies the customer segment for a particular competitive factor. <\/li>\n                <li>The <strong>Business\u2019 Relative Strength for the Customer Segment <\/strong> row (at the bottom of the matrix) is an aggregate approximation of how well each current business serves the market segment. <\/li>\n                <li>Gaps of two points or greater represent significant market opportunities for market entrants. <\/li>\n                <li>If a whole row has asterisks, then none of the current businesses is meeting the segment\u2019s expectations. <\/li>\n               <\/ul><br \/><strong>Example: Casual Dining Market <\/strong><p><strong>Market Segment \u2013<\/strong> University Faculty and Staff, Establishments within \u00bd mile of an urban university \n               <\/p><p><strong>Survey and\/or Focus Group Results<\/strong>\n               <br \/><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/252-internal-8.png\" alt=\"image\" \/><br \/><img src=\"https:\/\/capitalandgrowth.org\/storage\/attachments\/254-internal-9.png\" alt=\"image\" \/><br \/><strong>A Note of Caution in using the Competitive Matrix Approach <\/strong>\n               <br \/><\/p><ul><li>This method may take considerable time, money and market research to complete. If an entrepreneur uses this approach, he or she should confine the analysis to a few select market segments within a market. <\/li>\n                <li>The Competitive Market Approach can be a powerful tool, but its effectiveness is solely dependent on the quality of data collected from the customer segment\u2019s being analyzed. If the quality of the collected data is low, the resulting analysis could be wildly off the mark. <\/li>\n               <\/ul><br \/><strong>Why Red Oceans Matter <\/strong><br \/>In today\u2019s tech-centered vision of entrepreneurship, where everyone is intent on inventing and proving new business models and discovering new markets, its important to remember and carefully consider the proven value of Red Ocean ideas.<p>For every tech unicorn tech story like Uber or Airbnb, there are a hundred Red Ocean success stories (like Papa Johns, Peets Coffee or Motel 6). While these Red Ocean ideas may not (and probably won\u2019t) result in creating multibillion dollar companies within a ten-year period of time, they do offer entrepreneurs the opportunity to create excellent, successful businesses that can lead to riches and personal autonomy. \n               <\/p><p>[1] W. Chan Kim and Ren\u00e9e Mauborgne, \n               <em>Blue ocean Strategy: How to Create Uncontested Market Space and Make the Competition Irrelevant<\/em>, Harvard Business School Press, pg. 4, 2005.\n               <\/p><p>[2] John Case, \u201cThe Origins of Entrepreneurship,\u201d Inc. June, 1989, 51-62.\n               <\/p><p>[3] Stephen C. Harper, Extraordinary Entrepreneurship, John Wiley &amp; Sons, Inc., Hoboken, New Jersey, 2005, pg. 68. This is only pursuing opportunities that are within reach \u2013 namely what you know and like.\n               <\/p><p>[4] \n               <a href=\"http:\/\/www.paulgraham.com\/startupideas.html#f3n\">http:\/\/www.paulgraham.com\/startupideas.html#f3n<\/a> Graham counsels that the best way entrepreneurs can develop startup ideas is organically by \u201cnoticing\u201d market opportunities.\n               <\/p><p>[5] Peggy Lambing and Charles R. Kuehl, Entrepreneurship, Prentice-Hall, Inc., Upper Saddle River, New Jersey, 2000, pg. 92.\n               <\/p><p>[6] Stephen C. Harper, Extraordinary Entrepreneurship, John Wiley &amp; Sons, Inc., Hoboken, New Jersey, 2005, pg. 77.\n               <\/p><p>[7] Stephen C. Harper, Extraordinary Entrepreneurship, John Wiley &amp; Sons, Inc., Hoboken, New Jersey, 2005, pg. 80.\n               <br \/><\/p><hr \/>","created_at":"2018-11-07T16:45:21+0000","updated_at":"2018-11-07T16:45:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":20,"domain_name":"capitalandgrowth.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":347,"title":"How Generative Music Works","url":"http:\/\/teropa.info\/loop\/","content":"<div id=\"title\" class=\"step\" data-x=\"-500\" data-y=\"-900\"><h2>A Perspective<\/h2><footer>\n              Use space bar to advance\n              @teropa\n            <\/footer><\/div><p><\/p><h2>\n            This presentation is about making music<br \/>by designing systems that make music.\n          <\/h2><div id=\"systemdefinition\" class=\"step centered-content future\" data-x=\"-570\" data-y=\"80\" data-scale=\"0.2\"><div class=\"card\"><h3><em>System<\/em><\/h3><ol><li>A set of things working together as parts of a mechanism or an interconnecting network; a complex whole.<\/li>\n              <li class=\"scheme\">A set of principles or procedures according to which something is done; an organized scheme or method.<\/li>\n            <\/ol><em>Oxford English Dictionary<\/em><\/div><\/div><div id=\"itsgonnarain\" class=\"step centered-content future\" data-x=\"-2600\" data-y=\"-200\" data-scale=\"1\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              For his 1965 composition \u201cIt's Gonna Rain\u201d, Steve Reich\n              designed a generative mechanical system.\n            <\/p><p class=\"state state-init mode-present\">\n              Steve Reich: It's Gonna Rain (1965)\n            <\/p><\/div><footer class=\"clip-src\">\n              Clip from a <a href=\"https:\/\/www.youtube.com\/watch?v=yg7kKJsFpvc\" target=\"_blank\">Steve Reich interview on q on CBC<\/a>.\n          <\/footer><\/div><div id=\"reichphasing\" class=\"step centered-content future\" data-x=\"-3900\" data-y=\"25\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"mode-read\">Once Reich had discovered this generative process in tape, he applied it to other media.<\/p><p class=\"mode-present\">Reich's phasing process in other media<\/p><\/div><div class=\"videos\"><figure class=\"pendulum\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Steve Reich: Pendulum Music. <br \/><a href=\"https:\/\/www.youtube.com\/watch?v=fU6qDeJPT-w\" target=\"_blank\">Interpretation by Joan Cerver\u00f3, V\u00edctor Trescol\u00ed, Isabel Le\u00f3n, Estefan\u00eda S\u00e1nchez<\/a>.<\/figcaption><\/figure><figure class=\"pianophase\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Steve Reich: Piano Phase. <br \/><a href=\"http:\/\/www.chenalexander.com\/Piano-Phase\" target=\"_blank\">Visualization by Alexander Chen<\/a>.<\/figcaption><\/figure><\/div><\/div><div id=\"airports\" class=\"step future\" data-x=\"-2600\" data-y=\"800\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Brian Eno was inspired by \u201cIt's Gonna Rain\u201d, and used the same phasing process on \u201cMusic for Airports\u201d in 1978.\n            <\/p><p class=\"state state-init mode-present\">\n              Brian Eno: Music for Airports (1978)\n            <\/p><\/div><\/div><div id=\"inc\" class=\"step centered-content future\" data-x=\"-2600\" data-y=\"1700\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Terry Riley\u2019s \u201cIn C\u201d from 1964 is a composition expressed as a set of principles and procedures.\n            <\/p><p class=\"state state-init mode-present\">\n              Terry Riley: In C (1964)\n            <\/p><\/div><div class=\"vis\"><img class=\"score\" src=\"src\" alt=\"image\" \/><img class=\"instructions\" src=\"src\" alt=\"image\" \/><\/div><\/div><div id=\"incplayer\" class=\"step centered-content future\" data-x=\"-2600\" data-y=\"2600\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              With this simple blueprint you get a system that generates a new version every time it's played. <br \/>No two performances will ever be the same.\n            <\/p><p class=\"state state-init mode-present\">\n              A new realization of the score generated every time\n            <\/p><\/div><footer>\n            See also: <a href=\"https:\/\/teropa.info\/blog\/2017\/01\/23\/terry-rileys-in-c.html\" target=\"_blank\">Terry Riley's In C: A Journey Through a Musical Possibility Space<\/a>\n          <\/footer><\/div><div id=\"methodvsproduct\" class=\"step transient future\" data-x=\"-4600\" data-y=\"1200\" data-scale=\"4.5\"><div class=\"half\"><h2>Generative Method<\/h2><p class=\"mode-read\">\n              \u201cIt's Gonna Rain\u201d and \"Music for Airports\" were made using generative <em>methods<\/em>.\n            <\/p><p class=\"mode-read\">\n              But the results don't retain that generative nature. Every time you listen to these pieces, you hear the same thing.\n            <\/p><\/div><div class=\"half\"><h2>Generative Product<\/h2><p class=\"mode-read\">\n              \u201cIn C\u201c, by contrast, is a generative <em>product<\/em>. It was not just generated once by the composer, but is generated anew every time.\n            <\/p><p class=\"mode-read\">\n              For \u201cIn C\u201d the system itself is shared, not just its outputs. Anyone can use the score to generate a new version. That happens somewhere in the world almost every day.\n            <\/p><\/div><\/div><div id=\"reflection\" class=\"step future\" data-x=\"-500\" data-y=\"1500\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Eno's latest album, Reflection, is both:<br \/>A generative product made using generative methods.\n            <\/p><p class=\"state state-init mode-present\">\n              Brian Eno: Reflection (2017)<br \/>Both generative method and product\n            <\/p><\/div><footer><a href=\"http:\/\/www.generativemusic.com\/reflection.html\" target=\"_blank\">Brian Eno: Reflection<\/a>\n          <\/footer><\/div><div id=\"softwareoverview\" class=\"step transient\" data-x=\"0\" data-y=\"1300\" data-scale=\"4.5\"><p class=\"mode-read\">\n            \u201cReflection\u201d is a generative <em><strong>software system<\/strong><\/em>. That makes it easy to deliver to listeners.\n            <br \/>Much easier than reel-to-reel tape recorders.\n          <\/p><p class=\"mode-read\">\n            This presentation is another example of that convenience: Several generative software systems are embedded right onto this webpage.\n          <\/p><p class=\"mode-present\">\n            Software Systems\n          <\/p><\/div><div id=\"randomness\" class=\"step centered-content\" data-x=\"-1000\" data-y=\"2200\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Apart from the ease of delivery, software systems have a few other benefits.\n            <\/p><p class=\"state state-init mode-present\">\n              Other Benefits of Using Software\n            <\/p><\/div><\/div><div id=\"stochasticdrummachine\" class=\"step centered-content future\" data-x=\"-1000\" data-y=\"3000\"><div class=\"vis\"><div class=\"container\"><div class=\"controls\"><label class=\"control\">\n                    Deviations                    \n                  <\/label><\/div><p>Loading...<\/p><\/div><\/div><footer><a href=\"https:\/\/codepen.io\/teropa\/pen\/PKoYXM\" target=\"_blank\">Full version &amp; source code<\/a>\n          <\/footer><\/div><div id=\"cage\" class=\"step centered-content future\" data-x=\"100\" data-y=\"2800\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              John Cage was a pioneer in the use of randomness in music composition.\n            <\/p><p class=\"state state-init mode-present\">\n              John Cage's<br \/>Chance Procedures\n            <\/p><p>For example, <em>\u201cwhat duration should this note have?\u201d<\/em>\n            <\/p><p>This is also what the drum machine does. We ask <em>\"what drum pattern should I use?\"<\/em> and the system randomly selects an answer.\n              \n            <\/p><\/div><\/div><div id=\"algorithms\" class=\"step centered-content future\" data-x=\"1300\" data-y=\"2000\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Another thing software is good at is <em>algorithms<\/em>.\n            <\/p><p class=\"state state-init mode-present\">\n              Algorithms\n            <\/p><\/div><\/div><div id=\"generativegrammars\" class=\"step centered-content future\" data-x=\"1300\" data-y=\"2750\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              <em>Generative grammars<\/em> represent an algorithmic approach to music generation.\n            <\/p><p class=\"state state-init mode-present\">\n              Generative Grammars\n            <\/p><\/div><div class=\"vis\"><img src=\"src\" alt=\"image\" \/><\/div><\/div><div id=\"additiverhythmgrammar\" class=\"step centered-content future\" data-x=\"3400\" data-y=\"2550\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Let's define a generative grammar for <em>additive rhythm<\/em>.\n            <\/p><p class=\"state state-init mode-present\">\n              An additive rhythm grammar\n            <\/p><p>Now we have hierarchically defined what a \u201ccell\u201d\n                means in this language, all the way down to the atomic level.\n              \n            <\/p><p>This definition is <em>recursive<\/em>, and makes it possible to express multi-cell patterns.\n              \n            <\/p><\/div><\/div><div id=\"polyloops\" class=\"step centered-content future\" data-x=\"2550\" data-y=\"2750\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Once you have a grammar, you can generate rhythm patterns by expanding it.\n            <\/p><p class=\"state state-init mode-present\">\n              Generating patterns from the grammar\n            <\/p><\/div><footer><a href=\"https:\/\/codepen.io\/teropa\/pen\/rzQMEG\" target=\"_blank\">Full version &amp; source code<\/a>\n          <\/footer><\/div><div id=\"connected\" class=\"step centered-content future\" data-x=\"3900\" data-y=\"1750\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Software also makes it easy to connect systems to the external world.\n            <\/p><p class=\"state state-init mode-present\">\n              Connected Software\n            <\/p><\/div><\/div><div id=\"tramsofhelsinki\" class=\"step centered-content future\" data-x=\"1800\" data-y=\"1150\" data-scale=\"0.75\"><p class=\"state state-init mode-read\">\n            This system generates music from the Helsinki tram system.\n          <\/p><p class=\"state state-init mode-present\">\n            Generating music from the Helsinki tram system\n          <\/p><\/div><div id=\"tramszoomout\" class=\"step centered-content future\" data-x=\"1800\" data-y=\"1050\" data-scale=\"1.1\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              The trams are playing a pentatonic scale, mapped to central Helsinki.\n            <\/p><p class=\"state state-init mode-present\">\n              A pentatonic scale mapped to central Helsinki\n            <\/p><\/div><div class=\"the-rest\"><ul class=\"scale\"><li class=\"note\">E4<\/li>\n              <li class=\"note\">D4<\/li>\n              <li class=\"note\">C4<\/li>\n              <li class=\"note\">A3<\/li>\n              <li class=\"note\">G3<\/li>\n              <li class=\"note\">E3<\/li>\n              <li class=\"note\">D3<\/li>\n              <li class=\"note\">C3<\/li>\n              <li class=\"note\">A2<\/li>\n              <li class=\"note\">G2<\/li>\n            <\/ul><\/div><footer><a href=\"https:\/\/codepen.io\/teropa\/pen\/mBbPEe\" target=\"_blank\">Full version &amp; source code<\/a>\n          <\/footer><\/div><div id=\"datasonification\" class=\"step centered-content future\" data-x=\"1400\" data-y=\"350\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"mode-read\">This was an example of <em>data sonification<\/em>:<br \/>Generative musical systems riding on events or data produced by other systems.<\/p><p class=\"mode-present\">Data Sonification<\/p><\/div><div class=\"videos\"><figure class=\"atlas\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Brian Foo: <a href=\"http:\/\/eclipticalis.com\/\" target=\"_blank\">Music Eclipticalis<\/a>. Based on \u201cAtlas Eclipticalis\u201d by John Cage.<\/figcaption><\/figure><figure class=\"wiki\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Hatnote, Stephen LaPorte, and Mahmoud Hashemi: <a href=\"http:\/\/listen.hatnote.com\/\" target=\"_blank\">Listen to Wikipedia<\/a>.<\/figcaption><\/figure><\/div><\/div><div id=\"hooman\" class=\"step centered-content future\" data-x=\"3600\" data-y=\"900\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Speaking of connecting musical systems to other systems, how about the most interesting systems of them all: Human beings?\n            <\/p><p class=\"state state-init mode-present\">\n              The most interesting systems in the world: Human beings\n            <\/p><\/div><div class=\"vis\"><img src=\"src\" alt=\"image\" \/><footer>\n              Image: <a href=\"https:\/\/www.flickr.com\/photos\/bionicteaching\/14092973648\" target=\"_blank\">Tom Woodward<\/a>\n            <\/footer><\/div><\/div><div id=\"trope\" class=\"step centered-content future\" data-x=\"4350\" data-y=\"100\" data-scale=\"0.5\"><img src=\"src\" alt=\"image\" \/><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Eno's \u201cTrope\u201d is an example of a generative system that's also interactive.\n            <\/p><p class=\"state state-init mode-present\">\n              Brian Eno &amp; Peter Chilvers: Trope (2009)\n            <\/p><\/div><footer><a href=\"http:\/\/www.generativemusic.com\/trope.html\" target=\"_blank\">Trope<\/a> by Brian Eno and Peter Chilvers<\/footer><\/div><div id=\"musicmouse\" class=\"step centered-content future\" data-x=\"2800\" data-y=\"-100\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              These interactive generative systems can be designed to give the user varying degrees of freedom.\n            <\/p><p class=\"state state-init mode-present\">\n              Degrees of freedom for the user vary\n            <\/p><p>The result is a <em>collaboration<\/em> between you and Music Mouse.\n              \n            <\/p><\/div><div class=\"vis\"><footer>\n              A partial emulation of\n              <a href=\"http:\/\/retiary.org\/ls\/index.html\" target=\"_blank\">Laurie Spiegel<\/a>'s\n              <a href=\"http:\/\/retiary.org\/ls\/programs.html\" target=\"_blank\">\u201cMusic Mouse - An Intelligent Instrument\u201d<\/a>.\n              <br \/>Use 'z' and 'x' to transpose, 'c' to toggle generative pattern. <a href=\"https:\/\/teropa.info\/musicmouse\" target=\"_blank\">More features here<\/a>.\n            <\/footer><\/div><\/div><div id=\"models\" class=\"step centered-content future\" data-x=\"4400\" data-y=\"-1000\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Another form of human-machine interaction is <em>teaching<\/em> machines how to do things.\n            <\/p><p class=\"state state-init mode-present\">\n              Teaching a machine to do things\n            <\/p><\/div><div class=\"vis\"><img src=\"src\" alt=\"image\" \/><\/div><\/div><div id=\"markovplayer\" class=\"step centered-content future\" data-x=\"2800\" data-y=\"-1000\"><div class=\"vis\"><div class=\"controls\"><footer>\n                Play with mouse, keyboard, or MIDI keyboard (Chrome only)\n                \n                <br \/>See also: <a href=\"https:\/\/codepen.io\/teropa\/pen\/bRqYVj\" target=\"_blank\">Dancing Markov Gymnop\u00e9dies<\/a>\n              <\/footer><\/div><\/div><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              <em>Markov chains<\/em> are an old, but still effective, example of the machine learning approach.\n            <\/p><p class=\"state state-init mode-present\">\n              Markov Chains\n            <\/p><\/div><\/div><div id=\"deeplearning\" class=\"step centered-content future\" data-x=\"1300\" data-y=\"-850\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Markov chains are an ancestor of contemporary machine learning approaches to music generation, such as deep neural networks.\n            <\/p><p class=\"state state-init mode-present\">\n              Contemporary machine learning: Deep neural networks\n            <\/p><\/div><div class=\"vis\"><a href=\"https:\/\/magenta.tensorflow.org\/performance-rnn-browser\" target=\"_blank\"><img class=\"performancernn\" src=\"src\" alt=\"image\" \/><\/a><a href=\"https:\/\/magenta.tensorflow.org\/nsynth\" target=\"_blank\"><img class=\"nsynth\" src=\"src\" alt=\"image\" \/><\/a><footer>\n              See <a href=\"https:\/\/medium.com\/artists-and-machine-intelligence\/neural-nets-for-generating-music-f46dffac21c0\" target=\"_blank\">Neural Nets for Generating Music<\/a> by Kyle McDonald.\n            <\/footer><\/div><\/div><div id=\"overview\" class=\"step centered-content future\" data-x=\"250\" data-y=\"1400\" data-scale=\"6\"><h2>Systems<\/h2><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              This is what the generative music landscape looks like to me.\n            <\/p><p class=\"state state-init mode-present\">\n              The generative music landscape\n            <\/p><\/div><\/div>","created_at":"2018-11-07T16:45:21+0000","updated_at":"2018-11-07T16:45:21+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"teropa.info","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":348,"title":"How Generative Music Works","url":"http:\/\/teropa.info\/loop\/","content":"<div class=\"step\" data-x=\"-500\" data-y=\"-900\"><h2>A Perspective<\/h2><footer>\n              Use space bar to advance\n              @teropa\n            <\/footer><\/div><p><\/p><h2>\n            This presentation is about making music<br \/>by designing systems that make music.\n          <\/h2><div class=\"step centered-content future\" data-x=\"-570\" data-y=\"80\" data-scale=\"0.2\"><div class=\"card\"><h3><em>System<\/em><\/h3><ol><li>A set of things working together as parts of a mechanism or an interconnecting network; a complex whole.<\/li>\n              <li class=\"scheme\">A set of principles or procedures according to which something is done; an organized scheme or method.<\/li>\n            <\/ol><em>Oxford English Dictionary<\/em><\/div><\/div><div class=\"step centered-content future\" data-x=\"-2600\" data-y=\"-200\" data-scale=\"1\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              For his 1965 composition \u201cIt's Gonna Rain\u201d, Steve Reich\n              designed a generative mechanical system.\n            <\/p><p class=\"state state-init mode-present\">\n              Steve Reich: It's Gonna Rain (1965)\n            <\/p><\/div><footer class=\"clip-src\">\n              Clip from a <a href=\"https:\/\/www.youtube.com\/watch?v=yg7kKJsFpvc\" target=\"_blank\">Steve Reich interview on q on CBC<\/a>.\n          <\/footer><\/div><div class=\"step centered-content future\" data-x=\"-3900\" data-y=\"25\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"mode-read\">Once Reich had discovered this generative process in tape, he applied it to other media.<\/p><p class=\"mode-present\">Reich's phasing process in other media<\/p><\/div><div class=\"videos\"><figure class=\"pendulum\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Steve Reich: Pendulum Music. <br \/><a href=\"https:\/\/www.youtube.com\/watch?v=fU6qDeJPT-w\" target=\"_blank\">Interpretation by Joan Cerver\u00f3, V\u00edctor Trescol\u00ed, Isabel Le\u00f3n, Estefan\u00eda S\u00e1nchez<\/a>.<\/figcaption><\/figure><figure class=\"pianophase\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Steve Reich: Piano Phase. <br \/><a href=\"http:\/\/www.chenalexander.com\/Piano-Phase\" target=\"_blank\">Visualization by Alexander Chen<\/a>.<\/figcaption><\/figure><\/div><\/div><div class=\"step future\" data-x=\"-2600\" data-y=\"800\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Brian Eno was inspired by \u201cIt's Gonna Rain\u201d, and used the same phasing process on \u201cMusic for Airports\u201d in 1978.\n            <\/p><p class=\"state state-init mode-present\">\n              Brian Eno: Music for Airports (1978)\n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"-2600\" data-y=\"1700\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Terry Riley\u2019s \u201cIn C\u201d from 1964 is a composition expressed as a set of principles and procedures.\n            <\/p><p class=\"state state-init mode-present\">\n              Terry Riley: In C (1964)\n            <\/p><\/div><div class=\"vis\"><img class=\"score\" src=\"src\" alt=\"image\" \/><img class=\"instructions\" src=\"src\" alt=\"image\" \/><\/div><\/div><div class=\"step centered-content future\" data-x=\"-2600\" data-y=\"2600\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              With this simple blueprint you get a system that generates a new version every time it's played. <br \/>No two performances will ever be the same.\n            <\/p><p class=\"state state-init mode-present\">\n              A new realization of the score generated every time\n            <\/p><\/div><footer>\n            See also: <a href=\"https:\/\/teropa.info\/blog\/2017\/01\/23\/terry-rileys-in-c.html\" target=\"_blank\">Terry Riley's In C: A Journey Through a Musical Possibility Space<\/a>\n          <\/footer><\/div><div class=\"step transient future\" data-x=\"-4600\" data-y=\"1200\" data-scale=\"4.5\"><div class=\"half\"><h2>Generative Method<\/h2><p class=\"mode-read\">\n              \u201cIt's Gonna Rain\u201d and \"Music for Airports\" were made using generative <em>methods<\/em>.\n            <\/p><p class=\"mode-read\">\n              But the results don't retain that generative nature. Every time you listen to these pieces, you hear the same thing.\n            <\/p><\/div><div class=\"half\"><h2>Generative Product<\/h2><p class=\"mode-read\">\n              \u201cIn C\u201c, by contrast, is a generative <em>product<\/em>. It was not just generated once by the composer, but is generated anew every time.\n            <\/p><p class=\"mode-read\">\n              For \u201cIn C\u201d the system itself is shared, not just its outputs. Anyone can use the score to generate a new version. That happens somewhere in the world almost every day.\n            <\/p><\/div><\/div><div class=\"step future\" data-x=\"-500\" data-y=\"1500\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Eno's latest album, Reflection, is both:<br \/>A generative product made using generative methods.\n            <\/p><p class=\"state state-init mode-present\">\n              Brian Eno: Reflection (2017)<br \/>Both generative method and product\n            <\/p><\/div><footer><a href=\"http:\/\/www.generativemusic.com\/reflection.html\" target=\"_blank\">Brian Eno: Reflection<\/a>\n          <\/footer><\/div><div class=\"step transient\" data-x=\"0\" data-y=\"1300\" data-scale=\"4.5\"><p class=\"mode-read\">\n            \u201cReflection\u201d is a generative <em><strong>software system<\/strong><\/em>. That makes it easy to deliver to listeners.\n            <br \/>Much easier than reel-to-reel tape recorders.\n          <\/p><p class=\"mode-read\">\n            This presentation is another example of that convenience: Several generative software systems are embedded right onto this webpage.\n          <\/p><p class=\"mode-present\">\n            Software Systems\n          <\/p><\/div><div class=\"step centered-content\" data-x=\"-1000\" data-y=\"2200\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Apart from the ease of delivery, software systems have a few other benefits.\n            <\/p><p class=\"state state-init mode-present\">\n              Other Benefits of Using Software\n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"-1000\" data-y=\"3000\"><div class=\"vis\"><div class=\"container\"><div class=\"controls\"><label class=\"control\">\n                    Deviations                    \n                  <\/label><\/div><p>Loading...<\/p><\/div><\/div><footer><a href=\"https:\/\/codepen.io\/teropa\/pen\/PKoYXM\" target=\"_blank\">Full version &amp; source code<\/a>\n          <\/footer><\/div><div class=\"step centered-content future\" data-x=\"100\" data-y=\"2800\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              John Cage was a pioneer in the use of randomness in music composition.\n            <\/p><p class=\"state state-init mode-present\">\n              John Cage's<br \/>Chance Procedures\n            <\/p><p>For example, <em>\u201cwhat duration should this note have?\u201d<\/em>\n            <\/p><p>This is also what the drum machine does. We ask <em>\"what drum pattern should I use?\"<\/em> and the system randomly selects an answer.\n              \n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"1300\" data-y=\"2000\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Another thing software is good at is <em>algorithms<\/em>.\n            <\/p><p class=\"state state-init mode-present\">\n              Algorithms\n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"1300\" data-y=\"2750\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              <em>Generative grammars<\/em> represent an algorithmic approach to music generation.\n            <\/p><p class=\"state state-init mode-present\">\n              Generative Grammars\n            <\/p><\/div><div class=\"vis\"><img src=\"src\" alt=\"image\" \/><\/div><\/div><div class=\"step centered-content future\" data-x=\"3400\" data-y=\"2550\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Let's define a generative grammar for <em>additive rhythm<\/em>.\n            <\/p><p class=\"state state-init mode-present\">\n              An additive rhythm grammar\n            <\/p><p>Now we have hierarchically defined what a \u201ccell\u201d\n                means in this language, all the way down to the atomic level.\n              \n            <\/p><p>This definition is <em>recursive<\/em>, and makes it possible to express multi-cell patterns.\n              \n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"2550\" data-y=\"2750\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Once you have a grammar, you can generate rhythm patterns by expanding it.\n            <\/p><p class=\"state state-init mode-present\">\n              Generating patterns from the grammar\n            <\/p><\/div><footer><a href=\"https:\/\/codepen.io\/teropa\/pen\/rzQMEG\" target=\"_blank\">Full version &amp; source code<\/a>\n          <\/footer><\/div><div class=\"step centered-content future\" data-x=\"3900\" data-y=\"1750\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Software also makes it easy to connect systems to the external world.\n            <\/p><p class=\"state state-init mode-present\">\n              Connected Software\n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"1800\" data-y=\"1150\" data-scale=\"0.75\"><p class=\"state state-init mode-read\">\n            This system generates music from the Helsinki tram system.\n          <\/p><p class=\"state state-init mode-present\">\n            Generating music from the Helsinki tram system\n          <\/p><\/div><div class=\"step centered-content future\" data-x=\"1800\" data-y=\"1050\" data-scale=\"1.1\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              The trams are playing a pentatonic scale, mapped to central Helsinki.\n            <\/p><p class=\"state state-init mode-present\">\n              A pentatonic scale mapped to central Helsinki\n            <\/p><\/div><div class=\"the-rest\"><ul class=\"scale\"><li class=\"note\">E4<\/li>\n              <li class=\"note\">D4<\/li>\n              <li class=\"note\">C4<\/li>\n              <li class=\"note\">A3<\/li>\n              <li class=\"note\">G3<\/li>\n              <li class=\"note\">E3<\/li>\n              <li class=\"note\">D3<\/li>\n              <li class=\"note\">C3<\/li>\n              <li class=\"note\">A2<\/li>\n              <li class=\"note\">G2<\/li>\n            <\/ul><\/div><footer><a href=\"https:\/\/codepen.io\/teropa\/pen\/mBbPEe\" target=\"_blank\">Full version &amp; source code<\/a>\n          <\/footer><\/div><div class=\"step centered-content future\" data-x=\"1400\" data-y=\"350\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"mode-read\">This was an example of <em>data sonification<\/em>:<br \/>Generative musical systems riding on events or data produced by other systems.<\/p><p class=\"mode-present\">Data Sonification<\/p><\/div><div class=\"videos\"><figure class=\"atlas\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Brian Foo: <a href=\"http:\/\/eclipticalis.com\/\" target=\"_blank\">Music Eclipticalis<\/a>. Based on \u201cAtlas Eclipticalis\u201d by John Cage.<\/figcaption><\/figure><figure class=\"wiki\"><div class=\"vidcontainer\"><img src=\"src\" alt=\"image\" \/><\/div>\n              <figcaption>Hatnote, Stephen LaPorte, and Mahmoud Hashemi: <a href=\"http:\/\/listen.hatnote.com\/\" target=\"_blank\">Listen to Wikipedia<\/a>.<\/figcaption><\/figure><\/div><\/div><div class=\"step centered-content future\" data-x=\"3600\" data-y=\"900\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Speaking of connecting musical systems to other systems, how about the most interesting systems of them all: Human beings?\n            <\/p><p class=\"state state-init mode-present\">\n              The most interesting systems in the world: Human beings\n            <\/p><\/div><div class=\"vis\"><img src=\"src\" alt=\"image\" \/><footer>\n              Image: <a href=\"https:\/\/www.flickr.com\/photos\/bionicteaching\/14092973648\" target=\"_blank\">Tom Woodward<\/a>\n            <\/footer><\/div><\/div><div class=\"step centered-content future\" data-x=\"4350\" data-y=\"100\" data-scale=\"0.5\"><img src=\"src\" alt=\"image\" \/><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Eno's \u201cTrope\u201d is an example of a generative system that's also interactive.\n            <\/p><p class=\"state state-init mode-present\">\n              Brian Eno &amp; Peter Chilvers: Trope (2009)\n            <\/p><\/div><footer><a href=\"http:\/\/www.generativemusic.com\/trope.html\" target=\"_blank\">Trope<\/a> by Brian Eno and Peter Chilvers<\/footer><\/div><div class=\"step centered-content future\" data-x=\"2800\" data-y=\"-100\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              These interactive generative systems can be designed to give the user varying degrees of freedom.\n            <\/p><p class=\"state state-init mode-present\">\n              Degrees of freedom for the user vary\n            <\/p><p>The result is a <em>collaboration<\/em> between you and Music Mouse.\n              \n            <\/p><\/div><div class=\"vis\"><footer>\n              A partial emulation of\n              <a href=\"http:\/\/retiary.org\/ls\/index.html\" target=\"_blank\">Laurie Spiegel<\/a>'s\n              <a href=\"http:\/\/retiary.org\/ls\/programs.html\" target=\"_blank\">\u201cMusic Mouse - An Intelligent Instrument\u201d<\/a>.\n              <br \/>Use 'z' and 'x' to transpose, 'c' to toggle generative pattern. <a href=\"https:\/\/teropa.info\/musicmouse\" target=\"_blank\">More features here<\/a>.\n            <\/footer><\/div><\/div><div class=\"step centered-content future\" data-x=\"4400\" data-y=\"-1000\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Another form of human-machine interaction is <em>teaching<\/em> machines how to do things.\n            <\/p><p class=\"state state-init mode-present\">\n              Teaching a machine to do things\n            <\/p><\/div><div class=\"vis\"><img src=\"src\" alt=\"image\" \/><\/div><\/div><div class=\"step centered-content future\" data-x=\"2800\" data-y=\"-1000\"><div class=\"vis\"><div class=\"controls\"><footer>\n                Play with mouse, keyboard, or MIDI keyboard (Chrome only)\n                \n                <br \/>See also: <a href=\"https:\/\/codepen.io\/teropa\/pen\/bRqYVj\" target=\"_blank\">Dancing Markov Gymnop\u00e9dies<\/a>\n              <\/footer><\/div><\/div><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              <em>Markov chains<\/em> are an old, but still effective, example of the machine learning approach.\n            <\/p><p class=\"state state-init mode-present\">\n              Markov Chains\n            <\/p><\/div><\/div><div class=\"step centered-content future\" data-x=\"1300\" data-y=\"-850\" data-scale=\"0.5\"><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              Markov chains are an ancestor of contemporary machine learning approaches to music generation, such as deep neural networks.\n            <\/p><p class=\"state state-init mode-present\">\n              Contemporary machine learning: Deep neural networks\n            <\/p><\/div><div class=\"vis\"><a href=\"https:\/\/magenta.tensorflow.org\/performance-rnn-browser\" target=\"_blank\"><img class=\"performancernn\" src=\"src\" alt=\"image\" \/><\/a><a href=\"https:\/\/magenta.tensorflow.org\/nsynth\" target=\"_blank\"><img class=\"nsynth\" src=\"src\" alt=\"image\" \/><\/a><footer>\n              See <a href=\"https:\/\/medium.com\/artists-and-machine-intelligence\/neural-nets-for-generating-music-f46dffac21c0\" target=\"_blank\">Neural Nets for Generating Music<\/a> by Kyle McDonald.\n            <\/footer><\/div><\/div><div class=\"step centered-content future\" data-x=\"250\" data-y=\"1400\" data-scale=\"6\"><h2>Systems<\/h2><div class=\"text-content\"><p class=\"state state-init mode-read\">\n              This is what the generative music landscape looks like to me.\n            <\/p><p class=\"state state-init mode-present\">\n              The generative music landscape\n            <\/p><\/div><\/div>","created_at":"2018-11-07T16:45:21+0000","updated_at":"2018-11-07T16:45:21+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"teropa.info","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":345,"title":"Elements of Clojure by Zachary Tellman [Leanpub PDF\/iPad\/Kindle]","url":"https:\/\/leanpub.com\/elementsofclojure","content":"<section class=\"book-hero\" data-reactid=\"104\"><div class=\"container--large\" data-reactid=\"128\"><aside class=\"book-hero__summary ltr\" data-reactid=\"129\"><h3 class=\"book-hero__title ltr\" data-reactid=\"130\">Elements of Clojure<\/h3><div class=\"book-hero__blurb\" data-reactid=\"136\"><p>This is a book for anyone who knows Clojure, and wants to use it more effectively. It gives teams of Clojure developers the concepts and vocabulary necessary to weigh and discuss different approaches to a given problem.<\/p><\/div><div class=\"book-hero__actions\" data-reactid=\"137\"><p><a class=\"btn--sample-download btn--hollow\" data-reactid=\"139\">Read Free sample<i class=\"fa fa-file-pdf-o\" data-reactid=\"144\"><\/i><\/a><\/p><p><a class=\"btn--hollow\" data-reactid=\"146\">Table of Contents<i class=\"fa fa-list-ol\" data-reactid=\"148\"><\/i><\/a><\/p><\/div><\/aside><div class=\"book-hero__cover\" data-reactid=\"149\"><div class=\"book-cover hero\" data-reactid=\"150\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/elementsofclojure\/hero?1523239691\" alt=\"Elements of Clojure\" data-reactid=\"151\" \/><\/div><div class=\"book-progress\" data-reactid=\"152\"><p class=\"percent-complete\" data-reactid=\"155\">This book is 95% complete<\/p><\/div><\/div><aside class=\"book-hero__purchase\" data-reactid=\"159\"><div data-reactid=\"161\"><div class=\"pricing\" data-reactid=\"162\"><div class=\"price--minimum\" data-reactid=\"163\"><p>$25.00<\/p><p>Minimum price<\/p><\/div><div class=\"price--suggested\" data-reactid=\"169\"><p>$35.00<\/p><p>Suggested price<\/p><\/div><i class=\"fa fa-question-circle\" data-reactid=\"176\"><\/i><\/div><div class=\"cartable-form\" data-reactid=\"177\"><div class=\"cartable-form__actions\" data-reactid=\"179\"><div class=\"cartable-input\" data-reactid=\"180\"><div class=\"cartable-input__wrapper\" data-reactid=\"181\"><label data-reactid=\"182\">You pay (US$)<\/label><div class=\"cartable-form__action--with-hint\" data-reactid=\"186\"><p class=\"hint\" data-reactid=\"188\">EU customers: Price excludes VAT.<br data-reactid=\"191\" \/> VAT is added during checkout.<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/aside><\/div><\/section><section class=\"layer\" data-reactid=\"235\"><div class=\"container--large\" data-reactid=\"236\"><div class=\"container--small\" data-reactid=\"237\"><div class=\"truncate__text\" data-reactid=\"240\"><div class=\"truncate__text\" data-reactid=\"241\"><div data-reactid=\"242\"><div data-reactid=\"243\"><ul class=\"toc no-parts\"><li>\n    Introduction\n  <\/li>\n  <li>\n    Version\n  <\/li>\n  <li>\n    Names\n    <ul><li>\n        Naming Data\n      <\/li>\n      <li>\n        Naming Functions\n      <\/li>\n      <li>\n        Naming Macros\n      <\/li>\n    <\/ul><\/li>\n  <li>\n    Idioms\n    <ul><li>\n        When using inequalities, prefer <code>&lt;<\/code> and <code>&lt;=<\/code>\n      <\/li>\n      <li>\n        If a function accumulates values, support every arity\n      <\/li>\n      <li>\n        Use option maps, not named parameters\n      <\/li>\n      <li>\n        No one should have to know you\u2019ve used binding\n      <\/li>\n      <li>\n        If you have mutable state, use an atom\n      <\/li>\n    <\/ul><\/li>\n  <li>\n    Indirection\n    <ul><li>\n        Method Dispatch\n      <\/li>\n      <li>\n        What is an Abstraction?\n      <\/li>\n      <li>\n        A Model for Abstractions\n      <\/li>\n      <li>\n        Consequences of our Model\n      <\/li>\n      <li>\n        Systems of Abstractions\n      <\/li>\n    <\/ul><\/li>\n  <li>\n    Composition\n    <ul><li>\n        A Unit of Computation\n      <\/li>\n      <li>\n        Building a Process\n      <\/li>\n      <li>\n        Composing Processes\n      <\/li>\n    <\/ul><\/li>\n  <li>\n    Notes\n  <\/li>\n<\/ul><\/div><\/div><\/div><\/div><\/div><\/div><\/section><div class=\"happiness-guarantee\" data-reactid=\"244\"><div class=\"container--large\" data-reactid=\"245\"><h3 class=\"hg-title\" data-reactid=\"246\">The Leanpub 45-day 100% Happiness Guarantee<\/h3><p data-reactid=\"248\">Within <strong>45 days of purchase<\/strong> you can get a <strong>100% refund<\/strong> on any Leanpub purchase, in <strong>two clicks<\/strong>. <\/p><p><a href=\"https:\/\/leanpub.com\/terms#returns\">See full terms...<\/a><\/p><\/div><\/div><section class=\"layer dark\" data-reactid=\"259\"><section class=\"layer\" data-reactid=\"268\"><div class=\"container--large\" data-reactid=\"269\"><div class=\"carousel\" data-reactid=\"270\"><div data-reactid=\"271\"><div class=\"slick-slider slick-initialized\" dir=\"ltr\" data-reactid=\"277\"><a class=\"slick-arrow slick-prev slick-disabled\" data-reactid=\"278\"><\/a><div class=\"slick-list\" data-reactid=\"280\"><div class=\"slick-track\" data-reactid=\"281\"><div data-index=\"0\" class=\"slick-slide slick-active slick-current\" tabindex=\"-1\" aria-hidden=\"false\" data-reactid=\"282\"><div data-reactid=\"283\"><div tabindex=\"-1\" data-reactid=\"284\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"285\"><div class=\"lane-item--bundle\" data-reactid=\"286\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/software-architecture\" data-reactid=\"287\"><div class=\"book-cover medium\" data-reactid=\"288\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/software-architecture-for-developers\/medium?1540225551\" alt=\"Technical leadership and the balance with agility\" data-reactid=\"289\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"290\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/visualising-software-architecture\/medium?1537449617\" alt=\"Visualise, document and explore your software architecture\" data-reactid=\"291\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/software-architecture\" data-reactid=\"293\">Software Architecture for Developers: Volumes 1 &amp; 2 - Technical leadership and communication<\/a><\/p><p>2 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"1\" class=\"slick-slide slick-active\" tabindex=\"-1\" aria-hidden=\"false\" data-reactid=\"297\"><div data-reactid=\"298\"><div tabindex=\"-1\" data-reactid=\"299\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"300\"><div class=\"lane-item--bundle\" data-reactid=\"301\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/masteringcontainers\" data-reactid=\"302\"><div class=\"book-cover medium\" data-reactid=\"303\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/dockerdeepdive\/medium?1518078297\" alt=\"Docker Deep Dive\" data-reactid=\"304\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"305\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/thekubernetesbook\/medium?1541111836\" alt=\"The Kubernetes Book\" data-reactid=\"306\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/masteringcontainers\" data-reactid=\"308\">Mastering Containers<\/a><\/p><p>2 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"2\" class=\"slick-slide slick-active\" tabindex=\"-1\" aria-hidden=\"false\" data-reactid=\"312\"><div data-reactid=\"313\"><div tabindex=\"-1\" data-reactid=\"314\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"315\"><div class=\"lane-item--bundle\" data-reactid=\"316\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/python-craftsman\" data-reactid=\"317\"><div class=\"book-cover medium\" data-reactid=\"318\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/python-apprentice\/medium?1539350733\" alt=\"The Python Apprentice\" data-reactid=\"319\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"320\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/python-journeyman\/medium?1539934335\" alt=\"The Python Journeyman\" data-reactid=\"321\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"322\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/python-master\/medium?1537207866\" alt=\"The Python Master\" data-reactid=\"323\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/python-craftsman\" data-reactid=\"325\">The Python Craftsman<\/a><\/p><p>3 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"3\" class=\"slick-slide slick-active\" tabindex=\"-1\" aria-hidden=\"false\" data-reactid=\"329\"><div data-reactid=\"330\"><div tabindex=\"-1\" data-reactid=\"331\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"332\"><div class=\"lane-item--bundle\" data-reactid=\"333\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/thedevopstoolkitseriesvolumes4and5\" data-reactid=\"334\"><div class=\"book-cover medium\" data-reactid=\"335\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/the-devops-2-3-toolkit\/medium?1522332897\" alt=\"The DevOps 2.3 Toolkit: Kubernetes\" data-reactid=\"336\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"337\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/the-devops-2-4-toolkit\/medium?1532460436\" alt=\"The DevOps 2.4 Toolkit: Continuous Deployment To Kubernetes\" data-reactid=\"338\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/thedevopstoolkitseriesvolumes4and5\" data-reactid=\"340\">The DevOps Toolkit Series (volumes 4 and 5): Kubernetes<\/a><\/p><p>2 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"4\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"344\"><div data-reactid=\"345\"><div tabindex=\"-1\" data-reactid=\"346\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"347\"><div class=\"lane-item--bundle\" data-reactid=\"348\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/thecompletecoachsguide\" data-reactid=\"349\"><div class=\"book-cover medium\" data-reactid=\"350\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/CollaborationGamesToolbox\/medium?1518048264\" alt=\"Collaboration Games\" data-reactid=\"351\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"352\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/TrainingScrum\/medium?1518048438\" alt=\"Growing Agile: A Coach's Guide to Training Scrum\" data-reactid=\"353\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"354\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/agilerequirements\/medium?1518057036\" alt=\"Growing Agile: A Coach's Guide to Agile Requirements\" data-reactid=\"355\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"356\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/MasteringBacklogs\/medium?1518058731\" alt=\"Growing Agile: A Coach's Guide to Mastering Backlogs\" data-reactid=\"357\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"358\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/ReleasePlanning\/medium?1518058734\" alt=\"Growing Agile: A Coach's Guide to Release Planning\" data-reactid=\"359\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/thecompletecoachsguide\" data-reactid=\"361\">Growing Agile: The Complete Coach's Guide<\/a><\/p><p>7 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"5\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"365\"><div data-reactid=\"366\"><div tabindex=\"-1\" data-reactid=\"367\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"368\"><div class=\"lane-item--bundle\" data-reactid=\"369\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/fiftyquickbundle\" data-reactid=\"370\"><div class=\"book-cover medium\" data-reactid=\"371\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/50quickideas\/medium?1518055391\" alt=\"Fifty Quick Ideas to Improve your User Stories\" data-reactid=\"372\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"373\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/50quickideas-tests\/medium?1518060388\" alt=\"Fifty Quick Ideas to Improve Your Tests\" data-reactid=\"374\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"375\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/50quickretrospectives\/medium?1537436172\" alt=\"Fifty Quick Ideas To Improve Your Retrospectives\" data-reactid=\"376\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/fiftyquickbundle\" data-reactid=\"378\">Fifty Quick Ideas<\/a><\/p><p>3 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"6\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"382\"><div data-reactid=\"383\"><div tabindex=\"-1\" data-reactid=\"384\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"385\"><div class=\"lane-item--bundle\" data-reactid=\"386\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/thecstandardlibraryandconcurrencywithmodernc\" data-reactid=\"387\"><div class=\"book-cover medium\" data-reactid=\"388\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/cpplibrary\/medium?1518077427\" alt=\"The C++ Standard Library\" data-reactid=\"389\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"390\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/concurrencywithmodernc\/medium?1518079578\" alt=\"Concurrency with Modern C++\" data-reactid=\"391\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/thecstandardlibraryandconcurrencywithmodernc\" data-reactid=\"393\">\"The C++ Standard Library\" and \"Concurrency with Modern C++\"<\/a><\/p><p>2 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"7\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"397\"><div data-reactid=\"398\"><div tabindex=\"-1\" data-reactid=\"399\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"400\"><div class=\"lane-item--bundle\" data-reactid=\"401\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/node-beginner-and-craftsman-bundle\" data-reactid=\"402\"><div class=\"book-cover medium\" data-reactid=\"403\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/nodebeginner\/medium?1541533090\" alt=\"The Node Beginner Book\" data-reactid=\"404\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"405\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/nodecraftsman\/medium?1518053404\" alt=\"The Node Craftsman Book\" data-reactid=\"406\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/node-beginner-and-craftsman-bundle\" data-reactid=\"408\">The Node Beginner &amp; Craftsman Bundle<\/a><\/p><p>2 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"8\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"412\"><div data-reactid=\"413\"><div tabindex=\"-1\" data-reactid=\"414\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"415\"><div class=\"lane-item--bundle\" data-reactid=\"416\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/learngitbashandterraformthehardway\" data-reactid=\"417\"><div class=\"book-cover medium\" data-reactid=\"418\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/learnbashthehardway\/medium?1518080162\" alt=\"Learn Bash the Hard Way\" data-reactid=\"419\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"420\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/learngitthehardway\/medium?1518080276\" alt=\"Learn Git The Hard Way\" data-reactid=\"421\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"422\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/learnterraformthehardway\/medium?1528183930\" alt=\"Learn Terraform The Hard Way\" data-reactid=\"423\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/learngitbashandterraformthehardway\" data-reactid=\"425\">Learn Git, Bash, and Terraform the Hard Way<\/a><\/p><p>3 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"9\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"429\"><div data-reactid=\"430\"><div tabindex=\"-1\" data-reactid=\"431\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"432\"><div class=\"lane-item--bundle\" data-reactid=\"433\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/lean-agilepredictabilitybundle\" data-reactid=\"434\"><div class=\"book-cover medium\" data-reactid=\"435\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/actionableagilemetrics\/medium?1529583116\" alt=\"Actionable Agile Metrics for Predictability\" data-reactid=\"436\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"437\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/whenwillitbedone\/medium?1518079211\" alt=\"When Will It Be Done?\" data-reactid=\"438\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/lean-agilepredictabilitybundle\" data-reactid=\"440\">Lean-Agile Predictability Bundle<\/a><\/p><p>2 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"10\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"444\"><div data-reactid=\"445\"><div tabindex=\"-1\" data-reactid=\"446\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"447\"><div class=\"lane-item--bundle\" data-reactid=\"448\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/development-and-deployment-of-multiplayer-online-games-part-arch\" data-reactid=\"449\"><div class=\"book-cover medium\" data-reactid=\"450\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/development-and-deployment-of-multiplayer-online-games-vol1\/medium?1518079072\" alt=\"Development &amp; Deployment of Multiplayer Online Games Vol. I\" data-reactid=\"451\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"452\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/development-and-deployment-of-multiplayer-online-games-vol-ii\/medium?1518079539\" alt=\"Development and Deployment of Multiplayer Online Games, Vol. II.\" data-reactid=\"453\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"454\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/development-and-deployment-of-multiplayer-online-games-vol-iii\/medium?1518079540\" alt=\"Development and Deployment of Multiplayer Online Games, Vol. III.\" data-reactid=\"455\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/development-and-deployment-of-multiplayer-online-games-part-arch\" data-reactid=\"457\">Development and Deployment of Multiplayer Online Games, Part ARCH. Architecture (Vol. I-III)<\/a><\/p><p>3 Books<\/p><\/div><\/div><\/div><\/div><\/div><div data-index=\"11\" class=\"slick-slide\" tabindex=\"-1\" aria-hidden=\"true\" data-reactid=\"461\"><div data-reactid=\"462\"><div tabindex=\"-1\" data-reactid=\"463\"><div class=\"lane-item-wrapper--bundle\" data-reactid=\"464\"><div class=\"lane-item--bundle\" data-reactid=\"465\"><a class=\"book-cover-wrapper\" href=\"https:\/\/leanpub.com\/b\/complete-scala-bundle\" data-reactid=\"466\"><div class=\"book-cover medium\" data-reactid=\"467\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/modern-web-development-with-scala\/medium?1540001967\" alt=\"Modern Web Development with Scala\" data-reactid=\"468\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"469\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/mastering-advanced-scala\/medium?1518077402\" alt=\"Mastering Advanced Scala\" data-reactid=\"470\" \/><\/div><div class=\"book-cover medium\" data-reactid=\"471\"><img src=\"https:\/\/s3.amazonaws.com\/titlepages.leanpub.com\/practical-event-sourcing-with-scala\/medium?1530351752\" alt=\"Practical Event Sourcing with Scala\" data-reactid=\"472\" \/><\/div><\/a><p><a href=\"https:\/\/leanpub.com\/b\/complete-scala-bundle\" data-reactid=\"474\">Complete Scala Bundle<\/a><\/p><p>3 Books<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><a class=\"slick-arrow slick-next\" data-reactid=\"478\"><\/a><\/div><\/div><\/div><\/div><\/section><\/section>","created_at":"2018-11-07T16:45:20+0000","updated_at":"2018-11-07T16:45:20+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"leanpub.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":344,"title":"How to make a JAR file Linux executable (Example)","url":"https:\/\/coderwall.com\/p\/ssuaxa\/how-to-make-a-jar-file-linux-executable","content":"<p>Every Java programmer knows - or should known - that it is possible to create a runnable Java package archive (JAR), so that in order to launch an application it is enough to specify the jar file name on the Java interpreter command line along with the <code class=\"prettyprint\">-jar<\/code> parameter. For example:  <\/p><pre class=\"prettyprint\">$ java -jar helloworld.jar<\/pre>\n<p>There are plenty of tutorials showing how to implement this feature using Ant, Maven, Eclipse, Netbens, etc.<\/p>\n<p>Anyway in its basic form, it just requires to add a <code class=\"prettyprint\">MANIFEST.MF<\/code> file to the jar package. The manifest must contain an entry <code class=\"prettyprint\">Main-Class<\/code> that specifies which is the class defining the main method for your application. For example: <\/p>\n<pre class=\"prettyprint\">$ javac HelloWorld.java\n$ echo Main-Class: HelloWorld &gt; MANIFEST.MF\n$ jar -cvmf MANIFEST.MF helloworld.jar HelloWorld.class <\/pre>\n<p>But this still requires your users to invoke the Java interpreter with the <code class=\"prettyprint\">-jar<\/code> option. There are many reasons why it would be preferable to have your app runnable by simply invoking it on the terminal shell like any other command. <\/p>\n<p><strong>Here comes the protip!<\/strong><\/p>\n<p>This technique it is based on the ability to append a generic binary payload to a Linux shell script. Read more about this here: <a href=\"http:\/\/www.linuxjournal.com\/content\/add-binary-payload-your-shell-scripts\" rel=\"nofollow\">http:\/\/www.linuxjournal.com\/content\/add-binary-payload-your-shell-scripts<\/a><\/p>\n<p>Taking advantage of this possibility the trick is just to embed a runnable jar file into a Bash script file. The script when executed will launch the Java interpreter specifying itself as the jar to run. Too complex? Much more easier to do in practice, than to explain!<\/p>\n<p>Let's say that you have a runnable jar named <code class=\"prettyprint\">helloworld.jar<\/code> <\/p>\n<p>Copy the Bash script below to a file named <code class=\"prettyprint\">stub.sh<\/code><\/p>\n<pre class=\"prettyprint\">#!\/bin\/sh\nMYSELF=`which \"$0\" 2&gt;\/dev\/null`\n[ $? -gt 0 -a -f \"$0\" ] &amp;&amp; MYSELF=\".\/$0\"\njava=java\nif test -n \"$JAVA_HOME\"; then\n    java=\"$JAVA_HOME\/bin\/java\"\nfi\nexec \"$java\" $java_args -jar $MYSELF \"$@\"\nexit 1 <\/pre>\n<p>Than append the jar file to the saved script and grant the execute permission to the file resulting with the following command: <\/p>\n<pre class=\"prettyprint\">cat stub.sh helloworld.jar &gt; hello.run &amp;&amp; chmod +x helloworld.run <\/pre>\n<p><strong>That's all!<\/strong> <\/p>\n<p>Now you can execute the app just typing <code class=\"prettyprint\">helloworld.run<\/code> on your shell terminal. <\/p>\n<p>The script is smart enough to pass any command line parameters to the Java application transparently. Cool! Isn't it ?!<\/p>\n<p>In the case your are a Windows guy, obviously this will not work (except you will run a Linux compatibility layer like Cygwin). <\/p>\n<p>Anyway exist tools that are able to wrap a Java application into a native Windows <code class=\"prettyprint\">.exe<\/code> binary file, producing a result similar to the one explained in this tutorial. See for example <a href=\"http:\/\/launch4j.sourceforge.net\/\" rel=\"nofollow\">http:\/\/launch4j.sourceforge.net\/<\/a><\/p>","created_at":"2018-11-07T16:45:19+0000","updated_at":"2018-11-07T16:45:19+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"coderwall.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":341,"title":"r\/Steam_Link - Wireless Headset with steam link?","url":"https:\/\/www.reddit.com\/r\/Steam_Link\/comments\/56ovyo\/wireless_headset_with_steam_link\/","content":"<p class=\"yklcuq-10 hpxQMr\">Can I use my Logitech G933 wireless headset with the Steam Link? Moving in a few days and the computer will be far away so I have to use a Steam Link to stream games and videos\/pictures via Kodi. Can I use the bluetooth dongle for the G933 in the USB on the Steam Link? Do not need support for voice, only sound and hopefully not only stereo.<\/p>","created_at":"2018-11-07T16:45:18+0000","updated_at":"2018-11-07T16:45:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.reddit.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":342,"title":"esimov\/caire","url":"https:\/\/github.com\/esimov\/caire","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/travis-ci.org\/esimov\/caire\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7a7da78f87faa3539a1e40a8ac9c7e7b1ba47e67\/68747470733a2f2f7472617669732d63692e6f72672f6573696d6f762f63616972652e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/esimov\/caire.svg?branch=master\" \/><\/a>\n<a href=\"https:\/\/godoc.org\/github.com\/esimov\/caire\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/915b7be44ada53c290eb157634330494ebe3e30a\/68747470733a2f2f676f646f632e6f72672f6769746875622e636f6d2f676f6c616e672f6764646f3f7374617475732e737667\" alt=\"GoDoc\" data-canonical-src=\"https:\/\/godoc.org\/github.com\/golang\/gddo?status.svg\" \/><\/a>\n<a href=\"https:\/\/github.com\/esimov\/caire\/blob\/master\/LICENSE\"><img src=\"https:\/\/camo.githubusercontent.com\/22c72682b8a137c2797da5c5a434814f3362f196\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f6c6963656e73652f6d6173686170652f6170697374617475732e7376673f7374796c653d666c6174\" alt=\"license\" data-canonical-src=\"https:\/\/img.shields.io\/github\/license\/mashape\/apistatus.svg?style=flat\" \/><\/a>\n<a href=\"https:\/\/github.com\/esimov\/caire\/blob\/master\"><img src=\"https:\/\/camo.githubusercontent.com\/e6f667cc3ad1f3de25982bcc752fd81b0cf0b9d5\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f72656c656173652d76312e312e312d626c75652e737667\" alt=\"release\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/release-v1.1.1-blue.svg\" \/><\/a>\n<a href=\"https:\/\/github.com\/esimov\/caire\/blob\/master\"><img src=\"https:\/\/camo.githubusercontent.com\/d1c654bc8f7e50f563c966da0a1260589fd3758c\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f686f6d65627265772d76312e312e312d6f72616e67652e737667\" alt=\"homebrew\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/homebrew-v1.1.1-orange.svg\" \/><\/a><\/p>\n<p><strong>Caire<\/strong> is a content aware image resize library based on <em><a href=\"https:\/\/inst.eecs.berkeley.edu\/~cs194-26\/fa16\/hw\/proj4-seamcarving\/imret.pdf\" rel=\"nofollow\">Seam Carving for Content-Aware Image Resizing<\/a><\/em> paper.<\/p>\n<h3><a id=\"user-content-how-does-it-work\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-does-it-work\"><\/a>How does it work<\/h3>\n<ul><li>An energy map (edge detection) is generated from the provided image.<\/li>\n<li>The algorithm tries to find the least important parts of the image taking into account the lowest energy values.<\/li>\n<li>Using a dynamic programming approach the algorithm will generate individual seams accrossing the image from top to down, or from left to right (depending on the horizontal or vertical resizing) and will allocate for each seam a custom value, the least important pixels having the lowest energy cost and the most important ones having the highest cost.<\/li>\n<li>Traverse the image from the second row to the last row and compute the cumulative minimum energy for all possible connected seams for each entry.<\/li>\n<li>The minimum energy level is calculated by summing up the current pixel with the lowest value of the neighboring pixels from the previous row.<\/li>\n<li>Traverse the image from top to bottom and compute the minimum energy level. For each pixel in a row we compute the energy of the current pixel plus the energy of one of the three possible pixels above it.<\/li>\n<li>Find the lowest cost seam from the energy matrix starting from the last row and remove it.<\/li>\n<li>Repeat the process.<\/li>\n<\/ul><h4><a id=\"user-content-the-process-illustrated\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-process-illustrated\"><\/a>The process illustrated:<\/h4>\n<table><thead><tr><th align=\"center\">Original image<\/th>\n<th align=\"center\">Energy map<\/th>\n<th align=\"center\">Seams applied<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"center\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35481925-de130752-0435-11e8-9246-3950679b4fd6.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35481925-de130752-0435-11e8-9246-3950679b4fd6.jpg\" alt=\"original\" \/><\/a><\/td>\n<td align=\"center\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35481899-5d5096ca-0435-11e8-9f9b-a84fefc06470.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35481899-5d5096ca-0435-11e8-9f9b-a84fefc06470.jpg\" alt=\"sobel\" \/><\/a><\/td>\n<td align=\"center\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35481949-5c74dcb0-0436-11e8-97db-a6169cb150ca.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35481949-5c74dcb0-0436-11e8-97db-a6169cb150ca.jpg\" alt=\"debug\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<p>Key features which differentiates from the other existing open source solutions:<\/p>\n<ul class=\"contains-task-list\"><li class=\"task-list-item\"> Customizable command line support<\/li>\n<li class=\"task-list-item\"> Support for both shrinking or enlarging the image<\/li>\n<li class=\"task-list-item\"> Resize image both vertically and horizontally<\/li>\n<li class=\"task-list-item\"> Can resize all the images from a directory<\/li>\n<li class=\"task-list-item\"> Does not require any third party library<\/li>\n<li class=\"task-list-item\"> Use of sobel threshold for fine tuning<\/li>\n<li class=\"task-list-item\"> Use of blur filter for increased edge detection<\/li>\n<li class=\"task-list-item\"> Make the image square with a single command<\/li>\n<li class=\"task-list-item\"> Support for proportional scaling<\/li>\n<li class=\"task-list-item\"> Face detection to avoid face deformation<\/li>\n<\/ul><h2><a id=\"user-content-face-detection\" class=\"anchor\" aria-hidden=\"true\" href=\"#face-detection\"><\/a>Face detection<\/h2>\n<p>The library is capable detecting human faces prior resizing the images via <a href=\"https:\/\/github.com\/esimov\/pigo\">https:\/\/github.com\/esimov\/pigo<\/a>, which does not require to have OpenCV installed.<\/p>\n<p>The image below illustrates the application capabilities to detect human faces prior resizing. It's clearly visible from the image that with face detection activated the algorithm will avoid cropping pixels inside faces, retaining the face zone unaltered.<\/p>\n<table><thead><tr><th align=\"center\">Original image<\/th>\n<th align=\"center\">With face detection<\/th>\n<th align=\"center\">Without face detection<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"center\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/37569642-0c5f49e8-2aee-11e8-8ac1-d096c0387ca0.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/37569642-0c5f49e8-2aee-11e8-8ac1-d096c0387ca0.jpg\" alt=\"Original\" \/><\/a><\/td>\n<td align=\"center\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/41292871-6ca43280-6e5c-11e8-9d72-5b9a138228b6.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/41292871-6ca43280-6e5c-11e8-9d72-5b9a138228b6.jpg\" alt=\"With Face Detection\" \/><\/a><\/td>\n<td align=\"center\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/41292872-6cc90e8e-6e5c-11e8-8b41-5b4eb5042381.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/41292872-6cc90e8e-6e5c-11e8-8b41-5b4eb5042381.jpg\" alt=\"Without Face Detection\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><p><a href=\"http:\/\/www.lens-rumors.com\/wp-content\/uploads\/2014\/12\/EF-M-55-200mm-f4.5-6.3-IS-STM-sample.jpg\" rel=\"nofollow\">Sample image source<\/a><\/p>\n<h2><a id=\"user-content-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#install\"><\/a>Install<\/h2>\n<p>First, install Go, set your <code>GOPATH<\/code>, and make sure <code>$GOPATH\/bin<\/code> is on your <code>PATH<\/code>.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ export GOPATH=\"$HOME\/go\"\n$ export PATH=\"$PATH:$GOPATH\/bin\"<\/pre><\/div>\n<p>Next download the project and build the binary file.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ go get -u -f github.com\/esimov\/caire\/cmd\/caire\n$ go install<\/pre><\/div>\n<h2><a id=\"user-content-macos-brew-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#macos-brew-install\"><\/a>MacOS (Brew) install<\/h2>\n<p>The library now can be installed via Homebrew. The only thing you need is to run the commands below.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ brew tap esimov\/caire\n$ brew install caire<\/pre><\/div>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<div class=\"highlight highlight-source-shell\"><pre>$ caire -in input.jpg -out output.jpg<\/pre><\/div>\n<p>To detect faces prior rescaling use the <code>-face<\/code> flag and provide the face clasification binary file included in the <code>data<\/code> folder. The sample code below will rescale the provided image with 20% but will check for human faces prior rescaling.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ caire -in input.jpg -out output.jpg -face=1 -cc=\"data\/facefinder\" -perc=1 -width=20<\/pre><\/div>\n<h3><a id=\"user-content-supported-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-commands\"><\/a>Supported commands:<\/h3>\n<div class=\"highlight highlight-source-shell\"><pre>$ caire --help<\/pre><\/div>\n<p>The following flags are supported:<\/p>\n<table><thead><tr><th>Flag<\/th>\n<th>Default<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><code>in<\/code><\/td>\n<td>n\/a<\/td>\n<td>Input file<\/td>\n<\/tr><tr><td><code>out<\/code><\/td>\n<td>n\/a<\/td>\n<td>Output file<\/td>\n<\/tr><tr><td><code>width<\/code><\/td>\n<td>n\/a<\/td>\n<td>New width<\/td>\n<\/tr><tr><td><code>height<\/code><\/td>\n<td>n\/a<\/td>\n<td>New height<\/td>\n<\/tr><tr><td><code>perc<\/code><\/td>\n<td>false<\/td>\n<td>Reduce image by percentage<\/td>\n<\/tr><tr><td><code>square<\/code><\/td>\n<td>false<\/td>\n<td>Reduce image to square dimensions<\/td>\n<\/tr><tr><td><code>scale<\/code><\/td>\n<td>false<\/td>\n<td>Proportional scaling<\/td>\n<\/tr><tr><td><code>blur<\/code><\/td>\n<td>1<\/td>\n<td>Blur radius<\/td>\n<\/tr><tr><td><code>sobel<\/code><\/td>\n<td>10<\/td>\n<td>Sobel filter threshold<\/td>\n<\/tr><tr><td><code>debug<\/code><\/td>\n<td>false<\/td>\n<td>Use debugger<\/td>\n<\/tr><tr><td><code>face<\/code><\/td>\n<td>false<\/td>\n<td>Use face detection<\/td>\n<\/tr><tr><td><code>cc<\/code><\/td>\n<td>string<\/td>\n<td>Cascade classifier<\/td>\n<\/tr><\/tbody><\/table><p>In case you wish to scale down the image by a specific percentage, it can be used the <code>-perc<\/code> boolean flag. For example to reduce the image dimension by 20% both horizontally and vertically you can use the following command:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ caire -in input\/source.jpg -out .\/out.jpg -perc=1 -width=20 -height=20 -debug=false<\/pre><\/div>\n<p>Also the library supports the <code>-square<\/code> option. When this option is used the image will be resized to a squre, based on the shortest edge.<\/p>\n<p>The <code>-scale<\/code> option will resize the image proportionally. First the image is scaled down preserving the image aspect ratio, then the seam carving algorithm is applied only to the remaining points. Ex. : given an image of dimensions 2048x1536 if we want to resize to the 1024x500, the tool first rescale the image to 1024x768, then will remove only the remaining 268px. <strong>Using this option will drastically reduce the processing time.<\/strong><\/p>\n<p>The CLI command can process all the images from a specific directory too.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ caire -in .\/input-directory -out .\/output-directory<\/pre><\/div>\n<h2><a id=\"user-content-sample-images\" class=\"anchor\" aria-hidden=\"true\" href=\"#sample-images\"><\/a>Sample images<\/h2>\n<h4><a id=\"user-content-shrunk-images\" class=\"anchor\" aria-hidden=\"true\" href=\"#shrunk-images\"><\/a>Shrunk images<\/h4>\n<table><thead><tr><th>Original<\/th>\n<th>Shrunk<\/th>\n<\/tr><\/thead><tbody><tr><td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498083-83d6015e-04d5-11e8-936a-883e17b76f9d.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498083-83d6015e-04d5-11e8-936a-883e17b76f9d.jpg\" alt=\"broadway_tower_edit\" \/><\/a><\/td>\n<td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498110-a4a03328-04d5-11e8-9bf1-f526ef033d6a.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498110-a4a03328-04d5-11e8-9bf1-f526ef033d6a.jpg\" alt=\"broadway_tower_edit\" \/><\/a><\/td>\n<\/tr><tr><td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498250-2f31e202-04d6-11e8-8840-a78f40fc1a0c.png\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498250-2f31e202-04d6-11e8-8840-a78f40fc1a0c.png\" alt=\"waterfall\" \/><\/a><\/td>\n<td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498209-0411b16a-04d6-11e8-9ce2-ec4bce34828a.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498209-0411b16a-04d6-11e8-9ce2-ec4bce34828a.jpg\" alt=\"waterfall\" \/><\/a><\/td>\n<\/tr><tr><td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498466-1375b88a-04d7-11e8-8f8e-9d202da6a6b3.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498466-1375b88a-04d7-11e8-8f8e-9d202da6a6b3.jpg\" alt=\"dubai\" \/><\/a><\/td>\n<td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498499-3c32fc38-04d7-11e8-9f0d-07f63a8bd420.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498499-3c32fc38-04d7-11e8-9f0d-07f63a8bd420.jpg\" alt=\"dubai\" \/><\/a><\/td>\n<\/tr><tr><td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498465-1317a678-04d7-11e8-9185-ec92ea57f7c6.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498465-1317a678-04d7-11e8-9185-ec92ea57f7c6.jpg\" alt=\"boat\" \/><\/a><\/td>\n<td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498498-3c0f182c-04d7-11e8-9af8-695bc071e0f1.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498498-3c0f182c-04d7-11e8-9af8-695bc071e0f1.jpg\" alt=\"boat\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><h4><a id=\"user-content-enlarged-images\" class=\"anchor\" aria-hidden=\"true\" href=\"#enlarged-images\"><\/a>Enlarged images<\/h4>\n<table><thead><tr><th>Original<\/th>\n<th>Extended<\/th>\n<\/tr><\/thead><tbody><tr><td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498662-e11853c4-04d7-11e8-98d7-fcdb27207362.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498662-e11853c4-04d7-11e8-98d7-fcdb27207362.jpg\" alt=\"gasadalur\" \/><\/a><\/td>\n<td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498559-87eb6426-04d7-11e8-825c-2dd2abdfc112.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498559-87eb6426-04d7-11e8-825c-2dd2abdfc112.jpg\" alt=\"gasadalur\" \/><\/a><\/td>\n<\/tr><tr><td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498466-1375b88a-04d7-11e8-8f8e-9d202da6a6b3.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498466-1375b88a-04d7-11e8-8f8e-9d202da6a6b3.jpg\" alt=\"dubai\" \/><\/a><\/td>\n<td><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/883386\/35498827-8cee502c-04d8-11e8-8449-05805f196d60.jpg\"><img src=\"https:\/\/user-images.githubusercontent.com\/883386\/35498827-8cee502c-04d8-11e8-8449-05805f196d60.jpg\" alt=\"dubai\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-useful-resources\" class=\"anchor\" aria-hidden=\"true\" href=\"#useful-resources\"><\/a>Useful resources<\/h3>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Seam_carving\" rel=\"nofollow\">https:\/\/en.wikipedia.org\/wiki\/Seam_carving<\/a><\/li>\n<li><a href=\"https:\/\/inst.eecs.berkeley.edu\/~cs194-26\/fa16\/hw\/proj4-seamcarving\/imret.pdf\" rel=\"nofollow\">https:\/\/inst.eecs.berkeley.edu\/~cs194-26\/fa16\/hw\/proj4-seamcarving\/imret.pdf<\/a><\/li>\n<li><a href=\"http:\/\/pages.cs.wisc.edu\/~moayad\/cs766\/download_files\/alnammi_cs_766_final_report.pdf\" rel=\"nofollow\">http:\/\/pages.cs.wisc.edu\/~moayad\/cs766\/download_files\/alnammi_cs_766_final_report.pdf<\/a><\/li>\n<li><a href=\"https:\/\/stacks.stanford.edu\/file\/druid:my512gb2187\/Zargham_Nassirpour_Content_aware_image_resizing.pdf\" rel=\"nofollow\">https:\/\/stacks.stanford.edu\/file\/druid:my512gb2187\/Zargham_Nassirpour_Content_aware_image_resizing.pdf<\/a><\/li>\n<\/ul><h2><a id=\"user-content-author\" class=\"anchor\" aria-hidden=\"true\" href=\"#author\"><\/a>Author<\/h2>\n<p>Simo Endre <a href=\"https:\/\/twitter.com\/simo_endre\" rel=\"nofollow\">@simo_endre<\/a><\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2018 Endre Simo<\/p>\n<p>This project is under the MIT License. See the LICENSE file for the full license text.<\/p>\n<\/article>","created_at":"2018-11-07T16:45:18+0000","updated_at":"2018-11-07T16:45:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/883386?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":343,"title":"Xi: an editor for the next 20 years - Recurse Center","url":"https:\/\/www.recurse.com\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:45:18+0000","updated_at":"2018-11-07T16:45:18+0000","annotations":[],"reading_time":0,"domain_name":"www.recurse.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":340,"title":"The best wireless gaming headset","url":"https:\/\/www.pcgamer.com\/best-wireless-gaming-headset\/","content":"<p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/bDJDFyN66MRLCwdYP7kteX-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/bDJDFyN66MRLCwdYP7kteX-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/bDJDFyN66MRLCwdYP7kteX-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/bDJDFyN66MRLCwdYP7kteX.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/bDJDFyN66MRLCwdYP7kteX.jpg\" \/><\/p><p>Many years ago now, wireless headsets had a reputation for audio dropout and poor battery life. Those issues might have been a turnoff once, but with today's wireless tech they're no longer true. There\u2019s simply no downside to ditching the cables in 2018 for one of the best wireless gaming headsets.<\/p><div class=\"fancy-box\"><h5>Best gaming headsets<\/h5><div class=\"fancy_box_body\"><p><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/v4YykcNTg738ZbHcRikBF6-320-80.jpg\" alt=\"\" class=\"pinterest-pin-exclude lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/v4YykcNTg738ZbHcRikBF6-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/v4YykcNTg738ZbHcRikBF6-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/v4YykcNTg738ZbHcRikBF6.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/v4YykcNTg738ZbHcRikBF6.jpg\" data-pin-nopin=\"true\" \/><\/p><p>Looking for wired cans? Check out our guide to the<a href=\"https:\/\/www.pcgamer.com\/the-best-gaming-headsets\/\"> best gaming headsets<\/a>.\u00a0<\/p><\/div><\/div><p>The price premium that manufacturers used to slap on their wireless models has now been all but eradicated, which means gamers have options across the full budget gamut, from budget headsets to audiophile-grade models priced up in the stratosphere. But even cheap models from reliable manufacturers have good build quality and solid wireless signals.<\/p><p>Of course, this being PC gaming, there are a lot of headsets out there that offer a lot more than simply the fundamentals. Streaming quality mics, transmitter boxes with built-in EQ controls and, inevitably, RGB lighting are all common features on models across the gamut.\u00a0<\/p><p>So which one\u2019s right for you? Based on extensive testing and tons of gaming and music time, we\u2019ve picked out our favourite wireless PC gaming headsets.\u00a0<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/5nM7XuAxJv7Q7krTyVqACV-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/5nM7XuAxJv7Q7krTyVqACV-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/5nM7XuAxJv7Q7krTyVqACV-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/5nM7XuAxJv7Q7krTyVqACV.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/5nM7XuAxJv7Q7krTyVqACV.jpg\" \/><\/p><h3>1. Steelseries Arctis 7<\/h3><div class=\"_hawk subtitle\"><p>The best overall wireless gaming headset<\/p><\/div><p><strong>Drivers: <\/strong>40mm neodymium | <strong>Battery life: <\/strong>Up to 24 hours | <strong>Frequency response: <\/strong>20Hz-20,000Hz | <strong>Impedance: <\/strong>32 Ohm | <strong>Features: <\/strong>Retractable noise-cancelling mic, chat mix control, customisable earcup plates and headband<\/p><div class=\"icon icon-plus_circle _hawk\">Great battery life<\/div><div class=\"icon icon-plus_circle _hawk\">Comfortable floating headband<\/div><div class=\"icon icon-minus_circle _hawk\">Headband does slacken over time<\/div><div class=\"icon icon-minus_circle _hawk\">Average mic<\/div><p>Whatever you\u2019re listening to through the Arctis 7s, there\u2019s none of the muddiness or audio artifacts that have historically ruined the party for wireless headsets\u2014it sounds just as good as the best wired models we\u2019ve tested at this same $150 price range. The extraordinary battery life clocks in at over 20 hours out of the box, and after almost a year of heavy use that figure\u2019s hardly dropped off. You can keep playing while you charge, too, simply by connecting the headset to your PC with a USB cable.<\/p><p>The Arctis range\u2019s distinctive ski goggle headband is really effective at keeping the weight of the headset away from your head. After a year of daily usage, the headband does slacken which makes for a looser and slightly less comfortable fit, but the bands themselves are replaceable. We\u2019re big fans of the control placements at the rear of the headset, too: volume wheel and mic mute on the left, chat\/game mix and headset on\/off on the right. The retractable mic is a little quiet, but it remains perfectly usable.\u00a0<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/uqyuQS5D3bbcBRC3dWKw7i-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/uqyuQS5D3bbcBRC3dWKw7i-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/uqyuQS5D3bbcBRC3dWKw7i-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/uqyuQS5D3bbcBRC3dWKw7i.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/uqyuQS5D3bbcBRC3dWKw7i.jpg\" \/><\/p><h3>2. Corsair HS70<\/h3><div class=\"_hawk subtitle\"><p>The best budget wireless gaming headset<\/p><\/div><p><strong>Drivers: <\/strong>50mm | <strong>Battery life: <\/strong>Up to 16 hours | <strong>Frequency response: <\/strong>20Hz-20,000Hz | <strong>Impedance: <\/strong>32 Ohm | <strong>Features: <\/strong>Available in black or white finishes, detachable mic<\/p><div class=\"icon icon-plus_circle _hawk\">Great construction<\/div><div class=\"icon icon-plus_circle _hawk\">No-nonsense package<\/div><div class=\"icon icon-minus_circle _hawk\">Headband could be comfier<\/div><div class=\"icon icon-minus_circle _hawk\">Metallic finish<\/div><p>This is seriously impressive work from Corsair, which has channelled all its expertise in higher-end models and somehow kept costs right down without compromising... well, anything detectable. If you\u2019d told us the HS70 was a $150 headset when we first unboxed it, we\u2019d believe it.\u00a0<\/p><p>Stereo spread and overall sound articulation are the highlights here, the drivers tuned in line with the modern trend for flatter EQs and thus better versatility when you close down PUBG and bring up that doom metal playlist you\u2019ve been working on in Spotify. The build quality is what really baffles us about the HS70\u2019s wonderful budget pricing though\u2014they feel sturdy enough to last years, but light on the head and well-padded. The slightly under-padded headband is the only exception. We\u2019re not wholly onboard with that perforated metallic finish on the earcups either, but that\u2019s literally a small price to pay for nailing everything that counts.<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/DbkYZ7ks6vPETFAYF5dyr-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/DbkYZ7ks6vPETFAYF5dyr-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/DbkYZ7ks6vPETFAYF5dyr-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/DbkYZ7ks6vPETFAYF5dyr.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/DbkYZ7ks6vPETFAYF5dyr.jpg\" \/><\/p><h3>3. Steelseries Arctis Pro Wireless<\/h3><div class=\"_hawk subtitle\"><p>Best high-end wireless gaming headset<\/p><\/div><p><strong>Drivers: <\/strong>40mm neodymium | <strong>Battery life: <\/strong>Up to 10 hours (per battery) | <strong>Frequency response: <\/strong>10Hz-40,000Hz | <strong>Impedance: <\/strong>32 Ohm | <strong>Features: <\/strong>USB audio station, virtual surround, swappable batteries, Bluetooth<\/p><div class=\"icon icon-plus_circle _hawk\">Brimming with features<\/div><div class=\"icon icon-plus_circle _hawk\">Battery hot-swapping<\/div><div class=\"icon icon-minus_circle _hawk\">Expensive<\/div><div class=\"icon icon-minus_circle _hawk\">Little sound difference to cheaper Arctic models<\/div><p>Old meets new in the Arctis Pro Wireless, in the best possible way. The wireless transmitter station is almost identical to that of the older Steelseries 800 model, acting as a controls menu, wireless receiver via lossless 2.4G and Bluetooth, and battery charging station all at once. Meanwhile, the new Arctis headset design (lavished with a more luxurious finish than the 7 series) delivers on the comfort and sound. Winning combination.\u00a0<\/p><p>Being able to pull out a depleted battery and switch it for a fully charged one from the transmitter box isn\u2019t just convenient, it\u2019s a game changer in the wireless market. And with battery life times in excess of 9 hours, it\u2019s perfectly viable to make use of that Bluetooth functionality by pairing the headset to a smartphone. Although the excellent sound quality is largely identical to the 7-series, there are sufficient bonuses in the Pro Wireless package to reinforce the extra spend. If you're looking for the top of the line, this is it.<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/87E5Xvpc9wngMWTn9f45xg-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/87E5Xvpc9wngMWTn9f45xg-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/87E5Xvpc9wngMWTn9f45xg-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/87E5Xvpc9wngMWTn9f45xg.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/87E5Xvpc9wngMWTn9f45xg.jpg\" \/><\/p><h3>4. Asus ROG Strix Wireless<\/h3><div class=\"_hawk subtitle\"><p>A great wireless headset for real surround sound<\/p><\/div><p><strong>Drivers: <\/strong>60mm | <strong>Battery life: <\/strong>Up to 10 hours | <strong>Frequency response: <\/strong>20Hz-20,000Hz | <strong>Impedance: <\/strong>32 Ohm | <strong>Features: <\/strong>USB audio station, true 7.1 surround, detachable mic, on-headset controls<\/p><div class=\"icon icon-plus_circle _hawk\">Distinctive looks<\/div><div class=\"icon icon-plus_circle _hawk\">True 7.1 surround<\/div><div class=\"icon icon-minus_circle _hawk\">A bit bass heavy<\/div><div class=\"icon icon-minus_circle _hawk\">COnfusing audio station<\/div><p>If you want to stand out while wearing a pair of headphones, these are your dream. An unorthodox design intended to highlight its enormous drivers defines the ROG STRIX. While it might look a bit Angry Birds, there\u2019s no denying the power of its low end or spaciousness of the surround sound.\u00a0<\/p><p>Front and rear 40mm drivers are joined by a 30mm center and two 20mms at the side and rear. Dolby\u2019s virtual surround gives true surround a real run for its money these days, but the articulation within these earcups is really impressive. And when you want a more precise sound, you can pare it down to a stereo headset\u2014if you can figure out the USB audio station\u2019s controls. \u00a0One problem, unfortunately: even in this configuration, the bass ends up being a shade too powerful for everyday use, and can muddy music by being overbearing.\u00a0<\/p><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/uwTcaTg9EoCTx6HRmaXS7H-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/uwTcaTg9EoCTx6HRmaXS7H-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/uwTcaTg9EoCTx6HRmaXS7H-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/uwTcaTg9EoCTx6HRmaXS7H.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/uwTcaTg9EoCTx6HRmaXS7H.jpg\" \/><\/p><h3>5. Astro Gaming A50<\/h3><div class=\"_hawk subtitle\"><p>An alternative from a wireless mainstay<\/p><\/div><p><strong>Drivers: <\/strong>40mm neodymium | <strong>Battery life: <\/strong>Up to 15 hours | <strong>Frequency response: <\/strong>20Hz-20,000Hz | <strong>Impedance: <\/strong>32 Ohm | <strong>Features: <\/strong>Charging base station, virtual surround, detachable mic<\/p><div class=\"icon icon-plus_circle _hawk\">High class looks<\/div><div class=\"icon icon-plus_circle _hawk\">Flawless construction<\/div><div class=\"icon icon-minus_circle _hawk\">Battery issues<\/div><div class=\"icon icon-minus_circle _hawk\">Annoying short cables<\/div><p>Another candidate that missed out by a whisker for our \u2018best overall\u2019 pick. Flawless in-game sound and bomb-proof build quality are the headline acts here, as you\u2019d expect of a model bearing the Astro name. The A50 might be geared up for consoles as much as PC, but there\u2019s no real drawback to that when you do get it setup with the latter. As with several other of our picks, the A50 comprises a headset and USB base station which acts as a set of controls as well as a wireless receiver\/transmitter. Few do it with quite this much style, though.<\/p><p>The chink in its armor is in the battery charge, however. There have been issues with battery draining while the headset\u2019s off, and with some users unable to properly charge in the first place. It\u2019s hard to gauge how widespread these issues are, but they do tarnish an otherwise exemplary package.<\/p><h3 id=\"how-we-test-wireless-headsets-and-competitors\">How we test wireless headsets and competitors<\/h3><p>Many of the qualities you\u2019re looking for from a wireless headset are the same you\u2019d hope to find in any audio equipment\u2014tone, build quality, and reliability leading the charge. As such we listen to each review model while playing different genres of game, listening to music, and watching movies with bombastic sound effects and surround mixes\u2014think less Werner Herzog, more Chris Nolan. We also run a simple sine wave \u2018swoop\u2019 across the stated frequency response range (almost always the full 20Hz-20KHz these days), and in the case of surround headsets we\u2019ll listen to positional audio tests like <a href=\"http:\/\/listen.dts.com\/pages\/headphone-x\"><u>DTS Headphone-X test<\/u><\/a>. There\u2019s also our old favorite, <a href=\"https:\/\/www.youtube.com\/watch?v=IUDTlvagjJA\"><u>the Virtual Barber Shop<\/u><\/a>. YouTube\u2019s compression does limit the overall sound quality, but it\u2019s still a great way of separating the wheat from the chaff in surround sound earphones.<\/p><p>There are a few wireless-specific elements we need to test for, too: battery life, charge time, range and latency. The former is pretty self-explanatory, though in addition to an \u2018everyday use\u2019 battery life test we also run the headset at full volume to discover how quickly the charge drains under those conditions. To ascertain charge time, we\u2026 well, we charge the headsets and note how long it takes.<\/p><p>Range and latency are trickier to test in a scientific manner. However, having a good old walk around the house gives a good indication of range, and latency ultimately comes down to perception. With all that taken into account after several days of use, we\u2019re in a good place to make the call on a headset.<\/p><p>Most of the big players in USB\/3.5mm gaming headsets have a wireless option, but usually just one. As such the current market competitors list is a bit slim. The range expands when you look as far as console-specific wireless cans, but in the interest of ensuring full compatibility we\u2019ve stuck to officially supported PC models.\u00a0<\/p><p>The wireless end of PC gaming audio gear offers less choice than that of wired peripherals, but it\u2019s still a big marketplace\u2014and it\u2019s populated by models with incremental improvements and price hikes from the same manufacturers. We haven\u2019t tested every single model available, but done our best to seek out the cream of the crop. And from that cream, we\u2019ve cherry-picked a) the absolute best options available right now, and b) a really odd, food-themed mixing of metaphors.<\/p><p>It\u2019s also a marketplace that moves fast\u2014so we\u2019ll be keeping our eye on it and updating it as promising new models are released.<\/p><p><em>Some online stores give us a small cut if you buy something through one of our links. Read our<\/em> <a href=\"https:\/\/www.pcgamer.com\/a-note-on-affiliate-links\/\"><em>affiliate policy<\/em><\/a> <em>for more info.<\/em><\/p>","created_at":"2018-11-07T16:45:17+0000","updated_at":"2018-11-07T16:45:17+0000","published_by":["Phil Iwaniuk"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"www.pcgamer.com","preview_picture":"https:\/\/cdn.mos.cms.futurecdn.net\/bDJDFyN66MRLCwdYP7kteX-1200-80.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":339,"title":"The Gramarye","url":"http:\/\/zork.net\/~nick\/loyhargil\/gramarye.html","content":"The Gramarye\n<h3 style=\"text-align: center;\"> The Gramarye <br \/>A Magic System for FUDGE <br \/>by Carl D. Cravens <\/h3>\n<p style=\"text-align: center;\">Gramarye version 1.0, Copyright \u00a9 1997 by Carl\nD. Cravens.<\/p>\n<p><em><strong>Thanks to:<\/strong> Karen Cravens, Michael Feldhusen,\nJason Knight, Scott Raun, and the users of the <a href=\"http:\/\/fudge.phoenyx.net\/faq.html\">fudge-l mailing\nlist<\/a>.<\/em>\n<\/p>\n<p>\n<strong>Note:<\/strong> The original version of <b>The\nGramarye<\/b> was first published in <a href=\"http:\/\/www.fudgerpg.com\/fudge\/medley.html\"><i>A Magical\nMedley<\/i><\/a> by <a href=\"http:\/\/www.fudgerpg.com\/\">Grey Ghost Press,\nInc.<\/a> See <a href=\"http:\/\/zork.net\/~nick\/loyhargil\/license.html\">\nhttp:\/\/fudge.phoenyx.net\/gramarye\/license.html<\/a> for information on\nreproducing <b>The Gramarye<\/b>.\n<\/p>\n<hr \/><p> <b>gramarye<\/b> (GRAM-uh-re) n. [Archaic] magic; occult knowledge\n<\/p>\n<h3>Introduction<\/h3>\n<p> The magic system presented here represents the pseudo-scientific approach\nof a mythical medieval alchemy. Many GMs running games in pseudo-medieval\nEuropean settings will find this system adequate for their uses with little\nmodification. <\/p>\n<h4>The World-View<\/h4>\n<p> The world this system is designed for is similar to a mythical medieval Europe.\nThe typical sorcerer sees magic as a scientific force to be manipulated through\nscientific (to him) means. He divides the world into four basic realms --\nsolids, liquids, gases, and energy -- which he calls earth, water, air, and\nfire. All things are made up of one or more of these elements. (A man's flesh\nis made of earth, his blood of water, his breath of air, and his spirit of\nfire.) He divides living things into three different realms as well -- body,\nmind, and spirit. This world-view governs the way a sorcerer approaches magic.\n<\/p>\n<h4>How the Magic Works<\/h4>\n<p> The basic elements of magic are divided up into Realms (areas of control,\nsuch as Fire, Water, Spirit, and Body) and Colleges (types of control, such\nas Control, Healing, Creation, and Transformation). In order to cast spells\nthat manipulate one of these Realms, a character must have a skill for that\nRealm and a skill for the College that he desires an effect from. <\/p>\n<p> To cast a spell, the character chooses a College and a Realm and combines\nthem for the desired effect. The character's effective skill level for the\nspell is that of the lowest skill he is using. If a character has more than\none Realm that would affect the spell (e.g. Transform Body to Stone when\nthe caster has both Body and Earth), use the higher Realm skill with a +1\nlevel bonus to the skill. <\/p>\n<h3>Basic Rules<\/h3>\n<p> The character must have a knowledge skill (Easy) of his type of magic --\nsuch as Rune Lore, Sorcery, Elven Shamanism, Gypsy Dance Magic, etc. -- which\nis equal to or greater than his highest magic skill. This skill represents\nthe background knowledge that the mage must have before learning to cast\nspells. <\/p>\n<p> At the GM's option, a character may take only a sub-College or sub-Realm\nfor a decrease in the difficulty of the skill. <\/p>\n<p>Example: The Animal Realm allows the caster to affect all animals and is\nan Average skill. A character could take Animal\/Birds as an Easy skill and\nbe able to affect only birds. (The Animal Realm is optional; the GM may require\nthe purchase of sub-Realms of Body\/Animal, Mind\/Animal, and Spirit\/Animal\n-- all Easy skills -- to replace it.) <\/p>\n<h3>The Colleges and Realms<\/h3>\n<p> These Colleges and Realms are based on the world-view of the setting. See\nModifying the System, p. 95, for a discussion on altering the Colleges and\nRealms available in a given campaign. <\/p>\n<h4>Colleges<\/h4>\n<p> <b>Breaking (e.g. weaken, harm, dispel)<\/b> <br \/>Breaking is destruction and undoing. Almost anything can be broken, though\nsome things are harder to break than others. <\/p>\n<p> <b>Communication (e.g. speak, read)<\/b> <br \/>Communication is about willing communication. Thought Speech (Communicate\nwith Mind) is okay, but Read Minds is a Knowledge spell instead (Knowledge\nof Mind). <\/p>\n<p> <b>Control (H) (e.g. bind)<\/b> <br \/>Control is making something perform an action that it could perform of its\nown volition or through natural events. It cannot make a stream of water\nflow up-hill, for example. Control can, if the GM allows, influence natural\nevents, such as causing an earthquake along a fault line or causing a fire\nto spread more quickly. <\/p>\n<p> <b>Creation (H)<\/b> <br \/>Creation is making things. Create combined with any of the life-based Realms\nwill be off-limits in most campaigns. Create Spirit (or any other life) is\nusually something restricted to the gods, for instance. <\/p>\n<p> <b>Enhancement (e.g. strengthen, repair)<\/b> <br \/>Enhancement is making something better. It can repair broken items, make\na man stronger or faster, or grant a \"blessing\" that gives the party an edge\nin combat. Enhancement is usually temporary. <\/p>\n<p> <b>Healing (H) (e.g. fertility, growth)<\/b> <br \/>Healing is the repair, growth, and reproduction of living things. <\/p>\n<p> <b>Knowledge (H) (e.g. see, hear, find)<\/b> <br \/>Knowledge is learning things that are unknown. It allows the caster to see\nthings across long distances, read minds, or learn the history of an object\nor the true name of a dragon. <\/p>\n<p> <b>Movement<\/b> <br \/>Movement is the transportation of things through space. Examples are\npsychokinesis, flying, moving faster than normal, and teleportation. <\/p>\n<p> <b>Protection<\/b> <br \/>Protection is preventing the intrusion of damaging objects, spying spells,\nunwanted visitors, the weather, etc. <\/p>\n<p> <b>Transformation (H) (e.g. shape, transmute)<\/b> <br \/>Transformation is changing the size, shape, or chemical composition of something.\nThe GM should keep strict control over the Transformation College; it should\nnot be able to do things that are the domains of other Colleges. It cannot\nheal, hurt, control, or otherwise change anything but the physical nature\nof the target. Transformation is usually permanent. The more you transform\nsomething from its original state, the harder the spell. Transforming a man\ninto a goblin is fairly easy; they're both living humanoids. Transforming\na man into a mouse is harder, requiring both a mass and form change. Transforming\nthe man into a statue is even harder, requiring an element change. Transforming\nthe man's body into stone while leaving him mobile is harder yet because\nthis is a natural state for neither man nor stone. <\/p>\n<h4>Realms<\/h4>\n<p> <b>Air (e.g. gas)<\/b> <br \/>Air is any gaseous substance. <\/p>\n<p> <b>Animal<\/b> <br \/>Animal is any non-intelligent living creature. It overlaps with Body, Mind,\nand Spirit. <\/p>\n<p> <b>Body (H)<\/b> <br \/>Body is the physical body of any living creature. <\/p>\n<p> <b>Earth (H) (e.g. solid)<\/b> <br \/>Earth is any solid, non-living substance. <\/p>\n<p> <b>Fire (e.g. energy)<\/b> <br \/>Fire is any form of energy. <\/p>\n<p> <b>Illusion (H) (e.g. images, light, darkness, sound)<\/b> <br \/>Illusion is any form of sensory or mental illusion. <\/p>\n<p> <b>Magic (H) (e.g. mana)<\/b> <br \/>Magic is pure mystical force. If it can't be done with any other Realm, it\ncan be done with Magic. Beyond the obvious (Detect Magic, Dispel Magic),\nMagic is a catch-all area. If a mage wants to affect something and that something\nisn't one of the Realms, the mage probably needs to use Magic. A Telekinesis\nspell that can move anything is Move with Magic. The Magic Realm is flexible\nand can reproduce effects from other Realms (although with slightly different\nresults). For instance, Create Dog using Create Animal will produce a real\ndog while Create Magical Dog using Create Magic will produce a magical dog\nwith magical properties. (See Guardian Beast in the sample spells.) Spells\ncreated through the Magic Realm that reproduce effects from other Realms\nshould cost more mana, be more difficult, and\/or have radically different\neffects. The Magic Realm should not be a general substitute for other Realms.\n<\/p>\n<p>\n<b>Mind (H)<\/b> <br \/>Mind is the mind of any living creature. <\/p>\n<p> <b>Plant<\/b> <br \/>Plant is any plant matter, living or dead. <\/p>\n<p> <b>Spirit (H)<\/b> <br \/>Spirit is the life-force of any living creature. The Spirit College works\nwell for creating zombies, golems, or any other animated thing, provided\nthe setting allows for summoning and binding a spirit to animate it. <\/p>\n<p> <b>Time (H) (e.g. past, future, history)<\/b> <br \/>Time is the past, present, and future. The GM should keep careful control\nover this Realm. <\/p>\n<p> <b>Water (e.g. liquid)<\/b> <br \/>Water is any liquid substance. <\/p>\n<h3>Mana<\/h3>\n<p> It is assumed that characters (or at least mages) have a Mana attribute or\ngift which has a numerical value and represents the amount of mana the character\nhas available at one time (excluding magical devices). The character's mana\nis used up as he casts spells, and it recovers at a moderate rate. (In most\ncases, a wizard who exhausted his mana supply would recover it in a matter\nof minutes.) The recommended method for this system is a Mana attribute,\nwith Terrible = 1 mana and Superb = 7 mana, and a recovery rate of 1 mana\nevery 25 seconds (5 combat rounds). <\/p>\n<p> The basic system presented here assumes a certain level of power and frequency\nof magic. These levels can be easily adjusted by changing the base difficulty\nof spells, changing the cost of magical skills, making a gift of Magical\nAptitude a prerequisite to buying magical skill, or changing the availability\nand recharge rate of mana. <\/p>\n<p> Adjusting the amount of available mana is the GM's primary tool for adjusting\nthe overall power level of the system. For example, if the GM wants to eliminate\nthe need for props and lower the average casting times of spells, she simply\nneeds to raise the amount of mana available. <\/p>\n<h3>Choosing the Elements<\/h3>\n<p> The player first determines what he wants the spell to do, then looks at\nthe lists of Colleges and Realms for a pair that describes (very roughly)\nthe desired effect. The player should realize that not all spell effects\ncan be wrangled out of a limited set of Colleges and Realms. For example,\nCreate Animal is not a substitute for Locate (Know) Animal followed by Control\nAnimal; Create creates things. <\/p>\n<p> There are times when more than one Realm may produce the desired effect.\nIce is a solid and therefore affected by Earth, but it is also water and\naffected by Water. It's up to the GM to decide where she'll let this overlap\noccur. Some GMs might decide that ice is only water and is unaffected by\nEarth spells, or that ice is no longer a liquid and unaffected by Water spells.\n<\/p>\n<p> Also note that Move Body, for example, is not a spell; it is a combination\nof a College and Realm. The player should name the spell he desires before\ntelling the GM which elements he will use. A Haste spell, which makes the\ntarget act faster, would require Move Body. The player would say, \"I cast\nHaste using Move Body,\" not \"I cast Move Body.\" <\/p>\n<p> The player may want to keep a \"spell book\" of the spells his character uses\noften, and to give a copy to the GM. This way the spell doesn't have to be\ndescribed every time the character casts it. This will also save the time\nof recalculating the spell's casting parameters. <\/p>\n<p> The GM may require more than one Realm for certain effects (e.g., Transform\nBody to Stone may require both the Body and Earth Realms). Sometimes a spell\nwill consist of two or more spell effects strung together. A Summon Wolf\nspell might require two separate skill rolls -- Locate (Know) Wolf and Control\nWolf. <\/p><p> <em>Example: Baldor the Brilliant wants to cast a huge ball of fire at a group\nof advancing warriors. He combines his knowledge of Creation with his knowledge\nof Fire to throw a Fireball spell.<\/em> <\/p>\n<h3>Putting the Spell Together<\/h3>\n<p> A basic spell has close range, does zero damage or some other minor effect,\naffects one man-sized target or one cubic meter of material, takes five seconds\nto cast, lasts for five seconds (if appropriate), costs zero mana, has a\nDifficulty of Fair, and requires the caster to make magical gestures or\nincantations. <\/p>\n<p> The final cost, Difficulty, and casting time of a spell is determined by\nthe GM. See Spell Effects, below and right, for rough guidelines. (These\nare guidelines to help choose basic costs and casting times; they are not\nstrict rules.) The GM should not allow any spell that she thinks may be\nunbalancing. The GM may declare that certain spells' Difficulty Levels may\nnot be lowered below a certain point for game balance purposes. For example,\nthe GM may declare that Knowledge spells can never be easier than Fair,\nregardless of the time or mana spent casting them. Likewise, the GM may decide\nthat certain spells require a certain length of casting time or a certain\namount of mana regardless of resources available to the caster. <\/p>\n<p> It's difficult to give detailed guidelines for determining a spell's basic\nlevel of power. The best way to do so is to compare the desired effect with\nan existing spell or the suggested Spell Effects. (Note that the level of\neffect of a particular spell is influenced by the world. For example, if\ntalking to spirits is a common occurrence in your world, it probably shouldn't\nbe of Great Difficulty.) <\/p>\n<p> Once the GM determines the basic cost for a desired effect, see the Modifiers\ntable (p. 90) to refine the cost, duration, exact spell effect, etc. of a\nspecific spell. <\/p>\n<p> <em>Example: Back to Baldor's Fireball spell. Since he wants to incapacitate\nthese goons, he figures the fireball will need to do 10 damage. According\nto the Modifiers chart, that's +5 mana. He's in a hurry, so he won't mess\nwith the casting time, but it's long range (+2 mana) and will affect several\ntargets (+2 mana). The base Difficulty is Fair. Total mana required: 9. Since\nBaldor only has a Great Mana (6), he needs to come up with three more mana\nto help him cast the spell. <\/em> <\/p>\n<h4>Spell Effects<\/h4>\n<table style=\"border-spacing: 0px; width: 100%;\"><tr><td style=\"width: 15%;\">Cost<\/td>\n    <td>Spell Effect Is:<\/td>\n  <\/tr><tr><td>+0-1<\/td>\n    <td><b>Mediocre<b><\/b><\/b><\/td>\n  <\/tr><tr><td><em>Breaking<\/em>: Shatter fragile material (crystal). Cause minor discomfort.<\/td>\n  <\/tr><tr><td><em>Communication<\/em>: Speak local dialect of known language. Empathy.<\/td>\n  <\/tr><tr><td><em>Control<\/em>: Cause target to sneeze or itch.<\/td>\n  <\/tr><tr><td><em>Creation<\/em>: Simple illusion (simple stationary image or sound).<\/td>\n  <\/tr><tr><td><em>Enhancement<\/em>: Amplify voice. Make a person beautiful. Repair fragile material.<\/td>\n  <\/tr><tr><td><em>Healing<\/em>: Relieve minor fatigue or Scratches. Heal minor infection.<\/td>\n  <\/tr><tr><td><em>Knowledge<\/em>: Provide information obtainable through simple observation.<\/td>\n  <\/tr><tr><td><em>Movement<\/em>: Cause fire to spread more rapidly.<\/td>\n  <\/tr><tr><td><em>Protection<\/em>: Shield from weather. Act as cloth armor.<\/td>\n  <\/tr><tr><td><em>Transformation<\/em>: Make a man look like another man.<\/td>\n  <\/tr><\/table><br \/><table style=\"border-spacing: 0px; width: 100%;\"><tr><td style=\"width: 15%;\">+2-3<\/td>\n    <td><b>Fair<\/b><\/td>\n  <\/tr><tr><td><em>Breaking<\/em>: Shatter or tear soft material (leather, wood).<\/td>\n  <\/tr><tr><td><em>Communication<\/em>: Speak unknown language. Telepathy with like mind.<\/td>\n  <\/tr><tr><td><em>Control<\/em>: Cause target to do something it's inclined to do.<\/td>\n  <\/tr><tr><td><em>Creation<\/em>: Detailed illusion (stationary person).<\/td>\n  <\/tr><tr><td><em>Enhancement<\/em>: Temporarily increase one of target's attributes by one\n      level..Repair soft material.<\/td>\n  <\/tr><tr><td><em>Healing<\/em>: Cure major infection or minor disease.<\/td>\n  <\/tr><tr><td><em>Knowledge<\/em>: Provide information obtainable through simple mundane means.<\/td>\n  <\/tr><tr><td><em>Movement<\/em>: Move something telekinetically. Allow target to levitate. Allow\n      target to cling to walls like a spider. Cause target to travel faster.<\/td>\n  <\/tr><tr><td><em>Protection<\/em>: Shield from severe storm and light weapons.<\/td>\n  <\/tr><tr><td><em>Transformation<\/em>: Turn a man into a goblin.<\/td>\n  <\/tr><\/table><br \/><table style=\"border-spacing: 0px; width: 100%;\"><tr><td style=\"width: 15%;\">+4-7<\/td>\n    <td><b>Good<b><\/b><\/b><\/td>\n  <\/tr><tr><td><em>Breaking<\/em>: Shatter resistant material (stone, bone, fragile metal).<\/td>\n  <\/tr><tr><td><em>Communication<\/em>: Speak with being with no language (animals). Telepathy\n      with alien mind.<\/td>\n  <\/tr><tr><td><em>Control<\/em>: Cause target to do something it's mildly against.<\/td>\n  <\/tr><tr><td><em>Creation<\/em>: Complex illusion (person or dragon fighting).<\/td>\n  <\/tr><tr><td><em>Enhancement<\/em>: Temporarily increase one of target's attributes by two levels.\n      Repair resistant material.<\/td>\n  <\/tr><tr><td><em>Healing<\/em>: Cure major disease or poison.<\/td>\n  <\/tr><tr><td><em>Knowledge<\/em>: Provide information the caster could have obtained through\n      complex, lengthy, or dangerous mundane means.<\/td>\n  <\/tr><tr><td><em>Movement<\/em>: Allow target to fly. Allow target to walk through walls.<\/td>\n  <\/tr><tr><td><em>Protection<\/em>: Shield from heavy weapons.<\/td>\n  <\/tr><tr><td><em>Transformation<\/em>: Turn a man into a mouse.<\/td>\n  <\/tr><\/table><br \/><table style=\"border-spacing: 0px; width: 100%;\"><tr><td style=\"width: 15%;\">+8-15<\/td>\n    <td><b>Great<\/b><\/td>\n  <\/tr><tr><td><em>Breaking<\/em>: Shatter hard material (hardened metal).<\/td>\n  <\/tr><tr><td><em>Communication<\/em>: Speak with spirits, trees, or rocks.<\/td>\n  <\/tr><tr><td><em>Control<\/em>: Cause target to do something it's strongly against.<\/td>\n  <\/tr><tr><td><em>Creation<\/em>: Very complex illusion (band of knights fighting a dragon).<\/td>\n  <\/tr><tr><td><em>Enhancement<\/em>: Temporarily increase one of target's attributes by three\n      levels. Repair hard material.<\/td>\n  <\/tr><tr><td><em>Healing<\/em>: Reattach severed limbs.<\/td>\n  <\/tr><tr><td><em>Knowledge<\/em>: Provide information the caster could not have obtained otherwise.<\/td>\n  <\/tr><tr><td><em>Movement<\/em>: Allow target to move from one location to another instantly.<\/td>\n  <\/tr><tr><td><em>Protection<\/em>: Shield from most weapons.<\/td>\n  <\/tr><tr><td><em>Transformation<\/em>: Turn a man into stone.<\/td>\n  <\/tr><\/table><br \/><table style=\"border-spacing: 0px; width: 100%;\"><tr><td style=\"width: 15%;\">+16-31<\/td>\n    <td><b>Superb<\/b><\/td>\n  <\/tr><tr><td><em>Breaking<\/em>: Shatter mythical material (mithral).<\/td>\n  <\/tr><tr><td><em>Communication<\/em>: Speak with extra-planar being<\/td>\n  <\/tr><tr><td><em>Control<\/em>: Cause target to do something it would never do under its own\n      will.<\/td>\n  <\/tr><tr><td><em>Creation<\/em>: World-altering illusion (change entire setting).<\/td>\n  <\/tr><tr><td><em>Enhancement<\/em>: Temporarily increase one of target's attributes by four\n      levels. Repair mythical material.<\/td>\n  <\/tr><tr><td><em>Healing<\/em>: Regrow severed limbs.<\/td>\n  <\/tr><tr><td><em>Knowledge<\/em>: Provide information no one could have obtained otherwise.<\/td>\n  <\/tr><tr><td><em>Movement<\/em>: Allow target to move from one plane to another.<\/td>\n  <\/tr><tr><td><em>Protection<\/em>: Shield from most magical effects.<\/td>\n  <\/tr><tr><td><em>Transformation<\/em>: Turn a man into living stone.<\/td>\n  <\/tr><\/table><br \/><table style=\"border-spacing: 0px; width: 100%;\"><tr><td style=\"width: 15%;\">+32-<\/td>\n    <td><b>Legendary<\/b><\/td>\n  <\/tr><tr><td> These spells are earth-shaking and rare. They destroy powerful\n         entities, wreck nations, sink continents, and unleash terrible\n         evils on the world. <\/td>\n  <\/tr><\/table><h4>Modifiers<\/h4>\n<p> (add all that apply to specific spell's mana cost) <\/p>\n<b>Change the number of targets or area covered:<\/b>\n<ul><li>\n    -1 Small -- a bread box\n  <\/li><li>\n    +0 One -- one person\n  <\/li><li>\n    +1 Few -- 2-4 people, a closet-sized space\n  <\/li><li>\n    +2 Several -- 5-10 people, a whole room\n  <\/li><li>\n    +4 Many -- a small crowd, a small building\n  <\/li><li>\n    +8 Great -- large crowd, a large building (or more)\n<\/li><\/ul><p> Note: A spell that affects multiple targets should not be selective; it affects\neveryone in the area in which it is cast. In other words, the mage shouldn't\nget to selectively shoot multiple targets for +1 mana to avoid casting the\nspell once for each target. <\/p>\n<ul><li>\n    <b>Change the range:<\/b>\n<ul><li>\n    -1 Touch -- must touch target, requires to-hit roll\n  <\/li><li>\n    +0 Close -- in close fighting range\n  <\/li><li>\n    +1 Short -- within thrown weapon range\n  <\/li><li>\n    +2 Long -- within sight or powered weapon range\n  <\/li><li>\n    +4 Extreme -- out of sight of the caster (or more)\n<\/li><\/ul><\/li><li>\n    <b>+2 damage:<\/b> +1 mana\n  <\/li><li>\n    <b>+1 defense:<\/b> +1 mana\n  <\/li><li>\n    <b>Has no mundane defense<\/b> (armor, etc. doesn't stop the damage or effect): +2\n    mana\n  <\/li><li>\n    <b>Heal all wounds by 1 level:<\/b> +2 mana\n  <\/li><li>\n    <b>Each level of Health sacrificed:<\/b> -2 mana\n  <\/li><li>\n    <b>Increase Duration one level on Time Chart<\/b> (see below): +1 mana\n  <\/li><li>\n    <b>-1 level Difficulty:<\/b> +2 mana, or increase Casting Time level on Time Chart\n    (see below)\n  <\/li><li>\n    <b>+1 level Difficulty:<\/b> -2 mana, or decrease Casting Time level on Time Chart\n    (see below)\n  <\/li><li>\n    <b>Using appropriate props:<\/b> An appropriate decrease in mana cost\n  <\/li><li>\n    <b>Not using gestures or incantations:<\/b> Double mana cost (after\n    all other bonuses and penalties to mana cost, before prop modifiers)\n<\/li><\/ul><h4>Time Chart<\/h4>\n<p> Each level of the Time Chart is a rough approximation of five times the previous\nlevel. Use this chart to determine time modifiers both for spell duration\nand casting time. <\/p>\n<p>1 level = 2 mana or 1 Difficulty Level.\n<\/p><ul><li>1 Second<\/li>\n<li>5 Seconds (one combat round)<\/li>\n<li>25 Seconds (five combat rounds)<\/li>\n<li>2 Minutes<\/li>\n<li>10 Minutes<\/li>\n<li>1 Hour<\/li>\n<li>5 Hours<\/li>\n<li>1 Day<\/li>\n<li>1 Week<\/li>\n<li>1 Month<\/li>\n<li>1 Season<\/li>\n<li>1 Year<\/li>\n<\/ul><p> Extending the duration of a spell is at the GM's option; some spells are\ntoo powerful or make no sense if they last longer than a moment. <\/p>\n<h3>Props<\/h3>\n<p> Props are things the mage does or uses to help make the spell work. Props\nmay take the form of rituals, invocations, and gestures; or they may be physical\nprops that apply the magical laws of Similarity, Contact, or Sacrifice to\nprovide the caster with mana bonuses. <\/p>\n<p> Physical props are normally consumed in the casting of the spell. To create\na prop that is not consumed by the casting, the mage must enchant it. (See\nMaking Sources of Power and other Wizard Tools.) <\/p>\n<h4>Rituals, Invocations, and Gestures<\/h4>\n<p> A ritual is something the caster does to cast the spell before he actually\nbegins casting. Sample portions of rituals include fasting, special cleansing\nor confession, performing a special service or quest, praying at a specific\nlocation, etc. A simple ritual grants the mage a +1 mana bonus. The more\ninvolved and inconvenient the ritual, the greater the bonus the GM should\naward. <\/p>\n<p> Invocations are magical words or prayers spoken aloud during casting. Gestures\nare magical signs or motions made by the caster with his hands or body during\ncasting. To cast a spell with neither gestures nor invocations doubles the\nmana cost. (Casting a spell with both provides no bonus.) <\/p>\n<h4>Magical Laws and Physical Props<\/h4>\n<p> The mana cost of a spell can be paid for, in part or whole, by applying the\nLaws of Similarity, Contact, and Sacrifice. The Law of Similarity says that\nsimilar things affect each other, like produces like, or effect resembles\ncause. The Law of Contact says that things once in contact with each other\nstill have an effect on each other when they are separated. The Law of Sacrifice\nsays that giving up something of value to the caster or the target makes\nthe magic more likely. <\/p>\n<h4>The Law of Similarity<\/h4>\n<p> To apply the Law of Similarity, one must use material components that are\nsimilar to the desired effect. Examples are a feather for a flight spell,\na bit of sulfur for create flame, a drawing of a wolf to summon a real wolf,\na twig to be broken during casting for a spell that breaks down the wooden\ncastle gate, etc. <\/p>\n<p> The Law of Similarity depends completely on world-view. If the mage does\nnot view a particular item as being relevant to his spell, then that item\nis not Similar, regardless of the knowledge the player has. For instance,\nthe player may know that sulfur burns readily, but if his character knows\nnothing of this, a pinch of sulfur will do nothing for his flame spells. <\/p>\n<h4>The Law of Contact<\/h4>\n<p> To apply the Law of Contact, one must use material components that have once\nbeen in contact with the target of the spell; the longer or more personal\nthe contact, the better. Examples are the victim's hair and nail parings\nused in a death spell, a piece of jewelry worn by an enemy in a spell of\ndefense to protect the mage from that enemy, a piece of the castle gate to\nbe broken with a spell, etc. <\/p>\n<h4>The Law of Sacrifice<\/h4>\n<p> To apply the Law of Sacrifice, the component must be of personal value to\neither the caster or the target of the spell and fulfill one of the other\nLaws. Examples are a family ring, a lover's lock of hair, the tooth of a\ndragon, etc. (A dragon's tooth is valuable to the caster if he had to go\nto great pains to get it. It is hardly valuable if he walked down to the\ncorner store and picked one out of a bowl full of them.) <\/p>\n<h4>Determining Bonuses of Physical Props<\/h4>\n<p> The same object may meet the requirements of both of the first Laws, such\nas hair clippings used on a voodoo doll. The hair of the victim fulfills\nthe Law of Contact while the doll itself fulfills the Law of Similarity.\nUsing props will provide extra mana for the spell. The caster may use as\nmany different props as the GM feels reasonable, but duplicates (props that\nserve identical functions) are generally disallowed. <\/p>\n<pre>Base value:\nSimilar prop (Law of Similarity):       +1 mana\nContagious prop (Law of Contact):       +2 mana\nValue multiplier (multiply by base value):\nWorthless:      x 1 (the feather of a sparrow)\nMinor value:    x 2 (the feather of an eagle)\nMajor value:    x 3 (the feather of a gryphon)\nPriceless:      x 4 (the feather of unique creature)\nPersonal meaning bonus (add to multiplier):\nMeaningless to owner:   +0 (hair clippings)\nMinor meaning to owner: +1 (a ring)\nMajor meaning to owner: +2 (a ring given by a lover)\nPriceless to owner:     +3 (a hand, severed)\n<\/pre>\n<p> <em> Example: Using a gold chain stolen from the target, given to him by his wife\non their wedding night, a mage casts a spell of Binding (mind control) on\nhis target. Since the mage is binding (chaining) the target to his will,\nthe chain is Similar (+1); since the target wore it, it is Contagious (+2).\nThe chain is of minor value but has major meaning to the owner (x 4). The\ntotal value of the chain is 3 x 4 = 12 (the amount of mana it contributes\nto the spell -- a powerful prop!). If the chain has been enchanted properly,\nit can be used more than once; otherwise it is destroyed (or made magically\nworthless). <\/em> <\/p>\n<p> <em> Example: As previously stated, Baldor needs 9 mana to cast his Fireball spell,\nbut he has only 6. Luckily, he has with him a supply of small balls of bat\nguano mixed with sulfur, made just for this purpose. These are Similar (they\nburn readily) and gain him +1 mana. But he is still 2 mana short and cannot\ncast the spell without finding another source of mana. Fortunately for Baldor,\nhe has a 2-point mana stone (see Making Sources of Power, p. 93) to make\nup the difference. <\/em> <\/p>\n<h3>Casting the Spell<\/h3>\n<p>To begin casting a spell, the mage must have the desired components at hand,\nbe able to recite or perform any incantations or gestures, and have the necessary\nmana available. <\/p>\n<p> Once he has started casting the spell, an unintentional interruption of the\nspell results in failure as though he had failed his skill roll. If the mage\nis wounded, he should make a Willpower (or similar) check to avoid interruption\n(wound penalties apply; -1 for Hurt, -2 for Very Hurt). The mage may voluntarily\nrelease the spell at any time before actually finishing it at no penalty.\nThe source of mana -- whether the mage himself or some other source -- does\nnot recover mana while casting a spell. When the spell is complete, the player\nmust make a skill check against the Difficulty of the spell for the spell\nto succeed. <\/p>\n<p> Once cast, the mage may \"hold\" the spell and not release it right away. It\ncosts 1 mana every five combat rounds to hold a spell, during which time\nthe mage may cast no other spells and he, or any device providing him mana,\nregains no mana. <\/p>\n<p> Spells normally do not require a to-hit roll and get no bonuses to damage\nfor relative degree of success, that being the nature of magic. A spell either\ndoes its damage or it does not. If the GM desires, she may require to-hit\nrolls for all targeted spells and may allow increased damage based on the\nrelative degree. <\/p>\n<p> If a spell may have an extended duration, the mage may continue to maintain\nthat spell at a cost of one mana per base time unit of the spell. The base\ntime unit is the amount of time the spell was set to last when cast. There\nis no limit to the number of spells a mage can maintain except his available\nmana. The mage does recover mana while maintaining spells. <\/p>\n<h3>Spell Failure<\/h3>\n<p> The game master may use penalties for spell failure as a means to control\nthe more powerful spells. A player who would risk a 90% chance of failure\non an extremely powerful spell would be more reluctant if failure probably\nmeant the death of his character . Spell failure penalties are optional.\nIf it is extremely dangerous to use magic, this can greatly change the tone\nof the campaign. If the GM wishes to use penalties for spell failure, the\nrisks that their characters may take should be made clear to the players. <\/p>\n<p> Generally, the more powerful the spell, the greater the penalty for failure.\nExamples of penalties are loss of mana, loss of consciousness, temporary\nloss of magical ability, damage, unexpected spell effects, etc. The penalty\nshould be related to the type of spell cast. <\/p>\n<h3>Defense Against Spells<\/h3>\n<p> A character gets no defensive roll against a spell; whether or not he gets\nhit is a function of the mage's success at casting it. A character may take\nmultiple levels of the gift Magic Resistance, which acts like armor against\nmagical damage and reduces the effects of other magic at the GM's discretion.\n(The GM may allow Magic Resistance to reduce the mage's chances of success\nwith some spells.) <\/p>\n<h3>Creating Magic Items<\/h3>\n<p> This magic item creation system is simply a suggestion. The GM is encouraged\nto modify or replace the system as necessary to adapt it to her own campaign\nand personal preferences. <\/p>\n<p> To create a magic item, the mage must perform lengthy rituals involving the\nColleges and Realms necessary for the effects he wants the item to have.\nFor instance, to create a sword that increases the wielder's skill would\nrequire Enhancement and Body. <\/p>\n<p> A good guideline for item creation time is to take the mana cost of the desired\nspell and square it. The mana needed at the end of creation is the mana cost\nof the spell doubled. (If the final mana cost of the desired spell is 5,\nit would take 25 days to create a magic item that could cast that spell and\nwould cost 10 mana.) This produces a magical item that allows a person to\ncast the item's spell at the same casting time, mana cost, and other requirements\nas normal. Essentially, it grants the user the ability to cast that particular\nspell as if he were a mage with the proper skills, even if he isn't. To create\nan item that requires no mana, double the creation time. <\/p>\n<p> To create an item that requires no skill roll, double it again. If the player\nsuggests limitations on the item, such as single-use, the GM is encouraged\nto reduce the creation time and mana requirements. <\/p>\n<p> When the character has reached the end of the creation period, the chance\nof successfully finishing the item is the same for successfully casting the\nbasic spell. If the item has more than one spell, roll for each spell as\nit is enchanted into the item. To increase the chance of success, the caster\ncan modify the casting cost of the creation in the same ways as an ordinary\nspell. <\/p>\n<h3>Making Sources of Power and Other Wizard Tools<\/h3>\n<p> Mana stones and other items that store mana are important tools. They provide\nthe mana necessary to cast large spells, or to cast a large number of spells\nin a short time. The suggested average for a mana stone is three mana. <\/p>\n<p> Creating a mana stone usable with any form of magic requires Control and\nMagic for a Bind Mana spell. The more mana the stone can hold, the more mana\nthe Bind Mana spell will require. To create a mana stone that works with\na specific kind of magic require s only that particular College or Realm.\n(Thus a character with only Plant and Animal magic could create a \"mana stone\"\nuseable only with the Plant and Animal Realms.) <\/p>\n<p> A mana stone recharges at the same rate as the caster's personal mana. A\ncharacter may carry any number of mana stones, but can use only one at a\ntime. The user will recover all of his mana before any of his stones recharge,\nand when in close proximity to other stones, only one mana stone will recharge\nat a time, largest to smallest. <\/p>\n<p> Another important tool is the item which reduces the mana cost of every spell\nof a specific type. Generally, these are enchanted props. Their mana value\nis calculated as normal for a prop, but being enchanted they are not consumed\nin the casting of the sp ell. Enchanted props are made in the same way mana\nstones are, but enchanted props do not contain mana; they reduce the mana\ncost of any spell they apply to. <\/p>\n<p> To create a mana stone or enchanted prop requires a number of days equal\nto the mana bonus plus one for every College or Realm it can be used with,\ndoubled. The mana cost of the enchantment is equal to the number of days\nrequired to cast it. Sample Spells All sample spells have a base Difficulty\nof Fair. Most of them are low in detail; it's up to the GM to decide exactly\nwhat happens. Remember that props are always consumed unless they have been\nenchanted. <\/p>\n<h3>Sample Spells<\/h3>\n<h4>Fireball<\/h4>\n<p> Create Fire, Good effect (+5), long range (+2), several targets (+2). Total\nmana: 9. <\/p>\n<p> Creates a ball of flame, which streaks toward a point defined by the caster\nand explodes, causing 10 damage to all in its area. <\/p>\n<p> Sample prop: Small ball of bat guano and sulfur (-1). <\/p>\n<h4>Guardian Beast<\/h4>\n<p> Create Magic beast, Good effect (+4), lasts 1 day (+6), small area (+1).Total\nmana: 11. <\/p>\n<p> Creates an invisible guardian to guard an item or area. If anything enters\nthe area and moves toward the guarded item, the Beast will begin barking\nas a warning. If someone touches the item, the Beast will attack with the\nabilities of a powerful dog, with a fighting ability of Good due to being\ninvisible. The Beast will not stray more than a couple feet from the guarded\nitem. For +1 mana, the Beast can guard (and roam) an entire room. <\/p>\n<p> Sample prop: A crystal figurine of a dog or similar beast (-3). <\/p>\n<h4>Create Zombie or Golem<\/h4>\n<p> Find (Know) Spirit, Great effect (+8), long range (+2), takes 1 hour (-8).\nTotal mana: 2. <\/p>\n<p>Move Spirit, Good effect (+4), long range (+2), takes 10 minutes (-6). Total\nmana: 0. <\/p>\n<p> Control Spirit, Great effect (+8), lasts one day (+6), takes 1 hour (-8).\nTotal mana: 6. <\/p>\n<p> Causes a dead body, skeleton, or sculpture to be imbued with the spirit of\na dead man and come under the control of the caster. <\/p>\n<p> A three-part ritual magic, each part must be completed successfully before\nthe next part can begin, and the parts must be cast in immediate succession\nwith no recovery of mana in between. It is best to cast the spell in an area\nwhere a spirit is expected to be found, otherwise the range of the spell\nwill need to be increased. <\/p>\n<p> This spell assumes the use of spirits (tormented souls that have not left\nthis realm for one reason or another) to animate non-animate objects and\nis a combination of three spell effects. (To animate things purely by magical\nforce, the addition of an Animation College may be useful. Otherwise Move\nwith Magic (a Superb effect) should suffice.) <\/p>\n<p> If the GM wants to control the use of this spell, requiring valuable props\n(such as a body part of the spirit being captured) or simply raising the\nbasic cost of the spells should be effective. <\/p>\n<h4>Tanglefoot<\/h4>\n<p> Control Body, Mediocre effect (+1). Total mana: 1. <\/p>\n<p>\nCauses target to stumble and possibly fall. <\/p>\n<p> Sample prop: An ounce of strong drink (11). <\/p>\n<h4>Wizard's Eye<\/h4>\n<p> Know with Magic, Good effect (+4), small room (+2), extreme range (+4). Total\nmana: 10. <\/p>\n<p> Allows the caster to see things at a distance and through mundane obstacles\nas though he were standing at another place. <\/p>\n<p> Sample prop: A small magnifying crystal which the mage looks through (-2).\n<\/p>\n<h4>Scry Shield<\/h4>\n<p> Break Magic or Protect from Magic, Fair effect (+3), small room (+2), lasts\none hour (+4). Total mana: 9. <\/p>\n<p> Prevents spying by magical means anywhere inside its area. Treat attempts\nto spy through the shield as opposed actions, spell skill vs. spell skill. <\/p>\n<p> Sample props: A human eyeball pierced with an iron nail (-6), or a gem coated\nin lead (-2). <\/p>\n<h4>Grasping Hand<\/h4>\n<p> Move with Magic, Fair effect (+2), small target (-1), short range (+1), lasts\n2 minutes (+2). Total mana: 4. <\/p>\n<p> Allows the caster to move a small amount of any substance without touching\nit. <\/p>\n<h4>Portal <\/h4>\n<p> Move Body, Great effect (+8), extreme range (+4), lasts 10 minutes (+3),\ntakes 1 hour (-8). Total mana: 7. <br \/>or <br \/>Move with Magic, Great effect (+15), extreme range (+4), lasts 10 minutes\n(+3), takes 1 hour (-8). Total mana: 14. <\/p>\n<p> (Note: Move with Magic is a more costly spell because it can move anything.\nMove Body can move only a person or animal with personal possessions.) <\/p>\n<p> Creates a gate to anywhere the caster knows well, which anyone may pass through,\none person at a time. <\/p>\n<p> Sample prop: A doorway or appropriate passageway to pass through (-1), or\nif the target location is in a forest, a branch from a tree of that forest\n(-2). <\/p>\n<h4>Binding<\/h4>\n<p> Control Mind, Great effect (+8), extreme range (+4), lasts 1 season (+9),\ntakes 5 hours (-10). Total mana: 11. <\/p>\n<p> Allows the caster to bind the target to his will. The target will carry out\nthe caster's commands at all times, even when out of sight of the caster.\nThe target may break the control when told to do something he is strongly\nagainst if a Willpower or simil ar attribute check is successful. <\/p>\n<p> Sample prop: A gold chain, of great personal value, given to the target by\na loved one on a very special occasion (-12). <\/p>\n<h4>Tongues<\/h4>\n<p> Communicate with Body, Fair effect (+2), lasts 1 hour (+4). Total mana: 6.\n<\/p>\n<p> Allows target to speak to any living being that has a formal language. (To\nspeak to creatures without a formal language is a Good effect.) <\/p>\n<h4>Phantom Swarm<\/h4>\n<p> Control Mind, Good effect (+4), lasts 2 minutes (+2). Total mana: 6. <\/p>\n<p> Causes the target to believe he is being attacked by a huge swarm of stinging\ninsects. Unless target passes a Willpower or similar attribute check or would\nnormally be immune to stinging insects, he will take one Scratch of psychic\ndamage every round. (Psychic damage disappears at the end of the spell duration,\nand cannot cause death, but in all other respects acts as normal damage.)\nA critical success at the attribute check (+4 on the dice, or a result beyond\nSuperb) will cancel the spell prematurely. <\/p>\n<p> Sample prop: A hornets' nest with larvae inside (-4). <\/p>\n<h4>Thought Speech<\/h4>\n<p> Communicate with Mind, Fair effect (+2). Total mana: 2. <\/p>\n<p> Allows caster to mentally communicate with another intelligent humanoid.Note\nthat it works only at Close range for five seconds if the caster doesn't\nadd more mana to increase range and duration. <\/p>\n<p> Sample prop: A garment stained with the target's blood (-3). <\/p>\n<h3>Modifying the System<\/h3>\n<h4>World-View<\/h4>\n<p> The way magic works in a particular setting is a function of that\nsetting and how its populace views the world. If magic in one world is a\n\"scientific\" skill, requiring careful research and experimentation, a\nsorcerer in that world would have a different world-view than a sorcerer\nin a world where all magic is done by petitioning spirits to carry out\nthe sorcerer's task. The \"scientific\" sorcerer might see the world\ndivided into the classic elements of earth, air, fire, and water, while\nthe \"spirit\" sorcerer sees only animal, man, and divine spirits. Another\nsorcerer might petition his deity for magic and be able to cast only\nspells that fall into his deity's domain. <\/p>\n<p> Before the GM can customize this system to his setting, she must first\ndecide on a world-view and a \"philosophy\" of magic to base the system on.\nBasically, she must decide how or why magic works and how the sorcerers\nof the world think magic and the rest of the world works. This\nworld-view need not be universal. Just because the humans think in\nterms of written spell formulas, laboratories, and specific steps\ndoesn't mean the elves do. Feel free to create different magic systems\nfor different cultures and ways of thinking in your campaign. <\/p>\n<h4>Additional word ideas<\/h4>\n<p> One of the drastic modifications to the system is to change the set of\nwords available. The following are examples of words that the GM might\nfind useful in his campaign. <\/p>\n<p> <b>Animate<\/b> - If binding spirits to inanimate objects to\nanimate them doesn't work, you might need this College. <\/p>\n<p> <b>Matter, Energy<\/b> - If you need to manipulate generic\nmatter and energy directly but want the element Realms to be specific,\nthese Realms might be necessary. <\/p>\n<p> <b>Demonic, Divine<\/b> - If these supernatural elements are a big\npart of your campaign, you might want these as separate Realms. <\/p>\n<p> <b>Food<\/b> - This Realm is an example of just how specific the\nwords can get. <\/p>\n<p> <b>Language<\/b> - Another specific Realm that is currently covered\nunder the catch-all Magic Realm. <\/p>\n<p> <b>Machine<\/b> - If you're playing cyberfantasy, this is an\nimportant Realm. <\/p>","created_at":"2018-11-07T16:45:16+0000","updated_at":"2018-11-07T16:45:16+0000","annotations":[],"mimetype":"text\/html","reading_time":32,"domain_name":"zork.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":338,"title":"Four-by-Five Magic System for Fudge RPG","url":"http:\/\/www.panix.com\/~sos\/rpg\/4by5.html","content":"<br \/><div style=\"text-align: center;\">\nThis page last updated July 29, 2001<br \/>Copyright 1999 by Steffan O'Sullivan\n<\/div>\n<blockquote>\n<hr style=\"width: 90%; text-align: center;\" \/>\n<p>\n<i>Four-by-Five Magic<\/i> is a system similar to <a href=\"http:\/\/fudge.phoenyx.net\/gramarye\/\">The Gramarye<\/a>.  You might\nuse this system instead of <i>The Gramarye<\/i>, not in addition to\nit. <i>Four-by-Five Magic<\/i> derives its name from the Four Actions\nand Five Realms.  In this sense it is similar to <i>The Gramarye<\/i>,\nbut instead of buying individual Actions and Realms as skills, you\ncan only buy Spells from the resulting twenty combinations of\nActions and Realms.<\/p><p>\nThe magic system is written for <a href=\"http:\/\/www.panix.com\/~sos\/rpg\/fudfive.html\">Five-Point Fudge<\/a>,\na simple character creation system currently under playtest.  If\nyou are not using <i><b>Five-Point Fudge<\/b><\/i>, ignore any\nreferences to points.<\/p><blockquote><div>\n[<b>Note as of March 8, 2000:<\/b> <i>At this point, I have abandoned\nwork on this system.  I now realize it would require more work than I'm\nwilling to give to it to make it easy to use.  So this germ of the idea\nis all you get - if you can take it and run and expand it to something\nuseful, congratulations!  Everything else on this page was written May 7,\n1999, or earlier, except the final \"Grimoire\" section, added in 2001.<\/i>]<\/div><\/blockquote>\n<h3>Magical Talent Gift<\/h3>\n<p>\nThe Gift, <i>Magical Talent: Scholarly Magic<\/i>, is required to\nlearn spells in this system.  This may be bought in multiple levels\n- each level provides three points of mana, which is recovered as\ndescribed in <i>The Gramarye<\/i>.  In <i><b>Five-Point Fudge<\/b><\/i>,\nyou must have one level of Magical Talent for each point you spend in\nMagic skills.<\/p><h3>The Four Actions<\/h3>\n<ul><li><b>Enhance:<\/b> Strengthen, Heal, Enlarge, Repair, Sharpen, etc.\n<\/li><li><b>Diminish:<\/b> Weaken, Damage, Injure, Reduce, Break, Dull,\nDeteriorate, etc.\n<\/li><li><b>Communicate with, or Sense:<\/b> Read, Seek, Inform, Determine,\nUnderstand, etc.\n<\/li><li><b>Control:<\/b> Shape, Hold, Command, Form, Direct, Dictate, etc.\nControl is difficult to learn, and you may not learn Control spells\nunless you spend at least three points in the Magic Skill Group - <b>shown\nby [3] after the appropriate spells<\/b>.\n<\/li><\/ul><h3>The Five Realms<\/h3>\n<ul><li><b>Body:<\/b> Living body of sentient beings, animals, plants.\n<\/li><li><b>Mind:<\/b> That which normally inhabits and animates a body\n<\/li><li><b>Spirit:<\/b> That which no longer inhabits a body, or never did\nbut bears some resemblance to mind, or that unnaturally inhabits a body\nor other dwelling (undead, etc.).  Spirit is difficult to learn, and\nyou may not learn Spirit spells unless you spend at least three points\nin the Magic Skill Group - <b>shown by [3] after the appropriate\nspells<\/b>.\n<\/li><li><b>Energy:<\/b> Fire, water, air, magic, [electricity, etc.] - and,\nif the GM is willing, time.\n<\/li><li><b>Matter:<\/b> Solid material with no mind - stone, metal, wood,\nleather, paper, [plastic, etc.]\n<\/li><\/ul><h3>The Twenty Spells arranged by Realm<\/h3>\n<ul><li><b>Enhance Body:<\/b> Physical Healing, increase Strength or\nAgility, counter fatigue, increase size, increase movement ability,\netc.\n<\/li><li><b>Diminish Body:<\/b> Damage, shrink, weaken, make clumsy,\netc. Opposed by Health.\n<\/li><li><b>Sense Body:<\/b> \u201cRead\u201d an opponent, determine\nskill level, diagnose, determine whereabouts, etc.\n<\/li><li><b>Control Body:<\/b> Shapeshift, freeze in place, cause hand to\nattack owner, cause to stumble, enable to fly, etc. Opposed by Health.\n[3]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Mind:<\/b> Mental Healing, increase memory, temporarily\nraise intelligence, counter mental attacks, encourage, etc.\n<\/li><li><b>Diminish Mind:<\/b> Confuse, make stupid, induce headaches,\nmemory loss or distraction, discourage, etc. Opposed by Willpower.\n<\/li><li><b>Communicate with Mind:<\/b> Telepathy, both read and send,\nlanguages.  May be Opposed by Willpower.\n<\/li><li><b>Control Mind:<\/b> Dictate actions, emotions, attitudes,\nloyalties, beliefs, etc. Opposed by Willpower. [3]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Spirit:<\/b> Heal, increase size, encourage, counter\nattacks, etc. [3]\n<\/li><li><b>Diminish Spirit:<\/b> Injure, shrink, weaken, etc. Opposed by the\nspirit's Willpower.  [3]\n<\/li><li><b>Communicate with Spirit:<\/b> Communicate with spirits, summon.\nMay be Opposed by the Spirit's Willpower.  [3]\n<\/li><li><b>Control Spirit:<\/b> Dictate actions, emotions, banish from this\nplane, etc. Opposed by the spirit's Willpower.  [3]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Energy:<\/b> Increase the size of fire, a flow of water\nor air, aid a wizard in casting a spell, etc.\n<\/li><li><b>Diminish Energy:<\/b> Reduce fire, flow of water or air, counter\na magic spell, etc.\n<\/li><li><b>Sense Energy:<\/b> Sense types of energy, clairvoyance,\npsychometry, foresee, etc.\n<\/li><li><b>Control Energy:<\/b> Fireballs, jets of water or air, magical\nbolt, etc. [3]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Matter:<\/b> Repair, strengthen, sharpen, increase\nutility or aesthetics, enlarge, etc.\n<\/li><li><b>Diminish Matter:<\/b> Damage, shrink, dull, break,\nwarp, etc.\n<\/li><li><b>Sense Matter:<\/b> Sense type of matter, seek out closest of a\ngiven type, determine quality, strength, etc.\n<\/li><li><b>Control Matter:<\/b> Shape inanimate matter, telekinesis,\nanimation, open locks, etc. [3]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><\/ul><h3>The Twenty Spells arranged by Action<\/h3>\n<p>\nThese are the same descriptions found above, simply arranged by Action\ninstead of Realm in case you think that way.  However, sample spells,\nfound below, are only arranged by Realm.<\/p><ul><li><b>Enhance Body:<\/b> Physical Healing, increase Strength or\nAgility, counter fatigue, increase size, increase movement ability,\netc.\n<\/li><li><b>Enhance Mind:<\/b> Mental Healing, increase memory, temporarily\nraise intelligence, counter mental attacks, encourage, etc.\n<\/li><li><b>Enhance Spirit:<\/b> Heal, increase size, encourage, counter\nattacks, etc. [3]\n<\/li><li><b>Enhance Energy:<\/b> Increase the size of fire, a flow of water\nor air, aid a wizard in casting a spell, etc.\n<\/li><li><b>Enhance Matter:<\/b> Repair, strengthen, sharpen, increase\nutility or aesthetics, enlarge, etc.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Diminish Body:<\/b> Damage, shrink, weaken, make clumsy,\netc. Opposed by Health.\n<\/li><li><b>Diminish Mind:<\/b> Confuse, make stupid, induce headaches,\nmemory loss or distraction, discourage, etc. Opposed by Willpower.\n<\/li><li><b>Diminish Spirit:<\/b> Injure, shrink, weaken, etc. Opposed by the\nspirit's Willpower.  [3]\n<\/li><li><b>Diminish Energy:<\/b> Reduce fire, flow of water or air, counter\na magic spell, etc.\n<\/li><li><b>Diminish Matter:<\/b> Damage, shrink, dull, break,\nwarp, etc.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Sense Body:<\/b> \u201cRead\u201d an opponent, determine\nskill level, diagnose, determine whereabouts, etc.\n<\/li><li><b>Communicate with Mind:<\/b> Telepathy, both read and send,\nlanguages.  May be Opposed by Willpower.\n<\/li><li><b>Communicate with Spirit:<\/b> Communicate with spirits, summon.\nMay be Opposed by the Spirit's Willpower.  [3]\n<\/li><li><b>Sense Energy:<\/b> Sense types of energy, clairvoyance,\npsychometry, foresee, etc.\n<\/li><li><b>Sense Matter:<\/b> Sense type of matter, seek out closest of a\ngiven type, determine quality, strength, etc.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Control Body:<\/b> Shapeshift, freeze in place, cause hand to\nattack owner, cause to stumble, enable to fly, etc. Opposed by Health.\n[3]\n<\/li><li><b>Control Mind:<\/b> Dictate actions, emotions, attitudes,\nloyalties, beliefs, etc. Opposed by Willpower. [3]\n<\/li><li><b>Control Spirit:<\/b> Dictate actions, emotions, banish from this\nplane, etc. Opposed by the spirit's Willpower.  [3]\n<\/li><li><b>Control Energy:<\/b> Fireballs, jets of water or air, magical\nbolt, etc. [3]\n<\/li><li><b>Control Matter:<\/b> Shape inanimate matter, telekinesis,\nanimation, open locks, etc. [3]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><\/ul><div>\nIn addition to these spells, each of which is a skill, a character may\nalso learn:<\/div><ul><li><b>Alchemy *<\/b>\n<\/li><li><b>Thaumatology *<\/b>\n<\/li><\/ul><div>\nThere is no separate spell-casting skill in this system - each\nskill contains the ability to cast the spells inherent in it.<\/div><h3>Spell Potencies<\/h3>\n<p>\nSpells are divided into Potencies, rated at Fair through Superb.\nThe potency is the same as the target difficulty - that is, a Fairly\npotent spell needs a Fair or better result in order to succeed.\nIn addition, potency determines mana cost, at 2 mana per level -\nFair spells burn two mana, Good spells burn four, and so on.  These\nfigures can be adjusted up or down if you want a magic-rich or\nmagic-scarce campaign.<\/p><p>\nThe GM may use a system of bonuses and penalties to the skill roll.\nSomething similar to those found in Section 7.14 of <i><b>Fudge<\/b><\/i>\nitself can be adapted fairly easily.<\/p><p>\n<b>Fair Potency Spells<\/b> include things that don't seem like\nmagic to most people, but are nonetheless caused by a magician.\nSome <b>sample<\/b> Fair spells include:<\/p><ul><li><b>Enhance Body:<\/b> heal a minor wound in one day instead\nof five.\n<\/li><li><b>Diminish Body:<\/b> give an enemy a -1 to strike at you.\n<\/li><li><b>Sense Body:<\/b> discover an opponent's skill level by the way he\nholds his sword.\n<\/li><li><b>Control Body:<\/b> [none]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Mind:<\/b> being able to remember a fairly long list of\nitems.\n<\/li><li><b>Diminish Mind:<\/b> allow your weak disguise to pass all but\nintense inspection.\n<\/li><li><b>Communicate with Mind:<\/b> stare into a person's eyes and convey\na four-word message without speaking or moving.\n<\/li><li><b>Control Mind:<\/b> [none]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Spirit:<\/b> [none]\n<\/li><li><b>Diminish Spirit:<\/b> [none]\n<\/li><li><b>Communicate with Spirit:<\/b> [none]\n<\/li><li><b>Control Spirit:<\/b> [none]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Energy:<\/b> keep a fire going ten minutes after it\nshould have died out.\n<\/li><li><b>Diminish Energy:<\/b> kill a fire that should have kept burning.\n<\/li><li><b>Sense Energy:<\/b> sense which rooms in the inn you are at have a\nlit fire.\n<\/li><li><b>Control Energy:<\/b> [none]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Matter:<\/b> make a sword less likely to break.\n<\/li><li><b>Diminish Matter:<\/b> snap a taut rope by touching it.\n<\/li><li><b>Sense Matter:<\/b> sense the quality of a sword by touching it.\n<\/li><li><b>Control Matter:<\/b> [none]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><\/ul><div><b>Good Potency Spells<\/b> include still normal events, but those which\nhappen at strange times.  Some sample Good spells include:<ul><li><b>Enhance Body:<\/b> temporarily raise someone's Strength by two\nlevels.\n<\/li><li><b>Diminish Body:<\/b> cause a speaker to go hoarse.\n<\/li><li><b>Sense Body:<\/b> diagnose the most difficult disease accurately.\n<\/li><li><b>Control Body:<\/b> cause someone to stumble as they walk.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Mind:<\/b> temporarily raise someone's Reasoning power by\ntwo levels.\n<\/li><li><b>Diminish Mind:<\/b> make someone drunk after they've had only one\nsmall drink.\n<\/li><li><b>Communicate with Mind:<\/b> send a silent message to someone in\nthe same room.\n<\/li><li><b>Control Mind:<\/b> cause someone to see something out of the\ncorner of their eye - nothing really there.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Spirit:<\/b> strengthen a spirit by +1.\n<\/li><li><b>Diminish Spirit:<\/b> weaken a spirit by -1.\n<\/li><li><b>Communicate with Spirit:<\/b> ask a spirit \u201cyes\u201d\nor \u201cno\u201d questions and understand the answers.\n<\/li><li><b>Control Spirit:<\/b> [none]\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Energy:<\/b> increase a gust of wind to a strong blast.\n<\/li><li><b>Diminish Energy:<\/b> reduce an attack spell by -2 damage.\n<\/li><li><b>Sense Energy:<\/b> hear words spoken into the wind a few minutes\nago.\n<\/li><li><b>Control Energy:<\/b> cause fire to leap out of the fireplace a\nshort distance.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Matter:<\/b> allow a wooden staff to parry a sword\nwithout damage.\n<\/li><li><b>Diminish Matter:<\/b> snap a board with a light touch.\n<\/li><li><b>Sense Matter:<\/b> determine how much money a person is carrying.\n<\/li><li><b>Control Matter:<\/b> open a lock with a lockpick in a few seconds\n- without the lockpicking skill.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><\/ul><b>Great Potency Spells<\/b> include things that most people will\nacknowledge to be magic - things that go beyond the natural.  These are\nstill relatively minor effects, however.  Some sample Great spells\ninclude:<ul><li><b>Enhance Body:<\/b> run twice as fast as otherwise possible.\n<\/li><li><b>Diminish Body:<\/b> weaken a warrior so that he can't lift his\nsword.\n<\/li><li><b>Sense Body:<\/b> look at person and know their physical skills,\nincluding levels, at +\/-1 accuracy.\n<\/li><li><b>Control Body:<\/b> freeze someone in place.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Mind:<\/b> teach someone one of your skills at your\nlevel; lasts one day.\n<\/li><li><b>Diminish Mind:<\/b> make someone forget one mental skill for a\nday.\n<\/li><li><b>Communicate with Mind:<\/b> read surface thoughts of a subject.\n<\/li><li><b>Control Mind:<\/b> create illusions which can be seen, heard,\nsmelt, felt.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Spirit:<\/b> strengthen a spirit to twice its potency.\n<\/li><li><b>Diminish Spirit:<\/b> weaken a spirit to half its potency.\n<\/li><li><b>Communicate with Spirit:<\/b> talk freely with any spirit.\n<\/li><li><b>Control Spirit:<\/b> bind a spirit to specific object or allow a\nspirit bound in place to roam free.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Energy:<\/b> cause water to overflow a dam.\n<\/li><li><b>Diminish Energy:<\/b> stop a fireball from doing damage.\n<\/li><li><b>Sense Energy:<\/b> psychometry - feel the personality of the last\nperson to use an object.\n<\/li><li><b>Control Energy:<\/b> ride the wind.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Matter:<\/b> grow a knife into a sword.\n<\/li><li><b>Diminish Matter:<\/b> shrink a spear to an arrow.\n<\/li><li><b>Sense Matter:<\/b> \u201cimprint\u201d your signature on an object - be\naware of who touches it over the next week.\n<\/li><li><b>Control Matter:<\/b> animate a pair of scissors to operate at\nyour will from a distance.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><\/ul><b>Superb Potency Spells<\/b> include things that everyone will admit is\nmajor magic.  Some sample Superb spells include:<ul><li><b>Enhance Body:<\/b> restore a missing limb.\n<\/li><li><b>Diminish Body:<\/b> cause damage to an enemy as if struck by an\ninvisible sword.\n<\/li><li><b>Sense Body:<\/b> sense when a specific person walks into a room a\nmile away.\n<\/li><li><b>Control Body:<\/b> shapeshift into a hawk.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Mind:<\/b> restore an insane person to mental health\ninstantly.\n<\/li><li><b>Diminish Mind:<\/b> drive someone mad instantly.\n<\/li><li><b>Communicate with Mind:<\/b> read subsurface thoughts.\n<\/li><li><b>Control Mind:<\/b> cause a character to attack his friends.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Spirit:<\/b> cause a spirit to grow ten times its size.\n<\/li><li><b>Diminish Spirit:<\/b> cause a zombie to disintegrate, freeing\nthe spirit from the body forever.\n<\/li><li><b>Communicate with Spirit:<\/b> summon a spirit from another realm.\n<\/li><li><b>Control Spirit:<\/b> banish a spirit from this plane entirely.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Energy:<\/b> create a wind storm from a normal gust of\nwind.\n<\/li><li><b>Diminish Energy:<\/b> nullify another magician's powerful\nspell.\n<\/li><li><b>Sense Energy:<\/b> sense when magic is being used in a castle a\nmile away.\n<\/li><li><b>Control Energy:<\/b> cast a large, potent fireball.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><li><b>Enhance Matter:<\/b> make paper as hard as rock.\n<\/li><li><b>Diminish Matter:<\/b> snap a sword by touching it.\n<\/li><li><b>Sense Matter:<\/b> locate the nearest quantity of gold within a\nmile.\n<\/li><li><b>Control Matter:<\/b> shape stone as if it were soft putty.\n<hr style=\"border-style: none; border: 0; background-color: gray; color: gray; text-align: left; width: 25%;\" \/><\/li><\/ul>\nNote that there are certainly more potent spells than these, but they\nare not available to PC magicians.  Well, they might encounter such\nmagic ... but they can't cast it.<hr style=\"width: 90%; text-align: center;\" \/><h3>The Sample Spell Effects arranged as a Wizard's Grimoire<\/h3>\n<p><b>Communicate with Mind:<\/b>\n<\/p><ul><li>Fair: <\/li>             \n<li>Good:<\/li>          \n<li>Great:<\/li>      \n<li>Superb:<\/li>    <\/ul><p><b>Communicate with Spirit:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>           \n<li>Great: <\/li>    \n<li>Superb:<\/li>       <\/ul><p><b>Control Body:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>       \n<li>Great:<\/li>    \n<li>Superb:<\/li>     <\/ul><p><b>Control Energy:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>           \n<li>Great:<\/li>   \n<li>Superb:<\/li>      <\/ul><p><b>Control Matter:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>               \n<li>Great:<\/li>             \n<li>Superb:<\/li>         <\/ul><p><b>Control Mind:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>                \n<li>Great:<\/li>         \n<li>Superb:<\/li>        <\/ul><p><b>Control Spirit:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li> \n<li>Great:<\/li>                \n<li>Superb:<\/li>        <\/ul><p><b>Diminish Body:<\/b>\n<\/p><ul><li>Fair:<\/li>         \n<li>Good:<\/li>      \n<li>Great:<\/li>          \n<li>Superb:<\/li>             <\/ul><p><b>Diminish Energy:<\/b>\n<\/p><ul><li>Fair:<\/li>        \n<li>Good:<\/li>       \n<li>Great:<\/li>      \n<li>Superb:<\/li>      <\/ul><p><b>Diminish Matter:<\/b>\n<\/p><ul><li>Fair:<\/li>       \n<li>Good:<\/li>       \n<li>Great:<\/li>      \n<li>Superb:<\/li>       <\/ul><p><b>Diminish Mind:<\/b>\n<\/p><ul><li>Fair:<\/li>          \n<li>Good:<\/li>          \n<li>Great:<\/li>         \n<li>Superb:<\/li>     <\/ul><p><b>Diminish Spirit:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>     \n<li>Great:<\/li>       \n<li>Superb: <\/li>            <\/ul><p><b>Enhance Body:<\/b>\n<\/p><ul><li>Fair:<\/li>          \n<li>Good:<\/li>       \n<li>Great:<\/li>       \n<li>Superb:<\/li>     <\/ul><p><b>Enhance Energy:<\/b>\n<\/p><ul><li>Fair:<\/li>            \n<li>Good:<\/li>         \n<li>Great:<\/li>      \n<li>Superb:<\/li>           <\/ul><p><b>Enhance Matter:<\/b>\n<\/p><ul><li>Fair:<\/li>       \n<li>Good:<\/li>          \n<li>Great: <\/li>     \n<li>Superb:<\/li>       <\/ul><p><b>Enhance Mind:<\/b>\n<\/p><ul><li>Fair:<\/li>          \n<li>Good:<\/li>        \n<li>Great:<\/li>            \n<li>Superb:<\/li>         <\/ul><p><b>Enhance Spirit:<\/b>\n<\/p><ul><li>Fair:<\/li> \n<li>Good:<\/li>     \n<li>Great:<\/li>       \n<li>Superb:<\/li>          <\/ul><p><b>Sense Body:<\/b>\n<\/p><ul><li>Fair:<\/li>            \n<li>Good:<\/li>      \n<li>Great:<\/li>             \n<li>Superb: <\/li>            <\/ul><p><b>Sense Energy:<\/b>\n<\/p><ul><li>Fair:<\/li>             \n<li>Good: <\/li>         \n<li>Great:<\/li>             \n<li>Superb:<\/li>             <\/ul><p><b>Sense Matter:<\/b>\n<\/p><ul><li>Fair:<\/li>         \n<li>Good: <\/li>       \n<li>Great:<\/li>                 \n<li>Superb:<\/li>          <\/ul><hr style=\"width: 90%; text-align: center;\" \/><p>\nBack to <a href=\"http:\/\/www.panix.com\/~sos\/fudge.html\">The Fudge Page<\/a><br \/>Back to <a href=\"http:\/\/www.panix.com\/~sos\/gviews.html\">SOS' Gameviews<\/a><br \/>Back to <a href=\"http:\/\/www.panix.com\/~sos\/\">Steffan O'Sullivan's Home Page<\/a>\n<\/p><\/div><\/blockquote>","created_at":"2018-11-07T16:45:15+0000","updated_at":"2018-11-07T16:45:15+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"www.panix.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":335,"title":"modernish\/modernish","url":"https:\/\/github.com\/modernish\/modernish","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>modernish is an ambitious, as-yet experimental, cross-platform POSIX shell\nfeature detection and language extension library. It aims to extend the\nshell language with extensive\n<a href=\"#user-content-feature-testing\">feature testing<\/a>\nand language enhancements,\nusing the power of aliases and functions to extend the shell language\nusing the shell language itself.<\/p>\n<p>The name is a pun on Modernizr, the JavaScript feature testing library, -sh,\nthe common suffix for UNIX shell names, and -ish, still not quite a modern\nprogramming language but perhaps a little closer. jQuery is another source\nof general inspiration; like it, modernish adds a considerable feature set\nby using the power of the language it's implemented in to extend\/transcend\nthat same language.<\/p>\n<p>That said, the aim of modernish is to build a better shell language, and not\nto make the shell language into something it's not. Its feature set is aimed\nat solving specific and commonly experienced deficits and annoyances of the\nshell language, and not at adding\/faking things that are foreign to it, such\nas object orientation or functional programming. (However, since modernish\nis modular, nothing stops anyone from adding a module attempting to\nimplement these things.)<\/p>\n<p>The library builds on pure POSIX 2013 Edition (including full C-style shell\narithmetics with assignment, comparison and conditional expressions), so it\nshould run on any POSIX-compliant shell and operating system. But it does\nnot shy away from using non-standard extensions where available to enhance\nperformance or robustness.<\/p>\n<p>Some example programs are in <code>share\/doc\/modernish\/examples<\/code>.<\/p>\n<p>Modernish also comes with a suite of regression tests to detect bugs in\nmodernish itself. See <a href=\"#user-content-appendix-b\">Appendix B<\/a>.<\/p>\n<h2><a id=\"user-content-table-of-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-contents\"><\/a>Table of contents<\/h2>\n<ul><li><a href=\"#user-content-getting-started\">Getting started<\/a><\/li>\n<li><a href=\"#user-content-two-basic-forms-of-a-modernish-program\">Two basic forms of a modernish program<\/a>\n<ul><li><a href=\"#user-content-important-notes-regarding-the-system-locale\">Important notes regarding the system locale<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-interactive-use\">Interactive use<\/a><\/li>\n<li><a href=\"#user-content-non-interactive-command-line-use\">Non-interactive command line use<\/a>\n<ul><li><a href=\"#user-content-non-interactive-usage-examples\">Non-interactive usage examples<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-internal-namespace\">Internal namespace<\/a><\/li>\n<li><a href=\"#user-content-feature-testing\">Feature testing<\/a><\/li>\n<li><a href=\"#user-content-modernish-system-constants\">Modernish system constants<\/a>\n<ul><li><a href=\"#user-content-control-character-whitespace-and-shell-safe-character-constants\">Control character, whitespace and shell-safe character constants<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-legibility-aliases\">Legibility aliases<\/a><\/li>\n<li><a href=\"#user-content-enhanced-exit\">Enhanced exit<\/a><\/li>\n<li><a href=\"#user-content-reliable-emergency-halt\">Reliable emergency halt<\/a><\/li>\n<li><a href=\"#user-content-low-level-shell-utilities\">Low-level shell utilities<\/a><\/li>\n<li><a href=\"#user-content-quoting-strings-for-subsequent-parsing-by-the-shell\">Quoting strings for subsequent parsing by the shell<\/a><\/li>\n<li><a href=\"#user-content-the-stack\">The stack<\/a>\n<ul><li><a href=\"#user-content-the-shell-options-stack\">The shell options stack<\/a><\/li>\n<li><a href=\"#user-content-the-trap-stack\">The trap stack<\/a>\n<ul><li><a href=\"#user-content-trap-stack-compatibility-considerations\">Trap stack compatibility considerations<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-positional-parameters-stack\">Positional parameters stack<\/a><\/li>\n<li><a href=\"#user-content-stack-query-and-maintenance-functions\">Stack query and maintenance functions<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-hardening-emergency-halt-on-error\">Hardening: emergency halt on error<\/a>\n<ul><li><a href=\"#user-content-important-note-on-variable-assignments\">Important note on variable assignments<\/a><\/li>\n<li><a href=\"#user-content-hardening-while-allowing-for-broken-pipes\">Hardening while allowing for broken pipes<\/a><\/li>\n<li><a href=\"#user-content-tracing-the-execution-of-hardened-commands\">Tracing the execution of hardened commands<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-simple-tracing-of-commands\">Simple tracing of commands<\/a><\/li>\n<li><a href=\"#user-content-external-commands-without-full-path\">External commands without full path<\/a><\/li>\n<li><a href=\"#user-content-outputting-strings\">Outputting strings<\/a><\/li>\n<li><a href=\"#user-content-enhanced-dot-scripts\">Enhanced dot scripts<\/a><\/li>\n<li><a href=\"#user-content-testing-numbers-strings-and-files\">Testing numbers, strings and files<\/a>\n<ul><li><a href=\"#user-content-integer-number-arithmetic-tests-and-operations\">Integer number arithmetic tests and operations<\/a><\/li>\n<li><a href=\"#user-content-string-tests\">String tests<\/a><\/li>\n<li><a href=\"#user-content-file-type-tests\">File type tests<\/a><\/li>\n<li><a href=\"#user-content-file-comparison-tests\">File comparison tests<\/a><\/li>\n<li><a href=\"#user-content-file-status-tests\">File status tests<\/a><\/li>\n<li><a href=\"#user-content-io-tests\">I\/O tests<\/a><\/li>\n<li><a href=\"#user-content-file-permission-tests\">File permission tests<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-basic-string-operations\">Basic string operations<\/a>\n<ul><li><a href=\"#user-content-touppertolower\">toupper\/tolower<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-basic-system-utilities\">Basic system utilities<\/a><\/li>\n<li><a href=\"#user-content-modules\">Modules<\/a>\n<ul><li><a href=\"#user-content-use-safe\">use safe<\/a><\/li>\n<li><a href=\"#user-content-use-vararith\">use var\/arith<\/a>\n<ul><li><a href=\"#user-content-arithmetic-operator-shortcuts\">Arithmetic operator shortcuts<\/a><\/li>\n<li><a href=\"#user-content-arithmetic-comparison-shortcuts\">Arithmetic comparison shortcuts<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-use-varmapr\">use var\/mapr<\/a>\n<ul><li><a href=\"#user-content-differences-from-mapfile\">Differences from <code>mapfile<\/code><\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-use-varreadf\">use var\/readf<\/a><\/li>\n<li><a href=\"#user-content-use-varsetlocal\">use var\/setlocal<\/a><\/li>\n<li><a href=\"#user-content-use-varstring\">use var\/string<\/a><\/li>\n<li><a href=\"#user-content-use-sysbase\">use sys\/base<\/a>\n<ul><li><a href=\"#user-content-use-sysbasereadlink\">use sys\/base\/readlink<\/a><\/li>\n<li><a href=\"#user-content-use-sysbasewhich\">use sys\/base\/which<\/a><\/li>\n<li><a href=\"#user-content-use-sysbasemktemp\">use sys\/base\/mktemp<\/a><\/li>\n<li><a href=\"#user-content-use-sysbaseseq\">use sys\/base\/seq<\/a><\/li>\n<li><a href=\"#user-content-use-sysbaserev\">use sys\/base\/rev<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-use-sysdir\">use sys\/dir<\/a>\n<ul><li><a href=\"#user-content-use-sysdirtraverse\">use sys\/dir\/traverse<\/a><\/li>\n<li><a href=\"#user-content-use-sysdircountfiles\">use sys\/dir\/countfiles<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-use-systerm\">use sys\/term<\/a>\n<ul><li><a href=\"#user-content-use-systermreadkey\">use sys\/term\/readkey<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-use-optsparsergen\">use opts\/parsergen<\/a><\/li>\n<li><a href=\"#user-content-use-loopcfor\">use loop\/cfor<\/a><\/li>\n<li><a href=\"#user-content-use-loopsfor\">use loop\/sfor<\/a><\/li>\n<li><a href=\"#user-content-use-loopwith\">use loop\/with<\/a><\/li>\n<li><a href=\"#user-content-use-loopselect\">use loop\/select<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-appendix-a\">Appendix A<\/a>\n<ul><li><a href=\"#user-content-capabilities\">Capabilities<\/a><\/li>\n<li><a href=\"#user-content-quirks\">Quirks<\/a><\/li>\n<li><a href=\"#user-content-bugs\">Bugs<\/a><\/li>\n<li><a href=\"#user-content-warning-ids\">Warning IDs<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-content-appendix-b\">Appendix B<\/a>\n<ul><li><a href=\"#user-content-difference-between-bugquirkfeature-tests-and-regression-tests\">Difference between bug\/quirk\/feature tests and regression tests<\/a><\/li>\n<li><a href=\"#user-content-testing-modernish-on-all-your-shells\">Testing modernish on all your shells<\/a><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting started<\/h2>\n<p>Run <code>install.sh<\/code> and follow instructions, choosing your preferred shell\nand install location. After successful installation you can run modernish\nshell scripts and write your own. Run <code>uninstall.sh<\/code> to remove modernish.<\/p>\n<p>Both the install and uninstall scripts are interactive by default, but\nsupport fully automated (non-interactive) operation as well. Command\nline options are as follows:<\/p>\n<p><code>install.sh<\/code> [ <code>-n<\/code> ] [ <code>-s<\/code> <em>shell<\/em> ] [ <code>-f<\/code> ] [ <code>-d<\/code> <em>installroot<\/em> ] [ <code>-D<\/code> <em>prefix<\/em> ]<\/p>\n<ul><li><code>-n<\/code>: non-interactive operation<\/li>\n<li><code>-s<\/code>: specify default shell to execute modernish<\/li>\n<li><code>-f<\/code>: force unconditional installation on specified shell<\/li>\n<li><code>-d<\/code>: specify root directory for installation<\/li>\n<li><code>-D<\/code>: extra destination directory prefix (for packagers)<\/li>\n<\/ul><p><code>uninstall.sh<\/code> [ <code>-n<\/code> ] [ <code>-f<\/code> ] [ <code>-d<\/code> <em>installroot<\/em> ]<\/p>\n<ul><li><code>-n<\/code>: non-interactive operation<\/li>\n<li><code>-f<\/code>: delete *\/modernish directories even if files left<\/li>\n<li><code>-d<\/code>: specify root directory of modernish installation to uninstall<\/li>\n<\/ul><h2><a id=\"user-content-two-basic-forms-of-a-modernish-program\" class=\"anchor\" aria-hidden=\"true\" href=\"#two-basic-forms-of-a-modernish-program\"><\/a>Two basic forms of a modernish program<\/h2>\n<p>The <strong>simplest<\/strong> way to write a modernish program is to source modernish as a\ndot script. For example, if you write for bash:<\/p>\n<pre>#! \/bin\/bash\n. modernish\nuse safe\nuse sys\/base\n...your program starts here...\n<\/pre>\n<p>The modernish 'use' command load modules with optional functionality. <code>safe<\/code> is\na special module that introduces a new and safer way of shell programming, with\nfield splitting (word splitting) and pathname expansion (globbing) disabled by\ndefault. The <code>sys\/base<\/code> module contains modernish versions of certain basic but\nnon-standardised utilities (e.g. <code>readlink<\/code>, <code>mktemp<\/code>, <code>which<\/code>), guaranteeing\nthat modernish programs all have a known version at their disposal. There are\nmany other modules as well. See below for more information.<\/p>\n<p>The above method makes the program dependent on one particular shell (in this\ncase, bash). So it is okay to mix and match functionality specific to that\nparticular shell with modernish functionality.<\/p>\n<p>The <strong>most portable<\/strong> way to write a modernish program is to use the special\ngeneric hashbang path for modernish programs. For example:<\/p>\n<pre>#! \/usr\/bin\/env modernish\n#! use safe\n#! use sys\/base\n...your program begins here...\n<\/pre>\n<p>A program in this form is executed by whatever shell the user who installed\nmodernish on the local system chose as the default shell. Since you as the\nprogrammer can't know what shell this is (other than the fact that it passed\nsome rigorous POSIX compliance testing executed by modernish), a program in\nthis form <em>must be strictly POSIX compliant<\/em> -- except, of course, that it\nshould also make full use of the rich functionality offered by modernish.<\/p>\n<p>Note that modules are loaded in a different way: the <code>use<\/code> commands are part of\nhashbang comment (starting with <code>#!<\/code> like the initial hashbang path). Only such\nlines that <em>immediately<\/em> follow the initial hashbang path are evaluated; even\nan empty line in between causes the rest to be ignored.<\/p>\n<h3><a id=\"user-content-important-notes-regarding-the-system-locale\" class=\"anchor\" aria-hidden=\"true\" href=\"#important-notes-regarding-the-system-locale\"><\/a>Important notes regarding the system locale<\/h3>\n<ul><li>modernish, like most shells, fully supports two locales: POSIX (a.k.a.\nC, a.k.a. ASCII) and Unicode's UTF-8. It will work in others, but things\nlike converting to upper\/lower case, and matching single characters in\npatterns, are not guaranteed.<br \/><em>Caveat:<\/em> some shells or operating systems have bugs that prevent (or lack\nfeatures required for) full locale support. If portability is a concern,\ncheck for <code>thisshellhas BUG_MULTIBYTE<\/code> or <code>thisshellhas BUG_NOCHCLASS<\/code>\nwhere needed. See Appendix A under <a href=\"#user-content-bugs\">Bugs<\/a>.<\/li>\n<li>Scripts\/programs should <em>not<\/em> change the locale (<code>LC_*<\/code> or <code>LANG<\/code>) after\ninitialising modernish. Doing this might break various functions, as\nmodernish sets specific versions depending on your OS, shell and locale.\n(Temporarily changing the locale is fine as long as you don't use\nmodernish features that depend on it -- for example, setting a specific\nlocale just for an external command. However, if you use <code>harden()<\/code>, see\nthe <a href=\"#user-content-important-note-on-variable-assignments\">important note<\/a>\nin its documentation below!)<\/li>\n<\/ul><h2><a id=\"user-content-interactive-use\" class=\"anchor\" aria-hidden=\"true\" href=\"#interactive-use\"><\/a>Interactive use<\/h2>\n<p>Modernish is primarily designed to enhance shell programs\/scripts, but also\noffers features for use in interactive shells. For instance, the new <code>with<\/code>\nloop construct from the <code>loop\/with<\/code> module can be quite practical to repeat\nan action x times, and the <code>safe<\/code> module on interactive shells provides\nconvenience functions for manipulating, saving and restoring the state of\nfield splitting and globbing.<\/p>\n<p>To use modernish on your favourite interactive shell, you have to add it to\nyour <code>.profile<\/code>, <code>.bashrc<\/code> or similar init file.<\/p>\n<p><strong>Important:<\/strong> Upon initialising, modernish adapts itself to\nother settings, such as the locale. It also removes certain aliases that\nmay keep modernish from initialising properly. So you have to organise your\n<code>.profile<\/code> or similar file in the following order:<\/p>\n<ul><li><em>first<\/em>, define general system settings (<code>PATH<\/code>, locale, etc.);<\/li>\n<li><em>then<\/em>, <code>. modernish<\/code> and <code>use<\/code> any modules you want;<\/li>\n<li><em>then<\/em> define anything that may depend on modernish, and set your aliases.<\/li>\n<\/ul><h2><a id=\"user-content-non-interactive-command-line-use\" class=\"anchor\" aria-hidden=\"true\" href=\"#non-interactive-command-line-use\"><\/a>Non-interactive command line use<\/h2>\n<p>After installation, the <code>modernish<\/code> command can be invoked as if it were a\nshell, with the standard command line options from other shells (such as\n<code>-c<\/code> to specify a command or script directly on the command line), plus some\nenhancements. The effect is that the shell chosen at installation time will\nbe run enhanced with modernish functionality. It is not possible to use\nmodernish as an interactive shell in this way.<\/p>\n<p>Usage:<\/p>\n<ol><li><code>modernish<\/code> [ <code>--use=<\/code><em>module<\/em> | <em>option<\/em> ... ]\n[ <em>scriptfile<\/em> ] [ <em>arguments<\/em> ]<\/li>\n<li><code>modernish<\/code> [ <code>--use=<\/code><em>module<\/em> | <em>option<\/em> ... ]\n<code>-c<\/code> [ <em>script<\/em> [ <em>me-name<\/em> [ <em>arguments<\/em> ] ] ]<\/li>\n<li><code>modernish --test<\/code><\/li>\n<li><code>modernish --version<\/code><\/li>\n<\/ol><p>In the first form, the <code>--use<\/code> long-form option preloads any given modernish\n<a href=\"#user-content-modules\">modules<\/a>, any given short or long-form shell <em>option<\/em>s\nare set or unset (the syntax is identical to that of POSIX shells and the\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_25_03\" rel=\"nofollow\">shell options<\/a>\nsupported depend on the shell executing modernish), and then <em>scriptfile<\/em> is\nloaded and executed with any <em>arguments<\/em> assigned to the positional parameters.<\/p>\n<p>The <em>module<\/em> argument to each specified <code>--use<\/code> option is split using\nstandard shell field splitting. The first field is the module name and any\nfurther fields become arguments to that module's initialisation routine.<\/p>\n<p>Using the shell option <code>-e<\/code> or <code>-o errexit<\/code> is an error, because modernish\n<a href=\"#user-content-hardening-emergency-halt-on-error\">does not support it<\/a> and\nwould break. If the shell option <code>-x<\/code> or <code>-o xtrace<\/code> is given, modernish sets\nthe <code>PS4<\/code> prompt to a useful value that traces the line number and exit status,\nas well as the current file and function names if the shell is capable of this.<\/p>\n<p>In the second form, after pre-loading any <em>module<\/em>s and setting any shell\n<em>option<\/em>s as in the first form, <code>-c<\/code> executes the specified modernish\n<em>script<\/em>, optionally with the <em>me-name<\/em> assigned to <code>$ME<\/code> and the\n<em>arguments<\/em> assigned to the positional parameters. This is identical to the\n<code>-c<\/code> option on POSIX shells, except that the <em>me-name<\/em> is assigned to <code>$ME<\/code>\nand not <code>$0<\/code> (because POSIX shells do not allow changing <code>$0<\/code>).<\/p>\n<p>The <code>--test<\/code> option runs the regression test suite and exits. This verifies\nthat the modernish installation is functioning correctly.\nSee <a href=\"#user-content-appendix-b\">Appendix B<\/a> for more information.<\/p>\n<p>The <code>--version<\/code> option outputs the version of modernish and exits.<\/p>\n<h3><a id=\"user-content-non-interactive-usage-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#non-interactive-usage-examples\"><\/a>Non-interactive usage examples<\/h3>\n<ul><li>Count to 10 using <a href=\"#user-content-use-loopwith\">loop\/with<\/a>:<br \/><code>modernish --use=loop\/with -c 'with i=1 to 10; do putln \"$i\"; done'<\/code><\/li>\n<li>Run a <a href=\"#user-content-two-basic-forms-of-a-modernish-program\">portable-form<\/a>\nmodernish program using zsh and enhanced-prompt xtrace:<br \/><code>zsh \/usr\/local\/bin\/modernish -o xtrace \/path\/to\/program.sh<\/code><\/li>\n<\/ul><h2><a id=\"user-content-internal-namespace\" class=\"anchor\" aria-hidden=\"true\" href=\"#internal-namespace\"><\/a>Internal namespace<\/h2>\n<p>Function-local variables are not supported by the standard POSIX shell; only\nglobal variables are provided for. Modernish needs a way to store its\ninternal state without interfering with the program using it. So most of the\nmodernish functionality uses an internal namespace <code>_Msh_*<\/code> for variables,\nfunctions and aliases. All these names may change at any time without\nnotice. <em>Any names starting with <code>_Msh_<\/code> should be considered sacrosanct and\nuntouchable; modernish programs should never directly use them in any way.<\/em>\nOf course this is not enforceable, but names starting with <code>_Msh_<\/code> should be\nuncommon enough that no unintentional conflict is likely to occur.<\/p>\n<h2><a id=\"user-content-feature-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#feature-testing\"><\/a>Feature testing<\/h2>\n<p>Modernish includes a battery of shell bug, quirk and feature tests, each of\nwhich is given a special ID.\nSee <a href=\"#user-content-appendix-a\">Appendix A<\/a> below for a list of shell\ncapabilities, quirks and bugs that modernish currently tests for,\nas well as further general information on the feature testing framework.<\/p>\n<p><code>thisshellhas<\/code> is the central function of the modernish feature testing\nframework. It not only tests for the presence of modernish shell\ncapabilities\/quirks\/bugs on the current shell, but can also test for the\npresence of specific shell built-in commands, shell reserved words (a.k.a.\nkeywords), shell options (short or long form), and signals.<\/p>\n<p>Modernish itself extensively uses feature testing to adapt itself to the\nshell it's running on. This is how it works around shell bugs and takes\nadvantage of efficient features not all shells have. But any script using\nthe library can do this in the same way, with the help of this function.<\/p>\n<p>For those tests that have no standardised way of performing them,\n<code>thisshellhas<\/code> knows the shell-specific methods on all supported shells and\nautomatically initialises the correct version for the current shell.<\/p>\n<p>Test results are cached in memory, so repeated checks using <code>thisshellhas<\/code>\nare efficient and there is no need to avoid calling it to optimise\nperformance.<\/p>\n<p>Usage:<\/p>\n<p><code>thisshellhas<\/code> [ <code>--cache<\/code> | <code>--show<\/code> ] <em>item<\/em> [ <em>item<\/em> ... ]<\/p>\n<ul><li>If <em>item<\/em> contains only ASCII capital letters A-Z, digits 0-9 or <code>_<\/code>,\nreturn the result status of the associated modernish\n<a href=\"#user-content-appendix-a\">feature, quirk or bug test<\/a>.<\/li>\n<li>If <em>item<\/em> is an ASCII all-lowercase word, check if it's a shell reserved\nword or built-in command on the current shell.<\/li>\n<li>If <em>item<\/em> starts with <code>--rw=<\/code> or <code>--kw=<\/code>, check if the identifier\nimmediately following these characters is a shell reserved word\n(a.k.a. shell keyword).<\/li>\n<li>If <em>item<\/em> starts with <code>--bi=<\/code>, similarly check for a shell built-in command.<\/li>\n<li>If <em>item<\/em> starts with <code>--sig=<\/code>, check if the shell knows about a signal\n(usable by <code>kill<\/code>, <code>trap<\/code>, etc.) by the name or number following the <code>=<\/code>.\nIf a number &gt; 128 is given, the remainder of its division by 128 is checked.\nIf the signal is found, its canonicalised signal name is left in the\n<code>REPLY<\/code> variable, otherwise <code>REPLY<\/code> is unset. (If multiple <code>--sig=<\/code> items\nare given and all are found, <code>REPLY<\/code> contains only the last one.)<\/li>\n<li>If <em>item<\/em> is <code>-o<\/code> followed by a separate word, check if this shell has a\nlong-form shell option by that name.<\/li>\n<li>If <em>item<\/em> is any other letter or digit preceded by a single <code>-<\/code>, check if\nthis shell has a short-form shell option by that character.<\/li>\n<li>The <code>--cache<\/code> option runs all external modernish bug\/quirk\/feature tests\nthat have not yet been run, causing the cache to be complete.<\/li>\n<li>The <code>--show<\/code> option performs a <code>--cache<\/code> and then outputs all the IDs of\npositive results, one per line.<\/li>\n<\/ul><p><code>thisshellhas<\/code> continues to process <em>item<\/em>s until one of them produces a\nnegative result or is found invalid, at which point any further <em>item<\/em>s are\nignored. So the function only returns successfully if all the <em>item<\/em>s\nspecified were found on the current shell. (To check if either one <em>item<\/em> or\nanother is present, use separate <code>thisshellhas<\/code> invocations separated by the\n<code>||<\/code> shell operator.)<\/p>\n<p>Exit status: 0 if this shell has all the <em>items<\/em> in question; 1 if not; 2 if\nan <em>item<\/em> was encountered that is not recognised as a valid identifier.<\/p>\n<p><strong>Note:<\/strong> The tests for the presence of reserved words, built-in commands,\nshell options, and signals are different from feature\/quirk\/bug tests in an\nimportant way: they only check if an item by that name exists on this shell,\nand don't verify that it does the same thing as on another shell.<\/p>\n<h2><a id=\"user-content-modernish-system-constants\" class=\"anchor\" aria-hidden=\"true\" href=\"#modernish-system-constants\"><\/a>Modernish system constants<\/h2>\n<p>Modernish provides certain constants (read-only variables) to make life easier.\nThese include:<\/p>\n<ul><li><code>$MSH_VERSION<\/code>: The version of modernish.<\/li>\n<li><code>$MSH_PREFIX<\/code>: Installation prefix for this modernish installation (e.g.\n\/usr\/local).<\/li>\n<li><code>$MSH_CONFIG<\/code>: Path to modernish user configuration directory.<\/li>\n<li><code>$ME<\/code>: Path to the current program. Replacement for <code>$0<\/code>. This is\nnecessary if the hashbang path <code>#!\/usr\/bin\/env modernish<\/code> is used, or if\nthe program is launched like <code>sh \/path\/to\/bin\/modernish \/path\/to\/script.sh<\/code>, as these set <code>$0<\/code> to the path to bin\/modernish and\nnot your program's path.<\/li>\n<li><code>$MSH_SHELL<\/code>: Path to the default shell for this modernish installation,\nchosen at install time (e.g. \/bin\/sh). This is a shell that is known to\nhave passed all the modernish tests for fatal bugs. Cross-platform scripts\nshould use it instead of hard-coding \/bin\/sh, because on some operating\nsystems (NetBSD, OpenBSD, Solaris) \/bin\/sh is not POSIX compliant.<\/li>\n<li><code>$SIGPIPESTATUS<\/code>: The exit status of a command killed by <code>SIGPIPE<\/code> (a\nbroken pipe). For instance, if you use <code>grep something somefile.txt | more<\/code> and you quit <code>more<\/code> before <code>grep<\/code> is finished, <code>grep<\/code> is killed by\nSIGPIPE and exits with that particular status. Some modernish functions,\nsuch as <code>harden<\/code> and <code>traverse<\/code>, need to handle such a SIGPIPE exit\nspecially to avoid unduly killing the program. The exact value of this\nexit status is shell-specific, so modernish runs a quick test to determine\nit at initialisation time.<br \/>If <code>SIGPIPE<\/code> was set to ignore by the process that invoked the current\nshell, <code>SIGPIPESTATUS<\/code> can't be detected and is set to the special value\n99999. See also the description of the\n<a href=\"#user-content-warning-ids\"><code>WRN_NOSIGPIPE<\/code><\/a>\nID for\n<a href=\"#user-content-feature-testing\"><code>thisshellhas<\/code><\/a>.<\/li>\n<li><code>$DEFPATH<\/code>: The default system path guaranteed to find compliant POSIX\nutilities, as given by <code>getconf PATH<\/code>.<\/li>\n<\/ul><h3><a id=\"user-content-control-character-whitespace-and-shell-safe-character-constants\" class=\"anchor\" aria-hidden=\"true\" href=\"#control-character-whitespace-and-shell-safe-character-constants\"><\/a>Control character, whitespace and shell-safe character constants<\/h3>\n<p>POSIX does not provide for the quoted C-style escape codes commonly used in\nbash, ksh and zsh (such as <code>$'\\n'<\/code> to represent a newline character),\nleaving the standard shell without a convenient way to refer to control\ncharacters. Modernish provides control character constants (read-only\nvariables) with hexadecimal suffixes <code>$CC01<\/code> .. <code>$CC1F<\/code> and <code>$CC7F<\/code>, as well as <code>$CCe<\/code>,\n<code>$CCa<\/code>, <code>$CCb<\/code>, <code>$CCf<\/code>, <code>$CCn<\/code>, <code>$CCr<\/code>, <code>$CCt<\/code>, <code>$CCv<\/code> (corresponding with\n<code>printf<\/code> backslash escape codes). This makes it easy to insert control\ncharacters in double-quoted strings.<\/p>\n<p>More convenience constants, handy for use in bracket glob patterns for use\nwith <code>case<\/code> or modernish <code>match<\/code>:<\/p>\n<ul><li><code>$CONTROLCHARS<\/code>: All ASCII control characters.<\/li>\n<li><code>$WHITESPACE<\/code>: All ASCII whitespace characters.<\/li>\n<li><code>$ASCIIUPPER<\/code>: The ASCII uppercase letters A to Z.<\/li>\n<li><code>$ASCIILOWER<\/code>: The ASCII lowercase letters a to z.<\/li>\n<li><code>$ASCIIALNUM<\/code>: The ASCII alphanumeric characters 0-9, A-Z and a-z.<\/li>\n<li><code>$SHELLSAFECHARS<\/code>: Safelist for shell-quoting.<\/li>\n<li><code>$ASCIICHARS<\/code>: The complete set of ASCII characters (minus NUL).<\/li>\n<\/ul><h2><a id=\"user-content-legibility-aliases\" class=\"anchor\" aria-hidden=\"true\" href=\"#legibility-aliases\"><\/a>Legibility aliases<\/h2>\n<p>A few aliases that seem to make the shell language look slightly friendlier:<\/p>\n<pre>alias not='! '              # more legible synonym for '!'\nalias so='[ \"$?\" -eq 0 ]'   # test preceding command's success with\n                            # 'if so;' or 'if not so;'\nalias forever='while :;'    # indefinite loops: forever do &lt;stuff&gt;; done\n<\/pre>\n<h2><a id=\"user-content-enhanced-exit\" class=\"anchor\" aria-hidden=\"true\" href=\"#enhanced-exit\"><\/a>Enhanced exit<\/h2>\n<p><code>exit<\/code>: extended usage: <code>exit<\/code> [ <code>-u<\/code> ] [ <em>status<\/em> [ <em>message<\/em> ] ]<\/p>\n<ul><li>If the <code>-u<\/code> option is given, the shell function <code>showusage<\/code> is called, which\nhas a simple default but can be redefined by your script. This function is\nintended to print a message showing how the command should be invoked. The\nfunction is run in a subshell and its output is redirected to standard error.<\/li>\n<li>Any remaining arguments after <em>status<\/em> are combined, separated by spaces,\nand taken as a <em>message<\/em> to print on exit. It is printed to standard error,\npreceded by the name of the current program (<code>$ME<\/code> minus directories).<\/li>\n<\/ul><h2><a id=\"user-content-reliable-emergency-halt\" class=\"anchor\" aria-hidden=\"true\" href=\"#reliable-emergency-halt\"><\/a>Reliable emergency halt<\/h2>\n<p><code>die<\/code>: reliably halt program execution, even from within subshells, optionally\nprinting an error message. Note that <code>die<\/code> is meant for an emergency program\nhalt only, i.e. in situations were continuing would mean the program is in an\ninconsistent or undefined state. Shell scripts running in an inconsistent or\nundefined state may wreak all sorts of havoc. They are also notoriously\ndifficult to terminate correctly, especially if the fatal error occurs within\na subshell: <code>exit<\/code> won't work then. That's why <code>die<\/code> is optimised for\nkilling <em>all<\/em> the program's processes (including subshells and external\ncommands launched by it) as quickly as possible. It should never be used for\nexiting the program normally.<\/p>\n<p>On interactive shells, <code>die<\/code> behaves differently. It does not kill or exit your\nshell; instead, it issues <code>SIGINT<\/code> to the shell to abort the execution of your\nrunning command(s), which is equivalent to pressing Ctrl+C.<\/p>\n<p>Usage: <code>die<\/code> [ <em>message<\/em> ]<\/p>\n<p>A special <code>DIE<\/code> pseudosignal can be trapped (using plain old <code>trap<\/code> or\n<a href=\"#user-content-the-trap-stack\"><code>pushtrap<\/code><\/a>)\nto perform emergency cleanup commands upon invoking <code>die<\/code>.\nOn interactive shells, <code>DIE<\/code> is simply an alias for <code>SIGINT<\/code>.\nOn non-interactive shells, <code>DIE<\/code> traps are separate. In order to kill the\nmalfunctioning program as quickly as possible (hopefully before it has a chance\nto delete all your data), <code>die<\/code> doesn't wait for those traps to complete before\nkilling the program. Instead, it executes each <code>DIE<\/code> trap simultaneously as a\nbackground job, then gathers the process IDs of the main shell and all its\nsubprocesses, sending <code>SIGKILL<\/code> to all of them except any <code>DIE<\/code> trap processes.<\/p>\n<p>(One case where <code>die<\/code> is limited is when the main shell program has exited,\nbut several runaway background processes that it forked are still going. If\n<code>die<\/code> is called by one of those background processes, then it will kill that\nbackground process and its subshells, but not the others. This is due to an\ninherent limitation in the design of POSIX operating systems. When the main\nshell exits, its surviving background processes are detached from the\nprocess hierarchy and become independent from one another, with no way to\ndetermine that they once belonged to the same program.)<\/p>\n<h2><a id=\"user-content-low-level-shell-utilities\" class=\"anchor\" aria-hidden=\"true\" href=\"#low-level-shell-utilities\"><\/a>Low-level shell utilities<\/h2>\n<p><code>insubshell<\/code>: easily check if you're currently running in a\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_12\" rel=\"nofollow\">subshell environment<\/a>\n(usually called simply <em>subshell<\/em>), that is, a copy of the parent shell that\nstarts out as an exact duplicate except for traps. This is not to be confused\nwith a newly initialised shell that is merely a child process of the current\nshell, which is sometimes (erroneously) called a \"subshell\" as well.<br \/>Usage: <code>insubshell<\/code> [ <code>-p<\/code> | <code>-u<\/code> ]<br \/>This function returns success (0) if it was called from within a subshell\nand non-success (1) if not. One of two options can be given:<\/p>\n<ul><li><code>-p<\/code>: Store the process ID (PID) of the current subshell or main shell\nin <code>REPLY<\/code>. (Note that on AT&amp;T ksh93, which does not fork a new process\nfor non-background subshells, that PID is same as the main shell's except\nfor background jobs.)<\/li>\n<li><code>-u<\/code>: Store an identifier in <code>REPLY<\/code> that is useful for determining if\nyou've entered a subshell relative to a previously stored identifier. The\ncontent and format are unspecified and shell-dependent.<\/li>\n<\/ul><p><code>setstatus<\/code>: manually set the exit status <code>$?<\/code> to the desired value. The\nfunction exits with the status indicated. This is useful in conditional\nconstructs if you want to prepare a particular exit status for a subsequent\n'exit' or 'return' command to inherit under certain circumstances.<\/p>\n<p><code>isvarname<\/code>: Check if argument is valid portable identifier in the shell,\nthat is, a portable variable name, shell function name or long-form shell\noption name. (Modernish requires portable names everywhere; for example,\naccented or non-Latin characters in variable names are not supported.)<\/p>\n<p><code>isset<\/code>: check if a variable, shell function or option is set. Usage:<\/p>\n<ul><li><code>isset<\/code> <em>varname<\/em>: Check if a variable is set.<\/li>\n<li><code>isset -v<\/code> <em>varname<\/em>: Id.<\/li>\n<li><code>isset -x<\/code> <em>varname<\/em>: Check if variable is exported.<\/li>\n<li><code>isset -r<\/code> <em>varname<\/em>: Check if variable is read-only.<\/li>\n<li><code>isset -f<\/code> <em>funcname<\/em>: Check if a shell function is set.<\/li>\n<li><code>isset -<\/code><em>optionletter<\/em> (e.g. <code>isset -C<\/code>): Check if shell option is set.<\/li>\n<li><code>isset -o<\/code> <em>optionname<\/em>: Check if shell option is set by long name.<\/li>\n<\/ul><p>Exit status: 0 if the item is set; 1 if not; 2 if the argument is not\nrecognised as a syntactically valid identifier.<\/p>\n<p>When checking a shell option, a nonexistent shell option is not an error,\nbut returns the same result as an unset shell option. (To check if a shell\noption exists, use <a href=\"#user-content-feature-testing\"><code>thisshellhas<\/code><\/a>.<\/p>\n<p>Note: just <code>isset -f<\/code> checks if shell option <code>-f<\/code> (a.k.a. <code>-o noglob<\/code>) is\nset, but with an extra argument, it checks if a shell function is set.\nSimilarly, <code>isset -x<\/code> checks if shell option <code>-x<\/code> (a.k.a <code>-o xtrace<\/code>)\nis set, but <code>isset -x<\/code> <em>varname<\/em> checks if a variable is exported. If you\nuse unquoted variable expansions here, make sure they're not empty, or\nthe shell's empty removal mechanism will cause the wrong thing to be checked\n(even in <code>use safe<\/code> mode).<\/p>\n<p><code>unexport<\/code>: the opposite of <code>export<\/code>. Unexport a variable while preserving\nits value, or (while working under <code>set -a<\/code>) don't export it at all.\nUsage is like <code>export<\/code>, with the caveat that variable assignment arguments\ncontaining non-shellsafe characters or expansions must be quoted as\nappropriate, unlike in some specific shell implementations of <code>export<\/code>.\n(To get rid of that headache, <a href=\"#user-content-use-safe\"><code>use safe<\/code><\/a>.)<\/p>\n<h2><a id=\"user-content-quoting-strings-for-subsequent-parsing-by-the-shell\" class=\"anchor\" aria-hidden=\"true\" href=\"#quoting-strings-for-subsequent-parsing-by-the-shell\"><\/a>Quoting strings for subsequent parsing by the shell<\/h2>\n<p><code>shellquote<\/code>: Quote the values of specified variables in such a way that the\nvalues are suitable for parsing by the shell as string literals. This is\nessential for the safe use of <code>eval<\/code> or any other context where the shell\nmust parse untrusted input. <code>shellquote<\/code> only uses quoting mechanisms\nspecified by POSIX, so the quoted values it produces are safe to parse\nin any POSIX shell.<br \/>Usage: <code>shellquote<\/code> [ <code>-f<\/code>|<code>+f<\/code> ] <em>varname<\/em> [ [ <code>-f<\/code>|<code>+f<\/code> ] <em>varname<\/em> ... ]<br \/>The values of the variables specified by name are shell-quoted and stored\nback into those variables. By default, a value is only quoted if it contains\ncharacters not present in <code>$SHELLSAFECHARS<\/code>. An <code>-f<\/code> argument forces\nunconditional quoting for subsequent variables; an <code>+f<\/code> argument restores\ndefault behaviour. <code>shellquote<\/code> returns success (0) if all variables were\nprocessed successfully, and non-success (1) if any undefined (unset)\nvariables were encountered. In the latter case, any set variables still get\ntheir values quoted.<\/p>\n<p><code>shellquoteparams<\/code>: shell-quote the current shell's positional parameters\nin-place.<\/p>\n<p><code>storeparams<\/code>: store the positional parameters, or a sub-range of them,\nin a variable, in a shellquoted form suitable for restoration using\n<code>eval \"set -- $varname\"<\/code>. For instance: <code>storeparams -f2 -t6 VAR<\/code>\nquotes and stores <code>$2<\/code> to <code>$6<\/code> in <code>VAR<\/code>.<\/p>\n<h2><a id=\"user-content-the-stack\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-stack\"><\/a>The stack<\/h2>\n<p><code>push<\/code> &amp; <code>pop<\/code>: every variable and shell option gets its own stack. For\nvariables, both the value and the set\/unset state is (re)stored. Usage:<\/p>\n<ul><li><code>push<\/code> [ <code>--key=<\/code><em>value<\/em> ] <em>item<\/em> [ <em>item<\/em> ... ]<\/li>\n<li><code>pop<\/code> [ <code>--keepstatus<\/code> ] [ <code>--key=<\/code><em>value<\/em> ] <em>item<\/em> [ <em>item<\/em> ... ]<\/li>\n<\/ul><p>where <em>item<\/em> is a valid portable variable name, a short-form shell option\n(dash plus letter), or a long-form shell option (<code>-o<\/code> followed by an option\nname, as two arguments). The precise shell options supported (other than the\nones guaranteed by POSIX) depend on the shell modernish is running on. For\ncross-shell compatibility, nonexistent shell options are treated as unset.<\/p>\n<p>Before pushing or popping anything, both functions check if all the given\narguments are valid and <code>pop<\/code> checks all items have a non-empty stack. This\nallows pushing and popping groups of items with a check for the integrity of\nthe entire group. <code>pop<\/code> exits with status 0 if all items were popped\nsuccessfully, and with status 1 if one or more of the given items could not\nbe popped (and no action was taken at all).<\/p>\n<p>The <code>--key=<\/code> option is an advanced feature that can help different modules\nor funtions to use the same variable stack safely. If a key is given to\n<code>push<\/code>, then for each <em>item<\/em>, the given key <em>value<\/em> is stored along with the\nvariable's value for that position in the stack. Subsequently, restoring\nthat value with <code>pop<\/code> will only succeed if the key option with the same key\nvalue is given to the <code>pop<\/code> invocation. Similarly, popping a keyless value\nonly succeeds if no key is given to <code>pop<\/code>. If there is any key mismatch, no\nchanges are made and pop returns status 2. For instance, if a function\npushes all its values with something like <code>--key=myfunction<\/code>, it can do a\nloop like <code>while pop --key=myfunction var; do ...<\/code> even if <code>var<\/code> already has\nother items on its stack that shouldn't be tampered with. Note that this is\na robustness\/convenience feature, not a security feature; the keys are not\nhidden in any way. (The <a href=\"#user-content-use-varsetlocal\">var\/setlocal<\/a>\nmodule, which provides stack-based local variables, internally makes use of\nthis feature.)<\/p>\n<p>If the <code>--keepstatus<\/code> option is given, <code>pop<\/code> will exit with the\nexit status of the command executed immediately prior to calling <code>pop<\/code>. This\ncan avoid the need for awkward workarounds when restoring variables or shell\noptions at the end of a function. However, note that this makes failure to pop\n(stack empty or key mismatch) a fatal error that kills the program, as <code>pop<\/code>\nno longer has a way to communicate this through its exit status.<\/p>\n<h3><a id=\"user-content-the-shell-options-stack\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-shell-options-stack\"><\/a>The shell options stack<\/h3>\n<p>The shell options stack allows saving and restoring the state of any shell\noption available to the <code>set<\/code> builtin using <code>push<\/code> and <code>pop<\/code> commands with\na syntax similar to that of <code>set<\/code>.<\/p>\n<p>Long-form shell options are matched to their equivalent short-form shell\noptions, if they exist. For instance, on all POSIX shells, <code>-f<\/code> is\nequivalent to <code>-o noglob<\/code>, and <code>push -o noglob<\/code> followed by <code>pop -f<\/code> works\ncorrectly. (This works even for shell-specific short &amp; long option\nequivalents; modernish internally does a check to find any equivalent.)<\/p>\n<p>On shells with a dynamic <code>no<\/code> option name prefix, that is on ksh, zsh and\nyash (where, for example, <code>noglob<\/code> is the opposite of <code>glob<\/code>), the <code>no<\/code>\nprefix is ignored, so something like <code>push -o glob<\/code> followed by <code>pop -o noglob<\/code> does the right thing. But this depends on the shell and should never\nbe used in cross-shell scripts.<\/p>\n<p>To facilitate cross-shell scripting, it is fine to push and pop shell\noptions that may not exist on the current shell; nonexistent options are\nconsidered unset for stack purposes.<\/p>\n<h3><a id=\"user-content-the-trap-stack\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-trap-stack\"><\/a>The trap stack<\/h3>\n<p><code>pushtrap<\/code> and <code>poptrap<\/code>: traps are now also stack-based, so that each\nprogram component or library module can set its own trap commands\nwithout interfering with others.<\/p>\n<p>Note an important difference between the trap stack and stacks for variables\nand shell options: pushing traps does not save them for restoring later, but\nadds them alongside other traps on the same signal. All pushed traps are\nactive at the same time and are executed from last-pushed to first-pushed\nwhen the respective signal is triggered. Traps cannot be pushed and popped\nusing <code>push<\/code> and <code>pop<\/code> but use dedicated commands as follows.<\/p>\n<p>Usage:<\/p>\n<ul><li><code>pushtrap<\/code> [ <code>--key=<\/code><em>value<\/em> ] [ <code>--<\/code> ] <em>command<\/em> <em>sigspec<\/em> [ <em>sigspec<\/em> ... ]<\/li>\n<li><code>poptrap<\/code> [ <code>--key=<\/code><em>value<\/em> ] [ <code>--<\/code> ] <em>sigspec<\/em> [ <em>sigspec<\/em> ... ]<\/li>\n<\/ul><p><code>pushtrap<\/code> works like regular <code>trap<\/code>, with the following exceptions:<\/p>\n<ul><li>Adds traps for a signal without overwriting previous ones.\n(However, any traps set prior to initialising modernish, or by bypassing\nthe modernish 'trap' alias to access the system command directly, <em>will<\/em>\nbe overwritten by a <code>pushtrap<\/code> for the same signal. To remedy this, you\ncan issue a simple <code>trap<\/code> command; as modernish prints the traps, it will\nquietly detect ones it doesn't yet know about and make them work nicely\nwith the trap stack.)<\/li>\n<li>Unlike regular traps, a stack-based trap does not cause a signal to be\nignored. Setting one will cause it to be executed upon the shell receiving\nthat signal, but after the stack traps complete execution, modernish re-sends\nthe signal to the main shell, causing it to behave as if no trap were set\n(unless a regular POSIX trap is also active).\nThus, <code>pushtrap<\/code> does not accept an empty <em>command<\/em> as it would be pointless.<\/li>\n<li>Each stack trap is executed in a new subshell to keep it from interfering\nwith others. This means a stack trap cannot change variables except within\nits own environment, and 'exit' will only exit the trap and not the program.<\/li>\n<li>Each stack trap is executed with <code>$?<\/code> initially set to the exit status\nthat was active at the time the signal was triggered.<\/li>\n<li>Stack traps do not have access to the positional parameters.<\/li>\n<li><code>pushtrap<\/code> stores current <code>$IFS<\/code> (field splitting) and <code>$-<\/code> (shell options)\nalong with the pushed trap. Within the subshell executing each stack trap,\nmodernish restores <code>IFS<\/code> and the shell options <code>f<\/code> (<code>noglob<\/code>), <code>u<\/code>\n(<code>nounset<\/code>) and <code>C<\/code> (<code>noclobber<\/code>) to the values in effect during the\ncorresponding <code>pushtrap<\/code>. This is to avoid unexpected effects in case a trap\nis triggered while temporary settings are in effect.<\/li>\n<li>The <code>--key<\/code> option applies the keying functionality inherited from\n<a href=\"#user-content-the-stack\">plain <code>push<\/code><\/a> to the trap stack.\nIt works the same way, so the description is not repeated here.<\/li>\n<\/ul><p><code>poptrap<\/code> takes just signal names or numbers as arguments. It takes the\nlast-pushed trap for each signal off the stack, storing the commands that\nwas set for those signals into the REPLY variable, in a format suitable for\nre-entry into the shell. Again, the <code>--key<\/code> option works as in\n<a href=\"#user-content-the-stack\">plain <code>pop<\/code><\/a>.<\/p>\n<p>Stack-based traps, like native shell traps, are reset upon entering a\nsubshell (such as a command substitution or a series of commands enclosed in\nparentheses). However, commands for printing traps will print the traps for\nthe parent shell, until another <code>trap<\/code>, <code>pushtrap<\/code> or <code>poptrap<\/code> command is\ninvoked, at which point all memory of the parent shell's traps is erased.<\/p>\n<h4><a id=\"user-content-trap-stack-compatibility-considerations\" class=\"anchor\" aria-hidden=\"true\" href=\"#trap-stack-compatibility-considerations\"><\/a>Trap stack compatibility considerations<\/h4>\n<p>Modernish tries hard to avoid incompatibilities with existing trap practice.\nTo that end, it intercepts the regular POSIX 'trap' command using an alias,\nreimplementing and interfacing it with the shell's builtin trap facility\nso that plain old regular traps play nicely with the trap stack. You should\nnot notice any changes in the POSIX 'trap' command's behaviour, except for\nthe following:<\/p>\n<ul><li>The regular 'trap' command does not overwrite stack traps (but does\noverwrite previous regular traps).<\/li>\n<li>Unlike zsh's native trap command, signal names are case insensitive.<\/li>\n<li>Unlike dash's native trap command, signal names may have the <code>SIG<\/code> prefix;\nthat prefix is quietly accepted and discarded.<\/li>\n<li>Setting an empty trap action to ignore a signal only works fully (passing\nthe ignoring on to child processes) if there are no stack traps associated\nwith the signal; otherwise, an empty trap action merely suppresses the\nsignal's default action for the current process -- e.g., after executing\nthe stack traps, it keeps the shell from exiting.<\/li>\n<li>The 'trap' command with no arguments, which prints the traps that are set\nin a format suitable for re-entry into the shell, now also prints the\nstack traps as 'pushtrap' commands. (<code>bash<\/code> users might notice the <code>SIG<\/code>\nprefix is not included in the signal names written.)<\/li>\n<li>The bash\/yash-style '-p' option, including its yash-style <code>--print<\/code>\nequivalent, is now supported on all shells. If further arguments are\ngiven after that option, they are taken as signal specifications and\nonly the commands to recreate the traps for those signals are printed.<\/li>\n<li>Saving the traps to a variable using command substitution (as in:\n<code>var=$(trap)<\/code>) now works on every shell supported by modernish, including\n(d)ash, mksh and zsh.<\/li>\n<li>To reset (unset) a trap, the modernish 'trap' command accepts both\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_28_03\" rel=\"nofollow\">valid POSIX syntax<\/a>\nand legacy bash\/(d)ash\/zsh syntax, like <code>trap INT<\/code> to unset a SIGINT\ntrap (which only works if the 'trap' command is given exactly one\nargument). Note that this is for compatibility with existing scripts only.<\/li>\n<\/ul><p>POSIX traps for each signal are always executed after that signal's stack-based\ntraps; this means they should not rely on modernish modules that use the trap\nstack to clean up after themselves on exit, as those cleanups would already\nhave been done.<\/p>\n<p>Modernish introduces a new <code>DIE<\/code> pseudosignal whose traps are\nexecuted upon invoking <code>die<\/code>.\nSee the <a href=\"#user-content-reliable-emergency-halt\"><code>die<\/code> description<\/a> for\nmore information.<\/p>\n<p>On interactive shells, <code>INT<\/code> traps (both POSIX and stack) are cleared out\nafter executing them once. This is because <a href=\"#user-content-reliable-emergency-halt\"><code>die<\/code><\/a>\nuses SIGINT for cleanup and command interruption on interactive shells,\nwith the <code>DIE<\/code> signal name turned into an alias for <code>INT<\/code>.\nAs a side effect of this special handling, <code>INT<\/code> traps on interactive\nshells do not have access to the positional parameters and cannot\nreturn from shell functions.<\/p>\n<h3><a id=\"user-content-positional-parameters-stack\" class=\"anchor\" aria-hidden=\"true\" href=\"#positional-parameters-stack\"><\/a>Positional parameters stack<\/h3>\n<p><code>pushparams<\/code> and <code>popparams<\/code>: push and pop the complete set of positional\nparameters. No arguments are supported.<\/p>\n<h3><a id=\"user-content-stack-query-and-maintenance-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#stack-query-and-maintenance-functions\"><\/a>Stack query and maintenance functions<\/h3>\n<p>For the four functions below, <em>item<\/em> can be:<\/p>\n<ul><li>a valid portable variable name<\/li>\n<li>a short-form shell option: dash plus letter<\/li>\n<li>a long-form shell option: <code>-o<\/code> followed by an option name (two arguments)<\/li>\n<li><code>@<\/code> to refer to the <a href=\"#user-content-positional-parameters-stack\">positional parameters stack<\/a><\/li>\n<li><code>--trap=<\/code><em>SIGNAME<\/em> to refer to the trap stack for the indicated signal<\/li>\n<\/ul><p><code>stackempty<\/code> [ <code>--key=<\/code><em>value<\/em> ] [ <code>--force<\/code> ] <em>item<\/em>: Tests if the stack\nfor an item is empty. Returns status 0 if it is, 1 if it is not. The key\nfeature works as in <a href=\"#user-content-the-stack\"><code>pop<\/code><\/a>: by default, a key\nmismatch is considered equivalent to an empty stack. If <code>--force<\/code> is given,\nthis function ignores keys altogether.<\/p>\n<p><code>stacksize<\/code> [ <code>--silent<\/code> | <code>--quiet<\/code> ] <em>item<\/em>: Leaves the size of a stack in\nthe <code>REPLY<\/code> variable and, if option <code>--silent<\/code> or <code>--quiet<\/code> is not given,\nwrites it to standard output.\nThe size of the complete stack is returned, even if some values are keyed.<\/p>\n<p><code>printstack<\/code> [ <code>--quote<\/code> ] <em>item<\/em>: Outputs a stack's content.\nOption <code>--quote<\/code> shell-quotes each stack value before printing it, allowing\nfor parsing multi-line or otherwise complicated values.\nColumn 1 to 7 of the output contain the number of the item (down to 0).\nIf the item is set, column 8 and 9 contain a colon and a space, and\nif the value is non-empty or quoted, column 10 and up contain the value.\nSets of values that were pushed with a key are started with a special\nline containing <code>--- key: <\/code><em>value<\/em>. A subsequent set pushed with no key is\nstarted with a line containing <code>--- (key off)<\/code>.\nReturns status 0 on success, 1 if that stack is empty.<\/p>\n<p><code>clearstack<\/code> [ <code>--key=<\/code><em>value<\/em> ] [ <code>--force<\/code> ] <em>item<\/em> [ <em>item<\/em> ... ]:\nClears one or more stacks, discarding all items on it.\nIf (part of) the stack is keyed or a <code>--key<\/code> is given, only clears until a\nkey mismatch is encountered. The <code>--force<\/code> option overrides this and always\nclears the entire stack (be careful, e.g. don't use within\n<a href=\"#user-content-use-varsetlocal\"><code>setlocal<\/code> ... <code>endlocal<\/code><\/a>).\nReturns status 0 on success, 1 if that stack was already empty, 2 if\nthere was nothing to clear due to a key mismatch.<\/p>\n<h2><a id=\"user-content-hardening-emergency-halt-on-error\" class=\"anchor\" aria-hidden=\"true\" href=\"#hardening-emergency-halt-on-error\"><\/a>Hardening: emergency halt on error<\/h2>\n<p><code>harden<\/code>: modernish's replacement for <code>set -e<\/code> a.k.a. <code>set -o errexit<\/code> (which is\n<a href=\"https:\/\/lists.gnu.org\/archive\/html\/bug-bash\/2012-12\/msg00093.html\" rel=\"nofollow\">fundamentally<\/a>\n<a href=\"http:\/\/mywiki.wooledge.org\/BashFAQ\/105\" rel=\"nofollow\">flawed<\/a>,\nnot supported and will break the library).<\/p>\n<p><code>harden<\/code> installs a shell function that hardens a particular command by\nchecking its exit status against values indicating error or system failure.\nExactly what exit statuses signify an error or failure depends on the\ncommand in question; this should be looked up in the\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/contents.html\" rel=\"nofollow\">POSIX specification<\/a>\n(under \"Utilities\") or in the command's <code>man<\/code> page or other documentation.<\/p>\n<p>If the command fails, the function installed by <code>harden<\/code> calls <code>die<\/code>, so it\nwill reliably halt program execution, even if the failure occurred within a\nsubshell (for instance, in a pipe construct or command substitution).<\/p>\n<p><code>harden<\/code> (along with <code>use safe<\/code>) is an essential feature for robust shell\nprogramming that current shells lack. In shell programs without modernish,\nproper error checking is too inconvenient and therefore rarely done. It's often\nrecommended to use <code>set -e<\/code> a.k.a <code>set -o errexit<\/code>, but that is broken in\nvarious strange ways (see links above) and the idea is often abandoned. So,\nall too often, shell programs simply continue in an inconsistent state after a\ncritical error occurs, occasionally wreaking serious havoc on the system.\nModernish <code>harden<\/code> was designed to help solve that problem properly.<\/p>\n<p>Usage:<\/p>\n<p><code>harden<\/code> [ <code>-f<\/code> <em>funcname<\/em> ] [ <code>-[cSpXtPE]<\/code> ] [ <code>-e<\/code> <em>testexpr<\/em> ]\n[ <em>var<\/em><code>=<\/code><em>value<\/em> ... ] [ <code>-u<\/code> <em>var<\/em> ... ] <em>command_name_or_path<\/em>\n[ <em>command_argument<\/em> ... ]<\/p>\n<p>The <code>-f<\/code> option hardens the command as the shell function <em>funcname<\/em> instead\nof defaulting to <em>command_name_or_path<\/em> as the function name. (If the latter\nis a path, that's always an invalid function name, so the use of <code>-f<\/code> is\nmandatory.)<\/p>\n<p>The <code>-c<\/code> option causes <em>command_name_or_path<\/em> to be hardened and run\nimmediately instead of setting a shell function for later use. This option\nis meant for commands that run once; it is not efficient for repeated use.\nIt cannot be used together with the <code>-f<\/code> option.<\/p>\n<p>The <code>-S<\/code> option allows specifying several possible names\/paths for a\ncommand. It causes the <em>command_name_or_path<\/em> to be split by comma and\ninterpreted as multiple names or paths to search. The first name or path\nfound is used. Requires <code>-f<\/code>.<\/p>\n<p>The <code>-e<\/code> option, which defaults to <code>&gt;0<\/code>, indicates the exit statuses\ncorresponding to a fatal error. It depends on the command what these are;\nconsult the POSIX spec and the manual pages.\nThe status test expression <em>testexpr<\/em>, argument\nto the <code>-e<\/code> option, is like a shell arithmetic\nexpression, with the binary operators <code>==<\/code> <code>!=<\/code> <code>&lt;=<\/code> <code>&gt;=<\/code> <code>&lt;<\/code> <code>&gt;<\/code> turned\ninto unary operators referring to the exit status of the command in\nquestion. Assignment operators are disallowed. Everything else is the same,\nincluding <code>&amp;&amp;<\/code> (logical and) and <code>||<\/code> (logical or) and parentheses.\nNote that the expression needs to be quoted as the characters used in it\nclash with shell grammar tokens.<\/p>\n<p>The <code>-X<\/code> option causes <code>harden<\/code> to always search for and harden an external\ncommand, even if a built-in command by that name exists.<\/p>\n<p>The <code>-E<\/code> option causes the hardening function to consider it a fatal error\nif the hardened command writes anything to the standard error stream. This\noption allows hardening commands (such as\n<a href=\"https:\/\/github.com\/modernish\/modernish\/blob\/master\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/bc.html#tag_20_09_14\"><code>bc<\/code><\/a>)\nwhere you can't rely on the exit status to detect an error. The text written\nto standard error is passed on as part of the error message printed by\n<code>die<\/code>. Note that:<\/p>\n<ul><li>Intercepting standard error necessitates that the command be executed from a\nsubshell. This means any builtins or shell functions hardened with <code>-E<\/code> cannot\ninfluence the calling shell (e.g. <code>harden -E cd<\/code> renders <code>cd<\/code> ineffective).<\/li>\n<li><code>-E<\/code> does not disable exit status checks; by default, any exit status greater\nthan zero is still considered a fatal error as well. If your command does not\neven reliably return a 0 status upon success, then you may want to add <code>-e '&gt;125'<\/code>, limiting the exit status check to reserved values indicating errors\nlaunching the command and signals caught.<\/li>\n<\/ul><p>The <code>-p<\/code> option causes <code>harden<\/code> to search for commands using the\nsystem default path (as obtained with <code>getconf PATH<\/code>) as opposed to the\ncurrent <code>$PATH<\/code>. This ensures that you're using a known-good external\ncommand that came with your operating system. By default, the system-default\nPATH search only applies to the command itself, and not to any commands that\nthe command may search for in turn. But if the <code>-p<\/code> option is specified at\nleast twice, or if the command is a shell function (hardened under another name\nusing <code>-f<\/code>), the command is run in a subshell with <code>PATH<\/code> exported as the\ndefault path, which is equivalent to adding a <code>PATH=$DEFPATH<\/code> assignment\nargument (see <a href=\"#user-content-important-note-on-variable-assignments\">below<\/a>).<\/p>\n<p>Examples:<\/p>\n<pre>harden make                           # simple check for status &gt; 0\nharden -f tar '\/usr\/local\/bin\/gnutar' # id.; be sure to use this 'tar' version\nharden -e '&gt; 1' grep                  # for grep, status &gt; 1 means error\nharden -e '==1 || &gt;2' gzip            # 1 and &gt;2 are errors, but 2 isn't (see manual)\n<\/pre>\n<h3><a id=\"user-content-important-note-on-variable-assignments\" class=\"anchor\" aria-hidden=\"true\" href=\"#important-note-on-variable-assignments\"><\/a>Important note on variable assignments<\/h3>\n<p>As far as the shell is concerned, hardened commands are shell functions and\nnot external or builtin commands. This essentially changes one behaviour of\nthe shell: variable assignments preceding the command will not be local to\nthe command as usual, but <em>will persist<\/em> after the command completes.\n(POSIX technically makes that behaviour\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_09_01\" rel=\"nofollow\">optional<\/a>\nbut all current shells behave the same in POSIX mode.)<\/p>\n<p>For example, this means that something like<\/p>\n<pre>harden -e '&gt;1' grep\n# [...]\nLC_ALL=C grep regex some_ascii_file.txt\n<\/pre>\n<p>should never be done, because the meant-to-be-temporary <code>LC_ALL<\/code> locale\nassignment will persist and is likely to cause problems further on.<\/p>\n<p>To solve this problem, <code>harden<\/code> supports adding these assignments as\npart of the hardening command, so instead of the above you do:<\/p>\n<pre>harden -e '&gt;1' LC_ALL=C grep\n# [...]\ngrep regex some_ascii_file.txt\n<\/pre>\n<p>With the <code>-u<\/code> option, <code>harden<\/code> also supports unsetting variables for the\nduration of a command, e.g.:<\/p>\n<pre>harden -e '&gt;1' -u LC_ALL grep\n<\/pre>\n<p>Pitfall alert: if the <code>-u<\/code> option is used, this causes the hardened command to\nrun in a subshell with those variables unset, because using a subshell is the\nonly way to avoid altering those variables' state in the main shell. This is\nusually fine, but note that a builtin command hardened with use of <code>-u<\/code> cannot\ninfluence the calling shell. For instance, something like <code>harden -u LC_ALL cd<\/code>\nrenders <code>cd<\/code> ineffective: the working directory is only changed within the\nsubshell which is then immediately left.<\/p>\n<p>The same happens if you harden a shell function under another name using\n<code>-f<\/code> while adding environment variable assignments (or using the <code>-p<\/code>\noption, which effectively adds <code>PATH=$DEFPATH<\/code>). The hardened function\nwill not be able to influence the main shell. Also note that the hardening\nfunction will export the assigned environment variables for the duration of\nthat subshell, so those variables will be inherited by any external command\nrun from the hardened function. (While hardening shell functions using\n<code>harden<\/code> is possible, it's not really recommended and it's better to call\n<code>die<\/code> directly in your shell function upon detecting an error.)<\/p>\n<h3><a id=\"user-content-hardening-while-allowing-for-broken-pipes\" class=\"anchor\" aria-hidden=\"true\" href=\"#hardening-while-allowing-for-broken-pipes\"><\/a>Hardening while allowing for broken pipes<\/h3>\n<p>If you're piping a command's output into another command that may close\nthe pipe before the first command is finished, you can use the <code>-P<\/code> option\nto allow for this:<\/p>\n<pre>harden -e '==1 || &gt;2' -P gzip       # also tolerate gzip being killed by SIGPIPE\ngzip -dc file.txt.gz | head -n 10\t# show first 10 lines of decompressed file\n<\/pre>\n<p><code>head<\/code> will close the pipe of <code>gzip<\/code> input after ten lines; the operating\nsystem kernel then kills <code>gzip<\/code> with the PIPE signal before it's finished,\ncausing a particular exit status that is greater than 128. This exit status\nwould normally make <code>harden<\/code> kill your entire program, which in the example\nabove is clearly not the desired behaviour. If the exit status caused by a\nbroken pipe were known, you could specifically allow for that exit status in\nthe status expression. The trouble is that this exit status varies depending\non the shell and the operating system. The <code>-p<\/code> option was made to solve\nthis problem: it automatically detects and whitelists the correct exit\nstatus corresponding to SIGPIPE termination on the current system.<\/p>\n<p>Tolerating SIGPIPE is an option and not the default, because in many\ncontexts it may be entirely unexpected and a symptom of a severe error if a\ncommand is killed by a broken pipe. It is up to the programmer to decide\nwhich commands should expect SIGPIPE and which shouldn't.<\/p>\n<p><em>Tip:<\/em> It could happen that the same command should expect SIGPIPE in one\ncontext but not another. You can create two hardened versions of the same\ncommand, one that tolerates SIGPIPE and one that doesn't. For example:<\/p>\n<pre>harden -f hardGrep -e '&gt;1' grep     # hardGrep does not tolerate being aborted\nharden -f pipeGrep -e '&gt;1' -P grep  # pipeGrep for use in pipes that may break\n<\/pre>\n<p><em>Note:<\/em> If <code>SIGPIPE<\/code> was set to ignore by the process invoking the current\nshell, the <code>-p<\/code> option has no effect, because no process or subprocess of\nthe current shell can ever be killed by <code>SIGPIPE<\/code>. However, this may cause\nvarious other problems and you may want to refuse to let your program run\nunder that condition.\n<a href=\"#user-content-warning-ids\"><code>thisshellhas WRN_NOSIGPIPE<\/code><\/a> can help\nyou easily detect that condition so your program can make a decision. See\nthe <a href=\"#user-content-warnig-ids\">WRN_NOSIGPIPE description<\/a> for more information.<\/p>\n<h3><a id=\"user-content-tracing-the-execution-of-hardened-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#tracing-the-execution-of-hardened-commands\"><\/a>Tracing the execution of hardened commands<\/h3>\n<p>The <code>-t<\/code> option will trace command output. Each execution of a command\nhardened with <code>-t<\/code> causes the full command line to be output to standard\nerror, in the following format:<\/p>\n<pre>[functionname]&gt; commandline\n<\/pre>\n<p>where <code>functionname<\/code> is the name of the shell function used to harden the\ncommand and <code>commandline<\/code> is the complete and actual command executed. The\n<code>commandline<\/code> is properly shell-quoted in a format suitable for re-entry\ninto the shell (which is an enhancement over the builtin tracing facility on\nmost shells). If standard error is on a terminal that supports ANSI colours,\nthe tracing output will be colourised.<\/p>\n<p>The <code>-t<\/code> option was added to <code>harden<\/code> because the commands that you harden\nare often the same ones you would be particularly interested in tracing. The\nadvantage of using <code>harden -t<\/code> over the shell's builtin tracing facility\n(<code>set -x<\/code> or <code>set -o xtrace<\/code>) is that the output is a <em>lot<\/em> less noisy,\nespecially when using a shell library such as modernish.<\/p>\n<p><em>Note:<\/em> Internally, <code>-t<\/code> uses the shell file descriptor 9, redirecting it to\nstandard error (using <code>exec 9&gt;&amp;2<\/code>). This allows tracing to continue to work\nnormally even for commands that redirect standard error to a file (which is\nanother enhancement over <code>set -x<\/code> on most shells). However, this does mean\n<code>harden -t<\/code> conflicts with any other use of the file descriptor 9 in your\nshell program.<\/p>\n<p>If file descriptor 9 is already open before <code>harden<\/code> is called, <code>harden<\/code>\ndoes not attempt to override this. This means tracing may be redirected\nelsewhere by doing something like <code>exec 9&gt;trace.out<\/code> before calling\n<code>harden<\/code>. (Note that redirecting FD 9 on the <code>harden<\/code> command itself will\n<em>not<\/em> work as it won't survive the run of the command.)<\/p>\n<h2><a id=\"user-content-simple-tracing-of-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#simple-tracing-of-commands\"><\/a>Simple tracing of commands<\/h2>\n<p>Sometimes you just want to trace the execution of some specific commands as\nin <code>harden -t<\/code> (see above) without actually hardening them against command\nerrors; you might prefer to do your own error handling. <code>trace<\/code> makes this\neasy. It is modernish's replacement or complement for <code>set -x<\/code> a.k.a. <code>set -o xtrace<\/code>.<\/p>\n<p><code>trace<\/code> is actually a shortcut for\n<code>harden -t -P -e '&gt;125 &amp;&amp; !=255'<\/code> <em>commandname<\/em>. The\nresult is that the indicated command is automatically traced upon execution.\nOther options, including <code>-f<\/code>, <code>-c<\/code> and environment variable assignments, are\nas in <code>harden<\/code>.<\/p>\n<p>A bonus is that you still get minimal hardening against fatal system errors.\nErrors in the traced command itself are ignored, but your program is\nimmediately halted with an informative error message if the traced command:<\/p>\n<ul><li>cannot be found (exit status 127);<\/li>\n<li>was found but cannot be executed (exit status 126);<\/li>\n<li>was killed by a signal other than <code>SIGPIPE<\/code> (exit status &gt; 128, except\nthe shell-specific exit status for <code>SIGPIPE<\/code>, and except 255 which is\nused by some utilities, such as <code>ssh<\/code> and <code>rsync<\/code>, to return an error).<\/li>\n<\/ul><p><em>Note:<\/em> The caveat for command-local variable assignments for <code>harden<\/code> also\napplies to <code>trace<\/code>. See\n<a href=\"#user-content-important-note-on-variable-assignments\">Important note on variable assignments<\/a>\nabove.<\/p>\n<h2><a id=\"user-content-external-commands-without-full-path\" class=\"anchor\" aria-hidden=\"true\" href=\"#external-commands-without-full-path\"><\/a>External commands without full path<\/h2>\n<p><code>extern<\/code> is like <code>command<\/code> but always runs an external command, without\nhaving to know or determine its location. This provides an easy way to\nbypass a builtin, alias or function. It does the same <code>$PATH<\/code> search\nthe shell normally does when running an external command. For instance, to\nguarantee running external <code>printf<\/code> just do: <code>extern printf ...<\/code><\/p>\n<p>Usage: <code>extern<\/code> [ <code>-p<\/code> ] [ <code>-v<\/code> ] <em>command<\/em> [ <em>argument<\/em> ... ]<\/p>\n<ul><li><code>-p<\/code>: use the operating system's default <code>PATH<\/code> (as determined by <code>getconf PATH<\/code>) instead of your current <code>$PATH<\/code> for the command search. This guarantees\na path that finds all the standard utilities defined by POSIX, akin to\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/command.html#tag_20_22_04\" rel=\"nofollow\"><code>command -p<\/code><\/a>\nbut still guaranteeing an external command.\n(Note that <code>extern -p<\/code> is more reliable than <code>command -p<\/code> because many\nshell binaries don't ask the OS for the default path and have a wrong\ndefault path hard-coded in.)<\/li>\n<li><code>-v<\/code>: don't execute <em>command<\/em> but show the full path name of the command that\nwould have been executed. Any extra <em>argument<\/em>s are taken as more command\npaths to show, one per line. <code>extern<\/code> exits with status 0 if all the commands\nwere found, 1 otherwise. This option can be combined with <code>-p<\/code>.<\/li>\n<\/ul><h2><a id=\"user-content-outputting-strings\" class=\"anchor\" aria-hidden=\"true\" href=\"#outputting-strings\"><\/a>Outputting strings<\/h2>\n<p><code>putln<\/code>: prints each argument on a separate line. There is no processing of\noptions or escape codes. (Modernish constants <code>$CCn<\/code>, etc. can be used to insert\ncontrol characters in double-quoted strings. To process escape codes, use\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/printf.html\" rel=\"nofollow\"><code>printf<\/code><\/a>\ninstead.)<\/p>\n<p><code>put<\/code>: prints each argument separated by a space, without a trailing separator\nor newline. Again, there is no processing of options or escape codes.<\/p>\n<p><code>echo<\/code>: This command is notoriously unportable and kind of broken, so is\n<strong>deprecated<\/strong> in favour of <code>put<\/code> and <code>putln<\/code>. Modernish does provide its own\nversion of <code>echo<\/code>, but it is <em>only<\/em> activated if modernish is in the hashbang\npath or otherwise is itself used as the shell (the \"most portable\" way of\nrunning programs\n<a href=\"#user-content-two-basic-forms-of-a-modernish-program\">explained above<\/a>).\nIf your script runs on a specific shell and sources modernish as a dot script\n(<code>. modernish<\/code>), or if you use modernish interactively in your shell profile,\nthe shell-specific version of <code>echo<\/code> is left intact. This is to make it\npossible to add modernish to existing shell-specific scripts without breaking\nanything, while still providing one consistent <code>echo<\/code> for cross-shell scripts.<\/p>\n<p>The modernish version of <code>echo<\/code>, if active, does not interpret any escape codes\nand supports only one option, <code>-n<\/code>, which, like BSD <code>echo<\/code>, suppresses the\nfinal newline. However, unlike BSD <code>echo<\/code>, if <code>-n<\/code> is the only argument, it is\nnot interpreted as an option and the string <code>-n<\/code> is printed instead. This makes\nit safe to output arbitrary data using this version of <code>echo<\/code> as long as it is\ngiven as a single argument (using quoting if needed).<\/p>\n<h2><a id=\"user-content-enhanced-dot-scripts\" class=\"anchor\" aria-hidden=\"true\" href=\"#enhanced-dot-scripts\"><\/a>Enhanced dot scripts<\/h2>\n<p><code>source<\/code>: bash\/zsh-style <code>source<\/code> command now available to all POSIX\nshells, complete with optional positional parameters given as extra\narguments (which is not supported by POSIX <code>.<\/code>).<\/p>\n<h2><a id=\"user-content-testing-numbers-strings-and-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#testing-numbers-strings-and-files\"><\/a>Testing numbers, strings and files<\/h2>\n<p>Complete replacement for <code>test<\/code>\/<code>[<\/code> in the form of speed-optimised shell\nfunctions, so modernish scripts never need to use that <code>[<\/code> botch again.\nInstead of inherently ambiguous <code>[<\/code> syntax (or the nearly-as-confusing\n<code>[[<\/code> one), these familiar shell syntax to get more functionality, including:<\/p>\n<h3><a id=\"user-content-integer-number-arithmetic-tests-and-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#integer-number-arithmetic-tests-and-operations\"><\/a>Integer number arithmetic tests and operations<\/h3>\n<p><code>let<\/code>: implementation of <code>let<\/code> as in ksh, bash and zsh, now available to all\nPOSIX shells. This makes C-based signed integer arithmetic evaluation\navailable to every supported shell, with the exception of the unary \"++\" and\n\"--\" operators (which have been given the capability designation ARITHPP).\nThis means <code>let<\/code> should be used for operations and tests, e.g. both\n<code>let \"x=5\"<\/code> and <code>if let \"x==5\"; then<\/code>... are supported (note single = for\nassignment, double == for comparison).\nSee POSIX\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_06_04\" rel=\"nofollow\">2.6.4 Arithmetic Expansion<\/a>\nfor more information on the supported operators.<\/p>\n<p><code>isint<\/code>: test if a given argument is a decimal, octal or hexadecimal integer\nnumber in valid POSIX shell syntax, ignoring leading (but not trailing) spaces\nand tabs.<\/p>\n<h3><a id=\"user-content-string-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#string-tests\"><\/a>String tests<\/h3>\n<pre>empty:        test if string is empty\nidentic:      test if 2 strings are identical\ncontains:     test if string 1 contains string 2\nstartswith:   test if string 1 starts with string 2\nendswith:     test if string 1 ends with string 2\nmatch:        test if string matches a glob pattern\nematch:       test if string matches an extended regex\n<\/pre>\n<h3><a id=\"user-content-file-type-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#file-type-tests\"><\/a>File type tests<\/h3>\n<p>These avoid the snags with symlinks you get with <code>[<\/code> and <code>[[<\/code>.\nBy default, symlinks are <em>not<\/em> followed. Add <code>-L<\/code> to operate on files\npointed to by symlinks instead of symlinks themselves (the <code>-L<\/code> makes\nno difference if the operands are not symlinks).<\/p>\n<pre>is present:    test if file exists (yields true even if invalid symlink)\nis -L present: test if file exists and is not an invalid symlink\nis sym:        test if file is symlink\nis -L sym:     test if file is a valid symlink\nis reg:        test if file is a regular file\nis -L reg:     test if file is regular or a symlink pointing to a regular\nis dir:        test if file is a directory\nis -L dir:     test if file is dir or symlink pointing to dir\nis fifo, is -L fifo, is socket, is -L socket, is blockspecial,\n               is -L blockspecial, is charspecial, is -L charspecial:\n               same pattern, you figure it out :)\n<\/pre>\n<h3><a id=\"user-content-file-comparison-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#file-comparison-tests\"><\/a>File comparison tests<\/h3>\n<p>By default, symlinks are <em>not<\/em> followed. Again, add <code>-L<\/code> to follow them.<\/p>\n<pre>is newer:      test if file 1 is newer than file 2 (or if file 1 exists,\n               but file 2 doesn't)\nis older:      test if file 1 is older than file 2 (or if file 1 doesn't\n               exist, but file 2 does)\nis samefile:   test if file 1 and file 2 are the same file (hardlinks)\nis onsamefs:   test if file 1 and file 2 are on the same file system (for\n               non-regular, non-directory files, test the parent directory)\nis -L newer, is -L older, is -L samefile, is -L onsamefs:\n               same as above, but after resolving symlinks\n<\/pre>\n<h3><a id=\"user-content-file-status-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#file-status-tests\"><\/a>File status tests<\/h3>\n<p>Symlinks are followed.<\/p>\n<pre>is nonempty:   test is file exists, is not an invalid symlink, and is\n               not empty (also works for dirs with read permission)\nis setuid:     test if file has user ID bit set\nis setgid:     test if file has group ID bit set\n<\/pre>\n<h3><a id=\"user-content-io-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#io-tests\"><\/a>I\/O tests<\/h3>\n<pre>is onterminal: test if file descriptor is associated with a terminal\n<\/pre>\n<h3><a id=\"user-content-file-permission-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#file-permission-tests\"><\/a>File permission tests<\/h3>\n<p>These use a more straightforward logic than <code>[<\/code> and <code>[[<\/code>.\nAny symlinks given are resolved, as these tests would be meaningless\nfor a symlink itself.<\/p>\n<pre>can read:      test if we have read permission for a file\ncan write:     test if we have write permission for a file or directory\n               (for directories, only true if traverse permission as well)\ncan exec:      test if we have execute permission for a regular file\ncan traverse:  test if we can enter (traverse through) a directory\n<\/pre>\n<h2><a id=\"user-content-basic-string-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-string-operations\"><\/a>Basic string operations<\/h2>\n<p>The main modernish library contains functions for a few basic string\nmanipulation operations (because they are needed by other functions in the main\nlibrary). Currently these are:<\/p>\n<h3><a id=\"user-content-touppertolower\" class=\"anchor\" aria-hidden=\"true\" href=\"#touppertolower\"><\/a>toupper\/tolower<\/h3>\n<pre>toupper:       convert all letters to upper case\ntolower:       convert all letters to lower case\n<\/pre>\n<p>If no arguments are given, <code>toupper<\/code> and <code>tolower<\/code> copy standard input to\nstandard output, converting case.<\/p>\n<p>If one or more arguments are given, they are taken as variable names (note:\nthey should be given without the <code>$<\/code>) and case is converted in the contents\nof the specified variables, without reading input or writing output.<\/p>\n<p><code>toupper<\/code> and <code>tolower<\/code> try hard to use the fastest available method on the\nparticular shell your program is running on. They use built-in shell\nfunctionality where available and working correctly, otherwise they fall back\non running an external utility.<\/p>\n<p>Which external utility is chosen depends on whether the current locale uses\nthe Unicode UTF-8 character set or not. For non-UTF-8 locales, modernish\nassumes the POSIX\/C locale and <code>tr<\/code> is always used. For UTF-8 locales,\nmodernish tries hard to find a way to correctly convert case even for\nnon-Latin alphabets. A few shells have this functionality built in with\n<code>typeset<\/code>. The rest need an external utility. Even in 2017, it is a real\nchallenge to find an external utility on an arbitrary POSIX-compliant system\nthat will correctly convert case for all applicable UTF-8 characters.\nModernish initialisation tries <code>tr<\/code>, <code>awk<\/code>, GNU <code>awk<\/code> and GNU <code>sed<\/code> before\ngiving up and issuing a\n<a href=\"#user-content-warning-ids\">warning ID<\/a>.\nIf <code>thisshellhas WRN_2UP2LOW<\/code>, it\nmeans modernish is in a UTF-8 locale but has not found a way to convert\n<strong>C<\/strong>ase for <strong>NON ASCII<\/strong> characters, so <code>toupper<\/code> and <code>tolower<\/code> will convert\nonly ASCII characters and leave any other characters in the string alone.<\/p>\n<h2><a id=\"user-content-basic-system-utilities\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-system-utilities\"><\/a>Basic system utilities<\/h2>\n<p>Small utilities that should have been part of the standard shell, but\naren't. Since their implementation is inexpensive, they are part of the main\nlibrary instead of a module.<\/p>\n<p><code>mkcd<\/code>: make one or more directories, then, upon success, change into the\nlast-mentioned one. <code>mkcd<\/code> inherits <code>mkdir<\/code>'s usage, so options depend on\nyour system's <code>mkdir<\/code>; only the\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/mkdir.html#tag_20_79_04\" rel=\"nofollow\">POSIX options<\/a>\nare guaranteed.<\/p>\n<h2><a id=\"user-content-modules\" class=\"anchor\" aria-hidden=\"true\" href=\"#modules\"><\/a>Modules<\/h2>\n<p><code>use<\/code>: use a modernish module. It implements a simple Perl-like module\nsystem with names such as 'safe', 'var\/setlocal' and 'loop\/select'.\nThese correspond to files 'safe.mm', 'var\/setlocal.mm', etc. which are\ndot scripts defining functionality. Any extra arguments to the <code>use<\/code>\ncommand are passed on to the dot script unmodified, so modules can\nimplement option parsing to influence their initialisation.<\/p>\n<h3><a id=\"user-content-use-safe\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-safe\"><\/a>use safe<\/h3>\n<p>Does <code>IFS=''; set -f -u -C<\/code>, that is: field splitting and globbing are\ndisabled, variables must be defined before use, and output redirection using\n<code>&gt;<\/code> will not overwrite existing files (use <code>&gt;|<\/code> to allow overwriting).<\/p>\n<p>Essentially, this is a whole new way of shell programming,\neliminating most variable quoting headaches, protects against typos\nin variable names wreaking havoc, and protects files from being\naccidentally overwritten by output redirection.<\/p>\n<p>Of course, you don't get field splitting and globbing. But modernish\nprovides various ways of enabling one or both only for the commands\nthat need them, <code>setlocal<\/code>...<code>endlocal<\/code> blocks chief among them\n(see <a href=\"#user-content-use-varsetlocal\"><code>use var\/setlocal<\/code><\/a> below).<\/p>\n<p>On interactive shells (or if <code>use safe -i<\/code> is given), also loads\nconvenience functions <code>fsplit<\/code> and <code>glob<\/code> to control and inspect the\nstate of field splitting and globbing in a more user friendly way.<\/p>\n<p><em>It is highly recommended that new modernish scripts start out with <code>use safe<\/code>.<\/em>\nBut this mode is not enabled by default because it will totally break\ncompatibility with shell code written for default shell settings.<\/p>\n<h3><a id=\"user-content-use-vararith\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-vararith\"><\/a>use var\/arith<\/h3>\n<p>These shortcut functions are alternatives for using 'let'.<\/p>\n<h4><a id=\"user-content-arithmetic-operator-shortcuts\" class=\"anchor\" aria-hidden=\"true\" href=\"#arithmetic-operator-shortcuts\"><\/a>Arithmetic operator shortcuts<\/h4>\n<p><code>inc<\/code>, <code>dec<\/code>, <code>mult<\/code>, <code>div<\/code>, <code>mod<\/code>: simple integer arithmetic shortcuts. The first\nargument is a variable name. The optional second argument is an\narithmetic expression, but a sane default value is assumed (1 for inc\nand dec, 2 for mult and div, 256 for mod). For instance, <code>inc X<\/code> is\nequivalent to <code>X=$((X+1))<\/code> and <code>mult X Y-2<\/code> is equivalent to <code>X=$((X*(Y-2)))<\/code>.<\/p>\n<p><code>ndiv<\/code> is like <code>div<\/code> but with correct rounding down for negative numbers.\nStandard shell integer division simply chops off any digits after the\ndecimal point, which has the effect of rounding down for positive numbers\nand rounding up for negative numbers. <code>ndiv<\/code> consistently rounds down.<\/p>\n<h4><a id=\"user-content-arithmetic-comparison-shortcuts\" class=\"anchor\" aria-hidden=\"true\" href=\"#arithmetic-comparison-shortcuts\"><\/a>Arithmetic comparison shortcuts<\/h4>\n<p>These have the same name as their <code>test<\/code>\/<code>[<\/code> option equivalents. Unlike\nwith <code>test<\/code>, the arguments are shell integer arith expressions, which can be\nanything from simple numbers to complex expressions. As with <code>$(( ))<\/code>,\nvariable names are expanded to their values even without the <code>$<\/code>.<\/p>\n<pre>Function:         Returns successfully if:\neq &lt;expr&gt; &lt;expr&gt;  the two expressions evaluate to the same number\nne &lt;expr&gt; &lt;expr&gt;  the two expressions evaluate to different numbers\nlt &lt;expr&gt; &lt;expr&gt;  the 1st expr evaluates to a smaller number than the 2nd\nle &lt;expr&gt; &lt;expr&gt;  the 1st expr eval's to smaller than or equal to the 2nd\ngt &lt;expr&gt; &lt;expr&gt;  the 1st expr evaluates to a greater number than the 2nd\nge &lt;expr&gt; &lt;expr&gt;  the 1st expr eval's to greater than or equal to the 2nd\n<\/pre>\n<h3><a id=\"user-content-use-varmapr\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-varmapr\"><\/a>use var\/mapr<\/h3>\n<p><code>mapr<\/code> (map records): Read delimited records from the standard input, invoking\na <em>callback<\/em> command with each input record as an argument and with up to\n<em>quantum<\/em> arguments at a time. By default, an input record is one line of text.<\/p>\n<p>Usage: <code>mapr<\/code> [ <code>-d<\/code> <em>delimiter<\/em> | <code>-D<\/code> ] [ <code>-n<\/code> <em>count<\/em> ] [ -s <em>count<\/em> ]\n[ -c <em>quantum<\/em> ] <em>callback<\/em> [ <em>argument<\/em> ... ]<\/p>\n<p>Options:<\/p>\n<ul><li><code>-d<\/code> <em>delimiter<\/em>: Use the single character <em>delimiter<\/em> to delimit input records,\ninstead of the newline character. A <code>NUL<\/code> (0) character and multibyte\ncharacters are not supported.<\/li>\n<li><code>-P<\/code>: Paragraph mode. Input records are delimited by sequences consisting of\na newline plus one or more blank lines, and leading or trailing blank lines\nwill not result in empty records at the beginning or end of the input. Cannot\nbe used together with -d.<\/li>\n<li><code>-s<\/code> <em>number<\/em>: Skip and discard the first <em>count<\/em> records read.<\/li>\n<li><code>-n<\/code> <em>number<\/em>: Stop processing after passing a total of <em>number<\/em> records to\ninvocation(s) of <em>callback<\/em>. If <code>-n<\/code> is not supplied or <em>number<\/em> is 0, all\nrecords are passed, except those skipped using <code>-s<\/code>.<\/li>\n<li><code>-m<\/code> <em>length<\/em>: Pass at most <em>length<\/em> bytes of arguments to each call to\n<em>callback<\/em>. The default value depends on constraints set by the operating\nsystem. The length of each argument is rounded up to a multiple of 8 bytes.\nIf <em>length<\/em> is 0, this limit is disabled.<\/li>\n<li><code>-c<\/code> <em>quantum<\/em>: Pass at most <em>quantum<\/em> arguments at a time to each call to\n<em>callback<\/em>. If <code>-c<\/code> is not supplied or if <em>quantum<\/em> is 0, the number of\narguments per invocation is not limited except by <code>-m<\/code>; whichever limit is\nreached first applies.<\/li>\n<\/ul><p>Arguments:<\/p>\n<ul><li><em>callback<\/em>: Call the <em>callback<\/em> command with the collected arguments each\ntime <em>quantum<\/em> lines are read. The callback command may be a shell function or\nany other kind of command, and is executed from the same shell environment\nthat invoked <code>mapr<\/code>. If the callback command exits with status 255, <code>mapr<\/code>\naborts processing and exits with status 1. If the callback command is\ninterrupted by the SIGPIPE signal, <code>mapr<\/code> aborts processing and exits with\nstatus <code>$SIGPIPESTATUS<\/code>. Other than that, it is a\n<a href=\"#user-content-reliable-emergency-halt\">fatal error<\/a>\nfor the callback command to exit with a status &gt; 0.<\/li>\n<li><em>argument<\/em>:  If there are extra arguments supplied on the mapr command line,\nthey will be added before the collected arguments on each invocation on the\ncallback command.<\/li>\n<\/ul><h4><a id=\"user-content-differences-from-mapfile\" class=\"anchor\" aria-hidden=\"true\" href=\"#differences-from-mapfile\"><\/a>Differences from <code>mapfile<\/code><\/h4>\n<p><code>mapr<\/code> was inspired by the bash 4.x builtin command <code>mapfile<\/code> a.k.a.\n<code>readarray<\/code>, and uses similar options, but there are important differences.<\/p>\n<ul><li><code>mapr<\/code> passes all the records as arguments to the callback command.<\/li>\n<li><code>mapr<\/code> does not support assigning records directly to an array. Instead,\nall handling is done through the callback command (which could be a shell\nfunction that assigns its arguments to an array.)<\/li>\n<li>The callback command is specified directly instead of with a <code>-C<\/code> option,\nand it may consist of several arguments (as with <code>xargs<\/code>).<\/li>\n<li>The record separator itself is never included in the arguments passed\nto the callback command (so there is no <code>-t<\/code> option to remove it).<\/li>\n<li><code>mapr<\/code> supports paragraph mode.<\/li>\n<li>If the callback command exits unsuccessfully, this is treated as a fatal\nerror, except that status 255 merely aborts processing.<\/li>\n<\/ul><h4><a id=\"user-content-differences-from-xargs\" class=\"anchor\" aria-hidden=\"true\" href=\"#differences-from-xargs\"><\/a>Differences from <code>xargs<\/code><\/h4>\n<p><code>mapr<\/code> shares important characteristics with\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/xargs.html\" rel=\"nofollow\"><code>xargs<\/code><\/a>\nwhile avoiding its myriad pitfalls.<\/p>\n<ul><li>Instead of being an external utility, <code>mapr<\/code> is fully integrated into the\nshell. The callback command can be a shell function or builtin, which can\ndirectly modify the shell environment.<\/li>\n<li><code>mapr<\/code> is line-oriented by default, so it is safe to use for input\narguments that contain spaces or tabs.<\/li>\n<li><code>mapr<\/code> does not parse or modify the input arguments in any way, e.g. it\ndoes not process and remove quotes from them like <code>xargs<\/code> does.<\/li>\n<li><code>mapr<\/code> supports paragraph mode.<\/li>\n<li>If the callback command exits unsuccessfully, this is treated as a fatal\nerror, except that (like with <code>xargs<\/code>) status 255 merely aborts processing.<\/li>\n<\/ul><h3><a id=\"user-content-use-varreadf\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-varreadf\"><\/a>use var\/readf<\/h3>\n<p><code>readf<\/code> reads arbitrary data from standard input into a variable until end\nof file, converting it into a format suitable for passing to the\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/printf.html\" rel=\"nofollow\"><code>printf<\/code><\/a>\nutility. For example, <code>readf var &lt;foo; printf \"$var\" &gt;bar<\/code> will copy foo to\nbar. Thus, <code>readf<\/code> allows storing both text and binary files into shell\nvariables in a textual format suitable for manipulation with standard shell\nfacilities.<\/p>\n<p>All non-printable, non-ASCII characters are converted to <code>printf<\/code> octal or\none-letter escape codes, except newlines. Not encoding newline characters\nallows for better processing by line-based utilities such as <code>grep<\/code>, <code>sed<\/code>,\n<code>awk<\/code>, etc. However, if the file ends in a newline, that final newline is\nencoded to <code>\\\\n<\/code> to protect it from being stripped by command substitutions.<\/p>\n<p>Usage: <code>readf<\/code> <em>varname<\/em><\/p>\n<p>Caveats:<\/p>\n<ul><li>Best for small-ish files. The encoded file is stored in memory (a shell\nvariable). For a binary file, encoding in <code>printf<\/code> format typically\nabout doubles the size, though it could be up to four times as large.<\/li>\n<li>If the shell executing your program does not have <code>printf<\/code> as a builtin\ncommand, the external <code>printf<\/code> command will fail if the encoded file\nsize exceeds the maximum length of arguments to external commands\n(<code>getconf ARG_MAX<\/code> will obtain this limit for your system). Shell builtin\ncommands do not have this limit. Check for a <code>printf<\/code> builtin using\n<a href=\"#user-content-feature-testing\"><code>thisshellhas<\/code><\/a> if you need to be sure,\nand always <a href=\"#user-content-hardening-emergency-halt-on-error\"><code>harden<\/code><\/a>\n<code>printf<\/code>!<\/li>\n<\/ul><h3><a id=\"user-content-use-varsetlocal\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-varsetlocal\"><\/a>use var\/setlocal<\/h3>\n<p>Defines a new <code>setlocal<\/code>...<code>endlocal<\/code> shell code block construct with\narbitrary local variables and arbitrary local shell options, as well as\nsafe field splitting and pathname expansion operators.<\/p>\n<p>Usage: <code>setlocal<\/code> [ <em>localitem<\/em> ... ]\n[ [ <code>--split<\/code> | <code>--split=<\/code><em>characters<\/em> ] [ <code>--glob<\/code> | <code>--nglob<\/code> ] <code>--<\/code> [ <em>word<\/em> ... ] ]\n<code>;<\/code> <code>do<\/code> <em>commands<\/em> <code>;<\/code> <code>endlocal<\/code><\/p>\n<p>The <em>commands<\/em> are executed with the specified settings applied locally to\nthe <code>setlocal<\/code>...<code>endlocal<\/code> block.<\/p>\n<p>Each <em>localitem<\/em> can be:<\/p>\n<ul><li>A variable name with or without a <code>=<\/code> immediately followed by a value.\nThis renders that variable local to the block, initially either unsetting\nit or assigning the value, which may be empty.<\/li>\n<li>A shell option letter immediately preceded by a <code>-<\/code> or <code>+<\/code> sign. This\nlocally turns that shell option on or off, respectively. This follows the\ncounterintuitive syntax of <code>set<\/code>. Long-form shell options like <code>-o<\/code>\n<em>optionname<\/em> and <code>+o<\/code> <em>optionname<\/em> are also supported. It depends on the\nshell what options are supported. Specifying a nonexistent option is a\nfatal error. Use <a href=\"#user-content-feature-testing\"><code>thisshellhas<\/code><\/a> to check\nfor a non-POSIX option's existence on the current shell before using it.<\/li>\n<\/ul><p>The <code>return<\/code> command exits the block, causing the global variables and\nsettings to be restored and resuming execution at the point immmediately\nfollowing <code>endlocal<\/code>. This is like a shell function. In fact, internally,\n<code>setlocal<\/code> blocks <strong>are<\/strong> one-time shell functions that use\n<a href=\"#user-content-the-stack\">the stack<\/a>\nto save and restore variables and settings. Like any shell\nfunction, a <code>setlocal<\/code> block exits with the exit status of the last command\nexecuted within it or, with the status passed on by or given as an argument to\n<code>return<\/code>.<\/p>\n<p>The <code>break<\/code> and <code>continue<\/code> commands, when not used within a loop within the\nblock, also exit the block, but always with exit status 0. It's preferred to\nuse <code>return<\/code> instead. Note that <code>setlocal<\/code> creates a new loop context and\nyou cannot use <code>break<\/code> or <code>continue<\/code> to resume or break from enclosing loops\noutside the <code>setlocal<\/code> block. (Shells with\n<a href=\"#user-content-quirks\">QRK_BCDANGER<\/a> do in fact allow this, preventing\n<code>endlocal<\/code> from restoring the global settings! Shells without this quirk\nautomatically protect against this.)<\/p>\n<p>Within the block, the positional parameters (<code>$@<\/code>, <code>$1<\/code>, etc.) are always\nlocal. By default, a copy is inherited from outside the block. Any changes to\nthe positional parameters made within the block will be discarded upon\nexiting it.<\/p>\n<p>However, if a <code>--<\/code> is present, the set of <em>word<\/em>s after <code>--<\/code> becomes the\npositional parameters instead, after being modified by the <code>--split<\/code> or\n<code>--glob<\/code>\/<code>--nglob<\/code> operators if present, as follows:<\/p>\n<ul><li>A simple <code>--split<\/code> subjects the <em>word<\/em>s to default field splitting.<\/li>\n<li><code>--split=<\/code><em>string<\/em> subjects the <em>word<\/em>s to field splitting based on the\ncharacters given in <em>string<\/em>.<\/li>\n<li><code>--glob<\/code> causes each <em>word<\/em> (after split, if specified) to be treated\nas shell glob patterns and subjects each pattern to pathname expansion.\nUnmatched patterns are left untouched (i.e. resolve to themselves).<\/li>\n<li>The <code>--nglob<\/code> operator does the same as <code>--glob<\/code>, then also removes each\nunmatched pattern. Note that this includes patterns that do not contain\nany wildcard characters.<\/li>\n<\/ul><p><strong>Note:<\/strong> These <code>--split<\/code> and <code>--glob<\/code>\/<code>--nglob<\/code> operators do <strong>not<\/strong>\nenable field splitting or pathname expansion within the block itself, but\nonly subject the <em>word<\/em>s after the <code>--<\/code> to them. If field splitting and\nglobbing are <a href=\"#user-content-use-safe\">disabled globally<\/a>, this provides a\nsafe way to perform field splitting or globbing without actually enabling\nthem for any code. To illustrate this advantage, note the difference:<\/p>\n<pre># Field splitting is enabled for all unquoted expansions within the\n# setlocal block, which may be unsafe, so must quote \"$foo\" and \"$bar\".\nsetlocal dir IFS=':'; do\n    for dir in $PATH; do\n    somestuff \"$foo\" \"$bar\"\ndone\nendlocal\n# The value of PATH is split at ':' and assigned to the positional\n# parameters, without enabling field splitting within the setlocal block.\nsetlocal dir --split=':' -- $PATH; do\n    for dir do\n        somestuff $foo $bar\n    done\nendlocal\n<\/pre>\n<p><strong>Important:<\/strong>\nThe <code>--split<\/code> and <code>--glob<\/code>\/<code>--nglob<\/code> operators are designed to be\nused along with <a href=\"#user-content-use-safe\">safe mode<\/a>. If they are used in\ntraditional mode, i.e. with field splitting and\/or pathname expansion\nglobally active, you <em>must<\/em> make sure the <em>word<\/em>s after the <code>--<\/code> are\nproperly quoted as with any other command, otherwise you will have\nunexpected duplicate splitting or pathname expansion.<\/p>\n<p><strong>Other usage notes:<\/strong><\/p>\n<ul><li>Although the <code>setlocal<\/code> declaration ends with <code>; do<\/code> as in a <code>while<\/code> or\n<code>until<\/code> loop, the code block is terminated with <code>endlocal<\/code> and not with\n<code>done<\/code>. Terminating it with <code>done<\/code> results in a misleading shell syntax\nerror (end of file, or missing <code>}<\/code>), a side effect of how <code>setlocal<\/code> is\nimplemented.<\/li>\n<li><code>setlocal<\/code> blocks do <strong>not<\/strong> mix well with\n<a href=\"#user-content-user-content-capabilities\"><code>LOCAL<\/code><\/a>\n(shell-native functionality for local variables), especially not on shells\nwith <code>QRK_LOCALUNS<\/code> or <code>QRK_LOCALUNS2<\/code>. Use one or the other, but not\nboth.<\/li>\n<li>For maximum compatibility with shell bugs (particularly <code>BUG_FNSUBSH<\/code> on\nksh93, and an alias parsing oddity on mksh [up to\nR54 2016\/11\/11] that triggers a spurious syntax\nerror), <code>setlocal<\/code> blocks should not be used within subshells, including\ncommand substitution subshells. There is usually not much point to this\nanyway; the point of <code>setlocal<\/code> is to have certain settings local and keep\nthe rest global, all without the performance hit of forking a subshell\nprocess. (Forking new subshells within a <code>setlocal<\/code> block is fine.)<\/li>\n<li>A note of caution concerning loop constructs: Care should be taken not to\nuse <code>break<\/code> and <code>continue<\/code> in ways that would cause execution to continue\noutside the <code>setlocal<\/code> block. Some shells do not allow <code>break<\/code> and <code>continue<\/code>\nto break out of a shell function (including the internal one-time shell\nfunction employed by setlocal), so thankfully this fails on those shells.\nBut on others this succeeds, so global settings are not restored, wreaking\nhavoc on the rest of your program. One way to avoid the problem is to\nenvelop the entire loop in a <code>setlocal<\/code> block. Another is to exit the\ninternal shell function using <code>return 1<\/code> and then add <code>|| break<\/code> or\n<code>|| continue<\/code> immediately after <code>endlocal<\/code>.<\/li>\n<li>zsh programmers may recognise <code>setlocal<\/code> as pretty much the equivalent of\nzsh's anonymous functions -- functionality that is hereby brought to all\nPOSIX shells, albeit with a rather different syntax.<\/li>\n<\/ul><h3><a id=\"user-content-use-varstring\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-varstring\"><\/a>use var\/string<\/h3>\n<p>String comparison and manipulation functions.<\/p>\n<p><code>sortsbefore<\/code> and <code>sortsafter<\/code>: test if string 1 sorts before or after\nstring 2. The POSIX shell provides no standard built-in way of doing this.\nThese functions use built-in ways where available, or fall back on the\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/expr.html\" rel=\"nofollow\">expr<\/a>\nutility.<br \/>Usage: <code>sortsbefore<\/code>|<code>sortsafter<\/code> <em>string1<\/em> <em>string2<\/em><\/p>\n<p><code>trim<\/code>: strip whitespace from the beginning and end of a variable's value.\nWhitespace is defined by the <code>[:space:]<\/code> character class. In the POSIX\nlocale, this is tab, newline, vertical tab, form feed, carriage return, and\nspace, but in other locales it may be different.\n(On shells with <a href=\"#user-content-bugs\"><code>BUG_NOCHCLASS<\/code><\/a>,\n<a href=\"#user-content-control-character-whitespace-and-shell-safe-character-constants\"><code>$WHITESPACE<\/code><\/a>\nis used to define whitesapce instead.) Optionally, a string of literal\ncharacters can be provided in the second argument. Any characters appearing\nin that string will then be trimmed instead of whitespace.\nUsage: <code>trim<\/code> <em>varname<\/em> [ <em>characters<\/em> ]<\/p>\n<p><code>replacein<\/code>: Replace leading, <code>-t<\/code>railing or <code>-a<\/code>ll occurrences of a string by\nanother string in a variable.<br \/>Usage: <code>replacein<\/code> [ <code>-t<\/code> | <code>-a<\/code> ] <em>varname<\/em> <em>oldstring<\/em> <em>newstring<\/em><\/p>\n<p><code>append<\/code> and <code>prepend<\/code>: Append or prepend zero or more strings to a\nvariable, separated by a string of zero or more characters, avoiding the\nhairy problem of dangling separators.\nUsage: <code>append<\/code>|<code>prepend<\/code> [ <code>--sep=<\/code><em>separator<\/em> ] [ <code>-Q<\/code> ] <em>varname<\/em> [ <em>string<\/em> ... ]<br \/>If the separator is not specified, it defaults to a space character.\nIf the <code>-Q<\/code> option is given, each <em>string<\/em> is\n<a href=\"#user-content-quoting-strings-for-subsequent-parsing-by-the-shell\">shell-quoted<\/a>\nbefore appending or prepending.<\/p>\n<h3><a id=\"user-content-use-sysbase\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysbase\"><\/a>use sys\/base<\/h3>\n<p>Some very common external commands ought to be standardised, but aren't. For\ninstance, the <code>which<\/code> and <code>readlink<\/code> commands have incompatible options on\nvarious GNU and BSD variants and may be absent on other Unix-like systems.\nThis module provides a complete re-implementation of such basic utilities\nwritten as modernish shell functions. Scripts that use the modernish version\nof these utilities can expect to be fully cross-platform. They also have\nvarious enhancements over the GNU and BSD originals.<\/p>\n<h4><a id=\"user-content-use-sysbasereadlink\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysbasereadlink\"><\/a>use sys\/base\/readlink<\/h4>\n<p><code>readlink<\/code>: Read the target of a symbolic link. Robustly handles weird\nfilenames such as those containing newline characters. Stores result in the\n$REPLY variable and optionally writes it on standard output. Optionally\ncanonicalises each path, following all symlinks encountered (for this mode,\nall but the last component must exist). Optionally shell-quote each item of\noutput for later parsing by the shell, separating multiple items with spaces\ninstead of newlines.<\/p>\n<h4><a id=\"user-content-use-sysbasewhich\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysbasewhich\"><\/a>use sys\/base\/which<\/h4>\n<p><code>which<\/code>: Outputs, and\/or stores in the <code>REPLY<\/code> variable, either the first\navailable directory path to each given command, or all available paths,\naccording to the current <code>$PATH<\/code> or the system default path. Exits\nsuccessfully if at least one path was found for each command, or\nunsuccessfully if none were found for any given command.<\/p>\n<p>Usage: <code>which<\/code> [ <code>-[apqsnQ1]<\/code> ] [ <code>-P<\/code> <em>number<\/em> ] <em>program<\/em> [ <em>program<\/em> ... ]<\/p>\n<ul><li><code>-a<\/code>: List <em>a<\/em>ll executables found, not just the first one for each argument.<\/li>\n<li><code>-p<\/code>: Search the system default <em>p<\/em>ath, not the current <code>$PATH<\/code>. This is the\nminimal path, specified by POSIX, that is guaranteed to find all the standard\nutilities.<\/li>\n<li><code>-q<\/code>: Be <em>q<\/em>uiet: suppress all warnings.<\/li>\n<li><code>-s<\/code>: <em>S<\/em>ilent operation: don't write output, only store it in the <code>REPLY<\/code>\nvariable. Suppress warnings except, if you run <code>which -s<\/code> in a subshell,\nthe warning that the <code>REPLY<\/code> variable will not survive the subshell.<\/li>\n<li><code>-n<\/code>: When writing to standard output, do <em>n<\/em>ot write a final <em>n<\/em>ewline.<\/li>\n<li><code>-Q<\/code>: Shell-<em>q<\/em>uote each unit of output. Separate by spaces instead\nof newlines. This generates a list of arguments in shell syntax,\nguaranteed to be suitable for safe parsing by the shell, even if the\nresulting pathnames should contain strange characters such as spaces or\nnewlines and other control characters.<\/li>\n<li><code>-1<\/code> (one): Output the results for at most <em>one<\/em> of the arguments in\ndescending order of preference: once a search succeeds, ignore\nthe rest. Suppress warnings except a subshell warning for <code>-s<\/code>.\nThis is useful for finding a command that can exist under\nseveral names, for example:\n<code>which -f -1 gnutar gtar tar<\/code><br \/>This option modifies which's exit status behaviour: <code>which -1<\/code>\nreturns successfully if any match was found.<\/li>\n<li><code>-f<\/code>: Consider it a <a href=\"#user-content-reliable-emergency-halt\"><em>f<\/em>atal error<\/a>\nif at least one of the given <em>program<\/em>s is not found. But if option <code>-1<\/code>\nis also given, only throw a fatal error if none are found.<\/li>\n<li><code>-P<\/code>: Strip the indicated number of <em>p<\/em>athname elements from the output,\nstarting from the right.\n<code>-P1<\/code>: strip <code>\/program<\/code>;\n<code>-P2<\/code>: strip <code>\/*\/program<\/code>,\netc. This is useful for determining the installation root directory for\nan installed package.<\/li>\n<\/ul><h4><a id=\"user-content-use-sysbasemktemp\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysbasemktemp\"><\/a>use sys\/base\/mktemp<\/h4>\n<p>A cross-platform shell implementation of 'mktemp' that aims to be just as\nsafe as native <code>mktemp<\/code>(1) implementations, while avoiding the problem of\nhaving various mutually incompatible versions and adding several unique\nfeatures of its own.<\/p>\n<p>Creates one or more unique temporary files, directories or named pipes,\natomically (i.e. avoiding race conditions) and with safe permissions.\nThe path name(s) are stored in $REPLY and optionally written to stdout.<\/p>\n<p>Usage: <code>mktemp<\/code> [ <code>-dFsQCt<\/code> ] [ <em>template<\/em> ... ]<\/p>\n<ul><li><code>-d<\/code>: Create a directory instead of a regular file.<\/li>\n<li><code>-F<\/code>: Create a FIFO (named pipe) instead of a regular file.<\/li>\n<li><code>-s<\/code>: Silent. Store output in <code>$REPLY<\/code>, don't write any output or message.<\/li>\n<li><code>-Q<\/code>: Shell-quote each unit of output. Separate by spaces, not newlines.<\/li>\n<li><code>-C<\/code>: Automated cleanup.\n<a href=\"#user-content-the-trap-stack\">Pushes a trap<\/a>\nto remove the files\non exit. On an interactive shell, that's all this option does. On a\nnon-interactive shell, the following applies: Clean up on receiving\nSIGPIPE and SIGTERM as well. On receiving SIGINT, clean up if the\noption was given at least twice, otherwise notify the user of files\nleft. On the invocation of\n<a href=\"#user-content-reliable-emergency-halt\"><code>die<\/code><\/a>,\nclean up if the option was given at least three times, otherwise notify\nthe user of files left.<\/li>\n<li><code>-t<\/code>: Prefix one temporary files directory to all the <em>template<\/em>s:\n<code>$TMPDIR\/<\/code> if <code>TMPDIR<\/code> is set, <code>\/tmp\/<\/code> otherwise. The <em>template<\/em>s\nmay not contain any slashes. If the template has neither any trailing\n<code>X<\/code>es nor a trailing dot, a dot is added before the random suffix.<\/li>\n<\/ul><p>The template defaults to <code>\/tmp\/temp.<\/code>. An suffix of random shell-safe ASCII\ncharacters is added to the template to create the file. For compatibility with\nother <code>mktemp<\/code> implementations, any optional trailing <code>X<\/code> characters in the\ntemplate are removed. The length of the suffix will be equal to the amount of\n<code>X<\/code>es removed, or 10, whichever is more. The longer the random suffix, the\nhigher the security of using <code>mktemp<\/code> in a shared directory such as <code>tmp<\/code>.<\/p>\n<p>Since <code>\/tmp<\/code> is a world-writable directory shared by other users, for best\nsecurity it is recommended to create a private subdirectory using <code>mktemp -d<\/code>\nand work within that.<\/p>\n<p>Option <code>-C<\/code> cannot be used while invoking <code>mktemp<\/code> in a subshell, such as\nin a command substitution. Modernish will detect this and treat it as a\nfatal error. The reason is that a typical command substitution like\n<code>tmpfile=$(mktemp -C)<\/code>\nis incompatible with auto-cleanup, as the cleanup EXIT trap would be\ntriggered not upon exiting the program but upon exiting the command\nsubstitution subshell that just ran <code>mktemp<\/code>, thereby immediately undoing\nthe creation of the file. Instead, do something like:\n<code>mktemp -sC; tmpfile=$REPLY<\/code><\/p>\n<h4><a id=\"user-content-use-sysbaseseq\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysbaseseq\"><\/a>use sys\/base\/seq<\/h4>\n<p>A cross-platform implementation of <code>seq<\/code> that is more powerful and versatile\nthan native GNU and BSD <code>seq<\/code>(1) implementations. The core is written in\n<code>bc<\/code>, the POSIX arbitrary-presision calculator language. That means this\n<code>seq<\/code> inherits the capacity to handle numbers with a precision and size only\nlimited by computer memory, as well as the ability to handle input numbers\nin any base from 1 to 16 and produce output in any base 1 and up.<\/p>\n<p>Usage: <code>seq<\/code> [ <code>-w<\/code> ] [ <code>-f<\/code> <em>format<\/em> ] [ <code>-s<\/code> <em>string<\/em> ] [ <code>-S<\/code> <em>scale<\/em> ]\n[ <code>-B<\/code> <em>base<\/em> ] [ <code>-b<\/code> <em>base<\/em> ] [ <em>first<\/em> [ <em>incr<\/em> ] ] <em>last<\/em><\/p>\n<p><code>seq<\/code> prints a sequence of arbitrary-precision floating point numbers, one\nper line, from <em>first<\/em> (default 1), to as near <em>last<\/em> as possible, in increments of\n<em>incr<\/em> (default 1). If <em>first<\/em> is larger than <em>last<\/em>, the default <em>incr<\/em> is -1.<\/p>\n<ul><li><code>-w<\/code>: Equalise width by padding with leading zeros. The longest of the\n<em>first<\/em>, <em>incr<\/em> or <em>last<\/em> arguments is taken as the length that each\noutput number should be padded to.<\/li>\n<li><code>-f<\/code>: <code>printf<\/code>-style floating-point format. The format string is passed on\n(with an added <code>\\n<\/code>) to <code>awk<\/code>'s builtin <code>printf<\/code> function. Because\nof that, the <code>-f<\/code> option can only be used if the output base is 10.\nNote that <code>awk<\/code>'s floating point precision is limited, so very\nlarge or long numbers will be rounded.<\/li>\n<li><code>-s<\/code>: Use <em>string<\/em> to separate numbers. Default: newline. The terminator\ncharacter remains a newline in any case (which is like GNU <code>seq<\/code>\nand differs from BSD <code>seq<\/code>).<\/li>\n<li><code>-S<\/code>: Explicitly set the scale (number of digits after decimal point).\nDefaults to the largest number of digits after the decimal point\namong the <em>first<\/em>, <em>incr<\/em> or <em>last<\/em> arguments.<\/li>\n<li><code>-B<\/code>: Set input and output base from 1 to 16. Defaults to 10.<\/li>\n<li><code>-b<\/code>: Set arbitrary output base from 1. Defaults to input base.\nSee the <code>bc<\/code>(1) manual for more infromation on the output format\nfor bases greater than 16.<\/li>\n<\/ul><p>The <code>-S<\/code>, <code>-B<\/code> and <code>-b<\/code> options take shell integer numbers as operands. This\nmeans a leading <code>0X<\/code> or <code>0x<\/code> denotes a hexadecimal number and (except on\nshells with BUG_NOOCTAL) a leading <code>0<\/code> denotes an octal numnber.<\/p>\n<p>For portability reasons, modernish <code>seq<\/code> always uses a dot (.) for the\nfloating point, never a comma, regardless of the system locale. This applies\nboth to command arguments and to output.<\/p>\n<p>The <code>-w<\/code>, <code>-f<\/code> and <code>-s<\/code> options are inspired by GNU and BSD <code>seq<\/code>, mostly\nemulating GNU where they differ. The <code>-S<\/code>, <code>-B<\/code> and <code>-b<\/code> options are\nmodernish enhancements based on <code>bc<\/code>(1) functionality.<\/p>\n<h4><a id=\"user-content-use-sysbaserev\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysbaserev\"><\/a>use sys\/base\/rev<\/h4>\n<p><code>rev<\/code> copies the specified files to the standard output, reversing the order\nof characters in every line. If no files are specified, the standard input\nis read.<\/p>\n<p>Usage: like <code>rev<\/code> on Linux and BSD, which is like <code>cat<\/code> except that <code>-<\/code> is\na filename and does not denote standard input. No options are supported.<\/p>\n<h3><a id=\"user-content-use-sysdir\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysdir\"><\/a>use sys\/dir<\/h3>\n<p>Functions for working with directories. So far I have:<\/p>\n<h4><a id=\"user-content-use-sysdirtraverse\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysdirtraverse\"><\/a>use sys\/dir\/traverse<\/h4>\n<p><code>traverse<\/code> is a fully cross-platform, robust replacement for <code>find<\/code> without\nthe snags of the latter. It is not line oriented but handles all data\ninternally in the shell. Any weird characters in file names (including\nwhitespace and even newlines) \"just work\", provided either\n<a href=\"#user-content-use-safe\"><code>use safe<\/code><\/a> is active or shell expansions are\nproperly quoted. This avoids many hairy\n<a href=\"https:\/\/www.dwheeler.com\/essays\/filenames-in-shell.html\" rel=\"nofollow\">common pitfalls with <code>find<\/code><\/a>\nwhile remaining compatible with all POSIX systems.<\/p>\n<p><code>traverse<\/code> recursively walks through a directory, executing a command for\neach file and subdirectory found. That command is usually a handler shell\nfunction in your program.<\/p>\n<p>Unlike <code>find<\/code>, which is so smart its command line options are practically\ntheir own programming language, <code>traverse<\/code> is dumb: it has minimal\nfunctionality of its own. However, with a shell function as the command,\nany functionality of 'find' and anything else can be programmed in the\nshell language. Flexibility is unlimited. The <code>install.sh<\/code> script that comes\nwith modernish provides a good example of its practical use. See also the\n<a href=\"https:\/\/github.com\/modernish\/modernish\/blob\/master\/share\/doc\/modernish\/examples\/traverse-test\">traverse-test<\/a>\nexample program.<\/p>\n<p>Usage: <code>traverse<\/code> [ <code>-d<\/code> ] [ <code>-F<\/code>] [ <code>-X<\/code> ] <em>directory<\/em> <em>command<\/em><\/p>\n<p><code>traverse<\/code> calls <em>command<\/em>, once for each file found within the <em>directory<\/em>,\nwith one parameter containing the full pathname relative to the <em>directory<\/em>.\nAny directories found within are automatically entered and traversed\nrecursively unless the <em>command<\/em> exits with status 1. Symlinks to\ndirectories are not followed.<\/p>\n<p><code>find<\/code>'s <code>-prune<\/code> functionality is implemented by testing the command's exit\nstatus. If the command indicated exits with status 1 for a directory, this\nmeans: do not traverse the directory in question. For other types of files,\nexit status 1 is the same as 0 (success). Exit status 2 means: stop the\nexecution of <code>traverse<\/code> and resume program execution. An exit status greater\nthan 2 indicates system failure and causes the program to abort.<\/p>\n<p><code>find<\/code>'s <code>-depth<\/code> functionality is implemented using the <code>-d<\/code> option. By\ndefault, <code>traverse<\/code> handles directories first, before their contents. The\n<code>-d<\/code> option causes depth-first traversal, so all entries in a directory will\nbe acted on before the directory itself. This applies recursively to\nsubdirectories. That means depth-first traversal is incompatible with\npruning, so returning status 1 for directories will have no effect.<\/p>\n<p>find's <code>-xdev<\/code> functionality is implemented using the <code>-F<\/code> option. If this\nis given, <code>traverse<\/code> will not descend into directories that are on another\nfile system than that of the directory given in the argument.<\/p>\n<p><code>xargs<\/code>-like functionality is implemented using the <code>-X<\/code> option. As many\nitems as possible are saved up before being passed to the command all at\nonce. This is also incompatible with pruning. Unlike <code>xargs<\/code>, the command is\nonly executed if at least one item was found for it to handle.<\/p>\n<h4><a id=\"user-content-use-sysdircountfiles\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-sysdircountfiles\"><\/a>use sys\/dir\/countfiles<\/h4>\n<p><code>countfiles<\/code>: Count the files in a directory using nothing but shell\nfunctionality, so without external commands. (It's amazing how many pitfalls\nthis has, so a library function is needed to do it robustly.)<\/p>\n<p>Usage: <code>countfiles<\/code> [ <code>-s<\/code> ] <em>directory<\/em> [ <em>globpattern<\/em> ... ]<\/p>\n<p>Count the number of files in a directory, storing the number in <code>REPLY<\/code>\nand (unless <code>-s<\/code> is given) printing it to standard output.\nIf any <em>globpattern<\/em>s are given, only count the files matching them.<\/p>\n<h3><a id=\"user-content-use-systerm\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-systerm\"><\/a>use sys\/term<\/h3>\n<p>Utilities for working with the terminal.<\/p>\n<h4><a id=\"user-content-use-systermreadkey\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-systermreadkey\"><\/a>use sys\/term\/readkey<\/h4>\n<p><code>readkey<\/code>: read a single character from the keyboard without echoing back to\nthe terminal. Buffering is done so that multiple waiting characters are read\none at a time.<\/p>\n<p>Usage: <code>readkey<\/code> [ <code>-E<\/code> <em>ERE<\/em> ] [ <code>-t<\/code> <em>timeout<\/em> ] [ <code>-r<\/code> ] [ <em>varname<\/em> ]<\/p>\n<p><code>-E<\/code>: Only accept characters that match the extended regular expression\n<em>ERE<\/em> (the type of RE used by <code>grep -E<\/code>\/<code>egrep<\/code>). <code>readkey<\/code> will silently\nignore input not matching the ERE and wait for input matching it.<\/p>\n<p><code>-t<\/code>: Specify a <em>timeout<\/em> in seconds (one significant digit after the\ndecimal point). After the timeout expires, no character is read and\n<code>readkey<\/code> returns status 1.<\/p>\n<p><code>-r<\/code>: Raw mode. Disables INTR (Ctrl+C), QUIT, and SUSP (Ctrl+Z) processing\nas well as translation of carriage return (13) to linefeed (10).<\/p>\n<p>The character read is stored into the variable referenced by <em>varname<\/em>,\nwhich defaults to <code>REPLY<\/code> if not specified.<\/p>\n<h3><a id=\"user-content-use-optsparsergen\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-optsparsergen\"><\/a>use opts\/parsergen<\/h3>\n<p>Parsing of command line options for shell functions is a hairy problem.\nUsing <code>getopts<\/code> in shell functions is problematic at best, and manually\nwritten parsers are very hard to do right. That's why this module provides\n<code>generateoptionparser<\/code>, a command to generate an option parser: it takes\noptions specifying what variable names to use and what your function should\nsupport, and outputs code to parse options for your shell function. Options\ncan be specified to require or not take arguments. Combining\/stacking\noptions and arguments in the traditional UNIX manner is supported.<\/p>\n<p>Only short (one-character) options are supported. Each option gets a\ncorresponding variable with a name with a specified prefix, ending in the\noption character (hence, only option characters that are valid in variables\nare supported, namely, the ASCII characters A-Z, a-z, 0-9 and the\nunderscore). If the option was not specified on the command line, the\nvariable is set, otherwise it is set to the empty value, or, if the option\nrequires an argument, the variable will contain that argument.<\/p>\n<h3><a id=\"user-content-use-loopcfor\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-loopcfor\"><\/a>use loop\/cfor<\/h3>\n<p>A C-style for loop akin to <code>for (( ))<\/code> in bash\/ksh\/zsh, but unfortunately\nnot with the same syntax. For example, to count from 1 to 10:<\/p>\n<pre>cfor 'i=1' 'i&lt;=10' 'i+=1'; do\n    echo \"$i\"\ndone\n<\/pre>\n<p>(Note that <code>++i<\/code> and <code>i++<\/code> can only be used on shells with ARITHPP,\nbut <code>i+=1<\/code> or <code>i=i+1<\/code> can be used on all POSIX-compliant shells.)<\/p>\n<h3><a id=\"user-content-use-loopsfor\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-loopsfor\"><\/a>use loop\/sfor<\/h3>\n<p>A C-style for loop with arbitrary shell commands instead of arithmetic\nexpressions. For example, to count from 1 to 10 with traditional shell\ncommands:<\/p>\n<pre>sfor 'i=1' '[ \"$i\" -le 10 ]' 'i=$((i+1))'; do\n    print \"$i\"\ndone\n<\/pre>\n<p>or, with modernish commands:<\/p>\n<pre>sfor 'i=1' 'le i 10' 'inc i'; do\n    print \"$i\"\ndone\n<\/pre>\n<h3><a id=\"user-content-use-loopwith\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-loopwith\"><\/a>use loop\/with<\/h3>\n<p>The shell lacks a very simple and basic loop construct, so this module\nprovides for an old-fashioned MS BASIC-style <code>for<\/code> loop, renamed a <code>with<\/code>\nloop because we can't overload the reserved shell keyword <code>for<\/code>. Integer\narithmetic only. Usage:<\/p>\n<pre>with &lt;varname&gt;=&lt;value&gt; to &lt;limit&gt; [ step &lt;increment&gt; ]; do\n   # some commands\ndone\n<\/pre>\n<p>To count from 1 to 10:<\/p>\n<pre>with i=1 to 10; do\n    print \"$i\"\ndone\n<\/pre>\n<p>The value for <code>step<\/code> defaults to 1 if <em>limit<\/em> is equal to or greater\nthan <em>value<\/em>, and to -1 if <em>limit<\/em> is less than <em>value<\/em>. The latter is\na slight enhancement over the original BASIC <code>for<\/code> construct. So\ncounting backwards is as simple as <code>with i=10 to 1; do<\/code> (etc).<\/p>\n<h3><a id=\"user-content-use-loopselect\" class=\"anchor\" aria-hidden=\"true\" href=\"#use-loopselect\"><\/a>use loop\/select<\/h3>\n<p>A complete and nearly accurate reimplementation of the <code>select<\/code> loop from\nksh, zsh and bash for POSIX shells lacking it. Modernish scripts running\non any POSIX shell can now easily use interactive menus.<\/p>\n<p>(All the new loop constructs have one bug in common: as they start with\nan alias that expands to two commands, you can't pipe a command's output\ndirectly into such a loop. You have to enclose it in <code>{<\/code>...<code>}<\/code> as a\nworkaround. I have not found a way around this limitation that doesn't\ninvolve giving up the familiar <code>do<\/code>...<code>done<\/code> syntax.)<\/p>\n<hr \/><h2><a id=\"user-content-appendix-a\" class=\"anchor\" aria-hidden=\"true\" href=\"#appendix-a\"><\/a>Appendix A<\/h2>\n<p>This is a list of shell capabilities and bugs that modernish tests for, so\nthat both modernish itself and scripts can easily query the results of these\ntests. The all-caps IDs below are all usable with the\n<a href=\"#user-content-feature-testing\"><code>thisshellhas<\/code><\/a>\nfunction. This makes it easy for a cross-platform modernish script to write\noptimisations taking advantage of certain non-standard shell features,\nfalling back to a standard method on shells without these features. On the\nother hand, if universal compatibility is not a concern for your script, it\nis just as easy to require certain features and exit with an error message\nif they are not present, or to refuse shells with certain known bugs.<\/p>\n<p>Most feature\/quirk\/bug tests have their own little test script in the\n<code>libexec\/modernish\/cap<\/code> directory. These tests are executed on demand, the\nfirst time the capability or bug in question is queried using\n<code>thisshellhas<\/code>. See <code>README.md<\/code> in that directory for further information.\nThe test scripts also document themselves in the comments.<\/p>\n<p>Tests that are essential to the functioning of modernish are incorporated\ninto the main <code>bin\/modernish<\/code> script; these are considered built-in tests,\nand are immediately run and cached at initialisation time.\nIn the lists below, an ID in <em><code>ITALICS<\/code><\/em> denotes such a built-in test.<\/p>\n<h3><a id=\"user-content-capabilities\" class=\"anchor\" aria-hidden=\"true\" href=\"#capabilities\"><\/a>Capabilities<\/h3>\n<p>Non-standard shell capabilities currently tested for are:<\/p>\n<ul><li><code>LEPIPEMAIN<\/code>: execute last element of a pipe in the main shell, so that\nthings like <em>somecommand<\/em> <code>| read<\/code> <em>somevariable<\/em> work. (zsh, AT&amp;T ksh,\nbash 4.2+)<\/li>\n<li><em><code>RANDOM<\/code><\/em>: the <code>$RANDOM<\/code> pseudorandom generator.\nModernish seeds it if detected. The variable is then set it to read-only\nwhether the generator is detected or not, in order to block it from losing\nits special properties by being unset or overwritten, and to stop it being\nused if there is no generator. This is because some of modernish depends\non <code>RANDOM<\/code> either working properly or being unset.<br \/>(The use case for non-readonly <code>RANDOM<\/code> is setting a known seed to get\nreproducible pseudorandom sequences. To get that in a modernish script,\nuse <code>awk<\/code>'s <code>srand(yourseed)<\/code> and <code>int(rand()*32768)<\/code>.)<\/li>\n<li><code>LINENO<\/code>: the <code>$LINENO<\/code> variable contains the current shell script line\nnumber.<\/li>\n<li><em><code>LOCAL<\/code><\/em>: function-local variables, either using the <code>local<\/code> keyword, or\nby aliasing <code>local<\/code> to <code>typeset<\/code> (mksh, yash).<\/li>\n<li><em><code>KSH88FUNC<\/code><\/em>: define ksh88-style shell functions with the 'function' keyword,\nsupporting dynamically scoped local variables with the 'typeset' builtin.\n(mksh, bash, zsh, yash, et al)<\/li>\n<li><em><code>KSH93FUNC<\/code><\/em>: the same, but with static scoping for local variables. (ksh93 only)\nSee Q28 at the <a href=\"http:\/\/kornshell.com\/doc\/faq.html\" rel=\"nofollow\">ksh93 FAQ<\/a> for an explanation\nof the difference.<\/li>\n<li><code>ARITHPP<\/code>: support for the <code>++<\/code> and <code>--<\/code> unary operators in shell arithmetic.<\/li>\n<li><code>ARITHCMD<\/code>: standalone arithmetic evaluation using a command like\n<code>((<\/code><em>expression<\/em><code>))<\/code>.<\/li>\n<li><code>ARITHFOR<\/code>: ksh93\/C-style arithmetic 'for' loops of the form\n<code>for ((<\/code><em>exp1<\/em><code>; <\/code><em>exp2<\/em><code>; <\/code><em>exp3<\/em><code>)) do <\/code><em>commands<\/em><code>; done<\/code>.<\/li>\n<li><code>CESCQUOT<\/code>: Quoting with C-style escapes, like <code>$'\\n'<\/code> for newline.<\/li>\n<li><code>ADDASSIGN<\/code>: Add a string to a variable using additive assignment,\ne.g. <em>VAR<\/em><code>+=<\/code><em>string<\/em><\/li>\n<li><code>PSREPLACE<\/code>: Search and replace strings in variables using special parameter\nsubstitutions with a syntax vaguely resembling sed.<\/li>\n<li><code>ROFUNC<\/code>: Set functions to read-only with <code>readonly -f<\/code>. (bash, yash)<\/li>\n<li><code>DOTARG<\/code>: Dot scripts support arguments.<\/li>\n<li><code>HERESTR<\/code>: Here-strings, an abbreviated kind of here-document.<\/li>\n<li><code>TESTO<\/code>: The <code>test<\/code>\/<code>[<\/code> builtin supports the <code>-o<\/code> unary operator to check if\na shell option is set.<\/li>\n<li><code>PRINTFV<\/code>: The shell's <code>printf<\/code> builtin has the <code>-v<\/code> option to print to a variable,\nwhich avoids forking a command substitution subshell.<\/li>\n<li><code>ANONFUNC<\/code>: zsh anonymous functions (basically the native zsh equivalent\nof modernish's var\/setlocal module)<\/li>\n<li><code>KSHARRAY<\/code>: ksh88-style arrays. Supported on bash, zsh (under <code>emulate sh<\/code>),\nmksh, pdksh and ksh93.<\/li>\n<li><code>KSHARASGN<\/code>: ksh93-style mass array assignment in the style of\n<code>array=(one two three)<\/code>. Supported on the same shells as KSHARRAY except pdksh.<\/li>\n<li><code>TRAPZERR<\/code>: This feature ID is detected if the <code>ERR<\/code> trap is an alias for\nthe <code>ZERR<\/code> trap. According to the zsh manual, this is the case for zsh on\nmost systems, i.e. those that don't have a <code>SIGERR<\/code> signal. (The\n<a href=\"#user-content-the-trap-stack\">trap stack<\/a>\nuses this feature test.)<\/li>\n<li><code>TRAPPRSUBSH<\/code>: The ability to obtain a list of the current shell's native\ntraps from a command substitution subshell, for example: <code>var=$(trap)<\/code>.\nNote that modernish transparently reimplements this feature on shells\nwithout this native capability, so this feature ID is only relevant if you\nare bypassing modernish to access the <code>trap<\/code> builtin directly. Also, in\norder to be useful to modernish, this feature test only yields a positive\nif the <code>trap<\/code> command in <code>var=$(trap)<\/code> can be replaced by a shell function\nthat in turn calls the builtin <code>trap<\/code> command.<\/li>\n<li><code>OPTNOPREFIX<\/code>: Long-form shell option names use a dynamic <code>no<\/code> prefix for\nall options (including POSIX ones). For instance, <code>glob<\/code> is the opposite\nof <code>noglob<\/code>, and <code>nonotify<\/code> is the opposite of <code>notify<\/code>. (ksh93, yash, zsh)<\/li>\n<\/ul><h3><a id=\"user-content-quirks\" class=\"anchor\" aria-hidden=\"true\" href=\"#quirks\"><\/a>Quirks<\/h3>\n<p>Shell quirks currently tested for are:<\/p>\n<ul><li><em><code>QRK_IFSFINAL<\/code><\/em>: in field splitting, a final non-whitespace IFS delimiter\ncharacter is counted as an empty field (yash &lt; 2.42, zsh, pdksh). This is a QRK\n(quirk), not a BUG, because POSIX is ambiguous on this.<\/li>\n<li><code>QRK_32BIT<\/code>: mksh: the shell only has 32-bit arithmetics. Since every modern\nsystem these days supports 64-bit long integers even on 32-bit kernels, we\ncan now count this as a quirk.<\/li>\n<li><code>QRK_APIPEMAIN<\/code>: On zsh &lt; 5.3, any element of a pipeline (not just the\nlast element) that is nothing but a simple variable assignment is executed\nin the current shell environment, instead of a subshell. For instance, the\nassignment <code>var=foo<\/code> survives <code>SomeCommands | var=foo | SomeMoreCommands<\/code>.<\/li>\n<li><code>QRK_ARITHEMPT<\/code>: In yash, with POSIX mode turned off, a set but empty\nvariable yields an empty string when used in an arithmetic expression,\ninstead of 0. For example, <code>foo=''; echo $((foo))<\/code> outputs an empty line.<\/li>\n<li><code>QRK_ARITHWHSP<\/code>: In <a href=\"https:\/\/osdn.jp\/ticket\/browse.php?group_id=3863&amp;tid=36002\" rel=\"nofollow\">yash<\/a>\nand FreeBSD \/bin\/sh, trailing whitespace from variables is not trimmed in arithmetic\nexpansion, causing the shell to exit with an 'invalid number' error. POSIX is silent\non the issue. The modernish <code>isint<\/code> function (to determine if a string is a valid\ninteger number in shell syntax) is <code>QRK_ARITHWHSP<\/code> compatible, tolerating only\nleading whitespace.<\/li>\n<li><code>QRK_BCDANGER<\/code>: <code>break<\/code> and <code>continue<\/code> can affect non-enclosing loops,\neven across shell function barriers (zsh, Busybox ash; older versions\nof bash, dash and yash). (This is especially dangerous when using\n<a href=\"#user-content-use-varsetlocal\">var\/setlocal<\/a>\nwhich internally uses a temporary shell function to try to protect against\nbreaking out of the block without restoring global parameters and settings.)<\/li>\n<li><code>QRK_EMPTPPFLD<\/code>: Unquoted <code>$@<\/code> and <code>$*<\/code> do not discard empty fields.\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_05_02\" rel=\"nofollow\">POSIX says<\/a>\nfor both unquoted <code>$@<\/code> and unquoted <code>$*<\/code> that empty positional parameters\n<em>may<\/em> be discarded from the expansion. AFAIK, just one shell (yash)\ndoesn't.<\/li>\n<li><code>QRK_EMPTPPWRD<\/code>: <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_05_02\" rel=\"nofollow\">POSIX says<\/a>\nthat empty <code>\"$@\"<\/code> generates zero fields but empty <code>''<\/code> or <code>\"\"<\/code> or\n<code>\"$emptyvariable\"<\/code> generates one empty field. But it leaves unspecified\nwhether something like <code>\"$@$emptyvariable\"<\/code> generates zero fields or one\nfield. Zsh, pdksh\/mksh and (d)ash generate one field, as seems logical.\nBut bash, AT&amp;T ksh and yash generate zero fields, which we consider a\nquirk. (See also BUG_PP_01)<\/li>\n<li><code>QRK_EVALNOOPT<\/code>: <code>eval<\/code> does not parse options, not even <code>--<\/code>, which makes it\nincompatible with other shells: on the one hand, (d)ash does not accept<br \/><code>eval -- \"$command\"<\/code> whereas on other shells this is necessary if the command\nstarts with a <code>-<\/code>, or the command would be interpreted as an option to <code>eval<\/code>.\nA simple workaround is to prefix arbitrary commands with a space.\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_19_16\" rel=\"nofollow\">Both situations are POSIX compliant<\/a>,\nbut since they are incompatible without a workaround,the minority situation\nis labeled here as a QuiRK.<\/li>\n<li><code>QRK_EXECFNBI<\/code>: In pdksh and zsh, <code>exec<\/code> looks up shell functions and\nbuiltins before external commands, and if it finds one it does the\nequivalent of running the function or builtin followed by <code>exit<\/code>. This\nis probably a bug in POSIX terms; <code>exec<\/code> is supposed to launch a\nprogram that overlays the current shell, implying the program launched by\n<code>exec<\/code> is always external to the shell. However, since the\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_20\" rel=\"nofollow\">POSIX language<\/a>\nis rather\n<a href=\"https:\/\/www.mail-archive.com\/austin-group-l@opengroup.org\/msg01437.html\" rel=\"nofollow\">vague and possibly incorrect<\/a>,\nthis is labeled as a shell quirk instead of a shell bug.<\/li>\n<li><code>QRK_HDPARQUOT<\/code>: Double <strong>quot<\/strong>es within certain <strong>par<\/strong>ameter substitutions in\n<strong>h<\/strong>ere-<strong>d<\/strong>ocuments aren't removed (FreeBSD sh; bosh). For instance, if\n<code>var<\/code> is set, <code>${var+\"x\"}<\/code> in a here-document yields <code>\"x\"<\/code>, not <code>x<\/code>.\n<a href=\"https:\/\/www.mail-archive.com\/austin-group-l@opengroup.org\/msg01626.html\" rel=\"nofollow\">POSIX considers it undefined<\/a>\nto use double quotes there, so they should be avoided for a script to be\nfully POSIX compatible.\n(Note this quirk does <strong>not<\/strong> apply for substitutions that remove pattens,\nsuch as <code>${var#\"$x\"}<\/code> and <code>${var%\"$x\"}<\/code>; those are defined by POSIX\nand double quotes are fine to use.)\n(Note 2: single quotes produce widely varying behaviour and should never\nbe used within any form of parameter substitution in a here-document.)<\/li>\n<li><code>QRK_LOCALINH<\/code>: On a shell with LOCAL, local variables, when declared\nwithout assigning a value, inherit the state of their global namesake, if\nany. (dash, FreeBSD sh)<\/li>\n<li><code>QRK_LOCALSET<\/code>: On a shell with LOCAL, local variables are immediately set\nto the empty value upon being declared, instead of being initially without\na value. (zsh)<\/li>\n<li><code>QRK_LOCALSET2<\/code>: Like <code>QRK_LOCALSET<\/code>, but <em>only<\/em> if the variable by the\nsame name in the global\/parent scope is unset. If the global variable is\nset, then the local variable starts out unset. (bash 2 and 3)<\/li>\n<li><code>QRK_LOCALUNS<\/code>: On a shell with LOCAL, local variables lose their local\nstatus when unset. Since the variable name reverts to global, this means that\n<em><code>unset<\/code> will not necessarily unset the variable!<\/em> (yash, pdksh\/mksh. Note:\nthis is actually a behaviour of <code>typeset<\/code>, to which modernish aliases <code>local<\/code>\non these shells.)<\/li>\n<li><code>QRK_LOCALUNS2<\/code>: This is a more treacherous version of <code>QRK_LOCALUNS<\/code> that\nis unique to bash. The <code>unset<\/code> command works as expected when used on a local\nvariable in the same scope that variable was declared in, <strong>however<\/strong>, it\nmakes local variables global again if they are unset in a subscope of that\nlocal scope, such as a function called by the function where it is local.\n(Note: since <code>QRK_LOCALUNS2<\/code> is a special case of <code>QRK_LOCALUNS<\/code>, modernish\nwill not detect both.)<\/li>\n<li><code>QRK_OPTCASE<\/code>: Long-form shell option names are case-insensitive. (yash, zsh)<\/li>\n<li><code>QRK_OPTDASH<\/code>: Long-form shell option names ignore the <code>-<\/code>. (ksh93, yash)<\/li>\n<li><code>QRK_OPTULINE<\/code>: Long-form shell option names ignore the <code>_<\/code>. (yash, zsh)<\/li>\n<li><code>QRK_PPIPEMAIN<\/code>: On zsh, in all elements of a pipeline, parameter\nexpansions are evaluated in the current environment (with any changes they\nmake surviving the pipeline), though the commands themselves of every\nelement but the last are executed in a subshell. For instance, given unset\nor empty <code>v<\/code>, in the pipeline <code>cmd1 ${v:=foo} | cmd2<\/code>, the assignment to\n<code>v<\/code> survives, though <code>cmd1<\/code> itself is executed in a subshell.<\/li>\n<li><code>QRK_SPCBIXP<\/code>: Variable assignments directly preceding\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_14\" rel=\"nofollow\">special builtin commands<\/a>\nare exported, and persist as exported. (bash; yash)<\/li>\n<li><code>QRK_UNSETF<\/code>: If 'unset' is invoked without any option flag (-v or -f), and\nno variable by the given name exists but a function does, the shell unsets\nthe function. (bash)<\/li>\n<\/ul><h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#bugs\"><\/a>Bugs<\/h3>\n<p>Non-fatal shell bugs currently tested for are:<\/p>\n<ul><li><code>BUG_APPENDC<\/code>: When <code>set -C<\/code> (<code>noclobber<\/code>) is active, \"appending\" to a nonexistent\nfile with <code>&gt;&gt;<\/code> throws an error rather than creating the file. (zsh &lt; 5.1)\nThis is a bug making <code>use safe<\/code> less convenient to work with, as this sets\nthe <code>-C<\/code> (<code>-o noclobber<\/code>) option to reduce accidental overwriting of files.\nThe <code>safe<\/code> module requires an explicit override to tolerate this bug.<\/li>\n<li><code>BUG_ARITHINIT<\/code>: Using unset or empty variables (dash &lt;= 0.5.9.1)\nor unset variables (yash &lt;= 2.44) in arithmetic expressions causes the\nshell to exit, instead of taking them as a value of zero.<\/li>\n<li><code>BUG_ARITHLNNO<\/code>: The shell supports <code>$LINENO<\/code>, but the variable is\nconsidered unset in arithmetic contexts, like <code>$(( LINENO &gt; 0 ))<\/code>.\nThis makes it error out under <code>set -u<\/code> and default to zero otherwise.\nWorkaround: use shell expansion like <code>$(( $LINENO &gt; 0 ))<\/code>. (FreeBSD sh)<\/li>\n<li><code>BUG_ARITHSPLIT<\/code>: Unquoted <code>$((<\/code>arithmetic expressions<code>))<\/code> are not\nsubject to field splitting as expected. (zsh, pdksh, mksh&lt;=R49)<\/li>\n<li><code>BUG_ARITHTYPE<\/code>: In zsh, arithmetic assignments (using <code>let<\/code>, <code>$(( ))<\/code>,\netc.) on unset variables assign a numerical\/arithmetic type to a variable,\ncausing subsequent normal variable assignments to be interpreted as\narithmetic expressions and fail if they are not valid as such.<\/li>\n<li><code>BUG_BRACQUOT<\/code>: shell quoting within bracket patterns has no effect (zsh &lt; 5.3;\nksh93) This bug means the <code>-<\/code> retains it special meaning of 'character\nrange', and an initial <code>!<\/code> (and, on some shells, <code>^<\/code>) retains the meaning of\nnegation, even in quoted strings within bracket patterns, including quoted\nvariables.<\/li>\n<li><code>BUG_CASECC<\/code>: glob patterns as in <code>case<\/code> cannot match an escaped <code>^A<\/code>\n(<code>$CC01<\/code>) or <code>DEL<\/code> (<code>$CC7F<\/code>) control character. Found on: bash 2.05b, 3.0, 3.1<\/li>\n<li><code>BUG_CASELIT<\/code>: If a <code>case<\/code> pattern doesn't match as a pattern, it's tried\nagain as a literal string, even if the pattern isn't quoted. This can\nresult in false positives when a pattern doesn't match itself, like with\nbracket patterns. This contravenes POSIX and breaks use cases such as\ninput validation. (AT&amp;T ksh93) Note: modernish <code>match<\/code> works around this.<\/li>\n<li><code>BUG_CASESTAT<\/code>: The 'case' conditional construct prematurely clobbers the\nexit status <code>$?<\/code>. (found in zsh &lt; 5.3, Busybox ash &lt;= 1.25.0, dash &lt;\n0.5.9.1)<\/li>\n<li><code>BUG_CC7F<\/code>: Quoted expansions delete the DEL control character (<code>$CC7F<\/code>),\nexcept if that character is directly preceded in the value by the <code>\\1<\/code>\n(<code>$CC01<\/code>) control character. Note that this bug removes DEL from\n<code>$CONTROLCHARS<\/code> and <code>$ASCIICHARS<\/code>. (bash 2.05b, 3.0)<\/li>\n<li><code>BUG_CMDOPTEXP<\/code>: the <code>command<\/code> builtin does not recognise options if they\nresult from expansions. For instance, you cannot conditionally store <code>-p<\/code>\nin a variable like <code>defaultpath<\/code> and then do <code>command $defaultpath someCommand<\/code>. (found in zsh &lt; 5.3)<\/li>\n<li><code>BUG_CMDPV<\/code>: <code>command -pv<\/code> does not find builtins ({pd,m}ksh), does not\naccept the -p and -v options together (zsh &lt; 5.3) or ignores the '-p'\noption altogether (bash 3.2); in any case, it's not usable to find commands\nin the default system PATH.<\/li>\n<li><em><code>BUG_CMDSPASGN<\/code><\/em>: preceding a\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_14\" rel=\"nofollow\">special builtin<\/a>\nwith 'command' does not stop\npreceding invocation-local variable assignments from becoming global.\n(AT&amp;T ksh, 2010-ish versions; bash 2.05b and 3.0 in POSIX mode have a variant\nof this bug that only applies to the <code>eval<\/code>, <code>.<\/code>, and <code>source<\/code> builtins)<\/li>\n<li><code>BUG_CMDSPEXIT<\/code>: preceding a\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_14\" rel=\"nofollow\">special builtin<\/a>\nwith 'command' does not stop\nit from exiting the shell if the builtin encounters error.\n(zsh &lt; 5.2; mksh &lt; R50e)<\/li>\n<li><code>BUG_CMDVRESV<\/code>: 'command -v' does not find reserved words such as \"if\".\n(pdksh, mksh). This necessitates a workaround version of thisshellhas().<\/li>\n<li><code>BUG_CSCMTQUOT<\/code>: unbalanced single and double quotes and backticks in comments\nwithin command substitutions cause obscure and hard-to-trace syntax errors\nlater on in the script. (ksh88; pdksh, incl. {Open,Net}BSD ksh; bash 2.05b)<\/li>\n<li><code>BUG_CSNHDBKSL<\/code>: Backslashes within non-expanding here-documents within\ncommand substitutions are incorrectly expanded to perform newline joining,\nas opposed to left intact. (bash &lt;= 4.4, and pdksh)<\/li>\n<li><code>BUG_DOLRCSUB<\/code>: parsing problem where, inside a command substitution of\nthe form <code>$(...)<\/code>, the sequence <code>$$'...'<\/code> is treated as <code>$'...'<\/code> (i.e. as\na use of CESCQUOT), and <code>$$\"...\"<\/code> as <code>$\"...\"<\/code> (bash-specific translatable\nstring). (Found in bash up to 4.4)<\/li>\n<li><code>BUG_DQGLOB<\/code>: <em>glob<\/em>bing is not properly deactivated within\n<em>d<\/em>ouble-<em>q<\/em>uoted strings. Within double quotes, a <code>*<\/code> or <code>?<\/code> immediately\nfollowing a backslash is interpreted as a globbing character. This applies\nto both pathname expansion and pattern matching in <code>case<\/code>. Found in: dash.\n(The bug is not triggered when using modernish\n<a href=\"#user-content-string-tests\"><code>match<\/code><\/a>.)<\/li>\n<li><code>BUG_EMPTYBRE<\/code> is a <code>case<\/code> pattern matching bug in zsh &lt; 5.0.8: empty\nbracket expressions eat subsequent shell grammar, producing unexpected\nresults. This is particularly bad if you want to pass a bracket\nexpression using a variable or parameter, and that variable or parameter\ncould be empty. This means the grammar parsing depends on the contents\nof the variable!<\/li>\n<li><code>BUG_EVALCOBR<\/code>: <code>break<\/code> and <code>continue<\/code> do not work if they are within\n<code>eval<\/code>, wrongly causing loop execution to continue.\n(pdksh; mksh &lt; R55 2017\/04\/12)<\/li>\n<li><code>BUG_FNREDIRP<\/code>: I\/O redirections on function definitions are forgotten if the\nfunction is called as part of a pipeline with at least one <code>|<\/code>. (bash 2.05b)<\/li>\n<li><code>BUG_FNSUBSH<\/code>: Function definitions within subshells (including command\nsubstitutions) are ignored if a function by the same name exists in the\nmain shell, so the wrong function is executed. <code>unset -f<\/code> is also silently\nignored. ksh93 (all current versions as of June 2015) has this bug.<\/li>\n<li><code>BUG_HASHVAR<\/code>: On zsh, <code>$#var<\/code> means the length of <code>$var<\/code> - other shells and\nPOSIX require braces, as in <code>${#var}<\/code>. This causes interesting bugs when\ncombining <code>$#<\/code>, being the number of positional parameters, with other\nstrings. For example, in arithmetics: <code>$(($#-1))<\/code>, instead of the number of\npositional parameters minus one, is interpreted as <code>${#-}<\/code> concatenated with\n<code>1<\/code>. So, for zsh compatibility, always use <code>${#}<\/code> instead of <code>$#<\/code> unless it's\nstand-alone or followed by a space.<\/li>\n<li><code>BUG_HDOCBKSL<\/code>: Line continuation using <em>b<\/em>ac<em>ksl<\/em>ashes in expanding\n<em>h<\/em>ere-<em>doc<\/em>uments is handled incorrectly. (zsh up to 5.4.2)<\/li>\n<li><code>BUG_HDOCMASK<\/code>: Here-documents (and here-strings, see <code>HERESTRING<\/code>) use\ntemporary files. This fails if the current <code>umask<\/code> setting disallows the\nuser to read, so the here-document can't read from the shell's temporary\nfile. Workaround: ensure user-readable <code>umask<\/code> when using here-documents.\n(bash, pdksh, mksh, zsh)<\/li>\n<li><code>BUG_IFSGLOBC<\/code>: In glob pattern matching (such as in <code>case<\/code> and <code>[[<\/code>), if a\nwildcard character is part of <code>IFS<\/code>, it is matched literally instead of as a\nmatching character. This applies to glob characters <code>*<\/code>, <code>?<\/code>, <code>[<\/code> and <code>]<\/code>.\n<em>Since nearly all modernish functions use <code>case<\/code> for argument validation and\nother purposes, nearly every modernish function breaks on shells with this\nbug if IFS contains any of these three characters!<\/em>\n(Found in bash &lt; 4.4)<\/li>\n<li><code>BUG_IFSGLOBP<\/code>: In pathname expansion (filename globbing), if a\nwildcard character is part of <code>IFS<\/code>, it is matched literally instead of as a\nmatching character. This applies to glob characters <code>*<\/code>, <code>?<\/code>, <code>[<\/code> and <code>]<\/code>.\n(Bug found in bash, all versions up to at least 4.4)<\/li>\n<li><code>BUG_IFSGLOBS<\/code>: in glob pattern matching (as in <code>case<\/code> or paramter\nsubstitution with <code>#<\/code> and <code>%<\/code>), if <code>IFS<\/code> starts with <code>?<\/code> or <code>*<\/code> and the\n<code>\"$*\"<\/code> parameter expansion inserts any IFS separator characters, those\ncharacters are erroneously interpreted as wildcards when quoted \"$*\" is\nused as the glob pattern. (AT&amp;T ksh93)<\/li>\n<li><em><code>BUG_IFSISSET<\/code><\/em>: AT&amp;T ksh93 (recent versions): <code>${IFS+s}<\/code> always yields 's'\neven if IFS is unset. This applies to IFS only.<\/li>\n<li><code>BUG_ISSETLOOP<\/code>: AT&amp;T ksh93: Expansions like <code>${var+set}<\/code>\nremain static when used within a <code>for<\/code>, <code>while<\/code> or\n<code>until<\/code> loop; the expansions don't change along with the state of the\nvariable, so they cannot be used to check whether a variable is set\nwithin a loop if the state of that variable may change\nin the course of the loop.<\/li>\n<li><code>BUG_KUNSETIFS<\/code>: ksh93: Can't unset <code>IFS<\/code> under very specific\ncircumstances. <code>unset -v IFS<\/code> is a known POSIX shell idiom to activate\ndefault field splitting. With this bug, the <code>unset<\/code> builtin silently fails\nto unset IFS (i.e. fails to activate field splitting) if we're executing\nan <code>eval<\/code> or a trap and a number of specific conditions are met. See\n<a href=\"https:\/\/github.com\/modernish\/modernish\/blob\/master\/libexec\/modernish\/cap\/BUG_KUNSETIFS.t\">BUG_KUNSETIFS.t<\/a>\nfor more information.<\/li>\n<li><code>BUG_LNNOALIAS<\/code>: The shell has LINENO, but $LINENO is always expanded to 0\nwhen used within an alias. (pdksh variants, including mksh and oksh)<\/li>\n<li><code>BUG_LNNOEVAL<\/code>: The shell has LINENO, but $LINENO is always expanded to 0\nwhen used in 'eval'. (pdksh variants, including mksh and oksh)<\/li>\n<li><code>BUG_MULTIBIFS<\/code>: We're on a UTF-8 locale and the shell supports UTF-8\ncharacters in general (i.e. we don't have <code>BUG_MULTIBYTE<\/code>) -- however, using\nmultibyte characters as <code>IFS<\/code> field delimiters still doesn't work. For\nexample, <code>\"$*\"<\/code> joins positional parameters on the first byte of <code>$IFS<\/code>\ninstead of the first character. (ksh93, mksh, FreeBSD sh, Busybox ash)<\/li>\n<li><em><code>BUG_MULTIBYTE<\/code><\/em>: We're in a UTF-8 locale but the shell does not have\nmulti-byte\/variable-length character support. (Non-UTF-8 variable-length\nlocales are not yet supported.) Dash is a recent shell with this bug.<\/li>\n<li><code>BUG_NOCHCLASS<\/code>: POSIX-mandated character <code>[:<\/code>classes<code>:]<\/code> within bracket\n<code>[<\/code>expressions<code>]<\/code> are not supported in glob patterns. (pdksh, mksh, and\nfamily)<\/li>\n<li><code>BUG_NOEXPRO<\/code>: Cannot export read-only variables. (zsh 5.0.8-5.5.1 as sh)<\/li>\n<li><em><code>BUG_NOOCTAL<\/code><\/em>: Shell arithmetic does interpret numbers with leading\nzeroes as octal numbers; these are interpreted as decimal instead,\nthough POSIX specifies octal. (older mksh, 2013-ish versions)<\/li>\n<li><code>BUG_NOUNSETEX<\/code>: Cannot assign export attribute to variables in an unset\nstate; exporting a variable immediately sets it to the empty value.\n(zsh &lt; 5.3)<\/li>\n<li><code>BUG_NOUNSETRO<\/code>: Cannot freeze variables as readonly in an unset state.\nThis bug in zsh &lt; 5.0.8 makes the <code>readonly<\/code> command set them to the\nempty string instead.<\/li>\n<li><code>BUG_OPTNOLOG<\/code>: on dash, setting <code>-o nolog<\/code> causes <code>$-<\/code> to wreak havoc:\ntrying to expand <code>$-<\/code> silently aborts parsing of an entire argument,\nso e.g. <code>\"one,$-,two\"<\/code> yields <code>\"one,\"<\/code>. (Same applies to <code>-o debug<\/code>.)<\/li>\n<li><em><code>BUG_PARONEARG<\/code><\/em>: When <code>IFS<\/code> is empty on bash 3.x and 4.x (i.e. field\nsplitting is off), <code>${1+\"$@\"}<\/code> is counted as a single argument instead\nof each positional parameter as separate arguments. To avoid this bug,\nsimply use <code>\"$@\"<\/code> instead. (<code>${1+\"$@\"}<\/code> is an obsolete workaround for\na fatal shell bug, <code>FTL_UPP<\/code>.)<\/li>\n<li><code>BUG_PFRPAD<\/code>:  Negative padding value for strings in the <code>printf<\/code> builtin\ndoes not cause blank padding on the right-hand side, but inserts blank\npadding on the left-hand side as if the value were positive, e.g.\n<code>printf '[%-4s]' hi<\/code> outputs <code>[\u00a0\u00a0hi]<\/code>, not <code>[hi\u00a0\u00a0]<\/code>. (zsh 5.0.8)<\/li>\n<li><code>BUG_PP_01<\/code>: <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_05_02\" rel=\"nofollow\">POSIX says<\/a>\nthat empty <code>\"$@\"<\/code> generates zero fields but empty <code>''<\/code> or <code>\"\"<\/code> or\n<code>\"$emptyvariable\"<\/code> generates one empty field. This means concatenating\n<code>\"$@\"<\/code> with one or more other, separately quoted, empty strings (like\n<code>\"$@\"\"$emptyvariable\"<\/code>) should still produce one empty field. But on\nbash 3.x, this erroneously produces zero fields. (See also QRK_EMPTPPWRD)<\/li>\n<li><code>BUG_PP_02<\/code>: Like <code>BUG_PP_01<\/code>, but with unquoted <code>$@<\/code> and only\nwith <code>\"$emptyvariable\"$@<\/code>, not <code>$@\"$emptyvariable\"<\/code>. (pdksh)<\/li>\n<li><code>BUG_PP_03<\/code>: When IFS is unset or empty (zsh 5.3.1) or empty (pdksh),\nassigning <code>var=$*<\/code> only assigns the first field, failing to join and\ndiscarding the rest of the fields. Workaround: <code>var=\"$*\"<\/code>\n(POSIX leaves <code>var=$@<\/code>, etc. undefined, so we don't test for those.)<\/li>\n<li><code>BUG_PP_03A<\/code>: When IFS is unset, assignments like <code>var=$*<\/code>\nincorrectly remove leading and trailing spaces (but not tabs or\nnewlines) from the result. Workaround: quote the expansion. Found on:\nbash 4.3 and 4.4.<\/li>\n<li><code>BUG_PP_03B<\/code>: When IFS is unset, assignments like <code>var=${var+$*}<\/code>,\netc. incorrectly remove leading and trailing spaces (but not tabs or\nnewlines) from the result. Workaround: quote the expansion. Found on:\nbash 4.3 and 4.4.<\/li>\n<li><code>BUG_PP_03C<\/code>: When <code>IFS<\/code> is unset, assigning <code>var=${var-$*}<\/code> only assigns\nthe first field, failing to join and discarding the rest of the fields.\n(zsh 5.3, 5.3.1) Workaround: <code>var=${var-\"$*\"}<\/code><\/li>\n<li><code>BUG_PP_04<\/code>: If IFS is set and empty, assigning the positional parameters\nto a variable using a conditional assignment within a parameter substitution,\nsuch as <code>: ${var=$*}<\/code>, discards everything but the last field from the\nassigned value while incorrectly generating multiple fields for the\nexpansion. (pdksh, mksh)<\/li>\n<li><code>BUG_PP_04_S<\/code>: When IFS is null (empty), the result of a substitution\nlike <code>${var=$*}<\/code> is incorrectly field-split on spaces. The difference\nwith BUG_PP_04 is that the assignment itself succeeds normally.\nFound on: bash 4.2, 4.3<\/li>\n<li><code>BUG_PP_04A<\/code>: Like BUG_PP_03A, but for conditional assignments within\nparameter substitutions, as in <code>: ${var=$*}<\/code> or <code>: ${var:=$*}<\/code>.\nWorkaround: quote either <code>$*<\/code> within the expansion or the expansion\nitself. Found on: bash 2.05b through 4.4.<\/li>\n<li><code>BUG_PP_04B<\/code>: When assigning the positional parameters ($*) to a variable\nusing a conditional assignment within a parameter substitution, e.g.\n<code>: ${var:=$*}<\/code>, the fields are always joined and separated by spaces,\nregardless of the content or state of IFS. Workaround as in BUG_PP_04A.\n(bash 2.05b)<\/li>\n<li><code>BUG_PP_04D<\/code>: When field-splitting the result of an expansion such\nas <code>${var:=$*}<\/code>, if the first positional parameter starts with a space,\nan initial empty field is incorrectly generated. (mksh &lt;= R50)<\/li>\n<li><code>BUG_PP_04E<\/code>: Like <code>BUG_PP_04B<\/code> but not if IFS is set and empty. (bash 4.3)<\/li>\n<li><code>BUG_PP_05<\/code>: <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_05_02\" rel=\"nofollow\">POSIX says<\/a>\nthat empty <code>$@<\/code> and <code>$*<\/code> generate zero fields, but with null IFS, empty\nunquoted <code>$@<\/code> and <code>$*<\/code> yield one empty field. Found on: dash 0.5.9\nand 0.5.9.1; Busybox ash.<\/li>\n<li><code>BUG_PP_06<\/code>: <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_05_02\" rel=\"nofollow\">POSIX says<\/a>\nthat unquoted <code>$@<\/code> initially generates as many fields as there are\npositional parameters, and then (because <code>$@<\/code> is unquoted) each field is\nsplit further according to <code>IFS<\/code>. With this bug, the latter step is not\ndone. Found on: zsh &lt; 5.3<\/li>\n<li><code>BUG_PP_06A<\/code>: <a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_05_02\" rel=\"nofollow\">POSIX says<\/a>\nthat unquoted <code>$@<\/code> and <code>$*<\/code> initially generate as many fields as there are\npositional parameters, and then (because <code>$@<\/code> or <code>$*<\/code> is unquoted) each field is\nsplit further according to <code>IFS<\/code>. With this bug, the latter step is not\ndone if <code>IFS<\/code> is unset (i.e. default split). Found on: zsh &lt; 5.4<\/li>\n<li><code>BUG_PP_07<\/code>: unquoted <code>$*<\/code> and <code>$@<\/code> (including in substitutions like\n<code>${1+$@}<\/code> or <code>${var-$*}<\/code>) do not perform default field splitting if\n<code>IFS<\/code> is unset. Found on: zsh (up to 5.3.1) in sh mode<\/li>\n<li><code>BUG_PP_07A<\/code>: When <code>IFS<\/code> is unset, unquoted <code>$*<\/code> undergoes word splitting\nas if <code>IFS=' '<\/code>, and not the expected <code>IFS=\" ${CCt}${CCn}\"<\/code>.\nFound on: bash 4.4<\/li>\n<li><code>BUG_PP_08<\/code>: When <code>IFS<\/code> is empty, unquoted <code>$@<\/code> and <code>$*<\/code> do not generate\none field for each positional parameter as expected, but instead join\nthem into a single field without a separator. Found on: yash &lt; 2.44\nand dash &lt; 0.5.9 and Busybox ash &lt; 1.27.0<\/li>\n<li><code>BUG_PP_08B<\/code>: When <code>IFS<\/code> is empty, unquoted <code>$*<\/code> within a substitution (e.g.\n<code>${1+$*}<\/code> or <code>${var-$*}<\/code>) does not generate one field for each positional\nparameter as expected, but instead joins them into a single field without\na separator. Found on: bash 3 and 4<\/li>\n<li><code>BUG_PP_09<\/code>: When <code>IFS<\/code> is non-empty but does not contain a space,\nunquoted <code>$*<\/code> within a substitution (e.g. <code>${1+$*}<\/code> or <code>${var-$*}<\/code>) does\nnot generate one field for each positional parameter as expected,\nbut instead joins them into a single field separated by spaces\n(even though, as said, IFS does not contain a space).\nFound on: bash 2<\/li>\n<li><code>BUG_PP_10<\/code>: When <code>IFS<\/code> is null (empty), assigning <code>var=$*<\/code> removes any\n<code>$CC01<\/code> (^A) and <code>$CC7F<\/code> (DEL) characters. (bash 3, 4)<\/li>\n<li><code>BUG_PP_10A<\/code>: When <code>IFS<\/code> is non-empty, assigning <code>var=$*<\/code> prefixes each\n<code>$CC01<\/code> (^A) and <code>$CC7F<\/code> (DEL) character with a <code>$CC01<\/code> character. (bash 4.4)<\/li>\n<li><code>BUG_PP_11<\/code>: If <code>IFS<\/code> is set and empty (no field separator for <code>$*<\/code>),\nassigning unquoted <code>$*<\/code> to a variable (i.e.: <code>foo=$*<\/code>) causes the fields\nto be separated by a space in the variable, instead of joined together\nwithout a separator. (bash 2.05b, 3.0)<\/li>\n<li><code>BUG_PSUBBKSL1<\/code>: A backslash-escaped <code>}<\/code> character within a quoted parameter\nsubstitution is not unescaped. (bash 2 &amp; 3, standard dash, Busybox ash)<\/li>\n<li><code>BUG_PSUBNEWLN<\/code>: Due to a bug in the parser, parameter substitutions\nspread over more than one line cause a syntax error.\nWorkaround: instead of a literal newline, use <a href=\"#user-content-control-character-whitespace-and-shell-safe-character-constants\"><code>$CCn<\/code><\/a>.\n(found in dash &lt;= 0.5.9.1 and Busybox ash &lt;= 1.28.1)<\/li>\n<li><code>BUG_PSUBSQUOT<\/code>: in pattern matching parameter substitutions\n(<code>${param#pattern}<\/code>, <code>${param%pattern}<\/code>, <code>${param##pattern}<\/code> and\n<code>${param%%pattern}<\/code>), if the whole parameter substitution is quoted with\ndouble quotes, then single quotes in the <em>pattern<\/em> are not parsed. POSIX\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_06_02\" rel=\"nofollow\">says<\/a>\nthey are to keep their special meaning, so that glob characters may\nbe quoted. For example: <code>x=foobar; echo \"${x#'foo'}\"<\/code> should yield <code>bar<\/code>\nbut with this bug yields <code>foobar<\/code>. (dash; Busybox ash)<\/li>\n<li><code>BUG_PSUBSQHD<\/code>: Like BUG_PSUBSQUOT, but included a here-document instead of\nquoted with double quotes. (dash, pdksh, mksh)<\/li>\n<li><code>BUG_READWHSP<\/code>: If there is more than one field to read, <code>read<\/code> does not\ntrim trailing IFS whitespace (dash 0.5.7, 0.5.8) or any IFS whitespace\n(dash 0.5.6, 0.5.6.1).<\/li>\n<li><code>BUG_REDIRIO<\/code>: the I\/O redirection operator <code>&lt;&gt;<\/code> (open a file descriptor\nfor both read and write) defaults to opening standard output (i.e. is\nshort for <code>1&lt;&gt;<\/code>) instead of defaulting to opening standard input (<code>0&lt;&gt;<\/code>) as\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_07_07\" rel=\"nofollow\">POSIX specifies<\/a>.\n(AT&amp;T ksh93)<\/li>\n<li><code>BUG_REDIRPOS<\/code>: Buggy behaviour occurs if a <em>redir<\/em>ection is <em>pos<\/em>itioned\nin between to variable assignments in the same command. On zsh 5.0.x, a\nparse error is thrown. On zsh 5.1 to 5.4.2, anything following the\nredirection (other assignments or command arguments) is silently ignored.<\/li>\n<li><code>BUG_SCLOSEDFD<\/code>: bash &lt; 5.0 and dash &lt;= 0.5.9.1 fail to save a closed file\ndescriptor onto the shell-internal stack when added at the end of a block or\nloop (e.g. <code>} 8&lt;&amp;-<\/code> or <code>done 7&gt;&amp;-<\/code>), so any 'exec' of that descriptor will\nleak out of the block. However, pushing an open file descriptor works fine.\nWorkaround: enclose in another block that pushes the FD in an open state.<\/li>\n<li><code>BUG_SELECTEOF<\/code>: in a shell-native 'select' loop, the REPLY variable\nis not cleared if the user presses Ctrl-D to exit the loop. (zsh)<\/li>\n<li><code>BUG_SELECTRPL<\/code>: in a shell-native 'select' loop, input that is not a menu\nitem is not stored in the REPLY variable as it should be. (mksh R50 2014)<\/li>\n<li><code>BUG_SETOUTVAR<\/code>: The <code>set<\/code> builtin (with no arguments) only prints native\nfunction-local variables when called from a shell function. (yash &lt;= 2.46)<\/li>\n<li><code>BUG_SPCBILOC<\/code>: Variable assignments preceding\n<a href=\"http:\/\/pubs.opengroup.org\/onlinepubs\/9699919799\/utilities\/V3_chap02.html#tag_18_14\" rel=\"nofollow\">special builtins<\/a>\ncreate a partially function-local variable if a variable by the same name\nalready exists in the global scope. (bash &lt; 5.0 in POSIX mode)<\/li>\n<li><code>BUG_TESTERR0<\/code>: mksh: <code>test<\/code>\/<code>[<\/code> exits successfully (exit status 0) if\nan invalid argument is given to an operator. (mksh R52 fixes this)<\/li>\n<li><code>BUG_TESTERR1A<\/code>: AT&amp;T ksh: <code>test<\/code>\/<code>[<\/code> exits with a non-error 'false' status\n(1) if an invalid argument is given to an operator.<\/li>\n<li><code>BUG_TESTERR1B<\/code>: zsh: <code>test<\/code>\/<code>[<\/code> exits with status 1 (false) if there are\ntoo few or too many arguments, instead of a status &gt; 1 as it should do.<\/li>\n<li><code>BUG_TESTILNUM<\/code>: On dash (up to 0.5.8), giving an illegal number to <code>test -t<\/code>\nor <code>[ -t<\/code> causes some kind of corruption so the next <code>test<\/code>\/<code>[<\/code> invocation\nfails with an \"unexpected operator\" error even if it's legit.<\/li>\n<li><code>BUG_TESTONEG<\/code>: The <code>test<\/code>\/<code>[<\/code> builtin supports a <code>-o<\/code> unary operator to\ncheck if a shell option is set, but it ignores the <code>no<\/code> prefix on shell\noption names, so something like <code>[ -o noclobber ]<\/code> gives a false positive.\nBug found on yash up to 2.43. (The <code>TESTO<\/code> feature test implicitly checks\nagainst this bug and won't detect the feature if the bug is found.)<\/li>\n<li><code>BUG_TESTPAREN<\/code>: Incorrect exit status of <code>test -n<\/code>\/<code>-z<\/code> with values <code>(<\/code>,\n<code>)<\/code> or <code>!<\/code> in zsh 5.0.6 and 5.0.7. This can make scripts that process\narbitrary data (e.g. the shellquote function) take the wrong action unless\nworkarounds are implemented or modernish equivalents are used instead.\nAlso, spurious error message with both <code>test -n<\/code> and <code>test -z<\/code>.<\/li>\n<li><code>BUG_TESTRMPAR<\/code>: zsh: in binary operators with <code>test<\/code>\/<code>[<\/code>, if the first\nargument starts with <code>(<\/code> and the last with `)', both the first and the\nlast argument are completely removed, leaving only the operator, and the\nresult of the operation is incorrectly true because the operator is\nincorrectly parsed as a non-empty string. This applies to any operator.<\/li>\n<li><code>BUG_TRAPEMPT<\/code>: The <code>trap<\/code> builtin does not quote empty traps in its\noutput, rendering the output unsuitable for shell re-input. For instance,\n<code>trap '' INT; trap<\/code> outputs \"<code>trap --  INT<\/code>\" instead of \"<code>trap -- '' INT<\/code>\".\n(found in pdksh, mksh)<\/li>\n<li><code>BUG_TRAPEXIT<\/code>: the shell's <code>trap<\/code> builtin does not know the EXIT trap by\nname, but only by number (0). Using the name throws a \"bad trap\" error. Found in\n<a href=\"https:\/\/git.kernel.org\/pub\/scm\/libs\/klibc\/klibc.git\/tree\/usr\/dash\" rel=\"nofollow\">klibc 2.0.4 dash<\/a>.\nNote that the modernish <code>trap<\/code> and <code>pushtrap<\/code> commands effectively work\naround this bug, so it only affects scripts if they bypass modernish.<\/li>\n<li><code>BUG_TRAPRETIR<\/code>: Using 'return' within 'eval' triggers infinite recursion if\nboth a RETURN trap and the 'functrace' shell option are active. This bug in\nbash-only functionality triggers a crash when using modernish, so to avoid\nthis, modernish automatically disables the <code>functrace<\/code> shell option if a\n<code>RETURN<\/code> trap is set or pushed and this bug is detected. (bash 4.3, 4.4)<\/li>\n<\/ul><h3><a id=\"user-content-warning-ids\" class=\"anchor\" aria-hidden=\"true\" href=\"#warning-ids\"><\/a>Warning IDs<\/h3>\n<p>Warning IDs do not identify any characteristic of the shell, but instead\nwarn about a potentially problematic system condition that was detected at\ninitalisation time.<\/p>\n<ul><li><em><code>WRN_2UP2LOW<\/code><\/em>: This warning ID is issued if modernish is initialised\nin a UTF-8 locale and the modernish functions\n<a href=\"#user-content-touppertolower\"><code>toupper<\/code> and <code>tolower<\/code><\/a>\ncannot convert non-ASCII letters to upper or lower case -- e.g. accented\nLatin letters, Greek, cyrillic. Modernish tries hard to make it work,\nfalling back to using the external <code>tr<\/code>, <code>awk<\/code>, <code>gawk<\/code> or GNU <code>sed<\/code>\ncommand if the shell can't convert non-ASCII (or any) characters itself.\nSo if the shell cannot do it, then this warning ID is only issued if none\nof these external commands can convert them either. But if the shell can,\nthen this warning ID is not issued even if the external commands cannot.\nThis means that <em>the result of <code>thisshellhas WRN_2UP2LOW<\/code><\/em> <em><strong>only<\/strong><\/em>\n<em>applies to the modernish <code>toupper<\/code> and <code>tolower<\/code> functions<\/em> and not to\nyour shell or any external command in particular.<\/li>\n<li><em><code>WRN_NOSIGPIPE<\/code><\/em>: Modernish has detected that the process that launched\nthe current program has set <code>SIGPIPE<\/code> to ignore, an irreversible condition\nthat is in turn inherited by any process started by the current shell, and\ntheir subprocesses, and so on. This makes it impossible to detect\n<a href=\"#user-content-modernish-system-constants\"><code>$SIGPIPESTATUS<\/code><\/a>;\nit is set to the special\nvalue 99999 which is impossible as an exit status. But it also makes it\nirrelevant what that status is, because neither the current shell nor any\nprocess it spawns is now capable of receiving <code>SIGPIPE<\/code>. The\n<a href=\"#hardening-while-allowing-for-broken-pipes\"><code>-P<\/code> option to <code>harden<\/code><\/a>\nis also rendered irrelevant. Note that a command such as <code>yes | head -n 10<\/code> now never ends; the only way <code>yes<\/code> would ever stop trying to write\nlines is by receiving <code>SIGPIPE<\/code> from <code>head<\/code>, which is being ignored.\nPrograms that use commands in this fashion should check <code>if thisshellhas WRN_NOSIGPIPE<\/code> and either employ workarounds or refuse to run if so.<\/li>\n<\/ul><h2><a id=\"user-content-appendix-b\" class=\"anchor\" aria-hidden=\"true\" href=\"#appendix-b\"><\/a>Appendix B<\/h2>\n<p>Modernish comes with a suite of regression tests to detect bugs in modernish\nitself, which can be run using <code>modernish --test<\/code> after installation. By\ndefault, it will run all the tests verbosely but without tracing the command\nexecution. The <code>install.sh<\/code> installer will run the suite quietly on the\nselected shell before installation.<\/p>\n<p>A few options are available to specify after <code>--test<\/code>:<\/p>\n<ul><li><code>-e<\/code>: run expensive (i.e. slow or memory-hogging) tests that are disabled\nby default. Some tests run without this option but are more thorough\nif this option is given.<\/li>\n<li><code>-q<\/code>: quieter operation; report expected fails [known shell bugs]\nand unexpected fails [bugs in modernish]). Add <code>-q<\/code> again for\nquietest operation (report unexpected fails only).<\/li>\n<li><code>-s<\/code>: entirely silent operation.<\/li>\n<li><code>-t<\/code>: run only specific test sets or tests. Test sets are those listed\nin the full default output of <code>modernish --test<\/code>. This option requires\nan option-argument in the following format:<br \/><em>testset1<\/em><code>:<\/code><em>num1<\/em><code>,<\/code><em>num2<\/em><code>,<\/code>\u2026<code>\/<\/code><em>testset2<\/em><code>:``*num1*<\/code>,<code>*num2*<\/code>,<code>\u2026<\/code>\/<code>\u2026     The colon followed by numbers is optional; if omitted, the entire set will be run, otherwise the given numbered tests will be run in the given order. Example: <\/code>modernish --test -t match:2,4,7\/arith\/shellquote:1<code>runs test 2, 4 and 7 from the<\/code>match<code>set, the entire<\/code>arith<code>set, and only test 1 from the<\/code>shellquote` set.<\/li>\n<li><code>-x<\/code>: trace each test using the shell's <code>xtrace<\/code> facility. Each trace is\nstored in a separate file in a specially created temporary directory. By\ndefault, the trace is deleted if a test does not produce an unexpected\nfail. Add <code>-x<\/code> again to keep expected fails as well, and again to\nkeep all traces regardless of result. If any traces were saved,\nmodernish will tell you the location of the temporary directory at the\nend, otherwise it will silently remove the directory again.<\/li>\n<\/ul><p>These short options can be combined so, for example,\n<code>--test -qxx<\/code> is the same as <code>--test -q -x -x<\/code>.<\/p>\n<h3><a id=\"user-content-difference-between-bugquirkfeature-tests-and-regression-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#difference-between-bugquirkfeature-tests-and-regression-tests\"><\/a>Difference between bug\/quirk\/feature tests and regression tests<\/h3>\n<p>Note the difference between these regression tests and the tests listed\nabove in <a href=\"#user-content-appendix-a\">Appendix A<\/a>. The latter are tests for\nwhatever shell is executing modernish: they test for capabilities (features,\nquirks, bugs) of the current shell. They are meant to be run via\n<a href=\"#user-content-feature-testing\"><code>thisshellhas<\/code><\/a> and are designed to be\ntaken advantage of in scripts. On the other hand, these tests run by\n<code>modernish --test<\/code> are regression tests for modernish itself. It does not\nmake sense to use these in a script.<\/p>\n<p>New\/unknown shell bugs can still cause modernish regression tests to fail,\nof course. That's why some of the regression tests also check for\nconsistency with the results of the feature\/quirk\/bug tests: if there is a\nshell bug in a widespread release version that modernish doesn't know about\nyet, this in turn is considered to be a bug in modernish, because one of its\ngoals is to know about all the shell bugs in all released shell versions\ncurrently seeing significant use.<\/p>\n<h3><a id=\"user-content-testing-modernish-on-all-your-shells\" class=\"anchor\" aria-hidden=\"true\" href=\"#testing-modernish-on-all-your-shells\"><\/a>Testing modernish on all your shells<\/h3>\n<p>The <code>testshells.sh<\/code> program in <code>share\/doc\/modernish\/examples<\/code> can be used to\nrun the regression test suite on all the shells installed on your system.\nYou could put it as <code>testshells<\/code> in some convenient location in your\n<code>$PATH<\/code>, and then simply run:<\/p>\n<pre>testshells modernish --test\n<\/pre>\n<p>(adding any further options you like -- for instance, you might like to add\n<code>-q<\/code> to avoid very long terminal output). On first run, <code>testshells<\/code> will\ngenerate a list of shells it can find on your system and it will give you a\nchance to edit it before proceeding.<\/p>\n<hr \/><p><code>EOF<\/code><\/p>\n<\/article>","created_at":"2018-11-07T16:44:58+0000","updated_at":"2018-11-07T16:44:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":106,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/13722347?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":336,"title":"Plain text","url":"https:\/\/people.csail.mit.edu\/rivest\/Sexp.txt","content":"<pre>Network  Working Group                                         R. Rivest\r\nInternet Draft                                               May 4, 1997\r\nExpires November 4, 1997\r\n\r\n\r\n\t                      S-Expressions\r\n                        draft-rivest-sexp-00.txt\r\n\r\n\r\nStatus of this Memo\r\n\r\n   Distribution of this memo is unlimited.\r\n\r\n   This document is an Internet-Draft.  Internet Drafts are working\r\n   documents of the Internet Engineering Task Force (IETF), its Areas,\r\n   and its Working Groups.  Note that other groups may also distribute\r\n   working documents as Internet Drafts.\r\n\r\n   Internet Drafts are draft documents valid for a maximum of six\r\n   months, and may be updated, replaced, or obsoleted by other documents\r\n   at any time.  It is not appropriate to use Internet Drafts as\r\n   reference material, or to cite them other than as a ``working draft''\r\n   or ``work in progress.''\r\n\r\n   To learn the current status of any Internet-Draft, please check the\r\n   ``1id-abstracts.txt'' listing contained in the internet-drafts Shadow\r\n   Directories on: ftp.is.co.za (Africa), nic.nordu.net (Europe),\r\n   ds.internic.net (US East Coast), ftp.isi.edu (US West Coast),\r\n   or munnari.oz.au (Pacific Rim)\r\n\r\n\r\nAbstract\r\n\r\nThis memo describes a data structure called \"S-expressions\" that are\r\nsuitable for representing arbitrary complex data structures.  We make\r\nprecise the encodings of S-expressions: we give a \"canonical form\" for\r\nS-expressions, described two \"transport\" representations, and also\r\ndescribe an \"advanced\" format for display to people.\r\n\r\n\r\n\r\n1. Introduction\r\n\r\nS-expressions are data structures for representing complex data.  They\r\nare either byte-strings (\"octet-strings\") or lists of simpler\r\nS-expressions.  Here is a sample S-expression:\r\n\r\n\t(snicker \"abc\" (#03# |YWJj|))\r\n\r\nIt is a list of length three: \r\n\r\n\t-- the octet-string \"snicker\"\r\n\r\n\t-- the octet-string \"abc\"\r\n\r\n\t-- a sub-list containing two elements:\r\n\t\t- the hexadecimal constant #03#\r\n\t\t- the base-64 constant |YWJj| (which is the same as \"abc\")\r\n\r\nThis note gives a specific proposal for constructing and utilizing\r\nS-expressions.  The proposal is independent of any particular application.\r\n\r\nHere are the design goals for S-expressions:\r\n\r\n  -- generality: S-expressions should be good at representing arbitrary\r\n     data.\r\n\r\n  -- readability: it should be easy for someone to examine and \r\n     understand the structure of an S-expression.\r\n\r\n  -- economy: S-expressions should represent data compactly.\r\n\r\n  -- tranportability: S-expressions should be easy to transport\r\n     over communication media (such as email) that are known to be\r\n     less than perfect.\r\n\r\n  -- flexibility: S-expressions should make it relatively simple to\r\n     modify and extend data structures.\r\n\r\n  -- canonicalization: it should be easy to produce a unique \r\n     \"canonical\" form of an S-expression, for digital signature purposes.\r\n\r\n  -- efficiency: S-expressions should admit in-memory representations\r\n     that allow efficient processing.\r\n\r\n\r\nSection 2 gives an introduction to S-expressions.\r\nSection 3 discusses the character sets used.\r\nSection 4 presents the various representations of octet-strings.\r\nSection 5 describes how to represent lists.\r\nSection 6 discusses how S-expressions are represented for various uses.\r\nSection 7 gives a BNF syntax for S-expressions.\r\nSection 8 talks about how S-expressions might be represented in memory.\r\nSection 9 briefly describes implementations for handling S-expressions.\r\nSection 10 discusses how applications might utilize S-expressions.\r\nSection 11 gives historical notes on S-expressions.\r\nSection 12 gives references.\r\n\r\n2. S-expressions -- informal introduction\r\n\r\nInformally, an S-expression is either:\r\n\t-- an octet-string, or\r\n\t-- a finite list of simpler S-expressions.\r\n\r\nAn octet-string is a finite sequence of eight-bit octets.  There may be\r\nmany different but equivalent ways of representing an octet-string\r\n\r\n\tabc\t\t-- as a token\r\n\r\n\t\"abc\"\t\t-- as a quoted string\r\n\r\n\t#616263#\t-- as a hexadecimal string\r\n\r\n\t3:abc\t\t-- as a length-prefixed \"verbatim\" encoding\r\n\r\n\t{MzphYmM=}\t-- as a base-64 encoding of the verbatim encoding\r\n\t\t\t   (that is, an encoding of \"3:abc\")\r\n\r\n\t|YWJj|\t\t-- as a base-64 encoding of the octet-string \"abc\"\r\n\r\nThese encodings are all equivalent; they all denote the same octet string.\r\n\r\nWe will give details of these encodings later on, and also describe how to\r\ngive a \"display type\" to a byte string.\r\n\r\nA list is a finite sequence of zero or more simpler S-expressions.  A list\r\nmay be represented by using parentheses to surround the sequence of encodings\r\nof its elements, as in:\r\n\r\n\t(abc (de #6667#) \"ghi jkl\")\r\n\r\nAs we see, there is variability possible in the encoding of an\r\nS-expression.  In some cases, it is desirable to standardize or\r\nrestrict the encodings; in other cases it is desirable to have no\r\nrestrictions.  The following are the target cases we aim to handle:\r\n\r\n\t-- a \"transport\" encoding for transporting the S-expression between\r\n\t   computers.\r\n\r\n\t-- a \"canonical\" encoding, used when signing the S-expression.\r\n\r\n\t-- an \"advanced\" encoding used for input\/output to people.\r\n\r\n\t-- an \"in-memory\" encoding used for processing the S-expression in\r\n\t   the computer.\r\n\r\nThese need not be different; in this proposal the canonical encoding\r\nis the same as the transport encoding, for example.  In this note we\r\npropose (related) encoding techniques for each of these uses.\r\n\r\n3. Character set \r\n\r\nWe will be describing encodings of S-expressions.  Except when giving\r\n\"verbatim\" encodings, the character set used is limited to the following\r\ncharacters in US-ASCII:\r\n\tAlphabetic:\tA B ... Z a b ... z\r\n\tnumeric:\t0 1 ... 9\r\n\twhitespace:\tspace, horizontal tab, vertical tab, form-feed\r\n\t\t\tcarriage-return, line-feed\r\n\tThe following graphics characters, which we call \"pseudo-alphabetic\":\r\n\t\t\t- hyphen or minus\r\n\t\t\t. period\r\n\t\t\t\/ slash\r\n\t\t\t_ underscore\r\n\t\t\t: colon\r\n\t\t\t* asterisk\r\n\t\t\t+ plus\r\n\t\t\t= equal\r\n\tThe following graphics characters, which are \"reserved punctuation\":\r\n\t\t\t( left parenthesis\r\n\t\t\t) right parenthesis\r\n\t\t\t[ left bracket\r\n\t\t\t] right bracket\r\n\t\t\t{ left brace\r\n\t\t\t} right brace\r\n\t\t\t| vertical bar\r\n\t\t\t# number sign\r\n\t\t\t\" double quote\r\n\t\t\t&amp; ampersand\r\n\t\t\t\\ backslash\r\n\tThe following characters are unused and unavailable, except in\r\n\t\"verbatim\" encodings:\r\n\t\t\t! exclamation point\r\n\t\t\t% percent\r\n\t\t\t^ circumflex\r\n\t\t\t~ tilde\r\n\t\t\t; semicolon\r\n\t\t\t' apostrophe\r\n\t\t\t, comma\r\n\t\t\t&lt; less than\r\n\t\t\t&gt; greater than\r\n\t\t\t? question mark\r\n\r\n\r\n4. Octet string representations\r\n\r\nThis section describes in detail the ways in which an octet-string may\r\nbe represented.\r\n\r\nWe recall that an octet-string is any finite sequence of octets, and\r\nthat the octet-string may have length zero.\r\n\r\n\r\n4.1 Verbatim representation\r\n\r\nA verbatim encoding of an octet string consists of four parts:\r\n\r\n\t-- the length (number of octets) of the octet-string, \r\n\t   given in decimal most significant digit first, with\r\n\t   no leading zeros.\r\n\r\n\t-- a colon \":\"\r\n\r\n\t-- the octet string itself, verbatim.\r\n\r\nThere are no blanks or whitespace separating the parts.  No \"escape\r\nsequences\" are interpreted in the octet string.  This encoding is also\r\ncalled a \"binary\" or \"raw\" encoding.\r\n\t\r\nHere are some sample verbatim encodings:\r\n\r\n\t3:abc\r\n\t7:subject\r\n\t4:::::\r\n\t12:hello world!\r\n\t10:abcdefghij\r\n\t0:\r\n\r\n4.2 Quoted-string representation\r\n\r\nThe quoted-string representation of an octet-string consists of:\r\n\r\n\t-- an optional decimal length field\r\n\r\n\t-- an initial double-quote (\")\r\n\r\n\t-- the octet string with \"C\" escape conventions (\\n,etc)\r\n\r\n\t-- a final double-quote (\")\r\n\r\nThe specified length is the length of the resulting string after any\r\nescape sequences have been handled.  The string does not have any\r\n\"terminating NULL\" that C includes, and the length does not count such\r\na character.\r\n\r\nThe length is optional.\r\n\r\nThe escape conventions within the quoted string are as follows (these follow\r\nthe \"C\" programming language conventions, with an extension for \r\nignoring line terminators of just LF or CRLF):\r\n\t\\b\t\t-- backspace\r\n\t\\t\t\t-- horizontal tab\r\n\t\\v \t\t-- vertical tab\r\n\t\\n\t\t-- new-line\r\n\t\\f\t\t-- form-feed\r\n\t\\r\t\t-- carriage-return\r\n\t\\\"\t\t-- double-quote\r\n\t\\'\t\t-- single-quote\r\n\t\\\\\t\t-- back-slash\r\n\t\\ooo\t\t-- character with octal value ooo (all three digits\r\n\t\t\t   must be present)\r\n\t\\xhh\t\t-- character with hexadecimal value hh (both digits\r\n\t\t\t   must be present)\r\n\t\\ -- causes carriage-return to be ignored.\r\n\t\\       -- causes linefeed to be ignored\r\n\t\\ -- causes CRLF to be ignored.\r\n\t\\ -- causes LFCR to be ignored.\r\n\r\nHere are some examples of quoted-string encodings:\r\n\r\n\t\"subject\"\r\n\t\"hi there\"\r\n\t7\"subject\"\r\n\t3\"\\n\\n\\n\"\r\n\t\"This has\\n two lines.\"\r\n\t\"This has\\\r\n\tone.\"\r\n\t\"\"\r\n\r\n4.3 Token representation \r\n\r\nAn octet string that meets the following conditions may be given\r\ndirectly as a \"token\".\r\n\r\n\t-- it does not begin with a digit\r\n\r\n\t-- it contains only characters that are\r\n\t\t-- alphabetic (upper or lower case),\r\n\t\t-- numeric, or\r\n\t\t-- one of the eight \"pseudo-alphabetic\" punctuation marks:\r\n\t\t\t-   .   \/   _   :  *  +  =  \r\n\t(Note: upper and lower case are not equivalent.)\r\n\t(Note: A token may begin with punctuation, including \":\").\r\n\r\nHere are some examples of token representations:\r\n\r\n\tsubject\r\n\tnot-before\r\n\tclass-of-1997\r\n\t\/\/microsoft.com\/names\/smith\r\n\t*\r\n\r\n\r\n4.4 Hexadecimal representation\r\n\r\nAn octet-string may be represented with a hexadecimal encoding consisting of:\r\n\r\n\t-- an (optional) decimal length of the octet string\r\n\r\n\t-- a sharp-sign \"#\"\r\n\r\n\t-- a hexadecimal encoding of the octet string, with each octet\r\n\t   represented with two hexadecimal digits, most significant\r\n\t   digit first.\r\n\r\n\t-- a sharp-sign \"#\"\r\n\r\nThere may be whitespace inserted in the midst of the hexadecimal\r\nencoding arbitrarily; it is ignored.  It is an error to have\r\ncharacters other than whitespace and hexadecimal digits.\r\n\r\nHere are some examples of hexadecimal encodings:\r\n\r\n\t#616263#\t\t-- represents \"abc\"\r\n\t3#616263#\t\t-- also represents \"abc\"\r\n\t# 616\r\n\t  263 #                 -- also represents \"abc\"\r\n\t\r\n\r\n4.5 Base-64 representation\r\n\r\nAn octet-string may be represented in a base-64 coding consisting of:\r\n\r\n\t-- an (optional) decimal length of the octet string\r\n\r\n\t-- a vertical bar \"|\"\r\n\r\n\t-- the rfc 1521 base-64 encoding of the octet string.  \r\n\r\n\t-- a final vertical bar \"|\"\r\n\r\nThe base-64 encoding uses only the characters \r\n\tA-Z  a-z  0-9  +  \/  = \r\nIt produces four characters of output for each three octets of input.\r\nIf the input has one or two left-over octets of input, it produces an\r\noutput block of length four ending in two or one equals signs, respectively.\r\nOutput routines compliant with this standard MUST output the equals signs\r\nas specified.  Input routines MAY accept inputs where the equals signs are\r\ndropped.\r\n\r\nThere may be whitespace inserted in the midst of the base-64 encoding\r\narbitrarily; it is ignored.  It is an error to have characters other\r\nthan whitespace and base-64 characters.  \r\n\r\nHere are some examples of base-64 encodings:\r\n\r\n\t|YWJj|\t\t-- represents \"abc\"\r\n\t| Y W\r\n\t  J j |\t\t-- also represents \"abc\"\r\n\t3|YWJj|\t\t-- also represents \"abc\"\r\n\t|YWJjZA==|\t-- represents \"abcd\"\r\n\t|YWJjZA|\t-- also represents \"abcd\"\r\n\r\n\r\n4.6 Display hint\r\n\r\nAny octet string may be preceded by a single \"display hint\".\r\n\r\nThe purposes of the display hint is to provide information on how\r\nto display the octet string to a user.  It has no other function.\r\nMany of the MIME types work here.\r\n\r\nA display-hint is an octet string surrounded by square brackets.\r\nThere may be whitespace separating the octet string from the\r\nsurrounding brackets.  Any of the legal formats may be used for the\r\noctet string.\r\n\r\nHere are some examples of display-hints:\r\n\r\n\t[image\/gif]\r\n\t[URI]\r\n\t[charset=unicode-1-1]\r\n\t[text\/richtext]\r\n\t[application\/postscript]\r\n\t[audio\/basic]\r\n\t[\"http:\/\/abc.com\/display-types\/funky.html\"]\r\n\r\nIn applications an octet-string that is untyped may be considered to have\r\na pre-specified \"default\" mime type.  The mime type\r\n\t\t\"text\/plain; charset=iso-8859-1\" \r\nis the standard default.\r\n\r\n\r\n4.7  Equality of octet-strings\r\n\r\nTwo octet strings are considered to be \"equal\" if and only if they\r\nhave the same display hint and the same data octet strings.\r\n\r\nNote that octet-strings are \"case-sensitive\"; the octet-string \"abc\"\r\nis not equal to the octet-string \"ABC\".\r\n\r\nAn untyped octet-string can be compared to another octet-string (typed\r\nor not) by considering it as a typed octet-string with the default\r\nmime-type.\r\n\r\n\r\n5. Lists\r\n\r\nJust as with octet-strings, there are several ways to represent an\r\nS-expression.  Whitespace may be used to separate list elements, but\r\nthey are only required to separate two octet strings when otherwise\r\nthe two octet strings might be interpreted as one, as when one token\r\nfollows another.  Also,\twhitespace may follow the initial left\r\nparenthesis, or precede the final right parenthesis.\r\n\r\nHere are some examples of encodings of lists:\r\n\r\n\t(a b c)\r\n\r\n\t( a ( b c ) ( ( d e ) ( e f ) )  )\r\n\r\n\t(11:certificate(6:issuer3:bob)(7:subject5:alice))\r\n\r\n\t({3Rt=} \"1997\" murphy 3:{XC++})\r\n\r\n\r\n6. Representation types\r\n\r\nThere are three \"types\" of representations: \r\n\r\n\t-- canonical\r\n\r\n\t-- basic transport\r\n\r\n\t-- advanced transport\r\n\r\nThe first two MUST be supported by any implementation; the last is\r\noptional.\r\n\r\n\r\n6.1  Canonical representation\r\n\r\nThis canonical representation is used for digital signature purposes,\r\ntransmission, etc.  It is uniquely defined for each S-expression.  It\r\nis not particularly readable, but that is not the point.  It is\r\nintended to be very easy to parse, to be reasonably economical, and to\r\nbe unique for any S-expression.\r\n\r\nThe \"canonical\" form of an S-expression represents each octet-string\r\nin verbatim mode, and represents each list with no blanks separating\r\nelements from each other or from the surrounding parentheses.\r\n\r\nHere are some examples of canonical representations of S-expressions:\r\n\t\r\n\t(6:issuer3:bob)\r\n\r\n\t(4:icon[12:image\/bitmap]9:xxxxxxxxx)\r\n\r\n\t(7:subject(3:ref5:alice6:mother))\r\n\r\n\r\n6.2 Basic transport representation\r\n\r\nThere are two forms of the \"basic transport\" representation:\r\n\r\n\t-- the canonical representation\r\n\r\n\t-- an rfc-2045 base-64 representation of the canonical representation,\r\n           surrounded by braces.\r\n\r\nThe transport mechanism is intended to provide a universal means of\r\nrepresenting S-expressions for transport from one machine to another.\r\n\r\nHere are some examples of an S-expression represented in basic\r\ntransport mode:\r\n\r\n\t(1:a1:b1:c)\r\n\r\n\t{KDE6YTE6YjE6YykA}\r\n\t\t\t\t\r\n\t\t(this is the same S-expression encoded in base-64)\r\n\r\nThere is a difference between the brace notation for base-64 used here\r\nand the || notation for base-64'd octet-strings described above.  Here\r\nthe base-64 contents are converted to octets, and then re-scanned as\r\nif they were given originally as octets.  With the || notation, the\r\ncontents are just turned into an octet-string.\r\n\r\n\r\n6.3 Advanced transport representation\r\n\r\nThe \"advanced transport\" representation is intended to provide more\r\nflexible and readable notations for documentation, design, debugging,\r\nand (in some cases) user interface.\r\n\r\nThe advanced transport representation allows all of the representation\r\nforms described above, include quoted strings, base-64 and hexadecimal\r\nrepresentation of strings, tokens, representations of strings with\r\nomitted lengths, and so on.\r\n\r\n\r\n7. BNF for syntax\r\n\r\nWe give separate BNF's for canonical and advanced forms of S-expressions.\r\nWe use the following notation:\r\n\t* \t\tmeans 0 or more occurrences of \r\n\t+\t\tmeans 1 or more occurrences of   \r\n\t?\t\tmeans 0 or 1 occurrences of \r\n\tparentheses\tare used for grouping, as in ( | )*\r\n\r\nFor canonical and basic transport:\r\n\r\n    \t::  | \r\n   \t:: ?  ;\r\n\t::  ;\r\n  \t:: \"[\"  \"]\" ;\r\n      \t::  \":\"  ;\r\n  \t:: + ;\r\n\t\t-- decimal numbers should have no unnecessary leading zeros\r\n \t-- any string of bytes, of the indicated length\r\n     \t:: \"(\" * \")\" ;\r\n :: \"0\" | ... | \"9\" ;\r\n\r\nFor advanced transport:\r\n\r\n    \t::  | \r\n   \t:: ?  ;\r\n\t::  |  |  |  | \r\n\t\t            ;\r\n  \t:: \"[\"  \"]\" ;\r\n      \t::  \":\"  ;\r\n  \t:: + ;\r\n\t\t-- decimal numbers should have no unnecessary leading zeros\r\n \t-- any string of bytes, of the indicated length\r\n    \t:: + ;\r\n  \t:: ? \"|\" (  |  )* \"|\" ;\r\n   :: \"#\" (  |  )* \"#\" ;\r\n :: ?   \r\n :: \"\\\"\"  \"\\\"\"\r\n     \t:: \"(\" (  |  )* \")\" ;\r\n \t:: * ;\r\n  \t::  |  |  ;\r\n       \t::  |  |  ;\r\n  \t:: \"a\" | ... | \"z\" ;\r\n  \t:: \"A\" | ... | \"Z\" ;\r\n :: \"0\" | ... | \"9\" ;\r\n     ::  | \"A\" | ... | \"F\" | \"a\" | ... | \"f\" ;\r\n \t:: \"-\" | \".\" | \"\/\" | \"_\" | \":\" | \"*\" | \"+\" | \"=\" ;\r\n :: \" \" | \"\\t\" | \"\\r\" | \"\\n\" ;\r\n \t::  |  | \"+\" | \"\/\" | \"=\" ;\r\n        \t:: \"\" ;\r\n\r\n8. In-memory representations\r\n\r\nFor processing, the S-expression would typically be parsed and represented\r\nin memory in a more more amenable to efficient processing.  We suggest\r\ntwo alternatives:\r\n\r\n\t-- \"list-structure\"\r\n\r\n\t-- \"array-layout\"\r\n\r\nWe only sketch these here, as they are only suggestive.  The code referenced\r\nbelow illustrates these styles in more detail.\r\n\r\n\r\n8.1. List-structure memory representation\r\n\r\nHere there are separate records for simple-strings, strings, and\r\nlists.  An S-expression of the form (\"abc\" \"de\") would require two\r\nrecords for the simple strings, two for the strings, and two for the\r\nlist elements.  This is a fairly conventional representation, and\r\ndetails are omitted here.\r\n\r\n8.2 Array-layout memory representation\r\n\r\nHere each S-expression is represented as a contiguous array of bytes.\r\nThe first byte codes the \"type\" of the S-expression:\r\n\r\n\t01 \toctet-string\r\n\r\n\t02\toctet-string with display-hint\r\n\r\n\t03\tbeginning of list (and 00 is used for \"end of list\")\r\n\r\nEach of the three types is immediately followed by a k-byte integer\r\nindicating the size (in bytes) of the following representation.  Here\r\nk is an integer that depends on the implementation, it might be\r\nanywhere from 2 to 8, but would be fixed for a given implementation;\r\nit determines the size of the objects that can be handled.  The transport\r\nand canonical representations are independent of the choice of k made by\r\nthe implementation.\r\n\r\nAlthough the length of lists are not given in the usual S-expression\r\nnotations, it is easy to fill them in when parsing; when you reach a\r\nright-parenthesis you know how long the list representation was, and\r\nwhere to go back to fill in the missing length.\r\n\r\n\r\n8.2.1 Octet string\r\n\r\nThis is represented as follows:\r\n\r\n\t01  \r\n\r\nFor example (here k = 2)\r\n\r\n\t01 0003 a b c\r\n\r\n8.2.2 Octet-string with display-hint\r\n\r\nThis is represented as follows:\r\n\r\n\t02 \r\n\t  01      \/* for display-type *\/\r\n\t  01      \/* for octet-string *\/\r\n\r\nFor example, the S-expression \r\n\t\r\n\t[gif] #61626364# \r\n\r\nwould be represented as (with k = 2)\r\n\r\n\t02 000d\r\n\t  01 0003  g  i  f\r\n\t  01 0004 61 62 63 64\r\n\r\n8.2.3 List\r\n\r\nThis is represented as\r\n\r\n\t03     ...  00\r\n\r\nFor example, the list (abc [d]ef (g)) is represented in memory as (with k=2)\r\n\r\n\t03 001b\r\n\t  01 0003 a b c\r\n          02 0009 \r\n            01 0001 d\r\n            01 0002 e f\r\n          03 0005\r\n            01 0001 g\r\n          00\r\n        00\r\n\r\n9. Code\r\n\r\nThere is code available for reading and parsing the various\r\nS-expression formats proposed here.  \r\n\r\nSee http:\/\/theory.lcs.mit.edu\/~rivest\/sexp.html\r\n\r\n\r\n10. Utilization of S-expressions\r\n\r\nThis note has described S-expressions in general form.  Application writers\r\nmay wish to restrict their use of S-expressions in various ways.  Here are\r\nsome possible restrictions that might be considered:\r\n\r\n\t-- no display-hints\r\n\t-- no lengths on hexadecimal, quoted-strings, or base-64 encodings\r\n\t-- no empty lists\r\n\t-- no empty octet-strings\r\n\t-- no lists having another list as its first element\r\n\t-- no base-64 or hexadecimal encodings\r\n\t-- fixed limits on the size of octet-strings\r\n\r\n11. Historical note\r\n\r\nThe S-expression technology described here was originally developed\r\nfor ``SDSI'' (the Simple Distributed Security Infrastructure by\r\nLampson and Rivest [SDSI]) in 1996, although the origins clearly date\r\nback to McCarthy's LISP programming language.  It was further refined\r\nand improved during the merger of SDSI and SPKI [SPKI] during the\r\nfirst half of 1997.  S-expressions are similar to, but more readable\r\nand flexible than, Bernstein's \"net-strings\" [BERN].\r\n\r\n12. References\r\n\r\n[SDSI] \"A Simple Distributed Security Architecture\", by \r\n        Butler Lampson, and Ronald L. Rivest\r\n\thttp:\/\/theory.lcs.mit.edu\/~cis\/sdsi.html\r\n\r\n[SPKI] <a href=\"http:\/\/www.clark.net\/pub\/cme\/html\/spki.html\">SPKI--A\r\n       Simple Public Key Infrastructure<\/a>\r\n\r\n[BERN] Dan Bernstein's \"net-strings\"; Internet Draft\r\n       draft-bernstein-netstrings-02.txt\r\n\r\nAuthor's Address\r\n\r\n      Ronald L. Rivest\r\n      Room 324, 545 Technology Square\r\n      MIT Laboratory for Computer Science\r\n      Cambridge, MA 02139\r\n\r\n      rivest@theory.lcs.mit.edu\r\n\r\n\r\n<\/pre>","created_at":"2018-11-07T16:44:58+0000","updated_at":"2018-11-07T16:44:58+0000","annotations":[],"mimetype":"text\/plain","reading_time":15,"domain_name":"people.csail.mit.edu","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":337,"title":"Top 30 Game Mastering Articles","url":"https:\/\/gnomestew.com\/top-30-game-mastering-articles\/","content":"<div class=\"imgflownb\"><img src=\"https:\/\/gnomestew.com\/wp-content\/uploads\/2009\/04\/top-30.jpg\" alt=\"image\" \/><\/div><p>Welcome to Gnome Stew! If you\u2019re new here, this page will give you great introduction to what we\u2019re all about: making your time as a GM easier and more fun.<\/p><p>With more than 600 articles in our archives, we thought we\u2019d highlight the cream of the crop \u2014 the best game mastering advice, tips, ideas, tricks, and discussion on Gnome Stew, all in one spot.<\/p><p>If you like what you see here, we hope you\u2019ll <a href=\"http:\/\/www.gnomestew.com\/subscribe\">subscribe<\/a> to receive our articles for free via RSS or email (more than 3,100 readers already do), join over 1,700 members and <a href=\"http:\/\/www.gnomestew.com\/register\">register<\/a> for a free account, or <span style=\"text-decoration: line-through;\">build a creepy shrine to Gnome Stew in your basement<\/span> just come back every day to read our articles.<\/p><p>\u201c<em>I check Gnome Stew every day.<\/em>\u201d \u2014 Monte Cook<\/p><p>\u201c<em>fantastic blog for game masters, dungeon masters, and rpg fans<\/em>\u201d \u2014 Wil Wheaton (read <a href=\"http:\/\/www.gnomestew.com\/testimonials-smart-sexy-famous-people-love-gnome-stew\">more testimonials<\/a>)<\/p><h2>GMing Tricks and Tips<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/5e9005b1bc68240eac6db3b57a6896ae?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/island-design-theory\">Island Design Theory<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#johnarcadian\">John Arcadian<\/a><br \/> Have you ever prepped a ton of adventure material only to have your players ignore all of it? This flexible approach to plotting prevents that problem \u2014 and your players will love it.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/b3e34e260c51cca6671ea4e92453483a?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/restate-the-obvious\">Restate the Obvious<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#kurttelasschneider\">Kurt \u2018Telas\u2019 Schneider<\/a><br \/> What seems obvious to you, the GM, may not always be obvious to your players. This tip is incredibly simple, but has the power to completely change the course of your next session \u2014 try it and see.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/3c60956fecca50121a6302b57acfd330?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/player-characters-emerging-complexity-is-a-ok\">Player Characters: Emerging Complexity is A-OK<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#martinralya\">Martin Ralya<\/a><br \/> You don\u2019t need to force 10-page PC backstories out of your players. Instead, let their characters evolve organically during play \u2014 it can be less work and more fun, and everyone wins.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/249a68f7119b628ae4695f54b890dd04?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/tools-for-gms\/i-love-my-tiny-notebook\">I Love My Tiny Notebook<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#matthewjneagley\">Matthew J. Neagley<\/a><br \/> You carry a tiny notebook, right? If you don\u2019t, you\u2019re missing out on one of the most essential GMing tools ever created \u2014 yours for $0.99 at the office supply store.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/20c086c4f565dabfc1b8669ea3e48fb5?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/dd-burgoo-dontcha-got-a-job-or-something\">D&amp;D Burgoo (3.5): Don\u2019tcha got a job, or something?<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#troyetaylor\">Troy E. Taylor<\/a><br \/> For fantasy PCs, answering the question \u201cWhat did you do before you started adventuring?\u201d is one of the simplest tricks for building a compelling character backstory. Try it with your players.<\/p><h2>Game Mastering Techniques<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/542bbe056a4afdcd9ee2361709d7b3d4?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/introductory-games-for-new-roleplayers\">Introductory Games for New Roleplayers<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#scottmartin\">Scott Martin<\/a><br \/> When you GM a game for first-time roleplayers, there are some special considerations to take into account. It\u2019s all laid out in this informative article, from simplified character sheets to scenario structure.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/52c3b64340263edfc40d868952949f73?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/lessons-from-the-long-campaign-never-write-the-ending\">Lessons From the Long Campaign: Never Write the Ending<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#dnaphil\">DNAphil<\/a><br \/> If you\u2019re running an extended campaign, leaving individual scenes open-ended will make for a richer, more player-driven experience. Campaign structure tips galore.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/a7f330fdd078d9523ea127375453ccea?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/tools-for-gms\/preparing-to-improvise\">Preparing to Improvise<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#patrickbenson\">Patrick Benson<\/a><br \/> The best improvisation happens when you\u2019ve laid the foundation for successfully winging it in advance. Minimize your prep and maximize your fun at the gaming table with these techniques.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/058d5800431afd2e378160503130c22b?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/short-sessions\">Short Sessions<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#waltciechanowski\">Walt Ciechanowski<\/a><br \/> There\u2019s an art to running a fun short gaming session, and it\u2019s not just about doing everything faster. From game system choice to goals and rules, here are lots of tips for doing it well.<\/p><h2>Campaigns and Adventures<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/249a68f7119b628ae4695f54b890dd04?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/you-think-you-know-a-genre\">You Think You Know a Genre\u2026<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#matthewjneagley\">Matthew J. Neagley<\/a><br \/> What happens when you cross the Smurfs with zombies? The true origins of the survival horror genre, and tips for running bizarre and memorable games.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/52c3b64340263edfc40d868952949f73?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/why-you-cant-always-go-home-again\">Why You Can\u2019t Always Go Home Again<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#dnaphil\">DNAphil<\/a><br \/> Returning to a much-loved gaming system from your past can be a bittersweet experience. Here are the hallmarks of a one-campaign RPG, and how to make the most of them.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/058d5800431afd2e378160503130c22b?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/from-con-to-con-the-journey-ends\">From Con to Con: The Journey Ends<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#waltciechanowski\">Walt Ciechanowski<\/a><br \/> Planning on GMing an event at a gaming convention? These 13 tips from the trenches at GenCon will help you and your players have more fun with less stress.<\/p><h2>The Craft and Art of Game Mastering<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/a7f330fdd078d9523ea127375453ccea?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/5-mistakes-of-the-new-gm\">5 Mistakes of the New GM<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#patrickbenson\">Patrick Benson<\/a><br \/> New to game mastering? You\u2019re bound to make mistakes, just like all GMs. Here are five mistakes we all made when we started out \u2014 and how to avoid making them yourself.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/5e9005b1bc68240eac6db3b57a6896ae?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/first-time-gming-doing-a-job-that-you-dont-know-how-to-do\">First-Time GMing \u2014 Doing a Job You Don\u2019t Know How to Do<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#johnarcadian\">John Arcadian<\/a><br \/> GMing for the first time can be pretty intimidating: you feel like your entire group\u2019s fun depends on your actions, and the pressure is on. Fortunately, we\u2019re here to help.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/b3e34e260c51cca6671ea4e92453483a?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/beware-the-retcon\">Beware the Retcon<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#kurttelasschneider\">Kurt \u2018Telas\u2019 Schneider<\/a><br \/> Two practical reasons why retconning \u2014 \u201crewinding\u201d time in your game to fix a mistake \u2014 is always, always, always a bad idea. Yes, even just this one time.<\/p><h2>Players<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/542bbe056a4afdcd9ee2361709d7b3d4?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/planning-and-analysis-paralysis\">Planning and Analysis Paralysis<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#scottmartin\">Scott Martin<\/a><br \/> Building on two examples from actual play, here are the factors that contribute to analysis paralysis \u2014 aka wasting game time by talking in circles \u2014 and advice for keeping it from happening in your group.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/5e9005b1bc68240eac6db3b57a6896ae?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/player-narrative\">Player Narrative<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#johnarcadian\">John Arcadian<\/a><br \/> Knowing how and when to share narrative control \u2014 which you\u2019re already doing in every game you run, consciously or not \u2014 can get everyone at the table more involved in the game.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/20c086c4f565dabfc1b8669ea3e48fb5?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/specific-rpgs\/troys-crock-pot-a-little-thing-called-the-tpk\">Troy\u2019s Crock Pot: A Little Thing Called the TPK<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#troyetaylor\">Troy E. Taylor<\/a><br \/> At some point in your GMing career, you\u2019ll preside over a total party kill \u2014 a TPK. And despite surges and beefed-up hit points for lowbies, it can happen in D&amp;D 4th Edition.<\/p><h2>Touchy Subjects<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/249a68f7119b628ae4695f54b890dd04?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/hot-buttons\/the-concept-of-a-star-wars-rpg-should-die-in-a-fire\">The Concept of a Star Wars RPG Should Die In a Fire<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#matthewjneagley\">Matthew J. Neagley<\/a><br \/> Why is it that the people least suitable to run Star Wars games are often the ones you find behind the GM\u2019s screen? A painful truism that ignited one hell of a comment thread.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/3c60956fecca50121a6302b57acfd330?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/you-should-quit-gming-right-now\">You Should Quit GMing Right Now<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#martinralya\">Martin Ralya<\/a><br \/> The title says it all: You really should, in fact, just quit GMing right now. Ixnay on the amegay asteringmay. Cold turkey. Kaput. Done. Seriously \u2014 what are you still doing here?<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/058d5800431afd2e378160503130c22b?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/hot-buttons\/hot-button-whose-character-is-it-anyway\">Hot Button: Whose Character Is It Anyway?<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#waltciechanowski\">Walt Ciechanowski<\/a><br \/> Who has creative control over the player characters\u2019 backstories? Can you, the GM, add elements to them \u2014 or are they sacrosanct? Welcome to the world of hot button articles, and the discussions they spark.<\/p><h2>Running the Game<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/a7f330fdd078d9523ea127375453ccea?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/sometimes-to-run-a-fun-game-you-need-to-ignore-the-game\">Sometimes to Run a Fun Game You Need to Ignore the Game<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#patrickbenson\">Patrick Benson<\/a><br \/> It might sound zen, but it\u2019s also quite practical: At the end of the day, we run games so we can have fun with our friends \u2014 and if that means ignoring the rules to do what\u2019s best for the group, do it and don\u2019t look back.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/20c086c4f565dabfc1b8669ea3e48fb5?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/troys-crock-pot-oh-darn-the-npc-just-rolled-a-1\">Troy\u2019s Crock Pot: Oh Darn! The NPC just rolled a 1<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#troyetaylor\">Troy E. Taylor<\/a><br \/> The funny thing about dice is that they sometimes do exactly what you don\u2019t expect. If an NPC fails a major roll, don\u2019t automatically fudge it \u2014 instead, see where that opportunity takes you.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/542bbe056a4afdcd9ee2361709d7b3d4?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/compensating-for-failure\">Compensating for Failure<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#scottmartin\">Scott Martin<\/a><br \/> In-game failures can be planned for, mitigated with improvisation, and, on occasion, a whole lot of fun. Thinking about the meta aspects of failure can improve your game.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/15239cc615dbb28799984b64a2dd4442?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/spicing-up-combat\">Spicing Up Combat<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#adamnave\">Adam Nave<\/a><br \/> By considering focal points \u2014 the most interesting aspects of a battle \u2014 and focusing on vivid descriptions, your combats will go from humdrum to fantastic. Simple techniques, solid results.<\/p><h2>Social Issues<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/52c3b64340263edfc40d868952949f73?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/dont-fall-for-these-rpg-arguments\">Don\u2019t Fall For These RPG Arguments<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#dnaphil\">DNAphil<\/a><br \/> If you\u2019ve GMed more than once or twice, you\u2019ve heard these arguments. They derail games and reduce GMs to quivering blobs of jelly \u2014 but you can head them off at the pass.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/b3e34e260c51cca6671ea4e92453483a?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/gming-advice\/laying-the-ground-rules\">Laying the Ground Rules<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#kurttelasschneider\">Kurt \u2018Telas\u2019 Schneider<\/a><br \/> The foundation of a good social contract for gaming groups is built on establishing the basics and avoiding problems down the road \u2014 this article will show you how to do both.<\/p><h2>Game Mastering Tools<\/h2><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/15239cc615dbb28799984b64a2dd4442?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/tools-for-gms\/12-ways-to-use-google-apps-at-the-game-table\">12 Ways to Use Google Apps at the Game Table<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#adamnave\">Adam Nave<\/a><br \/> Google\u2019s free applications are a great way to organize, manage, and stay on top of your campaign \u2014 especially when you use several of them for the same game.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/addc129d28763ecd8a92c35b34a61072?s=55&amp;d=http%3A%2F%2Fhttp:\/\/www.gnomestew.com\/%2Fwp-content%2Fuploads%2F2008%2F05%2Fgnomatar1.jpg%3Fs%3D55&amp;r=PG\" width=\"60px\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/tools-for-gms\/gnomenclature-a-diminutive-rpg-glossary\">Gnomenclature: A Diminutive RPG Glossary<\/a><br \/> That title is a lie: There\u2019s nothing diminutive about this glossary \u2014 it\u2019s one of the largest on the web. If a gaming term is used by more than a handful of people, it\u2019s probably here.<\/p><div class=\"imgflow\"><img src=\"http:\/\/www.gravatar.com\/avatar\/3c60956fecca50121a6302b57acfd330?s=60&amp;d=http%3A%2F%2Fwww.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D60&amp;r=PG\" alt=\"image\" \/><\/div><p><a href=\"http:\/\/www.gnomestew.com\/reviews\/review-the-ultimate-dice-bag\">Review: The Ultimate Dice Bag<\/a>, by <a href=\"http:\/\/www.gnomestew.com\/meet-the-gnomes\/#martinralya\">Martin Ralya<\/a><br \/> The search for the best dice bag ever made is over. No, really \u2014 this bag is gorgeous, well-made, durable, and packed with features (and the right price, too). It earns the label \u201cultimate.\u201d<\/p><h2>Where To Next?<\/h2><p>If you made it all the way down here, chances are you\u2019ll enjoy reading our other articles, too. Why not <a href=\"http:\/\/www.gnomestew.com\/subscribe\">subscribe via RSS or email<\/a> or <a href=\"http:\/\/www.gnomestew.com\/register\">sign up for an account<\/a>?<\/p><p>You can also dive into the archives or browse your favorite author\u2019s articles using the drop-down menus up top, or search for articles on a specific topic \u2014 there\u2019s a lot here to explore.<\/p><p>And if you have an idea for an article on a topic we\u2019ve never covered, we\u2019d love to hear about it in the <a href=\"http:\/\/www.gnomestew.com\/suggestion-pot\">Suggestion Pot<\/a>.<\/p><div class=\"swp-content-locator\"><\/div>","created_at":"2018-11-07T16:44:58+0000","updated_at":"2018-11-07T16:44:58+0000","published_at":"2009-04-27T22:37:53+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"gnomestew.com","preview_picture":"https:\/\/gnomestew.com\/wp-content\/uploads\/2009\/04\/top-30.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":334,"title":"Introduction :: Eloquent JavaScript","url":"https:\/\/eloquentjavascript.net\/00_intro.html","content":"<blockquote>\n<p>We think we are creating the system for our own purposes. We believe we are making it in our own image... But the computer is not really like us. It is a projection of a very slim part of ourselves: that portion devoted to logic, order, rule, and clarity.<\/p>\n<footer>Ellen Ullman, <cite>Close to the Machine: Technophilia and its Discontents<\/cite><\/footer><\/blockquote><figure class=\"chapter framed\"><img src=\"https:\/\/eloquentjavascript.net\/img\/chapter_picture_00.jpg\" alt=\"Picture of a screwdriver and a circuit board\" \/><\/figure><p>This is a book about instructing computers. Computers are about as common as screwdrivers today, but they are quite a bit more complex, and making them do what you want them to do isn\u2019t always easy.<\/p><p>If the task you have for your computer is a common, well-understood one, such as showing you your email or acting like a calculator, you can open the appropriate application and get to work. But for unique or open-ended tasks, there probably is no application.<\/p><p>That is where programming may come in. <em>Programming<\/em> is the act of constructing a <em>program<\/em>\u2014a set of precise instructions telling a computer what to do. Because computers are dumb, pedantic beasts, programming is fundamentally tedious and frustrating.<\/p><p>Fortunately, if you can get over that fact, and maybe even enjoy the rigor of thinking in terms that dumb machines can deal with, programming can be rewarding. It allows you to do things in seconds that would take <em>forever<\/em> by hand. It is a way to make your computer tool do things that it couldn\u2019t do before. And it provides a wonderful exercise in abstract thinking.<\/p><p>Most programming is done with programming languages. A <em>programming language<\/em> is an artificially constructed language used to instruct computers. It is interesting that the most effective way we\u2019ve found to communicate with a computer borrows so heavily from the way we communicate with each other. Like human languages, computer languages allow words and phrases to be combined in new ways, making it possible to express ever new concepts.<\/p><p>At one point language-based interfaces, such as the BASIC and DOS prompts of the 1980s and 1990s, were the main method of interacting with computers. They have largely been replaced with visual interfaces, which are easier to learn but offer less freedom. Computer languages are still there, if you know where to look. One such language, JavaScript, is built into every modern web browser and is thus available on almost every device.<\/p><p>This book will try to make you familiar enough with this language to do useful and amusing things with it.<\/p><h2>On programming<\/h2><p>Besides explaining JavaScript, I will introduce the basic principles of programming. Programming, it turns out, is hard. The fundamental rules are simple and clear, but programs built on top of these rules tend to become complex enough to introduce their own rules and complexity. You\u2019re building your own maze, in a way, and you might just get lost in it.<\/p><p>There will be times when reading this book feels terribly frustrating. If you are new to programming, there will be a lot of new material to digest. Much of this material will then be <em>combined<\/em> in ways that require you to make additional connections.<\/p><p>It is up to you to make the necessary effort. When you are struggling to follow the book, do not jump to any conclusions about your own capabilities. You are fine\u2014you just need to keep at it. Take a break, reread some material, and make sure you read and understand the example programs and exercises. Learning is hard work, but everything you learn is yours and will make subsequent learning easier.<\/p><blockquote>\n<p>When action grows unprofitable, gather information; when information grows unprofitable, sleep.<\/p>\n<footer>Ursula K. Le Guin, <cite>The Left Hand of Darkness<\/cite><\/footer><\/blockquote><p>A program is many things. It is a piece of text typed by a programmer, it is the directing force that makes the computer do what it does, it is data in the computer\u2019s memory, yet it controls the actions performed on this same memory. Analogies that try to compare programs to objects we are familiar with tend to fall short. A superficially fitting one is that of a machine\u2014lots of separate parts tend to be involved, and to make the whole thing tick, we have to consider the ways in which these parts interconnect and contribute to the operation of the whole.<\/p><p>A computer is a physical machine that acts as a host for these immaterial machines. Computers themselves can do only stupidly straightforward things. The reason they are so useful is that they do these things at an incredibly high speed. A program can ingeniously combine an enormous number of these simple actions to do very complicated things.<\/p><p>A program is a building of thought. It is costless to build, it is weightless, and it grows easily under our typing hands.<\/p><p>But without care, a program\u2019s size and complexity will grow out of control, confusing even the person who created it. Keeping programs under control is the main problem of programming. When a program works, it is beautiful. The art of programming is the skill of controlling complexity. The great program is subdued\u2014made simple in its complexity.<\/p><p>Some programmers believe that this complexity is best managed by using only a small set of well-understood techniques in their programs. They have composed strict rules (\u201cbest practices\u201d) prescribing the form programs should have and carefully stay within their safe little zone.<\/p><p>This is not only boring, it is ineffective. New problems often require new solutions. The field of programming is young and still developing rapidly, and it is varied enough to have room for wildly different approaches. There are many terrible mistakes to make in program design, and you should go ahead and make them so that you understand them. A sense of what a good program looks like is developed in practice, not learned from a list of rules.<\/p><h2>Why language matters<\/h2><p>In the beginning, at the birth of computing, there were no programming languages. Programs looked something like this:<\/p><pre class=\"snippet cm-s-default\" data-language=\"null\">00110001 00000000 00000000\n00110001 00000001 00000001\n00110011 00000001 00000010\n01010001 00001011 00000010\n00100010 00000010 00001000\n01000011 00000001 00000000\n01000001 00000001 00000001\n00010000 00000010 00000000\n01100010 00000000 00000000<\/pre><p>That is a program to add the numbers from 1 to 10 together and print out the result: <code>1 + 2 + .<wbr \/>.<wbr \/>.<wbr \/> + 10 = 55<\/code>. It could run on a simple, hypothetical machine. To program early computers, it was necessary to set large arrays of switches in the right position or punch holes in strips of cardboard and feed them to the computer. You can probably imagine how tedious and error-prone this procedure was. Even writing simple programs required much cleverness and discipline. Complex ones were nearly inconceivable.<\/p><p>Of course, manually entering these arcane patterns of bits (the ones and zeros) did give the programmer a profound sense of being a mighty wizard. And that has to be worth something in terms of job satisfaction.<\/p><p>Each line of the previous program contains a single instruction. It could be written in English like this:<\/p><ol><li>\n<p>Store the number 0 in memory location 0.<\/p><\/li>\n<li>\n<p>Store the number 1 in memory location 1.<\/p><\/li>\n<li>\n<p>Store the value of memory location 1 in memory location 2.<\/p><\/li>\n<li>\n<p>Subtract the number 11 from the value in memory location 2.<\/p><\/li>\n<li>\n<p>If the value in memory location 2 is the number 0, continue with instruction 9.<\/p><\/li>\n<li>\n<p>Add the value of memory location 1 to memory location 0.<\/p><\/li>\n<li>\n<p>Add the number 1 to the value of memory location 1.<\/p><\/li>\n<li>\n<p>Continue with instruction 3.<\/p><\/li>\n<li>\n<p>Output the value of memory location 0.<\/p><\/li>\n<\/ol><p>Although that is already more readable than the soup of bits, it is still rather obscure. Using names instead of numbers for the instructions and memory locations helps.<\/p><pre class=\"snippet cm-s-default\" data-language=\"text\/plain\"> Set \u201ctotal\u201d to 0.\n Set \u201ccount\u201d to 1.\n[loop]\n Set \u201ccompare\u201d to \u201ccount\u201d.\n Subtract 11 from \u201ccompare\u201d.\n If \u201ccompare\u201d is zero, continue at [end].\n Add \u201ccount\u201d to \u201ctotal\u201d.\n Add 1 to \u201ccount\u201d.\n Continue at [loop].\n[end]\n Output \u201ctotal\u201d.<\/pre><p>Can you see how the program works at this point? The first two lines give two memory locations their starting values: <code>total<\/code> will be used to build up the result of the computation, and <code>count<\/code> will keep track of the number that we are currently looking at. The lines using <code>compare<\/code> are probably the weirdest ones. The program wants to see whether <code>count<\/code> is equal to 11 to decide whether it can stop running. Because our hypothetical machine is rather primitive, it can only test whether a number is zero and make a decision based on that. So it uses the memory location labeled <code>compare<\/code> to compute the value of <code>count - 11<\/code> and makes a decision based on that value. The next two lines add the value of <code>count<\/code> to the result and increment <code>count<\/code> by 1 every time the program has decided that <code>count<\/code> is not 11 yet.<\/p><p>Here is the same program in JavaScript:<\/p><pre class=\"snippet cm-s-default\" data-language=\"javascript\">let total = 0, count = 1;\nwhile (count &lt;= 10) {\n  total += count;\n  count += 1;\n}\nconsole.log(total);\n\/\/ \u2192 55<\/pre><p>This version gives us a few more improvements. Most important, there is no need to specify the way we want the program to jump back and forth anymore. The <code>while<\/code> construct takes care of that. It continues executing the block (wrapped in braces) below it as long as the condition it was given holds. That condition is <code>count &lt;= 10<\/code>, which means \u201c<em>count<\/em> is less than or equal to 10\u201d. We no longer have to create a temporary value and compare that to zero, which was just an uninteresting detail. Part of the power of programming languages is that they can take care of uninteresting details for us.<\/p><p>At the end of the program, after the <code>while<\/code> construct has finished, the <code>console.log<\/code> operation is used to write out the result.<\/p><p>Finally, here is what the program could look like if we happened to have the convenient operations <code>range<\/code> and <code>sum<\/code> available, which respectively create a collection of numbers within a range and compute the sum of a collection of numbers:<\/p><pre class=\"snippet cm-s-default\" data-language=\"javascript\">console.log(sum(range(1, 10)));\n\/\/ \u2192 55<\/pre><p>The moral of this story is that the same program can be expressed in both long and short, unreadable and readable ways. The first version of the program was extremely obscure, whereas this last one is almost English: <code>log<\/code> the <code>sum<\/code> of the <code>range<\/code> of numbers from 1 to 10. (We will see in <a href=\"https:\/\/eloquentjavascript.net\/04_data.html\">later chapters<\/a> how to define operations like <code>sum<\/code> and <code>range<\/code>.)<\/p><p>A good programming language helps the programmer by allowing them to talk about the actions that the computer has to perform on a higher level. It helps omit details, provides convenient building blocks (such as <code>while<\/code> and <code>console.log<\/code>), allows you to define your own building blocks (such as <code>sum<\/code> and <code>range<\/code>), and makes those blocks easy to compose.<\/p><h2>What is JavaScript?<\/h2><p>JavaScript was introduced in 1995 as a way to add programs to web pages in the Netscape Navigator browser. The language has since been adopted by all other major graphical web browsers. It has made modern web applications possible\u2014applications with which you can interact directly without doing a page reload for every action. JavaScript is also used in more traditional websites to provide various forms of interactivity and cleverness.<\/p><p>It is important to note that JavaScript has almost nothing to do with the programming language named Java. The similar name was inspired by marketing considerations rather than good judgment. When JavaScript was being introduced, the Java language was being heavily marketed and was gaining popularity. Someone thought it was a good idea to try to ride along on this success. Now we are stuck with the name.<\/p><p>After its adoption outside of Netscape, a standard document was written to describe the way the JavaScript language should work so that the various pieces of software that claimed to support JavaScript were actually talking about the same language. This is called the ECMAScript standard, after the Ecma International organization that did the standardization. In practice, the terms ECMAScript and JavaScript can be used interchangeably\u2014they are two names for the same language.<\/p><p>There are those who will say <em>terrible<\/em> things about JavaScript. Many of these things are true. When I was required to write something in JavaScript for the first time, I quickly came to despise it. It would accept almost anything I typed but interpret it in a way that was completely different from what I meant. This had a lot to do with the fact that I did not have a clue what I was doing, of course, but there is a real issue here: JavaScript is ridiculously liberal in what it allows. The idea behind this design was that it would make programming in JavaScript easier for beginners. In actuality, it mostly makes finding problems in your programs harder because the system will not point them out to you.<\/p><p>This flexibility also has its advantages, though. It leaves space for a lot of techniques that are impossible in more rigid languages, and as you will see (for example in <a href=\"https:\/\/eloquentjavascript.net\/10_modules.html\">Chapter 10<\/a>), it can be used to overcome some of JavaScript\u2019s shortcomings. After learning the language properly and working with it for a while, I have learned to actually <em>like<\/em> JavaScript.<\/p><p>There have been several versions of JavaScript. ECMAScript version 3 was the widely supported version in the time of JavaScript\u2019s ascent to dominance, roughly between 2000 and 2010. During this time, work was underway on an ambitious version 4, which planned a number of radical improvements and extensions to the language. Changing a living, widely used language in such a radical way turned out to be politically difficult, and work on the version 4 was abandoned in 2008, leading to a much less ambitious version 5, which made only some uncontroversial improvements, coming out in 2009. Then in 2015 version 6 came out, a major update that included some of the ideas planned for version 4. Since then we\u2019ve had new, small updates every year.<\/p><p>The fact that the language is evolving means that browsers have to constantly keep up, and if you\u2019re using an older browser, it may not support every feature. The language designers are careful to not make any changes that could break existing programs, so new browsers can still run old programs. In this book, I\u2019m using the 2017 version of JavaScript.<\/p><p>Web browsers are not the only platforms on which JavaScript is used. Some databases, such as MongoDB and CouchDB, use JavaScript as their scripting and query language. Several platforms for desktop and server programming, most notably the Node.js project (the subject of <a href=\"https:\/\/eloquentjavascript.net\/20_node.html\">Chapter 20<\/a>), provide an environment for programming JavaScript outside of the browser.<\/p><h2>Code, and what to do with it<\/h2><p><em>Code<\/em> is the text that makes up programs. Most chapters in this book contain quite a lot of code. I believe reading code and writing code are indispensable parts of learning to program. Try to not just glance over the examples\u2014read them attentively and understand them. This may be slow and confusing at first, but I promise that you\u2019ll quickly get the hang of it. The same goes for the exercises. Don\u2019t assume you understand them until you\u2019ve actually written a working solution.<\/p><p>I recommend you try your solutions to exercises in an actual JavaScript interpreter. That way, you\u2019ll get immediate feedback on whether what you are doing is working, and, I hope, you\u2019ll be tempted to experiment and go beyond the exercises.<\/p><p>When reading this book in your browser, you can edit (and run) all example programs by clicking them.<\/p><p>If you want to run the programs defined in this book outside of the book\u2019s website, some care will be required. Many examples stand on their own and should work in any JavaScript environment. But code in later chapters is often written for a specific environment (the browser or Node.js) and can run only there. In addition, many chapters define bigger programs, and the pieces of code that appear in them depend on each other or on external files. The <a href=\"https:\/\/eloquentjavascript.net\/code\">sandbox<\/a> on the website provides links to Zip files containing all the scripts and data files necessary to run the code for a given chapter.<\/p><h2>Overview of this book<\/h2><p>This book contains roughly three parts. The first 12 chapters discuss the JavaScript language. The next seven chapters are about web browsers and the way JavaScript is used to program them. Finally, two chapters are devoted to Node.js, another environment to program JavaScript in.<\/p><p>Throughout the book, there are five <em>project chapters<\/em>, which describe larger example programs to give you a taste of actual programming. In order of appearance, we will work through building a <a href=\"https:\/\/eloquentjavascript.net\/07_robot.html\">delivery robot<\/a>, a <a href=\"https:\/\/eloquentjavascript.net\/12_language.html\">programming language<\/a>, a <a href=\"https:\/\/eloquentjavascript.net\/16_game.html\">platform game<\/a>, a <a href=\"https:\/\/eloquentjavascript.net\/19_paint.html\">pixel paint program<\/a>, and a <a href=\"https:\/\/eloquentjavascript.net\/21_skillsharing.html\">dynamic website<\/a>.<\/p><p>The language part of the book starts with four chapters that introduce the basic structure of the JavaScript language. They introduce <a href=\"https:\/\/eloquentjavascript.net\/02_program_structure.html\">control structures<\/a> (such as the <code>while<\/code> word you saw in this introduction), <a href=\"https:\/\/eloquentjavascript.net\/03_functions.html\">functions<\/a> (writing your own building blocks), and <a href=\"https:\/\/eloquentjavascript.net\/04_data.html\">data structures<\/a>. After these, you will be able to write basic programs. Next, Chapters <a href=\"https:\/\/eloquentjavascript.net\/05_higher_order.html\">5<\/a> and <a href=\"https:\/\/eloquentjavascript.net\/06_object.html\">6<\/a> introduce techniques to use functions and objects to write more <em>abstract<\/em> code and keep complexity under control.<\/p><p>After a <a href=\"https:\/\/eloquentjavascript.net\/07_robot.html\">first project chapter<\/a>, the language part of the book continues with chapters on <a href=\"https:\/\/eloquentjavascript.net\/08_error.html\">error handling and bug fixing<\/a>, <a href=\"https:\/\/eloquentjavascript.net\/09_regexp.html\">regular expressions<\/a> (an important tool for working with text), <a href=\"https:\/\/eloquentjavascript.net\/10_modules.html\">modularity<\/a> (another defense against complexity), and <a href=\"https:\/\/eloquentjavascript.net\/11_async.html\">asynchronous programming<\/a> (dealing with events that take time). The <a href=\"https:\/\/eloquentjavascript.net\/12_language.html\">second project chapter<\/a> concludes the first part of the book.<\/p><p>The second part, Chapters <a href=\"https:\/\/eloquentjavascript.net\/13_browser.html\">13<\/a> to <a href=\"https:\/\/eloquentjavascript.net\/19_paint.html\">19<\/a>, describes the tools that browser JavaScript has access to. You\u2019ll learn to display things on the screen (Chapters <a href=\"https:\/\/eloquentjavascript.net\/14_dom.html\">14<\/a> and <a href=\"https:\/\/eloquentjavascript.net\/17_canvas.html\">17<\/a>), respond to user input (<a href=\"https:\/\/eloquentjavascript.net\/15_event.html\">Chapter 15<\/a>), and communicate over the network (<a href=\"https:\/\/eloquentjavascript.net\/18_http.html\">Chapter 18<\/a>). There are again two project chapters in this part.<\/p><p>After that, <a href=\"https:\/\/eloquentjavascript.net\/20_node.html\">Chapter 20<\/a> describes Node.js, and <a href=\"https:\/\/eloquentjavascript.net\/21_skillsharing.html\">Chapter 21<\/a> builds a small website using that tool.<\/p><h2>Typographic conventions<\/h2><p>In this book, text written in a <code>monospaced<\/code> font will represent elements of programs\u2014sometimes they are self-sufficient fragments, and sometimes they just refer to part of a nearby program. Programs (of which you have already seen a few) are written as follows:<\/p><pre class=\"snippet cm-s-default\" data-language=\"javascript\">function factorial(n) {\n  if (n == 0) {\n    return 1;\n  } else {\n    return factorial(n - 1) * n;\n  }\n}<\/pre><p>Sometimes, to show the output that a program produces, the expected output is written after it, with two slashes and an arrow in front.<\/p><pre class=\"snippet cm-s-default\" data-language=\"javascript\">console.log(factorial(8));\n\/\/ \u2192 40320<\/pre><p>Good luck!<\/p>","created_at":"2018-11-07T16:44:57+0000","updated_at":"2018-11-07T16:44:57+0000","annotations":[],"mimetype":"text\/html","reading_time":15,"domain_name":"eloquentjavascript.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":333,"title":"Mastering Eshell - Mastering Emacs","url":"https:\/\/masteringemacs.org\/article\/complete-guide-mastering-eshell","content":"<p>By <em>Mickey Petersen<\/em><\/p><img src=\"https:\/\/masteringemacs.org\/static\/img\/fleuron2.gif\" class=\"img-responsive center-block\" alt=\"image\" \/><p class=\"text-justify drop-cap\">There are several shells for Emacs, but none can match the versatility and integration with Emacs like Eshell. Eshell is a shell written entirely in Emacs-Lisp, and it replicates most of the features and commands from GNU CoreUtils and the Bourne-like shells. So by re-writing common commands like <code>ls<\/code> and <code>cp<\/code> in Emacs-Lisp, Eshell will function identically on any environment Emacs itself runs on.<\/p><p class=\"text-justify\">Unfortunately, there is a problem: Eshell is woefully underdocumented \u2013 a rare sight in GNU Emacs \u2013 so I\u2019ve compiled this guide to help people make full use of what Eshell has to offer.<\/p><h2 id=\"overview\">Overview<\/h2><p class=\"text-justify\">Unlike the other shells in Emacs, Eshell does not inherit from <em>comint-mode<\/em>, the default mode for interacting with inferior processes in Emacs. But because Eshell is not an inferior process, it does not have to use comint; but while that may seem like a good thing, it does mean that hooks and routines written for <em>comint-mode<\/em> won\u2019t work with Eshell.<\/p><p class=\"text-justify\">However, almost all the Emacs commands common to <em>comint-mode<\/em> are reimplemented natively in Eshell \u2013 and most share the same keybinds \u2013 but there are a few new advances that haven\u2019t been ported over to Eshell, like the spiffy <code>comint-history-isearch-backward-regexp<\/code> in Emacs 23.2, bound to <code>M-r<\/code>.<\/p><p class=\"text-justify\">Eshell works well on any platform Emacs itself runs on, as Eshell interacts with a common middleware (namely the Emacs-Lisp\/C source library) and that middleware will in turn communicate with your OS on how to go about copying files and what have you. That middleware support enables Eshell to take advantage of <a href=\"http:\/\/www.gnu.org\/software\/tramp\/\" class=\"article-link\">TRAMP<\/a> as well.<\/p><p class=\"text-justify\">Given Emacs\u2019 UNIX origin, Eshell emulates traditional UNIX shells like <em>bash<\/em> and the GNU toolchain. This is good news if you are using Windows and cannot be bothered fidgeting with cygwin, or if you require a completely portable Emacs with few or no external dependencies.<\/p><p class=\"text-justify\">And actually, the Windows support in Eshell is a lot better, in many ways, than cygwin\u2019s bash. You do not have the <code>\/cygdrive\/c<\/code> crud to contend with, as Eshell natively supports Windows\/MS-DOS drive paths (so <code>cd D:<\/code> and <code>D:<\/code> both work equally well.)<\/p><p class=\"text-justify\">Despite all the advantages offered by Eshell, there are some points I want to make that seem to confuse some people:<\/p><ol><li>Eshell is <em>not<\/em> a terminal emulator. It does not talk to a shell, for it <em>is<\/em> the shell. Everything it does \u2013 from displaying stuff on the screen, to fetching the contents of a directory \u2013 it does through Emacs, and Emacs in turn talks to your operating system.<\/li><li>Because of the way Eshell talks to other processes (asynchronous ones especially) there may be issues with the way it buffers text and how interrupts work.<\/li><li>Eshell does not support interactive (or \u201cvisual\u201d in Eshell parlance) programs, like <em>top<\/em>, directly; you must tell Eshell to launch them in a separate <code>ansi-term<\/code> instance instead.<\/li><li>It is not <em>bash<\/em> or <em>zsh<\/em> or even <em>csh<\/em>; do not treat it as such, even though it is heavily inspired by them. To use Eshell effectively you should treat it as if you are using a completely alien shell.<\/li><\/ol><h2 id=\"commands\">Commands<\/h2><p class=\"text-justify\">Eshell is capable of invoking almost any elisp function loaded in Emacs. That sort of flexibility is unmatched; there are no shells out there capable of approximating what Eshell can do. In fact, this functionality is heavily used (and encouraged!) by Eshell. If you want to open the file <code>foobar.txt<\/code> in Emacs you simply invoke <code>find-file foobar.txt<\/code> and Eshell will map that to the elisp call <code>(find-file \"foobar.txt\")<\/code> and open the file for you.<\/p><h3 id=\"technical-details\">Technical Details<\/h3><p class=\"text-justify\">All commands evaluated by Eshell have an <em>evaluation order<\/em>, which is an ordered list your command must pass through to determine what part of Eshell handles it. If there is nothing on the list that wants to evaluate your command, you will be told your command is invalid.<\/p><p class=\"text-justify\">Assuming you want to execute the command <code>cp<\/code>, the evaluation order is:<\/p><ol><li>A full filepath (e.g. <code>\/bin\/cp<\/code>) runs <code>cp<\/code> in <code>\/bin<\/code><\/li><li>Look for the command prefix, <code>eshell-explicit-command-char<\/code> (default is <code>*<\/code>), and if it is found then look for the command in the search path.<\/li><li>Look for a shell-defined alias (<code>alias<\/code> command)<\/li><li>Look for <code>cp<\/code> in the search path, <code>$PATH<\/code> (or <code>eshell-path-env<\/code>)<\/li><li>Look for a Lisp function named <code>cp<\/code> or the elisp function <code>eshell\/cp<\/code><\/li><\/ol><p class=\"text-justify\">The variable <code>eshell-prefer-lisp-functions<\/code> makes internal elisp calls take priority over external calls. What that means is when it\u2019s set to <code>t<\/code> Eshell will look for an elisp function <em>first<\/em>, instead of <em>last<\/em>. If the command prefix is specified, though, this directive is ignored.<\/p><h3 id=\"built-in-commands\">Built-In Commands<\/h3><p class=\"text-justify\">Eshell has a handful of commands written in Emacs-Lisp that closely emulate a large subset of what the real GNU Coreutils (or your favorite shell) has. Those commands are called \u201cAlias functions.\u201d<\/p><p class=\"text-justify\">EShell only implements a subset of the functionality provided by the real commands, but if you pass an unknown argument to Eshell it will defer to the <em>real<\/em> commandline tool (if it is installed) automatically.<\/p><p class=\"text-justify\">Here\u2019s what Eshell currently re-implements in elisp:<\/p><p class=\"text-justify\"><code>cat<\/code>, <code>cp<\/code>, <code>ls<\/code>, <code>cd<\/code>, <code>export<\/code>, <code>dirs<\/code>, <code>du<\/code>, <code>echo<\/code>, <code>env<\/code>, <code>kill<\/code>, <code>ln<\/code>, <code>mkdir<\/code>, <code>mv<\/code>, <code>alias<\/code>, <code>popd<\/code>, <code>pushd<\/code>, <code>pwd<\/code>, <code>rm<\/code>, <code>rmdir<\/code>, <code>time<\/code>, <code>umask<\/code>.<\/p><p class=\"text-justify\">There is a big emphasis on adhering to the original GNU functionality, so the fact they are emulated is unlikely to cause you any trouble.<\/p><h3 id=\"command-interception\">Command Interception<\/h3><p class=\"text-justify\">Eshell has a cool mechanism where certain commands are <em>intercepted<\/em> and passed on to Emacs proper. This enables you to invoke a command like <code>man ls<\/code> and have Emacs\u2019s built-in <code>man<\/code> formatter handle it instead. This functionality is especially important for interactive commands (as they will not work properly in Eshell) as Eshell would not be able to call them otherwise.<\/p><p class=\"text-justify\">But where the feature really shines is with complex commands like <code>grep<\/code>, or <code>diff<\/code> as Emacs comes with awesome grep and diff tools built in. This feature alone shows the power of Eshell.<\/p><p class=\"text-justify\">The following commands are redirected to Emacs proper:<\/p><p class=\"text-justify\"><code>agrep<\/code>, <code>diff<\/code>, <code>egrep<\/code>, <code>fgrep<\/code>, <code>glimpse<\/code>, <code>grep<\/code>, <code>info<\/code>, <code>jobs<\/code>, <code>locate<\/code>, <code>man<\/code>, <code>occur<\/code>, <code>su<\/code>, <code>sudo<\/code>, <code>whoami<\/code>.<\/p><p class=\"text-justify\">The commands <code>su<\/code>, <code>sudo<\/code> and <code>whoami<\/code> are TRAMP aware commands (in Emacs 23.2), so if you are connected to a remote shell they work as expected.<\/p><h3 id=\"subshells\">Subshells<\/h3><p class=\"text-justify\">You can use <code>$()<\/code> to in-line elisp calls and use their output as arguments, in much the same way as you would in bash. The only caveat here is you cannot use the backquote (backtick) to spawn a subshell, but that syntax was never universally supported anyway. It\u2019s also possible (though I would not recommend it, for there are cases where it does not work) to use a standard elisp form like this: <code>(form ...)<\/code> \u2013 so the same as the subshell syntax I explained before, but without the <code>$<\/code>.<\/p><h3 id=\"useful-elisp-commands\">Useful Elisp Commands<\/h3><p class=\"text-justify\">Eshell comes with a selection of helper functions that make your day-to-day life just **. That, combined with the power to invoke almost any elisp function, means you have incredible flexibility and control over your shell. Some of the commands I\u2019ve listed in the table below were written for Eshell specifically, and the rest are elisp commands I find useful.<\/p><p class=\"text-justify\">I\u2019ve compiled a table of elisp functions (some are made for Eshell; others are not.)<\/p><dl><dt><code>listify ARGS<\/code><\/dt><dd><p class=\"text-justify\">Parses an argument string into elisp list notation and prints it to the screen. It\u2019s clever enough to handle both MS-DOS\/Windows and POSIX-style argument syntax.<\/p><\/dd><dt><code>addpath PATH<\/code><\/dt><dd><p class=\"text-justify\">Adds the argument, which must be a path, to the <code>$PATH<\/code> environment variable. If no argument is specified the existing paths are pretty-printed to the screen.<\/p><\/dd><dt><code>unset ENV-VAR<\/code><\/dt><dd><p class=\"text-justify\">Unsets an existing environment variable<\/p><\/dd><dt><code>find-file FILE<\/code><\/dt><dd><p class=\"text-justify\">Finds the file FILE and opens it in Emacs. This function is TRAMP aware and will therefore work remotely.<\/p><\/dd><dt><code>dired DIRECTORY<\/code><\/dt><dd><p class=\"text-justify\">Opens a dired buffer in DIRECTORY.<\/p><\/dd><dt><code>calc-eval EXPR<\/code><\/dt><dd><p class=\"text-justify\">Runs EXPR through the Emacs calculator.<\/p><\/dd><dt><code>upcase STR<\/code>\/<code>downcase STR<\/code><\/dt><dd><p class=\"text-justify\">Converts STR to upper- or lowercase.<\/p><\/dd><dt><code>vc-dir DIRECTORY<\/code><\/dt><dd><p class=\"text-justify\">Reports the status of a version controlled directory (equivalent to the <code>status<\/code> command in most VCS)<\/p><\/dd><dt><code>ediff-files FILE1 FILE2<\/code><\/dt><dd><p class=\"text-justify\">Diffs FILE1 and FILE2 using ediff, Emacs\u2019 diff engine.<\/p><\/dd><\/dl><p class=\"text-justify\">If you\u2019re an Eshell user and you use elisp commands not listed in the table above, post a comment and let tell me what it is.<\/p><h3 id=\"aliasing\">Aliasing<\/h3><p class=\"text-justify\">Aliasing in Eshell works in much the same way as it does in other mainstream shells, except you can freely mix elisp and Eshell commands. The command <code>alias<\/code> takes an <code>alias-name<\/code> and a <code>definition<\/code>. The <code>definition<\/code> must be surrounded by <em>single quotes<\/em>. You can use the usual argument references known from other shells: <code>$1<\/code> for the first argument, <code>$2<\/code> for the second, \u2026, or <code>$*<\/code> to use all arguments, or omit them entirely as Eshell will magically append them on to the end of a command if they weren\u2019t referenced in the definition.<\/p><p class=\"text-justify\">To delete an alias, simply leave out the <code>definition<\/code> argument and it will be removed automagically. To list all the aliases, leave out both arguments.<\/p><p class=\"text-justify\">Eshell will write the alias definitions to <code>eshell-aliases-file<\/code>, which in turn is governed by the <code>Eshell-directory-name<\/code> and that put together means your alias file will be put in <code>~\/.Eshell\/alias<\/code> by default. This is done every time you alter an alias.<\/p><p class=\"text-justify\">Another useful thing to know is the <em>auto-correcting aliasing<\/em>. If you type an invalid command too many times (governed by <code>eshell-bad-command-tolerance<\/code>, which is 3 by default) Eshell will offer to alias it to its intended command for you. If you don\u2019t like that, you can bump up the aforementioned variable to a large number.<\/p><h4 id=\"useful-examples\">Useful Examples<\/h4><p class=\"text-justify\">Let\u2019s map the cumbersome command <code>find-file<\/code> to the more manageable <code>ff<\/code>:<\/p><pre class=\"sourceCode commonlisp\">alias ff 'find-file $1'<\/pre><p class=\"text-justify\">And let\u2019s map <code>dired<\/code> to <code>d<\/code>:<\/p><pre class=\"sourceCode commonlisp\">alias d 'dired $1'<\/pre><h3 id=\"visual-commands\">Visual Commands<\/h3><p class=\"text-justify\">Some commands are too complex to be displayed by Eshell directly, and require special handling. An example would be <code>top<\/code>, a program that won\u2019t work with a dumb terminal. To support these commands Eshell will run a <code>term<\/code> session when you invoke a command Eshell considers visual.<\/p><p class=\"text-justify\">To modify the list of visual commands, you can alter <code>eshell-visual-commands<\/code>.<\/p><h2 id=\"command-history\">Command History<\/h2><p class=\"text-justify\">Eshell comes with a feature-rich command history facility. Because Eshell does not use <code>comint-mode<\/code> it does not have <em>all<\/em> the history features available to it, but most of them are reimplemented.<\/p><dl><dt><code>M-r \/ M-s<\/code><\/dt><dd><p class=\"text-justify\">Search backwards or forwards for a command by regexp<\/p><\/dd><dt><code>M-p \/ M-n<\/code><\/dt><dd><p class=\"text-justify\">Goes backwards or forwards in the command history list<\/p><\/dd><dt><code>C-p \/ C-n<\/code><\/dt><dd><p class=\"text-justify\">Jump to the previous or next command position in Eshell<\/p><\/dd><dt><code>C-c M-r \/ C-c M-s<\/code><\/dt><dd><p class=\"text-justify\">Jumps to the previous or next command that shares the command currently used as input. So it jumps to other instances of the command <code>foo<\/code> if that is the current input.<\/p><\/dd><\/dl><p class=\"text-justify\">Unfortunately, the new-and-improved <code>comint-history-isearch-backward-regexp<\/code> (bound to <code>M-r<\/code> in comint) doesn\u2019t work in Eshell because it not inherit from <code>comint<\/code> (and therefore misses out on upgrades.).<\/p><h3 id=\"history-interaction\">History Interaction<\/h3><p class=\"text-justify\">Like bash and other shells, Eshell has support for history modification and interaction. It\u2019s probably easier to refer you to the bash info manual for detailed information on how the history interaction works. I\u2019ve included a small table below that describes most of the history syntax Eshell supports.<\/p><dl><dt><code>!!<\/code><\/dt><dd><p class=\"text-justify\">Repeats the last command<\/p><\/dd><dt><code>!ls<\/code><\/dt><dd><p class=\"text-justify\">Repeats the last command <strong>beginning<\/strong> with <code>ls<\/code><\/p><\/dd><dt><code>!?ls<\/code><\/dt><dd><p class=\"text-justify\">Repeats the last command <strong>containing<\/strong> <code>ls<\/code><\/p><\/dd><dt><code>!ls:n<\/code><\/dt><dd><p class=\"text-justify\">Extract the <em>nth<\/em> argument from the last command <strong>beginning<\/strong> with <code>ls<\/code><\/p><\/dd><dt><code>!ls&lt;tab&gt;<\/code><\/dt><dd><p class=\"text-justify\">Using <em>pcomplete<\/em>, show completion results matches <code>ls<\/code><\/p><\/dd><dt><code>^old^new<\/code><\/dt><dd><p class=\"text-justify\">Quick substitution. Using the last command, replace<code>old<\/code> with <code>new<\/code> and run it again. Appears to be buggy.<\/p><\/dd><dt><code>$_<\/code><\/dt><dd><p class=\"text-justify\">Returns the last parameter in the last executed command.<\/p><\/dd><\/dl><p class=\"text-justify\">Eshell also has some support for bash history modifiers (like <code>!!:s\/old\/new\/<\/code>) and <a href=\"http:\/\/www.gnu.org\/software\/bash\/manual\/bash.html#History-Interaction\" class=\"article-link\">the bash reference on history interaction<\/a> would be a good place to brush up on that.<\/p><h2 id=\"commandline-interaction\">Commandline Interaction<\/h2><h3 id=\"the-eshell-prompt\">The Eshell Prompt<\/h3><p class=\"text-justify\">You can customize the Eshell prompt by modifying <code>eshell-prompt-function<\/code>, a variable that takes a function that defines what the prompt should contain. By relegating prompt configuration to elisp you can do just about anything you like with it. The only problem is, of course, that Eshell will need to be told <em>what<\/em> the prompt \u201clooks\u201d like, so you must also edit the variable <code>eshell-prompt-regexp<\/code> so Eshell knows what the prompt is.<\/p><h3 id=\"the-commandline\">The Commandline<\/h3><p class=\"text-justify\">Eshell supports <code>\\<\/code> to escape newlines and supports rudimentary multi-line input that way. Another way of doing multi-line <em>literal strings<\/em> is with single quotes: begin a single quote and hit enter, and you are free to enter text until the closing quote delimiter is encountered. If you use double quotes Eshell will expand subshell commands and do variable expansion.<\/p><p class=\"text-justify\">Due to the way Eshell works, you can even go back and modify the text you entered, in quotes. This is very handy as you can go back and change stuff you don\u2019t like, and get it right the first time.<\/p><h3 id=\"useful-keybindings\">Useful Keybindings<\/h3><p class=\"text-justify\">Eshell comes equipped with a couple of quality-of-life improvements that make interacting with Emacs and Eshell a lot easier.<\/p><dl><dt><code>C-c M-b<\/code><\/dt><dd><p class=\"text-justify\">Inserts the printed buffer name at point<\/p><\/dd><dt><code>C-c M-i<\/code><\/dt><dd><p class=\"text-justify\">Inserts the printed process name at point<\/p><\/dd><dt><code>C-c M-v<\/code><\/dt><dd><p class=\"text-justify\">Inserts an environment variable name at point<\/p><\/dd><dt><code>C-c M-d<\/code><\/dt><dd><p class=\"text-justify\">Toggles between direct input and delayed input (send on RET).<\/p><\/dd><\/dl><p class=\"text-justify\">Useful for some programs that don\u2019t work correctly with buffered input.<\/p><h2 id=\"argument-predicates\">Argument Predicates<\/h2><p class=\"text-justify\">Argument predicates are a cool way of quickly filtering lists of files or even elisp lists. The predicate syntax is based on the one used in zsh, so if you are familiar with argument predication in zsh, you can apply most of your knowledge to Eshells\u2019 version.<\/p><p class=\"text-justify\">Unlike most other areas of Eshell, argument predicates are documented in Eshell itself. You can access the help files by typing <code>eshell-display-predicate-help<\/code> or <code>eshell-display-modifier-help<\/code>.<\/p><p class=\"text-justify\">Filtering globbed lists of files is very useful, as it saves you the hassle of using tools like <code>find<\/code> or abusing <code>ls<\/code> to do your thing.<\/p><p class=\"text-justify\">The help file is fairly spartan and only serves as a simple reference, so I\u2019ve included a small guide here; but actually, the only real way to learn something as flexible as argument predication is simply by trial and error.<\/p><h3 id=\"syntax-reference\">Syntax Reference<\/h3><p class=\"text-justify\">I\u2019ve opted not to reprint the sizeable list of predicates and modifiers, as the Eshell manual (see the commands above) do a good enough job of explaining how they work.<\/p><h3 id=\"globbing\">Globbing<\/h3><p class=\"text-justify\">Globbing in Eshell follow the same rules as it does in most other common shells: it is the shell that does the expansion of globs and <em>it<\/em> passes the expanded list of matches on to commands like <code>ls<\/code>. That\u2019s why when you use <code>find<\/code> and <code>xargs<\/code> together it\u2019s critical that you pass <code>-print0<\/code> to <code>find<\/code>, and <code>-0<\/code> to <code>xargs<\/code>. If you don\u2019t, filenames with obscure characters or spaces in them may trip up xargs; by using the NUL character as a separator ensures tokenization takes place correctly as the NUL character is an invalid character (along with <code>\/<\/code>) in files.<\/p><h3 id=\"elisp-lists\">Elisp Lists<\/h3><p class=\"text-justify\">Eshell\u2019s \u201clists\u201d are actually elisp lists in their printed form as well as internally. That makes life a lot simpler if you think about it, as Eshell can paw off list handling to elisp, which is something Lisp does well.<\/p><p class=\"text-justify\">Simplest glob example is <code>echo *<\/code>, which echos a list of all the wildcard matches in the current directory. Because \u2013 as I just mentioned above \u2013 wildcard expansion takes place <em>inline<\/em>, I can immediately apply a modifier to the <code>*<\/code> wildcard above.<\/p><p class=\"text-justify\">Let\u2019s uppercase the globbed result set:<\/p><pre>\/ $ echo *(:U)\n(\"BAR\" \"BIN\/\" \"DEV\/\" \"ETC\/\" \"FOO\" \"HOME\/\" \"LIB\/\" \"TMP\/\" \"USR\/\" \"VAR\/\")<\/pre><p class=\"text-justify\">Notice how I used <code>()<\/code> immediately following the glob pattern. The brackets are what makes argument modifiers or predicates possible. Modifiers are things that <em>modify<\/em> (big surprise!) the resulting list. Modifier commands always begin with <code>:<\/code>, and predicates do not.<\/p><p class=\"text-justify\">Another example, but this time I filter directories using a predicate:<\/p><pre>\/ $ echo *(^\/)\n(\"bar\" \"foo\")<\/pre><p class=\"text-justify\">The circumflex, <code>^<\/code>, in this case, like in regular expressions, is negation. The <code>\/<\/code> means \u201cdirectories\u201d only.<\/p><p class=\"text-justify\">But I don\u2019t have to use globs to apply modifiers or predicates to lists:<\/p><pre>\/ $ echo (\"foo\" \"bar\" \"baz\" \"foo\")(:gs\/foo\/blarg\/)\n(\"blarg\" \"bar\" \"baz\" \"blarg\")<\/pre><p class=\"text-justify\">This time I replaced all occurrences of <em>foo<\/em> with <em>blarg<\/em>. Observe that the syntax is identical, except instead of using globs to get a list of files, I used a list of my own choosing.<\/p><p class=\"text-justify\">The advantages provided by argument predicates and modifiers will greatly reduce commandline clutter as the predicates cover permissions, ownership, file attributes, and much more.<\/p><h3 id=\"adding-new-modifiers-and-predicates\">Adding New Modifiers and Predicates<\/h3><p class=\"text-justify\">You can even add your own predicates (<code>eshell-predicate-alist<\/code>) or modifiers (<code>eshell-modifier-alist<\/code>):<\/p><pre class=\"sourceCode commonlisp\">(add-to-list 'eshell-modifier-alist '(?X . '(lambda (lst) (mapcar 'rot13 lst))))<\/pre><p class=\"text-justify\">Here I\u2019ve bound <code>X<\/code> to <code>rot13<\/code>, the substitution cipher:<\/p><pre>\/ $ echo (\"foo\" \"bar\" \"baz\")(:X)\n(\"sbb\" \"one\" \"onm\")<\/pre><h2 id=\"plan-9-smart-shell\">Plan 9 Smart Shell<\/h2><p class=\"text-justify\">Eshell comes with a pared-down facsimile of Plan 9\u2019s terminal, called <em>the Eshell smart display<\/em>. The smart display is meant to improve the write-run-revise cycle all commandline hackers go through. It works by not letting the point follow the output of a command you execute, like a normal terminal would. Instead, the point is kept on the line of the command you executed, letting you revise it easily without having to use <code>M-p<\/code> and <code>M-n<\/code> or the history modification commands.<\/p><p class=\"text-justify\">If smart display is enabled it will also let you review the output of long-running commands by using <code>SPC<\/code> to move down a page and <code>BACKSPACE<\/code> to move up a page. If any other key is pressed it will jump the end of the buffer, essentially acting in the same way as if smart display wasn\u2019t enabled.<\/p><p class=\"text-justify\">Essentially, if Eshell detects that you want to <em>review<\/em> the last executed command, it will help you do so; if, on the other hand, you do not then Eshell will jump to the end of the buffer instead. It\u2019s pretty clever about it, and there are switches you can toggle to fine-tune the behavior.<\/p><p class=\"text-justify\">Where the smart display really shines is that it lets you modify the command you just executed by using the movement keys \u2013 like you normally would \u2013 to change the command, say to fix a typo or tweak an argument.<\/p><p class=\"text-justify\">The smart display can also be set not to use this extended \u201cedit mode\u201d if the command returns successfully, and without displaying output, like <code>chown<\/code> for instance. This is how I prefer it.<\/p><p class=\"text-justify\">To enable it put this in your .emacs file:<\/p><pre class=\"sourceCode commonlisp\">(require 'eshell)\n(require 'em-smart)\n(setq eshell-where-to-jump 'begin)\n(setq eshell-review-quick-commands nil)\n(setq eshell-smart-space-goes-to-end t)<\/pre><p class=\"text-justify\">If Eshell has already initialized (that is, you\u2019ve already launched an instance of Eshell in Emacs) then evaluating the changes above will not work. You must switch to the Eshell buffer and type <code>M-: (shell-smart-initialize)<\/code> (or restart Emacs.)<\/p><p class=\"text-justify\">The smart display is a pretty useful feature and it won\u2019t get in your way once you\u2019re used to it. Simply typing in new commands will make Eshell jump to the end of buffer as if the point was already there.<\/p><h2 id=\"redirection\">Redirection<\/h2><p class=\"text-justify\">Redirection in Eshell works in much the same way as it does in other shells. The key difference is that Eshell has to emulate the pseudo-devices as they may not be present (or may not be present in the same form) on platforms such as Windows where <code>\/dev\/null<\/code> is actually <code>NUL<\/code>.<\/p><p class=\"text-justify\">Another caveat is that Eshell does not support <em>input<\/em> redirection, though it does support output redirection. To skirt around the lack of input redirection you should use pipes instead.<\/p><p class=\"text-justify\">Redirection to stdout, stdin and stderr work as you would expect, and you can send things to multiple targets as well, which is very nice.<\/p><h3 id=\"to-emacs\">To Emacs<\/h3><p class=\"text-justify\">Because Eshell has to reimplement pseudo-devices internally it is not at the mercy of dealing with just UNIX device files \u2013 it is actually capable of implementing its own pseudo-devices.<\/p><p class=\"text-justify\">A good example would be redirection to a buffer of your choosing, and that can be done with the following syntax:<\/p><pre>\/ $ cat mylog.log &gt;&gt; #&lt;buffer *scratch*&gt;<\/pre><p class=\"text-justify\">The keybind I mentioned before <code>C-c M-b<\/code> will insert the printed name of a buffer.<\/p><p class=\"text-justify\">You can also output straight to an elisp symbol (but be careful you don\u2019t fry the wrong settings):<\/p><pre>\/ $ echo foo bar baz &gt; #'myvar\n\/ $ echo $(cadr myvar)\nbar<\/pre><p class=\"text-justify\">If you set <code>eshell-buffer-shorthand<\/code> to <code>t<\/code> you can use the shorthand <code>#'*scratch*<\/code> instead, but it means you will not be able to redirect straight to elisp symbols.<\/p><h3 id=\"to-pseudo-devices\">To Pseudo-Devices<\/h3><p class=\"text-justify\">Eshell reimplements the following pseudo-devices:<\/p><dl><dt><code>\/dev\/eshell<\/code><\/dt><dd><p class=\"text-justify\">Prints the output interactively to Eshell.<\/p><\/dd><dt><code>\/dev\/null<\/code><\/dt><dd><p class=\"text-justify\">Sends the output to the NULL device.<\/p><\/dd><dt><code>\/dev\/clip<\/code><\/dt><dd><p class=\"text-justify\">Sends the output to the clipboard.<\/p><\/dd><dt><code>\/dev\/kill<\/code><\/dt><dd><p class=\"text-justify\">Sends the output to the kill ring.<\/p><\/dd><\/dl><p class=\"text-justify\">The usual redirection rules like overwrite (<code>&gt;<\/code>) and append (<code>&gt;&gt;<\/code>) apply here.<\/p><h3 id=\"to-custom-virtual-targets\">To custom virtual targets<\/h3><p class=\"text-justify\">You can design your own virtual targets by modifying <code>eshell-virtual-targets<\/code>, an alist that takes the name of the pseudo-device you want to create, and a function that takes one parameter, <code>mode<\/code>, that determines if it\u2019s <code>overwrite<\/code>, <code>append<\/code> or <code>insert<\/code>.<\/p><h2 id=\"tramp\">TRAMP<\/h2><p class=\"text-justify\">Eshell now supports TRAMP natively, which means commands like <code>su<\/code>, <code>sudo<\/code> and <code>whoami<\/code> now query the remote system if the directory Eshell is in is <em>remote<\/em>.<\/p><p class=\"text-justify\">To use the TRAMP functionality simply enter the same TRAMP command string you\u2019d use in <code>C-x C-f<\/code> and off it goes. The TRAMP support in Eshell can be a bit flakey, but it <em>does<\/em> give you a remote shell courtesy of TRAMP. You don\u2019t have to limit your TRAMP use to <em>remote<\/em> shells, as TRAMP is also capable of using <code>sudo<\/code> and <code>su<\/code> for local use.<\/p><p class=\"text-justify\">I\u2019ll cover TRAMP in greater detail in a separate article, but the <a href=\"http:\/\/www.gnu.org\/software\/tramp\/\" class=\"article-link\">official manual<\/a> is a good place to start.<\/p><h2 id=\"startup-scripts\">Startup Scripts<\/h2><p class=\"text-justify\">Like most shells, Eshell supports both login and profile\/rc shell scripts. The full filepaths for both are stored in the variables <code>eshell-login-script<\/code> and <code>eshell-rc-script<\/code>, but by default the files <code>login<\/code> and <code>profile<\/code> are stored in <code>~\/.eshell\/<\/code>.<\/p><p class=\"text-justify\">It bears mention that the comment syntax is <code>#<\/code>.<\/p><h2 id=\"more-customization\">More Customization\u2026<\/h2><p class=\"text-justify\">Eshell has hundreds of options you can tweak to your liking. To configure Eshell, type <code>M-x customize-group RET eshell RET<\/code>.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p class=\"text-justify\">Phew. I think I\u2019ve covered all major areas of Eshell, and I hope it paints it in a good light. Eshell is remarkably versatile thanks to its tight integration with Emacs. It\u2019s not a full-on replacement for bash and your favorite terminal emulator, but it\u2019ll do most of the commandline stuff we all inevitably end up doing. If you use a lot of interactive programs then Eshell is probably not very useful, as it has to spawn a separate term instance for every visual program you run.<\/p><p class=\"text-justify\">Eshell has TRAMP support, custom pseudo-devices, a pocket-sized elisp REPL and lots of useful utility commands like being able to <code>find-file<\/code> or <code>dired<\/code> any directory or file you\u2019re in, and that makes it a trusty tool in my toolbox.<\/p>","created_at":"2018-11-07T16:44:56+0000","updated_at":"2018-11-07T16:44:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":19,"domain_name":"masteringemacs.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":331,"title":"The Power of Clojure: Debugging","url":"https:\/\/cambium.consulting\/articles\/2018\/2\/8\/the-power-of-clojure-debugging","content":"<div class=\"sqs-layout sqs-grid-12 columns-12\" data-layout-label=\"Post Body\" data-type=\"item\" data-updated-on=\"1518134480912\" id=\"item-5a7ce0ceec212dc959566681\"><div class=\"row sqs-row\"><div class=\"col sqs-col-8 span-8\"><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-86f606c1ff16226d56eb\"><div class=\"sqs-block-content\"><p>A common question we hear is \"How do I use Clojure for real?\" Not the language basics, but the practicalities of building software \u2013 questions like how to structure the project file tree and namespace hierarchy, how to write tests, or how to use the REPL.<\/p><p>This post tackles a small subset of this: debugging, and more specifically, REPL-based debugging in Clojure.<\/p><p>Debugging is fundamentally difficult. Clojure is a simple language, which helps with the cognitive load of debugging, and the debugging tools available to Clojure programmers are simple and powerful too.<\/p><p>Debugging in Clojure with a REPL involves using a number of small, simple tools in a systematic approach to achieve a thorough understanding of potentially complex situations. In this post, we will cover both these tools and the approach with which to apply to the tools to solve your debugging problems.<\/p><h2>Why Is This Important?<\/h2><p>A few years ago, a <a href=\"http:\/\/www.prweb.com\/releases\/2013\/1\/prweb10298185.htm\">Cambridge University study<\/a> found that \"[...] on average, software developers spend 50% of their programming time finding and fixing bugs.\"\u00a0Debugging is inevitable, necessary, important work.<\/p><p>For our purposes, we will take a look at two aspects of debugging: exploration of existing code and fixing bugs.<\/p><p><strong>Exploration<\/strong>: When exploring new projects, debugging is often how you start learning about \u00a0how the code works and how it's structured. When you need to alter code structure (e.g. breaking apart a large function before changing its behavior) you need to understand how it works. Learning how libraries work can be difficult. Often the documentation is insufficient and it\u2019s necessary to dive in to the code to answer questions.<\/p><p><strong>Fixing Bugs<\/strong>: It's unavoidable, it will be necessary to solve issues that present themselves in production. People will always write bugs, therefore you will always have to debug code.<\/p><p>Both of these contexts are an intrinsic part of the work of a software developer, and if these constitute 50% of your work (at least in time), debugging skills are important to hone. But note: debugging skills are just that: skills. Learnable, systematized, something to study.<\/p><h2>What We Won\u2019t Cover:<\/h2><p>Whilst important and and related, this post won\u2019t cover the following:<\/p><ul><li><p>Setting up logging frameworks<\/p><\/li><li><p>Reading and understanding Clojure\/Script exceptions<\/p><\/li><li><p>Profiling (understanding the performance characteristics of code)<\/p><\/li><li><p>Interactive debuggers such as <a href=\"https:\/\/github.com\/clojure-emacs\/cider\/blob\/master\/doc\/debugging.md\">CIDER\u2019s debugger<\/a> or Cursive\u2019s integration with IntelliJ\u2019s debugger.<\/p><\/li><\/ul><h2>Debugging Framework: The Scientific Method<\/h2><\/div><\/div><div class=\"sqs-block image-block sqs-block-image sqs-col-4 span-4 float float-right\" data-block-type=\"5\" id=\"block-yui_3_17_2_1_1518133363916_29256\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&amp;nbsp;&lt;em&gt;https:\/\/upload.wikimedia.org\/wikipedia\/commons\/c\/c7\/The_Scientific_Method.jpg&lt;\/em&gt;&lt;\/p&gt;\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce2cd71c10b289b5ff83d\/1518133965540\/The_Scientific_Method.jpg\" alt=\"&amp;nbsp; https:\/\/upload.wikimedia.org\/wikipedia\/commons\/c\/c7\/The_Scientific_Method.jpg\" \/><\/noscript><img class=\"thumb-image\" alt=\"&amp;nbsp; https:\/\/upload.wikimedia.org\/wikipedia\/commons\/c\/c7\/The_Scientific_Method.jpg\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce2cd71c10b289b5ff83d\/1518133965540\/The_Scientific_Method.jpg\" data-image-dimensions=\"364x294\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7ce2cd71c10b289b5ff83d\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce2cd71c10b289b5ff83d\/1518133965540\/The_Scientific_Method.jpg\" \/><\/div><div class=\"image-caption-wrapper\"><div class=\"image-caption\"><p>\u00a0<em>https:\/\/upload.wikimedia.org\/wikipedia\/commons\/c\/c7\/The_Scientific_Method.jpg<\/em><\/p><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_94187\"><div class=\"sqs-block-content\"><p>Debugging with the scientific method is hardly a novel idea and plenty of great resources exist, but if you\u2019re looking for some information specific to Clojure, check out Stu Halloway\u2019s 2015 Conj talk (<a href=\"https:\/\/www.youtube.com\/watch?v=FihU5JxmnBg\">video<\/a>, <a href=\"https:\/\/github.com\/stuarthalloway\/presentations\/wiki\/Debugging-with-the-Scientific-Method\">slides and links<\/a>) as well as Kyle Kingsbury\u2019s \"Clojure from the Ground Up\"\u00a0<a href=\"https:\/\/aphyr.com\/posts\/319-clojure-from-the-ground-up-debugging\">chapter on debugging<\/a>. Kyle\u2019s work specifically covers debugging and many other valuable techniques.<\/p><p>In short:<\/p><ul><li>Observe the situation (problem)<\/li><li>Form a hypothesis to explain the problem<\/li><li>Devise and execute a test to show the correctness of your hypothesis<ul><li>Inconclusive? Try again.<\/li><li>Conclusive? Good! The problem space has been reduced, try again with a smaller subset.<\/li><\/ul><\/li><\/ul><p>Contextually, we use debugging techniques to help observe and understand a problem, and to quantitatively evaluate the outcome of a test.<\/p><h2>Instrumenting: Inspecting, Tracing<\/h2><\/div><\/div><div class=\"sqs-block image-block sqs-block-image sqs-col-4 span-4 float float-right\" data-block-type=\"5\" id=\"block-yui_3_17_2_1_1518133363916_37708\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;https:\/\/pixabay.com\/en\/meters-armatures-pipes-factory-498677\/&lt;\/em&gt;&lt;\/p&gt;\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce3c58165f5831f68aca8\/1518186372199\/meters-498677_1920.jpg\" alt=\"https:\/\/pixabay.com\/en\/meters-armatures-pipes-factory-498677\/\" \/><\/noscript><img class=\"thumb-image\" alt=\"https:\/\/pixabay.com\/en\/meters-armatures-pipes-factory-498677\/\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce3c58165f5831f68aca8\/1518186372199\/meters-498677_1920.jpg\" data-image-dimensions=\"1614x1015\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7ce3c58165f5831f68aca8\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce3c58165f5831f68aca8\/1518186372199\/meters-498677_1920.jpg\" \/><\/div><div class=\"image-caption-wrapper\"><div class=\"image-caption\"><p><em>https:\/\/pixabay.com\/en\/meters-armatures-pipes-factory-498677\/<\/em><\/p><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_11558\"><div class=\"sqs-block-content\"><p>Adding instrumentation to code is like adding gauges and meters to a mechanical system.<\/p><p>Instrumentation can be added to code in two ways: by changing the code to explicitly instrument a particular value or control path, or by changing the runtime environment to provide insights into the code running within.<\/p><p>Changing the code is easy, precise, but often messy, overly redundant, and prone to leaving the instrumentation in by mistake.<\/p><p>Instrumenting the runtime is powerful and more easily able to paint a fuller picture, but it is generally more complex and harder to master, and can overwhelm you with data.<\/p><h2>Example Code<\/h2><p>The code samples used below are variations examples of a typical programming problems. The basic overview with lots of annotations and comments can be found here:\u00a0<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1518133363916_43054\"><div class=\"sqs-block-content\"><h2>Clojure Debugging Playbook<\/h2><p>The three following sections outline techniques for:<\/p><ul><li>instrumenting the code,\u00a0<\/li><li>Instrumenting the runtime, and<\/li><li>using the REPL itself as a debugging tool.\u00a0<\/li><\/ul><p>In our experience, problems are usually solved by applying a combination of two or three of these techniques.<\/p><h2>Instrumenting the Code<\/h2><p>Clojure is a dynamic language. It\u2019s very easy to make changes to a function, eval the code in the REPL, and immediately observe the effects. We can use this tight feedback loop to directly instrument the code.<\/p><h3>println debugging<\/h3><blockquote><div>\"The most effective debugging tool is still careful thought, coupled with judiciously placed print statements.\" \u2013 Brian Kernighan, Unix for Beginners (1979).<\/div><\/blockquote><p>First in our list is probably every programmer\u2019s first debugging technique: <em>println<\/em>. It\u2019s sometimes seen as a naive approach, perhaps because in compiled languages, like C or Java, the edit\/execute\/inspect cycle time can be so long. However, in an immutable, functional, dynamic language such as Clojure it is highly effective.<\/p><p>The technique is simple: to see value of an s-expression as the code is executed, <em>println<\/em> it. To trace the execution of the code, add <em>println<\/em> statements to act as a \u201cbreadcrumb trail\u201d through the code.<\/p><p>For example, the example code has an infinite loop. Knowing that recursion can be tricky, we add a little <em>println<\/em> into the loop to see what\u2019s happening:<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1518133363916_63476\"><div class=\"sqs-block-content\"><p>We never finish processing the first account. Looking at the loop\/recur, we never reduce the size of the <em>accounts<\/em> collection - we forgot a <em>(rest accounts)<\/em> in the recur call.<\/p><p>When you have more than a couple of print statements you will need to be careful to provide enough context in your <em>println<\/em> so that you know what each output value corresponds to. You might also have to change the structure of your code a little, e.g. introducing a let binding to capture a variable\u2019s value. You can make this easier with a short helper function:<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1518133363916_65471\"><div class=\"sqs-block-content\"><p>When using <em>println<\/em>s it\u2019s necessary to delete them prior to committing code. It\u2019s very easy to forget and leave them in. This problem can be mitigated by using a logging framework and adjusting environments to output the logs at appropriate levels. We cover more on logging below.<\/p><h3>inspecting let bindings<\/h3><p>We often want to see one or more of the values in a let binding. A specific instance of println value inspection is to use the _ (underscore) convention for an unused variable to act as the name to bind in order to add a <em>println<\/em> call directly into the let bindings.<\/p><p>Looking at the full stacktrace from the last error, we can see the error is thrown from inside the <em>as-currency<\/em> function. We can introduce a let and inspect all the intermediate values to see where the problem might be:<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1518133363916_67469\"><div class=\"sqs-block-content\"><p>A common error is forgetting the _ binding before each <em>println<\/em>.<\/p><p>A more elegant way to do this is with a debugging let macro. This example shows how it works, but you would likely put the <em>dlet<\/em> definition in your user namespace somewhere.<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1518133363916_69322\"><div class=\"sqs-block-content\"><p>The macro instruments every binding in the let form automatically, printing out all data with a single additional character. This is a useful timesaver for larger <em>let<\/em> forms.<\/p><h3>Inspecting threading intermediates<\/h3><p>Again, another special case of println debugging is to inspect the intermediate values in a chain of threaded functions.<\/p><p>This technique inserts an invocation of an identity function into a list of threaded forms, but the function has the additional side effect of printing out its argument. The following example shows both the direct inline anonymous function, and a quick helper function:<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_18102\"><div class=\"sqs-block-content\"><p>The double parens aren\u2019t very common, but here they are necessary - we use them to define and immediately execute a function. Also, we use them to make sure to return the value from the inspection function!<\/p><h3>logging<\/h3><p>Logging: it's basically println debugging but better. From a debugging point of view it is useful because you get a file name and line number for each log statement. From a maintenance point of view, you can leave the log statements in but disable them from printing via configuration.<\/p><p>Each log statement is assigned a log level (typically trace, debug, info, warning, and error) and the log framework is configured to emit log statements at a particular log level on a per-namespace basis. For example, you may want to configure your development environment to print all debug statements, however you just want warnings and errors in production.\u00a0<\/p><p>Like <em>println<\/em>, logger calls return nil, so if you are instrumenting a function that needs a return variable, you may need to capture the value to log and then return it. Here are a couple of ways to do that:<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_20830\"><div class=\"sqs-block-content\"><p>Spyscope<\/p><p>The <a href=\"https:\/\/github.com\/dgrnbrg\/spyscope\">Spyscope library<\/a>\u00a0is like a ninja version of println debugging. From its GitHub page:<\/p><blockquote><div>Spyscope includes 3 reader tools for debugging your Clojure code, which are exposed as reader tags: #spy\/p, #spy\/d, and #spy\/t, which stand for print, details, and trace, respectively. Reader tags were chosen because they allow one to use Spyscope by only writing 6 characters, and since they exist only to the left of the form one wants to debug, they require the fewest possible keystrokes, optimizing for developer happiness.<\/div><\/blockquote><p><em>#spy\/p<\/em> has the same uses and drawbacks of println debugging, but it\u2019s shorter and easier to add into code, and it outputs pretty printed colorized output.<\/p><p><em>#spy\/d <\/em>is even more powerful, showing one or more stack frames for where it was called from, timing data, and lots of other useful stuff. The <a href=\"https:\/\/github.com\/dgrnbrg\/spyscope\/blob\/master\/README.md\">project README<\/a> contains full documentation of all spyscope\u2019s functionality.<\/p><p>Here is a screencap of how you can quickly add spyscope tracing to your code, and what the output looks like:<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-yui_3_17_2_3_1518186156545_25410\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper lightbox  has-aspect-ratio\" data-description=\"\" data-lightbox-theme=\"dark\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db14a085229288120626d\/1518186884365\/spyscope.gif\" alt=\"spyscope.gif\" \/><\/noscript><img class=\"thumb-image\" alt=\"spyscope.gif\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db14a085229288120626d\/1518186884365\/spyscope.gif\" data-image-dimensions=\"1311x759\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7db14a085229288120626d\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db14a085229288120626d\/1518186884365\/spyscope.gif\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_25688\"><div class=\"sqs-block-content\"><h2>Instrumenting the runtime<\/h2><p>There are a number of more sophisticated tools that observe control flow and return values on your behalf, giving you ways to view and even interact with that data.<\/p><h3>clojure\/tools.trace (+ CIDER)<\/h3><p>The canonical tracing library, <a href=\"https:\/\/github.com\/clojure\/tools.trace\">clojure.tools.trace<\/a>, exposes two functions to add traces and two functions to remove them, one each for functions and for namespaces. Tracing is applied dynamically without modifying the code being instrumented. As traced functions are run, arguments and return values are printed. Nested function calls are printed with nesting indicated.\u00a0<\/p><p>This screencap shows the same faulty bank transactions example, first being run without tracing, tracing getting applied, and then running with tracing:<br \/>\u00a0<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-yui_3_17_2_3_1518186156545_59218\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db204c83025fff3e4acc7\/1518187018400\/clojure_tools_trace.gif\" alt=\"clojure_tools_trace.gif\" \/><\/noscript><img class=\"thumb-image\" alt=\"clojure_tools_trace.gif\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db204c83025fff3e4acc7\/1518187018400\/clojure_tools_trace.gif\" data-image-dimensions=\"850x598\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7db204c83025fff3e4acc7\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db204c83025fff3e4acc7\/1518187018400\/clojure_tools_trace.gif\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_57199\"><div class=\"sqs-block-content\"><p>Helpfully (for debugging), the function throwing an exception is printed last, clearly showing its arguments.\u00a0<\/p><p>CIDER has a <a href=\"https:\/\/github.com\/clojure-emacs\/cider\/blob\/master\/doc\/miscellaneous_features.md#tracing-function-execution\">couple of convenience functions<\/a> to help with tracing and untracing, plus highlights traced functions.<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-yui_3_17_2_3_1518186156545_72898\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db2570852292881209db0\/1518187099817\/cider-trace.gif\" alt=\"cider-trace.gif\" \/><\/noscript><img class=\"thumb-image\" alt=\"cider-trace.gif\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db2570852292881209db0\/1518187099817\/cider-trace.gif\" data-image-dimensions=\"1178x695\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7db2570852292881209db0\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db2570852292881209db0\/1518187099817\/cider-trace.gif\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_70748\"><div class=\"sqs-block-content\"><p>Clearly, these are powerful tools. You get to see argument and return values, and get a sense of the control flow, with no need to modify the source itself. A possible downside is that you can get an overwhelming amount of trace data back.<\/p><p>An additional noteworthy mention is for <a href=\"https:\/\/github.com\/bpiel\/sayid\">Sayid<\/a>. It is an inspecting debugger for the Clojure runtime and has a corresponding mode for Emacs. It offers even more advanced views into the data with the ability to drill down and even replay execution.\u00a0<\/p><p>One final advantage of inspecting the runtime instead of modifying the code is it is a safer technique: if you have to connect to a production server\u2019s nREPL to live debug a particularly tricky problem, adding traces to the vars is much less error prone than adding instrumentation to the code and re-evaluating live.<\/p><h3>re-frisk \/ re-frame-trace<\/h3><p><a href=\"https:\/\/github.com\/flexsurfer\/re-frisk\">Re-frisk<\/a>\u00a0and <a href=\"https:\/\/github.com\/Day8\/re-frame-trace\">re-frame-trace<\/a>\u00a0are specific to ClojureScript re-frame apps, but, given their current popularity, these tools are worth mentioning. Both provide inspection tools for the global application state, and an inspectable log containing the sequence of re-frame events triggered during the application\u2019s use. In combination with the other debugging techniques shown here, this specialized view of the re-frame events is very useful for solving the front end bugs that are typically hard to track down.<\/p><h2>Evaluative techniques<\/h2><p>The least obvious way to debug for new Clojure programmers is to leverage the language\u2019s dynamic nature and tight REPL integrations. Of the three general categories of debugging tools presented here, this is the most powerful and the most general purpose. It is also a key technique for effectively authoring Clojure code.<\/p><h3>Using the REPL to eval code<\/h3><p>Fast keystroke-based REPL integration with your editor is a must. Emacs, Atom, Cursive and Vim all support this workflow. We highly recommend you become familiar with this way of working. Is truly one of the most amazing aspects of programming in Clojure.<\/p><p>You can use your editor to evaluate code in the Clojure REPL directly from the source code. You might be evaluating a form to see what it evaluates as, or to observe any side effects (e.g. println output). Or you might be evaluating a new function, or a newly changed function, so that it is available for use by other code. Notably with all of these, you don\u2019t need to reload your project, restart the REPL, type directly into the REPL, or even save your source file. Your Clojure runtime is always running; you can change the program, execute parts of it, or write new parts and see how they work in context.<\/p><p>Perhaps the best way to explain this is by example: using the same code sample from earlier, I run the calculation and see an exception, add some instrumentation to observe the program flow, spot a problem case and run that manually, fix the code error and load the fixed function, then re-run the code. The pane on the right shows the Emacs keystrokes and more importantly the commands they correspond to.\u00a0<br \/>\u00a0<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-yui_3_17_2_3_1518186156545_144688\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db3acc83025fff3e50af1\/1518187447056\/repling.gif\" alt=\"repling.gif\" \/><\/noscript><img class=\"thumb-image\" alt=\"repling.gif\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db3acc83025fff3e50af1\/1518187447056\/repling.gif\" data-image-dimensions=\"990x728\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7db3acc83025fff3e50af1\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db3acc83025fff3e50af1\/1518187447056\/repling.gif\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_144966\"><div class=\"sqs-block-content\"><h3>repl, def, eval<\/h3><p>Many times you want to evaluate how some particular function might be breaking. We\u2019ve shown a number of ways how you can insert code to instrument a function\u2019s operation, but you can use the REPL instead.\u00a0<\/p><p>Stu Halloway has <a href=\"http:\/\/blog.cognitect.com\/blog\/2017\/6\/5\/repl-debugging-no-stacktrace-required\">another fantastic blog post<\/a> where he uses a combination of binary search, brainpower, and the REPL to debug an error without even having access to a stacktrace. The REPL technique he uses is this: temporarily define variables that correspond to the names used within your function, then evaluate the code in your function directly.<\/p><p>For example, say we came up with an as-currency function that would handle messier input, like comma separators for thousands:<\/p><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_156859\"><div class=\"sqs-block-content\"><p>Running the function throws an exception, so something isn\u2019t right. The following screencap shows how we test how the components of the <em>bigdec<\/em> conversion work assuming the <em>negative? <\/em>and <em>cleaned-amount<\/em> checks are correct, then narrowing the problem to the value of <em>cleaned-amount<\/em>. That\u2019s definitely not as expected, and so we further narrow the problem to the <em>str\/replace<\/em> call:<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-yui_3_17_2_3_1518186156545_160267\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db44ec83025fff3e53001\/1518187620209\/\" alt=\"image\" \/><\/noscript><img class=\"thumb-image\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db44ec83025fff3e53001\/1518187620209\/\" data-image-dimensions=\"1094x598\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7db44ec83025fff3e53001\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db44ec83025fff3e53001\/1518187620209\/\" alt=\"image\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_160546\"><div class=\"sqs-block-content\"><h3>Capture runtime data for post-run REPL analysis\u00a0<\/h3><p>In the previous example, by manually entering some appropriate data and binding a var we can evaluate subcomponents of a function independently. Sometimes the data we need to correctly operate a function is more complex than can reasonably be typed in. Similarly, it might be that you don\u2019t quite know the exact data to \u201cbreak\u201d a function, but you can run the code from its top level and get the code to break.<\/p><p>In these cases, we can capture runtime data, either by defining a var or by using an atom. Then, when the main execution flow has completed, we can use the data we captured and evaluate that data using our functions, under controlled conditions in the REPL.<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-yui_3_17_2_3_1518186156545_166601\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db491c83025fff3e53eb8\/1518187668608\/\" alt=\"image\" \/><\/noscript><img class=\"thumb-image\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db491c83025fff3e53eb8\/1518187668608\/\" data-image-dimensions=\"651x469\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5a7db491c83025fff3e53eb8\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7db491c83025fff3e53eb8\/1518187668608\/\" alt=\"image\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_3_1518186156545_166880\"><div class=\"sqs-block-content\"><h2>Other thoughts<\/h2><p>In wrapping up, another of Brian Kernighan\u2019s brilliant aphorisms is applicable:<\/p><blockquote><div>\"Everyone knows that debugging is twice as hard as writing a program in the first place. So if you're as clever as you can be when you write it, how will you ever debug it?\"\u00a0\u2013 Brian Kernighan, The Elements of Programming Style<\/div><\/blockquote><p>While this has been an entire article focused on inspecting and instrumenting code at run time, without an understanding of the code at hand these efforts are seriously hampered. We must always be able to read and comprehend the code that we are evaluating. When you write code, always strive for readability. Your 6-months-future self will thank you.<\/p><h2>Summary<\/h2><p>Hopefully we\u2019ve helped answer the question, \u201cHow do I debug in Clojure?\u201d\u00a0<\/p><p>Remember these three general techniques:<\/p><ul><li>instrument the code,\u00a0<\/li><li>instrument the runtime, and<\/li><li>use the REPL itself as a debugging tool.\u00a0<\/li><\/ul><p>You can modify code live in a REPL or editor (depending upon your build setup) to quickly add println, log, and trace statements. You can temporarily define variables and atoms to capture and inspect values in a connected REPL, or inspect a re-frame SPA\u2019s app-db and all events in the system from a browser window with tools like re-frame-trace and re-frisk. There are myriad techniques to help you understand what your Clojure and ClojureScript code is doing, and in practice you will probably use more than one of these at the same time<\/p><p>Once you are familiar with these techniques, you will learn which is the most effective to use to solve a particular problem. The techniques all overlap, each having its own particular pros and cons, and each having different maintainability, complexity, and flexibility characteristics.<\/p><p>Remember though, the easiest program to debug is the one that doesn\u2019t need it, because it is written so clearly that any errors are obvious. Decomposing code into simple, pure functions, covered by reasonable unit tests, goes a long way towards not needing to use the above techniques too often.<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image sqs-col-2 span-2 float float-left\" data-block-type=\"5\" id=\"block-yui_3_17_2_1_1535377202311_33089\"><div class=\"sqs-block-content\"><div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\"><div class=\"intrinsic\"><div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"\"><noscript><img src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5b83ff4e1ae6cfab3d2625bc\/1535377235629\/head.jpeg\" alt=\"head.jpeg\" \/><\/noscript><img class=\"thumb-image\" alt=\"head.jpeg\" data-image=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5b83ff4e1ae6cfab3d2625bc\/1535377235629\/head.jpeg\" data-image-dimensions=\"460x460\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b83ff4e1ae6cfab3d2625bc\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5b83ff4e1ae6cfab3d2625bc\/1535377235629\/head.jpeg\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-yui_3_17_2_1_1535377202311_28595\"><div class=\"sqs-block-content\"><p><em>Ian Truslove is a co-founder of <a target=\"_blank\" href=\"https:\/\/cambium.consulting\/\">Cambium Consulting<\/a>. He specializes in building large-scale resilient data processing systems using tools like Clojure and Elasticsearch. When not hunched over an Emacs terminal, you might find him on a bike in the wilds of Colorado.<\/em><\/p><\/div><\/div><\/div><\/div><\/div><p><a href=\"https:\/\/cambium.consulting\/articles\/?author=5a0a40336f935eb2a9322cf6\" class=\"Blog-meta-item Blog-meta-item--author\">Ian Truslove<\/a><time class=\"Blog-meta-item Blog-meta-item--date\" datetime=\"2018-02-09\">February 9, 2018<\/time><a href=\"https:\/\/cambium.consulting\/articles\/?tag=clojure\" class=\"Blog-meta-item-tag\">clojure<\/a>, <a href=\"https:\/\/cambium.consulting\/articles\/?tag=debugging\" class=\"Blog-meta-item-tag\">debugging<\/a>, <a href=\"https:\/\/cambium.consulting\/articles\/?tag=software+engineering\" class=\"Blog-meta-item-tag\">software engineering<\/a><\/p><section id=\"comments-5a7ce0ceec212dc959566681\" class=\"BlogItem-comments\"><\/section>","created_at":"2018-11-07T16:44:55+0000","updated_at":"2018-11-07T16:44:55+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":14,"domain_name":"cambium.consulting","preview_picture":"http:\/\/static1.squarespace.com\/static\/5a0a403451a58468b380ced9\/t\/5a7ce2cd71c10b289b5ff83d\/1518133965540\/The_Scientific_Method.jpg?format=1000w","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":332,"title":"sqlkorma","url":"http:\/\/sqlkorma.com\/","content":"<ul><li><div><h3>Composable from the ground up.<\/h3><p>All the parts of a query in Korma can be composed at will.\n                  This allows you to construct queries over time and in virtually\n                  any way you can dream of.<\/p><\/div><code>(def base (-&gt; (<em>select*<\/em> \"user\")\n            (<em>fields<\/em> :id:username)\n            (<em>where<\/em> {:email [<em>like<\/em> \"*@gmail.com\"]})))\n(def ordered-and-active (-&gt; base\n                          (<em>where<\/em> {:active true})\n                          (<em>order<\/em> :created)))\n(def constrained (-&gt; ordered-and-active\n                   (<em>limit<\/em> 20)))\n(<em>select<\/em> constrained)\n<\/code><\/li><li><div><h3>Take the pain out of relations.<\/h3><p>Korma entities handles relationships for you. Just specify\n                  what kind of relationship it is and all the joining will be\n                  done for you.<\/p><\/div><code>(<em>defentity<\/em> email)\n(<em>defentity<\/em> address)\n(<em>defentity<\/em> account)\n(<em>defentity<\/em> users\n  (<em>has-one<\/em> email)\n  (<em>has-many<\/em> address)\n  (<em>belongs-to<\/em> account))\n(<em>select<\/em> users\n  (<em>fields<\/em> :email.email:address.state:account.name)\n  (<em>with<\/em> email)\n  (<em>with<\/em> address)\n  (<em>with<\/em> account))\n<\/code><\/li><li><div><h3>Built for the real world.<\/h3><p>Real projects are never simple. As such, Korma was designed with\n                  flexibility in mind. Couple that with fast query generation\n                  and you have a library meant to play in the big leagues.<\/p><\/div><code>(<em>defentity<\/em> address\n  (<em>table<\/em> :__addresses:address))\n(<em>defentity<\/em> users\n  (<em>table<\/em> :somecrazy_table_name:users)\n  (<em>pk<\/em> :userID)\n  (<em>has-many<\/em> address {:fk:userID}))\n(<em>select<\/em> users\n  (<em>with<\/em> address)\n  (<em>where<\/em> {:last_login [<em> a-week-ago]}))\n(<em>select<\/em> users\n  (<em>aggregate<\/em> (<em>count<\/em> :*) :cnt)\n  (<em>where<\/em> (<em>or<\/em> (<em>&gt;<\/em> :visits 20)\n             (<em> :last_login a-year-ago))))\n<\/em><\/em><\/code><\/li><\/ul>","created_at":"2018-11-07T16:44:55+0000","updated_at":"2018-11-07T16:44:55+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"sqlkorma.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":329,"title":"CHICKEN Scheme","url":"https:\/\/www.call-cc.org\/","content":"<div class=\"grid\"><h4>Free<\/h4><p>Distributed free for use and modification under the <a href=\"http:\/\/opensource.org\/licenses\/BSD-3-Clause\">BSD License<\/a> (freedom to user and developer).<\/p><h4>Simple<\/h4><p>Lightweight on dependencies (a C toolchain and GNU Make) and <a href=\"https:\/\/code.call-cc.org\/\">easy to install<\/a>.<\/p><h4>Portable<\/h4><p>Runs on <a href=\"https:\/\/wiki.call-cc.org\/portability\">many platforms<\/a>, including x86, x86-64, ARM, MIPS, SPARC64 and PowerPC.<\/p><h4>Extensible<\/h4><p>Many libraries and extensions are available at <a href=\"https:\/\/wiki.call-cc.org\/egg-index\">Eggs Unlimited<\/a> and new ones can be easily created.<\/p><h4>Well documented<\/h4><p><a href=\"https:\/\/wiki.call-cc.org\/\">Wiki<\/a>, <a href=\"https:\/\/wiki.call-cc.org\/man\/4\/\">Manual<\/a> and <a href=\"http:\/\/api.call-cc.org\/doc\/\">API<\/a> reference provide updated and detailed documentation about CHICKEN use.<\/p><h4>Actively supported<\/h4><p><a href=\"https:\/\/wiki.call-cc.org\/discussion-groups\">Mailing lists<\/a> and an IRC channel (<em>#chicken<\/em> on Freenode) for useful and kind support.<\/p><\/div><div class=\"code\"><h3>A glimpse of CHICKEN<\/h3><pre><code class=\"highlight scheme-language\">;;; hello-world.scm\n(print \"Hello, world!\")\n;;; Running it interpreted:\n$ csi -s hello-world.scm \nHello, world!\n;;; Compiling and running the executable binary:\n$ csc hello-world.scm \n$ .\/hello-world\nHello, world!<\/code><\/pre><\/div><p><a href=\"https:\/\/wiki.call-cc.org\/man\/4\/Getting%20started\">Get started!<\/a><\/p>","created_at":"2018-11-07T16:44:54+0000","updated_at":"2018-11-07T16:44:54+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.call-cc.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":330,"title":"Build a Laptop Computer","url":"https:\/\/www.wikihow.com\/Build-a-Laptop-Computer","content":"<div id=\"intro\" class=\"section  sticky  hasad\">\n<h1 class=\"firstHeading\"><a href=\"https:\/\/www.wikihow.com\/Build-a-Laptop-Computer\">How to Build a Laptop Computer<\/a><\/h1>\n<p id=\"method_toc\" class=\"sp_method_toc\">In this Article:<a href=\"#Finding_the_Parts_sub\">Finding the Parts<\/a><a href=\"#Putting_it_Together_sub\">Putting it Together<\/a><a href=\"#Starting_it_Up_sub\">Starting it Up<\/a><a id=\"qa_toc\" href=\"#Questions_and_Answers_sub\">Community Q&amp;A<\/a><\/p>\n<p>Buying a prebuilt laptop from the store is typically an exercise in frustration. The features you want are usually not all available in one computer, and the price can be exorbitant. Not to mention all the software that the companies stuff into it. You can bypass all that if you're willing to get your hands a little dirty. Building your own laptop is a challenge, but incredibly rewarding. Follow this guide to learn how.<\/p>\n<div class=\"clearall\">\n<div class=\"wh_ad_inner\" data-service=\"adsense\" data-adtargetid=\"introad\" data-loaded=\"1\" data-lateload=\"0\" data-adsensewidth=\"728\" data-adsenseheight=\"120\" data-slot=\"7862589374\" data-channels=\"7275552975\" data-viewablerefresh=\"1\" data-aps-timeout=\"2000\" id=\"introad-outer\">\n<ins class=\"adsbygoogle ad_label ad_label_dollar\" data-ad-client=\"ca-pub-9543332082073187\" data-ad-slot=\"7862589374\">\n<\/ins><\/div>\n<div class=\"clearall adclear\">\n\n<\/div>\n<h2 class=\"hidden\">Steps<\/h2>\n<div class=\"section steps   sticky  steps_first\">\n<h3>\nPart 1\n<\/h3><\/div>\nFinding the Parts\n\n<div id=\"steps_1\" class=\"section_text\">\n<ol class=\"steps_list_2\"><li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-1.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-1.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 1\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_d3443b4f87\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/e\/e1\/Build-a-Laptop-Computer-Step-1.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-1.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 1\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/e\/e1\/Build-a-Laptop-Computer-Step-1.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-1.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript>\n<\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 1\">1<\/div>\n<div class=\"step\">\n<b class=\"whb\">Decide what the laptop\u2019s main purpose will be.<\/b> A laptop for word processing and email-checking will have much different specifications than a laptop for playing the latest games. Battery life is also an important consideration; if you plan to roam around unplugged, you\u2019ll want a laptop that doesn\u2019t draw much power.\n\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-2.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-2.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 2\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_3578cc6c7d\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/8d\/Build-a-Laptop-Computer-Step-2.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-2.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 2\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/8d\/Build-a-Laptop-Computer-Step-2.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-2.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 2\">2<\/div>\n<div class=\"step\">\n<b class=\"whb\">Choose a processor that meets your computer\u2019s needs.<\/b> The shell that you purchase will be dependent on the processor that you want to install, so pick your processor first. Compare processor models to determine which processor offers the best speed versus cooling and power consumption. Most major online retailers will let you compare processors side-by-side.\n\n<ul><li>Make sure that you are purchasing a mobile processor, and not a desktop processor.<\/li>\n<li>There are two major processor manufacturers: Intel and AMD. There are plenty of arguments for and against each brand, but generally AMD will be less expensive. Do as much research as possible on the processor models you are interested in to be certain it is worth the money.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-3.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-3.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 3\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_a8da97ae27\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/1\/1c\/Build-a-Laptop-Computer-Step-3.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-3.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 3\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/1\/1c\/Build-a-Laptop-Computer-Step-3.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-3.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 3\">3<\/div>\n<div class=\"step\">\n<b class=\"whb\">Choose your notebook shell.<\/b> The notebook shell will determine which parts you can use for the rest your laptop. The shell will come with the motherboard attached already, which will dictate what memory you can use.\n\n<ul><li>Take screen size and keyboard layout into account as well. Since the shell is not particularly customizable, you will be stuck with the screen and keyboard you choose. A larger laptop will be more difficult to carry around, and will most likely be significantly heavier.<\/li>\n<li>Finding shells for sale can be difficult. Enter \u201cbarebones notebook\u201d or \u201cwhitebook shell\u201d into your favorite search engine to track down retailers that stock shells. MSI is one of the few manufacturers that still makes barebones laptops.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-4.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-4.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 4\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_35533753c5\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/8e\/Build-a-Laptop-Computer-Step-4.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-4.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 4\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/8e\/Build-a-Laptop-Computer-Step-4.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-4.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 4\">4<\/div>\n<div class=\"step\">\n<b class=\"whb\">Purchase memory.<\/b> Your laptop will need memory to run, and the format of the memory is different than desktop. Look for SO-DIMM memory that will work with the motherboard in your shell. Faster memory will give better performance, but may lead to shorter battery life.\n\n<ul><li>Try to get 8 or 16 GB of memory for optimal everyday performance.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-5.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-5.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 5\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_eddd946f8e\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/1\/13\/Build-a-Laptop-Computer-Step-5.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-5.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 5\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/1\/13\/Build-a-Laptop-Computer-Step-5.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-5.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 5\">5<\/div>\n<div class=\"step\">\n<b class=\"whb\">Pick out a hard drive.<\/b> Laptops typically use 2.5\u201d drives, as opposed to the 3.5\u201d drives found in desktops. You can choose between a standard 5400 RPM or 7200 RPM drive, or pick a solid state drive with no moving parts. The solid state drive (SSD) will typically be faster and has no moving parts, but they can be more difficult to use over long periods.\n\nIf you want a laptop to roam around, a SSD drive would be best, as it will not be physically damaged by impacts as much as a hard disk drive.\n<ul><li>Get a hard drive with enough space for you to do what you want with the laptop. Most shells don\u2019t have space for more than one drive, so it can be difficult to upgrade later. Make sure that there is enough room on the hard drive after the operating system installation (usually between 15-20 GB). Nowadays, most people choose from a 800GB-1.5TB range for laptops.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-6.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-6.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 6\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_ea0585eeb9\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/f7\/Build-a-Laptop-Computer-Step-6.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-6.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 6\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/f7\/Build-a-Laptop-Computer-Step-6.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-6.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 6\">6<\/div>\n<div class=\"step\">\n<b class=\"whb\">Decide if you need a dedicated graphics card (optional).<\/b> Not all shells will fit a dedicated mobile graphics card. Instead, the graphics will be handled by the CPU's integrated graphics unit. If you can install a dedicated card, decide if you need one. They are most important for gamers and graphic designers.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-7.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-7.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 7\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_f9caa94157\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/e\/ee\/Build-a-Laptop-Computer-Step-7.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-7.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 7\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/e\/ee\/Build-a-Laptop-Computer-Step-7.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-7.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 7\">7<\/div>\n<div class=\"step\">\n<b class=\"whb\">Find an optical drive (optional).<\/b> This is becoming more of an optional step as computers advance, since you can install operating systems from USB drives and download most software. If you buy a new laptop today, most do not have an optical drive, as disc memory is now replaced with memory cards and removable drives.\n\n<ul><li>Some shells come with drives included. Not all notebook drives fit all shells, so make sure that the drive fits the shell you picked.<\/li>\n<li>Choosing whether or not to buy one is easy. Consider whether you use disc memory often. Remember, you can use a USB external optical drive instead of a built-in optical drive.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li>\n<div class=\"step_num\" aria-label=\"Step 8\">8<\/div>\n<div class=\"step\">\n<b class=\"whb\">Choose a battery.<\/b> You will need to find one that has the correct shape and uses the same connector (laptop batteries have multiple pins. The battery contains ICs and the IC notifies the computer on temperature, and notify the computer if the battery is not working and should not be charged, and the battery percentage). If you plan to move it around often, use a long life battery. You will need to try comparing many batteries to find one to buy.\n<ul><li>Buy one with good reviews. Read reviews on customer's experience with using those batteries.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<div class=\"wh_ad_inner step_ad\" data-service=\"dfp\" data-adtargetid=\"methodad\" data-loaded=\"0\" data-lateload=\"0\" data-adsensewidth=\"728\" data-adsenseheight=\"90\" data-channels=\"7275552975\" data-viewablerefresh=\"1\" data-apsload=\"1\" data-aps-timeout=\"2000\">\n<div id=\"methodad\">\n\n<\/div>\n<div class=\"clearall adclear\">\n\n<\/div><\/div><\/div><\/li>\n<\/ol><div class=\"clearall\">\n<\/div>\n<\/div>\n<div class=\"section steps   sticky\">\n<h3>\nPart 2\n<\/h3><\/div>\nPutting it Together\n\n<div id=\"steps_2\" class=\"section_text\">\n<ol class=\"steps_list_2\"><li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-8.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-8.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 8\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_f3dcd844b4\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/47\/Build-a-Laptop-Computer-Step-8.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-8.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 8\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/47\/Build-a-Laptop-Computer-Step-8.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-8.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 1\">1<\/div>\n<div class=\"step\">\n<b class=\"whb\">Get the tools.<\/b> You\u2019ll want a set of jeweler\u2019s screwdrivers, preferably magnetic. Laptop screws are much smaller and harder to work with than desktop screws. Find a pair of needle-nose pliers to reach any screws that fall into cracks.\n\n<ul><li>Keep your screws in plastic baggies until you need them. This will help keep them from rolling away or getting lost.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-9.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-9.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 9\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_bb2847e53a\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/30\/Build-a-Laptop-Computer-Step-9.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-9.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 9\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/30\/Build-a-Laptop-Computer-Step-9.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-9.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 2\">2<\/div>\n<div class=\"step\">\n<b class=\"whb\">Ground yourself.<\/b> Electrostatic discharge can quickly ruin computer components, so make sure that you are grounded before assembling your laptop. An antistatic wristband will keep you grounded and they are available for cheap.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-10.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-10.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 10\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_6a9a28974e\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/7\/7f\/Build-a-Laptop-Computer-Step-10.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-10.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 10\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/7\/7f\/Build-a-Laptop-Computer-Step-10.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-10.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 3\">3<\/div>\n<div class=\"step\">\n<b class=\"whb\">Turn over the shell so that the bottom is facing up.<\/b> You will be accessing the motherboard from several removable plates in the back of the unit.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-11.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-11.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 11\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_81e563f990\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/6f\/Build-a-Laptop-Computer-Step-11.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-11.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 11\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/6f\/Build-a-Laptop-Computer-Step-11.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-11.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 4\">4<\/div>\n<div class=\"step\">\n<b class=\"whb\">Remove the panel covering the drive bay.<\/b> This panel covers the 2.5\u201d bay that will hold your hard drive. The location varies depending on the shell, but the bay is typically located towards the front of the laptop.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-12.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-12.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 12\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_d8930e0a36\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/39\/Build-a-Laptop-Computer-Step-12.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-12.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 12\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/39\/Build-a-Laptop-Computer-Step-12.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-12.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 5\">5<\/div>\n<div class=\"step\">\n<b class=\"whb\">Mount the hard drive in the bracket.<\/b> Most notebooks require the hard drive to be mounted to a bracket that fits around the drive. Use four screws to make sure that the hard drive is secured to the bracket. The screw holes will typically ensure that you have it installed the correct direction.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-13.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-13.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 13\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_93af8e0ec3\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/61\/Build-a-Laptop-Computer-Step-13.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-13.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 13\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/61\/Build-a-Laptop-Computer-Step-13.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-13.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 6\">6<\/div>\n<div class=\"step\">\n<b class=\"whb\">Slide the bracketed hard drive into the bay.<\/b> Use the grip tape to apply enough pressure to seat the drive. Most brackets will line up with two screw holes once the drive is in place. Insert screws to secure the drive.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-14.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-14.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 14\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_d64e9f176a\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/88\/Build-a-Laptop-Computer-Step-14.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-14.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 14\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/8\/88\/Build-a-Laptop-Computer-Step-14.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-14.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 7\">7<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install the optical drive.<\/b> The method will vary depending on your shell, but they are typically inserted from the front of the bay opening, and they slide into SATA connectors.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-15.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-15.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 15\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_330c3983e6\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4f\/Build-a-Laptop-Computer-Step-15.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-15.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 15\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/4\/4f\/Build-a-Laptop-Computer-Step-15.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-15.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 8\">8<\/div>\n<div class=\"step\">\n<b class=\"whb\">Remove the panel covering the motherboard.<\/b> This panel will most likely be more difficult to remove than the hard drive panel. You may need to pry it open after removing all of the screws.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-16.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-16.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 16\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_72fcf508ed\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/2\/21\/Build-a-Laptop-Computer-Step-16.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-16.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 16\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/2\/21\/Build-a-Laptop-Computer-Step-16.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-16.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 9\">9<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install your memory.<\/b> Once the panel is open, you will have access to the motherboard and memory slots. Insert the SO-DIMM memory chips into their slots at an angle, and then push them down to click them into place. The memory sticks can only be installed in one direction, so don\u2019t try to force them in.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-17.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-17.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 17\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_7fad8ff9b5\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/f5\/Build-a-Laptop-Computer-Step-17.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-17.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 17\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/f\/f5\/Build-a-Laptop-Computer-Step-17.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-17.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 10\">10<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install the CPU.<\/b> There may be a CPU lock around the socket where the CPU is installed. You may need to use a flathead screwdriver to turn it to the \u201cunlocked\u201d position.\n\n<ul><li>Turn your CPU over so that you can see the pins. There should be one corner that is missing pins. This notch will line up with the notch on the socket.<\/li>\n<li>The CPU will only fit in the socket one way. If the CPU does not seat itself, do not force it, or you may bend the pins, ruining the processor.<\/li>\n<li>Once the CPU is inserted, put the CPU lock into the \u201clocked\u201d position.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-18.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-18.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 18\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_3b42a31376\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/d\/d5\/Build-a-Laptop-Computer-Step-18.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-18.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 18\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/d\/d5\/Build-a-Laptop-Computer-Step-18.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-18.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 11\">11<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install the cooling fan (most laptops use centrifugal fans).<\/b> This fan either cools the CPU or cool the CPU and multiple other parts. Your CPU should have come packaged with a cooling fan. Most fans will have thermal paste already applied to the bottom where it connects with CPU. If the fan does not have any paste, you will need to apply some before the fan is installed.\n\n<ul><li>Once the paste has been applied, you can install the fan. The exhaust must line up with the vents on your shell. This part can be tricky as you attempt to line everything up. Don\u2019t try to force the heatsink and fan assembly in, but wiggle it in instead. There may be also some mounting bolts to place. If your case includes a bay to place fan dust filter, place a dust filter to prevent dust from clogging the heatsink.<sup id=\"_ref-1\" class=\"reference\" aria-label=\"Link to Reference 1\"><a href=\"#_note-1\">[1]<\/a><\/sup><\/li>\n<li>Keep the heatsink angled until you find the right position. This will help keep the thermal paste from getting all over your components.<\/li>\n<li>Attach the fan\u2019s power cable to the motherboard once the fan is installed. If you don\u2019t connect the fan, the laptop will overheat and turn off after a few minutes of use.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-19.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-19.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 19\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_d85969a49a\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/a\/a8\/Build-a-Laptop-Computer-Step-19.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-19.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 19\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/a\/a8\/Build-a-Laptop-Computer-Step-19.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-19.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 12\">12<\/div>\n<div class=\"step\">\n<b class=\"whb\">Close your panels.<\/b> Once you have installed all of the components, you can place the panels back over the openings and secure them with screws. Your laptop is complete!\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<\/ol><div class=\"clearall\">\n<\/div>\n<\/div>\n<div class=\"section steps   sticky\">\n<h3>\nPart 3\n<\/h3><\/div>\nStarting it Up\n\n<div id=\"steps_3\" class=\"section_text\">\n<ol class=\"steps_list_2\"><li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-20.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-20.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 20\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_9921b5de60\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/7\/70\/Build-a-Laptop-Computer-Step-20.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-20.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 20\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/7\/70\/Build-a-Laptop-Computer-Step-20.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-20.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 1\">1<\/div>\n<div class=\"step\">\n<b class=\"whb\">Make sure that the battery is inserted.<\/b> It is easy to forget the battery in the build process, but make sure that it is inserted and charging properly before booting up the computer.\n\n<\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-21.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-21.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 21\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_ba47f8c08a\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/97\/Build-a-Laptop-Computer-Step-21.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-21.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 21\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/9\/97\/Build-a-Laptop-Computer-Step-21.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-21.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 2\">2<\/div>\n<div class=\"step\">\n<b class=\"whb\"><a href=\"https:\/\/www.wikihow.com\/Test-PC-RAM-with-MemTest86\" title=\"Test PC RAM with MemTest86\">Check your memory<\/a>.<\/b> Before you install an operating system, run Memtest86+ to ensure that your memory is working correctly, and that your computer is functioning in general. Memtest86+ can be downloaded for free online, and can be booted from a CD or USB drive.\n\n<ul><li>You can also check that the memory you installed is recognized by using the <a href=\"https:\/\/www.wikihow.com\/Enter-BIOS\" title=\"Enter BIOS\">BIOS<\/a>. Find the Hardware or Monitor section to see if your memory appears.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-22.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-22.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 22\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_077b959e54\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/6a\/Build-a-Laptop-Computer-Step-22.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-22.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 22\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/6\/6a\/Build-a-Laptop-Computer-Step-22.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-22.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 3\">3<\/div>\n<div class=\"step\">\n<b class=\"whb\">Install an operating system.<\/b> For self-built laptops, you can choose between <a href=\"https:\/\/www.wikihow.com\/Install-Windows-8\" title=\"Install Windows 8\">Microsoft Windows<\/a> or a <a href=\"https:\/\/www.wikihow.com\/Install-Linux\" title=\"Install Linux\">Linux distribution<\/a>. Windows costs money, but it offers a larger range of programs and hardware compatibility. Linux is free, secure, and supported by a community of volunteer developers.\n\n<ul><li>There are many versions of Linux to choose from, but some of the more popular include <a href=\"https:\/\/www.wikihow.com\/Install-Ubuntu-Linux\" title=\"Install Ubuntu Linux\">Ubuntu<\/a>, Mint, and Debian.<\/li>\n<li>It is recommended that you install the latest version of Windows released, as older versions lose support after enough time has passed.<\/li>\n<li>If you don\u2019t have an optical drive installed, you will need to create a <a href=\"https:\/\/www.wikihow.com\/Make-a-USB-Bootable\" title=\"Make a USB Bootable\">bootable USB drive<\/a> with your operating system files.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<li class=\"hasimage final_li\">\n<div class=\"mwimg  largeimage  floatcenter\"><a href=\"#\/Image:Build-a-Laptop-Computer-Step-23.jpg\" class=\"image lightbox\" data-href=\"\/Image:Build-a-Laptop-Computer-Step-23.jpg?ajax=true&amp;aid=1111395\"><div class=\"content-spacer\">\n<img alt=\"Image titled Build a Laptop Computer Step 23\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" id=\"img_2d54f434f1\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/36\/Build-a-Laptop-Computer-Step-23.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-23.jpg\" \/><\/div>\n<noscript><img alt=\"Image titled Build a Laptop Computer Step 23\" src=\"https:\/\/www.wikihow.com\/images\/thumb\/3\/36\/Build-a-Laptop-Computer-Step-23.jpg\/aid1111395-v4-728px-Build-a-Laptop-Computer-Step-23.jpg\" width=\"728\" height=\"546\" class=\"whcdn content-fill\" \/><\/noscript><\/a><\/div>\n<div class=\"step_num\" aria-label=\"Step 4\">4<\/div>\n<div class=\"step\">\n<b class=\"whb\"><a href=\"https:\/\/www.wikihow.com\/Find-and-Update-Drivers\" title=\"Find and Update Drivers\">Install your drivers<\/a>.<\/b> Once your operating system is installed, you will need to install drivers for your hardware. Most modern operating systems will do most of this automatically, but you may have one or two components that need to be installed manually.\n\n<ul><li>Most components will come with their drivers on a disc. Use the disc if your operating system cannot find the correct drivers.<\/li>\n<\/ul><\/div>\n<div class=\"clearall\">\n<\/div><\/li>\n<\/ol><div class=\"clearall\">\n<\/div>\n<\/div>\n<div id=\"donate_section\" data-image_count=\"5\" data-image_number=\"\" data-non_profit_name=\"\">\n\n<div class=\"qa section sticky qa_show_unanswered_questions\" data-few_contribs=\"1\" data-search_enabled=\"0\" data-aid=\"1111395\">\n\n\t<h2>\n\t\t\n\t\tCommunity Q&amp;A\n\t<\/h2>\n\t<div id=\"qa\" class=\"section_text\">\n\t\t<div id=\"qa_answered_questions_container\">\n\t\t\t<div id=\"qa_aq_search_container\">\n\t\t\t\t<div id=\"qa_aq_search_heading\" class=\"qa_section_heading\">Search<\/div>\n\t\t\t\t<input class=\"qa_aq_search\" placeholder=\"Type your query here\" \/><\/div>\n\t\t\t<a id=\"qa_add_curated_question\">Add New Question<\/a>\n\t\t\t<ul id=\"qa_article_questions\" class=\"qa_questions qa_aqs qa_waittheresmore\"><li class=\"qa_aq qa_block\" data-aqid=\"146726\" data-cqid=\"146743\" data-caid=\"147599\" data-sqid=\"817894\" data-qa_inactive=\"0\" data-votes_up=\"52\" data-votes_down=\"12\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Is it less expensive than buying a laptop?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">This depends on the parts you use, but generally the price will be less because of manufacturing and miscellaneous costs on the pre-built laptop.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 12<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 52<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"146728\" data-cqid=\"146745\" data-caid=\"147601\" data-sqid=\"1020153\" data-qa_inactive=\"0\" data-votes_up=\"42\" data-votes_down=\"12\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">How long does it take to build a laptop?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">It can take anywhere from hours to days depending on your skills set, materials and the complexity of the machine which you are building. <\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 12<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 42<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"231147\" data-cqid=\"231164\" data-caid=\"232020\" data-sqid=\"1447930\" data-qa_inactive=\"0\" data-votes_up=\"43\" data-votes_down=\"13\" data-verifier_id=\"0\" data-submitter_id=\"4150320\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Will the processor overheat without a fan?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\"><a href=\"https:\/\/www.wikihow.com\/User:Aroly-Seven\" target=\"_blank\">Aroly Seven<\/a><\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Yes.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 13<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 43<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"97186\" data-cqid=\"97203\" data-caid=\"98059\" data-sqid=\"568273\" data-qa_inactive=\"0\" data-votes_up=\"40\" data-votes_down=\"12\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Can a laptop computer connect to WiFi?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Only if you installed a WiFi modem into it. Most laptop motherboards already have that installed, but if it doesn't, you can install your own or buy a USB WiFi stick. <\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 12<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 40<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"249414\" data-cqid=\"249431\" data-caid=\"250287\" data-sqid=\"1574677\" data-qa_inactive=\"0\" data-votes_up=\"7\" data-votes_down=\"1\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Can I make my own case using a laser cutter?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Yes. If you can do it well in fact it would be a better idea. But don't try if you're not skilled.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 1<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 7<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"136277\" data-cqid=\"136294\" data-caid=\"137150\" data-sqid=\"820675\" data-qa_inactive=\"0\" data-votes_up=\"26\" data-votes_down=\"10\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Can I add an HDMI and a USB port?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Those connectors in a laptop usually come with the motherboard, so how many USB ports you have, for instance, depends on which motherboard you choose.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 10<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 26<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"274721\" data-cqid=\"274738\" data-caid=\"275594\" data-sqid=\"1742700\" data-qa_inactive=\"0\" data-votes_up=\"9\" data-votes_down=\"2\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Do I need to have a charger for the computer's battery?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Yes, if you are building a laptop, you need a charger for the battery. <\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 2<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 9<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"42709\" data-cqid=\"42726\" data-caid=\"43582\" data-sqid=\"52946\" data-qa_inactive=\"0\" data-votes_up=\"34\" data-votes_down=\"18\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">How much would it generally cost me to build my own laptop?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">It varies. A cheap one would be around 200 to 300 dollars, but it would have a tiny hard drive and only 2 gigs of RAM. An expensive one would be over 1000 dollars. It really depends on what you buy.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 18<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 34<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"62063\" data-cqid=\"62080\" data-caid=\"62936\" data-sqid=\"357868\" data-qa_inactive=\"0\" data-votes_up=\"41\" data-votes_down=\"24\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Where are bare bone notebook computers sold?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Bare bone laptops can be bought online through an independent vendor, or in an electronics store, like Fry's or Best Buy. <\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 24<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 41<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t\t<li class=\"qa_aq qa_block\" data-aqid=\"260491\" data-cqid=\"260508\" data-caid=\"261364\" data-sqid=\"1669420\" data-qa_inactive=\"0\" data-votes_up=\"14\" data-votes_down=\"6\" data-verifier_id=\"0\" data-submitter_id=\"0\">\n\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\n\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t<div class=\"qa_question_tab\">Question<\/div>\n\t\t\t\t\t\t<div class=\"qa_q_txt question\">Can I just get a cheaper computer and replace certain parts to make it faster?<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t<div class=\"qa_answers\">\n\t\t\t\t\t\t\t<div class=\"qa_user_area\">\n\t\t\t\t\t\t\t\t<div class=\"qa_user_hover hint_box\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_badge\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_person_circle\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t<div class=\"qa_answerer_info\">\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_name\">wikiHow Contributor<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t\t\t<div class=\"qa_user_label\">Community Answer<\/div>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\n\t\t\t\t\t\t\t<div class=\"qa_answer answer\">Unfortunately, most laptops have their components soldered. Unless you have the proper tools (or go risky and use a hair dryer), it's impossible to replace the parts. It's doable, but not feasible.<\/div>\n\t\t\t\t\t\t\t<div class=\"qa_answer_footer\">\n\t\t\t\t\t\t\t\t<div class=\"wh_vote_container qa_vote_buttons\">\n\t\t\t\t\t\t\t\t\tThanks!\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_links\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box button secondary\">Yes<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box button secondary\">No<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<div class=\"qa_widget_vote_buttons\">\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_down wh_vote_box\">Not Helpful 6<\/a>\n\t\t\t\t\t\t\t\t\t\t<a href=\"#\" class=\"wh_vote_up wh_vote_box\">Helpful 14<\/a>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div><\/div><\/div><\/div><\/li>\n\t\t\t<\/ul><a id=\"qa_show_more_answered\">Show more answers<\/a>\n\t\t<\/div>\n\t\t<div class=\"qa_divider qa_no_unpatrolled_questions\">\n\t\t<div class=\"qa_divider\">\n\n\t\t\t\t<div id=\"qa_submitted_questions_container\">\n\t\t\t\t\t<div class=\"qa_section_heading\">Unanswered Questions<\/div>\n\t\t\t\t\t<ul id=\"qa_submitted_questions\" class=\"qa_questions\"><li class=\"qa_sq\" data-sqid=\"1515263\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">How to mark a file on laptop?<\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t\t\t<li class=\"qa_sq\" data-sqid=\"2618931\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">How do I put the screen on the Laptop? I ask this because in the steps there wasn't anything of putting the screen.<\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t\t\t<li class=\"qa_sq\" data-sqid=\"2513999\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">What is the best shell to use?<\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t\t\t<li class=\"qa_sq\" data-sqid=\"2140689\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">How do I know which size shell to get band where the USB and charging ports will be?<\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t\t\t<li class=\"qa_sq\" data-sqid=\"1789854\">\n\t\t\t\t\t\t\t\t<div class=\"qa_divider\">\n\t\t\t\t\t\t\t\t<div class=\"qa_li_container\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_bullet\">\n\t\t\t\t\t\t\t\t\t<div class=\"qa_sq_question\">\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_q_txt\">What safety issues are there besides grounding and not plugging into an A\/C outlet should I worry about when building a laptop computer?<\/div>\n\t\t\t\t\t\t\t\t\t\t<div class=\"qa_editor_tools\">\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_edit_submitted button secondary\">Answer this question<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_flag_submitted\" data-type=\"sq_flag\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<a class=\"qa_ignore_submitted\" data-type=\"sq_ignore\">Flag as...<\/a>\n\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div><\/div><\/li>\n\t\t\t\t\t<\/ul><div id=\"qa_submitted_spinner\" class=\"qa_spinner\">\n\t\t\t\t\t\n\t\t\t\t\t\t<a id=\"qa_show_more_submitted\">Show more unanswered questions<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"qa_divider\">\n\n\t\t<div class=\"qa_section_heading\" id=\"qa_ask_heading\">Ask a Question<\/div>\n\t\t<div id=\"qa_submitted_question_form\">\n\t\t\t<input id=\"qa_asked_question\" type=\"text\" maxlength=\"200\" placeholder=\"What do you need to know? We'll do our best to find the answer.\" \/><div id=\"qa_asked_count\">200 characters left<\/div>\n\t\t\t<input id=\"qa_email\" type=\"text\" placeholder=\"Email address (optional)\" \/><div id=\"qa_email_prompt\">Include your email address to get a message when this question is answered.<\/div>\n\t\t\t<a id=\"qa_submit_button\" class=\"button primary\">Submit<\/a>\n\t\t<\/div>\n\n\t\t\t\n\t\t\t<div id=\"qa_flag_options\" data-type=\"\">\n\t\t\t<div class=\"hint_box\">\n\t\t\t\t<ul><li><a href=\"#\" data-option=\"answered\">Already answered<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#\" data-option=\"not_question\">Not a question<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#\" data-option=\"bad_question\">Bad question<\/a><\/li>\n\t\t\t\t\t<li><a href=\"#\" data-option=\"other\">Other<\/a><\/li>\n\t\t\t\t<\/ul><\/div>\n\t\t<\/div>\n\t\t<div id=\"qa_answer_flag_options\" data-type=\"\">\n\t\t\t<div class=\"hint_box\">\n\t\t\t\t<ul><\/ul><\/div>\n\t\t<\/div>\n\n\t\t\n\t\t\n<\/div>\n<\/div>\n<div id=\"qa_box\">\n\n\n<div class=\"section sourcesandcitations  sticky\">\n<h2>Sources and Citations<\/h2>\n<div id=\"sourcesandcitations\" class=\"section_text\">\n<div class=\"references-small\">\n<ol class=\"references sources\"><li id=\"_note-1\">\n<a href=\"#_ref-1\" target=\"_blank\">\u2191<\/a> <a rel=\"nofollow\" class=\"external free\" href=\"http:\/\/www.tomsguide.com\/us\/diy-laptop-whitebook,review-1286-10.html\" target=\"_blank\">http:\/\/www.tomsguide.com\/us\/diy-laptop-whitebook,review-1286-10.html<\/a>\n<\/li>\n<\/ol><\/div>\n<div class=\"clearall\">\n<\/div>\n<\/div>\n\n\n\n\n\n\n<div id=\"uci_section\" class=\"section uci_section_desktop sticky\">\n<h2 class=\"section_head nouciimages\" id=\"uci_header\">\nMade Recently\n\t\t<\/h2>\n\t<div id=\"uci_images_section\" class=\"uci_images_desktop section_text\">\n\t\t<div id=\"uci_images\">\n\t\t\t\t\t\t<div id=\"uci_fileinput_square_wrapper\" class=\"uci_thumbnail uci_empty\">\n\t\t\t\t<div id=\"uci_fileinput_square\" class=\"uci_fileinput_square_bg\">\n\t\t\t\t\t<div id=\"uci_fileinput_spin\">\n\t\t\t\t\t<div class=\"uci_fileinput_loading\">Loading...<\/div>\n\t\t\t\t\t<div class=\"uci_fileinput_center\">\n\t\t\t\t\t\tDid you try these steps?<br \/>Upload a picture for other readers to see.\t\t\t\t\t\t<div id=\"uci_upload_response_error\">Upload error<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t\t<input id=\"uci_fileupload\" class=\"op-action\" type=\"file\" name=\"wpUploadImage\" aria-label=\"Link to choose an image to upload\" \/><div id=\"files\" class=\"files\">\n\t\t\t\t\t<div class=\"uci_thumbnail_description\">\n\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n\t\t<div id=\"uci_upload_response\">\n\t\t<div id=\"uci_userreview_cta\" data-image=\"\">Awesome picture! Tell us more about it? <a href=\"#\">Click here to share your story.<\/a>\n<\/div>\n\t<\/div>\n<\/div>\n\n<\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:44:54+0000","updated_at":"2018-11-07T16:44:54+0000","published_at":"2010-11-05T00:00:00+0000","published_by":["wikiHow"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"www.wikihow.com","preview_picture":"https:\/\/www.wikihow.com\/images\/3\/36\/Build-a-Laptop-Computer-Step-23.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":328,"title":"r\/Clojure - Clojure for command line scripts idea. Feasibility?","url":"https:\/\/www.reddit.com\/r\/Clojure\/comments\/3qb9nz\/clojure_for_command_line_scripts_idea_feasibility\/","content":"<p class=\"yklcuq-10 hpxQMr\">Hi all,<\/p><p class=\"yklcuq-10 hpxQMr\">So ever since I began learning Clojure I've been thinking how I would love to use it as an alternative to bash \/ python \/ etc. for writing command line scripts. As we all know unfortunately, Clojure has a pretty long startup time which makes this pretty far from ideal for shorter scripts. An idea popped into my head recently though that I think could address this, but I'm not familiar enough with Clojure to attempt to go about properly implementing it<\/p><p class=\"yklcuq-10 hpxQMr\">My thinking is that rather than start up a new instance of the JVM to execute the Clojure script, have a single long running process that you forward scripts over to in order to be executed in a sandboxed environment. Basically:<\/p><ul class=\"yklcuq-12 kuqpqj\"><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">User executes: clj-bash myscript.clj<\/p><\/li><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">Small C program checks to see if long running Clojure\/JVM instance is up. If it is not, start it and wait<\/p><\/li><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">C program sends the script over to the long running instance<\/p><\/li><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">Long running instance executes the script in its own sandboxed environment and then sends back the results<\/p><\/li><\/ul><p class=\"yklcuq-10 hpxQMr\">From what I've seen in Clojure so far it seems like this is quite possible, the difficulty would come from a few things:<\/p><ul class=\"yklcuq-12 kuqpqj\"><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">Sandboxing the scripts so you don't get things like namespace collisions in the case of the same script (e.g. some makefile type thing) being run in two different directories at the same time.<\/p><\/li><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">Getting user input (readline) and such<\/p><\/li><li class=\"yklcuq-11 godfjz\"><p class=\"yklcuq-10 hpxQMr\">Handling errors gracefully so if one script breaks horribly, it doesn't bring down the server \/ any other running scripts<\/p><\/li><\/ul><p class=\"yklcuq-10 hpxQMr\">Would someone with more experience in Clojure be able to tell me if this is really even a viable idea? And if so, what should I start researching or which existing libraries are out there that might help me in accomplishing this?<\/p>","created_at":"2018-11-07T16:44:53+0000","updated_at":"2018-11-07T16:44:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.reddit.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":327,"title":"Best investments - 10 steps to investing money | uSwitch","url":"https:\/\/www.uswitch.com\/investments\/best-investment\/","content":"<p>When it comes to choosing what investment to go for, a one-size-fits-all approach just doesn\u2019t work \u2013 there isn\u2019t a single investment product that will work for everyone.<\/p><p>We all have a unique set of personal circumstances and individual savings goals, and it\u2019s only when you take a closer look at these that you can begin to answer the question \u2018what is the best investment for me?\u2019<\/p><h2>1. Investing money \u2013 how much do you have to invest?<\/h2><p>Are you looking to invest a lump sum, or to set aside a regular monthly amount? Maybe a short term investment or longer? And how much money do you have available? <\/p><p>Certain assets require a lump sum investment, such as\u00a0<a href=\"https:\/\/www.uswitch.com\/investments\/bonds-and-gilts\/\">corporate bonds<\/a>\u00a0or when you\u2019re putting down a deposit to purchase a\u00a0<a href=\"https:\/\/www.uswitch.com\/investments\/property-investment\/\">property<\/a>, and others offer the flexibility of either lump sum or regular contributions, such as a\u00a0<a title=\"Compare cash ISAs\" href=\"https:\/\/www.uswitch.com\/savings-isa\/\">cash ISA<\/a>\u00a0or\u00a0<a title=\"Find out more about stocks and shares ISAs\" href=\"https:\/\/www.uswitch.com\/investments\/stocks-and-shares-isa\/\">stocks and shares ISA<\/a>.<\/p><p>Some <a title=\"investments\" href=\"https:\/\/www.uswitch.com\/investments\/\">investments<\/a> also have a minimum financial commitment, so knowing what you can afford and whether you plan to make a one-off or an ongoing saving is a good starting point.<\/p><h2>2. How long do you want to invest money for?<\/h2><p>Or, put another way \u2013 when will you need access to your money? Certain investment products run for a fixed period of time, so if you have a specific date in mind as to when you need access to your\u00a0<a href=\"https:\/\/www.uswitch.com\/investments\/investment-glossary\/#step7\" target=\"_blank\">capital<\/a>, then some product types won\u2019t be right for you. In addition, certain investments, such as\u00a0<a title=\"Find out more about stocks and shares\" href=\"https:\/\/www.uswitch.com\/investments\/stocks-and-shares\/\">shares<\/a>, are much longer lasting and shouldn\u2019t be considered as short term investments.<\/p><p>That\u2019s because although shares have historically increased in value over the long term, they can fluctuate in value in the short term. It\u2019s recommended that you invest money at least five years to be in a good position to ride out these fluctuations.<\/p><h2>3. What are you planning to use the money for?<\/h2><p>We all have different reasons for saving, and the purpose of your investment can affect how much\u00a0<a title=\"Find out more about investment risk\" href=\"https:\/\/www.uswitch.com\/investments\/investment-risk\/\">risk<\/a> you\u2019re prepared to take with your money. If your investment is to pay for your children\u2019s education, then you may be investing over a long period of time, and looking for a higher return, as a result you may be inclined to choose a higher-risk investment option.<\/p><p>Conversely, if you\u2019re investing money to pay for an overseas trip, or a new car, you may be investing for a short period of time and want certainty about the outcome of your investment, and you may feel more comfortable with lower risk short term investments.<\/p><h2><img class=\"us-img--full\" title=\"Choosing the best investment\" src=\"https:\/\/assets0.uswitch.com\/s3\/uswitch-wp-cms-assets\/wp-content\/uploads\/2014\/06\/iStock_000019065788_Medium-copy-3.jpg\" alt=\"Choosing the best investment\" width=\"640\" height=\"250\" \/><\/h2><h2>4. Do you need an income from your investment?<\/h2><p>If you\u2019re looking for a regular income from your investment then this will influence your choice of product. A pension is probably the best-known investment vehicle for providing an income in retirement.<\/p><p>There are other investment products available that can also provide a regular income such as\u00a0<a href=\"https:\/\/www.uswitch.com\/investments\/investment-glossary\/#step2\" target=\"_blank\">annuities<\/a>, or corporate bond funds, alternatively you could choose to invest in a buy-to-let property to provide you with a rental income.<\/p><p>Consumers may wish to seek professional advice first before taking out such products as they often require a huge commitment.<\/p><h2>5. What age are you?<\/h2><p>Attitude to risk can change with age. Longer term, higher risk investment options may be more attractive to someone in their thirties than to someone who is getting close to retirement.<\/p><p>People tend to invest money in lower-risk products as their retirement approaches.<\/p><h2>6. What are your personal circumstances?<\/h2><p>If you\u2019re a parent with financially dependent children, then you\u2019re probably going to be more cautious with your savings than someone who\u2019s single and doesn\u2019t have any dependants, and therefore more likely to choose a low to medium risk and possibly short term investment. For someone who\u2019s self-employed the priority may be finding a product that allows flexible contributions to suit a more erratic income pattern.<\/p><p>It\u2019s important that you to take a close look at your circumstances and how they affect what\u00a0investment you opt for before\u00a0you make a commitment.<\/p><h2>7. Do you have other investments?<\/h2><p>If you already have a number of investments and feel that your future financial requirements are well taken care of, then you may be willing to take a higher risk with your next investment. However, if this is your first and only investment then you may be more conservative in your choice.<\/p><h2>8. What are your values?<\/h2><p>It\u2019s important that you feel comfortable with where your money is going, so if you have strong beliefs then it\u2019s worth seeking out an investment that fits with these. There are a number of green and\u00a0<a href=\"https:\/\/www.uswitch.com\/investments\/ethical-investments\/\">ethical investments<\/a>\u00a0\u00a0available as well as investments that are designed for specific cultural groups.<\/p><h2>9. What\u2019s your risk profile?<\/h2><p>How do you feel about\u00a0<a title=\"Find out more about investment risk\" href=\"https:\/\/www.uswitch.com\/investments\/investment-risk\/\">investment risk<\/a>? Not everyone is happy riding out the ups and downs of the stock market, and if the thought of a particular investment makes you lie awake at night, then it\u2019s probably too risky for you.<\/p><p><img class=\"alignnone wp-image-12565 us-img--full\" src=\"https:\/\/assets0.uswitch.com\/s3\/uswitch-wp-cms-assets\/wp-content\/uploads\/2014\/08\/money-cost-up-down-finance-interest-rates.jpg\" alt=\"best investment options - where to invest\" width=\"281\" height=\"281\" \/><\/p><h2>10. How much flexibility do you need?<\/h2><p>When you invest money, it gets tied up and is no longer easily accessible. But, if you have a sudden need for cash how quickly and easily can you liquidate your asset? And what\u2019s the penalty for doing this?<\/p><p>If you think this may be an important factor for you then it\u2019s worth knowing up front what the implication of getting out of an investment early is \u2013 if indeed it\u2019s possible.<\/p><p>Once you\u2019ve answered these questions you\u2019ll have a better idea of the type of investments that would suit you. Financial advisers tend to recommend having a\u00a0<a href=\"https:\/\/www.uswitch.com\/investments\/investment-glossary\/#step21\" target=\"_blank\">portfolio<\/a>\u00a0of investments, that way if one investment performs badly, you have others to fall back on. It also means you can plan so you have short term investments as well as long-term ones.<\/p><p>Your tax status is another important consideration. You need to factor in the tax implications of each investment option based on your personal circumstances \u2013 that way you\u2019ll get a true picture of the return you\u2019re likely to make. Find out more about\u00a0<a title=\"Find out more about investment and tax\" href=\"https:\/\/www.uswitch.com\/investments\/investment-tax\/\">investment and tax<\/a>.<\/p><p>You may need to consult an appropriate professional \u2013 financial adviser, accountant or tax specialist \u2013 about the tax implications of any particular investment in relation to your own circumstances.<\/p><p>It\u2019s also a good idea to review your portfolio on an annual basis after you\u2019ve invested money. Circumstances change, so it makes sense to check that it still represents the best investment for you.<\/p>","created_at":"2018-11-07T16:44:52+0000","updated_at":"2018-11-07T16:44:52+0000","published_by":["Jafar Hassan"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.uswitch.com","preview_picture":"https:\/\/assets0.uswitch.com\/s3\/uswitch-includes-production\/img\/logo_196-e91b5a5465.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":325,"title":"How to make extra money: 42 ideas that generate cash","url":"https:\/\/www.iwillteachyoutoberich.com\/make-extra-money\/","content":"<img width=\"800\" height=\"453\" src=\"https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/earn-extra-money-e1501542999693-800x453.png\" class=\"alignfull wp-post-image\" alt=\"\" \/><p>Learning how to make extra money is a great way to start living a Rich Life.<\/p><p>That\u2019s because it adheres to one key tenet about finances:<\/p><div style=\"text-align: center;\"><strong>There\u2019s a limit to how much you can save \u2014 but there\u2019s no limit to how much you can earn.<\/strong><\/div><p>Once you understand that your earning potential is limitless, you can start working toward the life you want today.<\/p><p>That\u2019s why I decided to update this article on how to make extra money for 2018. I want to give you my two favorite ways to make extra money \u2014 as well as a list of 42 side hustle ideas you can start today.<\/p><ul><li><u><a href=\"#1\">How to make extra money \u2014 the two best ways<\/a><\/u>\n<ul><li><u><a href=\"#2\">Earn more money at your job<\/a><\/u><\/li>\n<li><u><a href=\"#3\">Start a side hustle<\/a><\/u><\/li>\n<\/ul><\/li>\n<li><u><a href=\"#4\">Crush mental barriers to earning extra money<\/a><\/u>\n<ul><li><u><a href=\"#5\">Mental barrier #1: I don\u2019t have the time.<\/a><\/u><\/li>\n<li><u><a href=\"#6\">Mental barrier #2: But I\u2019m so tired after work.<\/a><\/u><\/li>\n<li><u><a href=\"#7\">Mental barrier #3: I don\u2019t have a good idea.<\/a><\/u><\/li>\n<\/ul><\/li>\n<li><u><a href=\"#8\">42 ways to make extra money on the side<\/a><\/u>\n<ul><li><u><a href=\"#9\">Side hustle ideas for beginners<\/a><\/u><\/li>\n<li><u><a href=\"#10\">Side hustle ideas for college students<\/a><\/u><\/li>\n<li><u><a href=\"#11\">Side hustle ideas for millennials<\/a><\/u><\/li>\n<li><u><a href=\"#12\">Side hustle ideas for stay-at-home parents<\/a><\/u><\/li>\n<li><u><a href=\"#13\">Side hustle ideas for working parents<\/a><\/u><\/li>\n<li><u><a href=\"#14\">Side hustle ideas for teachers<\/a><\/u><\/li>\n<\/ul><\/li>\n<li><u><a href=\"#15\">Need more ideas?<\/a><\/u><\/li>\n<\/ul><p>Let\u2019s get started.<\/p><h2 id=\"1\">How to make extra money \u2014 the two best ways<\/h2><p>There are plenty of ways to <u><a href=\"http:\/\/www.iwillteachyoutoberich.com\/blog\/how-to-make-money-fast\/\" target=\"_blank\">make more money<\/a><\/u>\u00a0\u2014 but my two favorite ways are:<\/p><ol><li>Earn more money at your job<\/li>\n<li>Start a side hustle<\/li>\n<\/ol><h3 id=\"2\">Earn more money at your job<\/h3><p>Salary negotiations are magic.<\/p><p>Seriously. That\u2019s because with a single conversation, you can make <em>thousands <\/em>more than you did before. <\/p><p>The best part? Your gains only add up year-after-year. <\/p><p>Check out how much you can earn over the course of 40 years with a single salary increase:<\/p><div style=\"text-align: center;\"><img width=\"700\" src=\"https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/one-of-the-best-ways-to-make-money-is-asking-for-a-raise-1024x601-2.png\" class=\"wp-image-26757\" alt=\"one of the best ways to make money is asking for a raise 1024x601 2\" srcset=\"https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/one-of-the-best-ways-to-make-money-is-asking-for-a-raise-1024x601-2.png 1024w, https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/one-of-the-best-ways-to-make-money-is-asking-for-a-raise-1024x601-2-300x176.png 300w, https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/one-of-the-best-ways-to-make-money-is-asking-for-a-raise-1024x601-2-768x451.png 768w, https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/one-of-the-best-ways-to-make-money-is-asking-for-a-raise-1024x601-2-106x62.png 106w, https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/one-of-the-best-ways-to-make-money-is-asking-for-a-raise-1024x601-2-101x60.png 101w\" \/><\/div><p>Negotiations can be <em>terrifying<\/em>,<em>\u00a0<\/em>though \u2014 especially if you\u2019ve never done it before.<\/p><p>Luckily, I have a system that\u2019ll help make it a little less terrifying while increasing your chances of nailing them and getting the salary you deserve.<\/p><p><strong>Three months before negotiations<br \/><\/strong>Before you even think about getting into your boss\u2019s office and hashing out what type of company car you want them to give you, you need to start tracking everything you do at work <em>and the results you get. <\/em><\/p><p>If you were on a team that sold 50,000 widgets, figure out what you did to help make that happen and quantify it. <\/p><p>By tracking all of your accomplishments and giving a hard number to them, you can adequately make your case as to 1) Why you\u2019re an indispensable employee and 2) Why you deserve to earn more. <\/p><p>During this time, ask your boss if you can schedule a time to sit down and discuss ways you can excel at work. Make it clear you want to exceed expectations and ask what that would entail. <\/p><p><strong>Two months before negotiations<\/strong><br \/>Now it\u2019s time to meet with your boss again and show him your tracking from the previous month \u2026 and ask him what you can do better.<\/p><p>You want to make sure you\u2019re on the right track with your work. More importantly, you\u2019re communicating your results. This shows your boss you\u2019re willing to take the initiative to make your work and therefore the company better.<\/p><p>During this time, you\u2019re also going to want to prep the Briefcase Technique. This is a tried-and-true system my students have used to earn thousands in immediate salary gains. <\/p><p>Here\u2019s how it works: You\u2019re going to create a one- to five-page proposal document showcasing specific areas in the company wherein you can add more value. <\/p><p>Then you\u2019re going to bring that proposal with you when you negotiate your salary. When compensation discussions come up, you\u2019re going to pull out this document and outline how you\u2019re going to solve the company\u2019s issues. <\/p><p>Check out the seven-minute video below for more. \u00a0 <\/p><iframe src=\"https:\/\/www.youtube.com\/embed\/3p28MFt8RBA?wmode=transparent&amp;modestbranding=1&amp;autohide=1&amp;showinfo=0&amp;rel=0\" width=\"100%\" height=\"100%\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p><strong>One month before negotiations<br \/><\/strong>If you haven\u2019t already, now\u2019s the time to mention to your boss that you\u2019d like to discuss compensation at a meeting next month.<\/p><p>Ask what you\u2019ll need to bring to make it a fruitful discussion and listen carefully to what your boss has to say.<\/p><p>During this time, it\u2019s time to do some research and practice.<\/p><p>If you don\u2019t have a hard number when you walk into salary negotiations with your boss or hiring manager, you\u2019re going to be at the mercy of someone who does this for a living and will control the conversation. <\/p><p>To that end, there are a few great resources for you to find a good number to start with:<\/p><ul><li>\n<strong><u><a href=\"http:\/\/www.salary.com\" target=\"_blank\">Salary.com<\/a><\/u><\/strong>:<strong>\u00a0<\/strong>This is a great website for both employers and job seekers to compare compensation rates for specific jobs across a huge variety of companies. <\/li>\n<li>\n<strong><u><a href=\"http:\/\/glassdoor.com\" target=\"_blank\">Glassdoor.com<\/a><\/u><\/strong>:\u00a0Though this site primarily acts as a \u201cYelp for jobs,\u201d it also includes an incredibly handy salary tool that allows you to look at the national average salary for your job as well as the average rate of compensation in your city. <\/li>\n<li>\n<strong><u><a href=\"https:\/\/www.payscale.com\/\" target=\"_blank\">PayScale.com<\/a><\/u><\/strong>:<strong>\u00a0<\/strong>This website sends you a personalized salary report based on a questionnaire you fill out regarding your career history. It\u2019s especially great for <u><a href=\"http:\/\/www.payscale.com\/college-salary-report\" target=\"_blank\">recent college grads<\/a><\/u>. <\/li>\n<li>\n<strong>Ask a friend<\/strong>:<strong>\u00a0<\/strong>Do you know anyone who has worked in this field before? Maybe a friend of yours has been in the industry for a few years. Ask them how much they were paid \u2014 as well as advice on how much you should ask for as well. \u00a0<\/li>\n<\/ul><p>Now it\u2019s time to practice, practice, practice, and practice some more.<\/p><p>After all, it\u2019s one thing to research and read about negotiations. It\u2019s another thing entirely to do it live and under pressure.<\/p><p>To practice effectively, you\u2019ll want to first brainstorm as many different discussion scenarios that might occur with your boss as you can. Then run through each of those scenarios with a friend and play them out. <\/p><p>For example:<\/p><ul><li>Your boss acts surprised or annoyed when you bring up a higher salary.<\/li>\n<li>Your boss asks you to name a number first.<\/li>\n<li>He gives you a BS excuse like \u201cIt\u2019s the economy\u201d or \u201cEveryone else is getting the same thing.\u201d<\/li>\n<\/ul><p>Then observe or have your friend give you feedback on the following:<\/p><ul><li>\n<strong>Your words.<\/strong>\u00a0They should be compelling and concise.<\/li>\n<li>\n<strong>Your body language. <\/strong>You want to be sitting up, attentive, but relaxed.<\/li>\n<li>\n<strong>Your tone. <\/strong>It should be professional, positive, and energetic. <\/li>\n<\/ul><p>Practice up until the day of negotiations. Once there, you\u2019ll be as ready as you\u2019ll ever be to negotiate your salary. <\/p><p>In my 6-minute video below, I go more in-depth on how exactly you can absolutely crush your salary negotiations. Check it out.<\/p><iframe src=\"https:\/\/www.youtube.com\/embed\/d3qfJeLQDoY?wmode=transparent&amp;modestbranding=1&amp;autohide=1&amp;showinfo=0&amp;rel=0\" width=\"100%\" height=\"100%\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><h3 id=\"3\">Start a side hustle<\/h3><p>Salary negotiations are great, but my all-time favorite way of making extra money is by <strong>starting a side hustle<\/strong>.<strong>\u00a0<\/strong><\/p><p>What\u2019s a side hustle, you ask?<\/p><p>Simply put, side hustles are freelance businesses that can generate you cash outside of your normal work hours like\u2026 <\/p><ul><li>Graphic design<\/li>\n<li>Copywriting and editing<\/li>\n<li>Web development and programming<\/li>\n<li>Whatever!!<\/li>\n<\/ul><p>The best part is freelance businesses are easily scalable, so you can make a lot of money if you\u2019re willing to devote just a little bit of time to it. Or if you get busy in other areas of your life, you can scale back. <\/p><h2>The 4 steps to make extra money on the side<\/h2><h3>Step 1:<strong>\u00a0Find a profitable business idea<\/strong><br \/><\/h3><p>As mentioned above, finding a profitable business idea is as simple as looking at your strengths. You can do that by answering four simple questions today and finding a side hustle that\u2019s perfect to you.<\/p><ol><li>\n<strong>What do you already pay for? \n<\/strong>We already pay people to do a lot of different things. Can you turn one of those things into your own online business?\n<p>Examples: Clean your home, walk your pet, cook you meals, etc.\n<\/p><\/li>\n<li>\n<strong>What skills do you have? \n<\/strong>Now, what do you know \u2014 and know <em>well<\/em>? These are the skills you have that you\u2019re great at \u2014 and people want to pay you to teach them.\n<p>Examples: Fluency in a foreign language, programming knowledge, cooking skills, etc.\n<\/p><\/li>\n<li>\n<strong>What do your friends say you\u2019re great at?\n<\/strong>I love this question. Not only can it be a nice little ego boost \u2014 but it can also be incredibly revealing.\n<p>Examples: Workout routines, relationship advice, great fashion sense, etc.\n<\/p><\/li>\n<li>\n<strong>What do you do on a Saturday morning?\n<\/strong>What do you do on a Saturday morning before everyone else is awake? This can be incredibly revealing to what you\u2019re passionate about and what you like to spend your time on.\n<p>Examples: Browsing fashion websites, working on your car, reading fitness subreddits, etc.<\/p><\/li>\n<\/ol><p>Find an answer to those questions and I promise you you\u2019ll find a profitable business idea.<\/p><h3>Step 2:<strong>\u00a0Use the the Craigslist Penis Effect to find clients<\/strong><br \/><\/h3><p>If you\u2019ve never heard of the <u><a href=\"http:\/\/www.iwillteachyoutoberich.com\/blog\/the-craigslist-penis-effect\/\" target=\"_blank\">Craigslist Penis Effect<\/a><\/u>\u00a0(CPE) before, let me say now that it\u2019s NOT about posting unsolicited and unwanted photos of yourself on Craigslist. That\u2019s an easy way to land yourself on a list you don\u2019t want to be on.<\/p><p>The CPE actually describes situations where everyone else is so horrible that, by being even half-decent, you can dominate everyone else and succeed.<\/p><p>Think of those idiots on Craigslist trying to pick up women in the personals section by sending out a picture of their mediocre manhood. They\u2019d be better served writing a few half-decent responses, testing to see which got the best response, and sending that out instead. <\/p><p>That\u2019s exactly what you can start doing to generate effective leads.<\/p><p>Check out the jobs section of Craigslist and take a look at the ads relevant to your business. <\/p><p>If you took the time to write a great email reaching out to these companies, you\u2019ll immediately separate yourself from 99.9999% of others on the site who are just sending boring boilerplate emails that\u2019ll get them nowhere.<\/p><p>This doesn\u2019t apply to just Craigslist either \u2014 you can use this with any job site. The key is to just be slightly better than average and willing to test a few different emails.<\/p><p>Here are a few suggestions of great sites freelancers can use to find business:<\/p><ul><li>\n<strong>Writers: <\/strong><u><a href=\"http:\/\/www.mediabistro.com\" target=\"_blank\">MediaBistro.com<\/a><\/u>, <u><a href=\"http:\/\/upwork.com\" target=\"_blank\">Upwork.com<\/a><\/u>, <u><a href=\"http:\/\/freelancewritinggigs.com\" target=\"_blank\">FreelanceWritingGigs.com<\/a><\/u>\u00a0<\/li>\n<li>\n<strong>Illustrators\/Designers: <\/strong><u><a href=\"https:\/\/99designs.com\/\" target=\"_blank\">99designs.com<\/a><\/u>, <u><a href=\"http:\/\/www.designs.net\/\" target=\"_blank\">Designs.net<\/a><\/u>\n<\/li>\n<li>\n<strong>Programmers: <\/strong><u><a href=\"https:\/\/www.toptal.com\/\" target=\"_blank\">Toptal.com<\/a><\/u>, <u><a href=\"https:\/\/www.gun.io\/\" target=\"_blank\">Gun.io<\/a><\/u>\n<\/li>\n<\/ul><h3>Step 3:<strong>\u00a0Use the 5-line email pitch to qualify your leads<\/strong><br \/><\/h3><p>After you find a few good leads on potential clients, you\u2019ll need to craft a pitch that is tailored to their needs and concerns.<\/p><p>In doing so, you\u2019ll want to stress the BENEFITS of working with you \u2014 while not giving away too much information as to how you\u2019ll help. I lost my fair share of clients while I was starting out by giving away exactly how I\u2019d help them in my initial reach out. Stupid. <\/p><p>Because I\u2019m a nerd, I\u2019ve developed a 5-line email template you can use to send the perfect email pitch. It includes:<\/p><ol><li>\n<strong>The introduction, <\/strong>You\u2019re going to want to build rapport by introducing yourself and how you know about the client.<\/li>\n<li>\n<strong>The offer. <\/strong>Talk about them. What do you want to do for them? Why are you good for that role? You\u2019re going to want to do some research on the organization to see what they need help with.<\/li>\n<li>\n<strong>The benefit. <\/strong>Walk them through how your work will benefit their company. Are you going to free up more time for them? Are you going to maximize profits by X amount?<\/li>\n<li>\n<strong>The foot-in-the-door. <\/strong>This is a classic technique that utilizes an old psychology trick to get the client to agree to a small agreement so you can ask for a larger agreement later. <\/li>\n<li>\n<strong>The call to action. <\/strong>Be clear with this and ask them if they would like to proceed. The call to action is a critical part of this script.<\/li>\n<\/ol><p>When it\u2019s all put together, it\u2019ll look something like this:<\/p><blockquote>\n<p><em>CLIENT\u2019S NAME,<\/em><\/p>\n<p><em><strong>[Introduction] <\/strong>I read your article about X and noticed that you\u2019ve recently started using videos on your website. <\/em><\/p>\n<p><em><strong>[The offer]<\/strong>\u00a0I\u2019ve been doing video editing for three years and I\u2019d like to offer to help you edit your videos and get them optimized for the web.<\/em><\/p>\n<p><em><strong>[The benefit] <\/strong>That would make them look more professional and load faster, which is important for your readers. And you\u2019d free up time that you could use to create new content.<\/em><\/p>\n<p><em><strong>[The foot-in-the-door] <\/strong>We can discuss the details, of course, but first I wanted to see if this is something you might be interested in.<\/em><\/p>\n<p><em><strong>[The call to action] <\/strong>If so, would it be okay if I sent you a few ideas on how to help?<\/em><\/p>\n<p><em>Best,<\/em><\/p>\n<p><em>Ramit Sethi<\/em><\/p>\n<\/blockquote><h3>Step 4:<strong>\u00a0The Tuner Strategy to charge the perfect price<\/strong><br \/><\/h3><p>People always get tripped up about pricing, which is why I always hear questions like:<\/p><ul><li>How much should I be charging?<\/li>\n<li>Is X amount too little\/too much?<\/li>\n<li>Should I be charging by the hour or per project?<\/li>\n<\/ul><p>These are questions every freelancer asks when they start out. And while there are no set rules for rates, there are actually a few methods you can use to find one that works for you.<\/p><ol><li>\n<strong>Drop Three Zeros Method\n<\/strong>Simply take your ideal (read: realistic) salary, drop three zeros from it, and voila, you have your hourly rate!\n<p>For example, say you\u2019d really like to earn at least $40,000. Just take out the three zeros from the end and you now have your rate: $40\/hour. \n<\/p><\/li>\n<li>\n<strong>Double your \u201cresentment number\u201d\n<\/strong>I love this one because it\u2019s both really interesting and effective. Ask yourself: What\u2019s the lowest rate you\u2019ll work for that\u2019ll leave you resentful of your work?\n<p>Say you\u2019ll work for $15\/hour at the VERY LEAST. Just double that number so now you\u2019ll earn $30\/hour.\n<\/p><\/li>\n<li>\n<strong>Do what the next guy does\n<\/strong>This method is incredibly simple: Go to Google and search for the average hourly rate for whatever service you\u2019re providing. You\u2019ll get a good sense of where to start when you\u2019re charging your clients.<\/li>\n<\/ol><p>The best part is after you start charging your clients, you can start to take on more or less work until you earn the amount you want.<\/p><p>For example, after you earn your first $1,000, it\u2019s incredibly easy to start dialing your prices up and charge even more money from your clients. <\/p><p>Start \u201ctuning\u201d your rates after your first few clients. Were you making $30\/hour? Start charging $40 or even $50. There\u2019s no hard and set rule for how much you should charge. Just start tuning until you find a rate you\u2019re happy with. <\/p><h3>Step 5:<strong>\u00a0Use the 85% Rule to invest in yourself<\/strong><br \/><\/h3><p>When it comes to starting a new business, too many of us will get overwhelmed thinking that we need to do everything \u201cperfectly.\u201d In fact, many will spin their tires doing things like:<\/p><ul><li>Starting a blog<\/li>\n<li>Creating a Facebook page<\/li>\n<li>Buying ads on Google<\/li>\n<li>Hiring skywriters to promote their business in the skies<\/li>\n<li>Whatever<\/li>\n<\/ul><p>This is an easy way to get burned out, leading us to do nothing at all.<\/p><p>Your goal isn\u2019t to start a blog or create a Facebook page with millions of likes \u2014 your goal is to FIND PAYING CLIENTS.<\/p><p>That\u2019s why the easiest way to start freelancing is to take things one step at a time and not worry about getting it perfect.<\/p><p>I\u2019d rather get it 85% right than do nothing at all. <\/p><p>Remember: it\u2019s okay to make mistakes. The important thing is that you get started and learn from those mistakes.<\/p><p>The best way to get started is to begin <em>investing in yourself.<\/em>\u00a0<\/p><p>Now, one thing that you will find very common with people who have not taken the time to invest in themselves and learn how this stuff works, is they will create what\u2019s called levels of abstraction. <\/p><p>Rather than just going directly to what they want, they will create all these different levels of abstraction \u2014 like making a Facebook page or a blog \u2014 that make them feel good, but that actually don\u2019t require them to do the hard work.<\/p><p>So, they\u2019ll spend six, nine, 12 months doing something frustrating and then give up because they never spent time buying a couple good books or buying a course.<\/p><p>That\u2019s why I\u2019d like to offer you somewhere to start: <u><a href=\"https:\/\/www.iwillteachyoutoberich.com\/newsletter-signup\/\" target=\"_blank\">my newsletter<\/a><\/u>.<\/p><p>Join this private list to see the tips, strategies, and techniques you WON\u2019T see on my blog. <\/p><h2 id=\"4\">Ramit, I\u2019m not ready to make extra money because of [issue]<\/h2><p>\u201cBut Ramit, I don\u2019t think I can possibly work on a side business. I don\u2019t have enough time and I don\u2019t think I even have a business idea!\u201d<\/p><p>There are always going to be excuses when it comes to trying something new \u2014 but when you take a look at some of them, you\u2019ll realize that they\u2019re almost always all MENTAL barriers.<\/p><p>Here\u2019s how you can break past the 3 most common ones I\u2019ve seen.<\/p><h3 id=\"5\">Mental Barrier #1: I don\u2019t have the time to make extra money!<\/h3><p>It\u2019s always funny to me when people tell me they don\u2019t have enough time for their regular job, let alone a freelance business on the side.<\/p><p>Well, I actually have a secret I\u2019d like to tell you that might blow your mind\u2026<\/p><p>\u2026YOU DO HAVE THE TIME!!!<\/p><p>There\u2019s no doubt about it: starting a freelance business will take time \u2014 but not nearly as much as you think.<\/p><p>Can you spare an hour a day to freelancing? How about a few hours on the weekends? <\/p><p>I\u2019m guessing that you can. Maybe instead of spending after-work time endlessly binging Netflix or browsing Reddit, you can devote yourself to starting a business that can earn you extra thousands this year. Hell, work from a laptop on the couch while streaming; Netflix and Making Money is the new Netflix and Chill.<\/p><p>Saying \u201cI don\u2019t have the time\u201d is really saying \u201cThis is not a priority.\u201d Try that out. \u201cMaking more money is not a priority.\u201d Compared to \u201cBrowsing Facebook is not a priority.\u201d Which one of those is true?<\/p><h3 id=\"6\">Mental Barrier #2: But I\u2019m so tired after work!<\/h3><p>This is actually a very valid concern. After all, many of us work long eight hour days \u2014 how are we expected to come home and work on freelancing when all we want to do is pour ourselves some wine and put on the latest episode of <em>The Bachelor<\/em>\u00a0(or is that just me..)?<\/p><p>It\u2019s difficult \u2014 but you can actually do it successfully with some simple psychology.<\/p><p>Imagine a smoker. Many smokers know that smoking is a bad habit and is terrible for their health \u2014 but they still do it because of cognitive dissonance. <\/p><p>What they\u2019ll say is, \u201cYeah, I know it\u2019s really bad for me but I work really hard and I deserve this pack of cigarettes.\u201d<\/p><p>It\u2019s the same thing as starting an online business. At the end of a long work day, you\u2019ll say to yourself, \u201cI worked all day and dealt with my dumb boss and blah blah blah. I really deserve to kick back and watch Netflix.\u201d<\/p><p>Instead of falling victim to cognitive dissonance, remember this question: <\/p><div style=\"text-align: center;\"><strong>Where do I want to be a year from now?<\/strong><\/div><p>That\u2019s actually something my parents told me when I was younger, and I never forgot it. Where do I want to be a year from now? And what small steps can I take to get there?<\/p><p>When you look at some successful freelancers who earn hundreds of dollars an hour, many will think, \u201cI\u2019ll never be able to earn that!\u201d<\/p><p>But you WILL as long as you start today. And a year from now, you\u2019ll be earning much more than you would if you didn\u2019t do anything at all.<\/p><h3 id=\"7\">Mental Barrier #3: I don\u2019t have a good business idea!<\/h3><p>This is the most common barrier preventing people from making money freelancing: <em>not knowing what type of work they want to do. <\/em><\/p><p>As I\u2019ve outlined in the steps above, though, it\u2019s simple to identify your profitable skills so long as you look at 4 things:<\/p><ol><li>What do you already pay for?<\/li>\n<li>What skills do you have?<\/li>\n<li>What do your friends say you\u2019re great at?<\/li>\n<li>What do you do on a Saturday morning?<\/li>\n<\/ol><p>And the answer to this mental barrier will serve as the very first step in creating your side business. <\/p><p>If you\u2019re still stuck, don\u2019t worry: I have just the thing for you \u2026<\/p><h2 id=\"8\">42 side hustle ideas for beginners, college students, parents, and more!<\/h2><p>No matter your employment, family status, or location, you can find the right business idea to pursue on the side. To help you get started even more, here are 42 examples of great hustles you can start today. <\/p><h3 id=\"9\">Side hustle ideas for beginners<\/h3><p>Here is a list of services that don\u2019t take any special skills or training at all that you can offer to people for money:<\/p><ul><li>Transcription services<\/li>\n<li>Courier services<\/li>\n<li>Cooking \/ personal chef<\/li>\n<li>Technical writing<\/li>\n<li>Small business marketing<\/li>\n<li>Proofreading \/ editing<\/li>\n<li>Business writing (business plans, grants, white papers)<\/li>\n<\/ul><h3 id=\"10\">Side hustle ideas for college students<\/h3><p>You can start a lucrative side hustle even if you\u2019re living in a dorm room right now. Actually, that\u2019s how I started IWT over a decade ago. Check out these hustles:<\/p><ul><li>Video production \/ editing<\/li>\n<li>Design (brochures, business cards, logos, newsletters)<\/li>\n<li>Web design<\/li>\n<li>Blogging \/ blog consulting<\/li>\n<li>Sports \/ personal training<\/li>\n<li>Child care<\/li>\n<li>Search engine optimization (SEO)<\/li>\n<\/ul><h3 id=\"11\">Side hustle ideas for millennials<\/h3><p>Millennials have a wealth of marketable skills just by virtue of being in a tech-centric generation. My first side hustle was consulting for venture capital firms on how young people were using social media. Here are some other things you can do:<\/p><ul><li>User interface &amp; user experience (UI\/UX)<\/li>\n<li>Google Analytics<\/li>\n<li>Programming \/ web development<\/li>\n<li>E-commerce consulting<\/li>\n<li>Tech tutoring (e.g., how to set up and use your tablet)<\/li>\n<li>Internet marketing consulting<\/li>\n<li>Fashion \/ image consulting<\/li>\n<\/ul><h3 id=\"12\">Side hustle ideas for stay-at-home parents<\/h3><p>Parents have some of the busiest schedules and irregular hours of anyone. However, on top of all those responsibilities, you can find a side hustle idea that makes you money. Case in point:<\/p><ul><li>Personal \/ virtual assistant<\/li>\n<li>Pet services (grooming, walking, etc.)<\/li>\n<li>Home maintenance<\/li>\n<li>Crafting \/ personalization services<\/li>\n<li>Human resources \/ payroll services<\/li>\n<li>Project management \/ productivity consulting<\/li>\n<li>Copywriting and editing<\/li>\n<\/ul><h3 id=\"13\">Side hustle ideas for working parents<\/h3><p>If you\u2019re a working parent, you probably have skills such as planning, budgeting, and productivity. You can easily transform one of those skills into a side hustle such as:<\/p><ul><li>Photography<\/li>\n<li>Travel planning<\/li>\n<li>Event planning \/ promotion<\/li>\n<li>Automobile \/ motorcycle repair<\/li>\n<li>Tax and financial planning<\/li>\n<li>Community management \/ promotion<\/li>\n<li>Career help (finding work, optimizing resumes, cover letters, etc.)<\/li>\n<\/ul><h3 id=\"14\">Side hustle ideas for teachers<\/h3><p>Teaching and the skills involved are at the core for a lot of thriving side hustles. That\u2019s why professional teachers have an advantage when it comes to hustles such as:<\/p><ul><li>Research<\/li>\n<li>Presentation design (e.g., PowerPoint)<\/li>\n<li>Tutoring and coaching<\/li>\n<li>SAT prep<\/li>\n<li>College application review<\/li>\n<li>Productivity coaching<\/li>\n<li>Accounting \/ bookkeeping<\/li>\n<\/ul><h2 id=\"15\">Need more ideas?<\/h2><p>If you want even more side hustle ideas, I want to offer you my FREE PDF of 30 successful and proven business ideas.<\/p><div style=\"text-align: center;\"><img src=\"https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/pasted-image-0-561.png\" class=\"wp-image-26758\" alt=\"pasted image 0 561\" srcset=\"https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/pasted-image-0-561.png 231w, https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/pasted-image-0-561-48x62.png 48w\" \/><\/div><p>These ideas can absolutely work for you. How do I know? That\u2019s because they\u2019re actual business ideas people have ran with to make thousands more each month.<\/p><p>Just enter your name and email below and I\u2019ll send the 30 ideas straight to your inbox.<\/p><div class=\"optin-large\"><div class=\"optin-inner\"><div class=\"box-gray\"><h2>Show me 30 successful business ideas!:<\/h2><div class=\"row\"><div class=\"col-lg-6\"><img src=\"https:\/\/www.iwillteachyoutoberich.com\/wp-content\/themes\/iwt\/assets\/img\/optin-books.png\" alt=\"\" \/><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:44:51+0000","updated_at":"2018-11-07T16:44:51+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":20,"domain_name":"www.iwillteachyoutoberich.com","preview_picture":"https:\/\/cdn.iwillteachyoutoberich.com\/wp-content\/uploads\/earn-extra-money-e1501542999693.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":326,"title":"Pardon the interruption","url":"https:\/\/www.moneyunder30.com\/how-to-invest","content":"Pardon the interruption\n  <img src=\"https:\/\/www.moneyunder30.com\/images\/logos\/MU30_Logo_Large.png\" alt=\"image\" \/>\n  <h3>Pardon the interruption...<\/h3>\n  <p>As you were browsing our site, something about your browser made us think you were a bot.<\/p>\n  <p>All you need to do is complete the Captcha below and we'll send you right back to Money Under 30.<\/p>\n  \u200b            \n            \n            <form id=\"distilCaptchaForm\" class=\"recaptcha2\" action=\"\/distil_r_captcha.html?requestId=2092e818-e03c-4c13-86a4-18acb9986933&amp;httpReferrer=%2Fhow-to-invest\" method=\"post\">\n                \n                <noscript>\n                    \n                <\/noscript>\n                <p>You reached this page when attempting to access https:\/\/www.moneyunder30.com\/how-to-invest from 104.236.87.208 on 2018-11-07 16:44:51 UTC.<br \/>Trace: 2092e818-e03c-4c13-86a4-18acb9986933 via 35a9edc9-dac9-459f-8854-dc3ec5d9864f<\/p>\n            <\/form>\n        \n  <p>If you keep seeing this page, you might want to consider checking your browser to make sure JavaScript is enabled. Or you might just be clicking through our site at superhuman speed. Check your browser, slow your clicking, and you probably won't ever have to do this again.<\/p>","created_at":"2018-11-07T16:44:51+0000","updated_at":"2018-11-07T16:44:51+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.moneyunder30.com","http_status":"405"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":324,"title":"CSS Grid for UI Layouts \u2013 Mozilla Hacks - the Web developer blog","url":"https:\/\/hacks.mozilla.org\/2018\/02\/css-grid-for-ui-layouts\/","content":"<p>CSS Grid is a great layout tool for content-driven websites that include long passages of text, and it has tremendous value for a variety of traditional UI layouts as well. In this article I\u2019ll show you how to use CSS Grid to improve application layouts that need to respond and adapt to user interactions and changing conditions, and always have your panels scroll properly.<\/p><p><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/CSS_Grid_Layout\" rel=\"noopener\" target=\"_blank\">CSS Grid<\/a> builds website layouts. It lets web designers create beautiful dynamic layouts using just a tiny bit of supported code instead of the endless float hacks we\u2019ve had to use for years. My friend and co-worker Jen Simmons has been <a href=\"http:\/\/jensimmons.com\/writing\">talking about<\/a> CSS Grid <a href=\"https:\/\/www.youtube.com\/watch?v=kRYrbcGWjzU\">for years<\/a>, tirelessly pushing to get it implemented in browsers, and her work has paid off. As of the end of last year, the current version of <i>every major browser, desktop and mobile, <\/i><a href=\"https:\/\/caniuse.com\/#feat=css-grid\"><i>supports CSS Grid<\/i><\/a>.<\/p><p>CSS Grid really is powerful, and you can build dynamic content driven websites easily, like in <a href=\"https:\/\/speckyboy.com\/creative-examples-css-grid-layouts\/\">these examples<\/a>. However, Grid is good for more than laying out pretty blocks of content. Grid gives you full control over both dimensions of your layout, including scrolling. This means features we take for granted in native applications like collapsing side-panels and fixed toolbars are now trivial to implement. No more hacks and no more debugging. Grid just works.<\/p><p>I\u2019ve been building web tools for years. Here\u2019s a screenshot of a game building tool I made for my retro RPGs. When <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Glossary\/Flexbox\" rel=\"noopener\" target=\"_blank\">Flexbox<\/a> first appeared I immediately started using it. I built complex layouts using nested horizontal and vertical boxes, with a few utility classes for things like scrolling and stretching.<\/p><p><a href=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image4.png\"><img class=\"alignnone wp-image-31917 size-large\" src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image4-500x309.png\" alt=\"\" width=\"500\" height=\"309\" srcset=\"https:\/\/hacks.mozilla.org\/files\/2018\/02\/image4-500x309.png 500w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image4-250x155.png 250w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image4-768x475.png 768w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image4.png 982w\" \/><\/a><\/p><p>Flexbox has certainly made me more productive than absolutely positioned divs and float hacks, but it still has problems. Look at this closeup where panels come together. See how the footers on the left and right don\u2019t line up?<\/p><p><a href=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image5.png\"><img class=\"alignnone wp-image-31918 size-full\" src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image5.png\" alt=\"\" width=\"348\" height=\"84\" srcset=\"https:\/\/hacks.mozilla.org\/files\/2018\/02\/image5.png 348w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image5-250x60.png 250w\" \/><\/a><\/p><p>Here\u2019s another screenshot. The toolbar is at the top of the drawing canvas, and according to my framework it should be fixed at the top, but the minute you start scrolling this happens. The toolbar disappears:<\/p><p><a href=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image3.png\"><img class=\"alignnone wp-image-31919 size-large\" src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image3-500x458.png\" alt=\"\" width=\"500\" height=\"458\" srcset=\"https:\/\/hacks.mozilla.org\/files\/2018\/02\/image3-500x458.png 500w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image3-250x229.png 250w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image3-768x703.png 768w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image3.png 1476w\" \/><\/a><\/p><p>Each of these problems can be fixed with more positioning and float hacks, but the result is always fragile. Every time I add a new panel I have to debug my layout all over again; searching to identify which <code>div<\/code> is grabbing the extra space during a resize. And the markup is ugly. The nested horizontal and vertical boxes become very complicated, and this example is only two levels deep. As interaction and functionality become more complex the design becomes even more challenging.<\/p><pre class=\"html\">&lt;div class='hbox'&gt;\n  &lt;div class='vbox'&gt;\n    &lt;div class='hbox'&gt;header&lt;\/div&gt;\n    &lt;div class='scroll'&gt;\n      &lt;div class='sidebar'&gt;sidebar&lt;\/div&gt;\n    &lt;\/div&gt;\n    &lt;div class='footer'&gt;footer&lt;\/div&gt;\n  &lt;\/div&gt;\n  &lt;div class=vbox&gt;\n    &lt;div class='hbox'&gt;button button \n        spacer label spacer button button &lt;\/div&gt;\n    &lt;div class='center'&gt;main content&lt;\/div&gt;\n    &lt;div class='hbox'&gt;the footer&lt;\/div&gt;\n  &lt;\/div&gt;\n  &lt;div class=vbox&gt;\n    &lt;div class=\u2019hbox\u2019&gt;header&lt;\/div&gt;\n    &lt;div class=\u2019scroll\u2019&gt;\n      &lt;div class=\u2019sidebar\u2019&gt;sidebar&lt;\/div&gt;\n    &lt;\/div&gt;\n    &lt;div class=\u2019footer\u2019&gt;footer&lt;\/div&gt;\n  &lt;\/div&gt;\n&lt;\/div&gt;\n<\/pre><h3>Entering the Second Dimension<\/h3><p>The fundamental problem with Flexbox is that it is one dimensional. This makes Flexbox great for one dimensional uses, like toolbars and navbars, but it begins to fail when I need to align content both horizontally and vertically at the same time. Instead I need real <b>two<\/b> dimensional layout, which is why I need CSS Grid. Fundamentally Grid is 2D.<\/p><p>Here\u2019s a similar kind of layout built with CSS Grid.<\/p><p><a href=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image1.png\"><img class=\"alignnone wp-image-31920 size-large\" src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image1-500x275.png\" alt=\"\" width=\"500\" height=\"275\" srcset=\"https:\/\/hacks.mozilla.org\/files\/2018\/02\/image1-500x275.png 500w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image1-250x138.png 250w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image1-768x422.png 768w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image1.png 1600w\" \/><\/a><\/p><p>Look closely at the bottom footers. They come together perfectly. And by using the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/grid-gap\"><code>grid-gap<\/code><\/a> for the lines instead of adding borders to each panel, I don\u2019t have to worry about inconsistent grid line widths. Everything just works.<\/p><p><a href=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image2.png\"><img class=\"alignnone wp-image-31921 size-large\" src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/02\/image2-500x204.png\" alt=\"\" width=\"500\" height=\"204\" srcset=\"https:\/\/hacks.mozilla.org\/files\/2018\/02\/image2-500x204.png 500w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image2-250x102.png 250w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image2-768x314.png 768w, https:\/\/hacks.mozilla.org\/files\/2018\/02\/image2.png 1038w\" \/><\/a><\/p><p>The biggest benefit I get from CSS Grid is adapting to changing conditions. My apps often have side panels. I need to make sure everything in the layout works regardless of whether the panels are expanded or collapsed, ideally without having to recalculate layout in JavaScript. Sidebars are made out of multiple components like headers and footers. All of these need to line up, regardless of which one is larger or smaller. \u00a0Grid can do this too using a magic function called <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/minmax\"><code>minmax()<\/code><\/a>.<\/p><p>If you\u2019ve studied CSS Grid before then you know you can define your layout using <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/grid-template-rows\">templates<\/a> for the rows and columns. \u00a0A template like <code>200px 1fr 200px<\/code> will give you 200px wide sidebars with a middle content area taking up the rest of the space. But what happens if the panel should collapse? Right now the column would stay at 200px, even though the content has shrunk. Instead we can use <code>minmax<\/code> with the <code>min-content<\/code> keyword for the max parameter.<\/p><pre class=\"css\"><b>#grid <\/b>{\n  <b>display<\/b>: <b>grid<\/b>;\n  <b>box-sizing<\/b>: <b>border-box<\/b>;\n  <b>width<\/b>: 100<b>vw<\/b>;\n  <b>height<\/b>: 100<b>vh<\/b>;\n  <b>grid-template-columns<\/b>: \n      [<b>start<\/b>] <b>minmax<\/b>(<b>auto<\/b>, <b>min-content<\/b>) \n      [<b>center<\/b>]1<b>fr <\/b>\n      [<b>end<\/b>] <b>minmax<\/b>(<b>auto<\/b>,<b>min-content<\/b>);\n  <b>grid-template-rows<\/b>: \n      [<b>header<\/b>]2<b>em <\/b>\n      [<b>content<\/b>]1<b>fr <\/b>\n      [<b>footer<\/b>]2<b>em<\/b>;\n  <b>grid-gap<\/b>: 1<b>px<\/b>;\n  <b>background-color<\/b>: <b>black<\/b>;\n}<\/pre><p>Now the grid column will be always be just wide enough to hold whatever is in any of the columns using their <i>minimum width<\/i>. Thus if one part of the column (say the header) is wider than the others, the column will expand to fit them all. If they become skinnier or disappear altogether, then the column will adjust accordingly. Essentially we have replicated the expanding\/contracting behavior of Flexbox, but made it work with <i>everything<\/i> in the column together, not just one item. This is <i>real<\/i> 2D layout.<\/p><p>Here is the code for the rest of the demo.<\/p><pre class=\"css\">\n.<b>start <\/b>{\n  <b>grid-column<\/b>: <b>start<\/b>;\n}\n.<b>center <\/b>{\n  <b>grid-column<\/b>: <b>center<\/b>;\n}\n.<b>end <\/b>{\n  <b>grid-column<\/b>: <b>end<\/b>;\n}\n<b>header <\/b>{\n  <b>grid-row<\/b>: <b>header<\/b>;\n}\n<b>footer <\/b>{\n  <b>grid-row<\/b>: <b>footer<\/b>;\n}\n.<b>sidebar <\/b>{\n  <b>overflow<\/b>: <b>auto<\/b>;\n}\n<\/pre><pre class=\"html\">\n&lt;<b>div <\/b><b>id=<\/b><b>\"grid\"<\/b>&gt;\n&lt;<b>header <\/b><b>class=<\/b><b>\"start\"<\/b>&gt;header&lt;\/<b>header<\/b>&gt;\n&lt;<b>header <\/b><b>class=<\/b><b>\"center\"<\/b>&gt;\n  &lt;<b>button <\/b><b>id=<\/b><b>\"toggle-left\"<\/b>&gt;toggle left&lt;\/<b>button<\/b>&gt;\n...\n&lt;\/<b>header<\/b>&gt;\n&lt;<b>header <\/b><b>class=<\/b><b>\"end\"<\/b>&gt;header&lt;\/<b>header<\/b>&gt;\n\u00a0\n&lt;<b>div <\/b><b>class=<\/b><b>\"start sidebar\"<\/b>&gt;sidebar&lt;\/<b>div<\/b>&gt;\n&lt;<b>div <\/b><b>class=<\/b><b>\"center content\"<\/b>&gt;the center content&lt;\/<b>div<\/b>&gt;\n&lt;<b>div <\/b><b>class=<\/b><b>\"end sidebar\"<\/b>&gt;\n  sidebar&lt;<b>br<\/b>\/&gt;\n...\n&lt;\/<b>div<\/b>&gt;\n\u00a0\n&lt;<b>footer <\/b><b>class=<\/b><b>\"start\"<\/b>&gt;left footer&lt;\/<b>footer<\/b>&gt;\n&lt;<b>footer <\/b><b>class=<\/b><b>\"center\"<\/b>&gt;center footer&lt;\/<b>footer<\/b>&gt;\n&lt;<b>footer <\/b><b>class=<\/b><b>\"end\"<\/b>&gt;right footer&lt;\/<b>footer<\/b>&gt;\n&lt;\/<b>div<\/b>&gt;\n<\/pre><p>To make the toggle buttons in the upper header actually hide the sidebars I added this code. Note that with modern DOM APIs and arrow functions we can essentially replicate JQuery in just a few lines:<\/p><pre class=\"js\"><b>const <\/b><i>$ <\/i>= (selector) =&gt; <b><i>document<\/i><\/b>.querySelector(selector)\n<b>const <\/b><i>$$ <\/i>= (selector) =&gt; <b><i>document<\/i><\/b>.querySelectorAll(selector)\n<b>const <\/b><i>on <\/i>= (elem, type, listener) =&gt; elem.addEventListener(type,listener)\n<i>on<\/i>(<i>$<\/i>(<b>'#toggle-left'<\/b>),<b>'click'<\/b>,()=&gt;{\n  <i>$$<\/i>(<b>\".start\"<\/b>).forEach((elem) =&gt; elem.classList.toggle(<b>'closed'<\/b>))\n})\n<i>on<\/i>(<i>$<\/i>(<b>'#toggle-right'<\/b>),<b>'click'<\/b>,()=&gt;{\n  <i>$$<\/i>(<b>\".end\"<\/b>).forEach((elem) =&gt; elem.classList.toggle(<b>'closed'<\/b>))\n})\n<\/pre><p>Also note that CSS Grid <i>does not<\/i> deprecate Flexbox. We still use Flexbox in the cases where it makes sense: namely one dimensional content like toolbars. Here are the styles that I\u2019m using for my toolbars made out of headers:<\/p><pre class=\"html\">\n&lt;<b>header <\/b><b>class=<\/b><b>\"center\"<\/b>&gt;\n  &lt;<b>button <\/b><b>id=<\/b><b>\"toggle-left\"<\/b>&gt;toggle left&lt;\/<b>button<\/b>&gt;\n  &lt;<b>button<\/b>&gt;open&lt;\/<b>button<\/b>&gt;\n  &lt;<b>button<\/b>&gt;save&lt;\/<b>button<\/b>&gt;\n  &lt;<b>span <\/b><b>class=<\/b><b>\"spacer\"<\/b>&gt;&lt;\/<b>span<\/b>&gt;\n  &lt;<b>span<\/b>&gt;filename.txt&lt;\/<b>span<\/b>&gt;\n  &lt;<b>span <\/b><b>class=<\/b><b>\"spacer\"<\/b>&gt;&lt;\/<b>span<\/b>&gt;\n  &lt;<b>button<\/b>&gt;delete&lt;\/<b>button<\/b>&gt;\n  &lt;<b>button <\/b><b>id=<\/b><b>\"toggle-right\"<\/b>&gt;toggle right&lt;\/<b>button<\/b>&gt;\n&lt;\/<b>header<\/b>&gt;\n<b>header <\/b>{\n  <b>background-color<\/b>: <b>#ccc<\/b>;\n  <b>display<\/b>: <b>flex<\/b>;\n  <b>flex-direction<\/b>: <b>row<\/b>;\n}\n.<b>spacer <\/b>{\n  <b>flex<\/b>: 1;\n}\n<\/pre><p>The <code>spacer<\/code> class makes an element take up all of the extra space. By using two spacers between the buttons I can make my toolbar shrink and grow as needed with the filename always in the middle. This is similar to native toolbars.<\/p><p>You can try out a demo live at this <a href=\"https:\/\/codepen.io\/joshmarinacci\/full\/EQXyvM\">Codepen<\/a>, then remix it to poke and prod.<\/p><p class=\"codepen\" data-height=\"459\" data-theme-id=\"0\" data-slug-hash=\"EQXyvM\" data-default-tab=\"result\" data-user=\"joshmarinacci\" data-embed-version=\"2\" data-pen-title=\"CSS Grid for UI Layouts\">See the Pen <a href=\"https:\/\/codepen.io\/joshmarinacci\/pen\/EQXyvM\/\">CSS Grid for UI Layouts<\/a> by Josh Marinacci (<a href=\"https:\/\/codepen.io\/joshmarinacci\">@joshmarinacci<\/a>) on <a href=\"https:\/\/codepen.io\">CodePen<\/a>.<\/p><p>CSS Grid is wonderful for designing interactive applications with two-dimensional complexity. We can keep the markup semantic. Panels and toolbar line up properly. The <code>grid-gap<\/code> gives us automatic borders. It adjusts our layout in complex ways without any JavaScript code, and it gives us control over both the horizontal and vertical. And we can do it all <b>without<\/b> using a heavy CSS framework.<\/p><p>Jen Simmons has started a new YouTube channel, <a href=\"https:\/\/www.youtube.com\/channel\/UC7TizprGknbDalbHplROtag\">Layout Land<\/a> to help you grok how Grid works. \u00a0If you work on web apps or any kind of richly interactive website, you should try out CSS Grid.<\/p>","created_at":"2018-11-07T16:44:50+0000","updated_at":"2018-11-07T16:44:50+0000","published_at":"2018-02-14T08:03:34+0000","published_by":["Josh Marinacci"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"hacks.mozilla.org","preview_picture":"https:\/\/hacks.mozilla.org\/files\/2018\/02\/hacksscreenshot-2_1c.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":322,"title":"Using IndexedDB","url":"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IndexedDB_API\/Using_IndexedDB","content":"<div class=\"summary\"><p>IndexedDB\u00a0is a way for you to persistently store data inside a user's browser. Because it lets you create web applications with rich query abilities regardless of network availability, your applications can work both online and offline.\u00a0<\/p><\/div><p>This tutorial walks you through using the asynchronous API of IndexedDB. If you are not familiar with IndexedDB, you should first read <a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\/Basic_Concepts_Behind_IndexedDB\" title=\"en\/IndexedDB\/Basic Concepts Behind IndexedDB\">Basic Concepts About IndexedDB<\/a>.<\/p><p>For the reference documentation on the IndexedDB API, see the <a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\" title=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\">IndexedDB<\/a> article and its subpages.\u00a0This article documents the types of objects used by IndexedDB, as well as the methods of the asynchronous API (synchronous API removed from spec). \u00a0<\/p><h2 id=\"pattern\">Basic pattern<\/h2><p>The basic pattern that IndexedDB encourages is the following:<\/p><ol><li>Open a database.<\/li>\n <li>Create an object store in the database.\u00a0<\/li>\n <li>Start a transaction and make a request to do some database operation, like adding or retrieving data.<\/li>\n <li>\n  <p>Wait for the operation to complete by listening to the right kind of DOM\u00a0event.<\/p>\n <\/li>\n <li>\n  <p>Do something with the results (which can be found on the request object).<\/p>\n <\/li>\n<\/ol><p>With these big concepts under our belts, we can get to more concrete stuff.<\/p><h2 id=\"open\">Creating and structuring the store<\/h2><h3 id=\"Using_an_experimental_version_of_IndexedDB\">Using an experimental version of IndexedDB<\/h3><p>In case you want to test your code in browsers that still use a prefix, you can use the following code: \u00a0<\/p><pre class=\"brush: js\">\/\/ In the following line, you should include the prefixes of implementations you want to test.\nwindow.indexedDB = window.indexedDB || window.mozIndexedDB || window.webkitIndexedDB || window.msIndexedDB;\n\/\/ DON'T use \"var indexedDB = ...\" if you're not in a function.\n\/\/ Moreover, you may need references to some window.IDB* objects:\nwindow.IDBTransaction = window.IDBTransaction || window.webkitIDBTransaction || window.msIDBTransaction || {READ_WRITE: \"readwrite\"}; \/\/ This line should only be needed if it is needed to support the object's constants for older browsers\nwindow.IDBKeyRange = window.IDBKeyRange || window.webkitIDBKeyRange || window.msIDBKeyRange;\n\/\/ (Mozilla has never prefixed these objects, so we don't need window.mozIDB*)<\/pre><p>Beware that implementations that use a prefix may be buggy, or incomplete, or following an old version of the specification. Therefore, it is not recommended to use it in production code. It may be preferable to not support a browser than to claim to support it and fail:<\/p><pre class=\"brush: js\">if (!window.indexedDB) {\n    window.alert(\"Your browser doesn't support a stable version of IndexedDB. Such and such feature will not be available.\");\n}\n<\/pre><h3 id=\"Opening_a_database\">Opening a database<\/h3><p>We start the whole process like this:<\/p><pre class=\"brush: js\">\/\/ Let us open our database\nvar request = window.indexedDB.open(\"MyTestDatabase\", 3);\n<\/pre><p>See that? Opening a database is just like any other operation \u2014 you have to \"request\" it.<\/p><p>The open request doesn't open the database or start the transaction right away. The call to the\u00a0<code>open()<\/code> function returns an\u00a0<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/IDBOpenDBRequest\" title=\"\/en-US\/docs\/IndexedDB\/IDBOpenDBRequest\"><code>IDBOpenDBRequest<\/code><\/a> object with a result (success) or error value that you handle as an event. Most other asynchronous functions in IndexedDB do the same thing - return\u00a0an\u00a0<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/IDBRequest\" title=\"\/en-US\/docs\/IndexedDB\/IDBRequest\"><code>IDBRequest<\/code><\/a>\u00a0object with the result or error. The result for the open function is an instance of an\u00a0<code><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/IDBDatabase\" title=\"\/en-US\/docs\/IndexedDB\/IDBDatabase\">IDBDatabase<\/a>.<\/code><\/p><p>The second parameter to the open method is the version of the database. The version of the database determines the database schema \u2014 the object stores in the database and their structure. If the database doesn't already exist, it is created by the <code>open<\/code> operation, then an <code>onupgradeneeded <\/code>event is triggered and you create the database schema in the handler for this event. If the database does exist but you are specifying an upgraded version number, an <code>onupgradeneeded <\/code>event is triggered straight away, allowing you to provide an updated schema in its handler. More on this later in <a href=\"#Updating_the_version_of_the_database\">Updating the version of the database<\/a> below, and the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBFactory\/open\" title=\"If an error occurs while the database connection is being opened, then an error event is fired on the request object returned from this method.\"><code>IDBFactory.open<\/code><\/a> reference page.\u00a0<\/p><div class=\"warning\"><p><strong>Important<\/strong>: The version number is an <code>unsigned long long<\/code> number, which means that it can be a very big integer. It also means that you can't use a float, otherwise it will be converted to the closest lower integer and the transaction may not start, nor the <code>upgradeneeded<\/code> event trigger. So for example, don't use 2.4 as a version number:<br \/><code>var request = indexedDB.open(\"MyTestDatabase\", 2.4); \/\/ don't do this, as the version will be rounded to 2<\/code><\/p><\/div><h4 id=\"Generating_handlers\">Generating handlers<\/h4><p>The first thing you'll want to do with almost all of the requests you generate is to add success and error handlers:<\/p><pre class=\"brush: js\">request.onerror = function(event) {\n  \/\/ Do something with request.errorCode!\n};\nrequest.onsuccess = function(event) {\n  \/\/ Do something with request.result!\n};<\/pre><p>Which of the two functions, <code>onsuccess()<\/code> or <code>onerror()<\/code>, gets called? If everything succeeds, a success event (that is, a DOM\u00a0event whose <code>type<\/code> property is set to <code>\"success\"<\/code>) is fired with <code>request<\/code> as its <code>target<\/code>. Once it is fired, the <code>onsuccess()<\/code> function on <code>request<\/code> is triggered with the success event as its argument. Otherwise, if there was any problem, an error event (that is, a DOM event whose <code>type<\/code> property is set to <code>\"error\"<\/code>) is fired at <code>request<\/code>. This triggers the <code><code>onerror()<\/code><\/code> function with the error event as its argument.<\/p><p>The IndexedDB API is designed to minimize the need for error handling, so you're not likely to see many error events (at least, not once you're used to the API!). In the case of opening a database, however, there are some common conditions that generate error events. The most likely problem is that the user decided not to give your web app permission to create a database. One of the main design goals of IndexedDB is to allow large amounts of data to be stored for offline use. (To learn more about how much storage you can have for each browser, see <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IndexedDB_API\/Browser_storage_limits_and_eviction_criteria#Storage_limits\" title=\"https:\/\/developer.mozilla.org\/en\/IndexedDB#Storage_limits\">Storage limits<\/a>.)\u00a0\u00a0<\/p><p>Obviously, browsers do not want to allow some advertising network or malicious website to pollute your computer, so browsers used to\u00a0prompt the user the first time any given web app attempts to open an IndexedDB for storage. The user could choose to allow or deny access. Also, IndexedDB storage in browsers' privacy modes\u00a0only lasts in-memory until the incognito session is closed\u00a0(Private Browsing mode for Firefox and\u00a0Incognito mode for Chrome, but in Firefox this is <a rel=\"noopener\" href=\"https:\/\/bugzilla.mozilla.org\/show_bug.cgi?id=781982\" class=\"external\">not implemented\u00a0yet<\/a>\u00a0as of Nov 2015 so you can't use IndexedDB in Firefox Private Browsing at all).<\/p><p>Now, assuming that the user allowed your request to create a database, and you've received a success event to trigger the success callback; What's next? The request here was generated with a call to <code>indexedDB.open()<\/code>, so <code>request.result<\/code> is an instance of <code>IDBDatabase<\/code>, and you definitely want to save that for later. Your code might look something like this:<\/p><pre class=\"brush: js\">var db;\nvar request = indexedDB.open(\"MyTestDatabase\");\nrequest.onerror = function(event) {\n  alert(\"Why didn't you allow my web app to use IndexedDB?!\");\n};\nrequest.onsuccess = function(event) {\n  db = event.target.result;\n};\n<\/pre><h4 id=\"Handling_Errors\">Handling Errors<\/h4><p>As mentioned above, error events bubble. Error events are targeted at the request that generated the error, then the event bubbles to the transaction, and then finally to the database object. If you want to avoid adding error handlers to every request, you can instead add a single error handler on the database object, like so:<\/p><pre class=\"brush: js\">db.onerror = function(event) {\n  \/\/ Generic error handler for all errors targeted at this database's\n  \/\/ requests!\n  alert(\"Database error: \" + event.target.errorCode);\n};\n<\/pre><p>One of the common possible errors when opening a database is <code>VER_ERR<\/code>. It indicates that the version of the database stored on the disk is <em>greater<\/em> than the version that you are trying to open. This is an error case that must always be handled by the error handler.<\/p><h3 id=\"Creating_or_updating_the_version_of_the_database\">Creating or updating the version of the database<\/h3><p>When you create a new database or increase the version number of an existing database (by specifying a higher version number than you did previously, when <a href=\"#Opening_a_database\">Opening a database<\/a>), the <code>onupgradeneeded<\/code> event will be triggered and an\u00a0<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBVersionChangeEvent\">IDBVersionChangeEvent<\/a> object will be passed to any <code>onversionchange<\/code> event handler set up on <code>request.result<\/code> (i.e., <code>db<\/code> in the example). In the handler for the <code>upgradeneeded<\/code> event, you should create the object stores needed for this version of the database:<\/p><pre class=\"brush:js;\">\/\/ This event is only implemented in recent browsers   \nrequest.onupgradeneeded = function(event) { \n  \/\/ Save the\u00a0IDBDatabase\u00a0interface\u00a0\n  var db = event.target.result;\n  \/\/ Create an objectStore for this database\n  var objectStore = db.createObjectStore(\"name\", { keyPath: \"myKey\" });\n};<\/pre><p>In this case, the database will already have the object stores from the previous version of the database, so you do not have to create these object stores again. You only need to create any new object stores, or delete object stores from the previous version that are no longer needed. If you need to change an existing object store (e.g., to change the <code>keyPath<\/code>), then you must delete the old object store and create it again with the new options. (Note that this will delete the information in the object store! If you need to save that information, you should read it out and save it somewhere else before upgrading the database.)<\/p><p>Trying to create an object store with a name that already exists (or trying to delete an object store with a name that does not already exist) will\u00a0throw an error.\u00a0<\/p><p>If the <code>onupgradeneeded<\/code> event exits successfully, the <code>onsuccess<\/code> handler of the open database request will then be triggered.\u00a0<\/p><p>Blink\/Webkit supports the current version of the spec, as shipped in Chrome 23+ and Opera 17+; IE10+ does too. Other and older implementations don't implement the current version of the spec, and thus do not support the <code>indexedDB.open(name, version).onupgradeneeded<\/code>\u00a0signature yet. For more information on how to upgrade the version of the database in older Webkit\/Blink, see the <a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\/IDBDatabase#setVersion()_.0A.0ADeprecated\" title=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\/IDBDatabase#setVersion()_.0A.0ADeprecated\">IDBDatabase reference article<\/a>.<\/p><h3 id=\"Structuring_the_database\">Structuring the database<\/h3><p>Now to structure the database. IndexedDB\u00a0uses object stores rather than tables, and a single database can contain any number of object stores. Whenever a value is stored in an object store, it is associated with a key. There are several different ways that a key can be supplied depending on whether the object store uses a <a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\/Basic_Concepts_Behind_IndexedDB#gloss_keypath\" title=\"https:\/\/developer.mozilla.org\/en\/IndexedDB#gloss_key_path\">key path<\/a> or a <a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\/Basic_Concepts_Behind_IndexedDB#gloss_keygenerator\" title=\"en\/IndexedDB#gloss key generator\">key generator<\/a>.<\/p><p>The following table shows the different ways the keys are supplied:<\/p><table class=\"standard-table\"><thead><tr><th scope=\"col\">Key Path (<code>keyPath<\/code>)<\/th>\n   <th scope=\"col\">Key Generator (<code>autoIncrement<\/code>)<\/th>\n   <th scope=\"col\">Description<\/th>\n  <\/tr><\/thead><tbody><tr><td>No<\/td>\n   <td>No<\/td>\n   <td>This object store can hold any kind of value, even primitive values like numbers and strings. You must supply a separate key argument whenever you want to add a new value.<\/td>\n  <\/tr><tr><td>Yes<\/td>\n   <td>No<\/td>\n   <td>This object store can only hold JavaScript objects. The objects must have a property with the same name as the key path.<\/td>\n  <\/tr><tr><td>No<\/td>\n   <td>Yes<\/td>\n   <td>This object store can hold any kind of value. The key is generated for you automatically, or you can supply a separate key argument if you want to use a specific key.<\/td>\n  <\/tr><tr><td>Yes<\/td>\n   <td>Yes<\/td>\n   <td>This object store can only hold JavaScript objects. Usually a key is generated and the value of the generated key is stored in the object in a property with the same name as the key path. However, if such a property already exists, the value of that property is used as key rather than generating a new key.<\/td>\n  <\/tr><\/tbody><\/table><p>You can also create indices on any object store, provided the object store holds objects, not primitives. An index lets you look up the values stored in an object store using the value of a property of the stored object, rather than the object's key.<\/p><p>Additionally, indexes have the ability to enforce simple constraints on the stored data. By setting the unique flag when creating the index, the index ensures that no two objects are stored with both having the same value for the index's key path. So, for example, if you have an object store which holds a set of people, and you want to ensure that no two people have the same email address, you can use an index with the unique flag set to enforce this.<\/p><p>That may sound confusing, but this simple example should illustrate the concepts. First, we'll define some customer data to use in our example:<\/p><pre class=\"brush: js\">\/\/ This is what our customer data looks like.\nconst customerData = [\n  { ssn: \"444-44-4444\", name: \"Bill\", age: 35, email: \"bill@company.com\" },\n  { ssn: \"555-55-5555\", name: \"Donna\", age: 32, email: \"donna@home.org\" }\n];\n<\/pre><p>Of course, you wouldn't use someone's social security number as the primary key to a customer table because not everyone has a social security number, and you would store their birth date instead of their age, but let's ignore those unfortunate choices for the sake of convenience and move along.<\/p><p>Now let's look at creating an IndexedDB to store our data:<\/p><pre class=\"brush: js\">const dbName = \"the_name\";\nvar request = indexedDB.open(dbName, 2);\nrequest.onerror = function(event) {\n  \/\/ Handle errors.\n};\nrequest.onupgradeneeded = function(event) {\n  var db = event.target.result;\n  \/\/ Create an objectStore to hold information about our customers. We're\n  \/\/ going to use \"ssn\" as our key path because it's guaranteed to be\n  \/\/ unique - or at least that's what I was told during the kickoff meeting.\n  var objectStore = db.createObjectStore(\"customers\", { keyPath: \"ssn\" });\n  \/\/ Create an index to search customers by name. We may have duplicates\n  \/\/ so we can't use a unique index.\n  objectStore.createIndex(\"name\", \"name\", { unique: false });\n  \/\/ Create an index to search customers by email. We want to ensure that\n  \/\/ no two customers have the same email, so use a unique index.\n  objectStore.createIndex(\"email\", \"email\", { unique: true });\n  \/\/ Use transaction oncomplete to make sure the objectStore creation is \n  \/\/ finished before adding data into it.\n  objectStore.transaction.oncomplete = function(event) {\n    \/\/ Store values in the newly created objectStore.\n    var customerObjectStore = db.transaction(\"customers\", \"readwrite\").objectStore(\"customers\");\n    customerData.forEach(function(customer) {\n      customerObjectStore.add(customer);\n    });\n  };\n};\n<\/pre><p>As indicated previously, <code>onupgradeneeded<\/code> is the only place where you can alter the structure of the database. In it, you can create and delete object stores and build and remove indices.<\/p><div>Object stores are created with a single call to<code>createObjectStore()<\/code>. The method takes a name of the store, and a parameter object. Even though the parameter object is optional, it is very important, because it lets you define important optional properties and refine the type of object store you want to create. In our case, we've asked for an object store named \"customers\" and defined a<code>keyPath<\/code>, which is the property that makes an individual object in the store unique. That property in this example is \"ssn\" since a social security number is guaranteed to be unique. \"ssn\" must be present on every object that is stored in the<code>objectStore<\/code>.\u00a0<\/div><p>We've also asked for an index named \"name\" that looks at the <code>name<\/code> property of the stored objects.\u00a0As with <code>createObjectStore()<\/code>, <code>createIndex()<\/code> takes an optional <code>options<\/code> object that refines the type of index that you want to create. Adding objects that don't have a <code>name<\/code> property still succeeds, but the objects won't appear in the \"name\" index.<\/p><p>We can now retrieve the stored customer objects using their <code>ssn<\/code> from the object store directly, or using their name by using the index. To learn how this is done, see the section on <a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\/Using_IndexedDB#Using_an_index\" title=\"Using IndexedDB#Using an index\">using an index<\/a>.<\/p><h3 id=\"Using_a_key_generator\">Using a key generator<\/h3><p>Setting up an <code>autoIncrement <\/code>flag when creating the object store would enable the key generator for that object store. By default this flag is not set.<\/p><p>With the key generator, the key would be generated automatically as you add the value to the object store. The current number of a key generator is always set to 1 when the object store for that key generator is first created. Basically the newly auto-generated key is increased by 1 based on the previous key. The current number for a key generator never decreases, other than as a result of database operations being reverted, for example, the database transaction is aborted. Therefore deleting a record or even clearing all records from an object store never affects the object store's key generator.<\/p><p>We can create another object store with the key generator as below:<\/p><pre class=\"brush: js\">\/\/ Open the indexedDB.\nvar request = indexedDB.open(dbName, 3);\nrequest.onupgradeneeded = function (event) {\n\u00a0\u00a0 \u00a0var db = event.target.result;\n\u00a0\u00a0 \u00a0\/\/ Create another object store called \"names\" with the autoIncrement flag set as true.\u00a0\u00a0 \u00a0\n\u00a0\u00a0 \u00a0var objStore = db.createObjectStore(\"names\", { autoIncrement : true });\n\u00a0\u00a0  \/\/ Because the \"names\" object store has the key generator, the key for the name value is generated automatically.\n\u00a0\u00a0  \/\/ The added records would be like:\n\u00a0\u00a0  \/\/ key : 1 =&gt; value : \"Bill\"\n\u00a0\u00a0  \/\/ key : 2 =&gt; value : \"Donna\"\n\u00a0\u00a0 \u00a0customerData.forEach(function(customer) {\n\u00a0\u00a0 \u00a0\u00a0\u00a0 \u00a0objStore.add(customer.name);\n\u00a0\u00a0 \u00a0});\n};<\/pre><p>For more details about the key generator, please see <a rel=\"noopener\" href=\"http:\/\/www.w3.org\/TR\/IndexedDB\/#key-generator-concept\" class=\"external\">\"W3C Key Generators\"<\/a>.<\/p><h2 id=\"Adding_retrieving_and_removing_data\">Adding, retrieving, and removing data<\/h2><p>Before you can do anything with your new database, you need to start a transaction. Transactions come from the database object, and you have to specify which object stores you want the transaction to span. Once you are inside the transaction, you can access the object stores that hold your data and make your requests. Next, you need to decide if you're going to make changes to the database or if you just need to read from it. Transactions have three available modes: <code>readonly<\/code>, <code>readwrite<\/code>, and <code>versionchange<\/code>.<\/p><p>To change the \"schema\" or structure of the database\u2014which involves creating or deleting object stores or indexes\u2014the transaction must be in <code>versionchange<\/code> mode. This transaction is opened by calling the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBFactory\/open\" title=\"If an error occurs while the database connection is being opened, then an error event is fired on the request object returned from this method.\"><code>IDBFactory.open<\/code><\/a> method with a <code>version<\/code> specified. (In WebKit browsers, which have not implemented the latest specification, the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBFactory\/open\" title=\"If an error occurs while the database connection is being opened, then an error event is fired on the request object returned from this method.\"><code>IDBFactory.open<\/code><\/a> method takes only one parameter, the <code>name<\/code> of the database; then you must call <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBVersionChangeRequest\/setVersion\" title=\"For older WebKit browsers, call this method before creating or deleting an object store.\"><code>IDBVersionChangeRequest.setVersion<\/code><\/a> to establish the <code>versionchange<\/code> transaction.)<\/p><p>To read the records of an existing object store, the transaction can either be in <code>readonly<\/code> or <code>readwrite<\/code> mode. To make changes to an existing object store, the transaction must be in <code>readwrite<\/code> mode. You open such transactions with <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBDatabase\/transaction\" title=\"An IDBTransaction object.\"><code>IDBDatabase.transaction<\/code><\/a>. The method accepts two parameters: the <code>storeNames<\/code> (the scope, defined as an array of object stores that you want to access) and the <code>mode<\/code> (<code>readonly<\/code> or <code>readwrite<\/code>) for the transaction. The method returns a transaction object containing the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBIndex\/objectStore\" title=\"An IDBObjectStore.\"><code>IDBIndex.objectStore<\/code><\/a> method, which you can use to access your object store. By default, where no mode is specified, transactions open in <code>readonly<\/code> mode.<\/p><div class=\"note\"><p><strong>Note<\/strong>: As of Firefox 40, IndexedDB transactions have relaxed durability guarantees to increase performance (see <a rel=\"noopener\" href=\"https:\/\/bugzilla.mozilla.org\/show_bug.cgi?id=1112702\" class=\"external\" title=\"FIXED: Switch IndexedDB transactions to be non-durable by default\">bug\u00a01112702<\/a>.) Previously in a <code>readwrite<\/code> transaction <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBTransaction\/oncomplete\" title=\"The oncomplete event handler of the IDBTransaction interface handles the complete event, fired\u00a0when the transaction successfully completes.\"><code>IDBTransaction.oncomplete<\/code><\/a> was fired only when all data was guaranteed to have been flushed to disk. In Firefox 40+ the <code>complete<\/code> event is fired after the OS has been told to write the data but potentially before that data has actually been flushed to disk. The <code>complete<\/code> event may thus be delivered quicker than before, however, there exists a small chance that the entire transaction will be lost if the OS crashes or there is a loss of system power before the data is flushed to disk. Since such catastrophic events are rare most consumers should not need to concern themselves further. If you must ensure durability for some reason (e.g. you're storing critical data that cannot be recomputed later) you can force a transaction to flush to disk before delivering the <code>complete<\/code> event by creating a transaction using the experimental (non-standard) <code>readwriteflush<\/code> mode (see <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBDatabase\/transaction\" title=\"An IDBTransaction object.\"><code>IDBDatabase.transaction<\/code><\/a>.<\/p><\/div><p>You can speed up data access by using the right scope and mode in the transaction. Here are a couple of tips:<\/p><ul><li>When defining the scope, specify only the object stores you need. This way, you can run multiple transactions with non-overlapping scopes concurrently.<\/li>\n <li>Only specify a <code>readwrite<\/code> transaction mode when necessary. You can concurrently run multiple <code>readonly<\/code> transactions with overlapping scopes, but you can have only one <code>readwrite<\/code> transaction for an object store. To learn more, see the definition for <dfn><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/Basic_Concepts_Behind_IndexedDB#Database\">transactions<\/a><\/dfn> in the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/Basic_Concepts_Behind_IndexedDB\">Basic Concepts<\/a> article.<\/li>\n<\/ul><h3 id=\"Adding_data_to_the_database\">Adding data to the database<\/h3><p>If you've just created a database, then you probably want to write to it. Here's what that looks like:<\/p><pre class=\"brush:js;\">var transaction = db.transaction([\"customers\"], \"readwrite\");\n\/\/ Note: Older experimental implementations use the deprecated constant IDBTransaction.READ_WRITE instead of \"readwrite\".\n\/\/ In case you want to support such an implementation, you can write: \n\/\/ var transaction = db.transaction([\"customers\"], IDBTransaction.READ_WRITE);<\/pre><p>The <code>transaction()<\/code> function takes two arguments (though one is optional) and returns a transaction object. The first argument is a list of object stores that the transaction will span. You can pass an empty array if you want the transaction to span all object stores, but don't do it because the spec says an empty array should generate an InvalidAccessError. If you don't specify anything for the second argument, you get a read-only transaction. Since you want to write to it here you need to pass the <code>\"readwrite\"<\/code> flag.<\/p><p>Now that you have a transaction you need to understand its lifetime. Transactions are tied very closely to the event loop. If you make a transaction and return to the event loop without using it then the transaction will become inactive. The only way to keep the transaction active is to make a request on it. When the request is finished you'll get a DOM\u00a0event and, assuming that the request succeeded, you'll have another opportunity to extend the transaction during that callback. If you return to the event loop without extending the transaction then it will become inactive, and so on. As long as there are pending requests the transaction remains active. Transaction lifetimes are really very simple but it might take a little time to get used to. A few more examples will help, too. If you start seeing <code>TRANSACTION_INACTIVE_ERR<\/code> error codes then you've messed something up.<\/p><p>Transactions can receive DOM\u00a0events of three different types: <code>error<\/code>, <code>abort<\/code>, and <code>complete<\/code>. We've talked about the way that <code>error<\/code> events bubble, so a transaction receives error events from any requests that are generated from it. A more subtle point here is that the default behavior of an error is to abort the transaction in which it occurred. Unless you handle the error by first calling <code>stopPropagation()<\/code> on the error event then doing something else, the entire transaction is rolled back. This design forces you to think about and handle errors, but you can always add a catchall error handler to the database if fine-grained error handling is too cumbersome. If you don't handle an error event or if you call <code>abort()<\/code> on the transaction, then the transaction is rolled back and an <code>abort<\/code> event is fired on the transaction. Otherwise, after all pending requests have completed, you'll get a <code>complete<\/code> event. If you're doing lots of database operations, then tracking the transaction rather than individual requests can certainly aid your sanity.<\/p><p>Now that you have a transaction, you'll need to get the object store from it. Transactions only let you have an object store that you specified when creating the transaction. Then you can add all the data you need.<\/p><pre class=\"brush: js\">\/\/ Do something when all the data is added to the database.\ntransaction.oncomplete = function(event) {\n  alert(\"All done!\");\n};\ntransaction.onerror = function(event) {\n  \/\/ Don't forget to handle errors!\n};\nvar objectStore = transaction.objectStore(\"customers\");\ncustomerData.forEach(function(customer) {\n  var request = objectStore.add(customer);\n  request.onsuccess = function(event) {\n    \/\/ event.target.result === customer.ssn;\n  };\n});<\/pre><p>The <code>result<\/code> of a request generated from a call to <code>add()<\/code> is the key of the value that was added. So in this case, it should equal the <code>ssn<\/code> property of the object that was added, since the object store uses the <code>ssn<\/code> property for the key path. Note that the <code>add()<\/code> function requires that no object already be in the database with the same key. If you're trying to modify an existing entry, or you don't care if one exists already, you can use the <code>put()<\/code> function, as shown below in the <a href=\"#Updating_an_entry_in_the_database\">Updating an entry in the database<\/a> section.<\/p><h3 id=\"Removing_data_from_the_database\">Removing data from the database<\/h3><p>Removing data is very similar:<\/p><pre class=\"brush: js\">var request = db.transaction([\"customers\"], \"readwrite\")\n                .objectStore(\"customers\")\n                .delete(\"444-44-4444\");\nrequest.onsuccess = function(event) {\n  \/\/ It's gone!\n};<\/pre><h3 id=\"Getting_data_from_the_database\">Getting data from the database<\/h3><p>Now that the database has some info in it, you can retrieve it in several ways. First, the simple <code>get()<\/code>.\u00a0You need to provide the key to retrieve the value, like so:<\/p><pre class=\"brush: js\">var transaction = db.transaction([\"customers\"]);\nvar objectStore = transaction.objectStore(\"customers\");\nvar request = objectStore.get(\"444-44-4444\");\nrequest.onerror = function(event) {\n  \/\/ Handle errors!\n};\nrequest.onsuccess = function(event) {\n  \/\/ Do something with the request.result!\n  alert(\"Name for SSN 444-44-4444 is \" + request.result.name);\n};<\/pre><p>That's a lot of code for a \"simple\" retrieval. Here's how you can shorten it up a bit, assuming that you handle errors at the database level:<\/p><pre class=\"brush: js\">db.transaction(\"customers\").objectStore(\"customers\").get(\"444-44-4444\").onsuccess = function(event) {\n  alert(\"Name for SSN 444-44-4444 is \" + event.target.result.name);\n};<\/pre><p>See how this works? Since there's only one object store, you can avoid passing a list of object stores you need in your transaction and just pass the name as a string. Also, you're only reading from the database, so you don't need a <code>\"readwrite\"<\/code> transaction. Calling <code>transaction()<\/code> with no mode specified gives you a <code>\"readonly\"<\/code> transaction. Another subtlety here is that you don't actually save the request object to a variable. Since the DOM event has the request as its target you can use the event to get to the <code>result<\/code> property.<\/p><p>Note that you can speed up data access by limiting the scope and mode in the transaction. Here are a couple of tips:<\/p><ul><li>When defining the <a href=\"#scope\">scope<\/a>, specify only the object stores you need. This way, you can run multiple transactions with non-overlapping scopes concurrently.<\/li>\n <li>Only specify a readwrite transaction mode when necessary. You can concurrently run multiple readonly transactions with overlapping scopes, but you can have only one readwrite transaction for an object store. To learn more, see the definition for <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/Basic_Concepts_Behind_IndexedDB#gloss_transaction\">transactions in the Basic Concepts article<\/a>.<\/li>\n<\/ul><h3 id=\"Updating_an_entry_in_the_database\">Updating an entry in the database<\/h3><p>Now we've retrieved some data, updating it and inserting it back into the IndexedDB is pretty simple. Let's update the previous example somewhat:<\/p><pre class=\"brush: js\">var objectStore = db.transaction([\"customers\"], \"readwrite\").objectStore(\"customers\");\nvar request = objectStore.get(\"444-44-4444\");\nrequest.onerror = function(event) {\n  \/\/ Handle errors!\n};\nrequest.onsuccess = function(event) {\n  \/\/ Get the old value that we want to update\n  var data = event.target.result;\n  \n  \/\/ update the value(s) in the object that you want to change\n  data.age = 42;\n  \/\/ Put this updated object back into the database.\n\u00a0 var requestUpdate = objectStore.put(data);\n\u00a0\u00a0 requestUpdate.onerror = function(event) {\n\u00a0\u00a0\u00a0\u00a0 \/\/ Do something with the error\n\u00a0\u00a0 };\n\u00a0\u00a0 requestUpdate.onsuccess = function(event) {\n\u00a0\u00a0\u00a0\u00a0 \/\/ Success - the data is updated!\n\u00a0\u00a0 };\n};<\/pre><p>So here we're creating an <code>objectStore<\/code> and requesting a customer record out of it, identified by its ssn value (<code>444-44-4444<\/code>). We then put the result of that request in a variable (<code>data<\/code>), update the <code>age<\/code> property of this object, then create a second request (<code>requestUpdate<\/code>) to put the customer record back into the <code>objectStore<\/code>, overwriting the previous value.<\/p><div class=\"note\"><p><strong>Note<\/strong>: In this case we've had to specify a <code>readwrite<\/code> transaction because we want to write to the database, not just read from it.<\/p><\/div><h3 id=\"Using_a_cursor\">Using a cursor<\/h3><p>Using <code>get()<\/code> requires that you know which key you want to retrieve. If you want to step through all the values in your object store, then you can use a cursor. Here's what it looks like:<\/p><pre class=\"brush: js\">var objectStore = db.transaction(\"customers\").objectStore(\"customers\");\nobjectStore.openCursor().onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    alert(\"Name for SSN \" + cursor.key + \" is \" + cursor.value.name);\n    cursor.continue();\n  }\n  else {\n    alert(\"No more entries!\");\n  }\n};<\/pre><p>The\u00a0<code>openCursor()<\/code> function takes several arguments. First, you can limit the range of items that are retrieved by using a key range object that we'll get to in a minute. Second, you can specify the direction that you want to iterate. In the above example, we're iterating over all objects in ascending order. The success callback for cursors is a little special. The cursor object itself is the <code>result<\/code> of the request (above we're using the shorthand, so it's <code>event.target.result<\/code>). Then the actual key and value can be found on the <code>key<\/code> and <code>value<\/code> properties of the cursor object. If you want to keep going, then you have to call <code>continue()<\/code> on the cursor. When you've reached the end of the data (or if there were no entries that matched your <code>openCursor()<\/code> request) you still get a success callback, but the <code>result<\/code> property is <code>undefined<\/code>.<\/p><p>One common pattern with cursors is to retrieve all objects in an object store and add them to an array, like this:<\/p><pre class=\"brush: js\">var customers = [];\nobjectStore.openCursor().onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    customers.push(cursor.value);\n    cursor.continue();\n  }\n  else {\n    alert(\"Got all customers: \" + customers);\n  }\n};<\/pre><div class=\"note\"><p><strong>Note<\/strong>: Alternatively, you can use <code>getAll()<\/code> to handle this case (and <code>getAllKeys()<\/code>) . The following code does precisely the same thing as above:<\/p><pre class=\"brush: js\">objectStore.getAll().onsuccess = function(event) {\n  alert(\"Got all customers: \" + event.target.result);\n};<\/pre><p>There is a performance cost associated with looking at the <code>value<\/code> property of a cursor, because the object is created lazily. When you use <code>getAll()<\/code> for example, the browser must create all the objects at once. If you're just interested in looking at each of the keys, for instance, it is much more efficient to use a cursor than to use <code>getAll()<\/code>. If you're trying to get an array of all the objects in an object store, though, use <code>getAll()<\/code>.<\/p><\/div><h3 id=\"Using_an_index\">Using an index<\/h3><p>Storing customer data using the SSN\u00a0as a key is logical since the SSN\u00a0uniquely identifies an individual. (Whether this is a good idea for privacy is a different question, and outside the scope of this article.) If you need to look up a customer by name, however, you'll need to iterate over every SSN\u00a0in the database until you find the right one. Searching in this fashion would be very slow, so instead you can use an index.<\/p><pre class=\"brush: js\">\/\/ First, make sure you created index in request.onupgradeneeded:\n\/\/ objectStore.createIndex(\"name\", \"name\");\n\/\/ Otherwize you will get DOMException.\nvar index = objectStore.index(\"name\");\nindex.get(\"Donna\").onsuccess = function(event) {\n\u00a0 alert(\"Donna's SSN is \" + event.target.result.ssn);\n};\n<\/pre><p>The \"name\" cursor isn't unique, so there could be more than one entry with the <code>name<\/code> set to <code>\"Donna\"<\/code>. In that case you always get the one with the lowest key value.<\/p><p>If you need to access all the entries with a given <code>name<\/code> you can use a cursor. You can open two different types of cursors on indexes. A normal cursor maps the index property to the object in the object store. A key cursor maps the index property to the key used to store the object in the object store. The differences are illustrated here:<\/p><pre class=\"brush: js\">\/\/ Using a normal cursor to grab whole customer record objects\nindex.openCursor().onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    \/\/ cursor.key is a name, like \"Bill\", and cursor.value is the whole object.\n    alert(\"Name: \" + cursor.key + \", SSN: \" + cursor.value.ssn + \", email: \" +\u00a0cursor.value.email);\n    cursor.continue();\n  }\n};\n\/\/ Using a key cursor to grab customer record object keys\nindex.openKeyCursor().onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    \/\/ cursor.key is a name, like \"Bill\", and cursor.value is the SSN.\n    \/\/ No way to directly get the rest of the stored object.\n    alert(\"Name: \" + cursor.key + \", SSN: \" + cursor.primaryKey);\n    cursor.continue();\n  }\n};<\/pre><h3 id=\"Specifying_the_range_and_direction_of_cursors\">Specifying the range and direction of cursors<\/h3><p>If you would like to limit the range of values you see in a cursor, you can use an <code>IDBKeyRange<\/code> object and pass it as the first argument to <code>openCursor()<\/code> or <code>openKeyCursor()<\/code>. You can make a key range that only allows a single key, or one that has a lower or upper bound, or one that has both a lower and upper bound. The bound may be \"closed\" (i.e., the key range includes the given value(s)) or \"open\" (i.e., the key range does not include the given value(s)). Here's how it works:<\/p><pre class=\"brush: js\">\/\/ Only match \"Donna\"\nvar singleKeyRange = IDBKeyRange.only(\"Donna\");\n\/\/ Match anything past \"Bill\", including \"Bill\"\nvar lowerBoundKeyRange = IDBKeyRange.lowerBound(\"Bill\");\n\/\/ Match anything past \"Bill\", but don't include \"Bill\"\nvar lowerBoundOpenKeyRange = IDBKeyRange.lowerBound(\"Bill\", true);\n\/\/ Match anything up to, but not including, \"Donna\"\nvar upperBoundOpenKeyRange = IDBKeyRange.upperBound(\"Donna\", true);\n\/\/ Match anything between \"Bill\" and \"Donna\", but not including \"Donna\"\nvar boundKeyRange = IDBKeyRange.bound(\"Bill\", \"Donna\", false, true);\n\/\/ To use one of the key ranges, pass it in as the first argument of openCursor()\/openKeyCursor()\nindex.openCursor(boundKeyRange).onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    \/\/ Do something with the matches.\n    cursor.continue();\n  }\n};<\/pre><p>Sometimes you may want to iterate in descending order rather than in ascending order (the default direction for all cursors). Switching direction is accomplished by passing <code>prev<\/code> to the <code>openCursor()<\/code> function as the second argument:<\/p><pre class=\"brush: js\">objectStore.openCursor(boundKeyRange, \"prev\").onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    \/\/ Do something with the entries.\n    cursor.continue();\n  }\n};<\/pre><p>If you just want to specify a change of direction but not constrain the results shown, you can just pass in null as the first argument:<\/p><pre class=\"brush: js\">objectStore.openCursor(null, \"prev\").onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    \/\/ Do something with the entries.\n    cursor.continue();\n  }\n};<\/pre><p>Since the \"name\" index isn't unique, there might be multiple entries where <code>name<\/code> is the same. Note that such a situation cannot occur with object stores since the key must always be unique. If you wish to filter out duplicates during cursor iteration over indexes, you can pass <code>nextunique<\/code>\u00a0(or <code>prevunique<\/code>\u00a0if you're going backwards) as the direction parameter. When <code>nextunique<\/code> or <code>prevunique<\/code>\u00a0is used, the entry with the lowest key is always the one returned.<\/p><pre class=\"brush: js\">index.openKeyCursor(null, \"nextunique\").onsuccess = function(event) {\n  var cursor = event.target.result;\n  if (cursor) {\n    \/\/ Do something with the entries.\n    cursor.continue();\n  }\n};<\/pre><p>Please see \"<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBCursor?redirectlocale=en-US&amp;redirectslug=IndexedDB%2FIDBCursor#Constants\">IDBCursor Constants<\/a>\" for the valid direction arguments.<\/p><h2 id=\"Version_changes_while_a_web_app_is_open_in_another_tab\">Version changes while a web app is open in another tab<\/h2><p>When your web app changes in such a way that a version change is required for your database, you need to consider what happens if the user has the old version of your app open in one tab and then loads the new version of your app in another. When you call <code>open()<\/code> with a greater version than the actual version of the database, all other open databases must explicitly acknowledge the request before you can start making changes to the database (an <code>onblocked<\/code> event is fired until they are closed or reloaded). Here's how it works:<\/p><pre class=\"brush: js\">var openReq = mozIndexedDB.open(\"MyTestDatabase\", 2);\nopenReq.onblocked = function(event) {\n  \/\/ If some other tab is loaded with the database, then it needs to be closed\n  \/\/ before we can proceed.\n  alert(\"Please close all other tabs with this site open!\");\n};\n  \nopenReq.onupgradeneeded = function(event) {\n  \/\/ All other databases have been closed. Set everything up.\n  db.createObjectStore(\/* ... *\/);\n  useDatabase(db);\n};\n  \nopenReq.onsuccess = function(event) {\n  var db = event.target.result;\n  useDatabase(db);\n  return;\n};\nfunction useDatabase(db) {\n  \/\/ Make sure to add a handler to be notified if another page requests a version\n  \/\/ change. We must close the database. This allows the other page to upgrade the database.\n  \/\/ If you don't do this then the upgrade won't happen until the user closes the tab.\n \u00a0db.onversionchange = function(event) {\n    db.close();\n    alert(\"A new version of this page is ready. Please reload or close this tab!\");\n  };\n  \/\/ Do stuff with the database.\n}\n<\/pre><p>You should also listen for <code>VersionError<\/code> errors to handle the situation where already opened apps may initiate code leading to a new attempt to open the database, but using an outdated version.<\/p><h2 id=\"Security\">Security<\/h2><p>IndexedDB uses the same-origin principle, which means that it ties the store to the origin of the site that creates it (typically, this is the site domain or subdomain), so it cannot be accessed by any other origin.<\/p><p>Third party window content (e.g. <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/HTML\/Element\/iframe\" title=\"The HTML Inline Frame element (&lt;iframe&gt;) represents a nested browsing context, effectively embedding another HTML page into the current page.\"><code>&lt;iframe&gt;<\/code><\/a> content) can access the IndexedDB store for the origin it is embedded into, unless the browser is set to <a rel=\"noopener\" href=\"https:\/\/support.mozilla.org\/en-US\/kb\/disable-third-party-cookies\" class=\"external\">never accept third party cookies<\/a> (see <a rel=\"noopener\" href=\"https:\/\/bugzilla.mozilla.org\/show_bug.cgi?id=1147821\" class=\"external\" title=\"FIXED: Only disable IndexedDB in third-party windows when the third-party cookie preference is set\">bug\u00a01147821<\/a>.)<\/p><p>When the browser shuts down (because the user chose the Quit or Exit option), the disk containing the database is removed unexpectedly, or permissions are lost to the database store, the following things happen:<\/p><ol><li>Each transaction on every affected database (or all open databases, in the case of browser shutdown) is aborted with an <code>AbortError<\/code>. The effect is the same as if <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBTransaction\/abort\" title=\"All pending IDBRequest objects created during this transaction have their IDBRequest.error attribute set to AbortError.\"><code>IDBTransaction.abort()<\/code><\/a> is called on each transaction.<\/li>\n <li>Once all of the transactions have completed, the database connection is closed.<\/li>\n <li>Finally, the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBDatabase\" title=\"The IDBDatabase interface of the IndexedDB API provides a connection to a database; you can use an IDBDatabase object to open a transaction on your database then create, manipulate, and delete objects (data) in that database. The interface provides the only way to get and manage versions of the database.\"><code>IDBDatabase<\/code><\/a> object representing the database connection receives a <code><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Events\/close\" title=\"\/en-US\/docs\/Web\/Events\/close\">close<\/a><\/code> event. You can use the <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBDatabase\/onclose\" title=\"A function which is called when the close event is fired.\"><code>IDBDatabase.onclose<\/code><\/a> event handler to listen for these events, so that you know when a database is unexpectedly closed.<\/li>\n<\/ol><p>The behavior described above is new, and is only available as of the following browser releases: Firefox 50, Google Chrome 31 (approximately).<\/p><p>Prior to these browser versions, the transactions are aborted silently, and no <code><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Events\/close\" title=\"\/en-US\/docs\/Web\/Events\/close\">close<\/a><\/code> event is fired, so there is no way to detect an unexpected database closure.<\/p><p>Since the user can exit the browser at any time, this means that you cannot rely upon any particular transaction to complete, and on older browsers, you don't even get told when they don't complete. There are several implications of this behavior.<\/p><p>First, you should take care to always leave your database in a consistent state at the end of every transaction. For example, suppose that you are using IndexedDB to store a list of items that you allow the user to edit. You save the list after the edit by clearing the object store and then writing out the new list. If you clear the object store in one transaction and write the new list in another transaction, there is a danger that the browser will close after the clear but before the write, leaving you with an empty database. To avoid this, you should combine the clear and the write into a single transaction.\u00a0<\/p><p>Second, you should never tie database transactions to unload events. If the unload event is triggered by the browser closing, any transactions created in the unload event handler will never complete. An intuitive approach to maintaining some information across browser sessions is to read it from the database when the browser (or a particular page) is opened, update it as the user interacts with the browser, and then save it to the database when the browser (or page) closes. However, this will not work. The database transactions will be created in the unload event handler, but because they are asynchronous they will be aborted before they can execute.<\/p><p>In fact, there is no way to guarantee that IndexedDB transactions will complete, even with normal browser shutdown. See <a rel=\"noopener\" href=\"https:\/\/bugzilla.mozilla.org\/show_bug.cgi?id=870645\" class=\"external\" title=\"IndexedDB asynchronous operations fail to complete when browser is closed\">bug\u00a0870645<\/a>. As a workaround for this normal shutdown notification, you might track your transactions and add a <code>beforeunload<\/code> event to warn the user if any transactions have not yet completed at the time of unloading.<\/p><p>At least with the addition of the abort notifications and <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBDatabase\/onclose\" title=\"A function which is called when the close event is fired.\"><code>IDBDatabase.onclose<\/code><\/a>, you can know when this has happened.<\/p><h2 id=\"Locale-aware_sorting\">Locale-aware sorting<\/h2><p>Mozilla has implemented the ability to perform locale-aware sorting of IndexedDB data in Firefox 43+. By default, IndexedDB didn\u2019t handle internationalization of sorting strings at all, and everything was sorted as if it were English text. For example, b, \u00e1, z, a would be sorted as:<\/p><ul><li>a<\/li>\n <li>b<\/li>\n <li>z<\/li>\n <li>\u00e1<\/li>\n<\/ul><p>which is obviously not how users want their data to be sorted \u2014 Aaron and \u00c1aron for example should go next to one another in a contacts list. Achieving proper international sorting therefore required the entire dataset to be called into memory, and sorting to be performed by client-side JavaScript, which is not very efficient.<\/p><p>This new functionality enables developers to specify a locale when creating an index using <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBObjectStore\/createIndex\" title=\"Note that this method must be called only from a VersionChange\u00a0transaction mode callback.\"><code>IDBObjectStore.createIndex()<\/code><\/a> (check out its parameters.) In such cases, when a cursor is then used to iterate through the dataset, and you want to specify locale-aware sorting, you can use a specialized <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBLocaleAwareKeyRange\" title=\"The IDBLocaleAwareKeyRange interface of the IndexedDB API is a Firefox-specific version of IDBKeyRange \u2014 it functions in exactly the same fashion, and has the same properties and methods, but it is intended for use with IDBIndex objects when the original index had a locale value specified upon its creation (see createIndex()'s optionalParameters) \u2014 that is, it has locale aware sorting enabled.\"><code>IDBLocaleAwareKeyRange<\/code><\/a>.<\/p><p><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IDBIndex\" title=\"IDBIndex interface of the IndexedDB API provides asynchronous access to an index in a database. An index is a kind of object store for looking up records in another object store, called the referenced object store. You use this interface to retrieve data.\"><code>IDBIndex<\/code><\/a> has also had new properties added to it to specify if it has a locale specified, and what it is: <code>locale<\/code> (returns the locale if any, or null if none is specified) and <code>isAutoLocale<\/code> (returns <code>true<\/code> if the index was created with an auto locale, meaning that the platform's default locale is used, <code>false<\/code> otherwise.)<\/p><div class=\"note\"><p><strong>Note<\/strong>: This feature is currently hidden behind a flag \u2014 to enable it and experiment, go to <a>about:config<\/a> and enable <code>dom.indexedDB.experimental<\/code>.<\/p><\/div><h2 id=\"Full_IndexedDB_example\">Full IndexedDB example<\/h2><h3 id=\"HTML_Content\">HTML Content<\/h3><pre class=\"brush: html\">&lt;script type=\"text\/javascript\" src=\"https:\/\/ajax.googleapis.com\/ajax\/libs\/jquery\/1.8.3\/jquery.min.js\"&gt;&lt;\/script&gt;    \n\u00a0\u00a0\u00a0 &lt;h1&gt;IndexedDB Demo: storing blobs, e-publication example&lt;\/h1&gt;\n\u00a0\u00a0\u00a0 &lt;div class=\"note\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;p&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Works and tested with:\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/p&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;div id=\"compat\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0 &lt;div id=\"msg\"&gt;\n\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0 &lt;form id=\"register-form\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;table&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tbody&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"pub-title\" class=\"required\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Title:\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"text\" id=\"pub-title\" name=\"pub-title\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"pub-biblioid\" class=\"required\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Bibliographic ID:&lt;br\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;span class=\"note\"&gt;(ISBN, ISSN, etc.)&lt;\/span&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"text\" id=\"pub-biblioid\" name=\"pub-biblioid\"\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"pub-year\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Year:\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"number\" id=\"pub-year\" name=\"pub-year\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tbody&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tbody&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"pub-file\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 File image:\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"file\" id=\"pub-file\"\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"pub-file-url\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Online-file image URL:&lt;br\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;span class=\"note\"&gt;(same origin URL)&lt;\/span&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"text\" id=\"pub-file-url\" name=\"pub-file-url\"\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tbody&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/table&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;div class=\"button-pane\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"button\" id=\"add-button\" value=\"Add Publication\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"reset\" id=\"register-form-reset\"\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0 &lt;\/form&gt;\n\u00a0\u00a0\u00a0 &lt;form id=\"delete-form\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;table&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tbody&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"pub-biblioid-to-delete\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Bibliographic ID:&lt;br\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;span class=\"note\"&gt;(ISBN, ISSN, etc.)&lt;\/span&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"text\" id=\"pub-biblioid-to-delete\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 name=\"pub-biblioid-to-delete\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;label for=\"key-to-delete\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Key:&lt;br\/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;span class=\"note\"&gt;(for example 1, 2, 3, etc.)&lt;\/span&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/label&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"text\" id=\"key-to-delete\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 name=\"key-to-delete\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/td&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tr&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/tbody&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/table&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;div class=\"button-pane\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"button\" id=\"delete-button\" value=\"Delete Publication\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"button\" id=\"clear-store-button\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 value=\"Clear the whole store\" class=\"destructive\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0 &lt;\/form&gt;\n\u00a0\u00a0\u00a0 &lt;form id=\"search-form\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;div class=\"button-pane\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;input type=\"button\" id=\"search-list-button\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 value=\"List database content\" \/&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0 &lt;\/form&gt;\n\u00a0\u00a0\u00a0 &lt;div&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;div id=\"pub-msg\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;div id=\"pub-viewer\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/div&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;ul id=\"pub-list\"&gt;\n\u00a0\u00a0\u00a0\u00a0\u00a0 &lt;\/ul&gt;\n\u00a0\u00a0\u00a0 &lt;\/div&gt;\n<\/pre><h3 id=\"CSS_Content\">CSS Content<\/h3><pre class=\"brush: css\">body {\n\u00a0 font-size: 0.8em;\n\u00a0 font-family: Sans-Serif;\n}\nform {\n\u00a0 background-color: #cccccc;\n\u00a0 border-radius: 0.3em;\n\u00a0 display: inline-block;\n\u00a0 margin-bottom: 0.5em;\n\u00a0 padding: 1em;\n}\ntable {\n\u00a0 border-collapse: collapse;\n}\ninput {\n\u00a0 padding: 0.3em;\n\u00a0 border-color: #cccccc;\n\u00a0 border-radius: 0.3em;\n}\n.required:after {\n\u00a0 content: \"*\";\n\u00a0 color: red;\n}\n.button-pane {\n\u00a0 margin-top: 1em;\n}\n#pub-viewer {\n\u00a0 float: right;\n\u00a0 width: 48%;\n\u00a0 height: 20em;\n\u00a0 border: solid #d092ff 0.1em;\n}\n#pub-viewer iframe {\n\u00a0 width: 100%;\n\u00a0 height: 100%;\n}\n#pub-list {\n\u00a0 width: 46%;\n\u00a0 background-color: #eeeeee;\n\u00a0 border-radius: 0.3em;\n}\n#pub-list li {\n\u00a0 padding-top: 0.5em;\n\u00a0 padding-bottom: 0.5em;\n\u00a0 padding-right: 0.5em;\n}\n#msg {\n\u00a0 margin-bottom: 1em;\n}\n.action-success {\n\u00a0 padding: 0.5em;\n\u00a0 color: #00d21e;\n\u00a0 background-color: #eeeeee;\n\u00a0 border-radius: 0.2em;\n}\n.action-failure {\n\u00a0 padding: 0.5em;\n\u00a0 color: #ff1408;\n\u00a0 background-color: #eeeeee;\n\u00a0 border-radius: 0.2em;\n}\n.note {\n\u00a0 font-size: smaller;\n}\n.destructive {\n\u00a0 background-color: orange;\n}\n.destructive:hover {\n\u00a0 background-color: #ff8000;\n}\n.destructive:active {\n\u00a0 background-color: red;\n}\n<\/pre><h3 id=\"JavaScript_Content\">JavaScript Content<\/h3><pre class=\"brush: js\">(function () {\n\u00a0 var COMPAT_ENVS = [\n\u00a0\u00a0\u00a0 ['Firefox', \"&gt;= 16.0\"],\n\u00a0\u00a0\u00a0 ['Google Chrome',\n\u00a0\u00a0\u00a0\u00a0 \"&gt;= 24.0 (you may need to get Google Chrome Canary), NO Blob storage support\"]\n\u00a0 ];\n\u00a0 var compat = $('#compat');\n\u00a0 compat.empty();\n\u00a0 compat.append('&lt;ul id=\"compat-list\"&gt;&lt;\/ul&gt;');\n\u00a0 COMPAT_ENVS.forEach(function(val, idx, array) {\n\u00a0\u00a0\u00a0 $('#compat-list').append('&lt;li&gt;' + val[0] + ': ' + val[1] + '&lt;\/li&gt;');\n\u00a0 });\n\u00a0 const DB_NAME = 'mdn-demo-indexeddb-epublications';\n\u00a0 const DB_VERSION = 1; \/\/ Use a long long for this value (don't use a float)\n\u00a0 const DB_STORE_NAME = 'publications';\n\u00a0 var db;\n\u00a0 \/\/ Used to keep track of which view is displayed to avoid uselessly reloading it\n\u00a0 var current_view_pub_key;\n\u00a0 function openDb() {\n\u00a0\u00a0\u00a0 console.log(\"openDb ...\");\n\u00a0\u00a0\u00a0 var req = indexedDB.open(DB_NAME, DB_VERSION);\n\u00a0\u00a0\u00a0 req.onsuccess = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ Better use \"this\" than \"req\" to get the result to avoid problems with\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ garbage collection.\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ db = req.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0 db = this.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"openDb DONE\");\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onerror = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"openDb:\", evt.target.errorCode);\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onupgradeneeded = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"openDb.onupgradeneeded\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 var store = evt.currentTarget.result.createObjectStore(\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 DB_STORE_NAME, { keyPath: 'id', autoIncrement: true });\n\u00a0\u00a0\u00a0\u00a0\u00a0 store.createIndex('biblioid', 'biblioid', { unique: true });\n\u00a0\u00a0\u00a0\u00a0\u00a0 store.createIndex('title', 'title', { unique: false });\n\u00a0\u00a0\u00a0\u00a0\u00a0 store.createIndex('year', 'year', { unique: false });\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 \/**\n\u00a0\u00a0 * @param {string} store_name\n\u00a0\u00a0 * @param {string} mode either \"readonly\" or \"readwrite\"\n\u00a0\u00a0 *\/\n\u00a0 function getObjectStore(store_name, mode) {\n\u00a0\u00a0\u00a0 var tx = db.transaction(store_name, mode);\n\u00a0\u00a0\u00a0 return tx.objectStore(store_name);\n\u00a0 }\n\u00a0 function clearObjectStore(store_name) {\n\u00a0\u00a0\u00a0 var store = getObjectStore(DB_STORE_NAME, 'readwrite');\n\u00a0\u00a0\u00a0 var req = store.clear();\n\u00a0\u00a0\u00a0 req.onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionSuccess(\"Store cleared\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayPubList(store);\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onerror = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"clearObjectStore:\", evt.target.errorCode);\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(this.error);\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 function getBlob(key, store, success_callback) {\n\u00a0\u00a0\u00a0 var req = store.get(key);\n\u00a0\u00a0\u00a0 req.onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 var value = evt.target.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (value)\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 success_callback(value.blob);\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 \/**\n\u00a0\u00a0 * @param {IDBObjectStore=} store\n\u00a0\u00a0 *\/\n\u00a0 function displayPubList(store) {\n\u00a0\u00a0\u00a0 console.log(\"displayPubList\");\n\u00a0\u00a0\u00a0 if (typeof store == 'undefined')\n\u00a0\u00a0\u00a0\u00a0\u00a0 store = getObjectStore(DB_STORE_NAME, 'readonly');\n\u00a0\u00a0\u00a0 var pub_msg = $('#pub-msg');\n\u00a0\u00a0\u00a0 pub_msg.empty();\n\u00a0\u00a0\u00a0 var pub_list = $('#pub-list');\n\u00a0\u00a0\u00a0 pub_list.empty();\n\u00a0\u00a0\u00a0 \/\/ Resetting the iframe so that it doesn't display previous content\n\u00a0\u00a0\u00a0 newViewerFrame();\n\u00a0\u00a0\u00a0 var req;\n\u00a0\u00a0\u00a0 req = store.count();\n\u00a0\u00a0\u00a0 \/\/ Requests are executed in the order in which they were made against the\n\u00a0\u00a0\u00a0 \/\/ transaction, and their results are returned in the same order.\n\u00a0\u00a0\u00a0 \/\/ Thus the count text below will be displayed before the actual pub list\n\u00a0\u00a0\u00a0 \/\/ (not that it is algorithmically important in this case).\n\u00a0\u00a0\u00a0 req.onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 pub_msg.append('&lt;p&gt;There are &lt;strong&gt;' + evt.target.result +\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 '&lt;\/strong&gt; record(s) in the object store.&lt;\/p&gt;');\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onerror = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"add error\", this.error);\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(this.error);\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 var i = 0;\n\u00a0\u00a0\u00a0 req = store.openCursor();\n\u00a0\u00a0\u00a0 req.onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 var cursor = evt.target.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ If the cursor is pointing at something, ask for the data\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (cursor) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"displayPubList cursor:\", cursor);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 req = store.get(cursor.key);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 req.onsuccess = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var value = evt.target.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var list_item = $('&lt;li&gt;' +\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 '[' + cursor.key + '] ' +\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 '(biblioid: ' + value.biblioid + ') ' +\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 value.title +\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 '&lt;\/li&gt;');\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (value.year != null)\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 list_item.append(' - ' + value.year);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (value.hasOwnProperty('blob') &amp;&amp;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 typeof value.blob != 'undefined') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var link = $('&lt;a href=\"' + cursor.key + '\"&gt;File&lt;\/a&gt;');\n            link.on('click', function() { return false; });\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 link.on('mouseenter', function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 setInViewer(evt.target.getAttribute('href')); });\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 list_item.append(' \/ ');\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 list_item.append(link);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 } else {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 list_item.append(\" \/ No attached file\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 pub_list.append(list_item);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ Move on to the next object in store\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 cursor.continue();\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ This counter serves only to create distinct ids\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 i++;\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"No more entries\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 function newViewerFrame() {\n\u00a0\u00a0\u00a0 var viewer = $('#pub-viewer');\n\u00a0\u00a0\u00a0 viewer.empty();\n\u00a0\u00a0\u00a0 var iframe = $('&lt;iframe \/&gt;');\n\u00a0\u00a0\u00a0 viewer.append(iframe);\n\u00a0\u00a0\u00a0 return iframe;\n\u00a0 }\n\u00a0 function setInViewer(key) {\n\u00a0\u00a0\u00a0 console.log(\"setInViewer:\", arguments);\n\u00a0\u00a0\u00a0 key = Number(key);\n\u00a0\u00a0\u00a0 if (key == current_view_pub_key)\n\u00a0\u00a0\u00a0\u00a0\u00a0 return;\n\u00a0\u00a0\u00a0 current_view_pub_key = key;\n\u00a0\u00a0\u00a0 var store = getObjectStore(DB_STORE_NAME, 'readonly');\n\u00a0\u00a0\u00a0 getBlob(key, store, function(blob) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"setInViewer blob:\", blob);\n\u00a0\u00a0\u00a0\u00a0\u00a0 var iframe = newViewerFrame();\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ It is not possible to set a direct link to the\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ blob to provide a mean to directly download it.\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (blob.type == 'text\/html') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var reader = new FileReader();\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 reader.onload = (function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var html = evt.target.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 iframe.load(function() {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 $(this).contents().find('html').html(html);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 reader.readAsText(blob);\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else if (blob.type.indexOf('image\/') == 0) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 iframe.load(function() {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var img_id = 'image-' + key;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var img = $('&lt;img id=\"' + img_id + '\"\/&gt;');\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 $(this).contents().find('body').html(img);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var obj_url = window.URL.createObjectURL(blob);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 $(this).contents().find('#' + img_id).attr('src', obj_url);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 window.URL.revokeObjectURL(obj_url);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else if (blob.type == 'application\/pdf') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 $('*').css('cursor', 'wait');\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 var obj_url = window.URL.createObjectURL(blob);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 iframe.load(function() {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 $('*').css('cursor', 'auto');\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 iframe.attr('src', obj_url);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 window.URL.revokeObjectURL(obj_url);\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 iframe.load(function() {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 $(this).contents().find('body').html(\"No view available\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0 });\n\u00a0 }\n\u00a0 \/**\n\u00a0\u00a0 * @param {string} biblioid\n\u00a0\u00a0 * @param {string} title\n\u00a0\u00a0 * @param {number} year\n\u00a0\u00a0 * @param {string} url the URL of the image to download and store in the local\n\u00a0\u00a0 *\u00a0\u00a0 IndexedDB database. The resource behind this URL is subjected to the\n\u00a0\u00a0 *\u00a0\u00a0 \"Same origin policy\", thus for this method to work, the URL must come from\n\u00a0\u00a0 *\u00a0\u00a0 the same origin as the web site\/app this code is deployed on.\n\u00a0\u00a0 *\/\n\u00a0 function addPublicationFromUrl(biblioid, title, year, url) {\n\u00a0\u00a0\u00a0 console.log(\"addPublicationFromUrl:\", arguments);\n\u00a0\u00a0\u00a0 var xhr = new XMLHttpRequest();\n\u00a0\u00a0\u00a0 xhr.open('GET', url, true);\n\u00a0\u00a0\u00a0 \/\/ Setting the wanted responseType to \"blob\"\n\u00a0\u00a0\u00a0 \/\/ http:\/\/www.w3.org\/TR\/XMLHttpRequest2\/#the-response-attribute\n\u00a0\u00a0\u00a0 xhr.responseType = 'blob';\n\u00a0\u00a0\u00a0 xhr.onload = function (evt) {\n\u00a0 \u00a0   if (xhr.status == 200) {\n\u00a0 \u00a0 \u00a0 \u00a0 console.log(\"Blob retrieved\");\n\u00a0 \u00a0 \u00a0 \u00a0 var blob = xhr.response;\n\u00a0 \u00a0 \u00a0 \u00a0 console.log(\"Blob:\", blob);\n\u00a0 \u00a0 \u00a0 \u00a0 addPublication(biblioid, title, year, blob);\n\u00a0 \u00a0 \u00a0 } else {\n\u00a0 \u00a0 \u00a0 \u00a0 console.error(\"addPublicationFromUrl error:\",\n\u00a0 \u00a0 \u00a0 \u00a0 xhr.responseText, xhr.status);\n\u00a0 \u00a0 \u00a0 }\n\u00a0\u00a0\u00a0\u00a0};\n\u00a0\u00a0\u00a0 xhr.send();\n\u00a0\u00a0\u00a0 \/\/ We can't use jQuery here because as of jQuery 1.8.3 the new \"blob\"\n\u00a0\u00a0\u00a0 \/\/ responseType is not handled.\n\u00a0\u00a0\u00a0 \/\/ http:\/\/bugs.jquery.com\/ticket\/11461\n\u00a0\u00a0\u00a0 \/\/ http:\/\/bugs.jquery.com\/ticket\/7248\n\u00a0\u00a0\u00a0 \/\/ $.ajax({\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 url: url,\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 type: 'GET',\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 xhrFields: { responseType: 'blob' },\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 success: function(data, textStatus, jqXHR) {\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0\u00a0\u00a0 console.log(\"Blob retrieved\");\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0\u00a0\u00a0 console.log(\"Blob:\", data);\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0\u00a0\u00a0 \/\/ addPublication(biblioid, title, year, data);\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 },\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 error: function(jqXHR, textStatus, errorThrown) {\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0\u00a0\u00a0 console.error(errorThrown);\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"Error during blob retrieval\");\n\u00a0\u00a0\u00a0 \/\/\u00a0\u00a0 }\n\u00a0\u00a0\u00a0 \/\/ });\n\u00a0 }\n\u00a0 \/**\n\u00a0\u00a0 * @param {string} biblioid\n\u00a0\u00a0 * @param {string} title\n\u00a0\u00a0 * @param {number} year\n\u00a0\u00a0 * @param {Blob=} blob\n\u00a0\u00a0 *\/\n\u00a0 function addPublication(biblioid, title, year, blob) {\n\u00a0\u00a0\u00a0 console.log(\"addPublication arguments:\", arguments);\n\u00a0\u00a0\u00a0 var obj = { biblioid: biblioid, title: title, year: year };\n\u00a0\u00a0\u00a0 if (typeof blob != 'undefined')\n\u00a0\u00a0\u00a0\u00a0\u00a0 obj.blob = blob;\n\u00a0\u00a0\u00a0 var store = getObjectStore(DB_STORE_NAME, 'readwrite');\n\u00a0\u00a0\u00a0 var req;\n\u00a0\u00a0\u00a0 try {\n\u00a0\u00a0\u00a0\u00a0\u00a0 req = store.add(obj);\n\u00a0\u00a0\u00a0 } catch (e) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (e.name == 'DataCloneError')\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"This engine doesn't know how to clone a Blob, \" +\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \"use Firefox\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 throw e;\n\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0 req.onsuccess = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"Insertion in DB successful\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionSuccess();\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayPubList(store);\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onerror = function() {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"addPublication error\", this.error);\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(this.error);\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 \/**\n\u00a0\u00a0 * @param {string} biblioid\n\u00a0\u00a0 *\/\n\u00a0 function deletePublicationFromBib(biblioid) {\n\u00a0\u00a0\u00a0 console.log(\"deletePublication:\", arguments);\n\u00a0\u00a0\u00a0 var store = getObjectStore(DB_STORE_NAME, 'readwrite');\n\u00a0\u00a0\u00a0 var req = store.index('biblioid');\n\u00a0\u00a0\u00a0 req.get(biblioid).onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (typeof evt.target.result == 'undefined') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"No matching record found\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 return;\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0 deletePublication(evt.target.result.id, store);\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onerror = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"deletePublicationFromBib:\", evt.target.errorCode);\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 \/**\n\u00a0\u00a0 * @param {number} key\n\u00a0\u00a0 * @param {IDBObjectStore=} store\n\u00a0\u00a0 *\/\n\u00a0 function deletePublication(key, store) {\n\u00a0\u00a0\u00a0 console.log(\"deletePublication:\", arguments);\n\u00a0\u00a0\u00a0 if (typeof store == 'undefined')\n\u00a0\u00a0\u00a0\u00a0\u00a0 store = getObjectStore(DB_STORE_NAME, 'readwrite');\n\u00a0\u00a0\u00a0 \/\/ As per spec http:\/\/www.w3.org\/TR\/IndexedDB\/#object-store-deletion-operation\n\u00a0\u00a0\u00a0 \/\/ the result of the Object Store Deletion Operation algorithm is\n\u00a0\u00a0\u00a0 \/\/ undefined, so it's not possible to know if some records were actually\n\u00a0\u00a0\u00a0 \/\/ deleted by looking at the request result.\n\u00a0\u00a0\u00a0 var req = store.get(key);\n\u00a0\u00a0\u00a0 req.onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 var record = evt.target.result;\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"record:\", record);\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (typeof record == 'undefined') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"No matching record found\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 return;\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ Warning: The exact same key used for creation needs to be passed for\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ the deletion. If the key was a Number for creation, then it needs to\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ be a Number for deletion.\n\u00a0\u00a0\u00a0\u00a0\u00a0 req = store.delete(key);\n\u00a0\u00a0\u00a0\u00a0\u00a0 req.onsuccess = function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"evt:\", evt);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"evt.target:\", evt.target);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"evt.target.result:\", evt.target.result);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"delete successful\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionSuccess(\"Deletion successful\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayPubList(store);\n\u00a0\u00a0\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0\u00a0\u00a0 req.onerror = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"deletePublication:\", evt.target.errorCode);\n\u00a0\u00a0\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 };\n\u00a0\u00a0\u00a0 req.onerror = function (evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.error(\"deletePublication:\", evt.target.errorCode);\n\u00a0\u00a0\u00a0 };\n\u00a0 }\n\u00a0 function displayActionSuccess(msg) {\n\u00a0\u00a0\u00a0 msg = typeof msg != 'undefined' ? \"Success: \" + msg : \"Success\";\n\u00a0\u00a0\u00a0 $('#msg').html('&lt;span class=\"action-success\"&gt;' + msg + '&lt;\/span&gt;');\n\u00a0 }\n\u00a0 function displayActionFailure(msg) {\n\u00a0\u00a0\u00a0 msg = typeof msg != 'undefined' ? \"Failure: \" + msg : \"Failure\";\n\u00a0\u00a0\u00a0 $('#msg').html('&lt;span class=\"action-failure\"&gt;' + msg + '&lt;\/span&gt;');\n\u00a0 }\n\u00a0 function resetActionStatus() {\n\u00a0\u00a0\u00a0 console.log(\"resetActionStatus ...\");\n\u00a0\u00a0\u00a0 $('#msg').empty();\n\u00a0\u00a0\u00a0 console.log(\"resetActionStatus DONE\");\n\u00a0 }\n\u00a0 function addEventListeners() {\n\u00a0\u00a0\u00a0 console.log(\"addEventListeners\");\n\u00a0\u00a0\u00a0 $('#register-form-reset').click(function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 resetActionStatus();\n\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0 $('#add-button').click(function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"add ...\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 var title = $('#pub-title').val();\n\u00a0\u00a0\u00a0\u00a0\u00a0 var biblioid = $('#pub-biblioid').val();\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (!title || !biblioid) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"Required field(s) missing\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 return;\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0 var year = $('#pub-year').val();\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (year != '') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ Better use Number.isInteger if the engine has EcmaScript 6\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (isNaN(year))\u00a0 {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"Invalid year\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 return;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 year = Number(year);\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 year = null;\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0 var file_input = $('#pub-file');\n\u00a0\u00a0\u00a0\u00a0\u00a0 var selected_file = file_input.get(0).files[0];\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"selected_file:\", selected_file);\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ Keeping a reference on how to reset the file input in the UI once we\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ have its value, but instead of doing that we rather use a \"reset\" type\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ input in the HTML form.\n\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/file_input.val(null);\n\u00a0\u00a0\u00a0\u00a0\u00a0 var file_url = $('#pub-file-url').val();\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (selected_file) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 addPublication(biblioid, title, year, selected_file);\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else if (file_url) {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 addPublicationFromUrl(biblioid, title, year, file_url);\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 addPublication(biblioid, title, year);\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0 $('#delete-button').click(function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 console.log(\"delete ...\");\n\u00a0\u00a0\u00a0\u00a0\u00a0 var biblioid = $('#pub-biblioid-to-delete').val();\n\u00a0\u00a0\u00a0\u00a0\u00a0 var key = $('#key-to-delete').val();\n\u00a0\u00a0\u00a0\u00a0\u00a0 if (biblioid != '') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 deletePublicationFromBib(biblioid);\n\u00a0\u00a0\u00a0\u00a0\u00a0 } else if (key != '') {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \/\/ Better use Number.isInteger if the engine has EcmaScript 6\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 if (key == '' || isNaN(key))\u00a0 {\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 displayActionFailure(\"Invalid key\");\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 return;\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 key = Number(key);\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 deletePublication(key);\n\u00a0\u00a0\u00a0\u00a0\u00a0 }\n\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0 $('#clear-store-button').click(function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 clearObjectStore();\n\u00a0\u00a0\u00a0 });\n\u00a0\u00a0\u00a0 var search_button = $('#search-list-button');\n\u00a0\u00a0\u00a0 search_button.click(function(evt) {\n\u00a0\u00a0\u00a0\u00a0\u00a0 displayPubList();\n\u00a0\u00a0\u00a0 });\n\u00a0 }\n\u00a0 openDb();\n\u00a0 addEventListeners();\n})(); \/\/ Immediately-Invoked Function Expression (IIFE)\n<\/pre><p><a rel=\"noopener\" href=\"https:\/\/mdn.mozillademos.org\/en-US\/docs\/Web\/API\/IndexedDB_API\/Using_IndexedDB$samples\/Full_IndexedDB_example?revision=1427067\" class=\"external\">Test the online live demo<\/a><\/p><h2 id=\"See_also\">See also<\/h2><p>Further reading for you to find out more information if desired.<\/p><h3 id=\"Reference\">Reference<\/h3><ul><li><a href=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\" title=\"https:\/\/developer.mozilla.org\/en\/IndexedDB\">IndexedDB\u00a0API\u00a0Reference<\/a><\/li>\n <li><a rel=\"noopener\" href=\"http:\/\/www.w3.org\/TR\/IndexedDB\/\" class=\"external\" title=\"http:\/\/www.w3.org\/TR\/IndexedDB\/\">Indexed Database API Specification<\/a><\/li>\n <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/IndexedDB\/Using_IndexedDB_in_chrome\" title=\"\/en-US\/docs\/IndexedDB\/Using_IndexedDB_in_chrome\">Using IndexedDB in chrome<\/a><\/li>\n <li><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/IndexedDB_API\/Using_JavaScript_Generators_in_Firefox\">Using JavaScript generators in Firefox<\/a><\/li>\n <li>IndexedDB <a rel=\"noopener\" href=\"https:\/\/mxr.mozilla.org\/mozilla-central\/find?text=&amp;string=dom%2FindexedDB%2F.*%5C.idl&amp;regexp=1\" class=\"external link-https\" title=\"https:\/\/mxr.mozilla.org\/mozilla-central\/find?text=&amp;string=dom\/indexedDB\/.*\\.idl&amp;regexp=1\">interface files<\/a> in the Firefox source code<\/li>\n<\/ul><h3 id=\"Tutorials_and_guides\">Tutorials and guides<\/h3><ul><li><a rel=\"noopener\" href=\"http:\/\/www.html5rocks.com\/en\/tutorials\/indexeddb\/uidatabinding\/\" class=\"external\" title=\"http:\/\/www.html5rocks.com\/en\/tutorials\/indexeddb\/uidatabinding\/\">Databinding UI Elements with IndexedDB<\/a><\/li>\n <li><a rel=\"noopener\" href=\"http:\/\/msdn.microsoft.com\/en-us\/scriptjunkie\/gg679063.aspx\" class=\"external\" title=\"http:\/\/msdn.microsoft.com\/en-us\/scriptjunkie\/gg679063.aspx\">IndexedDB\u00a0\u2014 The Store in Your Browser<\/a><\/li>\n<\/ul><h3 id=\"Libraries\">Libraries<\/h3><ul><li><a rel=\"noopener\" href=\"https:\/\/localforage.github.io\/localForage\/\" class=\"external\">localForage<\/a>: A Polyfill providing a simple name:value syntax for client-side data storage, which uses IndexedDB in the background, but falls back to WebSQL and then localStorage in browsers that don't support IndexedDB.<\/li>\n <li><a rel=\"noopener\" href=\"http:\/\/www.dexie.org\/\" class=\"external\">dexie.js<\/a>: A wrapper for IndexedDB that allows much faster code development via nice, simple syntax.<\/li>\n <li><a rel=\"noopener\" href=\"https:\/\/github.com\/erikolson186\/zangodb\" class=\"external\">ZangoDB<\/a>: A MongoDB-like interface for IndexedDB that supports most of the familiar filtering, projection, sorting, updating and aggregation features of MongoDB.<\/li>\n <li><a rel=\"noopener\" href=\"http:\/\/jsstore.net\/\" class=\"external\">JsStore<\/a>: A simple and advanced IndexedDB wrapper having SQL like syntax.\u00a0<\/li>\n<\/ul>","created_at":"2018-11-07T16:44:49+0000","updated_at":"2018-11-07T16:44:49+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":46,"domain_name":"developer.mozilla.org","preview_picture":"https:\/\/developer.mozilla.org\/static\/img\/opengraph-logo.72382e605ce3.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":323,"title":"Accidentally Turing-Complete","url":"http:\/\/beza1e1.tuxen.de\/articles\/accidentally_turing_complete.html","content":"<p>Some things were not supposed to be\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Turing_completeness\">Turing-complete<\/a>.\nThis is a collection of such accidents.<\/p><p>Stuff which is somehow limited\n(stack overflows, arbitrary configuration, etc)\nis still considered Turing complete,\nsince all \"physical\" Turing machines are resource limited.<\/p><h2>C++ Templates<\/h2><p>Although they were initially not supposed to,\nC++ templates are Turing-complete.\nFor proof look at this\n<a href=\"http:\/\/web.archive.org\/web\/20131101122512\/http:\/\/ubietylab.net\/ubigraph\/content\/Papers\/pdf\/CppTuring.pdf\">paper (pdf)<\/a>.<\/p><h2>TypeScript Type System<\/h2><p>In a very similar fashion to C++,\nthe type system of TypeScript can be used to <a href=\"https:\/\/github.com\/Microsoft\/TypeScript\/issues\/14833\">implement a prime check<\/a>.<\/p><h2>Java Generics<\/h2><p>Although Java set out to fix the errors of C++,\nit also fell into Turing-completeness eventually.\n<a href=\"https:\/\/arxiv.org\/pdf\/1605.05274.pdf\">Radu Grigore built a  parser generator for fluent interfaces<\/a>.<\/p><h2>X86 MMU<\/h2><p>The page fault handling in X86 can be used to <a href=\"https:\/\/github.com\/jbangert\/trapcc\">implement<\/a>\na simple machine.\nBasically, a page fault pushes a word to the stack, which might underflow generating another trap.\nThis mechanism provides a \"subtract and branch if less than or equal to zero\" instruction.\nEnough to implement a Turing machine.<\/p><iframe width=\"420\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/eSRcvrVs5ug\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><p>Also see <a href=\"http:\/\/www.youtube.com\/watch?v=NGXvJ1GKBKM\">the talk<\/a>.<\/p><h2>Magic: The Gathering<\/h2><p>Magic is a card game.\n<a href=\"http:\/\/boingboing.net\/2012\/09\/12\/magic-the-gathering.html\">Apparently<\/a>,\nthe rules are complex enough to reach Turing-completeness.<\/p><h2>HTML5 + CSS3<\/h2><p>While older HTML\/CSS versions are not,\nwith some new additions, you <a href=\"http:\/\/lambda-the-ultimate.org\/node\/4222\">can<\/a>\nwrite a\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Rule_110\">rule 110 automaton<\/a>.\nThe original page has <a href=\"http:\/\/eli.fox-epste.in\/rule110-full.html\">moved here<\/a>\nand there is also <a href=\"https:\/\/github.com\/elitheeli\/stupid-machines\/tree\/master\/rule110\">Github<\/a>\nand a video.<\/p><iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/Ak_sWZyHi3E\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><h2>Minecraft<\/h2><p>Maybe this one was intentional,\nbut the complexity of the computers people build in Minecraft is impressive.<\/p><iframe width=\"420\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/7sNge0Ywz-M\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><h2>Dwarf Fortress<\/h2><p>In a similar vein,\nthe infamous Dwarf Fortress game provides various ways\nto build logical AND, OR, or NAND gates for\n<a href=\"http:\/\/dwarffortresswiki.org\/index.php\/DF2014:Computing\">computing<\/a>.\nA very impressive example is a space invaders game\nin the following video.<\/p><h2>SQL<\/h2><p>SQL is usually not considered to be Turing-complete.\nHowever, with the features Common Table Expressions and Windowing,\n<a href=\"http:\/\/wiki.postgresql.org\/wiki\/Cyclic_Tag_System\">SQL is Turing Complete<\/a>.\nThe proof is in <a href=\"http:\/\/assets.en.oreilly.com\/1\/event\/27\/High%20Performance%20SQL%20with%20PostgreSQL%20Presentation.pdf\">these slides<\/a>.\nAlso, look a <a href=\"http:\/\/wiki.postgresql.org\/wiki\/Mandelbrot_set\">Mandelbrot in SQL:2008<\/a>.\nA <a href=\"http:\/\/blog.coelho.net\/database\/2013\/08\/17\/turing-sql-1\/\">broader exploration of various SQL-turing-machines<\/a>\nwas done by Fabien Coelho.<\/p><h2>(C Preprocessor)<\/h2><p>The C preprocessor is only Turing-complete\nif executed in a loop\nwhich feeds the output to the input ad infinitum.\nAn example has won the\n<a href=\"http:\/\/www.ioccc.org\/years.html#2001\">IOCCC 2001<\/a> contest.\nLook into the herrmann1 entry.\nNonetheless included in this list for coolness.<\/p><h2>Apache Rewrite Rules<\/h2><p>Apache comes with the mod_rewrite plugin to rewrite URLs.\nThese rules ultimately are a Turing complete\n<a href=\"http:\/\/olsner.se\/2008\/01\/21\/an-excursion-in-mod_rewrite\/\">string rewrite system<\/a>.\nThough the default configuration sets very low limits to recursion.<\/p><h2>(Pokemon Yellow)<\/h2><p>A Pokemon game, which is <a href=\"http:\/\/tasvideos.org\/2913S.html\">finished in 1minute 36seconds<\/a>.\nThe interesting point about this speedrun is the bug it exploits.\nTurns out the game logic itself is Turing-complete in the sense\nthat you can rewrite the assembly itself with game actions.\nFor example, someone turned the <a href=\"http:\/\/aurellem.org\/vba-clojure\/html\/total-control.html\">game into a MIDI player<\/a>.<\/p><h2>Scala Type System<\/h2><p>Apparently, you can <a href=\"http:\/\/michid.wordpress.com\/2010\/01\/29\/scala-type-level-encoding-of-the-ski-calculus\/\">implement the SKI calculus<\/a>\n(which is Turing complete)\nin Scala types.\nHowever, the compilers stack overflows at some point.<\/p><h2>MediaWiki Templates<\/h2><p>In MediaWiki you can define <a href=\"http:\/\/en.wikipedia.org\/wiki\/Wikipedia:Template_messages\">templates<\/a>.\nSince they provide recursion,\nyou can apparently <a href=\"https:\/\/web.archive.org\/web\/20130101055126\/http:\/\/mentalpolyphonics.com\/posts\/wikimedia-proves-greenspuns-tenth-law\">implement lambda calculus<\/a>.<\/p><h2>Little Big Planet<\/h2><p>Yet another game, where you can build a basic computer.\nFor example, Conways Game of Life:<\/p><iframe width=\"420\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/13GOFa1C4e4\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><h2>Server Side Includes<\/h2><p>While <a href=\"http:\/\/en.wikipedia.org\/wiki\/Server_Side_Includes\">SSI<\/a>\nwas design as a scripting language,\nloops were not planned.\nThe trick was to find <a href=\"http:\/\/www.janschejbal.de\/projekte\/ssituring\/\">a way for recursion<\/a>.\nThe recursion is limited in all web servers.<\/p><h2>Sendmail<\/h2><p>The venerable mail server is known for its arcane configuration.\n<a href=\"http:\/\/okmij.org\/ftp\/Computation\/sendmail-as-turing-machine.txt\">Turns out<\/a>\nthe configuration is also turing complete.<\/p><h2>Vim Normal-Mode<\/h2><p>Let me just quote verbatim from <a href=\"https:\/\/github.com\/ealter\/vim_turing_machine\">Github Readme<\/a>:<\/p><p>Ever wish you could run your code in your editor? Tired of installing huge dependencies like bash or python to run your scripts?\nLove Vim so much that you never want to leave it? Why not run your code... in your editor itself?\nEnter <a href=\"https:\/\/github.com\/ealter\/vim_turing_machine\">vim_turing_machine<\/a>:\na tool to allow you to run a Turing machine using only <strong>normal mode<\/strong> Vim commands.<\/p><p>And now you might ask, but what can we do on a Turing machine! To demonstrate its capabilities, we implemented a solution to the Merge Overlapping Intervals question and defined all the state transitions needed to solve this glorious problem.\nSo next time you need to merge some intervals, don't hand-write a 10-line python program.\nInstead, take out your favorite editor and watch it solve the problem in less than a minute with 1400 state transitions!<\/p><p>But a simple naysayer may say, 'We already have vimscript!\nWhy in God's name would I want to use a Turing machine instead?'\nTo that, we retort: our Turing machine only uses normal mode.\nSo you could theoretically just type in the program and then execute it without running a single script!\nNo ex mode either! This project proves that normal mode in Vim is as powerful as any computer!<\/p><h2>Border Gateway Protocol (BGP)<\/h2><p>BGP is an internet backbone technology,\nwhich manages the routes your data packages take.\nThe paper \"<a href=\"http:\/\/vanbever.eu\/pdfs\/vanbever_turing_icnp_2013.pdf\">Using Routers to Build Logic Circuits:\nHow Powerful is BGP?<\/a>\"\nprove Turing-Completeness via logic gate comparison.<\/p><h2>Excel<\/h2><p>It is no surprise, since Excel includes a scripting language.\nHowever,\nthere is a nice blog post,\nwhich shows how to <a href=\"http:\/\/www.felienne.com\/archives\/2974\">encode a Turing machine in Excel<\/a>\nusing only formulas.<\/p><h2>Super Mario World<\/h2><p>There are glitches in Super Mario World,\nsuch that you can write arbitrary values into some unused parts of the memory\nand then execute them.\nHere is a video,\nwhere a human (not a script!)\ncreates a playable Flappy Bird clone this way.<\/p><iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/hB6eY73sLV0\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><h2>PowerPoint<\/h2><p>Tom Wildenhain made a hilarious presentation (see video below)\nand also wrote <a href=\"https:\/\/www.andrew.cmu.edu\/user\/twildenh\/PowerPointTM\/Paper.pdf\">a paper \"On the Turing Completeness of MS PowerPoint<\/a>.\nIt answers the question \"whether any other applications are necessary at all,\nor if all computational tasks can be accomplished through the creation of\ndedicated .pptx files\" with \"Yes\".\nOf course, he does not use VBScript or Macros.\nWildenhain only used AutoShape, Animation, and hyperlinks.\nNevertheless, Powerpoint is not really Turing-complete\nas the viewer is required to click for each step of the machine.<\/p><iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/uNjxe8ShM-8\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><hr \/><p>Also interesting:\n<a href=\"http:\/\/beza1e1.tuxen.de\/articles\/when_goto_is_fine.html\">When Goto Is Just Fine<\/a>,\n<a href=\"http:\/\/beza1e1.tuxen.de\/one_letter_proglangs.html\">One Letter Programming Languages<\/a>,\nand <a href=\"http:\/\/beza1e1.tuxen.de\/articles\/titles.html\">Hacker Titles for Business Cards<\/a>.<\/p>","created_at":"2018-11-07T16:44:49+0000","updated_at":"2018-11-07T16:44:49+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"beza1e1.tuxen.de","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":320,"title":"Show HN: OpenBSD Email Service \u2013 A free-email alternative","url":"https:\/\/news.ycombinator.com\/item?id=16392096","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kennu\" class=\"hnuser\">kennu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392896\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jstanley\" class=\"hnuser\">jstanley<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393800\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rightos\" class=\"hnuser\">rightos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393844\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sfilargi\" class=\"hnuser\">sfilargi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395431\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=snksnk\" class=\"hnuser\">snksnk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395744\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sfilargi\" class=\"hnuser\">sfilargi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395988\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=snksnk\" class=\"hnuser\">snksnk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16396155\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sfilargi\" class=\"hnuser\">sfilargi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16396214\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tux1968\" class=\"hnuser\">tux1968<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16399419\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"360\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=justinclift\" class=\"hnuser\">justinclift<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16402484\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=phaer\" class=\"hnuser\">phaer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397052\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=npsimons\" class=\"hnuser\">npsimons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394354\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacobush\" class=\"hnuser\">jacobush<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393938\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=znpy\" class=\"hnuser\">znpy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394611\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thunfischbrot\" class=\"hnuser\">thunfischbrot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16396789\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=phaer\" class=\"hnuser\">phaer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397083\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=itsjloh\" class=\"hnuser\">itsjloh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16400585\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tux1968\" class=\"hnuser\">tux1968<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16399428\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelbuckbee\" class=\"hnuser\">michaelbuckbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394391\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394898\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=justinclift\" class=\"hnuser\">justinclift<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16402500\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393155\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mseebach\" class=\"hnuser\">mseebach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393672\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jbergstroem\" class=\"hnuser\">jbergstroem<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394411\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dasyatidprime\" class=\"hnuser\">dasyatidprime<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394673\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jbergstroem\" class=\"hnuser\">jbergstroem<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395927\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jbergstroem\" class=\"hnuser\">jbergstroem<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16405706\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cortesoft\" class=\"hnuser\">cortesoft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16396101\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ozim\" class=\"hnuser\">ozim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393813\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelmrose\" class=\"hnuser\">michaelmrose<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394874\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mseebach\" class=\"hnuser\">mseebach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394980\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ozim\" class=\"hnuser\">ozim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395377\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mseebach\" class=\"hnuser\">mseebach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395756\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mseebach\" class=\"hnuser\">mseebach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394192\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jms703\" class=\"hnuser\">jms703<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394885\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ozim\" class=\"hnuser\">ozim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395414\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cortesoft\" class=\"hnuser\">cortesoft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16396060\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mseebach\" class=\"hnuser\">mseebach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395768\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hueving\" class=\"hnuser\">hueving<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393785\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mseebach\" class=\"hnuser\">mseebach<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394179\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=deusex_\" class=\"hnuser\">deusex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393304\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eeks\" class=\"hnuser\">eeks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393689\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrismeller\" class=\"hnuser\">chrismeller<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394635\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eikenberry\" class=\"hnuser\">eikenberry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395505\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kiwijamo\" class=\"hnuser\">kiwijamo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395812\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eikenberry\" class=\"hnuser\">eikenberry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16409271\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=efficax\" class=\"hnuser\">efficax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395914\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=webaholic\" class=\"hnuser\">webaholic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393065\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=splitrocket\" class=\"hnuser\">splitrocket<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393334\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dqv\" class=\"hnuser\">dqv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393951\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=waz0wski\" class=\"hnuser\">waz0wski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16398275\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392937\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BartBoch\" class=\"hnuser\">BartBoch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16400895\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393088\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickpsecurity\" class=\"hnuser\">nickpsecurity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393355\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393514\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickpsecurity\" class=\"hnuser\">nickpsecurity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393993\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16411660\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mhd\" class=\"hnuser\">mhd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393244\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gtaylor\" class=\"hnuser\">gtaylor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393332\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LoSboccacc\" class=\"hnuser\">LoSboccacc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393089\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16411666\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=a012\" class=\"hnuser\">a012<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392143\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eikenberry\" class=\"hnuser\">eikenberry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395635\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pavon\" class=\"hnuser\">pavon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16396860\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392178\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392956\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=decasteve\" class=\"hnuser\">decasteve<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394362\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dozzie\" class=\"hnuser\">dozzie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393017\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gaadd33\" class=\"hnuser\">gaadd33<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393196\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kbenson\" class=\"hnuser\">kbenson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394237\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=haroldp\" class=\"hnuser\">haroldp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394676\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393184\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reaperducer\" class=\"hnuser\">reaperducer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393303\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393551\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=scruffyherder\" class=\"hnuser\">scruffyherder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393302\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Infernal\" class=\"hnuser\">Infernal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394569\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwaway2048\" class=\"hnuser\">throwaway2048<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395326\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dozzie\" class=\"hnuser\">dozzie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392105\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tharkun\" class=\"hnuser\">Tharkun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393145\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=semanticist\" class=\"hnuser\">semanticist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393195\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tharkun\" class=\"hnuser\">Tharkun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394379\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=semanticist\" class=\"hnuser\">semanticist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16406222\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amdavidson\" class=\"hnuser\">amdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394950\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=semanticist\" class=\"hnuser\">semanticist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16406264\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_eht\" class=\"hnuser\">_eht<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393604\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gsich\" class=\"hnuser\">gsich<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393733\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394998\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=a012\" class=\"hnuser\">a012<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16398425\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gsich\" class=\"hnuser\">gsich<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16395109\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CaptSpify\" class=\"hnuser\">CaptSpify<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397638\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=egwynn\" class=\"hnuser\">egwynn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393269\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tharkun\" class=\"hnuser\">Tharkun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394465\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lunatic666\" class=\"hnuser\">Lunatic666<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16399479\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tharkun\" class=\"hnuser\">Tharkun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16400476\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=peterwwillis\" class=\"hnuser\">peterwwillis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393513\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393265\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392961\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dozzie\" class=\"hnuser\">dozzie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393029\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16411684\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h0r14\" class=\"hnuser\">h0r14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16392172\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=giancarlostoro\" class=\"hnuser\">giancarlostoro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397072\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=emptybits\" class=\"hnuser\">emptybits<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397445\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16411671\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=elchief\" class=\"hnuser\">elchief<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397230\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=INTPenis\" class=\"hnuser\">INTPenis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397775\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=perlgod\" class=\"hnuser\">perlgod<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393463\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fwdpropaganda\" class=\"hnuser\">fwdpropaganda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393984\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tharkun\" class=\"hnuser\">Tharkun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394508\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=haroldp\" class=\"hnuser\">haroldp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394385\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=peterwwillis\" class=\"hnuser\">peterwwillis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393531\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=perlgod\" class=\"hnuser\">perlgod<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393587\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DrPhish\" class=\"hnuser\">DrPhish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393637\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bachmeier\" class=\"hnuser\">bachmeier<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393827\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=peterwwillis\" class=\"hnuser\">peterwwillis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393976\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DrPhish\" class=\"hnuser\">DrPhish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394327\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Milner08\" class=\"hnuser\">Milner08<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393597\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jlg23\" class=\"hnuser\">jlg23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393610\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=HighPlainsDrftr\" class=\"hnuser\">HighPlainsDrftr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393717\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dboreham\" class=\"hnuser\">dboreham<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393916\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=haroldp\" class=\"hnuser\">haroldp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394433\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CaptSpify\" class=\"hnuser\">CaptSpify<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397686\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Gracana\" class=\"hnuser\">Gracana<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394861\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hapless\" class=\"hnuser\">hapless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16393955\">8 months ago<\/a>  [flagged]<\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=drewpc\" class=\"hnuser\">drewpc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394742\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hapless\" class=\"hnuser\">hapless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16407060\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nbsd4lyfe\" class=\"hnuser\">nbsd4lyfe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394045\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hapless\" class=\"hnuser\">hapless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394221\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nbsd4lyfe\" class=\"hnuser\">nbsd4lyfe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16394744\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16397664\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hapless\" class=\"hnuser\">hapless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16407056\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16414018\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Tepix\" class=\"hnuser\">Tepix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16411694\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=upofadown\" class=\"hnuser\">upofadown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16414699\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16413930\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hapless\" class=\"hnuser\">hapless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16407054\">8 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:44:48+0000","updated_at":"2018-11-07T16:44:48+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":321,"title":"cknadler\/vim-anywhere","url":"https:\/\/github.com\/cknadler\/vim-anywhere","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Sometimes, you edit text outside of Vim. These are sad times. Enter\nvim-anywhere!<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/cknadler\/vim-anywhere\/blob\/master\/assets\/demo.gif\"><img src=\"https:\/\/github.com\/cknadler\/vim-anywhere\/raw\/master\/assets\/demo.gif\" alt=\"demo\" \/><\/a><\/p>\n<p>Once <a href=\"#keybinding\">invoked<\/a>, vim-anywhere will open a buffer. Close it and its\ncontents are copied to your <strong>clipboard<\/strong> and your previous application is\nrefocused.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<h4><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h4>\n<p><strong>OSX:<\/strong><\/p>\n<ul><li>MacVim (<code>brew cask install macvim<\/code>)<\/li>\n<\/ul><p><strong>Linux:<\/strong><\/p>\n<ul><li>Gnome (or a derivative)<\/li>\n<li>gVim<\/li>\n<\/ul><h4><a id=\"user-content-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#install\"><\/a>Install<\/h4>\n<div class=\"highlight highlight-source-shell\"><pre>curl -fsSL https:\/\/raw.github.com\/cknadler\/vim-anywhere\/master\/install | bash<\/pre><\/div>\n<h4><a id=\"user-content-update\" class=\"anchor\" aria-hidden=\"true\" href=\"#update\"><\/a>Update<\/h4>\n<div class=\"highlight highlight-source-shell\"><pre>~\/.vim-anywhere\/update<\/pre><\/div>\n<h4><a id=\"user-content-uninstall\" class=\"anchor\" aria-hidden=\"true\" href=\"#uninstall\"><\/a>Uninstall<\/h4>\n<div class=\"highlight highlight-source-shell\"><pre>~\/.vim-anywhere\/uninstall<\/pre><\/div>\n<h2><a id=\"user-content-keybinding\" class=\"anchor\" aria-hidden=\"true\" href=\"#keybinding\"><\/a>Keybinding<\/h2>\n<p><strong>OSX:<\/strong> ( default = <code>ctrl+cmd+v<\/code> )<\/p>\n<p>You can adjust the shortcut via <a href=\"https:\/\/github.com\/cknadler\/vim-anywhere\/blob\/master\/assets\/shortcut.png\">system preferences<\/a>.<\/p>\n<pre>System Preferences &gt; Keyboard &gt; Shortcuts &gt; Services &gt; Vim Anywhere\n<\/pre>\n<p><strong>Linux:<\/strong> ( default = <code>ctrl+alt+v<\/code> )<\/p>\n<p><em>Gnome<\/em><\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ gconftool -t str --set \/desktop\/gnome\/keybindings\/vim-anywhere\/binding &lt;custom binding&gt;<\/pre><\/div>\n<p><em>I3WM<\/em><\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ echo 'bindsym $mod+Alt+v exec ~\/.vim-anywhere\/bin\/run' &gt;&gt; ~\/.i3\/config # remember to reload your config after<\/pre><\/div>\n<p>Adjust in case <code>$mod<\/code> is not set to ctrl.<\/p>\n<h2><a id=\"user-content-history\" class=\"anchor\" aria-hidden=\"true\" href=\"#history\"><\/a>History<\/h2>\n<p>vim-anywhere creates a temporary file in <code>\/tmp\/vim-anywhere<\/code> when invoked. These\nfiles stick around until you restart your system, giving you a temporary\nhistory.<\/p>\n<p>View your history:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ ls \/tmp\/vim-anywhere<\/pre><\/div>\n<p>Reopen your most recent file:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ vim $( ls \/tmp\/vim-anywhere | sort -r | head -n 1 )<\/pre><\/div>\n<h2><a id=\"user-content-why\" class=\"anchor\" aria-hidden=\"true\" href=\"#why\"><\/a>Why?<\/h2>\n<p>I use Vim for <em>almost<\/em> everything. I wish I didn't have to say <em>almost<\/em>. My\nusual workflow is to open Vim, write, copy the text out of my current buffer\nand paste it into whatever application I was just using. vim-anywhere attempts\nto automate this process as much as possible, reducing the friction of using\nVim to do more than just edit code.<\/p>\n<h2><a id=\"user-content-bugs\" class=\"anchor\" aria-hidden=\"true\" href=\"#bugs\"><\/a>Bugs<\/h2>\n<p>First, make sure to <a href=\"https:\/\/github.com\/cknadler\/vim-anywhere\/blob\/master\/FAQ.md\">read the FAQ<\/a>. If you don't find the answer you're\nlooking for there, feel free to open an issue.<\/p>\n<p>Pull requests, suggestions and issues of any kind are welcome. <strong>Make sure\nto check out the <a href=\"https:\/\/github.com\/cknadler\/vim-anywhere\/blob\/master\/CONTRIBUTING.md\">contribution guidelines<\/a> before you submit a\npull request.<\/strong><\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>MIT.<\/p>\n<\/article>","created_at":"2018-11-07T16:44:48+0000","updated_at":"2018-11-07T16:44:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/476505?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":319,"title":"Install and Use GNU Command Line Tools on macOS\/OS X","url":"https:\/\/www.topbug.net\/blog\/2013\/04\/14\/install-and-use-gnu-command-line-tools-in-mac-os-x\/","content":"<p>If you are moving onto macOS\/OS X from GNU\/Linux, you would probably find out that the command line tools shipped with OS X are not as powerful and easy to use as the tools in Linux. The reason is that macOS\/Mac OS X uses the BSD version command line tools, which are different from the Linux version, while they are both compliant with <a class=\"reference external\" href=\"http:\/\/en.wikipedia.org\/wiki\/POSIX\">POSIX<\/a> standards. But we can easily install the GNU command line tools by using <a class=\"reference external\" href=\"http:\/\/brew.sh\">Homebrew<\/a> in Mac OS X and set them as\u00a0default.<\/p>\n<p>Note: you need to notice that you may have some compatibility issues with shell scripts written specifically for OS X after you have replaced your OS X commands with the GNU version. Although the very vast majority of shell scripts have no problem, you just need to be aware that when there comes a problem, this may be the spot to check\u00a0on.<\/p>\n<p>\n<\/p><div id=\"install-homebrew\" class=\"section\">\n<h2>Install\u00a0Homebrew<\/h2>\n<p>First, visit <a class=\"reference external\" href=\"http:\/\/brew.sh\">Homebrew<\/a> homepage and follow the installation instructions to install\u00a0Homebrew.<\/p>\n<p>Shortcut: install the latest <a class=\"reference external\" href=\"https:\/\/developer.apple.com\/xcode\/\">XCode<\/a> and then run the following command to\u00a0install:<\/p>\n<pre><code class=\"language-bash\">\/usr\/bin\/ruby -e \"$(curl -fsSL https:\/\/raw.githubusercontent.com\/Homebrew\/install\/master\/install)\"<\/code><\/pre>\n<p>Then add the following line to your <strong>.bashrc<\/strong> or <strong>.zshrc<\/strong>:<\/p>\n<pre><code class=\"language-bash\">export PATH=\"$(brew --prefix coreutils)\/libexec\/gnubin:\/usr\/local\/bin:$PATH\"<\/code><\/pre>\n<\/div>\n<div id=\"install-the-gnu-command-line-tools\" class=\"section\">\n<h2>Install the GNU Command Line\u00a0Tools<\/h2>\n<p>(<em>All the <code>brew install<\/code> commands below have been put into one script on <a class=\"reference external\" href=\"https:\/\/gist.github.com\/xuhdev\/8b1b16fb802f6870729038ce3789568f\">GitHub<\/a> for downloads.<\/em>)<br \/>\nFirst comes the most important one \u2014 <a class=\"reference external\" href=\"http:\/\/en.wikipedia.org\/wiki\/GNU_Core_Utilities\">GNU Coreutils<\/a>:<\/p>\n<pre><code class=\"language-bash\">brew install coreutils<\/code><\/pre>\n<p>GNU Coreutils contains the most essential UNIX commands, such as <strong>ls<\/strong>, <strong>cat<\/strong>.<\/p>\n<p>Then you may probably want to install the following ones (For some of the packages, you need to run <code class=\"docutils literal\">brew tap homebrew\/dupes<\/code> first, but only once for your\u00a0system):<\/p>\n<pre><code class=\"language-bash\">brew install binutils&#13;\nbrew install diffutils&#13;\nbrew install ed --with-default-names&#13;\nbrew install findutils --with-default-names&#13;\nbrew install gawk&#13;\nbrew install gnu-indent --with-default-names&#13;\nbrew install gnu-sed --with-default-names&#13;\nbrew install gnu-tar --with-default-names&#13;\nbrew install gnu-which --with-default-names&#13;\nbrew install gnutls&#13;\nbrew install grep --with-default-names&#13;\nbrew install gzip&#13;\nbrew install screen&#13;\nbrew install watch&#13;\nbrew install wdiff --with-gettext&#13;\nbrew install wget<\/code><\/pre>\n<p>The <code>--default-names<\/code> option will prevent Homebrew from prepending a <strong>g<\/strong> to each of the newly installed commands, thus we could use these commands as default commands over the ones shipped by OS\u00a0X.<\/p>\n<p>In addition, some GNU command line tools already exist by default on OS X, but you may want a newer\u00a0version:<\/p>\n<pre><code class=\"language-bash\">brew install bash&#13;\nbrew install emacs&#13;\nbrew install gdb  # gdb requires further actions to make it work. See `brew info gdb`.&#13;\nbrew install gpatch&#13;\nbrew install less&#13;\nbrew install m4&#13;\nbrew install make&#13;\nbrew install nano<\/code><\/pre>\n<p>As a complementary set of packages, the following ones are not from GNU, but you can install and use a newer version instead of the version shipped by OS X:<\/p>\n<pre><code class=\"language-bash\">brew install file-formula&#13;\nbrew install git&#13;\nbrew install openssh&#13;\nbrew install perl&#13;\nbrew install python&#13;\nbrew install rsync&#13;\nbrew install svn&#13;\nbrew install unzip&#13;\nbrew install vim --override-system-vi&#13;\nbrew install macvim --override-system-vim --custom-system-icons&#13;\nbrew install zsh<\/code><\/pre>\n<p>Now you should have an easier command line system in your OS X. Have fun with\u00a0them!<\/p>\n<\/div>\n<p><strong>Update 1<\/strong>: You may also want to add <code class=\"docutils literal\">$HOMEBREW_PREFIX\/opt\/coreutils\/libexec\/gnuman<\/code> to the <code class=\"docutils literal\">MANPATH<\/code> environmental variable, where <code class=\"docutils literal\">$HOMEBREW_PREFIX<\/code> is the prefix of Homebrew, which is <code class=\"docutils literal\">\/usr\/local<\/code> by default. (Thanks Matthew Walker!) Alternatively, there is also a one-line setup which you could put in your shell configuration files <a class=\"reference external\" href=\"https:\/\/gist.github.com\/quickshiftin\/9130153\">here<\/a> by\u00a0quickshiftin.<\/p>\n<div class=\"section\">\n<p><strong>Update 2<\/strong> All the <code>brew install<\/code> commands above have been put into one script on <a class=\"reference external\" href=\"https:\/\/gist.github.com\/xuhdev\/8b1b16fb802f6870729038ce3789568f\">GitHub<\/a> for\u00a0downloads.<\/p>\n<\/div>\n<div id=\"dpsp-content-bottom\" class=\"dpsp-content-wrapper dpsp-shape-rounded dpsp-size-medium dpsp-has-spacing dpsp-show-on-mobile dpsp-show-total-share-count dpsp-show-total-share-count-before dpsp-button-style-1\"><div class=\"dpsp-total-share-wrapper\">48shares<\/div><ul class=\"dpsp-networks-btns-wrapper dpsp-networks-btns-content dpsp-networks-btns-share dpsp-column-auto dpsp-has-button-icon-animation\"><li><\/li><li><\/li><li><\/li><li><\/li><li><\/li><li><\/li><\/ul><\/div><div class=\"crp_related\"><h3>Related Posts:<\/h3><ul><li><a href=\"https:\/\/www.topbug.net\/blog\/2014\/10\/13\/install-php-extensions-on-shared-hosts\/\">Install PHP Extensions on Shared Hosts<\/a><\/li><li><a href=\"https:\/\/www.topbug.net\/blog\/2016\/09\/27\/make-gnu-less-more-powerful\/\">Make the less Command More Powerful<\/a><\/li><li><a href=\"https:\/\/www.topbug.net\/blog\/2013\/12\/07\/back-up-homebrew-packages\/\">Back up (Migrate) Homebrew Packages<\/a><\/li><li><a href=\"https:\/\/www.topbug.net\/blog\/2016\/11\/28\/a-better-ls-command\/\">A Better ls Command<\/a><\/li><\/ul><div class=\"crp_clear\"><\/div> <\/div>","created_at":"2018-11-07T16:44:46+0000","updated_at":"2018-11-07T16:44:46+0000","published_at":"2013-04-14T00:32:00+0000","published_by":["Hong"],"annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"www.topbug.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":317,"title":"Books I had to read to develop a game engine","url":"https:\/\/www.haroldserrano.com\/blog\/books-i-used-to-develop-a-game-engine","content":"<p>If you have decided to develop your game engine, you may be wondering where to start, what books to start reading, etc. If you are in this situation, I recommend getting a copy of the following books found <a href=\"https:\/\/www.haroldserrano.com\/blog\/best-books-to-develop-a-game-engine?rq=books\">here<\/a>. Out of all the books I've read, they are the best of the best, so I strongly recommend them.  <\/p>\n<p>Now, if you want a comprehensive list, the list below should help you.   <\/p>\n<h3 id=\"books-to-develop-the-math-engine\">Books to develop the Math Engine<\/h3>\n<ol><li><a href=\"https:\/\/www.amazon.com\/Mathematics-Programming-Computer-Graphics-Third\/dp\/1435458869\/ref=sr_1_1?s=books&amp;ie=UTF8&amp;qid=1516083405&amp;sr=1-1&amp;keywords=3D+mathematics\">Mathematics for 3D Game Programming and Computer Graphics<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0691102988\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Quaternions and Rotation Sequences: A Primer with Applications to Orbits, Aerospace and Virtual Reality<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0857297597\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Quaternions for Computer Graphics<\/a> <\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1556229119\/ref=oh_aui_detailpage_o08_s00?ie=UTF8&amp;psc=1\">3D Math Primer For Graphics and Game Development<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1568812345\/ref=oh_aui_detailpage_o04_s00?ie=UTF8&amp;psc=1\">Practical Linear Algebra: A Geometry Toolbox <\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0120884003\/ref=oh_aui_detailpage_o05_s00?ie=UTF8&amp;psc=1\">Visualizing Quaternions<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0387241965\/ref=oh_aui_detailpage_o05_s01?ie=UTF8&amp;psc=1\">Curves and Surfaces for Computer Graphics<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0123742978\/ref=oh_aui_detailpage_o07_s00?ie=UTF8&amp;psc=1\">Essential Mathematics for Games and Interactive Applications: A Programmer's Guide<\/a><\/li>\n<\/ol><h3 id=\"books-to-develop-the-rendering-engine\">Books to develop the Rendering Engine<\/h3>\n<ol><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1849514763\/ref=oh_aui_detailpage_o05_s00?ie=UTF8&amp;psc=1\">OpenGL 4.0 Shading Language Cookbook<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1439893764\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">OpenGL Insights<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1568814240\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">Real-Time Rendering, Third Edition<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0321637631\/ref=oh_aui_detailpage_o08_s00?ie=UTF8&amp;psc=1\">OpenGL Shading Language<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0262017350\/ref=oh_aui_detailpage_o06_s00?ie=UTF8&amp;psc=1\">Foundations of 3D Computer Graphics<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0122287304\/ref=oh_aui_detailpage_o08_s00?ie=UTF8&amp;psc=1\">Texturing and Modeling, Second Edition: A Procedural Approach<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0672337479\/ref=oh_aui_detailpage_o05_s00?ie=UTF8&amp;psc=1\">OpenGL Superbible: Comprehensive Tutorial and Reference<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1568814348\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Graphics Shaders: Theory and Practice, Second Edition<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1584502991\/ref=oh_aui_detailpage_o03_s00?ie=UTF8&amp;psc=1\">Graphics Programming Methods<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/Real-Time-Shadows-Elmar-Eisemann\/dp\/1568814380\/ref=sr_1_1?ie=UTF8&amp;qid=1516083351&amp;sr=8-1&amp;keywords=real+time+shadows\">Real-Time Shadows<\/a><\/li>\n<\/ol><h3 id=\"books-to-develop-the-physics-engine\">Books to develop the Physics Engine<\/h3>\n<ol><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1449392512\/ref=oh_aui_detailpage_o04_s00?ie=UTF8&amp;psc=1\">Physics for Game Developers: Science, math, and code for realistic effects<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0486652416\/ref=oh_aui_detailpage_o07_s00?ie=UTF8&amp;psc=1\">Numerical Methods for Scientists and Engineers<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0521171903\/ref=oh_aui_detailpage_o08_s00?ie=UTF8&amp;psc=1\">A Student's Guide to Vectors and Tensors<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1558607323\/ref=oh_aui_detailpage_o09_s00?ie=UTF8&amp;psc=1\">Real-Time Collision Detection<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0123819768\/ref=oh_aui_detailpage_o00_s00?ie=UTF8&amp;psc=1\">Game Physics Engine Development: How to Build a Robust Commercial-Grade Physics Engine for your Game<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/3540779736\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">Computational Geometry: Algorithms and Applications<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1568814747\/ref=oh_aui_detailpage_o09_s00?ie=UTF8&amp;psc=1\">Game Physics Pearls<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1852338342\/ref=oh_aui_detailpage_o00_s00?ie=UTF8&amp;psc=1\">Geometry for Computer Graphics: Formulae, Examples and Proofs<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1558605940\/ref=oh_aui_detailpage_o03_s00?ie=UTF8&amp;psc=1\">Geometric Tools for Computer Graphics<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0521649765\/ref=oh_aui_detailpage_o09_s00?ie=UTF8&amp;psc=1\">Computational Geometry in C<\/a> <\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/155860801X\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Collision Detection in Interactive 3D Environments<\/a> <\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0123749034\/ref=oh_aui_detailpage_o03_s00?ie=UTF8&amp;psc=1\">Game Physics<\/a><\/li>\n<\/ol><h3 id=\"learning-3d-modeling-tools-blender-\">Learning 3D Modeling tools (Blender)<\/h3>\n<p>If you are wondering why you need to learn Blender, read the following <a href=\"https:\/\/www.haroldserrano.com\/blog\/game-character-modeling-for-game-engine-developers\">article<\/a>.<\/p>\n<ol><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1593274777\/ref=oh_aui_detailpage_o09_s00?ie=UTF8&amp;psc=1\">Blender Master Class: A Hands-On Guide to Modeling, Sculpting, Materials, and Rendering<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0240814304\/ref=oh_aui_detailpage_o00_s00?ie=UTF8&amp;psc=1\">Blender Foundations: The Essential Guide to Learning Blender 2.6<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0133886174\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Learning Blender: A Hands-On Guide to Creating 3D Animated Characters<\/a> <\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1849511268\/ref=oh_aui_detailpage_o06_s00?ie=UTF8&amp;psc=1\">Python 3 Object Oriented Programming<\/a>- Blender provides an API which allows you to extract 3D modeling information. To use the API, you need to know a bit of Python.<\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0321700899\/ref=oh_aui_detailpage_o09_s00?ie=UTF8&amp;psc=1\">Digital Modeling<\/a><\/li>\n<\/ol><h3 id=\"game-engine-architecture\">Game Engine Architecture<\/h3>\n<ol><li><a href=\"https:\/\/www.amazon.com\/Head-First-Design-Patterns-Brain-Friendly\/dp\/0596007124\/ref=sr_1_1?s=books&amp;ie=UTF8&amp;qid=1538145587&amp;sr=1-1&amp;keywords=head+first+design+patterns&amp;dpID=61ZG-hATOeL&amp;preST=_SX218_BO1,204,203,200_QL40_&amp;dpSrc=srch\">Head First Design Patterns<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1584500492\/ref=oh_aui_detailpage_o05_s00?ie=UTF8&amp;psc=1\">Game Programming Gems<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1584505710\/ref=oh_aui_detailpage_o06_s00?ie=UTF8&amp;psc=1\">Best of Game Programming Gems<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/012229064X\/ref=oh_aui_detailpage_o00_s00?ie=UTF8&amp;psc=1\">3D Game Engine Architecture<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0122290631\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">3D Game Engine Design: A Practical Approach<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0321321928\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">C++ Common Knowledge: Essential Intermediate Programming<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0471049980\/ref=oh_aui_detailpage_o01_s01?ie=UTF8&amp;psc=1\">C++ Pointers and Dynamic Memory Management<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0321334876\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Effective C++: 55 Specific Ways to Improve Your Programs and Designs<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0201734842\/ref=oh_aui_detailpage_o03_s00?ie=UTF8&amp;psc=1\">C++ Templates: The Complete Guide<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0201633612\/ref=oh_aui_detailpage_o04_s00?ie=UTF8&amp;psc=1\">Design Patterns: Elements of Reusable Object-Oriented Software<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0321623215\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">The C++ Standard Library: A Tutorial and Reference<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0139798099\/ref=oh_aui_detailpage_o03_s00?ie=UTF8&amp;psc=1\">Thinking in C++, Vol. 1: Introduction to Standard C++, 2nd Edition<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0130353132\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Thinking in C++, Volume 2: Practical Programming<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0321940156\/ref=oh_aui_detailpage_o08_s00?ie=UTF8&amp;psc=1\">Game Programming Algorithms and Techniques<\/a> <\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0123850037\/ref=oh_aui_detailpage_o01_s00?ie=UTF8&amp;psc=1\">API Design for C++<\/a> <\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0990582906\/ref=oh_aui_detailpage_o06_s00?ie=UTF8&amp;psc=1\">Game Programming Patterns<\/a> <\/li>\n<\/ol><h3 id=\"books-to-get-started\">Books to get started<\/h3>\n<p>You don't have to buy all these books at once. If you are starting out, then get the following books:<\/p>\n<p><strong>3D Math<\/strong><\/p>\n<ol><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1556229119\/ref=oh_aui_detailpage_o08_s00?ie=UTF8&amp;psc=1\">3D Math Primer For Graphics and Game Development<\/a><\/li>\n<\/ol><p><strong>Rendering Engine<\/strong><\/p>\n<ol><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0672337479\/ref=oh_aui_detailpage_o05_s00?ie=UTF8&amp;psc=1\">OpenGL Superbible: Comprehensive Tutorial and Reference<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1568814348\/ref=oh_aui_detailpage_o02_s00?ie=UTF8&amp;psc=1\">Graphics Shaders: Theory and Practice, Second Edition<\/a><\/li>\n<\/ol><p><strong>Physics Engine<\/strong><\/p>\n<ol><li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1449392512\/ref=oh_aui_detailpage_o04_s00?ie=UTF8&amp;psc=1\">Physics for Game Developers: Science, math, and code for realistic effects<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/0123819768\/ref=oh_aui_detailpage_o00_s00?ie=UTF8&amp;psc=1\">Game Physics Engine Development: How to Build a Robust Commercial-Grade Physics Engine for your Game<\/a><\/li>\n<li><a href=\"https:\/\/www.amazon.com\/gp\/product\/1558607323\/ref=oh_aui_detailpage_o09_s00?ie=UTF8&amp;psc=1\">Real-Time Collision Detection<\/a><\/li>\n<\/ol><p>Those books should get you started.<\/p>\n<p>Hope it helps.<\/p>","created_at":"2018-11-07T16:44:45+0000","updated_at":"2018-11-07T16:44:45+0000","published_at":"2018-01-15T00:00:00+0000","published_by":["Harold Serrano"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"www.haroldserrano.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":318,"title":"Reach for Markdown, not LaTeX","url":"https:\/\/blog.jez.io\/reach-for-markdown\/","content":"<p>I\u2019ve written <a href=\"https:\/\/blog.jez.io\/2014\/10\/06\/offline-latex-development\/\">in the past<\/a> (<a href=\"https:\/\/blog.jez.io\/2015\/01\/10\/offline-latex-development-part-2\/\">twice<\/a>) about how to\nstreamline the writing process when using LaTeX. Since then, I\u2019ve found that I\nenjoy writing even more when I don\u2019t have to reach for LaTeX at all. By reaching\nfirst for Markdown, then for LaTeX when necessary, writing is easier and more\nenjoyable.<\/p><h2>Writing at the Command Line<\/h2><p>Last year, I <a href=\"https:\/\/jez.io\/talks\/writing-at-the-command-line\/\">gave a talk<\/a> about the merits of writing primarily at\nthe command line. My main claims were that when writing we want:<\/p><ul><li>an open document format (so that our writings are future proof)<\/li>\n<li>to be using open source software (for considerations of privacy and cost)<\/li>\n<li>to optimize for the \u201ccommon case\u201d<\/li>\n<li>to be able to write for print and digital (PDFs, web pages, etc.)<\/li>\n<\/ul><p>Markdown solves these constraints nicely:<\/p><ul><li>It\u2019s a plain text format\u2014plain text has been around for decades and will\nbe for decades more.<\/li>\n<li>Given a plain text format, we can bring our own text editor.<\/li>\n<li>Plenty of open source programs manipulate Markdown.<\/li>\n<li>When we need advanced features, we can mix LaTeX into our Markdown documents.<\/li>\n<\/ul><p>For those unfamiliar with Markdown, it\u2019s super quick to pick up. If you only\nlook at one guide, see this one:<\/p><ul><li><a href=\"http:\/\/commonmark.org\/help\/\">CommonMark<\/a><\/li>\n<\/ul><p>If you want to start comparing features available in certain implementations of\nMarkdown:<\/p><ul><li><a href=\"https:\/\/guides.github.com\/features\/mastering-markdown\/\">GitHub Flavored Markdown<\/a><\/li>\n<li><a href=\"https:\/\/daringfireball.net\/projects\/markdown\/\">Markdown.pl<\/a><\/li>\n<li><a href=\"http:\/\/pandoc.org\/MANUAL.html#pandocs-markdown\">Pandoc Markdown<\/a><\/li>\n<\/ul><p>For more on why you should want to be writing at the command line, you can\n<a href=\"https:\/\/jez.io\/talks\/writing-at-the-command-line\/\">check out the talk slides<\/a>.<\/p><h2>Pandoc Starter<\/h2><p>The central tool I spoke about in <em>Writing at the Command Line<\/em> is <a href=\"https:\/\/pandoc.org\/\">Pandoc<\/a>.\nPandoc is an amazingly simple command line program that takes in Markdown files\nand spits out really anything you can think of.<\/p><p>To make using Pandoc even easier than it already is, I put together a\n<a href=\"https:\/\/github.com\/jez\/pandoc-starter\">collection of starter templates<\/a>. They\u2019re all available <a href=\"https:\/\/github.com\/jez\/pandoc-starter\">on\nGithub<\/a> if you\u2019d prefer to dive right in.<\/p><p>There are currently six different templates, specialized for the kind of\ndocument you\u2019d like to create. Each has a <code>README<\/code> for installation and usage\ninstructions, as well as a <code>Makefile<\/code> for invoking <code>pandoc<\/code> correctly.<\/p><p>All the templates generate PDFs from Markdown by way of LaTeX. In addition to\nPandoc, you\u2019ll also need LaTeX installed locally.<\/p><h3><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/tree\/master\/article\"><code>article<\/code><\/a><\/h3><p>This template uses the standard LaTeX <code>article<\/code> document class. It\u2019s a\nno frills, no nonsense choice.<\/p><p><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/blob\/master\/article\/src\/sample.pdf\"><img src=\"https:\/\/blog.jez.io\/images\/pandoc-starter-article.png\" alt=\"article template\" \/><\/a><\/p><h3><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/tree\/master\/tufte-handout\"><code>tufte-handout<\/code><\/a><\/h3><p>As an alternative to the <code>article<\/code> document class, there\u2019s also the\n<code>tufte-handout<\/code> document class. It originates from the style Edward Tufte\npopularized in his books and articles on visualization and design.<\/p><p>Apart from a different font (it uses Palatino instead of the default Computer\nModern), this template features the ability add side notes to your documents. I\noften find myself reaching for this template when I want to disguise the fact\nthat I\u2019m secretly using LaTeX.<\/p><p><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/blob\/master\/tufte-handout\/src\/sample.pdf\"><img src=\"https:\/\/blog.jez.io\/images\/pandoc-starter-tufte-handout.png\" alt=\"tufte-handout template\" \/><\/a><\/p><h3><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/tree\/master\/homework\"><code>homework<\/code><\/a><\/h3><p>A second alternative to the <code>article<\/code> document class is the <code>homework<\/code> document\nclass. It works nicely for homework assignments and problem sets. The class\nitself has a number of handy features, like:<\/p><ul><li>the option to put your name on every page, or only on the first page<\/li>\n<li>an option to use wide or narrow margins<\/li>\n<li>most of the AMS Math packages you\u2019d include in the process of typesetting a\nmath assignment<\/li>\n<li>a convenient environment for typesetting induction proofs<\/li>\n<\/ul><p>For more features and usage information, check out <a href=\"https:\/\/blog.jez.io\/2015\/01\/10\/the-latex-homework-document-class\/\">this blog post<\/a> or\n<a href=\"https:\/\/github.com\/jez\/latex-homework-class\">the source<\/a> on GitHub.<\/p><p><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/blob\/master\/homework\/src\/sample.pdf\"><img src=\"https:\/\/blog.jez.io\/images\/pandoc-starter-homework.png\" alt=\"homework template\" \/><\/a><\/p><h3><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/tree\/master\/beamer\"><code>beamer<\/code><\/a><\/h3><p>LaTeX provides the <code>beamer<\/code> document class for creating slides; this template\nmakes it even easier to use:<\/p><ul><li>Make a new slide with a \u201c<code>##<\/code>\u201d header<\/li>\n<li>Make a section divider with a \u201c<code>#<\/code>\u201d header<\/li>\n<li>Mix lists, links, code, and other Markdown features you\u2019re familiar with to\ncreate the content for a slide.<\/li>\n<\/ul><p>So basically, just write the outline for your talk, and Pandoc takes care of\nmaking the slides\u2014it doesn\u2019t get much simpler.<\/p><p><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/blob\/master\/beamer\/src\/sample.pdf\"><img src=\"https:\/\/blog.jez.io\/images\/pandoc-starter-beamer.png\" alt=\"beamer template\" \/><\/a><\/p><h3><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/tree\/master\/beamer-solarized\"><code>beamer-solarized<\/code><\/a><\/h3><p>The default beamer styles are pretty boring. To add a bit of flair and\npersonality to my slide decks, I made a Solarized theme for beamer.<\/p><p>In addition to the screenshot below, the <a href=\"https:\/\/jez.io\/talks\/writing-at-the-command-line\/\">Writing at the Command\nLine<\/a> slides I linked to earlier also use this theme, if you want\nto see a less contrived example.<\/p><p><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/blob\/master\/beamer-solarized\/src\/sample.pdf\"><img src=\"https:\/\/blog.jez.io\/images\/pandoc-starter-beamer-solarized.png\" alt=\"beamer solarized template\" \/><\/a><\/p><h3><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/tree\/master\/book-writeup\"><code>book-writeup<\/code><\/a><\/h3><p>Finally, sometimes a simple article or slide deck doesn\u2019t cut it. Usually this\nmeans I\u2019d like to group the writing into chapters. This template makes writing a\nchapter as easy as using a \u201c<code>#<\/code>\u201d Markdown header.<\/p><p><a href=\"https:\/\/github.com\/jez\/pandoc-starter\/blob\/master\/book-writeup\/src\/sample.pdf\"><img src=\"https:\/\/blog.jez.io\/images\/pandoc-starter-book-writeup.png\" alt=\"book writeup template\" \/><\/a><\/p><h2>Writing Plugins for Vim<\/h2><p>If you happen to use Vim, I\u2019d highly recommend installing <a href=\"https:\/\/github.com\/junegunn\/goyo.vim\">goyo.vim<\/a> for\nwriting. It removes all the visual frills Vim includes to make writing code\neasier so you can focus on your writing without distractions.<\/p><p>I also really enjoy <a href=\"https:\/\/github.com\/vim-pandoc\/vim-pandoc\">vim-pandoc<\/a> and <a href=\"https:\/\/github.com\/vim-pandoc\/vim-pandoc-syntax\">vim-pandoc-syntax<\/a>. They\u2019re a pair of\ncomplementary plugins for highlighting and working with Pandoc Markdown-flavored\ndocuments. They work so well that I use them for Markdown documents even when\nnot using Pandoc.<\/p><h2>Reach for Markdown<\/h2><p>Writing should be a pleasant experience. With the right tools, it can be. LaTeX\nis powerful but cumbersome to use. With Markdown, we can focus on our writing,\nand worry about the presentation later. Pandoc can take care of the presentation\nfor us, so the <a href=\"https:\/\/github.com\/jez\/pandoc-starter\">only thing left to do is start<\/a>.<\/p><footer class=\"entry-meta\"><time datetime=\"2017-02-26T21:26:53-05:00\">February 26, 2017<\/time><\/footer>","created_at":"2018-11-07T16:44:45+0000","updated_at":"2018-11-07T16:44:45+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"blog.jez.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":313,"title":"Lecture 1a: Overview and Introduction to Lisp","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=XYKRVNQ_MqE","content":"<iframe width=\"459\" height=\"344\" src=\"https:\/\/www.youtube.com\/embed\/XYKRVNQ_MqE?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:44:08+0000","updated_at":"2018-11-07T16:44:08+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/XYKRVNQ_MqE\/hqdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":314,"title":"Getting Acquainted With Clojure Zippers","url":"http:\/\/josf.info\/blog\/2014\/03\/21\/getting-acquainted-with-clojure-zippers\/","content":"<div class=\"paragraph\">\n<p>Zippers, in Clojure, are one of those things you end up running into\nthat can be rather confusing. A little googling of the topic seems to\nreveal a universe divided between people who understand zippers\nperfectly and those who don\u2019t get them at all. Zippers are a core\nlibrary, so they get mentioned in the Clojure books, but they are also\na special case that would take several pages to really explain in\ndetail.<\/p>\n<\/div><div class=\"paragraph\">\n<p>Most of all, they are, for a lot of us, slightly strange, a solution\nto the problem of \u201ctraversing\u201d a tree structure without any implied\nstate. In this post, I\u2019m going to try to help those who don\u2019t\nimmediately understand zippers to at least avoid some of the\nsubtleties that tripped me up.<\/p>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"why-zipper\">Why \u201czipper\u201d?<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p><code>I can\u2019t use zip files because I don\u2019t have a Zip drive''. Why does\neverything have to be called <\/code>zip\u201d? What is a zipper, anyway?<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>With their usual terseness, the docs simply say:<\/p>\n<\/div>\n<div class=\"quoteblock\">\n<blockquote><div>\nFunctional hierarchical zipper, with navigation, editing,\nand enumeration.  See Huet\n<\/div><\/blockquote>\n<div class=\"attribution\">\n\u2014 Rich Hickey<br \/><cite><a href=\"http:\/\/richhickey.github.io\/clojure\/clojure.zip-api.html\">API for clojure.zip<\/a><\/cite>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>Huet turns out to be G\u00e9rard Huet, who published\n<a href=\"http:\/\/www.st.cs.uni-saarland.de\/edu\/seminare\/2005\/advanced-fp\/docs\/huet-zipper.pdf\">Functional\npearl: the Zipper<\/a> in 1997. Huet insists on how simple the concept is,\nand almost asks forgiveness for writing a paper about something that\nreally should have been documented long ago:<\/p>\n<\/div>\n<div class=\"quoteblock\">\n<blockquote><div>\nThis simple idea must have been invented on numerous occasions by\ncreative programmers, and the only justification for presenting what\nought to be folklore is that it does not appear to have been\npublished, or even well-known.\n<\/div><\/blockquote>\n<div class=\"attribution\">\n\u2014 G\u00e9rard Huet<br \/><cite>Functional Pearl: the Zipper<\/cite>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>Huet is unsurprisingly one of those people who just \u201cgets it\u201d. Here is\nhow he explains the Zipper:<\/p>\n<\/div>\n<div class=\"quoteblock\">\n<blockquote>\n<div class=\"paragraph\">\n<p>The basic idea is simple: the tree is turned inside-out like a\nreturned glove, pointers from the root to the current position being\nreversed in a <em>path structure.<\/em> The current <em>location<\/em> holds both the\ndownward current subtree and the upward path.<\/p>\n<\/div>\n<\/blockquote>\n<\/div>\n<div class=\"paragraph\">\n<p>Now maybe this kind of spatial and logical torsion is intuitive\nfor you. Reading this now, I actually do understand what he means, and\nit might be <code>`simple'' in a Rich Hickey sense, as in\n<\/code>`uncomplected\u201d,\nbut it certainly isn\u2019t that easy to understand.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"a-few-basic-things\">A few basic things<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p>So here are a few things I would have liked to have understood to\nbegin with, and that I didn\u2019t fully grasp before going through lots of\ntrial and error and frustration.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"zipper-vs-tree-vs-node\">Zipper vs. tree vs. node<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p>We know what a tree is. In Clojure we would represent it as something\nlike this, with nested vectors for example:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>[1 [:a :b] 2 3 [40 50 60]]<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>This is just a tree, not a zipper, because a zipper is not a tree: it\nis a <em>tree plus your position in the tree.<\/em> Or to be more precise: <em>a\ntree seen from a particular location in the tree.<\/em><\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>So to make a zipper out of that tree, we do this:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>user&gt; (require '(clojure [zip :as zip]))\nnil\nuser&gt; (zip\/vector-zip [1 [:a :b] 2 3 [40 50 60]])\n[[1 [:a :b] 2 3 [40 50 60]] nil]<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>Notice how the result is just the same tree wrapped in a two-element\nvector? No big deal, right? Can it be that simple?<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>Well\u2026\u200b this is Huet\u2019s glove, but still right side out. If you want to\nsee the lining, just do this:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>user&gt; (-&gt;  [1 [:a :b] 2 3 [40 50 60]]\n\t   zip\/vector-zip\n\t   zip\/down\n\t   zip\/right)\n[[:a :b] {:l [1], :pnodes [[1 [:a :b] 2 3 [40 50 60]]], :ppath nil, :r (2 3 [40 50 60])}]<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>There\u2019s the lining, or at least the beginning of it. All that extra\nstuff is the zipper looking back up toward the root of the tree, to\nthe left and to the right. Moving around the tree means calling a\nfunction like zip\/down on the current zipper, and getting back a\n<em>new<\/em> zipper with some of the nodes shifted around to represent your\nnew position in the tree. This is functional programming. We are\nalways creating new values, not modifying old values. Position in a\ntree is typically a form of implicit state: we just <em>are<\/em>\nsomewhere. Here our location becomes explicit, and the zipper\ninfrastructure makes it so we don\u2019t have to think too much about how\nall of this happens. We don\u2019t ever need to know about :pnodes or\n:ppath, we just deal with the zipper interface. (Imagine having to\ndo this \u201cby hand\u201d, associng the left nodes over to the right\nnodes, dissocing them from the left, etc.)<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"two-forms-of-movement\">Two forms of movement<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p>While we are talking about movement, it\u2019s worth pointing out that\nthere are two ways of moving through the zipper. The first way is what\nwe\u2019ve just seen: zip\/down, zip\/up, zip\/left, zip\/right and\nfunctions like zip\/root that take us all the way back to\nthe top.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>The second way of getting around the tree, zipper-style, is to walk,\nas in tree-walk. One of the added benefits of using zippers is that\nyou get, for free, a depth-first tree traversal, just by calling\nzip\/next as many times as you need.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>Going back to our example tree, with zip\/next you can do stuff like\nthis:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>;; let's make a var to be easier to read.\n user&gt; (def zzz (zip\/vector-zip [1 [:a :b] 2 3 [40 50 60]]))\n#'user\/zzz\nuser&gt; (-&gt; zzz zip\/next zip\/node)\n1\nuser&gt; (-&gt; zzz zip\/next zip\/next zip\/node)\n[:a :b]\nuser&gt; (-&gt; zzz zip\/next zip\/next zip\/next zip\/node)\n:a\nuser&gt; (-&gt; zzz zip\/next zip\/next zip\/next zip\/next zip\/node)\n:b\nuser&gt; (-&gt; zzz zip\/next zip\/next zip\/next zip\/next zip\/next zip\/node)<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>(Notice that we call zip\/node at the end each time, so that we get\nthe value and not the whole zipper. More on that in a second.)<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>Each time we add a call to zip\/next, we get one more node. Without\neven trying very hard, we get a complete walk of the tree just by\ncalling zip\/next a bunch of times. The function just \u201cknows\u201d where\nto go to hit each node in the tree exactly once.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"end-games\">End games<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p>Now, when walking a zipper, like in any recursive situation, we\nwouldn\u2019t want to be in an infinite loop. zip\/next needs to know when\nto quit. Which is why the Clojure zip library provides an end?\nfunction that tells us if we are there yet.<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>user&gt; (defn fast-forward [loc]\n\t(if (zip\/end? loc)\n\t  (zip\/node loc)\n\t  (recur (zip\/next loc))))\n#'user\/fast-forward\nuser&gt; (fast-forward zzz)\n[1 [:a :b] 2 3 [40 50 60]]<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>Here we\u2019ve zipped all the way to the end, and we end up with the whole\ntree, that is the root node.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>Check out this slighly different version of our fast-forward function:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>user&gt; (defn fast-forward-z [loc]\n\t(if (zip\/end? loc)\n\t  loc\n\t  (recur (zip\/next loc))))\n#'user\/fast-forward-z\nuser&gt; (fast-forward-z zzz)\n[[1 [:a :b] 2 3 [40 50 60]] :end]<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>This time we get the zipper, and not the just the node, as in the\nprevious. And we have this strange little :end keyword as the second\npart of the zipper vector. The zipper has been \u201cconsumed\u201d. We can\u2019t\nwalk anymore. zip\/next will just keep returning the same thing, and\nzip\/prev will just return nil. The glove has been turned so far\ninside out that it can\u2019t go any more (yet, strangely, it looks right\nside out again).<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>So not only does the zipper represent our position in the tree, it\nalso \u201cknows\u201d where we\u2019ve been, or at least whether we can keep\nwalking or not.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"nodes-and-locs\">Nodes and locs<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p>So what about nodes and this zip\/node function? Really, it is\nactually fairly simple: a node is of course a tree, a subtree, a\nbranch or a leaf. The node is your data. But it doesn\u2019t know what it\u2019s\nparents or siblings are, or how far it is from the root.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>On the other hand, a <em>location<\/em> (usually pronounced loc) contains a\nnode (the node is the first element in the zipper vector), but a\nlocation is actually the entire zipper, zipped to a certain\npoint. Sometimes you need the node, and sometimes you need the <em>loc<\/em>.\nSomehow, figuring out when to look at a node and when to look at a\n<em>location<\/em> was one of the major conceptual hurdles that I encountered.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>What tripped me up, I think, looking back, was that zippers require\nyou to go back and forth between the zipper view of the world (that\nis: the whole tree seen from a particular position) and a node view of\nthe world, where we are back in a good old tree structure looking at\nindividual values.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>Say we wanted to find the first keyword (in preorder depth-first\ntraversal) in our tired zzz vector. We could do this:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>user&gt; (defn first-keyword [loc]\n\t(if (zip\/end? loc)\n\t  nil\n\t  (if (keyword? (zip\/node loc))\n\t    (zip\/node loc)\n\t    (recur (zip\/next loc)))))\n#'user\/first-keyword\nuser&gt; (first-keyword zzz)\n:a<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>The \u201ckeywordiness\u201d of a particular location in the tree is distinct\nfrom its location. To see if the current node is a keyword, we have to\nuse the zip\/node function to slip into nodespace. If, on the other\nhand, we want to know whether our node has a right or left sibling,\nfor example, we need to be in zipper space. If we want to find nodes\nof a certain kind that have parents or children of another kind, our\ntests have to dance between nodes and zippers. This is the kind of\nthing that seems completely obvious once you understand it, but, at\nleast for me, kept me scratching my head early on.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"sect1\">\n<h2 id=\"zippers-are-cheap\">Zippers are cheap<\/h2>\n<div class=\"sectionbody\">\n<div class=\"paragraph\">\n<p>One last thing that took a while to sink in: zippers are cheap. Huet\ninsists on how zippers are supposed to be fast since, in a very\nclojurey way, only the differences in path and position need to be\ntaken into account, leaving most of the original data intact.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>In a more concrete way, just looking at a zipper, we see that there\nisn\u2019t much to them. Here is your data:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>user&gt; (def your-data [:a :b :c])\n#'user\/your-data<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>Here is your data in a zipper:<\/p>\n<\/div>\n<div class=\"listingblock\">\n<div class=\"content\">\n<pre>[[:a :b :c] nil]<\/pre>\n<\/div>\n<\/div>\n<div class=\"paragraph\">\n<p>It\u2019s just a two-item vector. Well, almost: there is some metadata\nattached, so it doesn\u2019t work to just stuff some data in a raw vector\nfollowed by nil. You have to use functions like vector-zip or\nxml-zip to create them.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>But still, the data structure is very lightweight, especially when the\nzipper is at the root position of the tree. As a result, you can be\ncreative with when you decide to break out a zipper. Suppose for\ninstance that you know you only want to walk one subtree within a\nlarger tree. You need a zipper to navigate down to the top of the\nsubtree that interests you. Once you\u2019ve found it, rather than figuring\nout how to constrain your search to just that subtree, it is simpler\nto just take the node containing the subtree and plunk it into a new\nzipper. Then you can zip\/next to your heart\u2019s content, knowing that\nyou\u2019ll never end up in some weird place you didn\u2019t expect.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>I\u2019ll talk about \u201cediting\u201d zippers some other time, but I recently\ncame onto a problem where I wanted to walk a tree several times with\ndifferent functions each time. I couldn\u2019t just hold onto a reference\nto the original zipper because I was modifying as I went. I need to be\nable to chain my zipper-walking functions together in a sequence. The\nproblem was that after each walk, I was at the :end. My functions\nworked individually but couldn\u2019t be chained together. The original\nzipper had been walked too far and just couldn\u2019t take it anymore. What\nto do? <a href=\"https:\/\/github.com\/hiredman\">hiredman<\/a> helped me out on IRC with\nthis: just make a new zipper. In my case that meant just calling\nzip\/xml-zip again on the exhausted zipper.<\/p>\n<\/div>\n<div class=\"paragraph\">\n<p>There are some more gotchas that I\u2019ve run into with zippers,\nespecially when trying to edit trees, so I\u2019ll take that on later. My\nfeeling about zippers is that they are an extremely powerful tool that\ncan be used in lots of unusual situations, not just parsing XML, but\nanywhere you have a tree that starts to get a little too complicated.<\/p>\n<\/div>\n<\/div>\n<\/div>","created_at":"2018-11-07T16:44:08+0000","updated_at":"2018-11-07T16:44:08+0000","annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"josf.info","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":315,"title":"How to Name Clojure Functions","url":"https:\/\/stuartsierra.com\/2016\/01\/09\/how-to-name-clojure-functions","content":"<link href=\"\/css\/org-clojure-1.css\" type=\"text\/css\" rel=\"stylesheet\" \/><p> This is a guide on naming Clojure functions. There are exceptions to every rule. When you\u2019re defining something based on natural language, there are more exceptions than rules. I break these rules more often than I follow them. This guide is just a starting point for <b>thinking<\/b> about how to name things. <\/p>\n<div id=\"outline-container-orgheadline1\" class=\"outline-2\">\n<h2 id=\"orgheadline1\">Pure functions<\/h2>\n<div class=\"outline-text-2\" id=\"text-orgheadline1\">\n<p> Pure functions which return values are named with <b>nouns<\/b> describing the value they return. <\/p>\n<p> If I have a function to compute a user\u2019s age based on their birthdate, it is called <code>age<\/code>, not <code>calculate-age<\/code> or <code>get-age<\/code>. <\/p>\n<p> Think of the definition: a pure function is one which can be <b>replaced with its value<\/b> without affecting the result. So why not make that evident in the name? <\/p>\n<p> This is particularly good for constructors and accessors. No need to clutter up your function names with meaningless prefixes like <code>get-<\/code> and <code>make-<\/code>. <\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline2\" class=\"outline-2\">\n<h2 id=\"orgheadline2\">Don\u2019t repeat the name of the namespace<\/h2>\n<div class=\"outline-text-2\" id=\"text-orgheadline2\">\n<p> Function names should not repeat the name of the namespace. <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-clojure\">(ns products)&#13;\n&#13;\n;; Bad, redundant:&#13;\n(defn product-price [product]&#13;\n  ;; ...&#13;\n  )&#13;\n&#13;\n;; Good:&#13;\n(defn price [product]&#13;\n  ;; ...&#13;\n  )&#13;\n<\/pre>\n<\/div>\n<p> Assume that consumers of a function will use it with a <a href=\"http:\/\/stuartsierra.com\/2015\/05\/10\/clojure-namespace-aliases\">namespace alias<\/a>. <\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline3\" class=\"outline-2\">\n<h2 id=\"orgheadline3\">Conversions and coercions<\/h2>\n<div class=\"outline-text-2\" id=\"text-orgheadline3\">\n<p> I don\u2019t much like <code>-&gt;<\/code> arrows in function names, and I try to avoid them. <\/p>\n<p> If the function is a <b>coercion<\/b>, that is, it is meant to convert any of several input types into the desired output type, then name it for the output type. For example, in <code>clojure.java.io<\/code> the functions <code>file<\/code>, <code>reader<\/code>, and <code>writer<\/code> are all coercions. <\/p>\n<p> If there are different functions for different input types, then each one is a <b>conversion<\/b>. In that case, use <code>input-type-&gt;output-type<\/code> names. <\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline4\" class=\"outline-2\">\n<h2 id=\"orgheadline4\">Functions with side effects<\/h2>\n<div class=\"outline-text-2\" id=\"text-orgheadline4\">\n<p> Functions which have side-effects are named with <b>verbs<\/b> describing what they do. <\/p>\n<p> Constructor functions with side-effects, such as adding a record to a database, have names starting with <code>create-<\/code>. (I borrowed this idea from Stuart Halloway\u2019s Datomic tutorials.) <\/p>\n<p> Functions which perform side-effects to retrieve some information (e.g. query a web service) have names starting with <code>get-<\/code>. <\/p>\n<p> For words which could be either nouns or verbs, assume noun by default then add words to make verb phrases. E.g. <code>message<\/code> constructs a new object representing a message, <code>send-message<\/code> transmits it. <\/p>\n<p> I don\u2019t use the exclamation-mark convention (e.g. <code>swap!<\/code>) much. Different people use it to mean different things (side effect, state change, transaction-unsafe) so the meaning is vague at best. If I do use an exclamation mark, it\u2019s to signal a change to a mutable reference, not other side-effects such as I\/O. <\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline5\" class=\"outline-2\">\n<h2 id=\"orgheadline5\">Local name clashes<\/h2>\n<div class=\"outline-text-2\" id=\"text-orgheadline5\">\n<p> One problem I find is in <code>let<\/code> blocks, when the obvious name for a local is the same as the function which computes it. If you\u2019re not careful, this can lead to clashes: <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-clojure\">(defn shipping-label&#13;\n  \"Returns a new label to ship product to customer.\"&#13;\n  [customer product]&#13;\n  (let [address (address customer)&#13;\n        weight (weight product)&#13;\n        supplier (supplier product)]&#13;\n    {:from (address supplier)  ; oops, 'address' clashes!&#13;\n     :to address&#13;\n     :weight weight}))&#13;\n<\/pre>\n<\/div>\n<p> This is less of a problem when the functions are defined in a different namespace and referenced via an alias: <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-clojure\">(defn shipping-label&#13;\n  \"Returns a new label to ship product to customer.\"&#13;\n  [customer product]&#13;\n  (let [address (mailing\/address customer)&#13;\n        weight (product\/weight product)&#13;\n        supplier (product\/supplier product)]&#13;\n    {:from (mailing\/address supplier)  ; OK&#13;\n     :to address&#13;\n     :weight weight}))&#13;\n<\/pre>\n<\/div>\n<p> If name-clashes become a problem, add prefixes to the function names, <code>new-<\/code> for constructors and <code>get-<\/code> for accessors. If you are bothered that this contradicts the previous section, re-read the first paragraph of this article. <\/p>\n<\/div>\n<\/div>\n<div id=\"outline-container-orgheadline6\" class=\"outline-2\">\n<h2 id=\"orgheadline6\">Function returning functions<\/h2>\n<div class=\"outline-text-2\" id=\"text-orgheadline6\">\n<p> In general, I try to avoid defining top-level functions which return functions <b>if<\/b> I can write make the intent clearer using anonymous functions instead. <\/p>\n<p> For example, writing something like this makes me feel clever: <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-clojure\">(defn foo&#13;\n  \"Returns a function to compute foo of value.\"&#13;\n  [option]&#13;\n  (fn [value]&#13;\n    ;;... do stuff with value ...&#13;\n    ))&#13;\n&#13;\n(defn computation&#13;\n  \"Does stuff with values.\"&#13;\n  [option values]&#13;\n  (-&gt;&gt; values&#13;\n       (map (foo option))  ; look at me!&#13;\n       ;; ...&#13;\n       ))&#13;\n<\/pre>\n<\/div>\n<p> But it\u2019s easier for someone else to read when the closure is created close to where it\u2019s used: <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-clojure\">(defn foo&#13;\n  \"Returns the foo of value\"&#13;\n  [value option]&#13;\n  ;; ... &#13;\n  )&#13;\n&#13;\n(defn computation [option values]&#13;\n  (-&gt;&gt; values&#13;\n       (map #(foo % option))  ; I see what this does&#13;\n       ;; ...&#13;\n       ))&#13;\n<\/pre>\n<\/div>\n<p> I allow an exception to this rule when returning functions is part of a repeated pattern. For example, the transducer versions of <code>map<\/code>, <code>filter<\/code>, and other sequence functions all return functions, but that\u2019s a standard part of the language since Clojure 1.7 so users can be expected to know about it. Occasionally I discover a similar pattern in my own code. <\/p>\n<p> When functions returning functions are <b>not<\/b> part of a repeated pattern but for some reason I want them anyway, I call them out with a suffix <code>-fn<\/code>, like: <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-clojure\">(defn foo-fn&#13;\n  \"Returns a function to compute the foo of a value.\"&#13;\n  [param]&#13;\n  (fn [value]&#13;\n    ;; ...&#13;\n    ))&#13;\n<\/pre>\n<\/div>\n<\/div>\n<\/div>","created_at":"2018-11-07T16:44:08+0000","updated_at":"2018-11-07T16:44:08+0000","published_at":"2016-01-09T19:36:00+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"stuartsierra.com","preview_picture":"https:\/\/s0.wp.com\/i\/blank.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":316,"title":"What Made Lisp Different","url":"http:\/\/www.paulgraham.com\/diff.html","content":"<img src=\"http:\/\/ep.yimg.com\/ca\/I\/paulgraham_2202_11662025\" width=\"208\" height=\"18\" border=\"0\" alt=\"What Made Lisp Different\" style=\"margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><p><span>December 2001 (rev. May 2002)\n\n<i>(This article came about in response to some questions on\nthe <a href=\"http:\/\/ll1.mit.edu\">LL1<\/a> mailing list.  It is now\nincorporated in <a href=\"http:\/\/www.paulgraham.com\/icad.html\">Revenge of the Nerds<\/a>.)<\/i><\/span><\/p><p>When McCarthy designed Lisp in the late 1950s, it was\na radical departure from existing languages,\nthe most important of which was <a href=\"http:\/\/www.paulgraham.com\/history.html\">Fortran<\/a>.<\/p><p>Lisp embodied nine new ideas:<\/p><hr \/><b>1. Conditionals.<\/b>  A conditional is an if-then-else\nconstruct.  We take these for granted now.  They were \n<a href=\"http:\/\/www-formal.stanford.edu\/jmc\/history\/lisp\/node2.html\">invented<\/a>\nby McCarthy in the course of developing Lisp. \n(Fortran at that time only had a conditional\ngoto, closely based on the branch instruction in the \nunderlying hardware.)  McCarthy, who was on the Algol committee, got\nconditionals into Algol, whence they spread to most other\nlanguages.<p><b>2. A function type.<\/b> In Lisp, functions are first class \nobjects-- they're a data type just like integers, strings,\netc, and have a literal representation, can be stored in variables,\ncan be passed as arguments, and so on.<\/p><p><b>3. Recursion.<\/b>  Recursion existed as a mathematical concept\nbefore Lisp of course, but Lisp was the first programming language to support\nit.  (It's arguably implicit in making functions first class\nobjects.)<\/p><p><b>4. A new concept of variables.<\/b>  In Lisp, all variables\nare effectively pointers. Values are what\nhave types, not variables, and assigning or binding\nvariables means copying pointers, not what they point to.<\/p><p><b>5. Garbage-collection.<\/b><\/p><p><b>6. Programs composed of expressions.<\/b> Lisp programs are \ntrees of expressions, each of which returns a value.  \n(In some Lisps expressions\ncan return multiple values.)  This is in contrast to Fortran\nand most succeeding languages, which distinguish between\nexpressions and statements.<\/p><p>It was natural to have this\ndistinction in Fortran because (not surprisingly in a language\nwhere the input format was punched cards) the language was\nline-oriented.  You could not nest statements.  And\nso while you needed expressions for math to work, there was\nno point in making anything else return a value, because\nthere could not be anything waiting for it.<\/p><p>This limitation\nwent away with the arrival of block-structured languages,\nbut by then it was too late. The distinction between\nexpressions and statements was entrenched.  It spread from \nFortran into Algol and thence to both their descendants.<\/p><p>When a language is made entirely of expressions, you can\ncompose expressions however you want.  You can say either\n(using <a href=\"http:\/\/www.paulgraham.com\/arc.html\">Arc<\/a> syntax)<\/p><p>(if foo (= x 1) (= x 2))<\/p><p>or<\/p><p>(= x (if foo 1 2))<\/p><p><b>7. A symbol type.<\/b>  Symbols differ from strings in that\nyou can test equality by comparing a pointer.<\/p><p><b>8. A notation for code<\/b> using trees of symbols.<\/p><p><b>9. The whole language always available.<\/b>  \nThere is\nno real distinction between read-time, compile-time, and runtime.\nYou can compile or run code while reading, read or run code\nwhile compiling, and read or compile code at runtime.<\/p><p>Running code at read-time lets users reprogram Lisp's syntax;\nrunning code at compile-time is the basis of macros; compiling\nat runtime is the basis of Lisp's use as an extension\nlanguage in programs like Emacs; and reading at runtime\nenables programs to communicate using s-expressions, an\nidea recently reinvented as XML.\n<\/p><hr \/><p>When Lisp was first invented, all these ideas were far\nremoved from ordinary programming practice, which was\ndictated largely by the hardware available in the late 1950s.<\/p><p>Over time, the default language, embodied\nin a succession of popular languages, has\ngradually evolved toward Lisp.  1-5 are now widespread.\n6 is starting to appear in the mainstream.\nPython has a form of 7, though there doesn't seem to be\nany syntax for it.  \n8, which (with 9) is what makes Lisp macros\npossible, is so far still unique to Lisp,\nperhaps because (a) it requires those parens, or something \njust as bad, and (b) if you add that final increment of power, \nyou can no \nlonger claim to have invented a new language, but only\nto have designed a new dialect of Lisp ; -)<\/p><p>Though useful to present-day programmers, it's\nstrange to describe Lisp in terms of its\nvariation from the random expedients other languages\nadopted.  That was not, probably, how McCarthy\nthought of it.  Lisp wasn't designed to fix the mistakes\nin Fortran; it came about more as the byproduct of an\nattempt to <a href=\"http:\/\/www.paulgraham.com\/rootsoflisp.html\">axiomatize computation<\/a>.<\/p>","created_at":"2018-11-07T16:44:08+0000","updated_at":"2018-11-07T16:44:08+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"www.paulgraham.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":311,"title":"Common combinators in JavaScript","url":"https:\/\/gist.github.com\/Avaq\/1f0636ec5c8d6aed2e45","content":"<div id=\"readme\" class=\"readme blob instapaper_body\">\n    <article class=\"markdown-body entry-content\" itemprop=\"text\"><table><thead><tr><th align=\"right\">Name<\/th>\n<th>#<\/th>\n<th><a href=\"https:\/\/www.haskell.org\/\" rel=\"nofollow\">Haskell<\/a><\/th>\n<th><a href=\"http:\/\/ramdajs.com\/\" rel=\"nofollow\">Ramda<\/a><\/th>\n<th><a href=\"http:\/\/sanctuary.js.org\/#combinator\" rel=\"nofollow\">Sanctuary<\/a><\/th>\n<th>Signature<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"right\">identity<\/td>\n<td><strong>I<\/strong><\/td>\n<td><code>id<\/code><\/td>\n<td><code>identity<\/code><\/td>\n<td><code>I<\/code><\/td>\n<td><code>a \u2192 a<\/code><\/td>\n<\/tr><tr><td align=\"right\">constant<\/td>\n<td><strong>K<\/strong><\/td>\n<td><code>const<\/code><\/td>\n<td><code>always<\/code><\/td>\n<td><code>K<\/code><\/td>\n<td><code>a \u2192 b \u2192 a<\/code><\/td>\n<\/tr><tr><td align=\"right\">apply\u00b9<\/td>\n<td><strong>A<\/strong><\/td>\n<td><code>($)<\/code><\/td>\n<td><code>call<\/code><\/td>\n\n<td><code>(a \u2192 b) \u2192 a \u2192 b<\/code><\/td>\n<\/tr><tr><td align=\"right\">thrush<\/td>\n<td><strong>T<\/strong><\/td>\n<td><code>(&amp;)<\/code><\/td>\n<td><code>applyTo<\/code><\/td>\n<td><code>T<\/code><\/td>\n<td><code>a \u2192 (a \u2192 b) \u2192 b<\/code><\/td>\n<\/tr><tr><td align=\"right\">duplication<\/td>\n<td><strong>W<\/strong><\/td>\n<td><code>join<\/code>\u00b2<\/td>\n<td><code>unnest<\/code>\u00b2<\/td>\n<td><code>join<\/code>\u00b2<\/td>\n<td><code>(a \u2192 a \u2192 b) \u2192 a \u2192 b<\/code><\/td>\n<\/tr><tr><td align=\"right\">flip<\/td>\n<td><strong>C<\/strong><\/td>\n<td><code>flip<\/code><\/td>\n<td><code>flip<\/code><\/td>\n<td><code>flip<\/code><\/td>\n<td><code>(a \u2192 b \u2192 c) \u2192 b \u2192 a \u2192 c<\/code><\/td>\n<\/tr><tr><td align=\"right\">compose<\/td>\n<td><strong>B<\/strong><\/td>\n<td><code>(.)<\/code>, <code>fmap<\/code>\u00b2<\/td>\n<td><code>map<\/code>\u00b2<\/td>\n<td><code>compose<\/code>, <code>map<\/code>\u00b2<\/td>\n<td><code>(b \u2192 c) \u2192 (a \u2192 b) \u2192 a \u2192 c<\/code><\/td>\n<\/tr><tr><td align=\"right\">substitution<\/td>\n<td><strong>S<\/strong><\/td>\n<td><code>ap<\/code>\u00b2<\/td>\n<td><code>ap<\/code>\u00b2<\/td>\n<td><code>ap<\/code>\u00b2<\/td>\n<td><code>(a \u2192 b \u2192 c) \u2192 (a \u2192 b) \u2192 a \u2192 c<\/code><\/td>\n<\/tr><tr><td align=\"right\">psi<\/td>\n<td><strong>P<\/strong><\/td>\n<td><code>on<\/code><\/td>\n\n<td><code>on<\/code><\/td>\n<td><code>(b \u2192 b \u2192 c) \u2192 (a \u2192 b) \u2192 a \u2192 a \u2192 c<\/code><\/td>\n<\/tr><tr><td align=\"right\">fix-point\u00b3<\/td>\n<td><strong>Y<\/strong><\/td>\n<td><code>fix<\/code><\/td>\n\n\n<td><code>(a \u2192 a) \u2192 a<\/code><\/td>\n<\/tr><\/tbody><\/table><hr \/><p>\u00b9) The A-combinator can be implemented as an alias of the I-combinator.\nIts implementation in Haskell exists because the infix nature gives it\nsome utility. Its implementation in Ramda exists because it is\noverloaded with additional functionality.<\/p>\n<p>\u00b2) Algebras like <code>ap<\/code> have different implementations for different types.\nThey work like Function combinators only for Function inputs.<\/p>\n<p>\u00b3) In JavaScript and other non-lazy languages, it is impossible to implement the\nY-combinator. Instead a variant known as the <em>applicative<\/em> or <em>strict<\/em>\nfix-point combinator is implemented. This variant is sometimes rererred to as\nthe Z-combinator.<\/p>\n<\/article><\/div><div class=\"readme blob instapaper_body\">\n    <article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Note that when I use the word \"combinator\" in this context, it implies \"function combinator in the untyped lambda calculus\".<\/p>\n<ul><li><a href=\"https:\/\/www.amazon.com\/Mock-Mockingbird-Other-Logic-Puzzles\/dp\/0192801422\" rel=\"nofollow\">To Mock a Mocking Bird<\/a> (book of combinatory puzzles and examples)<\/li>\n<li><a href=\"http:\/\/www.angelfire.com\/tx4\/cus\/combinator\/birds.html\" rel=\"nofollow\">Combinator Birds<\/a> (summary of all combinators named by the book)<\/li>\n<li><a href=\"https:\/\/github.com\/benji6\/combinators-js\">CombinatorsJS<\/a> (implementation of all the same combinators in JavaScripts)<\/li>\n<li><a href=\"https:\/\/github.com\/fantasyland\/fantasy-combinators\">Fantasy Combinators<\/a> (implementation of common combinators in JavaScript)<\/li>\n<li><a href=\"http:\/\/jwodder.freeshell.org\/lambda.html\" rel=\"nofollow\">Collected Lambdas<\/a> (collection of noteworthy combinators)<\/li>\n<li><a href=\"http:\/\/rubymanor.org\/3\/videos\/programming_with_nothing\" rel=\"nofollow\">Programming With Nothing<\/a> (talk)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Church_encoding\" rel=\"nofollow\">Church Encoding<\/a> (wikipedia)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Fixed-point_combinator\" rel=\"nofollow\">Fixed-point combinator<\/a> (wikipedia)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Lambda_calculus\" rel=\"nofollow\">Lambda calculus<\/a> (wikipedia)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/SKI_combinator_calculus\" rel=\"nofollow\">SKI combinator calculus<\/a> (wikipedia)<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/B,_C,_K,_W_system\" rel=\"nofollow\">BCKW combinator calculus<\/a> (wikipedia)<\/li>\n<\/ul><\/article><\/div>","created_at":"2018-11-07T16:44:07+0000","updated_at":"2018-11-07T16:44:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"gist.github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/1217745?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":312,"title":"Lecture 1A | MIT 6.001 Structure and Interpretation, 1986","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=2Op3QLzMgSY&list=PLE18841CABEA24090","content":"<iframe id=\"video\" width=\"459\" height=\"344\" src=\"https:\/\/www.youtube.com\/embed\/videoseries?list=PLE18841CABEA24090\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:44:07+0000","updated_at":"2018-11-07T16:44:07+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/2Op3QLzMgSY\/hqdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":310,"title":"Reconstructing Windows EFI files in \/boot partition (Linux+Windows dual boot)","url":"https:\/\/superuser.com\/questions\/1110644\/reconstructing-windows-efi-files-in-boot-partition-linuxwindows-dual-boot","content":"<p>I'm dual booting Windows 10 (pre-Anniversary Update) and Fedora 24 on a Lenovo IdeaPad N580 (Intel Pentium B960 2.2GHz, 4GB RAM, UEFI with Secure Boot disabled). I originally had just Windows 10 and then installed Fedora 23 alongside it and everything worked fine - it worked as well when I upgraded to Fedora 24. A while back, I decided I wanted to try out Ubuntu, and installed it only overwriting the root, \/boot and SWAP partitions and keeping my home partition. Windows 10 can still boot here. Then I decided I didn't like Ubuntu and reinstalled Fedora, with the 24 installer, trying to do exactly what I did with Ubuntu. The interface was a bit different, but I'm sure I did it correctly: after I did all the partition configuring, it gave me a popup saying \"We will only touch these partitions,\" and the partitions it listed were root, \/boot and SWAP. Nothing else.<\/p><p>Then, when it finished installing, I rebooted. The only options in the GRUB menu were Fedora and Fedora Recovery - nothing else. No Windows.<\/p><p>I tried installing rEFInd, which has worked for me in the past. But it didn't help, it gave me those same two options.<\/p><p>The Windows partition still exists and I can access it from Fedora - important files such as \/bootmgr and \/Boot\/BCD seem intact. The Windows partition is at \/dev\/sda5.<\/p><p>As well, in the BIOS settings where Windows Boot Manager used to appear, it doesn't. It just has rEFInd and an entry with the model number of my hard drive - which I'm assuming just brings me to the GRUB menu. (I haven't actually tried bringing it to the top of the list and rebooting though - I'm a bit lazy :P)<\/p><p>Can someone help?<\/p><p><strong>UPDATE:<\/strong> Manoj Sawai identified the issue: by formatting the \/boot partition I was getting rid of the Windows EFI files. So I need to reconstruct the files at <code>\/boot\/efi\/EFI\/Microsoft<\/code>.\nI think I've reconstructed the directory <code>\/boot\/efi\/EFI\/Microsoft\/Boot<\/code>, but I need the other files in <code>Microsoft<\/code>.\nCould someone tell me what files I need there?<\/p>","created_at":"2018-11-07T16:44:06+0000","updated_at":"2018-11-07T16:44:06+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"superuser.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/superuser\/img\/apple-touch-icon@2.png?v=e869e4459439","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":309,"title":"Stripe Atlas: Guide to Software as a Service, as a business","url":"https:\/\/stripe.com\/atlas\/guides\/business-of-saas","content":"<p>Software-as-a-service (SaaS) is a billing and delivery model for software which is so superior to the traditional method for selling software licenses that it restructures businesses around itself. This has led SaaS businesses to have a distinct body of practice. Unfortunately, many entrepreneurs discover this body of practice the hard way, by making mistakes that have been made before, rather than by spending their mistake budget on newer, better\u00a0mistakes.<\/p><p>This shouldn\u2019t include you, so we\u2019ll take you through a whirlwind tour of the state of play of SaaS businesses. You should gain a better understanding of the SaaS business model, be able to anticipate whether to sell your product on a low-touch or high-touch model, and (if you\u2019re already operating a SaaS business) be able to evaluate its health and start improving\u00a0it.<\/p><p>If you are a software entrepreneur, and you do not sell mobile applications (which have a separate billing model, imposed by the platforms\u2019 app stores), you should thoroughly understand the business of SaaS. This will let you make better decisions for your product (and company), allow you to see business-threatening problems months or years in advance of them being obvious, and help you in <a href=\"https:\/\/stripe.com\/atlas\/guides\/pitching\">communicating with\u00a0investors<\/a>.<\/p><p><strong>Customers<\/strong> love SaaS because it \u201cjust works.\u201d There is typically nothing to install to access it. Hardware failures and operational errors, which are extraordinarily common among machines which are not maintained by professionals, do not result in meaningful data loss. SaaS companies achieve availability numbers (for example, percent of time where the software is accessible and operating correctly) which materially improve upon the numbers achievable by almost every IT department (and every individual,\u00a0full-stop).<\/p><p>SaaS also generally <em>appears<\/em> less expensive than software sold on other billing models, which matters for e.g. users who are not sure which software they should adopt over long terms, or who have only a short-term need for the\u00a0software.<\/p><p><strong>Developers<\/strong> love SaaS principally because of the delivery model, not the billing\u00a0model.<\/p><p>Most SaaS is developed continuously and run on the company\u2019s infrastructure. (There are significant exceptions in SaaS in the enterprise, but the overwhelming majority of B2C and B2B SaaS sold outside the enterprise is accessed over the internet from servers maintained by the software\u00a0company.)<\/p><p>Software companies historically have not controlled the environments their code executes in. This is historically a major source of both development friction and customer support cases. All software deployed on customers\u2019 hardware suffers from differences in configurations of systems, interactions with other installed software, and operator error. This has to be both accounted for in development and dealt with as a customer services issue. Companies which sell their software on both SaaS and installable models frequently see 10+ times more support requests per customer from customers who install the software\u00a0locally.<\/p><p><strong>Businesses and investors<\/strong> love SaaS because the economics of SaaS are impossibly attractive relative to selling software licenses. Revenue from SaaS is generally recurring and predictable; this makes cash flows in SaaS businesses impressively predictable, which allows businesses to plan against them and (via investors) trade future cash flows for money in the status quo, which allows them to (generously) fund present growth. This has made SaaS companies into some of the fastest growing software companies in\u00a0history.<\/p><h2 id=\"saas-sales-models\">SaaS sales models<\/h2><p>There are, broadly speaking, two ways to sell SaaS. The selling model dictates almost everything else about the SaaS company and the product, to a degree which is shocking to first-time entrepreneurs. One of the classic mistakes in SaaS, which can take years to correct, is a mismatch between a product or market and the selected model to sell it\u00a0on.<\/p><p>You will find that the sales model for SaaS defines <em>much<\/em> more about a product (and company) than other distinctions, like whether a company sells to customers (B2C) or businesses (B2B), whether it is bootstrapped or riding the VC rocket ship trajectory, or what technology stack it is built\u00a0on.<\/p><p>Some products sell\u00a0themselves.<\/p><p>Low-touch SaaS is designed for the majority of customers to purchase it without sustained one-on-one interaction with a human being. The primary sales channels are the software\u2019s website, email marketing, and (very frequently) a free trial for the software, with the trial being aggressively optimized to be very, very low-friction to start, onboard, and successfully make sustained use of the\u00a0SaaS.<\/p><p>Low-touch products sometimes involve sales teams, but they\u2019re frequently structured as so-called \u201cCustomer Success\u201d teams, which are less focused on convincing people to buy the software and more on ensuring that users of the free trial successfully onboard and convert to paying users by the end of their\u00a0trials.<\/p><p>Customer support in low-touch products is generally handled primarily in scalable fashions, by optimizing the product to avoid incidents which would require human intervention, by creating educational resources which scale across the customer base, and by using humans as a last-resort. That said, many low-touch companies have <em>excellent<\/em> customer support teams. The economics of SaaS depend on the long-term satisfaction of customers, so even a product which expects only one ticket (a countable discrete interaction with a customer) every 20 customer-months might invest comparatively heavily in their CS\u00a0team.<\/p><p>Low-touch SaaS is generally sold on a month-to-month subscription with price points clustering around $10 for B2C applications in the $20 to $500 range for B2B. This corresponds to an average contract value (ACV) of approximately $100 to $5k. The term ACV isn\u2019t commonly even used by low-touch SaaS businesses, which typically describe themselves by their monthly price points, but it is important to do comparisons to high-touch SaaS\u00a0applications.<\/p><p>If you asked a low-touch SaaS entrepreneur for their most important metric, they would say MRR\u2014monthly recurring\u00a0revenue.<\/p><p><a href=\"https:\/\/basecamp.com\">Basecamp<\/a> is the paradigmatic example of a low-touch SaaS business. <a href=\"https:\/\/atlassian.com\">Atlassian<\/a> (which makes JIRA, Trello, Confluence, and several other products) is possibly the publicly-traded company with the most success with the\u00a0model.<\/p><p>Some customers need some help in deciding whether or how to adopt certain\u00a0products.<\/p><p>High-touch SaaS is designed around there being a human-intensive process to convince businesses to adopt the software, successfully operationalize it, and continue using\u00a0it.<\/p><p>The beating heart of the organization is almost always the sales teams, which are often broken down into specialized roles: sales development representatives (SDRs) who find prospects for the software, account executives (AEs) who own the sales process against particular customers, and account managers (AMs) who are responsible for the happiness and continued performance of an individualized portfolio of\u00a0accounts.<\/p><p>The sales team is typically supported by marketing, whose primary job is generating a sufficient pipeline of qualified leads for the sales team to evaluate and\u00a0close.<\/p><p>There are many truly excellent products sold on the high-touch model, but to a first approximation, engineering and product are generally considered less important in high-touch SaaS businesses than the sales engine\u00a0is.<\/p><p>The organization of customer support is highly variable across high-touch SaaS companies; a commonality is that it is generally expected to be heavily utilized. The number of tickets per account per period is expected to be orders of magnitude higher than it is in low-touch\u00a0SaaS.<\/p><p>Note that while, in principle, one can make high-touch sales to consumers (for example, insurance has historically been sold primarily through commissioned agents), in SaaS, the overwhelming majority of high-touch businesses sell to businesses (B2B). Within B2B, there is a wide range of expected customer profiles, ACVs (defined variously as average contract value or annual contract value), and deal\u00a0complexity.<\/p><p>On the low-end, SaaS sold to \u201csmall and medium sized businesses\u201d (SMBs) on a high-touch model generally has an ACV of $6k to $15k, though this can range higher. The exact definition of an SMB varies heavily depending on who you ask; operationally, it is \u201cany business with sufficient sophistication to successfully adopt software which costs $10,000\u201d, which probably excludes your local flower shop but includes a dental practice with 2 partners and 4\u00a0employees.<\/p><p>The high end is usually called \u201cthe enterprise\u201d and targets extremely large businesses or governments. True enterprise deals start in the six figures; there is no ceiling. (There is a $70 million ACV customer in <a href=\"http:\/\/phx.corporate-ir.net\/External.File?item=UGFyZW50SUQ9NjYxODU5fENoaWxkSUQ9MzgyOTU1fFR5cGU9MQ==&amp;t=1\">Inovalon\u2019s annual report<\/a>, for\u00a0example.)<\/p><p>If you asked a high-touch SaaS entrepreneur for their most important metric, they would say ARR\u2014annually recurring revenue. (This is essentially all of the non-churned revenue of the company, minus certain non-recurring items such as one-time setup fees, consulting services, and similar. Since the economics of SaaS are attractive because of growth over time, one-off revenue, particularly comparatively low-margin one-off revenue, is not maximally interesting to entrepreneurs or\u00a0investors.)<\/p><p>Salesforce is the paradigmatic example of a high-touch SaaS business, and they literally <a href=\"https:\/\/www.amazon.com\/Predictable-Revenue-Business-Practices-Salesforce-com\/dp\/0984380213\">wrote the book<\/a> on the model. Small high-touch SaaS businesses exist in multitudes, though they\u2019re less visible than low-touch SaaS businesses, principally because visibility is a customer acquisition strategy in low-touch SaaS and not always optimal in high-touch SaaS. For example, there are many small SaaS businesses which quietly make six or seven figures a year selling services to a tightly defined\u00a0vertical.<\/p><h2 id=\"hybrid-sales-approaches\">Hybrid sales approaches<\/h2><p>There exist companies which successfully run a low-touch and high-touch business with functionally the same product. They are <em>exceedingly<\/em> rare relative to SaaS businesses. The most common result of attempting both models simultaneously is that only one of the models receives any traction, and (because these models weave themselves into all operations of the company) it typically strangles the\u00a0other.<\/p><p>A more common form of hybridization is adopting certain elements of the other sales model. For example, many low-touch SaaS businesses have customer success teams which, if you squint at them, look almost like inside sales. High-touch companies typically borrow fewer tactics than low-touch companies; the most common one is having a product that the company does not (materially) sell which they distribute in a low-touch fashion for the purpose of lead generation for the product the company actually\u00a0sells.<\/p><p>The SaaS model, fundamentally, works by financializing software: instead of selling software as a product with a sticker price, it sells the software as if it were a financial instrument, with a probabilistically forecastable cash\u00a0flow.<\/p><p>There are more sophisticated ways to model a SaaS business, but the no-MBA-required version just makes a few simplifying assumptions (like ignoring the time-value of money) and uses high-school math. If you only learn one thing about SaaS, learn this equation; it is the Rosetta Stone to understanding all material facts about a SaaS\u00a0business.<\/p><p>The core insight is really simple: one\u2019s revenue, over the long term, is the number of customers times the average lifetime revenue per\u00a0customer.<\/p><p>The number of customers you get is a product of two factors: <strong>acquisition<\/strong> (how effective you are at attracting the attention of prospects in low-touch SaaS or identifying and getting in front of them in high-touch SaaS) times your <strong>conversion rate<\/strong> (the percent of prospects you convert into paying\u00a0customers.)<\/p><p>The average lifetime revenue per customer (often called <strong>lifetime value (LTV)<\/strong>) is the product of how much they pay you for a particular period (such as one month) and how many periods they persist using your\u00a0service.<\/p><p>The <strong>average revenue per user (ARPU)<\/strong> is simply the average revenue for an account over any particular\u00a0period.<\/p><p>The <strong>churn<\/strong> is the percent of customers over a given period who do not continue paying for services. For example, if you have 200 customers pay you in January and only 190 of those pay you in February, the churn would be\u00a05%.<\/p><p>The lifetime of a customer can, with a few simplifiying assumptions, be calculated as the <a href=\"https:\/\/www.khanacademy.org\/math\/calculus-home\/series-calc\/geo-series-calc\/v\/infinite-geometric-series\">sum of an infinite geometric series<\/a>; this works out to simply taking the inverse of churn. A product which loses 5% of its customers per month has an expected customer lifetime of 20 months; if it charges each customer $30 a month, it has an expected lifetime revenue of $600 per new customer signed\u00a0up.<\/p><h2 id=\"implications-of-the-saas-business-model\">Implications of the SaaS business model<\/h2><p><strong>Improvements to a SaaS business are multiplicatively\u00a0effective.<\/strong><\/p><p>A 10% improvement to acquisition (via e.g. better marketing) and a 10% improvement to conversion rate (via e.g. product improvements or more effective sales techniques) sum to a 21% improvement (1.1 * 1.1), not a 20%\u00a0improvement.<\/p><p><strong>Improvements to a SaaS business are <em>incredibly<\/em>\u00a0leveraged.<\/strong><\/p><p>Because the margins in SaaS are so high, the long-term valuation of a SaaS business is effectively tied to some multiple of its long-term revenues. Thus, a 1% improvement in conversion rates doesn\u2019t simply mean a 1% increase in revenue next month or even over the long term\u2026 it implies a 1% increase in <em>enterprise value of the\u00a0company<\/em>.<\/p><p><strong>Price is the easiest lever to improve a SaaS\u00a0business.<\/strong><\/p><p>Aquisition, conversion, and churn often require major cross-functional efforts to improve. Pricing typically requires replacing a small number with a bigger one. (There exists enough nuance here that we wrote a guide to <a href=\"https:\/\/stripe.com\/atlas\/guides\/saas-pricing\">pricing\u00a0SaaS<\/a>.)<\/p><p><strong>SaaS businesses eventually\u00a0asymptote.<\/strong><\/p><p>Given fixed acquisition, conversion, and churn, there will be a point at which one\u2019s business hits a revenue plateau. This is predictable in advance: the number of customers at the plateau is equal to acquisition times conversion divided by churn\u00a0rate.<\/p><p>A SaaS business which loses ability to improve acquisition, conversion, or churn will, with almost mathematical certainty, stop growing. A SaaS business which stops growing before it can cover fixed costs (like e.g. salaries for the engineering team) dies ignominiously, <em>even if they did everything\u00a0right<\/em>.<\/p><p><strong>SaaS businesses can be capital-intensive to\u00a0grow.<\/strong><\/p><p>SaaS businesses have large front-loaded costs to grow, particularly when growing aggressively; marketing and sales dominates the marginal cost per customer and, often, the total expenditures of the company. The marketing and sales costs attributable to a particular customer occur very early in that customer\u2019s lifecycle; the revenue to eventually pay for those costs comes\u00a0later.<\/p><p>This means that a SaaS company <a href=\"http:\/\/www.paulgraham.com\/growth.html\">optimizing for growth<\/a> will <em>almost always<\/em> spend more money in a given period than they collect from customers. The money spent has to come from somewhere. Many SaaS companies choose to fund the growth via selling equity in the companies to investors. SaaS companies are <em>particularly<\/em> attractive to investors because the model is very well-understood: create a product, achieve some measure of product-market fit, spend a lot of money on marketing and sales according to a relatively repeatable playbook, and eventually sell one\u2019s stake in the business to someone else (the public markets, an acquirer, or another investor looking for a derisked business with good growth\u00a0potential).<\/p><p><strong>Margins, to a first approximation, don\u2019t\u00a0matter.<\/strong><\/p><p>Most businesses care quite a bit about their cost-of-goods-sold (COGS), the cost to satisfy a marginal\u00a0customer.<\/p><p>While some platform businesses (like AWS) have material COGS, at the typical SaaS company, the primary source of value is the software and it can be replicated at an extremely low COGS. SaaS companies frequently spend less than 5~10% of their marginal revenue per customer on delivering the underlying\u00a0service.<\/p><p>This allows SaaS entrepreneurs to almost ignore every factor of their unit economics except customer acquisition cost (CAC; the marginal spending on marketing and sales per customer added). If they\u2019re quickly growing, the company can ignore every expense that doesn\u2019t scale directly with the number of customers (i.e. engineering costs, general and administrative expenses, etc), on the assumption that growth at a sensible CAC will outrun anything on the expenses side of the\u00a0ledger.<\/p><p><strong>SaaS businesses take a while to\u00a0grow.<\/strong><\/p><p>While tales of so-called \u201chockeystick\u201d growth curves are common in the press, the representative experience of SaaS companies is that they take a very long time dialing in the product, marketing approaches, and sales approaches before things start to work very well. This has been referred to, memorably, as the <a href=\"http:\/\/businessofsoftware.org\/2013\/02\/gail-goodman-constant-contact-how-to-negotiate-the-long-slow-saas-ramp-of-death\/\">Long Slow SaaS Ramp of\u00a0Death<\/a>.<\/p><p><strong>Growth expectations vary widely in the SaaS\u00a0industry.<\/strong><\/p><p>Bootstrapped SaaS businesses often take 18 months before they\u2019re profitable enough to be competitive with reasonable wages for the founding team. After achieving that point, bootstrapped businesses have a wide range of acceptable outcomes for growth rates; 10~20% year over year growth rates in revenue can produce very, very happy outcomes for all\u00a0concerned.<\/p><p>Funded SaaS businesses are designed to trade cash for growth, which means they\u2019re designed to lose a lot of money upfront while perfecting their model; almost no funded SaaS business ever has failed at that\u00a0goal.<\/p><p>After they perfect the model, they scale it, which generally results in losing more money, faster. That this is a successful outcome for the business is counterintuitive to many observers of the software industry. If the business can continue growing, there is no size of accumulated deficit that it cannot eventually repay. If growth does not happen, the business\u00a0fails.<\/p><p>There exist many lower-stress businesses in life than SaaS companies being managed for aggressive growth; it\u2019s likened to riding a rocket ship, where you burn fuel aggressively to achieve acceleration and, by the way, if anything goes wrong you\u00a0explode.<\/p><p>The rule of thumb for growth rate expectations at a successful SaaS company being managed for aggressive growth is 3, 3, 2, 2, 2: starting from a material baseline (e.g. over $1 million in annual recurring revenue (ARR)), the business needs to triple annual revenues for two consecutive years and then double them for three consecutive years. A funded SaaS business which consistently grows by 20% per year early in its life is likely a failure in the eyes of its\u00a0investors.<\/p><p>One of the most popular questions for SaaS founders is \u201cAre my numbers any\u00a0good?\u201d<\/p><p>This is surprisingly difficult to answer, because of the differences across industries, business models, stages of a company, and goals of founders. In general, though, experienced SaaS entrepreneurs have a few rules of\u00a0thumb.<\/p><h2 id=\"low-touch-saas-benchmarks\">Low-touch SaaS benchmarks<\/h2><p><strong>Conversion\u00a0rate:<\/strong><\/p><p>Most low-touch SaaS uses a free trial, with the signup either requiring minimal information <em>or<\/em> a credit card that will be billed if the user doesn\u2019t cancel the trial. This decision dominates the character of the free trial: users who sign in to a relatively low-friction trial may not be very serious about evaluating the software and need to affirmatively decide to purchase the software later, while users who provide a credit card number generally have done more up-front research and are, essentially, committing to pay unless they affirmatively declare they are dissatisfied with the\u00a0product.<\/p><p>This results in cosmically different conversion\u00a0rates:<\/p><p><strong>Conversion rates of low-touch SaaS trials with credit card not\u00a0required:<\/strong><\/p><ul><li>substantially below 1%: generally evidence of poor product-market fit<\/li>\n  <li>~1%: roughly the baseline for competent execution<\/li>\n  <li>2%+: extremely good<\/li>\n<\/ul><p><strong>Conversion rates of low-touch SaaS trials with credit card\u00a0<em>required<\/em>:<\/strong><\/p><ul><li>substantially below 40%: generally evidence of poor product-market fit<\/li>\n  <li>40%: roughly the baseline for competent execution<\/li>\n  <li>60%: doing well!<\/li>\n<\/ul><p>In general, requiring a credit card upfront will, on net, increase the number of new paying customers you get (it increases the trial-to-paying-customer conversion rate by more than it decreases the number of trials started). This factor <em>reverses<\/em> as a company gets increasingly sophisticated about <strong>activating<\/strong> free trial users (ensuring they make meaningful use of the software), typically via better in-product experiences, lifecycle email, and customer success\u00a0teams.<\/p><p><strong>Conversion rate (to\u00a0trials):<\/strong><\/p><p>You should measure your conversion rate between unique page views and trials started, but it isn\u2019t the most actionable metric in your company, and it is difficult to give a good guideline for your expectations driven from this\u00a0number.<\/p><p>Conversion rate to the trial is incredibly sensitive to whether you are attracting high-quality visitors or not. Counterintuitively, companies which are better at marketing have <em>lower<\/em> conversion rates than companies which are worse at\u00a0it.<\/p><p>The companies with better marketing attract many more prospects, including typically a larger percentage and absolute number of prospects who are not a good fit for the offering. Companies that are worse at marketing are only discovered by the cognoscenti of their markets, who tend to be disproportionately good customers; they\u2019re so dissatisfied with the status quo that they\u2019re actively searching for solutions, often intensely, and they\u2019re willing to use a no-name company if it is possibly better than their current situation. The rest of the market might not be actively looking for a solution right now, might be satisfied with going with well-known players or only those who show up prominently on Google, and might not be incentivized to take on vendor risk for dealing with a newer\u00a0provider.<\/p><p><strong>Churn\u00a0rates:<\/strong><\/p><p>In low-touch SaaS, most customers are on month-to-month contracts, and churn rates are quoted monthly. (Selling annual accounts is certainly a good idea, too, both for the upfront cash collected and because they have lower churn rates. When reporting churn, though, typically the impact of them is blended in to produce a monthly\u00a0number.)<\/p><ul><li>2%: a very sticky product, with strong product-market fit and substantial investments in reducing involuntary churn<\/li>\n  <li>5%: roughly where you expect to start<\/li>\n  <li>7%: you likely have either low-hanging fruit for preventing voluntary churn or are selling to a difficult market<\/li>\n  <li>10%+: Evidence of very poor product market fit and <strong>an existential threat to the company<\/strong><\/li>\n<\/ul><p>Some markets structurally have higher churn than others: selling to \u201cpro-sumers\u201d or informal businesses such as freelancers exposes oneself to their high rate of exiting the business, which materially impacts churn rates. More established businesses fail far less frequently and have far less need to optimize their cash flows to the last\u00a0$50.<\/p><p>Since higher price points preferentially select for better customers, increasing prices is <strong>even more effective than entrepreneurs expect<\/strong>: increasing prices by 25% can result in \u201caccidentally\u201d decreasing churn by 20%, simply by changing the mix of customers who buy the product. <strong>This factor leads many, many low-touch SaaS businesses to march \u201cupmarket\u201d over\u00a0time.<\/strong><\/p><h2 id=\"high-touch-saas-benchmarks\">High-touch SaaS benchmarks<\/h2><p>High-touch SaaS businesses generally have much, much more heterogeneity with regards to both how they measure their conversion rates (largely due to differences in how they define an \u201copportunity\u201d) and in their realized conversion rates given similar definitions, due to differences in their industry, sales process, and so\u00a0forth.<\/p><p><strong>Churn rates<\/strong>, though, are closely clustered: roughly 10% <em>annualized churn<\/em> is reasonable for companies in their early years. 7% is an excellent churn rate. Note that mediocre high-touch SaaS businesses have materially lower churn rates than even the best low-touch SaaS businesses,\u00a0structurally.<\/p><p>High-touch businesses often measure so-called \u201clogo\u201d churn (one business counts as one logo, regardless of how many units at that business use one\u2019s software, how many seats they use, what they are paying, etc) and revenue churn. This is less important in low-touch SaaS, as those churn rates tend to be quite\u00a0similar.<\/p><p>Because high-touch SaaS businesses typically price their offerings such that they can increase the amount of revenue over the lifetime of a customer, by selling more seats or by offering additional products or similar, many of them track <strong>net revenue churn<\/strong>, which is the difference in revenue per cohort per year. The gold-standard for a high-touch SaaS business is <strong>negative net revenue churn<\/strong>: the impact of upgrades, increases in contract size on a year-to-year basis, and cross-selling to existing customers exceeds the revenue impact of customers deciding to terminate (or reduce) their use of the software. (Virtually no low-touch SaaS business achieves net negative churn; their churn rates are too high to\u00a0outrun.)<\/p><p>SaaS isn\u2019t just about the metrics. The hardest thing to put a number on early in the lifetime of a SaaS company is called <strong>product\/market fit<\/strong>, a term <a href=\"http:\/\/pmarchive.com\/guide_to_startups_part4.html\">coined<\/a> by Marc Andreessen, which informally means \u201cHave you found a group of people who <em>love<\/em> the thing you have built for\u00a0them?\u201d<\/p><p>Products which don\u2019t have product\/market fit yet are plagued by relatively low conversion rates and high churn rates. Products which achieve product\/market fit often accelerate their growth rates materially, have much higher conversion rates, and are generally more pleasant to work\u00a0on.<\/p><p>Serial SaaS entrepreneurs often struggle to describe product\/market fit other than to say \u201cIf you have it you will <em>know<\/em> that you have it, and if you have any doubt whether you have it, you do not.\u201d It\u2019s the difference between every sales conversation being you pushing a boulder up a hill and the customer practically pulling your hand off to get your\u00a0software.<\/p><p>Many SaaS with product\/market fit <strong>did not launch with it<\/strong>; it sometimes takes months or years of iterating to get there. The most important theme while iterating is to talk to many, many more customers than feels natural. Low-touch SaaS entrepreneurs can make an excuse to attempt to speak with <em>literally every<\/em> person who signs up for a free trial; the economics of this are unsustainable at the price point but <em>running a SaaS company without product\/market fit is also unsustainable<\/em>, so it\u2019s entirely justified by how much you\u00a0learn.<\/p><p>Achieving product\/market fit isn\u2019t just a matter of listening to feature requests and building those features. It is also listening closely to the commonalities of your best customers and leaning in on them. This can result in changes to the marketing, messaging, and design of the product to more closely target the needs of the best\u00a0customers.<\/p><p>Who are the \u201cbest\u201d customers? Generally speaking, they\u2019re the segments (by industry, size, user profile, or similar) where you have high conversion rates, low churn rates, and (almost always) relatively higher ACV. By far the most common change in emphasis of low-touch SaaS businesses is to launch with a product which serves a wide spectrum of users at a wide spectrum of sophistication, and then double-down on one or two niches for their most sophisticated\u00a0users.<\/p><p>Stripe Atlas is going to be publishing further guides on finding product\/market fit, interviewing users, and optimizing every facet of your online business. If you\u2019d like to hear about them, please give us your email address. If you have any thoughts about what other guides would be useful to your online business, please write us at atlas@stripe.com\u00a0.<\/p><aside class=\"collect-email-guide\">\n  <\/aside><a href=\"https:\/\/stripe.com\/atlas\/guides\" class=\"common-Link common-Link--arrowL\">Return to guides<\/a>","created_at":"2018-11-07T16:44:05+0000","updated_at":"2018-11-07T16:44:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":22,"domain_name":"stripe.com","preview_picture":"https:\/\/stripe.com\/img\/v3\/atlas\/guide\/social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":306,"title":"Data Science up and down the Ladder of Abstraction","url":"https:\/\/www.infoq.com\/articles\/data-science-abstraction","content":"<p><em>Data science is fast becoming a critical skill for developers and managers across industries, and it looks like a lot of fun as well. But it\u2019s pretty complicated - there are a lot of engineering and analytical options to navigate, and it\u2019s hard to know if you\u2019re doing it right or where the bear traps lie. In this series we explore ways of\u00a0making sense of data science - understanding where it\u2019s needed and where it\u2019s not, and how to make it an asset for you, from people who\u2019ve been there and done it.<\/em><\/p><p><em>This InfoQ article is part of the series \"<a href=\"https:\/\/www.infoq.com\/handle-data-science\" target=\"_blank\">Getting A Handle on Data Science<\/a>\" . You can subscribe to receive notifications via\u00a0<a href=\"https:\/\/www.infoq.com\/feed\/handle-data-science\" target=\"_blank\">RSS<\/a>.<\/em><\/p><table border=\"0\" cellpadding=\"1\" style=\"border-spacing: 1px;\"><tbody><tr><td>&#13;\n\t\t\t<h2>Key takeaways<\/h2>&#13;\n&#13;\n\t\t\t<p>As a data scientist working primarily in the Clojure programming language, I'm often asked why I don't use one of the more popular alternatives. In this article I'd like to set out my reasons for adopting Clojure, including its:<\/p>&#13;\n&#13;\n\t\t\t<ul><li>Excellent tooling for REPL-driven development and notebooks for exploratory analysis and live charting.<\/li>\n\t\t\t\t<li>Suitability for creating shareable browser-based visualisations and interactive simulations with modern UI frameworks like Facebook's React.<\/li>\n\t\t\t\t<li>Interoperability with Java, JavaScript, and Scala for making use of established and actively developed numerical optimisation, distributed computing, and machine learning libraries.<\/li>\n\t\t\t\t<li>Data-oriented functional core with a focus on composable sequence processing interfaces which scale with the volume of data you have.<\/li>\n\t\t\t\t<li>Clean and unambiguous syntax which even aids my comprehension of transcribed algorithms.<\/li>\n\t\t\t\t<li>Macro system which enables extensibility at the language level for custom algorithmic abstractions and cleaner code.<\/li>\n\t\t\t<\/ul><\/td>\n\t\t<\/tr><\/tbody><\/table><p>Although Clojure lacks the extensive toolbox and analytic community of the most popular data science languages, R and Python, it provides a powerful environment for developing statistical thinking and for practicing effective data science.<\/p><p>If you're already working as a data scientist, I hope to show what learning Clojure will bring to your toolchain. But if you're a programmer looking to learn the principles of data science for the first time, I hope that you'll consider Clojure as a language with which to explore this exciting analytical world.<\/p><h2>Highway to the Danger Zone<\/h2><p>Like so many others, my own journey to data science came via software engineering. In 2015, the\u00a0<a href=\"https:\/\/www.stitchdata.com\/resources\/reports\/the-state-of-data-science\/\">State of Data Science Survey<\/a>\u00a0found that the most popular major for data scientists on LinkedIn was computer science, and that most had added their skill within the previous four years. So, when Packt Publishing invited me to write a\u00a0<a href=\"https:\/\/www.amazon.com\/Clojure-Data-Science-Henry-Garner\/dp\/1784397180\/\">book introducing data science for Clojure programmers<\/a>, I gladly agreed. In addition to appreciating the need for such a book, I saw it as an opportunity to consolidate much of what I'd learned during the four years prior as chief technologist of my own data analytics company.<\/p><p>If you're thinking of developing your skills in data science, you've probably already considered Python or R. Python is an especially popular choice for those coming from a programming background since it's a good general-purpose scripting language which also provides access to excellent statistical and machine learning libraries. When I first started out in data science I used Python and\u00a0<a href=\"http:\/\/scikit-learn.org\/\">scikit-learn<\/a>\u00a0to tackle a clustering project. I had some data gathered from social media on users' interests and I was trying to determine if there were cohorts of users within the whole. I chose\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Spectral_clustering\">spectral clustering<\/a>\u00a0because it could identify non-globular clusters (so must be better, I reasoned), and the first results were promising. My confidence quickly evaporated when I re-ran the clustering and got different results. Behind the scenes, the library made use of\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/K-means_clustering\">k-means<\/a>, a\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Stochastic\">stochastic<\/a>\u00a0algorithm, and I was finding a new local minimum. Frustrated by the apparent lack of objective truth, but needing to show some results, I simply selected the result that looked most like what I expected and moved on.<\/p><p>I was wielding machine learning without a licence and using a black box I didn't really understand to justify the outcome I hoped for. I didn't know then about cluster evaluation metrics and methods for quantifying and controlling variance; let alone cross-validation, high bias or\u00a0the curse of dimensionality. I was sitting squarely in the segment\u00a0<a href=\"http:\/\/drewconway.com\/zia\/2013\/3\/26\/the-data-science-venn-diagram\">Drew Conway famously termed the \"danger zone\"<\/a>.<\/p><p>Back in 1711 Alexander Pope observed in his first major poem\u00a0<em>An Essay on Criticism<\/em>:<\/p><blockquote><div>&#13;\n<p>A little learning is a dangerous thing<\/p>&#13;\n<\/div><\/blockquote><p>His advice was directed at student literary critics, but it could readily apply to apprentice data scientists as well.<\/p><h2>How do we get people to understand data science?<\/h2><p>Advice abounds on the\u00a0<a href=\"http:\/\/blog.udacity.com\/2014\/11\/data-science-job-skills.html\">prerequisites for being a successful data scientist<\/a>: lists which often contain intimidating-sounding mathematical techniques such as multivariable calculus and linear algebra or probability theory. Whilst such suggestions are well intentioned, it's is not useful advice for most aspiring data scientists. Advanced practitioners making retrospective suggestions forget how overwhelming starting out can be.<\/p><p>Bret Victor's excellent article\u00a0<a href=\"http:\/\/worrydream.com\/LearnableProgramming\/\">Learnable Programming<\/a>\u00a0emphasises the importance of\u00a0following flow\u00a0when learning programming. Since it's also a highly technical skill, the guiding principles he presents can be easily paraphrased for data science:<\/p><ul><li>Data science\u00a0is a way of thinking, not a rote skill.\u00a0Learning about\u00a0clustering algorithms\u00a0is not learning to do data science any more than learning about pencils is to learning to draw.<\/li>\n\t<li>People understand what they can see.\u00a0If a\u00a0data scientist\u00a0cannot see what\u00a0an algorithm\u00a0is doing, she can't understand it.<\/li>\n<\/ul><p>I think the best way to get started with data science, as with programming, is by exploring a problem domain interactively with real-time feedback. This could involve writing little programs which make use of libraries that provide algorithms, or by implementing the algorithms yourself from tutorials and books.<\/p><p>The\u00a0<a href=\"http:\/\/www.tryclj.com\/\">REPL<\/a>\u00a0is a tool for experimenting with code and which allows you to interact with a running program and quickly try out ideas. It does this by presenting you with a prompt where you can enter code. It then reads your input, evaluates it, prints the result, and loops, presenting you with a prompt again. A REPL environment falls short of Bret Victor's ideal programming environment (by requiring the programmer to write a full expression at a time), but a well set up REPL environment comes a lot closer than a workflow which treats writing a script, perhaps compiling it, and then running it, as discrete operations.<\/p><p>Python's interactive interpreter and the R prompt also permit exploration of code expression-by-expression. But REPL-driven development is embedded in Clojure culture, and popular Clojure development environments such as\u00a0<a href=\"https:\/\/www.gnu.org\/software\/emacs\/\">Emacs<\/a>\u00a0and\u00a0<a href=\"https:\/\/cursive-ide.com\/\">Cursive<\/a>\u00a0give the REPL prominent status. Fantastic tools such as\u00a0<a href=\"https:\/\/github.com\/clojure-emacs\/cider\">Cider<\/a>\u00a0make interactive REPL development, debugging, and refactoring a real pleasure.<\/p><h2>Follow the flow<\/h2><p>The fundamental\u00a0way of thinking\u00a0which a programmer must develop as they learn data science is characterised by Richard Hamming's assertion:<\/p><blockquote><div>&#13;\n<p>In science, if you know what you are doing, you should not be doing it. In engineering, if you do not know what you are doing, you should not be doing it.<\/p>&#13;\n<\/div><\/blockquote><p>This is only a slight oversimplification. Doing data science will, by definition, involve open-ended exploration. For me, skills to manage and contain this uncertainty were developed when I began to understand descriptive and inferential statistics, as well as how the\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Relationships_among_probability_distributions\">common distributions relate to each other<\/a>. Along with data cleansing, they're the majority of what I bring to the data science projects I'm involved in, and I devoted the first two chapters of\u00a0<a href=\"https:\/\/www.amazon.com\/Clojure-Data-Science-Henry-Garner\/dp\/1784397180\/\">Clojure for Data Science<\/a>\u00a0simply to building a strong foundation in them. Unfortunately, they're often overlooked by programmers because they can't be understood by looking at code alone, and there are few resources introducing them to a programming audience.<\/p><p>The Incanter code examples at\u00a0<a href=\"https:\/\/data-sorcery.org\/contents\/\">data-scorcery.org<\/a>\u00a0(particularly the probability and statistics sections) provide useful starting points.\u00a0<a href=\"https:\/\/github.com\/incanter\/incanter\">Incanter<\/a>\u00a0itself provides Clojure's most complete set of statistical computing libraries and contains implementations of common statistical functions, common distributions, and charting capabilities. Following along the examples in the REPL and seeing the output at each stage is an excellent way to learn.<\/p><p>To consolidate this new knowledge, I read the fantastic free book\u00a0<a href=\"http:\/\/greenteapress.com\/thinkstats\/\">Think Stats<\/a>. Adapting the book's many exercises and solving them in Clojure helped to build my Clojure and statistics knowledge simultaneously. A web search for\u00a0<a href=\"https:\/\/www.google.co.uk\/?q=think+stats+in+clojure\">'think stats in Clojure'<\/a>will throw up many examples from others who've done the same thing.<\/p><blockquote><div>&#13;\n<p>The most exciting phrase to hear in science, the one that heralds new discoveries, is not 'Eureka!' but 'That's funny...'<\/p>&#13;\n&#13;\n<p>-- Isaac Asimov<\/p>&#13;\n<\/div><\/blockquote><p>Incidentally, some of the most counter-intuitive and interesting things I learned researching the book didn't involve any code at all. A prime example is\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Simpson%27s_paradox\">Simpson's Paradox<\/a>, in which the same trend appears in different groups of data but disappears or even reverses when these groups are combined. The Wikipedia article helpfully includes visualisations that neatly illustrate this apparent contradiction and makes clear its causes.<\/p><h2>Visualisation<\/h2><p>Visualisations provide useful insights into data that would be lost if you looked only at summary statistics and Incanter provides support for the most common charts and graphs. Unfortunately though, it's hard to overlook the fact that Incanter's charts are not particularly attractive. Additionally, when producing visuals from within a REPL, Incanter's charts will open up a separate window in the background.<\/p><p>The notebook format (which freely mixes interactive code and charts in a browser window) has become a popular way to do data science.\u00a0<a href=\"http:\/\/jupyter.org\/\">Jupyter<\/a>\u00a0is the most popular framework for producing notebooks and, with the\u00a0<a href=\"https:\/\/github.com\/roryk\/clojupyter\">clojupyter<\/a>\u00a0kernel, Jupyter notebooks can also run Clojure code too. For tighter integration between Clojure and charting\u00a0<a href=\"http:\/\/gorilla-repl.org\/\">Gorilla REPL<\/a>\u00a0provides an alternative notebook stack written entirely in Clojure. Although it ships with an attractive set of default charts, it also implements its\u00a0<a href=\"http:\/\/gorilla-repl.org\/renderer.html\">renderer<\/a>\u00a0in an extensible way. For example the\u00a0gg4clj\u00a0library extends Gorilla REPL to\u00a0<a href=\"http:\/\/viewer.gorilla-repl.org\/view.html?source=github&amp;user=JonyEpsilon&amp;repo=gg4clj&amp;path=ws\/demo.clj\">display visualisations using R's formidable ggplot2 charting library<\/a>. A modified version of Gorilla REPL is used by\u00a0<a href=\"https:\/\/github.com\/sbelak\/huri\">Huri<\/a>, a library providing \"tools for the lazy data scientist\".<\/p><p>The most flexible charting library is one that allows access to low-level drawing APIs.\u00a0<a href=\"https:\/\/github.com\/thi-ng\/geom\/blob\/master\/geom-viz\/src\/index.org\">thi.ng\/geom-viz<\/a>\u00a0provides this level of abstraction. It contains utility functions for drawing axes, mapping between chart and data scales, and functions for quickly rendering data as common chart types, but it otherwise offers a blank canvas and a set of primitives for rendering charts in the SVG graphics format. Since SVG can be rendered natively in web browsers,\u00a0thi.ng\/geom-viz\u00a0enables the creation of sophisticated custom web-based visualisations. Producing visualisations using the same language as the core analysis can mitigate the sorts of issues that arise when the visualisation is\u00a0<a href=\"http:\/\/viz.wtf\/\">done in a drawing program by someone with no understanding of the data<\/a>.<\/p><h2>Interactivity<\/h2><p>Like many Clojure libraries,\u00a0thi.ng\/geom-viz\u00a0doesn't depend on any Java host features and can be compiled to both JVM bytecode and JavaScript. Even low-level Clojure matrix manipulation libraries such as\u00a0<a href=\"https:\/\/github.com\/mikera\/core.matrix\">core.matrix<\/a>\u00a0can readily be compiled into JavaScript. This means that it's not only the visual output which can be rendered in the browser: almost any part of the Clojure code (with some exceptions relating to file access and Java-specific features such as threads) can be executed in the browser too. Thanks to JavaScript interoperability, it's possible to make use of native JavaScript libraries such as\u00a0<a href=\"https:\/\/github.com\/jstat\/jstat\">jStat<\/a>\u00a0and\u00a0<a href=\"https:\/\/d3js.org\/\">D3.js<\/a>\u00a0for your visualisations too.<\/p><p>One of the most helpful ways in which I built my understanding of inferential statistics was through interactive visualisation. In\u00a0<a href=\"https:\/\/github.com\/clojuredatascience\/ch2-inference\">chapter 2<\/a>, for example, I showed how to build a browser application which took repeated samples from several normal distributions and calculated the sample mean and the 95% confidence interval of the\u00a0<a href=\"http:\/\/stats.stackexchange.com\/questions\/269\/what-is-the-difference-between-a-population-and-a-sample\">population mean<\/a>\u00a0for each of them. It was clear to see the relationship between the confidence interval and sample size. And, since the confidence interval represents the bounds within which a value is expected to lie, repeated sampling demonstrated empirically that 5% of the time the population mean fell outside. Building simple tools like this can aid the development of deep understanding that might otherwise take months or years to develop.<\/p><p><a data-href=\"articles\/data-science-abstraction\/en\/resources\/Image-big.jpg\" href=\"https:\/\/res.infoq.com\/articles\/data-science-abstraction\/en\/resources\/Image-big.jpg\"><img alt=\"\" src=\"https:\/\/www.infoq.com\/articles\/articles\/data-science-abstraction\/en\/resources\/Image.jpg\" \/><\/a><\/p><p><small><strong>(Click on the image to enlarge it)<\/strong><\/small><\/p><p>Access to Facebook's\u00a0<a href=\"https:\/\/facebook.github.io\/react\/\">React<\/a>\u00a0library for user interfaces is provided by both the\u00a0<a href=\"https:\/\/reagent-project.github.io\/\">Reagent<\/a>\u00a0and\u00a0<a href=\"https:\/\/github.com\/omcljs\/om\/\">Om<\/a>\u00a0libraries. React Native opens the possibility to use the\u00a0<a href=\"http:\/\/cljsrn.org\/\">same Clojure code<\/a>\u00a0to create dedicated mobile applications.<\/p><h2>Data orientation<\/h2><p>The reasons I would recommend Clojure for data science are also why I think Clojure is an excellent language for any programmer to learn.<\/p><blockquote><div>&#13;\n<p>I'm confused by all these lines people draw between programming and data science. It's about encoding and representing information.<\/p>&#13;\n&#13;\n<p>--\u00a0<a href=\"https:\/\/twitter.com\/gigasquid\/status\/699556316637958144\">Carin Meier<\/a><\/p>&#13;\n<\/div><\/blockquote><p>Michael Nygard described Clojure as a\u00a0data-oriented\u00a0language in his article\u00a0<a href=\"http:\/\/blog.cognitect.com\/blog\/2016\/6\/28\/the-new-normal-data-leverage\">The New Normal: Data Leverage<\/a>. In his words:<\/p><blockquote><div>&#13;\n<p>For example, in Clojure, you deal with data as data. You don\u2019t try to force the data into an object-oriented semantic. Clojure has functions that work on any kind of sequential data, any kind of associative data and any kind of set data. Instead of writing a class that has ten or fifteen methods, you get to use hundreds of functions. (Over 640 in clojure.core alone.)<\/p>&#13;\n&#13;\n<p>So, it doesn\u2019t matter if you\u2019re using a map to represent a customer or an order or an order line item, you use the same functions to operate on all of them. The functions are more general, so they are more valuable because you can apply them more broadly.<\/p>&#13;\n<\/div><\/blockquote><p>The flexibility to model items in my domain simply as collections of associative or sequential data structures (or hierarchies thereof) provides ready access to data without prematurely flattening it into tables. It remains simple to extract numeric or categorical data from such structures to pass into the statistical functions of Incanter when required. The\u00a0<a href=\"https:\/\/github.com\/nathanmarz\/specter\">Specter<\/a>\u00a0library aims to augment Clojure's primitive data-manipulating functions specifically for extracting values from deeply nested data structures. As\u00a0<a href=\"https:\/\/twitter.com\/sbelak\">Simon Belak<\/a>, author of\u00a0<a href=\"https:\/\/github.com\/sbelak\/huri\">Huri<\/a>, succinctly put it:\u00a0<a href=\"http:\/\/www.slideshare.net\/simonbelak\/doing-data-science-with-clojure-65886938\">\"data frames considered harmful\"<\/a>.<\/p><p>Recently, Clojure 1.9 has been released with\u00a0<a href=\"http:\/\/clojure.org\/about\/spec\">clojure.spec<\/a>, a feature of the language which augments this flexible data-oriented approach to modelling with validation, error reporting, and generative test generation. Spec provides a means to encode the expected schema of data to catch errors and faulty assumptions early. Also, and perhaps more importantly for a data scientist, specs can serve as documentation:<\/p><blockquote><div>&#13;\n<p>The best and most useful specs (and interfaces) are related to purely information aspects. Only information specs work over wires and across systems. We will always prioritise, and where there is a conflict, prefer, the information approach.<\/p>&#13;\n<\/div><\/blockquote><p>Spec aims explicitly to make the code we write more comprehensible to our colleagues and also\u00a0to our future selves.<\/p><p>Even where Incanter has a function to perform some core statistical function, I found it useful when learning to implement my own version. I found this served two purposes: one was simply that writing my own version forced me to understand the function I was studying. The second was more surprising:\u00a0the Clojure implementation became a way for me to recall how the function behaved. It was something I could use to remind myself of what a function did much more effectively than by referring to the original mathematical notation.<\/p><p>Resources such as\u00a0<a href=\"http:\/\/www.evanmiller.org\/statistical-formulas-for-programmers.htm\">statistical formulas for programmers<\/a>\u00a0are useful only if you understand how to transform mathematical notation into code. Although it can appear obscure and upsetting, there are really only a handful of symbols that occur frequently in formulae. For example, \u03a3 is pronounced sigma and means sum. When you see it in mathematical notation it means that a sequence is being added up, and maps directly to the Clojure code:<\/p><p><code>(reduce + xs)<\/code><\/p><p>Where\u00a0xs\u00a0is the sequence over which to be summed. In my experience, Clojure's syntax aids the process of transcribing statistical approaches I encounter elsewhere.\u00a0<a href=\"https:\/\/twitter.com\/ztellman\">Zachary Tellman<\/a>\u00a0puts it like this in his book\u00a0<a href=\"http:\/\/elementsofclojure.com\/\">Elements of Clojure<\/a>:<\/p><blockquote><div>&#13;\n<p>This is possibly Clojure\u2019s most important property: the syntax expresses the code\u2019s semantic layers. An experienced reader of Clojure can skip over most of the code and have a lossless understanding of its high-level intent.<\/p>&#13;\n<\/div><\/blockquote><p>Assembling more sophisticated algorithms out of simple building blocks became a key feature of how I conveyed an intuition for machine learning algorithms in\u00a0<a href=\"https:\/\/github.com\/clojuredatascience\">Clojure for Data Science<\/a>. For example, I could start off writing small pure functions for calculating\u00a0<a href=\"https:\/\/github.com\/clojuredatascience\/ch4-classification\/blob\/0d6cdc041205177118aa403d9dfb00a76f66f4f2\/src\/cljds\/ch4\/stats.clj#L52\">information, entropy, weighted entropy, and information gain<\/a>, each in terms of the one before. Information gain quantifies how successful a partitioning of items is at grouping similar items together and different items apart. Applying this partitioning logic recursively yielded a\u00a0<a href=\"https:\/\/github.com\/clojuredatascience\/ch4-classification\/blob\/0d6cdc041205177118aa403d9dfb00a76f66f4f2\/src\/cljds\/ch4\/decision_tree.clj#L38\">simple tree classifier<\/a>. Clojure allowed me to express the essence of a particular approach because of its extremely succinct way of expressing such algorithms.<\/p><p>Bret Victor has written about how a deep understanding can be developed by moving up and down the\u00a0<a href=\"http:\/\/worrydream.com\/LadderOfAbstraction\/\">Ladder of Abstraction<\/a>:<\/p><blockquote><div>&#13;\n<p>Likewise, the most powerful way to gain insight into a system is by moving between levels of abstraction. Many designers do this instinctively. But it's easy to get stuck on the ground, experiencing concrete systems with no higher-level view. It's also easy to get stuck in the clouds, working entirely with abstract equations or aggregate statistics.<\/p>&#13;\n<\/div><\/blockquote><p>I found Clojure allowed me to easily move between low-level functions and high-level behaviours. Building my own versions of statistics functions gave me a much deeper understanding of how things worked and how they fit together. It also threw up some surprises that I might otherwise have glossed over, such as the fact that there is\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Quartile\">no universal agreement on selecting quartile values<\/a>.<\/p><p>No practicing data scientist has the time to routinely write their own algorithms from scratch. Even if they did, doing data science is about more than just algorithms; there's likely to be a component of data engineering too: getting your data out of a source system, scrubbing it, and perhaps capturing or presenting data to users of a live website too.<\/p><p>A fantastic aspect of using Clojure for data science is having access to the wealth of Java libraries available through Clojure's Java interop capabilities. For example, the parallel statistical computation library\u00a0<a href=\"https:\/\/github.com\/rwl\/ParallelColt\">ParallelColt<\/a>\u00a0powers much of Incanter's number crunching.\u00a0<a href=\"http:\/\/commons.apache.org\/proper\/commons-math\/\">Apache Commons Math<\/a>\u00a0also contains an enormous amount of functionality for flexible numeric analysis and optimisation, as well as genetic programming and machine learning with clustering. The documentation is excellent for describing how particular algorithms work. Likewise for neural networks, the\u00a0<a href=\"http:\/\/deeplearning4j.org\/\">Deeplearning4j<\/a>\u00a0library provides concrete implementations of many algorithms and its documentation is a veritable trove of useful tutorials on the subject.<\/p><p>Clojure wrappers already exist for some Java machine learning libraries. For example,\u00a0<a href=\"https:\/\/github.com\/joshuaeckroth\/clj-ml\">clj-ml<\/a>\u00a0provides Clojure interfaces for running linear regression as well as classification with logistic regression, naive Bayes, decision trees, and many other algorithms. It's built on top of\u00a0<a href=\"http:\/\/www.cs.waikato.ac.nz\/ml\/\">Weka<\/a>, an open source machine learning project written in Java. It's hard to beat Python's support for natural language processing, but the Clojure library\u00a0<a href=\"https:\/\/github.com\/dakrone\/clojure-opennlp\">clojure-opennlp<\/a>\u00a0wraps Java's Open NLP library.<\/p><p>If you really do have big data,\u00a0<a href=\"https:\/\/github.com\/damballa\/parkour\">Parkour<\/a>\u00a0and\u00a0<a href=\"http:\/\/cascalog.org\/\">Cascalog<\/a>\u00a0are Clojure libraries which take very different approaches to wrapping Hadoop's MapReduce framework. Cascalog was the reason that I adopted Clojure for data analysis in the first place. I was attracted by its terse syntax for expressing MapReduce jobs as solutions to logic programs. Parkour is now the Hadoop library I use most often, and it provides idiomatic Clojure wrappers around Hadoop's native capabilities. For example, it permits mappers and reducers to access to Hadoop's MapReduce context object at runtime, and access to the distributed cache to avoid costly joins for in-memory lookup data. It also allows custom schemas to be defined and serialised using Apache Avro using a pleasant and Clojure-friendly DSL.<\/p><p>Interop also permits access to distributed machine learning libraries such as\u00a0<a href=\"http:\/\/mahout.apache.org\/\">Mahout<\/a>\u00a0and Spark's\u00a0<a href=\"http:\/\/spark.apache.org\/mllib\/\">MLlib<\/a>\u00a0and\u00a0<a href=\"http:\/\/spark.apache.org\/graphx\/\">GraphX<\/a>\u00a0which are written in Scala. They're somewhat more convoluted to call from Clojure code since the bytecode emitted by Scala is more complicated than the equivalent Java. The\u00a0<a href=\"https:\/\/github.com\/t6\/from-scala\">from-scala<\/a>\u00a0library makes this easier, but it's still something of a dark art. I hope comprehensive wrappers emerge for them in the way that they have for Spark: Clojure has not one, but two libraries for interacting with Spark,\u00a0<a href=\"https:\/\/github.com\/gorillalabs\/sparkling\">Sparkling<\/a>\u00a0and\u00a0<a href=\"https:\/\/github.com\/yieldbot\/flambo\">Flambo<\/a>.<\/p><p>Recently, the\u00a0<a href=\"https:\/\/github.com\/thinktopic\/cortex\">Cortex<\/a>\u00a0library has been developed in collaboration with the author of\u00a0core.matrix. It provides neural networks, regression and feature learning implemented in pure Clojure. At the time of writing it's newly-released, but it already promises a coherent set of machine learning abstractions freed from Java's legacy, combined with fast GPU-based number crunching.\u00a0<a href=\"http:\/\/www.onyxplatform.org\/\">Onyx<\/a>\u00a0is a relatively new Clojure framework for distributed computation with an API based on immutable data structures.<\/p><p>The area of distributed machine learning addressed by\u00a0Mahout,\u00a0MLlib\u00a0and\u00a0GraphX\u00a0is where I hope the Clojure ecosystem will develop most over the next few years. Libraries such as\u00a0Cascalog\u00a0demonstrated the power that could come from applying Clojure's functional principles and terse syntax as a DSL for a large and complicated data processing framework. I would love to see this successful approach continue into the world of distributed machine learning too, and am following the development of\u00a0Cortex\u00a0and\u00a0Onyx\u00a0with interest.<\/p><h2>Functional programming<\/h2><p>Whilst Clojure provides access to tools for doing machine learning at significant scale, I'm usually asked to work on much smaller datasets. A colleague and good friend of mine\u00a0<a href=\"https:\/\/twitter.com\/otfrom\">Bruce Durling<\/a>\u00a0coined the term\u00a0<a href=\"https:\/\/rrees.me\/2016\/02\/25\/data-wrangling-with-clojure\/\">awkward-sized data<\/a>\u00a0for those volumes of data which are too large to be processed quickly on a desktop computer, but also too small to warrant a distributed framework such as Hadoop.<\/p><blockquote><div>&#13;\n<p>It is about hundreds of thousands of data points and not millions, it is about data sets that fit into the memory on a reasonably specified laptop.<\/p>&#13;\n<\/div><\/blockquote><p>This is where the majority of my work is done. Clojure provides an approach to solving this problem with\u00a0<a href=\"http:\/\/clojure.org\/reference\/reducers\">Reducers<\/a>.\u00a0Reducers provide implementations of\u00a0reducible collections\u00a0(a collection that can reduce itself across several machine cores) together with means of manipulating\u00a0reducing functions. What this means in practice is that you build up a sequence of computations which are then calculated simultaneously across your data. To address the problem of getting data into a reducible collection in the first place,\u00a0<a href=\"https:\/\/github.com\/thebusby\/iota\">Iota<\/a>\u00a0provides support for efficiently creating reducible collections from large files.<\/p><p>More recently Clojure has introduced\u00a0<a href=\"http:\/\/clojure.org\/reference\/transducers\">Transducers<\/a>, which separate the implementation of the reducing function from the collection to which they are applied. I gave a talk at London's Clojure eXchange on\u00a0<a href=\"http:\/\/henrygarner.com\/clojure\/data-science\/statistics\/transducers\/reducers\/talks\/2015\/12\/05\/expressive-parallel-analytics-clojure.html\">Expressive Parallel Analytics<\/a>\u00a0with transducers. The library\u00a0<a href=\"https:\/\/github.com\/mastodonc\/kixi.stats\">kixi.stats<\/a>\u00a0is the result of that talk: a library of statistical reducing functions. Each of the functions in\u00a0kixi.stats\u00a0will calculate its output in a single pass over the input data. Such algorithms are typically called streaming algorithms and are celebrated for their ability to produce approximate solutions in linear time.<\/p><p>Since reducing functions are\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Pure_function\">pure functions<\/a>, they can be combined together using functional composition. In this way, multiple statistics (such as the mean, standard deviation, skewness, etc) can all be calculated in a\u00a0<a href=\"http:\/\/clojuredatascience.com\/posts\/2015-12-05-expressive-parallel-analytics-with-clojure.html\">single pass over the data sequence<\/a>. Libraries such as\u00a0<a href=\"https:\/\/github.com\/cgrand\/xforms\">xforms<\/a>\u00a0and\u00a0<a href=\"https:\/\/github.com\/henrygarner\/redux\">redux<\/a>\u00a0provide reducing function combinators and will work with any reducing function, including the ones you write yourself.<\/p><p>The implementation of\u00a0kixi.stats\u00a0was heavily influenced by\u00a0<a href=\"https:\/\/github.com\/aphyr\/tesser\">Tesser<\/a>, a Clojure library which also provides implementations of streaming algorithms in such a way as they can either be run locally or on Hadoop.<\/p><h2>Code as data<\/h2><p>One of Clojure's most useful properties isn't a property of Clojure specifically, but of the Lisp language family of which it is a member. Lisp elicits high praise as the\u00a0<a href=\"http:\/\/www.eecs.wsu.edu\/~hauser\/teaching\/Arch-F07\/handouts\/jonesforth.s.txt\">ultimate high-level language<\/a>\u00a0or\u00a0<a href=\"http:\/\/onsmalltalk.com\/languages-of-the-gods\">language of the Gods<\/a>. It's the\u00a0<a href=\"https:\/\/en.wikiquote.org\/wiki\/Lisp_(programming_language)\">programmable programming language<\/a>\u00a0that was\u00a0<a href=\"http:\/\/www.paulgraham.com\/rootsoflisp.html\">discovered rather than invented<\/a>. And it inspires tributes from great computer scientists:<\/p><blockquote><div>&#13;\n<p>If you give someone Fortran, he has Fortran. If you give someone Lisp, he has any language he choses. -- Guy Steele<\/p>&#13;\n&#13;\n<p>Lisp is not a language, it's a building material. -- Alan Kay<\/p>&#13;\n<\/div><\/blockquote><p>These qualities of Lisp (and also of Clojure) can be the hardest to convey to programmers from other languages. Whilst I understand where the authors above are coming from, these statements are as alienating to newcomers as the suggestion that novice data scientists learn multivariate calculus before they run their first A\/B test.<\/p><p>Instead, I'll simply provide some examples about why having a\u00a0programmable programming language\u00a0is an asset.\u00a0<a href=\"http:\/\/blog.cognitect.com\/blog\/2016\/6\/28\/the-new-normal-data-leverage\">Michael Nygard<\/a>\u00a0again:<\/p><blockquote><div>&#13;\n<p>With Clojure, it\u2019s very easy to support multiple versions of our API. In fact, it so easy, we can even have an API about APIs. Now, our data-orientation approach takes everything up a level. By using data rather than classes or annotations we give ourselves the ability to \u201cgo meta\u201d [...]. If we wanted to do that in Java, we would need to build some kind of a new language that emitted Java as its bytecode. We don\u2019t need to do that with Clojure.<\/p>&#13;\n<\/div><\/blockquote><p>This capability is really the natural extension of Clojure's succinct way of representing data and its consistent and unambiguous syntax. Clojure code itself is data, expressed as lists of symbols which can be manipulated by those same data-manipulating functions. Through a language construct called a\u00a0<a href=\"http:\/\/clojure.org\/reference\/macros\">macro<\/a>, Clojure allows programmers to extend the language itself. Functions accept values and return new values. Macros accept code and return new code.<\/p><p>In practice, this is used by libraries such as\u00a0Tesser\u00a0to provide succinct ways of expressing transforms. Within the library is a macro called\u00a0deftransform\u00a0which allows for the elimination of boilerplate code for specifying transforms. Below is the implementation of\u00a0map:<\/p><div><pre>&#13;\n(deftransform map&#13;\n  \"Takes a function `f` and an optional fold. Returns a version of the fold&#13;\n  which finally calls (f input) to transform each element.&#13;\n      (-&gt;&gt; (t\/map inc) (t\/into []) (t\/Tesser [[1 2] [3 4]]))&#13;\n      ; =&gt; [2 3 4 5]\"&#13;\n  [f]&#13;\n  (assoc downstream :reducer (fn reducer [acc x]&#13;\n                               (reducer- acc (f x)))))<\/pre><\/div><p>The code which remains embodies the essence of a\u00a0map\u00a0transform in Tesser: it associates a downstream reducer function which calls\u00a0(f x). This internal representation enables Tesser to take the same computation and target either a local or Hadoop-based context. Incidentally, the author of Tesser has written a\u00a0<a href=\"https:\/\/aphyr.com\/posts\/305-clojure-from-the-ground-up-macros\">series of articles on learning Clojure, including macros<\/a>.<\/p><p>Macros aren't just used by advanced Clojurians to express new concepts. They enable you (even as a Clojure beginner) to get more done. For example, the\u00a0<a href=\"http:\/\/clojure.org\/guides\/threading_macros\">threading macros<\/a>\u00a0make code more legible by removing nested expressions. So:<\/p><div><pre>&#13;\n(reduce + (take 10 (filter even? (map square (range)))))&#13;\n<\/pre><\/div><p>with the thread-last macro\u00a0-&gt;&gt;\u00a0becomes:<\/p><div><pre>&#13;\n(-&gt;&gt; (range)&#13;\n     (map square)&#13;\n     (filter even?)&#13;\n     (take 10)&#13;\n     (reduce +))<\/pre><\/div><p>Whilst compiling to identical code as the previous version, the latter expresses the sequence of expressions more clearly.\u00a0<a href=\"https:\/\/twitter.com\/nonrecursive\">Daniel Higginbotham<\/a>\u00a0puts it like this in\u00a0<a href=\"http:\/\/www.braveclojure.com\/writing-macros\/\">Brave Clojure<\/a>:<\/p><blockquote><div>&#13;\n<p>Yes, macros are cooler than a polar bear\u2019s toenails, but you shouldn\u2019t think of macros as some esoteric tool you pull out when you feel like getting extra fancy with your code. In fact, macros allow Clojure to derive a lot of its built-in functionality from a tiny core of functions and special forms.<\/p>&#13;\n<\/div><\/blockquote><p>Macros drive some commentators into paroxysms of delight, but don't let that put you off. An appropriate use of macros will make your code easier, rather than harder, to comprehend.<\/p><h2>Learning to learn<\/h2><p>Learning a powerful language such as Clojure together with a technical discipline such as data science is an ambitious task. But don't be disheartened! U.S. Government data hacker\u00a0<a href=\"https:\/\/twitter.com\/bendystraw\/status\/768233651599446021\">Becky Sweger<\/a>\u00a0has said:<\/p><blockquote><div>&#13;\n<p>My best technical skill isn't coding. It's a willingness to ask questions, in front of everyone, about what I don't understand.<\/p>&#13;\n<\/div><\/blockquote><p>Besides, just by coming to data science via programming you bring a wealth of practical skills, borne of experience, which are very relevant. For example:<\/p><ul><li>User stories. You're used to decomposing large problems into smaller steps which deliver incremental, measurable business value. You know it's foolish to embark on a project without a clear statement of the desired outcomes and priorities first, even though the implementation may be unclear.<\/li>\n\t<li>Version control. You know you'll inevitably be required to repeat statistical analysis at some point in the future. Luckily you've already committed your code to version control in a branch specifying its intent. You'll also be able to point anyone who wants to scrutinise your approach at this code so they can replicate your methodology.<\/li>\n\t<li>Kanban boards. You track in-flight work to provide a context for discussions about progress, and archive completed work to remind yourself how far you've come (and the results of your attempts, whether they were successful or not).<\/li>\n\t<li>Devops. You're mindful when there's a need to be able to move algorithms into production. You can communicate with development and operations teams about how your model can be integrated into the live site in ways that they can understand.<\/li>\n<\/ul><p>Take heart. The diversity of your experience as a developer moving into data science is a tremendous asset.<\/p><h2>In summary<\/h2><p>Data scientists are expected to have a wide and deep level of technical understanding. The Clojure programming language enjoys an enormously broad array of applications and ClojureScript extends its reach even further.<\/p><p>If your data science project requires you to apply an off-the-shelf algorithm and move on, Clojure is not the language I would recommend. If instead you're looking to build a strong foundation in statistical thinking, Clojure's language, tooling, and community will repay your investment many times over. This approach will enable you to effectively apply fundamental principles in the diverse situations you will encounter as a practicing data scientist.<\/p><p>I quoted a line from Alexander Pope's\u00a0Essay on Criticism\u00a0at the beginning of this article. The full couplet is:<\/p><blockquote><div>&#13;\n<p>A little learning is a dangerous thing \/ Drink deep, or taste not the Pierian spring<\/p>&#13;\n<\/div><\/blockquote><p>The Pierian spring was the\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Pierian_Spring\">metaphorical source of knowledge of art and science<\/a>. If you're seeking to learn powerful analytical techniques, do so with a language that will move up and down the ladder of abstraction with you. This conceptual grounding will serve you well even as the technical options\u2014and perhaps even your choice of programming language itself\u2014continue to evolve in the future.<\/p><p><em>Thanks to Luke Snape for reviewing preliminary drafts of this article.<\/em><\/p><h2>About the author<\/h2><p><strong><a href=\"https:\/\/twitter.com\/henrygarner\"><img alt=\"\" src=\"https:\/\/www.infoq.com\/articles\/articles\/data-science-abstraction\/en\/resources\/HENRY.jpg\" \/>Henry Garner<\/a><\/strong>\u00a0is a freelance data scientist working primarily in Clojure. He\u2019s author of the Packt book\u00a0<a href=\"https:\/\/www.amazon.com\/Clojure-Data-Science-Henry-Garner\/dp\/1784397180\/\">Clojure for Data Science<\/a>\u00a0and he managed to squeeze the buzzwords\u00a0big data\u00a0and\u00a0machine learning\u00a0onto the cover. And also into this biography.<\/p><p><em>Data science is fast becoming a critical skill for developers and managers across industries, and it looks like a lot of fun as well. But it\u2019s pretty complicated - there are a lot of engineering and analytical options to navigate, and it\u2019s hard to know if you\u2019re doing it right or where the bear traps lie. In this series we explore ways of\u00a0making sense of data science - understanding where it\u2019s needed and where it\u2019s not, and how to make it an asset for you, from people who\u2019ve been there and done it.<\/em><\/p><p><em>This InfoQ article is part of the series \"<a href=\"https:\/\/www.infoq.com\/handle-data-science\" target=\"_blank\">Getting A Handle on Data Science<\/a>\" . You can subscribe to receive notifications via\u00a0<a href=\"https:\/\/www.infoq.com\/feed\/handle-data-science\" target=\"_blank\">RSS<\/a>.<\/em><\/p><div id=\"contentRatingWidget\"><div class=\"rate-article rate-section\"><div class=\"rate-article-header\"><p class=\"title\">Rate this Article<\/p><\/div><div class=\"rate-article-content clearfix\"><p><\/p><fieldset id=\"relevance_fieldset\" class=\"rating\"><legend>Adoption Stage<\/legend>\n                <label for=\"relevance-star5\" title=\"Innovator\"><\/label>\n                <label for=\"relevance-star4\" title=\"Early adopter\"><\/label>\n                <label for=\"relevance-star3\" title=\"Early majority\"><\/label>\n                <label for=\"relevance-star2\" title=\"Late majority\"><\/label>\n                <label for=\"relevance-star1\" title=\"Laggards\"><\/label>\n            <\/fieldset><p><\/p><fieldset id=\"style_fieldset\" class=\"rating\"><legend class=\"legend-style\">Style<\/legend>\n                <label for=\"style-star5\" title=\"Exceptional\"><\/label>\n                <label for=\"style-star4\" title=\"Good\"><\/label>\n                <label for=\"style-star3\" title=\"Average\"><\/label>\n                <label for=\"style-star2\" title=\"Acceptable\"><\/label>\n                <label for=\"style-star1\" title=\"Poor\"><\/label>\n            <\/fieldset><\/div><\/div><\/div>","created_at":"2018-11-07T16:44:04+0000","updated_at":"2018-11-07T16:44:04+0000","published_at":"2016-12-02T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":26,"domain_name":"www.infoq.com","preview_picture":"https:\/\/res.infoq.com\/articles\/data-science-abstraction\/en\/smallimage\/logo-clojure.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":307,"title":"My Python Development Environment, 2018 Edition \u00ab Jacob Kaplan-Moss","url":"https:\/\/jacobian.org\/writing\/python-environment-2018\/","content":"<p>For years I\u2019ve noodled around with various setups for a Python development\nenvironment, and never really found something I loved \u2013 until\u00a0now.<\/p><p>My setup pieces together <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a>, <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a>, and <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a>. It\u2019s probably a tad more\ncomplex that is ideal for most Python users, but for the things I need, it\u2019s\u00a0perfect.<\/p><div class=\"section\" id=\"my-requirements\"><h2>My\u00a0Requirements<\/h2><p>I do have somewhat specific (maybe unusual?)\u00a0requirements:<\/p><ul class=\"simple\"><li>I need to develop against multiple Python versions, including 2.7, various\nPython 3 versions (3.5 and 3.6, mostly), and\u00a0PyPy.<\/li>\n<li>I work on many projects simultaneously, each with different sets of\ndependencies, so some sort of virtual environment or isolation is\u00a0critical.<\/li>\n<li>I use multiple OSes: macOS at work, and Linux (well, Linux-ish - actually it\u2019s\n<a class=\"reference external\" href=\"https:\/\/docs.microsoft.com\/en-us\/windows\/wsl\/about\">WSL<\/a>) at\u00a0home.<\/li>\n<li>I want to avoid using the System-provided Python. On macOS it\u2019s too outdated.\nOn Linux, the system Python is used by the OS itself, so if you hose your\nPython you can hose your\u00a0system.<\/li>\n<li>I use a bunch Python-based CLI stuff, like <a class=\"reference external\" href=\"https:\/\/rg3.github.io\/youtube-dl\/\">youtube-dl<\/a>, <a class=\"reference external\" href=\"https:\/\/aws.amazon.com\/cli\/\">awscli<\/a>, <a class=\"reference external\" href=\"https:\/\/doc2dash.readthedocs.io\/en\/stable\/\">doc2dash<\/a>,\netc. I want to be able to install and use them without fussing around with\nactivating environments, but I also don\u2019t want their dependencies to clutter\nup a global\u00a0installation.<\/li>\n<li>I deploy stuff mostly to Heroku (personal) and <a class=\"reference external\" href=\"https:\/\/cloud.gov\/\">cloud.gov<\/a> (work). I expect this\nnot to change: I\u2019m spoiled, and never want to manage my own infrastructure\u00a0again.<\/li>\n<li>Although Docker meets all these requirements, I don\u2019t really like using it.\nI find it slow, frustrating, and overkill for my\u00a0purposes.<\/li>\n<\/ul><\/div><div class=\"section\" id=\"the-setup\"><h2>The\u00a0Setup<\/h2><div class=\"section\" id=\"id1\"><h3>1. <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a><\/h3><p><strong>Why?<\/strong> I need to run multiple Python versions, isolated from the system\nPython. <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a> makes it easy to install, manage, and switch between those\nmultiple Pythons. As a bonus, <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a> integrates with <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a> and will\nautomatically install missing Python versions if they\u2019re required by a\n<code class=\"docutils literal\">Pipfile<\/code>.<\/p><p>On my Mac, I installed <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a> from Homebrew (<code class=\"docutils literal\">brew install pyenv<\/code>). On Linux,\nI used the Github installation technique documented in the\n<a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv#installation\">installation instructions<\/a>,\nwhich was easy and went\u00a0smoothly.<\/p><p>Then, I installed some Python\u00a0versions:<\/p><pre class=\"literal-block\">pyenv install 3.6.4\npyenv install 3.5.4\npyenv install 2.7.14\npyenv install pypy3.5-5.10.0\n<\/pre><p>And made sure my default Python was set to the latest and\u00a0greatest:<\/p><pre class=\"literal-block\">pyenv global 3.6.4\n<\/pre><\/div><div class=\"section\" id=\"id2\"><h3>2. <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a><\/h3><p><strong>Why?<\/strong> <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> lets me install Python-based CLI stuff (like <a class=\"reference external\" href=\"https:\/\/rg3.github.io\/youtube-dl\/\">youtube-dl<\/a>,\n<a class=\"reference external\" href=\"https:\/\/aws.amazon.com\/cli\/\">awscli<\/a>, <a class=\"reference external\" href=\"https:\/\/doc2dash.readthedocs.io\/en\/stable\/\">doc2dash<\/a>, etc.) without those projects\u2019 dependencies messing up my\nglobal\u00a0Python.<\/p><p>Normally, installing <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> is\u00a0easy:<\/p><pre class=\"literal-block\">curl https:\/\/raw.githubusercontent.com\/mitsuhiko\/pipsi\/master\/get-pipsi.py | python\n<\/pre><p>However, at the time I\u2019m writing this (late February 2018), this didn\u2019t work\nfor me; I needed to do some tweaking to get it to work. I followed the workarounds\ndocumented in issues <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\/issues\/124\">#124<\/a> and\n<a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\/issues\/125\">#125<\/a>. Hopefully this\u2019ll be fixed\u00a0shortly.<\/p><p>I didn\u2019t do this, but it might be a good idea to install <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> using the system\nPython (i.e. run <code class=\"docutils literal\">pyenv local system<\/code> before running the installer). The\ndownside is that this will use a potentially old-ass Python, but the upside is\nthat you won\u2019t break your <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> install if you delete a pyenv version (like\nupon an\u00a0upgrade).<\/p><\/div><div class=\"section\" id=\"id5\"><h3>3. <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a><\/h3><p><strong>Why?<\/strong> <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a> handles dependency- and virtual-environment-management in a\nway that\u2019s very intuitive (to me), and fits perfectly with my desired\u00a0workflow.<\/p><p>The documentation covers a\n<a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/install\/#installing-pipenv\">few different ways to install pipenv<\/a>.\nBecause I\u2019m already using <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a>, I chose the\n<a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/install\/#fancy-installation-of-pipenv\">pipsi-based installation<\/a>:<\/p><pre class=\"literal-block\">pipsi install pew\npipsi install pipenv\n<\/pre><\/div><\/div><div class=\"section\" id=\"what-it-looks-like-in-use\"><h2>What it looks like in\u00a0use<\/h2><p>With this all together, all my use-cases are handled\u00a0simply:<\/p><p>To start new projects, I just make a directory and type <code class=\"docutils literal\">pipenv install ...<\/code>\nto start installing my dependencies. Pipenv creates a <code class=\"docutils literal\">Pipfile<\/code> for me, and\nmanages it, and I\u2019m up and\u00a0running.<\/p><p>To work on existing projects, I clone a repository and either run <code class=\"docutils literal\">pipenv\ninstall<\/code> (for projects that already have a <code class=\"docutils literal\">Pipfile<\/code>), or <code class=\"docutils literal\">pipenv install\n-r requirements.txt<\/code> (which as a side-effect automatically converts a the\nrequirements file to a <code class=\"docutils literal\">Pipfile<\/code>).<\/p><p>If I need to switch Python versions, I run <code class=\"docutils literal\">pyenv local &lt;version&gt;<\/code> in my\nproject directory. I can also\u00a0add:<\/p><pre class=\"literal-block\">[requires]\npython_version = \"&lt;version&gt;\"\n<\/pre><p>to my <code class=\"docutils literal\">Pipfile<\/code>, and <code class=\"docutils literal\">pipenv<\/code> will enforce that version\u00a0requirement.<\/p><p>When I want to install CLI stuff, I use\u00a0pipsi:<\/p><pre class=\"literal-block\">pipsi install awscli\npipsi install doc2dash\n# ... etc\n<\/pre><p>When it comes time to deploy, both Heroku and <a class=\"reference external\" href=\"https:\/\/cloud.gov\/\">cloud.gov<\/a> will read and\nunderstand my <code class=\"docutils literal\">Pipfile<\/code>. If I need to deploy to something that doesn\u2019t\ndo <code class=\"docutils literal\">Pipfile<\/code>-based installs, I create a <code class=\"docutils literal\">requirements.txt<\/code> by running\n<code class=\"docutils literal\">pipenv lock --requirements<\/code>.<\/p><\/div>","created_at":"2018-11-07T16:44:04+0000","updated_at":"2018-11-07T16:44:04+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"jacobian.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":308,"title":"My Python Development Environment, 2018 Edition \u00ab Jacob Kaplan-Moss","url":"https:\/\/jacobian.org\/writing\/python-environment-2018\/","content":"<p>For years I\u2019ve noodled around with various setups for a Python development\nenvironment, and never really found something I loved \u2013 until\u00a0now.<\/p><p>My setup pieces together <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a>, <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a>, and <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a>. It\u2019s probably a tad more\ncomplex that is ideal for most Python users, but for the things I need, it\u2019s\u00a0perfect.<\/p><div class=\"section\"><h2>My\u00a0Requirements<\/h2><p>I do have somewhat specific (maybe unusual?)\u00a0requirements:<\/p><ul class=\"simple\"><li>I need to develop against multiple Python versions, including 2.7, various\nPython 3 versions (3.5 and 3.6, mostly), and\u00a0PyPy.<\/li>\n<li>I work on many projects simultaneously, each with different sets of\ndependencies, so some sort of virtual environment or isolation is\u00a0critical.<\/li>\n<li>I use multiple OSes: macOS at work, and Linux (well, Linux-ish - actually it\u2019s\n<a class=\"reference external\" href=\"https:\/\/docs.microsoft.com\/en-us\/windows\/wsl\/about\">WSL<\/a>) at\u00a0home.<\/li>\n<li>I want to avoid using the System-provided Python. On macOS it\u2019s too outdated.\nOn Linux, the system Python is used by the OS itself, so if you hose your\nPython you can hose your\u00a0system.<\/li>\n<li>I use a bunch Python-based CLI stuff, like <a class=\"reference external\" href=\"https:\/\/rg3.github.io\/youtube-dl\/\">youtube-dl<\/a>, <a class=\"reference external\" href=\"https:\/\/aws.amazon.com\/cli\/\">awscli<\/a>, <a class=\"reference external\" href=\"https:\/\/doc2dash.readthedocs.io\/en\/stable\/\">doc2dash<\/a>,\netc. I want to be able to install and use them without fussing around with\nactivating environments, but I also don\u2019t want their dependencies to clutter\nup a global\u00a0installation.<\/li>\n<li>I deploy stuff mostly to Heroku (personal) and <a class=\"reference external\" href=\"https:\/\/cloud.gov\/\">cloud.gov<\/a> (work). I expect this\nnot to change: I\u2019m spoiled, and never want to manage my own infrastructure\u00a0again.<\/li>\n<li>Although Docker meets all these requirements, I don\u2019t really like using it.\nI find it slow, frustrating, and overkill for my\u00a0purposes.<\/li>\n<\/ul><\/div><div class=\"section\"><h2>The\u00a0Setup<\/h2><div class=\"section\"><h3>1. <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a><\/h3><p><strong>Why?<\/strong> I need to run multiple Python versions, isolated from the system\nPython. <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a> makes it easy to install, manage, and switch between those\nmultiple Pythons. As a bonus, <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a> integrates with <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a> and will\nautomatically install missing Python versions if they\u2019re required by a\n<code class=\"docutils literal\">Pipfile<\/code>.<\/p><p>On my Mac, I installed <a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv\">pyenv<\/a> from Homebrew (<code class=\"docutils literal\">brew install pyenv<\/code>). On Linux,\nI used the Github installation technique documented in the\n<a class=\"reference external\" href=\"https:\/\/github.com\/pyenv\/pyenv#installation\">installation instructions<\/a>,\nwhich was easy and went\u00a0smoothly.<\/p><p>Then, I installed some Python\u00a0versions:<\/p><pre class=\"literal-block\">pyenv install 3.6.4\npyenv install 3.5.4\npyenv install 2.7.14\npyenv install pypy3.5-5.10.0\n<\/pre><p>And made sure my default Python was set to the latest and\u00a0greatest:<\/p><pre class=\"literal-block\">pyenv global 3.6.4\n<\/pre><\/div><div class=\"section\"><h3>2. <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a><\/h3><p><strong>Why?<\/strong> <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> lets me install Python-based CLI stuff (like <a class=\"reference external\" href=\"https:\/\/rg3.github.io\/youtube-dl\/\">youtube-dl<\/a>,\n<a class=\"reference external\" href=\"https:\/\/aws.amazon.com\/cli\/\">awscli<\/a>, <a class=\"reference external\" href=\"https:\/\/doc2dash.readthedocs.io\/en\/stable\/\">doc2dash<\/a>, etc.) without those projects\u2019 dependencies messing up my\nglobal\u00a0Python.<\/p><p>Normally, installing <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> is\u00a0easy:<\/p><pre class=\"literal-block\">curl https:\/\/raw.githubusercontent.com\/mitsuhiko\/pipsi\/master\/get-pipsi.py | python\n<\/pre><p>However, at the time I\u2019m writing this (late February 2018), this didn\u2019t work\nfor me; I needed to do some tweaking to get it to work. I followed the workarounds\ndocumented in issues <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\/issues\/124\">#124<\/a> and\n<a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\/issues\/125\">#125<\/a>. Hopefully this\u2019ll be fixed\u00a0shortly.<\/p><p>I didn\u2019t do this, but it might be a good idea to install <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> using the system\nPython (i.e. run <code class=\"docutils literal\">pyenv local system<\/code> before running the installer). The\ndownside is that this will use a potentially old-ass Python, but the upside is\nthat you won\u2019t break your <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a> install if you delete a pyenv version (like\nupon an\u00a0upgrade).<\/p><\/div><div class=\"section\"><h3>3. <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a><\/h3><p><strong>Why?<\/strong> <a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/\">pipenv<\/a> handles dependency- and virtual-environment-management in a\nway that\u2019s very intuitive (to me), and fits perfectly with my desired\u00a0workflow.<\/p><p>The documentation covers a\n<a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/install\/#installing-pipenv\">few different ways to install pipenv<\/a>.\nBecause I\u2019m already using <a class=\"reference external\" href=\"https:\/\/github.com\/mitsuhiko\/pipsi\">pipsi<\/a>, I chose the\n<a class=\"reference external\" href=\"https:\/\/docs.pipenv.org\/install\/#fancy-installation-of-pipenv\">pipsi-based installation<\/a>:<\/p><pre class=\"literal-block\">pipsi install pew\npipsi install pipenv\n<\/pre><\/div><\/div><div class=\"section\"><h2>What it looks like in\u00a0use<\/h2><p>With this all together, all my use-cases are handled\u00a0simply:<\/p><p>To start new projects, I just make a directory and type <code class=\"docutils literal\">pipenv install ...<\/code>\nto start installing my dependencies. Pipenv creates a <code class=\"docutils literal\">Pipfile<\/code> for me, and\nmanages it, and I\u2019m up and\u00a0running.<\/p><p>To work on existing projects, I clone a repository and either run <code class=\"docutils literal\">pipenv\ninstall<\/code> (for projects that already have a <code class=\"docutils literal\">Pipfile<\/code>), or <code class=\"docutils literal\">pipenv install\n-r requirements.txt<\/code> (which as a side-effect automatically converts a the\nrequirements file to a <code class=\"docutils literal\">Pipfile<\/code>).<\/p><p>If I need to switch Python versions, I run <code class=\"docutils literal\">pyenv local &lt;version&gt;<\/code> in my\nproject directory. I can also\u00a0add:<\/p><pre class=\"literal-block\">[requires]\npython_version = \"&lt;version&gt;\"\n<\/pre><p>to my <code class=\"docutils literal\">Pipfile<\/code>, and <code class=\"docutils literal\">pipenv<\/code> will enforce that version\u00a0requirement.<\/p><p>When I want to install CLI stuff, I use\u00a0pipsi:<\/p><pre class=\"literal-block\">pipsi install awscli\npipsi install doc2dash\n# ... etc\n<\/pre><p>When it comes time to deploy, both Heroku and <a class=\"reference external\" href=\"https:\/\/cloud.gov\/\">cloud.gov<\/a> will read and\nunderstand my <code class=\"docutils literal\">Pipfile<\/code>. If I need to deploy to something that doesn\u2019t\ndo <code class=\"docutils literal\">Pipfile<\/code>-based installs, I create a <code class=\"docutils literal\">requirements.txt<\/code> by running\n<code class=\"docutils literal\">pipenv lock --requirements<\/code>.<\/p><\/div>","created_at":"2018-11-07T16:44:04+0000","updated_at":"2018-11-07T16:44:04+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"jacobian.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":303,"title":"clojure\/test.check","url":"https:\/\/github.com\/clojure\/test.check","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><em>test.check<\/em> is a Clojure property-based testing tool inspired by QuickCheck.\nThe core idea of <em>test.check<\/em> is that instead of enumerating expected input\nand output for unit tests, you write properties about your function that should\nhold true for all inputs. This lets you write concise, powerful tests.<\/p>\n<ul><li><a href=\"#releases-and-dependency-information\">Releases and Dependency Information<\/a><\/li>\n<li><a href=\"#documentation\">Documentation<\/a><\/li>\n<li><a href=\"#developer-information\">Developer Information<\/a><\/li>\n<li><a href=\"#see-also\">See also<\/a><\/li>\n<li><a href=\"#yourkit\">YourKit<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2><a id=\"user-content-releases-and-dependency-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#releases-and-dependency-information\"><\/a>Releases and Dependency Information<\/h2>\n<p>Release notes for each version are available in\n<a href=\"https:\/\/github.com\/clojure\/test.check\/blob\/master\/CHANGELOG.markdown\"><code>CHANGELOG.markdown<\/code><\/a>. Remember that prior to version\n0.5.7, <em>test.check<\/em> was called <em>simple-check<\/em>.<\/p>\n<p>As of version <code>0.9.0<\/code>, test.check requires Clojure &gt;= <code>1.7.0<\/code>.<\/p>\n<p>Please note a\n<a href=\"https:\/\/github.com\/clojure\/test.check\/blob\/master\/CHANGELOG.markdown#080\">breaking change for ClojureScript<\/a>\nin the <code>0.8.*<\/code> releases.<\/p>\n<h3><a id=\"user-content-latest-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#latest-version\"><\/a>Latest Version<\/h3>\n<h4><a id=\"user-content-leiningen\" class=\"anchor\" aria-hidden=\"true\" href=\"#leiningen\"><\/a>Leiningen<\/h4>\n<div class=\"highlight highlight-source-clojure\"><pre>[org.clojure\/test.check \"0.10.0-alpha3\"]<\/pre><\/div>\n<h4><a id=\"user-content-maven\" class=\"anchor\" aria-hidden=\"true\" href=\"#maven\"><\/a>Maven<\/h4>\n<div class=\"highlight highlight-text-xml\"><pre>&lt;dependency&gt;\n  &lt;groupId&gt;org.clojure&lt;\/groupId&gt;\n  &lt;artifactId&gt;test.check&lt;\/artifactId&gt;\n  &lt;version&gt;0.10.0-alpha3&lt;\/version&gt;\n&lt;\/dependency&gt;<\/pre><\/div>\n<h3><a id=\"user-content-stable-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#stable-version\"><\/a>Stable Version<\/h3>\n<h4><a id=\"user-content-leiningen-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#leiningen-1\"><\/a>Leiningen<\/h4>\n<div class=\"highlight highlight-source-clojure\"><pre>[org.clojure\/test.check \"0.9.0\"]<\/pre><\/div>\n<h4><a id=\"user-content-maven-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#maven-1\"><\/a>Maven<\/h4>\n<div class=\"highlight highlight-text-xml\"><pre>&lt;dependency&gt;\n  &lt;groupId&gt;org.clojure&lt;\/groupId&gt;\n  &lt;artifactId&gt;test.check&lt;\/artifactId&gt;\n  &lt;version&gt;0.9.0&lt;\/version&gt;\n&lt;\/dependency&gt;<\/pre><\/div>\n<p>If you'd like to try a SNAPSHOT version, <a href=\"http:\/\/dev.clojure.org\/display\/community\/Maven+Settings+and+Repositories\" rel=\"nofollow\">add the sonatype repository to your\nproject<\/a>.<\/p>\n<h2><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h2>\n<ul><li><a href=\"http:\/\/clojure.github.io\/test.check\/\" rel=\"nofollow\">API Docs<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/clojure\/test.check\/blob\/master\/doc\/cheatsheet.md\">Cheatsheet<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/clojure\/test.check\/blob\/master\/doc\/intro.md\">Generator writing guide<\/a><\/li>\n<li>Examples (some of these may refer to simple-check):\n<\/li>\n<li>Blog posts and videos (some of these may refer to simple-check):\n<ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=JMhNINPo__g\" rel=\"nofollow\">Powerful Testing with test.check - Clojure\/West 2014<\/a> -- <a href=\"https:\/\/speakerdeck.com\/reiddraper\/powerful-testing-with-test-dot-check\" rel=\"nofollow\">Slides<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=F4VZPxLZUdA\" rel=\"nofollow\">Building test.check Generators - Clojure\/Conj 2017<\/a> - <a href=\"https:\/\/gfredericks.com\/speaking\/2017-10-12-generators.pdf\" rel=\"nofollow\">Slides<\/a><\/li>\n<li><a href=\"http:\/\/blog.8thlight.com\/connor-mendenhall\/2013\/10\/31\/check-your-work.html\" rel=\"nofollow\">Check your work - 8th Light<\/a><\/li>\n<li><a href=\"http:\/\/reiddraper.com\/writing-simple-check\/\" rel=\"nofollow\">Writing simple-check - Reid Draper<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=u0TkAw8QqrQ\" rel=\"nofollow\">Generative testing in Clojure - Youtube<\/a><\/li>\n<li><a href=\"http:\/\/curtis.io\/using-simple-check-with-expectations\" rel=\"nofollow\">Using simple-check with Expectations - Curtis Gagliardi<\/a><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-useful-libraries\" class=\"anchor\" aria-hidden=\"true\" href=\"#useful-libraries\"><\/a>Useful libraries<\/h2>\n<h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<p>Let's say we're testing a sort function. We want to check that that our sort\nfunction is idempotent, that is, applying sort twice should be equivalent to\napplying it once: <code>(= (sort a) (sort (sort a)))<\/code>. Let's write a quick test to\nmake sure this is the case:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[clojure.test.check :as tc])\n(require '[clojure.test.check.generators :as gen])\n(require '[clojure.test.check.properties :as prop #?@(:cljs [:include-macros true])])\n(def sort-idempotent-prop\n  (prop\/for-all [v (gen\/vector gen\/int)]\n    (= (sort v) (sort (sort v)))))\n(tc\/quick-check 100 sort-idempotent-prop)\n;; =&gt; {:result true,\n;; =&gt;  :pass? true,\n;; =&gt;  :num-tests 100,\n;; =&gt;  :time-elapsed-ms 28,\n;; =&gt;  :seed 1528580707376}<\/pre><\/div>\n<p>In prose, this test reads: for all vectors of integers, <code>v<\/code>, sorting <code>v<\/code> is\nequal to sorting <code>v<\/code> twice.<\/p>\n<p>What happens if our test fails? <em>test.check<\/em> will try and find 'smaller'\ninputs that still fail. This process is called shrinking. Let's see it in\naction:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def prop-sorted-first-less-than-last\n  (prop\/for-all [v (gen\/not-empty (gen\/vector gen\/int))]\n    (let [s (sort v)]\n      (&lt; (first s) (last s)))))\n(tc\/quick-check 100 prop-sorted-first-less-than-last)\n;; =&gt; {:num-tests 5,\n;; =&gt;  :seed 1528580863556,\n;; =&gt;  :fail [[-3]],\n;; =&gt;  :failed-after-ms 1,\n;; =&gt;  :result false,\n;; =&gt;  :result-data nil,\n;; =&gt;  :failing-size 4,\n;; =&gt;  :pass? false,\n;; =&gt;  :shrunk\n;; =&gt;  {:total-nodes-visited 5,\n;; =&gt;   :depth 2,\n;; =&gt;   :pass? false,\n;; =&gt;   :result false,\n;; =&gt;   :result-data nil,\n;; =&gt;   :time-shrinking-ms 1,\n;; =&gt;   :smallest [[0]]}}<\/pre><\/div>\n<p>This test claims that the first element of a sorted vector should be less-than\nthe last. Of course, this isn't true: the test fails with input <code>[-3]<\/code>, which\ngets shrunk down to <code>[0]<\/code>, as seen in the output above. As your test functions\nrequire more sophisticated input, shrinking becomes critical to being able\nto understand exactly why a random test failed. To see how powerful shrinking\nis, let's come up with a contrived example: a function that fails if it's\npassed a sequence that contains the number 42:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def prop-no-42\n  (prop\/for-all [v (gen\/vector gen\/int)]\n    (not (some #{42} v))))\n(tc\/quick-check 100 prop-no-42)\n;; =&gt; {:num-tests 45,\n;; =&gt;  :seed 1528580964834,\n;; =&gt;  :fail\n;; =&gt;  [[-35 -9 -31 12 -30 -40 36 36 25 -2 -31 42 8 31 17 -19 3 -15 44 -1 -8 27 16]],\n;; =&gt;  :failed-after-ms 11,\n;; =&gt;  :result false,\n;; =&gt;  :result-data nil,\n;; =&gt;  :failing-size 44,\n;; =&gt;  :pass? false,\n;; =&gt;  :shrunk\n;; =&gt;  {:total-nodes-visited 16,\n;; =&gt;   :depth 5,\n;; =&gt;   :pass? false,\n;; =&gt;   :result false,\n;; =&gt;   :result-data nil,\n;; =&gt;   :time-shrinking-ms 1,\n;; =&gt;   :smallest [[42]]}}<\/pre><\/div>\n<p>We see that the test failed on a rather large vector, as seen in the <code>:fail<\/code>\nkey. But then <em>test.check<\/em> was able to shrink the input down to <code>[42]<\/code>, as\nseen in the keys <code>[:shrunk :smallest]<\/code>.<\/p>\n<p>To learn more, check out the <a href=\"#documentation\">documentation<\/a> links.<\/p>\n<h3><a id=\"user-content-clojuretest-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojuretest-integration\"><\/a><code>clojure.test<\/code> Integration<\/h3>\n<p>The macro <code>clojure.test.check.clojure-test\/defspec<\/code> allows you to succinctly\nwrite properties that run under the <code>clojure.test<\/code> runner, for example:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defspec first-element-is-min-after-sorting ;; the name of the test\n  100 ;; the number of iterations for test.check to test\n  (prop\/for-all [v (gen\/not-empty (gen\/vector gen\/int))]\n    (= (apply min v)\n       (first (sort v)))))<\/pre><\/div>\n<h3><a id=\"user-content-clojurescript\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojurescript\"><\/a>ClojureScript<\/h3>\n<p>ClojureScript support was added in version <code>0.7.0<\/code>.<\/p>\n<p>Integrating with <code>cljs.test<\/code> is via the\n<code>clojure.test.check.clojure-test\/defspec<\/code> macro, in the same fashion\nas integration with <code>clojure.test<\/code> on the jvm.<\/p>\n<h2><a id=\"user-content-developer-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#developer-information\"><\/a>Developer Information<\/h2>\n<p>We can not accept pull requests. Please see\n<a href=\"https:\/\/github.com\/clojure\/test.check\/blob\/master\/CONTRIBUTING.md\">CONTRIBUTING.md<\/a> for details.<\/p>\n<h3><a id=\"user-content-links\" class=\"anchor\" aria-hidden=\"true\" href=\"#links\"><\/a>Links<\/h3>\n<ul><li><a href=\"https:\/\/github.com\/clojure\/test.check\">GitHub project<\/a><\/li>\n<li><a href=\"http:\/\/dev.clojure.org\/jira\/browse\/TCHECK\" rel=\"nofollow\">Bug Tracker<\/a><\/li>\n<li><a href=\"http:\/\/build.clojure.org\/job\/test.check\/\" rel=\"nofollow\">Continuous Integration<\/a><\/li>\n<li><a href=\"http:\/\/build.clojure.org\/job\/test.check-test-matrix\/\" rel=\"nofollow\">Compatibility Test Matrix<\/a><\/li>\n<\/ul><h3><a id=\"user-content-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#tests\"><\/a>Tests<\/h3>\n<p>test.check runs in both jvm-clojure and clojurescript, so testing\ncomprehensively requires several steps:<\/p>\n<ul><li>Run <code>lein test<\/code> to run the JVM tests (requires <a href=\"https:\/\/leiningen.org\" rel=\"nofollow\">Leiningen<\/a>)<\/li>\n<li>Run <code>lein cljsbuild once<\/code> to run the ClojureScript tests (also requires <a href=\"https:\/\/nodejs.org\" rel=\"nofollow\">node.js<\/a>)<\/li>\n<li>To run the same tests in a web browser, open (after running the above command)\n<code>test-runners\/run_tests_dev.html<\/code> and <code>test-runners\/run_tests_adv.html<\/code> and watch the\njavascript console for output<\/li>\n<li>Run <code>script\/test-self-host<\/code> to run the self-hosted ClojureScript tests (also requires <a href=\"https:\/\/nodejs.org\" rel=\"nofollow\">node.js<\/a>)<\/li>\n<\/ul><h2><a id=\"user-content-see-also\" class=\"anchor\" aria-hidden=\"true\" href=\"#see-also\"><\/a>See also<\/h2>\n<h3><a id=\"user-content-other-implementations\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-implementations\"><\/a>Other implementations<\/h3>\n<ul><li><a href=\"http:\/\/hackage.haskell.org\/package\/QuickCheck\" rel=\"nofollow\">QC for Haskell<\/a><\/li>\n<li><a href=\"http:\/\/www.quviq.com\/index.html\" rel=\"nofollow\">The significantly more advanced QC for\nErlang<\/a><\/li>\n<\/ul><h3><a id=\"user-content-papers\" class=\"anchor\" aria-hidden=\"true\" href=\"#papers\"><\/a>Papers<\/h3>\n<ul><li><a href=\"http:\/\/www.eecs.northwestern.edu\/~robby\/courses\/395-495-2009-fall\/quick.pdf\" rel=\"nofollow\">QuickCheck: A Lightweight Tool for Random Testing of Haskell\nPrograms<\/a><\/li>\n<\/ul><h3><a id=\"user-content-simple-check\" class=\"anchor\" aria-hidden=\"true\" href=\"#simple-check\"><\/a>simple-check<\/h3>\n<p><em>test.check<\/em> used to be called\n<a href=\"https:\/\/github.com\/reiddraper\/simple-check\"><em>simple-check<\/em><\/a>.<\/p>\n<p>See <a href=\"https:\/\/github.com\/clojure\/test.check\/blob\/master\/doc\/migrating-from-simple-check.md\">migrating from simple-check<\/a>.<\/p>\n<h2><a id=\"user-content-yourkit\" class=\"anchor\" aria-hidden=\"true\" href=\"#yourkit\"><\/a>YourKit<\/h2>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/522ec07ba979f767228bfa281844244a64ed57b7\/687474703a2f2f7777772e796f75726b69742e636f6d2f696d616765732f796b6c6f676f2e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/522ec07ba979f767228bfa281844244a64ed57b7\/687474703a2f2f7777772e796f75726b69742e636f6d2f696d616765732f796b6c6f676f2e706e67\" alt=\"YourKit\" data-canonical-src=\"http:\/\/www.yourkit.com\/images\/yklogo.png\" \/><\/a><\/p>\n<p>YourKit is kindly supporting test.check and other open source projects with its\nfull-featured Java Profiler.  YourKit, LLC is the creator of innovative and\nintelligent tools for profiling Java and .NET applications. Take a look at\nYourKit's leading software products:<\/p>\n<ul><li><a href=\"http:\/\/www.yourkit.com\/java\/profiler\/index.jsp\" rel=\"nofollow\">YourKit Java Profiler<\/a> and<\/li>\n<li><a href=\"http:\/\/www.yourkit.com\/.net\/profiler\/index.jsp\" rel=\"nofollow\">YourKit .NET Profiler<\/a><\/li>\n<\/ul><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2014 Rich Hickey, Reid Draper and contributors<\/p>\n<p>Distributed under the Eclipse Public License, the same as Clojure.<\/p>\n<\/article>","created_at":"2018-11-07T16:44:03+0000","updated_at":"2018-11-07T16:44:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/317875?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":304,"title":"Getting started \u00b7 Gorilla REPL","url":"http:\/\/gorilla-repl.org\/start.html","content":"<h2 id=\"installation\">Installation<\/h2><p>The rest of these docs assume that you\u2019re familiar with the basics of Clojure, and have a working copy of Leiningen\n(version &gt;= 2) installed. If you\u2019re not yet familiar with Leiningen then you should head over to\n<a href=\"http:\/\/leiningen.org\">the Leiningen website<\/a> and get it installed first. It\u2019s a really nice tool that makes things very\neasy!<\/p><p>Gorilla is packaged as a Leiningen plugin. To use Gorilla in one of your Leiningen projects, then add the following to\nthe <code class=\"highlighter-rouge\">:plugins<\/code> section of that project\u2019s <code class=\"highlighter-rouge\">project.clj<\/code> file:<\/p><div class=\"language-clojure highlighter-rouge\"><pre>[lein-gorilla \"0.4.0\"]\n<\/pre><\/div><p>Your completed <code class=\"highlighter-rouge\">project.clj<\/code> file might look something like this:<\/p><div class=\"language-clojure highlighter-rouge\"><pre>(defproject gorilla-test \"0.1.0-SNAPSHOT\"\n  :description \"A test project for Gorilla REPL.\"\n  :dependencies [[org.clojure\/clojure \"1.6.0\"]]\n  :main ^:skip-aot gorilla-test.core\n  :target-path \"target\/%s\"\n  :plugins [[lein-gorilla \"0.4.0\"]]\n  :profiles {:uberjar {:aot :all}})\n<\/pre><\/div><p>That\u2019s it. You should now be able to run <code class=\"highlighter-rouge\">lein gorilla<\/code> from within the project directory and get started.<\/p><h2 id=\"usage\">Usage<\/h2><p>When you run <code class=\"highlighter-rouge\">lein gorilla<\/code> it will start up the REPL server, and print a web-link to the console. Point your\nweb-browser at this link to get going (hint for Mac users: try ctrl-clicking the link). You can open as many browser\nwindows as you like with this link, each will get its own nREPL session to work in, but share the same nREPL instance\n(in case you\u2019re not familiar with nREPL\u2019s terminology: this means all windows will share definitions etc, but each\nwindow will separately keep track of which namespace you\u2019re working in - try it, you\u2019ll see it\u2019s quite natural).<\/p><p>Once you\u2019ve got a web-browser pointed at Gorilla you can use it just like a REPL. Type some clojure code in, and hit\n<code class=\"highlighter-rouge\">shift+enter<\/code> to evaluate it. The results are displayed below the code, along with any console output or errors that\nwere generated. Gorilla offers nREPL\u2019s autocomplete function, hit <code class=\"highlighter-rouge\">ctrl+space<\/code> to see what nREPL has to suggest (unless\n you\u2019re using Firefox on Mac - see the commands section below for more info.)<\/p><h3 id=\"plotting-graphs\">Plotting graphs<\/h3><p>One of the most handy features of Gorilla is the ability to plot graphs. The plotting library\ngorilla-plot is integrated into Gorilla and is always available without\nexplicitly including it as a dependency in your <code class=\"highlighter-rouge\">project.clj<\/code>. Full documentation for gorilla-plot is available at the\n<a href=\"http:\/\/gorilla-repl.org\/plotting.html\">plotting<\/a> page, but to get you started, let\u2019s give a short summary.<\/p><p>There are five functions that should cover many plotting needs. These functions are in the <code class=\"highlighter-rouge\">gorilla-plot.core<\/code>\nnamespace, so you will need to <code class=\"highlighter-rouge\">use<\/code> or <code class=\"highlighter-rouge\">require<\/code> it before starting. The functions are:<\/p><ul><li>\n<code class=\"highlighter-rouge\">(list-plot data)<\/code> where <code class=\"highlighter-rouge\">data<\/code> can either be a sequence of y-values, or a sequence of <code class=\"highlighter-rouge\">(x y)<\/code> pairs.<\/li>\n  <li>\n<code class=\"highlighter-rouge\">(plot func [start end])<\/code> which will evaluate and plot <code class=\"highlighter-rouge\">func<\/code> over the given range.<\/li>\n  <li>\n<code class=\"highlighter-rouge\">(histogram data)<\/code> where <code class=\"highlighter-rouge\">data<\/code> is a list of values.<\/li>\n  <li>\n<code class=\"highlighter-rouge\">(bar-chart categories values)<\/code> where <code class=\"highlighter-rouge\">categories<\/code> are the category names, and <code class=\"highlighter-rouge\">values<\/code> their values.<\/li>\n  <li>\n<code class=\"highlighter-rouge\">(compose plot1 plot2 &amp; more)<\/code> which tries to compose together the given plots. Note that composing a bar-chart with\nother plots will give odd results, as it\u2019s not obvious how to compose category-scales.<\/li>\n<\/ul><p>These functions take many options, look at the detailed docs for more help.<\/p><p>There\u2019s a <a href=\"https:\/\/vimeo.com\/87139900\">short video<\/a> talking a little more about how the plots work, and how they try and\nfit nicely with the Clojure way of thinking (plots are values) which might interest you.<\/p><h3 id=\"other-kinds-of-output\">Other kinds of output<\/h3><p>Plots aren\u2019t the only way that Gorilla can give you a more useful view into your Clojure values. There are a number of\nother built in functions to view data:<\/p><ul><li>\n<code class=\"highlighter-rouge\">table-view<\/code> in the <code class=\"highlighter-rouge\">gorilla-repl.table<\/code> namespace lets you view lists-of-lists as tables. You can supply an\noptional vector as a <code class=\"highlighter-rouge\">:columns<\/code> argument to label the columns of the table.<\/li>\n  <li>\n<code class=\"highlighter-rouge\">latex-view<\/code> in the <code class=\"highlighter-rouge\">gorilla-repl.latex<\/code> namespace lets you view a string as its rendered LaTeX form.<\/li>\n  <li>\n<code class=\"highlighter-rouge\">html-view<\/code> in the <code class=\"highlighter-rouge\">gorilla-repl.html<\/code> namespace lets you view a string rendered as HTML.<\/li>\n<\/ul><p>These built-in view functions are just the beginning though. Gorilla REPL has a very\n<a href=\"https:\/\/vimeo.com\/89529751\">flexible<\/a>, <a href=\"https:\/\/vimeo.com\/89532785\">extensible<\/a> renderer so you can\nplug in new ways of viewing values. If your favourite library doesn\u2019t have the viewers for the data that you want, then\nfile a feature request, or even better write some code and contribute it!<\/p><p>A feature worth mentioning here is value-copy-and-paste. Although Gorilla shows some values (like graphs, tables) with\nfancy formatting, the underlying Clojure value is always there to be used. If you alt-click on any output in Gorilla\nit will give you the readable Clojure value to work with (if it exists).<\/p><h3 id=\"editor-commands\">Editor commands<\/h3><p>Before we can go much further we will need to introduce <em>editor commands<\/em>. You don\u2019t need them right now, but you can\nsee all of the editor commands by hitting <code class=\"highlighter-rouge\">ctrl+g<\/code> twice in succession (<code class=\"highlighter-rouge\">alt-g<\/code> on Windows and Linux, see below), or\nclicking on the feint menu icon in the top right hand corner of the Gorilla window. Hopefully they are all\nself-explanatory, or at least you can figure them out easily enough.<\/p><p>You\u2019ll probably want to use keyboard shortcuts to issue the editor commands if you use Gorilla much: they are usually a\nsequence of two keypresses. On Windows these are of the form <code class=\"highlighter-rouge\">alt+a alt+b<\/code> and on other platforms they are of the form\n<code class=\"highlighter-rouge\">ctrl+a ctrl+b<\/code>.\nThis document writes them in the format suitable for Mac - if you\u2019re on Windows\/Linux replace the <code class=\"highlighter-rouge\">ctrl<\/code>s with <code class=\"highlighter-rouge\">alt<\/code>s.\nAn attempt\nhas been made to make sure the commands work across popular browsers and operating systems (which is not as easy as you\nmight think). The one exception is the autocomplete command which doesn\u2019t work on Firefox on the Mac, as it\nsteals <code class=\"highlighter-rouge\">ctrl+space<\/code> for its own use, <a href=\"https:\/\/bugzilla.mozilla.org\/show_bug.cgi?id=435164\">somewhat controversially<\/a>.\nYou can instead use <code class=\"highlighter-rouge\">ctrl+g ctrl+a<\/code> if you\u2019re using Firefox.<\/p><p>Note that you can auto-indent code in the editor by selecting it and hitting shift-tab. This can be a nice way to spot\nsyntax errors :-)<\/p><h3 id=\"making-notes\">Making notes<\/h3><p>So far we\u2019ve used Gorilla as a fancy REPL, but we can also think of it as a tool for making documents, which we call\n\u2018worksheets\u2019. As well as including snippets of Clojure code, a Gorilla worksheet can include notes, which are written in\nMarkdown format. To add notes you need to first tell Gorilla that it should interpret a snippet of text as notes, rather\nthan Clojure code. To do this place the cursor in the segment (a segment being one of the boxes containing a snippet of\ncode\/notes) that you want to use for notes and hit, <code class=\"highlighter-rouge\">ctrl+g ctrl+m<\/code> (g for Gorilla, m for Markdown). You can then feel\nfree to put any Markdown you like in there. The notes segments also support LaTeX formulae. To write a formula you\nsimply surround the latex code with $$, or @@ if you want the formula to appear inline. So for instance, the contents of\na Markdown segment could be:<\/p><div class=\"highlighter-rouge\"><pre>This is an inline formula, @@\\sin(x)@@, and this is on its own line:\n$$\\int_0^{2\\pi}\\sin^2(x) \\textrm{d}x$$\n<\/pre><\/div><p><strong>Note:<\/strong> currently you will need to be online in order for LaTeX to render properly.<\/p><h4 id=\"spell-checking\">Spell checking<\/h4><p>Due to limitations of the underlying CodeMirror component, spell checking of the notes is currently not possible.\nA workaround is to open the worksheet file (see below) with a Clojure editor which supports spell checking. Emacs work\nvery well for\nthis, with its \u2018ispell-comments-and-strings\u201d command which ignores code blocks during the spell check. The document\nshould be saved without the output. Gorilla has a command to clear the output of all the segments, which\nmakes this easy.<\/p><h3 id=\"worksheet-files\">Worksheet files<\/h3><p>You can save the contents of a window to a worksheet file. This will include everything you see, the code, the output,\ngraphs, notes and mathematics, the lot. To save a file just hit <code class=\"highlighter-rouge\">ctrl+g ctrl+s<\/code>. If you haven\u2019t already saved the file it will\nprompt for a filename, which is given relative to the project. To load a file, use\n<code class=\"highlighter-rouge\">ctrl+g ctrl+l<\/code> (note that Gorilla files must end in <code class=\"highlighter-rouge\">.clj<\/code> or <code class=\"highlighter-rouge\">.cljw<\/code> to be loaded). By convention, I often find it\nconvenient to store my worksheets in a directory called <code class=\"highlighter-rouge\">ws<\/code> at the root\nof the project (alongside <code class=\"highlighter-rouge\">src<\/code> etc) but you, of course, can store them wherever you want. A neat feature is that these\nworksheet files are just plain Clojure files with some magic comments. This means it\u2019s really easy to interactively\ndevelop your code in Gorilla, and then turn it into a library when it stabilises.<\/p><h3 id=\"doc-source-and-other-repl-commands\">doc, source and other REPL commands<\/h3><p>You might be used to using <code class=\"highlighter-rouge\">doc<\/code> and <code class=\"highlighter-rouge\">source<\/code> at the command-line REPL. By default these are not imported into the\n<code class=\"highlighter-rouge\">user<\/code> namespace when Gorilla starts, but if you\u2019d like to use them then you just need to run <code class=\"highlighter-rouge\">(use 'clojure.repl)<\/code> to\nbring them into scope.<\/p><h3 id=\"other-editors\">Other editors<\/h3><p>Gorilla works well alongside other editors and there\u2019s <a href=\"http:\/\/gorilla-repl.org\/editors.html\">a page<\/a> with hints on setting up your favourite\nenvironment to work well with Gorilla.<\/p><h3 id=\"server-side-usage\">Server-side usage<\/h3><p>Many people find it useful to run Gorilla on a server. As all of Gorilla\u2019s communication happens over HTTP and\nwebsockets, this is usually very easy. A couple of notes:<\/p><ul><li>Any user who has access to Gorilla effectively has access to the server machine with the privileges of the JVM\nprocess. So, you will want to ensure that only trusted users have access. Gorilla does not offer any built-in \nauthentication mechanism. It is recommended to either use a reverse-proxy, network-level controls, or access through\nssh tunnels (see next point).<\/li>\n  <li>In the case there is a firewall or proxy between your browser and a remote Gorilla server, it might be possible that\nthe websocket connection to the Gorilla repl is blocked. One solution to this is to tunnel access over SSH, which can\nalso be used as a user authentication method.<\/li>\n  <li>By default Gorilla does not bind to external IPs, so is only accessible from the server itself. If you want to expose\nGorilla to the network, then see the advanced configuration options below.<\/li>\n<\/ul><h3 id=\"advanced-configuration\">Advanced configuration<\/h3><p>You can see what can be customised in Gorilla on the <a href=\"http:\/\/gorilla-repl.org\/configuration.html\">configuration<\/a> page.<\/p>","created_at":"2018-11-07T16:44:03+0000","updated_at":"2018-11-07T16:44:03+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"gorilla-repl.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":305,"title":"Clojure for Data Science: Clojure for Machine Learning","url":"http:\/\/clojuredatascience.com\/posts\/2016-12-01-clojure-machine-learning.html","content":"<div><p><a href=\"https:\/\/twitter.com\/henrygarner\">Henry<\/a> gave a talk at London's <a href=\"https:\/\/skillsmatter.com\/conferences\/7430-clojure-exchange-2016#program\">Clojure eXchange 2016<\/a> on the subject of machine learning in Clojure.<\/p><p>From the programme:<\/p><blockquote><p> You have heard that Clojure\u2019s ancestor language, Lisp, was developed for artificial intelligence research. Yet, until recently, Clojure\u2019s support for machine learning has mostly consisted of wrapped Java libraries.<\/p><p>Informed by his work as a freelance data scientist working primarily with Clojure, Henry will explore some newer libraries that enable sophisticated data analysis unencumbered by Java\u2019s legacy. You will also learn how Clojure\u2019s core sequence-processing strengths naturally lend themselves to simple machine learning techniques you can try at home. <\/p><\/blockquote><p>You can see a <a href=\"https:\/\/skillsmatter.com\/skillscasts\/9050-clojure-for-machine-learning\">video of the talk<\/a> on the Skills Matter website, the <a href=\"https:\/\/github.com\/henrygarner\/cljx-december-2016\">org-mode slides<\/a>, or a PDF of the presentation below.<\/p><\/div><p><b>Tags: <\/b>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/clojure.html\">clojure<\/a>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/data%20science.html\">data science<\/a>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/linear%20regression.html\">linear regression<\/a>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/transducers.html\">transducers<\/a>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/talks.html\">talks<\/a>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/machine%20learning.html\">machine learning<\/a>\n    \n    <a href=\"http:\/\/clojuredatascience.com\/tags\/classification.html\">classification<\/a><\/p><p><a href=\"http:\/\/clojuredatascience.com\/posts\/2016-12-02-data-science-ladder-abstraction.html\">\u00ab Data Science Up and Down the Ladder of Abstraction<\/a>\n        \n        \n        <a class=\"right\" href=\"http:\/\/clojuredatascience.com\/posts\/2015-12-05-expressive-parallel-analytics-with-clojure.html\">Expressive Parallel Analytics with Clojure \u00bb<\/a><\/p>","created_at":"2018-11-07T16:44:03+0000","updated_at":"2018-11-07T16:44:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"clojuredatascience.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":301,"title":"Emacs Lisp Lambda Expressions Are Not Self-Evaluating \u00ab null program","url":"https:\/\/nullprogram.com\/blog\/2018\/02\/22\/","content":"<time datetime=\"2018-02-22\">\n    February 22, 2018\n  <\/time><p>nullprogram.com\/blog\/2018\/02\/22\/<\/p><p>This week I made a mistake that ultimately enlightened me about the\nnature of function objects in Emacs Lisp. There are three kinds of\nfunction objects, but they each behave very differently when evaluated\nas objects.<\/p><p>But before we get to that, let\u2019s talk about one of Emacs\u2019\nembarrassing, old missteps: <code class=\"highlighter-rouge\">eval-after-load<\/code>.<\/p><h3 id=\"taming-an-old-dragon\">Taming an old dragon<\/h3><p>One of the long-standing issues with Emacs is that loading Emacs Lisp\nfiles (.el and .elc) is a slow process, even when those files have\nbeen byte compiled. There are a number of dirty hacks in place to deal\nwith this issue, and the biggest and nastiest of them all is the\n<a href=\"https:\/\/lwn.net\/Articles\/707615\/\"><em>dumper<\/em><\/a>, also known as <em>unexec<\/em>.<\/p><p>The Emacs you routinely use throughout the day is actually a previous\ninstance of Emacs that\u2019s been resurrected from the dead. Your undead\nEmacs was probably created months, if not years, earlier, back when it\nwas originally compiled. The first stage of compiling Emacs is to\ncompile a minimal C core called <code class=\"highlighter-rouge\">temacs<\/code>. The second stage is loading\na bunch of Emacs Lisp files, then dumping a memory image in an\nunportable, platform-dependent way. On Linux, this actually <a href=\"https:\/\/lwn.net\/Articles\/707615\/\">requires\nspecial hooks in glibc<\/a>. The Emacs you know and love is this\ndumped image loaded back into memory, continuing from where it left\noff just after it was compiled. Regardless of your own feelings on the\nmatter, you have to admit <a href=\"https:\/\/nullprogram.com\/blog\/2011\/01\/30\/\">this <em>is<\/em> a very lispy thing to do<\/a>.<\/p><p>There are two notable costs to Emacs\u2019 dumper:<\/p><ol><li>\n    <p>The dumped image contains hard-coded memory addresses. This means\nEmacs can\u2019t be a <em>Position Independent Executable<\/em> (PIE). It can\u2019t\ntake advantage of a security feature called <em>Address Space Layout\nRandomization<\/em> (ASLR), which would increase the difficulty of\n<a href=\"https:\/\/nullprogram.com\/blog\/2017\/07\/19\/\">exploiting<\/a> some <a href=\"https:\/\/nullprogram.com\/blog\/2012\/09\/28\/\">classes of bugs<\/a>. This might be\nimportant to you if Emacs processes untrusted data, such as when it\u2019s\nused as <a href=\"https:\/\/nullprogram.com\/blog\/2013\/09\/03\/\">a mail client<\/a>, <a href=\"https:\/\/github.com\/skeeto\/emacs-web-server\">a web server<\/a> or generally\n<a href=\"https:\/\/github.com\/skeeto\/elfeed\">parses data downloaded across the network<\/a>.<\/p>\n  <\/li>\n  <li>\n    <p>It\u2019s not possible to cross-compile Emacs since it can only be dumped\nby running <code class=\"highlighter-rouge\">temacs<\/code> on its target platform. As an experiment I\u2019ve\nattempted to dump the Windows version of Emacs on Linux using\n<a href=\"https:\/\/www.winehq.org\/\">Wine<\/a>, but was unsuccessful.<\/p>\n  <\/li>\n<\/ol><p>The good news is that there\u2019s <a href=\"https:\/\/lists.gnu.org\/archive\/html\/emacs-devel\/2018-02\/msg00347.html\">a portable dumper<\/a> in the works\nthat makes this a lot less nasty. If you\u2019re adventurous, you can\nalready disable dumping and run <code class=\"highlighter-rouge\">temacs<\/code> directly by setting\n<a href=\"https:\/\/lists.gnu.org\/archive\/html\/bug-gnu-emacs\/2016-11\/msg00729.html\"><code class=\"highlighter-rouge\">CANNOT_DUMP=yes<\/code> at compile time<\/a>. Be warned, though, that a\nnon-dumped Emacs takes several seconds, or worse, to initialize\n<em>before<\/em> it even begins loading your own configuration. It\u2019s also\nsomewhat buggy since it seems nobody ever runs it this way\nproductively.<\/p><p>The other major way Emacs users have worked around slow loading is\naggressive use of lazy loading, generally via <em>autoloads<\/em>. The major\npackage interactive entry points are defined ahead of time as stub\nfunctions. These stubs, when invoked, load the full package, which\noverrides the stub definition, then finally the stub re-invokes the\nnew definition with the same arguments.<\/p><p>To further assist with lazy loading, an evaluated <code class=\"highlighter-rouge\">defvar<\/code> form will\nnot override an existing global variable binding. This means you can,\nto a certain extent, configure a package before it\u2019s loaded. The\npackage will not clobber any existing configuration when it loads.\nThis also explains the bizarre interfaces for the various hook\nfunctions, like <code class=\"highlighter-rouge\">add-hook<\/code> and <code class=\"highlighter-rouge\">run-hooks<\/code>. These accept symbols \u2014 the\n<em>names<\/em> of the variables \u2014 rather than <em>values<\/em> of those variables as\nwould normally be the case. The <code class=\"highlighter-rouge\">add-to-list<\/code> function does the same\nthing. It\u2019s all intended to cooperate with lazy loading, where the\nvariable may not have been defined yet.<\/p><h4 id=\"eval-after-load\">eval-after-load<\/h4><p>Sometimes this isn\u2019t enough and you need some some configuration to\ntake place after the package has been loaded, but without forcing it\nto load early. That is, you need to tell Emacs \u201cevaluate this code\nafter this particular package loads.\u201d That\u2019s where <code class=\"highlighter-rouge\">eval-after-load<\/code>\ncomes into play, except for its fatal flaw: it takes the word \u201ceval\u201d\ncompletely literally.<\/p><p>The first argument to <code class=\"highlighter-rouge\">eval-after-load<\/code> is the name of a package. Fair\nenough. The second argument is a form that will be passed to <code class=\"highlighter-rouge\">eval<\/code>\nafter that package is loaded. Now hold on a minute. The general rule\nof thumb is that if you\u2019re calling <code class=\"highlighter-rouge\">eval<\/code>, you\u2019re probably doing\nsomething seriously wrong, and this function is no exception. This is\n<em>completely<\/em> the wrong mechanism for the task.<\/p><p>The second argument should have been a function \u2014 either a (sharp\nquoted) symbol or a function object. And then instead of <code class=\"highlighter-rouge\">eval<\/code> it\nwould be something more sensible, like <code class=\"highlighter-rouge\">funcall<\/code>. Perhaps this\nimproved version would be named <code class=\"highlighter-rouge\">call-after-load<\/code> or <code class=\"highlighter-rouge\">run-after-load<\/code>.<\/p><p>The big problem with passing an s-expression is that it will be left\nuncompiled due to being quoted. <a href=\"https:\/\/nullprogram.com\/blog\/2017\/12\/14\/\">I\u2019ve talked before about the\nimportance of evaluating your lambdas<\/a>. <code class=\"highlighter-rouge\">eval-after-load<\/code> not\nonly encourages badly written Emacs Lisp, it demands it.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;;; BAD!\n(eval-after-load 'simple-httpd\n                 '(push '(\"c\" . \"text\/plain\") httpd-mime-types))\n<\/pre><\/div><\/div><p>This was all corrected in Emacs 25. If the second argument to\n<code class=\"highlighter-rouge\">eval-after-load<\/code> is a function \u2014 the result of applying <code class=\"highlighter-rouge\">functionp<\/code> is\nnon-nil \u2014 then it uses <code class=\"highlighter-rouge\">funcall<\/code>. There\u2019s also a new macro,\n<code class=\"highlighter-rouge\">with-eval-after-load<\/code>, to package it all up nicely.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;;; Better (Emacs &gt;= 25 only)\n(eval-after-load 'simple-httpd\n  (lambda ()\n    (push '(\"c\" . \"text\/plain\") httpd-mime-types)))\n;;; Best (Emacs &gt;= 25 only)\n(with-eval-after-load 'simple-httpd\n  (push '(\"c\" . \"text\/plain\") httpd-mime-types))\n<\/pre><\/div><\/div><p>Though in both of these examples the compiler will likely warn about\n<code class=\"highlighter-rouge\">httpd-mime-types<\/code> not being defined. That\u2019s a problem for another\nday.<\/p><h4 id=\"a-workaround\">A workaround<\/h4><p>But what if you <em>need<\/em> to use Emacs 24, as was the <a href=\"https:\/\/github.com\/skeeto\/elfeed\/pull\/268\">situation that\nsparked this article<\/a>? What can we do with the bad version of\n<code class=\"highlighter-rouge\">eval-after-load<\/code>? We could situate a lambda such that it\u2019s evaluated,\nbut then smuggle the resulting function object into the form passed to\n<code class=\"highlighter-rouge\">eval-after-load<\/code>, all using a backquote.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;;; Note: this is subtly broken\n(eval-after-load 'simple-httpd\n  `(funcall\n    ,(lambda ()\n       (push '(\"c\" . \"text\/plain\") httpd-mime-types)))\n<\/pre><\/div><\/div><p>When everything is compiled, the backquoted form evalutes to this:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>(funcall #[0 &lt;bytecode&gt; [httpd-mime-types (\"c\" . \"text\/plain\")] 2])\n<\/pre><\/div><\/div><p>Where the second value (<code class=\"highlighter-rouge\">#[...]<\/code>) is a <a href=\"https:\/\/nullprogram.com\/blog\/2014\/01\/04\/\">byte-code object<\/a>.\nHowever, as the comment notes, this is subtly broken. A cleaner and\ncorrect way to solve all this is with a named function. The damage\ncaused by <code class=\"highlighter-rouge\">eval-after-load<\/code> will have been (mostly) minimized.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>(defun my-simple-httpd-hook ()\n  (push '(\"c\" . \"text\/plain\") httpd-mime-types))\n(eval-after-load 'simple-httpd\n  '(funcall #'my-simple-httpd-hook))\n<\/pre><\/div><\/div><p>But, let\u2019s go back to the anonymous function solution. What was broken\nabout it? It all has to do with evaluating function objects.<\/p><p>So what happens when we evaluate an expression like the one above with\n<code class=\"highlighter-rouge\">eval<\/code>? Here\u2019s what it looks like again.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>(funcall #[...])\n<\/pre><\/div><\/div><p>First, <code class=\"highlighter-rouge\">eval<\/code> notices it\u2019s been given a non-empty list, so it\u2019s probably\na function call. The first argument is the name of the function to be\ncalled (<code class=\"highlighter-rouge\">funcall<\/code>) and the remaining elements are its arguments. <em>But<\/em>\neach of these elements must be evaluated first, and the <em>result<\/em> of that\nevaluation becomes the arguments.<\/p><p>Any value that isn\u2019t a list or a symbol is <em>self-evaluating<\/em>. That is,\nit evaluates to its own value:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>(eval 10)\n;; =&gt; 10\n<\/pre><\/div><\/div><p>If the value is a symbol, it\u2019s treated as a variable. If the value is a\nlist, it goes through the function call process I\u2019m describing (or one\nof a number of other special cases, such as macro expansion, lambda\nexpressions, and special forms).<\/p><p>So, conceptually <code class=\"highlighter-rouge\">eval<\/code> recurses on the function object <code class=\"highlighter-rouge\">#[...]<\/code>. A\nfunction object is not a list or a symbol, so it\u2019s self-evaluating. No\nproblem.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;; Byte-code objects are self-evaluating\n(let ((x (byte-compile (lambda ()))))\n  (eq x (eval x)))\n;; =&gt; t\n<\/pre><\/div><\/div><p>What if this code <em>wasn\u2019t<\/em> compiled? Rather than a byte-code object,\nwe\u2019d have some other kind of function object for the interpreter.\nLet\u2019s examine the dynamic scope (<em>shudder<\/em>) case. Here, a lambda\n<em>appears<\/em> to evaluate to itself, but appearances can be deceiving:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>(eval (lambda ())\n;; =&gt; (lambda ())\n<\/pre><\/div><\/div><p>However, this is not self-evaluation. <strong>Lambda expressions are not\nself-evaluating<\/strong>. It\u2019s merely <em>coincidence<\/em> that the result of\nevaluating a lambda expression looks like the original expression.\nThis is just how the Emacs Lisp interpreter is currently implemented\nand, strictly speaking, it\u2019s an implementation detail that <em>just so\nhappens<\/em> to be mostly compatible with byte-code objects being\nself-evaluating. It would be a mistake to rely on this.<\/p><p>Instead, <strong>dynamic scope lambda expression evaluation is\n<a href=\"https:\/\/labs.spotify.com\/2013\/06\/18\/creative-usernames\/\">idempotent<\/a>.<\/strong> Applying <code class=\"highlighter-rouge\">eval<\/code> to the result will return\nan <code class=\"highlighter-rouge\">equal<\/code>, but not identical (<code class=\"highlighter-rouge\">eq<\/code>), expression. In contrast, a\nself-evaluating value is also idempotent under evaluation, but with\n<code class=\"highlighter-rouge\">eq<\/code> results.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;; Not self-evaluating:\n(let ((x '(lambda ())))\n  (eq x (eval x)))\n;; =&gt; nil\n;; Evaluation is idempotent:\n(let ((x '(lambda ())))\n  (equal x (eval x)))\n;; =&gt; t\n(let ((x '(lambda ())))\n  (equal x (eval (eval x))))\n;; =&gt; t\n<\/pre><\/div><\/div><p>So, with dynamic scope, the subtly broken backquote example will still\nwork, but only by sheer luck. Under lexical scope, the situation isn\u2019t\nso lucky:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;;; -*- lexical-scope: t; -*-\n(lambda ())\n;; =&gt; (closure (t) nil)\n<\/pre><\/div><\/div><p>These interpreted lambda functions are neither self-evaluating nor\nidempotent. Passing <code class=\"highlighter-rouge\">t<\/code> as the second argument to <code class=\"highlighter-rouge\">eval<\/code> tells it to\nuse lexical scope, as shown below:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;; Not self-evaluating:\n(let ((x '(lambda ())))\n  (eq x (eval x t)))\n;; =&gt; nil\n;; Not idempotent:\n(let ((x '(lambda ())))\n  (equal x (eval x t)))\n;; =&gt; nil\n(let ((x '(lambda ())))\n  (equal x (eval (eval x t) t)))\n;; error: (void-function closure)\n<\/pre><\/div><\/div><p>I can <a href=\"https:\/\/nullprogram.com\/blog\/2017\/05\/03\/\">imagine an implementation<\/a> of Emacs Lisp where dynamic\nscope lambda expressions are in the same boat, where they\u2019re not even\nidempotent. For example:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;;; -*- lexical-binding: nil; -*-\n(lambda ())\n;; =&gt; (totally-not-a-closure ())\n<\/pre><\/div><\/div><p>Most Emacs Lisp would work just fine under this change, and only code\nthat makes some kind of logical mistake \u2014 where there\u2019s nested\nevaluation of lambda expressions \u2014 would break. This essentially\nalready happened when lots of code was quietly switched over to\nlexical scope after Emacs 24. Lambda idempotency was lost and\nwell-written code didn\u2019t notice.<\/p><p>There\u2019s a temptation here for Emacs to define a <code class=\"highlighter-rouge\">closure<\/code> function or\nspecial form that would allow interpreter closure objects to be either\nself-evaluating or idempotent. This would be a mistake. It would only\nserve as a hack that covers up logical mistakes that lead to nested\nevaluation. Much better to catch those problems early.<\/p><h3 id=\"solving-the-problem-with-one-character\">Solving the problem with one character<\/h3><p>So how do we fix the subtly broken example? With a strategically\nplaced quote right before the comma.<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>(eval-after-load 'simple-httpd\n  `(funcall\n    ',(lambda ()\n        (push '(\"c\" . \"text\/plain\") httpd-mime-types)))\n<\/pre><\/div><\/div><p>So the form passed to <code class=\"highlighter-rouge\">eval-after-load<\/code> becomes:<\/p><div class=\"language-cl highlighter-rouge\"><div class=\"highlight\"><pre>;; Compiled:\n(funcall (quote #[...]))\n;; Dynamic scope:\n(funcall (quote (lambda () ...)))\n;; Lexical scope:\n(funcall (quote (closure (t) () ...)))\n<\/pre><\/div><\/div><p>The quote prevents <code class=\"highlighter-rouge\">eval<\/code> from evaluating the function object, which\nwould be either needless or harmful. There\u2019s also an argument to be\nmade that this is a perfect situation for a sharp-quote (<code class=\"highlighter-rouge\">#'<\/code>), which\nexists to quote functions.<\/p>","created_at":"2018-11-07T16:44:02+0000","updated_at":"2018-11-07T16:44:02+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"nullprogram.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":302,"title":"The Original 'Lambda Papers' by Guy Steele and Gerald Sussman","url":"http:\/\/library.readscheme.org\/page1.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:44:02+0000","updated_at":"2018-11-07T16:44:02+0000","annotations":[],"reading_time":0,"domain_name":"library.readscheme.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":299,"title":"antonlindstrom\/passpwn","url":"https:\/\/github.com\/antonlindstrom\/passpwn","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>This is a tool for checking your passwords in <a href=\"https:\/\/www.passwordstore.org\/\" rel=\"nofollow\">pass<\/a> against <a href=\"https:\/\/haveibeenpwned.com\/\" rel=\"nofollow\">Have I been pwned?<\/a>.<\/p>\n<p>passpwn uses the range API which means that we are only sending the first 5\ncharacters of a SHA1 hashed password to Have I been pwned? and are checking\nthe returned hashes against the full local hash.<\/p>\n<p>Inspiration for this tool comes from the <a href=\"https:\/\/blog.agilebits.com\/2018\/02\/22\/finding-pwned-passwords-with-1password\/\" rel=\"nofollow\">1Password\nblog<\/a>.<\/p>\n<p>A big thanks to haveibeenpwned.com for making this possible in a secure way.<\/p>\n<h2><a id=\"user-content-reporting-bugs\" class=\"anchor\" aria-hidden=\"true\" href=\"#reporting-bugs\"><\/a>Reporting bugs<\/h2>\n<p>If you find any bugs or want to provide feedback, you can file bugs in the project's <a href=\"https:\/\/github.com\/antonlindstrom\/passpwn\">GitHub Issues page<\/a>.<\/p>\n<h2><a id=\"user-content-author\" class=\"anchor\" aria-hidden=\"true\" href=\"#author\"><\/a>Author<\/h2>\n<p>This project is maintained by <a href=\"https:\/\/www.antonlindstrom.com\" rel=\"nofollow\">Anton Lindstr\u00f6m<\/a> (<a href=\"https:\/\/github.com\/antonlindstrom\">GitHub<\/a> | <a href=\"https:\/\/twitter.com\/mycap\" rel=\"nofollow\">Twitter<\/a>)<\/p>\n<h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>APACHE LICENSE 2.0\nCopyright 2018 Anton Lindstr\u00f6m<\/p>\n<p>Licensed under the Apache License, Version 2.0 (the \"License\");\nyou may not use this file except in compliance with the License.\nYou may obtain a copy of the License at<\/p>\n<p><a href=\"http:\/\/www.apache.org\/licenses\/LICENSE-2.0\" rel=\"nofollow\">http:\/\/www.apache.org\/licenses\/LICENSE-2.0<\/a><\/p>\n<p>Unless required by applicable law or agreed to in writing, software\ndistributed under the License is distributed on an \"AS IS\" BASIS,\nWITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\nSee the License for the specific language governing permissions and\nlimitations under the License.<\/p>\n<\/article>","created_at":"2018-11-07T16:44:01+0000","updated_at":"2018-11-07T16:44:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/10686?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":300,"title":"defmacro - The Nature of Lisp","url":"http:\/\/www.defmacro.org\/ramblings\/lisp.html","content":"<p>Monday, May 8, 2006<\/p><p class=\"first\">When I first stumbled into Lisp advocacy on various corners of the web I was already an experienced programmer. At that&#13;\n                point I had grokked what seemed at the time a wide range of programming languages. I was proud to have the usual suspects (C++, Java, C#,&#13;\n                etc.) on my service record and was under impression that I knew everything there is to know about programming languages. I couldn't have&#13;\n                possibly been more wrong.<\/p><p>My initial attempt to learn Lisp came to a crashing halt as soon as I saw some sample code. I suppose the same thought ran through my &#13;\n\t\t\t\tmind that ran through thousands of other minds who were ever in my shoes: \"Why on Earth would anyone want to use a language with such &#13;\n\t\t\t\thorrific syntax?!\" I couldn't be bothered to learn a language if its creators couldn't be bothered to give it a pleasant syntax. &#13;\n\t\t\t\tAfter all, I was almost blinded by the infamous Lisp parentheses!<\/p><p>The moment I regained my sight I communicated my frustrations to some members of the Lisp sect. Almost immediately I was bombarded by &#13;\n\t\t\t\ta standard set of responses: Lisp's parentheses are only a superficial matter, Lisp has a huge benefit of code and data being &#13;\n\t\t\t\texpressed in the same manner (which, obviously, is a huge improvement over XML), Lisp has tremendously powerful metaprogramming &#13;\n\t\t\t\tfacilities that allow programs to write code and modify themselves, Lisp allows for creation of mini-languages specific to the &#13;\n\t\t\t\tproblem at hand, Lisp blurs the distinction between run time and compile time, Lisp, Lisp, Lisp... The list was very impressive. &#13;\n\t\t\t\tNeedless to say none of it made sense. Nobody could illustrate the usefulness of these features with specific examples because these &#13;\n\t\t\t\ttechniques are supposedly only useful in large software systems. After many hours of debating that conventional programming languages &#13;\n\t\t\t\tdo the job just fine, I gave up. I wasn't about to invest months into learning a language with a terrible syntax in order to &#13;\n\t\t\t\tunderstand obscure features that had no useful examples. My time has not yet come.<\/p><p>For many months the Lisp advocates pressed on. I was baffled. Many extremely intelligent people I knew and had much respect for were &#13;\n\t\t\t\tpraising Lisp with almost religious dedication. There had to be something there, something I couldn't afford not to get my hands on! &#13;\n\t\t\t\tEventually my thirst for knowledge won me over. I took the plunge, bit the bullet, got my hands dirty, and began months of mind &#13;\n\t\t\t\tbending exercises. It was a journey on an endless lake of frustration. I turned my mind inside out, rinsed it, and put it back in &#13;\n\t\t\t\tplace. I went through seven rings of hell and came back. And then I got it.<\/p><p>The enlightenment came instantaneously. One moment I understood nothing, and the next moment everything clicked into place. I've &#13;\n\t\t\t\tachieved nirvana. Dozens of times I heard Eric Raymond's statement quoted by different people: \"Lisp is worth learning for the &#13;\n\t\t\t\tprofound enlightenment experience you will have when you finally get it; that experience will make you a better programmer for the &#13;\n\t\t\t\trest of your days, even if you never actually use Lisp itself a lot.\" I never understood this statement. I never believed it could be &#13;\n\t\t\t\ttrue. And finally, after all the pain, it made sense! There was more truth to it than I ever could have imagined. I've achieved an &#13;\n\t\t\t\talmost divine state of mind, an instantaneous enlightenment experience that turned my view of computer science on its head in less than &#13;\n\t\t\t\ta single second.<\/p><p>That very second I became a member of the Lisp cult. I felt something a ninjitsu master must feel: I had to spread my newfound &#13;\n\t\t\t\tknowledge to at least ten lost souls in the course of my lifetime. I took the usual path. I was rehashing the same arguments that &#13;\n\t\t\t\twere given to me for years (only now they actually made sense!), hoping to convert unsuspecting bystanders. It didn't work. My &#13;\n\t\t\t\tpersistence sparked a few people's interest but their curiosity dwindled at the mere sight of sample Lisp code. Perhaps years of &#13;\n\t\t\t\tadvocacy would forge a few new Lispers, but I wasn't satisfied. There had to be a better way.<\/p><p>I gave the matter careful thought. Is there something inherently hard about Lisp that prevents very intelligent, experienced &#13;\n\t\t\t\tprogrammers from understanding it? No, there isn't. After all, I got it, and if I can do it, anybody can. Then what is it that makes &#13;\n\t\t\t\tLisp so hard to understand? The answer, as such things usually do, came unexpectedly. Of course! Teaching anybody anything involves &#13;\n\t\t\t\tbuilding advanced concepts on top of concepts they already understand! If the process is made interesting and the matter is explained &#13;\n\t\t\t\tproperly the new concepts become as intuitive as the original building blocks that aided their understanding. That was the problem! &#13;\n\t\t\t\tMetaprogramming, code and data in one representation, self-modifying programs, domain specific mini-languages, none of the &#13;\n\t\t\t\texplanations for these concepts referenced familiar territory. How could I expect anyone to understand them! No wonder people wanted &#13;\n\t\t\t\tspecific examples. I could as well have been speaking in Martian!<\/p><p>I shared my ideas with fellow Lispers. \"Well, of course these concepts aren't explained in terms of familiar territory\", they said. &#13;\n\t\t\t\t\"They are so different, they're unlike anything these people have learned before.\" This was a poor excuse. \"I do not believe this to &#13;\n\t\t\t\tbe true\", I said. The response was unanimous: \"Why don't you give it a try?\" So I did. This article is a product of my efforts. It is &#13;\n\t\t\t\tmy attempt to explain Lisp in familiar, intuitive concepts. I urge brave souls to read on. Grab your favorite drink. Take a deep &#13;\n\t\t\t\tbreath. Prepare to be blown away. Oh, and may the Force be with you.<\/p><p class=\"first\">A thousand mile journey starts with a single step. A journey to enlightenment is no exception and our first step just happens to be &#13;\n\t\t\t\tXML. What more could possibly be said about XML that hasn't already been said? It turns out, quite a bit. While there's nothing &#13;\n\t\t\t\tparticularly interesting about XML itself, its relationship to Lisp is fascinating. XML is the all too familiar concept that Lisp &#13;\n\t\t\t\tadvocates need so much. It is our bridge to conveying understanding to regular programmers. So let's revive the dead horse, take out &#13;\n\t\t\t\tthe stick, and venture into XML wilderness that no one dared venture into before us. It's time to see the all too familiar moon from &#13;\n\t\t\t\tthe other side.<\/p><p>Superficially XML is nothing more than a standardized syntax used to express arbitrary hierarchical data in human readable form. &#13;\n\t\t\t\tTo-do lists, web pages, medical records, auto insurance claims, configuration files are all examples of potential XML use. Let's use &#13;\n\t\t\t\ta simple to-do list as an example (in a couple of sections you'll see it in a whole new light):<\/p><pre>&#13;\n&lt;todo name=\"housework\"&gt;&#13;\n    &lt;item priority=\"high\"&gt;Clean the house.&lt;\/item&gt;&#13;\n    &lt;item priority=\"medium\"&gt;Wash the dishes.&lt;\/item&gt;&#13;\n    &lt;item priority=\"medium\"&gt;Buy more soap.&lt;\/item&gt;&#13;\n&lt;\/todo&gt;&#13;\n<\/pre><p>What happens if we unleash our favorite XML parser on this to-do list? Once the data is parsed, how is it represented in memory? The &#13;\n\t\t\t\tmost natural representation is, of course, a tree - a perfect data structure for hierarchical data. After all is said and done, XML is &#13;\n\t\t\t\treally just a tree serialized to a human readable form. Anything that can be represented in a tree can be represented in XML and vice &#13;\n\t\t\t\tversa. I hope you understand this idea. It's very important for what's coming next.<\/p><p>Let's take this a little further. What other type of data is often represented as a tree? At this point the list is as good as &#13;\n\t\t\t\tinfinite so I'll give you a hint at what I'm getting at - try to remember your old compiler course. If you have a vague recollection &#13;\n\t\t\t\tthat source code is stored in a tree after it's parsed, you're on the right track. Any compiler inevitably parses the source code &#13;\n\t\t\t\tinto an abstract syntax tree. This isn't surprising since source code is hierarchical: functions contain arguments and blocks of code. &#13;\n\t\t\t\tBlocks of code contain expressions and statements. Expressions contain variables and operators. And so it goes.<\/p><p>Let's apply our corollary that any tree can easily be serialized into XML to this idea. If all source code is eventually represented &#13;\n\t\t\t\tas a tree, and any tree can be serialized into XML, then all source code can be converted to XML, right? Let's illustrate this &#13;\n\t\t\t\tinteresting property by a simple example. Consider the function below:<\/p><pre>\t\t\t\t&#13;\nint add(int arg1, int arg2)&#13;\n{&#13;\n    return arg1 + arg2;&#13;\n}&#13;\n<\/pre><p>Can you convert this function definition to its XML equivalent? Turns out, it's reasonably simple. Naturally there are many ways to &#13;\n\t\t\t\tdo this. Here is one way the resulting XML can look like:<\/p><pre>&#13;\n&lt;define-function return-type=\"int\" name=\"add\"&gt;&#13;\n    &lt;arguments&gt;&#13;\n        &lt;argument type=\"int\"&gt;arg1&lt;\/argument&gt;&#13;\n        &lt;argument type=\"int\"&gt;arg2&lt;\/argument&gt;&#13;\n    &lt;\/arguments&gt;&#13;\n    &lt;body&gt;&#13;\n        &lt;return&gt;&#13;\n            &lt;add value1=\"arg1\" value2=\"arg2\" \/&gt;&#13;\n        &lt;\/return&gt;&#13;\n    &lt;\/body&gt;&#13;\n&lt;\/define&gt;&#13;\n<\/pre><p>We can go through this relatively simple exercise with any language. We can turn any source code into XML, and we can transform the &#13;\n\t\t\t\tresulting XML back to original source code. We can write a converter that turns Java into XML and a converter that turns XML back to &#13;\n\t\t\t\tJava. We could do the same for C++. (In case you're wondering if anyone is crazy enough to do it, take a look at <a href=\"http:\/\/www.gccxml.org\">GCC-XML<\/a>). &#13;\n\t\t\t\tFurthermore, for languages that share common features but use different syntax (which to some extent is true about most mainstream &#13;\n\t\t\t\tlanguages) we could convert source code from one language to another using XML as an intermediary representation. We could use our &#13;\n\t\t\t\tJava2XML converter to convert a Java program to XML. We could then run an XML2CPP converter on the resulting XML and turn it into &#13;\n\t\t\t\tC++ code. With any luck (if we avoid using features of Java that don't exist in C++) we'll get a working C++ program. Neat, eh?<\/p><p>All this effectively means that we can use XML for generic storage of source code. We'd be able to create a whole class of &#13;\n\t\t\t\tprogramming languages that use uniform syntax, as well as write transformers that convert existing source code to XML. If we were to &#13;\n\t\t\t\tactually adopt this idea, compilers for different languages wouldn't need to implement parsers for their specific grammars - they'd &#13;\n\t\t\t\tsimply use an XML parser to turn XML directly into an abstract syntax tree.<\/p><p>By now you're probably wondering why I've embarked on the XML crusade and what it has to do with Lisp (after all, Lisp was created &#13;\n\t\t\t\tabout thirty years before XML). I promise that everything will become clear soon enough. But before we take our second step, let's go &#13;\n\t\t\t\tthrough a small philosophical exercise. Take a good look at the XML version of our \"add\" function above. How would you &#13;\n\t\t\t\tclassify it? Is it data or code? If you think about it for a moment you'll realize that there are good reasons to put this XML &#13;\n\t\t\t\tsnippet into both categories. It's XML and it's just information encoded in a standardized format. We've already determined that it &#13;\n\t\t\t\tcan be generated from a tree data structure in memory (that's effectively what GCC-XML does). It's lying around in a file with no &#13;\n\t\t\t\tapparent way to execute it. We can parse it into a tree of XML nodes and do various transformations on it. It's data. But wait a &#13;\n\t\t\t\tmoment! When all is said and done it's the same \"add\" function written with a different syntax, right? Once parsed, its tree could be &#13;\n\t\t\t\tfed into a compiler and we could execute it. We could easily write a small interpreter for this XML code and we could execute it &#13;\n\t\t\t\tdirectly. Alternatively, we could transform it into Java or C++ code, compile it, and run it. It's code.<\/p><p>So, where are we? Looks like we've just arrived to an interesting point. A concept that has traditionally been so hard to understand &#13;\n\t\t\t\tis now amazingly simple and intuitive. Code is also always data! Does it mean that data is also always code? As crazy as this sounds &#13;\n\t\t\t\tthis very well might be the case. Remember how I promised that you'll see our to-do list in a whole new light? Let me reiterate on &#13;\n\t\t\t\tthat promise. But we aren't ready to discuss this just yet. For now let's continue walking down our path.<\/p><p>A little earlier I mentioned that we could easily write an interpreter to execute our XML snippet of the add function. Of course this &#13;\n\t\t\t\tsounds like a purely theoretical exercise. Who in their right mind would want to do that for practical purposes? Well, it turns out &#13;\n\t\t\t\tquite a few people would disagree. You've likely encountered and used their work at least once in your career, too. Do I have you &#13;\n\t\t\t\tout on the edge of your seat? If so, let's move on!<\/p><p class=\"first\">Now that we've made the trip to the dark side of the moon, let's not leave quite yet. We may still learn something by exploring it a &#13;\n\t\t\t\tlittle more, so let's take another step. We begin by closing our eyes and remembering a cold rainy night in the winter of 2000. A &#13;\n\t\t\t\tprominent developer by the name of James Duncan Davidson<sup><a href=\"#note-james\">1<\/a><\/sup> was hacking his way through&#13;\n\t\t\t\t<a href=\"http:\/\/tomcat.apache.org\/\">Tomcat<\/a> servlet container. As the time came to build the changes he carefully saved all his files&#13;\n\t\t\t\tand ran <em>make<\/em>. Errors. Lots of errors. Something was wrong. After careful examination James exclaimed: \"Is my command not executing&#13;\n\t\t\t\tbecause I have a space in front of my tab?!\" Indeed, this was the problem. Again. James has had enough. He could sense the full moon through&#13;\n\t\t\t\tthe clouds and it made him adventurous. He created a fresh Java\tproject and quickly hacked together a simple but surprisingly useful&#13;\n\t\t\t\tutility. This spark of genius used Java property files for\tinformation on how to build the project. James could now write the equivalent&#13;\n\t\t\t\tof the makefile in a nice format without worrying about\tthe damned spaces ever again. His utility did all the hard work by interpreting&#13;\n\t\t\t\tthe property file and taking appropriate actions to build the project. It was neat. Another Neat Tool. <a href=\"http:\/\/ant.apache.org\/\">Ant<\/a>.<\/p><p>After using Ant to build Tomcat for a few months it became clear that Java property files are not sufficient to express complicated &#13;\n\t\t\t\tbuild instructions. Files needed to be checked out, copied, compiled, sent to another machine, and unit tested. In case of failure &#13;\n\t\t\t\te-mails needed to be sent out to appropriate people. In case of success \"Bad to the Bone\" needed to be played at the highest possible &#13;\n\t\t\t\tvolume. At the end of the track volume had to be restored to its original level. Yes, Java property files didn't cut it anymore. &#13;\n\t\t\t\tJames needed a more flexible solution. He didn't feel like writing his own parser (especially since he wanted an industry standard &#13;\n\t\t\t\tsolution). XML seemed like a reasonable alternative. In a couple of days Ant was ported to XML. It was the best thing since sliced &#13;\n\t\t\t\tbread.<\/p><p>So how does Ant work? It's pretty simple. It takes an XML file with specific build instructions (you decide if they're data or code) &#13;\n\t\t\t\tand interprets them by running specialized Java code for each XML element. It's actually much simpler than it sounds. A simple XML &#13;\n\t\t\t\tinstruction like the one below causes a Java class with an equivalent name to be loaded and its code to be executed.<\/p><pre>&#13;\n&lt;copy todir=\"..\/new\/dir\"&gt;&#13;\n    &lt;fileset dir=\"src_dir\"\/&gt;&#13;\n&lt;\/copy&gt;&#13;\n<\/pre><p>The snippet above copies a source directory to a destination directory. Ant locates a \"copy\" task (a Java class, really), sets &#13;\n\t\t\t\tappropriate parameters (todir and fileset) by calling appropriate Java methods and then executes the task. Ant comes with a set of &#13;\n\t\t\t\tcore tasks and anyone can extend it with tasks of their own simply by writing Java classes that follow certain conventions. Ant finds &#13;\n\t\t\t\tthese classes and executes them whenever XML elements with appropriate names are encountered. Pretty simple. Effectively Ant &#13;\n\t\t\t\taccomplishes what we were talking about in the previous section: it acts as an interpreter for a language that uses XML as its syntax &#13;\n\t\t\t\tby translating XML elements to appropriate Java instructions. We could write an \"add\" task and have Ant execute it when it encounters &#13;\n\t\t\t\tthe XML snippet for addition presented in the previous section! Considering that Ant is an extremely popular project, the ideas &#13;\n\t\t\t\tpresented in the previous section start looking more sane. After all, they're being used every day in what probably amounts to &#13;\n\t\t\t\tthousands of companies!<\/p><p>So far I've said nothing about why Ant actually goes through all the trouble of interpreting XML. Don't try to look for the answer on &#13;\n\t\t\t\tits website either - you'll find nothing of value. Nothing relevant to our discussion, anyway. Let's take another step. It's time to &#13;\n\t\t\t\tfind out why.<\/p><p class=\"first\">Sometimes right decisions are made without full conscious&#13;\n\t\t\t\tunderstanding of all the issues involved. I'm not sure if James&#13;\n\t\t\t\tknew why he chose XML - it was likely a subconscious decision.&#13;\n\t\t\t\tAt the very least, the reasons I saw on Ant's website for using&#13;\n\t\t\t\tXML are all the wrong reasons. It appears that the main concerns&#13;\n\t\t\t\trevolved around portability and extensibility. I fail to see how&#13;\n\t\t\t\tXML helps advance these goals in Ant's case. What is the&#13;\n\t\t\t\tadvantage of using interpreted XML over simple Java source code?&#13;\n\t\t\t\tWhy not create a set of classes with a nice API for commonly&#13;\n\t\t\t\tused tasks (copying directories, compiling, etc.) and using&#13;\n\t\t\t\tthose directly from Java source code? This would run on every&#13;\n\t\t\t\tplatform that runs Java (which Ant requires anyway), it's&#13;\n\t\t\t\tinfinitely extensible, and it has the benefit of having a more&#13;\n\t\t\t\tpleasant, familiar syntax. So why XML? Can we find a good reason&#13;\n\t\t\t\tfor using it?<\/p><p>It turns out that we can (although as I mentioned earlier I'm&#13;\n\t\t\t\tnot sure if James was consciously aware of it). XML has the&#13;\n\t\t\t\tproperty of being far more flexible in terms of introduction of&#13;\n\t\t\t\tsemantic constructs than Java could ever hope to be. Don't&#13;\n\t\t\t\tworry, I'm not falling into the trap of using big words to&#13;\n\t\t\t\tdescribe incomprehensible concepts. This is actually a&#13;\n\t\t\t\trelatively simple idea, though it may take some effort to&#13;\n\t\t\t\texplain. Buckle your seat-belt. We're about to make a giant leap&#13;\n\t\t\t\ttowards achieving nirvana.<\/p><p>How can we represent 'copy' example above in Java code? Here's one way to do it:<\/p><pre>&#13;\nCopyTask copy = new CopyTask();&#13;\nFileset fileset = new Fileset();&#13;\n&#13;\nfileset.setDir(\"src_dir\");&#13;\ncopy.setToDir(\"..\/new\/dir\");&#13;\ncopy.setFileset(fileset);&#13;\n&#13;\ncopy.execute();&#13;\n<\/pre><p>The code is almost the same, albeit a little longer than the original XML. So&#13;\n\t\t\t\twhat's different? The answer is that the XML snippet introduces a&#13;\n\t\t\t\tspecial semantic construct for copying. If we could do it in Java it would look&#13;\n\t\t\t\tlike this:<\/p><pre>\t\t&#13;\ncopy(\"..\/new\/dir\")&#13;\n{&#13;\n    fileset(\"src_dir\");&#13;\n}&#13;\n<\/pre><p>Can you see the difference? The code above (if it were possible&#13;\n\t\t\t\tin Java) is a special operator for copying files - similar to a&#13;\n\t\t\t\t<em>for<\/em> loop or a new <em>foreach<\/em> construct introduced in Java 5. If we&#13;\n\t\t\t\thad an automatic converter from XML to Java it would likely&#13;\n\t\t\t\tproduce the above gibberish. The reason for this is that Java's&#13;\n\t\t\t\taccepted syntax tree grammar is fixed by the language&#13;\n\t\t\t\tspecification - we have no way of modifying it. We can add&#13;\n\t\t\t\tpackages, classes, methods, but we cannot extend Java to make&#13;\n\t\t\t\taddition of new operators possible. Yet we can do it to our&#13;\n\t\t\t\theart's content in XML - its syntax tree isn't restricted by&#13;\n\t\t\t\tanything except our interpreter! If the idea is still unclear,&#13;\n\t\t\t\tconsider introducing a special operator 'unless' to Java:<\/p><pre>\t\t\t\t&#13;\nunless(someObject.canFly())&#13;\n{&#13;\n    someObject.transportByGround();&#13;\n}&#13;\n<\/pre><p>In the previous two examples we extend the Java language to introduce an&#13;\n\t\t\t\toperator for copying files and a conditional operator <em>unless<\/em>. We would do this&#13;\n\t\t\t\tby modifying the abstract syntax tree grammar that Java compiler accepts.&#13;\n\t\t\t\tNaturally we cannot do it with standard Java facilities, but we can easily do it&#13;\n\t\t\t\tin XML. Because our XML interpreter parses the abstract syntax tree that results&#13;\n\t\t\t\tfrom it, we can extend it to include any operator we like.<\/p><p>For complex operators this ability provides tremendous benefits. Can you imagine&#13;\n\t\t\t\twriting special operators for checking out source code, compiling files, running&#13;\n\t\t\t\tunit testing, sending email? Try to come up with some. If you're dealing with a&#13;\n\t\t\t\tspecialized problem (in our case it's building projects) these operators can do&#13;\n\t\t\t\twonders to decrease the amount of code you have to type and to increase clarity&#13;\n\t\t\t\tand code reuse. Interpreted XML makes this extremely easy to accomplish because&#13;\n\t\t\t\tit's a simple data file that stores hierarchical data. We do not have this option&#13;\n\t\t\t\tin Java because it's hierarchical structure is fixed (as you will soon find out,&#13;\n\t\t\t\twe do have this option in Lisp). Perhaps this is one of the reasons why Ant is&#13;\n\t\t\t\tso successful?<\/p><p>I urge you to take a look at recent evolution of Java and C# (especially the&#13;\n\t\t\t\trecently released specification for C# 3.0). The languages are being evolved by&#13;\n\t\t\t\tabstracting away commonly used functionality and adding it in the form of&#13;\n\t\t\t\toperators. New C# operators for built-in queries is one example. This is&#13;\n\t\t\t\taccomplished by relatively traditional means: language creators modify the&#13;\n\t\t\t\taccepted abstract syntax tree and add implementations of certain features.&#13;\n\t\t\t\tImagine the possibilities if the programmer could modify the abstract syntax&#13;\n\t\t\t\ttree himself! Whole new sub-languages could be built for specialized domains (for&#13;\n\t\t\t\texample a language for building projects, like Ant). Can you come up with other&#13;\n\t\t\t\texamples? Think about these concepts for a bit, but don't worry about them too&#13;\n\t\t\t\tmuch. We'll come back to these issues after introducing a few more ideas. By&#13;\n\t\t\t\tthen things will be a little more clear.<\/p><p class=\"first\">Let's forget about the operator business for the moment and try to expand our&#13;\n\t\t\t\thorizons beyond the constraints of Ant's design. I mentioned earlier that Ant&#13;\n\t\t\t\tcan be extended by writing conventional Java classes. Ant interpreter then&#13;\n\t\t\t\tattempts to match XML elements to appropriately named Java classes and if the&#13;\n\t\t\t\tmatch is found the task is executed. An interesting question begs to be asked.&#13;\n\t\t\t\tWhy not extend Ant in Ant itself? After all, core tasks contain a lot of&#13;\n\t\t\t\tconventional programming language constructs ('if' being a perfect example). If&#13;\n\t\t\t\tAnt provided constructs to develop tasks in Ant itself we'd reach a higher&#13;\n\t\t\t\tdegree of portability. We'd be dependent on a core set of tasks (a standard&#13;\n\t\t\t\tlibrary, if you will) and we wouldn't care if Java runtime is present: the core&#13;\n\t\t\t\tset could be implemented in anything. The rest of the tasks would be built on&#13;\n\t\t\t\ttop of the core using Ant-XML itself. Ant would then become a generic,&#13;\n\t\t\t\textensible, XML-based programming language. Consider the possibilities:<\/p><pre>&#13;\n&lt;task name=\"Test\"&gt;&#13;\n    &lt;echo message=\"Hello World!\"\/&gt;&#13;\n&lt;\/task&gt;&#13;\n&lt;Test \/&gt;&#13;\n<\/pre><p>If ant supported the \"task\" construct, the example above would print \"Hello&#13;\n\t\t\t\tWorld!\". In fact, we could write a \"task\" task in Java and make Ant able to&#13;\n\t\t\t\textend itself using Ant-XML! Ant would then be able to build more complicated&#13;\n\t\t\t\tprimitives on top of simple ones, just like any other programming language! This&#13;\n\t\t\t\tis an example of \"XML\" based programming language we were talking about in the&#13;\n\t\t\t\tbeginning of this tutorial. Not very useful (can you tell why?) but pretty damn&#13;\n\t\t\t\tcool.<\/p><p>By the way, take a look at our 'Test' task once again. Congratulations. You're&#13;\n\t\t\t\tlooking at Lisp code. What on Earth am I talking about? It doesn't look anything&#13;\n\t\t\t\tlike Lisp? Don't worry, we'll fix that in a bit. Confused? Good. Let's clear it&#13;\n\t\t\t\tall up!<\/p><p class=\"first\">I mentioned in the previous section that self-extending Ant wouldn't be very&#13;\n\t\t\t\tuseful. The reason for that is XML's verbosity. It's not too bad for data files&#13;\n\t\t\t\tbut the moment you try writing reasonably complex code the amount of typing you&#13;\n\t\t\t\thave to do quickly starts to get in the way and progresses to becoming unusable&#13;\n\t\t\t\tfor any real project. Have you ever tried writing Ant build scripts? I have, and&#13;\n\t\t\t\tonce they get complex enough having to do it in XML becomes really annoying.&#13;\n\t\t\t\tImagine having to type almost everything in Java twice because you have to close&#13;\n\t\t\t\tevery element. Wouldn't that drive you nuts?<\/p><p>The solution to this problem involves using a less verbose alternative to XML.&#13;\n\t\t\t\tRemember, XML is just a format for representing hierarchical data. We don't have&#13;\n\t\t\t\tto use XML's angle brackets to serialize trees. We could come up with many other&#13;\n\t\t\t\tformats. One such format (incidentally, the one Lisp uses) is called an&#13;\n\t\t\t\ts-expression. S-expressions accomplish the same goals as XML. They're just a lot&#13;\n\t\t\t\tless verbose, which makes them much better suited for typing code. I will&#13;\n\t\t\t\texplain s-expressions in a little while, but before I do I have to clear up a&#13;\n\t\t\t\tfew things about XML. Let's consider our XML example for copying files:<\/p><pre>&#13;\n&lt;copy todir=\"..\/new\/dir\"&gt;&#13;\n    &lt;fileset dir=\"src_dir\"\/&gt;&#13;\n&lt;\/copy&gt;&#13;\n<\/pre><p>Think of what the parse tree of this snippet would look like in memory. We'd&#13;\n\t\t\t\thave a 'copy' node that contains a fileset node. But what about attributes? How&#13;\n\t\t\t\tdo they fit into our picture? If you've ever used XML to describe data and&#13;\n\t\t\t\twondered whether you should use an element or an attribute, you're not alone.&#13;\n\t\t\t\tNobody can really figure this out and doing it right tends to be black magic&#13;\n\t\t\t\trather than science. The reason for that is that attributes are really subsets&#13;\n\t\t\t\tof elements. Anything attributes can do, elements can do as well. The reason&#13;\n\t\t\t\tattributes were introduced is to curb XML's verbosity. Take a look at another&#13;\n\t\t\t\tversion of our 'copy' snippet:<\/p><pre>&#13;\n&lt;copy&gt;&#13;\n    &lt;todir&gt;..\/new\/dir&lt;\/todir&gt;&#13;\n    &lt;fileset&gt;&#13;\n        &lt;dir&gt;src_dir&lt;\/dir&gt;&#13;\n    &lt;\/fileset&gt;&#13;\n&lt;\/copy&gt;&#13;\n<\/pre><p>The two snippets hold exactly the same information. However, we use attributes&#13;\n\t\t\t\tto avoid typing the same thing more than once. Imagine if attributes weren't&#13;\n\t\t\t\tpart of XML specification. Writing anything in XML would drive us nuts!<\/p><p>Now that we got attributes out of the way, let's look at s-expressions. The&#13;\n\t\t\t\treason we took this detour is that s-expressions do not have attributes. Because&#13;\n\t\t\t\tthey're a lot less verbose, attributes are simply unnecessary. This is one thing&#13;\n\t\t\t\twe need to keep in mind when transforming XML to s-expressions. Let's take a&#13;\n\t\t\t\tlook at an example. We could translate above snippet to s-expressions like this:<\/p><pre>&#13;\n(copy&#13;\n    (todir \"..\/new\/dir\")&#13;\n    (fileset (dir \"src_dir\")))&#13;\n<\/pre><p>Take a good look at this representation. What's different? Angle brackets seem&#13;\n\t\t\t\tto be replaced by parentheses. Instead of enclosing each element into a pair of&#13;\n\t\t\t\tparentheses and then closing each element with a \"(\/element)\" we simply skip the&#13;\n\t\t\t\tsecond parenthesis in \"(element\" and proceed. The element is then closed like&#13;\n\t\t\t\tthis: \")\". That's it! The translation is natural and very simple. It's also a&#13;\n\t\t\t\tlot easier to type. Do parentheses blind first time users? Maybe, but now that&#13;\n\t\t\t\twe're understand the reasoning behind them they're a lot easier to handle. At&#13;\n\t\t\t\tthe very least they're better than arthritis inducing verbosity of XML. After&#13;\n\t\t\t\tyou get used to s-expressions writing code in them is not only doable but very&#13;\n\t\t\t\tpleasant. And they provide all the benefits of writing code in XML (many of&#13;\n\t\t\t\twhich we're yet to explore). Let's take a look at our 'task' code in something&#13;\n\t\t\t\tthat looks a lot more like lisp:<\/p><pre>&#13;\n(task (name \"Test\")&#13;\n    (echo (message \"Hello World!\")))&#13;\n&#13;\n(Test)&#13;\n<\/pre><p>S-expressions are called lists in Lisp lingo. Consider our 'task' element above.&#13;\n\t\t\t\tIf we rewrite it without a line break and with comas instead of spaces it's&#13;\n\t\t\t\tstarting to look surprisingly like a list of elements and other lists (the formatting&#13;\n\t\t\t\tis added to make it easier to see nested lists):<\/p><pre>&#13;\n(task, (name, \"test\"), (echo, (message, \"Hello World!\")))&#13;\n<\/pre><p>We could do the same with XML. Of course the line above isn't really a list,&#13;\n\t\t\t\tit's a tree, just like its XML-alternative. Don't let references to lists&#13;\n\t\t\t\tconfuse you, it's just that lists that contain other lists and trees are&#13;\n\t\t\t\teffectively the same thing. Lisp may stand for List Processing, but it's really&#13;\n\t\t\t\ttree processing - no different than processing XML nodes.<\/p><p>Whew. After much rambling we finally got to something that looks like Lisp (and&#13;\n\t\t\t\tis Lisp, really). By now the mysterious Lisp parentheses as well as some claims&#13;\n\t\t\t\tmade by Lisp advocates should become more clear. But we still have a lot of&#13;\n\t\t\t\tground to cover. Ready? Let's move on!<\/p><p class=\"first\">By now you must be tired of all the XML talk. I'm tired of it as well. It's time&#13;\n\t\t\t\tto take a break from all the trees, s-expressions, and Ant business. Instead,&#13;\n\t\t\t\tlet's go back to every programmer's roots. It's time to talk about C&#13;\n\t\t\t\tpreprocessor. What's C got to do with anything, I hear you ask? Well, we now&#13;\n\t\t\t\tknow enough to get into metaprogramming and discuss code that writes other code.&#13;\n\t\t\t\tUnderstanding this tends to be hard since all tutorials discuss it in terms of&#13;\n\t\t\t\tlanguages that you don't know. But there is nothing hard about the concept. I&#13;\n\t\t\t\tbelieve that a metaprogramming discussion based on C will make the whole thing&#13;\n\t\t\t\tmuch easier to understand. So, let's see (pun intended).<\/p><p>Why would anyone want to write a program that writes programs? How can we use&#13;\n\t\t\t\tsomething like this in the real world? What on Earth is metaprogramming, anyway?&#13;\n\t\t\t\tYou already know all the answers, you just don't know it yet. In order to unlock&#13;\n\t\t\t\tthe hidden vault of divine knowledge let's consider a rather mundane task of&#13;\n\t\t\t\tsimple database access from code. We've all been there. Writing SQL queries all&#13;\n\t\t\t\tover the code to modify data within tables turns into repetitive hell soon&#13;\n\t\t\t\tenough. Even with the new C# 3.0 LINQ stuff this is a huge pain. Writing a full&#13;\n\t\t\t\tSQL query (albeit with a nice built in syntax) to get someone's name or to&#13;\n\t\t\t\tmodify someone's address isn't exactly a programmer's idea of comfort. What do&#13;\n\t\t\t\twe do to solve these problems? Enter data access layers.<\/p><p>The idea is simple enough. You abstract database access (at least trivial&#13;\n\t\t\t\tqueries, anyway) by creating a set of classes that mirror the tables in the&#13;\n\t\t\t\tdatabase and use accessor methods to execute actual queries. This simplifies&#13;\n\t\t\t\tdevelopment tremendously - instead of writing SQL queries we make simple method&#13;\n\t\t\t\tcalls (or property assignments, depending on your language of choice). Anyone&#13;\n\t\t\t\twho has ever used even the simplest of data access layers knows how much time it&#13;\n\t\t\t\tcan save. Of course anyone who has ever written one knows how much time it can&#13;\n\t\t\t\tkill - writing a set of classes that mirror tables and convert accessors to SQL&#13;\n\t\t\t\tqueries takes a considerable chunk of time. This seems especially silly since&#13;\n\t\t\t\tmost of the work is manual: once you figure out the design and develop a&#13;\n\t\t\t\ttemplate for your typical data access class you don't need to do any thinking.&#13;\n\t\t\t\tYou just write code based on the same template over and over and over and over&#13;\n\t\t\t\tagain. Many people figured out that there is a better way - there are plenty of&#13;\n\t\t\t\ttools that connect to the database, grab the schema, and write code for you based&#13;\n\t\t\t\ton a predefined (or a custom) template.<\/p><p>Anyone who has ever used such a tool knows what an amazing time saver it can be.&#13;\n\t\t\t\tIn a few clicks you connect the tool to the database, get it to generate the&#13;\n\t\t\t\tdata access layer source code, add the files to your project and voil\u00e0 - ten&#13;\n\t\t\t\tminutes worth of work do a better job than hundreds of man-hours that were&#13;\n\t\t\t\trequired previously. What happens if your database schema changes? Well, you&#13;\n\t\t\t\tjust have to go through this short process again. Of course some of the best&#13;\n\t\t\t\ttools let you automate this - you simply add them as a part of your build step&#13;\n\t\t\t\tand every time you compile your project everything is done for you&#13;\n\t\t\t\tautomatically. This is perfect! You barely have to do anything at all. If the&#13;\n\t\t\t\tschema ever changes your data access layer code updates automatically at compile&#13;\n\t\t\t\ttime and any obsolete access in your code will result in compiler errors!<\/p><p>Data access layers are one good example, but there are plenty of others. From&#13;\n\t\t\t\tboilerplate GUI code, to web code, to COM and CORBA stubs, to MFC and ATL, -&#13;\n\t\t\t\tthere are plenty of examples where the same code is written over and over again.&#13;\n\t\t\t\tSince writing this code is a task that can be automated completely and a&#13;\n\t\t\t\tprogrammer's time is far more expensive than CPU time, plenty of tools have been&#13;\n\t\t\t\tcreated that generate this boilerplate code automatically. What are these tools,&#13;\n\t\t\t\texactly? Well, they are programs that write programs. They perform a simple task&#13;\n\t\t\t\tthat has a mysterious name of metaprogramming. That's all there is to it.<\/p><p>We could create and use such tools in millions of scenarios but more often than&#13;\n\t\t\t\tnot we don't. What it boils down to is a subconscious calculation - is it worth it&#13;\n\t\t\t\tfor me to create a separate project, write a whole tool to generate something,&#13;\n\t\t\t\tand then use it, if I only have to write these very similar pieces about seven&#13;\n\t\t\t\ttimes? Of course not. Data access layers and COM stubs are written hundreds,&#13;\n\t\t\t\tthousands of times. This is why there are tools for them. For similar pieces of&#13;\n\t\t\t\tcode that repeat only a few times, or even a few dozen times, writing code&#13;\n\t\t\t\tgeneration tools isn't even considered. The trouble to create such a tool more&#13;\n\t\t\t\toften than not far outweighs the benefit of using one. If only creating such&#13;\n\t\t\t\ttools was much easier, we could use them more often, and perhaps save many hours&#13;\n\t\t\t\tof our time. Let's see if we can accomplish this in a reasonable manner.<\/p><p>Surprisingly C preprocessor comes to the rescue. We've all used it in C and C++.&#13;\n\t\t\t\tOn occasion we all wish Java had it. We use it to execute simple instructions&#13;\n\t\t\t\tat compile time to make small changes to our code (like selectively removing&#13;\n\t\t\t\tdebug statements). Let's look at a quick example:<\/p><pre>&#13;\n#define triple(X)    X + X + X&#13;\n<\/pre><p>What does this line do? It's a simple instruction written in the preprocessor&#13;\n\t\t\t\tlanguage that instructs it to replace all instances of <em>triple(X)<\/em> with <em>X + X +&#13;\n\t\t\t\tX<\/em>. For example all instances of '<em>triple(5)<\/em>' will be replaced with '<em>5 + 5 + 5<\/em>'&#13;\n\t\t\t\tand the resulting code will be compiled by the C compiler. We're really doing a&#13;\n\t\t\t\tvery primitive version of code generation here. If only C preprocessor was a&#13;\n\t\t\t\tlittle more powerful and included ways to connect to the database and a few more&#13;\n\t\t\t\tsimple constructs, we could use it to develop our data access layer right there,&#13;\n\t\t\t\tfrom within our program! Consider the following example that uses an imaginary&#13;\n\t\t\t\textension of the C preprocessor:<\/p><pre>&#13;\n#get-db-schema(\"127.0.0.1, un, pwd\");&#13;\n#iterate-through-tables&#13;\n#for-each-table&#13;\n    class #table-name&#13;\n    {&#13;\n    };&#13;\n#end-for-each&#13;\n<\/pre><p>We've just connected to the database schema, iterated through all the tables,&#13;\n\t\t\t\tand created an empty class for each. All in a couple of lines right within our&#13;\n\t\t\t\tsource code! Now every time we recompile the file where above code appears we'll&#13;\n\t\t\t\tget a freshly built set of classes that automatically update based on the&#13;\n\t\t\t\tschema. With a little imagination you can see how we could build a full data&#13;\n\t\t\t\taccess layer straight from within our program, without the use of any external&#13;\n\t\t\t\ttools! Of course this has a certain disadvantage (aside from the fact that such&#13;\n\t\t\t\tan advanced version of C preprocessor doesn't exist) - we'd have to learn a&#13;\n\t\t\t\twhole new \"compile-time language\" to do this sort of work. For complex code&#13;\n\t\t\t\tgeneration this language would have to be very complex as well, it would have to&#13;\n\t\t\t\tsupport many libraries and language constructs. For example, if our generated&#13;\n\t\t\t\tcode depended on some file located at some ftp server the preprocessor would&#13;\n\t\t\t\thave to be able to connect to ftp. It's a shame to create and learn a new&#13;\n\t\t\t\tlanguage just to do this. Especially since there are so many nice languages&#13;\n\t\t\t\talready out there. Of course if we add a little creativity we can easily avoid&#13;\n\t\t\t\tthis pitfall.<\/p><p>Why not replace the preprocessor language with C\/C++ itself? We'd have full&#13;\n\t\t\t\tpower of the language at compile time and we'd only need to learn a few simple&#13;\n\t\t\t\tdirectives to differentiate between compile time and runtime code!<\/p><pre>&#13;\n&lt;%&#13;\n    cout &lt;&lt; \"Enter a number: \";&#13;\n    cin &gt;&gt; n;&#13;\n%&gt;&#13;\nfor(int i = 0; i &lt; &lt;%= n %&gt;; i++)&#13;\n{&#13;\n    cout &lt;&lt; \"hello\" &lt;&lt; endl;&#13;\n}&#13;\n<\/pre><p>Can you see what happens here? Everything that's between &lt;% and %&gt; tags&#13;\n\t\t\t\truns when the program is compiled. Anything outside of these tags is normal&#13;\n\t\t\t\tcode. In the example above you'd start compiling your program in the development&#13;\n\t\t\t\tenvironment. The code between the tags would be compiled and then ran. You'd get&#13;\n\t\t\t\ta prompt to enter a number. You'd enter one and it would be placed inside the&#13;\n\t\t\t\tfor loop. The for loop would then be compiled as usual and you'd be able to&#13;\n\t\t\t\texecute it. For example, if you'd enter 5 during the compilation of your&#13;\n\t\t\t\tprogram, the resulting code would look like this:<\/p><pre>&#13;\nfor(int i = 0; i &lt; 5; i++)&#13;\n{&#13;\n    cout &lt;&lt; \"hello\" &lt;&lt; endl;&#13;\n}&#13;\n<\/pre><p>Simple and effective. No need for a special preprocessor language. We get full&#13;\n\t\t\t\tpower of our host language (in this case C\/C++) at compile time. We could easily&#13;\n\t\t\t\tconnect to a database and generate our data access layer source code at compile&#13;\n\t\t\t\ttime in the same way JSP or ASP generate HTML! Creating such tools would also be&#13;\n\t\t\t\ttremendously quick and simple. We'd never have to create new projects with&#13;\n\t\t\t\tspecialized GUIs. We could inline our tools right into our programs. We wouldn't&#13;\n\t\t\t\thave to worry about whether writing such tools is worth it because writing them&#13;\n\t\t\t\twould be so fast - we could save tremendous amounts of time by creating simple&#13;\n\t\t\t\tbits of code that do mundane code generation for us!<\/p><p class=\"first\">Everything we've learned about Lisp so far can be summarized by a single&#13;\n\t\t\t\tstatement: Lisp is executable XML with a friendlier syntax. We haven't said a&#13;\n\t\t\t\tsingle word about how Lisp actually operates. It's time to fill this&#13;\n\t\t\t\tgap<sup><a href=\"#note-blaise\">2<\/a><\/sup>.<\/p><p>Lisp has a number of built in data types. Integers and strings, for example,&#13;\n\t\t\t\taren't much different from what you're used to. The meaning of <em>71<\/em> or <em>\"hello\"<\/em> is&#13;\n\t\t\t\troughly the same in Lisp as in C++ or Java. What is of more interest to us are&#13;\n\t\t\t\t<em>symbols<\/em>, <em>lists<\/em>, and <em>functions<\/em>. I will spend the rest of this section describing&#13;\n\t\t\t\tthese data types as well as how a Lisp environment compiles and executes the&#13;\n\t\t\t\tsource code you type into it (this is called <em>evaluation<\/em> in Lisp lingo). Getting&#13;\n\t\t\t\tthrough this section in one piece is important for understanding true potential of&#13;\n\t\t\t\tLisp's metaprogramming, the unity of code and data, and the notion of domain&#13;\n\t\t\t\tspecific languages. Don't think of this section as a chore though, I'll try to&#13;\n\t\t\t\tmake it fun and accessible. Hopefully you can pick up a few interesting ideas on&#13;\n\t\t\t\tthe way. Ok. Let's start with Lisp's symbols.<\/p><p>A symbol in Lisp is roughly equivalent to C++ or Java's notion of an identifier.&#13;\n\t\t\t\tIt's a name you can use to access a variable (like <em>currentTime<\/em>, <em>arrayCount<\/em>, <em>n<\/em>,&#13;\n\t\t\t\tetc.) The difference is that a symbol in Lisp is a lot more liberal than its&#13;\n\t\t\t\tmainstream identifier alternative. In C++ or Java you're limited to&#13;\n\t\t\t\talphanumeric characters and an underscore. In Lisp, you are not. For example <em>+<\/em>&#13;\n\t\t\t\tis a valid symbol. So is <em>-<\/em>, <em>=<\/em>, <em>hello-world<\/em>, <em>hello+world<\/em>, <em>*<\/em>, etc. (you can find&#13;\n\t\t\t\tthe exact definition of valid Lisp symbols online). You can assign to these&#13;\n\t\t\t\tsymbols any data-type you like. Let's ignore Lisp syntax and use pseudo-code for&#13;\n\t\t\t\tnow. Assume that a function <em>set<\/em> assigns some value to a symbol (like <em>=<\/em> does&#13;\n\t\t\t\tin Java or C++). The following are all valid examples:<\/p><pre>&#13;\nset(test, 5)            \/\/ symbol 'test' will equal an integer 5&#13;\nset(=, 5)               \/\/ symbol '=' will equal an integer 5&#13;\nset(test, \"hello\")      \/\/ symbol 'test' will equal a string \"hello\"&#13;\nset(test, =)            \/\/ at this point symbol '=' is equal to 5&#13;\n                        \/\/ therefore symbol 'test' will equal to 5&#13;\nset(*, \"hello\")         \/\/ symbol '*' will equal a string \"hello\"&#13;\n<\/pre><p>At this point something must smell wrong. If we can assign strings and integers&#13;\n\t\t\t\tto symbols like <em>*<\/em>, how does Lisp do multiplication? After all, <em>*<\/em> means&#13;\n\t\t\t\tmultiply, right? The answer is pretty simple. Functions in Lisp aren't special.&#13;\n\t\t\t\tThere is a data-type, function, just like integer and string, that you assign to&#13;\n\t\t\t\tsymbols. A multiplication function is built into Lisp and is assigned to a&#13;\n\t\t\t\tsymbol <em>*<\/em>. You can reassign a different value to <em>*<\/em> and you'd lose the&#13;\n\t\t\t\tmultiplication function. Or you can store the value of the function in some&#13;\n\t\t\t\tother variable. Again, using pseudo-code:<\/p><pre>&#13;\n*(3, 4)                 \/\/ multiplies 3 by 4, resulting in 12&#13;\nset(temp, *)            \/\/ symbol '*' is equal to the multiply function&#13;\n                        \/\/ so temp will equal to the multiply function&#13;\nset(*, 3)               \/\/ sets symbol '*' to equal to 3&#13;\n*(3, 4)                 \/\/ error, symbol '*' no longer equals to a function&#13;\n                        \/\/ it's equal to 3&#13;\ntemp(3, 4)              \/\/ temp equals to a multiply function&#13;\n                        \/\/ so Lisp multiplies 3 by 4 resulting in 12&#13;\nset(*, temp)            \/\/ symbol '*' equals multiply function again&#13;\n*(3, 4)                 \/\/ multiplies 3 by 4, resulting in 12&#13;\n<\/pre><p>You can even do wacky stuff like reassigning plus to minus:<\/p><pre>&#13;\nset(+, -)               \/\/ the value of '-' is a built in minus function&#13;\n                        \/\/ so now symbol '+' equals to a minus function&#13;\n+(5, 4)                 \/\/ since symbol '+' is equal to the minus function&#13;\n                        \/\/ this results in 1&#13;\n<\/pre><p>I've used functions quite liberally in these examples but I didn't describe them&#13;\n\t\t\t\tyet. A function in Lisp is just a data-type like an integer, a string, or a&#13;\n\t\t\t\tsymbol. A function doesn't have a notion of a name like in Java or C++. Instead, it&#13;\n\t\t\t\tstands on its own. Effectively it is a pointer to a block of code along with&#13;\n\t\t\t\tsome information (like a number of parameters it accepts). You only give the&#13;\n\t\t\t\tfunction a name by assigning it to a symbol, just like you assign an integer or&#13;\n\t\t\t\ta string. You can create a function by using a built in function for creating&#13;\n\t\t\t\tfunctions, assigned to a symbol 'fn'. Using pseudo-code:<\/p><pre>&#13;\nfn [a]&#13;\n{&#13;\n    return *(a, 2);&#13;\n}&#13;\n<\/pre><p>This returns a function that takes a single parameter named <em>'a'<\/em> and doubles it.&#13;\n\t\t\t\tNote that the function has no name but you can assign it to a symbol:<\/p><pre>&#13;\nset(times-two, fn [a] { return *(a, 2); })&#13;\n<\/pre><p>We can now call this function:<\/p><pre>&#13;\ntimes-two(5)             \/\/ returns 10&#13;\n<\/pre><p>Now that we went over symbols and functions, what about lists? Well, you already&#13;\n\t\t\t\tknow a lot about them. Lists are simply pieces of XML written in s-expression&#13;\n\t\t\t\tform. A list is specified by parentheses and contains Lisp data-types (including&#13;\n\t\t\t\tother lists) separated by a space. For example (this is real Lisp, note that we&#13;\n\t\t\t\tuse semicolons for comments now):<\/p><pre>&#13;\n()                      ; an empty list&#13;\n(1)                     ; a list with a single element, 1&#13;\n(1 \"test\")              ; a list with two elements&#13;\n                        ; an integer 1 and a string \"test\"&#13;\n(test \"hello\")          ; a list with two elements&#13;\n                        ; a symbol test and a string \"hello\"&#13;\n(test (1 2) \"hello\")    ; a list with three elements, a symbol test&#13;\n                        ; a list of two integers 1 and 2&#13;\n                        ; and a string \"hello\"&#13;\n<\/pre><p>When a Lisp system encounters lists in the source code it acts exactly like Ant&#13;\n\t\t\t\tdoes when it encounters XML - it attempts to execute them. In fact, Lisp source&#13;\n\t\t\t\tcode is only specified using lists, just like Ant source code is only specified&#13;\n\t\t\t\tusing XML. Lisp executes lists in the following manner. The first element of the&#13;\n\t\t\t\tlist is treated as the name of a function. The rest of the elements are treated&#13;\n\t\t\t\tas functions parameters. If one of the parameters is another list it is&#13;\n\t\t\t\texecuted using the same principles and the result is passed as a parameter to&#13;\n\t\t\t\tthe original function. That's it. We can write real code now:<\/p><pre>&#13;\n(* 3 4)                 ; equivalent to pseudo-code *(3, 4).&#13;\n                        ; Symbol '*' is a function&#13;\n                        ; 3 and 4 are its parameters.&#13;\n                        ; Returns 12.&#13;\n(times-two 5)           ; returns 10&#13;\n(3 4)                   ; error: 3 is not a function&#13;\n(times-two)              ; error, times-two expects one parameter&#13;\n(times-two 3 4)          ; error, times-two expects one parameter&#13;\n(set + -)               ; sets symbol '+' to be equal to whatever symbol '-'&#13;\n                        ; equals to, which is a minus function&#13;\n(+ 5 4)                 ; returns 1 since symbol '+' is now equal&#13;\n                        ; to the minus function&#13;\n(* 3 (* 2 2))           ; multiplies 3 by the second parameter&#13;\n                        ; (which is a function call that returns 4).&#13;\n                        ; Returns 12.&#13;\n<\/pre><p>Note that so far every list we've specified was treated by a Lisp system as&#13;\n\t\t\t\tcode. But how can we treat a list as data? Again, imagine an Ant task that&#13;\n\t\t\t\taccepts XML as one of its parameters. In Lisp we do this using a quote operator&#13;\n\t\t\t\t<em>'<\/em> like so:<\/p><pre>&#13;\n(set test '(1 2))       ; test is equal to a list of two integers, 1 and 2&#13;\n(set test (1 2))        ; error, 1 is not a function&#13;\n(set test '(* 3 4))     ; sets test to a list of three elements,&#13;\n                        ; a symbol *, an integer 3, and an integer 4&#13;\n<\/pre><p>We can use a built in function <em>head<\/em> to return the first element of the&#13;\n\t\t\t\tlist, and a built in function <em>tail<\/em> to return the rest of the list's&#13;\n\t\t\t\telements:<\/p><pre>&#13;\n(head '(* 3 4))         ; returns a symbol '*'&#13;\n(tail '(* 3 4))         ; returns a list (3 4)&#13;\n(head (tail '( * 3 4))) ; (tail '(* 3 4)) returns a list (3 4)&#13;\n                        ; and (head '(3 4)) returns 3.&#13;\n(head test)             ; test was set to a list in previous example&#13;\n                        ; returns a symbol '*'&#13;\n<\/pre><p>You can think of built in Lisp functions as you think of Ant tasks. The difference&#13;\n\t\t\t\tis that we don't have to extend Lisp in another language (although we can), we can&#13;\n\t\t\t\textend it in Lisp itself as we did with the <em>times-two<\/em> example. Lisp comes with a very&#13;\n\t\t\t\tcompact set of built in functions - the necessary minimum. The rest of the language&#13;\n\t\t\t\tis implemented as a standard library in Lisp itself.<\/p><p class=\"first\">So far we've looked at metaprogramming in terms of a simple templating&#13;\n\t\t\t\tengine similar to JSP. We've done code generation using simple string&#13;\n\t\t\t\tmanipulations. This is generally how most code generation tools go&#13;\n\t\t\t\tabout doing this task. But we can do much better. To get on the&#13;\n\t\t\t\tright track, let's start off with a question. How would we write a&#13;\n\t\t\t\ttool that automatically generates Ant build scripts by looking at&#13;\n\t\t\t\tsource files in the directory structure?<\/p><p>We could take the easy way out and generate Ant XML by manipulating&#13;\n\t\t\t\tstrings. Of course a much more abstract, expressive and extensible way&#13;\n\t\t\t\tis to work with XML processing libraries to generate XML nodes&#13;\n\t\t\t\tdirectly in memory. The nodes can then be serialized to strings&#13;\n\t\t\t\tautomatically. Furthermore, our tool would be able to analyze and&#13;\n\t\t\t\ttransform existing Ant build scripts by loading them and dealing with&#13;\n\t\t\t\tthe XML nodes directly. We would abstract ourselves from strings and deal&#13;\n\t\t\t\twith higher level concepts which let us get the job done faster and&#13;\n\t\t\t\teasier.<\/p><p>Of course we could write Ant tasks that allow dealing with XML&#13;\n\t\t\t\ttransformations and write our generation tool in Ant itself. Or we&#13;\n\t\t\t\tcould just use Lisp. As we saw earlier, a list is a built in Lisp&#13;\n\t\t\t\tdata structure and Lisp has a number of facilities for processing lists&#13;\n\t\t\t\tquickly and effectively (<em>head<\/em> and <em>tail<\/em> being the simplest ones). Additionally Lisp&#13;\n\t\t\t\thas no semantic constraints - you can have your code (and data) have&#13;\n\t\t\t\tany structure you want.<\/p><p>Metaprogramming in Lisp is done using a&#13;\n\t\t\t\tconstruct called a \"macro\". Let's try to develop a set of macros that&#13;\n\t\t\t\ttransform data like, say, a to-do list (surprised?), into a language for dealing with&#13;\n\t\t\t\tto-do lists.<\/p><p>Let's recall our to-do list example. The XML looks like this:<\/p><pre>&#13;\n&lt;todo name=\"housework\"&gt;&#13;\n    &lt;item priority=\"high\"&gt;Clean the house.&lt;\/item&gt;&#13;\n    &lt;item priority=\"medium\"&gt;Wash the dishes.&lt;\/item&gt;&#13;\n    &lt;item priority=\"medium\"&gt;Buy more soap.&lt;\/item&gt;&#13;\n&lt;\/todo&gt;&#13;\n<\/pre><p>The corresponding s-expression version looks like this:<\/p><pre>&#13;\n(todo \"housework\"&#13;\n    (item (priority high) \"Clean the house.\")&#13;\n    (item (priority medium) \"Wash the dishes.\")&#13;\n    (item (priority medium) \"Buy more soap.\"))&#13;\n<\/pre><p>Suppose we're writing a to-do manager application. We keep our to-do&#13;\n\t\t\t\titems serialized in a set of files and when the program starts up we&#13;\n\t\t\t\twant to read them and display them to the user. How would we do this&#13;\n\t\t\t\twith XML and some other language (say, Java)? We'd parse our XML files&#13;\n\t\t\t\twith the to-do lists using some XML parser, write the code that walks&#13;\n\t\t\t\tthe XML tree and converts it to a Java data structure (because frankly,&#13;\n\t\t\t\tprocessing DOM in Java is a pain in the neck), and then use&#13;\n\t\t\t\tthis data structure to display the data. Now, how would we do the&#13;\n\t\t\t\tsame thing in Lisp?<\/p><p>If we were to adopt the same approach we'd parse the files using Lisp&#13;\n\t\t\t\tlibraries responsible for parsing XML. The XML would then be presented&#13;\n\t\t\t\tto us as a Lisp list (an s-expression) and we'd walk the list and&#13;\n\t\t\t\tpresent relevant data to the user. Of course if we used Lisp it would&#13;\n\t\t\t\tmake sense to persist the data as s-expressions directly as there's no&#13;\n\t\t\t\treason to do an XML conversion. We wouldn't need special parsing&#13;\n\t\t\t\tlibraries since data persisted as a set of s-expressions is valid Lisp and we&#13;\n\t\t\t\tcould use Lisp compiler to parse it and store it in memory as a Lisp&#13;\n\t\t\t\tlist. Note that Lisp compiler (much like .NET compiler) is available&#13;\n\t\t\t\tto a Lisp program at runtime.<\/p><p>But we can do better. Instead of writing code to walk the s-expression&#13;\n\t\t\t\tthat stores our data we could write a macro that allows us to treat&#13;\n\t\t\t\tdata as code! How do macros work? Pretty simple, really. Recall that a&#13;\n\t\t\t\tLisp function is called like this:<\/p><pre>&#13;\n(function-name arg1 arg2 arg3)&#13;\n<\/pre><p>Where each argument is a valid Lisp expression that's evaluated and&#13;\n\t\t\t\tpassed to the function. For example if we replace <em>arg1<\/em> above with <em>(+ 4&#13;\n\t\t\t\t5)<\/em>, it will be evaluated and <em>9<\/em> would be passed to the function. A&#13;\n\t\t\t\tmacro works the same way as a function, except its arguments are not&#13;\n\t\t\t\tevaluated.<\/p><pre>&#13;\n(macro-name (+ 4 5))&#13;\n<\/pre><p>In this case, (+ 4 5) is not evaluated and is passed to the macro as a&#13;\n\t\t\t\tlist. The macro is then free to do what it likes with it, including&#13;\n\t\t\t\tevaluating it. The return value of a macro is a Lisp list that's&#13;\n\t\t\t\ttreated as code. The original place with the macro is replaced with&#13;\n\t\t\t\tthis code. For example, we could define a macro plus that takes two&#13;\n\t\t\t\targuments and puts in the code that adds them.<\/p><p>What does it have to do with metaprogramming and our to-do list problem?&#13;\n\t\t\t\tWell, for one, macros are little bits of code that generate code using&#13;\n\t\t\t\ta list abstraction. Also, we could create macros named <em>to-do<\/em> and <em>item<\/em> that&#13;\n\t\t\t\treplace our data with whatever code we like, for instance code that&#13;\n\t\t\t\tdisplays the item to the user.<\/p><p>What benefits does this approach offer? We don't have to walk the list.&#13;\n\t\t\t\tThe compiler will do it for us and will invoke appropriate macros. All&#13;\n\t\t\t\twe need to do is create the macros that convert our data to&#13;\n\t\t\t\tappropriate code!<\/p><p>For example, a macro similar to our <em>triple<\/em> C macro we showed earlier&#13;\n\t\t\t\tlooks like this:<\/p><pre>&#13;\n(defmacro triple (x)&#13;\n    '(+ ~x ~x ~x))&#13;\n<\/pre><p>The quote prevents evaluation while the tilde allows it. Now every&#13;\n\t\t\t\ttime <em>triple<\/em> is encountered in lisp code:<\/p><pre>&#13;\n(triple 4)&#13;\n<\/pre><p>it is replaced with the following code:<\/p><pre>&#13;\n(+ 4 4 4)&#13;\n<\/pre><p>We can create macros for our to-do list items that will get called by&#13;\n\t\t\t\tlisp compiler and will transform the to-do list into code. Now our to-do&#13;\n\t\t\t\tlist will be treated as code and will be executed. Suppose all we want to do&#13;\n\t\t\t\tis print it to standard output for the user to read:<\/p><pre>&#13;\n(defmacro item (priority note)&#13;\n    '(block&#13;\n         (print stdout tab \"Priority: \"&#13;\n             ~(head (tail priority)) endl)&#13;\n         (print stdout tab \"Note: \" ~note endl endl)))&#13;\n<\/pre><p>We've just created a very small and limited language for managing to-do lists&#13;\n\t\t\t\tembedded in Lisp. Such languages are very specific to a particular problem domain&#13;\n\t\t\t\tand are often referred to as domain specific languages or <em>DSLs<\/em>.<\/p><p class=\"first\">In this article we've already encountered two domain specific languages: Ant&#13;\n\t\t\t\t(specific to dealing with project builds) and our unnamed mini-language for dealing with to-do&#13;\n\t\t\t\tlists. The difference is that Ant was written from scratch using XML, an XML parser, and Java while&#13;\n\t\t\t\tour language is embedded into Lisp and is easily created within a couple of minutes.<\/p><p>We've already discussed the benefits of DSLs, mainly why Ant is using XML, not Java source code.&#13;\n\t\t\t\tLisp lets us create as many DSLs as we need for our problem. We can create domain specific languages&#13;\n\t\t\t\tfor creating web applications, writing massively multiplayer games, doing fixed income trading,&#13;\n\t\t\t\tsolving the protein folding problem, dealing with transactions, etc. We can layer these languages on&#13;\n\t\t\t\ttop of each other and create a language for writing web-based trading applications by taking advantage&#13;\n\t\t\t\tof our web application language and bond trading language. Every day we'd reap the benefits of&#13;\n\t\t\t\tthis approach, much like we reap the benefits of Ant.<\/p><p>Using DSLs to solve problems results in much more compact, maintainable, flexible programs. In&#13;\n\t\t\t\ta way we create them in Java by creating classes that help us solve the problem. The difference&#13;\n\t\t\t\tis that Lisp allows us to take this abstraction to the next level: we're not limited by Java's parser.&#13;\n\t\t\t\tThink of writing build scripts in Java itself using some supporting library. Compare it to using Ant.&#13;\n\t\t\t\tNow apply this same comparison to every single problem you've ever worked on and you'll begin&#13;\n\t\t\t\tto glimpse a small share of the benefits offered by Lisp.<\/p><p class=\"first\">&#13;\n\t\t\t\tLearning Lisp is an uphill battle. Even though in Computer Science terms Lisp is an ancient language,&#13;\n\t\t\t\tfew people to date figured out how to teach it well enough&#13;\n\t\t\t\tto make it accessible. Despite great efforts by many Lisp advocates, &#13;\n\t\t\t\tlearning Lisp today is still hard. The good news is that this won't remain the case forever&#13;\n\t\t\t\tsince the amount of Lisp-related resources is rapidly increasing. Time is on Lisp's side.<\/p><p>Lisp is a way to escape mediocrity and to get ahead of the pack. Learning Lisp&#13;\n\t\t\t\tmeans you can get a better job today, because you can impress any reasonably intelligent interviewer&#13;\n\t\t\t\twith fresh insight into most aspects of software engineering. It also means&#13;\n\t\t\t\tyou're likely to get fired tomorrow because everyone is tired of you constantly mentioning how much better&#13;\n\t\t\t\tthe company could be doing if only its software was written in Lisp. Is it worth the effort? Everyone who has ever&#13;\n\t\t\t\tlearned Lisp says yes. The choice, of course, remains yours.<\/p><p class=\"first\">Whew. That's enough. I've been writing this article, on and off, for months. If you find it interesting,&#13;\n\t\t\t\thave any questions, comments, or suggestions, please drop a note at <a href=\"mailto:coffeemug@gmail.com\">coffeemug@gmail.com<\/a>.&#13;\n\t\t\t\tI'll be glad to hear your feedback.<\/p><div id=\"footnotes\"><p><sup><a id=\"note-james\">1<\/a><\/sup>I have never met James, nor does he know about my existence. The story is entirely fictional and is based on&#13;\n\t\t\t\t\t\ta few postings about Ant's history I found on the internet.<\/p><p><sup><a id=\"note-blaise\">2<\/a><\/sup>Lisp has many different dialects (the most popular of which are Common Lisp and Scheme). Each dialect deals&#13;\n\t\t\t\t\t\twith intricate details differently yet shares the same set of basic principles. Since the goal of this article is to give you an understanding of Lisp's&#13;\n\t\t\t\t\t\tprinciples I will use Blaise for examples (which at the time of this writing is vaporware). With some minor modifications these examples can be translated&#13;\n\t\t\t\t\t\tto other Lisp dialects.<\/p><\/div>","created_at":"2018-11-07T16:44:01+0000","updated_at":"2018-11-07T16:44:01+0000","annotations":[],"mimetype":"text\/html","reading_time":50,"domain_name":"www.defmacro.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":298,"title":"The Y Combinator (Slight Return) - Mike's World-O-Programming","url":"https:\/\/mvanier.livejournal.com\/2897.html","content":"<p>or:<\/p><p><u>How to Succeed at Recursion Without Really Recursing<\/u><\/p><div style=\"margin: auto;\"><table summary=\"\" style=\"width: 90%;\"><tr><td>\nTiger got to hunt,<br \/>Bird got to fly;<br \/>Lisper got to sit and wonder, (Y (Y Y))?<p>Tiger got to sleep,<br \/>Bird got to land;<br \/>Lisper got to tell himself he understand.<\/p><p>\u00a0\u00a0\u00a0\u00a0\u2014 Kurt Vonnegut, modified by Darius Bacon\n<\/p><\/td><\/tr><\/table><\/div><h2>Introduction<\/h2><p>I recently wrote a blog post about the Y combinator.  Since then, I've\n  received so many useful comments that I thought it was appropriate to\n  expand the post into a more complete article.  This article will go into\n  greater depth on the subject, but I hope it'll be more comprehensible as\n  well.  You don't need to have read the previous post to understand this one\n  (in fact, it's probably better if you haven't.)  The only background\n  knowledge I require is a tiny knowledge of the Scheme programming language\n  including recursion and first-class functions, which I will review.\n  Comments are (again) welcome.<\/p><h2>Why Y?<\/h2><p>Before I get into the details of what Y actually is, I'd like to address\nthe question of why you, as a programmer, should bother to learn about it. To\nbe honest, there aren't a lot of good nuts-and-bolts practical reasons for\nlearning about Y. Even though it does have a few practical applications, for\nthe most part it's mainly of interest to computer language theorists.\nNevertheless, I do think it's worth your while to know something about Y for\nthe following reasons:<\/p><ol><li><p>It's one of the most beautiful ideas in all of programming.  If you\n    have any sense of programming aesthetics, you're sure to be delighted\n    by Y.<\/p>\n<\/li>\n<li><p>It shows in a very stark way how amazingly powerful the simple ideas\n    of functional programming are.<\/p>\n<\/li>\n<\/ol><p>In 1959, the British scientist C. P. Snow gave a famous lecture\n  called <a href=\"http:\/\/en.wikipedia.org\/wiki\/The_Two_Cultures\" rel=\"nofollow\">The Two\n  Cultures<\/a> where he bemoaned the fact that many intelligent and\n  well-educated people of the time had almost no knowledge of science.  He\n  used knowledge of the Second Law of Thermodynamics as a kind of dividing\n  line between those who were scientifically literate and those who weren't.\n  I think we can similarly use knowledge of the Y combinator as a dividing\n  line between programmers who are \"functionally literate\" (<i>i.e.<\/i> have\n  a reasonably deep knowledge of functional programming) and those who\n  aren't.  There are other topics that could serve just as well as Y (notably\n  monads), but Y will do nicely.  So if you aspire to have the True\n  Lambda-Nature, read on.<\/p><p>By the way, <a href=\"http:\/\/www.paulgraham.com\" rel=\"nofollow\">Paul Graham<\/a> (the Lisp\n  hacker, Lisp book author, essayist, and now venture capitalist) apparently\n  thinks so highly of Y that he named his startup incubator\n  company <a href=\"http:\/\/www.ycombinator.com\/\" rel=\"nofollow\">Y Combinator<\/a>.  Paul got\n  rich from his knowledge of ideas like these; maybe someone else will too.\n  Maybe even you.<\/p><hr style=\"width: 75%;\" \/><h2>A puzzle<\/h2><h3>Factorials<\/h3><p>We'll start our exploration of the Y combinator by defining some functions\n  to compute factorials.  The factorial of a non-negative\n  integer <code>n<\/code> is the product of all integers starting\n  from <code>1<\/code> and going up to and including <code>n<\/code>.  Thus we\n  have:<\/p><pre>\n  factorial 1 = 1\n  factorial 2 = 2 * 1 = 2\n  factorial 3 = 3 * 2 * 1 = 6\n  factorial 4 = 4 * 3 * 2 * 1 = 24\n<\/pre><p>and so on. (I'm using a function notation without parentheses here, so\n\u00a0<code>factorial 3<\/code>\u00a0 is the same as what is usually written as\n\u00a0<code>factorial(3)<\/code>.\u00a0 Humor me.) Factorials increase very\nrapidly with increasing <code>n<\/code>; the factorial of <code>20<\/code> is\n<code>2432902008176640000<\/code>. The factorial of <code>0<\/code> is defined\nto be <code>1<\/code>; this turns out to be the appropriate definition for the kinds of things factorials are actually used for (like solving problems in\ncombinatorics).<\/p><h3>Recursive definitions of the factorial function<\/h3><p>It's easy to write a function in a programming language to compute\n  factorials using some kind of a looping control construct like\n  a <code>while<\/code> or <code>for<\/code> loop (<i>e.g.<\/i> in C or Java).\n  However, it's also easy to write a recursive function to compute\n  factorials, because factorials have a very natural recursive\n  definition:<\/p><pre>\n  factorial 0 = 1\n  factorial n = n * factorial (n - 1)\n<\/pre><p>where the second line applies for all <code>n<\/code> greater than zero. In\nfact, in the computer language <a href=\"http:\/\/www.haskell.org\" rel=\"nofollow\">Haskell<\/a>,\nthat's the way you actually define the factorial function. In <a href=\"http:\/\/www.schemers.org\" rel=\"nofollow\">Scheme<\/a>, the language we'll be using here,\nthis function would be written like this:<\/p><pre>\n  (define (factorial n)\n    (if (= n 0)\n        1\n        (* n (factorial (- n 1)))))\n<\/pre><p>Scheme uses a parenthesized prefix notation for everything, so something\n  like <code>\u00a0(- n 1)\u00a0<\/code> represents what is usually written \n  <code>\u00a0n - 1\u00a0<\/code> in most programming languages.  The reasons \n  for this are beyond the scope of this article, but getting used to this \n  notation isn't very hard.<\/p><p>In fact, the above definition of the factorial function in Scheme could\n  also be written in a slightly more explicit way as follows:<\/p><pre>\n  (define factorial\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n (factorial (- n 1))))))\n<\/pre><p>The keyword <code>lambda<\/code> simply indicates that the thing we're\ndefining (<i>i.e.<\/i> whatever is enclosed by the open parenthesis to the\nimmediate left of the <code>lambda<\/code> and its corresponding close\nparenthesis) is a function. What comes immediately after the word\n<code>lambda<\/code>, in parentheses, are the <i>formal arguments<\/i> of the\nfunction; here there is just one argument, which is <code>n<\/code>. The\n<i>body<\/i> of the function comes after the formal arguments, and here\nconsists of the expression <code>(if (= n 0) 1 (* n (factorial (- n\n1))))<\/code>. This kind of function is an <i>anonymous function<\/i>. Here you\ndo give the anonymous function the name <code>factorial<\/code> after you've\ndefined it, but you don't have to, and often it's handy not to if you're only\ngoing to be using it once. In Scheme and some other languages, anonymous\nfunctions are also called <i>lambda expressions<\/i>. Many programming\nlanguages besides Scheme allow you to define anonymous functions, including\nPython, Ruby, Javascript, Ocaml, and Haskell (but not C, C++, or Java,\nunfortunately). We'll be using lambda expressions a lot below.<\/p><p>In the Scheme language, the definition of <code>factorial<\/code> just given\nis identical to the one before it; Scheme simply translates the first\ndefinition into the second one before evaluating it. So all functions in\nScheme are really lambda expressions.<\/p><p>Note that the body of the function has a call to the <code>factorial<\/code>\nfunction (which we're in the process of defining) inside it, which makes this\na recursive definition. I will call this kind of definition, where the name of\nthe function being defined is used in the body of the function, an\n<i>explicitly recursive definition<\/i>. (You might wonder what an \"implicitly\nrecursive\" function would be. I'm not going to use that expression, but the\nnotion I have in mind is a recursive function which is generated through\nnon-recursive means \u2014 keep reading!)<\/p><p>For the sake of argument, we're going to assume that our version of Scheme\ndoesn't have the equivalent of <code>for<\/code> or <code>while<\/code> loops in\nC or Java (although in fact, real Scheme implementations do have such\nconstructs, but under a different name), so that in order to define a function\nlike <code>factorial<\/code>, we pretty much have to use recursion. Scheme is\noften used as a teaching language partly for this reason: it forces students\nto learn to think recursively.<\/p><h3>Functions as data and higher-order functions<\/h3><p>Scheme is a cool language for many reasons, but one that is relevant to us\nhere is that it allows you to use functions as \"first class\" data objects\n(this is often expressed by saying that Scheme supports <i>first-class\nfunctions<\/i>). This means that in Scheme, we can pass a function to another\nfunction as an argument, we can return a function as the result of evaluating\nanother function applied to its arguments, and we can create functions\non-the-fly as we need them (using the <code>lambda<\/code> notation shown\nabove). This is the essence of functional programming, and it will feature\nprominently in the ensuing discussion. Functions which take other functions as\narguments, and\/or which return other functions as their results, are usually\nreferred to as <i>higher-order functions<\/i>.<\/p><h3>Eliminating explicit recursion<\/h3><p>Now, here's the puzzle: what if you were asked to define the\n<code>factorial<\/code> function in Scheme, but were told that you could not\nuse recursive function calls in the definition (for instance, in the\n<code>factorial<\/code> function given above you cannot use the word\n<code>factorial<\/code> anywhere in the body of the function). However, you\n<i>are<\/i> allowed to use first-class functions and higher-order functions any\nway you see fit. With this knowledge, can you define the\n<code>factorial<\/code> function?<\/p><p>The answer to this question is yes, and it will lead us directly to the Y\ncombinator.<\/p><hr style=\"width: 75%;\" \/><h2>What the Y combinator is and what it does<\/h2><p>The Y combinator is a higher-order function. It takes a single argument,\nwhich is a function that isn't recursive. It returns a version of the function\nwhich is recursive. We will walk through this process of generating recursive\nfunctions from non-recursive ones using Y in great detail below, but that's\nthe basic idea.<\/p><p>More generally, Y gives us a way to get recursion in a programming language\nthat supports first-class functions but that doesn't have recursion built in\nto it. So what Y shows us is that such a language already allows us to define\nrecursive functions, even though the language definition itself says nothing\nabout recursion. This is a Beautiful Thing: it shows us that functional\nprogramming alone can allow us to do things that we would never expect to be\nable to do (and it's not the only example of this).<\/p><h3>Lazy or strict evaluation?<\/h3><p>We will be looking at two broad classes of computer languages: those that use <i>lazy evaluation<\/i> and those that use <i>strict evaluation<\/i>. Lazy\nevaluation means that in order to evaluate an expression in the language, you\nonly evaluate as much of the expression as is needed to get the final result.\nSo (for instance) if there is a part of the expression that doesn't need to\nget evaluated (because the result will not depend on it) it won't be\nevaluated. In contrast, strict evaluation means that all parts of an\nevaluation will be evaluated completely before the value of the expression as\na whole is determined (with some necessary exceptions, such as\n<code>if<\/code> expressions, which have to be lazy to work properly). In\npractice, lazy evaluation is more general, but strict evaluation is more\npredictable and often more efficient. Most programming languages use strict\nevaluation. The programming language <a href=\"http:\/\/www.haskell.org\" rel=\"nofollow\">Haskell<\/a> uses lazy evaluation, and this is\none of the most interesting things about that language. We will use both kinds\nof evaluation in what follows.<\/p><h3>One Y combinator or many?<\/h3><p>Even though we often refer to Y as \"the\" Y combinator, in actual fact there\nare an infinite number of Y combinators. We will only be concerned with two of\nthese, one lazy and one strict. We need two Y combinators because the Y\ncombinator we define for lazy languages will not work for strict languages.\nThe lazy Y combinator is often referred to as the <i>normal-order Y\ncombinator<\/i> and the strict one is referred to as the <i>applicative-order Y\ncombinator<\/i>. Basically, <i>normal-order<\/i> is another way of saying \"lazy\"\nand <i>applicative-order<\/i> is another way of saying \"strict\".<\/p><h3>Static or dynamic typing?<\/h3><p>Another big dividing line in programming languages is between <i>static\ntyping<\/i> and <i>dynamic typing<\/i>. A statically-typed language is one where\nthe types of all expressions are determined at compile time, and any type\nerrors cause the compilation to fail. A dynamically-typed language doesn't do\nany type checking until run time, and if a function is applied to arguments of\ninvalid types (<i>e.g.<\/i> by trying to add together an integer and a string),\nthen an error is reported. Among commonly-used programming languages, C, C++\nand Java are statically typed, and Perl, Python and Ruby are dynamically\ntyped. Scheme (the language we'll be using for our examples) is also\ndynamically typed. (There are also languages that straddle the border between\nstatically-typed and dynamically-typed, but I won't discuss this\nfurther.)<\/p><p>One often hears static typing referred to as <i>strong typing<\/i> and\ndynamic typing referred to as <i>weak typing<\/i>, but this is an abuse of\nterminology. Strong typing simply means that every value in the language has\none and only one type, whereas weak typing means that some values can have\nmultiple types. So Scheme, which is dynamically typed, is also strongly typed,\nwhile C, which is statically typed, is weakly typed (because you can cast a\npointer to one kind of object into a pointer to another type of object without\naltering the pointer's value). I will only be concerned with strongly typed\nlanguages here.<\/p><p>It turns out to be much simpler to define the Y combinator in dynamically\ntyped languages, so that's what I'll do. It is possible to define a Y\ncombinator in many statically typed languages, but (at least in the examples\nI've seen) such definitions usually require some non-obvious type hackery,\nbecause the Y combinator itself doesn't have a straightforward static type.\nThat's beyond the scope of this article, so I won't mention it further.<\/p><h3>What a \"combinator\" is<\/h3><p>A combinator is just a <i>lambda expression<\/i> with no <i>free\nvariables<\/i>. We saw above what lambda expressions are (they're just\nanonymous functions), but what's a free variable? It's a variable (<i>i.e.\n<\/i> a name or identifier in the language) which isn't a <i>bound\nvariable<\/i>. Happy now? No? OK, let me explain.<\/p><p>A bound variable is simply a variable which is contained inside the body of a lambda expression that has that variable name as one of its arguments.<\/p><p>Let's look at some examples of lambda expressions and free and bound variables:<\/p><ol><li><code>(lambda (x) x)<\/code><\/li>\n    <li><code>(lambda (x) y)<\/code><\/li>\n    <li><code>(lambda (x) (lambda (y) x))<\/code><\/li>\n    <li><code>(lambda (x) (lambda (y) (x y)))<\/code><\/li>\n    <li><code>(x (lambda (y) y))<\/code><\/li>\n    <li><code>((lambda (x) x) y)<\/code><\/li>\n<\/ol><p>Are the variables in the body of these lambda expressions free variables or bound variables?  We'll ignore the formal arguments of the lambda expressions, because only variables in the body of the lambda expression can be considered free or bound. As for the other variables, here are the answers:<\/p><ol><li>The <code>x<\/code> in the body of the lambda expression is a bound\nvariable, because the formal argument of the lambda expression is also\n<code>x<\/code>. This lambda expression has no other variables, therefore it\nhas no free variables, therefore it's a combinator.<\/li>\n    <li>The <code>y<\/code> in the lambda body is a free variable.  This lambda expression is therefore not a combinator.<\/li>\n    <li>Aside from the formal arguments of the lambda expression, there is only one variable, the final <code>x<\/code>, which is a bound variable (it's bound by the formal argument of the outer lambda expression).  Therefore, this lambda expression as a whole has no free variables, so this is a combinator.<\/li>\n    <li>Aside from the formal arguments of the lambda expression, there are two variables, the final <code>x<\/code> and <code>y<\/code>, both bound variables.  This is a combinator.<\/li>\n    <li>The entire expression is not a lambda expression, so it's by definition not a combinator.  Nevertheless, the <code>x<\/code> is a free variable and the final <code>y<\/code> is a bound variable.<\/li>\n    <li>Again, the entire expression isn't a lambda expression (it's a function application), so this isn't a combinator either.  The second <code>x<\/code> is a bound variable while the <code>y<\/code> is a free variable.<\/li>\n<\/ol><p>When you're wondering if a recursive function like\n<code>factorial<\/code>:<\/p><pre>\n   (define factorial\n     (lambda (n)\n       (if (= n 0)\n           1\n           (* n (factorial (- n 1))))))\n<\/pre><p>is a combinator, you don't consider the <code>define<\/code> part, so what you're really asking is if<\/p><pre>\n  (lambda (n)\n    (if (= n 0)\n        1\n        (* n (factorial (- n 1)))))\n<\/pre><p>is a combinator. Since in this lambda expression, the name\n<code>factorial<\/code> represents a free variable (the name\n<code>factorial<\/code> is not a formal argument of the lambda expression), this is not a combinator. This will be important below.  In fact, the names <code>=<\/code>, <code>*<\/code>, and <code>-<\/code> are also free variables, so even without the name <code>factorial<\/code> this would not be a combinator (to say nothing of the numbers!).<\/p><hr style=\"width: 75%;\" \/><h2>Back to the puzzle<\/h2><h3>Abstracting out the recursive function call<\/h3><p>Recall the factorial function we had previously:<\/p><pre>\n  (define factorial\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n (factorial (- n 1))))))\n<\/pre><p>What we want to do is to come up with a version of this that does the same\n  thing but doesn't have that pesky recursive call to <code>factorial<\/code>\n  in the body of the function.<\/p><p>Where do we start?  It would be nice if you could save all of the function\n  except for the offending recursive call, and put something else there.\n  That might look like this:<\/p><pre>\n  (define sort-of-factorial\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n (&lt;???&gt; (- n 1))))))\n<\/pre><p>This still leaves us with the problem of what to put in the place\n  marked <code>&lt;???&gt;<\/code>.  It's a tried-and-true principle of\n  functional programming that if you don't know exactly what you want to put\n  somewhere in a piece of code, just abstract it out and make it a parameter\n  of a function.  The easiest way to do this is as follows:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n<\/pre><p>What we've done here is to rename the recursive call to\n<code>factorial<\/code> to <code>f<\/code>, and to make <code>f<\/code> an\nargument to a function which we're calling <code>almost-factorial<\/code>.\nNotice that <code>almost-factorial<\/code> is not at all the factorial\nfunction. Instead, it's a higher-order function which takes a single argument\n<code>f<\/code>, which had better be a function (or else <code>(f (- n\n1))<\/code> won't make sense), and <i>returns<\/i> another function (the\n<code>(lambda (n) ...)<\/code> part) which (hopefully) will be a factorial\nfunction if we choose the right value for <code>f<\/code>.<\/p><p>It's important to realize that this trick is not in any way specific to\n  the <code>factorial<\/code> function.  We can do exactly the same trick with\n  any recursive function.  For instance, consider a recursive function to\n  compute fibonacci numbers.  The recursive definition of fibonacci numbers\n  is as follows:<\/p><pre>\n  fibonacci 0 = 0\n  fibonacci 1 = 1\n  fibonacci n = fibonacci (n - 1) + fibonacci (n - 2)\n<\/pre><p>(In fact, that's the definition of the fibonacci function in Haskell.)  In\n  Scheme, we can write the function this way:<\/p><pre>\n  (define fibonacci\n    (lambda (n)\n      (cond ((= n 0) 0)\n            ((= n 1) 1)\n            (else (+ (fibonacci (- n 1)) (fibonacci (- n 2)))))))\n<\/pre><p>(where <code>cond<\/code> is just a shorthand expression for\n  nested <code>if<\/code> expressions).  We can then remove the explicit\n  recursion just like we did for <code>factorial<\/code>:<\/p><pre>\n  (define almost-fibonacci\n    (lambda (f)\n      (lambda (n)\n        (cond ((= n 0) 0)\n              ((= n 1) 1)\n              (else (+ (f (- n 1)) (f (- n 2))))))))\n<\/pre><p>As you can see, the transformation from a recursive function to a\n  non-recursive <code>almost-<\/code> equivalent function is a purely\n  mechanical one: you rename the name of the recursive function inside the\n  body of the function to <code>f<\/code> and you wrap a <code>(lambda (f)\n  ...)<\/code> around the body.<\/p><p>If you've followed what I just did (never mind <i>why<\/i> I did it; we'll\n  see that later), then congratulations!  As Yoda says, you've just taken the\n  first step into a larger world.<\/p><h3>Sneak preview<\/h3><p>I probably shouldn't do this yet, but I'm going to give you a sneak\n  preview of where we're going.  Once we define the Y combinator, we'll be\n  able to define the factorial function using <code>almost-factorial<\/code>\n  as follows:<\/p><pre>\n  (define factorial (Y almost-factorial))\n<\/pre><p>where <code>Y<\/code> is the Y combinator. Note that this definition of\n<code>factorial<\/code> doesn't have any explicit recursion in it. Similarly,\nwe can define the <code>fibonacci<\/code> function using\n<code>almost-fibonacci<\/code> in the same way:<\/p><pre>\n  (define fibonacci (Y almost-fibonacci))\n<\/pre><p>So the Y combinator will give us recursion wherever we need it as long as\n  we have the appropriate <code>almost-<\/code> function available\n  (<i>i.e.<\/i> the non-recursive function derived from the recursive one by\n  abstracting out the recursive function calls).<\/p><p>Read on to see what's really going on here and why this will work.<\/p><h3>Recovering <code>factorial<\/code> from <code>almost-factorial<\/code><\/h3><p>Let's assume, for the sake of argument, that we already had a working\n  factorial function lying around (recursive or not, we don't care).  We'll\n  call that hypothetical factorial function <code>factorialA<\/code>.  Now\n  let's consider the following:<\/p><pre>\n  (define factorialB (almost-factorial factorialA))\n<\/pre><p>Question: does <code>factorialB<\/code> actually compute factorials?<\/p><p>To answer this, it's helpful to expand out the definition\n  of <code>almost-factorial<\/code>:<\/p><pre>\n  (define factorialB\n    ((lambda (f)\n       (lambda (n)\n         (if (= n 0)\n             1\n             (* n (f (- n 1))))))\n     factorialA))\n<\/pre><p>Now, by substituting <code>factorialA<\/code> for <code>f<\/code> inside the\n  body of the lambda expression we get:<\/p><pre>\n  (define factorialB\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n (factorialA (- n 1))))))\n<\/pre><p>This looks a lot like the recursive factorial function, but it isn't:\n<code>factorialA<\/code> is not the same function as <code>factorialB<\/code>.\nSo it's a non-recursive function that depends on a hypothetical\n<code>factorialA<\/code> function to work. Does it actually work? Well, it's\npretty obvious that it should work for <code>n = 0<\/code>, since\n<code>(factorialB 0)<\/code> will just return <code>1<\/code> (the factorial of\n<code>0<\/code>). If <code>n &gt; 0<\/code>, then the value of <code>(factorialB\nn)<\/code> will be <code>(* n (factorialA (- n 1)))<\/code>. Now, we assumed\nthat <code>factorialA<\/code> would correctly compute factorials, so\n<code>(factorialA (- n 1))<\/code> is the factorial of <code>n - 1<\/code>, and\ntherefore <code>(* n (factorialA (- n 1)))<\/code> is the factorial of\n<code>n<\/code> (by the definition of factorial), thus proving that\n<code>factorialB<\/code> computes the factorial function correctly as long as\n<code>factorialA<\/code> does. So this works. The only problem is that we don't\nactually have a <code>factorialA<\/code> lying around.<\/p><p>Now, if you're really clever, you might be asking yourself whether we can\n  just do this:<\/p><pre>\n  (define factorialA (almost-factorial factorialA))\n<\/pre><p>The idea is this: let's assume that <code>factorialA<\/code> is a valid\n  factorial function.  Then if we pass it as an argument\n  to <code>almost-factorial<\/code>, the resulting function will have to be a\n  valid factorial function, so why not just name that\n  function <code>factorialA<\/code>?<\/p><p>It looks like you've created a perpetual-motion machine (or perhaps I\n  should say a perpetual-calculation machine), and there must\n  be <i>something<\/i> wrong with this definition... mustn't there?<\/p><p>In fact, this definition will work fine as long as the Scheme language\n  you're using uses lazy evaluation!  Standard Scheme uses strict evaluation,\n  so it won't work (it'll go into an infinite loop).  If you\n  use <a href=\"http:\/\/www.drscheme.org\" rel=\"nofollow\">DrScheme<\/a> as your Scheme\n  interpreter (which you should), then you can use the \"lazy Scheme\" language\n  level, and the above code will actually work (huzzah!).  We'll see why\n  below, but for now I want to stick to standard (strict) Scheme and approach\n  the problem in a slightly different way.<\/p><p>Let's define a couple of functions:<\/p><pre>\n  (define identity (lambda (x) x))\n  (define factorial0 (almost-factorial identity))\n<\/pre><p>The <code>identity<\/code> function is pretty simple: it takes in a single\nargument and returns it unchanged (it's also a combinator, as I hope you can\ntell). We're basically going to use it as a placeholder when we need to pass a\nfunction as an argument and we don't know what function we should pass.<\/p><p><code>factorial0<\/code> is more interesting.  It's a function that can\n  compute <i>some<\/i>, but not <i>all<\/i> factorials.  Specifically, it can\n  compute the factorials up to and including the factorial of zero (which\n  means that it can only compute the factorial of zero, but you'll soon see\n  why I describe it this way).  Let's verify that:<\/p><pre>\n  (factorial0 0)\n  ==&gt; ((almost-factorial identity) 0)\n  ==&gt; (((lambda (f)\n          (lambda (n)\n             (if (= n 0)\n                 1\n                 (* n (f (- n 1))))))\n        identity)\n       0)\n  ==&gt; ((lambda (n)\n         (if (= n 0)\n             1\n             (* n (identity (- n 1)))))\n       0)\n  ==&gt; (if (= 0 0)\n          1\n          (* 0 (identity (- 0 1))))\n  ==&gt; (if #t\n          1\n          (* 0 (identity (- 0 1))))\n  ==&gt; 1\n<\/pre><p>OK, so it works.  Unfortunately, it won't work for <code>n &gt; 0<\/code>.\n  For instance, if <code>n = 1<\/code> then we'll have (skipping a few obvious\n  steps):<\/p><pre>\n  (factorial0 1)\n  ==&gt; (* 1 (identity (- 1 1)))\n  ==&gt; (* 1 (identity 0))\n  ==&gt; (* 1 0)\n  ==&gt; 0\n<\/pre><p>which is not the correct answer.<\/p><p>Now consider this spiffed-up version of <code>factorial0<\/code>:<\/p><pre>\n  (define factorial1\n    (almost-factorial factorial0))\n<\/pre><p>which is the same thing as:<\/p><pre>\n  (define factorial1\n    (almost-factorial \n      (almost-factorial identity)))\n<\/pre><p>This will correctly compute the factorials of <code>0<\/code>\n  and <code>1<\/code>, but it will be incorrect for any <code>n &gt; 1<\/code>.\n  Let's verify this as well, again skipping some obvious steps:<\/p><pre>\n  (factorial1 0)\n  ==&gt; ((almost-factorial factorial0) 0)\n  ==&gt; 1 (via essentially the same derivation we showed above)\n  (factorial1 1)\n  ==&gt; ((almost-factorial factorial0) 1)\n  ==&gt; (((lambda (f)\n          (lambda (n)\n            (if (= n 0)\n                1\n                (* n (f (- n 1))))))\n        factorial0)\n       1)\n  ==&gt; ((lambda (n)\n         (if (= n 0)\n             1\n             (* n (factorial0 (- n 1)))))\n       1)\n  ==&gt; (if (= 1 0)\n          1\n          (* 1 (factorial0 (- 1 1))))\n  ==&gt; (if #f\n          1\n          (* 1 (factorial0 (- 1 1))))\n  ==&gt; (* 1 (factorial0 (- 1 1)))\n  ==&gt; (* 1 (factorial0 0))\n  ==&gt; (* 1 1)\n  ==&gt; 1\n<\/pre><p>which is the correct answer.  So <code>factorial1<\/code> can compute\n  factorials for <code>n = 0<\/code> and <code>n = 1<\/code>.  You can verify,\n  though, that it won't be correct for <code>n &gt; 1<\/code>.<\/p><p>We can keep going, and define functions which can compute factorials up\n  to any particular limit:<\/p><pre>\n  (define factorial2 (almost-factorial factorial1))\n  (define factorial3 (almost-factorial factorial2))\n  (define factorial4 (almost-factorial factorial3))\n  (define factorial5 (almost-factorial factorial4))\n  etc.\n<\/pre><p><code>factorial2<\/code> will compute correct factorials for inputs\n  between <code>0<\/code> and <code>2<\/code>, <code>factorial3<\/code> will\n  compute correct factorials for inputs between <code>0<\/code>\n  and <code>3<\/code>, and so on.  You should be able to verify this for\n  yourself using the above derivations as models, though you probably won't\n  be able to do it in your head (at least, <i>I<\/i> can't do it in my\n  head).<\/p><p>One interesting way of looking at this is\n  that <code>almost-factorial<\/code> takes in a crappy factorial function and\n  outputs a factorial function that is slightly less crappy, in that it will\n  handle exactly one extra value of the input correctly.<\/p><p>Note that you can again rewrite the definitions of the factorial functions\n  like this:<\/p><pre>\n  (define factorial0 (almost-factorial identity))\n  (define factorial1 \n    (almost-factorial\n      (almost-factorial identity)))\n  (define factorial2\n    (almost-factorial\n      (almost-factorial\n        (almost-factorial identity))))\n  (define factorial3\n    (almost-factorial\n      (almost-factorial\n        (almost-factorial \n          (almost-factorial identity)))))\n  (define factorial4\n    (almost-factorial\n      (almost-factorial\n        (almost-factorial \n          (almost-factorial\n            (almost-factorial identity))))))\n  (define factorial5\n    (almost-factorial\n      (almost-factorial\n        (almost-factorial \n          (almost-factorial\n            (almost-factorial \n              (almost-factorial identity)))))))\n<\/pre><p>and so on.  Again, if you're very clever you might wonder if you could do\n  this:<\/p><pre>\n  (define factorial-infinity\n    (almost-factorial\n      (almost-factorial\n        (almost-factorial\n          ...))))\n<\/pre><p>where the <code>...<\/code> means that you're repeating the chain of\n<code>almost-factorials<\/code> an infinite number of times. If you did wonder\nthis, go to the head of the class! Unfortunately, we can't write this out\ndirectly, but we can define the equivalent of this. Note also that\n<code>factorial-infinity<\/code> is just the <code>factorial<\/code> function we\nwant: it works on all integers greater than or equal to zero.<\/p><p>What we have shown is that if we could define an infinite chain of\n<code>almost-factorials<\/code>, that would give us the factorial function.\nAnother way of saying this is that the factorial function is the\n<i>fixpoint<\/i> of <code>almost-factorial<\/code>, which is what I will explain\nnext.<\/p><h3>Fixpoints of functions<\/h3><p>The notion of a fixpoint should be familiar to anyone who has amused\nthemselves playing with a pocket calculator. You start with <code>0<\/code> and\nhit the <code>cos<\/code> (cosine) key repeatedly. What you find is that the\nanswer rapidly converges to a number which is (approximately)\n<code>0.73908513321516067<\/code>; hitting the <code>cos<\/code> key again\ndoesn't change anything because <code>cos(0.73908513321516067) =\n0.73908513321516067<\/code>. We say that the number\n<code>0.73908513321516067<\/code> is a <i>fixpoint<\/i> of the cosine\nfunction.<\/p><p>The cosine function takes a single input value (a real number) and produces\na single output value (also a real number). The fact that the input and output\nof the function are the same type is what allows you to apply it repeatedly,\nso that if <code>x<\/code> is a real number, we can calculate what\n<code>cos(x)<\/code> is, and since that will also be a real number, we can\ncalculate what <code>cos(cos(x))<\/code> is, and then what\n<code>cos(cos(cos(x)))<\/code> is, and so on. The fixpoint is the value\n<code>x<\/code> where <code>cos(x) = x<\/code>.<\/p><p>Fixpoints don't have to be real numbers. In fact, they can be any type of\nthing, as long as the function that generates them can take the same type of\nthing as input as it produces as output. Most importantly for our discussion,\nfixpoints can be functions. If you have a higher-order function like\n<code>almost-factorial<\/code> that takes in a function as its input and\nproduces a function as its output (with both input and output functions taking\na single integer argument as input and producing a single integer as output),\nthen it should be possible to compute its fixpoint (which will, naturally, be\na function which takes a single integer argument as input and produces a\nsingle integer as output). That fixpoint function will be the function for\nwhich<\/p><pre>\n  fixpoint-function = (almost-factorial fixpoint-function)\n<\/pre><p>By repeatedly substituting the right-hand side of that equation into\n  the <code>fixpoint-function<\/code> on the right, we get:<\/p><pre>\n  fixpoint-function =\n    (almost-factorial\n      (almost-factorial fixpoint-function))\n    = (almost-factorial\n        (almost-factorial\n          (almost-factorial fixpoint-function)))\n    = ...\n    = (almost-factorial\n        (almost-factorial\n          (almost-factorial\n            (almost-factorial\n              (almost-factorial ...)))))\n<\/pre><p>As we saw above, this will be the factorial function we want.  Thus, the\n  fixpoint of <code>almost-factorial<\/code> will be\n  the <code>factorial<\/code> function:<\/p><pre>\n  factorial = (almost-factorial factorial)\n    = (almost-factorial\n        (almost-factorial\n          (almost-factorial\n            (almost-factorial\n              (almost-factorial ...)))))\n<\/pre><p>That's all well and good, but just <i>knowing<\/i>\n  that <code>factorial<\/code> is the fixpoint\n  of <code>almost-factorial<\/code> doesn't tell us how to compute it.\n  Wouldn't it be nice if there was some magical higher-order function that\n  would take as its input a function like <code>almost-factorial<\/code>, and\n  would output its fixpoint function, which in that case would\n  be <code>factorial<\/code>?  Wouldn't that be really freakin' sweet?<\/p><p>That function exists, and it's the Y combinator.  Y is also known as\n  the <i>fixpoint combinator<\/i>: it takes in a function and returns its\n  fixpoint.<\/p><h3>Eliminating (most) explicit recursion (lazy version)<\/h3><p>OK, it's time to derive Y.<\/p><p>Let's start by specifying what Y does:<\/p><pre>\n  (Y f) = fixpoint-of-f\n<\/pre><p>What do we know about the fixpoint of <code>f<\/code>?  We know that<\/p><pre>\n  (f fixpoint-of-f) = fixpoint-of-f\n<\/pre><p>by the definition of what a fixpoint of a function is.  Therefore, we\n  have:<\/p><pre>\n  (Y f) = fixpoint-of-f = (f fixpoint-of-f)\n<\/pre><p>and we can substitute <code>(Y f)<\/code> for <code>fixpoint-of-f<\/code> to\n  get:<\/p><pre>\n  (Y f) = (f (Y f))\n<\/pre><p>Voila!  We've just defined Y.  If we want it to be expressed as a Scheme\n  function, we would have to write it like this:<\/p><pre>\n  (define (Y f) (f (Y f)))\n<\/pre><p>or, using an explicit <code>lambda<\/code> expression, as:<\/p><pre>\n  (define Y\n    (lambda (f)\n      (f (Y f))))\n<\/pre><p>However, there are two caveats regarding this definition of Y:<\/p><ol><li><p>It will only work in a lazy language (see below).<\/p><\/li>\n<li><p>It is not a combinator, because the <code>Y<\/code> in the body of the\n    definition is a free variable which is only bound once the definition is\n    complete.  In other words, we couldn't just take the body of this version\n    of <code>Y<\/code> and plop it in wherever we needed it, because it\n    requires that the name <code>Y<\/code> be defined somewhere.<\/p><\/li>\n<\/ol><p>Nevertheless, if you're using lazy Scheme, you can indeed define\n  factorials like this:<\/p><pre>\n  (define Y\n    (lambda (f)\n      (f (Y f))))\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial (Y almost-factorial))\n<\/pre><p>and it will work correctly.<\/p><p>What have we accomplished?  We originally wanted to be able to define the\n  factorial function without using any explicitly recursive functions at all.\n  We've <i>almost<\/i> done that.  Our definition of <code>Y<\/code> is still\n  explicitly recursive.  However, we've taken a giant step, because this is\n  the <i>only<\/i> function in our language that needs to be explicitly\n  recursive in order to define recursive functions.  With this version\n  of <code>Y<\/code> we can go ahead and define other recursive functions (for\n  instance, defining <code>fibonacci<\/code> as <code>(Y\n  almost-fibonacci)<\/code>).<\/p><h3>Eliminating (most) explicit recursion (strict version)<\/h3><p>I said above that the definition of Y that we derived wouldn't work in a\n  strict language (like standard Scheme).  In a strict language, we evaluate\n  all the arguments to a function call before applying the function to its\n  arguments, whether or not those arguments are needed.  So if we have a\n  function <code>f<\/code> and we try to evaluate <code>(Y f)<\/code> using the\n  above definition, we get:<\/p><pre>\n  (Y f)\n  = (f (Y f))\n  = (f (f (Y f)))\n  = (f (f (f (Y f))))\n  etc.\n<\/pre><p>and so on ad infinitum.  The evaluation of <code>(Y f)<\/code> will never\n  terminate, so we will never get a usable function out of it.  This\n  definition of Y doesn't work for strict languages.<\/p><p>However, there is a clever hack that we can use to save the day and define\na version of Y that works in strict languages. The trick is to realize that\n<code>(Y f)<\/code> is going to become a function of one argument. Therefore,\nthis equality will hold:<\/p><pre>\n  (Y f) = (lambda (x) ((Y f) x))\n<\/pre><p>Whatever one-argument function <code>(Y f)<\/code> is, <code>(lambda (x) ((Y\nf) x))<\/code> has to be the same function. All you're doing is taking in a\nsingle input value <code>x<\/code> and giving it to the function defined by\n<code>(Y f)<\/code>. In a similar way, this will be true:<\/p><pre>\n  cos = (lambda (x) (cos x))\n<\/pre><p>It doesn't matter whether you use <code>cos<\/code> or <code>(lambda (x)\n(cos x))<\/code> as your cosine function; they will both do the same thing.<\/p><p>However, it turns out that <code>(lambda (x) ((Y f) x))<\/code> has a big\nadvantage when defining Y in a strict language. By the reasoning given above,\nwe should be able to define Y as follows:<\/p><pre>\n  (define Y\n    (lambda (f)\n      (f (lambda (x) ((Y f) x)))))\n<\/pre><p>Since we know that <code>(lambda (x) ((Y f) x))<\/code> is the same function\nas <code>(Y f)<\/code>, this is a valid version of Y which will work just as\nwell as the previous version, even though it's a bit more complicated (and\nperhaps a tiny bit slower in practice). We could use this version of Y to\ndefine the <code>factorial<\/code> function in lazy Scheme, and it would work\nfine.<\/p><p>The cool thing about <i>this<\/i> version of Y is that it will also work in\na strict language (like standard Scheme)! The reason for this is that when you\ngive Y a particular <code>f<\/code> to find the fixpoint of, it will return<\/p><pre>\n  (Y f) = (f (lambda (x) ((Y f) x)))\n<\/pre><p>This time, there is no infinite loop, because the inner <code>(Y f)<\/code>\nis kept inside a <code>lambda<\/code> expression, where it sits until it's\nneeded (since the body of a lambda expression is never evaluated in Scheme\nuntil the lambda expression is applied to its arguments). Basically, you're\nusing the lambda to delay the evaluation of <code>(Y f)<\/code>. So if\n<code>f<\/code> was <code>almost-factorial<\/code>, we would have this:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial (Y almost-factorial))\n<\/pre><p>Expanding out the call to Y, we have:<\/p><pre>\n  (define factorial\n    ((lambda (f) (f (lambda (x) ((Y f) x))))\n     almost-factorial))\n  ==&gt;\n  (define factorial\n    (almost-factorial (lambda (x) ((Y almost-factorial) x))))\n  ==&gt;\n  (define factorial\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n ((lambda (x) ((Y almost-factorial) x)) (- n 1))))))\n<\/pre><p>Here again, <code>(lambda (x) ((Y almost-factorial) x))<\/code> is the same\nfunction as <code>(Y almost-factorial)<\/code>, which is the fixpoint of\n<code>almost-factorial<\/code>, which is just the factorial function. However,\nthe <code>(Y almost-factorial)<\/code> in <code>(lambda (x) ((Y\nalmost-factorial) x))<\/code> won't be evaluated until the entire lambda\nexpression is applied to its argument, which won't happen until later (or not\nat all, for the factorial of zero). Therefore this factorial function will\nwork in a strict language, and the version of Y used to define it will also\nwork in a strict language.<\/p><p>I realize that the preceding discussion and derivation is nontrivial, so\n  don't be discouraged if you don't get it right away.  Just sleep on it,\n  play with it in your mind and with your trusty DrScheme interpreter, and\n  you'll eventually get it.<\/p><p>At this point, we've accomplished everything we've set out to accomplish, except for one tiny little detail: we haven't yet derived the Y combinator itself.<\/p><hr style=\"width: 75%;\" \/><h2>Deriving the Y combinator<\/h2><h3>The lazy (normal-order) Y combinator<\/h3><p>At this point, we want to define not just Y, but a Y <i>combinator<\/i>.\n  Note that the previous (lazy) definition of Y:<\/p><pre>\n  (define Y \n    (lambda (f) \n      (f (Y f))))\n<\/pre><p>is a valid definition of Y but is not a Y combinator, since the definition\nof Y refers to Y itself. In other words, this definition is explicitly\nrecursive. A combinator isn't allowed to be explicitly recursive; it has to be\na lambda expression with no free variables (as I mentioned above), which means\nthat it can't refer to its own name (if it even has a name) in its definition.\nIf it did, the name would be a free variable in the definition, as we have in\nour definition of Y:<\/p><pre>\n  (lambda (f) \n    (f (Y f)))\n<\/pre><p>Note that Y in this definition is free; it isn't the bound variable of any\nlambda expression.  So this is not a combinator.<\/p><p>Another way to think about this is that you should be able to replace the\nname of a combinator with its definition everywhere it's found and have\neverything still work. (Can you see why this wouldn't work with the explicitly\nrecursive definition of Y? You would get into an infinite loop and you'd never\nbe able to replace all the Ys with their definitions.) So whatever the Y\ncombinator will be, it will not be explicitly recursive. From this\nnon-recursive function we will be able to define whatever recursive functions\nwe want.<\/p><p>I'm going to go back a bit to our original problem and derive a Y\ncombinator from the bottom up. After I've done that I'll check to make sure that it is a fixpoint combinator, like the versions of Y we've already seen. In what follows I will borrow (steal) liberally from a very elegant derivation of the Y combinator sent to me by Eli Barzilay (thanks, Eli!), who is one of the DrScheme developers and an all-around Scheme uberstud.<\/p><p>Recall our original recursive <code>factorial<\/code> function:<\/p><pre>\n  (define (factorial n)\n    (if (= n 0)\n        1\n        (* n (factorial (- n 1)))))\n<\/pre><p>Recall that we want to define a version of this without the explicit\n  recursion.  One way we could do this is to pass the factorial function\n  itself as an extra argument when you call the function:<\/p><pre>\n  ;; This won't work yet:\n  (define (part-factorial self n)\n    (if (= n 0)\n        1\n        (* n (self (- n 1)))))\n<\/pre><p>Note that <code>part-factorial<\/code> is not the same as\n  the <code>almost-factorial<\/code> function described above.  We would have\n  to call this <code>part-factorial<\/code> function in a different way to get\n  it to compute factorials:<\/p><pre>\n  (part-factorial part-factorial 5) ==&gt; 120\n<\/pre><p>This is not explicitly recursive because we send along an extra copy of the\n<code>part-factorial<\/code> function as the <code>self<\/code> argument.\nHowever, it won't work unless the point of recursion calls the function the\nexact same way:<\/p><pre>\n  (define (part-factorial self n)\n    (if (= n 0)\n        1\n        (* n (self self (- n 1)))))  ;; note the extra \"self\" here\n  (part-factorial part-factorial 5) ==&gt; 120\n<\/pre><p>This works, but now we have moved away from our original way of calling\n  the factorial function.  We can move back to something closer to our\n  original version by rewriting it like this:<\/p><pre>\n  (define (part-factorial self)\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n ((self self) (- n 1))))))\n  ((part-factorial part-factorial) 5) ==&gt; 120\n  (define factorial (part-factorial part-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>Pause for a second here. Notice that we've <i>already<\/i> defined a version\nof the factorial function without using explicit recursion anywhere! This is\nthe most crucial step. Everything else we do will be concerned with packaging\nwhat we've already done so that we can easily re-use it with other\nfunctions.<\/p><p>Now let's try to get back something like our <code>almost-factorial<\/code>\n  function by pulling out the <code>(self self)<\/code> call using\n  a <code>let<\/code> expression outside of a <code>lambda<\/code>:<\/p><pre>\n  (define (part-factorial self)\n    (let ((f (self self)))\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial (part-factorial part-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>This will work fine in a lazy language. In a strict language, the\n<code>(self self)<\/code> call in the <code>let<\/code> statement will send us\ninto an infinite loop, because in order to calculate <code>(part-factorial\npart-factorial)<\/code> (in the definition of <code>factorial<\/code>) you will\nfirst have to calculate <code>(part-factorial part-factorial)<\/code> (in the\n<code>let<\/code> expression). (For fun: figure out why this wasn't a problem\nwith the previous definition.) I'll let this go for now, because I want to\ndefine the lazy Y combinator, but in the next section I'll solve this problem\nin the same way we solved it before (by wrapping a <code>lambda<\/code> around\nthe <code>(self self)<\/code> call). Note that in a lazy language, the\n<code>(self self)<\/code> call in the <code>let<\/code> statement will never be\nevaluated unless <code>f<\/code> is actually needed (for instance, if <code>n =\n0<\/code> then <code>f<\/code> isn't needed to compute the answer, so\n<code>(self self)<\/code> won't be evaluated). Understanding how lazy languages\nevaluate expressions is not trivial, so don't worry if you find this a little\nconfusing. I recommend you experiment with the code using the lazy Scheme\nlanguage level of DrScheme to get a better feel for what's going on.<\/p><p>It turns out that any <code>let<\/code> expression can be converted into an\n  equivalent <code>lambda<\/code> expression using this equation:<\/p><pre>\n  (let ((x &lt;expr1&gt;)) &lt;expr2&gt;)\n  ==&gt; ((lambda (x) &lt;expr2&gt;) &lt;expr1&gt;)\n<\/pre><p>where <code>&lt;expr1&gt;<\/code> and <code>&lt;expr2&gt;<\/code> are\n  arbitrary Scheme expressions.  (I'm only considering <code>let<\/code>\n  expressions with a single binding and <code>lambda<\/code> expressions with\n  a single argument, but the principle can easily be generalized\n  to <code>lets<\/code> with multiple bindings and <code>lambdas<\/code> with\n  multiple arguments.)  This leads us to:<\/p><pre>\n  (define (part-factorial self)\n    ((lambda (f)\n       (lambda (n)\n         (if (= n 0)\n             1\n             (* n (f (- n 1))))))\n     (self self)))\n  (define factorial (part-factorial part-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>If you look closely, you'll see that we have our old friend the\n<code>almost-factorial<\/code> function embedded inside the\n<code>part-factorial<\/code> function. Let's pull it outside:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define (part-factorial self)\n    (almost-factorial\n      (self self)))\n  (define factorial (part-factorial part-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>I don't know about you, but I'm getting pretty fed up with this\n  whole <code>(part-factorial part-factorial)<\/code> thing, and I'm not going\n  to take it anymore!  Fortunately, I don't have to; I can first rewrite\n  the <code>part-factorial<\/code> function like this:<\/p><pre>\n  (define part-factorial\n    (lambda (self)\n      (almost-factorial\n        (self self))))\n<\/pre><p>Then I can rewrite the <code>factorial<\/code> function like this:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial\n    (let ((part-factorial (lambda (self) \n                            (almost-factorial \n                              (self self)))))\n      (part-factorial part-factorial)))\n  (factorial 5) ==&gt; 120\n<\/pre><p>The <code>factorial<\/code> function can be written a little more concisely\n  by changing the name of <code>part-factorial<\/code> to <code>x<\/code>\n  (since we aren't using this name anywhere else now):<\/p><pre>\n  (define factorial\n    (let ((x (lambda (self) \n               (almost-factorial (self self)))))\n      (x x)))\n<\/pre><p>Now let's use the same <code>let ==&gt; lambda<\/code> trick we used above to get:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial\n    ((lambda (x) (x x))\n     (lambda (self) \n       (almost-factorial (self self)))))\n  (factorial 5) ==&gt; 120\n<\/pre><p>And again, to make this definition a little more concise, we can\n  rename <code>self<\/code> to <code>x<\/code> to get:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial\n    ((lambda (x) (x x))\n     (lambda (x) \n       (almost-factorial (x x)))))\n  (factorial 5) ==&gt; 120\n<\/pre><p>Note that the two <code>lambda<\/code> expressions in the definition\n  of <code>factorial<\/code> both are functions of <code>x<\/code>, but the\n  two <code>x<\/code>'s don't conflict with each other.  In fact, we could\n  have renamed <code>self<\/code> to <code>y<\/code> or almost any other name,\n  but it'll be convenient to use <code>x<\/code> in what follows.<\/p><p>We're almost there!  This works fine, but it's too specific to\n  the <code>factorial<\/code> function.  Let's change it to a\n  generic <code>make-recursive<\/code> function that makes recursive functions\n  from non-recursive ones (sound familiar?):<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define (make-recursive f)\n    ((lambda (x) (x x))\n     (lambda (x) (f (x x)))))\n  (define factorial (make-recursive almost-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>The <code>make-recursive<\/code> function is in fact the long-sought lazy Y\n  combinator, also known as the <i>normal-order Y combinator<\/i>, so let's\n  write it that way:<\/p><pre>\n  (define almost-factorial\n    (lambda (f)\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define (Y f)\n    ((lambda (x) (x x))\n     (lambda (x) (f (x x)))))\n  (define factorial (Y almost-factorial))\n<\/pre><p>I'm going to expand out the definition of Y a little bit:<\/p><pre>\n  (define Y \n    (lambda (f)\n      ((lambda (x) (x x))\n       (lambda (x) (f (x x))))))\n<\/pre><p>Note that we can apply the inner <code>lambda<\/code> expression to\n  its argument to get an equivalent version of Y:<\/p><pre>\n  (define Y \n    (lambda (f)\n      ((lambda (x) (f (x x)))\n       (lambda (x) (f (x x))))))\n<\/pre><p>What this means is that, for a given function <code>f<\/code> (which is a\n  non-recursive function like <code>almost-factorial<\/code>), the\n  corresponding recursive function can be obtained first by\n  computing <code>(lambda (x) (f (x x)))<\/code>, and then applying\n  this <code>lambda<\/code> expression to itself.  This is the usual\n  definition of the normal-order Y combinator.<\/p><p>The only thing left to do is to check that this Y combinator is a fixpoint\ncombinator (which it has to be in order to compute the right thing). To do\nthis we have to demonstrate that this equation is correct:<\/p><pre>\n  (Y f) = (f (Y f))\n<\/pre><p>From the definition of the normal-order Y combinator given above, we\n  have:<\/p><pre>\n  (Y f)\n  = ((lambda (x) (f (x x)))\n     (lambda (x) (f (x x))))\n<\/pre><p>Now apply the first lambda expression to its argument, which is the second lambda expression, to get this:<\/p><pre>\n  = (f ((lambda (x) (f (x x)))\n        (lambda (x) (f (x x)))))\n  = (f (Y f))\n<\/pre><p>as desired.  So, not only is the normal-order Y combinator also a fixpoint\n  combinator, it's just about the most obvious fixpoint combinator there is,\n  in that the proof that it's a fixpoint combinator is so trivial.<\/p><p>If you've made it through all of this derivation, you should pat yourself\n  on the back and take a well-deserved break.  When you come back, we'll\n  finish off by deriving...<\/p><h3>The strict (applicative-order) Y combinator<\/h3><p>Let's pick up the previous derivation just before the point where it failed\nfor strict languages:<\/p><pre>\n  (define (part-factorial self)\n    (lambda (n)\n      (if (= n 0)\n          1\n          (* n ((self self) (- n 1))))))\n  ((part-factorial part-factorial) 5) ==&gt; 120\n  (define factorial (part-factorial part-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>Up to this point, everything works in a strict language.  Now if we pull\nthe <code>(self self)<\/code> out into a <code>let<\/code> expression as before,\nwe have:<\/p><pre>\n  (define (part-factorial self)\n    (let ((f (self self)))\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n  (define factorial (part-factorial part-factorial))\n  (factorial 5) ==&gt; 120\n<\/pre><p>As I said above, this will not work in a strict language, because whenever\nthe <code>factorial<\/code> function is called it will evaluate the function\ncall <code>(part-factorial part-factorial)<\/code>, and when that function call\nis evaluated it will first evaluate <code>(self self)<\/code> as part of the\n<code>let<\/code> expression, which in this case will be <code>(part-factorial\npart-factorial)<\/code>, leading to an infinite loop of <code>(part-factorial\npart-factorial)<\/code> calls.<\/p><p>We saw above that the way around problems like this is to realize that what\nwe are trying to evaluate are functions of one argument. In this case,\n<code>(self self)<\/code> will be a function of one argument (it's going to \nbe the same as <code>(part-factorial part-factorial)<\/code>, which is just\nthe <code>factorial<\/code> function).  We can wrap a lambda expression around\nthis function to get an equivalent function:<\/p><pre>\n  (define (part-factorial self)\n    (let ((f (lambda (y) ((self self) y))))\n      (lambda (n)\n        (if (= n 0)\n            1\n            (* n (f (- n 1)))))))\n<\/pre><p>All we've done here is convert <code>(self self)<\/code>, a function of one\nargument, to <code>(lambda (y) ((self self) y))<\/code>, an equivalent function\nof one argument (we saw this trick earlier). I'm using <code>y<\/code> instead\nof <code>x<\/code> as the variable binding of the new lambda expression so as\nnot to cause name conflicts later on in the derivation when <code>self<\/code>\ngets renamed to <code>x<\/code>, but I could have chosen another name as\nwell.<\/p><p>After we've done this, the <code>part-factorial<\/code> function will now\nwork even in a strict language. That's because once <code>(part-factorial\npart-factorial)<\/code> is evaluated, as part of evaluating the\n<code>let<\/code> expression the code <code>(lambda (x) ((self self) x))<\/code>\nwill be evaluated. Unlike before, this will <i>not<\/i> send us into an\ninfinite loop; the lambda expression won't be evaluated further until it's\napplied to its argument. This lambda wrapper doesn't change the value of the\nthing it wraps, but it does delay its evaluation, which is all we need to get\nthe definition of <code>part-factorial<\/code> to work in a strict\nlanguage.<\/p><p>And that's the trick. After that, we carry through every other step of the derivation in exactly the same way. We end up with this definition of the strict Y combinator:<\/p><pre>\n  (define Y \n    (lambda (f)\n      ((lambda (x) (f (lambda (y) ((x x) y))))\n       (lambda (x) (f (lambda (y) ((x x) y)))))))\n<\/pre><p>This can also be written in the equivalent form:<\/p><pre>\n  (define Y \n    (lambda (f)\n      ((lambda (x) (x x))\n       (lambda (x) (f (lambda (y) ((x x) y)))))))\n<\/pre><p>Hopefully, you can see why this is equivalent. Either of these are the\nstrict Y combinator, or as it's called in the technical literature, the\n<i>applicative-order Y combinator<\/i>. In a strict language (like standard\nScheme) you can use this to define the factorial function in the usual\nway:<\/p><pre>\n  (define factorial (Y almost-factorial))\n<\/pre><p>I recommend you try this out with DrScheme, and lo! marvel at the awesome power of the applicative-order Y combinator, that which hath created recursion where no recursion hath previously existed.<\/p><hr style=\"width: 75%;\" \/><h2>Other matters<\/h2><h3>Practical applications<\/h3><p>This article has (I hope) convinced you that you don't need to have\nexplicit recursion built in to a language in order for that language to allow\nyou to define recursive functions, as long as the language supports\nfirst-class functions so that you can define a Y combinator. However, I don't\nwant to leave you with the notion that recursion in real computer languages is\nimplemented this way. In practice, it's far more efficient to just implement\nrecursion directly in a computer language than to use the Y combinator. There\nare lots of other interesting issues that come up when considering how to\nimplement recursion efficiently, but those issues are beyond the scope of this\narticle. The point is that implementing recursion using the Y combinator is\nmainly of theoretical interest.<\/p><p>That said, in the paper <a href=\"http:\/\/citeseer.ist.psu.edu\/mcadams01practical.html\" rel=\"nofollow\">Y in Practical\nPrograms<\/a>, Bruce McAdams discusses a few ways in which Y can be used to\ndefine variants of recursive functions that <i>e.g.<\/i> print traces of their\nexecution or automatically memoize their execution to give greater efficiency\n(as well as a few more esoteric applications), so Y isn't <i>just<\/i> a\ntheoretical construct.<\/p><h3>Mutual Recursion<\/h3><p>Experienced functional programmers and\/or particularly astute readers\nmay have noticed that I didn't describe how to use the Y combinator to\nimplement <i>mutual recursion<\/i>, which is where you have two or more\nfunctions which all call each other.  The simplest example I can think of to\nillustrate mutual recursion are the following pair of functions which determine\nwhether a non-negative integer is even or odd:<\/p><pre>\n  (define (even? n)\n    (if (= n 0)\n        #t\n        (odd? (- n 1))))\n  (define (odd? n)\n    (if (= n 0)\n        #f\n        (even? (- n 1))))\n<\/pre><p>Before you start yelling at me, yes, I know that this isn't the most\nefficient way to compute evenness or oddness \u2014 it's just to illustrate\nwhat mutual recursion is. Any computer language that supports recursive\nfunction definitions has to support mutual recursion as well, but I haven't\nshown you how to use Y to define mutually-recursive functions. I'm going to\ncop out here because I think this article is long enough as it is, but rest\nassured that it is possible to define analogs of Y that can define\nmutually-recursive functions.<\/p><hr style=\"width: 75%;\" \/><h2>Further reading<\/h2><ul><li><p>The Wikipedia article on the <a href=\"http:\/\/en.wikipedia.org\/wiki\/Y_combinator\" rel=\"nofollow\"> Y combinator<\/a> is somewhat\ndifficult reading, but it has some interesting material I didn't cover\nhere.<\/p><\/li>\n<li><p><a href=\"http:\/\/www.amazon.com\/Little-Schemer-Daniel-P-Friedman\/dp\/0262560992\/ref=pd_bbs_sr_1?ie=UTF8&amp;s=books&amp;qid=1218099185&amp;sr=8-1\" rel=\"nofollow\">The\nLittle Schemer<\/a>, 4th. ed., by Dan Friedman and Matthias Felleisen. Chapter\n9 has a derivation of the Y combinator which is what got me interested in this\nsubject.<\/p><\/li>\n<li><p>The article <a href=\"http:\/\/citeseer.ist.psu.edu\/mcadams01practical.html\" rel=\"nofollow\">Y in Practical\nPrograms<\/a>, by Bruce McAdams, which was referred to in the previous\nsection.<\/p><\/li>\n<\/ul><hr style=\"width: 75%;\" \/><h2>Acknowledgments<\/h2><p>I would like to thank the following people:<\/p><ul><li><p>Everyone who commented on my first blog post on the Y combinator, and\nalso everyone who comments on this article.<\/p><\/li>\n<li><p>Eli Barzilay, for a very interesting email discussion on this subject.\nThe derivation of the normal-order Y combinator is taken directly from Eli\n(with permission).<\/p><\/li>\n<li><p>My friend Darius Bacon for the poem. I'd also like to apologize to the\nestate of Kurt Vonnegut for abusing his work. The original poem appeared in\nVonnegut's brilliant novel <a href=\"http:\/\/www.amazon.com\/Cats-Cradle-Kurt-Vonnegut\/dp\/038533348X\/ref=pd_bbs_sr_1?ie=UTF8&amp;s=books&amp;qid=1217491463&amp;sr=8-1\" rel=\"nofollow\">Cat's\nCradle<\/a>. If you haven't read it, you should do so as soon as\npossible.<\/p><\/li>\n<li><p>All the <a href=\"http:\/\/www.drscheme.org\" rel=\"nofollow\">DrScheme<\/a> implementors for\ngiving me a terrific tool with which to explore this subject.<\/p><\/li>\n<li><p>The authors of the book <a href=\"http:\/\/www.amazon.com\/Little-Schemer-Daniel-P-Friedman\/dp\/0262560992\/ref=pd_bbs_sr_1?ie=UTF8&amp;s=books&amp;qid=1217491595&amp;sr=8-1\" rel=\"nofollow\">The\nLittle Schemer<\/a>, Dan Friedman and Matthias Felleisen. This article is (in\nmy mind at least) a massive expansion of chapter 9 of their book.<\/p><\/li>\n<\/ul>","created_at":"2018-11-07T16:44:00+0000","updated_at":"2018-11-07T16:44:00+0000","annotations":[],"mimetype":"text\/html","reading_time":44,"domain_name":"mvanier.livejournal.com","preview_picture":"https:\/\/l-stat.livejournal.net\/img\/sign.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":296,"title":"Ledger CLI cheatsheet","url":"https:\/\/devhints.io\/ledger","content":"<h3 id=\"basic-usage\">Basic usage<\/h3><pre class=\"language-bash\">$ ledger bal\n$ ledger reg\n<\/pre>\n<pre class=\"language-bash\">$ ledger reg grocery  # show entries for 'grocery'\n$ ledger bal assets   # check if i'm broke\n<\/pre>\n<pre class=\"language-bash\">  -b 01\/01   # --begin\n  -e 01\/31   # --end\n  -S date    # --sort\n  -S amount\n<\/pre>\n<h3 id=\"examples\">Examples<\/h3>\n<pre class=\"language-bash\"># any\/all matches\n  ledger bal Rent Transportation  # any\n  ledger bal Income and Job       # all\n  ledger bal Expenses and not (Drinks or Food)\n<\/pre>\n<pre class=\"language-bash\"># what did I spend on most? (--sorted)\n  ledger reg Expenses -S amount\n<\/pre>\n<pre class=\"language-bash\"># how much did I have at this date? (--end)\n  ledger bal -e 01\/15 ^Assets ^Liabilities\n<\/pre>\n<pre class=\"language-bash\"># how much did I spend and earn this month?\n  ledger bal ^Expenses ^Income --invert\n<\/pre>\n<pre class=\"language-bash\"># how much was spent over the course of 3 days? (totalled)\n  ledger reg -b 01\/25 -e 01\/27 --subtotal\n  ledger reg -b 01\/25 -e 01\/27 --subtotal grocery\n<\/pre>\n<h2 id=\"format\">Format<\/h2>\n<h3 id=\"basic-format\">Basic format<\/h3>\n<pre>2013\/01\/03 * Rent for January\n  Expenses:Rent   $600.00\n  Assets:Savings\n<\/pre>\n<p><code>*<\/code> = cleared, <code>!<\/code> = pending<\/p>\n<pre>2008\/01\/01=2008\/01\/14 Client invoice\n<\/pre>\n<p>It can mean anything you want, eg, for the estimated date you\u2019ll be paid.<\/p>\n<h3 id=\"balance-assertions\">Balance assertions<\/h3>\n<pre>2008\/01\/01 * KFC\n  Expenses:Food    $20\n  Assets:Cash     $-20  = $500\n<\/pre>\n<p><code>Cash $X = $500<\/code> ensures Cash is at $500 after the transaction.<\/p>\n<h3 id=\"balance-assignment\">Balance assignment<\/h3>\n<pre class=\"language-bash\">2008\/01\/01 * Cash balance\n  Assets:Cash              = $500\n  Equity:Adjustments\n2008\/01\/01 * KFC\n  Expenses:Food            $20\n  Assets:Cash              = $500\n<\/pre>\n<p><code>ACCOUNT = $500<\/code> figures out what\u2019s needed to make it $500.<\/p>\n<h3 id=\"payables\">Payables<\/h3>\n<pre class=\"language-bash\">2008\/04\/25 * Rent\n  (Assets:Checking)  -$200\n  Expenses:Rent\n<\/pre>\n<h3 id=\"commodities\">Commodities<\/h3>\n<pre class=\"language-bash\">; cost per item\n2010\/05\/31 * Market\n  Assets:Fridge                35 apples @ $0.42\n  Assets:Cash\n<\/pre>\n<pre class=\"language-bash\">; total cost\n2010\/05\/31 * Market\n  Assets:Fridge                35 apples @@ $14.70\n  Assets:Cash\n<\/pre>\n<pre class=\"language-bash\">; fixed lot prices\n2010\/05\/31 * Gas\n  Expenses:Gasoline             11 GAL {=$2.299}\n  Assets:Cash\n<\/pre>\n<h3 id=\"commodity-definitions\">Commodity definitions<\/h3>\n<pre>commodity $\n  note American Dollars\n  format $1,000.00\n  nomarket\n  default\n<\/pre>\n<h3 id=\"budgeting\">Budgeting<\/h3>\n<pre>~ Monthly\n  Expenses:Rent  $500\n  Expenses:Food  $100\n  Expenses        $40 ; everything else\n  Assets\n~ Yearly\n<\/pre>\n<pre class=\"language-bash\">ledger bal --budget Expenses\nledger bal --unbudgeted Expenses\n<\/pre>\n<pre>; line comment\n# also line comment\n% also line comment\n| also line comment\n* also line comment\n<\/pre>\n<h2 id=\"querying\">Querying<\/h2>\n<h3 id=\"periods\">Periods<\/h3>\n<pre>[interval] [begin] [end]\n<\/pre>\n<pre>interval:\n  every day|week|month|quarter|year\n  every N days|weeks|...\n  daily|weekly|...\n<\/pre>\n<pre>begin:\n  from &lt;spec&gt;\nend:\n  to &lt;spec&gt;\n<\/pre>\n<pre>spec:\n  2004\n  2004\/10\/1\n<\/pre>\n<pre class=\"language-bash\">$ ledger bal|reg --period \"until aug\"\n$ ledger bal|reg --period \"last oct\"\n$ ledger bal|reg --period \"every week\"\n<\/pre>\n<h3 id=\"register\">Register<\/h3>\n<pre class=\"language-bash\">$ ledger reg\n<\/pre>\n<pre class=\"language-bash\">  -D, --daily\n  -W, --weekly\n  -M, --monthly\n      --quarterly\n  -Y, --yearly\n  -s, --subtotal\n  --start-of-week monday\n<\/pre>\n<pre class=\"language-bash\">  -S, --sort date\n  -S, --sort amount\n<\/pre>\n<h3 id=\"filters\">Filters<\/h3>\n<pre class=\"language-bash\">-b, --begin DATE\n-e, --end DATE\n-d payee =~ \/pioneer\/\n-C, --cleared    # (with *)\n-U, --uncleared  # (no *)\n    --pending    # (with !)\n-R, --real       # ignore virtual postings (eg: \"(Cash)  $-400\")\n-L, --actual     # no automated postings (eg: \"= \/^Income\/\")\n-r, --related   # show the other side\n                # \"reg -r savings\" shows where it comes from)\n<\/pre>\n<h3 id=\"queries\">Queries<\/h3>\n<pre>^regex$\n@payee\n%tag\n=note\n#code\nterm and term\nterm or term\nnot term\n\\( term \\)\n<\/pre>\n<p>Example:<\/p>\n<pre>ledger r ^expenses and @Denny's\nledger r food and @Starbucks and not dining\n<\/pre>\n<h3 id=\"display\">Display<\/h3>\n<pre class=\"language-bash\">-n, --collapse       # [register] collapse entries\n                     # [balance] no grand total\n-s, --subtotal       # [balance] show sub-accounts\n                     # [other] show subtotals\n--flat\n<\/pre>\n<pre class=\"language-bash\">2008\/01\/01=2008\/01\/14 Client invoice  ;  estimated date you'll be paid\n  Assets:Accounts Receivable            $100.00\n  Income: Client name\n<\/pre>\n<p>Say you\u2019re in business. If you bill a customer, you can enter something like above.\nThen, when you receive the payment, you change it to:<\/p>\n<pre class=\"language-bash\">2008\/01\/01=2008\/01\/15 Client invoice ;  actual date money received\n  Assets:Accounts Receivable            $100.00\n  Income: Client name\n<\/pre>\n<h2 class=\"-one-column\" id=\"references\">References<\/h2>\n<ul class=\"-also-see\"><li><a href=\"http:\/\/ledger-cli.org\/3.0\/doc\/ledger3.html\">http:\/\/ledger-cli.org\/3.0\/doc\/ledger3.html<\/a><\/li>\n<li><a href=\"https:\/\/gist.github.com\/agaviria\/3317397\">https:\/\/gist.github.com\/agaviria\/3317397<\/a><\/li>\n<\/ul>","created_at":"2018-11-07T16:43:58+0000","updated_at":"2018-11-07T16:43:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"devhints.io","preview_picture":"https:\/\/assets.devhints.io\/previews\/ledger.jpg?t=20180915070711","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":297,"title":"Ask HN: Where do you deploy to in 2018 and how?","url":"https:\/\/news.ycombinator.com\/item?id=16472887","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16472887\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16472887\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=16472887\" class=\"storylink\">Ask HN: Where do you deploy to in 2018 and how?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        77 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=Narzerus\" class=\"hnuser\">Narzerus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16472887\">8 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16472887&amp;goto=item%3Fid%3D16472887\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Where%20do%20you%20deploy%20to%20in%202018%20and%20how%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Where%20do%20you%20deploy%20to%20in%202018%20and%20how%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16472887&amp;auth=5ae5d53065fa92a6553f0a6c7f09b4e839671e5d\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16472887\">83\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>I've been using Heroku for several years now, and as it happens to everyone eventually it becomes just too expensive.<p>I'm curious to know what people here use to deploy, and where are you hosting your apps.<\/p><p>Personally I'm looking for an experience as similar as possible to Heroku, any recommendations?<\/p><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:43:58+0000","updated_at":"2018-11-07T16:43:58+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":294,"title":"Stripe Atlas: Guide to Software as a Service pricing","url":"https:\/\/stripe.com\/atlas\/guides\/saas-pricing","content":"<p><a href=\"https:\/\/stripe.com\/atlas\">Stripe Atlas<\/a>, which helps internet companies get up and running, frequently gets asked questions by software-as-a-service entrepreneurs on how to improve the pricing and packaging of their SaaS\u00a0products.<\/p><p>After publishing our recent guide to <a href=\"https:\/\/stripe.com\/atlas\/guides\/business-of-saas\">low-touch and high-touch SaaS businesses<\/a>, we had a Q&amp;A session in our private forums for SaaS entrepreneurs specific to pricing and\u00a0packaging.<\/p><p>Some of the entrepreneurs generously allowed us to talk about their pricing strategy publicly, to help the next generation of SaaS\u00a0companies.<\/p><p>This advice is distilled from my career in running and consulting in various SaaS companies. Your mileage may vary; I\u2019d encourage you to experiment often and boldly with pricing, as it is the easiest needle to move in your company. (The tendency of most SaaS companies is to set prices without much consideration and leave them alone for years at a time. I\u2019d encourage you to revisit them\u00a0quarterly.)<\/p><p>By coincidence, most of the companies which we\u2019ll be talking about sell SaaS on a <a href=\"https:\/\/stripe.com\/atlas\/guides\/business-of-saas#low-touch-saas-sales\">low-touch<\/a> model. Pricing for high-touch SaaS is similarly nuanced; we\u2019ll cover it in the\u00a0future.<\/p><p>We\u2019ve organized the case studies based on the primary market for the\u00a0products.<\/p><h2 id=\"pricing-and-packaging-for-things-with-no-physical-form\">Pricing and packaging for things with no physical form<\/h2><p><strong>Pricing<\/strong> is easy enough to\u00a0understand.<\/p><p><strong>Packaging<\/strong> is a jargon word coming from marketing of consumer packaged goods, like soap. In SaaS, it refers to a company which has one particular underlying product which they make available to different customers in slightly different ways at different price\u00a0points.<\/p><p>In low-touch SaaS, the most common way packages are presented are as different columns in a pricing grid, with each column corresponding to a <strong>plan<\/strong>, offered at a different price, with differential access to features or maximum allowable usage along some axis interesting to the business. This might be users (referred to pervasively, and unfortunately, in the industry as <strong>seats<\/strong>) or servers or something with a closer connection to the problem\u00a0domain.<\/p><h2 id=\"cirtru-case-study\">Cirtru case study<\/h2><p>Cirtru helps match people with new roommates. The customer is the person seeking a roommate, not the person seeking a new\u00a0apartment.<\/p><p>Their existing\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/cirtru.png\" alt=\"Cirtru pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"reduce-decision-fatigue-for-customers\">Reduce decision fatigue for customers<\/h3><p>You\u2019re asking the user to make a big decision here: Concierge or Premium Listing. Those achieve different goals for different\u00a0people.<\/p><p>The core insight here is great, which is that some people want to buy a tool and some people have more money and just want to buy an outcome. The desired outcome is \u201cFind me a paying tenant who probably won\u2019t skip out on rent or murder me in my\u00a0sleep.\u201d<\/p><p>I\u2019d probably position that choice\u00a0explicitly:<\/p><p><em>Find me a roommate \/ I want to do my own legwork finding a\u00a0roommate<\/em><\/p><p>I\u2019d default people into the first option and either deemphasize the details of the second (via greying them out or something) or hide them until someone interacted with the downsell\u00a0option.<\/p><p>This decreases my cognitive load to make the\u00a0decision:<\/p><p>\u201cDo I want you to find me a roommate?\u201d Of course yes, that is why I am\u00a0here.<\/p><p>\u201cHow long should we\u00a0look?\u201d<\/p><p>And my answer to that is probably \u201cI don\u2019t know, as much as required, but I want the roommate fast\u201d, which suggests to me as the user that I don\u2019t conceive of the value of this service along the same axis that you conceive of charging for it. You want to charge for your costs. I don\u2019t want to pay you more for concierging longer; I want to pay you more for concierging more efficiently. The ideal for me is probably paying you a success\u00a0fee.<\/p><p>Whether that is operationally sound for you to offer or not, I don\u2019t\u00a0know.<\/p><h3 id=\"sell-more-talk-less\">Sell more; talk less<\/h3><p>I\u2019d boil those bullet points down relentlessly. Consider this copy\u00a0instead:<\/p><ul><li>We manage your listing on Cirtu and other portals [like Craigslist, presumably? If so, name them.]<\/li>\n  <li>We screen communications; you get everything sorted in one place.<\/li>\n  <li>You talk to only those prospects you want to talk to; you pick your roommate.<\/li>\n  <li>~~$400~~ $250 (limited time offer)<\/li>\n<\/ul><p><strong>The Cirtru guarantee<\/strong>: We\u2019re very good at what we do. We\u2019ll refund 50% of the purchase price if we can\u2019t find you a roommate in 15 days or 75% if we can\u2019t find you a roommate in 30\u00a0days.<\/p><h2 id=\"cointracker-case-study\">Cointracker case study<\/h2><p>Consumers are notoriously price sensitive, but sometimes software is sold to individuals who, by construction, probably have a lot of money. <a href=\"https:\/\/www.cointracker.io\/\">Cointracker<\/a> is a great example; it assists cryptocurrency <a href=\"https:\/\/en.wikipedia.org\/wiki\/HODL\">hodlers<\/a> with tax\u00a0compliance.<\/p><p>Their original\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/cointracker.png\" alt=\"Cointracker pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"getting-the-aesthetics-of-pricing-correct-matters-in-b2c\">Getting the aesthetics of pricing correct matters in B2C<\/h3><p>$X.99 pricing is not generally used in B2B or prosumer (professional consumers) services because it communicates cheapness; I\u2019d suggest you drop the 0.99 accordingly here, for aesthetic reasons. (No series of A\/B tests I\u2019ve ever done generated as many strong opinions as whether it was better to price a product at $49, $49.99, or $50. Despite bikeshedding this to death at several clients, I\u2019ve yet to see it really matter, so aesthetics win the day. If your data says differently, trust the\u00a0data.)<\/p><p>Lite \/ Standard \/ Premium are weak names for SaaS plans because they don\u2019t help a user make an instant decision on which plan is right for them. I\u2019d consider Hobbyist \/ Active Trader \/ Financial Professional. After last year, cryptocurrency enthusiasts certainly shouldn\u2019t be hurting for money to pay for tax data\u00a0collection.<\/p><p>I\u2019d generally advise against having Unlimited anything available for $999. If a hedge fund comes to you and says \u201cOh, awesome, I want to adopt this for my funds\u2019 tax reporting\u201d, clearly $999 is the wrong price for that user. Don\u2019t anchor the custom services which they almost certainly need to\u00a0$999.<\/p><h3 id=\"charge-more-generally\">Charge more, generally<\/h3><p>Charge more! Your offering presently labeled Premium has the value proposition \u201cYou could spend $5k talking to a non-specialist accountant who would give you the wrong answer here, resulting in you paying tens of thousands of dollars extra taxes and then get audited anyway because his work product would look obviously fishy. Or you can use our services, get the right answer, and have clearly high-quality work\u00a0product.\u201d<\/p><p>I\u2019d probably increase your pricing to $99 \/ $499 \/ $2,499. This intentionally prices out <a href=\"https:\/\/hn.algolia.com\/?query=patio11%20%22pathological%20customers%22&amp;type=comment&amp;dateRange=all&amp;prefix=false&amp;page=0&amp;sort=byDate\">pathological customers<\/a> on the lower end, who will be exceedingly difficult to deal with. It is my impression that this market is rife with them and that they will require a lot of\u00a0handholding.<\/p><p>If a person didn\u2019t make tens of thousands of dollars while doing 1,000 transactions in 2017 in cryptocurrencies then, bluntly, there is something wrong with that person\u2019s ability to make good decisions, and you should probably not seek their\u00a0custom.<\/p><h3 id=\"charge-more-for-priority-customer-service\">Charge more for priority customer service<\/h3><p>What does \u201cDedicated\u201d customer service mean? Does that imply that the lower plans don\u2019t have CS available? That\u2019s probably not the optimal way to phrase it; I might phrase it as \u201cPriority\u201d customer\u00a0service.<\/p><p>This is an easy thing for any SaaS company to operationalize, by the way\u2014all you need is a special email address that e.g. adds a label in your ticketing tool or gmail inbox, and then you read that label first every day while you\u2019re answering email. That is all priority has to\u00a0mean!<\/p><p>This is one of the easiest ways available to get price-insensitive customers to self-select as price-insensitive. Many SaaS entrepreneurs are emotionally discomfit by optimizations like this. You shouldn\u2019t be, if you have confidence that your product works as described; a well-off person telling you \u201cI am willing to pay extra for the white glove service\u201d is not only doing that with regards to <em>your<\/em> firm. They are reaching back across their history of working with other firms and declining to use the cheapest available service because their experiences with commodity services have not been wonderful. Telling them that you have a premium service available gives them permission to adopt your service, which is in your mutual interest, <em>and<\/em> helps you subsidize the availability of your service to more price-sensitive customers on your other\u00a0plans.<\/p><p>(If you\u2019ve followed the cryptocurrency space in the last year, you probably understand why many investors might have the reasonable fear that your baseline expectation for customer service is \u201cFeel free to send an email to \/dev\/null.\u201d Making it clear that you answer your emails is a differentiating feature in this\u00a0space.)<\/p><p>The SaaS industry often describes sales of very horizontal applications as \u201cselling to SMBs\u201d, even though small-to-medium-sized businesses have a more official definition at various governmental agencies than \u201cWell it\u2019s not an individual person and not Boeing so it must be an\u00a0SMB.\u201d<\/p><p>Entrepreneurs new to SaaS often assume that most SaaS is sold to undifferentiated SMBs. This is probably not accurate; while there are many examples of software products which are very horizontal (sold across a variety of industries, customer archetypes, and sizes of target firm), much SaaS is actually very\u00a0targeted.<\/p><p>This makes advice for SMB pricing a bit harder to give in the abstract, and as you\u2019ll see, a lot of the opportunity is using pricing\/packaging to assist your efforts in honing in on a concrete understanding of a target customer and delighting them, as opposed to being mediocre for a larger\u00a0audience.<\/p><h2 id=\"formapi-case-study\">FormAPI case study<\/h2><p>FormAPI makes it easy for programmers to generate\u00a0PDFs.<\/p><p>Their original\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/formAPI.png\" alt=\"FormAPI pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"pricing-a-saas-for-technical-users\">Pricing a SaaS for technical users<\/h3><blockquote>\n  <p>Should I remove the credit card form on the sign up page? I\u2019ve found that most people who sign up continue paying for the service. If I offer a free trial that doesn\u2019t require a credit card, does that usually result in a higher number of\u00a0conversions?<\/p>\n<\/blockquote><p>An excellent question! Net on net, removing the CC required upfront decreases conversions in the early life of a startup, until you\u2019re sophisticated with regards to onboarding, in-app messaging, lifecycle email, and direct reachout by a customer success team. I would not suggest you remove the CC required from this free\u00a0trial.<\/p><blockquote>\n  <p>How many plans should a SaaS\u00a0have?<\/p>\n<\/blockquote><p>You should expose few decisions to your users at signup. Attempting to forecast their use in advance is challenging for many users, so you should have less gradations between usage and give plans names which help users bucket\u00a0themselves.<\/p><p>I count 8 plans. I would recommend you condense to 3 plus \u201ccall us for enterprise pricing\u201d presented on the page. If you do have users which appear in all of these buckets, present 3 plans but keep the other ones in reserve; you can hide their existence on the pricing page until a link is clicked on, run a window over the plans (showing only a subset at any given time),\u00a0etc.<\/p><p>For\u00a0example:<\/p><table><tbody><tr><td><strong>Quick hack<\/strong><\/td>\n      <td>(e.g. an internal application or pilot)<\/td>\n      <td>50 PDFs per month<\/td>\n      <td>$49 \/ mo<\/td>\n    <\/tr><tr><td><strong>Production<\/strong><\/td>\n      <td>(e.g. invoice generation for an e-commerce application)<\/td>\n      <td>1,000 PDFs per month<\/td>\n      <td>$99 \/ mo<\/td>\n    <\/tr><tr><td><strong>Deployed<\/strong><\/td>\n      <td>(e.g. line of business application)<\/td>\n      <td>5,000 PDFs per month<\/td>\n      <td>$249 \/ mo<\/td>\n    <\/tr><tr><td><strong>Enterprise<\/strong><\/td>\n      <td>We\u2019d love to talk about your use case. Call us and speak directly with a founding engineer.<\/td>\n      <td>\u00a0<\/td>\n      <td>\u00a0<\/td>\n    <\/tr><\/tbody><\/table><h3 id=\"your-pricing-page-should-continue-the-sales-message\">Your pricing page should continue the sales message<\/h3><p><em>From starter plans to on-site and enterprise solutions \u2014 Choose the plan that\u2019s best for\u00a0you.<\/em><\/p><p>Nothing here promises value to the user; it\u2019s all about mechanics which only you care about. Instead, you should reiterate what FormAPI does and why the plan they pick will be right for\u00a0them:<\/p><p><strong>Save weeks of development work on PDF generation; fits easily in the budget for any\u00a0project.<\/strong><\/p><h3 id=\"moving-upmarket-towards-the-enterprise\">Moving upmarket towards the enterprise<\/h3><p>You\u2019ve indicated a desire to attempt to sell this in a high-touch fashion, and that is a natural evolution for many SaaS apps which start targeting developers but find themselves adopted by materially sized businesses. Your goal for the first 10 times you attempt to sell this to a large customer is not to get the price correct, it is to thoroughly understand what your users want out of this software and what it is worth to them. Charge based on\u00a0that.<\/p><p>Knowing nothing about a particular enterprise user, I would assume that for most of them this simply saves them 2~4 weeks of dev time at a fully-loaded cost of $25k a month plus maintenance, and accordingly I would probably start offering them two\u00a0options:<\/p><ul><li>$10k a year for our SaaS offering, which we either host for you or let you use on-premises<\/li>\n  <li>$25k one-time license + $5k a year maintenance commitment for the source code; you host it.<\/li>\n<\/ul><p>I would probe a lot with regards to what their expectations were regarding support SLAs; anything better than what you\u2019d offer a $500 a month account gets a line item added to the invoice and that line item should cost\u00a0$10k.<\/p><p>You should certainly walk up your prices if you discover that FormAPI is key to the processes of materially-sized businesses. If an insurance company builds their entire business around shipping forms betwee them, suppliers, and regulators, and they are on FormAPI, then that is a six figure deal,\u00a0minimally.<\/p><h2 id=\"humble-dot-case-study\">Humble Dot case study<\/h2><p>Humble Dot is a broad B2B app which helps teams keep informed about what other team members are working on. Their existing\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/humbledot.png\" alt=\"HumbleDot pricing: $50 per month\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><p>Time for me to deploy <a href=\"https:\/\/hn.algolia.com\/?query=patio11%20%22charge%20more%22&amp;type=comment&amp;dateRange=all&amp;prefix&amp;page=0&amp;sort=byDate\">my catchphrase<\/a>: <strong>Charge\u00a0more!<\/strong><\/p><p>$50 per month is not an appropriate maximum price for this product, because it will be instituted at materially-sized businesses to manage teams of professionals, and no initiative worth the repeated effort of professionals caps out near $50 per month. The best customers of Humble Dot would spend more on coffee discussing buying it than they would on actually buying it; that seems like a poor allocation of\u00a0value.<\/p><h3 id=\"pricing-pages-should-continue-the-sales-message\">Pricing pages should continue the sales message<\/h3><p>Humble Dot\u2019s pricing page leads\u00a0with:<\/p><blockquote>\n  <p>Transparent and Flat\u00a0Pricing<\/p>\n<\/blockquote><p>I would strongly encourage you to have the title of your pricing page act like a sales representative. You have a mental of the mind for someone who has arrived at this page, right? What is your one-thought encapsulation of why they would buy Humble Dot? It can\u2019t be simply that you\u2019re offering Transparent and Flat Pricing, because I can offer you Transparent and Flat Pricing for a tuba, and you are not particularly enthusiastic about buying a tuba\u00a0today.<\/p><p>I would encourage you to try a headline\u00a0like:<\/p><p><strong>We cost a lot less than not knowing what happened this\u00a0week.<\/strong><\/p><p>Similarly, your sub-header is giving me a response to a pricing objection. Do you know I have a pricing objection? I am probably on the pricing page because I have a pricing question, not an objection. Additionally, that is a very niche objection: if my company grows to the point where it has 100 people using Humble Dot, that doesn\u2019t make it 50x as expensive to me as when there were 2 people using Humble Dot, it makes it 2,500x more valuable because the communication overhead of pairwise status updates increases at approximately\u00a0O(n^2).<\/p><h3 id=\"dont-sell-unlimited-anything\">Don\u2019t sell unlimited anything<\/h3><p>Think of the largest customer that you can conceive of using HumbleDot. How does the anchor of \u201cEverything we offer costs $50\/month, flat\u201d impact your pricing discussion with that customer? It decimates your ability to charge appropriate amounts for the\u00a0software.<\/p><p>The impact of a relatively small business division missing its targets this quarter is hundreds of thousands or millions of dollars. That is the number you want to anchor enterprise sales against, not $50. You could pull teeth to get \u201cten freaking times what you usually charged\u201d or you can step into the reality of B2B sales, where the cheapest service you can conveniently describe\u2014 taking out trash so your employees don\u2019t have to\u2014starts at about $800 a month because it isn\u2019t even worth discussing a number below that with a business large enough to afford an\u00a0office.<\/p><p>I would probably have the $50 plan flat for companies up to 10 employees, after which point their minimum possible payroll is, what, $30k a month or so? (Assumptions: full-time white collar employees in countries with costs similar to the US.) After that, $5 or $10 per employee sounds totally reasonable. You will likely \u201ccost\u201d your best customers more than that in staff time just filling out the updates\u2014and, again, imagine how valuable those updates\u00a0are.<\/p><h2 id=\"firma-case-study\">FirmA case study<\/h2><p>FirmA, whose name we\u2019ve obscured at their request, automates financial performance analysis for businesses. They sell directly to end-users and through accountants as a sales\u00a0channel.<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/firm-a.png\" alt=\"FirmA pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"minimize-distinct-products-sharing-a-single-pricing-page\">Minimize distinct products sharing a single pricing page<\/h3><p>I would probably pitch your product for end-users and your product for accountants separately, on two separate pricing pages. You can link each page to each other in case someone ends up in the wrong\u00a0place.<\/p><p>This would free up both mental bandwidth and visual real estate for having multiple pricing plans, which you should almost certainly have. This product creates 1000X+ as much value at a $10 million a month company as at a $10k a month company, right? Why would you charge them the same\u00a0amount?<\/p><h3 id=\"directly-price-discriminating-by-customer-success\">Directly price discriminating by customer success<\/h3><p>I\u2019d probably make the axis to differentiate plans from each other revenue tiers: up to $500k yearly, $5 million yearly, $50 million yearly, Hey I Just Met You And This Is Crazy But Your CFO Has Budget So Call Me\u00a0Maybe.<\/p><p>I think you\u2019re likely targeting ACVs of ~$500 (your current pricing), $5k, and $20k+ given those rough tiers, and that you\u2019ll likely have to increase your level of sophistication on sales and service delivery to get into the higher tiers, but positioning yourself as capable of servicing them makes the entry offering more\u00a0credible.<\/p><h3 id=\"using-channel-strategies\">Using channel strategies<\/h3><p>I love the idea of selling through accountants as a <strong>channel strategy<\/strong>. (This is when a SaaS company sells to a party with a relationship with a pool of customers in return for that party selling their product through those relationships. Typically, this results in the channel taking a cut of the price the end-user\u00a0pays.)<\/p><p>You do not explicitly make the case for it here, which is that it allows accountants to benefit from the upside of selling a high-margin software product in addition to their lower-margin services\u00a0work.<\/p><p>As a note for folks who haven\u2019t worked in a consultancy before: you can model the margins in an accountancy at about 30% given that the accountancy has to pay salaries and expenses related to the employee accountants who do most of the work. When the accountancy adds a line item to their invoice for access to your software, that line item has ~80% margins. This is pretty attractive as an upsell for them, and for SMB-focused accountants, which do a supermajority of their billable hours for the year in tax season, it gives them both things to bill for in Q2 through Q4 and also an ongoing reason to be in touch with their clients. Services businesses generally benefit from improvements in their offerings which increase the cadence at which they are in front of their clients. It leads to happier clients, more opportunities for follow-on work, and more opportunities to collect\u00a0referrals.<\/p><p>As this example demonstrates, it is helpful to <em>thoroughly<\/em> understand the business of one\u2019s customers and\/or channel partners. Understanding allows you to make relatively small changes to your offering which create outsized value for them. Often the changes are, by the standards of a software company which can quickly write code, almost trivial to deliver; for example, co-branded or white-labeled PDF reports would likely be very compelling for accountancies using your software to give their customers the impression of bespoke, authoritative guidance without costing professional time on high-quality one-off visual design. That feature is relatively easy for a software team to deliver but has impressive \u201cwow factor\u201d both for the channel partner and the\u00a0end-user.<\/p><h2 id=\"vempathy-case-study\">Vempathy case study<\/h2><p>Vempathy helps use AI to analyze user interviews and usability tests for their emotional\u00a0content.<\/p><p>Their existing\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/vempathy.png\" alt=\"Vempathy pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><p>Minor comment: nobody can \u201cChoose plan\u201d on Enterprise\/Agency. I prefer the call-to-action \u201cCall us\u201d or \u201cSpeak with us\u201d, in both cases directing someone to a contact form. You should not offer a 14 day free trial of your enterprise offering; you can build the demo into the sales\u00a0process.<\/p><p>Can I recommend that instead of the Free plan getting a column that you immediately demote it to text under the columns with similar visual weight to \u201cNo risk, 30 day money back guarantee\u201d\u00a0gets?<\/p><p>Consider the copy <em>We also have a free plan which does not offer analysis<\/em> and hyperlinking \u201cfree\u00a0plan.\u201d<\/p><blockquote>\n  <p>Is it clear from the pricing page what we charge customers\u00a0for?<\/p>\n<\/blockquote><p>I read pricing pages for a living, which is not the case of most users of pricing pages. That said, the plan names are reasonably good and descriptive, so I think I\u2019d be able to bucket my business accordingly based on them. (I might still tweak them, but they\u2019re a lot better than Bronze \/ Silver \/ Gold, which would make me have to actually read what was\u00a0included.)<\/p><p>I\u2019d encourage you to make the pricing page less about the transactional details and more about winning the\u00a0sale.<\/p><h3 id=\"dont-call-the-pricing-page-the-pricing-page-on-the-pricing-page\">Don\u2019t call the pricing page the pricing page on the pricing page<\/h3><p>Plans &amp; Pricing is a good title tag for SEO purposes, though I would generally put the product name in it as well. Plans &amp; Pricing is a <em>terrible<\/em> headline; it chews lots of visual weight and customer brainsweat without paying its rent in improved conversions. The UX designer you\u2019re selling to <em>knows<\/em> they\u2019re on the pricing page; they just clicked a link or Googled with the specific intention of opening exactly this\u00a0page.<\/p><p>A perhaps better headline, in a style which I\u2019ve used to effect in my own software companies, is: <strong>\u201cPays for itself with a single [CONCISELY ARTICULATE YOUR\u00a0VALUE].\u201d<\/strong><\/p><p>\u201cPays for itself with the time you\u2019ll save analyzing one UX experiment\u201d, for example. You might have wording that people connect to better than\u00a0that.<\/p><blockquote>\n  <p>Is the barrier to trial low\u00a0enough?<\/p>\n<\/blockquote><p>If anything I\u2019d be looking to raise the barrier! How would you communicate that this is the very valuable AI-assisted emotional analysis software used by the most sophisticated UX designers and researchers? I would demote the risk reducers to their proper status as minor\u00a0details.<\/p><blockquote>\n  <p>Do you think the pricing tiers make sense with regards to limitations and price\u00a0points?<\/p>\n<\/blockquote><p>Two unlimiteds is at least one unlimited too many; I think you could probably eliminate \u201cUnlimited Collaborators\u201d without affecting the vast majority of customers. You can clarify it later on the page for people who are worried they have to pay per\u00a0seat.<\/p><p>Charging based on AI-analyzed recordings as the primary axis makes sense. I think that is likely so much more valuable than having multiple teams\/projects that I\u2019d consider hiding the number of teams\/projects from this table and only exposing that limitation in the software. See if anyone hits it. (If they complain that it wasn\u2019t disclosed before purchase, upgrade them for free. I anticipate nobody will care, which means it doesn\u2019t need to be on this\u00a0page.)<\/p><blockquote>\n  <p>Should we include an exhaustive features\u00a0list?<\/p>\n<\/blockquote><p>Will it cause someone to choose to buy the software? Probably not. Instead, continue the sales message: how specifically does this reduce the amount of human work to pull stunning insights from user recordings? Put that on the page; supplement with details about features where\u00a0appropriate.<\/p><p>It is generally easier to sell into SMB verticals (a particularly industry or customer archetype) than across all SMBs, because it is easier to laser-focus on their specific needs, understand the factors which impact their business, and speak the language of your\u00a0users.<\/p><h2 id=\"kitchenwhiz-case-study\">KitchenWhiz case study<\/h2><p>KitchenWhiz creates kitchen plans for homeowners and professionals like\u00a0carpenters.<\/p><p>Their existing\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/kitchenwhiz.png\" alt=\"KitchenWhiz pricing: $9.95 per month or $60 per year\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"saas-entrepreneurs-overestimate-the-benefit-of-low-prices-early\">SaaS entrepreneurs overestimate the benefit of low prices early<\/h3><p>Many SaaS entrepreneurs describe their products as less mature than the incumbents and therefore they price far under the incumbents. This is almost always a\u00a0mistake.<\/p><p>Your customers are not buying KitchenWhiz because it is bad but attractively priced. They are buying it because your value proposition is \u201cIt produces beautifully designed professionally-produced kitchen plans easily.\u201d Your customer is likely in the trades and this is a revenue-producing tool in their business. I would encourage you to ask them what a single remodeling job typically costs; I suspect that they quite literally discard more materials as waste than they spend on KitchenWhiz at any of your contemplated price\u00a0points.<\/p><p>You already are a revenue-producing tool for your customers, and you\u2019re going to quickly iterate on this product and make it even better than it is now. You should price a bit ahead of where you think your maturity is; this lets you \u201cgrow into\u201d the new price as your customers discover and onboard with you, and prevents you from having to go back to them and ask them for more money in the near\u00a0future.<\/p><p>I would probably sell 3 packages of KitchenWhiz, with arbitrary contents in them that I leave to your best judgment. These price points are my default price points for a B2B SaaS sold on the low-touch model; they work across a range of firms and\u00a0industries.<\/p><p><strong>Monthly<\/strong>: $49 \/ $99 \/\u00a0$249<\/p><p><strong>Annually<\/strong>: Same as above, but with one month\u00a0\u201cfree.\u201d<\/p><p>If you sell literally none of the $249 a month version then you can do away with it if you feel that is useful, but I think the market may surprise\u00a0you.<\/p><p>Note that this proposed pricing is 10X to 50X the <em>maximum<\/em> pricing contemplated by the existing pricing. This demonstrates the power of doubling down on the B2B market when you have a product which could theoretically be used by a consumer or a business. AppAmaGooBookSoft are doing their best to convince customers worldwide that software should be free (or as close at makes no difference) and subsidized by extremely lucrative advertising\/hardware\/e-commerce\/etc ecosystems attached to it. That is a hard market expectation to go\u00a0against.<\/p><p>If you want to price SaaS sustainably, you probably want to target primarily business\u00a0users.<\/p><h3 id=\"generous-refund-policies-are-to-your-advantage\">Generous refund policies are to your advantage<\/h3><p>I would generally encourage software sold on the low-touch model to offer refunds to customers who are dissatisfied. Refunds do cause a short-term cashflow hit, but dissatisfied customers are a much bigger threat to your business than that cashflow hit, because dissatisfaction that you don\u2019t know about causes churn that is hard to\u00a0control.<\/p><p>You should eagerly encourage customers who are very unhappy about your services to tell you why and with specificity; consider the refund a small consulting fee you\u2019re giving them if it helps\u00a0you.<\/p><p>Additionally, customers who are dissatisfied with a SaaS company\u2019s services can always escalate the decision to not refund to their bank. The bank will often take their customer\u2019s side over yours, since you are not their customer. It will be difficult for you to prevail in a dispute along the lines of \u201cThey promised me the software would do X \/ Y \/ Z but it didn\u2019t so I asked for my money back and they said no.\u201d Answering that dispute is a poorer use of your time than talking to customers or building software, and you likely wouldn\u2019t win it even if you were entirely accurate in describing the\u00a0software.<\/p><h3 id=\"to-demo-or-not-to-demo\">To demo or not to demo?<\/h3><p>You\u2019re trying to strike a balance between conversion rates and learning with the decision to offer free trials or not. I would maximize for learning right now: offer a call to literally everyone who starts using the software. Use it 80% for learning about their business and 20% for showing them the important parts of your\u00a0software.<\/p><p>You should use these calls for customer development and for sales, not training\u2014if they have questions, note them for later and produce screencasts or similar that you can show to\u00a0everyone.<\/p><p>I would probably use the fact of the call and concierge onboarding to tell everyone that of course you charge for your services, just like they charge for their services, but that as a limited time offer you\u2019ll retroactively waive one month of fees if they can commit to a 15 minute conversation about why the software didn\u2019t work for your\u00a0needs.<\/p><p>This is effectively a demo but puts the onus on the user to organize themselves to actually use the software or feel embarrassed, as opposed to the typical free trial, which is just \u201cTry it, or don\u2019t, either way really.\u201d Committing to putting a software product through its paces is <strong>real work<\/strong> which software people, as inveterate early adopters, often underestimate. You\u2019re asking a busy professional to change the way they do business with respect to an actual project with an actual customer who has an actual kitchen on the line. If you do not project manage them in the adoption of your software and process, you can reasonably expect for them to <em>not<\/em> make effective use of your trial during the time limit, mooting any benefit of it being\u00a0available.<\/p><p>This is a <em>very common<\/em> failure case for free software trials; they carry a cost which, despite not being charged to a credit card, is much more material than software people give it credit for. We like them too much and we should be more considered about whether offering them is in the interests of our businesses or our\u00a0customers.<\/p><h2 id=\"publica-case-study\">Publica case study<\/h2><p>Publica helps authors sell digital publications online. Their current\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/publica.png\" alt=\"Publica pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"name-pricing-plans-to-sell-them-to-the-right-users\">Name pricing plans to sell them to the right users<\/h3><p>The only purpose of the name of a plan, as communicated on the pricing page, is to provide the right customer the right nudge to choose that particular\u00a0plan.<\/p><p>If you think of the sophistication spectrum among authors who publish their work online for money, I don\u2019t think \u201cInitial\u201d likely helps any to make the correct decision; I would call that plan Hobbyist, both because it helps people who have yet to sell their first book select into that plan and it helps professional authors, who are quite vocal about that being their primary career identity, select out of the cheapest\u00a0plan.<\/p><p>\u201cAuthor\u201d is a better name but perhaps not wonderful. How would you feel about Professional, Professional Author, or Midlist Author? (Midlist is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Midlist\">term of art<\/a> in traditional\u00a0publishing.)<\/p><p>Flagging a plan as recommended often increases the number of people who select it, but I would probably not name the plan \u201cRecommended.\u201d I might describe it as an Imprint (if I suspect my customers are either traditionally published authors or emotionally affiliated with the same) or something suggesting a small business (Cottage Publisher,\u00a0perhaps?)<\/p><p>You\u00a0asked:<\/p><blockquote>\n  <p>Right now we\u2019re giving a 14 day free trial period to any lead but we still ask for the customer credit card to start the free trial. Should we remove that requirement? I\u2019m afraid that by doing that we might get more trials but have them be less\u00a0qualified.<\/p>\n<\/blockquote><p>You are almost certainly correct: most B2B SaaS companies find that removing the credit card requirement increases the number of free trial signups they get but decreases the activation rate (the number of users who make material use of the software) and conversion rate to paying use. It is generally not worth it early in the lifecycle of your company. (Your business needs to be capable operationally of doing fairly sophisticated work to nurture customers to activation and then convince them to part with their credit card details later in the trials for it to win net\u00a0here.)<\/p><blockquote>\n  <p>Our plans have a lot of features, should we put them all on the pricing columns or only the ones that\u00a0matter?<\/p>\n<\/blockquote><p>You should only include the most salient details on your pricing page. If there is a difference between plans which is not salient to your customers or to you, it should not be a difference between plans; it creates engineering work for you and frustration for your customers without doing the one thing that feature differentials is supposed to do, which is convince people to pay more for the\u00a0software.<\/p><blockquote>\n  <p>Do we need to offer different features on each plan or only different amounts of the maximum usage? i.e: Right now we charge a monthly or annual fee based on the number of users and publications each account\u00a0has.<\/p>\n<\/blockquote><p>Fundamentally, the purpose of charging different amounts for software is that different users get different amounts of value from the software, and that you want to capture some portion of the differential value created between your least successful user and your most successful\u00a0user.<\/p><p>All decisions about the pricing page are optimizations to approximate the value creation curve and charge for\u00a0it.<\/p><p>The number of users \/ number of publications is an approximation for value created. Use of more sophisticated features might be another way to approximate. Is there another approximation\u00a0available?<\/p><p>I would suggest that the best approximation for value created by a publishing platform is probably the amount of money it enables your customers to earn, and therefore I would probably charge based on count or total sum of transactions facilitated. George RR Martin should pay more than a hobbyist recipe book author, even if they have the same number of\u00a0titles.<\/p><p>(Relatedly: I think your tiers are far, far too generous, if I understand them\u00a0correctly.)<\/p><blockquote>\n  <p>We have different customers with different needs. Is it OK to have different pricing models for each? For example, for publishers we charge based on how many publications and users they have. Universities have many, many more publications by nature of how they operate; we only want to charge them based on the number of\u00a0users.<\/p>\n<\/blockquote><p>Absolutely yes! This is not done much by smaller SaaS businesses, because of the operational complexity and the degree to which it requires you to understand and operate in different market segments, but it is entirely possible for you to have \u201cspecial academic pricing\u201d, \u201cthe University edition\u201d,\u00a0etc.<\/p><p>You do not need a differentiated feature set to justify this; it can be entirely a marketing decision. The fundamental insight is that universities have incentives very different than for-profit publishers and that, since their incentives are different, treating them as if they were equivalent is not wonderful for either group. They also likely respond to different marketing messages, will have different feature needs in the future, and have a different value proposition in mind for working with you. Packaging the software differently to them helps you to discover and address these differences in the\u00a0market.<\/p><h2 id=\"firmb-case-study\">FirmB case study<\/h2><p>FirmB, whose name we have obscured at their request, sells a Shopify plug-in which assists e-commerce businesses in optimizing their conversion\u00a0rates.<\/p><p>Their original\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/firm-b.png\" alt=\"Firm B pricing\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><p>I\u2019d also name the plans a bit differently: Small Store vs. Sophisticated E-Commerce Retailer vs. Enterprise, maybe. Who can resist saying that they\u2019re the sophisticated\u00a0one?<\/p><h3 id=\"anchor-your-prices-to-the-cost-of-experts\">Anchor your prices to the cost of experts<\/h3><p>Businesses buy benefits, not features, so \u201cUnlock advanced features\u201d is unlikely to convince someone to upgrade from Essential to\u00a0Business.<\/p><p>I might position that as: <strong>Like having your own automated CRO agency, at a fraction of the\u00a0cost.<\/strong><\/p><p>Do you know what e-commerce companies pay for conversion rate optimization specialists? I have some knowledge of what the typical rate cards are like in that industry. When I did CRO for software companies I charged $30k a week. That is probably above the price point for most firms which are hosted on Shopify but they certainly aren\u2019t paying only hundreds of dollars per engagement. Accordingly, I would walk up the price of the Business offering and explicitly compare it in your sales material to hiring a CRO agency (minimum buy-in: $5k per month) or adding a CRO specialist to the team ($100k or more a year even for relatively unproven junior employees, when you evaluate the total cost of\u00a0employment).<\/p><p>Incidentally, new business owners don\u2019t model the true cost of employees\u2019 time correctly. Assume the total cost of a white collar employee is about 140% of their quoted salary; that\u2019s a good enough approximation to price SaaS against their true fully-loaded cost rather than their salary. Businesses have to budget for fully-loaded costs, inclusive of taxes, benefits, required contributions to social insurance schemes, and\u00a0overhead.<\/p><p>It is to your advantage, when anchoring against the price of employees, to anchor off their <em>true<\/em> fully-loaded costs rather than their <em>imaginary<\/em> nominal costs (the salary\u00a0number).<\/p><h3 id=\"pricing-bundles-effectively\">Pricing bundles effectively<\/h3><p>I think that you effectively have an offering which bundles multiple discrete offerings which are popular on Shopify, correct? I would consider pricing your Essential plan up to $49 and then offering a less prominent (not a column, just a textual option below the main pricing grid) Getting Started plan at $29, which would allow only one widget (at any time). This lets price-conscious firms dip their toes into the water without cutting off the revenue from the substantial fraction of e-commerce shops which are actually legitimate businesses which pay for healthcare and everything. The difference between $29 and $49 is immaterial to them but certainly isn\u2019t to you, once you\u2019ve signed hundreds of\u00a0them.<\/p><p>I\u2019d price $199 to $249 immediately (since they\u2019re perceived as the same amount by users); I\u2019d probably price it higher than that, too, after you have dialed in the value provided via, for example, concierge\u00a0onboarding.<\/p><h3 id=\"concierge-onboarding\">Concierge onboarding<\/h3><p>Want a great way to increase your conversion rate to mid-tier plans like the one you described as Business? Offer free consulting services which are bundled in with the offering and focused on getting someone onboarded successfully; this is often called \u201cconcierge onboarding.\u201d If you\u2019re worried about the amount of time required, you can restrict them to companies which elect the annual plan. (You should have an annual\u00a0plan.)<\/p><p>Concierge onboarding makes your users more likely to successfully convert and more successful (so they are less likely to churn). In addition, early in the business it adds wonderful opportunities to talk with your customers and make sure you\u2019re building the right features, training material, and workflows to meet their\u00a0needs.<\/p><p>It also gives people who are less price sensitive a reason to hang their hat on your higher-tier offerings. What if I am making $500k a month of high-margin sales but simply am not sophisticated enough to know what multivariate testing is? Concierge onboarding helps me get into a detailed conversation with you, where you provide outstanding value by getting the plug-in set up to start increasing my sales automatically, and the opportunity for you to sell me on the value of multivariate testing.\u00a0Win\/win!<\/p><p>Sophisticated businesses have, generally, greater ability to pay for SaaS because the right SaaS creates more value in their businesses. This is orthogonal to its complexity, degree of \u201csecret sauce\u201d, or what it actually\u00a0does.<\/p><p>Accordingly, SaaS companies which sell directly into sophisticated businesses by construction should price at a premium to companies which are selling to the broader B2B market. Companies which sell across a variety of verticals should, where possible, segment out sophisticated businesses and package the offering so it delivers them premium services at the (substantial) price points they are used to paying for valuable\u00a0services.<\/p><h2 id=\"geomodelr-case-study\">Geomodelr case study<\/h2><p>Geomodelr sells into the oil &amp; gas industry. Their existing\u00a0pricing:<\/p><p><img src=\"https:\/\/stripe.com\/img\/atlas\/guides\/saas-pricing\/geomodelr.png\" alt=\"Geomodelr pricing grid; Monthly Professional $199; Yearly Professional $149.92 per month\" class=\"expandable\" \/><\/p><p>Our\u00a0commentary:<\/p><h3 id=\"free-is-not-a-compelling-value-proposition-to-well-monied-buyers\">Free is not a compelling value proposition to well-monied buyers<\/h3><p>Historically, companies have offered software for free or nearly free to benefit from the software spreading as wide as possible. This is of sharply limited utility in selling B2B services, because price is often a proxy for value, particularly among savvy customers who are in industries which are (to pardon a pun) gushing with\u00a0money.<\/p><p>Does \u201cfree\u201d communicate to an oil company that the data \/ analysis \/ etc produced by your report is reliable enough to place a $X00k capital investment or a $Y million mineral rights lease? Does free comport with the values of decisionmakers in mining companies? Do they expect to get accounting or legal or insurance\u00a0free?<\/p><p>I would discontinue the free version immediately. Send an email to existing free users saying that you\u2019ve discontinued the offering and that you will honor their pricing for forever and, as a special reward for their loyalty in the early days of your company, you will give them a 20% discount to the paid\u00a0version.<\/p><p>Going forward, while you can make tactical use of a limited free trial, you should probably not provide commercializable services for free\u00a0again.<\/p><h3 id=\"how-to-present-yearly-pricing\">How to present yearly pricing<\/h3><p>In the traditional SaaS pricing grid, there is a huge distinction between columns. Choosing between columns requires a Consequential Decision by your customer. You should minimize consequential\u00a0decisions.<\/p><p>Monthly versus annual billing shouldn\u2019t be a consequential\u00a0decision.<\/p><p>The majority of your customers should not have strong cash-flow considerations dominating their software budget, because they\u2019re highly-compensated professionals in a capital intensive industry which is happy to pay for engineering services. I would have the core offering be the yearly plan and have a toggle box to select\u00a0monthly.<\/p><h3 id=\"have-multiple-packages\">Have multiple packages<\/h3><p>We\u2019ve gotten rid of your free plan and consolidated the decision about billing cycle, so your company has essentially one package which it is selling now. There should probably be more than one packages available, to offer an appropriate incentive for companies which receive a lot of value to tell you \u201cActually, I receive a heck of a lot of value from this\u201d and, simultaneously, convince them to pay more than companies which are more marginal to\u00a0you.<\/p><p>I\u2019d probably adjust your price points to $2.5k annual \/ $250 monthly for the core package (the first of many moves upward) and then test another price point at $10k annual (with either no monthly option or a $1k per month but billed quarterly option, to minimize how many checks you have to\u00a0chase).<\/p><p>What goes in the more expensive offering? I don\u2019t know. I\u2019d start by putting priority support in that and then see what customers asked\u00a0for.<\/p><h3 id=\"consider-moving-up-market-to-the-enterprise\">Consider moving up-market to the enterprise<\/h3><p>I\u2019d also probably add an Enterprise offering where people have to call you for pricing. On your first dozen calls, you should just try to understand what they need out of the Enterprise offering and whether you can provide it. (If you are not comfortable doing that, call your existing paying customers, ask if they pay more than $50k a year for any software, and if so ask what they would need from you to justify paying more than that ballpark for\u00a0you.)<\/p><p>If you <a href=\"https:\/\/stripe.com\/atlas\">incorporate with Stripe Atlas<\/a>, you\u2019ll receive opportunities to get feedback on parts of your businesses from experienced entrepreneurs, investors, and subject-matter experts, similar to the Q&amp;A  which this article is excerpted from or our <a href=\"https:\/\/stripe.com\/atlas\/guides\/ama-marc-andreessen\">Q&amp;A with Marc\u00a0Andreessen<\/a>.<\/p><p>Whether you\u2019re running a company or just thinking about it, we\u2019d love to hear of what challenges we could help you solve. Drop us a line at atlas@stripe.com if there is a subject you\u2019d like to see us cover in more\u00a0detail.<\/p><aside class=\"collect-email-guide\">\n  <\/aside><a href=\"https:\/\/stripe.com\/atlas\/guides\" class=\"common-Link common-Link--arrowL\">Return to guides<\/a>","created_at":"2018-11-07T16:43:30+0000","updated_at":"2018-11-07T16:43:30+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":36,"domain_name":"stripe.com","preview_picture":"https:\/\/stripe.com\/img\/v3\/atlas\/guide\/social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":295,"title":"Free games of the week","url":"https:\/\/www.rockpapershotgun.com\/2018\/02\/24\/free-games-of-the-week-4\/","content":"<p>Ever wonder what happens out in space? Discovering lost recordings left inside alien structures by a previous astronaut might peak your curiosity. This astronaut wanted their final words and thoughts to be documented, but they don\u2019t seem to have made it back to Earth. While you float across space, listening to the different entries, try to discover what truly happened in this part of the void. If space is not your thing, perhaps trying to communicate as a raven in a world full of humans, or reliving memories around your favorite convenience store, is what you want to do this weekend. You could also raise a childhood hero, or even just eat some meatballs! It\u2019s all up to you \u2013 I just provide the games. <em>Need more no-money gaming? Check our list of the <a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/19\/best-free-pc-games\/\">best free games<\/a> you can download and play on PC right now.<\/em><\/p><p><strong><a href=\"https:\/\/xiotex.itch.io\/the-alpha-device\">The Alpha Device<\/a> by Xiotex Studios Ltd<\/strong><\/p>\n<p>Explore outer space, floating through vast, decaying structures seemingly constructed by extraterrestrials. These structures are everywhere \u2013 sometimes with gaps between them and other times tightly packed together. You are able to float in-between them, looking around for green dots and exploring the areas you find interesting. These dots are transmitting bits of story, told from an astronaut\u2019s perspective \u2013 a beautiful narrative voiced by David Hewlett for your ears to listen to as you float around.<\/p>\n<p>You do not change or affect the story in any way \u2013 there are no decisions to make. Instead you can enjoy the mystery around what has happened in the void of space. This space structure is lovely to explore as is, providing lots of little spaces to glide into and lots of openings to explore. The atmosphere and creepy overtones of listening to the various messages that evidently had to do with the reason why you are in space, now adds to the mystery and curiosity as you play. It is worth noting that The Alpha Device is only playable with a controller \u2013 so you will need to have one plugged in before going out into deep space.<\/p>\n<p><strong><a href=\"http:\/\/store.steampowered.com\/app\/744810\/A_Raven_Monologue\/\">A Raven Monologue<\/a> by Mojiken Studio<\/strong><\/p>\n<p>Instead of being unable to communicate with Earth, this Raven is unable to communicate in a world full of humans. He is silent, unable to ever cry out, but still attempts to interact with humans in his town without ever saying a word. A Raven Monologue is a short, experimental and silent story \u2013 a game that doesn\u2019t quite fit into any genre. You are able to move the story forward, seeing what each new page brings, or push it backwards and see what happens if you retrace your steps. The raven you are controlling is collecting different objects to build a beautiful cage sculpture.<\/p>\n<p>As you collect these different objects, you get to explore the entire town around you, interact with various townspeople, and listen to beautiful music. Once you have collected everything, you can head back, and see how the town and people have changed \u2013 living the silent life of this raven. Even though this game is a rather quick play, it is something extremely different and very worth taking a look at.<\/p>\n<p><strong><a href=\"https:\/\/kano-stuff.itch.io\/konbini\">Konbini<\/a> by Kano \/ Robin Beaurepaire \/ Andr\u00e9a Pedrono<\/strong><\/p>\n<p>Like the previous game, but with more words and minigames, Konbini takes you through the memories of someone who lived in Japan and frequented a convenience store. This store is average \u2013 typical even. It is by a bus stop, which is convenient if you get on and off the bus often. This shop isn\u2019t close to where you live, nor is it generally en route to wherever you are going. It just exists and sometimes crosses your path. You are able to click on various parts of the store \u2013 the vending machine, bus stop, and arcade machine inside of it, for example.<\/p>\n<p>These will show you memories and allow you to interact further with whatever is going on. You can put an obscene amount of coins in the vending machine or play through the arcade game while trying to read the story with it. These mini-games and tasks accompanying the story, add a lot to the game as a whole. These stories are very relatable \u2013 taking you to a memory you may even find relatable yourself\u2026<\/p>\n<p><strong><a href=\"https:\/\/gamejolt.com\/games\/dungeon-dad\/317041\">Dungeon Dad<\/a> by Flamebait<\/strong><\/p>\n<p>This action based game has you creating memories for your child \u2013 by sending them into the dungeon, coaching them on how to fight, and gambling away the money you make from them! You always wanted to be a hero, however, you didn\u2019t make it there. Now, you will live your dream through your very own child and become an ultimate hero and adventurer. You are able to send them into the dungeon and school them in defence and attack.<\/p>\n<p>They will defend and attack on their own, but your instruction will certainly help them along \u2013 and the only cost is their self-esteem. Once they want to come back from the dungeon, they will let you know! You can force them to stay longer, or allow them to come back before seeing the loot they collected. This loot can be spent in the shop on armor, swords, and ways to help the kid out. Or\u2026 it could be gambled into potentially even more cash! It is all up to you \u2013 you can be the parent your child needs or make your child into a hardened killer.<\/p>\n<p><strong><a href=\"https:\/\/aaronfimple.itch.io\/borgo-the-cat\">Borgo the Cat<\/a> by Aaron Fimple<\/strong><\/p>\n<p>Ever wanted to just eat meatballs? How about discovering and eating meatballs as a giant round cat \u2013 rolling over to your spherical friend to then chow down? Though this might not be a memory you had in your own head, it is surely something worth experiencing! In Borgo the Cat, you play a silly, round cat \u2013 solving physics based puzzles in this platformer. You can roll and jump your way through different levels, all in the name of spherical meat! You can move the camera around the level, viewing some hidden areas and items sometimes, before making your way to the goal.<\/p>\n<p>This meatball is very tempting, but sometimes you will need to push buttons, move boxes, or even break walls to get to your food. As Borgo, you are a very adorable little cat, well worth the cost of a single bit of food. Making your way through each level does take some time, but the game itself can be completed quite quickly \u2013 leaving you a bit hungry for more!<\/p>","created_at":"2018-11-07T16:43:30+0000","updated_at":"2018-11-07T16:43:30+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":5,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/assets.rockpapershotgun.com\/images\/\/2018\/02\/freegamesheader-620x316.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":293,"title":"Calibre: PDFs to ePub Conversion Tips","url":"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/","content":"<div class=\"da_blog_top_wrap\"><p><time>August 23, 2015<\/time><\/p><\/div><p class=\"post-meta post-meta-details\"><img alt=\"\" src=\"https:\/\/secure.gravatar.com\/avatar\/38b4d2b7351860a0e919f951674871ee?s=30&amp;d=https%3A%2F%2Fdearauthor.com%2Fwp-content%2Fthemes%2Fda-child%2Fimg%2Fgravatar_newest.png&amp;r=pg\" srcset=\"https:\/\/secure.gravatar.com\/avatar\/38b4d2b7351860a0e919f951674871ee?s=60&amp;d=https%3A%2F%2Fdearauthor.com%2Fwp-content%2Fthemes%2Fda-child%2Fimg%2Fgravatar_newest.png&amp;r=pg 2x\" class=\"avatar avatar-30 photo\" height=\"30\" width=\"30\" \/><a href=\"https:\/\/dearauthor.com\/category\/ebooks\/\" rel=\"category tag\">Ebooks<\/a><a href=\"https:\/\/dearauthor.com\/tag\/calibre\/\" rel=\"tag\">Calibre<\/a><a href=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/#comments\">28 Comments<\/a><\/p><p>A couple of weeks ago, after this post about inserting the blurb at the front of an ebook, a reader emailed me asking about cleaning up PDF files when converting to ePub. Often there will be stray letters or numbers or headers that affect the output of an ebook. There are some tips and tricks in Calibre\u2019s PDF conversion engine that can be used to produce very clean and readable PDFs. I\u2019m going to address three of the most common problems when converting a PDF to ePub and what you can do to address those problems.<\/p><p><strong>1) Line numbers.<\/strong><\/p><p><img data-attachment-id=\"75571\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-20-50-02\/\" data-orig-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-20.50.02.png?fit=260%2C406&amp;ssl=1\" data-orig-size=\"260,406\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 20.50.02\" data-image-description=\"\" data-medium-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-20.50.02.png?fit=260%2C406&amp;ssl=1\" data-large-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-20.50.02.png?fit=260%2C406&amp;ssl=1\" class=\"aligncenter  wp-image-75571\" src=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-20.50.02.png?resize=189%2C295&amp;ssl=1\" alt=\"Screenshot 2015-08-22 20.50.02\" width=\"189\" height=\"295\" data-recalc-dims=\"1\" \/><\/p><p>Some PDFs have line numbers that are on a hidden layer. When you read the PDF you can\u2019t see them, but when you convert to ePub or Mobi they appear within the text and render the converted book\u00a0unreadable. There\u2019s a very easy fix to this.<\/p><p>Under \u201cConvert books\u201d select \u201cSearch and Replace\u201d:<\/p><p><img data-attachment-id=\"75572\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/search-replace\/\" data-orig-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?fit=1596%2C1058&amp;ssl=1\" data-orig-size=\"1596,1058\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Search Replace\" data-image-description=\"\" data-medium-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?fit=300%2C199&amp;ssl=1\" data-large-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?fit=765%2C507&amp;ssl=1\" class=\"aligncenter  wp-image-75572\" src=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?resize=551%2C365&amp;ssl=1\" alt=\"Search Replace\" width=\"551\" height=\"365\" srcset=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?resize=765%2C507&amp;ssl=1 765w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?resize=300%2C199&amp;ssl=1 300w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?resize=768%2C509&amp;ssl=1 768w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Search-Replace.png?w=1596&amp;ssl=1 1596w\" data-recalc-dims=\"1\" \/><\/p><p>Make sure you select PDF as your input and then either Mobi or ePub (or whatever format you prefer) as output. As the screenshot says, search and replace uses regular expressions. You can read <a href=\"http:\/\/www.mobileread.com\/forums\/showthread.php?t=99258\" target=\"_blank\">more about it here<\/a>. Don\u2019t be afraid. You can do this!<\/p><p><img data-attachment-id=\"75573\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/wizard\/\" data-orig-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?fit=1562%2C260&amp;ssl=1\" data-orig-size=\"1562,260\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"wizard\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?fit=300%2C50&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?fit=765%2C127&amp;ssl=1\" class=\"aligncenter  wp-image-75573\" src=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?resize=458%2C76&amp;ssl=1\" alt=\"wizard\" width=\"458\" height=\"76\" srcset=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?resize=765%2C127&amp;ssl=1 765w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?resize=300%2C50&amp;ssl=1 300w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?resize=768%2C128&amp;ssl=1 768w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/wizard.png?w=1562&amp;ssl=1 1562w\" data-recalc-dims=\"1\" \/><\/p><p>The first thing you want to do is click on the wizard button. This will bring up a dialogue box that shows what the PDF looks like before conversion. There\u2019s a number followed by &lt;br&gt;, then a line break. You don\u2019t want any of that.<\/p><p><img data-attachment-id=\"75574\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-10-45\/\" data-orig-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.10.45.png?fit=512%2C1186&amp;ssl=1\" data-orig-size=\"512,1186\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.10.45\" data-image-description=\"\" data-medium-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.10.45.png?fit=300%2C695&amp;ssl=1\" data-large-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.10.45.png?fit=512%2C1186&amp;ssl=1\" class=\"aligncenter  wp-image-75574\" src=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.10.45.png?resize=337%2C781&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.10.45\" width=\"337\" height=\"781\" srcset=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.10.45.png?w=512&amp;ssl=1 512w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.10.45.png?resize=300%2C695&amp;ssl=1 300w\" data-recalc-dims=\"1\" \/><\/p><p>To remove the numbers and the &lt;br&gt; which is CSS for a line break, your code would be:<\/p><blockquote><p>\\d+&lt;br&gt;\\n<\/p><\/blockquote><ul><li>The \\d+ tells the program that you want to remove every digit whether it is 01 or\u00a0501. The letter \u201cd\u201d in regexpression stands for\u00a0single character that is a digit. The \u201c+\u201d is a greedy qualifier which means it will remove as many digits as possible. In a fiction\u00a0book, this is a good thing.<\/li>\n<li>&lt;br&gt; is the CSS code<\/li>\n<li>\\n is the line break or end of line.<\/li>\n<\/ul><p>You can click \u201cTest\u201d to determine whether your expression is going to strip out the right text. \u00a0\u00a0The yellow highlighted text will be removed:<\/p><p><img data-attachment-id=\"75575\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-15-20\/\" data-orig-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.15.20.png?fit=216%2C340&amp;ssl=1\" data-orig-size=\"216,340\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.15.20\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.15.20.png?fit=216%2C340&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.15.20.png?fit=216%2C340&amp;ssl=1\" class=\"aligncenter  wp-image-75575\" src=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.15.20.png?resize=119%2C187&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.15.20\" width=\"119\" height=\"187\" data-recalc-dims=\"1\" \/><\/p><p>As you scroll down, sometimes you will see these line numbers that start with S or N.<\/p><p><img data-attachment-id=\"75576\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-20-57\/\" data-orig-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?fit=862%2C112&amp;ssl=1\" data-orig-size=\"862,112\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.20.57\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?fit=300%2C39&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?fit=765%2C99&amp;ssl=1\" class=\"aligncenter size-large wp-image-75576\" src=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?resize=765%2C99&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.20.57\" width=\"765\" height=\"99\" srcset=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?resize=765%2C99&amp;ssl=1 765w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?resize=300%2C39&amp;ssl=1 300w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?resize=768%2C100&amp;ssl=1 768w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.20.57.png?w=862&amp;ssl=1 862w\" data-recalc-dims=\"1\" \/><\/p><p>or sometimes the line numbers will end with S and N.<\/p><p><img data-attachment-id=\"75577\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-22-24\/\" data-orig-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.22.24.png?fit=744%2C116&amp;ssl=1\" data-orig-size=\"744,116\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.22.24\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.22.24.png?fit=300%2C47&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.22.24.png?fit=744%2C116&amp;ssl=1\" class=\"aligncenter size-full wp-image-75577\" src=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.22.24.png?resize=744%2C116&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.22.24\" width=\"744\" height=\"116\" srcset=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.22.24.png?w=744&amp;ssl=1 744w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.22.24.png?resize=300%2C47&amp;ssl=1 300w\" data-recalc-dims=\"1\" \/><\/p><p>To construct the search text, you would simply add the S or N before or after the number.<\/p><blockquote><p>S\\d+&lt;br&gt;\\n<\/p>\n<p>N\\d+&lt;br&gt;\\n<\/p>\n<p>\\d+S&lt;br&gt;\\n<\/p>\n<p>\\d+N&lt;br&gt;\\n<\/p><\/blockquote><p>Or, if you\u2019re really savvy, you\u2019ll use the\u00a0<strong>\\w<\/strong> code. \\w removes a single word character.<\/p><blockquote><p><strong>\\d+&lt;br&gt;\\n<\/strong>|<strong>\\d+\\w&lt;br&gt;\\n<\/strong>|<strong>\\w\\d+&lt;br&gt;\\n<\/strong><\/p><\/blockquote><p>The\u00a0<strong>|<\/strong> is called a pipe and it is used to separate sets of regular expressions. I\u2019m sure there is a more sophisticated query but it works for me.<\/p><p><img data-attachment-id=\"75580\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-25-32\/\" data-orig-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.32.png?fit=170%2C118&amp;ssl=1\" data-orig-size=\"170,118\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.25.32\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.32.png?fit=170%2C118&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.32.png?fit=170%2C118&amp;ssl=1\" class=\"aligncenter size-full wp-image-75580\" src=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.32.png?resize=170%2C118&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.25.32\" width=\"170\" height=\"118\" data-recalc-dims=\"1\" \/><img data-attachment-id=\"75581\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-25-16\/\" data-orig-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.16.png?fit=204%2C114&amp;ssl=1\" data-orig-size=\"204,114\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.25.16\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.16.png?fit=204%2C114&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.16.png?fit=204%2C114&amp;ssl=1\" class=\"aligncenter size-large wp-image-75581\" src=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.25.16.png?resize=204%2C114&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.25.16\" width=\"204\" height=\"114\" data-recalc-dims=\"1\" \/><\/p><p>Now when you scroll down, those pesky stray line numbers should be highlighted. Press okay and you\u2019ll be sent back to the Search and Replace window.\u00a0The replace term\u00a0is left completely blank. Press \u201cAdd\u201d or your conversion won\u2019t include the search and replace you just constructed.<\/p><p><img data-attachment-id=\"75582\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-30-57\/\" data-orig-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?fit=1354%2C198&amp;ssl=1\" data-orig-size=\"1354,198\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.30.57\" data-image-description=\"\" data-medium-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?fit=300%2C44&amp;ssl=1\" data-large-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?fit=765%2C112&amp;ssl=1\" class=\"aligncenter size-large wp-image-75582\" src=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?resize=765%2C112&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.30.57\" width=\"765\" height=\"112\" srcset=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?resize=765%2C112&amp;ssl=1 765w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?resize=300%2C44&amp;ssl=1 300w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?resize=768%2C112&amp;ssl=1 768w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.30.57.png?w=1354&amp;ssl=1 1354w\" data-recalc-dims=\"1\" \/><\/p><p>Your search expression appears in the left with the replacement text (blank in this case) on the left.<\/p><p>If that\u2019s all you need to exclude, then press \u201cOK\u201d and the conversion will begin. Often, however, there will be headers that need to be removed.<\/p><p><strong>2. \u00a0Headers.<\/strong><\/p><p>In the following\u00a0book\u00a0there was an\u00a0alternating header with author name and title. \u00a0\u00a0Remember, the number and text will often change from page to page.<\/p><blockquote><p>&lt;a name=\u201d6\u2033&gt;&lt;\/a&gt;&lt;em&gt;Anne Calhoun &lt;\/em&gt;<\/p>\n<p>&lt;a name=\u201d7\u2033&gt;&lt;\/a&gt;&lt;em&gt;Liberating Lacey &lt;\/em&gt;<\/p><\/blockquote><p>and a footer with the page numbers:<\/p><blockquote><p>6 &lt;br&gt;<\/p>\n<p>&lt;hr \/&gt;<\/p><\/blockquote><p>Basically I copy any text that needs to be removed and then replace the number with <strong>\\d+<\/strong>. My reg expression is as follows.<\/p><blockquote><p>&lt;a name=\u201d\\d+\u201d&gt;&lt;\/a&gt;)(&lt;i&gt;Anne Calhoun &lt;\/i&gt;&lt;br&gt;<\/p><\/blockquote><p><a href=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?ssl=1\"><img data-attachment-id=\"23148\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/dear-jane-ebooks\/dear-jane-what-is-the-best-way-to-read-pdfs-on-a-6-inch-screen\/attachment\/screen-shot-2010-10-02-at-10-42-11-pm\/\" data-orig-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?fit=469%2C184&amp;ssl=1\" data-orig-size=\"469,184\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screen shot 2010-10-02 at 10.42.11 PM\" data-image-description=\"\" data-medium-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?fit=300%2C117&amp;ssl=1\" data-large-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?fit=469%2C184&amp;ssl=1\" class=\"aligncenter size-full wp-image-23148\" title=\"Screen shot 2010-10-02 at 10.42.11 PM\" src=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?resize=469%2C184&amp;ssl=1\" alt=\"\" width=\"469\" height=\"184\" srcset=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?w=469&amp;ssl=1 469w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.42.11-PM.png?resize=300%2C117&amp;ssl=1 300w\" data-recalc-dims=\"1\" \/><\/a><\/p><p>Highlight the text to be removed and then replace the numbers with <strong>\\d+<\/strong>.\u00a0You could add a \\s+ between \u201cLacey\u201d or \u201cCalhoun\u201d and the &lt;\/i&gt; just to be on the safe side:\u00a0\u00a0<strong>&lt;i&gt;Liberating Lacey\\s*&lt;\/i&gt;&lt;br&gt;<\/strong><\/p><p><strong><a href=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM.png?ssl=1\"><img data-attachment-id=\"23149\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/dear-jane-ebooks\/dear-jane-what-is-the-best-way-to-read-pdfs-on-a-6-inch-screen\/attachment\/screen-shot-2010-10-02-at-10-44-16-pm\/\" data-orig-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM-e1286077688164.png?fit=485%2C183&amp;ssl=1\" data-orig-size=\"485,183\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;}\" data-image-title=\"Screen shot 2010-10-02 at 10.44.16 PM\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM-e1286077688164.png?fit=300%2C113&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM-e1286077688164.png?fit=485%2C183&amp;ssl=1\" class=\"aligncenter size-full wp-image-23149\" title=\"Screen shot 2010-10-02 at 10.44.16 PM\" src=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM-e1286077688164.png?resize=485%2C183&amp;ssl=1\" alt=\"\" width=\"485\" height=\"183\" srcset=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM-e1286077688164.png?w=485&amp;ssl=1 485w, https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2010\/10\/Screen-shot-2010-10-02-at-10.44.16-PM-e1286077688164.png?resize=300%2C113&amp;ssl=1 300w\" data-recalc-dims=\"1\" \/><\/a><br \/><\/strong><\/p><p>The \\s+ removes excess white spaces.<\/p><p><strong>3. \u00a0Line spaces<\/strong><\/p><p>Calibre has a feature called Heuristic Processing which scans the ebook and tries to search for common errors and fixes them. I use this function primarily to unwrap lines. In this example, you can see the paragraph is broken up by these weird line wraps and reading a book in this format would<\/p><p>be<\/p><p>impossi<\/p><p>ble.<\/p><p><img data-attachment-id=\"75583\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-39-51\/\" data-orig-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?fit=946%2C642&amp;ssl=1\" data-orig-size=\"946,642\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.39.51\" data-image-description=\"\" data-medium-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?fit=300%2C204&amp;ssl=1\" data-large-file=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?fit=765%2C519&amp;ssl=1\" class=\"aligncenter  wp-image-75583\" src=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?resize=323%2C219&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.39.51\" width=\"323\" height=\"219\" srcset=\"https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?resize=765%2C519&amp;ssl=1 765w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?resize=300%2C204&amp;ssl=1 300w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?resize=768%2C521&amp;ssl=1 768w, https:\/\/i2.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.39.51.png?w=946&amp;ssl=1 946w\" data-recalc-dims=\"1\" \/><\/p><p>The default for Heuristic Processing is that it is disabled. So check the box and the options will become available. The default for the line unwrap is .40 and frankly that setting usually never works for me.<\/p><p><img data-attachment-id=\"75584\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-42-06\/\" data-orig-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?fit=1218%2C800&amp;ssl=1\" data-orig-size=\"1218,800\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.42.06\" data-image-description=\"\" data-medium-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?fit=300%2C197&amp;ssl=1\" data-large-file=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?fit=765%2C502&amp;ssl=1\" class=\"aligncenter  wp-image-75584\" src=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?resize=488%2C320&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.42.06\" width=\"488\" height=\"320\" srcset=\"https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?resize=765%2C502&amp;ssl=1 765w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?resize=300%2C197&amp;ssl=1 300w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?resize=768%2C504&amp;ssl=1 768w, https:\/\/i1.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06.png?w=1218&amp;ssl=1 1218w\" data-recalc-dims=\"1\" \/><\/p><p>You often will have to play around with this and reformat at different settings.\u00a0\u00a0In this example, I had to reduce the line unwrap factor to 0.5 to get the paragraphs to be readable.<\/p><p><img data-attachment-id=\"75585\" data-permalink=\"https:\/\/dearauthor.com\/ebooks\/calibre-pdfs-epub-conversion-tips\/attachment\/screenshot-2015-08-22-21-44-46\/\" data-orig-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?fit=1026%2C258&amp;ssl=1\" data-orig-size=\"1026,258\" data-comments-opened=\"0\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot 2015-08-22 21.44.46\" data-image-description=\"\" data-medium-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?fit=300%2C75&amp;ssl=1\" data-large-file=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?fit=765%2C192&amp;ssl=1\" class=\"aligncenter  wp-image-75585\" src=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?resize=594%2C149&amp;ssl=1\" alt=\"Screenshot 2015-08-22 21.44.46\" width=\"594\" height=\"149\" srcset=\"https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?resize=765%2C192&amp;ssl=1 765w, https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?resize=300%2C75&amp;ssl=1 300w, https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?resize=768%2C193&amp;ssl=1 768w, https:\/\/i0.wp.com\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.44.46.png?w=1026&amp;ssl=1 1026w\" data-recalc-dims=\"1\" \/><\/p><p>A soft scene break is when there is an extra space between different scenes in a book rather than the use of a wingding (???) or some small graphic such as hashmarks (###) or bullets (\u2022\u2022\u2022). You can replace the soft scene breaks with your own text or graphic to further customize the look and feel of your ebooks.<\/p><p>Hope this helps!<\/p>","created_at":"2018-11-07T16:43:26+0000","updated_at":"2018-11-07T16:43:26+0000","published_at":"2015-08-23T09:00:55+0000","published_by":["Jane"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"dearauthor.com","preview_picture":"https:\/\/dearauthor.com\/wp-content\/uploads\/2015\/08\/Screenshot-2015-08-22-21.42.06-765x502.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":292,"title":"Top 5 Apps for Personal Trainers and Health Coaches","url":"https:\/\/www.afpafitness.com\/blog\/top-5-apps-for-personal-trainers-and-health-coaches","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:43:24+0000","updated_at":"2018-11-07T16:43:24+0000","annotations":[],"reading_time":0,"domain_name":"www.afpafitness.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":289,"title":"zachcurry\/emacs-anywhere","url":"https:\/\/github.com\/zachcurry\/emacs-anywhere","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/07025e282fc67f96621ed0ae5a7376d157e1dacc\/68747470733a2f2f696d6775722e636f6d2f4b4576615654502e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/07025e282fc67f96621ed0ae5a7376d157e1dacc\/68747470733a2f2f696d6775722e636f6d2f4b4576615654502e6a7067\" width=\"250px\" data-canonical-src=\"https:\/\/imgur.com\/KEvaVTP.jpg\" alt=\"image\" \/><\/a>\n<\/p>\n<p style=\"text-align: center;\">\n  <em>Inspired by<\/em>\n  <a href=\"https:\/\/github.com\/cknadler\/vim-anywhere\">cknadler\/vim-anywhere<\/a>\n<\/p>\n<p style=\"text-align: center;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/01974a24723270cd2e22ea37c4a7d36e605baa0e\/68747470733a2f2f7468756d62732e6766796361742e636f6d2f48617264746f66696e6442656c6174656441746c616e74696373686172706e6f73657075666665722d73697a655f726573747269637465642e676966\"><img src=\"https:\/\/camo.githubusercontent.com\/01974a24723270cd2e22ea37c4a7d36e605baa0e\/68747470733a2f2f7468756d62732e6766796361742e636f6d2f48617264746f66696e6442656c6174656441746c616e74696373686172706e6f73657075666665722d73697a655f726573747269637465642e676966\" width=\"500px\" data-canonical-src=\"https:\/\/thumbs.gfycat.com\/HardtofindBelatedAtlanticsharpnosepuffer-size_restricted.gif\" alt=\"image\" \/><\/a>\n  \n<\/p>\n<p style=\"text-align: center;\">\n  <strong>Emacs Anywhere<\/strong> provides configurable automation and hooks containing window info, so you can bust moves anywhere in a quick, customizable fashion.\n<\/p>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#install\">Install<\/a>\n<ul><li><a href=\"#os-x\">OS X<\/a><\/li>\n<li><a href=\"#install\">Linux<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#usage\">Usage<\/a>\n<ul><li><a href=\"#try-it\">Try It<\/a><\/li>\n<li><a href=\"#environment\">Environment<\/a><\/li>\n<li><a href=\"#commands\">Commands<\/a><\/li>\n<li><a href=\"#variables\">Variables<\/a><\/li>\n<li><a href=\"#hooks\">Hooks<\/a><\/li>\n<li><a href=\"#examples\">Examples<\/a><\/li>\n<li><a href=\"#update\">Update<\/a><\/li>\n<li><a href=\"#uninstall\">Uninstall<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#todo\">Todo<\/a><\/li>\n<li><a href=\"#contributing\">Contributing<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><h2><a id=\"user-content-os-x\" class=\"anchor\" aria-hidden=\"true\" href=\"#os-x\"><\/a>OS X<\/h2>\n<pre>curl -fsSL https:\/\/raw.github.com\/zachcurry\/emacs-anywhere\/master\/install | bash\n<\/pre>\n<p>Open <strong>System Preferences<\/strong> and navigate to <strong>keyboard &gt; shortcuts &gt; Services<\/strong>.\nCheck the box beside \"Emacs Anywhere\", click \"Add Shortcut\" and key a shortcut.<\/p>\n<h2><a id=\"user-content-linux\" class=\"anchor\" aria-hidden=\"true\" href=\"#linux\"><\/a>Linux<\/h2>\n<blockquote>\n<p>In order to use <strong>Emacs Anywhere<\/strong>, you must use <strong>Xorg<\/strong> as your window manager. You can <strong>switch<\/strong> your window manager in <strong>Ubuntu<\/strong> by going to the login screen, clicking the cog icon, and selecting <strong>Xorg<\/strong>.<\/p>\n<\/blockquote>\n<h3><a id=\"user-content-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>dependencies<\/h3>\n<p>Install these dependencies as needed<\/p>\n<ul><li><strong>xclip<\/strong><\/li>\n<li><strong>xdotool<\/strong><\/li>\n<\/ul><pre>sudo apt-get install &lt;dependency&gt;\n<\/pre>\n<h3><a id=\"user-content-install-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#install-1\"><\/a>Install<\/h3>\n<pre>curl -fsSL https:\/\/raw.github.com\/zachcurry\/emacs-anywhere\/master\/install | bash\n<\/pre>\n<p>Create a keyboard shortcut to run <code>~\/.emacs_anywhere\/bin\/run<\/code>.<\/p>\n<h2><a id=\"user-content-windows\" class=\"anchor\" aria-hidden=\"true\" href=\"#windows\"><\/a>Windows<\/h2>\n<p>TBD<\/p>\n<h2><a id=\"user-content-try-it\" class=\"anchor\" aria-hidden=\"true\" href=\"#try-it\"><\/a>Try It<\/h2>\n<ol><li>Make sure your Emacs server is running. You can start the server like this<\/li>\n<\/ol><pre>emacs --daemon\n<\/pre>\n<ol start=\"2\"><li>Focus an application window's text input. A text area in your browser will do.<\/li>\n<li>Key the keyboard shortcut you've chosen for <strong>Emacs Anywhere<\/strong><\/li>\n<li>Write some text<\/li>\n<li>Delete the frame (<code>C-x 5 0<\/code>, or <code>:q<\/code> if you use evil-mode).\n<strong>Emacs Anywhere<\/strong> will copy the text from the buffer \"*Emacs Anywhere*\" to clipboard, delete the\nbuffer, switch focus to the window from step two and paste the buffers content into the text input.<\/li>\n<\/ol><h2><a id=\"user-content-environment\" class=\"anchor\" aria-hidden=\"true\" href=\"#environment\"><\/a>Environment<\/h2>\n<p>The <code>EA_EDITOR<\/code> environment variable can be used to override the way\n<strong>emacsclient<\/strong> is run. For example, you could put the following statement into\nyour <code>.bash_profile<\/code>.<\/p>\n<pre>export EA_EDITOR='emacsclient -a \"\" -c'\n<\/pre>\n<p>With the environment variable set, the invocation will look like this under the\nhood<\/p>\n<pre>emacsclient -a \"\" -c -e &lt;elisp code&gt;\n<\/pre>\n<p>You can prefix <strong>emacsclient<\/strong> with it's path if <strong>Emacs Anywhere<\/strong> can't find\nit in your <code>PATH<\/code> environment variable.<\/p>\n<h2><a id=\"user-content-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#commands\"><\/a>Commands<\/h2>\n<table><thead><tr><th>Command<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><strong>toggle-ea<\/strong><\/td>\n<td>If toggled to the \"off\" state, <strong>Emacs Anywhere<\/strong> will do nothing when the frame is deleted.<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-variables\" class=\"anchor\" aria-hidden=\"true\" href=\"#variables\"><\/a>Variables<\/h2>\n<table><thead><tr><th>Variable<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><strong>ea-on<\/strong><\/td>\n<td>Boolean where <code>t<\/code> and <code>nil<\/code> denote the \"on\" and \"off** states of <strong>Emacs Anywhere<\/strong>, respectively. The default value is <code>t<\/code><\/td>\n<\/tr><tr><td><strong>ea-copy<\/strong><\/td>\n<td>If true, the \"*Emacs Anywhere*\" buffer will be copied to clipboard. The default value is <code>t<\/code><\/td>\n<\/tr><tr><td><strong>ea-paste<\/strong><\/td>\n<td>If true, <strong>Emacs Anywhere<\/strong> will paste to the window from which it was launched. The default value is <code>t<\/code><\/td>\n<\/tr><tr><td><strong>ea-app-name<\/strong><\/td>\n<td>The name of the application process for the window from which <strong>Emacs Anywhere<\/strong> was launched.<\/td>\n<\/tr><tr><td><strong>ea-window-title<\/strong><\/td>\n<td>The title of the window from which <strong>Emacs Anywhere<\/strong> was launched.<\/td>\n<\/tr><tr><td><strong>ea-x<\/strong><\/td>\n<td>The x-coordinate of the upper-left corner of the window from which <strong>Emacs Anywhere<\/strong> was launched.<\/td>\n<\/tr><tr><td><strong>ea-y<\/strong><\/td>\n<td>The y-coordinate of the upper-left corner of the window from which <strong>Emacs Anywhere<\/strong> was launched.<\/td>\n<\/tr><tr><td><strong>ea-width<\/strong><\/td>\n<td>The width of the window from which <strong>Emacs Anywhere<\/strong> was launched.<\/td>\n<\/tr><tr><td><strong>ea-height<\/strong><\/td>\n<td>The height of the window from which <strong>Emacs Anywhere<\/strong> was launched.<\/td>\n<\/tr><\/tbody><\/table><blockquote>\n<p>Note: In OS X the window title, position and size information\nis only available if the application is has accessiblity permissions enabled.<\/p>\n<\/blockquote>\n<h2><a id=\"user-content-hooks\" class=\"anchor\" aria-hidden=\"true\" href=\"#hooks\"><\/a>Hooks<\/h2>\n<table><thead><tr><th>Hook<\/th>\n<th>Function Signature<\/th>\n<th>Description<\/th>\n<\/tr><\/thead><tbody><tr><td><strong>ea-popup-hook<\/strong><\/td>\n<td>(app-name window-title x y w h)<\/td>\n<td>Functions run after an <strong>Emacs Anywhere<\/strong> session starts.<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; Define a function or use a lambda of the same signature\n(defun popup-handler (app-name window-title x y w h)\n  ;; do stuff\n  )\n;; Hook your function\n(add-hook 'ea-popup-hook 'popup-handler)<\/pre><\/div>\n<p>In your hook function you can do things like set the major mode based on the\napplication name or window title...<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun github-conversation-p (window-title)\n  (or (string-match-p \"Pull Request\" window-title)\n      (string-match-p \"Issue\" window-title)\n      ;; ...\n      ))\n(defun popup-handler (app-name window-title x y w h)\n  ;; set major mode\n  (cond\n    ((github-conversation-p window-title) (gfm-mode))\n    ;; ...\n    (t (markdown-mode)) ; default major mode\n    ))<\/pre><\/div>\n<p>...or set the frame size and position to fit the bottom 400px of the window...<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun popup-handler (app-name window-title x y w h)\n  (set-frame-position (selected-frame) x (+ y (- h 400)))\n  (unless (zerop w)\n    (set-frame-size (selected-frame) w 400 t))\n  ;; ...\n  )<\/pre><\/div>\n<p>...or configure automation settings<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun popup-handler (app-name window-title x y w h)\n  (when (equal app-name \"Terminal\")\n    ;; Tell Emacs Anywhere not to paste if launched from Terminal\n    (setq ea-paste nil))\n  ;; ...\n  )<\/pre><\/div>\n<h2><a id=\"user-content-update\" class=\"anchor\" aria-hidden=\"true\" href=\"#update\"><\/a>Update<\/h2>\n<pre>~\/.emacs_anywhere\/update\n<\/pre>\n<h2><a id=\"user-content-uninstall\" class=\"anchor\" aria-hidden=\"true\" href=\"#uninstall\"><\/a>Uninstall<\/h2>\n<pre>\n~\/.emacs_anywhere\/uninstall\n<\/pre>\n<ul><li>Windows<\/li>\n<\/ul><p>TBD<\/p>\n<p>MIT<\/p>\n<p>Copyright \u00a9 2018, Zach Curry, All rights reserved.<\/p>\n<\/article>","created_at":"2018-11-07T16:43:23+0000","updated_at":"2018-11-07T16:43:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/7032632?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":290,"title":"Parameterizing ClojureScript Builds \u2014 Martin Klepsch","url":"https:\/\/www.martinklepsch.org\/posts\/parameterizing-clojurescript-builds.html","content":"<p><a class=\"link\" href=\"https:\/\/www.martinklepsch.org\/posts\/parameterizing-clojurescript-builds.html\" title=\"Permalink: Parameterizing ClojureScript Builds\">August 2015<\/a>\/<a class=\"link\" href=\"https:\/\/www.martinklepsch.org\/\" title=\"Home\">Home<\/a>\/<a class=\"link\" href=\"https:\/\/twitter.com\/martinklepsch\" title=\"@martinklepsch on Twitter\">@martinklepsch<\/a><\/p><section class=\"mkdwn lh-copy\"><p>Just like with most server side software we often want to make minor changes to the behaviour of the code depending on the environment it's run in. This post highlights language and compiler features of ClojureScript making parameterized builds easy peasy.<\/p><p>On servers environment variables are a go-to solution to set things like a database URI. In ClojureScript we don't have access to those. You can work around that with macros and emit code based on environment variables but this requires additional code and separate tools.<\/p><p>With ClojureScript 1.7.48 (<strong>Update:<\/strong> There was a bug in 1.7.48 <code>goog-define<\/code>. Use 1.7.107 instead.) a new macro <code>goog-define<\/code> has been added which allows build customization at compile time using plain compiler options. Let's walk through an example:<\/p><pre>(ns your.app)\n(goog-define api-uri \"http:\/\/your.api.com\")\n<\/pre><p><code>goog-define<\/code> emits code that looks something like this:<\/p><pre>\/** @define {string} *\/\ngoog.define(\"your.app.api_uri\",\"http:\/\/your.api.com\");\n<\/pre><p>The <code>goog.define<\/code> function from Closure's standard library plus the JSDoc <code>@define<\/code> annotation tell the Closure compiler that <code>your.app.api_uri<\/code> is a constant that can be overridden at compile time.  To do so you just need to pass the appropriate <code>:closure-defines<\/code> compiler option:<\/p><pre>:closure-defines {'your.app\/api-uri \"http:\/\/your-dev.api.com\"}\n<\/pre><p><strong>Note:<\/strong> When using Leinigen quoting is implicit so there is no quote necessary before the symbol.<\/p><p><strong>Note:<\/strong> Sometimes for debugging you may want to pass the Closure define as a string. If you decide to do so make sure it matches the string in the <code>goog.define<\/code> call in your emitted Javascript (i.e. account for name mangling).<\/p><aside>\nPrior to 1.7.48 you could annotate things with <code>@define<\/code> but without\nusing <code>goog.define<\/code> overriding those defines is not possible when\nusing optimizations <code>:none<\/code> effectively making them much less useful.\n<\/aside><h3>Under the hood<\/h3><p>When compiling with <code>:advanced<\/code> optimizations the Closure compiler will automatically replace all occurrences of your defined constants with their respective values. If this leads to unreachable branches in your code they will be removed as <a href=\"https:\/\/developers.google.com\/closure\/compiler\/docs\/compilation_levels?hl=en#advanced_optimizations\">dead code<\/a> by the Closure compiler. Very handy to elide things like logging!<\/p><p>Without any optimizations (<code>:none<\/code>) <code>goog.define<\/code> makes sure the right value is used. There are two global variables it takes into account for that: <code>CLOSURE_UNCOMPILED_DEFINES<\/code> and <code>CLOSURE_DEFINES<\/code>. When you override the default value using <code>:closure-defines<\/code> the ClojureScript compiler prepends <code>CLOSURE_UNCOMPILED_DEFINES<\/code> with your overridden define to your build causing <code>goog.define<\/code> to use the value in there instead of the default value you defined in your source files.<\/p><p>For details see <a href=\"https:\/\/github.com\/google\/closure-library\/blob\/master\/closure\/goog\/base.js#L147-L172\">the source of goog.define<\/a>.<\/p><\/section><p><a class=\"link\" href=\"https:\/\/twitter.com\/martinklepsch\">@martinklepsch<\/a>, August 2015<\/p>","created_at":"2018-11-07T16:43:23+0000","updated_at":"2018-11-07T16:43:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"www.martinklepsch.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":291,"title":"The beginners guide to React Native and Firebase","url":"https:\/\/firebase.googleblog.com\/2016\/01\/the-beginners-guide-to-react-native-and_84.html","content":"<div class=\"post-content\" itemprop=\"articleBody\">\n\n<noscript>\n <figure class=\"profile\"><div class=\"profile-picture\">    <img alt=\"David East\" src=\"https:\/\/4.bp.blogspot.com\/-wSmP7Jf6lQ4\/VxfJYQMZJqI\/AAAAAAAABLY\/9pkWTtArIrIPoU_auQUcXZKLXWGWAw6WwCLcB\/s1600\/david_east.png\" \/><\/div><figcaption><strong>David East<\/strong><\/figcaption><\/figure><\/noscript><\/div>       <em>Developer Advocate<\/em>   <blockquote><div>\nUPDATE: We updated this tutorial for <a href=\"https:\/\/firebase.googleblog.com\/2016\/07\/firebase-react-native.html\">the Firebase 3.1.0 SDK<\/a> which now supports the Firebase Realtime Database and Firebase Authentication.\n<\/div><\/blockquote>\n\n<p>Here at Firebase, <a href=\"https:\/\/www.firebase.com\/blog\/2015-05-29-react-native.html\">we\u2019re<\/a> <a href=\"https:\/\/www.firebase.com\/blog\/2015-07-15-reactfire-0-5-0.html\">big<\/a> <a href=\"https:\/\/www.firebase.com\/blog\/2014-05-01-using-firebase-with-react.html\">React fans<\/a>. Firebase synchronizes application state, and React re-renders the application UI based on state changes. It\u2019s a perfect fit.<\/p><p>And with React Native, app development became a lot easier for JavaScript developers. Using React Native you can build real native apps with just JavaScript. And that\u2019s just awesome-sauce. We live in the future. Let\u2019s get started.<\/p>\n\n<h2 id=\"setup\">Setup<\/h2><p>If you want to dive right into the code, <a href=\"https:\/\/github.com\/davideast\/firebase-react-native-sample\">you can check out the final GitHub repo here.<\/a> Otherwise, let's go through this step-by-step.<\/p><p>Getting started with React Native is fairly easy, but there are some gotchas you should be aware of. If you already have React Native set up, you can skip this section.<\/p><p>First, you\u2019ll need <a href=\"http:\/\/brew.sh\/\">Homebrew<\/a>, which is easy to install. You\u2019ll also need <a href=\"https:\/\/nodejs.org\/en\/\">Node.js 4.0<\/a> or higher. The React Native team recommends using <a href=\"https:\/\/github.com\/creationix\/nvm#installation\">nvm to manage your node versions<\/a> <em>(and I do too)<\/em>.<\/p><p>Once those tools have been installed, run the following commands:<\/p><pre class=\"prettyprint lang-bash\">brew install watchman\nnpm install -g react-native-cli\n<\/pre><p>Then finally, you can begin your project with the CLI command:<\/p><pre class=\"prettyprint lang-bash\">react-native init GroceryApp # or whatever you want\n<\/pre><p>Open the main folder in your favorite editor.<\/p>\n<pre class=\"prettyprint lang-bash\">\natom GroceryApp # if you\u2019re into Atom\n<\/pre>\n\n<h2 id=\"build-and-run\">Build and run<\/h2><p>To build a React Native project, run the following command:<\/p>\n\n<pre class=\"prettyprint lang-bash\">\nreact-native run-ios\n<\/pre>\n\n<p>This should launch the Simulator, and you should see the boilerplate screen.<\/p><p><br \/><img alt=\"Blank App\" src=\"https:\/\/fireassets.firebaseapp.com\/resources\/images\/blog\/guide-react-native\/rn-blank-app.png\" \/><br \/><\/p><p>React Native comes with <strong>hot-reloading<\/strong>, which means you can make an edit to the code, index.ios.js, and then hit <code>Cmd+R<\/code> and see your changes instantly update. And if you\u2019re an iOS developer, or really any kind of developer, you know how cool that is.<\/p><p>With the build setup done, let\u2019s get Firebase up and running.<\/p>\n\n<h2 id=\"firebase-setup\">Firebase Setup<\/h2><p>React Native manages dependencies through <code>npm<\/code>. To install Firebase, run the following command at the root of the project.<\/p>\n\n<pre class=\"prettyprint lang-bash\">\nnpm install firebase --save\n<\/pre>\n\n<p>Open <code>index.ios.js<\/code> and add the following line to the top:<\/p>\n\n<pre class=\"prettyprint lang-bash\">\nimport * as firebase from 'firebase';\n<\/pre>\n\n<p>\nThen right above the component, initialize Firebase with your config values:\n<\/p>\n\n<pre class=\"prettyprint lang-js\">\n\/\/ Initialize Firebase\nconst firebaseConfig = {\n  apiKey: \"&lt;your-api-key&gt;\",\n  authDomain: \"&lt;your-auth-domain&gt;\",\n  databaseURL: \"&lt;your-database-url&gt;\",\n  storageBucket: \"&lt;your-storage-bucket&gt;\",,\n};\nconst firebaseApp = firebase.initializeApp(firebaseConfig);\n<\/pre>\n\n<p><a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Statements\/const\">What\u2019s a <code>const<\/code>?<\/a> A <code>const<\/code> is a read-only reference to a value, which makes sense here, because you don\u2019t want to ever override the value of Firebase. You can use this keyword since you\u2019re using Node 4.0 or higher. If your editor yells at you, it\u2019s telling you a lie.<\/p><p>One ES2015 feature isn\u2019t enough. Rather than use <code>React.createClass()<\/code> to define components, let\u2019s use classes.<\/p>\n\n<h2 id=\"es2015-es6-components\">ES2015 (ES6) Components<\/h2><p>React is component-based. Which means an app is just a tree of components, starting with a root component. <a href=\"https:\/\/facebook.github.io\/react\/blog\/2015\/10\/07\/react-v0.14.html\">As of React 0.14<\/a>, you can use ES2015 classes to define React components.<\/p><p>In <code>index.ios.js<\/code>, let\u2019s change the component to use a class rather than <code>React.createClass()<\/code>.<\/p><pre class=\"prettyprint lang-bash\">class GroceryApp extends Component {\n  render() {\n    return (\n      &lt;View style=\"{styles.container}\"&gt;\n\n      &lt;\/View&gt;\n    );\n  }\n}\n<\/pre><p>Why ES2015 classes over <code>React.createClass()<\/code>? <a href=\"http:\/\/reactjsnews.com\/composing-components\/\">While there is quite the debate over the subject<\/a>, it boils down to a matter of taste.<\/p><p>The shell of the app is complete. Let\u2019s make it look good, or at least halfway decent.<\/p>\n\n<h2 id=\"styling-in-react-native\">Styling in React Native<\/h2><p>React Native uses JavaScript rather than CSS for styling. This may sound like an extreme divergence, but it\u2019s really not all that different. To declare a set of styles, you create a <code>StyleSheet<\/code>.<\/p><pre class=\"prettyprint lang-js\">var styles = StyleSheet.create({\n  container: {\n    backgroundColor: '#f2f2f2',\n    flex: 1,\n  },\n});\n<\/pre><p>A <code>StyleSheet<\/code> contains a set of objects that represent CSS-like styles. Then you use these styles on a React component:<\/p><pre class=\"prettyprint lang-js\">&lt;View style=\"{styles.container}\"&gt;\n  I\u2019m a container lol!\n&lt;\/View&gt;\n<\/pre><p>So don\u2019t worry about losing any of your CSS skills. If anything, <a href=\"https:\/\/css-tricks.com\/snippets\/css\/a-guide-to-flexbox\/\">you should really learn CSS Flexbox to make styling in React Native a breeze<\/a>.<\/p><p>Now that you\u2019re a pro at styling in React, let\u2019s declare the styles for the app.<\/p>\n\n<h2 id=\"add-styles\">Add Styles<\/h2><p>Create a file named <code>styles.js<\/code> and <a href=\"https:\/\/github.com\/davideast\/firebase-react-native-sample\/blob\/master\/styles.js\">add the following code from this file<\/a>. These are the styles we\u2019ll be using for the app.<\/p><p>You\u2019ll notice that React Native uses CommonJS modules. At the bottom of <code>styles.js<\/code>, the StyleSheet is exported using <code>module.exports<\/code>.<\/p><p>This will allow you to import these styles using <code>require()<\/code>. Open <code>index.ios.js<\/code>, and add the following line of code:<\/p><pre class=\"prettyprint lang-bash\">const styles = require('.\/styles.js')\n<\/pre><p><strong>Make sure to remove the styles variable at the bottom of the file.<\/strong><\/p><p>The styles are now in place. Let\u2019s take a look at the app\u2019s component structure.<\/p>\n\n<h2 id=\"break-the-ui-into-a-component-hierarchy\">Break the UI into a component hierarchy<\/h2><p>The best advice I have ever read on using React, <a href=\"https:\/\/facebook.github.io\/react\/docs\/thinking-in-react.html#step-1-break-the-ui-into-a-component-hierarchy\">is to start off by breaking the UI into a component hierarchy<\/a>. Below is a visual outline of the app's component hierarchy.<\/p><p><br \/><img alt=\"App Hierarchy\" src=\"https:\/\/fireassets.firebaseapp.com\/resources\/images\/blog\/guide-react-native\/rn-hierarchy-sm.png\" \/><br \/><\/p><p>The app is made up of five components:<\/p><ol><li>GroceryApp (orange): contains the entire app<\/li>\n<li>StatusBar (purple): displays the title of the view<\/li>\n<li>ListView (green): displays the list of grocery items<\/li>\n<li>ListItem <strong>(black)<\/strong>: displays an individual item from the list<\/li>\n<li>ActionButton (blue): adds an item to the list<\/li>\n<\/ol><p>Create a folder named <code>components<\/code>. Each one of these components is stored in the components folder. The exception is <code>GroceryApp<\/code>, since it is contained in <code>index.ios.js<\/code>.<\/p><p>Add each one of these components below to the <code>components<\/code> folder:<\/p><p><strong>ActionButton.js<\/strong><\/p>\n\n<pre class=\"prettyprint lang-js\">\n'use strict';\n\nimport React, {Component} from 'react';\nimport ReactNative from 'react-native';\nconst styles = require('..\/styles.js')\nconst constants = styles.constants;\nconst { StyleSheet, Text, View, TouchableHighlight} = ReactNative;\n\nclass ActionButton extends Component {\n  render() {\n    return (\n      &lt;View style={styles.action}&gt;\n        &lt;TouchableHighlight\n          underlayColor={constants.actionColor}\n          onPress={this.props.onPress}&gt;\n          &lt;Text style={styles.actionText}&gt;{this.props.title}&lt;\/Text&gt;\n        &lt;\/TouchableHighlight&gt;\n      &lt;\/View&gt;\n    );\n  }\n}\n\nmodule.exports = ActionButton;\n<\/pre>\n\n<p>\n<strong>ListItem.js<\/strong><\/p>\n<pre class=\"prettyprint lang-js\">\nimport React, {Component} from 'react';\nimport ReactNative from 'react-native';\nconst styles = require('..\/styles.js')\nconst { View, TouchableHighlight, Text } = ReactNative;\n\nclass ListItem extends Component {\n  render() {\n    return (\n      &lt;TouchableHighlight onPress={this.props.onPress}&gt;\n        &lt;View style={styles.li}&gt;\n          &lt;Text style={styles.liText}&gt;{this.props.item.title}&lt;\/Text&gt;\n        &lt;\/View&gt;\n      &lt;\/TouchableHighlight&gt;\n    );\n  }\n}\n\nmodule.exports = ListItem;\n<\/pre>\n<p><strong>StatusBar.js<\/strong><\/p>\n<pre class=\"prettyprint lang-js\">\n'use strict';\nimport React, {Component} from 'react';\nimport ReactNative from 'react-native';\nconst styles = require('..\/styles.js')\nconst { StyleSheet, Text, View} = ReactNative;\n\nclass StatusBar extends Component {\n  render() {\n    return (\n      &lt;View&gt;\n        &lt;View style={styles.statusbar}\/&gt;\n        &lt;View style={styles.navbar}&gt;\n          &lt;Text style={styles.navbarTitle}&gt;{this.props.title}&lt;\/Text&gt;\n        &lt;\/View&gt;\n      &lt;\/View&gt;\n    );\n  }\n}\n\nmodule.exports = StatusBar;\n<\/pre>\n\n<p>With the components added, let\u2019s make a static version of the app.<\/p>\n\n<h2 id=\"static-mockup\">Static Mockup<\/h2><p>In <code>index.ios.js<\/code>, add the following imports to the top of the page:<\/p>\n\n<pre class=\"prettyprint lang-js\">\nimport React, {Component} from 'react';\nimport ReactNative from 'react-native';\nimport * as firebase from 'firebase';\nconst StatusBar = require('.\/components\/StatusBar');\nconst ActionButton = require('.\/components\/ActionButton');\nconst ListItem = require('.\/components\/ListItem');\nconst styles = require('.\/styles.js');\n<\/pre>\n\n<p>Then add the following snippet of code:<\/p>\n\n<pre class=\"prettyprint lang-js\">\n_renderItem(item) {\n    return (\n      &lt;ListItem item=\"{item}\" onpress=\"{()\" ==\"\"&gt; {}} \/&gt;\n    );\n  }\n\n  render() {\n    return (\n      &lt;View style=\"{styles.container}\"&gt;\n\n        &lt;StatusBar title=\"Grocery List\"&gt;\n\n        &lt;ListView datasource=\"{this.state.dataSource}\" renderrow=\"{this._renderItem.bind(this)}\" style=\"{styles.listview}\/\"&gt;\n\n        &lt;ActionButton title=\"Add\" onpress=\"{()\" ==\"\"&gt; {}} \/&gt;\n\n      &lt;\/View&gt;\n    );\n  }\n<\/pre>\n\n<p>The <code>render()<\/code> function is the main view of the app, and <code>_renderItem()<\/code> sets the individual items in the list.<\/p><p>Next, create a constructor for the root component, <code>GroceryApp<\/code>.<\/p><pre class=\"prettyprint lang-js\">constructor(props) {\n  super(props);\n  this.state = {\n    dataSource: new ListView.DataSource({\n      rowHasChanged: (row1, row2) =&gt; row1 !== row2,\n    })\n  };\n}\n<\/pre><p>The component has a special property called <code>state<\/code>, which manages the entire data flow of the application. We\u2019ll dive deeper in the next section.<\/p><p>The app\u2019s state is a <code>ListView.DataSource<\/code>, which is a class that provides efficient data processing to a <code>ListView<\/code> component. The next step is to display this data.<\/p><p>Each component has a lifecycle, where certain functions are called at important events. When the component has first been rendered, <code>componentDidMount()<\/code> is called. This is where we want set any initial state of the app.<\/p><pre class=\"prettyprint lang-js\">componentDidMount() {\n    this.setState({\n      dataSource: this.state.dataSource.cloneWithRows([{ title: 'Pizza' }])\n    })\n  }\n<\/pre><p>Build and run the app, and you should see the following static app.<\/p><p><br \/><img src=\"https:\/\/fireassets.firebaseapp.com\/resources\/images\/blog\/guide-react-native\/rn-static-app.png\" alt=\"image\" \/><br \/><\/p><p>Each one of these components just simply displays data, or sets a callback function for a touch.<\/p><p>The key thing to understand is that these components are not state-ful. They have properties that are set by their root component, <code>GroceryApp<\/code>. To begin to understand React, you have to learn how to manage state.<\/p>\n\n<h2 id=\"state\">State<\/h2><p>State is simply data that can change. This data is called state, because it represents the \"state\" of your application. If this data changes, your application will likely look different, hence being in a different \"state\". State is usually things like a list of todos, or an enabled\/disabled button.<\/p><p>The root component will serve as the holder of state. State changes start at the root component, which is then responsible for updating the properties of its child components.<\/p><p>Component properties are immutable, meaning they can\u2019t be modified. <em>So, if you can\u2019t modify the properties, how do they ever change?<\/em> You re-render, the application by calling <code>setState()<\/code>, like seen in <code>componentDidMount()<\/code>.<\/p><p>The <code>setState()<\/code> function is special because every time it is called, it will attempt to re-render the entire app. This means that if a child property is different from the last state, it will be re-rendered with the new value.<\/p><p>This is where React fits perfectly with Firebase. Because the Firebase database synchronizes application state across multiple devices, and React efficiently re-renders application state changes.<\/p>\n\n<h2 id=\"firebase-listener\">Realtime Database listener<\/h2>\n<p>Create Realtime Database reference as a property in the constructor:<\/p>\n\n<pre class=\"prettyprint lang-js\">\nthis.itemsRef = firebaseApp.database().ref();\n<\/pre>\n\n<p>Then, add the following function to the <code>GroceryApp<\/code> component:<\/p>\n\n<pre class=\"prettyprint lang-js\">\nlistenForItems(itemsRef) {\n    itemsRef.on('value', (snap) =&gt; {\n\n      \/\/ get children as an array\n      var items = [];\n      snap.forEach((child) =&gt; {\n        items.push({\n          title: child.val().title,\n          _key: child.key\n        });\n      });\n\n      this.setState({\n        dataSource: this.state.dataSource.cloneWithRows(items)\n      });\n\n    });\n  }\n<\/pre><p>This function creates a value listener for all grocery items. Whenever an item is added, changed, or removed, you\u2019ll get the entire result set back as a <a href=\"https:\/\/www.firebase.com\/docs\/web\/api\/datasnapshot\"><code>DataSnapshot<\/code><\/a>, from the Firebase SDK. Using the <code>DataSnapshot<\/code>, you call <code>forEach(child)<\/code>, which iterates through all of the children and adds them to an array as a grocery list item. Notice in the <code>.forEach<\/code> function, a <code>_key<\/code> property is created from the DataSnapshot's <code>.key()<\/code> value. This makes life much easier when doing data operations later down the line.<\/p><p>Once the array is populated, you\u2019ll update the dataSource property on state using <code>dataSource.cloneWithRows(items)<\/code>. The <code>cloneWithRows()<\/code> function is just a convenience method for creating a new <code>ListView.DataSource<\/code> based on the same <code>DataSource<\/code> previously defined.<\/p><p>Next, you write the listen up to <code>componentDidMount()<\/code>:<\/p><pre class=\"prettyprint lang-js\">componentDidMount() {\n    this.listenForItems(this.itemsRef);\n  }\n<\/pre><p>Build and run the app. You should see an empty page, but try adding a few items using the Firebase App Dashboard or the <a href=\"https:\/\/chrome.google.com\/webstore\/detail\/vulcan-by-firebase\/oippbnlmebalopjbkemajgfbglcjhnbl?hl=en\">super awesome data viewer Vulcan<\/a>, and you\u2019ll see it update in realtime!<\/p><p><br \/><img alt=\"Vulcan\" src=\"https:\/\/fireassets.firebaseapp.com\/resources\/images\/blog\/guide-react-native\/rn-vulcan.png\" \/><br \/><\/p><p>This is awesome, but you\u2019d rather have the \"Add\" button working. Let\u2019s tackle that in the next section.<\/p>\n\n<h2 id=\"add-item\">Add Item<\/h2><p>When the <code>ActionButton<\/code> is tapped, an alert should pop up prompting the user to enter an item. The <code>AlertIOS<\/code> API is what you\u2019ll use to create this alert box.<\/p><p><br \/><img alt=\"Adding items\" src=\"https:\/\/fireassets.firebaseapp.com\/resources\/images\/blog\/guide-react-native\/rn-add-item.png\" \/><br \/><\/p><p>Add the following function to the <code>GroceryApp<\/code> component:<\/p><pre class=\"prettyprint lang-js\">_addItem() {\n    AlertIOS.prompt(\n      'Add New Item',\n      null,\n      [\n        {\n          text: 'Add',\n          onPress: (text) =&gt; {\n            this.itemsRef.push({ title: text })\n          }\n        },\n      ],\n      'plain-text'\n    );\n  }\n<\/pre><p><a href=\"https:\/\/facebook.github.io\/react-native\/docs\/alertios.html\">The <code>AlertIOS<\/code> API<\/a> is quite extensible when it comes to building alerts. The first two parameters are simple, they're just the title of the alert box and an optional message. The third parameter is the meat of the API. Here you create an array that specifies the buttons available to the user. Each button can have a <code>text<\/code>, <code>style<\/code>, and an <code>onPress<\/code> callback function. The last parameter is type the of input, whether <code>plain-text<\/code> or <code>secure-text<\/code>.<\/p><p>To add an item, create an object in the buttons array. This object can add items in the <code>onPress<\/code> callback. The callback returns the text the user has entered. Use this text to <code>.push()<\/code> a new child onto the <code>\/items<\/code> location.<\/p><p>Next, you'll need to update <code>render()<\/code> to assign the onPress property of the <code>ActionButton<\/code>:<\/p><pre class=\"prettyprint lang-js\">&lt;ActionButton title=\"Add\" onpress=\"{this._addItem.bind(this)}\"&gt;\n&lt;\/ActionButton&gt;<\/pre><p>Build and run. When you tap add, and enter an item title, you should see it update on the list.<\/p><p>Awesome, but what\u2019s a grocery list that can\u2019t complete items?<\/p>\n\n<h2 id=\"complete-items\">Complete Items<\/h2><p>The user can complete an item by tapping on it to open up an alert box. If they press the \"Complete\" option, you can delete it from the list.<\/p><p>Modify <code>_renderItem(item)<\/code> to include an <code>onPress<\/code> callback:<\/p><pre class=\"prettyprint lang-js\">_renderItem(item) {\n\n    const onPress = () =&gt; {\n      AlertIOS.prompt(\n        'Complete',\n        null,\n        [\n          {text: 'Complete', onPress: (text) =&gt; this.itemsRef.child(item._key).remove()},\n          {text: 'Cancel', onPress: (text) =&gt; console.log('Cancel')}\n        ],\n        'default'\n      );\n    };\n\n    return (\n      &lt;ListItem item=\"{item}\" onpress=\"{onPress}\"&gt;\n    );\n  }\n<\/pre><p>To \"complete\" an item, you'll need to delete it from the Firebase database. Using <code>.child(key)<\/code>, you can drill down into a specific item in the list. The <code>onPress<\/code> callback is a closure, therefore it has access to the outer environment which contains the <code>item<\/code> parameter. <strong>This is where the <code>_key<\/code> property comes in hand.<\/strong><\/p><p>When the \"Complete\" option is tapped, you can find the specific child by using the <code>_key<\/code> property on <code>item<\/code>. Then you can call <code>.remove()<\/code> to delete the item in the Firebase database.<\/p><p>Build and run, yet again. Tap any <code>ListItem<\/code> , tap \"Complete\", and you should see it removed from the list.<\/p>\n\n<h2 id=\"get-the-code-give-a-star\">Get the code, give a star<\/h2><p><a href=\"https:\/\/github.com\/davideast\/firebase-react-native-sample\">Check out the completed app on Github<\/a>. And, if you\u2019re feeling generous, we would love a star. Feel free to fork the repo, and even send in a PR if you want.<\/p>\n\n<h2 id=\"still-having-trouble-\">Still having trouble?<\/h2>\n<p>If you\u2019re running into issues, open up a question on Stackoverflow, we monitor the Firebase tag closely, or drop a line in our <a href=\"https:\/\/firebase-community.appspot.com\">community Slack Team<\/a>.<\/p>","created_at":"2018-11-07T16:43:23+0000","updated_at":"2018-11-07T16:43:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"firebase.googleblog.com","preview_picture":"https:\/\/4.bp.blogspot.com\/-wSmP7Jf6lQ4\/VxfJYQMZJqI\/AAAAAAAABLY\/9pkWTtArIrIPoU_auQUcXZKLXWGWAw6WwCLcB\/s1600\/david_east.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":288,"title":"The World of Linux Handhelds in 2018 \u2013 Giant Pockets","url":"https:\/\/www.giantpockets.com\/the-world-of-linux-handhelds-in-2018\/","content":"<p>So here is 2018. It\u2019s a kind of middle ground for Linux handhelds. After all, the concept of Linux handhelds preceded the advent of smartphones and since then there is probably less of a need for the general public \u2013 iOS and Android already seem to answer the needs of most users so far. In fact, it has clearly become a niche market, with only smaller companies operating in it. But the good news is, there are several efforts underway to get there, in several directions. There are several categories we can split them in. Note that in this article we will remain at relatively high level (and refrain from comparing every single spec under the radar) as this is meant to be an overview.<\/p>\n<p>\n<\/p><h3><strong>Gaming Handhelds<\/strong><\/h3>\n<p>Gaming Handhelds using Linux have been a thing for a long time by now. While we could expand our list, the most promising ones that should be available sometimes in the near future are the following.<\/p>\n<p><strong>The DragonBox Pyra<\/strong>: it is getting closer \u2013 for those who are not aware, <a href=\"https:\/\/pyra-handheld.com\/boards\/pages\/pyra\/\">it\u2019s a supercharged version of the Open Pandora<\/a>, with numerous improvements both in performance and hardware design. It features both gaming controls and a full keyboard so that it can be used in many ways (both in emulators as well as typical computing tasks including coding). It\u2019s been under development for years, and it\u2019s taking quite a while, yet in 2018 or 2019 it may finally come out. There is only a few people attached to the project so that explains why they are slower to reach the market compared to companies like GPD.<\/p>\n<p><img class=\"alignnone size-full wp-image-5623\" src=\"http:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/pyra.jpg\" alt=\"\" width=\"800\" height=\"600\" srcset=\"https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/pyra.jpg 800w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/pyra-300x225.jpg 300w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/pyra-768x576.jpg 768w\" \/><\/p>\n<p><em>Above Pyra is a render by <a href=\"https:\/\/pyra-handheld.com\/boards\/members\/erico.20949\/\">Erico<\/a> (active user in the OpenPandora boards)<\/em><\/p>\n<p>Its hardware is ARM-based (Texas Instrument OMAP5), relies on a 720p screen, and it is likely to be relatively obsolete by the time of its release, but its strength is that it has a modular design that should enable SoC upgrades down the road, and features many extensions ports in a very compact design. On top of that, the development is happening pretty much in the open, with good communication and good involvement with the interested parties. Anybody familiar with the Open Pandora community will also know that the software support should be pretty good down the road (the Open Pandora still gets ports of modern software as we speak in 2018, despite being almost 10 years old design wise!). OS-wise, it will run Debian (probably not mainline though), while it\u2019s very likely we will see community ports of many other distributions, just like what happened on the Pandora. There are potential areas of friction like actual software support at launch, availability of 3D drivers (since there are no mature open drivers for the PowerVR GPU), and a potentially high price point. On the other hand, it\u2019s going to be really unique in its category, so those interested may not be so price sensitive in the first place.<\/p>\n<p><strong>Smach Z<\/strong>: It was initially thought as a portable Steam Machine, back when the project started years ago. While this company has never delivered any hardware so far, this design is more focused towards gaming and it has no keyboard. It emulates the design of the Steam Controller with haptic controls for more flexibility. Its screen is going to be 1080p resolution, which is a curious choice for the size of this thing.<\/p>\n<p><img class=\"alignnone size-full wp-image-5624\" src=\"http:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/smachz.jpg\" alt=\"\" width=\"680\" height=\"381\" srcset=\"https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/smachz.jpg 680w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/smachz-300x168.jpg 300w\" \/><\/p>\n<p>It\u2019s still very unclear at this stage by what time it will be released as the project team has already missed several milestones since the start of their crowdfunding campaign. They have recently disclosed that<a href=\"https:\/\/boilingsteam.com\/the-smach-z-handheld-is-still-alive\/\"> they will use an AMD chip<\/a> (Ryzen\/Vega V1605B) which will bring full compatibility with x86 games (i.e. the world of PC gaming).<\/p>\n<p><iframe width=\"648\" height=\"365\" src=\"https:\/\/www.youtube.com\/embed\/V74zEOy4HPs?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>\n<p>It will be released with SteamOS or Windows10, and should be relatively powerful and capable, hardware-wise, but it may quite large (it needs a fan for cooling and a high capacity battery) and it\u2019s not clear how long you will be able to play with it continuously. You can also expect a fairly high price point, while no clear details have been communicated yet. The tentative ones are between <a href=\"https:\/\/liliputing.com\/2018\/03\/smach-z-handheld-gaming-pc-pre-orders-open-march-15th-629.html\">629 USD and 809 USD<\/a> depending on the options you get with it.<\/p>\n<p><strong>GPD Win 2<\/strong>: While there is apparently no official support for Linux on the GPD Win2, folks with prototypes have already shown it was apparently trivial to install and run Ubuntu on it with pretty much full hardware support (except pads). It features relatively powerful hardware for its size (Core m3-7Y30 with integrated HD Graphics 615, 8GB of RAM) and a 720p screen just like the Pyra.<\/p>\n<p><img class=\"alignnone size-full wp-image-5625\" src=\"http:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gpd-win2.jpg\" alt=\"\" width=\"800\" height=\"600\" srcset=\"https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gpd-win2.jpg 800w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gpd-win2-300x225.jpg 300w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gpd-win2-768x576.jpg 768w\" \/><\/p>\n<p>The GPD Win 1 could also run a modified version of Arch Linux (not mainline, since it needed numerous patches for wifi, touch screen and controls), so we can expect the same to be possible with the GPD Win 2. While it should be a decent machine for PC gaming on the go (apparently capable of running <strong>Overwatch<\/strong> on low settings at a very decent framerate), you should not have expect it to be really usable for typical computing tasks. If it is anything like the GPD Win 1, the keyboard is likely to be sub-par, for example. For those who are interested in Intel-based hardware equipped with keyboard, it may be the only option though. Actual retail price is said to be around 899 USD, and should be available for purchase around May 2018.<\/p>\n<p><iframe width=\"648\" height=\"365\" src=\"https:\/\/www.youtube.com\/embed\/jpWi4nUK6zY?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/p>\n<h3><strong>Non-gaming Handhelds<\/strong><\/h3>\n<p><strong>GPD Pocket:<\/strong> GPD has been pretty active in recent history and the GPD Pocket is a new type of device that can be best described as a micro-laptop. Imagine something like a Macbook, shrink it to your palm\u2019s size and you have an idea of how tiny it is. It has absolutely no gaming controls and no trackpad either, using a nub for the user to move the mouse around (as well as a touch screen). This is not for thumb-typing however and it made to be used just like a regular laptop despite its smaller form factor. It has very few expansion ports and limited memory, but it is a cool alternative for folks who are just interested to do some work on the go.<\/p>\n<p><img class=\"alignnone size-full wp-image-5620\" src=\"http:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/GPD-Pocket.png\" alt=\"\" width=\"640\" height=\"316\" srcset=\"https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/GPD-Pocket.png 640w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/GPD-Pocket-300x148.png 300w\" \/><\/p>\n<p>Note that the keyboard layout is far from ideal, and the lack of trackpad makes it somewhat awkward to use for long periods of time. It would be great to see a better designed version of the same concept, if sales figures were good enough for GPD to continue supporting it. In terms of OS, you can get it to run Windows or Ubuntu (officially) and GPD has published the diff of their changes made to the distribution. You can already find it online at about 510 USD while the official list price is rather in the 700s.<\/p>\n<p><strong>Gemini:<\/strong> The spiritual successor of the Psion portable computers<a href=\"https:\/\/www.indiegogo.com\/projects\/gemini-pda-android-linux-keyboard-mobile-device-phone#\/\"> is almost there<\/a>. It will run either Android or Linux, while the flavor of Linux is not yet confirmed. It should be very compact and have a great mechanical keyboard that you can use pretty much like a laptop one. There is absolutely no mouse controls so you have to use your fingers to emulate clicks on screen. This is fine if you use Android, but for the Linux version it still remains to be seen how comfortable this will be. It features a very high resolution screen (2160\u00d71080 for 6 inches) so don\u2019t expect any serious gaming on this unit with such high pixel density.<\/p>\n<p><img class=\"alignnone size-full wp-image-5621\" src=\"http:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gemini.jpg\" alt=\"\" width=\"770\" height=\"433\" srcset=\"https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gemini.jpg 770w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gemini-300x169.jpg 300w, https:\/\/www.giantpockets.com\/wp-content\/uploads\/2018\/03\/gemini-768x432.jpg 768w\" \/><\/p>\n<p>It\u2019s ARM-based as well, and while the official release date is not 100% confirmed it looks like <a href=\"https:\/\/www.indiegogo.com\/projects\/gemini-pda-android-linux-keyboard-mobile-device-phone#\/updates\/all\">they are shipping some first units to backers as soon as Q1 2018<\/a>, with a broader release in November 2018.<\/p>\n<h3>My take on all this<\/h3>\n<p>I have somewhat mixed feelings about where these devices go. First, I am really glad to see a lot more activity and official support of Linux on more and more hardware, because more choice means more competition and likely better options altogether for several types of users. However, I am quite concerned with the overall trend to make such devices more and more premium, price-wise. The original Pandora could be had (if you were lucky to get one) at less than 500 USD, but most of the newer devices are likely hitting way above and closer to 1000 USD while at the same time the hardware has become more and more commoditized, thanks to the advent of smartphones. On one hand we can get relatively powerful ARM boards like the Raspberry Pi or Odroid boards for less than 100 USD, but having fully packaged machines with screen, controls and keyboards which are somewhat custom made for each of these machines make them really expensive in the end. On top of that, most of them do not have replaceable parts, and cannot be upgraded easily. This comes with the form factor, but I rather like what Dragonbox is doing there with having as many parts user-replaceable, including the battery and the screen. On the same level, official Linux distribution support remains a question mark beyond release. Will GPD for example update its Ubuntu image to 18.04 in April this year? What will be Gemini\u2019s policy on that matter? Solid software support is as important as hardware design and we need to also compare such machines under that lens when they come out. In any case I would strongly advise any potential buyer to be extremely careful when it comes to pre-orders. Many have been burnt before by such practices (as in, no delivery of actual hardware in the end) so do not invest any money you are not ready to lose. Better wait until it\u2019s actually available, no matter how \u201cattractive\u201d early birds deal look like.<\/p>\n<div class=\"yarpp-related\">\n<h3>Related posts:<\/h3><ol><li><a href=\"https:\/\/www.giantpockets.com\/pandora-survey\/\" rel=\"bookmark\" title=\"So, What\u2019s the Open Pandora Community Made Of ?\">So, What\u2019s the Open Pandora Community Made Of ? <\/a><\/li>\n<\/ol><\/div>","created_at":"2018-11-07T16:43:22+0000","updated_at":"2018-11-07T16:43:22+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"www.giantpockets.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":287,"title":"What are your thoughts on building landing pages before the product is available?","url":"https:\/\/www.indiehackers.com\/forum\/what-are-your-thoughts-on-building-landing-pages-before-the-product-is-available-e45dd6eec1","content":"<p>There's lots of advice on the Internet which suggests that building a landing page before even starting your product is a good way to validate a new business idea . As I understand it, the idea is to build up a list of potential customers and talk with them about their needs as the product evolves. I see how this could be a smart strategy, but it feels like having the landing page up creates an expectation that the service will exist relatively soon and I'm not sure having that time pressure is a great idea.<\/p><p>I'm starting my first company, and I'm bootstrapping by myself while I continue to work full-time as a software engineer. I like the idea of generating leads ASAP with a landing page, but it will take a while (6+ months) to build the product because of my limited time investment. Should I hold off on creating a landing page until I have a usable product, or am I worrying too much and should just get it out there now?<\/p>","created_at":"2018-11-07T16:43:20+0000","updated_at":"2018-11-07T16:43:20+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/580157013239-us-east-1-chromeless.s3.amazonaws.com\/thread-shareables\/e45dd6eec1.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":286,"title":"Coding Your Own Algo Trading Robot | Investopedia","url":"https:\/\/www.investopedia.com\/articles\/active-trading\/081315\/how-code-your-own-algo-trading-robot.asp","content":"<p>Ever wanted to become an <a href=\"https:\/\/www.investopedia.com\/terms\/a\/algorithmictrading.asp\">algorithmic trader<\/a> with the ability to code your own trading robot? And yet, are you frustrated with the amount of disorganized, misleading information and false promises of overnight prosperity? Well, Lucas Liew, creator of the online algorithmic trading course AlgoTrading101, may have the solution for you. Having excellent reviews and garnering over 8,000 students since first launching in October 2014, Liew\u2019s course \u2013 aimed at presenting the fundamentals of algorithmic trading in an organized way \u2013 is proving to be quite popular. He is adamant about the fact that algorithmic trading is \u201cnot a get-rich-quick scheme.\u201d<\/p><p>Outlined below are the basics of what it takes to design, build and maintain your own <a href=\"https:\/\/www.investopedia.com\/university\/systemcoding\/\">algorithmic trading robot<\/a>\u00a0(drawn from Liew and his course).<\/p><h2>What an Algorithmic Trading Robot Is and Does<\/h2><p>At the most basic level, an algorithmic trading robot is a computer code that has the ability to generate and execute buy and sell signals in financial markets. The main components of such a robot include <a href=\"https:\/\/www.investopedia.com\/terms\/e\/entry-point.asp\">entry rules<\/a> that signal when to buy or sell, <a href=\"https:\/\/www.investopedia.com\/terms\/e\/exit-point.asp\">exit rules<\/a> indicating when to close the current position, and <a href=\"https:\/\/www.investopedia.com\/terms\/p\/positionsizing.asp\">position sizing<\/a> rules defining the quantities to buy or sell.<\/p><h2>The Main Tools<\/h2><p>Obviously, you\u2019re going to need a computer and an Internet connection. After that, a Windows or Mac operating system will be needed to run <a href=\"http:\/\/www.metatrader4.com\/\">MetaTrader 4<\/a> (MT4) \u2013 an electronic <a class=\"termLink\" data-inlink=\"link_11\" href=\"https:\/\/www.investopedia.com\/terms\/t\/trading-platform.asp\">trading platform<\/a> that uses the MetaQuotes Language 4 (MQL4) for coding trading strategies. Although MT4 is not the only software one could use to build a robot it has a number of significant benefits.<\/p><p>While MT4\u2019s main <a class=\"termLink\" data-inlink=\"link_12\" href=\"https:\/\/www.investopedia.com\/terms\/a\/assetclasses.asp\">asset class<\/a> is <a href=\"https:\/\/www.investopedia.com\/terms\/f\/foreign-exchange.asp\">foreign exchange<\/a> (FX), the platform can be used to trade equities, <a href=\"https:\/\/www.investopedia.com\/terms\/e\/equitymarket.asp\">equity indices<\/a>, commodities and Bitcoin\u00a0using <a href=\"https:\/\/www.investopedia.com\/terms\/c\/contractfordifferences.asp\">CFDs<\/a>. Other benefits of using MT4 as opposed to other platforms include being easy to learn, has numerous available FX data sources and it\u2019s free.<\/p><p>Unfortunately, MT4 does not allow for direct trading in stock and <a href=\"https:\/\/www.investopedia.com\/terms\/f\/futuresmarket.asp\">futures markets<\/a> and conducting <a href=\"https:\/\/www.investopedia.com\/terms\/a\/anova.asp\">statistical analysis<\/a> can be burdensome; however, MS Excel can be used as a supplementary statistical tool.<\/p><h2>Algorithmic Trading Strategies<\/h2><p>It is important to begin by reflecting on some core traits that every algorithmic <a class=\"termLink\" data-inlink=\"link_23\" href=\"https:\/\/www.investopedia.com\/terms\/t\/trading-strategy.asp\">trading strategy<\/a> should have. The strategy should be market prudent in that it is fundamentally sound from a market and economic standpoint. Also, the mathematical model used in developing the strategy should be based on <a href=\"https:\/\/www.investopedia.com\/articles\/financial-theory\/10\/gaussian-models-statistics.asp\">sound statistical methods<\/a>.<\/p><p>Next, it is crucial to determine what information your robot is aiming to capture. In order to have an automated strategy, your robot needs to be able to capture identifiable, persistent <a href=\"https:\/\/www.investopedia.com\/terms\/i\/inefficientmarket.asp\">market inefficiencies<\/a>. Algorithmic trading strategies follow a rigid set of rules that take advantage of market behavior and thus, the occurrence of a one-time market inefficiency is not enough to build a strategy around. Further, if the cause of the market inefficiency is unidentifiable, then there will be no way to know if the success or failure of the strategy was due to chance or not.<\/p><p>With the above in mind, there are a number of strategy types to inform the design of your algorithmic trading robot. These include strategies that take advantage of the following (or any combination thereof):<\/p><ul><li>Macroeconomic news (e.g. non-farm payroll or interest rate changes)<\/li>\n<li><a href=\"https:\/\/www.investopedia.com\/terms\/f\/fundamentalanalysis.asp\">Fundamental analysis<\/a> (e.g. using revenue data or earnings release notes)<\/li>\n<li>Statisticalanalysis (e.g. correlation or co-integration)<\/li>\n<li><a href=\"https:\/\/www.investopedia.com\/terms\/t\/technicalanalysis.asp\">Technical analysis<\/a> (e.g. <a href=\"https:\/\/www.investopedia.com\/terms\/m\/movingaverage.asp\">moving averages<\/a>)<\/li>\n<li>The market microstructure (e.g. arbitrage or trade infrastructure)<\/li>\n<\/ul><h2>Designing and Testing Your Robot<\/h2><p>There are essentially four steps needed to build and manage a trading robot:<\/p><h3>Preliminary Research<\/h3><p>This step focuses on developing a strategy that suits your own personal characteristics. Factors such as <a href=\"https:\/\/www.investopedia.com\/terms\/r\/risktolerance.asp\">personal risk profile<\/a>, time commitment and <a class=\"termLink\" data-inlink=\"link_36\" href=\"https:\/\/www.investopedia.com\/terms\/c\/capital.asp\">trading capital<\/a> are all important to think about when developing a strategy. You can then begin to identify the persistent market inefficiencies mentioned above. Having identified a market inefficiency you can begin to code a trading robot suited to your own personal characteristics.<\/p><h3>Backtesting<\/h3><p><a href=\"https:\/\/www.investopedia.com\/terms\/b\/backtesting.asp\">This step<\/a>\u00a0focuses on validating your trading robot. This includes checking the code to make sure it is doing what you want and understanding how it performs over different time frames, asset classes, or different market conditions, especially in <a href=\"https:\/\/www.investopedia.com\/terms\/b\/blackswan.asp\">black swan<\/a>\u00a0type events such as the <a class=\"termLink\" data-inlink=\"link_39\" href=\"https:\/\/www.investopedia.com\/articles\/economics\/09\/financial-crisis-review.asp\">2008 global financial crisis<\/a>.<\/p><h3>Optimization<\/h3><p>Now that you have coded a robot that works and at this stage, you want to maximize its performance while minimizing the\u00a0<a href=\"https:\/\/www.investopedia.com\/terms\/o\/overfitting.asp\">overfitting<\/a> bias. To maximize performance you first need to select a good performance measure that captures risk and reward elements, as well as consistency (e.g. <a href=\"https:\/\/www.investopedia.com\/terms\/s\/sharperatio.asp\">Sharpe ratio<\/a>). An overfitting bias occurs when your robot is too closely based on past data; such a robot will give off the illusion of high performance, but since the future never completely resembles the past, it may actually fail.<\/p><h3>Live Execution<\/h3><p>You are now ready to begin using real money. However, aside from being prepared for the emotional ups and downs that you might experience, there are a few technical issues that need to be addressed. These issues include selecting an appropriate <a class=\"termLink\" data-inlink=\"link_42\" href=\"https:\/\/www.investopedia.com\/terms\/b\/broker.asp\">broker<\/a> and implementing mechanisms to manage both <a class=\"termLink\" data-inlink=\"link_43\" href=\"https:\/\/www.investopedia.com\/terms\/m\/marketrisk.asp\">market risks<\/a> and <a class=\"termLink\" data-inlink=\"link_44\" href=\"https:\/\/www.investopedia.com\/terms\/o\/operational_risk.asp\">operational risks<\/a> such as potential hackers and technology downtime.<\/p><p>It is also important at this step to verify that the robot\u2019s performance is similar to that experienced in the testing stage. Finally, continual monitoring is needed to ensure that the market efficiency that the robot was designed for still exists.\u00a0<\/p><h2>The Bottom Line<\/h2><p>Considering that Richard Dennis, the <a href=\"https:\/\/www.investopedia.com\/articles\/trading\/08\/turtle-trading.asp\">legendary<\/a> commodity trader, taught a group of students his personal trading strategies who then went on to earn over $175 million in just five years, it is completely possible for inexperienced traders to be taught a strict set of guidelines and <a class=\"termLink\" data-inlink=\"link_48\" href=\"https:\/\/www.investopedia.com\/articles\/trading\/10\/top-ten-rules-for-trading.asp\">become successful traders<\/a>. However, this is one extraordinary example and beginners should definitely remember to have modest expectations.<\/p><p>In order to be successful, it is important to not just follow a set of guidelines but to understand how those guidelines are working. Liew stresses that the most important part of algorithmic trading is \u201cunderstanding under which types of market conditions your robot will work and when it will break down,\u201d and \u201cunderstanding when to intervene.\u201d Algorithmic trading can be rewarding but the key to success is understanding. Any course or teacher promising high rewards with minimal understanding should be a major warning sign.<\/p>","created_at":"2018-11-07T16:43:18+0000","updated_at":"2018-11-07T16:43:18+0000","published_at":"2015-08-13T05:33:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.investopedia.com","preview_picture":"https:\/\/i.investopedia.com\/dimages\/graphics\/shutterstock_268518818.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":281,"title":"Notes for new Make users","url":"http:\/\/gromnitsky.users.sourceforge.net\/articles\/notes-for-new-make-users\/","content":"Notes for new Make users<header><p><img src=\"http:\/\/gromnitsky.users.sourceforge.net\/articles\/notes-for-new-make-users\/man-on-horse.svg\" class=\"man-on-horse\" alt=\"\" \/><\/p><p>John Constable would have said that \u2018There is nothing ugly in Make; I never saw an ugly build system in my life: for let the number of targets be what it may,\u2013dependencies, recipes, and parallel jobs will always make it manageable.\u2019<\/p>\n<p>It\u2019s funny to see a sudden spike of interest in Make-like tools on <a href=\"https:\/\/news.ycombinator.com\/item?id=16483889\">HN<\/a> &amp; <a href=\"https:\/\/www.reddit.com\/r\/programming\/comments\/80x5z6\/\">reddit<\/a>, when many start joyfully sharing their favourite tricks. I predict some ppl will get overexcited &amp; start rewriting their existing build infrastructure to be more \u201cMake friendly\u201d only to realize it\u2019s not that straightforward as it sounds.<\/p>\n<p>The discussion about Make ordinaly ends up w\/ ideas how Make can be improved. Nevertheless, instead of <em>improving<\/em> Make, avid &amp; enthusiastic chaps invariably decide to start from ground zero, producing something unequivocally awesome, but incompatible w\/ everything that was written before.<\/p>\n<p>If you say Make is suboptimal, no one sane will disagree w\/ you. However, it\u2019s a solid, well known tool that (if used correctly, w\/o wonky religious zeal) works like the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Beverly_Clock\">Beverly Clock<\/a>.<\/p>\n<p>If you really curious about (GNU) Make, read on. I\u2019m not going to evangelize Make, I assume you\u2019ve already decided to try it.<\/p>\n<h2 id=\"4b6d995-dont-try-to-be-clever\">Don\u2019t try to be clever<\/h2>\n<p>Regardless of how you like your makefiles now, you\u2019ll cringe at them in 6 months &amp; laugh at them in 1 year. Your style will evolve.<\/p>\n<h2 id=\"4b6d995-the-official-docs\">The official docs<\/h2>\n<p>When in doubt, don\u2019t google until you read the <a href=\"https:\/\/www.gnu.org\/software\/make\/manual\/\">official manual<\/a> <del>from cover to cover<\/del>. It\u2019s a document of an astonishing quality. Go print the pdf.<\/p>\n<h2 id=\"4b6d995-adhere-to-the-common-terminology\">Adhere to the common terminology<\/h2>\n<pre>target: prerequisite(s)\n\trecipe<\/pre>\n<p>so in<\/p>\n<pre>out\/bundle.js: main.js foo.js bar.js\n\tmkdir $(dir $@)\n\tbrowserify main.js -o $@<\/pre>\n<p><code>out\/bundle.js<\/code> is a <em>target<\/em> that has 3 <em>prerequisites<\/em> (dependencies): <code>main.js<\/code>, <code>foo.js<\/code>, <code>bar.js<\/code>. The set of lines prefixed w\/ the TAB char is a <em>recipe<\/em>.<\/p>\n<p>Rules may have empty recipes:<\/p>\n<pre>deploy: rsync\nrsync: compile\n\trsync -a out\/ user@host:\/somewhere\/\ncompile: out\/bundle.js<\/pre>\n<p>The 1st rule that Make stumbles upon is called a <em>default goal<\/em>. Here, the default goal is <code>deploy<\/code>. You can override the default goal by passing the name of a desired target as an argument to the Make cmd:<\/p>\n<pre>$ make compile<\/pre>\n<p><em>Macros<\/em> serve a 2fold purpose: as variables whose values you can override from the command line, &amp; as a mechanism for writing custom functions. I won\u2019t talk about the latter much, but here\u2019s an example of the former, which you may find moderately amusing to play w\/:<\/p>\n<pre>date = $(shell date -d $(today) +%Y-%m-%d)\ntoday = now\ncomic: $(date).gif\n\txv $&lt;\n%.html:\n\twget -q http:\/\/dilbert.com\/strip\/$* -O $@\n%.gif: %.html\n\tnokogiri -e 'p $$_.css(\".comic-item-container\").first[\"data-image\"]' $&lt; | xargs wget -q -O $@<\/pre>\n<p>If you save it in the file named <code>Makefile<\/code>, then type <code>make<\/code> in the same directory, Make downloads the current Dilbert strip page, parses the html, downloads the .gif image &amp; displays it. If you run <code>make<\/code> again, it won\u2019t re-download or re-parse anything, but will display the .gif image right away.<\/p>\n<p>If you run<\/p>\n<pre>$ make today=2002-04-10\nwget -q http:\/\/dilbert.com\/strip\/2002-04-10 -O 2002-04-10.html\nnokogiri -e 'p $_.css(\".comic-item-container\").first[\"data-image\"]' 2002-04-10.html | xargs wget -q -O 2002-04-10.gif\nxv 2002-04-10.gif<\/pre>\n<p>the <code>today<\/code> macro gets overriden from its default <code>now<\/code> string to the supplied <code>2002-04-10<\/code> value. Notice how <code>date<\/code> macro recursively expands to get the properly formatted date string. It uses the internal <code>shell()<\/code> fn to get the output of the external <code>date<\/code> command.<\/p>\n<h2 id=\"4b6d995-most-important-concept-1-dag\">Most important concept #1: DAG<\/h2>\n<blockquote>\n<blockquote>\n<p>\u2018The <em>make<\/em> utility shall update files that are derived from other files.\u2019<br \/>\u2014 The Open Group Base Specifications Issue 7<\/p>\n<\/blockquote>\n<\/blockquote>\n<p>The 1st thing you need to understand about Make is that all you do is contruct <a href=\"https:\/\/en.wikipedia.org\/wiki\/Directed_acyclic_graph\">DAG<\/a>. You either do it by hand using simple explicit rules (I\u2019m skipping the recipes for brevity):<\/p>\n<pre>foo.js.min: foo.es5\nfoo.es5: foo.js<\/pre>\n<p>which means a straightforward <code>foo.min.js<\/code> \u2192 <code>foo.es5<\/code> \u2192 <code>foo.js<\/code> graph (where an arrow means <em>depends on<\/em>) or you use special <em>pattern rules<\/em> (smtms also called <em>metarules<\/em>) w\/ which Make can <em>generate<\/em> DAG vertices (nodes) for you automatically:<\/p>\n<pre>%.js.min: %.es5\n%.es5: %.js<\/pre>\n<p>In both cases, when you type <code>make foo.min.js<\/code> &amp; you indeed have <code>foo.js<\/code> file\u2013Make creates a proper dependency graph, checks that it doesn\u2019t contain cycles, takes <code>foo.min.js<\/code> as it\u2019s goal, sees that such a vertex has an out-degree number of 1 (i.e., it has exactly 1 <em>prerequisite<\/em>\u2013<code>foo.es5<\/code>), recursively jumps from vertex to vertex until it hits a vertex w\/ 0 prereqs (in compsci speak: w\/ an out-degree number of 0). Then it creates the lonely leaf &amp; unwinds itself until it reaches the orig goal. If anything goes wrong along the way, Make stops.<\/p>\n<p>In the case w\/ the pattern rules, Make cannot create the full DAG until it has all the vertices, hence it automatically searches for a so called <em>implicit rule<\/em> for file <code>foo.js<\/code>, finds a perfect match in <code>%.es5: %.js<\/code>, then looks for a rule for <code>foo.es5<\/code> &amp; so on.<\/p>\n<p>This is all you need to know. When Make looks at a single graph endpoint, it runs a recipe for a target (<em>remakes<\/em> it) iff its prereqs are \u201cnewer\u201d. This is how it avoids unnecessary work &amp; as any young lad should know, the only way to improve performance is not to do something faster but not do it at all.<\/p>\n<h2 id=\"4b6d995-most-important-concept-2-2-phases\">Most important concept #2: 2 phases<\/h2>\n<p>The 2nd thing you need to understand is how Make reads makefiles. It does it in 2 phases:<\/p>\n<ol type=\"I\"><li>Creates a DAG.<\/li>\n<li>Invokes the recipes during which it resolves the required macros.<\/li>\n<\/ol><p>E.g., after the phase I, a makefile<\/p>\n<pre>opts = -s inline\nbundle = foo\n$(bundle).es5: $(bundle).js\n\tbabel $(opts) $&lt; -o $@<\/pre>\n<p>to Make looks like<\/p>\n<pre>opts = ?\nbundle = ?\nfoo.es5: foo.js\n\t?<\/pre>\n<p>Make was forced to expand <code>bundle<\/code> macro in the rule def line, but it did nothing for <code>opts<\/code> in the recipe.<\/p>\n<p>If you forget about the 2 phases, it could lead to a misunderstanding when a rule creates some file that you expect Make to pick up later on.<\/p>\n<h2 id=\"4b6d995-the-order-of-the-garter\">The order <span style=\"text-decoration: line-through;\">of the garter<\/span><\/h2>\n<p>All macros &amp; rules are hoisted so the order it which they appear in makefiles is irrelevant, but Make has also ordinal variables (called <em>simply expanded variables<\/em>, SEVs or plainly <em>variables<\/em>):<\/p>\n<pre>src := foo.js bar.js<\/pre>\n<p>(note <code>:=<\/code> instead of <code>=<\/code>) that get expanded exactly 1 time during the phase I only. Hence this will work as expected:<\/p>\n<pre>bundle.deps = $(js) vendor\/baz.js\njs = foo.js bar.js<\/pre>\n<p>but this won\u2019t:<\/p>\n<pre>bundle.deps := $(js) vendor\/baz.js\njs := foo.js bar.js<\/pre>\n<p>here, <code>bundle.deps<\/code> will contain only <code>vendor\/baz.js<\/code> string. Remember that if you use SEVs, their order does matter as well as the order of rules\/macros that employ variables.<\/p>\n<p>To reiterate again: macros get expanded every time they are accessed. This could be during the phase I if a macro is referenced in a target\/prereq portion of a rule, or it can be expanded in the time of the phase II, when the DAG is ready and Make can invoke the recipes.<\/p>\n<p>Variables, on the other hand, get expanded immediately in the course of the phase I.<\/p>\n<pre>times_expanded :=\nmacro = $(shell date; sleep 2; $(eval times_expanded += 1))\nvar := $(macro)\nq:\n\t@echo var = $(var)\n\t@echo var = $(var)\n\t@echo macro = $(macro)\n\t@echo macro = $(macro)\n\t@echo times_expanded = $(times_expanded)<\/pre>\n<p>The macro that I conveniently named <code>macro<\/code> should be expanded precisely 3 times:<\/p>\n<pre>$ make\nvar = Tue Mar 6 15:45:57 EET 2018\nvar = Tue Mar 6 15:45:57 EET 2018\nmacro = Tue Mar 6 15:45:59 EET 2018\nmacro = Tue Mar 6 15:46:01 EET 2018\ntimes_expanded = 1 1 1<\/pre>\n<p>Notice how the output for the \u2018macro =\u2019 lines contains a different time, but for the \u2018var\u2019 lines it\u2019s always the same.<\/p>\n<p><code>$(eval ...)<\/code> in this example does exactly what you think: it combines a parser and an evaluator of the Make language.<\/p>\n<h2 id=\"4b6d995-automatic-variables\">Automatic variables<\/h2>\n<p>If you write metarules you can\u2019t do it w\/ the <em>automatic variables<\/em>, for in a rule<\/p>\n<pre>out\/.cache\/%.js: %.mjs\n\tbabel $&lt; -o $@<\/pre>\n<p>it\u2019s impossible to know beforehand the actual names of the files. The popular clamour is that the autovars are too short &amp; confusing. Unless you like to write magic makefiles (e.g., w\/ <code>.SECONDEXPANSION<\/code> nifty tricks), you\u2019ll need to remember only 3 autovar types:<\/p>\n<ul><li><code>$@<\/code> == target<\/li>\n<li><code>$&lt;<\/code> == the 1st prereq<\/li>\n<li><code>$^<\/code> == all prereqs<\/li>\n<\/ul><p><img src=\"http:\/\/gromnitsky.users.sourceforge.net\/articles\/notes-for-new-make-users\/autovars.svg\" alt=\"\" \/><\/p>\n<p>Are they really so hard to grasp?<\/p>\n<p>HN user <em>dahart<\/em> came up with the following <a href=\"https:\/\/news.ycombinator.com\/item?id=16525293\">visual mnemonics<\/a>:<\/p>\n<blockquote>\n<p>\u201c<code>$@<\/code> looks like a target, a ring with a bullseye. <code>$&lt;<\/code> is pointing left, so it\u2019s the first prereq. <code>$^<\/code> is like a horizontal bracket that groups all prereqs.\u201d<\/p>\n<\/blockquote>\n<p>Make comes w\/ a set of internal functions for text processing. Most of them are pure, idempotent &amp; don\u2019t do any IO. E.g., <code>dir<\/code> fn<\/p>\n<pre>$(dir lib\/foo.js vendor\/bar.js)<\/pre>\n<p>treats its argument as a string, returning <code>lib\/ vendor\/<\/code> regardless of whether <code>lib\/foo.js<\/code> or <code>vendor\/bar.js<\/code> exist.<\/p>\n<p>Some functions (<code>patsubst()<\/code>, <code>filter()<\/code>, <code>filter-out()<\/code>) support a tiny DSL: strings that use <code>%<\/code> char as a wildcard. If you don\u2019t fully get <a href=\"http:\/\/gromnitsky.blogspot.com\/2016\/10\/how-gnu-makes-patsubst-function-really.html\">the purpose of the <code>%<\/code><\/a>, the results are often confusing.<\/p>\n<p>The 2 most important fn that do IO are:<\/p>\n<ul><li><code>wildcard()<\/code> that in Linux\/Mac\/BSD uses glob(3) under the hood &amp; <a href=\"https:\/\/msdn.microsoft.com\/en-us\/library\/windows\/desktop\/aa364418(v=vs.85).aspx\">smthg else<\/a> in Windows;<\/li>\n<li><p><code>shell()<\/code> that is the equivalent of the bash command substitution (<code>`command`<\/code>). Only in reality, Make\u2019s<\/p>\n<pre>$(shell find . -name \\*.js)<\/pre>\n<p>is more like typing<\/p>\n<pre>$ \/bin\/sh -c 'find . -name \\*.js' | tr \\\\n ' '<\/pre>\n<p>in bash; i.e., after explicitly running the cmd under the particular shell, it replaces every newline w\/ a space.<\/p>\n<p>On Windows it obviously cannot assume that bash is installed, thence is uses the horrible <code>cmd.exe<\/code>.<\/p><\/li>\n<\/ul><p>Because of this historical Unix vs.\u00a0Windows diff, the pattern lang in <code>windcard()<\/code> &amp; the command syntax in <code>shell()<\/code> are both inherently non-portable, unfortunately.<\/p>\n<h2 id=\"4b6d995-why-doesnt-make-support-file-names-w-spaces\">Why doesn\u2019t Make support file names w\/ spaces?<\/h2>\n<p>Initially it was a cunning plan to remove the necessity of having a support for a special list data type. Perhaps <a href=\"https:\/\/en.wikipedia.org\/wiki\/Stuart_Feldman\">Stuart Feldman<\/a> could have used <code>,<\/code> instead of a space for a delimiter, but that ship has long sailed. Also recall that many Unix utils that deal w\/ file names return the list of file names join()\u2019ed by a space or a newline. Thus the choice of a space char was very natural.<\/p>\n<p>As Make doesn\u2019t have any concept of lists, rules get their prereqs as a string: after expanding all macros, a rule chops the string into pieces that all together look like an array of targets, where each target may or may not be a file name.<\/p>\n<p>In JS it would have looked like:<\/p>\n<pre>&gt; src = 'foo.js bar.js'\n'foo.js bar.js'\n&gt; ` ${src} baz.css`.trim().split(\/\\s+\/)\n[ 'foo.js', 'bar.js', 'baz.css' ]<\/pre>\n<p>The same goes for function arguments.<\/p>\n<h2 id=\"4b6d995-compile-everything-to-1-directory\">Compile everything to 1 directory<\/h2>\n<p>You may say it\u2019s a matter of style, but I don\u2019t like seeing the results of compilation scattered along the src files. I consider it a common antipattern. It\u2019s popular, for makefiles that produce such output are the easiest to write, especially for novices. Partially this is also Make\u2019s fault, for its collection of built-in metarules (observable via <code>make -p<\/code>) has schooled ppl to write similar rules for their tools.<\/p>\n<p>To remove the compilation results, ppl write <code>clean<\/code> targets, which are usually updated w\/ the same consistency as comments in the code.<\/p>\n<p>Don\u2019t write <code>clean<\/code> or <code>nuke<\/code> targets. If your output goes under 1 umbrella dir, then all you need to do for starting from the clean slate is to remove 1 directory.<\/p>\n<p>If you\u2019re writing an SPA, copy your static assets &amp; the relevant files from <code>node_modules<\/code> to the output dir too. A good makefile leaves a ready to deploy directory that doesn\u2019t depend on files left in the src dir.<\/p>\n<p>E.g., say the output directory is named <code>_out<\/code>. Our src tree:<\/p>\n<pre>.\n\u251c\u2500\u2500 src\n\u2502\u00a0\u00a0 \u251c\u2500\u2500 a.js\n\u2502\u00a0\u00a0 \u251c\u2500\u2500 b.js\n\u2502\u00a0\u00a0 \u251c\u2500\u2500 index.html\n\u2502\u00a0\u00a0 \u251c\u2500\u2500 style.css\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 main.js\n\u2514\u2500\u2500 Makefile<\/pre>\n<p>We need to transpile multiple <code>src\/*.js<\/code> files before combining them to a bundle.<\/p>\n<pre>$ cat Makefile\nout := _out\ncache := $(out)\/.cache\nbuild := $(out)\/development\nmkdir = @mkdir -p $(dir $@)\nall: $(build)\/main.js\njs.src := $(wildcard src\/*.js)\njs.dest := $(addprefix $(cache)\/, $(js.src))\n$(build)\/main.js: $(js.dest)\n\t$(mkdir)\n\tbrowserify $(cache)\/src\/main.js -o $@\n$(cache)\/%.js: %.js\n\t$(mkdir)\n\tbabel $&lt; -o $@<\/pre>\n<p>Run Make, &amp; it creates the umbrella dir &amp; populates it w\/ the compilation results:<\/p>\n<pre>$ make\nbabel src\/a.js -o _out\/.cache\/src\/a.js\nbabel src\/b.js -o _out\/.cache\/src\/b.js\nbabel src\/main.js -o _out\/.cache\/src\/main.js\nbrowserify _out\/.cache\/src\/main.js -o _out\/development\/main.js\n$ tree --noreport -a _out\n_out\/\n\u251c\u2500\u2500 .cache\n\u2502\u00a0\u00a0 \u2514\u2500\u2500 src\n\u2502\u00a0\u00a0     \u251c\u2500\u2500 a.js\n\u2502\u00a0\u00a0     \u251c\u2500\u2500 b.js\n\u2502\u00a0\u00a0     \u2514\u2500\u2500 main.js\n\u2514\u2500\u2500 development\n    \u2514\u2500\u2500 main.js<\/pre>\n<p>Of course to be able to test <code>_out\/development<\/code> in the browser we need to copy the static assets to the umbrella dir too. Adding this to the makefile<\/p>\n<pre>$(build)\/%: src\/%\n\t$(mkdir)\n\tcp $&lt; $@\nstatic.src := $(wildcard src\/*.css src\/*.html)\nstatic.dest := $(patsubst src\/%, $(build)\/%, $(static.src))\nall: $(static.dest)<\/pre>\n<p>\u2026 accomplishes our goal:<\/p>\n<pre>$ make\ncp src\/style.css _out\/development\/style.css\ncp src\/index.html _out\/development\/index.html<\/pre>\n<h2 id=\"4b6d995-multiple-builds\">Multiple builds<\/h2>\n<p>It\u2019s easy to add support for several builds from a single src directory. Make supports <em>conditional directives<\/em>, w\/ which you may alter parameters to transpilers, change the output dir, etc. E.g., the enhanced version of the makefile from the prev section:<\/p>\n<pre>NODE_ENV ?= development\nout := _out\ncache := $(out)\/.cache.$(NODE_ENV)\nbuild := $(out)\/$(NODE_ENV)\nbabel.opts := -s inline\nbrowserify.opts := -d\nifeq ($(NODE_ENV), production)\nbabel.opts := --minified\nbrowserify.opts :=\nendif\nmkdir = @mkdir -p $(dir $@)\nall: $(build)\/main.js\njs.src := $(wildcard src\/*.js)\njs.dest := $(addprefix $(cache)\/, $(js.src))\n$(build)\/main.js: $(js.dest)\n\t$(mkdir)\n\tbrowserify $(browserify.opts) $(cache)\/src\/main.js -o $@\n$(cache)\/%.js: %.js\n\t$(mkdir)\n\tbabel $(babel.opts) $&lt; -o $@<\/pre>\n<p>Conditional directives are evaluated during the phase I. There\u2019s also <code>if()<\/code> fn that can be used in macros, but there\u2019s no <code>eq()<\/code> fn (it exists in the Make\u2019s src code, but under \u2018experimental\u2019 flag), that limits the applicability of <code>if()<\/code>. You can play w\/ <code>filter()<\/code> inside <code>if()<\/code> (for <code>if()<\/code> treats a 0-length string or anything that expands to such a string as <em>false<\/em>) but it quickly gets unreadable.<\/p>\n<h2 id=\"4b6d995-chain-of-rules\">Chain of rules<\/h2>\n<p>There are 2 categories of ppl:<\/p>\n<ol type=\"1\"><li><p>those who write multiple rules, that work in a chain: e.g., first we transpile js, second we minify the transpiled output;<\/p>\n<pre>%.js.min: %.es5\n%.es5: %.js<\/pre><\/li>\n<li><p>those who write a single rule .min.js \u2192 js, doing all steps in 1 recipe; there are 2 subcategories of such ppl as well:<\/p>\n<ol type=\"a\"><li><p>those who create tmp files in the recipe;<\/p><\/li>\n<li><p>those who use pipes, avoiding tmp files whatsoever; unfortunately I cannot recoment this method, for \/bin\/sh doesn\u2019t signal an error if any of the cmds in a pipeline fail, except for the last one; with this approach you may (&amp; will) easily end up w\/ garbage in output or 0-length files.<\/p><\/li>\n<\/ol><\/li>\n<\/ol><p>When deciding what target to build next, Make is capable of recognising which targets are temporary.<\/p>\n<pre>foo.js.min: foo.es5\nfoo.es5: foo.js<\/pre>\n<p>Here, <code>foo.es5<\/code> is an <em>intermediary<\/em>, but as it\u2019s explicitly mentioned in makefile, it won\u2019t be recognized as such.<\/p>\n<p>The makefile below (taken from a simple <a href=\"https:\/\/github.com\/gromnitsky\/shopping-hours\">shopping-hours<\/a> program) uses the sequence of implicit metarules to create 2 UMD bundles: a minimized es5 &amp; a usual js parcel:<\/p>\n<pre>out := dist\nmkdir = @mkdir -p $(dir $@)\nbundle.name := $(out)\/shopping_hours\ncompile: $(bundle.name).min.js\n$(out)\/%.min.js: $(out)\/%.es5.js\n\tuglifyjs $&lt; -o $@ -mc\n$(out)\/%.es5.js: $(out)\/%.js\n\tbabel --presets `npm -g root`\/babel-preset-es2015 $&lt; -o $@\n$(bundle.name).js: index.js\n\t$(mkdir)\n\tbrowserify -s $(basename $(notdir $@)) $&lt; -o $@<\/pre>\n<pre>$ make\nbrowserify -s shopping_hours index.js -o dist\/shopping_hours.js\nbabel --presets `npm -g root`\/babel-preset-es2015 dist\/shopping_hours.js -o dist\/shopping_hours.es5.js\nuglifyjs dist\/shopping_hours.es5.js -o dist\/shopping_hours.min.js -mc\nrm dist\/shopping_hours.es5.js<\/pre>\n<p>Notice the last line (the rm command). There\u2019s nowhere such a line could be found in the makefile! Make has automatically deduced that <code>dist\/shopping_hours.es5.js<\/code> vertex is temporal &amp; auto removed it before exiting.<\/p>\n<h2 id=\"4b6d995-write-makefiles-in-shell-scripts-stead\">Write makefiles in shell scripts stead<\/h2>\n<p>If you already have a bunch of small .sh files\u2013move them to 1 makefile in the form of 1 .sh file == 1 target + recipe. By doint this you\u2019ll get for free:<\/p>\n<ol type=\"a\"><li>the dependency management (what target to run first);<\/li>\n<li>command line args processing (foo=bar args).<\/li>\n<\/ol><p>Make can run several recipes at once. Recall the Dilbert makefile. To download all the comics (starting from April 16, 1989) just generate the corresponding dates &amp; pass them as targets. There is no need to modify the makefile itself.<\/p>\n<p>Generate the target names:<\/p>\n<pre>$ seq `date -d 1989-04-16 +%s` $((60*60*24)) `date +%s` | xargs -Isec date -d @sec +%Y-%m-%d.gif &gt; targets.txt\n$ head -3 !$\n1989-04-16.gif\n1989-04-17.gif\n1989-04-18.gif<\/pre>\n<p>Run Make w\/ 50 parallel jobs (on a 2nd thought, don\u2019t type that):<\/p>\n<pre>$ make -j50 `cat targets.txt`<\/pre>\n<p><img src=\"http:\/\/gromnitsky.users.sourceforge.net\/articles\/notes-for-new-make-users\/nyan-cat.svg\" alt=\"\" \/><\/p>\n<p>(I\u2019ve got banned pretty quickly.)<\/p>\n<p>The beauty of the approach is that you may press Ctrl-C any time &amp; when you run Make again it won\u2019t re-download already processed pages.<\/p>\n<p>You can even mask the makefile for a standalone script: add a proper shebang line to the aforementioned Dilbert makefile:<\/p>\n<pre>$ printf '%s\\n\\n' '#!\/usr\/bin\/make -f' | cat - Makefile &gt; dilbert\n$ chmod +x !$<\/pre>\n<p>Now users can run <code>dilbert today=1999-02-21<\/code> w\/o ever suspecting they are using Make.<\/p>\n<h2 id=\"4b6d995-tabs-shell\">Tabs &amp; shell<\/h2>\n<blockquote>\n<blockquote>\n<p>\u2018Why the tab in column 1? Yacc was new, Lex was brand new. I hadn\u2019t tried either, so I figured this would be a good excuse to learn. After getting myself snarled up with my first stab at Lex, I just did something simple with the pattern newline-tab. It worked, it stayed. And then a few weeks later I had a user population of about a dozen, most of them friends, and I didn\u2019t want to screw up my embedded base. The rest, sadly, is history.\u2019<br \/>\u2014 Stuart Feldman<\/p>\n<\/blockquote>\n<\/blockquote>\n<p>First, fix your editor, any decent one can highlight tabs. Second, if you\u2019re still against tabs, redefine <code>.RECIPEPREFIX<\/code> variable:<\/p>\n<pre>space :=\nspace +=\n.RECIPEPREFIX := $(space)<\/pre>\n<p>w\/ which you can use spaces before any recipe line.<\/p>\n<p>If you\u2019re feeling mischievous, Make has divers \u2018secret\u2019 vars &amp; targets for you. For instance, you don\u2019t have to \u2018struggle\u2019 w\/ the ancient bash, for it\u2019s possible to program recipes in any language that supports evaluations straight from the command line. This\u2019ll help your prank get started:<\/p>\n<pre>$ cat Makefile\nSHELL := node\n.SHELLFLAGS := -e\ndate != console.log(new Date())\nq:\n\t@console.log('today is $(date)')<\/pre>\n<pre>$ make\ntoday is 2018-03-03T10:02:05.789Z<\/pre>\n<p><code>!=<\/code> is a syntactic sugar for the <code>shell()<\/code> fn.<\/p>\n<p>Think that running every recipe line in a sub-shell is \u2018expensive\u2019? Make\u2019s got you covered! Mentioning <code>.ONESHELL<\/code> <em>target<\/em> in a makefile sends all lines in a recipe to the sub-shell in bulk. If you resolve to profit by this circumstance, don\u2019t complain about the consequences of inability to auto-detect errors in the middle of recipes.<\/p>\n<p><img src=\"http:\/\/gromnitsky.users.sourceforge.net\/articles\/notes-for-new-make-users\/cans.svg\" alt=\"\" \/><\/p>\n<p>You can exploit macros as user-defined functions. Because a macro gets expanded every time it\u2019s being accesed, it\u2019s save to include autovars in it or refs to other macros. Perhaps, the 2 most common examples are:<\/p>\n<pre>copy = cp $&lt; $@\nmkdir = @mkdir -p $(dir $@)<\/pre>\n<p>as in<\/p>\n<pre>_build\/%.html: src\/%.html\n\t$(mkdir)\n\t$(copy)<\/pre>\n<p>Such macros are called <em>canned recipes<\/em>.<\/p>\n<p>If you need to pass a parameter to a macro, use a special <code>call()<\/code> fn to invoke it:<\/p>\n<pre>find = $(shell find . -name \\*.$1 -type f)\nsrc := $(call find,js)<\/pre>\n<p>Because <code>find<\/code> contains refs to params (<code>$1<\/code>), it\u2019s called a parametrised function.<\/p>\n<h2 id=\"4b6d995-deps\">Deps<\/h2>\n<p>Writing makefiles for small programs often means a manual dependency management. If some <code>out\/bundle.js<\/code> depends on several .js modules, it\u2019s not hard to specify its prereqs manualy, but for large programs it becomes unmanageable.<\/p>\n<p><a href=\"http:\/\/gromnitsky.blogspot.com\/2016\/07\/generating-dependencies-automatically.html\">There are several ways<\/a> to tackle the dependency problem, the most popular one is colloquially called Tromey\u2019s Way.<\/p>\n<p>I\u2019ve noticed that various (programming) books authors love to give examples that have little or no bearing w\/ the subject, via explaining how they\u2019ve solved some minuscule problem in the course of preparing their manuscript. This is just a ~29KB markdown file, but I feel that such a great tradition of \u2018here\u2019s an example from my book toolchain!\u2019 should be continued.<\/p>\n<p>Before handing down the .md file to pandoc, I preprocess it w\/ erb. This allows me to write<\/p>\n<pre>&lt;%= File.read 'dilbert\/Makefile' %&gt;<\/pre>\n<p>instead of a copy-pasting. But when I edit such a referenced makefile I want the .md file to be auto-recompiled. Manually specifying all included files as prereqs is lame, so I have a lilliputian script that reads the .md file &amp; prints:<\/p>\n<pre>_out\/web\/index.html: dilbert\/Makefile\n_out\/web\/index.html: umbrella\/ver1.mk\n...<\/pre>\n<p>I inject those lines into a makefile (via Make\u2019s <code>include<\/code> directive). A simplified version of the whole scheme looks like:<\/p>\n<pre>define make-depend\n@mkdir -p $(cache)\n.\/deps $@ &lt; $&lt; &gt; $(cache)\/$*.d\nendef\n$(out)\/%.html: %.md\n\t$(mkdir)\n\terb $&lt; | pandoc -o $@\n\t$(make-depend)\n-include $(cache)\/index.d<\/pre>\n<p><code>define<\/code> \u2026 <code>endef<\/code> is a multiline macro. \u2018deps\u2019 is the script in question. The main trick here is to invoke the macro after the successful .md to .html transformation. If I edit the .md\u2013it doesn\u2019t matter if there are new <code>File.read<\/code> commands in it, for Make rebuilds the .md file regardless, but if I change any of <code>File.read<\/code>\u2019ed files &amp; then run Make, it picks up the proper prereq list &amp; sees that it ought to recompile the .md.<\/p>\n<h2 id=\"4b6d995-watchers\">Watchers<\/h2>\n<p>To auto-recompile on changes w\/o manualy invoking Make means using an external file \u201cwatcher\u201d. I have <a href=\"https:\/\/github.com\/gromnitsky\/watchthis\">my own little watcher<\/a> that plays different sounds depending on the return value of Make.<\/p>\n<pre># npm i -g watchthis<\/pre>\n<p>Then in the project dir:<\/p>\n<pre>$ watchthis -e _out make test<\/pre>\n<p>See, how easy the life is, when you compile everything to 1 directory (<code>_out<\/code>, in this example).<\/p>\n<p>If you feel it\u2019s time to be frugal with inotify watchers, <a href=\"https:\/\/stackoverflow.com\/a\/48496716\/81081\">try<\/a><\/p>\n<pre>$ while true; do make -q || make; sleep 2; done<\/pre>\n<p>The reason for <code>make -q || make<\/code>, instead of a simple <code>make<\/code>, is to prevent a terminal from filling up with annoying \u2018make: Nothing to be done for <em>target<\/em>\u2019 messages. The downside is that you compute the DAG twice for each \u2018succesful\u2019 run, not to mention a cpu waste from constant polling.<\/p>\n<h2 id=\"4b6d995-auto-restarting\">Auto-restarting<\/h2>\n<p>Make can auto-restart itself during a DAG construction if<\/p>\n<ol type=\"a\"><li>it reads another makefile (via <code>include<\/code> directive), &amp;<\/li>\n<li>it has found a rule where the target == included makefile name.<\/li>\n<\/ol><pre>include _out\/.npm\n_out\/.npm: package.json\n\tnpm i\n\ttouch $@<\/pre>\n<p>How is this useful? If you have targets w\/ prereqs from <code>node_modules<\/code> dir (e.g., a css library or already minified bundles) those files must exist before Make runs, otherwise Make throws an error. The auto-restarting facility allows Make to set forth with a clean slate and read the makefile anew, this time having all vertices in the DAG you\u2019ve specified.<\/p>\n<p>Files similar to <code>_out\/.npm<\/code> are called <em>empty targets<\/em>.<\/p>\n<h2 id=\"4b6d995-microscope\">Microscope<\/h2>\n<p>By default, all rules, macros &amp; plain variables are placed in the global scope. Autovars are the obvious exception, for they are local to a particular recipe.<\/p>\n<p>You can provide <em>target-specific<\/em> macros\/vars for a rule. Say we want to compile .js files with different options and to a different directory, depending on the value of <code>NODE_ENV<\/code> environment variable:<\/p>\n<pre>NODE_ENV ?= development\ncache := _out\/.cache\/$(NODE_ENV)\n$(cache)\/%.js: %.js\n\t@mkdir -p $(dir $@)\n\tbabel $(babel.opts) $&lt; -o $@\n$(NODE_ENV): $(addprefix $(cache)\/, $(wildcard *.js))\nbabel.opts := -s inline\nproduction: babel.opts := --minified<\/pre>\n<p>The name of the default rule in this example is equal to the value of the <code>NODE_ENV<\/code> var. If there\u2019s no <code>NODE_ENV<\/code> in the environment, we presume it\u2019s equal to \u2018development\u2019:<\/p>\n<pre>$ make\nbabel -s inline a.js -o _out\/.cache\/development\/a.js\nbabel -s inline b.js -o _out\/.cache\/development\/b.js<\/pre>\n<p>Nothing interesting is going on, until we override the said <code>NODE_ENV<\/code> var:<\/p>\n<pre>$ make NODE_ENV=production\nbabel --minified a.js -o _out\/.cache\/production\/a.js\nbabel --minified b.js -o _out\/.cache\/production\/b.js<\/pre>\n<p>We have 0 conditional directives in the makefile, but the value of <code>babel.opts<\/code> has somehow changed nevertheless. The following 2 lines:<\/p>\n<pre>$(NODE_ENV): $(addprefix $(cache)\/, $(wildcard *.js))\nproduction: babel.opts := --minified<\/pre>\n<p>after the phase I become:<\/p>\n<pre>production: _out\/.cache\/production\/a.js _out\/.cache\/production\/b.js\nproduction: babel.opts := --minified<\/pre>\n<p>The last line overrides the value of the global <code>babel.opts<\/code> variable, but <em>only<\/em> for the target named <code>production<\/code>. That target, in its turn, has 2 prereqs, each of which inherits the value of <code>babel.opts<\/code> from their incoming vertex (<code>production<\/code>). I.e., Make internally generates a rule for <code>_out\/.cache\/production\/a.js<\/code> that looks quite similar to:<\/p>\n<pre>_out\/.cache\/production\/a.js: a.js  babel.opts := --minified\n\t@mkdir -p $(dir $@)\n\tbabel $(babel.opts) $&lt; -o $@<\/pre>\n\n<h2 id=\"4b6d995-debugging\">Debugging<\/h2>\n<p>Many saints have suffered martyrdom. Many users of Make have learned to tolerate the absence of debugging facilities. There\u2019s <code>--trace<\/code> CLO but reading its output is tedious.<\/p>\n<p>You\u2019d expect Make to have an option \u2018print how you see this makefile after the phase I\u2019 (or in another words, \u2018show me how you have constructed the DAG\u2019), but no such feature exists in a polished state. You resort to <code>-pk<\/code> &amp; grep:<\/p>\n<pre>$ make -rR -pk -q | grep -v ^#<\/pre>\n<p>(<code>-rR<\/code> disables a load of Make\u2019s built-in implicit metarules\/vars for C\/C++). This also prints every env variable (for Make converts every env var into a macro (yes!), which is an interesting choice security-wise). We can write a little sh wrapper that reads all the env var names &amp; filters them out alongside w\/ often tangential build-in vars:<\/p>\n<pre>$ cat make-phase1\n#!\/bin\/sh\nunset TERMCAP\nre_env=$(env | cut -d= -f1 | tr \\\\n '|')\nre_dot_smthg='\\.[^ ]+'\nre_autovars='[%*+&lt;?^@][^ ]'\nre_misc='GNUMAKEFLAGS|MAKE([^ ]+)?|MFLAGS|SUFFIXES|-\\*-command-variables-\\*-'\nmake -rR -pk -q \"$@\" \\\n    | egrep -v \"^(${re_env}#|$re_dot_smthg|$re_autovars|$re_misc) \" \\\n    | cat -s<\/pre>\n<p>Then type \u2018make-phase1\u2019 &amp; pass any params to it as you would\u2019ve done w\/ \u2018make\u2019.<\/p>\n<p>Even if you\u2019re not <a href=\"https:\/\/tenderlovemaking.com\/2016\/02\/05\/i-am-a-puts-debuggerer.html\">a puts debugger<\/a>, Make forces you to become one.<\/p>\n<p>Everybody knows the trick how to print the value of the macro\/variable, right? Create <code>debug.mk<\/code> file somewhere not far away, e.g., in <code>~\/lib\/<\/code>:<\/p>\n<pre>p-%:\n\t@echo \"$(strip $($*))\" | tr ' ' \\\\n<\/pre>\n<p>Then type:<\/p>\n<pre>$ make -f ~\/lib\/debug.mk p-SHELL\n\/bin\/sh<\/pre>\n<p>Adding <code>-f makefile<\/code> however many times you want &amp; passing <code>p-macroName<\/code>, as a target, prints the expanded value of the macro.<\/p>\n<p>There\u2019s also a build-in <code>warning()<\/code> fn that expands its param &amp; prints it to the stderr. It could be inserted anywhere, for its actual expanded value is an empty string.<\/p>\n<p>Metarules could be frustrating if you\u2019re unsure how the matching is done against provided targets. In the absence of a REPL of any kind, fall back to a helper .mk again:<\/p>\n<pre>$ cat ~\/lib\/metarule.mk\nmatch: $(T)\n$(P):\n\t@echo '$$@ == $@'\n\t@echo '$$* == $*'<\/pre>\n<p>To work properly it shoult be run in some tmp dir:<\/p>\n<pre>$ (cd \/tmp; make -f ~\/lib\/metarule.mk P=out\/%.js T=out\/foo\/bar.js)\n$@ == out\/foo\/bar.js\n$* == foo\/bar<\/pre>\n<p><code>$*<\/code> is a stem autovar that matches the <code>%<\/code> portion of the metarule. <code>P<\/code> is for \u2018pattern\u2019, <code>T<\/code> is for \u2018target\u2019. You know that the target doesn\u2019t match when Make fails:<\/p>\n<pre>$ (cd \/tmp; make -f ~\/lib\/metarule.mk P=out\/%.js T=bar.js)\nmake: *** No rule to make target 'bar.js', needed by 'match'.  Stop.<\/pre>\n<p>Make has a \u2018dry run\u2019 CLO (<code>-n<\/code>) that just prints recipes instead of executing them. Combined w\/ <code>-t<\/code> option, you get a terse view of what targets Make is going to rebuild. E.g., dry run the final makefile from <a href=\"#4b6d995-compile-everything-to-1-directory\">Compile everything to 1 directory<\/a> section:<\/p>\n<pre>$ make -tn -f ver2.mk\ntouch _out\/.cache\/src\/a.js\ntouch _out\/.cache\/src\/b.js\ntouch _out\/.cache\/src\/main.js\ntouch _out\/development\/main.js\ntouch _out\/development\/style.css\ntouch _out\/development\/index.html<\/pre>\n<p>Beware that even w\/ <code>-n<\/code>, Make expands all vars &amp; required macros, so if any of them contain the <code>shell()<\/code> fn, Make executes it irrespective of the <code>-n<\/code> presence.<\/p>\n<h2 id=\"4b6d995-additional-reading\">Additional reading<\/h2>\n<p><img src=\"http:\/\/gromnitsky.users.sourceforge.net\/articles\/notes-for-new-make-users\/cromwellian-pikeman.svg\" class=\"cromwellian-pikeman\" alt=\"\" \/><\/p>\n<p>Apart from the official manual, a <a href=\"http:\/\/make.mad-scientist.net\/papers\/\">series of essays<\/a> from the current GNU Make maintainer (Paul Smith) are super informative. His profile at <a href=\"https:\/\/stackoverflow.com\/users\/939557\/madscientist\">stackoverflow<\/a>, strangely enough, it\u2019s mostly about Make.<\/p>\n<p>Eric Melski\u2019s blog posts:<\/p>\n<ul><li><a href=\"https:\/\/blog.melski.net\/2010\/11\/15\/shell-commands-in-gnu-make\/\">Shell commands in GNU make<\/a><\/li>\n<li><a href=\"https:\/\/blog.melski.net\/2015\/01\/07\/howto-intro-to-gnu-make-variables\/\">HOWTO: Intro to GNU make variables<\/a><\/li>\n<\/ul><p>John Graham-Cumming\u2019s <a href=\"https:\/\/www.amazon.com\/GNU-Make-Book-John-Graham-Cumming\/dp\/1593276494\">The GNU Make Book<\/a>.<\/p>\n<p>The original Stuart Feldman\u2019s paper (pdf, Bell Labs, 1978): <a href=\"http:\/\/ewald.cas.usf.edu\/teaching\/2004F\/5156\/packet\/packet2.docs\/05.make.pdf\">Make\u2013A Program for Maintaining Computer Programs<\/a><\/p>\n<p><a href=\"https:\/\/archive.org\/details\/pdfy-MgN0H1joIoDVoIC7\">The AWK programming language<\/a> book has a Make-like program that fits on <em>one<\/em> page! (Ch.7, p.178.) Whilst it supports only static rules (no metarules, macros or functions), it\u2019s easy to imagine how one could combine it w\/ a decent preprocessor.<\/p>\n<p><a href=\"http:\/\/www.conifersystems.com\/whitepapers\/gnu-make\/\">What\u2019s Wrong With GNU make?<\/a> \u201cpaper\u201d lists divers valid points, although in a somewhat disgruntled style.<\/p>\n<p><a href=\"http:\/\/ftp.qucis.queensu.ca\/home\/cordy\/Papers\/MCAA_ICPC15_Makefiles.pdf\">Make It Simple \u2013 An Empirical Analysis of GNU Make Feature Use in Open Source Projects<\/a> (2015) \u2013 a study that looks at the diff between handwritten vs.\u00a0generated makefiles.<\/p>\n<p>If you have a compsci degree, you are accustomed to graphs, their terminology et al., thus when writing makefiles you should feel at home. My degree was in materials science\/nanomaterials, therefore I\u2019ve found the <em>Ch 14 Graph Algorithms<\/em> of <a href=\"https:\/\/www.amazon.com\/Data-Structures-Algorithms-Java-6th-ebook\/dp\/B00JDRQF8C\">Data Structures and Algorithms in Java<\/a> very helpful.<\/p><\/header>","created_at":"2018-11-07T16:43:17+0000","updated_at":"2018-11-07T16:43:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":25,"domain_name":"gromnitsky.users.sourceforge.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":282,"title":"Alli AI \u00ab Artificial Intelligence for SEO","url":"https:\/\/www.alliai.com\/","content":"<div class=\"col-xl-4 col-sm-6 b-l b-t feature-cell\"><h6 class=\"feature-title\">AI-powered SEO Strategy<\/h6><p>Google is investing hundreds of millions of dollars in RankBrain and the latest AI and Machine Learning tech to grow their business. Now you can too (for a lot less).<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-r b-l b-t feature-cell\"><h6 class=\"feature-title\">Adapt to Algorithm Changes<\/h6><p>As Google changes their algorithm, so should your approach. With Alli, your SEO strategy will update automatically. You don't want to get left behind, do you?<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-r b-t feature-cell\"><h6 class=\"feature-title\">Make SEO Fun Again<\/h6><p>Imagine logging on every day and knowing exactly what to do to increase your traffic as much as possible as fast as possible. That's what life is like with Alli AI.<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-l feature-cell\"><h6 class=\"feature-title\">Get Quick Wins<\/h6><p>Low-hanging fruit, 80\/20, whatever you want to call it. Alli finds the easiest links to build and recommends the code and content changes which will boost your rankings the fastest.<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-r b-l feature-cell\"><h6 class=\"feature-title\">Rank Higher for Real<\/h6><p>This isn't a magic pill, but it is pretty damn good. Search results are ordered by an algorithm \u2014 basically a mathematical equation. If you adjust certain variables in that equation, your rankings go up. Simple.<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-r feature-cell\"><h6 class=\"feature-title\">Traffic You Can Feel<\/h6><p>Remember how winning feels? Imagine the orders coming in, hear the phone ringing, see those KPIs getting blasted and your numbers going up and up. That's what life is like when your traffic goes up.<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-l feature-cell\"><h6 class=\"feature-title\">SEO Peace of Mind<\/h6><p>Do you ever feel stressed about SEO? Between Penguins, Pandas and Hummingbirds, SEO can feel like a maze of anxiety and information overload. Enough! With Alli, you'll know what to do and how to do it.<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-r b-l feature-cell\"><h6 class=\"feature-title\">Tackle Those KPIs<\/h6><p>Your judged on your results, and so are we. Focus your SEO campaign on what matters most \u2014 optimizations, ranking higher and getting more traffic. With the team dashboard, you can easily show your progress to the whole business.<\/p><\/div><div class=\"col-xl-4 col-sm-6 b-r feature-cell\"><h6 class=\"feature-title\">SEO Campaign In-A-Box<\/h6><p>Alli is your Swiss army knife of SEO. Plan your SEO strategy, find backlinks, get code and content optimizations and track your rankings and traffic progress all in one tool. Good deal, eh?<\/p><\/div>","created_at":"2018-11-07T16:43:17+0000","updated_at":"2018-11-07T16:43:17+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.alliai.com","preview_picture":"https:\/\/www.alliai.com\/assets\/homepage\/social-preview-c58d0929aec8f8e9587ed8bbbf66433ff0cdb7d6da6f1dea37a1306275c4b535.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":283,"title":"Freelance Logo Design, Web Design & Graphic Design","url":"https:\/\/www.designcrowd.com\/","content":"<section class=\"hero\"><div class=\"hero__content\"><div class=\"row column\"><h3 class=\"mbl\">With the world's #1 custom design marketplace<\/h3><a href=\"https:\/\/www.designcrowd.com\/registration\" class=\"button large button-cta button-next\" data-reveal-id=\"modal-sign-up\" data-reveal-ajax=\"\/modal\/register\/Customer\" data-reveal-size=\"tiny\" data-success=\"https:\/\/www.designcrowd.com\/start\">Get started now<\/a><a class=\"button button-stealth large\" href=\"https:\/\/www.designcrowd.com\/how-it-works\" data-reveal-id=\"modal-how-it-works\" data-reveal-ajax=\"\/modal\/howitworks\">&#13;\n                <img src=\"https:\/\/dcassetcdn.com\/assets\/images\/new-world\/icons\/common-medium\/common-medium-play-button-a69eca6393.svg\" alt=\"Play button\" \/>&#13;\n                See how it works&#13;\n            <\/a><div class=\"row align-center mtm big-numbers\"><div class=\"column small-12 medium-10 large-6\"><div class=\"row small-up-1 medium-up-3 text-center\"><div class=\"column\"><em>658,425<\/em><p class=\"mtn\">freelance designers<\/p><\/div><div class=\"column\"><em>25 - 100+<\/em><p class=\"mtn\">designs per project<\/p><\/div><div class=\"column\"><em>309,467<\/em><p class=\"mtn\">completed projects<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/section><section class=\"section\"><div class=\"row column\"><h2 class=\"text-center\">Categories<\/h2><section class=\"section\"><section class=\"section\"><p>&#13;\n            <a class=\"button large\" href=\"https:\/\/www.designcrowd.com\/registration\" data-reveal-id=\"modal-sign-up\" data-reveal-ajax=\"\/modal\/register\/Customer\" data-reveal-size=\"tiny\" data-success=\"https:\/\/www.designcrowd.com\/start\">Get started now<\/a>&#13;<\/p><div class=\"text-center\"><a class=\"button-expand-collapse\" id=\"categories-link-expand\" data-js-toggle=\"extra-categories extra-categories-button\">&#13;\n                &#13;\n                    <strong>&#13;\n                        See all categories&#13;\n                        <img src=\"https:\/\/dcassetcdn.com\/assets\/images\/new-world\/icons\/common-tiny\/common-tiny-arrow-down-blue-1a05d6f106.svg\" alt=\"Arrow down\" \/><\/strong>&#13;\n                &#13;\n                &#13;\n                    <strong>&#13;\n                        Hide all categories&#13;\n                        <img src=\"https:\/\/dcassetcdn.com\/assets\/images\/new-world\/icons\/common-tiny\/common-tiny-arrow-down-blue-1a05d6f106.svg\" alt=\"Arrow up\" \/><\/strong>&#13;\n                &#13;\n            <\/a><\/div><\/section><\/section><\/div><\/section><section class=\"section\"><div data-design-carousel=\"\"><h2 class=\"text-center\">Example projects<\/h2><\/div><\/section><section class=\"section\"><div class=\"row column\"><h2 class=\"text-center\">Let's get started<\/h2><\/div><\/section><section class=\"section section--transparent press-links\"><\/section>","created_at":"2018-11-07T16:43:17+0000","updated_at":"2018-11-07T16:43:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.designcrowd.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":284,"title":"Access to this page has been denied.","url":"https:\/\/www.crunchbase.com","content":"<p>\nAccess to this page has been denied because we believe you are using automation tools to browse the\nwebsite.\n<\/p><p>\nThis may happen as a result of the following:\n<\/p><ul><li>\nJavascript is disabled or blocked by an extension (ad blockers for example)\n<\/li>\n<li>\nYour browser does not support cookies\n<\/li>\n<\/ul><p>\nPlease make sure that Javascript and cookies are enabled on your browser and that you are not blocking\nthem from loading.\n<\/p><p>\nReference ID: #3a7616b0-e2ac-11e8-9d28-3f4ef1c8c34f\n<\/p>","created_at":"2018-11-07T16:43:17+0000","updated_at":"2018-11-07T16:43:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.crunchbase.com","http_status":"403"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":285,"title":"Coding Your Own Algo Trading Robot | Investopedia","url":"https:\/\/www.investopedia.com\/articles\/active-trading\/081315\/how-code-your-own-algo-trading-robot.asp","content":"<p>Ever wanted to become an <a href=\"https:\/\/www.investopedia.com\/terms\/a\/algorithmictrading.asp\">algorithmic trader<\/a> with the ability to code your own trading robot? And yet, are you frustrated with the amount of disorganized, misleading information and false promises of overnight prosperity? Well, Lucas Liew, creator of the online algorithmic trading course AlgoTrading101, may have the solution for you. Having excellent reviews and garnering over 8,000 students since first launching in October 2014, Liew\u2019s course \u2013 aimed at presenting the fundamentals of algorithmic trading in an organized way \u2013 is proving to be quite popular. He is adamant about the fact that algorithmic trading is \u201cnot a get-rich-quick scheme.\u201d<\/p><p>Outlined below are the basics of what it takes to design, build and maintain your own <a href=\"https:\/\/www.investopedia.com\/university\/systemcoding\/\">algorithmic trading robot<\/a>\u00a0(drawn from Liew and his course).<\/p><h2>What an Algorithmic Trading Robot Is and Does<\/h2><p>At the most basic level, an algorithmic trading robot is a computer code that has the ability to generate and execute buy and sell signals in financial markets. The main components of such a robot include <a href=\"https:\/\/www.investopedia.com\/terms\/e\/entry-point.asp\">entry rules<\/a> that signal when to buy or sell, <a href=\"https:\/\/www.investopedia.com\/terms\/e\/exit-point.asp\">exit rules<\/a> indicating when to close the current position, and <a href=\"https:\/\/www.investopedia.com\/terms\/p\/positionsizing.asp\">position sizing<\/a> rules defining the quantities to buy or sell.<\/p><h2>The Main Tools<\/h2><p>Obviously, you\u2019re going to need a computer and an Internet connection. After that, a Windows or Mac operating system will be needed to run <a href=\"http:\/\/www.metatrader4.com\/\">MetaTrader 4<\/a> (MT4) \u2013 an electronic <a class=\"termLink\" data-inlink=\"link_11\" href=\"https:\/\/www.investopedia.com\/terms\/t\/trading-platform.asp\">trading platform<\/a> that uses the MetaQuotes Language 4 (MQL4) for coding trading strategies. Although MT4 is not the only software one could use to build a robot it has a number of significant benefits.<\/p><p>While MT4\u2019s main <a class=\"termLink\" data-inlink=\"link_12\" href=\"https:\/\/www.investopedia.com\/terms\/a\/assetclasses.asp\">asset class<\/a> is <a href=\"https:\/\/www.investopedia.com\/terms\/f\/foreign-exchange.asp\">foreign exchange<\/a> (FX), the platform can be used to trade equities, <a href=\"https:\/\/www.investopedia.com\/terms\/e\/equitymarket.asp\">equity indices<\/a>, commodities and Bitcoin\u00a0using <a href=\"https:\/\/www.investopedia.com\/terms\/c\/contractfordifferences.asp\">CFDs<\/a>. Other benefits of using MT4 as opposed to other platforms include being easy to learn, has numerous available FX data sources and it\u2019s free.<\/p><p>Unfortunately, MT4 does not allow for direct trading in stock and <a href=\"https:\/\/www.investopedia.com\/terms\/f\/futuresmarket.asp\">futures markets<\/a> and conducting <a href=\"https:\/\/www.investopedia.com\/terms\/a\/anova.asp\">statistical analysis<\/a> can be burdensome; however, MS Excel can be used as a supplementary statistical tool.<\/p><h2>Algorithmic Trading Strategies<\/h2><p>It is important to begin by reflecting on some core traits that every algorithmic <a class=\"termLink\" data-inlink=\"link_23\" href=\"https:\/\/www.investopedia.com\/terms\/t\/trading-strategy.asp\">trading strategy<\/a> should have. The strategy should be market prudent in that it is fundamentally sound from a market and economic standpoint. Also, the mathematical model used in developing the strategy should be based on <a href=\"https:\/\/www.investopedia.com\/articles\/financial-theory\/10\/gaussian-models-statistics.asp\">sound statistical methods<\/a>.<\/p><p>Next, it is crucial to determine what information your robot is aiming to capture. In order to have an automated strategy, your robot needs to be able to capture identifiable, persistent <a href=\"https:\/\/www.investopedia.com\/terms\/i\/inefficientmarket.asp\">market inefficiencies<\/a>. Algorithmic trading strategies follow a rigid set of rules that take advantage of market behavior and thus, the occurrence of a one-time market inefficiency is not enough to build a strategy around. Further, if the cause of the market inefficiency is unidentifiable, then there will be no way to know if the success or failure of the strategy was due to chance or not.<\/p><p>With the above in mind, there are a number of strategy types to inform the design of your algorithmic trading robot. These include strategies that take advantage of the following (or any combination thereof):<\/p><ul><li>Macroeconomic news (e.g. non-farm payroll or interest rate changes)<\/li>\n<li><a href=\"https:\/\/www.investopedia.com\/terms\/f\/fundamentalanalysis.asp\">Fundamental analysis<\/a> (e.g. using revenue data or earnings release notes)<\/li>\n<li>Statisticalanalysis (e.g. correlation or co-integration)<\/li>\n<li><a href=\"https:\/\/www.investopedia.com\/terms\/t\/technicalanalysis.asp\">Technical analysis<\/a> (e.g. <a href=\"https:\/\/www.investopedia.com\/terms\/m\/movingaverage.asp\">moving averages<\/a>)<\/li>\n<li>The market microstructure (e.g. arbitrage or trade infrastructure)<\/li>\n<\/ul><h2>Designing and Testing Your Robot<\/h2><p>There are essentially four steps needed to build and manage a trading robot:<\/p><h3>Preliminary Research<\/h3><p>This step focuses on developing a strategy that suits your own personal characteristics. Factors such as <a href=\"https:\/\/www.investopedia.com\/terms\/r\/risktolerance.asp\">personal risk profile<\/a>, time commitment and <a class=\"termLink\" data-inlink=\"link_36\" href=\"https:\/\/www.investopedia.com\/terms\/c\/capital.asp\">trading capital<\/a> are all important to think about when developing a strategy. You can then begin to identify the persistent market inefficiencies mentioned above. Having identified a market inefficiency you can begin to code a trading robot suited to your own personal characteristics.<\/p><h3>Backtesting<\/h3><p><a href=\"https:\/\/www.investopedia.com\/terms\/b\/backtesting.asp\">This step<\/a>\u00a0focuses on validating your trading robot. This includes checking the code to make sure it is doing what you want and understanding how it performs over different time frames, asset classes, or different market conditions, especially in <a href=\"https:\/\/www.investopedia.com\/terms\/b\/blackswan.asp\">black swan<\/a>\u00a0type events such as the <a class=\"termLink\" data-inlink=\"link_39\" href=\"https:\/\/www.investopedia.com\/articles\/economics\/09\/financial-crisis-review.asp\">2008 global financial crisis<\/a>.<\/p><h3>Optimization<\/h3><p>Now that you have coded a robot that works and at this stage, you want to maximize its performance while minimizing the\u00a0<a href=\"https:\/\/www.investopedia.com\/terms\/o\/overfitting.asp\">overfitting<\/a> bias. To maximize performance you first need to select a good performance measure that captures risk and reward elements, as well as consistency (e.g. <a href=\"https:\/\/www.investopedia.com\/terms\/s\/sharperatio.asp\">Sharpe ratio<\/a>). An overfitting bias occurs when your robot is too closely based on past data; such a robot will give off the illusion of high performance, but since the future never completely resembles the past, it may actually fail.<\/p><h3>Live Execution<\/h3><p>You are now ready to begin using real money. However, aside from being prepared for the emotional ups and downs that you might experience, there are a few technical issues that need to be addressed. These issues include selecting an appropriate <a class=\"termLink\" data-inlink=\"link_42\" href=\"https:\/\/www.investopedia.com\/terms\/b\/broker.asp\">broker<\/a> and implementing mechanisms to manage both <a class=\"termLink\" data-inlink=\"link_43\" href=\"https:\/\/www.investopedia.com\/terms\/m\/marketrisk.asp\">market risks<\/a> and <a class=\"termLink\" data-inlink=\"link_44\" href=\"https:\/\/www.investopedia.com\/terms\/o\/operational_risk.asp\">operational risks<\/a> such as potential hackers and technology downtime.<\/p><p>It is also important at this step to verify that the robot\u2019s performance is similar to that experienced in the testing stage. Finally, continual monitoring is needed to ensure that the market efficiency that the robot was designed for still exists.\u00a0<\/p><h2>The Bottom Line<\/h2><p>Considering that Richard Dennis, the <a href=\"https:\/\/www.investopedia.com\/articles\/trading\/08\/turtle-trading.asp\">legendary<\/a> commodity trader, taught a group of students his personal trading strategies who then went on to earn over $175 million in just five years, it is completely possible for inexperienced traders to be taught a strict set of guidelines and <a class=\"termLink\" data-inlink=\"link_48\" href=\"https:\/\/www.investopedia.com\/articles\/trading\/10\/top-ten-rules-for-trading.asp\">become successful traders<\/a>. However, this is one extraordinary example and beginners should definitely remember to have modest expectations.<\/p><p>In order to be successful, it is important to not just follow a set of guidelines but to understand how those guidelines are working. Liew stresses that the most important part of algorithmic trading is \u201cunderstanding under which types of market conditions your robot will work and when it will break down,\u201d and \u201cunderstanding when to intervene.\u201d Algorithmic trading can be rewarding but the key to success is understanding. Any course or teacher promising high rewards with minimal understanding should be a major warning sign.<\/p>","created_at":"2018-11-07T16:43:17+0000","updated_at":"2018-11-07T16:43:17+0000","published_at":"2015-08-13T05:33:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.investopedia.com","preview_picture":"https:\/\/i.investopedia.com\/dimages\/graphics\/shutterstock_268518818.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":278,"title":"The best Minecraft mods","url":"https:\/\/www.pcgamer.com\/best-minecraft-mods\/","content":"<p>Here, we've collected the best Minecraft mods, according to our experts on the game. Mods will help you get more out of Minecraft, no matter what it is you need. In this list, we'll collect incremental improvements to the game, dazzling worlds and entirely new types of games you can play within Minecraft. It's worth noting, not all mods are compatible with the latest version of Minecraft, and some require the installation of legacy versions of Forge to get them running. Thankfully, installing older mods isn\u2019t too difficult. Most mods come with instructions, but if they don\u2019t, we\u2019ve got a <a href=\"https:\/\/www.pcgamer.com\/how-to-install-mods-for-older-versions-of-minecraft\/\">mini-guide<\/a> to get you up and running.\u00a0<\/p><p>These mods are great on their own, but it's worth noting that they might conflict with each other somewhat if you try and install several at once. If you do have issues with that or anything else, consider trying a preconfigured modpack out instead\u2014the launchers provided by <a href=\"http:\/\/www.feed-the-beast.com\/\">Feed the Beast<\/a> and the <a href=\"http:\/\/www.technicpack.net\/\">Technic Platform<\/a> are some of the simplest ways to quickly and cleanly get playing modded Minecraft.<\/p><p>If you're looking for <a href=\"https:\/\/www.pcgamer.com\/best-minecraft-seeds\/\" target=\"_blank\">best Minecraft seeds<\/a>, too, or <a href=\"https:\/\/www.pcgamer.com\/minecraft-console-commands-and-cheats-every-player-should-know\/\" target=\"_blank\">Minecraft console commands and cheats<\/a>, or <a href=\"https:\/\/www.pcgamer.com\/the-best-minecraft-servers\/\" target=\"_blank\">best Minecraft servers<\/a>, we can help with everything you need.<\/p><p>Here are the best Minecraft mods:<\/p><h3 id=\"it-x2019-s-portal-but-in-minecraft-xa0\">It\u2019s Portal, but in Minecraft\u00a0<\/h3><h3 id=\"1-10-2\">[1.10.2]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/S6Do6aqSWwc5CqRuQEscu3-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/S6Do6aqSWwc5CqRuQEscu3-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/S6Do6aqSWwc5CqRuQEscu3-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/S6Do6aqSWwc5CqRuQEscu3.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/S6Do6aqSWwc5CqRuQEscu3.jpg\" \/><\/p><p>How do you fancy creating your own Aperture testing lab? Included in this modpack are a wealth of different Portal-themed blocks and, more importantly, a selection of Portal appliances, which includes floor buttons for dumping companion cubes onto, pedestal buttons, and indicator lights. So if you decide to create a nefarious puzzle challenge, all the tools you\u2019ll need will be there from the off.<\/p><p>A quick note: For the full Portal experience, you\u2019ll first need to grab iChun\u2019s Portalgun modpack first from here [Link 1]. After that, you\u2019re safe to install Raptor\u2019s mod-extension from here [Link 2].<\/p><h3 id=\"mill-xe9-naire-xa0\">Mill\u00e9naire\u00a0<\/h3><h3 id=\"1-7-10\">[1.7.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/7bjppB3zPPyu984tpDJ4ET-320-80.jpg\" alt=\"\" class=\"expandable lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/7bjppB3zPPyu984tpDJ4ET-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/7bjppB3zPPyu984tpDJ4ET-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/7bjppB3zPPyu984tpDJ4ET.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/7bjppB3zPPyu984tpDJ4ET.jpg\" \/><\/p><p>How many times have you explored a new area only to be met with nothing but vast emptiness? The Mill\u00e9naire mod fixes that by introducing a wealth of new content into spaces where there\u2019d usually be nothing. Villagers are replaced by human men, women, and children, instead of the bog-standard villager. Even villages have had a makeover to incorporate 11th-century Norman, North Indian and Mayan themes. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/1272755-millenaire-npc-village-new-villagers-in-players\">here<\/a>.\u00a0<\/p><h3 id=\"fossil-and-archaeology-revival-mod\">Fossil and Archaeology revival mod<\/h3><h3 id=\"1-17-10\">[1.17.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/qxiWQV5SDMuNcb5Wg962M4-320-80.jpg\" alt=\"\" class=\"expandable lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/qxiWQV5SDMuNcb5Wg962M4-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/qxiWQV5SDMuNcb5Wg962M4-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/qxiWQV5SDMuNcb5Wg962M4.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/qxiWQV5SDMuNcb5Wg962M4.jpg\" \/><\/p><p>If there\u2019s one thing missing from Minecraft it\u2019s dinosaurs. Who doesn\u2019t want to ditch creepers in favour of riding around on a T-Rex? Not only is there a massive amount of dinosaurs to spawn in creative, but in survival, the player can hunt down fossils and bring these forgotten beasts back to life. As well as a texture overhaul, there\u2019s also a new mob boss in the Anu for those looking to put their prehistoric skills to the test. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/2398119-fossils-and-archeology-revival-build-7-3-2-the\">here<\/a>.\u00a0<\/p><h3 id=\"instant-massive-structures\">Instant Massive Structures<\/h3><h3 id=\"1-8-8-1-8-1-8-1-1-9-1-7-10\">[1.8.8, 1.8, 1.8.1, 1.9, 1.7.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/jgwpkxF6EHSpNe4m8i28Jo-320-80.jpg\" alt=\"\" class=\"expandable lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/jgwpkxF6EHSpNe4m8i28Jo-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/jgwpkxF6EHSpNe4m8i28Jo-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/jgwpkxF6EHSpNe4m8i28Jo.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/jgwpkxF6EHSpNe4m8i28Jo.jpg\" \/><\/p><p>Okay, okay, so this one may be cheating. We\u2019ve all seen monstrous builds online which make our 5x5 dirt house pail in comparison. But what happens if you want to turn the tables? What happens if you want an enormous city in a matter of seconds? That\u2019s where the Instant Structures mod comes in. Simply cycle through the creative menu, drop a block, right click it, and a building will magically spawn. Structures range from castles, to houses, all the way to tram stations. Yeah, sure, it\u2019s technically cheating, but we won\u2019t tell anyone. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/2111962-1-9-1-8-1-7-instant-massive-structures-mod\">here<\/a>.\u00a0<\/p><h3 id=\"the-lost-cities\">The Lost Cities<\/h3><h3 id=\"1-10-1-11-1-12\">[1.10, 1.11, 1.12]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph.jpg\" \/><\/p><p>Something has ended life as we know it. Buildings are in disrepair, and everyone has vanished, or so it seems. The Lost Cities, as the foreboding name suggests, spawns you into a city forgotten by time. The goal? See how long you can survive this post-apocalyptic wasteland without succumbing to death. If you fancy taking the survival one step further, this mod also interlinks with the Biomes O Plenty mod, meaning you\u2019ll have a lot of stunning new areas to explore alongside the empty cities. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/2828782-the-lost-cities-a-post-apocalyptic-city-world\">here<\/a>.\u00a0<\/p><h3 id=\"biosphere\">Biosphere<\/h3><h3 id=\"1-7-10-1-7-2\">[1.7.10, 1.7.2]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5.jpg\" \/><\/p><p>Creating a biosphere, of decent size, takes around eight-hours, give or take. I know this because I spent eight-painstaking-hours making one. By the end, I wanted to chuck Minecraft into a nearby volcano. Making something spherical rather than a jagged diamond isn\u2019t easy. Thankfully, the Biospheres mod removes all the annoying obstacles and spawns you into a world where the sky is overrun with various floating biomes. Just don\u2019t fall off, yeah? Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/wip-mods\/2236413-new-biospheres-mod\">here<\/a>.\u00a0<\/p><h3 id=\"the-aether\">The Aether<\/h3><h3 id=\"1-7-3\">[1.7.3]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/F4uVEYviWbYUZ4uUSVRk2R-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/F4uVEYviWbYUZ4uUSVRk2R-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/F4uVEYviWbYUZ4uUSVRk2R-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/F4uVEYviWbYUZ4uUSVRk2R.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/F4uVEYviWbYUZ4uUSVRk2R.jpg\" \/><\/p><p>The Aether (pronounced \u201cee-ther\u201d) is the opposite of the Nether. While the Nether is seen as Minecraft\u2019s equivalent of hell, the Aether is a sort of heavenly realm. Once transported via a portal, you\u2019ll find yourself above the clouds surrounded by gorgeous, newly-designed floating islands. There\u2019s also new mobs, including flying cows and flying pigs, some new boss enemies, as well as new block types to play with and new-fangled loot to find. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/1275355-1-7-3-aether-collaboration-mod-v1-02-new-mobs\">here<\/a>.\u00a0<\/p><h3 id=\"optifine-xa0\">Optifine\u00a0<\/h3><h3 id=\"1-8-1\">[1.8.1]<\/h3><p>Minecraft doesn't scale too well to the power of fast or slow machines. It runs surprisingly poorly on low-end laptops, and a high-end rig can't do much with its extra oomph. Enter Optifine\u2014a mod that not only makes Minecraft run faster but also look far better. It supports HD textures, smooth lighting, and more, and framerate doubling is not uncommon. It's one of the first things I usually add when installing Minecraft. Grab it <a href=\"https:\/\/optifine.net\">here<\/a>.<\/p><h3 id=\"twilight-forest\">Twilight Forest<\/h3><h3 id=\"1-7-10-2\">[1.7.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/534953a707ce317a847eeb4869a12ce7-320-80.png\" alt=\"Twilightforest\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/534953a707ce317a847eeb4869a12ce7-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/534953a707ce317a847eeb4869a12ce7-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/534953a707ce317a847eeb4869a12ce7.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/534953a707ce317a847eeb4869a12ce7.png\" \/><\/p><p>Love adventuring? This mod adds a new, densely-forested dimension shrouded in perpetual twilight that hides both valuable treasures and dangerous monsters. Throw a diamond into a pool of water surrounded by flowers to create a portal there, then spend a while roaming around. You'll find hedge mazes, hollow hills, enchanted groves, glaciers, lich towers and more with rich rewards for those that delve the deepest. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding\/minecraft-mods\/1276258-the-twilight-forest-v2-3-3-snow-queen-troll\">here<\/a>.<\/p><h3 id=\"biomes-o-plenty\">Biomes O'Plenty<\/h3><h3 id=\"1-7-10-3\">[1.7.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/eaa53698697c41017474080904746fd3-320-80.jpg\" alt=\"Biomesoplenty\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/eaa53698697c41017474080904746fd3-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/eaa53698697c41017474080904746fd3-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/eaa53698697c41017474080904746fd3.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/eaa53698697c41017474080904746fd3.jpg\" \/><\/p><p>Since the '<a href=\"http:\/\/minecraft.gamepedia.com\/The_Update_that_Changed_the_World\">Update That Changed The World<\/a>' in late 2013, there's been a bit more diversity in Minecraft worlds. But Biomes O'Plenty adds vastly more\u201475 to be exact\u2014from brushland and coral reefs, through lavender fields and ominous woods, to tundra and wasteland. You'll need to create a new world to use it (make sure to select the 'Biomes O'Plenty' world generation option), but it's worth it to see corners of Minecraft that you've never seen before. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding\/minecraft-mods\/1286162-biomes-o-plenty-over-75-new-biomes-plants-and-more\">here<\/a>.<\/p><h3 id=\"the-lost-cities-2\">The Lost Cities<\/h3><h3 id=\"1-10-1-11-1-12-2\">[1.10, 1.11, 1.12]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph-320-80.jpg\" alt=\"\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/WLzTuT5uWiCPG8SpQEAWph.jpg\" \/><\/p><p>Something has ended life as we know it. Buildings are in disrepair, and everyone has vanished, or so it seems. The Lost Cities, as the foreboding name suggests, spawns you into a city forgotten by time. The goal? See how long you can survive this post-apocalyptic wasteland without succumbing to death. If you fancy taking the survival one step further, this mod also interlinks with the Biomes O Plenty mod, meaning you\u2019ll have a lot of stunning new areas to explore alongside the empty cities. Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/2828782-the-lost-cities-a-post-apocalyptic-city-world\">here<\/a>.\u00a0<\/p><h3 id=\"biosphere-2\">Biosphere<\/h3><h3 id=\"1-7-10-1-7-2-2\">[1.7.10, 1.7.2]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5-320-80.jpg\" alt=\"\" class=\"expandable lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5-320-80.jpg 320w, https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5-650-80.jpg 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5.jpg\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/9z29pTGQjNkMvYeb5PcWx5.jpg\" \/><\/p><p>Creating a biosphere, of decent size, takes around eight-hours, give or take. I know this because I spent eight-painstaking-hours making one. By the end, I wanted to chuck Minecraft into a nearby volcano. Making something spherical rather than a jagged diamond isn\u2019t easy. Thankfully, the Biospheres mod removes all the annoying obstacles and spawns you into a world where the sky is overrun with various floating biomes. Just don\u2019t fall off, yeah? Grab it <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding-java-edition\/minecraft-mods\/wip-mods\/2236413-new-biospheres-mod\">here<\/a>.<\/p><h3 id=\"botania\">Botania<\/h3><h3 id=\"1-7-10-4\">[1.7.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/2574322bfe6d3a7b8cdbaeacd4719e2c-320-80.png\" alt=\"Botania\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/2574322bfe6d3a7b8cdbaeacd4719e2c-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/2574322bfe6d3a7b8cdbaeacd4719e2c-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/2574322bfe6d3a7b8cdbaeacd4719e2c.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/2574322bfe6d3a7b8cdbaeacd4719e2c.png\" \/><\/p><p>Some mods add powerful magical items. Others add intricate machinery. Botania just adds flowers\u2014but wow, what flowers. Flowers that heal you. Flowers that feed animals. Flowers that turn hostile mobs against each other. Flowers that eat cake. Oh, and did I mention that you've can also use flowers to create a magical portal to a world of elves? If you want to try something wildly different from most other mods, Botania is it. Grab it <a href=\"http:\/\/botaniamod.net\/index.php\">here<\/a>.<\/p><h3 id=\"inventory-tweaks-notenoughitems-waila\">Inventory Tweaks, NotEnoughItems &amp; Waila<\/h3><h3 id=\"1-7-10-5\">[1.7.10]<\/h3><p class=\"bordeaux-image-check\"><img src=\"https:\/\/cdn.mos.cms.futurecdn.net\/e2fb3c4528c7427d8a65aa8282924789-320-80.png\" alt=\"Nei\" class=\"lazy-image lazy-image-loading lazyload optional-image\" data-normal=\"https:\/\/vanilla.futurecdn.net\/pcgamer\/media\/img\/missing-image.svg\" data-srcset=\"https:\/\/cdn.mos.cms.futurecdn.net\/e2fb3c4528c7427d8a65aa8282924789-320-80.png 320w, https:\/\/cdn.mos.cms.futurecdn.net\/e2fb3c4528c7427d8a65aa8282924789-650-80.png 650w\" data-sizes=\"auto\" data-original-mos=\"https:\/\/cdn.mos.cms.futurecdn.net\/e2fb3c4528c7427d8a65aa8282924789.png\" data-pin-media=\"https:\/\/cdn.mos.cms.futurecdn.net\/e2fb3c4528c7427d8a65aa8282924789.png\" \/><\/p><p>This trio of mods are essential quality-of-life improvements, especially when you've got loads of mods installed at the same time. Inventory Tweaks allows you to sort your chests with a single click and automatically replace tools when they break. NotEnoughItems provides you with a searchable list of all the blocks available in the game, and the recipes for crafting them, and Waila lets you point your cursor at an unfamiliar blocks to find out what it is. Grab them <a href=\"https:\/\/inventory-tweaks.readthedocs.org\/en\/latest\/\">here<\/a>, <a href=\"http:\/\/chickenbones.net\/Pages\/links.html\">here<\/a> and <a href=\"http:\/\/www.minecraftforum.net\/forums\/mapping-and-modding\/minecraft-mods\/1289765-waila-1-5-7-for-1-7-10-1-6-0-for-1-8-1\">here<\/a> respectively.<\/p>","created_at":"2018-11-07T16:43:16+0000","updated_at":"2018-11-07T16:43:16+0000","published_by":["Duncan Geere"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"www.pcgamer.com","preview_picture":"https:\/\/cdn.mos.cms.futurecdn.net\/S6Do6aqSWwc5CqRuQEscu3-1200-80.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":279,"title":"Tweet from \n\n    \n      James Ryan","url":"https:\/\/mobile.twitter.com\/xfoml\/status\/894666621393936385","content":"<div class=\"dir-ltr\" dir=\"ltr\">  one of the first computer-generated plots, produced by Sheldon Klein's murder-mystery system (1973) <a href=\"https:\/\/t.co\/0y5y1dm2Xn\" data-pre-embedded=\"true\" rel=\"nofollow\" data-entity-id=\"894666288210993152\" dir=\"ltr\" data-url=\"https:\/\/twitter.com\/xfoml\/status\/894666621393936385\/photo\/1\" data-tco-id=\"0y5y1dm2Xn\" class=\"twitter_external_link dir-ltr tco-link has-expanded-path\" target=\"_top\" data-expanded-path=\"\/xfoml\/status\/894666621393936385\/photo\/1\">pic.twitter.com\/0y5y1dm2Xn<\/a>\n<\/div>","created_at":"2018-11-07T16:43:16+0000","updated_at":"2018-11-07T16:43:16+0000","published_by":["James Ryan"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"mobile.twitter.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":280,"title":"Projects","url":"https:\/\/www.jamesryan.world\/projects\/","content":"<p><a>Menu<\/a><\/p><section id=\"page\" role=\"main\" data-content-field=\"main-content\"><div id=\"projectPages\" data-collection-id=\"\"><div class=\"project-controls\"><div><p>prev \/ next<\/p><a href=\"#\">Back to Projects<\/a><\/div><\/div><\/div><div id=\"projectThumbs\"><div class=\"wrapper\"><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/sheldon-county\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/5a96f926e2c4837c160e46ad\/1519843673514\/logo_bw.jpg?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Sheldon County<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/badnews\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/58fd24581b631b1afffbc2b9\/1492985001403\/5.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Bad News<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/gamespace\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/58fd2bbeb8a79b0b91cfd08f\/1492986819122\/gamespace_angle1.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>GameSpace<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/talktown\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/58fd2061440243efc9d5e5cc\/1492984001706\/17.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Talk of the Town<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/expressionist\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/591114a4725e256f43ec94bb\/1494291622946\/expressionist.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Expressionist<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/conversation\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110bc0725e256f43ec202e\/1494289396998\/neural_nlu.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Videogame Conversation Tech<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/storygen-history\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/5930c013b8a79be27197e52d\/1496367126129\/grimes.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>History of Story Generation<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/branching-narrative\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59ab0342bebafb8981d7f283\/1504379719971\/corelli.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Origins of Branching Narrative<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/typewriter-art\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59a9b704cf81e015f86de0ab\/1504294664158\/col.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Early Typewriter Art<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/gamenet\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110ae7bf629a03c2eb3cbb\/1494289129067\/gamenet.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>GameNet<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/gamesage\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110e6d5016e1d346f8aa62\/1494290049772\/gamesage.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>GameSage<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/gameglobs\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110aa5bf629a03c2eb393e\/1494289063137\/gameglobs.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>GameGlobs<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/islanders\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110b22d2b857aab92d356d\/1494289193176\/islanders.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Islanders<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/diol\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110dc6893fc011d4e6eb93\/1494289867691\/diol.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Diol \/ Diel \/ Dial<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/archipelagos\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/590e1363e6f2e1fda306e423\/1494094882466\/example_game_world.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Archipelago Generator<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/jukejoint\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/58fd6919ebbd1a41121a681a\/1494293878265\/juke_joint_title.jpg?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>Juke Joint<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/nationalpastime\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110a53bebafbe5ba37097a\/1494288989324\/national_pastime.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>National Pastime<\/p><\/div>\n        <\/a><a class=\"project\" href=\"https:\/\/www.jamesryan.world\/vfclust\/\">\n          <div><div class=\"project-image\"><div class=\"intrinsic\"><div class=\"content-fill\"><img src=\"https:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/59110b446a49631ac1e697b4\/1494289221753\/vfclust.png?format=500w\" alt=\"image\" \/><\/div><\/div><\/div><p>VFClust<\/p><\/div>\n        <\/a><\/div><\/div><\/section>","created_at":"2018-11-07T16:43:16+0000","updated_at":"2018-11-07T16:43:16+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"www.jamesryan.world","preview_picture":"http:\/\/static1.squarespace.com\/static\/58fb957a3a041197bb7f37f9\/t\/590fe1421e5b6c8e16f89e0f\/1494212945056\/social_sharing.png?format=1000w","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":275,"title":"Generating $25,000\/mo Offering Wikis to Company Slack Teams - Indie Hackers","url":"https:\/\/www.indiehackers.com\/interview\/tettra-f7e811e0c2","content":"<h2>Hello! What's your background, and what are you working on?<\/h2><p>Hi, I'm Nelson Joyce, a product designer and cofounder of <a href=\"https:\/\/tettra.co\">Tettra<\/a>. I've been designing and building B2B SaaS products since 2010.<\/p><p>Tettra is a communication tool for teams to document processes, projects, and knowledge. Most work tools like chat, project management, and email are quick and transactional. Tettra is built for sharing information that requires a deeper and more well thought out write up like product specs or company-wide updates.<\/p><p>Making information available to your team allows them to make better decisions, since they have more context. You can think of it like a new take on a traditional company wiki.<\/p><p>We're about two years in, have just north of 300 customers, and are making $23,844 in monthly recurring revenue at the time I'm writing this.<\/p><a class=\"screenshot-wrapper\" href=\"https:\/\/tettra.co\">\n  <img alt=\"Tettra homepage\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/content\/tettra__website.jpg\" \/><\/a><h2>What motivated you to get started with Tettra?<\/h2><p>Before we started <a href=\"https:\/\/tettra.co\">Tettra<\/a>, my co-founder Andy and I worked on a \"secret project\" inside of a large public company called HubSpot. It was just the two of us working on this project for over a year, and we built up a ton of knowledge that was trapped in our heads.<\/p><p>Once our product started to get some traction, our bosses decided to add more people to our team. Since we had been so secretive with our project, it was extremely painful trying to communicate all this built-up knowledge to our new team members.<\/p><p>Anticipating having to do the same thing with new teammates, we looked for a simple knowledge base tool for sharing all that knowledge. We didn't find one that worked the way we wanted, so we decided to leave HubSpot and build our own instead.<\/p><p>Since HubSpot had gone public a few years earlier, we were able to sell some of the stock and self-fund the company for the first six months.<\/p><h2>What went into building the initial product?<\/h2><p>Despite the fact that my co-founder and I are both technical, we didn't go straight into building. We spent about six weeks on initial customer interviews before writing a single line of code.<\/p><p>As far as finding people to interview, we posted to our personal networks and did some good old cold outreach. Here's a <a href=\"https:\/\/medium.com\/@nelsonjoyce\/4-strategies-we-used-to-get-over-100-customer-interviews-c7577429453a\">more in-depth post<\/a> about those strategies that I wrote.<\/p><p>Once we got folks into interviews, we asked about pain points around sharing internal information, specifically involving current wiki and documentation solutions. Here are <a href=\"https:\/\/medium.com\/@nelsonjoyce\/4-customer-interview-questions-you-should-never-ask-a5596586313b\">some tips I wrote up<\/a> for interview questions to ask.<\/p><p>Based on that research, we were able to spend November of 2015 building a prototype that we could get some people using ASAP. We actually just forked a WordPress plugin called P2 and made some modifications for the initial beta users. This wasn't a perfect production environment that we planned on scaling to thousands of users; it was a simple WPEngine setup just to get the product into the hands of a few beta users to start learning.<\/p><p>The bad news was that once we rolled out that initial prototype, no one ended up engaging with it. The good news was that we only spent a few weeks building something instead of months. Coincidently, during the time we went back to the drawing board, Slack released their new platform ecosystem. When I looked back at all the initial research we did, I noticed that <strong>every<\/strong> team we talked to used and loved Slack.<\/p><p>We decided to make the switch and spent the two weeks around Christmas and New Years rebuilding a brand new prototype built 100% to work with Slack. Once we launched the new Slack prototype to the public in the first week of January, we had hundreds of people sign up in the first few days. This was the validation we needed to double down on this Slack-first approach. <a href=\"https:\/\/thinkgrowth.org\/building-on-slack-saved-our-startup-94953fdaf27a\">Here's a post<\/a> where you can read more details about that process and the results.<\/p><p>At this point we decided that in order to capture all this new demand around the Slack platform, we needed to increase our development speed. The problem was that we were almost out of the money we had put into the business. We either had to do contract work or try to raise a small angel round to fund continued development.<\/p><p>The angel round allowed us to hire a full-time engineer and rebuild the product (yet again!) to create a scalable platform for future development. We chose to <a href=\"https:\/\/medium.com\/team-tettra\/our-tech-stack-highlights-7a7c23013e9f\">build the product on Laravel<\/a> since there's so much already built in to get you up and running quickly and simply. For more details about the nitty gritty of getting the round done in only a few months, check out <a href=\"https:\/\/soundcloud.com\/techinboston\/andy-cook-tettra\">this podcast episode<\/a>.<\/p><p>We spent March and April building this new version on the Slack platform and signed up our first paying customer on May 10, 2016.<\/p><a class=\"screenshot-wrapper\" href=\"https:\/\/tettra.co\">\n  <img alt=\"Tettra first payment\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/content\/tettra__first-payment.png\" \/><\/a><p>So from start to first customer it took us a little over seven months of learning and iteration.<\/p><h2>How have you attracted users and grown Tettra?<\/h2><p>We publicly launched the new product in May of 2016 with a <a href=\"https:\/\/medium.com\/team-tettra\/tettra-an-internal-wiki-built-for-slack-teams-edbcf2d388f0\">Medium post<\/a> announcing the \"public beta\" and our funding.<\/p><p>That post immediately got <a href=\"https:\/\/www.americaninno.com\/boston\/slack-wiki-startup-tettra-raises-seed-round-from-hubspot-execs\/\">picked up<\/a> by the local tech publication Bostinno.<\/p><p>We did a brainstorming session around all the channels we could try out and followed the <a href=\"https:\/\/medium.com\/@yegg\/the-bullseye-framework-for-getting-traction-ef49d05bfd7e\">Bullseye framework<\/a> to prioritize a few of them. Here's the actual <a href=\"https:\/\/app.tettra.co\/teams\/tettra\/pages\/launch-channels\">Tettra page<\/a> where we did the brainstorming.<\/p><p>We did the typical sharing to all social networks (Linkedin, Facebook, Twitter, etc.).<\/p><p>Prior to launch, we <a href=\"https:\/\/medium.com\/team-tettra\/startup-first-steps-post-on-betalist-c53fccc9f982\">posted on Betalist<\/a> and had about 65 people on a waiting list. We also posted to all the typical directories like Stackshare, Siftery, Capterra, and Alternativeto. As you can see in the chart below, only Betalist and Alternativeto ended up driving any significant traffic.<\/p><p>We decided to wait to post on Product Hunt until the product was more fleshed out and our sales funnel was working on a small scale. We ended up launching on Product Hunt about six months later and got over 700 upvotes and was the <a href=\"https:\/\/www.producthunt.com\/posts\/tettra-2\">#3 spot for that day<\/a>. Here's the <a href=\"https:\/\/app.tettra.co\/teams\/tettra\/pages\/product-hunt-launch-plan\">actual Tettra page<\/a> we used to plan the Product Hunt launch.<\/p><p><strong>The Slack Platform<\/strong><\/p><p>Of course another massive source of trials was the Slack App Directory. After our experience, I highly recommend launching on a platform like Slack, WordPress, or Shopify.<\/p><p>There's a whole chunk of functionality that you can essentially \"outsource\" to the platform. For us, that was authentication, user management, and access to the \"work graph\". You also get free distribution from the parent platform and a clear target persona you can attract.<\/p><p>There are definitely risks involved, but the benefits outweighed the risks in our eyes. Here you can check out more details on <a href=\"https:\/\/thinkgrowth.org\/building-on-slack-saved-our-startup-94953fdaf27a\">why you should launch on a platform<\/a> from my co-founder Andy.<\/p><p><strong>\"Emerging\" Keywords<\/strong><\/p><p>We were also lucky enough to jump on the \"Slack Wiki\" keyword pretty early on an ranked in the top few slots from the get go. There wasn't much competition for that term at the time, and we've been able to maintain the top non-branded Google result for that. All we did was put the term \"Slack Wiki\" in our website title tags and headers, and that was good enough to rank for a new term.<\/p><p><strong>Results<\/strong><\/p><a class=\"screenshot-wrapper\" href=\"https:\/\/tettra.co\">\n  <img alt=\"Tettra traffic sources\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/content\/tettra__traffic-sources-may-july.png\" \/><\/a><p>These are the traffic sources between May 1 and July 31 to <a href=\"https:\/\/tettra.co\">tettra.co<\/a> after launch. As you can see, most of the traffic came from Google and Slack. Those top two channels make up about 50% of the trials as well.<\/p><p><strong>Bottom of the funnel<\/strong><\/p><p>Once we got people signed up for a trial, we got as many of them on a screen share as possible to walk them through the product and collected their credit card info right on the phone.<\/p><p>We tried to continue to scale a demo-based inside sales playbook, but after a lot of pushing, we couldn't get it to work. The problem was that the product was simple enough and we were selling to small enough teams that they just wanted to start using the product right way and didn't want to do demos. It was a great way to get the machine going initially, but it just didn't scale for us.<\/p><p>Our strategy now is to sell \"touchlessly\" with emails, getting started resources, and chat support. We're still in the early stages of this transition, so I'll have to report back on the forum with the data once we see the results.<\/p><h2>What's your business model, and how have you grown your revenue?<\/h2><p>We launched the initial product with no billing integration at all \u2014 we just collected credit cards over the phone with customers and entered them manually into Stripe. There was nothing preventing anyone from going to <a href=\"https:\/\/app.tettra.co\">app.tettra.co<\/a> and using the product without paying. But no-one did of course. It saved us a few days of dev time and allowed us to start selling right away.<\/p><p>Since then, we've tested a bunch of different pricing models.<\/p><p>We started with $5 per user per month. The rationale was that it was the same mechanism but a smaller amount than Slack's pricing. We ended up switching the pricing to flat \"bundles\" of users due to customer feedback. We also were seeing that it took months for folks to add their team members, so we were only collecting $5-15 dollars in the first few months of a new customer.<\/p><p>When we changed to the flat pricing, we saw a big bump in average sales price. We started making more money by doing less work, which was a win-win.<\/p><p>We tried to increase the prices a second time, and while we still saw customer signups, the trial numbers were way down. You end up trading the higher revenue for fewer customers, so we decided we wanted to be lower on the spectrum to attract the appropriately sized customer.<\/p><p>Currently, we use a lower \"base\" price based on features (starts at $60\/mo) and have a second axis of pricing around user bundles. We think it's a good balance between not being too intimidating to start and the ability for a customer account to grow as they use the product more over time.<\/p><a class=\"screenshot-wrapper\" href=\"https:\/\/tettra.co\">\n  <img alt=\"Tettra MRR growth\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/content\/tettra__mrr-growth.png\" \/><\/a><p>The hard thing about pricing is that you can't really A\/B test it, so it takes a long time to see the results of a change. Even with a 15-day free trial, it still takes us 6-8 weeks to see the full effect of any change.<\/p><p>My advice would be to keep gradually increasing the price over time until you see your signups drop off to a point where you're uncomfortable. Then lower the prices to just below that point.<\/p><h2>What are your goals for the future?<\/h2><p>Our next big milestone is getting to profitability. We project that we'll get there by Q2 of next year.<\/p><p>Beyond profitability, getting to 1 million ARR would be a great accomplishment that we think we can reach. Increasing top of the funnel traffic, optimizations throughout the onboarding funnel, and revenue growth from our existing customer base should all combine to help get us there.<\/p><p>Our biggest roadblock right now is acquisition (shocking, I know). Driving enough of the right type of person to your product or service is always tricky business.<\/p><p>Another challenge I foresee is learning how to run a profitable company. Everywhere I've worked in the past has followed the typical \"hypergrowth\" startup strategy which entails running at a loss using massive amounts of VC funding. There's nothing wrong with that approach, it just takes a different kind of mindset to be successful.<\/p><p>We're pretty bullish on the <a href=\"http:\/\/indie.vc\">IndieVC<\/a> model that <a href=\"https:\/\/twitter.com\/bryce\">Bryce Roberts<\/a> is working on, so I'm excited to learn more from folks like him in this community.<\/p><h2>If you had to start over, what would you do differently?<\/h2><p>Our biggest mistake was trying to do too much at once, which manifested in us raising our burn rate too high. (We've since lowered it.) This is such a common mistake I'm almost embarrassed to admit it.<\/p><p>One of the downsides of raising money is the false sense of security you get. This caused us to try to sell to too many target customers with both a sales-driven and product-driven model. This pulled us in too many directions.<\/p><p>Focus on one use case for one target user with one go-to-market strategy. This is even more important for small, resource-strapped companies.<\/p><h2>What's your advice for indie hackers who are just starting out?<\/h2><p>Sleep. Seriously, getting a good night's sleep is my top \"hack\" for staying happy and productive.<\/p><p>Don't burn yourself out. I see a lot of founders and tech folks grind themselves to the bone to get more done. In reality, those extra hours are probably hurting your productivity in the long run. When you're working, work hard, but when you're home, rest your body and mind to tackle what lies ahead.<\/p><p>Everything takes longer to get right than you think. Not much you can do to about this except to stay positive and patient! A lot of products win simply because they don't get acquired and\/or shut down.<\/p><p>Learn about pricing. One of the highest leverage changes you can make for your business is a proper pricing strategy. I <em>highly<\/em> recommend reading <a href=\"https:\/\/blog.profitwell.com\">everything written by Patrick Campbell<\/a> at PriceIntelligently \u2014 specifically the <a href=\"http:\/\/www.priceintelligently.com\/saas-pricing-strategy-ebook?__hssc=263140243.2.1482463892066&amp;__hstc=263140243.8b221627a570d9492f726532b7acd60e.1470410574392.1481197338430.1482463892066.7&amp;hsCtaTracking=fa9d47ed-c0a8-4e81-9e20-37d52b0eb473%7C21329f3c-657d-49e5-be40-9e1fdfc59f64\">SaaS pricing eBook<\/a>. It took us too long to get this right (and there's still room for improvement!).<\/p><p>Learn about \"market\/product\/channel\/model fit\". Most people focus on the integration between products and markets (aka product market fit), but <a href=\"https:\/\/brianbalfour.com\/essays\/product-market-fit-isnt-enough\">this concept by Brian Balfour<\/a> has been transformational for how I understand software businesses.<\/p><p>Acquisition will probably be your biggest challenge. Everyone underestimates the difficulty of acquiring customers, especially us hacker types. Almost all great companies win due to distribution advantages.<\/p><h2>Where can we go to learn more?<\/h2><p>If you're interested in improving communication and transparency on your team, <a href=\"https:\/\/tettra.co?utm_source=indiehackers&amp;utm_medium=referral\">give Tettra a try<\/a> and let us know what you think. Happy to answer any questions in the comments below <a href=\"https:\/\/twitter.com\/nelsonjoyce\">or on Twitter<\/a> as well.<\/p><p><strong>Bonus if you've made it this far:<\/strong> Mention this post to us and we'll give you 20% off your subscription for the first three months.<\/p><p>\n  \u2014\n  <a id=\"ember531777\" href=\"https:\/\/www.indiehackers.com\/Nelson?id=oS7LFpiW6PQTIoGaFrxou8Ad4hf2\" class=\"user-link__link ember-view\">    <img class=\"user-avatar__img\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/avatars\/72x72_oS7LFpiW6PQTIoGaFrxou8Ad4hf2\" alt=\"image\" \/>\n      \n          Nelson Joyce\n      \n\n    \n<\/a>\n, Co-founder of Tettra\n<\/p><h2>Want to build your own business like Tettra?<\/h2><p>\n  You should join the <a id=\"ember531779\" href=\"https:\/\/www.indiehackers.com\/\" class=\"ember-view\">Indie Hackers community<\/a>! \ud83e\udd17\n<\/p><p>\n  We're a few thousand founders helping each other build profitable businesses\n  and side projects. Come share what you're working on and get feedback from\n  your peers.\n<\/p><p>\n  Not ready to get started on your product yet? No problem. The community is a\n  great place to meet people, learn, and get your feet wet. Feel free to\n  <a id=\"ember531780\" href=\"https:\/\/www.indiehackers.com\/\" class=\"ember-view\">just browse<\/a>!\n<\/p><p>\n  \u2014\n  <a id=\"ember531782\" href=\"https:\/\/www.indiehackers.com\/csallen?id=ibTLPyjwVebnZjMGKvz6ztarnuV2\" class=\"user-link__link ember-view\">    <img class=\"user-avatar__img\" src=\"https:\/\/storage.googleapis.com\/indie-hackers.appspot.com\/avatars\/72x72_ibTLPyjwVebnZjMGKvz6ztarnuV2\" alt=\"image\" \/>\n      \n          Courtland Allen\n      \n\n    \n<\/a>\n, Indie Hackers founder\n<\/p>","created_at":"2018-11-07T16:43:14+0000","updated_at":"2018-11-07T16:43:14+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":12,"domain_name":"www.indiehackers.com","preview_picture":"https:\/\/580157013239-us-east-1-chromeless.s3.amazonaws.com\/interview-shareables\/f7e811e0c2.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":276,"title":"Qake","url":"http:\/\/qake.se\/demo\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:43:14+0000","updated_at":"2018-11-07T16:43:14+0000","annotations":[],"reading_time":0,"domain_name":"qake.se","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":277,"title":"Boston\u2019s Best Bars For Celebrating St. Patrick\u2019s\u00a0Day!","url":"https:\/\/boston.cbslocal.com\/top-lists\/bostons-best-bars-for-celebrating-st-patricks-day\/","content":"<div id=\"wrapper\" class=\"body-wrapper\"><section class=\"page-ad-container page-pencil-ad-container-top\"><div id=\"page-content\" class=\"main-white page-block-container and-w-sidebar\"><div class=\"cols\"><div class=\"big-left\"><div class=\"well\"><div class=\"post-wrap\"><p class=\"post-stamp\">March 17, 2012 2:02 PM<\/p><div class=\"post-entry\"><img width=\"420\" height=\"316\" src=\"https:\/\/cbsboston.files.wordpress.com\/2011\/03\/blackrosetavern.jpg?w=420\" class=\"attachment-single size-single wp-post-image\" alt=\"\" srcset=\"https:\/\/cbsboston.files.wordpress.com\/2011\/03\/blackrosetavern.jpg 420w, https:\/\/cbsboston.files.wordpress.com\/2011\/03\/blackrosetavern.jpg?w=200 200w\" \/><div class=\"wp-caption\"><p>Patrons at Black Rose Tavern in Boston celebrated St. Patrick\u2019s Day.<\/p><\/div><div class=\"description\"><div class=\"basicsection fifty50_rw divider\"><em>\u2013 By Colleen Mescall<\/em>Did you know that Boston is said to be where the American St. Patrick\u2019s Day started? Did you know that about 13 million pints of Guinness are served around the world on March 17th? Beware: Did you know that many people who wear the ever-popular \u201cKiss me, I\u2019m Irish!\u201d pins are actually not Irish? (How dare they!)Now onto the most important question\u2026out of all the Irish bars in Boston, which is the best spot to pick for your St. Patrick\u2019s Day celebration? Here are some suggestions that cover Boston, Davis Square, Dorchester and Harvard Square. Whether you\u2019re Irish year-round, or just one special day of the year \u2013 we\u2019ve got something for everyone. Erin go bragh!<p><em><strong>Read: <a href=\"https:\/\/boston.cbslocal.com\/top-lists\/bostons-best-st-patricks-day-party-spots\/\" target=\"_blank\">More of Boston\u2019s Best St. Patrick\u2019s Day Party Spots<\/a><\/strong><\/em><\/p><\/div><div class=\"basicsection fifty50_rw divider\"><div id=\"attachment_222738\" class=\"wp-caption alignnone\"><img class=\"size-full wp-image-222738\" title=\"3\/8 Nightlife &amp; Music Black Rose Boston\" alt=\"black rose e1324400467985 Boston\u2019s Best Bars For Celebrating St. Patrick\u2019s Day!\" src=\"https:\/\/cbsboston.files.wordpress.com\/2011\/12\/black-rose-e1324400467985.jpg?w=420\" srcset=\"https:\/\/cbsboston.files.wordpress.com\/2011\/12\/black-rose-e1324400467985.jpg 300w, https:\/\/cbsboston.files.wordpress.com\/2011\/12\/black-rose-e1324400467985.jpg?w=200 200w\" \/><p class=\"wp-caption-text\">Photo Credit: BlackRoseBoston.com<\/p><\/div><p>The Black Rose is set right in the heart of Boston, and without being able to take a trip over to Ireland, this is truly as authentic as it gets. How can you tell? Well, one \u2013 you\u2019ll be greeted by a staff member with a real brogue. Two \u2013 you can enjoy your perfectly poured Guinness with their signature (and delicious) corned beef, and three \u2013 you can bet your Lucky Charms that they\u2019ll have the best live Irish music around.<\/p><p><em>Insider Tip: they\u2019ve also been known to hand out free Irish breakfasts on March 17th. <\/em><\/p><\/div><div class=\"basicsection fifty50_rw divider\"><p><img class=\"alignnone\" title=\"3\/8 Nightlife &amp; Music St Patrick\u2019s Day\" alt=\"98159523 Boston\u2019s Best Bars For Celebrating St. Patrick\u2019s Day!\" src=\"https:\/\/cbsboston.files.wordpress.com\/2011\/12\/98159523.jpg?w=300&amp;h=225\" width=\"300\" height=\"225\" \/><\/p><p><br \/>Hours: Daily 11 a.m. \u2013 2 a.m.<\/p><p>If you\u2019re looking for a rowdier time in Boston \u2013 McGreevy\u2019s is your spot. Located in the heart of the Back Bay and \u2018home\u2019 to the Dropkick Murphy\u2019s, this sports bar means business on St. Patrick\u2019s Day. Their history dates all the way back to 1894, and they say they host the \u201cBest St. Patrick\u2019s Day parties in Boston\u201d\u2026we\u2019re thinking this year might be the year to see if that\u2019s true.<\/p><\/div><div class=\"basicsection fifty50_rw divider\"><p><img class=\"alignnone\" title=\"The Burren\" alt=\"the burren Boston\u2019s Best Bars For Celebrating St. Patrick\u2019s Day!\" src=\"https:\/\/cbsboston.files.wordpress.com\/2010\/10\/the-burren.jpg?w=300&amp;h=225\" width=\"300\" height=\"225\" \/><\/p><p>Looking for a Somerville spot to celebrate St. Paddy? We\u2019re thinking The Burren in Davis Square. With traditional Irish music offered seven nights a week, Guinness and Magners Irish Cider on tap, fish n\u2019 chips on the menu, a cozy atmosphere, and Irish accents galore\u2026how can you go wrong? Sl\u00e1inte!<\/p><\/div><div class=\"basicsection fifty50_rw divider\"><p><img class=\"alignnone\" title=\"3\/8 Nightlife &amp; Music Guinness Beer\" alt=\"105864957 Boston\u2019s Best Bars For Celebrating St. Patrick\u2019s Day!\" src=\"https:\/\/cbsboston.files.wordpress.com\/2011\/12\/105864957.jpg?w=300&amp;h=225\" width=\"300\" height=\"225\" \/><\/p><p>If you\u2019re south of Boston but don\u2019t necessarily want to venture into the Southie craziness, give The Banshee in Dorchester a try for a dose of \u201cIrish Boston\u201d. They\u2019re the number one spot in the area where you can enjoy watching European sports on one of the many flat screens, including Gaelic football and hurling (two amazing Irish sports) \u2013 not to mention the traditional Irish folk music they feature and the mean Irish breakfast they serve. We have to say, we trust them for our celebration!<\/p><\/div><div class=\"basicsection fifty50_rw divider\"><img class=\"alignnone\" title=\"3\/8 Nightlife &amp; Music Grafton Street\" alt=\"grafton 011 Boston\u2019s Best Bars For Celebrating St. Patrick\u2019s Day!\" src=\"https:\/\/cbsboston.files.wordpress.com\/2011\/12\/grafton_011.jpg?w=300&amp;h=225\" width=\"300\" height=\"225\" \/><br \/>Hours: Sun \u2013 Weds open until 12a.m. Thurs \u2013 Sat open until 1 a.m.Boston has the traditional St. Patrick\u2019s Day parade, step dancers all over the place, and a long-standing Irish history in general. Want to mix traditional with modern this year? Grafton Street in Harvard Square is a classy pub who can do just that. They boast that they\u2019re \u201canything but your typical Irish bar\u201d, and that\u2019s fine by us. Around since 1996, this spot is named after the lively shopping district in the Temple Bar area of Dublin \u2013 they\u2019ll give you the traditional Irish hospitality mixed with a contemporary menu and chic d\u00e9cor.<em>Colleen Mescall is the Promotions Manager for a website in the Boston area and holds a special place in her heart for the Celtics. She writes about her love of all things New England over at her blog, <a href=\"http:\/\/thedailycraic.com\/\" target=\"_blank\">the daily craic<\/a> (catch her on Twitter <a href=\"https:\/\/twitter.com\/#!\/thedailycraic\" target=\"_blank\">@thedailycraic<\/a>).<\/em><\/div><\/div><\/div><\/div><\/div><\/div><\/div><section class=\"page-ad-container page-pencil-ad-container-bottom\"><\/section><\/div><footer id=\"page-footer\" class=\"page-footer-type-menus-vertical\"><section class=\"top\"><div class=\"wrapper\"><a href=\"https:\/\/boston.cbslocal.com\" class=\"logo\"><img src=\"https:\/\/cbsboston.files.wordpress.com\/2017\/10\/boston.png\" alt=\"CBS Boston\" \/><\/a><\/div><\/section><section class=\"middle\"><section class=\"bottom\"><div class=\"wrapper\"><a href=\"https:\/\/boston.cbslocal.com\" class=\"logo\"><img src=\"https:\/\/s2.wp.com\/wp-content\/themes\/vip\/cbs-local\/plugins\/modules\/theme-manager\/images\/cbs-local-logo-white.png\" alt=\"CBS Boston\" \/><\/a><p>\u00a92018 CBS Broadcasting Inc. All Rights Reserved.\n\t\t\t\t\t\t\t\t\tPowered by <a href=\"https:\/\/vip.wordpress.com\/?utm_source=vip_powered_wpcom&amp;utm_medium=web&amp;utm_campaign=VIP%20Footer%20Credit&amp;utm_term=boston.cbslocal.com\" rel=\"generator nofollow\" class=\"powered-by-wpcom\">WordPress.com VIP<\/a><\/p><div class=\"right-wrapper\"><p>By viewing our video content, you are accepting the terms of our <a href=\"http:\/\/policies.cbslocal.com\/privacy-june-2015\/current.html#video-services\" target=\"_blank\" title=\"Video Services Policy\">Video Services Policy<\/a><\/p><\/div><\/div><\/section><\/section><\/footer><\/section><\/div>\n\t\n\t\t\n<noscript><\/noscript>","created_at":"2018-11-07T16:43:14+0000","updated_at":"2018-11-07T16:43:14+0000","published_at":"2012-03-17T14:02:30+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"boston.cbslocal.com","preview_picture":"https:\/\/cbsboston.files.wordpress.com\/2011\/03\/blackrosetavern.jpg?w=420","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":274,"title":"4 Customer Interview Questions You Should Never Ask","url":"https:\/\/medium.com\/@nelsonjoyce\/4-customer-interview-questions-you-should-never-ask-a5596586313b","content":"<section class=\"section section--body section--middleCenter section-image--aspectRatioViewport is-imageBackgrounded is-backgrounded is-darkBackgrounded u-imageSpectrum section--first\"><div class=\"js-sectionPoint u-ignoreBlock u-textColorTransparentWhite u-absolute u-top0 u-left0\"><div class=\"section-background u-width100pct u-absolute u-top0 u-cursorDefault u-bottom0\" data-scroll=\"aspect-ratio-viewport\" contenteditable=\"false\" data-image-id=\"1*nmVhZQyCXhpoXRmuhWBgLw.jpeg\" data-width=\"2692\" data-height=\"1830\"><div class=\"section-backgroundImage u-absolute u-top0 u-left0 u-right0 u-backgroundCover u-bottom0\"><div class=\"u-ignoreBlock u-absolute0 u-borderRadius3\"><br \/><\/div><div class=\"section-doubleWidthTable u-table u-ignoreBlock\"><div class=\"section-contentCell u-tableCell u-verticalAlignTop u-ignoreBlock\"><div class=\"section-aspectRatioViewportBottomSpacer u-ignoreBlock\"><div class=\"u-ignoreBlock\"><\/div><div class=\"section-doubleWidthTable u-table u-ignoreBlock\"><div class=\"section-contentCell u-tableCell u-ignoreBlock\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"1003\" class=\"graf graf--h2 graf--leading graf--title\">4 Customer Interview Questions You Should Never\u00a0Ask<\/h1><h2 id=\"5fbf\" class=\"graf graf--h4 graf-after--h2 graf--trailing graf--subtitle\">(and what to ask\u00a0instead)<\/h2><\/div><\/div><\/div><div class=\"u-tableCell u-verticalAlignTop u-ignoreBlock\"><div class=\"section-aspectRatioViewportPlaceholder u-ignoreBlock\"><div class=\"u-ignoreBlock\"><\/div><div class=\"section-aspectRatioViewportBottomPlaceholder u-ignoreBlock\"><div class=\"u-ignoreBlock\"><\/div><\/div><\/div><\/div><div class=\"u-tableCell u-ignoreBlock\"><div class=\"section-aspectRatioViewportCropPlaceholder u-ignoreBlock\"><div class=\"u-ignoreBlock\"><\/div><\/div><\/div><div class=\"section-captionContent u-ignoreBlock\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"9ae2\" class=\"graf graf--p graf--leading\">Interviewing customers is a fundamental skill every team of startup founders needs. Asking good interview questions can be the difference between creating something people want or wasting time building something useless.<\/p><p id=\"48af\" class=\"graf graf--p graf-after--p\">I\u2019ve sat in on and conducted <a href=\"http:\/\/leadin.com\/100-customer-interviews\/\" data-href=\"http:\/\/leadin.com\/100-customer-interviews\/\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">hundreds of customer interviews <\/a>over the past 3 years while working on the product team at <a href=\"http:\/\/HubSpot.com\" data-href=\"http:\/\/HubSpot.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">HubSpot<\/a> and as part of a two-man internal startup inside of HubSpot called <a href=\"http:\/\/leadin.com\" data-href=\"http:\/\/leadin.com\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">Leadin<\/a>. During all of those interviews, I have either heard or asked a lot of really bad questions.<\/p><p id=\"6d28\" class=\"graf graf--p graf-after--p\">When we started working on Leadin, we had no product ideas, just a target market we wanted to help. By doing those customer interviews, we were able to validate product ideas and build something that our <a href=\"https:\/\/wordpress.org\/support\/view\/plugin-reviews\/leadin\" data-href=\"https:\/\/wordpress.org\/support\/view\/plugin-reviews\/leadin\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">customers love using<\/a>.<\/p><p id=\"a9f8\" class=\"graf graf--p graf-after--p\">At HubSpot, we have a team of researchers that can make sure we\u2019re asking good questions, but if you\u2019re starting out, you\u2019ll need to learn these skills yourself. Here\u2019s four bad interview questions I often hear asked, and what to ask instead.<\/p><h4 id=\"d449\" class=\"graf graf--h4 graf-after--p\"><strong class=\"markup--strong markup--h4-strong\">What do you think of our\u00a0product?<\/strong><\/h4><p id=\"3a73\" class=\"graf graf--p graf-after--h4\">This question is bad for two reasons. One, it\u2019s much too broad. You probably only have 30 minutes to an hour with this customer and you need to make it count. Asking really broad questions leads to really broad, meandering answers.<\/p><p id=\"35cb\" class=\"graf graf--p graf-after--p\">The second problem is that people are probably going to be too nice. You\u2019ve taken the time to talk to them, so their probably hesitant to bash your ideas.<\/p><p id=\"d980\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Ask instead: <\/em><\/strong><em class=\"markup--em markup--p-em\">What\u2019s your workflow like when you use our product?<\/em><\/p><p id=\"32eb\" class=\"graf graf--p graf-after--p graf--trailing\">This question is much better because it focuses on real situation in which someone uses your product. If you\u2019re looking for improvements to the current product, having someone explain the steps they took to use it will help reveal the pain points. Also, it takes opinion out of the equation. Either someone was able to efficiently complete their task or they weren\u2019t.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h4 id=\"23f0\" class=\"graf graf--h4 graf--leading\"><strong class=\"markup--strong markup--h4-strong\">What features should we\u00a0add?<\/strong><\/h4><p id=\"6d91\" class=\"graf graf--p graf-after--h4\">If you ask this question, customers will always try to think of an answer, whether it\u2019s something they actually need or not. Everyone wants to be helpful. Also, in my experience, customers are notoriously bad at determining what features your product needs. Focus on customers problems, not on the solutions they come up with to those problems.<\/p><p id=\"9f5d\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Ask instead: <\/em><\/strong><em class=\"markup--em markup--p-em\">What tools do you use alongside our product?<\/em><\/p><p id=\"8a5c\" class=\"graf graf--p graf-after--p\">There\u2019s going to be a good amount of jobs related to the jobs your product already does. By focusing on what customers are already doing, you don\u2019t have to guess whether or not it\u2019s useful.<\/p><p id=\"25a6\" class=\"graf graf--p graf-after--p graf--trailing\">When you add these types of features, you need to make sure that the experience of doing that job inside of you product is better than they way your customers are currently doing it. There are some jobs that are too hard to replace, like sending emails outside of Gmail for example.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h4 id=\"4cb2\" class=\"graf graf--h4 graf--leading\"><strong class=\"markup--strong markup--h4-strong\">How often do you do\u00a0X?<\/strong><\/h4><p id=\"42d6\" class=\"graf graf--p graf-after--h4\">When answering this question, customers will tend to tell you an idealized version of their usage patterns instead of an accurate one. They also won\u2019t want to hurt your feelings if their usage is minimal.<\/p><p id=\"688c\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Ask instead: <\/em><\/strong><em class=\"markup--em markup--p-em\">When was the last time you did X?<\/em><\/p><p id=\"c233\" class=\"graf graf--p graf-after--p graf--trailing\">The difference is subtle here, but it\u2019s critical. Instead of having folks hyperbolize about usage, ground the answer in actual usage instead.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h4 id=\"66a0\" class=\"graf graf--h4 graf--leading\"><strong class=\"markup--strong markup--h4-strong\"><em class=\"markup--em markup--h4-em\">Would you use\/buy\u00a0this?<\/em><\/strong><\/h4><p id=\"09fe\" class=\"graf graf--p graf-after--h4\">Having people hypothesize about their usage and purchasing behavior is dangerous territory. There are so many factors that go into both usage and buying decisions, it\u2019s really hard for people to predict the outcome before they actually are faced with a real decision.<\/p><p id=\"a4b6\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\"><em class=\"markup--em markup--p-em\">Ask instead: <\/em><\/strong><em class=\"markup--em markup--p-em\">There\u2019s no good way to ask this.<\/em><\/p><p id=\"2251\" class=\"graf graf--p graf-after--p graf--trailing\">This is where your ability to gain insights from interviewing falls short. The only way to know if someone is going to use or buy something is to give them the ability to do so.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"212b\" class=\"graf graf--p graf--leading graf--trailing\">Customer interviews aren\u2019t easy, but the insights you gain from them are invaluable. Interviewing is a muscle you have to strengthen over time, so keep practicing and keep these questions in mind.<\/p><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><figure id=\"af65\" class=\"graf graf--figure graf--iframe graf--leading graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><div class=\"aspectRatioPlaceholder-fill\"><div class=\"iframeContainer\"><iframe data-width=\"800\" data-height=\"400\" width=\"700\" height=\"350\" src=\"https:\/\/medium.com\/media\/65cb64d42e63d50273ca7ac020cb1d16?postId=a5596586313b\" data-media-id=\"65cb64d42e63d50273ca7ac020cb1d16\" allowfullscreen=\"allowfullscreen\" frameborder=\"0\">[embedded content]<\/iframe><\/div><\/div><\/div><\/figure><\/div><\/div><\/section>","created_at":"2018-11-07T16:42:47+0000","updated_at":"2018-11-07T16:42:47+0000","published_at":"2014-10-31T20:23:48+0000","published_by":["Nelson Joyce \ud83d\udc1f"],"annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"medium.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*nmVhZQyCXhpoXRmuhWBgLw.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":271,"title":"Calm by Tones - Tones","url":"https:\/\/www.tones.fm\/tones\/calm","content":"<div id=\"rotate_message\"><div id=\"rotate_inner\"><img src=\"https:\/\/d8ev4i00v2x30.cloudfront.net\/assets\/rotate-48239cb5bc6d5e704c49d35d8fd6f7cbf3451a6f23750ae297fccdc45630e18d.png\" alt=\"image\" \/><\/div><\/div><div id=\"settings_browse\" class=\"settings-popover\"><div class=\"settings-popover-inner\"><div><div><table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td valign=\"top\">\n\t\t\t\n\t\t\t\t\t\t<p>Browse<\/p>\n\t\t\t\n\t\t\t\t\t<\/td>\n\t\t\t\t\t<td valign=\"top\" style=\"width: 100%;\">\n\t\t\t\n\t\t\t\t\t\t<div id=\"browse_switcher\" class=\"settings-switcher settings-switcher2\"><p>Saved<\/p><p>Favorites<\/p><\/div>\n\t\t\t\n\t\t\t\t\t<\/td>\n\t\t\t\t<\/tr><\/table><\/div><\/div><\/div><\/div>\n<div id=\"settings_bar\" class=\"settings-popover\"><div class=\"settings-popover-inner\"><table cellpadding=\"0\" border=\"0\" class=\"settings-table\" style=\"border-spacing: 0px; width: 100%;\"><tr><td valign=\"middle\">\n\t\t\t\t\n\t\t\t\t<p>Repeat<\/p>\n\t\t\t\t\n\t\t\t<\/td>\n\t\t\t<td valign=\"top\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t<\/td>\n\t\t<\/tr><tr><td valign=\"top\" colspan=\"2\">\n\t\t\t\t\n\t\t\t\t<p><i class=\"fa fa-clone fa-3x fa-fw\" aria-hidden=\"true\">Duplicate<\/i><\/p>\n\t\t\t\t<p>Delete<\/p>\n\t\t\t\t\n\t\t\t<\/td>\n\t\t<\/tr><\/table><\/div><\/div>\n<div id=\"settings_edit\" class=\"settings-popover\"><div class=\"settings-popover-inner\"><table cellpadding=\"0\" border=\"0\" class=\"settings-table\" style=\"border-spacing: 0px; width: 100%;\"><tr><td valign=\"top\" rowspan=\"3\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t<\/td>\n\t\t\t<td valign=\"top\">\n\t\t\t\t\n\t\t\t\t<p>Clear<\/p>\n\t\t\t\t<p><i class=\"fa fa-magic fa-3x -fa-fw\" aria-hidden=\"true\">Randomize<\/i><\/p>\n\t\t\t\t\n\t\t\t<\/td>\n\t\t<\/tr><tr>\n\t\t\t<td valign=\"top\">\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t<p>Undo<\/p>\n\t\t\t\t<p>Redo<\/p>\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t<\/td>\n\t\t<\/tr><\/table><\/div><\/div>\n<div id=\"settings_sound\" class=\"settings-popover\"><div class=\"settings-popover-inner\"><div class=\"sound-settings-tabs\"><p>Main<\/p><p>Shape<\/p><p>Wobble<\/p><p>Effects<\/p><\/div><\/div><\/div>","created_at":"2018-11-07T16:42:46+0000","updated_at":"2018-11-07T16:42:46+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.tones.fm","preview_picture":"https:\/\/d8ev4i00v2x30.cloudfront.net\/assets\/icon512-fd09ff0628b8dd77a2d977aa6ca63c70b9dcc7c5da1d4fef159c7f43b1445a61.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":272,"title":"3D Lightning","url":"http:\/\/calculatedimages.blogspot.com\/2013\/05\/3d-lightning.html","content":"<a href=\"http:\/\/www.reddit.com\/\">Reddit<\/a>\u00a0is a great website, where the ability to share and discuss things on the web gives some great little discoveries. Things that would otherwise seem impossibly unlikely, like two people in completely different places getting a photo of the same lightning bolt, suddenly pop up all the time.<p><\/p><div class=\"separator\">\n<a href=\"http:\/\/1.bp.blogspot.com\/-dAySK_SlUA4\/UYuXjjaxLuI\/AAAAAAAAIBI\/A2zg3MJwjag\/s1600\/Cwsnm6qh.jpg\"><img border=\"0\" height=\"480\" src=\"http:\/\/1.bp.blogspot.com\/-dAySK_SlUA4\/UYuXjjaxLuI\/AAAAAAAAIBI\/A2zg3MJwjag\/s640\/Cwsnm6qh.jpg\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<div>\n\u00a0Pic by <a href=\"http:\/\/www.reddit.com\/r\/Denver\/comments\/1dxs0s\/it_was_a_shocking_day_at_coors_field_xpost_from\/\">chordnine<\/a><\/div>\n<div>\n<br \/><\/div>\n<div class=\"separator\">\n<a href=\"http:\/\/2.bp.blogspot.com\/-Ye3azP5t09k\/UYuXj1L4eTI\/AAAAAAAAIBM\/GAUOkqjVv7M\/s1600\/SkJQBmFh.jpg\"><img border=\"0\" height=\"480\" src=\"http:\/\/2.bp.blogspot.com\/-Ye3azP5t09k\/UYuXj1L4eTI\/AAAAAAAAIBM\/GAUOkqjVv7M\/s640\/SkJQBmFh.jpg\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<div>\n\u00a0Pic by <a href=\"http:\/\/www.reddit.com\/r\/Denver\/comments\/1dwjp4\/maybe_the_tree_will_finally_turn_green\/\">Bobo1010<\/a><\/div>\n<div>\n<br \/><\/div>\n<div>\nHaving two pictures of the exact same lightning bolt lets you do something pretty amazing; reconstruct its path in 3D. In this case because the precise location and elevation of the photographers isn't known this is slightly more art than science, but it is still fun!<\/div>\n<div>\n<br \/><\/div>\n<div>\nThese are the two bolts, scaled to approximately the same size:<\/div>\n<div>\n<br \/><\/div>\n<div class=\"separator\">\n<\/div>\n<div class=\"separator\">\n<a href=\"http:\/\/2.bp.blogspot.com\/-Ci798ed0ks4\/UYuZyQn5cOI\/AAAAAAAAIBs\/TRsRsyBNYnY\/s1600\/BothBolts.jpg\"><img border=\"0\" height=\"345\" src=\"http:\/\/2.bp.blogspot.com\/-Ci798ed0ks4\/UYuZyQn5cOI\/AAAAAAAAIBs\/TRsRsyBNYnY\/s640\/BothBolts.jpg\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/>\nIt is immediately clear that they are taken from about the same <i>direction <\/i>but different <i>heights<\/i>: the second bolt looks squashed vertically. This means the pair of images are roughly a stereo pair, but with a vertical shift instead of a horizontal. This is just like the pair of images you would see with your eyes if you had two eyes positioned vertically instead of horizontally on your head.<p><\/p><div class=\"separator\">\n<a href=\"http:\/\/2.bp.blogspot.com\/-0vvxM5n60OM\/UYuZyTEfo4I\/AAAAAAAAIBw\/2FgaHSIkme4\/s1600\/PointsSelected.png\"><img border=\"0\" height=\"371\" src=\"http:\/\/2.bp.blogspot.com\/-0vvxM5n60OM\/UYuZyTEfo4I\/AAAAAAAAIBw\/2FgaHSIkme4\/s640\/PointsSelected.png\" width=\"640\" alt=\"image\" \/><\/a><\/div>\n<br \/><div class=\"separator\">\n<\/div>\n<div class=\"separator\">\nTo analyse this the first step is to trace the lightning bolt, making sure that every point in one image matches up to the corresponding point in the other image, then record the coordinates of all the points. This gives a nice table of numbers where you can calculate the <i>difference<\/i> in x and y position in the two images.<\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div class=\"separator\">\n<\/div>\n<div class=\"separator\">\n<a href=\"http:\/\/3.bp.blogspot.com\/-Arejd_FC3dk\/UYuqeOzMs_I\/AAAAAAAAICQ\/Zv0PaHphDBI\/s1600\/spreadsheet.png\"><img border=\"0\" src=\"http:\/\/3.bp.blogspot.com\/-Arejd_FC3dk\/UYuqeOzMs_I\/AAAAAAAAICQ\/Zv0PaHphDBI\/s1600\/spreadsheet.png\" alt=\"image\" \/><\/a><\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div class=\"separator\">\nNow we need to do some maths... except I don't like doing complicated maths and it turns out there is a big simplification you can make! If both pictures are taken from a long way away from the lightning bolt (i.e. the object has quite a small angular size in the image) then the shift in position between the images is proportional to the distance from the camera. Bigger shifts mean that bit of the bolt is closer to the camera. This approximation is pretty accurate for the majority of cameras, so I used it here.<\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div class=\"separator\">\nThe other problem is the proportionality factor. If one part of the lightning bolt shifts twice as much between the two images as another part that means it is twice as close. But twice as close as what? Without knowing exactly where the cameras were positioned that means only the <i>relative<\/i> distance, not <i>absolute<\/i> distance, can be calculated. Oh well, close enough!<\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div class=\"separator\">\nSo what does the lightning bolt look like in 3D? I plugged the coordinates into Blender and this is the result:<\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div class=\"separator\">\n<a href=\"http:\/\/2.bp.blogspot.com\/-5wBDJxha2x0\/UYuqerVr_JI\/AAAAAAAAICM\/d_avg9pHQz4\/s1600\/Ani2.gif\"><img border=\"0\" src=\"http:\/\/2.bp.blogspot.com\/-5wBDJxha2x0\/UYuqerVr_JI\/AAAAAAAAICM\/d_avg9pHQz4\/s1600\/Ani2.gif\" alt=\"image\" \/><\/a><\/div>\n<br \/><div class=\"separator\">\nPretty amazing really!<\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div class=\"separator\">\nSoftware used:<\/div>\n<div class=\"separator\">\n<a href=\"http:\/\/rsbweb.nih.gov\/ij\/\">ImageJ<\/a>: Image analysis.<\/div>\n<div class=\"separator\">\n<a href=\"http:\/\/www.blender.org\/\">Blender<\/a>: 3D modelling and rendering.<\/div>\n<div class=\"separator\">\n<br \/><\/div>\n<div>\n<\/div>","created_at":"2018-11-07T16:42:46+0000","updated_at":"2018-11-07T16:42:46+0000","published_at":"2013-05-09T00:00:00+0000","published_by":["Richard Wheeler"],"annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":2,"domain_name":"calculatedimages.blogspot.com","preview_picture":"http:\/\/1.bp.blogspot.com\/-dAySK_SlUA4\/UYuXjjaxLuI\/AAAAAAAAIBI\/A2zg3MJwjag\/w1200-h630-p-k-no-nu\/Cwsnm6qh.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":273,"title":"emacs2018 | The Changelog","url":"https:\/\/changelog.complete.org\/archives\/tag\/emacs2018","content":"<div class=\"entry-content\">\n\t\t&#13;\n<div id=\"outline-container-org09fccba\">&#13;\n<h3 id=\"org09fccba\">The Emacs series<\/h3>&#13;\n<div id=\"text-0-1\">&#13;\n<p>&#13;\nThis is fifth in a <a href=\"https:\/\/changelog.complete.org\/archives\/tag\/emacs2018\">series on Emacs and org-mode<\/a>.&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nThis blog post was generated from <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/blob\/master\/emacs\/emacs-org-beamer\/emacs-org-beamer.org\">an org-mode source<\/a> and is available&#13;\nas: <a href=\"http:\/\/changelog.complete.org\/archives\/9900-emacs-5-documents-and-presentations-with-org-mode\">a blog page<\/a>, <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/raw\/master\/emacs\/emacs-org-beamer\/emacs-org-beamer.pdf\">slides (PDF format)<\/a>, and a <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/raw\/master\/emacs\/emacs-org-beamer\/emacs-org-beamer-document.pdf\">PDF document<\/a>.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orga9c4cf3\">&#13;\n<h2 id=\"orga9c4cf3\">1 About org-mode exporting<\/h2>&#13;\n<div id=\"text-1\">&#13;\n<\/div><div id=\"outline-container-org789bc73\">&#13;\n<h3 id=\"org789bc73\">1.1 Background<\/h3>&#13;\n<div id=\"text-1-1\">&#13;\n<p>&#13;\norg-mode isn't just an agenda-making program.  It can also export to&#13;\nlots of formats: LaTeX, PDF, Beamer, iCalendar (agendas), HTML,&#13;\nMarkdown, ODT, plain text, man pages, and more complicated formats&#13;\nsuch as a set of web pages.&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nThis isn't just some afterthought either; it's a core part of the&#13;\nsystem and integrates very well.&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nOne file can be source code, automatically-generated output, task&#13;\nlist, documentation, and presentation, all at once.&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nSome use org-mode as their preferred markup format, even for things&#13;\nlike LaTeX documents.  The org-mode manual has an extensive <a href=\"https:\/\/orgmode.org\/manual\/Exporting.html#Exporting\">section on&#13;\nexporting<\/a>.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-org18e8ede\">&#13;\n<h3 id=\"org18e8ede\">1.2 Getting started<\/h3>&#13;\n<div id=\"text-1-2\">&#13;\n<p>&#13;\nFrom any org-mode document, just hit <code>C-c C-e<\/code>.  From there will come&#13;\nup a menu, letting you choose various export formats and options.&#13;\nThese are generally single-key options so it's easy to set and&#13;\nexecute.  For instance, to export a document to a PDF, use &#13;\n<code>C-c C-e l p<\/code> or for HTML export, <code>C-c C-e h h<\/code>.&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nThere are lots of settings available for all of these export options;&#13;\nsee the manual.  It is, in fact, quite possible to use LaTeX-format&#13;\nequations in both LaTeX and HTML modes, to insert arbitrary preambles&#13;\nand settings for different modes, etc.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-org2847d65\">&#13;\n<h3 id=\"org2847d65\">1.3 Add-on packages<\/h3>&#13;\n<div id=\"text-1-3\">&#13;\n<p>&#13;\nELPA containts many addition exporters for org-mode as well.  Check&#13;\nthere for details.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-org2a38105\">&#13;\n<h2 id=\"org2a38105\">2 Beamer slides with org-mode<\/h2>&#13;\n<div id=\"text-2\">&#13;\n<\/div><div id=\"outline-container-org4058ffb\">&#13;\n<h3 id=\"org4058ffb\">2.1 About Beamer<\/h3>&#13;\n<div id=\"text-2-1\">&#13;\n<p>&#13;\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Beamer_(LaTeX)\">Beamer<\/a> is a LaTeX environment for making presentations.  Its features&#13;\ninclude:&#13;\n<\/p>&#13;\n&#13;\n<ul><li>Automated generating of structural elements in the&#13;\npresentation (see, for example, <a href=\"https:\/\/hartwork.org\/beamer-theme-matrix\/all\/beamer-albatross-Marburg-1.png\">the Marburg theme<\/a>).  This provides&#13;\na visual reference for the audience of where they are in the&#13;\npresentation.<\/li>\n<li>Strong help for structuring the presentation<\/li>\n<li>Themes<\/li>\n<li>Full LaTeX available<\/li>\n<\/ul><\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-org6ad9a02\">&#13;\n<h3 id=\"org6ad9a02\">2.2 Benefits of Beamer in org-mode<\/h3>&#13;\n<div id=\"text-2-2\">&#13;\n<p>&#13;\norg-mode has a lot of benefits for working with Beamer.  Among them:&#13;\n<\/p>&#13;\n&#13;\n<ul><li>org-mode's very easy and strong support for visualizing and&#13;\nchanging the structure makes it very quick to reorganize your&#13;\nmaterial.<\/li>\n<li>Combined with org-babel, live source code (with syntax&#13;\nhighlighting) and results can be embedded.<\/li>\n<li>The syntax is often easier to work with.<\/li>\n<\/ul><p>&#13;\nI have completely replaced my usage of&#13;\nLibreOffice\/Powerpoint\/GoogleDocs with org-mode and beamer.  It is, in&#13;\nfact, rather frustrating when I have to use one of those tools, as&#13;\nthey are nowhere near as strong as org-mode for visualizing a&#13;\npresentation structure.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-org56abede\">&#13;\n<h3 id=\"org56abede\">2.3 Headline Levels<\/h3>&#13;\n<div id=\"text-2-3\">&#13;\n<p>&#13;\norg-mode's Beamer export will convert sections of your document&#13;\n(defined by headings) into slides.  The question, of course, is:&#13;\n<i>which<\/i> sections?  This is governed by the H <a href=\"https:\/\/orgmode.org\/manual\/Export-settings.html#Export-settings\">export setting<\/a>&#13;\n(<code>org-export-headline-levels<\/code>).&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nThere are many ways to go, which suit people.  I like to have my&#13;\npresentation like this:&#13;\n<\/p>&#13;\n&#13;\n<pre>&#13;\n#+OPTIONS: H:2&#13;\n#+BEAMER_HEADER: \\AtBeginSection{\\frame{\\sectionpage}}&#13;\n<\/pre>&#13;\n&#13;\n<p>&#13;\nThis gives a standalone section slide for each major topic, to&#13;\nhighlight major transitions, and then takes the level 2 (two&#13;\nasterisks) headings to set the slide.  Many Beamer themes expect a&#13;\nthird level of indirection, so you would set <code>H:3<\/code> for them.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orge4b18fa\">&#13;\n<h3 id=\"orge4b18fa\">2.4 Themes and settings<\/h3>&#13;\n<div id=\"text-2-4\">&#13;\n<p>&#13;\nYou can configure many Beamer and LaTeX settings in your document by&#13;\ninserting lines at the top of your org file.  This document, for&#13;\ninstance, defines:&#13;\n<\/p>&#13;\n&#13;\n<pre>&#13;\n#+TITLE:  Documents and presentations with org-mode&#13;\n#+AUTHOR: John Goerzen&#13;\n#+BEAMER_HEADER: \\institute{The Changelog}&#13;\n#+PROPERTY: comments yes&#13;\n#+PROPERTY: header-args :exports both :eval never-export&#13;\n#+OPTIONS: H:2&#13;\n#+BEAMER_THEME: CambridgeUS&#13;\n#+BEAMER_COLOR_THEME: default&#13;\n<\/pre>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orgd5f0e34\">&#13;\n<h3 id=\"orgd5f0e34\">2.5 Advanced settings<\/h3>&#13;\n<div id=\"text-2-5\">&#13;\n<p>&#13;\nI like to change some colors, bullet formatting, and the like.  I&#13;\nround out my document with:&#13;\n<\/p>&#13;\n&#13;\n<pre>&#13;\n# We can't just +BEAMER_INNER_THEME: default because that picks the theme default.&#13;\n# Override per https:\/\/tex.stackexchange.com\/questions\/11168\/change-bullet-style-formatting-in-beamer&#13;\n#+BEAMER_INNER_THEME: default&#13;\n#+LaTeX_CLASS_OPTIONS: [aspectratio=169]&#13;\n#+BEAMER_HEADER: \\definecolor{links}{HTML}{0000A0}&#13;\n#+BEAMER_HEADER: \\hypersetup{colorlinks=,linkcolor=,urlcolor=links}&#13;\n#+BEAMER_HEADER: \\setbeamertemplate{itemize items}[default]&#13;\n#+BEAMER_HEADER: \\setbeamertemplate{enumerate items}[default]&#13;\n#+BEAMER_HEADER: \\setbeamertemplate{items}[default]&#13;\n#+BEAMER_HEADER: \\setbeamercolor*{local structure}{fg=darkred}&#13;\n#+BEAMER_HEADER: \\setbeamercolor{section in toc}{fg=darkred}&#13;\n#+BEAMER_HEADER: \\setlength{\\parskip}{\\smallskipamount}&#13;\n<\/pre>&#13;\n&#13;\n<p>&#13;\nHere, <code>aspectratio=169<\/code> sets a&#13;\n16:9 aspect ratio, and the remaining are standard LaTeX\/Beamer&#13;\nconfiguration bits.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orgb9546cd\">&#13;\n<h3 id=\"orgb9546cd\">2.6 Shrink (to fit)<\/h3>&#13;\n<div id=\"text-2-6\">&#13;\n<p>&#13;\nSometimes you've got some really large code examples and you might&#13;\nprefer to just shrink the slide to fit.  &#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nJust type <code>C-c C-x p<\/code>, set the <code>BEAMER_opt<\/code> property to <code>shrink=15<\/code>.  &#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\n(Or a larger value of shrink).  The previous slide uses this here.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n<div id=\"outline-container-orgc077aa1\">&#13;\n<h3 id=\"orgc077aa1\">2.7 Result<\/h3>&#13;\n<div id=\"text-2-7\">&#13;\n<p>&#13;\nHere's the end result:&#13;\n<\/p>&#13;\n&#13;\n&#13;\n<div>&#13;\n<p><a href=\"https:\/\/www.flickr.com\/photos\/jgoerzen\/26366340577\/in\/dateposted\/\" title=\"screenshot1\"><img src=\"https:\/\/farm1.staticflickr.com\/889\/26366340577_fbde8ff266_o.png\" width=\"1283\" height=\"723\" alt=\"screenshot1\" \/><\/a>&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-org3768469\">&#13;\n<h2 id=\"org3768469\">3 Interactive Slides<\/h2>&#13;\n<div id=\"text-3\">&#13;\n<\/div><div id=\"outline-container-orgde7d3e3\">&#13;\n<h3 id=\"orgde7d3e3\">3.1 Interactive Emacs Slideshows<\/h3>&#13;\n<div id=\"text-3-1\">&#13;\n<p>&#13;\nWith the <a href=\"https:\/\/orgmode.org\/worg\/org-tutorials\/non-beamer-presentations.html#org-tree-slide\">org-tree-slide package<\/a>, you can display your slideshow from&#13;\nright within Emacs.  Just run <code>M-x org-tree-slide-mode<\/code>.  Then, use&#13;\n<code>C-&gt;<\/code> and <code>C-&lt;<\/code> to move between slides.&#13;\n<\/p>&#13;\n&#13;\n<p>&#13;\nYou might find <code>C-c C-x C-v<\/code> (which is <code>org-toggle-inline-images<\/code>)&#13;\nhelpful to cause the system to display embedded images.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orgd046819\">&#13;\n<h3 id=\"orgd046819\">3.2 HTML Slideshows<\/h3>&#13;\n<div id=\"text-3-2\">&#13;\n<p>&#13;\nThere are a lot of ways to export org-mode presentations to HTML, with&#13;\nvarious levels of JavaScript integration.  See the <a href=\"https:\/\/orgmode.org\/worg\/org-tutorials\/non-beamer-presentations.html\">non-beamer&#13;\npresentations section<\/a> of the org-mode wiki for details.&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orgaff1066\">&#13;\n<h2 id=\"orgaff1066\">4 Miscellaneous<\/h2>&#13;\n<div id=\"text-4\">&#13;\n<\/div><div id=\"outline-container-orgee09be5\">&#13;\n<h3 id=\"orgee09be5\">4.1 Additional resources to accompany this post<\/h3>&#13;\n<div id=\"text-4-1\">&#13;\n<ul><li><a href=\"https:\/\/orgmode.org\/worg\/exporters\/beamer\/tutorial.html\">orgmode.org beamer tutorial<\/a><\/li>\n<li><a href=\"https:\/\/en.wikibooks.org\/wiki\/LaTeX\/Presentations\">LaTeX wiki<\/a><\/li>\n<li><a href=\"https:\/\/tex.stackexchange.com\/questions\/117658\/automatically-generate-section-title-slides-in-beamer\/117661\">Generating section title slides<\/a><\/li>\n<li><a href=\"https:\/\/tex.stackexchange.com\/questions\/78514\/content-doesnt-fit-in-one-slide\">Shrinking content to fit on slide<\/a><\/li>\n<li>A <b>great<\/b> resource: refcard-org-beamer&#13;\n<ul><li>See its <a href=\"https:\/\/github.com\/fniessen\/refcard-org-beamer\">Github repo<\/a><\/li>\n<li>Make sure to check out both the PDF and the .org file<\/li>\n<\/ul><\/li>\n<li>A nice <a href=\"https:\/\/hartwork.org\/beamer-theme-matrix\/\">Theme matrix<\/a><\/li>\n<\/ul><\/div>&#13;\n<\/div>&#13;\n&#13;\n<div id=\"outline-container-orgf013de8\">&#13;\n<h3 id=\"orgf013de8\">4.2 Up next in my Emacs series\u2026<\/h3>&#13;\n<div id=\"text-4-2\">&#13;\n<p>&#13;\nmu4e for email!&#13;\n<\/p>&#13;\n<\/div>&#13;\n<\/div>&#13;\n<\/div>&#13;\n\n\t<\/div><div class=\"entry-content\">\n\t\t<p>This is fourth in <a href=\"https:\/\/changelog.complete.org\/archives\/tag\/emacs2018\">a series on Emacs and org-mode<\/a>.<\/p>\n<p>Hopefully by now you\u2019ve started to see how powerful and useful org-mode is.  If you\u2019re like me, you\u2019re thinking:<\/p>\n<p>\u201cI\u2019d really like to have this in sync across all my devices.\u201d<\/p>\n<p>and, perhaps:<\/p>\n<p>\u201cCan I forward emails into org-mode?\u201d<\/p>\n<p>This being Emacs, the answers, of course, are \u201cYes.\u201d<\/p>\n<p><b>Syncing<\/b><\/p>\n<p>Since org-mode just uses text files, syncing is pretty easily accomplished using any number of tools.  I use git with git-remote-gcrypt.  Due to some limitations of git-remote-gcrypt, each machine  tends to push to its own branch, and to master on command.  Each machine merges from all the other branches and pushes the result to master after a merge.  A cron job causes pushes to the machine\u2019s branch to happen, and a bit of elisp coordinates it all \u2014 making sure to save buffers before a sync, refresh them from disk after, etc.<\/p>\n<p>The code for this post is somewhat more extended, so I will be linking to it on github rather than posting inline.<\/p>\n<p>I have a directory $HOME\/org where all my org-stuff lives.  In ~\/org lives <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/blob\/master\/emacs\/org-tools\/Makefile\">a Makefile<\/a> that handles the syncing.  It defines these targets:<\/p>\n<ul><li>push: adds, commits, and pushes to a branch named after the machine\u2019s hostname<\/li>\n<li>fetch: does a simple git fetch<\/li>\n<li>sync: adds, commits, pulls remote changes, merges, and (assuming the merge was successful) pushes to the branch named after the machine\u2019s hostname plus master<\/li>\n<\/ul><p>Now, in my user\u2019s crontab, I have this:<\/p>\n<pre>&#13;\n*\/15   *   *  *   *      make -C $HOME\/org push fetch 2&gt;&amp;1 | logger --tag 'orgsync'&#13;\n<\/pre>\n<p>The <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/blob\/master\/emacs\/org-tools\/emacs-config.org\">accompanying elisp code<\/a> defines a shortcut (C-c s) to cause a sync to occur.  Thanks to the cronjob, as long as files were saved \u2014 even if I didn\u2019t explicitly sync on the other boxen \u2014 they\u2019ll be pulled in.<\/p>\n<p>I have found this setup to work really well.<\/p>\n<p><b>Emailing to org-mode<\/b><\/p>\n<p>Before going down this path, one should ask the question: do you really need it?  I use org-mode with mu4e, and the integration is excellent; any org task can link to an email by message-id, and this is ideal \u2014 it lets a person do things like make a reminder to reply to a message in a week.<\/p>\n<p>However, org is not just about reminders.  It\u2019s also a knowledge base, authoring system, etc.  And, not all of my mail clients use mu4e.  (Note: things like MobileOrg exist for mobile devices).  I don\u2019t actually use this as much as I thought I would, but it has its uses and I thought I\u2019d document it here too.<\/p>\n<p>Now I didn\u2019t want to just be able to accept plain text email.  I wanted to be able to handle attachments, HTML mail, etc.  This quickly starts to sound problematic \u2014 but with tools like ripmime and pandoc, it\u2019s not too bad.<\/p>\n<p>The first step is to set up some way to get mail into a specific folder.  A plus-extension, special user, whatever.  I then use a <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/blob\/master\/emacs\/org-tools\/fetchmailrc.orgmail\">fetchmail configuration<\/a> to pull it down and run my <a href=\"https:\/\/github.com\/jgoerzen\/public-snippets\/blob\/master\/emacs\/org-tools\/insorgmail\">insorgmail<\/a> script.<\/p>\n<p>This script is where all the interesting bits happen.  It starts with ripmime to process the message.  HTML bits are converted from HTML to org format using pandoc.  an org hierarchy is made to represent the structure of the email as best as possible.  emails can get pretty complicated, with HTML and the rest, but I have found this does an acceptable job with my use cases.<\/p>\n<p><b>Up next\u2026<\/b><\/p>\n<p>My last post on org-mode will talk about using it to write documents and prepare slides \u2014 a use for which I found myself surprisingly pleased with it, but which needed a bit of tweaking.<\/p>\n\t<\/div><div class=\"entry-content\">\n\t\t<p>This is third in <a href=\"https:\/\/changelog.complete.org\/archives\/tag\/emacs2018\">a series on Emacs and org-mode<\/a>.<\/p>\n<p><b>Todo tracking and keywords<\/b><\/p>\n<p>When using org-mode to <a href=\"https:\/\/orgmode.org\/guide\/TODO-Items.html#TODO-Items\">track your TODOs<\/a>, it can have multiple states.  You can press <code>C-c C-t<\/code> for a quick shift between states.  I have set this:<\/p>\n<pre>&#13;\n(setq org-todo-keywords '(&#13;\n  (sequence \"TODO(t!)\" \"NEXT(n!)\" \"STARTED(a!)\" \"WAIT(w@\/!)\" \"OTHERS(o!)\" \"|\" \"DONE(d)\" \"CANCELLED(c)\")&#13;\n))&#13;\n<\/pre>\n<p>Here, I set up 5 states that are for a task that is not yet done: TODO, NEXT, STARTED, WAIT, and OTHERS.  Each has a single-character shortcut (t, n, a, etc).  The states after the pipe symbol are ones that are considered \u201cdone\u201d.  I have two: DONE (for things that I have done) and CANCELED (for things that I haven\u2019t done, but for whatever reason, won\u2019t).<\/p>\n<p>The exclamation mark means to log the time when an item was changed to a state.  I don\u2019t add this to the done states because those are already logged anyhow.  The @ sign means to prompt for a reason; so when switching to WAIT, org-mode will ask me why and add this to the note.<\/p>\n<p>Here\u2019s an example of an entry that has had some state changes:<\/p>\n<pre>&#13;\n** DONE This is a test&#13;\n   CLOSED: [2018-03-02 Fri 03:05]&#13;\n  &#13;\n   - State \"DONE\"       from \"WAIT\"       [2018-03-02 Fri 03:05]&#13;\n   - State \"WAIT\"       from \"TODO\"       [2018-03-02 Fri 03:05] \\\\&#13;\n     waiting for pigs to fly&#13;\n   - State \"TODO\"       from \"NEXT\"       [2018-03-02 Fri 03:05]&#13;\n   - State \"NEXT\"       from \"TODO\"       [2018-03-02 Fri 03:05]&#13;\n<\/pre>\n<p>Here, the most recent items are on top.<\/p>\n<p><b>Agenda mode, schedules, and deadlines<\/b><\/p>\n<p>When you\u2019re in a todo item, C-c C-s or C-c C-d can set a schedule or a deadline for it, respectively.  These show up in agenda mode.  The difference is in intent and presentation.  A schedule is something that you expect to work on at around a time, while a deadline is something that is due at a specific time.  By default, the agenda view will start warning you about deadline items in advance.<\/p>\n<p>And while we\u2019re at it, the <a href=\"https:\/\/orgmode.org\/guide\/Agenda-Views.html#Agenda-Views\">agenda view<\/a> will show you the items that you have coming up, offers a nice way to search for items based on plain text or tags, and handles bulk manipulation of items even across multiple files.  I covered setting the files for agenda mode in <a href=\"https:\/\/changelog.complete.org\/archives\/9865-emacs-2-introducing-org-mode\">part 2<\/a> of this series.<\/p>\n<p><b>Tags<\/b><\/p>\n<p>Of course org-mode has <a href=\"https:\/\/orgmode.org\/guide\/Tags.html#Tags\">tags<\/a>.  You can quickly set them with C-c C-q.<\/p>\n<p>You can set shortcuts for tags you might like to use often.  Perhaps something like this:<\/p>\n<pre>&#13;\n  (setq org-tag-persistent-alist &#13;\n        '((\"@phone\" . ?p) &#13;\n          (\"@computer\" . ?c) &#13;\n          (\"@websurfing\" . ?w)&#13;\n          (\"@errands\" . ?e)&#13;\n          (\"@outdoors\" . ?o)&#13;\n          (\"MIT\" . ?m)&#13;\n          (\"BIGROCK\" . ?b)&#13;\n          (\"CONTACTS\" . ?C)&#13;\n          (\"INBOX\" . ?i)&#13;\n          ))&#13;\n<\/pre>\n<p>You can also add tags to this list on a per-file basis, and also set tags for something on a per-file basis.  I use that for my inbox.org and email.org files to set an INBOX tag.  I can then review all items tagged INBOX from the agenda view each day, and the simple act of refiling them into other files will cause them to lost the INBOX tag.<\/p>\n<p><b>Refiling<\/b><\/p>\n<p>\u201cRefiling\u201d is moving things around, either within a file or elsewhere.  It has completion using your headlines.  <code>C-c C-w<\/code> does this.  I like these settings:<\/p>\n<pre>&#13;\n(setq org-outline-path-complete-in-steps nil)         ; Refile in a single go&#13;\n(setq org-refile-use-outline-path 'file)&#13;\n<\/pre>\n<p><b>Archiving<\/b><\/p>\n<p>After awhile, you\u2019ll get your files all cluttered with things that are done.  org-mode has an <a href=\"https:\/\/orgmode.org\/guide\/Archiving.html#Archiving\">archive<\/a> feature to move things out of your main .org files and into some other files for future reference.  If you have your org files in git or something, you may wish to delete these other files since you\u2019d have things in history anyhow, but I find them handy for grepping and searching.<\/p>\n<p>I periodically want to go through and archive everything in my files.  Based on a <a href=\"https:\/\/stackoverflow.com\/questions\/6997387\/how-to-archive-all-the-done-tasks-using-a-single-command\">stackoverflow discussion<\/a>, I have this code:<\/p>\n<pre>&#13;\n(defun org-archive-done-tasks ()&#13;\n  (interactive)&#13;\n  (org-map-entries&#13;\n   (lambda ()&#13;\n     (org-archive-subtree)&#13;\n     (setq org-map-continue-from (outline-previous-heading)))&#13;\n   \"\/DONE\" 'file)&#13;\n  (org-map-entries&#13;\n   (lambda ()&#13;\n     (org-archive-subtree)&#13;\n     (setq org-map-continue-from (outline-previous-heading)))&#13;\n   \"\/CANCELLED\" 'file)&#13;\n)&#13;\n<\/pre>\n<p>This is based on <a href=\"https:\/\/stackoverflow.com\/a\/27043756\">a particular answer<\/a> \u2014 see the comments there for some additional hints.  Now you can run <code>M-x org-archive-done-tasks<\/code> and everything in the current file marked DONE or CANCELED will be pulled out into a different file.<\/p>\n<p><b>Up next<\/b><\/p>\n<p>I\u2019ll wrap up org-mode with a discussion of automatically receiving emails into org, and syncing org between machines.<\/p>\n<p><b>Resources to accompany this article<\/b><\/p>\n<ul><li><a href=\"https:\/\/orgmode.org\/orgcard.pdf\">Printable org-mode reference card<\/a><\/li>\n<\/ul>\t<\/div><div class=\"entry-content\">\n\t\t<p>In <a href=\"https:\/\/changelog.complete.org\/archives\/9861-emacs-1-ditching-a-bunch-of-stuff-and-moving-to-emacs-and-org-mode\">my first post in my series on Emacs<\/a>, I described returning to Emacs after over a decade of vim, and org-mode being the reason why.<\/p>\n<p>I really am astounded at the usefulness, and simplicity, of org-mode.  It is really a killer app.<\/p>\n<p><b>So what exactly is org-mode?<\/b><\/p>\n<p>I wrote yesterday:<\/p>\n<blockquote><p>It\u2019s an information organization platform. Its website says \u201cYour life in plain text: Org mode is for keeping notes, maintaining TODO lists, planning projects, and authoring documents with a fast and effective plain-text system.\u201d<\/p><\/blockquote>\n<p>That\u2019s true, but doesn\u2019t quite capture it.  org-mode is a toolkit for you to organize things.  It has reasonable out-of-the-box defaults, but it\u2019s designed throughout for you to customize.<\/p>\n<p>To highlight a few things:<\/p>\n<ul><li><b>Maintaining TODO lists<\/b>: items can be scattered across org-mode files, contain attachments, have tags, deadlines, schedules.  There is a convenient \u201cagenda\u201d view to show you what needs to be done.  Items can repeat.<\/li>\n<li><b>Authoring documents<\/b>: org-mode has special features for generating HTML, LaTeX, slides (with LaTeX beamer), and all sorts of other formats.  It also supports direct evaluation of code in-buffer and literate programming in virtually any Emacs-supported language.  If you want to bend your mind on this stuff, read <a href=\"http:\/\/www.howardism.org\/Technical\/Emacs\/literate-devops.html\">this article on literate devops<\/a>.  The <a href=\"https:\/\/orgmode.org\/worg\/\">entire Worg website<\/a><br \/>\nis made with org-mode.<\/li>\n<li><b>Keeping notes<\/b>: yep, it can do that too.  With full-text search, cross-referencing by file (as a wiki), by UUID, and even into other systems (into mu4e by Message-ID, into ERC logs, etc, etc.)<\/li>\n<\/ul><p><b>Getting started<\/b><\/p>\n<p>I highly recommend watching <a href=\"https:\/\/www.youtube.com\/watch?v=oJTwQvgfgMM\">Carsten Dominik\u2019s excellent Google Talk on org-mode<\/a>.  It is an excellent introduction.<\/p>\n<p>org-mode is included with Emacs, but you\u2019ll often want a more recent version.  Debian users can <code>apt-get install org-mode<\/code>, or it comes with the Emacs packaging system; <code>M-x package-install RET org-mode RET<\/code> may do it for you.<\/p>\n<p>Now, you\u2019ll probably want to start with the org-mode compact guide\u2019s <a href=\"https:\/\/orgmode.org\/guide\/Introduction.html#Introduction\">introduction section<\/a>, noting in particular to set the keybindings mentioned in the <a href=\"https:\/\/orgmode.org\/guide\/Activation.html#Activation\">activation section<\/a>.<\/p>\n<p><b>A good tutorial\u2026<\/b><\/p>\n<p>I\u2019ve linked to a number of excellent tutorials and introductory items; this post is not going to serve as a tutorial.  There are two good videos linked at the end of this post, in particular.<\/p>\n<p><b>Some of my configuration<\/b><\/p>\n<p>I\u2019ll document some of my configuration here, and go into a bit of what it does.  This isn\u2019t necessarily because you\u2019ll want to copy all of this verbatim \u2014 but just to give you a bit of an idea of some of what can be configured, an idea of what to look up in the manual, and maybe a reference for \u201cnow how do I do that?\u201d  <\/p>\n<p>First, I set up Emacs to work in UTF-8 by default.<br \/><code><br \/>\n(prefer-coding-system 'utf-8)<br \/>\n(set-language-environment \"UTF-8\")<br \/><\/code><\/p>\n<p>org-mode can follow URLs.  By default, it opens in Firefox, but I use Chromium.<br \/><code><br \/>\n(setq browse-url-browser-function 'browse-url-chromium)<br \/><\/code><\/p>\n<p>I set the basic key bindings as documented in the Guide, plus configure the M-RET behavior.<\/p>\n<p><code>(global-set-key \"\\C-cl\" 'org-store-link)<br \/>\n(global-set-key \"\\C-ca\" 'org-agenda)<br \/>\n(global-set-key \"\\C-cc\" 'org-capture)<br \/>\n(global-set-key \"\\C-cb\" 'org-iswitchb)<\/code><\/p>\n<p>(setq org-M-RET-may-split-line nil)<br \/><\/p>\n<p><b>Configuration: Capturing<\/b><\/p>\n<p>I can press <code>C-c c<\/code> from anywhere in Emacs.  It will <a href=\"https:\/\/orgmode.org\/guide\/Capture.html#Capture\">capture something for me<\/a>, and include a link back to whatever I was working on.<\/p>\n<p>You can define <a href=\"https:\/\/orgmode.org\/guide\/Capture-templates.html#Capture-templates\">capture templates<\/a> to set how this will work.  I am going to keep two journal files for general notes about meetings, phone calls, etc.  One for personal, one for work items.  If I press <code>C-c c j<\/code>, then it will capture a personal item.  The %a in all of these includes the link to where I was (or a link I had stored with <code>C-c l<\/code>).<\/p>\n<pre>&#13;\n(setq org-default-notes-file \"~\/org\/tasks.org\")&#13;\n(setq org-capture-templates&#13;\n      '(&#13;\n        (\"t\" \"Todo\" entry (file+headline \"inbox.org\" \"Tasks\")&#13;\n         \"* TODO %?\\n  %i\\n  %u\\n  %a\")&#13;\n        (\"n\" \"Note\/Data\" entry (file+headline \"inbox.org\" \"Notes\/Data\")&#13;\n         \"* %?   \\n  %i\\n  %u\\n  %a\")&#13;\n        (\"j\" \"Journal\" entry (file+datetree \"~\/org\/journal.org\")&#13;\n         \"* %?\\nEntered on %U\\n %i\\n %a\")&#13;\n        (\"J\" \"Work-Journal\" entry (file+datetree \"~\/org\/wjournal.org\")&#13;\n         \"* %?\\nEntered on %U\\n %i\\n %a\")&#13;\n        ))&#13;\n(setq org-irc-link-to-logs t)&#13;\n<\/pre>\n<p>I like to link by UUIDs, which lets me move things between files without breaking locations.  This helps generate UUIDs when I ask Org to store a link target for future insertion.<\/p>\n<p><code><br \/>\n(require 'org-id)<br \/>\n(setq org-id-link-to-org-use-id 'create-if-interactive)<br \/><\/code><\/p>\n<p><b>Configuration: agenda views<\/b><\/p>\n<p>I like my week to start on a Sunday, and for org to note the time when I mark something as done.<\/p>\n<p><code><br \/>\n(setq org-log-done 'time)<br \/>\n(setq org-agenda-start-on-weekday 0)<br \/><\/code><\/p>\n<p><b>Configuration: files and refiling<\/b><\/p>\n<p>Here I tell it what files to use in the agenda, and to add a few more to the plain text search.  I like to keep a general inbox (from which I can move, or \u201crefile\u201d, content), and then separate tasks, journal, and knowledge base for personal and work items.<\/p>\n<pre>&#13;\n  (setq org-agenda-files (list \"~\/org\/inbox.org\"&#13;\n                               \"~\/org\/email.org\"&#13;\n                               \"~\/org\/tasks.org\"&#13;\n                               \"~\/org\/wtasks.org\"&#13;\n                               \"~\/org\/journal.org\"&#13;\n                               \"~\/org\/wjournal.org\"&#13;\n                               \"~\/org\/kb.org\"&#13;\n                               \"~\/org\/wkb.org\"&#13;\n  ))&#13;\n  (setq org-agenda-text-search-extra-files&#13;\n        (list \"~\/org\/someday.org\"&#13;\n              \"~\/org\/config.org\"&#13;\n  ))&#13;\n&#13;\n  (setq org-refile-targets '((nil :maxlevel . 2)&#13;\n                             (org-agenda-files :maxlevel . 2)&#13;\n                             (\"~\/org\/someday.org\" :maxlevel . 2)&#13;\n                             (\"~\/org\/templates.org\" :maxlevel . 2)&#13;\n                             )&#13;\n        )&#13;\n(setq org-outline-path-complete-in-steps nil)         ; Refile in a single go&#13;\n(setq org-refile-use-outline-path 'file)&#13;\n<\/pre>\n<p><b>Configuration: Appearance<\/b><\/p>\n<p>I like a pretty screen.  After you\u2019ve gotten used to org a bit, you might try this.<\/p>\n<pre>&#13;\n(require 'org-bullets)&#13;\n(add-hook 'org-mode-hook&#13;\n          (lambda ()&#13;\n            (org-bullets-mode t)))&#13;\n(setq org-ellipsis \"\u2935\")&#13;\n<\/pre>\n<p><b>Coming up next\u2026<\/b><\/p>\n<p>This hopefully showed a few things that org-mode can do.  Coming up next, I\u2019ll cover how to customize TODO keywords and tags, archiving old tasks, forwarding emails to org-mode, and using git to synchronize between machines.<\/p>\n<p>You can also see a <a href=\"https:\/\/changelog.complete.org\/archives\/tag\/emacs2018\">list of all articles in this series<\/a>.<\/p>\n<p><b>Resources to accompany this article<\/b><\/p>\n<ul><li><a href=\"https:\/\/orgmode.org\/guide\/\">org-mode compact guide<\/a><\/li>\n<li><a href=\"https:\/\/orgmode.org\/org.html\">Full org manual<\/a><\/li>\n<li><a href=\"https:\/\/orgmode.org\/worg\/org-tutorials\/\">List of tutorials<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=oJTwQvgfgMM\">Carsten Dominik\u2019s excellent Google Talk on org-mode<\/a><\/li>\n<li>A great intro: <a href=\"http:\/\/cachestocaches.com\/2016\/9\/my-workflow-org-agenda\/\">\u201cMy workflow with org-agenda\u201d<\/a><\/li>\n<li><a href=\"http:\/\/doc.norang.ca\/org-mode.html\">A more detailed intro<\/a><\/li>\n<li><a href=\"https:\/\/orgmode.org\/worg\/org-tutorials\/orgtutorial_dto.html\">David O\u2019Toole\u2019s org-mode tutorial<\/a><\/li>\n<li><a href=\"https:\/\/www.youtube.com\/watch?v=SzA2YODtgK4\">Getting Started with org-mode video by Harry Schwartz<\/a> (another excellent watch).  He also has his <a href=\"https:\/\/github.com\/hrs\/dotfiles\/blob\/master\/emacs\/.emacs.d\/configuration.org\">example config files<\/a> available.<\/li>\n<\/ul>\t<\/div><div class=\"entry-content\">\n\t\t<p>I\u2019ll admit it.  After over a decade of vim, I\u2019m hooked on <a href=\"https:\/\/www.gnu.org\/software\/emacs\/\">Emacs<\/a>.<\/p>\n<p>I\u2019ve long had this frustration over how to organize things.  I\u2019ve followed approaches like <a href=\"https:\/\/gettingthingsdone.com\/\">GTD<\/a> and <a href=\"https:\/\/zenhabits.net\/zen-to-done-the-simple-productivity-e-book\/\">ZTD<\/a>, but things like email or large files are really hard to organize.<\/p>\n<p>I had been using Asana for tasks, Evernote for notes, Thunderbird for email, a combination of ikiwiki and some other items for a personal knowledge base, and various files in an archive directory on my PC.  When my new job added Slack to the mix, that was finally the last straw.<\/p>\n<p>A lot of todo-management tools integrate with email \u2014 poorly.  When you want to do something like \u201cremind me to reply to this in a week\u201d, a lot of times that\u2019s impossible because the tool doesn\u2019t store the email in a fashion you can easily reply to.  And that problem is even worse with Slack.<\/p>\n<p>It was right around then that I stumbled onto <a href=\"https:\/\/www.youtube.com\/watch?v=oJTwQvgfgMM\">Carsten Dominik\u2019s Google Talk on org-mode<\/a>.  Carsten was the author of org-mode, and although the talk is 10 years old, it is still highly relevant.<\/p>\n<p>I\u2019d stumbled across <a href=\"https:\/\/orgmode.org\/\">org-mode<\/a> before, but each time I didn\u2019t really dig in because I had the reaction of \u201can outliner?  But I need a todo list.\u201d  Turns out I was missing out.  org-mode is all that.<\/p>\n<p><b>Just what IS Emacs?  And org-mode?<\/b><\/p>\n<p>Emacs grew up as a text editor.  It still is, and that heritage is definitely present throughout.  But to say Emacs is an editor would be rather unfair.<\/p>\n<p>Emacs is something more like a platform or a toolkit.  Not only do you have source code to it, but the very configuration is a program, and there are hooks all over the place.  It\u2019s as if it was super easy to write a Firefox plugin.  A couple lines, and boom, behavior changed.<\/p>\n<p>org-mode is very similar.  Yes, it\u2019s an outliner, but that\u2019s not really what it is.  It\u2019s an information organization platform.  Its website says \u201cYour life in plain text: Org mode is for keeping notes, maintaining TODO lists, planning projects, and authoring documents with a fast and effective plain-text system.\u201d<\/p>\n<p><b>Capturing<\/b><\/p>\n<p>If you\u2019ve ever read productivity guides based on GTD, one of the things they stress is effortless capture of items.  The idea is that when something pops into your head, get it down into a trusted system quickly so you can get on with what you were doing.  org-mode has a capture system for just this.  I can press <code>C-c c<\/code> from anywhere in Emacs, and up pops a spot to type my note.  But, critically, automatically embedded in that note is a link back to what I was doing when I pressed <code>C-c c<\/code>.  If I was editing a file, it\u2019ll have a link back to that file and the line I was on.  If I was viewing an email, it\u2019ll link back to that email (by Message-Id, no less, so it finds it in any folder).  Same for participating in a chat, or even viewing another org-mode entry.<\/p>\n<p>So I can make a note that will remind me in a week to reply to a certain email, and when I click the link in that note, it\u2019ll bring up the email in my mail reader \u2014 even if I subsequently archived it out of my inbox.<\/p>\n<p>YES, this is what I was looking for!<\/p>\n<p><b>The tool suite<\/b><\/p>\n<p>Once you\u2019re using org-mode, pretty soon you want to integrate everything with it.  There are browser plugins for capturing things from the web.  Multiple Emacs mail or news readers integrate with it.  ERC (IRC client) does as well.  So I found myself switching from Thunderbird and mairix+mutt (for the mail archives) to mu4e, and from xchat+slack to ERC.<\/p>\n<p>And wouldn\u2019t you know it, I liked each of those Emacs-based tools <b>better<\/b> than the standalone they replaced.<\/p>\n<p>A small side tidbit: I\u2019m using OfflineIMAP again!  I even used it with GNUS way back when.<\/p>\n<p><b>One Emacs process to rule them<\/b><\/p>\n<p>I used to use Emacs extensively, way back.  Back then, Emacs was a \u201clarge\u201d program.  (Now my battery status applet literally uses more RAM than Emacs).  There was this problem of startup time back then, so there was a way to connect to a running Emacs process.<\/p>\n<p>I like to spawn programs with Mod-p (an xmonad shortcut to a dzen menubar, but Alt-F2 in more traditional DEs would do the trick).  It\u2019s convenient to not run several emacsen with this setup, so you don\u2019t run into issues with trying to capture to a file that\u2019s open in another one.  The solution is very simple: I created a script, named it <code>em<\/code>, and put it on my path.  All it does is this:<\/p>\n<p><code><br \/>\n#!\/bin\/bash<br \/>\nexec emacsclient -c -a \"\" \"$@\"<br \/><\/code><\/p>\n<p>It creates a new emacs process if one doesn\u2019t already exist; otherwise, it uses what you\u2019ve got.  A bonus here: parameters such as <code>-nw<\/code> work just fine, so it really acts just as if you\u2019d typed <code>emacs<\/code> at the shell prompt.  It\u2019s a suitable setting for <code>EDITOR<\/code>.<\/p>\n<p><b>Up next\u2026<\/b><\/p>\n<p>I\u2019ll be talking about my use of, and showing off configurations for:<\/p>\n<ul><li>org-mode, including syncing between computers, capturing, agenda and todos, files, linking, keywords and tags, various exporting (slideshows), etc.<\/li>\n<li>mu4e for email, including multiple accounts, bbdb integration<\/li>\n<li>ERC for IRC and IM<\/li>\n<\/ul><p>You can also see a <a href=\"https:\/\/changelog.complete.org\/archives\/tag\/emacs2018\">list of all articles in this series<\/a>.<\/p>\n\t<\/div>","created_at":"2018-11-07T16:42:46+0000","updated_at":"2018-11-07T16:42:46+0000","published_by":["John Goerzen","John Goerzen","John Goerzen","John Goerzen","John Goerzen"],"annotations":[],"mimetype":"text\/html","reading_time":24,"domain_name":"changelog.complete.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":270,"title":"Stripe Atlas: Guide to networking for founders","url":"https:\/\/stripe.com\/atlas\/guides\/networking","content":"<p>Networking isn\u2019t as mysterious as it is made out to be. It is\u00a0simply talking intentionally to someone who could potentially help you in\nbusiness.<\/p><p>That\u2019s\u00a0it.<\/p><p>It means being conscious of our interactions with others. You should\u00a0seek to understand people\/relationships better so you don\u2019t miss out on\na good relationship, a chance to help someone, or a chance to get help\nyourself. It\u2019s about being aware of how you come off to others. It\u2019s\nabout building community.<\/p><p>Asking \u201cWhy is networking important in business?\u201d is like asking \u201cWhy\u00a0are relationships important in business?\u201d Networking is simply the\ninitial stage in your relationship with others.<\/p><p>Networking has the capability to build walls, but it can also open doors\u00a0and help those with less privilege. You should always be aware that\nnetworking at its best is still a closed network, and doing your best to\nopen it up as much as possible and lift others up changes lives.<\/p><p>I co-founded Binomial, a company that makes a GPU-friendly image\u00a0compressor called Basis that improves download sizes and performance in\napps. I started my business a few years ago. When I started, I had two\nmonths of savings until I had to get a job, no experience, and no good\nplans.<\/p><p>Networking was the key. I had hundreds and hundreds of\u00a0conversations with potential customers, other entrepreneurs, experienced\nbusinesspeople, experienced programmers, and more.<\/p><p>I believe that was one of the single biggest factors in getting me to\u00a0where I am now\u2014running a successful company that serves some of the\nbiggest names in games, virtual reality, mapping applications,\ninteractive web apps, and more.<\/p><h2 id=\"what-are-different-ways-to-network\">What are different ways to network?<\/h2><p>\u201cWhat are all the ways to meet humans?\u201d is a similar question! Sometimes\u00a0we have a goal in mind when we interact with people. For instance, if\nyou want to make friends you can play video games with you might frequent\nan arcade or video game conference. If you want to date, you might go to\na bar or a singles mixer. If you want to organize a volunteer effort,\nyou might try being vocal about that effort while you volunteer. You\ncould just go to a public park for each of these efforts and start\ntalking to strangers, but you might be more effective (and bother others\nless!) with a more thoughtful approach.<\/p><p>In our case we want to build a tech business, one that is a positive\u00a0force in our communities with happy customers that allows us to make a\ngood living.<\/p><p>It\u2019s helpful to zoom out for an overview, and I\u2019ll go in more depth\u00a0later:<\/p><ul><li>\n    <p>Events: either online or in\u00a0person<\/p>\n    <ul><li>\n        <p>Conferences<\/p>\n      <\/li>\n      <li>\n        <p>Local professional\u00a0events<\/p>\n      <\/li>\n      <li>\n        <p>Social events people in the industry\u00a0frequent<\/p>\n      <\/li>\n    <\/ul><\/li>\n  <li>\n    <p>Social media (Twitter, Facebook, Instagram, Tumblr\u2026): both posting to and interacting with people\u00a0one-on-one<\/p>\n  <\/li>\n  <li>\n    <p>Online communities: Slack channels, open source projects, mailing\u00a0lists\u2026<\/p>\n  <\/li>\n  <li>\n    <p>Informational\u00a0interviews<\/p>\n  <\/li>\n  <li>\n    <p>Mentoring people in the\u00a0industry<\/p>\n  <\/li>\n  <li>\n    <p>Volunteering related to the\u00a0industry<\/p>\n  <\/li>\n  <li>\n    <p>Organizing your own events and communities: big or small, public or\u00a0private<\/p>\n  <\/li>\n  <li>\n    <p>\u2026 and\u00a0more!<\/p>\n  <\/li>\n<\/ul><h2 id=\"a-note-on-mentoring-giving-back\">A note on mentoring &amp; giving back<\/h2><p>I am passionate about making sure there\u2019s a healthy balance between\u00a0\u201cgiving\u201d and \u201cgetting\u201d in networking. The exact balance may differ\nbetween someone first starting out and someone who is very established,\nbut should always be kept in mind.<\/p><p>You should care about others and help others selflessly. I believe you\u00a0should do this without an explanation needed. You should do this even\nif it didn\u2019t help your business\u2014but it will.<\/p><p>You shouldn\u2019t start all your networking conversations with \u201cHello, I\u00a0need these things, can you do them for me?\u201d There are lots of layers to\nwhy that\u2019s a wrong blanket approach, which I\u2019ll dive into later, but\nhopefully you have an intuition for why it\u2019s not optimal. Since you\nwouldn\u2019t converse selfishly, you similarly shouldn\u2019t seek out\nconversations selfishly, based on how much you think people could help\nyou.<\/p><p>If you seek to create value for others, others will create value for you.\u00a0Nothing is more powerful than seeing reciprocity in action.<\/p><p>During the first year of running my business, I ran several fairly large\u00a0charitable initiatives to help others. I mentored people one-on-one,\norganized a mentoring program, and taught free classes in the community.<\/p><p>I thought these were purely charitable initiatives that wouldn\u2019t affect\u00a0the business. I was very, very wrong. In fact, helping others is one of\nthe single biggest drivers for new business I\u2019ve ever had.<\/p><p>Even having done it, it is difficult to fully measure the effect of it.\u00a0People talk, word gets around, you build a reputation.<\/p><p>But for me, it goes deeper than that: Balancing my time between helping\u00a0others and being helped\u2014and helping others without expecting anything\nin return\u2014feels good for the soul, and makes you a better human being\nin your interactions with the people who can help you.<\/p><p>If all you do is take, others will notice. You\u2019ll do worse in business\u00a0and be a less happy human being.<\/p><h2 id=\"considerations-for-different-industries-and-cultures\">Considerations for different industries and cultures<\/h2><p>Expectations are different across industries and\u00a0cultures.<\/p><p>Some industries will be more active on certain social media platforms.\u00a0Some conferences and events may be more or less popular. Some industries\nhave more prominent open source projects than others. Some industries\nare more used to networking with small business owners than others. Some\nmay have more private events you need to search out, others may have\nmany public events.<\/p><p>Industries vary not only in their tangible reality but also in their\u00a0intangible cultures. In video games, for instance, the culture is\nusually very casual\u2014people often show up in t-shirts and jeans, go\nout to drink beer, and socialize a lot outside of work. If you show up\nin a suit and speak with a formal tone, you\u2019ll be seen as strange\u2014but\nin other industries that won\u2019t be out of place.<\/p><p>Make sure you assess the culture of the industry you want to\u00a0target.<\/p><p>Set yourself up for success in a relationship by nailing the first impression,\u00a0online or offline.<\/p><h2 id=\"communicating-via-dress-and-appearance\">Communicating via dress and appearance<\/h2><p>We don\u2019t talk about fashion nearly enough in the tech industry. It is a\u00a0well-studied form of communication in itself. Many pretend to be\nindifferent to it, but because we are human there are still norms, and\nthose who are indifferent are either conforming to the norms\nunconsciously or at a disadvantage without realizing it.<\/p><p>I was a stylist at a clothing store for years. I would dress my clients\u00a0for job interviews, work, promotion conversations, and launch events. I\u2019d\nform long term relationships with them and watch just how big of an\nimpact their clothes &amp; appearance had on their career.<\/p><p>Clients often started from a place of believing \u201cClothes are so superficial.\u00a0I don\u2019t like this, I don\u2019t want to care about such superficial things!\nPeople should like me for me!\u201d<\/p><p>No one can know your whole life story. Even your closest confidants only\nknow it in a rough outline compared to its actual detail. Your\nperformance of <em>you<\/em> is a compression of everything you are and have\u00a0experienced.<\/p><p>Your physical presentation is a compression about the compression. It helps\u00a0other people understand how you expect to be related to. Your project has\na README irrespective whether you want to be read or not.<\/p><p>Taking control of your first impression is not bowing down to the\u00a0expectations of others and losing sight of your identity\u2014it\u2019s\naccepting this reality of the world, being knowledgeable about it, and\nbeing deliberate in how you want to come off to others. (And sometimes,\nit means refusing to conform, with intentionality.)<\/p><p>The key is simply being educated\u2014we should never be opposed to\u00a0learning about our world. I\u2019ll also include practical, easy-to-follow\ntips for people who don\u2019t want to study this in-depth.<\/p><p>Tips:<\/p><p><strong>Dress a notch or two more professionally than the average person in\nthe group you\u2019re talking to<\/strong>. Generally, be cautious about getting\u00a0more casual than \u201cbusiness casual.\u201d Remember that a business owner\nis not held to the same standard as an employee, and dressing a\nbit more professionally sets you apart and commands respect while\nstill fitting in. (This advice also works for employees looking to\nget promoted: look the part you\u2019re trying out for.)<\/p><p><strong>Maintain your clothes appropriately.<\/strong> They are a professional\u00a0tool.<\/p><p><strong>Hair matters.<\/strong> Again, get a professional to help if possible\u2014a\u00a0hair stylist or barber will know what to do, and can adjust the\nlook to how much effort you want to put into maintaining it while\nstill keeping it good looking. You can help your stylist pick\nappropriate looks for you by showing them photos of any styles you\nlike, and by telling them what your industry is or what the\noccasion you\u2019re preparing for is.<\/p><p><strong>Skincare also matters.<\/strong> Study skincare routines, and even if you\u2019re\u00a0going for a no-makeup look you might consider makeup. Skincare and\nproper makeup can make a big impact on clearing up blemishes,\nsmoothing out skin, and highlighting good features so people focus\non you and you look good &amp; polished\u2014not distracted by that red\nspot on your face. Once again, professionals know this area best.\nFor makeup, Sephora and MAC are both examples of stores with\nstylists walking around, happy to give advice\u2014you can watch\nYouTube and read extensively, or you can just take an hour to get\nhigher quality advice.<\/p><p><strong>Fit is hands-down the most important thing for clothing.<\/strong> If you take\none thing away from this section, remember that. If you have an\nunusual body shape or height, consider getting a tailor. Try on\n<em>lots<\/em> of clothes until you find ones that fit. Fit is\u00a0everything.<\/p><p><strong>Whenever possible, get styled by a professional<\/strong>. This does not need\u00a0to be expensive. Department stores like Nordstrom have experienced\nprofessionals who are happy to give you help. (This is generally\nfree with the expectation that you\u2019ll buy some of the clothes.)\nStitchfix is an example of a styling service that is purely\nonline. You will not outperform a professional at styling\nyourself, for the same reason that a fashion stylist is unlikely\nto run your tech business better than you would. Professional\nstylists are students of subtlety in presentation and have honed\nthat ability for years.<\/p><h3 id=\"know-the-norms-break-them-judiciously\">Know the norms; break them judiciously<\/h3><p>What about breaking the rules and going against social norms in\u00a0appearances? As a woman in a heavily male-dominated industry, I know all\nabout going against norms in this sense. Staying within norms helps\ninsulate you from discrimination, but selectively transgressing can make\nyou memorable in a good way.<\/p><p>The general advice is to stick mostly with the norm, but go against it\u00a0in one way or slowly over time. The advanced advice is to get educated\nabout fashion and how you present so that you can go against lots of\nnorms in deliberate, intentional ways.<\/p><p>Consider this pervasive example from the tech\u00a0industry:<\/p><p><strong>The norm<\/strong>: Male-dominated, t-shirt and\u00a0jeans<\/p><p><strong>Good outfits which slightly tweak the\u00a0norm<\/strong>:<\/p><p>Masculine\u00a0look:<\/p><ul><li>\n    <p>Dark good quality t-shirt that fits well, dark jeans that fit well,\u00a0and leather dress shoes that can work with a casual look<\/p>\n  <\/li>\n  <li>\n    <p>Neutral blazer that\u2019s not too formal, t-shirt, dark jeans, and nice\u00a0non-athletic sneakers that go with the color scheme<\/p>\n  <\/li>\n  <li>\n    <p>Dress shirt under a dark sweater, well fitted jeans, leather dress\u00a0shoes on the casual side<\/p>\n  <\/li>\n<\/ul><p>Feminine\u00a0look:<\/p><ul><li>\n    <p>Blazer that\u2019s not too formal, nice t-shirt, dark jeans,\u00a0flats<\/p>\n  <\/li>\n  <li>\n    <p>Shirt dress that isn\u2019t too formal, leather belt, tights, leather\u00a0boots<\/p>\n  <\/li>\n  <li>\n    <p>Not-too-formal blouse, simple jewelry, jeans,\u00a0flats<\/p>\n  <\/li>\n<\/ul><h2 id=\"presenting-as-a-professional-online\">Presenting as a professional online<\/h2><p>In person, someone only has so many things to go off of to judge you\u00a0before they talk with you. Your dress, your appearance, body language.<\/p><p>But online there\u2019s so much more! You can fully expect someone to at\u00a0least glance at your avatar or profile before responding to you, and\nmany people will do a Google search as well. Let\u2019s touch on making a\ngood first impression\u2014before you even get to talk to someone!\u2014\nonline.<\/p><h3 id=\"website\">Website<\/h3><p>You should have a website for your company, and one for yourself as\u00a0well! Even if your company hasn\u2019t gotten any customers or hasn\u2019t made\nany money, even if it\u2019s an early idea, that website can really help you\nland those initial deals.<\/p><p>This is another opportunity for you to improve your results by\u00a0consulting with a professional. A designer can help you create (or\nselect) a site appropriate to your industry, business model, stage in\ncompany development, marketing strategy, and so on.<\/p><p>Some high-level website\u00a0tips:<\/p><p><strong>The general design of your website should communicate\nprofessionalism.<\/strong> Design also communicates the type of industry\u00a0you\u2019re in or even how established you are. Early in my business, I\nwas told that my company\u2019s website looked a lot like a personal\nblog as opposed to a professional, established company\u2014that\ncouldn\u2019t have been good for business! Just like you should be\nintentional about the extent to which you don\u2019t conform with\nfashion, you should be selective about how your website would\nstand out from peers in your industry.<\/p><p><strong>Make it obvious and explicit how a first-time visitor engages with\nyou.<\/strong> What is your <em>process<\/em>? Can they do a free evaluation of your\u00a0product? How can they contact you? What happens after they contact\nyou? What are the steps from contacting you to purchasing? Who do\nyou respond to (are you open to working with both small and big\ncompanies)? Following this advice can completely transform your\nbusiness\u2014you may be surprised at how many people don\u2019t want to\nreach out if they\u2019re not sure what\u2019ll happen next or when they can\nexpect to hear back.<\/p><p><strong>Remove barriers to getting to talk to you.<\/strong> Especially in the early\u00a0stages of a company, when you are optimizing for learning about\ncustomers, you should facilitate conversations in as many ways as\npossible.<\/p><p><strong>Logos logos logos!<\/strong> Put company logos of customers you\u2019ve worked with\u00a0on your website if you have them, front and center, easy to see.\nIf you don\u2019t have permission to do this but have customers,\ndescribe them in as much detail as you can.<\/p><p><strong>Highlight and amplify your the credentials of your team.<\/strong> Have you\u00a0spoken at conferences (especially about your business)? Have you\nworked at well-known companies in the past? Have you done some\nimpressive work? Citing these will help you build trust with your\ncustomers, and provides an icebreaker for initial conversations.<\/p><p>There are lots of social media platforms: Twitter, Instagram, Snapchat,\u00a0Facebook, LinkedIn, and more. Each network has a different userbase,\nencourages different user behavior, and attracts different kinds of\npeople and industries.<\/p><p>Being active on a social media account can be a significant time\u00a0investment. Be intentional about how you spend this time; you probably\ndo not want to be open an account literally everywhere. I sell a B2B\nengineering product in the game industry. Twitter works very well in my\nspace since a lot of technology professionals use it.<\/p><p>You should be mindful of the first impression a user will have when\u00a0encountering you on a social network: your profile, description, and the\ncontent you post.<\/p><p>A corporate presence on a social media account is very different than a\npersonal account. You almost certainly want a personal account that\nkeeps in mind your company\u2019s audience is watching; whether you should\nalso have a corporate presence is a separate question. Some people have\nmore success with this, some people have less\u2014I think it depends\nagain on your industry and customer base. Do your customers feel a need\nto trust and like <em>you<\/em>, or do they just want company updates? There\u2019s\u00a0no harm in getting a company account to test how well that does.<\/p><p>You can choose to not follow the below advice and still be successful\u2014\u00a0just like with clothing choices outside norms, the key to success is\nacting with intentionality. You should be educated about online\ndynamics, social norms, and the first impression different behaviors\ngive off. These tips are meant to be easy-to-follow, safe guidelines.<\/p><p>Social media first impression\u00a0tips:<\/p><p><strong>Ensure your recent posts generally include something about your industry or business. For instance, my\npinned tweet is always about my business.<\/strong> When I first started,\nI\u2019d make sure at least 50% of my posts were about my product,\nwork, or industry. It\u2019s still at about that ratio these days,\nthough I\u2019m not so deliberate about that anymore. Keep in mind that\nthis ratio varies depending on industry and social media\nplatform\u2014for instance, on LinkedIn you might want to keep that\nratio closer to 100%. A high signal-to-noise ratio tends to help\n<em>converting<\/em> interested users into followers when your content\nhappens to end up in front of them, for example via social\namplification like retweets. It says \u201cYes, you <em>are<\/em> in the right\nplace; following me <em>will<\/em> get you more of what you are interested\u00a0in.\u201d<\/p><p><strong>Just like in fashion, you can go against the norms thoughtfully<\/strong> if you\u00a0understand the reasons they exist and what your actions\ncommunicate.<\/p><p><strong>Pause and think before posting.<\/strong> Some social media platforms (ahem, Twitter, ahem) are known for\u00a0negativity and ranting. Your\nprofessional presence on social media is equivalent to you\nspeaking to others in-person at a tech meetup. Would a stranger\nfeel comfortable walking up and opening a conversation with you\nbased on what they overheard?<\/p><p><strong>Present your business as overall stable and doing well, at\nall stages<\/strong>. If your business looks like it is struggling or you\u00a0look needy and desperate, potential customers will back away.<\/p><p>What if you\u2019re not stable or doing well? There is a temptation to\u00a0bluster or shade the truth in business. You should never\ncompromise your ethics and you should be open about failures, but\ndecide when and how to do this. You can talk about failures\nwithout dwelling on them. Do so in a way that makes you appear\nstrong and positive, or feel free to address them only once you\u2019re\nin a better place.<\/p><p><strong>Be authentic, particularly on personal accounts.<\/strong> Part of social media\u00a0is about building trust, showing that you\u2019re a likable person who\nis nice to work with. You should feel free to have your\npersonality and voice shine on your account.<\/p><p>If you need a private outlet that you would not feel comfortable\u00a0with customers seeing, make an intentionally discrete account for\nuse with close confidants, and check privacy settings carefully.<\/p><p><strong>Don\u2019t forget about helping others.<\/strong> Social media is a powerful tool\u00a0for helping others and giving back, and doing this benefits us\nall.<\/p><h3 id=\"other-online-first-impressions\">Other online first impressions<\/h3><p><strong>Online communities<\/strong>: Slack groups, open source communities, etc.\u00a0People are watching you interact in online communities and\nwatching what you contribute to them.<\/p><p><strong>GitHub\/code repositories<\/strong>: If open source is important to your\u00a0company, your code repository may be someone\u2019s first impression\nof your business. Have excellent documentation. Strongly consider\nhaving a designed website for the project; Github\u2019s branding makes\npeople associate your code with Github and their brand impressions\nof Github, not with you. Make sure the person can get a good sense\nof what it does from a 6-second skim of the description. Have an\neasy to build\/pre-built demo or sample code. If you accept\ncontributions, make it clear how others can submit toward it.\nTreat all web presences of your company with the same seriousness\nthat you would treat your website\u2019s home page.<\/p><h2 id=\"creating-helpful-and-meaningful-content\">Creating helpful and meaningful content<\/h2><p>Creating content is an amazing way to get people to come to\u00a0you!<\/p><p><strong>Speaking at local events:<\/strong> Events relevant to your industry may\u00a0be plentiful locally, depending on where you live. They\u2019ve been\nreally helpful for me! Resources like meetup.com can help you find\nevents. You can also browse social media and ask people you know\nin the industry for recommendations. Speaking at an intimate local\nevents can be nicer than speaking at a conference because there\u2019s\nmuch less pressure\u2014the groups are typically pretty small and if\nyou want to speak you can just message the organizer and ask!\nWe\u2019ll talk a little bit about how to make \u201casks\u201d in the coming\nsections.<\/p><p><strong>Speaking at conferences:<\/strong> I started getting invitations to\u00a0conferences once I built up enough of a social media presence and\nhad speaking at local events under my belt, but you can also\nsubmit at an application to speak at these. Conferences are\nexcellent for building credibility and are especially useful for\nbuilding credibility with developers if you sell to engineers.<\/p><p><strong>Blogging:<\/strong> You can have a company blog or a personal blog. A lot\u00a0of the advice on social media first impressions applies here.\nBlogs can be very good for PR, depending on your industry\u2014for\ninstance, getting blog posts featured in prominent publications or\non aggregation sites like Hacker News has driven a lot of business\nto my company, and is something not possible with a social media\npost. I recommend sharing all blog posts through your social media\naccounts.<\/p><p><strong>Mailing list<\/strong>: A mailing list delivers exclusive content, which\u00a0isn\u2019t available to the world at large, and is therefore can be a\ncloser relationship than just interacting publicly over social\nmedia. You can send updates or newsletters via a mailing list;\noffering people the option of signing up for it is an excellent\nway to see who is interested and build a closer-knit community.<\/p><p><strong>Podcasts, online conferences, and online events:<\/strong> Either being a\u00a0guest or hosting your own podcast can be great for PR. Online\nconferences are very similar but usually include video and chat\n(think: webinars). These all have a lot in common with local\nevents\u2014being included either means just building up your social\nmedia presence and getting an invite or reaching out directly to\nthe organizers. Interview podcasts in particular structurally need\nto have new guests to continue running the show, so don\u2019t be shy\nabout pitching yourself to them; you\u2019re helping them out!<\/p><p><strong>Online communities:<\/strong> There are so many! From public forums to\u00a0private slack groups, you can also contribute here and post\ncontent.<\/p><p><strong>Hosting your own events and communities:<\/strong> We\u2019ll touch on this\u00a0later!<\/p><p>No matter your charm, success, social media following size, or amount of\u00a0brilliant content posted\u2014you will sometimes need to make the first\nmove. In sales, this is called an \u201cask\u201d\u2014you\u2019re straightforwardly requesting\nthat someone do a specific something on your behalf.<\/p><p>You may have your own preferred way to network, but this is one of my\u00a0favorite ways to break the ice.<\/p><p>An informational interview has a simple theme and intent: Person 1 asks\u00a0Person 2 to coffee or a video chat, and Person 1 spends that time asking\nquestions and learning from Person 2. It is not a direct sales call.\nNeither party expects immediate business as a result of the meeting; it\nis just about learning.<\/p><p>I have participated in hundreds of informational interviews, on both\u00a0sides of the table. They are fun to do, easy to line up, and extremely\nhelpful.<\/p><p>This is basically how I learned business\u2014I\u2019d ask experienced business\u00a0people out to coffee. They\u2019re an excellent way to learn: business,\ndesign, coding, any topic! The internet has lots of resources to learn\nfrom, but having someone with experience give you customized advice and\ncurated learning resources is invaluable.<\/p><p>It is also how I learned what not to do and what I didn\u2019t want, to my\nsurprise. I\u2019d sometimes ask someone I really admired out for coffee, and\nrealize that a path they went down or a decision they made was not as\nideal as it seemed. One-on-one conversations are much more candid and\nauthentic than the more polished, selective view you get from reading\ninterviews. (Remember, the successful people you admire <em>also<\/em> struggle\u00a0with the tension between being authentic and making a good professional\nimpression. You\u2019ve likely only seen their highlight reel.)<\/p><p>Let\u2019s not forget about customer interviews too! Doing informational\u00a0interviews with potential customers is an excellent idea\u2014they\u2019re much\neasier to get than a sales call in the early stages of business, they\nhelp shape your product to be better, and they can turn into real sales\ndown the road.<\/p><p>Lastly, pay it forward and return the favor\u2014be open to questions and\u00a0let people take you out for an informational interview!<\/p><p>How do you ask someone to an informational interview? Here are some\u00a0examples for asking over social media where your profile clearly\nintroduces yourself\u2014if writing over e-mail or you\u2019re not sure your\nprofile does the best job as an intro, add more context in the message\ntext.<\/p><h4 id=\"student-asking-for-chat\">Student asking for chat<\/h4><blockquote>\n  <p>Hello,<\/p>\n  <p>I\u2019m a student at Imaginary University and I\u2019m really interested in\u00a0computer graphics. I really admire your work\u2014I found your blog\npost on particle systems so interesting! I would love to get a chance\nto ask you some questions on getting a career in computer graphics and\nwhat you do\u2014would you have time for a 30-minute Skype chat\nsometime?<\/p>\n  <p>Best,<\/p>\n  <p>Stephanie<\/p>\n  <p>[link to some kind of proof of work here\u2014demos you\u2019ve done, website, github,\u00a0etc]<\/p>\n<\/blockquote><h4 id=\"transitioning-careers\">Transitioning careers<\/h4><blockquote>\n  <p>Hello,<\/p>\n  <p>I\u2019m looking to transition into a career in web development (I\u2019m\u00a0currently a photographer!) and I have learned so much from following\nyou on Twitter\u2014your blog post with advice on [technology topic] was\nreally helpful! Would you be open to a Skype chat sometime? I\u2019d love\nto ask about career advice and what technologies you recommend\nlearning.<\/p>\n  <p>Best,<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><blockquote>\n  <p>Hello,<\/p>\n  <p>I\u2019m an aspiring entrepreneur and I\u2019ve loved reading your tweets,\u00a0they\u2019ve given me a lot of hope and ideas. I\u2019m currently working on\nbuilding out [brief description of company]. If you have time, I\u2019d\nlove to ask questions around bootstrapping a small business (that\u2019s\nbeen one of my biggest challenges) and how to scale\/grow in a smart\nway. Would you be free for a 30 minute Skype call?*<\/p>\n  <p>Stephanie<\/p>\n  <p>[Stephanie\u2019s website if she has\u00a0one]<\/p>\n<\/blockquote><blockquote>\n  <p>Hello,<\/p>\n  <p>Writing cause I really liked what you were saying on Twitter about\u00a0[technology topic they\u2019re experienced with]. This is something I\u2019d\nreally like to learn more about actually\u2014I work as an engineer at\nImaginary Company now but I never get to touch stuff like that\u2014\nwould you mind if I asked you some questions over Skype sometime?<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><h4 id=\"offering-unsolicited-help-to-someone-with-less-experience\">Offering unsolicited help to someone with less experience<\/h4><blockquote>\n  <p>Hello,<\/p>\n  <p>Saw your blog post on Technology Topic\u2014really enjoyed it! I\u2019ve\u00a0been working with that Technology Topic for several years now and just\nwanted to say if you ever need help feel free to ping me!<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><h2 id=\"meeting-people-at-conferences-events\">Meeting people at conferences &amp; events<\/h2><p>Conferences and events can feel like a hard thing to tackle when you\u2019re\u00a0not used to networking. (There are so many people, scary!)<\/p><p>But the beautiful thing about them is it is <em>expected<\/em> to network\u00a0professionally there. If you go up to a random person in a park and\nstart talking about your business idea, they are probably going to be\nconfused. But if you go up to a person at a tech event or conference and\nstart talking about your business idea, they\u2019re going to get excited\n(and probably relieved that you broke the ice first!)<\/p><p>A conference is typically a large event, planned in advance, that\nrequires travel and days blocked out of schedules. It\u2019s extremely normal\nto plan networking meetings and events in advance for conferences. <strong>I\u00a0usually start a month before the event; occasionally several months in\nadvance.<\/strong><\/p><p>How long should the meeting be? A 30-minute meeting does not leave much\u00a0room for relationship building, but you can still get a lot done and is\nideal for meeting with busy people or for when you\u2019re in the early stage\nof business and not many people know of you. An hour is good for giving\nbuffer between events and building up a relationship. A meeting over a\nmeal is typically longer than that.<\/p><p>Figuring out where to meet can be a challenge itself for large events. A\u00a0tip on that: coffee shops close to the conference can get quite\ncongested\u2014hotel lobbies have much more room and often have more open\nseats. If it\u2019s in a warm climate, it\u2019s also completely fine to meet at a\npark. If you need more privacy or even more professionalism, you can\nreserve a conference room, office, or professional space nearby for\nmeetings. If the person you\u2019re meeting with is busier than you are, feel\nfree to suggest a spot but always offer to meet where is most convenient\nfor them.<\/p><p>Since it\u2019s often the longest open timeslot, try to always grab a meal\u00a0with a group\u2014remember, people are there to network, so feel free to\ngather one or two people you\u2019d like to grab a meal with and tell them to\nfeel free to invite more of their friends\/contacts.<\/p><p>Be visible about meeting up and open to spur-of-the-moment invites.\u00a0Announce you\u2019ll be there on social media well in advance and again when\nyou get to the conference. Post updates on social media of talks you see\nor events you go to. Talk to people and ask where they\u2019ll be going next\nor what events are happening.<\/p><p>Feel free to message people asking to meet up! Here are some examples\u00a0for meeting up at conferences, again assuming you\u2019re asking over social\nmedia with a well-developed profile. Notice you need less introductory\ntext here, since it\u2019s very expected to network:<\/p><h4 id=\"meeting-people-you-otherwise-wouldnt-be-able-to\">Meeting people you otherwise wouldn\u2019t be able to<\/h4><blockquote>\n  <p>Hello,<\/p>\n  <p>I\u2019m a student at Imaginary University and I\u2019m really interested in\u00a0computer graphics, and I saw you were going to be at Technology\nConference! If you\u2019re free I\u2019d love to talk with you about getting\ninto the industry and your work in graphics\u2014would you have some\ntime?<\/p>\n  <p>Best,<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><h4 id=\"meeting-a-peer\">Meeting a peer<\/h4><blockquote>\n  <p>Hey,<\/p>\n  <p>I saw you\u2019re going to be at Technology Conference\u2014I\u2019ll be there\u00a0too Monday-Wednesday! I\u2019d love to grab coffee sometime if you have\nsome free time\u2014I work at Imaginary Company doing [technology\ndescription].<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><h4 id=\"starting-a-prospecting-conversation\">Starting a prospecting conversation<\/h4><blockquote>\n  <p>Hi,<\/p>\n  <p>I\u2019m building a company that does Problem-They-Need-Solved ([link to\u00a0company website]), and I\u2019d love to ask about your pipeline and see\nwhat features are important to you if you have some time at Technology\nConference. I can be flexible, just let me know if you have some time\nthat week!<\/p>\n  <p>Thanks,<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><h4 id=\"asking-for-advice\">Asking for advice<\/h4><blockquote>\n  <p>Hi,<\/p>\n  <p>I\u2019m building a small company right now and learning so much about\u00a0marketing and pricing. I really admire what you\u2019ve been able to build\nwith your company\u2014would you mind if I asked you some questions over\ncoffee at Technology Conference? Let me know if you have time open!<\/p>\n  <p>Best,<\/p>\n  <p>Stephanie<\/p>\n  <p>[link to company\u00a0website]<\/p>\n<\/blockquote><h3 id=\"comporting-yourself-at-events\">Comporting yourself at events<\/h3><p>What about for smaller events, or events within a conference? You can\u00a0always feel free to ask to meet up with someone in advance at an event,\nbut usually it\u2019s best to meet them before or after to give them time to\nnetwork in the short time of the event.<\/p><p>When you go to event, try to bring a business card, laptop with a demo\u00a0or video ready to play about your product\u2014any way to try to deepen a\nconversation or make it easier to keep in touch with someone.<\/p><p>Approaching people you don\u2019t know during an event can feel scary, and is\u00a0typically what people think of when they think of networking. I still\nget nervous about it and I do a massive amount of networking\u2014I much\nprefer meeting with people one-on-one or in small groups. When I do go\nto events I typically try to find the people standing alone, close my\neyes and take a deep breath, and introduce myself. I find that they\u2019re\nusually relieved that you broke the ice. Another way to approach someone\nis to stand in a group and attentively listen to the person speaking\u2014\ndon\u2019t worry, you won\u2019t look like you\u2019re intruding, this is quite normal\nat networking events.<\/p><p>One note on online events: These exist too! You\u2019ll typically see them in\u00a0the form of webinars and livestreams, originating on social media or in\nother online communities. Ask people you talk to if they know of any.\nApproaching people at these depends highly on the medium, but the same\ngeneral advice applies: feel free to join into conversations, direct\nmessage people you don\u2019t know, and remember to keep in touch.<\/p><p>Obviously doing informational interviews and meeting people at events\u00a0can originate from a social media interaction\u2014but how do you approach\nsomeone on social media in the first place? What\u2019s the general etiquette\nfor social media?<\/p><p>Here are some\u00a0tips:<\/p><p>The expectation in professional spaces is that after you are\u00a0following someone or otherwise connected, responding to their\nposts is both allowed and encouraged. Everyone, at all levels of success,\nlikes knowing that their work is appreciated.<\/p><p>Try to respect peoples\u2019 boundaries, and be especially cautious\u00a0if engagement isn\u2019t reciprocal and you don\u2019t have a relationship.\n(Persistence is helpful in business; \u201cping someone every other day\u201d\nis spammy.)<\/p><p>Remember a response is a first impression of you\u2014when you\u00a0interact, stay positive and try to contribute something\nmeaningful, like linking to an article they might find interesting\nor bringing up a point they might not have thought of.<\/p><p>It is okay to message people you don\u2019t know asking for an\u00a0informational interview.<\/p><p>It is okay to ask someone you have already established a\u00a0relationship with to share something you posted on their network,\nbut make more deposits in your karma bank than withdraws.<\/p><h2 id=\"host-your-own-events-and-communities\">Host your own events and communities!<\/h2><p>Hosting events and communities can be powerful ways to talk to people\u00a0and generate business. I\u2019ve hosted many\u2014sometimes they\u2019re focused\ndinners with potential customers, sometimes they\u2019re community events\nfocused around giving back.<\/p><p>At any event, I start by talking to a small group of people, each\u00a0one-on-one, to confirm their interest and be sure they\u2019ll show up. If\nit\u2019s an intimate dinner, this is all that\u2019s needed.<\/p><p>If it\u2019s a larger event this confirms that there is some interest and some\npeople committed to coming. Successful events <a href=\"https:\/\/sivers.org\/dancingguy\">nucleate around a core of early people<\/a>,\u00a0much like successful movements.<\/p><p>This is also true of product adoption,\nwhich is one reason why networking is so important. You can get your first\n<a href=\"https:\/\/stripe.com\/atlas\/guides\/starting-sales\">few users<\/a> scrappily, when\u00a0most users won\u2019t want anything to do with you, and then the fact that you\nhave users gives everyone else safety to come.<\/p><p>For public events, feel free to reach out to local organizations or\u00a0companies to partner with. Companies are usually happy to donate space\nfor events. It\u2019s free PR for them and helps with other business goals\nlike recruiting.<\/p><p>If it\u2019s critical that the event be professional and polished and it\u2019s a\nlarger event, hire an event planner. Or, if you have the budget, consult\nwith an event planner. Just trust me on this one. You <em>can<\/em> plan events\u00a0without professional help, but it\u2019s always better to have someone who\nknows what they\u2019re doing helping you.<\/p><p>Feel free to plan events around conferences, but you can also plan\u00a0events any time of the year and can travel to cities just to network at\nevents you\u2019ve planned. There are all kinds of events you can plan around\nyour business: free classes around your product, public or private\ncustomer events, trainings, dinners with high-profile potential\ncustomers, etc. There are many events you can plan for the community,\ntoo: free classes on technology or business, talk series, interactive\nworkshops, career advice panels, etc.<\/p><p>Congratulations, you\u2019ve got yourself a\u00a0meeting!<\/p><h2 id=\"how-to-approach-the-ask\">How to approach the ask<\/h2><p>Many people you talk to are going to want to help you build your\u00a0business. When and how should you ask for help?<\/p><p>There are lots of different styles of \u201casks,\u201d introductions, and\u00a0conversation styles. Let\u2019s go over some of them and discuss pros\/cons.<\/p><p>A cold ask is introducing yourself with an immediate request for something.\u00a0People who aren\u2019t very familiar with business believe sales to be all cold\nasks, all the time. That isn\u2019t accurate, but there is a place for them.<\/p><p>Most people won\u2019t be able to help you right away. A cold ask will filter\u00a0away anyone who can\u2019t immediately help you\u2014you won\u2019t form a\nrelationship with them, they probably won\u2019t come back to help you down\nthe road, they probably won\u2019t refer you to someone they do know who can\nhelp you.<\/p><p>It can be tricky to craft and get right, especially since you\u2019re\u00a0unlikely to get good feedback from the people you\u2019re sending it to. If\nyou\u2019re new to this, I recommend getting it reviewed by someone with\nexperience and starting with messaging a small number of people and\nslowly expanding after judging reaction.<\/p><p>With this ask, it\u2019s common advice to send several polite follow-ups.\nIf you do this <em>poorly<\/em>, it can be annoying. Done well, it is simply\u00a0helping another busy professional manage their schedule. Keep the follow-up\nsimple, concise, and polite.<\/p><blockquote>\n  <p>Hiya\u00a0Steve,<\/p>\n  <p>Do you have 30 minutes this week to discuss whether we can improve\u00a0your asset pipeline?<\/p>\n  <p>Best,<\/p>\n  <p>Stephanie<\/p>\n<\/blockquote><p>Making cold asks is hard and awkward, particularly for people who\u00a0are sensitive to avoiding other people. On the flipside, it is efficient,\nlets disinterest parties signal that easily, and makes sure that\neveryone is on the same page.<\/p><p>Consider a job fair. Everyone at the job fair is there with an objective:\u00a0they want to hire someone or they want to be hired. You don\u2019t need to\nspend a lot of time building a relationship prior to saying \u201cActually I\u2019d\nlike a job\u201d\u2014that would be borderline disingenuous. Just ask for the job.<\/p><p>Much more of life than early entrepreneurs is more like a job fair than\nit is like a dinner party. Businesses exist to buy things and sell things.\n<em>You<\/em> will buy things, including from cold asks, and often you will be\u00a0very happy with those things. Some interactions are very transactional,\nand that is okay, too.<\/p><p>In professional spaces, when you\u2019ve just met someone, you often don\u2019t have a particular\ngoal in mind and don\u2019t know if you can help them. You can break the ice by having both\nparties introduce yourselves with a quick pitch. Get good at this. <a href=\"https:\/\/stripe.com\/atlas\/guides\/pitching\">Pitching<\/a>\u00a0is a skill.<\/p><p>It took me probably six months to be able to describe my company in two compelling\u00a0sentences\u2014it\u2019s not easy.<\/p><p>I used to start pitches with an explanation of GPU formats to give them a basis of understanding\u00a0of our algorithm, then moving into the problem we solve. This wasn\u2019t an appropriate amount of detail for most people I spoke too; it went over the heads of some and wasn\u2019t obviously valuable to others. Now, I lead off with \u201cBasis reduces download size and improves performance of apps.\u201d Everyone involved in a software company in any capacity understands that those are both good things. If I know they have some familiarity with the field of compression, I\u2019ll add \u201cBasis reduces images to the size of JPEG, but 6-8 times smaller on the GPU. If you already use GPU-friendly compression, it can cut your image data in half!\u201d That gives elaboration on the engineering case for basis and, while remaining comprehensible to a non-specialists, signals that I am capable of going arbitrarily deep if they want to.<\/p><p><strong>Be brief.<\/strong> I can\u2019t stress this enough. You\u2019re starting a conversation\u00a0this way, and you have no idea how interested the person is going to\nbe\u2014they don\u2019t need your life story. This is why crafting a short,\ncompelling pitch is key.<\/p><p>After you give your pitch, feel free to answer any questions they have,\u00a0then usually it\u2019s a good idea to switch to showing interest in them. One\nof the easiest ways to keep up a conversation is to be genuinely\ninterested in someone and just enjoy getting to know them.<\/p><p>This is the trickiest one, and usually one you\u2019ll see in social media.\u00a0It\u2019s also an approach you\u2019ll see in industries who are very allergic to\npeople appearing too \u201csales-y\u201d\u2014you can try to craft a pitch that\ndoesn\u2019t sound sales-y, or you can just skirt around it and use this\napproach. Selling software to video game companies can be like this, for instance.<\/p><p>You start by talking about something that is perhaps related to what you\u00a0do, but you don\u2019t introduce yourself or your company. For instance, you\njoin a conversation talking about a programming topic in social media or\nat an event, just actively contributing ideas and generating interest in\nwhat you do by doing that.<\/p><p>It can also happen in one-on-one conversations. You meet with someone in\u00a0the industry \u201cjust to catch up\u201d or without a clear agenda\u2014perhaps you\nstart by talking about industry news or what\u2019s going on with them, and\nyou may never get around to talking about what you do.<\/p><p>This approach to networking reminds me a little bit of selling a luxury\ngood, which is also why it works well in places that don\u2019t like sales-y\napproaches. Part of the mystique of luxury goods is that they\u2019re so\ninnately desirable that they don\u2019t need to be sold. This isn\u2019t <em>true<\/em>,\u00a0but sellers of them affect a sort of indifference to whether any particular\ntransaction happens, because the product is so good that buyers exist\nin abundance. People learn to associate this aloofness with success; they\nlearn to associate seeming desperation with failure. You generally don\u2019t\nwant to seem desperate.<\/p><p>There are a couple keys to success with this approach that also can help\u00a0other approaches. One very important one is keeping it genuine. Don\u2019t\nact like you\u2019re personal friends with someone if you aren\u2019t. It is okay\nto keep a relationship 100% professional and be genuine in that sense.\nDon\u2019t pretend to get excited about something you aren\u2019t excited\nabout\u2014you can share an opinion or article just for the sake of sharing\nit and contributing to the conversation. Stay yourself, but also be\nconfident and positive.<\/p><p>The other part of being genuine is maintaining long-term relationships\u00a0even if someone can\u2019t immediately help you, which ties into the next\npoint. Consistently follow up and plan on keeping in touch, even beyond\nthe needs of this business. Care about them as people, even if you don\u2019t\ncheck in super often. I realize this can be hard when you\u2019re talking to\nlots of people\u2014that is one of the downsides of this approach\u2014but\neven checking in and chatting with someone every few months is extremely\npowerful.<\/p><h3 id=\"hard-versus-soft-asks\">Hard versus soft asks<\/h3><p>Let\u2019s say you didn\u2019t go with the cold ask right at the introduction. How\u00a0do you bring up what you need?<\/p><p>\u201cWhat you need\u201d can be all sorts of things\u2014a referral to other\u00a0customers, a sale from the person you\u2019re talking to, or consulting work.<\/p><p>There are two ways to approach\u00a0it:<\/p><p>Examples of soft\u00a0asks:<\/p><ul><li>\n    <p>\u201cWe\u2019re putting together our consulting calendar at the\u00a0moment.\u201d<\/p>\n  <\/li>\n  <li>\n    <p>\u201cWe\u2019re looking for customers interested in our\u00a0product.\u201d<\/p>\n  <\/li>\n  <li>\n    <p>\u201cWe\u2019re currently doing free evaluations of the\u00a0product.\u201d<\/p>\n  <\/li>\n<\/ul><p>Examples of hard\u00a0asks:<\/p><ul><li>\n    <p>\u201cDo you need any consulting work\u00a0done?\u201d<\/p>\n  <\/li>\n  <li>\n    <p>\u201cDo you know of anyone who could use this\u00a0product?\u201d<\/p>\n  <\/li>\n  <li>\n    <p>\u201cWould you be interested in a free\u00a0evaluation?\u201d<\/p>\n  <\/li>\n<\/ul><p>After reading the other sections, you can probably guess what the\u00a0pros\/cons of each are. A soft ask lets someone come to you with interest\nand is good for people who don\u2019t like sales-y approaches, the \u201cluxury\ngood\u201d mentality, and for someone who may not be able to help you. The\nhard ask can be great when someone is expecting it, it cuts straight to\nthe point, and it isn\u2019t ambiguous about your needs.<\/p><p>A note of caution for all asks\u2014you should respect people and their\u00a0time. If you genuinely aren\u2019t interested in keeping in touch or hearing\nabout a person\u2019s story or helping them out (which is often\nunderstandable\u2014for instance if you have to approach a huge number of\npeople in a limited time), it\u2019s often better to go with the cold hard\nask right in the intro. A hard ask later in the conversation requires a\ndegree of respect, trust, and prior relationship building so the person\ndoesn\u2019t feel their time leading up to the ask was a waste.<\/p><h2 id=\"courtesies-follow-ups\">Courtesies &amp; follow ups<\/h2><p>Feel free to hand someone a business card after a conversation or plan a\u00a0concrete follow-up if the meeting\u2019s going well or you both have more to\ntalk about.<\/p><p>It\u2019s completely okay to follow up with someone frequently and have a\u00a0100% professional relationship. In fact, it is a part of staying\ngenuine. With some people you\u2019ll catch up about their kids and home\nlife, with some people you\u2019ll be great friends, and with some people\nyou\u2019ll just talk business. Be aware of people\u2019s boundaries and be aware\nthat this is very normal.<\/p><h2 id=\"a-note-on-mentoring-relationships-and-power-dynamics\">A note on mentoring relationships and power dynamics<\/h2><p>When you are mentoring someone, talking with someone more junior, or\u00a0someone in a position of less power, the power dynamic is going to be\nvery different from networking with someone you\u2019d consider a colleague,\nat your experience level, or more senior\/powerful.<\/p><p>If the power dynamic is unbalanced in your favor, the person won\u2019t feel\nas comfortable saying no to your questions or resisting your\nsuggestions. Do your best to sense how they\u2019re feeling, make them feel\ncomfortable, and have <em>them<\/em> suggest things, leaving options\u00a0open.<\/p><p>Doing networking well, at its core, is about understanding people and\u00a0treating both them and yourself with respect.<\/p><p>Always consider your particular situation and circumstances. And never\u00a0forget to pay it forward and use this advice to connect people less\nexperienced &amp; privileged to opportunities.<\/p><p>I would love to hear about any networking tips you\u2019d add or your\nexperiences with networking\u2014<a href=\"https:\/\/www.twitter.com\/sehurlburt\">find me on\u00a0Twitter<\/a>!<\/p><p>Wishing you the\u00a0best!<\/p><aside class=\"collect-email-guide\">\n  <\/aside><a href=\"https:\/\/stripe.com\/atlas\/guides\" class=\"common-Link common-Link--arrowL\">Return to guides<\/a>","created_at":"2018-11-07T16:42:45+0000","updated_at":"2018-11-07T16:42:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":39,"domain_name":"stripe.com","preview_picture":"https:\/\/stripe.com\/img\/v3\/atlas\/guide\/social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":266,"title":"Building an Immersive Game with A-Frame and Low Poly Models \u2013 Mozilla Hacks - the Web developer blog","url":"https:\/\/hacks.mozilla.org\/2018\/03\/immersive-aframe-low-poly\/","content":"<p><em>Note: This is Part 1 of a two-part tutorial.<\/em><\/p><p>There is a big difference between <strong>immersion<\/strong> and <strong>realism<\/strong>.  A high-end computer game with detailed models and a powerful GPU can feel realistic, but still not feel immersive. There\u2019s more to creating a feeling of <em>being there<\/em> than polygon count.  A low poly experience can feel very immersive through careful set design and lighting choices, without being realistic at all.<\/p><p>Today I\u2019m going to show you how to build a simple but immersive game with <a href=\"https:\/\/aframe.io\/\" rel=\"noopener\" target=\"_blank\">A-Frame<\/a> and models from the previous Sketchfab design challenge. Unlike <a href=\"https:\/\/blog.mozvr.com\/using-gltf-models-with-a-frame\/\" rel=\"noopener\" target=\"_blank\">my previous tutorials<\/a>, in this one we will walk through creating the <strong>entire application<\/strong>. Not just the basic interaction, but also adding and positioning 3d models, programmatically building a landscape with rocks, adding sounds and lighting to make the player feel immersed in the environment, and finally interaction tweaks for different form factors.<\/p><p>Here\u2019s a condensed video version of my tutorial that introduces the steps you\u2019ll need to follow to build your WebVR game:<\/p><iframe width=\"500\" height=\"281\" src=\"https:\/\/www.youtube.com\/embed\/1PxHy9zl2uA?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>I hope this blog will inspire you to submit to the current <a href=\"https:\/\/challenges.mozilla.community\/webvr-challenge\/\">challenge<\/a> we are running with SketchFab. There\u2019s still time to enter before submissions close on April 2nd.<\/p><h2>Boilerplate<\/h2><p>Our WebVR Whack-an-Imp game is a variation on Whack-A-Mole, except in our case it will be an imp flying out of a bubbling cauldron. Before we get to fancy 3D models, however, we must begin with an empty HTML file that includes the A-Frame library.<\/p><pre class=\"html\">&lt;html&gt;\n  &lt;head&gt;\n    &lt;!-- aframe itself --&gt;\n    &lt;script src=\"https:\/\/aframe.io\/releases\/0.7.0\/aframe.min.js\"&gt;&lt;\/script&gt;\n  &lt;\/head&gt;\n  &lt;body&gt;\n  &lt;\/body&gt;\n&lt;\/html&gt;\n<\/pre><p>At first we won\u2019t make the scene pretty at all. We just want to prove that our concept will work, so we will keep it simple. That means no lighting, models, or sound effects. Once the underlying concept is proven we will make it pretty.<\/p><p>Let\u2019s start off with a scene with stats turned on, then add a camera with <code>look-controls<\/code> at a height of 1.5 m; which is a good camera height for VR interaction (roughly corresponding to the average eye height of most adult humans).<\/p><pre class=\"html\">&lt;a-scene stats&gt;\n    &lt;a-entity camera look-controls position=\"0 1.5 0\"&gt;\n        &lt;a-cursor&gt;&lt;\/a-cursor&gt;\n    &lt;\/a-entity&gt;\n&lt;\/a-scene&gt;\n<\/pre><p>Notice the <code>a-cursor<\/code> inside of the camera. This will draw a little circular cursor, which is important for displays that don\u2019t have controllers, such as Cardboard.<\/p><p>Our game will have an object that pops up from a cauldron, then falls back down as gravity takes hold. The player will have a paddle or stick to hit the object. If the player misses, then the object should fall on the ground. For now let\u2019s represent the object with a sphere and the ground with a simple plane. Put this code inside of the <code>a-scene<\/code>.<\/p><pre class=\"html\">&lt;a-entity id='ball'\n          position=\"0 1 -4\"\n          material=\"color:green;\"\n          geometry=\"primitive:sphere; radius: 0.5;\"\n&gt;&lt;\/a-entity&gt;\n&lt;a-plane color='red' rotation=\"-90 0 0\" \n         width=\"100\" height=\"100\"&gt;&lt;\/a-plane&gt;\n<\/pre><p>Note that I\u2019m using the long syntax of <code>a-entity<\/code> for the ball rather than <code>a-sphere<\/code>. That\u2019s because later we will switch the geometry to an externally loaded model. However, the plane will always be a plane, so I\u2019ll use the shorter <code>a-plane<\/code> syntax for that one.<\/p><p>We have an object to hit but nothing to hit it <em>with<\/em>. Now add a box for the paddle.  Instead of using a controller to swing the paddle, we will start with the simplest possible interaction: put the box inside of the camera. Then you can swing it by just turning your head (or dragging the scene camera w\/ the mouse on desktop). A little awkward but it works well enough for now.<\/p><p>Also note that I placed the paddle box at z -3. If I\u2019d left it at the default position it would seem to disappear, but it\u2019s actually still there. The paddle is too close to the camera for us to see. If I look down at my feet I can see it though.  Whenever you are working with VR and your object doesn\u2019t show up, first check if it\u2019s behind you or too close to the camera.<\/p><pre class=\"html\">&lt;a-entity position=\"0 0 -3\" id=\"weapon\"&gt;\n    &lt;a-box color='blue' width='0.25' height='0.5' depth='3'&gt;&lt;\/a-box&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>Great. Now all of the elements of our scene are here. If you followed along you should have a scene on your desktop that looks like this.<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-geometry.png\" alt=\"Basic Geometry\" \/><\/p><p>If you play with this demo you\u2019ll see that you can move your head and the paddle moves with it, but trying to hit the ball won\u2019t do anything. That\u2019s because we only have geometry. The computer knows how our objects look but nothing about how they <em>behave<\/em>. For that we need physics.<\/p><h2>Physics<\/h2><p>Physics engines can be complicated, but fortunately <a href=\"https:\/\/github.com\/donmccurdy\" rel=\"noopener\" target=\"_blank\">Don McCurdy<\/a> has created A-Frame bindings for the excellent <a href=\"http:\/\/www.cannonjs.org\/\">Cannon.js<\/a> open source physics framework. We just need to include his <a href=\"https:\/\/github.com\/donmccurdy\/aframe-physics-system\"><code>aframe-extras<\/code><\/a> library to start playing with physics.<\/p><p>Add this to the <code>head<\/code> of the html page:<\/p><pre class=\"html\">&lt;!-- physics and other extras --&gt;\n&lt;script src=\"\/\/cdn.rawgit.com\/donmccurdy\/aframe-extras\/v3.13.1\/dist\/aframe-extras.min.js\"&gt;&lt;\/script&gt;\n<\/pre><p>Now we can turn on physics by adding <code>physics=\"debug:true;\u201d<\/code> to the <code>a-scene<\/code>.<\/p><p>Of course merely turning on the physics engine won\u2019t do anything. We still have to tell it <em>which<\/em> objects in the scene should be affected by gravity and other forces.  We do this with dynamic and static bodies.  A dynamic body is an object with full physics. It can transmit force and be affected by other forces, including gravity. A static body can transmit force when something hits it, but is otherwise unaffected by forces.  Generally you will use a static body for something that doesn\u2019t move, like the ground or a wall, and a dynamic body for things which do move around the scene, such as our ball.<\/p><p>Let\u2019s make the ground static and the ball dynamic by adding <code>dynamic-body<\/code> and <code>static-body<\/code> to their components:<\/p><pre class=\"html\">&lt;a-entity id='ball'\n          position=\"0 1 -4\"\n          material=\"color:green;\"\n          geometry=\"primitive:sphere; radius: 0.5;\"\n          dynamic-body\n&gt;&lt;\/a-entity&gt;\n&lt;a-plane color='red'\n         static-body\n         rotation=\"-90 0 0\" width=\"100\" height=\"100\"&gt;&lt;\/a-plane&gt;\n<\/pre><p>Great. Now when you reload the page the ball will fall to the ground. You may also see grid lines or dots on the ball or plane. These are bits of debugging information from the physics engine to let us see the edges of our objects from a physics perspective. It is possible to have the physics engine use a size or shape for our objects that\u2019s different than the real drawn geometry. I know this sounds strange, but it\u2019s actually quite useful, as we will see later.<\/p><p>Now we need to make the paddle able to hit the ball. Since the paddle moves, you might think we should use a <code>dynamic-body<\/code>, but really we want <em>our code<\/em> (and the camera) to control the position of the paddle, <em>not<\/em> the physics engine. We just want the paddle to be there for exerting forces on the ball, not the other way around, so we will use a <code>static-body<\/code>.<\/p><pre class=\"html\">&lt;a-entity camera look-controls position=\"0 1.5 0\"&gt;\n    &lt;a-cursor&gt;&lt;\/a-cursor&gt;\n    &lt;a-entity position=\"0 0 -3\" id='weapon'&gt;\n        &lt;a-box color='blue' width='0.25' height='0.5' depth='3'\n               static-body&gt;&lt;\/a-box&gt;\n    &lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>Now we can move the camera to swing the paddle and hit the ball. If you hit it hard then it will fly off to the side instead of rolling, exactly what we want!<\/p><p>You might ask why not just turn on physics for everything. Two reasons: First, physics requires CPU time. If more objects have associated physics, the more CPU resources they will consume.<\/p><p>Second reason: For many objects in the scene, we don\u2019t actually want physics turned on. If I have a tree in my scene, I don\u2019t want it to fall down just because it\u2019s a millimeter above the ground. I don\u2019t want the moon to be able to fall from the sky just because it\u2019s above the ground. Only turn on physics for things that really need it for your application.<\/p><h2>Collisions<\/h2><p>Moving the ball by hitting it is fun, but for a real game we need to track when the paddle hits the ball to increase the player\u2019s score. We also need to reset the ball back to the middle for another shot. We use collisions to do this. The physics engine emits a <code>collide<\/code> event each time an object hits another object. By listening to this event we can find out when something has been hit, what it is, and we can manipulate it.<\/p><p>First, let\u2019s make some utility functions for accessing DOM elements. I\u2019ve put these at the top of the page so they will be available to code everywhere.<\/p><pre class=\"html\">&lt;script&gt;\n    $ = (sel) =&gt; document.querySelector(sel)\n    $$ = (sel) =&gt; document.querySelectorAll(sel)\n    on = (elem, type, hand) =&gt; elem.addEventListener(type,hand)\n&lt;\/script&gt;\n<\/pre><p>Let\u2019s talk about the functions we need. First, we want to reset the ball after the player has hit it or if they\u2019ve missed and a certain number of seconds have gone by. Resetting means moving the ball back to the center, setting the forces back to zero, and initializing a timeout.  Let\u2019s create the <code>resetBall<\/code> function to do this:<\/p><pre class=\"js\">let hit = false\nlet resetId = 0\nconst resetBall = () =&gt; {\n    clearTimeout(resetId)\n    $(\"#ball\").body.position.set(0, 0.6,-4)\n    $(\"#ball\").body.velocity.set(0, 5,0)\n    $(\"#ball\").body.angularVelocity.set(0, 0,0)\n    hit = false\n    resetId = setTimeout(resetBall,6000)\n}\n<\/pre><p>In the above code I\u2019m using the <code>$<\/code> function with a selector to find the ball element in the page. The physics engine adds a <code>body<\/code> property to the element containing all of the physics attributes. We can reset the position, velocity, and <code>angularVelocity<\/code> from here. The code above also sets a timeout to call <code>resetBall<\/code> again after six seconds, if nothing else happens.<\/p><p>There are two things to note here. First, I\u2019m setting <code>body.position<\/code> rather than the regular <code>position<\/code> component that all A-Frame entities have. That\u2019s because the physics engine is in charge of this object, so we need to tell the <em>physics engine<\/em> about the changes, not A-Frame.<\/p><p>The second thing to note\u2014the velocity is not reset to zero. Instead it\u2019s set to the vector <code>0,5,0<\/code>.  This means zero velocity in the x and z directions, but <em>5<\/em> in the y direction. This gives the ball an initial vertical velocity, shooting it up. Of course gravity will start to affect it as soon as the ball jumps, so the velocity will quickly slow down. If I wanted to make the game harder I could increase the initial velocity here, or point the vector in a random direction. Lots of opportunities for improvements.<\/p><p>Now we need to know when the collision actually happens so we can increment the score and trigger the reset. We\u2019ll do this by handling the <code>collide<\/code> event on the <code>#weapon<\/code> entity. Note, this code should go <em>after<\/em> the closing <code>&lt;\/a-scene&gt;<\/code> so it won\u2019t be loaded until after the scene is ready.<\/p><pre class=\"js\">&lt;script&gt;\nlet score = 0;\nlet hit = false\nlet resetId = 0\non($(\"#weapon\"),'collide',(e)=&gt;{\n    const ball = $(\"#ball\")\n    if(e.detail.body.id === ball.body.id &amp;&amp; !hit) {\n        hit = true\n        score = score + 1\n        clearTimeout(resetId)\n        resetId = setTimeout(resetBall,2000)\n    }\n})\nsetTimeout(resetBall,3000)\n&lt;\/script&gt;\n<\/pre><p>The code above checks if the collision event is for the ball by comparing the body ids. It also makes sure the player didn\u2019t already hit the ball, otherwise they could hit the ball over and over again before we reset it. If the ball was hit, then set hit to true, clear the reset timeout, and schedule a new one for two seconds in the future.<\/p><p>Great, now we can launch the ball over and over and keep track of score.  Of course a score isn\u2019t very useful if we can\u2019t see it. Let\u2019s add a text element <em>inside<\/em> of the camera, so it is always visible. This is called a <em>Heads Up Display<\/em> or HUD.<\/p><pre class=\"html\">&lt;a-entity camera ....\n        &lt;a-text id=\"score\" value=\"Score\" position=\"-0.2 -0.5 -1\" color=\"red\" width=\"5\" anchor=\"left\"&gt;&lt;\/a-text&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>We need to update the score text whenever the score changes. Let\u2019s add this to the end of the  <code>collide<\/code> event handler.<\/p><pre class=\"js\">on($(\"#weapon\"),'collide',(e)=&gt;{\n    const ball = $(\"#ball\")\n    if(e.detail.body.id === ball.body.id &amp;&amp; !hit) {\n...\n        $(\"#score\").setAttribute('text','value','Score '+score)\n    }\n})\n<\/pre><p>Now we can see the score on screen. It should look like this:<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-physic-score.png\" alt=\"Score and Physics\" \/><\/p><h2>Models<\/h2><p>We have a basic game running. The player can hit the ball with a paddle and get points.  It\u2019s time to make this look better with real 3D models. We need a cool-looking imp to whack with the stick.<\/p><p>The <a href=\"https:\/\/medium.com\/mozilla-open-innovation\/building-webvr-worlds-together-mozilla-and-sketchfab-launching-real-time-vr-design-challenge-18b3315efe03\">last challenge<\/a> resulted in tons of great 3D scenes built around the theme of Low-Poly Medieval Fantasy.  Many of these have already been split up into individual assets and tagged with <a href=\"https:\/\/sketchfab.com\/tags\/medievalfantasyassets\">medievalfantasyassets<\/a>.<\/p><p>For this project I chose to use this <a href=\"https:\/\/sketchfab.com\/models\/e4dd9c9d9ad14976a4ce0d1f6049292e?ref=related\">imp model<\/a> for the ball and this <a href=\"https:\/\/sketchfab.com\/models\/9d96aab0ddc34aae9f7bb5ae8dfc8bd6#\">staff mode<\/a> for the paddle.<\/p><p>Since we are going to be loading lots of models we should load them as assets.  Assets are large chunks of data (images, sounds, models) that are preloaded and cached automatically when the game starts. Put this at the top of the scene and adjust the <code>src<\/code> urls to point to wherever you downloaded the models.<\/p><pre class=\"html\">&lt;a-assets&gt;\n    &lt;a-asset-item id=\"imp\" src=\"models\/imp\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"staff\" src=\"models\/staff\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n&lt;\/a-assets&gt;\n<\/pre><p>Now we can swap the sphere with the imp and the paddle box for the staff. Update the weapon element like this:<\/p><pre class=\"html\">&lt;a-entity position=\"0 0 -3\" id=\"weapon\"&gt;\n    &lt;a-entity gltf-model=\"#staff\"&gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>And the ball element like this:<\/p><pre class=\"html\">&lt;a-entity id='ball'\n          position=\"0 1 -4\"\n          dynamic-body\n    &gt;\n    &lt;a-entity id='imp-model' gltf-model=\"#imp\"&gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-imp.png\" alt=\"Imp and missing staff\" \/><\/p><p>We can see the imp but the staff is missing. What happened?<\/p><p>The problem is the staff model itself.  The imp model is (mostly) centered inside of its coordinate system, so it is visually positioned where we put it. However the staff model\u2019s center is significantly off from the center of its coordinate system; roughly 15 to 20 meters away.  This is a common issue with models you find online.  To fix it we need to translate the model\u2019s position to account for the offset.  After playing around with the staff model I found that an offset of 2.3, -2.7, -16.3 did the trick. I also had to rotate it 90 degrees to make it horizontal and shift it forward by four meters to make it visible. Wrap the model with an additional entity to apply the translation and rotation.<\/p><pre class=\"html\">&lt;a-entity id=\u201cweapon\u201d  rotation=\"-90 0 0\" position=\"0 0 -4\"&gt;\n    &lt;a-entity position=\"2.3 -2.7 -16.3\" \n              gltf-model=\"#staff\"\n              static-body&gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>Now we can see the staff, but we still have a problem. The staff is not a simple geometric shape, it\u2019s a full 3d model. The physics engine can\u2019t work directly with a full mesh. Instead it needs to know which primitive object to use.  We could use a box like we did originally, but I chose to go with a <em>sphere<\/em> centered at the end of the staff. That\u2019s the part that the player should actually use to hit the imp, and by making it larger than the staff\u2019s diameter we can make the game easier than it would be in real life. We also need to move the <code>static-body<\/code> definition to the outer entity so that it isn\u2019t affected by the model offset.<\/p><pre class=\"html\">&lt;a-entity rotation=\"-90 0 0\" position=\"0 0 -4\" id='weapon'\n          static-body=\"shape:sphere; sphereRadius: 0.3;\"&gt;\n    &lt;a-entity position=\"2.3 -2.7 -16.3\" \n              gltf-model=\"#staff\" &gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-imp-staff.png\" alt=\"Imp and Staff\" \/><\/p><h2>Scenery<\/h2><p>We have the core game mechanics working correctly with the new models, let\u2019s add some decorations next. I grabbed more models from SketchFab for a <a href=\"https:\/\/sketchfab.com\/models\/83c39af4bc5c478296501b80ef993855#download\">moon<\/a>, a <a href=\"https:\/\/sketchfab.com\/models\/e6eb5febb0ef4b63b3a90206a8b96135#download\">cauldron<\/a>, a <a href=\"https:\/\/sketchfab.com\/models\/6a0fa67b84a44d51b929d8450fb1ec48\">rock<\/a>, and <a href=\"https:\/\/sketchfab.com\/models\/4ddc25af08b24509aa9aba53c5365967\">two<\/a> different <a href=\"https:\/\/sketchfab.com\/models\/71c0756bf001437ebab9955c12f10d77\">trees<\/a>.  Place them in the scene at different positions.<\/p><pre class=\"html\">&lt;a-assets&gt;\n    &lt;a-asset-item id=\"imp\" src=\"models\/imp\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"staff\" src=\"models\/staff\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"tree1\" src=\"models\/arbol1\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"tree2\" src=\"models\/arbol2\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"moon\" src=\"models\/moon\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"cauldron\" src=\"models\/cauldron\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"rock1\" src=\"models\/rock1\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n&lt;\/a-assets&gt;\n...\n&lt;!-- cauldron --&gt;\n&lt;a-entity  position=\"1.5 0 -3.5\" gltf-model=\"#cauldron\"&gt;&lt;\/a-entity&gt;\n&lt;!-- the moon --&gt;\n&lt;a-entity gltf-model=\"#moon\"&gt;&lt;\/a-entity&gt;\n&lt;!-- trees --&gt;\n&lt;a-entity gltf-model=\"#tree2\" position=\"38 8.5 -10\"&gt;&lt;\/a-entity&gt;\n&lt;a-entity gltf-model=\"#tree1\" position=\"33 5.5 -10\"&gt;&lt;\/a-entity&gt;\n&lt;a-entity gltf-model=\"#tree1\" position=\"33 5.5 -30\"&gt;&lt;\/a-entity&gt;\n<\/pre><p>Our little game is starting to look like a real scene!<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-trees-moon.png\" alt=\"Trees and Moon\" \/><\/p><p>The cauldron has bubbles which appeared to animate on SketchFab but they aren\u2019t animating here.  The animation is stored inside the model but it isn\u2019t automatically played without an additional component.  Just add <code>animation-mixer<\/code> to the entity for the cauldron.<\/p><p>The final game has rocks scattered around the field. However, we really don\u2019t want to manually position fifty different rocks. Instead we can write a component to randomly position them for us.<\/p><p>The A-Frame docs explain <a href=\"https:\/\/aframe.io\/docs\/0.7.0\/introduction\/writing-a-component.html\">how to create a component<\/a> so I won\u2019t recount it all here. The gist of it is this: A component has some input properties and then executes code when <code>init()<\/code> is called (and a few other functions). In this case, we want to accept the source of a model, some variables controlling how to distribute the model around the scene, and then have a function which will create N copies of the model.<\/p><p>Below is the code. I know it looks intimidating but it\u2019s actually pretty simple. We\u2019ll go through it step by step.<\/p><pre class=\"html\">&lt;!-- alternate random number generator --&gt;\n&lt;script src=\"js\/random.js\"&gt;&lt;\/script&gt;\n&lt;!-- our `distribute` component --&gt;\n&lt;script&gt;\n    AFRAME.registerComponent('distribute', {\n        schema: {\n            src: {type:'string'},\n            jitter: {type:'vec3'},\n            centerOffset: {type:'vec3'},\n            radius: {type:'number'}\n        },\n        init: function() {\n            const rg = new Random(Random.engines.mt19937().seed(10))\n            const center = new THREE.Vector3(this.data.centerOffset.x, \n                      this.data.centerOffset.y, this.data.centerOffset.z)\n            const jx = this.data.jitter.x\n            const jy = this.data.jitter.y\n            const jz = this.data.jitter.z\n            if($(this.data.src).hasLoaded) {\n                const s = this.data.radius\n                for(let i = -s; i&lt;s; i++) {\n                    for(let j=-s; j&lt;s; j++) {\n                        const el = document.createElement('a-entity')\n                        el.setAttribute('gltf-model', this.data.src)\n                        const offset = new THREE.Vector3(i*s + rg.real(-jx,jx), \n                               rg.real(-jy,jy), \n                               j*s - rg.real(-jz,jz));\n                        el.setAttribute('position', center.clone().add(offset));\n                        el.setAttribute('rotation',{x:0, y:rg.real(-45,45)*Math.PI\/180, z:0})\n                        const scale = rg.real(0.5,1.5)\n                        el.setAttribute('scale',{x:scale,y:scale,z:scale})\n                        $('a-scene').appendChild(el)\n                    }\n                }\n            }\n        }\n    })\n&lt;\/script&gt;\n<\/pre><p>First I import <code>random.js<\/code>. This is a random number generator from <a href=\"https:\/\/github.com\/ckknight\/random-js\">this random-js project<\/a> by Cameron Knight. We could use the standard <code>Math.random()<\/code> function built into Javascript, but I want to ensure that the rocks are always positioned the same way every time the game is run. This other generator lets us provide a seed.<\/p><p>In the first line of the <code>init()<\/code> code you can see that I used the seed 10. I actually tried several different seeds until I found one that I liked the look of.  If I <em>did<\/em> actually want each load to be different, say for different levels of the game, then I could provide a different seed for each level.<\/p><p>The core of the <em>distribute<\/em> component consists of the nested <code>for<\/code> loops. The code creates a grid of entities, each attached to the same model. For each copy, we will translate it from the natural center point of the original model (the <code>modelCenter<\/code> parameter) , adding a random offset using the <code>jitter<\/code> parameter. Jitter represents the maximum amount the rock should move from that grid point.  Using <code>0 0 0<\/code> would be no jitter.  Using <code>0 10 0<\/code> would make the rocks go vertically anywhere between -10 and 10, but not move at all in the horizontal plane.  For this game I used <code>2 0.5 2<\/code> to move them around mostly horizontally but move up and down a tiny bit.  The loop code also gives the rocks a random scale and rotation around the Y axis, just to make the scene look a bit more organic.<\/p><p>This is the final result.<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-rocks.png\" alt=\"Distributed Rocks\" \/><\/p><p>This blog has gotten pretty long and we still haven\u2019t worked on lighting, sounds, or polish.  Let\u2019s continue the game right now in <a href=\"https:\/\/hacks.mozilla.org\/2018\/03\/immersive-aframe-low-poly-part2\/\">Part 2<\/a>.<\/p><p>If you\u2019d like to look at the source to the finished project, I\u2019ve put it all in this <a href=\"https:\/\/github.com\/joshmarinacci\/whack-an-imp\">github repo<\/a>.<\/p>","created_at":"2018-11-07T16:42:44+0000","updated_at":"2018-11-07T16:42:44+0000","published_at":"2018-03-07T08:11:13+0000","published_by":["Josh Marinacci"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":19,"domain_name":"hacks.mozilla.org","preview_picture":"https:\/\/hacks.mozilla.org\/files\/2018\/03\/Screenshot-2018-3-7-http-localhost1.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":267,"title":"Pixar&#8217;s 22 Rules of Storytelling","url":"https:\/\/www.aerogrammestudio.com\/2013\/03\/07\/pixars-22-rules-of-storytelling\/","content":"<p data-imagelightbox=\"g\"><img class=\"aligncenter wp-image-665 size-medium\" src=\"https:\/\/i1.wp.com\/www.aerogrammestudio.com\/wp-content\/uploads\/2013\/03\/pixars-22-rules-for-storytelling.jpg?resize=800%2C500&amp;ssl=1\" alt=\"Pixar's Rules of Storytelling\" width=\"800\" height=\"500\" srcset=\"https:\/\/i1.wp.com\/www.aerogrammestudio.com\/wp-content\/uploads\/2013\/03\/pixars-22-rules-for-storytelling.jpg?resize=800%2C500&amp;ssl=1 800w, https:\/\/i1.wp.com\/www.aerogrammestudio.com\/wp-content\/uploads\/2013\/03\/pixars-22-rules-for-storytelling.jpg?resize=150%2C94&amp;ssl=1 150w, https:\/\/i1.wp.com\/www.aerogrammestudio.com\/wp-content\/uploads\/2013\/03\/pixars-22-rules-for-storytelling.jpg?w=1440&amp;ssl=1 1440w, https:\/\/i1.wp.com\/www.aerogrammestudio.com\/wp-content\/uploads\/2013\/03\/pixars-22-rules-for-storytelling.jpg?resize=1200%2C750&amp;ssl=1 1200w\" data-recalc-dims=\"1\" \/><br \/>\nPixar\u2019s Rules of Storytelling were originally tweeted by\u00a0<a href=\"https:\/\/twitter.com\/lawnrocket\" target=\"_blank\" rel=\"noopener\">Emma Coats<\/a>, Pixar\u2019s Story Artist. Number 9 on the list\u00a0\u2013\u00a0When you\u2019re stuck, make a list of what wouldn\u2019t happen next \u2013 is a great one and can apply to writers in\u00a0all\u00a0genres.<\/p>\n<ol><li>You admire a character for trying more than for their successes.<\/li>\n<li>You gotta keep in mind what\u2019s interesting to you as an audience, not what\u2019s fun to do as a writer. They can be very different.<\/li>\n<li>Trying for theme is important, but you won\u2019t see what the story is actually about til you\u2019re at the end of it. Now rewrite.<\/li>\n<li>Once upon a time there was ___. Every day, ___. One day ___. Because of that, ___. Because of that, ___. Until finally ___.<\/li>\n<li>Simplify. Focus. Combine characters. Hop over detours. You\u2019ll feel like you\u2019re losing valuable stuff but it sets you free.<\/li>\n<li>What is your character good at, comfortable with? Throw the polar opposite at them. Challenge them. How do they deal?<\/li>\n<li>Come up with your ending before you figure out your middle. Seriously. Endings are hard, get yours working up front.<\/li>\n<li>Finish your story, let go even if it\u2019s not perfect. In an ideal world you have both, but move on. Do better next time.<\/li>\n<li>When you\u2019re stuck, make a list of what WOULDN\u2019T happen next. Lots of times the material to get you unstuck will show up.<\/li>\n<li>Pull apart the stories you like. What you like in them is a part of you; you\u2019ve got to recognize it before you can use it.<\/li>\n<li>Putting it on paper lets you start fixing it. If it stays in your head, a perfect idea, you\u2019ll never share it with anyone.<\/li>\n<li>Discount the 1st thing that comes to mind. And the 2nd, 3rd, 4th, 5th \u2013 get the obvious out of the way. Surprise yourself.<\/li>\n<li>Give your characters opinions. Passive\/malleable might seem likable to you as you write, but it\u2019s poison to the audience.<\/li>\n<li>Why must you tell THIS story? What\u2019s the belief burning within you that your story feeds off of? That\u2019s the heart of it.<\/li>\n<li>If you were your character, in this situation, how would you feel? Honesty lends credibility to unbelievable situations.<\/li>\n<li>What are the stakes? Give us reason to root for the character. What happens if they don\u2019t succeed? Stack the odds against.<\/li>\n<li>No work is ever wasted. If it\u2019s not working, let go and move on \u2013 it\u2019ll come back around to be useful later.<\/li>\n<li>You have to know yourself: the difference between doing your best &amp; fussing. Story is testing, not refining.<\/li>\n<li>Coincidences to get characters into trouble are great; coincidences to get them out of it are cheating.<\/li>\n<li>Exercise: take the building blocks of a movie you dislike. How d\u2019you rearrange them into what you DO like?<\/li>\n<li>You gotta identify with your situation\/characters, can\u2019t just write \u2018cool\u2019. What would make YOU act that way?<\/li>\n<li>What\u2019s the essence of your story? Most economical telling of it? If you know that, you can build out from there.<\/li>\n<\/ol><ul><li><em><a href=\"https:\/\/www.aerogrammestudio.com\/2013\/03\/22\/the-story-spine-pixars-4th-rule-of-storytelling\/\">Read More About The Story Spine: Pixar\u2019s 4th Rule of Storytelling<\/a><\/em><\/li>\n<li><em><a href=\"https:\/\/www.aerogrammestudio.com\/2013\/03\/12\/wall-e-and-toy-story-screenwriter-reveals-the-clues-to-a-great-story\/\">Watch Pixar\u2019s Andrew Stanton Reveal the Clues to a Great Story<\/a><\/em><\/li>\n<\/ul><p><\/p><div style=\"text-align: center;\"><br \/><br \/><ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-9640407720164730\" data-ad-slot=\"4665341201\"><br \/><div class=\"pipdig-post-sig socialz nopin\">&#13;\n\t\t\t\t\t<\/div>&#13;\n\t\t&#13;\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t&#13;\n\t\t<div id=\"amzn-assoc-ad-bb9d81f4-8ea4-4477-a9e6-b160b37f907b\"><div class=\"clearfix\"><div class=\"clearfix\">\t\n\t\n\t\n\t\t\t&#13;\n\t<\/div><\/div><\/div><\/ins><\/div>","created_at":"2018-11-07T16:42:44+0000","updated_at":"2018-11-07T16:42:44+0000","published_at":"2013-03-07T00:00:00+0000","published_by":["Aerogramme Writers' Studio"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"www.aerogrammestudio.com","preview_picture":"https:\/\/www.aerogrammestudio.com\/wp-content\/uploads\/2013\/03\/pixars-22-rules-for-storytelling-800x500.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":268,"title":"How to Design Programs","url":"http:\/\/www.htdp.org\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:42:44+0000","updated_at":"2018-11-07T16:42:44+0000","annotations":[],"reading_time":0,"domain_name":"www.htdp.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":269,"title":"Literate DevOps","url":"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops.html","content":"<p>\nWhile investigations in operations and administrations (as I\u2019ve\ndescribed) are useful to oneself when understanding the problem\ndomain, I need to communicate the results with my team mates.\nSince my <a href=\"https:\/\/github.com\/howardabrams\/dot-files\/blob\/master\/emacs-mail.org#sending-email\">Emacs configuration<\/a> allows me to send mail messages, I kick\noff the function, <code>org-mime-org-buffer-htmlize<\/code>, which exports the\n<code>org-mode<\/code> file to an HTML mail message (This function is part of\nthe latest <code>org-plus-contrib<\/code> package).\n<\/p><p>\nHowever, some times the exported results are not quite perfect.\n<\/p><p>\nFor instance, some blocks may result in some JSON data, and since\nthe HTML output can colorize the syntax, if I could just specify\nthat the output results were JavaScript, then the JSON data would be\nmuch prettier. Just use the <code>wrap<\/code> parameter, as in:\n<\/p><figure><a href=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-20.txt\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-20.png\" title=\"Click for text\" alt=\"image\" \/><\/a><\/figure><p>\nWhich puts the following in my <code>org-mode<\/code> file:\n<\/p><figure><a href=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-21.txt\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-21.png\" title=\"Click for text\" alt=\"image\" \/><\/a><\/figure><p>\nWhich gets exported like:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-js\">{\"time\":{\"iso\":\"2015-05-19T23:12:40Z\",\"timestamp\":1432077160,\"date\":\"19 May 2015\",\"time\":\"7:12 PM\"}}\n<\/pre><\/div><p>\nFor another example, my current project involves working with OpenStack,\nand its <code>nova<\/code> command line utility attempts to format the data as a\ntable:\n<\/p><pre class=\"example\">\n+--------------------------------------+--------------------+--------+------------+-------------+------------------------+\n| ID                                   | Name               | Status | Task State | Power State | Networks               |\n+--------------------------------------+--------------------+--------+------------+-------------+------------------------+\n| f9e7aed8-e425-4808-aace-8758dadd91bf | chefserver         | ACTIVE | -          | Running     | WPC-private=10.0.1.73  |\n| 0432f8b1-7e6d-4fc1-b181-02fa768c38ac | ha-compute1        | ACTIVE | -          | Running     | WPC-private=10.0.1.104 |\n| a5bdd1d0-d4b3-4856-a657-5759356c186b | ha-controller1     | ACTIVE | -          | Running     | WPC-private=10.0.1.97  |\n| 16263972-609e-44c0-83e0-f3147336071c | ha-controller2     | ACTIVE | -          | Running     | WPC-private=10.0.1.99  |\n| 89a89d1f-7be5-4c4f-82db-64b751f15f3b | ha-controller3     | ACTIVE | -          | Running     | WPC-private=10.0.1.100 |\n| b740095a-3f89-45d0-a2a1-9cfcadfb4ca3 | ha-monitoring      | ACTIVE | -          | Running     | WPC-private=10.0.1.95  |\n| 6bebe823-1504-4cb1-a898-bbc7894b1a32 | ha-sdn-controller1 | ACTIVE | -          | Running     | WPC-private=10.0.1.101 |\n| 456bf417-580e-49fb-be08-1b0153710f86 | ha-sdn-controller2 | ACTIVE | -          | Running     | WPC-private=10.0.1.102 |\n| 7aab184c-5fb4-4996-8ab2-8a65ea7668cb | ha-sdn-controller3 | ACTIVE | -          | Running     | WPC-private=10.0.1.103 |\n| 0c90d7b0-dab4-4af8-a970-e2e90dd8b9e4 | ha-storage-1       | ACTIVE | -          | Running     | WPC-private=10.0.1.76  |\n| fda0666e-d656-48fd-928f-83fb47c923f2 | ha-storage-2       | ACTIVE | -          | Running     | WPC-private=10.0.1.81  |\n| 021fc9c1-8d79-4c09-b3d4-6014d242403a | ha-storage-3       | ACTIVE | -          | Running     | WPC-private=10.0.1.96  |\n| bc5ad0fe-9ef2-4966-8d2b-99892f3f94cd | yum-server         | ACTIVE | -          | Running     | WPC-private=10.0.1.74  |\n+--------------------------------------+--------------------+--------+------------+-------------+------------------------+\n<\/pre><p>\nIf you manually change the output, those changes will not be honored\nwhen the file is exported (since those are redone during the\nexporting process).\n<\/p><p>\nThe way to do that is with a little Emacs Lisp code block that you\njust need to place somewhere in your file, like:\n<\/p><figure><a href=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-22.txt\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-22.png\" title=\"Click for text\" alt=\"image\" \/><\/a><\/figure><p>\nWith this code block <i>named<\/i>, <code>nova-conv<\/code>, I can use it to\n<i>post-process<\/i> the results, as in:\n<\/p><figure><a href=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-23.txt\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-23.png\" title=\"Click for text\" alt=\"image\" \/><\/a><\/figure><p>\nIn my particular case, I also want to get rid of that first line of\ndashes to make it more <code>org-mode<\/code> like:\n<\/p><figure><a href=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-24.txt\"><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/literate-devops-24.png\" title=\"Click for text\" alt=\"image\" \/><\/a>\n<p>\nTo be truly re-useable, place this code in your <a href=\"http:\/\/orgmode.org\/manual\/Library-of-Babel.html\">Library of Babel<\/a>,\nand then it is available from any file.\n<\/p>\n<\/figure>","created_at":"2018-11-07T16:42:44+0000","updated_at":"2018-11-07T16:42:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"howardism.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":265,"title":"Building an Immersive Game with A-Frame and Low Poly Models \u2013 Mozilla Hacks - the Web developer blog","url":"https:\/\/hacks.mozilla.org\/2018\/03\/immersive-aframe-low-poly\/","content":"<p><em>Note: This is Part 1 of a two-part tutorial.<\/em><\/p><p>There is a big difference between <strong>immersion<\/strong> and <strong>realism<\/strong>.  A high-end computer game with detailed models and a powerful GPU can feel realistic, but still not feel immersive. There\u2019s more to creating a feeling of <em>being there<\/em> than polygon count.  A low poly experience can feel very immersive through careful set design and lighting choices, without being realistic at all.<\/p><p>Today I\u2019m going to show you how to build a simple but immersive game with <a href=\"https:\/\/aframe.io\/\" rel=\"noopener\" target=\"_blank\">A-Frame<\/a> and models from the previous Sketchfab design challenge. Unlike <a href=\"https:\/\/blog.mozvr.com\/using-gltf-models-with-a-frame\/\" rel=\"noopener\" target=\"_blank\">my previous tutorials<\/a>, in this one we will walk through creating the <strong>entire application<\/strong>. Not just the basic interaction, but also adding and positioning 3d models, programmatically building a landscape with rocks, adding sounds and lighting to make the player feel immersed in the environment, and finally interaction tweaks for different form factors.<\/p><p>Here\u2019s a condensed video version of my tutorial that introduces the steps you\u2019ll need to follow to build your WebVR game:<\/p><iframe width=\"500\" height=\"281\" src=\"https:\/\/www.youtube.com\/embed\/1PxHy9zl2uA?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>I hope this blog will inspire you to submit to the current <a href=\"https:\/\/challenges.mozilla.community\/webvr-challenge\/\">challenge<\/a> we are running with SketchFab. There\u2019s still time to enter before submissions close on April 2nd.<\/p><h2>Boilerplate<\/h2><p>Our WebVR Whack-an-Imp game is a variation on Whack-A-Mole, except in our case it will be an imp flying out of a bubbling cauldron. Before we get to fancy 3D models, however, we must begin with an empty HTML file that includes the A-Frame library.<\/p><pre class=\"html\">&lt;html&gt;\n  &lt;head&gt;\n    &lt;!-- aframe itself --&gt;\n    &lt;script src=\"https:\/\/aframe.io\/releases\/0.7.0\/aframe.min.js\"&gt;&lt;\/script&gt;\n  &lt;\/head&gt;\n  &lt;body&gt;\n  &lt;\/body&gt;\n&lt;\/html&gt;\n<\/pre><p>At first we won\u2019t make the scene pretty at all. We just want to prove that our concept will work, so we will keep it simple. That means no lighting, models, or sound effects. Once the underlying concept is proven we will make it pretty.<\/p><p>Let\u2019s start off with a scene with stats turned on, then add a camera with <code>look-controls<\/code> at a height of 1.5 m; which is a good camera height for VR interaction (roughly corresponding to the average eye height of most adult humans).<\/p><pre class=\"html\">&lt;a-scene stats&gt;\n    &lt;a-entity camera look-controls position=\"0 1.5 0\"&gt;\n        &lt;a-cursor&gt;&lt;\/a-cursor&gt;\n    &lt;\/a-entity&gt;\n&lt;\/a-scene&gt;\n<\/pre><p>Notice the <code>a-cursor<\/code> inside of the camera. This will draw a little circular cursor, which is important for displays that don\u2019t have controllers, such as Cardboard.<\/p><p>Our game will have an object that pops up from a cauldron, then falls back down as gravity takes hold. The player will have a paddle or stick to hit the object. If the player misses, then the object should fall on the ground. For now let\u2019s represent the object with a sphere and the ground with a simple plane. Put this code inside of the <code>a-scene<\/code>.<\/p><pre class=\"html\">&lt;a-entity id='ball'\n          position=\"0 1 -4\"\n          material=\"color:green;\"\n          geometry=\"primitive:sphere; radius: 0.5;\"\n&gt;&lt;\/a-entity&gt;\n&lt;a-plane color='red' rotation=\"-90 0 0\" \n         width=\"100\" height=\"100\"&gt;&lt;\/a-plane&gt;\n<\/pre><p>Note that I\u2019m using the long syntax of <code>a-entity<\/code> for the ball rather than <code>a-sphere<\/code>. That\u2019s because later we will switch the geometry to an externally loaded model. However, the plane will always be a plane, so I\u2019ll use the shorter <code>a-plane<\/code> syntax for that one.<\/p><p>We have an object to hit but nothing to hit it <em>with<\/em>. Now add a box for the paddle.  Instead of using a controller to swing the paddle, we will start with the simplest possible interaction: put the box inside of the camera. Then you can swing it by just turning your head (or dragging the scene camera w\/ the mouse on desktop). A little awkward but it works well enough for now.<\/p><p>Also note that I placed the paddle box at z -3. If I\u2019d left it at the default position it would seem to disappear, but it\u2019s actually still there. The paddle is too close to the camera for us to see. If I look down at my feet I can see it though.  Whenever you are working with VR and your object doesn\u2019t show up, first check if it\u2019s behind you or too close to the camera.<\/p><pre class=\"html\">&lt;a-entity position=\"0 0 -3\" id=\"weapon\"&gt;\n    &lt;a-box color='blue' width='0.25' height='0.5' depth='3'&gt;&lt;\/a-box&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>Great. Now all of the elements of our scene are here. If you followed along you should have a scene on your desktop that looks like this.<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-geometry.png\" alt=\"Basic Geometry\" \/><\/p><p>If you play with this demo you\u2019ll see that you can move your head and the paddle moves with it, but trying to hit the ball won\u2019t do anything. That\u2019s because we only have geometry. The computer knows how our objects look but nothing about how they <em>behave<\/em>. For that we need physics.<\/p><h2>Physics<\/h2><p>Physics engines can be complicated, but fortunately <a href=\"https:\/\/github.com\/donmccurdy\" rel=\"noopener\" target=\"_blank\">Don McCurdy<\/a> has created A-Frame bindings for the excellent <a href=\"http:\/\/www.cannonjs.org\/\">Cannon.js<\/a> open source physics framework. We just need to include his <a href=\"https:\/\/github.com\/donmccurdy\/aframe-physics-system\"><code>aframe-extras<\/code><\/a> library to start playing with physics.<\/p><p>Add this to the <code>head<\/code> of the html page:<\/p><pre class=\"html\">&lt;!-- physics and other extras --&gt;\n&lt;script src=\"\/\/cdn.rawgit.com\/donmccurdy\/aframe-extras\/v3.13.1\/dist\/aframe-extras.min.js\"&gt;&lt;\/script&gt;\n<\/pre><p>Now we can turn on physics by adding <code>physics=\"debug:true;\u201d<\/code> to the <code>a-scene<\/code>.<\/p><p>Of course merely turning on the physics engine won\u2019t do anything. We still have to tell it <em>which<\/em> objects in the scene should be affected by gravity and other forces.  We do this with dynamic and static bodies.  A dynamic body is an object with full physics. It can transmit force and be affected by other forces, including gravity. A static body can transmit force when something hits it, but is otherwise unaffected by forces.  Generally you will use a static body for something that doesn\u2019t move, like the ground or a wall, and a dynamic body for things which do move around the scene, such as our ball.<\/p><p>Let\u2019s make the ground static and the ball dynamic by adding <code>dynamic-body<\/code> and <code>static-body<\/code> to their components:<\/p><pre class=\"html\">&lt;a-entity id='ball'\n          position=\"0 1 -4\"\n          material=\"color:green;\"\n          geometry=\"primitive:sphere; radius: 0.5;\"\n          dynamic-body\n&gt;&lt;\/a-entity&gt;\n&lt;a-plane color='red'\n         static-body\n         rotation=\"-90 0 0\" width=\"100\" height=\"100\"&gt;&lt;\/a-plane&gt;\n<\/pre><p>Great. Now when you reload the page the ball will fall to the ground. You may also see grid lines or dots on the ball or plane. These are bits of debugging information from the physics engine to let us see the edges of our objects from a physics perspective. It is possible to have the physics engine use a size or shape for our objects that\u2019s different than the real drawn geometry. I know this sounds strange, but it\u2019s actually quite useful, as we will see later.<\/p><p>Now we need to make the paddle able to hit the ball. Since the paddle moves, you might think we should use a <code>dynamic-body<\/code>, but really we want <em>our code<\/em> (and the camera) to control the position of the paddle, <em>not<\/em> the physics engine. We just want the paddle to be there for exerting forces on the ball, not the other way around, so we will use a <code>static-body<\/code>.<\/p><pre class=\"html\">&lt;a-entity camera look-controls position=\"0 1.5 0\"&gt;\n    &lt;a-cursor&gt;&lt;\/a-cursor&gt;\n    &lt;a-entity position=\"0 0 -3\" id='weapon'&gt;\n        &lt;a-box color='blue' width='0.25' height='0.5' depth='3'\n               static-body&gt;&lt;\/a-box&gt;\n    &lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>Now we can move the camera to swing the paddle and hit the ball. If you hit it hard then it will fly off to the side instead of rolling, exactly what we want!<\/p><p>You might ask why not just turn on physics for everything. Two reasons: First, physics requires CPU time. If more objects have associated physics, the more CPU resources they will consume.<\/p><p>Second reason: For many objects in the scene, we don\u2019t actually want physics turned on. If I have a tree in my scene, I don\u2019t want it to fall down just because it\u2019s a millimeter above the ground. I don\u2019t want the moon to be able to fall from the sky just because it\u2019s above the ground. Only turn on physics for things that really need it for your application.<\/p><h2>Collisions<\/h2><p>Moving the ball by hitting it is fun, but for a real game we need to track when the paddle hits the ball to increase the player\u2019s score. We also need to reset the ball back to the middle for another shot. We use collisions to do this. The physics engine emits a <code>collide<\/code> event each time an object hits another object. By listening to this event we can find out when something has been hit, what it is, and we can manipulate it.<\/p><p>First, let\u2019s make some utility functions for accessing DOM elements. I\u2019ve put these at the top of the page so they will be available to code everywhere.<\/p><pre class=\"html\">&lt;script&gt;\n    $ = (sel) =&gt; document.querySelector(sel)\n    $$ = (sel) =&gt; document.querySelectorAll(sel)\n    on = (elem, type, hand) =&gt; elem.addEventListener(type,hand)\n&lt;\/script&gt;\n<\/pre><p>Let\u2019s talk about the functions we need. First, we want to reset the ball after the player has hit it or if they\u2019ve missed and a certain number of seconds have gone by. Resetting means moving the ball back to the center, setting the forces back to zero, and initializing a timeout.  Let\u2019s create the <code>resetBall<\/code> function to do this:<\/p><pre class=\"js\">let hit = false\nlet resetId = 0\nconst resetBall = () =&gt; {\n    clearTimeout(resetId)\n    $(\"#ball\").body.position.set(0, 0.6,-4)\n    $(\"#ball\").body.velocity.set(0, 5,0)\n    $(\"#ball\").body.angularVelocity.set(0, 0,0)\n    hit = false\n    resetId = setTimeout(resetBall,6000)\n}\n<\/pre><p>In the above code I\u2019m using the <code>$<\/code> function with a selector to find the ball element in the page. The physics engine adds a <code>body<\/code> property to the element containing all of the physics attributes. We can reset the position, velocity, and <code>angularVelocity<\/code> from here. The code above also sets a timeout to call <code>resetBall<\/code> again after six seconds, if nothing else happens.<\/p><p>There are two things to note here. First, I\u2019m setting <code>body.position<\/code> rather than the regular <code>position<\/code> component that all A-Frame entities have. That\u2019s because the physics engine is in charge of this object, so we need to tell the <em>physics engine<\/em> about the changes, not A-Frame.<\/p><p>The second thing to note\u2014the velocity is not reset to zero. Instead it\u2019s set to the vector <code>0,5,0<\/code>.  This means zero velocity in the x and z directions, but <em>5<\/em> in the y direction. This gives the ball an initial vertical velocity, shooting it up. Of course gravity will start to affect it as soon as the ball jumps, so the velocity will quickly slow down. If I wanted to make the game harder I could increase the initial velocity here, or point the vector in a random direction. Lots of opportunities for improvements.<\/p><p>Now we need to know when the collision actually happens so we can increment the score and trigger the reset. We\u2019ll do this by handling the <code>collide<\/code> event on the <code>#weapon<\/code> entity. Note, this code should go <em>after<\/em> the closing <code>&lt;\/a-scene&gt;<\/code> so it won\u2019t be loaded until after the scene is ready.<\/p><pre class=\"js\">&lt;script&gt;\nlet score = 0;\nlet hit = false\nlet resetId = 0\non($(\"#weapon\"),'collide',(e)=&gt;{\n    const ball = $(\"#ball\")\n    if(e.detail.body.id === ball.body.id &amp;&amp; !hit) {\n        hit = true\n        score = score + 1\n        clearTimeout(resetId)\n        resetId = setTimeout(resetBall,2000)\n    }\n})\nsetTimeout(resetBall,3000)\n&lt;\/script&gt;\n<\/pre><p>The code above checks if the collision event is for the ball by comparing the body ids. It also makes sure the player didn\u2019t already hit the ball, otherwise they could hit the ball over and over again before we reset it. If the ball was hit, then set hit to true, clear the reset timeout, and schedule a new one for two seconds in the future.<\/p><p>Great, now we can launch the ball over and over and keep track of score.  Of course a score isn\u2019t very useful if we can\u2019t see it. Let\u2019s add a text element <em>inside<\/em> of the camera, so it is always visible. This is called a <em>Heads Up Display<\/em> or HUD.<\/p><pre class=\"html\">&lt;a-entity camera ....\n        &lt;a-text id=\"score\" value=\"Score\" position=\"-0.2 -0.5 -1\" color=\"red\" width=\"5\" anchor=\"left\"&gt;&lt;\/a-text&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>We need to update the score text whenever the score changes. Let\u2019s add this to the end of the  <code>collide<\/code> event handler.<\/p><pre class=\"js\">on($(\"#weapon\"),'collide',(e)=&gt;{\n    const ball = $(\"#ball\")\n    if(e.detail.body.id === ball.body.id &amp;&amp; !hit) {\n...\n        $(\"#score\").setAttribute('text','value','Score '+score)\n    }\n})\n<\/pre><p>Now we can see the score on screen. It should look like this:<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-physic-score.png\" alt=\"Score and Physics\" \/><\/p><h2>Models<\/h2><p>We have a basic game running. The player can hit the ball with a paddle and get points.  It\u2019s time to make this look better with real 3D models. We need a cool-looking imp to whack with the stick.<\/p><p>The <a href=\"https:\/\/medium.com\/mozilla-open-innovation\/building-webvr-worlds-together-mozilla-and-sketchfab-launching-real-time-vr-design-challenge-18b3315efe03\">last challenge<\/a> resulted in tons of great 3D scenes built around the theme of Low-Poly Medieval Fantasy.  Many of these have already been split up into individual assets and tagged with <a href=\"https:\/\/sketchfab.com\/tags\/medievalfantasyassets\">medievalfantasyassets<\/a>.<\/p><p>For this project I chose to use this <a href=\"https:\/\/sketchfab.com\/models\/e4dd9c9d9ad14976a4ce0d1f6049292e?ref=related\">imp model<\/a> for the ball and this <a href=\"https:\/\/sketchfab.com\/models\/9d96aab0ddc34aae9f7bb5ae8dfc8bd6#\">staff mode<\/a> for the paddle.<\/p><p>Since we are going to be loading lots of models we should load them as assets.  Assets are large chunks of data (images, sounds, models) that are preloaded and cached automatically when the game starts. Put this at the top of the scene and adjust the <code>src<\/code> urls to point to wherever you downloaded the models.<\/p><pre class=\"html\">&lt;a-assets&gt;\n    &lt;a-asset-item id=\"imp\" src=\"models\/imp\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"staff\" src=\"models\/staff\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n&lt;\/a-assets&gt;\n<\/pre><p>Now we can swap the sphere with the imp and the paddle box for the staff. Update the weapon element like this:<\/p><pre class=\"html\">&lt;a-entity position=\"0 0 -3\" id=\"weapon\"&gt;\n    &lt;a-entity gltf-model=\"#staff\"&gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>And the ball element like this:<\/p><pre class=\"html\">&lt;a-entity id='ball'\n          position=\"0 1 -4\"\n          dynamic-body\n    &gt;\n    &lt;a-entity id='imp-model' gltf-model=\"#imp\"&gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-imp.png\" alt=\"Imp and missing staff\" \/><\/p><p>We can see the imp but the staff is missing. What happened?<\/p><p>The problem is the staff model itself.  The imp model is (mostly) centered inside of its coordinate system, so it is visually positioned where we put it. However the staff model\u2019s center is significantly off from the center of its coordinate system; roughly 15 to 20 meters away.  This is a common issue with models you find online.  To fix it we need to translate the model\u2019s position to account for the offset.  After playing around with the staff model I found that an offset of 2.3, -2.7, -16.3 did the trick. I also had to rotate it 90 degrees to make it horizontal and shift it forward by four meters to make it visible. Wrap the model with an additional entity to apply the translation and rotation.<\/p><pre class=\"html\">&lt;a-entity id=\u201cweapon\u201d  rotation=\"-90 0 0\" position=\"0 0 -4\"&gt;\n    &lt;a-entity position=\"2.3 -2.7 -16.3\" \n              gltf-model=\"#staff\"\n              static-body&gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p>Now we can see the staff, but we still have a problem. The staff is not a simple geometric shape, it\u2019s a full 3d model. The physics engine can\u2019t work directly with a full mesh. Instead it needs to know which primitive object to use.  We could use a box like we did originally, but I chose to go with a <em>sphere<\/em> centered at the end of the staff. That\u2019s the part that the player should actually use to hit the imp, and by making it larger than the staff\u2019s diameter we can make the game easier than it would be in real life. We also need to move the <code>static-body<\/code> definition to the outer entity so that it isn\u2019t affected by the model offset.<\/p><pre class=\"html\">&lt;a-entity rotation=\"-90 0 0\" position=\"0 0 -4\" id='weapon'\n          static-body=\"shape:sphere; sphereRadius: 0.3;\"&gt;\n    &lt;a-entity position=\"2.3 -2.7 -16.3\" \n              gltf-model=\"#staff\" &gt;&lt;\/a-entity&gt;\n&lt;\/a-entity&gt;\n<\/pre><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-imp-staff.png\" alt=\"Imp and Staff\" \/><\/p><h2>Scenery<\/h2><p>We have the core game mechanics working correctly with the new models, let\u2019s add some decorations next. I grabbed more models from SketchFab for a <a href=\"https:\/\/sketchfab.com\/models\/83c39af4bc5c478296501b80ef993855#download\">moon<\/a>, a <a href=\"https:\/\/sketchfab.com\/models\/e6eb5febb0ef4b63b3a90206a8b96135#download\">cauldron<\/a>, a <a href=\"https:\/\/sketchfab.com\/models\/6a0fa67b84a44d51b929d8450fb1ec48\">rock<\/a>, and <a href=\"https:\/\/sketchfab.com\/models\/4ddc25af08b24509aa9aba53c5365967\">two<\/a> different <a href=\"https:\/\/sketchfab.com\/models\/71c0756bf001437ebab9955c12f10d77\">trees<\/a>.  Place them in the scene at different positions.<\/p><pre class=\"html\">&lt;a-assets&gt;\n    &lt;a-asset-item id=\"imp\" src=\"models\/imp\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"staff\" src=\"models\/staff\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"tree1\" src=\"models\/arbol1\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"tree2\" src=\"models\/arbol2\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"moon\" src=\"models\/moon\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"cauldron\" src=\"models\/cauldron\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n    &lt;a-asset-item id=\"rock1\" src=\"models\/rock1\/scene.gltf\"&gt;&lt;\/a-asset-item&gt;\n&lt;\/a-assets&gt;\n...\n&lt;!-- cauldron --&gt;\n&lt;a-entity  position=\"1.5 0 -3.5\" gltf-model=\"#cauldron\"&gt;&lt;\/a-entity&gt;\n&lt;!-- the moon --&gt;\n&lt;a-entity gltf-model=\"#moon\"&gt;&lt;\/a-entity&gt;\n&lt;!-- trees --&gt;\n&lt;a-entity gltf-model=\"#tree2\" position=\"38 8.5 -10\"&gt;&lt;\/a-entity&gt;\n&lt;a-entity gltf-model=\"#tree1\" position=\"33 5.5 -10\"&gt;&lt;\/a-entity&gt;\n&lt;a-entity gltf-model=\"#tree1\" position=\"33 5.5 -30\"&gt;&lt;\/a-entity&gt;\n<\/pre><p>Our little game is starting to look like a real scene!<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-trees-moon.png\" alt=\"Trees and Moon\" \/><\/p><p>The cauldron has bubbles which appeared to animate on SketchFab but they aren\u2019t animating here.  The animation is stored inside the model but it isn\u2019t automatically played without an additional component.  Just add <code>animation-mixer<\/code> to the entity for the cauldron.<\/p><p>The final game has rocks scattered around the field. However, we really don\u2019t want to manually position fifty different rocks. Instead we can write a component to randomly position them for us.<\/p><p>The A-Frame docs explain <a href=\"https:\/\/aframe.io\/docs\/0.7.0\/introduction\/writing-a-component.html\">how to create a component<\/a> so I won\u2019t recount it all here. The gist of it is this: A component has some input properties and then executes code when <code>init()<\/code> is called (and a few other functions). In this case, we want to accept the source of a model, some variables controlling how to distribute the model around the scene, and then have a function which will create N copies of the model.<\/p><p>Below is the code. I know it looks intimidating but it\u2019s actually pretty simple. We\u2019ll go through it step by step.<\/p><pre class=\"html\">&lt;!-- alternate random number generator --&gt;\n&lt;script src=\"js\/random.js\"&gt;&lt;\/script&gt;\n&lt;!-- our `distribute` component --&gt;\n&lt;script&gt;\n    AFRAME.registerComponent('distribute', {\n        schema: {\n            src: {type:'string'},\n            jitter: {type:'vec3'},\n            centerOffset: {type:'vec3'},\n            radius: {type:'number'}\n        },\n        init: function() {\n            const rg = new Random(Random.engines.mt19937().seed(10))\n            const center = new THREE.Vector3(this.data.centerOffset.x, \n                      this.data.centerOffset.y, this.data.centerOffset.z)\n            const jx = this.data.jitter.x\n            const jy = this.data.jitter.y\n            const jz = this.data.jitter.z\n            if($(this.data.src).hasLoaded) {\n                const s = this.data.radius\n                for(let i = -s; i&lt;s; i++) {\n                    for(let j=-s; j&lt;s; j++) {\n                        const el = document.createElement('a-entity')\n                        el.setAttribute('gltf-model', this.data.src)\n                        const offset = new THREE.Vector3(i*s + rg.real(-jx,jx), \n                               rg.real(-jy,jy), \n                               j*s - rg.real(-jz,jz));\n                        el.setAttribute('position', center.clone().add(offset));\n                        el.setAttribute('rotation',{x:0, y:rg.real(-45,45)*Math.PI\/180, z:0})\n                        const scale = rg.real(0.5,1.5)\n                        el.setAttribute('scale',{x:scale,y:scale,z:scale})\n                        $('a-scene').appendChild(el)\n                    }\n                }\n            }\n        }\n    })\n&lt;\/script&gt;\n<\/pre><p>First I import <code>random.js<\/code>. This is a random number generator from <a href=\"https:\/\/github.com\/ckknight\/random-js\">this random-js project<\/a> by Cameron Knight. We could use the standard <code>Math.random()<\/code> function built into Javascript, but I want to ensure that the rocks are always positioned the same way every time the game is run. This other generator lets us provide a seed.<\/p><p>In the first line of the <code>init()<\/code> code you can see that I used the seed 10. I actually tried several different seeds until I found one that I liked the look of.  If I <em>did<\/em> actually want each load to be different, say for different levels of the game, then I could provide a different seed for each level.<\/p><p>The core of the <em>distribute<\/em> component consists of the nested <code>for<\/code> loops. The code creates a grid of entities, each attached to the same model. For each copy, we will translate it from the natural center point of the original model (the <code>modelCenter<\/code> parameter) , adding a random offset using the <code>jitter<\/code> parameter. Jitter represents the maximum amount the rock should move from that grid point.  Using <code>0 0 0<\/code> would be no jitter.  Using <code>0 10 0<\/code> would make the rocks go vertically anywhere between -10 and 10, but not move at all in the horizontal plane.  For this game I used <code>2 0.5 2<\/code> to move them around mostly horizontally but move up and down a tiny bit.  The loop code also gives the rocks a random scale and rotation around the Y axis, just to make the scene look a bit more organic.<\/p><p>This is the final result.<\/p><p><img src=\"https:\/\/2r4s9p1yi1fa2jd7j43zph8r-wpengine.netdna-ssl.com\/files\/2018\/03\/screenshot-rocks.png\" alt=\"Distributed Rocks\" \/><\/p><p>This blog has gotten pretty long and we still haven\u2019t worked on lighting, sounds, or polish.  Let\u2019s continue the game right now in <a href=\"https:\/\/hacks.mozilla.org\/2018\/03\/immersive-aframe-low-poly-part2\/\">Part 2<\/a>.<\/p><p>If you\u2019d like to look at the source to the finished project, I\u2019ve put it all in this <a href=\"https:\/\/github.com\/joshmarinacci\/whack-an-imp\">github repo<\/a>.<\/p>","created_at":"2018-11-07T16:42:42+0000","updated_at":"2018-11-07T16:42:42+0000","published_at":"2018-03-07T08:11:13+0000","published_by":["Josh Marinacci"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":19,"domain_name":"hacks.mozilla.org","preview_picture":"https:\/\/hacks.mozilla.org\/files\/2018\/03\/Screenshot-2018-3-7-http-localhost1.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":262,"title":"Marketing Yourself: Why You Need a Blog - Simple Programmer","url":"https:\/\/simpleprogrammer.com\/marketing-need-blog\/","content":"<p>What would you think if you were interested in buying some new product you heard about, but when you went to the company that created the product\u2019s website you found it wasn\u2019t there, because they didn\u2019t have a website?<\/p><p><strong>Today, we expect pretty much every reputable company to have a functioning website.<\/strong><\/p><p>But, many developers\u2014web developers included\u2014don\u2019t have any kind of online presence of their own.<\/p><p>Sure, you may have social networking accounts, like Facebook and Twitter, but do you have a website that you own which you can point people to as your castle on the web?<\/p><h2>The importance of having a home base for marketing yourself<\/h2><p>I\u2019ve talked before about <a href=\"https:\/\/simpleprogrammer.com\/2013\/11\/18\/become-valuable-software-developer\/\" target=\"_blank\">how important it is to market yourself as a software developer<\/a>, but I\u2019ve never really gone into the details of how.<\/p><p><a href=\"https:\/\/spzone-simpleprogrammer.netdna-ssl.com\/wp-content\/uploads\/2013\/12\/home-base.jpg\"><img title=\"Alicante castle\" src=\"https:\/\/spzone-simpleprogrammer.netdna-ssl.com\/wp-content\/uploads\/2013\/12\/home-base_thumb.jpg\" alt=\"Alicante castle\" width=\"238\" height=\"360\" border=\"0\" \/><\/a><\/p><p>I\u2019ll be creating a series of posts dealing with the subject of marketing yourself over the next few months, starting with this post on what I believe is the cornerstone to any success software developer\u2019s self promotion strategy, building a blog.<\/p><p>As you\u2019ll see in this series, it is actually pretty easy to get started creating a blog\u2014probably simpler than you may think.\u00a0 But, before we get into the details, let\u2019s take a moment to talk about why it is so important to have a home base on the web, especially for a software developer.<\/p><p>It really begins with how you view yourself as a software developer and your software development career.\u00a0 Many, if not most, developers view themselves as a software developer who does a job.\u00a0 For the most part, there is nothing really wrong with this view, but it is not the best way to think about what you do.<\/p><p><strong>Instead, you are better off thinking of yourself as a business.<\/strong>\u00a0 Sure, it may be a one man or one woman business, but the truth of the matter is that you are providing a service to a client, even if that client happens to be your boss.<\/p><p>When you think of yourself and your career as a business that you are building, you suddenly are no longer exempt from needing a web presence.\u00a0\u00a0 Just like we might think it would be pretty bizarre for a company that we do business with to not have a website, your clients and customers will think it is bizarre if you don\u2019t have one\u2014especially if you are a programmer that specializes in web development.<\/p><p><strong>For most developers, your blog will be your main presence, or your home base on the web.<\/strong>\u00a0 Your blog is a chance to tell the world about what you are doing and show what you can do, and to completely control the message and image you present.\u00a0 This is an extremely powerful concept, because it allows you to shape the way potential and present customers and clients see you and can really increase your exposure.<\/p><p>This really is the key to marketing yourself online.<\/p><h2>But, I don\u2019t have anything interesting to talk about<\/h2><p>Hogwash.\u00a0 That excuse is just no good.\u00a0 Everyone has something interesting to talk about.<\/p><p><a href=\"https:\/\/spzone-simpleprogrammer.netdna-ssl.com\/wp-content\/uploads\/2013\/12\/not-interesting.jpg\"><img title=\"Tired sleepless woman\" src=\"https:\/\/spzone-simpleprogrammer.netdna-ssl.com\/wp-content\/uploads\/2013\/12\/not-interesting_thumb.jpg\" alt=\"Tired sleepless woman\" width=\"338\" height=\"231\" border=\"0\" \/><\/a><\/p><p>This is an excuse I hear pretty often, and it seems like a good one\u2014until you really sit down and think about it.<\/p><p>As a beginner, it can seem like you are not good enough; like what you have to say isn\u2019t important; like there are so many other people that have much more valuable advice and opinions.\u00a0 <strong>But, the truth is different people at different levels in their knowledge of a subject, or with different kinds of combination of subjects they have knowledge about, can reach and provide value to different sets of people.<\/strong><\/p><p>Let me break that down a bit.<\/p><p>What I mean to say is that just because you are learning C++ and there are C++ gurus out there with 30 years of experience and more knowledge about C++ then you may ever have, doesn\u2019t mean that you don\u2019t have something valuable to offer.<\/p><p>Sure, <a href=\"http:\/\/herbsutter.com\/2012\/12\/03\/perspective-why-c-is-not-back\/\" target=\"_blank\">Herb Sutter might know more about C++ than you<\/a>, and other C++ experts may gain valuable information from his blog, but can he reach the C++ beginner, like you, that is just starting out as well as you can?\u00a0 Probably not.<\/p><p>The truth is sometimes an amateur can reach other amateurs better than a professional can.<\/p><p>The truth is sometimes a woman can reach other females better than a man can.<\/p><p>The truth is sometimes a younger 20 something person can reach other 20 something people better than a programming dinosaur can or vice versa.<\/p><p><strong>Chances are if you find it interesting, someone else does to.<\/strong>\u00a0 So, stop using that excuse.\u00a0 You can create a blog and it can provide value.\u00a0 You just have to be willing to put in the work.<\/p><h2>I\u2019m not looking to advance my career or sell something<\/h2><p>Again, I have to say this excuse is a bit short-sighted.\u00a0 You might not be looking for another job right now, or to move up the ladder, but chances are, at some time in the future, you will be.<\/p><p><strong>The biggest mistake I see developers make with career advancement is waiting until they need a job to start doing things like networking or blogging.<\/strong><\/p><p>This is a bad idea, because it reeks of desperation and building up momentum, be it with blogging, networking, or something else, requires time.<\/p><p>Ideally, you want to start your blog and start using it to market yourself and your skills, before you need to.\u00a0 Then, if the well ever dries up, you\u2019ll have plenty of prospects.<\/p><p>The same goes with selling something.\u00a0 You may think that you\u2019ll never have something to sell, but if you ever write a book or decide to sell some consulting hours, having a blog can bring you clients and prospects instead of you having to go out and search for them.<\/p><h2>Up next<\/h2><p>Ok, so hopefully, I\u2019ve convinced you to at least consider creating a blog that will serve as your home base on the web\u2013 which will be your primary tool for marketing yourself online.<\/p><p>I can\u2019t tell you how many opportunities have come to me from having this blog that I would have never expected.<\/p><p>But, you may be wondering how to get started with creating your blog.\u00a0 If you are like me, you probably want to know what options you have and how to pick the best one.<\/p><p><a href=\"https:\/\/simpleprogrammer.com\/2013\/12\/30\/marketing-create-blog\/\" target=\"_blank\">In my next post, I\u2019ll talk about the<strong> 3 main options for creating a blog<\/strong>, <strong>give you the one I personally recommend<\/strong>, and give you<strong> one really important piece of advice<\/strong> that you won\u2019t want to ignore.<\/a><\/p><p>Just check back next week, or <a href=\"https:\/\/simpleprogrammer.com\/email\" target=\"_blank\">you can sign up here to get updates<\/a>, so I can let you now when the next post goes live or when something else interesting is happening at Simple Programmer.<\/p><p>And if you can't wait till next week, take a look at this book: <a href=\"http:\/\/www.amazon.com\/gp\/product\/B00FAXNM52\/ref=as_li_ss_tl?ie=UTF8&amp;camp=1789&amp;creative=390957&amp;creativeASIN=B00FAXNM52&amp;linkCode=as2&amp;tag=makithecompsi-20\" target=\"_blank\">Technical Blogging: Turn Your Expertise into a Remarkable Online Presence.<\/a> \u00a0(It is from a fellow developer who gives some tips on creating a successful blog.) \u00a0I really enjoyed this book and found some great tips in it.<\/p><p>Oh, and if you are super excited about the idea of learning to market yourself as a software developer to boost your career,<a href=\"https:\/\/simpleprogrammer.com\/howtomarketyourself\"> I am taking limited pre-sales for my new complete course and package \u201cHow To Market Yourself as a Software Developer.\u201d <\/a>\u00a0I'll announce more about this later, when it is ready for an official launch, but if you are quick, you can get in early and help shape the course.<\/p>","created_at":"2018-11-07T16:42:41+0000","updated_at":"2018-11-07T16:42:41+0000","published_at":"2013-12-23T16:30:11+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"simpleprogrammer.com","preview_picture":"https:\/\/spzone-simpleprogrammer.netdna-ssl.com\/wp-content\/uploads\/2013\/12\/home-base_thumb.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":263,"title":"Breaking Jazz Tracks into Beats with \"Unsupervised\" Clustering","url":"https:\/\/www.edwardshu.com\/ml\/music\/2018\/03\/07\/2018-Breaking-Jazz-Tracks-into-Beats-with-Unsupervised-Clustering.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:42:41+0000","updated_at":"2018-11-07T16:42:41+0000","annotations":[],"reading_time":0,"domain_name":"www.edwardshu.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":264,"title":"Breaking Jazz Tracks into Beats with \"Unsupervised\" Clustering","url":"https:\/\/www.edwardshu.com\/ml\/music\/2018\/03\/07\/2018-Breaking-Jazz-Tracks-into-Beats-with-Unsupervised-Clustering.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:42:41+0000","updated_at":"2018-11-07T16:42:41+0000","annotations":[],"reading_time":0,"domain_name":"www.edwardshu.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":261,"title":"How to Decide What to Blog About - Amy Lynn Andrews","url":"https:\/\/amylynnandrews.com\/how-to-decide-what-to-blog-about\/","content":"<p>Unless your blog is strictly for your own enjoyment, you\u2019re probably hoping to gain readers. So, it\u2019s important to consider what others might want to read.<\/p><p>I\u2019ve been\u00a0watching the blogosphere for years. Below are five overarching areas\u00a0that seem to attract the most readers. Below that are some practical tips for choosing your own blog topic.<\/p><p><img class=\"aligncenter size-full wp-image-16452\" src=\"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/decide-what-to-blog-about.jpg\" alt=\"How to decide what to blog about.\" width=\"1450\" height=\"725\" srcset=\"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/decide-what-to-blog-about.jpg 1450w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/decide-what-to-blog-about-250x125.jpg 250w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/decide-what-to-blog-about-768x384.jpg 768w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/decide-what-to-blog-about-550x275.jpg 550w\" \/><\/p><p><br \/><img class=\"aligncenter size-full wp-image-14112\" src=\"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/what-readers-want.png\" alt=\"what readers want\" width=\"756\" height=\"90\" srcset=\"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/what-readers-want.png 756w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/what-readers-want-140x16.png 140w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/what-readers-want-550x65.png 550w\" \/><\/p><h3>1. Readers want to solve a problem<\/h3><p>What do people get frustrated about? Do you have a solution?\u00a0This is how my blogging took off. I talked to a lot of bloggers who loved to write but got frustrated with the techy side of blogging. I knew I could help solve this problem by <a title=\"How to Blog: A Step-By-Step Guide\" href=\"https:\/\/amylynnandrews.com\/how-to-blog\/\">sharing blogging tips, tools and tutorials in a non-techy way<\/a>.<\/p><h3>2. Readers want to relieve their fears<\/h3><p>What are people afraid of? How can you help ease those fears?\u00a0Maybe you\u2019re a parent who has lost a child. It\u2019s a real and valid fear for a lot of parents. Sharing your story of hope and healing could be very helpful for many people. Or maybe you can offer help to those facing job loss or financial disaster.<\/p><h3>3. Readers want to learn something new<\/h3><p>What would people love to do if only they knew how? What do you know that you could teach them?\u00a0Maybe you\u2019re a whiz at crocheting, you have a knack for writing or you have a unique way of teaching math that makes it easy to understand. A lot of people have projects around the house they would gladly tackle but aren\u2019t sure where to start. Teach them.<\/p><h3>4. Readers want to reach a goal<\/h3><p>What are common goals people have? Have you set and reached some significant goals? Can you spell out how you did it and inspire others on their journey?\u00a0Fitness and weight loss come to mind here, as well as getting out of debt. Pursuing big goals can be disheartening and lonely. Knowing someone else has been there does wonders.<\/p><h3>5. Readers want to be entertained<\/h3><p>Do you have a fascinating story? Do you lead a wildly interesting life? Are you outrageously funny?\u00a0Everyone needs down time and plenty of blogs exist purely to entertain. I\u2019d say this is a trickier path to pursue since there\u2019s no shortage of entertainment to be had on the internet, in magazines and on TV, but it\u2019s doable. The key is providing something totally unique. Of course as a bonus, you could be entertaining and helpful at the same time.\u00a0For example, if your family raises llamas, talk about\u00a0<em>how<\/em>\u00a0you raise llamas not just\u00a0<em>that<\/em>\u00a0you raise llamas. Entertaining + helpful = a great combination.<\/p><p><strong>Related: <a title=\"How to Blog: A Step-By-Step Guide\" href=\"https:\/\/amylynnandrews.com\/how-to-blog\/\">How to Blog: Step-By-Step Guide<\/a><\/strong><\/p><p><img class=\"aligncenter size-full wp-image-14111\" src=\"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/blog-topic-tips.png\" alt=\"blog topic tips\" width=\"551\" height=\"90\" srcset=\"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/blog-topic-tips.png 551w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/blog-topic-tips-140x22.png 140w, https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/blog-topic-tips-550x89.png 550w\" \/><\/p><p>Now that you know what others are looking for, here are some tips to decide where your interests might overlap and therefore make a good blog topic.<\/p><ul id=\"top\"><li><a href=\"#goodidea\">Is my blog idea a good one?<\/a><\/li>\n<li><a href=\"#make-money\">Hey Amy, will I be able to make money\u00a0writing about ____ topic?<\/a><\/li>\n<li><a href=\"#write\">Write for others<\/a><\/li>\n<li><a href=\"#niche\">Pick a niche<\/a><\/li>\n<li><a href=\"#broad-or-specific\">Is this niche too broad or too specific?<\/a><\/li>\n<li><a href=\"#saturated\">Is this niche saturated?<\/a><\/li>\n<li><a href=\"#spend-money\">Are readers in this niche willing to spend money?<\/a><\/li>\n<li><a href=\"#blog-about\">Do you have plenty to blog about?<\/a><\/li>\n<li><a href=\"#ninja\">Rock your ninja-ness<\/a><\/li>\n<li><a href=\"#passionate\">Choose a topic you are genuinely passionate about<\/a><\/li>\n<li><a href=\"#authority\">Choose a niche in which you can be an authority<\/a><\/li>\n<li><a href=\"#wish\">What kind of site do you wish you could find?<\/a><\/li>\n<li><a href=\"#missing\">What\u2019s missing for others?<\/a><\/li>\n<li><a href=\"#different\">Be different<\/a><\/li>\n<li><a href=\"#flexible\">Be flexible<\/a><\/li>\n<li><a href=\"#be-you\">Be you<\/a><\/li>\n<li><a href=\"#paralysis\">Resist paralysis of analysis<\/a><\/li>\n<\/ul><h3 id=\"goodidea\">Is my blog idea a good one?<\/h3><p>I get many, many emails from people asking me\u00a0if I think their blog topic idea is a good one. These\u00a0emails usually include:<\/p><ol><li>An explanation of\u00a0the topic, perhaps with some personal background<\/li>\n<li>Questions about whether\u00a0(a) there are people who would read it and\/or (b) if it\u2019s a topic\u00a0that could make money<\/li>\n<\/ol><p>I <em>completely<\/em>\u00a0understand why\u00a0it\u2019s\u00a0such a common question! I would want to know the same. After all, who wants\u00a0to put time and energy into something that gets no visitors and earns no money?<\/p><p>Here are my thoughts on the subject, and the answer I usually give in response\u2026<\/p><p>First, there are blogs on\u00a0almost every topic imaginable, from <a href=\"http:\/\/awkwardfamilyphotos.com\/\" target=\"_blank\">awkward family photos<\/a> to <a href=\"https:\/\/jellomoldmistress.com\/\" target=\"_blank\">jello mold recipes<\/a>\u00a0and everything in between. In short, the topic itself is usually not the problem. Most topics have the potential to attract visitors and make at least some money.<\/p><p>The main questions are:<\/p><ol><li>Is your passion or excitement about the topic infectious, so those who are interested will be drawn to you and those who aren\u2019t interested will become so?<\/li>\n<li>Can you write about the topic indefinitely? It takes most bloggers a long time to gain traction\u2026and then once they do, they have to keep going. So, you have to be in it for the long haul. Like years.<\/li>\n<li>How will you stand out among the other blogs in the same niche? There are millions of blogs online so it\u2019s very likely others\u00a0on the web are writing about the same topic (or very similar topics) despite how unique or obscure it seems to you.<\/li>\n<li>Are there more than a few\u00a0people in the world who share your interest or would find it entertaining?<\/li>\n<\/ol><p><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"make-money\">Will I be able to make money\u00a0writing about ____ topic?<\/h3><p>What\u2019s nice about blogging is that just about any topic <em>can<\/em> generate income. However there is no way to guarantee a particular topic <em>will<\/em> generate income. For example, there are plenty of photography blogs (or horse blogs or hunting blogs or craft blogs or decor blogs) that make a lot of income, and there are many more that don\u2019t.<\/p><p>It\u2019s not so much the topic that\u2019s important, it\u2019s becoming the go-to resource for that topic. Write in a way others are drawn to and want to read, and create or promote products your readers are willing to buy. Stand out. Give more. Build relationships with people who want to read about that topic.<\/p><p><strong>Pro tip:<\/strong>\u00a0A quick way\u00a0to tell if others\u00a0are making money in a particular topic or niche is to do a Google search. Do ads show up at the top of the search results page when you type in main topics? If so, people are making money in that niche (otherwise they wouldn\u2019t spend money on advertising) and there\u2019s potential for you too. <em>Note:<\/em> The absence of ads doesn\u2019t necessarily indicate you <em>couldn\u2019t<\/em> make money in a particular niche. Deeper digging often reveals sites that do make money but don\u2019t choose to advertise on Google.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"write\">Write for others<\/h3><p>A lot of new bloggers fail to think beyond their own interests when starting a blog (see above). Your blog should undoubtedly be an extension of you, but if you\u2019re not writing for the benefit of others at the same time, you might as well just keep a diary. \ud83d\ude42<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"niche\">Pick a niche<\/h3><p>Instead of just writing whatever comes to mind, try to write around a general topic. (This is called your niche.) Not only will it be easier to stay on task, it\u2019ll be a lot easier for readers to track with you.<\/p><p>While not required, a niche provides focus and direction, making your blog\u2019s purpose easily understood and defined, not only by you, but by your visitors as well.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><p>If your blog\u2019s topic is too broad, it\u2019s hard to compete with, and stand out from, all the other blogs and websites in your niche. On the other hand, if your topic is too narrow, the pool of interested readers will be too small to gain any traction.<\/p><p>For example, \u201cphotography\u201d is a very\u00a0broad topic. On the other hand, \u201cphotography in 50-Person Town, USA\u201d doesn\u2019t give you a very large audience. \u201cBlack and white photography\u201d is better. \u201cBlack and white photography in National Parks\u201d is\u00a0better still. \u201cBlack and white photography in Yellowstone\u201d might be even better. The goal is to find a topic with a good number of interested people and plenty of potential subtopics,\u00a0but\u00a0a topic that not so many other people are writing about. Do some research and googling to narrow it down.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"saturated\">Is this niche saturated?<\/h3><p>Back in the day, when there weren\u2019t so many blogs online, you could almost pick any topic and run with it. Now, not so much. There are definitely niches that are really, really full and therefore, difficult to break into. How do you know? If you can easily find several dozen popular blogs on the topic, you might rethink your topic.<\/p><p>However, just because a niche is big doesn\u2019t mean it\u2019s a bad idea to choose it. After all, a large niche means\u00a0there\u2019s a market for it. Spend time\u00a0watching the main players.\u00a0Knowing your way around will help you fine-tune the \u201cthing\u201d that will make you stand out.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"spend-money\">Are readers in this niche willing to spend money?<\/h3><p>If you <a title=\"How to Make Money Blogging \u2013 2014 Edition\" href=\"https:\/\/amylynnandrews.com\/how-to-make-money-blogging\/\">hope to generate income<\/a>, this is an important question. Think about the intersection between your niche and your audience. For example, if you\u2019re hoping to promote high-end clothing products, it\u2019s probably not a great idea to target struggling college students. Another way to look at it: are other blogs in this niche earning money? Finding this out is easier said than done, but keep your eyes and ears peeled. Do those blogs have advertisers? Are the blogs active, engaging and growing?<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"blog-about\">Do you have plenty to blog about?<\/h3><p>Choose a topic that you can write about regularly and indefinitely. Remember, you\u2019re in this for the long haul. If you post once a week, that amounts to 52 posts a year. Three times a week? 156 posts. Five times a week? 260 posts. \u00a0And that\u2019s just barely getting started!\u00a0Don\u2019t choose something so narrow that you run out of writing fodder after only a few weeks or months.<\/p><p>A good way to test this is to brainstorm possible posts or subtopics pertaining to your main blog topic.\u00a0If you can easily come up with a list of several dozen with additional ideas about how to branch out, it\u2019s probably a good sign. If, however, you can\u2019t think of many, you might need to rethink your choice.\u00a0Another way to work around this problem is to have \u00a0a multi-author blog.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"ninja\">Rock your ninja-ness<\/h3><p>If you aren\u2019t sure you have much to offer, I love what\u00a0<a href=\"http:\/\/www.copyblogger.com\/author\/sonia\/\" target=\"_blank\">Sonia Simone<\/a>\u00a0says,\u00a0\u201cEven if you\u2019re only pretty good, but not a ninja, you\u2019re still a ninja to someone.\u201d<strong>\u00a0<\/strong>Find that thing about which you have a decent amount of know-how and go with it. Chances are there are others who will appreciate what you have to say.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"passionate\">Choose a topic you are genuinely passionate about<\/h3><p>If you don\u2019t have a genuine interest in what you\u2019re writing about, it will be a drag and a burden. If you talk about the topic among your real life friends and they just want you to be quiet already, it\u2019s a great topic to blog about.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><p>The key here is to think smaller. I\u2019ve always been impressed with Carrie from <a href=\"http:\/\/springsbargains.com\/\" target=\"_blank\">Springs Bargains<\/a> who did just this. She started a deal blog, but instead of starting a general one like so many others, she purposefully targeted Colorado Springs. She is absolutely <em>the<\/em> authority on deals in Colorado Springs.<\/p><p>\u201cNiching down,\u201d as some say, or, choosing a narrower niche, may have a smaller pool of potential readers, but you might be able to gain a following quicker too.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"wish\">What kind of site do you wish you could find?<\/h3><p>Sometimes a good way to determine a viable blog niche is to ask yourself what you\u2019ve found to be lacking online.\u00a0After all, if you\u2019re looking for it, someone else might be looking for it too. This is how I started my first blog.\u00a0Back in 2004, I was a struggling pastor\u2019s wife. I knew I couldn\u2019t be the only pastor\u2019s wife having a difficult time, so I searched online for others with whom I could relate. I couldn\u2019t find any, so I started my own. Another way to look at it: what group is being ignored online?<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"missing\">What\u2019s missing on other blogs?<\/h3><p>When I asked this question in 2010, my blog started taking off. There are a gazillion blogs about blogging and making money blogging. What I noticed though, is that a lot of them say things like, \u201cWanna start a blog? Great! You\u2019ll need hosting and a domain and then here\u2019s how to blog\u2026\u201d\u00a0Not a lot of them explained exactly\u00a0<em>how\u00a0<\/em>to choose a domain\u00a0and\u00a0<a title=\"How to Start a Blog\" href=\"https:\/\/amylynnandrews.com\/how-to-start-a-blog\/\"><em>how<\/em>\u00a0to purchase &amp; set up hosting<\/a>\u00a0in a step-by-step way. So, even in this huge niche, I decided to tackle the basics where a lot of people seemed to get lost (like I did when I first started!). Find a hole and fill it.<\/p><p>As you hang out in your potential niche, continually ask yourself what\u2019s missing. What are people looking for? What are <em>you<\/em> looking for?\u00a0Read comments, get involved in forums, Twitter and Facebook and keep your ears peeled for hints about what people want, but can\u2019t find.<\/p><p><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"different\">Be different<\/h3><p>Bloggers tend to copy what other bloggers do. This is absolutely valuable in many ways, but it\u2019s not so good when it puts your blog right smack in the middle of average. Brainstorm ways to do things differently.<\/p><ul><li>Do most bloggers in your niche write long posts? Why not keep yours short?<\/li>\n<li>Do most bloggers in your niche write words? Why not vlog?<\/li>\n<li>Do most bloggers in your niche post a few times a week? Why not post every day?<\/li>\n<li>Or maybe you could start a unique feature or incorporate an interesting twist \u2014 something no one else has done or something you saw someone in a different niche do that you think might work in yours.<\/li>\n<\/ul><p>You may have heard of <a href=\"http:\/\/thepioneerwoman.com\/\" target=\"_blank\">The Pioneer Woman<\/a>. Way back when, she was one of many bloggers blogging about their lives as a mom. But one day she started recounting the tale of how her (a city girl) and her husband (a cowboy) met and fell in love. Her readers ate it up. Coupled with her outstanding photography and love for cooking, she subsequently rose to the top of her niche\u2026and the entire mommy blogging world.<\/p><p>What hasn\u2019t been done before? Try it and see if it propels you to the top!<\/p><p><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"flexible\">Be flexible<\/h3><p>Once you choose a niche, don\u2019t feel like you\u2019re committed to it for life. Blogging is very fluid and changes constantly.\u00a0Being flexible and taking advantage of ways to be different will serve you well.\u00a0In fact, expect it.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"be-you\">Be you<\/h3><p>One of the biggest mistakes you can make is to find a blog you love\u2026and then try to duplicate it. You must differentiate yourself. One of the greatest things about blogs is they afford us the opportunity to get to know the individual behind the blog.\u00a0Blogging is part of social media because it\u2019s just that\u2014social.<\/p><p>Let your personality come through. If you\u2019re goofy, be goofy. If you\u2019re feisty, be feisty. If you\u2019re contemplative, be contemplative.\u00a0Your readers will be drawn to what you have to say, but they will also be drawn to who you are.\u00a0BE YOU.<br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><h3 id=\"paralysis\">Resist paralysis of analysis<\/h3><p>Many people get stuck at this point in the process because they\u2019re terrified of making the \u201cwrong choice.\u201d While a well-chosen niche is a benefit, one wonderful thing about the internet is how forgiving it is.\u00a0<strong>Don\u2019t be afraid to dive in and figure it out as you go. We all do that.<\/strong>\u00a0Better to do that than to do nothing at all. Just start. <a title=\"How to Blog: A Step-By-Step Guide\" href=\"https:\/\/amylynnandrews.com\/how-to-blog\/\">How about right now?<\/a><br \/><small><a href=\"#top\">Back to top<\/a><\/small><\/p><p><strong>Related: <a title=\"How to Make Money Blogging\" href=\"https:\/\/amylynnandrews.com\/how-to-make-money-blogging\/\">How to Make Money Blogging<\/a><\/strong><\/p>","created_at":"2018-11-07T16:42:37+0000","updated_at":"2018-11-07T16:42:37+0000","published_at":"2013-12-13T00:47:09+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":13,"domain_name":"amylynnandrews.com","preview_picture":"https:\/\/amylynnandrews.com\/wp-content\/uploads\/2013\/12\/decide-what-to-blog-about.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":258,"title":"The Portacle Common Lisp Development Environment","url":"https:\/\/portacle.github.io","content":"<section id=\"what\"><h2>What This Is<\/h2><img class=\"logo portacle\" alt=\"portacle logo\" src=\"https:\/\/portacle.github.io\/portacle.svg\" \/><p>\n          Portacle is a complete IDE for Common Lisp that you can take with you on a USB stick. It is multi-platform and can be run on Windows, OS X, and Linux. Since it does not require any complicated installation process, it is <a class=\"xref\" href=\"#get\">set up and running<\/a> in no time.<\/p><p>\n          It lends itself very well both to absolute beginners of Lisp that just need a good starting point, as well as advanced users that want to minimise the time spent getting everything ready.\n        <\/p><p>\n          The following software packages are included in Portacle:\n        <\/p><article id=\"what-contents\"><section id=\"what-contents-emacs\"><img class=\"logo emacs\" alt=\"emacs logo\" src=\"https:\/\/portacle.github.io\/emacs.svg\" \/><h3>Emacs<\/h3><p>\n              Emacs is a very flexible and extensible editor that has been around for a long time. Due to its history and huge community, it is capable of covering pretty much any area of programming and more.\n            <\/p><p>\n              Bundled with Emacs are some very useful packages, the major ones being <a href=\"https:\/\/common-lisp.net\/project\/slime\/\">Slime<\/a>, <a href=\"https:\/\/magit.vc\/\">Magit<\/a>, and <a href=\"http:\/\/company-mode.github.io\/\">Company<\/a>.\n            <\/p><\/section><section id=\"what-contents-sbcl\"><img class=\"logo sbcl\" alt=\"sbcl logo\" src=\"https:\/\/portacle.github.io\/sbcl.svg\" \/><h3>SBCL<\/h3><p>\n              SBCL is one of the fastest and most capable Lisp implementations available today.\n            <\/p><\/section><section id=\"what-contents-quicklisp\"><img class=\"logo quicklisp\" alt=\"quicklisp logo\" src=\"https:\/\/portacle.github.io\/quicklisp.svg\" \/><h3>Quicklisp<\/h3><p>\n              Quicklisp is the standard Lisp package manager that allows you to retrieve and manage over a thousand libraries with a few keystrokes.\n            <\/p><\/section><section id=\"what-contents-git\"><img class=\"logo git\" alt=\"git logo\" src=\"https:\/\/portacle.github.io\/git.svg\" \/><h3>Git<\/h3><p>\n              Git is perhaps the most widely used version control system today. It allows you to manage changes in code over small and large projects fast and efficiently.\n            <\/p><\/section><\/article><\/section><section id=\"get\"><h2>How to Get It<\/h2><p>\n          Portacle is rather straight-forward to setup and use. All you need to do is extract an archive. After that, you can freely move its installation folder around and even take it with you on a memory stick.\n        <\/p><p>\n          If you are new to Emacs, Lisp, or both, you should also read the section <a class=\"xref\" href=\"#use\">after this one<\/a> once you successfully completed the installation.\n        <\/p><article id=\"get-win\"><img class=\"logo win\" alt=\"windows logo\" src=\"https:\/\/portacle.github.io\/win.svg\" \/><h3>Windows<\/h3><p>\n            Download the <a class=\"download win\" href=\"https:\/\/github.com\/portacle\/portacle\/releases\/download\/1.2c\/win-portacle.exe\">latest release<\/a> and run it. It will ask you where to install it to, defaulting to your home folder. Note that you do not need to append <code>portacle<\/code> to the end of the path. After extraction, you can launch it by double-clicking the <code>portacle.exe<\/code>.\n          <\/p><p>\n            Note that <code>portacle.exe<\/code> is tied to the <code>portacle<\/code> directory and needs everything within it to function properly. You can however create a shortcut to the exe to reach it more easily from your desktop.\n          <\/p><\/article><article id=\"get-mac\"><img class=\"logo mac\" alt=\"mac os x logo\" src=\"https:\/\/portacle.github.io\/mac.svg\" \/><h3>Mac OS X<\/h3><p>\n            Download the <a class=\"download mac\" href=\"https:\/\/github.com\/portacle\/portacle\/releases\/download\/1.2b\/mac-portacle.dmg\">latest release<\/a> and extract it. Due to \"security\" reasons on OS X you must then move the <code>Portacle.app<\/code> within the extracted directory into another directory like <code>projects\/<\/code> and back again using Finder. From then on you can launch it by double-clicking the <code>Portacle.app<\/code>. The first time you launch it, OS X is going to block the application as it is \"from an unidentified developer.\" You need to open System Preferences, go to Security, and click the Open Anyway button to mark the application as trusted. After that it should work straight away.\n          <\/p><p>\n            Note that you cannot copy the <code>Portacle.app<\/code> outside of the <code>portacle<\/code> directory. You must take the whole directory with you. You can however drag the app into your dock.\n          <\/p><\/article><article id=\"get-lin\"><img class=\"logo lin\" alt=\"linux logo\" src=\"https:\/\/portacle.github.io\/lin.svg\" \/><h3>Linux<\/h3><p>\n            Download the <a class=\"download lin\" href=\"https:\/\/github.com\/portacle\/portacle\/releases\/download\/1.2b\/lin-portacle.tar.xz\">latest release<\/a> and extract it. You can then launch it by double-clicking <code>portacle.desktop<\/code>. The file may also be presented to you as just <code>Portacle<\/code>.\n          <\/p><p>\n            Note that you cannot move or copy <code>portacle.desktop<\/code> elsewhere. It has to reside in the <code>portacle<\/code> directory for it to work.\n          <\/p><\/article><\/section><section id=\"use\"><h2>Using It<\/h2><p>\n          Upon launch it might fizzle around for a few seconds and then present you with a window that looks something like this:\n        <\/p><img class=\"full\" alt=\"portacle initial window\" src=\"https:\/\/portacle.github.io\/portacle-start.png\" \/><p>\n          The window should be divided up into two regions, each of which display something called a buffer. One of the buffers, the one displaying <code>CL-USER&gt;<\/code>, is the Lisp prompt. The other is the <code>scratch<\/code> buffer and should also contain some hints on how to reach a reference guide.\n        <\/p><p>\n          As a quick test, click inside the Lisp prompt buffer, and type in the following snippet:\n        <\/p><pre class=\"noselect\" title=\"Come on, just type it in real quick.\">(progn (format T \"Hey, what's your name?~%\")\n       (format T \"Hello, ~a!\" (read-line)))<\/pre><p>\n          You will notice that it automatically inserts closing parens and the end of the quote for you. To run it, move the cursor to the end by either clicking or using the arrow keys, and hit your <kbd title=\"Return\">\u23ce<\/kbd> key.\n        <\/p><img alt=\"portacle repl greeting\" src=\"https:\/\/portacle.github.io\/portacle-greet.png\" \/><p>\n          Congratulations! You just ran your first program in Portacle.\n        <\/p><p>\n          You should probably also configure Portacle's default settings. To do this, press <kbd>Alt<\/kbd>+<kbd>X<\/kbd> (<kbd>\u2318<\/kbd>+<kbd>X<\/kbd> on Mac), then type <code>portacle-configure<\/code> followed by <kbd title=\"Return\">\u23ce<\/kbd>. You should then see a prompt at the bottom of the window, in an area called the \"minibuffer\".\n        <\/p><img alt=\"portacle minibuffer\" src=\"https:\/\/portacle.github.io\/portacle-minibuffer.png\" \/><p>\n          Once you've answered all the questions, Portacle should be set and ready to go. Make sure to read the help document. It should explain all the basic terminology and commands on how to use the editor. To access it, press <kbd>Ctrl<\/kbd>+<kbd>H<\/kbd> <kbd>H<\/kbd>.\n        <\/p><\/section><section id=\"help\"><h2>Getting Help or Contributing<\/h2><p>\n          If you need help using Portacle, or would like to contribute to its development, please see the official <a href=\"https:\/\/github.com\/portacle\/portacle\/issues\">repository<\/a>. You can also chat directly with the developers over IRC, using the <a href=\"irc:\/\/irc.freenode.net\/%23shirakumo\">#shirakumo<\/a> channel on <a href=\"https:\/\/freenode.net\">Freenode<\/a>.\n        <\/p><\/section>","created_at":"2018-11-07T16:42:36+0000","updated_at":"2018-11-07T16:42:36+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"portacle.github.io","preview_picture":"https:\/\/portacle.github.io\/portacle.svg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":259,"title":"A Practical Introduction to Blockchain with Python \/\/ Adil Moujahid \/\/ Data Analytics and more","url":"http:\/\/adilmoujahid.com\/posts\/2018\/03\/intro-blockchain-bitcoin-python\/","content":"<section class=\"post\"><p>Blockchain is arguably one of the most significant and disruptive technologies that came into existence since the inception of the Internet. It's the core technology behind Bitcoin and other crypto-currencies that drew a lot of attention in the last few years.<\/p><p>As its core, a blockchain is a distributed database that allows direct transactions between two parties without the need of a central authority. This simple yet powerful concept has great implications for various institutions such as banks, governments and marketplaces, just to name a few. Any business or organization that relies on a centralized database as a core competitive advantage can potentially be disrupted by blockchain technology.<\/p><p>Putting aside all the hype around the price of Bitcoin and other cryptocurrencies, the goal of this blog post is to give you a practical introduction to blockchain technology. Sections 1 and 2 cover some core concepts behind blockchain, while section 3 shows how to implement a blockchain using Python. We will also implement 2 web applications to make it easy for end users to interact with our blockchain.<\/p><p>Please note that I'm using Bitcoin here as a medium for explaning the more general technology of \"Blockchain\", and most of the concepts described in this post are applicable to other blockchain use cases and crypto-currencies. <\/p><p>Below is an animated gif of the two web apps that we will build in section 3.<\/p><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-simulation.gif\" alt=\"image\" \/><\/div><p>It all started with a white paper released in 2008 by an unknown person or entity using the name Satoshi Nakamoto. The white paper was titled <a href=\"https:\/\/bitcoin.org\/bitcoin.pdf\">\u201cBitcoin: A Peer-to-Peer Electronic Cash System\u201d<\/a> and it laid the foundation of what later became known as Blockchain. In the original Bitcoin white paper, Satoshi described how to build a peer-to-peer electronic cash system that allows online payments to be sent directly from one party to another without going through a centralized institution. This system solves an important problem in digital money called double-spending.<\/p><h2>1.1. What is Double-Spending?<\/h2><p>Suppose that Alice wants to pay Bob 1$. If Alice and Bob use physical cash, then Alice will not longer have the 1$ after the transaction is executed. If Alice and Bob use digital money, then the problem gets more complicated. Digital money is in digital form and can be easily duplicated. If Alice sends a digital file worth 1$ to Bob by email for example, Bob cannot know for sure if Alice has deleted her copy of the file. If Alice still has the 1$ digital file, then she can choose to send the same file to Carol. This problem is called double-spending.<\/p><div><p><img alt=\"Alt Text\" src=\"http:\/\/adilmoujahid.com\/images\/blockchain-double-spending.png\" \/><\/p><\/div><p>One way of solving the double-spending problem is to have a trusted third party (a bank for example) between Alice, Bob and all other participants in the network. This third party is responsible for managing a centralized ledger that keeps track of and validates all the transactions in the network. The drawback of this solution is that for the system to function, it requires trust in a centralized third party.<\/p><h2>1.2. Bitcoin: A Decentralized Solution for the Double-Spending Problem<\/h2><p>To solve the double-spending problem, Satoshi proposed a public ledger, i.e., Bitcoin\u2019s blockchain to keep track  of all transactions in the network. Bitcoin\u2019s blockchain has the following characteristics:<\/p><ul><li>Distributed:  The ledger is replicated across a number of computers, rather than being stored on a central server. Any computer with an internet connection can download a full copy of the blockchain.<\/li>\n<li>Cryptographic: Cryptography is used to make sure that the sender owns the bitcoin that she's trying to send, and to decide how the transactions are added to the blockchain.<\/li>\n<li>Immutable: The blockchain can be changed in append only fashion. In other words, transactions can only be added to the blockchain but cannot be deleted or modified. <\/li>\n<li>Uses Proof of Work (PoW): A special type of participants in the network called miners compete on searching for the solution to a cryptographic puzzle that will allow them to add a block of transactions to Bitcoin\u2019s blockchain. This process is called Proof of Work and it allows the system to be secure (more on this later).<\/li>\n<\/ul><div><p><img alt=\"Alt Text\" src=\"http:\/\/adilmoujahid.com\/images\/blockchain-cash-bitcoin.png\" \/><\/p><\/div><p>Sending bitcoin money goes as follows: <\/p><ul><li>Step 1 (one-time effort): Create a bitcoin wallet. For a person to send or receive bitcoins, she needs to create a bitcoin wallet. A bitcoin wallet stores 2 pieces of information: A private key and a public key. The private key is a secret number that allows the owner to send bitcoin to another user, or spend bitcoins on services that accept them as payment method. The public key is a number that is needed to receive bitcoins. The public key is also referred to as bitcoin address (not entirely true, but for simplicity we will assume that the public key and the bitcoin address are the same). Note that the wallet doesn\u2019t store the bitcoins themselves. Information about bitcoins balances are stored on the Bitcoin\u2019s blockchain. <\/li>\n<li>Step 2: Create a bitcoin transaction. If Alice wants to send 1 BTC to Bob, Alice needs to connect to her bitcoin wallet using her private key, and create a transaction that contains the amount of bitcoins she wants to send and the address where she wants to send them (in this case Bob\u2019s public address). <\/li>\n<li>Step 3: Broadcast the transaction to Bitcoin\u2019s network. Once Alice creates the bitcoin transaction, she needs to broadcast this transaction to the entire Bitcoin\u2019s network.<\/li>\n<li>Step 4: Confirm the transaction. A miner listening to Bitcoin\u2019s network authenticates the transaction using Alice's public key, confirms that Alice has enough bitcoins in her wallet (in this case at least 1 BTC), and adds a new record to Bitcoin\u2019s Blockchain containing the details of the transaction. <\/li>\n<li>Step 5: Broadcast the blockchain change to all miners. Once the transaction is confirmed, the miner should broadcast the blockchain change to all miners to make sure that their copies of the blockchain are all in sync.<\/li>\n<\/ul><p>The goal of this section is to go deeper into the technical building blocks that power the blockchain. We will cover public key cryptography, hashing functions, mining and security of the blockchain.<\/p><h2>2.1. Public Key Cryptography<\/h2><p>Public-key cryptography, or asymmetrical cryptography, is any cryptographic system that uses pairs of keys: public keys which may be disseminated widely, and private keys which are known only to the owner. This accomplishes two functions: authentication, where the public key verifies a holder of the paired private key sent the message, and encryption, where only the paired private key holder can decrypt the message encrypted with the public key. [1]<\/p><p><a href=\"https:\/\/en.wikipedia.org\/wiki\/RSA_(cryptosystem)\">RSA<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Elliptic_Curve_Digital_Signature_Algorithm\">Elliptic Curve Digital Signature (ECDSA)<\/a> are the most popular public-key cryptography algorithms. <\/p><p>In the case of Bitcoin, ECDSA algorithm is used to generate Bitcoin wallets. Bitcoin uses a variety of keys and addresses, but for the sake of simplicity, we will assume in this blog post that each Bitcoin wallet has one pair of private\/public keys and that a Bitcoin address is the wallet's public key. I recommend this <a href=\"https:\/\/en.bitcoin.it\/wiki\/Technical_background_of_version_1_Bitcoin_addresses\">article<\/a>, if you're interested in the complete technical details of Bitcoin wallets. <\/p><p>To send or receive BTCs, a user starts by generating a wallet which contains a pair of private and public keys. If Alice wants to send Bob some BTCs, she creates a transaction in which she enters both her and Bob's public keys, and the amount of BTCs she wants to send. She then sign the transaction using her private key. A computer on the blockchain uses Alice's public key to verify that the transaction is authentic and adds the transaction to a block that will be later added to the blockchain. <\/p><div><p><img alt=\"Alt Text\" src=\"http:\/\/adilmoujahid.com\/images\/blockchain-public-crypto.png\" \/><\/p><\/div><h2>2.2. Hashing Functions and Mining<\/h2><p>All Bitcoin transactions are grouped in files called blocks.  Bitcoin adds a new block of transactions every 10 minutes. Once a new block is added to the blockchain, it becomes immutable and can't be deleted or modified.\nA special group of participants in the network called miners (computers connected to the blockchain) are responsible for creating new blocks of transactions. A miner has to authenticate each transaction using the sender's public key, confirm that the sender has enough balance for the requested transaction, and add the transaction to the block. \nMiners are completely free to choose which transactions to include in the blocks, therefore the senders need to include a transaction fee to incentivise the miners to add their transactions to the blocks. <\/p><p>For a block to be accepted by the blockchain, it needs to be \"mined\". To mine a block, miners need to find an extremely rare solution to a cryptographic puzzle. If a mined block is accepted by the blockchain, the miner receive a reward in bitcoins which is an additional incentive to transaction fees. The mining process is also referred to as Proof of Work (PoW), and it's the main mechanism that enables the blockchain to be trustless and secure (more on blockchain security later).<\/p><h3>Hashing and Blockchain's Cryptographic Puzzle<\/h3><p>To understand the blockchain's cryptographic puzzle, we need to start with hash functions. A hash function is any function that can be used to map data of arbitrary size to data of fixed size. The values returned by a hash function are called hashes. Hash functions are usually used to accelerate database lookup by detecting duplicated records, and they are also widely used in cryptography.  A cryptographic hash function allows one to easily verify that some input data maps to a given hash value, but if the input data is unknown, it is deliberately difficult to reconstruct it by knowing the stored hash value. [2]<\/p><p>Bitcoins uses a cryptographic hash function called SHA-256. SHA-256 is applied to a combination of the block's data (bitcoin transactions) and a number called nonce. By changing the block data or the nonce, we get completely different hashes. For a block to be considered valid or \"mined\", the hash value of the block and the nonce needs to meet a certain condition. For example, the four leading digits of the hash needs to be equal to \"0000\". We can increase the mining complexity by making the condition more complex, for example we can increase the number of 0s that the hash value needs to start with.<\/p><p>The cryptograhic puzzle that miners need to solve is to find a nonce value that makes the hash value satisfies the mining condition. You can use the app below to simulate block mining. When you type in the \"Data\" text box or change the nonce value, you can notice the change in the hash value. When you click the \"Mine\" button, the app starts with a nonce equals to zero, computes the hash value and checks if the leading four digits of the hash value is equal to \"0000\". If the leading  four digits are not equal to \"0000\", it increments the nonce by one and repeats the whole process until it finds a nonce value that satisify the condition. If the block is considered mined, the background color turns green.<\/p><div class=\"container col-sm-12\"><div id=\"block0chain0well\" class=\"well\"><form class=\"form-horizontal\" action=\"action\">\n  <div class=\"form-group\"><label for=\"block0chain0nonce\" class=\"col-sm-2 control-label\">\n      Nonce:\n    <\/label><\/div>\n  \n  \n<\/form><\/div><\/div><h2>2.3. From Blocks to Blockchain<\/h2><p>As discussed in the previous section, transactions are grouped in blocks and blocks are appended to the blockchain. In order to create a chain of blocks, each new block uses the previous block\u2019s hash as part of its data. To create a new block, a miner selects a set of transactions, adds the previous block\u2019s hash and mines the block in a similar fashion described above. <\/p><p>Any changes to the data in any block will affect all the hash values of the blocks that come after it and they will become invalid. This give the blockchain its immutability characteristic. <\/p><div><p><img alt=\"Alt Text\" src=\"http:\/\/adilmoujahid.com\/images\/blockchain-from-blocks.png\" \/><\/p><\/div><p>You can use the app below to simulate a blockchain with 3 blocks. When you type in the \"Data\" text box or change the nonce value, you can notice the change in the hash value and the \"Prev\" value (previous hash) of the next block. You can simulate the mining process by clicking on the \u201cMine\u201d button of each individual block. After mining the 3 blocks, try changing the data in block 1 or 2, and you will notice that all the blocks that come after become invalid.<\/p><div class=\"container col-sm-12\"><div id=\"block1chain1well\" class=\"well\"><form class=\"form-horizontal\" action=\"action\">\n      <div class=\"form-group\"><label for=\"block1chain1number\" class=\"col-sm-2 control-label\">Block:<\/label><\/div>\n      <div class=\"form-group\"><label for=\"block1chain1nonce\" class=\"col-sm-2 control-label\">Nonce:<\/label><\/div>\n      \n      \n      \n      \n    <\/form><\/div><div id=\"block2chain1well\" class=\"well\"><form class=\"form-horizontal\" action=\"action\">\n      <div class=\"form-group\"><label for=\"block2chain1number\" class=\"col-sm-2 control-label\">Block:<\/label><\/div>\n      <div class=\"form-group\"><label for=\"block2chain1nonce\" class=\"col-sm-2 control-label\">Nonce:<\/label><\/div>\n      \n      \n      \n      \n    <\/form><\/div><div id=\"block3chain1well\" class=\"well\"><form class=\"form-horizontal\" action=\"action\">\n      <div class=\"form-group\"><label for=\"block3chain1number\" class=\"col-sm-2 control-label\">Block:<\/label><\/div>\n      <div class=\"form-group\"><label for=\"block3chain1nonce\" class=\"col-sm-2 control-label\">Nonce:<\/label><\/div>\n      \n      \n      \n      \n    <\/form><\/div><\/div><p>Both mining simulators above were adapted from Anders Brownworth's excellent <a href=\"https:\/\/anders.com\/blockchain\/blockchain.html\">Blockchain Demo<\/a>.<\/p><h2>2.4. Adding Blocks to the Blockchain<\/h2><p>All the miners in the Bitcoin network compete with each other to find a valid block that will be added to the blockchain and get the reward from the network. Finding a nonce that validated a block is rare, but because of the number of miners, the probability of a miner in the network validating a block is extremely high. The first miner to submit a valid block gets his block added to the blockchain and receives the reward in bitcoins. But what happens if two miners or more submit their blocks at the same time?<\/p><h3>Resolving Conflicts<\/h3><p>If 2 miners solve a block at almost the same time, then we will have 2 different blockchains in the network, and we need to wait for the next block to resolve the conflict. Some miners will decide to mine on top of blockchain 1 and others on top of blockchain 2. The first miner to find a new block resolves the conflict. If the new block was mined on top of blockchain 1, then blockchain 2 becomes invalid, the reward of the previous block goes to the miner from blockchain 1 and the transactions that were part of blockchain 2 and weren\u2019t added to the blockchain go back to the transactions pool and get added to the next blocks. In short, if there is a conflict on the blockchain, then the the longest chain wins.<\/p><div><p><img alt=\"Alt Text\" src=\"http:\/\/adilmoujahid.com\/images\/blockchain-the-longest-chain-wins.png\" \/><\/p><\/div><h2>2.5. Blockchain and Double-Spending<\/h2><p>In this section, we will cover the most popular ways for performing double-spending attacks on the blockchain, and the measures that users should take to prevent damages from them.<\/p><h3>Race Attack<\/h3><p>An attacker sends the same coin in rapid succession to two different addresses. To prevent from this attack, it is recommended to wait for at least one block confirmation before accepting the payment. [3]<\/p><h3>Finney Attack<\/h3><p>An attacker pre-mines a block with a transaction, and spends the same coins in a second transaction before releasing the block. In this scenario, the second transaction will not be validated. To prevent from this attack, it is recommended to wait for at least 6 block confirmations before accepting the payment. [3] <\/p><h3>Majority Attack (also called 51% attack)<\/h3><p>In this attack, the attacker owns 51% of the computing power of the network. The attacker starts by making a transaction that is brodcasted to the entire network, and then mines a private blockchain where he double-spends the coins of the previous transaction. Since the attacker owns the majority of the computing power, he is guaranteed that he will have at some point a longer chain than the \"honest\" network. He can then release his longer blockchain that will replace the \"honest\" blockchain and cancel the original transaction. This attack is highly unlikely, as it\u2019s very expensive in blockchain networks like Bitcoin. [4]<\/p><p>In this section, we will implement a basic blockchain and a blockchain client using Python. Our blockchain will have the following features:<\/p><ul><li>Possibility of adding multiple nodes to the blockchain<\/li>\n<li>Proof of Work (PoW)<\/li>\n<li>Simple conflict resolution between nodes<\/li>\n<li>Transactions with RSA encryption<\/li>\n<\/ul><p>Our blockchain client will have the following features:<\/p><ul><li>Wallets generation using Public\/Private key encryption (based on RSA algorithm)<\/li>\n<li>Generation of transactions with RSA encryption <\/li>\n<\/ul><p>We will also implement 2 dashboards: <\/p><ul><li>\"Blockchain Frontend\" for miners <\/li>\n<li>\"Blockchain Client\" for users to generate wallets and send coins <\/li>\n<\/ul><p>The blockchain implementation is mostly based on this <a href=\"https:\/\/github.com\/dvf\/blockchain\">github project<\/a>. I made a few modifications to the original code in order to add RSA encryption to the transactions. Wallet generation and transaction encryption is based on this <a href=\"https:\/\/github.com\/julienr\/ipynb_playground\/blob\/master\/bitcoin\/dumbcoin\/dumbcoin.ipynb\">Jupyter notebook<\/a>. The 2 dashboard are implemented from scratch using HTML\/CSS\/JS.<\/p><p>You can download the complete source code from <a href=\"https:\/\/github.com\/adilmoujahid\/blockchain-python-tutorial\">https:\/\/github.com\/adilmoujahid\/blockchain-python-tutorial<\/a>.<\/p><p>Please note that this implementation is for educational purposes only and shouldn't be use in production as it doesn't have good security, doesn't scale well and lacks many important features.<\/p><h2>3.1. Blockchain Client Implementation<\/h2><p>You can start the blockchain client from the terminal by going to the <code>blockchain_client<\/code> folder, and typing <code>python blockchain_client.py<\/code>. In your browser, go to <code>http:\/\/localhost:8080<\/code> and you'll see the dashboard below.<\/p><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-client.png\" alt=\"image\" \/><\/div><p>The dashboard has 3 tabs in the navigation bar:<\/p><ul><li>Wallet Generator: To generate wallets (Public\/Private keys pair) using RSA encryption algorithm<\/li>\n<li>Make Transaction: To generate transactions and send them to a blockchain node<\/li>\n<li>View Transasctions: To view the transactions that are on the blockchain<\/li>\n<\/ul><p>In order to make or view transactions, you will need at least one blockchain node running (to be covered in next section).<\/p><p>Below is some explanation of the most important parts in the <code>blockchain_client.py<\/code> code.<\/p><p>We define a python class that we name <code>Transaction<\/code> that has 4 attributes <code>sender_address<\/code>, <code>sender_private_key<\/code>, <code>recipient_address<\/code>, <code>value<\/code>. These are the 4 pieces of information that a sender needs to create a transaction. <\/p><p>The <code>to_dict()<\/code> method returns the transaction information in a Python dictionary format (without the sender's private key). The <code>sign_transaction()<\/code> method takes the transaction information (without the sender's private key) and signs it using the sender's private key.<\/p><div class=\"highlight\"><pre>class Transaction:\n    def __init__(self, sender_address, sender_private_key, recipient_address, value):\n        self.sender_address = sender_address\n        self.sender_private_key = sender_private_key\n        self.recipient_address = recipient_address\n        self.value = value\n    def __getattr__(self, attr):\n        return self.data[attr]\n    def to_dict(self):\n        return OrderedDict({'sender_address': self.sender_address,\n                            'recipient_address': self.recipient_address,\n                            'value': self.value})\n    def sign_transaction(self):\n        \"\"\"\n        Sign transaction with private key\n        \"\"\"\n        private_key = RSA.importKey(binascii.unhexlify(self.sender_private_key))\n        signer = PKCS1_v1_5.new(private_key)\n        h = SHA.new(str(self.to_dict()).encode('utf8'))\n        return binascii.hexlify(signer.sign(h)).decode('ascii')\n<\/pre><\/div><p>The line below initate a <code>Python Flask<\/code> app that we will use to create different APIs to interact with the blockchain and its client.<\/p><div class=\"highlight\"><pre>app = Flask(__name__)\n<\/pre><\/div><p>Below we define the 3 <code>Flask<\/code> routes that returns html pages. One html page for each tab.<\/p><div class=\"highlight\"><pre>@app.route('\/')\ndef index():\n  return render_template('.\/index.html')\n@app.route('\/make\/transaction')\ndef make_transaction():\n    return render_template('.\/make_transaction.html')\n@app.route('\/view\/transactions')\ndef view_transaction():\n    return render_template('.\/view_transactions.html')\n<\/pre><\/div><p>Below we define an API that generates wallets (Private\/Public keys pairs).<\/p><div class=\"highlight\"><pre>@app.route('\/wallet\/new', methods=['GET'])\ndef new_wallet():\n  random_gen = Crypto.Random.new().read\n  private_key = RSA.generate(1024, random_gen)\n  public_key = private_key.publickey()\n  response = {\n    'private_key': binascii.hexlify(private_key.exportKey(format='DER')).decode('ascii'),\n    'public_key': binascii.hexlify(public_key.exportKey(format='DER')).decode('ascii')\n  }\n  return jsonify(response), 200\n<\/pre><\/div><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-generate-wallet.gif\" alt=\"image\" \/><\/div><p>Below we define an API that takes as input <code>sender_address<\/code>, <code>sender_private_key<\/code>, <code>recipient_address<\/code>, <code>value<\/code>, and returns the transaction (without private key) and the signature.<\/p><div class=\"highlight\"><pre>@app.route('\/generate\/transaction', methods=['POST'])\ndef generate_transaction():\n  sender_address = request.form['sender_address']\n  sender_private_key = request.form['sender_private_key']\n  recipient_address = request.form['recipient_address']\n  value = request.form['amount']\n  transaction = Transaction(sender_address, sender_private_key, recipient_address, value)\n  response = {'transaction': transaction.to_dict(), 'signature': transaction.sign_transaction()}\n  return jsonify(response), 200\n<\/pre><\/div><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-generate-transaction.gif\" alt=\"image\" \/><\/div><h2>3.2. Blockchain Implementation<\/h2><p>You can start a blockchain node from the terminal by going to the <code>blockchain<\/code> folder, and type <code>python blockchain_client.py<\/code> or <code>python blockchain_client.py  -p &lt;PORT NUMBER&gt;<\/code>. If you don't specify a port number, it will default to port 5000. In your browser, go to <code>http:\/\/localhost:&lt;PORT NUMBER&gt;<\/code> to see the blockchain frontend dashboard.<\/p><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-frontend.png\" alt=\"image\" \/><\/div><p>The dashboard has 2 tabs in the navigation bar:<\/p><ul><li>Mine: For viewing transactions and blockchain data, and for mining new blocks of transactions.<\/li>\n<li>Configure: For configuring connections between the different blockchain nodes.<\/li>\n<\/ul><p>Below is some explanation of the most important parts in the <code>blockchain.py<\/code> code.<\/p><p>We start by defining a <code>Blockchain<\/code> class that has the following attributes:<\/p><ul><li><code>transactions<\/code>: List of transactions that will be added to the next block.<\/li>\n<li><code>chain<\/code>: The actual blockchain which is an array of blocks.<\/li>\n<li><code>nodes<\/code>: A set containing node urls. The blockchain uses these nodes to retrieve blockchain data from other nodes and updates its blockchain if they're not in sync.<\/li>\n<li><code>node_id<\/code>: A random string to identify the blockchain node.<\/li>\n<\/ul><p>The <code>Blockchain<\/code> class also implements the following methods:<\/p><ul><li><code>register_node(node_url)<\/code>: Adds a new blockchain node to the list of nodes.<\/li>\n<li><code>verify_transaction_signature(sender_address, signature, transaction)<\/code>: Checks that the provided signature corresponds to transaction signed by the public key (sender_address).<\/li>\n<li><code>submit_transaction(sender_address, recipient_address, value, signature)<\/code>: Adds a transaction to list of transactions if the signature verified.<\/li>\n<li><code>create_block(nonce, previous_hash)<\/code>: Adds a block of transactions to the blockchain.<\/li>\n<li><code>hash(block)<\/code>: Create a SHA-256 hash of a block.<\/li>\n<li><code>proof_of_work()<\/code>: Proof of work algorithm. Looks for a nonce that satisfies the mining condition.<\/li>\n<li><code>valid_proof(transactions, last_hash, nonce, difficulty=MINING_DIFFICULTY)<\/code>: Checks if a hash value satisfies the mining conditions. This function is used within the proof_of_work function.<\/li>\n<li><code>valid_chain(chain)<\/code>: checks if a bockchain is valid.<\/li>\n<li><code>resolve_conflicts()<\/code>: Resolves conflicts between blockchain's nodes by replacing a chain with the longest one in the network.<\/li>\n<\/ul><div class=\"highlight\"><pre>class Blockchain:\n    def __init__(self):\n        self.transactions = []\n        self.chain = []\n        self.nodes = set()\n        #Generate random number to be used as node_id\n        self.node_id = str(uuid4()).replace('-', '')\n        #Create genesis block\n        self.create_block(0, '00')\n    def register_node(self, node_url):\n        \"\"\"\n        Add a new node to the list of nodes\n        \"\"\"\n        ...\n    def verify_transaction_signature(self, sender_address, signature, transaction):\n        \"\"\"\n        Check that the provided signature corresponds to transaction\n        signed by the public key (sender_address)\n        \"\"\"\n        ...\n    def submit_transaction(self, sender_address, recipient_address, value, signature):\n        \"\"\"\n        Add a transaction to transactions array if the signature verified\n        \"\"\"\n        ...\n    def create_block(self, nonce, previous_hash):\n        \"\"\"\n        Add a block of transactions to the blockchain\n        \"\"\"\n        ...\n    def hash(self, block):\n        \"\"\"\n        Create a SHA-256 hash of a block\n        \"\"\"\n        ...\n    def proof_of_work(self):\n        \"\"\"\n        Proof of work algorithm\n        \"\"\"\n        ...\n    def valid_proof(self, transactions, last_hash, nonce, difficulty=MINING_DIFFICULTY):\n        \"\"\"\n        Check if a hash value satisfies the mining conditions. This function is used within the proof_of_work function.\n        \"\"\"\n        ...\n    def valid_chain(self, chain):\n        \"\"\"\n        check if a bockchain is valid\n        \"\"\"\n        ...\n    def resolve_conflicts(self):\n        \"\"\"\n        Resolve conflicts between blockchain's nodes\n        by replacing our chain with the longest one in the network.\n        \"\"\"\n        ...\n<\/pre><\/div><p>The line below initate a <code>Python Flask<\/code> app that we will use to create different APIs to interact with the blockchain.<\/p><div class=\"highlight\"><pre>app = Flask(__name__)\nCORS(app)\n<\/pre><\/div><p>Next, we initiate a Blockchain instance.<\/p><div class=\"highlight\"><pre>blockchain = Blockchain()\n<\/pre><\/div><p>Below we define the 2 <code>Flask<\/code> routes that return the html pages for our blockchain frontend dashboard.<\/p><div class=\"highlight\"><pre>@app.route('\/')\ndef index():\n    return render_template('.\/index.html')\n@app.route('\/configure')\ndef configure():\n    return render_template('.\/configure.html')\n<\/pre><\/div><p>Below we define <code>Flask<\/code> APIs to manage transactions and mining the blockchain.<\/p><ul><li><code>'\/transactions\/new'<\/code>: This API takes as input <code>'sender_address'<\/code>, <code>'recipient_address'<\/code>, <code>'amount'<\/code> and <code>'signature'<\/code>, and adds the transaction to the list of transactions that will be added to next block if the signature is valid.<\/li>\n<li><code>'\/transactions\/get'<\/code>: This API returns all the transactions that will be added to the next block.<\/li>\n<li><code>'\/chain'<\/code>: This API returns all blockchain data.<\/li>\n<li><code>'\/mine'<\/code>: This API runs the proof of work algorithm, and adds the new block of transactions to the blockchain.<\/li>\n<\/ul><div class=\"highlight\"><pre>@app.route('\/transactions\/new', methods=['POST'])\ndef new_transaction():\n    values = request.form\n    # Check that the required fields are in the POST'ed data\n    required = ['sender_address', 'recipient_address', 'amount', 'signature']\n    if not all(k in values for k in required):\n        return 'Missing values', 400\n    # Create a new Transaction\n    transaction_result = blockchain.submit_transaction(values['sender_address'], values['recipient_address'], values['amount'], values['signature'])\n    if transaction_result == False:\n        response = {'message': 'Invalid Transaction!'}\n        return jsonify(response), 406\n    else:\n        response = {'message': 'Transaction will be added to Block '+ str(transaction_result)}\n        return jsonify(response), 201\n@app.route('\/transactions\/get', methods=['GET'])\ndef get_transactions():\n    #Get transactions from transactions pool\n    transactions = blockchain.transactions\n    response = {'transactions': transactions}\n    return jsonify(response), 200\n@app.route('\/chain', methods=['GET'])\ndef full_chain():\n    response = {\n        'chain': blockchain.chain,\n        'length': len(blockchain.chain),\n    }\n    return jsonify(response), 200\n@app.route('\/mine', methods=['GET'])\ndef mine():\n    # We run the proof of work algorithm to get the next proof...\n    last_block = blockchain.chain[-1]\n    nonce = blockchain.proof_of_work()\n    # We must receive a reward for finding the proof.\n    blockchain.submit_transaction(sender_address=MINING_SENDER, recipient_address=blockchain.node_id, value=MINING_REWARD, signature=\"\")\n    # Forge the new Block by adding it to the chain\n    previous_hash = blockchain.hash(last_block)\n    block = blockchain.create_block(nonce, previous_hash)\n    response = {\n        'message': \"New Block Forged\",\n        'block_number': block['block_number'],\n        'transactions': block['transactions'],\n        'nonce': block['nonce'],\n        'previous_hash': block['previous_hash'],\n    }\n    return jsonify(response), 200\n<\/pre><\/div><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-mine.gif\" alt=\"image\" \/><\/div><p>Below we define Flask APIs to manage blockchain nodes.<\/p><ul><li><code>'\/nodes\/register'<\/code>: This API takes as input a list of node urls, and adds them to the list of nodes.<\/li>\n<li><code>'\/nodes\/resolve'<\/code>: This API resolves conflicts between blockchain nodes by replacing a local chain with the longest one available in the network.<\/li>\n<li><code>'\/nodes\/get'<\/code>: This API returns the list of nodes.<\/li>\n<\/ul><div class=\"highlight\"><pre>@app.route('\/nodes\/register', methods=['POST'])\ndef register_nodes():\n    values = request.form\n    nodes = values.get('nodes').replace(\" \", \"\").split(',')\n    if nodes is None:\n        return \"Error: Please supply a valid list of nodes\", 400\n    for node in nodes:\n        blockchain.register_node(node)\n    response = {\n        'message': 'New nodes have been added',\n        'total_nodes': [node for node in blockchain.nodes],\n    }\n    return jsonify(response), 201\n@app.route('\/nodes\/resolve', methods=['GET'])\ndef consensus():\n    replaced = blockchain.resolve_conflicts()\n    if replaced:\n        response = {\n            'message': 'Our chain was replaced',\n            'new_chain': blockchain.chain\n        }\n    else:\n        response = {\n            'message': 'Our chain is authoritative',\n            'chain': blockchain.chain\n        }\n    return jsonify(response), 200\n@app.route('\/nodes\/get', methods=['GET'])\ndef get_nodes():\n    nodes = list(blockchain.nodes)\n    response = {'nodes': nodes}\n    return jsonify(response), 200\n<\/pre><\/div><div><img src=\"http:\/\/adilmoujahid.com\/images\/blockchain-manage-nodes.gif\" alt=\"image\" \/><\/div><p>In this blog post, we covered some core concepts behind blockchain and we learned how to implement one using Python. For the sake of simplicity, I didn't cover some technical details, for example: Wallet addresses and Merkel trees. If you want to learn more about the subject, I recommend reading the original Bitcoin white paper and follow up with <a href=\"https:\/\/en.bitcoin.it\/wiki\/Main_Page\">bitcoin wiki<\/a> and Andreas Antonopoulos's excellent book: <a href=\"https:\/\/www.amazon.com\/gp\/product\/1491954388\/ref=as_li_tl?ie=UTF8&amp;camp=1789&amp;creative=9325&amp;creativeASIN=1491954388&amp;linkCode=as2&amp;tag=adilmoujahid-20&amp;linkId=bd776f9224715e8a022d4984909d6a69\">Mastering Bitcoin: Programming the Open Blockchain<\/a>.<\/p><ul><li>1 - <a href=\"https:\/\/en.wikipedia.org\/wiki\/Public-key_cryptography\">Wikipedia - Public-key cryptography<\/a><\/li>\n<li>2 - <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hash_function\">Wikipedia - Hash function<\/a><\/li>\n<li>3 - <a href=\"https:\/\/bitcoin.stackexchange.com\/questions\/58687\/what-happens-to-a-transaction-once-generated\">Bitcoin Stackexchange - What happens to a transaction once generated?<\/a> <\/li>\n<li>4 - <a href=\"https:\/\/en.bitcoin.it\/wiki\/Majority_attack\">Bitcoin Wiki - Majority attack<\/a><\/li>\n<\/ul><a href=\"#\" class=\"go-top\">Go Top<\/a><\/section>","created_at":"2018-11-07T16:42:36+0000","updated_at":"2018-11-07T16:42:36+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":22,"domain_name":"adilmoujahid.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":260,"title":"How To Install Wordpress on Ubuntu 14.04","url":"https:\/\/www.digitalocean.com\/community\/tutorials\/how-to-install-wordpress-on-ubuntu-14-04","content":"<p>At this time, WordPress is the most popular CMS (content management system) on the internet.  It allows you to easily set up flexible blogs and websites on top of a MySQL backend with PHP processing.  WordPress has seen incredible adoption and is a great choice for getting a website up and running quickly.<\/p><p>In this guide, we'll focus on getting a WordPress instance set up with an Apache web server on Ubuntu 14.04.<\/p><h2 id=\"prerequisites\">Prerequisites<\/h2><p>Before you begin this guide, there are some important steps that you need to complete on your server.<\/p><p>We will be proceeding through these steps as a non-root user with sudo privileges, so you will need to have one available.  You can find out how to create a user with sudo privileges by following steps 1-4 in our <a href=\"https:\/\/www.digitalocean.com\/community\/articles\/initial-server-setup-with-ubuntu-14-04\">Ubuntu 14.04 initial server setup<\/a> guide.<\/p><p>Additionally, you'll need to have a LAMP (Linux, Apache, MySQL, and PHP) stack installed on your VPS instance.  If you don't have these components already installed and configured, you can use this guide to learn <a href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-install-linux-apache-mysql-php-lamp-stack-on-ubuntu-14-04\">how to install LAMP on Ubuntu 14.04<\/a>.<\/p><p>When you are finished with these steps, you can continue with this guide.<\/p><h2 id=\"step-one-&#x2014;-create-a-mysql-database-and-user-for-wordpress\">Step One \u2014 Create a MySQL Database and User for WordPress<\/h2><p>The first step that we will take is a preparatory one.  WordPress uses a relational database to manage and store site and user information.<\/p><p>We have MySQL installed, which can provide this functionality, but we need to make a database and a user for WordPress to work with.<\/p><p>To get started, log into the MySQL root (administrative) account by issuing this command:<\/p><pre>mysql -u root -p\n<\/pre><p>You will be prompted for the password you set for the MySQL root account when you installed the software.  You will then be given a MySQL command prompt.<\/p><p>First, we can create a separate database that WordPress can control.  You can call this whatever you would like, but I will be calling it <code>wordpress<\/code> because it is descriptive and simple.  Enter this command to create the database:<\/p><pre>\nCREATE DATABASE wordpress;\n<\/pre><p>Every MySQL statement must end in a semi-colon (;), so check to make sure this is present if you are running into any issues.<\/p><p>Next, we are going to create a separate MySQL user account that we will use exclusively to operate on our new database.  Creating one-function databases and accounts is a good idea from a management and security standpoint.<\/p><p>I am going to call the new account that I'm making <code>wordpressuser<\/code> and will assign it a password of <code>password<\/code>.  You should definitely change the password for your installation and can name the user whatever you'd like.  This is the command you need to create the user:<\/p><pre>\nCREATE USER wordpressuser@localhost IDENTIFIED BY 'password';\n<\/pre><p>At this point, you have a database and a user account, each made specifically for WordPress.  However, these two components have no relationship yet.  The user has no access to the database.<\/p><p>Let's fix that by granting our user account access to our database with this command:<\/p><pre>\nGRANT ALL PRIVILEGES ON wordpress.* TO wordpressuser@localhost;\n<\/pre><p>Now the user has access to the database.  We need to flush the privileges so that the current instance of MySQL knows about the recent privilege changes we've made:<\/p><pre>FLUSH PRIVILEGES;\n<\/pre><p>We're all set now.  We can exit out of the MySQL prompt by typing:<\/p><pre>exit\n<\/pre><p>You should now be back to your regular command prompt.<\/p><p>Next, we will download the actual WordPress files from the project's website.<\/p><p>Luckily, the WordPress team always links the most recent stable version of their software to the same URL, so we can get the most up-to-date version of WordPress by typing this:<\/p><pre>cd ~\nwget http:\/\/wordpress.org\/latest.tar.gz\n<\/pre><p>This will download a compressed file that contains the archived directory contents of the WordPress files to our home directory.<\/p><p>We can extract the files to rebuild the WordPress directory we need by typing:<\/p><pre>tar xzvf latest.tar.gz\n<\/pre><p>This will create a directory called <code>wordpress<\/code> in your home directory.<\/p><p>While we are downloading things, we should also get a few more packages that we need.  We can get these directly from Ubuntu's default repositories after we update our local package index:<\/p><pre>sudo apt-get update\nsudo apt-get install php5-gd libssh2-php\n<\/pre><p>This will allow you to work with images and will also allow you to install plugins and update portions of your site using your SSH login credentials.<\/p><h2 id=\"step-three-&#x2014;-configure-wordpress\">Step Three \u2014 Configure WordPress<\/h2><p>Most of the configuration that we will be doing will be through a web interface later on.  However, we do need to do some work from the command line before we can get this up and running.<\/p><p>Begin by moving into the WordPress directory that you just unpacked:<\/p><pre>cd ~\/wordpress\n<\/pre><p>A sample configuration file that mostly matches the configuration we need is included by default.  However, we need to copy it to the default configuration file location to get WordPress to recognize the file.  Do that now by typing:<\/p><pre>cp wp-config-sample.php wp-config.php\n<\/pre><p>Now that we have a configuration file to work with, we can generate some secret keys that help to secure the installation.  WordPress provides a secure generator for these values so that you do not have to try to come up with good values on your own. These are only used internally, so it won't hurt usability to have complex, secure values here.<\/p><p>To grab secure values from the WordPress secret key generator, type:<\/p><pre>curl -s https:\/\/api.wordpress.org\/secret-key\/1.1\/salt\/\n<\/pre><p>You will get back unique values that look something like this:<\/p><p><strong>Warning!<\/strong>  It is important that you request unique values each time.  Do <strong>NOT<\/strong> copy the values shown below!<br \/><\/p><pre>Outputdefine('AUTH_KEY',         '1jl\/vqfs&lt;XhdXoAPz9 DO NOT COPY THESE VALUES c_j{iwqD^&lt;+c9.k&lt;J@4H');\ndefine('SECURE_AUTH_KEY',  'E2N-h2]Dcvp+aS\/p7X DO NOT COPY THESE VALUES {Ka(f;rv?Pxf})CgLi-3');\ndefine('LOGGED_IN_KEY',    'W(50,{W^,OPB%PB&lt;JF DO NOT COPY THESE VALUES 2;y&amp;,2m%3]R6DUth[;88');\ndefine('NONCE_KEY',        'll,4UC)7ua+8&lt;!4VM+ DO NOT COPY THESE VALUES #`DXF+[$atzM7 o^-C7g');\ndefine('AUTH_SALT',        'koMrurzOA+|L_lG}kf DO NOT COPY THESE VALUES  07VC*Lj*lD&amp;?3w!BT#-');\ndefine('SECURE_AUTH_SALT', 'p32*p,]z%LZ+pAu:VY DO NOT COPY THESE VALUES C-?y+K0DK_+F|0h{!_xY');\ndefine('LOGGED_IN_SALT',   'i^\/G2W7!-1H2OQ+t$3 DO NOT COPY THESE VALUES t6**bRVFSD[Hi])-qS`|');\ndefine('NONCE_SALT',       'Q6]U:K?j4L%Z]}h^q7 DO NOT COPY THESE VALUES 1% ^qUswWgn+6&amp;xqHN&amp;%');\n<\/pre><p>These are configuration lines that we can paste directly in our configuration file to set secure keys.  Copy the output you received now.<\/p><p>Next, let's open the configuration file in a text editor:<\/p><pre>nano wp-config.php\n<\/pre><p>Find the section that contains the dummy values for those settings.  It will look something like this:<\/p><pre>. . .\n#define('AUTH_KEY',         'put your unique phrase here');\n#define('SECURE_AUTH_KEY',  'put your unique phrase here');\n#define('LOGGED_IN_KEY',    'put your unique phrase here');\n#define('NONCE_KEY',        'put your unique phrase here');\n#define('AUTH_SALT',        'put your unique phrase here');\n#define('SECURE_AUTH_SALT', 'put your unique phrase here');\n#define('LOGGED_IN_SALT',   'put your unique phrase here');\n#define('NONCE_SALT',       'put your unique phrase here');\n. . .\n<\/pre><p>Delete those lines and paste in the values you copied from the command line:<\/p><p>\/var\/www\/html\/wp-config.php<\/p><pre>. . .\ndefine('AUTH_KEY',         'VALUES COPIED FROM THE COMMAND LINE');\ndefine('SECURE_AUTH_KEY',  'VALUES COPIED FROM THE COMMAND LINE');\ndefine('LOGGED_IN_KEY',    'VALUES COPIED FROM THE COMMAND LINE');\ndefine('NONCE_KEY',        'VALUES COPIED FROM THE COMMAND LINE');\ndefine('AUTH_SALT',        'VALUES COPIED FROM THE COMMAND LINE');\ndefine('SECURE_AUTH_SALT', 'VALUES COPIED FROM THE COMMAND LINE');\ndefine('LOGGED_IN_SALT',   'VALUES COPIED FROM THE COMMAND LINE');\ndefine('NONCE_SALT',       'VALUES COPIED FROM THE COMMAND LINE');\n. . .\n<\/pre><p>After that, the only modifications we need to make are to the parameters that hold our database information.<\/p><p>We will need to find the settings for <code>DB_NAME<\/code>, <code>DB_USER<\/code>, and <code>DB_PASSWORD<\/code> in order for WordPress to correctly connect and authenticate to the database we created.<\/p><p>Fill in the values of these parameters with the information for the database you created.  It should look like this:<\/p><pre>\n\/\/ ** MySQL settings - You can get this info from your web host ** \/\/\n\/** The name of the database for WordPress *\/\ndefine('DB_NAME', 'wordpress');\n\/** MySQL database username *\/\ndefine('DB_USER', 'wordpressuser');\n\/** MySQL database password *\/\ndefine('DB_PASSWORD', 'password');\n<\/pre><p>These are the only values that you need to change.<\/p><p>When you are finished, save and close the file.<\/p><h2 id=\"step-four-&#x2014;-copy-files-to-the-document-root\">Step Four \u2014 Copy Files to the Document Root<\/h2><p>Now that we have our application configured, we need to copy it into Apache's document root, where it can be served to visitors of our website.<\/p><p>One of the easiest and most reliable way of transferring files from directory to directory is with the <code>rsync<\/code> command.  This preserves permissions and has good data integrity features.<\/p><p>The location of the document root in <a href=\"https:\/\/www.digitalocean.com\/community\/articles\/how-to-install-linux-apache-mysql-php-lamp-stack-on-ubuntu-14-04\">the Ubuntu 14.04 LAMP guide<\/a> is <code>\/var\/www\/html\/<\/code>.  We can transfer our WordPress files there by typing:<\/p><pre>sudo rsync -avP ~\/wordpress\/ \/var\/www\/html\/\n<\/pre><p>This will safely copy all of the contents from the directory you unpacked to the document root.<\/p><p>We should now move into the document root to make some final permissions changes<\/p><pre>cd \/var\/www\/html\n<\/pre><p>You will need to change the ownership of our files for increased security.<\/p><p>We want to give user ownership to the regular, non-root user (with sudo privileges) that you plan on using to interact with your site.  This can be your regular user if you wish, but some may suggest that you create an additional user for this process.  It is up to you which you choose.<\/p><p>For this guide, we will use the same account that we set up during the <a href=\"https:\/\/www.digitalocean.com\/community\/articles\/initial-server-setup-with-ubuntu-14-04\">initial server setup<\/a> guide, which we called <code>demo<\/code>.  This is the account I am performing all of the actions of this guide as.<\/p><p>The group ownership we will give to our web server process, which is <code>www-data<\/code>.  This will allow Apache to interact with the content as necessary.<\/p><p>We can quickly assign these ownership values by typing:<\/p><pre>\nsudo chown -R demo:www-data *\n<\/pre><p>This will set up the ownership properties that we are looking for.<\/p><p>While we are dealing with ownership and permissions, we should also look into assigning correct ownership on our uploads directory.  This will allow us to upload images and other content to our site.  Currently, the permissions are too restrictive.<\/p><p>First, let's manually create the <code>uploads<\/code> directory beneath the <code>wp-content<\/code> directory at our document root.  This will be the parent directory of our content:<\/p><pre>mkdir \/var\/www\/html\/wp-content\/uploads\n<\/pre><p>We have a directory now to house uploaded files, however the permissions are still too restrictive.  We need to allow the web server itself to write to this directory.  We can do this by assigning group ownership of this directory to our web server, like this:<\/p><pre>sudo chown -R :www-data \/var\/www\/html\/wp-content\/uploads\n<\/pre><p>This will allow the web server to create files and directories under this directory, which will permit us to upload content to the server.<\/p><h2 id=\"step-five-&#x2014;-complete-installation-through-the-web-interface\">Step Five \u2014 Complete Installation through the Web Interface<\/h2><p>Now that you have your files in place and your software is configured, you can complete the installation through the web interface.<\/p><p>In your web browser, navigate to your server's domain name or public IP address:<\/p><pre>\nhttp:\/\/server_domain_name_or_IP\n<\/pre><p>You will see the WordPress initial configuration page, where you will create an initial administrator account:<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/initial_config.png\" alt=\"Wordpress initial config\" \/><\/p><p>Fill out the information for the site and the administrative account you wish to make.  When you are finished, click on the install button at the bottom.<\/p><p>WordPress will confirm the installation, and then ask you to log in with the account you just created:<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/confirm_install.png\" alt=\"WordPress confirm install\" \/><\/p><p>Hit the button at the bottom and then fill out your account information:<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/login.png\" alt=\"WordPress login\" \/><\/p><p>You will be presented with the WordPress interface:<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/admin_interface.png\" alt=\"WordPress admin interface\" \/><\/p><p>By default, WordPress creates URLs dynamically that look something like this:<\/p><pre>\nhttp:\/\/server_domain_name_or_IP\/?p=1\n<\/pre><p>This isn't exactly the most useful interface for visitors or search engines, so most users want to modify this.  WordPress has the ability to create \"pretty\" permalinks which will clean up the URL into a more human-friendly format.<\/p><p>There are a few things we need to do to get this to work with Apache on Ubuntu 14.04.<\/p><h3 id=\"modifying-apache-to-allow-url-rewrites\">Modifying Apache to Allow URL Rewrites<\/h3><p>First, we need to modify the Apache virtual host file for WordPress to allow for <code>.htaccess<\/code> overrides.  You can do this by editing the virtual host file.<\/p><p>By default, this is <code>000-default.conf<\/code>, but your file might be different if you created another configuration file:<\/p><pre>sudo nano \/etc\/apache2\/sites-available\/000-default.conf\n<\/pre><p>Inside of this file, we want to set up a few things.  We should set the <code>ServerName<\/code> and create a directory section where we allow overrides.  This should look something like this:<\/p><pre>\n&lt;VirtualHost *:80&gt;\n    ServerAdmin webmaster@localhost\n    DocumentRoot \/var\/www\/html\n    ServerName server_domain_name_or_IP\n    &lt;Directory \/var\/www\/html\/&gt;\n        AllowOverride All\n    &lt;\/Directory&gt;\n    . . .\n<\/pre><p>When you are finished, save and close the file.<\/p><p>Next, we need to enable the rewrite module, which allows you to modify URLs.  You can do this by typing:<\/p><pre>sudo a2enmod rewrite\n<\/pre><p>After you have made these changes, restart Apache:<\/p><pre>sudo service apache2 restart\n<\/pre><h3 id=\"create-an-htaccess-file\">Create an .htaccess File<\/h3><p>Now that Apache is configured to allow rewrites through <code>.htaccess<\/code> files, we need to create an actual file.<\/p><p>You need to place this file in your document root.  Type this to create an empty file:<\/p><pre>touch \/var\/www\/html\/.htaccess\n<\/pre><p>This will be created with your username and user group.  We need the web server to be the group owner though, so we should adjust the ownership by typing:<\/p><pre>sudo chown :www-data \/var\/www\/html\/.htaccess\n<\/pre><p>We now have the correct ownership of this file.<\/p><p>We may need to adjust the permissions however.  This depends on how you prefer to work.  WordPress will generate the necessary rewrite rules for you.  If it has write permissions to this file, it can implement the rules automatically.  If it does not, you will have to manually edit this file to add the correct rules.<\/p><p>Which configuration you choose depends on how much you value convenience over security.  Allowing the web server write access to this file will definitely be more convenient, but some say that it is an unnecessary security risk.<\/p><p>If you want WordPress to automatically update this file with rewrite rules, you can ensure that it has the correct permissions to do so by typing:<\/p><pre>chmod 664 \/var\/www\/html\/.htaccess\n<\/pre><p>If you want to update this file manually for the sake of a small security gain, you can allow the web server only read privileges by typing:<\/p><pre>chmod 644 \/var\/www\/html\/.htaccess\n<\/pre><h3 id=\"change-the-permalink-settings-in-wordpress\">Change the Permalink Settings in WordPress<\/h3><p>When you are finished doing the server-side changes, you can easily adjust the permalink settings through the WordPress administration interface.<\/p><p>On the left-hand side, under the <code>Settings<\/code> menu, you can select <code>Permalinks<\/code>:<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/perma_settings.png\" alt=\"WordPress permalinks\" \/><\/p><p>You can choose any of the preconfigured settings to organize URLs, or you can create your own.<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/perma_options.png\" alt=\"WordPress perma options\" \/><\/p><p>When you have made your selection, click \"Save Changes\" to generate the rewrite rules.<\/p><p>If you allowed the web server write access to your <code>.htaccess<\/code> file, you should see a message like this:<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/wordpress_1404\/perma_update.png\" alt=\"WordPress perma update\" \/><\/p><p>If you did <em>not<\/em> allow the web server write access to your <code>.htaccess<\/code> file, you will be provided with the rewrite rules you need to add to the file manually.<\/p><p>Copy the lines that WordPress gives you and then edit file on your server:<\/p><pre>nano \/var\/www\/html\/.htaccess\n<\/pre><p>This should give you the same functionality.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>You should now have a WordPress instance up and running on your Ubuntu 14.04 VPS.  There are many avenues you can take from here.  Below we've listed some options:<\/p>","created_at":"2018-11-07T16:42:36+0000","updated_at":"2018-11-07T16:42:36+0000","published_at":"2014-04-24T18:49:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"www.digitalocean.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":256,"title":"7 Steps To Writing Perfect Blog Posts [Checklist]","url":"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/","content":"<input class=\"jpibfi\" type=\"hidden\" \/><p>Writing successful blog posts is not just a matter of creating content. To make sure, your blog posts get real attention, this 7-step-guide and checklist will help you to turn a good blog post into gold.<\/p><p>Content is still king. Considering that millions of blog posts are published every day, your target readers get millions of options to choose from. So, you have to make your content stands out among the masses. In the age of the social web and content marketing, commercial content produced by corporate blogs and bloggers is competing with traditional journalistic content. There is hardly a topic not yet covered. Mark Schaefer prepares us for <a href=\"https:\/\/www.businessesgrow.com\/2014\/01\/06\/content-shock\/\">The Content Marketing Shock<\/a>:<\/p><div>\u201c<i>The emerging marketing epoch defined when exponentially increasing volumes of content intersect our limited human capacity to consume it<\/i>?\u201d<\/div><p>What does this mean for bloggers? Is there still a chance to get to the top of an ever growing blogosphere? Yes, there is. But it also means you will have a lot of more things to consider before and after hitting the publish button.\u00a0\u00a0This article will help you with a <b>7-step-guide &amp; checklist<\/b>\u00a0to make sure you create valuable blog posts for your target audience, to optimize your posts for better readability\u00a0and search engine indexing and to maximize your promotional efforts for more outreach and performance.<\/p><div><div class=\"row\"><img class=\"size-medium wp-image-54668 aligncenter\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_header-icegram-600x194.png\" alt=\"7 Steps to Writing Perfect Blog Posts [Checklist]\" width=\"600\" height=\"194\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/\" data-jpibfi-post-title=\"7 Steps To Writing Perfect Blog Posts [Checklist]\" data-jpibfi-src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_header-icegram-600x194.png\" \/><br \/><strong>Download the Checklist 7 Steps to Writing Perfect Blog Posts<\/strong><div id=\"rainmaker_form_54675\" class=\"rm_form_container rainmaker_form\" data-type=\"rm_custom\" data-form-id=\"54675\"><form class=\"DSGVO\" action=\"https:\/\/e.asend.de\/index.php?entryPoint=WebToLeadCapture\" method=\"post\" id=\"pleaseValidate\"> <div><label for=\"salutation _frau\"><input type=\"radio\" name=\"salutation\" required=\"required\" value=\"Ms.\" \/> Ms.<\/label> <label for=\"salutation _herr\"><input type=\"radio\" name=\"salutation\" required=\"required\" value=\"Mr.\" \/> Mr.<\/label> <input type=\"text\" name=\"first_name\" required=\"required\" id=\"first_name\" placeholder=\"Firstname\" \/><input type=\"text\" name=\"last_name\" required=\"required\" id=\"last_name\" placeholder=\"Lastname\" \/><input type=\"email\" name=\"email1\" required=\"required\" id=\"email1\" placeholder=\"email\" \/><input id=\"sprache_c\" name=\"sprache_c\" type=\"hidden\" value=\"englisch\" \/><input id=\"assigned_user_id\" name=\"assigned_user_id\" type=\"hidden\" value=\"\" \/><input id=\"geeignet_fuer_produkt_c\" name=\"geeignet_fuer_produkt_c\" type=\"hidden\" value=\"prgateway,blog2social\" \/><input id=\"lead_source\" name=\"lead_source\" type=\"hidden\" value=\"Registration\" \/><input id=\"redirect_url\" name=\"redirect_url\" type=\"hidden\" value=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/?v=confirm\" \/><input id=\"campaign_id\" name=\"campaign_id\" type=\"hidden\" value=\"34325c86-f5d3-fe3b-8ee0-59ae82d8d005\" \/><input id=\"dokument_c\" name=\"dokument_c\" type=\"hidden\" value=\"Checklist 7 Steps to Writing Perfect Blog Posts\" \/><input id=\"dokumentenprojekt_c\" name=\"dokumentenprojekt_c\" type=\"hidden\" value=\"Checklist\" \/><input id=\"dokumentenurl_c\" name=\"dokumentenurl_c\" type=\"hidden\" value=\"https:\/\/www.blog2social.com\/docs\/checklist_the-perfect-blog-post.pdf\" \/><input id=\"dokumententyp_c\" name=\"dokumententyp_c\" type=\"hidden\" value=\"CaseStudy\" \/><div class=\"remark\"><p>We will never share or sell your email and personal data and we will always keep your data safe and secure.<\/p><p><label title=\"You may at any time revoke your consent\"><input type=\"checkbox\" class=\"DSGVO\" name=\"dsgvo_c\" value=\"eingewilligt\" \/> I agree to the <a href=\"https:\/\/www.adenion.de\/privacy-policy\" target=\"_blank\">privacy policy<\/a> and <a href=\"https:\/\/www.blog2social.com\/en\/terms\/\" target=\"_blank\">terms of service<\/a>. (You can unsubscribe any time.)<\/label><\/p><\/div> <input class=\"DSGVOsubmit\" disabled=\"disabled\" type=\"submit\" value=\"download for free\" \/><\/div><\/form><div class=\"rm-loader\"><\/div><\/div><\/div><div><h2><b>7 Steps for Writing Perfect Blog Posts<\/b><\/h2><h3><strong>Step 1: Make sure your blog post content is awesome<\/strong><\/h3><div>In the age of the semantic web (web 3.0), <b>relevancy<\/b> is more important than ever to hit your target audience. Take your time before writing on any topic of general interest. It is somehow very tempting to write a post on a topic of general interest. But you will be more likely to succeed if you focus a specific niche topic for a specific target group.\u00a0If you try to write for everyone, you will end with no one. The better you know your audience, the better your post will catch the attention of your target readers.<\/div><ul><li>Focusing on a very <b>specific topic or niche<\/b> will help you adapt your approach as well as your wording to be more relevant for the specific audience.<\/li><li>Make sure that you research thoroughly about the subject you are writing so that your work may be <b>relevant and up-to-date<\/b>.<\/li><li>Research what others have written on your topic and comment or pick up a specific aspect that has not been covered. By doing this, you will find many more ideas for your future posts.<\/li><li>Create awesome content, that will make your readers remember you as a point of interest and a <b>useful resource<\/b> for valuable information on your topic.<\/li><li><b>Provide tips, strategies, how to articles or entertaining content for your specific audience and topic<\/b>. Adding value for the readers will make your posts memorable.<\/li><\/ul><p>For more guidance on managing content marketing without overwhelming your audience, also\u00a0read \u201c<a href=\"http:\/\/www.amazon.com\/lm\/R2I4YZM047UOPI\/ref=cm_sw_em_r_lm_Ury3rb122SB4G?%20\"><i>Epic Content Marketing<\/i><\/a>,\u201d by Joe Pulizzi.<\/p><h3><strong>Step 2: Write a catchy headline and introduction for your blog post<\/strong><\/h3><div>Your headline and introduction are most important to grab the attention of your target readers, so make sure you spend enough time to create a catchy title and an introduction that will make your readers want to read the rest of your post. Summarize what your post is about. Outline the <b>topic<\/b> and your\u00a0<b>key arguments<\/b>, but keep the suspense.<\/div><div>Also, make sure your headline and introduction are optimized for SEO and\u00a0meet the requirements for the \u2018title\u2019 and \u2018meta description\u2019 displayed in the search engines:<\/div><div><ul><li>Write a <strong>title with less than 57 characters<\/strong> and include your <strong>keywords<\/strong>. This makes sure your title is properly displayed and gets not truncated in the search engines.<\/li><\/ul><\/div><ul><li>Make sure the <strong>first 160 characters of your introduction cover the essence<\/strong> of what your post is about and also <strong>include your focus keywords<\/strong> so that it can double act as meta description and snippet for your post.<\/li><\/ul><div><b>Tip<\/b>: Try the <a href=\"https:\/\/sumo.com\/kickass-headline-generator\/\">Sumo Kickass Headline Generator<\/a> to create engaging headlines and titles for your blog post<\/div><h3><strong>Step 3: Maximize readability for readers and search engines \u00a0<\/strong><\/h3><div>A well written and properly formatted blog post will\u00a0maximize readability for your readers as well as for search engines to index your blog post for matching searches. There is no perfect length for a blog post, however, make your blog post<b> as short and precise as possible<\/b> and <b>as long and detailed as necessary <\/b>to get to the bottom of your topic. Keep your <strong>sentences short and easy<\/strong> to understand. Keep your <strong>language simple<\/strong> and stay within the language usage of your target audience. Listings, <em>italics<\/em> and <strong>bold<\/strong> help to outline and highlight important aspects and keywords.<\/div><h4>To maximize readablility for your perfect blog post<\/h4><ul><li><b>Be as short as possible<\/b> and <b>as long as necessary<\/b>.<\/li><li>Segment your blog post into <b>paragraphs<\/b> with <b>strong sub headings.<\/b><\/li><li>Organize your blog post in <b>paragraphs with less than 150 words<\/b> and<\/li><li>Structure your post in<b> sections with less than 300 words<\/b>.<\/li><li>Use subheadings for\u00a0<b>each section focussing on a single aspect<\/b> of your topic.<\/li><li>Include at least 1 <b>internal\u00a0link<\/b>\u00a0to refer to related content in your past posts.<\/li><li>Include at least 1 <b>external links<\/b>\u00a0for quotes or high authority sources referencing on your topic.<\/li><li>Write in a <b>language and wording<\/b> that meets the tonality and language usage of your target audience.<\/li><li><b>Proofread<\/b> for spelling errors and grammar.<\/li><\/ul><h3><b>Step 4: Check your blog posts for keywords<\/b><\/h3><div>Identify the keywords for your topic that will most likely be keyed in by your target audience for the search to get information on your topic. Only if you apply relevant content with relevant keywords in your blog post, your blog post will get found by your target audience. Select synonymous keywords and check the search volume for each keyword or keywords phrase (you can use the <a href=\"https:\/\/adwords.google.com\/home\/tools\/keyword-planner\/\">Google Adwords Tool<\/a> for this).<\/div><ul><li>Use your focus keyword and synonymous keywords according to search volume and place them appropriately into your blog post.<\/li><li>Most importantly include your <b>focus keyword in your headline<\/b> and <b>introduction<\/b> as well as in <b>subheadings<\/b> and <b>copy<\/b>.<\/li><li>Use your focus keyword also in <b>image and video title and descriptions. <\/b><\/li><li>Apply your keywords<b>\u00a0<\/b>in your <b>anchor texts for internal links<\/b>.<\/li><li><strong>Note:<\/strong> Only use relevant context to place your keywords.<\/li><li>Don\u2019t overstuff your text, keep a <b>keyword density of 1-3%<\/b>.<\/li><li>Use synonyms to get more variation and always keep your text flow consistently. You are still writing for humans, not for engines.<\/li><\/ul><h3><strong>Step 5: Use the power of images and videos to visualize your content<\/strong><\/h3><div>\u201c<b>A picture is worth a thousand words<\/b>\u201c. This proverb has taken on a specific significance in the age of social media. <b>Visual content<\/b> has become more important than ever to capture the attention of today\u2019s media consumers. Photos, images, and graphics <b>trigger emotions<\/b> and make complex information easier to understand and more memorable.\u00a0Visual information is perceived on a higher emotional level than using plain texts. Your readers will understand better the concepts that you are outlining in your blog post if you visualize your key aspects.<\/div><ul><li>You can visualize <b>tutorials or technical instructions by using screenshots or screencast videos<\/b>.<\/li><li>Use <b>infographics for creating checklists, pros &amp; cons, step-by-step-guides, flow charts<\/b> or even for telling stories.<\/li><\/ul><div>Images are more likely to be liked and shared than plain text. So, images will drive more traffic to your post from <b>social media sites<\/b>. And, images with relevant keywords will also help you to get more matches in the <strong>search engines<\/strong> like Google image search or YouTube.\u00a0Only use relevant images for your content and don\u2019t overstuff your blog post with visuals.<\/div><div><b>\u00a0<\/b><\/div><div><b>Tip<\/b>: You can use <a href=\"http:\/\/www.canva.com\/\">Canva<\/a> to create easy infographics without being a designer<\/div><h3><strong>Step 6: Use links and strong CTAs for your blog posts<\/strong><\/h3><div>Once you have managed to drive people to your site, your want them to stay as long as possible. The best way to do this is to guide them to further readings on your site.<\/div><ul><li>Use sticky content and takeaways to make your readers subscribe to your blog or leave their e-mail address to follow-up.<\/li><li>Encourage your readers to read on or comment on your post or to grab a white paper, a checklist or an e-book by using <b>strong call-to-actions<\/b> (CTAs).<\/li><\/ul><h3><strong>Step 7: Promote your blog posts like hell and repurpose your content<\/strong><\/h3><div>Once you have published your new post, don\u2019t sit back and wait for your readers to come in in millions. You will need to promote your post to get outreach and performance. as hell. Start by sharing your blog post on your social media channels. Social media are very powerful sources to get more outreach for your content and to drive traffic to your blog. You should also re-purpose your content as much as possible and re-share it with all your media formats to extend the lifetime of your post.<\/div><ul><li>Share your post on <strong>Twitter, Facebook, Google+ and LinkedIn<\/strong>.<\/li><li>If you have <strong>business pages<\/strong> on Facebook, LinkedIn, and Google+, share your posts to your pages as well but at different times and days.<\/li><li>Also, share your blog post to relevant <strong>groups and communities<\/strong> on\u00a0Facebook, LinkedIn, and Google+, but share them on different days and times and use different messages tailored for each group.<\/li><li>Schedule your post <strong>multiple times<\/strong> on Twitter.<\/li><li>Share your blog post<strong> images on Pinterest, Instagram, and Flickr<\/strong>, using your focus image with your introduction, hashtags and a link to your post.<\/li><li>If you use multiple images, share them all, but share them on different days and at different times.<\/li><li>You can also share your <strong>images on Twitter, Facebook, LinkedIn and Google+<\/strong> to re-share your posts with different visual anchors.<\/li><li>Share your blog post in special interest communities such as <strong>Reddit<\/strong>.<\/li><li>Also share a link to your blog post to social bookmark networks such as <strong>Diigo, Digg, StumbleUpon<\/strong>.<\/li><li><strong>Schedule<\/strong> your posts at the <strong>best times for each network<\/strong> and across several days, weeks and month to extend the lifespan of your posts.<\/li><\/ul><h4>Re-purpose your content:<\/h4><ul><li><strong>Re-share<\/strong> and re-schedule your <strong>evergreen posts<\/strong> from time to time to revive them.<\/li><li><strong>Syndicate and re-publish<\/strong> your post on <strong>Medium, Tumblr, Linkedin Pulse, Bloglovin<\/strong>, but wait for 3-7 days to let Google index your original blog post first.<\/li><li>Turn your images and content into a<strong> slideshow or video to share<\/strong> on Slideshare and Youtube.<\/li><li>Turn longer post into a<strong> white paper, checklist or e-book<\/strong> as a takeaway for your blog and for re-sharing your content.<\/li><li>Reach out to <strong>influencers<\/strong> to present your post or link to in and to your post. Like, share and comment their posts to get attention, response, and backlinks to your post.<\/li><\/ul><div><strong>Note<\/strong>: Each social network offers\u00a0<a href=\"https:\/\/www.blog2social.com\/en\/blog\/a-complete-guide-to-social-media-sharing-what-when-how-to-share\/\"><strong>different options\u00a0for promoting your content<\/strong><\/a>\u00a0and <strong>communities react differently to post formats<\/strong>. Each network is used for different purposes and has rush hours at <strong>different times<\/strong>.\u00a0To effectively <strong><a href=\"https:\/\/www.blog2social.com\/en\/blog\/crossposting-cross-promoting-social-media\/\">cross-promote your blog posts<\/a><\/strong>, don\u2019t just cross-post the same message on every social network. <strong>Tailor the message<\/strong> for your posts to fit each social media platform and audience. Get the<strong> <a href=\"https:\/\/www.blog2social.com\/en\/blog\/complete-guide-social-media-sharing\/\">Complete\u00a0Guide to Social Media Sharing: What, When &amp; How to Share<\/a>\u00a0<\/strong>to help you set up your cross-promotion plan and how to use tools to help you with this task.<\/div><div><b>Tip<\/b>: The <a href=\"https:\/\/wordpress.org\/plugins\/blog2social\">Blog2Social WordPress Plugin<\/a> allows you to share individually tailored social media posts with one click on all social media platforms such as Twitter, Facebook, LinkedIn, Google+, Instagram, Pinterest, Reddit, Medium, Blogloving and much more.\u00a0 Learn more about the <a href=\"http:\/\/www.blog2social.com\/ rel=nofollow\">Blog2Social Plugin<\/a>\u00a0and the <a href=\"https:\/\/www.blog2social.com\/de\/webapp\/ rel=nofollow\">Blog2Social Web App<\/a>.<\/div><h2>Wrap up<\/h2><p>There are several things you can do to turn your blog posts into great content that stands out among millions of other reading options for your target audience. Writing about a very specific topic for a very specific audience by using the right keywords is one thing. Optimizing your post for the human reader by considering some basic SEO criteria will make sure your content gets better indexed for searches by your target audience. Promoting and seeding your blog post on multiple social media and influencer outlets will help you to maximize outreach and performance. Download your 7-step-checklist\u00a0to make sure, you think of everything you can do to write perfect blog posts.<\/p><div><figure id=\"attachment_54374\" class=\"wp-caption aligncenter\"><a href=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg\"><img class=\"wp-image-54374 size-full\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg\" alt=\"checklist the perfect blog post infographic\" width=\"800\" height=\"2000\" srcset=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg 800w, https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g-360x900.jpg 360w, https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g-768x1920.jpg 768w, https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g-410x1024.jpg 410w\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/\" data-jpibfi-post-title=\"7 Steps To Writing Perfect Blog Posts [Checklist]\" data-jpibfi-src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg\" \/><\/a><figcaption class=\"wp-caption-text\">Checklist: The perfect blog post<\/figcaption><\/figure><h3>Share the infographik on your blog and to your social media!<\/h3><p>Take this code to embed this infographic on your site:<\/p><p><textarea rows=\"10\" cols=\"50\">&lt;p&gt;&lt;strong&gt;Please include attribution to http:\/\/www.blog2social.com with this graphic&lt;\/strong&gt;&lt;a https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg\u201d target=\u201d_blank\u201d rel=\u201dhttps:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg\u201d&gt;&lt;img class=\u201dsize-full wp-image-54374\u2033 src=\u201dhttps:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_infographic_g.jpg\u201d alt=\u201dChecklist: The perfect blog post\u201d width=\u201d800\u2033 height=\u201d2000\u2033 \/&gt;&lt;\/a&gt;&lt;\/p&gt;<\/textarea><\/p><\/div><\/div><div><div class=\"row\"><img class=\"alignnone size-medium wp-image-54668\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_header-icegram-600x194.png\" alt=\"7 Steps to Writing Perfect Blog Posts [Checklist]\" width=\"600\" height=\"194\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/\" data-jpibfi-post-title=\"7 Steps To Writing Perfect Blog Posts [Checklist]\" data-jpibfi-src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/checklist-the-perfect-blog-post_header-icegram-600x194.png\" \/><br \/><strong>Download the Checklist 7 Steps to Writing Perfect Blog Posts<\/strong><div class=\"rm_form_container rainmaker_form\" data-type=\"rm_custom\" data-form-id=\"54675\"><form class=\"DSGVO\" action=\"https:\/\/e.asend.de\/index.php?entryPoint=WebToLeadCapture\" method=\"post\"> <div><label for=\"salutation _frau\"><input type=\"radio\" name=\"salutation\" required=\"required\" value=\"Ms.\" \/> Ms.<\/label> <label for=\"salutation _herr\"><input type=\"radio\" name=\"salutation\" required=\"required\" value=\"Mr.\" \/> Mr.<\/label> <input type=\"text\" name=\"first_name\" required=\"required\" placeholder=\"Firstname\" \/><input type=\"text\" name=\"last_name\" required=\"required\" placeholder=\"Lastname\" \/><input type=\"email\" name=\"email1\" required=\"required\" placeholder=\"email\" \/><input name=\"sprache_c\" type=\"hidden\" value=\"englisch\" \/><input name=\"assigned_user_id\" type=\"hidden\" value=\"\" \/><input name=\"geeignet_fuer_produkt_c\" type=\"hidden\" value=\"prgateway,blog2social\" \/><input name=\"lead_source\" type=\"hidden\" value=\"Registration\" \/><input name=\"redirect_url\" type=\"hidden\" value=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/?v=confirm\" \/><input name=\"campaign_id\" type=\"hidden\" value=\"34325c86-f5d3-fe3b-8ee0-59ae82d8d005\" \/><input name=\"dokument_c\" type=\"hidden\" value=\"Checklist 7 Steps to Writing Perfect Blog Posts\" \/><input name=\"dokumentenprojekt_c\" type=\"hidden\" value=\"Checklist\" \/><input name=\"dokumentenurl_c\" type=\"hidden\" value=\"https:\/\/www.blog2social.com\/docs\/checklist_the-perfect-blog-post.pdf\" \/><input name=\"dokumententyp_c\" type=\"hidden\" value=\"CaseStudy\" \/><div class=\"remark\"><p>We will never share or sell your email and personal data and we will always keep your data safe and secure.<\/p><p><label title=\"You may at any time revoke your consent\"><input type=\"checkbox\" class=\"DSGVO\" name=\"dsgvo_c\" value=\"eingewilligt\" \/> I agree to the <a href=\"https:\/\/www.adenion.de\/privacy-policy\" target=\"_blank\">privacy policy<\/a> and <a href=\"https:\/\/www.blog2social.com\/en\/terms\/\" target=\"_blank\">terms of service<\/a>. (You can unsubscribe any time.)<\/label><\/p><\/div> <input class=\"DSGVOsubmit\" disabled=\"disabled\" type=\"submit\" value=\"download for free\" \/><\/div><\/form><div class=\"rm-loader\"><\/div><\/div><\/div><div id=\"yasr_visitor_votes_11926\" class=\"yasr-visitor-votes\"><div class=\"yasr-container-custom-text-and-visitor-rating\"> Our Reader Score<\/div><div class=\"rateit bigstars yasr_visitor_votes_stars_div\" id=\"yasr_rateit_visitor_votes_11926\" data-postid=\"11926\" data-rateit-starwidth=\"32\" data-rateit-starheight=\"32\" data-rateit-value=\"5\" data-rateit-step=\"1\" data-rateit-resetable=\"false\" data-rateit-readonly=\"false\">[Average: 5]<\/div><div class=\"shariff shariff-align-flex-start shariff-widget-align-flex-start\" data-services=\"twitter%7Cfacebook%7Cxing%7Clinkedin%7Cvk%7Cpinterest\" data-url=\"https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F\" data-timestamp=\"1507124001\" data-backendurl=\"https:\/\/www.blog2social.com\/en\/blog\/wp-json\/shariff\/v1\/share_counts?\"><ul class=\"shariff-buttons theme-round orientation-horizontal buttonsize-medium\"><li class=\"shariff-button twitter shariff-nocustomcolor\"><a href=\"https:\/\/twitter.com\/share?url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F&amp;text=7%20Steps%20To%20Writing%20Perfect%20Blog%20Posts%20%5BChecklist%5D\" title=\"Share on Twitter\" aria-label=\"Share on Twitter\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">tweet\u00a0\u00a0<\/a><\/li><li class=\"shariff-button facebook shariff-nocustomcolor\"><a href=\"https:\/\/www.facebook.com\/sharer\/sharer.php?u=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F\" title=\"Share on Facebook\" aria-label=\"Share on Facebook\" role=\"button\" rel=\"nofollow\" class=\"shariff-link\" target=\"_blank\">share\u00a0\u00a0<\/a><\/li><li class=\"shariff-button googleplus shariff-nocustomcolor\"><a href=\"https:\/\/plus.google.com\/share?url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F\" title=\"Share on Google+\" aria-label=\"Share on Google+\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">share\u00a0<\/a><\/li><li class=\"shariff-button xing shariff-nocustomcolor\"><a href=\"https:\/\/www.xing.com\/social_plugins\/share?url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F\" title=\"Share on XING\" aria-label=\"Share on XING\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">share\u00a0 5\u00a0<\/a><\/li><li class=\"shariff-button linkedin shariff-nocustomcolor\"><a href=\"https:\/\/www.linkedin.com\/shareArticle?mini=true&amp;url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F&amp;title=7%20Steps%20To%20Writing%20Perfect%20Blog%20Posts%20%5BChecklist%5D\" title=\"Share on LinkedIn\" aria-label=\"Share on LinkedIn\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">share\u00a0 0\u00a0<\/a><\/li><li class=\"shariff-button vk shariff-nocustomcolor\"><a href=\"https:\/\/vk.com\/share.php?url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F\" title=\"Share on VK\" aria-label=\"Share on VK\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">share\u00a0 0\u00a0<\/a><\/li><li class=\"shariff-button pinterest shariff-nocustomcolor\"><a href=\"https:\/\/www.pinterest.com\/pin\/create\/link\/?url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F&amp;media=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2Fwp-content%2Fuploads%2F2017%2F09%2Fthe-perfect-blog-post_header.jpg&amp;description=7%20Steps%20To%20Writing%20Perfect%20Blog%20Posts%20%5BChecklist%5D\" title=\"Pin it on Pinterest\" aria-label=\"Pin it on Pinterest\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">save\u00a0 57\u00a0<\/a><\/li><li class=\"shariff-button pocket shariff-nocustomcolor\"><a href=\"https:\/\/getpocket.com\/edit?url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F&amp;title=7%20Steps%20To%20Writing%20Perfect%20Blog%20Posts%20%5BChecklist%5D\" title=\"Save to Pocket\" aria-label=\"Save to Pocket\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">pocket\u00a0<\/a><\/li><li class=\"shariff-button flipboard shariff-nocustomcolor\"><a href=\"https:\/\/share.flipboard.com\/bookmarklet\/popout?v=2&amp;title=7%20Steps%20To%20Writing%20Perfect%20Blog%20Posts%20%5BChecklist%5D&amp;url=https%3A%2F%2Fwww.blog2social.com%2Fen%2Fblog%2F7-steps-perfect-blog-post%2F\" title=\"Share on Flipboard\" aria-label=\"Share on Flipboard\" role=\"button\" rel=\"noopener nofollow\" class=\"shariff-link\" target=\"_blank\">share\u00a0<\/a><\/li><\/ul><\/div><div class=\"rp4wp-related-posts\"><h3>Related Posts<\/h3><ul><li><div class=\"rp4wp-related-post-content\"> <a href=\"https:\/\/www.blog2social.com\/en\/blog\/5-tips-for-successful-social-media-posts\/\">5 Tips for successful social media posts<\/a><p>Social media is in continuous change. Popular networks loose users and importance, meanwhile new networks\u2026<\/p><\/div><\/li><li><div class=\"rp4wp-related-post-image\"> <a href=\"https:\/\/www.blog2social.com\/en\/blog\/video-customized-social-media-postings-achieve-coverage\/\"><input class=\"jpibfi\" type=\"hidden\" \/><img width=\"150\" height=\"150\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2016\/04\/Gie\u00dfkanne-150x150.png\" class=\"attachment-thumbnail size-thumbnail wp-post-image\" alt=\"Customization instead of scattergun approach\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/\" data-jpibfi-post-title=\"7 Steps To Writing Perfect Blog Posts [Checklist]\" data-jpibfi-src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2016\/04\/Gie\u00dfkanne-150x150.png\" \/><\/a><\/div><div class=\"rp4wp-related-post-content\"> <a href=\"https:\/\/www.blog2social.com\/en\/blog\/video-customized-social-media-postings-achieve-coverage\/\">[Video]: Why Customizing Social Media Posts Leads To More Likes<\/a><p>Customization instead of scattergun approach: Why customized social media posts\u00a0lead to\u00a0more likes Studies show that\u2026<\/p><\/div><\/li><li><div class=\"rp4wp-related-post-image\"> <a href=\"https:\/\/www.blog2social.com\/en\/blog\/video-best-times-post-social-networking-sites\/\"><input class=\"jpibfi\" type=\"hidden\" \/><img width=\"150\" height=\"150\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2016\/07\/best-times-social-media-table-150x150.png\" class=\"attachment-thumbnail size-thumbnail wp-post-image\" alt=\"the best time to post on social media\" data-jpibfi-post-excerpt=\"\" data-jpibfi-post-url=\"https:\/\/www.blog2social.com\/en\/blog\/7-steps-perfect-blog-post\/\" data-jpibfi-post-title=\"7 Steps To Writing Perfect Blog Posts [Checklist]\" data-jpibfi-src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2016\/07\/best-times-social-media-table-150x150.png\" \/><\/a><\/div><div class=\"rp4wp-related-post-content\"> <a href=\"https:\/\/www.blog2social.com\/en\/blog\/video-best-times-post-social-networking-sites\/\">[Video]: The best times to post on social networking sites<\/a><p>When are the best times to post on social networking sites? When are the best\u2026<\/p><\/div><\/li><\/ul><\/div><div id=\"sexy-author-bio\" class=\"melanie-tambl\u00e9\"><div id=\"sab-author\"><a rel=\"nofollow author\" href=\"https:\/\/de.linkedin.com\/in\/melanietamble\" title=\"Melanie Tambl\u00e9\" target=\"_blank\">Melanie Tambl\u00e9<\/a><\/div><div id=\"sab-gravatar\"><a rel=\"nofollow\" href=\"https:\/\/de.gravatar.com\/mtamble\" target=\"_blank\"><img alt=\"Melanie Tambl\u00e9\" src=\"https:\/\/www.gravatar.com\/avatar\/b4a733c9d5a769199add6844160c6bb6\" \/><\/a><\/div><div id=\"sab-byline\">Co-Founder &amp; CEO at <a rel=\"nofollow\" href=\"http:\/\/www.adenion.de\" target=\"_blank\">Adenion GmbH<\/a><\/div><div id=\"sab-description\">Melanie Tambl\u00e9 is co-founder and co-CEO of <a href=\"http:\/\/www.adenion.de\" rel=\"nofollow\">Adenion GmbH<\/a>. She is an experienced expert in content marketing and social media. <br \/> Adenion GmbH specializes in online services and tools for bloggers, businesses and agencies of any size to support their online marketing and content seeding tasks on the web. <p><a href=\"http:\/\/www.blog2social.com\" rel=\"nofollow\">Blog2Social<\/a> as <a href=\"https:\/\/wordpress.org\/plugins\/blog2social\" rel=\"nofollow\">WordPress Plugin<\/a> and <a href=\"https:\/\/blog2social.com\/en\/webapp\" rel=\"nofollow\">WebApp<\/a> enable fast and easy auto-posting, scheduling and cross-promotion of blog posts, articles, links, images, videos and documents across multiple social media sites. <br \/> Social media posts will be automatically turned into a customized format for each social platform and auto-scheduled for the best time. Social media post can be previewed and tailored with individual post formats, images or personal comments - all in one easy step.<\/p><\/div><div id=\"sab-social-wrapper\"><a id=\"sab-Facebook\" rel=\"nofollow\" href=\"https:\/\/www.facebook.com\/melanie.tamble\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Facebook\" alt=\"Melanie Tambl\u00e9 on Facebook\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sabfacebook.png\" \/><\/a><a id=\"sab-Google\" rel=\"nofollow\" href=\"https:\/\/plus.google.com\/u\/0\/+MelanieTamble\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Google\" alt=\"Melanie Tambl\u00e9 on Google\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sabgoogle.png\" \/><\/a><a id=\"sab-Instagram\" rel=\"nofollow\" href=\"https:\/\/www.instagram.com\/melanietamble\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Instagram\" alt=\"Melanie Tambl\u00e9 on Instagram\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sabinstagram.png\" \/><\/a><a id=\"sab-Linkedin\" rel=\"nofollow\" href=\"https:\/\/de.linkedin.com\/in\/melanietamble\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Linkedin\" alt=\"Melanie Tambl\u00e9 on Linkedin\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sablinkedin.png\" \/><\/a><a id=\"sab-Pinterest\" rel=\"nofollow\" href=\"https:\/\/www.pinterest.com\/melanietamble\/\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Pinterest\" alt=\"Melanie Tambl\u00e9 on Pinterest\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sabpinterest.png\" \/><\/a><a id=\"sab-Tumblr\" rel=\"nofollow\" href=\"http:\/\/melanietamble.tumblr.com\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Tumblr\" alt=\"Melanie Tambl\u00e9 on Tumblr\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sabtumblr.png\" \/><\/a><a id=\"sab-Twitter\" rel=\"nofollow\" href=\"https:\/\/twitter.com\/melanie_tamble\" target=\"_blank\"><img class=\"bio-icon\" id=\"sig-Twitter\" alt=\"Melanie Tambl\u00e9 on Twitter\" src=\"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/plugins\/sexy-author-bio\/public\/assets\/images\/flat-square\/sabtwitter.png\" \/><\/a><\/div><\/div><div class=\"author-meta vcard author post-author\"><h2 class=\"author-heading\">About the author<\/h2><div class=\"author-description\"><h3 class=\"author-title\">Melanie Tambl\u00e9<\/h3><p class=\"author-bio\"> Melanie Tambl\u00e9 is co-founder and co-CEO of <a href=\"http:\/\/www.adenion.de\" rel=\"nofollow\">Adenion GmbH<\/a>. She is an experienced expert in content marketing and social media. \nAdenion GmbH specializes in online services and tools for bloggers, businesses and agencies of any size to support their online marketing and content seeding tasks on the web. <a href=\"http:\/\/www.blog2social.com\" rel=\"nofollow\">Blog2Social<\/a> as <a href=\"https:\/\/wordpress.org\/plugins\/blog2social\" rel=\"nofollow\">WordPress Plugin<\/a> and <a href=\"https:\/\/blog2social.com\/en\/webapp\" rel=\"nofollow\">WebApp<\/a> enable fast and easy auto-posting, scheduling and cross-promotion of blog posts, articles, links, images, videos and documents across multiple social media sites. \nSocial media posts will be automatically turned into a customized format for each social platform and auto-scheduled for the best time. Social media post can be previewed and tailored with individual post formats, images or personal comments - all in one easy step. <a class=\"author-link\" href=\"https:\/\/www.blog2social.com\/en\/blog\/author\/melanie-tamble\/\" rel=\"author\"> View all posts by Melanie Tambl\u00e9 <\/a><\/p><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:42:35+0000","updated_at":"2018-11-07T16:42:35+0000","published_at":"2017-09-01T13:35:03+0000","published_by":["About the authorMelanie Tambl\u00e9 Melanie Tambl\u00e9 is co-founder and co-CEO of Adenion GmbH. She is an experienced expert in content marketing and social media. \nAdenion GmbH specializes in online services and tools for bloggers, businesses and agencies of any size to support their online marketing and content seeding tasks on the web. Blog2Social as WordPress Plugin and WebApp enable fast and easy auto-posting, scheduling and cross-promotion of blog posts, articles, links, images, videos and documents across multiple social media sites. \nSocial media posts will be automatically turned into a customized format for each social platform and auto-scheduled for the best time. Social media post can be previewed and tailored with individual post formats, images or personal comments - all in one easy step.  View all posts by Melanie Tambl\u00e9"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":13,"domain_name":"www.blog2social.com","preview_picture":"https:\/\/www.blog2social.com\/en\/blog\/wp-content\/uploads\/2017\/09\/the-perfect-blog-post_header.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":257,"title":"Kungsgeten\/org-brain","url":"https:\/\/github.com\/Kungsgeten\/org-brain\/blob\/master\/README.org","content":"<p><code>org-brain<\/code> implements a variant of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Concept_map\" rel=\"nofollow\">concept mapping<\/a> in Emacs, using <a href=\"http:\/\/orgmode.org\/\" rel=\"nofollow\">org-mode<\/a>. It\n  is heavily inspired by a piece of software called <a href=\"http:\/\/thebrain.com\/\" rel=\"nofollow\">The Brain<\/a>, and you can view an\n  introduction to that program <a href=\"https:\/\/www.youtube.com\/watch?v=GFqLUBKCFdA\" rel=\"nofollow\">here<\/a>. They also provide <a href=\"https:\/\/www.thebrain.com\/blog\/\" rel=\"nofollow\">a blog<\/a> with great ideas of\n  how you can think when organizing your Brain.<\/p>\n<p>You can think of <code>org-brain<\/code> as a combination of a wiki and a mind map, where\n  each wiki page \/ mind map node is an <code>org-mode<\/code> file which resides in your\n  <code>org-brain-path<\/code>, or a headline with an ID property in one of those files. These\n  are called <i>entries<\/i>. Entries can be linked together, and you can then\n  view the network of links as a mind map, using <code>M-x org-brain-visualize<\/code>.<\/p>\n<pre>PINNED:  Index\n               +-Python              Game development-+-Game design\n               +-Programming books           |\n   Programming-+-Emacs                       |\n         |                                   |\n         +-----------------+-----------------+\n                           |\n                           \u25bd\n                    Game programming \u2190\u2192 Computer games\nGame Maker  Unity\n--- Resources ---------------------------------\n\u2022 https:\/\/en.wikipedia.org\/wiki\/Game_programming\n\u2022 Passing Through Ghosts in Pac-Man\n\u2022 In-House Engine Development: Technical Tips\n--- Text --------------------------------------\nGame programming is the art of programming computer games...\n<\/pre>\n<p>When visualizing an entry, you will see the entry\u2019s relationship to other\n  entries. There are four different types of relationships in <code>org-brain<\/code>:<\/p>\n<dl><dt>Parents<\/dt><dd>Entries above the visualized entry. If the visualized entry is a\n    headline, then the parent headline in the <code>org-mode<\/code> file will be\n    one of the parents. In the case of top level headlines, the file\n    itself will be considered a parent. Additional parents can be added\n    manually. In the example above, <i>Programming<\/i> and <i>Game\n    development<\/i> are parents of the visualized <i>Game programming<\/i>\n    entry.<\/dd>\n  <dt>Children<\/dt><dd>Entries below the visualized entry. This will by default be\n    subheadings of the visualized entry (or level one headlines, if\n    the visualized entry is a file). You can add other children,\n    residing elsewhere, manually. In the example above, <i>Game Maker<\/i>\n    and <i>Unity<\/i> are the children of <i>Game programming<\/i>.<\/dd>\n  <dt>Siblings<\/dt><dd>These appear to the right of the parent entries. Siblings are the\n    other children of the visualized entry\u2019s parents.<\/dd>\n  <dt>Friends<\/dt><dd>These appear to the right of the visualized entry. Friends provide\n    a way of adding a hierarchy independent two-way relationship\n    between two entries. Friends must be added manually. In the example\n    above, <i>Computer games<\/i> and <i>Game programming<\/i> are friends.<\/dd>\n<\/dl><p><a href=\"http:\/\/blogarchive.thebrain.com\/thought-relationships\/\" rel=\"nofollow\">Here\u2019s an article<\/a> describing how you can use the different relationships (The\n  Brain\u2019s <i>jump thoughts<\/i> are the equivalent of <i>friends<\/i> in <code>org-brain<\/code>).<\/p>\n<p>Apart from the visualized entry\u2019s relationships, <code>org-brain-visualize<\/code> also show\n  pinned entries, which are shown independently of the visualized entry; <i>Index<\/i>\n  is a pinned entry in the example above. <code>org-brain-visualize<\/code> also show a list\n  of the entry\u2019s resources (links and attachments), and the text in the entry. The\n  example above have three resources, and a short text. The resources and text is\n  gathered from <code>org-mode<\/code> automatically.<\/p>\n<p>There\u2019s also the option to visualize the entry as a tree, or similar to a\n  mind map, where you can zoom in order to show grandparents and grandchildren.<\/p>\n<p>The relationship entries, pinned entries and resources are all links; they can\n  be pressed\/clicked to visualize other entries, visit resources etc.<\/p>\n<p>The easiest way is to get <code>org-brain<\/code> from MELPA. If you do not want to do that,\n  clone this git repository or download <code>org-brain.el<\/code> and add it to your\n  load-path. The example below is using <a href=\"https:\/\/github.com\/jwiegley\/use-package\">use-package<\/a> and assumer that you\u2019re using\n  MELPA, but you could use <code>(require 'org-brain)<\/code> or add a <code>:load-path<\/code> to\n  <code>use-package<\/code> instead.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package org-brain :ensure t\n  :init\n  (setq org-brain-path \"directory\/path\/where-i-want-org-brain\")\n  ;; For Evil users\n  (with-eval-after-load 'evil\n    (evil-set-initial-state 'org-brain-visualize-mode 'emacs))\n  :config\n  (setq org-id-track-globally t)\n  (setq org-id-locations-file \"~\/.emacs.d\/.org-id-locations\")\n  (push '(\"b\" \"Brain\" plain (function org-brain-goto-end)\n          \"* %i%?\" :empty-lines 1)\n        org-capture-templates)\n  (setq org-brain-visualize-default-choices 'all)\n  (setq org-brain-title-max-length 12))<\/pre><\/div>\n<ol><li><code>org-brain<\/code> requires Emacs 25 and org-mode 9. These need to be part of your\n    Emacs.<\/li>\n  <li>Configure <code>org-brain-path<\/code> (defaults to <code>\/brain<\/code> in your <code>org-directory<\/code>) to\n    a directory where you want to put your <code>org-brain<\/code> files (which could be the\n    location where you already keep your org files if you wish to transform your\n    existing org files into <code>org-brain<\/code> files). You can set this with the example\n    config presented above or through <code>M-x customize-group RET org-brain<\/code>.<\/li>\n  <li>If you\u2019re an <a href=\"https:\/\/github.com\/emacs-evil\/evil\">evil<\/a> user, you\u2019ll want to add <code>(evil-set-initial-state\n    'org-brain-visualize-mode 'emacs)<\/code> to your <code>org-brain<\/code> configuration.<\/li>\n  <li><code>org-brain<\/code> use <code>org-id<\/code> in order to speed things up. Because of this, the\n    variable <code>org-id-track-globally<\/code> should be <code>t<\/code> (which it already is by\n    default). You may want to modify <code>org-id-locations-file<\/code> too.<\/li>\n  <li>You might want to add information at the end of an entry, without visiting\n    the file. A way to do this is to use a <a href=\"http:\/\/orgmode.org\/manual\/Capture.html\" rel=\"nofollow\">capture<\/a> template, such as the one\n    presented above.<\/li>\n  <li>If you have a lot of entries, it might take some time to gather information\n    about all entries when using <code>org-brain-visualize<\/code>. You could change the\n    value of <code>org-brain-visualize-default-choices<\/code> (which is =\u2019all= by default)\n    to only include files, or even just files in the direct root of\n    <code>org-brain-path<\/code>.<\/li>\n  <li>If you feel that <code>org-brain-visualize<\/code> is too cluttered, you may want to set\n    <code>org-brain-show-resources<\/code> and\/or <code>org-brain-show-text<\/code> to <code>nil<\/code>.<\/li>\n  <li>If you have very long entry names, <code>org-brain-visualize<\/code> may take a lot of\n    horizontal space. You can cap the shown length of entry titles, by setting\n    <code>org-brain-title-max-length<\/code>.<\/li>\n<\/ol>\n<p>There are two types of entries in <code>org-brain<\/code>: <i>headline<\/i> entries and <i>file<\/i>\n  entries. For the most part these are used the same way, and the main difference\n  between them is how their content is stored inside your <code>org-brain<\/code> directory.\n  All .org-files inside the <code>org-brain-path<\/code> are considered as <i>file<\/i> entries (the\n  content typically being the text before the first headline in the file) and all\n  headlines <i>with an ID property<\/i> inside these files are considered as <i>headline<\/i>\n  entries.<\/p>\n<p>If you have a headline entry, which you want to convert to a file entry, use\n  <code>M-x org-brain-headline-to-file<\/code>. Unfortunately there is currently no function\n  to convert a file entry into a headline entry.<\/p>\n<p>If you find that <code>org-brain<\/code> is missing entries, or list entries which doesn\u2019t\n  exist, try using <code>M-x org-brain-update-id-locations<\/code>, which syncs the\n  <code>org-brain<\/code> entries with the <code>org-id<\/code> caching system.<\/p>\n<h2><a id=\"user-content-org-brain-visualize\" class=\"anchor\" aria-hidden=\"true\" href=\"#org-brain-visualize\"><\/a><code>org-brain-visualize<\/code><\/h2>\n<p>The primary usage of <code>org-brain<\/code> is through <code>M-x org-brain-visualize<\/code> (which you\n  might want to bind to a key). From there you can browse entries, add\/remove\n  relationships, open entries for editing etc. The following keybindings are\n  available in <code>org-brain-visualize<\/code>:<\/p>\n<table><tbody><tr><th>Key<\/th><th>Command<\/th><th>Description<\/th><\/tr><tr><td>m<\/td><td><code>org-brain-visualize-mind-map<\/code><\/td><td>Toggle between normal and mind-map visualization.<\/td><\/tr><tr><td>j or TAB<\/td><td><code>forward-button<\/code><\/td><td>Goto next link<\/td><\/tr><tr><td>k or S-TAB<\/td><td><code>backward-button<\/code><\/td><td>Goto previous link<\/td><\/tr><tr><td>b<\/td><td><code>org-brain-visualize-back<\/code><\/td><td>Like the back button in a web browser.<\/td><\/tr><tr><td>h or *<\/td><td><code>org-brain-new-child<\/code><\/td><td>Add a new child <i>headline<\/i> to entry<\/td><\/tr><tr><td>c<\/td><td><code>org-brain-add-child<\/code><\/td><td>Add an existing entry, or a new <i>file<\/i>, as a child<\/td><\/tr><tr><td>C<\/td><td><code>org-brain-remove-child<\/code><\/td><td>Remove one the entry\u2019s child relations<\/td><\/tr><tr><td>p<\/td><td><code>org-brain-add-parent<\/code><\/td><td>Add an existing entry, or a new <i>file<\/i>, as a parent<\/td><\/tr><tr><td>P<\/td><td><code>org-brain-remove-parent<\/code><\/td><td>Remove one of the entry\u2019s parent relations<\/td><\/tr><tr><td>f<\/td><td><code>org-brain-add-friendship<\/code><\/td><td>Add an existing entry, or a new <i>file<\/i>, as a friend<\/td><\/tr><tr><td>F<\/td><td><code>org-brain-remove-friendship<\/code><\/td><td>Remove one of the entry\u2019s friend relations<\/td><\/tr><tr><td>n<\/td><td><code>org-brain-pin<\/code><\/td><td>Toggle if the entry is pinned or not<\/td><\/tr><tr><td>t<\/td><td><code>org-brain-set-title<\/code><\/td><td>Change the title of the entry.<\/td><\/tr><tr><td>T<\/td><td><code>org-brain-set-tags<\/code><\/td><td>Change the tags of the entry.<\/td><\/tr><tr><td>d<\/td><td><code>org-brain-delete-entry<\/code><\/td><td>Choose an entry to delete.<\/td><\/tr><tr><td>l<\/td><td><code>org-brain-visualize-add-resource<\/code><\/td><td>Add a new resource link in entry<\/td><\/tr><tr><td>C-y<\/td><td><code>org-brain-visualize-paste-resource<\/code><\/td><td>Add a new resource link from clipboard<\/td><\/tr><tr><td>a<\/td><td><code>org-brain-visualize-attach<\/code><\/td><td>Run <code>org-attach<\/code> on entry (headline entries only)<\/td><\/tr><tr><td>A<\/td><td><code>org-brain-archive<\/code><\/td><td>Archive the entry (headline entries only)<\/td><\/tr><tr><td>o<\/td><td><code>org-brain-goto-current<\/code><\/td><td>Open current entry for editing<\/td><\/tr><tr><td>O<\/td><td><code>org-brain-goto<\/code><\/td><td>Choose and edit one of your <code>org-brain<\/code> entries<\/td><\/tr><tr><td>v<\/td><td><code>org-brain-visualize<\/code><\/td><td>Choose and visualize a different entry<\/td><\/tr><tr><td>r<\/td><td><code>org-brain-visualize-random<\/code><\/td><td>Visualize one of your entries at random.<\/td><\/tr><tr><td>R<\/td><td><code>org-brain-visualize-wander<\/code><\/td><td>Visualize at random, in a set interval. <code>R<\/code> again to cancel.<\/td><\/tr><\/tbody><\/table><p>You can also use <code>org-store-link<\/code> inside of <code>org-brain-visualize<\/code> in order to\n  store a link to the currently visualized <code>org-brain<\/code> entry.<\/p>\n<p>When using the mind map visualization (toggle by pressing <code>m<\/code>), you can use the\n  following keybindings in order to add\/remove grandparents\/grandchildren:<\/p>\n<table><tbody><tr><th>Key<\/th><th>Command<\/th><th>Description<\/th><\/tr><tr><td>+<\/td><td><code>org-brain-visualize-add-grandchild<\/code><\/td><td>Add another grandchild level<\/td><\/tr><tr><td>-<\/td><td><code>org-brain-visualize-remove-grandchild<\/code><\/td><td>Remove a grandchild level<\/td><\/tr><tr><td>z<\/td><td><code>org-brain-visualize-add-grandparent<\/code><\/td><td>Add another grandparent level<\/td><\/tr><tr><td>Z<\/td><td><code>org-brain-visualize-remove-grandparent<\/code><\/td><td>Remove a grandparent level<\/td><\/tr><\/tbody><\/table><h2><a id=\"user-content-editing-from-org-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#editing-from-org-mode\"><\/a>Editing from <code>org-mode<\/code><\/h2>\n<p>You can edit <code>org-brain<\/code> entries directly from <code>org-mode<\/code>. You can use the\n  default <code>org-mode<\/code> outline structure to define parent\/children relationships,\n  but keep in mind that only entries with an <code>ID<\/code> property will be considered as\n  entries to <code>org-brain<\/code>; use <code>M-x org-id-get-create<\/code> to create an <code>ID<\/code>\n  property to the current <code>org-mode<\/code> headline. Another alternative is to use <code>M-x\n  org-brain-refile<\/code> which will create the ids for you.<\/p>\n<p>Most of the commands available in <code>org-brain-visualize<\/code> can also be used in\n  <code>org-mode<\/code> directly, in which case they will operate on the \u201centry at point\u201d. In\n  other words you can use <code>M-x org-brain-add-child<\/code> directly from <code>org-mode<\/code> in\n  order to add a child to the <code>org-brain<\/code> entry at point. You may also want to use\n  the commands <code>org-brain-goto-&lt;relationsship&gt;<\/code> to navigate between entries.<\/p>\n<p>You may want to create a link to an <code>org-brain<\/code> entry in an <code>org-mode<\/code> file (not\n  necessarily an <code>org-brain<\/code> file itself). <code>org-brain<\/code> provides several link types\n  for this purpose. You can use <code>org-insert-link<\/code> (bound to <code>C-c C-l<\/code> in\n  <code>org-mode<\/code> by default) to insert one of these links. They all have in common\n  that they, when clicked, will open the <code>org-brain<\/code> entry for editing. When\n  inserting a link like this, <code>org-brain<\/code> will run completion upon all your\n  entries.<\/p>\n<dl><dt><code>brain:<\/code><\/dt><dd>The default kind of link. Just let\u2019s you visit another <code>org-brain<\/code> entry when clicked.<\/dd>\n  <dt><code>brain-child:<\/code><\/dt><dd>When inserted using <code>org-insert-link<\/code> this will make\n    the linked entry a child to the current <code>org-brain<\/code> entry,\n    upon completion. Keep in mind that this doesn\u2019t work if you\n    type the link manually; only by completion through\n    <code>org-insert-link<\/code>.<\/dd>\n  <dt><code>brain-parent:<\/code><\/dt><dd>Like <code>brain-child:<\/code> but makes the linked entry a parent of\n    the current entry.<\/dd>\n  <dt><code>brain-friend:<\/code><\/dt><dd>Like <code>brain-child:<\/code> but adds the linked entry as a friend.<\/dd>\n<\/dl><p>The names of the relationship inserting links (<code>brain-child<\/code>, <code>brain-parent<\/code> and <code>brain-friend<\/code>) can be customized with the variables <code>org-brain-child-link-name<\/code>, <code>org-brain-parent-link-name<\/code>, and <code>org-brain-friend-link-name<\/code>. This customization should be done before loading <code>org-brain<\/code>. If you\u2019re using <code>use-package<\/code>, put the customization in the <code>:init<\/code> block.<\/p>\n<h2><a id=\"user-content-general-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-information\"><\/a>General information<\/h2>\n<p>If you try to add a child\/parent\/friend to an entry which doesn\u2019t exist, that\n  entry will be created. The name of a new entry can be written like this:\n  <code>file::headline<\/code>. The <code>headline<\/code> will be created as a level one headline in\n  <code>file<\/code>.<\/p>\n<p>When adding children, parents, or friends, multiple entries can be added at once\n  by separating their titles with <code>org-brain-entry-separator<\/code> (which is <code>;<\/code> by\n  default). For instance <code>M-x org-brain-add-parent RET music;artists<\/code> would add\n  both <code>music<\/code> and <code>artists<\/code> as parents.<\/p>\n<p>Another available command is <code>M-x org-brain-agenda<\/code>, which can be used to run\n  <code>org-agenda<\/code> on your <code>org-brain<\/code> files.<\/p>\n<h2><a id=\"user-content-slashes-in-file-entry-titles\" class=\"anchor\" aria-hidden=\"true\" href=\"#slashes-in-file-entry-titles\"><\/a>Slashes in file entry titles<\/h2>\n<p>When giving a file entry a title, the title can not contain slashes (<code>\/<\/code>) if\n  <code>org-brain-file-entries-use-title<\/code> is <code>t<\/code>.<\/p>\n<h2><a id=\"user-content-renaming-files-in-org-brain\" class=\"anchor\" aria-hidden=\"true\" href=\"#renaming-files-in-org-brain\"><\/a>Renaming files in <code>org-brain<\/code><\/h2>\n<p>Headline entries use <code>org-id<\/code> to identify themselves, so the headlines can be\n  manually renamed without worries. File entries, on the other hand, uses the\n  filename as the identifier. This will cause problems if you try to manually\n  rename files inside of <code>org-brain<\/code>.<\/p>\n<p>In order to rename a file, use <code>M-x org-brain-rename-file<\/code>.<\/p>\n<h2><a id=\"user-content-archiving-entries\" class=\"anchor\" aria-hidden=\"true\" href=\"#archiving-entries\"><\/a>Archiving entries<\/h2>\n<p><code>org-archive<\/code> has a problem in <code>org-brain<\/code>: relationships are maintained, even\n  though the entry should really be removed from the brain. Because of this,\n  please use <code>org-brain-archive<\/code> instead. This command removes relationships to\n  the entry in the brain, before archiving it. The command also inserts handy\n  links to the archived entry\u2019s relationships.<\/p>\n<h2><a id=\"user-content-special-tags\" class=\"anchor\" aria-hidden=\"true\" href=\"#special-tags\"><\/a>Special tags<\/h2>\n<p>You might have a headline which you do not really want as an entry in\n  <code>org-brain<\/code>. The most basic way to exclude such a headline is simply to not add\n  an <code>ID<\/code> property to it. However, <code>org-brain<\/code> also provide two tags, which you\n  can use to tag the headline:<\/p>\n<dl><dt><code>:nobrain:<\/code><\/dt><dd>This tag excludes the headline, and its subheadings, from your\n    <code>org-brain<\/code> entries. You can change the tag name by modifying\n    <code>org-brain-exclude-tree-tag<\/code>.<\/dd>\n  <dt><code>:childless:<\/code><\/dt><dd>This tag does not exclude the headline, but it excludes the\n    subheadings. You can change the tag name by modifying\n    <code>org-brain-exclude-children-tag<\/code>. Works on file entries.<\/dd>\n<\/dl><p>The following tags modifies the kind of information that is shown when an entry\n  is visualized:<\/p>\n<dl><dt><code>:notext:<\/code><\/dt><dd>Do not show the entry\u2019s text in <code>org-brain-visualize<\/code>. You can\n    change the tag name by modifying <code>org-brain-exclude-text-tag<\/code>.<\/dd>\n  <dt><code>:resourceless:<\/code><\/dt><dd>Do not show the entry\u2019s resources in\n    <code>org-brain-visualize<\/code>. You can change the tag name by modifying\n    <code>org-brain-exclude-resources-tag<\/code>.<\/dd>\n  <dt><code>:showchildren:<\/code><\/dt><dd>By default local child entries aren\u2019t shown as text. By\n    setting this tag the entry get the entire subtree as text. You can change\n    the tag name by modifying <code>org-brain-show-children-tag<\/code>. Works on file\n    entries.<\/dd>\n<\/dl><h2><a id=\"user-content-entry-descriptions\" class=\"anchor\" aria-hidden=\"true\" href=\"#entry-descriptions\"><\/a>Entry descriptions<\/h2>\n<p>You may add a <i>description<\/i> to an entry, which should be a short summary of what the entry is about. The description will be shown in <code>org-brain-visualize<\/code> as mouse tooltips for links, and also by <code>eldoc-mode<\/code> if point is at a link. Add a description by using a <code>#+BEGIN_description<\/code> block. Here\u2019s an example:<\/p>\n<pre lang=\"org\">* org-brain.el\n  :PROPERTIES:\n  :ID:       7cf9563e-24e7-4823-8742-64b7fcc5cafb\n  :END:\n  #+BEGIN_description\n  Package for concept mapping using org-mode.\n  #+END_description\n  ...\n<\/pre>\n<p>At the moment there\u2019s no way to add\/remove\/change descriptions directly from the <code>org-brain-visualize<\/code> interface, sorry!<\/p>\n<h2><a id=\"user-content-having-multiple-brains\" class=\"anchor\" aria-hidden=\"true\" href=\"#having-multiple-brains\"><\/a>Having multiple brains<\/h2>\n<p>You can have multiple brains simply by having more than one brain folder. In this way, each folder becomes a separate brain. You can switch between these using <code>M-x org-brain-switch-brain<\/code>.<\/p>\n<h2><a id=\"user-content-take-note\" class=\"anchor\" aria-hidden=\"true\" href=\"#take-note\"><\/a>Take note!<\/h2>\n<p><code>org-brain<\/code> creates and uses several headline properties in the <code>PROPERTIES<\/code>\n  drawer of <code>org-mode<\/code> headlines:<\/p>\n<ul><li><code>BRAIN_PARENTS<\/code><\/li>\n  <li><code>BRAIN_CHILDREN<\/code><\/li>\n  <li><code>BRAIN_FRIENDS<\/code><\/li>\n  <li><code>ID<\/code><\/li>\n<\/ul><p>These properties are also mirrored as file keywords at the top of file entries,\n  for instance <code>#+BRAIN_CHILDREN: 00c0f06c-9bd4-4c31-aed0-15bb3361d9a2<\/code>.<\/p>\n<p>These properties\/keywords are <i>not meant to be manipulated directly<\/i>! If you\n  want to remove these properties, use the corresponding command instead\n  (<code>org-brain-remove-child<\/code> or similar).<\/p>\n<p>You might also see that <code>org-brain<\/code> inserts a <code>RESOURCES<\/code> drawer. It is okay to\n  modify this drawer manually.<\/p>\n<h2><a id=\"user-content-org-brain-is-slow\" class=\"anchor\" aria-hidden=\"true\" href=\"#org-brain-is-slow\"><\/a><code>org-brain<\/code> is slow!<\/h2>\n<p>If you feel that <code>org-brain<\/code> is slow while indexing your entries (for instance when running <code>M-x org-brain-visualize<\/code>) you can customize <code>org-brain-file-entries-use-title<\/code>, and set it to <code>nil<\/code>. This will display file names when indexing, instead of the file entry\u2019s title, which is faster.<\/p>\n<p><i>This is only relevant if you\u2019ve been using org-brain before version 0.4<\/i><\/p>\n<p>As of version 0.4 (June 2017) <code>org-brain<\/code> has been rewritten, in order to\n  increase performance and add more options. Because of this, older setups are\n  considered obsolete. Prior to 0.4 only files were considered entries, but now\n  also headlines with an <code>ID<\/code> property are included as entries. Prior to 0.4\n  <code>org-brain<\/code> was using the <code>brain:<\/code> link and <code>#+BRAIN_PINNED:<\/code> file keyword to\n  connect files, which was slow due to the need of searching all files for links.\n  In version 0.4 <code>org-brain<\/code> uses a combination of headline properties, file\n  keywords, <code>org-id<\/code>, and a data file (<code>org-brain-data-file<\/code>).<\/p>\n<p>No data in old configurations should be lost, but you\u2019ll have to update the\n  connections between entries. This can be done by using <code>M-x\n  org-brain-create-relationships-from-links<\/code>, but please backup your <code>org-brain<\/code>\n  directory first.<\/p>\n<p>It is still possible to add children to an entry by using the <code>brain-child:<\/code> link, but\n  only if the link is inserted with <code>org-insert-link<\/code> (bound to <code>C-c C-l<\/code> in\n  <code>org-mode<\/code> by default). Linking to specific headlines in a file, via\n  <code>brain:filename::*Headline<\/code> is <b>deprecated<\/b> and will no longer work, instead you\n  can convert the headline to an entry and link directly to that.<\/p>\n<p>There\u2019s some missing functionality in <code>org-brain<\/code>, which you may find useful.\n  However there are other packages which might improve your <code>org-brain<\/code>\n  experience. Below are some suggestions (feel free to create an issue or send a\n  pull request if you have more examples), all of them should be available on\n  MELPA.<\/p>\n<blockquote>\n  <p>A simple command that takes a URL from the clipboard and inserts an org-mode link with a title of a page found by the URL into the current buffer.<\/p>\n<\/blockquote>\n<p>Here\u2019s a command which uses <code>org-cliplink<\/code> to add a link from the clipboard as an <code>org-brain<\/code> resource. It guesses the description from the URL title. Here I\u2019ve bound it to <code>L<\/code> in <code>org-brain-visualize<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun org-brain-cliplink-resource ()\n  \"Add a URL from the clipboard as an org-brain resource.\nSuggest the URL title as a description for resource.\"\n  (interactive)\n  (let ((url (org-cliplink-clipboard-content)))\n    (org-brain-add-resource\n     url\n     (org-cliplink-retrieve-title-synchronously url)\n     t)))\n(define-key org-brain-visualize-mode-map (kbd \"L\") #'org-brain-cliplink-resource)<\/pre><\/div>\n<blockquote>\n  <p>link-hint.el is inspired by the link hinting functionality in vim-like browsers\n    and browser plugins such as pentadactyl. It provides commands for using avy to\n    open or copy \u201clinks.\u201d<\/p>\n<\/blockquote>\n<p>After installing <code>link-hint<\/code> you could bind <code>link-hint-open-link<\/code> to a key, and\n  use it in <code>org-brain-visualize-mode<\/code>. If you only want to use <code>link-hint<\/code> in\n  <code>org-brain-visualize-mode<\/code>, you could add the following to your init-file:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(define-key org-brain-visualize-mode-map (kbd \"C-l\") #'link-hint-open-link)<\/pre><\/div>\n<blockquote>\n  <p>Converts simple ASCII art line drawings in the region of the current buffer to\n    Unicode.<\/p>\n<\/blockquote>\n<p><code>ascii-art-to-unicode<\/code> is useful if you want <code>org-brain-visualize-mode<\/code> to look\n  a bit nicer. After installing, add the following to your init-file:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun aa2u-buffer ()\n  (aa2u (point-min) (point-max)))\n(add-hook 'org-brain-after-visualize-hook #'aa2u-buffer)<\/pre><\/div>\n<blockquote>\n  <p>A utility package to collect various Icon Fonts and propertize them within Emacs.<\/p>\n<\/blockquote>\n<p>After installing <code>all-the-icons<\/code> you could decorate the resources in <code>org-brain<\/code>, by using\n  <code>org-brain-after-resource-button-functions<\/code>. Here\u2019s a small example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun org-brain-insert-resource-icon (link)\n  \"Insert an icon, based on content of org-mode LINK.\"\n  (insert (format \"%s \"\n                  (cond ((string-prefix-p \"http\" link)\n                         (cond ((string-match \"wikipedia\\\\.org\" link)\n                                (all-the-icons-faicon \"wikipedia-w\"))\n                               ((string-match \"github\\\\.com\" link)\n                                (all-the-icons-octicon \"mark-github\"))\n                               ((string-match \"vimeo\\\\.com\" link)\n                                (all-the-icons-faicon \"vimeo\"))\n                               ((string-match \"youtube\\\\.com\" link)\n                                (all-the-icons-faicon \"youtube\"))\n                               (t\n                                (all-the-icons-faicon \"globe\"))))\n                        ((string-prefix-p \"brain:\" link)\n                         (all-the-icons-fileicon \"brain\"))\n                        (t\n                         (all-the-icons-icon-for-file link))))))\n(add-hook 'org-brain-after-resource-button-functions #'org-brain-insert-resource-icon)<\/pre><\/div>\n<blockquote>\n  <p>An Emacs mode for quickly browsing, filtering, and editing directories of plain\n    text notes, inspired by Notational Velocity.<\/p>\n<\/blockquote>\n<p>After installing <code>deft<\/code>, you can add the function below to your init-file.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun org-brain-deft ()\n  \"Use `deft' for files in `org-brain-path'.\"\n  (interactive)\n  (let ((deft-directory org-brain-path)\n        (deft-recursive t)\n        (deft-extensions '(\"org\")))\n    (deft)))<\/pre><\/div>\n<blockquote>\n  <p>It searches both headings and contents of entries in Org buffers, and it\n    displays entries that match all search terms, whether the terms appear in the\n    heading, the contents, or both.<\/p>\n<\/blockquote>\n<p>After installing <code>helm-org-rifle<\/code>, you can add the function below to your\n  init-file.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun helm-org-rifle-brain ()\n  \"Rifle files in `org-brain-path'.\"\n  (interactive)\n  (helm-org-rifle-directories (list org-brain-path)))<\/pre><\/div>\n<blockquote>\n  <p>org-board is a bookmarking and web archival system for Emacs Org mode, building\n    on ideas from Pinboard. It archives your bookmarks so that you can access them\n    even when you\u2019re not online, or when the site hosting them goes down.<\/p>\n<\/blockquote>\n<p>The Emacs Wiki has an article about <a href=\"https:\/\/www.emacswiki.org\/emacs\/WikiModes\" rel=\"nofollow\">wiki modes in Emacs<\/a>.<\/p>\n<blockquote>\n  <p>Org-wiki is a org-mode extension that provides tools to manage and build\n    personal wiki or desktop wiki where each wiki page is a org-mode file.<\/p>\n<\/blockquote>\n<p>Emacs freex is a Pymacs\/SQLite\/Elisp system that implements a transcluding wiki.\n  Emacs-freex is not compatible at this time with org-mode. Despite this,\n  emacs-freex is an impressive system for maintaining a wiki. Further, because the\n  data is stored both in files on disk and in an SQLite database, it opens the\n  possibility for implementing something like <code>org-brain<\/code>\u2019s visualize interface\n  (ala TheBrain\u2019s \u201cplex\u201d) by talking with SQLite, via Pymacs, to return the\n  relationships between nodes. This would consistute a lot of work to implement\n  but would be very impressive. If someone was to also add LaTeX rendering inside\n  <code>emacs-freex<\/code> <code>nuggets<\/code> also, those two additional features would make\n  <code>emacs-freex<\/code> more compelling. As it is, practically speaking, you may think of\n  <code>org-brain<\/code> as implementing many of the features of <code>emacs-freex<\/code>, but with all\n  of <code>org-mode<\/code>\u2019s goodness included.<\/p>","created_at":"2018-11-07T16:42:35+0000","updated_at":"2018-11-07T16:42:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":18,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/3976377?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":254,"title":"Squashing Commits with Magit","url":"http:\/\/howardism.org\/Technical\/Emacs\/magit-squashing.html","content":"<header>\n<\/header><p>\nWhile <a href=\"http:\/\/howardism.org\/Technical\/Git\/squashing-commits.html\">squashing Git commits<\/a> isn\u2019t too difficult, with Magit, I don\u2019t\nwant to use Git on the command line. However, manipulating the commit\nhistory in Emacs isn\u2019t as straight-forward as other Magit activities,\nso I thought I would share my notes.\n<\/p><p>\nTo explain this, I figure I need a story. While in Magit, hit the <b><code>l<\/code><\/b>\n(lowercase L) twice to bring up the log history summary:\n<\/p><figure><img src=\"http:\/\/howardism.org\/Technical\/Emacs\/magit-rebase-1.png\" alt=\"magit-rebase-1.png\" \/><\/figure><p>\nNotice the line the cursor is on was my \u2018presentation\u2019 commit, but I\nforgot to commit the images, so the latest commit on top contains\nthose. Ideally, they should be one commit.\n<\/p><p>\nGame on.\n<\/p>","created_at":"2018-11-07T16:42:34+0000","updated_at":"2018-11-07T16:42:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"howardism.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":255,"title":"Rezi - Developing Smarter Resumes","url":"https:\/\/rezi.io\/","content":"Rezi - Developing Smarter Resumes<section class=\"your_mom\"><div class=\"container nopadding\"><div class=\"container\"><div class=\"row top80\"><div class=\"col-lg-6\" data-aos=\"fade-right\"><section id=\"start\"><p>Writing a resume is hard. So Rezi has created ways to build a resume to pass Applicant Tracking Systems quickly and easily.<\/p><p><a href=\"#start\" class=\"btn btn rezi radius6 spacing15 uppercase size50 font16 padding_sides15 right40\">All Services<\/a>\n\t\t\t\t\t<\/p><\/section><\/div><\/div><\/div><\/div><\/section><section id=\"build\"><section class=\"feature_19 padding_top120 padding_bottom40\"><div class=\"slider\"><div><div class=\"container nopadding max_width970 min_height660 flex align_center slide\"><div class=\"left400 inner\"><h2 class=\"font40 bold spacing10 tk-brandon-grotesque capitalize dark_blue\">Rezi Instant is The only ATS optimized  resume generator<\/h2><p>Applicant Tracking Systems are used by 90% of large corporations to manage submitted applications. You must make sure your resume is optimized to pass them.<\/p><a href=\"https:\/\/app.rezi.io\" class=\"top40 btn rezi radius6 spacing15 uppercase size50 font16 padding_sides15\">Start with Rezi Instant<\/a><\/div><img srcset=\"\/i\/rest2.png 1x, \/i\/rest2.png 2x\" src=\"https:\/\/rezi.io\/i\/rest2.png\" class=\"max_height_full absolute bg\" alt=\"\" \/><\/div><\/div><div><div class=\"container nopadding max_width970 min_height660 flex align_center slide\"><div class=\"left400 inner\"><h2 class=\"font40 bold spacing10 tk-brandon-grotesque capitalize dark_blue\">Ask us anything. Talk with a Rezi Expert as you go<\/h2><p>You'll never be alone, we'll help you identify your talents, build them into strengths, and professional successes. Ask us anything.<\/p><a href=\"https:\/\/app.rezi.io\" class=\"top40 btn rezi radius6 spacing15 uppercase size50 font16 padding_sides15\">Start with Rezi Instant<\/a><\/div><img srcset=\"\/i\/rest1.png 1x, \/i\/rest1.png 2x\" src=\"https:\/\/rezi.io\/i\/rest1.png\" class=\"max_height_full absolute bg\" alt=\"\" \/><\/div><\/div><\/div><\/section><section id=\"products\"><section class=\"call_to_action_14 padding_top40 padding_bottom75\"><div class=\"container nopadding max_width970 dark_gray text_center\"><h2 class=\"font40 bold spacing10 tk-brandon-grotesque capitalize dark_blue\">Rezi ATS Optimized  Resume Templates<\/h2><p>Rezi templates are a flexible way to improve your resume\u2014and get interviewed.<\/p><\/div><\/section><section id=\"review\"><section class=\"feature_14 padding_top90 padding_bottom20\"><div class=\"container nopadding\"><div class=\"max_width770 margin_auto text_center\"><h2 class=\"font40 bold spacing10 tk-brandon-grotesque capitalize dark_blue\">Rezi Format &amp; Content Review<\/h2><p>Rezi gives you unbiased, constructive feedback and reformats your resume based on the ATS optimized Rezi OMEGA format.<\/p><\/div><div class=\"row\"><div class=\"col-md-3 bottom50 text_right block_left\"><p>For everyone from students to professionals to exectives, Rezi is ready to turn your existing resume into your greatest work.<\/p><a href=\"https:\/\/rezi.io\/resume-review\/\" class=\"top40 btn rezi radius6 spacing15 uppercase size50 font16 padding_sides30\">See all Options<\/a><\/div><div class=\"col-md-6 bottom50 text_center block_center\" data-aos=\"fade-up\"><img srcset=\"\/i\/omega1.png 1x, \/i\/omega1.png 2x\" src=\"https:\/\/rezi.io\/i\/omega1.png\" class=\"max_width_full radius6 shadow\" alt=\"\" \/><\/div><\/div><\/div><\/section><section class=\"content_31 padding_top40\"><div class=\"container nopadding\"><h2 class=\"font40 bold spacing10 tk-brandon-grotesque dark_blue capitalize text_center\">Rezi worked for them.  It will work for you<\/h2><\/div><\/section><section class=\"content_36 padding_top130 padding_bottom100\"><div class=\"container nopadding\"><div class=\"max_width770 margin_auto text_center\"><h2 class=\"font40 bold spacing10 tk-brandon-grotesque capitalize dark_blue\">Rezi is honored to serve  our partners<\/h2><p>We provide resume solutions to inspiring institions that care about your career.<\/p><\/div><\/div><\/section><\/section><\/section><\/section>","created_at":"2018-11-07T16:42:34+0000","updated_at":"2018-11-07T16:42:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"rezi.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":252,"title":"Brandmark Logo Maker - the most advanced AI logo design tool","url":"https:\/\/brandmark.io\/","content":"Brandmark Logo Maker - the most advanced AI logo design tool\n\t<section id=\"main\"><div class=\"hero-wrapper\"><div class=\"container\"><div id=\"logo-container\"><div class=\"logo\"><img id=\"logo_brand\" src=\"https:\/\/brandmark.io\/svg-brand-2.svg\" width=\"455\" height=\"255\" alt=\"Brand\" \/><img id=\"logo_mark\" src=\"https:\/\/brandmark.io\/svg-mark-2.svg\" width=\"455\" height=\"255\" alt=\"Mark\" \/><\/div><div class=\"welcome\"><h4>complete with business card designs, social media graphics, app icons and more<\/h4><p><a href=\"https:\/\/app.brandmark.io\/v2\/\" target=\"_blank\" class=\"btn btn-primary\">Create your logo<\/a><\/p><\/div><\/div><\/div><\/div><\/section><section id=\"featured\"><section id=\"samples\"><div class=\"container\"><div class=\"row\"><div class=\"col-md-12\"><img src=\"https:\/\/brandmark.io\/img\/sample.jpg\" width=\"1600\" height=\"870\" class=\"big\" alt=\"real life logo\" \/><\/div><div class=\"col-md-5 col-md-offset-1 pull-up\"><h2>Latest customer creations<\/h2><p>See why over 108,704 users have created a logo with Brandmark.io<\/p><\/div><\/div><\/div><\/section><section id=\"bookend\"><div class=\"hero-wrapper\"><div class=\"container\"><div class=\"yellow-box\"><div class=\"row\"><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/design.svg\" alt=\"design icon\" \/><h3>No more designer anxiety<\/h3><p>Only pay when you're happy. Don't spend time and money on uncertain results.<\/p><\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/time.svg\" alt=\"time icon\" \/><h3>On time and on budget<\/h3>Get your logo and design assets instantly, not in weeks or months with a design agency.<\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/copyright.svg\" alt=\"copyright icon\" \/><h3>Full copyright<\/h3>All packages include full rights for use on your website, print media, products and more.<\/div><\/div><\/div><div class=\"row\"><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/coin.svg\" alt=\"coin icon\" \/><h3>Affordable one-time purchase<\/h3><p>Pay once, gain access to our branding tools forever. No monthly charges.<\/p><\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/infinite.svg\" alt=\"infinite icon\" \/><h3>Unlimited revisions<\/h3>Modify your logo at any time, even after purchase.<\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/support.svg\" alt=\"support icon\" \/><h3>Your personal design team<\/h3>Need a tweak to your design? We'll do it for you, absolutely free.<\/div><\/div><\/div><p><\/p><h2>Create your unique logo design, try it free!<\/h2>\t\t\t\n\t\t\t\t\t<a href=\"https:\/\/app.brandmark.io\/v2\/\" target=\"_blank\" class=\"btn btn-primary\">\n\t\t\t\t\t\tDesign your logo\n\t\t\t\t\t<\/a><\/div><\/div><\/div><\/section><\/section>","created_at":"2018-11-07T16:42:11+0000","updated_at":"2018-11-07T16:42:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"brandmark.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":253,"title":"Quickly find Emacs Lisp sources \u00b7 Emacs Redux","url":"http:\/\/emacsredux.com\/blog\/2014\/06\/18\/quickly-find-emacs-lisp-sources\/","content":"<p>One thing Emacs users deal all the time is looking for the source of\nvarious libraries and function or variable definitions.  Typically this\ninvolves the use of commands like <code class=\"highlighter-rouge\">C-h f<\/code> (<code class=\"highlighter-rouge\">describe-function<\/code>), <code class=\"highlighter-rouge\">C-h v<\/code>\n(<code class=\"highlighter-rouge\">describe-variable<\/code>) and <code class=\"highlighter-rouge\">C-h k<\/code> (<code class=\"highlighter-rouge\">describe-key<\/code>) and jumping to the source from the help buffer those commands produce.\nLet\u2019s see if we can improve upon this workflow.<\/p><h3 id=\"find-func\">find-func<\/h3><p>Enter the built-in Emacs library <code class=\"highlighter-rouge\">find-func<\/code>. It provides a myriad of useful commands to help us quickly locate sources.\nLet\u2019s take a brief look at some of them.<\/p><ul><li>You can find the source of a library with <code class=\"highlighter-rouge\">M-x find-library<\/code>. If you want to find the source of <code class=\"highlighter-rouge\">ido<\/code> you can do so like this:<\/li>\n<\/ul><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-library RET ido RET\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-l<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-l\") 'find-library)\n<\/pre><\/div><\/div><ul><li>You can find the source of a function definition with <code class=\"highlighter-rouge\">M-x find-function<\/code>. Here\u2019s an example:<\/li>\n<\/ul><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-function RET find-function RET\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-f<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-f\") 'find-function)\n<\/pre><\/div><\/div><p>We can actually do one better - we can directly jump to a command\ndefinition using a keybinding of the command with <code class=\"highlighter-rouge\">M-x\nfind-function-on-key<\/code>. Here\u2019s how we can find the source the command\nbound to <code class=\"highlighter-rouge\">C-a<\/code> (<code class=\"highlighter-rouge\">beginning-of-line<\/code>):<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-function-on-key RET C-a\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-k<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-k\") 'find-function-on-key)\n<\/pre><\/div><\/div><ul><li>You can find the source of a variable definition with <code class=\"highlighter-rouge\">M-x find-variable<\/code>. Here\u2019s an example:<\/li>\n<\/ul><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-variable RET large-file-warning-threshold RET\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-v<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-v\") 'find-variable)\n<\/pre><\/div><\/div><hr \/><p>The library provides other useful commands as well - like\n<code class=\"highlighter-rouge\">find-function-at-point<\/code> and <code class=\"highlighter-rouge\">find-variable-at-point<\/code>.<\/p><p>If you don\u2019t like the keybindings I suggested you can use\n<code class=\"highlighter-rouge\">find-function-setup-keys<\/code> instead.  This will give you keybindings\nlike <code class=\"highlighter-rouge\">C-x F<\/code>, <code class=\"highlighter-rouge\">C-x V<\/code>, <code class=\"highlighter-rouge\">C-f K<\/code> (plus a few extra for commands like\n<code class=\"highlighter-rouge\">find-funtion-other-window<\/code>).<\/p><p>Another really cool way to browse Elisp sources (and documentation) is the third-party package\n<a href=\"https:\/\/github.com\/purcell\/elisp-slime-nav\">elisp-slime-nav<\/a>. Assuming you\u2019ve installed it already you can\nenabled it like this:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(require 'elisp-slime-nav)\n(dolist (hook '(emacs-lisp-mode-hook ielm-mode-hook))\n(add-hook hook 'elisp-slime-nav-mode))\n<\/pre><\/div><\/div><p>Once this is done you\u2019ll be able to jump to the source of the Emacs\nLisp object at point (function or variable) with <code class=\"highlighter-rouge\">M-.<\/code> (as in SLIME\nand CIDER for Common Lisp and Clojure respectively) and jump back with\n<code class=\"highlighter-rouge\">M-,<\/code>. You can also see the description of the object at point using\n<code class=\"highlighter-rouge\">C-c C-d<\/code> (or <code class=\"highlighter-rouge\">C-c C-d d<\/code>).<\/p><p>That\u2019s all for today, folks!<\/p><p><strong>P.S.<\/strong><\/p><p>All the suggested keybindings are present out-of-the-box in\n<a href=\"https:\/\/github.com\/bbatsov\/prelude\">Prelude<\/a>. <code class=\"highlighter-rouge\">elisp-slime-nav<\/code> is also\nenabled out-of-the box in Prelude.<\/p>","created_at":"2018-11-07T16:42:11+0000","updated_at":"2018-11-07T16:42:11+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"emacsredux.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":251,"title":"Best Fallout 4 mods","url":"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/","content":"<div><main><p>Now that <a href=\"https:\/\/www.rockpapershotgun.com\/game\/fallout-4\/\">Fallout 4<\/a> is in its third year and the DLC has dried up, is there any point in returning to the Commonwealth Wasteland? Yes! There are loads, largely thanks to the still-growing list of mods, overhauls and user tweaks. Here, I\u2019ve gathered over 50 of my favourites, ranging from weird weapons to wild weather.<\/p>\n<p>Before we start, a couple of things to remember: some mods will require some or all of the DLC expansions, as well as additional mods, while others don\u2019t play nicely with each other. The mod descriptions on <a href=\"https:\/\/www.nexusmods.com\/fallout4\">Nexus<\/a> will usually tell you, so keep an eye out.<\/p>\n<p>Fallout 4 mods come in all shapes and sizes and so we\u2019ve split this guide into sections, so you can hop directly to the kind of mod you\u2019re looking for.<\/p>\n<p><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/2\/\">How to install Fallout 4 mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/3\/\">Fallout 4 graphics mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/4\/\">Fallout 4 settlement mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/5\/\">Fallout 4 weapon and armour mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/6\/\">Fallout 4 companion mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/7\/\">Fallout 4 quests and overhaul mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/8\/\">Fallout 4 UI mods<\/a><\/p>\n\t\t\t\t\t<p>Page 1<a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/2\/\">Page 2<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/3\/\">Page 3<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/4\/\">Page 4<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/5\/\">Page 5<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/6\/\">Page 6<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/7\/\">Page 7<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/8\/\">Page 8<\/a><\/p>\t\t\t\t<\/main><\/div><div class=\"container-690\"><p class=\"disclaimer\">\n\t\t\t\t\tIf you click our links to online stores and make a purchase we may receive a few pennies. Find more information <a href=\"https:\/\/www.rockpapershotgun.com\/affiliate-linking-policy\/\">here<\/a>.\t\t\t\t<\/p><a href=\"https:\/\/www.rockpapershotgun.com\/support-us?utm_source=article-promo&amp;utm_medium=web&amp;utm_campaign=supporter\" class=\"support-us-promo\" rel=\"nofollow\">\n\t\t\n\t\t\t<img src=\"https:\/\/www.rockpapershotgun.com\/wp-content\/themes\/rps-2018\/images\/logo-inverted.svg\" alt=\"Rock Paper Shotgun\" \/>\n\t\t\n\t\t\n\t\t\t\n\t\t\t\tSupport RPS and get an ad-free site, extra articles, and free stuff!\t\t\t\n\t\t\tTell me more\n\t\t\n\t\t\n\t<\/a><\/div>","created_at":"2018-11-07T16:42:10+0000","updated_at":"2018-11-07T16:42:10+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/assets.rockpapershotgun.com\/images\/2018\/10\/Homemaker-1212x682.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":250,"title":"Presenting the Eshell","url":"http:\/\/howardism.org\/Technical\/Emacs\/eshell-present.html","content":"<p>\nDemonstrated the iPython approach by entering the following into an\n<code>ipython<\/code> REPL:\n<\/p><pre class=\"example\">\nPython 2.7.12 (default, Nov 19 2016, 06:48:10)\nType \"copyright\", \"credits\" or \"license\" for more information.\nIPython 2.4.1 -- An enhanced Interactive Python.\n?         -&gt; Introduction and overview of IPython's features.\n%quickref -&gt; Quick reference.\nhelp      -&gt; Python's own help system.\nobject?   -&gt; Details about 'object', use 'object??' for extra details.\nIn [1]: 2 ** 60\nOut[1]: 1152921504606846976\nIn [2]: cd \/tmp\/testing\n\/tmp\/testing\nIn [3]: ls\npi.py  pi.rb  README  src\/  tests\/\nIn [4]: cat pi.py\nimport math\nprint(math.pi)\nprint(math.cos(math.pi))\nIn [5]: chmod a+x pi.py\n  File \"&lt;ipython-input-5-b9e30b9b8e31&gt;\", line 1\n    chmod a+x pi.py\n          ^\nSyntaxError: invalid syntax\nIn [6]: run pi\n3.14159265359\n-1.0\nIn [7]: system ruby pi.rb\nOut[7]: ['-0.9999999999964793', '-3.141592653589793']\nIn [8]: cat README\nCiao, all you cool cats.\nOh, and hey to all my dawgs.\nIn [9]: def cat(arg=None):\n   ...:     return 'Meow!'\n   ...:\nIn [10]: cat\nOut[10]: &lt;function __main__.cat&gt;\nIn [11]: cat()\nOut[11]: 'Meow!'\nIn [12]: cat README\n  File \"&lt;ipython-input-12-01ccd53eccb4&gt;\", line 1\n    cat README\n             ^\nSyntaxError: invalid syntax\n<\/pre><p>\nIn answer to your question, I haven\u2019t looked to see why we have an\narray when calling the <code>system<\/code> function.\n<\/p><p>\nIf you want to see what <i>shell-like<\/i> functions iPython has, type <code>%<\/code>\nand hit the Tab key, which shows something like:\n<\/p><pre class=\"example\">\nIn [13]: %\nDisplay all 122 possibilities? (y or n)\n%%!                      %doctest_mode            %pfile\n%%HTML                   %ed                      %pinfo\n%%SVG                    %edit                    %pinfo2\n%%bash                   %env                     %popd\n%%capture                %gui                     %pprint\n%%debug                  %hist                    %precision\n%%file                   %history                 %profile\n%%html                   %install_default_config  %prun\n%%javascript             %install_ext             %psearch\n%%latex                  %install_profiles        %psource\n%%perl                   %killbgscripts           %pushd\n%%prun                   %ldir                    %pwd\n%%pypy                   %less                    %pycat\n%%python                 %lf                      %pylab\n%%python2                %lk                      %quickref\n%%python3                %ll                      %recall\n%%ruby                   %load                    %rehashx\n%%script                 %load_ext                %reload_ext\n%%sh                     %loadpy                  %rep\n%%svg                    %logoff                  %rerun\n%%sx                     %logon                   %reset\n%%system                 %logstart                %reset_selective\n%%time                   %logstate                %rm\n%%timeit                 %logstop                 %rmdir\n%%writefile              %ls                      %run\n%alias                   %lsmagic                 %save\n%alias_magic             %lx                      %sc\n%autocall                %macro                   %store\n%autoindent              %magic                   %sx\n%automagic               %man                     %system\n%bookmark                %matplotlib              %tb\n%cat                     %mkdir                   %time\n%cd                      %more                    %timeit\n%clear                   %mv                      %unalias\n%colors                  %notebook                %unload_ext\n%config                  %page                    %who\n%cp                      %paste                   %who_ls\n%cpaste                  %pastebin                %whos\n%debug                   %pdb                     %xdel\n%dhist                   %pdef                    %xmode\n%dirs                    %pdoc\n<\/pre>","created_at":"2018-11-07T16:40:35+0000","updated_at":"2018-11-07T16:40:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"howardism.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":244,"title":"org2blog\/org2blog","url":"https:\/\/github.com\/org2blog\/org2blog\/blob\/master\/README.org","content":"<p><a href=\"https:\/\/gitter.im\/punchagan\/org2blog?utm_source=badge&amp;utm_medium=badge&amp;utm_campaign=pr-badge&amp;utm_content=badge\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/da2edb525cde1455a622c58c0effc3a90b9a181c\/68747470733a2f2f6261646765732e6769747465722e696d2f4a6f696e253230436861742e737667\" alt=\"https:\/\/badges.gitter.im\/Join%20Chat.svg\" data-canonical-src=\"https:\/\/badges.gitter.im\/Join%20Chat.svg\" \/><\/a> <a href=\"https:\/\/travis-ci.org\/punchagan\/org2blog\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/0999a3683a308c876f870e97f198e299c8fe8262\/68747470733a2f2f7472617669732d63692e6f72672f70756e63686167616e2f6f726732626c6f672e737667\" alt=\"https:\/\/travis-ci.org\/punchagan\/org2blog.svg\" data-canonical-src=\"https:\/\/travis-ci.org\/punchagan\/org2blog.svg\" \/><\/a><\/p>\n<p>org2blog\/wp is a tool to publish directly from Emacs\u2019s org-mode to\n  <a href=\"http:\/\/wordpress.org\/\" rel=\"nofollow\">WordPress<\/a> blogs.  org2blog\/wp is one of the two packages named\n  org2blog.  The other is <a href=\"http:\/\/repo.or.cz\/r\/org2blog.git\/\" rel=\"nofollow\">org2blog\/atom<\/a> by <a href=\"http:\/\/tehom-blog.blogspot.com\/\" rel=\"nofollow\">Tom Breton<\/a>.<\/p>\n<p>org2blog\/wp should, <i>ideally<\/i>, work with any platform that supports\n  xml-rpc but some of WordPress specific features may not work.<\/p>\n<p>org2blog was inspired by and based on a <a href=\"http:\/\/paste.lisp.org\/display\/69993\" rel=\"nofollow\">WordPress posting client for\n  muse-mode<\/a> written by <a href=\"http:\/\/www.emacswiki.org\/emacs\/AshishShukla\" rel=\"nofollow\">Ashish Shukla<\/a>.<\/p>\n<p>org2blog is licensed under GPLv3<\/p>\n<p>Download org2blog from GitHub<\/p>\n<pre>git clone http:\/\/github.com\/punchagan\/org2blog.git\n<\/pre>\n<p>Add the org2blog directory to your load path and then add<\/p>\n<pre>(setq load-path (cons \"~\/.emacs.d\/org2blog\/\" load-path))\n(require 'org2blog-autoloads)\n<\/pre>\n<p>to your dot emacs.<\/p>\n<h3><a id=\"user-content-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>Dependencies<\/h3>\n<ol><li>org2blog depends on\n    <ul><li><i>xml-rpc<\/i> available at <a href=\"https:\/\/github.com\/hexmode\/xml-rpc-el\">GitHub<\/a><\/li>\n      <li><i>metaweblog.el<\/i> available <a href=\"https:\/\/github.com\/punchagan\/metaweblog\">here<\/a><\/li>\n      <li><i>htmlize<\/i> available <a href=\"https:\/\/github.com\/hniksic\/emacs-htmlize\">here<\/a><\/li>\n    <\/ul><p>Add them to your <i>load-path<\/i> before using org2blog.<\/p>\n  <\/li>\n  <li>Make sure you have org-mode 8.3 or later installed.<\/li>\n<\/ol><h3><a id=\"user-content-unicode-character-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#unicode-character-support\"><\/a>Unicode Character Support<\/h3>\n<ol><li>As of &lt;2017-06-26 Mon&gt; if you use Unicode characters in your posts or\n    pages then you need to use <a href=\"https:\/\/github.com\/grettke\/xml-rpc-el\/blob\/master\/xml-rpc.el\">this<\/a> patched version of <code>xml-rpc<\/code><\/li>\n  <li>Installation steps\n    <ul><li>Back up your local copy of <code>xml-rpc.el<\/code><\/li>\n      <li>Copy <a href=\"https:\/\/raw.githubusercontent.com\/grettke\/xml-rpc-el\/master\/xml-rpc.el\" rel=\"nofollow\">this<\/a> patched version of <code>xml-rpc.el<\/code> to the <code>xml-rpc<\/code> directory<\/li>\n      <li>Delete <code>xml-rpc.elc<\/code><\/li>\n      <li>Restart Emacs<\/li>\n    <\/ul><\/li>\n<\/ol>\n<h3><a id=\"user-content-defining-your-blog-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#defining-your-blog-setup\"><\/a>Defining your blog setup<\/h3>\n<p><b>Note<\/b> - This setup works only for org2blog version &gt;=0.3. For\n  lower versions of org2blog check the README from the commit\n  e37126e<\/p>\n<p>You can (should) tell org2blog where your blogs are and optionally\n  specify different values for some options. This is done by\n  customising the variable <code>org2blog\/wp-blog-alist<\/code>.<\/p>\n<p>For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq org2blog\/wp-blog-alist\n      '((\"wordpress\"\n         :url \"https:\/\/username.wordpress.com\/xmlrpc.php\"\n         :username \"username\"\n         :default-title \"Hello World\"\n         :default-categories (\"org2blog\" \"emacs\")\n         :tags-as-categories nil)\n        (\"my-blog\"\n         :url \"http:\/\/username.server.com\/xmlrpc.php\"\n         :username \"admin\")))\n<\/pre><\/div>\n<p>This defines two blogs identified by the terms <code>wordpress<\/code> and\n  <code>my-blog<\/code>.<\/p>\n<p>The variables in the <code>plist<\/code> set for each blog over-ride the\n  global values of the corresponding variables. <code>url<\/code> and <code>username<\/code>\n  are mandatory variables. Others are optional. If you customize the\n  variable <code>org2blog\/wp-blog-alist<\/code> there is detail information about the\n  meaning of each property.<\/p>\n<p>You can also use <code>.netrc<\/code> or <code>.authinfo<\/code> files to save your usernames and\n  passwords.<\/p>\n<p>Your <code>.netrc<\/code> should look like this<\/p>\n<pre>machine myblog login myusername password myrealpassword\n<\/pre>\n<p>Then, use these details, as shown below.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'auth-source) ;; or nothing if already in the load-path\n(let (credentials)\n  ;; only required if your auth file is not already in the list of auth-sources\n  (add-to-list 'auth-sources \"~\/.netrc\")\n  (setq credentials (auth-source-user-and-password \"myblog\"))\n  (setq org2blog\/wp-blog-alist\n        `((\"my-blog\"\n           :url \"http:\/\/username.server.com\/xmlrpc.php\"\n           :username ,(car credentials)\n           :password ,(cadr credentials)))))\n<\/pre><\/div>\n<p>To browse or customize other variables use the customize group\n  function.<\/p>\n<pre>M-x customize-group org2blog\/wp RET\n<\/pre>\n<h3><a id=\"user-content-logging-in\" class=\"anchor\" aria-hidden=\"true\" href=\"#logging-in\"><\/a>Logging In<\/h3>\n<p>Use <b>M-x org2blog\/wp-login<\/b>.  This function ask you for the name of one\n  of the blogs you have configured in the org2blog\/wp-blog-alist\n  variable described above and will ask for your password. The\n  counterpart of this function is <b>M-x org2blog\/wp-logout<\/b>, which will\n  log you out of the active blog.<\/p>\n<h3><a id=\"user-content-writing-a-new-post\" class=\"anchor\" aria-hidden=\"true\" href=\"#writing-a-new-post\"><\/a>Writing a new post<\/h3>\n<p>Use <b>M-x org2blog\/wp-new-entry<\/b>.<\/p>\n<ul><li>Tags and Categories are comma or space separated values.\n    Completion (for both) is triggered using <b>C-c t<\/b>.  If a new\n    category is used you will be prompted before adding it.<\/li>\n  <li>The excerpt for a post can be written as <code>#+DESCRIPTION<\/code>.<\/li>\n  <li>The Permalink of a post can be set using <code>#+PERMALINK<\/code>.<\/li>\n  <li>The parent for a page can be specified using <code>#+PARENT<\/code>.  <b>C-c\n    t<\/b> again provides completion for the page name.<\/li>\n<\/ul><h3><a id=\"user-content-using-post-templates\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-post-templates\"><\/a>Using post templates<\/h3>\n<p>Custom post templates can be specified by modifying variables\n  <code>org2blog\/wp-buffer-template<\/code> and <code>org2blog\/wp-buffer-format-function<\/code>.<\/p>\n<p>For example:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq org2blog\/wp-buffer-template\n      \"-----------------------\n#+TITLE: %s\n#+DATE: %s\n-----------------------\\n\")\n(defun my-format-function (format-string)\n  (format format-string\n          org2blog\/wp-default-title\n          (format-time-string \"%d-%m-%Y\" (current-time)))))\n(setq org2blog\/wp-buffer-format-function 'my-format-function)\n<\/pre><\/div>\n<p><code>org2blog\/wp-new-entry<\/code> will now create buffers begining with a default\n  title and the current date.<\/p>\n<h3><a id=\"user-content-updating-an-existing-post\" class=\"anchor\" aria-hidden=\"true\" href=\"#updating-an-existing-post\"><\/a>Updating an existing post<\/h3>\n<p>To update a blog post, you can simply edit the content of the org\n  file and republish using <code>org2blog<\/code>.  The <code>POSTID<\/code> saved in the\n  file, ensures that the same blog post is updated, and a new post\n  is not made.<\/p>\n<h3><a id=\"user-content-publishing\" class=\"anchor\" aria-hidden=\"true\" href=\"#publishing\"><\/a>Publishing<\/h3>\n<h5><a id=\"user-content-keybindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#keybindings\"><\/a>Keybindings<\/h5>\n<table><tbody><tr><td>post buffer as draft<\/td><td><b>C-c M-p d<\/b><\/td><td><b>M-x     org2blog\/wp-post-buffer<\/b><\/td><\/tr><tr><td>publish buffer<\/td><td><b>C-c M-p p<\/b><\/td><td><b>C-u M-x org2blog\/wp-post-buffer<\/b><\/td><\/tr><tr><td>post buffer as page draft<\/td><td><b>C-c M-p D<\/b><\/td><td><b>M-x     org2blog\/wp-post-buffer-as-page<\/b><\/td><\/tr><tr><td>publish buffer as page<\/td><td><b>C-c M-p P<\/b><\/td><td><b>C-u M-x org2blog\/wp-post-buffer-as-page<\/b><\/td><\/tr><\/tbody><\/table><h5><a id=\"user-content-after-publishing-access-the-metadata\" class=\"anchor\" aria-hidden=\"true\" href=\"#after-publishing-access-the-metadata\"><\/a>After Publishing Access The Metadata<\/h5>\n<p>Now your post or page exists both in your Org-Mode file on your computer, and\n  also in WordPress itself. That page or post inside of WordPress contains a lot\n  of metadata and you might be interested in some of it. <a href=\"https:\/\/codex.wordpress.org\/XML-RPC_MetaWeblog_API\" rel=\"nofollow\">Here<\/a> is documentation\n  covering all of the fields. You can easily access that data using a hook function.<\/p>\n<p>After publishing your post or page, Org2Blog calls the functions in\n  <code>org2blog\/wp-after-new-post-or-page-functions<\/code> passing them the post or page\n  metadata. Maybe you\u2019ve never seen a hook function like this before because it\n  takes an argument. They are still just plain old functions. Here they need to\n  accept one argument so that Org2Blog can give you that metadata. It is pretty\n  simple.<\/p>\n<p>Here is an example that displays your post or page information in the\n  <code>*Messages*<\/code> buffer:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'org2blog\/wp-after-new-post-or-page-functions (lambda (p) (pp p)))<\/pre><\/div>\n<h3><a id=\"user-content-post-your-existing-org-files-too\" class=\"anchor\" aria-hidden=\"true\" href=\"#post-your-existing-org-files-too\"><\/a>Post your existing org-files too<\/h3>\n<ul><li>Add <code>#+TITLE<\/code>, <code>#+OPTIONS<\/code>, <code>#+TAGS<\/code>, <code>#+CATEGORY<\/code>,\n    <code>#+DESCRIPTION<\/code>, <code>#+DATE<\/code>, to your file.  (All of these are\n    Optional)<\/li>\n  <li>Add a <code>#+POSTID<\/code> to edit an existing post.<\/li>\n  <li>Post buffer using the function names, as above.<\/li>\n<\/ul><h3><a id=\"user-content-posting-a-subtree\" class=\"anchor\" aria-hidden=\"true\" href=\"#posting-a-subtree\"><\/a>Posting a subtree<\/h3>\n<ul><li>Use <b>M-x   org2blog\/wp-post-subtree<\/b>.<\/li>\n  <li>Tags of the subtree will be used for tags.<\/li>\n  <li>Other properties can be set using a property drawer.  The\n    properties use the same names as the buffer variables.<\/li>\n  <li><code>POST_DATE<\/code> sets the date to be used for the post. If it is not\n    present, <code>SCHEDULED<\/code>, <code>DEADLINE<\/code> or any other active or inactive\n    time-stamps are used.<\/li>\n  <li>The headline is used as the title of the post. But, <code>TITLE<\/code> can\n    be used to use a different title.<\/li>\n<\/ul><h3><a id=\"user-content-posting-source-code-blocks\" class=\"anchor\" aria-hidden=\"true\" href=\"#posting-source-code-blocks\"><\/a>Posting source code blocks<\/h3>\n<p>Babel source blocks or example code is automatically posted in\n  <code>&lt;pre&gt;<\/code> tags.  You can also ask <code>org2blog<\/code> to use WordPress\u2019s\n  sourcecode shortcode blocks.<\/p>\n<p>To use this <a href=\"https:\/\/en.support.wordpress.com\/code\/posting-source-code\/\" rel=\"nofollow\">this feature<\/a>, install a shortcode plugin. For example,\n  <a href=\"https:\/\/wordpress.org\/plugins\/syntaxhighlighter\/\" rel=\"nofollow\">SyntaxHighlighter Evolved<\/a> seems to work well.<\/p>\n<p>To use this, you need to set the variable\n  <code>org2blog\/wp-use-sourcecode-shortcode<\/code> to <code>t<\/code>. WordPress\u2019s sourcecode\n  shortcode blocks can be given various <a href=\"http:\/\/en.support.wordpress.com\/code\/posting-source-code\/#configuration-parameters\" rel=\"nofollow\">configuration parameters<\/a>. These can\n  be passed to the exported sourcecode shortcode blocks via an <code>#+ATTR_WP<\/code>\n  line immediately preceding the <code>#+BEGIN_SRC<\/code> line, e.g.<\/p>\n<pre>#+ATTR_WP: :syntaxhl light=true\n<\/pre>\n<h3><a id=\"user-content-delete-an-entry-or-a-page\" class=\"anchor\" aria-hidden=\"true\" href=\"#delete-an-entry-or-a-page\"><\/a>Delete an entry or a page<\/h3>\n<ul><li>If you are in the buffer of the entry\/page, with <code>#+POSTID<\/code>\n    present on the page, use:\n    <p><b>M-x org2blog\/wp-delete-entry<\/b> or <b>M-x  org2blog\/wp-delete-page<\/b><\/p>\n  <\/li>\n  <li>If you want to delete a particular post (whose post-id is known)\n    from any buffer, use\n    <p><b>C-u post-id M-x org2blog\/wp-delete-entry<\/b>.<\/p>\n  <\/li>\n<\/ul><p>Similarly, for a page.<\/p>\n<h3><a id=\"user-content-org2blog-mode-key-bindings-and-auto-loading\" class=\"anchor\" aria-hidden=\"true\" href=\"#org2blog-mode-key-bindings-and-auto-loading\"><\/a>org2blog mode key-bindings and auto loading<\/h3>\n<h5><a id=\"user-content-key-bindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#key-bindings\"><\/a>Key-Bindings<\/h5>\n<ul><li><code>org2blog\/wp-mode<\/code> defines the following key-bindings\n    <table><tbody><tr><td><b>C-c p<\/b><\/td><td>publish buffer<\/td><\/tr><tr><td><b>C-c d<\/b><\/td><td>post as draft<\/td><\/tr><tr><td><b>C-c t<\/b><\/td><td>complete tag or parent page name<\/td><\/tr><\/tbody><\/table><\/li>\n<\/ul><h5><a id=\"user-content-auto-loading-org2blog-mode\" class=\"anchor\" aria-hidden=\"true\" href=\"#auto-loading-org2blog-mode\"><\/a>Auto Loading org2blog Mode<\/h5>\n<ul><li>Add <code>org2blog\/wp-org-mode-hook-fn<\/code> to <code>org-mode-hook<\/code><\/li>\n  <li>So that <code>org2blog\/wp-mode<\/code> is automatically enabled when <code>#+ORG2BLOG:<\/code> is\n    present<\/li>\n  <li>Add code like this somewhere after you load Org-Mode and org2blog\n    <div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'org-mode-hook #'org2blog\/wp-org-mode-hook-fn)\n    <\/pre><\/div>\n  <\/li>\n  <li>Add <code>#+ORG2BLOG:<\/code> to your file\n    <ul><li>When you create a new entry it is automatically created<\/li>\n    <\/ul><\/li>\n<\/ul><h3><a id=\"user-content-uploading-images-or-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#uploading-images-or-files\"><\/a>Uploading Images or Files<\/h3>\n<p>Only in-line images and linked images (or files) with file:\/\/ URLs are\n  supported. They will be uploaded to the media library and the links will\n  be updated. Information about uploaded files is added as a comment to the\n  post itself.<\/p>\n<p>Captions and attributes as <a href=\"http:\/\/orgmode.org\/manual\/Images-in-HTML-export.html\" rel=\"nofollow\">defined<\/a> in org-mode will be preserved,\n  but these attributes are not saved with the image to the library\n  itself.<\/p>\n<h3><a id=\"user-content-dashboard-of-all-posts\" class=\"anchor\" aria-hidden=\"true\" href=\"#dashboard-of-all-posts\"><\/a>\u201cDashboard\u201d of all posts<\/h3>\n<p><code>org2blog<\/code> makes it easy to manage your blog-posts by actually\n  keeping track of all the posts you make from it, in a simple\n  org-table.  By default it is saved in a file <code>.org2blog.org<\/code> in\n  the <code>org-directory<\/code>.  This is meant to be a dashboard of sorts,\n  and is an optional feature that can be turned off.<\/p>\n<h3><a id=\"user-content-mathjax-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#mathjax-support\"><\/a>MathJax Support<\/h3>\n<p>\u201c<a href=\"http:\/\/docs.mathjax.org\/en\/latest\/mathjax.html\" rel=\"nofollow\">MathJax<\/a> is an open-source JavaScript display engine for LaTeX, MathML, and\n  AsciiMath notation that works in all modern browsers.\u201d<\/p>\n<ul><li>Whether you use MathJax with a WordPress plugin, manual inclusion, or any\n    other means you need to be aware of MathJax\u2019s <a href=\"http:\/\/docs.mathjax.org\/en\/latest\/start.html#using-a-content-delivery-network-cdn\" rel=\"nofollow\">CDN<\/a> options\n    <ul><li>If you want predictibility then reference a specific release number. For\n        example, this configuration uses 2.7.1<\/li>\n    <\/ul><\/li>\n  <li>Disable translation to <code>wp-latex<\/code> syntax\n    <ul><li><code>(setq org2blog\/wp-use-wp-latex nil)<\/code><\/li>\n    <\/ul><\/li>\n  <li>Setup the <a href=\"https:\/\/wordpress.org\/plugins\/mathjax-latex\/\" rel=\"nofollow\">MathJax-LaTeX<\/a> plugin\n    <ul><li>Install it<\/li>\n      <li>Configure it\n        <ul><li>Force Load: TRUE<\/li>\n          <li>Default [latex] syntax attribute: Inline<\/li>\n          <li>Use wp-latex syntax? TRUE<\/li>\n          <li>Use MathJax CDN Service? FALSE<\/li>\n          <li>Custom MathJax location? <code>https:\/\/cdnjs.cloudflare.com\/ajax\/libs\/mathjax\/2.7.1\/MathJax.js<\/code><\/li>\n          <li>MathJax Configuration: default<\/li>\n        <\/ul><\/li>\n    <\/ul><\/li>\n  <li>Test it out using these (<a href=\"https:\/\/math.meta.stackexchange.com\/questions\/5020\/mathjax-basic-tutorial-and-quick-reference\" rel=\"nofollow\">and more<\/a>) examples\n    <pre lang=\"org\">- The word LaTeX\n  - $\\LaTeX$\n- Inline\n  - $\\sum_{i=0}^n i^2 = \\frac{(n^2+n)(2n+1)}{6}$\n- Equation\n  - $$\\sum_{i=0}^n i^2 = \\frac{(n^2+n)(2n+1)}{6}$$\n    <\/pre>\n  <\/li>\n  <li>Test it locally using HTML export settings\n    <pre>#+HTML_HEAD: &lt;script type=\"text\/javascript\"\n#+HTML_HEAD:   src=\"http:\/\/cdn.mathjax.org\/mathjax\/latest\/MathJax.js\"&gt;\n#+HTML_HEAD: &lt;\/script&gt;\n    <\/pre>\n  <\/li>\n<\/ul>\n<ol><li>You may want to look at the <a href=\"http:\/\/orgmode.org\/manual\/Export-options.html#Export-options\" rel=\"nofollow\">Export options<\/a> and <a href=\"http:\/\/orgmode.org\/manual\/HTML-export.html#HTML-export\" rel=\"nofollow\">HTML export<\/a>\n    sections of the org-manual.<\/li>\n  <li>If you wish to post to blogger from org-mode, look at -\n    <ol><li><a href=\"http:\/\/repo.or.cz\/r\/org2blog.git\/\" rel=\"nofollow\">org2blog\/atom<\/a>, a g-client extension by <a href=\"http:\/\/tehom-blog.blogspot.com\/\" rel=\"nofollow\">Tom Breton<\/a><\/li>\n      <li><a href=\"http:\/\/github.com\/rileyrg\/org-googlecl\">org-googlecl<\/a> by Richard Riley \u2013 uses <a href=\"http:\/\/code.google.com\/p\/googlecl\/\" rel=\"nofollow\">googlecl<\/a><\/li>\n    <\/ol><\/li>\n  <li>Please go through the README and the FAQ, before writing to me.\n    Also, looking at <code>M-x customize-group org2blog-wp<\/code> might help.<\/li>\n  <li>If you have an issue\/bug\/feature request, use the issue tracker\n    on git or drop a mail to punchagan+org2blog[at]gmail[dot]com.\n    I\u2019d also appreciate patches\/suggestions to improve the\n    documentation.  Feel free to drop in with general comments, too.\n    I\u2019d love to hear from you!  NOTE: If you are using the package\n    from ELPA, please try using the latest <code>git<\/code> version\n    before filing a bug report.<\/li>\n  <li>Feel free to add your site to the list of sites using org2blog,\n    on the <a href=\"https:\/\/github.com\/punchagan\/org2blog\/wiki\/Blogs-using-org2blog\">wiki<\/a> at github.<\/li>\n<\/ol>\n<ul><li>How many blogs can I configure with org2blog?\n    <p>You can configure and use any number of blogs with org2blog\/wp.\n      Use the <code>org2blog\/wp-blog-alist<\/code> variable to configure each blog.\n      Look at the <a href=\"https:\/\/github.com\/org2blog\/org2blog\/blob\/master\/Defining%20your%20blog%20setup\">Defining your blog setup<\/a> section, in the README.  But,\n      note that you can be logged in, to just one blog at a time.<\/p>\n  <\/li>\n  <li>How do I change the default title of a new post?\n    <pre>(setq org2blog\/wp-default-title \"My New Title\")\n    <\/pre>\n  <\/li>\n  <li>How do I change the default title for one blog alone?\n    <p>Set the relevant (<code>:default-title<\/code>) variable in the\n      <code>org2blog\/wp-blog-alist<\/code> variable.<\/p>\n  <\/li>\n  <li>I do not use tags. I wish to use them as categories. How?\n    <p>Setting the <code>org2blog\/wp-use-tags-as-categories<\/code> variable will do\n      that for all the blogs.<\/p>\n    <pre>(setq org2blog\/wp-use-tags-as-categories t)\n    <\/pre>\n    <p>Instead, you can set the corresponding variable for each blog that\n      you are using.<\/p>\n  <\/li>\n  <li>Can I configure org2blog to confirm before publishing a post?\n    <pre>(setq org2blog\/wp-confirm-post t)\n    <\/pre>\n  <\/li>\n  <li>I wish to customize the default template of a new post. How do I\n    do it?\n    <p>Customize the variable <code>org2blog\/wp-buffer-template<\/code>.<\/p>\n  <\/li>\n  <li>New-lines are not handled properly. Why?\n    <p>The behaviour of new-lines has been working since a little before\n      version 0.3.  It is highly recommended that you use the latest git\n      version of org2blog.<\/p>\n    <p>By default, new lines are stripped off from the org buffer.  To\n      retain new-lines, unset the <code>org2blog\/wp-keep-new-lines<\/code>\n      variable.<\/p>\n  <\/li>\n  <li>Why aren\u2019t my SRC blocks not enclosed in <code>[sourcecode]\n    [\/sourcecode]<\/code>  block?\n    <p>Set the <code>org2blog\/wp-use-sourcecode-shortcode<\/code> variable to turn on\n      this behaviour.  You may also set it at a per-blog level, if you\n      choose.<\/p>\n  <\/li>\n  <li>Is wordpress the only CMS\/Blog-engine that org2blog\/wp supports?\n    <p>Any blog engine using the metaweblog API should work,\n      theoretically.  But, it is not tested with other blog engines.<\/p>\n    <p>One happy user reports that org2blog <a href=\"https:\/\/github.com\/punchagan\/org2blog\/issues\/issue\/37\">also works<\/a> with the <a href=\"http:\/\/dotclear.org\/\" rel=\"nofollow\">Dotclear<\/a>\n      weblog engine.<\/p>\n  <\/li>\n  <li>Is there a way to import from WordPress or other blogs into the\n    org2blog post format?\n    <p>I\u2019ve a simple <a href=\"https:\/\/github.com\/punchagan\/org2blog-importers\/blob\/master\/wp_to_org2blog.py\">python script<\/a> that uses <code>pandoc<\/code> to convert from\n      WordPress export xml to org2blog posts.  It could easily be\n      tweaked to write importers for other kinds of blogs.<\/p>\n  <\/li>\n  <li>How do I split a post into an introductory paragraph and a full\n    view.\n    <p>Just put in<\/p>\n    <pre>#+HTML: &lt;!--more--&gt;\n    <\/pre>\n    <p>at the location where you wish to split the post.<\/p>\n  <\/li>\n<\/ul>","created_at":"2018-11-07T16:40:34+0000","updated_at":"2018-11-07T16:41:07+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/29651589?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":245,"title":"Babel: Introduction","url":"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/intro.html","content":"<p>\nTangling functionality is controlled by the <code>tangle<\/code> family of tangle\nheader arguments.  These arguments can be used to turn tangling on or\noff (the default), either for the code block or the Org-mode\nheading level.\n<\/p><p>\nThe following code blocks demonstrate how to tangle them into a\nsingle source code file using <code>org-babel-tangle<\/code>.\n<\/p><p>\nThe following two code blocks have no <code>tangle<\/code> header arguments\nand so will not, by themselves, create source code files.  They are\nincluded in the source code file by the third code block, which\ndoes have a <code>tangle<\/code> header argument.\n<\/p><p>\nIn the Org-mode file:\n<\/p><pre class=\"example\">\n#+name: hello-world-prefix\n#+begin_src sh :exports none\n  echo \"\/-----------------------------------------------------------\\\\\"\n#+end_src\n<\/pre><p>\nHTML export of code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\" id=\"org248d479\">echo \"\/-----------------------------------------------------------\\\\\"\n<\/pre><\/div><p>\nIn the Org-mode file\n<\/p><pre class=\"example\">\n#+name: hello-world-postfix\n#+begin_src sh :exports none\n  echo \"\\-----------------------------------------------------------\/\"\n#+end_src\n<\/pre><p>\nHTML export of code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\" id=\"orgb34d9a2\">echo \"\\-----------------------------------------------------------\/\"\n<\/pre><\/div><p>\nThe third code block does have a <code>tangle<\/code> header argument\nindicating the name of the file to which the tangled source code will\nbe written.  It also has <a href=\"http:\/\/www.cs.tufts.edu\/~nr\/noweb\/\">Noweb<\/a> style references to the two previous\ncode blocks.  These references will be expanded during tangling\nto include them in the output file as well.\n<\/p><p>\nIn the Org-mode file:\n<\/p><pre class=\"example\">\n#+name: hello-world\n#+begin_src sh :tangle hello :exports none :noweb yes\n  &lt;&lt;hello-world-prefix&gt;&gt;\n  echo \"|                       hello world                         |\"\n  &lt;&lt;hello-world-postfix&gt;&gt;\n#+end_src\n<\/pre><p>\nHTML export of code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\" id=\"org572bc7b\">&lt;&lt;hello-world-prefix&gt;&gt;\necho \"|                       hello world                         |\"\n&lt;&lt;hello-world-postfix&gt;&gt;\n<\/pre><\/div><p>\nCalling <code>org-babel-tangle<\/code> will result in the following shell source\ncode being written to the <code>hello.sh<\/code> file:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\" id=\"org850af31\">#!\/usr\/bin\/env sh\n# [[file:~\/org\/temp\/index.org::*Noweb%20test][hello-world]]\necho \"\/-----------------------------------------------------------\\\\\"\necho \"|                       hello world                         |\"\necho \"\\-----------------------------------------------------------\/\"\n# hello-world ends here\n<\/pre><\/div><p>\nIn addition, the following syntax can be used to insert the <b>results<\/b>\nof evaluating a code block, in this case one named <code>example-block<\/code>.\n<\/p><pre class=\"example\">\n# &lt;&lt;example-block()&gt;&gt;\n<\/pre><p>\nAny optional arguments can be passed to <code>example-block()<\/code> by placing the\narguments inside the parentheses following the convention defined when\ncalling source block functions (see the <a href=\"#library-of-babel\">Library of babel<\/a>). For example,\n<\/p><pre class=\"example\">\n# &lt;&lt;example-block(a=9)&gt;&gt;\n<\/pre><p>\nsets the value of argument \\\"a\\\" equal to \\\"9\\\".  Note that\nthese arguments are not evaluated in the current source-code\nblock but are passed literally to <code>example-block()<\/code>.\n<\/p>","created_at":"2018-11-07T16:40:34+0000","updated_at":"2018-11-17T03:37:16+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":246,"title":"Babel: Introduction","url":"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/intro.html","content":"<p>\nTangling functionality is controlled by the <code>tangle<\/code> family of tangle\nheader arguments.  These arguments can be used to turn tangling on or\noff (the default), either for the code block or the Org-mode\nheading level.\n<\/p><p>\nThe following code blocks demonstrate how to tangle them into a\nsingle source code file using <code>org-babel-tangle<\/code>.\n<\/p><p>\nThe following two code blocks have no <code>tangle<\/code> header arguments\nand so will not, by themselves, create source code files.  They are\nincluded in the source code file by the third code block, which\ndoes have a <code>tangle<\/code> header argument.\n<\/p><p>\nIn the Org-mode file:\n<\/p><pre class=\"example\">\n#+name: hello-world-prefix\n#+begin_src sh :exports none\n  echo \"\/-----------------------------------------------------------\\\\\"\n#+end_src\n<\/pre><p>\nHTML export of code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">echo \"\/-----------------------------------------------------------\\\\\"\n<\/pre><\/div><p>\nIn the Org-mode file\n<\/p><pre class=\"example\">\n#+name: hello-world-postfix\n#+begin_src sh :exports none\n  echo \"\\-----------------------------------------------------------\/\"\n#+end_src\n<\/pre><p>\nHTML export of code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">echo \"\\-----------------------------------------------------------\/\"\n<\/pre><\/div><p>\nThe third code block does have a <code>tangle<\/code> header argument\nindicating the name of the file to which the tangled source code will\nbe written.  It also has <a href=\"http:\/\/www.cs.tufts.edu\/~nr\/noweb\/\">Noweb<\/a> style references to the two previous\ncode blocks.  These references will be expanded during tangling\nto include them in the output file as well.\n<\/p><p>\nIn the Org-mode file:\n<\/p><pre class=\"example\">\n#+name: hello-world\n#+begin_src sh :tangle hello :exports none :noweb yes\n  &lt;&lt;hello-world-prefix&gt;&gt;\n  echo \"|                       hello world                         |\"\n  &lt;&lt;hello-world-postfix&gt;&gt;\n#+end_src\n<\/pre><p>\nHTML export of code:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">&lt;&lt;hello-world-prefix&gt;&gt;\necho \"|                       hello world                         |\"\n&lt;&lt;hello-world-postfix&gt;&gt;\n<\/pre><\/div><p>\nCalling <code>org-babel-tangle<\/code> will result in the following shell source\ncode being written to the <code>hello.sh<\/code> file:\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">#!\/usr\/bin\/env sh\n# [[file:~\/org\/temp\/index.org::*Noweb%20test][hello-world]]\necho \"\/-----------------------------------------------------------\\\\\"\necho \"|                       hello world                         |\"\necho \"\\-----------------------------------------------------------\/\"\n# hello-world ends here\n<\/pre><\/div><p>\nIn addition, the following syntax can be used to insert the <b>results<\/b>\nof evaluating a code block, in this case one named <code>example-block<\/code>.\n<\/p><pre class=\"example\">\n# &lt;&lt;example-block()&gt;&gt;\n<\/pre><p>\nAny optional arguments can be passed to <code>example-block()<\/code> by placing the\narguments inside the parentheses following the convention defined when\ncalling source block functions (see the <a href=\"#library-of-babel\">Library of babel<\/a>). For example,\n<\/p><pre class=\"example\">\n# &lt;&lt;example-block(a=9)&gt;&gt;\n<\/pre><p>\nsets the value of argument \\\"a\\\" equal to \\\"9\\\".  Note that\nthese arguments are not evaluated in the current source-code\nblock but are passed literally to <code>example-block()<\/code>.\n<\/p>","created_at":"2018-11-07T16:40:34+0000","updated_at":"2018-11-07T16:41:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":247,"title":"Babel: active code in Org-mode","url":"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/index.html","content":"<p>\nBabel is Org-mode's ability to execute source code within Org-mode\ndocuments.  Org-mode is an <a href=\"http:\/\/www.gnu.org\/software\/emacs\/\">Emacs<\/a> major mode for doing almost anything\nwith plain text.  If you are not familiar with Org-mode please take a\nmoment to read <a href=\"https:\/\/orgmode.org\/\">the Org-mode homepage<\/a> before continuing.  Babel started\nlife as Org-babel, an extension to Org-mode.  It was integrated into\nthe Org-mode core at version 7.0. The author of Babel is Eric\nSchulte. The secondary author is Dan Davison.\n<\/p><ul class=\"org-ul\"><li><a href=\"https:\/\/www.jstatsoft.org\/article\/view\/v046i03\">A Multi-Language Computing Environment for Literate Programming and\nReproducible Research<\/a>: a journal paper providing the most up to date\nand complete introduction to using Org-mode and Babel for\nReproducible research and literate programming.<\/li>\n<li><a href=\"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/intro.html\">Introductory Tutorial<\/a>: a smaller online set of installation and\nsetup instructions, as well as an overview of the main features of\nBabel, with plenty of examples.<\/li>\n<li><a href=\"https:\/\/orgmode.org\/manual\/Working-with-source-code.html\">Reference and Documentation<\/a>: is available online and in Chapter 14\nof the Org-mode manual. Some more information on using header args\nand result types for Babel code blocks can be found <a href=\"https:\/\/orgmode.org\/manual\/Header-arguments.html#Header-arguments\">here<\/a>.<\/li>\n<li><a href=\"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/languages.html\">Languages<\/a>: list of supported languages with information on\nrequirements, configuration instructions, and pointers to language\nspecific documentation.<\/li>\n<li><a href=\"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/uses.html\">Uses<\/a>: user-submitted examples of Babel work flows \u2013 this is a good\nplace to get an idea of what can be done and see how Babel may fit\ninto your life.  In addition there is a collection of small code\nblock examples and scraps at <a href=\"http:\/\/eschulte.github.io\/org-scraps\/\">org-scraps<\/a>.<\/li>\n<\/ul><hr \/><div id=\"logo1\"><p>\n    <img src=\"https:\/\/orgmode.org\/worg\/images\/babel\/tower-of-babel.png\" alt=\"Tower of Babel\" \/><\/p><p>The Tower of Babel by\n      <a href=\"https:\/\/commons.wikimedia.org\/wiki\/Pieter_Brueghel_the_Elder\" title=\"\">\n        <b>Pieter Brueghel the Elder<\/b>\n      <\/a><\/p><p>\n      And the Lord said, Behold, the people is one, and they have all\n      one language; and this they begin to do; and now nothing will be\n      restrained from them, which they have imagined to do. Genesis\n      11:1-9\n    <\/p><\/div><div id=\"logo2\"><p>\n    <img src=\"https:\/\/orgmode.org\/worg\/images\/babel\/babelfish.png\" alt=\"Babel Fish\" \/><\/p><p>\n      The Babel Fish is small, yellow, and simultaneously translates\n      from one spoken language to another.\n    <\/p><p>\n      \u2013 The Hitchhiker's Guide to the Galaxy, Douglas Adams\n    <\/p><\/div>","created_at":"2018-11-07T16:40:34+0000","updated_at":"2018-11-07T16:41:10+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":248,"title":"Managing dotfiles with org mode?","url":"https:\/\/www.reddit.com\/r\/emacs\/comments\/5oimrs\/managing_dotfiles_with_org_mode\/","content":"<form action=\"#\" class=\"usertext warn-on-unload\" id=\"form-t3_5oimrs2dq\"><div class=\"usertext-body may-blank-within md-container\"><div class=\"md\"><p>Hey there,<\/p>\n<p>I have been playing around with emacs for a while now. What really got me hooked was org-mode's ability to do literal programming as well as file exporting. Naturally the first thing I did, when I heard about it was to plunder github for various pieces of emacs configuration and putting them into a single org file. So far this has been working great for me, I know exactly what the configuration is supposed to do and could also export the configuration anytime.<\/p>\n<p>My question is now, how I can adapt this literal programming approach to various other configuration files, e.g. in order to manage my dotfiles. I would like to keep an org file per application (e.g. i3wm, i3status, vim etc.) and to be able to export the source code blocks of the files into valid configuration files, which I could then deploy using emacs export. Also I'd like to know whether it is possible to export various different versions of the configuration based on tagging or properties?<\/p>\n<p>I have found a <a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/4b713m\">post<\/a> from 10 months ago, but it seems to be down.<\/p>\n<p>Do any of you use emacs for similiar tasks? <\/p>\n<p>Many thanks in advance!<\/p>\n<p>EDIT: Thanks for all your replies! This is exactly what I was looking for! <\/p>\n<\/div>\n<\/div><\/form>","created_at":"2018-11-07T16:40:34+0000","updated_at":"2018-11-07T16:41:12+0000","published_at":"2017-01-17T15:57:56+0000","published_by":["tigoesnumb3rs"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.reddit.com","preview_picture":"https:\/\/www.redditstatic.com\/new-icon.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":249,"title":"One Order of Operations for Starting a Startup","url":"https:\/\/blog.ycombinator.com\/one-order-of-operations-for-starting-a-startup\/","content":"<p>More often than not, when I talk to a talented technical person who\u2019s thinking about becoming a founder, their number one blocker is that they don\u2019t have an idea. At some point during their formative years they learned that every great startup started with a great idea and if the idea isn\u2019t amazing (usually as judged by peers, parents, or other people with little startup experience), the startup will fail.<\/p><p>My first startup\u2019s idea (really Justin Kan\u2019s idea) was to create a live reality TV show where people would walk around with cameras on their head livestreaming their lives 24\/7. This company eventually became Twitch.tv and sold to Amazon for almost a billion dollars (thanks to some amazing work by Emmett Shear, Kevin Lin, and a great group of employees). Airbnb started with the idea that when you want to attend a conference, sometimes all the hotels in that city are sold out. Therefore, wouldn\u2019t it be great to have a site where you can find an airbed to crash on. In 2017 over 200 million nights were booked on Airbnb, <a href=\"https:\/\/press.atairbnb.com\/the-u-s-midwest-ryokans-and-brazilian-beaches-airbnbs-2018-travel-trends\/\">almost none of them fitting the original purpose of the site<\/a>.<\/p><p>I\u2019d like to offer an alternative to the conventional \u201cgreat idea\u201d theory of startups for first time founders to consider. This guide can help you walk the path from nothing to a launched minimal viable product (MVP). This is by no means the only path to an MVP and I don\u2019t know if it\u2019s the best path, but it is a path that I\u2019ve seen work for a number of YC companies.<\/p><h2>The Order of Operations<\/h2><p>Here\u2019s where I\u2019d start: Is there any particular problem that you are passionate about? For younger founders, that problem can come from your experience growing up, from your family\/friends, your personal interests, or from school. For older founders it could also come from experiences at work, being married, or being someone\u2019s partner. For the Airbnb founders, their original problem was how to pay rent. For Justin.tv\/Twitch the original problem Justin and Emmett had was what startup should we work on next. Some problems seem important and impactful at the beginning and others feel like trivial toys. I\u2019ve seen both types grow into billion dollar companies.<\/p><p>Once you decide on a problem, find some friends and brainstorm potential solutions to the problem. Those friends can be other people who also have the same problem but they don\u2019t have to be. They can be people you went to school with, people from work, or people you\u2019ve met in your personal life. What\u2019s probably most important is that you pick friends who are smart and fun to brainstorm with. When someone is fun to brainstorm with \u2013 instead of rejecting potential solutions as they are said aloud \u2013 they riff on the solutions and improve upon them. If you\u2019re not technical I\u2019d encourage you to brainstorm with friends\/colleagues who are. If you pick the correct friends to talk to, sometimes the result of this brainstorm is a solution or idea that you can get really excited about. This is the spark that can make your startup come alive. Because the people you brainstormed with feel some ownership for the solution you came up with together, this is the perfect time to ask them to work with you to actually build that solution. A great brainstorm can often not only provide you with an idea, it can also provide you with co-founders (so pick your brainstorming partners carefully).<sup id=\"footnoteid1\"><a href=\"#footnote1\">1<\/a><\/sup><\/p><p>At this point the most important thing to make that spark turn into a fire is to work together to build and launch a minimal viable product (MVP). An MVP is the smallest possible product you can build that allows you to start offering your service to users. Often times you need very little software to start. Resist the temptation to build your complete solution because in reality you have no idea whether it will work. Better to build an extremely stripped down version of your solution and start seeing if users actually want to use it. Airbnb\u2019s first version had almost no features (no map view, no profiles, no messaging, no payments, etc) and they referred to it as \u201cAirbnb Lite\u201d and built it in under a month. It\u2019s totally okay if you need to do a bunch of manual work that doesn\u2019t scale on the backend, is users actually want what you are making you\u2019ll have time to fix everything and make it scale later. Once you have your first users of any kind, you have a startup. That being said, you still need someone with the technical skills to ship the MVP.<\/p><p>Once you\u2019ve put your MVP together, there\u2019s a ton of startup advice to follow. Here are a few resources I recommend:<br \/>\u2013 <a href=\"https:\/\/blog.ycombinator.com\/ycs-essential-startup-advice\/\">YC\u2019s Essential Advice<\/a><br \/>\u2013 <a href=\"http:\/\/www.paulgraham.com\/articles.html\">Paul Graham\u2019s Essays<\/a><br \/>\u2013 <a href=\"https:\/\/pmarchive.com\/\">Blog Pmarca<\/a><\/p><h2>Two Failed Orders of Operations<\/h2><p>There are a couple of paths I see first time founders walk that often lead them to failure.<\/p><p><strong>1) Come up with an idea and then pitch investors.<\/strong><br \/>Usually these founders think that investors, by investing, will validate their idea and that if they can\u2019t raise money \u2013 they don\u2019t have a good idea. Good investors are interested in teams who are actually in the process of executing their idea and tend to shy away from teams that are just pitching their idea.<\/p><p><strong>2) I have an idea but can\u2019t write code and none of my co-founders can either.<\/strong><br \/>These founders want to pay someone else to build their product. Then they\u2019ll launch it, try to raise money, and build out a tech team. The problem there is that good investors are actually investing in the team you\u2019ve built. If you are building a tech startup and no one on your team can write code or do the technical work necessary, the bar to raise investment is much much higher. In my experience you have to have 10x the amount of traction to get a good investor to overlook a non-technical team. Also, in general the problem with outsourcing engineering is that the people you outsource aren\u2019t as motivated as you are and are often fairly expensive.<\/p><h2>Counterexamples<\/h2><p>There are a lot of counterexamples to the \u201cfailure\u201d scenarios I described above, both for first time and repeat founders. What I will tell you is that building a successful startup is extremely hard and when you choose to follow one of the failure paths \u2013 often your road is much harder and less likely to lead to success. Why make things more painful on yourself?<\/p><p>By no means do I think this is the only path but when founders ask me to recommend a path, this is the one I recommend. It\u2019s the most effective one I\u2019ve seen in getting founders past the first major hurdles of putting something out and finding initial funding.<\/p><hr \/><p><strong>Note<\/strong><br \/><b id=\"footnote1\">1.<\/b> Now, what if you\u2019re not technical and don\u2019t have access to technical people? Often what I tell future founders is go work at a startup and make those friends. At the end of the day, having the right founding team is such a fast track to to building a successful startup that it\u2019s worth investing a year or two building those relationships before you start.<a href=\"#footnoteid1\">\u21a9<\/a><\/p><p><em>Thanks to Carolynn Levy, Adora Cheung, Daniel Gackle, and Craig Cannon for reading drafts of this post.<\/em><\/p>","created_at":"2018-11-07T16:40:34+0000","updated_at":"2018-11-07T16:41:17+0000","published_by":["Michael Seibel"],"annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"blog.ycombinator.com","preview_picture":"https:\/\/blog.ycombinator.com\/wp-content\/uploads\/2018\/03\/One-Order-of-Operations-for-Starting-a-Startup.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":242,"title":"A CEO's Guide to Emacs","url":"https:\/\/blog.fugue.co\/2015-11-11-guide-to-emacs.html","content":"<p>Years\u2014no, decades\u2014ago, I lived in Emacs. I wrote code and documents, managed email and calendar, and shelled all in the editor\/OS. I was quite happy. Years went by and I moved to newer, shinier things. As a result, I forgot how to do tasks as basic as efficiently navigating files without a mouse. About three months ago, noticing just how much of my time was spent switching between applications and computers, I decided to give Emacs another try. It was a good decision for several reasons that will be covered in this post. Covered too are <code>.emacs<\/code> and Dropbox tips so that you can set up a good, movable environment.<\/p><p>For those who haven't used Emacs, it's something you'll likely hate, but may love. It's sort of a Rube Goldberg machine the size of a house that, at first glance, performs all the functions of a toaster. That hardly sounds like an endorsement, but the key phrase is \"at first glance.\" Once you grok Emacs, you realize that it's a thermonuclear toaster that can also serve as the engine for... well, just about anything you want to do with text. When you think about how much your computing life revolves around text, this is a rather bold statement. Bold, but true.<\/p><p>Perhaps more importantly to me though, it's the one application I've ever used that makes me feel like I really own it instead of casting me as an anonymous \"user\" whose wallet is cleverly targeted by product marketing departments in fancy offices somewhere near <a href=\"http:\/\/www.huffingtonpost.com\/zachary-ehren\/soma-isnt-a-drug-san-fran_b_987841.html\">Soma<\/a> or Redmond. Modern productivity and authoring applications (e.g., Pages or IDEs) are like carbon fiber racing bikes. They come kitted out very nicely and fully assembled. Emacs is like a box of classic <a href=\"http:\/\/www.campagnolo.com\/US\/en\">Campagnolo<\/a> parts and a beautiful lugged steel frame that's missing one crank arm and a brake lever that you have to find in some tiny subculture on the Internet. The first one is faster and complete. The second is a source of endless joy or annoyance depending on your personality\u2014and will last until your dying day. I'm the sort of person who feels equal joy at finding an old stash of Campy parts or tweaking my editor with eLisp. YMMV.<\/p><figure><img src=\"https:\/\/blog.fugue.co\/assets\/images\/bicycle.jpg\" alt=\"1933 steel bicycle\" \/><figcaption>A 1933 steel bicycle that I still ride. Check out this comparison of frame tubes: <a href=\"https:\/\/www.youtube.com\/watch?v=khJQgRLKMU0\">https:\/\/www.youtube.com\/watch?v=khJQgRLKMU0<\/a>.<\/figcaption><\/figure><p>This may give the impression that Emacs is anachronistic or old-fashioned. It's not. It's powerful and timeless, but demands that you patiently understand it on its terms. The terms are pretty far off the beaten path and seem odd, but there is a logic to them that is both compelling and charming. To me, Emacs feels like the future rather than the past. Just as the lugged steel frame will be useful and comfortable in decades to come and the carbon fiber wunderbike will be in a landfill, having shattered on impact, so will Emacs persist as a useful tool when the latest trendy app is long forgotten.<\/p><p>If the notion of building your own personal working environment by editing Lisp code and having that fits-like-a-glove environment follow you to any computer is appealing to you, you may really like Emacs. If you like the new and shiny and want to get straight to work without much investment of time and mental cycles, it's likely not for you. I don't write code any more (other than Ludwig and Emacs Lisp), but many of the engineers at Fugue use Emacs to good effect. I'd say our engineers are about 30% Emacs, 40% IDEs, and 30% Vim users. But, this post is about Emacs for CEOs and other <a href=\"http:\/\/www.businessinsider.com\/best-pointy-haired-boss-moments-from-dilbert-2013-10\">Pointy-Haired Bosses<\/a> (PHB)<sup id=\"phbOrigin\"><a href=\"#phb\">1<\/a><\/sup> (and, hey, anyone who\u2019s curious), so I'm going to explain and\/or rationalize why I love it and how I use it. I also hope to provide you with enough detail that you can have a successful experience with it, without hours of Googling.<\/p><h2 id=\"lasting-advantages\">Lasting Advantages<\/h2><p>The long-term advantages that come with using Emacs just make life easier. The net gain makes the initial lift entirely worthwhile. Consider these:<\/p><h3 id=\"no-more-context-switching\">No More Context Switching<\/h3><p>Org Mode alone is worth investing some serious time in, but if you are like me, you are usually working on a dozen or so documents\u2014from blog posts to lists of what you need to do for a conference to employee reviews. In the modern world of computing, this generally means using several applications, all of which have distracting user interfaces and different ways to store, sort, and search. The result is that you need to constantly switch mental contexts and remember minutiae. I hate context switching because it is an imposition put on me due to a broken interface model<sup id=\"tufteOrigin\"><a href=\"#tufte\">2<\/a><\/sup> and I hate having to remember things my computer should remember for me in any rational world. In providing a single environment, Emacs is even more powerful for the PHB than the programmer, since programmers tend to spend a greater percentage of their day in a single application. Switching mental contexts has a higher cost than is often apparent. OS and application vendors have tarted up interfaces to distract us from this reality. If you\u2019re technical, having access to a powerful <a href=\"http:\/\/www.webopedia.com\/TERM\/I\/interpreter.html\">language interpreter<\/a> in a single keyboard shortcut (<code>M-:<\/code>) is especially useful.<sup id=\"interpreterOrigin\"><a href=\"#interpreter\">3<\/a><\/sup><\/p><p>Many applications can be full screened all day and used to edit text. Emacs is singular because it is both an editor and a Lisp interpreter. In essence, you have a Turing complete machine a keystroke or two away at all times, while you go about your business. If you know a little or a lot about programming, you'll recognize that this means you can do <em>anything<\/em> in Emacs. The full power of your computer is available to you in near real time while you work, once you have the commands in memory. You won't want to re-create Excel in eLisp, but most things you might do in Excel are smaller in scope and easy to accomplish in a line or two of code. If I need to crunch numbers, I'm more likely to jump over to the scratch buffer and write a little code than open a spreadsheet. Even if I have an email to write that isn't a one-liner, I'll usually just write it in Emacs and paste it into my email client. Why context switch when you can just flow? You might start with a simple calculation or two, but, over time, anything you need computed can be added with relative ease to Emacs. This is perhaps unique in applications that also provide rich features for creating things for other humans. Remember those magical terminals in Isaac Asimov's books? Emacs is the closest thing I've encountered to them.<sup id=\"ewwOrigin\"><a href=\"#eww\">4<\/a><\/sup> I no longer decide what app to use for this or that thing. Instead, I just work. There is real power and efficiency to having a great tool and committing to it.<\/p><h3 id=\"creating-things-in-peace-and-quiet\">Creating Things in Peace and Quiet<\/h3><p>What\u2019s the end result of having the best text editing features I've ever found? Having a community of people making all manner of useful additions? Having the full power of Lisp a keychord away? It\u2019s that I use Emacs for all my creative work, aside from making music or images.<\/p><p>I have a dual monitor set up at my desk. One of them is in portrait mode with Emacs full screened all day long. The other one has web browsers for researching and reading; it usually has a terminal open as well. I keep my calendar, email, etc., on another desktop in OS X, which is hidden while I'm in Emacs, and I keep all notifications turned off. This allows me to actually concentrate on what I'm doing. I've found eliminating distractions to be almost impossible in the more modern UI applications due to their efforts to be helpful and easy to use. I don't need to be constantly reminded how to do operations I've done tens of thousands of times, but I do need a nice, clean white sheet of paper to be thoughtful. Maybe I'm just bad at living in noisy environments due to age and abuse, but I\u2019d suggest it\u2019s worth a try for anyone. See what it's like to have some actual peace and quiet in your computing environment. Of course, lots of apps now have modes that hide the interface and, thankfully, both Apple and Microsoft now have meaningful full-screen modes. But, no other application is powerful enough to \u201clive in\u201d for most things. Unless you are writing code all day or perhaps working on a very long document like a book, you're still going to face the noise of other apps. Also, most modern applications seem simultaneously patronizing and lacking in functionality and usability.<sup id=\"usabilityOrigin\"><a href=\"#usability\">5<\/a><\/sup> The only applications I dislike more than office apps are the online versions.<\/p><figure><img src=\"https:\/\/blog.fugue.co\/assets\/images\/desktop.jpg\" alt=\"1933 steel bicycle\" \/><figcaption>My desktop arrangement. Emacs on the left.<\/figcaption><\/figure><p>But what about communicating? The difference between creating and communicating is substantial. I'm much more productive at both when I set aside distinct time for each. We use Slack at Fugue, which is both wonderful and hellish. I keep it on a messaging desktop alongside my calendar and email, so that, while I'm actually making things, I'm blissfully unaware of all the chatter in the world. It takes just one Slackstorm or an email from a VC or Board Director to immediately throw me out of my work. But, most things can usually wait an hour or two.<\/p><h3 id=\"taking-everything-with-you-and-keeping-it-forever\">Taking Everything with You and Keeping It Forever<\/h3><p>The third reason I find Emacs more advantageous than other environments is that it's easy to take all your stuff with you. By this, I mean that, rather than having a plethora of apps interacting and syncing in their own ways, all you need is one or two directories syncing via Dropbox or the like. Then, you can have all your work follow you anywhere in the environment you have crafted to suit your purposes. I do this across OS X, Windows, and sometimes Linux. It's dead simple and reliable. I've found this capability to be so useful that I dread dealing with Pages, GDocs, Office, or other kinds of files and applications that force me back into finding stuff somewhere on the filesystem or in the cloud.<\/p><p>The limiting factor in keeping things forever on a computer is file format. Assuming that humans have now solved the problem of storage <sup id=\"s3Origin\"><a href=\"#s3\">6<\/a><\/sup> for good, the issue we face over time is whether we can continue to access the information we've created. Text files are the most long-lived format for computing. You easily can open a text file from 1970 in Emacs. That\u2019s not so true for Office applications. Text files are also nice and small\u2014radically smaller than Office application data files. As a digital pack rat and as someone who makes lots of little notes as things pop into my head, having a simple, light, permanent collection of stuff that is always available is important to me.<\/p><p>If you\u2019re feeling ready to give Emacs a try, read on! The sections that follow don\u2019t take the place of a full tutorial, but will have you operational by the time you finish reading.<\/p><h2 id=\"learning-to-ride-emacs-a-technical-setup\">Learning To Ride Emacs - A Technical Setup<\/h2><p>The price of all this power and mental peace and quiet is that you have a steep learning curve with Emacs and it does everything differently than you're used to. At first, this will make you feel like you\u2019re wasting time on an archaic and strange application that the modern world passed by. It\u2019s a bit like learning to ride a bicycle<sup id=\"bicyclesOrigin\"><a href=\"#bicycles\">7<\/a><\/sup> if you've only driven cars. <\/p><h3 id=\"which-emacs-\">Which Emacs?<\/h3><p>I use the plain vanilla Emacs from GNU for OS X and Windows. You can get the OS X version at <a href=\"http:\/\/emacsformacosx.com\/\">http:\/\/emacsformacosx.com\/<\/a> and the Windows version at <a href=\"http:\/\/www.gnu.org\/software\/emacs\/\">http:\/\/www.gnu.org\/software\/emacs\/<\/a>. There are a bunch of other versions out there, especially for the Mac, but I've found the learning curve for doing powerful stuff (which involves Lisp and lots of modes) to be much lower with the real deal. So download it, and we can get started!<sup id=\"nottutorialOrigin\"><a href=\"#nottutorial\">8<\/a><\/sup><\/p><p>I use the Emacs conventions for keys and combinations in this document. These are 'C' for control, 'M' for meta (which is usually mapped to Alt or Option), and the hyphen for holding down the keys in combination. So <code>C-h t<\/code> means to hold down control and type h, then release control and type t. This is the command for bringing up the tutorial, which you should go ahead and do.<\/p><p>Don't use the arrow keys or the mouse. They work, but you should give yourself a week of using the native navigation commands in Emacs. Once you have them committed to muscle memory, you'll likely enjoy them and miss them badly everywhere else you go. The Emacs tutorial does a pretty good job of walking you through them, but I'll summarize so you don't need to read the whole thing. The boring stuff is that, instead of arrows, you use <code>C-b<\/code> for back, <code>C-f<\/code> for forward, <code>C-p<\/code> for previous (up), and <code>C-n<\/code> for next (down). You may be thinking \"why in the world would I do that, when I have perfectly good arrow keys?\" There are several reasons. First, you don't have to move your hands from the typing position, and the forward and back keys used with Alt (or Meta in Emacspeak) navigate a word at a time. This is more handy than is obvious. The third good reason is that, if you want to repeat a command, you can precede it with a number. I often use this when editing documents by estimating how many words I need to go back or lines up or down and doing something like <code>C-9 C-p<\/code> or <code>M-5 M-b<\/code>. The other really important navigation commands are based on <code>a<\/code> for the beginning of a thing and <code>e<\/code> for the end of a thing. Using <code>C-a|e<\/code> are on lines, and using <code>M-a|e<\/code>, are on sentences. For the sentence commands to work properly, you'll need to double space after periods, which simultaneously provides a useful feature and takes a shibboleth of <a href=\"http:\/\/www.huffingtonpost.com\/2015\/05\/29\/two-spaces-after-period-debate_n_7455660.html\">opinion<\/a> off the mental table. If you need to export the document to a single space <a href=\"http:\/\/practicaltypography.com\/one-space-between-sentences.html\">publication environment<\/a>, you can write a macro in moments to do so.<\/p><p>It genuinely is worth going through the tutorial that ships with Emacs. I'll cover a few important commands for the truly impatient, but the tutorial is gold. Reminder: <code>C-h t<\/code> for the tutorial.<\/p><h3 id=\"learn-to-copy-and-paste\">Learn To Copy and Paste<\/h3><p>You can put Emacs into <code>CUA<\/code> mode, which will work in familiar ways, but the native Emacs way is pretty great and plenty easy once you learn it. You mark regions (like selecting) by using Shift with the navigation commands. So <code>C-F<\/code> selects one character forward from the cursor, etc. You copy with <code>M-w<\/code>, you cut with <code>C-w<\/code>, and you paste with <code>C-y<\/code>. These are actually called killing and yanking, but it's very similar to cut and paste. There is magic under the hood here in the kill ring, but for now, just worry about cut, copy, and paste. If you start fumbling around at this point, <code>C-x u<\/code> is undo...<\/p><h3 id=\"next-learn-ido-mode\">Next, Learn Ido Mode<\/h3><p>Trust me. Ido makes working with files much easier. You don't generally use a separate Finder|Explorer window to work with files in Emacs. Instead you use the editor's commands to create, open, and save files. This is a bit of a pain without Ido, so I recommend installing it before learning the other way. Ido comes with Emacs beginning with version 22, but you'll want to make some tweaks to your <code>.emacs<\/code> file so that it is always used. This is a good excuse to get your environment set up.<\/p><p>Most features in Emacs come in modes. To install any given mode, you'll need to do two things. Well, at first you'll need to do a few extra things, but these only need to be done once, and thereafter only two things. So the extra things are that you'll need a single place to put all your eLisp files and you'll need to tell Emacs where that place is. I suggest you make a single directory in, say, Dropbox that is your Emacs home. Inside this, you'll want to create an <code>.emacs<\/code> file and an <code>.emacs.d<\/code> directory. Inside the <code>.emacs.d<\/code>, make a directory called <code>lisp<\/code>. So you should have:<\/p><pre>home\n|\n+.emacs\n|\n-.emacs.d\n  |\n  -lisp<\/pre><p>You'll put the <code>.el<\/code> files for things like modes into the <code>home\/.emacs.d\/lisp<\/code> directory, and you'll point to that in your <code>.emacs<\/code> like so:<\/p><p><code>(add-to-list 'load-path \"~\/.emacs.d\/lisp\/\")<\/code><\/p><p>Ido Mode comes with Emacs, so you won't need to put an <code>.el<\/code> file into your Lisp directory for this, but you'll be adding other stuff soon that will go in there.<\/p><h3 id=\"symlinks-are-your-friend\">Symlinks are Your Friend<\/h3><p>But wait, that says that <code>.emacs<\/code> and <code>.emacs.d<\/code> are in your home directory, and we put them in some dumb folder in Dropbox! Correct. This is how you make it easy to have your environment anywhere you go. Keep everything in Dropbox and make symbolic links to <code>.emacs<\/code>, <code>.emacs.d<\/code>, and your main document directories in <code>~<\/code>. On OS X, this is super easy with the <code>ln -s<\/code> command, but on Windows this is a pain. Fortunately, Emacs provides an easy alternative to symlinking on Windows, the HOME environment variable. Go into Environment Variables in Windows (as of Windows 10, you can just hit the Windows key and type \"Environment Variables\" to find this with search, which is the best part of Windows 10), and make a HOME environment variable in your account that points to the Dropbox folder you made for Emacs. If you want to make it easy to navigate to local files that aren't in Dropbox, you may instead want to make a symbolic link to the Dropbox Emacs home in your actual home directory.<\/p><p>So now you've done all the jiggery-pokery needed to get any machine pointed to your Emacs setup and files. If you get a new computer or use someone else's for an hour or a day, you get your entire work environment. This seems a little difficult the first time you do it, but it's about a ten minute (at most) operation once you know what you're doing.<\/p><p>But we were configuring Ido...<\/p><p><code>C-x<\/code> <code>C-f<\/code> and type <code>~\/.emacs RET RET<\/code> to create your <code>.emacs<\/code> file. Add these lines to it:<\/p><pre>;; set up ido mode\n(require `ido)\n(setq ido-enable-flex-matching t)\n(setq ido-everywhere t)\n(ido-mode 1)<\/pre><p>With the <code>.emacs<\/code> buffer open, do an <code>M-x evaluate-buffer<\/code> command, and you'll either get an error if you munged something or you'll get Ido. Ido changes how the minibuffer works when doing file operations. There is great documentation on it, but I'll point out a few tips. Use the <code>~\/<\/code> effectively; you can just type <code>~\/<\/code> at any point in the minibuffer and it'll jump back to home. Implicit in this is that you should have most of your stuff a short hop off your home. I use <code>~\/org<\/code> for all my non-code stuff and <code>~\/code<\/code> for code. Once you\u2019re in the right directory, you'll often have a collection of files with different extensions, especially if you use Org Mode and publish from it. You can type period and the extension you want no matter where you are in the file name and Ido will limit the choices to files with that extension. For example, I'm writing this blog post in Org Mode, so the main file is:<\/p><p><code>~\/org\/blog\/emacs.org<\/code><\/p><p>I also occasionally push it out to HTML using Org Mode publishing, so I've got an <code>emacs.html<\/code> file in the same directory. When I want to open the Org file, I will type:<\/p><p><code>C-x C-f ~\/o[RET]\/bl[RET].or[RET]<\/code><\/p><p>The [RET]s are me hitting return for auto completion for Ido Mode. So, that\u2019s 12 characters typed and, if you're used to it, a <em>lot<\/em> less time than opening Finder|Explorer and clicking around. Ido Mode is plenty useful, but really is a utility mode for operating Emacs. Let's explore some modes that are useful for getting work done.<\/p><h3 id=\"fonts-and-styles\">Fonts and Styles<\/h3><p>I recommend getting the excellent input family of typefaces for use in Emacs. They are customizable with different braces, zeroes, and other characters. You can build in extra line spacing into the font files themselves. I recommend a 1.5X line spacing and using their excellent proportional fonts for code and data. I use Input Serif for my writing, which has a funky but modern feel. You can find them on <a href=\"http:\/\/input.fontbureau.com\/\">http:\/\/input.fontbureau.com\/<\/a> where you can customize to your preferences. You can manually set the fonts using menus in Emacs, but this puts code into your <code>.emacs<\/code> file and, if you use multiple devices, you may find you want some different settings. I've set up my <code>.emacs<\/code> to look for the machine I'm using by name and configure the screen appropriately. The code for this is:<\/p><pre>;; set up fonts for different OSes. OSX toggles to full screen.\n(setq myfont \"InputSerif\")\n(cond\n((string-equal system-name \"Sampo.local\")\n (set-face-attribute 'default nil :font myfont :height 144)\n (toggle-frame-fullscreen))\n((string-equal system-name \"Morpheus.local\")\n (set-face-attribute 'default nil :font myfont :height 144))\n((string-equal system-name \"ILMARINEN\")\n (set-face-attribute 'default nil :font myfont :height 106))\n((string-equal system-name \"UKKO\")\n (set-face-attribute 'default nil :font myfont :height 104)))<\/pre><p>You should replace the <code>system-name<\/code> values with what you get when you evaluate <code>(system-name)<\/code> in your copy of Emacs. Note that on Sampo (my MacBook), I also set Emacs to full screen. I'd like to do this on Windows as well, but Windows and Emacs don't really love each other and it always ends up in some wonky state when I try this. Instead, I just fullscreen it manually after launch.<\/p><p>I also recommend getting rid of the awful toolbar that Emacs got sometime in the 90s when the cool thing to do was to have toolbars in your application. I also got rid of some other \"chrome\" so that I have a simple, productive interface. Add these to your <code>.emacs<\/code> file to get rid of the toolbar and scroll bars, but to keep your menu available (on OS X, it'll be hidden unless you mouse to the top of the screen anyway):<\/p><pre>(if (fboundp 'scroll-bar-mode) (scroll-bar-mode -1))\n(if (fboundp 'tool-bar-mode) (tool-bar-mode -1))\n(if (fboundp 'menu-bar-mode) (menu-bar-mode 1))<\/pre><h3 id=\"org-mode\">Org Mode<\/h3><p>I pretty much live in Org Mode. It is my go-to environment for authoring documents, keeping notes, making to-do lists and 90% of everything else I do. Org was originally conceived as a combination note-taking and to-do list utility by a fellow who is a laptop-in-meetings sort. I am against use of laptops in meetings and don't do it myself, so my use cases are a little different than his. For me, Org is primarily a way to handle all manner of content within a structure. There are heads and subheads, etc., in Org Mode, and they function like an outline. Org allows you to expand or hide the contents of the tree and also to rearrange the tree. This fits how I think very nicely and I find it to be just a pleasure to use in this way.<\/p><p>Org Mode also has a lot of little things that make life pleasant. For example, the footnote handling is excellent and the LaTeX\/PDF output is great. Org has the ability to generate agendas based on the to-do's in all your documents and a nice way to relate them to dates\/times. I don't use this for any sort of external commitments, which are handled on a shared calendar, but for creating things and keeping track of what I need to create in the future, it's invaluable. Installing it is as easy as adding the <code>org-mode.el<\/code> to your Lisp directory and adding these lines to your <code>.emacs<\/code>, if you want it to indent based on tree location and to open documents fully expanded:<\/p><pre>;; set up org mode\n(setq org-startup-indented t)\n(setq org-startup-folded \"showall\")\n(setq org-directory \"~\/org\")<\/pre><p>The last line is there so that Org knows where to look for files to include in agendas and some other things. I keep Org right in my home directory, i.e., a symlink to the directory that lives in Dropbox, as described earlier.<\/p><p>I have a <code>stuff.org<\/code> file that is always open in a buffer. I use it like a notepad. Org makes it easy to extract things like TODOs and stuff with deadlines. It's especially useful when you can inline Lisp code and evaluate it whenever you need. Having code with content is super handy. Again, you have access to the actual computer with Emacs, and this is a liberation.<\/p><h4 id=\"publishing-with-org-mode\">Publishing with Org Mode<\/h4><p>I care about the appearance and formatting of my documents. I started my career as a designer, and I think information can and should be presented clearly and beautifully. Org has great support for generating PDFs via LaTeX, which has a bit of its own learning curve, but doing simple things is pretty easy.<\/p><p>If you want to use fonts and styles other than the typical LaTeX ones, you've got a few things to do. First, you'll want XeLaTeX so you can use normal system fonts rather than LaTeX specific fonts. Next, you'll want to add this to <code>.emacs<\/code>:<\/p><pre>(setq org-latex-pdf-process\n '(\"xelatex -interaction nonstopmode %f\"\n  \"xelatex -interaction nonstopmode %f\"))<\/pre><p>I put this right at the end of my Org section of <code>.emacs<\/code> to keep things tidy. This will allow you to use more formatting options when publishing from Org. For example, I often use:<\/p><pre>#+LaTeX_HEADER: \\usepackage{fontspec}\n#+LATEX_HEADER: \\setmonofont[Scale=0.9]{Input Mono}\n#+LATEX_HEADER: \\setromanfont{Maison Neue}\n#+LATEX_HEADER: \\linespread{1.5}\n#+LATEX_HEADER: \\usepackage[margin=1.25in]{geometry}\n#+TITLE: Document Title Here<\/pre><p>These simply go somewhere in your <code>.org<\/code> file. Our corporate font for body copy is Maison Neue, but you can put whatever is appropriate here. I strongly discourage the use of Maison Neue. It\u2019s a terrible font and no one should ever use it.<\/p><p>This file is an example of PDF output using these settings. This is what out-of-the-box LaTeX always looks like. It's fine I suppose, but the fonts are boring and a little odd. Also, if you use the standard format, people will assume they are reading something that is or pretends to be an academic paper. You've been warned.<\/p><h3 id=\"ace-jump-mode\">Ace Jump Mode<\/h3><p>This is more of a gem than a major feature, but you want it. It works a bit like Jef Raskin's Leap feature from days gone by.<sup id=\"canoncatOrigin\"><a href=\"#canoncat\">9<\/a><\/sup> The way it works is you type <code>C-c<\/code> <code>C-SPC<\/code> and then type the first letter of the word you want to jump to. It highlights all occurrences of words with that initial character, replacing it with a letter of the alphabet. You simply type the letter of the alphabet for the location you want and your cursor jumps to it. I find myself using this as often as the more typical nav keys or search. Download the <code>.el<\/code> to your Lisp directory and put this in your <code>.emacs<\/code>:<\/p><pre>;; set up ace-jump-mode\n(add-to-list 'load-path \"which-folder-ace-jump-mode-file-in\/\")\n(require 'ace-jump-mode)\n(define-key global-map (kbd \"C-c C-SPC\" ) 'ace-jump-mode)<\/pre><h2 id=\"more-later\">More Later<\/h2><p>That's enough for one post\u2014this may get you somewhere you'd like to be. I'd love to hear about your uses for Emacs aside from programming (or for programming!) and whether this was useful at all. There are likely some boneheaded PHBisms in how I use Emacs, and if you want to point them out, I'd appreciate it. I'll probably write some updates over time to introduce additional features or modes. I'll certainly show you how to use Fugue with Emacs and Ludwig-mode as we evolve it into something more useful than code highlighting. Send your thoughts to <a href=\"https:\/\/twitter.com\/fugueHQ\">@fugueHQ<\/a> on Twitter.<\/p><hr \/><h4 id=\"footnotes\">Footnotes<\/h4><ol class=\"footnotes\"><li id=\"phb\"><a href=\"#phbOrigin\">^<\/a> If you are now a PHB of some sort, but were never technical, Emacs likely isn\u2019t for you. There may be a handful of folks for whom Emacs will form a path into the more technical aspects of computing, but this is probably a small population. It\u2019s helpful to know how to use a Unix or Windows terminal, to have edited a dotfile or two, and to have written some code at some point in your life for Emacs to make much sense.<\/li>\n    <li id=\"tufte\"><a href=\"#tufteOrigin\">^<\/a> <a href=\"http:\/\/archive.wired.com\/wired\/archive\/2.08\/tufte.html\">http:\/\/archive.wired.com\/wired\/archive\/2.08\/tufte.html<\/a><\/li>\n    <li id=\"interpreter\"><a href=\"#interpreterOrigin\">^<\/a> I mainly use this to perform calculations while writing. For example, I was writing an offer letter to a new employee and wanted to calculate how many options to include in the offer. Since I have a variable defined in my <code>.emacs<\/code> for outstanding-shares, I can simply type <code>M-: (* .001 outstanding-shares)<\/code> and get a tenth of a point without opening a calculator or spreadsheet. I keep <em>lots<\/em> of numbers in variables like this so I can avoid context switching.<\/li>\n    <li id=\"eww\"><a href=\"#ewwOrigin\">^<\/a> The missing piece of this is the web. There is an Emacs web browser called eww that will allow you to browse in Emacs. I actually use this, as it is both a great ad-blocker and removes most of the poor choices in readability from the web designer's hands. It's a bit like Reading Mode in Safari. Unfortunately, most websites have lots of annoying cruft and navigation that translates poorly into text.<\/li>\n    <li id=\"usability\"><a href=\"#usabilityOrigin\">^<\/a> Usability is often confused with learnability. Learnability is how difficult it is to learn a tool. Usability is how useful the tool is. Often, these are at odds, such as with the mouse and menus. Menus are highly learnable, but have poor usability, so there have been keyboard shortcuts from the earliest days. Raskin was right on many points where he was ignored about GUIs in general. Now, OSes are putting things like decent search onto a keyboard shortcut. On OS X and Windows, my default method of navigation is search. Ubuntu's search is badly broken, as is the rest of its GUI.<\/li>\n    <li id=\"s3\"><a href=\"#s3Origin\">^<\/a> AWS S3 has effectively solved file storage for as long as we have the Internet. Trillions of objects are stored in S3 and they've never lost one of them. Most every service out there that offers cloud storage is built on S3 or imitates it. No one has the scale of S3, so I keep important stuff there, via Dropbox.<\/li>\n    <li id=\"bicycles\"><a href=\"#bicyclesOrigin\">^<\/a> By now, you might be thinking \"what is it with this guy and bicycles?\" ... I love them on every level. They are the most mechanically efficient form of transportation ever invented. They can be objects of real beauty. And, with some care, they can last a lifetime. I had Rivendell Bicycle Works build a frame for me back in 2001 and it still makes me happy every time I look at it. Bicycles and UNIX are the two best inventions I've interacted with. Well, they and Emacs.<\/li>\n    <li id=\"nottutorial\"><a href=\"#nottutorialOrigin\">^<\/a> This is not a tutorial for Emacs. It comes with one and it's excellent. I do walk through some of the things that I find most important to getting a useful Emacs setup, but this is not a replacement in any way.<\/li>\n    <li id=\"canoncat\"><a href=\"#canoncatOrigin\">^<\/a> Jef Raskin designed the Canon Cat computer in the 1980s after falling out with Steve Jobs on the Macintosh project, which he originally led. The Cat had a document-centric interface (as all computers should) and used the keyboard in innovative ways that you can now imitate with Emacs. If I could have a modern, powerful Cat with a giant high-res screen and Unix underneath, I'd trade my Mac for it right away. <a href=\"https:\/\/youtu.be\/o_TlE_U_X3c?t=19s\">https:\/\/youtu.be\/o_TlE_U_X3c?t=19s<\/a><\/li>\n<\/ol><div class=\"shifty shifty--visible-base shifty--visible-small\" data-set=\"shifty-interrupt\"><div class=\"text-longform__interrupt text-longform__interrupt--align-right text-longform__interrupt--size-small shifty__item\"><aside class=\"supplement-blog text-longform__interrupt-item\"><header class=\"supplement-blog__header\"><h2>Related Posts<\/h2>\n                                                            <\/header><\/aside><\/div><\/div>","created_at":"2018-11-07T16:40:33+0000","updated_at":"2018-11-07T16:41:20+0000","annotations":[],"mimetype":"text\/html","reading_time":26,"domain_name":"blog.fugue.co","preview_picture":"https:\/\/blog.fugue.co\/assets\/images\/opengraph.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":243,"title":"Pinching the best bits from Spacemacs","url":"https:\/\/medium.com\/@CBowdon\/pinching-the-best-bits-from-spacemacs-869b8c793ad3","content":"<div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"7384\" class=\"graf graf--h3 graf--leading graf--title\">Pinching the best bits from Spacemacs<\/h1><p id=\"b6db\" class=\"graf graf--p graf-after--h3\">Have you heard the term \u201cVim refugee\u201d? There are those of use whose brains are hard-wired to hit ESC when we pause for breath, yet have had enough of VimScript. The idea of customising your editor with Lisp is pretty damn appealing \u2013 as is the extensive package ecosystem. So a good option is to make the switch to Emacs with Evil mode (blah blah converting the heathens blah blah blah).<\/p><p id=\"661e\" class=\"graf graf--p graf-after--p\">For the uninitiated, Evil mode is a Vim emulation package that is regarded as the most faithful available for any other editor. This is not just adding a normal mode and binding some movement keys, like most emulations. In Evil I can do things like \u201cC-v a t d\u201d to delete an XML node, \u201c.\u201d to repeat an action or \u201c:v\/evil\/d\u201d to delete all lines not containing \u201cevil\u201d. I haven\u2019t seen any emulation that comes close. It\u2019s a truly incredible effort.<\/p><p id=\"24c9\" class=\"graf graf--p graf-after--p\">The downside as a Vim refugee, is the boundary between the comfy Evil bubble and Emacs-land. As you explore various exciting new features (Org mode! Eshell! Ediff! Many more!) you\u2019ll inevitably get dumped into a buffer where Evil conflicts with the default key bindings for that buffer and instead of moving the cursor you accidentally delete something. A sort of frightening shadow dimension where everything you think you know is wrong.<\/p><p id=\"b412\" class=\"graf graf--p graf-after--p\">Spacemacs extends this boundary a lot, providing correct keybindings for many modes out of the box, and great discoverability by popping up a buffer with options when you start typing a command. It even has its own layers system for customisation, so you barely need to know about packages \u2013 a ton of the best ones are already included for you. The \u201cstock\u201d Emacs experience is hidden, most of the time.<\/p><p id=\"7d76\" class=\"graf graf--p graf-after--p\">I was a happy Spacemacs user for 6 months (even contributed a layer) but the more I explored, the more I kept running into that boundary. Searching the web to understand new functionality or error messages would always turn up far more help for stock Emacs than Spacemacs.<\/p><p id=\"704b\" class=\"graf graf--p graf-after--p\">So it was necessary to learn the basics: navigation, editing, customisation. And at that point, it felt like Spacemacs was an unnecessary layer of abstraction, so I decided to start my own init.el from scratch with just the following selected bits Spacemacs had introduced me to.<\/p><h4 id=\"9240\" class=\"graf graf--h4 graf-after--p\">Evil-leader<\/h4><p id=\"1e73\" class=\"graf graf--p graf-after--h4\">This allows you to define a leader key that initiates a key sequence \u2013 the \u201cspace\u201d in Spacemacs and usually mapped to \u201c,\u201d in Vim. You can then set up keybindings that are effectively tucked away under this key. The sequences can be arbitrarily long, a fact which you can use to group them. For example, all my shell-related commands are under \u201c, s\u201d. You can also define keys that work only in a particular mode, to keep your key space clean.<\/p><p id=\"da0b\" class=\"graf graf--p graf-after--p\">Rebinding everything is an uphill battle though \u2013 it\u2019s only worth the effort to rebind frequently-used keys. Having an RSI-inducing binding like \u201cC-x C-c v\u201d or what-have-you doesn\u2019t matter much if you use it once a week.<\/p><h4 id=\"5a34\" class=\"graf graf--h4 graf-after--p\">Which-key<\/h4><p id=\"2768\" class=\"graf graf--p graf-after--h4\">Which-key is the engine of Spacemacs\u2019 discoverability \u2013 if you initiate a key sequence (both Evil or Emacs style) it will pop up a buffer to show what\u2019s available to complete your command. Handy!<\/p><h4 id=\"a1d9\" class=\"graf graf--h4 graf-after--p\">Use-package<\/h4><p id=\"49ef\" class=\"graf graf--p graf-after--h4\">This is a single powerful macro for managing packages in Lisp. It abstracts functions like installing and checking for presence. Currently an external package, but rumour has it this will be folded into Emacs core in the near future.<\/p><h4 id=\"2671\" class=\"graf graf--h4 graf-after--p\">Magit<\/h4><p id=\"b390\" class=\"graf graf--p graf-after--h4\">This just provides a really nice and pretty comprehensive interface to Git. That said, on some systems I haven\u2019t been able to install it and resorted to the built-in VC package. It turns out VC is no slouch either. It doesn\u2019t seem to cover more advanced Git commands and you have to learn some new jargon (e.g. mark\/register for adding to index), but it\u2019s pretty close to equivalent to Magit for day to day stuff.<\/p><h4 id=\"775a\" class=\"graf graf--h4 graf-after--p\">Copying Spacemacs ideas<\/h4><p id=\"1705\" class=\"graf graf--p graf-after--h4\">I also copied a couple of neat ideas from Spacemacs:<\/p><ul class=\"postList\"><li id=\"4e86\" class=\"graf graf--li graf-after--p\">A shortcut to edit my init file (a literate init file with Org mode, which is great).<\/li><li id=\"a80b\" class=\"graf graf--li graf-after--li\">Using a mnemonic key mapping, e.g. keeping VC commands under leader-v and shell commands under leader-s.<\/li><li id=\"7143\" class=\"graf graf--li graf-after--li\">I reference Spacemacs to get ideas for language-specific packages to install.<\/li><li id=\"a575\" class=\"graf graf--li graf-after--li\">One key departure from Spacemacs: adding special modes like dired and VC to the evil-emacs-state-map, so that it uses the (usually consistent and mnemonic) default key bindings for those modes instead of mixing with Evil unpredictably.<\/li><\/ul><p id=\"5774\" class=\"graf graf--p graf-after--li\">The resulting Emacs setup is very stripped-down compared to Spacemacs, and nippier because of it \u2013 I\u2019ve left out some of the heavyweights like Helm and Projectile. But it still covers 90% of what was useful (to me) in Spacemacs and assembling it over time has greatly increased my Emacs-fu.<\/p><p id=\"e044\" class=\"graf graf--p graf-after--p\">It\u2019s been a fun and productive journey, so I\u2019d recommend it to anyone who\u2019s curious.<\/p><p id=\"9e69\" class=\"graf graf--p graf-after--p graf--trailing\">If you\u2019re interested, my init file is <a href=\"https:\/\/github.com\/cbowdon\/Config\/blob\/master\/emacs\/init.org\" data-href=\"https:\/\/github.com\/cbowdon\/Config\/blob\/master\/emacs\/init.org\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">here<\/a>.<\/p><\/div><\/div>","created_at":"2018-11-07T16:40:33+0000","updated_at":"2018-11-07T16:41:26+0000","published_at":"2017-03-24T08:04:24+0000","published_by":["Chris Bowdon"],"annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"medium.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":238,"title":"cbowdon\/Config","url":"https:\/\/github.com\/cbowdon\/Config\/blob\/master\/emacs\/init.org","content":"<p>Get packages from ELPA, MELPA <del>and org<\/del> by default. This can be customized (e.g. to use a local checkout of <code>spacemacs-elpa-mirror<\/code> by customizing <code>package-archives<\/code>).<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'package)\n(setq package-archives\n      '((\"gnu\" . \"https:\/\/elpa.gnu.org\/packages\/\")\n        (\"melpa\" . \"https:\/\/melpa.org\/packages\/\")\n        (\"melpa-stable\" . \"https:\/\/stable.melpa.org\/packages\/\")))\n(package-initialize)<\/pre><\/div>\n<p>GPG is not necessarily installed on Windows.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(when (eq system-type 'nt)\n  (setq package-check-signature nil))<\/pre><\/div>\n<p>From here-on out <code>use-package<\/code> loads packages. We almost always want to <code>:ensure<\/code> packages are present.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(unless (package-installed-p 'use-package)\n  (package-refresh-contents)\n  (package-install 'use-package))\n(require 'use-package)\n(setq use-package-always-ensure t)<\/pre><\/div>\n<p>UTF-8 everywhere, but don\u2019t mess with line-endings.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(if (eq system-type 'windows-nt)\n    (prefer-coding-system 'utf-8-dos)\n  (prefer-coding-system 'utf-8-unix))<\/pre><\/div>\n<p>It\u2019s also handy to programmatically convert files, mainly when Windows has UTF-16\u2019d something.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun convert-file-encoding (file coding-system)\n  \"Convert FILE to CODING-SYSTEM without visiting it.\"\n  (with-temp-buffer\n    (insert-file-contents file)\n    (let ((coding-system-for-write coding-system))\n      (write-region nil nil file))))<\/pre><\/div>\n<p>It\u2019s convenient to be able to get at this quickly.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun init-file ()\n  \"Return the path to my init file.\"\n  (if (eq system-type 'windows-nt)\n      \"C:\/Cygwin64\/home\/chris.bowdon\/init.org\"\n    \"~\/init.org\"))\n(defun find-init-file ()\n  (interactive)\n  (find-file (init-file)))\n(global-set-key (kbd \"C-x .\") 'find-init-file)<\/pre><\/div>\n<p>I prefer not to clutter <code>.emacs<\/code> with customizations, so create an <code>init-custom.el<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq custom-file \"~\/init-custom.el\")\n(unless (file-exists-p custom-file)\n  (with-temp-buffer\n    (write-file custom-file)))\n(load-file custom-file)<\/pre><\/div>\n<p>It\u2019s also convenient to allow loading some custom scripts per host, e.g. for work or home-specific setup.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defcustom init-file-extra-files nil\n  \"Extra files to load with init.org\"\n  :type '(repeat (file :must-match t)))\n(defun init-file-load-extra-files ()\n  \"Loads each .el or .org in init-file-extra-files\"\n  (interactive)\n  (dolist (extra-file init-file-extra-files)\n    (message \"Extra init file: %s\" extra-file)\n    (pcase (file-name-extension extra-file)\n      (\"el\" (load-file extra-file))\n      (\"org\" (org-babel-load-file extra-file))\n      (_ (message \"I have no idea what this is: %s\" extra-file)))))\n(init-file-load-extra-files)<\/pre><\/div>\n<p>I\u2019ve hacked together a module system so that most of these sections are opt-in. At the moment there is a macro <code>defmodule<\/code> that defines this, but it would be nice if I can modify <code>org-babel-tangle<\/code> so that it turns each section into a module (and automatically turns the text into module documentation).<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defgroup init-modules nil\n  \"The init modules customization group.\")\n(defvar init-modules-defs-plist nil\n  \"The module definitions plist.\")\n(defun init-module-load (name)\n  \"Load the module NAME.\"\n  (let ((exec-module (plist-get init-modules-defs-plist name)))\n    (message \"Init-module-load %s: %s\" name exec-module)\n    (when exec-module\n      (message \"Found def for %s\" name)\n      (funcall exec-module))))\n(defun init-module-set (name value)\n  \"Enable\/disable the module NAME by setting boolean VALUE.\"\n  (set-default name value)\n  (when value\n    (message \"%s set: evaluating\" name)\n    (init-module-load name)))\n(defmacro defmodule (name doc &amp;rest body)\n  \"Define an init module called NAME that executes BODY when enabled.\"\n  (message \"Defining module for %s\" name)\n  (let ((module-name (intern (format \"init-module-%s\" name))))\n    ;; Declare var if not already\n    (when (not (boundp module-name))\n      (custom-declare-variable module-name nil\n                               doc\n                               :group 'init-modules\n                               :set 'init-module-set\n                               :tag (symbol-name name)\n                               :type '(boolean)))\n    `(progn\n       ;; Stash the definition in the symbol's plist\n       (setq init-modules-defs-plist\n             (plist-put init-modules-defs-plist\n                        (quote ,module-name)\n                        (lambda ()\n                          (progn\n                            (message \"Evaluating definition of %s\" ,module-name)\n                            ,@body))))\n       ;; If set, evaluate it\n       (when ,module-name\n         (message \"Evaluating %s on first def\" (quote ,module-name))\n         ,@body))))<\/pre><\/div>\n<h2><a id=\"user-content-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#packages\"><\/a>Packages<\/h2>\n<p>Join the dark side.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package evil\n  :demand\n  :config (progn\n            (evil-mode 1)\n            (define-key evil-normal-state-map (kbd \";\") 'evil-ex)\n            (define-key evil-visual-state-map (kbd \";\") 'evil-ex)))\n(use-package evil-leader\n  :demand\n  :config (progn\n            (setq evil-leader\/in-all-states t)\n            (global-evil-leader-mode)))\n(setq-default indent-tabs-mode nil)<\/pre><\/div>\n<p>AND BECOME A GOD!<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package god-mode :demand)\n(use-package evil-god-state :demand)\n;; Try to remove C-w\n(global-unset-key (kbd \"C-w\"))\n(define-key global-map (kbd \"C-w\") nil)\n(define-key global-map (kbd \"C-&lt;escape&gt;\") 'evil-normal-state)\n(define-key global-map (kbd \"C-~\") 'evil-normal-state)\n(define-key global-map (kbd \"M-&lt;escape&gt;\") 'god-mode)\n(define-key global-map (kbd \"C-M-&lt;escape&gt;\") 'god-local-mode)\n(define-key evil-normal-state-map (kbd \"SPC\") 'evil-execute-in-god-state)\n(define-key evil-visual-state-map (kbd \"SPC\") 'evil-execute-in-god-state)<\/pre><\/div>\n<p>This evil\/god combination should be the most ergonomic possible without manually remapping most bindings.<\/p>\n<p>I\u2019ve bound <code>C-&lt;escape&gt;<\/code> and <code>C-~<\/code> in particular because in <code>term-mode<\/code> it\u2019s possible to get stuck in <code>evil-god-state<\/code> if you hit some combinations (I prefer <code>C-&lt;escape&gt;<\/code> but Windows intercepts this).<\/p>\n<p><a href=\"https:\/\/github.com\/chrisdone\/god-mode\">Basic <code>god-mode<\/code> usage:<\/a><\/p>\n<ul><li><code>abc<\/code> \u2192 <code>C-a C-b C-c<\/code><\/li>\n  <li><code>ab c<\/code> \u2192 <code>C-a C-b c<\/code><\/li>\n  <li><code>gf<\/code> \u2192 <code>M-f<\/code><\/li>\n  <li><code>Gf<\/code> \u2192 <code>C-M-f<\/code><\/li>\n  <li><code>12f<\/code> \u2192 <code>M-12 C-f<\/code><\/li>\n  <li><code>gf..<\/code> \u2192 <code>M-f M-f M-f<\/code><\/li>\n  <li><code>uco<\/code> \u2192 <code>C-u C-c C-o<\/code><\/li>\n<\/ul><p>It\u2019s nice to have a guide\u2026<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package which-key :demand)\n(which-key-mode)\n(which-key-enable-god-mode-support)<\/pre><\/div>\n<h2><a id=\"user-content-state-modes\" class=\"anchor\" aria-hidden=\"true\" href=\"#state-modes\"><\/a>State modes<\/h2>\n<p>I prefer not to clobber the default bindings with evil bindings in some modes.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'evil-emacs-state-modes 'dired-mode)\n(add-to-list 'evil-emacs-state-modes 'proced-mode)\n(add-to-list 'evil-emacs-state-modes 'eshell-mode)\n(add-to-list 'evil-emacs-state-modes 'term-mode)\n(add-to-list 'evil-emacs-state-modes 'diff-mode)\n(add-to-list 'evil-emacs-state-modes 'vc-mode)<\/pre><\/div>\n<h2><a id=\"user-content-keyboard-mapping\" class=\"anchor\" aria-hidden=\"true\" href=\"#keyboard-mapping\"><\/a>Keyboard mapping<\/h2>\n<p>Use tilde for leader key until I\u2019m sure it\u2019s no longer necessary.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-leader\/set-leader \"`\")\n(evil-leader\/set-key\n  \"\/\" 'comment-region\n  \"\\\\\\\\\" 'uncomment-region\n  \".\" 'find-init-file\n  \"W\" 'toggle-truncate-lines\n  \"X\" 'delete-trailing-whitespace)<\/pre><\/div>\n<p>Set CAPSLOCK as another ESC. In GNOME 3 and MacOS use the system settings GUI. In X, use the function below. In Windows\u2026 screw around with the registry?<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun set-x-caps-escape ()\n  \"Set CAPSLOCK to be another ESC key in X.\"\n  (interactive)\n  (shell-command \"setxkbmap -option caps:escape\"))<\/pre><\/div>\n<p>For future reference, find <code>rules\/base.lst<\/code> in the <code>xkb<\/code> directory to learn all the options. (See the man page.)<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(global-unset-key (kbd \"C-f\"))\n(global-unset-key (kbd \"C-b\"))\n(define-key evil-normal-state-map (kbd \"C-f\") nil)\n(define-key evil-visual-state-map (kbd \"C-f\") nil)\n(define-key evil-normal-state-map (kbd \"C-b\") nil)\n(define-key evil-visual-state-map (kbd \"C-b\") nil)\n(define-key global-map (kbd \"C-f C-s\") 'save-buffer)\n(define-key global-map (kbd \"C-b\") 'switch-to-buffer)\n(define-key evil-normal-state-map (kbd \"C-k\") 'kill-buffer)\n(define-key evil-visual-state-map (kbd \"C-k\") 'kill-buffer)\n(define-key evil-god-state-map (kbd \"C-k\") 'kill-buffer)\n(define-key evil-normal-state-map (kbd \"C-w\") 'evil-window-map)\n(define-key evil-visual-state-map (kbd \"C-w\") 'evil-window-map)\n(define-key evil-god-state-map (kbd \"C-w\") 'evil-window-map)\n(define-key global-map (kbd \"C-S-X\") 'delete-trailing-whitespace)\n(define-key global-map (kbd \"C-S-W\") 'toggle-truncate-lines)\n(define-key global-map (kbd \"C-S-P\") 'proced)<\/pre><\/div>\n<h2><a id=\"user-content-repl-shortcuts\" class=\"anchor\" aria-hidden=\"true\" href=\"#repl-shortcuts\"><\/a>REPL shortcuts<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar repl-map (make-sparse-keymap))\n(define-key repl-map (kbd \"C-c\") 'cider-jack-in)\n(define-key repl-map (kbd \"C-d\") 'run-dig)\n(define-key repl-map (kbd \"C-e\") 'eshell)\n(define-key repl-map (kbd \"C-i\") 'ielm)\n(define-key repl-map (kbd \"C-p\") 'run-python)\n(define-key repl-map (kbd \"C-s C-m\") 'sql-mysql)\n(define-key repl-map (kbd \"C-s C-p\") 'sql-postgres)\n(define-key repl-map (kbd \"C-s C-s\") 'sql-sqlite)\n(define-key repl-map (kbd \"C-t\") 'term)\n(evil-leader\/set-key\n  \"rd\" 'run-dig ;; not exactly a REPL, but fits nonetheless\n  \"rf\" 'run-fsharp\n  \"ri\" 'ielm\n  \"rp\" 'run-python\n  \"rr\" 'run-ruby\n  \"R\" 'repl-map)\n(global-unset-key (kbd \"C-r\"))\n(define-key global-map (kbd \"C-r\") nil)\n(define-key global-map (kbd \"C-r\") repl-map)\n(define-key evil-god-state-map (kbd \"C-r\") repl-map)<\/pre><\/div>\n<h2><a id=\"user-content-mode-shortcuts\" class=\"anchor\" aria-hidden=\"true\" href=\"#mode-shortcuts\"><\/a>Mode shortcuts<\/h2>\n<p>Sometimes the right mode isn\u2019t picked up by buffer name\/shebang.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-leader\/set-key\n  \"md\" 'markdown-mode\n  \"me\" 'ensime ;; ensime is slow, don't like to autoload\n  \"mh\" 'html-mode\n  \"mj\" 'javascript-mode\n  \"mnc\" 'column-number-mode\n  \"mnl\" 'linum-mode\n  \"mo\" 'org-mode\n  \"ms\" 'shell-script-mode\n  \"mx\" 'nxml-mode)<\/pre><\/div>\n<p>Company mode is my preferred auto-completion package because it seems to be best supported by the languages I use.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package company\n  :demand\n  :config (progn\n            (add-hook 'prog-mode-hook #'(lambda () (company-mode)))\n            (setq company-show-numbers t)))<\/pre><\/div>\n<p>Trying out Ido for Emacs command completion. Good for buffers so far.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(ido-mode)<\/pre><\/div>\n<h2><a id=\"user-content-general\" class=\"anchor\" aria-hidden=\"true\" href=\"#general\"><\/a>General<\/h2>\n<p>The Language Server Protocol (lsp) packages are rather good and worth setting up initially.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package lsp-mode)\n(use-package flycheck :config (add-hook 'prog-mode-hook 'flycheck-mode)) ;; used by lsp-ui for fancy displays\n(use-package lsp-ui :config (add-hook 'lsp-mode-hook 'lsp-ui-mode))\n(use-package company-lsp :config (push 'company-lsp company-backends))<\/pre><\/div>\n<p>By default every text editor should display line and column number, and not wrap text.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq-default truncate-lines t)\n(add-hook 'prog-mode-hook 'column-number-mode)\n(defun enable-line-numbers ()\n  (setq display-line-numbers t))\n(if (&lt; emacs-major-version 26)\n    (add-hook 'prog-mode-hook 'linum-mode)\n  (add-hook 'prog-mode-hook 'enable-line-numbers))<\/pre><\/div>\n<p>Electric indent interferes with lots of modes\u2019 own indenting, so disable it.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq electric-indent-inhibit t)<\/pre><\/div>\n<p><code>popup-imenu<\/code> is a nice document-overview tool.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package popup-imenu)\n(evil-leader\/set-key \"p\" 'popup-imenu)\n(define-key evil-normal-state-map (kbd \"C-p\") nil)\n(define-key evil-visual-state-map (kbd \"C-p\") nil)\n(define-key evil-god-state-map (kbd \"C-p\") nil)\n(define-key evil-normal-state-map (kbd \"C-p\") 'popup-imenu)\n(define-key evil-visual-state-map (kbd \"C-p\") 'popup-imenu)\n(define-key evil-god-state-map (kbd \"C-p\") 'popup-imenu)<\/pre><\/div>\n<p>I\u2019m slowly warming up to Projectile.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package projectile\n  :config (define-key global-map (kbd \"C-x C-p\") 'projectile-command-map))\n(projectile-mode)<\/pre><\/div>\n<h2><a id=\"user-content-emacs-lisp\" class=\"anchor\" aria-hidden=\"true\" href=\"#emacs-lisp\"><\/a>Emacs Lisp<\/h2>\n<p>Nobody likes dynamic binding by default.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq lexical-binding t)<\/pre><\/div>\n<h3><a id=\"user-content-helper-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#helper-functions\"><\/a>Helper functions<\/h3>\n<p>It\u2019s often handy to know if we\u2019re using *nix.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun is-nix ()\n  (or (equal system-type 'gnu)\n      (equal system-type 'gnu\/linux)\n      (equal system-type 'gnu\/kfreebsd)\n      (equal system-type 'darwin)))<\/pre><\/div>\n<p>A handy timer macro.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmacro time-sexp (body)\n  \"Run the BODY s-expression(s) and print the time between start and finish.\"\n  `(let ((t0 (float-time))\n         (result (progn ,body))\n         (t1 (float-time)))\n     (with-current-buffer (pop-to-buffer \"*time-sexp*\" nil t)\n       (goto-char (point-max))\n       (insert\n        (format \"time-sexp: %s\\n\" (quote ,body))\n        (format \"--&gt; %fs\\n\" (- t1 t0))))\n     result))<\/pre><\/div>\n<h3><a id=\"user-content-threading-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#threading-macros\"><\/a>Threading macros<\/h3>\n<p>Emacs 25 brought some mod cons to the language that are worth shimming in older Emacsen.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun emacs-version-less-than (major-number)\n  (and\n   (string-match \"\\\\([0-9]+\\\\)\\\\.[0-9]+.*\" emacs-version)\n   (&gt; major-number (string-to-number (match-string 1 emacs-version)))))\n(when (emacs-version-less-than 25)\n  (use-package seq )\n  (defmacro internal--thread-argument (first? &amp;rest forms)\n    \"Internal implementation for `thread-first' and `thread-last'.\nWhen Argument FIRST? is non-nil argument is threaded first, else\nlast.  FORMS are the expressions to be threaded.\"\n    (pcase forms\n      (`(,x (,f . ,args) . ,rest)\n       `(internal--thread-argument\n         ,first? ,(if first? `(,f ,x ,@args) `(,f ,@args ,x)) ,@rest))\n      (`(,x ,f . ,rest) `(internal--thread-argument ,first? (,f ,x) ,@rest))\n      (_ (car forms))))\n  (defmacro thread-first (&amp;rest forms)\n    \"Thread FORMS elements as the first argument of their succesor.\nExample:\n    (thread-first\n      5\n      (+ 20)\n      (\/ 25)\n      -\n      (+ 40))\nIs equivalent to:\n    (+ (- (\/ (+ 5 20) 25)) 40)\nNote how the single `-' got converted into a list before\nthreading.\"\n    (declare (indent 1)\n             (debug (form &amp;rest [&amp;or symbolp (sexp &amp;rest form)])))\n    `(internal--thread-argument t ,@forms))\n  (defmacro thread-last (&amp;rest forms)\n    \"Thread FORMS elements as the last argument of their succesor.\nExample:\n    (thread-last\n      5\n      (+ 20)\n      (\/ 25)\n      -\n      (+ 40))\nIs equivalent to:\n    (+ 40 (- (\/ 25 (+ 20 5))))\nNote how the single `-' got converted into a list before\nthreading.\"\n    (declare (indent 1) (debug thread-first))\n    `(internal--thread-argument nil ,@forms)))<\/pre><\/div>\n<h3><a id=\"user-content-comfort\" class=\"anchor\" aria-hidden=\"true\" href=\"#comfort\"><\/a>Comfort<\/h3>\n<p>Make ad-hoc lisping more comfortable.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package paredit)\n(use-package flycheck)\n(add-hook 'emacs-lisp-mode-hook 'prettify-symbols-mode)\n(add-hook 'emacs-lisp-mode-hook 'paredit-mode)\n(add-hook 'emacs-lisp-mode-hook 'eldoc-mode)\n(add-hook 'emacs-lisp-mode-hook 'flycheck-mode)\n(evil-leader\/set-key \"xE\" 'eval-buffer)\n;; Create some vimmish bindings for paredit functions\n(evil-define-key 'normal 'evil-normal-state-map\n  (kbd \",dd\") 'paredit-kill\n  (kbd \",dw\") 'paredit-forward-kill-word\n  (kbd \",dB\") 'paredit-backward-kill-word\n  (kbd \",l\") 'paredit-forward\n  (kbd \",h\") 'paredit-backward\n  (kbd \",k\") 'paredit-backward-up\n  (kbd \",j\") 'paredit-forward-down\n  (kbd \",J\") 'paredit-wrap-round\n  (kbd \",K\") 'paredit-splice-sexp)\n(defun change-brackets (type)\n  \"Change |(..) to |[..]. | is point position.\"\n  (cond ((eq type 'square) (paredit-open-square 1))\n        ((eq type 'curly) (paredit-open-curly 1))\n        ((eq type 'angled) (paredit-open-angled 1))\n        ((eq type 'round) (paredit-open-round 1))\n        ((eq type 'bracket) (paredit-open-bracket 1))\n        ((eq type 'parenthesis (paredit-open-parenthesis 1))))\n  (right-char 1)\n  (paredit-splice-sexp)\n  (left-char 1))\n(evil-define-key 'normal 'evil-normal-state-map\n  (kbd \",[\") #'(lambda () (interactive) (change-brackets 'square))\n  (kbd \",(\") #'(lambda () (interactive) (change-brackets 'round))\n  (kbd \",{\") #'(lambda () (interactive) (change-brackets 'curly)))<\/pre><\/div>\n<h3><a id=\"user-content-package-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#package-development\"><\/a>Package development<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package flycheck-package)\n(eval-after-load 'flycheck '(flycheck-package-setup))\n(use-package package-lint)<\/pre><\/div>\n<h3><a id=\"user-content-rest\" class=\"anchor\" aria-hidden=\"true\" href=\"#rest\"><\/a>REST<\/h3>\n<p>It\u2019s nice to wrap the built-in <code>url<\/code> functions into a higher-level API.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar rest--default-headers '(\"DNT\" . \"1\")\n  \"The default headers include a DNT.\")\n(defun rest-- (method url body &amp;rest headers)\n  \"Make an HTTP METHOD request to URL with BODY and optional HEADERS.\nShows the result in a new buffer.\"\n  (let ((url-request-method (upcase method))\n        (url-request-extra-headers headers)\n        (url-request-data body))\n    (message (format \"%s %s\" (upcase method) url))\n    (url-retrieve url\n                  (lambda (status)\n                    (when status\n                      (message (format \"RESPONSE STATUS %s &lt;-- %s\" status url)))\n                    (rename-buffer \"*rest response*\" t)\n                    (switch-to-buffer (current-buffer))))))\n(cl-defun rest (&amp;key url (method \"GET\") (body nil) (headers rest--default-headers))\n  \"Make an HTTP METHOD request to URL with BODY and HEADERS.\nDefaults to a GET request with no body and default headers (see `rest--default-headers').\nShows the result in a new buffer.\"\n  (rest-- method url body headers))\n(cl-defun rest-delete (url &amp;key (headers rest--default-headers))\n  \"Make an HTTP DELETE request to URL with optional HEADERS. Shows the result in a new buffer.\"\n  (rest-- \"delete\" url nil headers))\n(cl-defun rest-get (url &amp;key (headers rest--default-headers))\n  \"Make an HTTP GET request to URL with optional HEADERS. Shows the result in a new buffer.\"\n  (rest-- \"get\" url nil headers))\n(cl-defun rest-head (url &amp;key (headers rest--default-headers))\n  \"Make an HTTP HEAD request to URL with optional HEADERS. Shows the result in a new buffer.\"\n  (rest-- \"head\" url nil headers))\n(cl-defun rest-options (url &amp;key (headers rest--default-headers))\n  \"Make an HTTP OPTIONS request to URL with optional HEADERS. Shows the result in a new buffer.\"\n  (rest-- \"options\" url nil headers))\n(cl-defun rest-post (url &amp;key (body nil) (headers rest--default-headers))\n  \"Make an HTTP POST request to URL with BODY and optional HEADERS. Shows the result in a new buffer.\"\n  (rest-- \"post\" url body headers))\n(cl-defun rest-put (url &amp;key (body nil) (headers rest--default-headers))\n  \"Make an HTTP PUT request to URL with BODY and optional HEADERS. Shows the result in a new buffer.\"\n  (rest-- \"put\" url body headers))<\/pre><\/div>\n<h2><a id=\"user-content-f\" class=\"anchor\" aria-hidden=\"true\" href=\"#f\"><\/a>F#<\/h2>\n<p>Ensure there\u2019s a symlink to the F# bin dir!<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule fsharp\n  \"Init module for F# development. Installs fsharp-mode and sets up path to interpreter.\"\n  (use-package fsharp-mode\n    :mode \"\\\\.fsx?\\\\'\"\n    :config\n    (progn\n      (when (equal system-type 'windows-nt)\n        (let ((fsDir \"C:\\\\Program Files (x86)\\\\Microsoft SDKs\\\\F#\\\\3.1\\\\Framework\\\\v4.0\"))\n          ;; Doesn't necessarily work - just set your path!\n          (setenv \"PATH\"\n                  (concat (getenv \"PATH\")\n                          (format \";%s\" fsDir)))\n          (setq exec-path\n                (append exec-path '(fsDir)))))\n      (setq inferior-fsharp-program\n            (cond ((equal system-type 'windows-nt) \"Fsi.exe\")\n                  ((equal system-type 'cygwin) \"\/home\/chris.bowdon\/fs\/Fsi.exe\"))))))<\/pre><\/div>\n<h2><a id=\"user-content-python\" class=\"anchor\" aria-hidden=\"true\" href=\"#python\"><\/a>Python<\/h2>\n<p>I\u2019ve become a big fan of <code>lsp-mode<\/code>. The Python support requires <a href=\"https:\/\/github.com\/palantir\/python-language-server\">python-language-server<\/a> from PIP. For minimal stress, just install \u2018[all]\u2019.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule python3\n  \"Init module for Python 3 development.\"\n  (use-package pyvenv)\n  (use-package lsp-python\n    :config (add-hook 'python-mode-hook 'lsp-python-enable)))<\/pre><\/div>\n<h2><a id=\"user-content-javascript\" class=\"anchor\" aria-hidden=\"true\" href=\"#javascript\"><\/a>JavaScript<\/h2>\n<p>Though I remain suspicious of Facebook, React has brought a nice functional taste to JS. I\u2019ve configured <code>rjsx-mode<\/code> for all JS and JSX files since in various contexts (e.g. React Native) JSX files are still given a <code>.js<\/code> extension. It\u2019s built on <code>js2-mode<\/code>, which is very good for general JS programming anyway.<\/p>\n<p><code>prettier<\/code> (and any standardised formatting program, e.g. <code>yapf<\/code> for Python) should be mandatory. Not having to give a damn about formatting is a big weight off the mind.<\/p>\n<p>I\u2019m using <code>lsp-javascript<\/code> for completion\/checking support, which requires installing the <a href=\"https:\/\/github.com\/sourcegraph\/javascript-typescript-langserver\">javascript-typescript-langserver<\/a>, i.e. <code>yarn global add javascript-typescript-langserver<\/code>. Make sure <code>~\/.yarn\/bin<\/code> is on your path.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule javascript\n  \"Init module for JavaScript (including React\/JSX).\"\n  (use-package prettier-js)\n  (use-package rjsx-mode :mode \"\\\\.[jt]sx?$\")\n  (use-package lsp-javascript-typescript\n    :config (progn\n              (add-hook 'js-mode-hook #'lsp-javascript-typescript-enable)\n              (add-hook 'typescript-mode-hook #'lsp-javascript-typescript-enable)\n              (add-hook 'rjsx-mode #'lsp-javascript-typescript-enable))))<\/pre><\/div>\n<h2><a id=\"user-content-clojure\" class=\"anchor\" aria-hidden=\"true\" href=\"#clojure\"><\/a>Clojure<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule clojure\n  \"Init module for Clojure. A work in progress.\"\n  (use-package clojure-mode\n    :pin melpa-stable\n    :config (progn\n              (add-hook 'clojure-mode-hook 'paredit-mode)\n              (add-to-list 'auto-mode-alist '(\"\\\\.boot$\" . clojure-mode))\n              (add-to-list 'magic-mode-alist '(\".* boot\" . clojure-mode))))\n  (defun cider-repl-require-helpers ()\n    \"Insert require expressions into a Clojure REPL to save some typing.\"\n    (interactive)\n    (when (eq major-mode 'clojure-mode)\n      (cider-switch-to-repl-buffer))\n    (insert \"(require '[clojure.repl :refer :all] '[clojure.pprint :refer :all])\")\n    (cider-repl-return))\n  (use-package cider\n    :pin melpa-stable\n    :config (progn\n              (setq cider-auto-jump-to-error nil)\n              (dolist (m '(cider-stacktrace-mode\n                           cider-test-report-mode\n                           cider-classpath-mode\n                           cider-browse-ns-mode\n                           cider-browse-spec-mode\n                           cider-browse-spec-example-mode\n                           cider-browse-spec-view-mode))\n                (add-to-list 'evil-emacs-state-modes m))\n              (add-hook 'cider-repl-mode-hook 'paredit-mode)\n              (add-hook 'cider-repl-mode-hook 'company-mode)\n              (dolist (m (list cider-mode-map cider-repl-mode-map))\n                (define-key m (kbd \"C-c C-n\") 'cider-repl-set-ns))\n              (dolist (m '(clojure-mode cider-repl-mode))\n                (evil-leader\/set-key-for-mode m\n                  \"cc\" 'cider-eval-defun-at-point\n                  \"ce\" 'cider-eval-last-sexp\n                  \"ck\" 'cider-load-buffer\n                  \"cl\" 'cider-load-file\n                  \"cL\" 'cider-load-all-files\n                  \"cfv\" 'cider-find-var\n                  \"cfn\" 'cider-find-ns\n                  \"cfk\" 'cider-find-keyword\n                  \"cfp\" 'cider-find-property\n                  \"cfr\" 'cider-find-resource\n                  \"cda\" 'cider-apropos\n                  \"cdd\" 'cider-doc\n                  \"cdj\" 'cider-javadoc\n                  \"cdr\" 'cider-grimoire\n                  \"cD\" 'cider-debug-defun-at-point\n                  \"ctt\" 'cider-test-run-test\n                  \"ctn\" 'cider-test-run-ns-tests\n                  \"ctp\" 'cider-test-run-project-tests\n                  \"crn\" 'cider-repl-set-ns\n                  \"crh\" 'cider-repl-require-helpers\n                  \"cz\" 'cider-switch-to-repl-buffer\n                  \"cj\" 'cider-jack-in)))))<\/pre><\/div>\n<h2><a id=\"user-content-java\" class=\"anchor\" aria-hidden=\"true\" href=\"#java\"><\/a>Java<\/h2>\n<p><code>lsp-java<\/code> is the best Java mode I\u2019ve tried so far (beating <code>meghanada<\/code>, <code>eclim<\/code> and <code>ensime<\/code>). Make sure to download the <a href=\"http:\/\/download.eclipse.org\/jdtls\/snapshots\/jdt-language-server-latest.tar.gz\" rel=\"nofollow\">Eclipse JDT language server<\/a> and put it in <code>~\/.emacs.d\/eclipse.jdt.ls\/server\/<\/code>.<\/p>\n<p><code>maven-test-mode<\/code> is a convenient way to run individual tests rather than passing detailed arguments to <code>mvn<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule java\n  \"Init module for Java.\"\n  (use-package lsp-java\n    :config (progn\n              (add-hook 'java-mode-hook (lambda () (add-to-list 'lsp-java--workspace-folders (lsp-java--get-root))))\n              (add-hook 'java-mode-hook 'lsp-java-enable)))\n  (use-package maven-test-mode))<\/pre><\/div>\n<h2><a id=\"user-content-scala\" class=\"anchor\" aria-hidden=\"true\" href=\"#scala\"><\/a>Scala<\/h2>\n<p>Like there\u2019s any other choice but ENSIME?<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule scala\n  \"Fucking Scala init module\"\n  (use-package ensime)\n  (setq ensime-startup-notification nil\n        ensime-startup-snapshot-notification nil))<\/pre><\/div>\n<h2><a id=\"user-content-ansible\" class=\"anchor\" aria-hidden=\"true\" href=\"#ansible\"><\/a>Ansible<\/h2>\n<p>Ansible is my configuration management solution of choice. The Emacs support isn\u2019t bad.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule ansible\n  \"Init module for Ansible.\"\n  (use-package ansible)\n  (use-package ansible-doc)\n  (use-package company)\n  (use-package company-ansible))<\/pre><\/div>\n<h2><a id=\"user-content-ruby\" class=\"anchor\" aria-hidden=\"true\" href=\"#ruby\"><\/a>Ruby<\/h2>\n<p>I ended up coding some Ruby for use with Chef and Capistrano.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule ruby\n  \"Init module for Ruby. Uses Robe.\"\n  (use-package inf-ruby)\n  (add-hook 'after-init-hook 'inf-ruby-switch-setup)\n  (use-package robe)\n  (eval-after-load 'company\n    '(push 'company-robe company-backends))\n  (evil-leader\/set-key-for-mode 'ruby-mode\n    \"r\" 'robe-start\n    \"j\" 'robe-jump\n    \"c\" 'company-robe))<\/pre><\/div>\n<h2><a id=\"user-content-rust\" class=\"anchor\" aria-hidden=\"true\" href=\"#rust\"><\/a>Rust<\/h2>\n<p>Rust support is pretty basic at this stage. Autocompletion comes from a separate program, <code>racer<\/code> that reads the stdlib source code.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule rust\n  \"Init module for Rust. Uses racer for autocompletion.\"\n  (use-package rust-mode\n    :config (add-hook 'rust-mode-hook '(lambda ()\n                                         (racer-activate)\n                                         (racer-turn-on-eldoc)\n                                         (add-hook 'flycheck-mode-hook #'flycheck-rust-setup))))\n  (use-package company\n    :config (setq company-idle-delay 0.2\n                  company-minimum-prefix-length 1))\n  (use-package company-racer\n    :config (set (make-local-variable 'company-backends)\n                 '(company-racer)))\n  ;; racer reads the rust src code to suggest competions\n  ;; git clone https:\/\/github.com\/rust-lang\/rust.git ~\/.rust\n  ;; git clone https:\/\/github.com\/phildawes\/racer.git ~\/.racer\n  ;; cd ~\/.racer\n  ;; cargo build --release\n  (use-package racer\n    :config (setq racer-cmd (expand-file-name \"~\/.racer\/target\/release\/racer\")\n                  racer-rust-src-path (expand-file-name \"~\/.rust\/src\")))\n  (use-package flycheck  )\n  (use-package flycheck-rust  ))<\/pre><\/div>\n<h2><a id=\"user-content-php\" class=\"anchor\" aria-hidden=\"true\" href=\"#php\"><\/a>PHP<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule php\n  \"Init module for php. Uses web-mode for Cake templates.\"\n  (use-package php-mode :mode \"\\\\.php$\")\n  (use-package web-mode :mode \"\\\\.ctp$\"))<\/pre><\/div>\n<h2><a id=\"user-content-docker\" class=\"anchor\" aria-hidden=\"true\" href=\"#docker\"><\/a>Docker<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule docker\n  \"Init module for working with Docker and Compose.\"\n  (use-package docker)\n  (use-package docker-compose-mode)\n  (use-package dockerfile-mode :mode \"Dockerfile$\"))<\/pre><\/div>\n<h2><a id=\"user-content-sql\" class=\"anchor\" aria-hidden=\"true\" href=\"#sql\"><\/a>SQL<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule sql\n  \"Init module for SQL. Sets evil leader shortcuts for interpreters.\"\n  (evil-leader\/set-key\n    \"s q p\" 'sql-postgres\n    \"s q s\" 'sql-sqlite))<\/pre><\/div>\n<h2><a id=\"user-content-data-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#data-languages\"><\/a>Data languages<\/h2>\n<h3><a id=\"user-content-csv\" class=\"anchor\" aria-hidden=\"true\" href=\"#csv\"><\/a>CSV<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package csv-mode)<\/pre><\/div>\n<h3><a id=\"user-content-json\" class=\"anchor\" aria-hidden=\"true\" href=\"#json\"><\/a>JSON<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule json\n  \"Init module to support JSON syntax highlighting\/navigation\/formatting.\"\n  (use-package json-mode :mode \"\\\\.json\")\n  (use-package json-navigator)\n  (use-package json-reformat)\n  (evil-leader\/set-key-for-mode 'json-mode\n    \"jnp\" 'json-navigator-navigate-after-point\n    \"jnr\" 'json-navigator-navigate-region\n    \"jr\" 'json-reformat-region\n    \"jpr\" 'json-pretty-print\n    \"jpb\" 'json-pretty-print-buffer))<\/pre><\/div>\n<h3><a id=\"user-content-xml\" class=\"anchor\" aria-hidden=\"true\" href=\"#xml\"><\/a>XML<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule xml\n  \"Init module for XML. Adds helper functions and tag folding.\"\n  (defun split-xml-lines ()\n    (interactive)\n    ;; TODO use looking-at etc. because replace-regexp is interactive\n    (replace-regexp \"&gt; *&lt;\" \"&gt;\\n&lt;\"))\n  (require 'hideshow)\n  (require 'sgml-mode)\n  (require 'nxml-mode)\n  (add-to-list 'hs-special-modes-alist\n               '(nxml-mode\n                 \"&lt;!--\\\\|&lt;[^\/&gt;]*[^\/]&gt;\"\n                 \"--&gt;\\\\|&lt;\/[^\/&gt;]*[^\/]&gt;\"\n                 \"&lt;!--\"\n                 sgml-skip-tag-forward\n                 nil))\n  (add-hook 'nxml-mode-hook 'hs-minor-mode)\n  (evil-leader\/set-key-for-mode 'nxml-mode\n    \"h\" 'hs-toggle-hiding))<\/pre><\/div>\n<h3><a id=\"user-content-yaml\" class=\"anchor\" aria-hidden=\"true\" href=\"#yaml\"><\/a>YAML<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule yaml\n  \"Init module for YAML support.\"\n  (use-package yaml-mode :mode \"\\\\.ya?ml\"))<\/pre><\/div>\n<h3><a id=\"user-content-graphql\" class=\"anchor\" aria-hidden=\"true\" href=\"#graphql\"><\/a>GraphQL<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package graphql-mode :mode \"\\\\.sdl$|\\\\.graphql$\")<\/pre><\/div>\n<h3><a id=\"user-content-elasticsearch\" class=\"anchor\" aria-hidden=\"true\" href=\"#elasticsearch\"><\/a>ElasticSearch<\/h3>\n<pre>(use-package es-mode :mode \"\\\\.es$\")\n<\/pre>\n<h2><a id=\"user-content-config-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#config-languages\"><\/a>Config languages<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule configs\n  \"Init module for config languages (e.g. Apache, nginx configs).\"\n  (use-package apache-mode)\n  (use-package nginx-mode)\n  (use-package syslog-mode\n    :load-path \"~\/.emacs.d\/syslog-mode.el\"\n    :mode \"\\\\.log$\")\n  (evil-leader\/set-key\n    \"mca\" 'apache-mode\n    \"mcs\" 'syslog-mode\n    \"mcn\" 'nginx-mode))<\/pre><\/div>\n<h2><a id=\"user-content-text-markup-languages\" class=\"anchor\" aria-hidden=\"true\" href=\"#text-markup-languages\"><\/a>Text markup languages<\/h2>\n<p>The ubiquitous, but not actually-that-pleasant, Markdown. I\u2019ve looked at a couple of real-time preview modes and they\u2019re nice, but all have external dependencies requiring Go or Ruby, which I\u2019m not interested in installing. (I run a lean system. \ud83d\ude09)<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package markdown-mode\n  :mode \"\\\\.md$\")<\/pre><\/div>\n<h2><a id=\"user-content-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance\"><\/a>Performance<\/h2>\n<p>Line numbers are helpful in structured docs like org, but perform poorly (until the native line numbers in Emacs 26, anyway).<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-hook 'org-mode-hook\n          (lambda ()\n            (let ((lines (count-lines (point-min) (point-max))))\n              (when (&lt; lines 500)\n                (linum-mode)))))<\/pre><\/div>\n<h2><a id=\"user-content-evilification\" class=\"anchor\" aria-hidden=\"true\" href=\"#evilification\"><\/a>Evilification<\/h2>\n<p>It\u2019s nice to disable evil indentation, which doesn\u2019t play nice. I set a few leader bindings but really it\u2019s just easier to use God mode here.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-leader\/set-key-for-mode 'org-mode\n  \"oi\" 'org-clock-in\n  \"oo\" 'org-clock-out\n  \"or\" 'org-clock-report\n  \"ou\" 'org-update-all-dblocks\n  \"oa\" 'org-agenda\n  \"ot\" 'org-todo\n  \"oT\" 'org-set-tags\n  \"oe\" 'org-set-effort\n  \"ov\" 'org-columns\n  \"oq\" 'org-columns-quit\n  \"oc\" 'org-edit-special)\n(evil-leader\/set-key-for-mode 'org-mode\n  \"TAB\" 'org-cycle)\n(add-hook 'org-mode-hook #'(lambda () (electric-indent-local-mode 0)))\n(add-hook 'org-mode-hook #'(lambda () (setq evil-auto-indent nil)))<\/pre><\/div>\n<h2><a id=\"user-content-task-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#task-management\"><\/a>Task management<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq org-agenda-files '(\"~\/chris.org\")\n      org-enforce-todo-dependencies t\n      org-hide-emphasis-markers t\n      ;; Don't show days when summing times, just hours and minutes\n      org-time-clocksum-format (list :hours \"%d\"\n                                     :require-hours t\n                                     :minutes \":%02d\"\n                                     :require-minutes t))<\/pre><\/div>\n<h2><a id=\"user-content-babel\" class=\"anchor\" aria-hidden=\"true\" href=\"#babel\"><\/a>Babel<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq org-confirm-babel-evaluate #'(lambda (lang body)\n                                     (not (or (string= lang \"emacs-lisp\")\n                                              (string= lang \"dot\")\n                                              (string= lang \"python\")))))\n(org-babel-do-load-languages 'org-babel-load-languages\n                             '((emacs-lisp . t)\n                               (dot . t)\n                               (python . t)\n                               (ditaa . t)))<\/pre><\/div>\n<h2><a id=\"user-content-exporting\" class=\"anchor\" aria-hidden=\"true\" href=\"#exporting\"><\/a>Exporting<\/h2>\n<p>Because everyone else uses GitHub-flavored markdown\u2026<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package htmlize)\n(use-package ox-gfm)<\/pre><\/div>\n<h2><a id=\"user-content-diagrams-graphviz\" class=\"anchor\" aria-hidden=\"true\" href=\"#diagrams-graphviz\"><\/a>Diagrams (graphviz)<\/h2>\n<p>I prefer graphviz to ditaa. The binding for (re)displaying inline images helps iterate quickly on diagrams.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq default-tab-width tab-width)\n(use-package graphviz-dot-mode)\n(evil-leader\/set-key-for-mode 'org-mode\n  \"oI\" 'org-display-inline-images)<\/pre><\/div>\n<p>The dot language user guide is <a href=\"https:\/\/graphviz.gitlab.io\/_pages\/pdf\/dotguide.pdf\" rel=\"nofollow\">here<\/a>.<\/p>\n<h2><a id=\"user-content-daemons\" class=\"anchor\" aria-hidden=\"true\" href=\"#daemons\"><\/a>Daemons<\/h2>\n<p>Use <a href=\"https:\/\/github.com\/cbowdon\/daemons.el\">daemons<\/a> to manage system services and customize it for evil.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'evil-emacs-state-modes 'daemons-mode)\n(add-to-list 'evil-emacs-state-modes 'daemons-output-mode)\n(use-package daemons\n  :config\n  (progn\n    (add-to-list 'evil-emacs-state-modes 'daemons-mode)\n    (add-to-list 'evil-emacs-state-modes 'daemons-output-mode)\n    (evil-leader\/set-key\n      \"d d\" 'daemons\n      \"d RET\" 'daemons-status\n      \"d s\" 'daemons-start\n      \"d S\" 'daemons-stop\n      \"d r\" 'daemons-reload\n      \"d R\" 'daemons-restart)))<\/pre><\/div>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun set-frame-transparency (fg-percent bg-percent)\n  \"Set the transparency of this frame to FG-PERCENT when focused and BG-PERCENT when not focused.\"\n  (interactive \"nforeground percent:\\nnbackground percent:\")\n  (set-frame-parameter (selected-frame) 'alpha (list fg-percent bg-percent)))\n(set-frame-transparency 93 80)\n(evil-leader\/set-key\n  \"F\" 'toggle-frame-fullscreen)\n(define-key global-map (kbd \"C-S-F\") 'toggle-frame-fullscreen)\n(setq frame-title-format\n      (format \"Emacs %s (%s) - %s@%s\"\n              emacs-version\n              (cond ((eq system-type 'cygwin) \"Cygwin\")\n                    ((eq system-type 'windows-nt) \"Windows\")\n                    (t \"*nix\"))\n              (user-login-name)\n              (system-name)))<\/pre><\/div>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package golden-ratio)\n(evil-leader\/set-key\n  \"w\" 'evil-window-map\n  \"wg\" 'golden-ratio\n  \"wG\" 'golden-ratio-mode)\n(define-key evil-window-map (kbd \"C-l\") 'evil-window-right)\n(define-key evil-window-map (kbd \"C-h\") 'evil-window-left)\n(define-key evil-window-map (kbd \"C-k\") 'evil-window-up)\n(define-key evil-window-map (kbd \"C-j\") 'evil-window-down)<\/pre><\/div>\n<h2><a id=\"user-content-performance-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance-1\"><\/a>Performance<\/h2>\n<p>Attempt to improve long line performance.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq-default bidi-display-reordering nil)<\/pre><\/div>\n<h2><a id=\"user-content-themes\" class=\"anchor\" aria-hidden=\"true\" href=\"#themes\"><\/a>Themes<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; strong color\/dark themes\n(use-package django-theme :defer t)\n(use-package ubuntu-theme :defer t)\n(use-package monokai-theme :defer t)\n(use-package darkokai-theme :defer t)\n(use-package badger-theme :defer t)\n;; gentler themes\n(use-package color-theme-sanityinc-tomorrow :defer t)\n(use-package color-theme-sanityinc-solarized :defer t)\n(use-package creamsody-theme :defer t)\n(use-package arjen-grey-theme :defer t)\n;; moody themes\n(use-package gotham-theme :defer t)\n(use-package blackboard-theme :defer t)\n(use-package bubbleberry-theme :defer t)\n;; can't-remember-but-possibly-nice themes\n(use-package material-theme :defer t)\n(use-package gruvbox-theme :defer t)\n<\/pre><\/div>\n<h2><a id=\"user-content-widgets\" class=\"anchor\" aria-hidden=\"true\" href=\"#widgets\"><\/a>Widgets<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(tool-bar-mode -1)\n(menu-bar-mode -1)\n(unless (frame-parameter nil 'tty)\n    (scroll-bar-mode -1))\n(setq inhibit-splash-screen t\n      ring-bell-function 'ignore)<\/pre><\/div>\n<h2><a id=\"user-content-modeline\" class=\"anchor\" aria-hidden=\"true\" href=\"#modeline\"><\/a>Modeline<\/h2>\n<p>Telephone line is a sweet rewrite of Powerline.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package telephone-line\n  :config (progn\n            (require 'telephone-line-config)\n            (telephone-line-evil-config)\n            (setq telephone-line-height 24)))<\/pre><\/div>\n<h2><a id=\"user-content-dashboard\" class=\"anchor\" aria-hidden=\"true\" href=\"#dashboard\"><\/a>Dashboard<\/h2>\n<p>Another good idea lovingly ripped off from Spacemacs.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package dashboard\n  :demand\n  :config\n  (progn\n    (dashboard-setup-startup-hook)\n    (setq dashboard-banner-logo-title \"[ E M A C S ]\"\n          dashboard-startup-banner 'logo\n          dashboard-items '((recents  . 8)\n                            (bookmarks . 5)))))<\/pre><\/div>\n<h2><a id=\"user-content-icons\" class=\"anchor\" aria-hidden=\"true\" href=\"#icons\"><\/a>Icons<\/h2>\n<p>For a bit of snazz, <code>all-the-icons<\/code> is nice. Run <code>all-the-icons-install-fonts<\/code> after getting the package.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package all-the-icons)\n(use-package all-the-icons-dired\n  :config (add-hook 'dired-mode-hook 'all-the-icons-dired-mode))<\/pre><\/div>\n<p>Backups and lock files not required.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq make-backup-files nil\n      create-lockfiles nil)<\/pre><\/div>\n<p>Being in the correct mode (line or char) is the cure for all evils in <code>term-mode<\/code>, but for some reason I couldn\u2019t get a key binding to work on <code>term-mode-hook<\/code> so just shove it here.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-leader\/set-key\n  \"se\" 'eshell\n  \"st\" 'term\n  \"sc\" 'shell-command\n  \"sr\" 'shell-command-on-region\n  \"sl\" 'term-line-mode\n  \"sc\" 'term-char-mode)<\/pre><\/div>\n<h2><a id=\"user-content-eshell\" class=\"anchor\" aria-hidden=\"true\" href=\"#eshell\"><\/a>Eshell<\/h2>\n<p>Fix an irritating warning about the pager program.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setenv \"PAGER\" \"\/bin\/cat\") ;; fixes git terminal warning\n(add-hook 'eshell-mode-hook #'(lambda () (setenv \"PAGER\" \"\/bin\/cat\")))<\/pre><\/div>\n<p>I often manually list after changing directory, so let\u2019s automate it.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq eshell-list-files-after-cd t)\n(setq eshell-ls-initial-args \"-lh\")<\/pre><\/div>\n<p>Calling <code>vim<\/code> is deep in my muscle memory, so alias it to <code>find-file<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun eshell\/vim (file)\n  (find-file file))\n(defun eshell\/vimo (file)\n  (find-file-other-window file))\n(defun eshell\/emacs (file)\n  (find-file-other-window file))<\/pre><\/div>\n<h3><a id=\"user-content-pop-up-shell\" class=\"anchor\" aria-hidden=\"true\" href=\"#pop-up-shell\"><\/a>Pop-up shell<\/h3>\n<p>Thank you <a href=\"https:\/\/github.com\/cbowdon\/Config\/blob\/master\/emacs\/howardism.org\">howardism.org<\/a>. I no longer use this, but keep it around for reference.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun eshell-here ()\n  \"Opens up a new shell in the directory associated with the\ncurrent buffer's file. The eshell is renamed to match that\ndirectory to make multiple eshell windows easier.\"\n  (interactive)\n  (let* ((parent (if (buffer-file-name)\n                     (file-name-directory (buffer-file-name))\n                   default-directory))\n         (height (\/ (window-total-height) 3))\n         (name   (car (last (split-string parent \"\/\" t)))))\n    (split-window-vertically (- height))\n    (other-window 1)\n    (eshell \"new\")\n    (rename-buffer (concat \"*eshell: \" name \"*\"))\n    (insert (concat \"ls\"))\n    (eshell-send-input)))\n(defun eshell\/quit ()\n  (insert \"exit\")\n  (eshell-send-input)\n  (delete-window))\n(defun toggle-eshell ()\n  (interactive)\n  (if (string-match \"^\\\\*eshell: \" (buffer-name))\n      (eshell\/quit)\n    (eshell-here)))<\/pre><\/div>\n<h3><a id=\"user-content-prompt\" class=\"anchor\" aria-hidden=\"true\" href=\"#prompt\"><\/a>Prompt<\/h3>\n<p>You can\u2019t be a professional without a try-hard prompt.<\/p>\n<h4><a id=\"user-content-xpm-shapes\" class=\"anchor\" aria-hidden=\"true\" href=\"#xpm-shapes\"><\/a>XPM shapes<\/h4>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>;; pinched from powerline.el\n(defun curve-right-xpm (color1 color2)\n  \"Return an XPM right curve string representing.\"\n  (create-image\n   (format \"\/* XPM *\/\nstatic char * curve_right[] = {\n\\\"12 18 2 1\\\",\n\\\". c %s\\\",\n\\\"  c %s\\\",\n\\\"           .\\\",\n\\\"         ...\\\",\n\\\"         ...\\\",\n\\\"       .....\\\",\n\\\"       .....\\\",\n\\\"       .....\\\",\n\\\"      ......\\\",\n\\\"      ......\\\",\n\\\"      ......\\\",\n\\\"      ......\\\",\n\\\"      ......\\\",\n\\\"      ......\\\",\n\\\"       .....\\\",\n\\\"       .....\\\",\n\\\"       .....\\\",\n\\\"         ...\\\",\n\\\"         ...\\\",\n\\\"           .\\\"};\"\n           (if color2 color2 \"None\")\n           (if color1 color1 \"None\"))\n   'xpm t :ascent 'center))\n(defun curve-left-xpm (color1 color2)\n  \"Return an XPM left curve string representing.\"\n  (create-image\n   (format \"\/* XPM *\/\nstatic char * curve_left[] = {\n\\\"12 18 2 1\\\",\n\\\". c %s\\\",\n\\\"  c %s\\\",\n\\\".           \\\",\n\\\"...         \\\",\n\\\"...         \\\",\n\\\".....       \\\",\n\\\".....       \\\",\n\\\".....       \\\",\n\\\"......      \\\",\n\\\"......      \\\",\n\\\"......      \\\",\n\\\"......      \\\",\n\\\"......      \\\",\n\\\"......      \\\",\n\\\".....       \\\",\n\\\".....       \\\",\n\\\".....       \\\",\n\\\"...         \\\",\n\\\"...         \\\",\n\\\".           \\\"};\"\n           (if color1 color1 \"None\")\n           (if color2 color2 \"None\"))\n   'xpm t :ascent 'center))\n(defun arrow-left-xpm (color1 color2)\n  \"Return an XPM left arrow string representing.\"\n  (create-image\n   (format \"\/* XPM *\/\nstatic char * curve_right[] = {\n\\\"12 18 2 1\\\",\n\\\". c %s\\\",\n\\\"  c %s\\\",\n\\\"         ...\\\",\n\\\"        ....\\\",\n\\\"       .....\\\",\n\\\"      ......\\\",\n\\\"     .......\\\",\n\\\"    ........\\\",\n\\\"   .........\\\",\n\\\"  ..........\\\",\n\\\" ...........\\\",\n\\\" ...........\\\",\n\\\"  ..........\\\",\n\\\"   .........\\\",\n\\\"    ........\\\",\n\\\"     .......\\\",\n\\\"      ......\\\",\n\\\"       .....\\\",\n\\\"        ....\\\",\n\\\"         ...\\\"};\"\n           (if color2 color2 \"None\")\n           (if color1 color1 \"None\"))\n   'xpm t :ascent 'center))\n(defun arrow-right-xpm (color1 color2)\n  \"Return an XPM right arrow string representing.\"\n  (create-image\n   (format \"\/* XPM *\/\nstatic char * curve_left[] = {\n\\\"12 18 2 1\\\",\n\\\". c %s\\\",\n\\\"  c %s\\\",\n\\\"...         \\\",\n\\\"....        \\\",\n\\\".....       \\\",\n\\\"......      \\\",\n\\\".......     \\\",\n\\\"........    \\\",\n\\\".........   \\\",\n\\\"..........  \\\",\n\\\"........... \\\",\n\\\"........... \\\",\n\\\"..........  \\\",\n\\\".........   \\\",\n\\\"........    \\\",\n\\\".......     \\\",\n\\\"......      \\\",\n\\\".....       \\\",\n\\\"....        \\\",\n\\\"...         \\\"};\"\n           (if color1 color1 \"None\")\n           (if color2 color2 \"None\"))\n   'xpm t :ascent 'center))<\/pre><\/div>\n<h4><a id=\"user-content-prompt-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#prompt-functions\"><\/a>Prompt functions<\/h4>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defvar eshell-prompt-suffix\n  (if (or (eq system-type 'gnu\/linux)\n          (eq system-type 'darwin)) \"\ud83d\udd25 \" \" \")\n  \"String at end of prompt.\")\n(unless (frame-parameter nil 'tty)\n  ;; TODO fancy prompt in terminal mode also\n  (defvar eshell-fg-color\n    (let ((fg-color (frame-parameter nil 'foreground-color)))\n      (if (equal \"unspecified-fg\" fg-color) \"#ffffff\" fg-color)))\n  (defvar eshell-bg-color\n    (let ((bg-color (frame-parameter nil 'background-color)))\n      (if (equal \"unspecified-bg\" bg-color) \"#ffffff\" bg-color)))\n  (defvar eshell-prompt-bg-color\n    (color-complement-hex (if (equal \"unspecified-fg\" eshell-fg-color) \"#ffffff\" eshell-fg-color))\n    \"Color of the prompt block background.\")\n  (defvar eshell-prompt-fg-color\n    (color-complement-hex (if (equal \"unspecified-bg\" eshell-bg-color) \"#000000\" eshell-bg-color))\n    \"Color of the prompt block foreground.\")\n  (defvar eshell-prompt--separators (make-hash-table))\n  (defun eshell-prompt--lhs (fg-color bg-color)\n    (let ((hash-key (concat \"lhs_\" fg-color \"_\" bg-color)))\n      (or (gethash hash-key eshell-prompt--separators))\n      (puthash hash-key (arrow-right-xpm bg-color fg-color) eshell-prompt--separators)))\n  (defun eshell-prompt--rhs (fg-color bg-color)\n    (let ((hash-key (concat \"rhs_\" fg-color \"_\" bg-color)))\n      (or (gethash hash-key eshell-prompt--separators))\n      (puthash hash-key (arrow-right-xpm fg-color bg-color) eshell-prompt--separators)))\n  (defun eshell-blocky-prompt ()\n    \"Create a blocky eshell prompt.\"\n    (let ((bg eshell-bg-color)\n          (fg eshell-fg-color))\n      (concat\n       (propertize \" \" 'display (eshell-prompt--lhs eshell-prompt-bg-color bg))\n       (propertize (eshell\/pwd) 'face\n                   (list :foreground eshell-prompt-fg-color\n                         :background eshell-prompt-bg-color))\n       (propertize \" \" 'display (eshell-prompt--rhs eshell-prompt-bg-color bg))\n       eshell-prompt-suffix)))\n  (defconst eshell-blocky-prompt-regexp\n    (string-join (list \"^[^#\\n]* \" eshell-prompt-suffix)))\n  (setq eshell-prompt-function 'eshell-blocky-prompt\n        eshell-prompt-regexp eshell-blocky-prompt-regexp))<\/pre><\/div>\n<p>These days Treemacs is the coolest directory browser.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package treemacs\n  :config (add-to-list 'evil-emacs-state-modes  'treemacs-mode))\n(evil-leader\/set-key \"t\" 'treemacs)<\/pre><\/div>\n<p>Customize dired a bit and give it a key.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(evil-leader\/set-key \"D\" 'dired)\n(setq dired-listing-switches \"-lh --group-directories-first\")<\/pre><\/div>\n<p>\u201cBetter at git than git\u201d true, but apparently not better at compiling and running on multiple platforms\u2026<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defmodule magit\n  \"Init module for Magit.\"\n  (use-package magit\n    :config\n    (progn\n      (global-set-key (kbd \"C-x g\") 'magit-status)\n      (add-to-list 'evil-emacs-state-modes 'magit-mode)\n      (add-to-list 'evil-emacs-state-modes 'magit-blame-mode)\n      (evil-leader\/set-key \"g\" 'magit-status))))<\/pre><\/div>\n<h2><a id=\"user-content-gnulinux\" class=\"anchor\" aria-hidden=\"true\" href=\"#gnulinux\"><\/a>GNU\/Linux<\/h2>\n<h3><a id=\"user-content-terminal-colors\" class=\"anchor\" aria-hidden=\"true\" href=\"#terminal-colors\"><\/a>Terminal colors<\/h3>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(when (eq system-type 'gnu\/linux)\n  ;; (use-package eterm-256color :config (add-hook 'term-mode-hook #'eterm-256color-mode))\n  )<\/pre><\/div>\n<h2><a id=\"user-content-macos\" class=\"anchor\" aria-hidden=\"true\" href=\"#macos\"><\/a>MacOS<\/h2>\n<h3><a id=\"user-content-terminal-colors-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#terminal-colors-1\"><\/a>Terminal colors<\/h3>\n<p><a href=\"http:\/\/stackoverflow.com\/questions\/8918910\/weird-character-zsh-in-emacs-terminal\/8920373#8920373\" rel=\"nofollow\">It turns out <code>term<\/code> and <code>ansi-term<\/code> on MacOS needs a bit of setup.<\/a><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun fix-terminal-colors ()\n  \"Installs a copy of eterm-color terminfo.\"\n  (interactive)\n  (let ((path-to-emacs-app \"\/Applications\/Emacs.app\"))\n    (shell-command\n     (format \"tic -o ~\/.terminfo %s\/Contents\/Resources\/etc\/e\/eterm-color.ti\"\n             path-to-emacs-app))))<\/pre><\/div>\n<h2><a id=\"user-content-windows\" class=\"anchor\" aria-hidden=\"true\" href=\"#windows\"><\/a>Windows<\/h2>\n<h3><a id=\"user-content-performance-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#performance-2\"><\/a>Performance<\/h3>\n<p><a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/55ork0\/is_emacs_251_noticeably_slower_than_245_on_windows\/\" rel=\"nofollow\">Some tweaks are required on Windows.<\/a><\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(when (eq system-type 'windows-nt)\n  (if (&gt;= emacs-major-version 25)\n      (remove-hook 'find-file-hooks 'vc-refresh-state)\n    (remove-hook 'find-file-hooks 'vc-find-file-hook))\n  (progn\n    (setq gc-cons-threshold (* 511 1024 1024)\n          gc-cons-percentage 0.5\n          garbage-collection-messages t)\n    (run-with-idle-timer 5 t #'garbage-collect)))<\/pre><\/div>\n<h3><a id=\"user-content-tramp\" class=\"anchor\" aria-hidden=\"true\" href=\"#tramp\"><\/a>TRAMP<\/h3>\n<p>PuTTY (and so <code>plink.exe<\/code>) is basically your only choice on Windows.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(when (eq system-type 'windows-nt)\n  (require 'tramp)\n  (setq tramp-default-method \"plink\"))<\/pre><\/div>\n<h2><a id=\"user-content-helpers\" class=\"anchor\" aria-hidden=\"true\" href=\"#helpers\"><\/a>Helpers<\/h2>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun is-nix ()\n  (or (equal system-type \"gnu\")\n      (equal system-type \"gnu\/linux\")\n      (equal system-type \"gnu\/kfreebsd\")\n      (equal system-type \"darwin\")))<\/pre><\/div>\n<p>Provide functions to set all the necessary proxy variables for <code>$day_job<\/code>, assuming that <code>proxy-host<\/code> and <code>proxy-port<\/code> are set in <code>custom-file<\/code>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(defun proxy-make-url (host port &amp;optional username password)\n  (concat\n   (when (or username password)\n     (format \"%s:%s@\"\n             (if (not username) \"\" username)\n             (if (not password) \"\" password)))\n   (format \"%s:%s\" host port)))\n(defun proxy-set (http-proxy)\n  \"Set proxy variables that Emacs uses from the provided HTTP-PROXY string.\"\n  (setenv \"HTTP_PROXY\" (format \"https:\/\/%s\" http-proxy))\n  (setq url-proxy-services (list (cons \"http\" http-proxy)\n                                 (cons \"https\" http-proxy))))\n(defun proxy-set-with-user ()\n  \"Set proxy using current user login name and asking for password.\"\n  (interactive)\n  (proxy-set (proxy-make-url proxy-host\n                             proxy-port\n                             (user-login-name)\n                             (read-passwd \"Password: \"))))<\/pre><\/div>\n<p>For working with <a href=\"https:\/\/password-store.org\" rel=\"nofollow\">pass<\/a>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package password-store\n  :config (evil-leader\/set-key\n            \"Pc\" 'password-store-copy\n            \"Pg\" 'password-store-generate\n            \"Pr\" 'password-store-remove\n            \"Pe\" 'password-store-edit\n            \"PR\" 'password-store-rename\n            \"Pi\" 'password-store-insert\n            \"Pu\" 'password-store-url))<\/pre><\/div>","created_at":"2018-11-07T16:40:32+0000","updated_at":"2018-11-07T16:41:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":30,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/1069832?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":239,"title":"From helm, to ivy \u00b7 Samuel Barreto","url":"https:\/\/sam217pa.github.io\/2016\/09\/13\/from-helm-to-ivy\/","content":"<p>Recently, I have seen a lot of Emacs users showing interest in Ivy.\nMost of them already know Helm or Ido. When someone asked\n<a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/51lqn9\/helm_or_ivy\/\">Helm or Ivy?<\/a>\non reddit, I thought that I could give my opinion on Ivy, since I am\nan ex-Helm zealot.<\/p><p><strong>Helm and Ivy are completion frameworks<\/strong>. It means they are tools\nthat the Emacs ecosystem can use to narrow the field of possibilities\nafter a given user input. The most common example that comes to mind\nis searching for files. Helm and Ivy helps the user to quickly search\nfor files.<\/p>\n<p>They are frameworks. It means they can be used by much more complex\ncommands that needs narrowing or completion. As an example, Helm has a\ncommand that emulate the google search bar, and get google suggests as\nyou type (<code>helm-google-suggest<\/code>).<\/p>\n<p>Ivy and Helm have very similar goals, but the approach they take is\nvery different.<\/p>\n<p><strong>Here I want to compare the tools, from a user point of view.<\/strong> What\nI mean by user perspective is that I want to compare Ivy and Helm\nwithout knowing how they works internally. I do not know enough\n<code>elisp<\/code> to be able to compare their internals. But I used them both,\nand can give you my feedback on how I, as a <em>user<\/em>, felt the\ndifference between the tools. And finally, why I switched from Helm to\nIvy.<\/p>\n<p>I want to start with Helm. When I used Spacemacs, I learned the Helm\nway. I learned how to <em>use<\/em> it, how to <em>customize<\/em> it, and how to\n<em>tailor<\/em> it to my needs. I guess I would qualify as a medium-level\nHelm user. I have read\n<a href=\"http:\/\/tuhdo.github.io\/helm-intro.html\">this post<\/a> and\n<a href=\"http:\/\/tuhdo.github.io\/helm-projectile.html\">this one<\/a> and\n<a href=\"https:\/\/github.com\/emacs-helm\/helm\/wiki\">this wiki too<\/a>, and used it\nevery day for one year.<\/p>\n<p><strong>Helm is a very mature tool.<\/strong> According to its git commit history,\nwork on Helm really started around 2009. At the time of writing, the\nofficial Helm git repository has 26k lines of emacs-lisp code.<\/p>\n<pre class=\"language-bash\">git clone https:\/\/github.com\/emacs-helm\/helm.git\ncd helm\ncat *.el | wc -l\n# =&gt; 26431\n<\/pre>\n<p>This does not take into account the 142 packages that a query for\n\u201chelm\u201d returns on MELPA. You can do anything with Helm<sup class=\"footnote-ref\" id=\"fnref:1\"><a href=\"#fn:1\">1<\/a><\/sup>.<\/p>\n<p>Its main strength is that you can integrate and wrap many Emacs\nbehavior with it. You can build an interface totally centered around\nHelm, just like Spacemacs did. It allows for a really really\nconsistent interface. Everything goes through Helm.<\/p>\n<p>You can search files, search buffer, search colors, search project,\nsearch recently edited files, search system process, search music,\nsearch the internet, search completion, search snippets, search\nregexp, search commands, documentation, a-propos, infos, \u2026 You can\nmanage your project with Helm-projectile, a nice wrapping around\nProjectile. You can generate gitignore with\n<a href=\"https:\/\/www.gitignore.io\/\">gitignore.io<\/a>. You can manage your\nbibliography with Helm-bibtex. You can browse Firefox bookmarks.<\/p>\n<p>You can do anything with Helm.<\/p>\n<p>One feature I want to highlight, based on feedbacks from tuhdo\n<a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/52lnad\/from_helm_to_ivy_a_user_perspective\/d7lypeu\">here on Reddit<\/a>,\nis that Helm does not use the minibuffer<sup class=\"footnote-ref\" id=\"fnref:2\"><a href=\"#fn:2\">2<\/a><\/sup>, while Ivy does. So it\ncan be configured to open always in the currently opened window. It is\na very neat feature for users with large or ultrawide monitors. Eyes\ndo not have to travel back and forth to the minibuffer: completion\nresults are always printed in the same window.<\/p>\n<figure><a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/52lnad\/from_helm_to_ivy_a_user_perspective\/d7lypeu\">\n        <img src=\"http:\/\/i.imgur.com\/g1Oz9JY.png\" alt=\"Helm configured to open in the currently active window\" \/><\/a>\n    \n    <figcaption>\n        Helm configured to open in the currently active window\n        \n            \n        \n         \n    <\/figcaption><\/figure><p>The end result is a very convenient tool to use, as the great majority\nof Spacemacs users can tell you. The major drawback it has is that\nit is a big behemoth size package. I think it is an <code>elisp<\/code> exploit\nfrom the devs to have succeeded in making Helm a very fast tool\ndespite its size.<\/p>\n<p>And yet it sometimes feels like an overblown solution to simple\nproblems. It can be clumsy to configure. It can sometimes have strange\nbehaviors, causing lags or make Emacs use 100% of a core, even for\nsimple queries. Maybe an expert Helm user reading this can put those\nlags on me not being an <code>elisp<\/code> expert, but still. Despite having used\nit for more than a year, I did not find the way to make it more\nstable. I think it tells something about the \u201coverblown problem\u201d I was\ntalking about.<\/p>\n<p>You can do anything with Helm. But you don\u2019t need to. Just because you\ncan does not mean you should. After a year of using it, I can tell you\nthat I have only used a third or less of Helm\u2019s abilities. Some\nfunctions I thought were great, I re-discovered by reading\n<a href=\"http:\/\/tuhdo.github.io\/helm-intro.html\">this post<\/a> yesterday. Most of\nthe time, I used simple commands to switch buffers, or list files.<\/p>\n<blockquote>\n<p>Helm is only a package for completion, just like ido or ivy. It may\nbe easy to use, once someone has gone through the trouble of setting\nit up, but it\u2019s hard to get it to do exactly what you want. Some\npeople are fine with black boxes, as long as they give them nice\nthings, I\u2019m not.<\/p>\n<p>\u2014 abo-abo, developper of Ivy, on the \u201cWhy not Helm ?\u201d issue,\n<a href=\"https:\/\/github.com\/abo-abo\/swiper\/issues\/3\">here<\/a>.<\/p>\n<\/blockquote>\n<p><strong>Ivy strives for minimalism, simplicity, customizability and\ndiscoverability.<\/strong> Those four adjectives tells a lot about the\nparadigm difference between the two. Read\n<a href=\"http:\/\/oremacs.com\/swiper\/\">the introduction<\/a> of the Ivy\ndocumentation to have a good idea of what Ivy strives for.<\/p>\n<p>At the time of writing, Ivy only has 3.4k lines of code. The ecosystem\nbuilt around it \u2014 <em>ie<\/em> Swiper and Counsel \u2014 is 7.5k lines of code\nlong.<\/p>\n<pre class=\"language-bash\">git clone https:\/\/github.com\/abo-abo\/swiper.git\ncd swiper\n## Only ivy ?\ncat ivy.el | wc -l\n# =&gt; 3442\n## count lines of code into the whole swiper ecosystem\ncat *.el | wc -l\n# =&gt; 7526\n<\/pre>\n<p>Ivy is really simple to use. Here is my whole setup.<\/p>\n<pre class=\"language-lisp\">(use-package ivy :ensure t\n  :diminish (ivy-mode . \"\")\n  :bind\n  (:map ivy-mode-map\n   (\"C-'\" . ivy-avy))\n  :config\n  (ivy-mode 1)\n  ;; add \u2018recentf-mode\u2019 and bookmarks to \u2018ivy-switch-buffer\u2019.\n  (setq ivy-use-virtual-buffers t)\n  ;; number of result lines to display\n  (setq ivy-height 10)\n  ;; does not count candidates\n  (setq ivy-count-format \"\")\n  ;; no regexp by default\n  (setq ivy-initial-inputs-alist nil)\n  ;; configure regexp engine.\n  (setq ivy-re-builders-alist\n\t;; allow input not in order\n        '((t   . ivy--regex-ignore-order))))\n<\/pre>\n<p>Ivy is unobtrusive. It does not want you to integrate everything with\nit. It simply is the power house of your completion. You cannot do\neverything that the Helm ecosystem does with Ivy. So why did I\nswitched to Ivy ?<\/p>\n<p>Despite being minimalistic, I was able to replace most of my common\nHelm use cases with Ivy. Because Ivy is so minimalistic, abo-abo built\na package on top of it called Counsel. Counsel offers you many many\ntricks that you might like from Helm.<\/p>\n<p>You can switch buffers, search files, get project-wide search and\nreplace, get Projectile integration, search recently edited files,\nsearch Emacs command, search documentation, search keybindings, browse\nthe kill-ring \u2026<\/p>\n<p><strong>Let me describe how I replaced Helm with Ivy.<\/strong> Here is an overview\nof the most common Helm commands I needed to replace with Ivy. It was\nbasically the functions I used all the time. I <code>ivy-switch-buffer<\/code> like\nthree times per minute. I used <code>helm-swoop<\/code> fifty times a day, and\n<code>swiper<\/code> is just as good if not better. It is faster. For huge files,\nCounsel has <code>counsel-grep-or-swiper<\/code>. I have tested it on really\nreally huge made up files (1M lines or so). It did not tremble.<\/p>\n<table><thead><tr><th align=\"left\">Helm<\/th>\n<th align=\"left\">Ivy<\/th>\n<th align=\"left\">What ?<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\">helm-mini<\/td>\n<td align=\"left\">ivy-switch-buffer<\/td>\n<td align=\"left\">search for currently opened buffers<\/td>\n<\/tr><tr><td align=\"left\">helm-recentf<\/td>\n<td align=\"left\">counsel-recentf<\/td>\n<td align=\"left\">search for recently edited files<\/td>\n<\/tr><tr><td align=\"left\">helm-find-files<\/td>\n<td align=\"left\">counsel-find-files<\/td>\n<td align=\"left\">search files starting from <code>.\/<\/code><\/td>\n<\/tr><tr><td align=\"left\">helm-ag<\/td>\n<td align=\"left\">counsel-ag<\/td>\n<td align=\"left\">search regexp occurence in current project<\/td>\n<\/tr><tr><td align=\"left\">helm-grep-do-git-grep<\/td>\n<td align=\"left\">counsel-git-grep<\/td>\n<td align=\"left\">search regexp in current project<\/td>\n<\/tr><tr><td align=\"left\">helm-swoop<\/td>\n<td align=\"left\">swiper<\/td>\n<td align=\"left\">search string interactively in current buffer<\/td>\n<\/tr><tr><td align=\"left\">helm-show-kill-ring<\/td>\n<td align=\"left\">counsel-yank-pop<\/td>\n<td align=\"left\">search copy-paste history<\/td>\n<\/tr><tr><td align=\"left\">helm-projectile<\/td>\n<td align=\"left\">counsel-projectile<\/td>\n<td align=\"left\">search project and file in it<\/td>\n<\/tr><tr><td align=\"left\">helm-ls-git-ls<\/td>\n<td align=\"left\">counsel-git<\/td>\n<td align=\"left\">search file in current git project<\/td>\n<\/tr><tr><td align=\"left\">helm-themes<\/td>\n<td align=\"left\">counsel-load-theme<\/td>\n<td align=\"left\">switch themes<\/td>\n<\/tr><tr><td align=\"left\">helm-descbinds<\/td>\n<td align=\"left\">counsel-descbinds<\/td>\n<td align=\"left\">describe keybindings and associated functions<\/td>\n<\/tr><tr><td align=\"left\">helm-M-x<\/td>\n<td align=\"left\">counsel-M-x<\/td>\n<td align=\"left\">enhanced M-x command<\/td>\n<\/tr><\/tbody><\/table><p>I think you can see that Ivy based commands have nothing to be ashamed\nof. They can replace every common use cases of Helm. I am not saying\nthat you can do everything Helm does with Ivy. But it is already very\nconvenient. And as I said, I don\u2019t need to do everything Helm does.<\/p>\n<p>To speak of completion paradigm, the difference is not that obvious\nbetween Helm and Ivy. What I can tell you, as a <em>user<\/em>, is that Ivy\n<em>feels<\/em> less cluttered, snappier, more intuitive, self-explanatory and\nreally understandable. Every completion is predictable.<\/p>\n<p><strong>In the end, it really is a matter of personal taste.<\/strong> To me the\n\u201cIvy or Helm\u201d debate is very similar to debates likes Emacs or\nSpacemacs, Emacs or an IDE, C or Java, minimalistic or full-blown.\nThelonious or Duke. Van Der Rohe or Gaudi.<\/p>\n<p>With Helm you get a huge package, a full list of features you will\nnever use, a whole bunch of functions you will use occasionally, and a\nshort list of functions you will use fifty times an hour. With Ivy you\nget a small package with only essentials functions that does not get\nin your way, and that you can extend easily through Counsel or very\nsimple functions:<\/p>\n<pre class=\"language-lisp\">(ivy-read \"Pick:\" (mapcar #'number-to-string (number-sequence 1 10)))\n<\/pre>\n<p>With Helm:<\/p>\n<pre class=\"language-lisp\">(helm\n :sources\n  (helm-build-sync-source \"one-to-ten\"\n    :candidates\n    (mapcar #'number-to-string (number-sequence 1 10))\n    :fuzzy-match t)\n  :buffer\n  \"*helm one-to-ten*\")\n<\/pre>\n<p>Or the simpler form:<\/p>\n<pre class=\"language-lisp\">(helm-comp-read \"Pick:\" (mapcar #'number-to-string (number-sequence 1 10)))\n<\/pre>\n<p>Helm makes a lot of choices for the users. Ivy lets them tailor it to\ntheir needs. Helm uses a lot of memory to be very fast. Ivy stays\nsimple to be very fast. Helm is mature. Ivy is young. Helm offers\nconsistency across Emacs. Ivy offers simplicity and predictability.\nHelm ask you a rather involved setup. Ivy works out of the box.<\/p>\n<p>I am a little biased towards Ivy, since I use it. It suits my tastes\nbetter. But as a user, Helm and Ivy does not feel that different. They\nare both great packages, that chose a very different approach for the\nsame goal.<\/p>\n<div class=\"reddit-embed\" data-embed-media=\"www.redditmedia.com\" data-embed-parent=\"false\" data-embed-live=\"false\" data-embed-uuid=\"36b96121-6628-470a-82a9-dbacdc593a42\" data-embed-created=\"2016-09-13T11:22:25.426Z\"><a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/51lqn9\/helm_or_ivy\/d7dmed8\">Comment<\/a>\nfrom discussion <a href=\"https:\/\/www.reddit.com\/r\/emacs\/comments\/51lqn9\/helm_or_ivy\/\">Helm\nor ivy?<\/a>.<\/div>\n<div class=\"footnotes\">\n<hr \/><ol><li id=\"fn:1\">The package that inspired Helm is called Anything by the way.\n <a class=\"footnote-return\" href=\"#fnref:1\"><sup>[return]<\/sup><\/a><\/li>\n<li id=\"fn:2\">The line at the bottom of your screen that you use when you type M-x for example.\n <a class=\"footnote-return\" href=\"#fnref:2\"><sup>[return]<\/sup><\/a><\/li>\n<\/ol><\/div>","created_at":"2018-11-07T16:40:32+0000","updated_at":"2018-11-07T16:41:41+0000","published_by":["Samuel Barreto"],"annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"sam217pa.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":240,"title":"Re: helm vs ido","url":"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00441.html","content":"Re: helm vs ido\n<div style=\"text-align: center;\">\n<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 2px; background-color: #000000;\"><tr><td><table border=\"0\" style=\"background-color: #FFFFCC;\"><tr><td><span style=\"font-size: larger;\"><b>help-gnu-emacs<\/b><\/span><\/td><\/tr><\/table><\/td><\/tr><\/table><p><small>[<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/\">Top<\/a>][<a href=\"https:\/\/lists.gnu.org\/archive\/html\">All Lists<\/a>]<\/small><\/p>\n<form method=\"get\" action=\"\/archive\/cgi-bin\/namazu.cgi\">\n<a href=\"https:\/\/lists.gnu.org\/archive\/cgi-bin\/namazu.cgi?idxname=help-gnu-emacs\">Advanced<\/a>\n<\/form>\n<\/div>\n\n\n\n\n<hr \/>\n[<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00440.html\">Date Prev<\/a>][<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00442.html\">Date Next<\/a>][<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00440.html\">Thread Prev<\/a>][<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00442.html\">Thread Next<\/a>][<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/index.html#00441\">Date Index<\/a>][<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/threads.html#00441\">Thread Index<\/a>]\n\n\n\n<hr \/><table border=\"0\"><tbody><tr><td align=\"right\" valign=\"top\">\n<b>From<\/b>: <\/td>\n<td align=\"left\">\nThierry Volpiatto<\/td>\n<\/tr><tr><td align=\"right\" valign=\"top\">\n<b>Subject<\/b>: <\/td>\n<td align=\"left\">\nRe: helm vs ido<\/td>\n<\/tr><tr><td align=\"right\" valign=\"top\">\n<b>Date<\/b>: <\/td>\n<td align=\"left\">\nSun, 26 Jan 2014 12:01:43 +0100<\/td>\n<\/tr><tr><td align=\"right\" valign=\"top\">\n<b>User-agent<\/b>: <\/td>\n<td align=\"left\">\nGnus\/5.13 (Gnus v5.13) Emacs\/24.3.50 (gnu\/linux)<\/td>\n<\/tr><\/tbody><\/table><hr \/><pre>Thorsten Jolitz &lt;address@hidden&gt; writes:\n&gt;<i> Hi List, <\/i>\n&gt;\n&gt;<i> I used ido-everywhere and smex for quite a while now as a happy<\/i>\n&gt;<i> user. Recently I discovered helm as an impressive replacement for<\/i>\n&gt;<i> autocomplete. However, helm is so powerful it might replace ido, smex,<\/i>\n&gt;<i> regexp-builder and other useful libraries. <\/i>\n&gt;\n&gt;<i> There are some aspects I really like about helm, but others not so much,<\/i>\n&gt;<i> e.g.:<\/i>\n&gt;\n&gt;<i> - how to get rid of helm-completion in the mini-buffer? In Ido, there is<\/i>\n&gt;<i>   C-f, but helm seems to follow more a bondage&amp;submission approach with<\/i>\n&gt;<i>   no way out (or I'm just a new user and did not discover the<\/i>\n&gt;<i>   trick). C-x C-w (write-file) is a good example were completion gets in<\/i>\n&gt;<i>   your way. <\/i>\nIf you want to get rid of helm completion in e.g `write-file' when\n`helm-mode' is enabled, and you want `write-file' use ido, customize\n`helm-completing-read-handlers-alist':\nAdd write-file =&gt; ido.\nNow when using C-x C-w you will have ido completion instead of helm\ncompletion.\n&gt;<i> - is ido-completion smarter than helm-completion? Somehow I have the<\/i>\n&gt;<i>   (obviously subjective) feeling that I reach the searched<\/i>\n&gt;<i>   buffer\/file\/command ... with less effort using Ido. One example might<\/i>\n&gt;<i>   be M-x w3m - with Ido its typing M-x w3 RET, with helm the basic w3m<\/i>\n&gt;<i>   command is not necessary the first choice and some navigation in the<\/i>\n&gt;<i>   completion buffer or some more typing is necessary. <\/i>\nYou are using the M-x version enabled with `helm-mode' which is the same\nthan vanilla emacs M-x but with helm interface.\nUse `helm-M-x' instead.\n&gt;<i> Right now I'm back to ido-everywhere, but on the other hand I miss<\/i>\n&gt;<i> helm-grep, helm-regexp and other amazing stuff this package offers. Do I<\/i>\n&gt;<i> simply have to study helm more deeply<\/i>\nYes, probably.\n&gt;<i> to make it replace all its predecessors without regrets, or are there<\/i>\n&gt;<i> actually use cases where ido et.al. are still better than helm?<\/i>\n-- \nThierry\nGet my Gnupg key:\ngpg --keyserver pgp.mit.edu --recv-keys 59F29997 \n<\/pre>\n\n\n\n<hr \/><form method=\"post\" action=\"\/mp\/yyz.py\" enctype=\"multipart\/form-data\">\n<div style=\"text-align: center;\">reply via email to<br \/><\/div>\n<\/form>\n<hr \/><table style=\"width: 100%;\"><tr><td align=\"left\">[<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00440.html\">Prev in Thread<\/a>]<\/td>\n<td align=\"center\"><b>Current Thread<\/b><\/td>\n<td align=\"right\">[<a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00442.html\">Next in Thread<\/a>]<\/td><\/tr><\/table><ul><li><b><a name=\"00440\" href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00440.html\" id=\"00440\">helm vs ido<\/a><\/b>, <i>Thorsten Jolitz<\/i>, <code>2014\/01\/26<\/code>\n<ul><li><span><strong>Re: helm vs ido<\/strong>,\n<em>Thierry Volpiatto<\/em><\/span>\u00a0<b>&lt;=<\/b>\n<ul><li><b><a name=\"00442\" href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00442.html\" id=\"00442\">Re: helm vs ido<\/a><\/b>, <i>Thorsten Jolitz<\/i>, <code>2014\/01\/26<\/code>\n<\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><hr \/><ul><li>Prev by Date:\n<strong><a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00440.html\">helm vs ido<\/a><\/strong>\n<\/li>\n<li>Next by Date:\n<strong><a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00442.html\">Re: helm vs ido<\/a><\/strong>\n<\/li>\n<li>Previous by thread:\n<strong><a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00440.html\">helm vs ido<\/a><\/strong>\n<\/li>\n<li>Next by thread:\n<strong><a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/msg00442.html\">Re: helm vs ido<\/a><\/strong>\n<\/li>\n<li>Index(es):\n<ul><li><a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/index.html#00441\"><strong>Date<\/strong><\/a><\/li>\n<li><a href=\"https:\/\/lists.gnu.org\/archive\/html\/help-gnu-emacs\/2014-01\/threads.html#00441\"><strong>Thread<\/strong><\/a><\/li>\n<\/ul><\/li>\n<\/ul>","created_at":"2018-11-07T16:40:32+0000","updated_at":"2018-11-07T16:41:43+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"lists.gnu.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":241,"title":"Introduction to Ido Mode - Mastering Emacs","url":"https:\/\/masteringemacs.org\/article\/introduction-to-ido-mode","content":"<p>By <em>Mickey Petersen<\/em><\/p><img src=\"https:\/\/masteringemacs.org\/static\/img\/fleuron2.gif\" class=\"img-responsive center-block\" alt=\"image\" \/><p class=\"text-justify drop-cap\">There are many ways of improving your productivity when you use Emacs, and Ido (or \u201cInteractively DO things\u201d) is one of those packages that you enable and then never, ever turn off again. It\u2019s simply that useful. By super-charging Emacs\u2019s completion engine and improving the speed at which you open files and buffers, you will significantly cut down on the time spent doing these menial tasks. Think about how often you switch buffers, open files or navigate directories in Emacs.<\/p><p class=\"text-justify\">When Ido is enabled it replaces the existing key-, and command bindings with its own, \u201cIdo-powered\u201d, versions. Most ido commands are named \u201cido-xxxxx\u201d; so <code>find-file<\/code> becomes <code>ido-find-file<\/code>, and so on.<\/p><p class=\"text-justify\">The \u201cIdo-powered\u201d completion engine matches anywhere in a name, instead of just the beginning. That means if you were to search for the buffer \u201c*Customize Group: Foobar*\u201d without Ido you\u2019d have to contort your fingers and type the *, then TAB (and hope it completes) and if not, type in some more; then rinse and repeat.<\/p><p class=\"text-justify\">With Ido you\u2019d type a few characters until you find the match, or until you narrow down your list of matches to a manageable subset, and then press RET. Even better, with fuzzy matching you can find things even <em>more<\/em> efficiently. So the above buffer name could be found by typing \u201ccgf\u201d \u2013 c for Customize, g for Group and f for Foobar. Magic.<\/p><p class=\"text-justify\">There are some who refuse to use Ido because it \u201cgets in the way\u201d; I think a lot of those worries stem from not knowing that it is possible to disable Ido\u2019s completion mode when it makes sense to do so.<\/p><h2 id=\"enable-ido-everywhere\">Enable Ido Everywhere<\/h2><p class=\"text-justify\">Add this to your <a href=\"https:\/\/masteringemacs.org\/articles\/2010\/10\/04\/beginners-guide-to-emacs\/\" class=\"article-link\">.emacs file<\/a>:<\/p><pre class=\"sourceCode commonlisp\">(setq ido-enable-flex-matching t)\n(setq ido-everywhere t)\n(ido-mode 1)<\/pre><p class=\"text-justify\">That\u2019ll enable basic Ido support for files and buffers and the very useful \u201cflex matching\u201d as well.<\/p><h3 id=\"my-customizations\">My Customizations<\/h3><p class=\"text-justify\">There\u2019s more to Ido than just enabling it. I\u2019ve compiled a list of interesting settings you can change and I have thrown in some of my preferences for good measure.<\/p><h4 id=\"find-file-at-point\">Find File At Point<\/h4><p class=\"text-justify\">Find File At Point, also known generally as \u201cffap\u201d, is an intelligent (well, when I say that, I mean it uses simple heuristics to infer the context at point) system for opening files, and URLs.<\/p><p class=\"text-justify\">Some people hate it, and personally I tolerate it. It gets in the way sometimes and when it does it\u2019s really annoying, but mostly it helps me so it gets to stay. It works in surprisingly many contexts like inferring the filepath of a library in an elisp <code>(require 'foobar)<\/code> form.<\/p><p class=\"text-justify\">This will make Ido guess the context, which is what I use.<\/p><pre class=\"sourceCode commonlisp\">(setq ido-use-filename-at-point 'guess)<\/pre><p class=\"text-justify\">You can disable URL ffap support by toggling <code>ido-use-url-at-point<\/code>.<\/p><h4 id=\"eliminating-obstacles\">Eliminating Obstacles<\/h4><p class=\"text-justify\">If you create a ton of throw-away buffers like I do, this setting will force Ido to always create a new buffer (in <code>C-x b<\/code>) if the name does not exist. I create throw-away buffers all the time and the last thing I need is Ido asking my permission to create one. Other choices are <code>prompt<\/code> and <code>never<\/code>.<\/p><pre class=\"sourceCode commonlisp\">(setq ido-create-new-buffer 'always)<\/pre><p class=\"text-justify\">You can customize the order in which files are sorted when Ido displays them in the minibuffer. There are certain file extensions I use more than others, so I tell Ido to emphasize those.<\/p><pre class=\"sourceCode commonlisp\">(setq ido-file-extensions-order '(\".org\" \".txt\" \".py\" \".emacs\" \".xml\" \".el\" \".ini\" \".cfg\" \".cnf\"))<\/pre><h4 id=\"ignorance-is-bliss\">Ignorance is Bliss<\/h4><p class=\"text-justify\">Ido is capable of ignoring (that is, by omission) buffers, directories, files and extensions using regexp.<\/p><table class=\"table\"><thead><\/thead><tbody><tr class=\"odd\"><td><code>ido-ignore-buffers<\/code><\/td><td>Takes a list of buffers to ignore in <code>C-x b<\/code><\/td><\/tr><tr class=\"even\"><td><code>ido-ignore-directories<\/code><\/td><td>Takes a list of directories to ignore in <code>C-x d<\/code> and <code>C-x C-f<\/code><\/td><\/tr><tr class=\"odd\"><td><code>ido-ignore-files<\/code><\/td><td>Takes a list of files to ignore in <code>C-x C-f<\/code><\/td><\/tr><\/tbody><\/table><p class=\"text-justify\">I would recommend using <code>M-x customize-variable RET variable-name-here<\/code> to customize them unless you are comfortabe editing Emacs lisp directly. I don\u2019t ignore anything in Ido beyond the defaults, as I prefer to use the more general <code>completion-ignored-extensions<\/code> as it works with and without Ido.<\/p><p class=\"text-justify\">To make Ido use <code>completion-ignored-extensions<\/code> you need to enable it:<\/p><pre class=\"sourceCode commonlisp\">(setq ido-ignore-extensions t)<\/pre><p class=\"text-justify\">Now you can customize that variable as well: <code>M-x customize-variable RET completion-ignored-extensions<\/code>. Go ahead and add all the useless object files, backup, autosave and other computing flotsam you don\u2019t want Ido to show.<\/p><p class=\"text-justify\"><strong>NOTE:<\/strong>Ido will <em>still<\/em> complete the ignored elements if Ido would otherwise not show any other matches. So if you type out the name of an ignored file Ido will still let you open it just fine.<\/p><h2 id=\"ido-cheatsheet\">Ido Cheatsheet<\/h2><p class=\"text-justify\">Here\u2019s a quick cheatsheet of all the interesting Ido commands.<\/p><p class=\"text-justify\">The \u201cAvailable In\u201d column indicates where the keybinding is available. \u201cAll\u201d means that it is a feature inherent to <code>ido-completing-read<\/code> and as such is available in all the completion engines. \u201cFiles\u201d, \u201cDirs\u201d and \u201cBuffers\u201d means it is only available in one of those.<\/p><h3 id=\"general-purpose-commands\">General-purpose Commands<\/h3><p class=\"text-justify\"><code>C-b<\/code> Reverts to the old <code>switch-buffer<\/code> completion engine. <em>Available in Buffers<\/em>.<\/p><p class=\"text-justify\"><code>C-f<\/code> Reverts to the old <code>find-file<\/code> completion engine. <em>Available in Files<\/em><\/p><p class=\"text-justify\"><code>C-d<\/code> Opens a dired buffer in the current directory. <em>Available in Dirs \/ Files<\/em><\/p><p class=\"text-justify\"><code>C-a<\/code> Toggles showing ignored files (see <code>ido-ignore-files<\/code>). <em>Available in Files \/ Buffers<\/em><\/p><p class=\"text-justify\"><code>C-c<\/code> Toggles if searching of buffer and file names should ignore case. (see <code>ido-case-fold<\/code>). <em>Available in Dirs \/ Files \/ Buffers<\/em><\/p><p class=\"text-justify\"><code>TAB<\/code> Attempt to complete the input like the normal completing read functionality. <em>Available in Dirs \/ Files \/ Buffers<\/em><\/p><p class=\"text-justify\"><code>C-p<\/code> Toggles prefix matching; when it\u2019s on the input will only match the beginning of a filename instead of any part of it.<\/p><p class=\"text-justify\"><code>Files<\/code><\/p><p class=\"text-justify\"><code>C-s \/ C-r<\/code> Moves to the next and previous match, respectively. <em>Available everywhere<\/em><\/p><p class=\"text-justify\"><code>C-t<\/code> Toggles matching by Emacs regular expression.. <em>Available everywhere<\/em><\/p><p class=\"text-justify\"><code>Backspace<\/code> Deletes characters as usual or goes up one directory if it makes sense to do so.. <em>Available everywhere<\/em><\/p><p class=\"text-justify\"><code>C-SPC \/ C-@<\/code> Restricts the completion list to anything that matches your current input. * Available everywhere*<\/p><p class=\"text-justify\"><code>\/\/<\/code> Like most Linux shells two forward slashes in a path means \u201cignore the preceding path, and go back to the top-most directory\u201d. Works the same in Ido but it\u2019s more interactive: it will go to the root <code>\/<\/code> (or the root of the current drive in Windows) <em>Available in Files<\/em><\/p><p class=\"text-justify\"><code>~\/<\/code> Jumps to the home directory. On Windows this would be typically be <code>%USERPROFILE%<\/code> or <code>%HOME%<\/code>, if it is defined. <em>Available in Files \/ Dirs<\/em><\/p><p class=\"text-justify\"><code>M-d<\/code> Searches for the input in all sub-directories to the directory you\u2019re in.. <em>Available in Files<\/em><\/p><p class=\"text-justify\"><code>C-k<\/code> Kills the currently focused buffer or <strong>deletes the file<\/strong> depending on the mode.. <em>Available in Files \/ Buffers<\/em><\/p><p class=\"text-justify\"><code>M-m<\/code> Creates a new sub-directory to the directory you\u2019re in. <em>Available in Files<\/em><\/p><p class=\"text-justify\">OK, so you probably won\u2019t get in the habit of using <em>all<\/em> the commands; that\u2019s fine, but some are more important to remember than others, like: <code>Backspace<\/code>; <code>C-s<\/code> and <code>C-r<\/code>; <code>\/\/<\/code> and <code>~\/<\/code>; and <code>C-d<\/code>.<\/p><p class=\"text-justify\">If Ido is getting in your way, remember the fallback commands: <code>C-f<\/code> for files; <code>C-b<\/code> for buffers.<\/p><h3 id=\"work-directory-commands\">Work Directory Commands<\/h3><p class=\"text-justify\">Work directories are recently used directories that Ido caches so the files and directories in them can be quickly recalled. Work directory commands only function in <code>find-file<\/code> routines, but that goes without saying.<\/p><p class=\"text-justify\">You change the cached directory list by manipulating the variable <code>ido-work-directory-list<\/code>. If you\u2019re tired of Ido caching slow, transient or generally unwanted directories you can add regular expressions to the list in <code>ido-work-directory-list-ignore-regexps<\/code> to keep Ido from caching them.<\/p><p class=\"text-justify\">Similarly, when you go to the next or previous work directory you can force Ido to disregard directories that do not match your current input. To change this behavior, set <code>ido-work-directory-match-only<\/code>.<\/p><table class=\"table\"><thead><\/thead><tbody><tr class=\"odd\"><td>M-n \/ M-p<\/td><td>Cycles through the next or previous work directories<\/td><\/tr><tr class=\"even\"><td>M-k<\/td><td>Kills (removes) the active work directory from the list.<\/td><\/tr><tr class=\"odd\"><td>M-s<\/td><td>M-s forces ido to search the list of work directories for the current input.<\/td><\/tr><\/tbody><\/table><p class=\"text-justify\">Using <code>M-s<\/code> is an excellent way to quickly scan recently used directories for files. You do not have to explicitly type <code>M-s<\/code> as Ido will automagically start searching after a few seconds of idle time.<\/p><p class=\"text-justify\">You can change the work directory merge delay by modifying the <code>ido-auto-merge-delay-time<\/code> variable.<\/p>","created_at":"2018-11-07T16:40:32+0000","updated_at":"2018-11-07T16:41:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"masteringemacs.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":237,"title":"Where are the best places to launch other than PH, HN, etc","url":"https:\/\/www.indiehackers.com\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:40:31+0000","updated_at":"2018-11-07T16:41:46+0000","annotations":[],"reading_time":0,"domain_name":"www.indiehackers.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":235,"title":"Ask HN: How does a great resume look like? What are the best tools to make one?","url":"https:\/\/news.ycombinator.com\/item?id=16659255","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16659255\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16659255\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=16659255\" class=\"storylink\">Ask HN: How does a great resume look like? What are the best tools to make one?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        161 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=rayalez\" class=\"hnuser\">rayalez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16659255\">7 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16659255&amp;goto=item%3Fid%3D16659255\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20How%20does%20a%20great%20resume%20look%20like%3F%20What%20are%20the%20best%20tools%20to%20make%20one%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20How%20does%20a%20great%20resume%20look%20like%3F%20What%20are%20the%20best%20tools%20to%20make%20one%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16659255&amp;auth=f5ee2137e99936911f43f88130c678fb477de26d\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16659255\">96\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Hey everyone! I'm a self-taught full-stack web developer. I've spent the past few years learning webdev, working on personal projects, and building a portfolio (https:\/\/startuplab.io\/portfolio).<p>Now I want to find good clients, or work remotely for a startup, and to do that I need to make a resume(or CV?).<\/p><p>So I wanted to ask a few questions:<\/p><p>- What should I put on my resume?<\/p><p>- Can you share some examples of how a great resume should look like?<\/p><p>- What are some of the best tools for making one?<\/p><p>- What do you look for when deciding to hire a developer?<\/p><p>- Can you share some advice that would help me increase my chances of finding a good job?<\/p><\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"16660625\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jaysonelliot\" class=\"hnuser\">jaysonelliot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660625\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660862\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=make3\" class=\"hnuser\">make3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660862\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661248\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walterbell\" class=\"hnuser\">walterbell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661248\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661500\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ollerac\" class=\"hnuser\">ollerac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661500\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661520\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walterbell\" class=\"hnuser\">walterbell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661520\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661699\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ollerac\" class=\"hnuser\">ollerac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661699\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661530\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nsxwolf\" class=\"hnuser\">nsxwolf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661530\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661795\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lukeschlather\" class=\"hnuser\">lukeschlather<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661795\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661334\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ollerac\" class=\"hnuser\">ollerac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661334\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661207\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adamb_\" class=\"hnuser\">adamb_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661207\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660642\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rconti\" class=\"hnuser\">rconti<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660642\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662380\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mlevental\" class=\"hnuser\">mlevental<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662380\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662402\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662402\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661544\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacob_rezi\" class=\"hnuser\">jacob_rezi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661544\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16674193\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flanbiscuit\" class=\"hnuser\">flanbiscuit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16674193\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16677411\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jacob_rezi\" class=\"hnuser\">jacob_rezi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16677411\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661122\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cv230498\" class=\"hnuser\">cv230498<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661122\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662365\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fenwick67\" class=\"hnuser\">fenwick67<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662365\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662571\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662571\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16659506\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=itamarst\" class=\"hnuser\">itamarst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16659506\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660705\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=johnmcd3\" class=\"hnuser\">johnmcd3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660705\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661434\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmichal\" class=\"hnuser\">jdmichal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661434\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661744\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=treis\" class=\"hnuser\">treis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661744\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662769\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmichal\" class=\"hnuser\">jdmichal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662769\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662595\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662595\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16666461\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blablabla123\" class=\"hnuser\">blablabla123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16666461\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16664202\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w0rd-driven\" class=\"hnuser\">w0rd-driven<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16664202\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661328\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=greatamerican\" class=\"hnuser\">greatamerican<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661328\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661710\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmichal\" class=\"hnuser\">jdmichal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661710\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662598\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662598\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662755\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmichal\" class=\"hnuser\">jdmichal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662755\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661533\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=veritas3241\" class=\"hnuser\">veritas3241<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661533\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661099\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=EvRev\" class=\"hnuser\">EvRev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661099\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662028\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Humdeee\" class=\"hnuser\">Humdeee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662028\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16675981\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tuananh\" class=\"hnuser\">tuananh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16675981\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661472\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhonsby\" class=\"hnuser\">rhonsby<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661472\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661589\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wyattk\" class=\"hnuser\">wyattk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661589\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662067\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Humdeee\" class=\"hnuser\">Humdeee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662067\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662627\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662627\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662602\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrewmcwatters\" class=\"hnuser\">andrewmcwatters<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662602\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661269\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walterbell\" class=\"hnuser\">walterbell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661269\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16663243\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bb88\" class=\"hnuser\">bb88<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16663243\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16667084\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hypertexthero\" class=\"hnuser\">hypertexthero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16667084\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16664948\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=imauld\" class=\"hnuser\">imauld<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16664948\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661140\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=karmapolic\" class=\"hnuser\">karmapolic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661140\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661182\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sametmax\" class=\"hnuser\">sametmax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661182\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661263\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adamb_\" class=\"hnuser\">adamb_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661263\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661450\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maxk42\" class=\"hnuser\">maxk42<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661450\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660693\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Blackstone4\" class=\"hnuser\">Blackstone4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660693\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661066\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=minouye\" class=\"hnuser\">minouye<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661066\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661086\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=iamdave\" class=\"hnuser\">iamdave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661086\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662675\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662675\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661552\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhonsby\" class=\"hnuser\">rhonsby<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661552\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661695\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=monocasa\" class=\"hnuser\">monocasa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661695\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16663281\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bb88\" class=\"hnuser\">bb88<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16663281\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662678\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fyfy18\" class=\"hnuser\">fyfy18<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662678\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16666659\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cuchoi\" class=\"hnuser\">cuchoi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16666659\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16663690\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=abhishekjha\" class=\"hnuser\">abhishekjha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16663690\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16666506\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=astkaasa\" class=\"hnuser\">astkaasa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16666506\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660616\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lessclue\" class=\"hnuser\">lessclue<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660616\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662432\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=PenguinCoder\" class=\"hnuser\">PenguinCoder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662432\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661134\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BreakoutList\" class=\"hnuser\">BreakoutList<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661134\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662745\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nfriedly\" class=\"hnuser\">nfriedly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662745\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16664018\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rohanm93\" class=\"hnuser\">rohanm93<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16664018\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661833\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dilatedmind\" class=\"hnuser\">dilatedmind<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661833\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661181\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mosster\" class=\"hnuser\">mosster<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661181\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661027\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nathan_f77\" class=\"hnuser\">nathan_f77<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661027\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662537\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkya\" class=\"hnuser\">gkya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662537\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660872\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=otalp\" class=\"hnuser\">otalp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660872\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660800\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=peterwwillis\" class=\"hnuser\">peterwwillis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660800\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660907\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DoreenMichele\" class=\"hnuser\">DoreenMichele<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660907\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661057\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=realPubkey\" class=\"hnuser\">realPubkey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661057\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662682\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrewmcwatters\" class=\"hnuser\">andrewmcwatters<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662682\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661175\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=megadeth\" class=\"hnuser\">megadeth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661175\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661382\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelbuckbee\" class=\"hnuser\">michaelbuckbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661382\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662620\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Johnny555\" class=\"hnuser\">Johnny555<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662620\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661648\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wyattk\" class=\"hnuser\">wyattk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661648\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661671\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=monocasa\" class=\"hnuser\">monocasa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661671\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661252\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rapfaria\" class=\"hnuser\">rapfaria<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661252\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661585\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jedberg\" class=\"hnuser\">jedberg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661585\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661947\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kreeWall\" class=\"hnuser\">kreeWall<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661947\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661026\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=greenleafjacob\" class=\"hnuser\">greenleafjacob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661026\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661363\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jcadam\" class=\"hnuser\">jcadam<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661363\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16664250\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lugg\" class=\"hnuser\">lugg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16664250\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661380\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=soneca\" class=\"hnuser\">soneca<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661380\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662831\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ketan_anjaria\" class=\"hnuser\">ketan_anjaria<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662831\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661101\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Walkman\" class=\"hnuser\">Walkman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661101\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662134\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Humdeee\" class=\"hnuser\">Humdeee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662134\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16662934\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tranchms\" class=\"hnuser\">tranchms<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16662934\">7 months ago<\/a><\/p><br \/><div class=\"comment\">I approach the job search as a lengthy process of apply, reflect, refine, repeat.<p>I won\u2019t get into the resume writing aspect, since there are countless resources out there, so I\u2019ll just say this: identify jobs you\u2019d like to have, aggregate the required skills and qualifications, then write your resume that highlights and frames everything that fits those requirements.<\/p><p>I usually apply to hundreds of jobs. Some I may want. Some I may not want. When I get an interview, I try to learn everything I can about the position and the company, and I sell myself to the best of my ability. I always ask a lot of questions: what are the challenges of the job? What is the ideal candidate? What do you like about my skills and experiences? What don\u2019t you like? Etc. I then fill in the gaps in my resume and cover letter.<\/p><p>In the beginning of a new job search, i usually don\u2019t get many interview requests. My resume may need work, and I\u2019m usually not great at selling myself.<\/p><p>But due to the sheer volume of applications, I do get interviews which in turn allow me to collect data on companies and positions so I can refine my resume and my approach to the interview process,and become acquainted with the questions and how to pitch myself, etc.<\/p><p>I always try to get a final interview, even if I\u2019m not sold on the job. Recruiters and hiring managers probably dislike people like me who go the distance only to decline the offer. But it\u2019s invaluable experience, and you\u2019re interviewing companies as much as they are interviewing you. Remember that.<\/p><p>After a few weeks of getting warmed up, I apply to my choice companies\/ positions.<\/p><p>By this time I am polished, confident, and know the interview process inside and out, and am familiar with the positions and responsibilities that I\u2019m aiming for.<\/p><p>This method is highly effective, but lengthy, and time consuming.<\/p><p>The exposure to all the recruiters, going through the interview process, asking tons of questions and collecting information about how the business works, how the departments work, the responsibilities of the role, the processes, the culture, advancement, etc is invaluable experience that you can leverage when interviewing for the companies\/ positions you\u2019d really like.<\/p><p>As an example, I applied to about 10 companies a day for about 8 weeks before I landed my ideal job. Starting out I received maybe 1-3 interview requests the first week or two, and rarely getting past the first and second interview.<\/p><p>After refining the resume and polishing my interview skills, I was receiving 5-6 interviews a week, had to turn down many offers and final interviews requiring travel that I knew I wouldn\u2019t take.<\/p><p>This really gave me confidence. It allowed me to negotiate salary. And I knew if I walked away there were other offers waiting.<\/p><p>Starting out, the job search can be daunting. But just dive in. Apply, reflect, refine, repeat. You\u2019ll know your worth, and learn to sell yourself.<\/p><p>Best of luck!\n              <\/p><\/div><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660486\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neduma\" class=\"hnuser\">neduma<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660486\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660540\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flanbiscuit\" class=\"hnuser\">flanbiscuit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660540\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661859\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mythrwy\" class=\"hnuser\">mythrwy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661859\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16660521\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TremendousJudge\" class=\"hnuser\">TremendousJudge<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16660521\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661511\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=guessmyname\" class=\"hnuser\">guessmyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661511\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16665677\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lee101\" class=\"hnuser\">lee101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16665677\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16661307\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=greatamerican\" class=\"hnuser\">greatamerican<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16661307\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:40:26+0000","updated_at":"2018-11-07T16:41:47+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":236,"title":"Quickly find Emacs Lisp sources \u00b7 Emacs Redux","url":"http:\/\/emacsredux.com\/blog\/2014\/06\/18\/quickly-find-emacs-lisp-sources\/","content":"<p>One thing Emacs users deal all the time is looking for the source of\nvarious libraries and function or variable definitions.  Typically this\ninvolves the use of commands like <code class=\"highlighter-rouge\">C-h f<\/code> (<code class=\"highlighter-rouge\">describe-function<\/code>), <code class=\"highlighter-rouge\">C-h v<\/code>\n(<code class=\"highlighter-rouge\">describe-variable<\/code>) and <code class=\"highlighter-rouge\">C-h k<\/code> (<code class=\"highlighter-rouge\">describe-key<\/code>) and jumping to the source from the help buffer those commands produce.\nLet\u2019s see if we can improve upon this workflow.<\/p><h3 id=\"find-func\">find-func<\/h3><p>Enter the built-in Emacs library <code class=\"highlighter-rouge\">find-func<\/code>. It provides a myriad of useful commands to help us quickly locate sources.\nLet\u2019s take a brief look at some of them.<\/p><ul><li>You can find the source of a library with <code class=\"highlighter-rouge\">M-x find-library<\/code>. If you want to find the source of <code class=\"highlighter-rouge\">ido<\/code> you can do so like this:<\/li>\n<\/ul><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-library RET ido RET\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-l<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-l\") 'find-library)\n<\/pre><\/div><\/div><ul><li>You can find the source of a function definition with <code class=\"highlighter-rouge\">M-x find-function<\/code>. Here\u2019s an example:<\/li>\n<\/ul><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-function RET find-function RET\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-f<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-f\") 'find-function)\n<\/pre><\/div><\/div><p>We can actually do one better - we can directly jump to a command\ndefinition using a keybinding of the command with <code class=\"highlighter-rouge\">M-x\nfind-function-on-key<\/code>. Here\u2019s how we can find the source the command\nbound to <code class=\"highlighter-rouge\">C-a<\/code> (<code class=\"highlighter-rouge\">beginning-of-line<\/code>):<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-function-on-key RET C-a\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-k<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-k\") 'find-function-on-key)\n<\/pre><\/div><\/div><ul><li>You can find the source of a variable definition with <code class=\"highlighter-rouge\">M-x find-variable<\/code>. Here\u2019s an example:<\/li>\n<\/ul><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>M-x find-variable RET large-file-warning-threshold RET\n<\/pre><\/div><\/div><p>I\u2019d suggest binding the command to <code class=\"highlighter-rouge\">C-h C-v<\/code> for quick access:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(define-key 'help-command (kbd \"C-v\") 'find-variable)\n<\/pre><\/div><\/div><hr \/><p>The library provides other useful commands as well - like\n<code class=\"highlighter-rouge\">find-function-at-point<\/code> and <code class=\"highlighter-rouge\">find-variable-at-point<\/code>.<\/p><p>If you don\u2019t like the keybindings I suggested you can use\n<code class=\"highlighter-rouge\">find-function-setup-keys<\/code> instead.  This will give you keybindings\nlike <code class=\"highlighter-rouge\">C-x F<\/code>, <code class=\"highlighter-rouge\">C-x V<\/code>, <code class=\"highlighter-rouge\">C-f K<\/code> (plus a few extra for commands like\n<code class=\"highlighter-rouge\">find-funtion-other-window<\/code>).<\/p><p>Another really cool way to browse Elisp sources (and documentation) is the third-party package\n<a href=\"https:\/\/github.com\/purcell\/elisp-slime-nav\">elisp-slime-nav<\/a>. Assuming you\u2019ve installed it already you can\nenabled it like this:<\/p><div class=\"language-elisp highlighter-rouge\"><div class=\"highlight\"><pre>(require 'elisp-slime-nav)\n(dolist (hook '(emacs-lisp-mode-hook ielm-mode-hook))\n(add-hook hook 'elisp-slime-nav-mode))\n<\/pre><\/div><\/div><p>Once this is done you\u2019ll be able to jump to the source of the Emacs\nLisp object at point (function or variable) with <code class=\"highlighter-rouge\">M-.<\/code> (as in SLIME\nand CIDER for Common Lisp and Clojure respectively) and jump back with\n<code class=\"highlighter-rouge\">M-,<\/code>. You can also see the description of the object at point using\n<code class=\"highlighter-rouge\">C-c C-d<\/code> (or <code class=\"highlighter-rouge\">C-c C-d d<\/code>).<\/p><p>That\u2019s all for today, folks!<\/p><p><strong>P.S.<\/strong><\/p><p>All the suggested keybindings are present out-of-the-box in\n<a href=\"https:\/\/github.com\/bbatsov\/prelude\">Prelude<\/a>. <code class=\"highlighter-rouge\">elisp-slime-nav<\/code> is also\nenabled out-of-the box in Prelude.<\/p>","created_at":"2018-11-07T16:40:26+0000","updated_at":"2018-11-07T16:41:49+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"emacsredux.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":233,"title":"Lockbox \u2013 A stand-alone password manager that works with Firefox for desktop","url":"https:\/\/news.ycombinator.com\/item?id=16677905","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680150\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=acabal\" class=\"hnuser\">acabal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681529\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=333c\" class=\"hnuser\">333c<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682934\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=freedomben\" class=\"hnuser\">freedomben<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680942\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Deimorz\" class=\"hnuser\">Deimorz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681866\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=craftyguy\" class=\"hnuser\">craftyguy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681229\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681815\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xfeba\" class=\"hnuser\">0xfeba<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681898\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=craftyguy\" class=\"hnuser\">craftyguy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682516\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681787\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=some_account\" class=\"hnuser\">some_account<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681754\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=syx\" class=\"hnuser\">syx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680603\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexnewman\" class=\"hnuser\">alexnewman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683772\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=topfun\" class=\"hnuser\">topfun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681757\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kej\" class=\"hnuser\">kej<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681999\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=topfun\" class=\"hnuser\">topfun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682118\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681804\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JetSpiegel\" class=\"hnuser\">JetSpiegel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682970\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679643\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mastax\" class=\"hnuser\">mastax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679822\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679969\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=josteink\" class=\"hnuser\">josteink<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680769\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=akquise\" class=\"hnuser\">akquise<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681589\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mintplant\" class=\"hnuser\">mintplant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681963\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=the8472\" class=\"hnuser\">the8472<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682114\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mintplant\" class=\"hnuser\">mintplant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680954\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682689\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mintplant\" class=\"hnuser\">mintplant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683667\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mariusmg\" class=\"hnuser\">mariusmg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682928\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16687870\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e1ven\" class=\"hnuser\">e1ven<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16688585\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xTJ\" class=\"hnuser\">0xTJ<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679740\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=l1ambda\" class=\"hnuser\">l1ambda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681006\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=g4k\" class=\"hnuser\">g4k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682519\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=batisteo\" class=\"hnuser\">batisteo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682560\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paulryanrogers\" class=\"hnuser\">paulryanrogers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16684260\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pbreit\" class=\"hnuser\">pbreit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680181\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=floatingatoll\" class=\"hnuser\">floatingatoll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680487\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=laurent123456\" class=\"hnuser\">laurent123456<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679820\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanbura\" class=\"hnuser\">hanbura<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681239\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=laurent123456\" class=\"hnuser\">laurent123456<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16686998\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanklazard\" class=\"hnuser\">hanklazard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682148\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mariusmg\" class=\"hnuser\">mariusmg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682982\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanbura\" class=\"hnuser\">hanbura<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16684288\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zaarn\" class=\"hnuser\">zaarn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16696293\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paulryanrogers\" class=\"hnuser\">paulryanrogers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16684273\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=staticassertion\" class=\"hnuser\">staticassertion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683252\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mehrdadn\" class=\"hnuser\">mehrdadn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681672\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ealhad\" class=\"hnuser\">ealhad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679547\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maccard\" class=\"hnuser\">maccard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679594\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Digit-Al\" class=\"hnuser\">Digit-Al<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679684\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=slavik81\" class=\"hnuser\">slavik81<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681033\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maccard\" class=\"hnuser\">maccard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682726\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhasse\" class=\"hnuser\">jhasse<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682608\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maccard\" class=\"hnuser\">maccard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682706\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flanbiscuit\" class=\"hnuser\">flanbiscuit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680755\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neumann\" class=\"hnuser\">neumann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683493\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=binaryanomaly\" class=\"hnuser\">binaryanomaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679134\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bhhaskin\" class=\"hnuser\">bhhaskin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679832\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fatbob\" class=\"hnuser\">fatbob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681365\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=binaryanomaly\" class=\"hnuser\">binaryanomaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680259\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=remir\" class=\"hnuser\">remir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679790\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=binaryanomaly\" class=\"hnuser\">binaryanomaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680293\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e12e\" class=\"hnuser\">e12e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682453\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zie\" class=\"hnuser\">zie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683434\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=durin42\" class=\"hnuser\">durin42<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682276\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexghr\" class=\"hnuser\">alexghr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679838\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681289\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexghr\" class=\"hnuser\">alexghr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16686213\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=devinreams\" class=\"hnuser\">devinreams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681825\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tmd83\" class=\"hnuser\">tmd83<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679167\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sametmax\" class=\"hnuser\">sametmax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679296\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680692\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zaarn\" class=\"hnuser\">zaarn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679219\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tmd83\" class=\"hnuser\">tmd83<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679457\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rookwood102\" class=\"hnuser\">rookwood102<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679885\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=floatingatoll\" class=\"hnuser\">floatingatoll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679462\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmulv\" class=\"hnuser\">pmulv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678795\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=raimue\" class=\"hnuser\">raimue<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679678\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=conradev\" class=\"hnuser\">conradev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682015\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=333c\" class=\"hnuser\">333c<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683040\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Flimm\" class=\"hnuser\">Flimm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679804\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680670\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fwdpropaganda\" class=\"hnuser\">fwdpropaganda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678835\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blowski\" class=\"hnuser\">blowski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678920\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ealhad\" class=\"hnuser\">ealhad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678941\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bonyt\" class=\"hnuser\">bonyt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678987\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blowski\" class=\"hnuser\">blowski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679015\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xylakant\" class=\"hnuser\">Xylakant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678997\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679224\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xylakant\" class=\"hnuser\">Xylakant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679360\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679413\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ams6110\" class=\"hnuser\">ams6110<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678982\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679211\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blowski\" class=\"hnuser\">blowski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678980\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dwighttk\" class=\"hnuser\">dwighttk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678940\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daveFNbuck\" class=\"hnuser\">daveFNbuck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679780\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fwdpropaganda\" class=\"hnuser\">fwdpropaganda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680106\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daveFNbuck\" class=\"hnuser\">daveFNbuck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680300\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KirinDave\" class=\"hnuser\">KirinDave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679915\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680660\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=toyg\" class=\"hnuser\">toyg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678839\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678998\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tEMporality7\" class=\"hnuser\">tEMporality7<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679252\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679535\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xylakant\" class=\"hnuser\">Xylakant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679037\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=NikolaeVarius\" class=\"hnuser\">NikolaeVarius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679458\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rmurri\" class=\"hnuser\">rmurri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679742\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namitutonka\" class=\"hnuser\">namitutonka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16726537\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namitutonka\" class=\"hnuser\">namitutonka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16726607\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=byproxy\" class=\"hnuser\">byproxy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679387\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nullabillity\" class=\"hnuser\">Nullabillity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679163\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pibefision\" class=\"hnuser\">pibefision<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679018\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jopsen\" class=\"hnuser\">jopsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679667\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:40:25+0000","updated_at":"2018-11-07T16:41:50+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":234,"title":"Brandmark Logo Maker - the most advanced AI logo design tool","url":"https:\/\/brandmark.io\/","content":"Brandmark Logo Maker - the most advanced AI logo design tool\n\t<section id=\"main\"><div class=\"hero-wrapper\"><div class=\"container\"><div id=\"logo-container\"><div class=\"logo\"><img id=\"logo_brand\" src=\"https:\/\/brandmark.io\/svg-brand-2.svg\" width=\"455\" height=\"255\" alt=\"Brand\" \/><img id=\"logo_mark\" src=\"https:\/\/brandmark.io\/svg-mark-2.svg\" width=\"455\" height=\"255\" alt=\"Mark\" \/><\/div><div class=\"welcome\"><h4>complete with business card designs, social media graphics, app icons and more<\/h4><p><a href=\"https:\/\/app.brandmark.io\/v2\/\" target=\"_blank\" class=\"btn btn-primary\">Create your logo<\/a><\/p><\/div><\/div><\/div><\/div><\/section><section id=\"featured\"><section id=\"samples\"><div class=\"container\"><div class=\"row\"><div class=\"col-md-12\"><img src=\"https:\/\/brandmark.io\/img\/sample.jpg\" width=\"1600\" height=\"870\" class=\"big\" alt=\"real life logo\" \/><\/div><div class=\"col-md-5 col-md-offset-1 pull-up\"><h2>Latest customer creations<\/h2><p>See why over 108,704 users have created a logo with Brandmark.io<\/p><\/div><\/div><\/div><\/section><section id=\"bookend\"><div class=\"hero-wrapper\"><div class=\"container\"><div class=\"yellow-box\"><div class=\"row\"><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/design.svg\" alt=\"design icon\" \/><h3>No more designer anxiety<\/h3><p>Only pay when you're happy. Don't spend time and money on uncertain results.<\/p><\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/time.svg\" alt=\"time icon\" \/><h3>On time and on budget<\/h3>Get your logo and design assets instantly, not in weeks or months with a design agency.<\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/copyright.svg\" alt=\"copyright icon\" \/><h3>Full copyright<\/h3>All packages include full rights for use on your website, print media, products and more.<\/div><\/div><\/div><div class=\"row\"><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/coin.svg\" alt=\"coin icon\" \/><h3>Affordable one-time purchase<\/h3><p>Pay once, gain access to our branding tools forever. No monthly charges.<\/p><\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/infinite.svg\" alt=\"infinite icon\" \/><h3>Unlimited revisions<\/h3>Modify your logo at any time, even after purchase.<\/div><\/div><div class=\"col-sm-4\"><div class=\"icon-area\"><img src=\"https:\/\/brandmark.io\/img\/support.svg\" alt=\"support icon\" \/><h3>Your personal design team<\/h3>Need a tweak to your design? We'll do it for you, absolutely free.<\/div><\/div><\/div><p><\/p><h2>Create your unique logo design, try it free!<\/h2>\t\t\t\n\t\t\t\t\t<a href=\"https:\/\/app.brandmark.io\/v2\/\" target=\"_blank\" class=\"btn btn-primary\">\n\t\t\t\t\t\tDesign your logo\n\t\t\t\t\t<\/a><\/div><\/div><\/div><\/section><\/section>","created_at":"2018-11-07T16:40:25+0000","updated_at":"2018-11-07T16:41:51+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"brandmark.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":231,"title":"Best Fallout 4 mods","url":"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/","content":"<div><main><p>Now that <a href=\"https:\/\/www.rockpapershotgun.com\/game\/fallout-4\/\">Fallout 4<\/a> is in its third year and the DLC has dried up, is there any point in returning to the Commonwealth Wasteland? Yes! There are loads, largely thanks to the still-growing list of mods, overhauls and user tweaks. Here, I\u2019ve gathered over 50 of my favourites, ranging from weird weapons to wild weather.<\/p>\n<p>Before we start, a couple of things to remember: some mods will require some or all of the DLC expansions, as well as additional mods, while others don\u2019t play nicely with each other. The mod descriptions on <a href=\"https:\/\/www.nexusmods.com\/fallout4\">Nexus<\/a> will usually tell you, so keep an eye out.<\/p>\n<p>Fallout 4 mods come in all shapes and sizes and so we\u2019ve split this guide into sections, so you can hop directly to the kind of mod you\u2019re looking for.<\/p>\n<p><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/2\/\">How to install Fallout 4 mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/3\/\">Fallout 4 graphics mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/4\/\">Fallout 4 settlement mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/5\/\">Fallout 4 weapon and armour mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/6\/\">Fallout 4 companion mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/7\/\">Fallout 4 quests and overhaul mods<\/a><br \/><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/8\/\">Fallout 4 UI mods<\/a><\/p>\n\t\t\t\t\t<p>Page 1<a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/2\/\">Page 2<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/3\/\">Page 3<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/4\/\">Page 4<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/5\/\">Page 5<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/6\/\">Page 6<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/7\/\">Page 7<\/a><a href=\"https:\/\/www.rockpapershotgun.com\/2018\/10\/05\/best-fallout-4-mods-2\/8\/\">Page 8<\/a><\/p>\t\t\t\t<\/main><\/div><div class=\"container-690\"><p class=\"disclaimer\">\n\t\t\t\t\tIf you click our links to online stores and make a purchase we may receive a few pennies. Find more information <a href=\"https:\/\/www.rockpapershotgun.com\/affiliate-linking-policy\/\">here<\/a>.\t\t\t\t<\/p><a href=\"https:\/\/www.rockpapershotgun.com\/support-us?utm_source=article-promo&amp;utm_medium=web&amp;utm_campaign=supporter\" class=\"support-us-promo\" rel=\"nofollow\">\n\t\t\n\t\t\t<img src=\"https:\/\/www.rockpapershotgun.com\/wp-content\/themes\/rps-2018\/images\/logo-inverted.svg\" alt=\"Rock Paper Shotgun\" \/>\n\t\t\n\t\t\n\t\t\t\n\t\t\t\tSupport RPS and get an ad-free site, extra articles, and free stuff!\t\t\t\n\t\t\tTell me more\n\t\t\n\t\t\n\t<\/a><\/div>","created_at":"2018-11-07T16:40:23+0000","updated_at":"2018-11-07T16:41:53+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/assets.rockpapershotgun.com\/images\/2018\/10\/Homemaker-1212x682.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":232,"title":"Lockbox \u2013 A stand-alone password manager that works with Firefox for desktop","url":"https:\/\/news.ycombinator.com\/item?id=16677905","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680150\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=acabal\" class=\"hnuser\">acabal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681529\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=333c\" class=\"hnuser\">333c<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682934\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=freedomben\" class=\"hnuser\">freedomben<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680942\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Deimorz\" class=\"hnuser\">Deimorz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681866\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=craftyguy\" class=\"hnuser\">craftyguy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681229\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681815\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xfeba\" class=\"hnuser\">0xfeba<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681898\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=craftyguy\" class=\"hnuser\">craftyguy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682516\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681787\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=some_account\" class=\"hnuser\">some_account<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681754\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=syx\" class=\"hnuser\">syx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680603\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexnewman\" class=\"hnuser\">alexnewman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683772\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=topfun\" class=\"hnuser\">topfun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681757\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kej\" class=\"hnuser\">kej<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681999\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=topfun\" class=\"hnuser\">topfun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682118\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cajdavidson\" class=\"hnuser\">cajdavidson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681804\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JetSpiegel\" class=\"hnuser\">JetSpiegel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682970\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679643\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mastax\" class=\"hnuser\">mastax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679822\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679969\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=josteink\" class=\"hnuser\">josteink<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680769\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=akquise\" class=\"hnuser\">akquise<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681589\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mintplant\" class=\"hnuser\">mintplant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681963\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=the8472\" class=\"hnuser\">the8472<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682114\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mintplant\" class=\"hnuser\">mintplant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680954\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682689\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mintplant\" class=\"hnuser\">mintplant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683667\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mariusmg\" class=\"hnuser\">mariusmg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682928\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawl\" class=\"hnuser\">lawl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16687870\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e1ven\" class=\"hnuser\">e1ven<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16688585\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xTJ\" class=\"hnuser\">0xTJ<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679740\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=l1ambda\" class=\"hnuser\">l1ambda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681006\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=g4k\" class=\"hnuser\">g4k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682519\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=batisteo\" class=\"hnuser\">batisteo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682560\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paulryanrogers\" class=\"hnuser\">paulryanrogers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16684260\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pbreit\" class=\"hnuser\">pbreit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680181\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=floatingatoll\" class=\"hnuser\">floatingatoll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680487\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=laurent123456\" class=\"hnuser\">laurent123456<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679820\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanbura\" class=\"hnuser\">hanbura<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681239\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=laurent123456\" class=\"hnuser\">laurent123456<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16686998\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanklazard\" class=\"hnuser\">hanklazard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682148\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mariusmg\" class=\"hnuser\">mariusmg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682982\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hanbura\" class=\"hnuser\">hanbura<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16684288\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zaarn\" class=\"hnuser\">zaarn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16696293\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paulryanrogers\" class=\"hnuser\">paulryanrogers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16684273\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=staticassertion\" class=\"hnuser\">staticassertion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683252\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mehrdadn\" class=\"hnuser\">mehrdadn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681672\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ealhad\" class=\"hnuser\">ealhad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679547\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maccard\" class=\"hnuser\">maccard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679594\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Digit-Al\" class=\"hnuser\">Digit-Al<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679684\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=slavik81\" class=\"hnuser\">slavik81<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681033\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maccard\" class=\"hnuser\">maccard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682726\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhasse\" class=\"hnuser\">jhasse<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682608\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maccard\" class=\"hnuser\">maccard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682706\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flanbiscuit\" class=\"hnuser\">flanbiscuit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680755\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neumann\" class=\"hnuser\">neumann<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683493\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=binaryanomaly\" class=\"hnuser\">binaryanomaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679134\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bhhaskin\" class=\"hnuser\">bhhaskin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679832\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fatbob\" class=\"hnuser\">fatbob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681365\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=binaryanomaly\" class=\"hnuser\">binaryanomaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680259\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=remir\" class=\"hnuser\">remir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679790\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=binaryanomaly\" class=\"hnuser\">binaryanomaly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680293\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e12e\" class=\"hnuser\">e12e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682453\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zie\" class=\"hnuser\">zie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683434\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=durin42\" class=\"hnuser\">durin42<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682276\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexghr\" class=\"hnuser\">alexghr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679838\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681289\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexghr\" class=\"hnuser\">alexghr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16686213\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=devinreams\" class=\"hnuser\">devinreams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16681825\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tmd83\" class=\"hnuser\">tmd83<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679167\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sametmax\" class=\"hnuser\">sametmax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679296\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680692\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zaarn\" class=\"hnuser\">zaarn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679219\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tmd83\" class=\"hnuser\">tmd83<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679457\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rookwood102\" class=\"hnuser\">rookwood102<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679885\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=floatingatoll\" class=\"hnuser\">floatingatoll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679462\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmulv\" class=\"hnuser\">pmulv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678795\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=raimue\" class=\"hnuser\">raimue<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679678\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=conradev\" class=\"hnuser\">conradev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16682015\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=333c\" class=\"hnuser\">333c<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16683040\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Flimm\" class=\"hnuser\">Flimm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679804\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680670\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fwdpropaganda\" class=\"hnuser\">fwdpropaganda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678835\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blowski\" class=\"hnuser\">blowski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678920\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ealhad\" class=\"hnuser\">ealhad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678941\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bonyt\" class=\"hnuser\">bonyt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678987\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blowski\" class=\"hnuser\">blowski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679015\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xylakant\" class=\"hnuser\">Xylakant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678997\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679224\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xylakant\" class=\"hnuser\">Xylakant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679360\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679413\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ams6110\" class=\"hnuser\">ams6110<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678982\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zeveb\" class=\"hnuser\">zeveb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679211\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blowski\" class=\"hnuser\">blowski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678980\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dwighttk\" class=\"hnuser\">dwighttk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678940\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daveFNbuck\" class=\"hnuser\">daveFNbuck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679780\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fwdpropaganda\" class=\"hnuser\">fwdpropaganda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680106\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daveFNbuck\" class=\"hnuser\">daveFNbuck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680300\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KirinDave\" class=\"hnuser\">KirinDave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679915\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cpeterso\" class=\"hnuser\">cpeterso<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16680660\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=toyg\" class=\"hnuser\">toyg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678839\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16678998\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tEMporality7\" class=\"hnuser\">tEMporality7<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679252\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679535\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Xylakant\" class=\"hnuser\">Xylakant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679037\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=NikolaeVarius\" class=\"hnuser\">NikolaeVarius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679458\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rmurri\" class=\"hnuser\">rmurri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679742\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namitutonka\" class=\"hnuser\">namitutonka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16726537\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namitutonka\" class=\"hnuser\">namitutonka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16726607\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=byproxy\" class=\"hnuser\">byproxy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679387\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nullabillity\" class=\"hnuser\">Nullabillity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679163\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pibefision\" class=\"hnuser\">pibefision<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679018\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jopsen\" class=\"hnuser\">jopsen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16679667\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:40:23+0000","updated_at":"2018-11-07T16:41:54+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":230,"title":"jdee-emacs\/jdee","url":"https:\/\/github.com\/jdee-emacs\/jdee","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"http:\/\/melpa.org\/#\/jdee\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/569654b409b3297f90449a68fb1f8ea03e9beab0\/687474703a2f2f6d656c70612e6f72672f7061636b616765732f6a6465652d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"http:\/\/melpa.org\/packages\/jdee-badge.svg\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/jdee-emacs\/jdee\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/db27309327741fe84a82083761f1256e50a2d2f8\/68747470733a2f2f7472617669732d63692e6f72672f6a6465652d656d6163732f6a6465652e706e673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/jdee-emacs\/jdee.png?branch=master\" \/><\/a><\/p>\n<p>The JDEE is an add-on software package that turns Emacs into a\ncomprehensive system for creating, editing, debugging, and documenting\nJava applications.<\/p>\n<h2><a id=\"user-content-about-the-project\" class=\"anchor\" aria-hidden=\"true\" href=\"#about-the-project\"><\/a>About the project<\/h2>\n<p>As of 2015-07-13, <a href=\"https:\/\/github.com\/jdee-emacs\/jdee\/\">https:\/\/github.com\/jdee-emacs\/jdee\/<\/a> is the primary source repository.<\/p>\n<p>See <a href=\"https:\/\/github.com\/jdee-emacs\/jdee\/blob\/master\/CHANGES.md\">CHANGES.md<\/a> for migration instructions.<\/p>\n<p>Additional information can be found at <a href=\"http:\/\/jdee.sourceforge.net\/rootpage.html\" rel=\"nofollow\">http:\/\/jdee.sourceforge.net\/rootpage.html<\/a> and at <a href=\"https:\/\/github.com\/jdee-emacs\/jdee\/wiki\">https:\/\/github.com\/jdee-emacs\/jdee\/wiki<\/a><\/p>\n<h2><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h2>\n<p>Emacs 24.4 is the oldest version that JDEE can be expected to run in.<\/p>\n<p>If you install JDEE through the Emacs package system, it will take\ncare of installing the prerequisites for you:<\/p>\n<ul><li>flycheck<\/li>\n<li>memoize<\/li>\n<li>dash\nThese packages can be found on ELPA and\/or MELPA.<\/li>\n<\/ul><p>JDEE also requires <a href=\"http:\/\/github.com\/jdee-emacs\/jdee-server\">JDEE Server<\/a> for some operations.<\/p>\n<h2><a id=\"user-content-installing-with-the-emacs-package-system\" class=\"anchor\" aria-hidden=\"true\" href=\"#installing-with-the-emacs-package-system\"><\/a>Installing with the Emacs package system<\/h2>\n<p>The project is available in MELPA. To install it do the following:<\/p>\n<ol><li>Add MELPA to your <code>.emacs<\/code> or <code>init.el<\/code> if you don't have it:<\/li>\n<\/ol><div class=\"highlight highlight-source-emacs-lisp\"><pre>;;; Add this at the top of the init.el file:\n(require 'package)\n(add-to-list 'package-archives\n             '(\"melpa\" . \"https:\/\/melpa.org\/packages\/\"))\n(when (&lt; emacs-major-version 24)\n  ;; For important compatibility libraries like cl-lib\n  (add-to-list 'package-archives '(\"gnu\" . \"http:\/\/elpa.gnu.org\/packages\/\")))\n(package-initialize)<\/pre><\/div>\n<ol start=\"2\"><li>Install JDEE from Emacs using its package manager:<\/li>\n<\/ol><pre>M-x list-packages\n<\/pre>\n<p>There select JDEE and install: i x<\/p>\n<ol start=\"3\"><li>Install <a href=\"http:\/\/github.com\/jdee-emacs\/jdee-server\">JDEE Server<\/a> from github<\/li>\n<\/ol><p>Follow short readme at <a href=\"http:\/\/github.com\/jdee-emacs\/jdee-server\">JDEE Server<\/a>.<\/p>\n<ol start=\"4\"><li>Customize <code>jdee-server-dir<\/code> to make it point to directory with JDEE Server jars.<\/li>\n<\/ol><p><em>Warning!<\/em> Don't install JDEE Server inside JDEE installed from MELPA (<code>~\/.emacs.d\/elpa\/jdee-xxx<\/code>), because it will be deleted with next update of JDEE! Create a separate directory, for example: <code>~\/.emacs.d\/jdee-server<\/code>.<\/p>\n<p>Customized <code>jdee-server-dir<\/code> in your init file should look something like:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(custom-set-variables\n '(jdee-server-dir \"\/Users\/you\/.emacs.d\/jdee-server\"))<\/pre><\/div>\n<p>For additional information, see the <a href=\"http:\/\/htmlpreview.github.com\/?https:\/\/github.com\/peterwvj\/jdee\/blob\/master\/doc\/flat\/install.html\">old installation instructions<\/a>.<\/p>\n<h2><a id=\"user-content-installing-from-source-for-jdee-devs-only\" class=\"anchor\" aria-hidden=\"true\" href=\"#installing-from-source-for-jdee-devs-only\"><\/a>Installing from source (for JDEE devs only)<\/h2>\n<ol><li>Clone <a href=\"https:\/\/github.com\/jdee-emacs\/jdee\/\">the github repository<\/a>.<\/li>\n<li>Open the cloned JDEE directory in dired.<\/li>\n<li>Being in dired, install using command: <code>package-install-from-buffer<\/code><\/li>\n<\/ol><p>For development you will need to install Cask:<\/p>\n<ol><li>\n<p>Install <a href=\"http:\/\/cask.readthedocs.io\/en\/latest\/index.html\" rel=\"nofollow\">Cask<\/a>.<\/p>\n<\/li>\n<li>\n<p>Run <code>cask install<\/code> from the JDEE directory (make sure that the\n<code>cask<\/code> command is in your PATH).<\/p>\n<\/li>\n<li>\n<p>Run the tests: <code>make test<\/code><\/p>\n<\/li>\n<\/ol><p>To use this built distribution without installation, in your .emacs add:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>  (add-to-list 'load-path \"\/path\/to\/jdee\")\n  (require 'jdee)<\/pre><\/div>\n<h2><a id=\"user-content-building-the-documentation-in-other-formats\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-the-documentation-in-other-formats\"><\/a>Building the documentation in other formats<\/h2>\n<p>JDEE ships with documentation in Info format, but if you want you can generate the docs in other formats too.<\/p>\n<p>To generate Info, HTML and PDF documentation:<\/p>\n<ul><li>\n<p>Install texinfo.<\/p>\n<\/li>\n<li>\n<p>For PDF generation, you also need texi2dvi, texinfo-tex, texlive-ec\nand texlive-cm-super.<\/p>\n<\/li>\n<li>\n<p>In the doc\/ subdirectory, run: <code>makeinfo --info --html --pdf jdee.texi<\/code><\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<p>Some of the features of JDEE include:<\/p>\n<ul><li>source code editing with syntax highlighting, auto indentation using the\nnative GNU Emacs Java mode<\/li>\n<li>Symbol completion (jdee-complete-in-line)<\/li>\n<li>Code generation: templates, import insertion\/deletion\/ordering\netc.<\/li>\n<li>browse JDK doc, using the browser of your choice (jdee-help-symbol and\njdee-help-docsets)<\/li>\n<li>browse your source code (jdee-find-class* and semantic integration)<\/li>\n<li>compilation with automatic jump from error messages to responsible line in the\nsource code using ant (jdee-ant-build).<\/li>\n<li>run Java application in an interactive (comint) Emacs buffer<\/li>\n<li>integrated debugging with interactive debug command buffer and automatic\ndisplay of current source file\/line when stepping through code (jdee-debug)<\/li>\n<li>supports Oracle\/OpenJDK Java 7<\/li>\n<li>runs on any platform supported by GNU Emacs 24.3 and later<\/li>\n<li>easily and infinitely customizable<\/li>\n<\/ul><h3><a id=\"user-content-known-issues\" class=\"anchor\" aria-hidden=\"true\" href=\"#known-issues\"><\/a><a href=\"https:\/\/github.com\/jdee-emacs\/jdee\/issues\">Known Issues<\/a><\/h3>\n<ul><li>Static imports don't quite work<\/li>\n<li>Limited support for Java template beyond basic highlighting, parsing, and\nindenting.<\/li>\n<li>Indentation after multi-line annotations might be unexpected<\/li>\n<\/ul><h3><a id=\"user-content-troubleshooting\" class=\"anchor\" aria-hidden=\"true\" href=\"#troubleshooting\"><\/a>Troubleshooting<\/h3>\n<ul><li>If you notice a bug, open an issue on Github\n<a href=\"https:\/\/github.com\/jdee-emacs\/jdee\/issues\">Issues<\/a><\/li>\n<\/ul><h2><a id=\"user-content-authors\" class=\"anchor\" aria-hidden=\"true\" href=\"#authors\"><\/a>Authors<\/h2>\n<ul><li>Paul Kinnucan (original author and contributor)<\/li>\n<li>Przemys\u0142aw Wojnowski (primary maintainer\/owner)<\/li>\n<li>Paul Landes (maintainer)<\/li>\n<li>Shyamal Prasad (maintainer)<\/li>\n<li>Phil Lord (maintainer)<\/li>\n<\/ul><\/article>","created_at":"2018-11-07T16:40:22+0000","updated_at":"2018-11-07T16:41:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/13318517?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":229,"title":"JDEE User's Guide","url":"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html","content":"<div class=\"titlepage\"><div><p class=\"releaseinfo\">\n      <i>\n\tRevised for JDEE 2.3.4\n      <\/i>\n    <\/p><\/div><div><p class=\"copyright\">Copyright \u00a9 1997-2004 Paul Kinnucan<\/p><\/div><hr \/><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Introduction<\/h2><\/div><p>Welcome to the <i>JDEE User's Guide<\/i>. This\n      guide explains how to use the JDEE to develop Java applications and\n      applets. The guide assumes that you are familiar with Emacs, the\n      Java programming language, and Java development tools provided\n      by Sun Microsystems, Inc.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\tThis guide and the JDEE user interface use the term \"JDK\" to\n\trefer to Sun's Java 2 Software Development Kit (SDK) and its\n\tpredecessors.\n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">About the JDEE<\/h4><\/div><p>The Java Development Environment for Emacs (JDEE) is an\n\tEmacs Lisp package that interfaces Emacs to third-party Java\n\tapplication development tools, such as those provided by Sun\n\tMicrosystems as part of its JDK. The result is an integrated\n\tdevelopment environment (IDE) comparable in power to many\n\tcommercial Java IDEs. Features include: <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>source code editing with syntax highlighting and auto\n\t    indentation<\/p><\/li><li><p>automatic completion of class fields and methods    \n\t  <\/p><\/li><li><p>compilation with automatic jump from error messages\n\t    to responsible line in the source code.<\/p><\/li><li><p>\n\t    generates class and method skeletons automatically\n\t  <\/p><\/li><li><p>run Java application in an interactive (comint)\n\t    Emacs buffer\n\t  <\/p><\/li><li><p>\n\t    integrated debugging with interactive debug command\n\t    buffer and automatic display of current source file\/line\n\t    when stepping through code\n\t  <\/p><\/li><li><p>\n\t    browse JDK doc, using the browser of your choice\n\t  <\/p><\/li><li><p>\n\t    browse your source code, using the Emacs etags facility\n\t    or a tree-structured speedbar.\n\t  <\/p><\/li><li><p>\n\t    supports latest version of JavaSoft's Java Development\n\t  <\/p><\/li><li><p>\n\t    runs on any platform supported by Emacs and Sun's Java SDK (e.g.,\n\t    Win95\/NT and Solaris)\n\t  <\/p><\/li><li><p>\n\t    easily and infinitely customizable\n\t  <\/p><\/li><li><p>\n\t    works with FSF Emacs and XEmacs\n\t  <\/p><\/li><\/ul><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">JDE Requirements<\/h4><\/div><p>The JDE requires the following software: <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><a href=\"http:\/\/www.gnu.org\/\">FSF Emacs<\/a> or\n\t    <a href=\"http:\/\/www.xemacs.org\/\">XEmacs<\/a> on Unix\n\t    platforms; the 20.6.1 version (or later) of  \n\t    <a href=\"ftp:\/\/ftp.gnu.org\/gnu\/windows\/emacs\/latest\">NT\/Emacs<\/a> on\n\t    Windows platforms, for general source file editing and tool integration functions.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/java.sun.com\/j2se\">Java\n\t      Development Kit (JDK)<\/a> or compatible set of Java\n\t    development tools (compiler, virtual machine, debugger,\n\t    class libraries, etc.)\n\t  <\/p><\/li><li><p>\n\t    Web browser (e.g., Netscape or Internet Explorer) for\n\t    viewing documentation.\n\t  <\/p><\/li><li><p>\n\t    latest versions of Eric Ludlam's speedbar, \n\t    semantic, and\n\t    eieio packages available for download from\n\t    <a href=\"http:\/\/cedet.sourceforge.net\">the CDET home page<\/a>\n\t  <\/p><\/li><li><p>\n\t    The elib package available from the \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/index.html\">JDE web site<\/a>\n\t  <\/p><\/li><li><p>\n\t    <code>bash<\/code> or another Unix-style shell highly\n\t    recommended for Windows95\/NT environments. The <a href=\"http:\/\/sources.redhat.com\/cygwin\">cygwin<\/a>\n\t    Unix emulation package for Windows from Red Hat includes\n\t    <code>bash<\/code>.\n\t  <\/p><\/li><\/ul><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">JDEE Components<\/h4><\/div><p>\n\tThe JDEE distribution includes the following files:\n      <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p> \n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde.el\"><code>jde.el<\/code><\/a>. \n\t    Defines jde-mode, a major Emacs mode for developing Java code.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-run.el\"><code>jde-run.el<\/code><\/a>. \n\t    Runs Java applications and applets\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-db.el\"><code>jde-db.el<\/code><\/a>. \n\t    Interfaces Emacs to <code>jdb<\/code>, the command-line debugger that\n\t    comes with the JDK.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-gen.el\"><code>jde-gen.el<\/code><\/a>. \n\t    Contains code generation templates.\n\t  <\/p><\/li><li><p>\n\t    <code>bsh.jar<\/code>. Compiled files for the\n\t    <a href=\"http:\/\/www.beanshell.org\/\">BeanShell<\/a>,\n\t    a Java source code interpreter developed by Pat Neimeyer.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/beanshell.el\"><code>beanshell.el<\/code><\/a>. \n\t    Provides an Emacs interface to the BeanShell interpreter.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-wiz.el\"><code>jde-wiz.el<\/code><\/a>. \n\t    Provides \"wizards\" that generate skeleton\n\t    implementations of interfaces and skeleton overrides of\n\t    methods declared by superclasses.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-complete.el\"><code>jde-complete.el<\/code><\/a>. \n\t    Automatic field and method completion package.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-parse.el\"><code>jde-parse.el<\/code><\/a>. \n\t    Java parser package.<\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/java.bnf\"><code>java.bnf<\/code><\/a>. \n\t    Java grammar used to generate the JDE's lisp-based Java parser.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-bug.el\"><code>jde-bug.el\n\t      <\/code><\/a>. \n\t    JDEbug user interface package.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-dbs.el\"><code>jde-dbs.el<\/code><\/a>. \n\t    JDEbug low-level command interface package.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jde-dbo.el\"><code>jde-dbo.el<\/code><\/a>. \n\t    JDEbug output processing functions.\n\t  <\/p><\/li><li><p>\n\t    source code, jar files, and documentation for the <a href=\"http:\/\/jdee.sunsite.dk\/jdedoc\/java\/\">Java components<\/a> of the\n\t    JDE.<\/p><\/li><li><p><a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jtags\"><code>jtags<\/code><\/a> \n\t    is a bash shell script that tags Java source hierarchies.\n\t  <\/p><\/li><li><p>\n\t    <a href=\"http:\/\/jdee.sunsite.dk\/lisp\/jtags.csh\"><code>jtags.csh<\/code><\/a> \n\t    is a c shell script that tags Java source heierarchies.\n\t  <\/p><\/li><\/ul><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Latest Version<\/h4><\/div><p>See <a href=\"http:\/\/jdee.sourceforge.net\/rootpage.html#Downloading\">\n\t  Downloading the JDE<\/a> on the JDE website for instructions on\n\tdownloading the most recent version of the JDE. \n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Installing the JDE<\/h4><\/div><p>See the <a href=\"http:\/\/jdee.sourceforge.net\/install.html\">\n\tJDE Installation Guide<\/a> on the JDE website for information\n      on installing the JDE on your system.<\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Contacting the Author<\/h4><\/div><p>Please send bug reports and enhancement suggestions to \n\t<a href=\"mailto:pkinnucan@comcast.net\">Paul Kinnucan<\/a>.\n      <\/p><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Registering and Selecting a JDK<\/h2><\/div><p>The JDEE relies on command-line Java development tools, such\n      as those supplied by Sun Microsystems as part of its Java 2\n      Software Development Kit (SDK) also known as a Java Development\n      Kit (JDK). Before you can use the JDEE to compile, debug, or run\n      an application, you must tell it where the Java compiler,\n      debugger, and virtual machine are located on your system. The\n      JDEE assumes that the compiler, debugger, and vm are named\n      <code>javac<\/code>, <code>jdb<\/code>, and\n      <code>java<\/code>, respectively, and that they are\n      located on your system's command path. If these tools are not\n      located on the system command path or have different names, you\n      must tell the JDEE where they are located. The JDEE provides\n      customization variables for specifying the locations of\n      individual tools.\n    <\/p><p>\n      However, if you are using one or more versions of Sun's JDK or a\n      clone of the Sun JDK, the easiest way to specify the location of\n      the external development tools is to register the version of the\n      JDK that you intend to use. It is especially advantageous to\n      register a JDK, if you intend to use multiple versions of the\n      JDK. The reason? Registering each version of the JDK allows you\n      to change JDKs by changing a single customization variable. The\n      following sections explain how to register and select\n      JDKs.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Registering a JDK<\/h4><\/div><p>To register a version of the JDK with the JDEE:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Start Emacs.<\/p><\/li><li><p>Type <code>M-x customize-variable<\/code> and\n\t    press <code>Enter<\/code><\/p><p>Emacs prompts you to enter the name of a variable\n\t  to customize.<\/p><\/li><li><p>Enter <code>jde-jdk-registry<\/code> and\n\t  press <code>Enter<\/code>.<\/p><p>The <code>jde-jdk-registry<\/code> customization\n\t  buffer appears.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/regjdk1.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Click the <code>INS<\/code> button.<\/p><p>The entry for the JDK appears.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/regjdk2.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Enter the version number of the JDK and the path of the directory\n\t  in which the JDK is installed.<\/p><\/li><li><p>Repeat the preceding steps until you have entered the versions and\n\t  paths of all versions of the JDK that you wish to use with the JDEE.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/regjdk3.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Click the <code>State<\/code> button.<\/p><p>The <code>State<\/code> menu appears.<\/p><\/li><li><p>Select <code>Save for Future Sessions<\/code> from\n\t    the <code>State<\/code> menu.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>This saves the setting of <code>jde-jdk-registry<\/code> \n\t      in your <code>.emacs<\/code> file.<\/p><\/div><\/li><li><p>Click the <code>Finish<\/code> button to dismiss\n\t  the customization buffer.<\/p><\/li><\/ol><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Selecting a JDK<\/h4><\/div><p>To select a JDK:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Type <code>M-x customize-variable<\/code> and\n\t    press <code>Enter<\/code><\/p><p>Emacs prompts you to enter the name of a variable\n\t  to customize.<\/p><\/li><li><p>Enter <code>jde-jdk<\/code> and\n\t  press <code>Enter<\/code>.<\/p><p>The <code>jde-jdk<\/code> customization\n\t  buffer appears.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/seljdk1.gif\" alt=\"image\" \/><\/div><\/div><p>The buffer lists the JDKs that you have previously registered \n\t    (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#RegisterJDK\">Registering a JDK<\/a>).<\/p><\/li><li><p>Click the radio button next to the version of the JDK you\n\t  want to use.<\/p><\/li><li><p>Click the <code>State<\/code> button.<\/p><p>The <code>State<\/code> menu appears.<\/p><\/li><li><p>Select <code>Save for Future Sessions<\/code> from\n\t    the <code>State<\/code> menu.<\/p><\/li><li><p>Click the <code>Finish<\/code> button to dismiss\n\t  the customization buffer.<\/p><\/li><\/ol><\/div><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>This procedure saves your JDK selection in your\n\t  <code>.emacs<\/code> file so that it applies to all\n\t  projects. You can configure the JDEE to use different JDKs\n\t  for different projects by saving your customization in a\n\t  project file. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#UsingProjectFiles\">Using\n\t    Project Files<\/a> or more information.<\/p><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Editing Java Source Files<\/h2><\/div><p>\n      To edit an existing Java source file, load it into an Emacs\n      buffer by executing <code>C-x C-f<\/code>. Loading a Java source file\n      into a buffer causes the buffer to enter into <code>jde-mode<\/code>,\n      a variant of <code>java-mode<\/code>, which is in turn a variant of\n      the standard Emacs <code>cc-mode<\/code> for editing source files\n      written in C and related languages. <code>jde-mode<\/code> offers all the\n      specialized source file editing commands of <code>cc-mode<\/code> plus\n      functions for creating, compiling, building, running, and debugging Java source files.\n      You can invoke the commands by selecting them from the <b>JDE<\/b>\n      menu that <code>jde-mode<\/code> displays on the Emacs menu bar or by\n      typing the commands in the Emacs minibuffer.\n    <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/jdemenu.gif\" height=\"350\" width=\"544\" alt=\"Screenshot showing the JDE menu.\" \/><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Documenting Code<\/h2><\/div><p>\n      The JDE provides complete support for generating HTML\n      documentation for Java classes from comments inserted into the\n      source code for those classes.\n    <\/p><p>See also<\/p><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#DisplayingDoc\">Displaying Java\n\tDocumentation<\/a><\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Inserting Javadoc\n\tComments<\/h4><\/div><p>\n\tTo insert a skeleton javadoc comment for a class or method,\n\tposition point in the first line of the method or class and\n\tselect <b>JDE-&gt;Document<\/b> from the\n\tEmacs menubar  or type <code>C-c C-v j<\/code>.\n      <\/p><p>\n\tTo customize the javadoc skeletons, select\n\t<b>JDE-&gt;Options-&gt;Project-&gt;Javadoc<\/b> \n\tfrom the Emacs menubar.\n      <\/p><div class=\"note\"><h3 class=\"title\">Thanks...<\/h3><p>\n\t  to <a href=\"mailto:david@dponce.com\">David Ponce<\/a>\n\t  for developing the JDE's javadoc comment generation\n\t  facility.\n\t<\/p><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Generating the\n\tDocumentation<\/h4><\/div><p>\n\tTo generate documentation for the current project, open any\n\tsource file in the project and select <b>Make\n\t  Doc<\/b> from the <b>JDE<\/b> menu or\n\tenter <code>M-x jde-javadoc-make<\/code>.\n      <\/p><p>\n\tThe <code>jde-javadoc-make<\/code> command runs the JDK's\n\tjavadoc program to generate the documentation. The\n\t<code>javadoc<\/code> command must be on your system's\n\tcommand path.\n      <\/p><p>\n\t  The <code>jde-javadoc-make<\/code> command uses\n\t<code>jde-global-classpath<\/code> as the\n\t<code>-classpath<\/code> and <code>jde-sourcepath\n\t<\/code> as the <code>-sourcepath<\/code> option for\n\tgenerating the doc. You can specify all other javadoc options\n\tvia JDE customization variables. To specify the options,\n\tselect <b>Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>Javadoc<\/b> \n\t\n\t\n\t  from the \n\t  <b>JDE<\/b> menu. Use\n\t  <code>jde-javadoc-gen-packages<\/code> to specify the\n\t  packages, classes, or source files for which you want to\n\t  generate javadoc. If this variable is\n\t  <code>nil<\/code>, the\n\t  <code>jde-javadoc-make<\/code> generates javadoc for\n\t  the Java source file in the current buffer.\n      <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\t  Thanks to <a href=\"mailto:sakliban@cs.wustl.edu\">Sergey A\n\t    Klibanov <\/a> for developing the JDE's javadoc\n\t  generation facility.\n\t<\/p><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Abbreviations<\/h2><\/div><p>\n      The JDEE allows you to use abbreviations for Java keywords and\n      control flow constructs. The following sections explain how to\n      use these abbreviations.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Keyword\n\tAbbreviations<\/h4><\/div><p>The JDEE defines a set of abbreviations for Java keywords.\n\tWhen you type one of these abbreviations followed by a space\n\tin a Java source buffer, the JDE optionally expands the\n\tabbreviation into the keyword. For example, when the\n\tabbreviation mode is enabled, you need only type\n\t<code>fa<\/code> followed by a space to enter the Java\n\tkeyword <code>false<\/code>. To enable or disable\n\tabbreviation mode by default for a project, select\n\t<b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Options\n\t  <\/b>-&gt;<b>General<\/b>\n\tand toggle the boolean variable\n\t<code>jde-enable-abbrev-mode<\/code>. To toggle\n\tabbreviation mode on or off during a session, select\n\t<b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Modes<\/b>-&gt;<b>Abbrev<\/b>. To change, remove, or add an abbreviation, edit\n\tthe variable <code>jde-mode-abbreviations<\/code> in the\n\tproject customization buffer.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>The JDE's Java keyword expansion facility is based\n\t  on the Emacs <code>abbrev-mode<\/code> facility. For\n\t  more information, see the Emacs user manual.<\/p><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Control Flow\n\tAbbreviations<\/h4><\/div><p>The JDEE defines abbreviations for Java control flow\n\tstructures, such as if-then-else. When you enter the\n\tabbreviation followed by a space, the JDE expands the\n\tabbreviation into a corresponding control structure\n\ttemplate.<\/p><div class=\"example\"><p><b>Example 1. Expanding a Control Flow Abbreviation<\/b><\/p><p>The if-then-else abbreviation<\/p><pre class=\"programlisting\">\n\t  ife\n\t<\/pre><p>expands to the control flow structure<\/p><pre class=\"programlisting\">\n\t  if ( ) { } \/\/ end of if () else } \/\/ end of if ()else\n\t<\/pre><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Control Flow Abbreviations<\/h5><\/div><p>The JDE defines the following abbreviations for control\n\t  structures.<\/p><div class=\"table\"><p><b>Table 1. Control Flow Abbreviations<\/b><\/p><table summary=\"Control Flow Abbreviations\" border=\"1\"><colgroup><col \/><col \/><\/colgroup><thead><tr><th>Statement<\/th><th>Abbreviation<\/th><\/tr><\/thead><tbody><tr><td><code>if-then<\/code><\/td><td><code>if<\/code><\/td><\/tr><tr><td><code>else<\/code><\/td><td><code>else<\/code><\/td><\/tr><tr><td><code>if-then-else<\/code><\/td><td><code>ife<\/code><\/td><\/tr><tr><td><code>else-if<\/code><\/td><td><code>eif<\/code><\/td><\/tr><tr><td><code>while<\/code><\/td><td><code>while<\/code><\/td><\/tr><tr><td><code>for<\/code><\/td><td><code>for<\/code><\/td><\/tr><tr><td><code>for (int\n\t\t    I=0;I&lt;UL;I++)<\/code><\/td><td><code>fori<\/code><\/td><\/tr><tr><td><code>for (Iterator i = c.iterator();\n\t\t    i.hasNext();)<\/code><\/td><td><code>foriter<\/code><\/td><\/tr><tr><td><code>main method<\/code><\/td><td><code>main<\/code><\/td><\/tr><tr><td><code>switch<\/code><\/td><td><code>switch<\/code><\/td><\/tr><tr><td><code>case<\/code><\/td><td><code>case<\/code><\/td><\/tr><tr><td><code>try<\/code><\/td><td><code>try<\/code><\/td><\/tr><tr><td><code>catch<\/code><\/td><td><code>catch<\/code><\/td><\/tr><tr><td><code>try finally<\/code><\/td><td><code>tryf<\/code><\/td><\/tr><tr><td><code>finally<\/code><\/td><td><code>finally<\/code><\/td><\/tr><\/tbody><\/table><\/div><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\t    You must enable the JDE's Java keyword abbreviation mode\n\t    to use the control flow abbreviations. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#JavaKeywordCompletion\">Java Keyword\n\t      Completion<\/a> for more information.\n\t  <\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Left Brace Placement<\/h5><\/div><p>\n\t  The JDE's Java control flow templates support two options\n\t  for opening brace placement: placement on the first line of\n\t  the template (Kerningham &amp; Ritchie Style, the default)\n\t  and placement on a separate line. Type <code>M-x\n\t    customize-variable jde-gen-k&amp;r<\/code> to specify\n\t  the style you prefer.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Customizing the Control Flow Templates<\/h5><\/div><p>You can customize the templates to suit any indentation\n\t  style. To customize the templates, select \n\t    <b>Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>Autocode<\/b> from the JDE menu.<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Adding Your Own Control Flow Templates<\/h5><\/div><p>\n\t  Use the <code>jde-gen-define-abbrev-template<\/code>\n\t  function to define your own control flow abbreviation in\n\t  your  <code>.emacs<\/code> file. For example, the\n\t  following code defines an abbreviation  for an else clause.\n\t<\/p><pre class=\"programlisting\">\n\t  (jde-gen-define-abbrev-template \"melse\" '('&gt; \"else {\" '&gt; 'n\n\t  '&gt; 'r 'n \"}\"))\n\t<\/pre><p>The <code>jde-gen-define-abbrev-template<\/code>\n\t  function takes two arguments. The first argument is a string\n\t  that specifies the abbreviation. The section argument is a\n\t  list defining the text that replaces the abbreviation. The\n\t  list can contain any of the following elements.<\/p><div class=\"table\"><p><b>Table 2. Template Elements<\/b><\/p><table summary=\"Template Elements\" border=\"1\"><colgroup><col \/><col \/><\/colgroup><thead><tr><th>Element<\/th><th>Usage<\/th><\/tr><\/thead><tbody><tr><td>A string.<\/td><td>The string is inserted into the buffer at\n\t\t  point.<\/td><\/tr><tr><td>The symbol <code>'p<\/code>.<\/td><td>This position is saved in\n\t\t  <code>tempo-marks<\/code>.<\/td><\/tr><tr><td>The symbol <code>'r<\/code>.<\/td><td>If <code>tempo-insert<\/code> is called\n\t\t  with <code>ON-REGION<\/code> non-nil the\n\t\t  current region is placed here. Otherwise it works\n\t\t  like 'p.<\/td><\/tr><tr><td><code>(p PROMPT &lt;NAME&gt;\n\t\t    &lt;NOINSERT&gt;)<\/code><\/td><td>If <code>tempo-interactive<\/code> is\n\t\t  non-nil, the user is prompted in the minbuffer with\n\t\t  <code>PROMPT<\/code> for a string to be\n\t\t  inserted. If the optional parameter\n\t\t  <code>NAME<\/code> is non-nil, the text is\n\t\t  saved for later insertion with the\n\t\t  <code>s<\/code> tag. If there already is\n\t\t  something saved under <code>NAME<\/code> that\n\t\t  value is used instead and no prompting is made. If\n\t\t  <code>NOINSERT<\/code> is provided and non-nil,\n\t\t  nothing is inserted, but text is still saved when a\n\t\t  <code>NAME<\/code> is provided. For clarity,\n\t\t  the symbol <code>'noinsert<\/code> should be\n\t\t  used as argument.<\/td><\/tr><tr><td><code>(P PROMPT &lt;NAME&gt;\n\t\t    &lt;NOINSERT&gt;)<\/code><\/td><td>Works just like the previous element, but\n\t\t  forces <code>tempo-interactive<\/code> to be\n\t\t  true.<\/td><\/tr><tr><td><code>(r PROMPT &lt;NAME&gt;\n\t\t    &lt;NOINSERT&gt;)<\/code><\/td><td>Like the previous, but if\n\t\t  <code>tempo-interactive<\/code> is\n\t\t  <code>nil<\/code> and\n\t\t  <code>tempo-insert<\/code> is called with\n\t\t  <code>ON-REGION<\/code> non-nil, the current\n\t\t  region is placed here. This usually happens when you\n\t\t  call the template function with a prefix\n\t\t  argument.<\/td><\/tr><tr><td><code>(s NAME)<\/code><\/td><td>Inserts text previously read with the\n\t\t  <code>(p ..)<\/code> construct. Finds the\n\t\t  insertion saved under <code>NAME<\/code> and\n\t\t  inserts it. Acts like <code>'p<\/code> if\n\t\t  <code>tempo-interactive<\/code> is\n\t\t  <code>nil<\/code>.<\/td><\/tr><tr><td><code>'&amp;<\/code><\/td><td>If there is only whitespace between the line\n\t\t  start and point, nothing happens. Otherwise a\n\t\t  newline is inserted.<\/td><\/tr><tr><td><code>'%<\/code><\/td><td>If there is only whitespace between point and\n\t\t  end-of-line nothing happens. Otherwise a newline is\n\t\t  inserted.<\/td><\/tr><tr><td><code>'n<\/code><\/td><td>Inserts a newline.<\/td><\/tr><tr><td><code>'&gt;<\/code><\/td><td>The line is indented using\n\t\t  <code>indent-according-to-mode<\/code>. Note\n\t\t  that you often should place this item after the text\n\t\t  you want on the line.<\/td><\/tr><tr><td><code>'n&gt;<\/code><\/td><td>Inserts a newline and indents line.<\/td><\/tr><tr><td><code>'r&gt;<\/code><\/td><td>Like <code>r<\/code>, but it also indents\n\t\t  the region.<\/td><\/tr><tr><td><code>'o<\/code><\/td><td>Like <code>'%<\/code> but leaves the point\n\t\t  before the new line.<\/td><\/tr><tr><td>Anything else.<\/td><td>It is evaluated and the result is treated as an\n\t\t  element to be inserted. One additional tag is useful\n\t\t  for these cases. If an expression returns a list\n\t\t  <code>'(l foo bar)<\/code>, the elements after\n\t\t  <code>'l<\/code> will be inserted according to\n\t\t  the usual rules. This makes it possible to return\n\t\t  several elements from one expression.<\/td><\/tr><\/tbody><\/table><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Enabling Variable Content<\/h5><\/div><p>Some templates optionally prompt you for items to insert\n\t  into the template. To enable  prompting, set the variable\n\t  <code>tempo-interactive<\/code> to a\n\t  non-<code>nil<\/code> value in your\n\t  <code>.emacs<\/code> file.<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Disabling the Control Flow Abbreviations<\/h5><\/div><p>To disable the control flow abbreviations, set the\n\t  variable <code>jde-gen-cflow-enable<\/code> off.<\/p><div class=\"note\"><h3 class=\"title\">Thanks to...<\/h3><p>Thanks to Eric D. Friedman &gt;friedman@lmi.net&lt;\n\t    for contributing the control flow templates.<\/p><\/div><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Completing Expressions<\/h2><\/div><p>\n\tBoth Emacs and the JDEE provide commands for completing incomplete\n\texpressions. Emacs provides commands that work for any expression\n\tbut only if the completions exist in an open buffer. The JDEE\n\tprovides completion commands that work for any Java method or field\n\tname that exists on <code>jde-global-classpath<\/code>.\n\tThe following sections explain how to use completion facilities\n\tsupplied by the JDEE and by Emacs itself.\n\t\n      <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Completing Method\n\tand Field Names<\/h4><\/div><p>\n\tThe JDEE provides comamnds that can complete an incomplete\n\tfield or method name. All the methods determine the set of\n\tpossible completions for an incomplete field or method name at\n\tpoint. They differ in how they present the completions to you\n\tfor selection. The commands include:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>jde-complete-in-line<\/code><\/p><p>This command uses the first completion that it finds\n\t    to complete the method or field name at point. If multiple\n\t    completions exist, it allows you to cycle through the\n\t    completions, each completion replacing the previous in the\n\t    source buffer.  See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#InLineCompletion\">Using\n\t      In-Line Completion<\/a> for more information.\n\t  <\/p><\/li><li><p><code>jde-complete-menu<\/code><\/p><p>\n\t    This command displays a menu of possible completions for\n\t    the incomplete method or field name at point. It enters\n\t    the completion that you select in the source buffer. See\n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#MenuCompletion\">Using Menu-Based\n\t      Completion<\/a> for more information.\n\t  <\/p><\/li><li><p><code>jde-complete-minibuf<\/code><\/p><p>\n\t    This command uses the minibuffer to display possible\n\t    completions for the incomplete method or field name at\n\t    point. The comamnd enters the completion that you select\n\t    in the source buffer. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#MinibufferCompletion\">Using Minibuffer-Based\n\t      Completion<\/a> for more information.<\/p><\/li><li><p><code>jde-complete<\/code><\/p><p>This is the JDEE's user-definable \"standard\"\n\t    completion command. It delegates completion to one of the\n\t    JDEE's other completion commands\n\t    (<code>jde-complete-menu<\/code> by default) or to a\n\t    custom completion function that you specify (see <a href=\"http:\/\/jdee.sunsite.dk\/jdedoc\/html\/jde-ug\/SelectCompletionMethod\">Selecting a Completion\n\t      Method<\/a> for more information.) The\n\t    <code>jde-complete<\/code> command is bound by\n\t    default to the key combination <code>C-c C-v\n\t      C-.<\/code>. If you prefer another binding, use the\n\t    customization variable <code>jde-key-bindings<\/code>\n\t    to change the binding.\n\t  <\/p><\/li><\/ul><\/div><p>\n\tThe JDEE's field and method completion commands use the\n\tBeanshell to run Java code that in turn uses Java's reflection\n\t(class introspection) capability to determine the fields and\n\tmethods defined for the class of object at point. The commands\n\tstart the Beanshell if it is not running. This can cause a\n\tnoticeable delay in completion the first time it is used in a\n\tsession. The response can also be slow for classes containing\n\tmany methods and fields.\n      <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#SpeedingUpCompletion\">Speeding Up Completion<\/a>\n\tfor some tips on how you can dramatically speed up field and\n\tmethod completion.<\/p><\/div><p>\n\tCompletion works only for compiled classes that reside in the\n\tclasspath defined by <code>jde-global-classpath<\/code>,\n\tif set, otherwise by the <code>CLASSPATH<\/code>\n\tenvironment variable when the Beanshell starts. Thus, if the\n\tJDEE is unable to complete a method or field, make sure that\n\tthe class that defines the field or method is compiled and\n\texists on the classpath defined by\n\t<code>jde-global-classpath<\/code>, if set, otherwise by\n\tthe <code>CLASSPATH<\/code> environment variable.\n      <\/p><p>\n\tThe JDEE completion commands work for all of the following\n\tcases.\n      <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>objects referenced by variables declared in the\n\t    current buffer<\/p><\/li><li><p>static fields and methods<\/p><\/li><li><p>\n\t    fields and methods of objects defined by the current class\n\t    and its parent\n\t  <\/p><p>\n\t    Names of fields and methods of the current class must\n\t    start with <code>this<\/code> or <code>.<\/code>\n\t    (period). Names of fields and methods declared by the\n\t    parent of the current class must start with\n\t    <code>super<\/code>.\n\t  <\/p><p>For example, this command completes<\/p><pre class=\"programlisting\">\n\t    . ^ this. ^ super. ^\n\t  <\/pre><\/li><li><p>\n\t    objects referenced by the fields of objects referenced by\n\t    variables declared in the current class or its parent or\n\t    by static fields\n\t  <\/p><p>For example, this command completes<\/p><pre class=\"programlisting\">\n\t    System.out.prin ^\n\t  <\/pre><\/li><li><p>\n\t    objects returned by methods of objects referenced by\n\t    variables defined in the current \n\t  <\/p><p>\n\t    For example, this command completes\n\t  <\/p><pre class=\"programlisting\">\n\t    Toolkit.getDefaultToolkit().get ^\n\t  <\/pre><\/li><li><p>\n\t    objects referenced by method parameters<\/p><p>\n\t    For example, this command completes\n\t  <\/p><pre class=\"programlisting\">\n\t    void setColor(String color) { color.get ^ }\n\t  <\/pre><\/li><\/ul><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> Selecting a\n\t  Completion Method<\/h5><\/div><p>The JDEE's generic completion command,\n\t  <code>jde-complete<\/code> delegates completion to one\n\t  of several specific completion commands supplied by the JDEE\n\t  or to a custom completion method that you supply. This\n\t  allows you to customize <code>jde-complete<\/code> to\n\t  use the completion method that you prefer.<\/p><p>To specify your preferred completion method:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Open a customization buffer for\n\t      <code>jde-complete-function<\/code>.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/completion_sel_method.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Select the completion command that implements the\n\t      completion method you prefer.<\/p><\/li><li><p>If you selected <code>Custom<\/code> as your\n\t      preferred completion function, enter the name of the\n\t      custom completion function in the adjacent edit field,\n\t      replacing the existing contents\n\t      (<code>ignore<\/code> by default.).\n\t    <\/p><\/li><li><p>Save your selection in your\n\t      <code>.emacs<\/code> or project file.\n\t    <\/p><\/li><\/ol><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using Menu-Based\n\t  Completion<\/h5><\/div><p>The <code>jde-complete-menu<\/code> displays a\n\t  popup menu of completions for the method or field name at\n\t  point. Selecting a completion from the menu causes the\n\t  command to use it to complete the method or field at point.\n\t  This section assumes that you have selected\n\t  <code>jde-complete-menu<\/code> as your standard\n\t  completion command.\n\t<\/p><p>To select completions from a menu:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Position point immediately following the partially\n\t      complete field or method name.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/completion_menu1.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Type the <code>C-c C-v C-.<\/code> key combination.<\/p><p>A menu of completions pops up.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/completion_menu2.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Select the desired completion from the menu.<\/p><p>The JDEE inserts the completion at point in the\n\t      buffer.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/completion_menu3.gif\" alt=\"image\" \/><\/div><\/div><\/li><\/ol><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using In-Line\n\t  Completion<\/h5><\/div><p>\n\t  The <code>jde-complete-in-line<\/code> command allows\n\t  you to select completions at the point of insertion in the\n\t  source buffer. The command finds all the fields and methods\n\t  that complete the name at point. It then inserts the first\n\t  potential completion in the buffer at point. Repeatedly\n\t  executing the command causes  the JDEE to cycle through the\n\t  other potential completions. If the completion is a method\n\t  name, the command completes the method name and displays the\n\t  method signature in the minibuffer.\n\t<\/p><p>\n\t  The <code>jde-complete-in-line<\/code> command is bound\n\t  to <code>C-c C-v .<\/code>. You can use this\n\t  combination to invoke the command or <code>C-c C-v\n\t    C-.<\/code>, if you have selected this command as your\n\t  preferred completion method. The following examples use the\n\t  <code>C-c C-v .<\/code> combination.\t    \n\t<\/p><div class=\"example\"><p><b>Example 1. Completing a String Method Name<\/b><\/p><p>\n\t      Typing\n\t    C-c-C-v-.\n\t  <\/p><pre class=\"programlisting\">\n\t    String s; s.get ^\n\t  <\/pre><p>\n\t    completes the method name at point as follows<\/p><pre class=\"programlisting\">\n\t    String s; s.getClass( ^\n\t  <\/pre><p>and displays<\/p><pre class=\"programlisting\">\n\t    java.lang.Class getClass()\n\t  <\/pre><p>in the minibuffer. Repeatedly typing\n\t    C-c-C-v-. \n\t    cycles through all the other get methods for the Java\n\t    <code>String<\/code> class.<\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Speeding Up Completion<\/h5><\/div><p>Here are two ways you can significantly speed up field\n\t  and method completion:\n\t<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>Byte-compile the eieio and JDEE packages.<\/p><p>The beanshell package, the Emacs interface to the\n\t      BeanShell, is based on the eieio object-oriented Lisp\n\t      package for Emacs. Completion uses the BeanShell heavily\n\t      to determine the class of the field or method at point.\n\t      Thus compiling eieio and beanshell.el improves\n\t      completion's performance, dramatically in my tests.\n\t    <\/p><\/li><li><p>Import by class, not by package, in your Java files.\n\t    <\/p><p>Every package import, e.g.,\n\t      <code>java.util.*<\/code>, requires invoking the\n\t      BeanShell to determine the fully qualified name of the\n\t      class at point (see\n\t      <code>jde-parse-get-qualified-name<\/code>).\n\t      A Beanshell invocation is by far the most time-consuming\n\t      operation required for completion. Thus eliminating the\n\t      use of package-level imports can significantly\n\t      speed up completion.\n\t    <\/p><\/li><\/ul><\/div><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Dynamic Keyword Completion<\/h4><\/div><p>Emacs provides dynamic completion commands that attempt\n\t  to find completions for the word at point in the current\n\t  buffer or other buffers.\n\t<\/p><div class=\"example\"><p><b>Example 2. Using Dynamic Completion<\/b><\/p><p>\n\t    Suppose that the current source buffer contains the class\n\t    names <code>Component<\/code> and\n\t    <code>Container<\/code>. Now suppose you enter\n\t    <code>Co<\/code> somewhere in the buffer and type\n\t    <b>M<\/b>-<b>\/<\/b>.\n\t    \n\t    The text <code>Component<\/code> replaces\n\t    <code>Co<\/code> in the buffer. \n\t  <\/p><p>Typing\n\t    <b>M<\/b>-<b>\/<\/b>\n\t    again changes <code>Component<\/code> to\n\t    <code>Container<\/code>. In this way, you can cycle\n\t    through all the possible completions for\n\t    <code>Co<\/code> in the current buffer.\n\t  <\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Dynamic Completion Commands<\/h5><\/div><p>Emacs provides two dynamic completion commands.<\/p><p>\n\t    The <code>dabbrev-expand<\/code> command is bound to\n\t  the key combination\n\t  <b>M<\/b>-<b>\/<\/b>\n\t    by default. It searches the current buffer for completions\n\t    of the word at point and, if none are found, other buffers\n\t    of the same type. For example, if the current buffer is a\n\t    Java source buffer, it searches other Java source buffers\n\t    for completions if none are found in the active buffer.\n\t<\/p><p>\n\t  The <code>hippie-expand<\/code> command provides more\n\t  extensive search capabilities. See the docstrings for these\n\t  functions for more information.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Electric Return<\/h5><\/div><p>In electric return mode, pressing the\n\t  <b>Enter<\/b> key causes the JDEE to close open\n\t  braces at the end of a line. To enable or disable this \n\t  mode by default, customize <code>jde-electric-return-p<\/code>.\n\t  To turn the mode on or off during  a session, select <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Modes<\/b>-&gt;<b>Electric Return<\/b>.<\/p><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Generating Code<\/h2><\/div><p>The JDE provides the following code generation capabilities.<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>Code wizards<\/p><\/li><li><p>Code templates<\/p><\/li><\/ul><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Generating Import Statements<\/h4><\/div><p>The JDEE provides a set of commands that generate and\n      organize import statements in the current buffer. The commands\n      automate the following tasks:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>Generate an import statement for the class at point (see\n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ImportingClasses\">Importing Classes<\/a>).<\/p><\/li><li><p>Generate import statements for all classes in the current\n\t    buffer that need import statements (see \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ImportingAllClasses\">Importing All Classes<\/a>).<\/p><\/li><li><p>Expand a package import statement into specific class imports (see \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ExpandingPackageImports\">Expanding Package Imports<\/a>).<\/p><\/li><li><p>Collapse class imports into a package import statement (see \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#CollapsingClassImports\">Collapsing Class Imports<\/a>).<\/p><\/li><li><p>Group import statements (see \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#GroupingImports\">Grouping Imports<\/a>).<\/p><\/li><li><p>Sort import statements alphabetically (see \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#SortingImports\">Sorting Imports<\/a>).<\/p><\/li><li><p>Delete superfluous import statements (see \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#DeletingImports\">Deleting Imports<\/a>).<\/p><\/li><\/ul><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Importing Classes<\/h5><\/div><p>\n\t  The <b>JDE<\/b>-&gt;<b>Import<\/b>-&gt;<b>Class<\/b>\u00a0\n\t  (<code>jde-import-find-and-import<\/code>,\n\t  <b>C<\/b>-<b>c<\/b>\u00a0\n\t  <b>C<\/b>-<b>v<\/b>\u00a0\n\t  <b>C<\/b>-<b>z<\/b>)\u00a0command\n\t  generates an import statement for the\n\t  class at point in the current buffer. It inserts the import\n\t  statement at the head of the buffer.\n\t<\/p><p>This command uses the BeanShell to search the classpath\n\t  of the current project for classes that match the name of\n\t  the class at point. The class name may be unqualified.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\t    The classpath that the import wizard searches for import\n\t    candidates is the classpath specified by the value of\n\t    <code>jde-global-classpath<\/code> for the current\n\t    project.\n\t  <\/p><\/div><p>If the command finds more than one class of the same\n\t  unqualified name on the current classpath, it prompts you to\n\t  select one of the classes to import.<\/p><p>\n\t  The customization variable\n\t  <code>jde-import-excluded-classes<\/code> allows\n\t  you to specify rules for excluding classes from\n\t  consideration for import into the current source file. If\n\t  the import command finds any classes on the classpath that\n\t  belong to the list of prohibited classes, it removes them\n\t  from the list of classes considered for import into the\n\t  current buffer.\n\t<\/p><p>You can use either regular expressions or Lisp functions to\n\t  specify class exclusion rules. For each rule, you can also\n\t  specify that all classes whose unqualified names match the\n\t  rule be excluded from consideration for import. This is\n\t  useful for preventing the JDEE from importing alternate\n\t  implementations of standard Java classes that do not need to\n\t  be imported.<\/p><p>For example, the default setting for\n\t  <code>jde-import-excluded-classes<\/code> includes a\n\t  rule to exclude all classes belonging to the\n\t  <code>java.lang<\/code> package because every Java\n\t  class implicitly imports this package. The default setting\n\t  for <code>jde-import-excluded-classes<\/code> also\n\t  excludes all unqualified synonyms of classes belonging to\n\t  the <code>java.lang<\/code> package. This prevents\n\t  importation of alternate implementations of these classes,\n\t  e.g., alternate implementations of\n\t  <code>java.lang.String<\/code>, included in some\n\t  commonly used class libraries and intended for internal use.\n\t  See the docstring for\n\t  <code>jde-import-excluded-classes<\/code> for more\n\t  information.<\/p><div class=\"note\"><h3 class=\"title\">Thanks to ...<\/h3><p><a href=\"mailto:len@reeltwo.com\">Len\n\t      Trigg<\/a> for contributing the initial\n\t    implementation of the import wizard and to\n\t    Martin Schwarmberger for significantly\n\t    enhancing the import wizard.\n\t  <\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Importing All Classes<\/h5><\/div><p>The <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Import<\/b>-&gt;<b>All<\/b> (<code>jde-import-all<\/code>, \n\t  <b>C<\/b>-<b>c <\/b>\n\t  <b>C<\/b>-<b>v <\/b>\n\t  <b>z<\/b>\n\t  ) imports all the classes that need to be\n\t  imported into the current buffer. \n\t  <\/p><p>\n\t  The command imports all classes that meet the following\n\t  conditions:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>The class's base name begins with an uppercase letter.<\/p><\/li><li><p>The class's base name contains at least one noninitial\n\t    lowercase letter.<\/p><\/li><li><p>The class is not already imported into the buffer.<\/p><\/li><li><p>The class is not in \n\t      <code>jde-import-excluded-classes<\/code>.<\/p><\/li><li><p>The class exists on the current classpath.<\/p><\/li><\/ul><\/div><p>If more than one import candidate with the same base\n\t  name exists on the classpath, the JDEE displays a dialog\n\t  that allows you to select one of the candidates.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>You can use the <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Import<\/b>-&gt;<b>Import All Unique<\/b> (<code>jde-import-all-unique<\/code>\n\t  ) command to avoid the dialog box. This command imports all classes\n\t  whose unqualified names appear only once on the classpath.<\/p><\/div><div class=\"note\"><h3 class=\"title\">Thanks to ...<\/h3><p><a href=\"mailto:p.lord@russet.org.uk\">Phillip\n\t      Lord<\/a> for contributing the initial implementation\n\t    of the Import All Wizard.<\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Expanding Package Imports<\/h5><\/div><p>To expand a package import statement, select <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Import<\/b>-&gt;<b>Expand Package Imports<\/b>\n\t  (<code>jde-import-expand-imports<\/code>). This command replaces \n\t  a package import statement, e.g., \n\t<\/p><pre class=\"programlisting\">\n  import java.io.*;\n\t  <\/pre><p>with import statements for the members of that package referenced\n\tby the current buffer, e.g.,<\/p><pre class=\"programlisting\">\n  import java.io.BufferedReader;\n  import java.io.InputStreamReader;\n  import java.io.IOException;\n\t  <\/pre><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Collapsing Class Imports<\/h5><\/div><p>To collapse class import statements in  the current buffer into \n\t  package imports, select <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Import<\/b>-&gt;<b>Collapse Class Imports<\/b> (<code>jde-import-collapse-imports<\/code>). This command\n\t  collapses imports for any package that contains as many as or more than the\n\t  number of imports specified by <code>jde-import-collapse-imports-threshhold<\/code>.\n\t  The default value of this variable is 2. This causes this command to collapse\n\t  imports for any package from which the current buffer imports two or more classes.\n\t  For example, this command, by default, collapses\n\t<\/p><pre class=\"programlisting\">\n  import java.io.BufferedReader;\n  import java.io.InputStreamReader;\n  import java.io.IOException;\n<\/pre><p>to<\/p><pre class=\"programlisting\">\n  import java.io.*;\n<\/pre><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Grouping Imports<\/h5><\/div><p>To organize import statements into groups of related imports, select\n\t  <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Import<\/b>-&gt;<b>Organize Imports<\/b> (<code>jde-import-organize<\/code>). By default this command\n\t  groups import statements into two groups separated by a blank line. The first\n\t  group contains all imports from the java and javax packages. The other group\n\t  contains imports from all other packages. The imports are sorted alphabetically\n\t  in ascending order in each group, e.g., \n\t<\/p><pre class=\"programlisting\">\n  import java.io.InputStreamReader;\n  import java.util.Hashtable;\n  import javax.swing.JFrame;\n  import javax.swing.JPanel;\n  import jmath.LinearSystem;\n  import jmath.Test;\n<\/pre><p>The following variables allow you to customize the organization of imports\n\tinto groups:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>jde-import-group-function<\/code><\/p><p>This variable allows you to specify a function that assigns imports\n\t      to groups. The default group function is <code>jde-import-group-of<\/code>,\n\t    which uses rules specified by the following variable to assign imports to \n\t    groups.<\/p><\/li><li><p><code>jde-import-group-rules<\/code><\/p><p>This variable specifies a list of rules for assigning import statements to\n\t      group, e.g., <code>'((\"^javax\\\\.\" . \"Swing\"))<\/code>,\n\t      Each rule consists of two parts: a regular expression and an optional\n\t      group name specifier. The <code>jde-import-group-of<\/code> function assigns\n\t    an import statement to the group if it matches the regular expression. The group name\n\t    specifier is either a string that specifies the name or an index to a component of the regular\n\t    expression to be used as the group name.<\/p><\/li><li><p><code>jde-import-insert-group-names<\/code><\/p><p>Specifies whether to insert an import group's name in the buffer.<\/p><\/li><li><p><code>jde-import-default-group-name<\/code><\/p><p>Specifies the default name for an import group.<\/p><\/li><li><p><code>jde-import-blank-line-between-groups<\/code><\/p><p>Specifies whether to insert a blank line between import groups.<\/p><\/li><li><p><code>jde-import-sorted-groups<\/code><\/p><p>Specifies whether and how to sort groups of packages. Options include\n\t    no sorting, in the order specified by grouping rules, or in ascending or \n\t    descending alphabetical order.<\/p><\/li><\/ul><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Deleting Unneeded Imports<\/h5><\/div><p>\n\t  To delete superfluous import statements from the current buffer, select\n\t  <b>JDE<\/b>-&gt;<b>Code Generation<\/b>-&gt;<b>Import<\/b>-&gt;<b>Delete Unneeded<\/b>\n\t  (<code>jde-import-kill-extra-imports<\/code>). This command deletes all\n\t  import statements for classes that are not actually referenced by the class\n\t  in the current buffer.\n\t<\/p><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Code Wizards<\/h4><\/div><p>\n\tThe JDE provides a set of procedural code generators called\n\tcode wizards. They include:\n      <\/p><div class=\"table\"><p><b>Table 1. Code Wizards<\/b><\/p><table summary=\"Code Wizards\" border=\"1\"><colgroup><col \/><col \/><\/colgroup><thead><tr><th><p>Wizard<\/p><\/th><th><p>Generates...<\/p><\/th><\/tr><\/thead><tbody><tr><td><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#MethodOverrideWizard\">Method\n\t\t    Override Wizard<\/a><\/p><\/td><td><p> A method that overrides a method inherited by\n\t\t  the class containing point. <\/p><\/td><\/tr><tr><td><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#InterfaceWizard\">Interface\n\t\t    Wizard<\/a><\/p><\/td><td><p>Implementation of an interface in the class\n\t\t  containing point. <\/p><\/td><\/tr><tr><td><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#DelegateWizard\">Delegate\n\t\t    Wizard<\/a><\/p><\/td><td><p> Methods that delegate tasks to a specified\n\t\t  class. <\/p><\/td><\/tr><tr><td><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#AbstractClassWizard\">Abstract Class\n\t\t    Wizard<\/a><\/p><\/td><td><p>Implementations of the abstract methods\n\t\t  inherited by the class containing point. <\/p><\/td><\/tr><tr><td><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#GetSetWizard\">Get\/Set\n\t\t    Wizard<\/a><\/p><\/td><td><p>Get and set methods for the private fields of\n\t\t  the class containing point. <\/p><\/td><\/tr><\/tbody><\/table><\/div><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\t  Some of the wizards use the BeanShell to run Java code. A\n\t  wizard starts the Beanshell interpreter if it is not already\n\t  running. Thus, you may experience a slight delay when\n\t  invoking a wizard for the first time in a session.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Method Override\n\t  Wizard<\/h5><\/div><p>\n\t  The method override wizard generates a skeleton method that\n\t  overrides a similarly named method defined by a superclass.\n\t<\/p><p>To override a method of a superclass: <\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>\n\t      Position the Emacs point at the location in the buffer\n\t      where you want the generated method to appear.\n\t    <\/p><p>\n\t      The point must be within the class that is overriding\n\t      the   method. \n\t    <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/OverrideMeth1.gif\" height=\"328\" width=\"472\" alt=\"Screen capture showing point inside a&#10;&#9;&#9;      class.\" \/><\/div><\/div><\/li><li><p>\n\t      Select\n\t      <b>Wizards<\/b>-&gt;<b>Override \n\t\t  Method<\/b> from the\n\t      <b>JDE<\/b> menu or enter\n\t      <b>M<\/b>-<b>x<\/b> <code>jde-wiz-overrided-method<\/code>. \n\t    <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/OverrideMeth2.gif\" alt=\"Screen shot showing&#10;&#9;&#9;      Wizards menu.\" \/><\/div><\/div><p>\n\t      The JDE prompts you to enter the name of the method to\n\t      be overridden in the minibuffer. \n\t    <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/OverrideMeth3.gif\" alt=\"Screen shot showing prompt for&#10;&#9;&#9;      method name.\" \/><\/div><\/div><p>\n\t      The name must be the name of a method defined by an\n\t      ancestor of the class in which the Emacs point is\n\t      located. The compiled class of the ancestor must be on\n\t      the classpath specified by\n\t      <code>jde-global-classpath<\/code>. \n\t    <\/p><\/li><li><p>\n\t      Enter the name of the method to be overridden. \n\t    <\/p><p>\n\t      If the  ancestors of the class in which you are\n\t      overriding the method define more than one method of the\n\t      same name, the wizard displays a dialog buffer that\n\t      lists the methods. For example, Java's awt class\n\t      hierarchy defines several variants of the method\n\t      <code>repaint<\/code>. If you specify\n\t      <code>repaint<\/code> as the method to override,\n\t      the JDE displays the following dialog buffer: \n\t    <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/OverrideMeth4.gif\" alt=\"Screen shot showing dialog&#10;&#9;&#9;      buffer for selecting method.\" \/><\/div><\/div><p>\n\t      The dialog buffer lists the signature of each variant of\n\t      the method you specified. Next to each signature is a\n\t      radio button. The radio button of the currently selected\n\t      signature contains an asterisk. To select another\n\t      signature, right-click the radio button next to the\n\t      variant. To confirm your selection and dismiss the\n\t      dialog, right-click the [Ok]\n\t      button. \n\t    <\/p><\/li><li><p>\n\t      Select the method variant you want to override and click\n\t      the [Ok] button. \n\t    <\/p><p>The   wizard inserts a skeleton implementation of\n\t      the selected method at the current point in the Java\n\t      source buffer.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/OverrideMeth5.gif\" alt=\"Screen shot showing the inserted&#10;&#9;&#9;      method.\" \/><\/div><\/div><p>\n\t      The wizard also inserts import statements for any\n\t      classes referenced by the method that are not already\n\t      imported by the containing class either explicitly or\n\t      implicitly. The wizard inserts the import statements at\n\t      the head of the source buffer after any existing import\n\t      statements, or any package statement, or the first blank\n\t      line in the buffer. \n\t    <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\t\tThe method override wizard uses the BeanShell to\n\t\tcreate the interface implementation. If the BeanShell\n\t\tis not currently running, the wizard starts the\n\t\tBeanShell. Thus, if the BeanShell is not already\n\t\trunning, you may experience a short pause the first\n\t\ttime you   override a method.\n\t      <\/p><\/div><\/li><\/ol><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Interface Wizard<\/h5><\/div><p>\n\t  This wizard creates a skeleton implementation of any\n\t  interface defined on the classpath specified by\n\t  <code>jde-global-classpath<\/code>.\n\t<\/p><p>To create an implementation of an interface: <\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>\n\t      If the interface is to be implemented by a new class,\n\t      create the class in a buffer.\n\t    <\/p><\/li><li><p>\n\t      Position the Emacs point at the point in the class where\n\t      you want the implementation of the interface's methods\n\t      to appear.\n\t    <\/p><\/li><li><p>\n\t      Select\n\t      <b>JDE<\/b>-&gt;<b>Wizards<\/b>-&gt;<b>Implement Interface<\/b> or enter \n\t      <b>M<\/b>-<b>x\n\t\t<\/b><code>jde-wiz-implement-interface<\/code>. \n\t      \n\t      \n\t      \n\t      \n\t    <\/p><p>\n\t      The JDE prompts you   to enter the name of the interface\n\t      to be implemented. \n\t    <\/p><\/li><li><p>Enter the fully qualified name of the interface, for\n\t      example,\n\t      <code>java.awt.Event.MouseListener<\/code>.<\/p><p>\n\t      The wizard inserts skeleton implementations of the\n\t      methods declared by the interface at the current point\n\t      in the current buffer. It inserts import statements for\n\t      any classes required by the interface at the head of the\n\t      current buffer (only if import statements do not already\n\t      exist for the required classes). It also updates or\n\t      creates an implements clause for the class.\n\t    <\/p><\/li><\/ol><\/div><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>\n\t    The interface wizard uses the BeanShell to create the\n\t    interface implementation. If the BeanShell is not\n\t    currently running, it starts the BeanShell. Thus, if the\n\t    BeanShell is not already running, you may experience a\n\t    short pause the first time you use the wizard. <\/p><\/div><div class=\"note\"><h3 class=\"title\">Thanks to ...<\/h3><p><a href=\"mailto:friedman@lmi.net\">Eric\n\t      Friedman<\/a> for creating the framework for the\n\t    interface, method override, and abstract class wizards and\n\t    for contributing the initial implementation of the\n\t    interface wizard.<\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Delegate Wizard<\/h5><\/div><p>\n\t  This wizard generates methods that delegate calls to a class\n\t  in the current buffer to an attribute of the class, i.e., to\n\t  an object that is a field of the current class. For example,\n\t  if the current buffer contains class\n\t  <code>A<\/code> and  <code>A<\/code> has\n\t  an   attribute, <code>A.b<\/code>, that is an instance\n\t  of class <code>B<\/code>,  this wizard generates\n\t  all the public methods of class <code>B<\/code> in\n\t  <code>A<\/code> and delegates   handling of those\n\t  methods to <code>b<\/code>.\n\t<\/p><div class=\"note\"><h3 class=\"title\">Thanks to...<\/h3><p>\n\t    <a href=\"mailto:cfhart@Z-TEL.com\">Charles Hart<\/a>\n\t    for contributing this wizard.\n\t  <\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Get\/Set Wizard<\/h5><\/div><p>\n\t    This wizard generates get and set methods for the private\n\t  fields of the class at point, but only if the fields do not\n\t  already have get or set methods. To generate the methods,\n\t  move point to the point in the class where you want the get\n\t  and set methods to appear. Then select\n\t  <b>JDE<\/b>-&gt;<b>Code\n\t      Generation<\/b>-&gt;<b>Wizards<\/b>-&gt;<b>Generate Get\/Set Methods...<\/b> or enter \n\t    <b>M<\/b>-<b>x  <\/b> \u00a0\n\t    <code>jde-wiz-get-set-methods<\/code>\n\t<\/p><div class=\"note\"><h3 class=\"title\">Thanks to...<\/h3><p>\n\t    <a href=\"mailto:jslopez@forumsys.com\">Javier\n\t      Lopez<\/a> and <a href=\"mailto:sandip.chitale@brokat.com\">Sandip\n\t      Chitale<\/a> for contributing this wizard.\n\t  <\/p><\/div><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Code Templates<\/h4><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Buffer Templates<\/h5><\/div><p>These commands create buffers containing a skeleton Java class.\n\t    Each command prompts you to enter the path to a new Java source\n\t    file. They then create a buffer for the new file and insert a\n\t    template for a class of the same name as the newly created file.\n\t    In particular, the command <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>\n\t\t<b>Files<\/b>-&gt;<b>JDE New<\/b>-&gt;<b>Class<\/b> creates a buffer containing a generic Java public class.\n\t      <\/p><\/li><li><p>\n\t\t<b>Files<\/b>-&gt;<b>JDE New<\/b>-&gt;<b>Console<\/b> creates a buffer containing the main class of a\n\t\tJava console application.\n\t      <\/p><\/li><li><p>\n\t\t<b>Files<\/b>-&gt;<b>JDE New<\/b>-&gt;<b>Other<\/b> prompts you to create any of the above buffers or a\n\t\tcustom (user-defined) buffer.\n\t      <\/p><\/li><\/ul><\/div><p>\n\t    You can create an empty Java class buffer by selecting \n\t    <b>Files<\/b>-&gt;<b>Open<\/b> \n\t    (<b>C<\/b>-<b>x<\/b>\u00a0f) and\n\t    entering the path for a new file whose root name is the same as\n\t    the class you want to create and whose extension is <code>.java.<\/code>\n\t  <\/p><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Specifying Boilerplate Text<\/h6><\/div><p>\n\t    You can specify boilerplate text (for example, a copyright\n\t    notice) to be inserted at the head of class source files\n\t    created by the JDE. The JDE provides two ways to specify\n\t    the boilerplate text. The simplest way is to enter the\n\t    lines of boilerplate text as the value of the\n\t    customization variable\n\t    <code>jde-gen-buffer-boilerplate<\/code>. Another way\n\t    to specify the text is to set the value of the\n\t    customization variable\n\t    <code>jde-gen-boilerplate-function<\/code>to a\n\t    function that generates the boilerplate text. (The default\n\t    value of this variable is\n\t    <code>jde-gen-create-buffer-boilerplate<\/code>,\n\t    which returns the value of the boilerplate variable,\n\t    <code>jde-gen-buffer-boilerplate<\/code>). The\n\t    functional approach allows you to generate boilerplate\n\t    text dynamically by evaluating the appropriate Lisp code.\n\t    By saving the values of boilerplate variables in project\n\t    files, you can specify different boilerplate text for each\n\t    project. \n\t  <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Customization Variables<\/h6><\/div><p>The following JDE customization variables control creation of\n\t      autocoded Java source buffers: \n\t    <\/p><div class=\"table\"><p><b>Table 2. <\/b><\/p><table summary=\"\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Variable<\/th><th>Group<\/th><th>Usage<\/th><\/tr><\/thead><tbody><tr><td><code>jde-gen-class-buffer-template<\/code><\/td><td>Autocode<\/td><td>Template for a generic public class buffer.<\/td><\/tr><tr><td><code>jde-gen-console-buffer-template<\/code><\/td><td>Autocode<\/td><td>Template for a console application buffer.<\/td><\/tr><tr><td><code>jde-gen-jfc-app-buffer-template<\/code><\/td><td>Autocode<\/td><td>Template for a JFC (Swing) application.<\/td><\/tr><tr><td><code>jde-gen-junit-test-class-buffer-template<\/code><\/td><td>Autocode<\/td><td>Template for a JUnit test case class. This template requires\n\t\t    the <a href=\"http:\/\/www.junit.org\">JUnit<\/a> test framework.<\/td><\/tr><tr><td><code>jde-gen-buffer-templates<\/code><\/td><td>Autocode<\/td><td>Specifies templates available to create\n\t\t      Java buffers.<\/td><\/tr><tr><td><code>jde-gen-buffer-boilerplate<\/code><\/td><td>Autocode<\/td><td>Specifies lines of text to be inserted\n\t\t      at the head of class files.<\/td><\/tr><tr><td><code>jde-gen-boilerplate-function<\/code><\/td><td>Autocode<\/td><td>Specifies a function that returns a string\n\t\t      of boilerplate text. The default value is\n\t\t      <code>jde-gen-create-buffer-boilerplate<\/code>, which returns\n                      the value of <code>jde-gen-buffer-boilerplate<\/code>.<\/td><\/tr><\/tbody><\/table><\/div><p>\n\t      See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#CustomizingTemplates\">Customizing\n\t\tTemplates<\/a> for information on how to customize\n\t      the class creation templates. <\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Point Templates<\/h5><\/div><p>The following commands insert templates at the current\n\t  point in the buffer: <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>\n\t      <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Get\/Set Pair <\/b>\n\t\tgenerates an instance variable and a get and set method  .\n\t\tfor that variable\n\t    <\/p><\/li><li><p>\n\t      <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Println <\/b>\n\t      generates a\n\t      <\/p><pre class=\"programlisting\">System.out.println(...);<\/pre> \n\t      statement.\n\t    <\/li><li><p>\n\t      <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Action <\/b>\n\t\tgenerates and registers an action listener for a\n\t\tspecified component.\n\t    <\/p><\/li><li><p>\n\t      <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Listener<\/b>-&gt;<b>Window\n\t\t<\/b> \n\t\tgenerates and registers a window listener for a specified\n\t\twindow.\n\t    <\/p><\/li><li><p>\n\t      <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Listener<\/b>-&gt;<b>Mouse\n\t\t<\/b>\n\t\tgenerates and registers a mouse listener for a specified\n\t\tcomponent.\n\t    <\/p><\/li><li><p>\n\t      <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Other <\/b>\n\t\tallows you to select any of the above templates or a\n\t\tcustom (user-defined) template.\n\t    <\/p><\/li><\/ul><\/div><p>The following variables control generation of code at\n\t  point: \n\t<\/p><div class=\"table\"><p><b>Table 3. <\/b><\/p><table summary=\"\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Variable<\/th><th>Group<\/th><th>Usage<\/th><\/tr><\/thead><tbody><tr><td><code>jde-gen-get-set-var-template<\/code><\/td><td>Autocode<\/td><td>Defines a get\/set method pair template.<\/td><\/tr><tr><td><code>jde-gen-listener-action-template<\/code><\/td><td>Autocode<\/td><td>Defines an action listener template.<\/td><\/tr><tr><td><code>jde-gen-listener-window-template<\/code><\/td><td>Autocode<\/td><td>Defines a window listener template.<\/td><\/tr><tr><td><code>jde-gen-listener-mouse-template<\/code><\/td><td>Autocode<\/td><td>Defines a mouse listener template.<\/td><\/tr><tr><td><code>jde-gen-inner-class-template<\/code><\/td><td>Autocode<\/td><td>Defines a template for creating a class inside\n\t\t  another class or inside an existing source\n\t\t  buffer.<\/td><\/tr><tr><td><code>jde-gen-code-templates<\/code><\/td><td>Autocode<\/td><td>Specifies available code templates.<\/td><\/tr><\/tbody><\/table><\/div><p>See<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>\n\t      <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#CustomizingTemplates\">Customizing\n\t\tTemplates<\/a> for information on how to customize\n\t      templates provided by the JDE.\n\t    <\/p><\/li><li><p>\n\t      <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#CreatingTemplates\">Creating\n\t\tTemplates<\/a> for information on how to create\n\t      templates from scratch.\n\t    <\/p><\/li><\/ul><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Customizing\n\t  Templates<\/h5><\/div><p>You can customize the JDE's standard code templates,\n\t  using the Emacs customization feature. <\/p><p>To customize an autocode template: <\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Select <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Autocode<\/b><\/p><p>The JDE displays a customization buffer containing\n\t      the autocode templates. <\/p><\/li><li><p>Edit the template to suit your needs.<\/p><p>The JDE uses the template format defined by\n\t      <code>tempo.el<\/code> to represent class\n\t      templates. Each template consists of a list of strings,\n\t      symbols, and functions, each of which represents content\n\t      to be inserted successively into the buffer at the\n\t      current point. The strings represent fixed content. The\n\t      symbols and functions represent variable content. See\n\t      the docstring for the function\n\t      <code>tempo-define-template<\/code> for more\n\t      information, including the meaning of special symbols\n\t      such as <code>'n<\/code>. <\/p><\/li><li><p>Select the state button associated with the\n\t      template.<\/p><p>A menu pops up with a list of options for saving\n\t      your changes. <\/p><\/li><li><p>Save your changes.<\/p><p>Select \n\t      <b>Save for Future\n\t\t  Sessions<\/b> if you want \n\t\tyour changes to apply to all projects. If you want your changes to apply\n\t      only to the current projects, select <b>Set for Current\n\t\t  Session<\/b>. Then select \n\t      <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Save Project<\/b>\n\t      to save your changes in the current project's project\n\t      file.<\/p><\/li><\/ol><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Creating Templates<\/h5><\/div><p>The JDE considers any command (interactive function) that\n\t    inserts code into a buffer at point to be a template. The JDE\n\t    uses the Emacs tempo library to create built-in templates.\n\t    You can use tempo to create your own, add-on templates (see below\n\t    and the doc for the <code>tempo-define-template <\/code>for\n\t    more information) or create templates from scratch.\n\t    In either case, once you have created a template, you can add it\n\t    to the JDE's lists of available code and\/or buffer templates, \n\t    using the JDE's <code>jde-gen-code-templates<\/code> and\/or \n\t    <code>jde-gen-buffer-templates<\/code>\n\t    variables, respectively. Adding a template to these lists enables\n\t    you to invoke the templates from the JDE menus. When adding a\n\t    template, you need to specify a unique title for the template.\n\t    These titles enable you to specify the templates when invoking\n\t    them, using the JDE's custom code template commands  \n\t    (<b>Files<\/b>-&gt;<b>JDE New<\/b>-&gt;<b>Custom<\/b>  and  \n\t    <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Custom<\/b>). You can use\n\t    auto completion to enter a template title when invoking a custom\n\t    code generation command. Note that you can specify different sets\n\t    of templates for different projects, by setting and saving the\n\t    template list variables in project files. See the following\n\t    sections for more information: <\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#DefiningTemplate\">Defining a Template and\n\t\t  Template Insertion Function<\/a><\/p><\/li><li><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#RegisterTemplate\">Registering Custom\n\t\t  Templates<\/a><\/p><\/li><li><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#AssignKeyToTemplate\">Assigning Keys To\n\t\t  emplates<\/a><\/p><\/li><\/ul><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Defining a Template\n\t    and Template Insertion Function<\/h6><\/div><p>The <code>tempo-define-template<\/code> macro\n\t    enables you to define a template and a function that\n\t    inserts that template at the current point in the current\n\t    buffer. You specify the template as a list oftemplate\n\t    elements where each element is text, a special symbol, or\n\t    a Lisp expression. The function inserts each text element\n\t    exactly as specified in the buffer; it replaces special\n\t    symbols with some text (e.g., user input), and it replaces\n\t    Lisp expressions with the text that results from\n\t    evaluating them. <\/p><p>For example, the following Lisp code <\/p><pre class=\"programlisting\">\n\t    (tempo-define-template \"foo\"\n\t    ;; template name\n\t    '(\"System.out.println(\\\"foo\\\");\")\n\t    ;;template definition \"f\"\n\t    ;; abbreviation \"Inserts a print foo message\")\n\t    ;; template documentation\n\t  <\/pre><p>defines a template for Java code that always prints\n\t    \"foo\" to standard out: <\/p><pre class=\"programlisting\">\n\t    System.out.println(\"foo\");\n\t  <\/pre><p>Notice that the template definition uses the Lisp\n\t    string escape character to specify the string\n\t    \"foo\".  This is necessary when you want to\n\t    include quoted strings in a template definition.)\n\t  <\/p><p>The sample Lisp form also defines an interactive\n\t    template function <code>tempo-template-foo<\/code>.\n\t    Now suppose you insert the sample code in your\n\t    <code>.emacs<\/code> file. After Emacs starts up,\n\t    whenever you enter the command <b>M<\/b>-<b>x<\/b>\n\t    <code>tempo-template-foo,<\/code> Emacs inserts\n\t  <\/p><pre class=\"programlisting\">\n\t    System.out.println(\"foo\");\n\t  <\/pre><p>at the current point in your Java source buffer (or\n\t    any buffer, tempo doesn't care). <\/p><p>The preceding example is admittedly not vary useful\n\t    because it always prints the same text. You can create\n\t    more useful templates, using special tempo template\n\t    symbols and lisp forms. This approach, for example, allows\n\t    you to create a template that can print any user-defined\n\t    text to standard out: <\/p><pre class=\"programlisting\">\n\t    (tempo-define-template \"debug\"\n\t    ;; template name '(\"if (debug)\" n&gt;\n\t    ;; insert new line plus indent\n\t    \"System.out.println(\" (p \"Enter debug\n\t    message:\")       ;; Prompts for debug message\n\t    \");\") \"d\"\n\t    ;; abbreviation \"Inserts a print debug message\")\n\t    ;; template documentation \n\t  <\/pre><p>The template function produced by this example prompts\n\t    you to enter the text to be printed when inserting the\n\t    function into a buffer. In particular, it inserts <\/p><pre class=\"programlisting\">\n\t    if (debug) System.out.println(DEBUG-MESSAGE);\n\t  <\/pre><p>where <code>DEBUG-MESSAGE<\/code> is any text\n\t    that you enter. For example, suppose you enter <\/p><pre class=\"programlisting\">\n\t    \"Selected color = \" + color\n\t  <\/pre><p>at the prompt. The template function inserts <\/p><pre class=\"programlisting\">\n\t    if (debug) System.out.println(\"Selected color =\n\t    \" + color);\n\t  <\/pre><p>at the current point in the buffer. <\/p><p>See the documentation for\n\t    <code>tempo-define-template<\/code> (type\n\t    <b>c<\/b>-<b>h<\/b>\u00a0f\u00a0\n\t      <code>tempo-define-template<\/code>) for more\n\t    information on creating templates. <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Registering Custom\n\t    Templates<\/h6><\/div><p>You can register templates that you create with the\n\t    JDE. When you register a template with the JDE, it appears\n\t    among the list of templates that you can select when you\n\t    select <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Other....<\/b>. You\n\t    register a template by customizing the JDE variable\n\t    <code>jde-gen-code-templates<\/code>. The value of\n\t    this variable is a list of the template functions that the\n\t    JDE command  <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Other....<\/b> can\n\t    invoke. To register a custom template, add its name to the\n\t    list. For example, the following screen shows the\n\t    customization buffer for\n\t    <code>jde-gen-code-templates<\/code> after it has\n\t    been customized to include the template defined in the\n\t    previous example. <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/RegisterTemplate.gif\" alt=\"Screenshot showing customization buffer for&#10;&#9;&#9;    registering templates.\" \/><\/div><\/div><p>To insert a template that you have registered <\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Select <b>JDE<\/b>-&gt;<b>Generate<\/b>-&gt;<b>Other....<\/b><\/p><p>The JDE displays the prompt<\/p><pre class=\"programlisting\">\n\t\tEnter template:\n\t      <\/pre><p>in the minibuffer. <\/p><\/li><li><p>Enter the template's name and press\n\t\t<b>Enter<\/b><\/p><p>or,<\/p><p>press the <b>Tab<\/b> key to display a\n\t\tlist of templates in a completion buffer: <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/TemplateList.gif\" alt=\"Screenshot showing template completion&#10;&#9;&#9;&#9;buffer.\" \/><\/div><\/div><p>Select the template you want by double-clicking\n\t\tits name.<\/p><\/li><\/ol><\/div><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Assigning Keys to\n\t    Templates<\/h6><\/div><p>You can assign templates to keyboard keys to speed use\n\t    of frequently used templates. For example, insert this\n\t    form <\/p><pre class=\"programlisting\">\n\t    (global-set-key [f9] 'jde-gen-to-string-method)\n\t  <\/pre><p>in your <code>.emacs<\/code> file to assign the\n\t    <b>F9<\/b> function key to the JDE template that\n\t    generates a skeleton <code>toString<\/code> method.\n\t  <\/p><\/div><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Checking Coding Style<\/h2><\/div><p>The <b>JDE<\/b>-&gt;<b>Check Style<\/b> command (M-x <code>jde-checkstyle<\/code>) command checks\n      the current buffer for conformity to a Java coding standard. The\n      default standard is that specified by Sun Microsystems. Execute\n      <code>jde-checkstyle-customize<\/code> to customize the\n      command to support your own coding standard.\n    <\/p><p>The command displays a list of stylistic faults in a popup\n    buffer. Click the error message to display the line that violates\n    the standard.<\/p><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Compiling Java Programs<\/h2><\/div><p>To compile the source file in the current buffer, select\n      <b>JDE<\/b>-&gt;<b>Compile<\/b>, type\n      <b>C<\/b>-<b>c<\/b>-<b>C<\/b>-<b>v<\/b>-<b>C<\/b>-<b>c<\/b>,\n      or execute <code>M-x jde-compile<\/code>. The compiler uses\n      the version of <code>javac<\/code> included in the\n      currently selected JDK or on your system command path by\n      default. However, you can configure the JDEE to use another\n      compiler for the current project or all projects (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#SpecifyingCompiler\">Specifying a Compiler<\/a>). The\n      JDEE also allows you to configure the JDEE to invoke the\n      compiler with various <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#CompilationOptions\">compilation options<\/a>.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Compilation Buffer<\/h4><\/div><p>\n\tThe compile command displays the output of the compiler in a\n\tseparate compilation buffer.\n      <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/compile_buffer.gif\" alt=\"Image showing compilation buffer with&#10;&#9;      errors.\" \/><\/div><\/div><p>\n\tIf a compilation buffer does not exist, the compile command\n\tcreates the buffer; otherwise, it reuses the existing compile\n\toutput buffer. The compilation buffer operates in\n\t<code>compilation-mode<\/code>, a standard Emacs buffer\n\tmode. This mode greatly simplify locating compilation errors\n\tin the Java source code. For example, to find the line that\n\tcause a compilation error, simply click the error message in\n\tthe compilation buffer. \n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Specifying a Compiler<\/h4><\/div><p>The JDEE supports the following Java compilers.<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>javac<\/code><\/p><p>\n\t    This is the compiler that comes with Sun's JDK and the\n\t    JDK's clones. Note that <code>javac<\/code>\n\t    compiler is itself implemented in Java. The\n\t    <code>javac<\/code> executable that comes with the\n\t    JDK simply starts a vm to run <code>javac<\/code>.\n\t    The executable passes its command-line arguments to\n\t    <code>javac<\/code>'s main method.\n\t  <\/p><\/li><li><p>javac compile server<\/p><p>This is a JDEE feature that uses the BeanShell to\n\t  invoke the javac compiler on a source file, thus avoiding\n\t  the need to start a vm everytime you want to compile\n\t  a file. This option greatly reduces the time required\n\t  to compile individual files by eliminating the vm startup\n\t  time, which is usually much longer than the time required\n\t  to compile a file.<\/p><\/li><li><p><code>jikes<\/code><\/p><p>\n\t    Developed by IBM, this compiler is implemented in\n\t  C++ and hence avoids the vm startup time that slows\n\t  down javac.\n\t  <\/p><\/li><\/ul><\/div><p>\n\tTo specify one of these compilers for the current project or\n\tall projects:\n<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Type <code>M-x customize-variable<\/code> and\n\t    press <code>Enter<\/code><\/p><p>Emacs prompts you to enter the name of a variable\n\t  to customize.<\/p><\/li><li><p>Enter <code>jde-compiler<\/code> and\n\t  press <code>Enter<\/code>.<\/p><p>The <code>jde-compiler<\/code> customization\n\t  buffer appears.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/spec_compiler.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Select the compiler that you want to use for the\n\t  current project or all projects.<\/p><p>\n\t    You can skip the next step if you selected the compile server\n\t    or if you selected javac and you want to use the version of\n\t    javac that comes with the currently selected JDK, or you have\n\t    selected the javac executable or jikes and the executable for\n\t    the selected compiler is in your system's command path.\n\t  <\/p><\/li><li><p>Enter the path to the executable for the selected\n\t    compiler.\n\t  <\/p><\/li><li><p>If you want your selection to apply only to the current\n\t    project, select <code>Save for Current Session<\/code>\n\t    from the <code>State<\/code> menu. Otherwise select\n\t    <code>Save for Future Sessions<\/code>.<\/p><\/li><li><p>Select <code>Finish<\/code> to dismiss the\n\t  customization buffer.<\/p><\/li><li><p>If you want the selection to apply to the current\n\t    project, select <code>Project-&gt;Project\n\t      File-&gt;Save<\/code> from the <code>JDE<\/code>\n\t    menu.<\/p><\/li><\/ol><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Compilation Options<\/h4><\/div><p>The JDEE allows you to specify compilation options by\n\tsetting  compilation variables. You must use the Emacs\n\tcustomization feature. To use the customization feature,\n\tselect <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Compile<\/b>.  (See\n\t<a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ConfiguringJDE\">Configuring  the JDEE<\/a> for\n\tmore information on using the customization  feature). To save\n\tthe compilation settings in the project  file \n\t(see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#UsingProjectFiles\">Using Project Files<\/a>)  \n\tfor the current source buffer, select <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Update  Project<\/b>.\n      <\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> Setting\n\t  Compile Options Interactively<\/h5><\/div><p>If you set the customization variable\n\t  <code>jde-read-compile-args<\/code>  to a\n\t  non-<code>nil<\/code> value, the JDEE compile command\n\t  prompts you to enter compilation options in the minibuffer.\n\t  It appends the  options that you enter to the options\n\t  specified via customization  variables. The JDEE saves the\n\t  arguments that you enter in a  minibuffer history list. You\n\t  can recall previously entered  options by pressing the up or\n\t  down arrows on your keyboard. <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>The JDEE uses the values of the JDEE customization\n\t    variables to set the compiler's command-line option\n\t    switches. The JDEE assumes that the compiler you are using\n\t    (specified by the  customization variable\n\t    <code>jde-compiler<\/code>) has the same set of\n\t    command-line switches as the version of\n\t    <code>javac<\/code> or jikes that you have selected.\n\t    If the command-line  switch for a particular option\n\t    supported by the compiler  your are using is not the same\n\t    as that specified by <code>javac<\/code> or jikes,\n\t    you must use the variable\n\t    <code>jde-compile-option-command-line-args<\/code> to\n\t    select the option.<\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">\n\t  Compiler Customization Variables<\/h5><\/div><p>The following table lists the JDEE compilation variables and\n\t  the functions used to set them.<\/p><div class=\"table\"><p><b>Table 1. <\/b><\/p><table summary=\"\" border=\"1\"><colgroup><col align=\"left\" \/><col align=\"left\" \/><col align=\"left\" \/><\/colgroup><thead><tr><th align=\"left\"><p>Name<\/p><\/th><th align=\"left\"><p>Group<\/p><\/th><th align=\"left\"><p>Usage<\/p><\/th><\/tr><\/thead><tbody valign=\"top\"><tr><td align=\"left\"><p><code>jde-compiler<\/code><\/p><\/td><td align=\"left\">Project<\/td><td align=\"left\"><p>Specifies the compiler (javac server, by default)\n\t\t    to use to compile the code in  the current source buffer.\n\t\t  <\/p><\/td><\/tr><tr><td align=\"left\"><p>\n\t\t    <code>jde-compile-option-command-line-args<\/code>\n\t\t  <\/p><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Specifies a string of\n\t\t  command-line arguments to be passed to the compiler.<\/td><\/tr><tr><td align=\"left\"><code>jde-global-classpath<\/code><\/td><td align=\"left\">Project<\/td><td align=\"left\">Specify class paths\n\t\t  for compile, run, and debug commands.<\/td><\/tr><tr><td align=\"left\"><code>jde-read-compile-args<\/code><\/td><td align=\"left\">Project<\/td><td align=\"left\">Specify whether to read compile\n\t\t  options from the minibuffer.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-classpath<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Specifies the\n\t\t  classpath for compile command. If set, this variable\n\t\t  overrides <code>jde-global-classpath<\/code>.\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-sourcepath<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Specifies the path of source files\n\t\t  for classes required to compile the current class.<\/td><\/tr><tr><td align=\"left\"><code>jde-quote-classpath<\/code><\/td><td align=\"left\">Project\u00a0<\/td><td align=\"left\">Quote the classpath\n\t\t  argument.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-directory<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Specifies the\n\t\t  directory into which to place the compiled class.\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-deprecation<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Warn of use or\n\t\t  override of a deprecated member or class\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-debug<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Generate information\n\t\t  about local variables for debug tools.\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-optimize<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Directs the compiler\n\t\t  to try to generate faster varname.\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-depend<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Analyze dependencies.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-depend-switch<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Command line switch that causes the\n\t\t  compiler to analyze dependencies.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-vm-args<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Specify command-line\n\t\t  arguments for Java interpreter used to run the compiler.\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-verbose-path<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Print verbose messages.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-verbose<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">List directories searched to\n\t\t  compile current class.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-nowarn<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Turn off warnings.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-encoding<\/code><\/td><td align=\"left\">Compile\u00a0<\/td><td align=\"left\">Specify the source\n\t\t  file encoding name, such as EUCJIS\\SJIS.\u00a0<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-target<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Generate code compatible with a\n\t\t  specified vm version.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-bootclasspath<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Specify classpath of standard\n\t\t  libraries for target vm.<\/td><\/tr><tr><td align=\"left\"><code>jde-compile-option-bootclasspath<\/code><\/td><td align=\"left\">Compile<\/td><td align=\"left\">Specify path of directories\n\t\t  containing extensions for target vm.<\/td><\/tr><\/tbody><\/table><\/div><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Building Java Applications<\/h2><\/div><p>The <b>JDE<\/b>-&gt;<b>Build<\/b>\n      command builds an application. By default this command invokes the\n      Unix-style make utility to build the application. You can configure\n      the build command to invoke the Ant build system or a custom build\n      function.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Selecting a Build Method<\/h4><\/div><p>To select a build method:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Type <code>M-x customize-variable<\/code> and\n\t    press <code>Enter<\/code><\/p><p>Emacs prompts you to enter the name of a variable\n\t  to customize.<\/p><\/li><li><p>Enter <code>jde-build-function<\/code> and\n\t  press <code>Enter<\/code>.<\/p><p>The <code>jde-build-function<\/code> customization\n\t  buffer appears.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/build_fcn_cust_buff.gif\" alt=\"image\" \/><\/div><\/div><\/li><li><p>Select the build function that you want to use.<\/p><\/li><li><p>If you want your selection to apply only to the current\n\t    project, select <code>Save for Current Session<\/code>\n\t    from the <code>State<\/code> menu. Otherwise select\n\t    <code>Save for Future Sessions<\/code>.<\/p><\/li><li><p>Select <code>Finish<\/code> to dismiss the\n\t  customization buffer.<\/p><\/li><li><p>If you want the selection to apply to the current\n\t    project, select <code>Project-&gt;Project\n\t      File-&gt;Save<\/code> from the <code>JDE<\/code>\n\t    menu.<\/p><\/li><\/ol><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Building with Make<\/h4><\/div><p>\n\tThe <code>jde-make<\/code> command uses a make utility,\n\tsuch as  GNU make, to rebuild a project. You must supply the\n\tmakefiles used to build your projects. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#SampleMakefile\">Sample Makefile<\/a> for an example\n\tof a makefile for building a Java project. David Ponce's\n\t<a href=\"http:\/\/perso.wanadoo.fr\/david.ponce\/\">JMaker<\/a>\n\tpackage generates makefiles automatically for JDEE projects.\n\tIt uses JDEE variables to determine classpaths, compiler\n\toptions, and other information needed to create projects.\n\tJMaker recursively creates makefiles for all the Java packages\n\tin a project plus a toplevel makefile that invokes the other\n\tmakefiles to build the entire project.\n      <\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Make Options<\/h5><\/div><p>\n\t  The following customization variables  allow you to specify\n\t  various <code>jde-make<\/code> options:\n\t<\/p><div class=\"table\"><p><b>Table 1. <\/b><\/p><table summary=\"\" border=\"1\"><colgroup><col align=\"left\" \/><col align=\"left\" \/><col align=\"left\" \/><col align=\"left\" \/><\/colgroup><thead><tr><th align=\"left\">Name<\/th><th align=\"left\">Group<\/th><th align=\"left\">Default<\/th><th align=\"left\">Usage<\/th><\/tr><\/thead><tbody><tr><td align=\"left\" valign=\"top\"><code>jde-make-program<\/code><\/td><td align=\"left\" valign=\"top\">Project<\/td><td align=\"left\">make<\/td><td align=\"left\"><p>Specifies the pathname on your system of the\n\t\t    make utility that the <code>jde-make<\/code>\n\t\t    uses to build your project. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-make-args<\/code><\/td><td align=\"left\" valign=\"top\">Project<\/td><td align=\"left\">\u00a0<\/td><td align=\"left\"><p> Specifies a string of command-line arguments\n\t\t    (for  example, \"<code>-f mymakefile\n\t\t      all<\/code>\") to be passed to  the make\n\t\t    utility. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-read-make-args<\/code><\/td><td align=\"left\" valign=\"top\">Project<\/td><td align=\"left\"><code>nil<\/code><\/td><td align=\"left\"><p> If nonnil, this variable causes the\n\t\t    <code>jde-make<\/code> command to prompt you\n\t\t    to enter command-line arguments in the minibuffer.\n\t\t    <code>jde-make<\/code> passes these arguments\n\t\t    along with the arguments specified by\n\t\t    <code>jde-make-args<\/code> to the make\n\t\t    utility. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-make-working-directory<\/code><\/td><td align=\"left\" valign=\"top\">Project<\/td><td align=\"left\">\u00a0<\/td><td align=\"left\"><p> Specifies the directory from which\n\t\t    <code>jde-make<\/code> runs the make utility.\n\t\t  <\/p><\/td><\/tr><\/tbody><\/table><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Sample Makefile<\/h5><\/div><p>The following is a sample makefile that you can use as a\n\t  model for creating makefiles to build Java projects. <\/p><pre class=\"programlisting\">\n\t  JDK = d:\/jdk1.2\/lib\/classes.zip \n\t  JMATH = d:\/jmath\/src\n\t  CLASSPATH = $(JMATH);$(JDK) \n\t  COMPILER = javac VM = java\n\t  COPTIONS = -g -deprecation \n\t  ROPTIONS = \n          CLASSES = Test.class \\\n\t            JMathError.class \\ \n                    JMathException.class \\ \n                    LinearSystem.class\n\t  .SUFFIXES: .java .class \n          all : $(CLASSES) \n            # Rebuild (if necessary) and run the JMath test suite. run: $(CLASSES)\n\t    $(VM) -classpath \"$(CLASSPATH)\" jmath.Test \n          # Remove all class files from the project directory. \n          clean: rm *.class \n          \n          # Implicit rule for making Java class files from Java \n          # source files. \n          .java.class: $(COMPILER) $(COPTIONS) -classpath \"$(CLASSPATH)\" $? \n\t<\/pre><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Building with Ant<\/h4><\/div><p>\n\tThe <code>jde-ant-build<\/code> command uses the <a href=\"http:\/\/jakarta.apache.org\/ant\/\">Apache Ant<\/a>\n\tbuild tool to build a Java project.\n      <\/p><p>The following variables allow you to control the build\n\tprocess.<\/p><div class=\"table\"><p><b>Table 2. Ant Options<\/b><\/p><table summary=\"Ant Options\" border=\"1\"><colgroup><col align=\"left\" \/><col align=\"left\" \/><col align=\"left\" \/><col align=\"left\" \/><\/colgroup><thead><tr><th align=\"left\">Name<\/th><th align=\"left\">Group<\/th><th align=\"left\">Default<\/th><th align=\"left\">Usage<\/th><\/tr><\/thead><tbody><tr><td align=\"left\" valign=\"top\"><code>jde-ant-invocation-method<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\">script<\/td><td align=\"left\"><p> Specifies how to invoke ant. Ant can be invoked\n\t\t  in one of three ways. The first is via the ant\n\t\t  script\/program that comes with ant. The second is\n\t\t  via java and the third is via the Ant Server.\n\t\t<\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-home<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\">\u00a0<\/td><td align=\"left\"><p> Directory where ant is installed. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-user-jar-files<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><code>nil<\/code><\/td><td align=\"left\"><p> Specifies jar files that hold user-defined\n\t\t  tasks. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-program<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>ant<\/p><\/td><td align=\"left\"><p> Specifies name of ant program\/script. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-args<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>-emacs<\/p><\/td><td align=\"left\"><p> Specifies arguments to be passed to the Ant\n\t\t  program. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-buildfile<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>build.xml<\/p><\/td><td align=\"left\"><p> Specifies the default buildfile to use. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-read-buildfile<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p> Specify whether to prompt for a buildfile. If\n\t\t  non-nil, the jde-ant-build command prompts you for\n\t\t  an ant buildfile.  Note that\n\t\t  <code>jde-ant-read-buildfile<\/code> and\n\t\t  <code>jde-ant-enable-find<\/code> are\n\t\t  incompatible and that\n\t\t  <code>jde-ant-read-buildfile<\/code> will\n\t\t  override the former. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-read-target<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p> Specify whether to prompt for a build target.\n\t\t  If non-nil, the jde-ant-build command prompts you\n\t\t  for an ant target. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-interactive-buildfile<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p> Default buildfile to use when prompting\n\t\t  interactively. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-read-args<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p> Specify whether to prompt for additional\n\t\t  arguments to pass to Ant. If non-nil, the\n\t\t  <code>jde-ant-build<\/code> command prompts you\n\t\t  for the additional arguments. <\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-enable-find<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p> Specify whether jde-ant find the\n\t\t  <code>build.xml<\/code> file based on your\n\t\t  current directory. If non-nil, we will search up the\n\t\t  directory hierarchy from the current directory for\n\t\t  the build definition file. Also note that, if\n\t\t  non-nil, this will relax the requirement for an\n\t\t  explicit jde project file.  In order for this to\n\t\t  work <code>jde-ant-read-buildfile<\/code> must\n\t\t  be nil.<\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-complete-target<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p> Specify whether to enable completion of build\n\t\t  target names in the minibuffer. If non-nil, the\n\t\t  <code>jde-ant-build<\/code> command allows you\n\t\t  to use tab completion in the minibuffer to specify\n\t\t  the build target name.  This list of valid build\n\t\t  targets is determined by parsing the Ant build file.\n\t\t  This option has no effect if\n\t\t  <code>jde-ant-read-target<\/code> is\n\t\t  nil.<\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-use-global-classpath<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>off<\/p><\/td><td align=\"left\"><p>Specify whether to enable use of\n\t\t  <code>jde-global-classpath<\/code> when running\n\t\t  <code>jde-ant<\/code>.<\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-target-regexp<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>\n\t\t  &lt;\\s-*target\\s-[^...]*?name\\s-*=\\s-*\\\"\\s-*\\([^\\\"]+\\) \n\t\t<\/p><\/td><td align=\"left\"><p>Regular expression used to match target names in\n\t\t  Ant build files.<\/p><\/td><\/tr><tr><td align=\"left\" valign=\"top\"><code>jde-ant-build-hook<\/code><\/td><td align=\"left\" valign=\"top\">Ant<\/td><td align=\"left\"><p>nil<\/p><\/td><td align=\"left\"><p>List of hook functions run by\n\t\t  <code>jde-ant-build<\/code> (see\n\t\t  <code>run-hooks<\/code> in the Emacs\n\t\t  documentation for more information).<\/p><\/td><\/tr><\/tbody><\/table><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Running Java\n      Applications<\/h2><\/div><p>The JDE allows you to run a Java application as an Emacs\n      subprocess. You can run multiple applications concurrently, but\n      only one instance of each application at a time. The JDE\n      displays each application's standard and error output in an\n      Emacs command interface (<code>comint<\/code>) buffer. You\n      can interact with applications that accept command line input\n      via the <code>comint<\/code> buffer. When interacting with\n      an application, you can use any of\n      <code>comint-mode<\/code>'s extensive set of command-line\n      history commands to speed interaction. To run an application,\n      enter <\/p><pre class=\"programlisting\">\n      M-x jde-run\n    <\/pre><p>or select <b>Java<\/b>-&gt;<b>Run App<\/b> from the Emacs\n      menubar or type <b>C<\/b>-<b>c<\/b>\u00a0<b>C<\/b>-<b>v<\/b>\u00a0<b>C<\/b>-<b>r<\/b>. <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Specifying the\n\tApplication's Main Class<\/h4><\/div><p>The term <i>main class<\/i> refers to the\n\tclass that contains the application's main method. The JDE's\n\tRun App command assumes by default that the class in the\n\tcurrent buffer is the application's main class. This can be\n\tinconvenient if you have an application that has multiple\n\tclasses. You may want to be able to run the application from\n\tany buffer containing one of the application's classes. To do\n\tthis, set the variable\n\t<code>jde-run-application-class<\/code> to the\n\t<i>fully qualified name<\/i> of the application's\n\tmain class.<\/p><p>You can temporarily override the setting of\n\t<code>jde-run-application-class<\/code> by typing\n\t<code>C-u -<\/code> before executing the <code>jde-run<\/code> command, for example, by\n\ttyping <code>C-u - C-c C-r<\/code>. This causes the JDE to\n\tprompt you to enter the <i>fully qualified\n\t  name<\/i> of the application's main class (as well\n\tapplication arguments if any).<\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Specifying a\n\tStartup Directory<\/h4><\/div><p>The JDE can start an application from any directory that\n\tyou specify. By default, the JDE starts an application from\n\tthe default directory of the current source buffer. The\n\tdefault directory of the current source buffer is the\n\tdirectory containing the source file displayed in the buffer.\n\tYou can specify another directory as the startup directory by\n\tsetting the JDE customization variable\n\t<code>jde-run-working-directory<\/code>.<\/p><p>To set this variable, <\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Display its customization panel.<\/p><p>You can do this by  typing <\/p><p><b>M<\/b>-<b>x<\/b>\u00a0<code>customize-variable<\/code>\u00a0<code>jde-run-working-directory<\/code><\/p><p>or selecting <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Project<\/b> to display the project customization  panel\n\t    and searching this panel for\n\t    <code>jde-run-working-directory<\/code>.\n\t  <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/WorkingDirectory.gif\" alt=\"Screenshow showing customization buffer for&#10;&#9;&#9;    jde-run-working-directory.\" \/><\/div><\/div><\/li><li><p>Enter the working directory in the variable's edit\n\t    field. <\/p><\/li><li><p>Save the new setting.<\/p><p>To save the edited bindings,  right click the\n\t    [State]  button and choose\n\t    <b>Set for current\n\t\tsession<\/b> if you want the\n\t    settings to apply only  to the current project or\n\t    <b>Save for  future\n\t\tsessions<\/b> if you want the\n\t    settings to  apply to all projects. In either case, you\n\t    should save  the new settings in you project file if your\n\t    project has  a project file. To save the new settings in\n\t    your project  file, switch to a source buffer and choose\n\t    <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Save\n\t\tProject<\/b>.<\/p><\/li><\/ol><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Setting VM Command-Line\n\tArguments<\/h4><\/div><p>If you set the customization variable\n\t<code>jde-run-read-vm-args<\/code>  to a\n\tnon-<code>nil<\/code> value, the JDE compile command\n\tprompts you to enter virtual machine options in the\n\tminibuffer. It appends  the options that you enter to the\n\toptions specified via  customization variables. The JDE saves\n\tthe arguments that you  enter in a minibuffer history list.\n\tYou can recall previously  entered options by pressing the up\n\tor down arrows on your  keyboard. <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Setting Command-Line\n\tApplication Arguments<\/h4><\/div><p>Specifying a prefix argument before the\n\t<code>jde-run<\/code>, e.g., <code>C-u C-c C-v\n\t  C-r<\/code> causes the command to prompt you to enter\n\targuments to be passed to the application's main\n\tmethod. The command appends the options that you enter to the\n\targuments specified via  the customization variable\n\t<code>jde-run-option-application-args<\/code>.  The JDE\n\tsaves the arguments that you enter in a minibuffer  history\n\tlist. You can recall previously entered options by  pressing\n\tthe up or down arrows on your keyboard. <\/p><p>If you set the customization variable\n\t<code>jde-run-read-app-args<\/code>  to a\n\tnon-<code>nil<\/code> value, the JDE run command\n\tprompts you to enter command-line application arguments \n\tby default, i.e., you do not have to specify a prefix\n\targument each time you execute the command.<\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Navigating\n\tException Traces<\/h4><\/div><p>If an exception occurs while your program is running, the\n\tJava vm outputs an exception trace. This trace appears in the\n\trun buffer for the application.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/excepttrace.gif\" alt=\"Screenshot showing an exception trace in the&#10;&#9;      JDE's run buffer.\" \/><\/div><\/div><\/div><p>To view the source line corresponding to a point in the\n      exception stack trace, right-click the corresponding stack trace\n      line in the run buffer. You can move up and down the stack trace\n      by typing the key combinations <code>C-c C-v C-[<\/code> or\n      <code>C-c C-v C-]<\/code>.<\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Run Customization Variables\n      <\/h4><\/div><p>The JDE allows you to specify run-time options by setting\n\trun variables.You can use the Emacs customization feature to\n\tset run-time variables interactively. To use the customization\n\tfeature, select\n\t<b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>Run<\/b>. \n\t(See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ConfiguringJDE\">Configuring  the JDE<\/a>\n\tfor more information on using the customization  feature). To\n\tsave the compilation settings in the project file (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#UsingProjectFiles\">Using Project Files<\/a>) for\n\tthe  current source buffer, select <b>JDE<\/b>-&gt;<b> Project<\/b>-&gt;<b>Project\n\t    File<\/b>-&gt;<b>Save\n\t    Project<\/b>.\n      <\/p><p>The following table lists the JDE run variables and the\n\tfunctions used to set them. \u00a0 <\/p><div class=\"table\"><p><b>Table 1. <\/b><\/p><table summary=\"\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Variable<\/th><th>Group<\/th><th>Usage<\/th><\/tr><\/thead><tbody><tr><td valign=\"top\"><code>jde-run-mode-hook<\/code><\/td><td valign=\"top\">Project<\/td><td valign=\"top\">List of hook functions run by\n\t\t<code>jde-run-mode<\/code><\/td><\/tr><tr><td valign=\"top\"><code>jde-run-working-directory<\/code><\/td><td valign=\"top\">Project<\/td><td>Startup directory for running or debugging Java\n\t\tapplications.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-application-class<\/code><\/td><td valign=\"top\">Project\u00a0<\/td><td valign=\"top\">Name of the Java class to run. The\n\t\tcommand <code>jde-run-set-app<\/code> sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-java-vm<\/code><\/td><td valign=\"top\">Project\u00a0<\/td><td valign=\"top\">Specify Java interpreter for\n\t\tnon-Windows platforms. The command\n\t\t<code>jde-run-set-vm<\/code> sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-java-vm-w<\/code><\/td><td valign=\"top\">Project\u00a0<\/td><td valign=\"top\">Specify Java interpreter for Windows\n\t\tplatforms. The command\n\t\t<code>jde-run-set-vm-w<\/code> sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-global-classpath<\/code><\/td><td valign=\"top\">Project\u00a0<\/td><td valign=\"top\">Specify class paths for compile,\n\t\trun, and debug commands. The command<code>\n\t\t  jde-set-global-classpath<\/code> sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-classic-mode-vm<\/code><\/td><td valign=\"top\">Project<\/td><td>Specifies that the JDE should run the JDK's\n\t\tclassic version of the Java virtual machine (as\n\t\topposed to the HotSpot version). This option applies\n\t\tonly to versions of the JDK that include both a\n\t\tclassic and a HotSpot vm.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-read-vm-args<\/code><\/td><td valign=\"top\">Project<\/td><td>Specifies whether to read vm arguments from\n\t\tthe\u00a0 minibuffer.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-classpath<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Specifies the classpath for\u00a0\n\t\tthe Java interpreter. This option overrides the\n\t\tjde-global-classpath option.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-verbose<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Print messages about the running\n\t\tprocess.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-properties<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Specify property values.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-heap-size<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Specify the initial and maximum size\n\t\tof the interpreter heap.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-stack-size<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Specify size of the C and Java\n\t\tstacks.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-garbage-<\/code>\u00a0 \n\t\t<code>collection<\/code>\u00a0<\/td><td valign=\"top\">Run<\/td><td valign=\"top\">Specify garbage collection\n\t\toptions.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-java-profile<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Enable Java profiling.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-heap-profile<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Output heap profiling data.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-verify<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Verify classes.<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-vm-args<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Specify command-line arguments to be\n\t\tpassed to the Java vm. The command jde-run-set-args\n\t\tsets the buffer-local value of this\n\t\tvariable.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-option-application-args<\/code><\/td><td valign=\"top\">Run\u00a0<\/td><td valign=\"top\">Specify command-line arguments to\n\t\tpass to the application. The command\n\t\t<code>jde-run-set-app-args <\/code>sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><\/tbody><\/table><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\"> Working with Applets<\/h2><\/div><p>The JDE provides specialized commands for running and\n      debugging applets.<\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Running Applets<\/h4><\/div><p>To run an applet:<\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Open the applet's source file in a buffer. <\/p><\/li><li><p>Choose <b>JDE<\/b>-&gt;<b>Run\n\t\tApplet<\/b>. <\/p><p>The JDE searches for an html file in the directory\n\t    containing the applet source file and displays the first\n\t    file it finds, using your system's default browser. If\n\t    the JDE cannot find an html page in the applet's source\n\t    file directory, it signals an error by default.\n\t  <\/p><\/li><\/ol><\/div><p>The  <b>JDE<\/b>-&gt;<b>Run Applet<\/b> command\n\tassumes by default that the directory containing the  applet's\n\tsource also contains an html page for testing that  applet. If\n\tthis is not true in your case, you have two options.  You can\n\teither specify the path of the applet file via the JDE\n\tcustomization variable <code>jde-run-applet-doc<\/code>\n\tor you can use  the <code>jde-run-applet<\/code> command\n\tto run your applet. If the\n\t<code>jde-run-applet-doc<\/code>  variable is not a null\n\tstring, <b>JDE<\/b>-&gt;<b>Run Applet<\/b> displays\n\tthe\n\tdocument that the variable  specifies instead of searching the source\n\tfile directory for a  document to display. The\n\t<code>jde-run-applet<\/code> command prompts  you to enter in the\n\tminibuffer the path of an html file to  display. If you enter nothing,\n\tthe command defaults to the  behavior of  <b>JDE<\/b>-&gt;<b>Run Applet<\/b>.<\/p><p>The run commands chooses the viewer as follows. If\n\t<code>jde-run-applet-viewer<\/code> is a null string (the\n\tdefault) or browse-url, the JDE uses browse-url to launch the\n\tapplet doc in your system's default browser. Otherwise, the\n\tmenu command uses comint to launch the viewer (e.g.,\n\tappletviewer) specified by\n\t<code>jde-run-applet-viewer<\/code>.<\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Debugging Applets<\/h4><\/div><p>To debug an applet:<\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Open the applet's source file in a buffer.<\/p><\/li><li><p>Choose  <b>JDE<\/b>-&gt;<b>Debug\n\t\tApplet<\/b>.<\/p><p>The JDE searches for an html file in the directory\n\t    containing the applet source file. If the JDE cannot find\n\t    an html page in the applet's source file directory, it\n\t    signals an error by default. Otherwise, it runs\n\t    appletviewer in debug mode on the first file that it\n\t    finds.<\/p><p>The Emacs window splits into two panes.<\/p><\/li><\/ol><\/div><p>The top pane shows the applet's source file with the debug\n\tcursor pointing to the first line of the applet's init file.\n\tThe bottom pane displays the debugger's command line\n\tinterface. You can now set breakpoints, single-step, or\n\tcontinue running the applet. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#DebuggingApps\">Debugging Applications<\/a> for more\n\tinformation.<\/p><p>If you want to specify the document to be used to test the\n\tapplet, customize the variable jde-run-applet-doc or execute\n\t<code>M-x jde-db-applet<\/code>. This command prompts you\n\tto enter the test document's name.<\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Applet Customization\n\tVariables<\/h4><\/div><p>The JDE allows you to specify run-time options for applets\n\tby setting JDE configuration variables.You must use the Emacs\n\tcustomization feature to set applet run-time variables . To\n\tuse the customization feature, select <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Run<\/b>. \n\t(See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ConfiguringJDE\">Configuring the JDE<\/a>\n\tfor more information on using the customization feature). In\n\tsome cases, the JDE also provides functions for setting the\n\tbuffer-local values of the compilation variables. To save the\n\tcompilation settings in the project file (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#UsingProjectFiles\">Using Project Files<\/a>) for\n\tthe current source buffer, select\n\t<b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Update \n\t    Project<\/b>. <\/p><p>The following table lists the JDE applet run-time\n\tvariables and the functions used to set them.  <\/p><div class=\"table\"><p><b>Table 1. <\/b><\/p><table summary=\"\" border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Variable<\/th><th>Group<\/th><th>Usage<\/th><\/tr><\/thead><tbody><tr><td valign=\"top\"><code>jde-run-applet-viewer<\/code><\/td><td valign=\"top\">Project\u00a0<\/td><td valign=\"top\">Specify name of viewer to use to\n\t\tdisplay page containing the applet. The command\n\t\t<code>jde-run-set-applet-viewer<\/code> sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><tr><td valign=\"top\"><code>jde-run-applet-doc<\/code><\/td><td valign=\"top\">Project\u00a0<\/td><td valign=\"top\">Specify name of document containing\n\t\tapplet to be viewed. The command\n\t\t<code>jde-run-set-applet-doc<\/code> sets the\n\t\tbuffer-local value of this variable.\u00a0<\/td><\/tr><\/tbody><\/table><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Debugging Applications<\/h2><\/div><p>The JDE provides two options for debugging Java applications.<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>An Emacs interface to <code>jdb<\/code>, the command-line\n\t  debugger that comes with the JDK. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jdb-ug\/jdb-ug-frame.html\" target=\"jdb-ug\">Debugging with jdb<\/a> for more\n\t  information. <\/p><\/li><li><p>\n\t  JDEbug, a Java debugger developed specifically for use  with the JDE.\n\t  See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jdebug-ug\/jdebug-ug.html\" target=\"jdebug-ug\">JDEbug\n\t    User's Guide<\/a> for more information<\/p>.\n      <\/li><\/ul><\/div><p>JDEbug provides more debugging features but requires a JDK 1.2\n      or higher vm. You must use <code>jdb<\/code> to debug applications\n      running on earlier vms.<\/p><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Specifying\n      Classpaths<\/h2><\/div><p>The JDE provides customization variables that permit you to\n      specify a classpath for the JDE's compile, run, debug,\n      Beanshell, and other commands. The following table lists these\n      variables.<\/p><div class=\"table\"><p><b>Table 1. Classpath Variables<\/b><\/p><table summary=\"Classpath Variables\" border=\"1\"><colgroup><col \/><col \/><\/colgroup><thead><tr><th>Variable<\/th><th>Applies To ...<\/th><\/tr><\/thead><tbody><tr><td><p><code>jde-global-classpath<\/code><\/p><\/td><td><p>All JDE commands that use a\n\t\tclasspath.<\/p><\/td><\/tr><tr><td><p><code>jde-compile-option-classpath<\/code><\/p><\/td><td><p>Compile command<\/p><\/td><\/tr><tr><td><p><code>jde-run-option-classpath<\/code><\/p><\/td><td><p>Run App command<\/p><\/td><\/tr><tr><td><p><code>jde-db-option-classpath<\/code><\/p><\/td><td><p>Debug App command<\/p><\/td><\/tr><\/tbody><\/table><\/div><p>The <code>jde-global-classpath<\/code> variable is so\n      named because it specifies a classpath that is used by all JDE\n      commands unless overridden by a classpath variable specific to a\n      particular command. For example, the JDE's\n      <code>Compile<\/code> command uses\n      <code>jde-global-classpath<\/code> unless the value of\n      <code>jde-compile-option-classpath<\/code> is set.\n      Typically you want to use the same classpath to compile, run,\n      and debug an application. So typically\n      <code>jde-global-classpath<\/code> is the only variable you\n      need to set.<\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Setting a Classpath\n\tVariable<\/h4><\/div><p>As with other JDE customization variables, you must use\n\tEmacs' <code>customize-variable<\/code>\n\tcommand to set the JDE's classpath variables. You can use\n\trelative paths, cygwin paths, and paths with tilde notation\n\tand environment variables as values of JDE classpath\n\tvariables. See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#SpecifyingPaths\">Specifying\n\t  Paths<\/a> for more information.\n      <\/p><p>\n\tWhen setting a classpath variable, you must enter each path as\n\ta separate entry in the customization buffer. Do not enter the\n\tpaths as semicolon or colon-separated lists of paths. The\n\tfollowing examples show the right and wrong way to specify a\n\tclasspath.<\/p><div class=\"example\"><p><b>Example 1. Wrong Way to Specify a Classpath<\/b><\/p><pre class=\"programlisting\">\n          jde-global-classpath: [Hide] \n          [INS] [DEL] Path: c:\/jde-dev\/jmath\/classes;c:\/jde-dev\/jmath\/src \n          [INS] \n            [State]: this option has been set and saved.\n\t<\/pre><\/div><div class=\"example\"><p><b>Example 2. Right Way to Specify a Classpath<\/b><\/p><pre class=\"programlisting\">\n          jde-global-classpath: [Hide] \n          [INS] [DEL] Path: c:\/jde-dev\/jmath\/classes \n          [INS] [DEL] Path: c:\/jde-dev\/jmath\/src \n          [INS] \n            [State]: this option has been set and saved.\n\t<\/pre><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Including Class\n\tLibraries<\/h4><\/div><p>\n\tIf you store your class libraries as jar or zip files in\n\tseparate directory, you need specify only the path of the\n\tdirectory in a JDE classpath variable. Whenever the JDE\n\tencounters a library directory in a classpath variable, it\n\texpands the classpath to include all the jar and zip files in\n\tthe library directory.\n      <\/p><p>\n\tFor example, suppose you organize your project directory as\n\tfollows:\n      <\/p><pre class=\"programlisting\"> myproj classes lib\n\t  src <\/pre><p>\n\twhere the <code>src<\/code> directory contains all of\n\tyour source files, the <code>lib<\/code> directory all\n\tof your jar and zip files,  and the\n\t<code>classes<\/code> directory all of the classes\n\tcompiled from the <code>src<\/code> directory. Further,\n\tsuppose you store your project file (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#UsingProjectFiles\">Using Project Files<\/a>) at the\n\troot of your project directory. Then, you can specify\n\t<code>jde-global-classpath<\/code> simply as follows.\n      <\/p><pre class=\"programlisting\"> jde-global-classpath:\n\t  [Hide] [INS] [DEL] Path: .\/classes [INS] [DEL] Path: .\/src\n\t  [INS] [DEL] Path: .\/lib [INS] [State]: this option has been\n\t  set and saved. <\/pre><p>\n\tIf you do not want the JDE to expand library directories, set\n\tthe variable <code>jde-expand-classpath-p<\/code> off.\n\tThe JDE expands directories whose names match the regular\n\texpressions specified by\n\t<code>jde-lib-directory-names<\/code>. The default values\n\tof this variable are <code>^lib<\/code> and\n\t<code>^jar<\/code>. So by default, the JDE expands\n\tnames that begin with the string <code>lib<\/code> or\n\t<code>lib<\/code>, e.g., <code>lib<\/code>,\n\t<code>libraries<\/code>, etc. However, you can\n\tcustomize <code>jde-lib-directory-names<\/code> to\n\treflect any library naming scheme you prefer.\n      <\/p><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Using Project Files<\/h2><\/div><p>A project file is a Lisp file that the JDEE loads and\n      evaluates whenever you open a Java source file belonging to a\n      specific project. The JDEE has commands that save the current\n      settings of JDEE project variables in the project file for the\n      current project. Project files thus allow you to save and\n      restore project-specific settings of JDEE customization\n      variables. For example, you can use a project file to set the\n      value of the <code>jde-global-classpath<\/code> variable to\n      a project-specific classpath automatically whenever you load a\n      file belonging to that project. <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">How the JDE Finds Project Files<\/h4><\/div><p>To simplify the task of finding project files, the JDE makes\n\ttwo assumptions. First, it assumes that all Java source files\n\trelating to a particular project live in a single directory tree.\n\tSecond, it assumes that all project files have the same file name.\n\tThe name assumed by default is <code>prj.el<\/code>. You can use the JDE\n\tconfiguration variable <code>jde-project-file-name<\/code> to specify\n\tanother name.<\/p><p>\n\tWhen you open a Java source file, the JDE looks for\n\tproject files in the directory tree containing the source file.  \n\tIf the JDE finds one or more project files, it loads the\n        project files in the following manner.\n      <\/p><p>       \n\tThe JDE  first sets all the JDE variables to their\n\tEmacs startup values (i.e., the default value or the value\n\tsaved in your <code>.emacs<\/code> file). It then loads\n\tall the project files in the directory tree containing the\n\tcurrent source buffer, starting with the topmost file.\n      <\/p><p>\n\tWhat this means is that you can use project files to extend\n\tand\/or override the settings in your\n\t<code>.emacs<\/code> file and in other project files.\n\tFor example, your <code>.emacs<\/code> file can specify\n\tsettings that are common to all your projects. You can put\n\tsettings common to a group of projects at the top of the\n\tdirectory tree containing the projects, settings common to\n\teach project at the top of the directory containing each\n\tprojects, and so on.\n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Project File\n\tContents<\/h4><\/div><p>A project file can contain any Lisp code that can be\n\tevaluated correctly as the last step in creating a Java source\n\tbuffer. For example, a project file can contain code that sets\n\tthe value of JDE configuration variables. In general, a\n\tproject file can  contain any code that could be implemented\n\tas a <code>jde-mode<\/code> hook function. In fact, you\n\tcan think of a project file as a project-specific\n\t<code>jde-mode<\/code> hook function. <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Creating a Project\n\tFile<\/h4><\/div><p>The easiest way to create a project file is to use the\n\t<code>jde-create-new-project <\/code> command (<b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Project\n\t    File<\/b>-&gt;<b>Create\n\t    New<\/b>). This command creates a \n        new project file and saves the\n\tcurrent values of all customized JDE configuration variables\n\tin the project file for the selected Java buffer. (To be\n\tprecise,  the command inserts Lisp code in the project file\n\tthat restores  the current settings of the configuration\n\tvariables; if such code  already exists in the file, it\n\treplaces it with updated code.) Thus, to create\n\ta project file that sets JDE configuration  variables to\n\tproject-specific values:\n      <\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Open any source file belonging to the project.<\/p><\/li><li><p>Set the values of the JDE configuration variables to\n\t    the appropriate values for the project to which the source\n\t    file belongs.<\/p><\/li><li><p>See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ConfiguringJDE\">Configuring the\n\t      JDE<\/a>  for information on how to set the values of\n\t    the JDE  configuration variables.<\/p><\/li><li><p>Select <b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Project\n\t\tFile<\/b>-&gt;<b>Save\n\t\tProject<\/b>.\n\t  <\/p><\/li><li><p> The JDE prompts you to enter the path to a directory\n\t    in which to store the project file.\n\t  <\/p><\/li><li><p>\n\t    Enter the directory where you want to store the project\n\t    file or press <b>Enter<\/b> to store the project\n\t    file in the current directory.\n\t  <\/p><\/li><\/ol><\/div><p>Once you have created the project file, you can insert\n\tadditional configuration code into the file if desired. Note\n\tthat the <code>jde-save-project<\/code> command edits\n\trather than replaces existing project files. Thus, you can use\n\tthe command to update configuration variable settings without\n\tdisturbing any other configuration code that you have inserted\n\tmanually into the project file. \n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Creating Portable Projects<\/h4><\/div><p>\n\tYou can use relative paths and environment variables to \n\tconstruct portable project files. For example, suppose your\n\tproject directory has the following structure:\n      <\/p><pre class=\"programlisting\">      \n    myprj\n      classes\n        pkg1\n        pkg2\n      src\n        pkg1\n        pkg2\n      prj.el\n\t<\/pre><p>      \n\tFurther suppose that your project uses beans  from a \n\tshared library specified by the environment variable\n\t<code>BEANS<\/code>. With this setup, \n\tyou can specify <code>jde-global-classpath<\/code> as\n      <\/p><pre class=\"programlisting\">      \n      .\/src\n      .\/classes\n      $BEANS\/lib\/dials.jar\n\t<\/pre><p>  \n\tand <code>jde-compile-option-directory<\/code> as\n      <\/p><pre class=\"programlisting\">      \n      .\/classes\n\t<\/pre><p>  \n\tThis causes the JDE to store classes compiled from your src\n\tdirectory in the classes directory.  Note that you have not\n\tused any absolute paths in creating your project file. This\n\tmeans you can move your project anywhere without having to update\n\tthe project file.\n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Disabling Context Switching<\/h4><\/div><p>When you select a buffer that belongs to another\n\tproject, the JDE loads the project file for that project\n\tand may update the class list for the project. This \n        causes a slight delay before you can begin using the\n\tnew buffer. The delay can be irritating if you need\n        to switch frequently among open projects, for example,\n\twhen copying or diffing code from different projects.\n\tAccordingly, the JDE provides ways for you to disable\n\tcontext-switching either permanently or  temporarily.\n     <\/p><p>To disable context-switching permanently, toggle\n\tthe customizat variable \n\t<code>jde-project-context-switching-enabled-p<\/code>\n\toff. Be sure to use customize to do this and save the\n\tcustomization in your <code>\n\t  .emacs<\/code> or <code>prj.el<\/code>\n\tfile. Note that you can use the JDE's project file loading\n\tcommand, <code>JDE-&gt;Project-&gt;Project File-&gt;Load<\/code>,\n\tto switch contexts manually.\n      <\/p><p>\n\tTo disable context-switching temporarily during a session,\n\tselect <code>\n\tJDE-&gt;Project-&gt;Auto Switch<\/code> or enter\n\t<code>M-x jde-toggle-project-context-switching<\/code>. To reenable\n\tcontext-switching, execute the same command again.\n      <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Project Hook Functions<\/h4><\/div><p>The variable <code>jde-project-hooks<\/code> lists\n\tfunctions that execute after the JDEE loads a project file. By\n\twriting hook functions and adding them to this list, you can\n\tspecify actions to be taken every time the JDEE switches from\n\tone project to another.<\/p><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\"> Displaying Java Documentation<\/h2><\/div><p>\n      The JDE provides commands for displaying the JDK API documentation\n      and context-sensitive help for classes.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Browsing JDK\n\tDocumentation<\/h4><\/div><p>The <code>jde-browse-jdk-doc<\/code> command\n\t(<b>JDE<\/b>-&gt;<b>Help<\/b>-&gt;<b>JDK<\/b>, <b>C<\/b>-<b>c<\/b>\u00a0\n\t<b>C<\/b>-<b>v<\/b>\u00a0<b>C<\/b>-<b>n<\/b>) opens the\n\tJDK documentation in a browser. By default, this command\n\tdisplays the JDK documentation page at JavaSoft's web site. To\n\tdisplay a different copy, set the variable\n\t<code>jde-jdk-doc-url<\/code> to the url of the index\n\tpage of the copy you want to be displayed. <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Context-Sensitive Class\n\tHelp<\/h4><\/div><p>The JDK provides context-sensitive help for Java classes.\n\tTo use this facility, you must first customize the variable\n\t<code>jde-help-docsets<\/code> to specify the location of\n\tclass documentation on your system. The JDE class help\n\tfacility supports javadoc documentation by default but it\n\tworks with other types of documentation as well.<\/p><p>Once you have specified the location of class\n\tdocumentation on your system, you can get help for the class\n\tat point in the current buffer by selecting <b>JDE<\/b>-&gt;<b>Help<\/b>-&gt;<b>Symbol \n\t    at Point<\/b> . <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Specifying a Browser<\/h4><\/div><p>The JDEE uses the browsers or browsers specified by the customization \n\tvariable <code>browse-url-browser-function<\/code>. See the documentation\n      for this variable for more information.<\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using emacs-w3m to Browse Java Doc<\/h5><\/div><p>The <code>emacs-w3m<\/code> package uses the\n\t  <code>w3m<\/code> text-mode HTML browser to enable\n\t  Emacs to render text HTML pages. This allows you to use\n\t  Emacs not only to develop Java code but also to browse Java\n\t  API and tools documentation. For information on downloading\n\t  and installing emacs-w3m and w3m, see the <a href=\"http:\/\/emacs-w3m.namazu.org\/\">emacs-w3m<\/a> web\n\t  site. <\/p><p>While you're at it, you should also install the <a href=\"http:\/\/www.emacswiki.org\/cgi-bin\/wiki\/WThreeM\">\n\t    WThreeM<\/a> package. This package provides \n\tincremental searching on URLs in w3m HTML buffers, a nice feature\n\tfor browsing Javadoc class documentation, with its wealth of\n\tlinks.<\/p><p>One you have installed <code>emacs-w3m\/w3m<\/code>, you can\n\tconfigure Emacs (and hence the JDEE) to use it to display Javadoc pages \n\t  by customizing <code>jde-help-browser-function<\/code> to specify\n\t  <code>w3m-browse-url<\/code>. <\/p><p>The <code>w3m-browse-url<\/code> function displays its output\n\tby default in the current window, replacing the Java file in the buffer. \n\tTo cause the function to display its output in a popup window, customize\n\t  <code>w3m-pop-up-frames<\/code>.<\/p><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Browsing Source Code<\/h2><\/div><p>The JDE provides the following facilities for browsing Java source code:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#ClassesMenu\">Classes index menu<\/a><\/p><\/li><li><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#Speedbar\">Speedbar<\/a><\/p><\/li><li><p><a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#Tags\">Etags<\/a><\/p><\/li><\/ul><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Classes Index Menu<\/h4><\/div><p>The <b>Classes<\/b>\n\tindex menu appears by default in the Emacs menubar whenever a\n\tJava source buffer is active. The menu consists of a cascading\n\tlist of all classes, methods, fields, and imports defined in\n\tthe current buffer plus the package to which the current\n\tbuffer belongs. Selecting any item scrolls the buffer to the\n\tstatement that defines the item.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/classes_menu1.gif\" alt=\"Screenshot showing Classes menu.\" \/><\/div><\/div><p>The top-level menu contains an entry for each each class\n\tdefined by the active buffer followed by entries for the\n\timports and package of the current buffer. Selecting the entry\n\tfor a class displays a submenu listing the inner classes,\n\tmethods, and fields defined by the class. Selecting the entry\n\tfor an inner class displays another submenu for that class,\n\tand so on. Selecting the\n\t<b>imports<\/b> entry on\n\tthe top-level menu displays a submenu listing all the classes\n\tand packages imported by the active buffer.<\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> Special Index\n\t  Entries<\/h5><\/div><p>The index menu includes the following special index\n\t  entries:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><b>*Rescan*<\/b><\/p><p>Selecting this item causes the JDE to rebuild the\n\t      index menu.  You should rebuild the menu whenever you\n\t      edit the buffer.\n\t    <\/p><\/li><li><p><b>*class\n\t\t  def*<\/b><\/p><p>Selecting this item takes you to the start of the\n\t      definition of  the class on whose submenu it appears.\n\t      Turn the variable\n\t      <code>jde-imenu-include-classdef<\/code> off to\n\t      suppress inclusion  of these items in the menu. <\/p><\/li><\/ul><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Alphabetizing the\n\t  Classes Menu<\/h5><\/div><p>By default the\n\t  <b>Classes<\/b> menu\n\t  lists methods and fields in the order in which the active\n\t  buffer defines them. Use the variable\n\t  <code>jde-imenu-sort<\/code> to customize the menu to\n\t  list methods and fields alphabetically in either ascending\n\t  or descending order.<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Suppressing Method\n\t  Signatures and Field Types<\/h5><\/div><p>By default the\n\t  <b>Classes<\/b> menu\n\t  displays the signatures of methods and the types of fields\n\t  displayed in the active buffer. The method signatures enable\n\t  you to distinguish overloaded method names. The menu can\n\t  also display the names of methods and fields without\n\t  signatures and types. Use the variable\n\t  <code>jde-imenu-include-signature<\/code> to turn\n\t  signature and type display off or on. <\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Disabling the\n\t  Classes Menu<\/h5><\/div><p>Set the variable <code>jde-imenu-enable<\/code> off\n\t  to disable the\n\t  <b>Classes<\/b> menu.\n\t  When disabled, the menu does not appear in the Emacs\n\t  menubar.<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using the Keyboard to Navigate the Classes Menu<\/h5><\/div><p>\n\t  Execute <code>M-x imenu<\/code> to use the keyboard \n\t  to navigate the Classes menu. Emacs displays the \n\t  top-level items on the Classes menu in a buffer and\n\t  prompts you to enter the name of an an item in the\n\t  minibuffer. Enter the name of the item to select it.\n\t  Note that you can use completion to avoid having to\n\t  type the whole name. If you select a submenu, Emacs\n\t  displays the contents of the submenu and prompts\n\t  you to enter a selection in the keyboard. If you\n\t  select an index entry, Emacs scrolls the buffer\n\t  to the indexed point in the buffer.\n\t<\/p><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Using the Speedbar<\/h4><\/div><p>To display the speedbar, select <b>JDE<\/b>-&gt;<b> Speedbar<\/b>. The\n\tspeedbar opens in a separate frame.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/speedbar1.gif\" alt=\"Screenshot showing the speedbar.\" \/><\/div><\/div><p>The speedbar displays a list of the files and\n\tsubdirectories in the directory containing the file displayed\n\tin the current buffer. The speedbar highlights the file\n\tdisplayed in the current buffer.<\/p><p>Click on the expand (+) button in front of any file. The\n\tnode for the  file expands<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/speedbar2.gif\" alt=\"Screenshot showing a speedbar entry.\" \/><\/div><\/div><p>to show up to three entries, depending on the contents of\n\tthe source file. <\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Package<\/h5><\/div><p>\n\t  This item expands to show the package to which the file\n\t  belongs.\n\t<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/speedbar3.gif\" alt=\"Screenshot showing a speedbar entry.\" \/><\/div><\/div><p>\n\t  Clicking on the package scrolls the buffer to the\n\t  corresponding package declaration.\n\t<\/p><\/div><div class=\"sect4\"><p>\n\t  This item expands to show the classes in the selected file.\n\t<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/speedbar4.gif\" alt=\"Screenshot showing a speedbar entry.\" \/><\/div><\/div><p>\n\t  Each class expands to show the constructors, methods,\n\t  fields, and inner classes defined by the class and the\n\t  class's parent, if any. Inner classes also expand and their\n\t  inner classes, and so on. The constructors and methods\n\t  expand to show arguments and argument types and return\n\t  types. Fields expand to show their type.\n\t<\/p><p>\n\t  Clicking on any class, inner class, method, constructor, or\n\t  field scrolls the buffer to show that item.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Dependencies<\/h5><\/div>This item expands to show the classes and packages imported by the current source file.<div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/speedbar5.gif\" alt=\"Screenshot showing a speedbar entry.\" \/><\/div><\/div><p>Click on class or package to scroll the buffer to the\n\t  corresponding import statement.<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Updating the\n\t  Speedbar<\/h5><\/div><p>If you make changes to a source buffer, you must update\n\t  the speedbar view to reflect the changes. To update the\n\t  speedbar view of a buffer:<\/p><div class=\"procedure\"><ol type=\"1\"><li><p>\n\t      Collapse the speedbar view of the buffer.\n\t    <\/p><p>\n\t      This is necessary only if the speedbar view is expanded.\n\t      To collapse the speedbar view, click the collapse button\n\t      (-) next to the buffer's file name in the speedbar view.\n\t    <\/p><\/li><li><p>Hold the shift key down while clicking the expand\n\t      button (+) next to the buffer's name  in the speedbar\n\t      view.  <\/p><\/li><\/ol><\/div><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Searching Source Code<\/h2><\/div><p>The JDEE provides commands that enable you to find<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p>\n\t  All occurences of strings matching a regular expression anywhere\n\t  in your project's source path (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#FindExpressions\">Finding\n\t  Expressions<\/a>)\n\t<\/p><\/li><li><p>\n\t  Source code that defines the method, field,\n\t  or other symbol at point (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#FindSymbolDef\">Finding\n\t  Symbol Definitions<\/a>)\n\t<\/p><\/li><li><p>Source code for the class at point (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#FindClasses\">Finding\n\t  Classes<\/a>)<\/p><\/li><li><p>Source for the parent of the class at point<\/p><\/li><li><p>Source for an interface implemented by the class at point<\/p><\/li><li><p>All methods that invoke a specifed method (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#Xref\">Cross-Referencing \n\t    Classes<\/a>)<\/p><\/li><\/ul><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Finding Expressions<\/h4><\/div><p>The JDEE provides two commands for finding occurrences of strings\n      in Java source and other types of text files:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>JDE-&gt;Find-&gt;Expression<\/code> \n\t    (<code>jde-find<\/code>)<\/p><p>\n\t    This command uses the minibuffer to prompt you for the\n\t    search expression and the paths to be searched. You can\n\t    use customization variables to specify values for the\n\t    search paths and other search options. The default values\n\t    for the customization variables search the Java source \n            files on the classpaths and sourcepaths for your project.\n\t    See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#FindExpressionMinibuf\">Using the Minibuffer-Based\n\t    Find Command<\/a> for more information.\n\t  <\/p><\/li><li><p><code>JDE-&gt;Find-&gt;Expression...<\/code> \n\t    (<code>jde-find-dlg<\/code>)<\/p><p>This command uses a text dialog buffer to\n\t    prompt you for search arguments. The dialog buffer\n\t    allows you to specify all search options\n\t    interactively. It is thus useful for performing\n\t    searches that have special requirements.\n\t    See <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#FindExpressionDlg\">Using the Dialog-Based\n\t    Find Command<\/a> for more information. <\/p><\/li><\/ul><\/div><p>\n\tBoth commands use the Unix <code>grep<\/code> and\n\t<code>find<\/code> utilities to perform the searches that\n\tyou specify. Before executing a search, both commands search\n\tyour system for copies of <code>grep<\/code> and\n\t<code>find<\/code> and signal an error if the utilities\n\tare missing. All Unix systems include copies of\n\t<code>grep<\/code> and <code>find<\/code>. Versions\n\tof these utilities are available for Microsoft Windows, e.g.,\n\tas part of the Cygwin Unix emulation package for Microsoft\n\tWindows. If you want to use the JDEE's <code>find<\/code>\n\tcommands on Windows, you must install copies of\n\t<code>grep<\/code> and <code>find<\/code> on your\n\tsystem and either include them in your system's command path\n\tor use the variables <code>grep-command<\/code> and\n\t<code>find-program<\/code> to specify their respective\n\tlocations on your system.\n      <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p> \n\t  Microsoft Windows\/NT and Windows 2000 provide a\n\t  <code>find<\/code> utility that is incompatible with\n\t  the Unix <code>find<\/code> utility. To ensure that the\n\t  JDEE finds the right version, you should either put the\n\t  Unix-style <code>find<\/code> command ahead of the\n\t  Windows <code>find<\/code> command in your system\n\t  command path or use the <code>find-program<\/code>\n\t  variable to specify the Unix-style command's location.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using the\n\t  Minibuffer-Based Find (jde-find) Command<\/h5><\/div><p>To find an expression with the minibuffer-based find\n\t  command:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>\n\t      Select \n\t      <b>Find<\/b>-&gt;<b>Expression<\/b> (<b>C-c-C-v-C-f<\/b>)\n\t      from the <b>JDE<\/b> menu\n\t      or execute the <code>jde-find<\/code> command.\n\t    <\/p><p>The JDEE prompts you to enter a regular expression.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/find_minibuf_prompt_regex.gif\" alt=\"Screenshot showing the regular expression prompt.\" \/><\/div><\/div><\/li><li><p>Enter a regular expression that matches the\n\t      expressions you want to find and press <b>Enter<\/b>.<\/p><p>The JDEE prompts you to enter a list of directories to search.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/find_minibuf_prompt_dirs.gif\" alt=\"Screenshot showing the directories prompt.\" \/><\/div><\/div><p>The prompt may list a default search path. The JDEE determines the\n\t      default path by testing each of the following variables in the order\n\t      listed.<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>jde-sourcepath<\/code><\/p><\/li><li><p><code>jde-compile-option-sourcepath<\/code><\/p><\/li><li><p><code>jde-compile-option-classpath<\/code><\/p><\/li><li><p><code>jde-global-classpath<\/code><\/p><\/li><\/ul><\/div><p>The JDEE uses the directories\n\t      specified by the first path variable that has a nonnil\n\t      value as the default search path.<\/p><\/li><li><p>Edit the default search path, if desired, in the minibuffer\n\t      and press <b>Enter<\/b>.<\/p><p>The JDEE executes the find command that you have specified\n\t    and displays the command's output in a popup grep-mode buffer.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/find_minibuf_result.gif\" alt=\"Screenshot showing the search results.\" \/><\/div><\/div><\/li><\/ol><\/div><p>The search results buffer lists each instance of a string\n\tthat matches the specified regular expression in the specified\n\tsearch path. For each match, the listing shows the file\n\tand line number of the match, highlighted in red, and a snippet\n\tof the text in which the match occurred.\n\t<\/p><p>To visit the file containing the match, click the\n\tmatch message in the grep buffer with the middle mouse button\n\tor move point to the message and press <b>Enter<\/b><\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Customizing the\n\t  Minibuffer-Based Find Command<\/h5><\/div><p>\n\t  The following variables allow you to customize the \n\t  minibuffer-based search command:\n\t<\/p><div class=\"table\"><p><b>Table 1. <\/b><\/p><table border=\"1\"><colgroup><col \/><col \/><col \/><\/colgroup><thead><tr><th>Variable<\/th><th>Default Value<\/th><th>Description<\/th><\/tr><\/thead><tbody><tr><td><code>jde-find-case-sensitive<\/code><\/td><td><code>nil<\/code><\/td><td>Specifies whether the <code>jde-find<\/code> command\n\t\t  performs a case-sensitive search. If non-nil, the\n\t\t  search is case-sensitive; otherwise the search\n\t\t  ignores case.<\/td><\/tr><tr><td><code>jde-find-granularity<\/code><\/td><td><code>Character<\/code><\/td><td>Specifies the granularity of the expression\n\t\t  search conducted by <code>jde-find<\/code>: <code>Character<\/code>\n\t\t  (expression starting on any character), <code>Word<\/code> (match\n\t\t  words only), <code>Line<\/code> (match lines only).<\/td><\/tr><tr><td><code>jde-find-file-regexp<\/code><\/td><td><code>*.java<\/code><\/td><td>Specifies the regular expression that the\n\t\t  <code>jde-find<\/code> command uses to select\n\t\t  files to be searched. You can use any regular\n\t\t  expression supported by the <code>-name<\/code>\n\t\t  option of the GNU <code>find<\/code>\n\t\t  command.<\/td><\/tr><\/tbody><\/table><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using the Dialog-Based Find (jde-find-dlg) Command<\/h5><\/div><p>To find an expression with the dialog-based find command:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>\t    \n\t      Select \n\t      <b>Find<\/b>-&gt;<b>Expression...<\/b>\n\t      from the <b>JDE<\/b> menu\n\t      or execute the <code>jde-find<\/code> command.\n\t    <\/p><p>The Find Expression Options buffer appears.<\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/find_options_buffer.gif\" alt=\"Screenshot showing the Find Expression Options buffer.\" \/><\/div><\/div><\/li><li><p>Edit the buffer to reflect the search options that you desire.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>The options you select will appear the\n\t\tnext time you execute the <code>jde-find-dlg<\/code> command.<\/p><\/div><\/li><li><p>Select the Ok button.<\/p><\/li><\/ol><\/div><p>The <code>jde-find-dlg<\/code> invokes the\n\t  <code>find<\/code> and <code>grep<\/code> commands\n\t  with the options you specified and displays the results in a\n\t  buffer.\n\t<\/p><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Finding Symbol Definitions<\/h4><\/div><p>To find the source code that defines the class, interface,\n\tmethod, or field at point, select \n\t<b>Find<\/b>-&gt;<b>Symbol Definition<\/b> (<b>C-c\u00a0\n\t    C-v\u00a0\n\t    C-y<\/b>) from the <b>JDE<\/b> menu or\n\texecute M-x\n\t<code>jde-open-class-at-point<\/code>. The JDEE\n\tdisplays a buffer containing the source code that defines the\n\tsymbol at point in the original buffer.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>This command requires that the symbol definition be\n\t  located on <code>jde-sourcepath<\/code> and that the\n\t  class referenced at point be located on\n\t  <code>jde-global-classpath<\/code>. The command uses\n\t  the beanshell to determine the fully qualified name of the\n\t  class referenced at point. If the beanshell is not running,\n\t  it launches the beanshell.<\/p><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Finding Classes<\/h4><\/div><p>The <code>jde-open-class-source<\/code> command\n\tfinds and opens the Java source file for the class whose\n\tqualified or unqualified name appears at point.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>The JDE uses Java reflection to determine the fully\n\t  qualified name of the class at point. Therefore, the class\n\t  must be on <code>jde-global-classpath<\/code>.<\/p><\/div><p>If more than one class with the same name exists on\n\t<code>jde-global-classpath<\/code>, this command prompts\n\tyou to select one of the classes. The following conditions\n\tmust be true for this command to work. First, a class file for\n\tthis class must exist on the classpath specified by\n\t<code>jde-global-classpath<\/code>. Secondly, the source\n\tfor the class must exist on one of the paths specified by\n\t<code>jde-sourcepath<\/code>. This command uses the\n\tBeanshell to determine the fully qualified name of the class\n\tto be found. It starts the Beanshell if necessary. Thus, the\n\tcommand may respond somewhat slowly the first time you it in a\n\tsession. <\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Cross-Referencing Classes<\/h4><\/div><p>JDE includes a facility for creating and utilizing a\n\tcross-referencing database to enable you to quickly locate all\n\tthe callers of any particular function.  This functionality is\n\tvery useful for quickly figuring how unfamiliar code works,\n\tand useful for doing certain tasks such as renaming functions.\n\tBe advised that this only finds direct callers, and cannot\n\tdetect calls via Java's reflection mechanism.  The\n\tcross-reference database must be kept in sync with the project\n\tcode, however the database generation is generally quick. The\n\tremainder of this section explains how to configure and use\n\tthe cross-referencer.<\/p><p><\/p><h3 class=\"title\">Note<\/h3>The JDEE's cross-referencing facility was developed by Andrew Hyatt.<div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Configuring the Cross-Referencer<\/h5><\/div><p>JDE's cross-referencing database is built by examining\n\t  class files for a particular project.  To this end, JDE\n\t  needs to know where the built class files for a project\n\t  live.  This is usually a subset of the classpath.\n\t<\/p><p>Then select JDE-&gt;Options-&gt;Project-&gt;General, and go to\n\t  the \"Jde Built Class Path\" option.  Insert one path for each\n\t  place where built classes wind up.  You can add both\n\t  directories and zip files (jar\/bar\/etc).  Any file entered\n\t  will be assumed to be a zip file.<\/p><p>Now that you have the built class path defined, you can\n\t  configure the cross-referencer that uses it.  Type M-x\n\t  jde-xref-customize. The following are a list the variables\n\t  that can be set, and what they do:\n\t<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>jde-xref-db-base-directory<\/code><\/p><p>The cross-referencer creates a database to\n\t      store the cross-reference information. This variable\n\t      defines a directory where that information resides. It\n\t      might, for example, be at the base of your project tree,\n\t      in a directory called xrefdb\n\t    <\/p><\/li><li><p><code>jde-xref-store-prefixes<\/code><\/p><p>To reduce database size,\n\t      it's useful to ignore all information that doesn't refer\n\t      to code you care about. For example, most people don't\n\t      need to know who calls java.lang.System.out.println.  To\n\t      not store this information, insert a package prefix for\n\t      each package you are interested in. Typically, this may\n\t      be just one string, for example \"com.mycompany\", to\n\t      store only code that references com.mycompany\n\t      classes.<\/p><\/li><li><p><code>jde-xref-cache-size<\/code><\/p><p>The caller database can be quite large. However, it\n\t      is useful for efficiency reasons to keep part of it in\n\t      memory. This variable deterines how many packages to\n\t      keep in memory simulatenously. The higher this is, the\n\t      faster things will be, but also the more memory will be\n\t      taken up by the cross-reference information.\n\t    <\/p><\/li><\/ul><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Building the Cross-Reference Database<\/h5><\/div><p>The next step into getting the cross-reference\n\t  functionality working is to build the cross-reference\n\t  database.  The project must be fully built for this to work,\n\t  since it takes cross-reference information directly from the\n\t  built classes.\n\t<\/p><p>To build the database, type <code>M-x jde-xref-make-xref-db<\/code><\/p><p>If all the settings are properly configured, JDE will go\n\t  off and create the cross-reference database.  This can take\n\t  from a few seconds to several minutes depending on the size\n\t  of your project. Large class files may take long to parse,\n\t  so if it seems to be \"stuck\", please give it time.\n\t<\/p><p>When it is done, it will save the database in the base\n\t  directory specified in the jde-xref customization group.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Using the Cross-Reference Database<\/h5><\/div><p>There are several basic operations that use the\n\t  cross-reference database:\n\t<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>M-x jde-xref-first-caller (C-c C-v a)<\/code>\n\t    <\/p><p> If your cursor is in the body of a function, then\n\t      this will go to the first in a sequence of callers to\n\t      this function.  If there are no callers, there will be a\n\t      message saying so.\n\t    <\/p><\/li><li><p><code>M-x jde-xref-next-caller (C-c C-v n)<\/code><\/p><p> After calling <code>jde-xref-first-caller<\/code>, you can go to\n\t      each of the callers in sequence with this function.  The\n\t      first caller function creates a list of callers and goes\n\t      to the first one.  This navigates to the next caller on\n\t      the list.  You can keep calling this until there are no\n\t      more callers, in which case there will be a message\n\t      saying so.\n\t    <\/p><\/li><li><p><code>M-x jde-xref-display-call-tree<\/code>\n\t    <\/p><p>This displays an interactive tree view of the\n\t      function you are in, and it's callers.  Each level down\n\t      will show you who calls the parent function.  In this\n\t      way you can quickly get an overview of the context in\n\t      which the current function is used in.  Each item in the\n\t      tree will take you to it's definition if it is clicked\n\t      on.\n\t    <\/p><\/li><li><p><code>M-x jde-xref-list-uncalled-functions<\/code><\/p><p> Called from a java file, this lists all the\n\t      functions in the file that have no apparent callers.\n\t      It's important to remember that because of Java\n\t      reflection, there is no guarantee that the functions are\n\t      truly uncalled.  It only means they are not called\n\t      directly.\n\t    <\/p><\/li><\/ul><\/div><p>All of these functions have two modes, a strict and\n\t  non-strict mode. In non-strict mode, the default, more\n\t  potential callers are treated as callers, including those\n\t  that call an interface that the function implements, or who\n\t  call a superclass's function you are overriding. For\n\t  example, if you are in Foo.doSomething(), and foo implements\n\t  IFoo, and a caller calls IFoo.doSomething(), then this is\n\t  treated as a caller in non-strict mode.  Also, if you are in\n\t  Foo.doSomething, and Foo's superclass is Bar, and there is a\n\t  Bar.doSomething(), then Bar.doSomething() is treated as a\n\t  caller in non-strict mode.  In strict mode, however, the\n\t  caller must specify the function directly for it to be\n\t  considered a hit.   To run these functions with strict mode,\n\t  type C-u M-x and then the function-name.  So C-u M-x\n\t  jde-xref-list-uncalled-functions will list only functions\n\t  that are uncalled directly.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Updating the Cross-Reference Database<\/h5><\/div><p>As the code changes, the cross-referencer must be kept\n\t  up to date, or else the jde-xref functions will start taking\n\t  you to incorrect places in files, and the callers list\n\t  itself will become incorrect.  As you change the code, you\n\t  can call M-x jde-xref-update after a recompile, and it will\n\t  go through and recompute just the part of the database you\n\t  have changed.\n\t<\/p><p>Another, easier way to update, for those of us who leave\n\t  their emacs running all the time, is to add a line such as\n\t  this in their .emacs\n\t<\/p><pre class=\"programlisting\">\n\t  (run-at-time \"11:00pm\" 86400 'jde-xref-make-xref-db)\n\t<\/pre><\/div><\/div><div class=\"sect3\"><p>To use the <code>etags<\/code> facility, you must\n\tfirst construct a  TAGS file that indexes every symbol in your\n\tsource code. The JDE  package contains two shell scripts that\n\tyou can use to tag your  source code, one for\n\t<code>csh<\/code> shells and the other for\n\t<code>bash<\/code>.  The <code>bash<\/code> version\n\tis called <code>jtags<\/code>; the <code>csh<\/code>\n\tversion, <code>jtags.csh<\/code>. <\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> Tagging Java Source Code<\/h5><\/div><p>To tag your source code, first copy the appropriate shell\n\t  script to a directory in your Emacs path. Then start a shell (<b>M<\/b>-<b>x<\/b>\u00a0\n\t  shell). Change to the top-level directory containing your\n\t  source code and then enter <code>jtags<\/code>. The <code>jtags<\/code>\n\t  script tags every <code>.java<\/code> file in the current directory\n\t  and in all descendants of the current directory, storing the\n\t  result in a file called <code>TAGS<\/code> in the top-level directory.\n\t<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Finding the Definition of a Symbol<\/h5><\/div><p>To find the definition of a symbol, put your cursor anywhere\n\t  in the symbol and enter\n\t  <b>M<\/b>-<b>.<\/b>. Emacs\n\t  responds by locating\n\t  and opening (if necessary) the file containing the definition and\n\t  positioning the point at the definition. (The first time you type\n\t  <b>M<\/b>-<b>.<\/b>, Emacs\n\t  prompts you to load the <code>TAGS<\/code> file.) <\/p><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Customizing the JDE<\/h2><\/div><p>\n      This section describes various ways you can customize the JDE.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Customization Variables<\/h4><\/div><p>\n\tJDE customization variables allow you to specify compile, run, debug,\n\tand other JDE options. <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p> The traditional way of customizing Emacs is to use\n\t  <code>setq<\/code> forms to set the values of customization\n\t  variables in the Emacs initialization (<code>.emacs<\/code>)\n\t  file. This method does <i>not<\/i> work for JDE\n\t  customization variables. To set a JDE customization variable, you\n\t  <i>must<\/i> use  the customization buffer for a variable\n\t  (see the Emacs online manual for information on the  customization\n\t  feature). This is because the JDE context switching code resets all\n\t  JDE customization variables to their default or customized (via a\n\t  custom buffer) values whenever you open a Java source file or switch\n\t  projects.\n\t<\/p><\/div><p>  The following section explains  how to use the Emacs\n\tcustomization  feature to set the value of a JDE customization\n\tvariable.<\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> Setting a Customization Variable<\/h5><\/div><p>\n\t  To set a JDE customization variable:\n\t<\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Determine the name of the variable you want to customize.<\/p><p>  Refer to the section of this guide that documents the feature\n\t      you want to customize for the name of the corresponding variable.  Or\n\t    type <b>C<\/b>-<b>h<\/b>\u00a0\n\t    v followed by the JDE group prefix\n\t    (<code>jde-<\/code>) or subgroup prefix (e.g., \n\t    <code>jde-compile-option-<\/code>, \n\t    see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#JDECustomizationGroups\">JDE Customization Groups<\/a>). \n\t    Emacs displays all  variables belonging to the JDE group or subgroup. You \n\t    can then browse this list, using Emacs search, completion,  and\n\t    documentation display command, to find the applicable  variable.\n\t    <\/p><\/li><li><p>Display a customization buffer for the variable. <\/p><p>  If you know the name of the variable, the easiest way to\n\t    display a customization buffer for the variable is to select \n\t    <b>Help<\/b>-&gt;<b>Customize<\/b>-&gt;<b>Specific Option...<\/b> from the\n\t    Emacs  menubar or type <b>M<\/b>-<b>x<\/b>\u00a0<code>jde-customize-variable<\/code>.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p><code>jde-customize-variable<\/code> is a\n\t\tspecial version of the standard Emacs\n\t\t<code>customize-variable<\/code> command that\n\t\tprovides a customization buffer menu item for saving a\n\t\tcustomization in a JDEE project file (see below). You\n\t\tshould use this version of the command if you intend\n\t\tto save customizations in project files. You can also\n\t\tuse it to do any other type of customization that the\n\t\tstandard command supports. This allows you to use the\n\t\tJDEE version of the command for all your\n\t\tcustomizations regardless of whether you intend to\n\t\tsave them in project files. <\/p><\/div><p>\t      \n\t    If you know the  group to which the variable belongs (.e.g., compile\n\t    options), you  can display the customization buffer for the group. This\n\t    is useful when  you want to customize several related variables. See \n\t    <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#JDECustomizationGroups\">JDE Customization Groups<\/a> for \n\t    more information. \n\t  <\/p><\/li><li><p>Edit the value for the variable displayed in the customization\n\t    buffer. <\/p><\/li><li><p>Save the value for the variable in your <code>.emacs<\/code> \n\t    or <code>.prj<\/code> file. \n\t    <\/p><p>If you want the setting to apply to all projects that do not\n\t    have a project file  (see <a href=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/jde-ug-content.html#UsingProjectFiles\">Using Project\n\t      Files<\/a>), you should save the variable in your \n\t    <code>.emacs<\/code> file. To save the variable in your\n\t    <code>.emacs<\/code> file, select  <b>Save\n\t\tfor Future Sessions<\/b> from the\n\t    <b>State<\/b> menu for the\n\t    variable in the customization buffer.\n\t    <\/p><p>  If you want the setting to apply only to the current project, \n\t      select\n\t    <b>Save in JDEE Project File<\/b> from the <b>State<\/b> menu for the variable.\n\t    <\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p> \n\t\tThe <b>Save in JDEE Project File<\/b> menu item appears only if you used the \n\t\t<code>jde-customize-variable<\/code> command to create \n\t\tthe customization buffer.\n\t      <\/p><\/div><p>\n\t      If a project file does not exist for the project, the\n\t      JDEE prompts you to enter a name for the project.\n\t      It then creates a project file for the project with\n\t      the name that you enter.    \n\t    <\/p><\/li><\/ol><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Specifying Paths<\/h5><\/div><p>You can use paths containing environment variables and\n\t  or a tilde (~), cygwin paths, and  relative paths as a value\n\t  of any JDE customization variable that requires a\n\t  path.<\/p><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Environment Variables<\/h6><\/div><p>The JDE accepts paths that contain envirnoment variables, for\n\t    example, <\/p><pre class=\"programlisting\">\n\t      $JDK_HOME\/src\n\t    <\/pre><p>\n\t    The JDE replaces the environment variables with their\n\t    actual values before passing the paths to commands (e.g.,\n\t    javac) that require them. You must use Unix notation\n\t    (i.e., <code>$VARNAME<\/code> or\n\t    <code>${VARNAME}<\/code>) to  specify an environment\n\t    variable in a path even on Windows.\n\t<\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Tilde (~) Notation<\/h6><\/div><p>The JDE accepts paths that begin with a tilde, for example,\n\t    <code>~\/myproj\/classes<\/code>. The JDE replaces the tilde\n\t    with the path to your home directory.<\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Relative Paths<\/h6><\/div><p>A relative path is a path that begins with a period, for \n\t  example, <code>.\/src<\/code>. If \n\t  <code>jde-resolve-relative-paths<\/code> is\n\t  set to a non-nil value (the default), the JDE converts a relative \n\t  path to an absolute path by appending the relative path to the\n\t  path of the project file that set the path variable, or if no such\n\t  file exists, to the path of the current Java source buffer.<\/p><p>\n\t    You can use your <code>.emacs<\/code> file to\n\t    specify default relative paths for projects. For example,\n\t    setting <code>jde-global-classpath<\/code> to\n\t    <code>.\/classes<\/code> in your\n\t    <code>.emacs<\/code> file specifies that the\n\t    default location of class files is in a subdirectory of\n\t    the project file directory named <code>classes<\/code>.\n\t  <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Cygwin Paths<\/h6><\/div><p>You can use cygwin style paths in JDE classpath variables\n\t  on Windows. The JDE converts such paths to DOS paths before\n\t  using them, using a conversion function that you can specify.<\/p><p>The jde-cgywin-path-converter variable allows you\n\t  to choose the path conversion function used by the JDE.<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>\n\t\tjde-cygwin-path-converter-internal<\/code><\/p><p>This is the default path conversion function used by\n\t      the JDE. It converts any paths of the form <code>\/\/[a-z]\/<\/code>\n\t      or <code>\/\/cygdrive\/[a-z]\/<\/code> to the corresponding DOS form. For\n\t      example it converts  <code>\/c\/jde\/java\/classes<\/code> to \n\t      <code>c:\/jde\/java\/classes<\/code>.\t    \n\t      <\/p><\/li><li><p><code>jde-cygwin-path-converter-cygpath<\/code><\/p><p>This function invokes the cygwin\n\t\t<code>cygpath<\/code> utility to perform the\n\t\tpath conversion. The advantage is that the cygwin path\n\t\tutility can recognize and convert Unix style paths\n\t\tthat represent mount points in the cygwin mount table.\n\t\tFor example, suppose that you have mounted\n\t\t<code>d:\/javadev<\/code> as\n\t\t<code>\/javadev<\/code> in the cygwin mount table.\n\t\tThen choosing <code>jde-cygwin-path-converter-cygpath<\/code> \n\t\tallows you to use paths beginning with\n\t\t<code>\/javadev<\/code> in JDE classpaths. The\n\t\tdrawback is that the conversion is slow as it requires\n\t\trunning the cygwin utility for each Unix-style path to\n\t\tbe converted.<\/p><\/li><li><p>A custom conversion function that you supply.<\/p><\/li><\/ul><\/div><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> JDE Customization Groups<\/h5><\/div><p>The JDE defines a top-level customization group, the JDE\n\t  Group, for JDE customization variables. The JDE Group itself\n\t  contains the following subgroups:<\/p><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\"> Compile Option\n\t    Group<\/h6><\/div><p>Specifies compile options corresponding to the\n\t  command-line arguments (e.g., -d) accepted by the JDK\n\t  compiler, javac. When  you execute the JDE compile command,\n\t  the JDE uses the settings of  this group to construct a list\n\t  of command-line arguments that it  passes to the Java\n\t  compiler that the JDE uses to compile Java  source programs.\n\t  The compiler is itself an option that you can  specify (see\n\t  the Project Option Group). You can specify command-line\n\t  arguments directly by setting the value of the\n\t  <code>jde-compile-option-command-line-args<\/code>\n\t  variable. You can display the customization buffer for the\n\t  Compile Option Group by selecting\n\t  <b>JDE<\/b>-&gt;<b>\n\t      Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>Compile<\/b> from  the\n\t  Emacs\u2122 menubar. <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Run Option Group<\/h6><\/div><p>Specifies run-time options corresponding to the\n\t    command-line arguments (for example, -classpath) accepted\n\t    by the JDK virtual machine, java. When you execute the JDE\n\t    Run command,\u00a0 the JDE uses the settings of this group\n\t    to construct a list of command-line arguments that it\n\t    passes to the Java interpreter used by the JDE to run\n\t    applications. This group also contains options for\n\t    specifying non-java command-line arguments (in case you\n\t    want to use a VM that accepts a different set of arguments\n\t    than java does) and for specifying arguments to be passed\n\t    to the application (as opposed to the virtual machine\n\t    interpreting the application.) You can display the\n\t    customization buffer for the Run Option Group by selecting\n\t    <b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>Run<\/b> from  the\n\t    Emacs\u2122 menubar. <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Debug Option Group<\/h6><\/div><p>Specifies run-time options corresponding to the command-line\n\t  arguments (for example, -classpath) accepted by the JDK debugger,\n\t  jdb. jdb accepts the same command-line arguments as java. However,\n\t  the JDE maintains two parallel sets of configuration variables\n\t  for these programs in case you want to use a different set of\n\t  options for running a program in debug mode then you use for\n\t  running it normally. When you execute the JDE Debug command,\n\t  the JDE uses the settings of this group to construct a list of\n\t  command-line arguments that it passes to the Java debugger used\n\t  by the JDE to run applications in debug mode. You can display the\n\t  customization buffer for the Debug Option Group by selecting \n\t  <b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>Debug<\/b>\n\t  from  the Emacs\u2122 menubar. <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\"> Autocode Group<\/h6><\/div><p>Specifies templates used to generate code automatically. <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\"> General Options Group<\/h6><\/div><p>Specify all other JDE options. You can display the\n\t  customization buffer for the General Option Group by selecting \n\t  <b>JDE<\/b>-&gt;<b>Project<\/b>-&gt;<b>Options<\/b>-&gt;<b>General<\/b>\n\t  from  the Emacs\u2122 menubar. <\/p><\/div><\/div><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\"> Customizing jde-mode<\/h4><\/div><p>The JDE defines a major mode, named jde-mode, for editing\n\tJava source files. This mode derives from the standard Emacs\n\tJava source editing mode java-mode. In particular, it inherits\n\tall the functions and customization variables defined by\n\tjava-mode and adds its own customization variables. When you\n\tload a Java source file, Emacs runs a JDE mode initialization\n\tfunction called jde-mode. The jde-mode function in turn calls\n\tthe Java mode intialization function, java-mode. The last\n\tthing that jde-mode does before returning is to call a JDE\n\tmode function, if it exists. You can customize the JDE by\n\tdefining a JDE mode hook function in your .emacs file. The\n\tfollowing is an example of how to do this: <\/p><pre class=\"programlisting\">\n      (defun my-jde-mode-hook ()\n      (message \"my-jde-mode-hook function executed\"))\n       (add-hook 'jde-mode-hook 'my-jde-mode-hook)\n       <\/pre><p>The preceding example defines a JDE mode hook function named\n      my-jde-mode-hook and adds it to the list of JDE mode hook\n      functions, using the Emacs Lisp function add-hook. Now, whenever\n      you open a Java source file, jde-mode invokes the function my-jde-mode-hook.\n      Of course, the hook function defined in this example is trivial.\n      A real-world example might do something more useful, such as\n      setting a custom indentation style. <\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\"> Customizing Key Bindings<\/h5><\/div><p>A key binding establishes an equivalence between a keystroke\n\tor a sequence of keystrokes and an interactive Lisp function. The\n\tkeystroke or keystroke combination causes Emacs to execute the\n\tfunction to which the keystroke (combination) is bound. For\n\texample, the JDE by default binds the keystroke  <b>C<\/b>-<b>c<\/b>\u00a0<b>C<\/b>-<b>v<\/b>\u00a0<b>C<\/b>-<b>c<\/b>\n\tto the function <code>jde-compile<\/code>. You can use the JDE\n\tcustomization facility to change the standard JDE bindings or\n\tcreate bindings for functions that do not have default bindings. <\/p><p>To customize key bindings: <\/p><div class=\"procedure\"><ol type=\"1\"><li><p>Open a buffer on the JDE customization variable <code>jde-key-bindings<\/code>.<\/p><p>You can do this by typing\n\t    <b>M<\/b>-<b>x<\/b>\u00a0<code>customize-variable<\/code>\u00a0\n\t    <code>jde-key-bindings<\/code> or by selecting <b>JDE<\/b>-&gt;<b>Options<\/b>-&gt;<b>Project<\/b>  and searching the\n\t    resulting JDE project customization  buffer for\n\t    <code>jde-key-bindings<\/code>. <\/p><p>The key bindings panel looks like this: <\/p><div class=\"screenshot\"><div class=\"mediaobject\"><img src=\"http:\/\/jdee.sourceforge.net\/jdedoc\/html\/jde-ug\/images\/keybindings.gif\" alt=\"Screenshot showing key bindings buffer.\" \/><\/div><\/div><\/li><li><p>Edit the buffer to specify altered or new bindings. <\/p><p>For example, to add a binding, right click the\n\t    [INS] button, then enter the key  stroke in the\n\t    Key field and  the interactive function (command)\n\t    to which it is bound  in the Command field.<\/p><\/li><li><p>Save the edited bindings. <\/p><p>To save the edited bindings, right click the\n\t    [State] button and choose <b>Set for current session<\/b> if you\n\t    want the settings to apply only to the current project or\n\t    <b>Save for future\n\t\tsessions<\/b>if  you want the settings to apply\n\t    to all projects. In either  case, you should save the new settings in\n\t    you project  file if your project has a project file. To save the new\n\t    settings in your project file, switch to a source buffer  and choose\n\t    <b>JDE<\/b>-&gt;<b> Options<\/b>-&gt;<b>Save Project<\/b>  from the Emacs menubar. \n\t  <\/p><\/li><\/ol><\/div><\/div><\/div><\/div><div class=\"chapter\"><div class=\"titlepage\"><p><\/p><h2 class=\"title\">Creating and Installing Plugins<\/h2><\/div><p>Plugins are independently developed applications intended\n\tfor use with the JDEE and distributed in the form of a (g)zip or \n\tjar file. The following sections explain how to create and install\n\tplugins.\n    <\/p><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Installing a Plugin<\/h4><\/div><p>To install a plugin:<\/p><div class=\"orderedlist\"><ol type=\"1\"><li><p>Copy the zip or jar file containing the plugin in to\n\t    the JDEE's plugins directory.<\/p><div class=\"note\"><h3 class=\"title\">Note<\/h3><p>By default the JDEE plugin directory is a\n\t      subdirectory of the JDEE directory named\n\t      <code>plugins<\/code>. If you want to keep your\n\t      plugins in another directory, set\n\t      <code>jde-plugins-directory<\/code> to the path of\n\t      that directory.<\/p><\/div><\/li><li><p>Execute <code>M-h jde-pi-install-plugins<\/code>.<\/p><p>This command installs any plugins at the top level of the plugins\n\t    directory in the plugins directory.<\/p><\/li><\/ol><\/div><p>If the plugin adds any menu items to the JDEE's plugins menu\n\t(named <b>JDEpi<\/b>), you should now be able to see the \n\t  items on the plugins menu.<\/p><\/div><div class=\"sect3\"><div class=\"titlepage\"><p><\/p><h4 class=\"title\">Creating a Plugin<\/h4><\/div><p>The following sections explain how to create a JDEE plugin.<\/p><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Plugin Requirements<\/h5><\/div><p>A JDEE plugin must meet the following requirements:<\/p><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Packaging<\/h6><\/div><p>The plugin must be distributed as a compressed file that can be \n\t    decompressed by Java's jar program. Compressed file formats that meet this\n\t    requirement include zip, gzip, and jar.\n\t    <\/p><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Directory Structure<\/h6><\/div><p>Decompressing the plugin should create a subdirectory in the\n\t    current directory whose name is the plugin's name and that contains \n\t    at least one directory named <code>lisp<\/code>. The\n\t    plugin directory can contain any number of other directories.\n\t    The other directories can have any names. However, it would be\n\t    nice if plugins standardized on the following directory structure\n\t    and names.\n\t  <\/p><pre class=\"programlisting\">\n   plugin\n     bsh\n      doc\n      scripts\n     help\n       info\n       html\n       src      (e.g., xml files used to generate info and html)\n     java\n       class\n       doc\n       lib      (e.g, jar files)\n       src\n     lisp\n\t  <\/pre><\/div><div class=\"sect5\"><div class=\"titlepage\"><p><\/p><h6 class=\"title\">Bootstrap File<\/h6><\/div><p>The plugin's lisp directory must contain a Lisp file\n\t    named <code>jde-PLUGIN.el <\/code>where PLUGIN is\n\t    the same as the plugin's toplevel directory name. This\n\t    file is called a bootstrap file because it is the only\n\t    file directly loaded by the JDEE. It is responsible for\n\t    registering the plugin with the JDEE and loading any other\n\t    Lisp files required by the plugin.<\/p><\/div><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Registering a Plugin<\/h5><\/div><p>A plugin's bootstrap file must register the plugin with\n\t  the JDEE when it is loaded. The bootstrap file should do\n\t  this by invoking the JDEE's\n\t  <code>jde-pi-register<\/code> function. This function\n\t  takes a single argument: an instance of\n\t  <code>jde-plugin<\/code> class that optionally\n\t  specifies the plugin's menu and its contribution to the\n\t  classpath of the instance of the BeanShell run by the\n\t  JDEE. <\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Running Java Code from a Plugin<\/h5><\/div><p>A plugin can use the JDEE's instance of the BeanShell to\n\tinvoke Java methods and display their output in an Emacs\n\tbuffer. The JDEE provides the following functions for this\n\tpurpose:<\/p><div class=\"itemizedlist\"><ul type=\"disc\"><li><p><code>jde-jeval<\/code><\/p><p>Evaluate a Java statement in the JDEE's BeanShell\n\t      instance and return the resulting BeanShell\n\t      output.<\/p><\/li><li><p><code>jde-jeval-r<\/code><\/p><p>Evaluate a Java statement in the JDEE's BeanShell\n\t      instance, use the Emacs Lisp interpreter to evaluate the\n\t      resulting BeanShell output, and return the result. This\n\t      function assumes that the Java statement emits Emacs\n\t      Lisp code, i.e., that the Java code was developed\n\t      specifically to cause Emacs to do something.\n\t    <\/p><\/li><li><p><code>jde-jeval-cm<\/code><\/p><p>Evaluate a Java statement in the JDEE's BeanShell\n\t      instance and display the resulting BeanShell output in\n\t      an Emacs compilation buffer. This function is intended\n\t      to run Java applications, such as style checkers, that\n\t      output error messages keyed to specific lines in a\n\t      source file.\n\t    <\/p><\/li><\/ul><\/div><p>See the doc strings for these functions for more information.<\/p><\/div><div class=\"sect4\"><div class=\"titlepage\"><p><\/p><h5 class=\"title\">Bootstrap File Example<\/h5><\/div><p>\n\t  The following is an example of the bootstrap file\n\tfor a plugin named pi1.<\/p><pre class=\"programlisting\">\n  ;; jde-pi1.el\n  (defun jde-pi1-cmd ()\n    (interactive)\n    (message \"plugin 1\"))\n  (jde-pi-register\n   (jde-plugin \n    \"pi1\"\n    :bsh-cp (list\n\t   (expand-file-name \"lib\/lib1.jar\" (jde-pi-get-plugin-dir \"pi1\"))\n\t   (expand-file-name \"lib\/lib2.jar\" (jde-pi-get-plugin-dir \"pi1\")))\n    :menu-spec (list (list \"Pi1\" [\"cmd\" jde-pi1-cmd :active t]))))\n  (provide 'jde-pi1)\n\t<\/pre><\/div><\/div><\/div>","created_at":"2018-11-07T16:40:00+0000","updated_at":"2018-11-07T16:40:00+0000","annotations":[],"mimetype":"text\/html","reading_time":95,"domain_name":"jdee.sourceforge.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":228,"title":"No title found","url":"https:\/\/powerdns.org\/hello-dns\/","content":"<p>\n                            **A warm welcome to DNS**\n# Hello, and welcome to DNS!\nThis series of documents attempts to provide a correct introduction to the\nDomain Name System as of 2018.  The original RFCs remain the authoritative\nsource of normative text, but this document tries to make this venerable\nprotocol more accessible, while maintaining full alignment with all relevant\nand useful RFCs.\nThis effort is developed cooperatively on GitHub, the repository can be\nfound [https:\/\/github.com\/ahupowerdns\/hello-dns\/](here) and help is highly\nwelcome!  Feedback can also be sent to bert.hubert@powerdns.com or\n[@PowerDNS_Bert](https:\/\/twitter.com\/PowerDNS_Bert).\nContributors so far include: Micha\u0142 K\u0119pie\u0144, Jan-Piet Mens, Andrew Babichev,\nJacob Hoffman-Andrews, Peter van Dijk, Nathan Froyd, Gene McCulley,\nCharles-Henri Bruyand, jose nazario, Warren Kumari, Patrick Cloke, and\nAndrew Tunnell-Jones.  Thanks!\nAlthough we start from relatively basic principles, the reader is expected\nto know what IP addresses are, what a (stub) resolver is and what an\nauthoritative server is supposed to do.  When in doubt: authoritative\nservers 'host' DNS data, 'resolvers' look up things over at authoritative\nservers and clients run 'stub resolvers' to look things up over at\nresolvers.  This document is aimed at developers, but may also be of aid for\nadministrators.\nDNS was originally written down in August 1979 in '[IEN\n116](https:\/\/www.rfc-editor.org\/ien\/ien116.txt)', part of a parallel\nseries of documents describing the Internet.  IEN 116-era DNS is not\ncompatible with today's DNS.  In 1983, RFC 882 and 883 were released,\ndescribing a version of the DNS very similar but not quite interoperable\nwith the one we have today.\nDNS attained its modern form in 1987 when RFC 1034 and 1035 were published.\nAlthough much of 1034\/1035 remains valid, these standards are not that easy\nto read because they were written in a very different time. There are 100s\nof pages of updates that can only be found in later documents.\nThe main goal of this effort is not to contradict the DNS RFCs but to\nprovide an easier entrypoint into understanding the Domain Name System.\nIf you will, the goal is to be a mini \"[TCP\/IP\nIllustrated](https:\/\/en.wikipedia.org\/wiki\/TCP\/IP_Illustrated)\" of DNS. For\nmore about the philosophy of these documents, and how to contribute, please read\n[meta.md](meta.md.html).\nYour help &amp; insights are highly welcome!\nI want to thank \u00d3lafur Gu\u00f0mundsson and Job Snijders for their input and\nenthusiasm for improving the state of DNS.\n## Layout\nThe content is spread out over several documents:\n * [The core of DNS](basic.md.html)\n * [Relevant to stub resolvers and applications](stub.md.html)\n * [Relevant to authoritative servers](auth.md.html)\n * [Relevant to resolvers](resolver.md.html)\n * [tdns: a 'from scratch' DNS library](tdns\/README.md.html)\n   * [tauth: a minimal but feature complete authoritative server](tdns\/tauth.md.html)\n   * [tres: a minimal but feature complete DNS resolver](tdns\/tres.md.html)\n   * [C API: a C library for doing DNS queries](tdns\/c-api.md.html)\n * Optional elements: [EDNS, TSIG, Dynamic Updates, DNAME, DNS Cookies](optional.md.html)\n * [Privacy related](privacy.md.html): QName minimization, DNS-over-TLS, DNS-over-HTTPS, EDNS Padding\n * [DNSSEC](dnssec.md.html)\n * [non-IETF standards](non-ietf.md.html): RRL and RPZ\n * [Rare parts of DNS](rare.md.html) - not obsolete, but not frequently encountered in production\nWe start off with a general introduction of DNS basics: what is a resource\nrecord, what is an RRSET, what is a zone, what is a zone-cut, how are packets\nlaid out. This part is required reading for anyone ever wanting to query a\nnameserver or emit a valid response.\nWe then specialize into what applications can expect when they send\nquestions to a resolver, or what a stub resolver can expect.\nThe next part is about what an authoritative server is supposed to do. On\ntop of this, we describe in slightly less detail how a resolver could\noperate. Finally, there is a section on optional elements like EDNS, TSIG,\nDynamic Updates and DNSSEC.\nRFCs, especially earlier ones, tend to describe servers that perform both\nauthoritative and resolver functions. This turns out to make both code and\ntroubleshooting harder. Therefore, in these documents, the authoritative and\ncaching functions are described separately.\nNext up: [DNS Basics](basic.md.html).\n<\/p>","created_at":"2018-11-07T16:39:59+0000","updated_at":"2018-11-07T16:39:59+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"powerdns.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":225,"title":"Links change URL but not rendered component \u00b7 Issue #4975 \u00b7 ReactTraining\/react-router","url":"https:\/\/github.com\/ReactTraining\/react-router\/issues\/4975","content":"Links change URL but not rendered component \u00b7 Issue #4975 \u00b7 ReactTraining\/react-router \u00b7 GitHub\n    \n  \n  \n    \n  \n        \n  <div id=\"ajax-error-message\" class=\"ajax-error-message flash flash-error\">You can\u2019t perform that action at this time.<\/div>\n    \n  \n  \n  <details class=\"details-reset details-overlay details-overlay-dark lh-default text-gray-dark\" open=\"open\"><summary aria-haspopup=\"dialog\" aria-label=\"Close dialog\">\n      \n    <\/summary><\/details>\n<p>Press h to open a hovercard with more details.<\/p>","created_at":"2018-11-07T16:39:58+0000","updated_at":"2018-11-07T16:39:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/7783071?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":226,"title":"Don't Call Yourself A Programmer, And Other Career Advice","url":"https:\/\/www.kalzumeus.com\/2011\/10\/28\/dont-call-yourself-a-programmer\/","content":"<p>If there was one course I could add to every engineering education, it wouldn\u2019t involve compilers or gates or time complexity. \u00a0It would be Realities Of Your Industry 101, because we don\u2019t teach them and this results in lots of unnecessary pain and suffering. \u00a0This post aspires to be README.txt for your career as a young engineer. \u00a0The goal is to make you happy, by filling in the gaps in your education regarding how the \u201creal world\u201d actually works. \u00a0It took me about ten years and a lot of suffering to figure out some of this, starting from \u201cfairly bright engineer with low self-confidence and zero practical knowledge of business.\u201d \u00a0I wouldn\u2019t trust this as the definitive guide, but hopefully it will provide value over what your college Career Center isn\u2019t telling you.<\/p><p><strong>90% of programming jobs are in creating Line of Business software<\/strong>: Economics 101: the price for anything (including you) is a function of the supply of it and demand for it. \u00a0Let\u2019s talk about the demand side first. \u00a0Most software is not sold in boxes, available on the Internet, or downloaded from the App Store. \u00a0Most software is boring one-off applications in corporations, under-girding every imaginable facet of the global economy. \u00a0It tracks expenses, it optimizes shipping costs, it assists the accounting department in preparing projections, it helps design new widgets, it prices insurance policies, it flags orders for manual review by the fraud department, etc etc. \u00a0<em>Software solves business problems<\/em>. \u00a0Software often solves business problems despite being <em>soul-crushingly boring<\/em> and of <em>minimal technical complexity<\/em>. \u00a0For example, consider an internal travel expense reporting form. \u00a0Across a company with 2,000 employees, that might save 5,000 man-hours a year (at an average fully-loaded cost of $50 an hour) versus handling expenses on paper, for a savings of $250,000 a year. \u00a0It <em>does not matter<\/em> to the company that the reporting form is the world\u2019s simplest CRUD app, it only matters that it either saves the company costs or generates additional revenue.<\/p><p>There are companies which create software which actually gets used by customers, which describes almost everything that you probably think of when you think of software. \u00a0It is <em>unlikely that you will work at one<\/em> unless you work towards making this happen. \u00a0Even if you actually work at one, many of the programmers there <em>do not work on customer-facing software<\/em>, either.<\/p><p><strong>Engineers are hired to create business value, not to program things<\/strong>: \u00a0Businesses do things for irrational and political reasons all the time (see below), but in the main they converge on doing things which increase revenue or reduce costs. \u00a0Status in well-run businesses generally is awarded to people who successfully take credit for doing one of these things. \u00a0(That can, but does not necessarily, entail actually doing them.) \u00a0The person who has decided to bring on one more engineer is not doing it because they love having a geek around the room, they are doing it because adding the geek allows them to complete a project (or projects) which will add revenue or decrease costs. \u00a0Producing beautiful software is not a goal. \u00a0Solving complex technical problems is not a goal. \u00a0Writing bug-free code is not a goal. \u00a0Using sexy programming languages is not a goal. \u00a0Add revenue. \u00a0Reduce costs. \u00a0Those are your only goals.<\/p><p>Peter Drucker \u2014 you haven\u2019t heard of him, but he is a prophet among people who sign checks \u2014 came up with the terms Profit Center and Cost Center. \u00a0Profit Centers are the part of an organization that bring in the bacon: partners at law firms, sales at enterprise software companies, \u201cmasters of the universe\u201d on Wall Street, etc etc. \u00a0Cost Centers are, well, everybody else. \u00a0<strong>You really want to be attached to Profit Centers<\/strong> because it will bring you higher wages, more respect, and greater opportunities for everything of value to you. \u00a0It isn\u2019t hard: a bright high schooler, given a paragraph-long description of a business, can usually identify where the Profit Center is. \u00a0If you want to work there, work for that. \u00a0If you can\u2019t, either a) work elsewhere or b) engineer your transfer after joining the company.<\/p><p>Engineers in particular are usually\u00a0<em>very highly paid<\/em> Cost Centers, which sets MBA\u2019s optimization antennae to twitching. \u00a0This is what brings us wonderful ideas like outsourcing, which is \u201cLet\u2019s replace really expensive Cost Centers who do some magic which we kinda need but don\u2019t really care about with less expensive Cost Centers in a lower wage country\u201d. \u00a0(Quick sidenote: You can absolutely ignore outsourcing as a career threat if you read the rest of this guide.) \u00a0Nobody ever outsources Profit Centers. \u00a0Attempting to do so would be the setup for MBA humor. \u00a0It\u2019s like suggesting replacing your source control system with a bunch of copies maintained on floppy disks.<\/p><p><strong>Don\u2019t call yourself a programmer<\/strong>: \u201cProgrammer\u201d sounds like \u201canomalously high-cost peon who types some mumbo-jumbo into some other mumbo-jumbo.\u201d \u00a0If you call yourself a programmer, someone is <em>already<\/em> working on a way to get you fired. \u00a0You know Salesforce, widely perceived among engineers to be a Software as a Services company? \u00a0Their motto and sales point is \u201cNo Software\u201d, which conveys to their actual customers \u201cYou know those programmers you have working on your internal systems? \u00a0If you used Salesforce, you could fire half of them and pocket part of the difference in your bonus.\u201d \u00a0(There\u2019s nothing wrong with this, by the way. \u00a0You\u2019re in the business of unemploying people. \u00a0If you think that is unfair, go back to school and study something that doesn\u2019t matter.)<\/p><p>Instead, describe yourself by what you have accomplished for previously employers vis-a-vis increasing revenues or reducing costs. \u00a0If you have not had the opportunity to do this yet, describe things which suggest you have the ability to increase revenue or reduce costs, or ideas to do so.<\/p><p>There are many varieties of well-paid professionals who sling code but do not describe themselves as slinging code for a living. \u00a0Quants on Wall Street are the first and best-known example: they use computers and math as a lever to make high-consequence decisions better and faster than an unaided human could, and the punchline to those decisions is \u201cour firm make billions of dollars.\u201d \u00a0Successful quants make more in bonuses in a good year than many equivalently talented engineers will earn in a decade or lifetime.<\/p><p>Similarly, even though you might think Google sounds like a programmer-friendly company, there are programmers and then there\u2019s the people who are closely tied to 1% improvements in AdWords click-through rates. \u00a0(Hint: provably worth billions of dollars.) \u00a0I recently stumbled across a web-page from the guy whose professional bio is \u201cwrote the backend billing code that 97% of Google\u2019s revenue passes through.\u201d \u00a0He\u2019s now an angel investor (a polite synonym for \u201crich\u201d).<\/p><p><strong>You are not defined by your chosen software stack<\/strong>: I recently asked via Twitter what young engineers wanted to know about careers. \u00a0Many asked how to know what programming language or stack to study. \u00a0It doesn\u2019t matter. \u00a0There you go.<\/p><p>Do Java programmers make more money than .NET programmers? \u00a0Anyone describing themselves as either a Java programmer or .NET programmer has already lost, because a) they\u2019re a programmer (you\u2019re not, see above) and b) they\u2019re making themselves non-hireable for most programming jobs. \u00a0In the real world, picking up a new language takes a few weeks of effort and after 6 to 12 months nobody will ever notice you haven\u2019t been doing that one for your entire career. \u00a0I did back-end Big Freaking Java Web Application development as recently as March 2010. \u00a0Trust me, nobody cares about that. \u00a0If a Python shop was looking for somebody technical to make them a pile of money, the fact that I\u2019ve never written a line of Python would not get held against me.<\/p><p>Talented engineers are rare \u2014 vastly rarer than opportunities to use them \u2014 and it is a seller\u2019s market for talent right now in almost every facet of the field. \u00a0Everybody at Matasano uses Ruby. \u00a0If you don\u2019t, but are a good engineer, they\u2019ll hire you anyway. \u00a0(A good engineer has a track record of \u2014 repeat after me \u2014 increasing revenue or decreasing costs.) \u00a0Much of Fog Creek uses the Microsoft Stack. \u00a0I can\u2019t even <em>spell<\/em> ASP.NET and they\u2019d still hire me.<\/p><p>There are companies with broken HR policies where lack of a buzzword means you won\u2019t be selected. \u00a0You don\u2019t want to work for them, but if you really do, you can add the relevant buzzword to your resume for the costs of a few nights and weekends, or by controlling technology choices at your current job in such a manner that in advances your career interests. \u00a0Want to get trained on Ruby at a .NET shop? \u00a0Implement a one-off project in Ruby. \u00a0Bam, you are now a professional Ruby programmer \u2014 you coded Ruby and you took money for it. \u00a0(You laugh? \u00a0I did this at a Java shop. \u00a0The one-off Ruby project made the company $30,000. \u00a0My boss was, predictably, quite happy and <em>never even asked what produced the deliverable<\/em>.)<\/p><p><strong>Co-workers and bosses are not usually your friends<\/strong>: You will spend a lot of time with co-workers. \u00a0You may eventually become close friends with some of them, but in general, you will move on in three years and aside from maintaining cordial relations <em>you will not go out of your way to invite them over to dinner<\/em>. \u00a0They will treat you in exactly the same way. \u00a0You should be a good person to everyone you meet \u2014 it is the moral thing to do, and as a sidenote will really help your networking \u2014 but do not be under the delusion that everyone is your friend.<\/p><p>For example, at a job interview, even if you are talking to an affable 28 year old who feels like a slightly older version of you <strong>he is in a transaction<\/strong>. \u00a0You are not his friend, you are an input for an industrial process which he is trying to buy for the company at the lowest price. \u00a0That banter about World of Warcraft is just establishing a professional rapport, but he will (perfectly ethically) attempt to do things that none of your actual friends would ever do, like try to talk you down several thousand dollars in salary or guilt-trip you into spending more time with the company when you could be spending time with your actual friends. \u00a0You will have other coworkers who \u2014 affably and ethically \u2014 will suggest things which go against your interests, from \u201cI should get credit for that project you just did\u201d (probably not phrased in so many words) to \u201cWe should do this thing which advances my professional growth goals rather than yours.\u201d \u00a0Don\u2019t be surprised when this happens.<\/p><p><strong>You radically overestimate the average skill of the competition because of the crowd you hang around with<\/strong>: \u00a0Many people already successfully employed as senior engineers cannot actually implement\u00a0<a href=\"http:\/\/www.codinghorror.com\/blog\/2007\/02\/why-cant-programmers-program.html\">FizzBuzz<\/a>. \u00a0Just read it and weep. \u00a0Key takeaway: you probably\u00a0<em>are<\/em> good enough to work at that company you think you\u2019re not good enough for. \u00a0They hire better mortals, but they still hire mortals.<\/p><p><strong>\u201cRead ad. \u00a0Send in resume. \u00a0Go to job interview. \u00a0Receive offer.\u201d is the exception, not the typical case, for getting employment<\/strong>: Most jobs are never available publicly, just like most worthwhile candidates are not available publicly (see <a href=\"http:\/\/www.joelonsoftware.com\/articles\/FindingGreatDevelopers.html\">here<\/a>). \u00a0Information about the position travels at approximately the speed of beer, sometimes lubricated by email. \u00a0The decisionmaker at a company knows he needs someone. \u00a0He tells his friends and business contacts. \u00a0One of them knows someone \u2014 family, a roommate from college, someone they met at a conference, an ex-colleague, whatever. \u00a0Introductions are made, a meeting happens, and they achieve agreement in principle on the job offer. \u00a0<em>Then<\/em> the resume\/HR department\/formal offer dance comes about.<\/p><p>This is disproportionately true of jobs you actually <em>want<\/em> to get. \u00a0\u201cFirst employee at a successful startup\u201d has a certain cachet for a lot of geeks, and virtually none of those got placed by sending in a cover letter to an HR department, in part because two-man startups don\u2019t have enough scar tissue to form HR departments yet. \u00a0(P.S. You probably don\u2019t want to be first employee for a startup. \u00a0Be the last co-founder instead.) \u00a0Want to get a job at Google? \u00a0They have a formal process for giving you a leg up because a Googler likes you. \u00a0(They also have multiple informal ways for a Googler who likes you an awful lot to short-circuit that process. \u00a0One example: buy the company you work for. \u00a0When you have a couple of billion lying around you have many interesting options for solving problems.)<\/p><p>There are many reasons why most hiring happens privately. \u00a0One is that publicly visible job offers get spammed by hundreds of resumes (particularly in this economy) from people who are stunningly inappropriate for the position. \u00a0The other is that other companies are <em>so<\/em> bad at hiring that, if you don\u2019t have close personal knowledge about the candidate, you might accidentally hire a non-FizzBuzzer.<\/p><p><strong>Networking: it isn\u2019t just for TCP packets<\/strong>: Networking just means a) meeting people who at some point can do things for you (or vice versa) and b) making a favorable impression on them.<\/p><p>There are many places to meet people. \u00a0Events in your industry, such as conferences or academic symposia which get seen by non-academics, are one. \u00a0User groups are another. \u00a0Keep in mind that user groups draw a very different crowd than industry conferences and optimize accordingly.<\/p><p>Strive to help people. \u00a0It is the right thing to do, and people are keenly aware of who have in the past given them or theirs favors. \u00a0If you ever can\u2019t help someone but know someone who can, pass them to the appropriate person with a recommendation. \u00a0If you do this right, two people will be happy with you and favorably disposed to helping you out in the future.<\/p><p>You can meet people over the Internet (oh God, can you), but something in our monkey brains makes in-the-flesh meeting a bigger thing. \u00a0I\u2019ve Internet-met a great many people who I\u2019ve then gone on to meet in real life. \u00a0The physical handshake is a major step up in the relationship, even when Internet-meeting lead to very consequential things like \u201cMade them a lot of money through good advice.\u201d \u00a0Definitely blog and participate on your industry-appropriate watering holes like HN, but make it out to the meetups for it.<\/p><p><strong>Academia is not like the real world<\/strong>: Your GPA largely doesn\u2019t matter (modulo one high profile exception: a multinational advertising firm). \u00a0To the extent that it does matter, it only determines whether your resume gets selected for job interviews. \u00a0If you\u2019re reading the rest of this, you know that your resume isn\u2019t the primary way to get job interviews, so don\u2019t spend huge amount of efforts optimizing something that you either have sufficiently optimized already (since you\u2019ll get the same amount of interviews at 3.96 as you will at 3.8) or that you don\u2019t need at all (since you\u2019ll get job interviews because you\u2019re competent at asking the right people to have coffee with you).<\/p><p>Your major and minor don\u2019t matter. \u00a0Most decisionmakers in industry couldn\u2019t tell the difference between a major in Computer Science and a major in Mathematics if they tried. \u00a0I was once reduced to tears because a minor academic snafu threatened my ability to get a Bachelor of Science with a major in Computer Science, which my advisor told me was more prestigious than a Bachelor of Science in Computer Science. \u00a0Academia cares about distinctions like that. \u00a0The real world does not.<\/p><p>Your professors might understand how the academic job market works (short story: it is ridiculously inefficient in engineering and fubared beyond mortal comprehension in English) but they often have quixotic understandings of how the real world works. \u00a0For example, they may push you to get extra degrees because a) it sounds like a good idea to them and b) they enjoy having research-producing peons who work for ramen. \u00a0Remember, market wages for people capable of producing research are $80~100k+++ in your field. \u00a0That buys an <em>awful lot of ramen<\/em>.<\/p><p>The prof in charge of my research project offered me a spot in his lab, a tuition waiver, and <em>a whole $12,000 dollars<\/em> as a stipend if I would commit 4~6 years to him. \u00a0That\u2019s a great deal if, and only if, you have recently immigrated from a low-wage country and need someone to intervene with the government to get you a visa.<\/p><p>If you really like the atmosphere at universities, that is cool. \u00a0Put a backpack on and you can walk into any building at any university in the United States any time you want. \u00a0Backpacks are a lot cheaper than working in academia. \u00a0 You can lead the life of the mind in industry, too \u2014 and enjoy less politics and better pay. \u00a0You can even get published in journals, if that floats your boat. \u00a0(After you\u2019ve escaped the mind-warping miasma of academia, you might rightfully question whether Published In A Journal is really personally or societally significant as opposed to close approximations like Wrote A Blog Post And Showed It To Smart People.)<\/p><p><strong>How much money do engineers make?<\/strong><\/p><p>Wrong question. \u00a0The right question is \u201cWhat kind of offers do engineers routinely work for?\u201d, because salary is one of many levers that people can use to motivate you. \u00a0The answer to this is, less than helpfully, \u201cOffers are all over the map.\u201d<\/p><p>In general, big companies pay more (money, benefits, etc)\u00a0than startups. \u00a0Engineers with high perceived value make more than those with low perceived value. \u00a0Senior engineers make more than junior engineers. \u00a0People working in high-cost areas make more than people in low-cost areas. \u00a0<strong>People who are skilled in negotiation make more than those who are not<\/strong>.<\/p><p>We have strong cultural training to not ask about salary, ever. \u00a0This is not universal. \u00a0In many cultures, professional contexts are a perfectly appropriate time to discuss money. \u00a0(If you were a middle class Japanese man, you could reasonably be expected to reveal your exact salary to a 2nd date, anyone from your soccer club, or the guy who makes your sushi. \u00a0If you owned a company, you\u2019d probably be cagey about your net worth but you\u2019d talk about employee salaries the way programmers talk about compilers \u2014 quite frequently, without being\u00a0embarrassed.) \u00a0 If I were a Marxist academic or a conspiracy theorist, I might think that this bit of middle class American culture was specifically engineered to be in the interests of employers and against the interests of employees. \u00a0Prior to a discussion of salary at any particular target employer, you should speak to someone who works there in a similar situation and ask about the salary range for the position. \u00a0It is &lt;%= Date.today.year %&gt;; you can find these people online. \u00a0(LinkedIn, Facebook, Twitter, and your (non-graph-database) social networks are all good to lean on.)<\/p><p>Anyhow. \u00a0Engineers are routinely offered a suite of benefits. \u00a0It is worth worrying, in the United States, about health insurance (traditionally, you get it and your employer foots most or all of the costs) and your retirement program, which is some variant of \u201cwe will match contributions to your 401k up to X% of salary.\u201d \u00a0The value of that is easy to calculate: X% of salary. \u00a0(It is free money, so always max out your IRA up to the employer match. \u00a0Put it in index funds and forget about it for 40 years.)<\/p><p>There are other benefits like \u201cfree soda\u201d, \u201ccatered lunches\u201d, \u201cfree programming books\u201d, etc. \u00a0These are social signals more than anything else. \u00a0When I say that I\u2019m going to buy you soda, that says a specific thing about how I run my workplace, who I expect to work for me, and how I expect to treat them. \u00a0(It says \u201cI like to move the behavior of unsophisticated young engineers by making this job seem fun by buying 20 cent cans of soda, saving myself tens of thousands in compensation while simultaneously encouraging them to ruin their health.\u201d \u00a0And I like soda.) \u00a0Read social signals and react appropriately \u2014 someone who signals that, e.g., employee education is worth paying money for might very well be a great company to work for \u2014 but don\u2019t give up huge amounts of compensation in return for perks that you could trivially buy.<\/p><p><strong>How do I become better at negotiation? \u00a0<\/strong>This could be a post in itself. \u00a0Short version:<\/p><p>a) \u00a0Remember you\u2019re selling the solution to a business need (raise revenue or decrease costs) rather than programming skill or your beautiful face.<\/p><p>b) \u00a0Negotiate aggressively with appropriate confidence, like the ethical professional you are. \u00a0It is what your counterparty is probably doing. \u00a0You\u2019re aiming for a mutual beneficial offer, not for saying Yes every time they say something.<\/p><p>c) \u00a0\u201cWhat is your previous salary?\u201d is employer-speak for \u201cPlease give me reasons to pay you less money.\u201d \u00a0Answer appropriately.<\/p><p>d) \u00a0Always have a counteroffer. \u00a0Be comfortable counteroffering around axes you care about other than money. \u00a0If they can\u2019t go higher on salary then talk about vacation instead.<\/p><p>e) \u00a0The only time to ever discuss salary is after you have reached agreement in principle that they will hire you if you can strike a mutually beneficial deal. \u00a0This is late in the process after they have invested a lot of time and money in you, specifically, <em>not at the interview<\/em>. \u00a0Remember that there are large costs associated with them saying \u201cNo, we can\u2019t make that work\u201d and, appropriately, they will probably not scuttle the deal over comparatively small issues which matter quite a bit to you, like e.g. taking their offer and countering for that plus a few thousand bucks then sticking to it.<\/p><p>f) \u00a0Read a book. \u00a0Many have been written about negotiation. \u00a0I like Getting To Yes. \u00a0It is a little disconcerting that negotiation skills are worth thousands of dollars <em>per year for your entire career<\/em> but engineers think that directed effort to study them is crazy when that could be applied to trivialities about a technology that briefly caught their fancy.<\/p><p><strong>How to value an equity grant<\/strong>:<\/p><p>Roll d100. \u00a0(Not the right kind of geek? \u00a0Sorry. \u00a0rand(100) then.)<\/p><p>0~70: Your equity grant is worth nothing.<\/p><p>71~94: Your equity grant is worth a lump sum of money which makes you about as much money as you gave up working for the startup, instead of working for a megacorp at a higher salary with better benefits.<\/p><p>95~99: Your equity grant is a lifechanging amount of money. \u00a0You won\u2019t feel rich \u2014 you\u2019re not the richest person you know, because many of the people you spent the last several years with are now richer than you by definition \u2014 but your family will never again give you grief for not having gone into $FAVORED_FIELD like a proper $YOUR_INGROUP.<\/p><p>100: You worked at the next Google, and are rich beyond the dreams of avarice. \u00a0Congratulations.<\/p><p>Perceptive readers will note that 100 does not actually show up on a d100 or rand(100).<\/p><p><strong>Why are you so negative about equity grants?<\/strong><\/p><p>Because you radically overestimate the likelihood that your startup will succeed and radically overestimate the portion of the pie that will be allocated to you if the startup succeeds. \u00a0Read about dilution and liquidation preferences on Hacker News or Venture Hacks, then remember that there are people who know more about negotiating deals than you know about programming and imagine what you could do to a program if there were several hundred million on the line.<\/p><p><strong>Are startups great for your career as a fresh graduate?<\/strong><\/p><p>The high-percentage outcome is you work really hard for the next couple of years, fail ingloriously, and then be jobless and looking to get into another startup. \u00a0If you really wanted to get into a startup two years out of school, you could also just go work at a megacorp for the next two years, earn a bit of money, then take your warchest, domain knowledge, and contacts and found one.<\/p><p>Working at a startup, you tend to meet people doing startups. \u00a0Most of them will not be able to hire you in two years. \u00a0Working at a large corporation, you tend to meet other people in large corporations in your area. \u00a0Many of them either will be able to hire you or will have the ear of someone able to hire you in two years.<\/p><p><strong>So would you recommend working at a startup? \u00a0<\/strong>Working in a startup is a career path but, more than that, it is a lifestyle choice. \u00a0This is similar to working in investment banking or academia. \u00a0Those are three very different lifestyles. \u00a0Many people will attempt to sell you those lifestyles as being in your interests, for their own reasons. \u00a0If you genuinely would enjoy that lifestyle, go nuts. \u00a0If you only enjoy certain bits of it, remember that many things are available a la carte if you really want them. \u00a0For example, if you want to work on cutting-edge technology but also want to see your kids at 5:30 PM, you can work on cutting-edge technology at many, many, many megacorps.<\/p><p>(Yeah, really. \u00a0If it creates value for them, heck yes, they\u2019ll invest in it. \u00a0They\u2019ll also invest in a lot of CRUD apps, but then again, so do startups \u2014 they just market making CRUD apps better than most megacorps do. \u00a0The first hour of the Social Network is about making a CRUD app seem like sexy, the second is a Lifetime drama about a divorce improbably involving two heterosexual men.)<\/p><p><strong>Your most important professional skill is communication<\/strong>: Remember engineers are not hired to create programs and how they are hired to create business value? \u00a0The dominant quality which gets you jobs is the ability to <em>give people the perception that you will create value<\/em>. \u00a0This is not necessarily coextensive with ability to create value.<\/p><p>Some of the best programmers I know are pathologically incapable of carrying on a conversation. \u00a0People disproportionately a) wouldn\u2019t want to work with them or b) will underestimate their value-creation ability because they gain insight into that ability through conversation and the person just doesn\u2019t implement that protocol. \u00a0Conversely, people routinely assume that I am among the best programmers they know entirely because a) there exists observable evidence that I can program and b) I write and speak really, really well.<\/p><p>(Once upon a time I would have described myself as \u201cSlightly below average\u201d in programming skill. \u00a0I have since learned that I had a radically skewed impression of the skill distribution, that programming skill is not what people actually optimize for, and that modesty is against my interests. \u00a0These days if you ask me how good of a programmer I am I will start telling you stories about how I have programmed systems which helped millions of kids learn to read or which provably made companies millions. \u00a0The question of where I am on the bell curve matters to no one, so why bother worrying about it?)<\/p><p>Communication is a skill. \u00a0Practice it: you will get better. \u00a0One key sub-skill is being able to quickly, concisely, and confidently explain how you create value to someone who is not an expert in your field and who does not have a priori reasons to love you. \u00a0If when you attempt to do this technical buzzwords keep coming up (\u201cReduced 99th percentile query times by 200 ms by optimizing indexes on\u2026\u201d), take them out and try again. \u00a0You should be able to explain what you do to a bright 8 year old, the CFO of your company, or a programmer in a different specialty, at whatever the appropriate level of abstraction is.<\/p><p><strong>You will often be called to do Enterprise Sales and other stuff you got into engineering to avoid<\/strong>: Enterprise Sales is going into a corporation and trying to convince them to spend six or seven figures on buying a system which will either improve their revenue or reduce costs. \u00a0Every job interview you will ever have is Enterprise Sales. \u00a0Politics, relationships, and communication skills matter a heck of a lot, technical reality not quite so much.<\/p><p>When you have meetings with coworkers and are attempting to convince \u00a0them to implement your suggestions, you will also be doing Enterprise Sales. \u00a0If getting stuff done is your job description, then convincing people to get stuff done is a core job skill for you. \u00a0Spend appropriate effort on getting good at it. \u00a0This means being able to communicate effectively in memos, emails, conversations, meetings, and PowerPoint (when appropriate). \u00a0It means understanding how to make a business case for a technological\u00a0initiative. \u00a0It means knowing that sometimes you will make technological sacrifices in pursuit of business objectives and that this is the right call.<\/p><p><strong>Modesty is not a career-enhancing character trait<\/strong>: Many engineers have self-confidence issues (hello, self). \u00a0Many also come from upbringings where modesty with regards to one\u2019s accomplishments is culturally celebrated. \u00a0American businesses largely do not value modesty about one\u2019s accomplishments. \u00a0The right tone to aim for in interviews, interactions with other people, and life is closer to \u201crestrained, confident professionalism.\u201d<\/p><p>If you are part of a team effort and the team effort succeeds, the right note to hit is not \u201cI owe it all to my team\u201d unless your position is such that everyone will understand you are lying to be modest. \u00a0Try for \u201cIt was a privilege to assist my team by leading their efforts with regards to $YOUR_SPECIALTY.\u201d \u00a0Say it in a mirror a thousand times until you can say it with a straight face. \u00a0You might feel like you\u2019re overstating your accomplishments. \u00a0Screw that. \u00a0Someone who claims to Lead Efforts To Optimize Production while having the title\u00a0Sandwich\u00a0Artist is overstating their accomplishments. \u00a0You are an engineer. \u00a0You work magic which makes people\u2019s lives better. \u00a0If you were in charge of the database specifically on an important project involving people then <em>heck yes<\/em> you lead the database effort which was crucial for the success of the project. \u00a0This is how the game is played. \u00a0If you feel poorly about it, you\u2019re like a batter who feels poorly about stealing bases in baseball: you\u2019re not morally superior, you\u2019re just playing poorly<\/p><p><strong>All business decisions are ultimately made by one or a handful of multi-cellular organisms closely related to chimpanzees, not by rules or by algorithms<\/strong>: People are people. \u00a0Social grooming is a really important skill. \u00a0People will often back suggestions by friends because they are friends, even when other suggestions might actually be better. \u00a0People will often be favoritably disposed to people they have broken bread with. \u00a0(There is a business book called Never Eat Alone. \u00a0It might be worth reading, but that title is whatever the antonym of deceptive advertising is.) \u00a0People routinely favor people who they think are like them over people they think are not like them. \u00a0(This can be good, neutral, or invidious. \u00a0Accepting that it happens is the first step to profitably exploiting it.)<\/p><p>Actual grooming is at least moderately important, too, because people are <em>hilariously<\/em> easy to hack by expedients such as dressing appropriately for the situation, maintaining a professional appearance, speaking in a confident tone of voice, etc. \u00a0Your business suit will probably cost about as much as a computer monitor. \u00a0You only need it once in a blue moon, but when you need it you\u2019ll be really, really, really glad that you have it. \u00a0Take my word for it, if I wear everyday casual when I visit e.g. City Hall I get treated like a hapless awkward twenty-something, if I wear the suit I get treated like the CEO of a multinational company. \u00a0I\u2019m actually the awkward twenty-something CEO of a multinational company, but I get to pick which side to emphasize when I want favorable treatment from a\u00a0bureaucrat.<\/p><p>(People familiar with my business might object to me describing it as a multinational company because it is not what most people think of when \u201cmultinational company\u201d gets used in conversation. \u00a0Sorry \u2014 it is a simple conversational hack. \u00a0If you think people are pissed off at being manipulated when they find that out, well, some people passionately hate business suits, too. \u00a0That doesn\u2019t mean business suits are valueless. \u00a0Be appropriate to the circumstances. \u00a0Technically true answers are the best kind of answers when the alternative is Immigration deporting you, by the way.)<\/p><p><strong>At the end of the day, your life happiness will not be dominated by your career.<\/strong>\u00a0 Either talk to older people or trust the social scientists who have: family, faith, hobbies, etc etc generally swamp career achievements and money in terms of things which actually produce happiness. \u00a0Optimize appropriately. \u00a0Your career is important, and right now it might seem like the most important thing in your life, but odds are that is not what you\u2019ll believe forever. \u00a0Work to live, don\u2019t live to work.<\/p>","created_at":"2018-11-07T16:39:58+0000","updated_at":"2018-11-07T16:39:58+0000","annotations":[],"mimetype":"text\/html","reading_time":28,"domain_name":"www.kalzumeus.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":227,"title":"Procedural Worlds from Simple Tiles","url":"http:\/\/ijdykeman.github.io\/ml\/2017\/10\/12\/wang-tile-procedural-generation.html","content":"<p>In this post, I\u2019ll describe the two algorithms for creating complex procedural worlds from simple sets of colored tiles and constraints on the placements of those tiles.  I will show that by carefully designing these tile sets, you can create interesting procedurally generated content, such as landscapes with cities, or dungeons with complex internal structure.  The video below shows the system creating a procedural world based on the rules encoded in 43 colored tiles.<\/p><iframe width=\"100%\" height=\"500\" src=\"https:\/\/www.youtube.com\/embed\/XVIYY0AQF-8?rel=0&amp;showinfo=0\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>\n<p>The image below shows the tile set that generated the world in the video above.  The world is annotated to assist in the exercise of imagination required to see it as an actual environment.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/map_explanation.svg\" alt=\"map explanation\" \/><\/p>\n<p>We will define a tiling as a finite grid where one of those tiles lies in each grid square.  We will further define a valid world as one where the colors along the edges of adjacent tiles must be the same.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/pair_examples.svg\" alt=\"valid pairs\" \/><\/p>\n<p><em>This is the only concrete rule for a tiling: tile edge colors must match up.  All the higher level structure flows from this.<\/em><\/p>\n<p>A valid tiling looks like this:<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/example_valid_tiling.png\" alt=\"valid tiling\" \/><\/p>\n<p><em>This is a tiling which is supposed to represent a map with water, beaches, grass, towns with buildings (blue rectangles), and snow capped mountains.  The black lines show borders between tiles.<\/em><\/p>\n\n<p>I think this is an interesting way to describe and create worlds because very often, procedural generation algorithms take a top-down approach.  L-systems, for instance, rely on a recursive description of an object where the top level, large details are determined before lower level ones.  There is nothing wrong with this approach, but I think that it is interesting to create tile sets that are only able to encode simple low level relationships (e.g., ocean water and grass must be separated by a beach, buildings can only have convex, 90 degree corners) and see high level patterns emerge (e.g. square buildings).<\/p>\n<h2 id=\"tiling-is-an-np-complete-constraint-satisfaction-problem\">Tiling is an NP complete Constraint Satisfaction Problem<\/h2>\n<p>For the reader familiar with constraint satisfaction problems (CSPs), it will already be clear that tiling a finite world is a CSP.  In a CSP, we have a set of variables, a set of values that each variable can take on (called its domain), and a set of constraints.  For us, the variables are locations in the map, the domain of each variable is the tile set, and the constraints are that tiles must match along their edges with their neighbors.<\/p>\n<p>Intuitively, the problem of correctly creating a nontrivial tiling is hard because the tile sets can encode arbitrary, long range dependencies.  Formally, this is an NP complete constraint satisfaction problem when we consider tile sets in general.  A naive algorithm for creating tilings would exhaustively search the space of tilings and run in exponential time.  The hope is that we can create interesting worlds using tile sets that are solvable using search accelerated by heuristics.  The other option is to create tilings that are nearly correct, but may have a small number of incorrect placements.  I have found two algorithms that work well with some interesting tile sets, and I describe them below.<\/p>\n\n<h2 id=\"method-1-greedy-placement-with-backjumping\">Method 1: Greedy Placement with Backjumping<\/h2>\n<p><em>Keep choosing random locations and place valid tiles there.  If you get stuck, remove some and try again.<\/em><\/p>\n<div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>Initialize the entire map to UNDECIDED\nwhile UNDECIDED tiles remain on the map\n    if any valid tile can be placed on the map\n      t &lt;- collection of all possible valid tile placements\n      l &lt;- random selection from t weighted by tile probabilities\n      place l on the map\n    else\n      select a random UNDECIDED tile and set all its neighbors to UNDECIDED\n<\/pre><\/div><\/div>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/randomized_algo_diagram.svg\" alt=\"randomized algorithm example\" \/><\/p>\n<p>The first approach I took to creating a tiling from a tile set is to simply start with the entire grid in an undefined state, and to iteratively place a random tile in a location where it is valid, or, if no locations are valid, set a small region on near an undefined tile to be undefined and continue greedy placement.  The \u201cgreedy placement\u201d is the strategy of placing a tile as long as all its edges line up with existing tiles, without regard for whether this placement will create a partial tiling that cannot be completed without removing existing tiles.  When such a situation arises, and we cannot place any more tiles, we must remove some previously placed tiles.  But we can\u2019t say which are ideal to remove, because if we could solve that problem, we could probably also have solved the problem of placing tiles in a smart way in the first place.  To give the algorithm another chance at finding a valid tiling for a given area, we set all the tiles around the location that undefined and continue with the greedy placement strategy.  Eventually, the hope is, a valid tiling will be found, but this is not guaranteed.  The algorithm will continue to run until a valid tiling is found, which may be forever.  It has no ability to detect when a tile set is unsolvable.<\/p>\n<p>There is no guarantee that this algorithm will halt.  A simple tile set with two tiles that share no colors would cause this algorithm to loop forever.  An even simpler case would be one tile with different colors on the top and bottom.  It might make sense to somehow check for tile sets that cannot produce valid tilings.  We might say that a tile set is certainly valid if it can tile an infinite plane.  In some cases it is clearly possible to prove or disprove whether a tile set can tile an infinite plane, but the problem turns out to be undecidable in general.  Therefore, it is up to the designer of the tile set to create one which can yield a valid tiling.<\/p>\n<p>This algorithm is unable to find good solutions for the dungeon tile set in the video at the top of this post.  It performs well on simpler tile sets.  We would like to solve more complex tile sets where many types of transitions between tiles are possible and lots of rules (e.g. roads begin and end at buildings) are encoded.\nWe need an algorithm that is able to look ahead and take make tile placements while being somehow aware of what options that placement will leave open for future placements.  This will allow us to efficiently solve complex tile sets.<\/p>\n<h3 id=\"from-a-constraint-satisfaction-perspective\">From a constraint satisfaction perspective<\/h3>\n<p>This algorithm is equivalent to a backjumping search.  At each step, we attempt to assign a single variable.  If we cannot, we unassign a variable and all the variables which are connected by constraints to it.  This is called backjumping, which differs from backtracking where we unassign one variable at a time until we can continue making valid assignments.  In backtracking, we generally unassign variables in the reverse of the order we assigned them, but with backjumping we unassign variables according to the structure of the problem at hand.  It makes sense that if we can\u2019t place any tile at a particular location, we should change the placement of neighboring tiles, since their placement created an unsolvable situation.  Backtracking may instead cause us to unassign variables that are spatially far away but happened to be assigned recently.<\/p>\n<p>This search does not employ any local consistency method.  That is, we make no attempt to place tiles that won\u2019t cause an unsolvable situation later on, even as soon as one step of search in the future.  It may be possible to accelerate search by keeping track of the effects that a placement will have on possible placements a few tiles away.  Hopefully, this will keep our search from spending so much time undoing its work.  This is what the next algorithm does.<\/p>\n<p><em>Maintain a probability distribution over tiles at each location, making nonlocal updates to these distributions when a placement decision is made.  Never backtrack.<\/em><\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/wave_collapse_example.svg\" alt=\"wave collapse example\" \/><\/p>\n<p>Next, I\u2019ll describe an algorithm which is guaranteed to halt and produces better looking results for all the tile sets I have tried.  It is also able to produce nearly-valid tilings for much more complicated tile sets.  The tradeoff is that this algorithm does not guarantee that its output is always a valid tiling.  The Optimizations section describes optimizations that help this technique run fast even over large tile sets and maps.<\/p>\n<p>The difficulty of creating a valid tiling is largely determined by the number of transitions necessary to get between two tile types.  A simple tile set might contain only sand, water, and grass.  If grass and water cannot touch, then a transition to sand will be necessary between the two.  This is a simple case that the previous algorithm can solve easily.  A more complex case might involve many nested levels of tile types.  For instance, you might have deep water, water, sand, grass, high plain, mountain, and snow cap.  Seven transitions would need to be present in the map for all of these types to appear, assuming that these types cannot touch except in the order I stated them.  Further complexity can be introduced by creating tiles that naturally create long-distance dependencies between tiles, like roads that must start and end at certain types of tiles.<\/p>\n<p>Intuitively, it makes sense that an algorithm for this problem should have some ability to \u201clook ahead\u201d and consider at least a few transitions out what the consequences of its placement choices might be.  In order to do this, the tile map is considered to be a probability distribution over tiles at each location.  When the algorithm places a tile, it updates the probability distributions around that tile in response to the placement so that the probabilities of nearby tiles that are likely to be compatible with the placement are increased.<\/p>\n<p>For example, if a water tile is placed on the map, the tiles next to it must contain some water.  The tiles next to those tiles might also contain water, but there are other possibilities, such as grass if a beach was placed next to the original water.  The farther we get from the placed tile, the more types of tiles become possible.  To make concrete use of this observation, we can count the number of ways that we can arrive at each tile placement near the original tile.  In some cases, only a single sequence of transitions can cause one tile to transition to another over a given distance.  In other cases, there could be many possible transition sequences.  Once we have placed a tile, we can determine the probability distributions of tiles at nearby locations by counting the number of ways we can transition from the tile we have placed to nearby tiles.  The \u201clook ahead\u201d that the algorithm performs is tracking these transition counts and treating them as probability distributions from which to select future tiles to place.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/tile_transition_counts.svg\" alt=\"tile transition counts\" \/><\/p>\n<p>At each time step, the algorithm examines all non-decided tile locations, each of which is a probability distribution over tiles, and selects one location to \u201ccollapse\u201d into a tile.  It selects the distribution from the map with the lowest entropy.  Low entropy multinomial distributions tend to have their probability concentrated in a few modes, so collapsing these first yields the effect of placing tiles we already have some constraints for.  This is why the algorithm fills in tiles near tiles that are already decided first.<\/p>\n\n<p>This is the most effective algorithm I have managed to implement for this problem, and it has the added advantage of lending itself to nice visualizations as it runs.  It may be possible to improve this algorithm by adding some form of backtracking.  If an invalid location exists in the final tiling, undoing nearby tile placements and then resampling from the resulting distributions at their locations may allow a fix to be found for the final tiling.  Of course, if you were determined to always keep searching until a valid tiling is found, you would lose the bounded run time guarantee we have now.<\/p>\n<h3 id=\"optimizations\">Optimizations<\/h3>\n<p>The core operation of this method is updating the probabilities around a placed tile.  One approach would be to count the possible transitions outward from a placed tile each time a tile is placed.  This would be very slow, since many transition pairs would need to be considered for each map location to which the new probabilities are propagated.  One obvious optimization is not to propagate across the entire map.  A more interesting optimization is to cache the effect that each tile placement will have on the locations around it, so that each tile placement merely does a lookup to see what type of changes to the neighboring probabilities a placement makes, and then apply that change via some simple operation.  I will describe my technique for doing this below.<\/p>\n<p>Imagine a tile were placed on an otherwise empty map.  This placement would update the probabilities of tiles nearby.  We can think of these updated distributions as having a prior distribution which is informed by previous placements.  If multiple tiles are placed, this prior is a joint distribution.  I approximate the posterior given this joint prior as the product of the distributions given each placement in the past.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/sphere_example.svg\" alt=\"sphere example\" \/><\/p>\n<p>To implement this, I imagine that when a tile is placed on an empty map, it induces a significant change in the distributions in the map nearby.  I call these updates the <em>sphere<\/em> of the tile, as in, the sphere of influence the tile projects around it when it is placed on an empty map.  When two tiles are placed near each other, their spheres interact to create the final distributions that are affected by both placements.  Considering that many tiles may be placed near a given undecided location, there could be a large number of interacting constraints that would make a counting based approach to finding the probability of different tiles appearing at that location slow.  What if we instead only consider a simple model of the interaction between the precomputed spheres of the tiles that have already been placed?<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/labeled_sphere.svg\" alt=\"sphere diagram\" \/><\/p>\n<p>When a tile is placed, I update the probability map by elementwise multiplying the distribution of that tile\u2019s sphere at each location in the map with the distribution already stored in that location in the map.  It may be helpful to consider an example of what this might do to the distribution map.  Say a given location in the map currently has a distribution that considers grass and water likely, and we place a water tile next to that location.  The water tile\u2019s sphere will have a high probability of water right next to the water tile, and a low probability of grass.  When we multiply these distributions together elementwise, the probability of water in the result is high, because it is the product of two large probabilities, but the probability of grass will become low, because it is the product of the high probability stored in the map with the low probability stored in the sphere.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/interacting_spheres.svg\" alt=\"spheres interacting diagram\" \/><\/p>\n<p>This strategy allows us to efficiently approximate the effect that each tile placement should have on the probability map.\n&lt;!\u2013<\/p>\n<h3 id=\"from-a-constraint-satisfaction-perspective-1\">From a constraint satisfaction perspective<\/h3>\n<p>To solve constraint satisfaction problems efficiently, it often makes sense to keep track of what assignments of other variables become impossible when a particular variable is assigned.  This is known as enforcing local consistency conditions.  Enforcing some kind of local consistency helps prevent you from assigning a value to a variable then assigning an incompatible value to a nearby value right away and being forced to backtrack.  Such transformations are under the umbrella of constraint propagation methods in the CSP literature.  In this algorithm, we are propagating information through a small area of the map each time we place a tile about what tiles can or cannot appear nearby.  If we place a mountain tile, for instance, we know that there can\u2019t be an open ocean tile only two tiles away, so the probability of ocean at all locations on the map two tiles from the placement are set to zero.  This information is recorded in the spheres discussed above.  The spheres encode the local consistency we are interested in imposing.<\/p>\n<p>By reducing the number of possible assignments of nearby tiles, we significantly reduce the search space the algorithm needs to handle after with each placement.  We know that in that small neighborhood, all probabilities of tiles appearing that are incompatible with the places tile are zero.  This is equivalent to removing those values from the domains of the variables at those locations.  This means that each pair of neighboring locations in the region around the placement has some tile in its domain that is compatible with some tile still in its neighbors domain.  When two variables are connected by a constraint in a CSP problem and their domains contain only values that could satisfy the constraint, they are said to be arc consistent, so this method is really an efficient arc consistency enforcing strategy.<\/p>\n<p>In a CSP, the \u201cmost constrained\u201d variable in a given partial assignment is the one with the fewest possible values remaining in its domain.  The idea of placing a tile at the location of the lowest entropy distribution in the map is analogous to assigning a value to the most constrained variable in a CSP, which is a common variable ordering heuristic when solving CSPs by search.\n \u2013&gt;<\/p>\n<h2 id=\"manipulating-tilings-by-changing-tile-selection-probabilities\">Manipulating Tilings by Changing Tile Selection Probabilities.<\/h2>\n<p>So far, I\u2019ve talked about only about how to create valid tilings, but beyond being valid, there might be other properties we might like a tiling to have.  For instance, we might like it to have a certain ratio of one tile type to another, or we might like to ensure that it is not all one uniform type of tile, even if such a tiling is valid.  To accomplish this, both the algorithms I describe take as input a base probability associated with each tile.  The higher this probability, the more likely that tile should be in the final tiling.  Both algorithms make random choices over collections of tiles, and I simply weight these random choices by the base tile probabilities.<\/p>\n<p>Below, you can see how this is used.  By adjusting the probability of the solid water tile appearing, I can adjust the size and frequency of water features on the map.<\/p>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/water_probability.svg\" alt=\"water probability comparison\" \/><\/p>\n<h2 id=\"make-your-own-tile-sets\">Make Your Own Tile Sets<\/h2>\n<p>In short,<\/p>\n<ul><li>clone my <a href=\"https:\/\/github.com\/IJDykeman\/wangTiles\">repo on github<\/a><\/li>\n  <li>install Processing<\/li>\n  <li>in the data\/ folder of the repo, modify tiles.png<\/li>\n  <li>use processing to open wangTiles.pde and click the play button<\/li>\n<\/ul><p>Using the code (You can modify it, but do so at your own risk; I wrote most of it in high school) I\u2019ve put on github, you can create your own tilesets using an image editor, and see how the tiling solver creates worlds with them.  Simply clone the repo and edit the image named dungeon.png, then use Processing to run wangTiles.pde to see an animation of the map being generated. I\u2019ll now describe the \u201clanguage\u201d that the tiling solver expects.<\/p>\n<h3 id=\"the-tile-set-specification\">The Tile Set Specification<\/h3>\n<p><img src=\"http:\/\/ijdykeman.github.io\/assets\/wang_tiles\/tile_spec.svg\" alt=\"tile spec\" \/><\/p>\n<p>The tiles are laid out on a grid of 4x4 cells.  Each cell contains a colored tile in the upper left 3x3 region, and the remaining 7 pixel contain metadata about the tile.  The pixel below the center of the tile can be set to pure red to comment that tile out of the set.  The solvers will never include it in a map.  The upper pixel to the right of the tile can be set to pure black to add all 4 rotations of the tile to the tile set as well.  This is a nice shorthand when you want to add something like a corner, which can reasonably exist in 4 orientations.  Finally, the most important piece of markup is the pixel below and to the left of the tile.  This controls the base probability of that tile appearing in the map.  The darker the pixel is, the more likely that tile is to appear.<\/p>\n<p>Many people have explored <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wang_tile\">Wang tiles<\/a>, which are tile sets with colored edges that must match in edge color with tiles they are placed next to, just like the tiles I have talked about here.<\/p>\n<p>The Wave Function Collapse project by <a href=\"https:\/\/twitter.com\/ExUtumno\">@ExUtumno<\/a> explores creating textures from examples, including textures that represent maps of simple dungeons.  That algorithm maintains a map boolean vectors representing  possible pixel assignments and progressively removes possible assignments from consideration when they are discovered to be incompatible with pixels placed so far.  As is pointed in the paper <a href=\"https:\/\/dl.acm.org\/citation.cfm?id=3102071.3110566\">WaveFunctionCollapse is constraint solving in the wild<\/a>, this information propagation step is the same as the well-known <a href=\"https:\/\/en.wikipedia.org\/wiki\/AC-3_algorithm\">AC-3<\/a> algorithm for local consistency originally introduced in 1977.  The system as a whole is effectively a greedy constraint satisfaction solver with some modifications to its tie breaking to help the result texture match the statistics of the input texture.  This is a different approach than those discussed here, which don\u2019t use canonical local consistency methods like AC-3, and in the case of the second algorithm, maintain a continuous, as opposed to discrete, distribution over possible placement decisions at each location.<\/p>\n<p>Wave Function Collapse is predated by Paul Merrell\u2019s <a href=\"http:\/\/graphics.stanford.edu\/~pmerrell\/model_synthesis.pdf\">Model Synthesis<\/a> algorithm, which appears to also be an AC-3-based constraint satisfaction system, although AC-3 is not mentioned in this paper or Merrell\u2019s thesis.  Merrell considers the problem of generating 3D models from 3D tiles with adjacency constraints, like those I have in 2D above.   Merrel\u2019s work includes an interesting sampling strategy where he starts with a trivial solution and progressively replaces sections of it using his constraint satisfaction algorithm, maintaining a consistent model at each step.<\/p>\n<h2 id=\"conclusion-and-acknowledgments\">Conclusion and Acknowledgments<\/h2>\n<p>Enforcing local rules to create high level structure, such as having a rule that roads have no dead ends enforcing the fact that roads must lead from some non-road location to another, is an interesting perspective on the problem of procedural generation.  A surprising range of interesting rules can be encoded this way.  The quality of the resulting tilings is highly dependent on the solver, however, so a fruitful direction for future work would be improving the algorithms that do the tiling.  I have not explored using genetic algorithms, simulated annealing, or general purpose constraint satisfaction algorithms for this task yet.  Any of these might yield a tile set solver that is able to handle a wide range of tile sets with minimal parameter tuning.<\/p>\n<p>I would like to thank <a href=\"http:\/\/cannontwo.com\/\">Cannon Lewis<\/a> and Ronnie McLaren for helpful discussions during this project.  Ronnie came up with many interesting ideas about how to encode interesting behavior in the tiles.  The wave collapse algorithm resulted from a conversation with Cannon, and his advice was instrumental in creating the efficient approximate algorithm that I described in here.<\/p>","created_at":"2018-11-07T16:39:58+0000","updated_at":"2018-11-07T16:39:58+0000","annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"ijdykeman.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":224,"title":"r\/skyrimvr - HOW TO Install MODS in SKYRIM VR - EASY GUIDE - NEXUS MOD MANAGER","url":"https:\/\/www.reddit.com\/r\/skyrimvr\/comments\/89705e\/how_to_install_mods_in_skyrim_vr_easy_guide_nexus\/","content":"<header class=\"_2GyPfdsi-MbQFyHRECo9GO cx1ohrUAq6ARaXTX2u8YN  aswboo-0 isJDXQ s1czsxkv-0 bYsyvY\" data-redditstyle=\"true\"><div class=\"_2vkeRJojnV7cb9pMlPHy7d\"><div class=\"_3dnbqz69WJTFCss8wl7Wlk\">Press J to jump to the feed. Press question mark to learn the rest of the keyboard shortcuts<a aria-label=\"Home\" class=\"_30BbATRhFv3V83DHNDjJAO\" href=\"https:\/\/www.reddit.com\/\"><\/a><\/div><\/div><\/header><div class=\"_1nxEQl5D2Bx2jxDILRHemb  aswboo-1 WiUpz\" aria-hidden=\"false\"><div><div class=\"s1nj20jv-0 cvlibz\"><div class=\"s1nj20jv-1 LjQEP\"><div class=\"s1nj20jv-5 cVwMdj\"><div class=\"sogqxs-0 jlPabA\"><div class=\"s1enzdkr-4 gBQKdN\"><div class=\"s1enzdkr-4 gBQKdN\"><div class=\"s1knm1ot-9 itJXKz _2rszc84L136gWQrkwH6IaM Post t3_89705e  s1ukwo15-0 RqhAo\" id=\"t3_89705e\" tabindex=\"-1\"><div data-test-id=\"post-content\"><div class=\"_1KNG36IrXcP5X-eLQsMjZb\"><div class=\"s11bh4ne-0 hHHapT\"><div class=\"cZPZhMe-UCZ8htPodMyJ5\"><div class=\"_3AStxql1mQsrZuUIFP9xSg nU4Je7n-eSXStTBAPMYt8\">Posted by<div class=\"wx076j-0 hPglCh\"><a class=\"_2tbHP6ZydRpjI44J3syuqC s1461iz-1 gWXVVu\" href=\"https:\/\/www.reddit.com\/user\/tyrielwood\">u\/tyrielwood<\/a><\/div><\/div><\/div><\/div><div class=\"_1rcejqgj_laTEewtp2DbWG s1knm1ot-0 cKmWbx\"><div class=\"_1hLrLjnE1G_RBCNcN9MVQf\"><img src=\"https:\/\/www.redditstatic.com\/desktop2x\/img\/renderTimingPixel.png\" alt=\"image\" \/><\/div><\/div><p><a class=\"b5szba-0 dcClOD\" href=\"https:\/\/youtu.be\/Iw_lmNawuh0\" rel=\"noopener noreferrer\" target=\"_blank\">youtu.be\/Iw_lmN...<i class=\"icon icon-outboundLink jlrhi6-0 esUKm\"><\/i><\/a><\/p><div class=\"_1hwEKkB_38tIoal6fcdrt9\"><div class=\"_3-miAEojrCvx_4FQ8x3P-s kajbny-2 cVXrua\"><p>260 comments<\/p><\/div><p>97% Upvoted<\/p><\/div><\/div><\/div><\/div><div class=\"s6xxknp-6 jbHikj\"><p>Sort by<\/p><\/div><\/div><\/div><\/div><div class=\"s1nj20jv-6 kIGpJp\"><div class=\"RffBqG1xCvuy1r49wYhLr rled4m-1 kGkcLQ\"><div class=\"s73mpk5-0 duOLeC\"><div class=\"_ZhON3a3vplThB8NFwuJn\"><p>Community Details<\/p><\/div><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s6tnjvv-1 bvjZzf\"><a class=\"s6tnjvv-6 dRpzTj\" href=\"https:\/\/www.reddit.com\/r\/skyrimvr\/\">r\/skyrimvr<\/a><\/div><div class=\"s6tnjvv-10 eFxsKF\"><div class=\"s6tnjvv-11 eAIyWI\"><p class=\"s6tnjvv-12 gDgIlP\">140<\/p><p class=\"s6tnjvv-13 gnSfSX\" id=\"IdCard--CurrentlyViewing--undefined--t5_3m051\">Online<\/p><\/div><\/div><p class=\"s6tnjvv-14 bPUliZ\">Subreddit for discussing SkyrimVR. Unofficial\/Not associated with Bethesda.\n<\/p><p><a class=\"s6tnjvv-17 cvgsde s1l18q3c-2 duVyzU\" href=\"https:\/\/www.reddit.com\/r\/skyrimvr\/submit\">Create Post<\/a><\/p><\/div><\/div><div class=\"ii4q9d-0 fePAzF\"><div class=\"s73mpk5-0 hjCnwh\"><div class=\"TmgZY6tDcdErbE5d7E0HJ\"><div class=\"s1q6n0r9-1 bGifwX ser2k5-0 hIaYoE\"><div class=\"md\"><p>These are the most common links to relevant pages<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:39:57+0000","updated_at":"2018-11-07T16:39:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.reddit.com","preview_picture":"https:\/\/external-preview.redd.it\/icpXwoJAE1Hu58nnGBPnczKVQJVIiB2rrsfHKz5RZX4.jpg?auto=webp&s=553f88d912bab8ba7d728d932036473dc03f255b","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":221,"title":"noprompt\/garden","url":"https:\/\/github.com\/noprompt\/garden","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Garden is a library for rendering CSS in Clojure and ClojureScript.\nConceptually similar to <a href=\"https:\/\/github.com\/weavejester\/hiccup\">Hiccup<\/a>, it uses\nvectors to represent rules and maps to represent declarations. It is designed\nfor stylesheet authors who are interested in what's possible when you trade a\npreprocessor for a programming language.<\/p>\n<h2><a id=\"user-content-table-of-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#table-of-contents\"><\/a>Table of Contents<\/h2>\n<ul><li><a href=\"#getting-started\">Getting Started<\/a><\/li>\n<li><a href=\"#syntax\">Syntax<\/a><\/li>\n<li><a href=\"#development\">Development<\/a><\/li>\n<li><a href=\"#community\">Community<\/a><\/li>\n<li><a href=\"#help\">Help!<\/a><\/li>\n<\/ul><h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting Started<\/h2>\n<p>Add the following dependency to your <code>project.clj<\/code> file:<\/p>\n<p><a href=\"http:\/\/clojars.org\/garden\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/172931e4b3fc006de7f056b668fb7a447a485165\/687474703a2f2f636c6f6a6172732e6f72672f67617264656e2f6c61746573742d76657273696f6e2e737667\" alt=\"Clojars Project\" data-canonical-src=\"http:\/\/clojars.org\/garden\/latest-version.svg\" \/><\/a><\/p>\n<p>Garden 1.2.5 and below requires Clojure 1.6.0 and is known to work with\nClojureScript 0.0-2342. However, starting with Garden 1.3.0 Garden requires\nClojure 1.7 and ClojureScript 1.7.x to leverage a unified syntax with\n<a href=\"http:\/\/dev.clojure.org\/display\/design\/Reader+Conditionals\" rel=\"nofollow\">reader conditionals<\/a>,\nand other major changes in the compiler and repl in Clojurescript.<\/p>\n<h2><a id=\"user-content-syntax\" class=\"anchor\" aria-hidden=\"true\" href=\"#syntax\"><\/a>Syntax<\/h2>\n<p>Garden syntax is very similar to\n<a href=\"https:\/\/github.com\/weavejester\/hiccup\">Hiccup<\/a>. If you're familiar with Hiccup\nyou should feel right at home. If not, don't sweat it! Garden's syntax is\nsimple.<\/p>\n<p>From your project's root directory start up a new REPL and try the following:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (require '[garden.core :refer [css]])\nnil\nuser=&gt; (css [:body {:font-size \"16px\"}])\n\"body{font-size:16px}\"<\/pre><\/div>\n<p>First you'll notice the use of the <code>css<\/code> function. This function takes an\noptional map of compiler flags, any number of rules, and returns a string of\ncompiled CSS.<\/p>\n<p>Vectors represent rules in CSS. The first <em>n<\/em> <strong>non-collection<\/strong> elements of a\nvector depict the rule's selector where <em>n<\/em> &gt; 0. When <em>n<\/em> = 0 the rule is not\nrendered. To produce a rule which selects the <code>&lt;h1&gt;<\/code> and <code>&lt;h2&gt;<\/code> HTML elements\nfor example, we simply begin a vector with <code>[:h1 :h2]<\/code>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:h1 :h2 {:font-weight \"none\"}])\n\"h1,h2{font-weight:none}\"<\/pre><\/div>\n<p>To target <strong>child selectors<\/strong> nested vectors may be employed:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:h1 [:a {:text-decoration \"none\"}]])\n\"h1 a{text-decoration:none}\"\nuser=&gt; (css [:h1 :h2 [:a {:text-decoration \"none\"}]])\n\"h1 a, h2 a{text-decoration:none}\"<\/pre><\/div>\n<p>As in Less\/Sass, Garden also supports selectors prefixed with the <code>&amp;<\/code>\ncharacter allowing you to reference a <strong>parent selector<\/strong>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:a\n             {:font-weight 'normal\n              :text-decoration 'none}\n             [:&amp;:hover\n              {:font-weight 'bold\n               :text-decoration 'underline}]])\n\"a{text-decoration:none;font-weight:normal}a:hover{text-decoration:underline;font-weight:bold}\"<\/pre><\/div>\n<p>A slightly more complex example demonstrating nested vectors with multiple\nselectors:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:h1 :h2 {:font-weight \"normal\"}\n             [:strong :b {:font-weight \"bold\"}]])\n\"h1,h2{font-weight:normal}h1 strong,h1 b,h2 strong,h2 b{font-weight:bold}\"<\/pre><\/div>\n<p><code>garden.selectors<\/code> namespace defines a CSSSelector record. It doubles as both a\nfunction and a literal (when passed to the css-selector). When the function is\ncalled it will return a new instance that possesses the same properties. All\narguments to the function must satisfy ICSSSelector.<\/p>\n<p><code>garden.selectors<\/code> namespace also defines these macros that create a selector\nrecord: <code>defselector<\/code>, <code>defclass<\/code>, <code>defid<\/code>, <code>defpseudoclass<\/code> and\n<code>defpseudoelement<\/code>.<\/p>\n<p><code>garden.selectors<\/code> namespace also defines many CSSSelector instances such as:<\/p>\n<ul><li>Type selectors <code>a<\/code>, <code>abbr<\/code>, <code>address<\/code> and <a href=\"https:\/\/github.com\/noprompt\/garden\/blob\/master\/src\/garden\/selectors.cljc\">more<\/a><\/li>\n<li>Pseudo-classes <code>active<\/code>, <code>checked<\/code>, <code>disabled<\/code> and\n<a href=\"https:\/\/github.com\/noprompt\/garden\/blob\/master\/src\/garden\/selectors.cljc\">more<\/a><\/li>\n<li>Language and negation pseudo-classes <code>lang<\/code> and <code>not<\/code><\/li>\n<li>Structural pseudo-classes <code>nth-child<\/code>, <code>nth-last-child<\/code>, <code>nth-of-type<\/code> and\n<code>nth-last-of-type<\/code><\/li>\n<li>Pseudo-elements <code>after<\/code>, <code>before<\/code>, <code>first-letter<\/code> and <code>first-line<\/code><\/li>\n<li>Attribute selectors <code>attr=<\/code>, <code>attr-contains<\/code>, <code>attr-starts-with<\/code>,\n<code>attr-starts-with*<\/code>, <code>attr-ends-with<\/code> and <code>attr-matches<\/code><\/li>\n<li>Combinators <code>descendant<\/code>, <code>+<\/code>, <code>-<\/code> and <code>&gt;<\/code><\/li>\n<li>Special selector <code>&amp;<\/code><\/li>\n<\/ul><p>and allows to compose complex selectors such as this:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defselector *)\n(defpseudoclass host [x] x)\n(defpseudoelement content)\n(&gt; (host (attr :flipped)) content (* last-child))\n;; =&gt; :host([flipped]) &gt; ::content &gt; *:last-child<\/pre><\/div>\n<p><code>garden.selectors<\/code> namespace also defines a CSS3 selectors's <code>specificity<\/code>\nfunction:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(specificity \"#s12:not(FOO)\")\n;; =&gt; 101\n(specificity (a hover))\n;; =&gt; 10<\/pre><\/div>\n<p>Clojure maps represent CSS declarations where map keys and values represent CSS\nproperties and values respectively. Garden's declaration syntax is a bit more\ninvolved than rules and understanding it is important to make the most of the\nlibrary.<\/p>\n<p>Declaration map keys <em>should<\/em> either be a string, keyword, or symbol:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:h1 {\"font-weight\" \"normal\"}])\n\"h1{font-weight:normal}\"\nuser=&gt; (css [:h1 {:font-weight \"normal\"}])\n\"h1{font-weight:normal}\"\nuser=&gt; (css [:h1 {'font-weight \"normal\"}])\n\"h1{font-weight:normal}\"<\/pre><\/div>\n<p>Be aware, Garden makes no attempt to validate your declarations and\nwill not raise an error if other key types are used.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:h1 {30000 \"nom-nom\"}])\n\"h1{30000:nom-nom}\"<\/pre><\/div>\n<p>We've already seen strings used as declaration map values, but Garden also\nsupports keywords, symbols, numbers, maps, vectors, and lists in addition.<\/p>\n<h5><a id=\"user-content-custom-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#custom-functions\"><\/a>Custom functions<\/h5>\n<p>Since Garden doesn't have wrappers for all the possible CSS functions,\nsometimes you might need to define the function you need yourself.\nThis is where the <code>defcssfn<\/code> macro comes in handy.\nSuppose you want to use the <code>url<\/code> CSS function, even if it's not available\nin Garden directly you can just define it yourself by simply:<\/p>\n<pre>(defcssfn url)\n;; =&gt; #'user\/url\n<\/pre>\n<p>Which will render like this:<\/p>\n<pre>(css (url \"http:\/\/fonts.googleapis.com\/css?family=Lato\"))\n;; =&gt; url(http:\/\/fonts.googleapis.com\/css?family=Lato)\n<\/pre>\n<h5><a id=\"user-content-strings-keywords-symbols-and-numbers\" class=\"anchor\" aria-hidden=\"true\" href=\"#strings-keywords-symbols-and-numbers\"><\/a>Strings, keywords, symbols, and numbers<\/h5>\n<p>Strings, keywords, symbols, and numbers are rendered as literal CSS values:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:body {:font \"16px sans-serif\"}])\n\"body{font:16px sans-serif}\"<\/pre><\/div>\n<p>Be warned, you must escape literal string values yourself:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>user=&gt; (css [:pre {:font-family \"\\\"Liberation Mono\\\", Consolas, monospace\"}])\n\"pre{font-family:\\\"Liberation Mono\\\", Consolas, monospace}\"\n<\/pre><\/div>\n<h2><a id=\"user-content-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#development\"><\/a>Development<\/h2>\n<h3><a id=\"user-content-leiningen-commands\" class=\"anchor\" aria-hidden=\"true\" href=\"#leiningen-commands\"><\/a>Leiningen commands<\/h3>\n<p>Building ClojureScript<\/p>\n<pre>lein build-cljs\n<\/pre>\n<p>Starting a Node REPL<\/p>\n<pre>lein node-repl\n<\/pre>\n<p>Run Clojure tests, along with a test runner<\/p>\n<pre>lein test-clj\n<\/pre>\n<p>Run ClojureScript tests (on Node)<\/p>\n<pre>lein test-cljs\n<\/pre>\n<p>Run both Clojure <em>and<\/em> ClojureScript tests<\/p>\n<pre>lein test-cljc\n<\/pre>\n<h2><a id=\"user-content-further-reading--wiki\" class=\"anchor\" aria-hidden=\"true\" href=\"#further-reading--wiki\"><\/a>Further Reading &amp; Wiki<\/h2>\n<p>Detailed documentation and a developer guide for Syntax, Rules, Declarations,\nand Plugins is under the community-contributed\n<a href=\"https:\/\/github.com\/noprompt\/garden\/wiki\">wiki<\/a>.<\/p>\n<p>Please contribute!<\/p>\n<h2><a id=\"user-content-contributors\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributors\"><\/a>Contributors<\/h2>\n<p>Listed by first commit:<\/p>\n<h2><a id=\"user-content-help\" class=\"anchor\" aria-hidden=\"true\" href=\"#help\"><\/a>Help!<\/h2>\n<p>This project is looking for team members who can help this project succeed!\nIf you are interested in becoming a team member please open an issue.<\/p>\n<h2><a id=\"user-content-community\" class=\"anchor\" aria-hidden=\"true\" href=\"#community\"><\/a>Community<\/h2>\n<h3><a id=\"user-content-mailing-list\" class=\"anchor\" aria-hidden=\"true\" href=\"#mailing-list\"><\/a>Mailing List<\/h3>\n<ul><li><a href=\"https:\/\/groups.google.com\/forum\/#!forum\/garden-clojure\" rel=\"nofollow\">Google Groups<\/a><\/li>\n<\/ul><h3><a id=\"user-content-slack\" class=\"anchor\" aria-hidden=\"true\" href=\"#slack\"><\/a>Slack<\/h3>\n<ul><li>#clojurescript or #css channel on <a href=\"https:\/\/clojurians.slack.com\" rel=\"nofollow\">Clojurians<\/a><\/li>\n<\/ul><h2><a id=\"user-content-license\" class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Copyright \u00a9 2013-2016 Joel Holdbrooks.<\/p>\n<p>Distributed under the Eclipse Public License, the same as Clojure.<\/p>\n<\/article>","created_at":"2018-11-07T16:39:53+0000","updated_at":"2018-11-07T16:39:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/541996?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":222,"title":"Tweet from \n\n    \n      Jason Grinblat","url":"https:\/\/mobile.twitter.com\/ptychomancer\/status\/980968298002006016","content":"<div class=\"dir-ltr\" dir=\"ltr\">  I love maps &amp; their promise of fractal discovery.\nI love procedural generation and the aesthetics of the unauthored.\nWhere do these two loves intersect? Generated maps.\n\nI am the procgen map admirer. These are my favorite map generators and the folks who create them.\n\n\ud83d\udc47THREAD\ud83d\udc47 <a href=\"https:\/\/t.co\/OQJKRfYw9X\" data-pre-embedded=\"true\" rel=\"nofollow\" data-entity-id=\"980968100802654209\" dir=\"ltr\" data-url=\"https:\/\/twitter.com\/ptychomancer\/status\/980968298002006016\/photo\/1\" data-tco-id=\"OQJKRfYw9X\" class=\"twitter_external_link dir-ltr tco-link has-expanded-path\" target=\"_top\" data-expanded-path=\"\/ptychomancer\/status\/980968298002006016\/photo\/1\">pic.twitter.com\/OQJKRfYw9X<\/a>\n<\/div>","created_at":"2018-11-07T16:39:53+0000","updated_at":"2018-11-07T16:39:53+0000","published_by":["Jason Grinblat"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"mobile.twitter.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":223,"title":"Mathematics for Machine Learning","url":"https:\/\/mml-book.github.io\/","content":"Mathematics for Machine Learning | Companion webpage to the book \u201cMathematics for Machine Learning\u201d. Copyright 2018 by Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong. To be published by Cambridge University Press.\n    <div class=\"wrapper\"><header>\n        \n          <p>Companion webpage to the book \"Mathematics for Machine Learning\". Copyright 2018 by Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong. To be published by Cambridge University Press.<\/p>\n        \n        <p class=\"view\"><a href=\"https:\/\/github.com\/mml-book\/mml-book.github.io\">View the Project on GitHub <\/a><\/p>\n        <\/header><section><p>Please link to this site using <a href=\"https:\/\/mml-book.com\">https:\/\/mml-book.com<\/a>.<\/p><p>Twitter:<a href=\"https:\/\/twitter.com\/mpd37\">@mpd37<\/a>, <a href=\"https:\/\/twitter.com\/analogaldo\">@AnalogAldo<\/a>, <a href=\"https:\/\/twitter.com\/ChengSoonOng\">@ChengSoonOng<\/a>.<\/p><p>We are in the process of writing a book on Mathematics for Machine Learning that motivates people to learn mathematical concepts. The book is not intended to cover advanced machine learning techniques because there are already plenty of books doing this. Instead, we aim to provide the necessary mathematical skills to read those other books.<\/p><p>We split the book into two parts:<\/p><ul><li>Mathematical foundations<\/li>\n  <li>Example machine learning algorithms that use the mathematical foundations<\/li>\n<\/ul><p>We aim to keep this book fairly short (tried for 300 pages, now close to 400 pages),\nso we don\u2019t cover everything.<\/p><p><strong>We will keep PDFs of this book freely available after publication.<\/strong><\/p><h3 id=\"report-errata-and-feedback\"><a href=\"https:\/\/github.com\/mml-book\/mml-book.github.io\/issues\">Report errata and feedback<\/a>.<\/h3><ul><li><a href=\"https:\/\/mml-book.github.io\/book\/toc.pdf\"><strong>Table of Contents<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/notation.pdf\"><strong>Notation<\/strong><\/a><\/li>\n<\/ul><p><a href=\"https:\/\/mml-book.github.io\/book\/foreword.pdf\"><strong>Foreword<\/strong><\/a><\/p><p><strong>Part I: Mathematical Foundations<\/strong><\/p><ol><li><a href=\"https:\/\/mml-book.github.io\/book\/chapter01.pdf\"><strong>Introduction and Motivation<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter02.pdf\"><strong>Linear Algebra<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter03.pdf\"><strong>Analytic Geometry<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter04.pdf\"><strong>Matrix Decompositions<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter05.pdf\"><strong>Vector Calculus<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter06.pdf\"><strong>Probability and Distribution<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter07.pdf\"><strong>Continuous Optimization<\/strong><\/a><\/li>\n<\/ol><p><strong>Part II: Central Machine Learning Problems<\/strong><\/p><ol start=\"8\"><li><a href=\"https:\/\/mml-book.github.io\/book\/chapter08.pdf\"><strong>When Models Meet Data<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter09.pdf\"><strong>Linear Regression<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter10.pdf\"><strong>Dimensionality Reduction with Principal Component Analysis<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter11.pdf\"><strong>Density Estimation with Gaussian Mixture Models<\/strong><\/a><\/li>\n  <li><a href=\"https:\/\/mml-book.github.io\/book\/chapter12.pdf\"><strong>Classification with Support Vector Machines<\/strong><\/a><\/li>\n<\/ol><p><a href=\"https:\/\/mml-book.github.io\/book\/references.pdf\"><strong>references<\/strong><\/a><\/p><p><a href=\"https:\/\/mml-book.github.io\/book\/index.pdf\"><strong>index<\/strong><\/a><\/p><h3 id=\"report-errata-and-feedback-1\"><a href=\"https:\/\/github.com\/mml-book\/mml-book.github.io\/issues\">Report errata and feedback<\/a>.<\/h3><\/section><\/div>","created_at":"2018-11-07T16:39:53+0000","updated_at":"2018-11-07T16:39:53+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"mml-book.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":220,"title":"Best Image Sizes and How to Save Images For the Web (2018) - Flothemes","url":"https:\/\/flothemes.com\/flothemes-image-sizes\/","content":"<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19576\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web.jpg\" alt=\"Image sizes for web\" width=\"1500\" height=\"1050\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web.jpg 1500w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web-300x210.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web-768x538.jpg 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web-1024x717.jpg 1024w\" \/><\/p><h2>What are the best size images for the web and Flothemes?<\/h2>\n<p>When uploading images to the web there are a number of things to take into consideration, such as what is the image being used for, how large should the file be, and what type of image should be used for different items on your site. Remember when working with images on the web make sure to <a href=\"https:\/\/flothemes.com\/ultimate-guide-saving-images-web\/\" target=\"_blank\" rel=\"noopener\">optimize<\/a> them to increase loading time.<\/p>\n<h3>Image Use and Type<\/h3>\n<p>There are various cases where you will want to use different image sizes for different areas of your site, below I will cover the most common areas and the sizing options you can use for each. Note this is recommended sizing options, they can be increased or decreased depending on your site.<\/p>\n<h2>Slideshows<\/h2>\n<p>There are various slideshow types, <a href=\"http:\/\/demo.flothemes.com\/fiji\/\" target=\"_blank\" rel=\"noopener\">full screen<\/a>,\u00a0<a href=\"http:\/\/demo.flothemes.com\/rebecca\/home-2\/\" target=\"_blank\" rel=\"noopener\">simple slideshows<\/a>, <a href=\"http:\/\/demo.flothemes.com\/cozy\/\" target=\"_blank\" rel=\"noopener\">full width fixed height<\/a>, and <a href=\"http:\/\/demo.flothemes.com\/porto\/home-6\/\" target=\"_blank\" rel=\"noopener\">slick slideshows<\/a>. However we\u2019ll put them into 2 classes, full screen and regular.<\/p>\n<h3>Full Screen Slideshow<\/h3>\n<p><strong>2880 x 1500px<\/strong><\/p>\n<p>As a browser screen has a ratio of ~2:1 you\u2019ll need to select images in a different ratio that you normally shoot. You can select 5-10 images and then you should crop them to the sizing above to get the most out of your images with minimal cropping.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19569\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow.png\" alt=\"Full Screen Slideshow\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow-1024x643.png 1024w\" \/><\/p>\n<h3>Regular\u00a0Slideshow<\/h3>\n<p><strong>Width Auto x 1500px<\/strong><\/p>\n<p>For other slideshows, you can save all images to 1500px in height as width will always be auto.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19571\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow.png\" alt=\"Regular Slideshow\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow-1024x643.png 1024w\" \/><\/p>\n<h2>Gallery<\/h2>\n<p><strong>Width Auto x 1500px<\/strong><\/p>\n<p>Much like slideshows, you can save all images to 1500px in height.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19570\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery.png\" alt=\"Gallery\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery-1024x643.png 1024w\" \/><\/p>\n<h2>Blog<\/h2>\n<p><strong>1500px x Height Auto<\/strong><\/p>\n<p>For blog posts you\u2019ll want to think about the width attribute, most will be around 7-800px in width so 1500px in width will work great.<\/p>\n<p class=\"with-image\"><img class=\"alignnone wp-image-46088 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1.jpg\" alt=\"Blog image size\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1.jpg 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1-300x189.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1-768x483.jpg 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1-1024x643.jpg 1024w\" \/><\/p>\n<h2>Featured Images<\/h2>\n<p>There are 3 types of featured images, portrait, landscape and hero images. The key to having good featured images is consistency. If you want to use all portrait make sure they are the same size, the same goes for landscape and hero images. If using a masonry grid, the sizing above is not so relevant, just make sure that the images are at least 900px in width.<\/p>\n<p><strong>Hero Image \u2013 2880 x 1500px<\/strong><\/p>\n<p>The hero images are those that are full screen, examples can be found in <a href=\"http:\/\/demo.flothemes.com\/porto\/2013\/01\/10\/markup-image-alignment\/\" target=\"_blank\" rel=\"noopener\">Porto<\/a>, and <a href=\"http:\/\/demo.flothemes.com\/como\/2016\/04\/06\/nature-is-my-favourite-dj\/\" target=\"_blank\" rel=\"noopener\">Como<\/a>.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19566\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image.png\" alt=\"Featured Images Hero Image\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image-1024x643.png 1024w\" \/><\/p>\n<p><strong>Landscape \u2013 1200 x \u00a0900px<\/strong><\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19567\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape.png\" alt=\"Featured Images Landscape\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape-1024x643.png 1024w\" \/><\/p>\n<p><strong>Portrait \u2013 900 x 1200px<\/strong><\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19568\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait.png\" alt=\"Featured Images Portrait\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait-1024x643.png 1024w\" \/><\/p>\n<h3>Recommended Image Sizes \u2013 Overview<\/h3>\n<p><strong>Slideshow Full Screen<\/strong> \u20132880\u00a0x 1500px<\/p>\n<p><strong>Regular\u00a0Slideshows<\/strong> \u2013 1500px height, width auto.<\/p>\n<p><strong>Galleries<\/strong> \u2013 1500px\u00a0<em>height, width auto<\/em>.<\/p>\n<p><strong>Blog Posts<\/strong> \u2013 1500px <em>width, height auto<\/em>. (If using a wide blog, use 2000px in width)<\/p>\n<p><strong>Featured Images<\/strong><\/p>\n<ul><li><em>Hero images (full screen) \u20132880\u00a0x 1500px.<\/em><\/li>\n<li>Portrait \u2013 900 x 1200<\/li>\n<li>Landscape \u2013 1200 x 900<\/li>\n<\/ul><p>If you\u2019re having any problems with image sizing make sure to take a look at the image sizing options in your Flothemes settings to make sure they are set up correctly.<\/p>\n<h2>How to save images for the web?<\/h2>\n<p>Saving images for the web has always been bit of a mystery for many. Uploading unoptimized images will not only create issues with the user experience on your site, but will also affect your SEO strategy (loading speed, bounce rate, ranking, etc). In this Ultimate Guide to Saving Images for the web we aim to show you how to inspect the various sections on your website to identify the correct sizes for your images before uploading them to the web. We will also explain how to export \/ save images correctly, so that they\u2019re optimized for speed.<\/p>\n<p>Finally we will provide some Image SEO tips, how to title your images correctly when saving them and why using ALT tags will improve your SEO. For already uploaded images to your site, we\u2019ll show you how to find which images are too large and make your site slow down. Let\u2019s begin!<\/p>\n<h2>1. Choosing the right image dimensions<\/h2>\n<p>One of the hardest parts about creating a website has always been selecting the right images, then making sure they are optimized and have the correct image dimensions for each section on your site. Thankfully, there are a few ways that can help you easily inspect your website images, and identify the correct sizing needed for each element on your site.<\/p>\n<p>An important thing to remember is that for retina devices, you need to double the image pixel density. What that means is, you need images twice the size for retina devices so they look nice and crisp. So once you\u2019ve found the correct image sizes for each section on your site, you\u2019ll want to double those sizes, and only then export and upload them to your site.<\/p>\n<h3>a. Using the Image Size Info Extension<\/h3>\n<p>The first method to find image sizes for your site is by using the<a href=\"https:\/\/chrome.google.com\/webstore\/detail\/image-size-info\/oihdhfbfoagfkpcncinlbhfdgpegcigf?hl=en\" target=\"_blank\" rel=\"noopener\"> Image Size Info<\/a> extension for Chrome (If you don\u2019t use Chrome on your computer, we highly recommend using it while creating your site). Once you\u2019ve installed the extension, this tool will then allow you to find out the image sizes of every image element on your site.<\/p>\n<p>Check out the images on your site, make notes, and then with this knowledge you\u2019ll be able to tailor your photographs exactly to fit the spaces as they are intended in the theme design. Just remember, with retina screens you do require your images to be twice as large as the intended space. So if you have an image space of 750x500px, for retina it should be 1500x1000px when uploading to WordPress (or any other platform).<\/p>\n<p>For example: when we inspect an image with this extension, it\u2019s being displayed at 780x520px, but it\u2019s actual size is 1560x1040px. This is because we\u2019re using a retina screen when inspecting the image. In this case, you can use the dimensions sizes to pick your image size.<\/p>\n<p>On a non-retina device the displayed and dimensions sizes will remain the same. In this case, double up the numbers and upload images at that size.<\/p>\n<p>The main limitation of this tool, is that it can\u2019t inspect background images for sizing. On many themes, background images are used instead of image HTML elements, as they are more flexible when displaying images. If you\u2019re unable to inspect the images because they are background images, you can use the next method to review image sizes on your site.<\/p>\n<h3>b. Using the Inspector Tool<\/h3>\n<p>With Chrome and Safari you can also use the developers inspector tool to find image sizes on the site. The principle will be the same as before, inspect the various elements on the site, and note down the sizes. Double the image sizes for retina devices and then upload them to WordPress, once they have been optimized for web and titled correctly for SEO.<\/p>\n<p>The limitations of this method will be that if the site is using some responsive elements set in REM values, images may be shrunk when using the inspector in the same window. If the site is also using responsive elements, they\u2019ll be larger on bigger screens, and smaller on others. It\u2019s best to use a larger screen where possible when checking the site, so you may get the most accurate image sizing. In any case, using this method will give you a good indication of sizes for your site and will allow you to approach adding images to your website with more care.<\/p>\n<h3>Chrome<\/h3>\n<p>Chrome has built in developer tools, which is why it\u2019s the most popular browser on the web for developers and designers. The following video explains how you can inspect the images on your site to find out their sizes:<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/VBWaX_F7R8E\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<h3>Firefox<\/h3>\n<p>With Firefox, much like chrome, they have a built in inspector tool that allows you to check images for your site.<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/eDSrj7eooxA\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<h3>Safari<\/h3>\n<p>If you\u2019re using Safari, you can use the inspector tool to see the sizes of images. Check the following guide on how to enable the developer tools.<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/eSgZqL1E-1I\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<p>Now you can go to different sections in your site and prepare the image dimensions.<\/p>\n<p><strong>For galleries<\/strong> you\u2019ll want to look at the height of the images, as they are generally constrained by height (so width can be set to auto).<br \/><strong>For blog posts<\/strong> you\u2019ll want to note the width for your images, as they are constrained by width (the height can be auto).<\/p>\n<p><strong>For the rest<\/strong> of the elements on your site, you should use the method described above to get the correct dimensions, and upload them at optimal dimensions for your site.<\/p>\n<p>Now that you know what dimensions you need for all the elements on your site, we\u2019ll explain how to optimize all your images for size, that\u2019s in terms of data sizing. The file size of your images is also very important for the web. Uploading bloated and heavy images on your website will strain your server resources, slow down your site, and in effect create a negative user experience.<\/p>\n<h2>2. Optimizing and Saving Images for web<\/h2>\n<p>There are a number of things to consider when optimizing and saving images for the web. You should consider what file type to use, JPG, PNG or GIF depending on the purpose. You\u2019ll want to label your images correctly for SEO purposes and you\u2019ll want to save them in the right color format (sRGB for web).<\/p>\n<p>Depending on the purpose, you need to save images in one of the following formats .jpg, .gif, or .png. JPG will be the most common format when saving images for your portfolio website, they will also be the lightest in size.<\/p>\n<p>For filenames, please make sure they are appropriate for each particular image or blog post. Use only letters, numbers, underscores and hyphens. Try to use only English letters. Characters from other languages, question marks, spaces, percent signs may upload incorrectly or cause unexpected behavior in your galleries or blog posts.<\/p>\n<p>For example, if you\u2019ve just done a boho wedding in The Biltmore Santa Barbara, label the images, boho-wedding-biltmore-santa-barbara_0001.jpg this will be much more SEO friendly, and could provide better ranking for your site.<\/p>\n<h3>Optimizing images in Lightroom<\/h3>\n<p>Let\u2019s face it, there are lots of different settings to choose and endless combinations of numbers to input. However, if you already know the image format and required file image dimensions, it\u2019s much easier.<\/p>\n<p>In Lightroom, you can get to the Export Dialog in 3 ways:<\/p>\n<h3>1. Highlight and Export<\/h3>\n<p>You can simply highlight \/ select any photo from your Lightroom library and right-click on the photo. From the pop up menu select the Export option.<\/p>\n<h3>2. File and Export<\/h3>\n<p>The second way is to go to <strong>File -&gt; Export<\/strong><\/p>\n<h3>3. Using the keyboard shortcut<\/h3>\n<p>The third, and fastest way is to use a keyboard shortcut:<\/p>\n<p><strong>shift + command + E (\u21e7+\u2318+E)<\/strong><\/p>\n<p>It doesn\u2019t matter how you get there. Any of these will work, so choose the easiest one for you to remember.<\/p>\n<h3>Lightroom Export Dialog<\/h3>\n<p>Now that you know how to export the images for web, we\u2019ll cover the 4 sections and the various options to choose in the Lightroom export dialog to allow you to export optimized images for the web.<\/p>\n<p class=\"with-image\"><img class=\"aligncenter wp-image-30663 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom.png\" alt=\"Lightroom export dialog\" width=\"850\" height=\"902\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom.png 850w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom-283x300.png 283w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom-768x815.png 768w\" \/><\/p>\n<h4>1. File Naming<\/h4>\n<p>Each export can feature its own rename settings. If you stick to the default option, your images will have the same name as the original files, and saved as a JPEG image. However we would highly recommend that you tick the \u201crename to\u201d box and choose one of the many renaming templates that Lightroom offers to customize the exported images. In the File Naming, we renamed the image filenames to \u201cCustom Name \u2013 Sequence\u201d where the Custom Text is closely connected to our blog post title or blog post keywords.<\/p>\n<p class=\"with-image\"><img class=\"aligncenter wp-image-30669 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/file-naming-lightroom.png\" alt=\"Lightroom file naming\" width=\"678\" height=\"198\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/file-naming-lightroom.png 678w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/file-naming-lightroom-300x88.png 300w\" \/><\/p>\n<h4>2. In File Settings<\/h4>\n<p>Make sure that your Image Format is set to JPEG. This is the most web-friendly option. Your Quality should not be less than 65. Normally we set it to 75. Your Colour Space needs to be sRGB. This is the color space used by the web so it will render the best. <strong>DO NOT<\/strong> check the box \u201cLimit File Size to:\u201d. Checking that box would decrease file quality significantly.<\/p>\n<p>To summarize, when saving you images use these settings:<\/p>\n<ul><li>75% Quality<\/li>\n<li>.jpeg format (.png for logos and icons)<\/li>\n<li>sRGB colors<\/li>\n<\/ul><h4>3.<strong>\u00a0Image Sizing<\/strong><\/h4>\n<p>Check the box Resize to Fit and select Width &amp; Height. In the box W: H: specify W (Width in pixels) to the value needed (1500px for blog posts, and 2880 for slideshows). By leaving the value in H (height) blank it will be set to auto sizing.<\/p>\n<p>For Galleries, you can leave W as blank, and set H (Height) to 1500px, this should be optimal for most themes.<\/p>\n<p>The sizes above are already optimized for retina devices.<\/p>\n<h4>4. Output Sharpening<\/h4>\n<p>Check the box next to \u201cSharpen For\u201dand select the amount to \u201cHigh\u201d.<\/p>\n<p>You can now Export the images, and follow this process anytime you\u2019re exporting images.<\/p>\n<h3>Image Quality Settings<\/h3>\n<p>In a photographer\u2019s dream world, we would always export images at maximum quality. In reality, we find out that maximum quality exports aren\u2019t required. At web sizes, the differences in quality are often hard to perceive, and that means we can reduce the quality slightly, reducing the file size significantly.<\/p>\n<p>As long as the images aren\u2019t soft, you\u2019re clients won\u2019t know the difference between what picture you took and the image thats displayed on your site.<\/p>\n<p>This also applies to colors of images. Everyone\u2019s screen is calibrated differently, so worrying about small color variations between an image in a browser and in your editing tool of choice is a waste of time, unless it ruins the image quality.<\/p>\n<p>On the upper end of the quality scale, there are diminishing returns. The difference in perceived quality between an 75% and 100% quality setting is hard to see, but it increases the file size significantly.<\/p>\n<p class=\"with-image\"><img class=\"alignnone wp-image-30666 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart.jpg\" alt=\"image size optimization\" width=\"1600\" height=\"856\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart.jpg 1600w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart-300x161.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart-768x411.jpg 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart-1024x548.jpg 1024w\" \/><\/p>\n<p>Given the above results, selecting the export of 65-75 quality. It strikes the correct balance of quality and optimal image file size.<\/p>\n<p>Now it\u2019s your turn to, rename your files, export them for the web, save them for web and you\u2019ll have some really well optimized images for the web.<\/p>\n<h3>Bonus Step<\/h3>\n<p>There is one final step that can help drastically remove excess data from images, and reducing file size, is using tools such as\u00a0<a href=\"https:\/\/flothemes.com\/faster-website-jpegmini\/\" target=\"_blank\" rel=\"noopener\">JPEG Mini<\/a>\u00a0or <a href=\"https:\/\/tinypng.com\/\" target=\"_blank\" rel=\"noopener\">TinyPNG<\/a>.<\/p>\n<p>If you have JPEG Mini pro, you can add the <a href=\"http:\/\/help.jpegmini.com\/knowledge_base\/topics\/how-to-activate-lightroom-plugin-on-a-mac\" target=\"_blank\" rel=\"noopener\">lightroom extension<\/a>, optimizing images as you export them, simple!<\/p>\n<p>If not, just run your freshly exported images through JpegMini or TinyPNG.<\/p>\n<p>This will be the final step in creating fast web optimized images to keep your site\u2019s speed up, and images looking great!<\/p>\n<h2>3. Image SEO<\/h2>\n<p>Now that you\u2019ve got your images sized, optimized and uploaded to your site \u2013 now you\u2019ll want to add ALT tags! This is the final step in creating optimized images for SEO purposes. The ALT tag is what search engines use to read images. Without them, your images are pretty much non-existent on the web.<\/p>\n<p>Thankfully with WordPress you can go into your media library after upload, and label each image. One great plugin that will let you do this quickly is <a href=\"https:\/\/wordpress.org\/plugins\/wp-meta-seo\/\" target=\"_blank\" rel=\"noopener\">WP Meta SEO<\/a>, when added you can go to WP Meta SEO &gt; Image Information and add ALT Tags to any images where it isn\u2019t present, once completed the plugin can be de-activated:<\/p>\n<p class=\"with-image\"><img class=\"alignnone wp-image-43439 size-large\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-1024x355.jpg\" alt=\"alt tag wp plugin\" width=\"1024\" height=\"355\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-1024x355.jpg 1024w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-300x104.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-768x266.jpg 768w\" \/><\/p>\n<p>There are also several plugins that will add the alt tags automatically, and if you\u2019ve went ahead and renamed the images like we mentioned in the previous section, you\u2019ll have pretty well optimized ALT text.\u00a0<a href=\"https:\/\/wordpress.org\/plugins\/automatic-image-alt-attributes\/\" target=\"_blank\" rel=\"noopener\">Automatic Image ALT Attributes<\/a>\u00a0is a good plugin solution for this, simply install it and it will automatically add ALT tags upon upload from the image description, or title.<\/p>\n<p>So that\u2019s it folks. The 3 steps to having optimized images for your website.<\/p>\n<p>Start with figuring out your image dimensions, then save them correctly for the web, add proper titles and optimize their size. Finally, add alt tags to improve SEO via WordPress (It should be possible to add ALT tags in all content management systems, refer to the documentation to find out how, or reach out to support).<\/p>\n<h3>Finding Unoptimized Images<\/h3>\n<p>If you already uploaded images to your site and want to know how to find the ones that are slowing down your page, check out the following video. Tools such as GTMetrix or Pingdom, or using the browser console will easily let you find images that are too large for the web.<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/Dd5RnMbT2G0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<p>If you would like to learn more about optimizing your site for speed, check out this comprehensive <a href=\"https:\/\/flothemes.com\/flo-learn\/speed-optimization-for-wordpress\/\" target=\"_blank\" rel=\"noopener\"><strong>online course.<\/strong><\/a><\/p>\n<h3>Flothemes<\/h3>\n<h3>Empowering You<\/h3>","created_at":"2018-11-07T16:39:52+0000","updated_at":"2018-11-07T16:39:52+0000","published_at":"2015-09-11T15:49:30+0000","published_by":["Allen Mark"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":13,"domain_name":"flothemes.com","preview_picture":"https:\/\/flothemes.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":219,"title":"Some Thoughts on Generative Art \u00b7 inconvergent","url":"https:\/\/inconvergent.net\/thoughts-on-generative-art\/","content":"<div class=\"text\"><p>\nThe following is a combination of segments from a few different talks I've\ngiven . It touches\nbriefly on the history of generative art, as well as some of my experience\nwith it.\n<\/p><p>\nIf you are looking for a comprehensive introduction to generative art, you\nwill unfortunately have to keep looking, but I want to mention some things\nI wish I had been aware of earlier.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/head.jpg\" alt=\"Sand Splines\" title=\"Sand Splines\" \/><p><a href=\"https:\/\/inconvergent.net\/generative\/sand-spline\/\">Sand Splines<\/a><\/p><\/div><div class=\"text\"><p>\nAccording to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Generative_art\">Wikipedia<\/a>,\ngenerative art is made by a process that has some level of autonomy.\nThe role of the artist is to design or influence this process to some\ndegree. This process can be any number of things, and as such it is not\nlimited to the digital realm. It can just as easily be a system driven by a\nchemical reaction. Or a set of rules for coloring in a checkered sheet of\npaper.\n<\/p><p>\nWhen I use the word <em>generative<\/em> I tend to mean that some kind of\nrandomness is involved, however I do not believe this to be a requirement\n. I have little interest\nin a strict definition of these terms, so instead I will just mention that\nthere are several terms that seem to be used be people who work within\ndifferent disciplines. Often they refer to a set of intersecting ideas . This includes, but is\nnot limited to:\n<\/p><ul><li>Generative<\/li>\n<li>Algorithmic<\/li>\n<li>Procedural (Art)<\/li>\n<li>Procedural Generation (PCG\/ProcGen)<\/li>\n<li>Parametric or Computer-Aided Design<\/li>\n<li>Computational Art\/Architecture\/Design<\/li>\n<li>Computer (Art)<\/li>\n<\/ul><p>\nIn some sense I would argue that it does not really matter whether you are\ngenerating characters, dialogue or environments for a computer game,\ngeometry as part of a building, digital or analogue art, a musical\ncomposition, or a poem. In all cases you are using a system of some kind to\nhelp you along. Either towards a specific goal, some unidentified result,\nor somewhere in-between.\n<\/p><p>\nTo me the main point is usually to experiment with a small system that\nconsists of a set of relatively simple rules. Sometimes the goal is to\nrecreate a particular behaviour\u2014for instance a biological process or\nsome other natural phenomena\u2014sometimes not.\n<\/p><p>\nMy process is usually highly experimental. And while there is usually some\nkind of underlying goal there is not always any interest on my part in\nactually reaching that particular goal. I might be just as happy to\ndiscover some unexpected behaviour and explore that instead. Overall this\nmeans that the process is continuous, exploratory, and never really\ncomplete. Naturally this process has also changed quite a lot over time.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/nake-klee.jpg\" alt=\"Hommage \u00e0&#10;  Paul Klee\" title=\"Hommage \u00e0 Paul Klee\" \/><p>Hommage \u00e0 Paul Klee, Frieder Nake 1965<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nAt this point I have had some kind of interest in generative art for at\nleast ~ 7 years . I\nstarted working actively on it when I was at university, but really it\nstarted before that.\n<\/p><p>\nWhen I was about 16 I spent a year as an exchange student in Australia. At\nsome point I came across <a href=\"http:\/\/complexification.net\">complexification.net<\/a> by Jared\nTarbell. At the time I didn't really know enough about programming to\nrealise how it was done, or how to get started. As a result I mostly just\nspent a whole lot of time looking at the various animations. Then I\ngradually forgot about the site.\n<\/p><p>\nI'm not sure what reminded me of the site again, but I had learnt quite a\nlot of programming, so I started experimenting. Seeing as I was studying\nphysics and numerical mathematics it turned out I had a pretty good\nfoundation for this kind of programming. However, <em>because<\/em> I was\nstudying numerical mathematics I almost felt obliged to model some kind of\n<em>real thing<\/em>. Recently I've started to realise that there is no\nparticular reason why I should impose this limitation on my work.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/nake-walk-through-raster.jpg\" alt=\"Walk Through Raster\" title=\"Walk Through Raster\" \/><p>Walk Through Raster, Frieder Nake 1966<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nI got into generative art as a distraction at university, but people have\nbeen doing this for far longer. Among the people frequently credited with\nbeing part of the advent of computer art is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Frieder_Nake\">Frieder Nake<\/a>. Two\nothers include <a href=\"https:\/\/en.wikipedia.org\/wiki\/A._Michael_Noll\">Michael Noll<\/a> and\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Georg_Nees\">Georg Nees<\/a>.\n<\/p><p>\nUnbeknownst to each other these people had started making computer\ngenerated plotter drawings based on randomness and\/or rules in 1963-64.\nThen, in 1965, followed three separate exhibitions: Georg Nees exhibited\nwork in Stuttgart in February, Noll and <a href=\"https:\/\/en.wikipedia.org\/wiki\/B%C3%A9la_Julesz\">B\u00e9la Julesz<\/a> had\nan exhibition in New York City in April, and finally, Nake and Nees\nexhibited work in Stuttgart in November. If you would like a more authentic\nintroduction to this history I can not recommend Frieder Nake's <a href=\"https:\/\/vimeo.com\/104315361\">talk from Eyeo 2014<\/a> enough.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/molnar-desordres-3.jpg\" alt=\"(Des)Ordres\" title=\"(Des)Ordres\" \/><p>(Des)Ordres, Vera Moln\u00e1r 1974<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nThere are two other people who's work I also find thoroughly captivating.\nOne is pioneering artist <a href=\"https:\/\/en.wikipedia.org\/wiki\/Vera_Moln%C3%A1r\">Vera Moln\u00e1r<\/a>,\nwho began producing computer art in 1968. The other one is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Manfred_Mohr\">Manfred Mohr<\/a>, an\naction painter and jazz musician, who started working with computers in\n1969.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/mohr-p62.jpg\" alt=\"P62\" title=\"P62\" \/><p>P62, Manfred Mohr 1970<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/mohr-p197.jpg\" alt=\"P197\" title=\"P197\" \/><p>P197, Manfred Mohr 1977-79<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nLooking at these old works is quite stunning. In some ways They are rather\nminimalistic. It's easy to be tricked into believing that this is because\nthere is a limit to what you can do with <a href=\"https:\/\/en.wikipedia.org\/wiki\/Punched_card\">punch cards<\/a> and\nplotter. However, this minimalism is an excellent way to let you really pay\nattention to the intricacies of the algorithms behind it. You can clearly\nsee that there is some sort of pattern there, but the details are not at\nall obvious.\n<\/p><p>\nInstead of trying to tell you how they are done, I'm simply going to say\nthat while we might have better tools, and more processing power today,\nthat does not automatically mean we can make better (generative) art. \n<\/p><p>\nIf you are interested in my process and work in particular you could have a\nlook at another text I've written: <a href=\"https:\/\/inconvergent.net\/generative\/\">On Generative\nAlgorithms<\/a>.\n<\/p><\/div><div class=\"text\"><div class=\"footnotes\"><hr \/><ol id=\"footnotes\"><li>\nNone of it is verbatim, as I am completely incapable of writing a\nmanuscript and sticking to it.\n<\/li>\n<li>\nTechnically I use <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pseudorandom_number_generator\">\nPseudorandom Number Generators (PRNG)<\/a>. Seeing as these generators\nproduce deterministic results, there is no real difference between a\nprogram that includes this kind of randomness and a program that does\nnot use any randomness at all. There would be a conceptual difference\nif the random numbers used were <a href=\"https:\/\/www.random.org\/\">truly\nrandom<\/a>, but in the case of my work I am almost entirely certain\nthat you would not be able to inspect a work and decide whether it was\nmade using true randomness or a PRNG.\n<\/li>\n<li>\nI don't mean to diminish the importance of definitions, and there may\nvery well be a well formulated definition of some (or all) of these\nterms. However, if you are getting started it is useful to know the\nterms people use, and then you can look into them on your own.\n<\/li>\n<li>\nIt seems I registered this domain on the 5th of August 2010. This\nlends some doubt to my recollection that I registered this domain when\nI was supposed to be reading for my exams; the summer semester at my\nuniversity usually had exams in May. Unless I was reading for\ncontinuation exams, which did happen several times.\n<\/li>\n<li>\nInfo in this section is from the Wikipedia article about <a href=\"https:\/\/en.wikipedia.org\/wiki\/Frieder_Nake\">Frieder Nake<\/a>\nand <a href=\"https:\/\/vimeo.com\/104315361\">his presentation at Eyeo in\n2014<\/a>.\n<\/li>\n<li>\nNaturally we can, and should, try.\n<\/li>\n<\/ol><\/div><\/div>","created_at":"2018-11-07T16:39:50+0000","updated_at":"2018-11-07T16:39:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"inconvergent.net","preview_picture":"https:\/\/inconvergent.net\/img\/writing\/thoughts\/head.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":216,"title":"Unlearning toxic behaviors in a code review culture","url":"https:\/\/medium.freecodecamp.org\/unlearning-toxic-behaviors-in-a-code-review-culture-b7c295452a3c?gi=c1298e77c413","content":"<p id=\"f81c\" class=\"graf graf--p graf-after--figure\">Code reviews can be contentious. I had an incredible experience giving my first conference talk titled \u201c<a href=\"https:\/\/alterconf.com\/talks\/unlearning-toxic-behaviors-code-review-culture\" data-href=\"https:\/\/alterconf.com\/talks\/unlearning-toxic-behaviors-code-review-culture\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Unlearning Toxic Behaviors in a Code Review Culture<\/a>\u201dat <a href=\"https:\/\/alterconf.com\/\" data-href=\"https:\/\/alterconf.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">AlterConf<\/a>. I went into the experience preparing for any pushback and criticism that I might receive while encouraging software teams to create supportive code review environments.<\/p><p id=\"b40a\" class=\"graf graf--p graf-after--p\">But I totally underestimated how <strong class=\"markup--strong markup--p-strong\">supportive and receptive <\/strong>the AlterConf community is. I worried about pushback, but that community greeted me with kindness and encouragement instead. \u2764<\/p><figure id=\"9ab9\" class=\"graf graf--figure graf--iframe graf-after--p\"><figcaption class=\"imageCaption\">Twitter was extremely supportive.<\/figcaption><\/figure><p id=\"bb2e\" class=\"graf graf--p graf-after--figure\">I\u2019ve been asked to share my slides (<a href=\"https:\/\/speakerdeck.com\/sandyaaaa\/unlearning-toxic-behaviors-in-a-code-review-culture\" data-href=\"https:\/\/speakerdeck.com\/sandyaaaa\/unlearning-toxic-behaviors-in-a-code-review-culture\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">and I have<\/a>), <strong class=\"markup--strong markup--p-strong\">but I now worry that the slides alone are unhelpful and easily taken out of context<\/strong>, because the deck lacks context and explanations. Since my slides are floating around the internet without these key features, I decided to do a write-up to elaborate upon my talking points. The <a href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" data-href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">video recording<\/a> was released a few months after this article was published.<\/p><p id=\"fa5a\" class=\"graf graf--p graf-after--p\">I\u2019ve listed below <strong class=\"markup--strong markup--p-strong\">some common unhelpful behaviors<\/strong> people display during code reviews, and <strong class=\"markup--strong markup--p-strong\">some recommendations<\/strong> on how software teams can make their teams more supportive by refusing to normalize toxicity. All of the behaviors I describe were either witnessed by me or happened to an industry contact of mine. I\u2019ve been guilty of several of these behaviors in the past, too.<\/p><h3 id=\"3f0a\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #1: passing off opinion as\u00a0fact<\/h3><p id=\"bdb5\" class=\"graf graf--p graf-after--h3\">Don\u2019t make claims unless you can cite documentation, formalized guidelines, and coding examples to back those claims up. People need to know why they are being asked to make a change, and another developer\u2019s personal preference isn\u2019t a good enough argument.<\/p><p id=\"12ee\" class=\"graf graf--p graf-after--p\">Instead of saying:<\/p><blockquote id=\"a6d4\" class=\"graf graf--blockquote graf-after--p\"><div>This component should be stateless.<\/div><\/blockquote><p id=\"3608\" class=\"graf graf--p graf-after--blockquote\">\u2026provide some context behind the recommendation:<\/p><blockquote id=\"cad9\" class=\"graf graf--blockquote graf-after--p\"><div>Since this component doesn\u2019t have any lifecycle methods or state, it could be made a stateless functional component. This will improve performance and readability. *Here* is some documentation.<\/div><\/blockquote><p id=\"8a4b\" class=\"graf graf--p graf-after--blockquote\">Passing off opinion as fact is also common when developers discuss style and syntax preferences. These are really important discussions to have, but never in a code review, because style and syntax have nothing to do with the issue the developer was initially trying to solve.<\/p><p id=\"dbb2\" class=\"graf graf--p graf-after--p\">Have these discussions separately and decide which style guidelines you will be following as a group. Implement a <a href=\"http:\/\/www.jslint.com\/\" data-href=\"http:\/\/www.jslint.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">linter<\/a> and an <a href=\"https:\/\/atom.io\/packages\/atom-beautify\" data-href=\"https:\/\/atom.io\/packages\/atom-beautify\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">automatic code fixer<\/a>. This way, you can reference your style guidelines, and not your personal opinion, during the code review when you see inconsistencies.<\/p><p id=\"4211\" class=\"graf graf--p graf-after--p\">It is especially important not to pass off your opinion as fact when you have higher rank and authority within your team or company. If you do this, developers will feel they have no choice but to quietly implement your demands.<\/p><h3 id=\"50a0\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #2: overwhelming with an avalanche of\u00a0comments<\/h3><p id=\"b55a\" class=\"graf graf--p graf-after--h3\">When a person makes an error, chances are high that they have made the same error in several places in their changeset. I have noticed that reviewers sometimes point out every single one of an error\u2019s many occurrences instead of leaving one detailed note with links to helpful resources.<\/p><p id=\"611b\" class=\"graf graf--p graf-after--p\">Consolidating comments allows you to convey the same message without overwhelming the review seeker. An avalanche of comments for<em class=\"markup--em markup--p-em\"> one problem, duplicated,<\/em> comes off as nitpicking.<\/p><p id=\"b37f\" class=\"graf graf--p graf-after--p\">Unhelpful and overwhelming:<\/p><figure id=\"6e99\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*IOG-1oW1w6b1BkOZ.\" data-width=\"800\" data-height=\"727\" data-action=\"zoom\" data-action-value=\"0*IOG-1oW1w6b1BkOZ.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*IOG-1oW1w6b1BkOZ.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Leaving multiple comments for \u201cone error, duplicated\u201d (trailing whitespaces)<\/figcaption><\/figure><p id=\"f295\" class=\"graf graf--p graf-after--figure\">More helpful:<\/p><figure id=\"33f1\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*ivhV5hZXN6nSkyED.\" data-width=\"800\" data-height=\"158\" data-action=\"zoom\" data-action-value=\"0*ivhV5hZXN6nSkyED.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*ivhV5hZXN6nSkyED.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Consolidating feedback<\/figcaption><\/figure><p id=\"907f\" class=\"graf graf--p graf-after--figure\">I can understand the argument that pointing out every place an error occurs on a PR can sometimes be useful, since the comment disappears when the developer resolves the issue in subsequent commits. However, after this error has occurred across a changeset, it should be clear that the developer was not aware of a certain guideline and simply needs to be pointed to the correct resources.<\/p><h3 id=\"ccc9\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #3: asking engineers to solve problems they didn\u2019t cause \u201cwhile they are at\u00a0it\u201d<\/h3><p id=\"da78\" class=\"graf graf--p graf-after--h3\">Avoid asking developers to solve issues that aren\u2019t directly related to their changeset or the problem their changeset is trying to solve. Even if a developer is extending or modifying a messy part of the code that is rife with bad practices, don\u2019t ask the developer to fix them in that pull request just because their changeset happens to touch the messy code.<\/p><p id=\"2531\" class=\"graf graf--p graf-after--p\">I am <strong class=\"markup--strong markup--p-strong\">not <\/strong>suggesting that developers should feel no accountability for code just because they didn\u2019t initially introduce it. In fact, it is not collaborative to say something like, \u201cnot my code, not my problem.\u201d I am just suggesting that it would be more appropriate to create a separate ticket and PR to address the messy code. This way, you avoid drastically increasing the scope of someone\u2019s work, but the messy code still gets fixed.<\/p><p id=\"5c12\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">TL;DR: <\/strong>Don\u2019t ask devs to fix problems \u201cwhile they are at it.\u201d If the code solves the issue\/ticket the dev was working on and doesn\u2019t introduce any new issues to the codebase, give the pull request a \u201cthumbs up\u201d and then create a ticket to clean up the bad code.<\/p><h3 id=\"1ee0\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #4: asking judgmental questions<\/h3><p id=\"7338\" class=\"graf graf--p graf-after--h3\">Avoid asking judgmental questions like:<\/p><blockquote id=\"644b\" class=\"graf graf--blockquote graf-after--p\"><div>Why didn\u2019t you just do ___ here?\u201d<\/div><\/blockquote><p id=\"f52c\" class=\"graf graf--p graf-after--blockquote\">Asking such questions implies that a perceived simple solution should have been obvious. It also forces developers to have to defend themselves.<\/p><p id=\"bd12\" class=\"graf graf--p graf-after--p\">Oftentimes, these judgmental questions are just veiled demands. Instead, provide a recommendation (with documentation and citations) and leave out harsh words.<\/p><blockquote id=\"d652\" class=\"graf graf--blockquote graf-after--p\"><div>You can do ___, which has the benefit of ____.<\/div><\/blockquote><h3 id=\"b0af\" class=\"graf graf--h3 graf-after--blockquote\">Unhelpful behavior #5: being sarcastic<\/h3><p id=\"036d\" class=\"graf graf--p graf-after--h3\">There is no appropriate time to be sarcastic when offering someone feedback. Sarcastic comments tend not to provide context or actionable feedback. Instead, describe the issue with details and provide recommendations, but leave the caustic jokes out.<\/p><p id=\"db33\" class=\"graf graf--p graf-after--p\">Unhelpful:<\/p><blockquote id=\"d125\" class=\"graf graf--blockquote graf-after--p\"><div>Did you even test this code before you checked it in?<\/div><\/blockquote><p id=\"f396\" class=\"graf graf--p graf-after--blockquote\">Helpful:<\/p><blockquote id=\"a661\" class=\"graf graf--blockquote graf-after--p\"><div>This breaks when you enter a negative number. Can you please address this case?<\/div><\/blockquote><p id=\"6638\" class=\"graf graf--p graf-after--blockquote\">Here is another example of a <a href=\"https:\/\/techblog.workiva.com\/tech-blog\/we-are-ruthless-code-reviews\" data-href=\"https:\/\/techblog.workiva.com\/tech-blog\/we-are-ruthless-code-reviews\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">code review comment that isn\u2019t funny or helpful<\/a>:<\/p><blockquote id=\"0618\" class=\"graf graf--blockquote graf-after--p\"><div>I don\u2019t mean we\u2019re mean-spirited. I just mean that we are merciless. You\u2019ll notice that I left the comment \u201cBeep!\u201d on the imports of every file you touched. What I meant was, \u201cYour imports violate our standard convention\u200a\u2014\u200awe order them by built-ins, then third party, and then project level,\u201d but that was too much to type on every file.<\/div><\/blockquote><p id=\"db37\" class=\"graf graf--p graf-after--blockquote\">In the above example, \u201cbeep!\u201d is not helpful or descriptive. This practice is simply pedantic humor that does nothing to better the review seeker.<\/p><h3 id=\"e26a\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #6: using emojis instead of statements to point out\u00a0issues<\/h3><p id=\"391b\" class=\"graf graf--p graf-after--h3\">Avoid using the thumbs-down or puke emoji to point out issues in code. This is as unhelpful as sarcasm for similar reasons. Emojis are cryptic and easy to misconstrue. Emojis waste peoples\u2019 time as they try to figure out what you mean.<\/p><figure id=\"47fa\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*KcY0E6xRJ9Nt73vr.\" data-width=\"1060\" data-height=\"166\" data-action=\"zoom\" data-action-value=\"0*KcY0E6xRJ9Nt73vr.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*KcY0E6xRJ9Nt73vr.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Don\u2019t use emojis to point out coding\u00a0problems<\/figcaption><\/figure><p id=\"85d1\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">You shouldn\u2019t be having visceral reactions to peoples\u2019 coding mistakes anyway.<\/strong><\/p><p id=\"a0e3\" class=\"graf graf--p graf-after--p\">It\u2019s okay to use emojis like \u201cthumbs-up\u201d or \u201chooray\u201d to signify that code looks good, but don\u2019t use them to point out problems.<\/p><figure id=\"d301\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*zNh3gFLoNNv4kGXI.\" data-width=\"1032\" data-height=\"170\" data-action=\"zoom\" data-action-value=\"0*zNh3gFLoNNv4kGXI.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*zNh3gFLoNNv4kGXI.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Emojis are great when used to\u00a0praise<\/figcaption><\/figure><h3 id=\"310f\" class=\"graf graf--h3 graf-after--figure\">Unhelpful behavior #7: not replying to all\u00a0comments<\/h3><p id=\"b506\" class=\"graf graf--p graf-after--h3\">Review seekers can contribute to unsupportive environments, too. If you merge code without addressing all the feedback, people are left wondering why they bothered to help you, and you send the message that some opinions are worth more than others.<\/p><p id=\"8fe7\" class=\"graf graf--p graf-after--p\">If a comment is out of scope or you won\u2019t be taking action on the feedback, just provide a brief note explaining why.<\/p><p id=\"43a1\" class=\"graf graf--p graf-after--p\">Don\u2019t ghost people.<\/p><figure id=\"f09e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*6Mt-1qXjBnetnOQB.\" data-width=\"800\" data-height=\"287\" data-action=\"zoom\" data-action-value=\"0*6Mt-1qXjBnetnOQB.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*6Mt-1qXjBnetnOQB.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Merging code without addressing all the\u00a0feedback<\/figcaption><\/figure><h3 id=\"0f11\" class=\"graf graf--h3 graf-after--figure\">Unhelpful behavior #8: ignoring toxic behaviors from high performers<\/h3><p id=\"5f1d\" class=\"graf graf--p graf-after--h3\">Toxic behaviors should not be ignored or deemphasized because a developer is a high performer and extremely productive. Though this developer might be doing a fantastic job, it is important to keep in mind that this developer\u2019s toxic behaviors make them draining and stressful to work with for other developers.<\/p><p id=\"6c16\" class=\"graf graf--p graf-after--p\">On working with developers who commonly display toxic behaviors:<\/p><blockquote id=\"f666\" class=\"graf graf--pullquote graf-after--p\"><div>Others will find working with this individual draining and demotivating. They\u2019ll find themselves going to great lengths to avoid interacting with them, even if it means negatively impacting their own ability to complete tasks. Communication will shut down throughout your organization. If it becomes bad, your team will start looking for opportunities elsewhere. While you deal with the fallout of the exodus of talent and failing projects, this particular kind of developer will happily continue working as if nothing was wrong.\u200a\u2014\u200a<a href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" data-href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" class=\"markup--anchor markup--pullquote-anchor\" target=\"_blank\">Joseph\u00a0Gefroh<\/a><\/div><\/blockquote><p id=\"8259\" class=\"graf graf--p graf-after--pullquote\">Not taking steps to unlearn these toxic behaviors can have serious setbacks for the team, as developers feel overwhelmed, attacked, and demotivated. They start to dread the feedback processes that are supposed to help them grow.<\/p><p id=\"28fd\" class=\"graf graf--p graf-after--p\">I would personally feel a lot of anxiety whenever I got an email that a previous coworker (who was known for delivering toxic, unhelpful feedback) had commented on my pull request. Though toxic behaviors affect everyone differently, we can all agree that no one benefits from this toxicity.<\/p><p id=\"ce73\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Note:<\/strong> I want to be clear that slipping up and displaying one of the above toxic behaviors alone doesn\u2019t make a developer \u201ctoxic.\u201d Repeat offenses and patterns of caustic feedback say a lot, though.<\/p><h3 id=\"9d46\" class=\"graf graf--h3 graf-after--p\">Helpful Code Review Practices<\/h3><p id=\"5ea8\" class=\"graf graf--p graf-after--h3\">The following are some recommendations that can apply to any person in a collaborative, feedback-driven environment, even if they are filtered through the lens of a software engineer\u2019s eyes and are in the context of code reviews.<\/p><h3 id=\"3374\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #1: use questions or recommendations to drive\u00a0dialog<\/h3><p id=\"efb0\" class=\"graf graf--p graf-after--h3\">Never make demands for people to implement changes or ask judgmental questions, because this does not open up a dialog between you and the person whose code you are reviewing.<\/p><blockquote id=\"1ee7\" class=\"graf graf--blockquote graf-after--p\"><div>Why didn\u2019t you just pull these translations into a constants file?<\/div><\/blockquote><p id=\"2c4f\" class=\"graf graf--p graf-after--blockquote\">Though the above is a question, it doesn\u2019t create a dialog between you and your review seeker. It just forces them to defend themselves. Instead, ask what the review seeker thinks of your proposed solution, like:<\/p><blockquote id=\"dbe2\" class=\"graf graf--blockquote graf-after--p\"><div>What do you think about pulling these translations into a constants file? There are a lot, and a separate file might make sense at this point<\/div><\/blockquote><p id=\"f9d9\" class=\"graf graf--p graf-after--blockquote\">\u2026or make a recommendation like:<\/p><blockquote id=\"4080\" class=\"graf graf--blockquote graf-after--p\"><div>You introduced a lot of translation calls for \u201cfeature x\u201d in this file. It might make sense to create a separate file devoted to \u201cfeature x\u201d constants.<\/div><\/blockquote><h3 id=\"d53f\" class=\"graf graf--h3 graf-after--blockquote\">Helpful Behavior #2: collaborate, don\u2019t back-seat drive<\/h3><p id=\"60fc\" class=\"graf graf--p graf-after--h3\">When you pair-program with another person, you should be there to ask questions, debate, and point to resources.<\/p><blockquote id=\"e2a4\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201c\u2026when you want to help out or work with others, you should fully engage and not just butt in sporadically\u201d\u200a\u2014\u200aRecurse Center\u2019s <a href=\"https:\/\/www.recurse.com\/manual#no-backseat-driving\" data-href=\"https:\/\/www.recurse.com\/manual#no-backseat-driving\" class=\"markup--anchor markup--pullquote-anchor\" rel=\"noopener\" target=\"_blank\">User\u00a0Manual<\/a><\/div><\/blockquote><h3 id=\"e08f\" class=\"graf graf--h3 graf-after--pullquote\">Helpful Behavior #3: respond to every\u00a0comment<\/h3><p id=\"1b37\" class=\"graf graf--p graf-after--h3\">If you as a review seeker don\u2019t plan to apply a person\u2019s feedback, just leave a note letting them know. Don\u2019t ignore those who take time to help you.<\/p><p id=\"0e2a\" class=\"graf graf--p graf-after--p\">For example,<\/p><blockquote id=\"8cfb\" class=\"graf graf--blockquote graf-after--p\"><div>Person A: What do you think about creating a helper function for this api call? Looks good to me otherwise<\/div><\/blockquote><blockquote id=\"cda1\" class=\"graf graf--blockquote graf-after--blockquote\"><div>Person B: That line wasn\u2019t part of my changeset. I\u2019ll merge this code for now, but I\u2019ll also create a separate GitHub issue for the api call and put it in our group\u2019s backlog of to-do items.<\/div><\/blockquote><h3 id=\"398e\" class=\"graf graf--h3 graf-after--blockquote\">Helpful Behavior #4: know when to take a discussion offline<\/h3><p id=\"2d99\" class=\"graf graf--p graf-after--h3\">After dozens of contrasting PR comments and proposed solutions, it should be clear that online communication has become unproductive for the issue at hand. Send a meeting hit to involved members to discuss the issue offline.<\/p><p id=\"c481\" class=\"graf graf--p graf-after--p\">This way, your group can come to a decision more quickly and apply the solution.<\/p><blockquote id=\"28b3\" class=\"graf graf--pullquote graf-after--p\"><div>The kinds of problems that take hours and tons of comments can usually be resolved in a few minutes of productive talk.\u200a\u2014\u200a<a href=\"http:\/\/tidyjava.com\/better-code-reviewer\/\" data-href=\"http:\/\/tidyjava.com\/better-code-reviewer\/\" class=\"markup--anchor markup--pullquote-anchor\" rel=\"noopener\" target=\"_blank\">Tidy\u00a0Java<\/a><\/div><\/blockquote><h3 id=\"3cd1\" class=\"graf graf--h3 graf-after--pullquote\">Helpful Behavior #5: use opportunities to teach, and don\u2019t show\u00a0off<\/h3><p id=\"d985\" class=\"graf graf--p graf-after--h3\">Before choosing to participate in a code review, ask yourself:<\/p><blockquote id=\"d6d4\" class=\"graf graf--blockquote graf-after--p\"><div>Is your comment helping the other developer learn or are you nitpicking to participate?<\/div><\/blockquote><p id=\"f607\" class=\"graf graf--p graf-after--blockquote\">Reflect on why you are participating. Remember that the goal behind code reviews is to teach and help other developers grow. The goal isn\u2019t to grandstand.<\/p><h3 id=\"8b2f\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #6: don\u2019t show\u00a0surprise<\/h3><p id=\"4aab\" class=\"graf graf--p graf-after--h3\">Be careful not to show surprise that someone doesn\u2019t have the same knowledge that you do on a topic. Being comfortable admitting you lack experience with a topic is a great way to ask for help. Don\u2019t make people feel bad that they \u201cshould already have known\u201d this piece of info.<\/p><p id=\"d0bd\" class=\"graf graf--p graf-after--p\">For more on this, see Recurse Center\u2019s Social Rule \u201c<a href=\"https:\/\/www.recurse.com\/manual#no-feigned-surprise\" data-href=\"https:\/\/www.recurse.com\/manual#no-feigned-surprise\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Don\u2019t Feign Surprise<\/a>\u201d<\/p><h3 id=\"e166\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #7: automate what can\u00a0be<\/h3><p id=\"1051\" class=\"graf graf--p graf-after--h3\">Reviewing issues that can be caught by linters, git hooks, or automated tests are unhelpful because they often result in an avalanche of comments and come off as nitpicking. People are not particularly good at catching these issues, which is why automation tools exist.<\/p><p id=\"501c\" class=\"graf graf--p graf-after--p\">There are also tools out there that will run your tests when code is checked in and display warnings that a changeset is in violation of any of these tests. TeamCity and Jenkins CI both offer this functionality.<\/p><p id=\"109a\" class=\"graf graf--p graf-after--p\">Also, use git hooks. Git hooks run tests and linters when someone tries to check code in and they will intercept the commit if any buggy, unlinted code was introduced.<\/p><p id=\"edc4\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Let the tool point out issues so people don\u2019t have to.<\/strong><\/p><h3 id=\"bd49\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #8: refuse to normalize toxic\u00a0behavior<\/h3><p id=\"25cc\" class=\"graf graf--p graf-after--h3\">Don\u2019t join in and contribute to code review toxicity because it seems like the status quo, a hazing ritual for new developers, or just the way things are.<\/p><p id=\"51c2\" class=\"graf graf--p graf-after--p\">Search for teammates who are supportive and who will back you up.<\/p><p id=\"3bbc\" class=\"graf graf--p graf-after--p\">If you notice someone being unhelpful during code reviews, consider letting them know (if you feel safe in your role\/company doing so) and be direct.<\/p><h3 id=\"2833\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #9: managers\u200a\u2014\u200ahire carefully, listen to your team, and\u00a0enforce<\/h3><p id=\"cfaa\" class=\"graf graf--p graf-after--h3\">Managers have a lot of power to create positive and supportive cultures on their teams.<\/p><p id=\"6574\" class=\"graf graf--p graf-after--p\">To paraphrase the advice from \u201c<a href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" data-href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Toxic Developers Considered Harmful<\/a>\u201d:<\/p><ul class=\"postList\"><li id=\"4fc6\" class=\"graf graf--li graf-after--p\">Prevent toxic developers from joining your team. During the hiring process, don\u2019t just stress the technical. Learn how your candidate collaborates and communicates. Critically analyze their work and see how they react. Make sure every hire is a culture add, not just a culture fit.<\/li><li id=\"29e7\" class=\"graf graf--li graf-after--li\">When toxic developers trickle in or you inherit them, ask your direct reports how their experience is working with others on the team during 1:1\u2019s. If you have a toxic dev, your direct reports\u2019 feedback will show it.<\/li><li id=\"7f36\" class=\"graf graf--li graf-after--li\">Talk to your toxic dev. Give them specific examples and actionable feedback. Work with them throughout your manager\/direct report relationship.<\/li><li id=\"6f30\" class=\"graf graf--li graf-after--li\">Don\u2019t isolate the toxic developer (*)<\/li><li id=\"aef4\" class=\"graf graf--li graf-after--li\">Iterate that you expect your team to contribute to a supportive environment<\/li><\/ul><p id=\"9666\" class=\"graf graf--p graf-after--li\">(*) Though the article suggests to isolate toxic developers, I believe it\u2019s important to encourage the toxic developer to continue to collaborate with the rest of the team, but in more healthy ways. Isolating toxic workers won\u2019t solve the problem. Though giving this dev independent work allows the team to suffer less toxicity in the short run, this toxic developer has not unlearned unhelpful behaviors. They\u2019ve just lost the platform to display these behaviors.<\/p><h3 id=\"8d86\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #10: set the standard as your team is small and\u00a0growing<\/h3><p id=\"a992\" class=\"graf graf--p graf-after--h3\">Teams that are small enough can adopt ideas and implement them from the start. Even if you don\u2019t think it\u2019s necessary to set a standard of what is and isn\u2019t okay because your team is currently collaborating well, remember that you want your team to stay great as you invite new members in.<\/p><h3 id=\"39b8\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #11: understand you might be part of the\u00a0problem<\/h3><p id=\"233b\" class=\"graf graf--p graf-after--h3\">In order to foster more supportive environments, it\u2019s important for people to be honest with themselves and reflect on any unhelpful behaviors they may have ever displayed.<\/p><p id=\"6d7c\" class=\"graf graf--p graf-after--p\">As a junior engineer, there were several occasions when I noticed a bug in someone\u2019s code and delighted, because it meant that I could point it out and be part of the review. I know now that I was using the opportunity to show off, not teach. I am trying to be mindful of this going forward.<\/p><p id=\"d336\" class=\"graf graf--p graf-after--p\">I think everyone should have this tough moment of self-reflection too.<\/p><p id=\"9942\" class=\"graf graf--p graf-after--p\">One last thing\u2026<\/p><h3 id=\"9802\" class=\"graf graf--h3 graf-after--p\">I am not policing the content of feedback, just asking people to be mindful of their\u00a0tone<\/h3><p id=\"9f28\" class=\"graf graf--p graf-after--h3\">I know feedback is important and I am not waging a war on feedback. I\u2019m also not asking teams to compromise their code quality. A supportive code review culture and high code quality DON\u2019T have to be mutually exclusive.<\/p><p id=\"16f0\" class=\"graf graf--p graf-after--p\">I simply hope people take steps to provide constructive, actionable feedback and create more supportive environments, so developers feel comfortable to learn, grow, and make mistakes. We can all improve together.<\/p><h4 id=\"c4d2\" class=\"graf graf--h4 graf-after--p\">TL;DR:<\/h4><figure id=\"3a79\" class=\"graf graf--figure graf--iframe graf-after--h4\"><p id=\"db79\" class=\"graf graf--p graf-after--figure graf--trailing\">Questions or comments? Feel free to reach out on <a href=\"https:\/\/twitter.com\/sandyaaaas\" data-href=\"https:\/\/twitter.com\/sandyaaaas\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Twitter<\/a>. Video recording available <a href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" data-href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">here<\/a>.<\/p><\/figure>","created_at":"2018-11-07T16:39:25+0000","updated_at":"2018-11-07T16:39:25+0000","published_at":"2018-01-19T16:41:16+0000","published_by":["Sandya Sankarram"],"annotations":[],"mimetype":"text\/html","reading_time":13,"domain_name":"medium.freecodecamp.org","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*PU20rbKCMm3C2BLEM24dgg.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":217,"title":"Datomic: Look at all the things I'm not doing!","url":"https:\/\/augustl.com\/blog\/2018\/datomic_look_at_all_the_things_i_am_not_doing\/","content":"<p>This post is part of a series: <a href=\"https:\/\/augustl.com\/series\/datomic\">Datomic<\/a><\/p><p>This is the video that got me into programming:<\/p><small><em>(This embed should take you directly to 01:52, where it's all at)<\/em><\/small><iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/Gzj723LkRJY?start=112\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>It's DHH's epic first Ruby on Rails demo from 2005. His unique selling point (or one of them, at least) was \"look at all the things I'm not doing\"<\/p><p>At the time, this was ground breaking in its implementation and accessibility. \"Convention over configuration\" became a thing.<\/p><p>However, a portion of the developer community now seems to agree that hiding all the complexities of large CRUD apps has its trade offs, especially when it comes to large and long running projects. You end up in a situation where you have to be as knowledgeable about the conventions as if you wrote the code yourself.<\/p><p>Whoops!<\/p><p>Any sufficiently advanced technology is indistinguishable from magic. But is magic always bad? Do we like magic, as long as it hits the sweet spot where it is able to remain hidden from us?<\/p><p>Why do Clojure devs despise the magic in Ruby on Rails, but love the magic of auto-promoting persistent data structures?<\/p><p>Let's explore all the things you're <em>not<\/em> doing when you're using Datomic.<\/p><h2>Look at all the data modeling I'm not doing<\/h2><p>Datomic has made a decision for you. When inserting data, everything is completely and utterly flat, with no structure.<\/p><pre class=\"highlight\">[:db\/add -1 :person\/name \"August Lilleaas\"]\n[:db\/add -1 :person\/age 31]<\/pre><p>This is <em>it<\/em>. All you have for writing data are these facts. A fact consists of:<\/p><pre class=\"highlight\">[:db\/add \n  -1                 ;; Entity (the id of the thing we're talking about)\n  :person\/name       ;; Attribute (defines the type and some other properties)\n  \"August Lilleaas\"  ;; Value (self evident)\n  ]\n  \n  ;; No table, no nesting, no depth, no ...<\/pre><p>Any structure comes later, from queries and reads.<\/p><p>Datomic isn't <em>free<\/em> of data modeling. You have to choose if an attribute has a cardinality of <em>one<\/em> or <em>many<\/em>. You need to define the type of your attributes. But the API you get for writing data is extremely limited. There's no structured schema.<\/p><p>So, essentially: <strong>Look at all the up front decisions I'm not making about data modeling<\/strong>.<\/p><p>This has one <em>enormous<\/em> benefit. You create structure when you read, not when you write. Therefore, you get to come up with new ways of thinking about your data <em>arbitrarily, after the fact<\/em>. This is just a logical conclusion of having all the decisions about structure being put in the query engine, and not the write API. Contrary to traditional databases, where so much structure is defined up-front when you insert the data.<\/p><h2>Look at all the time modeling I'm not doing<\/h2><p>So you write these facts into the database, and there's almost no structure when you write.<\/p><p>There are 50 ways to get time modeling wrong in a traditional database, and a few ways to get it right. And even when you got it right, maybe your specific implementation of time management didn't cut it so you need to update it. And maybe you didn't add time management to your <em>entire<\/em> domain, just the crucial parts.<\/p><p>Given Datomic's extremely simple write API, it is able to implement time management for you, without <em>any<\/em> additional cost of data modeling or query hatching on your part.<\/p><p>So, essentially: <strong>Look at all the up front decisions I'm not making about time modeling<\/strong>.<\/p><p>It's difficult to talk about Datomic without at one point mentioning the time model. The writer is single threaded. You get a completely serialized time-line of your writes, without any special coordination or re-ordering happening. There's no such thing in Datomic as \"read this thing, write that thing, read this other thing\" and have that be inconsistent in any way shape or form. There's no such thing as \"isolation level\" (alternatively, Datomic is isolation level 9000), and transactions can't see what happens in other transactions. This is fundamentally solved by having only one transaction being processed at a time.<\/p><p>When you query Datomic, you query against what Datomic calls a \"database\". This is essentially a view of a point in time of the time-line of writes that have occurred so far. This view is completely immutable, and requires no synchronization with the writer process because Datomic is an append-only database with append-only semantics.<\/p><p>Are you thinking what I'm thinking?<\/p><p>Did you notice the detail there that makes Datomic amazing?<\/p><p>You ask Datomic for a \"database\", and you get an immutable reference to the value of that database at that point in time. You get to hold onto this value forever, no matter how many writes that occur in the future.<\/p><p>Getting it yet?<\/p><p><em>You get to hold onto a view of the database that's immutable, no matter what happens to it in the future.<\/em><\/p><p>This means that it's extremely cheap for Datomic to let you get a view of a <em>previous state of the database<\/em>. In fact, internally, the API is exactly the same. You get a reference to the database data structure, with a T value (a reference or id to a point in time in the transaction time-line), and all your queries will ensure that you get the data as of that T value.<\/p><p>So it's really cheap to just set the T value to an earlier point in time, say two weeks ago, or whatever, and get a <em>completely immutable and consistent view of the entire database as of two weeks ago<\/em>.<\/p><p>Woha.<\/p><h2>Look at all the consistency management I'm not doing<\/h2><p>In that vein, the consistency model in Datomic is also extremely simple. As mentioned above, the writes happen in a single thread, writing one transaction at a time. This is isolation level 9000. There's no extra management to be done on your part to ensure that you use the correct isolation level or that you do things in the correct order to get the expected result or any of that. Datomic is consistent, period. There's no way to screw that up by misunderstandings or bugs made by the developer that uses Datomic in their app.<\/p><p>The only trade-off here is that you need to resort to stored procedures, VoltDB style, when you want to do anything other than setting a value, deleting (retracting) a value, and doing a few built-in operations like compare-and-swap and identity based upsert. If you want to do things like \"increment by 10\" or \"take N from account X and put it into account Y\", you need to express this as a transaction function (Datomic's name for stored procedures), written in Clojure or Java, and ship it off to the transactor (Datomic's name for the database server).<\/p><h2>Look at all the time management I'm not doing<\/h2><p>This is a bit of a minor point, but I still like it. It is an example of the many things that are suddenly fundamentally solved when you have a database that manages time for you.<\/p><p>Exhibit A: Pagination. Pagination has the issue of being a non-atomic operation. You load the first 50 items. Then you go forward, and load the next 50 items. But in between those two fetches of items, the database might have changed. With Datomic, you can have the T value be a factor in your pagination. You can use this to get a consistent view of the database, so that page two is actually page two. You can also run queries to see if your data set has changed since the provided T value, and inform the user. \"Click here to update your view of the data to the most recent changes\".<\/p><p>Exhibit B: Return values. Let's say you have a method that takes a bunch of arguments and inserts a new thing into your database. What should it return? Maybe just the ID, so we don't force all users of the API to wait for us to reconstruct a view of the thing that was just inserted. With Datomic, however, this is a non-issue. You simply return the ID and a reference to the T-value as of after your insert (which is something Datomic provides automatically). Then it is up to the consumer to do what it wants with the T-value. And this is another example of the extreme agility you get with Datomic. Because time is managed for you, you get to make <em>a whole lot of decisions later on, since you don't have to worry about the rug getting pulled from underneath your feet<\/em>.<\/p><h2>Look at all the event sourcing I'm not doing<\/h2><p>This is a big one.<\/p><p>If you're familiar with event sourcing, you might wonder what the big deal is. \"One single threaded time-line of writes\" is exactly what event sourcing is? <\/p><p>What does Datomic have that event sourcing doesn't?<\/p><p>Answer: An absence of hand written views.<\/p><p>With event sourcing, you have raw data in your stream of events. This data is useless until you create a view of it. A view is typically a left fold reducing function that gets the previous state and the next event in the list passed to it, and returns a new and updated state.<\/p><p><img width=\"600\" src=\"https:\/\/augustl.com\/static\/posts\/datomic_look_at_all_the_things_i_am_not_doing\/event_sourcing.jpg\" alt=\"image\" \/><\/p><p>This means that you have to decide which views you want to have. And that if there's a bug in your view code, you have to traverse the entire time-line again to build up the state.<\/p><p>With Datomic, you <em>have<\/em> the stream of events (facts\/transactions), but there is one crucial but fundamentally important difference:<\/p><p><em>You don't have to decide what your views are up-front<\/em><\/p><p>So, essentially: <strong>Look at all the up front decisions I'm not making about which views to have<\/strong>.<\/p><p><img width=\"600\" src=\"https:\/\/augustl.com\/static\/posts\/datomic_look_at_all_the_things_i_am_not_doing\/event_sourcing_datomic.jpg\" alt=\"image\" \/><\/p><p>Again, Datomic provides us with extreme agility. Because the data we write is completely flat, but is <em>not<\/em> arbitrary data but a well known structure (facts), and Datomic have generic indices to query these facts, and because Datomic keeps these indices up to date when we add data to our system, you get the benefits of event sourcing (time travel, a consistent time-line, all your historical data easily traversable), without having to worry about creating views for your data. The views are already there, in form of Datomic's query engine, its built-in time travel features, and efficient indices of historical data.<\/p><h2>Look at all the file writing I'm not doing<\/h2><p>(Implementation detail)<\/p><p>Internally, Datomic doesn't write to the file system directly. It writes to existing storage engines and databases. You can set up Datomic to write to any JDBC database (mysql, postgres, h2, oracle, db\/2, ...), dynamodb, riak + zookeper, and a few other combinations.<\/p><p>This means that Datomic doesn't have to implement its own code for properly writing to disks, which is a huge pain to get right. And it means that if you already use postgres, you don't have to worry about backup routines and operations and what not, other than running the Datomic transactor itself. You can just use your existing routines for postgres operations.<\/p><h2>Look at all the database concept reinvention I'm not doing<\/h2><p>(Implementation detail)<\/p><p>Datomic doesn't actually invent a whole lot from scratch.<\/p><p>I've already mentioned VoltDB. VoltDB also writes in a single thread, after research revealed that <a href=\"http:\/\/nms.csail.mit.edu\/~stavros\/pubs\/OLTP_sigmod08.pdf\">multi-threaded RDBMS databases spend about 75% of their time doing coordination and only 25% of their time actually doing reads and writes<\/a>. There's this instinct that single threaded processes can't possibly be efficient. But you'd be surprised how much efficiency you can get out of a thread that spends close to 100% of its time doing useful work without much time spent on coordination between multiple threads.<\/p><p>There's also a concept borrowed from Google's Big Table. The indices in Datomic aren't updated live for every write, that would be way too much overhead. So a query in Datomic runs against the latest index + a list of transactions that are written to a transaction log but not yet merged into the index data structures. Then, Datomic will periodically take the unmerged transaction log and merge it into the main index. Big Table does the same, after Google learned that live-merging an index of <em>the entire internet<\/em> was maybe not going to work.<\/p><h2>Look at all the .... ok, I'm done<\/h2><p>In my experience, Datomic is indistinguishable from magic, in a good way. It's one level of abstraction above traditional databases, but manages to hide those abstractions since you at no point worry about postgres or zookeeper internals when you make use of Datomic on top of the existing storage engines. You get a whole lot of stuff for free, without Datomic getting in your way.<\/p><p>And to me, the most important point is all of the things you're <em>not<\/em> doing and all of the decisions you are <em>not<\/em> making up front, but rather in an agile fashion as your application grows.<\/p><h2>Update: even more things I'm not doing<\/h2><p>I wrote a follow-up, with even more things I'm not doing:<\/p><p><a href=\"http:\/\/augustl.com\/blog\/2018\/datomic_look_at_all_the_things_i_am_not_doing_cont\/\">http:\/\/augustl.com\/blog\/2018\/datomic_look_at_all_the_things_i_am_not_doing_cont\/<\/a><\/p><hr class=\"post-sep\" \/><p>Questions or comments?<\/p><p>Feel free to contact me on Twitter, <a href=\"http:\/\/twitter.com\/augustl\">@augustl<\/a>, or e-mail me at <a href=\"mailto:august@augustl.com\">august@augustl.com<\/a>.<\/p>","created_at":"2018-11-07T16:39:25+0000","updated_at":"2018-11-07T16:39:25+0000","annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"augustl.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":218,"title":"How to Remove the Sidebar in WordPress","url":"https:\/\/www.wpbeginner.com\/wp-themes\/how-to-remove-the-sidebar-in-wordpress\/","content":"<p>Do you want to remove the sidebar from your WordPress site? A sidebar is the widget-ready area in your WordPress theme. It often appears at the right or left side or below the content area on every page. In this article, we will show you how to easily remove the sidebar in WordPress. <\/p><p><img title=\"How to Remove the Sidebar in WordPress\" src=\"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/removewpsidebar.png\" alt=\"How to Remove the Sidebar in WordPress\" width=\"550\" height=\"340\" class=\"alignnone size-full wp-image-43547\" \/><\/p><h4>Why Remove the Sidebar in WordPress?<\/h4><p>Most <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/decide-premium-free-wordpress-themes\/\" title=\"Free vs Premium WordPress Themes (Pros and Cons)\">free and paid WordPress themes<\/a> come with multiple <a href=\"https:\/\/www.wpbeginner.com\/glossary\/sidebar\/\" title=\"What is Sidebar in WordPress?\">sidebars<\/a> or widget ready areas. <\/p><p>The purpose of sidebars in WordPress is to display items that are not part of the post or page content. Usually these are links to different areas of your website, advertisements, <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/12-tested-and-easy-ways-to-grow-your-email-list-faster\/\" title=\"12 Tested and Easy Ways to Grow Your Email List Faster\">email list signup forms<\/a>, social media profiles, etc. These items can be easily added to the sidebar of a theme using <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/how-to-add-and-use-widgets-in-wordpress\/\" title=\"How to Add and Use Widgets in WordPress\">WordPress widgets<\/a>. <\/p><p><img title=\"Sidebar on a WordPress powered website\" src=\"https:\/\/cdn3.wpbeginner.com\/wp-content\/uploads\/2017\/04\/sidebar.jpg\" alt=\"Sidebar on a WordPress powered website\" width=\"550\" height=\"344\" class=\"alignnone size-full wp-image-43540\" \/><\/p><p>In most WordPress themes, sidebar appears differently when a user is viewing your site from a mobile device. Due to reduced width of the screen size, sidebars that appear on right or left on the desktop are moved to the bottom on mobile screens. <\/p><p>Sidebars are immensely helpful and can help you grow your business. See these <a href=\"https:\/\/www.wpbeginner.com\/showcase\/12-wordpress-sidebar-tricks-to-get-maximum-results\/\" title=\"12 WordPress Sidebar Tricks to Get Maximum Results\">WordPress sidebar tricks<\/a> for maximum results. <\/p><p>However, some website owners feel that removing sidebar helps improve user experience. If you\u2019re looking to remove the sidebar on your WordPress site, then this tutorial is for you.<\/p><h4>Video Tutorial<\/h4><iframe class=\"youtube-player\" width=\"520\" height=\"323\" src=\"https:\/\/www.youtube.com\/embed\/mO-QIZYFNPo?version=3&amp;rel=0&amp;fs=1&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><div class=\"yt-sbscrb-bar\"><p>Subscribe to WPBeginner<\/p><iframe src=\"https:\/\/www.youtube.com\/subscribe_embed?usegapi=1&amp;channel=wpbeginner&amp;layout=default&amp;count=default&amp;origin=http%3A%2F%2Fwpbeginner.com&amp;gsrc=3p&amp;ic=1&amp;jsh=m%3B%2F_%2Fscs%2Fapps-static%2F_%2Fjs%2Fk%3Doz.gapi.en_GB.Rx51stRkYnQ.O%2Fm%3D__features__%2Fam%3DAQ%2Frt%3Dj%2Fd%3D1%2Ft%3Dzcms%2Frs%3DAGLTcCM5XUrLkPZ9aFXqtSLivsifryqsUA#_methods=onPlusOne%2C_ready%2C_close%2C_open%2C_resizeMe%2C_renderstart%2Concircled%2Cdrefresh%2Cerefresh%2Conload&amp;id=I0_1448178294715&amp;parent=http%3A%2F%2Fwpbeginner.com\"> <\/iframe><\/div><p>If you don\u2019t like the video or need more instructions, then continue reading.<\/p><h4>Removing Sidebars From Everywhere in WordPress<\/h4><p>This method allows you to simply remove sidebars from every <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/what-is-the-difference-between-posts-vs-pages-in-wordpress\/\" title=\"What is the Difference Between Posts vs. Pages in WordPress\">page and post<\/a> on your WordPress site. It requires you to edit WordPress theme files. If you haven\u2019t done this before then see our guide on <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/beginners-guide-to-pasting-snippets-from-the-web-into-wordpress\/\" title=\"Beginner\u2019s Guide to Pasting Snippets from the Web into WordPress\">how to copy and paste code in WordPress<\/a>. <\/p><p>First you need to connect to your WordPress site <a href=\"https:\/\/www.wpbeginner.com\/showcase\/6-best-ftp-clients-for-wordpress-users\/\" title=\"6 Best FTP Clients for Mac and Windows WordPress Users\">using an FTP client<\/a> and then go to \/wp-content\/themes\/your-current-theme\/ folder. <\/p><p>WordPress themes are made up of different templates, so you will need to edit all the templates where a sidebar is displayed. <\/p><p>For example, in a typical WordPress theme you may need to edit index.php, page.php, single.php, archive.php, home.php and so on. <\/p><p>Open a template file to edit it and then locate the line that looks like this: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php get_sidebar(); ?&gt;&#13;\n<\/pre><p>If your theme comes with multiple sidebars, then you will see different instances of this code with a sidebar name inside the function. For example: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php get_sidebar('footer-widget-area'); ?&gt;&#13;\n<\/pre><p>Delete the line that represents the sidebar that you don\u2019t want to display.<\/p><p>Now, save and upload the file back to your website. Repeat the process for all template files responsible for displaying different pages on your website. <\/p><p>Once finished, you can go to your website and see it in action. <\/p><p>You may notice that while sidebars are gone, your content area is still appearing with the same width leaving the sidebar area empty. <\/p><p><img title=\"Blank sidebar area in a WordPress theme\" src=\"https:\/\/cdn.wpbeginner.com\/wp-content\/uploads\/2017\/04\/blanksidebar.jpg\" alt=\"Blank sidebar area in a WordPress theme\" width=\"550\" height=\"344\" class=\"alignnone size-full wp-image-43542\" \/><\/p><p>That\u2019s because your theme has defined the width for the content area. Now that the sidebar is gone, you need to adjust the width of the content area by <a href=\"https:\/\/www.wpbeginner.com\/plugins\/how-to-easily-add-custom-css-to-your-wordpress-site\/\" title=\"How to Easily Add Custom CSS to Your WordPress Site\">adding custom CSS<\/a> to your theme. <\/p><p>First you will need to figure out the CSS class used by your theme to define the width of the content area. You can do that by <a href=\"https:\/\/www.wpbeginner.com\/wp-tutorials\/basics-of-inspect-element-with-your-wordpress-site\/\" title=\"Basics of Inspect Element: Customizing WordPress for DIY Users\">using the Inspect tool<\/a> in your browser. <\/p><p>Here is the sample CSS we used on our demo site using Twenty Sixteen theme. <\/p><pre class=\"brush: css; title: ; notranslate\" title=\"\">&#13;\n.content-area {&#13;\n    width: 100%;&#13;\n    margin: 0px;&#13;\n    border: 0px;&#13;\n    padding: 0px;&#13;\n}&#13;\n&#13;\n.content-area .site {&#13;\nmargin:0px;&#13;\n}&#13;\n<\/pre><p>Here is how it looked without sidebars: <\/p><p><img title=\"Sidebars removed from a WordPress theme\" src=\"https:\/\/cdn3.wpbeginner.com\/wp-content\/uploads\/2017\/04\/nosidebarwp.jpg\" alt=\"Sidebars removed from a WordPress theme\" width=\"550\" height=\"344\" class=\"alignnone size-full wp-image-43543\" \/><\/p><h4>Removing Sidebars From Individual Pages in WordPress<\/h4><p>It\u2019s possible that you only want to remove the sidebars on certain pages while displaying them on other sections of your website. <\/p><p>For example, a lot of websites display sidebars on their blog page but not on individual blog posts. Some prefer to show sidebars everywhere except static pages. <\/p><p>Let\u2019s take a look at different scenarios and how to remove sidebars from different areas. <\/p><h4>Removing Sidebars from a Static Page in WordPress<\/h4><p>Many WordPress themes come with ready-to use full width page templates. These templates can be used with any static page in WordPress and do not display sidebars on either side of the content. <\/p><p>You will just need to edit the page where you want to remove the sidebar and then select the full width template under \u2018Page Attributes\u2019 meta box. <\/p><p><img title=\"Full width page template with no sidebars\" src=\"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/fullwidth-template.png\" alt=\"Full width page template with no sidebars\" width=\"550\" height=\"355\" class=\"alignnone size-full wp-image-43537\" \/><\/p><p>If you do not see an option to select a template, then this means that your theme does not come with a full-width page template. <\/p><p>In that case, you will need to create a full-width page template manually. Open a plain text editor like Notepad and paste the following code in a blank file: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php&#13;\n\/*&#13;\n*&#13;\nTemplate Name: Full-Width&#13;\n*\/&#13;\nget_header(); ?&gt;&#13;\n<\/pre><p>Now save this file as full-width.php on your desktop. <\/p><p>Connect to your WordPress site using an FTP client and then go to \/wp-content\/themes\/your-current-theme\/ folder. Edit  the file called page.php and copy every thing that appears after <code>&lt;?php get_header(); ?&gt;<\/code> line. <\/p><p>Now paste this code in your full-width.php file. After that you need to locate and delete the line that looks like this: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php get_sidebar(); ?&gt;&#13;\n<\/pre><p>Save your changes and upload the full-width.php file to your current theme folder. <\/p><p>You can now edit any page on your WordPress site and you will be able to select your full width page template.<\/p><p>For more details, see our guide on <a href=\"https:\/\/www.wpbeginner.com\/wp-themes\/how-to-create-a-full-width-page-in-wordpress\/\" title=\"How to Create a Full Width Page in WordPress\">how to create a full width page template in WordPress<\/a>. <\/p><h4>Remove Sidebar from Blog Posts Page in WordPress<\/h4><p>WordPress allows you to use a static front page and a separate page for your blog posts. Usually, most WordPress themes display sidebars on the blog posts page. <\/p><p>If you don\u2019t want to display sidebars on the blog posts page, then you need to go to <strong>Settings \u00bb Reading<\/strong> page. Notice the name of the page you are using as your blog posts page. <\/p><p><img title=\"Find your WordPress blog page\" src=\"https:\/\/cdn3.wpbeginner.com\/wp-content\/uploads\/2017\/04\/blogpage.jpg\" alt=\"Find your WordPress blog page\" width=\"550\" height=\"267\" class=\"alignnone size-full wp-image-43544\" \/><\/p><p>Now go to <strong>Pages \u00bb All Pages<\/strong> and edit the page used to display your blog posts. Under the page attribute section select a full width template and then save your changes. <\/p><p><img title=\"Full width page template with no sidebars\" src=\"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/fullwidth-template.png\" alt=\"Full width page template with no sidebars\" width=\"550\" height=\"355\" class=\"alignnone size-full wp-image-43537\" \/><\/p><p>If you don\u2019t see a full width template option, then you will have to manually create a full width template. See the previous section in this article for detailed instructions. <\/p><h4>Remove Sidebar from a Single Post in WordPress<\/h4><p>If you want to remove sidebar from certain single posts, then you can create a custom single post template. <\/p><p>WordPress 4.7 came with built-in support for post templates. You\u2019d need to follow the steps outline in our guide on <a href=\"https:\/\/www.wpbeginner.com\/wp-themes\/create-custom-single-post-templates-for-specific-posts-or-sections-in-wordpress\/\" title=\"How to Create Custom Single Post Templates in WordPress\">how to create custom single post templates in WordPress<\/a>.<\/p><p>In your custom single post template all you need to do is make sure to remove the sidebar part of the code.<\/p><p>We hope this article helped you learn how to easily remove the sidebar in WordPress themes. You may also want to see our step by step guide on how to <a href=\"https:\/\/www.wpbeginner.com\/wordpress-performance-speed\/\" title=\"The Ultimate Guide to Boost WordPress Speed &amp; Performance\">boost WordPress speed and performance<\/a>. <\/p><p>If you liked this article, then please subscribe to our <a href=\"http:\/\/youtube.com\/wpbeginner?sub_confirmation=1\" title=\"WPBeginner on YouTube\" target=\"_blank\" rel=\"nofollow\">YouTube Channel<\/a> for WordPress video tutorials. You can also find us on <a href=\"http:\/\/twitter.com\/wpbeginner\" title=\"WPBeginner on Twitter\" target=\"_blank\" rel=\"nofollow\">Twitter<\/a> and <a href=\"https:\/\/www.facebook.com\/wpbeginner\" title=\"WPBeginner on Facebook\" target=\"_blank\" rel=\"nofollow\">Facebook<\/a>.<\/p>","created_at":"2018-11-07T16:39:25+0000","updated_at":"2018-11-07T16:39:25+0000","published_at":"2017-05-17T06:15:49+0000","published_by":["Editorial Staff"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"www.wpbeginner.com","preview_picture":"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/removewpsidebar.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":215,"title":"12 \u201cManager READMEs\u201d from Silicon Valley\u2019s Top Tech Companies","url":"https:\/\/hackernoon.com\/12-manager-readmes-from-silicon-valleys-top-tech-companies-26588a660afe?gi=a95ce6658fe4","content":"<section class=\"section section--body section--first\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"c750\" class=\"graf graf--figure graf--layoutFillWidth graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*JPGxBj3qjqgKctsp9oLqmQ.png\" data-width=\"1412\" data-height=\"790\" data-is-featured=\"true\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*JPGxBj3qjqgKctsp9oLqmQ.png\" alt=\"image\" \/><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><h2 id=\"522f\" class=\"graf graf--h4 graf-after--h3 graf--subtitle\">What does tech management at Slack, HubSpot, Netflix, Etsy, Shopify, InVision, and more have in common? A lot, apparently!<\/h2><p id=\"bb44\" class=\"graf graf--p graf-after--h4\">When we launched our first <a href=\"https:\/\/soapboxhq.com\/modern-managers\/heather\" data-href=\"https:\/\/soapboxhq.com\/modern-managers\/heather\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">SoapBox hero<\/a> I was surprised to discover a document used to introduce her working style to her team. I had never seen that before and it was awesome!<\/p><p id=\"007e\" class=\"graf graf--p graf-after--p\">Did others have these? I needed to know. Eventually, I found a treasure trove of examples by Engineering Leaders from companies with some of the best culture in tech.<\/p><p id=\"3725\" class=\"graf graf--p graf-after--p\">They called them Manager READMEs, or Manager User Manuals. And they\u2019ve graciously let me share them with you.<\/p><p id=\"bd10\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Click the images to read their READMEs.<\/strong> At the end I\u2019ve included my notes in case you\u2019re planning on building your own.<\/p><p id=\"15b0\" class=\"graf graf--p graf-after--p graf--trailing\">But first, the inspiration! \ud83d\udc40 \ud83c\udf81 \ud83c\udf89<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"9a23\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/rands\" data-href=\"https:\/\/twitter.com\/rands\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Michael Lopp<\/a> (aka rands), VP Eng. at <a href=\"https:\/\/slack.com\" data-href=\"https:\/\/slack.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Slack<\/a> ~~~~~~~~~<\/h3><figure id=\"f80c\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" data-href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*KvpV191-jSpgeWzhtYOQVA.png\" data-width=\"1490\" data-height=\"1050\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*KvpV191-jSpgeWzhtYOQVA.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" data-href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http:\/\/randsinrepose.com\/archives\/how-to-rands\/<\/a><\/figcaption><\/figure><h4 id=\"ddd4\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"d3bc\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--h4\">\u201cOur average week\u201d covers what a typical working week with Michael is like, and how to engage him in it.<\/li><li id=\"2d32\" class=\"graf graf--li graf-after--li\">He might chat you on weekends. Feel free to read it Monday.<\/li><li id=\"8784\" class=\"graf graf--li graf-after--li\">He\u2019s an introvert (\u201cDo not confuse my quiet with lack of engagement.\u201d)<\/li><li id=\"f330\" class=\"graf graf--li graf-after--li graf--trailing\">Gossip is a trigger for him.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"d933\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/bnmrrs?lang=en\" data-href=\"https:\/\/twitter.com\/bnmrrs?lang=en\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Ben Morris<\/a>, Sr. Developer Lead at <a href=\"https:\/\/shopify.com\" data-href=\"https:\/\/shopify.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Shopify<\/a> ~~~~~~~~~<\/h3><figure id=\"853b\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*pgzHHvQUgehI3tSL1OXOGg.png\" data-width=\"1802\" data-height=\"1030\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*pgzHHvQUgehI3tSL1OXOGg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85<\/a><\/figcaption><\/figure><h4 id=\"6e05\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"78a1\" class=\"graf graf--li graf-after--h4\">Super clear description of his role and goals.<\/li><li id=\"586d\" class=\"graf graf--li graf-after--li\">Explicitly stating how he can help you (and how you can help him do <em class=\"markup--em markup--li-em\">his<\/em> job).<\/li><li id=\"0bbe\" class=\"graf graf--li graf-after--li graf--trailing\">Specifically how he incorporates Shopify values into his role.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"652c\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/royrapoport\" data-href=\"https:\/\/twitter.com\/royrapoport\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Roy<\/a>, Director at <a href=\"https:\/\/slack.com\" data-href=\"https:\/\/slack.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Slack<\/a> (frmly Netflix.com) ~~~~~~~<\/h3><figure id=\"e8ad\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/bit.ly\/roy-slack-readme\" data-href=\"http:\/\/bit.ly\/roy-slack-readme\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*0UC40BUNV_y-yNYB2cBiUg.png\" data-width=\"1808\" data-height=\"1028\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*0UC40BUNV_y-yNYB2cBiUg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/bit.ly\/roy-slack-readme\" data-href=\"http:\/\/bit.ly\/roy-slack-readme\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http:\/\/bit.ly\/roy-slack-readme<\/a><\/figcaption><\/figure><h4 id=\"e115\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"27d1\" class=\"graf graf--li graf-after--h4\">It\u2019s a mix of work and life. He has three goats. That\u2019s neat. Something to chat about.<\/li><li id=\"ab35\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--li\">\u201cI don\u2019t believe in personality conflicts\u201d At least not as a manager at work.<\/li><li id=\"9da6\" class=\"graf graf--li graf-after--li\">One of the best descriptions of what a manager\u2019s job is I\u2019ve read in a long time.<\/li><\/ul><figure id=\"5b86\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--li\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/bit.ly\/welcometonetflixandme\" data-href=\"https:\/\/bit.ly\/welcometonetflixandme\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*I3Sx95TyxA4xvOa2GAJvyQ.png\" data-width=\"1808\" data-height=\"1028\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*I3Sx95TyxA4xvOa2GAJvyQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/bit.ly\/welcometonetflixandme\" data-href=\"https:\/\/bit.ly\/welcometonetflixandme\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener noopener\" target=\"_blank\">https:\/\/bit.ly\/welcometonetflixandme<\/a><\/figcaption><\/figure><p id=\"ffb2\" class=\"graf graf--p graf-after--figure\">Roy\u2019s first README was built at Netflix. I think it\u2019s great to see how Roy\u2019s document has changed between two companies. It\u2019s a great reminder that a manager adapts to the environment they are in.<\/p><p id=\"222f\" class=\"graf graf--p graf-after--p graf--trailing\">This is one of the first READMEs you can find online. Many have taken inspiration from Roy.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"4261\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/orenellenbogen\" data-href=\"https:\/\/twitter.com\/orenellenbogen\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Oren<\/a>, VP Eng. at <a href=\"https:\/\/forter.com\" data-href=\"https:\/\/forter.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Forter<\/a> (+<a href=\"http:\/\/softwareleadweekly.com\/\" data-href=\"http:\/\/softwareleadweekly.com\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"nofollow noopener\" target=\"_blank\">SoftwareLeadWeekly.com<\/a>)<\/h3><figure id=\"9de9\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" data-href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*LPKPBbi6NehCkkJ3Pn9HQQ.png\" data-width=\"1500\" data-height=\"924\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*LPKPBbi6NehCkkJ3Pn9HQQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" data-href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha<\/a><\/figcaption><\/figure><h4 id=\"4047\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"db08\" class=\"graf graf--li graf-after--h4\">Links, images, video\u200a\u2014\u200aFurther material is available to get a deeper understanding if you\u2019re curious!<\/li><li id=\"51ee\" class=\"graf graf--li graf-after--li\">Personal quirks (specifically the ones he doesn\u2019t like about himself!)<\/li><li id=\"5826\" class=\"graf graf--li graf-after--li graf--trailing\">Expected response times by communication channel (helpful especially for remote teams).<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"2476\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/aaronlerch\" data-href=\"https:\/\/twitter.com\/aaronlerch\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Aaron<\/a>, Eng. Manager at <a href=\"https:\/\/invisionapp.com\" data-href=\"https:\/\/invisionapp.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">InVision<\/a> ~~~~~~~~~~~~~~~~~<\/h3><figure id=\"977a\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*6z_AznTfTEsR8HNDds1oMQ.png\" data-width=\"1810\" data-height=\"1028\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*6z_AznTfTEsR8HNDds1oMQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p<\/a><\/figcaption><\/figure><h4 id=\"d3c2\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"daf2\" class=\"graf graf--li graf-after--h4\">His personal goal as a manager at InVision.<\/li><li id=\"2149\" class=\"graf graf--li graf-after--li\">What does an ideal team player look like on his team?<\/li><li id=\"39b4\" class=\"graf graf--li graf-after--li graf--trailing\">A transparent edit, commenting on his ability to be responsive.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"8e4a\" class=\"graf graf--h3 graf--leading\"><a href=\"http:\/\/www.mollywhite.net\/\" data-href=\"http:\/\/www.mollywhite.net\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Molly<\/a>, Tech Lead at <a href=\"https:\/\/hubspot.com\" data-href=\"https:\/\/hubspot.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">HubSpot<\/a> ~~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"d2fb\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/github.com\/molly\/manager-README\" data-href=\"https:\/\/github.com\/molly\/manager-README\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*kf3v83tAmWMl8cpe_BmjVg.png\" data-width=\"1988\" data-height=\"1078\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*kf3v83tAmWMl8cpe_BmjVg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/github.com\/molly\/manager-README\" data-href=\"https:\/\/github.com\/molly\/manager-README\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/github.com\/molly\/manager-README<\/a><\/figcaption><\/figure><h4 id=\"343f\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"84e2\" class=\"graf graf--li graf-after--h4\">It\u2019s on github! Super cool.<\/li><li id=\"d1d6\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--li\">\u201cI am accountable for the decisions the team make, even if I\u2019m not the one making them most of the time.\u201d<\/li><li id=\"21a0\" class=\"graf graf--li graf-after--li graf--trailing\">Specific call out for diversity and mental health.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"e470\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/matthewnewkirk.com\/\" data-href=\"https:\/\/matthewnewkirk.com\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Matt<\/a>, Sr. Eng. Manager at <a href=\"https:\/\/etsy.com\" data-href=\"https:\/\/etsy.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Etsy<\/a> ~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"5909\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" data-href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*CXHucUSFBXya7wVqntUPfg.png\" data-width=\"1434\" data-height=\"938\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*CXHucUSFBXya7wVqntUPfg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" data-href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/<\/a><\/figcaption><\/figure><h4 id=\"40a0\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"b34f\" class=\"graf graf--li graf-after--h4\">Guests can edit his calendar events by defaults!<\/li><li id=\"e811\" class=\"graf graf--li graf-after--li\">People &gt; Process (no matter the methodology). If agile\/scrum isn\u2019t working for the team, let\u2019s figure out what will.<\/li><li id=\"608d\" class=\"graf graf--li graf-after--li graf--trailing\">He\u2019s big on work-life balance\u2026 meaning you\u2019ll likely never hear from him on weekends unless it\u2019s a big deal!<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"e1b0\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/scottsburns\" data-href=\"https:\/\/twitter.com\/scottsburns\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Scott<\/a>, Vp Eng. at <a href=\"https:\/\/stratasan.com\" data-href=\"https:\/\/stratasan.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Stratasan<\/a> ~~~~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"226a\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*d6UqXWgq3iQqjtjFDMiYAA.png\" data-width=\"1798\" data-height=\"1024\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*d6UqXWgq3iQqjtjFDMiYAA.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0<\/a><\/figcaption><\/figure><h4 id=\"c04d\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"287e\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--h4\">\u201cI believe managers work for their direct reports.\u201d<\/li><li id=\"a107\" class=\"graf graf--li graf-after--li graf--trailing\">On Deep Work: \u201cThe maintenance of inboxes (email, Slack, social media, etc) is not what you\u2019re paid to do. Attention is a valuable resource\u201d Not true for all roles, but true for Scott\u2019s team.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"a509\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/github.com\/elliottcarlson\" data-href=\"https:\/\/github.com\/elliottcarlson\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Elliott<\/a>, Manager at <a href=\"https:\/\/renttherunway.com\" data-href=\"https:\/\/renttherunway.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">RentTheRunway<\/a> ~~~~~~~~~~~~~~<\/h3><figure id=\"3933\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" data-href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*CqfL5ujQmbHpZyWODh5irQ.png\" data-width=\"1630\" data-height=\"990\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*CqfL5ujQmbHpZyWODh5irQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" data-href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view<\/a><\/figcaption><\/figure><h4 id=\"a420\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"8f20\" class=\"graf graf--li graf-after--h4\">He treats the document as a \u201cpromise on how I will conduct myself as a manager\u201d and asks to be held accountable.<\/li><li id=\"2b14\" class=\"graf graf--li graf-after--li\">He\u2019s an internal thinker who is sometimes quieter in large groups.<\/li><li id=\"e84e\" class=\"graf graf--li graf-after--li graf--trailing\">He may have people pleaser tendencies that he\u2019s working on.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"3868\" class=\"graf graf--h3 graf--leading\"><a href=\"http:\/\/angryweasel.com\/blog\/\" data-href=\"http:\/\/angryweasel.com\/blog\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Alan<\/a>, Director at <a href=\"https:\/\/unity3d.com\/\" data-href=\"https:\/\/unity3d.com\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Unity<\/a> ~~~~~~~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"fc20\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/github.com\/angryweasel\/managerreadme\" data-href=\"https:\/\/github.com\/angryweasel\/managerreadme\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*NMH3j5te1Mxpcr7XCLNwPw.png\" data-width=\"1988\" data-height=\"940\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*NMH3j5te1Mxpcr7XCLNwPw.png\" alt=\"image\" \/><\/a><\/div><\/figure><h4 id=\"e7ca\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"ebd1\" class=\"graf graf--li graf-after--h4\">Love that its on github\u200a\u2014\u200acheck the commits for edits!<\/li><li id=\"ae7d\" class=\"graf graf--li graf-after--li\">How his ICs can become leaders within their organization.<\/li><li id=\"bd86\" class=\"graf graf--li graf-after--li graf--trailing\">References to further reading on his management influences!<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"e77e\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/medium.com\/@mikekleiman\" data-href=\"https:\/\/medium.com\/@mikekleiman\" data-anchor-type=\"2\" data-user-id=\"d0849441a93f\" data-action-value=\"d0849441a93f\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--h3-user\" target=\"_blank\">Mike<\/a>, Manager of Eng. at BiblioCommons ~~~~~~~~~<\/h3><figure id=\"38a5\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/medium.com\/@mikekleiman\/my-readme-md-35bd0197a5c5\" data-href=\"https:\/\/medium.com\/@mikekleiman\/my-readme-md-35bd0197a5c5\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*GqHG0Kxx4N1D7K6kq3oOiw.png\" data-width=\"1478\" data-height=\"966\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*GqHG0Kxx4N1D7K6kq3oOiw.png\" alt=\"image\" \/><\/a><\/div><\/figure><h4 id=\"debc\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"72ad\" class=\"graf graf--li graf-after--h4\">Mike included a <strong class=\"markup--strong markup--li-strong\">TLDR<\/strong>; ahh what a refreshing summary.<\/li><li id=\"c164\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--li\">\u201cYou should expect me to have your back 100% of the time.\u201d Love it!<\/li><li id=\"40c0\" class=\"graf graf--li graf-after--li\">It\u2019s on medium! Now, that\u2019s transparent!<\/li><\/ul><p id=\"cc93\" class=\"graf graf--p graf-after--li graf--trailing\">.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"afeb\" class=\"graf graf--h3 graf--leading\">Ready to write your\u00a0own?<\/h3><p id=\"4346\" class=\"graf graf--p graf-after--h3\">\ud83d\udca1\ud83d\udc49 <a href=\"https:\/\/soapboxhq.com\/blog\/modern-manager-community\/how-to-set-expectations-with-your-team\" data-href=\"https:\/\/soapboxhq.com\/blog\/modern-manager-community\/how-to-set-expectations-with-your-team\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">We (SoapBox) just wrote an article on how!<\/a><\/p><p id=\"c81d\" class=\"graf graf--p graf-after--p\">However, almost everyone above included the following:<\/p><ul class=\"postList\"><li id=\"ce3b\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">What is this? <\/strong>\u2014Let\u2019s face it, this is a weird document. Give context on what this is and how to interpret it: to supplement getting to know each other.<\/li><li id=\"1dce\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">About Me \/ My Job\u200a<\/strong>\u2014\u200aWhat are some common things people might like to know about you. Some people\u2019s are more personal than others. If you want to get to know your team\u2019s personal lives then start off by sharing yours.<\/li><li id=\"a630\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Personal Principles \/ Values<\/strong>\u200a\u2014\u200aWhat are your default assumptions about people and their intentions? What mindsets do you have and which do you hope other people adopt when working together on the team? What things trigger you?<\/li><li id=\"f374\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">One-on-Ones<\/strong>\u200a\u2014\u200aWhat style of 1:1s do you want out of your team? Most above follow the weekly\/bi-weekly, 30 min cadence, and the employee controls the agenda.<\/li><li id=\"8a79\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Feedback (how to give &amp; how you give)<\/strong>\u200a\u2014\u200aWhat type of feedback do you want? Are you comfortable with people being blunt with you? How do you prefer to give feedback? How do you expect your team to react to feedback?<\/li><li id=\"e150\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">How to interpret my calendar<\/strong>\u200a\u2014\u200aSometimes a manager\u2019s calendar can be packed. Almost every manager above wants their team to know that they are the most important part of their job, so they\u2019re explicit in saying so. Message them on slack if you need to talk. They\u2019ll make time.<\/li><\/ul><p id=\"24e9\" class=\"graf graf--p graf-after--li\">Many also include the following:<\/p><ul class=\"postList\"><li id=\"8ef2\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Personal performance scales<\/strong>\u200a\u2014\u200aIf an employee asks, \u201cWhere do I stand?\u201d How will you respond? Most seem to do Red\/Orange\/Green. While Green and Red are clear indicators, orange can open to interpretation. What does that mean and how do you intend for people to react?<\/li><li id=\"cf12\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">D.R.I. Principle<\/strong>\u200a\u2014\u200aNot everyone is a believer in <a href=\"https:\/\/www.forbes.com\/sites\/quora\/2012\/10\/02\/how-well-does-apples-directly-responsible-individual-dri-model-work-in-practice\/#1c3bdb35194c\" data-href=\"https:\/\/www.forbes.com\/sites\/quora\/2012\/10\/02\/how-well-does-apples-directly-responsible-individual-dri-model-work-in-practice\/#1c3bdb35194c\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">D<\/strong>irectly <strong class=\"markup--strong markup--li-strong\">R<\/strong>esponsible <strong class=\"markup--strong markup--li-strong\">I<\/strong>ndividual<\/a> principle, but if you are it can be very powerful. First, though, you need to set the expectation that they take that responsibly.<\/li><\/ul><h3 id=\"d2f1\" class=\"graf graf--h3 graf-after--li\">Done? Ship\u00a0it!<\/h3><p id=\"324a\" class=\"graf graf--p graf-after--h3\">How? Consensus from the authors I\u2019ve talked to seems to be:<\/p><ol class=\"postList\"><li id=\"ec34\" class=\"graf graf--li graf-after--p\">The most effective way to deliver the document is to take employees through it in a 1:1.<\/li><li id=\"af88\" class=\"graf graf--li graf-after--li graf--trailing\">Easiest time to start is during an employee\u2019s on-boarding. Their on-boarding or your own (if you\u2019re inheriting a team)<\/li><\/ol><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"aed4\" class=\"graf graf--h3 graf--leading\">You made it to the\u00a0end!<\/h3><blockquote id=\"ff83\" class=\"graf graf--blockquote graf-after--h3\"><div><strong class=\"markup--strong markup--blockquote-strong\">Here\u2019s my question for you: <\/strong>Anything that I missed in my notes? Anything that they missed in their documents? What are a must-haves in your doc?<\/div><\/blockquote><p id=\"38ce\" class=\"graf graf--p graf-after--blockquote\">Thanks for reading,<br \/>Brennan<\/p><p id=\"2dd6\" class=\"graf graf--p graf-after--p\">p.s. You should be having these conversations with your team at least bi-weekly. <strong class=\"markup--strong markup--p-strong\">If your conversations are stale, inconsistent, or unorganized, you should try out <\/strong><a href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" data-href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">SoapBox<\/strong><\/a><strong class=\"markup--strong markup--p-strong\"> now.<\/strong><\/p><figure id=\"a244\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*af8A62nrzIjp6gioRz3rrg.png\" data-width=\"1186\" data-height=\"414\" data-action=\"zoom\" data-action-value=\"1*af8A62nrzIjp6gioRz3rrg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*af8A62nrzIjp6gioRz3rrg.png\" alt=\"image\" \/><\/div><\/figure><h4 id=\"9d79\" class=\"graf graf--h4 graf-after--figure\">Brennan is the CEO &amp; Co-founder of <a href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" data-href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" class=\"markup--anchor markup--h4-anchor\" rel=\"noopener\" target=\"_blank\">SoapBox<\/a>, the #1 place to work in Canada. <a href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" data-href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" class=\"markup--anchor markup--h4-anchor\" rel=\"noopener\" target=\"_blank\">SoapBox<\/a>, is an app and assistant for managers to have better one-on-ones, team meetings, town-halls and more, with their\u00a0team.<\/h4><figure id=\"0bfb\" class=\"graf graf--figure graf-after--h4 graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/twitter.com\/i_am_brennan\" data-href=\"https:\/\/twitter.com\/i_am_brennan\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*-36b5jCwLYg0MVrPP0kXKg.png\" data-width=\"1408\" data-height=\"112\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*-36b5jCwLYg0MVrPP0kXKg.png\" alt=\"image\" \/><\/a><\/div><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"579c\" class=\"graf graf--p graf--leading\">If you liked this article, you should give it <strong class=\"markup--strong markup--p-strong\">12\ud83d\udc4f<\/strong>\u2019s (one for each readme) to help others find it!\ud83d\udc47<\/p><figure id=\"bf76\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/bit.ly\/2LLdEDb\" data-href=\"http:\/\/bit.ly\/2LLdEDb\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*Xv3XV-DgnRK9WjRs90xGMA@2x.png\" data-width=\"1400\" data-height=\"774\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Xv3XV-DgnRK9WjRs90xGMA@2x.png\" alt=\"image\" \/><\/a><\/div><\/figure><\/div><\/div><\/section>","created_at":"2018-11-07T16:39:24+0000","updated_at":"2018-11-07T16:39:24+0000","published_at":"2018-05-02T13:36:41+0000","published_by":["Brennan McEachran \ud83d\udc68\u200d\ud83d\ude80"],"annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"hackernoon.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*JPGxBj3qjqgKctsp9oLqmQ.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":214,"title":"Eshell as a main shell","url":"https:\/\/ambrevar.xyz\/emacs-eshell\/","content":"<p>\nIt\u2019s undeniable however that, as of Emacs 25.2, Eshell has numerous rough edges.\nQuite a few bug reports have been sent on debbugs and it\u2019s getting better.\n<\/p><p>\nNative completion is limited considering very little effort was ever made.  The\ncommunity can help changing that.  In the meantime, I\u2019ve implemented <a href=\"https:\/\/github.com\/Ambrevar\/emacs-fish-completion\">a fallback\nto Bash\/fish completion<\/a> so that I can keep completing just as much as I could\nwith Bash or fish.\n<\/p><p>\nNeed that curses-based program? Get a (much more) powerful Emacs alternative:\n<\/p><ul class=\"org-ul\"><li>git (CLI), gitk, tig, etc. -&gt; magit<\/li>\n<li>htop -&gt; symon, proced, helm-top, etc.<\/li>\n<li>abook -&gt; org-contacts<\/li>\n<li>mutt -&gt; mu4e, gnus, notmuch<\/li>\n<li>ncdu -&gt; dired-du<\/li>\n<li>cmus\/moc -&gt; EMMS<\/li>\n<li>newsbeuter -&gt; Elfeed, gnus<\/li>\n<li>weechat, irssi -&gt; ERC, etc.<\/li>\n<li>rtorrent, transmission-cli -&gt; transmission.el<\/li>\n<\/ul><p>\nAs opposed to their curses-like counterparts, Emacs alternatives share all those\nfeatures:\n<\/p><ul class=\"org-ul\"><li>Configure\/extend in Lisp.<\/li>\n<li>(Fuzzy-)search the buffer.  Copy anything to the clipboard.<\/li>\n<li>No need to change the color theme, it\u2019s already using your Emacs theme.  No\nneed for dircolors, lesspipe, Xresources\u2026<\/li>\n<li>Emacs bindings (Evil?).  All configurable. No more painful context-switch\nbetween Emacs-based and vi-based bindings.<\/li>\n<\/ul><p>\nIn the meantime, there might still be that annoying curses-based program you\ncannot replace.  Eshell allows you to run those programs by transparently\nswitching to a <code>*term*<\/code> buffer.  See <code>eshell-visual-commands<\/code>.\n<\/p><p>\nCurses-based programs are not the only ones that won\u2019t run well on Eshell: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Read%E2%80%93eval%E2%80%93print_loop\">REPL<\/a>\nare known for working poorly too.  REPLs are poorly designed though: from a\nUnix-philosophy standpoint, only the <code>E<\/code> is required for every language, while\nre-implementing the <code>R<\/code>, <code>P<\/code>, and <code>L<\/code>, i.e.\u00a0the user interface, for each one of\nthem is simply re-inventing the wheel.  Instead of that, Emacs has <code>comint-mode<\/code>,\nan REPL-lie user interface that can be plugged to any language.  With all the\npowerful features of Emacs that all those specialized REPLs are lacking.  For a\nstart, you can write code in Emacs and execute any piece of code from a single\nkeystroke.  It\u2019s configurable and extensible in a consistent manner across all\nlanguages.\n<\/p><p>\nSimilarly, \u201cenvironment setups\u201d, as commonly used by programming language tools\nsuch as Python\u2019s <code>virtualenv<\/code> or Ruby <code>rvm<\/code>, expect the user to be running a\nspecific shell.  This is broken by design.  Let me quote <a href=\"https:\/\/reddit.com\/u\/rpdillon\/\"><i>u\/rpdillon<\/i><\/a> as I could\nnot have said better:\n<\/p><blockquote>\n<p>\nTo OP\u2019s point, though, yes, these environments do make assumptions that you\u2019ll\nbe using some kind of bash-compatible shell, which tells you a little bit about\nhow sub-optimal the state-of-the-art is.  I suppose the only alternative I can\nthink of is to have the language itself manage many versions of itself so that\nevery developer\u2019s environment wouldn\u2019t have to be customized\/extended to\nnegotiate the maze.\n<\/p>\n<\/blockquote><p>\nThe <a href=\"https:\/\/golang.org\/\">Go<\/a> programming language for instance does not have such a tool as I know of.\n<\/p><p>\nA long-standing critique of Eshell is that it does not support input\nredirection.  This is hardly a major hindrance since you can always use a\nwork-around such as\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">cat file | command\n<\/pre><\/div><p>\nAnother common argument against Eshell is that it cannot run the widely popular\nTmux.  The latter is used for many things, out of which:\n<\/p><ul class=\"org-ul\"><li>Window management, tabs, etc.: you don\u2019t need that if you have Emacs.<\/li>\n<li>Leave long-running processes in the background after the user has logged out:\nthis feature is easily provided with <code>nohup<\/code> or <code>dtach<\/code>.<\/li>\n<\/ul><p>\nAll that being said, should Eshell not do the job for you, then maybe give <code>M-x\nshell<\/code> a try.  It has a similar design: it is a shell emancipated from any\nterminal; instead of being Lisp based, it runs an existing shell as an\ninterpreter.  <code>Bash<\/code> and <code>zsh<\/code> are known to work.  It has all the advantages an\nEmacs buffer has over terminal emulators.  The main difference then lies in the\nlanguage: <a href=\"https:\/\/www.gnu.org\/software\/emacs\/emacs-paper.html\">Emacs<\/a> <a href=\"http:\/\/paulgraham.com\/avg.html\">Lisp<\/a> vs.  <a href=\"http:\/\/mywiki.wooledge.org\/BashPitfalls\">Bash and related languages<\/a>.\n<\/p>","created_at":"2018-11-07T16:39:04+0000","updated_at":"2018-11-07T16:39:04+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"ambrevar.xyz","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":212,"title":"Experimental music notation resources","url":"https:\/\/llllllll.co\/t\/experimental-music-notation-resources\/149?_escaped_fragment_=","content":"<div class=\"post\" itemprop=\"articleBody\"><p>As somebody with one foot in the visual art world and the other in sound world, I have always been interested in the experimental music notation strategies. Aside from the fact that these are often beautiful artifacts, I love thinking about the \u201cutility\u201d of images: how\/when image becomes an \u201cinstruction\u201d.<\/p>\n<p>I think the new |||||||| forum, with it\u2019s expanded conceptual ambition, ought to have a thread devoted to this aspect of new and experimental music.<\/p>\n<p>Let\u2019s share resources, files, images, links and ahything you might have collected related to this esotheric and fascinating point on confluence between image, language ne sound.<\/p>\n<p>p.<\/p>\n<p>To get things started, here is a nice article from Simthsonian:<br \/><\/p><aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"https:\/\/static-media.smithsonianmag.com\/img\/favicon.ico\" class=\"site-icon\" width=\"16\" height=\"16\" alt=\"image\" \/><a href=\"https:\/\/www.smithsonianmag.com\/arts-culture\/5-12-examples-of-experimental-music-notation-92223646\/\" target=\"_blank\" rel=\"nofollow noopener\">Smithsonian<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/thumbs-prod.si-cdn.com\/YfEcEmAAFtq1KtIAZORrr_WpEMY=\/1072x720\/filters:no_upscale()\/https:\/\/public-media.smithsonianmag.com\/filer\/Design-Decoded-Experimental-Music-631.jpg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"https:\/\/www.smithsonianmag.com\/arts-culture\/5-12-examples-of-experimental-music-notation-92223646\/\" target=\"_blank\" rel=\"nofollow noopener\">5 1\/2 Examples of Experimental Music Notation<\/a><\/h3>\n<p>In the 1950s progressive composers broke from the 5 line music staff to experiment with new, more expressive forms of graphic music notation<\/p>\n  <\/article>\n  \n<\/aside>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/129\/2abc3a11c4fa3df8.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/1ba417cc7055250dab369d201624e1b8ed0354b6\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/1ad\/92b\/02c7b00782_666x500.jpg\" width=\"666\" height=\"500\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>immediately reminded me of jobina tinnemans\u2019 style of notation<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/130\/428c7328c89a148b.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/d58e3ad43c36ae68012d280e89c51c9a189a6693\" title=\"jt.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/ed8\/47c\/6c3d7c4289_500x500.jpg\" width=\"500\" height=\"500\" alt=\"image\" \/><\/a><\/div> \n<aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"https:\/\/s2.wp.com\/i\/favicon.ico\" class=\"site-icon\" width=\"32\" height=\"32\" alt=\"image\" \/><a href=\"https:\/\/jobinatinnemans.com\/\" target=\"_blank\" rel=\"nofollow noopener\">JOBINA TINNEMANS<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/jobinatinnemans.files.wordpress.com\/2015\/11\/jobina-tinnemans-owen-lucas-photography.jpeg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"https:\/\/jobinatinnemans.com\/\" target=\"_blank\" rel=\"nofollow noopener\">JOBINA TINNEMANS<\/a><\/h3>\n<p>Contemporary Composer<\/p>\n  <\/article>\n  \n<\/aside><\/div><div class=\"post\" itemprop=\"articleBody\"><p>This is exactly what I am hoping we could share here <a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/glia\">@glia<\/a> !<\/p>\n<p>I also should have mentioned specifically: the image I included in the first post is an example of many amazing experimantal notations by Cornelius Cardew.<\/p>\n<p>Here are a few more from Cardew:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/131\/1d566632c5f90a8c.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/506f699a77d6d9a91e57a9e4ca4d7fd35a51078e\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/cf8\/3c5\/58c23dae1a_690x451.jpg\" width=\"690\" height=\"451\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/132\/707501934fdae508.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/e9c1447ae7517229ea1a568de32cda8d0628d661\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/847\/db4\/9d3162a7fc_690x388.jpg\" width=\"690\" height=\"388\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/133\/6b0c53c95e2a8369.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/ddaa4fb6c0b1d41f3e6437e2e059553a517e547a\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/bb0\/57d\/cc9ae5441e_690x404.jpg\" width=\"690\" height=\"404\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>sweet looking stuff<\/p>\n<p>in the midst of read more on this\u2026<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>This kind of thing sends me.<\/p>\n<p>Leonardo Music Journal did an issue called \u201cBEYOND NOTATION: COMMUNICATING MUSIC\u201d:<\/p>\n<p><a href=\"http:\/\/www.leonardo.info\/isast\/journal\/toclmj21.html\" class=\"onebox\" target=\"_blank\" rel=\"nofollow noopener\">http:\/\/www.leonardo.info\/isast\/journal\/toclmj21.html<\/a><\/p>\n<p>It\u2019s not open access, and it\u2019s kind of expensive. I can comment further on particular articles if anyone\u2019s interested.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>beautiful<\/p>\n<p>Iannis Xenakis worked a lot around this : <a href=\"http:\/\/www.centre-iannis-xenakis.org\/cix_upic_presentation\">http:\/\/www.centre-iannis-xenakis.org\/cix_upic_presentation<\/a><br \/>(it\u2019s french but one can easily find resources about I. Xenakis)<br \/>the software he used is called UPIC.<\/p>\n<p>and there is <a href=\"http:\/\/www.iannix.org\/en\/whatisiannix\/\">http:\/\/www.iannix.org\/en\/whatisiannix\/<\/a><br \/>it\u2019s a sequencer software based on drawings. very very interesting compositionnal tool<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Wonderful references guys!<br \/><a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/beo\">@beo<\/a> great to see the UPIC legacy carried into the new dimensions with Iannix.<br \/>Here are a few images of Xenakis notations for visual pleasure:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/144\/1aabdffd84f29971.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/fdaca252d71f0a52a5f702c7e6f2a603e08edb55\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/f4d\/af2\/23837064ce_690x424.jpg\" width=\"690\" height=\"424\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/145\/3cb8b5f88ccaf2ff.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/36fa17127637bc35e89fc40136cd29a951a6f3e6\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/2d0\/3af\/fbd2826b42_633x499.jpg\" width=\"633\" height=\"499\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/146\/b14e50e190d4f1c2.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/61f6916c176bbbf51d5262871507711177f3f111\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/e26\/919\/d686532351_690x350.jpg\" width=\"690\" height=\"350\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Als, here is a nice collection of images with unorthodox notations:<\/p>\n<aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><a href=\"http:\/\/blog.wfmu.org\/freeform\/2007\/01\/gallery_of_musi.html\" target=\"_blank\" rel=\"nofollow noopener\">WFMU's Beware of the Blog<\/a>\n  <\/header><article class=\"onebox-body\"><img src=\"http:\/\/blog.wfmu.org\/photos\/uncategorized\/bernard_rands__as_all_thumb_1.jpg\" class=\"thumbnail onebox-avatar\" width=\"150\" height=\"150\" alt=\"image\" \/><h3><a href=\"http:\/\/blog.wfmu.org\/freeform\/2007\/01\/gallery_of_musi.html\" target=\"_blank\" rel=\"nofollow noopener\">Gallery of Graphic Musical Notation<\/a><\/h3>\n<p>Here's 24 examples of unusual or graphic musical notation, from the cuneiform markings of the oldest song in the world, to the graphic notation of Stockhausen, Crumb, Cage and Eno, to the painted musical inventions of Adolph Wolfli. Click each...<\/p>\n  <\/article>\n  \n<\/aside><p>Like this one from George Crumb:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/147\/7fe470a2b4fc0248.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/6d55016989181410e80584965695ed232d98b2ad\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/10e\/305\/6ccfcd031c_493x500.jpg\" width=\"493\" height=\"500\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/148\/06fac602da575cbb.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/50a385d6a0ed89adfb25bcca7da1cfbd067dd0e3\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/86e\/f10\/a3f4ba0fd2_570x500.jpg\" width=\"570\" height=\"500\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/149\/47243fb9970dbf91.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/e9f0b4704bbe6732a5e2ef0838efa601a4ab13d7\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/bae\/120\/fc11a6b73b_690x332.jpg\" width=\"690\" height=\"332\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>ohh, i especially like the circular notation ones. possible new profile picture <img src=\"https:\/\/llllllll.co\/images\/emoji\/twitter\/+1.png?v=5\" title=\":+1:\" class=\"emoji\" alt=\":+1:\" \/><\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p><a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/laborcamp\">@laborcamp<\/a> :   thanks for the picture <img src=\"https:\/\/llllllll.co\/images\/emoji\/twitter\/smile.png?v=5\" title=\":smile:\" class=\"emoji\" alt=\":smile:\" \/><br \/>actually a new UPIC is under dev by students in Rouen univ. the goal is to make a UPIC for win, mac and linux. they call it UPIX.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>This is a super cool\/useful book on the matter:<br \/><\/p><aside class=\"onebox amazon\"><header class=\"source\"><a href=\"https:\/\/www.amazon.co.uk\/Notations-21-Theresa-Sauer\/dp\/0979554640\" target=\"_blank\" rel=\"nofollow noopener\">amazon.co.uk<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/I\/512mPwCOojL._SY400_.jpg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"https:\/\/www.amazon.co.uk\/Notations-21-Theresa-Sauer\/dp\/0979554640\" target=\"_blank\" rel=\"nofollow noopener\">Notations 21: Amazon.co.uk: Theresa Sauer: 9780979554643: Books<\/a><\/h3>\n<p>\nDrawing inspiration from John Cages Notations, Notations 21 features illustrated musical scores from more than 100 composers from every continent, all of whom are making amazing breakthroughs in the a ...\n<\/p>\n  <\/article>\n  \n<\/aside>\n<p>A ton of fascinating scores in there.<\/p>\n<p>On a more digital front, there\u2019s lots of dynamic\/generative score stuff going on nowadays too.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p><a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/rodrigo\">@Rodrigo<\/a> yes!<br \/>Funny you posted this, as I was recently looking to get one of these, until I noticed the price! In US this book is listed as over $600!<br \/>Certainly prohibitive for me <img src=\"https:\/\/llllllll.co\/images\/emoji\/twitter\/smile.png?v=5\" title=\":smile:\" class=\"emoji\" alt=\":smile:\" \/><br \/>But it really looks like an awesome compendium of exactly the kind of materials I thought we could share here\u2026<br \/>Do you actually own a copy?<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Holy shit that\u2019s crazy!<\/p>\n<p>Yeah I own the book. Bought it (in the UK) a few years ago.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Lucky you <a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/rodrigo\">@Rodrigo<\/a> !<br \/>So, obviously, you must think it\u2019s worth the price?<br \/>I would LOVE to have this in my library!..<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>DEFINITELY not worth that price, lol.<\/p>\n<p>I don\u2019t remember how much it was, but it was when I got it but it was no more than like \u00a340-50 or so. I\u2019ve got the hardcover version too. Weird\u2026<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>this thread is simply amazing! i wish it was around a couple of years ago\u2026 i was working on a project and looked into experimental notation for a commission i was doing. its a pretty tenuous connection, but i wrote a little article about the piece here:<\/p>\n<aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"http:\/\/www.juggle.org\/\/juggle.org\/wp-content\/uploads\/2016\/07\/IJA_original.png\" class=\"site-icon\" width=\"\" height=\"\" alt=\"image\" \/><a href=\"http:\/\/www.juggle.org\/center-of-gravity\/\" target=\"_blank\" title=\"08:30AM - 28 February 2014\" rel=\"nofollow noopener\">IJA \u2013 28 Feb 14<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/www.juggle.org\/wp-content\/uploads\/2014\/02\/COGFeatured-1.jpg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"http:\/\/www.juggle.org\/center-of-gravity\/\" target=\"_blank\" rel=\"nofollow noopener\">Center of Gravity<\/a><\/h3>\n<p>22 jugglers and 2 musicians took the stage on November 12, 2013, to perform a show called<\/p>\n  <\/article>\n  \n<\/aside><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Hi <a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/instantjuggler\">@instantjuggler<\/a> , thank you for the link! It was very interesting to read such a detailed insight into the process of putting together a large scale juggling performance. It made me think of your collaboration with Pekka as well (thinking about themusic and juggling relationships, but also the ability to effortlessly move between scored events and improvisation).<br \/>I also love how you refer to the idea of constructing choreographed images on stage. I always think of performance in terms of images as well, regardless of if I am working with dancers or volunteer workers, or by myself.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Here are a few from John Cage:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/176\/b48a40477c97b326.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/db02ffd128beacd134d5d056f75865ff2f6f188d\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/aec\/efc\/9379abd1bb_690x447.jpg\" width=\"690\" height=\"447\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/177\/1705fbeee47a0cc1.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/d0941095fc14fa94e5c782750ac4c4f1c4100de6\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/b58\/409\/179a4326ea_690x449.jpg\" width=\"690\" height=\"449\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/178\/38b5633016e3568c.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/dbbd38eae19fe9d2469646e3b7d34409d4f52a2a\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/379\/e51\/41b1323e7b_690x441.jpg\" width=\"690\" height=\"441\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/179\/36b30b6e37030d40.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/ad6978445ecc003f028a6836ac89be34fcf22aae\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/39f\/a40\/25a1b65a1f_690x343.jpg\" width=\"690\" height=\"343\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/180\/88787074fecb636f.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/00521422466490af946811a3199bdeec56306f30\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/6d5\/ccc\/7087ce561c_690x479.jpg\" width=\"690\" height=\"479\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/181\/41ecaa742f9f9cfa.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/7450267ad87dbdf122e2e4d10aea4a94cf9232d7\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/c83\/e02\/11d8646db5_492x500.jpg\" width=\"492\" height=\"500\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Woah that last one is crazy\u2026Most of the ones posted so far are interesting if not still stuck in a linear left to right structure. How are you supposed to interpret what looks like molecules and an unfinished pentagram?<\/p>\n<p>Cool stuff!<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>The instructions from that Cage piece is fucking mad too. The notation is weird, but each one of those little diagrams(\/pieces) has equally crazy instructions.<\/p><\/div>","created_at":"2018-11-07T16:39:03+0000","updated_at":"2018-11-07T16:39:03+0000","published_at":"2015-03-28T16:43:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"llllllll.co","preview_picture":"https:\/\/static-media.smithsonianmag.com\/img\/favicon.ico","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":213,"title":"Figwheel with new Clojure 1.9 CLI tools","url":"http:\/\/www.functionalbytes.nl\/clojure\/nodejs\/figwheel\/repl\/clojurescript\/cli\/2017\/12\/20\/tools-deps-figwheel.html","content":"<p>With the <a href=\"http:\/\/blog.cognitect.com\/blog\/clojure19\">release<\/a> of Clojure 1.9, Cognitect has added command line tools for Clojure.\nUsing those tools it becomes quite easy to run a Clojure script\/application or start a REPL.\nNext to this, it can also be used to fetch project dependencies, form a classpath with those dependencies and start a script using this classpath.<\/p><p>Because of the latter feature, the new CLI tools seem suitable for building ClojureScript projects, instead of using leiningen or boot.\nAll you need is your own build script, started using the CLI tools.\nFor a recently started ClojureScript on NodeJS project, we set out to find a workable setup for building and developing it using such a build script.\nA workable setup for us would at least include being able to compile and test the project, having a <a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\">Figwheel REPL<\/a> available and being able to connect to it using <a href=\"https:\/\/cider.readthedocs.io\/\">Emacs CIDER<\/a>.<\/p><p>This blog post shows our minimal but workable setup.<\/p><h3 id=\"the-depsedn-file\">The deps.edn file<\/h3><p>To declare the dependencies of the project, a <code class=\"highlighter-rouge\">deps.edn<\/code> file is required by the CLI tools.\nA minimal file would have the following contents:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">{:deps {org.clojure\/clojurescript {:mvn\/version \"1.9.946\"}\n :aliases\n {:repl {:extra-deps\n         {;; Figwheel ClojureScript REPL\n          com.cemerick\/piggieback {:mvn\/version \"0.2.2\"\n                                   :exclusions  [com.google.javascript\/closure-compiler]}\n          figwheel-sidecar        {:mvn\/version \"0.5.14\"\n                                   :exclusions  [com.google.javascript\/closure-compiler]}\n          ;; CIDER compatible nREPL\n          cider\/cider-nrepl       {:mvn\/version \"0.15.1\"}\n          org.clojure\/tools.nrepl {:mvn\/version \"0.2.12\"}\n          refactor-nrepl          {:mvn\/version \"2.3.1\"}}}}}<\/pre><\/figure><p>The <code class=\"highlighter-rouge\">:deps<\/code> map declares the projects dependencies required to build it.\nNext to this, there is also the <code class=\"highlighter-rouge\">:repl<\/code> alias, which declares the dependencies required to start a CIDER-compatible Figwheel ClojureScript REPL.\nIf your project does not need CIDER compatibility, you could leave out the <code class=\"highlighter-rouge\">cider\/cider-nrepl<\/code>, <code class=\"highlighter-rouge\">org.clojure\/tools.nrepl<\/code> and <code class=\"highlighter-rouge\">refactor-nrepl<\/code> dependencies.<\/p><p>When you use the <code class=\"highlighter-rouge\">clojure<\/code> (or <code class=\"highlighter-rouge\">clj<\/code>) CLI tool, it will use the dependencies in the <code class=\"highlighter-rouge\">:deps<\/code> map.\nIf you want the REPL dependencies mixed in as well, you need to supply the <code class=\"highlighter-rouge\">:repl<\/code> alias using the <code class=\"highlighter-rouge\">-R<\/code> option, like so:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clj -R:repl &lt;your-script.clj&gt;<\/pre><\/figure><h3 id=\"the-build-script\">The build script<\/h3><p>Next is the build script.\nThe goal is to have the script perform a certain task, such as <code class=\"highlighter-rouge\">compile<\/code>, <code class=\"highlighter-rouge\">test<\/code> or <code class=\"highlighter-rouge\">figwheel<\/code>, and thus use the CLI tools like this:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clojure build.clj &lt;task&gt; [task args]<\/pre><\/figure><p>For this to work, the following \u201cframework\u201d is created in the <code class=\"highlighter-rouge\">build.clj<\/code> script:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">;;; Task dispatching.\n(defmulti task first)\n(defmethod task :default\n  [args]\n  (let [all-tasks  (-&gt; task methods (dissoc :default) keys sort)\n        interposed (-&gt;&gt; all-tasks (interpose \", \") (apply str)))]\n    (println \"Unknown or missing task. Choose one of:\" interposed)\n    (System\/exit 1)))\n;;; Build script entrypoint. This should be the last expression.\n(task *command-line-args*)<\/pre><\/figure><p>Now we can define a <code class=\"highlighter-rouge\">task<\/code> method for each of the tasks we want, dispatched on the the <code class=\"highlighter-rouge\">first<\/code> argument to the build script.<\/p><h4 id=\"compiling-the-project\">Compiling the project<\/h4><p>Let\u2019s add the first task for compiling the project.\nWe need to require the ClojureScript compiler API namespace and use it to call the ClojureScript compiler:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(require '[cljs.build.api :as api])\n(def source-dir \"src\")\n(def compiler-config {:main          'my-app.core\n                      :output-to     \"target\/my-app\/main.js\"\n                      :output-dir    \"target\/my-app\/main\"\n                      :target        :nodejs\n                      :optimizations :simple\n                      :source-map    \"target\/my-app\/main.js.map\"})\n(defmethod task \"compile\" [args]\n  (api\/build source-dir compiler-config))<\/pre><\/figure><p>Easy and effective.\nExecuting <code class=\"highlighter-rouge\">clojure build.clj compile<\/code> will compile the project for us.\nAs our project is a NodeJS project, we would now be able to run it by executing <code class=\"highlighter-rouge\">node target\/my-app\/main.js<\/code>.<\/p><h4 id=\"testing-the-project\">Testing the project<\/h4><p>Next, testing our project.\nFor testing a ClojureScript project, one generally uses a different entrypoint into the application -  the test runner.\nThe same is true for our NodeJS project.\nWe simply execute the NodeJS project from within our build script using the <code class=\"highlighter-rouge\">clojure.java.shell<\/code> namespace:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(require '[clojure.java.shell :as shell])\n(def test-dir \"test\")\n(def test-config {:main          'my-app.test-runner\n                  :output-to     \"target\/test.js\"\n                  :output-dir    \"target\/test\"\n                  :target        :nodejs\n                  :optimizations :none\n                  :source-map    true})\n(defn run-node-tests []\n  (let [{:keys [out err]} (shell\/sh \"node\" \"target\/test.js\")]\n    (println out err)))\n(defn test-once []\n  (api\/build (api\/inputs source-dir test-dir) test-config)\n  (run-node-tests)\n  (shutdown-agents))\n(defmethod task \"test\" [args]\n  (test-once))<\/pre><\/figure><p>Next to a one-off \u201ctest\u201d task, it is very convenient to have the project tested after every change in the source code.\nThis workflow can be had by extending the \u201ctest\u201d task a bit, using the <code class=\"highlighter-rouge\">:watch-fn<\/code> option of the ClojureScript compiler.\nOn every change, the tests will be re-run, without a new JVM spinning up.\nThe \u201ctest\u201d task can be extended as follows:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(defn test-refresh\n  (api\/watch (api\/inputs source-dir test-dir)\n             (assoc test-config :watch-fn run-node-tests)))\n(defmethod task \"test\" [[_ type]]\n  (case type\n    (nil \"once\") (test-once)\n    \"watch\"      (test-refresh)\n    (do (println \"Unknown argument to test task:\" type)\n        (System\/exit 1))))<\/pre><\/figure><p>Now you can execute <code class=\"highlighter-rouge\">clojure build.clj test watch<\/code> to start this continuously testing workflow.\nThe one-off test can still be executed by leaving out the <code class=\"highlighter-rouge\">watch<\/code> argument, or using the <code class=\"highlighter-rouge\">once<\/code> argument.<\/p><h4 id=\"figwheel-repl-driven-development\">Figwheel REPL-driven development<\/h4><p>Now let\u2019s setup a Figwheel REPL task.\nLuckily for us, the Figwheel project is nicely split up between leiningen-specific stuff and the actual compilation, code reloading and REPL mechanics.\nThe latter is what we need, and is available in the <a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\/tree\/master\/sidecar\">sidecar<\/a> library.\nThis library compiles the ClojureScript project with Figwheel client related code inserted and starts the Figwheel server.\nThe Figwheel server re-compiles the project on every source code change and signals the client to reload some files after compilation is complete.<\/p><p>The server can start a ClojureScript REPL for you, hooking into the running application.\nIt can also start an nREPL server for you, such that you can connect to it using CIDER, where you can start the ClojureScript REPL yourself.\nIn our setup, we want to be able to choose either one of these \u201cmodes\u201d, by optionally supplying an nREPL server port.\nThe \u201cfigwheel\u201d task definition then looks as follows:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(require '[figwheel-sidecar.repl-api :as repl-api :refer [cljs-repl]])\n(def dev-config (merge compiler-config\n                       {:optimizations :none\n                        :source-map    true}))\n(defmethod task \"figwheel\" [[_ port]]\n  (repl-api\/start-figwheel!\n   {:figwheel-options (when port\n                        {:nrepl-port       (some-&gt; port Long\/parseLong)\n                         :nrepl-middleware [\"cider.nrepl\/cider-middleware\"\n                                            \"refactor-nrepl.middleware\/wrap-refactor\"\n                                            \"cemerick.piggieback\/wrap-cljs-repl\"]})\n    :all-builds       [{:id           \"dev\"\n                        :figwheel     true\n                        :source-paths [source-dir]\n                        :compiler     dev-config}]})\n  (when-not port\n    (cljs-repl)))<\/pre><\/figure><p>If we launch our Figwheel setup as follows, we get a simple REPL in our terminal:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clj -R:repl build.clj figwheel<\/pre><\/figure><p>And if we launch as follows, we get an nREPL server, which we can connect to in Emacs by using the <code class=\"highlighter-rouge\">cider-connect<\/code> command.\nIn this CIDER repl, you can call <code class=\"highlighter-rouge\">(cljs-repl)<\/code> yourself, to open the ClojureScript REPL.<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clojure -R:repl build.clj figwheel &lt;nrepl-port&gt;<\/pre><\/figure><p>Very nice.\nBut, if you try to run our former tasks, such as \u201ccompile\u201d or \u201ctest\u201d, you will see that those won\u2019t work if you don\u2019t add the <code class=\"highlighter-rouge\">-R:repl<\/code> option.\nIt will complain about not being able to load the <code class=\"highlighter-rouge\">figwheel-sidecar.repl-api<\/code> namespace.\nWe don\u2019t want to add the <code class=\"highlighter-rouge\">repl<\/code> alias for all our tasks, for various reasons.\nTherefore we have to make our task definitions check whether the required namespaces are available.\nFor this we introduce a small macro:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(defn try-require [ns-sym]\n  (try (require ns-sym) true (catch Exception e false)))\n(defmacro with-namespaces\n  [namespaces &amp; body]\n  (if (every? try-require namespaces)\n    `(do ~@body)\n    `(println \"task not available - required dependencies not found\")))<\/pre><\/figure><p>Using this macro, we can update our \u201cfigwheel\u201d task a bit:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(defmethod task \"figwheel\" [[_ port]]\n  (with-namespaces [figwheel-sidecar.repl-api]\n    (figwheel-sidecar.repl-api\/start-figwheel!\n     {:figwheel-options (when port\n                          {:nrepl-port       (some-&gt; port Long\/parseLong)\n                           :nrepl-middleware [\"cider.nrepl\/cider-middleware\"\n                                              \"refactor-nrepl.middleware\/wrap-refactor\"\n                                              \"cemerick.piggieback\/wrap-cljs-repl\"]})\n      :all-builds       [{:id           \"dev\"\n                          :figwheel     true\n                          :source-paths [source-dir]\n                          :compiler     dev-config}]})\n    (when-not port\n      (figwheel-sidecar.repl-api\/cljs-repl)))<\/pre><\/figure><p>We can now remove the global require of <code class=\"highlighter-rouge\">figwheel-sidecar.repl-api<\/code> namespace.\nIf we then try to execute the \u201cfigwheel\u201d task without the <code class=\"highlighter-rouge\">:repl<\/code> alias, it will show a message that the required namespaces are missing.\nOther tasks now work again as expected.<\/p><h3 id=\"conclusions\">Conclusions<\/h3><p>Building, testing and developing a ClojureScript application using the new Clojure tools is certainly possible.\nIf anything, it was a nice excercise in how to work with the new tools.\nCreating your own build script makes it very explicit as to how the project is build and tested, instead of \u201cjust add these plugins, and it will probably work\u201d.\nIn other words, in may create insight in how it all works and fits together.<\/p><p>That said, the custom build script implements concepts like tasks, which tools like leiningen and boot have implemented and standardized for you.\nSimply being able to execute <code class=\"highlighter-rouge\">cider-jack-in-clojurescript<\/code> from within Emacs is easier and works because of those standardized build tools.\nThis is fine, as the new Clojure CLI tools and tools like leiningen and boot can complement each other.\nBecause of this, we will probably see more support for the Clojure tools, and the underlying <code class=\"highlighter-rouge\">tools.deps<\/code> library, in those build tools.\nInitial examples of such integration are the <a href=\"https:\/\/github.com\/seancorfield\/boot-tools-deps\">boot-tools-deps<\/a> task and <a href=\"https:\/\/github.com\/RickMoynihan\/lein-tools-deps\">lein-tools-deps<\/a> plugin.<\/p><p>A full build script can be found <a href=\"https:\/\/github.com\/aroemers\/clojure-cli-tools-build\">on github<\/a>.<\/p><p><em>As always, have fun!<\/em><\/p>","created_at":"2018-11-07T16:39:03+0000","updated_at":"2018-11-07T16:39:03+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"www.functionalbytes.nl","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":210,"title":"My favorite apps on F-Droid  - [Quaap.com]","url":"https:\/\/quaap.com\/D\/use-fdroid","content":"<p> I often recommend F-Droid to my Android-using friends. <a href=\"https:\/\/f-droid.org\/\" target=\"_blank\">F-Droid<\/a> is an alternative app store which contains only Free and Open Source Software (FOSS). It's great because the apps are almost all not evil. Did you know that a flashlight app <em>doesn't<\/em> need access to your contacts,  location, and pictures? Did you know that every app doesn't need to serve you ads?\n<\/p><p>\n Yeah, sometimes the apps aren't as polished, and sometimes they lack some functionality you'd like, but they are much less likely to contain\/be malware or spyware. And sometimes they fill a unique niche not filled by anything else. And sometimes they are amazingly good.\n<\/p><p>\n The apps are some individual's or team's passion projects, and most are not looking to earn money on the apps, they just want to make something they and others would find useful or fun.  I wrote several of them myself, and I wrote them because I didn't like what was available commercially, or just had a weird idea.\n<\/p><p>\n  Anyway, below is a current list of these FOSS apps I use (or have installed for my kids \/ future use).  And again, one of the best features of any of them is that they don't serve you ads, and don't report your location and browsing habits anywhere.  And don't steal your data and contact list.\n<\/p><p>\n  To install them, you'll need to go to <a href=\"https:\/\/f-droid.org\/\" target=\"_blank\">F-Droid<\/a> and install the main F-Droid app.  When you do that, you'll need to say \"yes\" or \"allow\" to the prompts asking if you want to install from unknown sources.\n<\/p><p>\nNote: there are a huge number of apps I've not tried\u2026 These are just the ones I've installed and found useful.\n<\/p><h2 id=\"system-apps\">System apps<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.menny.android.anysoftkeyboard\" target=\"_blank\">AnySoftKeyboard<\/a><\/li>\n<\/ul><p>\nThis is a replacement for the keyboard.  It provides a much easier-to-use experience (to me) that the built-in keyboards don't, like easy access to the single quote, for example.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.launchtime\" target=\"_blank\">LaunchTime<\/a><\/li>\n<\/ul><p>\nOk, full disclosure: I wrote this one. <a href=\"https:\/\/quaap.com\/D\/LaunchTime\">LaunchTime<\/a> is my replacement homescreen\/launcher.  I do like it sooo much better than the other homescreens though.  Before I wrote it, I struggled greatly with the \"app drawer\" model used by most launchers. See the <a href=\"https:\/\/github.com\/quaap\/LaunchTime\/blob\/master\/README.md\" target=\"_blank\">LaunchTime page on Github<\/a> for a full list of features.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.jmstudios.redmoon\" target=\"_blank\">Red Moon<\/a><\/li>\n<\/ul><p>\nThis app is designed to reduce the display's effect on your day-night cycle by removing blue light.  I don't know about that, but I use it as a \"super dimmer\" at night and it can run on a schedule, too.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.thosp.yourlocalweather\" target=\"_blank\">Your local weather<\/a><\/li>\n<\/ul><p>\nA replacement for whatever spyware\/crapware weather widget came with your phone.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.vrem.wifianalyzer\" target=\"_blank\">WiFiAnalyzer<\/a><\/li>\n<\/ul><p>\nCan display a list of WiFi access points and help you decide on what channel to use.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/link.standen.michael.phonesaver\" target=\"_blank\">Phone Saver<\/a><\/li>\n<\/ul><p>\nAdds a \"save to a directory\" option to the share menus.  Run the app by itself first to set a destination directory.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/de.tobiasbielefeld.searchbar\" target=\"_blank\">Simple Search<\/a><\/li>\n<\/ul><p>\nA generic search app widget which you can target different search engines. It's pretty slim and mostly useful so you don't have to have some company's widget always running on your phone.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/de.onyxbits.listmyapps\" target=\"_blank\">List My Apps<\/a><\/li>\n<\/ul><p>\nI used this to help me make this list! <img src=\"https:\/\/quaap.com\/D\/media\/emoticons\/smile.png\" alt=\"smile :)\" title=\"smile :)\" \/><\/p><h2 id=\"graphics\">Graphics<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.simplemobiletools.gallery\" target=\"_blank\">Gallery<\/a><\/li>\n<\/ul><p>\nA gallery of you photos, a photo viewer, and a photo editor.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.witness.sscphase1\" target=\"_blank\">ObscuraCam<\/a><\/li>\n<\/ul><p>\nLoads up your pictures and automatically detects and blurs faces so that you can share pictures that contain other people\/kids easier.  You can also manually blur (or unblur) regions of the photo.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/anupam.acrylic\" target=\"_blank\">Acrylic Paint<\/a><\/li>\n<\/ul><p>\nFinger painting\/drawing.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.google.zxing.client.android\" target=\"_blank\">Barcode Scanner<\/a><\/li>\n<\/ul><p>\nSimple barcode\/QR code\/etc scanner.  Just scans them and adds the code to its history so you can share them via text, email, file, etc.\n<\/p><h2 id=\"reading-and-writing\">Reading and Writing<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.indywidualni.fblite\" target=\"_blank\">Face Slim<\/a><\/li>\n<\/ul><p>\nA super slimmed-down Facebook app.  It's just a thin wrapper around facebook's mobile site, but it's great because it allows you to not have the two behemoth Facebook apps (Facebook proper and Messenger) to bog down your phone and gather data.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/it.feio.android.omninotes.foss\" target=\"_blank\">Omni Notes FOSS<\/a><\/li>\n<\/ul><p>\nKind of a hybrid notes and checklist app which allows adding all kinds of rich text and pictures.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.bookymcbookface\" target=\"_blank\">Booky McBookface<\/a><\/li>\n<\/ul><p>\nOk, it's mine again.  An almost <em>too<\/em>-simple epub reader. Very few features, but that's how I like it.\n<\/p><p>\nA simple daily journal. Can tie in with your calendar.\n<\/p><h2 id=\"utilities\">Utilities<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.woefe.shoppinglist\" target=\"_blank\">Shopping List<\/a><\/li>\n<\/ul><p>\nMy very favorite shopping list app.  You can manually sort the items (I have mine in ALDI-order), and have multiple lists.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/de.j4velin.pedometer\" target=\"_blank\">Pedometer<\/a><\/li>\n<\/ul><p>\nIf your device has a step counter sensor, this app can use it to keep track of your steps, and that's it.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.secuso.privacyfriendlyintervaltimer\" target=\"_blank\">Interval Timer<\/a><\/li>\n<\/ul><p>\nAn all-you-need app to help you if you do interval training.\n<\/p><p>\nVery simple app to generate dice rolls.  There are other, bigger apps available if you do tabletop gaming.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/openfoodfacts.github.scrachx.openfood\" target=\"_blank\">OpenFoodFacts<\/a><\/li>\n<\/ul><p>\nAccess data gathered on food products.  It can scan an item's UPC barcode and retrieve nutrition info (if available).\n<\/p><p>\nA simple text editor.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.danielkim.soundrecorder\" target=\"_blank\">Sound Recorder<\/a><\/li>\n<\/ul><p>\nAn easy way to record sound files and share them.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/net.programmierecke.radiodroid2\" target=\"_blank\">RadioDroid<\/a><\/li>\n<\/ul><p>\nA music player with a big list of free music and radio streams.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.gateshipone.odyssey\" target=\"_blank\">Odyssey<\/a><\/li>\n<\/ul><p>\nA kind of cool music player for mp3s and oggs on you phone.\n<\/p><p>\nTune your guitar, or just entertain a child who likes to see waveforms.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.billthefarmer.scope\" target=\"_blank\">Oscilloscope<\/a><\/li>\n<\/ul><p>\nView waveforms, or just entertain a child who likes to see waveforms.\n<\/p><h2 id=\"games-and-toys\">Games and toys<\/h2><p>Here's just a list of games I occasionally play, or, more likely, my kids play.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.primary\" target=\"_blank\">Primary<\/a><\/li>\n<\/ul><p>\nGamified educational app.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/net.sourceforge.solitaire_cg\" target=\"_blank\">SolitaireCG<\/a><\/li>\n<\/ul><p>\nStraightforward solitaire.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/name.boyle.chris.sgtpuzzles\" target=\"_blank\">Puzzles<\/a><\/li>\n<\/ul><p>\nLike 30 weird puzzle games.\n<\/p><p>\nYup 2048.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/ch.logixisland.anuto\" target=\"_blank\">Anuto TD<\/a><\/li>\n<\/ul><p>\nAnother tower-defense game.\n<\/p><p>\nAnd a bunch more I won't describe, but I found them fun enough to keep (or my kids did).\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.jfedor.frozenbubble\" target=\"_blank\">Frozen Bubble<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.gunshippenguin.openflood\" target=\"_blank\">Open Flood<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/net.artificialworlds.rabbitescape\" target=\"_blank\">Rabbit Escape<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.happysanta.gd\" target=\"_blank\">Gravity Defied<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.serwylo.lexica\" target=\"_blank\">Lexica<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/se.tube42.kidsmem.android\" target=\"_blank\">Candy Memory<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.secuso.privacyfriendlydame\" target=\"_blank\">Checkers<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.eehouse.android.xw4\" target=\"_blank\">CrossWords<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.dodatheexploda\" target=\"_blank\">Doda The Exploda<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.petero.droidfish\" target=\"_blank\">DroidFish<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/csd.qtproject.minesweeper\" target=\"_blank\">Minesweeper<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.moire.opensudoku\" target=\"_blank\">Open Sudoku<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.phonefonefun\" target=\"_blank\">PhoneFoneFun<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/br.usp.ime.retrobreaker\" target=\"_blank\">Retro Breaker<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.fishberserker\" target=\"_blank\">Seafood Berserker!<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.mobilepearls.sokoban\" target=\"_blank\">Sokoban<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/wsdfhjxc.taponium\" target=\"_blank\">Taponium<\/a><\/li>\n<\/ul><h2 id=\"end\">End<\/h2><p>So that's it.  There are hundreds of other apps there which might fit your needs. Try <a href=\"https:\/\/f-droid.org\/\" target=\"_blank\">it<\/a>!\n<\/p><p>2018-06-22\n <a href=\"https:\/\/quaap.com\/D\/programming-index\">#programming<\/a> \u00a0 <a href=\"https:\/\/quaap.com\/D\/app-index\">#app<\/a> \u00a0<\/p>","created_at":"2018-11-07T16:38:38+0000","updated_at":"2018-11-07T16:38:38+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"quaap.com","preview_picture":"https:\/\/quaap.com\/D\/media\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":211,"title":"Deriving the Y Combinator in JavaScript","url":"https:\/\/enlight.nyc\/projects\/y-combinator\/","content":"<h2>What is the Y Combinator?<\/h2><p>In order to understand the Y combinator, one must first understand recursion. Here\u2019s a definition.<\/p><blockquote>\n<p>recursion (noun)<br \/><em>see \u201crecursion\u201d<\/em><\/p>\n<\/blockquote><p>Kidding! But nonetheless, this definition captures the essence of recursion: self-reference. In our case \u2014 programming \u2014 recursion allows for the succint description of a repeating process.<\/p><p>Recursion can be puzzling at first, but oftentimes, it allows us to write code that is closer to the true intent of what we wish to do than an imperative, loopy style. Here are a few examples.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/**\n * Imperative style\n *\/\nfunction factorialA(x) {\n  let result = 1;\n  \/\/ What exactly is happening here?\n  for (let i = x; i--; i &gt;= 1) {\n    result *= x;\n  }\n  return x;\n}\n\/**\n * Recursive style\n *\/\nfunction factorialB(x) {\n  if (x === 0) {\n    return 1;\n  }\n  \/\/ Recursive call (function calls itself!)\n  \/\/ Stays true to definition of factorial: x times the factorial of x - 1\n  return x * factorial(x - 1);\n}<\/pre><\/div><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/**\n * Imperative style\n *\/\nfunction fibonacciA(n) {\n  \/\/ Term 0: 1, Term 1: 1, Term 2: 2\n  n = n - 1;\n  let prevTerm = 1;\n  let currTerm = 1;\n  \/\/ What does this loop do, exactly?\n  for (let i = 0; i &lt; n; i += 1) {\n    let temp = prevTerm;\n    prevTerm = currTerm;\n    currTerm += temp;\n  }\n  return currTerm;\n}\n\/**\n * Recursive style\n *\/\nfunction fibonacciB(n) {\n  if (n === 0 || n === 1) {\n    return 1;\n  }\n  \/\/ Definition of next Fibonacci term is sum of two previous\n  \/\/ Easy to follow!\n  return fibonacci(n - 2) + fibonacci(n - 1);\n}<\/pre><\/div><p>Many high-level languages, JavaScript included, support recursion. But some languages don\u2019t, and in the field of theoretical computer science, introducing recursion into non-recursive contexts is a fascinating problem.<\/p><p>This is where the Y combinator comes in: it provides a way to recurse \u2014 to self-reference \u2014 <em>even if self-reference is unsupported in the base language<\/em>!<\/p><h2>Eliminating Recursion<\/h2><p>The Y combinator bestows the power of self-referential recursion unto languages which don\u2019t support it. So to derive it in JavaScript, let\u2019s pretend a little \u2014 let\u2019s pretend that JavaScript doesn\u2019t support recursion.<\/p><p>I\u2019ve put the recursive factorial function given in the initial examples in the block below. If everything goes well, by the end of this writeup we\u2019ll have a tool that will allow this factorial function to self-reference again \u2014 but most importantly, without any self-reference in the function itself!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function factorial(x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * factorial(x - 1);\n}<\/pre><\/div><p>The best way to begin removing the recursion from <code class=\"language-text\">factorial<\/code> looks like factoring out the call to <code class=\"language-text\">factorial<\/code> from the function body and making it a parameter, so let\u2019s do that.<\/p><h3>Bringing Our Own<\/h3><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ Bring your own factorial function!\nfunction byoFactorial(f, x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * f(x - 1);\n}<\/pre><\/div><p>With this new factorial function, <code class=\"language-text\">byoFactorial<\/code>, we have to \u201cbring our own\u201d working <code class=\"language-text\">factorial<\/code> function <code class=\"language-text\">f<\/code> to pass into the new, non-recursive function as a parameter, to replace our old self-referential call.<\/p><p>Of course, the problem now is that we don\u2019t have a working <code class=\"language-text\">factorial<\/code> function anymore. But that doesn\u2019t mean we have to give up! We can start small \u2014 from zero.<\/p><p>Let\u2019s begin by creating a function <code class=\"language-text\">f<\/code>, that when passed to <code class=\"language-text\">byoFactorial<\/code>, will make it work, but only for <code class=\"language-text\">x = 0 (0! = 1)<\/code>.<\/p><blockquote>\n<p>0! is defined to be equal to 1 \u2014 but why? It comes from combinatorics, the study of counting and choosing. In combinatorics, the definition of <code class=\"language-text\">n!<\/code> is the number of orders and ways <code class=\"language-text\">n<\/code> items can be chosen (i.e. permutations).<\/p>\n<p>For instance, <code class=\"language-text\">2! = 2<\/code> is the number of ways you can select and order two items \u2014 the first item, then the second; or the second item, then the first. By that same token, <code class=\"language-text\">0!<\/code> is the number of ways you can select and order zero items \u2014 and you can do that in only <code class=\"language-text\">1<\/code> way.<\/p>\n<\/blockquote><p>How can we create this function <code class=\"language-text\">f<\/code> for which <code class=\"language-text\">byoFactorial<\/code> will work for <code class=\"language-text\">x = 0 (0! = 1)<\/code>? Well, closer inspection of our <code class=\"language-text\">byoFactorial<\/code> function shows that it already has a base case defined \u2014 when <code class=\"language-text\">x<\/code> is <code class=\"language-text\">0<\/code>, it returns <code class=\"language-text\">1<\/code>. Moreover, when <code class=\"language-text\">x = 0<\/code>, no self-reference occured anyway in the original <code class=\"language-text\">factorial<\/code> function \u2014 thus, <code class=\"language-text\">f<\/code> is not called.<\/p><p>That means we can pass any arbitrary function as <code class=\"language-text\">f<\/code>. When <code class=\"language-text\">x = 0<\/code>, <code class=\"language-text\">byoFactorial<\/code> will still return <code class=\"language-text\">1<\/code> regardless.<\/p><p>Here are some suggestions.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function identity(x) {\n  return x;\n}\nfunction one(x) {\n  return 1;\n}\nfunction random(x) {\n  return Math.random();\n}\n\/\/ I like to use `random` because you'll\n\/\/ immediately know if your function stops\n\/\/ working - you'll get a long and ugly\n\/\/ decimal. With the `identity` or `one` functions,\n\/\/ problems with output are more likely to slip\n\/\/ through. For the rest of this writeup, I'll\n\/\/ use `random`.\nbyoFactorial(random, 0);\n\/\/ =&gt; 1 - it works!\nbyoFactorial(random, 1);\n\/\/ =&gt; 0.07382473983422022<\/pre><\/div><p>So far, we\u2019ve been passing a function <code class=\"language-text\">f<\/code> to <code class=\"language-text\">byoFactorial<\/code>, but we can just as easily create replacement <code class=\"language-text\">factorial<\/code> functions from <code class=\"language-text\">byoFactorial<\/code> by <code class=\"language-text\">bind<\/code>ing <code class=\"language-text\">f<\/code>. Using this method, we\u2019ll create new functions which we can call with just one parameter \u2014 our desired factorial, <code class=\"language-text\">x<\/code>.<\/p><blockquote>\n<p>In our case, <code class=\"language-text\">bind<\/code> takes two parameters: a <code class=\"language-text\">this<\/code> value and then the actual first parameter for <code class=\"language-text\">byoFactorial<\/code>, <code class=\"language-text\">f<\/code>. Since the body of <code class=\"language-text\">byoFactorial<\/code> doesn\u2019t access <code class=\"language-text\">this<\/code> at all, we can just set it to <code class=\"language-text\">null<\/code>.<\/p>\n<\/blockquote><div class=\"gatsby-highlight\"><pre class=\"language-js\">function byoFactorial(f, x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * f(x - 1);\n}\n\/\/ Pass null because there's nothing specific we need to set `this` to\nconst factorialUpTo0 = byoFactorial.bind(null, Math.random);\n\/\/ factorialUpTo0 will work with x = 0, and can be called like a regular function:\nfactorialUpTo0(0);\n\/\/ =&gt; 1<\/pre><\/div><p>Here\u2019s what we get:<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">  factorialUpTo0(0) =&gt; 1\n  \/\/ I passed Math.random as the first parameter so you'll probably get different results\n  factorialUpTo0(1) =&gt; 0.07382473983422022\n  factorialUpTo0(2) =&gt; 0.7045168327647198\n  factorialUpTo0(3) =&gt; 1.4614935135773812\n  factorialUpTo0(4) =&gt; 0.3311529571140488<\/pre><\/div><p>As expected, it works for <code class=\"language-text\">0!<\/code>, but nothing else. But it\u2019s progress! We\u2019re <code class=\"language-text\">1\/infinity<\/code> done our complete <code class=\"language-text\">factorial<\/code> function!<\/p><blockquote>\n<p><code class=\"language-text\">1\/infinity<\/code> is bigger than you think! Ever wondered what the sum of <code class=\"language-text\">1\/2 + 1\/3 + 1\/4 ... + 1\/infinity + 1\/(infinity + 1)<\/code> was? Yep, it\u2019s <code class=\"language-text\">infinity<\/code>.<\/p>\n<\/blockquote><div class=\"gatsby-highlight\"><pre class=\"language-js\">function byoFactorial(f, x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * f(x - 1);\n}<\/pre><\/div><p>Let\u2019s make <code class=\"language-text\">byoFactorial<\/code> work for both <code class=\"language-text\">0!<\/code> and <code class=\"language-text\">1!<\/code>. In other words, how can we create a function <code class=\"language-text\">f<\/code> that we can bind to <code class=\"language-text\">byoFactorial<\/code> that\u2019ll allow it to return correct outputs for <code class=\"language-text\">x = 0<\/code> and <code class=\"language-text\">x = 1<\/code>?<\/p><p>First, let\u2019s walk through what happens when we pass <code class=\"language-text\">x = 1<\/code> into <code class=\"language-text\">byoFactorial<\/code>. For whatever <code class=\"language-text\">f<\/code> we come up with, <code class=\"language-text\">byoFactorial(f, 1)<\/code> will return <code class=\"language-text\">1 * f(0)<\/code>. So, all we need is a factorial function <code class=\"language-text\">f<\/code> which returns the correct value for the case when <code class=\"language-text\">x = 0<\/code>. Do we already have a factorial function which works for <code class=\"language-text\">0!<\/code>?<\/p><p>We do, and it\u2019s called <code class=\"language-text\">factorialUpTo0<\/code>! Let\u2019s bind it to <code class=\"language-text\">byoFactorial<\/code>. This will create a factorial function which will work for both <code class=\"language-text\">x = 0<\/code> and <code class=\"language-text\">1<\/code> \u2014 or rather, all integer values from 0 up to 1 (this process will be easier to reason about if we think about it this way). We\u2019ll call this function <code class=\"language-text\">factorialUpTo1<\/code>.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo1 = byoFactorial.bind(null, factorialUpTo0);<\/pre><\/div><p>Before we keep going, let\u2019s refactor our <code class=\"language-text\">byoFactorial<\/code> function. Calling <code class=\"language-text\">.bind<\/code> is clunky (especially with that <code class=\"language-text\">null<\/code>!) and we can clean up our syntax by returning a function instead (in essence, creating a \u201cfactory\u201d).<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function factorialFactory(f) {\n  return function factorial(x) {\n    if (x === 0) {\n      return 1;\n    }\n    return x * f(x - 1);\n  };\n}<\/pre><\/div><p>Now, we can create <code class=\"language-text\">factorialUpTo0<\/code> and <code class=\"language-text\">factorialUpTo1<\/code> functions with much cleaner syntax.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo0 = factorialFactory(Math.random);\nconst factorialUpTo1 = factorialFactory(factorialUpTo0);\n\/\/ Instead of\n\/\/ const factorialUpTo1 = byoFactorial.bind(null, factorialUpTo0);<\/pre><\/div><p>And we can make <code class=\"language-text\">factorialUpTo2<\/code>, too!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo2 = factorialFactory(factorialUpTo1);<\/pre><\/div><p><code class=\"language-text\">factorialUpTo2<\/code> will work with for <code class=\"language-text\">x = 2<\/code>, because it will return <code class=\"language-text\">2 * factorialUpTo1(1)<\/code>, and we already know that <code class=\"language-text\">factorialUpTo1<\/code> will work up to 1, so we also know that <code class=\"language-text\">factorialUpTo2<\/code> will return correct outputs for all values up to 1 as well!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo3 = factorialFactory(factorialUpTo2);<\/pre><\/div><p>By the same token, <code class=\"language-text\">factorialUpTo3<\/code> will work for <code class=\"language-text\">3!<\/code>, because it will return <code class=\"language-text\">3 * factorialUpTo2(2)<\/code>, and we already know that <code class=\"language-text\">factorialUpTo2<\/code> will work on all values up to 2!<\/p><p>We can continue this pattern to create our complete, working, non-recursive factorial function.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo3 = factorialFactory(factorialUpTo2);\nconst factorialUpTo4 = factorialFactory(factorialUpTo3);\nconst factorialUpTo5 = factorialFactory(factorialUpTo4);\nconst factorialUpTo6 = factorialFactory(factorialUpTo5);\nconst factorialUpTo7 = factorialFactory(factorialUpTo6);\nconst factorialUpTo8 = factorialFactory(factorialUpTo7);\nconst factorialUpTo9 = factorialFactory(factorialUpTo8);\nconst factorialUpTo10 = factorialFactory(factorialUpTo9);\n\/\/ ...\nfactorialUpTo10(5);\n\/\/ =&gt; 120\nfactorialUpTo10(8);\n\/\/ =&gt; 40320\nfactorialUpTo10(10);\n\/\/ =&gt; 3628800\nfactorialUpTo10(11);\n\/\/ =&gt; 28570479.681626465\nfactorialUpTo10(12);\n\/\/ =&gt; 79140708.41594502<\/pre><\/div><blockquote>\n<p>Note that the very first function <code class=\"language-text\">f<\/code> we passed to <code class=\"language-text\">factorialFactory<\/code> to create <code class=\"language-text\">factorialUpTo0<\/code> (in my case, <code class=\"language-text\">Math.random<\/code>) is only called if we pass an <code class=\"language-text\">x<\/code> greater than the upper bound of our <code class=\"language-text\">factorialUpTo{x}<\/code> function. If we keep passing our newly created <code class=\"language-text\">factorialUpTo{x}<\/code> functions back into <code class=\"language-text\">factorialFactory<\/code>, ad infinitum, <code class=\"language-text\">Math.random<\/code> will never be called (because the upper bound will be at infinity), and our function will work over all values of <code class=\"language-text\">x<\/code>!<\/p>\n<\/blockquote><p>Here\u2019s what <code class=\"language-text\">factorialUpTo10<\/code> looks like, expanded.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo10 = factorialFactory(\n  factorialFactory(\n    factorialFactory(\n      factorialFactory(\n        factorialFactory(\n          factorialFactory(\n            factorialFactory(\n              factorialFactory(\n                factorialFactory(\n                  factorialFactory(factorialFactory(Math.random))\n                )\n              )\n            )\n          )\n        )\n      )\n    )\n  )\n);<\/pre><\/div><p>So, it looks like we\u2019ll need to apply <code class=\"language-text\">factorialFactory<\/code> to itself ad infinitum to get our working <code class=\"language-text\">factorial<\/code> function. How can we do that though?<\/p><div class=\"gatsby-highlight\"><pre class=\"language-text\">const factorialUpToInfinity = ???<\/pre><\/div><h2>Here Comes Y!<\/h2><p>The Y combinator is here to save the day! The Y combinator is a function which applies another function infinite times, and in simple terms can be defined like below.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return f(f(f(f(f(f(f(f(f(f(f(f(f(...)))))))))))));\n}<\/pre><\/div><p>Unfortunately, JavaScript can\u2019t evaluate the above expression. We need to simplify it \u2014 and to start, it helps to think about the definition of <code class=\"language-text\">y<\/code> in the context of infinite sums.<\/p><h3>A Math Problem<\/h3><p>What\u2019s the sum of <code class=\"language-text\">sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 ...)))))<\/code>?<\/p><p>The solution is simple. Let\u2019s call the sum <code class=\"language-text\">S<\/code>, so <code class=\"language-text\">S = sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 ...)))))<\/code>.<\/p><p>Now, if we look closely at the expression, we can see <code class=\"language-text\">S<\/code> is a part of itself! The sum can alternatively be written as <code class=\"language-text\">S = sqrt(2 + S)<\/code>.<\/p><p>Squaring both sides, we get <code class=\"language-text\">S^2 = S + 2<\/code>. Moving all terms of <code class=\"language-text\">S<\/code> to the left side, we get <code class=\"language-text\">S^2 - S - 2 = 0<\/code>. Factored, this is <code class=\"language-text\">(S - 2)(S + 1) = 0<\/code>, so the two solutions to this equation are <code class=\"language-text\">S = -1<\/code> and <code class=\"language-text\">S = 2<\/code>. Since <code class=\"language-text\">sqrt<\/code> returns the principal (non-negative) square root, our solution must be <code class=\"language-text\">S = 2<\/code>.<\/p><p>We can apply this same principle to our original formulation of the Y-combinator: <code class=\"language-text\">y(f) = f(f(f(f(f(f(f(f(f(f(f(f(f(...)))))))))))))<\/code>. Looking closely, it\u2019s clear that <code class=\"language-text\">y(f)<\/code> is a part of the equation for <code class=\"language-text\">y(f)<\/code>! We can write the new equation for <code class=\"language-text\">y(f)<\/code> like so: <code class=\"language-text\">y(f) = f(y(f))<\/code>.<\/p><p>We can wrap that expression in another function so we can explicitly pass our <code class=\"language-text\">x<\/code> parameter, too.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return function(x) {\n    return f(y(f))(x);\n  };\n}<\/pre><\/div><p>There\u2019s another reason we can\u2019t define our version of <code class=\"language-text\">y<\/code> as simply <code class=\"language-text\">y(f) = f(y(f))<\/code>.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return f(y(f));\n}<\/pre><\/div><p>When we evaluate the above, we get this:<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const tenFactorial = y(factorialFactory)(10);\n\/\/ =&gt; RangeError: Maximum call stack size exceeded<\/pre><\/div><p>The problem is that JavaScript tries to evaluate <code class=\"language-text\">y(factorialFactory)<\/code>, which never finishes, before doing any computation. Let\u2019s quickly walk through the evaluation.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return f(y(f));\n}\ny(factorialFactory)(10);\nfactorialFactory(y(factorialFactory))(10);\nfactorialFactory(factorialFactory(y(factorialFactory)))(10);<\/pre><\/div><p>This continues until we get the stack overflow error.<\/p><p>On the other hand, when we wrap the expression in an outer function to pass the parameter <code class=\"language-text\">x<\/code>, we also defer evaluation of <code class=\"language-text\">y(f)<\/code>. The term for this type of outer function is a \u201cthunk\u201d, which you may have heard from libraries like redux-thunk, which does the same thing but for Redux, a popular state management library often paired with React.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return function thunk(x) {\n    return f(y(f))(x);\n  };\n}<\/pre><\/div><p>Let\u2019s walk through the thunked evaluation.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function factorialFactory(f) {\n  return function factorial(x) {\n    if (x === 0) {\n      return 1;\n    }\n    return x * f(x - 1);\n  };\n}\ny(factorialFactory)(10)\nthunk(10) {\n  return factorialFactory(y(factorialFactory))(10)\n}\nfactorialFactory(y(factorialFactory))(10)\nfactorialFactory(function thunk(x) {\n  return factorialFactory(y(factorialFactory))(x)\n})(10)\n\/\/ At this point, the first argument to factorialFactory\n\/\/ has been fully evaluated (the thunk is the result of\n\/\/ the evalation)! Now, the body of factorialFactory can\n\/\/ actually be run.\n10 * (function thunk(x) {\n  return factorialFactory(y(factorialFactory))(x);\n})(9)\n10 * factorialFactory(y(factorialFactory))(9);\n\/\/ And so on<\/pre><\/div><p>The evaluation semantics of JavaScript are really the only thing preventing us from defining the Y combinator as we had it originally (<code class=\"language-text\">y(f) = f(y(f))<\/code>). In other languages, it\u2019s completely acceptable!<\/p><p>In Haskell, a functional language, the Y combinator <em>can<\/em> be defined as we did above initially:\n<code class=\"language-text\">y(f) = f(y(f))<\/code>.\nThe reason this is possible is because Haskell\u2019s evaluation semantics are different \u2014 Haskell is a lazy language. By default, it doesn\u2019t fully evaluate expressions until they are truly needed.<\/p><p>Here\u2019s how Haskell would evaluate <code class=\"language-text\">y(factorialFactory)(10)<\/code> above, given that <code class=\"language-text\">y<\/code> was defined as we had it originally.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ Our factorialFactory, in JavaScript for reference\nfunction factorialFactory(f) {\n  return function factorial(x) {\n    if (x === 0) {\n      return 1;\n    }\n    return x * f(x - 1);\n  };\n}<\/pre><\/div><div class=\"gatsby-highlight\"><pre class=\"language-python\">y(factorialFactory)(10)\n# Haskell will initially unpack the call to y by one layer\nfactorialFactory(y(factorialFactory))(10)\n# At this point, Haskell will look at the call to factorialFactory, instead of continuing to unpack the original y (which is what JavaScript does)\nreturn 10 * y(factorialFactory)(9)\n# Haskell will repeat the above process for the rest of the calls to y\n# ...\nreturn 10 * 9 * 8 * 7 * 6 * 5 * 4 * 3 * 2 * 1 * factorialFactory(y(factorialFactory))(0)\n# At this point, factorialFactory encounters its base case and returns 1\n# Haskell never computes the actual value of y(f)!<\/pre><\/div><h2>But There\u2019s Still Recursion!<\/h2><p>Our current definition <code class=\"language-text\">y<\/code> of the Y combinator still contains recursion, though! One of the main selling points of the Y combinator is that it allows us to add support for recursion to languages which don\u2019t support it \u2014 so there must be a way to get rid of the recursive call.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return function thunk(x) {\n    return f(y(f))(x);\n  };\n}<\/pre><\/div><p>We could take the same approach as we did before, factoring out the recursive call into a parameter and creating a \u201cfactory\u201d function, but that won\u2019t fly \u2014 as we saw in the <code class=\"language-text\">factorial<\/code> example, we\u2019d still need to apply that factory function ad infinitum to create a functional <code class=\"language-text\">y<\/code>.<\/p><p>Let\u2019s take a closer look at <code class=\"language-text\">y<\/code>, then. How can we define <code class=\"language-text\">y<\/code> in non-recursive terms? Well, we can take advantage of the fact that while we are voluntarily shunning self-reference in function bodies, we can still pass functions to themselves as parameters!<\/p><p>We\u2019ll need to make a clever substitution, though. Instead of defining <code class=\"language-text\">y(f)<\/code> as we have it now, let\u2019s define it as <code class=\"language-text\">y(f) = x(x)(f)<\/code>, for some other function <code class=\"language-text\">x<\/code>. In simpler terms, <code class=\"language-text\">y = x(x)<\/code> \u2014 <code class=\"language-text\">y<\/code> is <code class=\"language-text\">x<\/code> applied to itself.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ Here's our current, recursively-defined y\nfunction y(f) {\n  return function thunk(n) {\n    return f(y(f))(n);\n  };\n}\n\/\/ Apply the substitution y = x(x)\n\/\/ Call the parameter otherX since we don't want\n\/\/ to confuse the outer x and inner x and to avoid\n\/\/ variable shadowing\nfunction x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}<\/pre><\/div><p>Let\u2019s try this out.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const newY = x(x);\n\/\/ x(x)(factorialFactory)\nconst newFactorial = newY(factorialFactory);\nnewFactorial(10);\n\/\/ =&gt; 3628800<\/pre><\/div><p>It works! Let\u2019s walk through the evaluation.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}\nx(x)(f);\nreturn function thunk(n) {\n  return f(x(x)(f))(n);\n};\n\/\/ Unthunked\nf(x(x)(f));\n\/\/ From first step above, we see that x(x)(f) = f(x(x)(f))\n\/\/ We can apply this substitution\nf(f(x(x)(f)));\n\/\/ And again, and again... look familiar?\nf(f(f(f(f(f(x(x)(f)))))));<\/pre><\/div><p>It works! By applying <code class=\"language-text\">x<\/code> to itself, we avoid self-reference in the function body. Everything the function needs to run comes from its parameters.<\/p><p>And with a little bit more refactoring and substitution, we\u2019ve got the non-recursive Y combinator!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ What we already have\nfunction x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      \/\/ By passing `x` to itself (in the form of `otherX`),\n      \/\/ we can technically \"call\" `x` inside the function\n      \/\/ body of `x` - but importantly, non-recursively! Of course,\n      \/\/ now that we've altered the parameters of `x` to take\n      \/\/ itself, we need to pass `otherX` to `otherX` in the\n      \/\/ return statement, too.\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}\n\/\/ Let's refactor\nconst y = x(x);\n\/\/ Here's `x` agian\nfunction x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}\n\/\/ Now, just substitute the body of `x`\n\/\/ for each `x` in the statement `y = x(x)`\nconst y = (function(otherX) {\n  return function(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n})(function(otherX) {\n  return function(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n});\n\/\/ If we want `y` to be usable on functions\n\/\/ with any number of arguments, we can use\n\/\/ the ES6 spread syntax over the arguments\nconst y = (function(otherX) {\n  return function(f) {\n    return function thunk(...args) {\n      return f(otherX(otherX)(f))(...args);\n    };\n  };\n})(function(otherX) {\n  return function(f) {\n    return function thunk(...args) {\n      return f(otherX(otherX)(f))(...args);\n    };\n  };\n});\n\/\/ We can shorten it even more with ES6 arrow functions\nconst y = (x =&gt; f =&gt; (...a) =&gt; f(x(x)(f))(...a))(x =&gt; f =&gt; (...a) =&gt;\n  f(x(x)(f))(...a)\n);<\/pre><\/div><p>And there you go! One of the most important concepts in theoretical computer science, sitting on your lap in just a few characters.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const y = (x =&gt; f =&gt; (...a) =&gt; f(x(x)(f))(...a))(x =&gt; f =&gt; (...a) =&gt;\n  f(x(x)(f))(...a)\n);<\/pre><\/div><h2>Appendix<\/h2><p>Try writing your own <code class=\"language-text\">y<\/code>-compatible, non-recursive implementations of your favorite recursive processes! Here\u2019s an example with the Fibonacci sequence mentioned earlier.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function fibonacciFactory(f) {\n  return function thunk(n) {\n    if (n === 0 || n === 1) {\n      return 1;\n    }\n    return f(n - 2) + f(n - 1);\n  };\n}\nconst fibonacci = y(fibonacciFactory);\nfibonacci(10);\n\/\/ =&gt; 89<\/pre><\/div><h2>About the Author<\/h2><p>Thanks for reading! My name is <a href=\"https:\/\/www.nathanhleung.com\">Nathan Leung<\/a>. I\u2019m the founder of location-based social messaging app <a href=\"https:\/\/www.getsuap.com\/\">Suap<\/a> and an incoming first-year student at the University of Michigan.<\/p><p>You can learn more about me on my <a href=\"https:\/\/www.linkedin.com\/in\/nathanhleung\/\">LinkedIn<\/a> or on my <a href=\"https:\/\/www.nathanhleung.com\">personal website<\/a>, and I encourage you to try <a href=\"https:\/\/www.getsuap.com\/\">Suap<\/a> at <a href=\"https:\/\/www.getsuap.com\/\">getsuap.com<\/a>!<\/p>","created_at":"2018-11-07T16:38:38+0000","updated_at":"2018-11-07T16:38:38+0000","annotations":[],"mimetype":"text\/html","reading_time":15,"domain_name":"enlight.nyc","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":208,"title":"How to implement a basic ActivityPub server","url":"https:\/\/blog.joinmastodon.org\/2018\/06\/how-to-implement-a-basic-activitypub-server\/","content":"<p>Today we\u2019ll be looking at how to connect the protocols powering Mastodon in the simplest way possible to enter the federated network. We will use static files, standard command-line tools, and some simple Ruby scripting, although the functionality should be easily adaptable to other programming languages.<\/p><p>First, what\u2019s the end goal of this exercise? We want to send a Mastodon user a message from our own, non-Mastodon server.<\/p>\n<p>So what are the ingredients required? The message itself will be formatted with ActivityPub, and it must be attributed to an ActivityPub actor. The actor must be discoverable via Webfinger, and the delivery itself must be cryptographically signed by the actor.<\/p>\n<h3 id=\"the-actor\">The actor<\/h3>\n<p>The actor is a publicly accessible JSON-LD document answering the question \u201cwho\u201d. JSON-LD itself is a quite complicated beast, but luckily for our purposes we can treat it as simple JSON with a <code>@context<\/code> attribute. Here is what an actor document could look like:<\/p>\n<div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n\t\"@context\": [\n\t\t\"https:\/\/www.w3.org\/ns\/activitystreams\",\n\t\t\"https:\/\/w3id.org\/security\/v1\"\n\t],\n\t\"id\": \"https:\/\/my-example.com\/actor\",\n\t\"type\": \"Person\",\n\t\"preferredUsername\": \"alice\",\n\t\"inbox\": \"https:\/\/my-example.com\/inbox\",\n\t\"publicKey\": {\n\t\t\"id\": \"https:\/\/my-example.com\/actor#main-key\",\n\t\t\"owner\": \"https:\/\/my-example.com\/actor\",\n\t\t\"publicKeyPem\": \"-----BEGIN PUBLIC KEY-----...-----END PUBLIC KEY-----\"\n\t}\n}<\/pre><\/div>\n<p>The <code>id<\/code> must be the URL of the document (it\u2019s a self-reference), and all URLs should be using HTTPS. You need to include an <code>inbox<\/code> even if you don\u2019t plan on receiving messages in response, because for legacy purposes Mastodon doesn\u2019t acknowledge inbox-less actors as compatible.<\/p>\n<p>The most complicated part of this document is the <code>publicKey<\/code> as it involves cryptography. The <code>id<\/code> will in this case refer to the actor itself, with a fragment (the part after <code>#<\/code>) to identify it\u2013this is because we are not going to host the key in a separate document (although we could). The <code>owner<\/code> must be the actor\u2019s <code>id<\/code>. Now to the hard part: You\u2019ll need to generate an RSA keypair.<\/p>\n<p>You can do this using OpenSSL:<\/p>\n<pre>openssl genrsa -out private.pem 2048\nopenssl rsa -in private.pem -outform PEM -pubout -out public.pem\n<\/pre>\n<p>The contents of the <code>public.pem<\/code> file is what you would put into the <code>publicKeyPem<\/code> property. However, JSON does not support verbatim line-breaks in strings, so you would first need to replace line-breaks with <code>\\n<\/code> instead.<\/p>\n<h3 id=\"webfinger\">Webfinger<\/h3>\n<p>What is Webfinger? It is what allows us to ask a website, \u201cDo you have a user with this username?\u201d and receive resource links in response. Implementing this in our case is really simple, since we\u2019re not messing with any databases and can hardcode what we want.<\/p>\n<p>The Webfinger endpoint is always under <code>\/.well-known\/webfinger<\/code>, and it receives queries such as <code>\/.well-known\/webfinger?resource=acct:bob@my-example.com<\/code>. Well, in our case we can cheat, and just make it a static file:<\/p>\n<div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n\t\"subject\": \"acct:alice@my-example.com\",\n\t\"links\": [\n\t\t{\n\t\t\t\"rel\": \"self\",\n\t\t\t\"type\": \"application\/activity+json\",\n\t\t\t\"href\": \"https:\/\/my-example.com\/actor\"\n\t\t}\n\t]\n}<\/pre><\/div>\n<p>The <code>subject<\/code> property here consists of the username (same as <code>preferredUsername<\/code> earlier) and the domain you\u2019re hosting on. This is how your actor will be stored on other Mastodon servers and how people will be able to mention it in toots. Only one link is required in the Webfinger response, and it\u2019s the link to the actor document.<\/p>\n<p>After this is uploaded to your webhost and available under your domain with a valid SSL certificate, you could already look up your actor from another Mastodon by entering <code>alice@my-example.com<\/code> into the search bar. Although it\u2019ll look quite barren.<\/p>\n<h3 id=\"the-message\">The message<\/h3>\n<p>ActivityPub messages practically consist of two parts, the message itself (the object) and a wrapper that communicates what\u2019s happening with the message (the activity). In our case, it\u2019s going to be a <code>Create<\/code> activity. Let\u2019s say \u201cHello world\u201d in response to my toot about writing this blog post:<\/p>\n<p>Here is how the document could look:<\/p>\n<div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n\t\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n\t\"id\": \"https:\/\/my-example.com\/create-hello-world\",\n\t\"type\": \"Create\",\n\t\"actor\": \"https:\/\/my-example.com\/actor\",\n\t\"object\": {\n\t\t\"id\": \"https:\/\/my-example.com\/hello-world\",\n\t\t\"type\": \"Note\",\n\t\t\"published\": \"2018-06-23T17:17:11Z\",\n\t\t\"attributedTo\": \"https:\/\/my-example.com\/actor\",\n\t\t\"inReplyTo\": \"https:\/\/mastodon.social\/@Gargron\/100254678717223630\",\n\t\t\"content\": \"&lt;p&gt;Hello world&lt;\/p&gt;\",\n\t\t\"to\": \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"\n\t}\n}<\/pre><\/div>\n<p>With the <code>inReplyTo<\/code> property we\u2019re chaining our message to a parent. The <code>content<\/code> property may contain HTML, although of course it will be sanitized by the receiving servers according to their needs \u2014 different implementations may find use for a different set of markup. Mastodon will only keep <code>p<\/code>, <code>br<\/code>, <code>a<\/code> and <code>span<\/code> tags. With the <code>to<\/code> property we are defining who should be able to view our message, in this case it\u2019s a special value to mean \u201ceveryone\u201d.<\/p>\n<p>For our purposes, we don\u2019t actually need to host this document publicly, although ideally both the activity and the object would be separately available under their respective <code>id<\/code>. Let\u2019s just save it under <code>create-hello-world.json<\/code> because we\u2019ll need it later.<\/p>\n<p>So the next question is, how do we send this document over, where do we send it, and how will Mastodon be able to trust it?<\/p>\n<h3 id=\"http-signatures\">HTTP signatures<\/h3>\n<p>To deliver our message, we will use POST it to the inbox of the person we are replying to (in this case, me). That inbox is <code>https:\/\/mastodon.social\/inbox<\/code>. But a simple POST will not do, for how would anyone know it comes from the real @alice@my-example.com and not literally anyone else? For that purpose, we need a HTTP signature. It\u2019s a HTTP header signed by the RSA keypair that we generated earlier, and that\u2019s associated with our actor.<\/p>\n<p>HTTP signatures is one of those things that are much easier to do with actual code instead of manually. The signature looks like this:<\/p>\n<pre>Signature: keyId=\"https:\/\/my-example.com\/actor#main-key\",headers=\"(request-target) host date\",signature=\"...\"\n<\/pre>\n<p>The <code>keyId<\/code> refers to public key of our actor, the <code>header<\/code> lists the headers that are used for building the signature, and then finally, the <code>signature<\/code> string itself. The order of the headers must be the same in plain-text and within the to-be-signed string, and header names are always lowercase. The <code>(request-target)<\/code> is a special, fake header that pins down the HTTP method and the path of the destination.<\/p>\n<p>The to-be-signed string would look something like this:<\/p>\n<pre>(request-target): post \/inbox\nhost: mastodon.social\ndate: Sun, 06 Nov 1994 08:49:37 GMT\n<\/pre>\n<p>Mind that there is only a \u00b130 seconds time window when that signature would be considered valid, which is a big reason why it\u2019s quite difficult to do manually. Anyway, assuming we\u2019ve got the valid date in there, we now need to build a signed string out of it. Let\u2019s put it all together:<\/p>\n<div class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\">require 'http'\nrequire 'openssl'\ndocument      = File.read('create-hello-world.json')\ndate          = Time.now.utc.httpdate\nkeypair       = OpenSSL::PKey::RSA.new(File.read('private.pem'))\nsigned_string = \"(request-target): post \/inbox\\nhost: mastodon.social\\ndate: #{date}\"\nsignature     = Base64.strict_encode64(keypair.sign(OpenSSL::Digest::SHA256.new, signed_string))\nheader        = 'keyId=\"https:\/\/my-example.com\/actor\",headers=\"(request-target) host date\",signature=\"' + signature + '\"'\nHTTP.headers({ 'Host': 'mastodon.social', 'Date': date, 'Signature': header })\n    .post('https:\/\/mastodon.social\/inbox', body: document)<\/pre><\/div>\n<p>Let\u2019s save it as <code>deliver.rb<\/code>. I am using the HTTP.rb gem here, so you\u2019ll need to have that installed (<code>gem install http<\/code>). Finally, run the file with <code>ruby deliver.rb<\/code>, and your message should appear as a reply on my toot!<\/p>\n<h3 id=\"conclusion\">Conclusion<\/h3>\n<p>We have covered how to create a discoverable ActivityPub actor and how to send replies to other people. But there is a lot we haven\u2019t covered: How to follow and be followed (it requires a working inbox), how to have a prettier profile, how to support document forwarding with LD-Signatures, and more. If there is demand, I will write more in-depth tutorials!<\/p>\n<p>Read more on:<\/p>\n<ul><li><a href=\"https:\/\/www.w3.org\/TR\/activitypub\/\">ActivityPub<\/a><\/li>\n<li><a href=\"https:\/\/tools.ietf.org\/html\/rfc7033\">Webfinger<\/a><\/li>\n<li><a href=\"https:\/\/tools.ietf.org\/html\/draft-cavage-http-signatures-10\">HTTP Signatures<\/a><\/li>\n<\/ul><p>\n            \u2014\n              \n                \n                Eugen Rochko\n              \n            ,\n            <\/p>","created_at":"2018-11-07T16:37:52+0000","updated_at":"2018-11-07T16:37:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"blog.joinmastodon.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":209,"title":"Browserless","url":"https:\/\/www.browserless.io\/blog\/2018\/08\/01\/running-an-indie-business.html","content":"<main id=\"main\" class=\"main main-1\"><div class=\"main-content\"><div class=\"container\"><div class=\"row d-flex align-items-center\"><div class=\"col-lg-6 offset-lg-0 col-md-10 offset-md-1 col-sm-10 offset-sm-1 text-center text-lg-left mb-5 mb-lg-0\"><div class=\"pr-0 pr-xl-5 pr-lg-0\"><p class=\"main-content__lead lead mb-4\">Stop spending time running and scaling Chrome, and more time on your application.<\/p><div class=\"row d-none d-lg-flex\"><div class=\"col-lg-6 col-md-6\"><ul class=\"features-list list-unstyled m-0\"><li class=\"features-list__item d-flex align-items-center\"><i class=\"features-list__item--icon mr-2\">Emoji's and fonts just work<\/i><\/li><li class=\"features-list__item d-flex align-items-center\"><i class=\"features-list__item--icon mr-2\">Zero-setup in your app<\/i><\/li><li class=\"features-list__item d-flex align-items-center\"><i class=\"features-list__item--icon mr-2\">Add or remove instances<\/i><\/li><\/ul><\/div><\/div><\/div><\/div><div class=\"col-lg-6 text-center text-lg-right position-relative\"><div class=\"code_frame__3LBQB\"><pre>\/\/ Replace puppeteer.launch with puppeteer.connect\nconst browser = await puppeteer.connect({\n  browserWSEndpoint: 'wss:\/\/chrome.browserless.io\/'\n});\n\/\/ Everything else stays the same\nconst page = await browser.newPage();\nawait page.goto('https:\/\/example.com\/');\nawait page.screenshot({ path: 'screenshot.png' });\nbrowser.close();<\/pre><\/div><\/div><\/div><\/div><\/div><\/main><section id=\"pricing\" class=\"pricing pricing-1\"><div class=\"container\"><div class=\"row\"><div class=\"col-lg-4\"><div class=\"pricing-card pricing-card\"><h5 class=\"pricing-card__title mb-4\">Small<\/h5><h3 class=\"pricing-card__value\">$30month<\/h3><ul class=\"pricing-card__list list-unstyled mb-0 pb-1\"><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>No maximum API or session calls<\/strong><i data-feather=\"help-circle\" data-toggle=\"tooltip\" data-placement=\"top\" title=\"Tooltip on top\"><\/i><\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>10<\/strong> running browsers<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Email support<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Static sites with images<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Dynamic sites with interactivity<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Large sites with webgl and video<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Live\/Chat support<\/i><\/li><\/ul><p><a href=\"https:\/\/www.browserless.io\/sign-up?size=small\" class=\"btn btn--blue btn--lg\">Get Started<\/a><\/p><\/div><\/div><div class=\"col-lg-4\"><div class=\"pricing-card pricing-card-highlighted\"><h6 class=\"text-uppercase mb-0\">Popular Option<\/h6><h5 class=\"pricing-card__title mb-4\">Medium<\/h5><h3 class=\"pricing-card__value\">$50month<\/h3><ul class=\"pricing-card__list list-unstyled mb-0 pb-1\"><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>No maximum API or session calls<\/strong><i data-feather=\"help-circle\" data-toggle=\"tooltip\" data-placement=\"top\" title=\"Tooltip on top\"><\/i><\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>15<\/strong> running browsers<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Email support<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Static sites with images<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Dynamic sites with interactivity<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Large sites with webgl and video<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Live\/Chat support<\/i><\/li><\/ul><p><a href=\"https:\/\/www.browserless.io\/sign-up?size=medium\" class=\"btn btn--green btn--lg\">Get Started<\/a><\/p><\/div><\/div><div class=\"col-lg-4\"><div class=\"pricing-card\"><h5 class=\"pricing-card__title mb-4\">Large<\/h5><h3 class=\"pricing-card__value\">$100month<\/h3><ul class=\"pricing-card__list list-unstyled mb-0 pb-1\"><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>No maximum API or session calls<\/strong><i data-feather=\"help-circle\" data-toggle=\"tooltip\" data-placement=\"top\" title=\"Tooltip on top\"><\/i><\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>20<\/strong> running browsers<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Email support<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Static sites with images<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Dynamic sites with interactivity<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Large sites with webgl and video<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Live\/Chat support<\/i><\/li><\/ul><p><a href=\"https:\/\/www.browserless.io\/sign-up?size=large\" class=\"btn btn--blue btn--lg\">Get Started<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"container\"><h5><a href=\"https:\/\/www.browserless.io\/commercial-license\">Click here<\/a> to purchase a commercial license.<\/h5><\/div><\/section><section id=\"features\" class=\"features-1 section section--padding__wmargin\"><div class=\"container\"><div class=\"row\"><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>Turn up the speed<\/h5><p class=\"pr-4\">Increase your app's performance by not having to manage Chrome or other browsers. Our smallest plan gives you access to 10 browser instances.<\/p><\/div><\/div><\/div><\/div><div class=\"row\"><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>No time restrictions<\/h5><p class=\"pr-4\">Sessions can run as long as you want, and even keep the browser open indefinitely.<\/p><\/div><\/div><\/div><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>No software or packages to install<\/h5><p class=\"pr-4\">Stop trying to get Chrome running in lambda, or fonts rendering properly and just use browserless.<\/p><\/div><\/div><\/div><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>Crucial notifications and metrics<\/h5><p class=\"pr-4\">Your account page shows crucial information like sessions and queues plus includes email notifications.<\/p><\/div><\/div><\/div><\/div><\/div><\/section><section id=\"integrations\" class=\"section section--padding section--gradient__1\"><div class=\"container\"><div class=\"row\"><div class=\"preamble preamble--light col-lg-8 offset-lg-2\"><h3 class=\"title\">How it works<\/h3><p class=\"lead\">browserless takes care of all the dependencies, sandboxing, and management of the web browser. You can connect remotely and drive it with open-source libraries, use our pre-built REST APIs, or write and run your own functions with webhooks.<\/p><\/div><\/div><div class=\"row d-flex align-items-center\"><div class=\"col-lg-5 order-2 order-lg-1 mt-5 mt-lg-0\"><ul class=\"features list-unstyled mb-0\"><li class=\"d-flex active\" data-example=\"puppeteer\"><div class=\"ml-4\"><h5>Puppeteer Connect<\/h5><p>If you're using puppeteer's launch method, simply change it to connect and specify it to use browserless.<\/p><\/div><\/li><li class=\"d-flex\" data-example=\"webdriver\"><div class=\"ml-4\"><h5>Webdriver<\/h5><p>Use libraries like selenium with our new \/webdriver route, which implements the webdriver protocol<\/p><\/div><\/li><li class=\"d-flex\" data-example=\"rest\"><div class=\"ml-4\"><h5>REST API<\/h5><p>Use our pre-made API's like \/screenshot or \/pdf to do common tasks, or write your own!<\/p><\/div><\/li><\/ul><\/div><div class=\"col-lg-7 order-1 order-lg-2\"><div class=\"code_frame__3LBQB\"><pre>\/\/ From inside your Node application\nimport puppeteer from 'puppeteer';\n\/\/ Replace puppeteer.launch with puppeteer.connect\nconst browser = await puppeteer.connect({\n  browserWSEndpoint: 'wss:\/\/chrome.browserless.io'\n});\n\/\/ The rest of your script remains the same\nconst page = await browser.newPage();\nawait page.goto('https:\/\/example.com\/');\nawait page.screenshot({ path: 'screenshot.png' });\npage.close();<\/pre><\/div><\/div><\/div><\/div><\/section><section id=\"clients\" class=\"clients-1 section section--padding\"><div class=\"container section--padding__bottom\"><div class=\"row\"><div class=\"preamble col-lg-8 offset-lg-2\"><h3 class=\"title\">Words from our users<\/h3><p class=\"lead\">With over 10 million sessions executed, here's what users have to say.<\/p><\/div><\/div><\/div><\/section><section id=\"questions\" class=\"section section--padding\"><div class=\"container\"><div class=\"row\"><h3 class=\"title\">Frequently Asked Questions<\/h3><\/div><div class=\"row\"><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>I'm not a developer, what's this for?<\/h5><p>You can use browserless to do <i>automatically<\/i> what you would normally do <i>manually<\/i>. For instance you can use it to take pictures of your web-site (for social media), or generate nice documents for your client's dashboards.<\/p><p>Getting Chrome, or any browser, to run on most web servers is pretty challenging as well. There's more than 20 additional packages you'll need to install, a way to manage the browser so it doesn't crash your servers, and a way to isolate your sessions so it's a fresh browser every time. We take care of all of that plus add notifications and a dashboard for you.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>Is there a way to check how my workers are doing?<\/h5><p>Yup! We have a GraphQL API that allows you to check your workers health prior to sending a request or connecting. You can use the online GraphiQL editor to see all the <a href=\"https:\/\/api.browserless.io\/graphql?query=%7B%0A%20%20pressure(apiToken%3A%20%22INSERT-YOUR-TOKEN-HERE%22)%20%7B%0A%20%20%20%20recentlyRejected%0A%20%20%20%20isAvailable%0A%20%20%20%20queued%0A%20%20%20%20running%0A%20%20%20%20date%0A%20%20%7D%0A%7D\">fields here<\/a>. We also have an admin dashboard that shows important statistics like rejected or queued request counts.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>Google Cloud supports Chrome, why use browserless?<\/h5><p>Cost is the first reason, as running a single browser can cost up to $42.77 a month at full utilization with Functions, and $44.38 using App Engine with their recommended settings. If you were to run 10 concurrently on Functions (which is what our $30 a month plan supports), your cost would be $427 a month. It's not uncommon to have cases where the browser won't close properly, meaning your functions over-run and end up costing more.<\/p><p>We also have numerous features over most cloud-providers, including: human-only support, numerous REST API's to do common actions, push-button upgrades, and (coming soon) support for other browsers plus Selenium web-drivers.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>I'm not using Node, can I still use this service?<\/h5><p>Even if you're not using NodeJS or puppeteer, you can still make use of our function API that allows you to send in a body of code to run for you. This allows you to keep your stack the same without making changes. You can read more about the <a href=\"https:\/\/docs.browserless.io\/blog\/2018\/05\/26\/metered-functions.html#browserless-functions\">function endpoint here<\/a>.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>What libraries do you support?<\/h5><p>Currently we support only puppeteer, but are looking to expand and integrate other libraries soon. If you can't use puppeteer due to the nature of your project you can use the <a href=\"https:\/\/docs.browserless.io\/blog\/2018\/05\/26\/metered-functions.html#browserless-functions\">function endpoint here<\/a>.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>How secure is my account?<\/h5><p>We don't store any sensitive information in browserless whatsoever. We use Stripe's payment API, so we never store credit card or other personal information, and we securely store your password with a one-way hashing algorithm. Your email is only used to contact you about updates to the service, and we don't sell or expose it to third-parties.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>How fast is the service?<\/h5><p>It generally takes about 5 minutes to generate the workers when you sign-up. When connecting via puppeteer.connect, this generally happens in less than 50 milliseconds. On average a screenshot capture happens in about a second for static pages, and PDFs take roughly 2 seconds.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>How reliable is the service?<\/h5><p>See for yourself on the <a href=\"https:\/\/status.browserless.io\/\">status page<\/a>. Since each account gets its own dedicated instances there's no over-crowding or congestion.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>Does this support proxies?<\/h5><p>Yes, we support using third-party proxies. See our documentation on <a href=\"https:\/\/docs.browserless.io\/docs\/chrome.html#launching-with-a-proxy\">how to use proxies<\/a>.<\/p><\/div><\/div><\/div><\/section>","created_at":"2018-11-07T16:37:52+0000","updated_at":"2018-11-07T16:37:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.browserless.io","preview_picture":"https:\/\/browserless.io\/assets\/browserless-social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":207,"title":"PHP: The Right Way","url":"https:\/\/phptherightway.com\/","content":"<section class=\"chapter\" id=\"welcome\"><p>There\u2019s a lot of outdated information on the Web that leads new PHP users astray,\npropagating bad practices and insecure code. <em>PHP: The Right Way<\/em> is an easy-to-read,\nquick reference for PHP popular coding standards, links to authoritative tutorials\naround the Web and what the contributors consider to be best practices at the present\ntime.<\/p><p><em>There is no canonical way to use PHP<\/em>. This website aims to introduce new PHP\ndevelopers to some topics which they may not discover until it is too late, and aims\nto give seasoned pros some fresh ideas on those topics they\u2019ve been doing for years\nwithout ever reconsidering. This website will also not tell you which tools to use, but\ninstead offer suggestions for multiple options, when possible explaining the differences\nin approach and use-case.<\/p><p>This is a living document and will continue to be updated with more helpful information\nand examples as they become available.<\/p><h2 id=\"translations\">Translations<\/h2><p><em>PHP: The Right Way<\/em> is translated into many different languages:<\/p><h2 id=\"book\">Book<\/h2><p>The most recent version of <em>PHP: The Right Way<\/em> is also available in PDF, EPUB and MOBI formats. <a href=\"https:\/\/leanpub.com\/phptherightway\">Go to Leanpub<\/a><\/p><h2 id=\"how-to-contribute\">How to Contribute<\/h2><p>Help make this website the best resource for new PHP programmers! <a href=\"https:\/\/github.com\/codeguy\/php-the-right-way\/tree\/gh-pages\">Contribute on GitHub<\/a><\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"use_the_current_stable_version\"><h2 id=\"use_the_current_stable_version_title\">Use the Current Stable Version (7.2)<\/h2><p>If you are getting started with PHP, start with the current stable release of <a href=\"http:\/\/php.net\/downloads.php\">PHP 7.2<\/a>. PHP 7.x adds many <a href=\"#language_highlights\">new features<\/a> over the older 5.x versions. The engine has been largely re-written, and PHP is now even quicker than older versions.<\/p><p>You should try to upgrade to the latest stable version quickly - PHP 5.6 <a href=\"http:\/\/php.net\/supported-versions.php\">will not receive security updates beyond 2018<\/a>.  Upgrading is easy, as there are not many <a href=\"http:\/\/php.net\/manual\/migration72.incompatible.php\">backwards compatibility breaks<\/a>. If you are not sure which version a function or feature is in, you can check the PHP documentation on the <a href=\"http:\/\/php.net\/manual\/\">php.net<\/a> website.<\/p><\/section><section class=\"chapter\" id=\"builtin_web_server\"><h2 id=\"builtin_web_server_title\">Built-in web server<\/h2><p>With PHP 5.4 or newer, you can start learning PHP without installing and configuring a full-fledged web server.\nTo start the server, run the following command from your terminal in your project\u2019s web root:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&gt; php -S localhost:8000<\/pre><\/figure><ul><li><a href=\"http:\/\/php.net\/features.commandline.webserver\">Learn about the built-in, command line web server<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"mac_setup\"><h2 id=\"mac_setup_title\">Mac Setup<\/h2><p>macOS comes prepackaged with PHP but it is normally a little behind the latest stable release. There are multiple ways to install the latest PHP version on macOS.<\/p><h3 id=\"install-php-via-homebrew\">Install PHP via Homebrew<\/h3><p><a href=\"https:\/\/brew.sh\/\">Homebrew<\/a> is a package manager for macOS that helps you easily install PHP and various extensions. The Homebrew core repository provides \u201cformulae\u201d for PHP 5.6, 7.0, 7.1, and 7.2. Install the latest version with this command:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>brew install php@7.2\n<\/pre><\/div><\/div><p>You can switch between Homebrew PHP versions by modifying your <code class=\"highlighter-rouge\">PATH<\/code> variable. Alternatively, you can use <a href=\"https:\/\/github.com\/philcook\/brew-php-switcher\">brew-php-switcher<\/a> to switch PHP versions automatically.<\/p><h3 id=\"install-php-via-macports\">Install PHP via Macports<\/h3><p>The <a href=\"https:\/\/www.macports.org\/install.php\">MacPorts<\/a> Project is an open-source community initiative to design an\neasy-to-use system for compiling, installing, and upgrading either\ncommand-line, X11 or Aqua based open-source software on the OS X operating\nsystem.<\/p><p>MacPorts supports pre-compiled binaries, so you don\u2019t need to recompile every\ndependency from the source tarball files, it saves your life if you don\u2019t\nhave any package installed on your system.<\/p><p>At this point, you can install <code class=\"highlighter-rouge\">php54<\/code>, <code class=\"highlighter-rouge\">php55<\/code>, <code class=\"highlighter-rouge\">php56<\/code>, <code class=\"highlighter-rouge\">php70<\/code> or <code class=\"highlighter-rouge\">php71<\/code> using the <code class=\"highlighter-rouge\">port install<\/code> command, for example:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>sudo port install php56\nsudo port install php71\n<\/pre><\/div><\/div><p>And you can run <code class=\"highlighter-rouge\">select<\/code> command to switch your active PHP:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>sudo port select --set php php71\n<\/pre><\/div><\/div><h3 id=\"install-php-via-phpbrew\">Install PHP via phpbrew<\/h3><p><a href=\"https:\/\/github.com\/phpbrew\/phpbrew\">phpbrew<\/a> is a tool for installing and managing multiple PHP versions. This can be really useful if two different\napplications\/projects require different versions of PHP, and you are not using virtual machines.<\/p><h3 id=\"install-php-via-liips-binary-installer\">Install PHP via Liip\u2019s binary installer<\/h3><p>Another popular option is <a href=\"https:\/\/php-osx.liip.ch\/\">php-osx.liip.ch<\/a> which provides one liner installation methods for versions 5.3 through 7.3.\nIt doesn\u2019t overwrite the PHP binaries installed by Apple, but installs everything in a separate location (\/usr\/local\/php5).<\/p><h3 id=\"compile-from-source\">Compile from Source<\/h3><p>Another option that gives you control over the version of PHP you install, is to <a href=\"https:\/\/secure.php.net\/install.macosx.compile\">compile it yourself<\/a>.\nIn that case be sure to have installed either <a href=\"https:\/\/github.com\/kennethreitz\/osx-gcc-installer\">Xcode<\/a> or Apple\u2019s substitute\n<a href=\"https:\/\/developer.apple.com\/downloads\">\u201cCommand Line Tools for XCode\u201d<\/a> downloadable from Apple\u2019s Mac Developer Center.<\/p><h3 id=\"all-in-one-installers\">All-in-One Installers<\/h3><p>The solutions listed above mainly handle PHP itself, and do not supply things like Apache, Nginx or a SQL server.\n\u201cAll-in-one\u201d solutions such as <a href=\"https:\/\/www.mamp.info\/en\/downloads\/\">MAMP<\/a> and <a href=\"https:\/\/www.apachefriends.org\/index.html\">XAMPP<\/a> will install these other bits of software for\nyou and tie them all together, but ease of setup comes with a trade-off of flexibility.<\/p><\/section><section class=\"chapter\" id=\"windows_setup\"><h2 id=\"windows_setup_title\">Windows Setup<\/h2><p>You can download the binaries from <a href=\"http:\/\/windows.php.net\/download\/\">windows.php.net\/download<\/a>. After the extraction of PHP, it is recommended to set the <a href=\"http:\/\/www.windows-commandline.com\/set-path-command-line\/\">PATH<\/a> to the root of your PHP folder (where php.exe is located) so you can execute PHP from anywhere.<\/p><p>For learning and local development, you can use the built in webserver with PHP 5.4+ so you don\u2019t need to worry about\nconfiguring it. If you would like an \u201call-in-one\u201d which includes a full-blown webserver and MySQL too then tools such\nas the <a href=\"https:\/\/www.microsoft.com\/web\/downloads\/platform.aspx\">Web Platform Installer<\/a>, <a href=\"http:\/\/www.apachefriends.org\/en\/xampp.html\">XAMPP<\/a>, <a href=\"http:\/\/www.easyphp.org\/\">EasyPHP<\/a>, <a href=\"http:\/\/open-server.ru\/\">OpenServer<\/a> and <a href=\"http:\/\/www.wampserver.com\/en\/\">WAMP<\/a> will\nhelp get a Windows development environment up and running fast. That said, these tools will be a little different from\nproduction so be careful of environment differences if you are working on Windows and deploying to Linux.<\/p><p>If you need to run your production system on Windows, then IIS7 will give you the most stable and best performance. You\ncan use <a href=\"http:\/\/phpmanager.codeplex.com\/\">phpmanager<\/a> (a GUI plugin for IIS7) to make configuring and managing PHP simple. IIS7 comes with\nFastCGI built in and ready to go, you just need to configure PHP as a handler. For support and additional resources\nthere is a <a href=\"http:\/\/php.iis.net\/\">dedicated area on iis.net<\/a> for PHP.<\/p><p>Generally running your application on different environment in development and production can lead to strange bugs popping up when you go\nlive. If you are developing on Windows and deploying to Linux (or anything non-Windows) then you should consider using a <a href=\"https:\/\/phptherightway.com\/#virtualization_title\">Virtual Machine<\/a>.<\/p><p>Chris Tankersley has a very helpful blog post on what tools he uses to do <a href=\"http:\/\/ctankersley.com\/2016\/11\/13\/developing-on-windows-2016\/\">PHP development using Windows<\/a>.<\/p><\/section><section class=\"chapter\" id=\"common_directory_structure\"><h2 id=\"common_directory_structure_title\">Common Directory structure<\/h2><p>A common question among those starting out with writing programs for the web is, \u201cwhere do I put my stuff?\u201d Over the years, this answer has consistently been \u201cwhere the <code class=\"highlighter-rouge\">DocumentRoot<\/code> is.\u201d Although this answer is not complete, it\u2019s a great place to start.<\/p><p>For security reasons, configuration files should not be accessible by a site\u2019s visitors; therefore, public scripts are kept in a public directory and private configurations and data are kept outside of that directory.<\/p><p>For each team, CMS, or framework one works in, a standard directory structure is used by each of those entities. However, if one is starting a project alone, knowing which filesystem structure to use can be daunting.<\/p><p><a href=\"https:\/\/twitter.com\/pmjones\">Paul M. Jones<\/a> has done some fantastic research into common practices of tens of thousands of github projects in the realm of PHP. He has compiled a standard file and directory structure, the <a href=\"https:\/\/github.com\/php-pds\/skeleton\">Standard PHP Package Skeleton<\/a>, based on this research. In this directory structure, <code class=\"highlighter-rouge\">DocumentRoot<\/code> should point to <code class=\"highlighter-rouge\">public\/<\/code>, unit tests should be in the <code class=\"highlighter-rouge\">tests\/<\/code> directory, and third party libraries, as installed by <a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">composer<\/a>, belong in the <code class=\"highlighter-rouge\">vendor\/<\/code> directory. For other files and directories, abiding by the <a href=\"https:\/\/github.com\/php-pds\/skeleton\">Standard PHP Package Skeleton<\/a> will make the most sense to contributors of a project.<\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"code_style_guide\"><p>The PHP community is large and diverse, composed of innumerable libraries, frameworks, and components. It is common for\nPHP developers to choose several of these and combine them into a single project. It is important that PHP code adhere\n(as close as possible) to a common code style to make it easy for developers to mix and match various libraries for\ntheir projects.<\/p><p>The <a href=\"https:\/\/www.php-fig.org\/\">Framework Interop Group<\/a> has proposed and approved a series of style recommendations. Not all of them related\nto code-style, but those that do are <a href=\"https:\/\/www.php-fig.org\/psr\/psr-1\/\">PSR-1<\/a>, <a href=\"https:\/\/www.php-fig.org\/psr\/psr-2\/\">PSR-2<\/a> and <a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">PSR-4<\/a>. These\nrecommendations are merely a set of rules that many projects like Drupal, Zend, Symfony, Laravel, CakePHP, phpBB, AWS SDK,\nFuelPHP, Lithium, etc are adopting. You can use them for your own projects, or continue to use your own\npersonal style.<\/p><p>Ideally, you should write PHP code that adheres to a known standard. This could be any combination of PSRs, or one\nof the coding standards made by PEAR or Zend. This means other developers can easily read and work with your code, and\napplications that implement the components can have consistency even when working with lots of third-party code.<\/p><ul><li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-1\/\">Read about PSR-1<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-2\/\">Read about PSR-2<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">Read about PSR-4<\/a><\/li>\n  <li><a href=\"https:\/\/pear.php.net\/manual\/en\/standards.php\">Read about PEAR Coding Standards<\/a><\/li>\n  <li><a href=\"https:\/\/symfony.com\/doc\/current\/contributing\/code\/standards.html\">Read about Symfony Coding Standards<\/a><\/li>\n<\/ul><p>You can use <a href=\"https:\/\/pear.php.net\/package\/PHP_CodeSniffer\/\">PHP_CodeSniffer<\/a> to check code against any one of these recommendations, and plugins for text\neditors like <a href=\"https:\/\/github.com\/benmatselby\/sublime-phpcs\">Sublime Text<\/a> to be given real-time feedback.<\/p><p>You can fix the code layout automatically by using one of the following tools:<\/p><ul><li>One is the <a href=\"https:\/\/cs.sensiolabs.org\/\">PHP Coding Standards Fixer<\/a> which has a very well tested codebase.<\/li>\n  <li>Also, the <a href=\"https:\/\/github.com\/squizlabs\/PHP_CodeSniffer\/wiki\/Fixing-Errors-Automatically\">PHP Code Beautifier and Fixer<\/a> tool which is included with PHP_CodeSniffer can be used to adjust your code accordingly.<\/li>\n<\/ul><p>And you can run phpcs manually from shell:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>phpcs -sw --standard=PSR2 file.php\n<\/pre><\/div><\/div><p>It will show errors and describe how to fix them.\nIt can also be helpful to include this command in a git hook.\nThat way, branches which contain violations against the chosen standard cannot enter the repository until those\nviolations have been fixed.<\/p><p>If you have PHP_CodeSniffer, then you can fix the code layout problems reported by it, automatically, with the\n<a href=\"https:\/\/github.com\/squizlabs\/PHP_CodeSniffer\/wiki\/Fixing-Errors-Automatically\">PHP Code Beautifier and Fixer<\/a>.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>phpcbf -w --standard=PSR2 file.php\n<\/pre><\/div><\/div><p>Another option is to use the <a href=\"https:\/\/cs.sensiolabs.org\/\">PHP Coding Standards Fixer<\/a>.\nIt will show which kind of errors the code structure had before it fixed them.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>php-cs-fixer fix -v --level=psr2 file.php\n<\/pre><\/div><\/div><p>English is preferred for all symbol names and code infrastructure. Comments may be written in any language easily\nreadable by all current and future parties who may be working on the codebase.<\/p><p>Finally, a good supplementary resource for writing clean PHP code is <a href=\"https:\/\/github.com\/jupeter\/clean-code-php\">Clean Code PHP<\/a>.<\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"programming_paradigms\"><h2 id=\"programming_paradigms_title\">Programming Paradigms<\/h2><p>PHP is a flexible, dynamic language that supports a variety of programming techniques. It has evolved dramatically over\nthe years, notably adding a solid object-oriented model in PHP 5.0 (2004), anonymous functions and namespaces in\nPHP 5.3 (2009), and traits in PHP 5.4 (2012).<\/p><h3 id=\"object-oriented-programming\">Object-oriented Programming<\/h3><p>PHP has a very complete set of object-oriented programming features including support for classes, abstract classes,\ninterfaces, inheritance, constructors, cloning, exceptions, and more.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.oop5\">Read about Object-oriented PHP<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/language.oop5.traits\">Read about Traits<\/a><\/li>\n<\/ul><p>PHP supports first-class functions, meaning that a function can be assigned to a variable. Both user-defined and\nbuilt-in functions can be referenced by a variable and invoked dynamically. Functions can be passed as arguments to\nother functions (a feature called <em>Higher-order Functions<\/em>) and functions can return other functions.<\/p><p>Recursion, a feature that allows a function to call itself, is supported by the language, but most PHP code\nis focused on iteration.<\/p><p>New anonymous functions (with support for closures) are present since PHP 5.3 (2009).<\/p><p>PHP 5.4 added the ability to bind closures to an object\u2019s scope and also improved support for callables such that they\ncan be used interchangeably with anonymous functions in almost all cases.<\/p><ul><li>Continue reading on <a href=\"https:\/\/phptherightway.com\/pages\/Functional-Programming.html\">Functional Programming in PHP<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/functions.anonymous\">Read about Anonymous Functions<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/class.closure\">Read about the Closure class<\/a><\/li>\n  <li><a href=\"https:\/\/wiki.php.net\/rfc\/closures\">More details in the Closures RFC<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/language.types.callable\">Read about Callables<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.call-user-func-array\">Read about dynamically invoking functions with <code class=\"highlighter-rouge\">call_user_func_array()<\/code><\/a><\/li>\n<\/ul><p>PHP supports various forms of meta-programming through mechanisms like the Reflection API and Magic Methods. There are\nmany Magic Methods available like <code class=\"highlighter-rouge\">__get()<\/code>, <code class=\"highlighter-rouge\">__set()<\/code>, <code class=\"highlighter-rouge\">__clone()<\/code>, <code class=\"highlighter-rouge\">__toString()<\/code>, <code class=\"highlighter-rouge\">__invoke()<\/code>, etc. that allow\ndevelopers to hook into class behavior. Ruby developers often say that PHP is lacking <code class=\"highlighter-rouge\">method_missing<\/code>, but it is\navailable as <code class=\"highlighter-rouge\">__call()<\/code> and <code class=\"highlighter-rouge\">__callStatic()<\/code>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.oop5.magic\">Read about Magic Methods<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/intro.reflection\">Read about Reflection<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/language.oop5.overloading\">Read about Overloading<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"namespaces\"><h2 id=\"namespaces_title\">Namespaces<\/h2><p>As mentioned above, the PHP community has a lot of developers creating lots of code. This means that one library\u2019s PHP\ncode might use the same class name as another. When both libraries are used in the same namespace, they collide\nand cause trouble.<\/p><p><em>Namespaces<\/em> solve this problem. As described in the PHP reference manual, namespaces may be compared to operating\nsystem directories that <em>namespace<\/em> files; two files with the same name may co-exist in separate directories. Likewise,\ntwo PHP classes with the same name may co-exist in separate PHP namespaces. It\u2019s as simple as that.<\/p><p>It is important for you to namespace your code so that it may be used by other developers without fear of colliding\nwith other libraries.<\/p><p>One recommended way to use namespaces is outlined in <a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">PSR-4<\/a>, which aims to provide a standard file, class and\nnamespace convention to allow plug-and-play code.<\/p><p>In October 2014 the PHP-FIG deprecated the previous autoloading standard: <a href=\"https:\/\/www.php-fig.org\/psr\/psr-0\/\">PSR-0<\/a>. Both PSR-0 and PSR-4 are still perfectly usable.  The latter requires PHP 5.3, so many PHP 5.2-only projects implement PSR-0.<\/p><p>If you\u2019re going to use an autoloader standard for a new application or package, look into PSR-4.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.namespaces\">Read about Namespaces<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-0\/\">Read about PSR-0<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">Read about PSR-4<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"standard_php_library\"><h2 id=\"standard_php_library_title\">Standard PHP Library<\/h2><p>The Standard PHP Library (SPL) is packaged with PHP and provides a collection of classes and interfaces. It is made up\nprimarily of commonly needed datastructure classes (stack, queue, heap, and so on), and iterators which can traverse\nover these datastructures or your own classes which implement SPL interfaces.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.spl\">Read about the SPL<\/a><\/li>\n  <li><a href=\"https:\/\/www.lynda.com\/PHP-tutorials\/Up-Running-Standard-PHP-Library\/175038-2.html\">SPL video course on Lynda.com(Paid)<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"command_line_interface\"><h2 id=\"command_line_interface_title\">Command Line Interface<\/h2><p>PHP was created to write web applications, but is also useful for scripting command line interface (CLI) programs.\nCommand line PHP programs can help automate common tasks like testing, deployment, and application administration.<\/p><p>CLI PHP programs are powerful because you can use your app\u2019s code directly without having to create and secure a web\nGUI for it. Just be sure <strong>not<\/strong> to put your CLI PHP scripts in your public web root!<\/p><p>Try running PHP from your command line:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&gt; php -i<\/pre><\/figure><p>The <code class=\"highlighter-rouge\">-i<\/code> option will print your PHP configuration just like the <a href=\"https:\/\/secure.php.net\/function.phpinfo\"><code class=\"highlighter-rouge\">phpinfo()<\/code><\/a> function.<\/p><p>The <code class=\"highlighter-rouge\">-a<\/code> option provides an interactive shell, similar to ruby\u2019s IRB or python\u2019s interactive shell. There are a number\nof other useful <a href=\"https:\/\/secure.php.net\/features.commandline.options\">command line options<\/a>, too.<\/p><p>Let\u2019s write a simple \u201cHello, $name\u201d CLI program. To try it out, create a file named <code class=\"highlighter-rouge\">hello.php<\/code>, as below.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nif ($argc !== 2) {\n    echo \"Usage: php hello.php [name].\\n\";\n    exit(1);\n}\n$name = $argv[1];\necho \"Hello, $name\\n\";<\/pre><\/figure><p>PHP sets up two special variables based on the arguments your script is run with. <a href=\"https:\/\/secure.php.net\/reserved.variables.argc\"><code class=\"highlighter-rouge\">$argc<\/code><\/a> is an integer\nvariable containing the argument <em>count<\/em> and <a href=\"https:\/\/secure.php.net\/reserved.variables.argv\"><code class=\"highlighter-rouge\">$argv<\/code><\/a> is an array variable containing each argument\u2019s <em>value<\/em>.\nThe first argument is always the name of your PHP script file, in this case <code class=\"highlighter-rouge\">hello.php<\/code>.<\/p><p>The <code class=\"highlighter-rouge\">exit()<\/code> expression is used with a non-zero number to let the shell know that the command failed. Commonly used\nexit codes can be found <a href=\"https:\/\/www.gsp.com\/cgi-bin\/man.cgi?section=3&amp;topic=sysexits\">here<\/a>.<\/p><p>To run our script, above, from the command line:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&gt; php hello.php\nUsage: php hello.php [name]\n&gt; php hello.php world\nHello, world<\/pre><\/figure><ul><li><a href=\"https:\/\/secure.php.net\/features.commandline.options\">Learn about running PHP from the command line<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"xdebug\"><h2 id=\"xdebug_title\">Xdebug<\/h2><p>One of the most useful tools in software development is a proper debugger. It allows you to trace the execution of your\ncode and monitor the contents of the stack. Xdebug, PHP\u2019s debugger, can be utilized by various IDEs to provide\nBreakpoints and stack inspection. It can also allow tools like PHPUnit and KCacheGrind to perform code coverage\nanalysis and code profiling.<\/p><p>If you find yourself in a bind, willing to resort to <code class=\"highlighter-rouge\">var_dump()<\/code>\/<code class=\"highlighter-rouge\">print_r()<\/code>, and you still can\u2019t find the solution -\nmaybe you need to use the debugger.<\/p><p><a href=\"https:\/\/xdebug.org\/docs\/install\">Installing Xdebug<\/a> can be tricky, but one of its most important features is \u201cRemote Debugging\u201d - if\nyou develop code locally and then test it inside a VM or on another server, Remote Debugging is the feature that you\nwill want to enable right away.<\/p><p>Traditionally, you will modify your Apache VHost or .htaccess file with these values:<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">php_value xdebug.remote_host 192.168.?.?\nphp_value xdebug.remote_port 9000<\/pre><\/figure><p>The \u201cremote host\u201d and \u201cremote port\u201d will correspond to your local computer and the port that you configure your IDE to\nlisten on. Then it\u2019s just a matter of putting your IDE into \u201clisten for connections\u201d mode, and loading the URL:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>http:\/\/your-website.example.com\/index.php?XDEBUG_SESSION_START=1\n<\/pre><\/div><\/div><p>Your IDE will now intercept the current state as the script executes, allowing you to set breakpoints and probe the\nvalues in memory.<\/p><p>Graphical debuggers make it very easy to step through code, inspect variables, and eval code against the live runtime.\nMany IDE\u2019s have built-in or plugin-based support for graphical debugging with Xdebug. MacGDBp is a free, open-source,\nstand-alone Xdebug GUI for Mac.<\/p><ul><li><a href=\"https:\/\/xdebug.org\/docs\/\">Learn more about Xdebug<\/a><\/li>\n  <li><a href=\"https:\/\/www.bluestatic.org\/software\/macgdbp\/\">Learn more about MacGDBp<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"dependency_management\"><p>There are a ton of PHP libraries, frameworks, and components to choose from. Your project will likely use \nseveral of them \u2014 these are project dependencies. Until recently, PHP did not have a good way to manage\nthese project dependencies. Even if you managed them manually, you still had to worry about autoloaders.\nThat is no longer an issue.<\/p><p>Currently there are two major package management systems for PHP - <a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">Composer<\/a> and <a href=\"https:\/\/phptherightway.com\/#pear\">PEAR<\/a>. Composer is currently\nthe most popular package manager for PHP, however for a long time PEAR was the primary package manager in use.\nKnowing PEAR\u2019s history is a good idea, since you may still find references to it even if you never use it.<\/p><\/section><section class=\"chapter\" id=\"composer_and_packagist\"><h2 id=\"composer_and_packagist_title\">Composer and Packagist<\/h2><p>Composer is the recommended dependency manager for PHP. List your project\u2019s dependencies in a <code class=\"highlighter-rouge\">composer.json<\/code> file and,\nwith a few simple commands, Composer will automatically download your project\u2019s dependencies and setup autoloading for\nyou. Composer is analogous to NPM in the node.js world, or Bundler in the Ruby world.<\/p><p>There is a plethora of PHP libraries that are compatible with Composer and ready to be used in your project. These\n\u201cpackages\u201d are listed on <a href=\"https:\/\/packagist.org\/\">Packagist<\/a>, the official repository for Composer-compatible PHP libraries.<\/p><h3 id=\"how-to-install-composer\">How to Install Composer<\/h3><p>The safest way to download composer is by <a href=\"https:\/\/getcomposer.org\/download\/\">following the official instructions<\/a>.\nThis will verify the installer is not corrupt or tampered with.\nThe installer installs a <code class=\"highlighter-rouge\">composer.phar<\/code> binary in your <em>current working directory<\/em>.<\/p><p>We recommend installing Composer <em>globally<\/em> (e.g. a single copy in <code class=\"highlighter-rouge\">\/usr\/local\/bin<\/code>). To do so, run this command next:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">mv composer.phar \/usr\/local\/bin\/composer<\/pre><\/figure><p><strong>Note:<\/strong> If the above fails due to permissions, prefix with <code class=\"highlighter-rouge\">sudo<\/code>.<\/p><p>To run a locally installed Composer you\u2019d use <code class=\"highlighter-rouge\">php composer.phar<\/code>, globally it\u2019s simply <code class=\"highlighter-rouge\">composer<\/code>.<\/p><h4 id=\"installing-on-windows\">Installing on Windows<\/h4><p>For Windows users the easiest way to get up and running is to use the <a href=\"https:\/\/getcomposer.org\/Composer-Setup.exe\">ComposerSetup<\/a> installer, which\nperforms a global install and sets up your <code class=\"highlighter-rouge\">$PATH<\/code> so that you can just call <code class=\"highlighter-rouge\">composer<\/code> from any\ndirectory in your command line.<\/p><h3 id=\"how-to-define-and-install-dependencies\">How to Define and Install Dependencies<\/h3><p>Composer keeps track of your project\u2019s dependencies in a file called <code class=\"highlighter-rouge\">composer.json<\/code>. You can manage it\nby hand if you like, or use Composer itself. The <code class=\"highlighter-rouge\">composer require<\/code> command adds a project dependency\nand if you don\u2019t have a <code class=\"highlighter-rouge\">composer.json<\/code> file, one will be created. Here\u2019s an example that adds <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a>\nas a dependency of your project.<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">composer require twig\/twig:^2.0<\/pre><\/figure><p>Alternatively, the <code class=\"highlighter-rouge\">composer init<\/code> command will guide you through creating a full <code class=\"highlighter-rouge\">composer.json<\/code> file\nfor your project. Either way, once you\u2019ve created your <code class=\"highlighter-rouge\">composer.json<\/code> file you can tell Composer to\ndownload and install your dependencies into the <code class=\"highlighter-rouge\">vendor\/<\/code> directory. This also applies to projects\nyou\u2019ve downloaded that already provide a <code class=\"highlighter-rouge\">composer.json<\/code> file:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">composer install<\/pre><\/figure><p>Next, add this line to your application\u2019s primary PHP file; this will tell PHP to use Composer\u2019s\nautoloader for your project dependencies.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nrequire 'vendor\/autoload.php';<\/pre><\/figure><p>Now you can use your project dependencies, and they\u2019ll be autoloaded on demand.<\/p><h3 id=\"updating-your-dependencies\">Updating your dependencies<\/h3><p>Composer creates a file called <code class=\"highlighter-rouge\">composer.lock<\/code> which stores the exact version of each package it\ndownloaded when you first ran <code class=\"highlighter-rouge\">composer install<\/code>. If you share your project with others,\nensure the <code class=\"highlighter-rouge\">composer.lock<\/code> file is included, so that when they run <code class=\"highlighter-rouge\">composer install<\/code> they\u2019ll\nget the same versions as you.  To update your dependencies, run <code class=\"highlighter-rouge\">composer update<\/code>. Don\u2019t use\n<code class=\"highlighter-rouge\">composer update<\/code> when deploying, only <code class=\"highlighter-rouge\">composer install<\/code>, otherwise you may end up with different\npackage versions on production.<\/p><p>This is most useful when you define your version requirements flexibly. For instance, a version\nrequirement of <code class=\"highlighter-rouge\">~1.8<\/code> means \u201canything newer than <code class=\"highlighter-rouge\">1.8.0<\/code>, but less than <code class=\"highlighter-rouge\">2.0.x-dev<\/code>\u201d. You can also use\nthe <code class=\"highlighter-rouge\">*<\/code> wildcard as in <code class=\"highlighter-rouge\">1.8.*<\/code>. Now Composer\u2019s <code class=\"highlighter-rouge\">composer update<\/code> command will upgrade all your\ndependencies to the newest version that fits the restrictions you define.<\/p><p>To receive notifications about new version releases you can sign up for <a href=\"https:\/\/www.versioneye.com\/\">VersionEye<\/a>, a web service\nthat can monitor your GitHub and BitBucket accounts for <code class=\"highlighter-rouge\">composer.json<\/code> files and send emails with new\npackage releases.<\/p><h3 id=\"checking-your-dependencies-for-security-issues\">Checking your dependencies for security issues<\/h3><p>The <a href=\"https:\/\/security.sensiolabs.org\/\">Security Advisories Checker<\/a> is a web service and a command-line tool, both will examine your <code class=\"highlighter-rouge\">composer.lock<\/code>\nfile and tell you if you need to update any of your dependencies.<\/p><h3 id=\"handling-global-dependencies-with-composer\">Handling global dependencies with Composer<\/h3><p>Composer can also handle global dependencies and their binaries. Usage is straight-forward, all you need\nto do is prefix your command with <code class=\"highlighter-rouge\">global<\/code>. If for example you wanted to install PHPUnit and have it\navailable globally, you\u2019d run the following command:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">composer global require phpunit\/phpunit<\/pre><\/figure><p>This will create a <code class=\"highlighter-rouge\">~\/.composer<\/code> folder where your global dependencies reside. To have the installed\npackages\u2019 binaries available everywhere, you\u2019d then add the <code class=\"highlighter-rouge\">~\/.composer\/vendor\/bin<\/code> folder to your\n<code class=\"highlighter-rouge\">$PATH<\/code> variable.<\/p><ul><li><a href=\"https:\/\/getcomposer.org\/doc\/00-intro.md\">Learn about Composer<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"pear\"><h2 id=\"pear_title\">PEAR<\/h2><p>A veteran package manager that some PHP developers enjoy is <a href=\"https:\/\/pear.php.net\/\">PEAR<\/a>. It behaves similarly to Composer,\nbut has some notable differences.<\/p><p>PEAR requires each package to have a specific structure, which means that the author of the package must prepare it for\nusage with PEAR. Using a project which was not prepared to work with PEAR is not possible.<\/p><p>PEAR installs packages globally, which means after installing them once they are available to all projects on that\nserver. This can be good if many projects rely on the same package with the same version but might lead to problems if\nversion conflicts between two projects arise.<\/p><h3 id=\"how-to-install-pear\">How to install PEAR<\/h3><p>You can install PEAR by downloading the <code class=\"highlighter-rouge\">.phar<\/code> installer and executing it. The PEAR documentation has\ndetailed <a href=\"https:\/\/pear.php.net\/manual\/installation.getting.php\">install instructions<\/a> for every operating system.<\/p><p>If you are using Linux, you can also have a look at your distribution package manager. Debian and Ubuntu,\nfor example, have an apt <code class=\"highlighter-rouge\">php-pear<\/code> package.<\/p><h3 id=\"how-to-install-a-package\">How to install a package<\/h3><p>If the package is listed on the <a href=\"https:\/\/pear.php.net\/packages.php\">PEAR packages list<\/a>, you can install it by specifying the official name:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">pear install foo<\/pre><\/figure><p>If the package is hosted on another channel, you need to <code class=\"highlighter-rouge\">discover<\/code> the channel first and also specify it when\ninstalling. See the <a href=\"https:\/\/pear.php.net\/manual\/guide.users.commandline.channels.php\">Using channel docs<\/a> for more information on this topic.<\/p><ul><li><a href=\"https:\/\/pear.php.net\/\">Learn about PEAR<\/a><\/li>\n<\/ul><h3 id=\"handling-pear-dependencies-with-composer\">Handling PEAR dependencies with Composer<\/h3><p>If you are already using <a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">Composer<\/a> and you would like to install some PEAR code too, you can use Composer to\nhandle your PEAR dependencies. This example will install code from <code class=\"highlighter-rouge\">pear2.php.net<\/code>:<\/p><figure class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n    \"repositories\": [\n        {\n            \"type\": \"pear\",\n            \"url\": \"https:\/\/pear2.php.net\"\n        }\n    ],\n    \"require\": {\n        \"pear-pear2\/PEAR2_Text_Markdown\": \"*\",\n        \"pear-pear2\/PEAR2_HTTP_Request\": \"*\"\n    }\n}<\/pre><\/figure><p>The first section <code class=\"highlighter-rouge\">\"repositories\"<\/code> will be used to let Composer know it should \u201cinitialize\u201d (or \u201cdiscover\u201d in PEAR\nterminology) the pear repo. Then the <code class=\"highlighter-rouge\">require<\/code> section will prefix the package name like this:<\/p><blockquote>\n  <p>pear-channel\/Package<\/p>\n<\/blockquote><p>The \u201cpear\u201d prefix is hardcoded to avoid any conflicts, as a pear channel could be the same as another packages vendor\nname for example, then the channel short name (or full URL) can be used to reference which channel the package is in.<\/p><p>When this code is installed it will be available in your vendor directory and automatically available through the\nComposer autoloader:<\/p><blockquote>\n  <p>vendor\/pear-pear2.php.net\/PEAR2_HTTP_Request\/pear2\/HTTP\/Request.php<\/p>\n<\/blockquote><p>To use this PEAR package simply reference it like so:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$request = new pear2\\HTTP\\Request();<\/pre><\/figure><ul><li><a href=\"https:\/\/getcomposer.org\/doc\/05-repositories.md#pear\">Learn more about using PEAR with Composer<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"the_basics\"><h2 id=\"the_basics_title\">The Basics<\/h2><p>PHP is a vast language that allows coders of all levels the ability to produce code not only quickly, but efficiently.\nHowever, while advancing through the language, we often forget the basics that we first learnt (or overlooked) in favor\nof short cuts and\/or bad habits. To help combat this common issue, this section is aimed at reminding coders of the\nbasic coding practices within PHP.<\/p><ul><li>Continue reading on <a href=\"https:\/\/phptherightway.com\/pages\/The-Basics.html\">The Basics<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"date_and_time\"><p>PHP has a class named DateTime to help you when reading, writing, comparing or calculating with date and time. There\nare many date and time related functions in PHP besides DateTime, but it provides nice object-oriented interface to\nmost common uses. It can handle time zones, but that is outside this short introduction.<\/p><p>To start working with DateTime, convert raw date and time string to an object with <code class=\"highlighter-rouge\">createFromFormat()<\/code> factory method\nor do <code class=\"highlighter-rouge\">new DateTime<\/code> to get the current date and time. Use <code class=\"highlighter-rouge\">format()<\/code> method to convert DateTime back to a string for\noutput.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$raw = '22. 11. 1968';\n$start = DateTime::createFromFormat('d. m. Y', $raw);\necho 'Start date: ' . $start-&gt;format('Y-m-d') . \"\\n\";<\/pre><\/figure><p>Calculating with DateTime is possible with the DateInterval class. DateTime has methods like <code class=\"highlighter-rouge\">add()<\/code> and <code class=\"highlighter-rouge\">sub()<\/code> that\ntake a DateInterval as an argument. Do not write code that expect same number of seconds in every day, both daylight\nsaving and timezone alterations will break that assumption. Use date intervals instead. To calculate date difference\nuse the <code class=\"highlighter-rouge\">diff()<\/code> method. It will return new DateInterval, which is super easy to display.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ create a copy of $start and add one month and 6 days\n$end = clone $start;\n$end-&gt;add(new DateInterval('P1M6D'));\n$diff = $end-&gt;diff($start);\necho 'Difference: ' . $diff-&gt;format('%m month, %d days (total: %a days)') . \"\\n\";\n\/\/ Difference: 1 month, 6 days (total: 37 days)<\/pre><\/figure><p>On DateTime objects you can use standard comparison:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nif ($start &lt; $end) {\n    echo \"Start is before the end!\\n\";\n}<\/pre><\/figure><p>One last example to demonstrate the DatePeriod class. It is used to iterate over recurring events. It can take two\nDateTime objects, start and end, and the interval for which it will return all events in between.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ output all thursdays between $start and $end\n$periodInterval = DateInterval::createFromDateString('first thursday');\n$periodIterator = new DatePeriod($start, $periodInterval, $end, DatePeriod::EXCLUDE_START_DATE);\nforeach ($periodIterator as $date) {\n    \/\/ output each date in the period\n    echo $date-&gt;format('Y-m-d') . ' ';\n}<\/pre><\/figure><p>A popular PHP API extension is <a href=\"https:\/\/carbon.nesbot.com\/\">Carbon<\/a>. It inherits everything in the DateTime class, so involves minimal code alterations, but extra features include Localization support, further ways to add, subtract and format a DateTime object, plus a means to test your code by simulating a date and time of your choosing.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.datetime\">Read about DateTime<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.date\">Read about date formatting<\/a> (accepted date format string options)<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"design_patterns\"><p>When you are building your application it is helpful to use common patterns in your code and common patterns for the\noverall structure of your project. Using common patterns is helpful because it makes it much easier to manage your code\nand lets other developers quickly understand how everything fits together.<\/p><p>If you use a framework then most of the higher level code and project structure will be based on that framework, so a\nlot of the pattern decisions are made for you. But it is still up to you to pick out the best patterns to follow in the\ncode you build on top of the framework. If, on the other hand, you are not using a framework to build your application\nthen you have to find the patterns that best suit the type and size of application that you\u2019re building.<\/p><p>You can learn more about PHP design patterns and see working examples at:<\/p><p><a href=\"https:\/\/designpatternsphp.readthedocs.io\/\">https:\/\/designpatternsphp.readthedocs.io\/<\/a><\/p><\/section><section class=\"chapter\" id=\"php_and_utf8\"><h2 id=\"php_and_utf8_title\">Working with UTF-8<\/h2><p><em>This section was originally written by <a href=\"https:\/\/alexcabal.com\/\">Alex Cabal<\/a> over at\n<a href=\"https:\/\/phpbestpractices.org\/#utf-8\">PHP Best Practices<\/a> and has been used as the basis for our own UTF-8 advice<\/em>.<\/p><h3 id=\"theres-no-one-liner-be-careful-detailed-and-consistent\">There\u2019s no one-liner. Be careful, detailed, and consistent.<\/h3><p>Right now PHP does not support Unicode at a low level. There are ways to ensure that UTF-8 strings are processed OK,\nbut it\u2019s not easy, and it requires digging in to almost all levels of the web app, from HTML to SQL to PHP. We\u2019ll aim\nfor a brief, practical summary.<\/p><h3 id=\"utf-8-at-the-php-level\">UTF-8 at the PHP level<\/h3><p>The basic string operations, like concatenating two strings and assigning strings to variables, don\u2019t need anything\nspecial for UTF-8. However, most string functions, like <code class=\"highlighter-rouge\">strpos()<\/code> and <code class=\"highlighter-rouge\">strlen()<\/code>, do need special consideration. These\nfunctions often have an <code class=\"highlighter-rouge\">mb_*<\/code> counterpart: for example, <code class=\"highlighter-rouge\">mb_strpos()<\/code> and <code class=\"highlighter-rouge\">mb_strlen()<\/code>. These <code class=\"highlighter-rouge\">mb_*<\/code> strings are made\navailable to you via the <a href=\"https:\/\/secure.php.net\/book.mbstring\">Multibyte String Extension<\/a>, and are specifically designed to operate on Unicode strings.<\/p><p>You must use the <code class=\"highlighter-rouge\">mb_*<\/code> functions whenever you operate on a Unicode string. For example, if you use <code class=\"highlighter-rouge\">substr()<\/code> on a\nUTF-8 string, there\u2019s a good chance the result will include some garbled half-characters. The correct function to use\nwould be the multibyte counterpart, <code class=\"highlighter-rouge\">mb_substr()<\/code>.<\/p><p>The hard part is remembering to use the <code class=\"highlighter-rouge\">mb_*<\/code> functions at all times. If you forget even just once, your Unicode\nstring has a chance of being garbled during further processing.<\/p><p>Not all string functions have an <code class=\"highlighter-rouge\">mb_*<\/code> counterpart. If there isn\u2019t one for what you want to do, then you might be out\nof luck.<\/p><p>You should use the <code class=\"highlighter-rouge\">mb_internal_encoding()<\/code> function at the top of every PHP script you write (or at the top of your\nglobal include script), and the <code class=\"highlighter-rouge\">mb_http_output()<\/code> function right after it if your script is outputting to a browser.\nExplicitly defining the encoding of your strings in every script will save you a lot of headaches down the road.<\/p><p>Additionally, many PHP functions that operate on strings have an optional parameter letting you specify the character\nencoding. You should always explicitly indicate UTF-8 when given the option. For example, <code class=\"highlighter-rouge\">htmlentities()<\/code> has an\noption for character encoding, and you should always specify UTF-8 if dealing with such strings. Note that as of PHP 5.4.0, UTF-8 is the default encoding for <code class=\"highlighter-rouge\">htmlentities()<\/code> and <code class=\"highlighter-rouge\">htmlspecialchars()<\/code>.<\/p><p>Finally, If you are building a distributed application and cannot be certain that the <code class=\"highlighter-rouge\">mbstring<\/code> extension will be\nenabled, then consider using the <a href=\"https:\/\/packagist.org\/packages\/patchwork\/utf8\">patchwork\/utf8<\/a> Composer package. This will use <code class=\"highlighter-rouge\">mbstring<\/code> if it is available, and\nfall back to non UTF-8 functions if not.<\/p><h3 id=\"utf-8-at-the-database-level\">UTF-8 at the Database level<\/h3><p>If your PHP script accesses MySQL, there\u2019s a chance your strings could be stored as non-UTF-8 strings in the database\neven if you follow all of the precautions above.<\/p><p>To make sure your strings go from PHP to MySQL as UTF-8, make sure your database and tables are all set to the\n<code class=\"highlighter-rouge\">utf8mb4<\/code> character set and collation, and that you use the <code class=\"highlighter-rouge\">utf8mb4<\/code> character set in the PDO connection string. See\nexample code below. This is <em>critically important<\/em>.<\/p><p>Note that you must use the <code class=\"highlighter-rouge\">utf8mb4<\/code> character set for complete UTF-8 support, not the <code class=\"highlighter-rouge\">utf8<\/code> character set! See\nFurther Reading for why.<\/p><h3 id=\"utf-8-at-the-browser-level\">UTF-8 at the browser level<\/h3><p>Use the <code class=\"highlighter-rouge\">mb_http_output()<\/code> function to ensure that your PHP script outputs UTF-8 strings to your browser.<\/p><p>The browser will then need to be told by the HTTP response that this page should be considered as UTF-8. Today, it is common to set the character set in the HTTP response header like this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nheader('Content-Type: text\/html; charset=UTF-8')<\/pre><\/figure><p>The historic approach to doing that was to include the <a href=\"http:\/\/htmlpurifier.org\/docs\/enduser-utf8.html\">charset <code class=\"highlighter-rouge\">&lt;meta&gt;<\/code> tag<\/a> in your page\u2019s <code class=\"highlighter-rouge\">&lt;head&gt;<\/code> tag.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ Tell PHP that we're using UTF-8 strings until the end of the script\nmb_internal_encoding('UTF-8');\n\/\/ Tell PHP that we'll be outputting UTF-8 to the browser\nmb_http_output('UTF-8');\n\/\/ Our UTF-8 test string\n$string = '\u00cal s\u00edla erin l\u00fb e-govaned v\u00een.';\n\/\/ Transform the string in some way with a multibyte function\n\/\/ Note how we cut the string at a non-Ascii character for demonstration purposes\n$string = mb_substr($string, 0, 15);\n\/\/ Connect to a database to store the transformed string\n\/\/ See the PDO example in this document for more information\n\/\/ Note the `charset=utf8mb4` in the Data Source Name (DSN)\n$link = new PDO(\n    'mysql:host=your-hostname;dbname=your-db;charset=utf8mb4',\n    'your-username',\n    'your-password',\n    array(\n        PDO::ATTR_ERRMODE =&gt; PDO::ERRMODE_EXCEPTION,\n        PDO::ATTR_PERSISTENT =&gt; false\n    )\n);\n\/\/ Store our transformed string as UTF-8 in our database\n\/\/ Your DB and tables are in the utf8mb4 character set and collation, right?\n$handle = $link-&gt;prepare('insert into ElvishSentences (Id, Body) values (?, ?)');\n$handle-&gt;bindValue(1, 1, PDO::PARAM_INT);\n$handle-&gt;bindValue(2, $string);\n$handle-&gt;execute();\n\/\/ Retrieve the string we just stored to prove it was stored correctly\n$handle = $link-&gt;prepare('select * from ElvishSentences where Id = ?');\n$handle-&gt;bindValue(1, 1, PDO::PARAM_INT);\n$handle-&gt;execute();\n\/\/ Store the result into an object that we'll output later in our HTML\n$result = $handle-&gt;fetchAll(\\PDO::FETCH_OBJ);\nheader('Content-Type: text\/html; charset=UTF-8');\n?&gt;&lt;!doctype html&gt;\n&lt;html&gt;\n    &lt;head&gt;\n        &lt;meta charset=\"UTF-8\"&gt;\n        &lt;title&gt;UTF-8 test page&lt;\/title&gt;\n    &lt;\/head&gt;\n    &lt;body&gt;\n        &lt;?php\n        foreach($result as $row){\n            print($row-&gt;Body);  \/\/ This should correctly output our transformed UTF-8 string to the browser\n        }\n        ?&gt;\n    &lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/figure><h3 id=\"further-reading\">Further reading<\/h3><ul><li><a href=\"https:\/\/secure.php.net\/language.operators.string\">PHP Manual: String Operations<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.strings\">PHP Manual: String Functions<\/a>\n    <ul><li><a href=\"https:\/\/secure.php.net\/function.strpos\"><code class=\"highlighter-rouge\">strpos()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.strlen\"><code class=\"highlighter-rouge\">strlen()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.substr\"><code class=\"highlighter-rouge\">substr()<\/code><\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.mbstring\">PHP Manual: Multibyte String Functions<\/a>\n    <ul><li><a href=\"https:\/\/secure.php.net\/function.mb-strpos\"><code class=\"highlighter-rouge\">mb_strpos()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-strlen\"><code class=\"highlighter-rouge\">mb_strlen()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-substr\"><code class=\"highlighter-rouge\">mb_substr()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-internal-encoding\"><code class=\"highlighter-rouge\">mb_internal_encoding()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-http-output\"><code class=\"highlighter-rouge\">mb_http_output()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.htmlentities\"><code class=\"highlighter-rouge\">htmlentities()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.htmlspecialchars\"><code class=\"highlighter-rouge\">htmlspecialchars()<\/code><\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"https:\/\/stackoverflow.com\/questions\/571694\/what-factors-make-php-unicode-incompatible\">Stack Overflow: What factors make PHP Unicode-incompatible?<\/a><\/li>\n  <li><a href=\"https:\/\/stackoverflow.com\/questions\/140728\/best-practices-in-php-and-mysql-with-international-strings\">Stack Overflow: Best practices in PHP and MySQL with international strings<\/a><\/li>\n  <li><a href=\"https:\/\/mathiasbynens.be\/notes\/mysql-utf8mb4\">How to support full Unicode in MySQL databases<\/a><\/li>\n  <li><a href=\"https:\/\/www.sitepoint.com\/bringing-unicode-to-php-with-portable-utf8\/\">Bringing Unicode to PHP with Portable UTF-8<\/a><\/li>\n  <li><a href=\"https:\/\/stackoverflow.com\/questions\/8218230\/php-domdocument-loadhtml-not-encoding-utf-8-correctly\">Stack Overflow: DOMDocument loadHTML does not encode UTF-8 correctly<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"i18n_l10n\"><h2 id=\"i18n_l10n_title\">Internationalization (i18n) and Localization (l10n)<\/h2><p><em>Disclaimer for newcomers: i18n and l10n are numeronyms, a kind of abbreviation where numbers are used to shorten\nwords - in our case, internationalization becomes i18n and localization, l10n.<\/em><\/p><p>First of all, we need to define those two similar concepts and other related things:<\/p><ul><li><strong>Internationalization<\/strong> is when you organize your code so it can be adapted to different languages or regions\nwithout refactorings. This action is usually done once - preferably, at the beginning of the project, or else you will\nprobably need some huge changes in the source!<\/li>\n  <li><strong>Localization<\/strong> happens when you adapt the interface (mainly) by translating contents, based on the i18n work done\nbefore. It usually is done every time a new language or region needs support and is updated when new interface pieces\nare added, as they need to be available in all supported languages.<\/li>\n  <li><strong>Pluralization<\/strong> defines the rules required between distinct languages to interoperate strings containing numbers and \ncounters. For instance, in English when you have only one item, it is singular, and anything different from that is \ncalled plural; plural in this language is indicated by adding an S after some words, and sometimes changes parts of it.\nIn other languages, such as Russian or Serbian, there are two plural forms in addition to the singular - you may even\nfind languages with a total of four, five or six forms, such as Slovenian, Irish or Arabic.<\/li>\n<\/ul><h2 id=\"common-ways-to-implement\">Common ways to implement<\/h2><p>The easiest way to internationalize PHP software is by using array files and using those strings in templates, such as\n<code class=\"highlighter-rouge\">&lt;h1&gt;&lt;?=$TRANS['title_about_page']?&gt;&lt;\/h1&gt;<\/code>. This way is, however, hardly recommended for serious projects, as it poses\nsome maintenance issues along the road - some might appear in the very beginning, such as pluralization. So, please,\ndon\u2019t try this if your project will contain more than a couple of pages.<\/p><p>The most classic way and often taken as reference for i18n and l10n is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gettext\">Unix tool called <code class=\"highlighter-rouge\">gettext<\/code><\/a>. It dates\nback to 1995 and is still a complete implementation for translating software. It is easy enough to get running, while\nstill sporting powerful supporting tools. It is about Gettext we will be talking here. Also, to help you not get messy\nover the command-line, we will be presenting a great GUI application that can be used to easily update your l10n source<\/p><h3 id=\"other-tools\">Other tools<\/h3><p>There are common libraries used that support Gettext and other implementations of i18n. Some of them may seem easier to\ninstall or sport additional features or i18n file formats. In this document, we focus on the tools provided with the\nPHP core, but here we list others for completion:<\/p><ul><li><a href=\"https:\/\/github.com\/auraphp\/Aura.Intl\">aura\/intl<\/a>: Provides internationalization (I18N) tools, specifically package-oriented per-locale message\ntranslation. It uses array formats for message. Does not provide a message extractor, but does provide advanced\nmessage formatting via the <code class=\"highlighter-rouge\">intl<\/code> extension (including pluralized messages).<\/li>\n  <li><a href=\"https:\/\/github.com\/oscarotero\/Gettext\">oscarotero\/Gettext<\/a>: Gettext support with an OO interface; includes improved helper functions, powerful\nextractors for several file formats (some of them not supported natively by the <code class=\"highlighter-rouge\">gettext<\/code> command), and can also export\nto other formats besides <code class=\"highlighter-rouge\">.mo\/.po<\/code> files. Can be useful if you need to integrate your translation files into other\nparts of the system, like a JavaScript interface.<\/li>\n  <li><a href=\"https:\/\/symfony.com\/doc\/current\/components\/translation.html\">symfony\/translation<\/a>: supports a lot of different formats, but recommends using verbose XLIFF\u2019s. Doesn\u2019t\ninclude helper functions nor a built-in extractor, but supports placeholders using <code class=\"highlighter-rouge\">strtr()<\/code> internally.<\/li>\n  <li><a href=\"https:\/\/docs.zendframework.com\/zend-i18n\/translation\">zend\/i18n<\/a>: supports array and INI files, or Gettext formats. Implements a caching layer to save you from\nreading the filesystem every time. It also includes view helpers, and locale-aware input filters and validators.\nHowever, it has no message extractor.<\/li>\n<\/ul><p>Other frameworks also include i18n modules, but those are not available outside of their codebases:<\/p><ul><li><a href=\"https:\/\/laravel.com\/docs\/master\/localization\">Laravel<\/a> supports basic array files, has no automatic extractor but includes a <code class=\"highlighter-rouge\">@lang<\/code> helper for template files.<\/li>\n  <li><a href=\"https:\/\/www.yiiframework.com\/doc\/guide\/2.0\/en\/tutorial-i18n\">Yii<\/a> supports array, Gettext, and database-based translation, and includes a messages extractor. It is backed by the\n<a href=\"https:\/\/secure.php.net\/manual\/intro.intl.php\"><code class=\"highlighter-rouge\">Intl<\/code><\/a> extension, available since PHP 5.3, and based on the <a href=\"http:\/\/www.icu-project.org\">ICU project<\/a>; this enables Yii to run powerful\nreplacements, like spelling out numbers, formatting dates, times, intervals, currency, and ordinals.<\/li>\n<\/ul><p>If you decide to go for one of the libraries that provide no extractors, you may want to use the gettext formats, so\nyou can use the original gettext toolchain (including Poedit) as described in the rest of the chapter.<\/p><h2 id=\"gettext\">Gettext<\/h2><h3 id=\"installation\">Installation<\/h3><p>You might need to install Gettext and the related PHP library by using your package manager, like <code class=\"highlighter-rouge\">apt-get<\/code> or <code class=\"highlighter-rouge\">yum<\/code>.\nAfter installed, enable it by adding <code class=\"highlighter-rouge\">extension=gettext.so<\/code> (Linux\/Unix) or <code class=\"highlighter-rouge\">extension=php_gettext.dll<\/code> (Windows) to\nyour <code class=\"highlighter-rouge\">php.ini<\/code>.<\/p><p>Here we will also be using <a href=\"https:\/\/poedit.net\">Poedit<\/a> to create translation files. You will probably find it in your system\u2019s package\nmanager; it is available for Unix, Mac, and Windows, and can be <a href=\"https:\/\/poedit.net\/download\">downloaded for free on their website<\/a>\nas well.<\/p><h3 id=\"structure\">Structure<\/h3><h4 id=\"types-of-files\">Types of files<\/h4><p>There are three files you usually deal with while working with gettext. The main ones are PO (Portable Object) and\nMO (Machine Object) files, the first being a list of readable \u201ctranslated objects\u201d and the second, the corresponding\nbinary to be interpreted by gettext when doing localization. There\u2019s also a POT (Template) file, which simply contains\nall existing keys from your source files, and can be used as a guide to generate and update all PO files. Those template\nfiles are not mandatory: depending on the tool you are using to do l10n, you can go just fine with only PO\/MO files.\nYou will always have one pair of PO\/MO files per language and region, but only one POT per domain.<\/p><h3 id=\"domains\">Domains<\/h3><p>There are some cases, in big projects, where you might need to separate translations when the same words convey \ndifferent meaning given a context. In those cases, you split them into different <em>domains<\/em>. They are, basically, named\ngroups of POT\/PO\/MO files, where the filename is the said <em>translation domain<\/em>. Small and medium-sized projects usually,\nfor simplicity, use only one domain; its name is arbitrary, but we will be using \u201cmain\u201d for our code samples.\nIn <a href=\"https:\/\/symfony.com\/doc\/current\/components\/translation.html\">Symfony<\/a> projects, for example, domains are used to separate the translation for validation messages.<\/p><h4 id=\"locale-code\">Locale code<\/h4><p>A locale is simply a code that identifies one version of a language. It is defined following the <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_ISO_639-1_codes\">ISO 639-1<\/a> and \n<a href=\"https:\/\/en.wikipedia.org\/wiki\/ISO_3166-1_alpha-2\">ISO 3166-1 alpha-2<\/a> specs: two lower-case letters for the language, optionally followed by an underline and two\nupper-case letters identifying the country or regional code. For <a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html#Rare-Language-Codes\">rare languages<\/a>, three letters are used.<\/p><p>For some speakers, the country part may seem redundant. In fact, some languages have dialects in different\ncountries, such as Austrian German (<code class=\"highlighter-rouge\">de_AT<\/code>) or Brazilian Portuguese (<code class=\"highlighter-rouge\">pt_BR<\/code>). The second part is used to distinguish\nbetween those dialects - when it is not present, it is taken as a \u201cgeneric\u201d or \u201chybrid\u201d version of the language.<\/p><h3 id=\"directory-structure\">Directory structure<\/h3><p>To use Gettext, we will need to adhere to a specific structure of folders. First, you will need to select an arbitrary\nroot for your l10n files in your source repository. Inside it, you will have a folder for each needed locale, and a\nfixed <code class=\"highlighter-rouge\">LC_MESSAGES<\/code> folder that will contain all your PO\/MO pairs. Example:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&lt;project root&gt;\n \u251c\u2500 src\/\n \u251c\u2500 templates\/\n \u2514\u2500 locales\/\n    \u251c\u2500 forum.pot\n    \u251c\u2500 site.pot\n    \u251c\u2500 de\/\n    \u2502  \u2514\u2500 LC_MESSAGES\/\n    \u2502     \u251c\u2500 forum.mo\n    \u2502     \u251c\u2500 forum.po\n    \u2502     \u251c\u2500 site.mo\n    \u2502     \u2514\u2500 site.po\n    \u251c\u2500 es_ES\/\n    \u2502  \u2514\u2500 LC_MESSAGES\/\n    \u2502     \u2514\u2500 ...\n    \u251c\u2500 fr\/\n    \u2502  \u2514\u2500 ...\n    \u251c\u2500 pt_BR\/\n    \u2502  \u2514\u2500 ...\n    \u2514\u2500 pt_PT\/\n       \u2514\u2500 ...<\/pre><\/figure><h3 id=\"plural-forms\">Plural forms<\/h3><p>As we said in the introduction, different languages might sport different plural rules. However, gettext saves us from\nthis trouble once again. When creating a new <code class=\"highlighter-rouge\">.po<\/code> file, you will have to declare the <a href=\"http:\/\/docs.translatehouse.org\/projects\/localization-guide\/en\/latest\/l10n\/pluralforms.html\">plural rules<\/a> for that\nlanguage, and translated pieces that are plural-sensitive will have a different form for each of those rules. When\ncalling Gettext in code, you will have to specify the number related to the sentence, and it will work out the correct\nform to use - even using string substitution if needed.<\/p><p>Plural rules include the number of plurals available and a boolean test with <code class=\"highlighter-rouge\">n<\/code> that would define in which rule the\ngiven number falls (starting the count with 0). For example:<\/p><ul><li>Japanese: <code class=\"highlighter-rouge\">nplurals=1; plural=0<\/code> - only one rule<\/li>\n  <li>English: <code class=\"highlighter-rouge\">nplurals=2; plural=(n != 1);<\/code> - two rules, first if N is one, second rule otherwise<\/li>\n  <li>Brazilian Portuguese: <code class=\"highlighter-rouge\">nplurals=2; plural=(n &gt; 1);<\/code> - two rules, second if N is bigger than one, first otherwise<\/li>\n<\/ul><p>Now that you understood the basis of how plural rules works - and if you didn\u2019t, please look at a deeper explanation\non the <a href=\"https:\/\/lingohub.com\/blog\/2013\/07\/php-internationalization-with-gettext-tutorial\/#Plurals\">LingoHub tutorial<\/a> -, you might want to copy the ones you need from a <a href=\"http:\/\/docs.translatehouse.org\/projects\/localization-guide\/en\/latest\/l10n\/pluralforms.html\">list<\/a> instead\nof writing them by hand.<\/p><p>When calling out Gettext to do localization on sentences with counters, you will have to give him the\nrelated number as well. Gettext will work out what rule should be in effect and use the correct localized version.\nYou will need to include in the <code class=\"highlighter-rouge\">.po<\/code> file a different sentence for each plural rule defined.<\/p><h3 id=\"sample-implementation\">Sample implementation<\/h3><p>After all that theory, let\u2019s get a little practical. Here\u2019s an excerpt of a <code class=\"highlighter-rouge\">.po<\/code> file - don\u2019t mind with its format,\nbut with the overall content instead; you will learn how to edit it easily later:<\/p><figure class=\"highlight\"><pre class=\"language-po\" data-lang=\"po\">msgid \"\"\nmsgstr \"\"\n\"Language: pt_BR\\n\"\n\"Content-Type: text\/plain; charset=UTF-8\\n\"\n\"Plural-Forms: nplurals=2; plural=(n &gt; 1);\\n\"\nmsgid \"We are now translating some strings\"\nmsgstr \"N\u00f3s estamos traduzindo algumas strings agora\"\nmsgid \"Hello %1$s! Your last visit was on %2$s\"\nmsgstr \"Ol\u00e1 %1$s! Sua \u00faltima visita foi em %2$s\"\nmsgid \"Only one unread message\"\nmsgid_plural \"%d unread messages\"\nmsgstr[0] \"S\u00f3 uma mensagem n\u00e3o lida\"\nmsgstr[1] \"%d mensagens n\u00e3o lidas\"<\/pre><\/figure><p>The first section works like a header, having the <code class=\"highlighter-rouge\">msgid<\/code> and <code class=\"highlighter-rouge\">msgstr<\/code> especially empty. It describes the file encoding,\nplural forms and other things that are less relevant.\nThe second section translates a simple string from English to\nBrazilian Portuguese, and the third does the same, but leveraging string replacement from <a href=\"https:\/\/secure.php.net\/manual\/function.sprintf.php\"><code class=\"highlighter-rouge\">sprintf<\/code><\/a> so the\ntranslation may contain the user name and visit date.\nThe last section is a sample of pluralization forms, displaying\nthe singular and plural version as <code class=\"highlighter-rouge\">msgid<\/code> in English and their corresponding translations as <code class=\"highlighter-rouge\">msgstr<\/code> 0 and 1\n(following the number given by the plural rule). There, string replacement is used as well so the number can be seen\ndirectly in the sentence, by using <code class=\"highlighter-rouge\">%d<\/code>. The plural forms always have two <code class=\"highlighter-rouge\">msgid<\/code> (singular and plural), so it is\nadvised not to use a complex language as the source of translation.<\/p><h3 id=\"discussion-on-l10n-keys\">Discussion on l10n keys<\/h3><p>As you might have noticed, we are using as source ID the actual sentence in English. That <code class=\"highlighter-rouge\">msgid<\/code> is the same used\nthroughout all your <code class=\"highlighter-rouge\">.po<\/code> files, meaning other languages will have the same format and the same <code class=\"highlighter-rouge\">msgid<\/code> fields but\ntranslated <code class=\"highlighter-rouge\">msgstr<\/code> lines.<\/p><p>Talking about translation keys, there are two main \u201cschools\u201d here:<\/p><ol><li><em><code class=\"highlighter-rouge\">msgid<\/code> as a real sentence<\/em>.\n The main advantages are:\n    <ul><li>if there are pieces of the software untranslated in any given language, the key displayed will still maintain some\n meaning. Example: if you happen to translate by heart from English to Spanish but need help to translate to French,\n you might publish the new page with missing French sentences, and parts of the website would be displayed in English\n instead;<\/li>\n      <li>it is much easier for the translator to understand what\u2019s going on and do a proper translation based on the\n <code class=\"highlighter-rouge\">msgid<\/code>;<\/li>\n      <li>it gives you \u201cfree\u201d l10n for one language - the source one;<\/li>\n      <li>The only disadvantage: if you need to change the actual text, you would need to replace the same <code class=\"highlighter-rouge\">msgid<\/code>\n across several language files.<\/li>\n    <\/ul><\/li>\n  <li><em><code class=\"highlighter-rouge\">msgid<\/code> as a unique, structured key<\/em>.\nIt would describe the sentence role in the application in a structured way, including the template or part where the\nstring is located instead of its content.\n    <ul><li>it is a great way to have the code organized, separating the text content from the template logic.<\/li>\n      <li>however, that could bring problems to the translator that would miss the context. A source language file would be\n needed as a basis for other translations. Example: the developer would ideally have an <code class=\"highlighter-rouge\">en.po<\/code> file, that\n translators would read to understand what to write in <code class=\"highlighter-rouge\">fr.po<\/code> for instance.<\/li>\n      <li>missing translations would display meaningless keys on screen (<code class=\"highlighter-rouge\">top_menu.welcome<\/code> instead of <code class=\"highlighter-rouge\">Hello there, User!<\/code>\n on the said untranslated French page). That is good it as would force translation to be complete before publishing -\n however, bad as translation issues would be remarkably awful in the interface. Some libraries, though, include an\n option to specify a given language as \u201cfallback\u201d, having a similar behavior as the other approach.<\/li>\n    <\/ul><\/li>\n<\/ol><p>The <a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html\">Gettext manual<\/a> favors the first approach as, in general, it is easier for translators and users in\ncase of trouble. That is how we will be working here as well. However, the <a href=\"https:\/\/symfony.com\/doc\/current\/components\/translation\/usage.html#creating-translations\">Symfony documentation<\/a> favors\nkeyword-based translation, to allow for independent changes of all translations without affecting templates as well.<\/p><h3 id=\"everyday-usage\">Everyday usage<\/h3><p>In a typical application, you would use some Gettext functions while writing static text in your pages. Those sentences\nwould then appear in <code class=\"highlighter-rouge\">.po<\/code> files, get translated, compiled into <code class=\"highlighter-rouge\">.mo<\/code> files and then, used by Gettext when rendering\nthe actual interface. Given that, let\u2019s tie together what we have discussed so far in a step-by-step example:<\/p><h4 id=\"1-a-sample-template-file-including-some-different-gettext-calls\">1. A sample template file, including some different gettext calls<\/h4><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php include 'i18n_setup.php' ?&gt;\n&lt;div id=\"header\"&gt;\n    &lt;h1&gt;&lt;?=sprintf(gettext('Welcome, %s!'), $name)?&gt;&lt;\/h1&gt;\n    &lt;!-- code indented this way only for legibility --&gt;\n    &lt;?php if ($unread): ?&gt;\n        &lt;h2&gt;&lt;?=sprintf(\n            ngettext('Only one unread message',\n                     '%d unread messages',\n                     $unread),\n            $unread)?&gt;\n        &lt;\/h2&gt;\n    &lt;?php endif ?&gt;\n&lt;\/div&gt;\n&lt;h1&gt;&lt;?=gettext('Introduction')?&gt;&lt;\/h1&gt;\n&lt;p&gt;&lt;?=gettext('We\\'re now translating some strings')?&gt;&lt;\/p&gt;<\/pre><\/figure><ul><li><a href=\"https:\/\/secure.php.net\/manual\/function.gettext.php\"><code class=\"highlighter-rouge\">gettext()<\/code><\/a> simply translates a <code class=\"highlighter-rouge\">msgid<\/code> into its corresponding <code class=\"highlighter-rouge\">msgstr<\/code> for a given language. There\u2019s also\nthe shorthand function <code class=\"highlighter-rouge\">_()<\/code> that works the same way;<\/li>\n  <li><a href=\"https:\/\/secure.php.net\/manual\/function.ngettext.php\"><code class=\"highlighter-rouge\">ngettext()<\/code><\/a> does the same but with plural rules;<\/li>\n  <li>there\u2019s also <a href=\"https:\/\/secure.php.net\/manual\/function.dgettext.php\"><code class=\"highlighter-rouge\">dgettext()<\/code><\/a> and <a href=\"https:\/\/secure.php.net\/manual\/function.dngettext.php\"><code class=\"highlighter-rouge\">dngettext()<\/code><\/a>, that allows you to override the domain for a single\ncall. More on domain configuration in the next example.<\/li>\n<\/ul><h4 id=\"2-a-sample-setup-file-i18n_setupphp-as-used-above-selecting-the-correct-locale-and-configuring-gettext\">2. A sample setup file (<code class=\"highlighter-rouge\">i18n_setup.php<\/code> as used above), selecting the correct locale and configuring Gettext<\/h4><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/**\n * Verifies if the given $locale is supported in the project\n * @param string $locale\n * @return bool\n *\/\nfunction valid($locale) {\n   return in_array($locale, ['en_US', 'en', 'pt_BR', 'pt', 'es_ES', 'es']);\n}\n\/\/setting the source\/default locale, for informational purposes\n$lang = 'en_US';\nif (isset($_GET['lang']) &amp;&amp; valid($_GET['lang'])) {\n    \/\/ the locale can be changed through the query-string\n    $lang = $_GET['lang'];    \/\/you should sanitize this!\n    setcookie('lang', $lang); \/\/it's stored in a cookie so it can be reused\n} elseif (isset($_COOKIE['lang']) &amp;&amp; valid($_COOKIE['lang'])) {\n    \/\/ if the cookie is present instead, let's just keep it\n    $lang = $_COOKIE['lang']; \/\/you should sanitize this!\n} elseif (isset($_SERVER['HTTP_ACCEPT_LANGUAGE'])) {\n    \/\/ default: look for the languages the browser says the user accepts\n    $langs = explode(',', $_SERVER['HTTP_ACCEPT_LANGUAGE']);\n    array_walk($langs, function (&amp;$lang) { $lang = strtr(strtok($lang, ';'), ['-' =&gt; '_']); });\n    foreach ($langs as $browser_lang) {\n        if (valid($browser_lang)) {\n            $lang = $browser_lang;\n            break;\n        }\n    }\n}\n\/\/ here we define the global system locale given the found language\nputenv(\"LANG=$lang\");\n\/\/ this might be useful for date functions (LC_TIME) or money formatting (LC_MONETARY), for instance\nsetlocale(LC_ALL, $lang);\n\/\/ this will make Gettext look for ..\/locales\/&lt;lang&gt;\/LC_MESSAGES\/main.mo\nbindtextdomain('main', '..\/locales');\n\/\/ indicates in what encoding the file should be read\nbind_textdomain_codeset('main', 'UTF-8');\n\/\/ if your application has additional domains, as cited before, you should bind them here as well\nbindtextdomain('forum', '..\/locales');\nbind_textdomain_codeset('forum', 'UTF-8');\n\/\/ here we indicate the default domain the gettext() calls will respond to\ntextdomain('main');\n\/\/ this would look for the string in forum.mo instead of main.mo\n\/\/ echo dgettext('forum', 'Welcome back!');\n?&gt;<\/pre><\/figure><h4 id=\"3-preparing-translation-for-the-first-run\">3. Preparing translation for the first run<\/h4><p>One of the great advantages Gettext has over custom framework i18n packages is its extensive and powerful file format.\n\u201cOh man, that\u2019s quite hard to understand and edit by hand, a simple array would be easier!\u201d Make no mistake,\napplications like <a href=\"https:\/\/poedit.net\">Poedit<\/a> are here to help - <em>a lot<\/em>. You can get the program from <a href=\"https:\/\/poedit.net\/download\">their website<\/a>,\nit\u2019s free and available for all platforms. It\u2019s a pretty easy tool to get used to, and a very powerful one at the same\ntime - using all features Gettext has available. This guide is based on PoEdit 1.8.<\/p><p>In the first run, you should select \u201cFile &gt; New\u2026\u201d from the menu. You\u2019ll be asked straight ahead for the language:\nhere you can select\/filter the language you want to translate to, or use that format we mentioned before, such as\n<code class=\"highlighter-rouge\">en_US<\/code> or <code class=\"highlighter-rouge\">pt_BR<\/code>.<\/p><p>Now, save the file - using that directory structure we mentioned as well. Then you should click \u201cExtract from sources\u201d,\nand here you\u2019ll configure various settings for the extraction and translation tasks. You\u2019ll be able to find all those\nlater through \u201cCatalog &gt; Properties\u201d:<\/p><ul><li>Source paths: here you must include all folders from the project where <code class=\"highlighter-rouge\">gettext()<\/code> (and siblings) are called - this\nis usually your templates\/views folder(s). This is the only mandatory setting;<\/li>\n  <li>Translation properties:\n    <ul><li>Project name and version, Team and Team\u2019s email address: useful information that goes in the .po file header;<\/li>\n      <li>Plural forms: here go those rules we mentioned before - there\u2019s a link in there with samples as well. You can\n  leave it with the default option most of the time, as PoEdit already includes a handy database of plural rules for\n  many languages.<\/li>\n      <li>Charsets: UTF-8, preferably;<\/li>\n      <li>Source code charset: set here the charset used by your codebase - probably UTF-8 as well, right?<\/li>\n    <\/ul><\/li>\n  <li>Source keywords: The underlying software knows how <code class=\"highlighter-rouge\">gettext()<\/code> and similar function calls look like in several\nprogramming languages, but you might as well create your own translation functions. It will be here you\u2019ll add those\nother methods. This will be discussed later in the \u201cTips\u201d section.<\/li>\n<\/ul><p>After setting those points it will run a scan through your source files to find all the localization calls. After every\nscan PoEdit will display a summary of what was found and what was removed from the source files. New entries will fed\nempty into the translation table, and you\u2019ll start typing in the localized versions of those strings. Save it and a .mo\nfile will be (re)compiled into the same folder and ta-dah: your project is internationalized.<\/p><h4 id=\"4-translating-strings\">4. Translating strings<\/h4><p>As you may have noticed before, there are two main types of localized strings: simple ones and those with plural\nforms. The first ones have simply two boxes: source and localized string. The source string cannot be modified as\nGettext\/Poedit do not include the powers to alter your source files - you should change the source itself and rescan\nthe files. Tip: you may right-click a translation line and it will hint you with the source files and lines where that\nstring is being used.\nOn the other hand, plural form strings include two boxes to show the two source strings, and tabs so you can configure\nthe different final forms.<\/p><p>Whenever you change your sources and need to update the translations, just hit Refresh and Poedit will rescan the code,\nremoving non-existent entries, merging the ones that changed and adding new ones. It may also try to guess some\ntranslations, based on other ones you did. Those guesses and the changed entries will receive a \u201cFuzzy\u201d marker,\nindicating it needs review, appearing golden in the list. It is also useful if you have a translation team and someone\ntries to write something they are not sure about: just mark Fuzzy, and someone else will review later.<\/p><p>Finally, it is advised to leave \u201cView &gt; Untranslated entries first\u201d marked, as it will help you <em>a lot<\/em> to not forget\nany entry. From that menu, you can also open parts of the UI that allow you to leave contextual information for\ntranslators if needed.<\/p><h3 id=\"tips--tricks\">Tips &amp; Tricks<\/h3><h4 id=\"possible-caching-issues\">Possible caching issues<\/h4><p>If you are running PHP as a module on Apache (<code class=\"highlighter-rouge\">mod_php<\/code>), you might face issues with the <code class=\"highlighter-rouge\">.mo<\/code> file being cached. It\nhappens the first time it is read, and then, to update it, you might need to restart the server. On Nginx and PHP5 it\nusually takes only a couple of page refreshes to refresh the translation cache, and on PHP7 it is rarely needed.<\/p><p>As preferred by many people, it is easier to use <code class=\"highlighter-rouge\">_()<\/code> instead of <code class=\"highlighter-rouge\">gettext()<\/code>. Many custom i18n libraries from\nframeworks use something similar to <code class=\"highlighter-rouge\">t()<\/code> as well, to make translated code shorter. However, that is the only function\nthat sports a shortcut. You might want to add in your project some others, such as <code class=\"highlighter-rouge\">__()<\/code> or <code class=\"highlighter-rouge\">_n()<\/code> for <code class=\"highlighter-rouge\">ngettext()<\/code>,\nor maybe a fancy <code class=\"highlighter-rouge\">_r()<\/code> that would join <code class=\"highlighter-rouge\">gettext()<\/code> and <code class=\"highlighter-rouge\">sprintf()<\/code> calls. Other libraries, such as\n<a href=\"https:\/\/github.com\/oscarotero\/Gettext\">oscarotero\u2019s Gettext<\/a> also provide helper functions like these.<\/p><p>In those cases, you\u2019ll need to instruct the Gettext utility on how to extract the strings from those new functions.\nDon\u2019t be afraid; it is very easy. It is just a field in the <code class=\"highlighter-rouge\">.po<\/code> file, or a Settings screen on Poedit. In the editor,\nthat option is inside \u201cCatalog &gt; Properties &gt; Source keywords\u201d. Remember: Gettext already knows the default functions\nfor many languages, so don\u2019t be afraid if that list seems empty. You need to include there the specifications of those\nnew functions, following <a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html#Language-specific-options\">a specific format<\/a>:<\/p><ul><li>if you create something like <code class=\"highlighter-rouge\">t()<\/code> that simply returns the translation for a string, you can specify it as <code class=\"highlighter-rouge\">t<\/code>.\nGettext will know the only function argument is the string to be translated;<\/li>\n  <li>if the function has more than one argument, you can specify in which one the first string is - and if needed, the\nplural form as well. For instance, if we call our function like this: <code class=\"highlighter-rouge\">__('one user', '%d users', $number)<\/code>, the\nspecification would be <code class=\"highlighter-rouge\">__:1,2<\/code>, meaning the first form is the first argument, and the second form is the second\nargument. If your number comes as the first argument instead, the spec would be <code class=\"highlighter-rouge\">__:2,3<\/code>, indicating the first form is\nthe second argument, and so on.<\/li>\n<\/ul><p>After including those new rules in the <code class=\"highlighter-rouge\">.po<\/code> file, a new scan will bring in your new strings just as easy as before.<\/p><h3 id=\"references\">References<\/h3><ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Internationalization_and_localization\">Wikipedia: i18n and l10n<\/a><\/li>\n  <li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gettext\">Wikipedia: Gettext<\/a><\/li>\n  <li><a href=\"https:\/\/lingohub.com\/blog\/2013\/07\/php-internationalization-with-gettext-tutorial\/\">LingoHub: PHP internationalization with gettext tutorial<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/manual\/book.gettext.php\">PHP Manual: Gettext<\/a><\/li>\n  <li><a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html\">Gettext Manual<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"dependency_injection\"><p>From <a href=\"https:\/\/wikipedia.org\/wiki\/Dependency_injection\">Wikipedia<\/a>:<\/p><blockquote>\n  <p>Dependency injection is a software design pattern that allows the removal of hard-coded dependencies and makes it\npossible to change them, whether at run-time or compile-time.<\/p>\n<\/blockquote><p>This quote makes the concept sound much more complicated than it actually is. Dependency Injection is providing a\ncomponent with its dependencies either through constructor injection, method calls or the setting of properties. It is\nthat simple.<\/p><\/section><section class=\"chapter\" id=\"basic_concept\"><h2 id=\"basic_concept_title\">Basic Concept<\/h2><p>We can demonstrate the concept with a simple, yet naive example.<\/p><p>Here we have a <code class=\"highlighter-rouge\">Database<\/code> class that requires an adapter to speak to the database. We instantiate the adapter in the\nconstructor and create a hard dependency. This makes testing difficult and means the <code class=\"highlighter-rouge\">Database<\/code> class is very tightly\ncoupled to the adapter.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nnamespace Database;\nclass Database\n{\n    protected $adapter;\n    public function __construct()\n    {\n        $this-&gt;adapter = new MySqlAdapter;\n    }\n}\nclass MysqlAdapter {}<\/pre><\/figure><p>This code can be refactored to use Dependency Injection and therefore loosen the dependency.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nnamespace Database;\nclass Database\n{\n    protected $adapter;\n    public function __construct(MySqlAdapter $adapter)\n    {\n        $this-&gt;adapter = $adapter;\n    }\n}\nclass MysqlAdapter {}<\/pre><\/figure><p>Now we are giving the <code class=\"highlighter-rouge\">Database<\/code> class its dependency rather than creating it itself. We could even create a method\nthat would accept an argument of the dependency and set it that way, or if the <code class=\"highlighter-rouge\">$adapter<\/code> property was <code class=\"highlighter-rouge\">public<\/code> we\ncould set it directly.<\/p><\/section><section class=\"chapter\" id=\"complex_problem\"><h2 id=\"complex_problem_title\">Complex Problem<\/h2><p>If you have ever read about Dependency Injection then you have probably seen the terms <em>\u201cInversion of Control\u201d<\/em> or\n<em>\u201cDependency Inversion Principle\u201d<\/em>. These are the complex problems that Dependency Injection solves.<\/p><h3 id=\"inversion-of-control\">Inversion of Control<\/h3><p>Inversion of Control is as it says, \u201cinverting the control\u201d of a system by keeping organizational control entirely\nseparate from our objects. In terms of Dependency Injection, this means loosening our dependencies by controlling and\ninstantiating them elsewhere in the system.<\/p><p>For years, PHP frameworks have been achieving Inversion of Control, however, the question became, which part of control\nare we inverting, and where to? For example, MVC frameworks would generally provide a super object or base controller\nthat other controllers must extend to gain access to its dependencies. This <strong>is<\/strong> Inversion of Control, however,\ninstead of loosening dependencies, this method simply moved them.<\/p><p>Dependency Injection allows us to more elegantly solve this problem by only injecting the dependencies we need, when we\nneed them, without the need for any hard coded dependencies at all.<\/p><h3 id=\"solid\">S.O.L.I.D.<\/h3><h4 id=\"single-responsibility-principle\">Single Responsibility Principle<\/h4><p>The Single Responsibility Principle is about actors and high-level architecture. It states that \u201cA class should have\nonly one reason to change.\u201d This means that every class should <em>only<\/em> have responsibility over a single part of the\nfunctionality provided by the software. The largest benefit of this approach is that it enables improved code\n<em>reusability<\/em>. By designing our class to do just one thing, we can use (or re-use) it in any other program without\nchanging it.<\/p><h4 id=\"openclosed-principle\">Open\/Closed Principle<\/h4><p>The Open\/Closed Principle is about class design and feature extensions. It states that \u201cSoftware entities (classes,\nmodules, functions, etc.) should be open for extension, but closed for modification.\u201d This means that we should design\nour modules, classes and functions in a way that when a new functionality is needed, we should not modify our existing\ncode but rather write new code that will be used by existing code. Practically speaking, this means that we should write\nclasses that implement and adhere to <em>interfaces<\/em>, then type-hint against those interfaces instead of specific classes.<\/p><p>The largest benefit of this approach is that we can very easily extend our code with support for something new without\nhaving to modify existing code, meaning that we can reduce QA time, and the risk for negative impact to the application\nis substantially reduced. We can deploy new code, faster, and with more confidence.<\/p><h4 id=\"liskov-substitution-principle\">Liskov Substitution Principle<\/h4><p>The Liskov Substitution Principle is about subtyping and inheritance. It states that \u201cChild classes should never break\nthe parent class\u2019 type definitions.\u201d Or, in Robert C. Martin\u2019s words, \u201cSubtypes must be substitutable for their base\ntypes.\u201d<\/p><p>For example, if we have a <code class=\"highlighter-rouge\">FileInterface<\/code> interface which defines an <code class=\"highlighter-rouge\">embed()<\/code> method, and we have <code class=\"highlighter-rouge\">Audio<\/code> and <code class=\"highlighter-rouge\">Video<\/code>\nclasses which both implement the <code class=\"highlighter-rouge\">embed()<\/code> method, then we can expect that the usage of the <code class=\"highlighter-rouge\">embed()<\/code> method will always\ndo the thing that we intend. If we later create a <code class=\"highlighter-rouge\">PDF<\/code> class or a <code class=\"highlighter-rouge\">Gist<\/code> class which implement the <code class=\"highlighter-rouge\">FileInterface<\/code>\ninterface, we will already know and understand what the <code class=\"highlighter-rouge\">embed()<\/code> method will do. The largest benefit of this approach\nis that we have the ability to build flexible and easily-configurable programs, because when we change one object of a\ntype (e.g., <code class=\"highlighter-rouge\">FileInterface<\/code>) to another we don\u2019t need to change anything else in our program.<\/p><h4 id=\"interface-segregation-principle\">Interface Segregation Principle<\/h4><p>The Interface Segregation Principle (ISP) is about <em>business-logic-to-clients<\/em> communication. It states that \u201cNo client\nshould be forced to depend on methods it does not use.\u201d This means that instead of having a single monolithic interface\nthat all conforming classes need to implement, we should instead provide a set of smaller, concept-specific interfaces\nthat a conforming class implements one or more of.<\/p><p>For example, a <code class=\"highlighter-rouge\">Car<\/code> or <code class=\"highlighter-rouge\">Bus<\/code> class would be interested in a <code class=\"highlighter-rouge\">steeringWheel()<\/code> method, but a <code class=\"highlighter-rouge\">Motorcycle<\/code> or <code class=\"highlighter-rouge\">Tricycle<\/code>\nclass would not. Conversely, a <code class=\"highlighter-rouge\">Motorcycle<\/code> or <code class=\"highlighter-rouge\">Tricycle<\/code> class would be interested in a <code class=\"highlighter-rouge\">handlebars()<\/code> method, but a\n<code class=\"highlighter-rouge\">Car<\/code> or <code class=\"highlighter-rouge\">Bus<\/code> class would not. There is no need to have all of these types of vehicles implement support for both\n<code class=\"highlighter-rouge\">steeringWheel()<\/code> as well as <code class=\"highlighter-rouge\">handlebars()<\/code>, so we should break-apart the source interface.<\/p><h4 id=\"dependency-inversion-principle\">Dependency Inversion Principle<\/h4><p>The Dependency Inversion Principle is about removing hard-links between discrete classes so that new functionality can\nbe leveraged by passing a different class. It states that one should <em>\u201cDepend on Abstractions. Do not depend on\nconcretions.\u201d<\/em>. Put simply, this means our dependencies should be interfaces\/contracts or abstract classes rather than\nconcrete implementations. We can easily refactor the above example to follow this principle.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nnamespace Database;\nclass Database\n{\n    protected $adapter;\n    public function __construct(AdapterInterface $adapter)\n    {\n        $this-&gt;adapter = $adapter;\n    }\n}\ninterface AdapterInterface {}\nclass MysqlAdapter implements AdapterInterface {}<\/pre><\/figure><p>There are several benefits to the <code class=\"highlighter-rouge\">Database<\/code> class now depending on an interface rather than a concretion.<\/p><p>Consider that we are working in a team and the adapter is being worked on by a colleague. In our first example, we\nwould have to wait for said colleague to finish the adapter before we could properly mock it for our unit tests. Now\nthat the dependency is an interface\/contract we can happily mock that interface knowing that our colleague will build\nthe adapter based on that contract.<\/p><p>An even bigger benefit to this method is that our code is now much more scalable. If a year down the line we decide\nthat we want to migrate to a different type of database, we can write an adapter that implements the original interface\nand injects that instead, no more refactoring would be required as we can ensure that the adapter follows the contract\nset by the interface.<\/p><\/section><section class=\"chapter\" id=\"containers\"><h2 id=\"containers_title\">Containers<\/h2><p>The first thing you should understand about Dependency Injection Containers is that they are not the same thing as\nDependency Injection. A container is a convenience utility that helps us implement Dependency Injection, however, they\ncan be and often are misused to implement an anti-pattern, Service Location. Injecting a DI container as a Service\nLocator in to your classes arguably creates a harder dependency on the container than the dependency you are replacing.\nIt also makes your code much less transparent and ultimately harder to test.<\/p><p>Most modern frameworks have their own Dependency Injection Container that allows you to wire your dependencies together\nthrough configuration. What this means in practice is that you can write application code that is as clean and de-\ncoupled as the framework it is built on.<\/p><\/section><section class=\"chapter\" id=\"further_reading\"><h2 id=\"further_reading_title\">Further Reading<\/h2><ul><li><a href=\"http:\/\/ralphschindler.com\/2011\/05\/18\/learning-about-dependency-injection-and-php\">Learning about Dependency Injection and PHP<\/a><\/li>\n  <li><a href=\"http:\/\/fabien.potencier.org\/article\/11\/what-is-dependency-injection\">What is Dependency Injection?<\/a><\/li>\n  <li><a href=\"https:\/\/mwop.net\/blog\/260-Dependency-Injection-An-analogy.html\">Dependency Injection: An analogy<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/dependency-injection-huh--net-26903\">Dependency Injection: Huh?<\/a><\/li>\n  <li><a href=\"https:\/\/medium.com\/philipobenito\/dependency-injection-as-a-tool-for-testing-902c21c147f1\">Dependency Injection as a tool for testing<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"databases\"><p>Many times your PHP code will use a database to persist information. You have a few options to connect and interact\nwith your database. The recommended option <strong>until PHP 5.1.0<\/strong> was to use native drivers such as <a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a>, <a href=\"https:\/\/secure.php.net\/pgsql\">pgsql<\/a>,\n<a href=\"https:\/\/secure.php.net\/mssql\">mssql<\/a>, etc.<\/p><p>Native drivers are great if you are only using <em>one<\/em> database in your application, but if, for example, you are using\nMySQL and a little bit of MSSQL, or you need to connect to an Oracle database, then you will not be able to use the\nsame drivers. You\u2019ll need to learn a brand new API for each database \u2014 and that can get silly.<\/p><\/section><section class=\"chapter\" id=\"mysql_extension\"><h2 id=\"mysql_extension_title\">MySQL Extension<\/h2><p>The <a href=\"https:\/\/secure.php.net\/mysqli\">mysql<\/a> extension for PHP is incredibly old and has been superseded by two other extensions:<\/p><ul><li><a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/pdo\">pdo<\/a><\/li>\n<\/ul><p>Not only did development stop long ago on <a href=\"https:\/\/secure.php.net\/mysqli\">mysql<\/a>, but it was <a href=\"https:\/\/secure.php.net\/migration55.deprecated\">deprecated as of PHP 5.5.0<\/a>, and <strong>has been <a href=\"https:\/\/secure.php.net\/manual\/migration70.removed-exts-sapis.php\">officially removed in PHP 7.0<\/a><\/strong>.<\/p><p>To save digging into your <code class=\"highlighter-rouge\">php.ini<\/code> settings to see which module you are using, one option is to search for <code class=\"highlighter-rouge\">mysql_*<\/code>\nin your editor of choice. If any functions such as <code class=\"highlighter-rouge\">mysql_connect()<\/code> and <code class=\"highlighter-rouge\">mysql_query()<\/code> show up, then <code class=\"highlighter-rouge\">mysql<\/code> is\nin use.<\/p><p>Even if you are not using PHP 7.x yet, failing to consider this upgrade as soon as possible will lead to greater\nhardship when the PHP 7.x upgrade does come about. The best option is to replace mysql usage with <a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a> or <a href=\"https:\/\/secure.php.net\/pdo\">PDO<\/a> in\nyour applications within your own development schedules so you won\u2019t be rushed later on.<\/p><p><strong>If you are upgrading from <a href=\"https:\/\/secure.php.net\/mysqli\">mysql<\/a> to <a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a>, beware lazy upgrade guides that suggest you can simply find and replace <code class=\"highlighter-rouge\">mysql_*<\/code> with <code class=\"highlighter-rouge\">mysqli_*<\/code>. Not only is that a gross oversimplification, it misses out on the advantages that mysqli provides, such as parameter binding, which is also offered in <a href=\"https:\/\/secure.php.net\/pdo\">PDO<\/a>.<\/strong><\/p><ul><li><a href=\"https:\/\/secure.php.net\/mysqlinfo.api.choosing\">PHP: Choosing an API for MySQL<\/a><\/li>\n  <li><a href=\"http:\/\/wiki.hashphp.org\/PDO_Tutorial_for_MySQL_Developers\">PDO Tutorial for MySQL Developers<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"pdo_extension\"><h2 id=\"pdo_extension_title\">PDO Extension<\/h2><p><a href=\"https:\/\/secure.php.net\/pdo\">PDO<\/a> is a database connection abstraction library \u2014 built into PHP since 5.1.0 \u2014 that provides a common\ninterface to talk with many different databases. For example, you can use basically identical code to interface with\nMySQL or SQLite:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ PDO + MySQL\n$pdo = new PDO('mysql:host=example.com;dbname=database', 'user', 'password');\n$statement = $pdo-&gt;query(\"SELECT some_field FROM some_table\");\n$row = $statement-&gt;fetch(PDO::FETCH_ASSOC);\necho htmlentities($row['some_field']);\n\/\/ PDO + SQLite\n$pdo = new PDO('sqlite:\/path\/db\/foo.sqlite');\n$statement = $pdo-&gt;query(\"SELECT some_field FROM some_table\");\n$row = $statement-&gt;fetch(PDO::FETCH_ASSOC);\necho htmlentities($row['some_field']);<\/pre><\/figure><p>PDO will not translate your SQL queries or emulate missing features; it is purely for connecting to multiple types of\ndatabase with the same API.<\/p><p>More importantly, <code class=\"highlighter-rouge\">PDO<\/code> allows you to safely inject foreign input (e.g. IDs) into your SQL queries without worrying\nabout database SQL injection attacks.\nThis is possible using PDO statements and bound parameters.<\/p><p>Let\u2019s assume a PHP script receives a numeric ID as a query parameter. This ID should be used to fetch a user record\nfrom a database. This is the <code class=\"highlighter-rouge\">wrong<\/code> way to do this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$pdo = new PDO('sqlite:\/path\/db\/users.db');\n$pdo-&gt;query(\"SELECT name FROM users WHERE id = \" . $_GET['id']); \/\/ &lt;-- NO!<\/pre><\/figure><p>This is terrible code. You are inserting a raw query parameter into a SQL query. This will get you hacked in a\nheartbeat, using a practice called <a href=\"http:\/\/wiki.hashphp.org\/Validation\">SQL Injection<\/a>. Just imagine if a hacker passes in an inventive <code class=\"highlighter-rouge\">id<\/code> parameter by\ncalling a URL like <code class=\"highlighter-rouge\">http:\/\/domain.com\/?id=1%3BDELETE+FROM+users<\/code>. This will set the <code class=\"highlighter-rouge\">$_GET['id']<\/code> variable to <code class=\"highlighter-rouge\">1;DELETE\nFROM users<\/code> which will delete all of your users! Instead, you should sanitize the ID input using PDO bound parameters.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$pdo = new PDO('sqlite:\/path\/db\/users.db');\n$stmt = $pdo-&gt;prepare('SELECT name FROM users WHERE id = :id');\n$id = filter_input(INPUT_GET, 'id', FILTER_SANITIZE_NUMBER_INT); \/\/ &lt;-- filter your data first (see [Data Filtering](#data_filtering)), especially important for INSERT, UPDATE, etc.\n$stmt-&gt;bindParam(':id', $id, PDO::PARAM_INT); \/\/ &lt;-- Automatically sanitized for SQL by PDO\n$stmt-&gt;execute();<\/pre><\/figure><p>This is correct code. It uses a bound parameter on a PDO statement. This escapes the foreign input ID before it is\nintroduced to the database preventing potential SQL injection attacks.<\/p><p>For writes, such as INSERT or UPDATE, it\u2019s especially critical to still <a href=\"#data_filtering\">filter your data<\/a> first and sanitize it for other things (removal of HTML tags, JavaScript, etc).  PDO will only sanitize it for SQL, not for your application.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.pdo\">Learn about PDO<\/a><\/li>\n<\/ul><p>You should also be aware that database connections use up resources and it was not unheard-of to have resources\nexhausted if connections were not implicitly closed, however this was more common in other languages. Using PDO you can\nimplicitly close the connection by destroying the object by ensuring all remaining references to it are deleted, i.e.\nset to NULL. If you don\u2019t do this explicitly, PHP will automatically close the connection when your script ends -\nunless of course you are using persistent connections.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/pdo.connections\">Learn about PDO connections<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"databases_interacting\"><h2 id=\"databases_interacting_title\">Interacting with Databases<\/h2><p>When developers first start to learn PHP, they often end up mixing their database interaction up with their\npresentation logic, using code that might look like this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;ul&gt;\n&lt;?php\nforeach ($db-&gt;query('SELECT * FROM table') as $row) {\n    echo \"&lt;li&gt;\".$row['field1'].\" - \".$row['field1'].\"&lt;\/li&gt;\";\n}\n?&gt;\n&lt;\/ul&gt;<\/pre><\/figure><p>This is bad practice for all sorts of reasons, mainly that it\u2019s hard to debug, hard to test, hard to read and it is\ngoing to output a lot of fields if you don\u2019t put a limit on there.<\/p><p>While there are many other solutions to doing this - depending on if you prefer <a href=\"https:\/\/phptherightway.com\/#object-oriented-programming\">OOP<\/a> or\n<a href=\"https:\/\/phptherightway.com\/#functional-programming\">functional programming<\/a> - there must be some element of separation.<\/p><p>Consider the most basic step:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nfunction getAllFoos($db) {\n    return $db-&gt;query('SELECT * FROM table');\n}\n$results = getAllFoos($db);\nforeach ($results as $row) {\n    echo \"&lt;li&gt;\".$row['field1'].\" - \".$row['field1'].\"&lt;\/li&gt;\"; \/\/ BAD!!\n}<\/pre><\/figure><p>That is a good start. Put those two items in two different files and you\u2019ve got some clean separation.<\/p><p>Create a class to place that method in and you have a \u201cModel\u201d. Create a simple <code class=\"highlighter-rouge\">.php<\/code> file to put the presentation\nlogic in and you have a \u201cView\u201d, which is very nearly <a href=\"https:\/\/code.tutsplus.com\/tutorials\/mvc-for-noobs--net-10488\">MVC<\/a> - a common OOP architecture for most\n<a href=\"https:\/\/phptherightway.com\/#frameworks\">frameworks<\/a>.<\/p><p><strong>foo.php<\/strong><\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$db = new PDO('mysql:host=localhost;dbname=testdb;charset=utf8mb4', 'username', 'password');\n\/\/ Make your model available\ninclude 'models\/FooModel.php';\n\/\/ Create an instance\n$fooModel = new FooModel($db);\n\/\/ Get the list of Foos\n$fooList = $fooModel-&gt;getAllFoos();\n\/\/ Show the view\ninclude 'views\/foo-list.php';<\/pre><\/figure><p><strong>models\/FooModel.php<\/strong><\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nclass FooModel\n{\n    protected $db;\n    public function __construct(PDO $db)\n    {\n        $this-&gt;db = $db;\n    }\n    public function getAllFoos() {\n        return $this-&gt;db-&gt;query('SELECT * FROM table');\n    }\n}<\/pre><\/figure><p><strong>views\/foo-list.php<\/strong><\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php foreach ($fooList as $row): ?&gt;\n    &lt;li&gt;&lt;?= $row['field1'] ?&gt; - &lt;?= $row['field1'] ?&gt;&lt;\/li&gt;\n&lt;?php endforeach ?&gt;<\/pre><\/figure><p>This is essentially the same as what most modern frameworks are doing, albeit a little more manual. You might not\nneed to do all of that every time, but mixing together too much presentation logic and database interaction can be a\nreal problem if you ever want to <a href=\"https:\/\/phptherightway.com\/#unit-testing\">unit-test<\/a> your application.<\/p><p><a href=\"https:\/\/phpbridge.org\/docs\/\">PHPBridge<\/a> has a great resource called <a href=\"https:\/\/phpbridge.org\/intro-to-php\/creating_a_data_class\">Creating a Data Class<\/a> which covers a very similar topic, and is great for\ndevelopers just getting used to the concept of interacting with databases.<\/p><\/section><section class=\"chapter\" id=\"databases_abstraction_layers\"><h2 id=\"databases_abstraction_layers_title\">Abstraction Layers<\/h2><p>Many frameworks provide their own abstraction layer which may or may not sit on top of <a href=\"https:\/\/secure.php.net\/book.pdo\">PDO<\/a>. These will often\nemulate features for one database system that is missing from another by wrapping your queries in PHP methods, giving\nyou actual database abstraction instead of just the connection abstraction that PDO provides. This will of course add a\nlittle overhead, but if you are building a portable application that needs to work with MySQL, PostgreSQL and SQLite\nthen a little overhead will be worth it the sake of code cleanliness.<\/p><p>Some abstraction layers have been built using the <a href=\"https:\/\/www.php-fig.org\/psr\/psr-0\/\">PSR-0<\/a> or <a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">PSR-4<\/a> namespace standards so can be\ninstalled in any application you like:<\/p><ul><li><a href=\"https:\/\/github.com\/auraphp\/Aura.Sql\">Aura SQL<\/a><\/li>\n  <li><a href=\"https:\/\/www.doctrine-project.org\/projects\/dbal.html\">Doctrine2 DBAL<\/a><\/li>\n  <li><a href=\"http:\/\/propelorm.org\/\">Propel<\/a><\/li>\n  <li><a href=\"https:\/\/packages.zendframework.com\/docs\/latest\/manual\/en\/index.html#zendframework\/zend-db\">Zend-db<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"templating\"><p>Templates provide a convenient way of separating your controller and domain logic from your presentation logic.\nTemplates typically contain the HTML of your application, but may also be used for other formats, such as XML.\nTemplates are often referred to as \u201cviews\u201d, which make up <strong>part of<\/strong> the second component of the\n<a href=\"https:\/\/phptherightway.com\/pages\/Design-Patterns.html#model-view-controller\">model\u2013view\u2013controller<\/a> (MVC) software architecture pattern.<\/p><\/section><section class=\"chapter\" id=\"templating_benefits\"><h2 id=\"templating_benefits_title\">Benefits<\/h2><p>The main benefit to using templates is the clear separation they create between the presentation logic and the rest of\nyour application. Templates have the sole responsibility of displaying formatted content. They are not responsible for\ndata lookup, persistence or other more complex tasks. This leads to cleaner, more readable code which is especially\nhelpful in a team environment where developers work on the server-side code (controllers, models) and designers work on\nthe client-side code (markup).<\/p><p>Templates also improve the organization of presentation code. Templates are typically placed in a \u201cviews\u201d folder, each\ndefined within a single file. This approach encourages code reuse where larger blocks of code are broken into smaller,\nreusable pieces, often called partials. For example, your site header and footer can each be defined as templates,\nwhich are then included before and after each page template.<\/p><p>Finally, depending on the library you use, templates can offer more security by automatically escaping user-generated\ncontent. Some libraries even offer sand-boxing, where template designers are only given access to white-listed\nvariables and functions.<\/p><\/section><section class=\"chapter\" id=\"plain_php_templates\"><h2 id=\"plain_php_templates_title\">Plain PHP Templates<\/h2><p>Plain PHP templates are simply templates that use native PHP code. They are a natural choice since PHP is actually a\ntemplate language itself. That simply means that you can combine PHP code within other code, like HTML. This is\nbeneficial to PHP developers as there is no new syntax to learn, they know the functions available to them, and their\ncode editors already have PHP syntax highlighting and auto-completion built-in. Further, plain PHP templates tend to be\nvery fast as no compiling stage is required.<\/p><p>Every modern PHP framework employs some kind of template system, most of which use plain PHP by default. Outside of\nframeworks, libraries like <a href=\"http:\/\/platesphp.com\/\">Plates<\/a> or <a href=\"https:\/\/github.com\/auraphp\/Aura.View\">Aura.View<\/a> make working with plain PHP templates easier by\noffering modern template functionality such as inheritance, layouts and extensions.<\/p><h3 id=\"simple-example-of-a-plain-php-template\">Simple example of a plain PHP template<\/h3><p>Using the <a href=\"http:\/\/platesphp.com\/\">Plates<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php \/\/ user_profile.php ?&gt;\n&lt;?php $this-&gt;insert('header', ['title' =&gt; 'User Profile']) ?&gt;\n&lt;h1&gt;User Profile&lt;\/h1&gt;\n&lt;p&gt;Hello, &lt;?=$this-&gt;escape($name)?&gt;&lt;\/p&gt;\n&lt;?php $this-&gt;insert('footer') ?&gt;<\/pre><\/figure><h3 id=\"example-of-plain-php-templates-using-inheritance\">Example of plain PHP templates using inheritance<\/h3><p>Using the <a href=\"http:\/\/platesphp.com\/\">Plates<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php \/\/ template.php ?&gt;\n&lt;html&gt;\n&lt;head&gt;\n    &lt;title&gt;&lt;?=$title?&gt;&lt;\/title&gt;\n&lt;\/head&gt;\n&lt;body&gt;\n&lt;main&gt;\n    &lt;?=$this-&gt;section('content')?&gt;\n&lt;\/main&gt;\n&lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/figure><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php \/\/ user_profile.php ?&gt;\n&lt;?php $this-&gt;layout('template', ['title' =&gt; 'User Profile']) ?&gt;\n&lt;h1&gt;User Profile&lt;\/h1&gt;\n&lt;p&gt;Hello, &lt;?=$this-&gt;escape($name)?&gt;&lt;\/p&gt;<\/pre><\/figure><\/section><section class=\"chapter\" id=\"compiled_templates\"><h2 id=\"compiled_templates_title\">Compiled Templates<\/h2><p>While PHP has evolved into a mature, object oriented language, it <a href=\"http:\/\/fabien.potencier.org\/article\/34\/templating-engines-in-php\">hasn\u2019t improved much<\/a> as\na templating language. Compiled templates, like <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a>, <a href=\"https:\/\/github.com\/box\/brainy\">Brainy<\/a>, or <a href=\"https:\/\/www.smarty.net\/\">Smarty<\/a>*, fill this void by offering a new syntax that has\nbeen geared specifically to templating. From automatic escaping, to inheritance and simplified control structures,\ncompiled templates are designed to be easier to write, cleaner to read and safer to use. Compiled templates can even be\nshared across different languages, <a href=\"https:\/\/mustache.github.io\/\">Mustache<\/a> being a good example of this. Since these templates must be compiled\nthere is a slight performance hit, however this is very minimal when proper caching is used.<\/p><p><em>*While Smarty offers automatic escaping, this feature is NOT enabled by default.<\/em><\/p><h3 id=\"simple-example-of-a-compiled-template\">Simple example of a compiled template<\/h3><p>Using the <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-html-jinja\" data-lang=\"html+jinja\">{% include 'header.html' with {'title': 'User Profile'} %}\n&lt;h1&gt;User Profile&lt;\/h1&gt;\n&lt;p&gt;Hello, {{ name }}&lt;\/p&gt;\n{% include 'footer.html' %}<\/pre><\/figure><h3 id=\"example-of-compiled-templates-using-inheritance\">Example of compiled templates using inheritance<\/h3><p>Using the <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-html-jinja\" data-lang=\"html+jinja\">\/\/ template.html\n&lt;html&gt;\n&lt;head&gt;\n    &lt;title&gt;{% block title %}{% endblock %}&lt;\/title&gt;\n&lt;\/head&gt;\n&lt;body&gt;\n&lt;main&gt;\n    {% block content %}{% endblock %}\n&lt;\/main&gt;\n&lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/figure><figure class=\"highlight\"><pre class=\"language-html-jinja\" data-lang=\"html+jinja\">\/\/ user_profile.html\n{% extends \"template.html\" %}\n{% block title %}User Profile{% endblock %}\n{% block content %}\n    &lt;h1&gt;User Profile&lt;\/h1&gt;\n    &lt;p&gt;Hello, {{ name }}&lt;\/p&gt;\n{% endblock %}<\/pre><\/figure><\/section><section class=\"chapter\" id=\"templating_further_reading\"><h2 id=\"templating_further_reading_title\">Further Reading<\/h2><h3 id=\"articles--tutorials\">Articles &amp; Tutorials<\/h3><ul><li><a href=\"http:\/\/fabien.potencier.org\/article\/34\/templating-engines-in-php\">Templating Engines in PHP<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/an-introduction-to-views-templating-in-codeigniter--net-25648\">An Introduction to Views &amp; Templating in CodeIgniter<\/a><\/li>\n  <li><a href=\"https:\/\/www.smashingmagazine.com\/2011\/10\/getting-started-with-php-templating\/\">Getting Started With PHP Templating<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/roll-your-own-templating-system-in-php--net-16596\">Roll Your Own Templating System in PHP<\/a><\/li>\n  <li><a href=\"https:\/\/laracasts.com\/series\/laravel-from-scratch\/episodes\/7\">Master Pages<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/working-with-templates-in-symfony-2--cms-21172\">Working With Templates in Symfony 2<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/box\/brainy\/wiki\/Writing-Safe-Templates\">Writing Safer Templates<\/a><\/li>\n<\/ul><h3 id=\"libraries\">Libraries<\/h3><ul><li><a href=\"https:\/\/github.com\/auraphp\/Aura.View\">Aura.View<\/a> <em>(native)<\/em><\/li>\n  <li><a href=\"https:\/\/laravel.com\/docs\/blade\">Blade<\/a> <em>(compiled, framework specific)<\/em><\/li>\n  <li><a href=\"https:\/\/github.com\/box\/brainy\">Brainy<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"http:\/\/dwoo.org\/\">Dwoo<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/github.com\/nette\/latte\">Latte<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/github.com\/bobthecow\/mustache.php\">Mustache<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/phptal.org\/\">PHPTAL<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"http:\/\/platesphp.com\/\">Plates<\/a> <em>(native)<\/em><\/li>\n  <li><a href=\"https:\/\/www.smarty.net\/\">Smarty<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/framework.zend.com\/manual\/2.3\/en\/modules\/zend.view.quick-start.html\">Zend\\View<\/a> <em>(native, framework specific)<\/em><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"errors\"><h2 id=\"errors_title\">Errors<\/h2><p>In many \u201cexception-heavy\u201d programming languages, whenever anything goes wrong an exception will be thrown. This is\ncertainly a viable way to do things, but PHP is an \u201cexception-light\u201d programming language. While it does have\nexceptions and more of the core is starting to use them when working with objects, most of PHP itself will try to keep\nprocessing regardless of what happens, unless a fatal error occurs.<\/p><p>For example:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">$ php -a\nphp &gt; echo $foo;\nNotice: Undefined variable: foo in php shell code on line 1<\/pre><\/figure><p>This is only a notice error, and PHP will happily carry on. This can be confusing for those coming from\n\u201cexception-heavy\u201d languages, because referencing a missing variable in Python for example will throw an exception:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">$ python\n&gt;&gt;&gt; print foo\nTraceback (most recent call last):\n  File \"&lt;stdin&gt;\", line 1, in &lt;module&gt;\nNameError: name 'foo' is not defined<\/pre><\/figure><p>The only real difference is that Python will freak out over any small thing, so that developers can be super sure any\npotential issue or edge-case is caught, whereas PHP will keep on processing unless something extreme happens, at which\npoint it will throw an error and report it.<\/p><h3 id=\"error-severity\">Error Severity<\/h3><p>PHP has several levels of error severity. The three most common types of messages are errors, notices and warnings.\nThese have different levels of severity; <code class=\"highlighter-rouge\">E_ERROR<\/code>, <code class=\"highlighter-rouge\">E_NOTICE<\/code>, and <code class=\"highlighter-rouge\">E_WARNING<\/code>. Errors are fatal run-time errors and\nare usually caused by faults in your code and need to be fixed as they\u2019ll cause PHP to stop executing. Notices are\nadvisory messages caused by code that may or may not cause problems during the execution of the script, execution is\nnot halted. Warnings are non-fatal errors, execution of the script will not be halted.<\/p><p>Another type of error message reported at compile time are <code class=\"highlighter-rouge\">E_STRICT<\/code> messages. These messages are used to suggest\nchanges to your code to help ensure best interoperability and forward compatibility with upcoming versions of PHP.<\/p><h3 id=\"changing-phps-error-reporting-behaviour\">Changing PHP\u2019s Error Reporting Behaviour<\/h3><p>Error Reporting can be changed by using PHP settings and\/or PHP function calls. Using the built in PHP function\n<code class=\"highlighter-rouge\">error_reporting()<\/code> you can set the level of errors for the duration of the script execution by passing one of the\npredefined error level constants, meaning if you only want to see Errors and Warnings - but not Notices - then you can\nconfigure that:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nerror_reporting(E_ERROR | E_WARNING);<\/pre><\/figure><p>You can also control whether or not errors are displayed to the screen (good for development) or hidden, and logged\n(good for production). For more information on this check out the <a href=\"https:\/\/phptherightway.com\/#error_reporting\">Error Reporting<\/a> section.<\/p><h3 id=\"inline-error-suppression\">Inline Error Suppression<\/h3><p>You can also tell PHP to suppress specific errors with the Error Control Operator <code class=\"highlighter-rouge\">@<\/code>. You put this operator at the\nbeginning of an expression, and any error that\u2019s a direct result of the expression is silenced.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\necho @$foo['bar'];<\/pre><\/figure><p>This will output <code class=\"highlighter-rouge\">$foo['bar']<\/code> if it exists, but will simply return a null and print nothing if the variable <code class=\"highlighter-rouge\">$foo<\/code> or\n<code class=\"highlighter-rouge\">'bar'<\/code> key does not exist. Without the error control operator, this expression could create a <code class=\"highlighter-rouge\">PHP Notice: Undefined\nvariable: foo<\/code> or <code class=\"highlighter-rouge\">PHP Notice: Undefined index: bar<\/code> error.<\/p><p>This might seem like a good idea, but there are a few undesirable tradeoffs. PHP handles expressions using an <code class=\"highlighter-rouge\">@<\/code> in a\nless performant way than expressions without an <code class=\"highlighter-rouge\">@<\/code>. Premature optimization may be the root of all programming\narguments, but if performance is particularly important for your application\/library it\u2019s important to understand the\nerror control operator\u2019s performance implications.<\/p><p>Secondly, the error control operator <strong>completely<\/strong> swallows the error. The error is not displayed, and the error is\nnot sent to the error log. Also, stock\/production PHP systems have no way to turn off the error control operator. While\nyou may be correct that the error you\u2019re seeing is harmless, a different, less harmless error will be just as silent.<\/p><p>If there\u2019s a way to avoid the error suppression operator, you should consider it. For example, our code above could be\nrewritten like this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/Null Coalescing Operator\necho $foo['bar'] ?? '';<\/pre><\/figure><p>One instance where error suppression might make sense is where <code class=\"highlighter-rouge\">fopen()<\/code> fails to find a file to load. You could check\nfor the existence of the file before you try to load it, but if the file is deleted after the check and before the\n<code class=\"highlighter-rouge\">fopen()<\/code> (which might sound impossible, but it can happen) then <code class=\"highlighter-rouge\">fopen()<\/code> will return false <em>and<\/em> throw an error. This\nis potentially something PHP should resolve, but is one case where error suppression might seem like the only valid\nsolution.<\/p><p>Earlier we mentioned there\u2019s no way in a stock PHP system to turn off the error control operator. However, <a href=\"https:\/\/xdebug.org\/docs\/basic\">Xdebug<\/a> has\nan <code class=\"highlighter-rouge\">xdebug.scream<\/code> ini setting which will disable the error control operator. You can set this via your <code class=\"highlighter-rouge\">php.ini<\/code> file\nwith the following.<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">xdebug.scream = On<\/pre><\/figure><p>You can also set this value at runtime with the <code class=\"highlighter-rouge\">ini_set<\/code> function<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nini_set('xdebug.scream', '1')<\/pre><\/figure><p>The \u201c<a href=\"https:\/\/secure.php.net\/book.scream\">Scream<\/a>\u201d PHP extension offers similar functionality to Xdebug\u2019s, although Scream\u2019s ini setting is named\n<code class=\"highlighter-rouge\">scream.enabled<\/code>.<\/p><p>This is most useful when you\u2019re debugging code and suspect an informative error is suppressed. Use scream with care,\nand as a temporary debugging tool. There\u2019s lots of PHP library code that may not work with the error control operator\ndisabled.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.operators.errorcontrol\">Error Control Operators<\/a><\/li>\n  <li><a href=\"https:\/\/www.sitepoint.com\/\">SitePoint<\/a><\/li>\n  <li><a href=\"https:\/\/xdebug.org\/docs\/basic\">Xdebug<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/book.scream\">Scream<\/a><\/li>\n<\/ul><h3 id=\"errorexception\">ErrorException<\/h3><p>PHP is perfectly capable of being an \u201cexception-heavy\u201d programming language, and only requires a few lines of code to\nmake the switch. Basically you can throw your \u201cerrors\u201d as \u201cexceptions\u201d using the <code class=\"highlighter-rouge\">ErrorException<\/code> class, which extends\nthe <code class=\"highlighter-rouge\">Exception<\/code> class.<\/p><p>This is a common practice implemented by a large number of modern frameworks such as Symfony and Laravel. In debug\nmode <em>(or dev mode)<\/em> both of these frameworks will display a nice and clean <em>stack trace<\/em>.<\/p><p>There are also some packages available for better error and exception handling and reporting. Like <a href=\"https:\/\/filp.github.io\/whoops\/\">Whoops!<\/a>, which comes with the default installation of Laravel and can be used in any framework as well.<\/p><p>By throwing errors as exceptions in development you can handle them better than the usual result, and if you see an\nexception during development you can wrap it in a catch statement with specific instructions on how to handle the\nsituation. Each exception you catch instantly makes your application that little bit more robust.<\/p><p>More information on this and details on how to use <code class=\"highlighter-rouge\">ErrorException<\/code> with error handling can be found at\n<a href=\"https:\/\/secure.php.net\/class.errorexception\">ErrorException Class<\/a>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.operators.errorcontrol\">Error Control Operators<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.constants\">Predefined Constants for Error Handling<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.error-reporting\"><code class=\"highlighter-rouge\">error_reporting()<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/phptherightway.com\/#error_reporting\">Reporting<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"exceptions\"><h2 id=\"exceptions_title\">Exceptions<\/h2><p>Exceptions are a standard part of most popular programming languages, but they are often overlooked by PHP programmers.\nLanguages like Ruby are extremely Exception heavy, so whenever something goes wrong such as a HTTP request failing, or\na DB query goes wrong, or even if an image asset could not be found, Ruby (or the gems being used) will throw an\nexception to the screen meaning you instantly know there is a mistake.<\/p><p>PHP itself is fairly lax with this, and a call to <code class=\"highlighter-rouge\">file_get_contents()<\/code> will usually just get you a <code class=\"highlighter-rouge\">FALSE<\/code> and a\nwarning.\nMany older PHP frameworks like CodeIgniter will just return a false, log a message to their proprietary logs and maybe\nlet you use a method like <code class=\"highlighter-rouge\">$this-&gt;upload-&gt;get_error()<\/code> to see what went wrong. The problem here is that you have to go\nlooking for a mistake and check the docs to see what the error method is for this class, instead of having it made\nextremely obvious.<\/p><p>Another problem is when classes automatically throw an error to the screen and exit the process. When you do this you\nstop another developer from being able to dynamically handle that error. Exceptions should be thrown to make a\ndeveloper aware of an error; they then can choose how to handle this. E.g.:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$email = new Fuel\\Email;\n$email-&gt;subject('My Subject');\n$email-&gt;body('How the heck are you?');\n$email-&gt;to('guy@example.com', 'Some Guy');\ntry\n{\n    $email-&gt;send();\n}\ncatch(Fuel\\Email\\ValidationFailedException $e)\n{\n    \/\/ The validation failed\n}\ncatch(Fuel\\Email\\SendingFailedException $e)\n{\n    \/\/ The driver could not send the email\n}\nfinally\n{\n    \/\/ Executed regardless of whether an exception has been thrown, and before normal execution resumes\n}<\/pre><\/figure><h3 id=\"spl-exceptions\">SPL Exceptions<\/h3><p>The generic <code class=\"highlighter-rouge\">Exception<\/code> class provides very little debugging context for the developer; however, to remedy this, it is\npossible to create a specialized <code class=\"highlighter-rouge\">Exception<\/code> type by sub-classing the generic <code class=\"highlighter-rouge\">Exception<\/code> class:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nclass ValidationException extends Exception {}<\/pre><\/figure><p>This means you can add multiple catch blocks and handle different Exceptions differently. This can lead to the\ncreation of a <em>lot<\/em> of custom Exceptions, some of which could have been avoided using the SPL Exceptions\nprovided in the <a href=\"https:\/\/phptherightway.com\/#standard_php_library\">SPL extension<\/a>.<\/p><p>If for example you use the <code class=\"highlighter-rouge\">__call()<\/code> Magic Method and an invalid method is requested then instead of throwing a\nstandard Exception which is vague, or creating a custom Exception just for that, you could just\n<code class=\"highlighter-rouge\">throw new BadMethodCallException;<\/code>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.exceptions\">Read about Exceptions<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/spl.exceptions\">Read about SPL Exceptions<\/a><\/li>\n  <li><a href=\"https:\/\/www.brandonsavage.net\/exceptional-php-nesting-exceptions-in-php\/\">Nesting Exceptions In PHP<\/a><\/li>\n  <li><a href=\"http:\/\/ralphschindler.com\/2010\/09\/15\/exception-best-practices-in-php-5-3\">Exception Best Practices in PHP 5.3<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"security\"><p>The best resource I\u2019ve found on PHP security is <a href=\"https:\/\/paragonie.com\/blog\/2017\/12\/2018-guide-building-secure-php-software\">The 2018 Guide to Building Secure PHP Software<\/a> by\n<a href=\"https:\/\/paragonie.com\/\">Paragon Initiative<\/a>.<\/p><\/section><section class=\"chapter\" id=\"web_application_security\"><h2 id=\"web_application_security_title\">Web Application Security<\/h2><p>It is very important for every PHP developer to learn <a href=\"https:\/\/paragonie.com\/blog\/2015\/08\/gentle-introduction-application-security\">the basics of web application security<\/a>, which can be broken\ndown into a handful of broad topics:<\/p><ol><li>Code-data separation.\n    <ul><li>When data is executed as code, you get SQL Injection, Cross-Site Scripting, Local\/Remote File Inclusion, etc.<\/li>\n      <li>When code is printed as data, you get information leaks (source code disclosure or, in the case of C programs,\nenough information to bypass <a href=\"http:\/\/searchsecurity.techtarget.com\/definition\/address-space-layout-randomization-ASLR\">ASLR<\/a>).<\/li>\n    <\/ul><\/li>\n  <li>Application logic.\n    <ul><li>Missing authentication or authorization controls.<\/li>\n      <li>Input validation.<\/li>\n    <\/ul><\/li>\n  <li>Operating environment.\n    <ul><li>PHP versions.<\/li>\n      <li>Third party libraries.<\/li>\n      <li>The operating system.<\/li>\n    <\/ul><\/li>\n  <li>Cryptography weaknesses.\n    <ul><li><a href=\"https:\/\/paragonie.com\/blog\/2016\/01\/on-design-and-implementation-stealth-backdoor-for-web-applications\">Weak random numbers<\/a>.<\/li>\n      <li><a href=\"https:\/\/paragonie.com\/blog\/2015\/05\/using-encryption-and-authentication-correctly\">Chosen-ciphertext attacks<\/a>.<\/li>\n      <li><a href=\"http:\/\/blog.ircmaxell.com\/2014\/11\/its-all-about-time.html\">Side-channel information leaks<\/a>.<\/li>\n    <\/ul><\/li>\n<\/ol><p>There are bad people ready and willing to exploit your web application. It is important that you take necessary\nprecautions to harden your web application\u2019s security. Luckily, the fine folks at\n<a href=\"https:\/\/www.owasp.org\/\">The Open Web Application Security Project<\/a> (OWASP) have compiled a comprehensive list of known security issues and\nmethods to protect yourself against them. This is a must read for the security-conscious developer. <a href=\"https:\/\/phpsecurity.readthedocs.io\/en\/latest\/index.html\">Survive The Deep End: PHP Security<\/a> by Padraic Brady is also another good web application security guide for PHP.<\/p><ul><li><a href=\"https:\/\/www.owasp.org\/index.php\/Guide_Table_of_Contents\">Read the OWASP Security Guide<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"password_hashing\"><h2 id=\"password_hashing_title\">Password Hashing<\/h2><p>Eventually everyone builds a PHP application that relies on user login. Usernames and passwords are stored in a\ndatabase and later used to authenticate users upon login.<\/p><p>It is important that you properly <a href=\"https:\/\/wikipedia.org\/wiki\/Cryptographic_hash_function\"><em>hash<\/em><\/a> passwords before storing them. Hashing and encrypting are <a href=\"https:\/\/paragonie.com\/blog\/2015\/08\/you-wouldnt-base64-a-password-cryptography-decoded\">two very different things<\/a>\nthat often get confused.<\/p><p>Hashing is an irreversible, one-way function. This produces a fixed-length string that cannot be feasibly reversed.\nThis means you can compare a hash against another to determine if they both came from the same source string, but you\ncannot determine the original string. If passwords are not hashed and your database is accessed by an unauthorized\nthird-party, all user accounts are now compromised.<\/p><p>Unlike hashing, encryption is reversible (provided you have the key). Encryption is useful in other areas, but is a poor\nstrategy for securely storing passwords.<\/p><p>Passwords should also be individually <a href=\"https:\/\/wikipedia.org\/wiki\/Salt_(cryptography)\"><em>salted<\/em><\/a> by adding a random string to each password before hashing. This prevents dictionary attacks and the use of \u201crainbow tables\u201d (a reverse list of crytographic hashes for common passwords.)<\/p><p>Hashing and salting are vital as often users use the same password for multiple services and password quality can be poor.<\/p><p>Additionally, you should use <a href=\"https:\/\/paragonie.com\/blog\/2016\/02\/how-safely-store-password-in-2016\">a specialized <em>password hashing<\/em> algoithm<\/a> rather than fast, general-purpose\ncryptographic hash function (e.g. SHA256). The short list of acceptable password hashing algorithms (as of June 2018)\nto use are:<\/p><ul><li>Argon2 (available in PHP 7.2 and newer)<\/li>\n  <li>Scrypt<\/li>\n  <li><strong>Bcrypt<\/strong> (PHP provides this one for you; see below)<\/li>\n  <li>PBKDF2 with HMAC-SHA256 or HMAC-SHA512<\/li>\n<\/ul><p>Fortunately, nowadays PHP makes this easy.<\/p><p><strong>Hashing passwords with <code class=\"highlighter-rouge\">password_hash<\/code><\/strong><\/p><p>In PHP 5.5 <code class=\"highlighter-rouge\">password_hash()<\/code> was introduced. At this time it is using BCrypt, the strongest algorithm currently\nsupported by PHP. It will be updated in the future to support more algorithms as needed though. The <code class=\"highlighter-rouge\">password_compat<\/code>\nlibrary was created to provide forward compatibility for PHP &gt;= 5.3.7.<\/p><p>Below we hash a string, and then check the hash against a new string. Because our two source strings are different\n(\u2018secret-password\u2019 vs. \u2018bad-password\u2019) this login will fail.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nrequire 'password.php';\n$passwordHash = password_hash('secret-password', PASSWORD_DEFAULT);\nif (password_verify('bad-password', $passwordHash)) {\n    \/\/ Correct Password\n} else {\n    \/\/ Wrong password\n}<\/pre><\/figure><p><code class=\"highlighter-rouge\">password_hash()<\/code> takes care of password salting for you. The salt is stored, along with the algorithm and \u201ccost\u201d, as part of the hash.  <code class=\"highlighter-rouge\">password_verify()<\/code> extracts this to determine how to check the password, so you don\u2019t need a separate database field to store your salts.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/function.password-hash\">Learn about <code class=\"highlighter-rouge\">password_hash()<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/ircmaxell\/password_compat\"><code class=\"highlighter-rouge\">password_compat<\/code> for PHP &gt;= 5.3.7 &amp;&amp; &lt; 5.5<\/a><\/li>\n  <li><a href=\"https:\/\/wikipedia.org\/wiki\/Cryptographic_hash_function\">Learn about hashing in regards to cryptography<\/a><\/li>\n  <li><a href=\"https:\/\/wikipedia.org\/wiki\/Salt_(cryptography)\">Learn about salts<\/a><\/li>\n  <li><a href=\"https:\/\/wiki.php.net\/rfc\/password_hash\">PHP <code class=\"highlighter-rouge\">password_hash()<\/code> RFC<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"data_filtering\"><h2 id=\"data_filtering_title\">Data Filtering<\/h2><p>Never ever (ever) trust foreign input introduced to your PHP code. Always sanitize and validate foreign input before\nusing it in code. The <code class=\"highlighter-rouge\">filter_var()<\/code> and <code class=\"highlighter-rouge\">filter_input()<\/code> functions can sanitize text and validate text formats (e.g.\nemail addresses).<\/p><p>Foreign input can be anything: <code class=\"highlighter-rouge\">$_GET<\/code> and <code class=\"highlighter-rouge\">$_POST<\/code> form input data, some values in the <code class=\"highlighter-rouge\">$_SERVER<\/code> superglobal, and the\nHTTP request body via <code class=\"highlighter-rouge\">fopen('php:\/\/input', 'r')<\/code>. Remember, foreign input is not limited to form data submitted by the\nuser. Uploaded and downloaded files, session values, cookie data, and data from third-party web services are foreign\ninput, too.<\/p><p>While foreign data can be stored, combined, and accessed later, it is still foreign input. Every time you process,\noutput, concatenate, or include data in your code, ask yourself if the data is filtered properly and can it be trusted.<\/p><p>Data may be <em>filtered<\/em> differently based on its purpose. For example, when unfiltered foreign input is passed into HTML\npage output, it can execute HTML and JavaScript on your site! This is known as Cross-Site Scripting (XSS) and can be a\nvery dangerous attack. One way to avoid XSS is to sanitize all user-generated data before outputting it to your page by\nremoving HTML tags with the <code class=\"highlighter-rouge\">strip_tags()<\/code> function or escaping characters with special meaning into their respective\nHTML entities with the <code class=\"highlighter-rouge\">htmlentities()<\/code> or <code class=\"highlighter-rouge\">htmlspecialchars()<\/code> functions.<\/p><p>Another example is passing options to be executed on the command line. This can be extremely dangerous (and is usually\na bad idea), but you can use the built-in <code class=\"highlighter-rouge\">escapeshellarg()<\/code> function to sanitize the executed command\u2019s arguments.<\/p><p>One last example is accepting foreign input to determine a file to load from the filesystem. This can be exploited by\nchanging the filename to a file path. You need to remove <code class=\"highlighter-rouge\">\"\/\"<\/code>, <code class=\"highlighter-rouge\">\"..\/\"<\/code>, <a href=\"https:\/\/secure.php.net\/security.filesystem.nullbytes\">null bytes<\/a>, or other characters from the\nfile path so it can\u2019t load hidden, non-public, or sensitive files.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.filter\">Learn about data filtering<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.filter-var\">Learn about <code class=\"highlighter-rouge\">filter_var<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.filter-input\">Learn about <code class=\"highlighter-rouge\">filter_input<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/security.filesystem.nullbytes\">Learn about handling null bytes<\/a><\/li>\n<\/ul><h3 id=\"sanitization\">Sanitization<\/h3><p>Sanitization removes (or escapes) illegal or unsafe characters from foreign input.<\/p><p>For example, you should sanitize foreign input before including the input in HTML or inserting it into a raw SQL query.\nWhen you use bound parameters with <a href=\"#databases\">PDO<\/a>, it will sanitize the input for you.<\/p><p>Sometimes it is required to allow some safe HTML tags in the input when including it in the HTML page. This is very\nhard to do and many avoid it by using other more restricted formatting like Markdown or BBCode, although whitelisting\nlibraries like <a href=\"http:\/\/htmlpurifier.org\/\">HTML Purifier<\/a> exists for this reason.<\/p><p><a href=\"https:\/\/secure.php.net\/filter.filters.sanitize\">See Sanitization Filters<\/a><\/p><h3 id=\"unserialization\">Unserialization<\/h3><p>It is dangerous to <code class=\"highlighter-rouge\">unserialize()<\/code> data from users or other untrusted sources.  Doing so can allow malicious users to instantiate objects (with user-defined properties) whose destructors will be executed, <strong>even if the objects themselves aren\u2019t used<\/strong>.  You should therefore avoid unserializing untrusted data.<\/p><p>If you absolutely must unserialize data from untrusted sources, use PHP 7\u2019s <a href=\"https:\/\/secure.php.net\/manual\/function.unserialize.php\"><code class=\"highlighter-rouge\">allowed_classes<\/code><\/a> option to restrict which object types are allowed to be unserialized.<\/p><h3 id=\"validation\">Validation<\/h3><p>Validation ensures that foreign input is what you expect. For example, you may want to validate an email address, a\nphone number, or age when processing a registration submission.<\/p><p><a href=\"https:\/\/secure.php.net\/filter.filters.validate\">See Validation Filters<\/a><\/p><\/section><section class=\"chapter\" id=\"configuration_files\"><h2 id=\"configuration_files_title\">Configuration Files<\/h2><p>When creating configuration files for your applications, best practices recommend that one of the following methods be\nfollowed:<\/p><ul><li>It is recommended that you store your configuration information where it cannot be accessed directly and pulled in\nvia the file system.<\/li>\n  <li>If you must store your configuration files in the document root, name the files with a <code class=\"highlighter-rouge\">.php<\/code> extension. This ensures\nthat, even if the script is accessed directly, it will not be output as plain text.<\/li>\n  <li>Information in configuration files should be protected accordingly, either through encryption or group\/user file\nsystem permissions.<\/li>\n  <li>It is a good idea to ensure that you do not commit configuration files containing sensitive information e.g. passwords or API tokens to source control.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"register_globals\"><h2 id=\"register_globals_title\">Register Globals<\/h2><p><strong>NOTE:<\/strong> As of PHP 5.4.0 the <code class=\"highlighter-rouge\">register_globals<\/code> setting has been removed and can no longer be used. This is only\nincluded as a warning for anyone in the process of upgrading a legacy application.<\/p><p>When enabled, the <code class=\"highlighter-rouge\">register_globals<\/code> configuration setting that makes several types of variables (including ones from\n<code class=\"highlighter-rouge\">$_POST<\/code>, <code class=\"highlighter-rouge\">$_GET<\/code> and <code class=\"highlighter-rouge\">$_REQUEST<\/code>) available in the global scope of your application. This can easily lead to security\nissues as your application cannot effectively tell where the data is coming from.<\/p><p>For example: <code class=\"highlighter-rouge\">$_GET['foo']<\/code> would be available via <code class=\"highlighter-rouge\">$foo<\/code>, which can override variables that have not been declared.\nIf you are using PHP &lt; 5.4.0 <strong>make sure<\/strong> that <code class=\"highlighter-rouge\">register_globals<\/code> is <strong>off<\/strong>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/security.globals\">Register_globals in the PHP manual<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"error_reporting\"><h2 id=\"error_reporting_title\">Error Reporting<\/h2><p>Error logging can be useful in finding the problem spots in your application, but it can also expose information about\nthe structure of your application to the outside world. To effectively protect your application from issues that could\nbe caused by the output of these messages, you need to configure your server differently in development versus\nproduction (live).<\/p><h3 id=\"development\">Development<\/h3><p>To show every possible error during <strong>development<\/strong>, configure the following settings in your <code class=\"highlighter-rouge\">php.ini<\/code>:<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">display_errors = On\ndisplay_startup_errors = On\nerror_reporting = -1\nlog_errors = On<\/pre><\/figure><blockquote>\n  <p>Passing in the value <code class=\"highlighter-rouge\">-1<\/code> will show every possible error, even when new levels and constants are added in future PHP\nversions. The <code class=\"highlighter-rouge\">E_ALL<\/code> constant also behaves this way as of PHP 5.4. -\n<a href=\"https:\/\/secure.php.net\/function.error-reporting\">php.net<\/a><\/p>\n<\/blockquote><p>The <code class=\"highlighter-rouge\">E_STRICT<\/code> error level constant was introduced in 5.3.0 and is not part of <code class=\"highlighter-rouge\">E_ALL<\/code>, however it became part of\n<code class=\"highlighter-rouge\">E_ALL<\/code> in 5.4.0. What does this mean? In terms of reporting every possible error in version 5.3 it means you must\nuse either <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL | E_STRICT<\/code>.<\/p><p><strong>Reporting every possible error by PHP version<\/strong><\/p><ul><li>&lt; 5.3 <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL<\/code><\/li>\n  <li>\u00a0 5.3 <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL | E_STRICT<\/code><\/li>\n  <li>&gt; 5.3 <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL<\/code><\/li>\n<\/ul><h3 id=\"production\">Production<\/h3><p>To hide errors on your <strong>production<\/strong> environment, configure your <code class=\"highlighter-rouge\">php.ini<\/code> as:<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">display_errors = Off\ndisplay_startup_errors = Off\nerror_reporting = E_ALL\nlog_errors = On<\/pre><\/figure><p>With these settings in production, errors will still be logged to the error logs for the web server, but will not be\nshown to the user. For more information on these settings, see the PHP manual:<\/p><ul><li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.error-reporting\">error_reporting<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.display-errors\">display_errors<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.display-startup-errors\">display_startup_errors<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.log-errors\">log_errors<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"testing\"><p>Writing automated tests for your PHP code is considered a best practice and can lead to well-built applications.\nAutomated tests are a great tool for making sure your application does not break when you are making changes or adding\nnew functionality and should not be ignored.<\/p><p>There are several different types of testing tools (or frameworks) available for PHP, which use different approaches -\nall of which are trying to avoid manual testing and the need for large Quality Assurance teams, just to make sure\nrecent changes didn\u2019t break existing functionality.<\/p><\/section><section class=\"chapter\" id=\"test_driven_development\"><h2 id=\"test_driven_development_title\">Test Driven Development<\/h2><p>From <a href=\"https:\/\/wikipedia.org\/wiki\/Test-driven_development\">Wikipedia<\/a>:<\/p><blockquote>\n  <p>Test-driven development (TDD) is a software development process that relies on the repetition of a very short\ndevelopment cycle: first the developer writes a failing automated test case that defines a desired improvement or new\nfunction, then produces code to pass that test and finally refactors the new code to acceptable standards. Kent Beck,\nwho is credited with having developed or \u2018rediscovered\u2019 the technique, stated in 2003 that TDD encourages simple\ndesigns and inspires confidence.<\/p>\n<\/blockquote><p>There are several different types of testing that you can do for your application:<\/p><h3 id=\"unit-testing\">Unit Testing<\/h3><p>Unit Testing is a programming approach to ensure functions, classes and methods are working as expected, from the point\nyou build them all the way through the development cycle. By checking values going in and out of various functions and\nmethods, you can make sure the internal logic is working correctly. By using Dependency Injection and building \u201cmock\u201d\nclasses and stubs you can verify that dependencies are correctly used for even better test coverage.<\/p><p>When you create a class or function you should create a unit test for each behavior it must have. At a very basic level\nyou should make sure it errors if you send it bad arguments and make sure it works if you send it valid arguments. This\nwill help ensure that when you make changes to this class or function later on in the development cycle that the old\nfunctionality continues to work as expected. The only alternative to this would be <code class=\"highlighter-rouge\">var_dump()<\/code> in a test.php, which is\nno way to build an application - large or small.<\/p><p>The other use for unit tests is contributing to open source. If you can write a test that shows broken functionality\n(i.e. fails), then fix it, and show the test passing, patches are much more likely to be accepted. If you run a project\nwhich accepts pull requests then you should suggest this as a requirement.<\/p><p><a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a> is the de-facto testing framework for writing unit tests for PHP applications, but there\nare several alternatives<\/p><ul><li><a href=\"https:\/\/github.com\/atoum\/atoum\">atoum<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/crysalead\/kahlan\">Kahlan<\/a><\/li>\n  <li><a href=\"https:\/\/peridot-php.github.io\/\">Peridot<\/a><\/li>\n  <li><a href=\"http:\/\/simpletest.org\">SimpleTest<\/a><\/li>\n<\/ul><h3 id=\"integration-testing\">Integration Testing<\/h3><p>From <a href=\"https:\/\/wikipedia.org\/wiki\/Integration_testing\">Wikipedia<\/a>:<\/p><blockquote>\n  <p>Integration testing (sometimes called Integration and Testing, abbreviated \u201cI&amp;T\u201d) is the phase in software testing in\nwhich individual software modules are combined and tested as a group. It occurs after unit testing and before\nvalidation testing. Integration testing takes as its input modules that have been unit tested, groups them in larger\naggregates, applies tests defined in an integration test plan to those aggregates, and delivers as its output the\nintegrated system ready for system testing.<\/p>\n<\/blockquote><p>Many of the same tools that can be used for unit testing can be used for integration testing as many of the same\nprinciples are used.<\/p><p>Sometimes also known as acceptance testing, functional testing consists of using tools to create automated tests that\nactually use your application instead of just verifying that individual units of code are behaving correctly and that\nindividual units can speak to each other correctly. These tools typically work using real data and simulating actual\nusers of the application.<\/p><ul><li><a href=\"https:\/\/docs.seleniumhq.org\/\">Selenium<\/a><\/li>\n  <li><a href=\"http:\/\/mink.behat.org\/\">Mink<\/a><\/li>\n  <li><a href=\"https:\/\/codeception.com\/\">Codeception<\/a> is a full-stack testing framework that includes acceptance testing tools<\/li>\n  <li><a href=\"https:\/\/datasift.github.io\/storyplayer\/\">Storyplayer<\/a> is a full-stack testing framework that includes support for creating and destroying test environments on demand<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"behavior_driven_development\"><h2 id=\"behavior_driven_development_title\">Behavior Driven Development<\/h2><p>There are two different types of Behavior-Driven Development (BDD): SpecBDD and StoryBDD. SpecBDD focuses on technical\nbehavior of code, while StoryBDD focuses on business or feature behaviors or interactions. PHP has frameworks for both\ntypes of BDD.<\/p><p>With StoryBDD, you write human-readable stories that describe the behavior of your application. These stories can then\nbe run as actual tests against your application. The framework used in PHP applications for StoryBDD is <a href=\"http:\/\/behat.org\/\">Behat<\/a>, which\nis inspired by Ruby\u2019s <a href=\"https:\/\/cucumber.io\/\">Cucumber<\/a> project and implements the Gherkin DSL for describing feature behavior.<\/p><p>With SpecBDD, you write specifications that describe how your actual code should behave. Instead of testing a function\nor method, you are describing how that function or method should behave. PHP offers the <a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a> framework for this\npurpose. This framework is inspired by the <a href=\"https:\/\/rspec.info\/\">RSpec project<\/a> for Ruby.<\/p><h3 id=\"bdd-links\">BDD Links<\/h3><ul><li><a href=\"http:\/\/behat.org\/\">Behat<\/a>, the StoryBDD framework for PHP, inspired by Ruby\u2019s <a href=\"https:\/\/cucumber.io\/\">Cucumber<\/a> project;<\/li>\n  <li><a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a>, the SpecBDD framework for PHP, inspired by Ruby\u2019s <a href=\"https:\/\/rspec.info\/\">RSpec<\/a> project;<\/li>\n  <li><a href=\"https:\/\/codeception.com\/\">Codeception<\/a> is a full-stack testing framework that uses BDD principles.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"complementary_testing_tools\"><h2 id=\"complementary_testing_tools_title\">Complementary Testing Tools<\/h2><p>Besides individual testing and behavior driven frameworks, there are also a number of generic frameworks and helper\nlibraries useful for any preferred approach taken.<\/p><ul><li><a href=\"https:\/\/www.seleniumhq.org\/\">Selenium<\/a> is a browser automation tool which can be <a href=\"https:\/\/github.com\/giorgiosironi\/phpunit-selenium\/\">integrated with PHPUnit<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/padraic\/mockery\">Mockery<\/a> is a Mock Object Framework which can be integrated with <a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a> or <a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/phpspec\/prophecy\">Prophecy<\/a> is a highly opinionated yet very powerful and flexible PHP object mocking framework. It\u2019s integrated with\n<a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a> and can be used with <a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a>.<\/li>\n  <li><a href=\"https:\/\/github.com\/php-mock\/php-mock\">php-mock<\/a>  is a library to help to mock PHP native functions.<\/li>\n  <li><a href=\"https:\/\/github.com\/infection\/infection\">Infection<\/a> is a PHP implementation of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mutation_testing\">Mutation Testing<\/a> to help to measure the effectiveness of your tests.<\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"servers_and_deployment\"><p>PHP applications can be deployed and run on production web servers in a number of ways.<\/p><\/section><section class=\"chapter\" id=\"platform_as_a_service\"><h2 id=\"platform_as_a_service_title\">Platform as a Service (PaaS)<\/h2><p>PaaS provides the system and network architecture necessary to run PHP applications on the web. This means little to no\nconfiguration for launching PHP applications or PHP frameworks.<\/p><p>Recently PaaS has become a popular method for deploying, hosting, and scaling PHP applications of all sizes. You can\nfind a list of <a href=\"#php_paas_providers\">PHP PaaS \u201cPlatform as a Service\u201d providers<\/a> in our <a href=\"#resources\">resources section<\/a>.<\/p><\/section><section class=\"chapter\" id=\"virtual_or_dedicated_servers\"><h2 id=\"virtual_or_dedicated_servers_title\">Virtual or Dedicated Servers<\/h2><p>If you are comfortable with systems administration, or are interested in learning it, virtual or dedicated servers give\nyou complete control of your application\u2019s production environment.<\/p><h3 id=\"nginx-and-php-fpm\">nginx and PHP-FPM<\/h3><p>PHP, via PHP\u2019s built-in FastCGI Process Manager (FPM), pairs really nicely with <a href=\"https:\/\/nginx.org\/\">nginx<\/a>, which is a lightweight,\nhigh-performance web server. It uses less memory than Apache and can better handle more concurrent requests. This is\nespecially important on virtual servers that don\u2019t have much memory to spare.<\/p><ul><li><a href=\"https:\/\/nginx.org\/\">Read more on nginx<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/install.fpm\">Read more on PHP-FPM<\/a><\/li>\n  <li><a href=\"https:\/\/nealpoole.com\/blog\/2011\/04\/setting-up-php-fastcgi-and-nginx-dont-trust-the-tutorials-check-your-configuration\/\">Read more on setting up nginx and PHP-FPM securely<\/a><\/li>\n<\/ul><h3 id=\"apache-and-php\">Apache and PHP<\/h3><p>PHP and Apache have a long history together. Apache is wildly configurable and has many available\n<a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/\">modules<\/a> to extend functionality. It is a popular choice for shared servers and an easy setup for PHP\nframeworks and open source apps like WordPress. Unfortunately, Apache uses more resources than nginx by default and\ncannot handle as many visitors at the same time.<\/p><p>Apache has several possible configurations for running PHP. The most common and easiest to setup is the <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/prefork.html\">prefork MPM<\/a>\nwith mod_php5. While it isn\u2019t the most memory efficient, it is the simplest to get working and to use. This is probably\nthe best choice if you don\u2019t want to dig too deeply into the server administration aspects. Note that if you use\nmod_php5 you MUST use the prefork MPM.<\/p><p>Alternatively, if you want to squeeze more performance and stability out of Apache then you can take advantage of the\nsame FPM system as nginx and run the <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/worker.html\">worker MPM<\/a> or <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/event.html\">event MPM<\/a> with mod_fastcgi or mod_fcgid. This configuration will\nbe significantly more memory efficient and much faster but it is more work to set up.<\/p><p>If you are running Apache 2.4 or later, you can use <a href=\"https:\/\/httpd.apache.org\/docs\/current\/mod\/mod_proxy_fcgi.html\">mod_proxy_fcgi<\/a> to get great performance that is easy to setup.<\/p><ul><li><a href=\"https:\/\/httpd.apache.org\/\">Read more on Apache<\/a><\/li>\n  <li><a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/mpm_common.html\">Read more on Multi-Processing Modules<\/a><\/li>\n  <li><a href=\"https:\/\/blogs.oracle.com\/opal\/entry\/php_fpm_fastcgi_process_manager\">Read more on mod_fastcgi<\/a><\/li>\n  <li><a href=\"denied:hhttps:\/\/httpd.apache.org\/mod_fcgid\/\">Read more on mod_fcgid<\/a><\/li>\n  <li><a href=\"https:\/\/httpd.apache.org\/docs\/current\/mod\/mod_proxy_fcgi.html\">Read more on mod_proxy_fcgi<\/a><\/li>\n  <li><a href=\"https:\/\/serversforhackers.com\/video\/apache-and-php-fpm\">Read more on setting up Apache and PHP-FPM with mod_proxy_fcgi<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"building_and_deploying_your_application\"><h2 id=\"building_and_deploying_your_application_title\">Building and Deploying your Application<\/h2><p>If you find yourself doing manual database schema changes or running your tests manually before updating your files\n(manually), think twice! With every additional manual task needed to deploy a new version of your app, the chances for\npotentially fatal mistakes increase. Whether you\u2019re dealing with a simple update, a comprehensive build process or even\na continuous integration strategy, <a href=\"https:\/\/wikipedia.org\/wiki\/Build_automation\">build automation<\/a> is your friend.<\/p><p>Among the tasks you might want to automate are:<\/p><ul><li>Dependency management<\/li>\n  <li>Compilation, minification of your assets<\/li>\n  <li>Running tests<\/li>\n  <li>Creation of documentation<\/li>\n  <li>Packaging<\/li>\n  <li>Deployment<\/li>\n<\/ul><h3 id=\"deployment-tools\">Deployment Tools<\/h3><p>Deployment tools can be described as a collection of scripts that handle common tasks of software deployment. The deployment tool is not a part of your software, it acts on your software from \u2018outside\u2019.<\/p><p>There are many open source tools available to help you with build automation and deployment, some are written in PHP others aren\u2019t. This shouldn\u2019t hold you back from using them, if they\u2019re better suited for the specific job. Here are a few examples:<\/p><p><a href=\"https:\/\/www.phing.info\/\">Phing<\/a> can control your packaging, deployment or testing process from within a XML build file. Phing (which is based on <a href=\"https:\/\/ant.apache.org\/\">Apache Ant<\/a>) provides a rich set of tasks usually needed to install or update a web application and can be extended with additional custom tasks, written in PHP. It\u2019s a solid and robust tool and has been around for a long time, however the tool could be perceived as a bit old fashioned because of the way it deals with configuration (XML files).<\/p><p><a href=\"http:\/\/capistranorb.com\/\">Capistrano<\/a> is a system for <em>intermediate-to-advanced programmers<\/em> to execute commands in a structured, repeatable way on one or more remote machines. It is pre-configured for deploying Ruby on Rails applications, however you can successfully deploy PHP systems with it. Successful use of Capistrano depends on a working knowledge of Ruby and Rake.<\/p><p><a href=\"https:\/\/ansistrano.com\">Ansistrano<\/a> is a couple of Ansible roles to easily manage the deployment process (deploy and rollback) for scripting applications such as PHP, Python and Ruby. It\u2019s an Ansible port for <a href=\"http:\/\/capistranorb.com\/\">Capistrano<\/a>. It\u2019s been used by quite a lot of PHP companies already.<\/p><p><a href=\"http:\/\/rocketeer.autopergamene.eu\/\">Rocketeer<\/a> gets its inspiration and philosophy from the Laravel framework. Its goal is to be fast, elegant and easy to use with smart defaults. It features multiple servers, multiple stages, atomic deploys and deployment can be performed in parallel. Everything in the tool can be hot swapped or extended, and everything is written in PHP.<\/p><p><a href=\"https:\/\/deployer.org\/\">Deployer<\/a> is a deployment tool written in PHP. It\u2019s simple and functional. Features include running tasks in parallel, atomic deployment and keeping consistency between servers. Recipes of common tasks for Symfony, Laravel, Zend Framework and Yii are available. Younes Rafie\u2019s article  <a href=\"https:\/\/www.sitepoint.com\/deploying-php-applications-with-deployer\/\">Easy Deployment of PHP Applications with Deployer<\/a> is a great tutorial for deploying your application with the tool.<\/p><p><a href=\"https:\/\/www.magephp.com\/\">Magallanes<\/a> is another tool written in PHP with simple configuration done in YAML files. It has support for multiple servers and environments, atomic deployment, and has some built in tasks that you can leverage for common tools and frameworks.<\/p><h4>Further reading:<\/h4><ul><li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/automate-your-projects-with-apache-ant--net-18595\">Automate your project with Apache Ant<\/a><\/li>\n  <li><a href=\"https:\/\/deployingphpapplications.com\/\">Deploying PHP Applications<\/a> - paid book on best practices and tools for PHP deployment.<\/li>\n<\/ul><h3 id=\"server-provisioning\">Server Provisioning<\/h3><p>Managing and configuring servers can be a daunting task when faced with many servers. There are tools for dealing with this so you can automate your infrastructure to make sure you have the right servers and that they\u2019re configured properly. They often integrate with the larger cloud hosting providers (Amazon Web Services, Heroku, DigitalOcean, etc) for managing instances, which makes scaling an application a lot easier.<\/p><p><a href=\"https:\/\/www.ansible.com\/\">Ansible<\/a> is a tool that manages your infrastructure through YAML files. It\u2019s simple to get started with and can manage complex and large scale applications. There is an API for managing cloud instances and it can manage them through a dynamic inventory using certain tools.<\/p><p><a href=\"https:\/\/puppet.com\/\">Puppet<\/a> is a tool that has its own language and file types for managing servers and configurations. It can be used in a master\/client setup or it can be used in a \u201cmaster-less\u201d mode. In the master\/client mode the clients will poll the central master(s) for new configuration on set intervals and update itself if necessary. In the master-less mode you can push changes to your nodes.<\/p><p><a href=\"https:\/\/www.chef.io\/\">Chef<\/a> is a powerful Ruby based system integration framework that you can build your whole server environment or virtual boxes with. It integrates well with Amazon Web Services through their service called OpsWorks.<\/p><h4 id=\"further-reading-1\">Further reading:<\/h4><ul><li><a href=\"https:\/\/serversforhackers.com\/an-ansible-tutorial\">An Ansible Tutorial<\/a><\/li>\n  <li><a href=\"https:\/\/leanpub.com\/ansible-for-devops\">Ansible for DevOps<\/a> - paid book on everything Ansible<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/ansible-for-aws\">Ansible for AWS<\/a> - paid book on integrating Ansible and Amazon Web Services<\/li>\n  <li><a href=\"http:\/\/www.jasongrimes.org\/2012\/06\/managing-lamp-environments-with-chef-vagrant-and-ec2-1-of-3\/\">Three part blog series about deploying a LAMP application with Chef, Vagrant, and EC2<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/chef-cookbooks\/php\">Chef Cookbook which installs and configures PHP and the PEAR package management system<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/playlist?list=PL11cZfNdwNyPnZA9D1MbVqldGuOWqbumZ\">Chef video tutorial series<\/a><\/li>\n<\/ul><h3 id=\"continuous-integration\">Continuous Integration<\/h3><blockquote>\n  <p>Continuous Integration is a software development practice where members of a team integrate their work frequently,\nusually each person integrates at least daily \u2014 leading to multiple integrations per day. Many teams find that this\napproach leads to significantly reduced integration problems and allows a team to develop cohesive software more\nrapidly.<\/p>\n<\/blockquote><p><em>\u2013 Martin Fowler<\/em><\/p><p>There are different ways to implement continuous integration for PHP. <a href=\"https:\/\/travis-ci.org\/\">Travis CI<\/a> has done a great job of\nmaking continuous integration a reality even for small projects. Travis CI is a hosted continuous integration service\nfor the open source community. It is integrated with GitHub and offers first class support for many languages including\nPHP.<\/p><h4 id=\"further-reading-2\">Further reading:<\/h4><ul><li><a href=\"https:\/\/jenkins.io\/\">Continuous Integration with Jenkins<\/a><\/li>\n  <li><a href=\"https:\/\/www.phptesting.org\/\">Continuous Integration with PHPCI<\/a><\/li>\n  <li><a href=\"https:\/\/www.jetbrains.com\/teamcity\/\">Continuous Integration with Teamcity<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"virtualization\"><p>Running your application on different environments in development and production can lead to strange bugs popping up\nwhen you go live. It\u2019s also tricky to keep different development environments up to date with the same version for all\nlibraries used when working with a team of developers.<\/p><p>If you are developing on Windows and deploying to Linux (or anything non-Windows) or are developing in a team, you\nshould consider using a virtual machine. This sounds tricky, but besides the widely known virtualization environments\nlike VMware or VirtualBox, there are additional tools that may help you setting up a virtual environment in a few easy\nsteps.<\/p><\/section><section class=\"chapter\" id=\"vagrant\"><h2 id=\"vagrant_title\">Vagrant<\/h2><p><a href=\"https:\/\/www.vagrantup.com\/\">Vagrant<\/a> helps you build your virtual boxes on top of the known virtual environments and will configure these\nenvironments based on a single configuration file. These boxes can be set up manually, or you can use \u201cprovisioning\u201d\nsoftware such as <a href=\"https:\/\/puppet.com\/\">Puppet<\/a> or <a href=\"https:\/\/www.chef.io\/\">Chef<\/a> to do this for you. Provisioning the base box is a great way to ensure that\nmultiple boxes are set up in an identical fashion and removes the need for you to maintain complicated \u201cset up\u201d\ncommand lists. You can also \u201cdestroy\u201d your base box and recreate it without many manual steps, making it easy to create\na \u201cfresh\u201d installation.<\/p><p>Vagrant creates folders for sharing your code between your host and your virtual machine, which means that you can\ncreate and edit your files on your host machine and then run the code inside your virtual machine.<\/p><h3 id=\"a-little-help\">A little help<\/h3><p>If you need a little help to start using Vagrant there are some services that might be useful:<\/p><ul><li><a href=\"http:\/\/rove.io\/\">Rove<\/a>: service that allows you to pre-generate typical Vagrant builds, PHP among the options. The provisioning is\nmade with Chef.<\/li>\n  <li><a href=\"https:\/\/puphpet.com\/\">Puphpet<\/a>: simple GUI to set up virtual machines for PHP development. <strong>Heavily focused in PHP<\/strong>. Besides local VMs,\nit can be used to deploy to cloud services as well. The provisioning is made with Puppet.<\/li>\n  <li><a href=\"https:\/\/www.getprotobox.com\/\">Protobox<\/a>: is a layer on top of vagrant and a web GUI to setup virtual machines for web development. A single YAML\ndocument controls everything that is installed on the virtual machine.<\/li>\n  <li><a href=\"http:\/\/phansible.com\/\">Phansible<\/a>: provides an easy to use interface that helps you generate Ansible Playbooks for PHP based projects.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"docker\"><h2 id=\"docker_title\">Docker<\/h2><p><a href=\"https:\/\/www.docker.com\/\">Docker<\/a> - a lightweight alternative to a full virtual machine - is so called because it\u2019s all about \u201ccontainers\u201d.  A container is a building block which, in the simplest case, does one specific job, e.g. running a web server.  An \u201cimage\u201d is the package you use to build the container - Docker has a repository full of them.<\/p><p>A typical LAMP application might have three containers: a web server, a PHP-FPM process and MySQL. As with shared folders in Vagrant, you can leave your application files where they are and tell Docker where to find them.<\/p><p>You can generate containers from the command line (see example below) or, for ease of maintenance, build a <code class=\"highlighter-rouge\">docker-compose.yml<\/code> file for your project specifying which to create and how they communicate with one another.<\/p><p>Docker may help if you\u2019re developing multiple websites and want the separation that comes from installing each on it\u2019s own virtual machine, but don\u2019t have the necessary disk space or the time to keep everything up to date.  It\u2019s efficient: the installation and downloads are quicker, you only need to store one copy of each image however often it\u2019s used, containers need less RAM and share the same OS kernel, so you can have more servers running simultaneously, and it takes a matter of seconds to stop and start them, no need to wait for a full server boot.<\/p><h3 id=\"example-running-your-php-applications-in-docker\">Example: Running your PHP Applications in Docker<\/h3><p>After <a href=\"https:\/\/docs.docker.com\/install\/\">installing docker<\/a> on your machine, you can start a web server with one command.\nThe following will download a fully functional Apache installation with the latest PHP version, map <code class=\"highlighter-rouge\">\/path\/to\/your\/php\/files<\/code> to the document root, which you can view at <code class=\"highlighter-rouge\">http:\/\/localhost:8080<\/code>:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">docker run -d --name my-php-webserver -p 8080:80 -v \/path\/to\/your\/php\/files:\/var\/www\/html\/ php:apache<\/pre><\/figure><p>This will initialize and launch your container. <code class=\"highlighter-rouge\">-d<\/code> makes it runs in the background. To stop and start it, simply run <code class=\"highlighter-rouge\">docker stop my-php-webserver<\/code> and <code class=\"highlighter-rouge\">docker start my-php-webserver<\/code> (the other parameters are not needed again).<\/p><p>The command above shows a quick way to run a basic server.  There\u2019s much more you can do (and thousands of pre-built images in the <a href=\"https:\/\/hub.docker.com\/\">Docker Hub<\/a>). Take time to learn the terminology and read the <a href=\"https:\/\/docs.docker.com\/\">Docker User Guide<\/a> to get the most from it, and don\u2019t run random code you\u2019ve downloaded without checking it\u2019s safe \u2013 unofficial images may not have the latest security patches. If in doubt, stick to the <a href=\"https:\/\/hub.docker.com\/explore\/\">official repositiories<\/a>.<\/p><p>The <a href=\"https:\/\/phpdocker.io\/generator\">PHPDocker.io<\/a> site will auto-generate all the files you need for a fully-featured LAMP\/LEMP stack, including your choice of PHP version and extensions.<\/p><ul><li><a href=\"https:\/\/www.docker.com\/\">Docker Website<\/a><\/li>\n  <li><a href=\"https:\/\/docs.docker.com\/install\/\">Docker Installation<\/a><\/li>\n  <li><a href=\"https:\/\/docs.docker.com\/\">Docker User Guide<\/a><\/li>\n  <li><a href=\"https:\/\/hub.docker.com\/\">Docker Hub<\/a><\/li>\n  <li><a href=\"https:\/\/hub.docker.com\/explore\/\">Docker Hub - official images<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"caching\"><p>PHP is pretty quick by itself, but bottlenecks can arise when you make remote connections, load files, etc. \nThankfully, there are various tools available to speed up certain parts of your application, or reduce the number of times these various time-consuming tasks need to run.<\/p><\/section><section class=\"chapter\" id=\"opcode_cache\"><h2 id=\"opcode_cache_title\">Opcode Cache<\/h2><p>When a PHP file is executed, it must first be compiled into <a href=\"https:\/\/secure.php.net\/manual\/internals2.opcodes.php\">opcodes<\/a> (machine language instructions for the CPU). If the source code is unchanged, the opcodes will be the same, so this compilation step becomes a waste of CPU resources.<\/p><p>An opcode cache prevents redundant compilation by storing opcodes in memory and reusing them on successive calls. It will typically check signature or modification time of the file first, in case there have been any changes.<\/p><p>It\u2019s likely an opcode cache will make a significant speed improvement to your application.  Since PHP 5.5 there is one built in - <a href=\"https:\/\/secure.php.net\/book.opcache\">Zend OPcache<\/a>. Depending on your PHP package\/distribution, it\u2019s usually turned on by default -\u00a0check <a href=\"https:\/\/secure.php.net\/manual\/opcache.configuration.php#ini.opcache.enable\">opcache.enable<\/a> and the output of <code class=\"highlighter-rouge\">phpinfo()<\/code> to make sure. For earlier versions there\u2019s a PECL extension.<\/p><p>Read more about opcode caches:<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.opcache\">Zend OPcache<\/a> (bundled with PHP since 5.5)<\/li>\n  <li>Zend OPcache (formerly known as Zend Optimizer+) is now <a href=\"https:\/\/github.com\/zendtech\/ZendOptimizerPlus\">open source<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/book.apc\">APC<\/a> - PHP 5.4 and earlier<\/li>\n  <li><a href=\"https:\/\/xcache.lighttpd.net\/\">XCache<\/a><\/li>\n  <li><a href=\"https:\/\/www.iis.net\/downloads\/microsoft\/wincache-extension\">WinCache<\/a> (extension for MS Windows Server)<\/li>\n  <li><a href=\"https:\/\/wikipedia.org\/wiki\/List_of_PHP_accelerators\">list of PHP accelerators on Wikipedia<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"object_caching\"><h2 id=\"object_caching_title\">Object Caching<\/h2><p>There are times when it can be beneficial to cache individual objects in your code, such as with data that is expensive\nto get or database calls where the result is unlikely to change. You can use object caching software to hold these\npieces of data in memory for extremely fast access later on. If you save these items to a data store after you retrieve\nthem, then pull them directly from the cache for following requests, you can gain a significant improvement in\nperformance as well as reduce the load on your database servers.<\/p><p>Many of the popular bytecode caching solutions let you cache custom data as well, so there\u2019s even more reason to take\nadvantage of them. APCu, XCache, and WinCache all provide APIs to save data from your PHP code to their memory cache.<\/p><p>The most commonly used memory object caching systems are APCu and memcached. APCu is an excellent choice for object\ncaching, it includes a simple API for adding your own data to its memory cache and is very easy to setup and use. The\none real limitation of APCu is that it is tied to the server it\u2019s installed on. Memcached on the other hand is\ninstalled as a separate service and can be accessed across the network, meaning that you can store objects in a\nhyper-fast data store in a central location and many different systems can pull from it.<\/p><p>Note that when running PHP as a (Fast-)CGI application inside your webserver, every PHP process will have its own cache,\ni.e. APCu data is not shared between your worker processes. In these cases, you might want to consider using memcached\ninstead, as it\u2019s not tied to the PHP processes.<\/p><p>In a networked configuration APCu will usually outperform memcached in terms of access speed, but memcached will be\nable to scale up faster and further. If you do not expect to have multiple servers running your application, or do not\nneed the extra features that memcached offers then APCu is probably your best choice for object caching.<\/p><p>Example logic using APCu:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ check if there is data saved as 'expensive_data' in cache\n$data = apc_fetch('expensive_data');\nif ($data === false) {\n    \/\/ data is not in cache; save result of expensive call for later use\n    apc_add('expensive_data', $data = get_expensive_data());\n}\nprint_r($data);<\/pre><\/figure><p>Note that prior to PHP 5.5, APC provides both an object cache and a bytecode cache. APCu is a project to bring APC\u2019s\nobject cache to PHP 5.5+, since PHP now has a built-in bytecode cache (OPcache).<\/p><ul><li><a href=\"https:\/\/github.com\/krakjoe\/apcu\">APCu<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.apc\">APC Functions<\/a><\/li>\n  <li><a href=\"https:\/\/memcached.org\/\">Memcached<\/a><\/li>\n  <li><a href=\"https:\/\/redis.io\/\">Redis<\/a><\/li>\n  <li><a href=\"https:\/\/xcache.lighttpd.net\/wiki\/XcacheApi\">XCache APIs<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.wincache\">WinCache Functions<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"phpdoc\"><h2 id=\"phpdoc_title\">PHPDoc<\/h2><p>PHPDoc is an informal standard for commenting PHP code. There are a <em>lot<\/em> of different <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/index.html\">tags<\/a> available. The full list\nof tags and examples can be found at the <a href=\"https:\/\/docs.phpdoc.org\/index.html\">PHPDoc manual<\/a>.<\/p><p>Below is an example of how you might document a class with a few methods;<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/**\n * @author A Name &lt;a.name@example.com&gt;\n * @link http:\/\/www.phpdoc.org\/docs\/latest\/index.html\n *\/\nclass DateTimeHelper\n{\n    \/**\n     * @param mixed $anything Anything that we can convert to a \\DateTime object\n     *\n     * @throws \\InvalidArgumentException\n     *\n     * @return \\DateTime\n     *\/\n    public function dateTimeFromAnything($anything)\n    {\n        $type = gettype($anything);\n        switch ($type) {\n            \/\/ Some code that tries to return a \\DateTime object\n        }\n        throw new \\InvalidArgumentException(\n            \"Failed Converting param of type '{$type}' to DateTime object\"\n        );\n    }\n    \/**\n     * @param mixed $date Anything that we can convert to a \\DateTime object\n     *\n     * @return void\n     *\/\n    public function printISO8601Date($date)\n    {\n        echo $this-&gt;dateTimeFromAnything($date)-&gt;format('c');\n    }\n    \/**\n     * @param mixed $date Anything that we can convert to a \\DateTime object\n     *\/\n    public function printRFC2822Date($date)\n    {\n        echo $this-&gt;dateTimeFromAnything($date)-&gt;format('r');\n    }\n}<\/pre><\/figure><p>The documentation for the class as a whole has the <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/author.html\">@author<\/a> tag and a <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/link.html\">@link<\/a> tag. The <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/author.html\">@author<\/a> tag is used to\ndocument the author of the code and can be repeated for documenting several authors. The <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/link.html\">@link<\/a> tag is used to link to a website indicating a relationship between the website and the code.<\/p><p>Inside the class, the first method has a <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/param.html\">@param<\/a> tag documenting the type, name and description of the parameter\nbeing passed to the method. Additionally it has the <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/return.html\">@return<\/a> and <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/throws.html\">@throws<\/a> tags for documenting the return type, and any exceptions that could be thrown respectively.<\/p><p>The second and third methods are very similar and have a single <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/param.html\">@param<\/a> tag as did the first method. The important\ndifference between the second and third methods\u2019 doc block is the inclusion\/exclusion of the <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/return.html\">@return<\/a> tag.\n<code class=\"highlighter-rouge\">@return void<\/code> explicitly informs us that there is no return; historically omitting the <code class=\"highlighter-rouge\">@return void<\/code> statement also results in the same (no return) action.<\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"from_the_source\"><h2 id=\"from_the_source_title\">From the Source<\/h2><ul><li><a href=\"https:\/\/secure.php.net\/\">PHP Website<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/docs.php\">PHP Documentation<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"people_to_follow\"><h2 id=\"people_to_follow_title\">People to Follow<\/h2><p>It\u2019s difficult to find interesting and knowledgeable PHP\ncommunity members when you are first starting out. You can\nfind an abbreviated list of PHP community members to get you started at:<\/p><ul><li><a href=\"https:\/\/www.ogprogrammer.com\/2017\/06\/28\/how-to-get-connected-with-the-php-community\/\">https:\/\/www.ogprogrammer.com\/2017\/06\/28\/how-to-get-connected-with-the-php-community\/<\/a><\/li>\n  <li><a href=\"https:\/\/twitter.com\/CalEvans\/lists\/phpeople\">https:\/\/twitter.com\/CalEvans\/lists\/phpeople<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"mentoring\"><h2 id=\"mentoring_title\">Mentoring<\/h2><ul><li><a href=\"https:\/\/php-mentoring.org\/\">php-mentoring.org<\/a> - Formal, peer to peer mentoring in the PHP community.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"php_paas_providers\"><h2 id=\"php_paas_providers_title\">PHP PaaS Providers<\/h2><ul><li><a href=\"https:\/\/www.ctl.io\/appfog\/\">AppFog<\/a><\/li>\n  <li><a href=\"https:\/\/aws.amazon.com\/elasticbeanstalk\/\">AWS Elastic Beanstalk<\/a><\/li>\n  <li><a href=\"https:\/\/www.cloudways.com\/\">Cloudways<\/a><\/li>\n  <li><a href=\"https:\/\/www.engineyard.com\/features\">Engine Yard Cloud<\/a><\/li>\n  <li><a href=\"https:\/\/www.fortrabbit.com\/\">fortrabbit<\/a><\/li>\n  <li><a href=\"https:\/\/cloud.google.com\/appengine\/docs\/php\/\">Google App Engine<\/a><\/li>\n  <li><a href=\"https:\/\/devcenter.heroku.com\/categories\/php-support\">Heroku<\/a><\/li>\n  <li><a href=\"https:\/\/console.bluemix.net\/docs\/runtimes\/php\/getting-started.html#getting_started\">IBM Cloud<\/a><\/li>\n  <li><a href=\"https:\/\/jelastic.com\/\">Jelastic<\/a><\/li>\n  <li><a href=\"https:\/\/azure.microsoft.com\/\">Microsoft Azure<\/a><\/li>\n  <li><a href=\"https:\/\/nanobox.io\/\">Nanobox<\/a><\/li>\n  <li><a href=\"https:\/\/run.pivotal.io\/\">Pivotal Web Services<\/a><\/li>\n  <li><a href=\"https:\/\/platform.sh\/\">Platform.sh<\/a><\/li>\n  <li><a href=\"https:\/\/www.openshift.com\/\">Red Hat OpenShift<\/a><\/li>\n<\/ul><p>To see which versions these PaaS hosts are running, head over to <a href=\"http:\/\/phpversions.info\/paas-hosting\/\">PHP Versions<\/a>.<\/p><\/section><section class=\"chapter\" id=\"frameworks\"><h2 id=\"frameworks_title\">Frameworks<\/h2><p>Rather than re-invent the wheel, many PHP developers use frameworks to build out web applications. Frameworks abstract\naway many of the low-level concerns and provide helpful, easy-to-use interfaces to complete common tasks.<\/p><p>You do not need to use a framework for every project. Sometimes plain PHP is the right way to go, but if you do need a\nframework then there are three main types available:<\/p><ul><li>Micro Frameworks<\/li>\n  <li>Full-Stack Frameworks<\/li>\n  <li>Component Frameworks<\/li>\n<\/ul><p>Micro-frameworks are essentially a wrapper to route a HTTP request to a callback, controller, method, etc as quickly as\npossible, and sometimes come with a few extra libraries to assist development such as basic database wrappers and the\nlike. They are prominently used to build remote HTTP services.<\/p><p>Many frameworks add a considerable number of features on top of what is available in a micro-framework; these are\ncalled Full-Stack Frameworks. These often come bundled with ORMs, Authentication packages, etc.<\/p><p>Component-based frameworks are collections of specialized and single-purpose libraries. Disparate component-based\nframeworks can be used together to make a micro- or full-stack framework.<\/p><\/section><section class=\"chapter\" id=\"components\"><h2 id=\"components_title\">Components<\/h2><p>As mentioned above \u201cComponents\u201d are another approach to the common goal of creating, distributing and implementing\nshared code. Various component repositories exist, the main two of which are:<\/p><ul><li><a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">Packagist<\/a><\/li>\n  <li><a href=\"https:\/\/phptherightway.com\/#pear\">PEAR<\/a><\/li>\n<\/ul><p>Both of these repositories have command line tools associated with them to help the installation and upgrade processes,\nand have been explained in more detail in the <a href=\"https:\/\/phptherightway.com\/#dependency_management\">Dependency Management<\/a> section.<\/p><p>There are also component-based frameworks and component-vendors that offer no framework at all. These projects provide\nanother source of packages which ideally have little to no dependencies on other packages, or specific frameworks.<\/p><p>For example, you can use the <a href=\"https:\/\/github.com\/fuelphp\/validation\">FuelPHP Validation package<\/a>, without needing to use the FuelPHP framework itself.<\/p><ul><li><a href=\"http:\/\/auraphp.com\/framework\/\">Aura<\/a><\/li>\n  <li>CakePHP Components\n    <ul><li><a href=\"https:\/\/github.com\/cakephp\/collection\">Collection<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/database\">Database<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/datasource\">Datasource<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/event\">Event<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/i18n\">I18n<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/orm\">ORM<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"https:\/\/github.com\/fuelphp\">FuelPHP<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/hoaproject\">Hoa Project<\/a><\/li>\n  <li><a href=\"https:\/\/symfony.com\/doc\/current\/components\/index.html\">Symfony Components<\/a><\/li>\n  <li><a href=\"https:\/\/thephpleague.com\/\">The League of Extraordinary Packages<\/a><\/li>\n  <li>Laravel\u2019s Illuminate Components\n    <ul><li><a href=\"https:\/\/github.com\/illuminate\/container\">IoC Container<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/illuminate\/database\">Eloquent ORM<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/illuminate\/queue\">Queue<\/a><\/li>\n    <\/ul><\/li>\n<\/ul><p><em>Laravel\u2019s <a href=\"https:\/\/github.com\/illuminate\">Illuminate components<\/a> will become better decoupled from the Laravel framework. For now, only the\ncomponents best decoupled from the Laravel framework are listed above.<\/em><\/p><\/section><section class=\"chapter\" id=\"other_resources\"><h2 id=\"other_resources_title\">Other Useful Resources<\/h2><h3 id=\"cheatsheets\">Cheatsheets<\/h3><ul><li><a href=\"http:\/\/phpcheatsheets.com\/\">PHP Cheatsheets<\/a> - for variable comparisons, arithmetics and variable testing in various\nPHP versions<\/li>\n  <li><a href=\"https:\/\/www.owasp.org\/index.php\/PHP_Security_Cheat_Sheet\">PHP Security Cheatsheet<\/a><\/li>\n<\/ul><h3 id=\"more-best-practices\">More best practices<\/h3><ul><li><a href=\"https:\/\/phpbestpractices.org\/\">PHP Best Practices<\/a><\/li>\n  <li><a href=\"https:\/\/www.airpair.com\/php\/posts\/best-practices-for-modern-php-development\">Best practices for Modern PHP Development<\/a><\/li>\n  <li><a href=\"https:\/\/kinsta.com\/blog\/php-versions\/\">Why You Should Be Using Supported PHP Versions<\/a><\/li>\n<\/ul><h3 id=\"news-around-the-php-and-web-development-communities\">News around the PHP and web development communities<\/h3><p>You can subscribe to weekly newsletters to keep yourself informed on new libraries, latest news, events and general\nannouncements, as well as additional resources being published every now and then:<\/p><ul><li><a href=\"http:\/\/www.phpweekly.com\">PHP Weekly<\/a><\/li>\n  <li><a href=\"https:\/\/javascriptweekly.com\/\">JavaScript Weekly<\/a><\/li>\n  <li><a href=\"https:\/\/frontendfoc.us\/\">Frontend Focus<\/a><\/li>\n  <li><a href=\"https:\/\/mobiledevweekly.com\/\">Mobile Web Weekly<\/a><\/li>\n<\/ul><p>There are also Weeklies on other platforms you might be interested in; here\u2019s <a href=\"https:\/\/github.com\/jondot\/awesome-weekly\">a list of some<\/a>.<\/p><h3 id=\"php-universe\">PHP universe<\/h3><ul><li><a href=\"https:\/\/blog.phpdeveloper.org\/\">PHP Developer blog<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"videos\"><h2>Video Tutorials<\/h2><h3 id=\"youtube-channels\">YouTube Channels<\/h3><ul><li><a href=\"https:\/\/www.youtube.com\/user\/phpacademy\">PHP Academy<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/user\/thenewboston\">The New Boston<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/user\/businessgeek\">Sherif Ramadan<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/user\/LevelUpTuts\">Level Up Tuts<\/a><\/li>\n<\/ul><h3 id=\"paid-videos\">Paid Videos<\/h3><ul><li><a href=\"https:\/\/teamtreehouse.com\/library\/php-standards-and-best-practices\">Standards and Best practices<\/a><\/li>\n  <li><a href=\"https:\/\/www.pluralsight.com\/search?q=php\">PHP Training on Pluralsight<\/a><\/li>\n  <li><a href=\"https:\/\/www.lynda.com\/search?q=php\">PHP Training on Lynda.com<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/categories\/php\/courses\">PHP Training on Tutsplus<\/a><\/li>\n  <li><a href=\"https:\/\/laracasts.com\/\">Laracasts<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"books\"><h2 id=\"books_title\">Books<\/h2><p>There are many PHP books; sadly some are now quite old and no longer accurate. In particular, avoid books on \u201cPHP 6\u201d, a version that will now never exist. The next major release of PHP after 5.6 was \u201cPHP 7\u201d, <a href=\"https:\/\/wiki.php.net\/rfc\/php6\">partly because of this<\/a>.<\/p><p>This section aims to be a living document for recommended books on PHP development in general. If you would like your\nbook to be added, send a PR and it will be reviewed for relevancy.<\/p><h3 id=\"free-books\">Free Books<\/h3><ul><li><a href=\"http:\/\/daylerees.com\/php-pandas\/\">PHP Pandas<\/a> - Aims to teach everyone how to be a web developer.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/phptherightway\/\">PHP The Right Way<\/a> - This website is available as a book completely for free.<\/li>\n  <li><a href=\"https:\/\/paragonie.com\/book\/pecl-libsodium\">Using Libsodium in PHP Projects<\/a> - Guide to using Libsodium PHP extension\nfor modern, secure, and fast cryptography.<\/li>\n<\/ul><h3 id=\"paid-books\">Paid Books<\/h3><ul><li><a href=\"https:\/\/apisyouwonthate.com\/\">Build APIs You Won\u2019t Hate<\/a> - Everyone and their dog wants an API,\nso you should probably learn how to build them.<\/li>\n  <li><a href=\"http:\/\/shop.oreilly.com\/product\/0636920033868.do\">Modern PHP<\/a> - covers modern PHP features, best practices, testing, tuning, deployment and setting up a dev environment.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/buildingsecurephpapps\">Building Secure PHP Apps<\/a> - Learn the security basics that a senior\ndeveloper usually acquires over years of experience, all condensed down into one quick and easy handbook<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/mlaphp\">Modernizing Legacy Applications In PHP<\/a> - Get your code under control in a series of\nsmall, specific steps<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/securingphp-coreconcepts\">Securing PHP: Core Concepts<\/a> - A guide to some of the most common\nsecurity terms and provides some examples of them in every day PHP<\/li>\n  <li><a href=\"http:\/\/www.scalingphpbook.com\/\">Scaling PHP<\/a> - Stop playing sysadmin and get back to coding<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/signalingphp\">Signaling PHP<\/a> - PCNLT signals are a great help when writing PHP scripts that\nrun from the command line.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/minimumviabletests\">Minimum Viable Tests<\/a> - Long-time PHP testing evangelist Chris Hartjes goes over what he feels is the minimum you need to know to get started.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/ddd-in-php\">Domain-Driven Design in PHP<\/a> - See real examples written in PHP showcasing Domain-Driven Design Architectural Styles (Hexagonal Architecture, CQRS or Event Sourcing), Tactical Design Patterns, and Bounded Context Integration.<\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"user_groups\"><h2 id=\"user_groups_title\">PHP User Groups<\/h2><p>If you live in a larger city, odds are there\u2019s a PHP user group nearby. You can easily find your local PUG at\n<a href=\"https:\/\/php.ug\/\">PHP.ug<\/a>. Alternate sources might be <a href=\"http:\/\/www.meetup.com\/find\/\">Meetup.com<\/a> or a search for <code class=\"highlighter-rouge\">php user group near me<\/code>\nusing your favorite search engine (i.e. <a href=\"https:\/\/www.google.com\/search?q=php+user+group+near+me\">Google<\/a>). If you live in a smaller town, there may not be a\nlocal PUG; if that\u2019s the case, start one!<\/p><p>Special mention should be made of two global user groups: <a href=\"https:\/\/nomadphp.com\/\">NomadPHP<\/a> and <a href=\"http:\/\/phpwomen.org\/\">PHPWomen<\/a>. <a href=\"https:\/\/nomadphp.com\/\">NomadPHP<\/a> offers twice monthly\nonline user group meetings with presentations by some of the top speakers in the PHP community.\n<a href=\"http:\/\/phpwomen.org\/\">PHPWomen<\/a> is a non-exclusive user group originally targeted towards the women in the PHP world. Membership is open to\neveryone who supports a more diverse community. PHPWomen provide a network for support, mentorship and education, and\ngenerally promote the creating of a \u201cfemale friendly\u201d and professional atmosphere.<\/p><p><a href=\"https:\/\/wiki.php.net\/usergroups\">Read about User Groups on the PHP Wiki<\/a><\/p><\/section><section class=\"chapter\" id=\"conferences\"><h2 id=\"conferences_title\">PHP Conferences<\/h2><p>The PHP community also hosts larger regional and national conferences in many countries around the world. Well-known\nmembers of the PHP community usually speak at these larger events, so it\u2019s a great opportunity to learn directly from\nindustry leaders.<\/p><p><a href=\"https:\/\/secure.php.net\/conferences\/index.php\">Find a PHP Conference<\/a><\/p><\/section><section class=\"chapter\" id=\"elephpants\"><h2 id=\"elephpants_title\">ElePHPants<\/h2><p><a href=\"https:\/\/secure.php.net\/elephpant.php\">ElePHPant<\/a> is that beautiful mascot of the PHP project with elephant in their design. It was originally designed for the PHP project in 1998 by <a href=\"http:\/\/www.elroubio.net\/\">Vincent Pontier<\/a> - spiritual father of thousands of elePHPants around the world and 10 years later adorable plush elephant toy came to birth as well. Now elePHPants are present at many PHP conferences and with many PHP developers at their computers for fun and inspiration.<\/p><p><a href=\"https:\/\/7php.com\/elephpant\/\">Interview with Vincent Pontier<\/a><\/p><\/section>","created_at":"2018-11-07T16:37:36+0000","updated_at":"2018-11-07T16:37:36+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":108,"domain_name":"phptherightway.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":205,"title":"cjbarber\/ToolsOfTheTrade","url":"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<blockquote>\n<p>Tools of The Trade, from <a href=\"https:\/\/news.ycombinator.com\" rel=\"nofollow\">Hacker News<\/a>.<\/p>\n<\/blockquote>\n<h5><a id=\"user-content-background\" class=\"anchor\" aria-hidden=\"true\" href=\"#background\"><\/a>Background<\/h5>\n<p>In 2010, Joshua Schachter, the founder of Delicious, <a href=\"https:\/\/news.ycombinator.com\/item?id=1769910\" rel=\"nofollow\">posted the following<\/a> on Hacker News:<\/p>\n<blockquote>\n<p>When I first started delicious, we had to host most of the services ourselves. CVS, mail, mailing lists, etc etc etc.<\/p>\n<p>These days, lots of that stuff is available as SaaS. What are the tools and services people use instead of hosting their own?<\/p>\n<p>(I'm not talking about actual production services like EC2 and Heroku and whatnot. We can go over this in another thread.)<\/p>\n<\/blockquote>\n<p>In 2013, Sharjeel Qureshi, <a href=\"https:\/\/news.ycombinator.com\/item?id=5235137\" rel=\"nofollow\">posted the following<\/a>:<\/p>\n<blockquote>\n<p>Few years ago, Joshua Schachter started this thread on HN for discussing hosted useful services: <a href=\"https:\/\/news.ycombinator.com\/item?id=1769910\" rel=\"nofollow\">https:\/\/news.ycombinator.com\/item?id=1769910<\/a><\/p>\n<p>The contribution in thread introduced many interesting SaaS services which can immensely help in deploying services as well as development.<\/p>\n<p>It's been three years since then. What do we have today?<\/p>\n<\/blockquote>\n<p>Many thanks to the big contributors to the previous threads, including garrettdimon, espeed, netshade, and cmadan, and many more that I haven't named.<\/p>\n<h5><a id=\"user-content-now\" class=\"anchor\" aria-hidden=\"true\" href=\"#now\"><\/a>Now<\/h5>\n<p>I've collected more data from Hacker News, AngelList and Quora, to make the 2015 and beyond version. This list also includes self-hosted as well as hosted services.<\/p>\n<p>It's on GitHub for a reason! Please submit pull requests.<\/p>\n<h2><a id=\"user-content-supported-with-\ufe0f-by-instabug\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-with-\ufe0f-by-instabug\"><\/a>Supported with \u2764\ufe0f by <a href=\"https:\/\/instabug.com\/?utm_source=toolsofthetrade&amp;utm_medium=spon&amp;utm_content=header\" rel=\"nofollow\">Instabug<\/a><\/h2>\n<p><a href=\"https:\/\/instabug.com\/?utm_source=toolsofthetrade&amp;utm_medium=spon&amp;utm_content=banner\" rel=\"nofollow\"><img src=\"https:\/\/user-images.githubusercontent.com\/9888943\/46256666-08746700-c4ae-11e8-8c64-406bbbfe2e0f.png\" alt=\"instabug\" \/><\/a><\/p>\n<blockquote>\n<p>Instabug will enable you to easily collect feedback from your beta testers and users. The SDK will automatically attach all the details you would need to debug and fix bugs and crashes faster. Join the top apps in the world that rely on Instabug to enhance their app quality and iterate faster. <a href=\"https:\/\/instabug.com\/?utm_source=toolsofthetrade&amp;utm_medium=spon&amp;utm_content=learnmore\" rel=\"nofollow\">learn more.<\/a>\ud83d\ude80<a href=\"https:\/\/instabug.com\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/cb3c1dfbdecd9b7411d57861eba9286374ce6638\/68747470733a2f2f696e7374616275672d67612e61707073706f742e636f6d2f55412d34313938323038382d362f6769746875622f546f6f6c734f6654686554726164653f706978656c\" alt=\"\" data-canonical-src=\"https:\/\/instabug-ga.appspot.com\/UA-41982088-6\/github\/ToolsOfTheTrade?pixel\" \/><\/a><\/p>\n<\/blockquote>\n<h2><a id=\"user-content-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#contents\"><\/a>Contents<\/h2>\n<ul><li><a href=\"#accountinginvoicing\">Accounting\/Invoicing<\/a><\/li>\n<li><a href=\"#api-builder\">API Builder<\/a><\/li>\n<li><a href=\"#app-developer-tools\">App Developer Tools<\/a><\/li>\n<li><a href=\"#application-distribution\">Application Distribution<\/a><\/li>\n<li><a href=\"#application-performance\">Application Performance<\/a><\/li>\n<li><a href=\"#aws-tools\">AWS Tools<\/a><\/li>\n<li><a href=\"#backend-aas\">Backend-aaS<\/a><\/li>\n<li><a href=\"#banking\">Banking<\/a><\/li>\n<li><a href=\"#billing--payment-processing\">Billing &amp; Payment Processing<\/a><\/li>\n<li><a href=\"#browseremail-testing\">Browser\/Email Testing<\/a><\/li>\n<li><a href=\"#bugissue-tracking\">Bug\/Issue Tracking<\/a><\/li>\n<li><a href=\"#business--traffic-analytics\">Business &amp; Traffic Analytics<\/a><\/li>\n<li><a href=\"#business-cards-and-print-material\">Business Cards and Print Material<\/a><\/li>\n<li><a href=\"#community-tools\">Community Tools<\/a><\/li>\n<li><a href=\"#content-creationinfographics\">Content Creation\/Infographics<\/a><\/li>\n<li><a href=\"#continuous-integrationcode-quality\">Continuous Integration\/Code Quality<\/a><\/li>\n<li><a href=\"#conversion-optimization--ab-testing\">Conversion Optimization &amp; A\/B Testing<\/a><\/li>\n<li><a href=\"#crmsales-tools\">CRM\/Sales Tools<\/a><\/li>\n<li><a href=\"#customer-feedback\">Customer Feedback<\/a><\/li>\n<li><a href=\"#customer-supporthelp-desks\">Customer Support\/Help Desks<\/a><\/li>\n<li><a href=\"#dashboards\">Dashboards<\/a><\/li>\n<li><a href=\"#data\">Data<\/a><\/li>\n<li><a href=\"#database-aas\">Database-aaS<\/a><\/li>\n<li><a href=\"#database\">Database<\/a><\/li>\n<li><a href=\"#deployment\">Deployment<\/a><\/li>\n<li><a href=\"#design-collaboration\">Design Collaboration<\/a><\/li>\n<li><a href=\"#designers\">Designers<\/a><\/li>\n<li><a href=\"#dns\">DNS<\/a><\/li>\n<li><a href=\"#documentation\">Documentation<\/a><\/li>\n<li><a href=\"#email-collectionlanding-page-apps\">Email Collection\/Landing Page Apps<\/a><\/li>\n<li><a href=\"#email-marketing\">Email Marketing<\/a><\/li>\n<li><a href=\"#engineering-metrics\">Engineering Metrics<\/a><\/li>\n<li><a href=\"#errorexception-handling\">Error\/Exception Handling<\/a><\/li>\n<li><a href=\"#forms--surveys\">Forms \/ Surveys<\/a><\/li>\n<li><a href=\"#fundraising--investor-relations\">Fundraising \/ Investor Relations<\/a><\/li>\n<li><a href=\"#group-communicationchat-tools\">Group Communication\/Chat Tools<\/a><\/li>\n<li><a href=\"#heroku-tools\">Heroku Tools<\/a><\/li>\n<li><a href=\"#hr\">HR<\/a><\/li>\n<li><a href=\"#identity-verification\">Identity Verification<\/a><\/li>\n<li><a href=\"#income-analytics\">Income Analytics<\/a><\/li>\n<li><a href=\"#knowledge-trackingwiki\">Knowledge Tracking\/Wiki<\/a><\/li>\n<li><a href=\"#load-testing\">Load Testing<\/a><\/li>\n<li><a href=\"#localization--internationalization\">Localization &amp; Internationalization<\/a><\/li>\n<li><a href=\"#log-monitoring\">Log Monitoring<\/a><\/li>\n<li><a href=\"#naming\">Naming<\/a><\/li>\n<li><a href=\"#notes\">Notes<\/a><\/li>\n<li><a href=\"#offsite-backups\">Offsite Backups<\/a><\/li>\n<li><a href=\"#ops-alerts-and-scheduling\">Ops Alerts and Scheduling<\/a><\/li>\n<li><a href=\"#other-apis\">Other APIs<\/a><\/li>\n<li><a href=\"#paas\">PaaS<\/a><\/li>\n<li><a href=\"#password-management\">Password Management<\/a><\/li>\n<li><a href=\"#payments-billing--downloads\">Payments, Billing &amp; Downloads<\/a><\/li>\n<li><a href=\"#payroll\">Payroll<\/a><\/li>\n<li><a href=\"#personal-machine-backups\">Personal Machine Backups<\/a><\/li>\n<li><a href=\"#personal-productivity\">Personal Productivity<\/a><\/li>\n<li><a href=\"#phonepbxsms\">Phone\/PBX\/SMS<\/a><\/li>\n<li><a href=\"#planning--project-management\">Planning &amp; Project Management<\/a><\/li>\n<li><a href=\"#presentations--slides\">Presentations \/ Slides<\/a><\/li>\n<li><a href=\"#privacy-policy-terms--conditions-legal-documents\">Privacy Policy, Terms &amp; Conditions, Legal Documents<\/a><\/li>\n<li><a href=\"#prototypingmockups\">Prototyping\/Mockups<\/a><\/li>\n<li><a href=\"#remote-collaboration\">Remote Collaboration<\/a><\/li>\n<li><a href=\"#remote-workers\">Remote Workers<\/a><\/li>\n<li><a href=\"#search\">Search<\/a><\/li>\n<li><a href=\"#security\">Security<\/a><\/li>\n<li><a href=\"#seo-tools\">SEO Tools<\/a><\/li>\n<li><a href=\"#server-monitoring\">Server Monitoring<\/a><\/li>\n<li><a href=\"#shipping\">Shipping<\/a><\/li>\n<li><a href=\"#site-search\">Site Search<\/a><\/li>\n<li><a href=\"#social-media-marketing\">Social Media Marketing<\/a><\/li>\n<li><a href=\"#source-code-hosting\">Source Code Hosting<\/a><\/li>\n<li><a href=\"#sources-of-clicksad-platforms\">Sources of Clicks\/Ad Platforms<\/a><\/li>\n<li><a href=\"#space-rental\">Space Rental<\/a><\/li>\n<li><a href=\"#status-blogsuser-alerts\">Status Blogs\/User Alerts<\/a><\/li>\n<li><a href=\"#storage\">Storage<\/a><\/li>\n<li><a href=\"#system-monitoring\">System Monitoring<\/a><\/li>\n<li><a href=\"#task-scheduling\">Task Scheduling<\/a><\/li>\n<li><a href=\"#time-tracking\">Time Tracking<\/a><\/li>\n<li><a href=\"#transactional-email\">Transactional Email<\/a><\/li>\n<li><a href=\"#user-feedback\">User Feedback<\/a><\/li>\n<li><a href=\"#user-management\">User Management<\/a><\/li>\n<li><a href=\"#user-testing\">User Testing<\/a><\/li>\n<li><a href=\"#video-hosting\">Video Hosting<\/a><\/li>\n<li><a href=\"#vps\">VPS<\/a><\/li>\n<li><a href=\"#websockets-aas\">WebSockets-aaS<\/a><\/li>\n<\/ul><h3><a id=\"user-content-identity-verification\" class=\"anchor\" aria-hidden=\"true\" href=\"#identity-verification\"><\/a>Identity Verification<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/cognitohq.com\" rel=\"nofollow\">Cognito<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getcognito\" rel=\"nofollow\">@getcognito<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Frictionless, modern identity verification that starts with just a phone number.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/blockscore.com\" rel=\"nofollow\">BlockScore<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getblockscore\" rel=\"nofollow\">@getblockscore<\/a><\/td>\n<td align=\"left\">$1\/verification - $5\/verification<\/td>\n<td align=\"left\">Individual ID verifications and company verifications.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/onfido.com\" rel=\"nofollow\">Onfido<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Onfido\" rel=\"nofollow\">@Onfido<\/a><\/td>\n<td align=\"left\">$2\/verification<\/td>\n<td align=\"left\">Individual ID verifications, Document Checks and Facial recognition.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-browseremail-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#browseremail-testing\"><\/a>Browser\/Email Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.browserstack.com\" rel=\"nofollow\">BrowserStack<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/browserstack\" rel=\"nofollow\">@browserstack<\/a><\/td>\n<td align=\"left\">$35\/mo - $359\/mo<\/td>\n<td align=\"left\">Live, web-based browser testing.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/litmus.com\" rel=\"nofollow\">Litmus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/litmusapp\" rel=\"nofollow\">@litmusapp<\/a><\/td>\n<td align=\"left\">$79\/mo - $399\/mo<\/td>\n<td align=\"left\">Lets you preview your campaigns across 30+ real email clients and devices in minutes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/saucelabs.com\" rel=\"nofollow\">Sauce Labs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/saucelabs\" rel=\"nofollow\">@saucelabs<\/a><\/td>\n<td align=\"left\">$12\/mo - $149\/mo<\/td>\n<td align=\"left\">Run your web and mobile app tests across hundreds of real browsers and platforms instantly.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.emailonacid.com\" rel=\"nofollow\">EmailOnAcid<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/EmailonAcid\" rel=\"nofollow\">@emailonacid<\/a><\/td>\n<td align=\"left\">$45\/mo - $295\/mo<\/td>\n<td align=\"left\">Quickly ensure that your emails don't look like crap in any email client.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.rainforestqa.com\" rel=\"nofollow\">Rainforest QA<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/rainforestqa\" rel=\"nofollow\">@rainforestqa<\/a><\/td>\n<td align=\"left\">$500\/mo - $2000\/mo<\/td>\n<td align=\"left\">Insanely simple integration testing.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/debugmail.io\" rel=\"nofollow\">DebugMail<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Easy to use fake mail (SMTP) server for developers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mailosaur.com\" rel=\"nofollow\">Mailosaur<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mailosaur\" rel=\"nofollow\">@mailosaur<\/a><\/td>\n<td align=\"left\">$19\/mo - $199\/mo<\/td>\n<td align=\"left\">Email, SMS and Spam testing with virtual SMTP servers for business-grade usage.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mailtrap.io\" rel=\"nofollow\">Mailtrap<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Mailtrap\" rel=\"nofollow\">@Mailtrap<\/a><\/td>\n<td align=\"left\">Free - $24.99\/mo<\/td>\n<td align=\"left\">A fake SMTP server for development and staging environments along with a REST API for testing.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-bugissue-tracking\" class=\"anchor\" aria-hidden=\"true\" href=\"#bugissue-tracking\"><\/a>Bug\/Issue Tracking<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/bitbucket.org\" rel=\"nofollow\">BitBucket Issues<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bitbucket\" rel=\"nofollow\">@bitbucket<\/a><\/td>\n<td align=\"left\">$10\/mo - $200\/mo, Unlimited private code repositories<\/td>\n<td align=\"left\">Host, manage, and share Git and Mercurial repositories in the cloud. Free, unlimited private repositories for up to 5 developers give teams the flexibility to grow and code without restrictions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/bugherd.com\" rel=\"nofollow\">BugHerd<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bugherd\" rel=\"nofollow\">@bugherd<\/a><\/td>\n<td align=\"left\">$29\/mo - $180\/mo<\/td>\n<td align=\"left\">Turn client feedback into actionable tasks. BugHerd lets you quickly see, at a glance, how your project is going and what everyone is working on. The task board lets you keep team members in sync by assigning and scheduling tasks with a simple drag and drop.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/bugify.com\" rel=\"nofollow\">Bugify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bugify\" rel=\"nofollow\">@bugify<\/a><\/td>\n<td align=\"left\">$59.00<\/td>\n<td align=\"left\">Self hosted issue management system. One-time payment. Written in PHP.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\">GitHub Issues<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/GitHub\" rel=\"nofollow\">@GitHub<\/a><\/td>\n<td align=\"left\">$7\/mo - $50\/mo<\/td>\n<td align=\"left\">Build software better, together. GitHub is the largest code host on the planet with over 13.2 million repositories. Large or small, every repository comes with the same powerful tools. These tools are open to the community for public projects and secure for private projects.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/about.gitlab.com\" rel=\"nofollow\">GitLab Issues<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gitlabhq\" rel=\"nofollow\">@gitlabhq<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">GitLab is open source software to collaborate on code that is used by more than 100,000 organisations worldwide. Unlimited private repositories on GitLab.com or host your own instance. Enterprise Edition includes deep LDAP support.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/huboard.com\" rel=\"nofollow\">Huboard<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/huboard\" rel=\"nofollow\">@huboard<\/a><\/td>\n<td align=\"left\">$7\/mo - $24\/mo<\/td>\n<td align=\"left\">GitHub issues made awesome Instant project management for GitHub repositories<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/jira\" rel=\"nofollow\">JIRA<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/JIRA\" rel=\"nofollow\">@JIRA<\/a><\/td>\n<td align=\"left\">$10\/mo hosted - $10\/yr self-hosted<\/td>\n<td align=\"left\">JIRA is the tracker for teams planning and building great products. Thousands of teams choose JIRA to capture and organize issues, assign work, and follow team activity. At your desk or on the go with the new mobile interface, JIRA helps your team get the job done.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/lighthouseapp.com\" rel=\"nofollow\">Lighthouse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/lighthouseapp\" rel=\"nofollow\">@lighthouseapp<\/a><\/td>\n<td align=\"left\">$25\/mo - $100\/mo<\/td>\n<td align=\"left\">Whether you're a large company or a small bootstrapped team, Lighthouse is the perfect ticket tracking solution. Collaborate effortlessly on projects. Whether you\u2019re a team of 5 or studio of 50, Lighthouse will help you keep track of your project development with ease.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pinitto.me\" rel=\"nofollow\">Pinitto.me<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Post It Notes on a virtual corkboard (OSS)<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sifterapp.com\" rel=\"nofollow\">Sifter<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sifterapp\" rel=\"nofollow\">@sifterapp<\/a><\/td>\n<td align=\"left\">$29\/mo - $149\/mo<\/td>\n<td align=\"left\">Less configuring. More doing. We've put in the time researching bug tracking to help create the simplest possible workflow for you to get work done. From time-to-time we even blog about some of our ideas around this optimal bug tracking process\u2026<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/usersnap.com\" rel=\"nofollow\">Usersnap<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/usersnap\" rel=\"nofollow\">@usersnap<\/a><\/td>\n<td align=\"left\">$19\/mo-$99\/mo<\/td>\n<td align=\"left\">Usersnap is visual bug reporting for anyone working on web projects. Get visual feedback and precious browser information with every bug report to reproduce and fix them even faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/instabug.com\" rel=\"nofollow\">Instabug<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/instabug\" rel=\"nofollow\">@instabug<\/a><\/td>\n<td align=\"left\">$49\/mo-$349\/mo<\/td>\n<td align=\"left\">Instabug is comprehensive bug reporting and in-app feedback tool for mobile apps. Receive detailed feedback attached with every bug report sent by your testers or users to debug  and build better apps, faster.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-planning--project-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#planning--project-management\"><\/a>Planning &amp; Project Management<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.aha.io\" rel=\"nofollow\">Aha!<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/aha_io\" rel=\"nofollow\">@aha_io<\/a><\/td>\n<td align=\"left\">$69\/mo, Ask about startup plan<\/td>\n<td align=\"left\">The new way to create brilliant product strategy and visual roadmaps.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sprint.ly\" rel=\"nofollow\">Sprintly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sprintly\" rel=\"nofollow\">@sprintly<\/a><\/td>\n<td align=\"left\">$49\/mo- $399\/mo<\/td>\n<td align=\"left\">Don't ask how projects are going. Watch how they're going in real-time. Use our elegant interface to prioritize, tag, manage, estimate, and measure your software developers' progress in real-time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/podio.com\" rel=\"nofollow\">Podio<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Podio\" rel=\"nofollow\">@Podio<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Teamwork made easy. News and views from the team behind Podio - changing the way the world works since 2009.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getflow.com\" rel=\"nofollow\">Flow<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/flowapp\" rel=\"nofollow\">@flowapp<\/a><\/td>\n<td align=\"left\">$19\/mo - $249\/mo<\/td>\n<td align=\"left\">Stop managing projects from your inbox. Flow is a collaborative task management app for the web and iPhone.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/basecamp.com\" rel=\"nofollow\">Basecamp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/37signals\" rel=\"nofollow\">@37signals<\/a><\/td>\n<td align=\"left\">$20\/mo - $150\/mo<\/td>\n<td align=\"left\">The official account for Basecamp\u00ae. Helping Basecamp customers every Mon-Fri 9am-6pm CT! Basecamp to help organize the store design, develop fixtures, and manage craftspeople. Primarily through word-of-mouth alone, Basecamp has become the world\u2019s #1 project management tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.apollohq.com\" rel=\"nofollow\">Apollo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/applicomhq\" rel=\"nofollow\">@applicomhq<\/a><\/td>\n<td align=\"left\">$23\/mo - $148\/mo<\/td>\n<td align=\"left\">Integrated Project and Contact Management Done Right Apollo is project and contact management done right. Using Apollo, you will realise that it's built to help you get things done, quickly and efficiently. With Apollo, you will always know where your projects, your contacts and your life are at and you will feel on top of everything \u2014 regardless of how hectic your schedule is.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pivotaltracker.com\" rel=\"nofollow\">Pivotal Tracker<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/pivotaltracker\" rel=\"nofollow\">@pivotaltracker<\/a><\/td>\n<td align=\"left\">$7\/mo - $175\/mo<\/td>\n<td align=\"left\">BUILD BETTER SOFTWARE FASTER. Break your project down into bite-sized stories, which get your product closer to the business goal. Use points to estimate each story\u2019s relative complexity and prioritize it in the backlog.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/asana.com\" rel=\"nofollow\">Asana<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/asana\" rel=\"nofollow\">@asana<\/a><\/td>\n<td align=\"left\">$50\/mo - $800\/mo<\/td>\n<td align=\"left\">Teamwork without email. Asana is our go-to for prioritizing projects, keeping up w\/orders &amp; staying on top of a growing to-do list<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/weekplan.net\" rel=\"nofollow\">WeekPlan<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/weekplan\" rel=\"nofollow\">@weekplan<\/a><\/td>\n<td align=\"left\">$7\/mo - $19\/mo<\/td>\n<td align=\"left\">Time management inspired by the \"7 habits of highly effective people\" Features: goals of the week, week view and quadrant matrix, pomodoro timer, shared workspaces, etc...<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/trello.com\" rel=\"nofollow\">Trello<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/trello\" rel=\"nofollow\">@trello<\/a><\/td>\n<td align=\"left\">$5\/mo<\/td>\n<td align=\"left\">Organize anything, together Trello is the fastest, easiest way to organize anything, from your day-to-day work, to a favorite side project, to your greatest life plans.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.blossom.co\" rel=\"nofollow\">Blossom<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/blossom_io\" rel=\"nofollow\">@blossom_io<\/a><\/td>\n<td align=\"left\">$19\/mo - $149\/mo<\/td>\n<td align=\"left\">Agile Project Management. Blossom gives each member of the team clear overview about who\u2019s doing what &amp; why and at the same time it helps you to focus on what matters most. With Blossom you can efficiently manage your whole development process in one place, built with simplicity in mind. Blossom is based on the principles ofKanban, a way of working that emphasizes iterative delivery cycles and continuously improves the workflow of your team or organization.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.redmine.org\" rel=\"nofollow\">Redmine<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Flexible project management web application. Written using Ruby on Rails framework, it is cross-platform and cross-database.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/jira\/agile\" rel=\"nofollow\">JIRA Agile<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/JIRA\" rel=\"nofollow\">@jira<\/a><\/td>\n<td align=\"left\">$10\/mo - $30\/mo<\/td>\n<td align=\"left\">Dream big, work smart, deliver fast. Makers of @JIRA, @Confluence, @Bitbucket and more. Software to plan, collaborate, code, and support. Built for teams<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tomsplanner.com\" rel=\"nofollow\">Tom's Planner<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tomsplanner\" rel=\"nofollow\">@tomsplanner<\/a><\/td>\n<td align=\"left\">$9\/mo - $19\/mo<\/td>\n<td align=\"left\">Tom's Planner is online Gantt chart software that allows anyone to create, collaborate and share Gantt Charts online with drag and drop simplicity. It's web based, extremely intuitive and easy-to-use.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/leankit.com\" rel=\"nofollow\">LeanKit<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/leankit\" rel=\"nofollow\">@leankit<\/a><\/td>\n<td align=\"left\">$15\/mo - $19\/mo<\/td>\n<td align=\"left\">Instant Project visibility. In LeanKit, you map your organization\u2019s processes onto virtual whiteboards. On each board the process steps are represented as vertical and horizontal lanes. Cards represent work items, which team members update and move from across the board as they complete their share of the work. Rather than having to ask for status reports, managers and customers can just look at the board. Board updates are visible in seconds around the globe and e-mail alerts and RSS feeds are available, so you and your team can take immediate action to resolve issues before they turn into serious problems.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.breeze.pm\" rel=\"nofollow\">Breeze<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/BreezeTeam\" rel=\"nofollow\">@BreezeTeam<\/a><\/td>\n<td align=\"left\">$29\/mo - $129\/mo<\/td>\n<td align=\"left\">Organize and track everything. Breeze shows you what's being worked on, who's working on what, where things are in the workflow and how much time it took.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/cushionapp.com\/\" rel=\"nofollow\">Cushion<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/cushionapp\" rel=\"nofollow\">@cushionapp<\/a><\/td>\n<td align=\"left\">$8\/mo - $48\/mo<\/td>\n<td align=\"left\">Piece of mind for freelancers. We built Cushion as a team of freelancers who always had trouble planning and managing our year\u2014constantly taking on too much work and never thinking to take a break.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/clickup.com\" rel=\"nofollow\">Clickup<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/clickup_app\" rel=\"nofollow\">@clickup_app<\/a><\/td>\n<td align=\"left\">$0 - $5\/mo<\/td>\n<td align=\"left\">ClickUp's #1 rated productivity software is making more productive projects with a beautifully designed and intuitive platform.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-time-tracking\" class=\"anchor\" aria-hidden=\"true\" href=\"#time-tracking\"><\/a>Time Tracking<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/toggl.com\" rel=\"nofollow\">Toggl<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/toggl\" rel=\"nofollow\">@toggl<\/a><\/td>\n<td align=\"left\">Free - $59\/mo<\/td>\n<td align=\"left\">Toggl\u2019s time tracker is built for speed and ease of use. Time logging with Toggl is so simple that you\u2019ll actually use it.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/clockify.me\" rel=\"nofollow\">Clockify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/clockify\" rel=\"nofollow\">@clockify<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Clockify is the only 100% Free time tracking software that lets you and your team track time with one click. It works like Toggl but offers unlimited features and unlimited users.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hubstaff.com\" rel=\"nofollow\">Hubstaff<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Hubstaff\" rel=\"nofollow\">@hubstaff<\/a><\/td>\n<td align=\"left\">Free - $9\/mo<\/td>\n<td align=\"left\">Hubstaff is time tracking software designed to make remote team management more effective and efficient. You just have to sign up, download our intuitive desktop app and push the start button to begin tracking time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tickspot.com\/\" rel=\"nofollow\">Tickspot<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tickspot\" rel=\"nofollow\">@tickspot<\/a><\/td>\n<td align=\"left\">Free - $149\/mo<\/td>\n<td align=\"left\">Straightforward time tracking software to help your team run more profitable projects. Whether you prefer iOS, Android, the Apple Watch, or your desktop computer, Tick is the easiest way to track your time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.kimai.org\/\" rel=\"nofollow\">Kimai<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kimai_org\" rel=\"nofollow\">@kimai_org<\/a><\/td>\n<td align=\"left\">self-hosted<\/td>\n<td align=\"left\">Kimai is a Free open source timetracker. It tracks work time and prints out a summary of your activities on demand. Yearly, monthly, daily, by customer, by project, by action, etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/samg\/timetrap\">Timetrap<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">simple open source command line time tracking<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-app-developer-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#app-developer-tools\"><\/a>App Developer Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.appannie.com\" rel=\"nofollow\">App Annie<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/appannie\/\" rel=\"nofollow\">@appannie\/<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Make better decisions with our app store data. We are the leader for app store analytics, app rankings, and market intelligence.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/appfigures.com\" rel=\"nofollow\">App Figures<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/appfigures\" rel=\"nofollow\">@appfigures<\/a><\/td>\n<td align=\"left\">$9\/mo<\/td>\n<td align=\"left\">A beautiful app tracking platform for app developers and publishers.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-localization--internationalization\" class=\"anchor\" aria-hidden=\"true\" href=\"#localization--internationalization\"><\/a>Localization &amp; Internationalization<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/localizejs.com\" rel=\"nofollow\">Localize.js<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$25\/mo - $150\/mo<\/td>\n<td align=\"left\">Website translation with just a few lines of code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gengo.com\" rel=\"nofollow\">Gengo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0.06 - $0.17\/word<\/td>\n<td align=\"left\">Human translation API<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/translate.google.com\/manager\/website\/\" rel=\"nofollow\">Google Website Translator<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Make your website instantly available in 90+ languages<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-business--traffic-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#business--traffic-analytics\"><\/a>Business &amp; Traffic Analytics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kissmetricshq.com\" rel=\"nofollow\">KISSmetrics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kissmetrics\/\" rel=\"nofollow\">@kissmetrics\/<\/a><\/td>\n<td align=\"left\">$150\/mo - $500\/mo<\/td>\n<td align=\"left\">KISSmetrics tells you who\u2019s doing it. Every last piece gets connected to a real person. All of it. It doesn\u2019t matter if people bounce around between different browsers and devices. Or even if it takes them 6 months to come back. You\u2019ll see what real people do.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.localytics.com\" rel=\"nofollow\">Localytics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/localytics\/\" rel=\"nofollow\">@localytics\/<\/a><\/td>\n<td align=\"left\">Free up to 10k MAUs, $200\/mo - $2700\/mo above that<\/td>\n<td align=\"left\">Find out what works in your mobile or web app. Do more of it. All in one place. Advanced analytics provide data and insight to help you build more successful apps. Integrated Marketing helps you easily engage and acquire more customers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mixpanel.com\" rel=\"nofollow\">Mixpanel<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mixpanel\/\" rel=\"nofollow\">@mixpanel\/<\/a><\/td>\n<td align=\"left\">$150\/mo - $2000\/mo<\/td>\n<td align=\"left\">Actions speak louder than page views. For years, companies have pushed page views as a primary measure of success. Page view counts are popular because they are easy to report, but ultimately cannot tell you how engaged your visitors are. Mixpanel lets you measure what customers do in your app by reporting actions, not page views.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/amplitude.com\" rel=\"nofollow\">Amplitude<\/a><\/td>\n<td align=\"left\">$299\/mo<\/td>\n<td align=\"left\">Mobile Analytics for decision makers<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/snowplowanalytics.com\" rel=\"nofollow\">Snowplow<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SnowPlowData\" rel=\"nofollow\">@SnowPlowData<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Snowplow is the most powerful, flexible, scalable web analytics platform in the world. Snowplow enables analysts to perform a wide variety of both simple and sophisticated analytics on your web analytics data.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/segment.com\" rel=\"nofollow\">Segment<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$29\/mo - $349\/mo<\/td>\n<td align=\"left\">The right way to manage your tools. The idea is simple: one pipeline for all your data. Send data to any third-party tool with a single integration.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/clicky.com\" rel=\"nofollow\">Clicky<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/clicky\" rel=\"nofollow\">@clicky<\/a><\/td>\n<td align=\"left\">$9.99\/mo - $19.99\/mo<\/td>\n<td align=\"left\">Real Time Web Analytics. Clicky lets you see every visitor and every action they take on your web site, with the option to attach custom data to visitors, such as usernames or email addresses. Analyze each visitor individually and see their full history.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.google.com\/analytics\/\" rel=\"nofollow\">Google Analytics<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Google Analytics lets you measure your advertising ROI as well as track your Flash, video, and social networking sites and applications.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/matomo.org\" rel=\"nofollow\">Matomo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/matomo_org\" rel=\"nofollow\">@matomo_org<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Whether you are an individual blogger, a small business, or a large corporation, Matomo helps you gain valuable insights to help your business or readership grow.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/chartio.com\" rel=\"nofollow\">Chartio<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chartio\" rel=\"nofollow\">@chartio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Visualize and explore your data with Chartio. Create interactive charts and perfect dashboards through an intuitive drag and drop interface. Switch from basic tables to sophisticated data visualizations in a single click. Powerful filters let you slice and dice your data, and you can drill down into most charts without configuring a thing.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/chartbeat.com\" rel=\"nofollow\">Chartbeat<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chartbeat\" rel=\"nofollow\">@chartbeat<\/a><\/td>\n<td align=\"left\">$9.95\/mo - $49.95\/mo<\/td>\n<td align=\"left\">Build a loyal and valuable audience for your site. Chartbeat's real-time traffic and audience-behavior data shows you who's on your site and how they're engaging with your content right now \u2014 so you can take action on what matters when it matters.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/calq.io\" rel=\"nofollow\">Calq<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/CalqAnalytics\" rel=\"nofollow\">@CalqAnalytics<\/a><\/td>\n<td align=\"left\">$0 - $2500\/mo<\/td>\n<td align=\"left\">Advanced custom analytics for mobile and web applications. Calq is an analytics platform that measures user actions rather than page views. An action can be anything a user does: reviewing a product, playing a level on a mobile game, making a purchase on your site, anything. Calq's ability to work with custom events AND custom data is what raises it above more traditional analytics platforms.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.gosquared.com\" rel=\"nofollow\">GoSquared<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/GoSquared\" rel=\"nofollow\">@gosquared<\/a><\/td>\n<td align=\"left\">\u00a321.60 - \u00a3396\/mo<\/td>\n<td align=\"left\">Easy to use real-time web analytics.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.improvely.com\" rel=\"nofollow\">Improvely<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/improvelycom\" rel=\"nofollow\">@improvelycom<\/a><\/td>\n<td align=\"left\">$29 - $899\/mo<\/td>\n<td align=\"left\">Conversion tracking and click fraud monitoring platform. The easiest way to track the performance of marketing campaigns and monitor them for click fraud.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/keen.io\" rel=\"nofollow\">Keen IO<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/keen_io\" rel=\"nofollow\">@keen_io<\/a><\/td>\n<td align=\"left\">$0 - $2000+\/mo<\/td>\n<td align=\"left\">Custom analytics shouldn't be a pain in the backend. Keen IO's powerful APIs do the heavy lifting for you, so you can gather all the data you want and start getting the answers you need.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/heapanalytics.com\" rel=\"nofollow\">Heap Analytics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/heap\" rel=\"nofollow\">@heap<\/a><\/td>\n<td align=\"left\">0 - $599+<\/td>\n<td align=\"left\">Instant, retroactive analytics for web and iOS. No code required.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/get.gaug.es\" rel=\"nofollow\">Gauges<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gaugesapp\" rel=\"nofollow\">@GaugesApp<\/a><\/td>\n<td align=\"left\">$6-$48\/mo<\/td>\n<td align=\"left\">Gauges provides real time web analytics such as how many people visit your site, where they come from, and where they go.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getWisdom.io\" rel=\"nofollow\">Wisdom<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free - $2000+\/mo<\/td>\n<td align=\"left\">Session Replay. Wisdom is the most accurate live visitor session recorder service available. Focusing only on session replay, Wisdom reconstructs a virtual desktop screen for every visitor, across all tabs, to capture the true feel of every visitor's experience.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-conversion-optimization--ab-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#conversion-optimization--ab-testing\"><\/a>Conversion Optimization &amp; A\/B Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.optimizely.com\" rel=\"nofollow\">Optimizely<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/optimizely\" rel=\"nofollow\">@optimizely<\/a><\/td>\n<td align=\"left\">$17\/mo - $359\/mo<\/td>\n<td align=\"left\">A\/B testing you'll actually use. Track engagement, clicks, conversions, sign ups, or anything else that matters to you and your business. Optimizely's custom goal tracking provides an endless range of measurable actions that you can define. Just tell Optimizely what to measure, and we will do the rest.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/vwo.com\" rel=\"nofollow\">Visual Website Optimizer<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wingify\" rel=\"nofollow\">@wingify<\/a><\/td>\n<td align=\"left\">$49\/mo - $129\/mo<\/td>\n<td align=\"left\">Increase your website sales and conversions. Using Visual Website Optimizer, they A\/B test different versions of their website and landing pages to find out which one works best. Made for marketers, our tool is incredibly easy to use, and doesn't need IT resources.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.eyequant.com\" rel=\"nofollow\">EyeQuant<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/eyequant\" rel=\"nofollow\">@eyequant<\/a><\/td>\n<td align=\"left\">$199\/mo - $999\/mo<\/td>\n<td align=\"left\">Instantly understand what your visitors will see and miss in their first seconds on your site, and improve your conversions. Analyse live sites or mockups within seconds, no code required.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/optimize.google.com\" rel=\"nofollow\">Optimize by Google<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Optimize allows you to test variants of web pages and see how they perform against an objective that you specify.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-user-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-management\"><\/a>User Management<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/developer.okta.com\/\" rel=\"nofollow\">Okta<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/OktaDev\" rel=\"nofollow\">@OktaDev<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Okta adds authentication, authorization, and user management to your web or mobile app within minutes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/auth0.com\" rel=\"nofollow\">Auth0<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/c2id.net\" rel=\"nofollow\">Connect2id<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/connect2id\" rel=\"nofollow\">@Connect2id<\/a><\/td>\n<td align=\"left\">\u20ac299\/mo - \u20ac999\/mo<\/td>\n<td align=\"left\">Hosted Connect2id server \u2013 a flexible, secure and certified OpenID Connect\/OAuth 2.0 identity provider.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-user-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-testing\"><\/a>User Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/silverbackapp.com\" rel=\"nofollow\">Silverback 2.0<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/silverbackapp\" rel=\"nofollow\">@silverbackapp<\/a><\/td>\n<td align=\"left\">69.95<\/td>\n<td align=\"left\">Guerrilla usability testing software for designers and developers. Silverback makes it easy, quick and cheap for everyone to perform guerrilla usability tests with no setup and no expense, using hardware already in your Mac.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hotjar.com\" rel=\"nofollow\">HotJar<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/hotjar\" rel=\"nofollow\">@hotjar<\/a><\/td>\n<td align=\"left\">Free - $29\/mo (personal)<\/td>\n<td align=\"left\">Records videos and collects heatmaps of your site visitor actions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getWisdom.io\" rel=\"nofollow\">Wisdom<\/a><\/td>\n<td align=\"left\">Free - $2000+\/mo<\/td>\n<td align=\"left\">Session Replay<\/td>\n<td align=\"left\">Wisdom is the most accurate live visitor session recorder service available. Focusing only on session replay, Wisdom reconstructs a virtual desktop screen for every visitor, across all tabs, to capture the true feel of every visitor's experience.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-hr\" class=\"anchor\" aria-hidden=\"true\" href=\"#hr\"><\/a>HR<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.workday.com\" rel=\"nofollow\">Workday<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/workday\" rel=\"nofollow\">@workday<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Workday works the way people work\u2014collaboratively, on the go, and in real-time. Explore the product previews below to learn how Workday can change the way you work. With powerful business applications and a user experience that's unmatched in enterprise software, Workday gives you everything you need to transform your business.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.lever.co\" rel=\"nofollow\">Lever<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/lever\" rel=\"nofollow\">@lever<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">A modern web app for hiring. Leverage your entire company \u2013 interviewers, managers, and recruiters \u2013 to source, vet, and close.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zenefits.com\" rel=\"nofollow\">Zenefits<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zenefits\" rel=\"nofollow\">@zenefits<\/a><\/td>\n<td align=\"left\">$0\/mo<\/td>\n<td align=\"left\">The #1 All-In-One HR Platform. Payroll. Benefits. Time. Compliance. All online, all in one place.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.testdome.com\/\" rel=\"nofollow\">TestDome<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/TestDome\" rel=\"nofollow\">@TestDome<\/a><\/td>\n<td align=\"left\">$8\/candidate - $20\/candidate<\/td>\n<td align=\"left\">Automated testing of programming skills, ask candidates to write real code before calling them for an interview.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hackerrank.com\/\" rel=\"nofollow\">HackerRank<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/hackerrank\" rel=\"nofollow\">@hackerrank<\/a><\/td>\n<td align=\"left\">paid<\/td>\n<td align=\"left\">End-to-end technical recruiting platform for hiring engineers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/peopledoc.com\" rel=\"nofollow\">PeopleDoc<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PeopleDoc_Inc\" rel=\"nofollow\">@PeopleDoc_Inc<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">HR Service delivery platform designed to simplify complex HR operations, compliance and can easily serve employees anytime, anywhere.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.bamboohr.com\/\" rel=\"nofollow\">BambooHR<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bamboohr\" rel=\"nofollow\">@bamboohr<\/a><\/td>\n<td align=\"left\">paid<\/td>\n<td align=\"left\">HR software with heart.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/peopledoc.com\" rel=\"nofollow\">PeopleDoc<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PeopleDoc_Inc\" rel=\"nofollow\">@PeopleDoc_Inc<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">HR Service delivery platform designed to simplify complex HR operations, compliance and can easily serve employees anytime, anywhere.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.bamboohr.com\" rel=\"nofollow\">BambooHR<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bamboohr\" rel=\"nofollow\">@bamboohr<\/a><\/td>\n<td align=\"left\">paid<\/td>\n<td align=\"left\">HR software with heart.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hiringplan.io\" rel=\"nofollow\">HiringPlan<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">HiringPlan is a headcount planning tool for startups with built-in market data. It helps founders adopt a structured compensation plan, pay people fairly and preserve cash and equity runway.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/tldroptions.io\" rel=\"nofollow\">TLDROptions<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">A tool for helping employees understand the potential value of their stock options.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-payroll\" class=\"anchor\" aria-hidden=\"true\" href=\"#payroll\"><\/a>Payroll<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/gusto.com\" rel=\"nofollow\">Gusto<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gustohq\" rel=\"nofollow\">@gustohq<\/a><\/td>\n<td align=\"left\">$29\/mo + $6\/user<\/td>\n<td align=\"left\">Payroll and benefits that put people first, easy setup, automated tax filings and thoughtful support.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/wagepoint.com\" rel=\"nofollow\">WagePoint<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wagepoint\" rel=\"nofollow\">@wagepoint<\/a><\/td>\n<td align=\"left\">$20 + $2\/employy<\/td>\n<td align=\"left\">The Simple, Fast &amp; Friendly way to pay your employees.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-continuous-integrationcode-quality\" class=\"anchor\" aria-hidden=\"true\" href=\"#continuous-integrationcode-quality\"><\/a>Continuous Integration\/Code Quality<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/travis-ci.org\" rel=\"nofollow\">Travis<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/travisci\" rel=\"nofollow\">@travisci<\/a><\/td>\n<td align=\"left\">Free - $489\/mo<\/td>\n<td align=\"left\">Hi I\u2019m Travis CI, a hosted continuous integration service for open source and private projects: <a href=\"https:\/\/travis-ci.com\/\" rel=\"nofollow\">travis-ci.com<\/a>  System status updates: <a href=\"https:\/\/twitter.com\/traviscistatus\" rel=\"nofollow\">@traviscistatus<\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.appveyor.com\" rel=\"nofollow\">AppVeyor<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/appveyor\" rel=\"nofollow\">@appveyor<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">AppVeyor automates building, testing and deployment of .NET applications, helping your team to focus on delivering great apps.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/codeship.com\" rel=\"nofollow\">Codeship<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/codeship\" rel=\"nofollow\">@codeship<\/a><\/td>\n<td align=\"left\">Start with 100 builds\/mo Free, Unlimited plans start at 49$<\/td>\n<td align=\"left\">Continuous Delivery as a service, start testing and deploying your code immediately<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/circleci.com\" rel=\"nofollow\">Circle<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/circleci\" rel=\"nofollow\">@circleci<\/a><\/td>\n<td align=\"left\">$19\/mo - $269\/mo<\/td>\n<td align=\"left\">Ship better code, faster. Easy, fast, continuous integration and deployment for web apps.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/nevercode.io\" rel=\"nofollow\">Nevercode<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/nevercodehq\" rel=\"nofollow\">@nevercodehq<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/houndci.com\" rel=\"nofollow\">Hound<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/houndci\" rel=\"nofollow\">@houndci<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Automated Code Review. Take care of pesky code reviews with a trusty Hound<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/codeclimate.com\" rel=\"nofollow\">CodeClimate<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/codeclimate\" rel=\"nofollow\">@codeclimate<\/a><\/td>\n<td align=\"left\">$0\/mo - $399\/mo<\/td>\n<td align=\"left\">Automated  Code Review. Code Climate hosted software metrics help you ship quality Ruby and JavaScript code faster. Get control of your technical debt with real time static analysis of your code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.codacy.com\" rel=\"nofollow\">Codacy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0 - $150\/mo<\/td>\n<td align=\"left\">Automated Code Review. Continuous Static Analysis designed to complement your unit tests. Similar to CodeClimate.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/codecov.io\" rel=\"nofollow\">Codecov<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0 - $5\/mo<\/td>\n<td align=\"left\">Hosted Code Coverage. Code coverage reporting done right.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/semaphoreci.com\" rel=\"nofollow\">Semaphore<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/semaphoreci\" rel=\"nofollow\">@semaphoreci<\/a><\/td>\n<td align=\"left\">$14\/mo - $899\/mo<\/td>\n<td align=\"left\">Create an Amazing Workflow. Semaphore assumes that your private or open source project is on GitHub. There are no new dependencies, hooks or SSH keys to manage. It works without any change in source code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.solanolabs.com\" rel=\"nofollow\">Solano CI<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/solanolabs\" rel=\"nofollow\">@SolanoLabs<\/a><\/td>\n<td align=\"left\">$15\/mo - $100\/mo<\/td>\n<td align=\"left\">Faster Continuous Integration and Deployment with patented auto-parallelization. Solano CI sets up Continuous Integration in minutes, Frees you from managing a build server, and lets you deploy software 10x - 80x faster by running tests in parallel safely and automatically. It also lets you use our massively scalable environment even before you push to CI. Seamlessly integrates into existing workflows. Free 14-day trial, no credit card required. Formerly loved as tddium.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/jenkins.io\" rel=\"nofollow\">Jenkins<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/jenkinsci\" rel=\"nofollow\">@jenkinsci<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Jenkins provides continuous integration services for software development. It is a server-based system that supports SCM tools including AccuRev, CVS, Subversion, Git, Mercurial, Perforce, Clearcase and RTC, and can execute Apache Ant and Apache Maven based projects as well as arbitrary shell scripts and Windows batch commands. Released under the MIT License, Jenkins is Free software.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/bamboo\" rel=\"nofollow\">Bamboo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/atlassian\" rel=\"nofollow\">@atlassian<\/a><\/td>\n<td align=\"left\">$10\/mo - $1000\/mo<\/td>\n<td align=\"left\">Bamboo does more than just run builds and tests. It connects issues, commits, test results, and deploys so the whole picture is available to your entire product team \u2013 from project managers, to devs and testers, to sys admins.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/buildkite.com\" rel=\"nofollow\">Buildkite (Buildbox)<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/buildkite\" rel=\"nofollow\">@buildkite<\/a><\/td>\n<td align=\"left\">$15\/dev\/mo<\/td>\n<td align=\"left\">Semi-hosted continuous integration and deployment. Buildkite uses your own infrastructure to run builds so you can test any language or run any deployment scripts. You can run as many parallel agents (and builds) as you want.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/crucible\" rel=\"nofollow\">Crucible<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/atlassian\" rel=\"nofollow\">@atlassian<\/a><\/td>\n<td align=\"left\">$10\/mo - $8000\/mo<\/td>\n<td align=\"left\">Code review system. Review code, discuss changes, share knowledge, and identify defects with Crucible's flexible review workflow. It's code review made easy for Git, Subversion, CVS, Perforce, and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/coveralls.io\" rel=\"nofollow\">Coveralls<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/coverallsapp\" rel=\"nofollow\">@coverallsapp<\/a><\/td>\n<td align=\"left\">$0\/mo - $50\/mo<\/td>\n<td align=\"left\">Coveralls works with your continuous integration server to give you test coverage history and statistics. It integrates with any langauge and is Free for open source.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/about.testributor.com\" rel=\"nofollow\">Testributor<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/testributor\" rel=\"nofollow\">@testributor<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Testributor is an open source Continuous Integration platform. A hosted version is available for Free, both for open source and private projects.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.wercker.com\" rel=\"nofollow\">Wercker<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wercker\" rel=\"nofollow\">@wercker<\/a><\/td>\n<td align=\"left\">$0\/mo - $350\/mo<\/td>\n<td align=\"left\">Wercker is a Docker-Native CI\/CD Automation platform for Kubernetes &amp; Microservice Deployments.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/monkeytest.it\" rel=\"nofollow\">Monkey Test It<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/monkeytestit\" rel=\"nofollow\">@monkeytestit<\/a><\/td>\n<td align=\"left\">$0\/mo - $199\/mo<\/td>\n<td align=\"left\">Automated website checking tool to catch bugs on your website (broken links, missing images, JS errors, etc) before your users do. Integrates with Slack, most CI systems, webooks, and supports built-in scheduling.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/concourse-ci.org\" rel=\"nofollow\">Concourse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/concourseci\" rel=\"nofollow\">@concourseci<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Concourse is an open-source continuous thing-doer. Built on the simple mechanics of resources, tasks, and jobs, Concourse presents a general approach to automation that makes it great for CI\/CD.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-dashboards\" class=\"anchor\" aria-hidden=\"true\" href=\"#dashboards\"><\/a>Dashboards<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.geckoboard.com\" rel=\"nofollow\">Geckoboard<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/geckoboard\" rel=\"nofollow\">@geckoboard<\/a><\/td>\n<td align=\"left\">$17\/mo - $899\/mo<\/td>\n<td align=\"left\">Meet Geckoboard. It's Your Key Data, In One Place. Geckoboard monitors your business\u2019s vital signs \u2013 don\u2019t wait, see it live on a business dashboard as it happens. Focus on what matters and react faster to important events.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.telemetrytv.com\" rel=\"nofollow\">Telemetry<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/telemetrytv\" rel=\"nofollow\">@telemetrytv<\/a><\/td>\n<td align=\"left\">$9\/mo - $749\/mo<\/td>\n<td align=\"left\">Build realtime dashboard with powerful visualizations that look beautiful on big screen TVs, desktop computers, mobile devices, and embedded systems\u2014all using a simple REST API that works with all modern languages.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/dashing.io\" rel=\"nofollow\">Dashing<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Dashing is a Sinatra based framework that lets you build beautiful dashboards.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.klipfolio.com\" rel=\"nofollow\">Klipfolio<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/klipfolio\" rel=\"nofollow\">@klipfolio<\/a><\/td>\n<td align=\"left\">$5\/user\/mo - $20\/user\/mo<\/td>\n<td align=\"left\">Meet Your Business Dashboard. Connect to any data service to bring your key numbers together on one dashboard. Assign your data to visualizations to show the story behind the numbers. Cultivate a data-driven culture by sharing dashboards with everyone on your team.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/grafana.com\" rel=\"nofollow\">Grafana<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/grafana\" rel=\"nofollow\">@grafana<\/a><\/td>\n<td align=\"left\">$0 - $90\/mo (+9\/user\/mo)<\/td>\n<td align=\"left\">The analytics platform for all your metrics. Grafana allows you to query, visualize, alert on and understand your metrics no matter where they are stored. Create, explore, and share dashboards with your team and foster a data driven culture.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/redash.io\" rel=\"nofollow\">Redash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Self-hosted or $49\/mo - 450\/mo<\/td>\n<td align=\"left\">\"Redash has support for querying multiple databases, including: Redshift, Google BigQuery, PostgreSQL, MySQL, Graphite, Presto, Google Spreadsheets, Cloudera Impala, Hive and custom scripts.\"<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cyfe.com\" rel=\"nofollow\">Cyfe<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Cyfe\" rel=\"nofollow\">@cyfe<\/a><\/td>\n<td align=\"left\">$0 - $29\/mo<\/td>\n<td align=\"left\">Cyfe offers all-in-one web dashboards for your clients to help them monitor and analyze data scattered across all their online services like Google Analytics, Salesforce, AdSense, MailChimp, Amazon, Facebook, WordPress, Twitter and more from one single location in real-time.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-errorexception-handling\" class=\"anchor\" aria-hidden=\"true\" href=\"#errorexception-handling\"><\/a>Error\/Exception Handling<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/catchjs.com\" rel=\"nofollow\">CatchJS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$49 - $499<\/td>\n<td align=\"left\">JavaScript error reporting for web apps. Can take screenshots and collect click trails, so reproducing bugs becomes easy.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/try.crashlytics.com\" rel=\"nofollow\">Crashlytics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/crashlytics\" rel=\"nofollow\">@crashlytics<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Crash reports and grouping for easier analysis. Basic analytics and reports. iOS &amp; Android<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sentry.io\/welcome\/\" rel=\"nofollow\">Sentry<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getsentry\" rel=\"nofollow\">@getsentry<\/a><\/td>\n<td align=\"left\">Open Source &amp; $24\/mo - $199\/mo<\/td>\n<td align=\"left\">Sentry notifies you when your users experience errors. Know immediately when things happen in your application. Engage users before they have a chance to report a problem.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.honeybadger.io\" rel=\"nofollow\">HoneyBadger<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/honeybadgerapp\" rel=\"nofollow\">@honeybadgerapp<\/a><\/td>\n<td align=\"left\">$39\/mo - $249\/mo<\/td>\n<td align=\"left\">Exception, uptime, and performance monitoring for Ruby. It tells you about errors, downtime and performance issues as they happen. And it gives you the tools you need to fix them ...without burying you in data. Without silly rate limits or per-server fees.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.bugsnag.com\" rel=\"nofollow\">BugSnag<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bugsnag\" rel=\"nofollow\">@bugsnag<\/a><\/td>\n<td align=\"left\">$29\/mo - $249\/mo<\/td>\n<td align=\"left\">Automatic, full-stack error monitoring. Web app monitoring for Rails, PHP, Node.js, Java, and every other leading platform.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/raygun.com\" rel=\"nofollow\">Raygun<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/raygunio\" rel=\"nofollow\">@raygunio<\/a><\/td>\n<td align=\"left\">$14\/mo - $199\/mo<\/td>\n<td align=\"left\">Exceptional Error Tracking. Your software faults get automatically sent to the Raygun service and analysis begins immediately. Raygun intelligently groups your errors so you're dealing with root causes, not every single error instance!<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/airbrake.io\" rel=\"nofollow\">Airbrake<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/airbrake\" rel=\"nofollow\">@airbrake<\/a><\/td>\n<td align=\"left\">$49 - $249\/mo<\/td>\n<td align=\"left\">No More Searching Log Files Capture and Track Your Application's Exceptions in 3 Minutes. Airbrake is the leading exception reporting service, currently providing error tracking for 50,000 applications with support for 18 programming languages.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/rollbar.com\" rel=\"nofollow\">Rollbar<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/rollbar\" rel=\"nofollow\">@rollbar<\/a><\/td>\n<td align=\"left\">$12\/mo - $1249\/mo<\/td>\n<td align=\"left\">Take control of your errors. Rollbar is platform-agnostic and can accept data from anything that can speak HTTP and JSON. You can use our official libraries for Ruby, Python, PHP, Node.js, JavaScript, Android, iOS, or Flash, or roll your own with our API.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/errorception.com\" rel=\"nofollow\">Errorception<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/errorception\" rel=\"nofollow\">@errorception<\/a><\/td>\n<td align=\"left\">$5\/mo - $59\/mo<\/td>\n<td align=\"left\">Painless JavaScript Error Tracking. Errorception is a simple and painless way to find out about JavaScript errors, as they occur in your users' browsers. All you need to do is insert a script tag on your page, and you will start recording errors as they happen in real-time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/errbit.com\" rel=\"nofollow\">Errbit<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">OSS<\/td>\n<td align=\"left\">The open source error catcher that's Airbrake API compliant.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.overops.com\" rel=\"nofollow\">OverOps<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/overopshq\" rel=\"nofollow\">@overopshq<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">God Mode in Production Code for java and scala applications.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-application-distribution\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-distribution\"><\/a>Application Distribution<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.hockeyapp.net\" rel=\"nofollow\">HockeyApp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/hockeyapp\" rel=\"nofollow\">@hockeyapp<\/a><\/td>\n<td align=\"left\">Free - $129\/mo depending on number of apps and number of owners<\/td>\n<td align=\"left\">Distribution of iOS, Android, Windows Phone and Mac OS apps. Includes analytics, user feedback and crash reports.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/setapp.com\" rel=\"nofollow\">Setapp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/setapp\" rel=\"nofollow\">@setapp<\/a><\/td>\n<td align=\"left\">$9.99\/mo<\/td>\n<td align=\"left\">Setapp is a single subscription with a huge collection of quality apps for macOS<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-log-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#log-monitoring\"><\/a>Log Monitoring<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.fluentd.org\" rel=\"nofollow\">Fluentd<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/fluentd\" rel=\"nofollow\">@fluentd<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Set Up Once, Collect More. Fluentd is an open source data collector designed for processing data streams. 150+ plugins instantly enable you to store the data for Log Management, Big Data Analytics, etc<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/cloudera\/flume\">Flume<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.graylog.org\" rel=\"nofollow\">Graylog<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/graylog2\" rel=\"nofollow\">@graylog2<\/a><\/td>\n<td align=\"left\">Field-tested open source data analytics system used and trusted all around the world. Search your logs, create charts, send reports and be alerted when something happens. All running on the existing JVM in your datacenter.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/logentries.com\" rel=\"nofollow\">LogEntries<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/logentries\" rel=\"nofollow\">@logentries<\/a><\/td>\n<td align=\"left\">$16\/mo - $245\/mo<\/td>\n<td align=\"left\">Log Management &amp; Analytics Made Easy. Logentries provides an easy-to-use cloud service for log management and analytics.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.loggly.com\" rel=\"nofollow\">Loggly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/loggly\" rel=\"nofollow\">@loggly<\/a><\/td>\n<td align=\"left\">$49\/mo - $349\/mo<\/td>\n<td align=\"left\">Solve operational problems faster. Loggly helps cloud-centric organizations\u2014organizations that build and manage cloud-facing applications\u2014to solve operational problems faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.elastic.co\/products\/logstash\" rel=\"nofollow\">Logstash<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/logstash\" rel=\"nofollow\">@logstash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Ship logs from any source, parse them, get the right timestamp, index them, and search them. Logstash is a tool for managing events and logs. You can use it to collect logs, parse them, and store them for later use (like, for searching). Speaking of searching, Logstash comes with a web interface for searching and drilling into all of your logs.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/papertrailapp.com\" rel=\"nofollow\">Papertrail<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/papertrailapp\" rel=\"nofollow\">@papertrailapp<\/a><\/td>\n<td align=\"left\">$7\/mo - $230\/mo<\/td>\n<td align=\"left\">Frustration-Free log management. Get started in seconds. Use Papertrail's time-saving log tools, flexible system groups, team-wide access, long-term archives, charts and analytics exports, monitoring webhooks, and 45-second setup.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/stackify.com\" rel=\"nofollow\">Stackify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Stackify\" rel=\"nofollow\">@Stackify<\/a><\/td>\n<td align=\"left\">$15\/mo<\/td>\n<td align=\"left\">Connecting the dots for you. Stackify focuses on application health, magnifying critical insights for developers, operations, and support teams.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/etsy\/statsd\/\">statsd<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sumologic.com\" rel=\"nofollow\">Sumo Logic<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SumoLogic\" rel=\"nofollow\">@SumoLogic<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Log Data is Big Data. Once enabled by the administrator, the new data will be searchable in the customer account. Sumo Logic  provides an Application for Data Volume for out-of\u2013the-box dashboards and searches that deliver a comprehensive view of data usage volume by category, collector, source name, and hosts.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-application-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-performance\"><\/a>Application Performance<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.appneta.com\" rel=\"nofollow\">AppNeta<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/AppNeta\" rel=\"nofollow\">@AppNeta<\/a><\/td>\n<td align=\"left\">Free to $119 \/ mo<\/td>\n<td align=\"left\">Full-stack application monitoring for web apps. Get visibility into code, network, and end user, especially for polyglot and service-oriented applications, by looking at transactions, errors, browser metrics, host metrics, and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dripstat.com\" rel=\"nofollow\">DripStat<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dripstat\" rel=\"nofollow\">@DripStat<\/a><\/td>\n<td align=\"left\">$20\/mo<\/td>\n<td align=\"left\">Application monitoring for Java. Next gen Java APM. Full visibility across your stack.  Specificially designed for Java. Supports SQL databases, MongoDB and Cassandra.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/newrelic.com\" rel=\"nofollow\">New Relic<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/NewRelic\" rel=\"nofollow\">@NewRelic<\/a><\/td>\n<td align=\"left\">$149\/mo<\/td>\n<td align=\"left\">Application monitoring for all your web apps. It\u2019s about gaining actionable, real-time business insights from the billions of metrics your software is producing, including user click streams, mobile activity, end user experiences and transactions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/appsignal.com\" rel=\"nofollow\">AppSignal<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/AppSignal\" rel=\"nofollow\">@AppSignal<\/a><\/td>\n<td align=\"left\">$49\/mo - $259\/mo<\/td>\n<td align=\"left\">Better monitoring for your Rails applications. Get detailled statistics on your site's performance with mean and 90th percentile measurements.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/instrumentalapp.com\" rel=\"nofollow\">Instrumental<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/instrumental\" rel=\"nofollow\">@instrumental<\/a><\/td>\n<td align=\"left\">$150\/mo - $750\/mo<\/td>\n<td align=\"left\">Monitor Your App in Realtime. Instrumental\u2019s made to monitor metrics at a ridiculously high scale. At rates of 500,000 metrics per second and higher, it doesn\u2019t even break a sweat<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atatus.com\" rel=\"nofollow\">Atatus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/atatusapp\" rel=\"nofollow\">@atatusapp<\/a><\/td>\n<td align=\"left\">$12\/mo - $159\/mo<\/td>\n<td align=\"left\">Simple JavaScript Error Tracking. Atatus is a simple error tracking and uptime monitoring system. Add two lines of code and get alerted on any errors that occurs in your application in realtime.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-load-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#load-testing\"><\/a>Load Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/github.com\/newsapps\/beeswithmachineguns\">Bees with Machine Guns!<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/flood.io\" rel=\"nofollow\">Flood.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/flood_io\" rel=\"nofollow\">@flood_io<\/a><\/td>\n<td align=\"left\">Free to $399\/mo<\/td>\n<td align=\"left\">Auto setup and results summaries\/graphs of JMeter and Gatling load tests.  Can easily scale to 100K+ reqs\/min.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.neustar.biz\/services\/web-performance\/load-testing\" rel=\"nofollow\">Neustar Website Load Testing<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Neustar\" rel=\"nofollow\">@Neustar<\/a><\/td>\n<td align=\"left\">$80\/mo<\/td>\n<td align=\"left\">Tackle performance problems such as bandwidth limitations, error rates exceeding thresholds, server PU limitations and much more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/loader.io\" rel=\"nofollow\">Loader.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free to 100.00$ \/ mo<\/td>\n<td align=\"left\">Loader.io is a Free load testing service that allows you to stress test your web-apps\/apis with thousands of concurrent connections.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/locust.io\" rel=\"nofollow\">Locust.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/locustio\" rel=\"nofollow\">@locustio<\/a><\/td>\n<td align=\"left\">Open Source<\/td>\n<td align=\"left\">Self-hosted load testing tool written in Python, with the tests themselves written in Python code<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/k6.io\" rel=\"nofollow\">k6.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/k6_io\" rel=\"nofollow\">@k6_io<\/a><\/td>\n<td align=\"left\">Free, Open Source<\/td>\n<td align=\"left\">k6 is a developer centric open source load testing tool for testing the performance of your backend infrastructure. It\u2019s built with Go and JavaScript to integrate well into your development workflow, so you can stay on top of performance without fuzz.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/loadimpact.com\/\" rel=\"nofollow\">loadimpact.com<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/loadimpact\" rel=\"nofollow\">@loadimpact<\/a><\/td>\n<td align=\"left\">$99.00+<\/td>\n<td align=\"left\">A SaaS performance testing solution for modern DevOps teams, built on top of the open source k6.io project.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-server-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#server-monitoring\"><\/a>Server Monitoring<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.serverdensity.com\" rel=\"nofollow\">Server Density<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/serverdensity\" rel=\"nofollow\">@serverdensity<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Premium hosted website and server monitoring tool. All your activity syncs in real time - from starting new instances to upgrading or deleting old ones. Work wherever you want - through web, mobile, API or directly with your provider. Everything stays in sync.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.datadoghq.com\" rel=\"nofollow\">Datadog<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/datadoghq\" rel=\"nofollow\">@datadoghq<\/a><\/td>\n<td align=\"left\">$0\/mo - $15\/host\/mo<\/td>\n<td align=\"left\">Datadog is a monitoring service for IT, Operations and Development teams who write and run applications at scale, and want to turn the massive amounts of data produced by their apps, tools and services into actionable insight.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.circonus.com\" rel=\"nofollow\">Circonus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/circonus\" rel=\"nofollow\">@circonus<\/a><\/td>\n<td align=\"left\">$15\/host\/mo - $25\/host\/mo<\/td>\n<td align=\"left\">Circonus combines multiple monitoring, alerting, event reporting, and analytical tools into one unified solution. Use any data, in any application, from any system, and visualize it in real-time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/truesightpulse.bmc.com\" rel=\"nofollow\">TrueSight Pulse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/truesightpulse\" rel=\"nofollow\">@truesightpulse<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Real-time visibility into cloud and server infrastructure<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.librato.com\" rel=\"nofollow\">Librato<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Librato\" rel=\"nofollow\">@Librato<\/a><\/td>\n<td align=\"left\">$0.05\/metric\/mo to $0.30\/metric\/mo<\/td>\n<td align=\"left\">Librato provides a complete solution for monitoring and understanding the metrics that impact your business at all levels of the stack.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/scoutapp.com\/\" rel=\"nofollow\">Scout<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/prometheus.io\" rel=\"nofollow\">Prometheus<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.site24x7.com\" rel=\"nofollow\">Site24x7<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Site24x7\" rel=\"nofollow\">@site24x7<\/a><\/td>\n<td align=\"left\">$9\/mo starter - $225\/enterprise<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/uptimerobot.com\" rel=\"nofollow\">Uptime Robot<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/uptimerobot\" rel=\"nofollow\">@uptimerobot<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free basic HTTP\/HTTPS website monitoring<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.binarycanary.com\" rel=\"nofollow\">BinaryCanary<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-customer-supporthelp-desks\" class=\"anchor\" aria-hidden=\"true\" href=\"#customer-supporthelp-desks\"><\/a>Customer Support\/Help Desks<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.desk.com\" rel=\"nofollow\">Desk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/desk\" rel=\"nofollow\">@desk<\/a><\/td>\n<td align=\"left\">$3\/mo - $50\/mo<\/td>\n<td align=\"left\">Deliver Customer Service That Wows. Desk.com creates the leading customer service application, which helps fast-growing companies deliver outstanding customer support. Desk.com's intuitive user interface and powerful features make solving customers' problems more efficient for the entire company. Plus, Desk.com is the only Customer Service Application backed by Salesforce, providing easy integration with other Salesforce services and robust security. There are thousands of companies using Desk.com as their help desk software application, from household names like Square and Instagram to the burrito shop down the street. Give it a try for Free.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.helpscout.net\" rel=\"nofollow\">HelpScout<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/helpscout\" rel=\"nofollow\">@helpscout<\/a><\/td>\n<td align=\"left\">$15\/mo<\/td>\n<td align=\"left\">Scalable customer support, no help desk headaches. Based on conditions you specify, Help Scout automatically performs one or more actions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zendesk.com\" rel=\"nofollow\">ZenDesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zendesk\" rel=\"nofollow\">@zendesk<\/a><\/td>\n<td align=\"left\">$1\/mo - $195\/mo<\/td>\n<td align=\"left\">Relationships between businesses and their customers can be hard. Better customer service starts with better communication Zendesk brings all your customer conversations into one place.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.groovehq.com\" rel=\"nofollow\">Groove<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/groove\" rel=\"nofollow\">@groove<\/a><\/td>\n<td align=\"left\">$15\/ mo<\/td>\n<td align=\"left\">Everything you need to deliver awesome, personal support to every customer. Your customers will never know that you\u2019re using a helpdesk. To them, your messages look like regular emails. It feels like personal support, and it helps you build deeper relationships with your customers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.intercom.com\" rel=\"nofollow\">Intercom<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/intercom\" rel=\"nofollow\">@intercom<\/a><\/td>\n<td align=\"left\">$49\/mo - $449\/mo<\/td>\n<td align=\"left\">The easiest way to see and talk to your users. Intercom is a single platform where you can see in real-time who is using your product and send personalized messages to the right users at the right time based on their behavior.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/tenderapp.com\" rel=\"nofollow\">Tender<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tenderapp\" rel=\"nofollow\">@tenderapp<\/a><\/td>\n<td align=\"left\">$9\/mo - $99\/mo<\/td>\n<td align=\"left\">Better, Simpler, Customer Support Software. Support your customers in the open! With public forums, you can offer a public space to your users to discuss common issues and get feedback, while still keeping certain categories private (billing, orders, ...). Power users can subscribe to categories and new discussions, and help out other customers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.enchant.com\" rel=\"nofollow\">Enchant<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/enchanthq\" rel=\"nofollow\">@enchanthq<\/a><\/td>\n<td align=\"left\">$9\/ mo<\/td>\n<td align=\"left\">It's like gmail on steroids! Enchant is a powerful helpdesk that helps your team deliver awesome support to each and every customer. To your customers, it's just email. They will never see a ticket number and will never have to log into anything!<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/freshdesk.com\" rel=\"nofollow\">Freshdesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/freshdesk\" rel=\"nofollow\">@freshdesk<\/a><\/td>\n<td align=\"left\">$16\/mo - $70\/mo<\/td>\n<td align=\"left\">Everything you need to deliver Exceptional Customer Support. Freshdesk keeps you from running behind issues blindly and gets your customer support issues under control.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/userdeck.com\" rel=\"nofollow\">UserDeck<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/user_deck\" rel=\"nofollow\">@user_deck<\/a><\/td>\n<td align=\"left\">$0 - $25\/mo<\/td>\n<td align=\"left\">Customer support software that embeds into your existing website.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sirportly.com\" rel=\"nofollow\">Sirportly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sirportly\" rel=\"nofollow\">@sirportly<\/a><\/td>\n<td align=\"left\">\u00a30 - \u00a315\/mo<\/td>\n<td align=\"left\">Grow your business and provide world class customer support. Simply setup your helpdesk in less than a few minutes. Integrate with your other software tools and take advantage of the automated rules and macros to scale your customer support to a new level, become more professional and customer focused and turn your customers into raving fans with Sirportly.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.olark.com\/\" rel=\"nofollow\">Olark<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/snapengage.com\/\" rel=\"nofollow\">SnapEngage<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Custom pricing per integration, or from $81\/month<\/td>\n<td align=\"left\">Highly customised live chat integrations<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getsatisfaction.com\/corp\/\" rel=\"nofollow\">Get Satisfaction!<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Customer communities for social support, social marketing &amp; customer feedback - online community software. Get Satisfaction is the leading customer engagement platform that helps companies build better relationships with their customers and prospects, through the best online customer community.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.reamaze.com\" rel=\"nofollow\">Reamaze<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/reamaze\" rel=\"nofollow\">@reamaze<\/a><\/td>\n<td align=\"left\">$15\/mo<\/td>\n<td align=\"left\">Lightweight, Lightspeed Help Desk. Email, Social, Branded, Integrated. Reamaze provides your team with helpdesk functionality that integrates with your application, as well as integrations with popular 3rd party tools in your workflow.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.jitbit.com\/helpdesk\/\" rel=\"nofollow\">Jitbit Helpdesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/jitbithelpdesk\" rel=\"nofollow\">@jitbithelpdesk<\/a><\/td>\n<td align=\"left\">$29\/mo - $199\/mo<\/td>\n<td align=\"left\">A help desk app that actually makes your work easier, not harder. Comes in both hosted and on-premise versions. Very well designed and easy to use. Has all the must-have features and doesn't get in your way.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.drift.com\" rel=\"nofollow\">Drift<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/drift\" rel=\"nofollow\">@drift<\/a><\/td>\n<td align=\"left\">Free for &lt; 100 contacts, paid from $49\/mo<\/td>\n<td align=\"left\">Stop wasting your website traffic. Sales-oriented live chat and in-app messaging, with chatbot automation.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zammad.com\" rel=\"nofollow\">Zammad<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zammadhq\" rel=\"nofollow\">@zammadhq<\/a><\/td>\n<td align=\"left\">$0\/mo - $24\/mo<\/td>\n<td align=\"left\">Zammad is a award winning open source helpdesk\/customer support system with a beautiful UI and many features to manage customer communication via several channels like telephone, facebook, twitter, chat and e-mails. It is distributed under the GNU AFFERO General Public License (AGPL).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.supporthero.io\" rel=\"nofollow\">Support Hero<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/supportheroapp\" rel=\"nofollow\">@supportheroapp<\/a><\/td>\n<td align=\"left\">$49\/mo - $199\/mo<\/td>\n<td align=\"left\">Knowledge base and tutorial management platform with smart analytics. Support Hero is a self-service knowledge base that helps companies educate their clients. Helping them to create and deliver a game-changing user education experience that increases retention and engagement, while reducing customer relationship costs.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-transactional-email\" class=\"anchor\" aria-hidden=\"true\" href=\"#transactional-email\"><\/a>Transactional Email<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/postmarkapp.com\" rel=\"nofollow\">Postmark<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/postmarkapp\" rel=\"nofollow\">@postmarkapp<\/a><\/td>\n<td align=\"left\">$1.50<\/td>\n<td align=\"left\">Email delivery for web apps \u2013 done right. Postmark removes the headaches of delivering and parsing transactional email for webapps with minimal setup time and zero maintenance. We have years of experience getting email to the inbox, so you can work and rest easier.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.mailgun.com\" rel=\"nofollow\">MailGun<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Mail_Gun\" rel=\"nofollow\">@Mail_Gun<\/a><\/td>\n<td align=\"left\">$20.00<\/td>\n<td align=\"left\">The Email Service For Developers. Easy SMTP integration and a simple, RESTful API abstracts away the messy details of email. Scale quickly, whether you need to send 10 or 10 million emails.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/ses\" rel=\"nofollow\">Amazon Simple Email Service<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/AWSSupport\" rel=\"nofollow\">@AWSSupport<\/a><\/td>\n<td align=\"left\">$0 for the first 62,000 emails you send each month<\/td>\n<td align=\"left\">Flexible, affordable, and highly-scalable email sending and receiving platform for businesses and developers<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sendgrid.com\" rel=\"nofollow\">SendGrid<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SendGrid\" rel=\"nofollow\">@SendGrid<\/a><\/td>\n<td align=\"left\">$9.95\/mo - $399.95<\/td>\n<td align=\"left\">Email Delivery. Simplified. SendGrid delivers billions of emails for companies of all sizes every month. Select the package that best fits with your sending volume, set-up your account, and let SendGrid take care of the rest!<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.critsend.com\" rel=\"nofollow\">CritSend<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/critsend\" rel=\"nofollow\">@critsend<\/a><\/td>\n<td align=\"left\">$50\/mo - $3000\/mo<\/td>\n<td align=\"left\">The Best SMTP Relay for Developers. Use the most reliable infrastructure for your transactional and bulk emails. It only takes 5 minutes to setup Critsend and start enjoying fast delivery time and automatic scalability.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/postageapp.com\" rel=\"nofollow\">Postage<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/postagebird\" rel=\"nofollow\">@postagebird<\/a><\/td>\n<td align=\"left\">$9\/mo - $399\/mo<\/td>\n<td align=\"left\">The easier way to send email from web apps. Postage app helps design, send, and analyze emails within minutes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sendwithus.com\" rel=\"nofollow\">Sendwithus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/send_with_us\" rel=\"nofollow\">@sendwithus<\/a><\/td>\n<td align=\"left\">Free 'Hacker' plan of 1000 messages\/mo<\/td>\n<td align=\"left\">Transactional email A\/B testing and drip campaigns<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sparkpost.com\" rel=\"nofollow\">SparkPost<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SparkPost\" rel=\"nofollow\">@SparkPost<\/a><\/td>\n<td align=\"left\">Free - $474\/mo<\/td>\n<td align=\"left\">The most performant email delivery service \u2014 whether your app or website sends hundreds, or billions, of messages.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-other-apis\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-apis\"><\/a>Other APIs<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.filestack.com\" rel=\"nofollow\">Filestack<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/FileStack\" rel=\"nofollow\">@FileStack<\/a><\/td>\n<td align=\"left\">$0\/mo - $49\/mo<\/td>\n<td align=\"left\">Easy, Powerful File Uploads. Connect, Store, and Process any file from anywhere on the Internet<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/openexchangerates.org\" rel=\"nofollow\">Open Exchange Rates<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$12\/mo - $97\/mo<\/td>\n<td align=\"left\">Real-time exchange rates &amp; currency conversion JSON API. A simple and easy-to-integrate exchange rates API in JSON format, with HTTPS and JSONP support, with examples, guides and full documentation.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/formapi.io\" rel=\"nofollow\">FormAPI<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/form_api\" rel=\"nofollow\">@form_api<\/a><\/td>\n<td align=\"left\">$49\/mo - $249\/mo<\/td>\n<td align=\"left\">PDF Service for Developers. Makes it easy for programmers to fill out and sign PDF documents. Saves a lot of development time if you need to generate a lot of contracts, invoices, etc.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-site-search\" class=\"anchor\" aria-hidden=\"true\" href=\"#site-search\"><\/a>Site Search<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.elastic.co\/products\/elasticsearch\" rel=\"nofollow\">Elasticsearch<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/elasticsearch\" rel=\"nofollow\">@elasticsearch<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">an end-to-end search and analytics platform. infinitely versatile. By combining the massively popular Elasticsearch, Logstash and Kibana we have created an end-to-end stack that delivers actionable insights in real-time from almost any type of structured and unstructured data source. Built and supported by the engineers behind each of these open source products, the Elasticsearch ELK stack makes searching and analyzing data easier than ever before.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/swiftype.com\" rel=\"nofollow\">Swiftype<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/swiftype\" rel=\"nofollow\">@Swiftype<\/a><\/td>\n<td align=\"left\">Free - $250\/mo, Enterprise Plans available<\/td>\n<td align=\"left\">Swiftype is a hosted website search service, available as a web crawler or as an API integration. API clients are available for major frameworks and languages, plugins are available for major third party platforms.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.algolia.com\" rel=\"nofollow\">Algolia<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Algolia\" rel=\"nofollow\">@Algolia<\/a><\/td>\n<td align=\"left\">$49\/mo - $449\/mo<\/td>\n<td align=\"left\">Build Realtime Search. Algolia is a fully hosted search service, available as a REST API. API clients are also available for all major frameworks, platforms and languages.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/lucene.apache.org\/solr\/\" rel=\"nofollow\">Apache Solr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">An open source search platform, based and co-released with Apache Lucene.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/cloudsearch\/\" rel=\"nofollow\">Amazon Cloudsearch<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Search SaaS<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-email-marketing\" class=\"anchor\" aria-hidden=\"true\" href=\"#email-marketing\"><\/a>Email Marketing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.mailcharts.com\" rel=\"nofollow\">MailCharts<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mailcharts\" rel=\"nofollow\">@mailcharts<\/a><\/td>\n<td align=\"left\">$30\/mo<\/td>\n<td align=\"left\">Track &amp; understand how your competitors use email marketing. MailCharts tracks over a thousand companies, bringing you actionable insights to help you improve your email marketing strategy, make data-driven decisions and gain design and content inspiration.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/customer.io\" rel=\"nofollow\">Customer.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$50\/mo - $1250\/Mo<\/td>\n<td align=\"left\">Send email based on what people do or don't do in your app. Customer.io lets you send newsletters to segments of customers using data from your site.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getvero.com\" rel=\"nofollow\">Vero<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/veroapp\" rel=\"nofollow\">@veroapp<\/a><\/td>\n<td align=\"left\">$99\/mo<\/td>\n<td align=\"left\">Send emails based on what your customers do. Vero makes it easy to create segments of customers based on the attributes you capture (e.g. age, location, gender) and the actions your customers take (e.g. logged in, used feature x, checked out, etc.).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mailchimp.com\" rel=\"nofollow\">Mailchimp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Mailchimp\" rel=\"nofollow\">@Mailchimp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.campaignmonitor.com\" rel=\"nofollow\">Campaign Monitor<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/CampaignMonitor\" rel=\"nofollow\">@CampaignMonitor<\/a><\/td>\n<td align=\"left\">$9\/mo - $699\/mo<\/td>\n<td align=\"left\">CAMPUnbounce Feature Tour<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sendy.co\" rel=\"nofollow\">Sendy<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getSendy\" rel=\"nofollow\">@getsendy<\/a><\/td>\n<td align=\"left\">$59 one time fee, host yourselfor use a hosted sendy instance from a variety of providers.<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/image-charts.com\" rel=\"nofollow\">Image-Charts<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/imagecharts\" rel=\"nofollow\">@imagecharts<\/a><\/td>\n<td align=\"left\">Free - $49\/mo, self-hosted plan available<\/td>\n<td align=\"left\">Include animated charts as image into emails, no server-side rendering, 1 URL = 1 chart, compatible with Google Image Charts.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.drip.com\" rel=\"nofollow\">Drip<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getdrip\" rel=\"nofollow\">@getdrip<\/a><\/td>\n<td align=\"left\">Free for &lt; 100 subscribers, paid plans from $49\/mo<\/td>\n<td align=\"left\">Advanced marketing automation with workflows, drip campaigns, conversion tracking etc..<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.mailerlite.com\" rel=\"nofollow\">MailerLite<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mailerlite\" rel=\"nofollow\">@mailerlite<\/a><\/td>\n<td align=\"left\">Free - $140\/mo<\/td>\n<td align=\"left\">Free plan offers a lot of features<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/markethero.io\/\" rel=\"nofollow\">MarketHero<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$99+\/mo<\/td>\n<td align=\"left\">Autoresponder \/ email marketing tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/emailoctopus.com\" rel=\"nofollow\">EmailOctopus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/emailoctopus\" rel=\"nofollow\">@emailoctopus<\/a><\/td>\n<td align=\"left\">Free up to 2,500 subscribers, paid plans from $19\/mo<\/td>\n<td align=\"left\">Low cost email marketing using Amazon SES.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-email-collectionlanding-page-apps\" class=\"anchor\" aria-hidden=\"true\" href=\"#email-collectionlanding-page-apps\"><\/a>Email Collection\/Landing Page Apps<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.launchrock.com\" rel=\"nofollow\">Launchrock<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/launchrock\" rel=\"nofollow\">@launchrock<\/a><\/td>\n<td align=\"left\">$49\/mo - $199\/mo<\/td>\n<td align=\"left\">Even if you know how to code a web page with HTML, you'll love how much faster it is with our landing page builder. Load up your logo and graphic assets, set up a few base colors from your brand palette and publish away.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/unbounce.com\" rel=\"nofollow\">Unbounce<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/unbounce\" rel=\"nofollow\">@unbounce<\/a><\/td>\n<td align=\"left\">$49\/mo - $199\/mo<\/td>\n<td align=\"left\">Unbounce empowers marketers to act independently from technical teams, improving their efficiency and their ability to generate sales. Produce high-converting landing pages without dealing with I.T. bottlenecks. See how Unbounce can enhance your campaigns and maximize your marketing spend.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.leadpages.net\/\" rel=\"nofollow\">LeadPages<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$25\/mo - $199\/mo<\/td>\n<td align=\"left\">Generate leads and increase revenue using the industry-leading landing page builder with accompanying suite of lead generation and opt-in tools.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/instapage.com\/\" rel=\"nofollow\">Instapge<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$29\/mo - $127\/mo<\/td>\n<td align=\"left\">A landing page solution for optimizing your ad spend. Build, Integrate, Collaborate &amp; Optimize.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/kickofflabs.com\" rel=\"nofollow\">KickoffLabs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kickofflabs\" rel=\"nofollow\">@kickofflabs<\/a><\/td>\n<td align=\"left\">$29\/mo - $99\/mo<\/td>\n<td align=\"left\">Stop building landing pages. Start building smarter campaigns. Send customers to a tailored page that speaks to them. Keep them engaged with signup-forms, newsletters and \u2013 best of all \u2013 a very cool, unique REWARD system for customer referrals.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.prefinery.com\" rel=\"nofollow\">Prefinery<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/prefinery\" rel=\"nofollow\">@prefinery<\/a><\/td>\n<td align=\"left\">$19\/mo-  $399\/mo<\/td>\n<td align=\"left\">Stress-Free Beta Invitation and Management. Prefinery is beta invitations &amp; management for serious product launches. You need more than just a landing page -- outsource your beta to Prefinery!<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-crmsales-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#crmsales-tools\"><\/a>CRM\/Sales Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kloudless.com\" rel=\"nofollow\">Kloudless<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kloudless\" rel=\"nofollow\">@kloudless<\/a><\/td>\n<td align=\"left\">Free - $500\/mo<\/td>\n<td align=\"left\">Kloudless allows you to integrate with several CRM solutions via a single &amp; comprehensive RESTful API. The Kloudless CRM Unified API allows you to integrate with Salesforce, SugarCRM, Microsoft Dynamics, Zoho, Oracle Sales Cloud, HubSpot, Pipeliner, and more with a single code-base. Kloudless also maintains all the integrations so you can focus on building awesome solutions faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.salesforce.com\" rel=\"nofollow\">Salesforce<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/salesforce\" rel=\"nofollow\">@salesforce<\/a><\/td>\n<td align=\"left\">TRANSFORM CONSUMERS INTO CUSTOMERS ACROSS ALL DIGITAL CHANNELS. The ExactTarget Marketing Cloud, built on the Salesforce1 Platform, allows marketers to create 1:1 campaigns like never before. So you can combine traditional digital channels like email, mobile, social, and the web with any conceivable IP-addressable product to turn consumers into customers.<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.salesforceiq.com\" rel=\"nofollow\">SalesforceIQ<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/salesforceiq\" rel=\"nofollow\">@salesforceiq<\/a><\/td>\n<td align=\"left\">\u00a317-85\/user\/mo<\/td>\n<td align=\"left\">Helps businesses build stronger relationships using Relationship Intelligence.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sugarcrm.com\" rel=\"nofollow\">SugarCRM<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sugarcrm\" rel=\"nofollow\">@sugarcrm<\/a><\/td>\n<td align=\"left\">$35\/mo - $150\/mo<\/td>\n<td align=\"left\">Sugar's open, flexible platform easily solves real business problems. From automating sales, marketing and customer support to creating a custom CRM application, we've got you covered. By placing the individual at the center of its solution, SugarCRM puts the \"i\" in CRM and empowers every customer-facing individual to create extraordinary customer relationships.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.insightly.com\" rel=\"nofollow\">Insight.ly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/insightly\" rel=\"nofollow\">@insightly<\/a><\/td>\n<td align=\"left\">$7\/mo<\/td>\n<td align=\"left\">Small Business CRM. Manage contacts, organizations, partners, vendors and suppliers. See everything from background, email history, events, projects or opportunities<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/close.io\" rel=\"nofollow\">Close.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/closeio\" rel=\"nofollow\">@closeio<\/a><\/td>\n<td align=\"left\">$59\/mo - $299\/mo<\/td>\n<td align=\"left\">Sales Communication Platform. Close.io automatically logs sent and received emails with your leads. You can send\/receive emails painlessly from Close.io. Or, by entering your IMAP and SMTP mail settings, Close.io can track emails that you send from Gmail or any email client.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.streak.com\" rel=\"nofollow\">Streak<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/streak\" rel=\"nofollow\">@streak<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Manage email support inside Gmail. Streak lets you keep track of all your deals right from your inbox. We let you group emails from the same customer together into one view and push that customer through your pipeline. When a new email comes in, you'll have all the context you need.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getbase.com\" rel=\"nofollow\">Base<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getbase\" rel=\"nofollow\">@getbase<\/a><\/td>\n<td align=\"left\">$15\/mo - $125\/mo<\/td>\n<td align=\"left\">The Sales and CRM Software Your Team Will Actually Use. Base is designed to significantly boost your team\u2019s sales productivity and give you the core sales tools you need to grow your business. Your leads come from a variety of sources. Lead management in Base helps you organize leads and assign them to the right sales reps so they can be followed up on and qualified quickly. After being qualified, convert a lead in Base and all of the contact information you have about your lead is transferred your new customer contact card. Optionally, you can specify follow up tasks and even create a deal at the same time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pipedrive.com\" rel=\"nofollow\">Pipedrive<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$9\/mo<\/td>\n<td align=\"left\">Pipeline software that gets you organized. Pipedrive is built for salespeople who need to put in serious effort to turn leads into sales. It helps to organize the work and spend less time on admin.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.contactually.com\" rel=\"nofollow\">Contactually<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/contactually\" rel=\"nofollow\">@Contactually<\/a><\/td>\n<td align=\"left\">$17.99\/mo - $99.99\/mo<\/td>\n<td align=\"left\">Maximize your network ROI. More referrals. More repeat business. Contactually helps businesses follow up with the right people, at the right time, to maximize relationship ROI.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-social-media-marketing\" class=\"anchor\" aria-hidden=\"true\" href=\"#social-media-marketing\"><\/a>Social Media Marketing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/buffer.com\" rel=\"nofollow\">Buffer<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/buffer\" rel=\"nofollow\">@buffer<\/a><\/td>\n<td align=\"left\">$10\/mo<\/td>\n<td align=\"left\">Buffer is the easiest way to publish on social media. Buffer helps you share to Twitter, Facebook and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hootsuite.com\" rel=\"nofollow\">HootSuite<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/HootSuite\" rel=\"nofollow\">@HootSuite<\/a><\/td>\n<td align=\"left\">$8.99\/mo<\/td>\n<td align=\"left\">Add speed and agility to your social media strategy. From one dashboard you\u2019ll schedule Tweets and Facebook posts, monitor conversations, and more. When you need to prove your social ROI, quickly create and customize ready-to-present analytics reports.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/claim.io\/welcome\" rel=\"nofollow\">Claim.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/claimio\" rel=\"nofollow\">@claimio<\/a><\/td>\n<td align=\"left\">$189\/mo - $499\/mo<\/td>\n<td align=\"left\">Owning your name on 300 Social media sites not only makes it easier for people to find you or your business online - it also works as a \"social media identity insurance\", protecting you from name squatting and identity fraud, minimizing risk to your brand.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.salesforce.com\/products\/marketing-cloud\/overview\/\" rel=\"nofollow\">Exacttarget Marketing Cloud\/Buddy Media<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/marketingcloud\" rel=\"nofollow\">@marketingcloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Bringing you closer to your social customers with an integrated social content solution. Run integrated campaigns across Facebook, Twitter, YouTube and your websites Increase fans, followers and advocates. Publish engaging and interactive social apps. Easily create landing pages and microsites and extend social to your websites. Understand engagement trends, demographics, conversions and business metrics with powerful analytics.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sproutsocial.com\" rel=\"nofollow\">Sprout Social<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sproutsocial\" rel=\"nofollow\">@sproutsocial<\/a><\/td>\n<td align=\"left\">$59\/mo - $1500\/mo<\/td>\n<td align=\"left\">Powerful Social Media Software. Sprout lets you post messages on Facebook, Twitter, Google+ and LinkedIn simultaneously from one easy-to-use message composition tool. Shorten links, attach photos, target your audience on Facebook, customize your posts and much more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/f5bot.com\" rel=\"nofollow\">F5Bot<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">F5Bot is Free social media keyword monitoring. F5Bot will send you an email when your startup, your product, or your competitors are mentioned on Hacker News or Reddit.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-naming\" class=\"anchor\" aria-hidden=\"true\" href=\"#naming\"><\/a>Naming<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.trademarkia.com\" rel=\"nofollow\">Trademarkia<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/trademarkia\" rel=\"nofollow\">@trademarkia<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Trademarkia is one of the largest trademark search engines in the world. LegalForce's network of licensed patent attorneys and agents have filed hundreds of patent applications for companies of every size.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.namerobot.com\" rel=\"nofollow\">NameRobot<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/namerobotEN\" rel=\"nofollow\">@namerobotEN<\/a>]<\/td>\n<td align=\"left\">0$ - 300$\/mo<\/td>\n<td align=\"left\">Find, create and check the name for your project. NameRobot offers everything you need to create suitable naming ideas in a short time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/whois.domaintools.com\" rel=\"nofollow\">DomainTools Whois Lookup<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/DomainTools\" rel=\"nofollow\">@DomainTools<\/a><\/td>\n<td align=\"left\">Free - $99\/mo<\/td>\n<td align=\"left\">Go beyond ordinary Whois to discover the people or organizations behind a domain name or IP address.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-space-rental\" class=\"anchor\" aria-hidden=\"true\" href=\"#space-rental\"><\/a>Space Rental<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/42floors.com\" rel=\"nofollow\">42Floors<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/42floors\" rel=\"nofollow\">@42floors<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The Best Place to Find Office &amp; Commercial Space Rentals. We're gathering listing data from everyone in the market. Even including off-market listings that landlords haven't yet posted on 42Floors.com or anywhere else.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/liquidspace.com\" rel=\"nofollow\">Liquidspace<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/LiquidSpace\" rel=\"nofollow\">@LiquidSpace<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Optimize Real Estate for Your Enterprise. Whether planning ahead or booking on the fly, find and reserve a great place to work. Rent professional conference and meeting rooms, private offices, or coworking spaces daily or hourly.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pivotdesk.com\" rel=\"nofollow\">PivotDesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PivotDesk\" rel=\"nofollow\">@PivotDesk<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">An office sharing marketplace that helps startups that need space find host companies that have excess space.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-community-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#community-tools\"><\/a>Community Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.discourse.org\" rel=\"nofollow\">Discourse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/discourse\" rel=\"nofollow\">@discourse<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Ready for a new discussion platform? Discussion software is a group of people interested in a common topic who are willing to type paragraphs to each other on a web page.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-personal-productivity\" class=\"anchor\" aria-hidden=\"true\" href=\"#personal-productivity\"><\/a>Personal Productivity<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/www.tomato.es\/\" rel=\"nofollow\">Tomatoes<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tomatoesapp\" rel=\"nofollow\">@tomatoesapp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Track your time and be productive with the Pomodoro Technique. Tomatoes is a \"pomodoro tracker\", a Pomodoro technique\u00ae driven time tracker. Track your time using 25 minutes slots called \"pomodoros\".<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.rescuetime.com\" rel=\"nofollow\">RescueTime<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/rescuetime\" rel=\"nofollow\">@rescuetime<\/a><\/td>\n<td align=\"left\">Free - $9\/mo<\/td>\n<td align=\"left\">RescueTime gives you an accurate picture of how you spend your time to help you become more productive every day.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/qotoqot.com\/qbserve\/\" rel=\"nofollow\">Qbserve<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Qbserve\" rel=\"nofollow\">@Qbserve<\/a><\/td>\n<td align=\"left\">$40 one-time<\/td>\n<td align=\"left\">Qbserve is a time tracker for Mac that does whatever RescueTime can but also tracks project time automatically, generates invoices, and stores tracked data locally.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/timingapp.com\/\" rel=\"nofollow\">Timing<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/TimingApp\" rel=\"nofollow\">@TimingApp<\/a><\/td>\n<td align=\"left\">$29 - $79<\/td>\n<td align=\"left\">Automatic time and productivity tracking for Mac. Helps you stay on track with your work and ensures no billable hours get lost (if you are billing hourly).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/fman.io\" rel=\"nofollow\">fman<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/m_herrmann\" rel=\"nofollow\">@m_herrmann<\/a><\/td>\n<td align=\"left\">$14.00<\/td>\n<td align=\"left\">Manage and transfer your files with ease. For Windows, Mac and Linux.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/wakatime.com\" rel=\"nofollow\">WakaTime<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/WakaTime\" rel=\"nofollow\">@WakaTime<\/a><\/td>\n<td align=\"left\">Free - $9\/mo<\/td>\n<td align=\"left\">Automatic productivity metrics using text editor plugins. Goals, leaderboards, GitHub integration, automatic project\/language\/branch detection.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-prototypingmockups\" class=\"anchor\" aria-hidden=\"true\" href=\"#prototypingmockups\"><\/a>Prototyping\/Mockups<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/creately.com\" rel=\"nofollow\">Creately<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/creately\" rel=\"nofollow\">@creately<\/a><\/td>\n<td align=\"left\">Free - $750\/mo<\/td>\n<td align=\"left\">Web based diagramming tool for fast easy diagrams. Supports flowcharts, mock-ups, wire-frames, mind maps, organizational charts, network diagrams, AWS diagrams, UML diagrams and many other diagram types. Real-time collaboration plus innovative productivity features to create diagrams 3 times faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.apple.com\/keynote\/\" rel=\"nofollow\">Keynote<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$19.99<\/td>\n<td align=\"left\">Easily create gorgeous presentations with the all-new Keynote, featuring powerful yet easy-to-use tools and dazzling effects that will make you a very hard act to follow. Also checkout the <a href=\"https:\/\/keynotopia.com\" rel=\"nofollow\">Keynotopia Themes<\/a> to get all the common UI elements for iOS, Android etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.omnigroup.com\/omniGraffle\" rel=\"nofollow\">OmniGraffle<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/omniGraffle\" rel=\"nofollow\">@omniGraffle<\/a><\/td>\n<td align=\"left\">$99.99<\/td>\n<td align=\"left\">Diagramming Worth a Thousand Words. Need a diagram, process chart, quick page-layout, website wireframe or graphic design? OmniGraffle can help you make eye-popping graphic documents quickly by keeping lines connected to shapes even when they're moved, providing powerful styling tools, and magically organizing diagrams with just one click. Whether you need a quick<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/moqups.com\" rel=\"nofollow\">moqups<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/moqups\" rel=\"nofollow\">@moqups<\/a><\/td>\n<td align=\"left\">$9\/mo -  $39\/mo<\/td>\n<td align=\"left\">Moqups is a nifty HTML5 App used to create wireframes, mockups or UI concepts, prototypes depending on how you like to call them. The most stunning HTML5 app for creating resolution-independent SVG mockups &amp; wireframes for your next project.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/balsamiq.com\" rel=\"nofollow\">Balsamiq<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/balsamiq\" rel=\"nofollow\">@balsamiq<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Life's too short for bad software! Balsamiq Mockups is a rapid wireframing tool that helps you Work Faster &amp; Smarter. It reproduces the experience of sketching on a whiteboard, but using a computer.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/proto.io\" rel=\"nofollow\">Proto.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/protoio\" rel=\"nofollow\">@protoio<\/a><\/td>\n<td align=\"left\">$24\/mo - $199\/mo<\/td>\n<td align=\"left\">Silly-fast mobile prototyping. Build high-fidelity fully interactive mobile app prototypes in minutes. Prototypes can be viewed on browser or device giving a real experience to the user how the app will look like and behave. Multiple devices like smart phones and tablets\/pads are supported including iPhone, iPad and Android devices.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.invisionapp.com\" rel=\"nofollow\">invision<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/InVisionApp\" rel=\"nofollow\">@InVisionApp<\/a><\/td>\n<td align=\"left\">$0\/mo - $100+\/mo<\/td>\n<td align=\"left\">Free Web &amp; Mobile (iOS, Android) Prototyping and UI Mockup Tool. Transform your Web &amp; Mobile (iOS, Android) designs into clickable, interactive Prototypes and Mockups. Share and Collaborate on them with others.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sketchapp.com\" rel=\"nofollow\">Sketch<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sketchapp\" rel=\"nofollow\">@sketchapp<\/a><\/td>\n<td align=\"left\">$99.00<\/td>\n<td align=\"left\">Professional digital design for Mac. Sketch gives you the power, flexibility and speed you always wanted in a lightweight and easy-to-use package. Finally you can focus on what you do best: Design.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.figma.com\" rel=\"nofollow\">Figma<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/figmadesign\" rel=\"nofollow\">@figmadesign<\/a><\/td>\n<td align=\"left\">Free - $12\/editor\/mo<\/td>\n<td align=\"left\">Design and prototyping tool, with a focus on collaboration and teams.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.animaapp.com\" rel=\"nofollow\">Anima<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/animaapp\" rel=\"nofollow\">@animaapp<\/a><\/td>\n<td align=\"left\">$0 - $39\/mo<\/td>\n<td align=\"left\">Sketch to Code Toolkit. Anima converts web design to code. Make your Sketch design responsive, interactive and alive, and export to HTML\/JS\/CSS.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/framer.com\" rel=\"nofollow\">Framer<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/framer\" rel=\"nofollow\">@framer<\/a><\/td>\n<td align=\"left\">Free Trial - $12\/mo<\/td>\n<td align=\"left\">Pixel-perfect, responsive designs. Truly high-fidelity prototypes. Seamless collaboration. Framer powers the product teams at Dropbox, Pinterest, Twitter, and thousands more.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-content-creationinfographics\" class=\"anchor\" aria-hidden=\"true\" href=\"#content-creationinfographics\"><\/a>Content Creation\/Infographics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/visual.ly\" rel=\"nofollow\">Visual.ly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Visually\" rel=\"nofollow\">@Visually<\/a><\/td>\n<td align=\"left\">$195\/mo - $994\/mo<\/td>\n<td align=\"left\">ORIGINAL VISUAL CONTENT FOR BRANDS. We only work with the best creative talent available. Thousands of designers, journalists, animators and developers are standing by to help you achieve your goals and take your project to the next level.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.canva.com\" rel=\"nofollow\">Canva<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/canva\" rel=\"nofollow\">@canva<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Design great social media images with text and graphics for Free or a few bucks depending on the images you select<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-customer-feedback\" class=\"anchor\" aria-hidden=\"true\" href=\"#customer-feedback\"><\/a>Customer Feedback<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.pickfu.com\" rel=\"nofollow\">PickFu<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/pickfu\" rel=\"nofollow\">@pickfu<\/a><\/td>\n<td align=\"left\">$20\/mo - $299\/mo<\/td>\n<td align=\"left\">REAL CONSUMER FEEDBACK IN MINUTES. PickFu is a tool that provides instant, unbiased and insightful public opinion on questions that you care about.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.promoter.io\" rel=\"nofollow\">Promoter.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/promoter_io\" rel=\"nofollow\">@Promoter_io<\/a><\/td>\n<td align=\"left\">$50\/mo - $500\/mo<\/td>\n<td align=\"left\">Helps companies quickly gain predictive customer intelligence &amp; insights driven by NPS (Net Promoter) to increase revenue and reduce churn.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-data\" class=\"anchor\" aria-hidden=\"true\" href=\"#data\"><\/a>Data<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.factual.com\" rel=\"nofollow\">Factual<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Factual\" rel=\"nofollow\">@Factual<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">GLOBAL DATA. LOCAL CONTEXT. Factual\u2019s location platform enriches mobile location signals with definitive global data, enabling personalized and contextually relevant mobile experiences. Built from billions of inputs, the data is constantly updated by Factual\u2019s real-time data stack.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-database\" class=\"anchor\" aria-hidden=\"true\" href=\"#database\"><\/a>Database<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.citusdata.com\" rel=\"nofollow\">Citus<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Citus scales out PostgreSQL through sharding, replication, and query parallelization. So you can focus on your application, not worry about your database. Citus is available as open source, enterprise software, and fully managed database as a service.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/bulbflow.org\" rel=\"nofollow\">Bulbs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Python persistence framework for graph databases.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.datomic.com\" rel=\"nofollow\">Datomic<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/datomic_team\" rel=\"nofollow\">@datomic_team<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The fully transactional, cloud-ready, immutable database. Immutable data means strong consistency combined with horizontal read scalability, plus built-in caching. Datomic is a distributed database designed to enable scalable, flexible and intelligent applications, running on next-generation cloud architectures.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/tinkerpop.apache.org\" rel=\"nofollow\">Tinkerpop<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Open source software products in the graph space.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.vertabelo.com\" rel=\"nofollow\">Vertabelo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Vertabelo\" rel=\"nofollow\">@vertabelo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Web-based tool for database design. Vertabelo allows you to visually design database models for PostgreSQL, MySQL, Oracle, SQL Server, SQLite, and IBM DB2. You can import the existing database structure from SQL, XML, or using reverse engineering tool. After you design a model, you can generate SQL script or ready-to-use code for various ORMs (Propel, jOOQ, SQLAlchemy, or <a href=\"http:\/\/mobile-orm.vertabelo.com\/\" rel=\"nofollow\">Vertabelo Mobile ORM<\/a>).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/tableplus.io\" rel=\"nofollow\">TablePlus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/TablePlus\" rel=\"nofollow\">@TablePlus<\/a><\/td>\n<td align=\"left\">$0 - $49<\/td>\n<td align=\"left\">Modern, native client with intuitive GUI tools to create, access, query &amp; edit multiple databases: MySQL, PostgreSQL, SQLite, Microsoft SQL Server, Amazon Redshift, MariaDB, CockroachDB, Vertica, Cassandra, Oracle, and Redis.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dbngin.com\" rel=\"nofollow\">DBngin<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dbngin\" rel=\"nofollow\">@DBngin<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">An all-in-one tool for database version management. It allows you to set up a local database server of any versions within a click.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-accountinginvoicing\" class=\"anchor\" aria-hidden=\"true\" href=\"#accountinginvoicing\"><\/a>Accounting\/Invoicing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.getharvest.com\" rel=\"nofollow\">Harvest<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/harvest\" rel=\"nofollow\">@harvest<\/a><\/td>\n<td align=\"left\">$12\/mo - $99\/mo<\/td>\n<td align=\"left\">Time Tracking Made Easy. Time tracking is simple and lightning fast with Harvest. Set up takes seconds, and there\u2019s nothing to install. We\u2019ve simplified the timesheet and timesheets approval process so you can stay focused on work.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getballpark.com\" rel=\"nofollow\">Ballpark<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$12\/mo - $99\/mo<\/td>\n<td align=\"left\">Stop sending your clients ugly paper invoices. Go paperless today, with Ballpark. Our beautiful, web-based invoices and estimates make it easier than ever to get paid and discuss projects with your clients and colleagues.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paysimple.com\" rel=\"nofollow\">PaySimple<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PaySimple\" rel=\"nofollow\">@PaySimple<\/a><\/td>\n<td align=\"left\">$34.95\/mo<\/td>\n<td align=\"left\">Simplify how you bill and collect. PaySimple is an industry-leading provider of payment management solutions. PaySimple simplifies billing and collection processes by enabling you to bill, collect and deposit all of your payments automatically. Our customized, secure ASP solution includes auto-recurring billing, electronic check processing, direct debit and credit card processing at some of the lowest rates available.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.freshbooks.com\" rel=\"nofollow\">FreshBooks<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/freshbooks\" rel=\"nofollow\">@freshbooks<\/a><\/td>\n<td align=\"left\">$19.95\/mo - $39.95\/mo<\/td>\n<td align=\"left\">Accounting Made for You, the Non-Accountant. FreshBooks is simple and intuitive, so accounting isn't intimidating. Plus you can talk to a real, live person anytime you have a question, 9am-6pm EDT, Mon-Fri.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.Freeagent.com\" rel=\"nofollow\">FreeAgent<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Freeagent\" rel=\"nofollow\">@Freeagent<\/a><\/td>\n<td align=\"left\">US $20\/mo<\/td>\n<td align=\"left\">Accounting software trusted by over 35,000 Freelancers and small businesses. Reconcile money in and out of the business via your electronic bank statements and build monthly balance charts.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.blinksale.com\" rel=\"nofollow\">Blinksale<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/blinksale\" rel=\"nofollow\">@blinksale<\/a><\/td>\n<td align=\"left\">$15.00<\/td>\n<td align=\"left\">INTRODUCING BLINKSALE UNLIMITED. ONE PLAN. ONE PRICE. UNLIMITED EVERYTHING. Blinksale makes you look your best. With over a dozen invoice designs &amp; thank you notes, you\u2019re sure to put your best foot forward. Every time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/cashboardapp.com\" rel=\"nofollow\">Cashboard<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/cashboard\" rel=\"nofollow\">@cashboard<\/a><\/td>\n<td align=\"left\">$8.25\/mo - $250\/mo<\/td>\n<td align=\"left\">FreeLANCE TIME TRACKING &amp; INVOICE SOFTWARE TRUSTED BY THOUSANDS, WORLDWIDE. Cashboard is the tool we designed to remedy that situation. It works for our software consultancy and we think it\u2019ll work for you too. We launched Cashboard in Spring of 2007after many sleepless nights of hard work. It was the first solution to combine estimates, invoices, time tracking, and online payments into one tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paydirtapp.com\" rel=\"nofollow\">Paydirt<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/paydirtapp\" rel=\"nofollow\">@paydirtapp<\/a><\/td>\n<td align=\"left\">$8\/mo - $149\/mo<\/td>\n<td align=\"left\">Smart Time Tracking Easier Invoicing Online Payments. In Paydirt, you can start a timer from any page in one click. No fiddly menus. No navigating around. Just a start button for each task.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.indinero.com\" rel=\"nofollow\">inDinero<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/qbo.intuit.com\/qbo27\/login?webredir\" rel=\"nofollow\">QuickBooks Online<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.xero.com\" rel=\"nofollow\">Xero<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/Fast409A.io\" rel=\"nofollow\">Fast409A<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">$2000 (depending on startup funding stage)<\/td>\n<td align=\"left\">The Fast409A.io process blends the best of software and human expertise. Each 409A valuation takes an hour or less of your time to submit, is delivered in days not weeks, and is customised to the nuances of your individual startup.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/startuprunway.io\" rel=\"nofollow\">Runway<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Most startups die from running out of cash. Runway is an easy, visual way to plan and manage your startup's most precious resource, explore scenarios and track planned versus actual spending.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-privacy-policy-terms--conditions-legal-documents\" class=\"anchor\" aria-hidden=\"true\" href=\"#privacy-policy-terms--conditions-legal-documents\"><\/a>Privacy Policy, Terms &amp; Conditions, Legal Documents<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/www.iubenda.com\" rel=\"nofollow\">iubenda<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/iubenda\" rel=\"nofollow\">@iubenda<\/a><\/td>\n<td align=\"left\">Free - $27\/year - customization services<\/td>\n<td align=\"left\">The easiest way to generate a professional, customizable, self-updating privacy policy. Choose between 6 languages. Documents hosted and kept up to date. Backed by real lawyers. Additional assistance service with premium legal team for custom Privacy Policy and Terms &amp; Conditions.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-income-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#income-analytics\"><\/a>Income Analytics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/baremetrics.com\" rel=\"nofollow\">Baremetrics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Baremetrics\" rel=\"nofollow\">@Baremetrics<\/a><\/td>\n<td align=\"left\">$79\/mo - $249\/mo<\/td>\n<td align=\"left\">One click and you get dozens of valuable metrics from your Stripe account<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getmetrics.io\" rel=\"nofollow\">getMetrics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getMEtrics\" rel=\"nofollow\">@getMetrics<\/a><\/td>\n<td align=\"left\">$25\/mo - $39\/mo<\/td>\n<td align=\"left\">Stripe metrics and notifications for Stripe events<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-payments-billing--downloads\" class=\"anchor\" aria-hidden=\"true\" href=\"#payments-billing--downloads\"><\/a>Payments, Billing &amp; Downloads<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.paypal.com\" rel=\"nofollow\">PayPal<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PayPal\" rel=\"nofollow\">@PayPal<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">PayPal is an international e-commerce business allowing payments and money transfers to be made through the Internet. Online money transfers serve as electronic alternatives to paying with traditional paper methods, such as checks and money orders<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gumroad.com\" rel=\"nofollow\">Gumroad<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gumroad\" rel=\"nofollow\">@gumroad<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">See higher conversion, lower fees, and more customer control. Sell films directly to your viewers. Creating digital products is hard, selling them shouldn't be. We let you start selling downloads in seconds.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.fetchapp.com\" rel=\"nofollow\">FetchApp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/fetchapp\" rel=\"nofollow\">@fetchapp<\/a><\/td>\n<td align=\"left\">$5\/mo - $500\/mo<\/td>\n<td align=\"left\">The Simpler way to Fetch. Simply put, FetchApp allows you to sell and digitally delivery downloadable goods<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.chargify.com\" rel=\"nofollow\">Chargify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chargify\" rel=\"nofollow\">@chargify<\/a><\/td>\n<td align=\"left\">$459\/mo - $65\/mo<\/td>\n<td align=\"left\">Easily Manage Your Recurring Revenue Business. Customers sign up, make payments, use coupons, upgrade... You bill one-time &amp; recurring fees using whatever pricing model you need, charge cards, send invoices &amp; reminders, etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/recurly.com\" rel=\"nofollow\">Recurly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Recurly\" rel=\"nofollow\">@Recurly<\/a><\/td>\n<td align=\"left\">$99\/mo - $259\/mo<\/td>\n<td align=\"left\">Subscription Billing Automation. As the leading recurring billing platform, Recurly ensures setup is easy, integrations are quick, and our service scales with the needs of your business. With Recurly you'll be ready to accept payments and focus on growing your sales in no time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.chargebee.com\" rel=\"nofollow\">ChargeBee<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chargebee\" rel=\"nofollow\">@chargebee<\/a><\/td>\n<td align=\"left\">$49\/mo - $249\/mo<\/td>\n<td align=\"left\">ChargeBee is an easy to use recurring billing and invoicing solution for online businesses<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/chargeover.com\" rel=\"nofollow\">ChargeOver<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chargeover\" rel=\"nofollow\">@ChargeOver<\/a><\/td>\n<td align=\"left\">$65\/mo- $549\/mo<\/td>\n<td align=\"left\">ChargeOver is a full featured invoicing platform supporting payment plans, recurring billing, one-time billing, and all the major gateway backends.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-billing--payment-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#billing--payment-processing\"><\/a>Billing &amp; Payment Processing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.braintreepayments.com\" rel=\"nofollow\">Braintree<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/braintree\" rel=\"nofollow\">@braintree<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Accept payments in your app or website. Braintree handles transactions for some of the fastest growing mobile companies like Uber, Airbnb, HotelTonight and Fab. With native, easy-to-follow SDKs for iOS, Android and Windows Phone you can quickly add native payments to your app.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.dwolla.com\" rel=\"nofollow\">Dwolla<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dwolla\" rel=\"nofollow\">@dwolla<\/a><\/td>\n<td align=\"left\">25\u00a2\/transaction<\/td>\n<td align=\"left\">The best way to move money. Dwolla, Inc. is an agent of Veridian Credit Union and all funds associated with your account in the Dwolla network are held in a pooled account at Veridian Credit Union. These funds are not eligible for individual insurance, and may not be eligible for share insurance by the National Credit Union Share Insurance Fund. Dwolla, Inc. is the operator of a software platform that communicates user instructions for funds transfers to Veridian Credit Union.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/stripe.com\" rel=\"nofollow\">Stripe<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/stripe\" rel=\"nofollow\">@stripe<\/a><\/td>\n<td align=\"left\">2.9% + 30\u00a2\/successful charge.<\/td>\n<td align=\"left\">Feature-packed payments. No need to design payment forms from scratch. Stripe Checkout offers a beautiful, customizable payment flow that works great across desktop and mobile. When you use Checkout, you\u2019re always up-to-date, with no extra code required.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pinpayments.com\" rel=\"nofollow\">Pin<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/pin_payments\" rel=\"nofollow\">@pin_payments<\/a><\/td>\n<td align=\"left\">2.9% + 30\u00a2\/successful charge.<\/td>\n<td align=\"left\">Payments, Rebooted. Accepting credit card payments from a global audience typically requires a merchant account. The process of establishing a merchant account for each currency can be too difficult and costly for small businesses.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.paymill.com\" rel=\"nofollow\">PayMill<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Paymill\" rel=\"nofollow\">@Paymill<\/a><\/td>\n<td align=\"left\">0.28 \u20ac - 0.25 \u20ac<\/td>\n<td align=\"left\">Online payments made easy. Make payments personal by customising the checkout according to the flow of your website<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.spreedly.com\" rel=\"nofollow\">Spreedly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/spreedly\" rel=\"nofollow\">@spreedly<\/a><\/td>\n<td align=\"left\">$150\/mo - $1500\/mo<\/td>\n<td align=\"left\">Payments as a Platform. One of Spreedly's major benefits is reaching a large number of merchant accounts by working across multiple payment gateways. As a direct merchant you can transact globally but deposit funds in unique merchant accounts based on geographic or other business rules. As a SaaS platform you can support the unique merchant accounts of your individual customers. A payment gateway token is your way to indicate to us which unique merchant account this particular transaction will go against. Each unique merchant account = one unique payment gateway token.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/go.wepay.com\" rel=\"nofollow\">WePay<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wepay\" rel=\"nofollow\">@wepay<\/a><\/td>\n<td align=\"left\">2.9% + 30\u00a2\/transaction.<\/td>\n<td align=\"left\">WePay is the first payments engine to offer platforms \u2014 marketplaces, crowdfunding, and business software\/tools \u2014 a way to own their customer experience while still shielding them from 100 percent of fraud and regulatory risk.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paddle.com\" rel=\"nofollow\">Paddle<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PaddleHQ\" rel=\"nofollow\">@PaddleHQ<\/a><\/td>\n<td align=\"left\">5% + 50\u00a2\/transaction.<\/td>\n<td align=\"left\">Payment processing and fulfillment, specialized for desktop apps and SaaS subscription services. Handles VAT and invoicing for you, so your accounting becomes easier.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/fattmerchant.com\/\" rel=\"nofollow\">Fattmerchant<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/fattmerchant\" rel=\"nofollow\">@fattmerchant<\/a><\/td>\n<td align=\"left\">0% + interchange + $99+\/mo<\/td>\n<td align=\"left\">Payments unified.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-banking\" class=\"anchor\" aria-hidden=\"true\" href=\"#banking\"><\/a>Banking<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.simple.com\" rel=\"nofollow\">Simple<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/simple\" rel=\"nofollow\">@simple<\/a>)<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The way banking should be.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-phonepbxsms\" class=\"anchor\" aria-hidden=\"true\" href=\"#phonepbxsms\"><\/a>Phone\/PBX\/SMS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.plivo.com\/\" rel=\"nofollow\">Plivo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tropo.com\/\" rel=\"nofollow\">Tropo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.twilio.com\/\" rel=\"nofollow\">Twilio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.phonebooth.com\/\" rel=\"nofollow\">PhoneBooth<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.talkdesk.com\/\" rel=\"nofollow\">TalkDesk<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hellofax.com\/\" rel=\"nofollow\">HelloFax<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.dialpad.com\" rel=\"nofollow\">Dialpad<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dialpadHQ\" rel=\"nofollow\">@DialpadHQ<\/a><\/td>\n<td align=\"left\">$15\/user\/mo<\/td>\n<td align=\"left\">people-first phone and meeting system that includes internet fax and follows you where you work<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/smith.ai\/keypad\" rel=\"nofollow\">Keypad by Smith.ai<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/smithdotai\" rel=\"nofollow\">@smithdotai<\/a><\/td>\n<td align=\"left\">$25\/mo<\/td>\n<td align=\"left\">The most advanced &amp; affordable cloud phone system for small businesses.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-system-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#system-monitoring\"><\/a>System Monitoring<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.appneta.com\/\" rel=\"nofollow\">AppNeta (Tracelytics)<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sensuapp.org\/\" rel=\"nofollow\">Sensu<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/riemann.io\" rel=\"nofollow\">Riemann<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pingdom.com\/\" rel=\"nofollow\">Pingdom<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/uptimerobot.com\/\" rel=\"nofollow\">UptimeRobot<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/wheresitup.com\/\" rel=\"nofollow\">Where's It Up?<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Nagios<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Smokeping<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-search\" class=\"anchor\" aria-hidden=\"true\" href=\"#search\"><\/a>Search<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/bonsai.io\/\" rel=\"nofollow\">Bonsai<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.websolr.com\" rel=\"nofollow\">WebSolr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.searchify.com\/\" rel=\"nofollow\">Searchify<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.searchblox.com\" rel=\"nofollow\">SearchBlox<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-security\" class=\"anchor\" aria-hidden=\"true\" href=\"#security\"><\/a>Security<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/portswigger.net\/burp\" rel=\"nofollow\">Burp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/duo.com\/\" rel=\"nofollow\">DuoSecurity<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/authy.com\/\" rel=\"nofollow\">Authy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.anchorFree.com\" rel=\"nofollow\">AnchorFree<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/encrypt.me\" rel=\"nofollow\">Encrypt.me<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tinfoilsecurity.com\/\" rel=\"nofollow\">Tinfoil Security<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-shipping\" class=\"anchor\" aria-hidden=\"true\" href=\"#shipping\"><\/a>Shipping<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.shipwire.com\" rel=\"nofollow\">Shipwire<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.shyp.com\" rel=\"nofollow\">Shyp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getwhiplash.com\" rel=\"nofollow\">Whiplash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-user-feedback\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-feedback\"><\/a>User Feedback<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.uservoice.com\/\" rel=\"nofollow\">Uservoice<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/userecho.com\/\" rel=\"nofollow\">Userecho<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-designers\" class=\"anchor\" aria-hidden=\"true\" href=\"#designers\"><\/a>Designers<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/dribbble.com\" rel=\"nofollow\">Dribbble<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sortfolio.com\" rel=\"nofollow\">Sortfolio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.behance.net\" rel=\"nofollow\">Behance<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#notes\"><\/a>Notes<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.google.com\/docs\/about\/\" rel=\"nofollow\">Google Docs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/evernote.com\/\" rel=\"nofollow\">Evernote<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.google.com\/keep\/\" rel=\"nofollow\">Google Keep<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/workflowy.com\/\" rel=\"nofollow\">Workflowy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/quip.com\/\" rel=\"nofollow\">Quip<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/etherpad.org\/\" rel=\"nofollow\">Etherpad<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.kamihq.com\" rel=\"nofollow\">Kami<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">View, edit, annotate, and collaborate on documents on your browser<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.onenote.com\/\" rel=\"nofollow\">OneNote<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/msonenote\" rel=\"nofollow\">@msonenote<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">The digital note-taking app for your devices<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.taskade.com\" rel=\"nofollow\">Taskade<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/taskade\" rel=\"nofollow\">@taskade<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Real-time collaborative notes, checklists, and outlines for teams.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/nulis.io\" rel=\"nofollow\">Nulis<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Outliner \/ mind-mapping tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/writingstreak.io\" rel=\"nofollow\">WritingStreak<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Develop a writing habit, take daily notes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dynalist.io\/\" rel=\"nofollow\">Dynalist<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dynalisthq\" rel=\"nofollow\">@dynalisthq<\/a><\/td>\n<td align=\"left\">Free - $8\/mo<\/td>\n<td align=\"left\">Outliner, note-taking app, checklists.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/notion.so\" rel=\"nofollow\">Notion<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/notionhq\" rel=\"nofollow\">@notionhq<\/a><\/td>\n<td align=\"left\">Free -  $8\/user\/mo<\/td>\n<td align=\"left\">Write, plan, collaborate, and get organized. Notion is all you need \u2014 in one tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.inkdrop.info\" rel=\"nofollow\">Inkdrop<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$4.99\/mo<\/td>\n<td align=\"left\">The Note-Taking App for Markdown Lovers.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-group-communicationchat-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#group-communicationchat-tools\"><\/a>Group Communication\/Chat Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/getkandan.com\" rel=\"nofollow\">Kandan<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.yammer.com\/\" rel=\"nofollow\">Yammer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/limechat.net\/\" rel=\"nofollow\">Limechat<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.flowdock.com\/\" rel=\"nofollow\">Flowdock<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free for teams up to 5 people<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.stride.com\" rel=\"nofollow\">Stride<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/slack.com\/\" rel=\"nofollow\">Slack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free - $12.5\/user<\/td>\n<td align=\"left\">Team collaboration and chat tool. Supports webhooks and has integrations with other tools.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.skype.com\" rel=\"nofollow\">Skype<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Video conference, chat and voice call VOIP application.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hangouts.google.com\/\" rel=\"nofollow\">Google Hangouts<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Video conference and chat application by Google. It is browser based and supports multiple participants on the same video call.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.gotomeeting.com\" rel=\"nofollow\">GoToMeeting<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.irccloud.com\/\" rel=\"nofollow\">IRCCloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/buddycloud.com\" rel=\"nofollow\">Buddycloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gitter.im\" rel=\"nofollow\">Gitter<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Chat for open-source and private development teams, with strong Github integration features<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/appear.in\/\" rel=\"nofollow\">appear.in<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">one click video conversations<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.convo.com\/\" rel=\"nofollow\">Convo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zoom.us\/\" rel=\"nofollow\">Zoom<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/telegram.org\/\" rel=\"nofollow\">Telegram<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/matrix.org\/\" rel=\"nofollow\">Matrix<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Decentralised Open Source chat protocol with <a href=\"https:\/\/matrix.org\/docs\/projects\/try-matrix-now.html#application-services\" rel=\"nofollow\">bridges to everywhere<\/a><\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-remote-collaboration\" class=\"anchor\" aria-hidden=\"true\" href=\"#remote-collaboration\"><\/a>Remote Collaboration<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.teamviewer.com\" rel=\"nofollow\">TeamViewer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.screenmailer.com\" rel=\"nofollow\">Screenmailer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-dns\" class=\"anchor\" aria-hidden=\"true\" href=\"#dns\"><\/a>DNS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/dyn.com\/dns\/\" rel=\"nofollow\">DynDNS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudflare.com\/\" rel=\"nofollow\">Cloudflare<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/route53\/\" rel=\"nofollow\">Amazon Route 53<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dnsimple.com\/\" rel=\"nofollow\">DNSimple<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudns.net\/\" rel=\"nofollow\">ClouDNS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/Freedns.afraid.org\" rel=\"nofollow\">FreeDNS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-status-blogsuser-alerts\" class=\"anchor\" aria-hidden=\"true\" href=\"#status-blogsuser-alerts\"><\/a>Status Blogs\/User Alerts<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.statuspage.io\/\" rel=\"nofollow\">StatusPage.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tumblr.com\/\" rel=\"nofollow\">Tumblr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hellobar.com\" rel=\"nofollow\">HelloBar<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/status.io\" rel=\"nofollow\">Status.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/statusio\" rel=\"nofollow\">@statusio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Hosted system status pages<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-forms--surveys\" class=\"anchor\" aria-hidden=\"true\" href=\"#forms--surveys\"><\/a>Forms \/ Surveys<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.wufoo.com\" rel=\"nofollow\">Wufoo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Google Forms<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.typeform.com\" rel=\"nofollow\">Typeform<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/typeform\" rel=\"nofollow\">@typeform<\/a><\/td>\n<td align=\"left\">$0\/mo - $25\/mo<\/td>\n<td align=\"left\">Ask Awesomely! Typeform makes asking questions easy, human &amp; beautiful. A user experience that makes your questions look &amp; feel great everywhere. Stimulated, inspired, excited, happy respondents boost completion rates. Gain insights with integrated analysis tools.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/qualaroo.com\" rel=\"nofollow\">Qualaroo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/qualarooinc\" rel=\"nofollow\">@qualarooinc<\/a><\/td>\n<td align=\"left\">$63\/mo - 499\/mo<\/td>\n<td align=\"left\">Qualaroo website surveys uncover customer insights that lead to better business results.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/formcarry.com\" rel=\"nofollow\">Formcarry<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0\/mo - $99\/mo<\/td>\n<td align=\"left\">Handle forms without a single line of back-end code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/formkeep.com\" rel=\"nofollow\">FormKeep<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/formkeep\" rel=\"nofollow\">@formkeep<\/a><\/td>\n<td align=\"left\">$59\/mo - $199\/mo<\/td>\n<td align=\"left\">Premium Forms Backend service with built in spam protection, data integrations to over a 1000 services, data retention policies and more.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-source-code-hosting\" class=\"anchor\" aria-hidden=\"true\" href=\"#source-code-hosting\"><\/a>Source Code Hosting<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/bitbucket.org\/\" rel=\"nofollow\">BitBucket<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.codebasehq.com\/\" rel=\"nofollow\">Codebase<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/\">GitHub<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/unfuddle.com\/\" rel=\"nofollow\">Unfuddle<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/about.gitlab.com\/\" rel=\"nofollow\">GitLab<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/launchpad.net\/\" rel=\"nofollow\">Launchpad<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tuxfamily.org\" rel=\"nofollow\">TuxFamily<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pypi.python.org\/pypi\/kforge\" rel=\"nofollow\">KForge<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.versionshelf.com\" rel=\"nofollow\">VersionShelf<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">30-day Free trial, <a href=\"https:\/\/www.versionshelf.com\/about_Free_plan\" rel=\"nofollow\">Free for students\/teachers<\/a>.<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.assembla.com\/home\" rel=\"nofollow\">Assembla<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/assembla\" rel=\"nofollow\">@assembla<\/a><\/td>\n<td align=\"left\">$24\/m - $99\/m<\/td>\n<td align=\"left\">Git hosting, all tightly integrated with your tasks. Browse files online, compare revisions, merge code and maintain a sustainable codebase.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-design-collaboration\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-collaboration\"><\/a>Design Collaboration<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.conceptshare.com\" rel=\"nofollow\">ConceptShare<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.framebench.com\" rel=\"nofollow\">Framebench<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/framebench\" rel=\"nofollow\">@framebench<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The 'smooth' way to Share, Review &amp; Discuss any file online.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.notableapp.com\/\" rel=\"nofollow\">Notable<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-paas\" class=\"anchor\" aria-hidden=\"true\" href=\"#paas\"><\/a>PaaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.heroku.com\" rel=\"nofollow\">Heroku<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">1000 Free hours\/mo for verified accounts<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudfoundry.org\" rel=\"nofollow\">Cloud Foundry<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">host it yourself, or use one of MANY hosting services<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pagodabox.io\" rel=\"nofollow\">Pagodabox<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.clever-cloud.com\" rel=\"nofollow\">Clever Cloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/cloud.google.com\/appengine\/docs\/\" rel=\"nofollow\">Google App Engine<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.openshift.com\" rel=\"nofollow\">OpenShift<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">OpenShift gives you all the tools you need to develop, host and scale your apps in the public or private cloud. Get started today.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.engineyard.com\" rel=\"nofollow\">Engine Yard<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.ctl.io\/appfog\/\" rel=\"nofollow\">AppFog<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/jelastic.com\/\" rel=\"nofollow\">Jelastic<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudbees.com\" rel=\"nofollow\">CloudBees<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/azure.microsoft.com\" rel=\"nofollow\">Microsoft Azure<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Known for IaaS<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/elasticbeanstalk\/\" rel=\"nofollow\">Amazon Web Services<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Like Azure, it's more known for IaaS<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/deis.io\" rel=\"nofollow\">Deis<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/scalingo.com\" rel=\"nofollow\">Scalingo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">European PAAS<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/eldarion.cloud\/\" rel=\"nofollow\">Eldarion<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Service for open source Kelproject that they develop in Python and Go<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-vps\" class=\"anchor\" aria-hidden=\"true\" href=\"#vps\"><\/a>VPS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.linode.com\/\" rel=\"nofollow\">Linode<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.ramnode.com\/\" rel=\"nofollow\">Ramnode<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.digitalocean.com\/\" rel=\"nofollow\">DigitalOcean<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.vultr.com\/\" rel=\"nofollow\">Vultr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.ovh.com\" rel=\"nofollow\">OVH<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.scaleway.com\/\" rel=\"nofollow\">Scaleway<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">\n<\/td><\/tr><\/tbody><\/table><h3><a id=\"user-content-heroku-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#heroku-tools\"><\/a>Heroku Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/hirefire.io\/\" rel=\"nofollow\">Hirefire<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-aws-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#aws-tools\"><\/a>AWS Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/ylastic.com\/\" rel=\"nofollow\">Ylastic<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.skeddly.com\" rel=\"nofollow\">Skeddly<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.gorillastack.com\" rel=\"nofollow\">GorillaStack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Tool for automating processes inside AWS including Cost Optimisation, Backup and more!<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-database-aas\" class=\"anchor\" aria-hidden=\"true\" href=\"#database-aas\"><\/a>Database-aaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.humongous.io\" rel=\"nofollow\">HumongouS.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">HumongouS.io is a web-based user interface (GUI) for MongoDB.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mlab.com\" rel=\"nofollow\">mLab<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.compose.com\" rel=\"nofollow\">Compose<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Compose is a fully-managed platform used by developers to deploy, host and scale databases (Elasticsearch and MongoDB.)<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/redislabs.com\" rel=\"nofollow\">RedisLabs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/RedisLabsInc\" rel=\"nofollow\">@RedisLabsInc<\/a><\/td>\n<td align=\"left\">Free - $338+\/mo<\/td>\n<td align=\"left\">RedisLabs offers fully-managed cloud service for hosting and running your redis or memcache datasets in a highly-available and scalable manner, with predictable and stable top performance.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-backend-aas\" class=\"anchor\" aria-hidden=\"true\" href=\"#backend-aas\"><\/a>Backend-aaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/firebase.google.com\" rel=\"nofollow\">Firebase<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Firebase\" rel=\"nofollow\">@Firebase<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">A powerful API to store and sync data in realtime. Build realtime mobile and web apps in minutes using client-side code and the powerful Firebase API. When data changes, apps built with Firebase update instantly across every device -- web or mobile.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/hood.ie\/\" rel=\"nofollow\">Hoodie<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.baasbox.com\" rel=\"nofollow\">BaasBox<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/loopback.io\/\" rel=\"nofollow\">LoopBack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paraio.com\" rel=\"nofollow\">Para<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/para_io\" rel=\"nofollow\">@Para_IO<\/a><\/td>\n<td align=\"left\">Free - $99\/mo<\/td>\n<td align=\"left\">A flexible and affordable backend API for busy developers.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-websockets-aas\" class=\"anchor\" aria-hidden=\"true\" href=\"#websockets-aas\"><\/a>WebSockets-aaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/pusher.com\" rel=\"nofollow\">Pusher<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.scaledrone.com\" rel=\"nofollow\">ScaleDrone<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-ops-alerts-and-scheduling\" class=\"anchor\" aria-hidden=\"true\" href=\"#ops-alerts-and-scheduling\"><\/a>Ops Alerts and Scheduling<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.pagerduty.com\/\" rel=\"nofollow\">PagerDuty<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.opsgenie.com\" rel=\"nofollow\">Opsgenie<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/opsgenie\" rel=\"nofollow\">@opsgenie<\/a><\/td>\n<td align=\"left\">$0 - $16 user\/mo<\/td>\n<td align=\"left\">We make alerts work for you. We provide the tools you need to design meaningful, actionable alerts and ensure the right people are notified.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.victorops.com\" rel=\"nofollow\">VictorOps<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$9 - $49 user\/month<\/td>\n<td align=\"left\">Incident Management to Support the DevOps Lifecycle, from First Alert to Post-Incident Review<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-video-hosting\" class=\"anchor\" aria-hidden=\"true\" href=\"#video-hosting\"><\/a>Video Hosting<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/wistia.com\" rel=\"nofollow\">Wistia<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/developer.jwplayer.com\" rel=\"nofollow\">JW Player<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Powerful suite of APIs and tools that enable video hosting, content delivery, playback, and data insights.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-knowledge-trackingwiki\" class=\"anchor\" aria-hidden=\"true\" href=\"#knowledge-trackingwiki\"><\/a>Knowledge Tracking\/Wiki<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/confluence\" rel=\"nofollow\">Confluence<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/confluence\/questions\" rel=\"nofollow\">Confluence Questions (Q&amp;A for your team)<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">bitbucket's default Wiki<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">comes with own repo, markdown\/creole, much lighter than Confluence.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/slimwiki.com\" rel=\"nofollow\">SlimWiki<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/documize.com\" rel=\"nofollow\">Documize<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-offsite-backups\" class=\"anchor\" aria-hidden=\"true\" href=\"#offsite-backups\"><\/a>Offsite Backups<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/www.tarsnap.com\" rel=\"nofollow\">Tarsnap<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.rsync.net\/\" rel=\"nofollow\">Rsync.net <\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/spideroak.com\" rel=\"nofollow\">SpiderOak<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/spideroak\" rel=\"nofollow\">@spideroak<\/a><\/td>\n<td align=\"left\">First 2gb Free - $100\/100gb\/year (check <a href=\"https:\/\/twitter.com\/spideroak\" rel=\"nofollow\">twitter<\/a> for deals)<\/td>\n<td align=\"left\">Store. Sync. Share. Privately.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-personal-machine-backups\" class=\"anchor\" aria-hidden=\"true\" href=\"#personal-machine-backups\"><\/a>Personal Machine Backups<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.crashplan.com\" rel=\"nofollow\">Crashplan<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Arq + S3\/Glacier<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-remote-workers\" class=\"anchor\" aria-hidden=\"true\" href=\"#remote-workers\"><\/a>Remote Workers<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.upwork.com\" rel=\"nofollow\">oDesk<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.Freelancer.com\/\" rel=\"nofollow\">Freelancer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/taskarmy.com\" rel=\"nofollow\">TaskArmy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/99designs.com\/\" rel=\"nofollow\">99Designs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.fiverr.com\" rel=\"nofollow\">Fiverr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.remotelyawesomejobs.com\" rel=\"nofollow\">Remotely Awesome Jobs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-deployment\" class=\"anchor\" aria-hidden=\"true\" href=\"#deployment\"><\/a>Deployment<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.chef.io\" rel=\"nofollow\">Chef<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.fabfile.org\" rel=\"nofollow\">Fabric<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/puppet.com\" rel=\"nofollow\">Puppet<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.ansible.com\" rel=\"nofollow\">Ansible<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.vagrantup.com\" rel=\"nofollow\">Vagrant<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/saltstack.com\/community\/\" rel=\"nofollow\">Salt<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/prelang.com\" rel=\"nofollow\">Prelang<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Automatic Rails Development<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ngrok.com\" rel=\"nofollow\">Ngrok<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-seo-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#seo-tools\"><\/a>SEO Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.accuranker.com\/\" rel=\"nofollow\">AccuRanker<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ahrefs.com\" rel=\"nofollow\">Ahrefs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ahrefs\" rel=\"nofollow\">@ahrefs<\/a><\/td>\n<td align=\"left\">$79\/mo - $2500\/mo<\/td>\n<td align=\"left\">Ahrefs provide a complete digital marketing suite with tools for analyzing back links, analizing websites, rank tracking, content exploring and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/serpbook.com\" rel=\"nofollow\">SerpBook<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.woorank.com\" rel=\"nofollow\">WooRank<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/woorank\" rel=\"nofollow\">@woorank<\/a><\/td>\n<td align=\"left\">Free - $49\/mo<\/td>\n<td align=\"left\">WooRank analyzes your website for optimization best practices and shows how it ranks against your competition. Its real-time brandable reports, consisting of over 150 data-points, help you to instantly spot critical issues that impact traffic, usability and conversions. Sync your analytics, social accounts and keywords for even more robust tracking.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/moz.com\" rel=\"nofollow\">Moz<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/moz\" rel=\"nofollow\">@moz<\/a><\/td>\n<td align=\"left\">$99\/mo - $599\/mo<\/td>\n<td align=\"left\">Moz provides you all the tools you need to effectively do search engine optimization. On site optimization graders, competitor tracking, back link analysis, rank trakcing and many more features available to users.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/kwfinder.com\/\" rel=\"nofollow\">KWFinder<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mangools_com\" rel=\"nofollow\">@mangools_com<\/a><\/td>\n<td align=\"left\">$30+\/mo<\/td>\n<td align=\"left\">Keyword research tool<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-api-builder\" class=\"anchor\" aria-hidden=\"true\" href=\"#api-builder\"><\/a>API Builder<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.getpostman.com\/\" rel=\"nofollow\">Postman<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/postmanclient\" rel=\"nofollow\">@postmanclient<\/a><\/td>\n<td align=\"left\">Free - $8\/mo<\/td>\n<td align=\"left\">Build, test, document and monitor your APIs faster. Used by 5 million developers and more than 100,000 companies.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/deployd.com\" rel=\"nofollow\">Deployd<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/deploydapp\" rel=\"nofollow\">@deploydapp<\/a><\/td>\n<td align=\"left\">Free (OSS)<\/td>\n<td align=\"left\">Design, build, and scale APIs for web and mobile apps in minutes instead of days.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/apiary.io\" rel=\"nofollow\">Apiary<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/apiaryio\" rel=\"nofollow\">@apiary<\/a><\/td>\n<td align=\"left\">Free - $99\/mo<\/td>\n<td align=\"left\">Powerful API Design Stack. Built for Developers. Work together to quickly design, prototype, document and test APIs.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-password-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#password-management\"><\/a>Password Management<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.lastpass.com\" rel=\"nofollow\">Lastpass<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/1password.com\" rel=\"nofollow\">1Password<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.passpack.com\" rel=\"nofollow\">Passpack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.keepassx.org\" rel=\"nofollow\">KeePassX<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\"><a href=\"https:\/\/www.eff.org\/deeplinks\/2014\/07\/protecting-your-anonymity-how-sex-workers\" rel=\"nofollow\">Recommended by EFF<\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/keepassxc.org\/\" rel=\"nofollow\">KeePassXC<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/KeePassXC\" rel=\"nofollow\">@KeePassXC<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">further development of KeePassX<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.enpass.io\" rel=\"nofollow\">Enpass<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.dashlane.com\" rel=\"nofollow\">Dashlane<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/bitwarden.com\/\" rel=\"nofollow\">Bitwarden<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free for personal use \/ Self-hosted<\/td>\n<td align=\"left\">Open Source Password Management Solutions<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-sources-of-clicksad-platforms\" class=\"anchor\" aria-hidden=\"true\" href=\"#sources-of-clicksad-platforms\"><\/a>Sources of Clicks\/Ad Platforms<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/neoreach.com\/\" rel=\"nofollow\">NeoReach<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.sitescout.com\/\" rel=\"nofollow\">SiteScout<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/advertise.bingads.microsoft.com\" rel=\"nofollow\">Bing Ads<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.taboola.com\/\" rel=\"nofollow\">Taboola<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.outbrain.com\/\" rel=\"nofollow\">Outbrain<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.google.com\/adwords\/\" rel=\"nofollow\">Google Adwords<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.facebook.com\/business\/\" rel=\"nofollow\">Facebook Advertising<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-storage\" class=\"anchor\" aria-hidden=\"true\" href=\"#storage\"><\/a>Storage<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kloudless.com\" rel=\"nofollow\">Kloudless<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kloudless\" rel=\"nofollow\">@Kloudless<\/a><\/td>\n<td align=\"left\">Free - $500\/mo<\/td>\n<td align=\"left\">Kloudless provides developers with a single cloud storage API in the place of many<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/cloudbuddy.cloud\" rel=\"nofollow\">CloudBuddy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$1\/10GB\/month<\/td>\n<td align=\"left\">SFTP Online Storage<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-task-scheduling\" class=\"anchor\" aria-hidden=\"true\" href=\"#task-scheduling\"><\/a>Task Scheduling<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.easycron.com\" rel=\"nofollow\">EasyCron<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ifttt.com\" rel=\"nofollow\">IFTTT<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/zapier.com\" rel=\"nofollow\">Zapier<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zapier\" rel=\"nofollow\">Zapier<\/a><\/td>\n<td align=\"left\">$99\/mo - $15\/mo<\/td>\n<td align=\"left\">Superpowers to get your work done. A Zap is a link between two apps (a \"trigger\" and an \"action\"). Zaps run automatically in the background every few minutes to move and manage data on your behalf. Only live Zaps count against your limit \u2014 you can have as many paused and unfinished Zaps as you'd like. For example, one Zap might be \"Send me an SMS every time I get a new email\".<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.integromat.com\" rel=\"nofollow\">Integromat<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/integromat\" rel=\"nofollow\">@integromat<\/a><\/td>\n<td align=\"left\">Free for &lt; 1000 operations or 100 MB\/mo, paid from $9\/mo<\/td>\n<td align=\"left\">Automates your tasks &amp; connects to almost any app or service. Very powerful toolset with features like error handling, iterators, aggregators, routers, functions, data stores, etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/deadmanssnitch.com\" rel=\"nofollow\">Dead Man's Snitch<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/deadmanssnitch\" rel=\"nofollow\">@DeadMansSnitch<\/a><\/td>\n<td align=\"left\">Free - $49\/mo<\/td>\n<td align=\"left\">Monitoring for scheduled tasks (like cron) and service heartbeats. Makes it easy to know when and why a particular job failed.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-documentation\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kapeli.com\/dash\" rel=\"nofollow\">Dash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/zealdocs.org\" rel=\"nofollow\">Zeal<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zealdocs\" rel=\"nofollow\">@zealdocs<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Offline documentation browser<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/devdocs.io\/\" rel=\"nofollow\">DevDocs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/DevDocs\" rel=\"nofollow\">@DevDocs<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">DevDocs combines multiple API documentations in a fast, organized, and searchable interface. You don't have to use your mouse \u2014 use keyboard shortcuts. The search supports fuzzy matching. You can search using your browser's address bar. DevDocs is Free and open source.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-engineering-metrics\" class=\"anchor\" aria-hidden=\"true\" href=\"#engineering-metrics\"><\/a>Engineering Metrics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.staticobject.com\" rel=\"nofollow\">Static Object<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/StaticObjectDev\" rel=\"nofollow\">@StaticObjectDev<\/a><\/td>\n<td align=\"left\">$295\/mo - $495\/mo<\/td>\n<td align=\"left\">Metrics that enhance transparency for engineering teams<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gitprime.com\" rel=\"nofollow\">Gitprime<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/GitPrime\" rel=\"nofollow\">@GitPrime<\/a><\/td>\n<td align=\"left\">$749\/mo - $2,549\/mo<\/td>\n<td align=\"left\">Metrics in context for data-driven engineering leaders<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-business-cards-and-print-material\" class=\"anchor\" aria-hidden=\"true\" href=\"#business-cards-and-print-material\"><\/a>Business Cards and Print Material<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.moo.com\" rel=\"nofollow\">Moo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.psprint.com\" rel=\"nofollow\">PSPrint<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/vistaprint.com\" rel=\"nofollow\">Vista Print<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-presentations--slides\" class=\"anchor\" aria-hidden=\"true\" href=\"#presentations--slides\"><\/a>Presentations \/ Slides<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/prezi.com\/\" rel=\"nofollow\">Prezi<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/gamell\/markpress\">markpress<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">HTML presentations from Markdown files<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/hakimel\/reveal.js\">Reveal.js<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Flexible HTML presenation framework for creating beautiful slides<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/slides.com\/\" rel=\"nofollow\">Slides<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free - $20\/mo<\/td>\n<td align=\"left\">Online editor built with Reveal.js that includes a graphical interface<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-fundraising--investor-relations\" class=\"anchor\" aria-hidden=\"true\" href=\"#fundraising--investor-relations\"><\/a>Fundraising \/ Investor Relations<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/captable.io\" rel=\"nofollow\">Captable.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/captable_io\" rel=\"nofollow\">@captable_io<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Captable.io is a free, fully-featured cap table management tool that enables step-by-step creation, easy collaboration &amp; sharing, convertible note &amp; option calculators, and powerful round &amp; exit modeling.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/startupdisclosure.io\" rel=\"nofollow\">Disclosure<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Helps startups fulfill investor update requests, so investors can manage and aggregate portfolio information. It replaces today's cumbersome email based processes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/notegenie.io\" rel=\"nofollow\">NoteGenie<\/a>  \/ <a href=\"https:\/\/safegenie.io\" rel=\"nofollow\">SAFEGenie<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">A calculator for exploring the impact of convertible notes and SAFEs on founder ownership.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ipo-ready.com\" rel=\"nofollow\">IPO Ready<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Quickly assess your startup's IPO readiness. Learn what it takes to prepare for an IPO.<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-use\" class=\"anchor\" aria-hidden=\"true\" href=\"#use\"><\/a>Use<\/h2>\n<p>The best ways to use this list are:<\/p>\n<ul><li>by browing the <a href=\"#contents\">contents<\/a><\/li>\n<li>by using <kbd>command<\/kbd> + <kbd>F<\/kbd> to search the contents<\/li>\n<\/ul><p>This list also uses tags to help when searching the contents:<\/p>\n<ul><li><strong>Hosted?<\/strong> <code>Hosted<\/code>, <code>Self-hosted<\/code><\/li>\n<\/ul><h2><a id=\"user-content-see-also\" class=\"anchor\" aria-hidden=\"true\" href=\"#see-also\"><\/a>See also<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/styfle\/awesome-online-ide\">Awesome Online IDEs<\/a> - A list of awesome online development environments<\/li>\n<li><a href=\"https:\/\/www.apify.com\/metamn\/2MS8r-api-https-github-com-cjbarber-toolsofthetrade\" rel=\"nofollow\">Data Extractor for Tools of The Trade<\/a> - Data extracted with Apify to JSON\/CSV\/XLS<\/li>\n<\/ul><h2><a id=\"user-content-credits\" class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h2>\n<p><strong><a href=\"https:\/\/www.chrisbarber.co\" rel=\"nofollow\">Chris Barber<\/a><\/strong><\/p>\n<p><a href=\"https:\/\/github.com\/cjbarber\"><img src=\"https:\/\/camo.githubusercontent.com\/61c3d43199fc0b9a40b5a0817d4316f273598d89\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f666f6c6c6f776572732f636a6261726265722e7376673f7374796c653d736f6369616c266c6162656c3d466f6c6c6f77253230636a626172626572\" alt=\"GitHub\" data-canonical-src=\"https:\/\/img.shields.io\/github\/followers\/cjbarber.svg?style=social&amp;label=Follow%20cjbarber\" \/><\/a> <a href=\"https:\/\/twitter.com\/chrisbarber\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d61fd278cca6002cb53a3b245074e18f5e9a6bc0\/68747470733a2f2f696d672e736869656c64732e696f2f747769747465722f666f6c6c6f772f63687269736261726265722e7376673f7374796c653d736f6369616c\" alt=\"Twitter\" data-canonical-src=\"https:\/\/img.shields.io\/twitter\/follow\/chrisbarber.svg?style=social\" \/><\/a><\/p>\n<p><strong><a href=\"https:\/\/davison.io\" rel=\"nofollow\">Craig Davison<\/a><\/strong><\/p>\n<p><a href=\"https:\/\/github.com\/davisonio\"><img src=\"https:\/\/camo.githubusercontent.com\/6b985f93d58bac679fa7f728b3eec92ea82f12ba\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f666f6c6c6f776572732f64617669736f6e696f2e7376673f7374796c653d736f6369616c266c6162656c3d466f6c6c6f772532304064617669736f6e696f\" alt=\"GitHub\" data-canonical-src=\"https:\/\/img.shields.io\/github\/followers\/davisonio.svg?style=social&amp;label=Follow%20@davisonio\" \/><\/a> <a href=\"https:\/\/twitter.com\/davisonio\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/9dc94489441aea9763d5ef56641c172de7dbca2c\/68747470733a2f2f696d672e736869656c64732e696f2f747769747465722f666f6c6c6f772f64617669736f6e696f2e7376673f7374796c653d736f6369616c\" alt=\"Twitter\" data-canonical-src=\"https:\/\/img.shields.io\/twitter\/follow\/davisonio.svg?style=social\" \/><\/a><\/p>\n<p>...and <a href=\"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade\/graphs\/contributors\">contributors<\/a>.<\/p>\n<p><strong><a href=\"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade\/blob\/master\/contributing.md\">Contributions are welcome!<\/a><\/strong><\/p>\n<\/article>","created_at":"2018-11-07T16:37:35+0000","updated_at":"2018-11-07T16:37:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":62,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/3237905?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":206,"title":"Inklet Trackpad Tablet for creativity on the go","url":"https:\/\/tenonedesign.com\/inklet.php","content":"<h2>How do I install and run Inklet?<\/h2><p><b>Step 1:<\/b> When it\u2019s done downloading, drag it to your Applications folder.<br \/><b>Step 2:<\/b> Launch Inklet. You will see a black droplet on the toolbar in the upper right-hand corner or your screen. Inklet will run in the background while you go about your business elsewhere on your laptop.<br \/><\/p><h2>How do I use Inklet with Photoshop or another drawing program?<\/h2><p><b>Step 1:<\/b> Open a drawing program such as Photoshop.<br \/><b>Step 2:<\/b> When you're ready to start Inklet, click on the droplet icon and select 'Start Inklet,' or use the hotkey (the default hotkey is control+option+i). Draw with your Pogo on the trackpad. If you don't have a Pogo, you can change the drawing tool to 'finger' in the Inklet drawing preferences.<br \/><b>Step 3:<\/b> Move your Pogo lightly around on the trackpad to allow the cursor to hover without inking. Increase pressure to draw. heavier pressure will result in a thicker stroke.<\/p><h2>I\u2019m drawing, but nothing is showing up?!<\/h2><p>Make sure you have a drawing program open under Inklet.\n\t\t\t\t Compatible programs are listed a few questions down.\n\t\t\t\t Also, make sure you have a pen tool selected in your program, and you\u2019re not trying to draw with white on a white background.\n\t\t\t\t That never works.<\/p><h2>How do I register Inklet?<\/h2><p>You will be given an activation key in your purchase confirmation email.\n\t\t\t\tActivating with this key will remove the demo message in Inklet.\n\t\t\t\tClick the black inklet icon in your top right toolbar, and select 'Preferences...'\n\t\t\t\tNow, select the Register tab, and paste your key into the text field.\n\t\t\t\tIf you don't see your confirmation email, make sure to check your spam folder.<\/p><h2>Can I draw without a Pogo?<\/h2><p>Yes, just change the drawing tool to 'finger' in the Inklet drawing preferences.\n\t\t\t\t It\u2019s not as easy, but it\u2019s way possible.<\/p><h2>Can I use Inklet with an external Magic Trackpad?<\/h2><p>Yes. If an internal trackpad is not available, Inklet will automatically detect and use a connected Magic Trackpad.\n\t\t\t\tInklet works best on internal trackpads, as pressure sensitivity is limited on the Magic Trackpad.<\/p><h2>How might I uninstall Inklet?<\/h2><p>Drag Inklet from your Applications folder into the trash<\/p><p>If you don\u2019t see an answer to your question here, <a href=\"https:\/\/tenonedesign.com\/about.php\">let us know!<\/a><\/p>","created_at":"2018-11-07T16:37:35+0000","updated_at":"2018-11-07T16:37:35+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"tenonedesign.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":204,"title":"Structure and Interpretation of Computer Programs, 2e: Top","url":"https:\/\/sarabander.github.io\/sicp\/html\/index.xhtml","content":"Structure and Interpretation of Computer Programs, 2e: Top\n<section><a href=\"#pagetop\" accesskey=\"t\">\u21e1<\/a><p><abbr>Second Edition<\/abbr><br \/>Unofficial Texinfo Format 2.andresraba6.6<\/p><strong>Harold Abelson and Gerald Jay Sussman with Julie Sussman, foreword by Alan J. Perlis<\/strong><p>\u00a91996 by The Massachusetts Institute of Technology\n<\/p><p>This work is licensed under a Creative Commons Attribution-ShareAlike 4.0\nInternational License (<a href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/4.0\/\"><abbr>CC BY-SA 4.0<\/abbr><\/a>).\nBased on a work at <a href=\"http:\/\/mitpress.mit.edu\/sicp\/\">mitpress.mit.edu<\/a>.\n<\/p><p>The <abbr>MIT<\/abbr> Press<br \/>Cambridge, Massachusetts<br \/>London, England\n<\/p><p>McGraw-Hill Book Company<br \/>New York, St. Louis, San Francisco,<br \/>Montreal, Toronto\n<\/p><p>This book is one of a series of texts written by faculty of the Electrical\nEngineering and Computer Science Department at the Massachusetts Institute of\nTechnology.  It was edited and produced by The <abbr>MIT<\/abbr> Press under a\njoint production-distribution arrangement with the McGraw-Hill Book Company.\n<\/p><p>Unofficial Texinfo Format <a href=\"http:\/\/sicpebook.wordpress.com\">2.andresraba6.6<\/a> (September 16, 2015),\nbased on <a href=\"http:\/\/www.neilvandyke.org\/sicp-texi\/\">2.neilvandyke4<\/a> (January 10, 2007).\n<\/p><div class=\"shortcontents\"><ul class=\"no-bullet\"><li><a id=\"stoc-Unofficial-Texinfo-Format\" href=\"#toc-Unofficial-Texinfo-Format\">Unofficial Texinfo Format<\/a><\/li>\n<li><a id=\"stoc-Dedication-1\" href=\"#toc-Dedication-1\">Dedication<\/a><\/li>\n<li><a id=\"stoc-Foreword-1\" href=\"#toc-Foreword-1\">Foreword<\/a><\/li>\n<li><a id=\"stoc-Preface-to-the-Second-Edition\" href=\"#toc-Preface-to-the-Second-Edition\">Preface to the Second Edition<\/a><\/li>\n<li><a id=\"stoc-Preface-to-the-First-Edition\" href=\"#toc-Preface-to-the-First-Edition\">Preface to the First Edition<\/a><\/li>\n<li><a id=\"stoc-Acknowledgments-1\" href=\"#toc-Acknowledgments-1\">Acknowledgments<\/a><\/li>\n<li><a id=\"stoc-Building-Abstractions-with-Procedures\" href=\"#toc-Building-Abstractions-with-Procedures\">1 Building Abstractions with Procedures<\/a><\/li>\n<li><a id=\"stoc-Building-Abstractions-with-Data\" href=\"#toc-Building-Abstractions-with-Data\">2 Building Abstractions with Data<\/a><\/li>\n<li><a id=\"stoc-Modularity_002c-Objects_002c-and-State\" href=\"#toc-Modularity_002c-Objects_002c-and-State\">3 Modularity, Objects, and State<\/a><\/li>\n<li>\n<\/li><li><a id=\"stoc-Computing-with-Register-Machines\" href=\"#toc-Computing-with-Register-Machines\">5 Computing with Register Machines<\/a><\/li>\n<li><a id=\"stoc-References-1\" href=\"#toc-References-1\">References<\/a><\/li>\n<li><a id=\"stoc-List-of-Exercises\" href=\"#toc-List-of-Exercises\">List of Exercises<\/a><\/li>\n<li><a id=\"stoc-List-of-Figures\" href=\"#toc-List-of-Figures\">List of Figures<\/a><\/li>\n<li><a id=\"stoc-Term-Index-1\" href=\"#toc-Term-Index-1\">Term Index<\/a><\/li>\n<li><a id=\"stoc-Colophon-1\" href=\"#toc-Colophon-1\">Colophon<\/a><\/li>\n<\/ul><\/div><\/section><a href=\"#pagebottom\" accesskey=\"b\">\u21e3<\/a>","created_at":"2018-11-07T16:37:10+0000","updated_at":"2018-11-07T16:37:10+0000","annotations":[],"mimetype":"application\/xhtml+xml","language":"en","reading_time":0,"domain_name":"sarabander.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":203,"title":"wordpress rest api custom endpoint don't serialize response at DuckDuckGo","url":"https:\/\/duckduckgo.com\/?ia=web&q=wordpress+rest+api+custom+endpoint+don%27t+serialize+response&t=ffab","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:36:35+0000","updated_at":"2018-11-07T16:36:35+0000","annotations":[],"reading_time":0,"domain_name":"duckduckgo.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":202,"title":"Howdy, Cow Pokes: Original Star Wars Trilogy Imagined As Spaghetti Western Movie Posters (I Love Spaghetti FYI)","url":"https:\/\/geekologie.com\/2012\/08\/howdy-cow-poke-original-star-wars-trilog.php","content":"<p><img alt=\"star-wars-western-1.jpg\" src=\"http:\/\/www.geekologie.com\/2012\/08\/20\/star-wars-western-1.jpg\" width=\"640\" height=\"960\" class=\"mt-image-none\" \/><\/p><p>This is a series of posters by artist <a href=\"http:\/\/timothyandersonart-printstore.blogspot.com\/\">Tim Anderson<\/a> (links to personal website where you can buy prints, $40) imagining<a href=\"http:\/\/blastr.com\/2012\/08\/3-awesome-star-wars-poste.php\"> the first three Star Wars movies as spaghetti westerns<\/a> by Italian director Sergio Leone.  Would Leone have had enough decency to not make any prequels?   Probably.  *spins gun on finger, accidentally shoots Jar Jar in the face*  Woopsie doopsie!  \"That wasn't an accident, was it?\" <em> Is a tauntaun penis ever warm to the touch?<\/em><\/p><p>Hit the jump for the other two.<\/p><p><img alt=\"star-wars-western-2.jpg\" src=\"http:\/\/www.geekologie.com\/2012\/08\/20\/star-wars-western-2.jpg\" width=\"640\" height=\"960\" class=\"mt-image-none\" \/><\/p><p><img alt=\"star-wars-western-3.jpg\" src=\"http:\/\/www.geekologie.com\/2012\/08\/20\/star-wars-western-3.jpg\" width=\"640\" height=\"960\" class=\"mt-image-none\" \/><\/p><p>Thanks to XtraFeely, who has like, more nerve endings than a regular person or something.<\/p><p><a href=\"https:\/\/facebook.com\/sharer\/sharer.php?text=Howdy, Cow Pokes: Original Star Wars Trilogy Imagined As Spaghetti Western Movie Posters (I Love Spaghetti FYI)&amp;u=http:\/\/geekologie.com\/2012\/08\/howdy-cow-poke-original-star-wars-trilog.php\" class=\"fb-circle\">Facebook<\/a>\n        \n        <a href=\"http:\/\/www.pinterest.com\/pin\/create\/button\/?description=Howdy, Cow Pokes: Original Star Wars Trilogy Imagined As Spaghetti Western Movie Posters (I Love Spaghetti FYI)&amp;media=http:\/\/geekologie.com\/2012\/08\/20\/star-wars-western-1.jpg&amp;url=http:\/\/geekologie.com\/2012\/08\/howdy-cow-poke-original-star-wars-trilog.php\" class=\"pinterest-circle\">Pinterest<\/a><\/p><p>Read More: , , , , , , , , , <\/p>","created_at":"2018-11-07T16:36:34+0000","updated_at":"2018-11-07T16:36:34+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"geekologie.com","preview_picture":"http:\/\/geekologie.com\/2012\/08\/20\/star-wars-western-3.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":201,"title":"You Suck at Excel with Joel Spolsky","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=0nbkaYsR94c","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/0nbkaYsR94c?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:35:14+0000","updated_at":"2018-11-07T16:35:14+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/0nbkaYsR94c\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":200,"title":"Photography SEO \u2022 Yoast","url":"https:\/\/yoast.com\/photography-seo\/","content":"<p>Have you read my post on <a title=\"Optimizing images for SEO\" href=\"https:\/\/yoast.com\/image-seo\/\">image SEO<\/a>? On that post, I got this\u00a0comment from <a href=\"http:\/\/www.jerryandlois.com\/\">Jerry<\/a>, a professional photographer: \u201c<em>In galleries, do you have any suggestions about what the ALT text should look like? The main concern is that we don\u2019t want to be perceived as keyword spamming; using the same keyword in all the images of a given gallery.\u201d<\/em><\/p><p>That\u2019s a good question about <em>photography SEO<\/em>. When thinking about the optimization of photo galleries for Google, I could actually only come up with one main comment: write more text to accompany these photos. In this post, I will go over a number of random photos and explain what direction I\u2019d go with that text and photography SEO. There might be some personal preference in this, but I\u2019m sure this will help you to find a way that will allow you to easily add that text yourself as well. Keep in mind that you don\u2019t need to be a copywriter to <a title=\"Content SEO\" href=\"https:\/\/yoast.com\/content-seo\/\">write appealing text<\/a> ;)<\/p><h3>Table of Contents<\/h3><ol class=\"toc\"><li><a href=\"https:\/\/yoast.com\/photography-seo\/#whyrank\">Why would a photographer want to rank with\u00a0images<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#singlephotoseo\">Single photos and SEO<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#photogalleryseo\">Photo galleries and SEO<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#weddingphotographyseo\">Wedding photography SEO<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#stockphotoseo\">Stockphotos and SEO<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#foodphotographyseo\">Food photography and SEO<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#exifimageobject\">EXIF and ImageObject<\/a><\/li>\n<li><a href=\"https:\/\/yoast.com\/photography-seo\/#whyphotographershateseo\">Why photographers hate SEO<\/a><\/li>\n<\/ol><h2 id=\"whyrank\">Why would a photographer want to rank with\u00a0images<\/h2><p>Recently, a friend of mine received a claim\u00a0for using an apparently copyright protected image. This photographer stuffed his website with copyright warnings, but also exposed all of his photos of very common objects\u00a0on <a href=\"http:\/\/flickr.com\">Flickr.com<\/a>\u00a0(although using an \u2018All rights reserved\u2019 status). I\u2019m curious what the outcome will be. We all know an image will be used at some point\u00a0if you don\u2019t make sure it\u2019s useless unless bought.\u00a0Watermark your photos if you don\u2019t want people to use them. Otherwise, people will find your photo and use it on a blog or whatever. If you\u2019re not watermarking, that\u00a0could be one reason to make \u2019em rank: making money of people stealing your photo. But I think that\u2019s an odd business model ;-)<\/p><p>I think the main reason you want your images to rank is because you are proud of your photos. And you want people to notice these, so they will hire you. That should be the main reason to invest in photography SEO in my book anyway. Your photos sell you as a photographer.<\/p><p>In the sections below, I\u2019ll go over five common photography cases\u00a0and explain how I would optimize these for Google.<\/p><p><img class=\"nopadding aligncenter wp-image-326118 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-550x367.jpg\" alt=\"Colorful Apiary\" width=\"550\" height=\"367\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-550x367.jpg 550w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-250x167.jpg 250w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-171x114.jpg 171w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-342x228.jpg 342w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-205x137.jpg 205w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/apiary-beautiful-sunny-day-bees-winter-sleep-411x274.jpg 411w\" \/><\/p><p>I mentioned optimizing single images \/ photos in my <a title=\"Optimizing images for SEO\" href=\"https:\/\/yoast.com\/image-seo\/\">Image SEO<\/a> post already. For photography SEO, obviously a lot is the same. Here\u2019s a quick summary:<\/p><ul><li>Pick the right file name for your image<\/li>\n<li>Make sure image dimension match the image size as displayed<\/li>\n<li>Reduce file size for faster loading<\/li>\n<li>Add a caption for easier scanning of the page<\/li>\n<li>Use an image alt text, title text is optional<\/li>\n<li>Add OpenGraph and Twitter Card tags for the image<\/li>\n<li>Don\u2019t break the left reading line using an image<\/li>\n<li>Use images in your XML sitemaps<\/li>\n<\/ul><p>In all this, it really helps to add text with that photo. The article\u00a0mainly mentions images to make\u00a0text richer, but it works\u00a0the other way around as well. If you want to rank your image, you need to add relevant text.<\/p><h2 id=\"photogalleryseo\">Photo galleries and SEO<\/h2><p><img class=\"nopadding aligncenter wp-image-326117 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery-550x200.jpg\" alt=\"BMX Skate Park Gallery\" width=\"550\" height=\"200\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery.jpg 550w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery-250x91.jpg 250w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery-180x65.jpg 180w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery-360x131.jpg 360w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery-235x85.jpg 235w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/bmx-skate-park-gallery-470x171.jpg 470w\" \/><\/p><p>You must have wondered how to rank all of these photos. They are pretty much the same. In the example, you see a BMX rider in a skate park. Jumping. I\u2019m sorry if you are a BMX rider and know all the names and nick names of these jumps, for the larger part of the world it\u2019s a BMX rider jumping. How to optimize\u00a0your photography SEO for that?<\/p><p>Well, this is\u00a0a great reason to add text! You have probably spent the day with these guys, tested a number of lenses, tried to play with\u00a0white balance and perhaps even some ISO values. You have created great stills from dynamic poses. Changed\u00a0Motion blur and shutter speed during the day\u00a0(and that is where my photography knowledge ends). There is a story in this and that story shows your knowledge. It could show your passion for this niche of sports photography. It will tell the visitor in words that if they want great photos like that, they should contact\u00a0you. The gallery page could already have a\u00a0<a title=\"Calling to the next action\" href=\"https:\/\/yoast.com\/calling-next-action\/\">call-to-action<\/a> for that reason.<\/p><p>In short: If you have created a gallery, write about it.<\/p><p>Back to Jerry\u2019s specific question about the ALT text in galleries, as mentioned in the first paragraph of this article. In an ideal world, you would create unique ALT texts for all these images. Would that be keyword spamming? Not if you can use <a title=\"Why focus on long tail keywords?\" href=\"https:\/\/yoast.com\/focus-on-long-tail-keywords\/\">long tail keywords<\/a> per photo, instead of repeating the same ALT text over and over again. Note that if your gallery is a hundred photos, you should ask yourself a) if that gallery isn\u2019t too large, and b) if it wouldn\u2019t be easier to set up a separate page to rank for the keyword (if any) and not try to rank that specific gallery page at all.<\/p><h2 id=\"weddingphotographyseo\">Wedding photography SEO<\/h2><p><img class=\"nopadding noborder aligncenter wp-image-326105 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-550x367.jpg\" alt=\"wedding day stealing a moment alone\" width=\"550\" height=\"367\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-550x367.jpg 550w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-250x167.jpg 250w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-171x114.jpg 171w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-342x228.jpg 342w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-205x137.jpg 205w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/wedding-day-stealing-a-moment-alone-411x274.jpg 411w\" \/><\/p><p>Oh, the emotion on that wedding day. So beautifully captured in photos. A couple stealing a quiet moment with only the photographer as a witness. It\u2019s just that, I think. What makes you a great wedding photographer? Are you the always present kind that parties along with the guests and is invisible for that reason? Or are you actively involved in the day and present photo moments along the way? I can\u2019t be the judge of your USP, but your subjects, the bride and groom, probably can.<\/p><p>We have done <a title=\"Testimonials: increase your visitor\u2019s trust\" href=\"https:\/\/yoast.com\/creating-trust-testimonials\/\">a nice post on testimonials and increasing trust<\/a> a while back, and I think it works the same with wedding photography. If your wedding couple\u00a0doesn\u2019t mind their photos used on the internet, these great images could lure people to your website looking for the same style of photography. It\u2019s not per se SEO that is optimized this\u00a0way, but if you add a genuine testimonial to the wedding gallery \/ photo(s), that would help your conversion, right?<\/p><p>Of course you will also write an additional text about\u00a0how you \u2018operated\u2019 during the day.<\/p><h2 id=\"stockphotoseo\">Stockphotos and SEO<\/h2><p><img class=\"nopadding aligncenter wp-image-326108 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-550x367.jpg\" alt=\"Couple in front of a Graffiti wall\" width=\"550\" height=\"367\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-550x367.jpg 550w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-250x167.jpg 250w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-171x114.jpg 171w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-342x228.jpg 342w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-205x137.jpg 205w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/couple-graffiti-wall-411x274.jpg 411w\" \/><\/p><p>If you\u2019re a professional in the stock photo business, you have probably done some photography SEO\u00a0already. There are some general guidelines:<\/p><ul><li>You\u2019ve probably chosen one or more\u00a0niches;<\/li>\n<li>you\u2019re doing all the right image optimization and more, as\u00a0mentioned in <a title=\"Optimizing images for SEO\" href=\"https:\/\/yoast.com\/image-seo\/\">our article on image seo<\/a>;<\/li>\n<li>of course, you\u2019re trying to rank watermarked images, not the originals.<\/li>\n<\/ul><p>There is one more issue to consider when thinking\u00a0about photography SEO. If your main business is selling stock photos, you probably don\u2019t want to do that via your own website. There are plenty of well visited websites that specialize in selling stock photos. Images in this article\u00a0are\u00a0from various photographers at\u00a0<a href=\"https:\/\/yayimages.com\/\">YAY Images<\/a>, a source for photos we like to use. Using websites like that, and for instance <a href=\"http:\/\/www.shutterstock.com\">Shutterstock<\/a> or <a href=\"http:\/\/www.istockphoto.com\/\">iStockphoto<\/a> will bring much more attention to people looking for these kind of photos.<\/p><p>If you want your photo to do well in these on-site search engines, you want to focus on the image description. Use the right keywords. The file name will probably be changed anyway. If you can tag the image, do so using the right tags, not a surplus of tags.<\/p><p>One more thing: most websites allow for a photographer\u2019s bio as well. Tell me who\u00a0you are. If I can relate to your story, it\u2019s so much easier to spend money on your photos.<\/p><h2 id=\"foodphotographyseo\">Food photography and SEO<\/h2><p><img class=\"nopadding aligncenter wp-image-326119 size-large\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-550x367.jpg\" alt=\"Blueberry Muffins\" width=\"550\" height=\"367\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-550x367.jpg 550w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-250x167.jpg 250w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-171x114.jpg 171w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-342x228.jpg 342w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-205x137.jpg 205w, https:\/\/yoast.com\/app\/uploads\/2015\/02\/delicious-homemade-blueberry-muffins-fresh-blueberries-spilling-from-wooden-spoon-wicker-basket-411x274.jpg 411w\" \/><\/p><p>Food photography SEO is a whole different ball game. The topic is so popular that you want to make sure all images are set for sharing. Yes, that means giving away (a number of) these for free. Especially for photos that have baked, cooked, or otherwise processed food, the recipe will help your rankings a lot. Even if it\u2019s a magazine shoot, you can probably\u00a0either use the recipe or come up with a related one.<\/p><p>Social platforms that help a lot with \u2018spreading\u2019 your image on the web are obviously <a title=\"Pinterest marketing for your business\" href=\"https:\/\/yoast.com\/pinterest-marketing\/\">Pinterest<\/a> and <a title=\"Instagram for business?\" href=\"https:\/\/yoast.com\/instagram-business\/\">Instagram<\/a>. For for instance Pinterest, it pays to add OpenGraph data to your image. If you have never heard of that, you are probably new on this website. OpenGraph image tags look something like this:<\/p><p><code>&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/photo.jpg\" \/&gt;<\/code><\/p><p>Be sure to add at least that one. On a side note: for the recipe, you should use <a href=\"http:\/\/schema.org\/Recipe\">Schema.org markup<\/a>\u00a0to improve your photography SEO as well.<\/p><h2 id=\"exifimageobject\">EXIF and ImageObject<\/h2><p>Although I wanted to write a post that would make photographers think about texts, I can\u2019t leave out\u00a0a bit of technical optimization. Especially for photographers, that understand EXIF\u00a0and other data and would like to include that as well in their online images, please check <a href=\"http:\/\/schema.org\/ImageObject\">http:\/\/schema.org\/ImageObject<\/a>.<\/p><p>There is another reason to include that EXIF data: for SEO. In the video below, Matt Cutts from Google clearly states that Google is very capable of reading this EXIF data and according to him, Google \u2018reserves the right to use it as a ranking factor\u2019. So make sure that your EXIF data is optimized for your keyword, or the subject of your photo, as well.<\/p><div class=\"videowrapper\"><iframe width=\"600\" height=\"338\" src=\"https:\/\/www.youtube.com\/embed\/GMf6FmRus2M?feature=oembed&amp;enablejsapi=1&amp;origin=https:\/\/yoast.com\/wp&amp;modestbranding=1&amp;hd=1&amp;rel=0&amp;\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><\/div><p>Besides <code>exifData<\/code>, you can also include things like\u00a0<code>copyrightHolder<\/code> and <code>copyrightYear<\/code>. It will make it easier for\u00a0search engines to grab that data. Will the copyright\u00a0make you rank better? Probably not. But I would understand the wish to include that data.<\/p><p>One more thing regarding EXIF data: we frequently recommend using\u00a0file size reducing software like <a href=\"http:\/\/www.jpegmini.com\/\">JPEGMini<\/a> and <a href=\"https:\/\/imageoptim.com\/\">ImageOptim<\/a>. Exif data is preserved in most of these applications,\u00a0but \u00a0please check the FAQ or changelog on their websites, <a href=\"http:\/\/www.jpegmini.com\/main\/faq\">like here<\/a>, to make sure of this.<\/p><h2 id=\"whyphotographershateseo\">Why photographers hate SEO<\/h2><p>I might be mistaken, but in my opinion most photographers hate SEO. Why do I think photographers hate SEO? Because it simply has less to do with photography, and loads to do with extra, textual content or code. Maintaining a website isn\u2019t simple. You preferably need to know a bit of everything. We have\u00a0<a title=\"Optimize your WordPress site or shop with our Gold SEO Review\" href=\"https:\/\/yoast.com\/hire-us\/website-review\/gold-seo-review\/\">reviewed<\/a> a lot of photography websites over the past few years. To be honest, it would be a lot if two out of all these websites actually had decent content accompanying the images.<\/p><p>It all comes down to this: if you want your photos to tell a story, please tell that story to Google in writing.<\/p><p class=\"readmore\">Read more: <a data-prefix=\"Read more\" href=\"https:\/\/yoast.com\/image-seo\/\">Image SEO \u00bb<\/a><\/p>","created_at":"2018-11-07T16:24:13+0000","updated_at":"2018-11-07T16:24:13+0000","published_at":"2015-02-18T14:10:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"yoast.com","preview_picture":"https:\/\/yoast.com\/app\/uploads\/2015\/02\/Photography-SEO-1.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":197,"title":"Ask HN: Clever ways to run a media-heavy website on a budget?","url":"https:\/\/news.ycombinator.com\/item?id=16783852","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16783852\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16783852\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=16783852\" class=\"storylink\">Ask HN: Clever ways to run a media-heavy website on a budget?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        181 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=wild_preference\" class=\"hnuser\">wild_preference<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16783852\">7 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16783852&amp;goto=item%3Fid%3D16783852\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Clever%20ways%20to%20run%20a%20media-heavy%20website%20on%20a%20budget%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Clever%20ways%20to%20run%20a%20media-heavy%20website%20on%20a%20budget%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16783852&amp;auth=f545e81315347c0bc0713cda0f878c8e186d84a8\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16783852\">97\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>I've been taking a break from working to relax and visit some friends in a few countries. So I don't have a lot of money, but with my newfound free time, I've been inspired to take a shot at a few ideas I've always wanted to build.<p>One idea is particularly media heavy. I have 5+ TB of media that I'd like to build something around.<\/p><p>Services like S3 are pretty expensive, especially bandwidth. Backblaze has been trying to market their storage solution (https:\/\/www.backblaze.com\/b2\/cloud-storage-pricing.html) against S3.<\/p><p>But I can get even cheaper by using low quality hardware and bandwidth like Kimsufi's 2TB machines for $10\/mo (https:\/\/www.kimsufi.com\/en\/servers.xml).<\/p><p>Does anyone have another ideas for stretching a dollar?<\/p><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:24:12+0000","updated_at":"2018-11-07T16:24:12+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":198,"title":"A thread on on the beauty of procedurally generated maps","url":"https:\/\/news.ycombinator.com\/item?id=16744575","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16744575\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16744575\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/twitter.com\/ptychomancer\/status\/980968298002006016\" class=\"storylink\">A thread on on the beauty of procedurally generated maps<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=twitter.com\">twitter.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        293 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=hownottowrite\" class=\"hnuser\">hownottowrite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16744575\">7 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16744575&amp;goto=item%3Fid%3D16744575\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=A%20thread%20on%20on%20the%20beauty%20of%20procedurally%20generated%20maps&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=A%20thread%20on%20on%20the%20beauty%20of%20procedurally%20generated%20maps\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16744575&amp;auth=29a5eb07f43f1957aacb74369ddf5f14c9a88f72\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16744575\">27\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:24:12+0000","updated_at":"2018-11-07T16:24:12+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":199,"title":"mxgmn\/WaveFunctionCollapse","url":"https:\/\/github.com\/mxgmn\/WaveFunctionCollapse\/blob\/master\/README.md","content":"<p>This program generates bitmaps that are locally similar to the input bitmap.<\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/3ea3b800ee9de783abea49612a40beeda97879de\/687474703a2f2f692e696d6775722e636f6d2f67317947764c372e706e67\"><img alt=\"main collage\" src=\"https:\/\/camo.githubusercontent.com\/3ea3b800ee9de783abea49612a40beeda97879de\/687474703a2f2f692e696d6775722e636f6d2f67317947764c372e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/g1yGvL7.png\" \/><\/a><\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/dc39c61e02aa67abd0f923628cf241120d14f517\/687474703a2f2f692e696d6775722e636f6d2f734e75425653722e676966\"><img alt=\"main gif\" src=\"https:\/\/camo.githubusercontent.com\/dc39c61e02aa67abd0f923628cf241120d14f517\/687474703a2f2f692e696d6775722e636f6d2f734e75425653722e676966\" data-canonical-src=\"http:\/\/i.imgur.com\/sNuBVSr.gif\" \/><\/a><\/p>\n<p>Local similarity means that<\/p>\n<ul><li>(C1) Each NxN pattern of pixels in the output should occur at least once in the input.<\/li>\n<li>(Weak C2) Distribution of NxN patterns in the input should be similar to the distribution of NxN patterns over a sufficiently large number of outputs. In other words, probability to meet a particular pattern in the output should be close to the density of such patterns in the input.<\/li>\n<\/ul><p>In the examples typical value of N is 3.<\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/c9a04da9ed7815de79b3f2236cd32d3e2dffc68f\/687474703a2f2f692e696d6775722e636f6d2f4b554c475838362e706e67\"><img alt=\"local similarity\" src=\"https:\/\/camo.githubusercontent.com\/c9a04da9ed7815de79b3f2236cd32d3e2dffc68f\/687474703a2f2f692e696d6775722e636f6d2f4b554c475838362e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/KULGX86.png\" \/><\/a><\/p>\n<p>WFC initializes output bitmap in a completely unobserved state, where each pixel value is in superposition of colors of the input bitmap (so if the input was black &amp; white then the unobserved states are shown in different shades of grey). The coefficients in these superpositions are real numbers, not complex numbers, so it doesn't do the actual quantum mechanics, but it was inspired by QM. Then the program goes into the observation-propagation cycle:<\/p>\n<ul><li>On each observation step an NxN region is chosen among the unobserved which has the lowest Shannon entropy. This region's state then collapses into a definite state according to its coefficients and the distribution of NxN patterns in the input.<\/li>\n<li>On each propagation step new information gained from the collapse on the previous step propagates through the output.<\/li>\n<\/ul><p>On each step the overall entropy decreases and in the end we have a completely observed state, the wave function has collapsed.<\/p>\n<p>It may happen that during propagation all the coefficients for a certain pixel become zero. That means that the algorithm has run into a contradiction and can not continue. The problem of determining whether a certain bitmap allows other nontrivial bitmaps satisfying condition (C1) is NP-hard, so it's impossible to create a fast solution that always finishes. In practice, however, the algorithm runs into contradictions surprisingly rarely.<\/p>\n<p>Wave Function Collapse algorithm has been implemented in <a href=\"https:\/\/github.com\/math-fehr\/fast-wfc\">C++<\/a>, <a href=\"https:\/\/github.com\/ikarth\/wfc_python\">Python<\/a>, <a href=\"https:\/\/gitlab.com\/nanodeath\/kollapse\" rel=\"nofollow\">Kotlin<\/a>, <a href=\"https:\/\/github.com\/sdleffler\/collapse\">Rust<\/a>, <a href=\"https:\/\/github.com\/Mitim-84\/WFC-Gen\">Haxe<\/a>, <a href=\"https:\/\/github.com\/kchapelier\/wavefunctioncollapse\">JavaScript<\/a> and adapted to <a href=\"https:\/\/selfsame.itch.io\/unitywfc\" rel=\"nofollow\">Unity<\/a>. You can download official executables from <a href=\"https:\/\/exutumno.itch.io\/wavefunctioncollapse\" rel=\"nofollow\">itch.io<\/a> or <a href=\"http:\/\/www.kchapelier.com\/wfc-example\/overlapping-model.html\" rel=\"nofollow\">run it in the browser<\/a>. WFC generates levels in <a href=\"https:\/\/www.badnorth.com\/\" rel=\"nofollow\">Bad North<\/a>, <a href=\"https:\/\/store.steampowered.com\/app\/333640\/Caves_of_Qud\/\" rel=\"nofollow\">Caves of Qud<\/a>, <a href=\"https:\/\/arcadia-clojure.itch.io\/proc-skater-2016\" rel=\"nofollow\">several<\/a> <a href=\"https:\/\/arcadia-clojure.itch.io\/swapland\" rel=\"nofollow\">smaller<\/a> <a href=\"https:\/\/marian42.itch.io\/wfc\" rel=\"nofollow\">games<\/a> and many prototypes. It led to <a href=\"https:\/\/adamsmith.as\/papers\/wfc_is_constraint_solving_in_the_wild.pdf\" rel=\"nofollow\">new<\/a> <a href=\"https:\/\/hal.inria.fr\/hal-01706539v3\/document\" rel=\"nofollow\">research<\/a>. For <a href=\"https:\/\/twitter.com\/OskSta\/status\/784847588893814785\" rel=\"nofollow\">more<\/a> <a href=\"https:\/\/twitter.com\/dwtw\/status\/810166761270243328\" rel=\"nofollow\">related<\/a> <a href=\"https:\/\/github.com\/mewo2\/oisin\">work<\/a>, <a href=\"https:\/\/trasevol.dog\/2017\/09\/01\/di19\/\" rel=\"nofollow\">explanations<\/a>, <a href=\"http:\/\/oskarstalberg.com\/game\/wave\/wave.html\" rel=\"nofollow\">interactive demos<\/a>, <a href=\"https:\/\/www.dropbox.com\/s\/zeiat1w8zre9ro8\/Knots%20breakdown.png?dl=0\" rel=\"nofollow\">guides<\/a>, <a href=\"http:\/\/www.procjam.com\/tutorials\/wfc\/\" rel=\"nofollow\">tutorials<\/a> and <a href=\"https:\/\/twitter.com\/ExUtumno\/status\/895684431477747715\" rel=\"nofollow\">examples<\/a> see the <a href=\"https:\/\/github.com\/mxgmn\/WaveFunctionCollapse#notable-ports-forks-and-spinoffs\">ports, forks and spinoffs section<\/a>.<\/p>\n<p>Watch a video demonstration of WFC algorithm on YouTube: <a href=\"https:\/\/youtu.be\/DOQTr2Xmlz0\" rel=\"nofollow\">https:\/\/youtu.be\/DOQTr2Xmlz0<\/a><\/p>\n<h2><a id=\"user-content-algorithm\" class=\"anchor\" aria-hidden=\"true\" href=\"#algorithm\"><\/a>Algorithm<\/h2>\n<ol><li>Read the input bitmap and count NxN patterns.\n<ol><li>(optional) Augment pattern data with rotations and reflections.<\/li>\n<\/ol><\/li>\n<li>Create an array with the dimensions of the output (called \"wave\" in the source). Each element of this array represents a state of an NxN region in the output. A state of an NxN region is a superpostion of NxN patterns of the input with boolean coefficients (so a state of a pixel in the output is a superposition of input colors with real coefficients). False coefficient means that the corresponding pattern is forbidden, true coefficient means that the corresponding pattern is not yet forbidden.<\/li>\n<li>Initialize the wave in the completely unobserved state, i.e. with all the boolean coefficients being true.<\/li>\n<li>Repeat the following steps:\n<ol><li>Observation:\n<ol><li>Find a wave element with the minimal nonzero entropy. If there is no such elements (if all elements have zero or undefined entropy) then break the cycle (4) and go to step (5).<\/li>\n<li>Collapse this element into a definite state according to its coefficients and the distribution of NxN patterns in the input.<\/li>\n<\/ol><\/li>\n<li>Propagation: propagate information gained on the previous observation step.<\/li>\n<\/ol><\/li>\n<li>By now all the wave elements are either in a completely observed state (all the coefficients except one being zero) or in the contradictive state (all the coefficients being zero). In the first case return the output. In the second case finish the work without returning anything.<\/li>\n<\/ol><h2><a id=\"user-content-tilemap-generation\" class=\"anchor\" aria-hidden=\"true\" href=\"#tilemap-generation\"><\/a>Tilemap generation<\/h2>\n<p>The simplest nontrivial case of the algorithm is when NxN=1x2 (well, NxM). If we simplify it even further by storing not the probabilities of pairs of colors but the probabilities of colors themselves, we get what we call a \"simple tiled model\". The propagation phase in this model is just adjacency constraint propagation. It's convenient to initialize the simple tiled model with a list of tiles and their adjacency data (adjacency data can be viewed as a large set of very small samples) rather than a sample bitmap.<\/p>\n<p style=\"text-align: center;\">\n  <a href=\"http:\/\/i.imgur.com\/jIctSoT.gif\" rel=\"nofollow\">GIF<\/a> |\n  <a href=\"http:\/\/i.imgur.com\/jIctSoT.gifv\" rel=\"nofollow\">GIFV<\/a>\n<\/p>\n<p>Lists of all the possible pairs of adjacent tiles in practical tilesets can be quite long, so I implemented a symmetry system for tiles to shorten the enumeration. In this system each tile should be assigned with its symmetry type.<\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/ac990b5e822b3b8059ce2d73cd8ecb48fe41af3a\/687474703a2f2f692e696d6775722e636f6d2f39483066726d4b2e706e67\"><img alt=\"symmetries\" src=\"https:\/\/camo.githubusercontent.com\/ac990b5e822b3b8059ce2d73cd8ecb48fe41af3a\/687474703a2f2f692e696d6775722e636f6d2f39483066726d4b2e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/9H0frmK.png\" \/><\/a><\/p>\n<p>Note that the tiles have the same symmetry type as their assigned letters (or, in other words, actions of the\ndihedral group D4 are isomorphic for tiles and their corresponding letters). With this system it's enough to enumerate pairs of adjacent tiles only up to symmetry, which makes lists of adjacencies for tilesets with many symmetrical tiles (even the summer tileset, despite drawings not being symmetrical the system considers such tiles to be symmetrical) several times shorter.<\/p>\n<p style=\"text-align: center;\">\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/557a472a05162f15ae981bccd96489e1424778ea\/687474703a2f2f692e696d6775722e636f6d2f456e426b63564e2e706e67\"><img alt=\"knots\" src=\"https:\/\/camo.githubusercontent.com\/557a472a05162f15ae981bccd96489e1424778ea\/687474703a2f2f692e696d6775722e636f6d2f456e426b63564e2e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/EnBkcVN.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/7663e897070ef8a13781fa52c98968ad821ac059\/687474703a2f2f692e696d6775722e636f6d2f427275784f78392e706e67\"><img alt=\"tiled rooms\" src=\"https:\/\/camo.githubusercontent.com\/7663e897070ef8a13781fa52c98968ad821ac059\/687474703a2f2f692e696d6775722e636f6d2f427275784f78392e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/BruxOx9.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/25034a24c90a7c9b9aa861f96ad9a76a048cc465\/687474703a2f2f692e696d6775722e636f6d2f425974374152362e706e67\"><img alt=\"circuit 1\" src=\"https:\/\/camo.githubusercontent.com\/25034a24c90a7c9b9aa861f96ad9a76a048cc465\/687474703a2f2f692e696d6775722e636f6d2f425974374152362e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/BYt7AR6.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/e990b2539fe6f9b9cb6e45512f37cc61a1cf0448\/687474703a2f2f692e696d6775722e636f6d2f795948624d78382e706e67\"><img alt=\"circuit 2\" src=\"https:\/\/camo.githubusercontent.com\/e990b2539fe6f9b9cb6e45512f37cc61a1cf0448\/687474703a2f2f692e696d6775722e636f6d2f795948624d78382e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/yYHbMx8.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/d3c71b0b087ccf3607bbad6c715108f2dff91523\/687474703a2f2f692e696d6775722e636f6d2f487273304972382e706e67\"><img alt=\"circles\" src=\"https:\/\/camo.githubusercontent.com\/d3c71b0b087ccf3607bbad6c715108f2dff91523\/687474703a2f2f692e696d6775722e636f6d2f487273304972382e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/Hrs0Ir8.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/c958c2371ec28fd3d4e16bbff8a28e6abf828b6c\/687474703a2f2f692e696d6775722e636f6d2f4e64326d514f432e706e67\"><img alt=\"castle\" src=\"https:\/\/camo.githubusercontent.com\/c958c2371ec28fd3d4e16bbff8a28e6abf828b6c\/687474703a2f2f692e696d6775722e636f6d2f4e64326d514f432e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/Nd2mQOC.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/95a9d84e4eea7e4223ae13cc7b8ceb0a5dcf8a06\/687474703a2f2f692e696d6775722e636f6d2f726538574275642e706e67\"><img alt=\"summer 1\" src=\"https:\/\/camo.githubusercontent.com\/95a9d84e4eea7e4223ae13cc7b8ceb0a5dcf8a06\/687474703a2f2f692e696d6775722e636f6d2f726538574275642e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/re8WBud.png\" \/><\/a>\n\t<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/1a39ae57f195bf36446c76b8d26c4057758765b4\/687474703a2f2f692e696d6775722e636f6d2f4f6d55486b31742e706e67\"><img alt=\"summer 2\" src=\"https:\/\/camo.githubusercontent.com\/1a39ae57f195bf36446c76b8d26c4057758765b4\/687474703a2f2f692e696d6775722e636f6d2f4f6d55486b31742e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/OmUHk1t.png\" \/><\/a>\n<\/p>\n<p>Note that the unrestrained knot tileset (with all 5 tiles being allowed) is not interesting for WFC, because you can't run into a situation where you can't place a tile. We call tilesets with this property \"easy\". Without special heuristics easy tilesets don't produce interesting global arrangements, because correlations of tiles in easy tilesets quickly fall off with a distance. Many easy tilesets can be found on <a href=\"http:\/\/cr31.co.uk\/stagecast\/wang\/tiles_e.html\" rel=\"nofollow\">cr31's site<\/a>. Consider the \"Dual\" 2-edge tileset there. How can it generate knots (without t-junctions, not easy) while being easy? The answer is, it can only generate a narrow class of knots, it can't produce an arbitrary knot.<\/p>\n<p>Note also that Circuit, Summer and Rooms tilesets are non-Wang. That is, their adjacency data cannot be induced from edge coloring. For example, in Circuit two Corners cannot be adjacent, yet they can be connected with a Connection tile, and diagonal tracks cannot change direction.<\/p>\n<h2><a id=\"user-content-higher-dimensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#higher-dimensions\"><\/a>Higher dimensions<\/h2>\n<p>WFC algorithm in higher dimensions works completely the same way as in dimension 2, though performance becomes an issue. These voxel models were generated with N=2 overlapping tiled model using 5x5x5 and 5x5x2 blocks and additional heuristics (height, density, curvature, ...).<\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/9065caed08b7793b7775cd53976ded88789de4dd\/687474703a2f2f692e696d6775722e636f6d2f6873715064516c2e706e67\"><img alt=\"voxels\" src=\"https:\/\/camo.githubusercontent.com\/9065caed08b7793b7775cd53976ded88789de4dd\/687474703a2f2f692e696d6775722e636f6d2f6873715064516c2e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/hsqPdQl.png\" \/><\/a><\/p>\n<p>Higher resolution screenshots: <a href=\"http:\/\/i.imgur.com\/0bsjlBY.png\" rel=\"nofollow\">1<\/a>, <a href=\"http:\/\/i.imgur.com\/GduN0Vr.png\" rel=\"nofollow\">2<\/a>, <a href=\"http:\/\/i.imgur.com\/IEOsbIy.png\" rel=\"nofollow\">3<\/a>.<\/p>\n<p>Voxel models generated with WFC and other algorithms will be in a separate repo.<\/p>\n<h2><a id=\"user-content-constrained-synthesis\" class=\"anchor\" aria-hidden=\"true\" href=\"#constrained-synthesis\"><\/a>Constrained synthesis<\/h2>\n<p>WFC algorithm supports constraints. Therefore, it can be easely combined with other generative algorithms or with manual creation.<\/p>\n<p>Here is WFC autocompleting a level started by a human:<\/p>\n<p style=\"text-align: center;\">\n  <a href=\"http:\/\/i.imgur.com\/X3aNDUv.gif\" rel=\"nofollow\">GIF<\/a> |\n  <a href=\"http:\/\/i.imgur.com\/X3aNDUv.gifv\" rel=\"nofollow\">GIFV<\/a>\n<\/p>\n<p><a href=\"https:\/\/github.com\/mxgmn\/ConvChain\">ConvChain<\/a> algorithm satisfies the strong version of the condition (C2): the limit distribution of NxN patterns in the outputs it is producing is exactly the same as the distributions of patterns in the input. However, ConvChain doesn't satisfy (C1): it often produces noticeable artefacts. It makes sense to run ConvChain first to get a well-sampled configuration and then run WFC to correct local artefacts. This is similar to a common strategy in optimization: first run a Monte-Carlo method to find a point close to a global optimum and then run a gradient descent from that point for greater accuracy.<\/p>\n<p>P. F. Harrison's <a href=\"https:\/\/github.com\/mxgmn\/SynTex\">texture synthesis<\/a> algorithm is significantly faster than WFC, but it has trouble with long correlations (for example, it's difficult for this algorithm to synthesize brick wall textures with correctly aligned bricks). But this is exactly where WFC shines, and Harrison's algorithm supports constraints. It makes sense first to generate a perfect brick wall blueprint with WFC and then run a constrained texture synthesis algorithm on that blueprint.<\/p>\n<h2><a id=\"user-content-comments\" class=\"anchor\" aria-hidden=\"true\" href=\"#comments\"><\/a>Comments<\/h2>\n<p>Why the minimal entropy heuristic? I noticed that when humans draw something they often follow the minimal entropy heuristic themselves. That's why the algorithm is so enjoyable to watch.<\/p>\n<p>The overlapping model relates to the simple tiled model the same way higher order Markov chains relate to order one Markov chains.<\/p>\n<p>Note that the entropy of any node can't increase during the propagation phase, i.e. possibilities are not arising, but can be canceled. When propagation step can not decrease entropy further, we activate observation step. If the observation step can not decrease entropy, that means that the algorithm has finished working.<\/p>\n<p>WFC's propagation phase is very similar to the loopy belief propagation algorithm. In fact, I first programmed belief propagation, but then switched to constraint propagation with a saved stationary distribution, because BP is significantly slower without a massive parallelization (on a CPU) and didn't produce significantly better results in my problems.<\/p>\n<p>Note that the \"Simple Knot\" and \"Trick Knot\" samples have 3 colors, not 2.<\/p>\n<p>One of the dimensions can be time. In particular, d-dimensional WFC captures the behaviour of any (d-1)-dimensional cellular automata.<\/p>\n<h2><a id=\"user-content-references\" class=\"anchor\" aria-hidden=\"true\" href=\"#references\"><\/a>References<\/h2>\n<p>This project builds upon Paul Merrell's work on model synthesis, in particular discrete model synthesis chapter of <a href=\"http:\/\/graphics.stanford.edu\/~pmerrell\/thesis.pdf\" rel=\"nofollow\">his dissertation<\/a>. Paul propagates adjacency constraints in what we call a simple tiled model with a heuristic that tries to complete propagation in a small moving region.<\/p>\n<p>It was also heavily influenced by declarative texture synthesis chapter of <a href=\"http:\/\/logarithmic.net\/pfh-files\/thesis\/dissertation.pdf\" rel=\"nofollow\">Paul F. Harrison's dissertation<\/a>. Paul defines adjacency data of tiles by labeling their borders and uses backtracking search to fill the tilemap.<\/p>\n<h2><a id=\"user-content-notable-ports-forks-and-spinoffs\" class=\"anchor\" aria-hidden=\"true\" href=\"#notable-ports-forks-and-spinoffs\"><\/a>Notable ports, forks and spinoffs<\/h2>\n<ul><li>Emil Ernerfeldt made a <a href=\"https:\/\/github.com\/emilk\/wfc\">C++ port<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/nanodeath\">Max Aller<\/a> made a Kotlin (JVM) library, <a href=\"https:\/\/gitlab.com\/nanodeath\/kollapse\" rel=\"nofollow\">Kollapse<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/kchapelier\">Kevin Chapelier<\/a> made a <a href=\"http:\/\/www.kchapelier.com\/wfc-example\/overlapping-model.html\" rel=\"nofollow\">JavaScript port<\/a>.<\/li>\n<li>Oskar Stalberg programmed a 3d tiled model, a 2d tiled model for irregular grids on a sphere and is building beautiful 3d tilesets for them: <a href=\"https:\/\/twitter.com\/OskSta\/status\/787319655648100352\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/784847588893814785\" rel=\"nofollow\">2<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/784847933686575104\" rel=\"nofollow\">3<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/784848286272327680\" rel=\"nofollow\">4<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/793545297376972801\" rel=\"nofollow\">5<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/793806535898136576\" rel=\"nofollow\">6<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/802496920790777856\" rel=\"nofollow\">7<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/804291629561577472\" rel=\"nofollow\">8<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/806856212260278272\" rel=\"nofollow\">9<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/806904557502464000\" rel=\"nofollow\">10<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/818857408848130048\" rel=\"nofollow\">11<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/832633189277409280\" rel=\"nofollow\">12<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/851170356530475008\" rel=\"nofollow\">13<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/858301207936458752\" rel=\"nofollow\">14<\/a>, <a href=\"https:\/\/twitter.com\/OskSta\/status\/863019585162932224\" rel=\"nofollow\">15<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/selfsame\">Joseph Parker<\/a> adapted <a href=\"https:\/\/selfsame.itch.io\/unitywfc\" rel=\"nofollow\">WFC to Unity<\/a> and used it generate skateparks in the <a href=\"https:\/\/arcadia-clojure.itch.io\/proc-skater-2016\" rel=\"nofollow\">Proc Skater 2016<\/a> game, <a href=\"https:\/\/twitter.com\/jplur_\/status\/929482200034226176\" rel=\"nofollow\">fantastic plateaus<\/a> in the 2017 game <a href=\"https:\/\/arcadia-clojure.itch.io\/swapland\" rel=\"nofollow\">Swapland<\/a> and platform levels in the 2018 game <a href=\"https:\/\/selfsame.itch.io\/bug-with-a-gun\" rel=\"nofollow\">Bug with a Gun<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/mewo2\">Martin O'Leary<\/a> applied a <a href=\"https:\/\/github.com\/mewo2\/oisin\">WFC-like algorithm<\/a> to poetry generation: <a href=\"https:\/\/twitter.com\/mewo2\/status\/789167437518217216\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/twitter.com\/mewo2\/status\/789177702620114945\" rel=\"nofollow\">2<\/a>, <a href=\"https:\/\/twitter.com\/mewo2\/status\/789187174683987968\" rel=\"nofollow\">3<\/a>, <a href=\"https:\/\/twitter.com\/mewo2\/status\/789897712372183041\" rel=\"nofollow\">4<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/NNNenov\">Nick Nenov<\/a> made a <a href=\"https:\/\/twitter.com\/NNNenov\/status\/789903180226301953\" rel=\"nofollow\">3d voxel tileset<\/a> based on my Castle tileset. Nick uses text output option in the tiled model to reconstruct 3d models in Cinema 4D.<\/li>\n<li>Sean Leffler implemented the <a href=\"https:\/\/github.com\/sdleffler\/collapse\">overlapping model in Rust<\/a>.<\/li>\n<li>rid5x is making an <a href=\"https:\/\/twitter.com\/rid5x\/status\/782442620459114496\" rel=\"nofollow\">OCaml version of WFC<\/a>.<\/li>\n<li>I published a very basic <a href=\"https:\/\/bitbucket.org\/mxgmn\/basic3dwfc\/overview\" rel=\"nofollow\">3d tiled model<\/a> so people could make their own 3d tilesets without waiting for the full 3d repository.<\/li>\n<li>I made an <a href=\"https:\/\/twitter.com\/ExUtumno\/status\/798571284342837249\" rel=\"nofollow\">interactive version<\/a> of the overlapping model, you can download the GUI executable from the <a href=\"https:\/\/exutumno.itch.io\/wavefunctioncollapse\" rel=\"nofollow\">WFC itch.io page<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/unormal\">Brian Bucklew<\/a> built a level generation pipeline that applies WFC in multiple passes for the <a href=\"http:\/\/store.steampowered.com\/app\/333640\" rel=\"nofollow\">Caves of Qud<\/a> game: <a href=\"https:\/\/twitter.com\/unormal\/status\/805987523596091392\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808566029387448320\" rel=\"nofollow\">2<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808523056259993601\" rel=\"nofollow\">3<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808523493994364928\" rel=\"nofollow\">4<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808519575264497666\" rel=\"nofollow\">5<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808519216185876480\" rel=\"nofollow\">6<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808795396508123136\" rel=\"nofollow\">7<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/808860105093632001\" rel=\"nofollow\">8<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/809637856432033792\" rel=\"nofollow\">9<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/810239794433425408\" rel=\"nofollow\">10<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/811034574973243393\" rel=\"nofollow\">11<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/811720423419314176\" rel=\"nofollow\">12<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/811034037259276290\" rel=\"nofollow\">13<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/810971337309224960\" rel=\"nofollow\">14<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/811405368777723909\" rel=\"nofollow\">15<\/a>, <a href=\"https:\/\/twitter.com\/ptychomancer\/status\/812053801544757248\" rel=\"nofollow\">16<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/812159308263788544\" rel=\"nofollow\">17<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/812158749838340096\" rel=\"nofollow\">18<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/814569437181476864\" rel=\"nofollow\">19<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/814570383189876738\" rel=\"nofollow\">20<\/a>, <a href=\"https:\/\/twitter.com\/unormal\/status\/819725864623603712\" rel=\"nofollow\">21<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/dannywynne\">Danny Wynne<\/a> implemented a <a href=\"https:\/\/twitter.com\/dwtw\/status\/810166761270243328\" rel=\"nofollow\">3d tiled model<\/a>.<\/li>\n<li>Arvi Teikari programmed a <a href=\"http:\/\/www.hempuli.com\/blogblog\/archives\/1598\" rel=\"nofollow\">texture synthesis algorithm with the entropy heuristic<\/a> in Lua. Headchant <a href=\"https:\/\/github.com\/headchant\/iga\">ported<\/a> it to work with L\u00d6VE.<\/li>\n<li>Isaac Karth made a <a href=\"https:\/\/github.com\/ikarth\/wfc_python\">Python port<\/a> of the overlapping model.<\/li>\n<li>Oskar Stalberg made an <a href=\"http:\/\/oskarstalberg.com\/game\/wave\/wave.html\" rel=\"nofollow\">interactive version<\/a> of the tiled model that runs in the browser.<\/li>\n<li><a href=\"https:\/\/github.com\/MattRix\">Matt Rix<\/a> implemented a 3d tiled model (<a href=\"https:\/\/twitter.com\/MattRix\/status\/869403586664570880\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/870999185167962113\" rel=\"nofollow\">2<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/871054734018453505\" rel=\"nofollow\">3<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/871056805761359872\" rel=\"nofollow\">4<\/a>) and made a 3-dimensional tiled model where one of the dimensions is time (<a href=\"https:\/\/twitter.com\/MattRix\/status\/872674537799913472\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/872648369625325568\" rel=\"nofollow\">2<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/872645716660891648\" rel=\"nofollow\">3<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/872641331956518914\" rel=\"nofollow\">4<\/a>, <a href=\"https:\/\/twitter.com\/MattRix\/status\/979020989181890560\" rel=\"nofollow\">5<\/a>).<\/li>\n<li><a href=\"https:\/\/github.com\/NNNenov\">Nick Nenov<\/a> made a <a href=\"https:\/\/www.dropbox.com\/s\/zeiat1w8zre9ro8\/Knots%20breakdown.png?dl=0\" rel=\"nofollow\">visual guide<\/a> to the tile symmetry system.<\/li>\n<li><a href=\"https:\/\/github.com\/ikarth\">Isaac Karth<\/a> and <a href=\"https:\/\/github.com\/rndmcnlly\">Adam M. Smith<\/a> wrote a <a href=\"https:\/\/adamsmith.as\/papers\/wfc_is_constraint_solving_in_the_wild.pdf\" rel=\"nofollow\">research paper<\/a> where they formulate WFC as an ASP problem, use general constraint solver <a href=\"https:\/\/github.com\/potassco\/clingo\">clingo<\/a> to generate bitmaps, experiment with global constraints, trace WFC's history and give detailed explanation of the algorithm.<\/li>\n<li>Sylvain Lefebvre made a <a href=\"https:\/\/github.com\/sylefeb\/VoxModSynth\">C++ implementation<\/a> of 3d model synthesis, described the thought process of designing a sample and provided an example where adjacency constraints ensure that the output is connected (walkable).<\/li>\n<li>I generalized 3d WFC to work with cube symmetry group and made a tileset that generates <a href=\"https:\/\/twitter.com\/ExUtumno\/status\/895684431477747715\" rel=\"nofollow\">Escheresque scenes<\/a>.<\/li>\n<li>There are many ways to visualize partially observed wave states. In the code, color values of possible options are averaged to produce the resulting color. Oskar Stalberg <a href=\"https:\/\/twitter.com\/OskSta\/status\/863019585162932224\" rel=\"nofollow\">shows<\/a> partially observed states as semi-transparent boxes, where the box is bigger for a state with more options. In the voxel setting I <a href=\"https:\/\/twitter.com\/ExUtumno\/status\/900395635412787202\" rel=\"nofollow\">visualize<\/a> wave states with per-voxel voting.<\/li>\n<li>Remy Devaux implemented the tiled model in PICO-8 and wrote an <a href=\"https:\/\/trasevol.dog\/2017\/09\/01\/di19\/\" rel=\"nofollow\">article<\/a> about generation of coherent data with the explanation of WFC.<\/li>\n<li>For the upcoming game <a href=\"https:\/\/www.badnorth.com\/\" rel=\"nofollow\">Bad North<\/a> Oskar Stalberg <a href=\"https:\/\/twitter.com\/OskSta\/status\/917405214638006273\" rel=\"nofollow\">uses<\/a> a heuristic that tries to select such tiles\nthat the resulting observed zone is navigable at each step.<\/li>\n<li>William Manning <a href=\"https:\/\/github.com\/heyx3\/easywfc\">implemented<\/a> the overlapping model in C# with the primary goal of making code readable, and provided it with WPF GUI.<\/li>\n<li><a href=\"https:\/\/gist.github.com\/selfsame\">Joseph Parker<\/a> wrote a WFC <a href=\"http:\/\/www.procjam.com\/tutorials\/wfc\/\" rel=\"nofollow\">tutorial<\/a> for Procjam 2017.<\/li>\n<li><a href=\"https:\/\/github.com\/aman-tiwari\">Aman Tiwari<\/a> formulated the connectivity constraint as an <a href=\"https:\/\/gist.github.com\/aman-tiwari\/8a7b874cb1fd1270adc203b2af293f4c\">ASP problem<\/a> for clingo.<\/li>\n<li>MatveyK programmed a <a href=\"https:\/\/github.com\/MatveyK\/Kazimir\">3d overlapping model<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/sylefeb\">Sylvain Lefebvre<\/a>, <a href=\"https:\/\/github.com\/1iyiwei\">Li-Yi Wei<\/a> and <a href=\"https:\/\/github.com\/connellybarnes\">Connelly Barnes<\/a> are <a href=\"https:\/\/hal.archives-ouvertes.fr\/hal-01706539\/\" rel=\"nofollow\">investigating<\/a> the possibility of hiding information inside textures. They made a <a href=\"https:\/\/members.loria.fr\/Sylvain.Lefebvre\/infotexsyn\/\" rel=\"nofollow\">tool<\/a> that can encode text messages as WFC tilings and decode them back. This technique allows to use WFC tilings as QR codes.<\/li>\n<li><a href=\"https:\/\/github.com\/math-fehr\">Mathieu Fehr<\/a> and <a href=\"https:\/\/github.com\/Ekdohibs\">Nathanael Courant<\/a> significantly <a href=\"https:\/\/github.com\/math-fehr\/fast-wfc\">improved<\/a> the running time of WFC, by an order of magnitude for the overlapping model. I <a href=\"https:\/\/github.com\/mxgmn\/WaveFunctionCollapse\/commit\/fad1066b5000f7e9fbda0ef81bbea56799686670\">integrated<\/a> their improvements into the code.<\/li>\n<li>Vasu Mahesh <a href=\"https:\/\/github.com\/vasumahesh1\/WFC_WebGL\">ported<\/a> 3d tiled model to TypeScript, made a new tileset and <a href=\"https:\/\/vasumahesh1.github.io\/WFC_WebGL\" rel=\"nofollow\">visualised<\/a> the generation process in WebGL.<\/li>\n<li><a href=\"https:\/\/github.com\/greentec\">Kim Hwanhee<\/a> experimented with 3d WFC and created\/adapted many voxel tilesets: <a href=\"https:\/\/twitter.com\/greentecq\/status\/1025348928634408960\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1004068394553913344\" rel=\"nofollow\">2<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1005835830802305024\" rel=\"nofollow\">3<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1022851327041265664\" rel=\"nofollow\">4<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1011351814216736769\" rel=\"nofollow\">5<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1008210550944387077\" rel=\"nofollow\">6<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1006390606875070464\" rel=\"nofollow\">7<\/a>, <a href=\"https:\/\/twitter.com\/greentecq\/status\/1015182718810841088\" rel=\"nofollow\">8<\/a>.<\/li>\n<li>Oskar Stalberg gave a <a href=\"https:\/\/www.youtube.com\/watch?v=0bcZb-SsnrA\" rel=\"nofollow\">talk<\/a> about level generation in Bad North at the Everything Procedural Conference 2018.<\/li>\n<li>I <a href=\"https:\/\/twitter.com\/ExUtumno\/status\/1024314661951467521\" rel=\"nofollow\">wrote<\/a> about how to generate (approximately) unbiased paths between 2 points with WFC and other algorithms.<\/li>\n<li><a href=\"https:\/\/github.com\/ikarth\">Isaac Karth<\/a> and <a href=\"https:\/\/github.com\/rndmcnlly\">Adam M. Smith<\/a> published a <a href=\"https:\/\/arxiv.org\/abs\/1809.04432\" rel=\"nofollow\">preprint<\/a> where they describe a system based on WFC that learns from both positive and negative examples, and discuss it in a general context of dialogs with example-driven generators.<\/li>\n<li>Brendan Anthony <a href=\"https:\/\/steamcommunity.com\/games\/314230\/announcements\/detail\/3369147113795750369\" rel=\"nofollow\">uses<\/a> WFC to generate wall decorations in the game <a href=\"https:\/\/store.steampowered.com\/app\/314230\/Rodina\/\" rel=\"nofollow\">Rodina<\/a>.<\/li>\n<li>Tim Kong implemented the <a href=\"https:\/\/github.com\/Mitim-84\/WFC-Gen\">overlapping model in Haxe<\/a>.<\/li>\n<li>In order to generate connected structures, Boris the Brave applied the <a href=\"https:\/\/www.boristhebrave.com\/2018\/04\/28\/random-paths-via-chiseling\" rel=\"nofollow\">chiseling method<\/a> to WFC. He published a <a href=\"https:\/\/boristhebrave.github.io\/DeBroglie\" rel=\"nofollow\">library<\/a> that supports hex grids, additional constraints and backtracking.<\/li>\n<li><a href=\"https:\/\/github.com\/marian42\">Marian Kleineberg<\/a> created a <a href=\"https:\/\/marian42.itch.io\/wfc\" rel=\"nofollow\">city generator<\/a> based on the tiled model for Procjam 2018.<\/li>\n<li>Sol Bekic <a href=\"https:\/\/github.com\/s-ol\/gpWFC\">programmed<\/a> the tiled model that runs on GPU using PyOpenCL. Instead of keeping a queue of nodes to propagate from, it propagates from every node on the grid in parallel.<\/li>\n<li>Wouter van Oortmerssen <a href=\"https:\/\/github.com\/aardappel\/lobster\/commit\/703f67472bfd80c26bb626e1d5c22ec91047da98\">implemented<\/a> the tiled model in a single C++ function, with a structure similar to a priority queue for faster observation.<\/li>\n<\/ul><h2><a id=\"user-content-how-to-build\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-build\"><\/a>How to build<\/h2>\n<p>WFC is a console application that depends only on the standard library. Build instructions from the community for various platforms can be found in the <a href=\"https:\/\/github.com\/mxgmn\/WaveFunctionCollapse\/issues\/3\">relevant issue<\/a>. Casey Marshall made a <a href=\"https:\/\/github.com\/mxgmn\/WaveFunctionCollapse\/pull\/18\">pull request<\/a> that makes using the program with the command line more convenient and includes snap packaging.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h2>\n<p>Some samples are taken from the games Ultima IV and <a href=\"https:\/\/github.com\/crawl\/crawl\">Dungeon Crawl<\/a>. Circles tileset is taken from <a href=\"https:\/\/twitter.com\/quasimondo\/status\/778196128957403136\" rel=\"nofollow\">Mario Klingemann<\/a>. Idea of generating integrated circuits was suggested to me by <a href=\"https:\/\/twitter.com\/Moonasaur\/status\/759890746350731264\" rel=\"nofollow\">Moonasaur<\/a> and their style was taken from Zachtronics' <a href=\"http:\/\/www.zachtronics.com\/ruckingenur-ii\/\" rel=\"nofollow\">Ruckingenur II<\/a>. Cat overlapping sample is taken from the Nyan Cat video, Qud sample was made by <a href=\"https:\/\/github.com\/unormal\">Brian Bucklew<\/a>, Magic Office + Spirals samples - by rid5x, Colored City + Link + Link 2 + Mazelike + Red Dot + Smile City overlapping samples - by Arvi Teikari. Summer tileset was made by Hermann Hillmann. Voxel models were rendered in <a href=\"http:\/\/ephtracy.github.io\/\" rel=\"nofollow\">MagicaVoxel<\/a>.<\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/eab73e9939d22708d2d539832034e4a8d37aab05\/687474703a2f2f692e696d6775722e636f6d2f435a73766e63372e706e67\"><img alt=\"second collage\" src=\"https:\/\/camo.githubusercontent.com\/eab73e9939d22708d2d539832034e4a8d37aab05\/687474703a2f2f692e696d6775722e636f6d2f435a73766e63372e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/CZsvnc7.png\" \/><\/a><\/p>\n<p style=\"text-align: center;\"><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/e4033a112e1748bbe99857736fb0ac6714b6780c\/687474703a2f2f692e696d6775722e636f6d2f5279775843486e2e706e67\"><img alt=\"voxel perspective\" src=\"https:\/\/camo.githubusercontent.com\/e4033a112e1748bbe99857736fb0ac6714b6780c\/687474703a2f2f692e696d6775722e636f6d2f5279775843486e2e706e67\" data-canonical-src=\"http:\/\/i.imgur.com\/RywXCHn.png\" \/><\/a><\/p>","created_at":"2018-11-07T16:24:12+0000","updated_at":"2018-11-07T16:24:12+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/14931730?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":196,"title":"Ask HN: Are there any reasonable alternatives to MacBook Pro for developer?","url":"https:\/\/news.ycombinator.com\/item?id=16785507","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16785507\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16785507\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=16785507\" class=\"storylink\">Ask HN: Are there any reasonable alternatives to MacBook Pro for developer?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        472 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=robsun\" class=\"hnuser\">robsun<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785507\">7 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16785507&amp;goto=item%3Fid%3D16785507\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Are%20there%20any%20reasonable%20alternatives%20to%20MacBook%20Pro%20for%20developer%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Are%20there%20any%20reasonable%20alternatives%20to%20MacBook%20Pro%20for%20developer%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16785507&amp;auth=fc710586ddc5537741acc2ffd27ed13e19d2da39\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16785507\">496\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Hi,<p>I'm writing this on my late 2012 MacBook Pro. Time goes by and I know rather sooner than later I'll need to replace it with a new machine. In 2012 I paid around 1000$ for MacBook Pro + Samsung SSD (256GB) + 16GB RAM, I made modifications on my own.<\/p><p>I check notebookcheck from time to time. I read reviews, opinions about new laptops. The point is, I don't know if there is any machine that could be recommended in reasonable price. At work I'm using some new MacBook Pro which (i5\/16GB\/128GB SSD) which is noticeably slower than my current machine.<\/p><p>Performance of the computer is quite important for me. I'm an Android developer, compilation of a big project I'm working on takes enormous amount of RAM and CPU nowadays (with new Android Studio it's even worse). From time to time I work on web projects, so handling several instances of docker shouldn't be a problem for a new machine. I prefer Linux over MacOs over Windows, so good support for Ubuntu\/Fedora would be nice.<\/p><p>I checked some computers in details but most of them fail in one or more aspects:\n- hinges - MacBook has superior hinges, if I pay more than 1000 - 1500$ I expect to have great hinges\n- price\n- performance\n- Linux support<\/p><p>Price is quite important for me, I'm from Eastern Europe. What computer would you recommend in, let's say, &lt;2000$ ?<\/p><\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"16787138\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkop\" class=\"hnuser\">gkop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787138\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788170\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=closeparen\" class=\"hnuser\">closeparen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788170\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788568\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=verbify\" class=\"hnuser\">verbify<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788568\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788719\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=closeparen\" class=\"hnuser\">closeparen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788719\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789263\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kenning\" class=\"hnuser\">kenning<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789263\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791087\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tischler\" class=\"hnuser\">tischler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791087\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16793832\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=closeparen\" class=\"hnuser\">closeparen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16793832\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788985\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788985\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789016\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amerine\" class=\"hnuser\">amerine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789016\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789331\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fjsolwmv\" class=\"hnuser\">fjsolwmv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789331\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789053\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789053\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789026\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brian-armstrong\" class=\"hnuser\">brian-armstrong<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789026\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788438\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloeki\" class=\"hnuser\">lloeki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788438\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16793214\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thetwan\" class=\"hnuser\">thetwan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16793214\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788314\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwafk81\" class=\"hnuser\">throwafk81<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788314\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788852\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ClassyJacket\" class=\"hnuser\">ClassyJacket<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788852\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789211\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sevensor\" class=\"hnuser\">sevensor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789211\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788367\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwatte\" class=\"hnuser\">jwatte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788367\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787308\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aviraldg\" class=\"hnuser\">aviraldg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787308\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787633\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787633\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787990\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eof\" class=\"hnuser\">eof<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787990\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16797577\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gurkendoktor\" class=\"hnuser\">gurkendoktor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16797577\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787488\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhoechtl\" class=\"hnuser\">jhoechtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787488\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789346\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rossy\" class=\"hnuser\">rossy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789346\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788063\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saltcured\" class=\"hnuser\">saltcured<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788063\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788461\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloeki\" class=\"hnuser\">lloeki<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788461\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788587\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saltcured\" class=\"hnuser\">saltcured<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788587\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788623\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhoechtl\" class=\"hnuser\">jhoechtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788623\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789121\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saltcured\" class=\"hnuser\">saltcured<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789121\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788340\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kuschku\" class=\"hnuser\">kuschku<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788340\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788594\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhoechtl\" class=\"hnuser\">jhoechtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788594\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788694\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kuschku\" class=\"hnuser\">kuschku<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788694\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788739\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhoechtl\" class=\"hnuser\">jhoechtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788739\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16797652\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gurkendoktor\" class=\"hnuser\">gurkendoktor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16797652\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787986\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shmerl\" class=\"hnuser\">shmerl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787986\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787674\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=denimnerd\" class=\"hnuser\">denimnerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787674\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788501\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merb\" class=\"hnuser\">merb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788501\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789079\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nitrogen\" class=\"hnuser\">nitrogen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789079\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791458\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merb\" class=\"hnuser\">merb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791458\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16799243\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nitrogen\" class=\"hnuser\">nitrogen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16799243\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787414\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pedrocr\" class=\"hnuser\">pedrocr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787414\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787421\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkop\" class=\"hnuser\">gkop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787421\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787886\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=indemnity\" class=\"hnuser\">indemnity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787886\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790048\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maxnoe\" class=\"hnuser\">maxnoe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790048\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16866342\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=indemnity\" class=\"hnuser\">indemnity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16866342\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788257\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MightySCollins\" class=\"hnuser\">MightySCollins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788257\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791902\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SZJX\" class=\"hnuser\">SZJX<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791902\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787353\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Symbiote\" class=\"hnuser\">Symbiote<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787353\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787663\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=digi_owl\" class=\"hnuser\">digi_owl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787663\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787379\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mixmastamyk\" class=\"hnuser\">mixmastamyk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787379\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787336\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bonoetmalo\" class=\"hnuser\">bonoetmalo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787336\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789291\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sridca\" class=\"hnuser\">sridca<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789291\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787826\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thomastjeffery\" class=\"hnuser\">thomastjeffery<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787826\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787467\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paride5745\" class=\"hnuser\">paride5745<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787467\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789319\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ledgerdev\" class=\"hnuser\">ledgerdev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789319\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787579\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smoyer\" class=\"hnuser\">smoyer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787579\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788919\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zepearl\" class=\"hnuser\">zepearl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788919\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787980\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brockers\" class=\"hnuser\">brockers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787980\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788599\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rkangel\" class=\"hnuser\">rkangel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788599\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789302\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eat_veggies\" class=\"hnuser\">eat_veggies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789302\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16806934\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=therealdrag0\" class=\"hnuser\">therealdrag0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16806934\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789748\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=organsnyder\" class=\"hnuser\">organsnyder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789748\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791837\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Light2Yellow\" class=\"hnuser\">Light2Yellow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791837\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16796907\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gkop\" class=\"hnuser\">gkop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16796907\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16798486\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=locusm\" class=\"hnuser\">locusm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16798486\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16798271\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aagha\" class=\"hnuser\">aagha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16798271\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787622\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=linkmotif\" class=\"hnuser\">linkmotif<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787622\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787989\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LordDragonfang\" class=\"hnuser\">LordDragonfang<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787989\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788040\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=forapurpose\" class=\"hnuser\">forapurpose<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788040\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791930\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SZJX\" class=\"hnuser\">SZJX<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791930\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787901\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=confounded\" class=\"hnuser\">confounded<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787901\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788053\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=forapurpose\" class=\"hnuser\">forapurpose<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788053\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788231\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=EpicEng\" class=\"hnuser\">EpicEng<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788231\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789908\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=forapurpose\" class=\"hnuser\">forapurpose<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789908\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787695\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SkyMarshal\" class=\"hnuser\">SkyMarshal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787695\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791758\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=O_H_E\" class=\"hnuser\">O_H_E<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791758\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785700\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noir_lord\" class=\"hnuser\">noir_lord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785700\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786943\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelangerman\" class=\"hnuser\">michaelangerman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786943\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787120\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chillydawg\" class=\"hnuser\">chillydawg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787120\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787524\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_sven_\" class=\"hnuser\">_sven_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787524\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787225\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noir_lord\" class=\"hnuser\">noir_lord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787225\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789049\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=soperj\" class=\"hnuser\">soperj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789049\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787068\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrischen\" class=\"hnuser\">chrischen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787068\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16837676\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=coatmatter\" class=\"hnuser\">coatmatter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16837676\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787500\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noir_lord\" class=\"hnuser\">noir_lord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787500\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16793104\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gargravarr\" class=\"hnuser\">gargravarr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16793104\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16797980\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sxldier\" class=\"hnuser\">sxldier<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16797980\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16898454\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=john_knight\" class=\"hnuser\">john_knight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16898454\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786505\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anotheryou\" class=\"hnuser\">anotheryou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786505\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786927\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=birksherty\" class=\"hnuser\">birksherty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786927\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787279\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anotheryou\" class=\"hnuser\">anotheryou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787279\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786551\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=contingencies\" class=\"hnuser\">contingencies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786551\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786607\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anotheryou\" class=\"hnuser\">anotheryou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786607\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790777\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flamtap\" class=\"hnuser\">flamtap<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790777\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791996\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SZJX\" class=\"hnuser\">SZJX<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791996\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16795456\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Qub3d\" class=\"hnuser\">Qub3d<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16795456\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786924\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=leetcrew\" class=\"hnuser\">leetcrew<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786924\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787286\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anotheryou\" class=\"hnuser\">anotheryou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787286\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786856\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DoofusOfDeath\" class=\"hnuser\">DoofusOfDeath<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786856\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787274\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anotheryou\" class=\"hnuser\">anotheryou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787274\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789878\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DoofusOfDeath\" class=\"hnuser\">DoofusOfDeath<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789878\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786800\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jnwatson\" class=\"hnuser\">jnwatson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786800\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786949\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sametmax\" class=\"hnuser\">sametmax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786949\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787064\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stefantheard\" class=\"hnuser\">stefantheard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787064\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790805\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flamtap\" class=\"hnuser\">flamtap<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790805\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787106\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickthemagicman\" class=\"hnuser\">nickthemagicman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787106\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787149\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thriptic\" class=\"hnuser\">Thriptic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787149\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787163\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stefantheard\" class=\"hnuser\">stefantheard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787163\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787958\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=musicaldope\" class=\"hnuser\">musicaldope<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787958\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16792860\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pbrumm\" class=\"hnuser\">pbrumm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16792860\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789744\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=caleblloyd\" class=\"hnuser\">caleblloyd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789744\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790939\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sametmax\" class=\"hnuser\">sametmax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790939\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16825833\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=caleblloyd\" class=\"hnuser\">caleblloyd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16825833\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788707\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j88439h84\" class=\"hnuser\">j88439h84<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788707\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788801\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tamrix\" class=\"hnuser\">tamrix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788801\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787265\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Boulth\" class=\"hnuser\">Boulth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787265\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787503\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Cakez0r\" class=\"hnuser\">Cakez0r<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787503\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789397\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lasagnaphil\" class=\"hnuser\">lasagnaphil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789397\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16812081\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gormz\" class=\"hnuser\">gormz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16812081\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787589\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BlackjackCF\" class=\"hnuser\">BlackjackCF<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787589\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788444\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwatte\" class=\"hnuser\">jwatte<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788444\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788766\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jjoonathan\" class=\"hnuser\">jjoonathan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788766\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786862\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DoofusOfDeath\" class=\"hnuser\">DoofusOfDeath<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786862\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787587\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=raquo\" class=\"hnuser\">raquo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787587\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789159\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StillBored\" class=\"hnuser\">StillBored<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789159\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16802611\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=theyinwhy\" class=\"hnuser\">theyinwhy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16802611\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788537\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gumballhead\" class=\"hnuser\">gumballhead<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788537\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788586\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dijit\" class=\"hnuser\">dijit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788586\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786469\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786469\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786812\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sxg\" class=\"hnuser\">sxg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786812\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787530\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787530\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789648\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gst\" class=\"hnuser\">gst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789648\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790996\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=willtim\" class=\"hnuser\">willtim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790996\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788066\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Rotareti\" class=\"hnuser\">Rotareti<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788066\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788032\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mst\" class=\"hnuser\">mst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788032\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786873\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ogig\" class=\"hnuser\">ogig<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786873\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787134\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=roryisok\" class=\"hnuser\">roryisok<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787134\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786592\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blinkingled\" class=\"hnuser\">blinkingled<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786592\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786653\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786653\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787107\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericd\" class=\"hnuser\">ericd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787107\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788140\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blinkingled\" class=\"hnuser\">blinkingled<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788140\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16795227\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericd\" class=\"hnuser\">ericd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16795227\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791139\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cmcginty\" class=\"hnuser\">cmcginty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791139\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787257\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=foepys\" class=\"hnuser\">foepys<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787257\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787544\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787544\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786602\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dannyw\" class=\"hnuser\">dannyw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786602\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786641\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786641\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786850\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelangerman\" class=\"hnuser\">michaelangerman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786850\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787549\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shazow\" class=\"hnuser\">shazow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787549\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786845\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelangerman\" class=\"hnuser\">michaelangerman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786845\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786870\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=flyingfences\" class=\"hnuser\">flyingfences<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786870\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786934\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=michaelangerman\" class=\"hnuser\">michaelangerman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786934\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787081\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eat_veggies\" class=\"hnuser\">eat_veggies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787081\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787407\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=feriancek\" class=\"hnuser\">feriancek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787407\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787874\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thomastjeffery\" class=\"hnuser\">thomastjeffery<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787874\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787245\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thinkmassive\" class=\"hnuser\">thinkmassive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787245\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788718\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saidmasoud\" class=\"hnuser\">saidmasoud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788718\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786859\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786859\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787164\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=peatmoss\" class=\"hnuser\">peatmoss<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787164\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791980\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SZJX\" class=\"hnuser\">SZJX<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791980\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786457\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786457\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789324\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tomxor\" class=\"hnuser\">tomxor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789324\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787130\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=coredog64\" class=\"hnuser\">coredog64<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787130\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787185\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787185\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787200\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=roywiggins\" class=\"hnuser\">roywiggins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787200\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788771\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saidmasoud\" class=\"hnuser\">saidmasoud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788771\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786696\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gilbetron\" class=\"hnuser\">gilbetron<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786696\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786846\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786846\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787661\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gilbetron\" class=\"hnuser\">gilbetron<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787661\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16792043\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16792043\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786920\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=oaiey\" class=\"hnuser\">oaiey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786920\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787041\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walterstucco\" class=\"hnuser\">walterstucco<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787041\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787253\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787253\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787557\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianai\" class=\"hnuser\">ianai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787557\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787559\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jaxn\" class=\"hnuser\">jaxn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787559\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789206\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mattbee\" class=\"hnuser\">mattbee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789206\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16797249\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ljm\" class=\"hnuser\">ljm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16797249\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787243\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=147\" class=\"hnuser\">147<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787243\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787295\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Boulth\" class=\"hnuser\">Boulth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787295\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787450\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cmonfeat\" class=\"hnuser\">cmonfeat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787450\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789354\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Per_Bothner\" class=\"hnuser\">Per_Bothner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789354\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16792729\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=theshrike79\" class=\"hnuser\">theshrike79<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16792729\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787368\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dyukqu\" class=\"hnuser\">dyukqu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787368\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787513\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dawnerd\" class=\"hnuser\">dawnerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787513\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787996\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rcarmo\" class=\"hnuser\">rcarmo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787996\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787284\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=modzu\" class=\"hnuser\">modzu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787284\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786610\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=philsnow\" class=\"hnuser\">philsnow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786610\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788874\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e12e\" class=\"hnuser\">e12e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788874\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16794984\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gpvos\" class=\"hnuser\">gpvos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16794984\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786883\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ronsor\" class=\"hnuser\">ronsor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786883\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786676\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=baq\" class=\"hnuser\">baq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786676\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786702\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yulaow\" class=\"hnuser\">yulaow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786702\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788703\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merb\" class=\"hnuser\">merb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788703\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787170\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joezydeco\" class=\"hnuser\">joezydeco<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787170\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787536\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=147\" class=\"hnuser\">147<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787536\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788316\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jaxn\" class=\"hnuser\">jaxn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788316\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787635\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787635\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786940\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=donatj\" class=\"hnuser\">donatj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786940\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787293\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tootie\" class=\"hnuser\">tootie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787293\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785621\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Giako\" class=\"hnuser\">Giako<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785621\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785725\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lorenzhs\" class=\"hnuser\">lorenzhs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785725\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785688\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=latch\" class=\"hnuser\">latch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785688\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786339\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericabiz\" class=\"hnuser\">ericabiz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786339\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16795504\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Qub3d\" class=\"hnuser\">Qub3d<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16795504\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786444\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=latch\" class=\"hnuser\">latch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786444\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786858\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pentium166\" class=\"hnuser\">pentium166<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786858\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790716\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=latch\" class=\"hnuser\">latch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790716\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787872\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmontra\" class=\"hnuser\">pmontra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787872\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16794463\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericabiz\" class=\"hnuser\">ericabiz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16794463\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785707\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevesimmons\" class=\"hnuser\">stevesimmons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785707\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785646\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=robotmay\" class=\"hnuser\">robotmay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785646\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786346\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bhouston\" class=\"hnuser\">bhouston<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786346\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785654\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dawnerd\" class=\"hnuser\">dawnerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785654\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785682\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=koehr\" class=\"hnuser\">koehr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785682\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786214\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chongli\" class=\"hnuser\">chongli<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786214\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786333\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bhouston\" class=\"hnuser\">bhouston<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786333\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786605\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lardo\" class=\"hnuser\">lardo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786605\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786724\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=foodstances\" class=\"hnuser\">foodstances<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786724\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16795514\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Qub3d\" class=\"hnuser\">Qub3d<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16795514\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786567\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=owenwil\" class=\"hnuser\">owenwil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786567\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16785770\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=owenwil\" class=\"hnuser\">owenwil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16785770\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786447\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eknkc\" class=\"hnuser\">eknkc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786447\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786374\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=owenwil\" class=\"hnuser\">owenwil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786374\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16789004\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seltzered_\" class=\"hnuser\">seltzered_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16789004\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787042\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=camflan\" class=\"hnuser\">camflan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787042\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786325\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MiddleEndian\" class=\"hnuser\">MiddleEndian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786325\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786998\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=013a\" class=\"hnuser\">013a<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786998\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787180\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonathanfoster\" class=\"hnuser\">jonathanfoster<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787180\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787558\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=theflagbug\" class=\"hnuser\">theflagbug<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787558\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788570\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonathanfoster\" class=\"hnuser\">jonathanfoster<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788570\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790495\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dlsniper\" class=\"hnuser\">dlsniper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790495\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787013\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knappe\" class=\"hnuser\">knappe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787013\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16788364\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lifeformed\" class=\"hnuser\">lifeformed<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16788364\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787080\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aschampion\" class=\"hnuser\">aschampion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787080\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787301\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=013a\" class=\"hnuser\">013a<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787301\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787126\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thriptic\" class=\"hnuser\">Thriptic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787126\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786989\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jasonvorhe\" class=\"hnuser\">jasonvorhe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786989\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787699\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sethrin\" class=\"hnuser\">sethrin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787699\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787948\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=antoineMoPa\" class=\"hnuser\">antoineMoPa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787948\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16790598\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tbyehl\" class=\"hnuser\">tbyehl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16790598\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16791665\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bigmanwalter\" class=\"hnuser\">bigmanwalter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16791665\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786318\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=storrgie\" class=\"hnuser\">storrgie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786318\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786378\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=owenwil\" class=\"hnuser\">owenwil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786378\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786687\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Majestic121\" class=\"hnuser\">Majestic121<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786687\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786821\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jnwatson\" class=\"hnuser\">jnwatson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786821\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787532\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhoechtl\" class=\"hnuser\">jhoechtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787532\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786482\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=storrgie\" class=\"hnuser\">storrgie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786482\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786566\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=owenwil\" class=\"hnuser\">owenwil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786566\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786411\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TheForumTroll\" class=\"hnuser\">TheForumTroll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786411\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16786759\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexdeloy\" class=\"hnuser\">alexdeloy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16786759\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787574\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=theyinwhy\" class=\"hnuser\">theyinwhy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787574\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"16787731\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SahAssar\" class=\"hnuser\">SahAssar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16787731\">7 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"morespace\"><\/tr><tr><td><table border=\"0\"><tr><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=16785507&amp;p=2\" class=\"morelink\" rel=\"nofollow\">More<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:24:10+0000","updated_at":"2018-11-07T16:24:10+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":195,"title":"Best Image Sizes and How to Save Images For the Web (2018) - Flothemes","url":"https:\/\/flothemes.com\/flothemes-image-sizes\/","content":"<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19576\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web.jpg\" alt=\"Image sizes for web\" width=\"1500\" height=\"1050\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web.jpg 1500w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web-300x210.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web-768x538.jpg 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web-1024x717.jpg 1024w\" \/><\/p><h2>What are the best size images for the web and Flothemes?<\/h2>\n<p>When uploading images to the web there are a number of things to take into consideration, such as what is the image being used for, how large should the file be, and what type of image should be used for different items on your site. Remember when working with images on the web make sure to <a href=\"https:\/\/flothemes.com\/ultimate-guide-saving-images-web\/\" target=\"_blank\" rel=\"noopener\">optimize<\/a> them to increase loading time.<\/p>\n<h3>Image Use and Type<\/h3>\n<p>There are various cases where you will want to use different image sizes for different areas of your site, below I will cover the most common areas and the sizing options you can use for each. Note this is recommended sizing options, they can be increased or decreased depending on your site.<\/p>\n<h2>Slideshows<\/h2>\n<p>There are various slideshow types, <a href=\"http:\/\/demo.flothemes.com\/fiji\/\" target=\"_blank\" rel=\"noopener\">full screen<\/a>,\u00a0<a href=\"http:\/\/demo.flothemes.com\/rebecca\/home-2\/\" target=\"_blank\" rel=\"noopener\">simple slideshows<\/a>, <a href=\"http:\/\/demo.flothemes.com\/cozy\/\" target=\"_blank\" rel=\"noopener\">full width fixed height<\/a>, and <a href=\"http:\/\/demo.flothemes.com\/porto\/home-6\/\" target=\"_blank\" rel=\"noopener\">slick slideshows<\/a>. However we\u2019ll put them into 2 classes, full screen and regular.<\/p>\n<h3>Full Screen Slideshow<\/h3>\n<p><strong>2880 x 1500px<\/strong><\/p>\n<p>As a browser screen has a ratio of ~2:1 you\u2019ll need to select images in a different ratio that you normally shoot. You can select 5-10 images and then you should crop them to the sizing above to get the most out of your images with minimal cropping.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19569\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow.png\" alt=\"Full Screen Slideshow\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Full-Screen-Slideshow-1024x643.png 1024w\" \/><\/p>\n<h3>Regular\u00a0Slideshow<\/h3>\n<p><strong>Width Auto x 1500px<\/strong><\/p>\n<p>For other slideshows, you can save all images to 1500px in height as width will always be auto.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19571\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow.png\" alt=\"Regular Slideshow\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Regular-Slideshow-1024x643.png 1024w\" \/><\/p>\n<h2>Gallery<\/h2>\n<p><strong>Width Auto x 1500px<\/strong><\/p>\n<p>Much like slideshows, you can save all images to 1500px in height.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19570\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery.png\" alt=\"Gallery\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Gallery-1024x643.png 1024w\" \/><\/p>\n<h2>Blog<\/h2>\n<p><strong>1500px x Height Auto<\/strong><\/p>\n<p>For blog posts you\u2019ll want to think about the width attribute, most will be around 7-800px in width so 1500px in width will work great.<\/p>\n<p class=\"with-image\"><img class=\"alignnone wp-image-46088 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1.jpg\" alt=\"Blog image size\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1.jpg 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1-300x189.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1-768x483.jpg 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2015\/09\/Untitled-1-1024x643.jpg 1024w\" \/><\/p>\n<h2>Featured Images<\/h2>\n<p>There are 3 types of featured images, portrait, landscape and hero images. The key to having good featured images is consistency. If you want to use all portrait make sure they are the same size, the same goes for landscape and hero images. If using a masonry grid, the sizing above is not so relevant, just make sure that the images are at least 900px in width.<\/p>\n<p><strong>Hero Image \u2013 2880 x 1500px<\/strong><\/p>\n<p>The hero images are those that are full screen, examples can be found in <a href=\"http:\/\/demo.flothemes.com\/porto\/2013\/01\/10\/markup-image-alignment\/\" target=\"_blank\" rel=\"noopener\">Porto<\/a>, and <a href=\"http:\/\/demo.flothemes.com\/como\/2016\/04\/06\/nature-is-my-favourite-dj\/\" target=\"_blank\" rel=\"noopener\">Como<\/a>.<\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19566\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image.png\" alt=\"Featured Images Hero Image\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Hero-Image-1024x643.png 1024w\" \/><\/p>\n<p><strong>Landscape \u2013 1200 x \u00a0900px<\/strong><\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19567\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape.png\" alt=\"Featured Images Landscape\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Landscape-1024x643.png 1024w\" \/><\/p>\n<p><strong>Portrait \u2013 900 x 1200px<\/strong><\/p>\n<p class=\"with-image\"><img class=\"alignnone size-full wp-image-19568\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait.png\" alt=\"Featured Images Portrait\" width=\"1200\" height=\"754\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait.png 1200w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait-300x189.png 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait-768x483.png 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2013\/11\/Featured-Images-Portrait-1024x643.png 1024w\" \/><\/p>\n<h3>Recommended Image Sizes \u2013 Overview<\/h3>\n<p><strong>Slideshow Full Screen<\/strong> \u20132880\u00a0x 1500px<\/p>\n<p><strong>Regular\u00a0Slideshows<\/strong> \u2013 1500px height, width auto.<\/p>\n<p><strong>Galleries<\/strong> \u2013 1500px\u00a0<em>height, width auto<\/em>.<\/p>\n<p><strong>Blog Posts<\/strong> \u2013 1500px <em>width, height auto<\/em>. (If using a wide blog, use 2000px in width)<\/p>\n<p><strong>Featured Images<\/strong><\/p>\n<ul><li><em>Hero images (full screen) \u20132880\u00a0x 1500px.<\/em><\/li>\n<li>Portrait \u2013 900 x 1200<\/li>\n<li>Landscape \u2013 1200 x 900<\/li>\n<\/ul><p>If you\u2019re having any problems with image sizing make sure to take a look at the image sizing options in your Flothemes settings to make sure they are set up correctly.<\/p>\n<h2>How to save images for the web?<\/h2>\n<p>Saving images for the web has always been bit of a mystery for many. Uploading unoptimized images will not only create issues with the user experience on your site, but will also affect your SEO strategy (loading speed, bounce rate, ranking, etc). In this Ultimate Guide to Saving Images for the web we aim to show you how to inspect the various sections on your website to identify the correct sizes for your images before uploading them to the web. We will also explain how to export \/ save images correctly, so that they\u2019re optimized for speed.<\/p>\n<p>Finally we will provide some Image SEO tips, how to title your images correctly when saving them and why using ALT tags will improve your SEO. For already uploaded images to your site, we\u2019ll show you how to find which images are too large and make your site slow down. Let\u2019s begin!<\/p>\n<h2>1. Choosing the right image dimensions<\/h2>\n<p>One of the hardest parts about creating a website has always been selecting the right images, then making sure they are optimized and have the correct image dimensions for each section on your site. Thankfully, there are a few ways that can help you easily inspect your website images, and identify the correct sizing needed for each element on your site.<\/p>\n<p>An important thing to remember is that for retina devices, you need to double the image pixel density. What that means is, you need images twice the size for retina devices so they look nice and crisp. So once you\u2019ve found the correct image sizes for each section on your site, you\u2019ll want to double those sizes, and only then export and upload them to your site.<\/p>\n<h3>a. Using the Image Size Info Extension<\/h3>\n<p>The first method to find image sizes for your site is by using the<a href=\"https:\/\/chrome.google.com\/webstore\/detail\/image-size-info\/oihdhfbfoagfkpcncinlbhfdgpegcigf?hl=en\" target=\"_blank\" rel=\"noopener\"> Image Size Info<\/a> extension for Chrome (If you don\u2019t use Chrome on your computer, we highly recommend using it while creating your site). Once you\u2019ve installed the extension, this tool will then allow you to find out the image sizes of every image element on your site.<\/p>\n<p>Check out the images on your site, make notes, and then with this knowledge you\u2019ll be able to tailor your photographs exactly to fit the spaces as they are intended in the theme design. Just remember, with retina screens you do require your images to be twice as large as the intended space. So if you have an image space of 750x500px, for retina it should be 1500x1000px when uploading to WordPress (or any other platform).<\/p>\n<p>For example: when we inspect an image with this extension, it\u2019s being displayed at 780x520px, but it\u2019s actual size is 1560x1040px. This is because we\u2019re using a retina screen when inspecting the image. In this case, you can use the dimensions sizes to pick your image size.<\/p>\n<p>On a non-retina device the displayed and dimensions sizes will remain the same. In this case, double up the numbers and upload images at that size.<\/p>\n<p>The main limitation of this tool, is that it can\u2019t inspect background images for sizing. On many themes, background images are used instead of image HTML elements, as they are more flexible when displaying images. If you\u2019re unable to inspect the images because they are background images, you can use the next method to review image sizes on your site.<\/p>\n<h3>b. Using the Inspector Tool<\/h3>\n<p>With Chrome and Safari you can also use the developers inspector tool to find image sizes on the site. The principle will be the same as before, inspect the various elements on the site, and note down the sizes. Double the image sizes for retina devices and then upload them to WordPress, once they have been optimized for web and titled correctly for SEO.<\/p>\n<p>The limitations of this method will be that if the site is using some responsive elements set in REM values, images may be shrunk when using the inspector in the same window. If the site is also using responsive elements, they\u2019ll be larger on bigger screens, and smaller on others. It\u2019s best to use a larger screen where possible when checking the site, so you may get the most accurate image sizing. In any case, using this method will give you a good indication of sizes for your site and will allow you to approach adding images to your website with more care.<\/p>\n<h3>Chrome<\/h3>\n<p>Chrome has built in developer tools, which is why it\u2019s the most popular browser on the web for developers and designers. The following video explains how you can inspect the images on your site to find out their sizes:<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/VBWaX_F7R8E\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<h3>Firefox<\/h3>\n<p>With Firefox, much like chrome, they have a built in inspector tool that allows you to check images for your site.<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/eDSrj7eooxA\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<h3>Safari<\/h3>\n<p>If you\u2019re using Safari, you can use the inspector tool to see the sizes of images. Check the following guide on how to enable the developer tools.<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/eSgZqL1E-1I\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<p>Now you can go to different sections in your site and prepare the image dimensions.<\/p>\n<p><strong>For galleries<\/strong> you\u2019ll want to look at the height of the images, as they are generally constrained by height (so width can be set to auto).<br \/><strong>For blog posts<\/strong> you\u2019ll want to note the width for your images, as they are constrained by width (the height can be auto).<\/p>\n<p><strong>For the rest<\/strong> of the elements on your site, you should use the method described above to get the correct dimensions, and upload them at optimal dimensions for your site.<\/p>\n<p>Now that you know what dimensions you need for all the elements on your site, we\u2019ll explain how to optimize all your images for size, that\u2019s in terms of data sizing. The file size of your images is also very important for the web. Uploading bloated and heavy images on your website will strain your server resources, slow down your site, and in effect create a negative user experience.<\/p>\n<h2>2. Optimizing and Saving Images for web<\/h2>\n<p>There are a number of things to consider when optimizing and saving images for the web. You should consider what file type to use, JPG, PNG or GIF depending on the purpose. You\u2019ll want to label your images correctly for SEO purposes and you\u2019ll want to save them in the right color format (sRGB for web).<\/p>\n<p>Depending on the purpose, you need to save images in one of the following formats .jpg, .gif, or .png. JPG will be the most common format when saving images for your portfolio website, they will also be the lightest in size.<\/p>\n<p>For filenames, please make sure they are appropriate for each particular image or blog post. Use only letters, numbers, underscores and hyphens. Try to use only English letters. Characters from other languages, question marks, spaces, percent signs may upload incorrectly or cause unexpected behavior in your galleries or blog posts.<\/p>\n<p>For example, if you\u2019ve just done a boho wedding in The Biltmore Santa Barbara, label the images, boho-wedding-biltmore-santa-barbara_0001.jpg this will be much more SEO friendly, and could provide better ranking for your site.<\/p>\n<h3>Optimizing images in Lightroom<\/h3>\n<p>Let\u2019s face it, there are lots of different settings to choose and endless combinations of numbers to input. However, if you already know the image format and required file image dimensions, it\u2019s much easier.<\/p>\n<p>In Lightroom, you can get to the Export Dialog in 3 ways:<\/p>\n<h3>1. Highlight and Export<\/h3>\n<p>You can simply highlight \/ select any photo from your Lightroom library and right-click on the photo. From the pop up menu select the Export option.<\/p>\n<h3>2. File and Export<\/h3>\n<p>The second way is to go to <strong>File -&gt; Export<\/strong><\/p>\n<h3>3. Using the keyboard shortcut<\/h3>\n<p>The third, and fastest way is to use a keyboard shortcut:<\/p>\n<p><strong>shift + command + E (\u21e7+\u2318+E)<\/strong><\/p>\n<p>It doesn\u2019t matter how you get there. Any of these will work, so choose the easiest one for you to remember.<\/p>\n<h3>Lightroom Export Dialog<\/h3>\n<p>Now that you know how to export the images for web, we\u2019ll cover the 4 sections and the various options to choose in the Lightroom export dialog to allow you to export optimized images for the web.<\/p>\n<p class=\"with-image\"><img class=\"aligncenter wp-image-30663 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom.png\" alt=\"Lightroom export dialog\" width=\"850\" height=\"902\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom.png 850w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom-283x300.png 283w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/export-dialog-lightroom-768x815.png 768w\" \/><\/p>\n<h4>1. File Naming<\/h4>\n<p>Each export can feature its own rename settings. If you stick to the default option, your images will have the same name as the original files, and saved as a JPEG image. However we would highly recommend that you tick the \u201crename to\u201d box and choose one of the many renaming templates that Lightroom offers to customize the exported images. In the File Naming, we renamed the image filenames to \u201cCustom Name \u2013 Sequence\u201d where the Custom Text is closely connected to our blog post title or blog post keywords.<\/p>\n<p class=\"with-image\"><img class=\"aligncenter wp-image-30669 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/file-naming-lightroom.png\" alt=\"Lightroom file naming\" width=\"678\" height=\"198\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/file-naming-lightroom.png 678w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/file-naming-lightroom-300x88.png 300w\" \/><\/p>\n<h4>2. In File Settings<\/h4>\n<p>Make sure that your Image Format is set to JPEG. This is the most web-friendly option. Your Quality should not be less than 65. Normally we set it to 75. Your Colour Space needs to be sRGB. This is the color space used by the web so it will render the best. <strong>DO NOT<\/strong> check the box \u201cLimit File Size to:\u201d. Checking that box would decrease file quality significantly.<\/p>\n<p>To summarize, when saving you images use these settings:<\/p>\n<ul><li>75% Quality<\/li>\n<li>.jpeg format (.png for logos and icons)<\/li>\n<li>sRGB colors<\/li>\n<\/ul><h4>3.<strong>\u00a0Image Sizing<\/strong><\/h4>\n<p>Check the box Resize to Fit and select Width &amp; Height. In the box W: H: specify W (Width in pixels) to the value needed (1500px for blog posts, and 2880 for slideshows). By leaving the value in H (height) blank it will be set to auto sizing.<\/p>\n<p>For Galleries, you can leave W as blank, and set H (Height) to 1500px, this should be optimal for most themes.<\/p>\n<p>The sizes above are already optimized for retina devices.<\/p>\n<h4>4. Output Sharpening<\/h4>\n<p>Check the box next to \u201cSharpen For\u201dand select the amount to \u201cHigh\u201d.<\/p>\n<p>You can now Export the images, and follow this process anytime you\u2019re exporting images.<\/p>\n<h3>Image Quality Settings<\/h3>\n<p>In a photographer\u2019s dream world, we would always export images at maximum quality. In reality, we find out that maximum quality exports aren\u2019t required. At web sizes, the differences in quality are often hard to perceive, and that means we can reduce the quality slightly, reducing the file size significantly.<\/p>\n<p>As long as the images aren\u2019t soft, you\u2019re clients won\u2019t know the difference between what picture you took and the image thats displayed on your site.<\/p>\n<p>This also applies to colors of images. Everyone\u2019s screen is calibrated differently, so worrying about small color variations between an image in a browser and in your editing tool of choice is a waste of time, unless it ruins the image quality.<\/p>\n<p>On the upper end of the quality scale, there are diminishing returns. The difference in perceived quality between an 75% and 100% quality setting is hard to see, but it increases the file size significantly.<\/p>\n<p class=\"with-image\"><img class=\"alignnone wp-image-30666 size-full\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart.jpg\" alt=\"image size optimization\" width=\"1600\" height=\"856\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart.jpg 1600w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart-300x161.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart-768x411.jpg 768w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/lightroom-export-chart-1024x548.jpg 1024w\" \/><\/p>\n<p>Given the above results, selecting the export of 65-75 quality. It strikes the correct balance of quality and optimal image file size.<\/p>\n<p>Now it\u2019s your turn to, rename your files, export them for the web, save them for web and you\u2019ll have some really well optimized images for the web.<\/p>\n<h3>Bonus Step<\/h3>\n<p>There is one final step that can help drastically remove excess data from images, and reducing file size, is using tools such as\u00a0<a href=\"https:\/\/flothemes.com\/faster-website-jpegmini\/\" target=\"_blank\" rel=\"noopener\">JPEG Mini<\/a>\u00a0or <a href=\"https:\/\/tinypng.com\/\" target=\"_blank\" rel=\"noopener\">TinyPNG<\/a>.<\/p>\n<p>If you have JPEG Mini pro, you can add the <a href=\"http:\/\/help.jpegmini.com\/knowledge_base\/topics\/how-to-activate-lightroom-plugin-on-a-mac\" target=\"_blank\" rel=\"noopener\">lightroom extension<\/a>, optimizing images as you export them, simple!<\/p>\n<p>If not, just run your freshly exported images through JpegMini or TinyPNG.<\/p>\n<p>This will be the final step in creating fast web optimized images to keep your site\u2019s speed up, and images looking great!<\/p>\n<h2>3. Image SEO<\/h2>\n<p>Now that you\u2019ve got your images sized, optimized and uploaded to your site \u2013 now you\u2019ll want to add ALT tags! This is the final step in creating optimized images for SEO purposes. The ALT tag is what search engines use to read images. Without them, your images are pretty much non-existent on the web.<\/p>\n<p>Thankfully with WordPress you can go into your media library after upload, and label each image. One great plugin that will let you do this quickly is <a href=\"https:\/\/wordpress.org\/plugins\/wp-meta-seo\/\" target=\"_blank\" rel=\"noopener\">WP Meta SEO<\/a>, when added you can go to WP Meta SEO &gt; Image Information and add ALT Tags to any images where it isn\u2019t present, once completed the plugin can be de-activated:<\/p>\n<p class=\"with-image\"><img class=\"alignnone wp-image-43439 size-large\" src=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-1024x355.jpg\" alt=\"alt tag wp plugin\" width=\"1024\" height=\"355\" srcset=\"https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-1024x355.jpg 1024w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-300x104.jpg 300w, https:\/\/l5vd03xwb5125jimp1nwab7r-wpengine.netdna-ssl.com\/wp-content\/uploads\/2017\/08\/image-seo-alt-tags-768x266.jpg 768w\" \/><\/p>\n<p>There are also several plugins that will add the alt tags automatically, and if you\u2019ve went ahead and renamed the images like we mentioned in the previous section, you\u2019ll have pretty well optimized ALT text.\u00a0<a href=\"https:\/\/wordpress.org\/plugins\/automatic-image-alt-attributes\/\" target=\"_blank\" rel=\"noopener\">Automatic Image ALT Attributes<\/a>\u00a0is a good plugin solution for this, simply install it and it will automatically add ALT tags upon upload from the image description, or title.<\/p>\n<p>So that\u2019s it folks. The 3 steps to having optimized images for your website.<\/p>\n<p>Start with figuring out your image dimensions, then save them correctly for the web, add proper titles and optimize their size. Finally, add alt tags to improve SEO via WordPress (It should be possible to add ALT tags in all content management systems, refer to the documentation to find out how, or reach out to support).<\/p>\n<h3>Finding Unoptimized Images<\/h3>\n<p>If you already uploaded images to your site and want to know how to find the ones that are slowing down your page, check out the following video. Tools such as GTMetrix or Pingdom, or using the browser console will easily let you find images that are too large for the web.<\/p>\n<iframe src=\"https:\/\/www.youtube.com\/embed\/Dd5RnMbT2G0\" width=\"640\" height=\"360\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\" data-mce-fragment=\"1\"> <\/iframe>\n<p>If you would like to learn more about optimizing your site for speed, check out this comprehensive <a href=\"https:\/\/flothemes.com\/flo-learn\/speed-optimization-for-wordpress\/\" target=\"_blank\" rel=\"noopener\"><strong>online course.<\/strong><\/a><\/p>\n<h3>Flothemes<\/h3>\n<h3>Empowering You<\/h3>","created_at":"2018-11-07T16:24:07+0000","updated_at":"2018-11-07T16:24:07+0000","published_at":"2015-09-11T15:49:30+0000","published_by":["Allen Mark"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":13,"domain_name":"flothemes.com","preview_picture":"https:\/\/flothemes.com\/wp-content\/uploads\/2015\/09\/image-sizes-for-web.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":192,"title":"How to install finger print driver on lenovo X1 carbon on Ubuntu 16.04","url":"https:\/\/askubuntu.com\/questions\/948083\/how-to-install-finger-print-driver-on-lenovo-x1-carbon-on-ubuntu-16-04","content":"<p>I need to  install the finger print driver of my X1 carbon 5th gen on Ubuntu 16.04. Where can I find drivers for Linux distributions?<\/p><p>Here are the outputs of <code>lspci<\/code><\/p><pre>kasun@kasun-ThinkPad-X1-Carbon-5th:~$ lspci\n00:00.0 Host bridge: Intel Corporation Device 5904 (rev 02)\n00:02.0 VGA compatible controller: Intel Corporation Device 5916 (rev 02)\n00:08.0 System peripheral: Intel Corporation Sky Lake Gaussian Mixture Model\n00:14.0 USB controller: Intel Corporation Sunrise Point-LP USB 3.0 xHCI Controller (rev 21)\n00:14.2 Signal processing controller: Intel Corporation Sunrise Point-LP Thermal subsystem (rev 21)\n00:16.0 Communication controller: Intel Corporation Sunrise Point-LP CSME HECI (rev 21)\n00:1c.0 PCI bridge: Intel Corporation Device 9d10 (rev f1)\n00:1c.2 PCI bridge: Intel Corporation Device 9d12 (rev f1)\n00:1c.4 PCI bridge: Intel Corporation Sunrise Point-LP PCI Express Root Port (rev f1)\n00:1d.0 PCI bridge: Intel Corporation Device 9d18 (rev f1)\n00:1f.0 ISA bridge: Intel Corporation Device 9d4e (rev 21)\n00:1f.2 Memory controller: Intel Corporation Sunrise Point-LP PMC (rev 21)\n00:1f.3 Audio device: Intel Corporation Device 9d71 (rev 21)\n00:1f.4 SMBus: Intel Corporation Sunrise Point-LP SMBus (rev 21)\n00:1f.6 Ethernet controller: Intel Corporation Ethernet Connection (4) I219-LM (rev 21)\n02:00.0 Unassigned class [ff00]: Realtek Semiconductor Co., Ltd. RTS525A PCI Express Card Reader (rev 01)\n04:00.0 Network controller: Intel Corporation Device 24fd (rev 88)\n05:00.0 Non-Volatile memory controller: Samsung Electronics Co Ltd Device a804\nkasun@kasun-ThinkPad-X1-Carbon-5th:~$ \n<\/pre><p>and the <code>lsusb<\/code> output<\/p><pre>kasun@kasun-ThinkPad-X1-Carbon-5th:~$ lsusb\nBus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub\nBus 001 Device 004: ID 138a:0097 Validity Sensors, Inc. \nBus 001 Device 003: ID 13d3:5682 IMC Networks \nBus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub\nkasun@kasun-ThinkPad-X1-Carbon-5th:~$ \n<\/pre>","created_at":"2018-11-07T16:24:03+0000","updated_at":"2018-11-07T16:24:03+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"askubuntu.com","preview_picture":"https:\/\/cdn.sstatic.net\/Sites\/askubuntu\/img\/apple-touch-icon@2.png?v=c492c9229955","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":193,"title":"Some Thoughts on Generative Art \u00b7 inconvergent","url":"https:\/\/inconvergent.net\/thoughts-on-generative-art\/","content":"<div class=\"text\"><p>\nThe following is a combination of segments from a few different talks I've\ngiven . It touches\nbriefly on the history of generative art, as well as some of my experience\nwith it.\n<\/p><p>\nIf you are looking for a comprehensive introduction to generative art, you\nwill unfortunately have to keep looking, but I want to mention some things\nI wish I had been aware of earlier.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/head.jpg\" alt=\"Sand Splines\" title=\"Sand Splines\" \/><p><a href=\"https:\/\/inconvergent.net\/generative\/sand-spline\/\">Sand Splines<\/a><\/p><\/div><div class=\"text\"><p>\nAccording to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Generative_art\">Wikipedia<\/a>,\ngenerative art is made by a process that has some level of autonomy.\nThe role of the artist is to design or influence this process to some\ndegree. This process can be any number of things, and as such it is not\nlimited to the digital realm. It can just as easily be a system driven by a\nchemical reaction. Or a set of rules for coloring in a checkered sheet of\npaper.\n<\/p><p>\nWhen I use the word <em>generative<\/em> I tend to mean that some kind of\nrandomness is involved, however I do not believe this to be a requirement\n. I have little interest\nin a strict definition of these terms, so instead I will just mention that\nthere are several terms that seem to be used be people who work within\ndifferent disciplines. Often they refer to a set of intersecting ideas . This includes, but is\nnot limited to:\n<\/p><ul><li>Generative<\/li>\n<li>Algorithmic<\/li>\n<li>Procedural (Art)<\/li>\n<li>Procedural Generation (PCG\/ProcGen)<\/li>\n<li>Parametric or Computer-Aided Design<\/li>\n<li>Computational Art\/Architecture\/Design<\/li>\n<li>Computer (Art)<\/li>\n<\/ul><p>\nIn some sense I would argue that it does not really matter whether you are\ngenerating characters, dialogue or environments for a computer game,\ngeometry as part of a building, digital or analogue art, a musical\ncomposition, or a poem. In all cases you are using a system of some kind to\nhelp you along. Either towards a specific goal, some unidentified result,\nor somewhere in-between.\n<\/p><p>\nTo me the main point is usually to experiment with a small system that\nconsists of a set of relatively simple rules. Sometimes the goal is to\nrecreate a particular behaviour\u2014for instance a biological process or\nsome other natural phenomena\u2014sometimes not.\n<\/p><p>\nMy process is usually highly experimental. And while there is usually some\nkind of underlying goal there is not always any interest on my part in\nactually reaching that particular goal. I might be just as happy to\ndiscover some unexpected behaviour and explore that instead. Overall this\nmeans that the process is continuous, exploratory, and never really\ncomplete. Naturally this process has also changed quite a lot over time.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/nake-klee.jpg\" alt=\"Hommage \u00e0&#10;  Paul Klee\" title=\"Hommage \u00e0 Paul Klee\" \/><p>Hommage \u00e0 Paul Klee, Frieder Nake 1965<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nAt this point I have had some kind of interest in generative art for at\nleast ~ 7 years . I\nstarted working actively on it when I was at university, but really it\nstarted before that.\n<\/p><p>\nWhen I was about 16 I spent a year as an exchange student in Australia. At\nsome point I came across <a href=\"http:\/\/complexification.net\">complexification.net<\/a> by Jared\nTarbell. At the time I didn't really know enough about programming to\nrealise how it was done, or how to get started. As a result I mostly just\nspent a whole lot of time looking at the various animations. Then I\ngradually forgot about the site.\n<\/p><p>\nI'm not sure what reminded me of the site again, but I had learnt quite a\nlot of programming, so I started experimenting. Seeing as I was studying\nphysics and numerical mathematics it turned out I had a pretty good\nfoundation for this kind of programming. However, <em>because<\/em> I was\nstudying numerical mathematics I almost felt obliged to model some kind of\n<em>real thing<\/em>. Recently I've started to realise that there is no\nparticular reason why I should impose this limitation on my work.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/nake-walk-through-raster.jpg\" alt=\"Walk Through Raster\" title=\"Walk Through Raster\" \/><p>Walk Through Raster, Frieder Nake 1966<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nI got into generative art as a distraction at university, but people have\nbeen doing this for far longer. Among the people frequently credited with\nbeing part of the advent of computer art is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Frieder_Nake\">Frieder Nake<\/a>. Two\nothers include <a href=\"https:\/\/en.wikipedia.org\/wiki\/A._Michael_Noll\">Michael Noll<\/a> and\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Georg_Nees\">Georg Nees<\/a>.\n<\/p><p>\nUnbeknownst to each other these people had started making computer\ngenerated plotter drawings based on randomness and\/or rules in 1963-64.\nThen, in 1965, followed three separate exhibitions: Georg Nees exhibited\nwork in Stuttgart in February, Noll and <a href=\"https:\/\/en.wikipedia.org\/wiki\/B%C3%A9la_Julesz\">B\u00e9la Julesz<\/a> had\nan exhibition in New York City in April, and finally, Nake and Nees\nexhibited work in Stuttgart in November. If you would like a more authentic\nintroduction to this history I can not recommend Frieder Nake's <a href=\"https:\/\/vimeo.com\/104315361\">talk from Eyeo 2014<\/a> enough.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/molnar-desordres-3.jpg\" alt=\"(Des)Ordres\" title=\"(Des)Ordres\" \/><p>(Des)Ordres, Vera Moln\u00e1r 1974<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nThere are two other people who's work I also find thoroughly captivating.\nOne is pioneering artist <a href=\"https:\/\/en.wikipedia.org\/wiki\/Vera_Moln%C3%A1r\">Vera Moln\u00e1r<\/a>,\nwho began producing computer art in 1968. The other one is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Manfred_Mohr\">Manfred Mohr<\/a>, an\naction painter and jazz musician, who started working with computers in\n1969.\n<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/mohr-p62.jpg\" alt=\"P62\" title=\"P62\" \/><p>P62, Manfred Mohr 1970<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text-img\" itemscope=\"itemscope\" itemtype=\"https:\/\/schema.org\/VisualArtwork\"><img itemprop=\"image\" src=\"https:\/\/inconvergent.net\/img\/writing\/thoughts\/mohr-p197.jpg\" alt=\"P197\" title=\"P197\" \/><p>P197, Manfred Mohr 1977-79<br \/>Victoria and Albert Museum online collection<\/p><\/div><div class=\"text\"><p>\nLooking at these old works is quite stunning. In some ways They are rather\nminimalistic. It's easy to be tricked into believing that this is because\nthere is a limit to what you can do with <a href=\"https:\/\/en.wikipedia.org\/wiki\/Punched_card\">punch cards<\/a> and\nplotter. However, this minimalism is an excellent way to let you really pay\nattention to the intricacies of the algorithms behind it. You can clearly\nsee that there is some sort of pattern there, but the details are not at\nall obvious.\n<\/p><p>\nInstead of trying to tell you how they are done, I'm simply going to say\nthat while we might have better tools, and more processing power today,\nthat does not automatically mean we can make better (generative) art. \n<\/p><p>\nIf you are interested in my process and work in particular you could have a\nlook at another text I've written: <a href=\"https:\/\/inconvergent.net\/generative\/\">On Generative\nAlgorithms<\/a>.\n<\/p><\/div><div class=\"text\"><div class=\"footnotes\"><hr \/><ol><li>\nNone of it is verbatim, as I am completely incapable of writing a\nmanuscript and sticking to it.\n<\/li>\n<li>\nTechnically I use <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pseudorandom_number_generator\">\nPseudorandom Number Generators (PRNG)<\/a>. Seeing as these generators\nproduce deterministic results, there is no real difference between a\nprogram that includes this kind of randomness and a program that does\nnot use any randomness at all. There would be a conceptual difference\nif the random numbers used were <a href=\"https:\/\/www.random.org\/\">truly\nrandom<\/a>, but in the case of my work I am almost entirely certain\nthat you would not be able to inspect a work and decide whether it was\nmade using true randomness or a PRNG.\n<\/li>\n<li>\nI don't mean to diminish the importance of definitions, and there may\nvery well be a well formulated definition of some (or all) of these\nterms. However, if you are getting started it is useful to know the\nterms people use, and then you can look into them on your own.\n<\/li>\n<li>\nIt seems I registered this domain on the 5th of August 2010. This\nlends some doubt to my recollection that I registered this domain when\nI was supposed to be reading for my exams; the summer semester at my\nuniversity usually had exams in May. Unless I was reading for\ncontinuation exams, which did happen several times.\n<\/li>\n<li>\nInfo in this section is from the Wikipedia article about <a href=\"https:\/\/en.wikipedia.org\/wiki\/Frieder_Nake\">Frieder Nake<\/a>\nand <a href=\"https:\/\/vimeo.com\/104315361\">his presentation at Eyeo in\n2014<\/a>.\n<\/li>\n<li>\nNaturally we can, and should, try.\n<\/li>\n<\/ol><\/div><\/div>","created_at":"2018-11-07T16:24:03+0000","updated_at":"2018-11-07T16:24:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"inconvergent.net","preview_picture":"https:\/\/inconvergent.net\/img\/writing\/thoughts\/head.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":194,"title":"Image SEO: alt tag and title tag optimization \u2022 Yoast","url":"https:\/\/yoast.com\/image-seo-alt-tag-and-title-tag-optimization\/","content":"<p>Adding images to your articles encourages people to read them, and well-chosen images can also back up your message. But you should always remember to give your images good <code>alt<\/code> attributes: alt tags and title tags strengthen the message of your articles with search engine spiders and improve the accessibility of your website. This article explains all about alt tags and title tags and why you should optimize them.<\/p><p>Note: the term \u201calt tag\u201d is a commonly used abbreviation of what\u2019s actually an alt attribute on an <code>img<\/code> tag. The alt tag of any image on your site should describe what\u2019s on it. Screen readers for the blind and visually impaired will read out this text and therefore make your image accessible.<\/p><h2>What are <code>alt<\/code> tags and <code>title<\/code> tags?<\/h2><p>This is a complete HTML image tag:<\/p><p><code>&lt;img src=\u201cimage.jpg\u201d alt=\u201cimage description\u201d title=\u201cimage tooltip\u201d&gt;<\/code><\/p><p>The alt and title attributes of an image are commonly referred to as alt tag or alt text and title tag \u2013 even though they\u2019re not technically tags. The alt text describes what\u2019s on the image and the function of the image on the page. So if you are using an image as a button to buy product X, the alt text should say: \u201cbutton to buy product X.\u201d<\/p><p>The alt tag is used by screen readers, which are browsers used by blind and visually impaired people, to tell them what is on the image. The title attribute is shown as a tooltip when you hover over the element, so in the case of an image button, the image title could contain an extra call-to-action, like \u201cBuy product X now for $19!\u201d.<\/p><p>Each image\u00a0<em>should<\/em>\u00a0have an alt text, not just for\u00a0<a href=\"https:\/\/yoast.com\/what-is-seo\/\">SEO<\/a>\u00a0purposes but also because blind and visually impaired people won\u2019t otherwise know what the image is about, but a title attribute is\u00a0<em>not<\/em>\u00a0required. It can be useful but in most cases, leaving it out shouldn\u2019t be too much of an issue.<\/p><h3>But what if an image doesn\u2019t have a purpose?<\/h3><p>If you have images in your design that are\u00a0<em>purely<\/em>\u00a0there for design reasons, you\u2019re doing it wrong, as those images should be in your CSS and not in your HTML. If you really can\u2019t change these images, give them an empty alt attribute, like so:<\/p><p><code>&lt;img src=\"image.png\" alt=\"\"&gt;<\/code><\/p><p>The empty alt attribute makes sure that screen readers skip over the image.<\/p><h2>alt text and SEO<\/h2><p>Google\u2019s\u00a0<a href=\"https:\/\/support.google.com\/webmasters\/answer\/114016?hl=en\">article about images<\/a>\u00a0has a heading \u201cCreate great alt text\u201d. This is no coincidence because Google places a relatively high value on alt text to determine not only what is on the image but also the topic of the surrounding text. This is why, in our\u00a0<a href=\"https:\/\/yoast.com\/wordpress\/plugins\/seo\/\">Yoast SEO<\/a>\u00a0content analysis, we have a feature that specifically checks that you have at least\u00a0<em>one<\/em>\u00a0image with an alt tag that contains your focus keyword.<\/p><p>Yoast SEO has the following three stages for images and their alt text in your posts:<\/p><p><img class=\"aligncenter wp-image-502513 size-full\" title=\"Image of the error Yoast SEO throws when there is no image in a post\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-no-image-error.png\" alt=\"Image of the error Yoast SEO throws when there is no image in a post\" width=\"457\" height=\"24\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-no-image-error.png 457w, https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-no-image-error-250x13.png 250w\" \/><\/p><p><img class=\"aligncenter wp-image-502516 size-full\" title=\"Image of the error Yoast SEO throws when there is no image with an alt tag containing the focus keyword in the post\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-no-image-alt-error.png\" alt=\"Image of the error Yoast SEO throws when there is no image with an alt tag containing the focus keyword in the post\" width=\"525\" height=\"24\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-no-image-alt-error.png 525w, https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-no-image-alt-error-250x11.png 250w\" \/><\/p><p><img class=\"aligncenter wp-image-502519 size-full\" title=\"Image of the green bullet notice Yoast SEO gives when it has found an image in your post containing the posts focus keyword in the alt tag\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-image-focus-keyword-notice.png\" alt=\"Image of the green bullet notice Yoast SEO gives when it has found an image in your post containing the posts focus keyword in the alt tag\" width=\"488\" height=\"25\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-image-focus-keyword-notice.png 488w, https:\/\/yoast.com\/app\/uploads\/2015\/10\/yoast-seo-image-focus-keyword-notice-250x13.png 250w\" \/><\/p><p>We\u2019re definitely\u00a0<em>not<\/em>\u00a0saying you should spam your keyword into every alt tag. You need good, high quality, related images for your posts, where it makes sense to have the focus keyword in the alt text.<\/p><p>If you look at the three images above, they have the same alt and title tags so you can hover to see them \u2014 you\u2019ll see each of them has fairly long and descriptive alt text. That\u2019s what a good alt text looks like.<\/p><p>If your image is of a specific product, include both the full product name and the product ID in the alt tag so that it can be more easily found. In general: if a keyword could be useful for finding something that\u00a0<em>is on the image<\/em>, include it in the alt tag if you can.<\/p><h2><code>alt<\/code> and <code>title<\/code> attributes in WordPress<\/h2><p>When you upload an image to WordPress, you can set a title and an alt attribute. By default, it uses the image filename in the title attribute, which, if you don\u2019t enter an alt attribute, it copies to the alt attribute. While this is better than writing nothing, it\u2019s pretty poor practice. You really need to take the time to craft a proper alt text for every image you add to a post. The interface makes it easy: click an image, hit the edit button, and you\u2019ll see this:<\/p><p><a href=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/image-alt-title-wordpress.png\"><img class=\"alignnone size-large wp-image-502523\" src=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/image-alt-title-wordpress-600x332.png\" alt=\"image-alt-title-wordpress\" width=\"600\" height=\"332\" srcset=\"https:\/\/yoast.com\/app\/uploads\/2015\/10\/image-alt-title-wordpress-600x332.png 600w, https:\/\/yoast.com\/app\/uploads\/2015\/10\/image-alt-title-wordpress-250x138.png 250w, https:\/\/yoast.com\/app\/uploads\/2015\/10\/image-alt-title-wordpress.png 917w\" \/><\/a><\/p><p>There\u2019s no excuse for not doing this right, other than laziness. Your (image) SEO will truly benefit if you get these tiny details right. Visually challenged users will also like you all the more for it.<\/p><h2>Read more about image SEO?<\/h2><p>We have a popular (and longer) article about\u00a0<a href=\"https:\/\/yoast.com\/image-seo\/\">Image SEO<\/a>. That post goes into a ton of different ways to optimize images but is relatively lacking in detail when it comes to <code>alt<\/code> and <code>title<\/code> tags \u2013 think of this as an add-on to that article. I recommend reading it when you\u2019re done here.<\/p><p class=\"readmore\">Read more: <a data-prefix=\"Read more\" href=\"https:\/\/yoast.com\/image-seo\/\">Optimizing images for SEO \u00bb<\/a><\/p>","created_at":"2018-11-07T16:24:03+0000","updated_at":"2018-11-07T16:24:03+0000","published_at":"2015-10-20T14:00:51+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"yoast.com","preview_picture":"https:\/\/yoast.com\/app\/uploads\/2017\/12\/most_read_5_image_seo_alt_tags.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":190,"title":"Pixieset - Client photo gallery for modern photographers.","url":"https:\/\/pixieset.com\/","content":"<div class=\"container-fluid index intro\"><p><a class=\"logo\" href=\"https:\/\/pixieset.com\/\">PIXIESET<\/a><\/p><div class=\"row content\"><div class=\"col-md-12 text-center\"><a href=\"https:\/\/pixieset.com\/registration\/\" class=\"btn btn-primary get-started\">GET STARTED<\/a><p class=\"btn-desc\">Free forever. Upgrade when you need to.<\/p><\/div><\/div><\/div><div class=\"container-fluid sales-section social\"><div class=\"row\"><div class=\"col-md-12 text-center\"><div class=\"mobile-features visible-sm visible-xs row\"><div class=\"col-sm-4\"><p>\n<i class=\"fas fa-bullhorn fa-lg\">\nYOUR BRAND ON EVERY GALLERY\n<\/i><\/p><\/div><div class=\"col-sm-4 sharing\"><p>\n<i class=\"fab fa-facebook-square fa-lg facebook\">\n<i class=\"fab fa-google-plus-square fa-lg google\">\n<i class=\"fab fa-pinterest-square fa-lg pinterest\">\nSOCIAL SHARING BUILT-IN\n<\/i><\/i><\/i><\/p><\/div><div class=\"col-sm-4\"><p>\n<i class=\"fas fa-mobile-alt fa-lg\">\nDESIGNED FOR THE MOBILE GENERATION\n<\/i><\/p><\/div><\/div><\/div><\/div><\/div><div class=\"container sales-section workflow\"><div class=\"row\"><div class=\"col-md-12 text-center\"><div class=\"workflow-download-content row\"><div class=\"download\"><div class=\"col-sm-6 animBlock\" data-position=\"left\"><img src=\"https:\/\/assets.pixieset.com\/images\/site\/sales\/index-workflow-download-computer.jpg\" alt=\"Digital Delivery\" \/><\/div><\/div><\/div><\/div><div class=\"workflow-favorite-content\"><div class=\"col-sm-6 text-center favorite-text animBlock\" data-position=\"left\"><div class=\"list\"><ul><li><i class=\"fas fa-check\">\u00a0\u00a0 Track client favorites<\/i><\/li>\n<li><i class=\"fas fa-check\">\u00a0\u00a0 Export filenames<\/i><\/li>\n<\/ul><ul><li><i class=\"fas fa-check\">\u00a0\u00a0 Download client favorites<\/i><\/li>\n<li><i class=\"fas fa-check\">\u00a0\u00a0 Share favorites<\/i><\/li>\n<\/ul><\/div><\/div><div class=\"col-sm-6 animBlock\" data-position=\"right\"><img src=\"https:\/\/assets.pixieset.com\/images\/site\/sales\/index-workflow-favorite-computer.jpg\" alt=\"Favorite way to Proof\" \/><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:24:02+0000","updated_at":"2018-11-07T16:24:02+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"pixieset.com","preview_picture":"https:\/\/assets.pixieset.com\/images\/site\/sales\/sales-share.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":191,"title":"Hosting multiple websites with Apache2","url":"https:\/\/debian-administration.org\/article\/412\/Hosting_multiple_websites_with_Apache2","content":"<p>This site is <a href=\"https:\/\/debian-administration.org\/article\/730\/This_site_is_going_to_go_read-only\">now 100% read-only<\/a>, and retired.<\/p><div class=\"article\"><div class=\"tagline\"><p>Posted by <a href=\"https:\/\/debian-administration.org\/users\/Steve\">Steve<\/a> on Thu 6 Jul 2006 at 22:03<\/p><\/div><p>Tags: <a href=\"https:\/\/debian-administration.org\/tag\/apache2\" title=\"Find more content with this tag\" rel=\"tag\">apache2<\/a>, <a href=\"https:\/\/debian-administration.org\/tag\/virtual%20domains\" title=\"Find more content with this tag\" rel=\"tag\">virtual domains<\/a><\/p><p>One of the most common Apache2 questions I've seen on Debian mailing lists is from users who wonder how to host multiple websites with a single server. This is very straightforward, especially with the additional tools the Debian package provides.<\/p><p><a href=\"https:\/\/www.debian-administration.org\/articles\/207\">We've previously discussed<\/a> some of the tools which are included in the <a href=\"https:\/\/packages.debian.org\/cgi-bin\/search_packages.pl?searchon=names&amp;version=all&amp;exact=1&amp;keywords=apache2\">Apache2 package<\/a>, but what we didn't do was show they're used from start to finish.<\/p><p>There are many different ways you can configure Apache to host multiple sites, ranging from the simple to the complex. Here we're only going to cover the basics with the use of the <a href=\"https:\/\/httpd.apache.org\/docs\/2.0\/mod\/core.html#namevirtualhost\">NameVirtualHost<\/a> directive. The advantage of this approach is that you don't need to hard-wire any IP addresses, and it will just work<sup>tm<\/sup>. The only thing you need is for your domain names to resolve to the IP address of your webserver.<\/p><p>For example if you have an Apache server running upon the IP address <code>192.168.1.1<\/code> and you wish to host the three sites <code>example.com<\/code>, <code>example.net<\/code>, and <code>example.org<\/code> you'll need to make sure that these names resolve to the IP address of your server.<\/p><p>(This might mean that you need <code>example.com<\/code> <i>and<\/i> <code>www.example.com<\/code> to resolve to the same address. However that is a choice you'll need to make for yourself).<\/p><p>Since we'll be hosting multiple websites on the same host it makes a lot of sense to be very clear on the location of each sites files upon the filesystem. The way I suggest you manage this is to create a completely seperate document root, cgi-bin directory, and logfile directory for each host. You can place these beneath the standard Debian prefix of <code>\/var\/www<\/code> or you may use a completely different root - I use <code>\/home\/www<\/code>.<\/p><p>If you've not already done create the directories to contain your content, etc, as follows:<\/p><pre>\nroot@irony:~# mkdir \/home\/www\nroot@irony:~# mkdir \/home\/www\/www.example.com\nroot@irony:~# mkdir \/home\/www\/www.example.com\/htdocs\nroot@irony:~# mkdir \/home\/www\/www.example.com\/cgi-bin\nroot@irony:~# mkdir \/home\/www\/www.example.com\/logs\nroot@irony:~# mkdir \/home\/www\/www.example.net \nroot@irony:~# mkdir \/home\/www\/www.example.net\/htdocs\nroot@irony:~# mkdir \/home\/www\/www.example.net\/logs\nroot@irony:~# mkdir \/home\/www\/www.example.net\/cgi-bin\nroot@irony:~# mkdir \/home\/www\/www.example.org \nroot@irony:~# mkdir \/home\/www\/www.example.org\/htdocs\nroot@irony:~# mkdir \/home\/www\/www.example.org\/logs\nroot@irony:~# mkdir \/home\/www\/www.example.org\/cgi-bin\n<\/pre><p>Here we've setup three different directory trees, one for each site. If you wanted to have identical content it might make sense to only create one, and then use symbolic links instead.<\/p><p>The next thing to do is to enable virtual hosts in your Apache configuration. The simplest way to do this is to create a file called <code>\/etc\/apache2\/conf.d\/virtual.conf<\/code> and include the following content in it:<\/p><pre>\n#\n#  We're running multiple virtual hosts.\n#\nNameVirtualHost *\n<\/pre><p>(When Apache starts up it reads the contents of all files included in <code>\/etc\/apache2\/conf.d<\/code>, and files you create here won't get trashed on package upgrades.)<\/p><p>Once we've done this we can create the individual host configuration files. The Apache2 setup you'll find on Debian GNU\/Linux includes two directories for locating your site configuration files:<\/p><dl><dt><code>\/etc\/apache2\/sites-available<\/code><\/dt>\n<dd>\n<p>This contains configuration files for sites which are available but not necessarily enabled.<\/p>\n<\/dd>\n<dt><code>\/etc\/apache2\/sites-enabled<\/code><\/dt>\n<dd>\n<p>This directory contains site files which are enabled.<\/p>\n<\/dd>\n<\/dl><p>As with the <code>conf.d<\/code> directory each configuration file in the <code>sites-enabled<\/code> directory is loaded when the server starts - whilst the files in <code>sites-available<\/code> are completely ignored.<\/p><p>You are expected to create your host configuration files in <code>\/etc\/apache2\/sites-available<\/code>, then create a symbolic link to those files in the <code>sites-enabled<\/code> directory - this will cause them to be actually loaded\/read.<\/p><p>Rather than actually messing around with symbolic links the Debian package includes two utility commands <a href=\"https:\/\/www.debian-administration.org\/articles\/207\"><code>a2ensite<\/code> and <code>a2dissite<\/code><\/a> which will do the necessary work for you as we will demonstrate shortly.<\/p><p>Lets start with a real example. Create <code>\/etc\/apache2\/sites-available\/www.example.com<\/code> with the following contents:<\/p><pre>\n#\n#  Example.com (\/etc\/apache2\/sites-available\/www.example.com)\n#\n&lt;VirtualHost *&gt;\n        ServerAdmin webmaster@example.com\n        ServerName  www.example.com\n        ServerAlias example.com\n        # Indexes + Directory Root.\n        DirectoryIndex index.html\n        DocumentRoot \/home\/www\/www.example.com\/htdocs\/\n        # CGI Directory\n        ScriptAlias \/cgi-bin\/ \/home\/www\/www.example.com\/cgi-bin\/\n        &lt;Location \/cgi-bin&gt;\n                Options +ExecCGI\n        &lt;\/Location&gt;\n        # Logfiles\n        ErrorLog  \/home\/www\/www.example.com\/logs\/error.log\n        CustomLog \/home\/www\/www.example.com\/logs\/access.log combined\n&lt;\/VirtualHost&gt;\n<\/pre><p>Next create the file <code>www.example.net<\/code>:<\/p><pre>\n#\n#  Example.net (\/etc\/apache2\/sites-available\/www.example.net)\n#\n&lt;VirtualHost *&gt;\n        ServerAdmin webmaster@example.net\n        ServerName  www.example.net\n        ServerAlias example.net\n        # Indexes + Directory Root.\n        DirectoryIndex index.html\n        DocumentRoot \/home\/www\/www.example.net\/htdocs\/\n        # CGI Directory\n        ScriptAlias \/cgi-bin\/ \/home\/www\/www.example.net\/cgi-bin\/\n        &lt;Location \/cgi-bin&gt;\n                Options +ExecCGI\n        &lt;\/Location&gt;\n        # Logfiles\n        ErrorLog  \/home\/www\/www.example.net\/logs\/error.log\n        CustomLog \/home\/www\/www.example.net\/logs\/access.log combined\n&lt;\/VirtualHost&gt;\n<\/pre><p>Finally create the file <code>www.example.org<\/code>:<\/p><pre>\n#\n#  Example.org (\/etc\/apache2\/sites-available\/www.example.org)\n#\n&lt;VirtualHost *&gt;\n        ServerAdmin webmaster@example.org\n        ServerName  www.example.org\n        ServerAlias example.org\n        # Indexes + Directory Root.\n        DirectoryIndex index.html\n        DocumentRoot \/home\/www\/www.example.org\/htdocs\/\n        # CGI Directory\n        ScriptAlias \/cgi-bin\/ \/home\/www\/www.example.org\/cgi-bin\/\n        &lt;Location \/cgi-bin&gt;\n                Options +ExecCGI\n        &lt;\/Location&gt;\n        # Logfiles\n        ErrorLog  \/home\/www\/www.example.org\/logs\/error.log\n        CustomLog \/home\/www\/www.example.org\/logs\/access.log combined\n&lt;\/VirtualHost&gt;\n<\/pre><p>Now we've got:<\/p><ul><li>Three directories which can be used to contain our content.<\/li>\n<li>Three directories which can be used to contain our logfiles.<\/li>\n<li>Three directories which can be used to contain our dynamic CGI scripts.<\/li>\n<li>Three configuration files which are being ignored by Apache.<\/li>\n<\/ul><p>To enable the sites simply run:<\/p><pre>\nroot@irony:~# a2ensite www.example.com\nSite www.example.com installed; run \/etc\/init.d\/apache2 reload to enable.\nroot@irony:~# a2ensite www.example.net\nSite www.example.net installed; run \/etc\/init.d\/apache2 reload to enable.\nroot@irony:~# a2ensite www.example.org\nSite www.example.org installed; run \/etc\/init.d\/apache2 reload to enable.\n<\/pre><p>This will now create the symbolic links so that <code>\/etc\/apache2\/sites-enabled\/www.example.org<\/code>, etc, now exist and will be read.<\/p><p>Once we've finished our setup we can restart, or reload, the webserver as the output above instructed us to do with:<\/p><pre>\nroot@irony:~# \/etc\/init.d\/apache2 reload\nReloading web server config...done.\nroot@irony:~# \n<\/pre><\/div><hr \/>","created_at":"2018-11-07T16:24:02+0000","updated_at":"2018-11-07T16:24:02+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"debian-administration.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":186,"title":"How to get initial customers\/users for a new startup","url":"https:\/\/www.quora.com\/How-do-I-get-initial-customers-users-for-a-new-startup","content":"<div class=\"header\"><div id=\"wHMFRYvA1\"><div class=\"QuestionArea\"><div id=\"w2Onw3D51\"><div id=\"w2Onw3D53\"><div class=\"TopicList QuestionTopicHorizontalList\" id=\"__w2_w2Onw3D54_topics_list\"><div id=\"w2Onw3D515\"><div class=\"QuestionTopicListItems TopicListItems2 Toggle QuestionTopicHorizontalListItems SimpleToggle\" id=\"__w2_w2Onw3D516__truncated\"><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_w2Onw3D518_topic_list_item\"><div class=\"u-inline-block\"><div id=\"w2Onw3D525\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/Customers\" id=\"__w2_w2Onw3D526_link\">Customers<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_w2Onw3D520_topic_list_item\"><div class=\"u-inline-block\"><div id=\"w2Onw3D527\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/Startups\" id=\"__w2_w2Onw3D528_link\">Startups<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_w2Onw3D522_topic_list_item\"><div class=\"u-inline-block\"><div id=\"w2Onw3D529\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/E-Commerce\" id=\"__w2_w2Onw3D530_link\">E-Commerce<\/a><\/div><\/div><\/div><div class=\"TopicListItem QuestionTopicListItem topic_pill\" id=\"__w2_w2Onw3D524_topic_list_item\"><div class=\"u-inline-block\"><div id=\"w2Onw3D531\"><a class=\"TopicNameLink HoverMenu topic_name\" href=\"https:\/\/www.quora.com\/topic\/Startup-Strategy-2\" id=\"__w2_w2Onw3D532_link\">Startup Strategy<\/a><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"w2Onw3D55\"><div class=\"question_text_edit\"><h1>How do I get initial customers\/users for a new startup?<\/h1><\/div><\/div><div id=\"w2Onw3D57\"><\/div><\/div><\/div><div class=\"Answer AnswerBase\" data-clog-metadata=\"{}\" id=\"__w2_wAojDlxL42_answer\"><div><div id=\"wAojDlxL43\"><div class=\"ContentHeader AnswerHeader\"><div class=\"ui_layout_photo_text u-flex ui_layout_size--small\"><div class=\"ui_layout_photo_wrapper u-flex-none\"><div class=\"ui_layout_photo u-relative\"><div id=\"wAojDlxL52\"><\/div><\/div><div class=\"ui_layout_text u-flex-auto u-flex-align-self--center\"><div class=\"u-flex u-flex-justify--between\"><div class=\"u-margin-right--sm\"><div class=\"feed_item_answer_user\"><a class=\"user\" href=\"https:\/\/www.quora.com\/profile\/Art-Zabalov\" id=\"__w2_wAojDlxL65_name_link\">Art Zabalov<\/a>, Head of Marketing at ChatPirate.com<\/div><div class=\"u-inline-block\" id=\"wAojDlxL47\"><div class=\"CredibilityFacts pass_color_to_child_links\"><a class=\"answer_permalink\" href=\"https:\/\/www.quora.com\/How-do-I-get-initial-customers-users-for-a-new-startup\/answer\/Art-Zabalov\" id=\"__w2_wAojDlxL77_link\">Answered 194w ago<\/a><\/div><\/div><\/div><div id=\"__w2_wAojDlxL45_follow_button_section\"><div class=\"u-relative icon_follow_button_wrapper\"><div class=\"u-absolute u-absolute--center u-border-radius--ellipse u-padding-right--xs u-padding-left--xs icon_follow_button\" id=\"__w2_wAojDlxL45_follow_button\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><div id=\"__w2_wAojDlxL42_answer_content\"><div id=\"__w2_wAojDlxL46_expanded_content\"><div id=\"wAojDlxL49\"><div class=\"TranslatedAnswerBanner\"><div class=\"u-serif-font-main--large\"><div class=\"ui_qtext_expanded\">\"Consumer Web Application\" is not much to go on, but I will try my best.<br \/>PPC, SEO, Email, Content and Social are a given.<br \/><b>1. Customer Feedback. <\/b><br \/>You've probably heard of Slack. Well, it's Founder, Stewart Butterfield, has recently shared their epic launch strategy (<a href=\"http:\/\/firstround.com\/article\/From-0-to-1B-Slacks-Founder-Shares-Their-Epic-Launch-Strategy\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"firstround.com\">From 0 to $1B - Slack's Founder Shares Their Epic Launch Strategy<\/a>), where he outlined that pre-launch, customer feedback was the epicenter of their efforts. <br \/>Takeaway: First offer your product to anyone in your target audience who is willing to take it(i'm talking personal outreach), in return for feedback, then make the necessary changes.\u00a0 In my opinion, this is the most critical part of being successful - delivering a product your customers will actually want. These first testers might just become your first paying users. <br \/>In fact, feedback doesn't have to come from personal outreach only. As your application is web based, you can get your feedback from online visitors. Here's how: <a href=\"http:\/\/customericare.com\/feedback-before-ab-testing-on-landing-pages\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"customericare.com\">The Best Way To Find Out If Your Landing Page Sucks (Before Losing Potential Customers)<\/a><br \/><b>2. Referrals.<\/b><br \/>Start with personally asking your contacts to refer you to a friend, either of their own free will, and kindness of the heart, or for a small incentive. Further down the line you could even launch a referral campaign, like Harry's\u00a0 <a href=\"http:\/\/fourhourworkweek.com\/2014\/07\/21\/harrys-prelaunchr-email\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"fourhourworkweek.com\">How to Gather 100,000 Emails in One Week (Includes Successful Templates, Code, Everything You Need)<\/a><br \/><b>3. Partnerships.<\/b><br \/>Strategic partnerships have helped numerous billion-dollar giants. One of them is Square, who used a partnership with Apple to build its own client base. Read more here: <a href=\"https:\/\/growthhackers.com\/companies\/square\/\" rel=\"noopener nofollow\" target=\"_blank\" class=\"external_link\" data-qt-tooltip=\"growthhackers.com\">Square - How Did Square Grow So Quickly? - GrowthHackers<\/a><br \/><b>4. Check out Quora's<\/b> <a href=\"https:\/\/www.quora.com\/topic\/How-Did-X-Get-Traction-3\">How Did X Get Traction<\/a> thread. Lots of interesting insight there.<\/div><\/div><div id=\"wAojDlxL57\"><div id=\"__w2_wAojDlxL46_survey_wrapper\"><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:24:01+0000","updated_at":"2018-11-07T16:24:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.quora.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":187,"title":"Juicy SEO tools you will love - Mangools","url":"https:\/\/mangools.com\/","content":"<section class=\"mg__hero bg-grey\"><div class=\"uk-container uk-container-center uk-text-center\"><p class=\"font-21\">One account with access to a bundle of juicy SEO tools.<\/p><a href=\"#discover\" class=\"btn is--green is--rounded uk-text-bold\" data-uk-smooth-scroll=\"{offset: 90}\">\n            Explore our tools\n        <\/a><\/div><\/section><div class=\"bg-fff uk-position-relative uk-clearfix bg-icons-repeat\"><div class=\"mg__margin-t-45\"><div class=\"uk-container uk-container-center\" id=\"package\"><h2 class=\"font-36 font-small-24 uk-text-bold uk-text-center\">What's in the package<\/h2><div id=\"discover\" class=\"mg__margin-t-60 uk-visible-large\"><div class=\"uk-grid uk-grid-collapse uk-flex uk-flex-middle mg__package\"><div class=\"uk-width-large-4-10 uk-text-right\"><div class=\"mg__margin-b-45 mg__padding-0-30 kwfinder\"><h3 class=\"font-24 uk-margin-remove uk-text-bold uk-flex-inline uk-flex-middle\">Keyword research<\/h3><p class=\"color-grey\">Find hundreds of long tail keywords with low SEO\u00a0difficulty for your website.<\/p><\/div><div class=\"mg__margin-b-45 mg__padding-0-30 serpchecker\"><h3 class=\"font-24 uk-margin-remove uk-text-bold uk-flex-inline uk-flex-middle\">SERP analysis<\/h3><p class=\"color-grey\">Get local Google search results, analyze SERP features and scan your competitors.<\/p><\/div><\/div><div class=\"uk-width-large-2-10 uk-text-center animated-once-slower\"><img alt=\"mangools\" width=\"80\" src=\"https:\/\/mangools.com\/assets\/app-logos\/mangools-icon-db2b3889a1797af9bf0af8adc757cc43954bc6ba835b56af648c1410fce8c13e.svg\" \/><\/div><div class=\"uk-width-large-4-10\"><div class=\"mg__margin-b-45 mg__padding-0-30 serpwatcher\"><h3 class=\"font-24 uk-margin-remove uk-text-bold uk-flex-inline uk-flex-middle\">Rank tracking<\/h3><p class=\"color-grey\">Track the overall progress of your organic search positions with one key metric.<\/p><\/div><div class=\"mg__margin-b-45 mg__padding-0-30 siteprofiler\"><h3 class=\"font-24 uk-margin-remove uk-text-bold uk-flex-inline uk-flex-middle\">Website analysis <strong class=\"mg__label is--small is--orange font-10 mg__margin-l-5\">NEW<\/strong><\/h3><p class=\"color-grey\">Inspect the overall online presence of any website with the most important SEO\u00a0metrics &amp; insights.<\/p><\/div><div class=\"mg__margin-b-45 mg__padding-0-30 extension\"><h3 class=\"font-24 uk-margin-remove uk-text-bold uk-flex-inline uk-flex-middle\">Browser SEO extension <strong class=\"mg__label is--small is--orange font-10 mg__margin-l-5\">NEW<\/strong><\/h3><p class=\"color-grey\">See the essential data from Mangools SEO tools and useful on-page insights in one click.<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><section class=\"mg__padding-90-0 bg-lightgrey mg__testimonials mg__border-t mg__border-b uk-position-relative\" id=\"testimonials\"><div class=\"uk-container uk-container-center\"><h2 class=\"font-36 uk-text-bold mg__margin-b-45 uk-text-center\">What people are saying about us<\/h2><\/div><div class=\"mg__testimonial-logos uk-text-center\"><div class=\"uk-container uk-container-center mg__padding-45-0\"><h2 class=\"section__title mg__margin-b-30\"><strong>Join our happy customers<\/strong><\/h2><\/div><\/div><div class=\"bg-fff mg__padding-30-0\"><div class=\"uk-container uk-container-center\"><div class=\"uk-grid uk-text-center spacing-init\"><p><strong>4.8<\/strong>\/5\n                    <a href=\"https:\/\/woorkup.com\/best-keyword-research-tool\/\" class=\"font-18 mg__margin-r-10\" target=\"_blank\">woorkup.com<\/a>\n                    <\/p><p><strong>4.0<\/strong>\/5\n                    <a href=\"https:\/\/www.pcmag.com\/review\/348802\/kwfinder-com\" class=\"font-18 mg__margin-r-10\" target=\"_blank\">pcmag.com<\/a>\n                    <\/p><p><strong>4.8<\/strong>\/5\n                    <a href=\"https:\/\/www.authorityhacker.com\/kwfinder-review\/\" class=\"font-18 mg__margin-r-10\" target=\"_blank\">authorityhacker.com<\/a>\n                    <\/p><\/div><\/div><\/div><p><a href=\"https:\/\/mangools.com\/testimonials\" class=\"btn is--ghost is--small\">More testimonials<\/a><\/p><\/section><section class=\"mg__padding-90-0 bg-lightblue uk-text-center mg__border-b\" id=\"plans-and-pricing\"><header class=\"uk-container uk-container-center\"><h2 class=\"font-36 uk-text-bold uk-margin-remove\"><strong>\n                Plans &amp; pricing\n        <\/strong><\/h2>\n    <\/header><div class=\"mg__pricing-table uk-container uk-container-center\"><div class=\"uk-grid uk-grid-small uk-display-block\"><div class=\"mg__flipper-container uk-width-large-1-3 uk-width-medium-1-1\"><div class=\"mg__flipper\"><div class=\"mg__plan mg__flip-front mg__padding-60-0 bg-lightgrey\"><img src=\"https:\/\/mangools.com\/assets\/subscription-basic-7ba2486b1f7e39c1a2213a23a01f36e86f12d71d50cc3fc83f3f259a3205c443.svg\" alt=\"Subscription basic\" \/><h3 class=\"font-24\">\n                Mangools\n                <strong class=\"color-orange\">Basic<\/strong>\n              <\/h3><strong class=\"font-36\">\n                  $2990\n                <\/strong><p>per month<\/p><p class=\"font-16\">\n                    <strong class=\"mg__label is--grey\">Save $229.20 per year<\/strong>\n                <\/p><p class=\"font-16 color-grey\">$358.80 billed upon purchase\n                <\/p><a class=\"btn is--green uk-text-bold\" data-no-turbolink=\"true\" data-turboliks=\"false\" data-disable-with=\"Processing...\" href=\"https:\/\/mangools.com\/subscriptions\/checkout?currency=usd&amp;plan_id=56991d5e8ce44af72b70bd75\">Subscribe now<\/a><\/div><div class=\"mg__plan mg__flip-back mg__padding-60-0 bg-lightgrey\"><img src=\"https:\/\/mangools.com\/assets\/subscription-basic-7ba2486b1f7e39c1a2213a23a01f36e86f12d71d50cc3fc83f3f259a3205c443.svg\" alt=\"Subscription basic\" \/><h3 class=\"font-24\">\n                Mangools\n                <strong class=\"color-orange\">Basic<\/strong>\n              <\/h3><strong class=\"font-36\">\n                  $4900\n                <\/strong><p>per month<\/p><a class=\"btn is--green uk-text-bold\" data-no-turbolink=\"true\" data-turboliks=\"false\" data-disable-with=\"Processing...\" href=\"https:\/\/mangools.com\/subscriptions\/checkout?currency=usd&amp;plan_id=5698f9028ce44af72b70bd6a\">Subscribe now<\/a><\/div><\/div><\/div><div class=\"mg__flipper-container uk-width-large-1-3 uk-width-medium-1-1\"><div class=\"mg__flipper\"><div class=\"mg__plan mg__flip-front mg__padding-60-0 bg-lightgrey\"><img src=\"https:\/\/mangools.com\/assets\/subscription-premium-8817d0eae9c839bae5634ad04b15582c80b8af1cfc8728ab82cd888a273e1f2f.svg\" alt=\"Subscription premium\" \/><h3 class=\"font-24\">\n                Mangools\n                <strong class=\"color-orange\">Premium<\/strong>\n              <\/h3><strong class=\"font-36\">\n                  $3990\n                <\/strong><p>per month<\/p><p class=\"font-16\">\n                    <strong class=\"mg__label is--orange\">Save $349.20 per year<\/strong>\n                <\/p><p class=\"font-16 color-grey\">$478.80 billed upon purchase\n                <\/p><a class=\"btn is--green uk-text-bold\" data-no-turbolink=\"true\" data-turboliks=\"false\" data-disable-with=\"Processing...\" href=\"https:\/\/mangools.com\/subscriptions\/checkout?currency=usd&amp;plan_id=56991d458ce44af72b70bd74\">Subscribe now<\/a><\/div><div class=\"mg__plan mg__flip-back mg__padding-60-0 bg-lightgrey\"><img src=\"https:\/\/mangools.com\/assets\/subscription-premium-8817d0eae9c839bae5634ad04b15582c80b8af1cfc8728ab82cd888a273e1f2f.svg\" alt=\"Subscription premium\" \/><h3 class=\"font-24\">\n                Mangools\n                <strong class=\"color-orange\">Premium<\/strong>\n              <\/h3><strong class=\"font-36\">\n                  $6900\n                <\/strong><p>per month<\/p><a class=\"btn is--green uk-text-bold\" data-no-turbolink=\"true\" data-turboliks=\"false\" data-disable-with=\"Processing...\" href=\"https:\/\/mangools.com\/subscriptions\/checkout?currency=usd&amp;plan_id=5698f9418ce44af72b70bd6b\">Subscribe now<\/a><\/div><\/div><\/div><div class=\"mg__flipper-container uk-width-large-1-3 uk-width-medium-1-1\"><div class=\"mg__flipper\"><div class=\"mg__plan mg__flip-front mg__padding-60-0 bg-lightgrey\"><img src=\"https:\/\/mangools.com\/assets\/subscription-agency-0589eaef1a72fb8d30d8737034edfef0c545da084302970f1d3c405194e29fbb.svg\" alt=\"Subscription agency\" \/><h3 class=\"font-24\">\n                Mangools\n                <strong class=\"color-orange\">Agency<\/strong>\n              <\/h3><strong class=\"font-36\">\n                  $7990\n                <\/strong><p>per month<\/p><p class=\"font-16\">\n                    <strong class=\"mg__label is--grey\">Save $589.20 per year<\/strong>\n                <\/p><p class=\"font-16 color-grey\">$958.80 billed upon purchase\n                <\/p><a class=\"btn is--green uk-text-bold\" data-no-turbolink=\"true\" data-turboliks=\"false\" data-disable-with=\"Processing...\" href=\"https:\/\/mangools.com\/subscriptions\/checkout?currency=usd&amp;plan_id=597b0c904b13940c1fd9d9a7\">Subscribe now<\/a><\/div><div class=\"mg__plan mg__flip-back mg__padding-60-0 bg-lightgrey\"><img src=\"https:\/\/mangools.com\/assets\/subscription-agency-0589eaef1a72fb8d30d8737034edfef0c545da084302970f1d3c405194e29fbb.svg\" alt=\"Subscription agency\" \/><h3 class=\"font-24\">\n                Mangools\n                <strong class=\"color-orange\">Agency<\/strong>\n              <\/h3><strong class=\"font-36\">\n                  $12900\n                <\/strong><p>per month<\/p><a class=\"btn is--green uk-text-bold\" data-no-turbolink=\"true\" data-turboliks=\"false\" data-disable-with=\"Processing...\" href=\"https:\/\/mangools.com\/subscriptions\/checkout?currency=usd&amp;plan_id=597b0df74b13940c1fd9d9a8\">Subscribe now<\/a><\/div><\/div><\/div><div class=\"uk-text-center uk-width-1-1\"><div class=\"uk-grid mg__margin-b-30\"><div class=\"uk-width-medium-1-2 uk-text-left uk-text-center-small color-grey\" id=\"currency-switcher\">Currency:\n            USD<a href=\"https:\/\/mangools.com\/?currency=eur#plans-and-pricing\" class=\"uk-link uk-padding-remove active\">EUR<\/a><\/div><div class=\"uk-width-medium-1-2 uk-text-right uk-text-center-small\"><p class=\"color-grey font-14\">* limit is shared with SERP lookups in KWFinder<\/p><\/div><\/div><\/div><\/div><div class=\"uk-grid uk-flex uk-flex-center uk-flex-wrap mg__margin-t-30\" id=\"pricing-more\"><div class=\"uk-width-medium-1-2\"><div class=\"bg-lightgrey mg__padding-45\"><h3 class=\"font-24 uk-text-bold\">Mangools FREE account<\/h3><p>\n              Get 5 lookups per 24 hours, 50 keyword suggestions per search and 100 backlinks per 24 hours.\n              <strong>No credit card needed!<\/strong>\n            <\/p><a class=\"btn is--blue is--small uk-text-bold\" href=\"https:\/\/mangools.com\/users\/sign_up?ref=pricing-mngls\">Register now<\/a><\/div><\/div><div class=\"uk-width-medium-1-2\"><div class=\"mg__padding-45\"><h3 class=\"font-24 uk-text-bold\">Need something special?<\/h3><p>\n            Feel free to contact our support with your query.<br \/>We will find a solution that suits your needs.\n          <\/p><a class=\"btn is--ghost is--small uk-text-bold\" href=\"#contact-form\" data-uk-modal=\"\">Contact us<\/a><\/div><\/div><\/div><\/div><\/section><div class=\"mg__padding-90-0 uk-text-center bg-fff\" id=\"more\"><section class=\"mg__promotions\"><h2 class=\"font-36 uk-text-bold mg__margin-b-45 uk-text-bold\">Get even more!<\/h2><\/section><section class=\"mg__testimonial-logos\"><div class=\"uk-container uk-container-center\"><h2 class=\"font-24 uk-text-bold mg__margin-b-30 uk-text-bold\">We've been featured on<\/h2><\/div><\/section><\/div>","created_at":"2018-11-07T16:24:01+0000","updated_at":"2018-11-07T16:24:01+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"mangools.com","preview_picture":"https:\/\/mangools.com\/assets\/mangools-facebook-72268ff1182711566d7a2c64d1896f976600b285b5b9ce030673c114c656d729.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":188,"title":"Emacs - Dev - HowTo: Detect when a frame's size changes","url":"http:\/\/emacs.1067599.n8.nabble.com\/HowTo-Detect-when-a-frame-s-size-changes-td424155.html","content":"Thanks Martin!<p>frame-size-changed-p would be nice --- Basically I was writing myself&#13;\n<br \/>some code in Emacspeak so it would cue me via sounds to the position&#13;\n<br \/>of the current window relative to the overall layout.&#13;\n<\/p><p>I wrote myself this function: -- and was looking to see if I could&#13;\n<br \/>avoid some of the computes by cacheing values --- though for now it&#13;\n<br \/>doesn't look like the code I have is sufficiently inefficient to&#13;\n<br \/>deserve cacheing.&#13;\n<\/p><p>(sox-multiwindow ...) used below uses SoX to produce a short tone, and&#13;\n<br \/>the tone's pitch and left\/right pan get paramaeterized \u00a0appropriately &#13;\n<\/p><p>(defun emacspeak--sox-multiwindow (corners)&#13;\n<br \/>\u00a0 \"Takes `window-edges' and plays a sound cue based on position of current window with respect to&#13;\n<br \/>the overall window layout.\"&#13;\n<br \/>\u00a0 (let&#13;\n<br \/>\u00a0 \u00a0 \u00a0 ((tr 0)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 \u00a0(mr (\/ (frame-height) 2))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 \u00a0(br (1- (frame-height)))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 \u00a0(lc 0)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 \u00a0(mc (\/ (frame-width) 2))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 \u00a0(rc (frame-width)))&#13;\n<br \/>\u00a0 \u00a0 (cond&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,lc ,tr ,mc ,br))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'left-half)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,mc ,tr ,rc, br))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow 'swap)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'right-half)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,lc ,tr ,rc ,mr))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow \u00a0nil 2)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'top-half)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,lc ,mr ,rc ,br))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow nil 1.3)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'bottom-half)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,lc ,tr ,mc ,mr)) &#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow nil 2.5)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'top-left)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,mc ,tr ,rc ,mr))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow t 2.5)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'top-right)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,lc ,mr ,mc ,br))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow nil 0.9)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'bottom-left)&#13;\n<br \/>\u00a0 \u00a0 \u00a0((equal corners `(,mc ,mr ,rc ,br))&#13;\n<br \/>\u00a0 \u00a0 \u00a0 (sox-multiwindow 'swap 0.9)&#13;\n<br \/>\u00a0 \u00a0 \u00a0 'bottom-right)&#13;\n<br \/>\u00a0 \u00a0 \u00a0(t \"\"))))&#13;\n<\/p><p>martin rudalics writes:&#13;\n<br \/>\u00a0&gt; \u00a0&gt; Thanks! will try that&#13;\n<br \/>\u00a0&gt; &#13;\n<br \/>\u00a0&gt; Hmmm... rather not. \u00a0You will still get false positives when the&#13;\n<br \/>\u00a0&gt; minibuffer window is resized. \u00a0For the moment stick to my earlier&#13;\n<br \/>\u00a0&gt; proposal: When you run your \u2018window-size-change-functions\u2019 function&#13;\n<br \/>\u00a0&gt; remember the old values of \u2018frame-pixel-width\u2019 and \u2018frame-pixel-height\u2019&#13;\n<br \/>\u00a0&gt; (maybe in a frame parameter) and the next time you run it compare them&#13;\n<br \/>\u00a0&gt; with the current values.&#13;\n<br \/>\u00a0&gt; &#13;\n<br \/>\u00a0&gt; I will fix that soon probably by providing a \u2018frame-size-changed-p\u2019&#13;\n<br \/>\u00a0&gt; function that you can use during \u2018window-size-change-functions\u2019.&#13;\n<br \/>\u00a0&gt; Providing a separate \u2018frame-size-change-functions\u2019 hook would be also&#13;\n<br \/>\u00a0&gt; possible but the interaction of these could be tricky.&#13;\n<br \/>\u00a0&gt; &#13;\n<br \/>\u00a0&gt; martin&#13;\n<\/p><p>-- &#13;\n<\/p><p>--&#13;\n<\/p>","created_at":"2018-11-07T16:24:01+0000","updated_at":"2018-11-07T16:24:01+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"emacs.1067599.n8.nabble.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":189,"title":"Productivity","url":"https:\/\/blog.samaltman.com\/productivity","content":"<p>I think I am\nat least somewhat more productive than average, and people sometimes ask me for\nproductivity tips. \u00a0So I decided to just write them all down in one place.<\/p><p>Compound\ngrowth gets discussed as a financial concept, but it works in careers as well,\nand it is magic.\u00a0 A small productivity\ngain, compounded over 50 years, is worth a lot.\u00a0\nSo it\u2019s worth figuring out how to optimize productivity. If you get 10%\nmore done and 1% better every day compared to someone else, the compounded\ndifference is massive.\u00a0<\/p><p>WHAT YOU WORK\nON<\/p><p>It doesn\u2019t\nmatter how fast you move if it\u2019s in a worthless direction.\u00a0 Picking the right thing to work on is the\nmost important element of productivity and usually almost ignored.\u00a0 So think about it more!\u00a0 Independent thought is hard but it\u2019s\nsomething you can get better at with practice.<\/p><p>The most\nimpressive people I know have strong beliefs about the world, which is rare in\nthe general population. \u00a0If you find yourself always agreeing with whomever\nyou last spoke with, that\u2019s bad.\u00a0 You\nwill of course be wrong sometimes, but develop the confidence to stick with\nyour convictions.\u00a0 It will let you be\ncourageous when you\u2019re right about something important that most people don\u2019t\nsee.<\/p><p>I make sure to\nleave enough time in my schedule to think about what to work on. \u00a0The best\nways for me to do this are reading books, hanging out with interesting people,\nand spending time in nature.<\/p><p>I\u2019ve learned\nthat I can\u2019t be very productive working on things I don\u2019t care about or don\u2019t\nlike. \u00a0So I just try not to put myself in a position where I have to do\nthem (by delegating, avoiding, or something else). \u00a0Stuff that you don\u2019t\nlike is a painful drag on morale and momentum.<\/p><p>By the way,\nhere is an important lesson about delegation: remember that everyone else is also\nmost productive when they\u2019re doing what they like, and do what you\u2019d want other\npeople to do for you\u2014try to figure out who likes (and is good at) doing what,\nand delegate that way. \u00a0<\/p><p>If you find\nyourself not liking what you\u2019re doing for a long period of time, seriously\nconsider a major job change. \u00a0Short-term burnout happens, but if it isn\u2019t\nresolved with some time off, maybe it\u2019s time to do something you\u2019re more\ninterested in.\u00a0<\/p><p>I\u2019ve been very\nfortunate to find work I like so much I\u2019d do it for free, which makes it easy\nto be really productive.<\/p><p>It\u2019s important\nto learn that you can learn anything you want, and that you can get better\nquickly. \u00a0This feels like an unlikely miracle the first few times it\nhappens, but eventually you learn to trust that you can do it.<\/p><p>Doing great\nwork usually requires colleagues of some sort. \u00a0Try to be around smart,\nproductive, happy, and positive people that don\u2019t belittle your ambitions.\u00a0 I love being around people who push me and\ninspire me to be better.\u00a0 To the degree\nyou able to, avoid the opposite kind of people\u2014the cost of letting them take up\nyour mental cycles is horrific.\u00a0<\/p><p>You have to\nboth pick the right problem and do the work. \u00a0There aren\u2019t many\nshortcuts.\u00a0 If you\u2019re going to do\nsomething really important, you are very likely going to work both smart and\nhard. \u00a0The biggest prizes are heavily competed for.\u00a0 This isn\u2019t true in every field (there are\ngreat mathematicians who never spend that many hours a week working) but it is\nin most.<\/p><p>PRIORITIZATION<\/p><p>My system has\nthree key pillars: \u201cMake sure to get the important shit done\u201d, \u201cDon\u2019t waste\ntime on stupid shit\u201d, and \u201cmake a lot of lists\u201d.<\/p><p>I highly\nrecommend using lists. \u00a0I make lists of what I want to accomplish each\nyear, each month, and each day. \u00a0Lists are very focusing, and they help me\nwith multitasking because I don\u2019t have to keep as much in my head.\u00a0 If I\u2019m not in the mood for some particular\ntask, I can always find something else I\u2019m excited to do.<\/p><p>I prefer lists\nwritten down on paper.\u00a0 It\u2019s easy to add\nand remove tasks.\u00a0 I can access them\nduring meetings without feeling rude.\u00a0 I\nre-transcribe lists frequently, which forces me to think about everything on the\nlist and gives me an opportunity to add and remove items.<\/p><p>I don\u2019t bother\nwith categorization or trying to size tasks or anything like that (the most I\ndo is put a star next to really important items). \u00a0<\/p><p>I try to\nprioritize in a way that generates momentum. \u00a0The more I get done, the\nbetter I feel, and then the more I get done.\u00a0\nI like to start and end each day with something I can really make\nprogress on.<\/p><p>I am\nrelentless about getting my most important projects done\u2014I\u2019ve found that if I\nreally want something to happen and I push hard enough, it usually happens.\u00a0<\/p><p>I try to be\nruthless about saying no to stuff, and doing non-critical things in the\nquickest way possible. \u00a0I probably take this too far\u2014for example, I am\nalmost sure I am terse to the point of rudeness when replying to emails.<\/p><p>I generally\ntry to avoid meetings and conferences as I find the time cost to be huge\u2014I get\nthe most value out of time in my office. \u00a0However, it is critical that you\nkeep enough space in your schedule to allow for chance encounters and exposure\nto new people and ideas. \u00a0Having an open network is valuable; though\nprobably 90% of the random meetings I take are a waste of time, the other 10%\nreally make up for it.<\/p><p>I find most meetings are best scheduled for 15-20 minutes, or 2 hours. \u00a0The default of 1 hour is usually wrong, and leads to a lot of wasted time.<\/p><p>I have\ndifferent times of day I try to use for different kinds of work. \u00a0The\nfirst few hours of the morning are definitely my most productive time of the\nday, so I don\u2019t let anyone schedule anything then. \u00a0I try to do meetings\nin the afternoon.\u00a0 I take a break, or\nswitch tasks, whenever I feel my attention starting to fade.\u00a0<\/p><p>I don\u2019t think\nmost people value their time enough\u2014I am surprised by the number of people I\nknow who make $100 an hour and yet will spend a couple of hours doing something\nthey don\u2019t want to do to save $20.<\/p><p>Also, don\u2019t\nfall into the trap of productivity porn\u2014chasing productivity for its own sake\nisn\u2019t helpful. \u00a0Many people spend too much time thinking about how to perfectly\noptimize their system, and not nearly enough asking if they\u2019re working on the\nright problems. \u00a0It doesn\u2019t matter what system you use or if you squeeze\nout every second if you\u2019re working on the wrong thing.<\/p><p>The right goal\nis to allocate your year optimally, not your day.<\/p><p>PHYSICAL\nFACTORS<\/p><p>Very likely\nwhat is optimal for me won\u2019t be optimal for you. \u00a0You\u2019ll have to\nexperiment to find out what works best for your body.\u00a0 It\u2019s definitely worth doing\u2014it helps in all\naspects of life, and you\u2019ll feel a lot better and happier overall.<\/p><p>It probably\ntook a little bit of my time every week for a few years to arrive at what works\nbest for me, but my sense is if I do a good job at all the below I\u2019m at least\n1.5x more productive than if not.<\/p><p>Sleep seems to\nbe the most important physical factor in productivity for me. \u00a0Some sort\nof sleep tracker to figure out how to sleep best is helpful. \u00a0I\u2019ve found the only\nthing I\u2019m consistent with are in the set-it-and-forget-it category, and I\nreally like the <a href=\"https:\/\/www.amazon.com\/gp\/product\/B0158W3E2A\/\">Emfit QS+Active<\/a>.<\/p><p>I like a cold,\ndark, quiet room, and a great mattress (I resisted spending a bunch of money on\na great mattress for years, which was stupid\u2014it makes a huge difference to my\nsleep quality. \u00a0I love <a href=\"https:\/\/www.tempurpedic.com\/shop-mattresses\/tempur-contour-elite\/v\/715\/\" title=\"Link: https:\/\/www.tempurpedic.com\/shop-mattresses\/tempur-contour-elite\/v\/715\/\">this one<\/a>). \u00a0Not eating a lot in the few hours before sleep\nhelps.\u00a0 Not drinking alcohol helps a lot, though\nI\u2019m not willing to do that all the time.<\/p><p>I use a <a href=\"https:\/\/www.chilitechnology.com\">Chili\nPad<\/a> to be cold while I sleep if I can\u2019t get the room cold enough, which is\ngreat but loud (I set it up to have the cooler unit outside my room).<\/p><p>When traveling,\nI use an <a href=\"https:\/\/www.amazon.com\/gp\/product\/B00GSO1D9O\/\">eye mask<\/a>\u00a0and <a href=\"https:\/\/www.amazon.com\/gp\/product\/B001J4HB1C\">ear plugs<\/a>.<\/p><p>This is likely\nto be controversial, but I take a low dose of sleeping pills (like a third of a\nnormal dose) or a very low dose of cannabis whenever I can\u2019t sleep. \u00a0I am\na bad sleeper in general, and a particularly bad sleeper when I travel.\u00a0 It likely has tradeoffs,\nbut so does not sleeping well.\u00a0 If you\ncan already sleep well, I wouldn\u2019t recommend this.<\/p><p>I use a full\nspectrum LED light most mornings for about 10-15 minutes while I catch up on\nemail. \u00a0It\u2019s great\u2014if you try nothing else in here, this is the thing I\u2019d\ntry.\u00a0 It\u2019s a ridiculous gain for me.\n\u00a0I like <a href=\"https:\/\/www.amazon.com\/gp\/product\/B075H39NDL\/\">this one<\/a>, and it\u2019s easy to travel with.<\/p><p>Exercise is\nprobably the second most important physical factor. \u00a0I tried a number of\ndifferent exercise programs for a few months each and the one that seemed best\nwas lifting heavy weights 3x a week for an hour, and high intensity interval\ntraining occasionally. \u00a0In addition to productivity gains, this is also\nthe exercise program that makes me feel the best overall. \u00a0<\/p><p>The third area\nis nutrition. \u00a0I very rarely eat breakfast, so I get about 15 hours of\nfasting most days (except an espresso when I wake up). \u00a0I know this is contrary\nto most advice, and I suspect it\u2019s not optimal for most people, but it\ndefinitely works well for me.<\/p><p>Eating lots of\nsugar is the thing that makes me feel the worst and that I try hardest to\navoid. \u00a0I also try to avoid foods that aggravate my digestion or spike up\ninflammation (for example, very spicy foods). \u00a0I don\u2019t have much willpower when it comes to\nsweet things, so I mostly just try to keep junk food out of the house.<\/p><p>I have one big shot of espresso immediately when I wake up and one after lunch. \u00a0I assume this is\nabout 200mg total of caffeine per day.\u00a0 I\ntried a few other configurations; this was the one that worked by far the best.\n\u00a0I otherwise aggressively avoid stimulants, but I will have more coffee if\nI\u2019m super tired and really need to get something done.<\/p><p>I\u2019m\nvegetarian and have been since I was a kid, and I supplement methyl B-12, Omega-3, Iron, and Vitamin D-3. \u00a0I\ngot to this list with a year or so of quarterly blood tests; it\u2019s worked for me\never since (I re-test maybe every year and a half or so). \u00a0There are many\ndoctors who will happily work with you on a super comprehensive blood test (and\nservices like <a href=\"https:\/\/www.wellnessfx.com\">WellnessFX<\/a>).\u00a0 I also go out\nof my way to drink a lot of protein shakes, which I hate and I wouldn\u2019t do if I\nweren\u2019t vegetarian.<\/p><p>OTHER STUFF<\/p><p>Here\u2019s what I\nlike in a workspace: natural light, quiet, knowing that I won\u2019t be interrupted\nif I don\u2019t want to be, long blocks of time, and being comfortable and relaxed\n(I\u2019ve got a beautiful desk with a couple of 4k monitors on it in my office, but\nI spend almost all my time on my couch with my laptop).<\/p><p>I wrote custom\nsoftware for the annoying things I have to do frequently, which is great.\n\u00a0I also made an effort to learn to type really fast and the keyboard\nshortcuts that help with my workflow.<\/p><p>Like most\npeople, I sometimes go through periods of a week or two where I just have no\nmotivation to do anything (I suspect it may have something to do with\nnutrition). \u00a0This sucks and always seems to happen at inconvenient\ntimes.\u00a0 I have not figured out what to do\nabout it besides wait for the fog to lift, and to trust that eventually it\nalways does.\u00a0 And I generally try to\navoid people and situations that put me in bad moods, which is good advice\nwhether you care about productivity or not.<\/p><p>In general, I\nthink it\u2019s good to overcommit a little bit. \u00a0I find that I generally get\ndone what I take on, and if I have a little bit too much to do it makes me more\nefficient at everything, which is a way to train to avoid distractions (a great\nhabit to build!). \u00a0However, overcommitting a lot is disastrous.<\/p><p>Don\u2019t neglect\nyour family and friends for the sake of productivity\u2014that\u2019s a very stupid\ntradeoff (and very likely a net productivity loss, because you\u2019ll be less\nhappy). \u00a0Don\u2019t neglect doing things you love or that clear your head\neither.<\/p><p>Finally, to\nrepeat one more time: productivity in the wrong direction isn\u2019t worth anything\nat all. \u00a0Think more about what to work on.<\/p>","created_at":"2018-11-07T16:24:01+0000","updated_at":"2018-11-07T16:24:01+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"blog.samaltman.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":185,"title":"In the Beginning was the Command Line","url":"http:\/\/cristal.inria.fr\/~weis\/info\/commandline.html","content":"<p> About twenty years ago Jobs and Wozniak, the founders of Apple, came\nup with the very strange idea of selling information processing machines\nfor use in the home. The business took off, and its founders made a lot of\nmoney and received the credit they deserved for being daring visionaries.\nBut around the same time, Bill Gates and Paul Allen came up with an idea\neven stranger and more fantastical: selling computer operating systems.\nThis was much weirder than the idea of Jobs and Wozniak. A computer at\nleast had some sort of physical reality to it. It came in a box, you could\nopen it up and plug it in and watch lights blink. An operating system had\nno tangible incarnation at all. It arrived on a disk, of course, but the\ndisk was, in effect, nothing more than the box that the OS came in. The\nproduct itself was a very long string of ones and zeroes that, when\nproperly installed and coddled, gave you the ability to manipulate other\nvery long strings of ones and zeroes. Even those few who actually\nunderstood what a computer operating system was were apt to think of it as\na fantastically arcane engineering prodigy, like a breeder reactor or a\nU-2 spy plane, and not something that could ever be (in the parlance of\nhigh-tech) \"productized.\" \n<\/p><p> Yet now the company that Gates and Allen founded is selling operating\nsystems like Gillette sells razor blades. New releases of operating\nsystems are launched as if they were Hollywood blockbusters, with\ncelebrity endorsements, talk show appearances, and world tours. The market\nfor them is vast enough that people worry about whether it has been\nmonopolized by one company. Even the least technically-minded people in\nour society now have at least a hazy idea of what operating systems do;\nwhat is more, they have strong opinions about their relative merits. It is\ncommonly understood, even by technically unsophisticated computer users,\nthat if you have a piece of software that works on your Macintosh, and you\nmove it over onto a Windows machine, it will not run. That this would, in\nfact, be a laughable and idiotic mistake, like nailing horseshoes to the\ntires of a Buick. \n<\/p><p> A person who went into a coma before Microsoft was founded, and woke\nup now, could pick up this morning's New York Times and understand\neverything in it--almost: \n<\/p><ul><li>Item: the richest man in the world made his fortune\n     from-what? Railways? Shipping? Oil? No, operating systems.\n<\/li><li>Item: the Department of Justice is tackling Microsoft's\n     supposed OS monopoly with legal tools that were invented to\n     restrain the power of Nineteenth-Century robber barons.\n<\/li><li>Item: a woman friend of mine recently told me that she'd\n     broken off a (hitherto) stimulating exchange of e-mail with\n     a young man. At first he had seemed like such an intelligent\n     and interesting guy, she said, but then \"he started going\n     all PC-versus-Mac on me.\"\n<\/li><\/ul><p> What the hell is going on here? And does the operating system business\nhave a future, or only a past? Here is my view, which is entirely\nsubjective; but since I have spent a fair amount of time not only using,\nbut programming, Macintoshes, Windows machines, Linux boxes and the BeOS,\nperhaps it is not so ill-informed as to be completely worthless. This is a\nsubjective essay, more review than research paper, and so it might seem\nunfair or biased compared to the technical reviews you can find in PC\nmagazines.  But ever since the Mac came out, our operating systems have\nbeen based on metaphors, and anything with metaphors in it is fair game as\nfar as I'm concerned. \n<\/p><p>MGBs, TANKS, AND BATMOBILES\n<\/p><p> Around the time that Jobs, Wozniak, Gates, and Allen were dreaming up\nthese unlikely schemes, I was a teenager living in Ames, Iowa. One of my\nfriends' dads had an old MGB sports car rusting away in his garage.\nSometimes he would actually manage to get it running and then he would\ntake us for a spin around the block, with a memorable look of wild\nyouthful exhiliration on his face; to his worried passengers, he was a\nmadman, stalling and backfiring around Ames, Iowa and eating the dust of\nrusty Gremlins and Pintos, but in his own mind he was Dustin Hoffman\ntooling across the Bay Bridge with the wind in his hair. \n<\/p><p> In retrospect, this was telling me two things about people's\nrelationship to technology. One was that romance and image go a long way\ntowards shaping their opinions. If you doubt it (and if you have a lot of\nspare time on your hands) just ask anyone who owns a Macintosh and who, on\nthose grounds, imagines him- or herself to be a member of an oppressed\nminority group. \n<\/p><p> The other, somewhat subtler point, was that interface is very\nimportant. Sure, the MGB was a lousy car in almost every way that counted:\nbalky, unreliable, underpowered. But it was fun to drive. It was\nresponsive. Every pebble on the road was felt in the bones, every nuance\nin the pavement transmitted instantly to the driver's hands. He could\nlisten to the engine and tell what was wrong with it. The steering\nresponded immediately to commands from his hands. To us passengers it was\na pointless exercise in going nowhere--about as interesting as peering\nover someone's shoulder while he punches numbers into a spreadsheet. But\nto the driver it was an experience. For a short time he was extending his\nbody and his senses into a larger realm, and doing things that he couldn't\ndo unassisted. \n<\/p><p> The analogy between cars and operating systems is not half bad, and so\nlet me run with it for a moment, as a way of giving an executive summary\nof our situation today. \n<\/p><p> Imagine a crossroads where four competing auto dealerships are\nsituated. One of them (Microsoft) is much, much bigger than the others. It\nstarted out years ago selling three-speed bicycles (MS-DOS); these were\nnot perfect, but they worked, and when they broke you could easily fix\nthem. \n<\/p><p> There was a competing bicycle dealership next door (Apple) that one\nday began selling motorized vehicles--expensive but attractively styled\ncars with their innards hermetically sealed, so that how they worked was\nsomething of a mystery. \n<\/p><p> The big dealership responded by rushing a moped upgrade kit (the\noriginal Windows) onto the market. This was a Rube Goldberg contraption\nthat, when bolted onto a three-speed bicycle, enabled it to keep up, just\nbarely, with Apple-cars. The users had to wear goggles and were always\npicking bugs out of their teeth while Apple owners sped along in\nhermetically sealed comfort, sneering out the windows. But the\nMicro-mopeds were cheap, and easy to fix compared with the Apple-cars, and\ntheir market share waxed. \n<\/p><p> Eventually the big dealership came out with a full-fledged car: a\ncolossal station wagon (Windows 95). It had all the aesthetic appeal of a\nSoviet worker housing block, it leaked oil and blew gaskets, and it was an\nenormous success. A little later, they also came out with a hulking\noff-road vehicle intended for industrial users (Windows NT) which was no\nmore beautiful than the station wagon, and only a little more reliable. \n<\/p><p> Since then there has been a lot of noise and shouting, but little has\nchanged. The smaller dealership continues to sell sleek Euro-styled sedans\nand to spend a lot of money on advertising campaigns. They have had GOING\nOUT OF BUSINESS! signs taped up in their windows for so long that they\nhave gotten all yellow and curly. The big one keeps making bigger and\nbigger station wagons and ORVs. \n<\/p><p> On the other side of the road are two competitors that have come along\nmore recently. \n<\/p><p> One of them (Be, Inc.) is selling fully operational Batmobiles (the\nBeOS). They are more beautiful and stylish even than the Euro-sedans,\nbetter designed, more technologically advanced, and at least as reliable\nas anything else on the market--and yet cheaper than the others. \n<\/p><p> With one exception, that is: Linux, which is right next door, and\nwhich is not a business at all. It's a bunch of RVs, yurts, tepees, and\ngeodesic domes set up in a field and organized by consensus. The people\nwho live there are making tanks. These are not old-fashioned, cast-iron\nSoviet tanks; these are more like the M1 tanks of the U.S. Army, made of\nspace-age materials and jammed with sophisticated technology from one end\nto the other.  But they are better than Army tanks. They've been modified\nin such a way that they never, ever break down, are light and maneuverable\nenough to use on ordinary streets, and use no more fuel than a subcompact\ncar. These tanks are being cranked out, on the spot, at a terrific pace,\nand a vast number of them are lined up along the edge of the road with\nkeys in the ignition. Anyone who wants can simply climb into one and drive\nit away for free. \n<\/p><p> Customers come to this crossroads in throngs, day and night.  Ninety\npercent of them go straight to the biggest dealership and buy station\nwagons or off-road vehicles. They do not even look at the other\ndealerships. \n<\/p><p> Of the remaining ten percent, most go and buy a sleek Euro-sedan,\npausing only to turn up their noses at the philistines going to buy the\nstation wagons and ORVs. If they even notice the people on the opposite\nside of the road, selling the cheaper, technically superior vehicles,\nthese customers deride them cranks and half-wits. \n<\/p><p> The Batmobile outlet sells a few vehicles to the occasional car nut\nwho wants a second vehicle to go with his station wagon, but seems to\naccept, at least for now, that it's a fringe player. \n<\/p><p> The group giving away the free tanks only stays alive because it is\nstaffed by volunteers, who are lined up at the edge of the street with\nbullhorns, trying to draw customers' attention to this incredible\nsituation. A typical conversation goes something like this: \n<\/p><p> Hacker with bullhorn: \"Save your money! Accept one of our free tanks!\nIt is invulnerable, and can drive across rocks and swamps at ninety miles\nan hour while getting a hundred miles to the gallon!\" \n<\/p><p> Prospective station wagon buyer: \"I know what you say is\ntrue...but...er...I don't know how to maintain a tank!\" \n<\/p><p> Bullhorn: \"You don't know how to maintain a station wagon either!\" \n<\/p><p> Buyer: \"But this dealership has mechanics on staff. If something goes\nwrong with my station wagon, I can take a day off work, bring it here, and\npay them to work on it while I sit in the waiting room for hours,\nlistening to elevator music.\" \n<\/p><p> Bullhorn: \"But if you accept one of our free tanks we will send\nvolunteers to your house to fix it for free while you sleep!\" \n<\/p><p> Buyer: \"Stay away from my house, you freak!\" \n<\/p><p> Bullhorn: \"But...\" \n<\/p><p> Buyer: \"Can't you see that everyone is buying station wagons?\" \n<\/p><p> BIT-FLINGER\n<\/p><p> The connection between cars, and ways of interacting with computers,\nwouldn't have occurred to me at the time I was being taken for rides in\nthat MGB. I had signed up to take a computer programming class at Ames\nHigh School. After a few introductory lectures, we students were granted\nadmission into a tiny room containing a teletype, a telephone, and an\nold-fashioned modem consisting of a metal box with a pair of rubber cups\non the top (note: many readers, making their way through that last\nsentence, probably felt an initial pang of dread that this essay was about\nto turn into a tedious, codgerly reminiscence about how tough we had it\nback in the old days; rest assured that I am actually positioning my\npieces on the chessboard, as it were, in preparation to make a point about\ntruly hip and up-to-the minute topics like Open Source Software). The\nteletype was exactly the same sort of machine that had been used, for\ndecades, to send and receive telegrams. It was basically a loud typewriter\nthat could only produce UPPERCASE LETTERS. Mounted to one side of it was a\nsmaller machine with a long reel of paper tape on it, and a clear plastic\nhopper underneath. \n<\/p><p> In order to connect this device (which was not a computer at all)  to\nthe Iowa State University mainframe across town, you would pick up the\nphone, dial the computer's number, listen for strange noises, and then\nslam the handset down into the rubber cups. If your aim was true, one\nwould wrap its neoprene lips around the earpiece and the other around the\nmouthpiece, consummating a kind of informational soixante-neuf.  The\nteletype would shudder as it was possessed by the spirit of the distant\nmainframe, and begin to hammer out cryptic messages. \n<\/p><p> Since computer time was a scarce resource, we used a sort of batch\nprocessing technique. Before dialing the phone, we would turn on the tape\npuncher (a subsidiary machine bolted to the side of the teletype) and type\nin our programs. Each time we depressed a key, the teletype would bash out\na letter on the paper in front of us, so we could read what we'd typed;\nbut at the same time it would convert the letter into a set of eight\nbinary digits, or bits, and punch a corresponding pattern of holes across\nthe width of a paper tape. The tiny disks of paper knocked out of the tape\nwould flutter down into the clear plastic hopper, which would slowly fill\nup what can only be described as actual bits. On the last day of the\nschool year, the smartest kid in the class (not me) jumped out from behind\nhis desk and flung several quarts of these bits over the head of our\nteacher, like confetti, as a sort of semi-affectionate practical joke. The\nimage of this man sitting there, gripped in the opening stages of an\natavistic fight-or-flight reaction, with millions of bits (megabytes) \nsifting down out of his hair and into his nostrils and mouth, his face\ngradually turning purple as he built up to an explosion, is the single\nmost memorable scene from my formal education. \n<\/p><p> Anyway, it will have been obvious that my interaction with the\ncomputer was of an extremely formal nature, being sharply divided up into\ndifferent phases, viz.: (1) sitting at home with paper and pencil, miles\nand miles from any computer, I would think very, very hard about what I\nwanted the computer to do, and translate my intentions into a computer\nlanguage--a series of alphanumeric symbols on a page. (2) I would carry\nthis across a sort of informational cordon sanitaire (three miles of\nsnowdrifts) to school and type those letters into a machine--not a\ncomputer--which would convert the symbols into binary numbers and record\nthem visibly on a tape. (3) Then, through the rubber-cup modem, I would\ncause those numbers to be sent to the university mainframe, which would\n(4) do arithmetic on them and send different numbers back to the teletype.\n(5) The teletype would convert these numbers back into letters and hammer\nthem out on a page and (6) I, watching, would construe the letters as\nmeaningful symbols. \n<\/p><p> The division of responsibilities implied by all of this is admirably\nclean: computers do arithmetic on bits of information.  Humans construe\nthe bits as meaningful symbols. But this distinction is now being blurred,\nor at least complicated, by the advent of modern operating systems that\nuse, and frequently abuse, the power of metaphor to make computers\naccessible to a larger audience. Along the way--possibly because of those\nmetaphors, which make an operating system a sort of work of art--people\nstart to get emotional, and grow attached to pieces of software in the way\nthat my friend's dad did to his MGB. \n<\/p><p> People who have only interacted with computers through graphical user\ninterfaces like the MacOS or Windows--which is to say, almost everyone who\nhas ever used a computer--may have been startled, or at least bemused, to\nhear about the telegraph machine that I used to communicate with a\ncomputer in 1973. But there was, and is, a good reason for using this\nparticular kind of technology. Human beings have various ways of\ncommunicating to each other, such as music, art, dance, and facial\nexpressions, but some of these are more amenable than others to being\nexpressed as strings of symbols. Written language is the easiest of all,\nbecause, of course, it consists of strings of symbols to begin with. If\nthe symbols happen to belong to a phonetic alphabet (as opposed to, say,\nideograms), converting them into bits is a trivial procedure, and one that\nwas nailed, technologically, in the early nineteenth century, with the\nintroduction of Morse code and other forms of telegraphy. \n<\/p><p> We had a human\/computer interface a hundred years before we had\ncomputers. When computers came into being around the time of the Second\nWorld War, humans, quite naturally, communicated with them by simply\ngrafting them on to the already-existing technologies for translating\nletters into bits and vice versa: teletypes and punch card machines. \n<\/p><p> These embodied two fundamentally different approaches to computing.\nWhen you were using cards, you'd punch a whole stack of them and run them\nthrough the reader all at once, which was called batch processing. You\ncould also do batch processing with a teletype, as I have already\ndescribed, by using the paper tape reader, and we were certainly\nencouraged to use this approach when I was in high school. But--though\nefforts were made to keep us unaware of this--the teletype could do\nsomething that the card reader could not. On the teletype, once the modem\nlink was established, you could just type in a line and hit the return\nkey. The teletype would send that line to the computer, which might or\nmight not respond with some lines of its own, which the teletype would\nhammer out--producing, over time, a transcript of your exchange with the\nmachine. This way of doing it did not even have a name at the time, but\nwhen, much later, an alternative became available, it was retroactively\ndubbed the Command Line Interface. \n<\/p><p> When I moved on to college, I did my computing in large, stifling\nrooms where scores of students would sit in front of slightly updated\nversions of the same machines and write computer programs: these used\ndot-matrix printing mechanisms, but were (from the computer's point of\nview) identical to the old teletypes. By that point, computers were better\nat time-sharing--that is, mainframes were still mainframes, but they were\nbetter at communicating with a large number of terminals at once.\nConsequently, it was no longer necessary to use batch processing. Card\nreaders were shoved out into hallways and boiler rooms, and batch\nprocessing became a nerds-only kind of thing, and consequently took on a\ncertain eldritch flavor among those of us who even knew it existed. We\nwere all off the Batch, and on the Command Line, interface now--my very\nfirst shift in operating system paradigms, if only I'd known it. \n<\/p><p> A huge stack of accordion-fold paper sat on the floor underneath each\none of these glorified teletypes, and miles of paper shuddered through\ntheir platens. Almost all of this paper was thrown away or recycled\nwithout ever having been touched by ink--an ecological atrocity so glaring\nthat those machines soon replaced by video terminals--so-called \"glass\nteletypes\"--which were quieter and didn't waste paper. Again, though, from\nthe computer's point of view these were indistinguishable from World War\nII-era teletype machines. In effect we still used Victorian technology to\ncommunicate with computers until about 1984, when the Macintosh was\nintroduced with its Graphical User Interface.  Even after that, the\nCommand Line continued to exist as an underlying stratum--a sort of\nbrainstem reflex--of many modern computer systems all through the heyday\nof Graphical User Interfaces, or GUIs as I will call them from now on. \n<\/p><p> GUIs\n<\/p><p> Now the first job that any coder needs to do when writing a new piece\nof software is to figure out how to take the information that is being\nworked with (in a graphics program, an image; in a spreadsheet, a grid of\nnumbers) and turn it into a linear string of bytes. These strings of bytes\nare commonly called files or (somewhat more hiply) streams. They are to\ntelegrams what modern humans are to Cro-Magnon man, which is to say the\nsame thing under a different name. All that you see on your computer\nscreen--your Tomb Raider, your digitized voice mail messages, faxes, and\nword processing documents written in thirty-seven different typefaces--is\nstill, from the computer's point of view, just like telegrams, except much\nlonger, and demanding of more arithmetic. \n<\/p><p> The quickest way to get a taste of this is to fire up your web\nbrowser, visit a site, and then select the View\/Document Source menu item.\nYou will get a bunch of computer code that looks something like this: \n<\/p><pre>\n&lt;HTML&gt;\n&lt;HEAD&gt;\n&lt;TITLE&gt;Welcome to the Avon Books Homepage&lt;\/TITLE&gt;\n&lt;\/HEAD&gt;\n&lt;MAP NAME=\"left0199\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"16,56,111,67\" HREF=\"\/bard\/\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"14,77,111,89\" HREF=\"\/eos\/\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"17,98,112,110\" HREF=\"\/twilight\/\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"18,119,112,131\" \n\tHREF=\"\/avon_user\/category.html?category_id=271\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"19,140,112,152\" \n\tHREF=\"http:\/\/www.goners.com\/\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"18,161,111,173\" \n\tHREF=\"http:\/\/www.spikebooks.com\/\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"2,181,112,195\" \n\tHREF=\"\/avon_user\/category.html?category_id=277\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"9,203,112,216\" \n\tHREF=\"\/chathamisland\/\"&gt;\n     &lt;AREA SHAPE=\"rect\" COORDS=\"7,223,112,236\" \n\tHREF=\"\/avon_user\/search.html\"&gt;\n&lt;\/MAP&gt;\n&lt;BODY TEXT=\"#478CFF\" LINK=\"#FFFFFF\" VLINK=\"#000000\" ALINK=\"#478CFF\" BGCOLOR=\"#003399\"&gt;\n&lt;TABLE BORDER=\"0\" WIDTH=\"600\" CELLPADDING=\"0\" CELLSPACING=\"0\"&gt;\n&lt;TR VALIGN=TOP&gt;\n     &lt;TD ROWSPAN=\"3\"&gt;\n     &lt;A HREF=\"\/cgi-bin\/imagemap\/maps\/left.gif.map\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/nav\/left0199.gif\" WIDTH=\"113\" HEIGHT=\"280\"\nBORDER=\"0\" USEMAP=\"#left0199\"&gt;&lt;\/A&gt;&lt;\/TD&gt;&lt;TD ROWSPAN=\"3\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/homepagejan98\/2ndleft.gif\" WIDTH=\"144\"\nHEIGHT=\"280\" BORDER=\"0\"&gt;&lt;\/TD&gt;&lt;TD&gt;&lt;A HREF=\"\/avon\/about.html\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/homepagejan98\/aboutavon.gif\" ALT=\"About\nAvon Books\" WIDTH=\"199\" HEIGHT=\"44\" BORDER=\"0\"&gt;&lt;\/A&gt;&lt;\/TD&gt;&lt;TD\nROWSPAN=\"3\"&gt;&lt;A HREF=\"\/avon\/fiction\/guides.html\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/feb98\/right1.gif\" ALT=\"Reading Groups\"\nWIDTH=\"165\" HEIGHT=\"121\" BORDER=\"0\"&gt;&lt;\/A&gt;&lt;BR&gt;&lt;A\nHREF=\"\/avon\/feature\/feb99\/crook.html\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/feb99\/crook_text.gif\" ALT=\"The Crook\nFactory\" WIDTH=\"165\" HEIGHT=\"96\" BORDER=\"0\"&gt;&lt;\/A&gt;&lt;BR&gt;&lt;A\nHREF=\"http:\/\/apps.hearstnewmedia.com\/cgi-bin\/gx.cgi\/AppLogic+APPSSURVEYS\nQuestionnaire?domain_id=182&amp;survey_id=541\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/feb99\/env_text.gif\" ALT=\"The Envelope\nPlease\" WIDTH=\"165\" HEIGHT=\"63\" BORDER=\"0\"&gt;&lt;\/A&gt;&lt;\/TD&gt;\n&lt;\/TR&gt;\n&lt;TR VALIGN=TOP&gt;&lt;TD&gt;&lt;IMG SRC=\"\/avon\/images\/home\/feb98\/main.gif\"\nWIDTH=\"199\" HEIGHT=\"182\" BORDER=\"0\"&gt;&lt;\/TD&gt;&lt;\/TR&gt;&lt;TR\nVALIGN=TOP&gt;&lt;TD&gt;&lt;A HREF=\"\/avon\/feature\/jan99\/sitchin.html\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/jan99\/sitchin_text.gif\" WIDTH=\"199\"\nHEIGHT=\"54\" BORDER=\"0\"&gt;&lt;\/A&gt;&lt;\/TD&gt;&lt;\/TR&gt;&lt;TR VALIGN=TOP&gt;&lt;TD\nCOLSPAN=\"4\"&gt;&lt;IMG\nSRC=\"\/avon\/images\/home\/jan99\/avon_bottom_beau.gif\" WIDTH=\"622\"\nHEIGHT=\"179\" BORDER=\"0\" USEMAP=\"#bottom\"&gt;&lt;\/TD&gt;&lt;\/TR&gt;&lt;TR&gt;&lt;TD\nALIGN=CENTER VALIGN=TOP COLSPAN=\"4\"&gt;&lt;FONT SIZE=\"2\"\nFACE=\"ARIAL,COURIER\"&gt;&lt;PRE&gt;\n&lt;\/PRE&gt;&lt;A HREF=\"\/avon\/ordering.html\"&gt;How to order&lt;\/A&gt; | &lt;A\nHREF=\"\/avon\/faq.html#manu\"&gt;How to submit a Manuscript&lt;\/A&gt; | &lt;A\nHREF=\"mailto:avonweb@hearst.com\"&gt;Contact us&lt;\/A&gt; | &lt;A\nHREF=\"\/avon\/policy.html\"&gt;Privacy Policy&lt;\/A&gt;&lt;\/FONT&gt;\n&lt;P&gt;\n&lt;\/FONT&gt;&lt;\/TD&gt;\n&lt;\/TR&gt;\n&lt;\/TABLE&gt;\n&lt;\/BODY&gt;\n&lt;\/HTML&gt;\n<\/pre><p> This crud is called HTML (HyperText Markup Language) and it is\nbasically a very simple programming language instructing your web browser\nhow to draw a page on a screen. Anyone can learn HTML and many people do.\nThe important thing is that no matter what splendid multimedia web pages\nthey might represent, HTML files are just telegrams. \n<\/p><p> When Ronald Reagan was a radio announcer, he used to call baseball\ngames by reading the terse descriptions that trickled in over the\ntelegraph wire and were printed out on a paper tape. He would sit there,\nall by himself in a padded room with a microphone, and the paper tape\nwould eke out of the machine and crawl over the palm of his hand printed\nwith cryptic abbreviations. If the count went to three and two, Reagan\nwould describe the scene as he saw it in his mind's eye: \"The brawny\nleft-hander steps out of the batter's box to wipe the sweat from his brow.\nThe umpire steps forward to sweep the dirt from home plate.\" and so on.\nWhen the cryptogram on the paper tape announced a base hit, he would whack\nthe edge of the table with a pencil, creating a little sound effect, and\ndescribe the arc of the ball as if he could actually see it. His\nlisteners, many of whom presumably thought that Reagan was actually at the\nballpark watching the game, would reconstruct the scene in their minds\naccording to his descriptions. \n<\/p><p> This is exactly how the World Wide Web works: the HTML files are the\npithy description on the paper tape, and your Web browser is Ronald\nReagan. The same is true of Graphical User Interfaces in general. \n<\/p><p> So an OS is a stack of metaphors and abstractions that stands between\nyou and the telegrams, and embodying various tricks the programmer used to\nconvert the information you're working with--be it images, e-mail\nmessages, movies, or word processing documents--into the necklaces of\nbytes that are the only things computers know how to work with. When we\nused actual telegraph equipment (teletypes) or their higher-tech\nsubstitutes (\"glass teletypes,\" or the MS-DOS command line) to work with\nour computers, we were very close to the bottom of that stack. When we use\nmost modern operating systems, though, our interaction with the machine is\nheavily mediated. Everything we do is interpreted and translated time and\nagain as it works its way down through all of the metaphors and\nabstractions. \n<\/p><p> The Macintosh OS was a revolution in both the good and bad senses of\nthat word. Obviously it was true that command line interfaces were not for\neveryone, and that it would be a good thing to make computers more\naccessible to a less technical audience--if not for altruistic reasons,\nthen because those sorts of people constituted an incomparably vaster\nmarket. It was clear the the Mac's engineers saw a whole new country\nstretching out before them; you could almost hear them muttering, \"Wow! We\ndon't have to be bound by files as linear streams of bytes anymore, vive\nla revolution, let's see how far we can take this!\" No command line\ninterface was available on the Macintosh; you talked to it with the mouse,\nor not at all. This was a statement of sorts, a credential of\nrevolutionary purity. It seemed that the designers of the Mac intended to\nsweep Command Line Interfaces into the dustbin of history. \n<\/p><p> My own personal love affair with the Macintosh began in the spring of\n1984 in a computer store in Cedar Rapids, Iowa, when a friend of\nmine--coincidentally, the son of the MGB owner--showed me a Macintosh\nrunning MacPaint, the revolutionary drawing program. It ended in July of\n1995 when I tried to save a big important file on my Macintosh Powerbook\nand instead instead of doing so, it annihilated the data so thoroughly\nthat two different disk crash utility programs were unable to find any\ntrace that it had ever existed. During the intervening ten years, I had a\npassion for the MacOS that seemed righteous and reasonable at the time but\nin retrospect strikes me as being exactly the same sort of goofy\ninfatuation that my friend's dad had with his car. \n<\/p><p> The introduction of the Mac triggered a sort of holy war in the\ncomputer world. Were GUIs a brilliant design innovation that made\ncomputers more human-centered and therefore accessible to the masses,\nleading us toward an unprecedented revolution in human society, or an\ninsulting bit of audiovisual gimcrackery dreamed up by flaky Bay Area\nhacker types that stripped computers of their power and flexibility and\nturned the noble and serious work of computing into a childish video game? \n<\/p><p> This debate actually seems more interesting to me today than it did in\nthe mid-1980s. But people more or less stopped debating it when Microsoft\nendorsed the idea of GUIs by coming out with the first Windows. At this\npoint, command-line partisans were relegated to the status of silly old\ngrouches, and a new conflict was touched off, between users of MacOS and\nusers of Windows. \n<\/p><p> There was plenty to argue about. The first Macintoshes looked\ndifferent from other PCs even when they were turned off: they consisted of\none box containing both CPU (the part of the computer that does arithmetic\non bits) and monitor screen. This was billed, at the time, as a\nphilosophical statement of sorts:  Apple wanted to make the personal\ncomputer into an appliance, like a toaster. But it also reflected the\npurely technical demands of running a graphical user interface. In a GUI\nmachine, the chips that draw things on the screen have to be integrated\nwith the computer's central processing unit, or CPU, to a far greater\nextent than is the case with command-line interfaces, which until recently\ndidn't even know that they weren't just talking to teletypes. \n<\/p><p> This distinction was of a technical and abstract nature, but it became\nclearer when the machine crashed (it is commonly the case with\ntechnologies that you can get the best insight about how they work by\nwatching them fail). When everything went to hell and the CPU began\nspewing out random bits, the result, on a CLI machine, was lines and lines\nof perfectly formed but random characters on the screen--known to\ncognoscenti as \"going Cyrillic.\" But to the MacOS, the screen was not a\nteletype, but a place to put graphics; the image on the screen was a\nbitmap, a literal rendering of the contents of a particular portion of the\ncomputer's memory. When the computer crashed and wrote gibberish into the\nbitmap, the result was something that looked vaguely like static on a\nbroken television set--a \"snow crash.\" \n<\/p><p> And even after the introduction of Windows, the underlying differences\nendured; when a Windows machine got into trouble, the old command-line\ninterface would fall down over the GUI like an asbestos fire curtain\nsealing off the proscenium of a burning opera. When a Macintosh got into\ntrouble it presented you with a cartoon of a bomb, which was funny the\nfirst time you saw it. \n<\/p><p> And these were by no means superficial differences. The reversion of\nWindows to a CLI when it was in distress proved to Mac partisans that\nWindows was nothing more than a cheap facade, like a garish afghan flung\nover a rotted-out sofa. They were disturbed and annoyed by the sense that\nlurking underneath Windows' ostensibly user-friendly interface\nwas--literally--a subtext. \n<\/p><p> For their part, Windows fans might have made the sour observation that\nall computers, even Macintoshes, were built on that same subtext, and that\nthe refusal of Mac owners to admit that fact to themselves seemed to\nsignal a willingness, almost an eagerness, to be duped. \n<\/p><p> Anyway, a Macintosh had to switch individual bits in the memory chips\non the video card, and it had to do it very fast, and in arbitrarily\ncomplicated patterns. Nowadays this is cheap and easy, but in the\ntechnological regime that prevailed in the early 1980s, the only realistic\nway to do it was to build the motherboard (which contained the CPU) and\nthe video system (which contained the memory that was mapped onto the\nscreen) as a tightly integrated whole--hence the single, hermetically\nsealed case that made the Macintosh so distinctive. \n<\/p><p> When Windows came out, it was conspicuous for its ugliness, and its\ncurrent successors, Windows 95 and Windows NT, are not things that people\nwould pay money to look at either. Microsoft's complete disregard for\naesthetics gave all of us Mac-lovers plenty of opportunities to look down\nour noses at them. That Windows looked an awful lot like a direct ripoff\nof MacOS gave us a burning sense of moral outrage to go with it. Among\npeople who really knew and appreciated computers (hackers, in Steven\nLevy's non-pejorative sense of that word) and in a few other niches such\nas professional musicians, graphic artists and schoolteachers, the\nMacintosh, for a while, was simply the computer. It was seen as not only a\nsuperb piece of engineering, but an embodiment of certain ideals about the\nuse of technology to benefit mankind, while Windows was seen as a\npathetically clumsy imitation and a sinister world domination plot rolled\ninto one. So very early, a pattern had been established that endures to\nthis day: people dislike Microsoft, which is okay; but they dislike it for\nreasons that are poorly considered, and in the end, self-defeating. \n<\/p><p> CLASS STRUGGLE ON THE DESKTOP\n<\/p><p> Now that the Third Rail has been firmly grasped, it is worth reviewing\nsome basic facts here: like any other publicly traded, for-profit\ncorporation, Microsoft has, in effect, borrowed a bunch of money from some\npeople (its stockholders) in order to be in the bit business. As an\nofficer of that corporation, Bill Gates has one responsibility only, which\nis to maximize return on investment. He has done this incredibly well. Any\nactions taken in the world by Microsoft-any software released by them, for\nexample--are basically epiphenomena, which can't be interpreted or\nunderstood except insofar as they reflect Bill Gates's execution of his\none and only responsibility. \n<\/p><p> It follows that if Microsoft sells goods that are aesthetically\nunappealing, or that don't work very well, it does not mean that they are\n(respectively) philistines or half-wits. It is because Microsoft's\nexcellent management has figured out that they can make more money for\ntheir stockholders by releasing stuff with obvious, known imperfections\nthan they can by making it beautiful or bug-free. This is annoying, but\n(in the end) not half so annoying as watching Apple inscrutably and\nrelentlessly destroy itself. \n<\/p><p> Hostility towards Microsoft is not difficult to find on the Net, and\nit blends two strains: resentful people who feel Microsoft is too\npowerful, and disdainful people who think it's tacky. This is all strongly\nreminiscent of the heyday of Communism and Socialism, when the bourgeoisie\nwere hated from both ends: by the proles, because they had all the money,\nand by the intelligentsia, because of their tendency to spend it on lawn\nornaments. Microsoft is the very embodiment of modern high-tech\nprosperity--it is, in a word, bourgeois--and so it attracts all of the\nsame gripes. \n<\/p><p> The opening \"splash screen\" for Microsoft Word 6.0 summed it up pretty\nneatly: when you started up the program you were treated to a picture of\nan expensive enamel pen lying across a couple of sheets of fancy-looking\nhandmade writing paper. It was obviously a bid to make the software look\nclassy, and it might have worked for some, but it failed for me, because\nthe pen was a ballpoint, and I'm a fountain pen man. If Apple had done it,\nthey would've used a Mont Blanc fountain pen, or maybe a Chinese\ncalligraphy brush. And I doubt that this was an accident. Recently I spent\na while re-installing Windows NT on one of my home computers, and many\ntimes had to double-click on the \"Control Panel\" icon. For reasons that\nare difficult to fathom, this icon consists of a picture of a clawhammer\nand a chisel or screwdriver resting on top of a file folder. \n<\/p><p> These aesthetic gaffes give one an almost uncontrollable urge to make\nfun of Microsoft, but again, it is all beside the point--if Microsoft had\ndone focus group testing of possible alternative graphics, they probably\nwould have found that the average mid-level office worker associated\nfountain pens with effete upper management toffs and was more comfortable\nwith ballpoints.  Likewise, the regular guys, the balding dads of the\nworld who probably bear the brunt of setting up and maintaining home\ncomputers, can probably relate better to a picture of a clawhammer--while\nperhaps harboring fantasies of taking a real one to their balky computers. \n<\/p><p> This is the only way I can explain certain peculiar facts about the\ncurrent market for operating systems, such as that ninety percent of all\ncustomers continue to buy station wagons off the Microsoft lot while free\ntanks are there for the taking, right across the street. \n<\/p><p> A string of ones and zeroes was not a difficult thing for Bill Gates\nto distribute, one he'd thought of the idea. The hard part was selling\nit--reassuring customers that they were actually getting something in\nreturn for their money. \n<\/p><p> Anyone who has ever bought a piece of software in a store has had the\ncuriously deflating experience of taking the bright shrink-wrapped box\nhome, tearing it open, finding that it's 95 percent air, throwing away all\nthe little cards, party favors, and bits of trash, and loading the disk\ninto the computer. The end result (after you've lost the disk) is nothing\nexcept some images on a computer screen, and some capabilities that\nweren't there before. Sometimes you don't even have that--you have a\nstring of error messages instead. But your money is definitely gone. Now\nwe are almost accustomed to this, but twenty years ago it was a very dicey\nbusiness proposition. Bill Gates made it work anyway. He didn't make it\nwork by selling the best software or offering the cheapest price. Instead\nhe somehow got people to believe that they were receiving something in\nexchange for their money. \n<\/p><p> The streets of every city in the world are filled with those hulking,\nrattling station wagons. Anyone who doesn't own one feels a little weird,\nand wonders, in spite of himself, whether it might not be time to cease\nresistance and buy one; anyone who does, feels confident that he has\nacquired some meaningful possession, even on those days when the vehicle\nis up on a lift in an auto repair shop. \n<\/p><p> All of this is perfectly congruent with membership in the bourgeoisie,\nwhich is as much a mental, as a material state. And it explains why\nMicrosoft is regularly attacked, on the Net, from both sides. People who\nare inclined to feel poor and oppressed construe everything Microsoft does\nas some sinister Orwellian plot. People who like to think of themselves as\nintelligent and informed technology users are driven crazy by the\nclunkiness of Windows. \n<\/p><p> Nothing is more annoying to sophisticated people to see someone who is\nrich enough to know better being tacky--unless it is to realize, a moment\nlater, that they probably know they are tacky and they simply don't care\nand they are going to go on being tacky, and rich, and happy, forever.\nMicrosoft therefore bears the same relationship to the Silicon Valley\nelite as the Beverly Hillbillies did to their fussy banker, Mr.\nDrysdale--who is irritated not so much by the fact that the Clampetts\nmoved to his neighborhood as by the knowledge that, when Jethro is seventy\nyears old, he's still going to be talking like a hillbilly and wearing bib\noveralls, and he's still going to be a lot richer than Mr. Drysdale. \n<\/p><p> Even the hardware that Windows ran on, when compared to the machines\nput out by Apple, looked like white-trash stuff, and still mostly does.\nThe reason was that Apple was and is a hardware company, while Microsoft\nwas and is a software company.  Apple therefore had a monopoly on hardware\nthat could run MacOS, whereas Windows-compatible hardware came out of a\nfree market.  The free market seems to have decided that people will not\npay for cool-looking computers; PC hardware makers who hire designers to\nmake their stuff look distinctive get their clocks cleaned by Taiwanese\nclone makers punching out boxes that look as if they belong on\ncinderblocks in front of someone's trailer. But Apple could make their\nhardware as pretty as they wanted to and simply pass the higher prices on\nto their besotted consumers, like me.  Only last week (I am writing this\nsentence in early Jan. 1999)  the technology sections of all the\nnewspapers were filled with adulatory press coverage of how Apple had\nreleased the iMac in several happenin' new colors like Blueberry and\nTangerine. \n<\/p><p> Apple has always insisted on having a hardware monopoly, except for a\nbrief period in the mid-1990s when they allowed clone-makers to compete\nwith them, before subsequently putting them out of business. Macintosh\nhardware was, consequently, expensive. You didn't open it up and fool\naround with it because doing so would void the warranty. In fact the first\nMac was specifically designed to be difficult to open--you needed a kit of\nexotic tools, which you could buy through little ads that began to appear\nin the back pages of magazines a few months after the Mac came out on the\nmarket. These ads always had a certain disreputable air about them, like\npitches for lock-picking tools in the backs of lurid detective magazines. \n<\/p><p> This monopolistic policy can be explained in at least three different\nways. \n<\/p><p> THE CHARITABLE EXPLANATION is that the hardware monopoly policy\nreflected a drive on Apple's part to provide a seamless, unified blending\nof hardware, operating system, and software. There is something to this.\nIt is hard enough to make an OS that works well on one specific piece of\nhardware, designed and tested by engineers who work down the hallway from\nyou, in the same company. Making an OS to work on arbitrary pieces of\nhardware, cranked out by rabidly entrepeneurial clonemakers on the other\nside of the International Date Line, is very difficult, and accounts for\nmuch of the troubles people have using Windows. \n<\/p><p> THE FINANCIAL EXPLANATION is that Apple, unlike Microsoft, is and\nalways has been a hardware company. It simply depends on revenue from\nselling hardware, and cannot exist without it. \n<\/p><p> THE NOT-SO-CHARITABLE EXPLANATION has to do with Apple's corporate\nculture, which is rooted in Bay Area Baby Boomdom. \n<\/p><p> Now, since I'm going to talk for a moment about culture, full\ndisclosure is probably in order, to protect myself against allegations of\nconflict of interest and ethical turpitude: (1)  Geographically I am a\nSeattleite, of a Saturnine temperament, and inclined to take a sour view\nof the Dionysian Bay Area, just as they tend to be annoyed and appalled by\nus. (2) Chronologically I am a post-Baby Boomer. I feel that way, at\nleast, because I never experienced the fun and exciting parts of the whole\nBoomer scene--just spent a lot of time dutifully chuckling at Boomers'\nmaddeningly pointless anecdotes about just how stoned they got on various\noccasions, and politely fielding their assertions about how great their\nmusic was. But even from this remove it was possible to glean certain\npatterns, and one that recurred as regularly as an urban legend was the\none about how someone would move into a commune populated by\nsandal-wearing, peace-sign flashing flower children, and eventually\ndiscover that, underneath this facade, the guys who ran it were actually\ncontrol freaks; and that, as living in a commune, where much lip service\nwas paid to ideals of peace, love and harmony, had deprived them of\nnormal, socially approved outlets for their control-freakdom, it tended to\ncome out in other, invariably more sinister, ways. \n<\/p><p> Applying this to the case of Apple Computer will be left as an\nexercise for the reader, and not a very difficult exercise. \n<\/p><p> It is a bit unsettling, at first, to think of Apple as a control\nfreak, because it is completely at odds with their corporate image.\nWeren't these the guys who aired the famous Super Bowl ads showing suited,\nblindfolded executives marching like lemmings off a cliff? Isn't this the\ncompany that even now runs ads picturing the Dalai Lama (except in Hong\nKong) and Einstein and other offbeat rebels? \n<\/p><p> It is indeed the same company, and the fact that they have been able\nto plant this image of themselves as creative and rebellious free-thinkers\nin the minds of so many intelligent and media-hardened skeptics really\ngives one pause. It is testimony to the insidious power of expensive slick\nad campaigns and, perhaps, to a certain amount of wishful thinking in the\nminds of people who fall for them. It also raises the question of why\nMicrosoft is so bad at PR, when the history of Apple demonstrates that, by\nwriting large checks to good ad agencies, you can plant a corporate image\nin the minds of intelligent people that is completely at odds with\nreality. (The answer, for people who don't like Damoclean questions, is\nthat since Microsoft has won the hearts and minds of the silent\nmajority--the bourgeoisie--they don't give a damn about having a slick\nimage, any more then Dick Nixon did. \"I want to believe,\"--the mantra that\nFox Mulder has pinned to his office wall in The X-Files--applies in\ndifferent ways to these two companies; Mac partisans want to believe in\nthe image of Apple purveyed in those ads, and in the notion that Macs are\nsomehow fundamentally different from other computers, while Windows people\nwant to believe that they are getting something for their money, engaging\nin a respectable business transaction). \n<\/p><p> In any event, as of 1987, both MacOS and Windows were out on the\nmarket, running on hardware platforms that were radically different from\neach other--not only in the sense that MacOS used Motorola CPU chips while\nWindows used Intel, but in the sense--then overlooked, but in the long\nrun, vastly more significant--that the Apple hardware business was a rigid\nmonopoly and the Windows side was a churning free-for-all. \n<\/p><p> But the full ramifications of this did not become clear until very\nrecently--in fact, they are still unfolding, in remarkably strange ways,\nas I'll explain when we get to Linux. The upshot is that millions of\npeople got accustomed to using GUIs in one form or another. By doing so,\nthey made Apple\/Microsoft a lot of money. The fortunes of many people have\nbecome bound up with the ability of these companies to continue selling\nproducts whose salability is very much open to question. \n<\/p><p> HONEY-POT, TAR-PIT, WHATEVER\n<\/p><p> When Gates and Allen invented the idea of selling software, they ran\ninto criticism from both hackers and sober-sided businesspeople. Hackers\nunderstood that software was just information, and objected to the idea of\nselling it. These objections were partly moral. The hackers were coming\nout of the scientific and academic world where it is imperative to make\nthe results of one's work freely available to the public. They were also\npartly practical; how can you sell something that can be easily copied?\nBusinesspeople, who are polar opposites of hackers in so many ways, had\nobjections of their own. Accustomed to selling toasters and insurance\npolicies, they naturally had a difficult time understanding how a long\ncollection of ones and zeroes could constitute a salable product. \n<\/p><p> Obviously Microsoft prevailed over these objections, and so did Apple.\nBut the objections still exist. The most hackerish of all the hackers, the\nUr-hacker as it were, was and is Richard Stallman, who became so annoyed\nwith the evil practice of selling software that, in 1984 (the same year\nthat the Macintosh went on sale) he went off and founded something called\nthe Free Software Foundation, which commenced work on something called\nGNU. Gnu is an acronym for Gnu's Not Unix, but this is a joke in more ways\nthan one, because GNU most certainly IS Unix,. Because of trademark\nconcerns (\"Unix\" is trademarked by AT&amp;T) they simply could not claim that\nit was Unix, and so, just to be extra safe, they claimed that it wasn't.\nNotwithstanding the incomparable talent and drive possessed by Mr.\nStallman and other GNU adherents, their project to build a free Unix to\ncompete against Microsoft and Apple's OSes was a little bit like trying to\ndig a subway system with a teaspoon. Until, that is, the advent of Linux,\nwhich I will get to later. \n<\/p><p> But the basic idea of re-creating an operating system from scratch was\nperfectly sound and completely doable. It has been done many times. It is\ninherent in the very nature of operating systems. \n<\/p><p> Operating systems are not strictly necessary. There is no reason why a\nsufficiently dedicated coder could not start from nothing with every\nproject and write fresh code to handle such basic, low-level operations as\ncontrolling the read\/write heads on the disk drives and lighting up pixels\non the screen. The very first computers had to be programmed in this way.\nBut since nearly every program needs to carry out those same basic\noperations, this approach would lead to vast duplication of effort. \n<\/p><p> Nothing is more disagreeable to the hacker than duplication of effort.\nThe first and most important mental habit that people develop when they\nlearn how to write computer programs is to generalize, generalize,\ngeneralize. To make their code as modular and flexible as possible,\nbreaking large problems down into small subroutines that can be used over\nand over again in different contexts. Consequently, the development of\noperating systems, despite being technically unnecessary, was inevitable.\nBecause at its heart, an operating system is nothing more than a library\ncontaining the most commonly used code, written once (and hopefully\nwritten well) and then made available to every coder who needs it. \n<\/p><p> So a proprietary, closed, secret operating system is a contradiction\nin terms. It goes against the whole point of having an operating system.\nAnd it is impossible to keep them secret anyway. The source code--the\noriginal lines of text written by the programmers--can be kept secret. But\nan OS as a whole is a collection of small subroutines that do very\nspecific, very clearly defined jobs. Exactly what those subroutines do has\nto be made public, quite explicitly and exactly, or else the OS is\ncompletely useless to programmers; they can't make use of those\nsubroutines if they don't have a complete and perfect understanding of\nwhat the subroutines do. \n<\/p><p> The only thing that isn't made public is exactly how the subroutines\ndo what they do. But once you know what a subroutine does, it's generally\nquite easy (if you are a hacker) to write one of your own that does\nexactly the same thing. It might take a while, and it is tedious and\nunrewarding, but in most cases it's not really hard. \n<\/p><p> What's hard, in hacking as in fiction, is not writing; it's deciding\nwhat to write. And the vendors of commercial OSes have already decided,\nand published their decisions. \n<\/p><p> This has been generally understood for a long time. MS-DOS was\nduplicated, functionally, by a rival product, written from scratch, called\nProDOS, that did all of the same things in pretty much the same way. In\nother words, another company was able to write code that did all of the\nsame things as MS-DOS and sell it at a profit. If you are using the Linux\nOS, you can get a free program called WINE which is a windows emulator;\nthat is, you can open up a window on your desktop that runs windows\nprograms. It means that a completely functional Windows OS has been\nrecreated inside of Unix, like a ship in a bottle. And Unix itself, which\nis vastly more sophisticated than MS-DOS, has been built up from scratch\nmany times over. Versions of it are sold by Sun, Hewlett-Packard, AT&amp;T,\nSilicon Graphics, IBM, and others. \n<\/p><p> People have, in other words, been re-writing basic OS code for so long\nthat all of the technology that constituted an \"operating system\" in the\ntraditional (pre-GUI) sense of that phrase is now so cheap and common that\nit's literally free. Not only could Gates and Allen not sell MS-DOS today,\nthey could not even give it away, because much more powerful OSes are\nalready being given away. Even the original Windows (which was the only\nwindows until 1995) has become worthless, in that there is no point in\nowning something that can be emulated inside of Linux--which is, itself,\nfree. \n<\/p><p> In this way the OS business is very different from, say, the car\nbusiness. Even an old rundown car has some value. You can use it for\nmaking runs to the dump, or strip it for parts. It is the fate of\nmanufactured goods to slowly and gently depreciate as they get old and\nhave to compete against more modern products. \n<\/p><p> But it is the fate of operating systems to become free. \n<\/p><p> Microsoft is a great software applications company. \nApplications--such as Microsoft Word--are an area where innovation brings\nreal, direct, tangible benefits to users. The innovations might be new\ntechnology straight from the research department, or they might be in the\ncategory of bells and whistles, but in any event they are frequently\nuseful and they seem to make users happy. And Microsoft is in the process\nof becoming a great research company. But Microsoft is not such a great\noperating systems company. And this is not necessarily because their\noperating systems are all that bad from a purely technological standpoint.\nMicrosoft's OSes do have their problems, sure, but they are vastly better\nthan they used to be, and they are adequate for most people. \n<\/p><p> Why, then, do I say that Microsoft is not such a great operating\nsystems company? Because the very nature of operating systems is such that\nit is senseless for them to be developed and owned by a specific company.\nIt's a thankless job to begin with.  Applications create possibilities for\nmillions of credulous users, whereas OSes impose limitations on thousands\nof grumpy coders, and so OS-makers will forever be on the shit-list of\nanyone who counts for anything in the high-tech world.  Applications get\nused by people whose big problem is understanding all of their features,\nwhereas OSes get hacked by coders who are annoyed by their limitations.\nThe OS business has been good to Microsoft only insofar as it has given\nthem the money they needed to launch a really good applications software\nbusiness and to hire a lot of smart researchers. Now it really ought to be\njettisoned, like a spent booster stage from a rocket.  The big question is\nwhether Microsoft is capable of doing this.  Or is it addicted to OS sales\nin the same way as Apple is to selling hardware? \n<\/p><p> Keep in mind that Apple's ability to monopolize its own hardware\nsupply was once cited, by learned observers, as a great advantage over\nMicrosoft. At the time, it seemed to place them in a much stronger\nposition. In the end, it nearly killed them, and may kill them yet. The\nproblem, for Apple, was that most of the world's computer users ended up\nowning cheaper hardware. But cheap hardware couldn't run MacOS, and so\nthese people switched to Windows.\n<\/p><p> Replace \"hardware\" with \"operating systems,\" and \"Apple\" with\n\"Microsoft\" and you can see the same thing about to happen all over again.\nMicrosoft dominates the OS market, which makes them money and seems like a\ngreat idea for now. But cheaper and better OSes are available, and they\nare growingly popular in parts of the world that are not so saturated with\ncomputers as the US. Ten years from now, most of the world's computer\nusers may end up owning these cheaper OSes. But these OSes do not, for the\ntime being, run any Microsoft applications, and so these people will use\nsomething else. \n<\/p><p> To put it more directly: every time someone decides to use a\nnon-Microsoft OS, Microsoft's OS division, obviously, loses a customer.\nBut, as things stand now, Microsoft's applications division loses a\ncustomer too. This is not such a big deal as long as almost everyone uses\nMicrosoft OSes. But as soon as Windows' market share begins to slip, the\nmath starts to look pretty dismal for the people in Redmond. \n<\/p><p> This argument could be countered by saying that Microsoft could simply\nre-compile its applications to run under other OSes. But this strategy\ngoes against most normal corporate instincts. Again the case of Apple is\ninstructive. When things started to go south for Apple, they should have\nported their OS to cheap PC hardware.  But they didn't. Instead, they\ntried to make the most of their brilliant hardware, adding new features\nand expanding the product line. But this only had the effect of making\ntheir OS more dependent on these special hardware features, which made it\nworse for them in the end. \n<\/p><p> Likewise, when Microsoft's position in the OS world is threatened,\ntheir corporate instincts will tell them to pile more new features into\ntheir operating systems, and then re-jigger their software applications to\nexploit those special features.  But this will only have the effect of\nmaking their applications dependent on an OS with declining market share,\nand make it worse for them in the end. \n<\/p><p> The operating system market is a death-trap, a tar-pit, a slough of\ndespond. There are only two reasons to invest in Apple and Microsoft. (1)\neach of these companies is in what we would call a co-dependency\nrelationship with their customers. The customers Want To Believe, and\nApple and Microsoft know how to give them what they want. (2) each company\nworks very hard to add new features to their OSes, which works to secure\ncustomer loyalty, at least for a little while. \n<\/p><p> Accordingly, most of the remainder of this essay will be about those\ntwo topics. \n<\/p><p> THE TECHNOSPHERE\n<\/p><p> Unix is the only OS remaining whose GUI (a vast suite of code called\nthe X Windows System) is separate from the OS in the old sense of the\nphrase. This is to say that you can run Unix in pure command-line mode if\nyou want to, with no windows, icons, mouses, etc. whatsoever, and it will\nstill be Unix and capable of doing everything Unix is supposed to do. But\nthe other OSes: MacOS, the Windows family, and BeOS, have their GUIs\ntangled up with the old-fashioned OS functions to the extent that they\nhave to run in GUI mode, or else they are not really running. So it's no\nlonger really possible to think of GUIs as being distinct from the OS; \nthey're now an inextricable part of the OSes that they belong to--and they\nare by far the largest part, and by far the most expensive and difficult\npart to create. \n<\/p><p> There are only two ways to sell a product: price and features.  When\nOSes are free, OS companies cannot compete on price, and so they compete\non features. This means that they are always trying to outdo each other\nwriting code that, until recently, was not considered to be part of an OS\nat all: stuff like GUIs. This explains a lot about how these companies\nbehave. \n<\/p><p> It explains why Microsoft added a browser to their OS, for example. It\nis easy to get free browsers, just as to get free OSes. If browsers are\nfree, and OSes are free, it would seem that there is no way to make money\nfrom browsers or OSes. But if you can integrate a browser into the OS and\nthereby imbue both of them with new features, you have a salable product. \n<\/p><p> Setting aside, for the moment, the fact that this makes government\nanti-trust lawyers really mad, this strategy makes sense. At least, it\nmakes sense if you assume (as Microsoft's management appears to) that the\nOS has to be protected at all costs. The real question is whether every\nnew technological trend that comes down the pike ought to be used as a\ncrutch to maintain the OS's dominant position. Confronted with the Web\nphenomenon, Microsoft had to develop a really good web browser, and they\ndid.  But then they had a choice: they could have made that browser work\non many different OSes, which would give Microsoft a strong position in\nthe Internet world no matter what happened to their OS market share. Or\nthey could make the browser one with the OS, gambling that this would make\nthe OS look so modern and sexy that it would help to preserve their\ndominance in that market. The problem is that when Microsoft's OS position\nbegins to erode (and since it is currently at something like ninety\npercent, it can't go anywhere but down) it will drag everything else down\nwith it. \n<\/p><p> In your high school geology class you probably were taught that all\nlife on earth exists in a paper-thin shell called the biosphere, which is\ntrapped between thousands of miles of dead rock underfoot, and cold dead\nradioactive empty space above.  Companies that sell OSes exist in a sort\nof technosphere.  Underneath is technology that has already become free.\nAbove is technology that has yet to be developed, or that is too crazy and\nspeculative to be productized just yet. Like the Earth's biosphere, the\ntechnosphere is very thin compared to what is above and what is below. \n<\/p><p> But it moves a lot faster. In various parts of our world, it is\npossible to go and visit rich fossil beds where skeleton lies piled upon\nskeleton, recent ones on top and more ancient ones below. In theory they\ngo all the way back to the first single-celled organisms. And if you use\nyour imagination a bit, you can understand that, if you hang around long\nenough, you'll become fossilized there too, and in time some more advanced\norganism will become fossilized on top of you. \n<\/p><p> The fossil record--the La Brea Tar Pit--of software technology is the\nInternet. Anything that shows up there is free for the taking (possibly\nillegal, but free). Executives at companies like Microsoft must get used\nto the experience--unthinkable in other industries--of throwing millions\nof dollars into the development of new technologies, such as Web browsers,\nand then seeing the same or equivalent software show up on the Internet\ntwo years, or a year, or even just a few months, later. \n<\/p><p> By continuing to develop new technologies and add features onto their\nproducts they can keep one step ahead of the fossilization process, but on\ncertain days they must feel like mammoths caught at La Brea, using all\ntheir energies to pull their feet, over and over again, out of the sucking\nhot tar that wants to cover and envelop them. \n<\/p><p> Survival in this biosphere demands sharp tusks and heavy, stomping\nfeet at one end of the organization, and Microsoft famously has those. But\ntrampling the other mammoths into the tar can only keep you alive for so\nlong. The danger is that in their obsession with staying out of the fossil\nbeds, these companies will forget about what lies above the biosphere: the\nrealm of new technology. In other words, they must hang onto their\nprimitive weapons and crude competitive instincts, but also evolve\npowerful brains. This appears to be what Microsoft is doing with its\nresearch division, which has been hiring smart people right and left (Here\nI should mention that although I know, and socialize with, several people\nin that company's research division, we never talk about business issues\nand I have little to no idea what the hell they are up to. I have learned\nmuch more about Microsoft by using the Linux operating system than I ever\nwould have done by using Windows). \n<\/p><p> Never mind how Microsoft used to make money; today, it is making its\nmoney on a kind of temporal arbitrage. \"Arbitrage,\" in the usual sense,\nmeans to make money by taking advantage of differences in the price of\nsomething between different markets.  It is spatial, in other words, and\nhinges on the arbitrageur knowing what is going on simultaneously in\ndifferent places.  Microsoft is making money by taking advantage of\ndifferences in the price of technology in different times. Temporal\narbitrage, if I may coin a phrase, hinges on the arbitrageur knowing what\ntechnologies people will pay money for next year, and how soon afterwards\nthose same technologies will become free. What spatial and temporal\narbitrage have in common is that both hinge on the arbitrageur's being\nextremely well-informed; one about price gradients across space at a given\ntime, and the other about price gradients over time in a given place. \n<\/p><p> So Apple\/Microsoft shower new features upon their users almost daily,\nin the hopes that a steady stream of genuine technical innovations,\ncombined with the \"I want to believe\" phenomenon, will prevent their\ncustomers from looking across the road towards the cheaper and better OSes\nthat are available to them. The question is whether this makes sense in\nthe long run. If Microsoft is addicted to OSes as Apple is to hardware,\nthen they will bet the whole farm on their OSes, and tie all of their new\napplications and technologies to them. Their continued survival will then\ndepend on these two things: adding more features to their OSes so that\ncustomers will not switch to the cheaper alternatives, and maintaining the\nimage that, in some mysterious way, gives those customers the feeling that\nthey are getting something for their money. \n<\/p><p> The latter is a truly strange and interesting cultural phenomenon. \n<\/p><p> THE INTERFACE CULTURE\n<\/p><p> A few years ago I walked into a grocery store somewhere and was\npresented with the following tableau vivant: near the entrance a young\ncouple were standing in front of a large cosmetics display.  The man was\nstolidly holding a shopping basket between his hands while his mate raked\nblister-packs of makeup off the display and piled them in. Since then I've\nalways thought of that man as the personification of an interesting human\ntendency: not only are we not offended to be dazzled by manufactured\nimages, but we like it. We practically insist on it. We are eager to be\ncomplicit in our own dazzlement: to pay money for a theme park ride, vote\nfor a guy who's obviously lying to us, or stand there holding the basket\nas it's filled up with cosmetics. \n<\/p><p> I was in Disney World recently, specifically the part of it called the\nMagic Kingdom, walking up Main Street USA. This is a perfect gingerbready\nVictorian small town that culminates in a Disney castle. It was very\ncrowded; we shuffled rather than walked. Directly in front of me was a man\nwith a camcorder. It was one of the new breed of camcorders where instead\nof peering through a viewfinder you gaze at a flat-panel color screen\nabout the size of a playing card, which televises live coverage of\nwhatever the camcorder is seeing. He was holding the appliance close to\nhis face, so that it obstructed his view. Rather than go see a real small\ntown for free, he had paid money to see a pretend one, and rather than see\nit with the naked eye he was watching it on television. \n<\/p><p> And rather than stay home and read a book, I was watching him.\n<\/p><p> Americans' preference for mediated experiences is obvious enough, and\nI'm not going to keep pounding it into the ground. I'm not even going to\nmake snotty comments about it--after all, I was at Disney World as a\npaying customer. But it clearly relates to the colossal success of GUIs\nand so I have to talk about it some.  Disney does mediated experiences\nbetter than anyone. If they understood what OSes are, and why people use\nthem, they could crush Microsoft in a year or two. \n<\/p><p> In the part of Disney World called the Animal Kingdom there is a new\nattraction, slated to open in March 1999, called the Maharajah Jungle\nTrek. It was open for sneak previews when I was there. This is a complete\nstone-by-stone reproduction of a hypothetical ruin in the jungles of\nIndia. According to its backstory, it was built by a local rajah in the\n16th Century as a game reserve. He would go there with his princely guests\nto hunt Bengal tigers. As time went on it fell into disrepair and the\ntigers and monkeys took it over; eventually, around the time of India's\nindependence, it became a government wildlife reserve, now open to\nvisitors. \n<\/p><p> The place looks more like what I have just described than any actual\nbuilding you might find in India. All the stones in the broken walls are\nweathered as if monsoon rains had been trickling down them for centuries,\nthe paint on the gorgeous murals is flaked and faded just so, and Bengal\ntigers loll amid stumps of broken columns. Where modern repairs have been\nmade to the ancient structure, they've been done, not as Disney's\nengineers would do them, but as thrifty Indian janitors would--with hunks\nof bamboo and rust-spotted hunks of rebar. The rust is painted on, or\ncourse, and protected from real rust by a plastic clear-coat, but you\ncan't tell unless you get down on your knees. \n<\/p><p> In one place you walk along a stone wall with a series of old pitted\nfriezes carved into it. One end of the wall has broken off and settled\ninto the earth, perhaps because of some long-forgotten earthquake, and so\na broad jagged crack runs across a panel or two, but the story is still\nreadable: first, primordial chaos leads to a flourishing of many animal\nspecies.  Next, we see the Tree of Life surrounded by diverse animals.\nThis is an obvious allusion (or, in showbiz lingo, a tie-in) to the\ngigantic Tree of Life that dominates the center of Disney's Animal Kingdom\njust as the Castle dominates the Magic Kingdom or the Sphere does Epcot.\nBut it's rendered in historically correct style and could probably fool\nanyone who didn't have a Ph.D. in Indian art history. \n<\/p><p> The next panel shows a mustachioed H. sapiens chopping down the Tree\nof Life with a scimitar, and the animals fleeing every which way. The one\nafter that shows the misguided human getting walloped by a tidal wave,\npart of a latter-day Deluge presumably brought on by his stupidity. \n<\/p><p> The final panel, then, portrays the Sapling of Life beginning to grow\nback, but now Man has ditched the edged weapon and joined the other\nanimals in standing around to adore and praise it. \n<\/p><p> It is, in other words, a prophecy of the Bottleneck: the scenario,\ncommonly espoused among modern-day environmentalists, that the world faces\nan upcoming period of grave ecological tribulations that will last for a\nfew decades or centuries and end when we find a new harmonious modus\nvivendi with Nature. \n<\/p><p> Taken as a whole the frieze is a pretty brilliant piece of work. \nObviously it's not an ancient Indian ruin, and some person or people now\nliving deserve credit for it. But there are no signatures on the\nMaharajah's game reserve at Disney World. There are no signatures on\nanything, because it would ruin the whole effect to have long strings of\nproduction credits dangling from every custom-worn brick, as they do from\nHollywood movies. \n<\/p><p> Among Hollywood writers, Disney has the reputation of being a real\nwicked stepmother. It's not hard to see why. Disney is in the business of\nputting out a product of seamless illusion--a magic mirror that reflects\nthe world back better than it really is. But a writer is literally talking\nto his or her readers, not just creating an ambience or presenting them\nwith something to look at; and just as the command-line interface opens a\nmuch more direct and explicit channel from user to machine than the GUI,\nso it is with words, writer, and reader. \n<\/p><p> The word, in the end, is the only system of encoding thoughts--the\nonly medium--that is not fungible, that refuses to dissolve in the\ndevouring torrent of electronic media (the richer tourists at Disney World\nwear t-shirts printed with the names of famous designers, because designs\nthemselves can be bootlegged easily and with impunity. The only way to\nmake clothing that cannot be legally bootlegged is to print copyrighted\nand trademarked words on it; once you have taken that step, the clothing\nitself doesn't really matter, and so a t-shirt is as good as anything\nelse. T-shirts with expensive words on them are now the insignia of the\nupper class. T-shirts with cheap words, or no words at all, are for the\ncommoners). \n<\/p><p> But this special quality of words and of written communication would\nhave the same effect on Disney's product as spray-painted graffiti on a\nmagic mirror. So Disney does most of its communication without resorting\nto words, and for the most part, the words aren't missed. Some of Disney's\nolder properties, such as Peter Pan, Winnie the Pooh, and Alice in\nWonderland, came out of books. But the authors' names are rarely if ever\nmentioned, and you can't buy the original books at the Disney store. If\nyou could, they would all seem old and queer, like very bad knockoffs of\nthe purer, more authentic Disney versions. Compared to more recent\nproductions like Beauty and the Beast and Mulan, the Disney movies based\non these books (particularly Alice in Wonderland and Peter Pan) seem\ndeeply bizarre, and not wholly appropriate for children. That stands to\nreason, because Lewis Carroll and J.M. Barrie were very strange men, and\nsuch is the nature of the written word that their personal strangeness\nshines straight through all the layers of Disneyfication like x-rays\nthrough a wall. Probably for this very reason, Disney seems to have\nstopped buying books altogether, and now finds its themes and characters\nin folk tales, which have the lapidary, time-worn quality of the ancient\nbricks in the Maharajah's ruins. \n<\/p><p> If I can risk a broad generalization, most of the people who go to\nDisney World have zero interest in absorbing new ideas from books. Which\nsounds snide, but listen: they have no qualms about being presented with\nideas in other forms. Disney World is stuffed with environmental messages\nnow, and the guides at Animal Kingdom can talk your ear off about biology. \n<\/p><p> If you followed those tourists home, you might find art, but it would\nbe the sort of unsigned folk art that's for sale in Disney World's\nAfrican- and Asian-themed stores. In general they only seem comfortable\nwith media that have been ratified by great age, massive popular\nacceptance, or both. \n<\/p><p> In this world, artists are like the anonymous, illiterate stone\ncarvers who built the great cathedrals of Europe and then faded away into\nunmarked graves in the churchyard. The cathedral as a whole is awesome and\nstirring in spite, and possibly because, of the fact that we have no idea\nwho built it. When we walk through it we are communing not with individual\nstone carvers but with an entire culture. \n<\/p><p> Disney World works the same way. If you are an intellectual type, a\nreader or writer of books, the nicest thing you can say about this is that\nthe execution is superb. But it's easy to find the whole environment a\nlittle creepy, because something is missing:  the translation of all its\ncontent into clear explicit written words, the attribution of the ideas to\nspecific people. You can't argue with it. It seems as if a hell of a lot\nmight be being glossed over, as if Disney World might be putting one over\non us, and possibly getting away with all kinds of buried assumptions and\nmuddled thinking. \n<\/p><p> But this is precisely the same as what is lost in the transition from\nthe command-line interface to the GUI. \n<\/p><p> Disney and Apple\/Microsoft are in the same business:  short-circuiting\nlaborious, explicit verbal communication with expensively designed\ninterfaces. Disney is a sort of user interface unto itself--and more than\njust graphical. Let's call it a Sensorial Interface. It can be applied to\nanything in the world, real or imagined, albeit at staggering expense. \n<\/p><p> Why are we rejecting explicit word-based interfaces, and embracing\ngraphical or sensorial ones--a trend that accounts for the success of both\nMicrosoft and Disney? \n<\/p><p> Part of it is simply that the world is very complicated now--much more\ncomplicated than the hunter-gatherer world that our brains evolved to cope\nwith--and we simply can't handle all of the details. We have to delegate.\nWe have no choice but to trust some nameless artist at Disney or\nprogrammer at Apple or Microsoft to make a few choices for us, close off\nsome options, and give us a conveniently packaged executive summary. \n<\/p><p> But more importantly, it comes out of the fact that, during this\ncentury, intellectualism failed, and everyone knows it. In places like\nRussia and Germany, the common people agreed to loosen their grip on\ntraditional folkways, mores, and religion, and let the intellectuals run\nwith the ball, and they screwed everything up and turned the century into\nan abbatoir. Those wordy intellectuals used to be merely tedious; now they\nseem kind of dangerous as well. \n<\/p><p> We Americans are the only ones who didn't get creamed at some point\nduring all of this. We are free and prosperous because we have inherited\npolitical and values systems fabricated by a particular set of\neighteenth-century intellectuals who happened to get it right. But we have\nlost touch with those intellectuals, and with anything like\nintellectualism, even to the point of not reading books any more, though\nwe are literate. We seem much more comfortable with propagating those\nvalues to future generations nonverbally, through a process of being\nsteeped in media.  Apparently this actually works to some degree, for\npolice in many lands are now complaining that local arrestees are\ninsisting on having their Miranda rights read to them, just like perps in\nAmerican TV cop shows. When it's explained to them that they are in a\ndifferent country, where those rights do not exist, they become outraged.\nStarsky and Hutch reruns, dubbed into diverse languages, may turn out, in\nthe long run, to be a greater force for human rights than the Declaration\nof Independence. \n<\/p><p> A huge, rich, nuclear-tipped culture that propagates its core values\nthrough media steepage seems like a bad idea. There is an obvious risk of\nrunning astray here. Words are the only immutable medium we have, which is\nwhy they are the vehicle of choice for extremely important concepts like\nthe Ten Commandments, the Koran, and the Bill of Rights. Unless the\nmessages conveyed by our media are somehow pegged to a fixed, written set\nof precepts, they can wander all over the place and possibly dump loads of\ncrap into people's minds. \n<\/p><p> Orlando used to have a military installation called McCoy Air Force\nBase, with long runways from which B-52s could take off and reach Cuba, or\njust about anywhere else, with loads of nukes. But now McCoy has been\nscrapped and repurposed. It has been absorbed into Orlando's civilian\nairport. The long runways are being used to land 747-loads of tourists\nfrom Brazil, Italy, Russia and Japan, so that they can come to Disney\nWorld and steep in our media for a while. \n<\/p><p> To traditional cultures, especially word-based ones such as Islam,\nthis is infinitely more threatening than the B-52s ever were. It is\nobvious, to everyone outside of the United States, that our\narch-buzzwords, multiculturalism and diversity, are false fronts that are\nbeing used (in many cases unwittingly) to conceal a global trend to\neradicate cultural differences. The basic tenet of multiculturalism (or\n\"honoring diversity\" or whatever you want to call it) is that people need\nto stop judging each other-to stop asserting (and, eventually, to stop\nbelieving)  that this is right and that is wrong, this true and that\nfalse, one thing ugly and another thing beautiful, that God exists and has\nthis or that set of qualities. \n<\/p><p> The lesson most people are taking home from the Twentieth Century is\nthat, in order for a large number of different cultures to coexist\npeacefully on the globe (or even in a neighborhood) it is necessary for\npeople to suspend judgment in this way. Hence (I would argue) our\nsuspicion of, and hostility towards, all authority figures in modern\nculture. As David Foster Wallace has explained in his essay \"E Unibus\nPluram,\" this is the fundamental message of television; it is the message\nthat people take home, anyway, after they have steeped in our media long\nenough. It's not expressed in these highfalutin terms, of course. It comes\nthrough as the presumption that all authority figures--teachers, generals,\ncops, ministers, politicians--are hypocritical buffoons, and that hip\njaded coolness is the only way to be. \n<\/p><p> The problem is that once you have done away with the ability to make\njudgments as to right and wrong, true and false, etc., there's no real\nculture left. All that remains is clog dancing and macrame. The ability to\nmake judgments, to believe things, is the entire it point of having a\nculture. I think this is why guys with machine guns sometimes pop up in\nplaces like Luxor, and begin pumping bullets into Westerners. They\nperfectly understand the lesson of McCoy Air Force Base. When their sons\ncome home wearing Chicago Bulls caps with the bills turned sideways, the\ndads go out of their minds. \n<\/p><p> The global anti-culture that has been conveyed into every cranny of\nthe world by television is a culture unto itself, and by the standards of\ngreat and ancient cultures like Islam and France, it seems grossly\ninferior, at least at first. The only good thing you can say about it is\nthat it makes world wars and Holocausts less likely--and that is actually\na pretty good thing! \n<\/p><p> The only real problem is that anyone who has no culture, other than\nthis global monoculture, is completely screwed. Anyone who grows up\nwatching TV, never sees any religion or philosophy, is raised in an\natmosphere of moral relativism, learns about civics from watching bimbo\neruptions on network TV news, and attends a university where\npostmodernists vie to outdo each other in demolishing traditional notions\nof truth and quality, is going to come out into the world as one pretty\nfeckless human being.  And--again--perhaps the goal of all this is to make\nus feckless so we won't nuke each other. \n<\/p><p> On the other hand, if you are raised within some specific culture, you\nend up with a basic set of tools that you can use to think about and\nunderstand the world. You might use those tools to reject the culture you\nwere raised in, but at least you've got some tools. \n<\/p><p> In this country, the people who run things--who populate major law\nfirms and corporate boards--understand all of this at some level. They pay\nlip service to multiculturalism and diversity and non-judgmentalness, but\nthey don't raise their own children that way. I have highly educated,\ntechnically sophisticated friends who have moved to small towns in Iowa to\nlive and raise their children, and there are Hasidic Jewish enclaves in\nNew York where large numbers of kids are being brought up according to\ntraditional beliefs. Any suburban community might be thought of as a place\nwhere people who hold certain (mostly implicit)  beliefs go to live among\nothers who think the same way. \n<\/p><p> And not only do these people feel some responsibility to their own\nchildren, but to the country as a whole. Some of the upper class are vile\nand cynical, of course, but many spend at least part of their time\nfretting about what direction the country is going in, and what\nresponsibilities they have. And so issues that are important to\nbook-reading intellectuals, such as global environmental collapse,\neventually percolate through the porous buffer of mass culture and show up\nas ancient Hindu ruins in Orlando. \n<\/p><p> You may be asking: what the hell does all this have to do with\noperating systems? As I've explained, there is no way to explain the\ndomination of the OS market by Apple\/Microsoft without looking to cultural\nexplanations, and so I can't get anywhere, in this essay, without first\nletting you know where I'm coming from vis-a-vis contemporary culture. \n<\/p><p> Contemporary culture is a two-tiered system, like the Morlocks and the\nEloi in H.G. Wells's The Time Machine, except that it's been turned upside\ndown. In The Time Machine the Eloi were an effete upper class, supported\nby lots of subterranean Morlocks who kept the technological wheels\nturning. But in our world it's the other way round. The Morlocks are in\nthe minority, and they are running the show, because they understand how\neverything works. The much more numerous Eloi learn everything they know\nfrom being steeped from birth in electronic media directed and controlled\nby book-reading Morlocks. So many ignorant people could be dangerous if\nthey got pointed in the wrong direction, and so we've evolved a popular\nculture that is (a) almost unbelievably infectious and (b) neuters every\nperson who gets infected by it, by rendering them unwilling to make\njudgments and incapable of taking stands. \n<\/p><p> Morlocks, who have the energy and intelligence to comprehend details,\ngo out and master complex subjects and produce Disney-like Sensorial\nInterfaces so that Eloi can get the gist without having to strain their\nminds or endure boredom. Those Morlocks will go to India and tediously\nexplore a hundred ruins, then come home and built sanitary bug-free\nversions: highlight films, as it were. This costs a lot, because Morlocks\ninsist on good coffee and first-class airline tickets, but that's no\nproblem because Eloi like to be dazzled and will gladly pay for it all. \n<\/p><p> Now I realize that most of this probably sounds snide and bitter to\nthe point of absurdity: your basic snotty intellectual throwing a tantrum\nabout those unlettered philistines. As if I were a self-styled Moses,\ncoming down from the mountain all alone, carrying the stone tablets\nbearing the Ten Commandments carved in immutable stone--the original\ncommand-line interface--and blowing his stack at the weak, unenlightened\nHebrews worshipping images. Not only that, but it sounds like I'm pumping\nsome sort of conspiracy theory. \n<\/p><p> But that is not where I'm going with this. The situation I describe,\nhere, could be bad, but doesn't have to be bad and isn't necessarily bad\nnow: \n<\/p><ul><li>It simply is the case that we are way too busy, nowadays, to\ncomprehend everything in detail. And it's better to comprehend it dimly,\nthrough an interface, than not at all.  Better for ten million Eloi to go\non the Kilimanjaro Safari at Disney World than for a thousand\ncardiovascular surgeons and mutual fund managers to go on \"real\" ones in\nKenya. \n<\/li><li>The boundary between these two classes is more porous than I've\nmade it sound. I'm always running into regular dudes--construction\nworkers, auto mechanics, taxi drivers, galoots in general--who were\nlargely aliterate until something made it necessary for them to become\nreaders and start actually thinking about things. Perhaps they had to come\nto grips with alcoholism, perhaps they got sent to jail, or came down with\na disease, or suffered a crisis in religious faith, or simply got bored.\nSuch people can get up to speed on particular subjects quite rapidly.\nSometimes their lack of a broad education makes them over-apt to go off on\nintellectual wild goose chases, but, hey, at least a wild goose chase\ngives you some exercise. \n<\/li><li>The spectre of a polity controlled by the fads and whims of voters\nwho actually believe that there are significant differences between Bud\nLite and Miller Lite, and who think that professional wrestling is for\nreal, is naturally alarming to people who don't. But then countries\ncontrolled via the command-line interface, as it were, by double-domed\nintellectuals, be they religious or secular, are generally miserable\nplaces to live. \n<\/li><li>Sophisticated people deride Disneyesque entertainments as pat and\nsaccharine, but, hey, if the result of that is to instill basically warm\nand sympathetic reflexes, at a preverbal level, into hundreds of millions\nof unlettered media-steepers, then how bad can it be? We killed a lobster\nin our kitchen last night and my daughter cried for an hour.  The\nJapanese, who used to be just about the fiercest people on earth, have\nbecome infatuated with cuddly adorable cartoon characters. \n<\/li><li>My own family--the people I know best--is divided about evenly\nbetween people who will probably read this essay and people who almost\ncertainly won't, and I can't say for sure that one group is necessarily\nwarmer, happier, or better-adjusted than the other. \n<\/li><\/ul><p> MORLOCKS AND ELOI AT THE KEYBOARD\n<\/p><p> Back in the days of the command-line interface, users were all\nMorlocks who had to convert their thoughts into alphanumeric symbols and\ntype them in, a grindingly tedious process that stripped away all\nambiguity, laid bare all hidden assumptions, and cruelly punished laziness\nand imprecision. Then the interface-makers went to work on their GUIs, and\nintroduced a new semiotic layer between people and machines. People who\nuse such systems have abdicated the responsibility, and surrendered the\npower, of sending bits directly to the chip that's doing the arithmetic,\nand handed that responsibility and power over to the OS. This is tempting\nbecause giving clear instructions, to anyone or anything, is difficult. We\ncannot do it without thinking, and depending on the complexity of the\nsituation, we may have to think hard about abstract things, and consider\nany number of ramifications, in order to do a good job of it. For most of\nus, this is hard work. We want things to be easier. How badly we want it\ncan be measured by the size of Bill Gates's fortune. \n<\/p><p> The OS has (therefore) become a sort of intellectual labor-saving\ndevice that tries to translate humans' vaguely expressed intentions into\nbits. In effect we are asking our computers to shoulder responsibilities\nthat have always been considered the province of human beings--we want\nthem to understand our desires, to anticipate our needs, to foresee\nconsequences, to make connections, to handle routine chores without being\nasked, to remind us of what we ought to be reminded of while filtering out\nnoise. \n<\/p><p> At the upper (which is to say, closer to the user) levels, this is\ndone through a set of conventions--menus, buttons, and so on.  These work\nin the sense that analogies work: they help Eloi understand abstract or\nunfamiliar concepts by likening them to something known. But the loftier\nword \"metaphor\" is used. \n<\/p><p> The overarching concept of the MacOS was the \"desktop metaphor\"  and\nit subsumed any number of lesser (and frequently conflicting, or at least\nmixed) metaphors. Under a GUI, a file (frequently called \"document\") is\nmetaphrased as a window on the screen (which is called a \"desktop\"). The\nwindow is almost always too small to contain the document and so you \"move\naround,\" or, more pretentiously, \"navigate\" in the document by \"clicking\nand dragging\" the \"thumb\" on the \"scroll bar.\" When you \"type\" (using a\nkeyboard) or \"draw\" (using a \"mouse\") into the \"window\" or use pull-down\n\"menus\" and \"dialog boxes\" to manipulate its contents, the results of your\nlabors get stored (at least in theory) in a \"file,\" and later you can pull\nthe same information back up into another \"window.\" When you don't want it\nanymore, you \"drag\" it into the \"trash.\" \n<\/p><p> There is massively promiscuous metaphor-mixing going on here, and I\ncould deconstruct it 'til the cows come home, but I won't.  Consider only\none word: \"document.\" When we document something in the real world, we\nmake fixed, permanent, immutable records of it. But computer documents are\nvolatile, ephemeral constellations of data. Sometimes (as when you've just\nopened or saved them) the document as portrayed in the window is identical\nto what is stored, under the same name, in a file on the disk, but other\ntimes (as when you have made changes without saving them) it is completely\ndifferent. In any case, every time you hit \"Save\" you annihilate the\nprevious version of the \"document\" and replace it with whatever happens to\nbe in the window at the moment. So even the word \"save\" is being used in a\nsense that is grotesquely misleading---\"destroy one version, save another\"\nwould be more accurate. \n<\/p><p> Anyone who uses a word processor for very long inevitably has the\nexperience of putting hours of work into a long document and then losing\nit because the computer crashes or the power goes out.  Until the moment\nthat it disappears from the screen, the document seems every bit as solid\nand real as if it had been typed out in ink on paper. But in the next\nmoment, without warning, it is completely and irretrievably gone, as if it\nhad never existed.  The user is left with a feeling of disorientation (to\nsay nothing of annoyance) stemming from a kind of metaphor shear--you\nrealize that you've been living and thinking inside of a metaphor that is\nessentially bogus. \n<\/p><p> So GUIs use metaphors to make computing easier, but they are bad\nmetaphors. Learning to use them is essentially a word game, a process of\nlearning new definitions of words like \"window\" and \"document\" and \"save\"\nthat are different from, and in many cases almost diametrically opposed\nto, the old. Somewhat improbably, this has worked very well, at least from\na commercial standpoint, which is to say that Apple\/Microsoft have made a\nlot of money off of it. All of the other modern operating systems have\nlearned that in order to be accepted by users they must conceal their\nunderlying gutwork beneath the same sort of spackle. This has some\nadvantages: if you know how to use one GUI operating system, you can\nprobably work out how to use any other in a few minutes.  Everything works\na little differently, like European plumbing--but with some fiddling\naround, you can type a memo or surf the web. \n<\/p><p> Most people who shop for OSes (if they bother to shop at all) are\ncomparing not the underlying functions but the superficial look and feel.\nThe average buyer of an OS is not really paying for, and is not especially\ninterested in, the low-level code that allocates memory or writes bytes\nonto the disk. What we're really buying is a system of metaphors.\nAnd--much more important--what we're buying into is the underlying\nassumption that metaphors are a good way to deal with the world. \n<\/p><p> Recently a lot of new hardware has become available that gives\ncomputers numerous interesting ways of affecting the real world:  making\npaper spew out of printers, causing words to appear on screens thousands\nof miles away, shooting beams of radiation through cancer patients,\ncreating realistic moving pictures of the Titanic. Windows is now used as\nan OS for cash registers and bank tellers' terminals. My satellite TV\nsystem uses a sort of GUI to change channels and show program guides.\nModern cellular telephones have a crude GUI built into a tiny LCD screen.\nEven Legos now have a GUI: you can buy a Lego set called Mindstorms that\nenables you to build little Lego robots and program them through a GUI on\nyour computer. \n<\/p><p> So we are now asking the GUI to do a lot more than serve as a\nglorified typewriter. Now we want to become a generalized tool for dealing\nwith reality. This has become a bonanza for companies that make a living\nout of bringing new technology to the mass market. \n<\/p><p> Obviously you cannot sell a complicated technological system to people\nwithout some sort of interface that enables them to use it. The internal\ncombustion engine was a technological marvel in its day, but useless as a\nconsumer good until a clutch, transmission, steering wheel and throttle\nwere connected to it.  That odd collection of gizmos, which survives to\nthis day in every car on the road, made up what we would today call a user\ninterface. But if cars had been invented after Macintoshes, carmakers\nwould not have bothered to gin up all of these arcane devices. We would\nhave a computer screen instead of a dashboard, and a mouse (or at best a\njoystick) instead of a steering wheel, and we'd shift gears by pulling\ndown a menu: \n<\/p><pre>\nPARK\n---\nREVERSE\n---\nNEUTRAL\n----\n3\n2\n1\n---\nHelp...\n<\/pre>\n<p> A few lines of computer code can thus be made to substitute for any\nimaginable mechanical interface. The problem is that in many cases the\nsubstitute is a poor one. Driving a car through a GUI would be a miserable\nexperience. Even if the GUI were perfectly bug-free, it would be\nincredibly dangerous, because menus and buttons simply can't be as\nresponsive as direct mechanical controls. My friend's dad, the gentleman\nwho was restoring the MGB, never would have bothered with it if it had\nbeen equipped with a GUI. It wouldn't have been any fun. \n<\/p><p> The steering wheel and gearshift lever were invented during an era\nwhen the most complicated technology in most homes was a butter churn.\nThose early carmakers were simply lucky, in that they could dream up\nwhatever interface was best suited to the task of driving an automobile,\nand people would learn it.  Likewise with the dial telephone and the AM\nradio. By the time of the Second World War, most people knew several\ninterfaces: they could not only churn butter but also drive a car, dial a\ntelephone, turn on a radio, summon flame from a cigarette lighter, and\nchange a light bulb. \n<\/p><p> But now every little thing--wristwatches, VCRs, stoves--is jammed with\nfeatures, and every feature is useless without an interface.  If you are\nlike me, and like most other consumers, you have never used ninety percent\nof the available features on your microwave oven, VCR, or cellphone. You\ndon't even know that these features exist. The small benefit they might\nbring you is outweighed by the sheer hassle of having to learn about them.\nThis has got to be a big problem for makers of consumer goods, because\nthey can't compete without offering features. \n<\/p><p> It's no longer acceptable for engineers to invent a wholly novel user\ninterface for every new product, as they did in the case of the\nautomobile, partly because it's too expensive and partly because ordinary\npeople can only learn so much. If the VCR had been invented a hundred\nyears ago, it would have come with a thumbwheel to adjust the tracking and\na gearshift to change between forward and reverse and a big cast-iron\nhandle to load or to eject the cassettes. It would have had a big analog\nclock on the front of it, and you would have set the time by moving the\nhands around on the dial. But because the VCR was invented when it\nwas--during a sort of awkward transitional period between the era of\nmechanical interfaces and GUIs--it just had a bunch of pushbuttons on the\nfront, and in order to set the time you had to push the buttons in just\nthe right way. This must have seemed reasonable enough to the engineers\nresponsible for it, but to many users it was simply impossible. Thus the\nfamous blinking 12:00 that appears on so many VCRs. Computer people call\nthis \"the blinking twelve problem\". When they talk about it, though, they\nusually aren't talking about VCRs. \n<\/p><p> Modern VCRs usually have some kind of on-screen programming, which\nmeans that you can set the time and control other features through a sort\nof primitive GUI. GUIs have virtual pushbuttons too, of course, but they\nalso have other types of virtual controls, like radio buttons, checkboxes,\ntext entry boxes, dials, and scrollbars. Interfaces made out of these\ncomponents seem to be a lot easier, for many people, than pushing those\nlittle buttons on the front of the machine, and so the blinking 12:00\nitself is slowly disappearing from America's living rooms.  The blinking\ntwelve problem has moved on to plague other technologies. \n<\/p><p> So the GUI has gone beyond being an interface to personal computers,\nand become a sort of meta-interface that is pressed into service for every\nnew piece of consumer technology. It is rarely an ideal fit, but having an\nideal, or even a good interface is no longer the priority; the important\nthing now is having some kind of interface that customers will actually\nuse, so that manufacturers can claim, with a straight face, that they are\noffering new features. \n<\/p><p> We want GUIs largely because they are convenient and because they are\neasy-- or at least the GUI makes it seem that way Of course, nothing is\nreally easy and simple, and putting a nice interface on top of it does not\nchange that fact. A car controlled through a GUI would be easier to drive\nthan one controlled through pedals and steering wheel, but it would be\nincredibly dangerous. \n<\/p><p> By using GUIs all the time we have insensibly bought into a premise\nthat few people would have accepted if it were presented to them bluntly:\nnamely, that hard things can be made easy, and complicated things simple,\nby putting the right interface on them. In order to understand how bizarre\nthis is, imagine that book reviews were written according to the same\nvalues system that we apply to user interfaces: \"The writing in this book\nis marvelously simple-minded and glib; the author glosses over complicated\nsubjects and employs facile generalizations in almost every sentence.\nReaders rarely have to think, and are spared all of the difficulty and\ntedium typically involved in reading old-fashioned books.\" As long as we\nstick to simple operations like setting the clocks on our VCRs, this is\nnot so bad. But as we try to do more ambitious things with our\ntechnologies, we inevitably run into the problem of: \n<\/p><p> METAPHOR SHEAR\n<\/p><p> I began using Microsoft Word as soon as the first version was released\naround 1985. After some initial hassles I found it to be a better tool\nthan MacWrite, which was its only competition at the time. I wrote a lot\nof stuff in early versions of Word, storing it all on floppies, and\ntransferred the contents of all my floppies to my first hard drive, which\nI acquired around 1987.  As new versions of Word came out I faithfully\nupgraded, reasoning that as a writer it made sense for me to spend a\ncertain amount of money on tools. \n<\/p><p> Sometime in the mid-1980's I attempted to open one of my old,\ncirca-1985 Word documents using the version of Word then current:  6.0 It\ndidn't work. Word 6.0 did not recognize a document created by an earlier\nversion of itself. By opening it as a text file, I was able to recover the\nsequences of letters that made up the text of the document. My words were\nstill there. But the formatting had been run through a log chipper--the\nwords I'd written were interrupted by spates of empty rectangular boxes\nand gibberish. \n<\/p><p> Now, in the context of a business (the chief market for Word)  this\nsort of thing is only an annoyance--one of the routine hassles that go\nalong with using computers. It's easy to buy little file converter\nprograms that will take care of this problem. But if you are a writer\nwhose career is words, whose professional identity is a corpus of written\ndocuments, this kind of thing is extremely disquieting. There are very few\nfixed assumptions in my line of work, but one of them is that once you\nhave written a word, it is written, and cannot be unwritten. The ink\nstains the paper, the chisel cuts the stone, the stylus marks the clay,\nand something has irrevocably happened (my brother-in-law is a theologian\nwho reads 3250-year-old cuneiform tablets--he can recognize the\nhandwriting of particular scribes, and identify them by name). But\nword-processing software--particularly the sort that employs special,\ncomplex file formats--has the eldritch power to unwrite things. A small\nchange in file formats, or a few twiddled bits, and months' or years'\nliterary output can cease to exist. \n<\/p><p> Now this was technically a fault in the application (Word 6.0 for the\nMacintosh) not the operating system (MacOS 7 point something)  and so the\ninitial target of my annoyance was the people who were responsible for\nWord. But. On the other hand, I could have chosen the \"save as text\"\noption in Word and saved all of my documents as simple telegrams, and this\nproblem would not have arisen.  Instead I had allowed myself to be seduced\nby all of those flashy formatting options that hadn't even existed until\nGUIs had come along to make them practicable. I had gotten into the habit\nof using them to make my documents look pretty (perhaps prettier than they\ndeserved to look; all of the old documents on those floppies turned out to\nbe more or less crap). Now I was paying the price for that\nself-indulgence. Technology had moved on and found ways to make my\ndocuments look even prettier, and the consequence of it was that all old\nugly documents had ceased to exist. \n<\/p><p> It was--if you'll pardon me for a moment's strange little fantasy--as\nif I'd gone to stay at some resort, some exquisitely designed and\nart-directed hotel, placing myself in the hands of past masters of the\nSensorial Interface, and had sat down in my room and written a story in\nballpoint pen on a yellow legal pad, and when I returned from dinner,\ndiscovered that the maid had taken my work away and left behind in its\nplace a quill pen and a stack of fine parchment--explaining that the room\nlooked ever so much finer this way, and it was all part of a routine\nupgrade.  But written on these sheets of paper, in flawless penmanship,\nwere long sequences of words chosen at random from the dictionary.\nAppalling, sure, but I couldn't really lodge a complaint with the\nmanagement, because by staying at this resort I had given my consent to\nit. I had surrendered my Morlock credentials and become an Eloi. \n<\/p><p> LINUX\n<\/p><p> During the late 1980's and early 1990's I spent a lot of time\nprogramming Macintoshes, and eventually decided for fork over several\nhundred dollars for an Apple product called the Macintosh Programmer's\nWorkshop, or MPW. MPW had competitors, but it was unquestionably the\npremier software development system for the Mac. It was what Apple's own\nengineers used to write Macintosh code. Given that MacOS was far more\ntechnologically advanced, at the time, than its competition, and that\nLinux did not even exist yet, and given that this was the actual program\nused by Apple's world-class team of creative engineers, I had high\nexpectations.  It arrived on a stack of floppy disks about a foot high,\nand so there was plenty of time for my excitement to build during the\nendless installation process. The first time I launched MPW, I was\nprobably expecting some kind of touch-feely multimedia showcase. Instead\nit was austere, almost to the point of being intimidating. It was a\nscrolling window into which you could type simple, unformatted text. The\nsystem would then interpret these lines of text as commands, and try to\nexecute them. \n<\/p><p> It was, in other words, a glass teletype running a command line\ninterface. It came with all sorts of cryptic but powerful commands, which\ncould be invoked by typing their names, and which I learned to use only\ngradually. It was not until a few years later, when I began messing around\nwith Unix, that I understood that the command line interface embodied in\nMPW was a re-creation of Unix. \n<\/p><p> In other words, the first thing that Apple's hackers had done when\nthey'd got the MacOS up and running--probably even before they'd gotten it\nup and running--was to re-create the Unix interface, so that they would be\nable to get some useful work done. At the time, I simply couldn't get my\nmind around this, but: as far as Apple's hackers were concerned, the Mac's\nvaunted Graphical User Interface was an impediment, something to be\ncircumvented before the little toaster even came out onto the market. \n<\/p><p> Even before my Powerbook crashed and obliterated my big file in July\n1995, there had been danger signs. An old college buddy of mine, who\nstarts and runs high-tech companies in Boston, had developed a commercial\nproduct using Macintoshes as the front end. Basically the Macs were\nhigh-performance graphics terminals, chosen for their sweet user\ninterface, giving users access to a large database of graphical\ninformation stored on a network of much more powerful, but less\nuser-friendly, computers. This fellow was the second person who turned me\non to Macintoshes, by the way, and through the mid-1980's we had shared\nthe thrill of being high-tech cognoscenti, using superior Apple technology\nin a world of DOS-using knuckleheads. Early versions of my friend's system\nhad worked well, he told me, but when several machines joined the network,\nmysterious crashes began to occur; sometimes the whole network would just\nfreeze. It was one of those bugs that could not be reproduced easily.\nFinally they figured out that these network crashes were triggered\nwhenever a user, scanning the menus for a particular item, held down the\nmouse button for more than a couple of seconds. \n<\/p><p> Fundamentally, the MacOS could only do one thing at a time.  Drawing a\nmenu on the screen is one thing. So when a menu was pulled down, the\nMacintosh was not capable of doing anything else until that indecisive\nuser released the button. \n<\/p><p> This is not such a bad thing in a single-user, single-process machine\n(although it's a fairly bad thing), but it's no good in a machine that is\non a network, because being on a network implies some kind of continual\nlow-level interaction with other machines.  By failing to respond to the\nnetwork, the Mac caused a network-wide crash. \n<\/p><p> In order to work with other computers, and with networks, and with\nvarious different types of hardware, an OS must be incomparably more\ncomplicated and powerful than either MS-DOS or the original MacOS. The\nonly way of connecting to the Internet that's worth taking seriously is\nPPP, the Point-to-Point Protocol, which (never mind the details) makes\nyour computer--temporarily--a full-fledged member of the Global Internet,\nwith its own unique address, and various privileges, powers, and\nresponsibilities appertaining thereunto. Technically it means your machine\nis running the TCP\/IP protocol, which, to make a long story short,\nrevolves around sending packets of data back and forth, in no particular\norder, and at unpredictable times, according to a clever and elegant set\nof rules. But sending a packet of data is one thing, and so an OS that can\nonly do one thing at a time cannot simultaneously be part of the Internet\nand do anything else. When TCP\/IP was invented, running it was an honor\nreserved for Serious Computers--mainframes and high-powered minicomputers\nused in technical and commercial settings--and so the protocol is\nengineered around the assumption that every computer using it is a serious\nmachine, capable of doing many things at once. Not to put too fine a point\non it, a Unix machine. Neither MacOS nor MS-DOS was originally built with\nthat in mind, and so when the Internet got hot, radical changes had to be\nmade. \n<\/p><p> When my Powerbook broke my heart, and when Word stopped recognizing my\nold files, I jumped to Unix. The obvious alternative to MacOS would have\nbeen Windows. I didn't really have anything against Microsoft, or Windows.\nBut it was pretty obvious, now, that old PC operating systems were\noverreaching, and showing the strain, and, perhaps, were best avoided\nuntil they had learned to walk and chew gum at the same time. \n<\/p><p> The changeover took place on a particular day in the summer of 1995. I\nhad been San Francisco for a couple of weeks, using my PowerBook to work\non a document. The document was too big to fit onto a single floppy, and\nso I hadn't made a backup since leaving home. The PowerBook crashed and\nwiped out the entire file. \n<\/p><p> It happened just as I was on my way out the door to visit a company\ncalled Electric Communities, which in those days was in Los Altos. I took\nmy PowerBook with me. My friends at Electric Communities were Mac users\nwho had all sorts of utility software for unerasing files and recovering\nfrom disk crashes, and I was certain I could get most of the file back. \n<\/p><p> As it turned out, two different Mac crash recovery utilities were\nunable to find any trace that my file had ever existed. It was completely\nand systematically wiped out. We went through that hard disk block by\nblock and found disjointed fragments of countless old, discarded,\nforgotten files, but none of what I wanted. The metaphor shear was\nespecially brutal that day. It was sort of like watching the girl you've\nbeen in love with for ten years get killed in a car wreck, and then\nattending her autopsy, and learning that underneath the clothes and makeup\nshe was just flesh and blood. \n<\/p><p> I must have been reeling around the offices of Electric Communities in\nsome kind of primal Jungian fugue, because at this moment three weirdly\nsynchronistic things happened. \n<\/p><p> (1) Randy Farmer, a co-founder of the company, came in for a quick\nvisit along with his family--he was recovering from back surgery at the\ntime. He had some hot gossip: \"Windows 95 mastered today.\" What this meant\nwas that Microsoft's new operating system had, on this day, been placed on\na special compact disk known as a golden master, which would be used to\nstamp out a jintillion copies in preparation for its thunderous release a\nfew weeks later. This news was received peevishly by the staff of Electric\nCommunities, including one whose office door was plastered with the usual\nassortment of cartoons and novelties, e.g. \n<\/p><p> (2) a copy of a Dilbert cartoon in which Dilbert, the long-suffering\ncorporate software engineer, encounters a portly, bearded, hairy man of a\ncertain age--a bit like Santa Claus, but darker, with a certain edge about\nhim. Dilbert recognizes this man, based upon his appearance and affect, as\na Unix hacker, and reacts with a certain mixture of nervousness, awe, and\nhostility.  Dilbert jabs weakly at the disturbing interloper for a couple\nof frames; the Unix hacker listens with a kind of infuriating, beatific\ncalm, then, in the last frame, reaches into his pocket.  \"Here's a nickel,\nkid,\" he says, \"go buy yourself a real computer.\" \n<\/p><p> (3) the owner of the door, and the cartoon, was one Doug Barnes. \nBarnes was known to harbor certain heretical opinions on the subject of\noperating systems. Unlike most Bay Area techies who revered the Macintosh,\nconsidering it to be a true hacker's machine, Barnes was fond of pointing\nout that the Mac, with its hermetically sealed architecture, was actually\nhostile to hackers, who are prone to tinkering and dogmatic about\nopenness.  By contrast, the IBM-compatible line of machines, which can\neasily be taken apart and plugged back together, was much more hackable. \n<\/p><p> So when I got home I began messing around with Linux, which is one of\nmany, many different concrete implementations of the abstract, Platonic\nideal called Unix. I was not looking forward to changing over to a new OS,\nbecause my credit cards were still smoking from all the money I'd spent on\nMac hardware over the years. But Linux's great virtue was, and is, that it\nwould run on exactly the same sort of hardware as the Microsoft\nOSes--which is to say, the cheapest hardware in existence. As if to\ndemonstrate why this was a great idea, I was, within a week or two of\nreturning home, able to get my hand on a then-decent computer (a 33-MHz\n486 box) for free, because I knew a guy who worked in an office where they\nwere simply being thrown away. Once I got it home, I yanked the hood off,\nstuck my hands in, and began switching cards around. If something didn't\nwork, I went to a used-computer outlet and pawed through a bin full of\ncomponents and bought a new card for a few bucks. \n<\/p><p> The availability of all this cheap but effective hardware was an\nunintended consequence of decisions that had been made more than a decade\nearlier by IBM and Microsoft. When Windows came out, and brought the GUI\nto a much larger market, the hardware regime changed: the cost of color\nvideo cards and high-resolution monitors began to drop, and is dropping\nstill. This free-for-all approach to hardware meant that Windows was\nunavoidably clunky compared to MacOS. But the GUI brought computing to\nsuch a vast audience that volume went way up and prices collapsed.\nMeanwhile Apple, which so badly wanted a clean, integrated OS with video\nneatly integrated into processing hardware, had fallen far behind in\nmarket share, at least partly because their beautiful hardware cost so\nmuch. \n<\/p><p> But the price that we Mac owners had to pay for superior aesthetics\nand engineering was not merely a financial one. There was a cultural price\ntoo, stemming from the fact that we couldn't open up the hood and mess\naround with it. Doug Barnes was right.  Apple, in spite of its reputation\nas the machine of choice of scruffy, creative hacker types, had actually\ncreated a machine that discouraged hacking, while Microsoft, viewed as a\ntechnological laggard and copycat, had created a vast, disorderly parts\nbazaar--a primordial soup that eventually self-assembled into Linux. \n<\/p><p> THE HOLE HAWG OF OPERATING SYSTEMS\n<\/p><p> Unix has always lurked provocatively in the background of the\noperating system wars, like the Russian Army. Most people know it only by\nreputation, and its reputation, as the Dilbert cartoon suggests, is mixed.\nBut everyone seems to agree that if it could only get its act together and\nstop surrendering vast tracts of rich agricultural land and hundreds of\nthousands of prisoners of war to the onrushing invaders, it could stomp\nthem (and all other opposition) flat. \n<\/p><p> It is difficult to explain how Unix has earned this respect without\ngoing into mind-smashing technical detail. Perhaps the gist of it can be\nexplained by telling a story about drills. \n<\/p><p> The Hole Hawg is a drill made by the Milwaukee Tool Company. If you\nlook in a typical hardware store you may find smaller Milwaukee drills but\nnot the Hole Hawg, which is too powerful and too expensive for homeowners.\nThe Hole Hawg does not have the pistol-like design of a cheap homeowner's\ndrill. It is a cube of solid metal with a handle sticking out of one face\nand a chuck mounted in another. The cube contains a disconcertingly potent\nelectric motor. You can hold the handle and operate the trigger with your\nindex finger, but unless you are exceptionally strong you cannot control\nthe weight of the Hole Hawg with one hand; it is a two-hander all the way.\nIn order to fight off the counter-torque of the Hole Hawg you use a\nseparate handle (provided), which you screw into one side of the iron cube\nor the other depending on whether you are using your left or right hand to\noperate the trigger. This handle is not a sleek, ergonomically designed\nitem as it would be in a homeowner's drill. It is simply a foot-long chunk\nof regular galvanized pipe, threaded on one end, with a black rubber\nhandle on the other. If you lose it, you just go to the local plumbing\nsupply store and buy another chunk of pipe. \n<\/p><p> During the Eighties I did some construction work. One day, another\nworker leaned a ladder against the outside of the building that we were\nputting up, climbed up to the second-story level, and used the Hole Hawg\nto drill a hole through the exterior wall. At some point, the drill bit\ncaught in the wall.  The Hole Hawg, following its one and only imperative,\nkept going.  It spun the worker's body around like a rag doll, causing him\nto knock his own ladder down. Fortunately he kept his grip on the Hole\nHawg, which remained lodged in the wall, and he simply dangled from it and\nshouted for help until someone came along and reinstated the ladder. \n<\/p><p> I myself used a Hole Hawg to drill many holes through studs, which it\ndid as a blender chops cabbage. I also used it to cut a few\nsix-inch-diameter holes through an old lath-and-plaster ceiling. I chucked\nin a new hole saw, went up to the second story, reached down between the\nnewly installed floor joists, and began to cut through the first-floor\nceiling below. Where my homeowner's drill had labored and whined to spin\nthe huge bit around, and had stalled at the slightest obstruction, the\nHole Hawg rotated with the stupid consistency of a spinning planet.  When\nthe hole saw seized up, the Hole Hawg spun itself and me around, and\ncrushed one of my hands between the steel pipe handle and a joist,\nproducing a few lacerations, each surrounded by a wide corona of deeply\nbruised flesh. It also bent the hole saw itself, though not so badly that\nI couldn't use it. After a few such run-ins, when I got ready to use the\nHole Hawg my heart actually began to pound with atavistic terror. \n<\/p><p> But I never blamed the Hole Hawg; I blamed myself. The Hole Hawg is\ndangerous because it does exactly what you tell it to. It is not bound by\nthe physical limitations that are inherent in a cheap drill, and neither\nis it limited by safety interlocks that might be built into a homeowner's\nproduct by a liability-conscious manufacturer. The danger lies not in the\nmachine itself but in the user's failure to envision the full consequences\nof the instructions he gives to it. \n<\/p><p> A smaller tool is dangerous too, but for a completely different\nreason: it tries to do what you tell it to, and fails in some way that is\nunpredictable and almost always undesirable. But the Hole Hawg is like the\ngenie of the ancient fairy tales, who carries out his master's\ninstructions literally and precisely and with unlimited power, often with\ndisastrous, unforeseen consequences. \n<\/p><p> Pre-Hole Hawg, I used to examine the drill selection in hardware\nstores with what I thought was a judicious eye, scorning the smaller\nlow-end models and hefting the big expensive ones appreciatively, wishing\nI could afford one of them babies. Now I view them all with such contempt\nthat I do not even consider them to be real drills--merely scaled-up toys\ndesigned to exploit the self-delusional tendencies of soft-handed\nhomeowners who want to believe that they have purchased an actual tool.\nTheir plastic casings, carefully designed and focus-group-tested to convey\na feeling of solidity and power, seem disgustingly flimsy and cheap to me,\nand I am ashamed that I was ever bamboozled into buying such knicknacks. \n<\/p><p> It is not hard to imagine what the world would look like to someone\nwho had been raised by contractors and who had never used any drill other\nthan a Hole Hawg. Such a person, presented with the best and most\nexpensive hardware-store drill, would not even recognize it as such. He\nmight instead misidentify it as a child's toy, or some kind of motorized\nscrewdriver. If a salesperson or a deluded homeowner referred to it as a\ndrill, he would laugh and tell them that they were mistaken--they simply\nhad their terminology wrong. His interlocutor would go away irritated, and\nprobably feeling rather defensive about his basement full of cheap,\ndangerous, flashy, colorful tools. \n<\/p><p> Unix is the Hole Hawg of operating systems, and Unix hackers, like\nDoug Barnes and the guy in the Dilbert cartoon and many of the other\npeople who populate Silicon Valley, are like contractor's sons who grew up\nusing only Hole Hawgs. They might use Apple\/Microsoft OSes to write\nletters, play video games, or balance their checkbooks, but they cannot\nreally bring themselves to take these operating systems seriously. \n<\/p><p> THE ORAL TRADITION\n<\/p><p> Unix is hard to learn. The process of learning it is one of multiple\nsmall epiphanies. Typically you are just on the verge of inventing some\nnecessary tool or utility when you realize that someone else has already\ninvented it, and built it in, and this explains some odd file or directory\nor command that you have noticed but never really understood before. \n<\/p><p> For example there is a command (a small program, part of the OS) \ncalled whoami, which enables you to ask the computer who it thinks you\nare. On a Unix machine, you are always logged in under some name--possibly\neven your own! What files you may work with, and what software you may\nuse, depends on your identity. When I started out using Linux, I was on a\nnon-networked machine in my basement, with only one user account, and so\nwhen I became aware of the whoami command it struck me as ludicrous. But\nonce you are logged in as one person, you can temporarily switch over to a\npseudonym in order to access different files. If your machine is on the\nInternet, you can log onto other computers, provided you have a user name\nand a password. At that point the distant machine becomes no different in\npractice from the one right in front of you. These changes in identity and\nlocation can easily become nested inside each other, many layers deep,\neven if you aren't doing anything nefarious. Once you have forgotten who\nand where you are, the whoami command is indispensible. I use it all the\ntime. \n<\/p><p> The file systems of Unix machines all have the same general structure.\nOn your flimsy operating systems, you can create directories (folders) and\ngive them names like Frodo or My Stuff and put them pretty much anywhere\nyou like. But under Unix the highest level--the root--of the filesystem is\nalways designated with the single character \"\/\" and it always contains the\nsame set of top-level directories: \n<\/p><pre>\n\/usr\n\/etc\n\/var\n\/bin\n\/proc\n\/boot\n\/home\n\/root\n\/sbin\n\/dev\n\/lib\n\/tmp\n<\/pre>\n<p> and each of these directories typically has its own distinct structure\nof subdirectories. Note the obsessive use of abbreviations and avoidance\nof capital letters; this is a system invented by people to whom repetitive\nstress disorder is what black lung is to miners. Long names get worn down\nto three-letter nubbins, like stones smoothed by a river. \n<\/p><p> This is not the place to try to explain why each of the above\ndirectories exists, and what is contained in it. At first it all seems\nobscure; worse, it seems deliberately obscure. When I started using Linux\nI was accustomed to being able to create directories wherever I wanted and\nto give them whatever names struck my fancy. Under Unix you are free to do\nthat, of course (you are free to do anything) but as you gain experience\nwith the system you come to understand that the directories listed above\nwere created for the best of reasons and that your life will be much\neasier if you follow along (within \/home, by the way, you have pretty much\nunlimited freedom). \n<\/p><p> After this kind of thing has happened several hundred or thousand\ntimes, the hacker understands why Unix is the way it is, and agrees that\nit wouldn't be the same any other way. It is this sort of acculturation\nthat gives Unix hackers their confidence in the system, and the attitude\nof calm, unshakable, annoying superiority captured in the Dilbert cartoon.\nWindows 95 and MacOS are products, contrived by engineers in the service\nof specific companies. Unix, by contrast, is not so much a product as it\nis a painstakingly compiled oral history of the hacker subculture. It is\nour Gilgamesh epic. \n<\/p><p> What made old epics like Gilgamesh so powerful and so long-lived was\nthat they were living bodies of narrative that many people knew by heart,\nand told over and over again--making their own personal embellishments\nwhenever it struck their fancy. The bad embellishments were shouted down,\nthe good ones picked up by others, polished, improved, and, over time,\nincorporated into the story. Likewise, Unix is known, loved, and\nunderstood by so many hackers that it can be re-created from scratch\nwhenever someone needs it. This is very difficult to understand for people\nwho are accustomed to thinking of OSes as things that absolutely have to\nbe bought. \n<\/p><p> Many hackers have launched more or less successful re-implementations\nof the Unix ideal. Each one brings in new embellishments. Some of them die\nout quickly, some are merged with similar, parallel innovations created by\ndifferent hackers attacking the same problem, others still are embraced,\nand adopted into the epic. Thus Unix has slowly accreted around a simple\nkernel and acquired a kind of complexity and asymmetry about it that is\norganic, like the roots of a tree, or the branchings of a coronary artery.\nUnderstanding it is more like anatomy than physics. \n<\/p><p> For at least a year, prior to my adoption of Linux, I had been hearing\nabout it. Credible, well-informed people kept telling me that a bunch of\nhackers had got together an implentation of Unix that could be downloaded,\nfree of charge, from the Internet. For a long time I could not bring\nmyself to take the notion seriously. It was like hearing rumors that a\ngroup of model rocket enthusiasts had created a completely functional\nSaturn V by exchanging blueprints on the Net and mailing valves and\nflanges to each other. \n<\/p><p> But it's true. Credit for Linux generally goes to its human namesake,\none Linus Torvalds, a Finn who got the whole thing rolling in 1991 when he\nused some of the GNU tools to write the beginnings of a Unix kernel that\ncould run on PC-compatible hardware. And indeed Torvalds deserves all the\ncredit he has ever gotten, and a whole lot more. But he could not have\nmade it happen by himself, any more than Richard Stallman could have. To\nwrite code at all, Torvalds had to have cheap but powerful development\ntools, and these he got from Stallman's GNU project. \n<\/p><p> And he had to have cheap hardware on which to write that code.  Cheap\nhardware is a much harder thing to arrange than cheap software; a single\nperson (Stallman) can write software and put it up on the Net for free,\nbut in order to make hardware it's necessary to have a whole industrial\ninfrastructure, which is not cheap by any stretch of the imagination.\nReally the only way to make hardware cheap is to punch out an incredible\nnumber of copies of it, so that the unit cost eventually drops. For\nreasons already explained, Apple had no desire to see the cost of hardware\ndrop. The only reason Torvalds had cheap hardware was Microsoft. \n<\/p><p> Microsoft refused to go into the hardware business, insisted on making\nits software run on hardware that anyone could build, and thereby created\nthe market conditions that allowed hardware prices to plummet. In trying\nto understand the Linux phenomenon, then, we have to look not to a single\ninnovator but to a sort of bizarre Trinity: Linus Torvalds, Richard\nStallman, and Bill Gates. Take away any of these three and Linux would not\nexist. \n<\/p><p> OS SHOCK\n<\/p><p> Young Americans who leave their great big homogeneous country and\nvisit some other part of the world typically go through several stages of\nculture shock: first, dumb wide-eyed astonishment. Then a tentative\nengagement with the new country's manners, cuisine, public transit systems\nand toilets, leading to a brief period of fatuous confidence that they are\ninstant experts on the new country. As the visit wears on, homesickness\nbegins to set in, and the traveler begins to appreciate, for the first\ntime, how much he or she took for granted at home. At the same time it\nbegins to seem obvious that many of one's own cultures and traditions are\nessentially arbitrary, and could have been different; driving on the right\nside of the road, for example.  When the traveler returns home and takes\nstock of the experience, he or she may have learned a good deal more about\nAmerica than about the country they went to visit. \n<\/p><p> For the same reasons, Linux is worth trying. It is a strange country\nindeed, but you don't have to live there; a brief sojourn suffices to give\nsome flavor of the place and--more importantly--to lay bare everything\nthat is taken for granted, and all that could have been done differently,\nunder Windows or MacOS. \n<\/p><p> You can't try it unless you install it. With any other OS, installing\nit would be a straightforward transaction: in exchange for money, some\ncompany would give you a CD-ROM, and you would be on your way. But a lot\nis subsumed in that kind of transaction, and has to be gone through and\npicked apart. \n<\/p><p> We like plain dealings and straightforward transactions in America. If\nyou go to Egypt and, say, take a taxi somewhere, you become a part of the\ntaxi driver's life; he refuses to take your money because it would demean\nyour friendship, he follows you around town, and weeps hot tears when you\nget in some other guy's taxi. You end up meeting his kids at some point,\nand have to devote all sort of ingenuity to finding some way to compensate\nhim without insulting his honor. It is exhausting. Sometimes you just want\na simple Manhattan-style taxi ride. \n<\/p><p> But in order to have an American-style setup, where you can just go\nout and hail a taxi and be on your way, there must exist a whole hidden\napparatus of medallions, inspectors, commissions, and so forth--which is\nfine as long as taxis are cheap and you can always get one. When the\nsystem fails to work in some way, it is mysterious and infuriating and\nturns otherwise reasonable people into conspiracy theorists. But when the\nEgyptian system breaks down, it breaks down transparently. You can't get a\ntaxi, but your driver's nephew will show up, on foot, to explain the\nproblem and apologize. \n<\/p><p> Microsoft and Apple do things the Manhattan way, with vast complexity\nhidden behind a wall of interface. Linux does things the Egypt way, with\nvast complexity strewn about all over the landscape. If you've just flown\nin from Manhattan, your first impulse will be to throw up your hands and\nsay \"For crying out loud! Will you people get a grip on yourselves!?\" But\nthis does not make friends in Linux-land any better than it would in\nEgypt. \n<\/p><p> You can suck Linux right out of the air, as it were, by downloading\nthe right files and putting them in the right places, but there probably\nare not more than a few hundred people in the world who could create a\nfunctioning Linux system in that way.  What you really need is a\ndistribution of Linux, which means a prepackaged set of files. But\ndistributions are a separate thing from Linux per se. \n<\/p><p> Linux per se is not a specific set of ones and zeroes, but a\nself-organizing Net subculture. The end result of its collective\nlucubrations is a vast body of source code, almost all written in C (the\ndominant computer programming language). \"Source code\"  just means a\ncomputer program as typed in and edited by some hacker. If it's in C, the\nfile name will probably have .c or .cpp on the end of it, depending on\nwhich dialect was used; if it's in some other language it will have some\nother suffix. Frequently these sorts of files can be found in a directory\nwith the name \/src which is the hacker's Hebraic abbreviation of \"source.\" \n<\/p><p> Source files are useless to your computer, and of little interest to\nmost users, but they are of gigantic cultural and political significance,\nbecause Microsoft and Apple keep them secret while Linux makes them\npublic. They are the family jewels. They are the sort of thing that in\nHollywood thrillers is used as a McGuffin:  the plutonium bomb core, the\ntop-secret blueprints, the suitcase of bearer bonds, the reel of\nmicrofilm. If the source files for Windows or MacOS were made public on\nthe Net, then those OSes would become free, like Linux--only not as good,\nbecause no one would be around to fix bugs and answer questions. Linux is\n\"open source\" software meaning, simply, that anyone can get copies of its\nsource code files. \n<\/p><p> Your computer doesn't want source code any more than you do; it wants\nobject code. Object code files typically have the suffix .o and are\nunreadable all but a few, highly strange humans, because they consist of\nones and zeroes. Accordingly, this sort of file commonly shows up in a\ndirectory with the name \/bin, for \"binary.\" \n<\/p><p> Source files are simply ASCII text files. ASCII denotes a particular\nway of encoding letters into bit patterns. In an ASCII file, each\ncharacter has eight bits all to itself. This creates a potential\n\"alphabet\" of 256 distinct characters, in that eight binary digits can\nform that many unique patterns. In practice, of course, we tend to limit\nourselves to the familiar letters and digits. The bit-patterns used to\nrepresent those letters and digits are the same ones that were physically\npunched into the paper tape by my high school teletype, which in turn were\nthe same one used by the telegraph industry for decades previously.  ASCII\ntext files, in other words, are telegrams, and as such they have no\ntypographical frills. But for the same reason they are eternal, because\nthe code never changes, and universal, because every text editing and word\nprocessing software ever written knows about this code. \n<\/p><p> Therefore just about any software can be used to create, edit, and\nread source code files. Object code files, then, are created from these\nsource files by a piece of software called a compiler, and forged into a\nworking application by another piece of software called a linker. \n<\/p><p> The triad of editor, compiler, and linker, taken together, form the\ncore of a software development system. Now, it is possible to spend a lot\nof money on shrink-wrapped development systems with lovely graphical user\ninterfaces and various ergonomic enhancements. In some cases it might even\nbe a good and reasonable way to spend money. But on this side of the road,\nas it were, the very best software is usually the free stuff.  Editor,\ncompiler and linker are to hackers what ponies, stirrups, and archery sets\nwere to the Mongols. Hackers live in the saddle, and hack on their own\ntools even while they are using them to create new applications. It is\nquite inconceivable that superior hacking tools could have been created\nfrom a blank sheet of paper by product engineers. Even if they are the\nbrightest engineers in the world they are simply outnumbered. \n<\/p><p> In the GNU\/Linux world there are two major text editing programs:  the\nminimalist vi (known in some implementations as elvis) and the maximalist\nemacs. I use emacs, which might be thought of as a thermonuclear word\nprocessor. It was created by Richard Stallman;  enough said. It is written\nin Lisp, which is the only computer language that is beautiful. It is\ncolossal, and yet it only edits straight ASCII text files, which is to\nsay, no fonts, no boldface, no underlining. In other words, the\nengineer-hours that, in the case of Microsoft Word, were devoted to\nfeatures like mail merge, and the ability to embed feature-length motion\npictures in corporate memoranda, were, in the case of emacs, focused with\nmaniacal intensity on the deceptively simple-seeming problem of editing\ntext. If you are a professional writer--i.e., if someone else is getting\npaid to worry about how your words are formatted and printed--emacs\noutshines all other editing software in approximately the same way that\nthe noonday sun does the stars. It is not just bigger and brighter; it\nsimply makes everything else vanish. For page layout and printing you can\nuse TeX: a vast corpus of typesetting lore written in C and also available\non the Net for free. \n<\/p><p> I could say a lot about emacs and TeX, but right now I am trying to\ntell a story about how to actually install Linux on your machine. The\nhard-core survivalist approach would be to download an editor like emacs,\nand the GNU Tools--the compiler and linker--which are polished and\nexcellent to the same degree as emacs. Equipped with these, one would be\nable to start downloading ASCII source code files (\/src) and compiling\nthem into binary object code files (\/bin) that would run on the machine.\nBut in order to even arrive at this point--to get emacs running, for\nexample--you have to have Linux actually up and running on your machine.\nAnd even a minimal Linux operating system requires thousands of binary\nfiles all acting in concert, and arranged and linked together just so. \n<\/p><p> Several entities have therefore taken it upon themselves to create\n\"distributions\" of Linux. If I may extend the Egypt analogy slightly,\nthese entities are a bit like tour guides who meet you at the airport, who\nspeak your language, and who help guide you through the initial culture\nshock. If you are an Egyptian, of course, you see it the other way; tour\nguides exist to keep brutish outlanders from traipsing through your\nmosques and asking you the same questions over and over and over again. \n<\/p><p> Some of these tour guides are commercial organizations, such as Red\nHat Software, which makes a Linux distribution called Red Hat that has a\nrelatively commercial sheen to it. In most cases you put a Red Hat CD-ROM\ninto your PC and reboot and it handles the rest. Just as a tour guide in\nEgypt will expect some sort of compensation for his services, commercial\ndistributions need to be paid for. In most cases they cost almost nothing\nand are well worth it. \n<\/p><p> I use a distribution called Debian (the word is a contraction of\n\"Deborah\" and \"Ian\") which is non-commercial. It is organized (or perhaps\nI should say \"it has organized itself\") along the same lines as Linux in\ngeneral, which is to say that it consists of volunteers who collaborate\nover the Net, each responsible for looking after a different chunk of the\nsystem. These people have broken Linux down into a number of packages,\nwhich are compressed files that can be downloaded to an already\nfunctioning Debian Linux system, then opened up and unpacked using a free\ninstaller application. Of course, as such, Debian has no commercial\narm--no distribution mechanism. You can download all Debian packages over\nthe Net, but most people will want to have them on a CD-ROM.  Several\ndifferent companies have taken it upon themselves to decoct all of the\ncurrent Debian packages onto CD-ROMs and then sell them. I buy mine from\nLinux Systems Labs. The cost for a three-disc set, containing Debian in\nits entirety, is less than three dollars. But (and this is an important\ndistinction) not a single penny of that three dollars is going to any of\nthe coders who created Linux, nor to the Debian packagers. It goes to\nLinux Systems Labs and it pays, not for the software, or the packages, but\nfor the cost of stamping out the CD-ROMs. \n<\/p><p> Every Linux distribution embodies some more or less clever hack for\ncircumventing the normal boot process and causing your computer, when it\nis turned on, to organize itself, not as a PC running Windows, but as a\n\"host\" running Unix. This is slightly alarming the first time you see it,\nbut completely harmless. When a PC boots up, it goes through a little\nself-test routine, taking an inventory of available disks and memory, and\nthen begins looking around for a disk to boot up from. In any normal\nWindows computer that disk will be a hard drive. But if you have your\nsystem configured right, it will look first for a floppy or CD-ROM disk,\nand boot from that if one is available. \n<\/p><p> Linux exploits this chink in the defenses. Your computer notices a\nbootable disk in the floppy or CD-ROM drive, loads in some object code\nfrom that disk, and blindly begins to execute it. But this is not\nMicrosoft or Apple code, this is Linux code, and so at this point your\ncomputer begins to behave very differently from what you are accustomed\nto. Cryptic messages began to scroll up the screen. If you had booted a\ncommercial OS, you would, at this point, be seeing a \"Welcome to MacOS\"\ncartoon, or a screen filled with clouds in a blue sky, and a Windows logo.\nBut under Linux you get a long telegram printed in stark white letters on\na black screen. There is no \"welcome!\" message. Most of the telegram has\nthe semi-inscrutable menace of graffiti tags. \n<\/p><pre>\nDec 14 15:04:15 theRev syslogd 1.3-3#17: restart. \nDec 14 15:04:15 theRev kernel: klogd 1.3-3, log source = \/proc\/kmsg started. \nDec 14 15:04:15 theRev kernel: Loaded 3535 symbols from \/System.map.\nDec 14 15:04:15 theRev kernel: Symbols match kernel version 2.0.30. Dec 14 15:04:15 theRev kernel: No module symbols loaded.\nDec 14 15:04:15 theRev kernel: Intel MultiProcessor Specification v1.4 \nDec 14 15:04:15 theRev kernel: Virtual Wire compatibility mode. \nDec 14 15:04:15 theRev kernel: OEM ID: INTEL Product ID: 440FX APIC at: 0xFEE00000 \nDec 14 15:04:15 theRev kernel: Processor #0 Pentium(tm) Pro APIC version 17 \nDec 14 15:04:15 theRev kernel: Processor #1 Pentium(tm) Pro APIC version 17 \nDec 14 15:04:15 theRev kernel: I\/O APIC #2 Version 17 at 0xFEC00000.\nDec 14 15:04:15 theRev kernel: Processors: 2\nDec 14 15:04:15 theRev kernel: Console: 16 point font, 400 scans \nDec 14 15:04:15 theRev kernel: Console: colour VGA+ 80x25, 1 virtual console (max 63) \nDec 14 15:04:15 theRev kernel: pcibios_init : BIOS32 Service Directory structure at 0x000fdb70 \nDec 14 15:04:15 theRev kernel: pcibios_init : BIOS32 Service Directory entry at 0xfdb80 \nDec 14 15:04:15 theRev kernel: pcibios_init : PCI BIOS revision 2.10 entry at 0xfdba1 \nDec 14 15:04:15 theRev kernel: Probing PCI hardware. \nDec 14 15:04:15 theRev kernel: Warning : Unknown PCI device (10b7:9001). Please read include\/linux\/pci.h \nDec 14 15:04:15 theRev kernel: Calibrating delay loop.. ok - 179.40 BogoMIPS \nDec 14 15:04:15 theRev kernel: Memory: 64268k\/66556k available (700k kernel code, 384k reserved, 1204k data) \nDec 14 15:04:15 theRev kernel: Swansea University Computer Society NET3.035 for Linux 2.0 \nDec 14 15:04:15 theRev kernel: NET3: Unix domain sockets 0.13 for Linux NET3.035. \nDec 14 15:04:15 theRev kernel: Swansea University Computer Society TCP\/IP for NET3.034\nDec 14 15:04:15 theRev kernel: IP Protocols: ICMP, UDP, TCP \nDec 14 15:04:15 theRev kernel: Checking 386\/387 coupling... Ok, fpu using exception 16 error reporting. \nDec 14 15:04:15 theRev kernel: Checking 'hlt' instruction... Ok. \nDec 14 15:04:15 theRev kernel: Linux version 2.0.30 (root@theRev) (gcc version 2.7.2.1) #15 Fri Mar 27 16:37:24 PST 1998 \nDec 14 15:04:15 theRev kernel: Booting processor 1 stack 00002000: Calibrating delay loop.. ok - 179.40 BogoMIPS \nDec 14 15:04:15 theRev kernel: Total of 2 processors activated (358.81 BogoMIPS). \nDec 14 15:04:15 theRev kernel: Serial driver version 4.13 with no serial options enabled\nDec 14 15:04:15 theRev kernel: tty00 at 0x03f8 (irq = 4) is a 16550A \nDec 14 15:04:15 theRev kernel: tty01 at 0x02f8 (irq = 3) is a 16550A \nDec 14 15:04:15 theRev kernel: lp1 at 0x0378, (polling) \nDec 14 15:04:15 theRev kernel: PS\/2 auxiliary pointing device detected -- driver installed. \nDec 14 15:04:15 theRev kernel: Real Time Clock Driver v1.07 \nDec 14 15:04:15 theRev kernel: loop: registered device at major 7 \nDec 14 15:04:15 theRev kernel: ide: i82371 PIIX (Triton) on PCI bus 0 function 57 \nDec 14 15:04:15 theRev kernel: ide0: BM-DMA at 0xffa0-0xffa7 \nDec 14 15:04:15 theRev kernel: ide1: BM-DMA at 0xffa8-0xffaf \nDec 14 15:04:15 theRev kernel: hda: Conner Peripherals 1275MB - CFS1275A, 1219MB w\/64kB Cache, LBA, CHS=619\/64\/63 \nDec 14 15:04:15 theRev kernel: hdb: Maxtor 84320A5, 4119MB w\/256kB Cache, LBA, CHS=8928\/15\/63, DMA \nDec 14 15:04:15 theRev kernel: hdc: , ATAPI CDROM drive \nDec 15 11:58:06 theRev kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 \nDec 15 11:58:06 theRev kernel: ide1 at 0x170-0x177,0x376 on irq 15 \nDec 15 11:58:06 theRev kernel: Floppy drive(s): fd0 is 1.44M \nDec 15 11:58:06 theRev kernel: Started kswapd v 1.4.2.2 \nDec 15 11:58:06 theRev kernel: FDC 0 is a National Semiconductor PC87306 \nDec 15 11:58:06 theRev kernel: md driver 0.35 MAX_MD_DEV=4, MAX_REAL=8 \nDec 15 11:58:06 theRev kernel: PPP: version 2.2.0 (dynamic channel allocation) \nDec 15 11:58:06 theRev kernel: TCP compression code copyright 1989 Regents of the University of California \nDec 15 11:58:06 theRev kernel: PPP Dynamic channel allocation code copyright 1995 Caldera, Inc. \nDec 15 11:58:06 theRev kernel: PPP line discipline registered. \nDec 15 11:58:06 theRev kernel: SLIP: version 0.8.4-NET3.019-NEWTTY (dynamic channels, max=256). \nDec 15 11:58:06 theRev kernel: eth0: 3Com 3c900 Boomerang 10Mbps\/Combo at 0xef00, 00:60:08:a4:3c:db, IRQ 10 \nDec 15 11:58:06 theRev kernel: 8K word-wide RAM 3:5 Rx:Tx split, 10base2 interface. \nDec 15 11:58:06 theRev kernel: Enabling bus-master transmits and whole-frame receives. \nDec 15 11:58:06 theRev kernel: 3c59x.c:v0.49 1\/2\/98 Donald Becker http:\/\/cesdis.gsfc.nasa.gov\/linux\/drivers\/vortex.html \nDec 15 11:58:06 theRev kernel: Partition check: \nDec 15 11:58:06 theRev kernel: hda: hda1 hda2 hda3 \nDec 15 11:58:06 theRev kernel: hdb: hdb1 hdb2 \nDec 15 11:58:06 theRev kernel: VFS: Mounted root (ext2 filesystem) readonly. \nDec 15 11:58:06 theRev kernel: Adding Swap: 16124k swap-space (priority -1) \nDec 15 11:58:06 theRev kernel: EXT2-fs warning: maximal mount count reached, running e2fsck is recommended \nDec 15 11:58:06 theRev kernel: hdc: media changed \nDec 15 11:58:06 theRev kernel: ISO9660 Extensions: RRIP_1991A \nDec 15 11:58:07 theRev syslogd 1.3-3#17: restart. \nDec 15 11:58:09 theRev diald[87]: Unable to open options file \/etc\/diald\/diald.options: No such file or directory \nDec 15 11:58:09 theRev diald[87]: No device specified. You must have at least one device! \nDec 15 11:58:09 theRev diald[87]: You must define a connector script (option 'connect'). \nDec 15 11:58:09 theRev diald[87]: You must define the remote ip address. \nDec 15 11:58:09 theRev diald[87]: You must define the local ip address. \nDec 15 11:58:09 theRev diald[87]: Terminating due to damaged reconfigure.\n<\/pre>\n\n<p> The only parts of this that are readable, for normal people, are the\nerror messages and warnings. And yet it's noteworthy that Linux doesn't\nstop, or crash, when it encounters an error; it spits out a pithy\ncomplaint, gives up on whatever processes were damaged, and keeps on\nrolling. This was decidedly not true of the early versions of Apple and\nMicrosoft OSes, for the simple reason that an OS that is not capable of\nwalking and chewing gum at the same time cannot possibly recover from\nerrors. Looking for, and dealing with, errors requires a separate process\nrunning in parallel with the one that has erred. A kind of superego, if\nyou will, that keeps an eye on all of the others, and jumps in when one\ngoes astray. Now that MacOS and Windows can do more than one thing at a\ntime they are much better at dealing with errors than they used to be, but\nthey are not even close to Linux or other Unices in this respect; and\ntheir greater complexity has made them vulnerable to new types of errors. \n<\/p><p> FALLIBILITY, ATONEMENT, REDEMPTION, TRUST, AND OTHER ARCANE TECHNICAL\nCONCEPTS\n<\/p><p> Linux is not capable of having any centrally organized policies\ndictating how to write error messages and documentation, and so each\nprogrammer writes his own. Usually they are in English even though tons of\nLinux programmers are Europeans. Frequently they are funny. Always they\nare honest. If something bad has happened because the software simply\nisn't finished yet, or because the user screwed something up, this will be\nstated forthrightly. The command line interface makes it easy for programs\nto dribble out little comments, warnings, and messages here and there.\nEven if the application is imploding like a damaged submarine, it can\nstill usually eke out a little S.O.S. message. Sometimes when you finish\nworking with a program and shut it down, you find that it has left behind\na series of mild warnings and low-grade error messages in the command-line\ninterface window from which you launched it. As if the software were\nchatting to you about how it was doing the whole time you were working\nwith it. \n<\/p><p> Documentation, under Linux, comes in the form of man (short for\nmanual) pages. You can access these either through a GUI (xman)  or from\nthe command line (man). Here is a sample from the man page for a program\ncalled rsh: \n<\/p><p> \"Stop signals stop the local rsh process only; this is arguably wrong,\nbut currently hard to fix for reasons too complicated to explain here.\" \n<\/p><p> The man pages contain a lot of such material, which reads like the\nterse mutterings of pilots wrestling with the controls of damaged\nairplanes. The general feel is of a thousand monumental but obscure\nstruggles seen in the stop-action light of a strobe.  Each programmer is\ndealing with his own obstacles and bugs; he is too busy fixing them, and\nimproving the software, to explain things at great length or to maintain\nelaborate pretensions. \n<\/p><p> In practice you hardly ever encounter a serious bug while running\nLinux. When you do, it is almost always with commercial software (several\nvendors sell software that runs under Linux). The operating system and its\nfundamental utility programs are too important to contain serious bugs. I\nhave been running Linux every day since late 1995 and have seen many\napplication programs go down in flames, but I have never seen the\noperating system crash. Never. Not once. There are quite a few Linux\nsystems that have been running continuously and working hard for months or\nyears without needing to be rebooted. \n<\/p><p> Commercial OSes have to adopt the same official stance towards errors\nas Communist countries had towards poverty. For doctrinal reasons it was\nnot possible to admit that poverty was a serious problem in Communist\ncountries, because the whole point of Communism was to eradicate poverty.\nLikewise, commercial OS companies like Apple and Microsoft can't go around\nadmitting that their software has bugs and that it crashes all the time,\nany more than Disney can issue press releases stating that Mickey Mouse is\nan actor in a suit. \n<\/p><p> This is a problem, because errors do exist and bugs do happen.  Every\nfew months Bill Gates tries to demo a new Microsoft product in front of a\nlarge audience only to have it blow up in his face.  Commercial OS\nvendors, as a direct consequence of being commercial, are forced to adopt\nthe grossly disingenuous position that bugs are rare aberrations, usually\nsomeone else's fault, and therefore not really worth talking about in any\ndetail. This posture, which everyone knows to be absurd, is not limited to\npress releases and ad campaigns. It informs the whole way these companies\ndo business and relate to their customers. If the documentation were\nproperly written, it would mention bugs, errors, and crashes on every\nsingle page. If the on-line help systems that come with these OSes\nreflected the experiences and concerns of their users, they would largely\nbe devoted to instructions on how to cope with crashes and errors. \n<\/p><p> But this does not happen. Joint stock corporations are wonderful\ninventions that have given us many excellent goods and services.  They are\ngood at many things. Admitting failure is not one of them. Hell, they\ncan't even admit minor shortcomings. \n<\/p><p> Of course, this behavior is not as pathological in a corporation as it\nwould be in a human being. Most people, nowadays, understand that\ncorporate press releases are issued for the benefit of the corporation's\nshareholders and not for the enlightenment of the public. Sometimes the\nresults of this institutional dishonesty can be dreadful, as with tobacco\nand asbestos. In the case of commercial OS vendors it is nothing of the\nkind, of course; it is merely annoying. \n<\/p><p> Some might argue that consumer annoyance, over time, builds up into a\nkind of hardened plaque that can conceal serious decay, and that honesty\nmight therefore be the best policy in the long run; the jury is still out\non this in the operating system market. The business is expanding fast\nenough that it's still much better to have billions of chronically annoyed\ncustomers than millions of happy ones. \n<\/p><p> Most system administrators I know who work with Windows NT all the\ntime agree that when it hits a snag, it has to be re-booted, and when it\ngets seriously messed up, the only way to fix it is to re-install the\noperating system from scratch. Or at least this is the only way that they\nknow of to fix it, which amounts to the same thing. It is quite possible\nthat the engineers at Microsoft have all sorts of insider knowledge on how\nto fix the system when it goes awry, but if they do, they do not seem to\nbe getting the message out to any of the actual system administrators I\nknow. \n<\/p><p> Because Linux is not commercial--because it is, in fact, free, as well\nas rather difficult to obtain, install, and operate--it does not have to\nmaintain any pretensions as to its reliability.  Consequently, it is much\nmore reliable. When something goes wrong with Linux, the error is noticed\nand loudly discussed right away.  Anyone with the requisite technical\nknowledge can go straight to the source code and point out the source of\nthe error, which is then rapidly fixed by whichever hacker has carved out\nresponsibility for that particular program. \n<\/p><p> As far as I know, Debian is the only Linux distribution that has its\nown constitution (http:\/\/www.debian.org\/devel\/constitution), but what\nreally sold me on it was its phenomenal bug database\n(http:\/\/www.debian.org\/Bugs), which is a sort of interactive Doomsday Book\nof error, fallibility, and redemption. It is simplicity itself. When had a\nproblem with Debian in early January of 1997, I sent in a message\ndescribing the problem to submit@bugs.debian.org. My problem was promptly\nassigned a bug report number (#6518) and a severity level (the available\nchoices being critical, grave, important, normal, fixed, and wishlist) \nand forwarded to mailing lists where Debian people hang out.  Within\ntwenty-four hours I had received five e-mails telling me how to fix the\nproblem: two from North America, two from Europe, and one from Australia.\nAll of these e-mails gave me the same suggestion, which worked, and made\nmy problem go away. But at the same time, a transcript of this exchange\nwas posted to Debian's bug database, so that if other users had the same\nproblem later, they would be able to search through and find the solution\nwithout having to enter a new, redundant bug report. \n<\/p><p> Contrast this with the experience that I had when I tried to install\nWindows NT 4.0 on the very same machine about ten months later, in late\n1997. The installation program simply stopped in the middle with no error\nmessages. I went to the Microsoft Support website and tried to perform a\nsearch for existing help documents that would address my problem. The\nsearch engine was completely nonfunctional; it did nothing at all. It did\nnot even give me a message telling me that it was not working. \n<\/p><p> Eventually I decided that my motherboard must be at fault; it was of a\nslightly unusual make and model, and NT did not support as many different\nmotherboards as Linux. I am always looking for excuses, no matter how\nfeeble, to buy new hardware, so I bought a new motherboard that was\nWindows NT logo-compatible, meaning that the Windows NT logo was printed\nright on the box. I installed this into my computer and got Linux running\nright away, then attempted to install Windows NT again. Again, the\ninstallation died without any error message or explanation. By this time a\ncouple of weeks had gone by and I thought that perhaps the search engine\non the Microsoft Support website might be up and running.  I gave that a\ntry but it still didn't work. \n<\/p><p> So I created a new Microsoft support account, then logged on to submit\nthe incident. I supplied my product ID number when asked, and then began\nto follow the instructions on a series of help screens. In other words, I\nwas submitting a bug report just as with the Debian bug tracking system.\nIt's just that the interface was slicker--I was typing my complaint into\nlittle text-editing boxes on Web forms, doing it all through the GUI,\nwhereas with Debian you send in an e-mail telegram. I knew that when I was\nfinished submitting the bug report, it would become proprietary Microsoft\ninformation, and other users wouldn't be able to see it. Many Linux users\nwould refuse to participate in such a scheme on ethical grounds, but I was\nwilling to give it a shot as an experiment. In the end, though I was never\nable to submit my bug report, because the series of linked web pages that\nI was filling out eventually led me to a completely blank page: a dead\nend. \n<\/p><p> So I went back and clicked on the buttons for \"phone support\" and\neventually was given a Microsoft telephone number. When I dialed this\nnumber I got a series of piercing beeps and a recorded message from the\nphone company saying \"We're sorry, your call cannot be completed as\ndialed.\" \n<\/p><p> I tried the search page again--it was still completely nonfunctional.\nThen I tried PPI (Pay Per Incident) again. This led me through another\nseries of Web pages until I dead-ended at one reading: \"Notice-there is no\nWeb page matching your request.\" \n<\/p><p> I tried it again, and eventually got to a Pay Per Incident screen\nreading: \"OUT OF INCIDENTS. There are no unused incidents left in your\naccount. If you would like to purchase a support incident, click OK-you\nwill then be able to prepay for an incident....\" The cost per incident was\n$95. \n<\/p><p> The experiment was beginning to seem rather expensive, so I gave up on\nthe PPI approach and decided to have a go at the FAQs posted on\nMicrosoft's website. None of the available FAQs had anything to do with my\nproblem except for one entitled \"I am having some problems installing NT\"\nwhich appeared to have been written by flacks, not engineers. \n<\/p><p> So I gave up and still, to this day, have never gotten Windows NT\ninstalled on that particular machine. For me, the path of least resistance\nwas simply to use Debian Linux. \n<\/p><p> In the world of open source software, bug reports are useful\ninformation. Making them public is a service to other users, and improves\nthe OS. Making them public systematically is so important that highly\nintelligent people voluntarily put time and money into running bug\ndatabases. In the commercial OS world, however, reporting a bug is a\nprivilege that you have to pay lots of money for. But if you pay for it,\nit follows that the bug report must be kept confidential--otherwise anyone\ncould get the benefit of your ninety-five bucks! And yet nothing prevents\nNT users from setting up their own public bug database. \n<\/p><p> This is, in other words, another feature of the OS market that simply\nmakes no sense unless you view it in the context of culture. What\nMicrosoft is selling through Pay Per Incident isn't technical support so\nmuch as the continued illusion that its customers are engaging in some\nkind of rational business transaction. It is a sort of routine maintenance\nfee for the upkeep of the fantasy. If people really wanted a solid OS they\nwould use Linux, and if they really wanted tech support they would find a\nway to get it; Microsoft's customers want something else. \n<\/p><p> As of this writing (Jan. 1999), something like 32,000 bugs have been\nreported to the Debian Linux bug database. Almost all of them have been\nfixed a long time ago. There are twelve \"critical\"  bugs still\noutstanding, of which the oldest was posted 79 days ago. There are 20\noutstanding \"grave\" bugs of which the oldest is 1166 days old. There are\n48 \"important\" bugs and hundreds of \"normal\" and less important ones. \n<\/p><p> Likewise, BeOS (which I'll get to in a minute) has its own bug\ndatabase (http:\/\/www.be.com\/developers\/bugs\/index.html) with its own\nclassification system, including such categories as \"Not a Bug,\"\n\"Acknowledged Feature,\" and \"Will Not Fix.\" Some of the \"bugs\" here are\nnothing more than Be hackers blowing off steam, and are classified as\n\"Input Acknowledged.\" For example, I found one that was posted on December\n30th, 1998. It's in the middle of a long list of bugs, wedged between one\nentitled \"Mouse working in very strange fashion\" and another called\n\"Change of BView frame does not affect, if BView not attached to a\nBWindow.\" \n<\/p><p> This one is entitled\n<\/p><p><code> R4: BeOS missing megalomaniacal figurehead to harness and focus\ndeveloper rage<\/code>\n<\/p><p> and it goes like this:\n<\/p><hr \/><pre>\nBe Status: Input Acknowledged\nBeOS Version: R3.2\nComponent: unknown\nFull Description:\nThe BeOS needs a megalomaniacal egomaniac sitting on its throne\nto give it a human character which everyone loves to hate.\nWithout this, the BeOS will languish in the impersonifiable realm\nof OSs that people can never quite get a handle on. You can judge\nthe success of an OS not by the quality of its features, but by\nhow infamous and disliked the leaders behind them are.\nI believe this is a side-effect of developer comraderie under\nmiserable conditions. After all, misery loves company. I believe\nthat making the BeOS less conceptually accessible and far less\nreliable will require developers to band together, thus\ndeveloping the kind of community where strangers talk to one-\nanother, kind of like at a grocery store before a huge snowstorm.\nFollowing this same program, it will likely be necessary to move\nthe BeOS headquarters to a far-less-comfortable climate. General\nenvironmental discomfort will breed this attitude within and\nthere truly is no greater recipe for success. I would suggest\nSeattle, but I think it's already taken. You might try\nWashington, DC, but definitely not somewhere like San Diego or\nTucson.\n<\/pre>\n<hr \/><p> Unfortunately, the Be bug reporting system strips off the names of the\npeople who report the bugs (to protect them from retribution!?) and so I\ndon't know who wrote this. \n<\/p><p> So it would appear that I'm in the middle of crowing about the\ntechnical and moral superiority of Debian Linux. But as almost always\nhappens in the OS world, it's more complicated than that.  I have Windows\nNT running on another machine, and the other day (Jan. 1999), when I had a\nproblem with it, I decided to have another go at Microsoft Support. This\ntime the search engine actually worked (though in order to reach it I had\nto identify myself as \"advanced\"). And instead of coughing up some useless\nFAQ, it located about two hundred documents (I was using very vague search\ncriteria) that were obviously bug reports--though they were called\nsomething else. Microsoft, in other words, has got a system up and running\nthat is functionally equivalent to Debian's bug database. It looks and\nfeels different, of course, but it contains technical nitty-gritty and\nmakes no bones about the existence of errors. \n<\/p><p> As I've explained, selling OSes for money is a basically untenable\nposition, and the only way Apple and Microsoft can get away with it is by\npursuing technological advancements as aggressively as they can, and by\ngetting people to believe in, and to pay for, a particular image: in the\ncase of Apple, that of the creative free thinker, and in the case of\nMicrosoft, that of the respectable techno-bourgeois. Just like Disney,\nthey're making money from selling an interface, a magic mirror. It has to\nbe polished and seamless or else the whole illusion is ruined and the\nbusiness plan vanishes like a mirage. \n<\/p><p> Accordingly, it was the case until recently that the people who wrote\nmanuals and created customer support websites for commercial OSes seemed\nto have been barred, by their employers' legal or PR departments, from\nadmitting, even obliquely, that the software might contain bugs or that\nthe interface might be suffering from the blinking twelve problem. They\ncouldn't address users' actual difficulties. The manuals and websites were\ntherefore useless, and caused even technically self-assured users to\nwonder whether they were going subtly insane. \n<\/p><p> When Apple engages in this sort of corporate behavior, one wants to\nbelieve that they are really trying their best. We all want to give Apple\nthe benefit of the doubt, because mean old Bill Gates kicked the crap out\nof them, and because they have good PR. But when Microsoft does it, one\nalmost cannot help becoming a paranoid conspiracist. Obviously they are\nhiding something from us! And yet they are so powerful! They are trying to\ndrive us crazy! \n<\/p><p> This approach to dealing with one's customers was straight out of the\nCentral European totalitarianism of the mid-Twentieth Century. The\nadjectives \"Kafkaesque\" and \"Orwellian\" come to mind. It couldn't last,\nany more than the Berlin Wall could, and so now Microsoft has a publicly\navailable bug database. It's called something else, and it takes a while\nto find it, but it's there. \n<\/p><p> They have, in other words, adapted to the two-tiered Eloi\/Morlock\nstructure of technological society. If you're an Eloi you install Windows,\nfollow the instructions, hope for the best, and dumbly suffer when it\nbreaks. If you're a Morlock you go to the website, tell it that you are\n\"advanced,\" find the bug database, and get the truth straight from some\nanonymous Microsoft engineer. \n<\/p><p> But once Microsoft has taken this step, it raises the question, once\nagain, of whether there is any point to being in the OS business at all.\nCustomers might be willing to pay $95 to report a problem to Microsoft if,\nin return, they get some advice that no other user is getting. This has\nthe useful side effect of keeping the users alienated from one another,\nwhich helps maintain the illusion that bugs are rare aberrations. But once\nthe results of those bug reports become openly available on the Microsoft\nwebsite, everything changes. No one is going to cough up $95 to report a\nproblem when chances are good that some other sucker will do it first, and\nthat instructions on how to fix the bug will then show up, for free, on a\npublic website. And as the size of the bug database grows, it eventually\nbecomes an open admission, on Microsoft's part, that their OSes have just\nas many bugs as their competitors'. There is no shame in that; as I\nmentioned, Debian's bug database has logged 32,000 reports so far. But it\nputs Microsoft on an equal footing with the others and makes it a lot\nharder for their customers--who want to believe--to believe. \n<\/p><p> MEMENTO MORI\n<\/p><p> Once the Linux machine has finished spitting out its jargonic opening\ntelegram, it prompts me to log in with a user name and a password. At this\npoint the machine is still running the command line interface, with white\nletters on a black screen. There are no windows, menus, or buttons. It\ndoes not respond to the mouse;  it doesn't even know that the mouse is\nthere. It is still possible to run a lot of software at this point. Emacs,\nfor example, exists in both a CLI and a GUI version (actually there are\ntwo GUI versions, reflecting some sort of doctrinal schism between Richard\nStallman and some hackers who got fed up with him). The same is true of\nmany other Unix programs. Many don't have a GUI at all, and many that do\nare capable of running from the command line. \n<\/p><p> Of course, since my computer only has one monitor screen, I can only\nsee one command line, and so you might think that I could only interact\nwith one program at a time. But if I hold down the Alt key and then hit\nthe F2 function button at the top of my keyboard, I am presented with a\nfresh, blank, black screen with a login prompt at the top of it. I can log\nin here and start some other program, then hit Alt-F1 and go back to the\nfirst screen, which is still doing whatever it was when I left it. Or I\ncan do Alt-F3 and log in to a third screen, or a fourth, or a fifth. On\none of these screens I might be logged in as myself, on another as root\n(the system administrator), on yet another I might be logged on to some\nother computer over the Internet. \n<\/p><p> Each of these screens is called, in Unix-speak, a tty, which is an\nabbreviation for teletype. So when I use my Linux system in this way I am\ngoing right back to that small room at Ames High School where I first\nwrote code twenty-five years ago, except that a tty is quieter and faster\nthan a teletype, and capable of running vastly superior software, such as\nemacs or the GNU development tools. \n<\/p><p> It is easy (easy by Unix, not Apple\/Microsoft standards) to configure\na Linux machine so that it will go directly into a GUI when you boot it\nup. This way, you never see a tty screen at all.  I still have mine boot\ninto the white-on-black teletype screen however, as a computational\nmemento mori. It used to be fashionable for a writer to keep a human skull\non his desk as a reminder that he was mortal, that all about him was\nvanity. The tty screen reminds me that the same thing is true of slick\nuser interfaces. \n<\/p><p> The X Windows System, which is the GUI of Unix, has to be capable of\nrunning on hundreds of different video cards with different chipsets,\namounts of onboard memory, and motherboard buses.  Likewise, there are\nhundreds of different types of monitors on the new and used market, each\nwith different specifications, and so there are probably upwards of a\nmillion different possible combinations of card and monitor. The only\nthing they all have in common is that they all work in VGA mode, which is\nthe old command-line screen that you see for a few seconds when you launch\nWindows. So Linux always starts in VGA, with a teletype interface, because\nat first it has no idea what sort of hardware is attached to your\ncomputer. In order to get beyond the glass teletype and into the GUI, you\nhave to tell Linux exactly what kinds of hardware you have. If you get it\nwrong, you'll get a blank screen at best, and at worst you might actually\ndestroy your monitor by feeding it signals it can't handle. \n<\/p><p> When I started using Linux this had to be done by hand. I once spent\nthe better part of a month trying to get an oddball monitor to work for\nme, and filled the better part of a composition book with increasingly\ndesperate scrawled notes. Nowadays, most Linux distributions ship with a\nprogram that automatically scans the video card and self-configures the\nsystem, so getting X Windows up and running is nearly as easy as\ninstalling an Apple\/Microsoft GUI. The crucial information goes into a\nfile (an ASCII text file, naturally) called XF86Config, which is worth\nlooking at even if your distribution creates it for you automatically. For\nmost people it looks like meaningless cryptic incantations, which is the\nwhole point of looking at it. An Apple\/Microsoft system needs to have the\nsame information in order to launch its GUI, but it's apt to be deeply\nhidden somewhere, and it's probably in a file that can't even be opened\nand read by a text editor. All of the important files that make Linux\nsystems work are right out in the open. They are always ASCII text files,\nso you don't need special tools to read them. You can look at them any\ntime you want, which is good, and you can mess them up and render your\nsystem totally dysfunctional, which is not so good. \n<\/p><p> At any rate, assuming that my XF86Config file is just so, I enter the\ncommand \"startx\" to launch the X Windows System. The screen blanks out for\na minute, the monitor makes strange twitching noises, then reconstitutes\nitself as a blank gray desktop with a mouse cursor in the middle. At the\nsame time it is launching a window manager. X Windows is pretty low-level\nsoftware; it provides the infrastructure for a GUI, and it's a heavy\nindustrial infrastructure. But it doesn't do windows. That's handled by\nanother category of application that sits atop X Windows, called a window\nmanager. Several of these are available, all free of course. The classic\nis twm (Tom's Window Manager) but there is a smaller and supposedly more\nefficient variant of it called fvwm, which is what I use. I have my eye on\na completely different window manager called Enlightenment, which may be\nthe hippest single technology product I have ever seen, in that (a)  it is\nfor Linux, (b) it is freeware, (c) it is being developed by a very small\nnumber of obsessed hackers, and (d) it looks amazingly cool; it is the\nsort of window manager that might show up in the backdrop of an Aliens\nmovie. \n<\/p><p> Anyway, the window manager acts as an intermediary between X Windows\nand whatever software you want to use. It draws the window frames, menus,\nand so on, while the applications themselves draw the actual content in\nthe windows. The applications might be of any sort: text editors, Web\nbrowsers, graphics packages, or utility programs, such as a clock or\ncalculator. In other words, from this point on, you feel as if you have\nbeen shunted into a parallel universe that is quite similar to the\nfamiliar Apple or Microsoft one, but slightly and pervasively different.\nThe premier graphics program under Apple\/Microsoft is Adobe Photoshop, but\nunder Linux it's something called The GIMP. Instead of the Microsoft\nOffice Suite, you can buy something called ApplixWare. Many commercial\nsoftware packages, such as Mathematica, Netscape Communicator, and Adobe\nAcrobat, are available in Linux versions, and depending on how you set up\nyour window manager you can make them look and behave just as they would\nunder MacOS or Windows. \n<\/p><p> But there is one type of window you'll see on Linux GUI that is rare\nor nonexistent under other OSes. These windows are called \"xterm\" and\ncontain nothing but lines of text--this time, black text on a white\nbackground, though you can make them be different colors if you choose.\nEach xterm window is a separate command line interface--a tty in a window.\nSo even when you are in full GUI mode, you can still talk to your Linux\nmachine through a command-line interface. \n<\/p><p> There are many good pieces of Unix software that do not have GUIs at\nall. This might be because they were developed before X Windows was\navailable, or because the people who wrote them did not want to suffer\nthrough all the hassle of creating a GUI, or because they simply do not\nneed one. In any event, those programs can be invoked by typing their\nnames into the command line of an xterm window. The whoami command,\nmentioned earlier, is a good example. There is another called wc (\"word\ncount\") which simply returns the number of lines, words, and characters in\na text file. \n<\/p><p> The ability to run these little utility programs on the command line\nis a great virtue of Unix, and one that is unlikely to be duplicated by\npure GUI operating systems. The wc command, for example, is the sort of\nthing that is easy to write with a command line interface. It probably\ndoes not consist of more than a few lines of code, and a clever programmer\ncould probably write it in a single line. In compiled form it takes up\njust a few bytes of disk space. But the code required to give the same\nprogram a graphical user interface would probably run into hundreds or\neven thousands of lines, depending on how fancy the programmer wanted to\nmake it. Compiled into a runnable piece of software, it would have a large\noverhead of GUI code. It would be slow to launch and it would use up a lot\nof memory. This would simply not be worth the effort, and so \"wc\" would\nnever be written as an independent program at all. Instead users would\nhave to wait for a word count feature to appear in a commercial software\npackage. \n<\/p><p> GUIs tend to impose a large overhead on every single piece of\nsoftware, even the smallest, and this overhead completely changes the\nprogramming environment. Small utility programs are no longer worth\nwriting. Their functions, instead, tend to get swallowed up into omnibus\nsoftware packages. As GUIs get more complex, and impose more and more\noverhead, this tendency becomes more pervasive, and the software packages\ngrow ever more colossal;  after a point they begin to merge with each\nother, as Microsoft Word and Excel and PowerPoint have merged into\nMicrosoft Office:  a stupendous software Wal-Mart sitting on the edge of a\ntown filled with tiny shops that are all boarded up. \n<\/p><p> It is an unfair analogy, because when a tiny shop gets boarded up it\nmeans that some small shopkeeper has lost his business. Of course nothing\nof the kind happens when \"wc\" becomes subsumed into one of Microsoft\nWord's countless menu items. The only real drawback is a loss of\nflexibility for the user, but it is a loss that most customers obviously\ndo not notice or care about. The most serious drawback to the Wal-Mart\napproach is that most users only want or need a tiny fraction of what is\ncontained in these giant software packages. The remainder is clutter, dead\nweight.  And yet the user in the next cubicle over will have completely\ndifferent opinions as to what is useful and what isn't. \n<\/p><p> The other important thing to mention, here, is that Microsoft has\nincluded a genuinely cool feature in the Office package: a Basic\nprogramming package. Basic is the first computer language that I learned,\nback when I was using the paper tape and the teletype.  By using the\nversion of Basic that comes with Office you can write your own little\nutility programs that know how to interact with all of the little\ndoohickeys, gewgaws, bells, and whistles in Office. Basic is easier to use\nthan the languages typically employed in Unix command-line programming,\nand Office has reached many, many more people than the GNU tools. And so\nit is quite possible that this feature of Office will, in the end, spawn\nmore hacking than GNU. \n<\/p><p> But now I'm talking about application software, not operating systems.\nAnd as I've said, Microsoft's application software tends to be very good\nstuff. I don't use it very much, because I am nowhere near their target\nmarket. If Microsoft ever makes a software package that I use and like,\nthen it really will be time to dump their stock, because I am a market\nsegment of one. \n<\/p><p> GEEK FATIGUE\n<\/p><p> Over the years that I've been working with Linux I have filled three\nand a half notebooks logging my experiences. I only begin writing things\ndown when I'm doing something complicated, like setting up X Windows or\nfooling around with my Internet connection, and so these notebooks contain\nonly the record of my struggles and frustrations. When things are going\nwell for me, I'll work along happily for many months without jotting down\na single note. So these notebooks make for pretty bleak reading.  Changing\nanything under Linux is a matter of opening up various of those little\nASCII text files and changing a word here and a character there, in ways\nthat are extremely significant to how the system operates. \n<\/p><p> Many of the files that control how Linux operates are nothing more\nthan command lines that became so long and complicated that not even Linux\nhackers could type them correctly. When working with something as powerful\nas Linux, you can easily devote a full half-hour to engineering a single\ncommand line. For example, the \"find\" command, which searches your file\nsystem for files that match certain criteria, is fantastically powerful\nand general.  Its \"man\" is eleven pages long, and these are pithy pages;\nyou could easily expand them into a whole book. And if that is not\ncomplicated enough in and of itself, you can always pipe the output of one\nUnix command to the input of another, equally complicated one. The \"pon\"\ncommand, which is used to fire up a PPP connection to the Internet,\nrequires so much detailed information that it is basically impossible to\nlaunch it entirely from the command line. Instead you abstract big chunks\nof its input into three or four different files. You need a dialing\nscript, which is effectively a little program telling it how to dial the\nphone and respond to various events; an options file, which lists up to\nabout sixty different options on how the PPP connection is to be set up;\nand a secrets file, giving information about your password. \n<\/p><p> Presumably there are godlike Unix hackers somewhere in the world who\ndon't need to use these little scripts and options files as crutches, and\nwho can simply pound out fantastically complex command lines without\nmaking typographical errors and without having to spend hours flipping\nthrough documentation. But I'm not one of them. Like almost all Linux\nusers, I depend on having all of those details hidden away in thousands of\nlittle ASCII text files, which are in turn wedged into the recesses of the\nUnix filesystem. When I want to change something about the way my system\nworks, I edit those files. I know that if I don't keep track of every\nlittle change I've made, I won't be able to get your system back in\nworking order after I've gotten it all messed up. Keeping hand-written\nlogs is tedious, not to mention kind of anachronistic. But it's necessary. \n<\/p><p> I probably could have saved myself a lot of headaches by doing\nbusiness with a company called Cygnus Support, which exists to provide\nassistance to users of free software. But I didn't, because I wanted to\nsee if I could do it myself. The answer turned out to be yes, but just\nbarely. And there are many tweaks and optimizations that I could probably\nmake in my system that I have never gotten around to attempting, partly\nbecause I get tired of being a Morlock some days, and partly because I am\nafraid of fouling up a system that generally works well. \n<\/p><p> Though Linux works for me and many other users, its sheer power and\ngenerality is its Achilles' heel. If you know what you are doing, you can\nbuy a cheap PC from any computer store, throw away the Windows discs that\ncome with it, turn it into a Linux system of mind-boggling complexity and\npower. You can hook it up to twelve other Linux boxes and make it into\npart of a parallel computer. You can configure it so that a hundred\ndifferent people can be logged onto it at once over the Internet, via as\nmany modem lines, Ethernet cards, TCP\/IP sockets, and packet radio links.\nYou can hang half a dozen different monitors off of it and play DOOM with\nsomeone in Australia while tracking communications satellites in orbit and\ncontrolling your house's lights and thermostats and streaming live video\nfrom your web-cam and surfing the Net and designing circuit boards on the\nother screens. But the sheer power and complexity of the system--the\nqualities that make it so vastly technically superior to other\nOSes--sometimes make it seem too formidable for routine day-to-day use. \n<\/p><p> Sometimes, in other words, I just want to go to Disneyland. \n<\/p><p> The ideal OS for me would be one that had a well-designed GUI that was\neasy to set up and use, but that included terminal windows where I could\nrevert to the command line interface, and run GNU software, when it made\nsense. A few years ago, Be Inc.  invented exactly that OS. It is called\nthe BeOS. \n<\/p><p> ETRE\n<\/p><p> Many people in the computer business have had a difficult time\ngrappling with Be, Incorporated, for the simple reason that nothing about\nit seems to make any sense whatsoever. It was launched in late 1990, which\nmakes it roughly contemporary with Linux. From the beginning it has been\ndevoted to creating a new operating system that is, by design,\nincompatible with all the others (though, as we shall see, it is\ncompatible with Unix in some very important ways). If a definition of\n\"celebrity\" is someone who is famous for being famous, then Be is an\nanti-celebrity. It is famous for not being famous; it is famous for being\ndoomed. But it has been doomed for an awfully long time. \n<\/p><p> Be's mission might make more sense to hackers than to other people. In\norder to explain why I need to explain the concept of cruft, which, to\npeople who write code, is nearly as abhorrent as unnecessary repetition. \n<\/p><p> If you've been to San Francisco you may have seen older buildings that\nhave undergone \"seismic upgrades,\" which frequently means that grotesque\nsuperstructures of modern steelwork are erected around buildings made in,\nsay, a Classical style. When new threats arrive--if we have an Ice Age,\nfor example--additional layers of even more high-tech stuff may be\nconstructed, in turn, around these, until the original building is like a\nholy relic in a cathedral--a shard of yellowed bone enshrined in half a\nton of fancy protective junk. \n<\/p><p> Analogous measures can be taken to keep creaky old operating systems\nworking. It happens all the time. Ditching an worn-out old OS ought to be\nsimplified by the fact that, unlike old buildings, OSes have no aesthetic\nor cultural merit that makes them intrinsically worth saving. But it\ndoesn't work that way in practice. If you work with a computer, you have\nprobably customized your \"desktop,\" the environment in which you sit down\nto work every day, and spent a lot of money on software that works in that\nenvironment, and devoted much time to familiarizing yourself with how it\nall works. This takes a lot of time, and time is money. As already\nmentioned, the desire to have one's interactions with complex technologies\nsimplified through the interface, and to surround yourself with virtual\ntchotchkes and lawn ornaments, is natural and pervasive--presumably a\nreaction against the complexity and formidable abstraction of the computer\nworld. Computers give us more choices than we really want. We prefer to\nmake those choices once, or accept the defaults handed to us by software\ncompanies, and let sleeping dogs lie. But when an OS gets changed, all the\ndogs jump up and start barking. \n<\/p><p> The average computer user is a technological antiquarian who doesn't\nreally like things to change. He or she is like an urban professional who\nhas just bought a charming fixer-upper and is now moving the furniture and\nknicknacks around, and reorganizing the kitchen cupboards, so that\neverything's just right. If it is necessary for a bunch of engineers to\nscurry around in the basement shoring up the foundation so that it can\nsupport the new cast-iron claw-foot bathtub, and snaking new wires and\npipes through the walls to supply modern appliances, why, so be\nit--engineers are cheap, at least when millions of OS users split the cost\nof their services. \n<\/p><p> Likewise, computer users want to have the latest Pentium in their\nmachines, and to be able to surf the web, without messing up all the stuff\nthat makes them feel as if they know what the hell is going on. Sometimes\nthis is actually possible. Adding more RAM to your system is a good\nexample of an upgrade that is not likely to screw anything up. \n<\/p><p> Alas, very few upgrades are this clean and simple. Lawrence Lessig,\nthe whilom Special Master in the Justice Department's antitrust suit\nagainst Microsoft, complained that he had installed Internet Explorer on\nhis computer, and in so doing, lost all of his bookmarks--his personal\nlist of signposts that he used to navigate through the maze of the\nInternet. It was as if he'd bought a new set of tires for his car, and\nthen, when pulling away from the garage, discovered that, owing to some\ninscrutable side-effect, every signpost and road map in the world had been\ndestroyed. If he's like most of us, he had put a lot of work into\ncompiling that list of bookmarks. This is only a small taste of the sort\nof trouble that upgrades can cause. Crappy old OSes have value in the\nbasically negative sense that changing to new ones makes us wish we'd\nnever been born. \n<\/p><p> All of the fixing and patching that engineers must do in order to give\nus the benefits of new technology without forcing us to think about it, or\nto change our ways, produces a lot of code that, over time, turns into a\ngiant clot of bubble gum, spackle, baling wire and duct tape surrounding\nevery operating system. In the jargon of hackers, it is called \"cruft.\" An\noperating system that has many, many layers of it is described as\n\"crufty.\"  Hackers hate to do things twice, but when they see something\ncrufty, their first impulse is to rip it out, throw it away, and start\nanew. \n<\/p><p> If Mark Twain were brought back to San Francisco today and dropped\ninto one of these old seismically upgraded buildings, it would look just\nthe same to him, with all the doors and windows in the same places--but if\nhe stepped outside, he wouldn't recognize it. And--if he'd been brought\nback with his wits intact--he might question whether the building had been\nworth going to so much trouble to save. At some point, one must ask the\nquestion: is this really worth it, or should we maybe just tear it down\nand put up a good one? Should we throw another human wave of structural\nengineers at stabilizing the Leaning Tower of Pisa, or should we just let\nthe damn thing fall over and build a tower that doesn't suck? \n<\/p><p> Like an upgrade to an old building, cruft always seems like a good\nidea when the first layers of it go on--just routine maintenance, sound\nprudent management. This is especially true if (as it were) you never look\ninto the cellar, or behind the drywall. But if you are a hacker who spends\nall his time looking at it from that point of view, cruft is fundamentally\ndisgusting, and you can't avoid wanting to go after it with a crowbar. Or,\nbetter yet, simply walk out of the building--let the Leaning Tower of Pisa\nfall over--and go make a new one THAT DOESN'T LEAN. \n<\/p><p> For a long time it was obvious to Apple, Microsoft, and their\ncustomers that the first generation of GUI operating systems was doomed,\nand that they would eventually need to be ditched and replaced with\ncompletely fresh ones. During the late Eighties and early Nineties, Apple\nlaunched a few abortive efforts to make fundamentally new post-Mac OSes\nsuch as Pink and Taligent. When those efforts failed they launched a new\nproject called Copland which also failed. In 1997 they flirted with the\nidea of acquiring Be, but instead they acquired Next, which has an OS\ncalled NextStep that is, in effect, a variant of Unix. As these efforts\nwent on, and on, and on, and failed and failed and failed, Apple's\nengineers, who were among the best in the business, kept layering on the\ncruft. They were gamely trying to turn the little toaster into a\nmulti-tasking, Internet-savvy machine, and did an amazingly good job of it\nfor a while--sort of like a movie hero running across a jungle river by\nhopping across crocodiles' backs. But in the real world you eventually run\nout of crocodiles, or step on a really smart one. \n<\/p><p> Speaking of which, Microsoft tackled the same problem in a\nconsiderably more orderly way by creating a new OS called Windows NT,\nwhich is explicitly intended to be a direct competitor of Unix. NT stands\nfor \"New Technology\" which might be read as an explicit rejection of\ncruft. And indeed, NT is reputed to be a lot less crufty than what MacOS\neventually turned into; at one point the documentation needed to write\ncode on the Mac filled something like 24 binders. Windows 95 was, and\nWindows 98 is, crufty because they have to be backward-compatible with\nolder Microsoft OSes. Linux deals with the cruft problem in the same way\nthat Eskimos supposedly dealt with senior citizens: if you insist on using\nold versions of Linux software, you will sooner or later find yourself\ndrifting through the Bering Straits on a dwindling ice floe. They can get\naway with this because most of the software is free, so it costs nothing\nto download up-to-date versions, and because most Linux users are\nMorlocks. \n<\/p><p> The great idea behind BeOS was to start from a clean sheet of paper\nand design an OS the right way. And that is exactly what they did. This\nwas obviously a good idea from an aesthetic standpoint, but does not a\nsound business plan make. Some people I know in the GNU\/Linux world are\nannoyed with Be for going off on this quixotic adventure when their\nformidable skills could have been put to work helping to promulgate Linux. \n<\/p><p> Indeed, none of it makes sense until you remember that the founder of\nthe company, Jean-Louis Gassee, is from France--a country that for many\nyears maintained its own separate and independent version of the English\nmonarchy at a court in St.  Germaines, complete with courtiers, coronation\nceremonies, a state religion and a foreign policy. Now, the same annoying\nyet admirable stiff-neckedness that gave us the Jacobites, the force de\nfrappe, Airbus, and ARRET signs in Quebec, has brought us a really cool\noperating system. I fart in your general direction, Anglo-Saxon pig-dogs! \n<\/p><p> To create an entirely new OS from scratch, just because none of the\nexisting ones was exactly right, struck me as an act of such colossal\nnerve that I felt compelled to support it. I bought a BeBox as soon as I\ncould. The BeBox was a dual-processor machine, powered by Motorola chips,\nmade specifically to run the BeOS; it could not run any other operating\nsystem. That's why I bought it.  I felt it was a way to burn my bridges.\nIts most distinctive feature is two columns of LEDs on the front panel\nthat zip up and down like tachometers to convey a sense of how hard each\nprocessor is working. I thought it looked cool, and besides, I reckoned\nthat when the company went out of business in a few months, my BeBox would\nbe a valuable collector's item. \n<\/p><p> Now it is about two years later and I am typing this on my BeBox.  The\nLEDs (Das Blinkenlights, as they are called in the Be community) flash\nmerrily next to my right elbow as I hit the keys. Be, Inc. is still in\nbusiness, though they stopped making BeBoxes almost immediately after I\nbought mine. They made the sad, but probably quite wise decision that\nhardware was a sucker's game, and ported the BeOS to Macintoshes and Mac\nclones.  Since these used the same sort of Motorola chips that powered the\nBeBox, this wasn't especially hard. \n<\/p><p> Very soon afterwards, Apple strangled the Mac-clone makers and\nrestored its hardware monopoly. So, for a while, the only new machines\nthat could run BeOS were made by Apple. \n<\/p><p> By this point Be, like Spiderman with his Spider-sense, had developed\na keen sense of when they were about to get crushed like a bug. Even if\nthey hadn't, the notion of being dependent on Apple--so frail and yet so\nvicious--for their continued existence should have put a fright into\nanyone. Now engaged in their own crocodile-hopping adventure, they ported\nthe BeOS to Intel chips--the same chips used in Windows machines. And not\na moment too soon, for when Apple came out with its new top-of-the-line\nhardware, based on the Motorola G3 chip, they withheld the technical data\nthat Be's engineers would need to make the BeOS run on those machines.\nThis would have killed Be, just like a slug between the eyes, if they\nhadn't made the jump to Intel. \n<\/p><p> So now BeOS runs on an assortment of hardware that is almost\nincredibly motley: BeBoxes, aging Macs and Mac orphan-clones, and Intel\nmachines that are intended to be used for Windows. Of course the latter\ntype are ubiquitous and shockingly cheap nowadays, so it would appear that\nBe's hardware troubles are finally over. Some German hackers have even\ncome up with a Das Blinkenlights replacement: it's a circuit board kit\nthat you can plug into PC-compatible machines running BeOS. It gives you\nthe zooming LED tachometers that were such a popular feature of the BeBox. \n<\/p><p> My BeBox is already showing its age, as all computers do after a\ncouple of years, and sooner or later I'll probably have to replace it with\nan Intel machine. Even after that, though, I will still be able to use it.\nBecause, inevitably, someone has now ported Linux to the BeBox. \n<\/p><p> At any rate, BeOS has an extremely well-thought-out GUI built on a\ntechnological framework that is solid. It is based from the ground up on\nmodern object-oriented software principles. BeOS software consists of\nquasi-independent software entities called objects, which communicate by\nsending messages to each other. The OS itself is made up of such objects,\nand serves as a kind of post office or Internet that routes messages to\nand fro, from object to object. The OS is multi-threaded, which means that\nlike all other modern OSes it can walk and chew gum at the same time;  but\nit gives programmers a lot of power over spawning and terminating threads,\nor independent sub-processes. It is also a multi-processing OS, which\nmeans that it is inherently good at running on computers that have more\nthan one CPU (Linux and Windows NT can also do this proficiently). \n<\/p><p> For this user, a big selling point of BeOS is the built-in Terminal\napplication, which enables you to open up windows that are equivalent to\nthe xterm windows in Linux. In other words, the command line interface is\navailable if you want it. And because BeOS hews to a certain standard\ncalled POSIX, it is capable of running most of the GNU software. That is\nto say that the vast array of command-line software developed by the GNU\ncrowd will work in BeOS terminal windows without complaint. This includes\nthe GNU development tools-the compiler and linker. And it includes all of\nthe handy little utility programs. I'm writing this using a modern sort of\nuser-friendly text editor called Pe, written by a Dutchman named Maarten\nHekkelman, but when I want to find out how long it is, I jump to a\nterminal window and run \"wc.\" \n<\/p><p> As is suggested by the sample bug report I quoted earlier, people who\nwork for Be, and developers who write code for BeOS, seem to be enjoying\nthemselves more than their counterparts in other OSes. They also seem to\nbe a more diverse lot in general. A couple of years ago I went to an\nauditorium at a local university to see some representatives of Be put on\na dog-and-pony show. I went because I assumed that the place would be\nempty and echoing, and I felt that they deserved an audience of at least\none. In fact, I ended up standing in an aisle, for hundreds of students\nhad packed the place. It was like a rock concert. One of the two Be\nengineers on the stage was a black man, which unfortunately is a very odd\nthing in the high-tech world. The other made a ringing denunciation of\ncruft, and extolled BeOS for its cruft-free qualities, and actually came\nout and said that in ten or fifteen years, when BeOS had become all crufty\nlike MacOS and Windows 95, it would be time to simply throw it away and\ncreate a new OS from scratch. I doubt that this is an official Be, Inc.\npolicy, but it sure made a big impression on everyone in the room! During\nthe late Eighties, the MacOS was, for a time, the OS of cool\npeople-artists and creative-minded hackers-and BeOS seems to have the\npotential to attract the same crowd now. Be mailing lists are crowded with\nhackers with names like Vladimir and Olaf and Pierre, sending flames to\neach other in fractured techno-English. \n<\/p><p> The only real question about BeOS is whether or not it is doomed. \n<\/p><p> Of late, Be has responded to the tiresome accusation that they are\ndoomed with the assertion that BeOS is \"a media operating system\" made for\nmedia content creators, and hence is not really in competition with\nWindows at all. This is a little bit disingenuous. To go back to the car\ndealership analogy, it is like the Batmobile dealer claiming that he is\nnot really in competition with the others because his car can go three\ntimes as fast as theirs and is also capable of flying. \n<\/p><p> Be has an office in Paris, and, as mentioned, the conversation on Be\nmailing lists has a strongly European flavor. At the same time they have\nmade strenuous efforts to find a niche in Japan, and Hitachi has recently\nbegun bundling BeOS with their PCs. So if I had to make wild guess I'd say\nthat they are playing Go while Microsoft is playing chess. They are\nstaying clear, for now, of Microsoft's overwhelmingly strong position in\nNorth America. They are trying to get themselves established around the\nedges of the board, as it were, in Europe and Japan, where people may be\nmore open to alternative OSes, or at least more hostile to Microsoft, than\nthey are in the United States. \n<\/p><p> What holds Be back in this country is that the smart people are afraid\nto look like suckers. You run the risk of looking naive when you say \"I've\ntried the BeOS and here's what I think of it.\"  It seems much more\nsophisticated to say \"Be's chances of carving out a new niche in the\nhighly competitive OS market are close to nil.\" \n<\/p><p> It is, in techno-speak, a problem of mindshare. And in the OS\nbusiness, mindshare is more than just a PR issue; it has direct effects on\nthe technology itself. All of the peripheral gizmos that can be hung off\nof a personal computer--the printers, scanners, PalmPilot interfaces, and\nLego Mindstorms--require pieces of software called drivers. Likewise,\nvideo cards and (to a lesser extent) monitors need drivers. Even the\ndifferent types of motherboards on the market relate to the OS in\ndifferent ways, and separate code is required for each one. All of this\nhardware-specific code must not only written but also tested, debugged,\nupgraded, maintained, and supported. Because the hardware market has\nbecome so vast and complicated, what really determines an OS's fate is not\nhow good the OS is technically, or how much it costs, but rather the\navailability of hardware-specific code. Linux hackers have to write that\ncode themselves, and they have done an amazingly good job of keeping up to\nspeed. Be, Inc. has to write all their own drivers, though as BeOS has\nbegun gathering momentum, third-party developers have begun to contribute\ndrivers, which are available on Be's web site. \n<\/p><p> But Microsoft owns the high ground at the moment, because it doesn't\nhave to write its own drivers. Any hardware maker bringing a new video\ncard or peripheral device to market today knows that it will be unsalable\nunless it comes with the hardware-specific code that will make it work\nunder Windows, and so each hardware maker has accepted the burden of\ncreating and maintaining its own library of drivers. \n<\/p><p> MINDSHARE\n<\/p><p> The U.S. Government's assertion that Microsoft has a monopoly in the\nOS market might be the most patently absurd claim ever advanced by the\nlegal mind. Linux, a technically superior operating system, is being given\naway for free, and BeOS is available at a nominal price. This is simply a\nfact, which has to be accepted whether or not you like Microsoft. \n<\/p><p> Microsoft is really big and rich, and if some of the government's\nwitnesses are to be believed, they are not nice guys. But the accusation\nof a monopoly simply does not make any sense. \n<\/p><p> What is really going on is that Microsoft has seized, for the time\nbeing, a certain type of high ground: they dominate in the competition for\nmindshare, and so any hardware or software maker who wants to be taken\nseriously feels compelled to make a product that is compatible with their\noperating systems. Since Windows-compatible drivers get written by the\nhardware makers, Microsoft doesn't have to write them; in effect, the\nhardware makers are adding new components to Windows, making it a more\ncapable OS, without charging Microsoft for the service. It is a very good\nposition to be in. The only way to fight such an opponent is to have an\narmy of highly competetent coders who write equivalent drivers for free,\nwhich Linux does. \n<\/p><p> But possession of this psychological high ground is different from a\nmonopoly in any normal sense of that word, because here the dominance has\nnothing to do with technical performance or price. The old robber-baron\nmonopolies were monopolies because they physically controlled means of\nproduction and\/or distribution. But in the software business, the means of\nproduction is hackers typing code, and the means of distribution is the\nInternet, and no one is claiming that Microsoft controls those. \n<\/p><p> Here, instead, the dominance is inside the minds of people who buy\nsoftware. Microsoft has power because people believe it does.  This power\nis very real. It makes lots of money. Judging from recent legal\nproceedings in both Washingtons, it would appear that this power and this\nmoney have inspired some very peculiar executives to come out and work for\nMicrosoft, and that Bill Gates should have administered saliva tests to\nsome of them before issuing them Microsoft ID cards. \n<\/p><p> But this is not the sort of power that fits any normal definition of\nthe word \"monopoly,\" and it's not amenable to a legal fix. The courts may\norder Microsoft to do things differently. They might even split the\ncompany up. But they can't really do anything about a mindshare monopoly,\nshort of taking every man, woman, and child in the developed world and\nsubjecting them to a lengthy brainwashing procedure. \n<\/p><p> Mindshare dominance is, in other words, a really odd sort of beast,\nsomething that the framers of our antitrust laws couldn't possibly have\nimagined. It looks like one of these modern, wacky chaos-theory phenomena,\na complexity thing, in which a whole lot of independent but connected\nentities (the world's computer users), making decisions on their own,\naccording to a few simple rules of thumb, generate a large phenomenon\n(total domination of the market by one company) that cannot be made sense\nof through any kind of rational analysis. Such phenomena are fraught with\nconcealed tipping-points and all a-tangle with bizarre feedback loops, and\ncannot be understood; people who try, end up (a) going crazy, (b) giving\nup, (c) forming crackpot theories, or (d)  becoming high-paid chaos theory\nconsultants. \n<\/p><p> Now, there might be one or two people at Microsoft who are dense\nenough to believe that mindshare dominance is some kind of stable and\nenduring position. Maybe that even accounts for some of the weirdos\nthey've hired in the pure-business end of the operation, the zealots who\nkeep getting hauled into court by enraged judges.  But most of them must\nhave the wit to understand that phenomena like these are maddeningly\nunstable, and that there's no telling what weird, seemingly\ninconsequential event might cause the system to shift into a radically\ndifferent configuration. \n<\/p><p> To put it another way, Microsoft can be confident that Thomas Penfield\nJackson will not hand down an order that the brains of everyone in the\ndeveloped world are to be summarily re-programmed. But there's no way to\npredict when people will decide, en masse, to re-program their own brains.\nThis might explain some of Microsoft's behavior, such as their policy of\nkeeping eerily large reserves of cash sitting around, and the extreme\nanxiety that they display whenever something like Java comes along. \n<\/p><p> I have never seen the inside of the building at Microsoft where the\ntop executives hang out, but I have this fantasy that in the hallways, at\nregular intervals, big red alarm boxes are bolted to the wall. Each\ncontains a large red button protected by a windowpane. A metal hammer\ndangles on a chain next to it. Above is a big sign reading: IN THE EVENT\nOF A CRASH IN MARKET SHARE, BREAK GLASS. \n<\/p><p> What happens when someone shatters the glass and hits the button, I\ndon't know, but it sure would be interesting to find out. One imagines\nbanks collapsing all over the world as Microsoft withdraws its cash\nreserves, and shrink-wrapped pallet-loads of hundred-dollar bills dropping\nfrom the skies. No doubt, Microsoft has a plan. But what I would really\nlike to know is whether, at some level, their programmers might heave a\nbig sigh of relief if the burden of writing the One Universal Interface to\nEverything were suddenly lifted from their shoulders. \n<\/p><p> THE RIGHT PINKY OF GOD\n<\/p><p> In his book The Life of the Cosmos, which everyone should read, Lee\nSmolin gives the best description I've ever read of how our universe\nemerged from an uncannily precise balancing of different fundamental\nconstants. The mass of the proton, the strength of gravity, the range of\nthe weak nuclear force, and a few dozen other fundamental constants\ncompletely determine what sort of universe will emerge from a Big Bang. If\nthese values had been even slightly different, the universe would have\nbeen a vast ocean of tepid gas or a hot knot of plasma or some other\nbasically uninteresting thing--a dud, in other words. The only way to get\na universe that's not a dud--that has stars, heavy elements, planets, and\nlife--is to get the basic numbers just right. If there were some machine,\nsomewhere, that could spit out universes with randomly chosen values for\ntheir fundamental constants, then for every universe like ours it would\nproduce 10^229 duds. \n<\/p><p> Though I haven't sat down and run the numbers on it, to me this seems\ncomparable to the probability of making a Unix computer do something\nuseful by logging into a tty and typing in command lines when you have\nforgotten all of the little options and keywords. Every time your right\npinky slams that ENTER key, you are making another try. In some cases the\noperating system does nothing. In other cases it wipes out all of your\nfiles. In most cases it just gives you an error message. In other words,\nyou get many duds. But sometimes, if you have it all just right, the\ncomputer grinds away for a while and then produces something like emacs.\nIt actually generates complexity, which is Smolin's criterion for\ninterestingness. \n<\/p><p> Not only that, but it's beginning to look as if, once you get below a\ncertain size--way below the level of quarks, down into the realm of string\ntheory--the universe can't be described very well by physics as it has\nbeen practiced since the days of Newton. If you look at a small enough\nscale, you see processes that look almost computational in nature. \n<\/p><p> I think that the message is very clear here: somewhere outside of and\nbeyond our universe is an operating system, coded up over incalculable\nspans of time by some kind of hacker-demiurge. The cosmic operating system\nuses a command-line interface. It runs on something like a teletype, with\nlots of noise and heat;  punched-out bits flutter down into its hopper\nlike drifting stars. The demiurge sits at his teletype, pounding out one\ncommand line after another, specifying the values of fundamental constants\nof physics: \n<\/p><p><code> universe -G 6.672e-11 -e 1.602e-19 -h 6.626e-34 -protonmass\n1.673e-27.... <\/code>\n<\/p><p> and when he's finished typing out the command line, his right pinky\nhesitates above the ENTER key for an aeon or two, wondering what's going\nto happen; then down it comes--and the WHACK you hear is another Big Bang. \n<\/p><p> Now THAT is a cool operating system, and if such a thing were actually\nmade available on the Internet (for free, of course)  every hacker in the\nworld would download it right away and then stay up all night long messing\nwith it, spitting out universes right and left. Most of them would be\npretty dull universes but some of them would be simply amazing. Because\nwhat those hackers would be aiming for would be much more ambitious than a\nuniverse that had a few stars and galaxies in it. Any run-of-the-mill\nhacker would be able to do that. No, the way to gain a towering reputation\non the Internet would be to get so good at tweaking your command line that\nyour universes would spontaneously develop life. And once the way to do\nthat became common knowledge, those hackers would move on, trying to make\ntheir universes develop the right kind of life, trying to find the one\nchange in the Nth decimal place of some physical constant that would give\nus an Earth in which, say, Hitler had been accepted into art school after\nall, and had ended up his days as a street artist with cranky political\nopinions. \n<\/p><p> Even if that fantasy came true, though, most users (including myself,\non certain days) wouldn't want to bother learning to use all of those\narcane commands, and struggling with all of the failures; a few dud\nuniverses can really clutter up your basement. After we'd spent a while\npounding out command lines and hitting that ENTER key and spawning dull,\nfailed universes, we would start to long for an OS that would go all the\nway to the opposite extreme: an OS that had the power to do everything--to\nlive our life for us. In this OS, all of the possible decisions we could\never want to make would have been anticipated by clever programmers, and\ncondensed into a series of dialog boxes. By clicking on radio buttons we\ncould choose from among mutually exclusive choices\n(HETEROSEXUAL\/HOMOSEXUAL). Columns of check boxes would enable us to\nselect the things that we wanted in our life (GET MARRIED\/WRITE GREAT\nAMERICAN NOVEL) and for more complicated options we could fill in little\ntext boxes (NUMBER OF DAUGHTERS: NUMBER OF SONS:). \n<\/p><p> Even this user interface would begin to look awfully complicated after\na while, with so many choices, and so many hidden interactions between\nchoices. It could become damn near unmanageable--the blinking twelve\nproblem all over again. The people who brought us this operating system\nwould have to provide templates and wizards, giving us a few default lives\nthat we could use as starting places for designing our own. Chances are\nthat these default lives would actually look pretty damn good to most\npeople, good enough, anyway, that they'd be reluctant to tear them open\nand mess around with them for fear of making them worse. So after a few\nreleases the software would begin to look even simpler: you would boot it\nup and it would present you with a dialog box with a single large button\nin the middle labeled:  LIVE. Once you had clicked that button, your life\nwould begin. If anything got out of whack, or failed to meet your\nexpectations, you could complain about it to Microsoft's Customer Support\nDepartment. If you got a flack on the line, he or she would tell you that\nyour life was actually fine, that there was not a thing wrong with it, and\nin any event it would be a lot better after the next upgrade was rolled\nout. But if you persisted, and identified yourself as Advanced, you might\nget through to an actual engineer. \n<\/p><p> What would the engineer say, after you had explained your problem, and\nenumerated all of the dissatisfactions in your life?  He would probably\ntell you that life is a very hard and complicated thing; that no interface\ncan change that; that anyone who believes otherwise is a sucker; and that\nif you don't like having choices made for you, you should start making\nyour own. \n<\/p><p> Copyright \u00a9 1999 by Neal Stephenson\n<\/p>","created_at":"2018-11-07T16:24:00+0000","updated_at":"2018-11-07T16:24:00+0000","annotations":[],"mimetype":"text\/html","reading_time":188,"domain_name":"cristal.inria.fr","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":182,"title":"Free, Open Source Mail Server Solution","url":"https:\/\/www.iredmail.org\/","content":"<p>\ufeff&#13;\n&#13;\n    <\/p>iRedMail - Free, Open Source Mail Server Solution<div class=\"pageWrapper animsition\"><div id=\"contentWrapper\"><div class=\"pageContent\"><div class=\"section\"><div class=\"container\"><p>With iRedMail, you can deploy an <strong class=\"main-color\">OPEN SOURCE<\/strong>, <strong class=\"main-color\">FULLY FLEDGED<\/strong>, <strong class=\"main-color\">FULL-FEATURED<\/strong> mail server in several minutes, for free.<\/p><p>We did the heavy lifting of putting all the open source components together and applying best practices. Our product does all the major tasks for you. Furthermore we offer professional support to back you up in case you have some problems.<\/p><\/div><\/div><div class=\"section\"><div class=\"container\"><div class=\"heading main-heading centered\"><h3 class=\"bold\">Why iRedMail<\/h3><\/div><div class=\"row\"><div class=\"col-md-4\"><i class=\"fa fa-lock fa-6x centered\">PrivacyYou have all personal data on your own hard disk, you can control the email security, inspect transaction log. No other organization can see the content of all messages.<\/i><\/div><div class=\"col-md-4\"><i class=\"fa fa-code fa-6x centered\">Open SourceAll components used in iRedMail are open source softwares, and you get the bug fixes and updates from the Linux\/BSD venders you trust. iRedMail is the right way to build your mail server with open source softwares.<\/i><\/div><div class=\"col-md-4\"><i class=\"fa fa-umbrella fa-6x centered\">Secure By DefaultEnd users are forced to use mail services through secure connections (POP3\/IMAP\/SMTP over TLS, webmail with HTTPS). Emails are encrypted in transit using TLS if possible. Passwords are stored in SSHA512 or BCRYPT (BSD).<\/i><\/div><\/div><\/div><\/div><div class=\"section\"><div class=\"container\"><div class=\"row\"><div class=\"col-md-4\"><i class=\"fa fa-envelope-o fa-6x centered\">WebmailManage mails, folders, sieve filters, vacation directly on a intuitive and easy to use web UI (Roundcube webmail or\/and SOGo groupware).<\/i><\/div><div class=\"col-md-4\"><i class=\"fa fa-calendar fa-6x centered\">Calendars\/Contacts\/ActiveSync<sup>*<\/sup>Manage your calendars (CalDAV), address books (CardDAV), tasks on a easy to use web UI or your mobile devices (iOS, Android, BlackBerry 10, Windows Phone).<br \/>* CalDAV, CardDAV and ActiveSync are offered by SOGo Groupware.<\/i><\/div><div class=\"col-md-4\"><i class=\"fa fa-user-plus fa-6x centered\">Unlimited AccountsForget about the products which pricing based on number of mailboxes, you can create as many mail accounts (domains, users, mailing lists, admins) as you want.<\/i><\/div><\/div><\/div><\/div><div class=\"section\"><div class=\"container\"><div class=\"row\"><div class=\"col-md-4\"><i class=\"fa fa-flag-o fa-6x centered\">Main Stream Linux\/BSDiRedMail works on Red Hat Enterprise Linux, CentOS, Debian, Ubuntu, FreeBSD, OpenBSD. No matter you switch to which Linux\/BSD distribution supported by iRedMail, you get the same setup in just few minutes.<\/i><\/div><div class=\"col-md-4\"><i class=\"fa fa-database fa-6x centered\">BackendsStores mail accounts in your favourte backend: OpenLDAP, MySQL, MariaDB, PostgreSQL.<\/i><\/div><div class=\"col-md-4\"><i class=\"fa fa-shield fa-6x centered\">Antispam &amp; AntivirusSpamAssassin, ClamAV, SPF, DKIM, greylisting, whitelisting, blacklisting. Quarantining detected spam into SQL database for further review.<\/i><\/div><\/div><\/div><\/div><div class=\"section\"><div class=\"container\"><div class=\"row\"><div class=\"col-md-4\"><i class=\"fa fa-tachometer fa-6x centered\">Web Admin PanelManage your mail accounts with a web admin panel. iRedMail ships a free and easy to use web admin panel along with it's product. - iRedMail also offers a separate, paid edition iRedAdmin-Pro with more features. <a href=\"https:\/\/www.iredmail.org\/admin_panel.html\" class=\"main-color\">Features list of iRedAdmin-Pro<\/a>.<\/i><\/div><div class=\"col-md-4\"><h3 class=\"bold main-color centered\">Reproduceable Deployment<\/h3><p>Get a reproduceable, easy to use, flexible, stable mail server in just few minutes, easy to migrate an old server, or restore a crashed server to new iRedMail server.<\/p><\/div><div class=\"col-md-4\"><i class=\"fa fa-life-ring fa-6x centered\">Professional Tech SupportWe back you up with fast and professional tech support. Check our <a href=\"https:\/\/www.iredmail.org\/support.html\" class=\"main-color\">full support services list<\/a>.&#13;\n                                <\/i><\/div><\/div><\/div><\/div><div class=\"section\"><div class=\"container\"><div class=\"heading main-heading centered\"><h3>Testimonials<\/h3><h4 class=\"sub-title\">What People Said<\/h4><\/div><div class=\"testimonials testimonials-3 horizontal-slider\"><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>Our (iRedMail) mail servers have been running flawlessly for years, I've never been disappointed and will continue to use iRedMail for even more years. Various mail systems promise a lot but are restricted on many points, where iRedMail just works however you need it to work.<\/p><p>&#13;\n                                            <strong class=\"main-color\">Dylan Ysmal <\/strong>, Sep 22, 2017&#13;<\/p><\/div><\/div><\/div><div class=\"testimonials testimonials-3 horizontal-slider\"><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>Having search for a solution to replace an Exchange + AD system with an OpenSource solution for a while. Did look at Kerio, Zentyal, Kolab, Kopano, Zimbra and Sogo from their respective free\/open version, and also made a deep investigation into the available documentations and forum contents.<\/p><p>So far the iRedmail packaging is the far best I've seen, and answers to almost every questions I come across can be found in the existing iRedMail documentation or on the forum site.<\/p><p>Since all modules are \"standard\" open source projects, it's easy to find  more information for specific issues, like OpenLdap or Postfix elsewhere.<\/p><p>Once I tried the installation and got it up and running in an hour or so, I didn't bother to check the other test systems again. We have now started a project to migrate our Exchange server, and to use OpenLDAP instead of Active directory for our application servers (Gitlab, Confluence, Jira, etc).<\/p><p>Thanks for a really good product.<\/p><p>&#13;\n                                            <strong class=\"main-color\"><a href=\"https:\/\/forum.iredmail.org\/post57620.html#p57620\" target=\"_blank\" rel=\"noopener\">swejun<\/a><\/strong>, Jul 26, 2017&#13;<\/p><\/div><\/div><\/div><div class=\"testimonials testimonials-3 horizontal-slider\"><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>As a long time user of iRedMail, we are constantly amazed by the stability and flexibility of the product. iRedMail consistently outperforms our clients expectations and the professional support we have received is second-to-none. We have absolutely no hesitation to recommend iRedMail to others!<\/p><p>&#13;\n                                            <strong class=\"main-color\">Mark Pike<\/strong>, March 3, 2016&#13;<\/p><\/div><\/div><\/div><div class=\"testimonials testimonials-3 horizontal-slider\"><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>We are really happy with the system, excellent work, throughout the years we have used many mail systems but this is the best.<\/p><p>&#13;\n                                            <strong class=\"main-color\">Juan Pablo Uggino<\/strong>, Dec 3, 2015&#13;<\/p><\/div><\/div><\/div><div class=\"testimonials testimonials-3 horizontal-slider\"><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>So this is my story: the more time passed, the more it felt wrong for me having a lot of my personnal data stored inside servers of giant US companies like microsoft, google... So I started to build my own personnal cloud server by renting a 7$\/month small dedicated server.<\/p><p>iRedMail made that dream possible for the mail side. It's easy to install, easy to use, support latest mail technologies and the support provided even for free by Zhang in the forum is better than professional support of other paid companies like IBM.<\/p><p>So to everyone else looking for a private mail server, iredmail is the way to go.<\/p><p>&#13;\n                                            <strong class=\"main-color\">@Neutro<\/strong>, <a href=\"https:\/\/forum.iredmail.org\/topic9819-success-stories-best-private-personnal-mail-server-for-free.html\">Sep 7, 2015 <i class=\"fa fa-external-link-square\"><\/i><\/a>&#13;<\/p><\/div><\/div><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>Great solution and a great time saver for a system administrator. iRedMail will be my first choice from now on ...<br \/>You can pay and buy a software but you cant buy such support. Thanks again.<\/p><p>&#13;\n                                            <strong class=\"main-color\">Bahadir Tecer<\/strong>, Turkiye, Call Center, <a href=\"https:\/\/forum.iredmail.org\/topic2772-turkiye-call-center-around-2000-users.html\">Dec 19, 2011 <i class=\"fa fa-external-link-square\"><\/i><\/a>&#13;<\/p><\/div><\/div><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>iRedMail is the our choice, thank's to the use of standard packeges and solutions for e-mail management. Before we used sendmail and we was searching for a system that implemented maildir + Dovecot upon LDAP for the managemet of user accounts, so iRedMail is perfect for us. Moreover it is free, has a very good support - and I emphasize *very good* - and the Admin Panel - iRedAdmin-Pro, we bought as soon as deployed iRedMail, is a very valid tool to manage the system.<\/p><p>&#13;\n                                            <strong class=\"main-color\">Andrea Di Dato<\/strong>, <a href=\"https:\/\/forum.iredmail.org\/topic1342-italy-naples-inaf-oac-research-institute-133.html\">Oct 21, 2010 <i class=\"fa fa-external-link-square\"><\/i><\/a>&#13;<\/p><\/div><\/div><div class=\"col-md-12\"><div class=\"testimonials-bg shape lg square\"><p>iRedMail is the easiest and reliable mail server with all of the features that a mail server needed for the administrators.I appreciate the developers of iRedMail who have really supported tremendously to deploy the mail server by solving all of the problem of us quickly. 3 cheers to iRedMail.<\/p><p>&#13;\n                                            <strong class=\"main-color\">Raskin Paul<\/strong>, <a href=\"https:\/\/forum.iredmail.org\/topic930-bangladeshdhakaitconnectwwwitconnectbdcom100-users.html\">May 22, 2010 <i class=\"fa fa-external-link-square\"><\/i><\/a>&#13;<\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>&#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n        &#13;\n        &#13;\n&#13;\n        &#13;\n&#13;\n&#13;\n    &#13;","created_at":"2018-11-07T16:23:59+0000","updated_at":"2018-11-07T16:23:59+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"www.iredmail.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":183,"title":"Running Lisp in Production","url":"https:\/\/tech.grammarly.com\/blog\/running-lisp-in-production","content":"<p>At Grammarly, the foundation of our business, our core grammar engine, is written\nin Common Lisp. It currently processes more than a thousand sentences\nper second, is horizontally scalable, and has reliably served in\nproduction for almost 3 years.<\/p><p>We noticed that there are very few, if any, accounts of how to deploy Lisp software to\nmodern cloud infrastructure, so we thought that it would be a good\nidea to share our experience. The Lisp runtime and programming environment\nprovides several unique, albeit obscure, capabilities to\nsupport production systems (for the impatient, they are described in\nthe final chapter).<\/p><h2 id=\"wut-lisp\">Wut Lisp?!!<\/h2><p><img src=\"https:\/\/tech.grammarly.com\/assets\/articles\/b2f-1.jpg\" alt=\"Why not?\" \/><\/p><p>Contrary to popular opinion, Lisp is an incredibly practical language\nfor building production systems. There are, in fact, many Lisp systems\nout there: when you search for an airline ticket on Hipmunk or take a\nLondon tube train, Lisp programs are being called.<\/p><p>Our Lisp services are conceptually a classical AI application\nthat operates on huge piles of knowledge created by linguists and\nresearchers. It\u2019s mostly a CPU-bound program, and is one of the\nbiggest consumers of computing resources in our network.<\/p><p>We run them on stock Linux images deployed to AWS. We use\n<a href=\"http:\/\/sbcl.org\/\">SBCL<\/a> for production deployment and\n<a href=\"http:\/\/www.clozure.com\/clozure-cl.html\">CCL<\/a> on most of the\ndevelopers\u2019 machines. One of the nice things about Lisp is that you\nhave an option of choosing from several mature implementations with\ndifferent strengths and weaknesses: in our case, we optimized for\nprocessing speed on the server, and compilation speed in the dev\nenvironment (why this is critical for us is described in the later\nsection).<\/p><h2 id=\"a-stranger-in-a-strange-land\">A stranger in a strange land<\/h2><p><img src=\"https:\/\/tech.grammarly.com\/assets\/articles\/b2f-2.jpg\" alt=\"Hello world\" \/><\/p><p>At Grammarly, we use many programming languages for developing our\nservices: in addition to JVM languages and Javascript, we also develop\nin Erlang, Python, and Go. Proper service encapsulation enables us to\nuse whatever language and platform makes the most sense. There is a\ncost to maintenance, but we value choice and freedom over rules and\nprocesses.<\/p><p>We also try to rely on simple language-agnostic infrastructure tools.\nThis approach spares us a lot of trouble integrating this zoo of\ntechnologies in our platform. For instance, statsd is a great example\nof both amazingly simple and useful service that is extremely easy to\nuse. Another one is Graylog2; it provides a brilliant\n<a href=\"https:\/\/www.graylog.org\/resources\/gelf-2\/\">specification<\/a> for\nlogging, and, although there was no ready-made library for working\nwith it from CL it was really easy to assemble from the building\nblocks already available in the Lisp ecosystem. This is all the code\nthat was needed (and most of it is just \u201cword-by-word\u201d translation of\nthe spec):<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>(defun graylog (message &amp;key level backtrace file line-no)\n  (let ((msg (salza2:compress-data\n              (babel:string-to-octets\n               (json:encode-json-to-string #{\n                 :version \"1.0\"\n                 :facility \"lisp\"\n                 :host *hostname*\n                 :|short_message| message\n                 :|full_message| backtrace\n                 :timestamp (local-time:timestamp-to-unix (local-time:now))\n                 :level level\n                 :file file\n                 :line line-no\n                })\n               :encoding :utf-8)\n              'salza2:zlib-compressor)))\n    (usocket:socket-send (usocket:socket-connect\n                          *graylog-host* *graylog-port*\n                          :protocol :datagram :element-type '(unsigned-byte 8))\n                         msg (length msg))))\n<\/pre><\/div><\/div><p>One of the common complaints about Lisp that there are no libraries in\nthe ecosystem. As you see, 5 libraries are used just in this example\nfor such things as encoding, compression, getting Unix time, and\nsocket connections.<\/p><p>Lisp libraries indeed exist, but, like all library integrations, we\nhave challenges with them as well. For instance, to plug into the\nJenkins CI system we had to use xUnit, and it was not so\nstraightforward to find the spec for it. Fortunately, this\n<a href=\"http:\/\/stackoverflow.com\/questions\/442556\/spec-for-junit-xml-output\">obscure Stackoverflow question<\/a> helped \u2014 we ended up having to build this into our own testing library\n<a href=\"https:\/\/github.com\/vseloved\/should-test\/blob\/master\/xunit.lisp\">should-test<\/a>.<\/p><p>Another example is using <a href=\"https:\/\/www.hdfgroup.org\/HDF5\/\">HDF5<\/a> for\nmachine learning models exchange: it took us some work to adapt the\nlow-level <a href=\"https:\/\/github.com\/HDFGroup\/hdf5-cffi\">hdf5-cffi<\/a> library\nto our use case, but we had to spend much more time upgrading our\nAMIs to support the current version of the C library.<\/p><p>Another principle that we try to follow in Grammarly platform is\nmaximal decoupling of different services to ensure horizontal\nscalability and operational independence \u2014 see\n<a href=\"https:\/\/tech.grammarly.com\/blog\/decoupling-a-monolithic-server-application\">my colleague\u2019s blog post<\/a>.\nThis way, we do not need to interact with databases in the critical\npaths in our core services. We do, however, use MySQL, Postgres,\nRedis, and Mongo, for internal storage, and we\u2019ve successfully used\n<a href=\"http:\/\/clsql.b9.com\/\">CLSQL<\/a>,\n<a href=\"http:\/\/marijnhaverbeke.nl\/postmodern\/\">postmodern<\/a>,\n<a href=\"https:\/\/github.com\/vseloved\/cl-redis\">cl-redis<\/a>, and\n<a href=\"https:\/\/github.com\/fons\/cl-mongo\">cl-mongo<\/a>\nto access them from the Lisp side.<\/p><p>We rely on <a href=\"https:\/\/www.quicklisp.org\">Quicklisp<\/a> for managing\nexternal dependencies and a simple system of bundling library source\ncode with the project for our internal libraries or forks. Quicklisp\nrepository hosts more than 1,000 Lisp libraries \u2014 not a mind-blowing\nnumber but quite enough for covering all of our production needs.<\/p><p>For deployment into production, we use a universal stack: the\napplication is tested and bundled by <code class=\"highlighter-rouge\">Jenkins<\/code>, put on the servers by\n<code class=\"highlighter-rouge\">Rundeck<\/code>, and run there as a regular Unix process by <code class=\"highlighter-rouge\">upstart<\/code>.<\/p><p>Overall, the problems we face with integrating a Lisp app into the\ncloud world are not radically different from the ones we encounter\nwith many other technologies. If you want to use Lisp in production,\nand to experience the joy of writing Lisp code, there is no valid\ntechnical reason not to!<\/p><h2 id=\"the-hardest-bug-ive-ever-debugged\">The hardest bug I\u2019ve ever debugged<\/h2><p><img src=\"https:\/\/tech.grammarly.com\/assets\/articles\/b2f-3.jpg\" alt=\"Call me maybe\" \/><\/p><p>As ideal as this story is so far, it has not been all rainbows and\nunicorns.<\/p><p>We\u2019ve built an esoteric application (even by Lisp standards), and in\nthe process have hit some limits of our platform. One unexpected thing\nwas heap exhaustion during compilation. We rely heavily on macros, and\nsome of the largest ones expand into thousands of lines of low-level\ncode. It turned out that SBCL compiler implements a lot of\noptimizations that allow us to enjoy quite fast generated code, but\nsome of which require exponential time and memory resources.\nUnfortunately, there\u2019s no way to influence that by turning\nthem off or tuning somehow. However, there exists a well-known general\nsolution, <a href=\"http:\/\/random-state.net\/log\/3390120648.html\">call-with-* style<\/a>,\nin which you trade off a little performance for better modularity (which\nturned out crucial for our use case) and debuggability.<\/p><p>Less surprising than compiler taming, we have spent some time with\nGC tuning to improve the latency and resource utilization in our\nsystem. SBCL provides a decent generational garbage collector,\nalthough the system is not nearly as sophisticated as in the JVM.\nWe had to tune the generation sizes, and it turned out that the best\noption was to use an over-sized heap: our application consumes 2-4\ngigabytes of memory, but we run it with 25G heap size which\nautomatically results in a huge volume for the nursery. Yet another\ncustomization we had to make \u2014 a much less obvious one \u2014 was to\nrun full GC programmatically every N minutes. With a large heap, we\nhave noticed a gradual memory usage buildup over periods of tens of\nminutes which resulted in spans of more time spent in GC and decreased\napplication throughput. Our periodic GC approach got the system into\na much more stable state with almost constant memory usage. On the\nleft you can see how an untuned system performs, and on the right \u2014\nthe effect of periodic collection.<\/p><p><img src=\"https:\/\/tech.grammarly.com\/assets\/articles\/lisp-mem.jpg\" alt=\"Running without and with periodic GC\" \/><\/p><p>Of all these challenges, the worst bug I\u2019ve ever seen was a network bug.\nAs usual with such stories, it was not a bug in the application but\n<a href=\"http:\/\/lisp-univ-etc.blogspot.com\/2013\/04\/errors-in-your-platform.html\">a problem in the underlying platform<\/a> (this time \u2014 SBCL). And, moreover, I was bitten by it\ntwice in two different services, but the first time I couldn\u2019t figure\nit out, so I had to develop a workaround.<\/p><p>As we were just beginning running our service under substantial load\nin production, after some period of normal operation all of the\nservers would suddenly start to slow down and then would become\nunresponsive. After much investigation centering on our input data, we\ndiscovered that the problem was instead a race condition in low-level\nSBCL network code, specifically in the way the socket function\n<a href=\"http:\/\/www.lehman.cuny.edu\/cgi-bin\/man-cgi?getprotobyname\">getprotobyname<\/a>,\nwhich is non-reentrant, was called. It was quite an unlikely race, so\nit manifested itself only in the high-load network service setup when\nthis function was called tens of thousands of times. It knocked off\none worker thread after another, eventually rendering the service\ncomatose.<\/p><p>Here\u2019s the fix we settled on; unfortunately, it can\u2019t be used in a\nbroader context as a library. (The bug was reported to SBCL\nmaintainers, and there was a fix there as well, but we are still\nrunning with this hack, just to be sure :)<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>#+unix\n(defun sb-bsd-sockets:get-protocol-by-name (name)\n  (case (mkeyw name)\n    (:tcp 6)\n    (:udp 17)))\n<\/pre><\/div><\/div><h2 id=\"back-to-the-future\">Back to the future<\/h2><p><img src=\"https:\/\/tech.grammarly.com\/assets\/articles\/b2f-4.jpg\" alt=\"A Lisp Machine\" \/><\/p><p>Common Lisp systems implement a lot of the ideas of the venerable Lisp\nmachines. One of the most prominent ones is the SLIME interactive\nenvironment. While the industry waits for LightTable and similar tools\nto mature, Lisp programmers have been silently and haughtily enjoying\nsuch capabilities with SLIME for many years. Witness the power of this\nfully armed and operational battlestation\n<a href=\"https:\/\/www.youtube.com\/playlist?list=PL2VAYZE_4wRKKr5pJzfYD1w4tKCXARs5y\">in action<\/a>.<\/p><p>But SLIME is not just a Lisp\u2019s take on an IDE. Being a client-server\napplication, it allows to run its backend on the remote machine and\nconnect to it from your local Emacs (or Vim, if you must, with SLIMV).\nJava programmers can think of JConsole, but here you\u2019re\nnot constrained by the predefined set of operations and can do any\nkind of introspection and changes you want. We could not have debugged\nthe socket race condition without this functionality.<\/p><p>Furthermore, the remote console is not the only useful tool provided\nby SLIME. Like many IDEs it has a jump-to-source function, but unlike\nJava or Python I have SBCL\u2019s source code on my machine, so I often\nconsult the implementation\u2019s sources, and this helps understand\nwhat\u2019s going on much better. For the socket bug case, this was also an\nimportant part of the debugging process.<\/p><p>Finally, another super-useful introspection and debugging tool we use\nis Lisp\u2019s <a href=\"http:\/\/clhs.lisp.se\/Body\/m_tracec.htm\">TRACE<\/a> facility.\nIt has completely changed my approach to debugging: from tedious local\nstepping to exploring the bigger picture. It was also instrumental in\nnailing our nasty bug.<\/p><p>With <code class=\"highlighter-rouge\">trace<\/code>, you define a function to trace, run the code, and Lisp\nprints all calls to that functions with arguments and all of its\nreturns with results. It is somewhat similar to stacktrace, but you\ndon\u2019t get to see the whole stack and you dynamically get a stream of\ntraces, not stopping the application. <code class=\"highlighter-rouge\">trace<\/code> is like <code class=\"highlighter-rouge\">print<\/code> on\nsteroids that allows you to quickly get into the inner workings of\narbitrary complex code and monitor complicated flows. The only\nshortcoming is that you can\u2019t trace macros.<\/p><p>Here\u2019s a snippet of tracing I did just today to ensure that a JSON\nrequest to one of our services is properly formatted and returns an\nexpected result:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>0: (GET-DEPS\n    (\"you think that's bad, hehe, i remember once i had an old 100MHZ dell unit i was using as a server in my room\"))\n  1: (JSON:ENCODE-JSON-TO-STRING\n      #&lt;HASH-TABLE :TEST EQL :COUNT 2 {1037DD9383}&gt;)\n    2: (JSON:ENCODE-JSON-TO-STRING \"action\")\n    2: JSON:ENCODE-JSON-TO-STRING returned \"\\\"action\\\"\"\n    2: (JSON:ENCODE-JSON-TO-STRING \"sentences\")\n    2: JSON:ENCODE-JSON-TO-STRING returned \"\\\"sentences\\\"\"\n  1: JSON:ENCODE-JSON-TO-STRING returned\n       \"{\\\"action\\\":\\\"deps\\\",\\\"sentences\\\":[\\\"you think that's bad, hehe, i remember once i had an old 100MHZ dell unit i was using as a server in my room\\\"]}\"\n0: GET-DEPS returned\n     (((\"nsubj\" 1 0) (\"ccomp\" 9 1) (\"nsubj\" 3 2) (\"ccomp\" 1 3) (\"acomp\" 3 4)\n       (\"punct\" 9 5) (\"intj\" 9 6) (\"punct\" 9 7) (\"nsubj\" 9 8) (\"root\" -1 9)\n       (\"advmod\" 9 10) (\"nsubj\" 12 11) (\"ccomp\" 9 12) (\"det\" 17 13)\n       (\"amod\" 17 14) (\"nn\" 16 15) (\"nn\" 17 16) (\"dobj\" 12 17)\n       (\"nsubj\" 20 18) (\"aux\" 20 19) (\"rcmod\" 17 20) (\"prep\" 20 21)\n       (\"det\" 23 22) (\"pobj\" 21 23) (\"prep\" 23 24)\n       (\"poss\" 26 25) (\"pobj\" 24 26)))\n     ((&lt;you 0,3&gt; &lt;think 4,9&gt; &lt;that 10,14&gt; &lt;'s 14,16&gt; &lt;bad 17,20&gt; &lt;, 20,21&gt;\n       &lt;hehe 22,26&gt; &lt;, 26,27&gt; &lt;i 28,29&gt; &lt;remember 30,38&gt; &lt;once 39,43&gt;\n       &lt;i 44,45&gt; &lt;had 46,49&gt; &lt;an 50,52&gt; &lt;old 53,56&gt; &lt;100MHZ 57,63&gt;\n       &lt;dell 64,68&gt; &lt;unit 69,73&gt; &lt;i 74,75&gt; &lt;was 76,79&gt; &lt;using 80,85&gt;\n       &lt;as 86,88&gt; &lt;a 89,90&gt; &lt;server 91,97&gt; &lt;in 98,100&gt; &lt;my 101,103&gt;\n       &lt;room 104,108&gt;))\n<\/pre><\/div><\/div><p>So, to debug our nasty socket bug, I had to dig deep into the SBCL\nnetwork code and study the functions being called, then connect via\nSLIME to the failing server and try tracing one function after\nanother. And when I got a call that didn\u2019t return, that was it.\nFinally, looking at <code class=\"highlighter-rouge\">man<\/code> to find out that this function isn\u2019t\nre-entrant and encountering some references to that in the SBCL\u2019s\nsource code comments allowed me to verify this hypothesis.<\/p><p>That said, Lisp proved to be a remarkably reliable platform for one of\nour most critical projects. It is quite fit for the common\nrequirements of a modern cloud infrastructure, and although this stack\nis not very well-known and popular, it has its own strong points \u2014\nyou just have to learn how to use them. Not to mention the power of\nthe Lisp approach to solving difficult problems for which we love it.\nBut that\u2019s a whole different story\u2026<\/p>","created_at":"2018-11-07T16:23:59+0000","updated_at":"2018-11-07T16:23:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"tech.grammarly.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":184,"title":"How to Remove the Sidebar in WordPress","url":"https:\/\/www.wpbeginner.com\/wp-themes\/how-to-remove-the-sidebar-in-wordpress\/","content":"<p>Do you want to remove the sidebar from your WordPress site? A sidebar is the widget-ready area in your WordPress theme. It often appears at the right or left side or below the content area on every page. In this article, we will show you how to easily remove the sidebar in WordPress. <\/p><p><img title=\"How to Remove the Sidebar in WordPress\" src=\"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/removewpsidebar.png\" alt=\"How to Remove the Sidebar in WordPress\" width=\"550\" height=\"340\" class=\"alignnone size-full wp-image-43547\" \/><\/p><h4>Why Remove the Sidebar in WordPress?<\/h4><p>Most <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/decide-premium-free-wordpress-themes\/\" title=\"Free vs Premium WordPress Themes (Pros and Cons)\">free and paid WordPress themes<\/a> come with multiple <a href=\"https:\/\/www.wpbeginner.com\/glossary\/sidebar\/\" title=\"What is Sidebar in WordPress?\">sidebars<\/a> or widget ready areas. <\/p><p>The purpose of sidebars in WordPress is to display items that are not part of the post or page content. Usually these are links to different areas of your website, advertisements, <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/12-tested-and-easy-ways-to-grow-your-email-list-faster\/\" title=\"12 Tested and Easy Ways to Grow Your Email List Faster\">email list signup forms<\/a>, social media profiles, etc. These items can be easily added to the sidebar of a theme using <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/how-to-add-and-use-widgets-in-wordpress\/\" title=\"How to Add and Use Widgets in WordPress\">WordPress widgets<\/a>. <\/p><p><img title=\"Sidebar on a WordPress powered website\" src=\"https:\/\/cdn3.wpbeginner.com\/wp-content\/uploads\/2017\/04\/sidebar.jpg\" alt=\"Sidebar on a WordPress powered website\" width=\"550\" height=\"344\" class=\"alignnone size-full wp-image-43540\" \/><\/p><p>In most WordPress themes, sidebar appears differently when a user is viewing your site from a mobile device. Due to reduced width of the screen size, sidebars that appear on right or left on the desktop are moved to the bottom on mobile screens. <\/p><p>Sidebars are immensely helpful and can help you grow your business. See these <a href=\"https:\/\/www.wpbeginner.com\/showcase\/12-wordpress-sidebar-tricks-to-get-maximum-results\/\" title=\"12 WordPress Sidebar Tricks to Get Maximum Results\">WordPress sidebar tricks<\/a> for maximum results. <\/p><p>However, some website owners feel that removing sidebar helps improve user experience. If you\u2019re looking to remove the sidebar on your WordPress site, then this tutorial is for you.<\/p><h4>Video Tutorial<\/h4><iframe class=\"youtube-player\" width=\"520\" height=\"323\" src=\"https:\/\/www.youtube.com\/embed\/mO-QIZYFNPo?version=3&amp;rel=0&amp;fs=1&amp;showsearch=0&amp;showinfo=1&amp;iv_load_policy=1&amp;wmode=transparent\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><div class=\"yt-sbscrb-bar\"><p>Subscribe to WPBeginner<\/p><iframe src=\"https:\/\/www.youtube.com\/subscribe_embed?usegapi=1&amp;channel=wpbeginner&amp;layout=default&amp;count=default&amp;origin=http%3A%2F%2Fwpbeginner.com&amp;gsrc=3p&amp;ic=1&amp;jsh=m%3B%2F_%2Fscs%2Fapps-static%2F_%2Fjs%2Fk%3Doz.gapi.en_GB.Rx51stRkYnQ.O%2Fm%3D__features__%2Fam%3DAQ%2Frt%3Dj%2Fd%3D1%2Ft%3Dzcms%2Frs%3DAGLTcCM5XUrLkPZ9aFXqtSLivsifryqsUA#_methods=onPlusOne%2C_ready%2C_close%2C_open%2C_resizeMe%2C_renderstart%2Concircled%2Cdrefresh%2Cerefresh%2Conload&amp;id=I0_1448178294715&amp;parent=http%3A%2F%2Fwpbeginner.com\"> <\/iframe><\/div><p>If you don\u2019t like the video or need more instructions, then continue reading.<\/p><h4>Removing Sidebars From Everywhere in WordPress<\/h4><p>This method allows you to simply remove sidebars from every <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/what-is-the-difference-between-posts-vs-pages-in-wordpress\/\" title=\"What is the Difference Between Posts vs. Pages in WordPress\">page and post<\/a> on your WordPress site. It requires you to edit WordPress theme files. If you haven\u2019t done this before then see our guide on <a href=\"https:\/\/www.wpbeginner.com\/beginners-guide\/beginners-guide-to-pasting-snippets-from-the-web-into-wordpress\/\" title=\"Beginner\u2019s Guide to Pasting Snippets from the Web into WordPress\">how to copy and paste code in WordPress<\/a>. <\/p><p>First you need to connect to your WordPress site <a href=\"https:\/\/www.wpbeginner.com\/showcase\/6-best-ftp-clients-for-wordpress-users\/\" title=\"6 Best FTP Clients for Mac and Windows WordPress Users\">using an FTP client<\/a> and then go to \/wp-content\/themes\/your-current-theme\/ folder. <\/p><p>WordPress themes are made up of different templates, so you will need to edit all the templates where a sidebar is displayed. <\/p><p>For example, in a typical WordPress theme you may need to edit index.php, page.php, single.php, archive.php, home.php and so on. <\/p><p>Open a template file to edit it and then locate the line that looks like this: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php get_sidebar(); ?&gt;&#13;\n<\/pre><p>If your theme comes with multiple sidebars, then you will see different instances of this code with a sidebar name inside the function. For example: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php get_sidebar('footer-widget-area'); ?&gt;&#13;\n<\/pre><p>Delete the line that represents the sidebar that you don\u2019t want to display.<\/p><p>Now, save and upload the file back to your website. Repeat the process for all template files responsible for displaying different pages on your website. <\/p><p>Once finished, you can go to your website and see it in action. <\/p><p>You may notice that while sidebars are gone, your content area is still appearing with the same width leaving the sidebar area empty. <\/p><p><img title=\"Blank sidebar area in a WordPress theme\" src=\"https:\/\/cdn.wpbeginner.com\/wp-content\/uploads\/2017\/04\/blanksidebar.jpg\" alt=\"Blank sidebar area in a WordPress theme\" width=\"550\" height=\"344\" class=\"alignnone size-full wp-image-43542\" \/><\/p><p>That\u2019s because your theme has defined the width for the content area. Now that the sidebar is gone, you need to adjust the width of the content area by <a href=\"https:\/\/www.wpbeginner.com\/plugins\/how-to-easily-add-custom-css-to-your-wordpress-site\/\" title=\"How to Easily Add Custom CSS to Your WordPress Site\">adding custom CSS<\/a> to your theme. <\/p><p>First you will need to figure out the CSS class used by your theme to define the width of the content area. You can do that by <a href=\"https:\/\/www.wpbeginner.com\/wp-tutorials\/basics-of-inspect-element-with-your-wordpress-site\/\" title=\"Basics of Inspect Element: Customizing WordPress for DIY Users\">using the Inspect tool<\/a> in your browser. <\/p><p>Here is the sample CSS we used on our demo site using Twenty Sixteen theme. <\/p><pre class=\"brush: css; title: ; notranslate\" title=\"\">&#13;\n.content-area {&#13;\n    width: 100%;&#13;\n    margin: 0px;&#13;\n    border: 0px;&#13;\n    padding: 0px;&#13;\n}&#13;\n&#13;\n.content-area .site {&#13;\nmargin:0px;&#13;\n}&#13;\n<\/pre><p>Here is how it looked without sidebars: <\/p><p><img title=\"Sidebars removed from a WordPress theme\" src=\"https:\/\/cdn3.wpbeginner.com\/wp-content\/uploads\/2017\/04\/nosidebarwp.jpg\" alt=\"Sidebars removed from a WordPress theme\" width=\"550\" height=\"344\" class=\"alignnone size-full wp-image-43543\" \/><\/p><h4>Removing Sidebars From Individual Pages in WordPress<\/h4><p>It\u2019s possible that you only want to remove the sidebars on certain pages while displaying them on other sections of your website. <\/p><p>For example, a lot of websites display sidebars on their blog page but not on individual blog posts. Some prefer to show sidebars everywhere except static pages. <\/p><p>Let\u2019s take a look at different scenarios and how to remove sidebars from different areas. <\/p><h4>Removing Sidebars from a Static Page in WordPress<\/h4><p>Many WordPress themes come with ready-to use full width page templates. These templates can be used with any static page in WordPress and do not display sidebars on either side of the content. <\/p><p>You will just need to edit the page where you want to remove the sidebar and then select the full width template under \u2018Page Attributes\u2019 meta box. <\/p><p><img title=\"Full width page template with no sidebars\" src=\"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/fullwidth-template.png\" alt=\"Full width page template with no sidebars\" width=\"550\" height=\"355\" class=\"alignnone size-full wp-image-43537\" \/><\/p><p>If you do not see an option to select a template, then this means that your theme does not come with a full-width page template. <\/p><p>In that case, you will need to create a full-width page template manually. Open a plain text editor like Notepad and paste the following code in a blank file: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php&#13;\n\/*&#13;\n*&#13;\nTemplate Name: Full-Width&#13;\n*\/&#13;\nget_header(); ?&gt;&#13;\n<\/pre><p>Now save this file as full-width.php on your desktop. <\/p><p>Connect to your WordPress site using an FTP client and then go to \/wp-content\/themes\/your-current-theme\/ folder. Edit  the file called page.php and copy every thing that appears after <code>&lt;?php get_header(); ?&gt;<\/code> line. <\/p><p>Now paste this code in your full-width.php file. After that you need to locate and delete the line that looks like this: <\/p><pre class=\"brush: php; title: ; notranslate\" title=\"\">&#13;\n&lt;?php get_sidebar(); ?&gt;&#13;\n<\/pre><p>Save your changes and upload the full-width.php file to your current theme folder. <\/p><p>You can now edit any page on your WordPress site and you will be able to select your full width page template.<\/p><p>For more details, see our guide on <a href=\"https:\/\/www.wpbeginner.com\/wp-themes\/how-to-create-a-full-width-page-in-wordpress\/\" title=\"How to Create a Full Width Page in WordPress\">how to create a full width page template in WordPress<\/a>. <\/p><h4>Remove Sidebar from Blog Posts Page in WordPress<\/h4><p>WordPress allows you to use a static front page and a separate page for your blog posts. Usually, most WordPress themes display sidebars on the blog posts page. <\/p><p>If you don\u2019t want to display sidebars on the blog posts page, then you need to go to <strong>Settings \u00bb Reading<\/strong> page. Notice the name of the page you are using as your blog posts page. <\/p><p><img title=\"Find your WordPress blog page\" src=\"https:\/\/cdn3.wpbeginner.com\/wp-content\/uploads\/2017\/04\/blogpage.jpg\" alt=\"Find your WordPress blog page\" width=\"550\" height=\"267\" class=\"alignnone size-full wp-image-43544\" \/><\/p><p>Now go to <strong>Pages \u00bb All Pages<\/strong> and edit the page used to display your blog posts. Under the page attribute section select a full width template and then save your changes. <\/p><p><img title=\"Full width page template with no sidebars\" src=\"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/fullwidth-template.png\" alt=\"Full width page template with no sidebars\" width=\"550\" height=\"355\" class=\"alignnone size-full wp-image-43537\" \/><\/p><p>If you don\u2019t see a full width template option, then you will have to manually create a full width template. See the previous section in this article for detailed instructions. <\/p><h4>Remove Sidebar from a Single Post in WordPress<\/h4><p>If you want to remove sidebar from certain single posts, then you can create a custom single post template. <\/p><p>WordPress 4.7 came with built-in support for post templates. You\u2019d need to follow the steps outline in our guide on <a href=\"https:\/\/www.wpbeginner.com\/wp-themes\/create-custom-single-post-templates-for-specific-posts-or-sections-in-wordpress\/\" title=\"How to Create Custom Single Post Templates in WordPress\">how to create custom single post templates in WordPress<\/a>.<\/p><p>In your custom single post template all you need to do is make sure to remove the sidebar part of the code.<\/p><p>We hope this article helped you learn how to easily remove the sidebar in WordPress themes. You may also want to see our step by step guide on how to <a href=\"https:\/\/www.wpbeginner.com\/wordpress-performance-speed\/\" title=\"The Ultimate Guide to Boost WordPress Speed &amp; Performance\">boost WordPress speed and performance<\/a>. <\/p><p>If you liked this article, then please subscribe to our <a href=\"http:\/\/youtube.com\/wpbeginner?sub_confirmation=1\" title=\"WPBeginner on YouTube\" target=\"_blank\" rel=\"nofollow\">YouTube Channel<\/a> for WordPress video tutorials. You can also find us on <a href=\"http:\/\/twitter.com\/wpbeginner\" title=\"WPBeginner on Twitter\" target=\"_blank\" rel=\"nofollow\">Twitter<\/a> and <a href=\"https:\/\/www.facebook.com\/wpbeginner\" title=\"WPBeginner on Facebook\" target=\"_blank\" rel=\"nofollow\">Facebook<\/a>.<\/p>","created_at":"2018-11-07T16:23:59+0000","updated_at":"2018-11-07T16:23:59+0000","published_at":"2017-05-17T06:15:49+0000","published_by":["Editorial Staff"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"www.wpbeginner.com","preview_picture":"https:\/\/cdn2.wpbeginner.com\/wp-content\/uploads\/2017\/04\/removewpsidebar.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":180,"title":"Evaluating Elisp in Emacs - Mastering Emacs","url":"https:\/\/masteringemacs.org\/article\/evaluating-elisp-emacs","content":"<p>By <em>Mickey Petersen<\/em><\/p><img src=\"https:\/\/masteringemacs.org\/static\/img\/fleuron2.gif\" class=\"img-responsive center-block\" alt=\"image\" \/><p class=\"text-justify drop-cap\">There are several ways of evaluating elisp code in Emacs, and picking the right approach will help you get your job done faster and more efficiently. If you\u2019re new to elisp you will quickly realize that Emacs has many shortcuts and features that makes writing, inspecting or debugging elisp code very easy.<\/p><h2 id=\"evaluating-chunks-of-code\">Evaluating Chunks of Code<\/h2><p class=\"text-justify\">If you have a body of code that you want to evaluate, this approach is by far the simplest. To quickly evaluate any amount of code \u2013 from one to many <em>forms<\/em>, where a <em>form<\/em> is essentially something that Emacs can evaluate \u2013 simply yank the code into an empty buffer and run <code>M-x eval-buffer<\/code>.<\/p><p class=\"text-justify\">If you want something less hamfisted you can tell Emacs to only evaluate the <em>region<\/em> by marking what you want to run and executing <code>M-x eval-region<\/code>.<\/p><h3 id=\"concerning-regions\">Concerning Regions\u2026<\/h3><p class=\"text-justify\">It goes without saying that I assume <code>transient-mark-mode<\/code> is active, which is it in newer Emacsen by default; but chances are it is enabled in your <code>.emacs<\/code> anyway, as <em>transient<\/em> makes the region visible on your screen, and not surprisingly that is what most people want.<\/p><p class=\"text-justify\">If you are a disbeliever and prefer a simpler life without <em>transient<\/em> then you\u2019re familiar with narrowing and exchanging the point and the mark to get what you want anyway. I should mention, by the way, that the vanguard of Emacs conversatism, Richard Stallman himself, <em>does<\/em> use <code>transient-mark-mode<\/code>.<\/p><h2 id=\"evaluating-expressions\">Evaluating Expressions<\/h2><h3 id=\"by-s-expression\">By S-Expression<\/h3><p class=\"text-justify\">This is where Emacs starts to shine, for it has special tools available to it to evaluate single <em>s-expressions<\/em>. The facilities that are in place in Emacs to do this depend on the major mode you are using (and, rarely, your mode\u2019s <em>syntax table<\/em>) but most modes still expose <code>C-x C-e<\/code>, which is bound to <code>eval-last-sexp<\/code>.<\/p><p class=\"text-justify\">The <code>eval-last-sexp<\/code> command is very interesting, because it is one of few commands that are globally available to most modes, and it only works if you have the point at the <strong>end<\/strong> of an s-expression (<em>\u201csexp\u201d<\/em>), which is anything enclosed by a pair of parentheses, like this: <code>(message \"Hello, World!\")|<\/code> where <code>|<\/code> is your point.<\/p><p class=\"text-justify\">There are a couple of limitations that make <code>C-x C-e<\/code> a poor tool to use if you are writing and testing elisp code, because it cannot update variables declared with <code>defvar<\/code> or <code>defcustom<\/code>, and it goes with saying that having to move point to the end of the expression every time you want to eval the sexp is annoying as well.<\/p><h3 id=\"by-form\">By Form<\/h3><p class=\"text-justify\">Arguably the best way to evaluate a form is <code>eval-defun<\/code>, bound to <code>C-M-x<\/code>. This command does not suffer from the problems I mentioned before, as it is designed for a typical hacker\u2019s write-eval-test cycle. You can type <code>C-M-x<\/code> anywhere in a form, and it will evaluate the outer-most form, ensuring that you won\u2019t accidentally evaluate only a part of a larger form.<\/p><p class=\"text-justify\">What makes it even better is its built-in support for <em>edebug<\/em>, Emacs\u2019 elisp debugger. Using the universal argument <code>C-u<\/code> Emacs will evaluate the form as normal, but enable debug instrumentation. The next time it is run it kicks off the debugger and lets you step through the code, with every evaluation showing its value (if any) in the echo area.<\/p><p class=\"text-justify\">You can also use this functionality for exploratory purposes by enabling debugging on Emacs\u2019 own routines.<\/p><p class=\"text-justify\">This is by far the best way of quickly re-evaluating code you\u2019ve updated, and it should be muscle memory for any serious elisp hacker.<\/p><h3 id=\"by-expression\">By Expression<\/h3><p class=\"text-justify\">Another quick way of evaluating an expression is to use <code>eval-expression<\/code>, bound to <code>M-:<\/code>. Evaluated expressions are printed to the echo area and the value prepended to the list variable <code>values<\/code>. It is important to note that as it takes <em>expressions<\/em> and not <em>sexps<\/em> or <em>forms<\/em> you can also print the value of any variable in Emacs \u2013 just be sure <strong>not<\/strong> to wrap the variable in parenthesis.<\/p><p class=\"text-justify\">This command is particularly handy if you quickly want to change a setting or invoke an elisp function directly, as it evaluates code in the context of the buffer it was invoked from. If you use the universal argument <code>C-u<\/code>, the output of the expression will be inserted into the buffer. In a similar vein, you can tell Emacs to redo the previous command it did and let you edit the command beforehand by invoking <code>C-x M-:<\/code>. Another useful way of exploring Emacs.<\/p><p class=\"text-justify\">One thing you cannot do is invoke multiple expressions; to get around this, you can wrap the expressions in the <code>progn<\/code> form, like so: <code>(progn (foo) (bar) (baz))<\/code>.<\/p><h2 id=\"interactive-evaluation\">Interactive Evaluation<\/h2><h3 id=\"the-scratch-buffer\">The Scratch Buffer<\/h3><p class=\"text-justify\">The Emacs <code>*scratch*<\/code> buffer comes with a few useful additions over the regular Emacs-Lisp mode that makes it useful for throw-away scripts and quick code evaluation. If you put point at the end of an sexp and type <code>C-j<\/code> Emacs will evaluate the expression and print the result straight into the buffer. <em>Note: If you use paredit, be aware that it rebinds the key!<\/em><\/p><p class=\"text-justify\">The scratch buffer is a poor substitute for a proper REPL, but thankfully Emacs does come with one of those too.<\/p><h3 id=\"the-interactive-emacs-lisp-mode\">The Interactive Emacs-Lisp Mode<\/h3><p class=\"text-justify\">This is a fantastic, and hidden, gem in Emacs: a proper REPL. Type <code>M-x ielm<\/code> to run the elisp REPL. It comes with all the useful features you would expect from an interactive shell, like <code>*<\/code>, <code>**<\/code> and <code>***<\/code> to get the last three outputs from the shell. Multi-line commands are fully supported as well, of course.<\/p><p class=\"text-justify\">Because <code>ielm<\/code> inherits from comint you get all the advantages that provides, like a command history and so on. You also get limited completion support with <code>TAB<\/code>, which is very nice.<\/p><p class=\"text-justify\">Another useful feature is the concept of a <em>working buffer<\/em> - a buffer through which your changes are evaluated. If you type <code>C-c C-b<\/code> you can change ielm\u2019s working buffer to one of your choosing and then all the code you evaluate thereafter will be treated as if you executed it in the context of that buffer. This functionality comes in handy if you are dealing with buffer-local variables or changes that\u2019re specific to one buffer only. Very, very powerful.<\/p><h4 id=\"enabling-auto-complete-in-ielm\">Enabling Auto Complete in IELM<\/h4><p class=\"text-justify\">By default ielm is not supported in <a href=\"http:\/\/cx4a.org\/software\/auto-complete\/\" class=\"article-link\">auto complete<\/a>, but that is easily rectified. Add this to your .emacs:<\/p><pre class=\"sourceCode commonlisp\">(defun ielm-auto-complete ()\n  \"Enables `auto-complete' support in \\\\[ielm].\"\n  (setq ac-sources '(ac-source-functions\n                     ac-source-variables\n                     ac-source-features\n                     ac-source-symbols\n                     ac-source-words-in-same-mode-buffers))\n  (add-to-list 'ac-modes 'inferior-emacs-lisp-mode)\n  (auto-complete-mode 1))\n(add-hook 'ielm-mode-hook 'ielm-auto-complete)<\/pre><h3 id=\"honorable-mention-eshell\">Honorable Mention: Eshell<\/h3><p class=\"text-justify\">You can evaluate elisp in Eshell directly from the prompt, with some limitations. I have covered this in greater detail in my article on Mastering Eshell.<\/p><h2>Conclusion<\/h2><p class=\"text-justify\">That just about covers evaluating code in Emacs, from one-liners to entire buffers of code. If you\u2019re hacking elisp you want <code>C-M-x<\/code> for your everyday editing and quick debugging; if you\u2019re inspecting state or want to change something in Emacs real quick, use <code>M-:<\/code>; if you want to evaluate a buffer or region, use <code>eval-buffer<\/code> or <code>eval-region<\/code>; if you\u2019re exploring elisp and want an interactive environment, use <code>ielm<\/code>. The rest serve more specialized roles, but you cannot go wrong if you stick to these recommendations.<\/p>","created_at":"2018-11-07T16:23:58+0000","updated_at":"2018-11-07T16:23:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"masteringemacs.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":181,"title":"A modern open-source Wiki software","url":"https:\/\/wiki.js.org\/get-started.html","content":"<header class=\"short\" id=\"root\">\n\t<\/header><div class=\"get-started\"><div class=\"container\"><div class=\"columns\"><aside class=\"column is-narrow\"><i>1<\/i>\n\t\t\t\t<\/aside><div class=\"column\"><h3>Install Dependencies<\/h3><p>\n\t\t\t\t\t\tBefore you can install Wiki.js, you need to make sure your server meets all the requirements.<\/p><p><i class=\"icon-document-text\"> <a href=\"https:\/\/docs.requarks.io\/wiki\/install\/prerequisites\">Prerequisites Guide<\/a>\n\t\t\t\t\t<\/i><\/p><\/div><\/div><\/div><\/div>\n\t<div class=\"get-started\"><div class=\"container\"><div class=\"columns\"><aside class=\"column is-narrow\"><i>2<\/i>\n\t\t\t\t<\/aside><div class=\"column\"><h3>Install Wiki.js<\/h3><p>\n\t\t\t\t\t\tYou are now ready to install the actual software.<\/p><p><i class=\"icon-document-text\"> <a href=\"https:\/\/docs.requarks.io\/wiki\/install\/installation\">Installation Guide<\/a><br \/><i class=\"icon-document-text\"> <a href=\"https:\/\/docs.requarks.io\/wiki\/install\/installation\/authentication\">Authentication Guide<\/a><br \/><i class=\"icon-document-text\"> <a href=\"https:\/\/docs.requarks.io\/wiki\/install\/installation\/database\">Database Guide<\/a><br \/><i class=\"icon-document-text\"> <a href=\"https:\/\/docs.requarks.io\/wiki\/install\/installation\/git\">Git Repository Guide<\/a><br \/><\/i><\/i><\/i><\/i><\/p><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:23:58+0000","updated_at":"2018-11-07T16:23:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"wiki.js.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":178,"title":"Datomic: Look at all the things I'm not doing!","url":"https:\/\/augustl.com\/blog\/2018\/datomic_look_at_all_the_things_i_am_not_doing\/","content":"<p>This post is part of a series: <a href=\"https:\/\/augustl.com\/series\/datomic\">Datomic<\/a><\/p><p>This is the video that got me into programming:<\/p><small><em>(This embed should take you directly to 01:52, where it's all at)<\/em><\/small><iframe width=\"560\" height=\"315\" src=\"https:\/\/www.youtube.com\/embed\/Gzj723LkRJY?start=112\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\"> <\/iframe><p>It's DHH's epic first Ruby on Rails demo from 2005. His unique selling point (or one of them, at least) was \"look at all the things I'm not doing\"<\/p><p>At the time, this was ground breaking in its implementation and accessibility. \"Convention over configuration\" became a thing.<\/p><p>However, a portion of the developer community now seems to agree that hiding all the complexities of large CRUD apps has its trade offs, especially when it comes to large and long running projects. You end up in a situation where you have to be as knowledgeable about the conventions as if you wrote the code yourself.<\/p><p>Whoops!<\/p><p>Any sufficiently advanced technology is indistinguishable from magic. But is magic always bad? Do we like magic, as long as it hits the sweet spot where it is able to remain hidden from us?<\/p><p>Why do Clojure devs despise the magic in Ruby on Rails, but love the magic of auto-promoting persistent data structures?<\/p><p>Let's explore all the things you're <em>not<\/em> doing when you're using Datomic.<\/p><h2>Look at all the data modeling I'm not doing<\/h2><p>Datomic has made a decision for you. When inserting data, everything is completely and utterly flat, with no structure.<\/p><pre class=\"highlight\">[:db\/add -1 :person\/name \"August Lilleaas\"]\n[:db\/add -1 :person\/age 31]<\/pre><p>This is <em>it<\/em>. All you have for writing data are these facts. A fact consists of:<\/p><pre class=\"highlight\">[:db\/add \n  -1                 ;; Entity (the id of the thing we're talking about)\n  :person\/name       ;; Attribute (defines the type and some other properties)\n  \"August Lilleaas\"  ;; Value (self evident)\n  ]\n  \n  ;; No table, no nesting, no depth, no ...<\/pre><p>Any structure comes later, from queries and reads.<\/p><p>Datomic isn't <em>free<\/em> of data modeling. You have to choose if an attribute has a cardinality of <em>one<\/em> or <em>many<\/em>. You need to define the type of your attributes. But the API you get for writing data is extremely limited. There's no structured schema.<\/p><p>So, essentially: <strong>Look at all the up front decisions I'm not making about data modeling<\/strong>.<\/p><p>This has one <em>enormous<\/em> benefit. You create structure when you read, not when you write. Therefore, you get to come up with new ways of thinking about your data <em>arbitrarily, after the fact<\/em>. This is just a logical conclusion of having all the decisions about structure being put in the query engine, and not the write API. Contrary to traditional databases, where so much structure is defined up-front when you insert the data.<\/p><h2>Look at all the time modeling I'm not doing<\/h2><p>So you write these facts into the database, and there's almost no structure when you write.<\/p><p>There are 50 ways to get time modeling wrong in a traditional database, and a few ways to get it right. And even when you got it right, maybe your specific implementation of time management didn't cut it so you need to update it. And maybe you didn't add time management to your <em>entire<\/em> domain, just the crucial parts.<\/p><p>Given Datomic's extremely simple write API, it is able to implement time management for you, without <em>any<\/em> additional cost of data modeling or query hatching on your part.<\/p><p>So, essentially: <strong>Look at all the up front decisions I'm not making about time modeling<\/strong>.<\/p><p>It's difficult to talk about Datomic without at one point mentioning the time model. The writer is single threaded. You get a completely serialized time-line of your writes, without any special coordination or re-ordering happening. There's no such thing in Datomic as \"read this thing, write that thing, read this other thing\" and have that be inconsistent in any way shape or form. There's no such thing as \"isolation level\" (alternatively, Datomic is isolation level 9000), and transactions can't see what happens in other transactions. This is fundamentally solved by having only one transaction being processed at a time.<\/p><p>When you query Datomic, you query against what Datomic calls a \"database\". This is essentially a view of a point in time of the time-line of writes that have occurred so far. This view is completely immutable, and requires no synchronization with the writer process because Datomic is an append-only database with append-only semantics.<\/p><p>Are you thinking what I'm thinking?<\/p><p>Did you notice the detail there that makes Datomic amazing?<\/p><p>You ask Datomic for a \"database\", and you get an immutable reference to the value of that database at that point in time. You get to hold onto this value forever, no matter how many writes that occur in the future.<\/p><p>Getting it yet?<\/p><p><em>You get to hold onto a view of the database that's immutable, no matter what happens to it in the future.<\/em><\/p><p>This means that it's extremely cheap for Datomic to let you get a view of a <em>previous state of the database<\/em>. In fact, internally, the API is exactly the same. You get a reference to the database data structure, with a T value (a reference or id to a point in time in the transaction time-line), and all your queries will ensure that you get the data as of that T value.<\/p><p>So it's really cheap to just set the T value to an earlier point in time, say two weeks ago, or whatever, and get a <em>completely immutable and consistent view of the entire database as of two weeks ago<\/em>.<\/p><p>Woha.<\/p><h2>Look at all the consistency management I'm not doing<\/h2><p>In that vein, the consistency model in Datomic is also extremely simple. As mentioned above, the writes happen in a single thread, writing one transaction at a time. This is isolation level 9000. There's no extra management to be done on your part to ensure that you use the correct isolation level or that you do things in the correct order to get the expected result or any of that. Datomic is consistent, period. There's no way to screw that up by misunderstandings or bugs made by the developer that uses Datomic in their app.<\/p><p>The only trade-off here is that you need to resort to stored procedures, VoltDB style, when you want to do anything other than setting a value, deleting (retracting) a value, and doing a few built-in operations like compare-and-swap and identity based upsert. If you want to do things like \"increment by 10\" or \"take N from account X and put it into account Y\", you need to express this as a transaction function (Datomic's name for stored procedures), written in Clojure or Java, and ship it off to the transactor (Datomic's name for the database server).<\/p><h2>Look at all the time management I'm not doing<\/h2><p>This is a bit of a minor point, but I still like it. It is an example of the many things that are suddenly fundamentally solved when you have a database that manages time for you.<\/p><p>Exhibit A: Pagination. Pagination has the issue of being a non-atomic operation. You load the first 50 items. Then you go forward, and load the next 50 items. But in between those two fetches of items, the database might have changed. With Datomic, you can have the T value be a factor in your pagination. You can use this to get a consistent view of the database, so that page two is actually page two. You can also run queries to see if your data set has changed since the provided T value, and inform the user. \"Click here to update your view of the data to the most recent changes\".<\/p><p>Exhibit B: Return values. Let's say you have a method that takes a bunch of arguments and inserts a new thing into your database. What should it return? Maybe just the ID, so we don't force all users of the API to wait for us to reconstruct a view of the thing that was just inserted. With Datomic, however, this is a non-issue. You simply return the ID and a reference to the T-value as of after your insert (which is something Datomic provides automatically). Then it is up to the consumer to do what it wants with the T-value. And this is another example of the extreme agility you get with Datomic. Because time is managed for you, you get to make <em>a whole lot of decisions later on, since you don't have to worry about the rug getting pulled from underneath your feet<\/em>.<\/p><h2>Look at all the event sourcing I'm not doing<\/h2><p>This is a big one.<\/p><p>If you're familiar with event sourcing, you might wonder what the big deal is. \"One single threaded time-line of writes\" is exactly what event sourcing is? <\/p><p>What does Datomic have that event sourcing doesn't?<\/p><p>Answer: An absence of hand written views.<\/p><p>With event sourcing, you have raw data in your stream of events. This data is useless until you create a view of it. A view is typically a left fold reducing function that gets the previous state and the next event in the list passed to it, and returns a new and updated state.<\/p><p><img width=\"600\" src=\"https:\/\/augustl.com\/static\/posts\/datomic_look_at_all_the_things_i_am_not_doing\/event_sourcing.jpg\" alt=\"image\" \/><\/p><p>This means that you have to decide which views you want to have. And that if there's a bug in your view code, you have to traverse the entire time-line again to build up the state.<\/p><p>With Datomic, you <em>have<\/em> the stream of events (facts\/transactions), but there is one crucial but fundamentally important difference:<\/p><p><em>You don't have to decide what your views are up-front<\/em><\/p><p>So, essentially: <strong>Look at all the up front decisions I'm not making about which views to have<\/strong>.<\/p><p><img width=\"600\" src=\"https:\/\/augustl.com\/static\/posts\/datomic_look_at_all_the_things_i_am_not_doing\/event_sourcing_datomic.jpg\" alt=\"image\" \/><\/p><p>Again, Datomic provides us with extreme agility. Because the data we write is completely flat, but is <em>not<\/em> arbitrary data but a well known structure (facts), and Datomic have generic indices to query these facts, and because Datomic keeps these indices up to date when we add data to our system, you get the benefits of event sourcing (time travel, a consistent time-line, all your historical data easily traversable), without having to worry about creating views for your data. The views are already there, in form of Datomic's query engine, its built-in time travel features, and efficient indices of historical data.<\/p><h2>Look at all the file writing I'm not doing<\/h2><p>(Implementation detail)<\/p><p>Internally, Datomic doesn't write to the file system directly. It writes to existing storage engines and databases. You can set up Datomic to write to any JDBC database (mysql, postgres, h2, oracle, db\/2, ...), dynamodb, riak + zookeper, and a few other combinations.<\/p><p>This means that Datomic doesn't have to implement its own code for properly writing to disks, which is a huge pain to get right. And it means that if you already use postgres, you don't have to worry about backup routines and operations and what not, other than running the Datomic transactor itself. You can just use your existing routines for postgres operations.<\/p><h2>Look at all the database concept reinvention I'm not doing<\/h2><p>(Implementation detail)<\/p><p>Datomic doesn't actually invent a whole lot from scratch.<\/p><p>I've already mentioned VoltDB. VoltDB also writes in a single thread, after research revealed that <a href=\"http:\/\/nms.csail.mit.edu\/~stavros\/pubs\/OLTP_sigmod08.pdf\">multi-threaded RDBMS databases spend about 75% of their time doing coordination and only 25% of their time actually doing reads and writes<\/a>. There's this instinct that single threaded processes can't possibly be efficient. But you'd be surprised how much efficiency you can get out of a thread that spends close to 100% of its time doing useful work without much time spent on coordination between multiple threads.<\/p><p>There's also a concept borrowed from Google's Big Table. The indices in Datomic aren't updated live for every write, that would be way too much overhead. So a query in Datomic runs against the latest index + a list of transactions that are written to a transaction log but not yet merged into the index data structures. Then, Datomic will periodically take the unmerged transaction log and merge it into the main index. Big Table does the same, after Google learned that live-merging an index of <em>the entire internet<\/em> was maybe not going to work.<\/p><h2>Look at all the .... ok, I'm done<\/h2><p>In my experience, Datomic is indistinguishable from magic, in a good way. It's one level of abstraction above traditional databases, but manages to hide those abstractions since you at no point worry about postgres or zookeeper internals when you make use of Datomic on top of the existing storage engines. You get a whole lot of stuff for free, without Datomic getting in your way.<\/p><p>And to me, the most important point is all of the things you're <em>not<\/em> doing and all of the decisions you are <em>not<\/em> making up front, but rather in an agile fashion as your application grows.<\/p><h2>Update: even more things I'm not doing<\/h2><p>I wrote a follow-up, with even more things I'm not doing:<\/p><p><a href=\"http:\/\/augustl.com\/blog\/2018\/datomic_look_at_all_the_things_i_am_not_doing_cont\/\">http:\/\/augustl.com\/blog\/2018\/datomic_look_at_all_the_things_i_am_not_doing_cont\/<\/a><\/p><hr class=\"post-sep\" \/><p>Questions or comments?<\/p><p>Feel free to contact me on Twitter, <a href=\"http:\/\/twitter.com\/augustl\">@augustl<\/a>, or e-mail me at <a href=\"mailto:august@augustl.com\">august@augustl.com<\/a>.<\/p>","created_at":"2018-11-07T16:23:57+0000","updated_at":"2018-11-07T16:23:57+0000","annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"augustl.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":179,"title":"Woody Woodclick","url":"https:\/\/www.woodywoodclick.com","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:23:57+0000","updated_at":"2018-11-07T16:23:57+0000","annotations":[],"reading_time":0,"domain_name":"www.woodywoodclick.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":176,"title":"onurakpolat\/awesome-analytics","url":"https:\/\/github.com\/onurakpolat\/awesome-analytics\/blob\/master\/README.md","content":"<p>A curated list of awesome analytics platforms, resources and other awesomeness. Inspired by <a href=\"https:\/\/github.com\/onurakpolat\/awesome-bigdata\">awesome-bigdata<\/a>.<\/p>\n<p>Your feedback and contributions are always welcome!<\/p>\n<ul><li><a href=\"#awesome-analytics\">Awesome Analytics<\/a>\n<ul><li>[General analytics](#General analytics)<\/li>\n<li><a href=\"#Heatmap-analytics\">Heatmap analytics<\/a><\/li>\n<li><a href=\"#Analytics-layers\">Analytics layers<\/a><\/li>\n<li><a href=\"#Mobile-analytics\">Mobile analytics<\/a><\/li>\n<li><a href=\"#App-store-analytics\">App store analytics<\/a><\/li>\n<li><a href=\"#Attribution-tracking\">Attribution tracking<\/a><\/li>\n<li><a href=\"#Social-media-analytics\">Social media analytics<\/a><\/li>\n<li><a href=\"#Analytics-dashboards\">Analytics dashboards<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#other-awesome-lists\">Other Awesome Lists<\/a><\/li>\n<\/ul><h2><a id=\"user-content-general-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-analytics\"><\/a>General analytics<\/h2>\n<ul><li><a href=\"https:\/\/www.hotjar.com\" rel=\"nofollow\">Hotjar<\/a> - new and easy way to truly understand your web and mobile site visitors.<\/li>\n<li><a href=\"http:\/\/piwik.org\/\" rel=\"nofollow\">Piwik<\/a> - real-time open source app analytics tool.<\/li>\n<li><a href=\"https:\/\/calq.io\/\" rel=\"nofollow\">Calq<\/a> - simple, beautiful, and real-time cross-platform app analytics tool.<\/li>\n<li><a href=\"https:\/\/heapanalytics.com\/\" rel=\"nofollow\">Heap<\/a> - tracks your app users, clicks, form submissions, and anything else.<\/li>\n<li><a href=\"http:\/\/www.opentracker.net\/\" rel=\"nofollow\">Opentracker<\/a> - real time reporting, geo-location user tracking.<\/li>\n<li><a href=\"http:\/\/foxmetrics.com\/%20\" rel=\"nofollow\">Foxmetrics<\/a> - analytics to track your user\u2019s actions and activities.<\/li>\n<li><a href=\"http:\/\/www.google.com\/analytics\/\" rel=\"nofollow\">Google Analytics<\/a> - de facto standard for analytics in the web analytics space.<\/li>\n<li><a href=\"http:\/\/www.clicktale.com\/\" rel=\"nofollow\">Clicktale<\/a> - record and watch exactly how a visitor used your website.<\/li>\n<li><a href=\"https:\/\/www.gosquared.com\/\" rel=\"nofollow\">Go Squared<\/a> - analytics with visitor tagging to help you dig deeper into one user\u2019s visit.<\/li>\n<li><a href=\"http:\/\/clicky.com\/\" rel=\"nofollow\">Clicky<\/a> - track visits and conversions, you can also track your video and audio analytics.<\/li>\n<li><a href=\"https:\/\/www.woopra.com\/\" rel=\"nofollow\">Woopra<\/a> - track where your users are coming from.<\/li>\n<li><a href=\"http:\/\/haveamint.com\/\" rel=\"nofollow\">Mint<\/a> - elf-hosted analytics solution.<\/li>\n<li><a href=\"http:\/\/www.goingup.com\/\" rel=\"nofollow\">Going Up<\/a> - manage SEO analytics and web app analytics with one tool.<\/li>\n<li><a href=\"https:\/\/chartbeat.com\/\" rel=\"nofollow\">Chartbeat<\/a> - beautiful, real-time app analytics tool for web apps.<\/li>\n<li><a href=\"http:\/\/get.gaug.es\/\" rel=\"nofollow\">Gauges<\/a> - real-time web analytics tool.<\/li>\n<li><a href=\"http:\/\/www.indicative.com\/\" rel=\"nofollow\">Indicative<\/a> - Web &amp; mobile  analytics tool.<\/li>\n<li><a href=\"http:\/\/www.openwebanalytics.com\/\" rel=\"nofollow\">Open Web Analytics<\/a> - open source web app analytics tool.<\/li>\n<li><a href=\"http:\/\/statcounter.com\/\" rel=\"nofollow\">Statcountr<\/a> - one of the ORIGINAL web analytics tools available.<\/li>\n<li><a href=\"http:\/\/snowplowanalytics.com\/\" rel=\"nofollow\">Snow Plow<\/a> - analytics tool for web apps with a lot of data.<\/li>\n<li><a href=\"http:\/\/www.adobe.com\/solutions\/digital-analytics\/marketing-reports-analytics.html\" rel=\"nofollow\">Adobe Digital Analytics<\/a> - standard analytics tools plus some that large organizations can use.<\/li>\n<li><a href=\"https:\/\/paditrack.com\/\" rel=\"nofollow\">Paditrack.com\/<\/a> - analyze any data from Google Analytics, including the data you had before.<\/li>\n<li><a href=\"http:\/\/www.hitslink.com\/\" rel=\"nofollow\">Hitslink.com\/<\/a> - real-time analytics, social media traffic reporting, and real-time dynamic segmentation.<\/li>\n<li><a href=\"http:\/\/parse.ly%20\" rel=\"nofollow\">parse.ly <\/a> - real-time web analytics tool with a focus on tracking content.<\/li>\n<li><a href=\"http:\/\/loggr.net\/\" rel=\"nofollow\">Loggr<\/a> -  track your user events and monitor your web app.<\/li>\n<li><a href=\"https:\/\/www.kissmetrics.com\/\" rel=\"nofollow\">Kissmetrics<\/a> - real-time standard cohort analysis tool.<\/li>\n<li><a href=\"http:\/\/sitemeter.com\/\" rel=\"nofollow\">Sitemeter<\/a> - old analytics tool.<\/li>\n<li><a href=\"http:\/\/www.crawltrack.net\/\" rel=\"nofollow\">Crawl Track<\/a> - another old analytics tool.<\/li>\n<li><a href=\"http:\/\/www.sitespect.com\/\" rel=\"nofollow\">Sitespect<\/a> - full-suite web app analytics tool including A\/B testing.<\/li>\n<\/ul><h2><a id=\"user-content-heatmap-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#heatmap-analytics\"><\/a>Heatmap analytics<\/h2>\n<ul><li><a href=\"https:\/\/www.reinvigorate.net\/\" rel=\"nofollow\">Reinviorate<\/a> - real-time analytics tool with heatmaps.<\/li>\n<li><a href=\"http:\/\/www.crazyegg.com\/\" rel=\"nofollow\">Crazyegg<\/a> - a heatmaps only analtyics tool.<\/li>\n<li><a href=\"https:\/\/www.inspectlet.com\/\" rel=\"nofollow\">Inspeclet<\/a> - another web app heatmaps tool.<\/li>\n<li><a href=\"http:\/\/mouseflow.com\/%20\" rel=\"nofollow\">Mouseflow<\/a> - live analytics and heatmaps.<\/li>\n<li><a href=\"http:\/\/www.sessioncam.com\/\" rel=\"nofollow\">Session Cam<\/a> - heatmaps analytics tool.<\/li>\n<li><a href=\"https:\/\/heatma.ps\/\" rel=\"nofollow\">Heatmaps for your iOS apps<\/a> - mobile app heatmaps for mobile apps.<\/li>\n<li><a href=\"http:\/\/heatdata.com\/\" rel=\"nofollow\">Heatmaps for Mobile Websites<\/a> - heatmaps tool for mobile apps and mobile websites.<\/li>\n<\/ul><h2><a id=\"user-content-analytics-layers\" class=\"anchor\" aria-hidden=\"true\" href=\"#analytics-layers\"><\/a>Analytics layers<\/h2>\n<ul><li><a href=\"http:\/\/adjust.com\/\" rel=\"nofollow\">Keen.io<\/a> - custom-analytics API.<\/li>\n<li><a href=\"http:\/\/www.popcornmetrics.com\/\" rel=\"nofollow\">Popcorn metrics<\/a> - visual editor to capture events and send to toher platforns.<\/li>\n<li><a href=\"http:\/\/Segment.io\" rel=\"nofollow\">Segment.io<\/a> - helps you integrate multiple app analytics tool with one piece of code.<\/li>\n<\/ul><h2><a id=\"user-content-mobile-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#mobile-analytics\"><\/a>Mobile analytics<\/h2>\n<p>The tools listed here are not necessarily mobile analytics tools only. However they show a strong mobile focus.<\/p>\n<ul><li><a href=\"http:\/\/www.upsight.com\/\" rel=\"nofollow\">Upsight<\/a> - mobile app analytics tool for developers.<\/li>\n<li><a href=\"http:\/\/www.appsflyer.com\/\" rel=\"nofollow\">Appsflyer<\/a> - all-in-one marketing tool with analytics.<\/li>\n<li><a href=\"http:\/\/aws.amazon.com\/mobileanalytics\/\" rel=\"nofollow\">Amazon Mobile Analytics<\/a> - Amazons multi-platform, basic mobile analytics tool.<\/li>\n<li><a href=\"https:\/\/tapstream.com\/\" rel=\"nofollow\">Tapstream<\/a> - user lifecycle analytics.<\/li>\n<li><a href=\"https:\/\/honeytracks.com\/\" rel=\"nofollow\">Honeytracks<\/a> - mobile app analytics for games.<\/li>\n<li><a href=\"https:\/\/apsalar.com\/\" rel=\"nofollow\">Apsalar<\/a> - analytics tool for larger app shops.<\/li>\n<li><a href=\"http:\/\/www.roambi.com\/\" rel=\"nofollow\">Roambi<\/a> - 3-in-1 analytics tool that helps you track analytics, handle mobile app business intelligence, and app reporting.<\/li>\n<li><a href=\"http:\/\/www.appcelerator.com\/platform\/appcelerator-analytics\/\" rel=\"nofollow\">Appcelerator<\/a> - entire mobile app marketing suite.<\/li>\n<li><a href=\"http:\/\/www.flurry.com\/\" rel=\"nofollow\">Flurry<\/a> - pretty much the \u201cindustry standard\u201d for mobile app analytics.<\/li>\n<li><a href=\"http:\/\/count.ly\/\" rel=\"nofollow\">Count.ly<\/a> - open-source mobile app analytics tool.<\/li>\n<li><a href=\"http:\/\/playtomic.org\/\" rel=\"nofollow\">Playtomatic<\/a> - mobile app open source analytics tool for games.<\/li>\n<li><a href=\"http:\/\/www.capptain.com\/\" rel=\"nofollow\">Capptain<\/a> - real-time analytics tool with segmentation and push.<\/li>\n<li><a href=\"https:\/\/amplitude.com\/\" rel=\"nofollow\">Amplitude<\/a> - real-time mobile analytics with all data provided in redshift.<\/li>\n<li><a href=\"http:\/\/www.appsee.com\/\" rel=\"nofollow\">Appsee<\/a> - mobile app analytics platform automatically tracks all users' interactions in your app<\/li>\n<li><a href=\"https:\/\/mixpanel.com\/\" rel=\"nofollow\">Mixpanel<\/a> - fully featured mobile analytics platform with segmentation and push.<\/li>\n<li><a href=\"http:\/\/www.localytics.com\/\" rel=\"nofollow\">Localytics<\/a> - fast and beautiful real-time mobile analytics platform with in-app and push.<\/li>\n<li><a href=\"http:\/\/www.gameanalytics.com\/\" rel=\"nofollow\">GameAnalytics<\/a> - leading game analytics platform.<\/li>\n<\/ul><h2><a id=\"user-content-app-store-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#app-store-analytics\"><\/a>App store analytics<\/h2>\n<ul><li><a href=\"http:\/\/appfigures.com\/\" rel=\"nofollow\">Appfigures<\/a> - app store analytics to track sales, reviews and rankings with an API.<\/li>\n<li><a href=\"http:\/\/www.appannie.com\/\" rel=\"nofollow\">Appannie<\/a> - track your app data from iTunes, Google Play &amp; Amazon.<\/li>\n<li><a href=\"http:\/\/www.distimo.com\/\" rel=\"nofollow\">Distimo<\/a> - free app store analytics (acquired by <a href=\"http:\/\/www.appannie.com\/\" rel=\"nofollow\">Appannie<\/a>).<\/li>\n<li><a href=\"https:\/\/prioridata.com\/\" rel=\"nofollow\">Priori Data<\/a> - track and benchmark the performance of apps on Apple- and Play store.<\/li>\n<li><a href=\"http:\/\/www.askingpoint.com\/mobile-app-rating-widget\" rel=\"nofollow\">Asking Point<\/a> - track your mobile app user ratings.<\/li>\n<li><a href=\"http:\/\/www.apptrace.com\/\" rel=\"nofollow\">Apptrace<\/a> - fast and free app store analytics platform.<\/li>\n<\/ul><h2><a id=\"user-content-attribution-tracking\" class=\"anchor\" aria-hidden=\"true\" href=\"#attribution-tracking\"><\/a>Attribution tracking<\/h2>\n<ul><li><a href=\"http:\/\/adjust.com\/\" rel=\"nofollow\">Adjust<\/a> - open-source SDK with sophisticated analysis and campaign tracking.<\/li>\n<li><a href=\"http:\/\/clickmeter.com\/\" rel=\"nofollow\">Clickmeter<\/a> - analytics tool that helps you track marketing campaigns.<\/li>\n<li><a href=\"http:\/\/www.mobileapptracking.com\/\" rel=\"nofollow\">HasOffers Mobile app tracking<\/a> - attribution analytics platform.<\/li>\n<\/ul><h2><a id=\"user-content-social-media-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#social-media-analytics\"><\/a>Social media analytics<\/h2>\n<p>Often there is no clear differentiation between social media management and analytics as most the tools provide analytics.<\/p>\n<ul><li><a href=\"https:\/\/klout.com\/\" rel=\"nofollow\">Klout<\/a> - Social media analytics and planning platform.<\/li>\n<li><a href=\"http:\/\/www.falconsocial.com\/\" rel=\"nofollow\">Falconsocial<\/a> - communications platform built on social media with analytics.<\/li>\n<li><a href=\"https:\/\/www.quintly.com\/\" rel=\"nofollow\">Quintly<\/a> - web-based tool to help you track, benchmark and optimize your social media performance.<\/li>\n<li><a href=\"http:\/\/kred.com\/\" rel=\"nofollow\">Kred<\/a> - Klout-like social media analytics platform.<\/li>\n<li><a href=\"https:\/\/bufferapp.com\/\" rel=\"nofollow\">Buffer<\/a> - Social media publishing and analytics platform.<\/li>\n<li><a href=\"http:\/\/topsy.com\/\" rel=\"nofollow\">Topsy<\/a> - Social analytics tool with search.<\/li>\n<li><a href=\"http:\/\/socialblade.com\/\" rel=\"nofollow\">SocialBlade<\/a> - premiere YouTube statistics tracking.<\/li>\n<li><a href=\"https:\/\/klout.com\/\" rel=\"nofollow\">SocialBench<\/a> - Social media analytics platform.<\/li>\n<li><a href=\"https:\/\/hootsuite.com\/\" rel=\"nofollow\">Hootsuite<\/a> - Social media management dashbaord.<\/li>\n<li><a href=\"http:\/\/www.brandwatch.com\/\" rel=\"nofollow\">Brandwatch<\/a> - Social media monitoring and analytics.<\/li>\n<li><a href=\"http:\/\/sproutsocial.com\/\" rel=\"nofollow\">Sproutsocial<\/a> - Social media management and analytics platform.<\/li>\n<\/ul><h2><a id=\"user-content-analytics-dashboards\" class=\"anchor\" aria-hidden=\"true\" href=\"#analytics-dashboards\"><\/a>Analytics dashboards<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/Freeboard\/freeboard\">Freeboard<\/a> - open source real-time dashboard builder for IOT and other web mashups.<\/li>\n<li><a href=\"https:\/\/www.geckoboard.com\/\" rel=\"nofollow\">Geckboard<\/a> - dashboard for key metrics in one place.<\/li>\n<li><a href=\"https:\/\/ducksboard.com\/\" rel=\"nofollow\">Ducksboard<\/a> - visualize and share the data that matters most to you and your team.<\/li>\n<\/ul>","created_at":"2018-11-07T16:23:56+0000","updated_at":"2018-11-07T16:23:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1712926?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":177,"title":"ink - inkle's narrative scripting language","url":"https:\/\/www.inklestudios.com\/ink\/","content":"<h3>Getting Started<\/h3><p>If you're new to <strong>ink<\/strong>, why not download <strong>Inky<\/strong>, our ink editor that allows you to play and test your stories.<\/p><p>The basics tutorial below has been written with a non-technical audience in mind, covering just the basics for people who want to write stories that they can put on the web.<\/p><p>And to see everything that ink has to offer, read the full guide.<\/p><a class=\"ink-section-button\" href=\"https:\/\/www.inklestudios.com\/ink\/web-tutorial\/\">Basics tutorial<\/a><a class=\"ink-section-button\" href=\"https:\/\/github.com\/inkle\/ink\/blob\/master\/Documentation\/WritingWithInk.md\">Full guide<\/a>","created_at":"2018-11-07T16:23:56+0000","updated_at":"2018-11-07T16:23:56+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.inklestudios.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":174,"title":"Unlearning toxic behaviors in a code review culture","url":"https:\/\/medium.freecodecamp.org\/unlearning-toxic-behaviors-in-a-code-review-culture-b7c295452a3c?gi=14f444d5cf9a","content":"<p id=\"f81c\" class=\"graf graf--p graf-after--figure\">Code reviews can be contentious. I had an incredible experience giving my first conference talk titled \u201c<a href=\"https:\/\/alterconf.com\/talks\/unlearning-toxic-behaviors-code-review-culture\" data-href=\"https:\/\/alterconf.com\/talks\/unlearning-toxic-behaviors-code-review-culture\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Unlearning Toxic Behaviors in a Code Review Culture<\/a>\u201dat <a href=\"https:\/\/alterconf.com\/\" data-href=\"https:\/\/alterconf.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">AlterConf<\/a>. I went into the experience preparing for any pushback and criticism that I might receive while encouraging software teams to create supportive code review environments.<\/p><p id=\"b40a\" class=\"graf graf--p graf-after--p\">But I totally underestimated how <strong class=\"markup--strong markup--p-strong\">supportive and receptive <\/strong>the AlterConf community is. I worried about pushback, but that community greeted me with kindness and encouragement instead. \u2764<\/p><figure id=\"9ab9\" class=\"graf graf--figure graf--iframe graf-after--p\"><figcaption class=\"imageCaption\">Twitter was extremely supportive.<\/figcaption><\/figure><p id=\"bb2e\" class=\"graf graf--p graf-after--figure\">I\u2019ve been asked to share my slides (<a href=\"https:\/\/speakerdeck.com\/sandyaaaa\/unlearning-toxic-behaviors-in-a-code-review-culture\" data-href=\"https:\/\/speakerdeck.com\/sandyaaaa\/unlearning-toxic-behaviors-in-a-code-review-culture\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">and I have<\/a>), <strong class=\"markup--strong markup--p-strong\">but I now worry that the slides alone are unhelpful and easily taken out of context<\/strong>, because the deck lacks context and explanations. Since my slides are floating around the internet without these key features, I decided to do a write-up to elaborate upon my talking points. The <a href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" data-href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">video recording<\/a> was released a few months after this article was published.<\/p><p id=\"fa5a\" class=\"graf graf--p graf-after--p\">I\u2019ve listed below <strong class=\"markup--strong markup--p-strong\">some common unhelpful behaviors<\/strong> people display during code reviews, and <strong class=\"markup--strong markup--p-strong\">some recommendations<\/strong> on how software teams can make their teams more supportive by refusing to normalize toxicity. All of the behaviors I describe were either witnessed by me or happened to an industry contact of mine. I\u2019ve been guilty of several of these behaviors in the past, too.<\/p><h3 id=\"3f0a\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #1: passing off opinion as\u00a0fact<\/h3><p id=\"bdb5\" class=\"graf graf--p graf-after--h3\">Don\u2019t make claims unless you can cite documentation, formalized guidelines, and coding examples to back those claims up. People need to know why they are being asked to make a change, and another developer\u2019s personal preference isn\u2019t a good enough argument.<\/p><p id=\"12ee\" class=\"graf graf--p graf-after--p\">Instead of saying:<\/p><blockquote id=\"a6d4\" class=\"graf graf--blockquote graf-after--p\"><div>This component should be stateless.<\/div><\/blockquote><p id=\"3608\" class=\"graf graf--p graf-after--blockquote\">\u2026provide some context behind the recommendation:<\/p><blockquote id=\"cad9\" class=\"graf graf--blockquote graf-after--p\"><div>Since this component doesn\u2019t have any lifecycle methods or state, it could be made a stateless functional component. This will improve performance and readability. *Here* is some documentation.<\/div><\/blockquote><p id=\"8a4b\" class=\"graf graf--p graf-after--blockquote\">Passing off opinion as fact is also common when developers discuss style and syntax preferences. These are really important discussions to have, but never in a code review, because style and syntax have nothing to do with the issue the developer was initially trying to solve.<\/p><p id=\"dbb2\" class=\"graf graf--p graf-after--p\">Have these discussions separately and decide which style guidelines you will be following as a group. Implement a <a href=\"http:\/\/www.jslint.com\/\" data-href=\"http:\/\/www.jslint.com\/\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">linter<\/a> and an <a href=\"https:\/\/atom.io\/packages\/atom-beautify\" data-href=\"https:\/\/atom.io\/packages\/atom-beautify\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">automatic code fixer<\/a>. This way, you can reference your style guidelines, and not your personal opinion, during the code review when you see inconsistencies.<\/p><p id=\"4211\" class=\"graf graf--p graf-after--p\">It is especially important not to pass off your opinion as fact when you have higher rank and authority within your team or company. If you do this, developers will feel they have no choice but to quietly implement your demands.<\/p><h3 id=\"50a0\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #2: overwhelming with an avalanche of\u00a0comments<\/h3><p id=\"b55a\" class=\"graf graf--p graf-after--h3\">When a person makes an error, chances are high that they have made the same error in several places in their changeset. I have noticed that reviewers sometimes point out every single one of an error\u2019s many occurrences instead of leaving one detailed note with links to helpful resources.<\/p><p id=\"611b\" class=\"graf graf--p graf-after--p\">Consolidating comments allows you to convey the same message without overwhelming the review seeker. An avalanche of comments for<em class=\"markup--em markup--p-em\"> one problem, duplicated,<\/em> comes off as nitpicking.<\/p><p id=\"b37f\" class=\"graf graf--p graf-after--p\">Unhelpful and overwhelming:<\/p><figure id=\"6e99\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*IOG-1oW1w6b1BkOZ.\" data-width=\"800\" data-height=\"727\" data-action=\"zoom\" data-action-value=\"0*IOG-1oW1w6b1BkOZ.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*IOG-1oW1w6b1BkOZ.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Leaving multiple comments for \u201cone error, duplicated\u201d (trailing whitespaces)<\/figcaption><\/figure><p id=\"f295\" class=\"graf graf--p graf-after--figure\">More helpful:<\/p><figure id=\"33f1\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*ivhV5hZXN6nSkyED.\" data-width=\"800\" data-height=\"158\" data-action=\"zoom\" data-action-value=\"0*ivhV5hZXN6nSkyED.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*ivhV5hZXN6nSkyED.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Consolidating feedback<\/figcaption><\/figure><p id=\"907f\" class=\"graf graf--p graf-after--figure\">I can understand the argument that pointing out every place an error occurs on a PR can sometimes be useful, since the comment disappears when the developer resolves the issue in subsequent commits. However, after this error has occurred across a changeset, it should be clear that the developer was not aware of a certain guideline and simply needs to be pointed to the correct resources.<\/p><h3 id=\"ccc9\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #3: asking engineers to solve problems they didn\u2019t cause \u201cwhile they are at\u00a0it\u201d<\/h3><p id=\"da78\" class=\"graf graf--p graf-after--h3\">Avoid asking developers to solve issues that aren\u2019t directly related to their changeset or the problem their changeset is trying to solve. Even if a developer is extending or modifying a messy part of the code that is rife with bad practices, don\u2019t ask the developer to fix them in that pull request just because their changeset happens to touch the messy code.<\/p><p id=\"2531\" class=\"graf graf--p graf-after--p\">I am <strong class=\"markup--strong markup--p-strong\">not <\/strong>suggesting that developers should feel no accountability for code just because they didn\u2019t initially introduce it. In fact, it is not collaborative to say something like, \u201cnot my code, not my problem.\u201d I am just suggesting that it would be more appropriate to create a separate ticket and PR to address the messy code. This way, you avoid drastically increasing the scope of someone\u2019s work, but the messy code still gets fixed.<\/p><p id=\"5c12\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">TL;DR: <\/strong>Don\u2019t ask devs to fix problems \u201cwhile they are at it.\u201d If the code solves the issue\/ticket the dev was working on and doesn\u2019t introduce any new issues to the codebase, give the pull request a \u201cthumbs up\u201d and then create a ticket to clean up the bad code.<\/p><h3 id=\"1ee0\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #4: asking judgmental questions<\/h3><p id=\"7338\" class=\"graf graf--p graf-after--h3\">Avoid asking judgmental questions like:<\/p><blockquote id=\"644b\" class=\"graf graf--blockquote graf-after--p\"><div>Why didn\u2019t you just do ___ here?\u201d<\/div><\/blockquote><p id=\"f52c\" class=\"graf graf--p graf-after--blockquote\">Asking such questions implies that a perceived simple solution should have been obvious. It also forces developers to have to defend themselves.<\/p><p id=\"bd12\" class=\"graf graf--p graf-after--p\">Oftentimes, these judgmental questions are just veiled demands. Instead, provide a recommendation (with documentation and citations) and leave out harsh words.<\/p><blockquote id=\"d652\" class=\"graf graf--blockquote graf-after--p\"><div>You can do ___, which has the benefit of ____.<\/div><\/blockquote><h3 id=\"b0af\" class=\"graf graf--h3 graf-after--blockquote\">Unhelpful behavior #5: being sarcastic<\/h3><p id=\"036d\" class=\"graf graf--p graf-after--h3\">There is no appropriate time to be sarcastic when offering someone feedback. Sarcastic comments tend not to provide context or actionable feedback. Instead, describe the issue with details and provide recommendations, but leave the caustic jokes out.<\/p><p id=\"db33\" class=\"graf graf--p graf-after--p\">Unhelpful:<\/p><blockquote id=\"d125\" class=\"graf graf--blockquote graf-after--p\"><div>Did you even test this code before you checked it in?<\/div><\/blockquote><p id=\"f396\" class=\"graf graf--p graf-after--blockquote\">Helpful:<\/p><blockquote id=\"a661\" class=\"graf graf--blockquote graf-after--p\"><div>This breaks when you enter a negative number. Can you please address this case?<\/div><\/blockquote><p id=\"6638\" class=\"graf graf--p graf-after--blockquote\">Here is another example of a <a href=\"https:\/\/techblog.workiva.com\/tech-blog\/we-are-ruthless-code-reviews\" data-href=\"https:\/\/techblog.workiva.com\/tech-blog\/we-are-ruthless-code-reviews\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">code review comment that isn\u2019t funny or helpful<\/a>:<\/p><blockquote id=\"0618\" class=\"graf graf--blockquote graf-after--p\"><div>I don\u2019t mean we\u2019re mean-spirited. I just mean that we are merciless. You\u2019ll notice that I left the comment \u201cBeep!\u201d on the imports of every file you touched. What I meant was, \u201cYour imports violate our standard convention\u200a\u2014\u200awe order them by built-ins, then third party, and then project level,\u201d but that was too much to type on every file.<\/div><\/blockquote><p id=\"db37\" class=\"graf graf--p graf-after--blockquote\">In the above example, \u201cbeep!\u201d is not helpful or descriptive. This practice is simply pedantic humor that does nothing to better the review seeker.<\/p><h3 id=\"e26a\" class=\"graf graf--h3 graf-after--p\">Unhelpful behavior #6: using emojis instead of statements to point out\u00a0issues<\/h3><p id=\"391b\" class=\"graf graf--p graf-after--h3\">Avoid using the thumbs-down or puke emoji to point out issues in code. This is as unhelpful as sarcasm for similar reasons. Emojis are cryptic and easy to misconstrue. Emojis waste peoples\u2019 time as they try to figure out what you mean.<\/p><figure id=\"47fa\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*KcY0E6xRJ9Nt73vr.\" data-width=\"1060\" data-height=\"166\" data-action=\"zoom\" data-action-value=\"0*KcY0E6xRJ9Nt73vr.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*KcY0E6xRJ9Nt73vr.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Don\u2019t use emojis to point out coding\u00a0problems<\/figcaption><\/figure><p id=\"85d1\" class=\"graf graf--p graf-after--figure\"><strong class=\"markup--strong markup--p-strong\">You shouldn\u2019t be having visceral reactions to peoples\u2019 coding mistakes anyway.<\/strong><\/p><p id=\"a0e3\" class=\"graf graf--p graf-after--p\">It\u2019s okay to use emojis like \u201cthumbs-up\u201d or \u201chooray\u201d to signify that code looks good, but don\u2019t use them to point out problems.<\/p><figure id=\"d301\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*zNh3gFLoNNv4kGXI.\" data-width=\"1032\" data-height=\"170\" data-action=\"zoom\" data-action-value=\"0*zNh3gFLoNNv4kGXI.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*zNh3gFLoNNv4kGXI.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Emojis are great when used to\u00a0praise<\/figcaption><\/figure><h3 id=\"310f\" class=\"graf graf--h3 graf-after--figure\">Unhelpful behavior #7: not replying to all\u00a0comments<\/h3><p id=\"b506\" class=\"graf graf--p graf-after--h3\">Review seekers can contribute to unsupportive environments, too. If you merge code without addressing all the feedback, people are left wondering why they bothered to help you, and you send the message that some opinions are worth more than others.<\/p><p id=\"8fe7\" class=\"graf graf--p graf-after--p\">If a comment is out of scope or you won\u2019t be taking action on the feedback, just provide a brief note explaining why.<\/p><p id=\"43a1\" class=\"graf graf--p graf-after--p\">Don\u2019t ghost people.<\/p><figure id=\"f09e\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"0*6Mt-1qXjBnetnOQB.\" data-width=\"800\" data-height=\"287\" data-action=\"zoom\" data-action-value=\"0*6Mt-1qXjBnetnOQB.\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/0*6Mt-1qXjBnetnOQB.\" alt=\"image\" \/><\/div><figcaption class=\"imageCaption\">Merging code without addressing all the\u00a0feedback<\/figcaption><\/figure><h3 id=\"0f11\" class=\"graf graf--h3 graf-after--figure\">Unhelpful behavior #8: ignoring toxic behaviors from high performers<\/h3><p id=\"5f1d\" class=\"graf graf--p graf-after--h3\">Toxic behaviors should not be ignored or deemphasized because a developer is a high performer and extremely productive. Though this developer might be doing a fantastic job, it is important to keep in mind that this developer\u2019s toxic behaviors make them draining and stressful to work with for other developers.<\/p><p id=\"6c16\" class=\"graf graf--p graf-after--p\">On working with developers who commonly display toxic behaviors:<\/p><blockquote id=\"f666\" class=\"graf graf--pullquote graf-after--p\"><div>Others will find working with this individual draining and demotivating. They\u2019ll find themselves going to great lengths to avoid interacting with them, even if it means negatively impacting their own ability to complete tasks. Communication will shut down throughout your organization. If it becomes bad, your team will start looking for opportunities elsewhere. While you deal with the fallout of the exodus of talent and failing projects, this particular kind of developer will happily continue working as if nothing was wrong.\u200a\u2014\u200a<a href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" data-href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" class=\"markup--anchor markup--pullquote-anchor\" target=\"_blank\">Joseph\u00a0Gefroh<\/a><\/div><\/blockquote><p id=\"8259\" class=\"graf graf--p graf-after--pullquote\">Not taking steps to unlearn these toxic behaviors can have serious setbacks for the team, as developers feel overwhelmed, attacked, and demotivated. They start to dread the feedback processes that are supposed to help them grow.<\/p><p id=\"28fd\" class=\"graf graf--p graf-after--p\">I would personally feel a lot of anxiety whenever I got an email that a previous coworker (who was known for delivering toxic, unhelpful feedback) had commented on my pull request. Though toxic behaviors affect everyone differently, we can all agree that no one benefits from this toxicity.<\/p><p id=\"ce73\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Note:<\/strong> I want to be clear that slipping up and displaying one of the above toxic behaviors alone doesn\u2019t make a developer \u201ctoxic.\u201d Repeat offenses and patterns of caustic feedback say a lot, though.<\/p><h3 id=\"9d46\" class=\"graf graf--h3 graf-after--p\">Helpful Code Review Practices<\/h3><p id=\"5ea8\" class=\"graf graf--p graf-after--h3\">The following are some recommendations that can apply to any person in a collaborative, feedback-driven environment, even if they are filtered through the lens of a software engineer\u2019s eyes and are in the context of code reviews.<\/p><h3 id=\"3374\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #1: use questions or recommendations to drive\u00a0dialog<\/h3><p id=\"efb0\" class=\"graf graf--p graf-after--h3\">Never make demands for people to implement changes or ask judgmental questions, because this does not open up a dialog between you and the person whose code you are reviewing.<\/p><blockquote id=\"1ee7\" class=\"graf graf--blockquote graf-after--p\"><div>Why didn\u2019t you just pull these translations into a constants file?<\/div><\/blockquote><p id=\"2c4f\" class=\"graf graf--p graf-after--blockquote\">Though the above is a question, it doesn\u2019t create a dialog between you and your review seeker. It just forces them to defend themselves. Instead, ask what the review seeker thinks of your proposed solution, like:<\/p><blockquote id=\"dbe2\" class=\"graf graf--blockquote graf-after--p\"><div>What do you think about pulling these translations into a constants file? There are a lot, and a separate file might make sense at this point<\/div><\/blockquote><p id=\"f9d9\" class=\"graf graf--p graf-after--blockquote\">\u2026or make a recommendation like:<\/p><blockquote id=\"4080\" class=\"graf graf--blockquote graf-after--p\"><div>You introduced a lot of translation calls for \u201cfeature x\u201d in this file. It might make sense to create a separate file devoted to \u201cfeature x\u201d constants.<\/div><\/blockquote><h3 id=\"d53f\" class=\"graf graf--h3 graf-after--blockquote\">Helpful Behavior #2: collaborate, don\u2019t back-seat drive<\/h3><p id=\"60fc\" class=\"graf graf--p graf-after--h3\">When you pair-program with another person, you should be there to ask questions, debate, and point to resources.<\/p><blockquote id=\"e2a4\" class=\"graf graf--pullquote graf--startsWithDoubleQuote graf-after--p\"><div>\u201c\u2026when you want to help out or work with others, you should fully engage and not just butt in sporadically\u201d\u200a\u2014\u200aRecurse Center\u2019s <a href=\"https:\/\/www.recurse.com\/manual#no-backseat-driving\" data-href=\"https:\/\/www.recurse.com\/manual#no-backseat-driving\" class=\"markup--anchor markup--pullquote-anchor\" rel=\"noopener\" target=\"_blank\">User\u00a0Manual<\/a><\/div><\/blockquote><h3 id=\"e08f\" class=\"graf graf--h3 graf-after--pullquote\">Helpful Behavior #3: respond to every\u00a0comment<\/h3><p id=\"1b37\" class=\"graf graf--p graf-after--h3\">If you as a review seeker don\u2019t plan to apply a person\u2019s feedback, just leave a note letting them know. Don\u2019t ignore those who take time to help you.<\/p><p id=\"0e2a\" class=\"graf graf--p graf-after--p\">For example,<\/p><blockquote id=\"8cfb\" class=\"graf graf--blockquote graf-after--p\"><div>Person A: What do you think about creating a helper function for this api call? Looks good to me otherwise<\/div><\/blockquote><blockquote id=\"cda1\" class=\"graf graf--blockquote graf-after--blockquote\"><div>Person B: That line wasn\u2019t part of my changeset. I\u2019ll merge this code for now, but I\u2019ll also create a separate GitHub issue for the api call and put it in our group\u2019s backlog of to-do items.<\/div><\/blockquote><h3 id=\"398e\" class=\"graf graf--h3 graf-after--blockquote\">Helpful Behavior #4: know when to take a discussion offline<\/h3><p id=\"2d99\" class=\"graf graf--p graf-after--h3\">After dozens of contrasting PR comments and proposed solutions, it should be clear that online communication has become unproductive for the issue at hand. Send a meeting hit to involved members to discuss the issue offline.<\/p><p id=\"c481\" class=\"graf graf--p graf-after--p\">This way, your group can come to a decision more quickly and apply the solution.<\/p><blockquote id=\"28b3\" class=\"graf graf--pullquote graf-after--p\"><div>The kinds of problems that take hours and tons of comments can usually be resolved in a few minutes of productive talk.\u200a\u2014\u200a<a href=\"http:\/\/tidyjava.com\/better-code-reviewer\/\" data-href=\"http:\/\/tidyjava.com\/better-code-reviewer\/\" class=\"markup--anchor markup--pullquote-anchor\" rel=\"noopener\" target=\"_blank\">Tidy\u00a0Java<\/a><\/div><\/blockquote><h3 id=\"3cd1\" class=\"graf graf--h3 graf-after--pullquote\">Helpful Behavior #5: use opportunities to teach, and don\u2019t show\u00a0off<\/h3><p id=\"d985\" class=\"graf graf--p graf-after--h3\">Before choosing to participate in a code review, ask yourself:<\/p><blockquote id=\"d6d4\" class=\"graf graf--blockquote graf-after--p\"><div>Is your comment helping the other developer learn or are you nitpicking to participate?<\/div><\/blockquote><p id=\"f607\" class=\"graf graf--p graf-after--blockquote\">Reflect on why you are participating. Remember that the goal behind code reviews is to teach and help other developers grow. The goal isn\u2019t to grandstand.<\/p><h3 id=\"8b2f\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #6: don\u2019t show\u00a0surprise<\/h3><p id=\"4aab\" class=\"graf graf--p graf-after--h3\">Be careful not to show surprise that someone doesn\u2019t have the same knowledge that you do on a topic. Being comfortable admitting you lack experience with a topic is a great way to ask for help. Don\u2019t make people feel bad that they \u201cshould already have known\u201d this piece of info.<\/p><p id=\"d0bd\" class=\"graf graf--p graf-after--p\">For more on this, see Recurse Center\u2019s Social Rule \u201c<a href=\"https:\/\/www.recurse.com\/manual#no-feigned-surprise\" data-href=\"https:\/\/www.recurse.com\/manual#no-feigned-surprise\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Don\u2019t Feign Surprise<\/a>\u201d<\/p><h3 id=\"e166\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #7: automate what can\u00a0be<\/h3><p id=\"1051\" class=\"graf graf--p graf-after--h3\">Reviewing issues that can be caught by linters, git hooks, or automated tests are unhelpful because they often result in an avalanche of comments and come off as nitpicking. People are not particularly good at catching these issues, which is why automation tools exist.<\/p><p id=\"501c\" class=\"graf graf--p graf-after--p\">There are also tools out there that will run your tests when code is checked in and display warnings that a changeset is in violation of any of these tests. TeamCity and Jenkins CI both offer this functionality.<\/p><p id=\"109a\" class=\"graf graf--p graf-after--p\">Also, use git hooks. Git hooks run tests and linters when someone tries to check code in and they will intercept the commit if any buggy, unlinted code was introduced.<\/p><p id=\"edc4\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Let the tool point out issues so people don\u2019t have to.<\/strong><\/p><h3 id=\"bd49\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #8: refuse to normalize toxic\u00a0behavior<\/h3><p id=\"25cc\" class=\"graf graf--p graf-after--h3\">Don\u2019t join in and contribute to code review toxicity because it seems like the status quo, a hazing ritual for new developers, or just the way things are.<\/p><p id=\"51c2\" class=\"graf graf--p graf-after--p\">Search for teammates who are supportive and who will back you up.<\/p><p id=\"3bbc\" class=\"graf graf--p graf-after--p\">If you notice someone being unhelpful during code reviews, consider letting them know (if you feel safe in your role\/company doing so) and be direct.<\/p><h3 id=\"2833\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #9: managers\u200a\u2014\u200ahire carefully, listen to your team, and\u00a0enforce<\/h3><p id=\"cfaa\" class=\"graf graf--p graf-after--h3\">Managers have a lot of power to create positive and supportive cultures on their teams.<\/p><p id=\"6574\" class=\"graf graf--p graf-after--p\">To paraphrase the advice from \u201c<a href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" data-href=\"https:\/\/medium.com\/@jgefroh\/toxic-developers-considered-harmful-f7ea1494d4c0\" class=\"markup--anchor markup--p-anchor\" target=\"_blank\">Toxic Developers Considered Harmful<\/a>\u201d:<\/p><ul class=\"postList\"><li id=\"4fc6\" class=\"graf graf--li graf-after--p\">Prevent toxic developers from joining your team. During the hiring process, don\u2019t just stress the technical. Learn how your candidate collaborates and communicates. Critically analyze their work and see how they react. Make sure every hire is a culture add, not just a culture fit.<\/li><li id=\"29e7\" class=\"graf graf--li graf-after--li\">When toxic developers trickle in or you inherit them, ask your direct reports how their experience is working with others on the team during 1:1\u2019s. If you have a toxic dev, your direct reports\u2019 feedback will show it.<\/li><li id=\"7f36\" class=\"graf graf--li graf-after--li\">Talk to your toxic dev. Give them specific examples and actionable feedback. Work with them throughout your manager\/direct report relationship.<\/li><li id=\"6f30\" class=\"graf graf--li graf-after--li\">Don\u2019t isolate the toxic developer (*)<\/li><li id=\"aef4\" class=\"graf graf--li graf-after--li\">Iterate that you expect your team to contribute to a supportive environment<\/li><\/ul><p id=\"9666\" class=\"graf graf--p graf-after--li\">(*) Though the article suggests to isolate toxic developers, I believe it\u2019s important to encourage the toxic developer to continue to collaborate with the rest of the team, but in more healthy ways. Isolating toxic workers won\u2019t solve the problem. Though giving this dev independent work allows the team to suffer less toxicity in the short run, this toxic developer has not unlearned unhelpful behaviors. They\u2019ve just lost the platform to display these behaviors.<\/p><h3 id=\"8d86\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #10: set the standard as your team is small and\u00a0growing<\/h3><p id=\"a992\" class=\"graf graf--p graf-after--h3\">Teams that are small enough can adopt ideas and implement them from the start. Even if you don\u2019t think it\u2019s necessary to set a standard of what is and isn\u2019t okay because your team is currently collaborating well, remember that you want your team to stay great as you invite new members in.<\/p><h3 id=\"39b8\" class=\"graf graf--h3 graf-after--p\">Helpful Behavior #11: understand you might be part of the\u00a0problem<\/h3><p id=\"233b\" class=\"graf graf--p graf-after--h3\">In order to foster more supportive environments, it\u2019s important for people to be honest with themselves and reflect on any unhelpful behaviors they may have ever displayed.<\/p><p id=\"6d7c\" class=\"graf graf--p graf-after--p\">As a junior engineer, there were several occasions when I noticed a bug in someone\u2019s code and delighted, because it meant that I could point it out and be part of the review. I know now that I was using the opportunity to show off, not teach. I am trying to be mindful of this going forward.<\/p><p id=\"d336\" class=\"graf graf--p graf-after--p\">I think everyone should have this tough moment of self-reflection too.<\/p><p id=\"9942\" class=\"graf graf--p graf-after--p\">One last thing\u2026<\/p><h3 id=\"9802\" class=\"graf graf--h3 graf-after--p\">I am not policing the content of feedback, just asking people to be mindful of their\u00a0tone<\/h3><p id=\"9f28\" class=\"graf graf--p graf-after--h3\">I know feedback is important and I am not waging a war on feedback. I\u2019m also not asking teams to compromise their code quality. A supportive code review culture and high code quality DON\u2019T have to be mutually exclusive.<\/p><p id=\"16f0\" class=\"graf graf--p graf-after--p\">I simply hope people take steps to provide constructive, actionable feedback and create more supportive environments, so developers feel comfortable to learn, grow, and make mistakes. We can all improve together.<\/p><h4 id=\"c4d2\" class=\"graf graf--h4 graf-after--p\">TL;DR:<\/h4><figure id=\"3a79\" class=\"graf graf--figure graf--iframe graf-after--h4\"><p id=\"db79\" class=\"graf graf--p graf-after--figure graf--trailing\">Questions or comments? Feel free to reach out on <a href=\"https:\/\/twitter.com\/sandyaaaas\" data-href=\"https:\/\/twitter.com\/sandyaaaas\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">Twitter<\/a>. Video recording available <a href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" data-href=\"https:\/\/www.youtube.com\/watch?v=QIUwGa-MttQ\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">here<\/a>.<\/p><\/figure>","created_at":"2018-11-07T16:23:55+0000","updated_at":"2018-11-07T16:23:55+0000","published_at":"2018-01-19T16:41:16+0000","published_by":["Sandya Sankarram"],"annotations":[],"mimetype":"text\/html","reading_time":13,"domain_name":"medium.freecodecamp.org","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*PU20rbKCMm3C2BLEM24dgg.jpeg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":175,"title":"Lisp in Small Pieces of Clojure","url":"http:\/\/gmorpheme.net\/lisp-in-small-pieces-of-clojure-part-1.html","content":"<p><img alt=\"L.I.S.P. cover\" class=\"right\" src=\"http:\/\/www.gmorpheme.net\/images\/lispcover.jpg\" \/><\/p><p><a href=\"http:\/\/www.amazon.co.uk\/Lisp-Small-Pieces-Christian-Queinnec\/dp\/0521545668\/ref=sr_1_1?ie=UTF8&amp;qid=1384103047&amp;sr=8-1&amp;keywords=lisp+in+small+pieces\">Lisp in Small Pieces<\/a>\nis a deep exploration of some of the fundamentals of programming that proceeds by describing a series of lisp evaluators and compilers over the course of 400 pages or\u00a0so.<\/p><p>Personally I find the text pretty baffling at times and the English is\noften rather curious. Also the Kindle version I have garbles and\nshrinks some of the code which adds a certain challenge to the\nreading. Nevertheless the depth and detail of the content are\u00a0awe-inspiring.<\/p><p>It appears in Rich Hickey\u2019s list of\n<a href=\"http:\/\/www.amazon.com\/Clojure-Bookshelf\/lm\/R3LG3ZBZS4GCTH\/ref=cm_lm_byauthor_title_full\">\u201cbooks that influenced Clojure, in no particular order\u201d<\/a>.\nIt\u2019s mentioned a few times in Lambda the Ultimate\u2019s\n<a href=\"http:\/\/lambda-the-ultimate.org\/node\/492\">\u201cGetting Started\u201d<\/a> page\nwhich is a great resource for PL texts. And Peter Norvig no less\ndescribes it as \u201cThe best book available on Lisp\u00a0implementation\u201d.<\/p><h3>\u2026of\u00a0Clojure<\/h3><p>There have probably been a few gestures towards porting examples from\nthe book to Clojure already (I know of a\n<a href=\"https:\/\/gist.github.com\/fogus\/3354936\">gist<\/a> by Fogus at\u00a0least).<\/p><p>I don\u2019t intend to produce a complete port of the code from LiSP (like\n<a href=\"http:\/\/juliangamble.com\/blog\/2012\/07\/13\/amazing-lisp-books-living-again-in-clojure\/\">these<\/a>).\nI\u2019m probably not interested\u00a0enough.<\/p><p>However I have the book and have worked through some of the\nimplementations. And, my implementation language of choice is for the\nmost part Clojure. So it\u2019s at least possible a few gists will\u00a0appear.<\/p><p>In the following, I\u2019ll have no hesitation in commiting crimes against\nidiomatic Clojure in the name of faithfulness to the book and crimes\nagainst the book in the name of idiomatic Clojure. Just a\u00a0warning.<\/p><h3>Chapter\u00a01<\/h3><p>The gist for chapter one is <a href=\"https:\/\/gist.github.com\/gmorpheme\/9293548#file-eval-clj\">here<\/a>.<\/p><p>Chapter one delivers a simple interpreter that is heavily parasitic on\nthe underlying lisp. As the implementation language in the book is\nassumed to be Scheme we need to build a little bit of a compatibility\nshim in Clojure in order to stay close to the implementation in the\u00a0book.<\/p><p>In particular, we\u00a0need:<\/p><ul><li>Mutable\u00a0environments<\/li>\n<li>Mutable cons\u00a0cells<\/li>\n<\/ul><h4>Mutable\u00a0Environments<\/h4><p>The interpreted language provides a <code>set!<\/code> special form for mutating\nvariables. The environment data structures used in chapter one are\nsimple A-lists that support <code>set!<\/code> by mutating the cdr of the cons\u00a0cells.<\/p><p>At this stage I decided to stick with a similar structure in my\nClojure implementation (atom wrapped round seq of seqs) and provided\n<code>lookup<\/code>, <code>update!<\/code> and <code>extend<\/code> operations as per the\u00a0book.<\/p><p>Fogus chose a more idiomatic map implementation which I think is\nequivalent except for pathological cases\u00a0like:<\/p><div class=\"highlight\"><pre>((lambda (x x x) (+ x x)) 1 2 3))\n<\/pre><\/div><p>\u2026and in that it provides for <code>set!<\/code> adding new bindings\u00a0automatically.<\/p><p>A large part of the exploration of the book regards the behaviour of\nthe environment and the global environment in particular so I decided\nto stick with the book on this. At this stage there is no dynamic\ncreation of global variables so the implementation creates some\ninitial variables in the environment (<code>foo<\/code>, <code>bar<\/code>, <code>fib<\/code>, <code>fact<\/code>) for\nuse by\u00a0programs.<\/p><h4>The\u00a0Evaluator<\/h4><p>The evaluator itself is a straight translation from the book differing\nonly in language specifics like <code>atom?<\/code>. <\/p><h4>Mutable Cons\u00a0Cells<\/h4><p>Just as in the scheme implementation, functionality is exposed to the\nchild lisp using <code>definitial<\/code> and <code>defprimitive<\/code> macros (substituting\nClojure\u2019s <code>defmacro<\/code> for the Scheme \u201chygienic\u201d macro\u00a0approach).<\/p><p>However Clojure is very different from Scheme. If we want to provide\nsimilar facilities to the child language (and mutability is essential\nto many of the problems discussed in the book) there\u2019s some work to\ndo. In particular we need to provide mutable cons cells, which means\nthe <code>cons<\/code> we need to expose to the child lisp is most certainly not\nthe <code>cons<\/code> of\u00a0Clojure.<\/p><p>We could work around this by putting atoms around pairs (and lists\nwould contain nested atoms in their cdr positions) but I chose instead\nto brute-force my way\u00a0out:<\/p><div class=\"highlight\"><pre>(defprotocol MutableCons\n  (mcar [self])\n  (mcdr [self])\n  (mset-car! [self val])\n  (mset-cdr! [self val]))\n(deftype Cell [^{:unsynchronized-mutable true} a ^{:unsynchronized-mutable true} b]\n  MutableCons\n  (mcar [self] a)\n  (mcdr [self] b)\n  (mset-car! [self val] (set! a val))\n  (mset-cdr! [self val] (set! b val))\n  Object\n  (toString [self] (str \"(\" a \" . \" b \")\")))\n(defn mcons [a b]\n  (Cell. a b))\n(defprimitive cons mcons 2)\n(defprimitive car mcar 1)\n(defprimitive cdr mcdr 1)\n(defprimitive set-car! mset-car! 2)\n(defprimitive set-cdr! mset-cdr! 2)\n<\/pre><\/div><h4>At the REPL<\/h4><p>It\u00a0works.<\/p><div class=\"highlight\"><pre>;; we have mutable cons cells, but printing leaves much to be desired\n(set! foo (cons 1 2))\n;; =&gt; #&lt;Cell (1 . 2)&gt;\n(set-cdr! foo (cons 2 nil))\n;; =&gt; #&lt;Cell (2 . )&gt;\nfoo\n;; =&gt; #&lt;Cell (1 . (2 . ))&gt;\n;; truthiness is parasitic on the underlying lisp, in clojure 0 is truthy\n(if 0 (quote truthy) (quote falsey))\n;; =&gt; truthy\n(begin (set! bar 1) (set! bar 2) bar)\n;; =&gt; 2\n((lambda (x) (+ x 1)) 12)\n;; =&gt; 13\n(set! fib (lambda (x) (if (&lt; x 2) 1 (+ (fib (- x 1)) (fib (- x 2))))))\n;; =&gt; #&lt;eval$make_function$fn__576 lisp.chapter1.eval$make_function$fn__576@5702cbd3&gt;\n(fib 8)\n;; =&gt; 34\n<\/pre><\/div><h4>Further\u00a0Discussion<\/h4><p>Chapter one contains some discussion of lexical closure. The lisp we\nhave create works as\u00a0expected:<\/p><div class=\"highlight\"><pre>((lambda (a) ((lambda (b) (cons a b)) (+ 2 a))) 1)\n;; =&gt; #&lt;Cell (1 . 3)&gt;\n<\/pre><\/div><p>Note also that our lisp is a lexical lisp in that the y in bar\u2019s\nenvironment below is not visible from within\u00a0foo.<\/p><div class=\"highlight\"><pre>(set! foo (lambda (x) (cons x y)))\n;; =&gt; #&lt;eval$make_function$fn__576 lisp.chapter1.eval$make_function$fn__576@42367c17&gt;\n(set! bar (lambda (y) (foo 1991)))\n;; =&gt; #&lt;eval$make_function$fn__576 lisp.chapter1.eval$make_function$fn__576@70cebf7&gt;\n(set! y 0)\n;; =&gt; 0\n(bar 100)\n;; =&gt; #&lt;Cell (1991 . 0)&gt;\n(foo 3)\n;; =&gt; #&lt;Cell (3 . 0)&gt;\n<\/pre><\/div><p>As the gist illustrates, providing cross-cutting enhancements like\ntracing, or enhancing any of the component parts (like the\nenvironment) is currently difficult because of the inflexible\u00a0design.<\/p><p>There are numerous ways in which the design might be\u00a0improved:<\/p><ul><li>\n<p>the environment functions (<code>lookup<\/code>, <code>update!<\/code>, <code>extend<\/code>) are crying\n  out to become a protocol, allowing alternative implementations to be\n  more easily\u00a0substituted<\/p>\n<\/li>\n<li>\n<p>the evaluator itself could be pulled apart as a multimethod which\n  dispatches on the symbol in the car position - a similar approach is\n  used in the core.async go macro implementation (see <code>-item-to-ssa<\/code>\n  <a href=\"https:\/\/github.com\/clojure\/core.async\/blob\/master\/src\/main\/clojure\/clojure\/core\/async\/impl\/ioc_macros.clj\">here<\/a>\n  and some explanation\n  <a href=\"http:\/\/www.youtube.com\/watch?v=HXfDK1OYpco\">here<\/a>)<\/p>\n<\/li>\n<li>\n<p>a monadic approach to the evaluation could be introduced to provide\n  for the clean addition of cross-cutting concerns (see Wadler\u2019s \u201cThe\n  essence of functional programming\u201d\n  <a href=\"http:\/\/homepages.inf.ed.ac.uk\/wadler\/topics\/monads.html\">here<\/a> for\n  an example of iterative enhancements to an interpreter by altering\n  the\u00a0monad)<\/p>\n<\/li>\n<\/ul>","created_at":"2018-11-07T16:23:55+0000","updated_at":"2018-11-07T16:23:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"gmorpheme.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":173,"title":"r\/emacs - Advanced usage of eshell?","url":"https:\/\/www.reddit.com\/r\/emacs\/comments\/1zkj2d\/advanced_usage_of_eshell\/","content":"<p class=\"yklcuq-10 hpxQMr\">In <a target=\"_blank\" class=\"yklcuq-27 blazbt\" href=\"http:\/\/www.reddit.com\/r\/emacs\/comments\/1yzk6n\/emacs_shells_or_tmux\/\">a previous discussion<\/a>, I asked about what kinds of shells people like to use with Emacs (eshell \/ ansi-term, or an external term and Emacs GUI or tmux and Emacs in a terminal window).<\/p><p class=\"yklcuq-10 hpxQMr\">I have been experimenting with the two, but I'm also curious about eshell, and I had a look at <a target=\"_blank\" class=\"yklcuq-27 blazbt\" href=\"http:\/\/www.masteringemacs.org\/articles\/2010\/12\/13\/complete-guide-mastering-eshell\/\">the introduction to eshell<\/a> on Mastering Emacs... Some people say it's very powerful, you can access all Emacs functions, etc. I like the idea of being able to easily open a file in a buffer with ff, or a dired buffer with d (after defining aliases), having man\/locate automatically open buffers etc. But I feel like there's probably a lot more power\/potential...<\/p><p class=\"yklcuq-10 hpxQMr\">So I'd love to hear about some power-users use of eshell - do you use Elisp as a scripting language, piping lists of files into a small function that you define on the command line, do you have some great aliases defined, what else do you do that would be difficult to do in a regular shell?<\/p><p class=\"yklcuq-10 hpxQMr\">Thinking about what I usually do on the command line, part of it is just moving around, looking at files, moving them around etc (I should probably be better at using dired, etc for that kind of stuff), and part of it is running interactive applications, mainly ipython (which I need to run in ansi-term to get tab-completion and up-arrow to work, I think).<\/p>","created_at":"2018-11-07T16:23:54+0000","updated_at":"2018-11-07T16:23:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.reddit.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":172,"title":"beetbox\/beets","url":"https:\/\/github.com\/beetbox\/beets","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><a href=\"https:\/\/pypi.python.org\/pypi\/beets\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/955b3f95a77c616f651880bc74f5f3ca3d63a8d1\/687474703a2f2f696d672e736869656c64732e696f2f707970692f762f62656574732e737667\" data-canonical-src=\"http:\/\/img.shields.io\/pypi\/v\/beets.svg\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/codecov.io\/github\/beetbox\/beets\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a107dcf92735fd8354477f7fc1499501fe63c5cd\/687474703a2f2f696d672e736869656c64732e696f2f636f6465636f762f632f6769746875622f62656574626f782f62656574732e737667\" data-canonical-src=\"http:\/\/img.shields.io\/codecov\/c\/github\/beetbox\/beets.svg\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/beetbox\/beets\" rel=\"nofollow\"><img alt=\"https:\/\/travis-ci.org\/beetbox\/beets.svg?branch=master\" src=\"https:\/\/camo.githubusercontent.com\/3f09e9db80ad3b031f6072884fc15cbc0dc8d848\/68747470733a2f2f7472617669732d63692e6f72672f62656574626f782f62656574732e7376673f6272616e63683d6d6173746572\" data-canonical-src=\"https:\/\/travis-ci.org\/beetbox\/beets.svg?branch=master\" \/><\/a>\n<p>Beets is the media library management system for obsessive-compulsive music\ngeeks.<\/p>\n<p>The purpose of beets is to get your music collection right once and for all.\nIt catalogs your collection, automatically improving its metadata as it goes.\nIt then provides a bouquet of tools for manipulating and accessing your music.<\/p>\n<p>Here's an example of beets' brainy tag corrector doing its thing:<\/p>\n<pre>$ beet import ~\/music\/ladytron\nTagging:\n    Ladytron - Witching Hour\n(Similarity: 98.4%)\n * Last One Standing      -&gt; The Last One Standing\n * Beauty                 -&gt; Beauty*2\n * White Light Generation -&gt; Whitelightgenerator\n * All the Way            -&gt; All the Way...\n<\/pre>\n<p>Because beets is designed as a library, it can do almost anything you can\nimagine for your music collection. Via <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/\" rel=\"nofollow\">plugins<\/a>, beets becomes a panacea:<\/p>\n<ul><li>Fetch or calculate all the metadata you could possibly need: <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/fetchart.html\" rel=\"nofollow\">album art<\/a>,\n<a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/lyrics.html\" rel=\"nofollow\">lyrics<\/a>, <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/lastgenre.html\" rel=\"nofollow\">genres<\/a>, <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/acousticbrainz.html\" rel=\"nofollow\">tempos<\/a>, <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/replaygain.html\" rel=\"nofollow\">ReplayGain<\/a> levels, or <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/chroma.html\" rel=\"nofollow\">acoustic\nfingerprints<\/a>.<\/li>\n<li>Get metadata from <a href=\"http:\/\/musicbrainz.org\/\" rel=\"nofollow\">MusicBrainz<\/a>, <a href=\"http:\/\/www.discogs.com\/\" rel=\"nofollow\">Discogs<\/a>, and <a href=\"https:\/\/www.beatport.com\" rel=\"nofollow\">Beatport<\/a>. Or guess\nmetadata using songs' filenames or their acoustic fingerprints.<\/li>\n<li><a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/convert.html\" rel=\"nofollow\">Transcode audio<\/a> to any format you like.<\/li>\n<li>Check your library for <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/duplicates.html\" rel=\"nofollow\">duplicate tracks and albums<\/a> or for <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/missing.html\" rel=\"nofollow\">albums that\nare missing tracks<\/a>.<\/li>\n<li>Clean up crufty tags left behind by other, less-awesome tools.<\/li>\n<li>Embed and extract album art from files' metadata.<\/li>\n<li>Browse your music library graphically through a Web browser and play it in any\nbrowser that supports <a href=\"http:\/\/www.w3.org\/TR\/html-markup\/audio.html\" rel=\"nofollow\">HTML5 Audio<\/a>.<\/li>\n<li>Analyze music files' metadata from the command line.<\/li>\n<li>Listen to your library with a music player that speaks the <a href=\"http:\/\/www.musicpd.org\/\" rel=\"nofollow\">MPD<\/a> protocol\nand works with a staggering variety of interfaces.<\/li>\n<\/ul><p>If beets doesn't do what you want yet, <a href=\"http:\/\/beets.readthedocs.org\/page\/dev\/plugins.html\" rel=\"nofollow\">writing your own plugin<\/a> is\nshockingly simple if you know a little Python.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#install\"><\/a>Install<\/h3>\n<p>You can install beets by typing <code>pip install beets<\/code>. Then check out the\n<a href=\"http:\/\/beets.readthedocs.org\/page\/guides\/main.html\" rel=\"nofollow\">Getting Started<\/a> guide.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#contribute\"><\/a>Contribute<\/h3>\n<p>Check out the <a href=\"https:\/\/github.com\/beetbox\/beets\/wiki\/Hacking\">Hacking<\/a> page on the wiki for tips on how to help out.\nYou might also be interested in the <a href=\"http:\/\/docs.beets.io\/page\/dev\/\" rel=\"nofollow\">For Developers<\/a> section in the docs.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#read-more\"><\/a>Read More<\/h3>\n<p>Learn more about beets at <a href=\"http:\/\/beets.io\/\" rel=\"nofollow\">its Web site<\/a>. Follow <a href=\"http:\/\/twitter.com\/b33ts\/\" rel=\"nofollow\">@b33ts<\/a> on Twitter for\nnews and updates.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#authors\"><\/a>Authors<\/h3>\n<p>Beets is by <a href=\"http:\/\/www.cs.cornell.edu\/~asampson\/\" rel=\"nofollow\">Adrian Sampson<\/a> with a supporting cast of thousands. For help,\nplease visit our <a href=\"https:\/\/discourse.beets.io\" rel=\"nofollow\">forum<\/a>.<\/p>\n<\/article>","created_at":"2018-11-07T16:23:53+0000","updated_at":"2018-11-07T16:23:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/15920313?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":170,"title":"Never Write Your Own Database (2017)","url":"https:\/\/news.ycombinator.com\/item?id=16995612","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D16995612\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"16995612\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/medium.com\/@terrycrowley\/never-write-your-own-database-736f704c780\" class=\"storylink\">Never Write Your Own Database (2017)<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=medium.com\">medium.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        445 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=ahiknsr\" class=\"hnuser\">ahiknsr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16995612\">6 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=16995612&amp;goto=item%3Fid%3D16995612\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Never%20Write%20Your%20Own%20Database&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Never%20Write%20Your%20Own%20Database\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=16995612&amp;auth=ea01482c96854559ec6cf1d9c66457d2dc4916ae\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=16995612\">154\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:23:52+0000","updated_at":"2018-11-07T16:23:52+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":171,"title":"beetbox\/beets","url":"https:\/\/github.com\/beetbox\/beets","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><a href=\"https:\/\/pypi.python.org\/pypi\/beets\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/955b3f95a77c616f651880bc74f5f3ca3d63a8d1\/687474703a2f2f696d672e736869656c64732e696f2f707970692f762f62656574732e737667\" data-canonical-src=\"http:\/\/img.shields.io\/pypi\/v\/beets.svg\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/codecov.io\/github\/beetbox\/beets\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a107dcf92735fd8354477f7fc1499501fe63c5cd\/687474703a2f2f696d672e736869656c64732e696f2f636f6465636f762f632f6769746875622f62656574626f782f62656574732e737667\" data-canonical-src=\"http:\/\/img.shields.io\/codecov\/c\/github\/beetbox\/beets.svg\" alt=\"image\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/beetbox\/beets\" rel=\"nofollow\"><img alt=\"https:\/\/travis-ci.org\/beetbox\/beets.svg?branch=master\" src=\"https:\/\/camo.githubusercontent.com\/3f09e9db80ad3b031f6072884fc15cbc0dc8d848\/68747470733a2f2f7472617669732d63692e6f72672f62656574626f782f62656574732e7376673f6272616e63683d6d6173746572\" data-canonical-src=\"https:\/\/travis-ci.org\/beetbox\/beets.svg?branch=master\" \/><\/a>\n<p>Beets is the media library management system for obsessive-compulsive music\ngeeks.<\/p>\n<p>The purpose of beets is to get your music collection right once and for all.\nIt catalogs your collection, automatically improving its metadata as it goes.\nIt then provides a bouquet of tools for manipulating and accessing your music.<\/p>\n<p>Here's an example of beets' brainy tag corrector doing its thing:<\/p>\n<pre>$ beet import ~\/music\/ladytron\nTagging:\n    Ladytron - Witching Hour\n(Similarity: 98.4%)\n * Last One Standing      -&gt; The Last One Standing\n * Beauty                 -&gt; Beauty*2\n * White Light Generation -&gt; Whitelightgenerator\n * All the Way            -&gt; All the Way...\n<\/pre>\n<p>Because beets is designed as a library, it can do almost anything you can\nimagine for your music collection. Via <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/\" rel=\"nofollow\">plugins<\/a>, beets becomes a panacea:<\/p>\n<ul><li>Fetch or calculate all the metadata you could possibly need: <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/fetchart.html\" rel=\"nofollow\">album art<\/a>,\n<a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/lyrics.html\" rel=\"nofollow\">lyrics<\/a>, <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/lastgenre.html\" rel=\"nofollow\">genres<\/a>, <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/acousticbrainz.html\" rel=\"nofollow\">tempos<\/a>, <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/replaygain.html\" rel=\"nofollow\">ReplayGain<\/a> levels, or <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/chroma.html\" rel=\"nofollow\">acoustic\nfingerprints<\/a>.<\/li>\n<li>Get metadata from <a href=\"http:\/\/musicbrainz.org\/\" rel=\"nofollow\">MusicBrainz<\/a>, <a href=\"http:\/\/www.discogs.com\/\" rel=\"nofollow\">Discogs<\/a>, and <a href=\"https:\/\/www.beatport.com\" rel=\"nofollow\">Beatport<\/a>. Or guess\nmetadata using songs' filenames or their acoustic fingerprints.<\/li>\n<li><a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/convert.html\" rel=\"nofollow\">Transcode audio<\/a> to any format you like.<\/li>\n<li>Check your library for <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/duplicates.html\" rel=\"nofollow\">duplicate tracks and albums<\/a> or for <a href=\"http:\/\/beets.readthedocs.org\/page\/plugins\/missing.html\" rel=\"nofollow\">albums that\nare missing tracks<\/a>.<\/li>\n<li>Clean up crufty tags left behind by other, less-awesome tools.<\/li>\n<li>Embed and extract album art from files' metadata.<\/li>\n<li>Browse your music library graphically through a Web browser and play it in any\nbrowser that supports <a href=\"http:\/\/www.w3.org\/TR\/html-markup\/audio.html\" rel=\"nofollow\">HTML5 Audio<\/a>.<\/li>\n<li>Analyze music files' metadata from the command line.<\/li>\n<li>Listen to your library with a music player that speaks the <a href=\"http:\/\/www.musicpd.org\/\" rel=\"nofollow\">MPD<\/a> protocol\nand works with a staggering variety of interfaces.<\/li>\n<\/ul><p>If beets doesn't do what you want yet, <a href=\"http:\/\/beets.readthedocs.org\/page\/dev\/plugins.html\" rel=\"nofollow\">writing your own plugin<\/a> is\nshockingly simple if you know a little Python.<\/p>\n<h3><a id=\"user-content-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#install\"><\/a>Install<\/h3>\n<p>You can install beets by typing <code>pip install beets<\/code>. Then check out the\n<a href=\"http:\/\/beets.readthedocs.org\/page\/guides\/main.html\" rel=\"nofollow\">Getting Started<\/a> guide.<\/p>\n<h3><a id=\"user-content-contribute\" class=\"anchor\" aria-hidden=\"true\" href=\"#contribute\"><\/a>Contribute<\/h3>\n<p>Check out the <a href=\"https:\/\/github.com\/beetbox\/beets\/wiki\/Hacking\">Hacking<\/a> page on the wiki for tips on how to help out.\nYou might also be interested in the <a href=\"http:\/\/docs.beets.io\/page\/dev\/\" rel=\"nofollow\">For Developers<\/a> section in the docs.<\/p>\n<h3><a id=\"user-content-read-more\" class=\"anchor\" aria-hidden=\"true\" href=\"#read-more\"><\/a>Read More<\/h3>\n<p>Learn more about beets at <a href=\"http:\/\/beets.io\/\" rel=\"nofollow\">its Web site<\/a>. Follow <a href=\"http:\/\/twitter.com\/b33ts\/\" rel=\"nofollow\">@b33ts<\/a> on Twitter for\nnews and updates.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#authors\"><\/a>Authors<\/h3>\n<p>Beets is by <a href=\"http:\/\/www.cs.cornell.edu\/~asampson\/\" rel=\"nofollow\">Adrian Sampson<\/a> with a supporting cast of thousands. For help,\nplease visit our <a href=\"https:\/\/discourse.beets.io\" rel=\"nofollow\">forum<\/a>.<\/p>\n<\/article>","created_at":"2018-11-07T16:23:52+0000","updated_at":"2018-11-07T16:23:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/15920313?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":168,"title":"12 \u201cManager READMEs\u201d from Silicon Valley\u2019s Top Tech Companies","url":"https:\/\/hackernoon.com\/12-manager-readmes-from-silicon-valleys-top-tech-companies-26588a660afe?gi=9c14abb8ac9d","content":"<section class=\"section section--body section--first\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--fullWidth\"><figure id=\"c750\" class=\"graf graf--figure graf--layoutFillWidth graf--leading\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*JPGxBj3qjqgKctsp9oLqmQ.png\" data-width=\"1412\" data-height=\"790\" data-is-featured=\"true\" src=\"https:\/\/cdn-images-1.medium.com\/max\/2000\/1*JPGxBj3qjqgKctsp9oLqmQ.png\" alt=\"image\" \/><\/div><\/figure><\/div><div class=\"section-inner sectionLayout--insetColumn\"><h2 id=\"522f\" class=\"graf graf--h4 graf-after--h3 graf--subtitle\">What does tech management at Slack, HubSpot, Netflix, Etsy, Shopify, InVision, and more have in common? A lot, apparently!<\/h2><p id=\"bb44\" class=\"graf graf--p graf-after--h4\">When we launched our first <a href=\"https:\/\/soapboxhq.com\/modern-managers\/heather\" data-href=\"https:\/\/soapboxhq.com\/modern-managers\/heather\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">SoapBox hero<\/a> I was surprised to discover a document used to introduce her working style to her team. I had never seen that before and it was awesome!<\/p><p id=\"007e\" class=\"graf graf--p graf-after--p\">Did others have these? I needed to know. Eventually, I found a treasure trove of examples by Engineering Leaders from companies with some of the best culture in tech.<\/p><p id=\"3725\" class=\"graf graf--p graf-after--p\">They called them Manager READMEs, or Manager User Manuals. And they\u2019ve graciously let me share them with you.<\/p><p id=\"bd10\" class=\"graf graf--p graf-after--p\"><strong class=\"markup--strong markup--p-strong\">Click the images to read their READMEs.<\/strong> At the end I\u2019ve included my notes in case you\u2019re planning on building your own.<\/p><p id=\"15b0\" class=\"graf graf--p graf-after--p graf--trailing\">But first, the inspiration! \ud83d\udc40 \ud83c\udf81 \ud83c\udf89<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"9a23\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/rands\" data-href=\"https:\/\/twitter.com\/rands\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Michael Lopp<\/a> (aka rands), VP Eng. at <a href=\"https:\/\/slack.com\" data-href=\"https:\/\/slack.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Slack<\/a> ~~~~~~~~~<\/h3><figure id=\"f80c\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" data-href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*KvpV191-jSpgeWzhtYOQVA.png\" data-width=\"1490\" data-height=\"1050\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*KvpV191-jSpgeWzhtYOQVA.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" data-href=\"http:\/\/randsinrepose.com\/archives\/how-to-rands\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http:\/\/randsinrepose.com\/archives\/how-to-rands\/<\/a><\/figcaption><\/figure><h4 id=\"ddd4\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"d3bc\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--h4\">\u201cOur average week\u201d covers what a typical working week with Michael is like, and how to engage him in it.<\/li><li id=\"2d32\" class=\"graf graf--li graf-after--li\">He might chat you on weekends. Feel free to read it Monday.<\/li><li id=\"8784\" class=\"graf graf--li graf-after--li\">He\u2019s an introvert (\u201cDo not confuse my quiet with lack of engagement.\u201d)<\/li><li id=\"f330\" class=\"graf graf--li graf-after--li graf--trailing\">Gossip is a trigger for him.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"d933\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/bnmrrs?lang=en\" data-href=\"https:\/\/twitter.com\/bnmrrs?lang=en\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Ben Morris<\/a>, Sr. Developer Lead at <a href=\"https:\/\/shopify.com\" data-href=\"https:\/\/shopify.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Shopify<\/a> ~~~~~~~~~<\/h3><figure id=\"853b\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*pgzHHvQUgehI3tSL1OXOGg.png\" data-width=\"1802\" data-height=\"1030\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*pgzHHvQUgehI3tSL1OXOGg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/presentation\/d\/1LGL7fh5zWx8XqHRBra51LcMIHCUluqrdXZ_-XBTXqlg\/edit#slide=id.g23dc7fe4e1_2_85<\/a><\/figcaption><\/figure><h4 id=\"6e05\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"78a1\" class=\"graf graf--li graf-after--h4\">Super clear description of his role and goals.<\/li><li id=\"586d\" class=\"graf graf--li graf-after--li\">Explicitly stating how he can help you (and how you can help him do <em class=\"markup--em markup--li-em\">his<\/em> job).<\/li><li id=\"0bbe\" class=\"graf graf--li graf-after--li graf--trailing\">Specifically how he incorporates Shopify values into his role.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"652c\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/royrapoport\" data-href=\"https:\/\/twitter.com\/royrapoport\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Roy<\/a>, Director at <a href=\"https:\/\/slack.com\" data-href=\"https:\/\/slack.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Slack<\/a> (frmly Netflix.com) ~~~~~~~<\/h3><figure id=\"e8ad\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/bit.ly\/roy-slack-readme\" data-href=\"http:\/\/bit.ly\/roy-slack-readme\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*0UC40BUNV_y-yNYB2cBiUg.png\" data-width=\"1808\" data-height=\"1028\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*0UC40BUNV_y-yNYB2cBiUg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"http:\/\/bit.ly\/roy-slack-readme\" data-href=\"http:\/\/bit.ly\/roy-slack-readme\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">http:\/\/bit.ly\/roy-slack-readme<\/a><\/figcaption><\/figure><h4 id=\"e115\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"27d1\" class=\"graf graf--li graf-after--h4\">It\u2019s a mix of work and life. He has three goats. That\u2019s neat. Something to chat about.<\/li><li id=\"ab35\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--li\">\u201cI don\u2019t believe in personality conflicts\u201d At least not as a manager at work.<\/li><li id=\"9da6\" class=\"graf graf--li graf-after--li\">One of the best descriptions of what a manager\u2019s job is I\u2019ve read in a long time.<\/li><\/ul><figure id=\"5b86\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--li\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/bit.ly\/welcometonetflixandme\" data-href=\"https:\/\/bit.ly\/welcometonetflixandme\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*I3Sx95TyxA4xvOa2GAJvyQ.png\" data-width=\"1808\" data-height=\"1028\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*I3Sx95TyxA4xvOa2GAJvyQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/bit.ly\/welcometonetflixandme\" data-href=\"https:\/\/bit.ly\/welcometonetflixandme\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener noopener\" target=\"_blank\">https:\/\/bit.ly\/welcometonetflixandme<\/a><\/figcaption><\/figure><p id=\"ffb2\" class=\"graf graf--p graf-after--figure\">Roy\u2019s first README was built at Netflix. I think it\u2019s great to see how Roy\u2019s document has changed between two companies. It\u2019s a great reminder that a manager adapts to the environment they are in.<\/p><p id=\"222f\" class=\"graf graf--p graf-after--p graf--trailing\">This is one of the first READMEs you can find online. Many have taken inspiration from Roy.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"4261\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/orenellenbogen\" data-href=\"https:\/\/twitter.com\/orenellenbogen\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Oren<\/a>, VP Eng. at <a href=\"https:\/\/forter.com\" data-href=\"https:\/\/forter.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Forter<\/a> (+<a href=\"http:\/\/softwareleadweekly.com\/\" data-href=\"http:\/\/softwareleadweekly.com\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"nofollow noopener\" target=\"_blank\">SoftwareLeadWeekly.com<\/a>)<\/h3><figure id=\"9de9\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" data-href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*LPKPBbi6NehCkkJ3Pn9HQQ.png\" data-width=\"1500\" data-height=\"924\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*LPKPBbi6NehCkkJ3Pn9HQQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" data-href=\"https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/document\/d\/1sx5ssYb_xMrmwPpyjD5xP7RvQ7cHweDYlRGn2SXztKw\/edit#heading=h.g1q7u9cc1uha<\/a><\/figcaption><\/figure><h4 id=\"4047\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"db08\" class=\"graf graf--li graf-after--h4\">Links, images, video\u200a\u2014\u200aFurther material is available to get a deeper understanding if you\u2019re curious!<\/li><li id=\"51ee\" class=\"graf graf--li graf-after--li\">Personal quirks (specifically the ones he doesn\u2019t like about himself!)<\/li><li id=\"5826\" class=\"graf graf--li graf-after--li graf--trailing\">Expected response times by communication channel (helpful especially for remote teams).<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"2476\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/aaronlerch\" data-href=\"https:\/\/twitter.com\/aaronlerch\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Aaron<\/a>, Eng. Manager at <a href=\"https:\/\/invisionapp.com\" data-href=\"https:\/\/invisionapp.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">InVision<\/a> ~~~~~~~~~~~~~~~~~<\/h3><figure id=\"977a\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*6z_AznTfTEsR8HNDds1oMQ.png\" data-width=\"1810\" data-height=\"1028\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*6z_AznTfTEsR8HNDds1oMQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/presentation\/d\/1F2PWxQ-sKJ1uAlrhU9ULXOVmH-CW6dw_ufMzbFLAWHA\/edit#slide=id.p<\/a><\/figcaption><\/figure><h4 id=\"d3c2\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"daf2\" class=\"graf graf--li graf-after--h4\">His personal goal as a manager at InVision.<\/li><li id=\"2149\" class=\"graf graf--li graf-after--li\">What does an ideal team player look like on his team?<\/li><li id=\"39b4\" class=\"graf graf--li graf-after--li graf--trailing\">A transparent edit, commenting on his ability to be responsive.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"8e4a\" class=\"graf graf--h3 graf--leading\"><a href=\"http:\/\/www.mollywhite.net\/\" data-href=\"http:\/\/www.mollywhite.net\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Molly<\/a>, Tech Lead at <a href=\"https:\/\/hubspot.com\" data-href=\"https:\/\/hubspot.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">HubSpot<\/a> ~~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"d2fb\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/github.com\/molly\/manager-README\" data-href=\"https:\/\/github.com\/molly\/manager-README\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*kf3v83tAmWMl8cpe_BmjVg.png\" data-width=\"1988\" data-height=\"1078\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*kf3v83tAmWMl8cpe_BmjVg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/github.com\/molly\/manager-README\" data-href=\"https:\/\/github.com\/molly\/manager-README\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/github.com\/molly\/manager-README<\/a><\/figcaption><\/figure><h4 id=\"343f\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"84e2\" class=\"graf graf--li graf-after--h4\">It\u2019s on github! Super cool.<\/li><li id=\"d1d6\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--li\">\u201cI am accountable for the decisions the team make, even if I\u2019m not the one making them most of the time.\u201d<\/li><li id=\"21a0\" class=\"graf graf--li graf-after--li graf--trailing\">Specific call out for diversity and mental health.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"e470\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/matthewnewkirk.com\/\" data-href=\"https:\/\/matthewnewkirk.com\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Matt<\/a>, Sr. Eng. Manager at <a href=\"https:\/\/etsy.com\" data-href=\"https:\/\/etsy.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Etsy<\/a> ~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"5909\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" data-href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*CXHucUSFBXya7wVqntUPfg.png\" data-width=\"1434\" data-height=\"938\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*CXHucUSFBXya7wVqntUPfg.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" data-href=\"https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/matthewnewkirk.com\/2017\/09\/20\/share-your-manager-readme\/<\/a><\/figcaption><\/figure><h4 id=\"40a0\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"b34f\" class=\"graf graf--li graf-after--h4\">Guests can edit his calendar events by defaults!<\/li><li id=\"e811\" class=\"graf graf--li graf-after--li\">People &gt; Process (no matter the methodology). If agile\/scrum isn\u2019t working for the team, let\u2019s figure out what will.<\/li><li id=\"608d\" class=\"graf graf--li graf-after--li graf--trailing\">He\u2019s big on work-life balance\u2026 meaning you\u2019ll likely never hear from him on weekends unless it\u2019s a big deal!<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"e1b0\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/twitter.com\/scottsburns\" data-href=\"https:\/\/twitter.com\/scottsburns\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Scott<\/a>, Vp Eng. at <a href=\"https:\/\/stratasan.com\" data-href=\"https:\/\/stratasan.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Stratasan<\/a> ~~~~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"226a\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*d6UqXWgq3iQqjtjFDMiYAA.png\" data-width=\"1798\" data-height=\"1024\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*d6UqXWgq3iQqjtjFDMiYAA.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" data-href=\"https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/docs.google.com\/presentation\/d\/1PE2OmkVykdZYF2QzmJ-ZUHqBAxnnsbnMsd19z3qLWNI\/edit#slide=id.gcb9a0b074_1_0<\/a><\/figcaption><\/figure><h4 id=\"c04d\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"287e\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--h4\">\u201cI believe managers work for their direct reports.\u201d<\/li><li id=\"a107\" class=\"graf graf--li graf-after--li graf--trailing\">On Deep Work: \u201cThe maintenance of inboxes (email, Slack, social media, etc) is not what you\u2019re paid to do. Attention is a valuable resource\u201d Not true for all roles, but true for Scott\u2019s team.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"a509\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/github.com\/elliottcarlson\" data-href=\"https:\/\/github.com\/elliottcarlson\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Elliott<\/a>, Manager at <a href=\"https:\/\/renttherunway.com\" data-href=\"https:\/\/renttherunway.com\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">RentTheRunway<\/a> ~~~~~~~~~~~~~~<\/h3><figure id=\"3933\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" data-href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*CqfL5ujQmbHpZyWODh5irQ.png\" data-width=\"1630\" data-height=\"990\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*CqfL5ujQmbHpZyWODh5irQ.png\" alt=\"image\" \/><\/a><\/div><figcaption class=\"imageCaption\"><a href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" data-href=\"https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view\" class=\"markup--anchor markup--figure-anchor\" rel=\"nofollow noopener\" target=\"_blank\">https:\/\/drive.google.com\/file\/d\/1mxUR69VBVBGPy-rq13jsgsgr67HosZOx\/view<\/a><\/figcaption><\/figure><h4 id=\"a420\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"8f20\" class=\"graf graf--li graf-after--h4\">He treats the document as a \u201cpromise on how I will conduct myself as a manager\u201d and asks to be held accountable.<\/li><li id=\"2b14\" class=\"graf graf--li graf-after--li\">He\u2019s an internal thinker who is sometimes quieter in large groups.<\/li><li id=\"e84e\" class=\"graf graf--li graf-after--li graf--trailing\">He may have people pleaser tendencies that he\u2019s working on.<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"3868\" class=\"graf graf--h3 graf--leading\"><a href=\"http:\/\/angryweasel.com\/blog\/\" data-href=\"http:\/\/angryweasel.com\/blog\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Alan<\/a>, Director at <a href=\"https:\/\/unity3d.com\/\" data-href=\"https:\/\/unity3d.com\/\" class=\"markup--anchor markup--h3-anchor\" rel=\"noopener\" target=\"_blank\">Unity<\/a> ~~~~~~~~~~~~~~~~~~~~~~~~~<\/h3><figure id=\"fc20\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/github.com\/angryweasel\/managerreadme\" data-href=\"https:\/\/github.com\/angryweasel\/managerreadme\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*NMH3j5te1Mxpcr7XCLNwPw.png\" data-width=\"1988\" data-height=\"940\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*NMH3j5te1Mxpcr7XCLNwPw.png\" alt=\"image\" \/><\/a><\/div><\/figure><h4 id=\"e7ca\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"ebd1\" class=\"graf graf--li graf-after--h4\">Love that its on github\u200a\u2014\u200acheck the commits for edits!<\/li><li id=\"ae7d\" class=\"graf graf--li graf-after--li\">How his ICs can become leaders within their organization.<\/li><li id=\"bd86\" class=\"graf graf--li graf-after--li graf--trailing\">References to further reading on his management influences!<\/li><\/ul><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"e77e\" class=\"graf graf--h3 graf--leading\"><a href=\"https:\/\/medium.com\/@mikekleiman\" data-href=\"https:\/\/medium.com\/@mikekleiman\" data-anchor-type=\"2\" data-user-id=\"d0849441a93f\" data-action-value=\"d0849441a93f\" data-action=\"show-user-card\" data-action-type=\"hover\" class=\"markup--user markup--h3-user\" target=\"_blank\">Mike<\/a>, Manager of Eng. at BiblioCommons ~~~~~~~~~<\/h3><figure id=\"38a5\" class=\"graf graf--figure graf--layoutOutsetLeft graf-after--h3\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/medium.com\/@mikekleiman\/my-readme-md-35bd0197a5c5\" data-href=\"https:\/\/medium.com\/@mikekleiman\/my-readme-md-35bd0197a5c5\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*GqHG0Kxx4N1D7K6kq3oOiw.png\" data-width=\"1478\" data-height=\"966\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*GqHG0Kxx4N1D7K6kq3oOiw.png\" alt=\"image\" \/><\/a><\/div><\/figure><h4 id=\"debc\" class=\"graf graf--h4 graf-after--figure\">My favourite parts:<\/h4><ul class=\"postList\"><li id=\"72ad\" class=\"graf graf--li graf-after--h4\">Mike included a <strong class=\"markup--strong markup--li-strong\">TLDR<\/strong>; ahh what a refreshing summary.<\/li><li id=\"c164\" class=\"graf graf--li graf--startsWithDoubleQuote graf-after--li\">\u201cYou should expect me to have your back 100% of the time.\u201d Love it!<\/li><li id=\"40c0\" class=\"graf graf--li graf-after--li\">It\u2019s on medium! Now, that\u2019s transparent!<\/li><\/ul><p id=\"cc93\" class=\"graf graf--p graf-after--li graf--trailing\">.<\/p><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"afeb\" class=\"graf graf--h3 graf--leading\">Ready to write your\u00a0own?<\/h3><p id=\"4346\" class=\"graf graf--p graf-after--h3\">\ud83d\udca1\ud83d\udc49 <a href=\"https:\/\/soapboxhq.com\/blog\/modern-manager-community\/how-to-set-expectations-with-your-team\" data-href=\"https:\/\/soapboxhq.com\/blog\/modern-manager-community\/how-to-set-expectations-with-your-team\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\">We (SoapBox) just wrote an article on how!<\/a><\/p><p id=\"c81d\" class=\"graf graf--p graf-after--p\">However, almost everyone above included the following:<\/p><ul class=\"postList\"><li id=\"ce3b\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">What is this? <\/strong>\u2014Let\u2019s face it, this is a weird document. Give context on what this is and how to interpret it: to supplement getting to know each other.<\/li><li id=\"1dce\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">About Me \/ My Job\u200a<\/strong>\u2014\u200aWhat are some common things people might like to know about you. Some people\u2019s are more personal than others. If you want to get to know your team\u2019s personal lives then start off by sharing yours.<\/li><li id=\"a630\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Personal Principles \/ Values<\/strong>\u200a\u2014\u200aWhat are your default assumptions about people and their intentions? What mindsets do you have and which do you hope other people adopt when working together on the team? What things trigger you?<\/li><li id=\"f374\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">One-on-Ones<\/strong>\u200a\u2014\u200aWhat style of 1:1s do you want out of your team? Most above follow the weekly\/bi-weekly, 30 min cadence, and the employee controls the agenda.<\/li><li id=\"8a79\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">Feedback (how to give &amp; how you give)<\/strong>\u200a\u2014\u200aWhat type of feedback do you want? Are you comfortable with people being blunt with you? How do you prefer to give feedback? How do you expect your team to react to feedback?<\/li><li id=\"e150\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">How to interpret my calendar<\/strong>\u200a\u2014\u200aSometimes a manager\u2019s calendar can be packed. Almost every manager above wants their team to know that they are the most important part of their job, so they\u2019re explicit in saying so. Message them on slack if you need to talk. They\u2019ll make time.<\/li><\/ul><p id=\"24e9\" class=\"graf graf--p graf-after--li\">Many also include the following:<\/p><ul class=\"postList\"><li id=\"8ef2\" class=\"graf graf--li graf-after--p\"><strong class=\"markup--strong markup--li-strong\">Personal performance scales<\/strong>\u200a\u2014\u200aIf an employee asks, \u201cWhere do I stand?\u201d How will you respond? Most seem to do Red\/Orange\/Green. While Green and Red are clear indicators, orange can open to interpretation. What does that mean and how do you intend for people to react?<\/li><li id=\"cf12\" class=\"graf graf--li graf-after--li\"><strong class=\"markup--strong markup--li-strong\">D.R.I. Principle<\/strong>\u200a\u2014\u200aNot everyone is a believer in <a href=\"https:\/\/www.forbes.com\/sites\/quora\/2012\/10\/02\/how-well-does-apples-directly-responsible-individual-dri-model-work-in-practice\/#1c3bdb35194c\" data-href=\"https:\/\/www.forbes.com\/sites\/quora\/2012\/10\/02\/how-well-does-apples-directly-responsible-individual-dri-model-work-in-practice\/#1c3bdb35194c\" class=\"markup--anchor markup--li-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--li-strong\">D<\/strong>irectly <strong class=\"markup--strong markup--li-strong\">R<\/strong>esponsible <strong class=\"markup--strong markup--li-strong\">I<\/strong>ndividual<\/a> principle, but if you are it can be very powerful. First, though, you need to set the expectation that they take that responsibly.<\/li><\/ul><h3 id=\"d2f1\" class=\"graf graf--h3 graf-after--li\">Done? Ship\u00a0it!<\/h3><p id=\"324a\" class=\"graf graf--p graf-after--h3\">How? Consensus from the authors I\u2019ve talked to seems to be:<\/p><ol class=\"postList\"><li id=\"ec34\" class=\"graf graf--li graf-after--p\">The most effective way to deliver the document is to take employees through it in a 1:1.<\/li><li id=\"af88\" class=\"graf graf--li graf-after--li graf--trailing\">Easiest time to start is during an employee\u2019s on-boarding. Their on-boarding or your own (if you\u2019re inheriting a team)<\/li><\/ol><\/div><\/div><\/section><section class=\"section section--body\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h3 id=\"aed4\" class=\"graf graf--h3 graf--leading\">You made it to the\u00a0end!<\/h3><blockquote id=\"ff83\" class=\"graf graf--blockquote graf-after--h3\"><div><strong class=\"markup--strong markup--blockquote-strong\">Here\u2019s my question for you: <\/strong>Anything that I missed in my notes? Anything that they missed in their documents? What are a must-haves in your doc?<\/div><\/blockquote><p id=\"38ce\" class=\"graf graf--p graf-after--blockquote\">Thanks for reading,<br \/>Brennan<\/p><p id=\"2dd6\" class=\"graf graf--p graf-after--p\">p.s. You should be having these conversations with your team at least bi-weekly. <strong class=\"markup--strong markup--p-strong\">If your conversations are stale, inconsistent, or unorganized, you should try out <\/strong><a href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" data-href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" class=\"markup--anchor markup--p-anchor\" rel=\"noopener\" target=\"_blank\"><strong class=\"markup--strong markup--p-strong\">SoapBox<\/strong><\/a><strong class=\"markup--strong markup--p-strong\"> now.<\/strong><\/p><figure id=\"a244\" class=\"graf graf--figure graf-after--p\"><div class=\"aspectRatioPlaceholder is-locked\"><img class=\"graf-image\" data-image-id=\"1*af8A62nrzIjp6gioRz3rrg.png\" data-width=\"1186\" data-height=\"414\" data-action=\"zoom\" data-action-value=\"1*af8A62nrzIjp6gioRz3rrg.png\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*af8A62nrzIjp6gioRz3rrg.png\" alt=\"image\" \/><\/div><\/figure><h4 id=\"9d79\" class=\"graf graf--h4 graf-after--figure\">Brennan is the CEO &amp; Co-founder of <a href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" data-href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" class=\"markup--anchor markup--h4-anchor\" rel=\"noopener\" target=\"_blank\">SoapBox<\/a>, the #1 place to work in Canada. <a href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" data-href=\"https:\/\/soapboxhq.com\/?utm_source=hackernoon&amp;utm_medium=referral&amp;utm_campaign=manager-readmes\" class=\"markup--anchor markup--h4-anchor\" rel=\"noopener\" target=\"_blank\">SoapBox<\/a>, is an app and assistant for managers to have better one-on-ones, team meetings, town-halls and more, with their\u00a0team.<\/h4><figure id=\"0bfb\" class=\"graf graf--figure graf-after--h4 graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"https:\/\/twitter.com\/i_am_brennan\" data-href=\"https:\/\/twitter.com\/i_am_brennan\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*-36b5jCwLYg0MVrPP0kXKg.png\" data-width=\"1408\" data-height=\"112\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*-36b5jCwLYg0MVrPP0kXKg.png\" alt=\"image\" \/><\/a><\/div><\/figure><\/div><\/div><\/section><section class=\"section section--body section--last\"><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><p id=\"579c\" class=\"graf graf--p graf--leading\">If you liked this article, you should give it <strong class=\"markup--strong markup--p-strong\">12\ud83d\udc4f<\/strong>\u2019s (one for each readme) to help others find it!\ud83d\udc47<\/p><figure id=\"c90a\" class=\"graf graf--figure graf-after--p graf--trailing\"><div class=\"aspectRatioPlaceholder is-locked\"><a href=\"http:\/\/bit.ly\/2LLdEDb\" data-href=\"http:\/\/bit.ly\/2LLdEDb\" class=\"graf-imageAnchor\" data-action=\"image-link\" data-action-observe-only=\"true\"><img class=\"graf-image\" data-image-id=\"1*Xv3XV-DgnRK9WjRs90xGMA@2x.png\" data-width=\"1400\" data-height=\"774\" src=\"https:\/\/cdn-images-1.medium.com\/max\/1600\/1*Xv3XV-DgnRK9WjRs90xGMA@2x.png\" alt=\"image\" \/><\/a><\/div><\/figure><\/div><\/div><\/section>","created_at":"2018-11-07T16:23:50+0000","updated_at":"2018-11-07T16:23:50+0000","published_at":"2018-05-02T13:36:41+0000","published_by":["Brennan McEachran \ud83d\udc68\u200d\ud83d\ude80"],"annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"hackernoon.com","preview_picture":"https:\/\/cdn-images-1.medium.com\/max\/1200\/1*JPGxBj3qjqgKctsp9oLqmQ.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":169,"title":"Never Write Your Own Database","url":"https:\/\/medium.com\/@terrycrowley\/never-write-your-own-database-736f704c780","content":"<div class=\"section-divider\"><hr class=\"section-divider\" \/><\/div><div class=\"section-content\"><div class=\"section-inner sectionLayout--insetColumn\"><h1 id=\"0609\" class=\"graf graf--h3 graf--leading graf--title\">Never Write Your Own\u00a0Database<\/h1><p id=\"7da1\" class=\"graf graf--p graf-after--h3\"><em class=\"markup--em markup--p-em\">Note: I originally authored this post while at Microsoft with Vincent Lascaux (the engineer primarily responsible for the work in OneNote described here) and decided to republish in this forum since that version no longer seems to be available. Thanks Vincent for all the detailed information and feedback!<\/em><\/p><p id=\"2b51\" class=\"graf graf--p graf-after--p\">There\u2019s an old adage in software engineering\u200a\u2014\u200a\u201cnever write your own database\u201d. So why did the OneNote team go ahead and write one as part of upgrading OneNote\u2019s local cache implementation?<\/p><p id=\"7603\" class=\"graf graf--p graf-after--p\">First, it\u2019s probably necessary to more carefully define what I mean by a \u201cdatabase\u201d in this context.<\/p><p id=\"15b9\" class=\"graf graf--p graf-after--p\">Most interesting applications need to be able to persist their application state and the user content they create and manage to durable storage. An application can either use its own custom logic for persistence or make use of some pre-existing database management software (DBMS) to handle storing its data. That pre-existing DBMS might be some kind of relational store (e.g. SQLLite, ESENT) or a simpler key-value store (or frequently, a relational store used as a key-value store).<\/p><p id=\"0041\" class=\"graf graf--p graf-after--p\">If you start exploring these technologies, it doesn\u2019t require much time browsing the Wikipedia page for <a href=\"http:\/\/en.wikipedia.org\/wiki\/Extensible_Storage_Engine\" data-href=\"http:\/\/en.wikipedia.org\/wiki\/Extensible_Storage_Engine\" class=\"markup--anchor markup--p-anchor\" rel=\"nofollow noopener\" target=\"_blank\">ESENT<\/a> (for example) to get a sense for some of the complexity involved in building a database. While reading through that description, you can almost do an archeological analysis of what were the initial core capabilities and what were additional capabilities developed over time, clearly driven from evolving client requirements (clustered indexes, sparse indexes, covering indexes, versioned columns, etc. etc.). It gives the clear sense that this is a component that has stood the test of time and served as the basis of some key products and how they evolved in the market place.<\/p><p id=\"f642\" class=\"graf graf--p graf-after--p\">There are two rather counter-balancing reactions when evaluating whether to use such a robust and functional component. The first reaction is \u201choly crap, there is no way that I\u2019m smart enough to duplicate that tech\u201d! That\u2019s a good and righteous reaction! \u201cIf I see so far, it is only because I stand on the shoulders of giants\u2026\u201d<\/p><p id=\"a52f\" class=\"graf graf--p graf-after--p\">The other reaction is: \u201cWow, there\u2019s a lot of stuff there that I don\u2019t need. I wonder what I\u2019m paying for that?\u201d One of the key principles of engineering is TANSTAAFL\u200a\u2014\u200a\u201cthere ain\u2019t no such thing as a free lunch\u201d. Ultimately, when you look under the covers, that extra capability is implemented with extra code, IO, memory and CPU, just as it would be if you implemented it on your own. Sometimes (rarely!) it\u2019s completely pay-for-play and you only pay the cost when you use those extra features. Other times the component\u2019s implementers have built in certain assumptions and trade-offs and those features are likely to constrain future evolution or place constraints on how the component is integrated into a larger system. It is the complexity of analyzing this, even when you understand the core design characteristics of the component, that often leads to the practice of building a full functional prototype to validate your specific performance requirements under your own simulated (or actual) load patterns.<\/p><p id=\"0b1c\" class=\"graf graf--p graf-after--p\">Let\u2019s get back to the problem of persisting application state.<\/p><p id=\"85c7\" class=\"graf graf--p graf-after--p\">Some applications operate over storage in a relatively simple manner. A file on disk is completely read into data structures in memory, modified, and then the entire file is written back to disk. A simple trick of writing the new file under a separate name and then renaming the new file to replace the old file only after it has been successfully written provides safety against storage corruption.<\/p><p id=\"1c06\" class=\"graf graf--p graf-after--p\">The next approach is to represent the application\u2019s state using multiple files on disk; the step up in complexity is surprisingly steep. An important challenge is that generally there are interdependencies between the files (e.g. an index of some kind) and file systems do not generally provide any transactional guarantees (or the tradeoffs in using the support they do provide is too great). In the face of failures (guaranteed to occur at scale) the state of the different files can become inconsistent. This means the application must build in techniques for recognizing and repairing this inconsistent \u201ccorrupted\u201d state. The very fact that the files are independently accessible to other applications and services through the file system also makes this kind of inconsistency a state that needs to be designed for. However, the interaction with any single file can be still relatively simple\u200a\u2014\u200ajust open and rewrite it in toto. In aggregate, there may be other challenges depending on the scale at which the file system is leveraged (e.g. if 10s, 100s, 1000s of files need to be managed and potentially split across directories to optimize file system performance characteristics that scale with number of files per directory). This approach of using independent files can work well when the files are independently consistent and where the typical file size is large enough that storage space efficiency is not a special concern (due to fragmentation or wasted space due to file system metadata and page size). An example use that meets those requirements is storing email attachments as independent files on disk rather than within a single mail cache file. Storing large independent blobs is a problem file systems have been tuned for, so this offloading can narrow the scope of design points a single file database needs to be targeted at.<\/p><p id=\"963b\" class=\"graf graf--p graf-after--p\">The key differentiating approach in a database as I am using the term here is that the single file is only partially read into memory and only incrementally rewritten. For example, OneNote provides a virtualized experience over a potentially very large space (10s of GBs) of notebooks. They cannot be fully loaded into memory and yet the application would like to provide the illusion that all the data is available for browsing and editing at any time. The application would like to boot to a working state quickly (that is, only do a small amount of IO on startup). Modifications typically only involve a small subset of the data and the application would like to write those changes to disk quickly and at low cost. As I browse over new parts of the data model, I would like to quickly fault that data in (e.g. navigate to a new section in OneNote). An additional requirement is that while most writes are small and incremental, the special case of syncing new content needs to be very fast and efficient for bulk transfers.<\/p><p id=\"da67\" class=\"graf graf--p graf-after--p\">Even with this rather simple description of requirements, we already have somewhat conflicting constraints (which is the nature of engineering!). We would like the data packed together so we can reduce total number and size of IOs, but we would like to avoid having to repack it all together when saving back changes to only a small subset of that data. If we only write the changed data our data ends up getting more fragmented over time while if we always repack we end up doing more IO\u200a\u2014\u200aessentially having to rewrite data that hasn\u2019t changed in order to get better packing. One fundamental change in approach that we\u2019ve seen in the move away from desktop AC-powered machines to battery powered (that is, power constrained) mobile devices is moving away from the assumption that we can handle some of these tradeoffs through sporadic high-cost maintenance processes (e.g. running a whole file compaction or defragmentation process \u201cat 3AM\u201d). Instead, steady-state performance is the focus and costs need to be amortized over time. It was already the case even before the rise of smartphones that our assumptions about our ability to run code at these times was suspect (battery-constrained laptops became dominant starting in the mid-2000\u2019s; \u201cgreen\u201d AC-powered machines are automatically shut down at night; as disk capacities exploded but IO per second did not, just rewriting all the data could take so long that it could not complete in any maintenance window that was available).<\/p><p id=\"c66b\" class=\"graf graf--p graf-after--p\">We also need to think about how to implement the core database ACID properties (Atomic, Consistent, Isolated, Durable). While there is a long technical history on approaches and techniques here, they are generally embedded in specific implementations rather than available as a simple set of APIs to call. Trying to roll your means joining a mysterious cabal of wizards familiar with file system arcana including things like when is a flush really a flush, what bizarre things virus checkers can do, where and when IO requests can be reordered, how backup systems really work, how sync engines work, how this varies across OS type, etc.<\/p><p id=\"01b6\" class=\"graf graf--p graf-after--p\">The advantage of rolling your own is that you get to optimize and control things at a level that is just not possible when working through a complex external component. In some sense this is the eternal component dependency problem. When does an external component solve a hard problem better than you could yourself? Determining \u201cbetter\u201d is frequently a hard question that includes aspects like performance, functionality and development time\/cost as well as a significant element of \u201cbetter over time\u201d. Over time your own requirements will change, your view of the importance to your product of optimizing this layer will change, the underlying technology landscape will change, and the component itself will change (or frequently, <strong class=\"markup--strong markup--p-strong\">not<\/strong> change, which is problematic when layered on top of that changing technology landscape).<\/p><p id=\"88b8\" class=\"graf graf--p graf-after--p\">OK, that was all just background. Let\u2019s dive into the details of the OneNote work.<\/p><p id=\"8bb2\" class=\"graf graf--p graf-after--p\">OneNote has been morphing for some time into a purely cloud-connected application with data stored in the cloud rather than operating on files stored on local disk. This work was focused on optimizing the OneNote \u201ccache\u201d\u200a\u2014\u200athe set of files that store a cached copy of cloud notebooks locally to support offline usage and fast and responsive read and write. The original OneNote persistence format stored each section in a separate file (where each section contains a set of pages). A notebook was represented as a folder hierarchy containing section files and other nested folders. A table of contents file stored certain metadata about the folders themselves. OneNote\u2019s original cache format merged all cached notebooks into a single file on disk. A section file is a \u201cdatabase\u201d in the meaning we are using it here (it was designed to be partially read and incrementally rewritten) and the cache file built on top of the core file and object management primitives that were used for reading and writing section files.<\/p><p id=\"81d9\" class=\"graf graf--p graf-after--p\">The cache rewrite effort had some clear goals and one significant constraint. As I talked about above, OneNote\u2019s cache file was designed to have periodic large-scale cleanup processes run that are not appropriate in a more mobile resource-constrained world. So an important goal was eliminating any need for periodic maintenance so that in practical usage the cache would use less storage and require less memory and CPU (since for many devices that maintenance never actually succeeded in running). Another goal was to address a long-time inefficiency related to OneNote\u2019s revision-based storage model. OneNote supports a \u201csave-less\u201d model where all changes are automatically persisted to storage. In order to do this efficiently, it appends a new revision to the page that only contains the changed content. For heavily edited pages (a small but highly valuable subset), these revision chains could get long, resulting in pages that were slower to load and required more memory and CPU. An important goal was to ensure that the cost of loading a page was in line with the actual complexity of the page presented to the user rather than proportional to its editing history.<\/p><p id=\"b9c5\" class=\"graf graf--p graf-after--p\">The significant constraint was that the work should be limited to the storage layer itself and not require a massive rewrite of the overall application. This ended up driving many of the requirements. When dealing with large, complex systems, a key problem is how to evolve the architecture without just throwing everything up in the air and hoping it lands on time. That approach is a thing of the past in our world of continuous engineering and delivery. In fact, even in the past it was problematic since it deferred many of the trickiest and time-consuming integration tasks and was a source of many a late project.<\/p><p id=\"d910\" class=\"graf graf--p graf-after--p\">In OneNote\u2019s case, this incremental approach drove significant requirements:<\/p><ul class=\"postList\"><li id=\"7461\" class=\"graf graf--li graf-after--p\">The new system should equal or improve on the performance (memory, CPU, storage) of the existing solution (that was really the whole point of the project).<\/li><li id=\"8880\" class=\"graf graf--li graf-after--li\">It needed to support the ability to snapshot pages (while continuing to allow creation of new page versions concurrently). The sync infrastructure, as well as versioning, indexing and other features rely on the ability to cheaply create snapshots of pages so that these features can operate over a page at the same time it might be being edited by the user. Storing multiple versions of a graph that contains dozens to tens of thousands of nodes in an efficient manner turned out to be a tricky endeavor. This requirement is a good example where a complex system comes to depend on some key operation being highly efficient. As a system evolves, maintaining these implicit performance guarantees is hugely important. Word has a similar requirement where it requires that \u201ccloning\u201d a document is very cheap. It is a non-trivial constraint for new features to maintain that characteristic.<\/li><li id=\"3f5d\" class=\"graf graf--li graf-after--li\">Efficient computation of changes between versions. The sync engine relies on fast change computation to upload user changes to the server.<\/li><li id=\"37db\" class=\"graf graf--li graf-after--li\">Multiple concurrent writers. Sync happens concurrently with editing and requires the ability for multiple writers of the same page.<\/li><li id=\"0de5\" class=\"graf graf--li graf-after--li\">Other subtle odd requirements. The existing code required that data could go unreferenced for a short period of time without being deleted. It also required that embedded files be exposed as real files while still maintaining ACID contracts. In practice, one of the main challenges of re-architecting a complex component or layer in a software system is even recognizing these subtle requirements since they are often implicit in the way the component is used rather than an explicit part of the API contract.<\/li><\/ul><p id=\"b3a8\" class=\"graf graf--p graf-after--li\">The team looked carefully at existing database solutions (remember that old adage about never writing your own database?). Ultimately it was really the list of integration requirements above that drove them to write their own database rather than results of an analysis around raw performance or robustness. These kinds of evaluations are notoriously tricky, since you can really fail in both directions\u200a\u2014\u200ayou can underestimate the cost of stabilizing your own low level code in the wild world of real machines and real usage patterns or you can underestimate the cost of delivering your full solution on top of an existing complex component that has its own design point and constraints, especially if you are not starting from scratch but rather are trying to match those component constraints to the existing constraints of your own application. The fact that the team was building on their established expertise in this space gave them confidence that they had a good understanding of where the bear traps were in building their own.<\/p><p id=\"ae82\" class=\"graf graf--p graf-after--p\">One of the earliest decisions they made was to split the cache into a separate file per page (with a single master index file). Garbage collecting space allocated to deleted pages and deleted content or closed notebooks had been a major issue with the single file cache design. By splitting out pages, this garbage collection problem was offloaded to the file system, which is highly tuned to deal with it. Additionally, the typical size of a single page is relatively small (although in line with sizes that the file system is tuned for) so that opened opportunities for other simplifications. There were other benefits\u200a\u2014\u200ait also avoided problems that had arisen with cache files that were exceptionally large (&gt; 4GB). Fortuitously, it also made integrating with Enterprise Data Protection APIs (a new Windows 10 feature) much simpler since each page could be separately encrypted according to the appropriate polices for that content. I said \u201cfortuitously\u201d but in reality this is a good example where having a low-level design that is more in line with the user\u2019s high level model of their content ends up avoiding internal implementation complexity that doesn\u2019t align well with value delivered. This is the kind of \u201caccidental complexity\u201d that can slow down development over time in a large code base.<\/p><p id=\"9c8d\" class=\"graf graf--p graf-after--p\">As alluded to above, this does introduce a separate challenge with multi-file consistency. The team took an interesting approach here. The index file points to the location of the root of the page graph and free space map in each page cache file. When persisting a multi-file transaction, all page cache files are flushed to disk before the index file. Each page file now has <strong class=\"markup--strong markup--p-strong\">two<\/strong> graph roots (and their related free space map) the old graph and the new graph (which typically share many children). Both are fully consistent and available in the file, depending on where you start looking. The index file is then flushed to disk to point to the new graph roots and new free maps. Transactional consistency for creation and deletion of pages requires an additional small twist. A data structure in the index file (flushed to disk as part of the transaction) describes the files to delete and the files created. New files are created in a separate folder and then moved to the main cache folder after the transaction completes. On restart, these data structures are processed to ensure any actions that didn\u2019t fully complete at last run are finalized. Deletion is idempotent (deleting multiple times results in the same result) so is fairly straightforward. Finalizing the creation transaction is also idempotent (any files no longer in the sub-folder have already been successfully moved to the main folder) and files present in the subfolder but not referenced in the creation list are related to an unfinished new page transaction and can be safely deleted (or made available as part of a \u201clost pages\u201d recovery feature).<\/p><p id=\"6d65\" class=\"graf graf--p graf-after--p\">To address the problem of long revision chains, they made a fundamental change in the page-level data structures. Pages are still defined as revision chains, but each revision now stores the lookup table (now a BTree) for all the objects the revision references. The old version only referenced changed objects in the revision and needing to walk down the revision chain explicitly to find any objects that were not changed in that revision.<\/p><p id=\"1019\" class=\"graf graf--p graf-after--p\">The effect is that the cost of loading a version of a page (especially the current version) now reflects the complexity of the page, rather than its editing history. This change does make the computation of changes between revisions different. As noted above, this is one of the operations that much of the system assumes is fast. The team spent quite a bit of time redesigning the differencing algorithm (including building out an extensive set of unit and component tests) to ensure that it had both the reliability and the performance characteristics that the rest of the system required, especially over a broad set of use cases. This is another good example of investing a lot of effort (and surrounding engineering infrastructure to ensure the property is maintained over time as the product evolves) in one particular feature\u200a\u2014\u200acompute the difference between two page versions\u200a\u2014\u200aso that the rest of the system can be written to just assume it is reliable and fast. In fact, by writing the differencing algorithm as a general capability to compare any two revisions rather than assuming a chaining relationship, it opened up the feature for other uses (e.g. user-level compare of page versions that were not related by simple revision chaining but more complex branching).<\/p><p id=\"ab3e\" class=\"graf graf--p graf-after--p\">Fast page snapshot was another feature required of the new implementation. The transactional consistency mechanism that required having multiple instances of the page graph be valid at the same time also could support a robust snapshot mechanism (simply by continuing to hold on to both instances of modified page graphs during the scope of the transaction). By building on the whole cache transaction mechanism it also expanded the capability of the snapshot from a single page in the old system to allowing snapshots of the entire notebook. That is not a mechanism that is currently being used but there are multiple features in early stages of planning that are considering making use of this more fully functional capability.<\/p><p id=\"5537\" class=\"graf graf--p graf-after--p\">Support for multiple writers allows both transactions to proceed independently but then are merged into a single transaction at commit time. This depends on leveraging OneNote\u2019s robust merge capability, including the ability to represent conflicting edits directly in the content model. There are a few cases where the content model cannot be merged and in that case the transaction is backed out. Combining write transactions does have the consequence that a long-running transaction could delay other transactions from committing so care needs to be taken about how this mechanism is used by higher level processes (like sync).<\/p><p id=\"72fe\" class=\"graf graf--p graf-after--p\">One of the key approaches they took to minimizing memory was to use memory mapping for the page cache file and minimize the number of additional data structures that needed to be allocated and maintained in order to represent the page in memory. So this storage layer both requires less memory and less CPU since the storage data structures can be directly referenced without additional allocations and data copying. OneNote typically then builds an in-memory graph from the storage structures for ease of overall manipulation. The team is now looking at more opportunities where certain processing (especially bulk operations like sync and indexing with very specific functional requirements) can be implemented directly on the storage layers rather than requiring a graph to be \u201chydrated\u201d. Experience shows that optimizations like this can be surprisingly effective.<\/p><p id=\"2f64\" class=\"graf graf--p graf-after--p graf--trailing\">In the longer term, these changes around flexible threading models and highly efficient processing models sets OneNote up for building on their investment in owning the low-level storage layers. For the OneNote team, it seems clear that at least in this case, that old adage proved false.<\/p><\/div><\/div>","created_at":"2018-11-07T16:23:50+0000","updated_at":"2018-11-07T16:23:50+0000","published_at":"2017-02-13T16:39:52+0000","published_by":["Terry Crowley"],"annotations":[],"mimetype":"text\/html","reading_time":18,"domain_name":"medium.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":167,"title":"Ask HN: What problem in your industry is a potential startup?","url":"https:\/\/news.ycombinator.com\/item?id=16995260","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16995970\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=franciscop\" class=\"hnuser\">franciscop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996303\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mikepurvis\" class=\"hnuser\">mikepurvis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996427\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996457\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=atmosx\" class=\"hnuser\">atmosx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000902\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merinowool\" class=\"hnuser\">merinowool<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004562\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BjoernKW\" class=\"hnuser\">BjoernKW<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005882\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merinowool\" class=\"hnuser\">merinowool<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007424\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vimal7370\" class=\"hnuser\">vimal7370<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007610\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=awmartin\" class=\"hnuser\">awmartin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999207\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=iopeak\" class=\"hnuser\">iopeak<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000698\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mon1key\" class=\"hnuser\">mon1key<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000019\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w_t_payne\" class=\"hnuser\">w_t_payne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001296\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mon1key\" class=\"hnuser\">mon1key<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003851\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w_t_payne\" class=\"hnuser\">w_t_payne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008077\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ravi096\" class=\"hnuser\">ravi096<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17017642\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ThomPete\" class=\"hnuser\">ThomPete<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005998\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nathan_f77\" class=\"hnuser\">nathan_f77<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998145\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwilliams\" class=\"hnuser\">jwilliams<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000084\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=herlitzj\" class=\"hnuser\">herlitzj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996369\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996486\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=biswajitsharma\" class=\"hnuser\">biswajitsharma<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007245\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daxfohl\" class=\"hnuser\">daxfohl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997448\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hashkb\" class=\"hnuser\">hashkb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996050\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996127\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hashkb\" class=\"hnuser\">hashkb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996004\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996027\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998477\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hashkb\" class=\"hnuser\">hashkb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996038\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996067\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=orf\" class=\"hnuser\">orf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996123\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lwansbrough\" class=\"hnuser\">lwansbrough<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996336\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996551\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dchuk\" class=\"hnuser\">dchuk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999410\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w_t_payne\" class=\"hnuser\">w_t_payne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001300\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kowdermeister\" class=\"hnuser\">kowdermeister<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996063\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996146\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kowdermeister\" class=\"hnuser\">kowdermeister<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997248\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jononor\" class=\"hnuser\">jononor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001862\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j45\" class=\"hnuser\">j45<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006822\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maerF0x0\" class=\"hnuser\">maerF0x0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996467\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996532\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w_t_payne\" class=\"hnuser\">w_t_payne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001289\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=beamatronic\" class=\"hnuser\">beamatronic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996133\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kodablah\" class=\"hnuser\">kodablah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996176\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dyim\" class=\"hnuser\">dyim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997014\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wicha\" class=\"hnuser\">wicha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998354\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mlevental\" class=\"hnuser\">mlevental<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000059\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aerovistae\" class=\"hnuser\">aerovistae<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001043\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dabockster\" class=\"hnuser\">dabockster<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17010141\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jackgolding\" class=\"hnuser\">jackgolding<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17010792\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmode\" class=\"hnuser\">dmode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996367\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=idax\" class=\"hnuser\">idax<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996954\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=herbst\" class=\"hnuser\">herbst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006350\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thriptic\" class=\"hnuser\">Thriptic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997039\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Balgair\" class=\"hnuser\">Balgair<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002915\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w_t_payne\" class=\"hnuser\">w_t_payne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17025798\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yourpalkeith\" class=\"hnuser\">yourpalkeith<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999436\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nelsontkq\" class=\"hnuser\">nelsontkq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999242\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thriptic\" class=\"hnuser\">Thriptic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999333\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w_t_payne\" class=\"hnuser\">w_t_payne<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001474\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jgamman\" class=\"hnuser\">jgamman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999225\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thriptic\" class=\"hnuser\">Thriptic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999328\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=akhilcacharya\" class=\"hnuser\">akhilcacharya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000617\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdiq\" class=\"hnuser\">sdiq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996136\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=adz_6891\" class=\"hnuser\">adz_6891<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006137\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noobhacker\" class=\"hnuser\">noobhacker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999529\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=godelski\" class=\"hnuser\">godelski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996705\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdiq\" class=\"hnuser\">sdiq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997172\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=herlitzj\" class=\"hnuser\">herlitzj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996423\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=riezebos\" class=\"hnuser\">riezebos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002665\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdiq\" class=\"hnuser\">sdiq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997149\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kathe-rine\" class=\"hnuser\">kathe-rine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003200\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Immortalin\" class=\"hnuser\">Immortalin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996788\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sdiq\" class=\"hnuser\">sdiq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997184\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rayiner\" class=\"hnuser\">rayiner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996002\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simon_acca\" class=\"hnuser\">simon_acca<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998403\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SOLAR_FIELDS\" class=\"hnuser\">SOLAR_FIELDS<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001569\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jokh\" class=\"hnuser\">jokh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003120\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=iamdave\" class=\"hnuser\">iamdave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996494\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mabbo\" class=\"hnuser\">mabbo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996899\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=akhilcacharya\" class=\"hnuser\">akhilcacharya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000609\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mullen\" class=\"hnuser\">mullen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999446\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rb808\" class=\"hnuser\">rb808<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997189\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KillerRabbitt\" class=\"hnuser\">KillerRabbitt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006469\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cakefactory\" class=\"hnuser\">cakefactory<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997518\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tixocloud\" class=\"hnuser\">tixocloud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996075\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=realPubkey\" class=\"hnuser\">realPubkey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996284\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Joeri\" class=\"hnuser\">Joeri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996666\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=robbyt\" class=\"hnuser\">robbyt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000683\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jcfrei\" class=\"hnuser\">jcfrei<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996716\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Semkas\" class=\"hnuser\">Semkas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996228\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eerikkivistik\" class=\"hnuser\">eerikkivistik<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005949\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seige\" class=\"hnuser\">seige<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996443\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=billconan\" class=\"hnuser\">billconan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999159\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kichuku\" class=\"hnuser\">kichuku<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996862\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=irundebian\" class=\"hnuser\">irundebian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008382\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=screye\" class=\"hnuser\">screye<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004870\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kakwa_\" class=\"hnuser\">kakwa_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998536\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Something1234\" class=\"hnuser\">Something1234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004998\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eksemplar\" class=\"hnuser\">eksemplar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996069\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jakebrereton\" class=\"hnuser\">jakebrereton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996998\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Joeri\" class=\"hnuser\">Joeri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996811\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bojo\" class=\"hnuser\">bojo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999559\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=beckler\" class=\"hnuser\">beckler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000243\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AutoEngineer\" class=\"hnuser\">AutoEngineer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997292\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=praeconium\" class=\"hnuser\">praeconium<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998121\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=screye\" class=\"hnuser\">screye<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004879\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AskewEgret\" class=\"hnuser\">AskewEgret<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997650\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AutoEngineer\" class=\"hnuser\">AutoEngineer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997910\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ilikeatari\" class=\"hnuser\">ilikeatari<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002061\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ILikeConemowk\" class=\"hnuser\">ILikeConemowk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000231\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rb808\" class=\"hnuser\">rb808<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997212\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=csomar\" class=\"hnuser\">csomar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001104\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gleenn\" class=\"hnuser\">gleenn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997579\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=warriormonk5\" class=\"hnuser\">warriormonk5<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001898\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rb808\" class=\"hnuser\">rb808<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002149\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=herbst\" class=\"hnuser\">herbst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006367\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hunter23\" class=\"hnuser\">hunter23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996106\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=twunde\" class=\"hnuser\">twunde<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996769\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hunter23\" class=\"hnuser\">hunter23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996869\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Regardsyjc\" class=\"hnuser\">Regardsyjc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999631\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hunter23\" class=\"hnuser\">hunter23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17009640\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=slow_donkey\" class=\"hnuser\">slow_donkey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996186\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hunter23\" class=\"hnuser\">hunter23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996691\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cvaidya1986\" class=\"hnuser\">cvaidya1986<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996203\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hunter23\" class=\"hnuser\">hunter23<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996704\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cvaidya1986\" class=\"hnuser\">cvaidya1986<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17068151\">5 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Thriptic\" class=\"hnuser\">Thriptic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996923\">6 months ago<\/a><\/p><br \/><div class=\"comment\">&gt; Helping patients select the right doctor. Currently most people use Yelp or through referrals. The problem is that Yelp has little correlation to quality of care. It's very difficult for patients to evaluate a doctor - usually what they end up doing is evaluating the customer service aspect of the doctor (did they speak to me nicely, did they make me wait for an appointment, etc.) but no one is able to evaluate doctors based on the quality of care.<p>We actually explored this for a startup idea. The problem is that it is difficult to find a group of people who can do the evaluations in a truthful and holistic way:<\/p><p>* Hospitals will never want to give out outcome data because outcome data will be used against them for ratings by people who don't understand it (for example, some community hospital in Montana may be rated higher than Mass General because of case complexity issues). Or worse, it will be used by people who DO understand it :D<\/p><p>* We explored having doctors rate other doctors in a variety of ways (which I think would reflect the \"truest\" measure of quality). Residents and fellows could rate attendings, but they might not know how attendings in their hospital compare to attendings in most other hospitals. Additionally, attendings or hospitals might apply pressure to these groups to provide good ratings. Specialists could rate other specialists in their field, but then you might see collusion, false negative reviews, or retaliation. How you would avoid these problems is not immediately clear to me.<\/p><p>* As you point out, patients are really only able to evaluate bedside manner and not quality of care.<\/p><p>One way we thought about it is having a rating system which would have public profiles for physicians and anonynmous reviews from other physicians and members of the care team. Ratings coming from other physicians in the specialty and providers at their institutions would be weighted more than ratings from other physicians. The highest rated physicians would also have more weight within their specialty than an average rated physician. You could bootstrap the system by asking specialists to provide the names of the top X people in their field. These people would automatically be rated highly.<\/p><p>Patients could log in and provide comments about patient experience; hospitals could log in and provide outcome data if they wanted.<\/p><p>I am not entirely sure how you would really monetize it. It's the equivalent of the dating app problem; the better you are at matching, the less money you make as users exit the platform. I do agree that it would be great if osmeone could solve this problem though.<\/p><p>&gt; Helping doctors and patients estimate costs - Neither doctors or patients understand costs. it is very routine for a doctor to suggest getting a lab test from X place because they have experience working with the center. They have no idea that for your specific insurance plan this will cost you 2x another place and so you with an unhappy patient who blames their doctor for ripping them off. There should be tools to patients and doctors estimate costs.<\/p><p>We had a startup that came at this in an indirect way. We were trying to make it easier for labs and other providers to perform eligibility checks and facilitate prior auths in real time. Our proposed solution would have involved running the check during ordering, and then having the phlebotomist or lab tech at the hospital doing the sample collection contact the doctor and inform them that something would or would not be covered by insurance. The provider could then talk to the patient about out of pocket costs etc. I think this actually could have worked; our team imploded before we could prototype it however :(<\/p><p>While I like the other ideas, glhf trying to get people to practice lifestyle management or be adherent to a treatment regimen XD\n              <\/p><\/div><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Arete31415\" class=\"hnuser\">Arete31415<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004981\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaicianflone\" class=\"hnuser\">kaicianflone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002627\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gimlids\" class=\"hnuser\">gimlids<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996180\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seige\" class=\"hnuser\">seige<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996454\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kamarg\" class=\"hnuser\">kamarg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996925\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=laken\" class=\"hnuser\">laken<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996929\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nelsontkq\" class=\"hnuser\">nelsontkq<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999245\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=creatornator\" class=\"hnuser\">creatornator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004946\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Pinbenterjamin\" class=\"hnuser\">Pinbenterjamin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16995974\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tixocloud\" class=\"hnuser\">tixocloud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996062\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=choogi\" class=\"hnuser\">choogi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996345\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shapiro92\" class=\"hnuser\">shapiro92<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17011616\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=awshepard\" class=\"hnuser\">awshepard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996808\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nugget\" class=\"hnuser\">nugget<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996784\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=misiti3780\" class=\"hnuser\">misiti3780<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996041\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevesimmons\" class=\"hnuser\">stevesimmons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001065\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=contingencies\" class=\"hnuser\">contingencies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003780\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ryanSrich\" class=\"hnuser\">ryanSrich<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996660\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=austenallred\" class=\"hnuser\">austenallred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996783\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaicianflone\" class=\"hnuser\">kaicianflone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002693\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=austenallred\" class=\"hnuser\">austenallred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17012546\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=techsin101\" class=\"hnuser\">techsin101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17063753\">5 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cloogshicer\" class=\"hnuser\">cloogshicer<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997284\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=austenallred\" class=\"hnuser\">austenallred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999266\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aerovistae\" class=\"hnuser\">aerovistae<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001051\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ColonelPhantom\" class=\"hnuser\">ColonelPhantom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997131\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=austenallred\" class=\"hnuser\">austenallred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999269\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=javiramos\" class=\"hnuser\">javiramos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002631\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moftz\" class=\"hnuser\">moftz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003429\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=solveeng\" class=\"hnuser\">solveeng<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007949\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Procuriffy\" class=\"hnuser\">Procuriffy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17016936\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lastofus\" class=\"hnuser\">lastofus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003254\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Procuriffy\" class=\"hnuser\">Procuriffy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17017038\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hackerews\" class=\"hnuser\">hackerews<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002737\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=contingencies\" class=\"hnuser\">contingencies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003663\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ChristianGeek\" class=\"hnuser\">ChristianGeek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005228\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kaicianflone\" class=\"hnuser\">kaicianflone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996767\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nulagrithom\" class=\"hnuser\">nulagrithom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999075\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dchuk\" class=\"hnuser\">dchuk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999421\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nulagrithom\" class=\"hnuser\">nulagrithom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999715\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hboon\" class=\"hnuser\">hboon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001995\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dchuk\" class=\"hnuser\">dchuk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002290\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nulagrithom\" class=\"hnuser\">nulagrithom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005103\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dchuk\" class=\"hnuser\">dchuk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005435\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dhruvkar\" class=\"hnuser\">dhruvkar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000479\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nulagrithom\" class=\"hnuser\">nulagrithom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005071\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dragonwriter\" class=\"hnuser\">dragonwriter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002467\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=querulous\" class=\"hnuser\">querulous<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17010189\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=greenleafjacob\" class=\"hnuser\">greenleafjacob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001091\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CM30\" class=\"hnuser\">CM30<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999771\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=herbst\" class=\"hnuser\">herbst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006383\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moorhosj\" class=\"hnuser\">moorhosj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996826\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rando444\" class=\"hnuser\">rando444<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005718\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shakkhar\" class=\"hnuser\">shakkhar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997482\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moorhosj\" class=\"hnuser\">moorhosj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997625\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tenzo\" class=\"hnuser\">tenzo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008935\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lkrubner\" class=\"hnuser\">lkrubner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008192\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aaavl2821\" class=\"hnuser\">aaavl2821<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996610\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=canadianwriter\" class=\"hnuser\">canadianwriter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996208\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=raleigh_user\" class=\"hnuser\">raleigh_user<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996815\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=awshepard\" class=\"hnuser\">awshepard<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997055\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrisfrantz\" class=\"hnuser\">chrisfrantz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999853\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hackerews\" class=\"hnuser\">hackerews<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002710\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=canadianwriter\" class=\"hnuser\">canadianwriter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997502\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Joeri\" class=\"hnuser\">Joeri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996732\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hackerews\" class=\"hnuser\">hackerews<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002731\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mlevental\" class=\"hnuser\">mlevental<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000124\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Joeri\" class=\"hnuser\">Joeri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007008\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Raj7k\" class=\"hnuser\">Raj7k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001802\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=basch\" class=\"hnuser\">basch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997217\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LukeB42\" class=\"hnuser\">LukeB42<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001120\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=raleigh_user\" class=\"hnuser\">raleigh_user<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996441\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=abhimir\" class=\"hnuser\">abhimir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997743\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=scrollaway\" class=\"hnuser\">scrollaway<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996029\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dividuum\" class=\"hnuser\">dividuum<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996273\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tixocloud\" class=\"hnuser\">tixocloud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16995810\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=techsin101\" class=\"hnuser\">techsin101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17063800\">5 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tixocloud\" class=\"hnuser\">tixocloud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17101236\">5 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SteveNuts\" class=\"hnuser\">SteveNuts<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16995947\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fermienrico\" class=\"hnuser\">fermienrico<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996318\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Something1234\" class=\"hnuser\">Something1234<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17005031\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hackerews\" class=\"hnuser\">hackerews<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002701\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cashsterling\" class=\"hnuser\">cashsterling<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17010515\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Aeolus98\" class=\"hnuser\">Aeolus98<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004648\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=snerbles\" class=\"hnuser\">snerbles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996866\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SteveNuts\" class=\"hnuser\">SteveNuts<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997907\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fermienrico\" class=\"hnuser\">fermienrico<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16997306\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16998520\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fiatjaf\" class=\"hnuser\">fiatjaf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002235\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007349\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nulagrithom\" class=\"hnuser\">nulagrithom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16999115\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001208\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hazz99\" class=\"hnuser\">hazz99<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001735\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=i_made_a_booboo\" class=\"hnuser\">i_made_a_booboo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004677\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=PerfectElement\" class=\"hnuser\">PerfectElement<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008741\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=karag\" class=\"hnuser\">karag<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17004245\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=solveeng\" class=\"hnuser\">solveeng<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008048\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eurticket\" class=\"hnuser\">eurticket<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996569\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joddystreet\" class=\"hnuser\">joddystreet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001399\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrisgd\" class=\"hnuser\">chrisgd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17007224\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrei_says_\" class=\"hnuser\">andrei_says_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996695\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=khamba\" class=\"hnuser\">khamba<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17003127\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrei_says_\" class=\"hnuser\">andrei_says_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17023921\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xdeadbeefbabe\" class=\"hnuser\">0xdeadbeefbabe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996606\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wiml\" class=\"hnuser\">wiml<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17000791\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Bombthecat\" class=\"hnuser\">Bombthecat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17001565\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zok3102\" class=\"hnuser\">zok3102<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006741\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Bombthecat\" class=\"hnuser\">Bombthecat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17006948\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chirau\" class=\"hnuser\">chirau<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17002528\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CodeKommissar\" class=\"hnuser\">CodeKommissar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17008481\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aloukissas\" class=\"hnuser\">aloukissas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17010069\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hullsean\" class=\"hnuser\">hullsean<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17010299\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=0xdeadbeefbabe\" class=\"hnuser\">0xdeadbeefbabe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996630\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=toephu2\" class=\"hnuser\">toephu2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=16996651\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fuzzfactor\" class=\"hnuser\">fuzzfactor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17012685\">6 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=techsin101\" class=\"hnuser\">techsin101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17063822\">5 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:23:48+0000","updated_at":"2018-11-07T16:23:48+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":166,"title":"De-Googling my phone","url":"https:\/\/piware.de\/post\/2018-05-01-android-degoogle\/","content":"<p>I\u2019ve been a professional Free Software developer in the GNU\/Linux area for 14\nyears now, and a hobbyist developer and user for much longer. For some reason\nthat never extended much to the smartphone world, beyond running\n<a href=\"https:\/\/lineageos.org\/\">LineageOS<\/a> on my older phones (my current Sony Xperia\nis still under warranty and I\u2019m fine with the officially supported Android),\nand various stabs at using the Ubuntu phone (RIP!).<\/p><p>On a few long weekends this year it got a hold of me, and I had a look over the\nGoogle fence to see how Free Software is doing on Android and how to reduce my\ndependency on Google Play Services and Google apps. Less because I would\nactually severely distrust Google, as they have a lot of business and goodwill\nto lose if they ever majorly screw up; but more because of simple curiosity and\nfor learning new things. I want to note down my experience here for sharing and\ndiscussing.<\/p><p>I started experimenting on my old Nexus 4 by completely blanking it and\ninstalling current LineageOS 14.1 without the Google apps. This provides a nice\ntesting ground that is completely free of any proprietary Google stuff. From\nthat I can apply good solutions on my \u201cproduction\u201d Xperia.<\/p><h2 id=\"alternative-app-stores-and-free-apps\">Alternative app stores and free apps<\/h2><p>The number one must have is of course to install the\n<a href=\"https:\/\/f-droid.org\/\">F-Droid<\/a> app, which opens the door to a surprisingly\nlarge and good world of Free Software apps. With that I was already able to replace a\ngood deal of my installed apps with free alternatives:<\/p><ul><li><p><a href=\"https:\/\/f-droid.org\/en\/packages\/com.fsck.k9\/\">K9 Mail<\/a>,\n<a href=\"https:\/\/f-droid.org\/en\/packages\/org.telegram.messenger\/\">Telegram<\/a>,\nand <a href=\"https:\/\/f-droid.org\/en\/packages\/net.osmand.plus\/\">OsmAnd+<\/a> are also\navailable on F-Droid, which I find preferrable as they rebuild everything\nfrom source and thus binaries seem a bit more trustworthy.<\/p><\/li>\n<li><p>I previously used Opera as a web browser, because it is relatively\nlightweight (important on my previous phone) and the really good builtin ad\nblocker. But these days Firefox is really fast and good enough, so I\nreplaced it with\n<a href=\"https:\/\/f-droid.org\/en\/packages\/org.mozilla.fennec_fdroid\/\">Fennec<\/a>, which\nis more or less Firefox with some non-free bits removed. After installing\n<a href=\"https:\/\/addons.mozilla.org\/de\/firefox\/addon\/ublock-origin\/\">uBlock Origin<\/a>\nI\u2019ve never looked back.<\/p><\/li>\n<li><p>Some apps have fairly straightforward replacements:<\/p>\n<ul><li>Google Authenticator \u2192 <a href=\"https:\/\/f-droid.org\/en\/packages\/org.fedorahosted.freeotp\/\">FreeOTP<\/a><\/li>\n<li>YouTube \u2192 <a href=\"https:\/\/f-droid.org\/en\/packages\/org.schabi.newpipe\/\">NewPipe<\/a><\/li>\n<li>BlueFilter \u2192 <a href=\"https:\/\/f-droid.org\/en\/packages\/info.papdt.blackblub\/\">Night Screen<\/a><\/li>\n<li>QR Scanner \u2192 <a href=\"https:\/\/f-droid.org\/en\/packages\/de.t_dankworth.secscanqr\/\">SecScanQR<\/a><\/li>\n<li>ConnectBot \u2192 <a href=\"https:\/\/f-droid.org\/en\/packages\/com.termux\/\">Termux<\/a> (which\nis actually a magnitude more powerful and also has a better terminal\nimplementation)<\/li>\n<li>GPS Status \u2192 <a href=\"https:\/\/f-droid.org\/en\/packages\/com.vonglasow.michael.satstat\/\">SatStat<\/a><\/li>\n<\/ul><\/li>\n<li><p>Some other apps like Google+ and Pocketliga just got replaced with bookmarks\nto web sites, which are good enough. This includes Google Maps; I usually\nprefer and use OsmAnd for walking, biking, and general navigation\/routing.\nBut for finding commercial things Maps is usually better; however, I use it\nrarely enough to not get too annoyed by the much higher latency of the web\npage.<\/p><\/li>\n<\/ul><p>There are also some apps that I want (like WhatsApp, Nextbike, or Quizduell)\nthat don\u2019t have free software alternatives. <a href=\"https:\/\/uptodown.com\/\">Uptodown<\/a>\nis an alternative app store that does not require Play Services and has a lot\nof apps like the above or the Deutsche Bahn Navigator (but that requires Play\nServices by itself, so that doesn\u2019t help that much).<\/p><p>This was also easy. I\u2019ve run my own Mail server forever, use\n<a href=\"http:\/\/www.mutt.org\/\">mutt<\/a> on the desktop,\n<a href=\"https:\/\/www.rainloop.net\/\">Rainloop<\/a> for the occasional web mail access, and\nthe gorgeous <a href=\"https:\/\/f-droid.org\/en\/packages\/com.fsck.k9\/\">K9<\/a> on the phone.\nI have always loathed the GMail web UI, so I never got into the habit of using\nGMail much, except as a sacrificial spam address for random web shops. But it\nturns out that this is surprisingly simple to use without the GMail app and\nPlay services - just create an <a href=\"https:\/\/support.google.com\/mail\/answer\/185833\">app password<\/a>\nand use it in a new K9 mail account. This also simplifies things to not having\nto deal with two different email apps.<\/p><h2 id=\"notes-and-todo-lists\">Notes and TODO lists<\/h2><p>I\u2019ve used <a href=\"https:\/\/keep.google.com\">Google Keep<\/a> for a long time for notes and\nTODO lists. The Android app and widget, and the web UI have a nice no-nonsense\napproach and work just fine. Perhaps surprisingly this was the thing that took\nme longest to replace: there is a plethora of notes\/TODO apps around, and I\u2019ve\ntested quite a few of them.<\/p><p>During that I discovered <a href=\"https:\/\/orgmode.org\/\">Org mode<\/a>, a very powerful and\nflexible plaintext-based system (a bit like Markdown in spirit, but different\nsyntax and much more powerful). Primarily it is an Emacs extension, but there\nis a reasonably good <a href=\"https:\/\/github.com\/jceb\/vim-orgmode\">vim-orgmode<\/a> plugin\nfor writing on the desktop (which I much prefer over typing in the Keep web\nUI!), and <a href=\"https:\/\/f-droid.org\/en\/packages\/com.orgzly\/\">Orgzly<\/a> for Android\nthat has a pleasant UI for maintaining notes, and also a nice widget for\nputting my TODO list on the starter screen.<\/p><p>Neither of these can sync with other devices, but fortunately this (rather\nindependent) task also has a nice solution, see below.<\/p><p>I\u2019ve happily used these tools for two months now. vim-orgmode fits my usual\nworkflow on the desktop much better, and on the phone the experience is not\nsignificantly different (I don\u2019t use the more powerful Org features - yet).<\/p><h2 id=\"file-syncing\">File syncing<\/h2><p>While looking for how to synchronize my Org notes between phone and laptop, I\nstumbled over the really magnificient <a href=\"https:\/\/syncthing.net\/\">Syncthing<\/a>\nproject. It\u2019s a general-purpose many-to-many file sync solution using\nmodern and standard encryption and authentication technology, and is readily\navailable on Debian (for my server), Fedora (for my laptop), and F-Droid (for\nmy phone). Installation and configuration on the web and Android UIs are\nstraightforward, and it can just as easily be used to backing up your photos or\nsyncing music or ebooks (for you Dropbox or Google Drive using types - I use\nneither, though).<\/p><p>The main disadvantage to Keep or other Google apps is that there is no push\nnotification on changes - all instances of Syncthing poll, so there is some\ndelay in propagating updates. Also, Android keeps nagging with a \u201cSyncthing is\nrunning in the background\u201d notification, but fortunately it can be toned down\nenough in Android 8 to not get in the way any more. Power usage is obviously\nsomething to watch out for here, but I haven\u2019t noticed any significantly worse\nbattery drain so far.<\/p><p>A few years ago I had run <a href=\"https:\/\/owncloud.org\/\">ownCloud<\/a> for quite some\ntime, mostly to store my contacts and calendar. But it takes a non-trivial\namount of setup, maintance, and resources on the server and has a rather high\ninflux of <a href=\"http:\/\/cve.mitre.org\/cgi-bin\/cvekey.cgi?keyword=owncloud\">security issues<\/a>,\nso I went back to Google eventually.<\/p><p>Today I discovered <a href=\"http:\/\/radicale.org\/\">Radicale<\/a>, which is much more\nappealing: It is a lightweight Python module that provides a CalDAV\/CardDAV API\nfor a bunch of standard <code>*.ics<\/code> (calendar) and <code>*.vcf<\/code> (contact) files. For the\nmost part its <a href=\"http:\/\/radicale.org\/documentation\/\">documentation<\/a> is excellent,\nand I set up the service and Apache reverse proxying on my server in no time.<\/p><p>The only thing that is not documented well is how to import existing ics\/vcf\nfiles (exported from Google Calendar\/Contacts). Turns out one can just use a\nstandard HTTP\/WebDAV <code>PUT<\/code> request:<\/p><pre>curl -u 'martin:myPassWord' -X PUT https:\/\/my.server.de\/radicale\/martin\/contacts\/ --data-binary @contacts.vcf\n<\/pre><p>(the same thing with exported <code>*.ics<\/code> for calendar).<\/p><p>On Android I installed\n<a href=\"https:\/\/f-droid.org\/en\/packages\/at.bitfire.davdroid\/\">DAVdroid<\/a> where I just\nadded the Radicale root URL (with user\/password), then saw my two\ncontact\/calendar sources, and enabled them. I disabled the phone\u2019s Google\naccount on the phone instead. Yay, once again I own my data. :-)<\/p><p>This does not provide a web UI for the calendar\/contacts. I rarely used\n<a href=\"https:\/\/calendar.google.com\">https:\/\/calendar.google.com<\/a>, and essentially never <a href=\"https:\/\/contacts.google.com\">https:\/\/contacts.google.com<\/a>,\nso I suppose I won\u2019t actually need to find a replacement. But if so, I might\ntake a look at Roundcube to replace Rainloop, as the former has a plugin for\ncalendar. I don\u2019t have Thunderbird, Evolution, or anything similar installed,\nas that\u2019s just too heavyweight for the rare cases where I deal with my personal\ncalendar on the laptop.<\/p><h2 id=\"remaining-items\">Remaining items<\/h2><p>The one remaining app on my phone that needs Google Play Services is now the\nDeutsche Bahn Navigator. I like and use that a lot, both for information and\nfor buying tickets. On my next trip I\u2019ll give <a href=\"https:\/\/www.trainline.eu\/\">trainline.eu<\/a>\na go, which looks very promising: same price as <a href=\"https:\/\/bahn.de\">https:\/\/bahn.de<\/a> directly, they\nconsider my BahnCard, don\u2019t need Play Services, and can also book trains in\nother countries.<\/p><p>There are also some warts with other apps (Quizduell and swa Fahrinfo) that\noccasionally complain about missing Play Services, although they work well\nenough without. In Nextbike one loses the life map where bikes are available,\nwhich is quite important; a reasonable alternative is to look that up on their\nwebsite, which provides the same map. Actually booking a bike and all the other\nfunctionality works fine without Play Services.<\/p><h2>Conclusion<\/h2><p>I\u2019ve learned a lot of new things during the above, so this was a worthwhile\nexperiment beyond mere political and idealistic reasons. However, setting up\nsuch a \u201cGoogle free\u201d phone requires a significant amount of time, dedication,\nand skill (not to mention a personal server), definitively far beyond to what\nI\u2019d trust my mother to comprehend. But it\u2019s reassuring to know that, if push\ncomes to shove, it\u2019s definitively still possible to have an useful smartphone\nthat\u2019s running mostly FOSS and no Google proprietary stuff!<\/p><p><strong>Update<\/strong>: I posted a\n<a href=\"https:\/\/piware.de\/post\/2018-05-21-android-degoogle-2\/\">follow-up<\/a> that\nresolves and amends some issues above.<\/p>","created_at":"2018-11-07T16:23:44+0000","updated_at":"2018-11-07T16:23:44+0000","published_at":"2018-05-01T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"piware.de","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":165,"title":"Mail-in-a-Box \u2013 one-click, easy-to-deploy email server","url":"https:\/\/news.ycombinator.com\/item?id=13050500","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D13050500\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"13050500\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/github.com\/mail-in-a-box\/mailinabox\" class=\"storylink\">Mail-in-a-Box \u2013 one-click, easy-to-deploy email server<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=github.com\">github.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        188 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=avinassh\" class=\"hnuser\">avinassh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=13050500\">on Nov 27, 2016<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=13050500&amp;goto=item%3Fid%3D13050500\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Mail-in-a-Box%20%E2%80%93%20one-click%2C%20easy-to-deploy%20email%20server&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Mail-in-a-Box%20%E2%80%93%20one-click%2C%20easy-to-deploy%20email%20server\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=13050500&amp;auth=f5e637538bba2253623893ebb4fd410629906535\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=13050500\">90\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:23:43+0000","updated_at":"2018-11-07T16:23:43+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":164,"title":"ISPmail guide for Debian Jessie","url":"https:\/\/workaround.org\/ispmail\/jessie","content":"<p><a href=\"https:\/\/workaround.org\/wp-content\/uploads\/2015\/10\/christoph-haas.png\" class=\"simplelightbox\" rel=\"lightbox[504]\"><img class=\"size-full wp-image-1456 alignright\" src=\"https:\/\/workaround.org\/wp-content\/uploads\/2015\/10\/christoph-haas.png\" alt=\"christoph-haas\" width=\"99\" height=\"99\" \/><\/a>Welcome to the ISPmail tutorial for Debian Jessie. Since Debian Woody I have provided step-by-step-instructions for setting up and operating your own mail server. So this is already the seventh issue. Yay.<\/p><p>A personal note: I have been working on this issue for three weeks in my spare time. And I deliberately removed any advertisements from workaround.org because I\u2019m opposed to that. So if this guide helps you in any way please consider donating. You can find the appropriate links at the very bottom of this page. Thank you. And of course your feedback or questions are\u00a0welcome \u2013 comment on any page of the ISPmail guide that you find appropriate.<\/p>\n<h2>What this tutorial is about<\/h2>\n<p>This tutorial teaches you all the basics about running a mail server that is completely controlled by you, as secure as you like and fully powered by open-source software. And while you are setting up your mail server you will learn a lot about communication protocols like IMAP or SMTP. You will also learn how emails find their way through the internet using DNS records.\u00a0You can use it to run a free mail server for your company, your familiy, your friends or just for yourself. You can put any number of accounts and domains on it. So screw \u201cfree\u201d mail services and take control of your email communication again. Essentially all you need is a computer hooked to the internet, an installation medium of Debian Jessie and a little\u00a0time.\u00a0When you are done your mail server will be able to\u2026<\/p>\n<ul><li>receive and store email for your users from other mail servers on the internet<\/li>\n<li>let your users retrieve the email through IMAP and POP3 using TLS-encrypted connections<\/li>\n<li>let your users define rules what to do with incoming email \u2013 e.g. distribute them to their mail folders or reply automatically<\/li>\n<li>let your users send (\u201crelay\u201d) email to the internet using SMTP authentication<\/li>\n<li>offer a beautiful webmail interface so that your users can manage their email just using a web browser<\/li>\n<li>deal with spam effectively\u00a0without losing emails<\/li>\n<\/ul><h2>What this tutorial is not\u00a0about<\/h2>\n<p>If you are lazy or in a hurry or just don\u2019t want to learn anything about mail servers \u2013 then this guide is not for you.\u00a0There are ready solutions like\u00a0<a href=\"http:\/\/www.iredmail.org\/\">iRedMail<\/a>\u00a0that you may want to consider. But running a mail server requires a certain\u00a0technical understanding. And that\u2019s what the ISPmail guide is for. Experience from giving support in the #postfix channel on IRC shows that sysadmins fail because some detail in a complex setup goes wrong and they have no idea how to track it down. Go with ready solutions if you like. But I have a feeling that we meet again. And you will probably not save time either taking the easy route.<\/p>\n<h2>Why is it called ISPmail?<\/h2>\n<p>The software components that are used in this guide are proven and powerful enough so that you could become an ISP (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Internet_service_provider\">internet service provider<\/a>) yourself and sell email services. You will need powerful servers, a backup solution, proper monitoring, staff and a bigger internet connection. But essentially the system will work for 10 or 10,000 users.<\/p>\n<h2>Does this guide work with other flavors of Linux?<\/h2>\n<p>The software components are available on most Linux distributions. However the versions will vary slightly and the installation paths in your filesystem will be different.\u00a0So feel free to work through this guide to learn the basics. But be prepared if pieces do not work out of the box. I am considering to offer this tutorial for Ubuntu 14.04 LTS because it gets increasingly popular \u2013 but currently the priority is upon Debian Jessie.<\/p>\n<h2>Things you will need<\/h2>\n<p>The server setup described here is standard work for a professional system administrator. Depending on your level of knowledge and experience you may walk through this document easily or need more patience to get to grips with the basics. It works best if are familiar with\u00a0basic system administration tasks like editing text files or starting\/stopping services, DNS and MySQL. Make sure you have:<\/p>\n<ul><li>One or more internet<strong> domains<\/strong> to receive email at. Either you manage the DNS yourself or you let domain providers do that for you. However you need to be able to change DNS records. If your ISP just offers \u201cweb sites\u201d or \u201cweb redirects\u201d then transfer your domain to a decent provider.<\/li>\n<li>A <strong>computer<\/strong> capable of running Debian Jessie. Unless you have special needs I would suggest you get a low-cost virtual server from your favorite hosting company. Those start at $5 for example at <a href=\"https:\/\/vexxhost.com\/\">Vexxhost<\/a>. Start with 10 GB of mailbox space per user, 1 GB of RAM and any CPU.<\/li>\n<li>A \u00a0permanent <strong>internet connection <\/strong>and an<strong> IP address<\/strong>\u00a0for your mail server that does not change. The IP address\u00a0should <strong>not be\u00a0blacklisted<\/strong> on the internet. Check the IP address at web sites like\u00a0<a href=\"http:\/\/www.anti-abuse.org\/multi-rbl-check\/\" class=\"broken_link\">http:\/\/www.anti-abuse.org\/multi-rbl-check\/<\/a>\u00a0or\u00a0<a href=\"http:\/\/rbl-check.org\/\">http:\/\/rbl-check.org\/<\/a>. If your IP address shows up on black lists then other mail servers on the internet will likely refuse your emails or consider them spam.<\/li>\n<li>Experience with <strong>operating a Linux system<\/strong> \u2013 preferably even a Debian\/Ubuntu system.<\/li>\n<li>Basic understanding of <strong>DNS<\/strong>. You will need to set up DNS records of the domains that you like to receive email for. At least you need someone who can manage \u201cMX\u201d and \u201cA\u201d records for your domain.<\/li>\n<li>Enough skills to install and configure a basic <strong>MySQL database server<\/strong>. You will not need to tune the server. But you should understand the basics of relational databases (databases, tables, rows, fields\/columns) and the meaning of simple SQL queries.<\/li>\n<li><strong>Control over the firewall<\/strong> that connects your server to the internet. You will need to allow\u00a0certain TCP ports through to your mail server. Also make sure that your internet provider does not block any ports. Professional ISPs are usually not a problem. But cheap home ISPs often block ports like TCP\/25 which is used for the SMTP protocol.<\/li>\n<li><strong>Time<\/strong>.\u00a02 hours of time if you are a seasoned Linux administrator. Or 2 days if running your own Linux server on the internet is a new topic for you.<\/li>\n<\/ul><h2>To do<\/h2>\n<p>Although the tutorial is complete as it is I am working on these additional sections:<\/p>\n<ul><li>Automatic deployment with Ansible<\/li>\n<li>Automatic signing of outgoing email using domain keys (DKIM)<\/li>\n<li>A Ruby-on-Rails web application that helps you manage your mail server<\/li>\n<\/ul><h2>Please respect the copyright<\/h2>\n<p>Please ask if you would like to use parts on your own web site.\u00a0Feel free to contact me if you have any questions or leave a donation if you found the tutorial useful. I am maintaining these tutorials in my spare time and am not making any money with it. It annoys me to find copies of this tutorial or parts of it on other web sites. I think you will understand that.<\/p>\n<h2>About this tutorial<\/h2>\n<p>Over a decade ago I wanted to turn my Debian server into a mail server with virus scanning, spam detection, email forwarding, POP3 and IMAP access and a webmail service. All the components were available but it took me quite a while until they worked properly together. So I summed up my desk full of scrawly notes into a tutorial that has become pretty famous over ten years ago. This document is not a simple copy-and-paste tutorial where you just copy the commands from the web site and run it on your server and hope they will work. Instead it will make you understand the different components that you are setting up. In the end you will be skilled enough to debug problems yourself. Unlike many other Postfix tutorials on the internet this is already the seventh edition. Writing this tutorial took a lot of work so these are not just quick draft notes thrown together but a consistent document guiding you. The whole tutorial is split into several pages. Please use the links on the right side\u00a0to\u00a0navigate through the pages. You are also invited to comment on the pages or ask questions.\u00a0Just use the comment feature at the bottom of each page.<\/p>\n<div class=\"post-views post-504 entry-meta\">\n\t\t\t\n\t\t\tViews: \n\t\t\t78,761\n\t\t\t<\/div>","created_at":"2018-11-07T16:23:41+0000","updated_at":"2018-11-07T16:23:41+0000","published_by":["ThemeGrill"],"annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":6,"domain_name":"workaround.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":161,"title":"Figwheel with new Clojure 1.9 CLI tools","url":"http:\/\/www.functionalbytes.nl\/clojure\/nodejs\/figwheel\/repl\/clojurescript\/cli\/2017\/12\/20\/tools-deps-figwheel.html","content":"<p>With the <a href=\"http:\/\/blog.cognitect.com\/blog\/clojure19\">release<\/a> of Clojure 1.9, Cognitect has added command line tools for Clojure.\nUsing those tools it becomes quite easy to run a Clojure script\/application or start a REPL.\nNext to this, it can also be used to fetch project dependencies, form a classpath with those dependencies and start a script using this classpath.<\/p><p>Because of the latter feature, the new CLI tools seem suitable for building ClojureScript projects, instead of using leiningen or boot.\nAll you need is your own build script, started using the CLI tools.\nFor a recently started ClojureScript on NodeJS project, we set out to find a workable setup for building and developing it using such a build script.\nA workable setup for us would at least include being able to compile and test the project, having a <a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\">Figwheel REPL<\/a> available and being able to connect to it using <a href=\"https:\/\/cider.readthedocs.io\/\">Emacs CIDER<\/a>.<\/p><p>This blog post shows our minimal but workable setup.<\/p><h3 id=\"the-depsedn-file\">The deps.edn file<\/h3><p>To declare the dependencies of the project, a <code class=\"highlighter-rouge\">deps.edn<\/code> file is required by the CLI tools.\nA minimal file would have the following contents:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">{:deps {org.clojure\/clojurescript {:mvn\/version \"1.9.946\"}\n :aliases\n {:repl {:extra-deps\n         {;; Figwheel ClojureScript REPL\n          com.cemerick\/piggieback {:mvn\/version \"0.2.2\"\n                                   :exclusions  [com.google.javascript\/closure-compiler]}\n          figwheel-sidecar        {:mvn\/version \"0.5.14\"\n                                   :exclusions  [com.google.javascript\/closure-compiler]}\n          ;; CIDER compatible nREPL\n          cider\/cider-nrepl       {:mvn\/version \"0.15.1\"}\n          org.clojure\/tools.nrepl {:mvn\/version \"0.2.12\"}\n          refactor-nrepl          {:mvn\/version \"2.3.1\"}}}}}<\/pre><\/figure><p>The <code class=\"highlighter-rouge\">:deps<\/code> map declares the projects dependencies required to build it.\nNext to this, there is also the <code class=\"highlighter-rouge\">:repl<\/code> alias, which declares the dependencies required to start a CIDER-compatible Figwheel ClojureScript REPL.\nIf your project does not need CIDER compatibility, you could leave out the <code class=\"highlighter-rouge\">cider\/cider-nrepl<\/code>, <code class=\"highlighter-rouge\">org.clojure\/tools.nrepl<\/code> and <code class=\"highlighter-rouge\">refactor-nrepl<\/code> dependencies.<\/p><p>When you use the <code class=\"highlighter-rouge\">clojure<\/code> (or <code class=\"highlighter-rouge\">clj<\/code>) CLI tool, it will use the dependencies in the <code class=\"highlighter-rouge\">:deps<\/code> map.\nIf you want the REPL dependencies mixed in as well, you need to supply the <code class=\"highlighter-rouge\">:repl<\/code> alias using the <code class=\"highlighter-rouge\">-R<\/code> option, like so:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clj -R:repl &lt;your-script.clj&gt;<\/pre><\/figure><h3 id=\"the-build-script\">The build script<\/h3><p>Next is the build script.\nThe goal is to have the script perform a certain task, such as <code class=\"highlighter-rouge\">compile<\/code>, <code class=\"highlighter-rouge\">test<\/code> or <code class=\"highlighter-rouge\">figwheel<\/code>, and thus use the CLI tools like this:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clojure build.clj &lt;task&gt; [task args]<\/pre><\/figure><p>For this to work, the following \u201cframework\u201d is created in the <code class=\"highlighter-rouge\">build.clj<\/code> script:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">;;; Task dispatching.\n(defmulti task first)\n(defmethod task :default\n  [args]\n  (let [all-tasks  (-&gt; task methods (dissoc :default) keys sort)\n        interposed (-&gt;&gt; all-tasks (interpose \", \") (apply str)))]\n    (println \"Unknown or missing task. Choose one of:\" interposed)\n    (System\/exit 1)))\n;;; Build script entrypoint. This should be the last expression.\n(task *command-line-args*)<\/pre><\/figure><p>Now we can define a <code class=\"highlighter-rouge\">task<\/code> method for each of the tasks we want, dispatched on the the <code class=\"highlighter-rouge\">first<\/code> argument to the build script.<\/p><h4 id=\"compiling-the-project\">Compiling the project<\/h4><p>Let\u2019s add the first task for compiling the project.\nWe need to require the ClojureScript compiler API namespace and use it to call the ClojureScript compiler:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(require '[cljs.build.api :as api])\n(def source-dir \"src\")\n(def compiler-config {:main          'my-app.core\n                      :output-to     \"target\/my-app\/main.js\"\n                      :output-dir    \"target\/my-app\/main\"\n                      :target        :nodejs\n                      :optimizations :simple\n                      :source-map    \"target\/my-app\/main.js.map\"})\n(defmethod task \"compile\" [args]\n  (api\/build source-dir compiler-config))<\/pre><\/figure><p>Easy and effective.\nExecuting <code class=\"highlighter-rouge\">clojure build.clj compile<\/code> will compile the project for us.\nAs our project is a NodeJS project, we would now be able to run it by executing <code class=\"highlighter-rouge\">node target\/my-app\/main.js<\/code>.<\/p><h4 id=\"testing-the-project\">Testing the project<\/h4><p>Next, testing our project.\nFor testing a ClojureScript project, one generally uses a different entrypoint into the application -  the test runner.\nThe same is true for our NodeJS project.\nWe simply execute the NodeJS project from within our build script using the <code class=\"highlighter-rouge\">clojure.java.shell<\/code> namespace:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(require '[clojure.java.shell :as shell])\n(def test-dir \"test\")\n(def test-config {:main          'my-app.test-runner\n                  :output-to     \"target\/test.js\"\n                  :output-dir    \"target\/test\"\n                  :target        :nodejs\n                  :optimizations :none\n                  :source-map    true})\n(defn run-node-tests []\n  (let [{:keys [out err]} (shell\/sh \"node\" \"target\/test.js\")]\n    (println out err)))\n(defn test-once []\n  (api\/build (api\/inputs source-dir test-dir) test-config)\n  (run-node-tests)\n  (shutdown-agents))\n(defmethod task \"test\" [args]\n  (test-once))<\/pre><\/figure><p>Next to a one-off \u201ctest\u201d task, it is very convenient to have the project tested after every change in the source code.\nThis workflow can be had by extending the \u201ctest\u201d task a bit, using the <code class=\"highlighter-rouge\">:watch-fn<\/code> option of the ClojureScript compiler.\nOn every change, the tests will be re-run, without a new JVM spinning up.\nThe \u201ctest\u201d task can be extended as follows:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(defn test-refresh\n  (api\/watch (api\/inputs source-dir test-dir)\n             (assoc test-config :watch-fn run-node-tests)))\n(defmethod task \"test\" [[_ type]]\n  (case type\n    (nil \"once\") (test-once)\n    \"watch\"      (test-refresh)\n    (do (println \"Unknown argument to test task:\" type)\n        (System\/exit 1))))<\/pre><\/figure><p>Now you can execute <code class=\"highlighter-rouge\">clojure build.clj test watch<\/code> to start this continuously testing workflow.\nThe one-off test can still be executed by leaving out the <code class=\"highlighter-rouge\">watch<\/code> argument, or using the <code class=\"highlighter-rouge\">once<\/code> argument.<\/p><h4 id=\"figwheel-repl-driven-development\">Figwheel REPL-driven development<\/h4><p>Now let\u2019s setup a Figwheel REPL task.\nLuckily for us, the Figwheel project is nicely split up between leiningen-specific stuff and the actual compilation, code reloading and REPL mechanics.\nThe latter is what we need, and is available in the <a href=\"https:\/\/github.com\/bhauman\/lein-figwheel\/tree\/master\/sidecar\">sidecar<\/a> library.\nThis library compiles the ClojureScript project with Figwheel client related code inserted and starts the Figwheel server.\nThe Figwheel server re-compiles the project on every source code change and signals the client to reload some files after compilation is complete.<\/p><p>The server can start a ClojureScript REPL for you, hooking into the running application.\nIt can also start an nREPL server for you, such that you can connect to it using CIDER, where you can start the ClojureScript REPL yourself.\nIn our setup, we want to be able to choose either one of these \u201cmodes\u201d, by optionally supplying an nREPL server port.\nThe \u201cfigwheel\u201d task definition then looks as follows:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(require '[figwheel-sidecar.repl-api :as repl-api :refer [cljs-repl]])\n(def dev-config (merge compiler-config\n                       {:optimizations :none\n                        :source-map    true}))\n(defmethod task \"figwheel\" [[_ port]]\n  (repl-api\/start-figwheel!\n   {:figwheel-options (when port\n                        {:nrepl-port       (some-&gt; port Long\/parseLong)\n                         :nrepl-middleware [\"cider.nrepl\/cider-middleware\"\n                                            \"refactor-nrepl.middleware\/wrap-refactor\"\n                                            \"cemerick.piggieback\/wrap-cljs-repl\"]})\n    :all-builds       [{:id           \"dev\"\n                        :figwheel     true\n                        :source-paths [source-dir]\n                        :compiler     dev-config}]})\n  (when-not port\n    (cljs-repl)))<\/pre><\/figure><p>If we launch our Figwheel setup as follows, we get a simple REPL in our terminal:<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clj -R:repl build.clj figwheel<\/pre><\/figure><p>And if we launch as follows, we get an nREPL server, which we can connect to in Emacs by using the <code class=\"highlighter-rouge\">cider-connect<\/code> command.\nIn this CIDER repl, you can call <code class=\"highlighter-rouge\">(cljs-repl)<\/code> yourself, to open the ClojureScript REPL.<\/p><figure class=\"highlight\"><pre class=\"language-sh\" data-lang=\"sh\">$ clojure -R:repl build.clj figwheel &lt;nrepl-port&gt;<\/pre><\/figure><p>Very nice.\nBut, if you try to run our former tasks, such as \u201ccompile\u201d or \u201ctest\u201d, you will see that those won\u2019t work if you don\u2019t add the <code class=\"highlighter-rouge\">-R:repl<\/code> option.\nIt will complain about not being able to load the <code class=\"highlighter-rouge\">figwheel-sidecar.repl-api<\/code> namespace.\nWe don\u2019t want to add the <code class=\"highlighter-rouge\">repl<\/code> alias for all our tasks, for various reasons.\nTherefore we have to make our task definitions check whether the required namespaces are available.\nFor this we introduce a small macro:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(defn try-require [ns-sym]\n  (try (require ns-sym) true (catch Exception e false)))\n(defmacro with-namespaces\n  [namespaces &amp; body]\n  (if (every? try-require namespaces)\n    `(do ~@body)\n    `(println \"task not available - required dependencies not found\")))<\/pre><\/figure><p>Using this macro, we can update our \u201cfigwheel\u201d task a bit:<\/p><figure class=\"highlight\"><pre class=\"language-clj\" data-lang=\"clj\">(defmethod task \"figwheel\" [[_ port]]\n  (with-namespaces [figwheel-sidecar.repl-api]\n    (figwheel-sidecar.repl-api\/start-figwheel!\n     {:figwheel-options (when port\n                          {:nrepl-port       (some-&gt; port Long\/parseLong)\n                           :nrepl-middleware [\"cider.nrepl\/cider-middleware\"\n                                              \"refactor-nrepl.middleware\/wrap-refactor\"\n                                              \"cemerick.piggieback\/wrap-cljs-repl\"]})\n      :all-builds       [{:id           \"dev\"\n                          :figwheel     true\n                          :source-paths [source-dir]\n                          :compiler     dev-config}]})\n    (when-not port\n      (figwheel-sidecar.repl-api\/cljs-repl)))<\/pre><\/figure><p>We can now remove the global require of <code class=\"highlighter-rouge\">figwheel-sidecar.repl-api<\/code> namespace.\nIf we then try to execute the \u201cfigwheel\u201d task without the <code class=\"highlighter-rouge\">:repl<\/code> alias, it will show a message that the required namespaces are missing.\nOther tasks now work again as expected.<\/p><h3 id=\"conclusions\">Conclusions<\/h3><p>Building, testing and developing a ClojureScript application using the new Clojure tools is certainly possible.\nIf anything, it was a nice excercise in how to work with the new tools.\nCreating your own build script makes it very explicit as to how the project is build and tested, instead of \u201cjust add these plugins, and it will probably work\u201d.\nIn other words, in may create insight in how it all works and fits together.<\/p><p>That said, the custom build script implements concepts like tasks, which tools like leiningen and boot have implemented and standardized for you.\nSimply being able to execute <code class=\"highlighter-rouge\">cider-jack-in-clojurescript<\/code> from within Emacs is easier and works because of those standardized build tools.\nThis is fine, as the new Clojure CLI tools and tools like leiningen and boot can complement each other.\nBecause of this, we will probably see more support for the Clojure tools, and the underlying <code class=\"highlighter-rouge\">tools.deps<\/code> library, in those build tools.\nInitial examples of such integration are the <a href=\"https:\/\/github.com\/seancorfield\/boot-tools-deps\">boot-tools-deps<\/a> task and <a href=\"https:\/\/github.com\/RickMoynihan\/lein-tools-deps\">lein-tools-deps<\/a> plugin.<\/p><p>A full build script can be found <a href=\"https:\/\/github.com\/aroemers\/clojure-cli-tools-build\">on github<\/a>.<\/p><p><em>As always, have fun!<\/em><\/p>","created_at":"2018-11-07T16:23:40+0000","updated_at":"2018-11-07T16:23:40+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"www.functionalbytes.nl","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":162,"title":"Eshell as a main shell","url":"https:\/\/ambrevar.xyz\/emacs-eshell\/","content":"<p>\nIt\u2019s undeniable however that, as of Emacs 25.2, Eshell has numerous rough edges.\nQuite a few bug reports have been sent on debbugs and it\u2019s getting better.\n<\/p><p>\nNative completion is limited considering very little effort was ever made.  The\ncommunity can help changing that.  In the meantime, I\u2019ve implemented <a href=\"https:\/\/github.com\/Ambrevar\/emacs-fish-completion\">a fallback\nto Bash\/fish completion<\/a> so that I can keep completing just as much as I could\nwith Bash or fish.\n<\/p><p>\nNeed that curses-based program? Get a (much more) powerful Emacs alternative:\n<\/p><ul class=\"org-ul\"><li>git (CLI), gitk, tig, etc. -&gt; magit<\/li>\n<li>htop -&gt; symon, proced, helm-top, etc.<\/li>\n<li>abook -&gt; org-contacts<\/li>\n<li>mutt -&gt; mu4e, gnus, notmuch<\/li>\n<li>ncdu -&gt; dired-du<\/li>\n<li>cmus\/moc -&gt; EMMS<\/li>\n<li>newsbeuter -&gt; Elfeed, gnus<\/li>\n<li>weechat, irssi -&gt; ERC, etc.<\/li>\n<li>rtorrent, transmission-cli -&gt; transmission.el<\/li>\n<\/ul><p>\nAs opposed to their curses-like counterparts, Emacs alternatives share all those\nfeatures:\n<\/p><ul class=\"org-ul\"><li>Configure\/extend in Lisp.<\/li>\n<li>(Fuzzy-)search the buffer.  Copy anything to the clipboard.<\/li>\n<li>No need to change the color theme, it\u2019s already using your Emacs theme.  No\nneed for dircolors, lesspipe, Xresources\u2026<\/li>\n<li>Emacs bindings (Evil?).  All configurable. No more painful context-switch\nbetween Emacs-based and vi-based bindings.<\/li>\n<\/ul><p>\nIn the meantime, there might still be that annoying curses-based program you\ncannot replace.  Eshell allows you to run those programs by transparently\nswitching to a <code>*term*<\/code> buffer.  See <code>eshell-visual-commands<\/code>.\n<\/p><p>\nCurses-based programs are not the only ones that won\u2019t run well on Eshell: <a href=\"https:\/\/en.wikipedia.org\/wiki\/Read%E2%80%93eval%E2%80%93print_loop\">REPL<\/a>\nare known for working poorly too.  REPLs are poorly designed though: from a\nUnix-philosophy standpoint, only the <code>E<\/code> is required for every language, while\nre-implementing the <code>R<\/code>, <code>P<\/code>, and <code>L<\/code>, i.e.\u00a0the user interface, for each one of\nthem is simply re-inventing the wheel.  Instead of that, Emacs has <code>comint-mode<\/code>,\nan REPL-lie user interface that can be plugged to any language.  With all the\npowerful features of Emacs that all those specialized REPLs are lacking.  For a\nstart, you can write code in Emacs and execute any piece of code from a single\nkeystroke.  It\u2019s configurable and extensible in a consistent manner across all\nlanguages.\n<\/p><p>\nSimilarly, \u201cenvironment setups\u201d, as commonly used by programming language tools\nsuch as Python\u2019s <code>virtualenv<\/code> or Ruby <code>rvm<\/code>, expect the user to be running a\nspecific shell.  This is broken by design.  Let me quote <a href=\"https:\/\/reddit.com\/u\/rpdillon\/\"><i>u\/rpdillon<\/i><\/a> as I could\nnot have said better:\n<\/p><blockquote>\n<p>\nTo OP\u2019s point, though, yes, these environments do make assumptions that you\u2019ll\nbe using some kind of bash-compatible shell, which tells you a little bit about\nhow sub-optimal the state-of-the-art is.  I suppose the only alternative I can\nthink of is to have the language itself manage many versions of itself so that\nevery developer\u2019s environment wouldn\u2019t have to be customized\/extended to\nnegotiate the maze.\n<\/p>\n<\/blockquote><p>\nThe <a href=\"https:\/\/golang.org\/\">Go<\/a> programming language for instance does not have such a tool as I know of.\n<\/p><p>\nA long-standing critique of Eshell is that it does not support input\nredirection.  This is hardly a major hindrance since you can always use a\nwork-around such as\n<\/p><div class=\"org-src-container\"><pre class=\"src src-sh\">cat file | command\n<\/pre><\/div><p>\nAnother common argument against Eshell is that it cannot run the widely popular\nTmux.  The latter is used for many things, out of which:\n<\/p><ul class=\"org-ul\"><li>Window management, tabs, etc.: you don\u2019t need that if you have Emacs.<\/li>\n<li>Leave long-running processes in the background after the user has logged out:\nthis feature is easily provided with <code>nohup<\/code> or <code>dtach<\/code>.<\/li>\n<\/ul><p>\nAll that being said, should Eshell not do the job for you, then maybe give <code>M-x\nshell<\/code> a try.  It has a similar design: it is a shell emancipated from any\nterminal; instead of being Lisp based, it runs an existing shell as an\ninterpreter.  <code>Bash<\/code> and <code>zsh<\/code> are known to work.  It has all the advantages an\nEmacs buffer has over terminal emulators.  The main difference then lies in the\nlanguage: <a href=\"https:\/\/www.gnu.org\/software\/emacs\/emacs-paper.html\">Emacs<\/a> <a href=\"http:\/\/paulgraham.com\/avg.html\">Lisp<\/a> vs.  <a href=\"http:\/\/mywiki.wooledge.org\/BashPitfalls\">Bash and related languages<\/a>.\n<\/p>","created_at":"2018-11-07T16:23:40+0000","updated_at":"2018-11-07T16:23:40+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"ambrevar.xyz","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":163,"title":"Multi-RBL Check Results | The Anti-Abuse Project","url":"http:\/\/www.anti-abuse.org\/multi-rbl-check-results\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:23:40+0000","updated_at":"2018-11-07T16:23:40+0000","annotations":[],"reading_time":0,"domain_name":"www.anti-abuse.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":159,"title":"Essays","url":"http:\/\/paulgraham.com\/articles.html","content":"<img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/sun.html\">General and Surprising<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/pow.html\">Charisma \/ Power<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/disc.html\">The Risk of Discovery<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/prop62.html\">This Year We Can End the Death Penalty in California<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/pgh.html\">How to Make Pittsburgh a Startup Hub<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/vb.html\">Life is Short<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ineq.html\">Economic Inequality<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/re.html\">The Refragmentation<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/jessica.html\">Jessica Livingston<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/bias.html\">A Way to Detect Bias<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/talk.html\">Write Like You Talk<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/aord.html\">Default Alive or Default Dead?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/safe.html\">Why It's Safe for Founders to Be Nice<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/name.html\">Change Your Name<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/altair.html\">What Microsoft Is this the Altair Basic of?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ronco.html\">The Ronco Principle<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/work.html\">What Doesn't Seem Like Work?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/corpdev.html\">Don't Talk to Corp Dev<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/95.html\">Let the Other 95% of Great Programmers In<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ecw.html\">How to Be an Expert in a Changing World<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/know.html\">How You Know<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/pinch.html\">The Fatal Pinch<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/mean.html\">Mean People Fail<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/before.html\">Before the Startup<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/fr.html\">How to Raise Money<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/herd.html\">Investor Herd Dynamics<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/convince.html\">How to Convince Investors<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ds.html\">Do Things that Don't Scale<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/invtrend.html\">Startup Investing Trends<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/startupideas.html\">How to Get Startup Ideas<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hw.html\">The Hardware Renaissance<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/growth.html\">Startup = Growth<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/swan.html\">Black Swan Farming<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/todo.html\">The Top of My Todo List<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/speak.html\">Writing and Speaking<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/property.html\">Defining Property<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ambitious.html\">Frighteningly Ambitious Startup Ideas<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/word.html\">A Word to the Resourceful<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/schlep.html\">Schlep Blindness<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/vw.html\">Snapshot: Viaweb, June 1998<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hubs.html\">Why Startup Hubs Work<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/patentpledge.html\">The Patent Pledge<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/airbnb.html\">Subject: Airbnb<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/control.html\">Founder Control<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/tablets.html\">Tablets<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/founders.html\">What We Look for in Founders<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/superangels.html\">The New Funding Landscape<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/seesv.html\">Where to See Silicon Valley<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hiresfund.html\">High Resolution Fundraising <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/yahoo.html\">What Happened to Yahoo <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/future.html\">The Future of Startup Funding <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/addiction.html\">The Acceleration of Addictiveness<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/top.html\">The Top Idea in Your Mind <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/selfindulgence.html\">How to Lose Time and Money <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/organic.html\">Organic Startup Ideas<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/apple.html\">Apple's Mistake<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/really.html\">What Startups Are Really Like<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/discover.html\">Persuade xor Discover <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/publishing.html\">Post-Medium Publishing<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/nthings.html\">The List of N Things<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/determination.html\">The Anatomy of Determination <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/kate.html\">What Kate Saw in Silicon Valley  <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/segway.html\">The Trouble with the Segway<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ramenprofitable.html\">Ramen Profitable<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/makersschedule.html\">Maker's Schedule, Manager's Schedule <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/revolution.html\">A Local Revolution?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/twitter.html\">Why Twitter is a Big Deal<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/foundervisa.html\">The Founder Visa<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/5founders.html\">Five Founders<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/relres.html\">Relentlessly Resourceful<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/angelinvesting.html\">How to Be an Angel Investor<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/convergence.html\">Why TV Lost<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/maybe.html\">Can You Buy a Silicon Valley?  Maybe.<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hackernews.html\">What I've Learned from Hacker News<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/13sentences.html\">Startups in 13 Sentences<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/identity.html\">Keep Your Identity Small  <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/credentials.html\">After Credentials<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/divergence.html\">Could VC be a Casualty of the Recession?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/highres.html\">The High-Res Society<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/artistsship.html\">The Other Half of \"Artists Ship\"  <\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/badeconomy.html\">Why to Start a Startup in a Bad Economy<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/fundraising.html\">A Fundraising Survival Guide<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/prcmc.html\">The Pooled-Risk Company Management Company<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/cities.html\">Cities and Ambition<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/distraction.html\">Disconnecting Distraction<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/lies.html\">Lies We Tell Kids<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/good.html\">Be Good<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/googles.html\">Why There Aren't More Googles<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/heroes.html\">Some Heroes<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/disagree.html\">How to Disagree<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/boss.html\">You Weren't Meant to Have a Boss<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ycombinator.html\">A New Venture Animal<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/trolls.html\">Trolls<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/newthings.html\">Six Principles for Making New Things<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/startuphubs.html\">Why to Move to a Startup Hub<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/webstartups.html\">The Future of Web Startups<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/philosophy.html\">How to Do Philosophy<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/colleges.html\">News from the Front<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/die.html\">How Not to Die<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/head.html\">Holding a Program in One's Head<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/stuff.html\">Stuff<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/equity.html\">The Equity Equation<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/unions.html\">An Alternative Theory of Unions<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/guidetoinvestors.html\">The Hacker's Guide to Investors<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/judgement.html\">Two Kinds of Judgement<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/microsoft.html\">Microsoft is Dead<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/notnot.html\">Why to Not Not Start a Startup<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/wisdom.html\">Is It Worth Being Wise?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/foundersatwork.html\">Learning from Founders<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/goodart.html\">How Art Can Be Good<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/startupmistakes.html\">The 18 Mistakes That Kill Startups<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/mit.html\">A Student's Guide to Startups<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/investors.html\">How to Present to Investors<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/copy.html\">Copy What You Like<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/island.html\">The Island Test<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/marginal.html\">The Power of the Marginal<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/america.html\">Why Startups Condense in America<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/siliconvalley.html\">How to Be Silicon Valley<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/startuplessons.html\">The Hardest Lessons for Startups to Learn<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/randomness.html\">See Randomness<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/softwarepatents.html\">Are Software Patents Evil?<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/6631327.html\">6,631,372<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/whyyc.html\">Why YC<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/love.html\">How to Do What You Love<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/procrastination.html\">Good and Bad Procrastination<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/web20.html\">Web 2.0<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/startupfunding.html\">How to Fund a Startup<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/vcsqueeze.html\">The Venture Capital Squeeze<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ideas.html\">Ideas for Startups<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/sfp.html\">What I Did this Summer<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/inequality.html\">Inequality and Risk<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ladder.html\">After the Ladder<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/opensource.html\">What Business Can Learn from Open Source<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hiring.html\">Hiring is Obsolete<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/submarine.html\">The Submarine<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/bronze.html\">Why Smart People Have Bad Ideas<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/mac.html\">Return of the Mac<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/writing44.html\">Writing,  Briefly<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/college.html\">Undergraduation<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/venturecapital.html\">A Unified Theory of VC Suckage<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/start.html\">How to Start a Startup<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hs.html\">What You'll Wish You'd Known<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/usa.html\">Made in USA<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/charisma.html\">It's Charisma, Stupid<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/polls.html\">Bradley's Ghost<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/laundry.html\">A Version 1.0<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/bubble.html\">What the Bubble Got Right<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/essay.html\">The Age of the Essay<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/pypar.html\">The Python Paradox<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/gh.html\">Great Hackers<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/gap.html\">Mind the Gap<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/wealth.html\">How to Make Wealth<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/gba.html\">The Word \"Hacker\"<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/say.html\">What You Can't Say<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/ffb.html\">Filters that Fight Back<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hp.html\">Hackers and Painters<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/iflisp.html\">If Lisp is So Great<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/hundred.html\">The Hundred-Year Language<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/nerds.html\">Why Nerds are Unpopular<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/better.html\">Better Bayesian Filtering<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/desres.html\">Design and Research<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/spam.html\">A Plan for Spam<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/icad.html\">Revenge of the Nerds<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/power.html\">Succinctness is Power<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/fix.html\">What Languages Fix<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/taste.html\">Taste for Makers<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/noop.html\">Why Arc Isn't Especially Object-Oriented<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/diff.html\">What Made Lisp Different<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/road.html\">The Other Road Ahead<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/rootsoflisp.html\">The Roots of Lisp<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/langdes.html\">Five Questions about Language Design<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/popular.html\">Being Popular<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/javacover.html\">Java's Cover<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/avg.html\">Beating the Averages<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/lwba.html\">Lisp for Web-Based Applications<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"https:\/\/sep.yimg.com\/ty\/cdn\/paulgraham\/acl1.txt?t=1536426018&amp;\">Chapter 1 of Ansi Common Lisp<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"https:\/\/sep.yimg.com\/ty\/cdn\/paulgraham\/acl2.txt?t=1536426018&amp;\">Chapter 2 of Ansi Common Lisp<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"5\" width=\"1\" border=\"0\" alt=\"image\" \/><img src=\"https:\/\/sep.yimg.com\/ca\/I\/paulgraham_2272_1423\" width=\"12\" height=\"14\" border=\"0\" alt=\"image\" style=\"float: left; margin-left: 0px; margin-right: 0px; margin-top: 0px; margin-bottom: 0px;\" \/><span><a href=\"http:\/\/paulgraham.com\/progbot.html\">Programming Bottom-Up<\/a><img src=\"https:\/\/sep.yimg.com\/ca\/Img\/trans_1x1.gif\" height=\"2\" width=\"1\" border=\"0\" alt=\"image\" \/><br \/><\/span>","created_at":"2018-11-07T16:23:39+0000","updated_at":"2018-11-07T16:23:39+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"paulgraham.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":160,"title":"The Many Uses of Org Mode","url":"https:\/\/robots.thoughtbot.com\/the-many-uses-of-org-mode","content":"<header>\n<\/header><div class=\"post-body\"><p>I\u2019m a relatively recent convert to GNU Emacs; \n<a href=\"https:\/\/github.com\/haskell\/haskell-mode\">Haskell mode<\/a>\nis what made me make the switch, but <a href=\"https:\/\/orgmode.org\/\">Org mode<\/a>\nis what has kept me here.<\/p><p>I\u2019ve tried several applications for todo management before,\nbut have not been able to make one stick.\nOrg mode works for me because I spend a large portion of my day in my\neditor anyway, so there is minimal friction in adding to or checking it.<\/p><p>On client work, I use Org mode in combination with\n<a href=\"https:\/\/github.com\/IvanMalison\/org-projectile\">org-projectile<\/a> to get\nthings out of my head and onto paper. As soon as I feel the need to \n\u2018remember to do something later\u2019, I use org-projectile to open an Org file \nscoped to my current git repository, and jot it down. I\u2019ve noticed an \nimprovement in my focus because of the lack of context switch. I also use \norg-projectile to leave myself todos at the end of each day,\nmaking picking up where I left off much smoother.<\/p><p><img src=\"https:\/\/images.thoughtbot.com\/blog-vellum-image-uploads\/eD2jsIqRQ5ekDMy5Kdvw_Screen%20Shot%202018-05-07%20at%206.30.16%20PM.png\" alt=\"\" \/><\/p><p>When I\u2019m advising a project, I  use it during weekly retrospectives to take \nnotes and capture action items. Org comes with extensive \n<a href=\"https:\/\/orgmode.org\/manual\/Exporting.html\">exporting<\/a> functionality -\nI usually export to markdown, then share a \n<a href=\"https:\/\/gist.github.com\/sidraval\/58f252f00099d0a5486b28b33f1edd4c\">gist of the retro notes<\/a>\nwith the team, complete with checkboxes for action items! <\/p><p>I also use Org mode in my life outside of work. I use it for all my personal \ntodos \u2013 Org mode comes with the ability to schedule tasks, and \n<a href=\"https:\/\/orgmode.org\/worg\/org-tutorials\/orgtutorial_dto.html#org171994c\">Org agenda<\/a>\nallows me to see only the tasks that are due for the current week.\nIt alleviates the feeling of overwhelmedness I get when looking at a huge,\nnever ending list of tasks.<\/p><p>When I finish tasks, I move them to a DONE.org file so I have a list of \nall my completed items, dated and with any comments I may have thought \nto leave myself:<\/p><p><img src=\"https:\/\/images.thoughtbot.com\/blog-vellum-image-uploads\/NgcO96YTde2bcFrLRu2I_Screen%20Shot%202018-05-07%20at%206.42.21%20PM.png\" alt=\"\" \/><\/p><p>And of course, no todo list would be complete without a mobile application;\nI use <a href=\"https:\/\/github.com\/orgzly\/orgzly-android\">Orgzly<\/a> along with Dropbox\nto keep my org files synced across devices.<\/p><p>I\u2019m writing this blog post at a hack night during\n<a href=\"http:\/\/emacsnyc.org\/\">Emacs NYC<\/a>. I asked attendees to come mention their\ncreative uses of Org mode, and lo and behold <a href=\"http:\/\/zck.me\/\">Zachary<\/a>\nshowed me a package he wrote called\n<a href=\"https:\/\/bitbucket.org\/zck\/zpresent.el\">zpresent<\/a>. He writes presentations,\nconference talks, etc. as plain org files and then uses <code>zpresent<\/code> to transform\nthem into slides that he can present from within Emacs.<\/p><p>If managing your life inside a text editor appeals to you, I highly recommend giving\nOrg mode a try.\n<a href=\"https:\/\/robots.thoughtbot.com\/getting-started-with-org-mode\">Harry Schwartz presented a great intro<\/a>\nat the Boston Emacs meetup that serves as a good entry point!<\/p><\/div><aside class=\"related-articles\">\n<ol><li>\n<a href=\"https:\/\/robots.thoughtbot.com\/and-then-there-was\">And Then There Was...<\/a>\n<\/li>\n<li>\n<a href=\"https:\/\/robots.thoughtbot.com\/paper\">Paper<\/a>\n<\/li>\n<li>\n<a href=\"https:\/\/robots.thoughtbot.com\/tips-for-clojure-beginners\">Tips for Clojure Beginners<\/a>\n<\/li>\n<\/ol><\/aside>","created_at":"2018-11-07T16:23:39+0000","updated_at":"2018-11-07T16:23:39+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"robots.thoughtbot.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":157,"title":"Escaping the Competitive Herd: Youngme Moon: 8601400553657: Amazon.com: Books","url":"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon\/dp\/030746086X","content":"<div id=\"universal-detail-ilm\"><div class=\"a-section a-spacing-none uilm-section\"><a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/b?node=17276826011\">\n      <img alt=\"\" src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/G\/01\/kindle\/merch\/2018\/DCLP\/Drivers\/KCD-1231_DCLP_UILM-Desktop_650x45._CB480319419_.jpg\" height=\"45\" width=\"650\" data-a-hires=\"https:\/\/images-na.ssl-images-amazon.com\/images\/G\/01\/kindle\/merch\/2018\/DCLP\/Drivers\/KCD-1231_DCLP_UILM-Desktop_1300x90._CB480319419_.jpg\" \/><\/a><\/div><\/div><div id=\"rightCol\"><div id=\"unifiedBuyBox_feature_div\" class=\"feature\" data-feature-name=\"unifiedBuyBox\"><div id=\"combinedBuyBox\" class=\"a-section a-spacing-medium\"><form method=\"post\" id=\"addToCart\" action=\"\/gp\/product\/handle-buy-box\/ref=dp_start-bbf_1_glance\" class=\"a-content\">\n            \n            \n    <div id=\"buybox\" class=\"a-row a-spacing-medium\"><div id=\"rbbContainer\" class=\"a-box-group dp-accordion\"><div class=\"a-box rbbSection selected dp-accordion-active\"><div class=\"a-box-inner\"><div class=\"a-section a-spacing-none a-padding-none\"><div id=\"buyNewInner\" class=\"rbbContent dp-accordion-inner\"><div id=\"buyBoxInner\" class=\"a-section a-spacing-none\"><div class=\"a-row\"><div class=\"a-column a-span7 a-text-right a-span-last\"><ul class=\"a-unordered-list a-nostyle a-vertical\"><li>\n                    \n    \n    \n    \n        \n        \n        \n            \n            \n            \n        \n    \n        \n        \n        \n    \n        \n        \n        \n    \n    \n    \n    \n        \n        \n            \nList Price: $16.00\n        \n    \n                <\/li>\n            \n            \n        \n        \n        \n        \n        \n            <li>\n                \n                    Save: $3.21\n                     (20%)\n                \n            <\/li>\n        \n    \n    \n        \n        \n    \n    \n        \n        \n        \n        \n        \n        \n        \n        \n        \n        \n        \n    <\/ul><\/div><div class=\"a-column a-span5\"><div id=\"selectQuantity\" class=\"a-section a-spacing-none a-padding-none\"><div class=\"a-row a-spacing-base\"><p><label for=\"quantity\" class=\"a-native-dropdown\">Qty:<\/label>Qty:1<i class=\"a-icon a-icon-dropdown\"><\/i><\/p><\/div><\/div><\/div><\/div><\/div><div class=\"a-section a-spacing-small a-spacing-top-micro\"><p><a href=\"https:\/\/www.amazon.com\/gp\/help\/customer\/display.html\/ref=mk_sss_dp_1\/?ie=UTF8&amp;pop-up=1&amp;nodeId=527692\" class=\"\" target=\"AmazonHelp\">FREE Shipping<\/a>.<\/p><\/div><div class=\"a-section a-spacing-small\"><p>In Stock.<\/p><p>Ships from and sold by Amazon.com.\n    \n    \n        \n        \n        \n        \n        \n        \n        \n            Gift-wrap available.<\/p><\/div><div id=\"bbopAndCartBox\" class=\"a-box\"><div class=\"a-box-inner\"><p><i class=\"a-icon a-icon-cart\">Add to Cart<\/i><\/p><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"a-box rbbSection unselected\"><div class=\"a-box-inner\"><div class=\"a-section a-spacing-none a-padding-none\"><div id=\"usedbuyBox\" class=\"a-section rbbContent dp-accordion-inner\"><div class=\"a-section a-spacing-mini\"><p><a href=\"https:\/\/www.amazon.com\/gp\/help\/customer\/display.html\/ref=mk_sss_dp_1?ie=UTF8&amp;pop-up=1&amp;nodeId=527692\" target=\"AmazonHelp\">FREE Shipping<\/a><\/p><\/div><div class=\"a-section a-spacing-base\"><p><strong> \n    \n    \n    \n    \n        \n    \n    \n    \n    \nUsed: \n    \n        \n    \n    \n    \n    \n    \nVery Good\n\t\t<\/strong>\n\t\t\n\t\t\t | \n\t\t\t<a id=\"usedItemConditionInfoLink\" class=\"a-link-normal\" href=\"#\">\n\t\t\t\tDetails\n\t\t\t<\/a><\/p><p>Sold by\n\t\t\n\t\t\t\n\t\t\t\n\t\t\t\t<a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/help\/seller\/at-a-glance.html?ie=UTF8&amp;seller=A6Y9RGWJY2DRD&amp;isAmazonFulfilled=1\">Deadstock_Gear<\/a><\/p><p><a id=\"SSOFpopoverLink_ubb\" class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/help\/customer\/display.html?ie=UTF8&amp;ref=dp_ubb_fulfillment&amp;nodeId=106096011\">\n\t\t\t\tFulfilled by Amazon\n\t\t\t<\/a><\/p><\/div><div class=\"a-popover-preload\" id=\"a-popover-usedItemConditionDetailsPopover\"><p><strong>Condition:<\/strong>\n      Used: Very Good<\/p><p><strong>Comment:<\/strong> Unmarked copy. Cover shows some wear.<\/p><\/div><div class=\"a-popover-preload\" id=\"a-popover-SSOFpopoverLink_ubb-content\"><p>Fulfillment by Amazon (FBA) is a service we offer sellers that lets them store their products in Amazon's fulfillment centers, and we directly pack, ship, and provide customer service for these products. Something we hope you'll especially enjoy: <em>FBA items qualify for FREE Shipping and Amazon Prime.<\/em><\/p><p>If you're a seller, Fulfillment by Amazon can help you increase your sales. We invite you to <a class=\"a-link-normal\" target=\"_blank\" rel=\"noopener\" href=\"https:\/\/www.amazon.com\/gp\/seller\/fba\/fulfillment-by-amazon.html?ie=UTF8&amp;ld=AZFBADP&amp;ref=dp_fba_pop\">\n    learn more about Fulfillment by Amazon\n  <\/a>.<\/p><\/div><p><i class=\"a-icon a-icon-cart\">Add to Cart<\/i><\/p><\/div><\/div><\/div><\/div><\/div><\/div>\n                    \n                    \n                    \n                    \n                        \n                        \n                        \n                        \n                    \n                \n                \n            \n        <\/form><\/div><\/div><div id=\"moreBuyingChoices_feature_div\" class=\"feature\" data-feature-name=\"moreBuyingChoices\"><div class=\"a-section a-spacing-medium a-spacing-top-base\"><div id=\"mbc-action-panel-wrapper\"><div id=\"mbc\" data-asin=\"030746086X\" data-brand=\"Crown Business\" data-product-group=\"book_display_on_website\" data-timeout=\"Sorry we encountered a problem.\" class=\"a-box-group\"><div class=\"a-box a-text-center\"><p><\/p><h5>\n                                    Other Sellers on Amazon\n                                <\/h5><\/div><div class=\"a-box mbc-offer-row pa_mbc_on_amazon_offer\"><div class=\"a-box-inner a-padding-small\"><a id=\"mbc-buybutton-addtocart-1-announce\" href=\"https:\/\/www.amazon.com\/gp\/item-dispatch\/ref=dp_ebb_1?ie=UTF8&amp;registryItemID.1=&amp;submit.addToCart=addToCart&amp;registryID.1=&amp;offeringID.1=V1h%252FIROAsZydqxJdVOeEiaJJ1VzoFg6Fmf0BLfMktnWMf%252FBPZcxg%252Fl2QpByzLV%252FgaEZOhex4%252BVOEwaiy4TsAe370%252Fj4guCTBOrQW%252F0rFolDn2zvrUPwrxxGkgkckMmeZM3EprgjZkkPG2nB6zPLvSw%253D%253D&amp;storeID=books&amp;nodeID=&amp;signInToHUC=0&amp;session-id=000-0000000-0000000&amp;itemCount=1\" class=\"a-button-text\" role=\"button\">Add to Cart<\/a><div class=\"a-row\"><p>$11.31\n                \n                \n                \n                <br \/>+\u00a0Free Shipping<\/p><p>Sold by:\n        \n        \n            kozybooks85<\/p><\/div><\/div><\/div><div class=\"a-box mbc-offer-row pa_mbc_on_amazon_offer\"><div class=\"a-box-inner a-padding-small\"><a id=\"mbc-buybutton-addtocart-2-announce\" href=\"https:\/\/www.amazon.com\/gp\/item-dispatch\/ref=dp_ebb_2?ie=UTF8&amp;registryItemID.1=&amp;submit.addToCart=addToCart&amp;registryID.1=&amp;offeringID.1=V1h%252FIROAsZydqxJdVOeEiaJJ1VzoFg6FwstETWOrlxdoM0OfwzLNdikljWu06iGl9e5QBIB1G2xWZZ4bC5hFJqj024ZrwFof5tq2Jfa81BnTjyiag6UWFcuiKaNLQQr7O%252FfRfDsQChXLSWdnZTUrTnZbYeJpu9lb&amp;storeID=books&amp;nodeID=&amp;signInToHUC=0&amp;session-id=000-0000000-0000000&amp;itemCount=1\" class=\"a-button-text\" role=\"button\">Add to Cart<\/a><div class=\"a-row\"><p>$7.33\n                \n                \n                \n                <br \/>+\u00a0$3.99\u00a0shipping<\/p><p>Sold by:\n        \n        \n            PHIMARREM Enterprises LLC<\/p><\/div><\/div><\/div><div class=\"a-box mbc-offer-row pa_mbc_on_amazon_offer\"><div class=\"a-box-inner a-padding-small\"><a id=\"mbc-buybutton-addtocart-3-announce\" href=\"https:\/\/www.amazon.com\/gp\/item-dispatch\/ref=dp_ebb_3?ie=UTF8&amp;registryItemID.1=&amp;submit.addToCart=addToCart&amp;registryID.1=&amp;offeringID.1=V1h%252FIROAsZydqxJdVOeEiaJJ1VzoFg6FTKhdpHC5ApZyT0bLY%252FeiOLfbCs3Qcr5JRGoDR6HLeFfqrBvA15UVHxKHXNp0Dq3sHUkPJ3e5w8kK0I9Y7OEglmDwZodFDgkXcDszQGm18flUIJz7hFSpURFpYlf7ZpvQ&amp;storeID=books&amp;nodeID=&amp;signInToHUC=0&amp;session-id=000-0000000-0000000&amp;itemCount=1\" class=\"a-button-text\" role=\"button\">Add to Cart<\/a><div class=\"a-row\"><p>$7.53\n                \n                \n                \n                <br \/>+\u00a0$4.21\u00a0shipping<\/p><p>Sold by:\n        \n        \n            u_pick<\/p><\/div><\/div><\/div><\/div><\/div><p>Have one to sell?\n        <a href=\"https:\/\/catalog-retail.amazon.com\/abis\/syh\/DisplayCondition?_encoding=UTF8&amp;ld=AMZDP&amp;coliid=&amp;asin=030746086X&amp;colid=&amp;qid=&amp;sr=\" class=\"a-button-text\" role=\"button\">\n            Sell on Amazon\n        <\/a><\/p><\/div><\/div><\/div><div id=\"leftCol\"><div id=\"booksImageBlock_feature_div\" class=\"feature\" data-feature-name=\"booksImageBlock\"><div id=\"imageBlockOuter\" class=\"a-row\"><div id=\"sitbLogo\"><img id=\"sitbLogoImg\" src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/G\/01\/digital\/sitb\/sticker\/sitb-sticker-v3-small._CB341765094_.png\" class=\"a-hidden litb-on-click\" alt=\"image\" \/><\/div><div id=\"imageBlock\" class=\"a-row\"><div id=\"imageBlockContainer\" class=\"a-row a-grid-vertical-align a-grid-center maintain-height\"><div id=\"main-image-container\" class=\"a-column a-span12 a-text-center maintain-height a-span-last\"><div id=\"img-wrapper\" class=\"maintain-height\"><div id=\"img-canvas\" class=\"maintain-height litb-on-click\"><img alt=\"\" src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/I\/41yeUrnC1vL._SY291_BO1,204,203,200_QL40_.jpg\" class=\"a-dynamic-image image-stretch-vertical frontImage\" id=\"imgBlkFront\" data-a-dynamic-image=\"{&quot;https:\/\/images-na.ssl-images-amazon.com\/images\/I\/41yeUrnC1vL._SY344_BO1,204,203,200_.jpg&quot;:[225,346],&quot;https:\/\/images-na.ssl-images-amazon.com\/images\/I\/41yeUrnC1vL._SX322_BO1,204,203,200_.jpg&quot;:[324,499]}\" \/><\/div><\/div><\/div><\/div><div id=\"flipAndSampleAudio\" class=\"a-row\"><p>Flip to back\n          Flip to front<\/p><p>Listen\n            \n          \n          \n            \n              \n              Playing...\n            \n          \n          \n            \n              \n              Paused\n            \n          \n        \n        \u00a0\n        \n          You're listening to a sample of the Audible audio edition.\n          <br \/><a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/dp\/B003FOOG6W\/ref=dp_bookdesc_audio\">Learn more<\/a><\/p><\/div><\/div><div id=\"imageBlockThumbs\" class=\"a-row a-spacing-mini a-spacing-top-micro\"><div class=\"a-column a-span3 a-spacing-micro imageThumb thumb\"><img src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/I\/41yeUrnC1vL._AC_SX60_CR,0,0,60,60_.jpg\" alt=\"image\" \/><\/div><div class=\"a-column a-span3 a-spacing-micro imageThumb thumb\"><img src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/I\/516R5GH7ZRL._AC_SX60_CR,0,0,60,60_.jpg\" alt=\"image\" \/><\/div>See all 2 images<\/div><\/div><\/div><\/div><div id=\"centerCol\" class=\"centerColumn\"><div id=\"averageCustomerReviews_feature_div\" class=\"feature\" data-feature-name=\"averageCustomerReviews\"><p><a href=\"denied:javascript:void(0)\" class=\"a-popover-trigger a-declarative\">\n                \n<i class=\"a-icon a-icon-star a-star-4-5\">4.3 out of 5 stars<\/i>\n                \n            <i class=\"a-icon a-icon-popover\"><\/i><\/a>\n        \n        \n    \n        \n        \n        \n        \n        \n        \n            \n        \n        \n        \n        \n        \n        \n        \n        \n            \n            \n                \n                    <a id=\"acrCustomerReviewLink\" class=\"a-link-normal\" href=\"#customerReviews\">\n                        114 customer reviews\n                    <\/a><\/p><\/div><hr \/><div id=\"MediaMatrix\" class=\"feature\" data-feature-name=\"MediaMatrix\"><div id=\"formats\" class=\"a-section a-spacing-large responsive\"><a class=\"a-link-expander a-spacing-top-micro a-spacing-small a-size-small\">\n    \n    \n        \n            <i id=\"formatsIcon\" class=\"a-icon a-icon-arrow\">\n        \n        \n    \n    See all 5 formats and editions\n    Hide other formats and editions\n<\/i><\/a><div id=\"twister\" class=\"a-section a-spacing-base a-spacing-top-small a-size-mini\"><table class=\"a-normal a-spacing-none title\"><tr><td class=\"dp-title-col\">\n            \n        <\/td>\n            \n        \n        <td class=\"a-color-secondary a-text-right dp-price-col\">\n       \t\t \n        <\/td>\n        \n            <td class=\"a-color-secondary a-text-right dp-new-col\">\n              New from\n            <\/td>\n        \n        \n            <td class=\"a-color-secondary a-text-right dp-used-col\">\n              Used from\n            <\/td>\n        \n        \n      <\/tr><\/table><div class=\"top-level unselected-row\"><table class=\"a-normal a-spacing-none\"><tr><td class=\"dp-title-col\">\n        \n        \n        \n            \n            \n            \n                \n            \n            \n          \n            \n                \n                    \n                    \n                        <a href=\"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon-ebook\/dp\/B0036S4CNE\/ref=tmm_kin_title_0?_encoding=UTF8&amp;qid=&amp;sr=\" class=\"title-text\">\n                            Kindle\n                            \n                            \n                            \n                        <\/a>\n                        \n                   \n                \n            \n            \n                <div class=\"a-box a-alert-inline a-alert-inline-error\" aria-live=\"assertive\" role=\"alert\"><div class=\"a-box-inner a-alert-container\"><i class=\"a-icon a-icon-alert\">\"Please retry\"<\/i><\/div><\/div>\n            \n        <\/td>\n         \n        \n        \n        \n\t        \n\t        \n\t        \n\t        \n        \n        \n        <td class=\"a-text-right dp-price-col\">\n            <div class=\"a-fixed-right-grid\"><div class=\"a-fixed-right-grid-inner\"><p><a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon-ebook\/dp\/B0036S4CNE\/ref=tmm_kin_title_0?_encoding=UTF8&amp;qid=&amp;sr=\">\n                                         \n                                         \n                                         \n                                         $11.99 \n                                     <\/a><\/p><\/div><\/div>\n        <\/td>\n        \n          \n          \n          \n                \n                    \n                    \n                        <td class=\"a-color-tertiary a-text-right dp-new-col\">\n                            \u2014\n                        <\/td>\n                    \n                \n            \n            \n                \n                    \n                    \n                        <td class=\"a-color-tertiary a-text-right dp-used-col\">\n                            \u2014\n                        <\/td>\n                    \n                \n            \n            \n          \n        \n    <\/tr><\/table><\/div><div class=\"top-level unselected-row\"><table class=\"a-normal a-spacing-none\"><tr><td class=\"dp-title-col\">\n        \n        \n        \n            \n            \n                <a class=\"a-link-expander title-expander\">\n                    \n                    \n                    \n                      \n                      \n                          \n                              \n                                  <i id=\"editionsIcon_2\" class=\"a-icon a-icon-arrow tmm-row-icon\">\n                              \n                              \n                          \n                      \n                       \n                    \n                <\/i><\/a>\n            \n            \n            \n          \n            \n                \n                    \n                    \n                        <a href=\"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon\/dp\/0307460851\/ref=tmm_hrd_title_0?_encoding=UTF8&amp;qid=&amp;sr=\" class=\"title-text\">\n                            Hardcover\n                            \n                            \n                            \n                        <\/a>\n                        \n                   \n                \n            \n            \n                <div class=\"a-box a-alert-inline a-alert-inline-error\" aria-live=\"assertive\" role=\"alert\"><div class=\"a-box-inner a-alert-container\"><i class=\"a-icon a-icon-alert\">\"Please retry\"<\/i><\/div><\/div>\n            \n        <\/td>\n         \n        \n        \n        \n\t        \n\t        \n\t        \n\t        \n        \n        \n        <td class=\"a-text-right dp-price-col\">\n            <div class=\"a-fixed-right-grid\"><div class=\"a-fixed-right-grid-inner\"><p><a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon\/dp\/0307460851\/ref=tmm_hrd_title_0?_encoding=UTF8&amp;qid=&amp;sr=\">\n                                         \n                                         \n                                         \n                                         $29.58 \n                                     <\/a><\/p><\/div><\/div>\n        <\/td>\n        \n          \n          \n          \n                \n                    \n                        <td class=\"a-text-right dp-new-col\">\n                          <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/0307460851\/ref=tmm_hrd_new_olp_0?ie=UTF8&amp;condition=new&amp;qid=&amp;sr=\">\n                              $27.72\n                          <\/a>\n                        <\/td>\n                    \n                    \n                \n            \n            \n                \n                    \n                        <td class=\"a-text-right dp-used-col\">\n                           <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/0307460851\/ref=tmm_hrd_used_olp_0?ie=UTF8&amp;condition=used&amp;qid=&amp;sr=\">\n                               $1.42\n                           <\/a>\n                        <\/td>\n                    \n                    \n                \n            \n            \n          \n        \n    <\/tr><\/table><\/div><div class=\"top-level selected-row\"><table class=\"a-normal a-spacing-none\"><tr><td class=\"dp-title-col\">\n        \n        \n        \n            \n            \n            \n                \n            \n            \n          \n            \n                \n                    \n                        \n                            Paperback\n                            \n                            \n                        \n                    \n                    \n                \n            \n            \n                <div class=\"a-box a-alert-inline a-alert-inline-error\" aria-live=\"assertive\" role=\"alert\"><div class=\"a-box-inner a-alert-container\"><i class=\"a-icon a-icon-alert\">\"Please retry\"<\/i><\/div><\/div>\n            \n        <\/td>\n         \n        \n        \n        \n\t        \n\t        \n\t        \n\t        \n        \n        \n        <td class=\"a-text-right dp-price-col\">\n            <div class=\"a-fixed-right-grid\"><div class=\"a-fixed-right-grid-inner\"><p>$12.79<\/p><\/div><\/div>\n        <\/td>\n        \n          \n          \n          \n                \n                    \n                        <td class=\"a-text-right dp-new-col\">\n                          <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/030746086X\/ref=tmm_pap_new_olp_sr?ie=UTF8&amp;condition=new&amp;qid=&amp;sr=\">\n                              $6.00\n                          <\/a>\n                        <\/td>\n                    \n                    \n                \n            \n            \n                \n                    \n                        <td class=\"a-text-right dp-used-col\">\n                           <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/030746086X\/ref=tmm_pap_used_olp_sr?ie=UTF8&amp;condition=used&amp;qid=&amp;sr=\">\n                               $1.73\n                           <\/a>\n                        <\/td>\n                    \n                    \n                \n            \n            \n          \n        \n    <\/tr><\/table><\/div><div class=\"top-level unselected-row\"><table class=\"a-normal a-spacing-none\"><tr><td class=\"dp-title-col\">\n        \n        \n        \n            \n            \n            \n                \n            \n            \n          \n            \n                \n                    \n                    \n                        <a href=\"https:\/\/www.amazon.com\/Different-Escaping-the-Competitive-Herd\/dp\/B003FOOG6W\/ref=tmm_aud_title_0?_encoding=UTF8&amp;qid=&amp;sr=\" class=\"title-text\">\n                            Audible Audiobook, Unabridged\n                            \n                            \n                            \n                        <\/a>\n                        \n                   \n                \n            \n            \n                <div class=\"a-box a-alert-inline a-alert-inline-error\" aria-live=\"assertive\" role=\"alert\"><div class=\"a-box-inner a-alert-container\"><i class=\"a-icon a-icon-alert\">\"Please retry\"<\/i><\/div><\/div>\n            \n        <\/td>\n         \n        \n        \n        \n\t        \n\t        \n\t        \n\t        \n        \n        \n        <td class=\"a-text-right dp-price-col\">\n            \n        <\/td>\n        \n          \n            \n          \n          \n        \n    <\/tr><\/table><\/div><\/div><div id=\"tmmSwatches\" class=\"a-row nonJSFormats\"><ul class=\"a-unordered-list a-nostyle a-button-list a-horizontal\"><li class=\"swatchElement unselected\">\n\t                \n     \n<a href=\"https:\/\/www.amazon.com\/Different-Escaping-the-Competitive-Herd\/dp\/B003FOOG6W\/ref=tmm_aud_swatch_0?_encoding=UTF8&amp;qid=&amp;sr=\" class=\"a-button-text\" role=\"button\">\n  \n  \n    \n        \n      \n<img src=\"https:\/\/m.media-amazon.com\/images\/G\/01\/Audible\/Homestead\/logo\/audiblelogo.png\" alt=\"image\" \/> Audiobook\n<br \/>\n    \n    \n    \n    \n      \n          \n        \n  \n  \n    \n      \n        $0.00\n      \n      \n    \n  \n  \n    \n      \n      \n      \n        \n  \n    \n      \n    \n    \n    \n      \n \n     \n    \n  \n<\/a>\n\t                \n\t                  \n        \n\t                  \n    \n        \n    \n    \n\n                      \n\t\n\t\t\n      \t\t\n        \t\t\n          \t\t\t<a class=\"a-size-mini a-link-normal\" href=\"https:\/\/www.amazon.com\/Different-Escaping-the-Competitive-Herd\/dp\/B003FOOG6W\/ref=tmm_aud_title_0?_encoding=UTF8&amp;qid=&amp;sr=\">\n            \t\t\tFree with your Audible trial\n          \t\t\t<\/a>\n        \t\t\n\t      \t\n                \n                \n\t\t\n \t\n \n                      \n    \n                    \n\t              <\/li>\n\t            \n\t            \n\t            \n                \n\t          \n\t        \n\t      \n\t    \n\t      \n            \n\t        \n\t          \n\t          \n\t            \n\t            \n                \n\t            \n\t              <li class=\"swatchElement unselected\">\n\t                \n     \n<a href=\"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon-ebook\/dp\/B0036S4CNE\/ref=tmm_kin_swatch_0?_encoding=UTF8&amp;qid=&amp;sr=\" class=\"a-button-text\" role=\"button\">\n  \n  \n    \n        \n      \n \nKindle <br \/>$11.99\n    \n    \n  \n  \n  \n    \n      \n      \n      \n        \n    \n      \n    \n    \n    \n  \n<\/a>\n\t                \n\t                  \n        \n\t                  \n    \n        \n          \n          \n          \n          \n          \n          \n          \n            \n          \n          \n            \n          \n          \n          \n          \n          \n          \n          \n            \n        \n    \n    \n\n    \n    \n        \n        \n            \n        \n        \n    \n    \n        \n        \n        \n        \n        \n            \n                \n                \n                <a id=\"kcpAppsPopOver\" class=\"a-size-mini a-link-normal kcpAppsPopOver\" href=\"denied:javascript:void(0);\">\n                    \n                        \n                            Read with Our\n                            \n                                Free App\n                            \n                        \n                    \n                <\/a>\n            \n        \n    \n                      \n\t\n\t\t\n      \t\t\n                \n                  \n                    \n                        \n                            \n                          \n \n  \n  \n  \n    \n                        \n                      \n                    \n                    \n                  \n                \n                \n\t\t\n \t\n \n                      \n    \n                    \n\t              <\/li>\n\t            \n\t            \n\t            \n                \n\t          \n\t        \n\t      \n\t    \n\t      \n            \n\t        \n\t          \n\t          \n\t            \n\t            \n                \n\t            \n\t              <li class=\"swatchElement unselected\">\n\t                \n     \n<a href=\"https:\/\/www.amazon.com\/Different-Escaping-Competitive-Youngme-Moon\/dp\/0307460851\/ref=tmm_hrd_swatch_0?_encoding=UTF8&amp;qid=&amp;sr=\" class=\"a-button-text\" role=\"button\">\n  \n  \n    \n        \n      \n \nHardcover <br \/>$29.58\n    \n    \n  \n  \n  \n    \n      \n      \n      \n        \n    \n      \n    \n    \n    \n      \n \n     \n    \n  \n<\/a>\n\t                \n\t                  \n        \n\t                  \n    \n        \n    \n    \n\n                      \n\t\n\t\t\n      \t\t\n                \n                  \n                    \n                        \n                            \n                          \n \n  \n    \n      <a class=\"a-size-mini a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/0307460851\/ref=tmm_hrd_used_olp_0?ie=UTF8&amp;condition=used&amp;qid=&amp;sr=\">\n        \n          46 Used from $1.42\n        \n      <\/a>        \n    \n  \n  \n    \n      <a class=\"a-size-mini a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/0307460851\/ref=tmm_hrd_new_olp_0?ie=UTF8&amp;condition=new&amp;qid=&amp;sr=\">\n        \n          6 New from $27.72\n        \n      <\/a>        \n    \n  \n  \n    \n                        \n                      \n                    \n                    \n                  \n                \n                \n\t\t\n \t\n \n                      \n    \n                    \n\t              <\/li>\n\t            \n\t            \n\t            \n                \n\t          \n\t        \n\t      \n\t    \n\t      \n            \n\t        \n\t          \n\t            \n\t            \n                \n\t            \n\t              <li class=\"swatchElement selected\">\n\t                \n     \n<a href=\"denied:javascript:void(0)\" class=\"a-button-text\" role=\"button\">\n  \n  \n    \n        \n      \n \nPaperback <br \/>$12.79\n    \n    \n  \n  \n  \n    \n      \n      \n      \n        \n    \n      \n    \n    \n    \n      \n \n     \n    \n  \n<\/a>\n\t                \n\t                  \n        \n\t                  \n    \n        \n    \n    \n\n\t                  \n\t\n\t\t\n      \t\t\n                \n                  \n                    \n                        \n                            \n                          \n \n  \n    \n      <a class=\"a-size-mini a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/030746086X\/ref=tmm_pap_used_olp_sr?ie=UTF8&amp;condition=used&amp;qid=&amp;sr=\">\n        \n          58 Used from $1.73\n        \n      <\/a>        \n    \n  \n  \n    \n      <a class=\"a-size-mini a-link-normal\" href=\"https:\/\/www.amazon.com\/gp\/offer-listing\/030746086X\/ref=tmm_pap_new_olp_sr?ie=UTF8&amp;condition=new&amp;qid=&amp;sr=\">\n        \n          31 New from $6.00\n        \n      <\/a>        \n    \n  \n  \n    \n                        \n                      \n                    \n                    \n                  \n                \n                \n\t\t\n \t\n \n\t                  \n    \n\t                \n\t              <\/li>\n\t            \n\t            \n\t            \n                \n\t          \n\t          \n\t        \n\t      \n\t     \n      \n      \n    \n  \n<\/ul><\/div><\/div><\/div><div id=\"bookDescription_feature_div\" class=\"feature\" data-feature-name=\"bookDescription\"><noscript>\n \t<div><b>What if working like crazy to beat the competition did exactly the opposite, making you mediocre and more like the competition?<\/b><p>In today\u2019s world of overabundant consumer choices and superfluous apps, upgrades, add-ons, and features, brands have become nearly identical, as their efforts to outdo one another have pushed them into a dizzying herd of indistinct options. <\/p><p>Youngme Moon identifies the outliers, the mavericks, the iconoclasts\u2014the players who have thoughtfully rejected orthodoxy in favor of an approach that is more adventurous. Some are even \u201chostile,\u201d almost daring you to buy what they are selling.\u00a0<\/p><p>Using her original research on companies such as IKEA and Google, Moon will inspire you to be counterintuitive and\u00a0<i>meaningfully<\/i>\u00a0different\u2014to rethink your business strategy, to stop conforming and start deviating, to stop emulating and start innovating. Because to stand out you must become the exception, not the rule.<\/p><\/div>\n \t\n <\/noscript><\/div><\/div><div id=\"descriptionAndDetails\" class=\"a-section a-spacing-extra-large\"><div id=\"productDescription_feature_div\" class=\"feature\" data-feature-name=\"productDescription\"><div data-feature-name=\"productDescription\" data-template-name=\"productDescription\" class=\"a-row feature\"><h2 class=\"books\">\n                    Editorial Reviews\n                    \n                 <\/h2><div id=\"productDescription\" class=\"a-section a-spacing-small prodDescWidth\"><h3>Review<\/h3><div aria-live=\"polite\" data-a-expander-collapsed-height=\"300\" class=\"a-expander-collapsed-height a-row a-expander-container a-expander-partial-collapse-container\"><div aria-expanded=\"false\" class=\"a-expander-content a-expander-partial-collapse-content\"><p>\u201c...to give a bullet-point summation of takeaways is to deny the real value of this lovely book.\u201d--<b>Harvard Business Review<\/b><\/p><p><i>From the Hardcover edition.<\/i>\n\t\t\t\t\t\t    \t\n\t\t\t\t\t    \t<\/p><\/div><p><a href=\"denied:javascript:void(0)\" data-action=\"a-expander-toggle\" class=\"a-declarative\" data-a-expander-toggle=\"{&quot;allowLinkDefault&quot;:true, &quot;expand_prompt&quot;:&quot;Read more&quot;, &quot;collapse_prompt&quot;:&quot;Read less&quot;}\"><i class=\"a-icon a-icon-extender-expand\">Read more<\/i><\/a><\/p><\/div><h3>About the Author<\/h3><div aria-live=\"polite\" data-a-expander-collapsed-height=\"300\" class=\"a-expander-collapsed-height a-row a-expander-container a-expander-partial-collapse-container\"><div aria-expanded=\"false\" class=\"a-expander-content a-expander-partial-collapse-content\"><p><b>Youngme Moon <\/b>is the Donald K. David Professor at Harvard Business School. One of HBS\u2019s most popular teachers, Dr. Moon has received the Student Association Faculty Award for teaching excellence on multiple occasions. Dr. Moon\u2019s research focuses on innovative consumer-marketing strategies and her work has been published widely, including in <i>Harvard Business Review<\/i>.\n\t\t\t\t\t\t    \t\n\t\t\t\t\t    \t<\/p><\/div><p><a href=\"denied:javascript:void(0)\" data-action=\"a-expander-toggle\" class=\"a-declarative\" data-a-expander-toggle=\"{&quot;allowLinkDefault&quot;:true, &quot;expand_prompt&quot;:&quot;Read more&quot;, &quot;collapse_prompt&quot;:&quot;Read less&quot;}\"><i class=\"a-icon a-icon-extender-expand\">Read more<\/i><\/a><\/p><\/div><\/div><\/div><\/div><\/div><div id=\"rightRail\" class=\"a-box btfRightRail\"><div class=\"a-box-inner\"><p>NO_CONTENT_IN_FEATURE<\/p><hr class=\"a-spacing-base a-divider-normal\" \/><\/div><\/div><hr class=\"a-spacing-large a-divider-normal\" \/><p>Pages with related products.\n        See and discover other items: <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/slp\/building-a-brand\/xxevttgm6t8q8w9\">building a brand<\/a>, <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/slp\/harvard-business\/9v9ud6pnk52c6ud\">harvard business<\/a>, <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/slp\/startup-company\/vdfa7jrxhwxc977\">startup company<\/a>, <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/slp\/startup-idea\/6zckxhgu2f4z9ps\">startup idea<\/a>, <a class=\"a-link-normal\" href=\"https:\/\/www.amazon.com\/slp\/startup-business-ideas\/fp8xjfzy2gah9j2\">startup business ideas<\/a><\/p>","created_at":"2018-11-07T16:23:38+0000","updated_at":"2018-11-07T16:23:38+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"www.amazon.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":158,"title":"The Power of Prolog","url":"https:\/\/www.metalevel.at\/prolog","content":"The Power of Prolog\n    \n    <div style=\"text-align: center;\">\u00a9 2005-2018 <a href=\"https:\/\/www.metalevel.at\/\">Markus Triska<\/a><\/div>\n    <div><ol start=\"0\"><li><a href=\"https:\/\/www.metalevel.at\/prolog\/introduction\"><b>Introduction<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/facets\"><b>Facets of Prolog<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/concepts\"><b>Basic Concepts<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/data\"><b>Data Structures<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/reading\"><b>Reading Prolog Programs<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/writing\"><b>Writing Prolog Programs<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/termination\"><b>Termination<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/nontermination\"><b>Nontermination<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/clpfd\"><b>Integer Arithmetic<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/metapredicates\"><b>Higher-order Predicates<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/purity\"><b>Logical Purity<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/testing\"><b>Declarative Testing<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/debugging\"><b>Declarative Debugging<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/dcg\"><b>Definite Clause Grammars<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/sorting\"><b>Sorting and Searching<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/global\"><b>Global Variables<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/tist\/\"><b>Thinking in States<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/acomip\/\"><b>Meta-interpreters<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/macros\"><b>Macros<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/optimization\"><b>Combinatorial Optimization<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/expertsystems\"><b>Expert Systems<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/web\"><b>Web Applications<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/cryptography\"><b>Cryptography<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/business\"><b>Business Cases<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/theoremproving\"><b>Theorem Proving<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/puzzles\"><b>Logic Puzzles<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/efficiency\"><b>Efficiency<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/memoization\"><b>Memoization<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/ai\"><b>Artificial Intelligence<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/horror\"><b>Horror Stories<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/fun\"><b>Fun Facts<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/engineering\"><b>Engineering Aspects<\/b><\/a><\/li>\n      <li><a href=\"https:\/\/www.metalevel.at\/prolog\/future\"><b>The Future<\/b><\/a><\/li>\n    <\/ol><\/div>\n    <p>Selected example programs are\n    available as <a href=\"https:\/\/www.metalevel.at\/prolog\/showcases\/\"><b>showcases<\/b><\/a>.\n    <\/p><p>For offline access, you can download this book from a public git\n    repository:\n    <\/p><p><b><a href=\"https:\/\/github.com\/triska\/the-power-of-prolog\">https:\/\/github.com\/triska\/the-power-of-prolog<\/a><\/b>\n    <\/p><p>If you have any comments, suggestions or questions, you can also file\n    them as issues in that repository.\n    <\/p><p><a href=\"https:\/\/www.metalevel.at\/prolog\"><b>Recursion<\/b><\/a>\n    <\/p><p><a href=\"https:\/\/www.metalevel.at\/\"><b>Main page<\/b><\/a>\n  <\/p>","created_at":"2018-11-07T16:23:38+0000","updated_at":"2018-11-07T16:23:38+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.metalevel.at","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":155,"title":"Y Not? - Adventures in Functional Programming","url":"https:\/\/www.infoq.com\/presentations\/Y-Combinator","content":"<h3 id=\"biolabel\">Bio<\/h3><p id=\"biotext\">\n\t\t\t\t\t\t\t\t\t\tJim Weirich has been developing now for over 25 years, working with everything from crunching rocket launch data on supercomputers to wiring up servos and LEDs on micro-controllers. Currently he loves working in Ruby and Rails as the Chief Scientist at EdgeCase, but you can also find him strumming on his ukulele as time permits.\n\t\t\t\t\t\t\t\t\t<\/p><p id=\"conference\">\n\t\t\t\t\t\t\t\t\t\t<a title=\"\" href=\"https:\/\/thestrangeloop.com\/\">\n\t\t\t\t\t                        <img border=\"0\" alt=\"\" src=\"https:\/\/res.infoq.com\/presentations\/Y-Combinator\/en\/promoimage\/Strangeloop+2012.jpg\" \/><\/a>\t\t\t                            \n\t\t\t                            Strange Loop is a multi-disciplinary conference that aims to bring together the developers and thinkers building tomorrow's technology in fields such as emerging languages, alternative databases, concurrency, distributed systems, mobile development, and the web.  &#13;\nStrange Loop was created in 2009 by software developer Alex Miller and is now run by a team of St. Louis-based friends and developers under Strange Loop LLC, a for-profit but not particularly profitable venture.\n\t\t\t\t\t\t\t\t\t<\/p>","created_at":"2018-11-07T16:23:35+0000","updated_at":"2018-11-07T16:23:35+0000","published_at":"2012-11-01T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.infoq.com","preview_picture":"https:\/\/res.infoq.com\/presentations\/Y-Combinator\/en\/smallimage\/Jim.JPG","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":156,"title":"Leather Binding a Paperback: a New and Improved Guide","url":"http:\/\/www.instructables.com\/id\/Leather-binding-a-Paperback-A-New-and-Improved-Gui\/","content":"<section class=\"step\" data-stepid=\"SU5Y977IEX5464N\"><h2 class=\"step-title\">Step 1: Tear Off the Old Cover<\/h2><noscript><img alt=\"Picture of Tear Off the Old Cover\" src=\"https:\/\/cdn.instructables.com\/F1V\/JZ1D\/IF9T9D6Q\/F1VJZ1DIF9T9D6Q.LARGE.jpg\" width=\"1327\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist: <\/strong>Carefully peel the existing paper cover off of your book.<\/p><p><strong>Detail: <\/strong><\/p><ul><li>Put the book down on the table  <\/li><li>Open it  <\/li><li>Put one palm on the first page to hold the book down  <\/li><li>With the other hand, peel the cover back towards the spine like you're trying to shuck a giant ear of corn  <\/li><li>Turn the book over and repeat if necessary<\/li><\/ul><\/div><\/section><section class=\"step\" data-stepid=\"SA2ASSOIF3SSF63\"><h2 class=\"step-title\">Step 2: Clean Up the Spine<\/h2><noscript><img alt=\"Picture of Clean Up the Spine\" src=\"https:\/\/cdn.instructables.com\/F9Z\/2UDP\/IF9T9DAE\/F9Z2UDPIF9T9DAE.LARGE.jpg\" width=\"1631\" height=\"1080\" \/><img alt=\"Picture of Clean Up the Spine\" src=\"https:\/\/cdn.instructables.com\/FVO\/G2UU\/IF9T9DAF\/FVOG2UUIF9T9DAF.LARGE.jpg\" width=\"1101\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Rub off the fuzzies.<\/p><p><strong>Detail<\/strong>: Rub them off. Use your thumb, or an eraser, or maybe a <em>little<\/em> bit of sponged-on water to soften them. Don't get too crazy, just clean it up a little. <\/p><\/div><\/section><section class=\"step\" data-stepid=\"S7LOFYIIF9T9DDX\"><h2 class=\"step-title\">Step 3: Pick Out Some Cool Papers<\/h2><noscript><img alt=\"Picture of Pick Out Some Cool Papers\" src=\"https:\/\/cdn.instructables.com\/F3V\/642C\/IF9T9EGS\/F3V642CIF9T9EGS.LARGE.jpg\" width=\"1631\" height=\"1080\" \/><img alt=\"Picture of Pick Out Some Cool Papers\" src=\"https:\/\/cdn.instructables.com\/FB0\/JUZG\/IF9T9EHY\/FB0JUZGIF9T9EHY.LARGE.jpg\" width=\"1273\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Get two pieces of paper you'd like to see inside the front and back cover of your book.<\/p><p><strong>Detail<\/strong>: <\/p><ul><li><strong>Size:<\/strong> Paper height = book height. Paper width = 2 x book width. <ul><li>Example: My book is 9 inches tall and 6 inches wide, so my paper has to be at least 9 inches tall and <em>12<\/em> inches wide (see how I did that complex math in my head?!) <\/li><\/ul><\/li><li><strong>Type<\/strong>: Doesn't really matter. Normal-ish. Not tissue or cardboard. <\/li><li><strong>Where: <\/strong>Craft stores, art stores, Amazon. I used <a href=\"http:\/\/www.amazon.com\/gp\/product\/B00JON7F5C?psc=1&amp;redirect=true&amp;ref_=oh_aui_search_detailpage\">these<\/a>.<\/li><\/ul><p><strong>Unnecessary Trivia:<\/strong> If you want to sound smart, you can call these pieces of paper \"endpapers\", \"end-papers\", or \"end-sheets\". The side glued to the cover is (creatively) called the \"paste-down\". The other part is called the \"flyleaf\". <\/p><\/div><\/section><section class=\"step\" data-stepid=\"SPJSG8AIF9T9EQE\"><h2 class=\"step-title\">Step 4: Fold Papers in Half<\/h2><noscript><img alt=\"Picture of Fold Papers in Half\" src=\"https:\/\/cdn.instructables.com\/F6M\/MHG7\/IFDSTQS1\/F6MMHG7IFDSTQS1.LARGE.jpg\" width=\"767\" height=\"733\" \/><\/noscript><div class=\"step-body\"><p><strong>Gist<\/strong>: Fold each piece of paper in half. Crease goes <em>with the grain<\/em>. <\/p><p><strong>Detail<\/strong>:<\/p><ul><li>Paper has \"grain\" (kind of like wood). Find the grain by gently bending the paper. It's very slightly easier to bend the paper \"with the grain\". See <a href=\"https:\/\/www.youtube.com\/watch?v=HlA8icCHl_w\">this video<\/a> (the useful part is 0:50 to 1:30). <\/li><\/ul><p><strong>Why \"with the grain\"?<\/strong><\/p><ul><li>Paper grows and shrinks when humidity changes. It mostly grows and shrinks <em>across the grain<\/em>.  <\/li><li>We're going to glue down the folded part of the paper. That means the fold won't be able to grow or shrink.  <\/li><li>If the fold is <em>across the grain<\/em> rather than <em>with the grain<\/em>, the growing and shrinking paper will be stressing the glue and crinkling the paper. <\/li><\/ul><img alt=\"\" class=\"lazy-img-onload\" src=\"http:\/\/i2.photobucket.com\/albums\/y10\/solobo\/paper%20grain%20animation.gif\" \/><\/div><\/section><section class=\"step\" data-stepid=\"SBT7NOXIF9T9G39\"><h2 class=\"step-title\">Step 5: Trim Papers to Book Size<\/h2><noscript><img alt=\"Picture of Trim Papers to Book Size\" src=\"https:\/\/cdn.instructables.com\/F7I\/61DD\/IF9T9GL4\/F7I61DDIF9T9GL4.LARGE.jpg\" width=\"1526\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Trim to match the size of your book.<\/p><p><strong>Detail<\/strong>: If your book is 6 inches wide and 9 inches tall, trim the folded paper to 6 inches wide and 9 inches tall. <\/p><\/div><\/section><section class=\"step\" data-stepid=\"SVZ4C8WIF9T9GN0\"><h2 class=\"step-title\">Step 6: Glue on the Papers<\/h2><noscript><img alt=\"Picture of Glue on the Papers\" src=\"https:\/\/cdn.instructables.com\/FX3\/ON6R\/IF9T9HUO\/FX3ON6RIF9T9HUO.LARGE.jpg\" width=\"1440\" height=\"1080\" \/><img alt=\"Picture of Glue on the Papers\" src=\"https:\/\/cdn.instructables.com\/FGZ\/EMDQ\/IF9T9HUP\/FGZEMDQIF9T9HUP.LARGE.jpg\" width=\"1605\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Glue the folded paper onto your book. Glue should only extend about 1\/4-inch from the spine of the book.<\/p><p><strong>Detail<\/strong>:<\/p><ul><li>Lay your folded paper on some scrap paper.  <\/li><li>Spread some glue on a small section near the fold (use more scrap paper or masking tape to mask it off if you want).   <\/li><li>Stick the folded paper to your book. Crease goes right next to the spine.<\/li><\/ul><p><em><strong>Glue: <\/strong>Elmer's glue (<a href=\"http:\/\/www.amazon.com\/Elmers-Washable-No-Run-School-E304\/dp\/B000Q3KHCM\">$3<\/a>) is probably fine. There's also special bookbinding PVA (<a href=\"http:\/\/www.amazon.com\/Books-Hand-Neutral-PVA-Adhesive\/dp\/B0025TZ26Q\/ref=sr_1_cc_1?s=aps&amp;ie=UTF8&amp;qid=1443906556&amp;sr=1-1-catcorr&amp;keywords=bookbinding+pva\">$8<\/a>) that's essentially the same thing, but marketed for book binding. Weigh your risk-averseness \/ perfectionism against your cheapness \/ laziness. <\/em><\/p><\/div><\/section><section id=\"step7\" class=\"step\" data-stepid=\"SPVHFH8IF3SSFC3\"><h2 class=\"step-title\">Step 7: Beef Up the Spine<\/h2><noscript><img alt=\"Picture of Beef Up the Spine\" src=\"https:\/\/cdn.instructables.com\/FLP\/GHY4\/IF9T9HW4\/FLPGHY4IF9T9HW4.LARGE.jpg\" width=\"1172\" height=\"1080\" \/><img alt=\"Picture of Beef Up the Spine\" src=\"https:\/\/cdn.instructables.com\/FRN\/4JDH\/IF9T9HWC\/FRN4JDHIF9T9HWC.LARGE.jpg\" width=\"1445\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Glue a piece of fabric onto the spine. <\/p><p><strong>Detail<\/strong>: <\/p><ul><li>Get some thin, stiff fabric (color doesn't matter).     <\/li><li>Cut it so that it is about 1 inch shorter than the spine and 2 or 3 inches wider.    <\/li><li>Center the fabric and glue it to the spine of the book (with Elmer's or PVA).    <ul><li>Note: In this picture, the ends of the fabric are still dangling. You can either leave them like this and glue them to the paper later, or you can glue them down now. <\/li><\/ul><\/li><\/ul><p><strong>Unnecessary Details: <\/strong>This piece of cloth is called a \"super\". It adds a little bit of strength to the spine and helps keep the papers from peeling away from the book when you put your covers on. The old-school way was to make it out of cheese cloth, but some people say cheese cloth is too weak. Making it out of muslin works well, or you can even buy yourself some <a href=\"http:\/\/www.amazon.com\/Binding-Super-Cloth-18x30-Sheet\/dp\/B0007LS8G0\">special cloth<\/a> marketed specifically for making supers out of. The important thing is for the cloth not to be too thick (like not felt or velvet) and to be fairly stiff (not too stretchy; don't use your mom's old nylons). I use old pillowcases\/sheets. <\/p><\/div><\/section><section id=\"step8\" class=\"step\" data-stepid=\"SKHDP5HIF3SSHWK\"><h2 class=\"step-title\">Step 8: Add \"Headbands\"<\/h2><noscript><img alt=\"Picture of Add &quot;Headbands&quot;\" src=\"https:\/\/cdn.instructables.com\/FLF\/4PL9\/IF9TC5G5\/FLF4PL9IF9TC5G5.LARGE.jpg\" width=\"1830\" height=\"1080\" \/><img alt=\"Picture of Add &quot;Headbands&quot;\" src=\"https:\/\/cdn.instructables.com\/FSW\/4T2O\/IF9TC5G7\/FSW4T2OIF9TC5G7.LARGE.jpg\" width=\"1389\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Glue some cool-looking folded cloth at the ends of the spine.<\/p><p><strong>Detail<\/strong>: <\/p><ul><li>Get some \"headband\" material <br \/><ul><li>OPTION 1: Buy it on <a href=\"http:\/\/www.amazon.com\/Bookbinding-Headband--3-Assorted-Designs\/dp\/B00D0CPBOU\/ref=sr_1_4?ie=UTF8&amp;qid=1443471890&amp;sr=8-4&amp;keywords=headbands+book+binding\">amazon<\/a> ($5 for over 100 books worth)     <\/li><li>OPTION 2: Make your own (fold a piece of fabric over a piece of string and glue it)<br \/><ul><li><img alt=\"\" class=\"lazy-img-onload\" src=\"https:\/\/www.instructables.com\/files\/deriv\/FAW\/NZ2Q\/IF9TC4XN\/FAWNZ2QIF9TC4XN.SQUARE2.jpg\" \/><\/li><\/ul><\/li><\/ul><\/li><\/ul>Cut two pieces of headband as wide as the spine of your book     Glue the two headband pieces onto the spine of your book     Make sure the bump side of the headband just barely hangs over the top\/bottom of the book<p><strong>Unnecessary Trivia: <\/strong>Headbands used to have a structural purpose. In most modern books, they don't. But they sure look cool. <\/p><\/div><\/section><section id=\"step9\" class=\"step\" data-stepid=\"SNCH724IF9TC4UJ\"><h2 class=\"step-title\">Step 9: Get Some Cardboard<\/h2><noscript><img alt=\"Picture of Get Some Cardboard\" src=\"https:\/\/cdn.instructables.com\/FT6\/O7SV\/IF9TC5XZ\/FT6O7SVIF9TC5XZ.LARGE.jpg\" width=\"1273\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Get some cardboard to make your covers out of. <\/p><p><strong>Detail<\/strong>: You want something sturdy, about 1\/16 inch thick. I used <a href=\"http:\/\/www.amazon.com\/gp\/product\/B0013JRFUA?psc=1&amp;redirect=true&amp;ref_=oh_aui_search_detailpage\">chipboard from Amazon<\/a> for this project, but I've used the back of sketch pads before too. I wouldn't recommend corrugated cardboard (like you'd get from a box), because it seems like it would crease and dent too easily, but I've never actually tried it... maybe it's awesome!<\/p><p>You'll need enough for 2-4 pieces that are just a hair bigger than your book.<\/p><\/div><\/section><section id=\"step10\" class=\"step\" data-stepid=\"S6IZ1JTIF9TC646\"><h2 class=\"step-title\">Step 10: Cut Two Pieces of Cardboard to Size<\/h2><noscript><img alt=\"Picture of Cut Two Pieces of Cardboard to Size\" src=\"https:\/\/cdn.instructables.com\/FMI\/MSPI\/IF9TC7P4\/FMIMSPIIF9TC7P4.LARGE.jpg\" width=\"1362\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Cut the cardboard about 1\/4-inch <em>taller<\/em> and 1\/4-inch <em>narrower<\/em> than your book. <em>Grain should run from top to bottom<\/em> (just like the papers). <\/p><ul><li>Example: 6 x 9-inch book =&gt; 5-3\/4 x 9-1\/4 -inch cardboard pieces<\/li><\/ul><p><strong>Detail<\/strong>: <\/p><ul><li>The cardboard needs to overhang the book by about 1\/8-inch on the top, bottom, and one side.    <\/li><li>It also needs to have a gap (say 3\/8-inch) near the spine so that the leather can fold.    <\/li><li>This means:  <ul><li>It has to be about 1\/4-inch <strong>taller<\/strong> than your book (1\/8 + 1\/8 = 1\/4)   <\/li><li>It has to be about 1\/4-inch <strong>narrower<\/strong> than your book (1\/8-inch bigger for the overhang, 3\/8-inch smaller for the gap)<\/li><\/ul><\/li><\/ul>Check for the grain direction of the cardboard the same way you did for the paper. Make sure the grain (the easy-to-bend direction) runs from top to bottom (just like it did with the papers). <\/div><\/section><section id=\"step11\" class=\"step\" data-stepid=\"SVAUN81IF9TC8DS\"><h2 class=\"step-title\">Step 11: [Optional]: Glue on Some Cool Designs<\/h2><noscript><img alt=\"Picture of [Optional]: Glue on Some Cool Designs\" src=\"https:\/\/cdn.instructables.com\/FAQ\/PAME\/IF9TCAHR\/FAQPAMEIF9TCAHR.LARGE.jpg\" width=\"1010\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Use extra cardboard to glue a second partial-layer on each cover piece.<\/p><p><strong>Details<\/strong>: The sky's the limit. You could even X-ACTO the edges of the second layer to get rounded curves and stuff. Just realize that any gaps that are really small (like less than 1\/8-inch) will be hard to get the leather to conform to.<\/p><\/div><\/section><section id=\"step12\" class=\"step\" data-stepid=\"S9D07MRIF9TCAKL\"><h2 class=\"step-title\">Step 12: Make a Spine<\/h2><noscript><img alt=\"Picture of Make a Spine\" src=\"https:\/\/cdn.instructables.com\/FED\/1OU9\/IJSUXCPB\/FED1OU9IJSUXCPB.LARGE.jpg\" width=\"960\" height=\"720\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Cut a piece of paper or card stock to give a little shape to the spine of your new leather cover. <\/p><p><strong>Detail<\/strong>:(see the red piece in the diagram)<\/p><ul><li>Measure the thickness of your book. Add the thickness of your cardboard covers. That's how wide your spine piece should be.  <ul><li>Example: 3\/4-inch thick book + two 1\/16-inch thick cardboard covers =&gt; a 7\/8-inch wide spine.<\/li><\/ul><\/li><li>The spine should have the same height as your cardboard covers (which are 1\/4-inch taller than the book itself).  <ul><li>Example: 9-inch tall book has 9-1\/4-inch tall covers and a 9-1\/4-inch tall spine.<\/li><\/ul><\/li><li><strong>Final spine dimensions (example): <\/strong>7\/8 inch x 9-1\/4 inches. <\/li><\/ul><\/div><\/section><section id=\"step13\" class=\"step\" data-stepid=\"SIVREX2IFCDF4NZ\"><h2 class=\"step-title\">Step 13: Use Masking Tape to Line Things Up<\/h2><noscript><img alt=\"Picture of Use Masking Tape to Line Things Up\" src=\"https:\/\/cdn.instructables.com\/FHP\/FL7J\/IFCDF6WH\/FHPFL7JIFCDF6WH.LARGE.jpg\" width=\"2078\" height=\"1080\" \/><img alt=\"Picture of Use Masking Tape to Line Things Up\" src=\"https:\/\/cdn.instructables.com\/FTI\/9CFG\/IFCDF4VV\/FTI9CFGIFCDF4VV.LARGE.jpg\" width=\"1887\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Lay out your covers and spine and tape them to keep them that way.<\/p><p><strong>Detail<\/strong>: <\/p><ul><li>Lay the front cover face down on the table.  <\/li><li>Lay the spine 3\/8-inch to the right of the front cover.  <\/li><li>Lay the back cover face down 3\/8-inch to the right of the spine.  <\/li><li>Tape.<\/li><\/ul><\/div><\/section><section id=\"step14\" class=\"step\" data-stepid=\"SDLV1K2IFCDF6YA\"><h2 class=\"step-title\">Step 14: Pick Out Some Leather<\/h2><noscript><img alt=\"Picture of Pick Out Some Leather\" src=\"https:\/\/cdn.instructables.com\/FMX\/NC1Y\/IFCDF98X\/FMXNC1YIFCDF98X.LARGE.jpg\" width=\"1631\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist: <\/strong>Pick out some leather (or other cloth).<\/p><p><strong>Detail:<\/strong><\/p><ul><li>Get a piece that's a little bigger than the total size of your laid-out covers and spine.   <\/li><li>Thin, flexible, chromium-tanned leather is best for this type of binding.   <\/li><li>If you don't want leather (or can't find some), just use some cool fabric! I've done this before, and it looks pretty nice. <\/li><\/ul><p><strong>Where to get leather:  <\/strong>If you live close to a leather store like Tandy Leather, go in and see if they have anything for sale (I found a nice big sheet of this black-dyed  pigskin for like $20). Otherwise, check Ebay or Tandy Leather's online  catalog, or get a free leather couch on craigslist and cannibalize it. I would stay away from vegetable-tanned leather (unless you're  planning to finish it yourself). Anything you find on furniture or upholstery is probably fine, as long as it's flexible and you like it. <\/p><\/div><\/section><section id=\"step15\" class=\"step\" data-stepid=\"S2P6N5YIFCDF9RY\"><h2 class=\"step-title\">Step 15: Glue the Leather to the Covers<\/h2><noscript><img alt=\"Picture of Glue the Leather to the Covers\" src=\"https:\/\/cdn.instructables.com\/FS3\/DE40\/IFCDFAYN\/FS3DE40IFCDFAYN.LARGE.jpg\" width=\"1695\" height=\"1080\" \/><img alt=\"Picture of Glue the Leather to the Covers\" src=\"https:\/\/cdn.instructables.com\/F0R\/FS7X\/IFCDFAWO\/F0RFS7XIFCDFAWO.LARGE.jpg\" width=\"1699\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Glue the leather \/ fabric onto the covers. Press it into all the nooks and crannies of your 3D cardboard design.<\/p><p><strong>Detail<\/strong>: <\/p><ul><li>Coat part of your cover (say the front cover) with glue. (if you're Lightning McQueen, you could do the whole thing at once).    <\/li><li>Lay a big piece of leather down on top with enough to overhang your cover by about an inch on all sides.    <\/li><li>Use something pointy-ish but smooth-ish to push the leather down into all the little 3D parts of your design (if you have them). I've used the butt-end of a toothbrush, a capped Sharpie, the blunt end of a ceramic chopstick, the handle of a butterknife... whatever's handy and not sharp enough to mark the leather.  <\/li><li>Repeat with the other part of your cover and the spine. <\/li><\/ul><\/div><\/section><section id=\"step16\" class=\"step\" data-stepid=\"S5QYYVUIFCDFB08\"><h2 class=\"step-title\">Step 16: Trim the Leather<\/h2><noscript><img alt=\"Picture of Trim the Leather\" src=\"https:\/\/cdn.instructables.com\/FC4\/0FN8\/IFCDFBS1\/FC40FN8IFCDFBS1.LARGE.jpg\" width=\"1934\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Leave a margin of about 1-inch on all sides. Trim the corners at a 45\u00b0 angle, leaving an overhang about as big as the cover is thick. <\/p><p><strong>Detail<\/strong>: Use scissors, or an X-ACTO knife, or an axe (if you're Paul Bunyan).<\/p><\/div><\/section><section id=\"step17\" class=\"step\" data-stepid=\"SFU247HIFCDFC6G\"><h2 class=\"step-title\">Step 17: Fold and Glue Two Sides of the Leather<\/h2><noscript><img alt=\"Picture of Fold and Glue Two Sides of the Leather\" src=\"https:\/\/cdn.instructables.com\/FHO\/P261\/IFCDFCLY\/FHOP261IFCDFCLY.LARGE.jpg\" width=\"1992\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p>Fold and glue the top and bottom.Keep things nice and tight. Set a weight on top for a few minutes to make sure the glue has set.<\/p><\/div><\/section><section id=\"step18\" class=\"step\" data-stepid=\"S8NTSW8IFCDFCSE\"><h2 class=\"step-title\">Step 18: Make Sure Corners Are Creased Well<\/h2><noscript><img alt=\"Picture of Make Sure Corners Are Creased Well\" src=\"https:\/\/cdn.instructables.com\/FD0\/6LBY\/IFCDFDBE\/FD06LBYIFCDFDBE.LARGE.jpg\" width=\"1289\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p>Pay special attention to the corners of the two folded slides. Add some extra glue and crease with a table knife or something if you have to.<\/p><\/div><\/section><section id=\"step19\" class=\"step\" data-stepid=\"S47MVX4IFCDFDED\"><h2 class=\"step-title\">Step 19: Fold and Glue the Remaining Two Sides<\/h2><noscript><img alt=\"Picture of Fold and Glue the Remaining Two Sides\" src=\"https:\/\/cdn.instructables.com\/FBE\/RUOZ\/IFCDFW88\/FBERUOZIFCDFW88.LARGE.jpg\" width=\"1706\" height=\"1080\" \/><img alt=\"Picture of Fold and Glue the Remaining Two Sides\" src=\"https:\/\/cdn.instructables.com\/F91\/HS6K\/IFCDFW89\/F91HS6KIFCDFW89.LARGE.jpg\" width=\"1142\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p>Keep things tight. Pay close attention to the corners and make sure they're pressed down well.<\/p><\/div><\/section><section id=\"step20\" class=\"step\" data-stepid=\"SOXJ9WIIFCDFVUF\"><h2 class=\"step-title\">Step 20: Check the Fit<\/h2><noscript><img alt=\"Picture of Check the Fit\" src=\"https:\/\/cdn.instructables.com\/F20\/LPDJ\/IFDSTJNJ\/F20LPDJIFDSTJNJ.LARGE.jpg\" width=\"1631\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p>Lay your book on one half of the cover. Make sure the cover overhangs by about 1\/8-inch. Fold the cover over the book and make sure everything fits nicely.<\/p><p>If you left the ends of the fabric on the spine dangling in step 7, glue them down to the papers now. Pull them nice and tight. <\/p><\/div><\/section><section id=\"step21\" class=\"step\" data-stepid=\"S6DAKAFIFDSTJP5\"><h2 class=\"step-title\">Step 21: Glue on the Covers<\/h2><noscript><img alt=\"Picture of Glue on the Covers\" src=\"https:\/\/cdn.instructables.com\/FT4\/LCCD\/IFDSTMHP\/FT4LCCDIFDSTMHP.LARGE.jpg\" width=\"1879\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Glue on the front and back covers, but<em> not<\/em> the spine or the gap next to the spine.<\/p><p><strong>Detail<\/strong>:<\/p><ul><li>Put a piece of scrap paper just inside the outer paper to protect the pages from glue.<\/li><li>Get the book all lined up.<\/li><li>Put glue on the paper, but leave a 3\/8-inch gap near the spine un-glued.<\/li><li>Close the cover on the glued paper and press firmly for a minute or so.<\/li><li>Repeat on the other side of the book.<\/li><\/ul><\/div><\/section><section id=\"step22\" class=\"step\" data-stepid=\"S2IKRT8IFDSTMJF\"><h2 class=\"step-title\">Step 22: Put Everything Under Weights to Dry<\/h2><noscript><img alt=\"Picture of Put Everything Under Weights to Dry\" src=\"https:\/\/cdn.instructables.com\/F86\/8QOT\/IFDSTND6\/F868QOTIFDSTND6.LARGE.jpg\" width=\"1103\" height=\"1080\" \/><img alt=\"Picture of Put Everything Under Weights to Dry\" src=\"https:\/\/cdn.instructables.com\/FMU\/CQK9\/IFDSTND9\/FMUCQK9IFDSTND9.LARGE.jpg\" width=\"1680\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p><strong>The Gist<\/strong>: Stack some books or weights on top of your book for a good day or two while everything dries.<\/p><p><strong>Detail<\/strong>: <\/p><ul><li>Make sure your book is sitting on something smooth and clean... little bumps or specks of dirt will get imprinted into the leather.<\/li><li>[Optional]: Use knitting needles, two pencils end-to-end, or something else to press the leather into the gap next to the spine while everything dries.<\/li><li>Put a smooth board on top of your book.<\/li><li>Stack some books \/ weights \/ big heavy rocks on top.<\/li><li>Wait for a day or two so that everything can dry without warping.<\/li><\/ul><\/div><\/section><section id=\"step23\" class=\"step\" data-stepid=\"SPVO7XCIFDSTNF1\"><h2 class=\"step-title\">Step 23: Enjoy<\/h2><noscript><img alt=\"Picture of Enjoy\" src=\"https:\/\/cdn.instructables.com\/FNA\/S9EX\/IFDSTNP3\/FNAS9EXIFDSTNP3.LARGE.jpg\" width=\"1631\" height=\"1080\" \/><img alt=\"Picture of Enjoy\" src=\"https:\/\/cdn.instructables.com\/FJB\/0K42\/IFDSTNP4\/FJB0K42IFDSTNP4.LARGE.jpg\" width=\"1125\" height=\"1080\" \/><img alt=\"Picture of Enjoy\" src=\"https:\/\/cdn.instructables.com\/FAV\/GO3S\/IFDSTNP5\/FAVGO3SIFDSTNP5.LARGE.jpg\" width=\"1440\" height=\"1080\" \/><img alt=\"Picture of Enjoy\" src=\"https:\/\/cdn.instructables.com\/F0Z\/9GPG\/IFDSTNP6\/F0Z9GPGIFDSTNP6.LARGE.jpg\" width=\"1747\" height=\"1080\" \/><img alt=\"Picture of Enjoy\" src=\"https:\/\/cdn.instructables.com\/FUC\/HPBP\/IFDSTNP7\/FUCHPBPIFDSTNP7.LARGE.jpg\" width=\"1440\" height=\"1080\" \/><\/noscript><div class=\"step-body\"><p>You now have a leather-bound version of your favorite paperback book! <\/p><p>Want to check out my novels (about teenage medieval secret agents), short stories (about everything from dystopian futures to sword-and-sorcery fantasy) food experiments (about biscuits, hash browns, and pancakes), and book reviews? Take a look at my <a href=\"http:\/\/travisdanielbow.weebly.com\/\">website<\/a> and like my <a href=\"https:\/\/www.facebook.com\/thanenovel\">facebook page<\/a>, where you can get updates about giveaways and other news.<\/p><p>Thanks for reading.<\/p><\/div><\/section><div class=\"contest-entries finalized-contests\"><div class=\"contest-entry\"><img alt=\"Leather Contest\" src=\"https:\/\/cdn.instructables.com\/F0A\/G9K7\/IESTSW3S\/F0AG9K7IESTSW3S.SMALL.jpg\" \/><p>Second Prize in the <br \/><a href=\"http:\/\/www.instructables.com\/contest\/leather2015\/\">Leather Contest<\/a><\/p><\/div><div class=\"contest-entry\"><img alt=\"Papercraft Contest 2015\" src=\"https:\/\/cdn.instructables.com\/FGI\/9C07\/IFI37YIB\/FGI9C07IFI37YIB.SMALL.jpg\" \/><p>Runner Up in the <br \/><a href=\"http:\/\/www.instructables.com\/contest\/papercraft2015\/\">Papercraft Contest 2015<\/a><\/p><\/div><\/div><section class=\"imadeits\"><h2>6 People Made This Project!<\/h2><ul class=\"list-unstyled unstyled imadeit-list js-posts\"><li class=\"imadeit js-post\" data-id=\"C0LNQYEJMGKY4DS\"><img alt=\"I did this on a one-...-CaseyL52\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F75\/YA3J\/JMGKY3ZK\/F75YA3JJMGKY3ZK.SQUARE3.jpg?auto=webp\" \/><h3 class=\"made-by\"><a class=\"user\" href=\"http:\/\/www.instructables.com\/member\/CaseyL52\/\">CaseyL52<\/a> made it!<\/h3><\/li><li class=\"imadeit js-post\" data-id=\"CWMP1RTJ4SQ7NU7\"><img alt=\"Great tutorial, I pr...-PseudonymousD\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FNI\/NJ5I\/J4SQ7NKL\/FNINJ5IJ4SQ7NKL.SQUARE3.jpg?auto=webp\" \/><h3 class=\"made-by\"><a class=\"user\" href=\"http:\/\/www.instructables.com\/member\/PseudonymousD\/\">PseudonymousD<\/a> made it!<\/h3><\/li><li class=\"imadeit js-post\" data-id=\"CDRCK6MIXQFPXUX\"><img alt=\"Hi, I finally got ro...-EvanA25\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FOD\/ZXWC\/IXQFPXOZ\/FODZXWCIXQFPXOZ.SQUARE3.jpg?auto=webp\" \/><h3 class=\"made-by\"><a class=\"user\" href=\"http:\/\/www.instructables.com\/member\/EvanA25\/\">EvanA25<\/a> made it!<\/h3><\/li><li class=\"imadeit js-post\" data-id=\"CQBTNQUIO8U2LHR\"><img alt=\"Update: I'm finally ...-Soccer_Kates\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FOI\/ZKNS\/IO8U2LHN\/FOIZKNSIO8U2LHN.SQUARE3.jpg?auto=webp\" \/><h3 class=\"made-by\"><a class=\"user\" href=\"http:\/\/www.instructables.com\/member\/Soccer_Kates\/\">Soccer_Kates<\/a> made it!<\/h3><\/li><li class=\"see-more\"><div class=\"see-more-inner\"><p>See 2 More<\/p><\/div><\/li><\/ul><\/section><section class=\"recommendations\"><h2>Recommendations<\/h2><ul class=\"list-unstyled projects\"><li><article><a href=\"http:\/\/www.instructables.com\/id\/Wood-and-Resin-Coffee-Table\/\"><img alt=\"Wood and Resin Coffee Table\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F6W\/2KGN\/JNZ4EEI2\/F6W2KGNJNZ4EEI2.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><h3><a href=\"http:\/\/www.instructables.com\/id\/Wood-and-Resin-Coffee-Table\/\">Wood and Resin Coffee Table<\/a><\/h3><footer><p>by <a href=\"http:\/\/www.instructables.com\/member\/FreemanF1\/\">FreemanF1<\/a> in\n                        \n                            <a href=\"http:\/\/www.instructables.com\/craft\/\">Craft<\/a><\/p><\/footer><\/div><\/article><\/li><li><article><a href=\"http:\/\/www.instructables.com\/id\/Black-Magic-Leather-Grimoire-Bookbinding-Tutorial\/\"><img alt=\"Black Magic Leather Grimoire - Bookbinding Tutorial\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FYP\/974N\/JNIRDTWZ\/FYP974NJNIRDTWZ.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><h3><a href=\"http:\/\/www.instructables.com\/id\/Black-Magic-Leather-Grimoire-Bookbinding-Tutorial\/\">Black Magic Leather Grimoire - Bookbinding Tutorial<\/a><\/h3><footer><p>by <a href=\"http:\/\/www.instructables.com\/member\/Nerdforge\/\">Nerdforge<\/a> in\n                        \n                            <a href=\"http:\/\/www.instructables.com\/craft\/books-and-journals\/\">Books &amp; Journals<\/a><\/p><\/footer><\/div><\/article><\/li><li><article><a href=\"http:\/\/www.instructables.com\/id\/Coptic-Letter-Binding-Notepad-2-Sketchbook\/\"><img alt=\"Coptic Letter Binding -  Notepad 2 Sketchbook\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FTO\/OSA7\/JL6UXM77\/FTOOSA7JL6UXM77.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><h3><a href=\"http:\/\/www.instructables.com\/id\/Coptic-Letter-Binding-Notepad-2-Sketchbook\/\">Coptic Letter Binding -  Notepad 2 Sketchbook<\/a><\/h3><footer><p>by <a href=\"http:\/\/www.instructables.com\/member\/ruudcreates\/\">ruudcreates<\/a> in\n                        \n                            <a href=\"http:\/\/www.instructables.com\/craft\/books-and-journals\/\">Books &amp; Journals<\/a><\/p><\/footer><\/div><\/article><\/li><li><article><a href=\"http:\/\/www.instructables.com\/class\/Starting-a-Handmade-Business\/\"><img alt=\"Starting a Handmade Business\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F5M\/0OFD\/J9EGZ1XT\/F5M0OFDJ9EGZ1XT.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><img alt=\"Starting a Handmade Business\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FXN\/VIAT\/J98R7KK6\/FXNVIATJ98R7KK6.SQUARE2.jpg?auto=webp\" \/><h3><a href=\"http:\/\/www.instructables.com\/id\/Starting-a-Handmade-Business\/\">Starting a Handmade Business<\/a><\/h3><footer><p>5,741 Enrolled<\/p><\/footer><\/div><\/article><\/li><\/ul><ul class=\"list-unstyled contests\"><li><h3 class=\"sr-only\">Halloween Contest 2018<\/h3><a href=\"http:\/\/www.instructables.com\/contest\/halloween18\/\"><img class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FSA\/Q4G4\/JMQKX2XX\/FSAQ4G4JMQKX2XX.MEDIUM.jpg?auto=webp\" alt=\"Halloween Contest 2018\" \/><\/a><\/li><li><h3 class=\"sr-only\">Plastics Contest<\/h3><a href=\"http:\/\/www.instructables.com\/contest\/plastics2018\/\"><img class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FLR\/AT0U\/JMKVA284\/FLRAT0UJMKVA284.MEDIUM.jpg?auto=webp\" alt=\"Plastics Contest\" \/><\/a><\/li><li><h3 class=\"sr-only\">Optics Contest<\/h3><a href=\"http:\/\/www.instructables.com\/contest\/optics\/\"><img class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F4H\/U1WQ\/JMKV2FGF\/F4HU1WQJMKV2FGF.MEDIUM.jpg?auto=webp\" alt=\"Optics Contest\" \/><\/a><\/li><\/ul><\/section>","created_at":"2018-11-07T16:23:35+0000","updated_at":"2018-11-07T16:23:35+0000","published_at":"2015-10-05T00:00:00+0000","published_by":["Instructables"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"www.instructables.com","preview_picture":"https:\/\/cdn.instructables.com\/ORIG\/FQH\/UZS9\/IFDSTPQ0\/FQHUZS9IFDSTPQ0.jpg?width=1297","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":153,"title":"Lobotomizing GNOME","url":"https:\/\/eklitzke.org\/lobotomizing-gnome","content":"<p>I have pretty basic needs from my desktop. I do 99% of my work using just three\nprograms: a terminal emulator, Emacs, and Firefox. I don\u2019t want a lot of bells\nand whistles in my desktop, and I really just want it to get out of the way so I\ncan do my work.<\/p><p>But just because my needs are basic doesn\u2019t mean that I want a 1990\u2019s window\nmanager experience. I want good text rendering and windows and buttons with\nrounded corners. I want my laptop to work correctly when connecting it to\nexternal displays or projectors without a lot of futzing around. I want vsync to\nwork with my monitor out of the box, I want to be able to watch video without\ntearing, and I want a desktop that has first class support for high-DPI\ndisplays. I also want to have some basic integration with the other system\nfeatures provided by my distro, which increasingly means high-quality\nintegration with NetworkManager and different systemd components. I want to get\nintegrated notifications when a program segfaults on my computer or in case\nthere\u2019s an SELinux AVC denial.<\/p><p>In my opinion, in these areas GNOME is far ahead of everything else. The stock\nGNOME configuration is beautiful but minimal\u2014I get a discreet black bar across\nthe top of my screen that shows the time and date, and lets me control the\nfundamentals like network access. I think that GNOME Shell is the most\nattractive and useful window manager for any operating system out there. And\nGNOME has <em>really good<\/em> integration with the other parts of my system, which\nmakes sense because it\u2019s the default desktop environment on my distro (Fedora)\nand most others, including Debian and Ubuntu. GNOME is also light-years ahead of\neverything else in terms of Wayland support. Fedora has been shipping Wayland as\nthe default GNOME display backend since Fedora 25 (2016), and it works\nincredibly well. The most compelling user-visible feature that has come out of\nthis is GNOME\u2019s \u201cfractional scaling\u201d feature, which is a quantum leap in terms\nof how content is scaled on high-DPI screens.<\/p><p>But I\u2019ll be honest: GNOME is huge and kind of bloated, and it\u2019s hard to disable\nvarious unwanted components. GNOME Shell is amazing, but a lot of the other\ncomponents of GNOME are simply unwanted. This is what turns a lot of power users\naway from GNOME, which I think is a shame given all of the other amazing things\nabout GNOME. While you won\u2019t find these instructions in the GNOME manuals, if\nyou know what you\u2019re doing modern GNOME releases make it very easy to lobotomize\na lot of the unneeded and unwanted features.<\/p><h2>Configuration With <code>dconf<\/code><\/h2><p>The first step in improving the GNOME user experience is toggling basic features\non or off. You can do this either by clicking around in the GUI, or using the\n<code>dconf<\/code> command line tool. One of the interesting things about GNOME is that it\nhas a lot of hidden options that either not at all exposed in the GUI, or only\naccessible through a tool like <code>gnome-tweaks<\/code>. But if you know what you\u2019re doing\nyou can change a lot of GNOME behavior.<\/p><p>Most people are already in the habit of keeping their dot files in a git repo\nsomewhere. I recently took this up a notch, and I now maintain my whole desktop\nconfiguration (including my dot files) using\n<a href=\"https:\/\/www.ansible.com\/\">Ansible<\/a>. To illustrate my GNOME configuration I\u2019ll\nbe posting snippets from current Ansible configuration. If you\u2019re not familiar\nwith Ansible don\u2019t worry: it\u2019s just YAML, and can easily be translated into\ncommand line invocations.<\/p><p>Here\u2019s what my GNOME <code>dconf<\/code> settings look like in Ansible:<\/p><pre class=\"language-yaml\">- name: update gnome dconf settings\n  dconf:\n    key: \"{{item.key}}\"\n    value: \"{{item.value}}\"\n  loop:\n    - {key: '\/org\/gnome\/desktop\/input-sources\/xkb-options', value: \"['caps:ctrl_modifier']\"}\n    - {key: '\/org\/gnome\/desktop\/interface\/clock-show-date', value: 'true'}\n    - {key: '\/org\/gnome\/desktop\/interface\/cursor-blink', value: 'false'}\n    - {key: '\/org\/gnome\/desktop\/interface\/gtk-theme', value: \"'Adwaita-dark'\"}\n    - {key: '\/org\/gnome\/desktop\/privacy\/old-files-age', value: 'uint32 7'}\n    - {key: '\/org\/gnome\/desktop\/privacy\/remove-old-trash-files', value: 'true'}\n    - {key: '\/org\/gnome\/desktop\/privacy\/report-technical-problems', value: 'false'}\n    - {key: '\/org\/gnome\/desktop\/search-providers\/disable-external', value: 'true'}\n    - {key: '\/org\/gnome\/desktop\/wm\/preferences\/audible-bell', value: 'false'}\n    - {key: '\/org\/gnome\/desktop\/wm\/preferences\/focus-mode', value: \"'sloppy'\"}\n    - {key: '\/org\/gnome\/settings-daemon\/plugins\/color\/night-light-enabled', value: 'true'}\n    - {key: '\/org\/gnome\/settings-daemon\/plugins\/xsettings\/antialiasing', value: \"'rgba'\"}\n    - {key: '\/org\/gnome\/shell\/disable-user-extensions', value: 'true'}\n    - {key: '\/org\/gnome\/software\/allow-updates', value: 'false'}\n    - {key: '\/org\/gnome\/software\/download-updates', value: 'false'}\n    - {key: '\/org\/gnome\/terminal\/legacy\/default-show-menubar', value: 'false'}\n<\/pre><p>You can probably guess what most of these options do based on the key name. A\nlot of them turn off various options, or change basic features of the window\nmanager (such as setting \u201csloppy\u201d focus, a must-have for me).<\/p><p>There\u2019s one I want to call out in particular though: I set\n<code>\/org\/gnome\/shell\/disable-user-extensions<\/code> to <code>false<\/code>, which completely disables\nthe user extensions feature of GNOME. User extensions are a mechanism that allow\nusers to write GNOME extensions in Javascript, similar to how Chrome and Firefox\nextensions work. In my opinion this idea has dubious merit, and my personal\nfeeling is the less Javascript in my life the better. I felt somewhat vindicated\nabout this decision during recent\n<a href=\"https:\/\/phoronix.com\/scan.php?page=news_item&amp;px=GNOME-Shell-Memory-Leak-Fix\">coverage<\/a>\nof a memory leak in GNOME Shell. The underlying issue was related to the\nJavascript garbage collector in GNOME Shell not collecting object references in\na timely manner. I\u2019m not sure that disabling user extensions actually disables\nthe Javascript engine completely, but it definitely minimizes it to the least\npossible scope.<\/p><p>You might wonder how you\u2019re actually supposed to find these options in the first\nplace\u2014how are you supposed to know what keys are available, and what values\nthey take? In practice I found most of these in a kind of hacky way. The <code>dconf<\/code>\ncommand lets you dump the entire database using <code>dconf dump \/<\/code>. So what I did to\nfind most of these options was some process kind of like the following:<\/p><pre class=\"language-bash\"># Get the initial dconf state.\n$ dconf dump \/ &gt;a\n# Click around in the GUI, change things in gnome-tweaks, etc.\n# Dump the final dconf state.\n$ dconf dump \/ &gt;b\n# Manually look at the diff of the two states.\n$ diff -u a b\n<\/pre><p>This is simple and effective, and just takes a minute or so of work. The more\nrobust (but much more labor intensive) way to do this is to look at the GNOME\nsettings schemas. GNOME packages install files ending in a <code>.gschema.xml<\/code>\nextension, which are XML descriptions of all of the possible options supported\nby the application, a description of what the options do, and their default\nvalues. In principle you could discover all of these options by digging around\nin <code>\/usr\/share\/glib-2.0\/schemas<\/code> and reading a bunch of XML schema descriptions.\nIn practice I do this very rarely, really only when if I need to read the\ncomplete documentation for an option.<\/p><h2>Removing Unneeded Components<\/h2><p>GNOME bundles a lot of components that I find to be really annoying and\nunwelcome. These are the biggest offenders:<\/p><ul><li><p><a href=\"https:\/\/wiki.gnome.org\/Projects\/Tracker\/\">Tracker<\/a> is a file daemon that is\nsupposed to work like Spotlight on macOS. I don\u2019t want this. Between <code>find<\/code>\nand <code>locate<\/code> I can find all of my files quickly. I don\u2019t need yet another\nthing trying to index my files.<\/p><\/li><li><p><a href=\"https:\/\/wiki.gnome.org\/Apps\/Evolution\">Evolution<\/a> is the standard GNOME\nemail application, which I don\u2019t want to use. If Evolution was just an email\nprogram I could ignore that would be fine. Unfortunately, Evolution evolved\ninto a weird half-baked clone of Microsoft Exchange complete with contact\nintegration, calendar capabilities, and a host of other office productivity\nfeatures that are not useful to me. Evolution provides a component called\n<code>evolution-data-server<\/code> that is a daemon that ties these things together.<\/p><\/li><li><p><a href=\"https:\/\/wiki.gnome.org\/Projects\/GnomeOnlineAccounts\">GNOME Online Accounts<\/a>\nis a feature that lets you connect your GNOME account to things like your\nGoogle or Facebook login. The idea is supposed to be that you can connect\nthings like your Google calendar and email to the GNOME calendar and\nEvolution. It\u2019s a cool idea, but I\u2019m not comfortable with the idea of giving\nmy online credentials to a bundle of tens or hundreds of thousands of lines\nof hand-written C code with manual memory management.<\/p><\/li><li><p><a href=\"https:\/\/wiki.gnome.org\/Apps\/Software\">GNOME Software<\/a> is this weird app\nstore that I don\u2019t care to use. It always gives me notifications about GNOME\nupdates that don\u2019t actually correlate with updates in my distro package\nmanager. The <code>dconf<\/code> configuration I showed above will actually tell GNOME\nSoftware not to do this, but I don\u2019t really want GNOME Software running at\nall.<\/p><\/li><\/ul><p>It turns out that you can remove a <em>lot<\/em> of these components without breaking\nanything. I remove as much of these as possible, as well as a number of programs\nthat I know I don\u2019t plan on using. The following list is specific to Fedora 28,\nbut shows what I remove:<\/p><pre class=\"language-yaml\"># remove unwanted gnome packages\n- name: dnf remove gnome bloat\n  dnf:\n    name:\n      - cheese\n      - evolution\n      - evolution-ews\n      - evolution-help\n      - gfbgraph\n      - gnome-boxes\n      - gnome-calendar\n      - gnome-contacts\n      - gnome-dictionary\n      - gnome-documents\n      - gnome-getting-started-docs\n      - gnome-initial-setup\n      - gnome-maps\n      - gnome-online-miners\n      - gnome-photos\n      - gnome-software\n      - gnome-user-docs\n      - gnome-user-share\n      - gnome-video-effects\n      - gnome-weather\n      - simple-scan\n      - totem\n      - tracker-miners\n      - yelp\n    state: absent\n<\/pre><p>Some of these are harmless programs that don\u2019t actually install session daemons,\nbut are also programs I don\u2019t plan on using. In other cases (e.g.\n<code>gnome-software<\/code>) uninstalling the program actually removes daemons that are run\nby default in the session, so removing those components actually causes less\nprograms to run and waste memory.<\/p><p>They key thing here is to make sure you don\u2019t remove dependencies of any of the\ncore components, which includes <code>gnome-shell<\/code> but also some other programs like\n<code>nautilus<\/code>. To double check that I didn\u2019t actually break anything, I also have\nthe following stanza in my config. This is just a sanity check to make sure that\nnothing critical was accidentally removed.<\/p><pre class=\"language-yaml\"># double-check that we still have the basics\n- name: ensure core gnome packages are installed\n  dnf:\n    name:\n      - eog\n      - evince\n      - evolution-data-server\n      - flatpak\n      - gdm\n      - gnome-keyring\n      - gnome-menus\n      - gnome-screenshot\n      - gnome-shell\n      - gnome-terminal\n      - gnome-tweaks\n      - nautilus\n      - redhat-menus\n    state: present\n<\/pre><p>I\u2019m removing a lot of stuff here, including some things you might want to keep.\nSo look at the list for yourself and apply caution before removing anything.\nAfter removing these packages there are still a few annoying components like\n<code>evolution-data-server<\/code> and <code>tracker<\/code> that still remain because they can\u2019t be\ncompletely uninstalled without breaking critical components. But they can still\nbe disabled in user sessions, as I\u2019ll explain below.<\/p><h2>Disabling <code>evolution-data-server<\/code>, Tracker, and other bloat<\/h2><p>GNOME used to have a horribly complicated system that was used to manage start\nuser session services, and in the past it was very difficult or impossible to\ndisable certain components. As part of the Wayland effort, GNOME moved to\nmanaging user services using systemd. This is <em>really awesome<\/em> because it means\nyou can control GNOME startup services using <code>systemctl<\/code> just like regular\nsystem services.<\/p><p>On Wayland GNOME services run as part of your systemd user session. To see these\nservices, use one of the following commands:<\/p><pre class=\"language-bash[\"># Show running user services.\n$ systemctl --user status\n# Show all user units and their status.\n$ systemctl --user list-unit-files\n<\/pre><p>From these commands I identified various services I want to disable. The trick\nto disabling these is to use the <code>systemctl --user mask<\/code> command, which inhibits\nthe system installed service units from running. In theory masking services\ncould cause problems, but at least for the set presented here I haven\u2019t observed\nany issues.<\/p><p>Here\u2019s my Ansible code:<\/p><pre class=\"language-yaml\">- name: mask unwanted gnome services\n  systemd:\n    name: \"{{item}}\"\n    user: yes\n    masked: yes\n    state: stopped\n  loop:\n    - evolution-addressbook-factory.service\n    - evolution-calendar-factory.service\n    - evolution-source-registry.service\n    - gvfs-goa-volume-monitor.service\n    - tracker-store.service\n<\/pre><p>This list lobotomizes <code>evolution-data-server<\/code>, GNOME Online Accounts, and\nTracker. I also have an Ansible command to ensure that the tracker database is\ncompletely removed from my system:<\/p><pre class=\"language-yaml\">- name: remove tracker databases\n  file:\n    name: ~\/.cache\/tracker\n    state: absent\n<\/pre><p>I\u2019ll note that even after removing these services there are still a lot of GNOME\nuser services running in my session. If you wanted to you could mask a lot more\nservices. In general the approach I take is to only mask services whose behavior\nI understand. Most of the remaining services in my GNOME session are daemons\nthat use very little memory and zero CPU, so I don\u2019t mind keeping them around.<\/p><h2>Firefox\/GNOME Integration<\/h2><p>I version control my Firefox preferences using Firefox\u2019s obscure\n<a href=\"http:\/\/kb.mozillazine.org\/User.js_file\">user.js<\/a> feature, which essentially\nlets you put <code>about:config<\/code> preferences in a Javascript file. I have a fairly\nextensive user.js file that is a relaxed version of options suggested by\n<a href=\"https:\/\/github.com\/pyllyukko\/user.js\">pyllyukko\/user.js<\/a>, which I might write\nabout in more detail another time. Two of the options I set in this file are\nspecific to GNOME, so I\u2019ll cover them here.<\/p><p>The <a href=\"https:\/\/extensions.gnome.org\/\">GNOME Extensions<\/a> site lets you install\nGNOME extensions from Firefox. This is done using a native plugin, meaning that\nunlike normal extensions it runs in a privileged context that allows it to\ninteract with GNOME. Since I have user extensions disabled anyway this serves no\npurpose, and could possibly do nasty things if a security flaw is found in it\nlater. Therefore I disable this plugin:<\/p><pre class=\"language-js\">\/\/ Disable GNOME browser plugin.\nuser_pref(\"plugin.state.libgnome-shell-browser-plugin\", 0);\n<\/pre><p>The second option is related to a Firefox bug (present as of Firefox 60.0)\nrelated to GNOME 3.28 when using the Adwaita dark theme. When using this theme\nFirefox will render certain widgets\/text areas with inverted colors. This is\nmore than just strange looking: in some cases it can cause things to be\ncompletely unreadable (e.g. a white text color can get applied to text in a\nwhite text area). This can be fixed by forcing Firefox to render widgets use the\nAdwaita light theme:<\/p><pre class=\"language-js\">\/\/ Force Adwaita light theme.\nuser_pref(\"widget.content.gtk-theme-override\", \"Adwaita:light\");\n<\/pre><h2>Parting Thoughts<\/h2><p>The configuration listed above is the accumulation of a few years of efforts for\nme. Things change from release to release, and therefore I revisit things from\ntime to time. Generally though I\u2019m very happy with the direction GNOME has been\ndoing, and it\u2019s a much more pleasant experience dealing with certain aspects of\nit today than it was a few years ago. I hope the information presented here is\nuseful, and I hope to see more great things from the GNOME project in the\nfuture.<\/p>","created_at":"2018-11-07T16:23:34+0000","updated_at":"2018-11-07T16:23:34+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"eklitzke.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":154,"title":"Experimental music notation resources","url":"https:\/\/llllllll.co\/t\/experimental-music-notation-resources\/149?_escaped_fragment_=","content":"<div class=\"post\" itemprop=\"articleBody\"><p>As somebody with one foot in the visual art world and the other in sound world, I have always been interested in the experimental music notation strategies. Aside from the fact that these are often beautiful artifacts, I love thinking about the \u201cutility\u201d of images: how\/when image becomes an \u201cinstruction\u201d.<\/p>\n<p>I think the new |||||||| forum, with it\u2019s expanded conceptual ambition, ought to have a thread devoted to this aspect of new and experimental music.<\/p>\n<p>Let\u2019s share resources, files, images, links and ahything you might have collected related to this esotheric and fascinating point on confluence between image, language ne sound.<\/p>\n<p>p.<\/p>\n<p>To get things started, here is a nice article from Simthsonian:<br \/><\/p><aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"https:\/\/static-media.smithsonianmag.com\/img\/favicon.ico\" class=\"site-icon\" width=\"16\" height=\"16\" alt=\"image\" \/><a href=\"https:\/\/www.smithsonianmag.com\/arts-culture\/5-12-examples-of-experimental-music-notation-92223646\/\" target=\"_blank\" rel=\"nofollow noopener\">Smithsonian<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/thumbs-prod.si-cdn.com\/YfEcEmAAFtq1KtIAZORrr_WpEMY=\/1072x720\/filters:no_upscale()\/https:\/\/public-media.smithsonianmag.com\/filer\/Design-Decoded-Experimental-Music-631.jpg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"https:\/\/www.smithsonianmag.com\/arts-culture\/5-12-examples-of-experimental-music-notation-92223646\/\" target=\"_blank\" rel=\"nofollow noopener\">5 1\/2 Examples of Experimental Music Notation<\/a><\/h3>\n<p>In the 1950s progressive composers broke from the 5 line music staff to experiment with new, more expressive forms of graphic music notation<\/p>\n  <\/article>\n  \n<\/aside>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/129\/2abc3a11c4fa3df8.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/1ba417cc7055250dab369d201624e1b8ed0354b6\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/1ad\/92b\/02c7b00782_666x500.jpg\" width=\"666\" height=\"500\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>immediately reminded me of jobina tinnemans\u2019 style of notation<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/130\/428c7328c89a148b.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/d58e3ad43c36ae68012d280e89c51c9a189a6693\" title=\"jt.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/ed8\/47c\/6c3d7c4289_500x500.jpg\" width=\"500\" height=\"500\" alt=\"image\" \/><\/a><\/div> \n<aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"https:\/\/s2.wp.com\/i\/favicon.ico\" class=\"site-icon\" width=\"32\" height=\"32\" alt=\"image\" \/><a href=\"https:\/\/jobinatinnemans.com\/\" target=\"_blank\" rel=\"nofollow noopener\">JOBINA TINNEMANS<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/jobinatinnemans.files.wordpress.com\/2015\/11\/jobina-tinnemans-owen-lucas-photography.jpeg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"https:\/\/jobinatinnemans.com\/\" target=\"_blank\" rel=\"nofollow noopener\">JOBINA TINNEMANS<\/a><\/h3>\n<p>Contemporary Composer<\/p>\n  <\/article>\n  \n<\/aside><\/div><div class=\"post\" itemprop=\"articleBody\"><p>This is exactly what I am hoping we could share here <a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/glia\">@glia<\/a> !<\/p>\n<p>I also should have mentioned specifically: the image I included in the first post is an example of many amazing experimantal notations by Cornelius Cardew.<\/p>\n<p>Here are a few more from Cardew:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/131\/1d566632c5f90a8c.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/506f699a77d6d9a91e57a9e4ca4d7fd35a51078e\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/cf8\/3c5\/58c23dae1a_690x451.jpg\" width=\"690\" height=\"451\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/132\/707501934fdae508.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/e9c1447ae7517229ea1a568de32cda8d0628d661\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/847\/db4\/9d3162a7fc_690x388.jpg\" width=\"690\" height=\"388\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/133\/6b0c53c95e2a8369.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/ddaa4fb6c0b1d41f3e6437e2e059553a517e547a\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/bb0\/57d\/cc9ae5441e_690x404.jpg\" width=\"690\" height=\"404\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>sweet looking stuff<\/p>\n<p>in the midst of read more on this\u2026<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>This kind of thing sends me.<\/p>\n<p>Leonardo Music Journal did an issue called \u201cBEYOND NOTATION: COMMUNICATING MUSIC\u201d:<\/p>\n<p><a href=\"http:\/\/www.leonardo.info\/isast\/journal\/toclmj21.html\" class=\"onebox\" target=\"_blank\" rel=\"nofollow noopener\">http:\/\/www.leonardo.info\/isast\/journal\/toclmj21.html<\/a><\/p>\n<p>It\u2019s not open access, and it\u2019s kind of expensive. I can comment further on particular articles if anyone\u2019s interested.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>beautiful<\/p>\n<p>Iannis Xenakis worked a lot around this : <a href=\"http:\/\/www.centre-iannis-xenakis.org\/cix_upic_presentation\">http:\/\/www.centre-iannis-xenakis.org\/cix_upic_presentation<\/a><br \/>(it\u2019s french but one can easily find resources about I. Xenakis)<br \/>the software he used is called UPIC.<\/p>\n<p>and there is <a href=\"http:\/\/www.iannix.org\/en\/whatisiannix\/\">http:\/\/www.iannix.org\/en\/whatisiannix\/<\/a><br \/>it\u2019s a sequencer software based on drawings. very very interesting compositionnal tool<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Wonderful references guys!<br \/><a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/beo\">@beo<\/a> great to see the UPIC legacy carried into the new dimensions with Iannix.<br \/>Here are a few images of Xenakis notations for visual pleasure:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/144\/1aabdffd84f29971.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/fdaca252d71f0a52a5f702c7e6f2a603e08edb55\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/f4d\/af2\/23837064ce_690x424.jpg\" width=\"690\" height=\"424\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/145\/3cb8b5f88ccaf2ff.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/36fa17127637bc35e89fc40136cd29a951a6f3e6\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/2d0\/3af\/fbd2826b42_633x499.jpg\" width=\"633\" height=\"499\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/146\/b14e50e190d4f1c2.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/61f6916c176bbbf51d5262871507711177f3f111\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/e26\/919\/d686532351_690x350.jpg\" width=\"690\" height=\"350\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Als, here is a nice collection of images with unorthodox notations:<\/p>\n<aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><a href=\"http:\/\/blog.wfmu.org\/freeform\/2007\/01\/gallery_of_musi.html\" target=\"_blank\" rel=\"nofollow noopener\">WFMU's Beware of the Blog<\/a>\n  <\/header><article class=\"onebox-body\"><img src=\"http:\/\/blog.wfmu.org\/photos\/uncategorized\/bernard_rands__as_all_thumb_1.jpg\" class=\"thumbnail onebox-avatar\" width=\"150\" height=\"150\" alt=\"image\" \/><h3><a href=\"http:\/\/blog.wfmu.org\/freeform\/2007\/01\/gallery_of_musi.html\" target=\"_blank\" rel=\"nofollow noopener\">Gallery of Graphic Musical Notation<\/a><\/h3>\n<p>Here's 24 examples of unusual or graphic musical notation, from the cuneiform markings of the oldest song in the world, to the graphic notation of Stockhausen, Crumb, Cage and Eno, to the painted musical inventions of Adolph Wolfli. Click each...<\/p>\n  <\/article>\n  \n<\/aside><p>Like this one from George Crumb:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/147\/7fe470a2b4fc0248.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/6d55016989181410e80584965695ed232d98b2ad\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/10e\/305\/6ccfcd031c_493x500.jpg\" width=\"493\" height=\"500\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/148\/06fac602da575cbb.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/50a385d6a0ed89adfb25bcca7da1cfbd067dd0e3\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/86e\/f10\/a3f4ba0fd2_570x500.jpg\" width=\"570\" height=\"500\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/149\/47243fb9970dbf91.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/e9f0b4704bbe6732a5e2ef0838efa601a4ab13d7\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/bae\/120\/fc11a6b73b_690x332.jpg\" width=\"690\" height=\"332\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>ohh, i especially like the circular notation ones. possible new profile picture <img src=\"https:\/\/llllllll.co\/images\/emoji\/twitter\/+1.png?v=5\" title=\":+1:\" class=\"emoji\" alt=\":+1:\" \/><\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p><a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/laborcamp\">@laborcamp<\/a> :   thanks for the picture <img src=\"https:\/\/llllllll.co\/images\/emoji\/twitter\/smile.png?v=5\" title=\":smile:\" class=\"emoji\" alt=\":smile:\" \/><br \/>actually a new UPIC is under dev by students in Rouen univ. the goal is to make a UPIC for win, mac and linux. they call it UPIX.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>This is a super cool\/useful book on the matter:<br \/><\/p><aside class=\"onebox amazon\"><header class=\"source\"><a href=\"https:\/\/www.amazon.co.uk\/Notations-21-Theresa-Sauer\/dp\/0979554640\" target=\"_blank\" rel=\"nofollow noopener\">amazon.co.uk<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/images-na.ssl-images-amazon.com\/images\/I\/512mPwCOojL._SY400_.jpg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"https:\/\/www.amazon.co.uk\/Notations-21-Theresa-Sauer\/dp\/0979554640\" target=\"_blank\" rel=\"nofollow noopener\">Notations 21: Amazon.co.uk: Theresa Sauer: 9780979554643: Books<\/a><\/h3>\n<p>\nDrawing inspiration from John Cages Notations, Notations 21 features illustrated musical scores from more than 100 composers from every continent, all of whom are making amazing breakthroughs in the a ...\n<\/p>\n  <\/article>\n  \n<\/aside>\n<p>A ton of fascinating scores in there.<\/p>\n<p>On a more digital front, there\u2019s lots of dynamic\/generative score stuff going on nowadays too.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p><a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/rodrigo\">@Rodrigo<\/a> yes!<br \/>Funny you posted this, as I was recently looking to get one of these, until I noticed the price! In US this book is listed as over $600!<br \/>Certainly prohibitive for me <img src=\"https:\/\/llllllll.co\/images\/emoji\/twitter\/smile.png?v=5\" title=\":smile:\" class=\"emoji\" alt=\":smile:\" \/><br \/>But it really looks like an awesome compendium of exactly the kind of materials I thought we could share here\u2026<br \/>Do you actually own a copy?<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Holy shit that\u2019s crazy!<\/p>\n<p>Yeah I own the book. Bought it (in the UK) a few years ago.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Lucky you <a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/rodrigo\">@Rodrigo<\/a> !<br \/>So, obviously, you must think it\u2019s worth the price?<br \/>I would LOVE to have this in my library!..<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>DEFINITELY not worth that price, lol.<\/p>\n<p>I don\u2019t remember how much it was, but it was when I got it but it was no more than like \u00a340-50 or so. I\u2019ve got the hardcover version too. Weird\u2026<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>this thread is simply amazing! i wish it was around a couple of years ago\u2026 i was working on a project and looked into experimental notation for a commission i was doing. its a pretty tenuous connection, but i wrote a little article about the piece here:<\/p>\n<aside class=\"onebox whitelistedgeneric\"><header class=\"source\"><img src=\"http:\/\/www.juggle.org\/\/juggle.org\/wp-content\/uploads\/2016\/07\/IJA_original.png\" class=\"site-icon\" width=\"\" height=\"\" alt=\"image\" \/><a href=\"http:\/\/www.juggle.org\/center-of-gravity\/\" target=\"_blank\" title=\"08:30AM - 28 February 2014\" rel=\"nofollow noopener\">IJA \u2013 28 Feb 14<\/a>\n  <\/header><article class=\"onebox-body\"><div class=\"aspect-image\"><img src=\"https:\/\/www.juggle.org\/wp-content\/uploads\/2014\/02\/COGFeatured-1.jpg\" class=\"thumbnail\" alt=\"image\" \/><\/div>\n<h3><a href=\"http:\/\/www.juggle.org\/center-of-gravity\/\" target=\"_blank\" rel=\"nofollow noopener\">Center of Gravity<\/a><\/h3>\n<p>22 jugglers and 2 musicians took the stage on November 12, 2013, to perform a show called<\/p>\n  <\/article>\n  \n<\/aside><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Hi <a class=\"mention\" href=\"https:\/\/llllllll.co\/u\/instantjuggler\">@instantjuggler<\/a> , thank you for the link! It was very interesting to read such a detailed insight into the process of putting together a large scale juggling performance. It made me think of your collaboration with Pekka as well (thinking about themusic and juggling relationships, but also the ability to effortlessly move between scored events and improvisation).<br \/>I also love how you refer to the idea of constructing choreographed images on stage. I always think of performance in terms of images as well, regardless of if I am working with dancers or volunteer workers, or by myself.<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Here are a few from John Cage:<\/p>\n<div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/176\/b48a40477c97b326.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/db02ffd128beacd134d5d056f75865ff2f6f188d\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/aec\/efc\/9379abd1bb_690x447.jpg\" width=\"690\" height=\"447\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/177\/1705fbeee47a0cc1.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/d0941095fc14fa94e5c782750ac4c4f1c4100de6\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/b58\/409\/179a4326ea_690x449.jpg\" width=\"690\" height=\"449\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/178\/38b5633016e3568c.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/dbbd38eae19fe9d2469646e3b7d34409d4f52a2a\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/379\/e51\/41b1323e7b_690x441.jpg\" width=\"690\" height=\"441\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/179\/36b30b6e37030d40.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/ad6978445ecc003f028a6836ac89be34fcf22aae\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/39f\/a40\/25a1b65a1f_690x343.jpg\" width=\"690\" height=\"343\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/180\/88787074fecb636f.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/00521422466490af946811a3199bdeec56306f30\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/6d5\/ccc\/7087ce561c_690x479.jpg\" width=\"690\" height=\"479\" alt=\"image\" \/><\/a><\/div><br \/><div class=\"lightbox-wrapper\"><a class=\"lightbox\" href=\"https:\/\/llllllll.co\/uploads\/default\/181\/41ecaa742f9f9cfa.jpg\" data-download-href=\"\/\/llllllll.co\/uploads\/default\/7450267ad87dbdf122e2e4d10aea4a94cf9232d7\" title=\"image.jpg\"><img src=\"https:\/\/llllllll.co\/uploads\/default\/_optimized\/c83\/e02\/11d8646db5_492x500.jpg\" width=\"492\" height=\"500\" alt=\"image\" \/><\/a><\/div><\/div><div class=\"post\" itemprop=\"articleBody\"><p>Woah that last one is crazy\u2026Most of the ones posted so far are interesting if not still stuck in a linear left to right structure. How are you supposed to interpret what looks like molecules and an unfinished pentagram?<\/p>\n<p>Cool stuff!<\/p><\/div><div class=\"post\" itemprop=\"articleBody\"><p>The instructions from that Cage piece is fucking mad too. The notation is weird, but each one of those little diagrams(\/pieces) has equally crazy instructions.<\/p><\/div>","created_at":"2018-11-07T16:23:34+0000","updated_at":"2018-11-07T16:23:34+0000","published_at":"2015-03-28T16:43:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"llllllll.co","preview_picture":"https:\/\/static-media.smithsonianmag.com\/img\/favicon.ico","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":151,"title":"Comodo PositiveSSL Certificates for as low as $5.88\/yr","url":"https:\/\/www.ssls.com\/ssl-certificates\/{{ngMeta['og:url']}}","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:23:33+0000","updated_at":"2018-11-07T16:23:33+0000","annotations":[],"reading_time":0,"domain_name":"www.ssls.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":152,"title":"CA certificates trusted by Gmail for S\/MIME","url":"https:\/\/support.google.com\/a\/answer\/7448393?hl=en","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:23:33+0000","updated_at":"2018-11-07T16:23:33+0000","annotations":[],"reading_time":0,"domain_name":"support.google.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":148,"title":"Use Emacs Org Mode and REST APIs for an up-to-date Stock Portfolio","url":"http:\/\/www.sastibe.de\/2018\/05\/2018-05-11-emacs-org-mode-rest-apis-stocks\/","content":"<p>A couple of weeks ago, I started to work with <a href=\"https:\/\/www.gnu.org\/software\/emacs\/\" title=\"Emacs\">Emacs<\/a>, and I grow fonder of it every day. During a very short time period, it has become my go-to editor for nearly everything I do on my computer, including (but not limited to)<\/p>\n<ul><li>planning my Todos (in <a href=\"https:\/\/orgmode.org\/\" title=\"org-mode\">org-mode<\/a>, to be precise),<\/li>\n<li>setting up my agenda (org-mode again),<\/li>\n<li>taking memos during meetings<\/li>\n<li>writing my (longer) e-mails<\/li>\n<li>play around with new stuff<\/li>\n<li>write blog posts (this is the first of these...)<\/li>\n<\/ul><p>It is difficult to pin down exactly why Emacs is taking over so much. My main influences for starting with EMACS were blogposts, <a href=\"https:\/\/blog.fugue.co\/2015-11-11-guide-to-emacs.html\" title=\"the first\">the first<\/a> describing a general EMACS setup, <a href=\"https:\/\/emacs.cafe\/emacs\/orgmode\/gtd\/2017\/06\/30\/orgmode-gtd.html\" title=\"the second\">the second<\/a> detailing how to implement GTD, i.e. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Getting_Things_Done\" title=\"Getting Things Done\">Getting Things Done<\/a> in Emacs org-mode.<\/p>\n<p>In this post, I will demonsrate the strengths of using Emacs in a very specific use case: Getting up-to-date financial data to use in a spread-sheet including all your financial data. Applications like these are usually provided by online banks themselves, so that I don't show you anything particularly <strong>new<\/strong> or <strong>shiny<\/strong>. However, the ability to customize every step of the way brings with a number of advantages.<\/p>\n<h2 id=\"finding-a-rest-api-for-stock-quotes\">Finding a REST API for Stock Quotes<\/h2>\n<p>First, we need to get up-to-date financial data from a REST API. I decided to use Alpha Vantage, a site I first stumbled upon after reading <a href=\"http:\/\/www.financial-hacker.com\/bye-yahoo-and-thank-you-for-the-fish\/\" title=\"this blog post.\">this blog post.<\/a> The other APIs listed on that page had various issues, either being deprecated in the near future (google, yahoo, stooq) or not having a number of symbols (IEX). The API of Alphavantage is rather easy to understand, even though the naming convention is terrible. Try for instance <a href=\"https:\/\/www.alphavantage.co\/query?function=TIME_SERIES_INTRADAY&amp;symbol=MSFT&amp;interval=1min&amp;apikey=demo\" title=\"this link\">this link<\/a> using the demo API key, yielding the following result for the Microsoft stock:<\/p>\n<pre class=\"language-json\">{\n    \"Meta Data\": {\n        \"1. Information\": \"Intraday (1min) prices and volumes\",\n        \"2. Symbol\": \"MSFT\",\n        \"3. Last Refreshed\": \"2018-05-11 16:00:00\",\n        \"4. Interval\": \"1min\",\n        \"5. Output Size\": \"Compact\",\n        \"6. Time Zone\": \"US\/Eastern\"\n    },\n    \"Time Series (1min)\": {\n        \"2018-05-11 16:00:00\": {\n            \"1. open\": \"97.5900\",\n            \"2. high\": \"97.7300\",\n            \"3. low\": \"97.5750\",\n            \"4. close\": \"97.7000\",\n            \"5. volume\": \"3776187\"\n        },\n        \"2018-05-11 15:59:00\": {\n            \"1. open\": \"97.4800\",\n            \"2. high\": \"97.5900\",\n            \"3. low\": \"97.4700\",\n            \"4. close\": \"97.5900\",\n            \"5. volume\": \"257615\"\n        },...\n<\/pre>\n<h2 id=\"reading-api-requests-into-emacs-variables\">Reading API Requests into Emacs Variables<\/h2>\n<p>Having located the data out in the internet was a good first step, but now we need to figure out a way how to use this information. Luckily, most of the work needed for this can be found in various places on the net, for instance in <a href=\"https:\/\/vxlabs.com\/2017\/06\/03\/querying-restful-webservices-into-emacs-orgmode-tables\/\" title=\"this blog post\">this blog post<\/a>. I decided to follow their general setup, using the following packages:<\/p>\n<ul><li>org<\/li>\n<li>org-babel<\/li>\n<li>request<\/li>\n<li>json<\/li>\n<\/ul><p>The API used in their scenario gave different results with a much cleaner nomenclature. For the Alphavantage API, I had to become a little creative with the eLisp code.<\/p>\n<pre>(require 'request)\n(require 'json)\n(require 'cl)\n(request\n \"https:\/\/www.alphavantage.co\/query\"\n :params `((\"function\" . \"TIME_SERIES_INTRADAY\")\n           (\"symbol\" . \"SC0J\")\n           (\"interval\" . \"1min\")\n           (\"apikey\" . \"...\"))\n :parser 'json-read\n :success (function*\n           (lambda (&amp;key data &amp;allow-other-keys)\n             (setq open_sc0j (string-to-number (cdr (elt (elt (elt data 1) 1) 1)))))))\n<\/pre>\n<p>The variable <code>open_sc0j<\/code> is evaluated as follows: From the received json, take the second entry of the second element of the second element. Not very nice, but it works...<\/p>\n<p>I encountered a second difficulty in my portfolio: I have both European stocks (in EUR) and American stocks (traded in USD). In order to keep my balances comparable, I added yet another variable <code>rate_usd_eur<\/code>, which receives up-to-date exchange rates from USD to EUR from the appropriate query. All in all, my requests to Alphavantage look like this:<\/p>\n<pre>(request\n \"https:\/\/www.alphavantage.co\/query\"\n :params `((\"function\" . \"TIME_SERIES_INTRADAY\")\n           (\"symbol\" . \"SC0J\")\n           (\"interval\" . \"1min\")\n           (\"apikey\" . \"...\"))\n :parser 'json-read\n :success (function*\n           (lambda (&amp;key data &amp;allow-other-keys)\n             (setq open_sc0j (string-to-number (cdr (elt (elt (elt data 1) 1) 1)))))))\n(request\n \"https:\/\/www.alphavantage.co\/query\"\n :params `((\"function\" . \"CURRENCY_EXCHANGE_RATE\")\n           (\"from_currency\" . \"USD\")\n           (\"to_currency\" . \"EUR\")\n           (\"apikey\" . \"...\"))\n :parser 'json-read\n :success (function*\n           (lambda (&amp;key data &amp;allow-other-keys)\n             (setq rate_usd_eur (string-to-number (cdr (elt (elt data 0) 5)))))))\n(request\n \"https:\/\/www.alphavantage.co\/query\"\n :params `((\"function\" . \"TIME_SERIES_INTRADAY\")\n           (\"symbol\" . \"PG\")\n           (\"interval\" . \"1min\")\n           (\"apikey\" . \"...\"))\n :parser 'json-read\n :success (function*\n           (lambda (&amp;key data &amp;allow-other-keys)\n             (org-table-iterate-buffer-tables)\n             (setq open_prg (* rate_usd_eur (string-to-number (cdr (elt (elt (elt data 1) 1) 1))))))))\n<\/pre>\n<p>Putting this code inside an org-mode file, bracketing it by code blocks <code>#+BEGIN_SRC emacs-lisp :results none<\/code> and <code>#+END_SRC<\/code>, and hitting \"C-c C-c\" inside it leads to the evaluation of the code block and thus the filling of the variables <code>open_sc0j<\/code>, <code>rate_usd_eur<\/code> and <code>open_prg<\/code>. Since we included the wonderful little function <code>org-table-iterate-buffer-tables<\/code>, the evaluation also repeats until all the columns in the table below are calculated correctly. This neat little trick I also copied from <a href=\"https:\/\/vxlabs.com\/2017\/06\/03\/querying-restful-webservices-into-emacs-orgmode-tables\/\" title=\"here.\">here.<\/a><\/p>\n<h2 id=\"setting-up-a-custom-stock-portfolio-org-table\">Setting up a Custom Stock Portfolio Org Table<\/h2>\n<p>After these steps, we now set up an org-table to give us a customizable overview of how our stocks are doing. That means setting up an org-table with columns for historic data, such as the date of the buy. Additionally, we use the <code>#+TBLFM<\/code> function to calculate appropriate performance indicators. An example for such functions:<\/p>\n<pre>| Stock         | Symbol | Amt. |    Buy | Date Bought     |  Fees | Dividends | Close |   Gain | Gain Perc | Gain per Day |\n|---------------+--------+------+--------+-----------------+-------+-----------+-------+--------+-----------+--------------|\n| MSCI ETF      | SC0J   |   10 |  47.11 | [2018-04-16 Mo] | 12.35 |         0 |       |        |           |              |\n| ProcterGamble | PG     |    5 | 65.014 | [2015-10-01 Do] | 10.61 |     72.03 |       |        |           |              |\n#+TBLFM: $9=(-$4 + $8)*$3 - $6 + $7;%0.3f::$10=100*$9\/($4*$3)::$11=$9\/(now() - $5)::@2$8='(format \"%f\" open_sc0j)::@3$7=17.78 + 17.70 + 17.94 + 18.61::@3$8='(format \"%f\" open_prg)\n<\/pre>\n<p>This code leads to the following result:<\/p>\n<table><thead><tr><th>Stock<\/th>\n<th>Symbol<\/th>\n<th>Amt.<\/th>\n<th>Buy<\/th>\n<th>Date Bought<\/th>\n<th>Fees<\/th>\n<th>Dividends<\/th>\n<th>Close<\/th>\n<th>Gain<\/th>\n<th>Gain Perc<\/th>\n<th>Gain per Day<\/th>\n<\/tr><\/thead><tbody><tr><td>MSCI ETF<\/td>\n<td>SC0J<\/td>\n<td>10<\/td>\n<td>47.11<\/td>\n<td>[2018-04-16 Mo]<\/td>\n<td>12.35<\/td>\n<td>0<\/td>\n<td>49.290000<\/td>\n<td>9.450<\/td>\n<td>2.0059435<\/td>\n<td>0.35552367<\/td>\n<\/tr><tr><td>ProcterGamble<\/td>\n<td>PG<\/td>\n<td>5<\/td>\n<td>65.014<\/td>\n<td>[2015-10-01 Do]<\/td>\n<td>10.61<\/td>\n<td>72.03<\/td>\n<td>61.342605<\/td>\n<td>43.063<\/td>\n<td>13.247301<\/td>\n<td>0.045111962<\/td>\n<\/tr><\/tbody><\/table><p>In this example, <code>Gain<\/code> is first calculated by multiplying <code>-$4 + $8<\/code>, i.e. the difference between Buy and (today's) Close by the amount of stocks bought. Additionally, any dividends are added and any fees are subtracted, yielding a \"net gain\" for the stock. In column <code>Gain per Day<\/code>, this number is broken down per day since I bought the stock, highlighting my most efficient assets.<\/p>\n<p>There is no limit to what types of functions one can use, and no limit on the sophistication of analysis. And all of this within a very light-weight, easy-to-use interface, without any unnecessary over-head. It is not only convenient, but also educational: while writing this article, I learned a lot about REST APIs and financial data.<\/p>\n<p>Let me conclude this article by picking up a picture from <a href=\"https:\/\/blog.fugue.co\/2015-11-11-guide-to-emacs.html\" title=\"one of the blog posts that got me into Emacs in the first place\">one of the blog posts that got me into Emacs in the first place<\/a>: Emacs is like a classical steel frame road bike, reliant, robust, nothing fancy but easy to repair. It is the ideal tool to explore the wilderness of the internet. And I can only invite everybody else to come along for the ride.<\/p>\n<p><img src=\"http:\/\/vg09.met.vgwort.de\/na\/d161742c2abc4c1da54c58f5b3c2e753\" width=\"1\" height=\"1\" alt=\"\" \/><\/p>","created_at":"2018-11-07T16:23:26+0000","updated_at":"2018-11-07T16:23:26+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"www.sastibe.de","preview_picture":"http:\/\/res.cloudinary.com\/dlprdrxib\/image\/upload\/v1515794260\/profile_picture_xjw3jw.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":149,"title":"Multi-Provider Email API and Bounce Handler","url":"https:\/\/www.flutemail.com\/","content":"Flute: Multi-Provider Email API and Bounce Handler\n        <p>Flute Mail has shut down. <a href=\"https:\/\/www.flutemail.com\/blog\/2018\/11\/06\/flute-has-shut-down.html\">Learn more.<\/a>\n    Flute Mail has shut down. <a href=\"https:\/\/www.flutemail.com\/blog\/2018\/11\/06\/flute-has-shut-down.html\">learn more<\/a><\/p>\n\n        <div class=\"wrapper\"><div class=\"homepage\"><div class=\"hero\"><div class=\"container\"><div class=\"row\"><div class=\"col-sm-7 col-xs-12\"><div class=\"iphone\"><img src=\"https:\/\/www.flutemail.com\/img\/iphone.png\" alt=\"image\" \/><\/div><\/div><\/div><\/div><\/div><div class=\"provider-list\"><h2>A good email provider isn't enough. <small>You also need delivery optimization services.<\/small><\/h2><small>Bounce handling and retry functionality should be independent from your email provider. <a href=\"https:\/\/www.flutemail.com\/features#unbouncer\" class=\"text-gold\">Learn more \u203a<\/a><\/small><\/div><div class=\"deliverability\" id=\"how-it-works\"><div class=\"lines\"><img src=\"https:\/\/www.flutemail.com\/img\/lines.png\" alt=\"image\" \/><\/div><div class=\"container\"><div class=\"row flex\"><div class=\"col-md-6\"><h2> How It Works<\/h2><\/div><\/div><div class=\"row flex\"><div class=\"col-md-9 col-sm-12 no-pad\"><div class=\"row\"><div class=\"col-md-5 col-sm-5 col-xs-12 environment-section\"><p>Critical Emails<\/p><\/div><div class=\"col-md-3 col-sm-3 col-xs-4 provider\"><h4 class=\"visible-xs\">Providers<\/h4><p><i class=\"material-icons\">close<\/i>SendGrid<\/p><p><i class=\"material-icons\">check<\/i>Private Servers<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-5 col-sm-5 col-xs-12 environment-section\"><p>Transactional Emails<\/p><\/div><div class=\"col-md-3 col-sm-3 col-xs-4 provider\"><h4 class=\"visible-xs\">Providers<\/h4><p><i class=\"material-icons\">close<\/i>Amazon SES<\/p><p><i class=\"material-icons\">check<\/i>Postmark<\/p><\/div><\/div><div class=\"row\"><div class=\"col-md-5 col-sm-5 col-xs-12 environment-section\"><p>Promotional Emails<\/p><\/div><div class=\"col-md-3 col-sm-3 col-xs-4 provider\"><h4 class=\"visible-xs\">Providers<\/h4><p><i class=\"material-icons\">close<\/i>Postmark<\/p><p><i class=\"material-icons\">check<\/i>SparkPost<\/p><\/div><\/div><\/div><\/div><\/div><\/div><div class=\"request-demo\"><div class=\"container\"><div class=\"col-md-7 col-md-offset-3 col-sm-11 col-sm-offset-1\"><h2>Ready to optimize your delivery?\n            <small>Hook up your provider in minutes.<\/small>\n        <\/h2><p><a href=\"https:\/\/www.flutemail.com\/pricing\" class=\"btn btn-md btn-gold-gradient\">See plans &amp; pricing<\/a><\/p><small>Drop-in dashboard for your provider.<\/small><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:23:26+0000","updated_at":"2018-11-07T16:23:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.flutemail.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":150,"title":"Mailing lists vs Github","url":"https:\/\/begriffs.com\/posts\/2018-06-05-mailing-list-vs-github.html","content":"<p>Most fledgling open source projects use Github or Gitlab to collaborate on code. However there\u2019s an older method that developers ought to know about because it offers some advantages.<\/p><div class=\"figure\"><img src=\"https:\/\/begriffs.com\/images\/gh-vs-mail.png\" alt=\"Github vs Email\" \/><p class=\"caption\">Github vs Email<\/p><\/div><p>The alternative method is the developer mailing list. It arose in the late eighties to early nineties, and predates the popularity of the web browser. But far from being a mere historical curiosity, the discussion list is still the primary method of development in many important open source projects, from databases to operating systems to web browsers.<\/p><p>In this article I\u2019ll carefully compare the use of mailing lists with code collaboration web sites such as Github. I\u2019ll do my best to present the pros and cons of each, so that projects assessing the two can make an informed decision.<\/p><h3 id=\"briefly-how-does-each-work\">Briefly, how does each work?<\/h3><p>Github hosts a git server, a bug tracker, a wiki, as well as release artifacts. Participation in discussions or code changes requires a Github identity. Code changes are proposed by making another Github-hosted project (a \u201cfork\u201d), modifying a remote branch, and using the GUI to open a pull request from your branch to the original. Project members debate the change in comments, with subsequent changes pushed as commits to the pull request branch. Ultimately a project owner can decide to merge the commits.<\/p><p>A mailing list maintains a list of subscribers who receive all messages sent to a \u201creflector\u201d email address. A project must host its own version control server elsewhere. Code changes are proposed by sending an email to the reflector, with a textual diff of the changes either attached or inline. List members debate the change with email replies, sending subsequent changes as more diffs, either to replace or amend the original. Ultimately a project committer can decide to apply the patch(es) to the codebase.<\/p><h3 id=\"advantages-of-mailing-lists\">Advantages of mailing lists<\/h3><h4 id=\"precise-flexible-communication\">Precise, flexible communication<\/h4><p>Mailing lists have three properties that allow developers to communicate with greater precision and flexibility than is possible in the Github UI: the threaded nature of emails, the user generation of patches, and the free mixing of discussion and patches.<\/p><h5 id=\"threading\">Threading<\/h5><p>The Github web UI has a limited model of threading, where there is one thread for the pull request as a whole, and other threads attached to single changed lines of code. The UI shows a few lines of code for context around the line chosen for discussion, but the full chunk of code under discussion must be inferred.<\/p><p>Email sub-threads allow specialized discussion about different aspects or sections of the code. A linear Github-style discussion would mix those conversations.<\/p><p>Conversations over email have a greater permanence and reference value than Github comments. On Github, comments continually change. They become \u201coutdated\u201d and disappear when attached to a line that has been changed. Same for the commits, which vanish after a force-push to the pull request branch. In an email thread, by contrast, the original messages and proposed changes remain for comparison with later messages and patches.<\/p><p>Furthermore, patches from multiple authors can\u2019t mix in a Github pull request. The person opening the pull request \u201cowns\u201d it. Other participants can drop code suggestions as comments, but what gets merged by the UI is ultimately the commits on the branch. The pull requester must turn those comments into commits on the branch if he or she wants to incorporate the suggested changes.<\/p><p>On a mailing list, anyone is free to reply to an existing thread with a new proposed set of patches. Another nice effect is that other people can carry the patch to the finish line if the original author stops caring or being involved. (I\u2019ve encountered a number of MIA authors in pull requests on my projects over the years.)<\/p><h5 id=\"patch-format\">Patch Format<\/h5><p>Custom patches are another way that the email process improves communication over the Github UI. Contributors can shape the patches to make them more readable for other people.<\/p><p>With Github you commit your changes and then Github chooses how to render the diff to the reviewer, but when creating the patch yourself you can tweak the settings. Diffs can be in <em>unified<\/em> or <em>context<\/em> format, and for the latter the number of context lines is configurable. (To be fair, Github allows expanding the view of context around changes too.)<\/p><p>Some people feel that git-style unified diffs are clearest for small changes of a few lines, and context diffs are more suitable for changing larger ranges, especially lines with indentation changes. (To be fair, I should note that Github will hide indentation changes if you add <code>?w=1<\/code> to the URL displaying the change.)<\/p><p>Finally, creating patches by hand with <code>diff<\/code> can take advantage of other options, like annotating the hunk headers with which C functions are affected (using the <code>-p<\/code> flag), and the choice of <a href=\"http:\/\/fabiensanglard.net\/git_code_review\/diff.php\">various algorithms<\/a> such as \u201cpatience.\u201d<\/p><p>Patience does a little more work to calculate its results, but the output makes more sense for humans. For instance, a standard (non-patience) diff might look like this:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode diff\"> void func1() {\n     x += 1\n+}\n+\n+void functhreehalves() {\n+    x += 1.5\n }\n void func2() {\n     x += 2\n }<\/pre><\/div><p>Whereas a patience diff shows the function added as a unit:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode diff\"> void func1() {\n     x += 1\n }\n+void functhreehalves() {\n+    x += 1.5\n+}\n+\n void func2() {\n     x += 2\n }<\/pre><\/div><p>Finally, there\u2019s nothing to stop the mailing list style of collaboration from adapting to emerging diff formats, such as <a href=\"https:\/\/begriffs.com\/posts\/2014-04-08-pilgrimage-report-structural-merging.html\">structural diffs<\/a>. That\u2019s because patches in emails directly describe changes whereas Github renders commit deltas without nuance.<\/p><h5 id=\"patchdiscussion-mix\">Patch\/Discussion mix<\/h5><p>The bulk of communication around a new feature is in its proposal, not around the subsequent code. A contributor proposes the feature, justifies it, and confirms the scope before writing any code. The community responds, and often by the time anyone implements anything, the idea has shifted significantly from the start of the process.<\/p><p>This is again where emails fit the group discussion well, with inline replies and the threaded structure as mentioned above. When it\u2019s time to get coding there\u2019s no discontinuity of switching from an \u201cissue\u201d to a \u201cpull request.\u201d The author simply replies again with patches attached.<\/p><h4 id=\"control-and-customization\">Control and customization<\/h4><p>While web apps deliver a centrally-controlled user interface, native applications allow each person to customize their own experience. Open protocols like SMTP encourage a proliferation of clients. So it is with with mail clients, each providing its own functions, filters, folders, flags and fun. Mail clients provide ways to mark a message important, or set it back as unread. In general you don\u2019t have to unsuccessfully beg a central committee over five years for an interface change. (It happens, see this <a href=\"https:\/\/github.com\/isaacs\/github\/issues\/50\">petition<\/a> in an unofficial repo designed to supplicate the Github deities.)<\/p><p>Some people script their mail client so that they can apply patches with a keyboard shortcut, others go minimalist, and still others even use webmail. Each person is different, and so is their software, but the nature of the mailing list allows them all to work together.<\/p><p>Another area of control is the ability to search and interact with a mailing list while offline. This is often by choice rather than necessity. Urban and suburban areas usually have internet access available anywhere at any time, using cellular service if need be. Things are no longer as they once were with intermittent dial-up connections, but working offline is a choice that some people simply prefer.<\/p><p>Although git is itself distributed and operates locally without a network connection, Github requires connectivity to review issues and pull requests. Demanding connectivity for an essentially reflective solitary task is unreasonable.<\/p><p>Remember that being online applies in two directions: the developer\u2019s connection to the internet and also Github\u2019s uptime. Although the latter has been pretty stable for a while, I do remember a handful of times where my entire office got derailed by Github downtime. Taking Github down is a big attractive target for angry nations and ambitious hackers.<\/p><p>With a native email client you can review all emails and attachments offline. You can even send replies to messages offline and the client will queue them until internet access becomes available.<\/p><p>For the other side, the uptime of the mailing list server, mail sent to an offline server will eventually go through. Mail transfer agents along the email chain are designed to retry transmissions. In addition, developer lists recommend that users \u201cReply All\u201d to messages so that everyone involved in a thread as well as the list reflector address get the message. That way even if the reflector is offline, your email can go through to those most recently involved.<\/p><p>Ultimately the control and customizability enjoyed by the mailing list style of development comes from diverse tools built for open standards. Patchfiles are universally supported (git itself directly <a href=\"https:\/\/git-scm.com\/docs\/git-format-patch\">supports<\/a> <a href=\"https:\/\/git-scm.com\/docs\/git-am\">patches<\/a>), SMTP has an RFC. Tools can work together, rather than having a GUI locked in the browser.<\/p><h4 id=\"politics-and-profit\">Politics and Profit<\/h4><p>Git was originally created as an escape from BitKeeper, a version control system with associated centralized hosting. BitKeeper granted a free license for open source projects, and required projects to store their metadata on company servers.<\/p><p>What a twist of history, then, that users of git chose Github\u2026 a centralized host granting free licenses for open source projects, and requiring projects to store their metadata on company servers.<\/p><p>Github can legally delete projects or users with or without cause. This makes sense of course, since the projects are using computers owned by Github. Also, while not explicitly prohibiting development of competing projects like BitKeeper did, Github can still view the private source code of all companies who choose to host with them.<\/p><p>Really, what\u2019s the future of Github (or any profit-driven yet free code hosting service)? History provides the example of SourceForge, the premier code collaboration platform at the turn of the century. After acquisition by dice.com, and in a last-ditch effort to grab some more money, SourceForge hijacked hosted projects so that their installers would include adware and spyware.<\/p><p>When I first began notes for this article, the implications for Github were purely hypothetical, but even as of this writing news has emerged that Microsoft is buying Github. It doesn\u2019t feel far-fetched to suggest that Github may be moving into its \u201cSourceForge stage.\u201d<\/p><p>Let this lesson be broader than Github though. The same logic applies to other web platforms that centralize and combine programmer communication and hosting.<\/p><p>Some major open source projects like the PostgreSQL database played it safe and explicitly chose never to put their code on external company servers (other than for mirroring). Postgres also takes pains to place control in a governing body spanning multiple companies. No company is allowed a majority of committers, and the organization is designed to survive independently.<\/p><p>Not all projects are important enough to require this level of neutrality, but all projects deserve to control their own destiny.<\/p><p>The fundamental task we\u2019re talking about is asynchronous group communication of code changes. Patches are a universally understood way of describing code changes, and email is a universally understood method of communication. So it seems that this approach matches the problem at hand.<\/p><p>In fact emailing patches works with any version control system, not just Git. The OpenBSD project still happily uses mailing lists with CVS. Patch authors don\u2019t need any version control software installed at all. An author can download the source release tarball, make changes in the copy, capture the diff, and email it.<\/p><p>Sending and applying patches cuts out busywork like cleaning up remote branches after merge, or creating a local branch in the forked repo in preparation for a pull request. For comparison, I remember teaching a group of new programmers how to use Github, and was conscious at the time of all the weird steps I asked them to do.<\/p><p>There\u2019s also less busywork for finished communications. There aren\u2019t things to \u201cclean up\u201d like abandoned pull requests, merged branches, or issues to mark closed. The replies just stop on those threads. More on some downsides of this later, but for now look on the bright side.<\/p><p>Using email also decouples digital identity from the accounts on a site like Github, and ultimately places that trust in the DNS system managed by an international organization, ICANN. For more about this, see my article <a href=\"https:\/\/begriffs.com\/posts\/2016-07-08-returning-original-social-network.html\">Returning to the Original Social Network<\/a>. PGP provides a further guarantee of identity, verified through a decentralized web of trust. See my recording of Neal Walfield\u2019s talk <a href=\"https:\/\/begriffs.com\/posts\/2016-11-05-advanced-intro-gnupg.html\">An Advanced Intro to GnuPG<\/a>.<\/p><h3 id=\"challenges-for-mailing-lists\">Challenges for mailing lists<\/h3><p>The mailing list approach raises quite a few questions and difficulties. I\u2019ve done my best to try to acknowledge them and suggest possible solutions.<\/p><h4 id=\"unfamiliarity-and-obscurity\">Unfamiliarity and obscurity<\/h4><h5 id=\"variety-of-process\">Variety of process<\/h5><p>When you browse a Github project you know exactly how it works. Sure, the addons\/CI\/hooks may differ, and setting up the project might involve some contortions, but the process of browsing code, finding known bugs, and submitting changes is always the same. This is most definitely not true for projects with their own web presence and mailing lists.<\/p><p>I\u2019ve heard people go so far as to say, \u201cIf a project is not on Github then there is a 0% chance that I will contribute.\u201d To be honest I can\u2019t say that I\u2019ve contributed to non-Github projects either. After I migrated to Github from Google Code my only non-Github contributions (if you want to call them that) were filing a bug report for Chrome, and another for a tool on Gitlab.<\/p><p>One more source of unfamiliarity is that many people, especially younger devs, have never seen a real mail client like <code>mutt<\/code>. In their minds they may picture an overflowing GMail inbox full of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Posting_style#Top-posting\">top posted<\/a> replies all mixed together from different projects. That\u2019s not an enticing image. Perhaps this article can start these developers on the path to rediscovering the care and engineering that went into classic email clients (\u201cMUAs\u201d as they are called).<\/p><p>Given the unfamiliarity, a self-hosted project needs to communicate very clearly how it does collaboration. A mailing list alone is not enough; a project needs to explain itself on the web, and provide information about getting involved.<\/p><p>A final note about unfamiliarity. Beware that offline users can be misunderstood in today\u2019s connected age. It can look silly to write an email while offline, only to have it queue, send, and arrive after another better response renders yours unnecessary. It\u2019s probably wise to include a signature line saying something like, \u201cNote: composed offline, may lack recent context.\u201d<\/p><p>The next problem that projects outside Github face is the lack of social proof through star voting. It\u2019s harder for new projects to build reputation without stars. People take notice of a project with thousands of stars in a way that the mere project description doesn\u2019t evoke. While much established software like OSes, browsers, and databases do fine without the stars, they won their acceptance years ago, or are promoted by big companies.<\/p><p>One way to help self-hosted project popularity is to hook into Github without going too deep. What some projects like git and Linux do is run their own servers, but offer a read-only mirror on Github for public admiration and code browsing.<\/p><p>Two tools that can help are the <a href=\"http:\/\/nopullrequests.com\/\">Pull Request Rejection Bot<\/a> and the built-in setting to <a href=\"https:\/\/help.github.com\/articles\/disabling-issues\/\">disable issues<\/a>. After disabling pull requests and issues, the project README needs to include information about how and where to contribute.<\/p><p>The self-hosted git server will also need a post-commit hook to keep Github up to date. This command should do the trick in the hook:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode bash\">git push github_remote -f --mirror<\/pre><\/div><h4 id=\"disorganization\">Disorganization<\/h4><h5 id=\"patch-state\">Patch state<\/h5><p>A big difficulty for mailing lists is sharing the global state of email threads. Not the content of the messages, obviously, but the decisions that require action. Individual contributors can flag messages for action in their mail clients, but this status remains private for each individual. It\u2019s helpful to record publicly that a patch has been accepted for future application, or that a previously reported bug was fixed.<\/p><p>Projects do this in one of two ways: adding metadata to the git repository itself, or using an external tool that ties into the mailing list.<\/p><p>For instance PostgreSQL had a problem where patches hung around without being accepted or rejected (there are up to ~250 proposed patches to consider at one time!). So the team created <a href=\"https:\/\/commitfest.postgresql.org\/\">commitfest.postgresql.org<\/a>. Contributors now register the patches on this site for final review. Several times per year committers stop their own work and make time to accept\/defer\/reject\/apply the patches.<\/p><p><a href=\"http:\/\/jk.ozlabs.org\/projects\/patchwork\/\">Patchwork<\/a> is a similar tool that is not tied to Postgres and is more suitable for general use. Patchwork supplements a mailing list by subscribing to the list just like a person would, and capturing patches from the emails. For each patch it creates a web page. It doesn\u2019t fragment discussion because it doesn\u2019t allow commenting through the web interface, it merely reflects any comments from the emails, and allows maintainers to mark patches with a state such as Accepted, Rejected, or Under Review.<\/p><h5 id=\"comprehensive-bug-list\">Comprehensive bug list<\/h5><p>Here\u2019s something surprising I just learned: the PostgreSQL project does not have any place a new contributor can go to find a list of open bugs for the project. There is a separate mailing list, pgsql-bugs, where people report them, but there\u2019s nothing that ties that with the activity on pgsql-hackers to indicate whether the bug is resolved. Interestingly, OpenBSD also has a <code>bugs<\/code> email list and no central bug status.<\/p><p>Postgres is still searching for a system which matches how they like to work and which doesn\u2019t detract from the mailing lists. I don\u2019t believe OpenBSD desires a bug tracker. As they quipped on IRC, \u201cIf we did [have one] we would be FreeBSD, wouldn\u2019t we?\u201d<\/p><p>For projects that do want a bug tracker, one good contender is <a href=\"https:\/\/www.debian.org\/Bugs\/\">Debbugs<\/a> from the Debian project. It provides an email interface to manipulate the bugs and can be used for projects other than Debian.<\/p><p>Other bug systems work by adding files or objects into the project git repo itself. Thus fixing a bug on a branch could mean deleting the bug from the repo along with the other changes on that branch. Like email, this preserves offline access to bug information. Probably one of the best is <a href=\"http:\/\/bugseverywhere.org\/\">Bugs Everywhere<\/a>. Another pretty clean one is simply called <a href=\"https:\/\/github.com\/driusan\/bug\">bug<\/a>.<\/p><h5 id=\"message-history\">Message history<\/h5><p>Once you subscribe to a mailing list, new messages flow into your client and are available for searching. Hooray! But what about messages sent before you joined the list? Furthermore, how do you \u201clink\u201d to old messages from newer ones? Every issue on Github has its own URL, so issues can refer to each other, but how would an email say something like, \u201cwe already discussed this in message X, look there.\u201d?<\/p><p>The way this works is with the <code>Message-ID<\/code> header. Every email has one, and they are part of how replies work. They are big GUIDs that look like <code>8953.1527887111@sss.pgh.pa.us<\/code> or <code>CAFjFpReKyYrsUF8sP5GoPYyyp9ZSqm_mLeD8kQLigX-3CzDiUg@mail.gmail.com<\/code>. So an email could say, \u201cWe discussed this issue in 8953.1527887111@sss.pgh.pa.us, let\u2019s move the discussion there.\u201d<\/p><p>Reading a message sent prior to your joining the list requires going to the list archive on the web. Good archives provide links to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mbox\">mbox<\/a> files which can be imported into a native mail client. This makes the old messages available for search\/reply in the mail client. Archives also usually provide a web interface to read the messages as HTML, but navigating threads in these old interfaces is usually way more painful than in a good mail client.<\/p><p>As an example, check out the <a href=\"https:\/\/www.postgresql.org\/list\/pgsql-hackers\/\">archives<\/a> for pgsql-hackers. Notice the mbox file links. You can download an mbox for a month\u2019s worth of messages, for individual messages or for full threads. To navigate to a certain message id, use the url <code>https:\/\/www.postgresql.org\/message-id\/:the-id<\/code>. Try one of the ids I listed earlier. Similarly the Linux kernel mailing list archives can display a message by id via <code>lkml.kernel.org<\/code>.<\/p><p>Remember that all emails have a Message-ID header, this isn\u2019t something special having to do with mailing lists. If you want to <em>reply<\/em> to a message that\u2019s not imported in your mail client, you need to include the email header <code>In-Reply-To: &lt;message-id&gt;<\/code> or the thread will be split, regardless of the subject line you use.<\/p><h5 id=\"external-tools\">External tools<\/h5><p>Email and patch files are both old interoperable standards, so, theoretically, running patches through continuous integration ought not to be a problem. However I don\u2019t know that the tooling here is well-developed. Where Gitlab has a CI server built in, and several addons exist for Github, I don\u2019t know of any clearly dominant thing for a patch based workflow. IBM built a tool called <a href=\"https:\/\/github.com\/ruscur\/snowpatch\">snowpatch<\/a> to do this, but I don\u2019t know what tomfoolery is involved.<\/p><p>It is possible to read patches from a list and then script an integration to a CI tool via patching and pushing to Github. Thomas Munro recently created <a href=\"http:\/\/cfbot.cputube.org\/\">cfbot<\/a> to do this for Postgres. (See the interesting <a href=\"https:\/\/www.pgcon.org\/2018\/schedule\/attachments\/499_pgcon%202018%20-%20continuous%20integration.pdf\">slides<\/a> about its development.) It\u2019s a creative use for Github that doesn\u2019t compromise the mailing list flow.<\/p><h4 id=\"barrier-to-entry\">Barrier to entry<\/h4><p>Related to the unfamiliarity mentioned earlier, using a mailing list well requires new skills for contributors, and infrastructure for maintainers.<\/p><h5 id=\"tooling-and-the-protocol\">Tooling and the protocol<\/h5><p>Although you can subscribe to a list with Outlook or GMail and call it a day, neither you nor the people you interact with will be happy about it. Email is a precise tool from ye olde Unix days, where people care about things like the plain text lines you send over the wire, or the way you use MIME types.<\/p><p>But even before crafting these little RFC-821 gems you have to go through a dance every time you want to subscribe to a list, confirming your address and adding a folder and filter to the mail client. It\u2019s not like jotting a note in the issues of another Github repo, unless the mailing list is \u201copen\u201d meaning accepts mail from non-registered addresses. Most lists are \u201cclosed\u201d because of spam. Such is the price we pay for basing identity on the DNS system and not on a corporate collaborative coding site profile.<\/p><p>Once on the list and ready to send some messages or patches, there are a few guidelines to follow that pretty much all lists ask. They help accommodate the diversity of mail clients:<\/p><ul><li>Send with plain text<\/li>\n<li>Disable all HTML \u201cenhancements\u201d<\/li>\n<li>At most 72 characters per line<\/li>\n<li>Do not top-post<\/li>\n<li>Keep quoted text small and relevant<\/li>\n<li>Strip out company legal footers<\/li>\n<li>Shorten lengthy signatures<\/li>\n<li>Use Reply-All<\/li>\n<li>If replying, ensure <code>In-Reply-To<\/code> is set<\/li>\n<\/ul><p>If your message contains a patch then there are more rules to observe, and they vary by community. Linux, OpenBSD and FreeBSD say, \u201cNo MIME, no links, no compression, no attachments. Just plain text.\u201d The patch goes right in the body of the message at the bottom. If the mail client would mangle the code (such as by wrapping long lines), then attaching with MIME is permissible. For patches above 300kb, host them online and include a link.<\/p><p>Postgres, on the other hand, wants patches attached as type <code>text\/x-patch<\/code>, with <code>disposition: inline<\/code> for tiny patches and <code>disposition: attachment<\/code> for substantial ones. And in both cases including a unique filename for the patch in the Content-Disposition. Also different reviewers prefer different patch formats, like Tom Lane will only accept context diffs, but other people want unified diffs.<\/p><h5 id=\"hosting-requirements\">Hosting requirements<\/h5><p>On the maintainer side, it\u2019s not as easy to make a new project as it is clicking a button on Github. Maintainers need to host the mailing list and <em>optionally<\/em> a patch status app and a bug tracker.<\/p><p>I\u2019d advise creating a general purpose listserv for your personal use, which can run lists for all your small projects. Then create project pages off your homepage or something. I just can\u2019t see taking the overhead to spin up new infrastructure for a dinky experimental open source utility.<\/p><p>There are probably commercially hosted listservs for rent, or you can host your own. Probably the best ones for self-hosting are <a href=\"https:\/\/www.gnu.org\/software\/mailman\/\">GNU Mailman<\/a> or <a href=\"http:\/\/mlmmj.org\/\">Mlmmj<\/a>. There\u2019s also a service called <a href=\"https:\/\/www.mail-archive.com\/\">The Mail Archive<\/a> which can add an archive to any list. Might also be worth looking into <a href=\"https:\/\/public-inbox.org\/README.html\">public-inbox<\/a> which is a new approach that shares a mailbox itself over Git. Git to hold the mailing list to talk about git \u2013 inception!<\/p><p>For your listserv you\u2019ll also want to install Spamassassin and restrict allowed MIME types to plain text and patches. Also remember to configure the list to use <code>Reply-To<\/code> rather than <code>From<\/code> for DMARC compatibility. This whole ordeal isn\u2019t something you want to have to do twice!<\/p><p>Optionally you may want to host a lightweight web interface for browsing git. The simplest seem to be one of these:<\/p><ul><li><a href=\"https:\/\/git.zx2c4.com\/cgit\/about\/\">cgit<\/a><\/li>\n<li><a href=\"https:\/\/git.2f30.org\/stagit\/\">stagit<\/a><\/li>\n<li><a href=\"http:\/\/klausdemo.lophus.org\/klaus\/\">klaus<\/a><\/li>\n<li><a href=\"https:\/\/git-scm.com\/book\/en\/v2\/Git-on-the-Server-GitWeb\">GitWeb<\/a><\/li>\n<\/ul><p>Alternately you could just use Github for the actual git server and code browsing. Or mirror there.<\/p><p>Finally, you could go really hard-core and let people use git\u2019s builtin functionality for pulling a shallow clone rather than providing any web interface:<\/p><div class=\"sourceCode\"><pre class=\"sourceCode bash\"># get the latest commit\ngit clone --depth=1 --single-branch --branch master &lt;remote&gt;\n# or get it and don't even include the .git folder\ngit archive --remote=&lt;remote&gt; master<\/pre><\/div><p>This is rather unfriendly though.<\/p><h4 id=\"topic-specialization\">Topic specialization<\/h4><p>Sometimes it\u2019s useful to restrict the general purpose nature of email messages. For instance bug reports should follow a certain format and include details about the environment. Github offers <a href=\"https:\/\/blog.github.com\/2016-02-17-issue-and-pull-request-templates\/\">issue templates<\/a>; what\u2019s the equivalent for email?<\/p><p>Some projects address this with helper programs that collect information and format an email on your behalf. The information is thus specialized for the type of software involved. GNU has <a href=\"https:\/\/www.gnu.org\/software\/gnats\/\">GNATS<\/a>, OpenBSD has <a href=\"https:\/\/man.openbsd.org\/sendbug\">sendbug<\/a>, Debian has <a href=\"https:\/\/www.debian.org\/Bugs\/\">debbugs<\/a> and Postgres has a bug <a href=\"https:\/\/www.postgresql.org\/account\/submitbug\/\">report form<\/a>.<\/p><p>Noob questions are another kind of specialized topic, as are topics that interest only a small subset of the list. Lists full of kernel developers or database hackers tend to be\u2026 stern. Communities which still use mailing lists are usually the ones full of scary wizards who invent the world that the rest of us play in. Thus it can be intimidating to ask basic questions there. What projects usually do is create a beginner list separate from the main developer zone. Also IRC is a good way to ask questions that fade away.<\/p><h4 id=\"why-not-nntp\">Why not NNTP?<\/h4><p>Isn\u2019t the whole email list concept a bit of a hack for turning a protocol designed for small-group communication into a behemoth? After all, NNTP is made for threaded large group discussion.<\/p><p>NNTP has slightly worse offline characteristics though. While email allows the Reply-All trick to keep things limping along during a list outage, replies through NNTP have to route through a central server. Also identity verification needs to happen on the server, whereas email can delegate to SPF\/DKIM\/DMARC. There are also a greater number of clients for email than NNTP, and clients that continue to be supported.<\/p><p>Finally, if you like reading a mailing list with a newsreader, go right ahead. Use a bridge like <a href=\"http:\/\/www.gmane.org\/\">GMANE<\/a>.<\/p><h4>Conclusion<\/h4><p>I hope that I\u2019ve given mailing lists a fair comparison with today\u2019s more common web interfaces. It seems that mailing lists offer real advantages, and that some projects could benefit from making the switch. The biggest difficulty with mailing lists seems to be simply setting them up. Perhaps if there was a deployable server image including necessary software and configuration, that would help boost adoption.<\/p><h3 id=\"related-posts\">Related Posts<\/h3><ul><li><a href=\"https:\/\/begriffs.com\/posts\/2016-07-08-returning-original-social-network.html\">Returning to the Original Social Network<\/a><\/li>\n<li><a href=\"https:\/\/begriffs.com\/posts\/2018-09-18-dmarc-mailing-list.html\">Actually, DMARC works fine with mailing lists<\/a><\/li>\n<\/ul>","created_at":"2018-11-07T16:23:26+0000","updated_at":"2018-11-07T16:23:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":23,"domain_name":"begriffs.com","preview_picture":"https:\/\/begriffs.com\/images\/gh-vs-mail.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":142,"title":"Ask HN: Solo app devs, how do you do user testing?","url":"https:\/\/news.ycombinator.com\/item?id=17336984","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D17336984\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"17336984\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=17336984\" class=\"storylink\">Ask HN: Solo app devs, how do you do user testing?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        262 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=rhardih\" class=\"hnuser\">rhardih<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17336984\">4 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=17336984&amp;goto=item%3Fid%3D17336984\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Solo%20app%20devs%2C%20how%20do%20you%20do%20user%20testing%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Solo%20app%20devs%2C%20how%20do%20you%20do%20user%20testing%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=17336984&amp;auth=25539ea1c416497e10ea858fffd01e22706f84b5\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=17336984\">81\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Enlisting the help of friends and family can only get you so far, so how do you gather quality feedback pre\/post-launch, without shelling out for a professional agency to do the testing?<\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:23:25+0000","updated_at":"2018-11-07T16:23:25+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":143,"title":"Guide to Angel Investing","url":"https:\/\/anishgodha.com\/2018\/06\/15\/guide-to-angel-investing\/","content":"<p>I\u2019ve been dabbling in startup investing for a few years now. Nothing extraordinary, and I\u2019ve definitely not made it a career, but I have developed a guide of sorts that I follow when making the final decision of whether or not to invest.<\/p>\n<p>This is written from the viewpoint of a single angel investor. Here it is:<\/p>\n<ol><li><strong><u>Risk<\/u><\/strong>: Understand the high risk in angel investing. Mentally, assume you\u2019ll write off all investments as soon as you make them. Invest only what you can afford to lose.<\/li>\n<\/ol><ol start=\"2\"><li><strong><u>Avoid life-support investing<\/u><\/strong>: If a company comes for funding, where the founder says\/implies something to the extent of: \u201cfund us or we\u2019re going out of business\u201d \u2013 usually gravitate to not investing. Life support investing is a bad idea (not to mention stressful!).<\/li>\n<\/ol><p><em>Lean towards companies that will survive regardless of whether or not you invest in them.<\/em><\/p>\n<ol start=\"3\"><li><strong><u>Three companies come in for funding<\/u><\/strong>:<\/li>\n<\/ol><p>A \u2013 Entrepreneur has the best idea ever! They need funds to build the product.<\/p>\n<p>B \u2013 Entrepreneur has the best product ever! They need funds to market.<\/p>\n<p>C \u2013 Entrepreneur has a product which shows great traction. They need funds to\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 scale.<\/p>\n<p>Everything else equal, investing in C has the most promise.<\/p>\n<ol start=\"4\"><li><strong><u>Follow-up investor:<\/u><\/strong> For the new investor, prefer being a follow up investor rather than leading rounds.<\/li>\n<\/ol><ol start=\"5\"><li><strong><u>Investment hierarchy<\/u><\/strong><strong>:<\/strong> First invest in companies with great traction. If that does not exist in the options, then invest in those with great people, followed by those with a great product.<\/li>\n<\/ol><ol start=\"6\"><li><strong><u>Bet on people rather than product<\/u><\/strong>: Early stages, funding is a bet on people rather than product. Products can change more easily than people.<\/li>\n<\/ol><ol start=\"7\"><li><strong><u>Never invest in something you don\u2019t understand<\/u><\/strong>: Doesn\u2019t matter what it is, who is involved, who has invested or how well the market is going.<\/li>\n<\/ol><ol start=\"8\"><li><strong><u>Entrepreneurial grit matters<\/u><\/strong>: Invest in entrepreneurs who show they\u2019ll simply never give up on the company \u2013 no matter what. Perseverance is crucial. When things get hard (and in startups they will!), you want someone who has the grit to stick through it to figure a way out.<\/li>\n<\/ol><ol start=\"9\"><li><strong><u>Scaling<\/u><\/strong>: Always think about how the company can scale. If scaling is expensive or hard, lean towards not investing.<\/li>\n<\/ol><ol start=\"10\"><li><strong><u>Avoid niches<\/u><\/strong>: Niche markets are starting points \u2013 but that\u2019s it. The ultimate market cannot remain in a niche space. Successful companies need to scale.<\/li>\n<\/ol><ol start=\"11\"><li><strong><u>Viable business or science project?<\/u><\/strong>: Is the company capable of being a business or will it remain a technology\/project? Not all technologies can become businesses.<\/li>\n<\/ol><ol start=\"12\"><li><strong><u>The Fence<\/u><\/strong>: If you\u2019re on the fence about a company for a long time \u2013 defer the investment.<\/li>\n<\/ol><ol start=\"13\"><li><strong><u>Entrepreneur\u2019s knowledge<\/u><\/strong>: The entrepreneur needs to know more about the space than you. If you seem to know more about his own space \u2013 that\u2019s a red flag.<\/li>\n<\/ol><ol start=\"14\"><li><strong><u>First versions will suck:<\/u><\/strong> Products require several iterations of trials and failure to become great. The key is if the entrepreneur\/team is unfazed by failure to keep at it despite the obstacles.<\/li>\n<\/ol><ol start=\"15\"><li><strong><u>What if Google\/FB does it?:<\/u><\/strong> Assume the incumbent is always going to copy the entrepreneur if he does well. What is his strategy?<\/li>\n<\/ol><ol start=\"16\"><li><strong><u>Does the product create delight?:<\/u><\/strong> Do you feel delighted when you use the entrepreneur\u2019s product or service? If not, at least do the customers absolutely love the product? Defer the investment until that is true.<\/li>\n<\/ol><ol start=\"17\"><li><strong><u>Simple enough for your mother: <\/u><\/strong>Can your mother use the entrepreneur\u2019s product or service? Simplicity is key for widespread adoption.<\/li>\n<\/ol><ol start=\"18\"><li><strong><u>Failure is a learning process<\/u><\/strong>: Past failures &gt; No failures.<\/li>\n<\/ol><ol start=\"19\"><li><strong><u>Return time period:<\/u><\/strong> Expect returns no less than 5-7 years.<\/li>\n<\/ol><ol start=\"20\"><li><strong><u>Differentiate between the decision and outcome:<\/u><\/strong> Decisions must be taken using all information available to you\u00a0<em>at that time.<\/em> You can\u2019t predict the future and hindsight is always 20\/20. A bad outcome does not necessarily mean a bad decision and vice versa. Don\u2019t beat yourself up too much over a bad outcome, and then again don\u2019t let a good outcome go to your head.<\/li>\n<\/ol><p>Good luck! And as always, if you do discover the next Google or Facebook, do let me know! \ud83d\ude09<\/p>\n<p><em>With that said, here is the obligatory legal disclaimer that goes with any investment advice \u2013 Please invest at your own risk!\u00a0<\/em><\/p>","created_at":"2018-11-07T16:23:25+0000","updated_at":"2018-11-07T16:23:25+0000","published_at":"2018-06-15T23:36:01+0000","published_by":["Anish Godha"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"anishgodha.com","preview_picture":"https:\/\/s0.wp.com\/i\/blank.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":144,"title":"a newcomer's (angry) guide to R","url":"http:\/\/arrgh.tim-smith.us\/","content":"<p><a href=\"http:\/\/tim-smith.us\">Tim Smith<\/a> &lt;arrgh@tim-smith.us&gt;, <a href=\"http:\/\/twitter.com\/biotimylated\">@biotimylated<\/a><br \/>with Kevin Ushey &lt;kevinushey@gmail.com&gt;, <a href=\"http:\/\/twitter.com\/kevin_ushey\">@kevin_ushey<\/a><\/p><p>Every once in a while, this document catches the wind.<\/p><p>Since I first wrote it, I have become less convinced of the fairness, importance,\nand sophistication of many of the criticisms I\u2019ve leveled.<\/p><p>I\u2019ve become quite productive in R.\nI\u2019ve also spent some time as a maintainer of a popular open-source tool\nand learned how it feels to be the target of abuse from Internet Randos.\n(I survived, but it wasn\u2019t motivating.)<\/p><p>I\u2019ve long intended to reconsider my arguments and tone\nbut engaging with my own embarrassing writing\nwas never the most interesting way to spend a weekend.<\/p><p>I\u2019m grateful to the R community for their labors.\nI use R and benefit from R and I apologize for the lack of empathy I show here.\nI\u2019m leaving the rest of the document unchanged for the moment in hopes that the technical content\nis useful to new users but I expect to revise it soon and remove many of the \u201cgood parts.\u201d<\/p><p>R is fine. (Do use the tidyverse.)<\/p><p>R is a shockingly dreadful language for an exceptionally useful data analysis environment. The more you learn about the R language, the worse it will feel. The development environment suffers from literally decades of accretion of stupid hacks from a community containing, to a first-order approximation, zero software engineers.<sup id=\"fnref:engineers\"><a href=\"#fn:engineers\" class=\"footnote\">1<\/a><\/sup> R makes me want to kick things almost every time I use it.<\/p><p>But there are a lot of great tools that are built in R. <a href=\"http:\/\/docs.ggplot2.org\/current\/\">ggplot2<\/a> is first-in-class and <a href=\"https:\/\/bioconductor.org\/\">Bioconductor<\/a> packages are often essential. Sometimes there\u2019s aught to do but grin and bear (though never without a side of piss and moan).<\/p><p>The documentation is inanely bad. I can\u2019t explain it. aRrgh is my attempt to explain the language to myself. aRrgh exists as a living document and will continue to grow \u2013 it is not complete, but it got to a point where it seemed like it was probably useful so I decided to toss it on the web. It should be correct and it\u2019s a bug if it isn\u2019t. Please email me or <a href=\"https:\/\/github.com\/tdsmith\/aRrgh\/issues\">file issues on Github<\/a>.<\/p><p>The goal of the document is to describe R\u2019s data types and structures while offering enough help with the syntax to get a programmer coming from another, saner language into a more comfortable place.<\/p><ul><li><a href=\"http:\/\/arrgh.tim-smith.us\/syntax.html\">Basic syntax<\/a>: crash course and gotchas; finding help<\/li>\n  <li><a href=\"http:\/\/arrgh.tim-smith.us\/atomic.html\">Atomic vectors<\/a>: R\u2019s simplest data types; logic values; vectorizing; arrays<\/li>\n  <li><a href=\"http:\/\/arrgh.tim-smith.us\/factors.html\">Factors<\/a>: A useful and misunderstood data type. Where they come from, how to handle them<\/li>\n  <li><a href=\"http:\/\/arrgh.tim-smith.us\/data_frames.html\">Data frames<\/a>: R\u2019s structure for tabular data. How to create them, access semantics<\/li>\n<\/ul><p>To come?:<\/p><ul><li>Indexing<\/li>\n  <li>Lists<\/li>\n  <li>Namespaces<\/li>\n  <li>Beyond Base R, Or: How I Learned To Stop Worrying and Love the Hadleyverse<\/li>\n<\/ul><p>\u00a9 Tim Smith 2012-6. This work is made available under a <a href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/3.0\/deed.en_US\">Creative Commons Attribution-ShareAlike 3.0 Unported<\/a> License.<\/p><p>If you enjoyed this, you will probably enjoy <a href=\"http:\/\/me.veekun.com\/blog\/2012\/04\/09\/php-a-fractal-of-bad-design\/\">PHP: A Fractal of Bad Design<\/a>, which is even more cathartic.<\/p><div class=\"footnotes\"><ol><li id=\"fn:engineers\">\n      <p>Many brilliant and talented people have contributed to R over the years, many of whom are renowned computer scientists for whom I have tremendous respect! But a computer scientist is not the same as a software engineer.\u00a0<a href=\"#fnref:engineers\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>","created_at":"2018-11-07T16:23:25+0000","updated_at":"2018-11-07T16:23:25+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"arrgh.tim-smith.us","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":145,"title":"Math and front-end","url":"https:\/\/www.chenhuijing.com\/blog\/math-and-front-end\/","content":"<p>If you were one of those kids that wondered why you had to attend all those math classes and learn about angles and algebra, well, so was I. I\u2019ll admit, I was too restless and easily bored to be a good student in a classroom setting. But as I got older, I realised that certain mathematical principles actually had practical applications in my day-to-day work (Not calculus though, I mean, I\u2019m not a rocket scientist).<\/p><p>The question of whether you need to be good at math to be a good web developer boils down to what your definition of being good at math entails. If your definition of being good at math refers to understanding pure mathematical concepts like analysis or calculus, which involves a lot of proofs and theorems, then I have to admit, you don\u2019t use those skills in web design\/development.<\/p><p>But although the practical skills required are probably at the high school level, it\u2019s a matter of applying those arithmetic skills and knowledge of geometry in the context of web design.<\/p><p>For example, using the border-width hack to create triangles involves both a knowledge of how the browser renders borders and the application of trigonometry to realise that the property is \u201chackable\u201d.<\/p><p>So let\u2019s take a look at some of the aspects of web development where math can come in handy.<\/p><h2 id=\"arithmetic\">Arithmetic<\/h2><blockquote>\n<p>Arithmetic is a branch of mathematics that deals with properties of the counting (and also whole) numbers and fractions and the basic operations applied to these numbers.<br \/>\u2014<a href=\"https:\/\/www.cut-the-knot.org\/WhatIs\/WhatIsArithmetic.shtml\">Alexander Bogomolny<\/a><\/p>\n<\/blockquote><p>This seems unremarkably basic, but if you\u2019re doing any type of responsive design, and let\u2019s be honest, responsive design IS the norm nowadays, arithmetic is totally relevant to you. What does arithmetic have to do with design, you may ask? Plenty. But before that, let\u2019s talk about CSS units and values, because that\u2019s what we\u2019re going to be counting.<\/p><p>Numbers feature quite heavily in CSS, as property values, mostly. And the specification that covers this is the <a href=\"https:\/\/www.w3.org\/TR\/css-values-3\/\">CSS Values and Units Module Level 3<\/a>. The key function behind the arithmetic of CSS is the <code>calc()<\/code> function, which supports the four basic operations of addition, subtraction, multiplication and division.<\/p><p>Back in the day when web design was mostly fixed- width, designers and developers would produce pixel-perfect designs that only worked when viewed at specific viewport widths. But as the number of different screen sizes that users would use to browse the web grew exponentially, fixed- width designs didn\u2019t really fit the bill anymore.<\/p><p>It made a lot more sense to let the browser figure out the sizing of elements on the page depending on the viewport size instead, using percentages, font-relative units like <code>em<\/code>s or <code>ch<\/code>s, and more recently, viewport units. The <code>calc<\/code> function works with combinations of CSS values of different units, handling the tricky computation for us so we can focus on designing and building the layouts and components we want.<\/p><p>A common requirement in many designs is to make sure the footer is \u201cstuck\u201d to the bottom of the viewport even if there isn\u2019t enough content to fill the height of the viewport.<\/p><p>Now, there are several different ways to achieve such an effect, but using <code>calc<\/code> is elegant enough to ensure that the footer never ends up floating when there isn\u2019t enough content, yet still remain within the document flow.<\/p><p>Here\u2019s some very basic markup, consisting of a <code>header<\/code>, <code>main<\/code> and <code>footer<\/code> elements.<\/p><pre>&lt;header&gt;Header&lt;\/header&gt;\n&lt;main&gt;Main&lt;\/main&gt;\n&lt;footer&gt;Footer&lt;\/footer&gt;<\/pre><p>To make sure the footer stays at the bottom of the page regardless of the amount of content within the <code>main<\/code> element, the <code>main<\/code> element needs to have a minimum height of 100% of the viewport height less the height of the header and footer, or:<\/p><pre>100% viewport height \u2013 ( height of header + height of footer)<\/pre><p>Translating that into CSS (this assumes browser styles have been reset):<\/p><pre>main {\n  min-height: calc(100vh - 2em);\n}<\/pre><p>Without additional styling, the height of the header and footer should both be <code>1em<\/code> due to the text within them. Using <code>min-height<\/code> instead of <code>height<\/code> ensures that if there is more content than the height of the viewport, it will flow as per normal.<\/p><figure><figcaption>The <code>footer<\/code> is always at the bottom.<\/figcaption><\/figure><h3 id=\"fluid-typography\">Fluid typography<\/h3><p>While we\u2019re on the topic of responsive design, let\u2019s also use some math to figure out our font sizes for different viewport widths. Font sizes can take more than just px or em units, we can use viewport units to define font sizes as well. A problem with this approach is that if the viewport gets too small, your font could potentially shrink to an illegible size.<\/p><p>To combat this, we can use <code>calc<\/code> to provide a minimum font-size like the example below:<\/p><pre>body { font-size: calc(1em + 1vw); }<\/pre><p>If you do require more control over your font sizes, then more variables need to be added to the equation. The concept of precise fluid font sizes was pioneered by Mike Riethmuller in his article, <a href=\"https:\/\/www.madebymike.com.au\/writing\/precise-control-responsive-typography\/\">Precise control over responsive typography<\/a>, and was an extension of the idea of <a href=\"http:\/\/nicewebtype.com\/notes\/2012\/02\/03\/molten-leading-or-fluid-line-height\/\">molten leading<\/a> by <a href=\"http:\/\/tbrown.org\/\">Tim Brown<\/a>. Mike Riethmuller\u2019s equation looks something like this:<\/p><p><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/fluid-typography.svg\" alt=\"Fluid typography annotation\" \/><\/p><p><a href=\"https:\/\/fvsch.com\/\">Florens Verschelde<\/a> then did a deep dive into the mathematics behind CSS locks in his article <a href=\"https:\/\/fvsch.com\/code\/css-locks\/\">The math of CSS locks<\/a>, by expressing the font-size \/ line-height calculation as a linear function. Linear functions can be plotted on graphs, which makes it easier to visualise relationship between font-size \/ line-height with the viewport size.<\/p><figure><figcaption>Graph from <a href=\"https:\/\/fvsch.com\/code\/css-locks\/\">The math of CSS locks<\/a> by <a href=\"https:\/\/fvsch.com\/\">Florens Verschelde<\/a><\/figcaption><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/linear-function.png\" srcset=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/linear-function@2x.png 2x\" alt=\"Font-size against viewport size in pixels\" \/><\/figure><h2 id=\"geometry\">Geometry<\/h2><blockquote>\n<p>Geometry is a branch of mathematics that is concerned with the properties of configurations of geometric objects - points, (straight) lines, and circles being the most basic of these.<br \/>\u2014<a href=\"https:\/\/www.cut-the-knot.org\/WhatIs\/WhatIsGeometry.shtml\">Alexander Bogomolny<\/a><\/p>\n<\/blockquote><h3 id=\"drawing-shapes-with-css\">Drawing shapes with CSS<\/h3><p>Geometry can help with understanding how to create shapes with just CSS. Let\u2019s take the simple <code>border-radius<\/code> property, for example, which is used to round the corners of an element\u2019s outside borders. Most of us just put in a single value and call it a day, but the <code>border-radius<\/code> property is a little more complicated than that.<\/p><p><code>border-radius<\/code> is actually a shorthand for all 4 <code>border-*-radius<\/code> properties, where <code>*<\/code> refers to <code>top-left<\/code>, <code>top-right<\/code>, <code>bottom-left<\/code> or <code>bottom-right<\/code>. And it can take up to 2 values, separated with a <code>\/<\/code>, where the first value is the horizontal radius, while the second value is the vertical radius. Here\u2019s a diagram to visualisation purposes:<\/p><p><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/border-radius.svg\" alt=\"Border radius values\" \/><\/p><p>And when we use percentages as values, the horizontal radius will be a percentage of the <strong>width<\/strong> of the border box while the vertical radius will be a percentage of the <strong>height<\/strong> of the border box. So that\u2019s why setting a <code>border-radius: 50%<\/code> gives us a perfect circle or ellipse.<\/p><p>Moving onto something more interesting, but also requires borders, we have triangles. Pure CSS triangles are made possible by \u201chacking\u201d the borders of an element. When we create borders around an element, the edges of these borders meet diagonally, and we can see this if we apply a sufficiently thick border width to our element. They are trapeziums.<\/p><p><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/border-box.svg\" alt=\"Thick border widths\" \/><\/p><p>If we set the width and height of the element to 0, the trapeziums then become triangles, and voila, we\u2019ve got our pure CSS triangles.<\/p><p><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/borders-only.svg\" alt=\"CSS triangles from borders\" \/><\/p><p>So let\u2019s say we don\u2019t want the triangles in a set of four, which is probably the usual case, the other three borders should be made invisible, by setting the adjacent borders\u2019 colour to <code>transparent<\/code> and omitting the opposite border altogether.<\/p><pre>.triangle-up {\n  width: 0; \n  height: 0; \n  border-left: 30px solid transparent;\n  border-right: 30px solid transparent;\n  border-bottom: 30px solid gray;\n}<\/pre><p><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/border-colour.svg\" alt=\"Single pure CSS triangle\" \/><\/p><p>And maybe we don\u2019t always want isosceles triangles (that\u2019s what you get by setting all the border widths to the same value), so some geometry comes into play. The handy <a href=\"https:\/\/en.wikipedia.org\/wiki\/Pythagorean_theorem\">Pythagorean theorem<\/a> can be used to calculate what the height of the triangle should be like so:<\/p><p><img src=\"https:\/\/www.chenhuijing.com\/assets\/images\/posts\/math-for-fed\/equilateral.svg\" alt=\"Equilateral CSS triangle\" \/><\/p><pre>.triangle-up {\n  width: 0; \n  height: 0; \n  border-left: 50px solid transparent;\n  border-right: 50px solid transparent;\n  border-bottom: calc(100px * 0.866) solid gray;\n}<\/pre><h2 id=\"wrapping-up\">Wrapping up<\/h2><p>Mathematics may seem like something that is far from the creative visual aspect of web design and development but it does have a number of practical applications, so why not brush off that high school math textbook of yours and see if there is anything in there that can inspire you to explore CSS in ways you never thought of before?<\/p>","created_at":"2018-11-07T16:23:25+0000","updated_at":"2018-11-07T16:23:25+0000","published_at":"2018-05-10T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"www.chenhuijing.com","preview_picture":"https:\/\/www.chenhuijing.com\/assets\/images\/avatar.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":146,"title":"JVM Internals","url":"http:\/\/blog.jamesdbloom.com\/JVMInternals.html","content":"<p>This article explains the internal architecture of the Java Virtual Machine (JVM).  The following diagram show the key internal components of a typical JVM that conforms to <a href=\"http:\/\/www.amazon.co.uk\/Virtual-Machine-Specification-Edition-Series\/dp\/0133260445\">The Java Virtual Machine Specification Java SE 7 Edition<\/a>.<\/p><a class=\"image_link\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/JVM_Internal_Architecture.png\"><img itemprop=\"image\" id=\"JVM_Internal_Architecture_image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/JVM_Internal_Architecture_small.png\" alt=\"internal architecture of Java Virtual Machine (JVM)\" \/><\/a><p>The components shown on this diagram are each explained below in two sections.  <a href=\"#threads\">First section<\/a> covers the components that are created for each thread and the <a href=\"#shared_between_threads\">second section<\/a> covers the components that are created independently of threads.<\/p><div><ul><li><a href=\"#threads\">Threads<\/a>\n                    <div><ul><li><a href=\"#jvm_system_threads\">JVM System Threads<\/a><\/li>\n                            <li><a href=\"#per_thread\">Per Thread<\/a><\/li>\n                            <li><a href=\"#program_counter\">program Counter (PC)<\/a><\/li>\n                            <li><a href=\"#stack\">Stack<\/a><\/li>\n                            <li><a href=\"#native_stack\">Native Stack<\/a><\/li>\n                            <li><a href=\"#stack_restrictions\">Stack Restrictions<\/a><\/li>\n                            <li><a href=\"#frame\">Frame<\/a><\/li>\n                            <li><a href=\"#local_variables_array\">Local Variables Array<\/a><\/li>\n                            <li><a href=\"#operand_stack\">Operand Stack<\/a><\/li>\n                            <li><a href=\"#dynamic_linking\">Dynamic Linking<\/a><\/li>\n                        <\/ul><\/div>\n                <\/li>\n                <li><a href=\"#shared_between_threads\">Shared Between Threads<\/a>\n                    <div><ul><li><a href=\"#heap\">Heap<\/a><\/li>\n                            <li><a href=\"#memory_management\">Memory Management<\/a><\/li>\n                            <li><a href=\"#non_heap_memory\">Non-Heap Memory<\/a><\/li>\n                            <li><a href=\"#jit_compilation\">Just In Time (JIT) Compilation<\/a><\/li>\n                            <li><a href=\"#method_area\">Method Area<\/a><\/li>\n                            <li><a href=\"#class_file_structure\">Class File Structure<\/a><\/li>\n                            \n                            <li><a href=\"#classloader\">Classloader<\/a><\/li>\n                            <li><a href=\"#faster_class_loading\">Faster Class Loading<\/a><\/li>\n                            <li><a href=\"#where_is_the_method_area\">Where Is The Method Area<\/a><\/li>\n                            <li><a href=\"#classloader_reference\">Classloader Reference<\/a><\/li>\n                            <li><a href=\"#constant_pool\">Run Time Constant Pool<\/a><\/li>\n                            <li><a href=\"#exception_table\">Exception Table<\/a><\/li>\n                            <li><a href=\"#symbol_table\">Symbol Table<\/a><\/li>\n                            <li><a href=\"#string_table\">Interned Strings (String Table)<\/a><\/li>\n                        <\/ul><\/div>\n                <\/li>\n            <\/ul><\/div><h2 id=\"threads\">Thread<\/h2><p>A thread is a thread of execution in a program. The JVM allows an application to have multiple threads of execution running concurrently.  In the Hotspot JVM there is a direct mapping between a Java Thread and a native operating system Thread.  After preparing all of the state for a Java thread such as thread-local storage, allocation buffers, synchronization objects, stacks and the program counter, the native thread is created.  The native thread is reclaimed once the Java thread terminates.  The operating system is therefore responsible for scheduling all threads and dispatching them to any available CPU.  Once the native thread has initialized it invokes the run() method in the Java thread.  When the run() method returns, uncaught exceptions are handled, then the native thread confirms if the JVM needs to be terminated as a result of the thread terminating (i.e. is it the last non-deamon thread).  When the thread terminates all resources for both the native and Java thread are released.<\/p><h3 id=\"jvm_system_threads\">JVM System Threads<\/h3><p>If you use jconsole or any debugger it is possible to see there are numerous threads running in the background.  These background threads run in addition to the main thread, which is created as part of invoking public static void main(String[]), and any threads created by the main thread.  The main background system threads in the Hotspot JVM are:<\/p><div class=\"columns\"><div class=\"row\"><p>VM thread<\/p><p>This thread waits for operations to appear that require the JVM to reach a safe-point.  The reason these operations have to happen on a separate thread is because they all require the JVM to be at a safe point where modifications to the heap can not occur.  The type of operations performed by this thread are \"stop-the-world\" garbage collections, thread stack dumps, thread suspension and biased locking revocation.<\/p><\/div><div class=\"row\"><p>Periodic task thread<\/p><p>This thread is responsible for timer events (i.e. interrupts) that are used to schedule execution of periodic operations<\/p><\/div><div class=\"row\"><p>GC threads<\/p><p>These threads support the different types of garbage collection activities that occur in the JVM<\/p><\/div><div class=\"row\"><p>Compiler threads<\/p><p>These threads compile byte code to native code at runtime<\/p><\/div><div class=\"row\"><p>Signal dispatcher thread<\/p><p>This thread receives signals sent to the JVM process and handle them inside the JVM by calling the appropriate JVM methods.<\/p><\/div><\/div><h2 id=\"per_thread\">Per Thread<\/h2><p>Each thread of execution has the following components:<\/p><h3 id=\"program_counter\">Program Counter (PC)<\/h3><p>Address of the current instruction (or opcode) unless it is native.  If the current method is native then the PC is undefined.  All CPUs have a PC, typically the PC is incremented after each instruction and therefore holds the address of the next instruction to be executed.  The JVM uses the PC to keep track of where it is executing instructions, the PC will in fact be pointing at a memory address in the Method Area.<\/p><h3 id=\"stack\">Stack<\/h3><p>Each thread has its own stack that holds a frame for each method executing on that thread.  The stack is a Last In First Out (LIFO) data structure, so the currently executing method is at the top of the stack. A new frame is created and added (pushed) to the top of stack for every method invocation.  The frame is removed (popped) when the method returns normally or if an uncaught exception is thrown during the method invocation. The stack is not directly manipulated, except to push and pop frame objects, and therefore the frame objects may be allocated in the Heap and the memory does not need to be contiguous.<\/p><h3 id=\"native_stack\">Native Stack<\/h3><p>Not all JVMs support native methods, however, those that do typically create a per thread native method stack.  If a JVM has been implemented using a C-linkage model for Java Native Invocation (JNI) then the native stack will be a C stack.  In this case the order of arguments and return value will be identical in the native stack to typical C program. A native method can typically (depending on the JVM implementation) call back into the JVM and invoke a Java method.  Such a native to Java invocation will occur on the stack (normal Java stack); the thread will leave the native stack and create a new frame on the stack (normal Java stack).<\/p><h3 id=\"stack_restrictions\">Stack Restrictions<\/h3><p>A stack can be a dynamic or fixed size.  If a thread requires a larger stack than allowed a StackOverflowError is thrown.  If a thread requires a new frame and there isn\u2019t enough memory to allocate it then an OutOfMemoryError is thrown.<\/p><h3 id=\"frame\">Frame<\/h3><p>A new frame is created and added (pushed) to the top of stack for every method invocation.  The frame is removed (popped) when the method returns normally or if an uncaught exception is thrown during the method invocation.  For more detail on exception handling <a href=\"#exception_table\">see the section on Exception Tables below<\/a>.<\/p><p>Each frame contains:<\/p><div><ul class=\"list\"><li>Local variable array<\/li>\n            <li>Return value<\/li>\n            <li>Operand stack<\/li>\n            <li>Reference to runtime constant pool for class of the current method<\/li>\n        <\/ul><\/div><h3 id=\"local_variables_array\">Local Variables Array<\/h3><p>The array of local variables contains all the variables used during the execution of the method, including a reference to this, all method parameters and other locally defined variables.  For class methods (i.e. static methods) the method parameters start from zero, however, for instance method the zero slot is reserved for this.<\/p><p>A local variable can be:<\/p><div><ul class=\"list\"><li>boolean<\/li>\n            <li>byte<\/li>\n            <li>char<\/li>\n            <li>long<\/li>\n            <li>short<\/li>\n            <li>int<\/li>\n            <li>float<\/li>\n            <li>double<\/li>\n            <li>reference<\/li>\n            <li>returnAddress<\/li>\n        <\/ul><\/div><p>All types take a single slot in the local variable array except long and double which both take two consecutive slots because these types are double width (64-bit instead of 32-bit).<\/p><h3 id=\"operand_stack\">Operand Stack<\/h3><p>The operand stack is used during the execution of byte code instructions in a similar way that general-purpose registers are used in a native CPU.  Most JVM byte code spends its time manipulating the operand stack by pushing, popping, duplicating, swapping, or executing operations that produce or consume values.  Therefore, instructions that move values between the array of local variables and the operand stack are very frequent in byte code. For example, a simple variable initialization results in two byte codes that interact with the operand stack.<\/p><pre class=\"code\">int i;<\/pre><p>Gets compiled to the following byte code:<\/p><pre class=\"code\"> 0:\ticonst_0\t\/\/ Push 0 to top of the operand stack\n 1:\tistore_1\t\/\/ Pop value from top of operand stack and store as local variable 1<\/pre><p>For more detail explaining interactions between the local variables array, operand stack and run time constant pool <a href=\"#class_file_structure\">see the section on Class File Structure below<\/a>.<\/p><h3 id=\"dynamic_linking\">Dynamic Linking<\/h3><p>Each frame contains a reference to the runtime constant pool.  The reference points to the constant pool for the class of the method being executed for that frame.  This reference helps to support dynamic linking.<\/p><p>C\/C++ code is typically compiled to an object file then multiple object files are linked together to product a usable artifact such as an executable or dll.  During the linking phase symbolic references in each object file are replaced with an actual memory address relative to the final executable.  In Java this linking phase is done dynamically at runtime.<\/p><p>When a Java class is compiled, all references to variables and methods are stored in the class's constant pool as a symbolic reference. A symbolic reference is a logical reference not a reference that actually points to a physical memory location.  The JVM implementation can choose when to resolve symbolic references, this can happen when the class file is verified, after being loaded, called eager or static resolution, instead this can happen when the symbolic reference is used for the first time called lazy or late resolution. However the JVM has to behave as if the resolution occurred when each reference is first used and throw any resolution errors at this point.  Binding is the process of the field, method or class identified by the symbolic reference being replaced by a direct reference, this only happens once because the symbolic reference is completely replaced.  If the symbolic reference refers to a class that has not yet been resolved then this class will be loaded.  Each direct reference is stored as an offset against the storage structure associated with the runtime location of the variable or method.<\/p><h3 id=\"heap\">Heap<\/h3><p>The Heap is used to allocate class instances and arrays at runtime. Arrays and objects can never be stored on the stack because a frame is not designed to change in size after it has been created.  The frame only stores references that point to objects or arrays on the heap.  Unlike primitive variables and references in the local variable array (in each frame) objects are always stored on the heap so they are not removed when a method ends.  Instead objects are only removed by the garbage collector.<\/p><p>To support garbage collection the heap is divided into three sections:<\/p><div><ul class=\"list\"><li><strong>Young Generation<\/strong>\n            <div><ul><li>Often split between Eden and Survivor<\/li><\/ul><\/div><\/li>\n            <li><strong>Old Generation<\/strong> (also called Tenured Generation)<\/li>\n            <li><strong>Permanent Generation<\/strong><\/li>\n        <\/ul><\/div><h3 id=\"memory_management\">Memory Management<\/h3><p>Objects and Arrays are never explicitly de-allocated instead the garbage collector automatically reclaims them.<\/p><p>Typically this works as follows:<\/p><div><ol class=\"list\"><li>New objects and arrays are created into the young generation<\/li>\n            <li>Minor garbage collection will operate in the young generation.  Objects, that are still alive, will be moved from the eden space to the survivor space.<\/li>\n            <li>Major garbage collection, which typically causes the application threads to pause, will move objects between generations.  Objects, that are still alive, will be moved from the young generation to the old (tenured) generation.<\/li>\n            <li>The permanent generation is collected every time the old generation is collected.  They are both collected when either becomes full.<\/li>\n        <\/ol><\/div><h3 id=\"non_heap_memory\">Non-Heap Memory<\/h3><p>Objects that are logically considered as part of the JVM mechanics are not created on the Heap.<\/p><p>The non-heap memory includes:<\/p><div><ul class=\"list\"><li><strong>Permanent Generation<\/strong> that contains\n                <div><ul><li>the method area<\/li>\n                    <li>interned strings<\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Code Cache<\/strong> used for compilation and storage of methods that have been compiled to native code by the JIT compiler<\/li>\n        <\/ul><\/div><h3 id=\"jit_compilation\">Just In Time (JIT) Compilation<\/h3><p>Java byte code is interpreted however this is not as fast as directly executing native code on the JVM\u2019s host CPU.  To improve performance the Oracle Hotspot VM looks for \u201chot\u201d areas of byte code that are executed regularly and compiles these to native code.  The native code is then stored in the code cache in non-heap memory.  In this way the Hotspot VM tries to choose the most appropriate way to trade-off the extra time it takes to compile code verses the extra time it take to execute interpreted code.<\/p><h3 id=\"method_area\">Method Area<\/h3><p>The method area stores per-class information such as:<\/p><div><ul class=\"list\"><li><strong>Classloader Reference<\/strong><\/li>\n            <li><strong>Run Time Constant Pool<\/strong>\n                <div><ul><li>Numeric constants<\/li>\n                    <li>Field references<\/li>\n                    <li>Method References<\/li>\n                    <li>Attributes<\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Field data<\/strong>\n                <div><ul><li>Per field\n                        <div><ul><li>Name<\/li>\n                            <li>Type<\/li>\n                            <li>Modifiers<\/li>\n                            <li>Attributes<\/li>\n                        <\/ul><\/div>\n                    <\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Method data<\/strong>\n                <div><ul><li>Per method\n                        <div><ul><li>Name<\/li>\n                            <li>Return Type<\/li>\n                            <li>Parameter Types (in order)<\/li>\n                            <li>Modifiers<\/li>\n                            <li>Attributes<\/li>\n                        <\/ul><\/div>\n                    <\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Method code<\/strong>\n                <div><ul><li>Per method\n                        <div><ul><li>Bytecodes<\/li>\n                            <li>Operand stack size<\/li>\n                            <li>Local variable size<\/li>\n                            <li>Local variable table<\/li>\n                            <li>Exception table\n                                <div><ul><li>Per exception handler\n                                        <div><ul><li>Start point<\/li>\n                                            <li>End point<\/li>\n                                            <li>PC offset for handler code<\/li>\n                                            <li>Constant pool index for exception class being caught<\/li>\n                                        <\/ul><\/div>\n                                    <\/li>\n                                <\/ul><\/div>\n                            <\/li>\n                        <\/ul><\/div>\n                    <\/li>\n                <\/ul><\/div>\n            <\/li>\n        <\/ul><\/div><p>All threads share the same method area, so access to the method area data and the process of dynamic linking must be thread safe.  If two threads attempt to access a field or method on a class that has not yet been loaded it must only be loaded once and both threads must not continue execution until it has been loaded.<\/p><h3 id=\"class_file_structure\">Class File Structure<\/h3><p>A compiled class file consists of the following structure:<\/p><pre class=\"code\">ClassFile {\n    u4\t\t\tmagic;\n    u2\t\t\tminor_version;\n    u2\t\t\tmajor_version;\n    u2\t\t\tconstant_pool_count;\n    cp_info\t\tcontant_pool[constant_pool_count \u2013 1];\n    u2\t\t\taccess_flags;\n    u2\t\t\tthis_class;\n    u2\t\t\tsuper_class;\n    u2\t\t\tinterfaces_count;\n    u2\t\t\tinterfaces[interfaces_count];\n    u2\t\t\tfields_count;\n    field_info\t\tfields[fields_count];\n    u2\t\t\tmethods_count;\n    method_info\t\tmethods[methods_count];\n    u2\t\t\tattributes_count;\n    attribute_info\tattributes[attributes_count];\n}<\/pre><div class=\"columns\"><div class=\"row\"><p>magic,\n minor_version,\n major_version<\/p><p>specifies information about the version of the class and the version of the JDK this class was compiled for.<\/p><\/div><div class=\"row\"><p>constant_pool<\/p><p>similar to a symbol table although it contains more data <a href=\"#constant_pool\">this is described in more detail below.<\/a><\/p><\/div><div class=\"row\"><p>access_flags<\/p><p>provides the list of modifiers for this class.<\/p><\/div><div class=\"row\"><p>this_class<\/p><p>index into the constant_pool providing the fully qualified name of this class i.e. org\/jamesdbloom\/foo\/Bar<\/p><\/div><div class=\"row\"><p>super_class<\/p><p>index into the constant_pool providing a symbolic reference to the super class i.e. java\/lang\/Object<\/p><\/div><div class=\"row\"><p>interfaces<\/p><p>array of indexes into the constant_pool providing a symbolic references to all interfaces that have been implemented.<\/p><\/div><div class=\"row\"><p>fields<\/p><p>array of indexes into the constant_pool giving a complete description of each field.<\/p><\/div><div class=\"row\"><p>methods<\/p><p>array of indexes into the constant_pool giving a complete description of each method signature, if the method is not abstract or native then the bytecode is also present.<\/p><\/div><div class=\"row\"><p>attributes<\/p><p>array of different value that provide additional information about the class including any annotations with RetentionPolicy.CLASS or RetentionPolicy.RUNTIME<\/p><\/div><\/div><p>It is possible to view the byte code in a compiled Java class by using the javap command.<\/p><p>If you compile the following simple class:<\/p><pre class=\"code\">package org.jvminternals;\npublic class SimpleClass {\n    public void sayHello() {\n        System.out.println(\"Hello\");\n    }\n}<\/pre><p>Then you get the following output if you run:<\/p><p>javap -v -p -s -sysinfo -constants classes\/org\/jvminternals\/SimpleClass.class<\/p><pre class=\"code\">public class org.jvminternals.SimpleClass\n  SourceFile: \"SimpleClass.java\"\n  minor version: 0\n  major version: 51\n  flags: ACC_PUBLIC, ACC_SUPER\nConstant pool:\n   #1 = Methodref          #6.#17         \/\/  java\/lang\/Object.\"&lt;init&gt;\":()V\n   #2 = Fieldref           #18.#19        \/\/  java\/lang\/System.out:Ljava\/io\/PrintStream;\n   #3 = String             #20            \/\/  \"Hello\"\n   #4 = Methodref          #21.#22        \/\/  java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n   #5 = Class              #23            \/\/  org\/jvminternals\/SimpleClass\n   #6 = Class              #24            \/\/  java\/lang\/Object\n   #7 = Utf8               &lt;init&gt;\n   #8 = Utf8               ()V\n   #9 = Utf8               Code\n  #10 = Utf8               LineNumberTable\n  #11 = Utf8               LocalVariableTable\n  #12 = Utf8               this\n  #13 = Utf8               Lorg\/jvminternals\/SimpleClass;\n  #14 = Utf8               sayHello\n  #15 = Utf8               SourceFile\n  #16 = Utf8               SimpleClass.java\n  #17 = NameAndType        #7:#8          \/\/  \"&lt;init&gt;\":()V\n  #18 = Class              #25            \/\/  java\/lang\/System\n  #19 = NameAndType        #26:#27        \/\/  out:Ljava\/io\/PrintStream;\n  #20 = Utf8               Hello\n  #21 = Class              #28            \/\/  java\/io\/PrintStream\n  #22 = NameAndType        #29:#30        \/\/  println:(Ljava\/lang\/String;)V\n  #23 = Utf8               org\/jvminternals\/SimpleClass\n  #24 = Utf8               java\/lang\/Object\n  #25 = Utf8               java\/lang\/System\n  #26 = Utf8               out\n  #27 = Utf8               Ljava\/io\/PrintStream;\n  #28 = Utf8               java\/io\/PrintStream\n  #29 = Utf8               println\n  #30 = Utf8               (Ljava\/lang\/String;)V\n{\n  public org.jvminternals.SimpleClass();\n    Signature: ()V\n    flags: ACC_PUBLIC\n    Code:\n      stack=1, locals=1, args_size=1\n        0: aload_0\n        1: invokespecial #1    \/\/ Method java\/lang\/Object.\"&lt;init&gt;\":()V\n        4: return\n      LineNumberTable:\n        line 3: 0\n      LocalVariableTable:\n        Start  Length  Slot  Name   Signature\n          0      5      0    this   Lorg\/jvminternals\/SimpleClass;\n  public void sayHello();\n    Signature: ()V\n    flags: ACC_PUBLIC\n    Code:\n      stack=2, locals=1, args_size=1\n        0: getstatic      #2    \/\/ Field java\/lang\/System.out:Ljava\/io\/PrintStream;\n        3: ldc            #3    \/\/ String \"Hello\"\n        5: invokevirtual  #4    \/\/ Method java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n        8: return\n      LineNumberTable:\n        line 6: 0\n        line 7: 8\n      LocalVariableTable:\n        Start  Length  Slot  Name   Signature\n          0      9      0    this   Lorg\/jvminternals\/SimpleClass;\n}<\/pre><p>This class file shows three main sections the constant pool, the constructor and the sayHello method.<\/p><div><ul class=\"list\"><li><strong>Constant Pool<\/strong> \u2013 this provides the same information that a symbol table typically provides and is <a href=\"#constant_pool\">described in more detail below.<\/a><\/li>\n            <li><strong>Methods<\/strong> \u2013 each containing four areas:\n                <div><ul><li>signature and access flags<\/li>\n                    <li>byte code<\/li>\n                    <li>LineNumberTable \u2013 this provides information to a debugger to indicate which line corresponds to which byte code instruction, for example line 6 in the Java code corresponds to byte code 0 in the sayHello method and line 7 corresponds to byte code 8.<\/li>\n                    <li>LocalVariableTable \u2013 this lists all local variables provided in the frame, in both examples the only local variable is this.<\/li>\n                <\/ul><\/div>\n            <\/li>\n        <\/ul><\/div><p>The following byte code operands are used in this class file<\/p><div class=\"columns\"><div class=\"row\"><p>aload_0<\/p><p>This opcode is one of a group of opcodes with the format aload_&lt;n&gt;.  They all load an object reference into the operand stack.  The &lt;n&gt; refers to the location in the local variable array that is being accessed but can only be 0, 1, 2 or 3.  There are other similar opcodes for loading values that are not an object reference iload_&lt;n&amp;gt, lload_&lt;n&amp;gt, float_&lt;n&amp;gt and dload_&lt;n&amp;gt where i is for int, l is for long, f is for float and d is for double.  Local variables with an index higher than 3 can be loaded using iload, lload, float, dload and aload.  These opcodes all take a single operand that specifies the index of local variable to load.<\/p><\/div><div class=\"row\"><p>ldc<\/p><p>This opcode is used to push a constant from the run time constant pool into the operand stack.<\/p><\/div><div class=\"row\"><p>getstatic<\/p><p>This opcode is used to push a static value from a static field listed in the run time constant pool into the operand stack.<\/p><\/div><div class=\"row\"><p>invokespecial, invokevirtual<\/p><p>These opcodes are in a group of opcodes that invoke methods these are invokedynamic, invokeinterface, invokespecial, invokestatic, invokevirtual.  In this class file invokespecial and invokevirutal are both used the difference between these is that invokevirutal invokes a method based on the class of the object. The invokespecial instruction is used to invoke instance initialization methods as well as private methods and methods of a superclass of the current class.<\/p><\/div><div class=\"row\"><p>return<\/p><p>This opcode is in a group of opcodes ireturn, lreturn, freturn, dreturn, areturn and return.  Each of these opcodes are a typed return statement that returns a different type where i is for int, l is for long, f is for float, d is for double and a is for an object reference.  The opcode with no leading type letter return only returns void.<\/p><\/div><\/div><p>As in any typical byte code the majority of the operands interact with the local variables, operand stack and run time constant pool as follows.<\/p><p>The constructor has two instructions first this is pushed onto the operand stack, next the constructor for the super class is invoked which consumes the value off this and therefore pops it off the operand stack.<\/p><a class=\"image_link\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_SimpleClass.png\"><img itemprop=\"image\" id=\"bytecode_explanation_SimpleClass_image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_SimpleClass.png\" alt=\"local variables, operand stack and run time constant pool changes when invoking a static method in Java Virtual Machine (JVM)\" \/><\/a><p>The sayHello() method is more complex as it has to resolve symbolic references to actual references using the run time constant pool, <a href=\"#dynamic_linking\">as explained in more detail above<\/a>.  The first operand getstatic is used to push a reference to the static field out of the System class on to the operand stack.  The next operand ldc pushes the string \"Hello\" onto the operand stack.  The final operand invokevirtual invokes the println method of System.out which pops \"Hello\" off the operand stack as an argument and creates a new frame for the current thread.<\/p><a class=\"image_link\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_sayHello.png\"><img itemprop=\"image\" id=\"bytecode_explanation_sayHello_image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_sayHello_smaller.png\" alt=\"local variables, operand stack and run time constant pool changes when calling System.out in Java Virtual Machine (JVM)\" \/><\/a><h3 id=\"classloader\">Classloader<\/h3><p>The JVM starts up by loading an initial class using the bootstrap classloader.  The class is then linked and initialized before public static void main(String[]) is invoked.  The execution of this method will in turn drive the loading, linking and initialization of additional classes and interfaces as required.<\/p><p><strong>Loading<\/strong> is the process of finding the class file that represents the class or interface type with a particular name and reading it into a byte array.  Next the bytes are parsed to confirm they represent a Class object and have the correct major and minor versions.  Any class or interface named as a direct superclass is also loaded.  Once this is completed a class or interface object is created from the binary representation.<\/p><p><strong>Linking<\/strong> is the process of taking a class or interface verifying and preparing the type and its direct superclass and superinterfaces.  Linking consists of three steps verifying, preparing and optionally resolving.<\/p><div><p><strong><em>Verifying<\/em><\/strong> is the process of confirming the class or interface representation is structurally correct and obeys the semantic requirements of the Java programming language and JVM, for example the following checks are performed:<\/p><div><ol><li>consistent and correctly formatted symbol table<\/li>\n                <li>final methods \/ classes not overridden<\/li>\n                <li>methods respect access control keywords<\/li>\n                <li>methods have correct number and type of parameters<\/li>\n                <li>bytecode doesn't manipulate stack incorrectly<\/li>\n                <li>variables are initialized before being read<\/li>\n                <li>variables are a value of the correct type<\/li>\n            <\/ol><\/div><p>Performing these checks during the verifying stages means these checks do not need to be performed at runtime.  Verification during linking slows down class loading however it avoids the need to perform these checks multiple when executing the bytecode.<\/p><p><strong><em>Preparing<\/em><\/strong> involves allocation of memory for static storage and any data structures used by the JVM such as method tables.  Static fields are created and initialized to their default values, however, no initializers or code is executed at this stage as that happens as part of initialization.<\/p><p><strong><em>Resolving<\/em><\/strong> is an optional stage which involves checking symbolic references by loading the referenced classes or interfaces and checking the references are correct.  If this does not take place at this point the resolution of symbolic references can be deferred until just prior to their use by a byte code instruction.<\/p><\/div><p><strong>Initialization<\/strong> of a class or interface consists of executing the class or interface initialization method &lt;clinit&gt;<\/p><a id=\"Class_Loading_Linking_Initializing_image\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/Class_Loading_Linking_Initializing.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/Class_Loading_Linking_Initializing.png\" alt=\"class loading, linking and initialization in the Java Virtual Machine (JVM)\" \/><\/a><p>In the JVM there are multiple classloaders with different roles.  Each classloader delegates to its parent classloader (that loaded it) except the <strong>bootstrap classloader<\/strong> which is the top classloader.<\/p><div><p><strong><em>Bootstrap Classloader<\/em><\/strong> is usually implemented as native code because it is instantiated very early as the JVM is loaded.  The bootstrap classloader is responsible for loading the basic Java APIs, including for example rt.jar.  It only loads classes found on the boot classpath which have a higher level of trust; as a result it skips much of the validation that gets done for normal classes.<\/p><p><strong><em>Extension Classloader<\/em><\/strong> loads classes from standard Java extension APIs such as security extension functions.<\/p><p><strong><em>System Classloader<\/em><\/strong> is the default application classloader, which loads application classes from the classpath.<\/p><p><strong><em>User Defined Classloaders<\/em><\/strong> can alternatively be used to load application classes.  A user defined classloader is used for a number of special reasons including run time reloading of classes or separation between different groups of loaded classes typically required by web servers such as Tomcat.<\/p><\/div><a id=\"class_loader_hierarchy_image\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/class_loader_hierarchy.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/class_loader_hierarchy.png\" alt=\"classloader hierarchy in the Java Virtual Machine (JVM)\" \/><\/a><h3 id=\"faster_class_loading\">Faster Class Loading<\/h3><p>A feature called Class Data Sharing (CDS) was introduce in HotSpot JMV from version 5.0.  During the installation process of the JVM the installer loads a set of key JVM classes, such as rt.jar, into a memory-mapped shared archive.  CDS reduces the time it takes to load these classes improving JVM start-up speed and allows these classes to be shared between different instances of the JVM reducing the memory footprint.<\/p><h3 id=\"where_is_the_method_area\">Where Is The Method Area<\/h3><p><a href=\"http:\/\/www.amazon.co.uk\/Virtual-Machine-Specification-Edition-Series\/dp\/0133260445\">The Java Virtual Machine Specification Java SE 7 Edition<\/a> clearly states:  \u201cAlthough the method area is logically part of the heap, simple implementations may choose not to either garbage collect or compact it.\u201d  In contradiction to this jconsole for the Oracle JVM shows the method area (and code cache) as being non-heap.  The <a href=\"http:\/\/openjdk.java.net\/\">OpenJDK<\/a> code shows that the CodeCache is a separate field of the VM to the ObjectHeap.<\/p><h3 id=\"classloader_reference\">Classloader Reference<\/h3><p>All classes that are loaded contain a reference to the classloader that loaded them.  In turn the classloader also contains a reference to all classes that it has loaded.<\/p><h3 id=\"constant_pool\">Run Time Constant Pool<\/h3><p>The JVM maintains a per-type constant pool, a run time data structure that is similar to a symbol table although it contains more data.  Byte codes in Java require data, often this data is too large to store directly in the byte codes, instead it is stored in the constant pool and the byte code contains a reference to the constant pool. The run time constant pool is used in dynamic linking <a href=\"#dynamic_linking\">as described above<\/a><\/p><p>Several types of data is stored in the constant pool including<\/p><div><ul class=\"list\"><li>numeric literals<\/li>\n            <li>string literals<\/li>\n            <li>class references<\/li>\n            <li>field references<\/li>\n            <li>method references<\/li>\n        <\/ul><\/div><p>For example the following code:<\/p><pre class=\"code\">Object foo = new Object();<\/pre><p>Would be written in byte code as follows:<\/p><pre class=\"code\"> 0: \tnew #2 \t\t    \/\/ Class java\/lang\/Object\n 1:\tdup\n 2:\tinvokespecial #3    \/\/ Method java\/ lang\/Object \"&lt;init&gt;\"( ) V<\/pre><p>The new opcode (operand code) is followed by the #2 operand.  This operand is an index into the constant pool and therefore is referencing the second entry in the constant pool.  The second entry is a class reference, this entry in turn references another entry in the constant pool containing the name of the class as a constant UTF8 string with the value \/\/ Class java\/lang\/Object.  This symbolic link can then be used to lookup the class for java.lang.Object.  The new opcode creates a class instance and initializes its variables.  A reference to the new class instance is then added to the operand stack.  The dup opcode then creates an extra copy of the top reference on the operand stack and adds this to the top of the operand stack.  Finally an instance initialization method is called on line 2 by invokespecial.  This operand also contains a reference to the constant pool.  The initialization method consumes (pops) the top reference off the operand pool as an argument to the method.  At the end there is one reference to the new object that has been both created and initialized.<\/p><p>If you compile the following simple class:<\/p><pre class=\"code\">package org.jvminternals;\npublic class SimpleClass {\n    public void sayHello() {\n        System.out.println(\"Hello\");\n    }\n}<\/pre><p>The constant pool in the generated class file would look like:<\/p><pre class=\"code\">Constant pool:\n   #1 = Methodref          #6.#17         \/\/  java\/lang\/Object.\"&lt;init&gt;\":()V\n   #2 = Fieldref           #18.#19        \/\/  java\/lang\/System.out:Ljava\/io\/PrintStream;\n   #3 = String             #20            \/\/  \"Hello\"\n   #4 = Methodref          #21.#22        \/\/  java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n   #5 = Class              #23            \/\/  org\/jvminternals\/SimpleClass\n   #6 = Class              #24            \/\/  java\/lang\/Object\n   #7 = Utf8               &lt;init&gt;\n   #8 = Utf8               ()V\n   #9 = Utf8               Code\n  #10 = Utf8               LineNumberTable\n  #11 = Utf8               LocalVariableTable\n  #12 = Utf8               this\n  #13 = Utf8               Lorg\/jvminternals\/SimpleClass;\n  #14 = Utf8               sayHello\n  #15 = Utf8               SourceFile\n  #16 = Utf8               SimpleClass.java\n  #17 = NameAndType        #7:#8          \/\/  \"&lt;init&gt;\":()V\n  #18 = Class              #25            \/\/  java\/lang\/System\n  #19 = NameAndType        #26:#27        \/\/  out:Ljava\/io\/PrintStream;\n  #20 = Utf8               Hello\n  #21 = Class              #28            \/\/  java\/io\/PrintStream\n  #22 = NameAndType        #29:#30        \/\/  println:(Ljava\/lang\/String;)V\n  #23 = Utf8               org\/jvminternals\/SimpleClass\n  #24 = Utf8               java\/lang\/Object\n  #25 = Utf8               java\/lang\/System\n  #26 = Utf8               out\n  #27 = Utf8               Ljava\/io\/PrintStream;\n  #28 = Utf8               java\/io\/PrintStream\n  #29 = Utf8               println\n  #30 = Utf8               (Ljava\/lang\/String;)V\n<\/pre><p>The constant pool contains the following types:<\/p><div class=\"columns\"><div class=\"row\"><p>Integer<\/p><p>A 4 byte int constant<\/p><\/div><div class=\"row\"><p>Long<\/p><p>An 8 byte long constant<\/p><\/div><div class=\"row\"><p>Float<\/p><p>A 4 byte float constant<\/p><\/div><div class=\"row\"><p>Double<\/p><p>A 8 byte double constant<\/p><\/div><div class=\"row\"><p>String<\/p><p>A String constant that points at another Utf8 entry in the constant pool which contains the actual bytes<\/p><\/div><div class=\"row\"><p>Utf8<\/p><p>A stream of bytes representing a Utf8 encoded sequence of characters<\/p><\/div><div class=\"row\"><p>Class<\/p><p>A Class constant that points at another Utf8 entry in the constant pool which contains the fully qualified class name in the internal JVM format (this is used by the <a href=\"#dynamic_linking\">dynamic linking process<\/a>)<\/p><\/div><div class=\"row\"><p>NameAndType<\/p><p>A colon separated pair of values each pointing at other entries in the constant pool.  The first value (before the colon) points at a Utf8 string entry that is the method or field name.  The second value points at a Utf8 entry that represents the type, in the case of a field this is the fully qualified class name, in the case of a method this is a list of fully qualified class names one per parameter.<\/p><\/div><div class=\"row\"><p>Fieldref,\nMethodref,\nInterfaceMethodref<\/p><p>A dot separated pair of values each pointing at other entries in the constant pool.  The first value (before the dot) points at a Class entry.  The second value points at a NameAndType entry.<\/p><\/div><\/div><h3 id=\"exception_table\">Exception Table<\/h3><p>The exception table stores per-exception handler information such as:<\/p><div><ul class=\"list\"><li>Start point<\/li>\n            <li>End point<\/li>\n            <li>PC offset for handler code<\/li>\n            <li>Constant pool index for exception class being caught<\/li>\n        <\/ul><\/div><p>If a method has defined a try-catch or a try-finally exception handler then an Exception Table will be created.  This contains information for each exception handler or finally block including the range over which the handler applies, what type of exception is being handled and where the handler code is.<\/p><p>When an exception is thrown the JVM looks for a matching handler in the current method, if none is found the method ends abruptly popping the current stack frame and the exception is re-thrown in the calling method (the new current frame).  If no exception handler is found before all frames have been popped then the thread is terminated.   This can also cause the JVM itself to terminate if the exception is thrown in the last non-daemon thread, for example if the thread is the main thread.<\/p><p>Finally exception handlers match all types of exceptions and so always execute whenever an exception is thrown.  In the case when no exception is thrown a finally block is still executed at the end of a method, this is achieved by jumping to the finally handler code immediately before the return statement is executed.<\/p><h3 id=\"symbol_table\">Symbol Table<\/h3><p>In addition to per-type run-time constant pools the Hotspot JVM has a symbol table held in the permanent generation.  The symbol table is a Hashtable mapping symbol pointers to symbols (i.e. Hashtable&lt;Symbol*, Symbol&gt;) and includes a pointer to all symbols including those held in run time constant pools in each class.<\/p><p>Reference counting is used to control when a symbol is removed from the symbol table.  For example when a class is unloaded the reference count of all symbols held in its run time constant pool are decremented.  When the reference count of a symbol in the the symbol table goes to zero then the symbol table knows that symbol is not being referenced anymore and the symbol is unloaded from the symbol table. For both the symbol table and the string table (see below) all entries are held in a canonicalized form to improve efficiency and ensure each entry only appears once.<\/p><h3 id=\"string_table\">Interned Strings (String Table)<\/h3><p>The Java Language Specification requires that identical string literals, that contain the same sequence of Unicode code points, must refer to the same instance of String.  In addition if String.intern() is called on an instance of String a reference must be returned that would be identical to the reference return if the string was a literal. The following therefore holds true:<\/p><pre class=\"code\">(\"j\" + \"v\" + \"m\").intern() == \"jvm\"<\/pre><p>In the Hotspot JVM interned string are held in the string table, which is a Hashtable mapping object pointers to symbols (i.e. Hashtable&lt;oop, Symbol&gt;), and is held in the permanent generation.  For both the symbol table (see above) and the string table all entries are held in a canonicalized form to improve efficiency and ensure each entry only appears once.<\/p><p>String literals are automatically interned by the compiler and added into the symbol table when the class is loaded.  In addition instances of the String class can be explicitly interned by calling String.intern().  When String.intern() is called, if the symbol table already contains the string then a reference to this is returned, if not the string is added to the string table and its reference is returned.<\/p>","created_at":"2018-11-07T16:23:25+0000","updated_at":"2018-11-07T16:23:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":26,"domain_name":"blog.jamesdbloom.com","preview_picture":"http:\/\/blog.jamesdbloom.com\/apple-touch-icon_2013_11_17_17_56.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":147,"title":"JVM Internals","url":"http:\/\/blog.jamesdbloom.com\/JVMInternals.html","content":"<p>This article explains the internal architecture of the Java Virtual Machine (JVM).  The following diagram show the key internal components of a typical JVM that conforms to <a href=\"http:\/\/www.amazon.co.uk\/Virtual-Machine-Specification-Edition-Series\/dp\/0133260445\">The Java Virtual Machine Specification Java SE 7 Edition<\/a>.<\/p><a class=\"image_link\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/JVM_Internal_Architecture.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/JVM_Internal_Architecture_small.png\" alt=\"internal architecture of Java Virtual Machine (JVM)\" \/><\/a><p>The components shown on this diagram are each explained below in two sections.  <a href=\"#threads\">First section<\/a> covers the components that are created for each thread and the <a href=\"#shared_between_threads\">second section<\/a> covers the components that are created independently of threads.<\/p><div><ul><li><a href=\"#threads\">Threads<\/a>\n                    <div><ul><li><a href=\"#jvm_system_threads\">JVM System Threads<\/a><\/li>\n                            <li><a href=\"#per_thread\">Per Thread<\/a><\/li>\n                            <li><a href=\"#program_counter\">program Counter (PC)<\/a><\/li>\n                            <li><a href=\"#stack\">Stack<\/a><\/li>\n                            <li><a href=\"#native_stack\">Native Stack<\/a><\/li>\n                            <li><a href=\"#stack_restrictions\">Stack Restrictions<\/a><\/li>\n                            <li><a href=\"#frame\">Frame<\/a><\/li>\n                            <li><a href=\"#local_variables_array\">Local Variables Array<\/a><\/li>\n                            <li><a href=\"#operand_stack\">Operand Stack<\/a><\/li>\n                            <li><a href=\"#dynamic_linking\">Dynamic Linking<\/a><\/li>\n                        <\/ul><\/div>\n                <\/li>\n                <li><a href=\"#shared_between_threads\">Shared Between Threads<\/a>\n                    <div><ul><li><a href=\"#heap\">Heap<\/a><\/li>\n                            <li><a href=\"#memory_management\">Memory Management<\/a><\/li>\n                            <li><a href=\"#non_heap_memory\">Non-Heap Memory<\/a><\/li>\n                            <li><a href=\"#jit_compilation\">Just In Time (JIT) Compilation<\/a><\/li>\n                            <li><a href=\"#method_area\">Method Area<\/a><\/li>\n                            <li><a href=\"#class_file_structure\">Class File Structure<\/a><\/li>\n                            \n                            <li><a href=\"#classloader\">Classloader<\/a><\/li>\n                            <li><a href=\"#faster_class_loading\">Faster Class Loading<\/a><\/li>\n                            <li><a href=\"#where_is_the_method_area\">Where Is The Method Area<\/a><\/li>\n                            <li><a href=\"#classloader_reference\">Classloader Reference<\/a><\/li>\n                            <li><a href=\"#constant_pool\">Run Time Constant Pool<\/a><\/li>\n                            <li><a href=\"#exception_table\">Exception Table<\/a><\/li>\n                            <li><a href=\"#symbol_table\">Symbol Table<\/a><\/li>\n                            <li><a href=\"#string_table\">Interned Strings (String Table)<\/a><\/li>\n                        <\/ul><\/div>\n                <\/li>\n            <\/ul><\/div><h2>Thread<\/h2><p>A thread is a thread of execution in a program. The JVM allows an application to have multiple threads of execution running concurrently.  In the Hotspot JVM there is a direct mapping between a Java Thread and a native operating system Thread.  After preparing all of the state for a Java thread such as thread-local storage, allocation buffers, synchronization objects, stacks and the program counter, the native thread is created.  The native thread is reclaimed once the Java thread terminates.  The operating system is therefore responsible for scheduling all threads and dispatching them to any available CPU.  Once the native thread has initialized it invokes the run() method in the Java thread.  When the run() method returns, uncaught exceptions are handled, then the native thread confirms if the JVM needs to be terminated as a result of the thread terminating (i.e. is it the last non-deamon thread).  When the thread terminates all resources for both the native and Java thread are released.<\/p><h3>JVM System Threads<\/h3><p>If you use jconsole or any debugger it is possible to see there are numerous threads running in the background.  These background threads run in addition to the main thread, which is created as part of invoking public static void main(String[]), and any threads created by the main thread.  The main background system threads in the Hotspot JVM are:<\/p><div class=\"columns\"><div class=\"row\"><p>VM thread<\/p><p>This thread waits for operations to appear that require the JVM to reach a safe-point.  The reason these operations have to happen on a separate thread is because they all require the JVM to be at a safe point where modifications to the heap can not occur.  The type of operations performed by this thread are \"stop-the-world\" garbage collections, thread stack dumps, thread suspension and biased locking revocation.<\/p><\/div><div class=\"row\"><p>Periodic task thread<\/p><p>This thread is responsible for timer events (i.e. interrupts) that are used to schedule execution of periodic operations<\/p><\/div><div class=\"row\"><p>GC threads<\/p><p>These threads support the different types of garbage collection activities that occur in the JVM<\/p><\/div><div class=\"row\"><p>Compiler threads<\/p><p>These threads compile byte code to native code at runtime<\/p><\/div><div class=\"row\"><p>Signal dispatcher thread<\/p><p>This thread receives signals sent to the JVM process and handle them inside the JVM by calling the appropriate JVM methods.<\/p><\/div><\/div><h2>Per Thread<\/h2><p>Each thread of execution has the following components:<\/p><h3>Program Counter (PC)<\/h3><p>Address of the current instruction (or opcode) unless it is native.  If the current method is native then the PC is undefined.  All CPUs have a PC, typically the PC is incremented after each instruction and therefore holds the address of the next instruction to be executed.  The JVM uses the PC to keep track of where it is executing instructions, the PC will in fact be pointing at a memory address in the Method Area.<\/p><h3>Stack<\/h3><p>Each thread has its own stack that holds a frame for each method executing on that thread.  The stack is a Last In First Out (LIFO) data structure, so the currently executing method is at the top of the stack. A new frame is created and added (pushed) to the top of stack for every method invocation.  The frame is removed (popped) when the method returns normally or if an uncaught exception is thrown during the method invocation. The stack is not directly manipulated, except to push and pop frame objects, and therefore the frame objects may be allocated in the Heap and the memory does not need to be contiguous.<\/p><h3>Native Stack<\/h3><p>Not all JVMs support native methods, however, those that do typically create a per thread native method stack.  If a JVM has been implemented using a C-linkage model for Java Native Invocation (JNI) then the native stack will be a C stack.  In this case the order of arguments and return value will be identical in the native stack to typical C program. A native method can typically (depending on the JVM implementation) call back into the JVM and invoke a Java method.  Such a native to Java invocation will occur on the stack (normal Java stack); the thread will leave the native stack and create a new frame on the stack (normal Java stack).<\/p><h3>Stack Restrictions<\/h3><p>A stack can be a dynamic or fixed size.  If a thread requires a larger stack than allowed a StackOverflowError is thrown.  If a thread requires a new frame and there isn\u2019t enough memory to allocate it then an OutOfMemoryError is thrown.<\/p><h3>Frame<\/h3><p>A new frame is created and added (pushed) to the top of stack for every method invocation.  The frame is removed (popped) when the method returns normally or if an uncaught exception is thrown during the method invocation.  For more detail on exception handling <a href=\"#exception_table\">see the section on Exception Tables below<\/a>.<\/p><p>Each frame contains:<\/p><div><ul class=\"list\"><li>Local variable array<\/li>\n            <li>Return value<\/li>\n            <li>Operand stack<\/li>\n            <li>Reference to runtime constant pool for class of the current method<\/li>\n        <\/ul><\/div><h3>Local Variables Array<\/h3><p>The array of local variables contains all the variables used during the execution of the method, including a reference to this, all method parameters and other locally defined variables.  For class methods (i.e. static methods) the method parameters start from zero, however, for instance method the zero slot is reserved for this.<\/p><p>A local variable can be:<\/p><div><ul class=\"list\"><li>boolean<\/li>\n            <li>byte<\/li>\n            <li>char<\/li>\n            <li>long<\/li>\n            <li>short<\/li>\n            <li>int<\/li>\n            <li>float<\/li>\n            <li>double<\/li>\n            <li>reference<\/li>\n            <li>returnAddress<\/li>\n        <\/ul><\/div><p>All types take a single slot in the local variable array except long and double which both take two consecutive slots because these types are double width (64-bit instead of 32-bit).<\/p><h3>Operand Stack<\/h3><p>The operand stack is used during the execution of byte code instructions in a similar way that general-purpose registers are used in a native CPU.  Most JVM byte code spends its time manipulating the operand stack by pushing, popping, duplicating, swapping, or executing operations that produce or consume values.  Therefore, instructions that move values between the array of local variables and the operand stack are very frequent in byte code. For example, a simple variable initialization results in two byte codes that interact with the operand stack.<\/p><pre class=\"code\">int i;<\/pre><p>Gets compiled to the following byte code:<\/p><pre class=\"code\"> 0:\ticonst_0\t\/\/ Push 0 to top of the operand stack\n 1:\tistore_1\t\/\/ Pop value from top of operand stack and store as local variable 1<\/pre><p>For more detail explaining interactions between the local variables array, operand stack and run time constant pool <a href=\"#class_file_structure\">see the section on Class File Structure below<\/a>.<\/p><h3>Dynamic Linking<\/h3><p>Each frame contains a reference to the runtime constant pool.  The reference points to the constant pool for the class of the method being executed for that frame.  This reference helps to support dynamic linking.<\/p><p>C\/C++ code is typically compiled to an object file then multiple object files are linked together to product a usable artifact such as an executable or dll.  During the linking phase symbolic references in each object file are replaced with an actual memory address relative to the final executable.  In Java this linking phase is done dynamically at runtime.<\/p><p>When a Java class is compiled, all references to variables and methods are stored in the class's constant pool as a symbolic reference. A symbolic reference is a logical reference not a reference that actually points to a physical memory location.  The JVM implementation can choose when to resolve symbolic references, this can happen when the class file is verified, after being loaded, called eager or static resolution, instead this can happen when the symbolic reference is used for the first time called lazy or late resolution. However the JVM has to behave as if the resolution occurred when each reference is first used and throw any resolution errors at this point.  Binding is the process of the field, method or class identified by the symbolic reference being replaced by a direct reference, this only happens once because the symbolic reference is completely replaced.  If the symbolic reference refers to a class that has not yet been resolved then this class will be loaded.  Each direct reference is stored as an offset against the storage structure associated with the runtime location of the variable or method.<\/p><h3>Heap<\/h3><p>The Heap is used to allocate class instances and arrays at runtime. Arrays and objects can never be stored on the stack because a frame is not designed to change in size after it has been created.  The frame only stores references that point to objects or arrays on the heap.  Unlike primitive variables and references in the local variable array (in each frame) objects are always stored on the heap so they are not removed when a method ends.  Instead objects are only removed by the garbage collector.<\/p><p>To support garbage collection the heap is divided into three sections:<\/p><div><ul class=\"list\"><li><strong>Young Generation<\/strong>\n            <div><ul><li>Often split between Eden and Survivor<\/li><\/ul><\/div><\/li>\n            <li><strong>Old Generation<\/strong> (also called Tenured Generation)<\/li>\n            <li><strong>Permanent Generation<\/strong><\/li>\n        <\/ul><\/div><h3>Memory Management<\/h3><p>Objects and Arrays are never explicitly de-allocated instead the garbage collector automatically reclaims them.<\/p><p>Typically this works as follows:<\/p><div><ol class=\"list\"><li>New objects and arrays are created into the young generation<\/li>\n            <li>Minor garbage collection will operate in the young generation.  Objects, that are still alive, will be moved from the eden space to the survivor space.<\/li>\n            <li>Major garbage collection, which typically causes the application threads to pause, will move objects between generations.  Objects, that are still alive, will be moved from the young generation to the old (tenured) generation.<\/li>\n            <li>The permanent generation is collected every time the old generation is collected.  They are both collected when either becomes full.<\/li>\n        <\/ol><\/div><h3>Non-Heap Memory<\/h3><p>Objects that are logically considered as part of the JVM mechanics are not created on the Heap.<\/p><p>The non-heap memory includes:<\/p><div><ul class=\"list\"><li><strong>Permanent Generation<\/strong> that contains\n                <div><ul><li>the method area<\/li>\n                    <li>interned strings<\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Code Cache<\/strong> used for compilation and storage of methods that have been compiled to native code by the JIT compiler<\/li>\n        <\/ul><\/div><h3>Just In Time (JIT) Compilation<\/h3><p>Java byte code is interpreted however this is not as fast as directly executing native code on the JVM\u2019s host CPU.  To improve performance the Oracle Hotspot VM looks for \u201chot\u201d areas of byte code that are executed regularly and compiles these to native code.  The native code is then stored in the code cache in non-heap memory.  In this way the Hotspot VM tries to choose the most appropriate way to trade-off the extra time it takes to compile code verses the extra time it take to execute interpreted code.<\/p><h3>Method Area<\/h3><p>The method area stores per-class information such as:<\/p><div><ul class=\"list\"><li><strong>Classloader Reference<\/strong><\/li>\n            <li><strong>Run Time Constant Pool<\/strong>\n                <div><ul><li>Numeric constants<\/li>\n                    <li>Field references<\/li>\n                    <li>Method References<\/li>\n                    <li>Attributes<\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Field data<\/strong>\n                <div><ul><li>Per field\n                        <div><ul><li>Name<\/li>\n                            <li>Type<\/li>\n                            <li>Modifiers<\/li>\n                            <li>Attributes<\/li>\n                        <\/ul><\/div>\n                    <\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Method data<\/strong>\n                <div><ul><li>Per method\n                        <div><ul><li>Name<\/li>\n                            <li>Return Type<\/li>\n                            <li>Parameter Types (in order)<\/li>\n                            <li>Modifiers<\/li>\n                            <li>Attributes<\/li>\n                        <\/ul><\/div>\n                    <\/li>\n                <\/ul><\/div>\n            <\/li>\n            <li><strong>Method code<\/strong>\n                <div><ul><li>Per method\n                        <div><ul><li>Bytecodes<\/li>\n                            <li>Operand stack size<\/li>\n                            <li>Local variable size<\/li>\n                            <li>Local variable table<\/li>\n                            <li>Exception table\n                                <div><ul><li>Per exception handler\n                                        <div><ul><li>Start point<\/li>\n                                            <li>End point<\/li>\n                                            <li>PC offset for handler code<\/li>\n                                            <li>Constant pool index for exception class being caught<\/li>\n                                        <\/ul><\/div>\n                                    <\/li>\n                                <\/ul><\/div>\n                            <\/li>\n                        <\/ul><\/div>\n                    <\/li>\n                <\/ul><\/div>\n            <\/li>\n        <\/ul><\/div><p>All threads share the same method area, so access to the method area data and the process of dynamic linking must be thread safe.  If two threads attempt to access a field or method on a class that has not yet been loaded it must only be loaded once and both threads must not continue execution until it has been loaded.<\/p><h3>Class File Structure<\/h3><p>A compiled class file consists of the following structure:<\/p><pre class=\"code\">ClassFile {\n    u4\t\t\tmagic;\n    u2\t\t\tminor_version;\n    u2\t\t\tmajor_version;\n    u2\t\t\tconstant_pool_count;\n    cp_info\t\tcontant_pool[constant_pool_count \u2013 1];\n    u2\t\t\taccess_flags;\n    u2\t\t\tthis_class;\n    u2\t\t\tsuper_class;\n    u2\t\t\tinterfaces_count;\n    u2\t\t\tinterfaces[interfaces_count];\n    u2\t\t\tfields_count;\n    field_info\t\tfields[fields_count];\n    u2\t\t\tmethods_count;\n    method_info\t\tmethods[methods_count];\n    u2\t\t\tattributes_count;\n    attribute_info\tattributes[attributes_count];\n}<\/pre><div class=\"columns\"><div class=\"row\"><p>magic,\n minor_version,\n major_version<\/p><p>specifies information about the version of the class and the version of the JDK this class was compiled for.<\/p><\/div><div class=\"row\"><p>constant_pool<\/p><p>similar to a symbol table although it contains more data <a href=\"#constant_pool\">this is described in more detail below.<\/a><\/p><\/div><div class=\"row\"><p>access_flags<\/p><p>provides the list of modifiers for this class.<\/p><\/div><div class=\"row\"><p>this_class<\/p><p>index into the constant_pool providing the fully qualified name of this class i.e. org\/jamesdbloom\/foo\/Bar<\/p><\/div><div class=\"row\"><p>super_class<\/p><p>index into the constant_pool providing a symbolic reference to the super class i.e. java\/lang\/Object<\/p><\/div><div class=\"row\"><p>interfaces<\/p><p>array of indexes into the constant_pool providing a symbolic references to all interfaces that have been implemented.<\/p><\/div><div class=\"row\"><p>fields<\/p><p>array of indexes into the constant_pool giving a complete description of each field.<\/p><\/div><div class=\"row\"><p>methods<\/p><p>array of indexes into the constant_pool giving a complete description of each method signature, if the method is not abstract or native then the bytecode is also present.<\/p><\/div><div class=\"row\"><p>attributes<\/p><p>array of different value that provide additional information about the class including any annotations with RetentionPolicy.CLASS or RetentionPolicy.RUNTIME<\/p><\/div><\/div><p>It is possible to view the byte code in a compiled Java class by using the javap command.<\/p><p>If you compile the following simple class:<\/p><pre class=\"code\">package org.jvminternals;\npublic class SimpleClass {\n    public void sayHello() {\n        System.out.println(\"Hello\");\n    }\n}<\/pre><p>Then you get the following output if you run:<\/p><p>javap -v -p -s -sysinfo -constants classes\/org\/jvminternals\/SimpleClass.class<\/p><pre class=\"code\">public class org.jvminternals.SimpleClass\n  SourceFile: \"SimpleClass.java\"\n  minor version: 0\n  major version: 51\n  flags: ACC_PUBLIC, ACC_SUPER\nConstant pool:\n   #1 = Methodref          #6.#17         \/\/  java\/lang\/Object.\"&lt;init&gt;\":()V\n   #2 = Fieldref           #18.#19        \/\/  java\/lang\/System.out:Ljava\/io\/PrintStream;\n   #3 = String             #20            \/\/  \"Hello\"\n   #4 = Methodref          #21.#22        \/\/  java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n   #5 = Class              #23            \/\/  org\/jvminternals\/SimpleClass\n   #6 = Class              #24            \/\/  java\/lang\/Object\n   #7 = Utf8               &lt;init&gt;\n   #8 = Utf8               ()V\n   #9 = Utf8               Code\n  #10 = Utf8               LineNumberTable\n  #11 = Utf8               LocalVariableTable\n  #12 = Utf8               this\n  #13 = Utf8               Lorg\/jvminternals\/SimpleClass;\n  #14 = Utf8               sayHello\n  #15 = Utf8               SourceFile\n  #16 = Utf8               SimpleClass.java\n  #17 = NameAndType        #7:#8          \/\/  \"&lt;init&gt;\":()V\n  #18 = Class              #25            \/\/  java\/lang\/System\n  #19 = NameAndType        #26:#27        \/\/  out:Ljava\/io\/PrintStream;\n  #20 = Utf8               Hello\n  #21 = Class              #28            \/\/  java\/io\/PrintStream\n  #22 = NameAndType        #29:#30        \/\/  println:(Ljava\/lang\/String;)V\n  #23 = Utf8               org\/jvminternals\/SimpleClass\n  #24 = Utf8               java\/lang\/Object\n  #25 = Utf8               java\/lang\/System\n  #26 = Utf8               out\n  #27 = Utf8               Ljava\/io\/PrintStream;\n  #28 = Utf8               java\/io\/PrintStream\n  #29 = Utf8               println\n  #30 = Utf8               (Ljava\/lang\/String;)V\n{\n  public org.jvminternals.SimpleClass();\n    Signature: ()V\n    flags: ACC_PUBLIC\n    Code:\n      stack=1, locals=1, args_size=1\n        0: aload_0\n        1: invokespecial #1    \/\/ Method java\/lang\/Object.\"&lt;init&gt;\":()V\n        4: return\n      LineNumberTable:\n        line 3: 0\n      LocalVariableTable:\n        Start  Length  Slot  Name   Signature\n          0      5      0    this   Lorg\/jvminternals\/SimpleClass;\n  public void sayHello();\n    Signature: ()V\n    flags: ACC_PUBLIC\n    Code:\n      stack=2, locals=1, args_size=1\n        0: getstatic      #2    \/\/ Field java\/lang\/System.out:Ljava\/io\/PrintStream;\n        3: ldc            #3    \/\/ String \"Hello\"\n        5: invokevirtual  #4    \/\/ Method java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n        8: return\n      LineNumberTable:\n        line 6: 0\n        line 7: 8\n      LocalVariableTable:\n        Start  Length  Slot  Name   Signature\n          0      9      0    this   Lorg\/jvminternals\/SimpleClass;\n}<\/pre><p>This class file shows three main sections the constant pool, the constructor and the sayHello method.<\/p><div><ul class=\"list\"><li><strong>Constant Pool<\/strong> \u2013 this provides the same information that a symbol table typically provides and is <a href=\"#constant_pool\">described in more detail below.<\/a><\/li>\n            <li><strong>Methods<\/strong> \u2013 each containing four areas:\n                <div><ul><li>signature and access flags<\/li>\n                    <li>byte code<\/li>\n                    <li>LineNumberTable \u2013 this provides information to a debugger to indicate which line corresponds to which byte code instruction, for example line 6 in the Java code corresponds to byte code 0 in the sayHello method and line 7 corresponds to byte code 8.<\/li>\n                    <li>LocalVariableTable \u2013 this lists all local variables provided in the frame, in both examples the only local variable is this.<\/li>\n                <\/ul><\/div>\n            <\/li>\n        <\/ul><\/div><p>The following byte code operands are used in this class file<\/p><div class=\"columns\"><div class=\"row\"><p>aload_0<\/p><p>This opcode is one of a group of opcodes with the format aload_&lt;n&gt;.  They all load an object reference into the operand stack.  The &lt;n&gt; refers to the location in the local variable array that is being accessed but can only be 0, 1, 2 or 3.  There are other similar opcodes for loading values that are not an object reference iload_&lt;n&amp;gt, lload_&lt;n&amp;gt, float_&lt;n&amp;gt and dload_&lt;n&amp;gt where i is for int, l is for long, f is for float and d is for double.  Local variables with an index higher than 3 can be loaded using iload, lload, float, dload and aload.  These opcodes all take a single operand that specifies the index of local variable to load.<\/p><\/div><div class=\"row\"><p>ldc<\/p><p>This opcode is used to push a constant from the run time constant pool into the operand stack.<\/p><\/div><div class=\"row\"><p>getstatic<\/p><p>This opcode is used to push a static value from a static field listed in the run time constant pool into the operand stack.<\/p><\/div><div class=\"row\"><p>invokespecial, invokevirtual<\/p><p>These opcodes are in a group of opcodes that invoke methods these are invokedynamic, invokeinterface, invokespecial, invokestatic, invokevirtual.  In this class file invokespecial and invokevirutal are both used the difference between these is that invokevirutal invokes a method based on the class of the object. The invokespecial instruction is used to invoke instance initialization methods as well as private methods and methods of a superclass of the current class.<\/p><\/div><div class=\"row\"><p>return<\/p><p>This opcode is in a group of opcodes ireturn, lreturn, freturn, dreturn, areturn and return.  Each of these opcodes are a typed return statement that returns a different type where i is for int, l is for long, f is for float, d is for double and a is for an object reference.  The opcode with no leading type letter return only returns void.<\/p><\/div><\/div><p>As in any typical byte code the majority of the operands interact with the local variables, operand stack and run time constant pool as follows.<\/p><p>The constructor has two instructions first this is pushed onto the operand stack, next the constructor for the super class is invoked which consumes the value off this and therefore pops it off the operand stack.<\/p><a class=\"image_link\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_SimpleClass.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_SimpleClass.png\" alt=\"local variables, operand stack and run time constant pool changes when invoking a static method in Java Virtual Machine (JVM)\" \/><\/a><p>The sayHello() method is more complex as it has to resolve symbolic references to actual references using the run time constant pool, <a href=\"#dynamic_linking\">as explained in more detail above<\/a>.  The first operand getstatic is used to push a reference to the static field out of the System class on to the operand stack.  The next operand ldc pushes the string \"Hello\" onto the operand stack.  The final operand invokevirtual invokes the println method of System.out which pops \"Hello\" off the operand stack as an argument and creates a new frame for the current thread.<\/p><a class=\"image_link\" href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_sayHello.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/bytecode_explanation_sayHello_smaller.png\" alt=\"local variables, operand stack and run time constant pool changes when calling System.out in Java Virtual Machine (JVM)\" \/><\/a><h3>Classloader<\/h3><p>The JVM starts up by loading an initial class using the bootstrap classloader.  The class is then linked and initialized before public static void main(String[]) is invoked.  The execution of this method will in turn drive the loading, linking and initialization of additional classes and interfaces as required.<\/p><p><strong>Loading<\/strong> is the process of finding the class file that represents the class or interface type with a particular name and reading it into a byte array.  Next the bytes are parsed to confirm they represent a Class object and have the correct major and minor versions.  Any class or interface named as a direct superclass is also loaded.  Once this is completed a class or interface object is created from the binary representation.<\/p><p><strong>Linking<\/strong> is the process of taking a class or interface verifying and preparing the type and its direct superclass and superinterfaces.  Linking consists of three steps verifying, preparing and optionally resolving.<\/p><div><p><strong><em>Verifying<\/em><\/strong> is the process of confirming the class or interface representation is structurally correct and obeys the semantic requirements of the Java programming language and JVM, for example the following checks are performed:<\/p><div><ol><li>consistent and correctly formatted symbol table<\/li>\n                <li>final methods \/ classes not overridden<\/li>\n                <li>methods respect access control keywords<\/li>\n                <li>methods have correct number and type of parameters<\/li>\n                <li>bytecode doesn't manipulate stack incorrectly<\/li>\n                <li>variables are initialized before being read<\/li>\n                <li>variables are a value of the correct type<\/li>\n            <\/ol><\/div><p>Performing these checks during the verifying stages means these checks do not need to be performed at runtime.  Verification during linking slows down class loading however it avoids the need to perform these checks multiple when executing the bytecode.<\/p><p><strong><em>Preparing<\/em><\/strong> involves allocation of memory for static storage and any data structures used by the JVM such as method tables.  Static fields are created and initialized to their default values, however, no initializers or code is executed at this stage as that happens as part of initialization.<\/p><p><strong><em>Resolving<\/em><\/strong> is an optional stage which involves checking symbolic references by loading the referenced classes or interfaces and checking the references are correct.  If this does not take place at this point the resolution of symbolic references can be deferred until just prior to their use by a byte code instruction.<\/p><\/div><p><strong>Initialization<\/strong> of a class or interface consists of executing the class or interface initialization method &lt;clinit&gt;<\/p><a href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/Class_Loading_Linking_Initializing.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/Class_Loading_Linking_Initializing.png\" alt=\"class loading, linking and initialization in the Java Virtual Machine (JVM)\" \/><\/a><p>In the JVM there are multiple classloaders with different roles.  Each classloader delegates to its parent classloader (that loaded it) except the <strong>bootstrap classloader<\/strong> which is the top classloader.<\/p><div><p><strong><em>Bootstrap Classloader<\/em><\/strong> is usually implemented as native code because it is instantiated very early as the JVM is loaded.  The bootstrap classloader is responsible for loading the basic Java APIs, including for example rt.jar.  It only loads classes found on the boot classpath which have a higher level of trust; as a result it skips much of the validation that gets done for normal classes.<\/p><p><strong><em>Extension Classloader<\/em><\/strong> loads classes from standard Java extension APIs such as security extension functions.<\/p><p><strong><em>System Classloader<\/em><\/strong> is the default application classloader, which loads application classes from the classpath.<\/p><p><strong><em>User Defined Classloaders<\/em><\/strong> can alternatively be used to load application classes.  A user defined classloader is used for a number of special reasons including run time reloading of classes or separation between different groups of loaded classes typically required by web servers such as Tomcat.<\/p><\/div><a href=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/class_loader_hierarchy.png\"><img itemprop=\"image\" src=\"http:\/\/blog.jamesdbloom.com\/images_2013_11_17_17_56\/class_loader_hierarchy.png\" alt=\"classloader hierarchy in the Java Virtual Machine (JVM)\" \/><\/a><h3>Faster Class Loading<\/h3><p>A feature called Class Data Sharing (CDS) was introduce in HotSpot JMV from version 5.0.  During the installation process of the JVM the installer loads a set of key JVM classes, such as rt.jar, into a memory-mapped shared archive.  CDS reduces the time it takes to load these classes improving JVM start-up speed and allows these classes to be shared between different instances of the JVM reducing the memory footprint.<\/p><h3>Where Is The Method Area<\/h3><p><a href=\"http:\/\/www.amazon.co.uk\/Virtual-Machine-Specification-Edition-Series\/dp\/0133260445\">The Java Virtual Machine Specification Java SE 7 Edition<\/a> clearly states:  \u201cAlthough the method area is logically part of the heap, simple implementations may choose not to either garbage collect or compact it.\u201d  In contradiction to this jconsole for the Oracle JVM shows the method area (and code cache) as being non-heap.  The <a href=\"http:\/\/openjdk.java.net\/\">OpenJDK<\/a> code shows that the CodeCache is a separate field of the VM to the ObjectHeap.<\/p><h3>Classloader Reference<\/h3><p>All classes that are loaded contain a reference to the classloader that loaded them.  In turn the classloader also contains a reference to all classes that it has loaded.<\/p><h3>Run Time Constant Pool<\/h3><p>The JVM maintains a per-type constant pool, a run time data structure that is similar to a symbol table although it contains more data.  Byte codes in Java require data, often this data is too large to store directly in the byte codes, instead it is stored in the constant pool and the byte code contains a reference to the constant pool. The run time constant pool is used in dynamic linking <a href=\"#dynamic_linking\">as described above<\/a><\/p><p>Several types of data is stored in the constant pool including<\/p><div><ul class=\"list\"><li>numeric literals<\/li>\n            <li>string literals<\/li>\n            <li>class references<\/li>\n            <li>field references<\/li>\n            <li>method references<\/li>\n        <\/ul><\/div><p>For example the following code:<\/p><pre class=\"code\">Object foo = new Object();<\/pre><p>Would be written in byte code as follows:<\/p><pre class=\"code\"> 0: \tnew #2 \t\t    \/\/ Class java\/lang\/Object\n 1:\tdup\n 2:\tinvokespecial #3    \/\/ Method java\/ lang\/Object \"&lt;init&gt;\"( ) V<\/pre><p>The new opcode (operand code) is followed by the #2 operand.  This operand is an index into the constant pool and therefore is referencing the second entry in the constant pool.  The second entry is a class reference, this entry in turn references another entry in the constant pool containing the name of the class as a constant UTF8 string with the value \/\/ Class java\/lang\/Object.  This symbolic link can then be used to lookup the class for java.lang.Object.  The new opcode creates a class instance and initializes its variables.  A reference to the new class instance is then added to the operand stack.  The dup opcode then creates an extra copy of the top reference on the operand stack and adds this to the top of the operand stack.  Finally an instance initialization method is called on line 2 by invokespecial.  This operand also contains a reference to the constant pool.  The initialization method consumes (pops) the top reference off the operand pool as an argument to the method.  At the end there is one reference to the new object that has been both created and initialized.<\/p><p>If you compile the following simple class:<\/p><pre class=\"code\">package org.jvminternals;\npublic class SimpleClass {\n    public void sayHello() {\n        System.out.println(\"Hello\");\n    }\n}<\/pre><p>The constant pool in the generated class file would look like:<\/p><pre class=\"code\">Constant pool:\n   #1 = Methodref          #6.#17         \/\/  java\/lang\/Object.\"&lt;init&gt;\":()V\n   #2 = Fieldref           #18.#19        \/\/  java\/lang\/System.out:Ljava\/io\/PrintStream;\n   #3 = String             #20            \/\/  \"Hello\"\n   #4 = Methodref          #21.#22        \/\/  java\/io\/PrintStream.println:(Ljava\/lang\/String;)V\n   #5 = Class              #23            \/\/  org\/jvminternals\/SimpleClass\n   #6 = Class              #24            \/\/  java\/lang\/Object\n   #7 = Utf8               &lt;init&gt;\n   #8 = Utf8               ()V\n   #9 = Utf8               Code\n  #10 = Utf8               LineNumberTable\n  #11 = Utf8               LocalVariableTable\n  #12 = Utf8               this\n  #13 = Utf8               Lorg\/jvminternals\/SimpleClass;\n  #14 = Utf8               sayHello\n  #15 = Utf8               SourceFile\n  #16 = Utf8               SimpleClass.java\n  #17 = NameAndType        #7:#8          \/\/  \"&lt;init&gt;\":()V\n  #18 = Class              #25            \/\/  java\/lang\/System\n  #19 = NameAndType        #26:#27        \/\/  out:Ljava\/io\/PrintStream;\n  #20 = Utf8               Hello\n  #21 = Class              #28            \/\/  java\/io\/PrintStream\n  #22 = NameAndType        #29:#30        \/\/  println:(Ljava\/lang\/String;)V\n  #23 = Utf8               org\/jvminternals\/SimpleClass\n  #24 = Utf8               java\/lang\/Object\n  #25 = Utf8               java\/lang\/System\n  #26 = Utf8               out\n  #27 = Utf8               Ljava\/io\/PrintStream;\n  #28 = Utf8               java\/io\/PrintStream\n  #29 = Utf8               println\n  #30 = Utf8               (Ljava\/lang\/String;)V\n<\/pre><p>The constant pool contains the following types:<\/p><div class=\"columns\"><div class=\"row\"><p>Integer<\/p><p>A 4 byte int constant<\/p><\/div><div class=\"row\"><p>Long<\/p><p>An 8 byte long constant<\/p><\/div><div class=\"row\"><p>Float<\/p><p>A 4 byte float constant<\/p><\/div><div class=\"row\"><p>Double<\/p><p>A 8 byte double constant<\/p><\/div><div class=\"row\"><p>String<\/p><p>A String constant that points at another Utf8 entry in the constant pool which contains the actual bytes<\/p><\/div><div class=\"row\"><p>Utf8<\/p><p>A stream of bytes representing a Utf8 encoded sequence of characters<\/p><\/div><div class=\"row\"><p>Class<\/p><p>A Class constant that points at another Utf8 entry in the constant pool which contains the fully qualified class name in the internal JVM format (this is used by the <a href=\"#dynamic_linking\">dynamic linking process<\/a>)<\/p><\/div><div class=\"row\"><p>NameAndType<\/p><p>A colon separated pair of values each pointing at other entries in the constant pool.  The first value (before the colon) points at a Utf8 string entry that is the method or field name.  The second value points at a Utf8 entry that represents the type, in the case of a field this is the fully qualified class name, in the case of a method this is a list of fully qualified class names one per parameter.<\/p><\/div><div class=\"row\"><p>Fieldref,\nMethodref,\nInterfaceMethodref<\/p><p>A dot separated pair of values each pointing at other entries in the constant pool.  The first value (before the dot) points at a Class entry.  The second value points at a NameAndType entry.<\/p><\/div><\/div><h3>Exception Table<\/h3><p>The exception table stores per-exception handler information such as:<\/p><div><ul class=\"list\"><li>Start point<\/li>\n            <li>End point<\/li>\n            <li>PC offset for handler code<\/li>\n            <li>Constant pool index for exception class being caught<\/li>\n        <\/ul><\/div><p>If a method has defined a try-catch or a try-finally exception handler then an Exception Table will be created.  This contains information for each exception handler or finally block including the range over which the handler applies, what type of exception is being handled and where the handler code is.<\/p><p>When an exception is thrown the JVM looks for a matching handler in the current method, if none is found the method ends abruptly popping the current stack frame and the exception is re-thrown in the calling method (the new current frame).  If no exception handler is found before all frames have been popped then the thread is terminated.   This can also cause the JVM itself to terminate if the exception is thrown in the last non-daemon thread, for example if the thread is the main thread.<\/p><p>Finally exception handlers match all types of exceptions and so always execute whenever an exception is thrown.  In the case when no exception is thrown a finally block is still executed at the end of a method, this is achieved by jumping to the finally handler code immediately before the return statement is executed.<\/p><h3>Symbol Table<\/h3><p>In addition to per-type run-time constant pools the Hotspot JVM has a symbol table held in the permanent generation.  The symbol table is a Hashtable mapping symbol pointers to symbols (i.e. Hashtable&lt;Symbol*, Symbol&gt;) and includes a pointer to all symbols including those held in run time constant pools in each class.<\/p><p>Reference counting is used to control when a symbol is removed from the symbol table.  For example when a class is unloaded the reference count of all symbols held in its run time constant pool are decremented.  When the reference count of a symbol in the the symbol table goes to zero then the symbol table knows that symbol is not being referenced anymore and the symbol is unloaded from the symbol table. For both the symbol table and the string table (see below) all entries are held in a canonicalized form to improve efficiency and ensure each entry only appears once.<\/p><h3>Interned Strings (String Table)<\/h3><p>The Java Language Specification requires that identical string literals, that contain the same sequence of Unicode code points, must refer to the same instance of String.  In addition if String.intern() is called on an instance of String a reference must be returned that would be identical to the reference return if the string was a literal. The following therefore holds true:<\/p><pre class=\"code\">(\"j\" + \"v\" + \"m\").intern() == \"jvm\"<\/pre><p>In the Hotspot JVM interned string are held in the string table, which is a Hashtable mapping object pointers to symbols (i.e. Hashtable&lt;oop, Symbol&gt;), and is held in the permanent generation.  For both the symbol table (see above) and the string table all entries are held in a canonicalized form to improve efficiency and ensure each entry only appears once.<\/p><p>String literals are automatically interned by the compiler and added into the symbol table when the class is loaded.  In addition instances of the String class can be explicitly interned by calling String.intern().  When String.intern() is called, if the symbol table already contains the string then a reference to this is returned, if not the string is added to the string table and its reference is returned.<\/p>","created_at":"2018-11-07T16:23:25+0000","updated_at":"2018-11-07T16:23:25+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":26,"domain_name":"blog.jamesdbloom.com","preview_picture":"http:\/\/blog.jamesdbloom.com\/apple-touch-icon_2013_11_17_17_56.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":140,"title":"Linux Fu: The Great Power of Make","url":"https:\/\/hackaday.com\/2018\/06\/22\/linux-fu-the-great-power-of-make\/","content":"<p>Over the years, Linux (well, the operating system that is commonly known as Linux which is the Linux kernel and the GNU tools) has become much more complicated than its Unix roots. That\u2019s inevitable, of course. However, it means old-timers get to slowly grow into new features while new people have to learn all in one gulp. A good example of this is how software is typically built on a Linux system. Fundamentally, most projects use <code>make<\/code> \u2014 a program that tries to be smart about running compiles. This was especially important when your 100 MHz CPU connected to a very slow disk drive would take a day to build a significant piece of software. On the face of it, make is pretty simple. But today, looking at a typical makefile will give you a headache, and many projects use an abstraction over make that further obscures things.<\/p>\n<p>In this article, I want to show you how simple a makefile can be. If you can build a simple makefile, you\u2019ll find they have more uses than you might think. I\u2019m going to focus on C, but only because that\u2019s sort of the least common denominator. A make file can build just about anything from a shell prompt.<\/p>\n<p>\n<\/p><h2>No IDE?<\/h2>\n<p>You may scoff and say that you don\u2019t need a makefile. Your IDE builds your software for you. That may be true, but a lot of IDEs actually use makefiles under the covers. Even IDEs that do not often have a way to export a makefile. This allows you to more easily write build scripts or integrate with continuous release tools which usually expect a makefile.<\/p>\n<p>One exception to this is Java. Java has its own set of build tools that people use and so make isn\u2019t as common. However, it is certainly possible to use it along with command line tools like <code>javac<\/code> or <code>gcj<\/code>.<\/p>\n<h2>The Simplest Makefile<\/h2>\n<p><a href=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png\" target=\"_blank\"><img data-attachment-id=\"309834\" data-permalink=\"https:\/\/hackaday.com\/2018\/06\/22\/linux-fu-the-great-power-of-make\/tux1\/\" data-orig-file=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png\" data-orig-size=\"640,635\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"tux1\" data-image-description=\"\" data-medium-file=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png?w=400\" data-large-file=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png?w=630\" class=\"alignleft wp-image-309834 size-thumbnail\" src=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png?w=250&amp;h=248\" alt=\"\" width=\"250\" height=\"248\" srcset=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png?w=250&amp;h=248 250w, https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png?w=500&amp;h=496 500w, https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux1.png?w=400&amp;h=397 400w\" \/><\/a>Here\u2019s a simple makefile:<\/p>\n<pre>hello: hello.c<\/pre>\n<p>That\u2019s it. Really. This is a rule that says there is a file called hello and it depends on the file hello.c. If hello.c is newer than hello, then build hello. Wait, what? How does it know what to do? Usually, you\u2019ll add instructions about how to build a target (hello) from its dependencies. However, there are default rules.<\/p>\n<p>If you create a file called Makefile along with <a href=\"https:\/\/gist.github.com\/szczys\/3af48cdb165094945badc92de4c9d897\" target=\"_blank\">hello.c<\/a> in a directory and then issue the <code>make<\/code> command in that directory you\u2019ll see the command run will be:<\/p>\n<pre>cc hello.c -o hello<\/pre>\n<p>That\u2019s perfectly fine since most distributions will have cc pointing to the default C compiler. If you run make again, you\u2019ll see it doesn\u2019t build anything. Instead, it will report something like this:<\/p>\n<pre>make: 'hello' is up to date.<\/pre>\n<p>To get it to build again, you\u2019ll need to make a change to hello.c, use the touch command on hello.c, or delete hello. By the way, the <code>-n<\/code> command line option will tell make to tell you what it will do without actually doing it. That\u2019s handy when you are trying to figure out what a particular makefile is doing.<\/p>\n<p>If you like, you can customize the defaults using variables. The variables can come from the command line, the environment, or be set in the makefile itself. For example:<\/p>\n<pre>CC=gcc&#13;\nCFLAGS=-g&#13;\nhello : hello.c<\/pre>\n<p>Now you\u2019ll get a command like:<\/p>\n<pre>gcc -g hello.c -o hello<\/pre>\n<p>In a complex makefile you might want to add options to ones you already have and you can do that too (the number sign, #, is a comment):<\/p>\n<pre>CC=gcc&#13;\nCFLAGS=-g&#13;\n# Comment next line to turn off optimization&#13;\nCFLAGS+=-O&#13;\nhello : hello.c<\/pre>\n<p>In fact, the implicit rule being used is:<\/p>\n<pre>$(CC) $(CFLAGS) $(CPPFLAGS) hello.c -o hello<\/pre>\n<p>Note the variables always use the <code>$()<\/code> syntax by convention. In reality, if you have a single character variable name you could omit it (e.g., <code>$X<\/code>) but that may not be portable, so the best thing to do is always use the parenthesis.<\/p>\n<h2>Custom Rules<\/h2>\n<p>Maybe you don\u2019t want to use the default rule. That\u2019s no problem. The make program is a little bit of stickler for file format, though. If you start subsequent lines with a tab character (not just a few spaces; a real tab) then that line (along with others) will run like a script instead of the default rules. So while this isn\u2019t very elegant, here\u2019s a perfectly fine way to write our makefile:<\/p>\n<pre>hello : hello.c&#13;\n\u00a0 \u00a0 gcc -g -O hello.c<\/pre>\n<p>In fact, you ought to use variables to make your rules more flexible just like the default ones do. You can also use wildcards. Consider this:<\/p>\n<pre>% : %.c&#13;\n    $(CC) $(CPPFLAGS) $(CFLAGS) $&lt; -o $@<\/pre>\n<p>This is more or less the same as the default rule. The percent signs match anything and the <code>$&lt;<\/code> gets the name of the first (and in this case, only) prerequisite which is hello.c. The <code>$@<\/code> variable gives you the name of the target (hello, for this example). There are many more variables available, but these will get you started.<\/p>\n<p>You can have multiple script lines (processed by the default system shell, although you can change that) as long as they all start with a tab. You can also add more prerequisites. For example:<\/p>\n<pre>hello : hello.c hello.h mylocallib.h<\/pre>\n<p>This gets tedious to maintain, though. For C and C++, most compilers (including gcc) have a way to create .d files that can automatically tell make what files an object depends on. That\u2019s beyond the scope of this post, but look up the <code>-MMD<\/code> option for gcc, if you want to know more.<\/p>\n<h2>I Object<\/h2>\n<p>Normally, in a significant project, you won\u2019t just compile C files (or whatever you have). You\u2019ll compile the source files to object files and then link the object files at the end. The default rules understand this and, of course, you can write your own:<\/p>\n<pre>hello : hello.o mylib.o&#13;\nhello.o : hello.c hello.h mylib.h&#13;\nmylib.o : mylib.c mylib.h<\/pre>\n<p>Thanks to the default rules, that\u2019s all you need. The make program is smart enough to see that it needs hello.o, it will go find the rule for hello.o and so on. Of course, you can add one or more script lines after any of these if you want to control what happens for the build process.<\/p>\n<p>By default, make only tries to build the first thing it finds. Sometimes you\u2019ll see a fake target as the first thing like this:<\/p>\n<pre>all : hello libtest config-editor<\/pre>\n<p>Presumably, then, you\u2019ll have rules for those three programs elsewhere and make will build all three. This works as long as you never have a file name \u201call\u201d in the working directory. To prevent that being a problem, you can declare that target as phony by using this statement in the makefile:<\/p>\n<pre>.PHONY: all<\/pre>\n<p>You can attach actions to phony targets, too. For example, you\u2019ll often see something like:<\/p>\n<pre>.PHONY: clean&#13;\nclean:\u00a0&#13;\n\u00a0 \u00a0 \u00a0rm *.o<\/pre>\n<p>Then you can issue <code>make clean<\/code> from the command line to remove all the object files. You probably won\u2019t make that the first target or add it to something like all. Another common thing to do is create a phony target that can burn your code to a microcontroller so you can issue a command like \u201cmake program\u201d to download code to the chip.<\/p>\n<h2>Stupid Make Tricks<\/h2>\n<p><a href=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png\" target=\"_blank\"><img data-attachment-id=\"309835\" data-permalink=\"https:\/\/hackaday.com\/2018\/06\/22\/linux-fu-the-great-power-of-make\/tux2\/\" data-orig-file=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png\" data-orig-size=\"958,1245\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"tux2\" data-image-description=\"\" data-medium-file=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png?w=308\" data-large-file=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png?w=481\" class=\"alignright size-thumbnail wp-image-309835\" src=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png?w=192&amp;h=250\" alt=\"\" width=\"192\" height=\"250\" srcset=\"https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png?w=192&amp;h=250 192w, https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png?w=384&amp;h=500 384w, https:\/\/hackadaycom.files.wordpress.com\/2018\/05\/tux2.png?w=308&amp;h=400 308w\" \/><\/a>Usually, you use make to build software. But like any other Unix\/Linux tool, you\u2019ll find people using it for all sorts of things. For example, it would be easy to build a makefile that uses <a href=\"https:\/\/hackaday.com\/2017\/05\/23\/stupid-git-tricks\/\">pandoc<\/a> to convert a document to multiple formats any time it changes.<\/p>\n<p>The key is to realize that make is going to build a dependency tree, figure out what pieces need attention, and run the scripts for those pieces. The scripts don\u2019t have to be build steps. They can copy files (even remotely using something like scp), delete files, or do anything else a Linux shell script could do.<\/p>\n<h2>But Wait, There\u2019s More<\/h2>\n<p>There\u2019s a lot more to learn about make, but with what you already know you can do a surprising amount. You can read <a href=\"https:\/\/www.gnu.org\/software\/make\/manual\/make.html\" target=\"_blank\">the manual<\/a>, of course, to learn more. If you look at a significant makefile like the one for <a href=\"https:\/\/hackaday.com\/2015\/10\/01\/arduino-development-theres-a-makefile-for-that\/\">Arduino projects<\/a>, you\u2019ll see there\u2019s a lot there we didn\u2019t talk about. But you still should be able to pick through a lot of it.<\/p>\n<p>Next time, I\u2019ll talk a little about schemes that take a different input format and generate a makefile. There are a few reasons people want to do this. One is the abstraction software might determine dependencies or make other changes to the generated makefile. In addition, most of these tools can generate build scripts for different platforms. As an example, if you\u2019ve seen<code>automake<\/code>or<code>cmake<\/code>, these are the kinds of tools I\u2019m talking about.<\/p>\n<p>However, you don\u2019t need anything that fancy for a lot of small projects like we often do on a Raspberry Pi or other embedded system. Basic make functionality will take you a long way. Especially for a project targeting a single kind of computer with a small number of files and dependencies.<\/p>","created_at":"2018-11-07T16:23:23+0000","updated_at":"2018-11-07T16:23:23+0000","published_at":"2018-06-22T14:01:21+0000","published_by":["Al Williams"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"hackaday.com","preview_picture":"https:\/\/hackadaycom.files.wordpress.com\/2018\/06\/linuxfu.jpg?w=1200","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":141,"title":"A visual introduction to machine learning, Part II","url":"http:\/\/www.r2d3.us\/visual-intro-to-machine-learning-part-2\/","content":"<h2>Flukes are normal<\/h2><p>This is not an isolated incident. For example, we could grow additional trees from random subsets of the training data.<\/p><p>From the training set of 250 homes\u2026<\/p><div id=\"mini-tree-sampling\"><hr class=\"whitespace\" \/><p>\u2026 let's draw four <strong>random samples<\/strong> of 200 homes each and grow trees based on them.<\/p><hr class=\"whitespace\" \/><\/div><p>The resulting trees all look reasonably different and also have single-home leaf nodes.<\/p><p>These seemingly-esoteric homes that may result in single-data-point leaf nodes are actually a normal part of any data set. They are an outcome of the <em>method<\/em> for fitting the model.  <\/p><p>\n          When the <em>minimum node size<\/em> parameter is one, the tree grows until every branch has a homogeneous leaf node.<\/p><p>For a given data set, growing the tree on a different set of homes changes <em>what<\/em> the branches overfit to, but overfitting still occurs. \n          \n             \n              The data can also be a source of error. An example of <em>data<\/em> resulting in a biased model is <a href=\"https:\/\/dism.ssri.duke.edu\/survey-help\/tipsheets\/tipsheet-nonresponse-error\" target=\"_blank\">non-response<\/a> in polling. \n            \n          \n          Models that overfit are unstable and sensitive to small changes in the training data and thus high variance.\n        <\/p><hr class=\"whitespace\" \/><div id=\"increase-minimal-node-size\"><h2>The Bias-Variance Tradeoff<\/h2><p>One way to address errors from overfitting is to impose limits on how a tree grows by changing the minimum-node-size threshold.<\/p><hr class=\"whitespace\" \/><p>As the minimum-node-size threshold increases, there are fewer splits. The trees get less <em>bushy<\/em>. <\/p><div id=\"pause-minimal-node-size\"><hr class=\"whitespace\" \/><p>The accuracy of the each tree improves as errors due to variance decrease.<\/p><hr class=\"whitespace\" \/><\/div><hr class=\"whitespace\" \/><p>As the minimum-node-size threshold continues to increase, the accuracy begins to deteriorate from error due to bias.<\/p><hr class=\"whitespace\" \/><p>Until you get back to a stump.<\/p><\/div><hr class=\"whitespace\" \/><p> A model that is overly-simplistic is just as problematic as one that is overly-scrupulous. Errors due to bias and those due to variance are distinct. \n          Understanding the tradeoff between bias and variance (and how different model types let you balance the two) is foundational to modeling well.<\/p><hr class=\"whitespace\" \/>","created_at":"2018-11-07T16:23:23+0000","updated_at":"2018-11-07T16:23:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.r2d3.us","preview_picture":"http:\/\/www.r2d3.us\/static\/pages\/decision-trees-part2-v2\/part2-preview-en.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":139,"title":"H3: Uber\u2019s Hexagonal Hierarchical Spatial Index","url":"https:\/\/eng.uber.com\/h3\/","content":"<div class=\"td-post-featured-image\"><a href=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3.png\" data-caption=\"\"><img width=\"1068\" height=\"534\" class=\"entry-thumb td-modal-image\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3.png\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3.png 2160w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3-300x150.png 300w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3-768x384.png 768w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3-1024x512.png 1024w\" alt=\"\" title=\"Twitter H3\" \/><\/a><\/div><p><em>By Isaac Brodsky<\/em><\/p><p>Grid systems are critical to analyzing large spatial data sets, partitioning areas of the Earth into identifiable grid cells. <\/p><p>With this in mind, Uber developed <a href=\"https:\/\/github.com\/uber\/h3\" target=\"_blank\" rel=\"noopener\">H3<\/a>, our grid system for efficiently optimizing ride pricing and dispatch, for visualizing and exploring spatial data. H3 enables us to analyze geographic information to set dynamic prices and make other decisions on a city-wide level. We use H3 as the grid system for analysis and optimization throughout our marketplaces. H3 was designed for this purpose, and led us to make some choices such as using hexagonal, hierarchical indexes. <\/p><p>Earlier this year, we <a href=\"https:\/\/uber.github.io\/h3\/\" target=\"_blank\" rel=\"noopener\">open sourced H3<\/a> on Github, giving others access to this powerful solution, and last week, we open sourced our <a href=\"https:\/\/github.com\/uber\/h3-js\" target=\"_blank\" rel=\"noopener\">H3 JavaScript bindings<\/a>. In this article, we discuss why we use a grid system, some of the unique properties of H3, and how you can get started using H3.<\/p><figure id=\"attachment_3461\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image12.png\" target=\"_blank\" rel=\"noopener\"><img class=\"wp-image-3461\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image12.png\" alt=\"\" width=\"400\" height=\"370\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image12.png 577w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image12-300x278.png 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Figure 1. H3 enables users to partition the globe into hexagons for more accurate analysis.<\/figcaption><\/figure><h3>Analysis using grids<\/h3><p>Every day, millions of events occur in the Uber marketplace. Every minute, riders request rides, driver-partners start trips, and hungry users request food, among other actions on the platform. Each event happens at a specific location, for example, a rider requests a ride from home, and a driver accepts that request in their car just a few miles away.<\/p><p>These events empower Uber to better understand and optimize the marketplace for users across our services. For instance, these events might tell us that there is more demand than supply in a certain part of a city and adjust pricing in response, or inform the platform that there are two ride requests within close proximity to a specific driver on uberPOOL.<\/p><p>Deriving information and insights from data in the Uber marketplace requires analyzing data across an entire city. Because cities are geographically diverse, this analysis needs to happen at a fine granularity. Analysis at the finest granularity, the exact location where an event happens, is very difficult and expensive. Analysis on areas, such as neighborhoods within a city, is much more practical.<\/p><table style=\"width: 564px;\"><tbody><tr><td><i>Figure 2. The maps above depict the process of bucketing points with H3: cars in a city; cars in hexagons; and hexagons shaded by number of cars.<\/i><\/td>\n<\/tr><\/tbody><\/table><p>We use a grid system to bucket events into hexagonal areas, in other words, cells. Data points are bucketed into hexagons and can be written using the hexagonally bucketed data. For example, we calculate surge pricing by measuring supply and demand in hexagons in each city that we serve. These hexagons form the basis for our analysis of the Uber marketplace.<\/p><p>Hexagons were an important choice because people in a city are often in motion, and hexagons minimize the <a href=\"https:\/\/www.semanticscholar.org\/paper\/Quantization-Information-Theory%2C-IEEE-Transactions-Gray-Neuhoff\/cf36ef494d96dfcf8b3c1af3cfc67147f6f8adee\" target=\"_blank\" rel=\"noopener\">quantization error<\/a> introduced when users move through a city. Hexagons also allow us to approximate radiuses easily, such as in <a href=\"http:\/\/eng.uber.com\/elk\/\" target=\"_blank\" rel=\"noopener\">this example using Elasticsearch<\/a>.<\/p><figure id=\"attachment_3470\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image21.png\" target=\"_blank\" rel=\"noopener\"><img class=\"wp-image-3470\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image21.png\" alt=\"\" width=\"600\" height=\"313\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image21.png 1643w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image21-300x156.png 300w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image21-768x400.png 768w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image21-1024x534.png 1024w\" \/><\/a><figcaption class=\"wp-caption-text\">Figure 3. Zip code map of New York City (Manhattan).<\/figcaption><\/figure><p>There are other choices we could use for bucketing events into areas, for instance polygonal zones around areas. These could be postal code areas, but such areas have unusual shapes and sizes which are not helpful for analysis, and are subject to change for reasons <a href=\"https:\/\/fas.org\/sgp\/crs\/misc\/RL33488.pdf\" target=\"_blank\" rel=\"noopener\">entirely unrelated<\/a> to what we would use them for. Zones could also be drawn by Uber operations teams based on their knowledge of the city, but such zones require frequent updating as cities change and often define the edges of areas arbitrarily.<\/p><figure id=\"attachment_3462\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image13-e1529950174302.png\" target=\"_blank\" rel=\"noopener\"><img class=\"wp-image-3462\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image13-e1529950174302.png\" alt=\"\" width=\"600\" height=\"324\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image13-e1529950174302.png 949w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image13-e1529950174302-300x162.png 300w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image13-e1529950174302-768x415.png 768w\" \/><\/a><figcaption class=\"wp-caption-text\">Figure 4. Randomly generated hexagonal clusters cover the city of San Francisco.<\/figcaption><\/figure><p>Grid systems can have comparable shapes and sizes across the cities that Uber operates in and are not subject to arbitrary changes. While grid systems do not align to streets and neighborhoods in cities, they can be used to efficiently represent neighborhoods by clustering grid cells. Clustering can be done using objective functions, producing shapes much more useful for analysis. Determining membership of a cluster is as efficient as a set lookup operation.<\/p><h3>H3<\/h3><p>We decided to create H3 to combine the benefits of a hexagonal global grid system with a hierarchical indexing system.<\/p><p>A global grid system usually requires at least two things: a map projection and a grid laid on top of the map. A map projection is needed for going from a three-dimensional location on Earth to a two dimensional point on a map. A grid is then overlaid on the map, forming a global grid system.<\/p><p>This process can be accomplished in innumerable ways by combining different map projections and grids, for example, the widely recognized <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mercator_projection\" target=\"_blank\" rel=\"noopener\">Mercator projection<\/a> and a square grid. While this simple method would work, it has a number of drawbacks. To start, the Mercator projection has significant size distortion, so some cells will have vastly different areas. Square grids also have drawbacks, requiring multiple sets of coefficients when used for analysis. This disadvantage is a result of squares having two different types of neighbors, one type with which they share an edge (in the four cardinal directions) and another type with which they share a vertex (in four diagonal directions).<\/p><table style=\"width: 100%;\"><tbody><tr><td><i>Figure 5. We chose to use gnomonic projections centered on icosahedron faces (left) for H3\u2019s map projection, projecting Earth as a spherical icosahedron (right).<\/i><\/td>\n<\/tr><\/tbody><\/table><p>For map projection, we chose to use gnomonic projections centered on <a href=\"https:\/\/en.wikipedia.org\/wiki\/Icosahedron\" target=\"_blank\" rel=\"noopener\">icosahedron<\/a> faces. This projects from Earth as a sphere to an icosahedron, a twenty-sided platonic solid. An icosahedron-based map projection results in twenty separate two-dimensional planes rather than a single plane. The icosahedron can be <a href=\"https:\/\/books.google.com\/books?id=WlEEAAAAMBAJ&amp;lpg=PA2&amp;dq=March%201%2C%201943%20Life&amp;pg=PA41#v=onepage&amp;q=dymaxion&amp;f=false\" target=\"_blank\" rel=\"noopener\">unfolded in many ways<\/a>, producing a two-dimensional map each time. H3, however, does not unfold the icosahedron to build its grid system, and instead lays its grid out on the icosahedron faces themselves, forming a <a href=\"http:\/\/www.discreteglobalgrids.org\/wp-content\/uploads\/sites\/33\/2016\/01\/gdggs03.pdf\" target=\"_blank\" rel=\"noopener\">geodesic discrete global grid system<\/a>.<\/p><table style=\"width: 600px;\"><tbody><tr><td><i>Figure 6. Distances from a triangle to its neighbors (left), a square to its neighbors (center), and a hexagon to its neighbors (right).<\/i><\/td>\n<\/tr><\/tbody><\/table><p>Using a hexagon as the cell shape is critical for H3. As depicted in Figure 6, hexagons have only one distance between a hexagon centerpoint and its neighbors\u2019, compared to two distances for squares or three distances for triangles. This property greatly simplifies performing analysis and smoothing over gradients.<\/p><figure id=\"attachment_3452\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image3-1.png\" target=\"_blank\" rel=\"noopener\"><img class=\"wp-image-3452\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image3-1.png\" alt=\"\" width=\"400\" height=\"370\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image3-1.png 578w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image3-1-300x277.png 300w\" \/><\/a><figcaption class=\"wp-caption-text\">Figure 7. H3 used to create a grid on a single icosahedron face.<\/figcaption><\/figure><p>The H3 grid is constructed by laying out 122 base cells over the Earth, with ten cells per face. Some cells are contained by more than one face. Since it is not possible to tile the icosahedron with only hexagons, we chose to introduce twelve pentagons, one at each of the icosahedron vertices. These vertices were positioned using the spherical icosahedron orientation by <a href=\"https:\/\/exhibits.stanford.edu\/bucky\/about\/about-r-buckminster-fuller\" target=\"_blank\" rel=\"noopener\">R. Buckminster Fuller<\/a>, which places all the vertices in the water. This helps avoid pentagons surfacing in our work.<\/p><table style=\"width: 683px;\"><tbody><tr><td><i>Figure 8. H3 enables the user to subdivide areas into smaller and smaller hexagons. <\/i><\/td>\n<\/tr><\/tbody><\/table><p>H3 supports sixteen resolutions. Each finer resolution has cells with one seventh the area of the coarser resolution. Hexagons cannot be perfectly subdivided into seven hexagons, so the finer cells are only approximately contained within a parent cell.<\/p><p>The identifiers for these child cells can be easily truncated to find their ancestor cell at a coarser resolution, enabling efficient indexing. Because the children cells are only approximately contained, the truncation process produces a fixed amount of shape distortion. This distortion is only present when performing truncation of a cell identifier; when indexing locations at a specific resolution, the cell boundaries are exact.<\/p><h3>Getting started with H3<\/h3><p>The <a href=\"https:\/\/uber.github.io\/h3\/#\/\" target=\"_blank\" rel=\"noopener\">H3 indexing system<\/a> is open source and available on GitHub. The <a href=\"https:\/\/github.com\/uber\/h3\" target=\"_blank\" rel=\"noopener\">H3 library<\/a> itself is written in C, and bindings are available for a number of languages. Using bindings is the recommended way to start using H3. Uber has published bindings for <a href=\"https:\/\/github.com\/uber\/h3-java\" target=\"_blank\" rel=\"noopener\">Java<\/a> and <a href=\"https:\/\/github.com\/uber\/h3-js\" target=\"_blank\" rel=\"noopener\">JavaScript<\/a>, and the community has contributed bindings for <a href=\"https:\/\/uber.github.io\/h3\/#\/documentation\/community\/bindings\" target=\"_blank\" rel=\"noopener\">more languages<\/a>. Bindings are coming soon for Python and Go.<\/p><table style=\"width: 600px;\"><tbody><tr><td><i>Figure 9. The point (blue, left) and its containing hexagon, the centroid (lighter blue, middle) of its containing hexagon, and both the point and centroid of the containing hexagon (right) depict a possible difference between the two.<\/i><\/td>\n<\/tr><\/tbody><\/table><p><i>Code: <\/i><a href=\"https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/index.c\" target=\"_blank\" rel=\"noopener\"><i>https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/index.c<\/i><\/a><\/p><p>The basic functions of the H3 library are for indexing locations, which transforms latitude and longitude pairs to a <a href=\"https:\/\/uber.github.io\/h3\/#\/documentation\/core-library\/h3-index-representations\" target=\"_blank\" rel=\"noopener\">64-bit H3 index<\/a>, identifying a grid cell. The function <a href=\"https:\/\/uber.github.io\/h3\/#\/documentation\/api-reference\/indexing\" target=\"_blank\" rel=\"noopener\"><i>geoToH3<\/i><\/a> takes a latitude, longitude, and resolution (between 0 and 15, with 0 being coarsest and 15 being finest), and returns an index. <i>h3ToGeo<\/i> and <i>h3ToGeoBoundary<\/i> are the inverse of this function, providing the center coordinates and outline of the grid cell specified by the H3 index, respectively.<\/p><p>Once you have data indexed using H3, the H3 API has functions for working with indexes.<\/p><table style=\"width: 598px;\"><tbody><tr><td><i>Figure 10. Hexagons neighboring an index, with distance of 0 (left; just the original index), 1 (middle; with neighbors of the original index), and 2 (right; with neighbors of neighbors).<\/i><\/td>\n<\/tr><\/tbody><\/table><p><i>Code: <\/i><a href=\"https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/neighbors.c\" target=\"_blank\" rel=\"noopener\"><i>https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/neighbors.c<\/i><\/a><\/p><p>Neighboring hexagons have the useful property of approximating circles using the grid system. The <a href=\"https:\/\/uber.github.io\/h3\/#\/documentation\/api-reference\/neighbors\" target=\"_blank\" rel=\"noopener\"><i>kRing<\/i><\/a> function provides grid cells within grid distance <i>k <\/i>of an origin index.<\/p><table style=\"width: 600px;\"><tbody><tr><td><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786.png\" target=\"_blank\" rel=\"noopener\"><img class=\"aligncenter wp-image-3481 size-medium\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786-294x300.png\" alt=\"\" width=\"294\" height=\"300\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786-294x300.png 294w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786-768x784.png 768w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786-32x32.png 32w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786-50x50.png 50w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786-64x64.png 64w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image23-1-e1530032642786.png 803w\" \/><\/a><\/td>\n<td><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image11-1.png\" target=\"_blank\" rel=\"noopener\"><img class=\"aligncenter wp-image-3460 size-medium\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image11-1-286x300.png\" alt=\"\" width=\"286\" height=\"300\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image11-1-286x300.png 286w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image11-1-768x805.png 768w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image11-1.png 787w\" \/><\/a><\/td>\n<\/tr><\/tbody><\/table><table style=\"width: 600px;\"><tbody><tr><td><i>Figure 11. Dense hexagons representing California stand in stark contrast to compact hexagons representing the state, requiring many fewer hexagons to represent the same area.<\/i><\/td>\n<\/tr><\/tbody><\/table><p><i>Code: <\/i><a href=\"https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/compact.c\" target=\"_blank\" rel=\"noopener\"><i>https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/compact.c<\/i><\/a><\/p><p>The hierarchical nature of H3 allows for efficiently truncating the precision (or resolution) of an index and recovering the original indexes. The <a href=\"https:\/\/uber.github.io\/h3\/#\/documentation\/api-reference\/hierarchy\" target=\"_blank\" rel=\"noopener\">uncompact and compact<\/a> representations of a set of hexagons are shown above. The uncompact representation has 10,633 hexagons at resolution 6, but the compact representation has 901 hexagons at resolutions up to 6. In both cases, a hexagon index is a 64-bit integer.<\/p><p>The precision of a single index can also be efficiently truncated as a bitwise operation, or expanded to a higher precision set of indexes.<\/p><figure id=\"attachment_3466\" class=\"wp-caption aligncenter\"><a href=\"http:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image17.png\" target=\"_blank\" rel=\"noopener\"><img class=\"wp-image-3466\" src=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image17.png\" alt=\"\" width=\"500\" height=\"351\" srcset=\"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image17.png 983w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image17-300x211.png 300w, https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/image17-768x539.png 768w\" \/><\/a><figcaption class=\"wp-caption-text\">Figure 12. H3 can represent the movement from one cell to another, here represented by drawing an arrow from the origin to destination cell.<\/figcaption><\/figure><p><i>Code: <\/i><a href=\"https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/edge.c\" target=\"_blank\" rel=\"noopener\"><i>https:\/\/github.com\/uber\/h3\/blob\/master\/examples\/edge.c<\/i><\/a><\/p><p>H3 has functions for <a href=\"https:\/\/uber.github.io\/h3\/#\/documentation\/api-reference\/unidirectional-edges\" target=\"_blank\" rel=\"noopener\">directed edges of grid cells<\/a> which can represent the movement from one cell to another. Directed edges can be stored as 64-bit integers, and can be obtained from two neighboring cells, or by finding all edges of a cell. When needed, an edge can be converted back to the origin or destination indexes.<\/p><h3>Moving forward<\/h3><p>H3 is used throughout Uber to support quantitative analysis of our marketplace, and now that it\u2019s open sourced, you too can hexagonify the world! We look forward to you joining the H3 community by following our <a href=\"https:\/\/github.com\/uber\/h3\" target=\"_blank\" rel=\"noopener\">repository on Github<\/a> and tweeting with the <a href=\"https:\/\/twitter.com\/hashtag\/uberh3\" target=\"_blank\" rel=\"noopener\">#uberH3<\/a> hashtag.<\/p><p><em>Maps generated using <a href=\"https:\/\/www.mapbox.com\/help\/define-mapbox-gl\/\" target=\"_blank\" rel=\"noopener\">MapboxGL<\/a>.<\/em><\/p><p><em>Learn more information about <a href=\"https:\/\/www.uber.com\/legal\/other\/data-providers\/\">our mapping data providers<\/a>.<\/em><\/p><p><i>Major contributors to the H3 library also include Kevin Sahr (Southern Oregon University), Joseph Gilley (Uber), Nick Rabinowitz (Uber), and David Ellis (CarDash).<\/i><\/p><p><a href=\"http:\/\/uber.us11.list-manage1.com\/subscribe?u=092a95bfe05dfa7c27877ca59&amp;id=381801863c\" target=\"_blank\" rel=\"noopener\"><i>Subscribe to our newsletter<\/i><\/a><i>\u00a0to keep up with the latest innovations from Uber Engineering.<\/i><\/p>","created_at":"2018-11-07T16:23:22+0000","updated_at":"2018-11-07T16:23:22+0000","published_at":"2018-06-27T09:00:41+0000","published_by":["Isaac Brodsky"],"annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"eng.uber.com","preview_picture":"https:\/\/eng.uber.com\/wp-content\/uploads\/2018\/06\/Twitter-H3.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":135,"title":"ncredinburgh\/QuickTheories","url":"https:\/\/github.com\/ncredinburgh\/QuickTheories","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/ncredinburgh\/QuickTheories\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/aff0e64603287929ab4f2a7ef0b1156e1f68d3bd\/68747470733a2f2f7472617669732d63692e6f72672f6e63726564696e62757267682f517569636b5468656f726965732e737667\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/ncredinburgh\/QuickTheories.svg\" \/><\/a>\n<a href=\"https:\/\/maven-badges.herokuapp.com\/maven-central\/org.quicktheories\/quicktheories\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f93314fc9b5419b91ce4c4e2cf56a84311c1069b\/68747470733a2f2f6d6176656e2d6261646765732e6865726f6b756170702e636f6d2f6d6176656e2d63656e7472616c2f6f72672e717569636b7468656f726965732f717569636b7468656f726965732f62616467652e7376673f7374796c653d666c6174\" alt=\"Maven Central\" data-canonical-src=\"https:\/\/maven-badges.herokuapp.com\/maven-central\/org.quicktheories\/quicktheories\/badge.svg?style=flat\" \/><\/a><\/p>\n<p>Property-based testing for Java 8.<\/p>\n<p>If you were looking for QuickCheck for Java you just found it.<\/p>\n<p>Unlike many other systems QuickTheories supports both auto-magical shrinking and targeted search using coverage data.<\/p>\n<h2><a id=\"user-content-what-is-property-based-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-is-property-based-testing\"><\/a>What is property based testing<\/h2>\n<p>Traditional unit testing is performed by specifying a series of concrete examples and asserting on the outputs\/behaviour of the unit under test.<\/p>\n<p>Property based testing moves away from concrete examples and instead checks that certain properties hold true for all possible inputs. It does this by automatically generating\na random sample of valid inputs from the possible values.<\/p>\n<p>This can be a good way to uncover bad assumptions made by you and your code.<\/p>\n<p>If the word \"random\" is making you feel a little nervous, don't worry QuickTheories provides ways to keep your tests repeatable.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#quick-start\"><\/a>Quick Start<\/h2>\n<p>Add the QuickTheories jar to your build (see the badge at the top of the page for the maven coordinates of the latest version).<\/p>\n<p>You can run QuickTheories from JUnit, TestNG or any other test framework.<\/p>\n<p>Here we are using JUnit<\/p>\n<div class=\"highlight highlight-source-java\"><pre>import static org.quicktheories.QuickTheory.qt;\nimport static org.quicktheories.generators.SourceDSL.*;\npublic class SomeTests {\n  @Test\n  public void addingTwoPositiveIntegersAlwaysGivesAPositiveInteger(){\n    qt()\n    .forAll(integers().allPositive()\n          , integers().allPositive())\n    .check((i,j) -&gt; i + j &gt; 0); \n  }\n}<\/pre><\/div>\n<p>The static import <code>org.quicktheories.QuickTheory.qt<\/code> provides access to the QuickTheories DSL.<\/p>\n<p>The static import <code>org.quicktheories.generators.SourceDSL.*<\/code> provides access to a DSL that allows valid inputs to be defined.<\/p>\n<p>This property looks pretty simple, it just checks that adding two integers always produces a number greater than 0.<\/p>\n<p>This couldn't possibly fail could it? That would mean math was broken.<\/p>\n<p>If we run this test we get something like :-<\/p>\n<pre>java.lang.AssertionError: Property falsified after 1 example(s) \nSmallest found falsifying value(s) :-\n{840226137, 1309274625}\nOther found falsifying value(s) :- \n{848253830, 1320535400}\n{841714728, 1317667877}\n{840894251, 1310141916}\n{840226137, 1309274625}\n \nSeed was 29678088851250\t\n<\/pre>\n<p>The falsified theory has highlighted something that we forgot.<\/p>\n<p>Math works just fine, but in Java integers can overflow.<\/p>\n<h3><a id=\"user-content-without-static-imports\" class=\"anchor\" aria-hidden=\"true\" href=\"#without-static-imports\"><\/a>Without static imports<\/h3>\n<p>If you prefer the QuickTheories entry points can be brought into scope by implementing an interface, removing the need for static imports.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>public class SomeTests implements WithQuickTheories {\n  @Test\n  public void addingTwoPositiveIntegersAlwaysGivesAPositiveInteger(){\n    qt()\n    .forAll(integers().allPositive()\n          , integers().allPositive())\n    .check((i,j) -&gt; i + j &gt; 0); \n  }\n}<\/pre><\/div>\n<h3><a id=\"user-content-less-verbose\" class=\"anchor\" aria-hidden=\"true\" href=\"#less-verbose\"><\/a>Less verbose<\/h3>\n<p>The source DSL reads nicely but can be a little verbose. Most of the core generators can also be accessed by importing <code>org.quicktheories.generators.Generate<\/code>. This provides simple static methods that return generators of core types.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>import static org.quicktheories.generators.Generate.*;\n@Test\npublic void someProperty() {\n  qt()\n  .forAll(range(1, 102), constant(7))\n  .check((i,c) -&gt; i + c &gt;= 7);\n}\n<\/pre><\/div>\n<h3><a id=\"user-content-shrinking\" class=\"anchor\" aria-hidden=\"true\" href=\"#shrinking\"><\/a>Shrinking<\/h3>\n<p>QuickTheories supports shrinking.<\/p>\n<p>This means that it doesn't just find a falsifying value and stop. Instead it will try to find other smaller (or \"simpler\") values that also invalidate the theory.<\/p>\n<p>By default QuickTheories will spend about 100 times more effort looking for smaller values than it did looking for the original falsifying value.<\/p>\n<p>The smallest found value is reported along with a sample of any other falsifying values found along the way.<\/p>\n<p>There is no guarantee that this is the smallest possible falsifying value or that others don't exist. Generally the shrunk values will be easier to understand and work with than the original un-shrunk ones \u2013 patterns might be visible in the reported values.<\/p>\n<p>Unlike straight QuickCheck clones QuickTheories does not require you to supply your own shrinking implementation for each type. Shrinking is performed automatically for any and all types. The mechanism by which this is achieved does not make any assumptions about the structure or implementation of the type or break encapsulation.<\/p>\n<h3><a id=\"user-content-seeds-and-repeatable-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#seeds-and-repeatable-tests\"><\/a>Seeds and repeatable tests<\/h3>\n<p>At the end of the report the Seed is reported.<\/p>\n<p>This is the value from which all randomness is derived in QuickTheories.<\/p>\n<p>By default it is set to the <code>System.nanoTime()<\/code> so the values will be different each time QuickTheories is run, however the seed can also be set explicitly so runs can be reproduced and deterministic.<\/p>\n<p>Whenever a property is falsified the seed used is reported so you can always reproduce the exact same run.<\/p>\n<p>It is therefore always possible to recreate a run, and you can opt for a fully deterministic behaviour by using a single fixed seed.<\/p>\n<p>Two methods are provided to set the seed.<\/p>\n<p>Directly using the DSL<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  qt()\n  .withFixedSeed(0)\n  .forAll( . . .)<\/pre><\/div>\n<p>Or using the <code>QT_SEED<\/code> system property.<\/p>\n<p>The same tests can therefore be run with a fixed seed for the purpose of catching regression, or with a changing seed so that falsifying values are constantly being searched for.<\/p>\n<h3><a id=\"user-content-assertions\" class=\"anchor\" aria-hidden=\"true\" href=\"#assertions\"><\/a>Assertions<\/h3>\n<p>Our example theory used a simple predicate, but sometimes it would be nice to take advantage of the functionality provided by assertion libraries such as <a href=\"http:\/\/joel-costigliola.github.io\/assertj\/\" rel=\"nofollow\">assertj<\/a> and <a href=\"https:\/\/github.com\/hamcrest\">hamcrest<\/a>.<\/p>\n<p>This can be done using the <code>checkAssert<\/code> method.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void someTheory() {\n    qt().forAll(longs().all())\n        .checkAssert(i -&gt; assertThat(i).isEqualsTo(42));\n  }<\/pre><\/div>\n<p>Any block of code that returns void can be passed to <code>checkAssert<\/code>. Any unchecked exception will be interpreted as falsifying the theory.<\/p>\n<h3><a id=\"user-content-assumptions\" class=\"anchor\" aria-hidden=\"true\" href=\"#assumptions\"><\/a>Assumptions<\/h3>\n<p>As we've seen we can create theories from a pair of Gens - which produce a pair of values.<\/p>\n<p>In fact we can create theories about any number of values between 1 and 4.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>   @Test\n  public void someTheoryOrOther(){\n    qt()\n    .forAll(integers().allPositive()\n          , strings().basicLatinAlphabet().ofLengthBetween(0, 10)\n          , lists().allListsOf(integers().all()).ofSize(42))\n    .check((i,s,l) -&gt; l.contains(i) &amp;&amp; s.equals(\"\"));\n  }<\/pre><\/div>\n<p>In the example above we use three Gens, as you can see QuickTheories provides ways of generating most common Java types.<\/p>\n<p>A Gen is just a simple function from a random number generator to a value. As we can see, the DSL provides a way to put constraints on the values we generate (e.g we will only generate positive integers and the lists in this example will only be of size 42).<\/p>\n<p>Whenever possible you should use the DSL to provide constraints, but sometimes you might need to constrain the domain in ways that cannot be expressed with the DSL.<\/p>\n<p>When this happens use assumptions.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void someTheoryOrOther(){\n    qt()\n    .forAll(integers().allPositive()\n          , strings().basicLatinAlphabet().ofLengthBetween(0, 10)\n          , lists().allListsOf(integers().all()).ofSize(42))\n    .assuming((i,s,l) -&gt; s.contains(i.toString())) \/\/ &lt;-- an assumption\n    .check((i,s,l) -&gt; l.contains(i) &amp;&amp; s.contains(i.toString()));\n  }<\/pre><\/div>\n<p>Assumptions further constrain the values which form the subject of the theory.<\/p>\n<p>Although we could always replace the constraints we created in the DSL with assumptions, this would be very inefficient. QuickTheories would have to spend a lot of effort just trying to find valid values before it could try to invalidate a theory.<\/p>\n<p>As difficult to find values probably represent a coding error, QuickTheories will throw an error if less than 10% of the generated values pass the assumptions:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void badUseOfAssumptions() {\n    qt()\n    .forAll(integers().allPositive())\n    .assuming(i -&gt; i &lt; 30000)\n    .check( i -&gt; i &lt; 3000);\n  }<\/pre><\/div>\n<p>Gives<\/p>\n<pre>java.lang.IllegalStateException: Gave up after finding only 107 example(s) matching the assumptions\n\tat org.quicktheories.quicktheories.core.ExceptionReporter.valuesExhausted(ExceptionReporter.java:20)\n<\/pre>\n<p>(Note: this assumption could have been replaced by the following:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>   @Test\n  public void goodUseOfSource(){\n    qt().forAll(integers().from(1).upTo(30000))\n    .check( i -&gt; i &lt; 3000);\n  }<\/pre><\/div>\n<p>Which gives the following failure message: )<\/p>\n<pre>java.lang.AssertionError: Property falsified after 1 example(s) \nSmallest found falsifying value(s) :-\n3000\nOther found falsifying value(s) :- \n13723\n13722\n13721\n13720\n13719\n13718\n13717\n13716\n13715\n13714\n \nSeed was 2563360080237\n<\/pre>\n<h3><a id=\"user-content-gens\" class=\"anchor\" aria-hidden=\"true\" href=\"#gens\"><\/a>Gens<\/h3>\n<p>It is likely that you will want to construct instances of your own types. You could do this within each check, but this would result in a lot of code duplication.<\/p>\n<p>Instead you can define a conversion function. This can be done inline, or placed somewhere convenient for reuse.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void someTheoryOrOther(){\n    qt()\n    .forAll(integers().allPositive()\n          , integers().allPositive())\n    .as( (width,height) -&gt; new Widget(width,height) ) \/\/ &lt;-- convert to our own type here\n    .check( widget -&gt; widget.isValid());\n  }<\/pre><\/div>\n<p>This works well for simple cases, but there are two problems.<\/p>\n<ol><li>We cannot refer to the original width and height integers in our theory. So we couldn't (for example) check that the widget had the expected size.<\/li>\n<li>If our widget doesn't define a <code>toString<\/code> method it is hard to know what the falsifying values were<\/li>\n<\/ol><p>Both of these problems are solved by the <code>asWithPrecursors<\/code> method<\/p>\n<pre>  @Test\n  public void someTheoryOrOther(){\n     qt()\n    .forAll(integers().allPositive()\n          , integers().allPositive())\n    .asWithPrecursor( (width,height) -&gt; new Widget(width,height) )\n    .check( (width,height,widget) -&gt; widget.size() &gt; width * height ); \n  }\n<\/pre>\n<p>When this fails it gives us<\/p>\n<pre>java.lang.AssertionError: Property falsified after 2 example(s)\nSmallest found falsifying value(s) :-\n{43, 23259, com.example.QuickTheoriesExample$Widget@9e89d68}\nOther found falsifying value(s) :- \n{536238991, 619642140, com.example.QuickTheoriesExample$Widget@59f95c5d}\n{2891501, 215920967, com.example.QuickTheoriesExample$Widget@5ccd43c2}\n{1479099, 47930205, com.example.QuickTheoriesExample$Widget@4aa8f0b4}\n{297099, 11425635, com.example.QuickTheoriesExample$Widget@7960847b}\n{288582, 10972429, com.example.QuickTheoriesExample$Widget@6a6824be}\n{14457, 5650202, com.example.QuickTheoriesExample$Widget@5c8da962}\n{14456, 393098, com.example.QuickTheoriesExample$Widget@512ddf17}\n{14454, 38038, com.example.QuickTheoriesExample$Widget@2c13da15}\n{14453, 38037, com.example.QuickTheoriesExample$Widget@77556fd}\n{14452, 38036, com.example.QuickTheoriesExample$Widget@368239c8}\n \nSeed was 4314310398163\n<\/pre>\n<p>Notice that shrinking works for our custom type without any effort on our part.<\/p>\n<p>Defining the values that make up the valid domain for your objects might not be straightforward and could result in a lot of repeated code between theories.<\/p>\n<p>Fortunately the Gen objects that produce the random values can be freely reused and combined with other Gens.<\/p>\n<p>e.g.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void cylindersHavePositiveAreas() {\n    qt()\n    .forAll(cylinders())\n    .check( cylinder -&gt; cylinder.area().compareTo(BigDecimal.ZERO) &gt; 0);\n  }\n  \n  private Gen&lt;Cylinder&gt; cylinders() {\n    return radii().zip(heights(),\n        (radius, height) -&gt; new Cylinder(radius, height))\n        .assuming(cylinder -&gt; some sort of validation );\n  }\n  private Gen&lt;Integer&gt; heights() {\n    return integers().from(79).upToAndIncluding(1004856);\n  }\n  private Gen&lt;Integer&gt; radii() {\n    return integers().allPositive();\n  }<\/pre><\/div>\n<p>Gens provide a number of methods that allows them to be mapped to different types or combined with other Gens. All these operations preserve assumptions and allow the resulting types to be shrunk without the need for any additional code.<\/p>\n<h2><a id=\"user-content-modifying-the-falsification-output\" class=\"anchor\" aria-hidden=\"true\" href=\"#modifying-the-falsification-output\"><\/a>Modifying the falsification output<\/h2>\n<p>Values produces by the sources DSL should provide clear falsification messages.<\/p>\n<p>If you are working with your own Gens, or would like to modify the defaults, you can supply your own function to be used when describing the falsifying values.<\/p>\n<p>For example:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void someTestInvolvingCylinders() {\n      qt()\n      .forAll(integers().allPositive().describedAs(r -&gt; \"Radius = \" + r)\n             , integers().allPositive().describedAs(h -&gt; \"Height = \" + h))\n      .check((r,h) -&gt; whatever);\n  }<\/pre><\/div>\n<p>Custom description functions will be retained when converting to a type with precursors. A description function for the converted type can be optionally passed to the asWithPrecursor function.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void someTestInvolvingCylinders() {\n      qt()\n      .forAll(integers().allPositive().describedAs(r -&gt; \"Radius = \" + r)\n             , integers().allPositive().describedAs(h -&gt; \"Height = \" + h))\n      .asWithPrecursor((r,h) -&gt; new Cylinder(r,h)\n                       , cylinder -&gt; \"Cylinder r =\" + cylinder.radius() + \" h =\" + cylinder.height())        \n      .check((i,j,k) -&gt; whatever);\n  }<\/pre><\/div>\n<p>A description function can be provided for a type converted without precursors as follows:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void someTestInvolvingCylinders() {\n      qt()\n      .forAll(integers().allPositive().describedAs(r -&gt; \"Radius = \" + r)\n             , integers().allPositive().describesAs(h -&gt; \"Height = \" + h))\n      .as( (r,h) -&gt; new Cylinder(r,h))\n      .describedAs(cylinder -&gt; \"Cylinder r =\" + cylinder.radius() + \" h =\" + cylinder.height())        \n      .check(l -&gt; whatever);\n  }<\/pre><\/div>\n<h2><a id=\"user-content-coverage-guidance\" class=\"anchor\" aria-hidden=\"true\" href=\"#coverage-guidance\"><\/a>Coverage guidance<\/h2>\n<p>QuickTheories includes an experimental feature to improve the efficiency of the targeted search. If the coverage jar is included on the classpath QuickTheories will insert probes into the code under test. When these probes reveal that a particular example has exercised new code paths QuickTheories will concentrate the search in this area.<\/p>\n<p>This approach has been demonstrated to be far more effective at falsifying branched code in simple example cases. It is not yet known how well it performs in real world scenarios.<\/p>\n<p>There are some disadvantages to coverage guidance. In order to measure coverage QuickTheories must attach an agent to the JVM. The agent will be active from the moment it is installed until the JVM exits - this means it may be active while non QuickTheory tests are running. This will result in an ~10% reduction in performance, and may interfere with JaCoCo and other coverage systems if they are also active. For this reason we recommend running QuickTheories tests in a separate suite if you are using coverage guidance.<\/p>\n<p>Coverage guidance can be disabled on a per test basis.<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  qt() \n  .withGuidance(noGuidance())\n  .etc<\/pre><\/div>\n<h2><a id=\"user-content-configuration-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration-properties\"><\/a>Configuration properties<\/h2>\n<p>Three system properties can be set that determine QuickTheories behaviour:<\/p>\n<ul><li><code>QT_SEED<\/code> - the random seed to use<\/li>\n<li><code>QT_EXAMPLES<\/code> - the number of examples to try for each theory<\/li>\n<li><code>QT_SHRINKS<\/code> - the number of shrink attempts to make<\/li>\n<\/ul><h2><a id=\"user-content-writing-good-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#writing-good-properties\"><\/a>Writing good properties<\/h2>\n<p>Properties should not just duplicate the logic of your code under test (this is equally true for the example based testing).<\/p>\n<p>Instead properties should try to specify very simple but general invariants that should hold true. Start with very simple general properties and get more specific as\nyou go along.<\/p>\n<p>Some common patterns that produce good properties include :-<\/p>\n<p>(note, these patterns are largely a summary of the material at <a href=\"http:\/\/fsharpforfunandprofit.com\/posts\/property-based-testing-2\/\" rel=\"nofollow\">fsharpforfunandprofit<\/a>)<\/p>\n<h3><a id=\"user-content-the-invariant-pattern-aka-some-things-never-change\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-invariant-pattern-aka-some-things-never-change\"><\/a>The invariant pattern aka \"Some things never change\"<\/h3>\n<p>Some things are expected to remain constant, e.g a map operation should produce the same number of items that it was given, the total balance across two bank accounts should\nremain constant after a transfer etc.<\/p>\n<h3><a id=\"user-content-the-inverse-function-pattern-aka-there-and-back-again\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-inverse-function-pattern-aka-there-and-back-again\"><\/a>The inverse function pattern aka \"There and back again\"<\/h3>\n<p>If we have two functions that are the inverse of each other then applying the input of one to the other should result in no change.<\/p>\n<p>Common examples of inverse function pairs include<\/p>\n<ul><li>serialisation \/ deserialisation<\/li>\n<li>compression \/ decompression<\/li>\n<li>encryption \/ decryption<\/li>\n<li>create \/ delete<\/li>\n<\/ul><h3><a id=\"user-content-analogous-function-pattern-aka-different-paths-same-destination\" class=\"anchor\" aria-hidden=\"true\" href=\"#analogous-function-pattern-aka-different-paths-same-destination\"><\/a>Analogous function pattern aka \"Different paths same destination\"<\/h3>\n<p>If you have two functions that implement the same logic, but differ in some other property (perhaps one is inefficient, insecure or implemented in a third party library) then a property can be defined that checks the outputs of the functions match given the same input.<\/p>\n<h3><a id=\"user-content-idempotence-aka-the-more-things-change-the-more-they-stay-the-same\" class=\"anchor\" aria-hidden=\"true\" href=\"#idempotence-aka-the-more-things-change-the-more-they-stay-the-same\"><\/a>Idempotence aka \"The more things change, the more they stay the same\"<\/h3>\n<p>Sometimes it's important\/logical that performing an operation multiple times has no effect. e.g if you trim the whitespace from a string multiple times, only the first call to trim should have any observable effect.<\/p>\n<h2><a id=\"user-content-simple-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#simple-examples\"><\/a>Simple examples<\/h2>\n<p>An example test that is falsifying, showing that adding two positive integers in Java does not always give a positive integer:<\/p>\n<div class=\"highlight highlight-source-java\"><pre>@Test\n  public void addingTwoPositiveIntegersAlwaysGivesAPositiveInteger(){\n    qt()\n    .forAll(integers().allPositive()\n          , integers().allPositive())\n    .check((i,j) -&gt; i + j &gt; 0);  \/\/fails\n  }\n<\/pre><\/div>\n<p>An example of multiple tests for code that claims to find the greatest common divisor between two integers. The first property test fails due to a <code>java.lang.StackOverflowError<\/code> error (caused by attempting to take the absolute value of <code>Integer.MIN_VALUE<\/code>).<\/p>\n<div class=\"highlight highlight-source-java\"><pre>  @Test\n  public void shouldFindThatAllIntegersHaveGcdOfOneWithOne() {\n    qt().forAll(integers().all()).check(n -&gt; gcd(n, 1) == 1); \/\/ fails on\n                                                              \/\/ -2147483648\n  }\n  @Test\n  public void shouldFindThatAllIntegersInRangeHaveGcdOfOneWithOne() {\n    qt().forAll(integers().between(-Integer.MAX_VALUE, Integer.MAX_VALUE))\n        .check(n -&gt; gcd(n, 1) == 1);\n  }\n  @Test\n  public void shouldFindThatAllIntegersHaveGcdThemselvesWithThemselves() {\n    qt().forAll(integers().between(-Integer.MAX_VALUE, Integer.MAX_VALUE))\n        .check(n -&gt; gcd(n, n) == Math.abs(n));\n  }\n  @Test\n  public void shouldFindThatGcdOfNAndMEqualsGcdMModNAndN() {\n    qt().forAll(integers().between(-Integer.MAX_VALUE, Integer.MAX_VALUE)\n               ,integers().between(-Integer.MAX_VALUE, Integer.MAX_VALUE))\n        .check((n, m) -&gt; gcd(n, m) == gcd(m % n, n));\n  }\n  private int gcd(int n, int m) {\n    if (n == 0) {\n      return Math.abs(m);\n    }\n    if (m == 0) {\n      return Math.abs(n);\n    }\n    if (n &lt; 0) {\n      return gcd(-n, m);\n    }\n    if (m &lt; 0) {\n      return gcd(n, -m);\n    }\n    if (n &gt; m) {\n      return gcd(m, n);\n    }\n    return gcd(m % n, n);\n  }<\/pre><\/div>\n<h2><a id=\"user-content-design-goals\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-goals\"><\/a>Design Goals<\/h2>\n<p>QuickTheories was written with the following design goals<\/p>\n<ol><li>Random by default, but builds must be repeatable<\/li>\n<li>Support for shrinking<\/li>\n<li>Independent of test api (JUnit, TestNG etc)<\/li>\n<\/ol><p>It turned out that number 2 was the hard bit as it had many implications for the design. The approach was completely changed between releases in the 0.1x and 0.2x series.<\/p>\n<p>As of 0.20 shrinking uses an approach similar to the python library hypothesis, whereby shrinking is unaware of the type it is generating. This approach is less flexible than the original approach but allows Gens to be freely composed together while greatly reducing the size of the codebase.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#background\"><\/a>Background<\/h2>\n<p>QuickTheories was produced at <a href=\"http:\/\/ncredinburgh.com\/\" rel=\"nofollow\">NCR Edinburgh<\/a> as part of our graduate training program.<\/p>\n<p>We like to do training a little differently \u2013 our <a href=\"http:\/\/github.com\/katyrae\">new graduates<\/a> get to work on an interesting project for a weeks with a more worn in and <a href=\"http:\/\/github.com\/hcoles\">weathered member of our staff<\/a>. Our motto for these projects is \"software that can fail\" \u2013 so we get to play with interesting ideas that may come to nothing.<\/p>\n<p>We're happy to share the results as open source when we think they're successful.<\/p>\n<h2><a id=\"user-content-other-property-based-testing-systems-for-java\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-property-based-testing-systems-for-java\"><\/a>Other property based testing systems for Java<\/h2>\n<p>If you don't like QuickTheories you might want to try one of the other systems below which have different design goals. Besides junit-quickcheck, none of them look at implementing shrinking, but all provide ways of generating random values and should work on earlier versions of Java.<\/p>\n<ul><li><a href=\"http:\/\/pholser.github.io\/junit-quickcheck\" rel=\"nofollow\">JUnit-quickcheck<\/a>. Tightly integrated with JUnit, uses annotations to configure generators.\n<ul><li>As of v0.6 junit-quickcheck also supports shrinking.<\/li>\n<\/ul><\/li>\n<li><a href=\"http:\/\/www.jcheck.org\/\" rel=\"nofollow\">JCheck<\/a>. Tightly integrated with JUnit. Does not look to be maintained.<\/li>\n<li><a href=\"https:\/\/bitbucket.org\/blob79\/quickcheck\" rel=\"nofollow\">QuickCheck<\/a>. Not tied to a test framework - provides generators of random values to be used in tests.<\/li>\n<li><a href=\"http:\/\/www.functionaljava.org\/\" rel=\"nofollow\">FunctionalJava<\/a>. Apparently contains a property based testing system, but appears to be completely undocumented.<\/li>\n<li><a href=\"http:\/\/www.scalacheck.org\/\" rel=\"nofollow\">ScalaCheck<\/a>. Mature property based testing system with shrinking, but requires Scala rather than Java. Also seem to be <a href=\"https:\/\/github.com\/rickynils\/scalacheck\/issues\/317\">design level issues with how shrinking works<\/a>.<\/li>\n<li><a href=\"http:\/\/jqwik.net\/\" rel=\"nofollow\">jqwik<\/a>. JUnit 5 based implementation using annotations. Supports shrinking.<\/li>\n<\/ul><\/article>","created_at":"2018-11-07T16:23:21+0000","updated_at":"2018-11-07T16:23:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/5927858?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":136,"title":"harababurel\/gcsf","url":"https:\/\/github.com\/harababurel\/gcsf","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/a717eca9ae37d365ea051ce95394e24d686b4125\/68747470733a2f2f692e696d6775722e636f6d2f397864467751712e706e67\"><img width=\"300px\" height=\"300px\" title=\"Size Limit logo\" src=\"https:\/\/camo.githubusercontent.com\/a717eca9ae37d365ea051ce95394e24d686b4125\/68747470733a2f2f692e696d6775722e636f6d2f397864467751712e706e67\" data-canonical-src=\"https:\/\/i.imgur.com\/9xdFwQq.png\" alt=\"image\" style=\"float: right;\" \/><\/a><\/p>\n<p><a href=\"https:\/\/travis-ci.org\/harababurel\/gcsf\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/b8188e6dbb0dd185d8f0b3198f687f7f3a3e5685\/68747470733a2f2f7472617669732d63692e6f72672f686172616261627572656c2f676373662e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/harababurel\/gcsf.svg?branch=master\" \/><\/a>\n<a href=\"https:\/\/crates.io\/crates\/gcsf\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d131846d2e3fedf014573e9a0a0cdaacccad1c4c\/68747470733a2f2f696d672e736869656c64732e696f2f6372617465732f762f676373662e737667\" alt=\"Crates.io\" data-canonical-src=\"https:\/\/img.shields.io\/crates\/v\/gcsf.svg\" \/><\/a>\n<a href=\"https:\/\/docs.rs\/gcsf\/latest\/gcsf\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/b097bd084042048194bda2ce64a42d7ceb9530a9\/68747470733a2f2f646f63732e72732f676373662f62616467652e737667\" alt=\"Docs\" data-canonical-src=\"https:\/\/docs.rs\/gcsf\/badge.svg\" \/><\/a><br \/><a href=\"https:\/\/github.com\/harababurel\/gcsf\/issues\"><img src=\"https:\/\/camo.githubusercontent.com\/9d9920a0e65525428e6d546b1118659ea01e8934\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f6973737565732f686172616261627572656c2f676373662e737667\" alt=\"GitHub Issues\" data-canonical-src=\"https:\/\/img.shields.io\/github\/issues\/harababurel\/gcsf.svg\" \/><\/a>\n<a href=\"https:\/\/crates.io\/crates\/gcsf\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/bb81703bd2aa86178df93ad3389544828fa1ec98\/68747470733a2f2f696d672e736869656c64732e696f2f6372617465732f642f676373662e737667\" alt=\"Downloads\" data-canonical-src=\"https:\/\/img.shields.io\/crates\/d\/gcsf.svg\" \/><\/a>\n<a href=\"https:\/\/github.com\/harababurel\/gcsf\/blob\/master\/LICENSE\"><img src=\"https:\/\/camo.githubusercontent.com\/7eda69a169c140b99b6c943b858d5f930107f600\/68747470733a2f2f696d672e736869656c64732e696f2f6372617465732f6c2f676373662e737667\" alt=\"MIT License\" data-canonical-src=\"https:\/\/img.shields.io\/crates\/l\/gcsf.svg\" \/><\/a><\/p>\n<p>GCSF is a virtual filesystem that allows users to mount their Google Drive account locally and interact with it as a regular disk partition. You can find out more in this <a href=\"https:\/\/sergiu.ml\/~sergiu\/thesis.pdf\" rel=\"nofollow\">paper<\/a>.<\/p>\n<h3><a id=\"user-content-requirements\" class=\"anchor\" aria-hidden=\"true\" href=\"#requirements\"><\/a>Requirements<\/h3>\n<p>GCSF requires the stable branch of the Rust programming language, which can be installed following the instructions on <a href=\"https:\/\/rustup.rs\" rel=\"nofollow\">rustup.rs<\/a>. If you already have Rust installed, make sure that it is updated to the latest version (\u22651.26):<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ rustup update stable<\/pre><\/div>\n<h4><a id=\"user-content-osx\" class=\"anchor\" aria-hidden=\"true\" href=\"#osx\"><\/a>OSX<\/h4>\n<p>On Mac OSX, GCSF requires <a href=\"https:\/\/osxfuse.github.io\/\" rel=\"nofollow\">osxfuse<\/a> and <a href=\"http:\/\/macappstore.org\/pkg-config\/\" rel=\"nofollow\">pkg-config<\/a>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ brew update; brew install pkg-config; brew tap homebrew\/cask; brew cask install osxfuse<\/pre><\/div>\n<h4><a id=\"user-content-ubuntu\" class=\"anchor\" aria-hidden=\"true\" href=\"#ubuntu\"><\/a>Ubuntu<\/h4>\n<p>On Ubuntu, GCSF requires <a href=\"https:\/\/packages.ubuntu.com\/trusty\/libfuse-dev\" rel=\"nofollow\">libfuse-dev<\/a> and <a href=\"https:\/\/packages.ubuntu.com\/xenial\/pkg-config\" rel=\"nofollow\">pkg-config<\/a>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>sudo apt-get install -y libfuse-dev pkg-config<\/pre><\/div>\n<h4><a id=\"user-content-arch-linux\" class=\"anchor\" aria-hidden=\"true\" href=\"#arch-linux\"><\/a>Arch Linux<\/h4>\n<p>An AUR package is maintained by <a href=\"https:\/\/github.com\/axionl\">axionl<\/a>: <a href=\"https:\/\/aur.archlinux.org\/packages\/gcsf-git\/\" rel=\"nofollow\">gcsf-git<\/a>.<\/p>\n<h4><a id=\"user-content-other-linux-distros\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-linux-distros\"><\/a>Other linux distros<\/h4>\n<p>Make sure you have <code>pkg-config<\/code> and the <code>fuse<\/code> library installed. These are usually found in the package repositories of major distributions.<\/p>\n<h4><a id=\"user-content-windows\" class=\"anchor\" aria-hidden=\"true\" href=\"#windows\"><\/a>Windows<\/h4>\n<p>Unfortunately, Windows is not supported at the time being. See issue <a href=\"https:\/\/github.com\/harababurel\/gcsf\/issues\/19\">#19<\/a>.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h3>\n<p>After all requirements are met, GCSF can be installed using <code>cargo<\/code>:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ cargo install gcsf<\/pre><\/div>\n<p>This will generate the <code>gcsf<\/code> binary in <code>$HOME\/.cargo\/bin<\/code>. Make sure that this directory is in your <code>PATH<\/code> variable: <code>export PATH=$PATH:$HOME\/.cargo\/bin<\/code><\/p>\n<p>Alternatively, you can download a <a href=\"https:\/\/github.com\/harababurel\/gcsf\/releases\">release binary<\/a> for your platform.<\/p>\n<h3><a id=\"user-content-configuration\" class=\"anchor\" aria-hidden=\"true\" href=\"#configuration\"><\/a>Configuration<\/h3>\n<p>GCSF will attempt to create a configuration file in <code>$XDG_CONFIG_HOME\/gcsf\/gcsf.toml<\/code>, which is usually defined as <code>$HOME\/.config\/gcsf\/gcsf.toml<\/code>. Credentials are stored in the same directory.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h3>\n<p>The first step is to log in to Drive and authorize the application. A name must be provided for the session:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ gcsf login some_session_name\nPlease direct your browser to https:\/\/accounts.google.com\/o\/oauth2\/[...] and follow the instructions displayed there.\nSuccessfully logged in. Saved credentials to \"$HOME\/.config\/gcsf\/some_session_name\"<\/pre><\/div>\n<p>You can also list all existing sessions:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ gcsf list\nSessions:\n        - personal\n        - some_session_name\n        - work<\/pre><\/div>\n<p>And then mount one (or more) of them:<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ gcsf mount \/mnt\/gcsf -s some_session_name\nINFO  gcsf &gt; Creating and populating file system...\nINFO  gcsf &gt; File sytem created.\nINFO  gcsf &gt; Mounting to \/mnt\/gcsf\nINFO  gcsf &gt; Mounted to \/mnt\/gcsf<\/pre><\/div>\n<p>You can now find the contents of your Drive account in <code>\/mnt\/gcsf<\/code>:<\/p>\n<p style=\"text-align: left;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/aafbebafd44ae72e4d39475546d599c3c7c3f3d5\/68747470733a2f2f692e696d6775722e636f6d2f6a6446497535592e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/aafbebafd44ae72e4d39475546d599c3c7c3f3d5\/68747470733a2f2f692e696d6775722e636f6d2f6a6446497535592e706e67\" alt=\"GCSF ls\" width=\"530px\" height=\"165px\" data-canonical-src=\"https:\/\/i.imgur.com\/jdFIu5Y.png\" \/><\/a>\n<\/p>\n<p>Using Ranger:<\/p>\n<p style=\"text-align: left;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/f226704b7c223787c29eb673537a6676b0c9a936\/68747470733a2f2f692e696d6775722e636f6d2f427553394244442e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/f226704b7c223787c29eb673537a6676b0c9a936\/68747470733a2f2f692e696d6775722e636f6d2f427553394244442e706e67\" alt=\"GCSF in Ranger\" width=\"616px\" height=\"351px\" data-canonical-src=\"https:\/\/i.imgur.com\/BuS9BDD.png\" \/><\/a>\n<\/p>\n<p>Or Thunar:<\/p>\n<p style=\"text-align: left;\">\n  <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b3cdcd432ab8fb3d2eb23ca2bbf5b251d975d8b2\/68747470733a2f2f692e696d6775722e636f6d2f394a534471657a2e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/b3cdcd432ab8fb3d2eb23ca2bbf5b251d975d8b2\/68747470733a2f2f692e696d6775722e636f6d2f394a534471657a2e6a7067\" alt=\"GCSF in Thunar\" width=\"746px\" height=\"176px\" data-canonical-src=\"https:\/\/i.imgur.com\/9JSDqez.jpg\" \/><\/a>\n<\/p>\n<h3><a id=\"user-content-why-gcsf\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-gcsf\"><\/a>Why GCSF?<\/h3>\n<p>GCSF stands for \"Google Conduce Sistem de Fi\u0219iere\" which translated from Romanian is \"Google Drive Filesystem\". However <a href=\"https:\/\/github.com\/robin-thomas\/GDFS\">GDFS<\/a> already exists so it remains GCSF.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h3>\n<p>Contributions are welcome. You can also help by reporting or fixing <a href=\"https:\/\/github.com\/harababurel\/gcsf\/issues\">issues<\/a>.<\/p>\n<\/article>","created_at":"2018-11-07T16:23:21+0000","updated_at":"2018-11-07T16:23:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/8619823?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":137,"title":"Featured products","url":"https:\/\/ergodox-ez.com\/collections\/frontpage","content":"<div class=\"site-wrapper\"><main class=\"app-content\"><div class=\"hero-products\"><div class=\"container\"><article class=\"product-snippet\"><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-original\" title=\"ErgoDox EZ Original\" class=\"display\">\n            <img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1152\/3264\/products\/e1fbe3297b62fb3a394029694f7d3d90_large.jpg?v=1492776831\" alt=\"ErgoDox EZ Original\" class=\"image-product\" \/><\/a><div class=\"content\"><p><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-original\" title=\"ErgoDox EZ Original\" class=\"title\">ErgoDox EZ Original<\/a>\n              \n                \n                  \n  \n    \n      \n    \n    \n    $325.00\n  \n                \n                \n                \n                   (free shipping)<\/p><p>MOST POPULAR<\/p><\/div><\/article><article class=\"product-snippet\"><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-shine\" title=\"ErgoDox EZ Shine\" class=\"display\">\n            <img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1152\/3264\/products\/99aca2643827f677e1d74ab2a69f5741_large.jpg?v=1492776855\" alt=\"ErgoDox EZ Shine\" class=\"image-product\" \/><\/a><div class=\"content\"><p><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-shine\" title=\"ErgoDox EZ Shine\" class=\"title\">ErgoDox EZ Shine<\/a>\n              \n                \n                  \n  \n    \n      \n    \n    \n    $355.00\n  \n                \n                \n                \n                   (free shipping)<\/p><\/div><\/article><article class=\"product-snippet\"><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-glow-standalone\" title=\"Pre-order [Nov shipment] - ErgoDox EZ Glow Standalone\" class=\"display\">\n            <img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1152\/3264\/products\/P1100608_large.jpg?v=1541016486\" alt=\"Pre-order [Nov shipment] - ErgoDox EZ Glow Standalone\" class=\"image-product\" \/><\/a><div class=\"content\"><p><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-glow-standalone\" title=\"Pre-order [Nov shipment] - ErgoDox EZ Glow Standalone\" class=\"title\">Pre-order [Nov shipment] - ErgoDox EZ Glow Standalone<\/a>\n              \n                \n                  \n  \n    \n      \n    \n    \n    $299.00\n  \n                \n                \n                \n                   (free shipping)<\/p><\/div><\/article><article class=\"product-snippet\"><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-glow\" title=\"Pre-order [Nov shipment] - ErgoDox EZ Glow\" class=\"display\">\n            <img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1152\/3264\/products\/P1100593_large.jpg?v=1540477151\" alt=\"Pre-order [Nov shipment] - ErgoDox EZ Glow\" class=\"image-product\" \/><\/a><div class=\"content\"><p><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-glow\" title=\"Pre-order [Nov shipment] - ErgoDox EZ Glow\" class=\"title\">Pre-order [Nov shipment] - ErgoDox EZ Glow<\/a>\n              \n                \n                  \n  \n    \n      \n    \n    \n    $354.00\n  \n                \n                \n                \n                   (free shipping)<\/p><\/div><\/article><article class=\"product-snippet\"><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-original-standalone\" title=\"ErgoDox EZ Original Standalone\" class=\"display\">\n            <img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1152\/3264\/products\/8695e3be4c0983607b10034b7555489d_large.jpg?v=1492776796\" alt=\"ErgoDox EZ Original Standalone\" class=\"image-product\" \/><\/a><div class=\"content\"><p><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-original-standalone\" title=\"ErgoDox EZ Original Standalone\" class=\"title\">ErgoDox EZ Original Standalone<\/a>\n              \n                \n                  \n  \n    \n      \n    \n    \n    $270.00\n  \n                \n                \n                \n                   (free shipping)<\/p><\/div><\/article><article class=\"product-snippet\"><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-shine-standalone\" title=\"ErgoDox EZ Shine Standalone\" class=\"display\">\n            <img src=\"https:\/\/cdn.shopify.com\/s\/files\/1\/1152\/3264\/products\/98171a1dcc4bbc3dca675b62e4d66746_large.jpg?v=1492776864\" alt=\"ErgoDox EZ Shine Standalone\" class=\"image-product\" \/><\/a><div class=\"content\"><p><a href=\"https:\/\/ergodox-ez.com\/collections\/frontpage\/products\/ergodox-ez-shine-standalone\" title=\"ErgoDox EZ Shine Standalone\" class=\"title\">ErgoDox EZ Shine Standalone<\/a>\n              \n                \n                  \n  \n    \n      \n    \n    \n    $299.00\n  \n                \n                \n                \n                   (free shipping)<\/p><\/div><\/article><p><a href=\"https:\/\/ergodox-ez.com\/collections\/accessories\/\" class=\"item\">Purchase standalone accessories<\/a><\/p><\/div><\/div>\n  \n            <\/main><footer><div class=\"hero-companies\"><div class=\"container\">Used and loved by thousands, including those at...<div class=\"grid\"><p><a class=\"button\" href=\"https:\/\/ergodox-ez.com\/pages\/testimonials\">See what people say<\/a><\/p><\/div><\/div><\/div>\n  \n  \n    \n      <div class=\"hero-crowdfund\"><div class=\"container\"><div class=\"content\">A<a href=\"https:\/\/www.indiegogo.com\/projects\/ergodox-ez-an-incredible-mechanical-keyboard-computers-health#\/\" class=\"link\">crowdfunding success<\/a>- over $400,000... and we shipped on time! :)<\/div><\/div><\/div>\n    \n  \n  \n  \n<\/footer><\/div>","created_at":"2018-11-07T16:23:21+0000","updated_at":"2018-11-07T16:23:21+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"ergodox-ez.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":138,"title":"Clojure Error Messages are Accidental - LispCast","url":"https:\/\/lispcast.com\/clojure-error-messages-accidental\/","content":"<p><em>I\u2019ve recently shifted my thinking about Clojure error messages. It is more useful to think of them as non-existent than to think of them as bad. We end with the role Spec can play in improving error messages.\u00a0<\/em><\/p><p>I have the good fortune of helping many people learn Clojure. One of the most common complaints is the bad error messages. I have been looking for different ways to improve the error messages for beginners. I\u2019ve tried better stack trace printers. I\u2019ve tried implementing some prototype solutions to catch bad arguments to functions. I\u2019ve been looking at other languages, particularly Elm, for inspiration. I\u2019ve also just been exploring how error messages are actually implemented in Clojure. What I\u2019ve discovered has been kind of surprising. I have a completely new perspective on error messages in Clojure, and I want to share that with you.<\/p><p>Before I go on, I want to make something clear: Clojure error messages are bad. They make it harder for beginners. I don\u2019t think anyone really argues with that. I like Clojure and think it\u2019s well designed in general. I wanted to figure out what was going on with error messages.<\/p><p>It will be important to distinguish between the error messages and the stack traces. The stack traces are those horribly long printouts you get when you get an exception thrown. It\u2019s one of those cases where Clojure is relying on the host platform. That\u2019s just what JVM stack traces look like.<\/p><p>The worst sin of stack traces is that the most important information is printed first, then followed by less and less important information, usually a few screenfuls of it. In essence, the stack traces are printed backwards in the terminal. You have to scroll way back, looking for the beginning of the trace. <strong>The default stack traces are ergonomically difficult.<\/strong><\/p><p>The second worst sin is that lots of details from Clojure\u2019s implementation is included in the stack trace, making them longer and noisier, and, frankly, more intimidating. Both sins suck. I lived with them for years. And they are easily solvable. I happen to use and like <a href=\"https:\/\/github.com\/AvisoNovate\/pretty\">Pretty<\/a>, which reverses the stack trace and filters out noise.<\/p><p>But both of these sins combined are not as bad as the Exceptions themselves. Let\u2019s dig into those.<\/p><p>There are actually two major types of errors in Clojure. There are compile-time errors and runtime errors. Compile-time errors are things like syntax errors, unresolvable symbols (missing variables), and errors during macro expansion. These errors tend to be decent, on average. The Exception type makes sense and the error message points to the problem.<\/p><pre class=\"lang-clojure\">(f)\n;; throws clojure.lang.Compiler$CompilerException: \n           java.lang.RuntimeException: \n              Unable to resolve symbol: f in this context, compiling:(*cider-repl pftv*:2780:15)<\/pre><p>It\u2019s not beautiful, but it\u2019s acceptable. Some Clojure macros had incomprehensible error messages, but that seems to have gotten better. It used to be:<\/p><pre>(defn foo a)\n;; throws IllegalArgumentException:\n            Don't know how to create ISeq from clojure.lang.Symbol, compiling:(*cider-repl pftv*:2780:15)<\/pre><p>It seems like Clojure just assumed the parameter list was a <a href=\"https:\/\/purelyfunctional.tv\/guide\/clojure-collections\/#vector\">vector<\/a> and tried to iterate over it. Now (Clojure 1.8 even) it\u2019s much nicer:<\/p><pre>(defn foo a)\n;; throws\u00a0java.lang.IllegalArgumentException:\n            Parameter declaration \"a\" should be a vector<\/pre><p>Runtime errors, on the other hand, are not as good. There has never seemed to be any consistency to them. The Exception type is sometimes correct, but sometimes it seems to have nothing to do with what I am doing. The error messages rarely describe the actual problem. Here\u2019s an example where the error message is just wrong:<\/p><pre>(def my-val 1)\n@my-val\n;; throws java.lang.ClassCastException: \n            java.lang.Long cannot be cast to java.util.concurrent.Future<\/pre><p>Who said anything about a Future? Why is Clojure bringing that up? You don\u2019t know how many times I\u2019ve looked for Futures in my code, only to realize I had a stray @ somewhere.<\/p><p>And what\u2019s more, very often, no error is thrown at all. Maybe <code>nil<\/code> is returned, or some other unexpected value. Here\u2019s an example:<\/p><pre>(keyword 5)   ;=&gt; nil\n(keyword nil) ;=&gt; nil<\/pre><p>That\u2019s not very helpful. Try passing different types to <code>keyword<\/code> to see what happens. Wouldn\u2019t you expect an error?<\/p><p>Before I started exploring, these behaviors seemed lazy and neglectful. However, they have been something I\u2019ve learned to live with. Sometimes the error is something I\u2019ve seen before. But more frequently, I don\u2019t even read the error message. I work in small increments. Any error must be somewhere in the code I just wrote. I didn\u2019t pay much attention to error message design until I started exploring it more.<\/p><p>Through that exploration, I\u2019ve realized that <strong>it\u2019s not so much that Clojure\u2019s errors are bad. It\u2019s more that they\u2019re accidental.<\/strong> Clojure\u2019s core functions are, for the most part, implemented without checks on the arguments. They only code the \u201chappy path\u201d. They assume that the arguments are of the correct type and shape, and proceed without caution. When they do explicitly throw an Exception, it\u2019s deep inside a conditional where there clearly isn\u2019t a way to proceed.<\/p><p>What happens if they\u2019re not correct? That\u2019s up to chance. Sometimes you get an Exception from something that does implicitly check its arguments. For example, trying to call any method on <code>nil<\/code> will throw an Exception. And numeric functions will die on non-numbers. But sometimes, after trying all of the tests in a conditional, the conditional fall through to <code>nil<\/code>. That\u2019s <a href=\"https:\/\/github.com\/clojure\/clojure\/blob\/clojure-1.9.0\/src\/clj\/clojure\/core.clj#L606\">what happens with <code>keyword<\/code><\/a>. It appears that it worked, when nothing was done at all, and <code>nil<\/code> is returned.<\/p><p>I\u2019ve come to believe that Clojure\u2019s errors aren\u2019t bad by design. No, it\u2019s something totally different. Clojure\u2019s errors are actually missing.<\/p><p>We\u2019re used to languages doing runtime argument checks. One robust way to implement checks is to implement them in the most central core of the language. Then you build new functions on top of that core. Those new functions can choose to check their own arguments and throw meaningful errors, or rely on on the underlying core\u2019s errors if they\u2019re sufficient (or you\u2019re lazy). It\u2019s a systematic way to ensure that bad runtime behavior throws an error. Even a language like JavaScript will eventually bottom out with \u201cundefined is not a function\u201d (meaning method not found) or a null pointer exception. Even if the errors are bad like JavaScript\u2019s, at least they exist. That\u2019s one way to do it, but that\u2019s not what Clojure does.<\/p><p>What has Clojure done?<\/p><p>Clojure, in typical <a href=\"http:\/\/www.infoq.com\/presentations\/Simple-Made-Easy\">\u201cde-complecting\u201d style<\/a>, has separated out the implementation of a function from enforcing the preconditions of that function. We have the implementations, which are the \u201chappy paths\u201d. What is missing\u2013and has always been missing\u2013are the preconditions. It\u2019s almost as if Clojure\u2019s functions\u2019 implementations assumed some kind of external check on their arguments would happen.<\/p><p>Enter Spec. Spec is that external check. Spec is not there to make the error messages better, as is widely believed. Spec is there to have error messages at all. Since the runtime error messages you see are accidental, any consistently applied error checks will be beneficial. Once we have error messages, we can begin the work of making them better.<\/p><h2>Speculation<\/h2><p>There\u2019s a concern that Spec may make error messages worse, and those concerned point to some macros\u2019 error messages where it actually did get worse once the macros were Specced. This is a valid concern. However, I believe that it is far worse to have accidental error messages (like we have now) than consistent error messages, even if they\u2019re bad. Spec\u2019s messages may be bad (I don\u2019t like them that much; they remind me a lot of Haskell\u2019s type error messages), but they will cover the core functions of the language. Functions that call those will also get error messages. I used to be skeptical of the benefit of Spec, but now I\u2019m looking forward to its release.<\/p><p>I think Spec is going to surprise us. Once <a href=\"https:\/\/github.com\/clojure\/core.specs.alpha\">Clojure\u2019s core functions are specced<\/a>, we will be surprised by how much of our code violates the assumptions of the functions we use, but somehow worked anyway by accident. For instance, I can imagine <code>nil<\/code>s flowing through functions like <code>get<\/code> that seem to tolerate them and returning <code>nil<\/code> themselves. Since <code>nil<\/code> is a valid return from <code>get<\/code>, we handle it and it may work out. But is <code>nil<\/code> really a valid argument to <code>get<\/code>? Probably not, and that would be reflected in the Spec. We\u2019ve probably got a lot of code like that. When we turn on Spec instrumentation and run our tests, we\u2019ll have to face all of these violated assumptions that happened to work. There will be many errors in our existing code.<\/p><p>Even those of us who have worked in Clojure for a long time will have to internalize those assumptions. We\u2019re not used to a pre-condition checks. We\u2019re used to thinking of Clojure\u2019s functions as loose and dynamically typed. Before we internalize the logic of Spec, we will write code that doesn\u2019t pass the spec the first time. We will be like beginners. And some may not like the language that it will have become because it will be picky in unfamiliar ways. Lucky for them, they can completely turn it off. But we will be better off with instrumentation on.<\/p><p>I think we will be equally surprised by how simple the types are. For example, it\u2019s easy to see <code>clojure.set\/union<\/code> as a really complex function. It takes two <a href=\"https:\/\/purelyfunctional.tv\/guide\/clojure-collections\/\">collections<\/a> and returns a new collection with all of the elements of both collections. However, sometimes the type it returns is from the first argument, and sometimes from the second argument, depending on which one is bigger. That\u2019s so complicated.<\/p><pre>(clojure.set\/union #{0} [1 2 3 4])   ;=&gt; [1 2 3 4 0]\n(clojure.set\/union #{0 1 2 3} [4])   ;=&gt; #{0 1 4 3 2}\n(clojure.set\/union '(0 1 2 3 4) [5]) ;=&gt; (5 0 1 2 3 4)<\/pre><p>When we first use <code>union<\/code>, we are trying to apply rules that make sense in other areas of Clojure. Shouldn\u2019t the arguments be coerced to Sets, like all collections are coerced to seqs in the <a href=\"https:\/\/purelyfunctional.tv\/guide\/clojure-collections\/#subvectors\">sequence<\/a> functions? Or shouldn\u2019t it really <em>always<\/em> use the type of the first argument, like in protocol methods? And there can\u2019t be something wrong with the types, otherwise it would have thrown an error, right?<\/p><p>But these rules don\u2019t apply. All of the <code>union<\/code> calls above are wrong. The fact that they do anything at all is an accident. The type of <code>clojure.set\/union<\/code> is so simple. It simply assumes all of the arguments are Sets. If that\u2019s true, it will always return a Set. In Haskell, you\u2019d say <code>Set -&gt; Set -&gt; Set<\/code>. And that type does work with Clojure\u2019s <code>union<\/code>. We can predict that Spec will choose exactly that type. Most experienced Clojure programmers will agree to that. So what was once complex behavior will be replaced by simple behavior or an error message.<\/p><p>But I think there will be some cases where we don\u2019t agree, or at least we will have to rebuild our understanding.<\/p><p>Finally, there will be times when speccing a function will inform the implementation. Currently, the <a href=\"https:\/\/twitter.com\/ericnormand\/status\/996770936547414016\">type for clojure.core\/nth<\/a> is quite complicated. There are eight different types that it can take. And there\u2019s no nice abstraction that covers them all. I mentioned this on Twitter and <a href=\"https:\/\/twitter.com\/puredanger\/status\/996831977901842433\">Alex Miller pointed out<\/a> that we might need a new notion here, perhaps called <code>nthable?<\/code>, which would simplify the spec.<\/p><p>The least obvious consequence of my new perspective (that we are adding error messages where none really existed) is that we have an opportunity to correct some mistakes in the implementation. For example, should you really be able to <code>get<\/code> out of anything?<\/p><pre>(get (java.io.File. \"hello.txt\") :foo) ;=&gt; nil<\/pre><p>Does that make any sense? Is that the expected behavior? Maybe we should stop wondering about the expected behavior when you pass in garbage input, and talk instead about the expected input. Spec will let us talk about that.<\/p><p>Maybe Rich has learned a lot since a lot of these functions were defined ten years ago and he wants to use the release of Spec as a way to correct those mistakes without breaking backwards compatibility. Or maybe the core specs will be extremely conservative. It wouldn\u2019t surprise me either way. But if the official spec doesn\u2019t restrict <code>get<\/code>, Spec will let people redefine the spec for <code>get<\/code> for themselves. They can choose whatever subset of the type they want. I can imagine many companies adopting strict specs that catch bugs they find in their code, like some companies <a href=\"https:\/\/github.com\/airbnb\/javascript\/tree\/master\/packages\/eslint-config-airbnb\">release their linter configs<\/a> today.<\/p><p>In fact, you can do this right now in your code. If <code>get<\/code> is a source of bugs for you, define a spec for it, instrument, and run your tests.<\/p><p>Finally, this new perspective has helped me understand Cognitect\u2019s stance on the error message issue. People have been complaining about error messages for years, and the response from Cognitect has been alienating. Perhaps I\u2019m dense, but it hasn\u2019t been until I made this mental shift that I\u2019ve understood some of what they\u2019re getting at.<\/p><h2>Conclusions<\/h2><p>I\u2019m finding it useful to see Clojure error messages as <em>missing<\/em>. In practice, they\u2019re still bad, but this perspective helps me understand why and gives Spec more meaning. Clojure\u2019s core specs, where every function has a spec, will finally give us error messages. However they are, they are better than what we have now. Will core specs also uncover lots of problems in our existing code? Will core specs change our understanding of the difference between runtime behavior and valid input? Who knows. But I\u2019m looking forward to a release of clojure.core.specs. It will contain lots of insights into Rich\u2019s design and about how to best understand Clojure. And once we have error messages, we can finally begin the work of making them better.<\/p>","created_at":"2018-11-07T16:23:21+0000","updated_at":"2018-11-07T16:23:21+0000","published_at":"2018-06-28T16:20:08+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":12,"domain_name":"lispcast.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":133,"title":"Ask HN: Who are some unsung heroes in open source that need more support?","url":"https:\/\/news.ycombinator.com\/item?id=17445280","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=koala_man\" class=\"hnuser\">koala_man<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446058\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=voltagex_\" class=\"hnuser\">voltagex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446624\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ymse\" class=\"hnuser\">ymse<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446696\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmt\" class=\"hnuser\">mmt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446981\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=voltagex_\" class=\"hnuser\">voltagex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17447335\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmt\" class=\"hnuser\">mmt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17447640\">4 months ago<\/a><\/p><br \/><div class=\"comment\">&gt; There's a big gap between making a working .deb file and<p>I assume none of us is even talking about this case. That is, a trivially working .deb could provide little more functionality than a tarball, so that's hardly interesting.<\/p><p>&gt; getting something to pass QA to get into the archive.<\/p><p>The last part I have no knowledge of, since I only ever put my packages into internal archives. Were there political aspects, or only technical?<\/p><p>For the technical, I'm still looking for (ideally from the GP, but from anyone who shares the opinion), <i>specifics<\/i>.<\/p><p>&gt; the documentation was fragmented and not up to date.<\/p><p>This is a common complaint I hear\/read, and I recall it was part of my initial learning investment, figuring out where to go for what information. Now it's mostly bookmarked, and changes tend not to be drastic.<\/p><p>&gt; Things have gotten better now you can use git-buildpackage<\/p><p>One thing I found, repeatedly and consistently, was that every single external (i.e. not from Debian) utility that tried to make the process \"better\" or \"easier\" ultimately did the opposite, since it would hide or abstract away an important aspect of the packaging process.<\/p><p>&gt; Challenge - take a package like wget and try to update it to the latest version from upstream Git (or whatever). It should be one or two commands but I've never been able to make it work easily.<\/p><p>I can't recall if I've done wget specifically, but I've done this kind of thing before without much issue, <i>assuming<\/i> that latest version actually builds and runs on that platform without porting work and without needing a ton of new dependencies (or new versions of existing ones).<\/p><p>I tended to find most of my time was spent in chasing down and re-packaging various libraries whose older versions were no longer good enough.<\/p><p>&gt; It's even harder if you're trying to port something across from Ubuntu<\/p><p>I'm not sure what you mean, since Ubuntu is already using Debian packaging.<\/p><p>&gt;  or to use a fork like wget-lua<\/p><p>Certainly forks are going to be more effort, but my experience is that this is true even in their unpackaged state, if they require more exotic (or specific) dependencies be pre-installed, a custom build process, or any porting work. If someone had already done all that work, comprehensively documented it, but merely not translated it into debianization, it would save me a ton of time in creating a custom package.\n              <\/p><\/div><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=voltagex_\" class=\"hnuser\">voltagex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17447971\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmt\" class=\"hnuser\">mmt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17448304\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdc\" class=\"hnuser\">jdc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17450786\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AnIdiotOnTheNet\" class=\"hnuser\">AnIdiotOnTheNet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446522\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kosma\" class=\"hnuser\">kosma<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446584\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=probonopd\" class=\"hnuser\">probonopd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17452513\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jessaustin\" class=\"hnuser\">jessaustin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446697\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=belorn\" class=\"hnuser\">belorn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17448278\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=probonopd\" class=\"hnuser\">probonopd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17452553\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmt\" class=\"hnuser\">mmt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17453236\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mmt\" class=\"hnuser\">mmt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446867\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=heyoni\" class=\"hnuser\">heyoni<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446565\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=probonopd\" class=\"hnuser\">probonopd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17452562\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jquast\" class=\"hnuser\">jquast<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446590\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jancsika\" class=\"hnuser\">jancsika<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446632\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aaronarduino\" class=\"hnuser\">aaronarduino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445492\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dvirsky\" class=\"hnuser\">dvirsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445543\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aaronarduino\" class=\"hnuser\">aaronarduino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445598\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dvirsky\" class=\"hnuser\">dvirsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445626\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aaronarduino\" class=\"hnuser\">aaronarduino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445673\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j45\" class=\"hnuser\">j45<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445535\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dvirsky\" class=\"hnuser\">dvirsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445553\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=j45\" class=\"hnuser\">j45<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446432\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aaronarduino\" class=\"hnuser\">aaronarduino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445631\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=koala_man\" class=\"hnuser\">koala_man<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17447034\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BoysenberryPi\" class=\"hnuser\">BoysenberryPi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445846\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tribidi\" class=\"hnuser\">tribidi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446098\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sincerely\" class=\"hnuser\">sincerely<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446437\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=georgecalm\" class=\"hnuser\">georgecalm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445878\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TAForObvReasons\" class=\"hnuser\">TAForObvReasons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446406\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=FLGMwt\" class=\"hnuser\">FLGMwt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446312\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swyx\" class=\"hnuser\">swyx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446465\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sheetjs\" class=\"hnuser\">sheetjs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445880\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=toufka\" class=\"hnuser\">toufka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445777\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dbrgn\" class=\"hnuser\">dbrgn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445860\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cbHXBY1D\" class=\"hnuser\">cbHXBY1D<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446175\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=phkahler\" class=\"hnuser\">phkahler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446708\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cbHXBY1D\" class=\"hnuser\">cbHXBY1D<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17451868\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nategri\" class=\"hnuser\">nategri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445428\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bradknowles\" class=\"hnuser\">bradknowles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445718\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bradknowles\" class=\"hnuser\">bradknowles<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445994\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=robertk\" class=\"hnuser\">robertk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445775\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nkkollaw\" class=\"hnuser\">nkkollaw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445901\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cperciva\" class=\"hnuser\">cperciva<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446136\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wafflesindeed\" class=\"hnuser\">wafflesindeed<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446231\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cperciva\" class=\"hnuser\">cperciva<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17448122\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dingaling\" class=\"hnuser\">dingaling<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17449621\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nategri\" class=\"hnuser\">nategri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445810\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sldenazis\" class=\"hnuser\">sldenazis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445609\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ravenstine\" class=\"hnuser\">ravenstine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445542\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dvirsky\" class=\"hnuser\">dvirsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445568\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bufferoverflow\" class=\"hnuser\">bufferoverflow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445927\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ravenstine\" class=\"hnuser\">ravenstine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446122\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tvlooy\" class=\"hnuser\">tvlooy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445965\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neom\" class=\"hnuser\">neom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445419\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brynet\" class=\"hnuser\">brynet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445784\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brynet\" class=\"hnuser\">brynet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445828\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=busterarm\" class=\"hnuser\">busterarm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446153\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brynet\" class=\"hnuser\">brynet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446301\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrburton\" class=\"hnuser\">mrburton<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446462\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=willow9886\" class=\"hnuser\">willow9886<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445850\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ljw1001\" class=\"hnuser\">ljw1001<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446556\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brutus1213\" class=\"hnuser\">brutus1213<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446980\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dualbus\" class=\"hnuser\">dualbus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446869\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blzaugg\" class=\"hnuser\">blzaugg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446208\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TAForObvReasons\" class=\"hnuser\">TAForObvReasons<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446234\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sam0x17\" class=\"hnuser\">sam0x17<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446570\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=petrohi\" class=\"hnuser\">petrohi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445887\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=makx\" class=\"hnuser\">makx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445882\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=inspector14\" class=\"hnuser\">inspector14<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445563\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445633\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zzzeek\" class=\"hnuser\">zzzeek<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445911\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kondro\" class=\"hnuser\">kondro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446781\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=warp\" class=\"hnuser\">warp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17449985\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nagVenkat\" class=\"hnuser\">nagVenkat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445402\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fiveFeet\" class=\"hnuser\">fiveFeet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445487\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=poster123\" class=\"hnuser\">poster123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445545\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ljw1001\" class=\"hnuser\">ljw1001<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446577\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=poster123\" class=\"hnuser\">poster123<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17450750\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ljw1001\" class=\"hnuser\">ljw1001<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17454498\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jhasse\" class=\"hnuser\">jhasse<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17448956\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dnautics\" class=\"hnuser\">dnautics<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445547\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nunez\" class=\"hnuser\">nunez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446979\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=porsager\" class=\"hnuser\">porsager<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445914\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=panoply\" class=\"hnuser\">panoply<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17447051\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dvirsky\" class=\"hnuser\">dvirsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445458\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dvirsky\" class=\"hnuser\">dvirsky<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445513\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reeme\" class=\"hnuser\">reeme<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446683\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sandebert\" class=\"hnuser\">sandebert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17447812\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zacharycarter\" class=\"hnuser\">zacharycarter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445977\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=electricslpnsld\" class=\"hnuser\">electricslpnsld<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446486\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bckmn\" class=\"hnuser\">bckmn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445955\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mabynogy\" class=\"hnuser\">mabynogy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17454186\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=threeseed\" class=\"hnuser\">threeseed<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446274\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=justinclift\" class=\"hnuser\">justinclift<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17477866\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kartiksura\" class=\"hnuser\">kartiksura<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446191\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=phkahler\" class=\"hnuser\">phkahler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445593\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=voltagex_\" class=\"hnuser\">voltagex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446614\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JdeBP\" class=\"hnuser\">JdeBP<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17449031\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=voltagex_\" class=\"hnuser\">voltagex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17449948\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=settings11\" class=\"hnuser\">settings11<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17460433\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=haddadda\" class=\"hnuser\">haddadda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445822\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=notatoad\" class=\"hnuser\">notatoad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445830\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ljw1001\" class=\"hnuser\">ljw1001<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446497\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=willow9886\" class=\"hnuser\">willow9886<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445944\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hutzlibu\" class=\"hnuser\">hutzlibu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446062\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jstarfish\" class=\"hnuser\">jstarfish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17446007\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=justinclift\" class=\"hnuser\">justinclift<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17477875\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brynet\" class=\"hnuser\">brynet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445867\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rajeshbatralive\" class=\"hnuser\">rajeshbatralive<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445398\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sattoshi\" class=\"hnuser\">sattoshi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17445983\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rajeshbatra\" class=\"hnuser\">rajeshbatra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17449690\">4 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:23:20+0000","updated_at":"2018-11-07T16:23:20+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":134,"title":"The advantages of an email-driven git workflow","url":"https:\/\/drewdevault.com\/2018\/07\/02\/Email-driven-git.html","content":"<p><a href=\"https:\/\/raw.githubusercontent.com\/git\/git\/master\/Documentation\/RelNotes\/2.18.0.txt\">git 2.18.0<\/a> has been released, and with it my first contribution to\ngit has shipped! My patch was for a git feature which remains disappointingly\nobscure: <a href=\"https:\/\/git-scm.com\/docs\/git-send-email\">git send-email<\/a>. I want to\nintroduce my readers to this feature and speak to the benefits of using an\nemail-driven git workflow - the workflow git was originally designed for.<\/p><p>Email isn\u2019t as sexy as GitHub (and its imitators), but it has several\nadvantages over the latter. Email is standardized, federated, well-understood,\nand venerable. A very large body of email-related software exists and is equally\nreliable and well-understood. You can interact with email using only open source\nsoftware and customize your workflow at every level of the stack - filtering,\norganizing, forwarding, replying, and so on; in any manner you choose.<\/p><p>Git has several built-in tools for leveraging email. The first one of note is\n<a href=\"https:\/\/git-scm.com\/docs\/git-format-patch\">format-patch<\/a>. This can take a git\ncommit (or series of commits) and format them as plaintext emails with embedded\ndiffs. Here\u2019s a small example of its output:<\/p><pre class=\"language-mail\">From 8f5045c871c3060ff5f5f99ce1ada09f4b4cd105 Mon Sep 17 00:00:00 2001\nFrom: Drew DeVault &lt;sir@cmpwn.com&gt;\nDate: Wed, 2 May 2018 08:59:27 -0400\nSubject: [PATCH] Silently ignore touch_{motion,up} for unknown ids\n---\n types\/wlr_seat.c | 2 --\n 1 file changed, 2 deletions(-)\ndiff --git a\/types\/wlr_seat.c b\/types\/wlr_seat.c\nindex f77a492d..975746db 100644\n--- a\/types\/wlr_seat.c\n+++ b\/types\/wlr_seat.c\n@@ -1113,7 +1113,6 @@ void wlr_seat_touch_notify_up(struct wlr_seat *seat, uint32_t time,\n \tstruct wlr_seat_touch_grab *grab = seat-&gt;touch_state.grab;\n \tstruct wlr_touch_point *point = wlr_seat_touch_get_point(seat, touch_id);\n \tif (!point) {\n-\t\twlr_log(L_ERROR, \"got touch up for unknown touch point\");\n \t\treturn;\n \t}\n \n@@ -1128,7 +1127,6 @@ void wlr_seat_touch_notify_motion(struct wlr_seat *seat, uint32_t time,\n \tstruct wlr_seat_touch_grab *grab = seat-&gt;touch_state.grab;\n \tstruct wlr_touch_point *point = wlr_seat_touch_get_point(seat, touch_id);\n \tif (!point) {\n-\t\twlr_log(L_ERROR, \"got touch motion for unknown touch point\");\n \t\treturn;\n \t}\n \n-- \n2.18.0\n<\/pre><p>git format-patch is at the bottom of git\u2019s stack of outgoing email features. You\ncan send the emails it generates manually, but usually you\u2019ll use git send-email\ninstead. It logs into the SMTP server of your choice and sends the email for\nyou, after running git format-patch for you and giving you an opportunity to\nmake any edits you like. Given that most popular email clients these days are\nawful and can\u2019t handle basic tasks like \u201csending email\u201d properly, I strongly\nrecommend this tool over attempting to send format-patch\u2019s output yourself.<\/p><p><img src=\"https:\/\/sr.ht\/wmKv.jpg\" alt=\"image\" \/><\/p><p>\n    <em>\n        I put a notch in my keyboard for each person who ignores my advice,\n        struggles through sending emails manually, and eventually comes around\n        to letting git send-email do it for them.\n    <\/em>\n<\/p><p>I recommend a few settings to apply to git send-email to make your workflow a\nbit easier. One is <code class=\"highlighter-rouge\">git config --global sendemail.verify off<\/code>, which turns off\na sometimes-annoying and always-confusing validation step which checks for\nfeatures only supported by newer SMTP servers - newer, in this case, meaning\nmore recent than November of 1995. I started a thread on the git mailing list\nthis week to discuss changing this option to off by default.<\/p><p>You can also set the default recipient for a given repository by using a local\ngit config: <code class=\"highlighter-rouge\">git config sendemail.to admin@example.org<\/code>. This lets you skip a\nstep if you send your patches to a consistent destination for that project, like\na mailing list. I also recommend <code class=\"highlighter-rouge\">git config --global sendemail.annotate yes<\/code>,\nwhich will always open the emails in your editor to allow you to make changes\n(you can get this with <code class=\"highlighter-rouge\">--annotate<\/code> if you don\u2019t want it every time).<\/p><p>The main edit you\u2019ll want to make when annotating is to provide what some call\n\u201ctimely commentary\u201d on your patch. Immediately following the <code class=\"highlighter-rouge\">---<\/code> after your\ncommit message, you can add a summary of your changes which can be seen by the\nrecipient, but doesn\u2019t appear in the final commit log. This is a useful place to\ntalk about anything useful regarding the testing, review, or integration of your\nchanges. You may also want to edit the <code class=\"highlighter-rouge\">[PATCH]<\/code> text in the subject line to\nsomething like <code class=\"highlighter-rouge\">[PATCH v2]<\/code> - this can also be done with the <code class=\"highlighter-rouge\">-v<\/code> flag as well.\nI also like to add additional To\u2019s, Cc\u2019s, etc at this time.<\/p><p>Git also provides tools for the recipient of your messages. One such tool is\n<a href=\"https:\/\/git-scm.com\/docs\/git-am\">git am<\/a>, which accepts an email prepared with\nformat-patch and integrates it into their repository. Several flags are provided\nto assist with common integration activities, like signing off on the commit or\nattempting a 3-way merge. The difficult part can be getting the email to git am\nin the first place. If you simply use the GMail web UI, this can be difficult. I\nuse <a href=\"http:\/\/www.mutt.org\/\">mutt<\/a>, a TUI email client, to manage incoming\npatches. This is useful for being able to compose replies with vim rather than\nfighting some other mail client to write emails the way I want, but more\nimportantly it has the <code class=\"highlighter-rouge\">|<\/code> key, which prompts you for a command to pipe the\nemail into. Other tools like <a href=\"http:\/\/www.offlineimap.org\/\">OfflineIMAP<\/a> are also\nuseful here.<\/p><p>On the subject of composing replies, reviewing patches is quite easy with the\nemail approach as well. Many bad, yet sadly popular email clients have\npopularized the idea that the sender\u2019s message is immutable, encouraging you to\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Posting_style#Top-posting\">top post<\/a> and leave an endlessly growing chain of replies\nunderneath your message. A secret these email clients have kept from you is that\nyou are, in fact, permitted by the mail RFCs to edit the sender\u2019s message as you\nplease when replying - a style called <a href=\"https:\/\/en.wikipedia.org\/wiki\/Posting_style#Bottom-posting\">bottom posting<\/a>. I\nstrongly encourage you to get comfortable doing this in general, but it\u2019s\nessential when reviewing patches received over email.<\/p><p>In this manner, you can dissect the patch and respond to specific parts of it\nrequesting changes or clarifications. It\u2019s just email - you can reply, forward\nthe message, Cc interested parties, start several chains of discussion, and so\non. I recently sent the following feedback on a patch I received:<\/p><pre class=\"language-mail\">Date: Mon, 11 Jun 2018 14:19:22 -0400\nFrom: Drew DeVault &lt;sir@cmpwn.com&gt;\nTo: Gregory Mullen &lt;omitted&gt;\nSubject: Re: [PATCH 2\/3 todo] Filter private events from events feed\nOn 2018-06-11  9:14 AM, Gregory Mullen wrote:\n&gt; diff --git a\/todosrht\/alembic\/versions\/cb9732f3364c_clear_defaults_from_tickets_to_support_.py b\/todosrht\/alembic\/versions\/cb9732f3364c_clear_defaults_from_tickets_to_support_.py\n&gt; -%&lt;-\n&gt; +class FlagType(types.TypeDecorator):\nI think you can safely import the srht FlagType here without implicating\nthe entire sr.ht database support code\n&gt; diff --git a\/todosrht\/blueprints\/html.py b\/todosrht\/blueprints\/html.py\n&gt; -%&lt;-\n&gt; +def collect_events(target, count):\n&gt; +    events = []\n&gt; +    for e in EventNotification.query.filter(EventNotification.user_id == target.id).order_by(EventNotification.created.desc()):\n80 cols\nI suspect this 'collect_events' function can be done entirely in SQL\nwithout having to process permissions in Python and do several SQL\nround-trips\n&gt;  @html.route(\"\/~&lt;username&gt;\")\n&gt;  def user_GET(username):\n&gt; -    print(username)\nWhoops! Nice catch.\n&gt;      user = User.query.filter(User.username == username.lower()).first()\n&gt;      if not user:\n&gt;          abort(404)\n&gt;      trackers, _ = get_tracker(username, None)\n&gt;      # TODO: only show public events (or events the current user can see)\nCan remove the comment\n<\/pre><p>Obviously this isn\u2019t the whole patch we\u2019re seeing - I\u2019ve edited it down to just\nthe parts I want to talk about. I also chose to leave the file names in to aid\nin navigating my feedback, with casual <code class=\"highlighter-rouge\">-%&lt;-<\/code> symbols indicating where I had\ntrimmed out parts of the patch. This approach is common and effective.<\/p><p>The main disadvantage of email driven development is that some people are more\ncomfortable working with email in clients which are not well-suited to this kind\nof work. Popular email clients have caused terrible ideas like HTML email to\nproliferate, not only enabling spam, privacy leaks, and security\nvulnerabilities, but also making it more difficult for people to write emails\nthat can be understood by git or tolerated by advanced email users.<\/p><p>I don\u2019t think that the solution to these problems is to leave these powerful\ntools hanging in the wind and move to less powerful models like GitHub\u2019s pull\nrequests. This is why on my own platform, <a href=\"https:\/\/sr.ht\">sr.ht<\/a>, I chose to\nembrace git\u2019s email-driven approach, and extend it with new tools that make it\neasier to participate without directly using email. For those like me, I still\nwant the email to be there so you can dig my heels in and do it old-school, but\nI appreciate that it\u2019s not for everyone.<\/p><p>I started working on the sr.ht mailing list service a couple of weeks ago, which\nis where these goals will be realized with new email-driven code review tools.\nMy friend <a href=\"https:\/\/emersion.fr\">Simon<\/a> has been helping out with a Python module\nnamed <a href=\"https:\/\/git.sr.ht\/~emersion\/python-emailthreads\/\">emailthreads<\/a> which can\nbe used to parse email discussions - with a surprising degree of accuracy,\nconsidering the flexibility of email. Once I get these tools into a usable\nstate, we\u2019ll likely see sr.ht registrations finally opened to the general public\n(interested in trying it earlier? <a href=\"mailto:sir@cmpwn.com\">Email me<\/a>). Of course,\nit\u2019s all <a href=\"https:\/\/git.sr.ht\/~sircmpwn\/?search=sr.ht\">open source<\/a>, so you can\nfollow along and try it on your own infrastructure if you like.<\/p><p>Using email for git scales extremely well. The canonical project, of course, is\nthe Linux kernel. A change is made to the Linux kernel an average of 7 times per\nhour, constantly. It is maintained by dozens of veritable clans of software\nengineers hacking on dozens of modules, and email allows these changes to\nefficiently flow code throughout the system. Without email, Linux\u2019s maintenance\nmodel would be impossible. It\u2019s worth noting that git was designed for\nmaintaining Linux, of course.<\/p><p>With the right setup, it\u2019s well suited to small projects as well. Sending a\npatch along for review is a single git command. It lands directly in the\nmaintainer\u2019s inbox and can be integrated with a handful of keystrokes. All of\nthis works without any centralization or proprietary software involved. We\nshould embrace this!<\/p><hr \/><p>Related articles sent in by readers:<\/p><p><a href=\"https:\/\/begriffs.com\/posts\/2018-06-05-mailing-list-vs-github.html\">Mailing lists vs Github<\/a>\nby Joe Nelson<\/p><p><a href=\"https:\/\/dpc.pw\/blog\/2017\/08\/youre-using-git-wrong\/\">You\u2019re using git wrong<\/a> by\nDawid Ci\u0119\u017carkiewicz<\/p><hr \/><p>\n  Have a comment on one of my posts? Start a discussion in my\n  <a href=\"https:\/\/lists.sr.ht\/~sircmpwn\/public-inbox\">public inbox<\/a>\n  by sending an email to\n  <a href=\"mailto:~sircmpwn\/public-inbox@lists.sr.ht?Subject=Re%3A%20The%20advantages%20of%20an%20email-driven%20git%20workflow\">\n    ~sircmpwn\/public-inbox@lists.sr.ht\n  <\/a>\n  <small>\n    [<a href=\"https:\/\/man.sr.ht\/lists.sr.ht\/etiquette.md\">mailing list etiquette<\/a>]\n  <\/small>\n<\/p>","created_at":"2018-11-07T16:23:20+0000","updated_at":"2018-11-07T16:23:20+0000","published_at":"2018-07-02T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":9,"domain_name":"drewdevault.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":131,"title":"nbedos\/termtosvg","url":"https:\/\/github.com\/nbedos\/termtosvg","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/nbedos\/termtosvg\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/117a15a683e9f3156029fc3f5cb35800effa739f\/68747470733a2f2f7472617669732d63692e6f72672f6e6265646f732f7465726d746f7376672e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/nbedos\/termtosvg.svg?branch=master\" \/><\/a> <a href=\"https:\/\/circleci.com\/gh\/nbedos\/termtosvg\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/28be5f18fd01f63ea45f0930e5facff99a06c4c8\/68747470733a2f2f636972636c6563692e636f6d2f67682f6e6265646f732f7465726d746f7376672e7376673f7374796c653d737667\" alt=\"CircleCI\" data-canonical-src=\"https:\/\/circleci.com\/gh\/nbedos\/termtosvg.svg?style=svg\" \/><\/a><\/p>\n<p>termtosvg is a Unix terminal recorder written in Python that renders your command\nline sessions as standalone SVG animations.<\/p>\n<p style=\"text-align: center;\">\n    <a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/13f6d1b02a34a39a5836045d88ddac63cf7c48b0\/68747470733a2f2f6e6265646f732e6769746875622e696f2f7465726d746f7376672f6578616d706c65732f617765736f6d655f77696e646f775f6672616d652e737667\"><img src=\"https:\/\/camo.githubusercontent.com\/13f6d1b02a34a39a5836045d88ddac63cf7c48b0\/68747470733a2f2f6e6265646f732e6769746875622e696f2f7465726d746f7376672f6578616d706c65732f617765736f6d655f77696e646f775f6672616d652e737667\" data-canonical-src=\"https:\/\/nbedos.github.io\/termtosvg\/examples\/awesome_window_frame.svg\" alt=\"image\" \/><\/a>\n<\/p>\n<ul><li><a href=\"https:\/\/nbedos.github.io\/termtosvg\/pages\/examples.html\" rel=\"nofollow\">Gallery of examples<\/a><\/li>\n<li><a href=\"https:\/\/nbedos.github.io\/termtosvg\/pages\/templates.html\" rel=\"nofollow\">Gallery of templates<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/nbedos\/termtosvg\/blob\/develop\/man\/termtosvg.md\">Manual page<\/a><\/li>\n<\/ul><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<ul><li>Produce lightweight and clean looking animations embeddable on a project page<\/li>\n<li>Custom color themes, terminal UI and animation controls via <a href=\"https:\/\/github.com\/nbedos\/termtosvg\/blob\/develop\/man\/termtosvg-templates.md\">SVG templates<\/a><\/li>\n<li>Compatible with asciinema recording format<\/li>\n<\/ul><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>termtosvg is compatible with Linux, Mac OS and BSD OSes, requires Python &gt;= 3.5 and can be installed using pip:<\/p>\n<pre>pip3 install --user termtosvg\n<\/pre>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#basic-usage\"><\/a>Basic usage<\/h2>\n<p>Start recording with:<\/p>\n<pre>$ termtosvg\nRecording started, enter \"exit\" command or Control-D to end\n<\/pre>\n<p>You are now in a subshell where you can type your commands as usual.\nOnce you are done, exit the shell to end the recording:<\/p>\n<pre>$ exit\nRecording ended, file is \/tmp\/termtosvg_exp5nsr4.svg\n<\/pre>\n<p>Finally, use your favorite web browser to play the animation:<\/p>\n<pre>$ firefox \/tmp\/termtosvg_exp5nsr4.svg\n<\/pre>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>Dependencies<\/h2>\n<p>termtosvg uses:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/selectel\/pyte\">pyte<\/a> to render the terminal screen<\/li>\n<li><a href=\"https:\/\/github.com\/lxml\/lxml\">lxml<\/a> to work with SVG data<\/li>\n<\/ul><\/article>","created_at":"2018-11-07T16:23:18+0000","updated_at":"2018-11-07T16:23:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/33371440?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":132,"title":"ardour - the digital audio workstation","url":"https:\/\/ardour.org\/","content":"<header id=\"homepage\">\n<\/header><section id=\"activities\" class=\"unshaded\"><div class=\"container\"><div class=\"row\"><div class=\"span4 record\"><h2><img class=\"activity-icon\" src=\"https:\/\/ardour.org\/images\/icons\/record.png\" alt=\"\" \/>Record<\/h2><p>Bring it to the computer<\/p><p>\n          Grab yourself a nice audio interface. Plug in a microphone\n          or a keyboard. Add a track. Press record. Got it.\n        <\/p><p>Use what you've got<\/p><p>\n          Import audio or\n          MIDI from your hard drive or the Freesound database.\n        <\/p><\/div><div class=\"span4 edit\"><h2><img class=\"activity-icon\" src=\"https:\/\/ardour.org\/images\/icons\/edit.png\" alt=\"\" \/>Edit<\/h2><p>The work to be done<\/p><p>\n          Cut, move, stretch, copy, paste, delete, align, trim,\n          crossfade, rename, snapshot, zoom, transpose, quantize,\n          swing, drag, drop.\n        <\/p><p>The tools to do it<\/p><p>\n          All with unlimited undo\/redo. All in the same window.\n        <\/p><\/div><div class=\"span4 mix\"><h2><img class=\"activity-icon\" src=\"https:\/\/ardour.org\/images\/icons\/mix.png\" alt=\"\" \/>Mix<\/h2><p>All the buzzwords<\/p><p>Mute, solo, fader, automate, EQ, dynamics,\n          insert, send, pre-fader, post-fader, sync, monitor,\n          isolate.\n        <\/p><p>Less of the pain<\/p><p>Floating point fidelity, the\n            most flexible mixer architecture in the industry, hundreds of\n            plugins, and external control surfaces.\n          <\/p><\/div><\/div><p><\/p><h3>\u2026and then go on to master and export your work.<\/h3><\/div><\/section><section class=\"darkshaded\" id=\"users\"><section class=\"users film\" id=\"users-engineer\"><div class=\"container\"><div class=\"row\"><div class=\"span9 engineer\"><h2>Audio Engineers<\/h2><p>Ardour's core user group: people who want to record,\n          edit, mix and master audio and MIDI projects. When you need\n          complete control over your tools, when the limitations of\n          other designs get in the way, when you plan to spend hours\n          or days working on a session, Ardour is there to make\n          things work the way you want them to.<\/p><\/div><div class=\"span3 engineer-icon right\"><img src=\"https:\/\/ardour.org\/images\/mixer4.jpg\" alt=\"mixer\" \/><\/div><\/div><\/div><\/section><section class=\"users film\" id=\"users-musician\"><div class=\"container\"><div class=\"row\"><div class=\"span3 musician-icon left\"><img src=\"https:\/\/ardour.org\/images\/jazzdrums.jpg\" alt=\"jazzdrums\" \/><\/div><div class=\"span9 musician\"><h2>Musicians<\/h2><p>Being the best tool to record talented performers on actual\n          instruments has always been a top priority for\n          Ardour. Rather than being focused on electronic and pop music\n          idioms, Ardour steps out of the way to encourage the\n          creative process to remain where it always has been: a musician\n            playing a carefully designed and well built instrument.\n          <\/p><\/div><\/div><\/div><\/section><section class=\"users film\" id=\"users-film\"><div class=\"container\"><div class=\"row\"><div class=\"span9 film\"><h2>Soundtrack Editors<\/h2><p>Sample accurate sync and shared transport control with\n          video playback tools allows Ardour to provide a fast and\n          natural environment for creating and editing soundtracks\n          for film and video projects.<\/p><\/div><div class=\"span3 film-icon left\"><img src=\"https:\/\/ardour.org\/images\/icons\/filmslate.png\" alt=\"film clapper\" \/><\/div><\/div><\/div><\/section><section class=\"users film\" id=\"users-composer\"><div class=\"container\"><div class=\"row\"><div class=\"span3 composer-icon center\"><img src=\"https:\/\/ardour.org\/images\/edogkeys.png\" alt=\"piano\" \/><\/div><div class=\"span9 composer\"><h2>Composers<\/h2><p>Arrange audio and MIDI using the same tools and same\n            workflow. Use external hardware synthesizers or software\n            instruments as sound sources. From sound design to\n            electro-acoustic composition to dense multitrack MIDI\n            editing, Ardour can help.\n          <\/p><\/div><\/div><\/div><\/section><\/section><section class=\"unshaded\"><div class=\"container\"><div class=\"row\"><div class=\"span8\"><h2>Audio and MIDI Multi-Track Recording and Editing<\/h2><p>\n      Any number of tracks and busses. Non-linear\n      editing. Non-destructive (and destructive!) recording. Any bit\n      depth, any sample rate. Dozens of file formats.\n      <\/p><\/div><div class=\"span4\"><img src=\"https:\/\/ardour.org\/images\/fp-edits.png\" alt=\"\" \/><\/div><\/div><div class=\"row\"><div class=\"span4\"><img src=\"https:\/\/ardour.org\/images\/plugin-collage-small.png\" alt=\"\" \/><\/div><div class=\"span8\"><h2>Plugins with Full Sample Accurate Automation<\/h2><p>AudioUnit, LV2, LinuxVST and LADSPA formats. FX\n          plugins. Software instruments. MIDI processors. Automate any\n          parameters. Physically manipulate them via control surfaces.\n          Distribute processing across as many (or as few) cores as\n          you want.\n        <\/p><\/div><\/div><div class=\"row\"><div class=\"span8\"><h2>Transport Sync and External Control Surfaces<\/h2><p>\n          Best-in-industry sync to MIDI timecode and LTC. Send and\n          receive MIDI Machine Control. Sync with JACK transport and\n          MIDI clock. Dedicated Mackie Control protocol support,\n          pre-defined mappings for many MIDI controllers plus\n          dynamic MIDI learn. Use OSC to drive almost any operation in\n          Ardour.\n\t<\/p><\/div><div class=\"span4\"><img src=\"https:\/\/ardour.org\/images\/nucleus-small.jpg\" alt=\"\" \/><\/div><\/div><div class=\"row\"><div class=\"span4\"><img src=\"https:\/\/ardour.org\/images\/audiomatrixsmall.png\" alt=\"\" \/><\/div><div class=\"span8\"><h2>Powerful Anywhere-to-Anywhere Signal Routing<\/h2><p>\n\t  Complex signal flows are simple and elegant. Inputs and outputs\n\t  connect to your hardware and\/or other applications. Use\n\t  sends, inserts and returns freely. Connections can be\n\t  one-to-many, many-to-one or many-to-many. Tap signal flows \n\t  at any point. If you can't connect in the way you want with Ardour, it probably\n\t  can't be done.\n\t<\/p><\/div><\/div><div class=\"row\"><div class=\"span8\"><h2>Video Timeline<\/h2><p>\n    Import a single video and optionally extract the soundtrack from\n    it. Display a frame-by-frame (thumbnail) timeline of the video.\n    Use a Video-monitor window, or full-screen display, of the imported video in sync with any of the available ardour timecode sources.\n    Lock audio-regions to the video: Move audio-regions with the video at video-frame granularity.\n    Export the video, cut start\/end, add blank frames and\/or mux it with the soundtrack of the current-session.\n\t<\/p><\/div><div class=\"span4\"><img src=\"https:\/\/ardour.org\/images\/vtl.png\" alt=\"\" \/><\/div><\/div><\/div><\/section><section class=\"darkshaded\" id=\"open-source\"><div class=\"container\"><div class=\"row\"><p><\/p><h2>Ardour Is Open<\/h2>\n        <i class=\"general foundicon-globe\">\n          Ardour is an open source, collaborative effort of a worldwide team\n          including musicians, programmers, and professional recording\n          engineers. Development is transparent \u2014 anyone can\n          watch our work as it happens. Like a good piece of vintage\n          hardware, you can open the box and look inside. Of course,\n          you don't <em>have to<\/em> \u2026 but one day the fact\n          that anybody can will be useful.\n        \n          This openness forces a kind of integrity on the project that\n          is hard to find in proprietary software, and helps us to\n          focus on issues and features that matter to our users rather\n          than stuff that just looks good in advertisements.\n        <\/i><\/div><\/div><\/section>","created_at":"2018-11-07T16:23:18+0000","updated_at":"2018-11-07T16:23:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"ardour.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":129,"title":"Deriving the Y Combinator in JavaScript","url":"https:\/\/enlight.nyc\/projects\/y-combinator\/","content":"<h2>What is the Y Combinator?<\/h2><p>In order to understand the Y combinator, one must first understand recursion. Here\u2019s a definition.<\/p><blockquote>\n<p>recursion (noun)<br \/><em>see \u201crecursion\u201d<\/em><\/p>\n<\/blockquote><p>Kidding! But nonetheless, this definition captures the essence of recursion: self-reference. In our case \u2014 programming \u2014 recursion allows for the succint description of a repeating process.<\/p><p>Recursion can be puzzling at first, but oftentimes, it allows us to write code that is closer to the true intent of what we wish to do than an imperative, loopy style. Here are a few examples.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/**\n * Imperative style\n *\/\nfunction factorialA(x) {\n  let result = 1;\n  \/\/ What exactly is happening here?\n  for (let i = x; i--; i &gt;= 1) {\n    result *= x;\n  }\n  return x;\n}\n\/**\n * Recursive style\n *\/\nfunction factorialB(x) {\n  if (x === 0) {\n    return 1;\n  }\n  \/\/ Recursive call (function calls itself!)\n  \/\/ Stays true to definition of factorial: x times the factorial of x - 1\n  return x * factorial(x - 1);\n}<\/pre><\/div><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/**\n * Imperative style\n *\/\nfunction fibonacciA(n) {\n  \/\/ Term 0: 1, Term 1: 1, Term 2: 2\n  n = n - 1;\n  let prevTerm = 1;\n  let currTerm = 1;\n  \/\/ What does this loop do, exactly?\n  for (let i = 0; i &lt; n; i += 1) {\n    let temp = prevTerm;\n    prevTerm = currTerm;\n    currTerm += temp;\n  }\n  return currTerm;\n}\n\/**\n * Recursive style\n *\/\nfunction fibonacciB(n) {\n  if (n === 0 || n === 1) {\n    return 1;\n  }\n  \/\/ Definition of next Fibonacci term is sum of two previous\n  \/\/ Easy to follow!\n  return fibonacci(n - 2) + fibonacci(n - 1);\n}<\/pre><\/div><p>Many high-level languages, JavaScript included, support recursion. But some languages don\u2019t, and in the field of theoretical computer science, introducing recursion into non-recursive contexts is a fascinating problem.<\/p><p>This is where the Y combinator comes in: it provides a way to recurse \u2014 to self-reference \u2014 <em>even if self-reference is unsupported in the base language<\/em>!<\/p><h2>Eliminating Recursion<\/h2><p>The Y combinator bestows the power of self-referential recursion unto languages which don\u2019t support it. So to derive it in JavaScript, let\u2019s pretend a little \u2014 let\u2019s pretend that JavaScript doesn\u2019t support recursion.<\/p><p>I\u2019ve put the recursive factorial function given in the initial examples in the block below. If everything goes well, by the end of this writeup we\u2019ll have a tool that will allow this factorial function to self-reference again \u2014 but most importantly, without any self-reference in the function itself!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function factorial(x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * factorial(x - 1);\n}<\/pre><\/div><p>The best way to begin removing the recursion from <code class=\"language-text\">factorial<\/code> looks like factoring out the call to <code class=\"language-text\">factorial<\/code> from the function body and making it a parameter, so let\u2019s do that.<\/p><h3>Bringing Our Own<\/h3><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ Bring your own factorial function!\nfunction byoFactorial(f, x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * f(x - 1);\n}<\/pre><\/div><p>With this new factorial function, <code class=\"language-text\">byoFactorial<\/code>, we have to \u201cbring our own\u201d working <code class=\"language-text\">factorial<\/code> function <code class=\"language-text\">f<\/code> to pass into the new, non-recursive function as a parameter, to replace our old self-referential call.<\/p><p>Of course, the problem now is that we don\u2019t have a working <code class=\"language-text\">factorial<\/code> function anymore. But that doesn\u2019t mean we have to give up! We can start small \u2014 from zero.<\/p><p>Let\u2019s begin by creating a function <code class=\"language-text\">f<\/code>, that when passed to <code class=\"language-text\">byoFactorial<\/code>, will make it work, but only for <code class=\"language-text\">x = 0 (0! = 1)<\/code>.<\/p><blockquote>\n<p>0! is defined to be equal to 1 \u2014 but why? It comes from combinatorics, the study of counting and choosing. In combinatorics, the definition of <code class=\"language-text\">n!<\/code> is the number of orders and ways <code class=\"language-text\">n<\/code> items can be chosen (i.e. permutations).<\/p>\n<p>For instance, <code class=\"language-text\">2! = 2<\/code> is the number of ways you can select and order two items \u2014 the first item, then the second; or the second item, then the first. By that same token, <code class=\"language-text\">0!<\/code> is the number of ways you can select and order zero items \u2014 and you can do that in only <code class=\"language-text\">1<\/code> way.<\/p>\n<\/blockquote><p>How can we create this function <code class=\"language-text\">f<\/code> for which <code class=\"language-text\">byoFactorial<\/code> will work for <code class=\"language-text\">x = 0 (0! = 1)<\/code>? Well, closer inspection of our <code class=\"language-text\">byoFactorial<\/code> function shows that it already has a base case defined \u2014 when <code class=\"language-text\">x<\/code> is <code class=\"language-text\">0<\/code>, it returns <code class=\"language-text\">1<\/code>. Moreover, when <code class=\"language-text\">x = 0<\/code>, no self-reference occured anyway in the original <code class=\"language-text\">factorial<\/code> function \u2014 thus, <code class=\"language-text\">f<\/code> is not called.<\/p><p>That means we can pass any arbitrary function as <code class=\"language-text\">f<\/code>. When <code class=\"language-text\">x = 0<\/code>, <code class=\"language-text\">byoFactorial<\/code> will still return <code class=\"language-text\">1<\/code> regardless.<\/p><p>Here are some suggestions.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function identity(x) {\n  return x;\n}\nfunction one(x) {\n  return 1;\n}\nfunction random(x) {\n  return Math.random();\n}\n\/\/ I like to use `random` because you'll\n\/\/ immediately know if your function stops\n\/\/ working - you'll get a long and ugly\n\/\/ decimal. With the `identity` or `one` functions,\n\/\/ problems with output are more likely to slip\n\/\/ through. For the rest of this writeup, I'll\n\/\/ use `random`.\nbyoFactorial(random, 0);\n\/\/ =&gt; 1 - it works!\nbyoFactorial(random, 1);\n\/\/ =&gt; 0.07382473983422022<\/pre><\/div><p>So far, we\u2019ve been passing a function <code class=\"language-text\">f<\/code> to <code class=\"language-text\">byoFactorial<\/code>, but we can just as easily create replacement <code class=\"language-text\">factorial<\/code> functions from <code class=\"language-text\">byoFactorial<\/code> by <code class=\"language-text\">bind<\/code>ing <code class=\"language-text\">f<\/code>. Using this method, we\u2019ll create new functions which we can call with just one parameter \u2014 our desired factorial, <code class=\"language-text\">x<\/code>.<\/p><blockquote>\n<p>In our case, <code class=\"language-text\">bind<\/code> takes two parameters: a <code class=\"language-text\">this<\/code> value and then the actual first parameter for <code class=\"language-text\">byoFactorial<\/code>, <code class=\"language-text\">f<\/code>. Since the body of <code class=\"language-text\">byoFactorial<\/code> doesn\u2019t access <code class=\"language-text\">this<\/code> at all, we can just set it to <code class=\"language-text\">null<\/code>.<\/p>\n<\/blockquote><div class=\"gatsby-highlight\"><pre class=\"language-js\">function byoFactorial(f, x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * f(x - 1);\n}\n\/\/ Pass null because there's nothing specific we need to set `this` to\nconst factorialUpTo0 = byoFactorial.bind(null, Math.random);\n\/\/ factorialUpTo0 will work with x = 0, and can be called like a regular function:\nfactorialUpTo0(0);\n\/\/ =&gt; 1<\/pre><\/div><p>Here\u2019s what we get:<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">  factorialUpTo0(0) =&gt; 1\n  \/\/ I passed Math.random as the first parameter so you'll probably get different results\n  factorialUpTo0(1) =&gt; 0.07382473983422022\n  factorialUpTo0(2) =&gt; 0.7045168327647198\n  factorialUpTo0(3) =&gt; 1.4614935135773812\n  factorialUpTo0(4) =&gt; 0.3311529571140488<\/pre><\/div><p>As expected, it works for <code class=\"language-text\">0!<\/code>, but nothing else. But it\u2019s progress! We\u2019re <code class=\"language-text\">1\/infinity<\/code> done our complete <code class=\"language-text\">factorial<\/code> function!<\/p><blockquote>\n<p><code class=\"language-text\">1\/infinity<\/code> is bigger than you think! Ever wondered what the sum of <code class=\"language-text\">1\/2 + 1\/3 + 1\/4 ... + 1\/infinity + 1\/(infinity + 1)<\/code> was? Yep, it\u2019s <code class=\"language-text\">infinity<\/code>.<\/p>\n<\/blockquote><div class=\"gatsby-highlight\"><pre class=\"language-js\">function byoFactorial(f, x) {\n  if (x === 0) {\n    return 1;\n  }\n  return x * f(x - 1);\n}<\/pre><\/div><p>Let\u2019s make <code class=\"language-text\">byoFactorial<\/code> work for both <code class=\"language-text\">0!<\/code> and <code class=\"language-text\">1!<\/code>. In other words, how can we create a function <code class=\"language-text\">f<\/code> that we can bind to <code class=\"language-text\">byoFactorial<\/code> that\u2019ll allow it to return correct outputs for <code class=\"language-text\">x = 0<\/code> and <code class=\"language-text\">x = 1<\/code>?<\/p><p>First, let\u2019s walk through what happens when we pass <code class=\"language-text\">x = 1<\/code> into <code class=\"language-text\">byoFactorial<\/code>. For whatever <code class=\"language-text\">f<\/code> we come up with, <code class=\"language-text\">byoFactorial(f, 1)<\/code> will return <code class=\"language-text\">1 * f(0)<\/code>. So, all we need is a factorial function <code class=\"language-text\">f<\/code> which returns the correct value for the case when <code class=\"language-text\">x = 0<\/code>. Do we already have a factorial function which works for <code class=\"language-text\">0!<\/code>?<\/p><p>We do, and it\u2019s called <code class=\"language-text\">factorialUpTo0<\/code>! Let\u2019s bind it to <code class=\"language-text\">byoFactorial<\/code>. This will create a factorial function which will work for both <code class=\"language-text\">x = 0<\/code> and <code class=\"language-text\">1<\/code> \u2014 or rather, all integer values from 0 up to 1 (this process will be easier to reason about if we think about it this way). We\u2019ll call this function <code class=\"language-text\">factorialUpTo1<\/code>.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo1 = byoFactorial.bind(null, factorialUpTo0);<\/pre><\/div><p>Before we keep going, let\u2019s refactor our <code class=\"language-text\">byoFactorial<\/code> function. Calling <code class=\"language-text\">.bind<\/code> is clunky (especially with that <code class=\"language-text\">null<\/code>!) and we can clean up our syntax by returning a function instead (in essence, creating a \u201cfactory\u201d).<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function factorialFactory(f) {\n  return function factorial(x) {\n    if (x === 0) {\n      return 1;\n    }\n    return x * f(x - 1);\n  };\n}<\/pre><\/div><p>Now, we can create <code class=\"language-text\">factorialUpTo0<\/code> and <code class=\"language-text\">factorialUpTo1<\/code> functions with much cleaner syntax.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo0 = factorialFactory(Math.random);\nconst factorialUpTo1 = factorialFactory(factorialUpTo0);\n\/\/ Instead of\n\/\/ const factorialUpTo1 = byoFactorial.bind(null, factorialUpTo0);<\/pre><\/div><p>And we can make <code class=\"language-text\">factorialUpTo2<\/code>, too!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo2 = factorialFactory(factorialUpTo1);<\/pre><\/div><p><code class=\"language-text\">factorialUpTo2<\/code> will work with for <code class=\"language-text\">x = 2<\/code>, because it will return <code class=\"language-text\">2 * factorialUpTo1(1)<\/code>, and we already know that <code class=\"language-text\">factorialUpTo1<\/code> will work up to 1, so we also know that <code class=\"language-text\">factorialUpTo2<\/code> will return correct outputs for all values up to 1 as well!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo3 = factorialFactory(factorialUpTo2);<\/pre><\/div><p>By the same token, <code class=\"language-text\">factorialUpTo3<\/code> will work for <code class=\"language-text\">3!<\/code>, because it will return <code class=\"language-text\">3 * factorialUpTo2(2)<\/code>, and we already know that <code class=\"language-text\">factorialUpTo2<\/code> will work on all values up to 2!<\/p><p>We can continue this pattern to create our complete, working, non-recursive factorial function.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo3 = factorialFactory(factorialUpTo2);\nconst factorialUpTo4 = factorialFactory(factorialUpTo3);\nconst factorialUpTo5 = factorialFactory(factorialUpTo4);\nconst factorialUpTo6 = factorialFactory(factorialUpTo5);\nconst factorialUpTo7 = factorialFactory(factorialUpTo6);\nconst factorialUpTo8 = factorialFactory(factorialUpTo7);\nconst factorialUpTo9 = factorialFactory(factorialUpTo8);\nconst factorialUpTo10 = factorialFactory(factorialUpTo9);\n\/\/ ...\nfactorialUpTo10(5);\n\/\/ =&gt; 120\nfactorialUpTo10(8);\n\/\/ =&gt; 40320\nfactorialUpTo10(10);\n\/\/ =&gt; 3628800\nfactorialUpTo10(11);\n\/\/ =&gt; 28570479.681626465\nfactorialUpTo10(12);\n\/\/ =&gt; 79140708.41594502<\/pre><\/div><blockquote>\n<p>Note that the very first function <code class=\"language-text\">f<\/code> we passed to <code class=\"language-text\">factorialFactory<\/code> to create <code class=\"language-text\">factorialUpTo0<\/code> (in my case, <code class=\"language-text\">Math.random<\/code>) is only called if we pass an <code class=\"language-text\">x<\/code> greater than the upper bound of our <code class=\"language-text\">factorialUpTo{x}<\/code> function. If we keep passing our newly created <code class=\"language-text\">factorialUpTo{x}<\/code> functions back into <code class=\"language-text\">factorialFactory<\/code>, ad infinitum, <code class=\"language-text\">Math.random<\/code> will never be called (because the upper bound will be at infinity), and our function will work over all values of <code class=\"language-text\">x<\/code>!<\/p>\n<\/blockquote><p>Here\u2019s what <code class=\"language-text\">factorialUpTo10<\/code> looks like, expanded.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const factorialUpTo10 = factorialFactory(\n  factorialFactory(\n    factorialFactory(\n      factorialFactory(\n        factorialFactory(\n          factorialFactory(\n            factorialFactory(\n              factorialFactory(\n                factorialFactory(\n                  factorialFactory(factorialFactory(Math.random))\n                )\n              )\n            )\n          )\n        )\n      )\n    )\n  )\n);<\/pre><\/div><p>So, it looks like we\u2019ll need to apply <code class=\"language-text\">factorialFactory<\/code> to itself ad infinitum to get our working <code class=\"language-text\">factorial<\/code> function. How can we do that though?<\/p><div class=\"gatsby-highlight\"><pre class=\"language-text\">const factorialUpToInfinity = ???<\/pre><\/div><h2>Here Comes Y!<\/h2><p>The Y combinator is here to save the day! The Y combinator is a function which applies another function infinite times, and in simple terms can be defined like below.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return f(f(f(f(f(f(f(f(f(f(f(f(f(...)))))))))))));\n}<\/pre><\/div><p>Unfortunately, JavaScript can\u2019t evaluate the above expression. We need to simplify it \u2014 and to start, it helps to think about the definition of <code class=\"language-text\">y<\/code> in the context of infinite sums.<\/p><h3>A Math Problem<\/h3><p>What\u2019s the sum of <code class=\"language-text\">sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 ...)))))<\/code>?<\/p><p>The solution is simple. Let\u2019s call the sum <code class=\"language-text\">S<\/code>, so <code class=\"language-text\">S = sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 + sqrt(2 ...)))))<\/code>.<\/p><p>Now, if we look closely at the expression, we can see <code class=\"language-text\">S<\/code> is a part of itself! The sum can alternatively be written as <code class=\"language-text\">S = sqrt(2 + S)<\/code>.<\/p><p>Squaring both sides, we get <code class=\"language-text\">S^2 = S + 2<\/code>. Moving all terms of <code class=\"language-text\">S<\/code> to the left side, we get <code class=\"language-text\">S^2 - S - 2 = 0<\/code>. Factored, this is <code class=\"language-text\">(S - 2)(S + 1) = 0<\/code>, so the two solutions to this equation are <code class=\"language-text\">S = -1<\/code> and <code class=\"language-text\">S = 2<\/code>. Since <code class=\"language-text\">sqrt<\/code> returns the principal (non-negative) square root, our solution must be <code class=\"language-text\">S = 2<\/code>.<\/p><p>We can apply this same principle to our original formulation of the Y-combinator: <code class=\"language-text\">y(f) = f(f(f(f(f(f(f(f(f(f(f(f(f(...)))))))))))))<\/code>. Looking closely, it\u2019s clear that <code class=\"language-text\">y(f)<\/code> is a part of the equation for <code class=\"language-text\">y(f)<\/code>! We can write the new equation for <code class=\"language-text\">y(f)<\/code> like so: <code class=\"language-text\">y(f) = f(y(f))<\/code>.<\/p><p>We can wrap that expression in another function so we can explicitly pass our <code class=\"language-text\">x<\/code> parameter, too.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return function(x) {\n    return f(y(f))(x);\n  };\n}<\/pre><\/div><p>There\u2019s another reason we can\u2019t define our version of <code class=\"language-text\">y<\/code> as simply <code class=\"language-text\">y(f) = f(y(f))<\/code>.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return f(y(f));\n}<\/pre><\/div><p>When we evaluate the above, we get this:<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const tenFactorial = y(factorialFactory)(10);\n\/\/ =&gt; RangeError: Maximum call stack size exceeded<\/pre><\/div><p>The problem is that JavaScript tries to evaluate <code class=\"language-text\">y(factorialFactory)<\/code>, which never finishes, before doing any computation. Let\u2019s quickly walk through the evaluation.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return f(y(f));\n}\ny(factorialFactory)(10);\nfactorialFactory(y(factorialFactory))(10);\nfactorialFactory(factorialFactory(y(factorialFactory)))(10);<\/pre><\/div><p>This continues until we get the stack overflow error.<\/p><p>On the other hand, when we wrap the expression in an outer function to pass the parameter <code class=\"language-text\">x<\/code>, we also defer evaluation of <code class=\"language-text\">y(f)<\/code>. The term for this type of outer function is a \u201cthunk\u201d, which you may have heard from libraries like redux-thunk, which does the same thing but for Redux, a popular state management library often paired with React.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return function thunk(x) {\n    return f(y(f))(x);\n  };\n}<\/pre><\/div><p>Let\u2019s walk through the thunked evaluation.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function factorialFactory(f) {\n  return function factorial(x) {\n    if (x === 0) {\n      return 1;\n    }\n    return x * f(x - 1);\n  };\n}\ny(factorialFactory)(10)\nthunk(10) {\n  return factorialFactory(y(factorialFactory))(10)\n}\nfactorialFactory(y(factorialFactory))(10)\nfactorialFactory(function thunk(x) {\n  return factorialFactory(y(factorialFactory))(x)\n})(10)\n\/\/ At this point, the first argument to factorialFactory\n\/\/ has been fully evaluated (the thunk is the result of\n\/\/ the evalation)! Now, the body of factorialFactory can\n\/\/ actually be run.\n10 * (function thunk(x) {\n  return factorialFactory(y(factorialFactory))(x);\n})(9)\n10 * factorialFactory(y(factorialFactory))(9);\n\/\/ And so on<\/pre><\/div><p>The evaluation semantics of JavaScript are really the only thing preventing us from defining the Y combinator as we had it originally (<code class=\"language-text\">y(f) = f(y(f))<\/code>). In other languages, it\u2019s completely acceptable!<\/p><p>In Haskell, a functional language, the Y combinator <em>can<\/em> be defined as we did above initially:\n<code class=\"language-text\">y(f) = f(y(f))<\/code>.\nThe reason this is possible is because Haskell\u2019s evaluation semantics are different \u2014 Haskell is a lazy language. By default, it doesn\u2019t fully evaluate expressions until they are truly needed.<\/p><p>Here\u2019s how Haskell would evaluate <code class=\"language-text\">y(factorialFactory)(10)<\/code> above, given that <code class=\"language-text\">y<\/code> was defined as we had it originally.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ Our factorialFactory, in JavaScript for reference\nfunction factorialFactory(f) {\n  return function factorial(x) {\n    if (x === 0) {\n      return 1;\n    }\n    return x * f(x - 1);\n  };\n}<\/pre><\/div><div class=\"gatsby-highlight\"><pre class=\"language-python\">y(factorialFactory)(10)\n# Haskell will initially unpack the call to y by one layer\nfactorialFactory(y(factorialFactory))(10)\n# At this point, Haskell will look at the call to factorialFactory, instead of continuing to unpack the original y (which is what JavaScript does)\nreturn 10 * y(factorialFactory)(9)\n# Haskell will repeat the above process for the rest of the calls to y\n# ...\nreturn 10 * 9 * 8 * 7 * 6 * 5 * 4 * 3 * 2 * 1 * factorialFactory(y(factorialFactory))(0)\n# At this point, factorialFactory encounters its base case and returns 1\n# Haskell never computes the actual value of y(f)!<\/pre><\/div><h2>But There\u2019s Still Recursion!<\/h2><p>Our current definition <code class=\"language-text\">y<\/code> of the Y combinator still contains recursion, though! One of the main selling points of the Y combinator is that it allows us to add support for recursion to languages which don\u2019t support it \u2014 so there must be a way to get rid of the recursive call.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function y(f) {\n  return function thunk(x) {\n    return f(y(f))(x);\n  };\n}<\/pre><\/div><p>We could take the same approach as we did before, factoring out the recursive call into a parameter and creating a \u201cfactory\u201d function, but that won\u2019t fly \u2014 as we saw in the <code class=\"language-text\">factorial<\/code> example, we\u2019d still need to apply that factory function ad infinitum to create a functional <code class=\"language-text\">y<\/code>.<\/p><p>Let\u2019s take a closer look at <code class=\"language-text\">y<\/code>, then. How can we define <code class=\"language-text\">y<\/code> in non-recursive terms? Well, we can take advantage of the fact that while we are voluntarily shunning self-reference in function bodies, we can still pass functions to themselves as parameters!<\/p><p>We\u2019ll need to make a clever substitution, though. Instead of defining <code class=\"language-text\">y(f)<\/code> as we have it now, let\u2019s define it as <code class=\"language-text\">y(f) = x(x)(f)<\/code>, for some other function <code class=\"language-text\">x<\/code>. In simpler terms, <code class=\"language-text\">y = x(x)<\/code> \u2014 <code class=\"language-text\">y<\/code> is <code class=\"language-text\">x<\/code> applied to itself.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ Here's our current, recursively-defined y\nfunction y(f) {\n  return function thunk(n) {\n    return f(y(f))(n);\n  };\n}\n\/\/ Apply the substitution y = x(x)\n\/\/ Call the parameter otherX since we don't want\n\/\/ to confuse the outer x and inner x and to avoid\n\/\/ variable shadowing\nfunction x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}<\/pre><\/div><p>Let\u2019s try this out.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const newY = x(x);\n\/\/ x(x)(factorialFactory)\nconst newFactorial = newY(factorialFactory);\nnewFactorial(10);\n\/\/ =&gt; 3628800<\/pre><\/div><p>It works! Let\u2019s walk through the evaluation.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}\nx(x)(f);\nreturn function thunk(n) {\n  return f(x(x)(f))(n);\n};\n\/\/ Unthunked\nf(x(x)(f));\n\/\/ From first step above, we see that x(x)(f) = f(x(x)(f))\n\/\/ We can apply this substitution\nf(f(x(x)(f)));\n\/\/ And again, and again... look familiar?\nf(f(f(f(f(f(x(x)(f)))))));<\/pre><\/div><p>It works! By applying <code class=\"language-text\">x<\/code> to itself, we avoid self-reference in the function body. Everything the function needs to run comes from its parameters.<\/p><p>And with a little bit more refactoring and substitution, we\u2019ve got the non-recursive Y combinator!<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">\/\/ What we already have\nfunction x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      \/\/ By passing `x` to itself (in the form of `otherX`),\n      \/\/ we can technically \"call\" `x` inside the function\n      \/\/ body of `x` - but importantly, non-recursively! Of course,\n      \/\/ now that we've altered the parameters of `x` to take\n      \/\/ itself, we need to pass `otherX` to `otherX` in the\n      \/\/ return statement, too.\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}\n\/\/ Let's refactor\nconst y = x(x);\n\/\/ Here's `x` agian\nfunction x(otherX) {\n  return function y(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n}\n\/\/ Now, just substitute the body of `x`\n\/\/ for each `x` in the statement `y = x(x)`\nconst y = (function(otherX) {\n  return function(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n})(function(otherX) {\n  return function(f) {\n    return function thunk(n) {\n      return f(otherX(otherX)(f))(n);\n    };\n  };\n});\n\/\/ If we want `y` to be usable on functions\n\/\/ with any number of arguments, we can use\n\/\/ the ES6 spread syntax over the arguments\nconst y = (function(otherX) {\n  return function(f) {\n    return function thunk(...args) {\n      return f(otherX(otherX)(f))(...args);\n    };\n  };\n})(function(otherX) {\n  return function(f) {\n    return function thunk(...args) {\n      return f(otherX(otherX)(f))(...args);\n    };\n  };\n});\n\/\/ We can shorten it even more with ES6 arrow functions\nconst y = (x =&gt; f =&gt; (...a) =&gt; f(x(x)(f))(...a))(x =&gt; f =&gt; (...a) =&gt;\n  f(x(x)(f))(...a)\n);<\/pre><\/div><p>And there you go! One of the most important concepts in theoretical computer science, sitting on your lap in just a few characters.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">const y = (x =&gt; f =&gt; (...a) =&gt; f(x(x)(f))(...a))(x =&gt; f =&gt; (...a) =&gt;\n  f(x(x)(f))(...a)\n);<\/pre><\/div><h2>Appendix<\/h2><p>Try writing your own <code class=\"language-text\">y<\/code>-compatible, non-recursive implementations of your favorite recursive processes! Here\u2019s an example with the Fibonacci sequence mentioned earlier.<\/p><div class=\"gatsby-highlight\"><pre class=\"language-js\">function fibonacciFactory(f) {\n  return function thunk(n) {\n    if (n === 0 || n === 1) {\n      return 1;\n    }\n    return f(n - 2) + f(n - 1);\n  };\n}\nconst fibonacci = y(fibonacciFactory);\nfibonacci(10);\n\/\/ =&gt; 89<\/pre><\/div><h2>About the Author<\/h2><p>Thanks for reading! My name is <a href=\"https:\/\/www.nathanhleung.com\">Nathan Leung<\/a>. I\u2019m the founder of location-based social messaging app <a href=\"https:\/\/www.getsuap.com\/\">Suap<\/a> and an incoming first-year student at the University of Michigan.<\/p><p>You can learn more about me on my <a href=\"https:\/\/www.linkedin.com\/in\/nathanhleung\/\">LinkedIn<\/a> or on my <a href=\"https:\/\/www.nathanhleung.com\">personal website<\/a>, and I encourage you to try <a href=\"https:\/\/www.getsuap.com\/\">Suap<\/a> at <a href=\"https:\/\/www.getsuap.com\/\">getsuap.com<\/a>!<\/p>","created_at":"2018-11-07T16:23:17+0000","updated_at":"2018-11-07T16:23:17+0000","annotations":[],"mimetype":"text\/html","reading_time":15,"domain_name":"enlight.nyc","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":130,"title":"Ask HN: What's on your RSS reader?","url":"https:\/\/news.ycombinator.com\/item?id=12194875","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D12194875\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"12194875\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=12194875\" class=\"storylink\">Ask HN: What's on your RSS reader?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        148 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=Guyag\" class=\"hnuser\">Guyag<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12194875\">on July 30, 2016<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=12194875&amp;goto=item%3Fid%3D12194875\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20What's%20on%20your%20RSS%20reader%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20What's%20on%20your%20RSS%20reader%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=12194875&amp;auth=d5fd8c14ece6b6fd330aa77cbfd0094f96fb51b1\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=12194875\">72\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>I'm interested to hear what you think is good\/important enough to warrant following, admittedly with the aim of expanding my list. I'd also be interested to hear what reader you use.<p>Here's my list (most of these are easily searchable so I won't link them):<\/p><p>Comics:\nCube Drone, Dilbert daily, Garfield daily, Invisible Bread, lolnein, Poorly Drawn Lines, The Oatmeal, xkcd<\/p><p>Misc:\nxkcd what if?, John A De Goes' blog [1], Still Drinking [2], Krebs on Security<\/p><p>I use a Telegram bot [3] as a 'reader' of sorts (no affiliation).<\/p><p>[1] http:\/\/degoes.net\/<\/p><p>[2] https:\/\/www.stilldrinking.org\/<\/p><p>[3] https:\/\/telegram.me\/TheFeedReaderBot<\/p><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:23:17+0000","updated_at":"2018-11-07T16:23:17+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":128,"title":"Ask HN: Is it possible to run your own mail server for personal use?","url":"https:\/\/news.ycombinator.com\/item?id=12282231","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D12282231\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"12282231\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=12282231\" class=\"storylink\">Ask HN: Is it possible to run your own mail server for personal use?<\/a><\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        648 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282231\">on Aug 13, 2016<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=12282231&amp;goto=item%3Fid%3D12282231\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Ask%20HN%3A%20Is%20it%20possible%20to%20run%20your%20own%20mail%20server%20for%20personal%20use%3F&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Ask%20HN%3A%20Is%20it%20possible%20to%20run%20your%20own%20mail%20server%20for%20personal%20use%3F\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=12282231&amp;auth=9d13573b98f38964434265992146fc5e3430e77b\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=12282231\">299\u00a0comments<\/a>              <\/td><\/tr><tr><\/tr><tr><td colspan=\"2\"><\/td><td>Recently I decided that I wanted to run my own MTA. Downloaded qmail, applied a couple of patches and it was done. The problem is making sure my mail is not marked as spam by the major MTAs out there, gmail and hotmail both mark my mails as spam. So far I've:<p>- made sure I'm not running an open relay (obviously)<\/p><p>- made sure a reverse lookup to my IP matches my host<\/p><p>- made sure my IP wasn't already in any spam IP blacklist<\/p><p>- added SPF<\/p><p>- added DKIM<\/p><p>- run a lot of tests at glockapps.com<\/p><p>It's still not enough. I don't know what else to do, is it possible to run your own MTA, for personal use in 2016? Who, here, is doing it successfully?<\/p><\/td><\/tr><\/table><table border=\"0\" class=\"comment-tree\"><tr class=\"athing comtr\" id=\"12282667\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nux\" class=\"hnuser\">Nux<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282667\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283060\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwr\" class=\"hnuser\">jwr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283060\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283270\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simon_acca\" class=\"hnuser\">simon_acca<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283270\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283613\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kequc\" class=\"hnuser\">Kequc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283613\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285277\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MatthewWilkes\" class=\"hnuser\">MatthewWilkes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285277\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283649\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=OJFord\" class=\"hnuser\">OJFord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283649\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283755\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=teddyh\" class=\"hnuser\">teddyh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283755\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284275\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=robertely\" class=\"hnuser\">robertely<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284275\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284805\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=philtar\" class=\"hnuser\">philtar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284805\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12286892\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sbierwagen\" class=\"hnuser\">sbierwagen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12286892\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285443\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"320\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bigp3t3\" class=\"hnuser\">bigp3t3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285443\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12286458\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=OJFord\" class=\"hnuser\">OJFord<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12286458\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12289206\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=teddyh\" class=\"hnuser\">teddyh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12289206\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284834\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=benev\" class=\"hnuser\">benev<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284834\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12289510\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=simon_acca\" class=\"hnuser\">simon_acca<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12289510\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12289236\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=benbristow\" class=\"hnuser\">benbristow<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12289236\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12307275\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bedros\" class=\"hnuser\">bedros<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12307275\">on Aug 17, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283155\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=reallyimran\" class=\"hnuser\">reallyimran<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283155\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285707\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tqwhite\" class=\"hnuser\">tqwhite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285707\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285727\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daxorid\" class=\"hnuser\">daxorid<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285727\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12286463\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blktiger\" class=\"hnuser\">blktiger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12286463\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12298065\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gentoo\" class=\"hnuser\">gentoo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12298065\">on Aug 16, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283332\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickpsecurity\" class=\"hnuser\">nickpsecurity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283332\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283555\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bbunix\" class=\"hnuser\">bbunix<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283555\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285723\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tqwhite\" class=\"hnuser\">tqwhite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285723\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283939\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=malbs\" class=\"hnuser\">malbs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283939\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283297\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=seanp2k2\" class=\"hnuser\">seanp2k2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283297\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283641\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=falcolas\" class=\"hnuser\">falcolas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283641\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283699\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mozumder\" class=\"hnuser\">mozumder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283699\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284195\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=XorNot\" class=\"hnuser\">XorNot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284195\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285239\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Karlozkiller\" class=\"hnuser\">Karlozkiller<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285239\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283691\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mozumder\" class=\"hnuser\">mozumder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283691\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284262\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dasil003\" class=\"hnuser\">dasil003<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284262\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284498\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ommunist\" class=\"hnuser\">ommunist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284498\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284505\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dasil003\" class=\"hnuser\">dasil003<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284505\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285035\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gambiting\" class=\"hnuser\">gambiting<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285035\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285112\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ommunist\" class=\"hnuser\">ommunist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285112\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284602\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mozumder\" class=\"hnuser\">mozumder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284602\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283485\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=joosters\" class=\"hnuser\">joosters<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283485\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282968\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noinsight\" class=\"hnuser\">noinsight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282968\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282989\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=X86BSD\" class=\"hnuser\">X86BSD<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282989\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283886\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MustardTiger\" class=\"hnuser\">MustardTiger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283886\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284371\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=X86BSD\" class=\"hnuser\">X86BSD<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284371\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285289\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MustardTiger\" class=\"hnuser\">MustardTiger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285289\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283175\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pflanze\" class=\"hnuser\">pflanze<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283175\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283265\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=noinsight\" class=\"hnuser\">noinsight<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283265\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283343\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickpsecurity\" class=\"hnuser\">nickpsecurity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283343\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283020\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merb\" class=\"hnuser\">merb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283020\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283151\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283151\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283180\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=matthewking\" class=\"hnuser\">matthewking<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283180\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284405\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=voltagex_\" class=\"hnuser\">voltagex_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284405\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284610\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=merb\" class=\"hnuser\">merb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284610\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283052\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cddotdotslash\" class=\"hnuser\">cddotdotslash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283052\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282714\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=colejohnson66\" class=\"hnuser\">colejohnson66<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282714\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282751\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rasjani\" class=\"hnuser\">rasjani<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282751\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284751\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lazyjones\" class=\"hnuser\">lazyjones<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284751\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12286802\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mancerayder\" class=\"hnuser\">mancerayder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12286802\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12290365\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zatkin\" class=\"hnuser\">zatkin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12290365\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282972\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thomasjames\" class=\"hnuser\">thomasjames<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282972\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283512\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mgbmtl\" class=\"hnuser\">mgbmtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283512\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282851\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=darkhorn\" class=\"hnuser\">darkhorn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282851\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284504\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ashitlerferad\" class=\"hnuser\">ashitlerferad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284504\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283126\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chrischen\" class=\"hnuser\">chrischen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283126\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284430\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=NetOpWibby\" class=\"hnuser\">NetOpWibby<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284430\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283872\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MustardTiger\" class=\"hnuser\">MustardTiger<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283872\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282893\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=madez\" class=\"hnuser\">madez<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282893\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282952\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=a3n\" class=\"hnuser\">a3n<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282952\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283420\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=quickben\" class=\"hnuser\">quickben<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283420\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283458\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=systemfreund\" class=\"hnuser\">systemfreund<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283458\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284038\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mastazi\" class=\"hnuser\">mastazi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284038\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284645\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmontra\" class=\"hnuser\">pmontra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284645\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282808\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrb\" class=\"hnuser\">mrb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282808\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283070\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bks\" class=\"hnuser\">bks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283070\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285519\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StavrosK\" class=\"hnuser\">StavrosK<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285519\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12288061\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dhruvkar\" class=\"hnuser\">dhruvkar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12288061\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285673\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bkuri\" class=\"hnuser\">bkuri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285673\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283014\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nokya\" class=\"hnuser\">nokya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283014\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284160\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=huskyr\" class=\"hnuser\">huskyr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284160\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284511\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ashitlerferad\" class=\"hnuser\">ashitlerferad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284511\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284824\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fnordje\" class=\"hnuser\">Fnordje<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284824\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285074\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=huskyr\" class=\"hnuser\">huskyr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285074\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283054\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tezza\" class=\"hnuser\">tezza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283054\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283122\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mhurron\" class=\"hnuser\">mhurron<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283122\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282880\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nilved\" class=\"hnuser\">nilved<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282880\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283256\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=downandout\" class=\"hnuser\">downandout<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283256\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282881\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=djsumdog\" class=\"hnuser\">djsumdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282881\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283026\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=x0x0\" class=\"hnuser\">x0x0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283026\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283195\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=PaulHoule\" class=\"hnuser\">PaulHoule<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283195\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12287464\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SteveCr48\" class=\"hnuser\">SteveCr48<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12287464\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283722\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LinuxBender\" class=\"hnuser\">LinuxBender<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283722\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283098\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrb\" class=\"hnuser\">mrb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283098\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283545\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mgbmtl\" class=\"hnuser\">mgbmtl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283545\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283350\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ptero\" class=\"hnuser\">ptero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283350\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282925\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walrus01\" class=\"hnuser\">walrus01<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282925\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283204\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283204\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283242\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=semanticist\" class=\"hnuser\">semanticist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283242\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283312\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283312\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283729\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LinuxBender\" class=\"hnuser\">LinuxBender<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283729\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283471\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=quonn\" class=\"hnuser\">quonn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283471\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285412\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crasm\" class=\"hnuser\">crasm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285412\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12318705\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Spakman\" class=\"hnuser\">Spakman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12318705\">on Aug 19, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283443\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=havetocharge\" class=\"hnuser\">havetocharge<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283443\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283608\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283608\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12287260\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=havetocharge\" class=\"hnuser\">havetocharge<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12287260\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12288773\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12288773\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282562\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hannob\" class=\"hnuser\">hannob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282562\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282839\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ChuckMcM\" class=\"hnuser\">ChuckMcM<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282839\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283187\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283187\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282522\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TheMog\" class=\"hnuser\">TheMog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282522\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283161\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrewaylett\" class=\"hnuser\">andrewaylett<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283161\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283846\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Biganon\" class=\"hnuser\">Biganon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283846\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282958\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dsr_\" class=\"hnuser\">dsr_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282958\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283462\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrbill\" class=\"hnuser\">mrbill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283462\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283686\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dsr_\" class=\"hnuser\">dsr_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283686\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283218\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283218\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283674\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dsr_\" class=\"hnuser\">dsr_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283674\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283970\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283970\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284581\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=angry_octet\" class=\"hnuser\">angry_octet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284581\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285097\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zaxomi\" class=\"hnuser\">zaxomi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285097\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282656\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282656\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285176\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=syshum\" class=\"hnuser\">syshum<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285176\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285216\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285216\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285644\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=syshum\" class=\"hnuser\">syshum<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285644\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282707\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TheMog\" class=\"hnuser\">TheMog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282707\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285227\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285227\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12286414\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TheMog\" class=\"hnuser\">TheMog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12286414\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282474\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pflanze\" class=\"hnuser\">pflanze<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282474\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282502\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282502\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282638\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pflanze\" class=\"hnuser\">pflanze<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282638\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282553\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=derefr\" class=\"hnuser\">derefr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282553\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282565\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282565\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282586\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=calpaterson\" class=\"hnuser\">calpaterson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282586\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282748\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pflanze\" class=\"hnuser\">pflanze<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282748\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282894\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rahkiin\" class=\"hnuser\">rahkiin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282894\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283043\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwr\" class=\"hnuser\">jwr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283043\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282614\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tezza\" class=\"hnuser\">tezza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282614\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282716\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ctrlc-root\" class=\"hnuser\">ctrlc-root<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282716\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284180\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=superswordfish\" class=\"hnuser\">superswordfish<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284180\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282765\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tristor\" class=\"hnuser\">tristor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282765\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282450\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bensbox\" class=\"hnuser\">bensbox<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282450\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282468\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282468\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282480\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=techsupporter\" class=\"hnuser\">techsupporter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282480\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282509\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bensbox\" class=\"hnuser\">bensbox<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282509\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282683\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282683\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282469\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tantalor\" class=\"hnuser\">tantalor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282469\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282477\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282477\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284061\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kadin\" class=\"hnuser\">Kadin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284061\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282771\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=soneil\" class=\"hnuser\">soneil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282771\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283248\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283248\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283529\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TheOtherHobbes\" class=\"hnuser\">TheOtherHobbes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283529\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283598\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283598\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283646\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=agildehaus\" class=\"hnuser\">agildehaus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283646\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283944\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=loup-vaillant\" class=\"hnuser\">loup-vaillant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283944\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282878\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=darkhorn\" class=\"hnuser\">darkhorn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282878\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283018\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=soneil\" class=\"hnuser\">soneil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283018\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282922\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lwhalen\" class=\"hnuser\">lwhalen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282922\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282888\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sigil\" class=\"hnuser\">sigil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282888\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284492\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=njt\" class=\"hnuser\">njt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284492\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282836\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=acd\" class=\"hnuser\">acd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282836\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282457\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ebbv\" class=\"hnuser\">ebbv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282457\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282717\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282717\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282534\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhizome\" class=\"hnuser\">rhizome<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282534\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282735\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ebbv\" class=\"hnuser\">ebbv<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282735\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283075\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jjnoakes\" class=\"hnuser\">jjnoakes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283075\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283134\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jghn\" class=\"hnuser\">jghn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283134\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283329\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=johnrob\" class=\"hnuser\">johnrob<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283329\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283173\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sippeangelo\" class=\"hnuser\">sippeangelo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283173\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284122\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=__david__\" class=\"hnuser\">__david__<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284122\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283068\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mehdym\" class=\"hnuser\">mehdym<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283068\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283382\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=FollowSteph3\" class=\"hnuser\">FollowSteph3<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283382\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283402\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wslh\" class=\"hnuser\">wslh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283402\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282505\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=waits\" class=\"hnuser\">waits<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282505\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282528\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=timdeneau\" class=\"hnuser\">timdeneau<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282528\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282700\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282700\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282642\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zzzcpan\" class=\"hnuser\">zzzcpan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282642\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282802\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pcl\" class=\"hnuser\">pcl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282802\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282519\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=deftnerd\" class=\"hnuser\">deftnerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282519\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282575\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282575\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282625\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=deftnerd\" class=\"hnuser\">deftnerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282625\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284318\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tacon\" class=\"hnuser\">tacon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284318\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282418\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mifreewil\" class=\"hnuser\">mifreewil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282418\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282430\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282430\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282501\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pflanze\" class=\"hnuser\">pflanze<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282501\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282428\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jeffmould\" class=\"hnuser\">jeffmould<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282428\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284772\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=e12e\" class=\"hnuser\">e12e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284772\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284681\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=galori\" class=\"hnuser\">galori<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284681\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282779\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pja\" class=\"hnuser\">pja<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282779\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282641\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282641\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282788\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=takeda\" class=\"hnuser\">takeda<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282788\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282816\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282816\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283793\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dbcurtis\" class=\"hnuser\">dbcurtis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283793\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282994\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=virtualio\" class=\"hnuser\">virtualio<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282994\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282517\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=calpaterson\" class=\"hnuser\">calpaterson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282517\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282525\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=calpaterson\" class=\"hnuser\">calpaterson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282525\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282547\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=TheMog\" class=\"hnuser\">TheMog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282547\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283878\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jrnichols\" class=\"hnuser\">jrnichols<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283878\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282617\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=diego_moita\" class=\"hnuser\">diego_moita<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282617\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282652\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=extrapolate\" class=\"hnuser\">extrapolate<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282652\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12288928\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hukl\" class=\"hnuser\">hukl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12288928\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283044\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=perakojotgenije\" class=\"hnuser\">perakojotgenije<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283044\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282495\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=timdeneau\" class=\"hnuser\">timdeneau<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282495\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282595\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hstrauss\" class=\"hnuser\">hstrauss<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282595\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12287953\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shirro\" class=\"hnuser\">shirro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12287953\">on Aug 15, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283114\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hawat\" class=\"hnuser\">hawat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283114\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283318\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Jaruzel\" class=\"hnuser\">Jaruzel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283318\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283448\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrbill\" class=\"hnuser\">mrbill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283448\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282827\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swenn\" class=\"hnuser\">swenn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282827\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284747\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmlnr\" class=\"hnuser\">pmlnr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284747\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283758\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aabajian\" class=\"hnuser\">aabajian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283758\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284446\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fusiongyro\" class=\"hnuser\">fusiongyro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284446\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283123\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jghn\" class=\"hnuser\">jghn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283123\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283663\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lisper\" class=\"hnuser\">lisper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283663\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283084\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vancan1ty\" class=\"hnuser\">vancan1ty<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283084\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283532\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jloughry\" class=\"hnuser\">jloughry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283532\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282497\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cookiecaper\" class=\"hnuser\">cookiecaper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282497\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282864\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jcadam\" class=\"hnuser\">jcadam<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282864\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282755\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282755\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284208\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cookiecaper\" class=\"hnuser\">cookiecaper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284208\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282629\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cmdrfred\" class=\"hnuser\">cmdrfred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282629\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282743\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282743\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282887\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cmdrfred\" class=\"hnuser\">cmdrfred<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282887\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282999\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lucb1e\" class=\"hnuser\">lucb1e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282999\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283412\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bluejekyll\" class=\"hnuser\">bluejekyll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283412\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283436\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=havetocharge\" class=\"hnuser\">havetocharge<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283436\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282846\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lazyant\" class=\"hnuser\">lazyant<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282846\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282712\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=saynsedit\" class=\"hnuser\">saynsedit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282712\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285186\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cdysthe\" class=\"hnuser\">cdysthe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285186\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283188\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lrusnac\" class=\"hnuser\">lrusnac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283188\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283266\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sliverstorm\" class=\"hnuser\">sliverstorm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283266\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282848\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bluedino\" class=\"hnuser\">bluedino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282848\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282588\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gbuk2013\" class=\"hnuser\">gbuk2013<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282588\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282476\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mverwijs\" class=\"hnuser\">mverwijs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282476\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282819\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=soneil\" class=\"hnuser\">soneil<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282819\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282527\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhizome\" class=\"hnuser\">rhizome<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282527\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12283427\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=zbuf\" class=\"hnuser\">zbuf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12283427\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12287405\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mverwijs\" class=\"hnuser\">mverwijs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12287405\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282950\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=walrus01\" class=\"hnuser\">walrus01<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282950\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282662\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=z3t4\" class=\"hnuser\">z3t4<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282662\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284243\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dboreham\" class=\"hnuser\">dboreham<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284243\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284242\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dboreham\" class=\"hnuser\">dboreham<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284242\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282953\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=darkhorn\" class=\"hnuser\">darkhorn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282953\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282388\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=trentmb\" class=\"hnuser\">trentmb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282388\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282426\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282426\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282510\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ViViDboarder\" class=\"hnuser\">ViViDboarder<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282510\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282544\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mfjordvald\" class=\"hnuser\">mfjordvald<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282544\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282593\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jdmoreira\" class=\"hnuser\">jdmoreira<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282593\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12285964\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mfjordvald\" class=\"hnuser\">mfjordvald<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12285964\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282538\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhizome\" class=\"hnuser\">rhizome<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282538\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12282730\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nchelluri\" class=\"hnuser\">nchelluri<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12282730\">on Aug 13, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"athing comtr\" id=\"12284470\"><td>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhizome\" class=\"hnuser\">rhizome<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=12284470\">on Aug 14, 2016<\/a><\/p><br \/><\/td><\/tr><\/table><\/td><\/tr><tr class=\"morespace\"><\/tr><tr><td><table border=\"0\"><tr><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=12282231&amp;p=2\" class=\"morelink\" rel=\"nofollow\">More<\/a><\/td><\/tr><\/table><\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:23:16+0000","updated_at":"2018-11-07T16:23:16+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":127,"title":"My favorite apps on F-Droid  - [Quaap.com]","url":"https:\/\/quaap.com\/D\/use-fdroid","content":"<p> I often recommend F-Droid to my Android-using friends. <a href=\"https:\/\/f-droid.org\/\" target=\"_blank\">F-Droid<\/a> is an alternative app store which contains only Free and Open Source Software (FOSS). It's great because the apps are almost all not evil. Did you know that a flashlight app <em>doesn't<\/em> need access to your contacts,  location, and pictures? Did you know that every app doesn't need to serve you ads?\n<\/p><p>\n Yeah, sometimes the apps aren't as polished, and sometimes they lack some functionality you'd like, but they are much less likely to contain\/be malware or spyware. And sometimes they fill a unique niche not filled by anything else. And sometimes they are amazingly good.\n<\/p><p>\n The apps are some individual's or team's passion projects, and most are not looking to earn money on the apps, they just want to make something they and others would find useful or fun.  I wrote several of them myself, and I wrote them because I didn't like what was available commercially, or just had a weird idea.\n<\/p><p>\n  Anyway, below is a current list of these FOSS apps I use (or have installed for my kids \/ future use).  And again, one of the best features of any of them is that they don't serve you ads, and don't report your location and browsing habits anywhere.  And don't steal your data and contact list.\n<\/p><p>\n  To install them, you'll need to go to <a href=\"https:\/\/f-droid.org\/\" target=\"_blank\">F-Droid<\/a> and install the main F-Droid app.  When you do that, you'll need to say \"yes\" or \"allow\" to the prompts asking if you want to install from unknown sources.\n<\/p><p>\nNote: there are a huge number of apps I've not tried\u2026 These are just the ones I've installed and found useful.\n<\/p><h2 id=\"system-apps\">System apps<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.menny.android.anysoftkeyboard\" target=\"_blank\">AnySoftKeyboard<\/a><\/li>\n<\/ul><p>\nThis is a replacement for the keyboard.  It provides a much easier-to-use experience (to me) that the built-in keyboards don't, like easy access to the single quote, for example.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.launchtime\" target=\"_blank\">LaunchTime<\/a><\/li>\n<\/ul><p>\nOk, full disclosure: I wrote this one. <a href=\"https:\/\/quaap.com\/D\/LaunchTime\">LaunchTime<\/a> is my replacement homescreen\/launcher.  I do like it sooo much better than the other homescreens though.  Before I wrote it, I struggled greatly with the \"app drawer\" model used by most launchers. See the <a href=\"https:\/\/github.com\/quaap\/LaunchTime\/blob\/master\/README.md\" target=\"_blank\">LaunchTime page on Github<\/a> for a full list of features.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.jmstudios.redmoon\" target=\"_blank\">Red Moon<\/a><\/li>\n<\/ul><p>\nThis app is designed to reduce the display's effect on your day-night cycle by removing blue light.  I don't know about that, but I use it as a \"super dimmer\" at night and it can run on a schedule, too.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.thosp.yourlocalweather\" target=\"_blank\">Your local weather<\/a><\/li>\n<\/ul><p>\nA replacement for whatever spyware\/crapware weather widget came with your phone.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.vrem.wifianalyzer\" target=\"_blank\">WiFiAnalyzer<\/a><\/li>\n<\/ul><p>\nCan display a list of WiFi access points and help you decide on what channel to use.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/link.standen.michael.phonesaver\" target=\"_blank\">Phone Saver<\/a><\/li>\n<\/ul><p>\nAdds a \"save to a directory\" option to the share menus.  Run the app by itself first to set a destination directory.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/de.tobiasbielefeld.searchbar\" target=\"_blank\">Simple Search<\/a><\/li>\n<\/ul><p>\nA generic search app widget which you can target different search engines. It's pretty slim and mostly useful so you don't have to have some company's widget always running on your phone.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/de.onyxbits.listmyapps\" target=\"_blank\">List My Apps<\/a><\/li>\n<\/ul><p>\nI used this to help me make this list! <img src=\"https:\/\/quaap.com\/D\/media\/emoticons\/smile.png\" alt=\"smile :)\" title=\"smile :)\" \/><\/p><h2 id=\"graphics\">Graphics<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.simplemobiletools.gallery\" target=\"_blank\">Gallery<\/a><\/li>\n<\/ul><p>\nA gallery of you photos, a photo viewer, and a photo editor.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.witness.sscphase1\" target=\"_blank\">ObscuraCam<\/a><\/li>\n<\/ul><p>\nLoads up your pictures and automatically detects and blurs faces so that you can share pictures that contain other people\/kids easier.  You can also manually blur (or unblur) regions of the photo.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/anupam.acrylic\" target=\"_blank\">Acrylic Paint<\/a><\/li>\n<\/ul><p>\nFinger painting\/drawing.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.google.zxing.client.android\" target=\"_blank\">Barcode Scanner<\/a><\/li>\n<\/ul><p>\nSimple barcode\/QR code\/etc scanner.  Just scans them and adds the code to its history so you can share them via text, email, file, etc.\n<\/p><h2 id=\"reading-and-writing\">Reading and Writing<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.indywidualni.fblite\" target=\"_blank\">Face Slim<\/a><\/li>\n<\/ul><p>\nA super slimmed-down Facebook app.  It's just a thin wrapper around facebook's mobile site, but it's great because it allows you to not have the two behemoth Facebook apps (Facebook proper and Messenger) to bog down your phone and gather data.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/it.feio.android.omninotes.foss\" target=\"_blank\">Omni Notes FOSS<\/a><\/li>\n<\/ul><p>\nKind of a hybrid notes and checklist app which allows adding all kinds of rich text and pictures.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.bookymcbookface\" target=\"_blank\">Booky McBookface<\/a><\/li>\n<\/ul><p>\nOk, it's mine again.  An almost <em>too<\/em>-simple epub reader. Very few features, but that's how I like it.\n<\/p><p>\nA simple daily journal. Can tie in with your calendar.\n<\/p><h2 id=\"utilities\">Utilities<\/h2><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.woefe.shoppinglist\" target=\"_blank\">Shopping List<\/a><\/li>\n<\/ul><p>\nMy very favorite shopping list app.  You can manually sort the items (I have mine in ALDI-order), and have multiple lists.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/de.j4velin.pedometer\" target=\"_blank\">Pedometer<\/a><\/li>\n<\/ul><p>\nIf your device has a step counter sensor, this app can use it to keep track of your steps, and that's it.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.secuso.privacyfriendlyintervaltimer\" target=\"_blank\">Interval Timer<\/a><\/li>\n<\/ul><p>\nAn all-you-need app to help you if you do interval training.\n<\/p><p>\nVery simple app to generate dice rolls.  There are other, bigger apps available if you do tabletop gaming.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/openfoodfacts.github.scrachx.openfood\" target=\"_blank\">OpenFoodFacts<\/a><\/li>\n<\/ul><p>\nAccess data gathered on food products.  It can scan an item's UPC barcode and retrieve nutrition info (if available).\n<\/p><p>\nA simple text editor.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.danielkim.soundrecorder\" target=\"_blank\">Sound Recorder<\/a><\/li>\n<\/ul><p>\nAn easy way to record sound files and share them.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/net.programmierecke.radiodroid2\" target=\"_blank\">RadioDroid<\/a><\/li>\n<\/ul><p>\nA music player with a big list of free music and radio streams.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.gateshipone.odyssey\" target=\"_blank\">Odyssey<\/a><\/li>\n<\/ul><p>\nA kind of cool music player for mp3s and oggs on you phone.\n<\/p><p>\nTune your guitar, or just entertain a child who likes to see waveforms.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.billthefarmer.scope\" target=\"_blank\">Oscilloscope<\/a><\/li>\n<\/ul><p>\nView waveforms, or just entertain a child who likes to see waveforms.\n<\/p><h2 id=\"games-and-toys\">Games and toys<\/h2><p>Here's just a list of games I occasionally play, or, more likely, my kids play.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.primary\" target=\"_blank\">Primary<\/a><\/li>\n<\/ul><p>\nGamified educational app.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/net.sourceforge.solitaire_cg\" target=\"_blank\">SolitaireCG<\/a><\/li>\n<\/ul><p>\nStraightforward solitaire.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/name.boyle.chris.sgtpuzzles\" target=\"_blank\">Puzzles<\/a><\/li>\n<\/ul><p>\nLike 30 weird puzzle games.\n<\/p><p>\nYup 2048.\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/ch.logixisland.anuto\" target=\"_blank\">Anuto TD<\/a><\/li>\n<\/ul><p>\nAnother tower-defense game.\n<\/p><p>\nAnd a bunch more I won't describe, but I found them fun enough to keep (or my kids did).\n<\/p><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.jfedor.frozenbubble\" target=\"_blank\">Frozen Bubble<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.gunshippenguin.openflood\" target=\"_blank\">Open Flood<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/net.artificialworlds.rabbitescape\" target=\"_blank\">Rabbit Escape<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.happysanta.gd\" target=\"_blank\">Gravity Defied<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.serwylo.lexica\" target=\"_blank\">Lexica<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/se.tube42.kidsmem.android\" target=\"_blank\">Candy Memory<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.secuso.privacyfriendlydame\" target=\"_blank\">Checkers<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.eehouse.android.xw4\" target=\"_blank\">CrossWords<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.dodatheexploda\" target=\"_blank\">Doda The Exploda<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.petero.droidfish\" target=\"_blank\">DroidFish<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/csd.qtproject.minesweeper\" target=\"_blank\">Minesweeper<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/org.moire.opensudoku\" target=\"_blank\">Open Sudoku<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.phonefonefun\" target=\"_blank\">PhoneFoneFun<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/br.usp.ime.retrobreaker\" target=\"_blank\">Retro Breaker<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.quaap.fishberserker\" target=\"_blank\">Seafood Berserker!<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/com.mobilepearls.sokoban\" target=\"_blank\">Sokoban<\/a><\/li>\n<\/ul><ul><li> <a href=\"https:\/\/f-droid.org\/en\/packages\/wsdfhjxc.taponium\" target=\"_blank\">Taponium<\/a><\/li>\n<\/ul><h2 id=\"end\">End<\/h2><p>So that's it.  There are hundreds of other apps there which might fit your needs. Try <a href=\"https:\/\/f-droid.org\/\" target=\"_blank\">it<\/a>!\n<\/p><p>2018-06-22\n <a href=\"https:\/\/quaap.com\/D\/programming-index\">#programming<\/a> \u00a0 <a href=\"https:\/\/quaap.com\/D\/app-index\">#app<\/a> \u00a0<\/p>","created_at":"2018-11-07T16:23:12+0000","updated_at":"2018-11-07T16:23:12+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"quaap.com","preview_picture":"https:\/\/quaap.com\/D\/media\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":126,"title":"Review: 5 folding keyboards for your smartphone","url":"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html","content":"<p>While thumb-typing or finger-swiping on your smartphone screen may be good enough for tweeting or texting, more wordy activities -- like typing long email messages, taking notes at conference sessions or writing documents like this one -- are grueling at best. For that, you need a physical keyboard.<\/p><p>There's no shortage of Bluetooth keyboards intended for tablet users -- some freestanding, some also acting as a screen cover or carry case. Many of them can be used with smartphones, but since most are as big as a tablet, they are less than convenient.<\/p><p>One solution is a folding Bluetooth keyboard, which is not only useable by most smartphones, but small enough to easily carry around. In this roundup, I report on five folding keyboards intended for use with smartphones: Cervantes Mobile Jorno, iwerkz Universal Foldable Bluetooth Keyboard, Perixx Periboard-805L, Targus Universal Foldable Keyboard for Android and Zagg Pocket Keyboard.<\/p><p>(One recently announced keyboard that wasn't available at the time of review is the <a href=\"http:\/\/www.microsoft.com\/hardware\/en-us\/p\/universal-foldable-keyboard\" target=\"_blank\" rel=\"nofollow\">Microsoft Universal Foldable Keyboard<\/a>, introduced at the Mobile World Congress in March 2015 and due to ship in July. According to Microsoft, the keyboard will measure 11.6 in. open and 5.8 in. closed, weigh 7.1 oz. and work with Android, iOS and Windows 10 phones.)<\/p><p>To test the keyboards, I carried them around for several days and used them to type email and documents -- including the first draft of the review for the keyboard in question -- at home at my desk and dining room table, at the public library and at a few other places.<\/p><p>In addition to the sizes of the keyboard in open and closed positions (closed size includes the case\/stand, if any), I've included the the full width of the keys, from the left edge of the leftmost keys to the right edge of the rightmost keys (e.g., the Shift keys at either side of the keyboard). For comparison, the Shift-to-Shift width of a standard desktop keyboard is about 11 in.<\/p><p>All keyboards in this round-up recharge via a micro-USB port and include a USB\/micro-USB cable. All fold vertically except for the Zagg, which folds horizontally. The Targus has a dedicated top row of function keys (Home, Search, Cut\/Paste, media control, etc.), while the other four keyboards let the top row of numbers\/symbols also act as function keys when accessed in combination with the FN key. The iwerkz, Jorno and Perixx include an FN\/keyboard key combination that lets you suppress the virtual keyboard on the display to maximize display space.<\/p><div id=\"drr-container\" class=\"cat\" itemprop=\"reviewBody\"><section class=\"pagination\"><a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link prev\" rel=\"prev\"><i class=\"ss-icon ss-navigateleft\"> Previous<\/i><\/a>\n\t\t\n\t    \n\t\t    \n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link\">1<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link current\">2<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link\">3<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link\">4<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link\">5<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link\">6<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link\">7<\/a>\n\t\t\t    \n\t\t    \n\t\t    \n\t\t    \tPage 2\n\t\t    \n\t\t\n\t\t\n\t\t\t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link next\" rel=\"next\">Next <i class=\"ss-icon ss-navigateright\"><\/i><\/a>\n\t\t\n\t<\/section>\n\t\n\t<p><strong>Price:<\/strong> $99 (<a href=\"http:\/\/jornostore.com\/products\/jorno\" target=\"_blank\" rel=\"nofollow\">vendor price<\/a>) <br \/><strong>Size:<\/strong> 9.91 x 3.53 x 0.26 in. (open); 5.87 x 3.87 x 1.0 in. (closed) <br \/><strong>Full width of keys:<\/strong> 9.19 in. <br \/><strong>Weight:<\/strong> 6.5 oz. <br \/><strong>Compatible with:<\/strong> Android, iOS, Windows (tablets only)<\/p><figure class=\"large\"><img src=\"https:\/\/images.techhive.com\/images\/article\/2015\/04\/jorno-sam_0268-2-100577501-large.idge.jpg\" border=\"0\" alt=\"Cervantes Mobile's Jorno\" width=\"620\" height=\"414\" \/><small class=\"credit\">Cervantes<\/small> <figcaption>Cervantes Mobile Jorno\n<\/figcaption><\/figure><p>Cervantes Mobile's Jorno is a trifold keyboard with the middle segment being half the total width, and the left and right sides folding in.<\/p><p>What's distinctive about Jorno's design is the impressive metal hinges, which allow the keys in the folds to retain a standard keyboard's non-columnated positions -- exactly what touch-typists want. In addition, its \"shorter sides, longer middle\" configuration lets you type with it on your lap -- if you can keep your legs close enough together.<\/p><p>Jorno began as a Kickstarter project back in 2012, reaching about $5,000 over its $100,000 goal. Interestingly, the Jorno's initial design had four hinges, folding into a 3.5-in. square, with a much smaller case, but the design changed over time to the current rectangular-closed shape.<\/p><p>Jorno's <a href=\"http:\/\/jornostore.com\/pages\/specifications\" target=\"_blank\" rel=\"nofollow\">compatibility list<\/a> is the longest of all the keyboards in this roundup, including not just iOS and Android but also Windows\/Mac\/Linux-based tablets, BlackBerry, PlayStation, WebOS, and Symbian. (Of course, since all the keyboards reviewed here are Bluetooth, it's entirely possible that the other keyboards work with all these systems as well.)<\/p><p>The Jorno case opens up to form a stand, and has a 1.5-in. tab that can slide out in front if you want your phone or tablet to recline back at a lower viewing angle than the folded-case proper provides. Magnets in the case keep it in place in both the closed and stand configurations.<\/p><p>The Jorno has no visible switches -- opening and closing the keyboard turns it on. Pairing is initiated using the FN-C key combination; the C key also has a Bluetooth symbol. You tell the keyboard which OS you are pairing to via FN plus the labeled Q (for Android), W (for Windows) or E (for iOS) key. The FN key combined with the top row of keys also provides Home, Search, Cut\/Paste and media control.<\/p><p>According to the company, the keys are \"just 9% smaller than standard keys.\" The keyboard itself measures 9 in. Shift-to-Shift vs. 11 in. on a standard desktop keyboard, which makes the Jorno slightly over 80% of desktop size, according to my calculation. One quirk: The Delete key is at the upper right hand corner, not where my fingers expect it. I found I could be reasonably accurate typing on the Jorno (although, as was the case with all of these keyboards, I had to keep my eyes on things).<\/p><p>There is at least one distinct negative: When open, the underside of the hinges raise the keyboard like a pair of mini-speed bumps, so the keyboard doesn't lie flat. Every time I typed the keys on one edge, the other side seesawed up slightly. (I was able to cure this by folding two business cards in four, and taping one under each of the sides).<\/p><p>Even so, I recommend the Jorno, particularly if you are a touch-typist -- despite the problem with the hinges, having keys where your fingers expect them makes a big difference in ease of use.<\/p><\/div><div class=\"cat\" itemprop=\"reviewBody\"><section class=\"pagination\"><a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link prev\" rel=\"prev\"><i class=\"ss-icon ss-navigateleft\"> Previous<\/i><\/a>\n\t\t\n\t    \n\t\t    \n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link\">1<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link\">2<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link current\">3<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link\">4<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link\">5<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link\">6<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link\">7<\/a>\n\t\t\t    \n\t\t    \n\t\t    \n\t\t    \tPage 3\n\t\t    \n\t\t\n\t\t\n\t\t\t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link next\" rel=\"next\">Next <i class=\"ss-icon ss-navigateright\"><\/i><\/a>\n\t\t\n\t<\/section>\n\t\n\t<p><strong>Price:<\/strong> $30 (<a href=\"http:\/\/www.amazon.com\/Universal-Foldable-Bluetooth-Keyboard-44652BW\/dp\/B00FEYJOTK\" target=\"_blank\" rel=\"nofollow\">starting Amazon price<\/a>) <br \/><strong>Size:<\/strong> 11.5 x 3.5 x 0.25 in. (open); 5.87 x 3.5 x .88 in. (closed) <br \/><strong>Full width of keys:<\/strong> 9.56 in. <br \/><strong>Weight:<\/strong> 7.0 oz. <br \/><strong>Compatible with:<\/strong> Android, iOS, Windows (tablets only)<\/p><figure class=\"large\"><img src=\"https:\/\/images.techhive.com\/images\/article\/2015\/04\/black_iwerkz_bluetooth_keyboard-100577496-large.idge.jpg\" border=\"0\" alt=\"iwerkz Universal Foldable Bluetooth Keyboard\" width=\"620\" height=\"620\" \/><small class=\"credit\">iwerkz<\/small> <figcaption>iwerkz Universal Foldable Bluetooth Keyboard\n<\/figcaption><\/figure><p>The iwerkz Universal Foldable Bluetooth Keyboard includes a hard-plastic protective sleeve that also serves as a stand for smartphones and tablets, which you can prop up either via a pullout tab or a slot in the case. When folded up, the keyboard halves stay together with a magnetic latch -- useful if you aren't toting the protector sleeve.<\/p><p>The iwerkz keyboard is slightly smaller than the Targus -- almost the same width but slightly shorter. It's available in five colors -- white, black, blue, purple and red; the keys on all models are white. It offers a number of shortcut key combinations, including functions like search and media player control. There are also several Android\/iOS\/Windows-specific key combinations, which are listed in the single-sheet instruction guide.<\/p><p>When the keyboard's open, the on\/off switch is located on the top, with On and Off clearly labeled (which is not the case with some other keyboards such as the Targus). Pairing is initiated by a dedicated button labeled Connect.<\/p><p>The iwerkz lays nicely flat on a flat surface, making it easy to type on. However, because it has a straight-fold hinge in the middle, some keys aren't where touch-typists' fingers will expect them -- notably the \"B\" key, which is below the \"Y\" and \"H\" keys, rather than below the \"T\" and \"G\" keys. Similarly, the key for both a question mark and slash is at the bottom right, under the Shift key, and the arrow\/cursor keys are closer to the right edge than usual. This positioning meant I often typed an arrow key by accident, moving the cursor up to a previous line.<\/p><p>Otherwise, the iwerkz Universal Foldable Bluetooth Keyboard is a reasonable choice and is available at a nicely low price at many retailers.<\/p><h2>\u00a0<\/h2><\/div><div class=\"cat\" itemprop=\"reviewBody\"><section class=\"pagination\"><a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link prev\" rel=\"prev\"><i class=\"ss-icon ss-navigateleft\"> Previous<\/i><\/a>\n\t\t\n\t    \n\t\t    \n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link\">1<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link\">2<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link\">3<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link current\">4<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link\">5<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link\">6<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link\">7<\/a>\n\t\t\t    \n\t\t    \n\t\t    \n\t\t    \tPage 4\n\t\t    \n\t\t\n\t\t\n\t\t\t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link next\" rel=\"next\">Next <i class=\"ss-icon ss-navigateright\"><\/i><\/a>\n\t\t\n\t<\/section>\n\t\n\t<p><strong>Price:<\/strong> $30 (<a href=\"http:\/\/www.amazon.com\/Perixx-Bluetooth-Keyboard-PERIBOARD-805LII-US\/dp\/B00BU4OV50\/\" target=\"_blank\" rel=\"nofollow\">Amazon price<\/a>) <br \/><strong>Size:<\/strong> 12.5 x 3.75 x 0.31 in. (open); 6.75 x 3.75 x 0.62 in. (closed) <br \/><strong>Full width of keys:<\/strong> 11.0 in. <br \/><strong>Weight:<\/strong> 8.8 oz. <br \/><strong>Compatible with:<\/strong> Android, iOS, Windows (tablets only)<\/p><figure class=\"large\"><img src=\"https:\/\/images.techhive.com\/images\/article\/2015\/04\/periboard-805l-2us-100577497-large.idge.jpg\" border=\"0\" alt=\"Perixx Periboard-805L\" width=\"620\" height=\"357\" \/><small class=\"credit\">Perixx<\/small> <figcaption>Perixx Periboard-805L\n<\/figcaption><\/figure><p>The Perixx Periboard-805L is the closest to a full-size keyboard of all those reviewed here. (The 805L is the black model; it's available in white as the 805L W.) It measures the same 11 in. Shift-to-Shift as my desktop keyboard, although one-quarter inch of the Periboard's 11 in. goes to the hinge\/fold in the center.<\/p><p>The Periboard-805L's halves magnetically clasp together when folded up. There is no case, which is good for carry-ability, but also means there's no stand for your phone.<\/p><p>Like the other keyboards in this roundup, the Perixx Periboard-805L doesn't lock open. Interestingly, Perixx used to have a model that did called the Periboard-806, which is no longer available. (Perixx says it is working on a new model that will lock open.)<\/p><p>The power on\/off switch is on the side, clearly labeled, with the pairing button adjacent.<\/p><p>Unlike the iwerkz keyboard, the Periboard-805L's \"B\" key is on the left, under the \"T\" and \"G\" keys, just like on a standard desktop keyboard -- a more normal, usable position. However, the hinge fold creates a gap between the keys in the center, which takes a while to get used to. Overall, it felt like my fingers weren't automatically settling into what touch-typing calls the \"home position.\"<\/p><p>However, on the whole, if you prefer a full-width keyboard, don't mind the position shifts of some keys and don't need a case\/stand, the Perixx Periboard-805L may be a good choice.<\/p><\/div><div class=\"cat\" itemprop=\"reviewBody\"><section class=\"pagination\"><a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link prev\" rel=\"prev\"><i class=\"ss-icon ss-navigateleft\"> Previous<\/i><\/a>\n\t\t\n\t    \n\t\t    \n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link\">1<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link\">2<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link\">3<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link\">4<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link current\">5<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link\">6<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link\">7<\/a>\n\t\t\t    \n\t\t    \n\t\t    \n\t\t    \tPage 5\n\t\t    \n\t\t\n\t\t\n\t\t\t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link next\" rel=\"next\">Next <i class=\"ss-icon ss-navigateright\"><\/i><\/a>\n\t\t\n\t<\/section>\n\t\n\t<p><strong>Price:<\/strong> $29 (<a href=\"http:\/\/www.amazon.com\/Targus-Universal-Foldable-Keyboard-AKF001US\/dp\/B00D953A2I\/\" target=\"_blank\" rel=\"nofollow\">Android price<\/a>) <br \/><strong>Size:<\/strong> 11.6 x 3.75 x 0.25 in. (open); 6.0 x 4.0 x 0.75 in. (closed) <br \/><strong>Full width of keys:<\/strong> 10.1 in. <br \/><strong>Weight:<\/strong> 10.6 oz. <br \/><strong>Compatible with:<\/strong> Android, iOS<\/p><figure class=\"large\"><img src=\"https:\/\/images.techhive.com\/images\/article\/2015\/04\/targus_universal-foldable-keyboard-for-android-100577499-large.idge.jpeg\" border=\"0\" alt=\"targus universal foldable keyboard for android\" width=\"620\" height=\"491\" \/><small class=\"credit\">Targus<\/small> <figcaption>Targus Universal Foldable Keyboard for Android\n<\/figcaption><\/figure><p>Sometimes, a product's marketing can actually work against it. Despite its OS-specific name, the Targus Universal Foldable Keyboard for Android does work with iOS devices -- at least, it worked with my iPhone 4 and iPad 2, both running iOS 7.<\/p><p>The power switch is along the top edge, with the switch's positions labeled using extremely small raised letters; I found them difficult to read because they were the same color as the background. The Bluetooth pairing button is located on the underside on the keyboard; granted, you only tend to use it once or twice per session, but I found it annoying. The case\/stand took a bit of practice until I could open it smoothly; there's a latch to keep the keyboard closed.<\/p><p>Unlike the other four keyboards, which have five rows of keys (so the top number keys also act as function keys), the Targus has a dedicated top row of function keys for such duties as Search, Cut\/Paste and media control.<\/p><p>However, the keyboard has some quirks as well. For example, one key had a symbol that I didn't recognize and there's no \"what's what\" chart in the user manual. (I eventually called tech support, which said it lets you \"finger-mark a display region for copying.\") And even though the keyboard has an FN key, I didn't see any function markings on any of the other keys -- and anyway, the dedicated row of function keys makes the FN key moot.<\/p><p>Like the iwerkz keyboard, the Targus banishes the \"B\" key from under the \"T\" and \"G\" keys, to the right side of the fold, under the \"Y\" and \"H\" keys, which can slow down typing.<\/p><p>At 10.12 in. Shift-to-Shift, the Targus is the second widest keyboard in this roundup. Despite that, however, I found it the least comfortable to type on -- a combination of the center gap, the moved \"B\" key, and the overall size and space of the keys.<\/p><p>However, although I found the Targus too awkward for my taste, if you need a wide keyboard because you have a naturally wide reach, this one may work for you.<\/p><h2>\u00a0<\/h2><\/div><div class=\"cat\" itemprop=\"reviewBody\"><section class=\"pagination\"><a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link prev\" rel=\"prev\"><i class=\"ss-icon ss-navigateleft\"> Previous<\/i><\/a>\n\t\t\n\t    \n\t\t    \n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link\">1<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link\">2<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link\">3<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link\">4<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link\">5<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link current\">6<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link\">7<\/a>\n\t\t\t    \n\t\t    \n\t\t    \n\t\t    \tPage 6\n\t\t    \n\t\t\n\t\t\n\t\t\t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link next\" rel=\"next\">Next <i class=\"ss-icon ss-navigateright\"><\/i><\/a>\n\t\t\n\t<\/section>\n\t\n\t<p><strong>Price:<\/strong> $70 (<a href=\"http:\/\/www.zagg.com\/us\/en_us\/keyboards\/universal-tablet-keyboard\/pocket\" target=\"_blank\" rel=\"nofollow\">vendor price<\/a>); $52 (<a href=\"http:\/\/www.amazon.com\/ZAGG-Foldable-Wireless-Keyboard-Smartphones\/dp\/B00P7TQO18\/\" target=\"_blank\" rel=\"nofollow\">Amazon price<\/a>) <br \/><strong>Size:<\/strong> 8.75 x 6.87 x 0.3 in. (open); 8.75 x 2.12 x 0.5 in. (closed) <br \/><strong>Full width of keys:<\/strong> 7.8 in. <br \/><strong>Weight:<\/strong> 6.9 oz. <br \/><strong>Compatible with:<\/strong> Android, iOS<\/p><figure class=\"large\"><img src=\"https:\/\/images.techhive.com\/images\/article\/2015\/04\/zagg-pocket_quarter-100577498-large.idge.png\" border=\"0\" alt=\"zagg pocket keyboard\" width=\"620\" height=\"294\" \/><small class=\"credit\">Zagg<\/small> <figcaption>Zagg Pocket Keyboard\n<\/figcaption><\/figure><p>Zagg has a reputation for innovative design, so it's not surprising that the Zagg Pocket wireless Bluetooth keyboard folds horizontally rather than vertically, and incorporates a fold-out stand into the keyboard unit itself -- no separate case\/stand to carry (or lose).<\/p><p>Closed, the Zagg Pocket is a neat, slim shape (somewhat reminiscent of a slide rule, if you know what one of those is). The long-fold design means the keyboard won't fit into a pants pocket or a jacket side pocket, but will fit in any but the smallest of carry bags.<\/p><p>Like the Jorno keyboard, the Zagg powers on and off automatically when you open and close the keyboard, which is nice. Pairing is simple: press FN plus the \"Bluetooth\" key (which, on the Zagg, is on the Right-Arrow key), and, if need be, follow up with FN-A for Android, or FN-I for iOS.<\/p><p>The Zagg is smallest of the keyboards in this roundup. Although Zagg says that it delivers \"over 85% the size of a desktop keyboard,\" I measured the Zagg's width Shift-to-Shift at 7.8 in., which is closer to 70%. That's a bit narrow for my fingers, so I had to type a little slower than usual, and pay careful attention.<\/p><p>Like the Targus, the Zagg lacks a \"suppress virtual keyboard\" function key.<\/p><p>That said, the Zagg is a good choice -- no separate case, no odd-position keys, instant-on and it's the least bulky overall of the keyboards in this roundup. Just make sure you don't find the keyboard too small to type on.<\/p><\/div><div class=\"cat\" itemprop=\"reviewBody\"><section class=\"pagination\"><a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link prev\" rel=\"prev\"><i class=\"ss-icon ss-navigateleft\"> Previous<\/i><\/a>\n\t\t\n\t    \n\t\t    \n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html\" class=\"page-link\">1<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=2\" class=\"page-link\">2<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=3\" class=\"page-link\">3<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=4\" class=\"page-link\">4<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=5\" class=\"page-link\">5<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=6\" class=\"page-link\">6<\/a>\n\t\t\t    \n\t\t\t    \t<a href=\"https:\/\/www.computerworld.com\/article\/2903950\/mobile-wireless\/pocketable-productivity-5-folding-keyboards-for-your-smartphone.html?page=7\" class=\"page-link current\">7<\/a>\n\t\t\t    \n\t\t    \n\t\t    \n\t\t    \tPage 7\n\t\t    \n\t\t\n\t\t\n\t<\/section>\n\t\n\t<p>Because much of the usability of a keyboard depends on how large or small your fingers are, whether or not you're a touch typist and whether you need to be able to work without having a continuous flat surface to rest the keyboard on, there's no single winner in this roundup.<\/p><p>For touch typists, the Jorno and Zagg offer the least compromised key layouts -- and although both are a little smaller than I'm usually comfortable with (something that can't be avoided with a keyboard this portable), I'd call it a tie between the two. Of the others, the Perixx lost points with me for not having any kind of case (although I'm eager to see Perixx's new \"lock open\" model when it's available). The iwerkz and Targus lost points for sundry bad design choices.<\/p><p>But in the end, any of these keyboards beat the heck out of thumbing or finger-swiping.<\/p><\/div>","created_at":"2018-11-07T16:23:11+0000","updated_at":"2018-11-07T16:23:11+0000","published_at":"2015-04-10T03:30:00+0000","published_by":["Daniel P. Dern"],"annotations":[],"mimetype":"text\/html","reading_time":12,"domain_name":"www.computerworld.com","preview_picture":"https:\/\/images.techhive.com\/images\/article\/2015\/04\/dern-computerworld-foldingkeyboards-dsc03761-comparison-open-100577053-primary.idge.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":122,"title":"mrakgr\/The-Spiral-Language","url":"https:\/\/github.com\/mrakgr\/The-Spiral-Language","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><ul><li><a href=\"#table-of-contents\">Table of Contents<\/a><\/li>\n<li><a href=\"#the-spiral-language\">The Spiral Language<\/a>\n<ul><li><a href=\"#overview\">Overview<\/a>\n<ul><li><a href=\"#intro\">Intro<\/a><\/li>\n<li><a href=\"#design-philosophy\">Design Philosophy<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#dependencies\">Dependencies<\/a>\n- <a href=\"#for-the-compiler\">For the compiler:<\/a>\n- <a href=\"#for-the-cuda-using-spiral-libraries\">For the Cuda using Spiral libraries:<\/a><\/li>\n<li><a href=\"#tutorials-introduction-to-spiral\">Tutorials: Introduction to Spiral<\/a>\n<ul><li><a href=\"#0-the-way-to-use-the-language\">0: The way to use the language<\/a><\/li>\n<li><a href=\"#1-inlineables-methods-and-join-points\">1: Inlineables, Methods and Join Points<\/a>\n<ul><li><a href=\"#recursion-destructuring-and-pattern-matching\">Recursion, Destructuring and Pattern Matching<\/a>\n<ul><li><a href=\"#intensional-recursion\">Intensional Recursion<\/a><\/li>\n<li><a href=\"#term-casting-of-functions\">Term Casting of Functions<\/a><\/li>\n<li><a href=\"#function-error-message\"><code>&lt;function&gt;<\/code> error message<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#2-modules-macros-and-interop\">2: Modules, Macros and Interop<\/a>\n<ul><li><a href=\"#modules\">Modules<\/a><\/li>\n<li><a href=\"#macros\">Macros<\/a>\n<ul><li><a href=\"#solve-me\">Solve Me<\/a><\/li>\n<li><a href=\"#simple-array-sum-macro-version\">Simple Array Sum (macro version)<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#spiral-libraries\">Spiral libraries<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#3-loops-and-arrays\">3: Loops and Arrays<\/a>\n<ul><li><a href=\"#loops\">Loops<\/a><\/li>\n<li><a href=\"#arrays\">Arrays<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#3-union-types-and-lists\">3: Union Types and Lists<\/a>\n<ul><li><a href=\"#type-splitting-and-generic-parameters\">Type Splitting and Generic Parameters<\/a><\/li>\n<li><a href=\"#warning-on-combining-union-types-partial-active-patterns-and-join-points\">Warning on combining union types, partial active patterns and join points<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#4-continuation-passing-style-monadic-computation-and-parsing\">4: Continuation Passing Style, Monadic Computation and Parsing<\/a>\n<ul><li><a href=\"#parsing-benchmark\">Parsing Benchmark<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#5-tensors-and-structural-reflection\">5: Tensors and Structural Reflection<\/a>\n<ul><li><a href=\"#under-the-hood\">Under the Hood<\/a>\n<ul><li><a href=\"#layout-polymorphism\">Layout Polymorphism<\/a><\/li>\n<li><a href=\"#dimensionality-polymorphism\">Dimensionality Polymorphism<\/a>\n<ul><li><a href=\"#design-of-the-tensor\">Design of the Tensor<\/a><\/li>\n<li><a href=\"#the-tensor-facade\">The Tensor Facade<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#closing-comments\">Closing Comments<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#6-the-cuda-backend\">6: The Cuda Backend<\/a>\n<ul><li><a href=\"#intro-1\">Intro<\/a><\/li>\n<li><a href=\"#how-cuda-kernels-are-compiled\">How Cuda Kernels Are Compiled<\/a><\/li>\n<li><a href=\"#tour-of-the-standard-library-cuda-module\">Tour Of The Standard Library Cuda Module<\/a><\/li>\n<li><a href=\"#why-spiral-was-created\">Why Spiral Was Created<\/a>\n<ul><li><a href=\"#how-it-used-to-be-done\">How It Used To Be Done<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#7-object-orientation\">7: Object Orientation<\/a>\n<ul><li><a href=\"#motivation\">Motivation<\/a><\/li>\n<li><a href=\"#the-object\">The Object<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#8-gpu-programming-basics\">8: GPU Programming Basics<\/a>\n<ul><li><a href=\"#map\">map<\/a><\/li>\n<li><a href=\"#map_redo_map\">map_redo_map<\/a>\n<ul><li><a href=\"#flatten\">flatten<\/a><\/li>\n<li><a href=\"#cuda-loops\">Cuda Loops<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#d2_replicate_map\">d2_replicate_map<\/a>\n<ul><li><a href=\"#example\">Example<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#mapi_d2_redo_map\">mapi_d2_redo_map<\/a>\n<ul><li><a href=\"#example-1\">Example<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#mapi_d1_seq_broadcast\">mapi_d1_seq_broadcast<\/a>\n<ul><li><a href=\"#the-softmax-activation\">The Softmax Activation<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#9-deep-learning-basics\">9: Deep Learning Basics<\/a>\n<ul><li><a href=\"#primitives\">Primitives<\/a>\n<ul><li><a href=\"#map-1\">map<\/a><\/li>\n<li><a href=\"#map_redo_map-1\">map_redo_map<\/a><\/li>\n<li><a href=\"#d2_replicate_map-1\">d2_replicate_map<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#optimizers\">Optimizers<\/a><\/li>\n<li><a href=\"#operations\">Operations<\/a><\/li>\n<li><a href=\"#layers\">Layers<\/a>\n<ul><li><a href=\"#multiplicative-integration-rnn\">Multiplicative Integration RNN<\/a><\/li>\n<li><a href=\"#map--layer-norm--relu\">Map + Layer Norm + Relu<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#layer-combinators\">Layer Combinators<\/a>\n<ul><li><a href=\"#a-note-on-compilation-times\">A Note On Compilation Times<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#loops-1\">Loops<\/a><\/li>\n<li><a href=\"#example---feedforward-net-on-mnist\">Example - Feedforward Net On Mnist<\/a><\/li>\n<li><a href=\"#example---recurrent-net-on-tiny_shakespeare\">Example - Recurrent Net on tiny_shakespeare<\/a><\/li>\n<li><a href=\"#future-work\">Future Work<\/a>\n<ul><li><a href=\"#generic-matrix-multiplication\">Generic Matrix Multiplication<\/a><\/li>\n<li><a href=\"#fuse-the-weight-updates\">Fuse The Weight Updates<\/a><\/li>\n<li><a href=\"#metalearning-via-optimization\">Metalearning Via Optimization<\/a><\/li>\n<li><a href=\"#dilated-rnns\">Dilated RNNs<\/a><\/li>\n<li><a href=\"#dense-connections\">Dense Connections<\/a><\/li>\n<li><a href=\"#deep-layer-compilation-optimizations\">Deep Layer Compilation Optimizations<\/a><\/li>\n<li><a href=\"#generic-convolutions\">Generic Convolutions<\/a><\/li>\n<li><a href=\"#1d-block-scan-and-reduce\">1D Block Scan and Reduce<\/a><\/li>\n<li><a href=\"#softmax-cost\">Softmax Cost<\/a><\/li>\n<li><a href=\"#checkpointing\">Checkpointing<\/a><\/li>\n<li><a href=\"#improve-the-allocator\">Improve The Allocator<\/a><\/li>\n<li><a href=\"#plastic-neurons\">Plastic Neurons<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#user-guide-the-spiral-power\">User Guide: The Spiral Power<\/a>\n<ul><li><a href=\"#1-data-structures-abstraction-and-destructuring\">1: Data Structures, Abstraction and Destructuring<\/a><\/li>\n<li><a href=\"#2-let-insertion-and-common-subexpression-elimination\">2: Let Insertion and Common Subexpression Elimination<\/a><\/li>\n<li><a href=\"#3-the-if-statement\">3: The If Statement<\/a>\n<ul><li><a href=\"#raising-type-errors\">Raising Type Errors<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#4-boxing-of-union-types\">4: Boxing of Union Types<\/a><\/li>\n<li><a href=\"#5-unboxing-of-union-types\">5: Unboxing of Union Types<\/a><\/li>\n<li><a href=\"#6-join-points\">6: Join Points<\/a><\/li>\n<li><a href=\"#7-the-prepass-and-unused-variable-filtering\">7: The Prepass and Unused Variable Filtering<\/a><\/li>\n<li><a href=\"#8-pattern-compilation\">8: Pattern Compilation<\/a>\n<ul><li><a href=\"#type-equality-pattern\">Type Equality Pattern<\/a><\/li>\n<li><a href=\"#tuples\">Tuples<\/a><\/li>\n<li><a href=\"#active-patterns\">Active Patterns<\/a><\/li>\n<li><a href=\"#boolean-patterns\">Boolean Patterns<\/a><\/li>\n<li><a href=\"#literal-patterns\">Literal Patterns<\/a><\/li>\n<li><a href=\"#module-patterns\">Module Patterns<\/a><\/li>\n<li><a href=\"#when-pattern\">When Pattern<\/a><\/li>\n<li><a href=\"#term-function-type-pattern\">Term Function Type Pattern<\/a><\/li>\n<li><a href=\"#the-pos-pattern\">The Pos Pattern<\/a><\/li>\n<li><a href=\"#a-note-on-case\">A Note On Case<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#9-layout-types\">9: Layout Types<\/a><\/li>\n<li><a href=\"#10-macros\">10: Macros<\/a>\n<ul><li><a href=\"#parser-macros\">Parser Macros<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#11-operatives-and-the-core-library\">11: Operatives and the Core Library<\/a>\n<ul><li><a href=\"#virtualization\">Virtualization<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#known-bugs\">Known Bugs<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-overview\" class=\"anchor\" aria-hidden=\"true\" href=\"#overview\"><\/a>Overview<\/h2>\n<h3><a id=\"user-content-intro\" class=\"anchor\" aria-hidden=\"true\" href=\"#intro\"><\/a>Intro<\/h3>\n<p>As the world inexorably hurls towards the black maw of tomorrow, the power to face it is needed.<\/p>\n<p>Throughout the history of programming languages, the choice was between fast or expressive; the two traditions are crystallized by the C and the Lisp family of languages. There has been a lot of effort into this, but always as languages developed and moved forward they stepped away from the bare metal and in turn lost some of that core vitality that is needed for performance.<\/p>\n<p>The culprit for this is the heap allocation by default dogma introduced by Lisp decades ago. It is a crutch for languages with weak type systems.<\/p>\n<p>Abstraction by heap allocation is a dead end. It works moderately well on the current generation of computers where CPU is still the dominant driver of computation.<\/p>\n<p>It cannot work for devices like GPUs and the rest coming down the line. Many of the most important computational devices of the future won't support heap allocation so an alternative is needed to draw out their full power. It is of absolute importance that a language for that task have excellent control over inlining. Inlining, therefore, must come as a guarantee in the language and be a part of the type system.<\/p>\n<p>Inlining is a trade-off that expresses the exchange of memory for computation. It should be the default instead of heap allocating.<\/p>\n<p>A language good enough at propagating information so as to be capable of expressing inlining guarantees is also powerful enough for expressing a lot of other things well - without any abstraction overhead.<\/p>\n<ol><li>First class types.<\/li>\n<li>Structural reflection through pattern matching.<\/li>\n<li>Interoperability between different languages (such as F# and Cuda.)<\/li>\n<li>First class functions.<\/li>\n<li>Tuples as heterogeneous lists.<\/li>\n<li>First class modules.<\/li>\n<li>First class layouts of data structures.<\/li>\n<\/ol><p>Spiral is such a language.<\/p>\n<p>Statically typed and with a lightweight, very powerful type system giving it expressiveness of dynamic languages and the speed of C, Spiral is the crystallization of staged functional programming. It boasts of having intensional polymorphism and first-class staging. It was made for the sake of making a deep learning library which was too difficult to do in F# itself for its author.<\/p>\n<h3><a id=\"user-content-design-philosophy\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-philosophy\"><\/a>Design Philosophy<\/h3>\n<p>Automatically doing type inference, inlining and other optimizations require restrictions and heuristics in order to ensure termination. Languages that make the choice of automating the important parts of their internals invariably hamstring their expressiveness. Even if they end up doing well on some low-level benchmarks, they perform poorly when high-level abstractions are required.<\/p>\n<p>Spiral is different. The power of Spiral lies in its novel kind of language design, not compiler smartness. Spiral is a static language without any restrictions on either type inference or optimizations.<\/p>\n<p>It introduces novel constructs such as inlineables and join points, exposes optimizations as polymorphism in its type system, and then ties them together and with the rest of the features in such a manner so as to allow programs to be written so they terminate.<\/p>\n<p>The halting problem is the primary obstacle preventing the bridging of expressiveness and performance. Ultimately, C is a competing style more than it is a competing language. It is something other higher level languages regress to once they start worrying about performance.<\/p>\n<p>In Spiral, inlining by hand will never be necessary.<\/p>\n<p>In Spiral, the most abstract way of writing a program is also the optimal one from a performance standpoint.<\/p>\n<p>Spiral exists to abstract away optimization.<\/p>\n<h2><a id=\"user-content-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#dependencies\"><\/a>Dependencies<\/h2>\n<h5><a id=\"user-content-for-the-compiler\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-the-compiler\"><\/a>For the compiler:<\/h5>\n<p>FParsec<\/p>\n<p>Visual Studio 2017 F# template (.NET desktop development)<\/p>\n<h5><a id=\"user-content-for-the-cuda-using-spiral-libraries\" class=\"anchor\" aria-hidden=\"true\" href=\"#for-the-cuda-using-spiral-libraries\"><\/a>For the Cuda using Spiral libraries:<\/h5>\n<p>ManagedCuda 8.0 + (CUBLAS,CURAND)<\/p>\n<p>Cuda SDK 8.0 + 9.0 (8.0 for the libraries and 9.0 for the NVCC compiler)<\/p>\n<p>The Cuda Unbound library<\/p>\n<p>Visual C++ version 15.4 v14.11 toolset<\/p>\n<h2><a id=\"user-content-tutorials-introduction-to-spiral\" class=\"anchor\" aria-hidden=\"true\" href=\"#tutorials-introduction-to-spiral\"><\/a>Tutorials: Introduction to Spiral<\/h2>\n<h3><a id=\"user-content-0-the-way-to-use-the-language\" class=\"anchor\" aria-hidden=\"true\" href=\"#0-the-way-to-use-the-language\"><\/a>0: The way to use the language<\/h3>\n<p>The easiest way to do it right now would be to clone this repo. In the Testing project, look at <code>run.fs<\/code>. It has the latest example used for the tutorial. Select the <code>Testing<\/code> project as the starter one and point the output to the <code>output.fs<\/code> file in the <code>Temporary<\/code> project. No need to worry about getting it wrong - at worst an exception will be raised.<\/p>\n<p>Modifying the Cuda configuration options in the <code>run.fs<\/code> file unless usage of libraries related to that is required.<\/p>\n<h3><a id=\"user-content-1-inlineables-methods-and-join-points\" class=\"anchor\" aria-hidden=\"true\" href=\"#1-inlineables-methods-and-join-points\"><\/a>1: Inlineables, Methods and Join Points<\/h3>\n<p>Spiral has a great many similarities to other languages of the ML family, most notably F# with whom it shares the most similarity and a great deal of syntax, but in terms of semantics, it is different at its core.<\/p>\n<pre>inl x = 2 \/\/ Define a 64-bit integer in Spiral.\n<\/pre>\n<pre>let x = 2 \/\/ Define a 32-bit integer in F#.\n<\/pre>\n<p>Unlike in F#, statements and function definitions in Spiral are preceded by <code>inl<\/code> instead of <code>let<\/code> which is short for <code>inl<\/code>ineable.<\/p>\n<pre>let x = 2\n()\n<\/pre>\n<p>If a program like the above was disassembled, <code>x<\/code> would compile down to a public method in F#. In Spiral, by contrast, there would be nothing.<\/p>\n<pre>module SpiralExample.Main\nlet cuda_kernels = \"\"\"\nextern \"C\" {\n    \n}\n\"\"\"\n<\/pre>\n<p>It is not absolutely nothing though. If the program used Cuda kernels in it, they would be gathered at the top of the file inside the <code>cuda_kernels<\/code> variable. For interests of brevity, the unremarkable top part will be cut out during the tutorials.<\/p>\n<pre>inl x = dyn 2\n<\/pre>\n<pre>let (var_0: int64) = 2L \/\/ Generated F# code.\n<\/pre>\n<p>The reason Spiral is generating nothing is that <code>2<\/code> as defined is a literal and gets tracked as such by the partial evaluator inside the environment. In order to have it appear in the generated code, it is necessary to cast it from the type to the term level using <code>dyn<\/code>amize function. From here on out, the literal will be bound to a variable and the binding <code>x<\/code> will track <code>var_0<\/code> instead.<\/p>\n<p>Being able to do this is useful for a variety of reasons. For example, without it constructs such as runtime loops would be impossible to write in Spiral because the partial evaluator would diverge. Despite its static typing features, the language would essentially be constrained to being an interpreter for a pure dynamic functional language.<\/p>\n<pre>inl x = dyn 2\ninl y = 3\n(x + y) * (x + y)\n<\/pre>\n<pre>let (var_0: int64) = 2L\nlet (var_1: int64) = (var_0 + 3L)\n(var_1 * var_1)\n<\/pre>\n<p>To get a sense of how <code>dyn<\/code> works, here is a slightly more complex example. The evaluator does common subexpression elimination in the local scope and so it is smart enough to optimize the <code>x+y<\/code> into a single addition and a multiplication.<\/p>\n<pre>inl x = 2\ninl y = 3\n(x + y) * (x + y)\n<\/pre>\n<pre>25L\n<\/pre>\n<p>Without <code>dyn<\/code>, all the arithmetic operations get evaluated at compile time. This is due to the simple fact that a variable added to a literal is a variable. In general, if an operation has a variable as one of its inputs, then its output will also be a variable. The evaluator term casts the literals when necessary. For an operation to be evaluated at compile time, the partial evaluator must have support for it internally.<\/p>\n<p><code>inl<\/code> can also be used to define functions.<\/p>\n<pre>inl add a b = a + b\nadd 1 2, add 3 4\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nTuple0(3L, 7L)\n<\/pre>\n<p>Tuples are used in Spiral much in the same way as in other functional languages. As per their name, <code>inl<\/code>ineables are always inlined. There are no heuristics at play here in the evaluator. Spiral's staged functions are tracked at the type level and both their exact body and environments are known at every point of compilation. This is an absolute guarantee and there is no point at which they can be automatically converted into heap allocated closures.<\/p>\n<pre>inl add a b = a + b\ninl f = add 1\nf 2, f 3\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nTuple0(3L, 4L)\n<\/pre>\n<p>Of course, they can be partially applied.<\/p>\n<p>Besides being staged, Spiral's functions allow more powerful forms of polymorphism than F#'s.<\/p>\n<pre>inl mult a b = a * b\ninl f g = g 1 2, g 3.0 4.0 \/\/ Would give a type error in F#.\nf mult\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nTuple0(2L, 12.000000)\n<\/pre>\n<p>Haskell could manage something like that using typeclasses and higher ranked types, but it would be nowhere as simple. Having to declare new types and then putting in the necessary annotations into the function would be needed.<\/p>\n<p>On the other hand, being able to do this makes type inference for Spiral undecidable.<\/p>\n<pre>inl f _ = 1 + \"qwe\" \/\/ Does not give a type error\n()\n<\/pre>\n<p>Undecidability manifests in Spiral like so - the body of the function is not evaluated until it is applied. That means that type errors can lurk in functions that are unused.<\/p>\n<p>That being said, Spiral is a statically typed language and any type errors for code on the execution path will get reported at compile time along with the trace for it.<\/p>\n<pre>inl f _ = 1 + \"qwe\" \/\/ Does not give a type error\ninl _ = 1\ninl _ = 2\ninl _ = 3\nf ()\n<\/pre>\n<pre>Error trace on line: 3, column: 5 in file \"example1\".\ninl _ = 1\n    ^\nError trace on line: 4, column: 5 in file \"example1\".\ninl _ = 2\n    ^\nError trace on line: 5, column: 5 in file \"example1\".\ninl _ = 3\n    ^\nError trace on line: 6, column: 1 in file \"example1\".\nf ()\n^\nError trace on line: 2, column: 7 in file \"example1\".\ninl f _ = 1 + \"qwe\" \/\/ Does not give a type error\n      ^\nError trace on line: 51, column: 11 in file \"Core\".\ninl (+) a b = !Add(a,b)\n          ^\n`is_numeric a &amp;&amp; get_type a = get_type b` is false.\na=lit 1i64, b=lit qwe\n<\/pre>\n<p>Since methods are necessary to achieve code reuse, Spiral makes it possible to make use of them with the <code>met<\/code> keyword. It works much like <code>inl<\/code>.<\/p>\n<pre>inl mult a b = a * b\nmet f g = g 1 2, g 3.0 4.0\nf mult\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0(): Tuple0 =\n    Tuple0(2L, 12.000000)\nmethod_0()\n<\/pre>\n<p><code>mult<\/code> can also be defined using <code>met<\/code> and passed into the other function.<\/p>\n<pre>met mult a b = a * b\nmet f g = g 1 2, g 3.0 4.0\nf mult\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0(): Tuple0 =\n    let (var_0: int64) = method_1()\n    let (var_1: float) = method_2()\n    Tuple0(var_0, var_1)\nand method_1(): int64 =\n    2L\nand method_2(): float =\n    12.000000\nmethod_0()\n<\/pre>\n<p>The result would not be as one might expect since the methods would get specialized to the literal arguments passed to them. Instead, it would be better to use <code>dyn<\/code> here. But rather than letting the caller <code>f<\/code> do the term casting operation, it would be better if the callee <code>mult<\/code> did it.<\/p>\n<pre>met mult (!dyn a) (!dyn b) = a * b\nmet f g = g 1 2, g 3.0 4.0\nf mult\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0(): Tuple0 =\n    let (var_0: int64) = 1L\n    let (var_1: int64) = 2L\n    let (var_2: int64) = method_1((var_0: int64), (var_1: int64))\n    let (var_3: float) = 3.000000\n    let (var_4: float) = 4.000000\n    let (var_5: float) = method_2((var_3: float), (var_4: float))\n    Tuple0(var_2, var_5)\nand method_1((var_0: int64), (var_1: int64)): int64 =\n    (var_0 * var_1)\nand method_2((var_0: float), (var_1: float)): float =\n    (var_0 * var_1)\nmethod_0()\n<\/pre>\n<p><code>!<\/code> on the left (the pattern) side of the expression is the active pattern unary operator. It takes a function as its first argument, applies the input to it and rebinds the result to the second argument of the pattern (in this case <code>a<\/code> and <code>b<\/code> respectively) before the body is evaluated.<\/p>\n<h4><a id=\"user-content-recursion-destructuring-and-pattern-matching\" class=\"anchor\" aria-hidden=\"true\" href=\"#recursion-destructuring-and-pattern-matching\"><\/a>Recursion, Destructuring and Pattern Matching<\/h4>\n<p>Much like in F#, recursive functions can be defined using <code>rec<\/code>.<\/p>\n<pre>inl rec foldl f s = function\n    | x :: xs -&gt; foldl f (f s x) xs\n    | () -&gt; s\ninl sum = foldl (+) 0\nsum (1,2,3)\n<\/pre>\n<pre>6L\n<\/pre>\n<p>In the ML family of languages, <code>::<\/code> is the list cons pattern. In Spiral it is the tuple cons pattern. Tuples are fully fledged heterogeneous lists in Spiral and can be treated as such.<\/p>\n<pre>inl a = 2,3\n1 :: a\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nTuple0(1L, 2L, 3L)\n<\/pre>\n<p>There are some interesting applications of this.<\/p>\n<pre>inl rec foldl f s = function\n    | x :: xs -&gt; foldl f (f s x) xs\n    | () -&gt; s\nmet sum (!dyn l) = foldl (+) 0 l\nsum (1,2,3), sum (1,2,3,4)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0((var_0: int64), (var_1: int64), (var_2: int64)): int64 =\n    let (var_3: int64) = (var_0 + var_1)\n    (var_3 + var_2)\nand method_1((var_0: int64), (var_1: int64), (var_2: int64), (var_3: int64)): int64 =\n    let (var_4: int64) = (var_0 + var_1)\n    let (var_5: int64) = (var_4 + var_2)\n    (var_5 + var_3)\nlet (var_0: int64) = 1L\nlet (var_1: int64) = 2L\nlet (var_2: int64) = 3L\nlet (var_3: int64) = method_0((var_0: int64), (var_1: int64), (var_2: int64))\nlet (var_4: int64) = 1L\nlet (var_5: int64) = 2L\nlet (var_6: int64) = 3L\nlet (var_7: int64) = 4L\nlet (var_8: int64) = method_1((var_4: int64), (var_5: int64), (var_6: int64), (var_7: int64))\nTuple0(var_3, var_8)\n<\/pre>\n<p>While the language allows variant arguments, Spiral has the ability to specialize methods to their exact arguments and in combination with destructuring to implement variant arguments in a typesafe manner.<\/p>\n<pre>inl rec foldl f s = function\n    | x :: xs -&gt; foldl f (f s x) xs\n    | () -&gt; s\nmet sum l = foldl (+) 0 l\nsum (1,2,3,dyn 4), sum (2,2,3,dyn 4)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0((var_0: int64)): int64 =\n    (6L + var_0)\nand method_1((var_0: int64)): int64 =\n    (7L + var_0)\nlet (var_0: int64) = 4L\nlet (var_1: int64) = method_0((var_0: int64))\nlet (var_2: int64) = 4L\nlet (var_3: int64) = method_1((var_2: int64))\nTuple0(var_1, var_3)\n<\/pre>\n<p>By default, in Spiral tuples and modules are flattened and have their variables tracked individually. As can be seen in the generated code, when a tuple is passed into a function it is not the actual tuple that is being passed into it, but its arguments instead.<\/p>\n<p>The specialization is exact to the structure, not just the types. If literals are being passed through the method call, the method will get specialized to them.<\/p>\n<pre>met f _ = 1,2,3\ninl x = f ()\n0\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nlet rec method_0(): Tuple0 =\n    Tuple0(1L, 2L, 3L)\nlet (var_0: Tuple0) = method_0()\nlet (var_1: int64) = var_0.mem_0\nlet (var_2: int64) = var_0.mem_1\nlet (var_3: int64) = var_0.mem_2\n0L\n<\/pre>\n<p>Tuples themselves are only manifested in the generated code on join point and branch returns. They get destructured right away and tracked by their individual variables on binding and function application. Had the method return not been bound to <code>x<\/code>, it would not have been destructured. This is the desired behavior because otherwise destructuring might block tail call optimizations.<\/p>\n<pre>met f _ = 1\n<\/pre>\n<pre>inl f _ = join 1\n<\/pre>\n<p>The above two code fragments are identical in Spiral. <code>met<\/code> is just syntax sugar for a function with a join point around its body.<\/p>\n<p>Being able to do this is quite powerful as it allows more fine-grained control over inlining.<\/p>\n<pre>inl rec foldl f s = function\n    | x :: xs -&gt; foldl f (f s x) xs\n    | () -&gt; s\ninl rec forall f = function\n    | x :: xs -&gt; f x &amp;&amp; forall f xs\n    | () -&gt; true\ninl sum l = \n    if forall lit_is l then foldl (+) 0 l\n    else join foldl (+) 0 (dyn l)\nsum (1,2,3,4), sum (1,2,3,dyn 4)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0((var_0: int64)): int64 =\n    let (var_1: int64) = 1L\n    let (var_2: int64) = 2L\n    let (var_3: int64) = 3L\n    let (var_4: int64) = (var_1 + var_2)\n    let (var_5: int64) = (var_4 + var_3)\n    (var_5 + var_0)\nlet (var_0: int64) = 4L\nlet (var_1: int64) = method_0((var_0: int64))\nTuple0(10L, var_1)\n<\/pre>\n<p>The <code>lit_is<\/code> always resolves at compile time to either <code>true<\/code> or <code>false<\/code> just like other structure testing functions. In combination with <code>forall<\/code> that allows for testing of whether all the arguments of <code>l<\/code> are known at compile time. Then using a static if, the two branches amount to either summing them all at compile time, or term casting them and pushing the work to runtime.<\/p>\n<p>This ensures that the sum function does not get specialized for every arbitrary literal passed into it.<\/p>\n<pre>if true then 1 else \"qwe\" \/\/ Not a type error.\n<\/pre>\n<pre>if dyn true then 1 else \"qwe\" \/\/ A type error.\n<\/pre>\n<p>If statements in Spiral default to evaluating only one branch if their conditional is known at compile time meaning they are static by default. This is the bedrock of its intensional (structural) polymorphism. Under the hood, the patterns get compiled to static if statements which allow it to branch on structures and types.<\/p>\n<p><code>if<\/code> goes hand in hand with join point specialization.<\/p>\n<pre>inl default_of = function\n    | _: int64 -&gt; 0\n    | _: float64 -&gt; 0.0\n    | _: string -&gt; \"\"\ndefault_of 1, default_of 1.0, default_of \"qwe\"\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    val mem_2: string\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nTuple0(0L, 0.000000, \"\")\n<\/pre>\n<p>Unlike ML languages which use Hindley-Milner global type inference, Spiral does not infer as much as propagate. A consequence of that besides undecidability is that it knows the exact structure and type of everything at all times. When done on non-union types as done up to now, this sort of branching has no runtime overhead whatsoever and can be readily seen by looking into the generated code.<\/p>\n<p><a href=\"https:\/\/stackoverflow.com\/questions\/1839016\/f-explicit-match-vs-function-syntax\" rel=\"nofollow\"><code>function<\/code><\/a> is just shorthand for matching on the immediate argument like in F#.<\/p>\n<p>One other thing that is different from F# is that <code>int64<\/code>,<code>float64<\/code> and <code>string<\/code> on the right side of the <code>:<\/code> operators are not type annotations, but standard variables. The types in Spiral are first class much like everything else.<\/p>\n<pre>inl int64_type = type 1\ninl float64_type = type 1.0\ninl string_type = type \"qwe\"\ninl default_of = function\n    | _: int64_type -&gt; 0\n    | _: float64_type -&gt; 0.0\n    | _: string_type -&gt; \"\"\ndefault_of 1, default_of 1.0, default_of \"qwe\"\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    val mem_2: string\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nTuple0(0L, 0.000000, \"\")\n<\/pre>\n<p>As can be seen, the two generated code fragments are identical. <code>:<\/code> on the pattern side is the type equality operator. It can be invoked outside the pattern using the <code>eq_type<\/code> function.<\/p>\n<p><code>type<\/code> is a keyword and like <code>join<\/code> it enters a new scope.<\/p>\n<p>The types themselves can do more than being passed around or matched on.<\/p>\n<pre>inl int64_type = type 1\ninl float64_type = type 1.0\ninl string_type = type \"qwe\"\ninl default_of = function\n    | _: int64_type -&gt; 0\n    | _: float64_type -&gt; 0.0\n    | _: string_type -&gt; \"\"\ninl a = type int64_type + 1\ninl b = type float64_type + 1.0\ninl c = type string_format \"{0}, {1}\" (string_type, \"rty\")\ndefault_of a, default_of b, default_of c\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    val mem_2: string\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nTuple0(0L, 0.000000, \"\")\n<\/pre>\n<p>The above is just to illustrate that inside the evaluator naked types are treated just like variables of the same type. If they should happen to slip on the term level that would cause an error.<\/p>\n<pre>int64 + 3\n<\/pre>\n<pre>(naked_type (*int64*) + 3L)\n<\/pre>\n<p>These kinds of errors are easier to locate when they are shown in generated code. When they happen it is usually because of a missed argument to a curried function which causes its environment to spill into the generated code. This makes the usual error messages unhelpful, but looking at the error code gives a good indication of what is happening.<\/p>\n<h5><a id=\"user-content-intensional-recursion\" class=\"anchor\" aria-hidden=\"true\" href=\"#intensional-recursion\"><\/a>Intensional Recursion<\/h5>\n<p>Spiral, in general, does not need type annotations. The only exceptions are recursive functions when used in tandem with join points.<\/p>\n<pre>met rec fact (!dyn x) = if x &gt; 1 then x * fact (x-1) else 1\nfact 3\n<\/pre>\n<pre>Process is terminated due to StackOverflowException.\n<\/pre>\n<p>The correct way to write the above would be.<\/p>\n<pre>met rec fact (!dyn x) = \n    if x &gt; 1 then x * fact (x-1) else 1\n    : int64 \/\/ or alternatively `: x`\nfact 3\n<\/pre>\n<pre>let rec method_0((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &gt; 1L)\n    if var_1 then\n        let (var_2: int64) = (var_0 - 1L)\n        let (var_3: int64) = method_0((var_2: int64))\n        (var_0 * var_3)\n    else\n        1L\nlet (var_0: int64) = 3L\nmethod_0((var_0: int64))\n<\/pre>\n<p><code>:<\/code> has the lowest precedence of all Spiral's constructs so it will get applied before any of the statements. It does not necessarily have to be put directly into the function. As a reminder, on the pattern side <code>:<\/code> is not a type annotation, but a type equality test.<\/p>\n<pre>inl rec fact x =\n    inl body x = if x &gt; 1 then x * fact (x-1) else 1\n    if lit_is x then body x\n    else join (body x : int64)\nfact 3, fact (dyn 3)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &gt; 1L)\n    if var_1 then\n        let (var_2: int64) = (var_0 - 1L)\n        let (var_3: int64) = method_0((var_2: int64))\n        (var_0 * var_3)\n    else\n        1L\nlet (var_0: int64) = 3L\nlet (var_1: int64) = method_0((var_0: int64))\nTuple0(6L, var_1)\n<\/pre>\n<p>It takes some work, but it is not difficult to make functions stage polymorphic in Spiral.<\/p>\n<p>Mutual recursion can also be done using join points.<\/p>\n<pre>\/\/ https:\/\/en.wikipedia.org\/wiki\/Hofstadter_sequence#Hofstadter_Female_and_Male_sequences\ninl rec hof x = \n    inl male n = if n &gt; 0 then n - hof.female (hof.male (n-1)) else 0\n    inl female n = if n &gt; 0 then n - hof.male (hof.female (n-1)) else 1\n    match x with\n    | .male (!dyn n) -&gt; join male n : int64\n    | .female (!dyn n) -&gt; join female n : int64\nhof.male 3\n<\/pre>\n<pre>let rec method_0((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &gt; 0L)\n    if var_1 then\n        let (var_2: int64) = (var_0 - 1L)\n        let (var_3: int64) = method_0((var_2: int64))\n        let (var_4: int64) = method_1((var_3: int64))\n        (var_0 - var_4)\n    else\n        0L\nand method_1((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &gt; 0L)\n    if var_1 then\n        let (var_2: int64) = (var_0 - 1L)\n        let (var_3: int64) = method_1((var_2: int64))\n        let (var_4: int64) = method_0((var_3: int64))\n        (var_0 - var_4)\n    else\n        1L\nlet (var_0: int64) = 3L\nmethod_0((var_0: int64))\n<\/pre>\n<p><code>.<\/code> here is the type literal lift operator. It has special syntax for strings and when used directly next to an expression, it binds more tightly than application similar to how F#'s method access works. It also has its own dedicated pattern as shown above.<\/p>\n<pre>inl f x = .(x)\ninl a = f \"asd\"\ninl b = .asd\neq_type a b\n<\/pre>\n<pre>true\n<\/pre>\n<p>It works on any kind of literal, not just strings. Type literals can be converted to ordinary literals as well.<\/p>\n<pre>inl a = .1\ninl b = .2\nmatch a,b with\n| .(a), .(b) -&gt; a + b\n<\/pre>\n<pre>3L\n<\/pre>\n<p>The difference between type literals and ordinary literals is that type literals will always be erased in generated code and it is impossible to push them at runtime by <code>dyn<\/code>ing them.<\/p>\n<pre>dyn (.a,\"b\",.false,true)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: string\n    val mem_1: bool\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: string) = \"b\"\nlet (var_1: bool) = true\nTuple0(var_0, var_1)\n<\/pre>\n<p>Using <code>print_static<\/code> can be used to inspect what the evaluator sees at compile time.<\/p>\n<pre>print_static (dyn (.a,\"b\",.false,true))\n<\/pre>\n<pre>[type (type_lit (a)), var (string), type (type_lit (false)), var (bool)]\n<\/pre>\n<p>All the information in type literals is preserved at all times.<\/p>\n<h5><a id=\"user-content-term-casting-of-functions\" class=\"anchor\" aria-hidden=\"true\" href=\"#term-casting-of-functions\"><\/a>Term Casting of Functions<\/h5>\n<p>Spiral's functions, as flexible as they are, have the notable weakness of not being able to emulate recursive datatypes. For that, they need to be cast to the term level.<\/p>\n<p>Consider a silly example like the following where the function is used as a counter.<\/p>\n<pre>met rec loop f (!dyn i) =\n    if i &lt; 10 then loop (inl _ -&gt; f() + 1) (i + 1)\n    else f()\n    : int64\nloop (inl _ -&gt; 0) 0\n<\/pre>\n<p>This will never compile for the reason that <code>f<\/code> continually expands its environment.<\/p>\n<p>At first it tries to specialize the function for just <code>inl _ -&gt; 0<\/code> -&gt; <code>int64<\/code> -&gt; <code>int64<\/code>. The second specialization it tries is <code>[inl _ -&gt; 0; inl _ -&gt; f() + 1]<\/code> -&gt; <code>int64<\/code> -&gt; <code>int64<\/code>. During the third it is trying to specialize it for <code>[inl _ -&gt; 0; inl _ -&gt; f() + 1; inl _ -&gt; f() + 1]<\/code> -&gt; <code>int64<\/code> -&gt; <code>int64<\/code>. The syntax used here is just for the sake of description. The problem is the it is impossible for the compiler to ever terminate on the above program. The only way to do it would be to cast the function to the term level and track it as a variable.<\/p>\n<pre>inl rec loop f i =\n    inl f, i = term_cast f (), dyn i\n    join \n        if i &lt; 10 then loop (inl _ -&gt; f() + 1) (i + 1) else f()\n        : int64\nloop (inl _ -&gt; 0) 0\n<\/pre>\n<pre>let rec method_0 (): int64 =\n    0L\nand method_1((var_0: (unit -&gt; int64)), (var_1: int64)): int64 =\n    let (var_2: bool) = (var_1 &lt; 10L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_4: (unit -&gt; int64)) = method_2((var_0: (unit -&gt; int64)))\n        method_1((var_4: (unit -&gt; int64)), (var_3: int64))\n    else\n        var_0()\nand method_2 ((var_0: (unit -&gt; int64))) (): int64 =\n    let (var_1: int64) = var_0()\n    (var_1 + 1L)\nlet (var_0: (unit -&gt; int64)) = method_0\nlet (var_1: int64) = 0L\nmethod_1((var_0: (unit -&gt; int64)), (var_1: int64))\n<\/pre>\n<p><code>term_cast<\/code> works by taking a function as its first argument and a type as its second. It emulates a function call, gets the return type of the term function from the result of that, and set the input type to the first argument. In the generated code, it flattens the arguments a single tuple level.<\/p>\n<p>Term level functions have their environments hidden and the only information available to the evaluator is its type.<\/p>\n<p>On the Cuda side, term functions are also allowed with the restriction that their environments be empty. Meaning, they cannot capture variables in their lexical scope and can only be used as function pointers. Despite that restriction, they are useful for interop with Cuda libraries.<\/p>\n<p>All the features of Spiral with the exception of heap allocated modules and closures can be used on the Cuda side.<\/p>\n<h5><a id=\"user-content-function-error-message\" class=\"anchor\" aria-hidden=\"true\" href=\"#function-error-message\"><\/a><code>&lt;function&gt;<\/code> error message<\/h5>\n<p>Don't be fooled by the <code>&lt;function&gt;<\/code> during type errors. As was repeatedly stated, functions are not at all opaque - they are fully transparent to the evaluator. The reason why they get printed like that is simply that they have a tendency to suck everything into the environment. And except for very small examples, trying to print out the raw AST of its body is worthless even for debugging as it is so convoluted.<\/p>\n<pre>if dyn true then\n    inl a,b = dyn (1,2)\n    inl _ -&gt; a + b\nelse\n    inl a,b = dyn (3.0,4.0)\n    inl _ -&gt; a + b\n<\/pre>\n<pre>if dyn true then\n^\nTypes in branches of If do not match.\nGot: &lt;function&gt; and &lt;function&gt;\n<\/pre>\n<p>If functions have the same bodies, they can be returned from branches of a dynamic if statement if they also have the same environments.<\/p>\n<pre>if dyn true then\n    inl a,b = dyn (1,2)\n    inl _ -&gt; a + b\nelse\n    inl a,b = dyn (3,4)\n    inl _ -&gt; a + b\n<\/pre>\n<pre>type Env0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: bool) = true\nif var_0 then\n    let (var_1: int64) = 1L\n    let (var_2: int64) = 2L\n    (Env0(var_1, var_2))\nelse\n    let (var_3: int64) = 3L\n    let (var_4: int64) = 4L\n    (Env0(var_3, var_4))\n<\/pre>\n<h3><a id=\"user-content-2-modules-macros-and-interop\" class=\"anchor\" aria-hidden=\"true\" href=\"#2-modules-macros-and-interop\"><\/a>2: Modules, Macros and Interop<\/h3>\n<h4><a id=\"user-content-modules\" class=\"anchor\" aria-hidden=\"true\" href=\"#modules\"><\/a>Modules<\/h4>\n<p>Owing to Spiral's relatively dynamic nature, modules work much like <a href=\"https:\/\/docs.microsoft.com\/en-us\/dotnet\/fsharp\/language-reference\/records\" rel=\"nofollow\">records<\/a> do in F# albeit with greatly expanded functionality. Unlike in F#, they do not need to be predefined.<\/p>\n<pre>inl f m =\n    open m\n    q + w + e\ninl m1 = {q=1; w=2; e=3}\ninl m2 = {q=1.0; w=2.0; e=3.0}\nf m1, f m2\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: float\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nTuple0(6L, 6.000000)\n<\/pre>\n<p>As per their namesake, they can be opened and passed as arguments. They have their own dedicated patterns.<\/p>\n<pre>inl f {q w e} = q + w + e\ninl q = 1\ninl w = 2\ninl e = 3\ninl m = {q w e}\nf m\n<\/pre>\n<pre>6L\n<\/pre>\n<p>They allow functional lens updates. Note that in the generated code their fields are ordered by their names.<\/p>\n<pre>inl f d = {d.data with a = self + 10}\ninl a = 1\ninl b = 2\ninl c = 3\ninl m = {data={a b c}}\nf m\n<\/pre>\n<pre>type Env0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nand Env1 =\n    struct\n    val mem_0: Env0\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\n(Env1((Env0(11L, 2L, 3L))))\n<\/pre>\n<p>Fields can be added to them and removed arbitrarily in an immutable fashion. Using <code>without<\/code> on a non-existing field will not do anything.<\/p>\n<pre>inl a = 1\ninl b = 2\ninl c = 3\ninl m = {a b c}\n{m with d = 4; without a}\n<\/pre>\n<pre>type Env0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\n(Env0(2L, 3L, 4L))\n<\/pre>\n<p>Like with tuples which are represented by immutable lists in Spiral, the modules in Spiral allow anything immutable maps might do. For example, they can be mapped over(<code>module_map<\/code>), folded(<code>module_foldl<\/code>,<code>module_fold<\/code>), and filtered(<code>module_filter<\/code>). Here is the fold example.<\/p>\n<pre>inl m = {a=1; b=2; c=3}\nmodule_foldl (inl key state value -&gt; state + value) 0 m\n<\/pre>\n<pre>6L\n<\/pre>\n<p>They support more powerful patterns than F# allows on records like not(<code>!<\/code>) and xor(<code>^<\/code>).<\/p>\n<pre>inl f {!nope (a ^ b)=s} = s\n\/\/ f {nope=()} \/\/ Would trigger a type error\ninl m = {a=1; b=2}\n\/\/ f m \/\/ Without trigger a type error\nf {m without a}, f {m without b}\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nTuple0(2L, 1L)\n<\/pre>\n<p>Last, but not least, Spiral's modules and functions support several kinds of layouts. By default, like tuples, they have a transparent structure whose variables are tracked on an individual basis. Here is the heap layout.<\/p>\n<pre>{a=1; b=2; c=3} |&gt; dyn |&gt; heap\n<\/pre>\n<pre>type EnvHeap0 =\n    {\n    mem_0: int64\n    mem_1: int64\n    mem_2: int64\n    }\nlet (var_0: int64) = 1L\nlet (var_1: int64) = 2L\nlet (var_2: int64) = 3L\n({mem_0 = (var_0: int64); mem_1 = (var_1: int64); mem_2 = (var_2: int64)} : EnvHeap0)\n<\/pre>\n<p>Here are the 5 layouts in order: <code>indiv<\/code>,<code>heap<\/code>,<code>heapm<\/code>,<code>stack<\/code>,<code>packed_stack<\/code>.<\/p>\n<pre>{a=1; b=2; c=3} |&gt; dyn |&gt; heap |&gt; heapm |&gt; stack |&gt; packed_stack\n<\/pre>\n<pre>type EnvHeap0 =\n    {\n    mem_0: int64\n    mem_1: int64\n    mem_2: int64\n    }\nand EnvHeapMutable1 =\n    {\n    mutable mem_0: int64\n    mutable mem_1: int64\n    mutable mem_2: int64\n    }\nand EnvStack2 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\n[&lt;System.Runtime.InteropServices.StructLayout(System.Runtime.InteropServices.LayoutKind.Sequential,Pack=1)&gt;]\nand EnvPackedStack3 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nlet (var_0: int64) = 1L\nlet (var_1: int64) = 2L\nlet (var_2: int64) = 3L\nlet (var_3: EnvHeap0) = ({mem_0 = (var_0: int64); mem_1 = (var_1: int64); mem_2 = (var_2: int64)} : EnvHeap0)\nlet (var_4: int64) = var_3.mem_0\nlet (var_5: int64) = var_3.mem_1\nlet (var_6: int64) = var_3.mem_2\nlet (var_7: EnvHeapMutable1) = ({mem_0 = (var_4: int64); mem_1 = (var_5: int64); mem_2 = (var_6: int64)} : EnvHeapMutable1)\nlet (var_8: int64) = var_7.mem_0\nlet (var_9: int64) = var_7.mem_1\nlet (var_10: int64) = var_7.mem_2\nlet (var_11: EnvStack2) = EnvStack2((var_8: int64), (var_9: int64), (var_10: int64))\nlet (var_12: int64) = var_11.mem_0\nlet (var_13: int64) = var_11.mem_1\nlet (var_14: int64) = var_11.mem_2\nEnvPackedStack3((var_12: int64), (var_13: int64), (var_14: int64))\n<\/pre>\n<p>That is the tour of them, but it does not yet demonstrate their true power. The essence of modules converted to layout types is that they capture scope.<\/p>\n<pre>inl npc = \n    {\n    health = dyn 0\n    mana = dyn 0\n    max_health = 40\n    max_mana = 30\n    }\ninl ar = array_create (type stack npc) 3\nar 0 &lt;- stack {npc with health = dyn 10; mana = dyn 20}\nar 1 &lt;- stack {npc with health = dyn 20; mana = dyn 10}\n\/\/ar 2 &lt;- {npc with health = dyn 10; mana = dyn 20} \/\/ Gives a type error as the module here is not a layout type.\n\/\/ar 2 &lt;- stack {npc with health = dyn 10; mana = dyn 20; max_health = 50} \/\/ Gives a type error as max health is an incorrect literal.\n()\n<\/pre>\n<pre>type EnvStack0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: int64) = 0L\nlet (var_1: int64) = 0L\nlet (var_3: (EnvStack0 [])) = Array.zeroCreate&lt;EnvStack0&gt; (System.Convert.ToInt32(3L))\nlet (var_4: int64) = 10L\nlet (var_5: int64) = 20L\nlet (var_6: EnvStack0) = EnvStack0((var_4: int64), (var_5: int64))\nvar_3.[int32 0L] &lt;- var_6\nlet (var_7: int64) = 20L\nlet (var_8: int64) = 10L\nlet (var_9: EnvStack0) = EnvStack0((var_7: int64), (var_8: int64))\nvar_3.[int32 1L] &lt;- var_9\n<\/pre>\n<p>In layout types, literals and naked types become a part of the bigger type and are tracked at the type level. The individual variables are flattened and the intermediate structures are erased in the generated code, very similarly to how the arguments are handled at join points.<\/p>\n<p>The <code>packed_stack<\/code> layout is just there in case it might be necessary to pass a tuple over to the Cuda side. In most cases though, it makes more sense to use the default (non)layout and pass them as individual arguments.<\/p>\n<p><code>heapm<\/code> layout is useful for mutably updating individual fields of a heap allocated module.<\/p>\n<p>Layout types are there in order to allow finer control of the boxed representations of modules and functions. Without <code>heap<\/code> it would be impossible to heap allocate modules directly for example.<\/p>\n<h4><a id=\"user-content-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#macros\"><\/a>Macros<\/h4>\n<h5><a id=\"user-content-solve-me\" class=\"anchor\" aria-hidden=\"true\" href=\"#solve-me\"><\/a>Solve Me<\/h5>\n<p>Modules are beautiful and elegant part of Spiral. Macros are definitely ugly, but they are the only way for Spiral to interop with other languages' libraries and as such are indispensable.<\/p>\n<p>In Spiral they have the interesting property of also acting as types.<\/p>\n<p>So far all the examples given in the tutorial were relatively unmotivated. Macros make it is possible to do IO among other things which allows the language to be applied to real-world problems.<\/p>\n<p>As a very basic demonstration of them, let us start with this HackerRank problem.<\/p>\n<pre>\/\/ https:\/\/www.hackerrank.com\/challenges\/solve-me-first\/problem\n\/\/ The entire code is given, you can just review and submit!\nopen System\n[&lt;EntryPoint&gt;]\nlet main argv = \n    let a = Console.ReadLine() |&gt; int\n    let b = Console.ReadLine() |&gt; int\n    printfn \"%d\" (a+b)\n    0 \/\/ return an integer exit code\n<\/pre>\n<p>The above is the F# solution given directly. It just reads two ints from input, sums them and returns the sum. Doing it in Spiral without the IDE support or even direct language support for .NET constructs make it more complicated.<\/p>\n<p>First the <code>System.Console<\/code> type needs to be defined.<\/p>\n<pre>\/\/inl console = fs ((.text, \"System.Console\") :: ())\ninl console = fs [text: \"System.Console\"]\nprint_static console\n<\/pre>\n<pre>type (dotnet_type (System.Console))\n<\/pre>\n<p>What this has done is create the <code>[text: \"System.Console\"]<\/code> naked type. The type shown in the output is just how it gets printed - the actual type is determined by its body, namely <code>[text: \"System.Console\"]<\/code>. This is equivalent to <code>(.text, \"System.Console\") :: ()<\/code><\/p>\n<pre>inl a = (.text, \"System.Console\") :: () |&gt; fs\ninl b = [text: \"System.Console\"] |&gt; fs\neq_type a b\n<\/pre>\n<pre>true\n<\/pre>\n<p><code>[]<\/code> is just syntax sugar for named tuples. It has no extra functionality apart from what is provided by the standard constructs.<\/p>\n<p>Since types are just macros it is possible to make nonsensical types.<\/p>\n<pre>print_static (fs [text: \"1 + 2 + 3\"])\n<\/pre>\n<pre>type (dotnet_type (1 + 2 + 3))\n<\/pre>\n<p>Hence mistakes with macros will have to be responsibility of the downwards languages. But in the worst they will just lead to a type error.<\/p>\n<p>Unlike in other languages where they are used for abstraction, macros in Spiral are only to be used for interop. They would not be good at all for that anyway given that at most they can print text.<\/p>\n<pre>inl console = fs [text: \"System.Console\"]\ninl static_method static_type method_name args return_type = \n    macro.fs return_type [\n        type: static_type\n        text: \".\"\n        text: method_name\n        args: args\n        ]\ninl readline() = static_method console .ReadLine() string\ninl a, b = readline(), readline()\n()\n<\/pre>\n<pre>let (var_0: string) = System.Console.ReadLine()\nlet (var_1: string) = System.Console.ReadLine()\n<\/pre>\n<p>The above program succinctly captures Spiral's approach to language interop. The facilities used for defining macro-based types and printing them are interwoven with one another. One extra ingredient macros evaluation require over macro type definitions is the return type.<\/p>\n<p>What the <code>macro.fs<\/code> function is doing is printing the macro based on the second argument and returning a variable of the type in the first argument.<\/p>\n<p>Now all the pieces are in place to finish the exercise.<\/p>\n<pre>inl console = fs [text: \"System.Console\"]\ninl static_method static_type method_name args return_type = \n    macro.fs return_type [\n        type: static_type\n        text: \".\"\n        text: method_name\n        args: args\n        ]\ninl unop name arg return_type = \n    macro.fs return_type [\n        text: name\n        text: \" \"\n        arg: arg\n        ]\ninl readline() = static_method console .ReadLine() string\ninl writeline x = static_method console .WriteLine x string\ninl int x = unop \"int\" x int32\ninl a, b = readline(), readline()\nwriteline (int a + int b)\n<\/pre>\n<pre>let (var_0: string) = System.Console.ReadLine()\nlet (var_1: string) = System.Console.ReadLine()\nlet (var_2: int32) = int var_0\nlet (var_3: int32) = int var_1\nlet (var_4: int32) = (var_2 + var_3)\nSystem.Console.WriteLine(var_4)\n<\/pre>\n<h5><a id=\"user-content-simple-array-sum-macro-version\" class=\"anchor\" aria-hidden=\"true\" href=\"#simple-array-sum-macro-version\"><\/a>Simple Array Sum (macro version)<\/h5>\n<p>This example is to demonstrate how macros can be used to interop with F# libraries which often take in functions as arguments.<\/p>\n<p>The code fragments will be split into two. The first part loads the numbers into a Spiral array, splits them based on the whitespace char and convert them to ints.<\/p>\n<pre>\/\/https:\/\/www.hackerrank.com\/challenges\/simple-array-sum\/problem\ninl console = fs [text: \"System.Console\"]\ninl static_method static_type method_name args return_type = \n    macro.fs return_type [\n        type: static_type\n        text: \".\"\n        text: method_name\n        args: args\n        ]\ninl readline() = static_method console .ReadLine() string\ninl writeline x = static_method console .WriteLine x string\ninl array t = type (array_create t 0)\ninl _, ar = readline(), macro.fs (array int32) [arg: readline(); text: \".Split [|' '|] |&gt; Array.map int\"]\n<\/pre>\n<pre>let (var_0: string) = System.Console.ReadLine()\nlet (var_2: string) = System.Console.ReadLine()\nlet (var_3: (int32 [])) = var_2.Split [|' '|] |&gt; Array.map int\n<\/pre>\n<p>The next part could also be done using macros, but is here to demonstrate an aspect of Spiral's intensional polymorphism.<\/p>\n<pre>\/\/ Converts a type level function to a term level function based on a type.\ninl rec closure_of f tys = \n    match tys with\n    | x =&gt; xs -&gt; term_cast (inl x -&gt; closure_of (f x) xs) x\n    | x: f -&gt; f\n    | _ -&gt; error_type \"The tail of the closure does not correspond to the one being casted to.\"\ninl add a b = a + b\ninl add_closure = closure_of add (int32 =&gt; int32 =&gt; int32)\nmacro.fs int32 [text: \"Array.fold \"; arg: add_closure; text: \" 0 \"; arg: ar]\n|&gt; writeline\n<\/pre>\n<pre>let (var_5: (int32 -&gt; (int32 -&gt; int32))) = method_0\nlet (var_6: int32) = Array.fold var_5 0 var_3\nSystem.Console.WriteLine(var_6)\n<\/pre>\n<p><code>closure_of<\/code> is an expanded version of <code>term_cast<\/code> that instead of converting by applying using the input argument converts a (staged) type level function to a term level using a target type thereby unstaging it. Term level functions have their own dedicated pattern for destructuring their types.<\/p>\n<p>Naked types for them can be constructed with the <code>=&gt;<\/code> operator. The <code>error_type<\/code> raises a type error with the specified message whenever it is evaluated.<\/p>\n<p>What the <code>closure_of<\/code> function does can be better understood by rewriting it to a specific instance with two arguments.<\/p>\n<pre>inl closure_of_2 f (a' =&gt; b' =&gt; c') = \n    term_cast (inl a -&gt; term_cast (inl b -&gt; f a b : c') b') a'\nclosure_of_2 (+) (int32 =&gt; int32 =&gt; int32)\n<\/pre>\n<pre>let rec method_0 ((var_0: int32)): (int32 -&gt; int32) =\n    method_1((var_0: int32))\nand method_1 ((var_1: int32)) ((var_0: int32)): int32 =\n    (var_1 + var_0)\nmethod_0\n<\/pre>\n<p>The original version is just a more generic version of <code>closure_of_2<\/code> that loops over the arguments while both accumulating the results of the application of the closure and term casting it.<\/p>\n<p>That is roughly it with regards to interop. Spiral of course does have its own libraries.<\/p>\n<p><code>closure_of<\/code> and other macro related functions can be found in the <code>Extern<\/code> module.<\/p>\n<h4><a id=\"user-content-spiral-libraries\" class=\"anchor\" aria-hidden=\"true\" href=\"#spiral-libraries\"><\/a>Spiral libraries<\/h4>\n<pre>let example1 = \n    \"example1\",[array;console],\"Module description.\",\n    \"\"\"\nopen Console\ninl _, b = readline(), macro.fs (array int32) [arg: readline(); text: \".Split [|' '|] |&gt; Array.map int\"]\nArray.foldl (+) (dyn 0i32) b |&gt; writeline\n    \"\"\"\n<\/pre>\n<p>The way Spiral is currently meant to be used is as a scripting language inside F#. The module argument is the list in the middle and the <code>array<\/code> and <code>console<\/code> are the modules of the same name respectively.<\/p>\n<pre>inl Array = ...\ninl Console = ...\nopen Console\ninl _, b = readline(), macro.fs (array int32) [arg: readline(); text: \".Split [|' '|] |&gt; Array.map int\"]\nArray.foldl (+) (dyn 0i32) b |&gt; writeline\n<\/pre>\n<p>The above is roughly how the program would be unfolded after parsing, but before typing and partial evaluation. Modules are unfolded in a flattened manner in the sequence they are input. Duplicate modules are ignored.<\/p>\n<p>Much like F#, Spiral imposes a top down ordering of the program and modules cannot refer to each other recursively. If that functionality is required, it can be achieved using join points, but in general it should not be necessary.<\/p>\n<p>This kind of constrained architecture cuts down on circular referencing and encourages purposeful laying out of programs.<\/p>\n<p>Spiral libraries are (to be) covered in depth in the user guide and the reference.<\/p>\n<h3><a id=\"user-content-3-loops-and-arrays\" class=\"anchor\" aria-hidden=\"true\" href=\"#3-loops-and-arrays\"><\/a>3: Loops and Arrays<\/h3>\n<h4><a id=\"user-content-loops\" class=\"anchor\" aria-hidden=\"true\" href=\"#loops\"><\/a>Loops<\/h4>\n<p>Most languages make it trivial to write loops and the user does not have to worry about them diverging except at runtime.<\/p>\n<p>Spiral's staging abilities introduce new complexities into the mix. In Spiral, for every loop one writes, it is necessary to keep in mind whether it is intended to run at compile or at runtime.<\/p>\n<p>Making functions stage polymorphic takes even more effort. Furthermore for recursive runtime functions it is easy to forget to put in the type annotation and to dynamize the counter.<\/p>\n<p>For that reason, the bog standard <code>for<\/code> and <code>while<\/code> loops exist as a part of the <code>Loops<\/code> module in Spiral.<\/p>\n<p>This chapter will be on building up the basic loop and then using it to implement the array library functions from first principles.<\/p>\n<p>At this point, apart from union types and the Cuda backend, all the main language features have been introduced albeit shallowly.<\/p>\n<p>This makes it possible to demonstrate how the architecture of a Spiral program differs from those in other languages.<\/p>\n<pre>let example = \n    \"example\",[console],\"Module description.\",\n    \"\"\"\nopen Console\nmet rec for {d with from=(!dyn from) to by body} =\n    if from &lt;= to then body from; for {d with from=from+by}\n    else ()\n    : ()\nfor {from=0; to=5; by=1; body=inl i -&gt;\n    string_format \"The loop is on iteration {0}\" i |&gt; writeline\n    }\n    \"\"\"\n<\/pre>\n<pre>let rec method_0((var_0: int64)): unit =\n    let (var_1: bool) = (var_0 &lt;= 5L)\n    if var_1 then\n        let (var_2: string) = System.String.Format(\"The loop is on iteration {0}\",var_0)\n        let (var_3: string) = System.String.Format(\"{0}\",var_2)\n        System.Console.WriteLine(var_3)\n        let (var_4: int64) = (var_0 + 1L)\n        method_1((var_4: int64))\n    else\n        ()\nand method_1((var_0: int64)): unit =\n    let (var_1: bool) = (var_0 &lt;= 5L)\n    if var_1 then\n        let (var_2: string) = System.String.Format(\"The loop is on iteration {0}\",var_0)\n        let (var_3: string) = System.String.Format(\"{0}\",var_2)\n        System.Console.WriteLine(var_3)\n        let (var_4: int64) = (var_0 + 1L)\n        method_1((var_4: int64))\n    else\n        ()\nlet (var_0: int64) = 0L\nmethod_0((var_0: int64))\n<\/pre>\n<p>Somewhat inadvertently, the first example become a good lesson in why loops would be desirable as a part of the library. The first example was careful to <code>dyn<\/code> the counter and did not forget the annotation, but for some reason the loop got specialized to two functions one which only got called once.<\/p>\n<p>It is not a compiler bug.<\/p>\n<pre>met rec for {d with from=(!dyn from) to by body} =\n    if from &lt;= to then body from; for {d with from=from+by}\n<\/pre>\n<p>The way join points work is that they specialize the call by their arguments. By rewriting the fragment highlighted above to an equivalent form it will be easy to demonstrate what is happening.<\/p>\n<pre>inl rec for d =\n    inl from = dyn d.from\n    inl {to by body} = d\n    join\n        if from &lt;= to then body from; for {d with from=from+by}\n        else ()\n        : ()\n<\/pre>\n<p>What is going on is that <code>d<\/code> - the old one with the <code>from<\/code> field still as literal is getting passed through the join point and causes the redundant specialization to happen.<\/p>\n<p>Here is the way to write the <code>for<\/code> function correctly.<\/p>\n<p>Out of all the mistakes to make in Spiral, accidentally passing old state through the join point is the easiest one to make. With missed return type annotations and such the compiler will diverge and warn the user that way, but but this one has a way of preying on laziness.<\/p>\n<p>In fact, this kind of error can happen in any language that supports records with mutable updates, not just Spiral. Spiral in particular just makes it obvious by looking at the argument count in the generated code.<\/p>\n<pre>met rec for {from=(!dyn from) to by body} =\n    if from &lt;= to then body from; for {from=from+by; to by body}\n    else ()\n    : ()\n<\/pre>\n<pre>let rec method_0((var_0: int64)): unit =\n    let (var_1: bool) = (var_0 &lt;= 5L)\n    if var_1 then\n        let (var_2: string) = System.String.Format(\"The loop is on iteration {0}\",var_0)\n        let (var_3: string) = System.String.Format(\"{0}\",var_2)\n        System.Console.WriteLine(var_3)\n        let (var_4: int64) = (var_0 + 1L)\n        method_0((var_4: int64))\n    else\n        ()\nlet (var_0: int64) = 0L\nmethod_0((var_0: int64))\n<\/pre>\n<p>The above output is the ideal for this kind of loop. Only the <code>var_0<\/code> varies, the other literals all get passed through the boundary and specialized along with the body.<\/p>\n<p>This is kind of specialization important to do with Cuda kernels as using too many variables in place of literals can cause register spillage into global memory and cause drastic degradations of performance. Spiral makes it easy to keep such data static and propagate it through the program.<\/p>\n<p>In addition, Spiral makes it trivial to this kind of specialization even across language boundaries. Partial evaluation is commonly refereed to as specialization. Staging makes it user directed. And being able to use staging constructs through the the type system as the basis of abstraction rather than being restricted to a second class macro inspired system is what makes Spiral's staging first class. That is a desirable trait as it increases uniformity of the language and with it, its power. It also simplifies its implementation greatly, so it is a good design principle to follow at all times.<\/p>\n<p>More concretely, one of the main motivations for writing Spiral for its author is avoiding having to write unending litanies of wrappers for simple Cuda kernels.<\/p>\n<p>Moving on, here is the static version of the loop.<\/p>\n<pre>inl rec for {from to by body} =\n    if from &lt;= to then body from; for {from=from+by; to by body}\n    else ()\nfor {from=0; to=5; by=1; body=inl i -&gt;\n    string_format \"The loop is on iteration {0}\" i |&gt; writeline\n    }\n<\/pre>\n<pre>System.Console.WriteLine(\"The loop is on iteration 0\")\nSystem.Console.WriteLine(\"The loop is on iteration 1\")\nSystem.Console.WriteLine(\"The loop is on iteration 2\")\nSystem.Console.WriteLine(\"The loop is on iteration 3\")\nSystem.Console.WriteLine(\"The loop is on iteration 4\")\nSystem.Console.WriteLine(\"The loop is on iteration 5\")\n<\/pre>\n<p>Now what remains is to make the function stage polymorphic.<\/p>\n<pre>let example = \n    \"example\",[tuple;console],\"Module description.\",\n    \"\"\"\nopen Console\ninl rec for {from to by body} =\n    inl body from = \n        if from &lt;= to then body from; for {from=from+by; to by body}\n        else ()\n    if Tuple.forall lit_is (from,to,by) then body from\n    else \n        inl from = dyn from\n        join body from : ()\nfor {from=0; to=5; by=1; body=inl i -&gt;\n    string_format \"The loop is on iteration {0}\" i |&gt; writeline\n    }\nfor {from=dyn 0; to=5; by=1; body=inl i -&gt;\n    string_format \"The loop is on iteration {0}\" i |&gt; writeline\n    }\n    \"\"\"\n<\/pre>\n<pre>let rec method_0((var_0: int64)): unit =\n    let (var_1: bool) = (var_0 &lt;= 5L)\n    if var_1 then\n        let (var_2: string) = System.String.Format(\"The loop is on iteration {0}\",var_0)\n        let (var_3: string) = System.String.Format(\"{0}\",var_2)\n        System.Console.WriteLine(var_3)\n        let (var_4: int64) = (var_0 + 1L)\n        method_0((var_4: int64))\n    else\n        ()\nSystem.Console.WriteLine(\"The loop is on iteration 0\")\nSystem.Console.WriteLine(\"The loop is on iteration 1\")\nSystem.Console.WriteLine(\"The loop is on iteration 2\")\nSystem.Console.WriteLine(\"The loop is on iteration 3\")\nSystem.Console.WriteLine(\"The loop is on iteration 4\")\nSystem.Console.WriteLine(\"The loop is on iteration 5\")\nlet (var_0: int64) = 0L\nmethod_0((var_0: int64))\n<\/pre>\n<p>The above loop can further improved in terms of functionality. Notice that its body has a type <code>unit<\/code> which is represented by an empty tuple in Spiral. That is a throwback to C that has no place in modern language such as Spiral.<\/p>\n<pre>open Console\ninl rec for {from to by state body} =\n    inl body from = \n        if from &lt;= to then for {to by body from=from+by; state=body {state i=from}}\n        else state\n    if Tuple.forall lit_is (from,to,by) then body from\n    else \n        inl from = dyn from\n        join body from : state\ninl power a to = for {from=2; to by=1; state=a; body=inl {state} -&gt; state * a}\npower 2 3\n<\/pre>\n<pre>8L\n<\/pre>\n<p>The above works, but various criticisms of the program could be made. For one, is it really necessary to give <code>by<\/code> every time? Vast majority of loops will in fact have it as <code>1<\/code> so if it is not given it makes sense to use that default instead of giving a type error.<\/p>\n<p>Speaking of defaults, a decent guess would be that most loops are not intended to be unrolled and that a user is more likely to just forget to <code>dyn<\/code> the <code>from<\/code> field by accident.<\/p>\n<pre>open Console\ninl rec for {d with to state body} =\n    inl body {from by} = \n        if from &lt;= to then for {to by body from=from+by; state=body {state i=from}}\n        else state\n    inl from =\n        match d with\n        | {from} -&gt; dyn from\n        | {static_from} -&gt; static_from\n    inl by =\n        match d with\n        | {by} -&gt; by\n        | _ -&gt; 1\n    if Tuple.forall lit_is (from,to,by) then body {from}\n    else \n        inl from = dyn from\n        join body {from by} : state\ninl power a to = for {from=2; to state=a; body=inl {state} -&gt; state * a}\npower 2 3\n<\/pre>\n<pre>let rec method_0((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &lt;= 3L)\n    if var_1 then\n        let (var_2: int64) = (var_0 + 1L)\n        method_1((var_2: int64))\n    else\n        2L\nand method_1((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &lt;= 3L)\n    if var_1 then\n        let (var_2: int64) = (var_0 + 1L)\n        method_2((var_2: int64))\n    else\n        4L\nand method_2((var_0: int64)): int64 =\n    let (var_1: bool) = (var_0 &lt;= 3L)\n    if var_1 then\n        let (var_2: int64) = (var_0 + 1L)\n        method_3((var_2: int64))\n    else\n        8L\n\/\/ ...and so on up to method_63\n<\/pre>\n<p>Not quite as planned. An error made now is that the state gets specialized for every different power of 2.<\/p>\n<p>With a single added <code>dyn<\/code> that can be fixed.<\/p>\n<pre>inl power a to = for {from=2; to state=dyn a; body=inl {state} -&gt; state * a}\n<\/pre>\n<pre>let rec method_0((var_0: int64), (var_1: int64)): int64 =\n    let (var_2: bool) = (var_1 &lt;= 3L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_4: int64) = (var_0 * 2L)\n        method_0((var_4: int64), (var_3: int64))\n    else\n        var_0\nlet (var_0: int64) = 2L\nlet (var_1: int64) = 2L\nmethod_0((var_0: int64), (var_1: int64))\n<\/pre>\n<p>As a matter of convention, Spiral library functions that take in <code>state<\/code> never <code>dyn<\/code> it directly. That responsibility should fall onto the user.<\/p>\n<p>As an example of the reason for that, the state might be an option type so it might be better to specialize it for both of its states without instantiating it directly. Or it might be a tuple with some fields which would be desirable to remain as literals.<\/p>\n<p><code>for<\/code> is intended to be used as a primitive and so requires some flexibility; it would not do to block functionality.<\/p>\n<p>Looking over the function now it seems fine, but it is a bit uncomfortable how from has to start from <code>2<\/code>. It is not like the loop has to use the <code>&lt;=<\/code> operator for comparison in the conditional. In a lot of cases <code>&lt;<\/code> make a lot more sense.<\/p>\n<p>Furthermore, an user might want to iterate downwards. That can be accommodated.<\/p>\n<pre>open Console\ninl rec for {d with state body} =\n    inl check =\n        match d with\n        | {near_to} from -&gt; from &lt; near_to \n        | {to} from -&gt; from &lt;= to\n        | {down_to} from -&gt; from &gt;= down_to\n        | {near_down_to} from -&gt; from &gt; near_down_to\n    inl from =\n        match d with\n        | {from} -&gt; dyn from\n        | {static_from} -&gt; static_from\n    inl {(to ^ near_to ^ down_to ^ near_down_to)=to} = d\n    inl by =\n        match d with\n        | {by} -&gt; by\n        | _ -&gt; 1\n    inl rec loop {from state} =\n        inl body {from} = \n            if check from then loop {from=from+by; state=body {state i=from}}\n            else state\n        if Tuple.forall lit_is (from,to,by) then body {from}\n        else \n            inl from = dyn from\n            join body {from} : state\n    loop {from state}\ninl power a near_to = for {from=1; near_to state=dyn a; body=inl {state} -&gt; state * a}\npower 2 3\n<\/pre>\n<pre>let rec method_0((var_0: int64), (var_1: int64)): int64 =\n    let (var_2: bool) = (var_1 &lt; 3L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_4: int64) = (var_0 * 2L)\n        method_0((var_4: int64), (var_3: int64))\n    else\n        var_0\nlet (var_0: int64) = 2L\nlet (var_1: int64) = 1L\nmethod_0((var_0: int64), (var_1: int64))\n<\/pre>\n<p>The module member queries are all done statically and so maximum polymorphism is attained. The above program also demonstrates why lexical scope is so great.<\/p>\n<p>The above is starting to near the functionality of the <code>for<\/code> function in the actual library. To make it more professional, rather than returning a pattern miss error on when a field is missed, it would be better to tell the user what the problem is.<\/p>\n<pre>open Console\ninl for {d with state body} =\n    inl check =\n        match d with\n        | {near_to} from -&gt; from &lt; near_to \n        | {to} from -&gt; from &lt;= to\n        | {down_to} from -&gt; from &gt;= down_to\n        | {near_down_to} from -&gt; from &gt; near_down_to\n        | _ -&gt; error_type \"One of `to`,`near_to`,`down_to`,`near_down_to` needs be present.\"\n    inl from =\n        match d with\n        | {from=(!dyn from) ^ static_from=from} -&gt; from\n        | _ -&gt; error_type \"Only one of `from` and `static_from` field to loop needs to be present.\"\n    inl to =\n        match d with\n        | {(to ^ near_to ^ down_to ^ near_down_to)=to} -&gt; to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` is allowed.\"\n    inl by =\n        match d with\n        | {by} -&gt; by\n        | _ -&gt; 1\n    inl rec loop {from state} =\n        inl body {from} = \n            if check from then loop {from=from+by; state=body {state i=from}}\n            else state\n        if Tuple.forall lit_is (from,to,by) then body {from}\n        else \n            inl from = dyn from\n            join body {from} : state\n    loop {from state}\ninl power a near_to = for {static_from=1; near_to state=a; body=inl {state} -&gt; state * a}\npower 2 3\n<\/pre>\n<pre>8L\n<\/pre>\n<p>The above design is in fact superior to what is currently in the standard library. A lot of features of the language were developed along with the library and some parts of it did not keep up. The author also got to fancy with the design of it. At the time that was useful for pushing the language, but not so much from a design perspective.<\/p>\n<p>Another issue with the standard library as it stands is that in fact its author did not know how to program in the language he was making and had to learn it as he going along.<\/p>\n<pre>    inl from =\n        match d with\n        | {from=(!dyn from) ^ static_from=from} -&gt; from\n        | _ -&gt; error_type \"Only one of `from` and `static_from` field to loop needs to be present.\"\n<\/pre>\n<p>This part here is highlighted in order to show the xor (<code>^<\/code>) pattern might be used in tandem with active patterns.<\/p>\n<p>All the features in the making of the loop so far have been covered in the previous chapters and now it can be seen how they come together.<\/p>\n<p>It is not done yet.<\/p>\n<p>In order to attain the full functionality of C style loops, Spiral's loops also need the ability to break out. Strictly speaking, this cannot be done in a functional language and having <code>return<\/code> would make even less sense in Spiral than it does in ML variants, but the same functionality can be achieved instead by writing the loop body and calling the continuation for the next iteration in tail position.<\/p>\n<p>As motivating example, imagine trying to iterate over nested arrays trying to find a specific item before breaking out. With the loop as was written above, there is no way to stop before reaching the end.<\/p>\n<p>To start things off, first the nested arrays need to be created.<\/p>\n<pre>open Console\ninl for {d with body} =\n    inl state = \n        match d with\n        | {state} -&gt; state\n        | _ -&gt; ()\n\/\/\/ ...\ninl array_init near_to f =\n    assert (near_to &gt;= 0) \"The input to init needs to be greater or equal to 0.\"\n    \/\/ Somewhat of an ugly practice in order to infer the type in a language that doesn't support inference. \n    \/\/ For large functions, it is recommended to put them in a join point otherwise compile times could \n    \/\/ become exponential if the function contains branches.\n    \/\/ For a simple map for an array like here, it does not matter.\n    inl typ = type (f 0) \n    inl ar = array_create typ near_to\n    for {from=0; near_to; body=inl {i} -&gt; ar i &lt;- f i}\n    ar\ninl rec zeroes = function\n    | x :: x' -&gt; array_init x (inl _ -&gt; zeroes x')\n    | () -&gt; \"\"\ninl ar = zeroes (4,4,4,4)\nar 0 0 0 2 &lt;- \"princess\"\n<\/pre>\n<pre>let rec method_3((var_0: ((((string []) []) []) [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_8: (((string []) []) [])) = Array.zeroCreate&lt;((string []) [])&gt; (System.Convert.ToInt32(4L))\n        let (var_9: int64) = 0L\n        method_2((var_8: (((string []) []) [])), (var_9: int64))\n        var_0.[int32 var_1] &lt;- var_8\n        method_3((var_0: ((((string []) []) []) [])), (var_3: int64))\n    else\n        ()\nand method_2((var_0: (((string []) []) [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_6: ((string []) [])) = Array.zeroCreate&lt;(string [])&gt; (System.Convert.ToInt32(4L))\n        let (var_7: int64) = 0L\n        method_1((var_6: ((string []) [])), (var_7: int64))\n        var_0.[int32 var_1] &lt;- var_6\n        method_2((var_0: (((string []) []) [])), (var_3: int64))\n    else\n        ()\nand method_1((var_0: ((string []) [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_4: (string [])) = Array.zeroCreate&lt;string&gt; (System.Convert.ToInt32(4L))\n        let (var_5: int64) = 0L\n        method_0((var_4: (string [])), (var_5: int64))\n        var_0.[int32 var_1] &lt;- var_4\n        method_1((var_0: ((string []) [])), (var_3: int64))\n    else\n        ()\nand method_0((var_0: (string [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        var_0.[int32 var_1] &lt;- \"\"\n        method_0((var_0: (string [])), (var_3: int64))\n    else\n        ()\nlet (var_6: ((((string []) []) []) [])) = Array.zeroCreate&lt;(((string []) []) [])&gt; (System.Convert.ToInt32(4L))\nlet (var_7: int64) = 0L\nmethod_3((var_6: ((((string []) []) []) [])), (var_7: int64))\nlet (var_8: (((string []) []) [])) = var_6.[int32 0L]\nlet (var_9: ((string []) [])) = var_8.[int32 0L]\nlet (var_10: (string [])) = var_9.[int32 0L]\nvar_10.[int32 2L] &lt;- \"princess\"\n<\/pre>\n<p>First off, the case for <code>state<\/code> that was forgotten is added at the top of the <code>for<\/code> function. As the big comment states inferring the type returned by <code>f<\/code> involves evaluating it twice which might not be a good idea depending on what it is.<\/p>\n<p>The <code>Lazy<\/code> module's only function <code>lazy<\/code> for example puts a join point before evaluating the function because it might otherwise repeat long evaluations and those long evaluation in combination with branching (such as when nesting lazy values) might make compilation time take exponential time.<\/p>\n<p>The <code>assert<\/code> function if its conditional can be statically determined and is true gives a type error at compile time instead of triggering at runtime. Spiral has support for throwing exceptions, but not catching or cleaning up after them, so they are intended to be used only for unrecoverable errors.<\/p>\n<p>As the example shows, nesting loops is straightforward in Spiral. It is a decent bit more elegant than doing it with macros which are the only choice in languages with weaker type systems. In Spiral, type inference and partial evaluation are one.<\/p>\n<p>Its type system is extremely powerful, and yet it does not have parametric polymorphism. Adding parametric polymorphism would significantly increase the complexity of both the language and its implementation, would not make the language any more expressive and would make it a lot harder to integrate the partial evaluator with the type system. This would make the language quite a bit slower.<\/p>\n<p>It is interesting to consider the implication of this - in Lisp languages, its raw AST flavored syntax is there for the reason of supporting its macro meta-programming feature. Maybe a really powerful type system does require the absence of parametricity?<\/p>\n<p>In light of what Spiral can do, it might be worth considering whether the programming language community at large collectively missed a whole evolutionary branch of languages with static typing, but without parametric polymorphism.<\/p>\n<p>If that is not convincing yet, maybe it will be after the tutorials are through.<\/p>\n<p>Here is how to write a breakable version of the <code>for<\/code> function in to take advantage of the continuation passing style.<\/p>\n<pre>inl for' {d with body} =\n    inl finally =\n        match d with\n        | {finally} -&gt; finally\n        | _ -&gt; id\n    inl state = \n        match d with\n        | {state} -&gt; state\n        | _ -&gt; ()\n    inl check =\n        match d with\n        | {near_to} from -&gt; from &lt; near_to \n        | {to} from -&gt; from &lt;= to\n        | {down_to} from -&gt; from &gt;= down_to\n        | {near_down_to} from -&gt; from &gt; near_down_to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` needs be present.\"\n    inl from =\n        match d with\n        | {from=(!dyn from) ^ static_from=from} -&gt; from\n        | _ -&gt; error_type \"Only one of `from` and `static_from` field to loop needs to be present.\"\n    inl to =\n        match d with\n        | {(to ^ near_to ^ down_to ^ near_down_to)=to} -&gt; to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` is allowed.\"\n    inl by =\n        match d with\n        | {by} -&gt; by\n        | _ -&gt; 1\n    inl rec loop {from state} =\n        inl body {from} = \n            if check from then \n                inl next state = loop {state from=from+by}\n                body {next state i=from}\n            else finally state\n        if Tuple.forall lit_is (from,to,by) then body {from}\n        else \n            inl from = dyn from\n            join body {from} : finally state\n    loop {from state}\n<\/pre>\n<p>There is a significant amount of duplication now that will need to be eliminated. The highlights are the addition of the <code>finally<\/code> field and the parts inside <code>loop<\/code>.<\/p>\n<pre>            if check from then \n                inl next state = loop {state from=from+by}\n                body {next state i=from}\n            else finally state\n<\/pre>\n<p>Instead of the loop calling itself, it instead passes a function to the body and lets it do it instead.<\/p>\n<p>The <code>finally<\/code> field is useful for resuming the outer loop. It can also be used to set the state to <code>unit<\/code>, which would allow the loop to change states without having to resort to union types.<\/p>\n<p>First a utility function for reversing a tuple is needed. This is standard fare for functional programmers and closely mirrors how one would reverse a list in ML styled languages.<\/p>\n<pre>\/\/ Reverses a tuple\ninl tuple_rev = \n    inl rec loop state = function\n        | x :: xs -&gt; loop (x :: state) xs\n        | () -&gt; state\n    loop ()\n<\/pre>\n<p>Here is how to apply the breakable for loop function. The goal is to find the coordinates of \"princess\". The method is generalized to an arbitrary of number of nested arrays.<\/p>\n<pre>\/\/ Correct version\ninl rec find_index {next state} = function\n    | ar &amp; @array_is _ -&gt; \n        inl body {next i} = find_index {next state=i::state} (ar i)\n        for' {from=0; near_to=array_length ar; finally=next; body}\n    | \"princess\" -&gt; tuple_rev state\n    | _ -&gt; next ()\nfind_index {state=(); next = inl _ -&gt; failwith (type (dim)) \"The princess is in another castle.\"} ar\n<\/pre>\n<p><code>failwith<\/code> unlike in F#, requires the return type in Spiral but otherwise functions the same.<\/p>\n<p>The <code>@<\/code> operator on the pattern side is a partial active pattern. Unlike F#'s which expect an option type, what <code>@<\/code> takes in is a function with three arguments in the <code>inl arg on_fail on_succ -&gt; ...<\/code> form. <code>on_fail<\/code> and <code>on_succ<\/code> are to be called in tail position and possibly with join points around them when done so multiple times. They represent pattern failure and pattern success respectively.<\/p>\n<p>Here is a small example.<\/p>\n<pre>inl f pat = function\n    | @pat x -&gt; x\n    | _ -&gt; error_type \"The pattern failed to trigger.\"\ninl pat x on_fail on_succ =\n    match x with\n    | x: string | x: int64 -&gt; join on_succ (string_format \"{0} joined\" x)\n    | _ -&gt; on_fail()\nf pat \"qwe\", f pat 123\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: string\n    val mem_1: string\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_0(): string =\n    \"qwe joined\"\nand method_1(): string =\n    \"123 joined\"\nlet (var_0: string) = method_0()\nlet (var_1: string) = method_1()\nTuple0(var_0, var_1)\n<\/pre>\n<p>Anything in Spiral can be passed as an argument, and since that includes functions it also applies to partial active patterns.<\/p>\n<p>The output of the compiled program is rather large, but it will be reproduced in bulk as an example this time to show that all the loops are being unfolded correctly into tail recursive functions.<\/p>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    val mem_3: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2, arg_mem_3) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2; mem_3 = arg_mem_3}\n    end\nlet rec method_3((var_0: ((((string []) []) []) [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_8: (((string []) []) [])) = Array.zeroCreate&lt;((string []) [])&gt; (System.Convert.ToInt32(4L))\n        let (var_9: int64) = 0L\n        method_2((var_8: (((string []) []) [])), (var_9: int64))\n        var_0.[int32 var_1] &lt;- var_8\n        method_3((var_0: ((((string []) []) []) [])), (var_3: int64))\n    else\n        ()\nand method_4((var_0: ((((string []) []) []) [])), (var_1: int64), (var_2: int64)): Tuple0 =\n    let (var_3: bool) = (var_2 &lt; var_1)\n    if var_3 then\n        let (var_4: (((string []) []) [])) = var_0.[int32 var_2]\n        let (var_5: int64) = var_4.LongLength\n        let (var_6: int64) = 0L\n        method_5((var_4: (((string []) []) [])), (var_2: int64), (var_5: int64), (var_0: ((((string []) []) []) [])), (var_1: int64), (var_6: int64))\n    else\n        (failwith \"The princess is in another castle.\")\nand method_2((var_0: (((string []) []) [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_6: ((string []) [])) = Array.zeroCreate&lt;(string [])&gt; (System.Convert.ToInt32(4L))\n        let (var_7: int64) = 0L\n        method_1((var_6: ((string []) [])), (var_7: int64))\n        var_0.[int32 var_1] &lt;- var_6\n        method_2((var_0: (((string []) []) [])), (var_3: int64))\n    else\n        ()\nand method_5((var_0: (((string []) []) [])), (var_1: int64), (var_2: int64), (var_3: ((((string []) []) []) [])), (var_4: int64), (var_5: int64)): Tuple0 =\n    let (var_6: bool) = (var_5 &lt; var_2)\n    if var_6 then\n        let (var_7: ((string []) [])) = var_0.[int32 var_5]\n        let (var_8: int64) = var_7.LongLength\n        let (var_9: int64) = 0L\n        method_6((var_7: ((string []) [])), (var_5: int64), (var_1: int64), (var_8: int64), (var_0: (((string []) []) [])), (var_2: int64), (var_3: ((((string []) []) []) [])), (var_4: int64), (var_9: int64))\n    else\n        let (var_11: int64) = (var_1 + 1L)\n        method_4((var_3: ((((string []) []) []) [])), (var_4: int64), (var_11: int64))\nand method_1((var_0: ((string []) [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        let (var_4: (string [])) = Array.zeroCreate&lt;string&gt; (System.Convert.ToInt32(4L))\n        let (var_5: int64) = 0L\n        method_0((var_4: (string [])), (var_5: int64))\n        var_0.[int32 var_1] &lt;- var_4\n        method_1((var_0: ((string []) [])), (var_3: int64))\n    else\n        ()\nand method_6((var_0: ((string []) [])), (var_1: int64), (var_2: int64), (var_3: int64), (var_4: (((string []) []) [])), (var_5: int64), (var_6: ((((string []) []) []) [])), (var_7: int64), (var_8: int64)): Tuple0 =\n    let (var_9: bool) = (var_8 &lt; var_3)\n    if var_9 then\n        let (var_10: (string [])) = var_0.[int32 var_8]\n        let (var_11: int64) = var_10.LongLength\n        let (var_12: int64) = 0L\n        method_7((var_10: (string [])), (var_8: int64), (var_1: int64), (var_2: int64), (var_11: int64), (var_0: ((string []) [])), (var_3: int64), (var_4: (((string []) []) [])), (var_5: int64), (var_6: ((((string []) []) []) [])), (var_7: int64), (var_12: int64))\n    else\n        let (var_14: int64) = (var_1 + 1L)\n        method_5((var_4: (((string []) []) [])), (var_2: int64), (var_5: int64), (var_6: ((((string []) []) []) [])), (var_7: int64), (var_14: int64))\nand method_0((var_0: (string [])), (var_1: int64)): unit =\n    let (var_2: bool) = (var_1 &lt; 4L)\n    if var_2 then\n        let (var_3: int64) = (var_1 + 1L)\n        var_0.[int32 var_1] &lt;- \"\"\n        method_0((var_0: (string [])), (var_3: int64))\n    else\n        ()\nand method_7((var_0: (string [])), (var_1: int64), (var_2: int64), (var_3: int64), (var_4: int64), (var_5: ((string []) [])), (var_6: int64), (var_7: (((string []) []) [])), (var_8: int64), (var_9: ((((string []) []) []) [])), (var_10: int64), (var_11: int64)): Tuple0 =\n    let (var_12: bool) = (var_11 &lt; var_4)\n    if var_12 then\n        let (var_13: string) = var_0.[int32 var_11]\n        let (var_14: bool) = (var_13 = \"princess\")\n        if var_14 then\n            Tuple0(var_3, var_2, var_1, var_11)\n        else\n            let (var_15: int64) = (var_11 + 1L)\n            method_7((var_0: (string [])), (var_1: int64), (var_2: int64), (var_3: int64), (var_4: int64), (var_5: ((string []) [])), (var_6: int64), (var_7: (((string []) []) [])), (var_8: int64), (var_9: ((((string []) []) []) [])), (var_10: int64), (var_15: int64))\n    else\n        let (var_18: int64) = (var_1 + 1L)\n        method_6((var_5: ((string []) [])), (var_2: int64), (var_3: int64), (var_6: int64), (var_7: (((string []) []) [])), (var_8: int64), (var_9: ((((string []) []) []) [])), (var_10: int64), (var_18: int64))\nlet (var_6: ((((string []) []) []) [])) = Array.zeroCreate&lt;(((string []) []) [])&gt; (System.Convert.ToInt32(4L))\nlet (var_7: int64) = 0L\nmethod_3((var_6: ((((string []) []) []) [])), (var_7: int64))\nlet (var_8: (((string []) []) [])) = var_6.[int32 0L]\nlet (var_9: ((string []) [])) = var_8.[int32 0L]\nlet (var_10: (string [])) = var_9.[int32 0L]\nvar_10.[int32 2L] &lt;- \"princess\"\nlet (var_11: int64) = var_6.LongLength\nlet (var_12: int64) = 0L\nmethod_4((var_6: ((((string []) []) []) [])), (var_11: int64), (var_12: int64))\n<\/pre>\n<p>The continuation passing style is the key to a significant amount of abstractive power. It is difficult to understand in terms of what the program does, instead what is needed is to focus on what the program is.<\/p>\n<p>There are numerous ways of writing <code>find_index<\/code> incorrectly that would not get immediately caught by the type system.<\/p>\n<ol><li>Forgetting to pass in the array.<\/li>\n<\/ol><pre>...\nfind_index {state=(); next = inl _ -&gt; failwith (type (dim)) \"The princess is in another castle.\"}\n<\/pre>\n<pre>...\nmethod_3((var_6: ((((string []) []) []) [])), (var_7: int64))\nlet (var_8: (((string []) []) [])) = var_6.[int32 0L]\nlet (var_9: ((string []) [])) = var_8.[int32 0L]\nlet (var_10: (string [])) = var_9.[int32 0L]\nvar_10.[int32 2L] &lt;- \"princess\"\n(Env7((Env4((Env3((Env2((Env1((Env0(naked_type (*bool*))))))))))), (Env3((Env2((Env1((Env0(naked_type (*bool*))))))))), (Env6(Tuple5(4L, 4L, 4L, 4L)))))\n<\/pre>\n<p>Seeing a dozen nested <code>Env<\/code>s along with a naked type in the generated code is almost always a sign of forgetting to apply an argument somewhere.<\/p>\n<ol start=\"2\"><li>Passing the state in incorrectly in the else branch.<\/li>\n<\/ol><pre>...\n| _ -&gt; next state\n<\/pre>\n<p>This would still have it compile and run correctly, but the code would have 40 lines more of useless specializations. It won't be shown here.<\/p>\n<ol start=\"3\"><li>Passing the state even more incorrectly.<\/li>\n<\/ol><pre>inl rec find_index {next state} = function\n    | ar &amp; @array_is _ -&gt; \n        inl body {next state i} = find_index {next state=i::state} (ar i)\n        for' {from=0; near_to=array_length ar; state finally=next; body}\n    | \"princess\" -&gt; tuple_rev state\n    | _ -&gt; next state\n<\/pre>\n<p>This would cause it to diverge as it would continually append to <code>state<\/code> inside the loop.<\/p>\n<p>In general though, programs written in a higher order style tend to work well after they typecheck much like in F# despite the language feeling more dynamic. And it is not necessarily the case that Spiral is less typesafe than F#.<\/p>\n<p>In fact it is the opposite for tasks that require union types due to F# having insufficient polymorphism in its type system. Many tasks that would otherwise require writing an interpreter in other languages can be done at compile time in Spiral.<\/p>\n<p>Loop unrolling is just one of those examples.<\/p>\n<p>Before the section on Loops can be finished there is just one bit of cleaning up left to do. That would be to merge <code>for<\/code> and <code>for'<\/code> into one function. Here is the full example in its completed form with a last minute change to <code>by<\/code>. The new loops are going to go into the standard library.<\/p>\n<pre>inl for_template kind {d with body} =\n    inl finally =\n        match d with\n        | {finally} -&gt; finally\n        | _ -&gt; id\n    inl state = \n        match d with\n        | {state} -&gt; state\n        | _ -&gt; ()\n    inl check =\n        match d with\n        | {near_to} from -&gt; from &lt; near_to \n        | {to} from -&gt; from &lt;= to\n        | {down_to} from -&gt; from &gt;= down_to\n        | {near_down_to} from -&gt; from &gt; near_down_to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` needs be present.\"\n    inl from =\n        match d with\n        | {from=(!dyn from) ^ static_from=from} -&gt; from\n        | _ -&gt; error_type \"Only one of `from` and `static_from` field to loop needs to be present.\"\n    inl to =\n        match d with\n        | {(to ^ near_to ^ down_to ^ near_down_to)=to} -&gt; to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` is allowed.\"\n    inl by =\n        match d with\n        | {by} -&gt; by\n        | {to | near_to} -&gt; 1\n        | {down_to | near_down_to} -&gt; -1\n    inl rec loop {from state} =\n        inl body {from} = \n            if check from then \n                match kind with\n                | .Standard -&gt;\n                    loop {from=from+by; state=body {state i=from}}\n                | .CPSd -&gt;\n                    inl next state = loop {state from=from+by}\n                    body {next state i=from}\n            else finally state\n        if Tuple.forall lit_is (from,to,by) then body {from}\n        else \n            inl from = dyn from\n            join (body {from} : finally state)\n    loop {from state}\ninl for' = for_template .CPSd\ninl for = for_template .Standard\ninl array_init near_to f =\n    assert (near_to &gt;= 0) \"The input to init needs to be greater or equal to 0.\"\n    \/\/ Somewhat of an ugly practice in order to infer the type in a language that doesn't support inference. \n    \/\/ For large functions, it is recomended to put them in a join point otherwise compile times could \n    \/\/ become exponential if the function contains branches.\n    \/\/ For a simple map for an array like here, it does not matter.\n    inl typ = type (f 0) \n    inl ar = array_create typ near_to\n    for {from=0; near_to; body=inl {i} -&gt; ar i &lt;- f i}\n    ar\ninl rec zeroes = function\n    | x :: x' -&gt; array_init x (inl _ -&gt; zeroes x')\n    | () -&gt; \"\"\ninl dim = (4,4,4,4)\ninl ar = zeroes dim\nar 0 0 0 2 &lt;- \"princess\"\n\/\/ Reverses a tuple\ninl tuple_rev = \n    inl rec loop state = function\n        | x :: xs -&gt; loop (x :: state) xs\n        | () -&gt; state\n    loop ()\n\/\/ Correct version\ninl rec find_index {next state} = function\n    | ar &amp; @array_is _ -&gt; \n        inl body {next i} = find_index {next state=i::state} (ar i)\n        for' {from=0; near_to=array_length ar; finally=next; body}\n    | \"princess\" -&gt; tuple_rev state\n    | _ -&gt; next ()\nfind_index {state=(); next = inl _ -&gt; failwith (type (dim)) \"The princess is in another castle.\"} ar\n<\/pre>\n<p>There was a lot of material covered here. The logic of <code>find_index<\/code> as well as the other loop unrolling functions might seem confusing to the uninitiated, and would no doubt be to the author had he encountered this over a year ago. But ultimately the function is just 5 lines long and there is nothing particular magical about it; the function is fully explicit. Thinking about it for a long time will help and so will mentally rehearsing the motions until the pieces fall into place.<\/p>\n<p>One useful tool in gaining understanding is trying to manually expand the loop. Here is what happens if <code>find_index<\/code> is expanded a single step of recursion.<\/p>\n<pre>inl rec find_index {next state} = function\n    | ar &amp; @array_is _ -&gt; \n        inl body {next i} = \n            inl state = i :: state\n            match ar i with\n            | ar &amp; @array_is _ -&gt; \n                inl body {next i} = find_index {next state=i::state} (ar i)\n                for' {from=0; near_to=array_length ar; finally=next; body}\n            | \"princess\" -&gt; tuple_rev state\n            | _ -&gt; next ()\n        for' {from=0; near_to=array_length ar; finally=next; body}\n    | \"princess\" -&gt; tuple_rev state\n    | _ -&gt; next ()\n<\/pre>\n<p>Supposing the input is one dimensional, that is if the type of the array was <code>string []<\/code> it become possible to do more partial evaluation by hand.<\/p>\n<pre>inl rec find_index {next state} = function\n    | ar &amp; @array_is _ -&gt; \n        inl body {next i} = \n            inl state = i :: state\n            match ar i with\n            | \"princess\" -&gt; tuple_rev state\n            | _ -&gt; next ()\n        for' {from=0; near_to=array_length ar; finally=next; body}\n<\/pre>\n<p>This program corresponds to a single loop and is in fact what the program would get specialized to had it been given only a string array as input.<\/p>\n<p>Seeing similar examples of this pattern will no doubt help and there will be a significant number of them throughout these tutorials.<\/p>\n<h4><a id=\"user-content-arrays\" class=\"anchor\" aria-hidden=\"true\" href=\"#arrays\"><\/a>Arrays<\/h4>\n<p>Compared to the intensity of the previous section, this one should be a breeze in comparison. The most important of the array functions <code>init<\/code> was already introduced.<\/p>\n<pre>let array =\n    (\n    \"Array\",[tuple;loops],\"The array module\",\n    \"\"\"\nopen Loops\n\/\/\/ Creates an empty array with the given type.\n\/\/\/ t -&gt; t array\ninl empty t = array_create t 0\n\/\/\/ Creates a singleton array with the given element.\n\/\/\/ x -&gt; t array\ninl singleton x =\n    inl ar = array_create x 1\n    ar 0 &lt;- x\n    ar\n\/\/\/ Applies a function to each elements of the collection, threading an accumulator argument through the computation.\n\/\/\/ If the input function is f and the elements are i0..iN then computes f..(f i0 s)..iN.\n\/\/\/ (s -&gt; a -&gt; s) -&gt; s -&gt; a array -&gt; s\ninl foldl f state ar = for {from=0; near_to=array_length ar; state; body=inl {state i} -&gt; f state (ar i)}\n\/\/\/ Applies a function to each element of the array, threading an accumulator argument through the computation. \n\/\/\/ If the input function is f and the elements are i0...iN then computes f i0 (...(f iN s)).\n\/\/\/ (a -&gt; s -&gt; a) -&gt; a array -&gt; s -&gt; s\ninl foldr f ar state = for {from=array_length ar-1; down_to=0; state; body=inl {state i} -&gt; f (ar i) state}\n...\n<\/pre>\n<p>Here are some of the basic ones. Having a loop as a part of the standard library makes it really easy to implement the two <code>fold<\/code> functions. Unlike in F# where <code>foldl<\/code> and <code>foldr<\/code> are <code>fold<\/code> and <code>foldBack<\/code>, here the Haskell naming convention has been followed for no special reason apart from <code>foldr<\/code> being more elegant to write than <code>foldBack<\/code>.<\/p>\n<pre>\/\/ Creates an array given a dimension and a generator function to compute the elements.\n\/\/ ?(.is_static) -&gt; int -&gt; (int -&gt; a) -&gt; a array\ninl init = \n    inl body is_static n f =\n        assert (n &gt;= 0) \"The input to init needs to be greater or equal to 0.\"\n        inl typ = type (f 0)\n        inl ar = array_create typ n\n        inl d = \n            inl d = {near_to=n; body=inl {i} -&gt; ar i &lt;- f i}\n            if is_static then {d with from = 0} else {d with static_from = 0}\n        for d\n        ar\n    function\n    | .static -&gt; body true\n    | n -&gt; body false n\n<\/pre>\n<p><code>init<\/code> here is a example how the architecture of a Spiral function differs from that in F#. The interesting part is that <code>.static<\/code> can be passed into it as an optional argument that will allow it to be run statically. Otherwise this is the same as the <code>init<\/code> from the previous section so no demonstration of it should be necessary.<\/p>\n<pre>\/\/\/ Builds a new array that contains elements of a given array.\n\/\/\/ a array -&gt; a array\nmet copy ar = init (array_length ar) ar\n\/\/\/ Builds a new array whose elements are the result of applying a given function to each of the elements of the array.\n\/\/\/ (a -&gt; b) -&gt; a array -&gt; a array\ninl map f ar = init (array_length ar) (ar &gt;&gt; f)\n<\/pre>\n<p><code>init<\/code> is useful as it is easy to derive other functions from it. These two function exactly as in F# and other functional languages.<\/p>\n<pre>\/\/\/ Returns a new array containing only elements of the array for which the predicate function returns `true`.\n\/\/\/ (a -&gt; bool) -&gt; a array -&gt; a array\ninl filter f ar =\n    inl ar' = array_create ar.elem_type (array_length ar)\n    inl count = foldl (inl s x -&gt; if f x then ar' s &lt;- x; s+1 else s) (dyn 0) ar\n    init count ar'\n<\/pre>\n<p>Being able to apply arrays directly instead of having to index them allows them to be used more like functions. Also worthy of note is the <code>ar.elem_type<\/code>. In Spiral there is no inference, only propagation so the type of an array must be extracted directly. In Spiral, types are first class and can be used as values. This can be exploited to get around the lack of inference in most cases.<\/p>\n<p>Arrays are a simple examples of how types might be held in structures.<\/p>\n<p>Unlike other .NET types, arrays are built into the language directly.<\/p>\n<pre>\/\/\/ Merges all the arrays in a tuple into a single one.\n\/\/\/ a array tuple -&gt; a array\ninl append l =\n    inl ar' = array_create ((fst l).elem_type) (Tuple.foldl (inl s l -&gt; s + array_length l) 0 l)\n    inl ap s ar = foldl (inl i x -&gt; ar' i &lt;- x; i+1) s ar\n    Tuple.foldl ap (dyn 0) l |&gt; ignore\n    ar'\n<\/pre>\n<p>Like how <code>init<\/code> can match on its first arguments before deciding whether to run statically or not, being able to iterate over tuples in order to merge the arrays is a standard use case for intensional polymorphism.<\/p>\n<pre>\/\/\/ Flattens an array of arrays into a single one.\n\/\/\/ a array array -&gt; a array\ninl concat ar =\n    inl count = foldl (inl s ar -&gt; s + array_length ar) (dyn 0) ar\n    inl ar' = array_create ar.elem_type.elem_type count\n    (foldl &lt;&lt; foldl) (inl i x -&gt; ar' i &lt;- x; i+1) (dyn 0) ar |&gt; ignore\n    ar'\n<\/pre>\n<p><code>foldl &lt;&lt; foldl<\/code> is a good way to compose folds for nested arrays.<\/p>\n<p>Writing functions in this higher order style is the optimal way to program in Spiral. For contrast, here is how <code>concat<\/code> is implemented in F#'s source.<\/p>\n<pre>let concatArrays (arrs : 'T[][]) : 'T[] =\n    let mutable acc = 0    \n    for h in arrs do\n        acc &lt;- acc + h.Length        \n        \n    let res = Microsoft.FSharp.Primitives.Basics.Array.zeroCreateUnchecked acc  \n        \n    let mutable j = 0\n    for i = 0 to arrs.Length-1 do     \n        let h = arrs.[i]\n        let len = h.Length\n        Array.Copy(h,0,res,j,len)        \n        j &lt;- j + len\n    res               \n<\/pre>\n<p>It is not quite C, but it is the same style inherited from it. All fast languages tend to regress to that particular kind of programming when performance or just the guarantee of it becomes a necessity.<\/p>\n<p>Even in a pure and lazy language like Haskell, looking under the hood of some of its fast libraries will reveal this and other kinds of regressions.<\/p>\n<p>On the strength of its inlining guarantees, the goal of Spiral is to liberate programmers from that gravitic impulse towards C.<\/p>\n<p>During the last 45 years there have been numerous attempts at bridging the expressiveness of dynamic languages with the performance of C, none of which have borne fruit.<\/p>\n<p>Assuming Spiral can be scaled, it or some other language of similar design with powerful first class types and staging features will finally break beyond the atmosphere to bring light of civilization into the cold, dead space that lies beyond.<\/p>\n<p>The above example is not that bad actually. It is only 12 lines in F# vs 5 in Spiral. It is hardly a reason to create a new language and propose the jettison of parametric polymoprhism.<\/p>\n<p>In the following chapters there will be examples of programs, most notably of Spiral's tensors, whose functionalities have such requirements that would pretty much break any existing language.<\/p>\n<p>The next two functions are all that remains of the module.<\/p>\n<pre>\/\/\/ Tests if all the elements of the array satisfy the given predicate.\n\/\/\/ (a -&gt; bool) -&gt; a array -&gt; bool\ninl forall f ar = for' {from=0; near_to=array_length ar; state=true; body = inl {next state i} -&gt; f (ar i) &amp;&amp; next state}\n\/\/\/ Tests if any the element of the array satisfies the given predicate.\n\/\/\/ (a -&gt; bool) -&gt; a array -&gt; bool\ninl exists f ar = for' {from=0; near_to=array_length ar; state=false; body = inl {next state i} -&gt; f (ar i) || next state}\n{empty singleton foldl foldr init copy map filter append concat forall exists} \n|&gt; stack\n    \"\"\") |&gt; module_\n<\/pre>\n<p>On the F# side it is necessary to wrap the module in a type using the <code>module_<\/code> function. That <code>|&gt; stack<\/code> at the end is not necessary and only has something to do with the way the language is currently implemented. Omitting the conversion of the module to a layout type would not break anything, at most there might be a minor compilation slowdown. More details are (to be) provided in the user guide.<\/p>\n<p>Modules with no free variables such as the <code>Array<\/code> module whose fields are entirely made of combinators always get converted into naked types rather than variables and hence have no overhead.<\/p>\n<h3><a id=\"user-content-3-union-types-and-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#3-union-types-and-lists\"><\/a>3: Union Types and Lists<\/h3>\n<p>Discriminated union types in Spiral take direct inspiration from F#'s own. Having said that, the lack of type inference and the aggressive unboxing of them by the Spiral evaluator makes them less convenient to work with. Nonetheless, union types capture the essence of dynamism and are absolutely essential in a modern language.<\/p>\n<p>Since Spiral has first class types, type string literals take the place of case names. Furthermore, types can be defined anywhere in the program rather than only at the top level like in F#.<\/p>\n<p>A non-recursive union type like the Option can be defined like the following. <code>\\\/<\/code> is the type union keyword operator. It has a lower precedence than tuples.<\/p>\n<pre>inl Option x = .Some, x \\\/ .None\n\/\/ constructors\ninl some x = box (Option x) (.Some, x)\ninl none x = box (Option x) (.None)\nnone int64\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0 of Tuple1\n    | Union0Case1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\nUnion0Case1\n<\/pre>\n<p>Commentary on the quality of the generated code will be left for the user guide. Pattern matching on the boxed union values can be done the same way as in F#.<\/p>\n<pre>match none int64 with\n| .Some, x -&gt; x\n| .None -&gt; -11\n<\/pre>\n<pre>11L\n<\/pre>\n<p>The word 'staging' means 'defering for later'. Just like literals, the creation of union types is deferred for as long as possible in Spiral.<\/p>\n<p>In order to actually instantiate the type, it is necessary to <code>dyn<\/code> it or return it from a join point or an if branch. The end of the entire program also qualifies for instantiation.<\/p>\n<pre>match none int64 |&gt; dyn with\n| .Some, x -&gt; x\n| .None -&gt; -11\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0 of Tuple1\n    | Union0Case1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\nlet (var_0: Union0) = Union0Case1\nmatch var_0 with\n| Union0Case0(var_1) -&gt;\n    var_1.mem_0\n| Union0Case1 -&gt;\n    -11L\n<\/pre>\n<p>The above is roughly what would be expect to in F# or the MLs. Spiral's pattern matching is more flexible though.<\/p>\n<pre>inl TypeA = .A \\\/ .B\ninl TypeB = .B \\\/ .C\ninl f = function\n    | .A -&gt; 1\n    | .B -&gt; 2\n    | .C -&gt; 3\nbox TypeA .A |&gt; dyn |&gt; f |&gt; ignore\nbox TypeB .C |&gt; dyn |&gt; f |&gt; ignore\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\nand Union1 =\n    | Union1Case0\n    | Union1Case1\nlet (var_0: Union0) = Union0Case0\nlet (var_1: int64) =\n    match var_0 with\n    | Union0Case0 -&gt;\n        1L\n    | Union0Case1 -&gt;\n        2L\nlet (var_2: Union1) = Union1Case1\nlet (var_3: int64) =\n    match var_2 with\n    | Union1Case0 -&gt;\n        2L\n    | Union1Case1 -&gt;\n        3L\n<\/pre>\n<p>Despite this added flexibility, it is in fact exhaustive. Unlike in F#, this is not a warning, but an error as Spiral's union types are intended to be used on devices which have no capabilities for raising exceptions.<\/p>\n<pre>inl TypeA = .A \\\/ .B\ninl f = function\n    | .A -&gt; 1\n    | .C -&gt; 3\nbox TypeA .A |&gt; dyn |&gt; f |&gt; ignore\n<\/pre>\n<pre>...\nError trace on line: 35, column: 7 in file \"example\".\n    | .A -&gt; 1\n      ^\nError trace on line: 36, column: 7 in file \"example\".\n    | .C -&gt; 3\n      ^\nPattern miss error. The argument is type (type_lit (B))\n<\/pre>\n<p>As it never matches <code>.B<\/code> it goes over the edge and returns a type error.<\/p>\n<p>Here is how recursive datatypes like lists might be defined.<\/p>\n<pre>let example = \n    \"example\",[option;tuple;loops],\"Module description.\",\n    \"\"\"\nopen Loops\ninl rec List x = join_type () \\\/ x, List x\n\/\/\/ Creates an empty list with the given type.\n\/\/\/ t -&gt; List t\ninl empty x = box (List x) ()\n\/\/\/ Creates a single element list with the given type.\n\/\/\/ x -&gt; List x\ninl singleton x = box (List x) (x, empty x)\n\/\/\/ Immutable appends an element to the head of the list.\n\/\/\/ x -&gt; List x -&gt; List x\ninl cons a b = \n    inl t = List a\n    box t (a, box t b)\nsingleton 3 |&gt; cons 2 |&gt; cons 1\n    \"\"\"\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\n(Rec0Case1(Tuple1(1L, (Rec0Case1(Tuple1(2L, (Rec0Case1(Tuple1(3L, Rec0Case0)))))))))\n<\/pre>\n<p><code>join_type<\/code> is similar to the standard <code>join<\/code> except it is used to define types. It always returns a naked type and on entry converts everything in the environment to their types and that includes literals. That means that passing literals through the type join point requires doing it on the type level or inside a layout type.<\/p>\n<pre>\/\/\/ Creates a list by calling the given generator on each index.\n\/\/\/ ?(.static) -&gt; int -&gt; (int -&gt; a) -&gt; List a\ninl init =\n    inl body is_static n f =\n        inl t = type (f 0)\n        inl d = {near_to=n; state=empty t; body=inl {next i state} -&gt; cons (f i) (next state)}\n        if is_static then for' {d with static_from=0}\n        else for' {d with from=0}\n    function\n    | .static -&gt; body true\n    | x -&gt; body false x\n<\/pre>\n<p>The above function resembles the <code>init<\/code> in the <code>Array<\/code> module in structure. There is an interesting usage of the breakable <code>for'<\/code> here. Usually the <code>next<\/code> is intended to be called in tail position, but here it is not. Instead the <code>state<\/code> is used merely to ship the empty list to the end of it.<\/p>\n<pre>inl x = init.static 3 id |&gt; dyn\n()\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(0L, (Rec0Case1(Tuple1(1L, (Rec0Case1(Tuple1(2L, Rec0Case0)))))))))\n<\/pre>\n<p>The above is nearly identical to the <code>singleton 3 |&gt; cons 2 |&gt; cons 1<\/code> example.<\/p>\n<p>The next function on the list would be the <code>map<\/code>. This is where things start to get tricky. Here is an example of it that does not work.<\/p>\n<pre>inl rec map f l = \n    inl loop l =\n        match l with\n        | x,xs -&gt; cons (f x) (map f xs)\n        | () -&gt; l \/\/ Error #1\n        : ??? \/\/ Error #2\n    if box_is l then loop l\n    else join loop l\n<\/pre>\n<p>Error #2 should be obvious - there is no return type. Error #1 is more subtle - and is related to the way pattern matching is compiled.<\/p>\n<p>Backtracking to the earlier example.<\/p>\n<pre>inl TypeA = .A \\\/ .B\ninl f = function\n    | x -&gt; x\nbox TypeA .A |&gt; dyn |&gt; f |&gt; ignore\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\nlet (var_0: Union0) = Union0Case0\n<\/pre>\n<p>The above is as one would expect.<\/p>\n<pre>inl TypeA = .A \\\/ .B\ninl f = function\n    | \"qwe\" -&gt; ()\n    | x -&gt; x\nbox TypeA .A |&gt; dyn |&gt; f |&gt; ignore\n<\/pre>\n<pre>...\nError trace on line: 35, column: 7 in file \"example\".\n    | \"qwe\" -&gt; ()\n      ^\nAll the cases in pattern matching clause with dynamic data must have the same type.\nGot: [type_lit (A), type_lit (B)]\n<\/pre>\n<p>The error message gives an indication of what is wrong. In Spiral, the match case is not what triggers unboxing - the operations that actually need to unbox the union type are what do it. That means literal, type literal, type equality, tuple and module patterns. This applies even to those patterns that have nothing to do with the variable's type and would have been expected to be skipped.<\/p>\n<p>It gets worse. Spiral is really aggressive at rewriting the terms it is unboxing even if they are outside its intended scope.<\/p>\n<pre>inl x = box TypeA .A |&gt; dyn\nprint_static x \/\/ var (union {type_lit (A) | type_lit (B)})\nmatch x with\n| \"qwe\" -&gt; ()\n| _ -&gt; \n    \/\/ prints twice\n    \/\/ type (type_lit (A))\n    \/\/ type (type_lit (B))\n    print_static x \n    x\n<\/pre>\n<pre>...\nAll the cases in pattern matching clause with dynamic data must have the same type.\nGot: [type_lit (A), type_lit (B)]\n<\/pre>\n<p>The way things are currently is the fault of whoever wrote the pattern matching compiler. Since patterns would be difficult to compile otherwise, internally Spiral uses the same mechanism used to do common subexpression elimination to pass information over multiple branches. There is no issue at all with this when not dealing with union types, but here there is some friction there.<\/p>\n<p>There is something good about the current arrangement that MLs do not have.<\/p>\n<pre>inl TypeA = .A \\\/ .B \\\/ .C \\\/ .D\ninl f g = function\n    | .A -&gt; 1\n    | .B -&gt; 2\n    | x -&gt; \n        dyn \"Just passing through.\" |&gt; ignore\n        g x\nf (function\n    | .C -&gt; 3\n    | .D -&gt; 4)\n    (box TypeA .A |&gt; dyn )\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\n    | Union0Case2\n    | Union0Case3\nlet (var_0: Union0) = Union0Case0\nmatch var_0 with\n| Union0Case0 -&gt;\n    1L\n| Union0Case1 -&gt;\n    2L\n| Union0Case2 -&gt;\n    let (var_1: string) = \"Just passing through.\"\n    3L\n| Union0Case3 -&gt;\n    let (var_2: string) = \"Just passing through.\"\n    4L\n<\/pre>\n<p>That would be that the exhaustiveness check is not local to the pattern. As long as all the branches of it are properly handled, the pattern does not have to be squeezed all into one place and can be composed. This is one of the safety aspects at compile time that F# does not have.<\/p>\n<p>Regardless of the merits and demerits of this approach, in order to complete the map function some kind of method for getting what would be the generic parameter of the list in a parametric language is needed.<\/p>\n<h4><a id=\"user-content-type-splitting-and-generic-parameters\" class=\"anchor\" aria-hidden=\"true\" href=\"#type-splitting-and-generic-parameters\"><\/a>Type Splitting and Generic Parameters<\/h4>\n<pre>\/\/\/ Returns the element type of the list.\n\/\/\/ a List -&gt; a type\ninl elem_type l =\n    match split l with\n    | (), (a,b) when eq_type (List a) l -&gt; a\n    | _ -&gt; error_type \"Expected a List in elem_type.\"\n<\/pre>\n<p>The way the <code>split<\/code> function works is that it splits an union type into its individual components and returns them as a tuple. After it has been split, this makes it possible to match on the types of it directly.<\/p>\n<pre>inl Option x = .Some, x \\\/ .None\nprint_static (Option int64) \/\/ type (union {[type_lit (Some), int64] | type_lit (None)})\nprint_static (Option int64 |&gt; split) \/\/ [type ([type_lit (Some), int64]), type (type_lit (None))]\n<\/pre>\n<p>This is a bit of a hack. Spiral has union and not sum types, meaning they are not ordered. Or better put, they are ordered, just not based on how they were entered.<\/p>\n<p>The above example works for lists and is how they are implemented in the standard library, but there are alternative ways of implementing the basic list.<\/p>\n<pre>inl rec List x = join_type \n    inl el = stack {elem_type=x}\n    el, () \\\/ el, (x, List x)\n<\/pre>\n<p>They all involve sticking the type in directly by using layout types. Since layout types capture the scope by the expression instead of type and since <code>x<\/code> can only ever be a naked type once it passes the <code>join_type<\/code> point, that assures that it will always be instantiated.<\/p>\n<p>If adding <code>el<\/code> to all the branches of a larger type by hand is tedious, it is possible to automate that. It needs to be done inside the type join point. Here is how it would be done on a tuple.<\/p>\n<pre>inl rec List x = join_type \n    inl el = stack {elem_type=x}\n    inl typ = () \\\/ x, List x\n    split typ\n    |&gt; Tuple.map (inl x -&gt; el, x)\n    |&gt; Tuple.reducel (inl a b -&gt; a \\\/ b)\n\/\/ [type ([layout_stack {elem_type=type (int64)}, []]), type ([layout_stack {elem_type=type (int64)}, [int64, rec_type 0]])]\nprint_static (split (List int64))\n<\/pre>\n<p>Using first class types Spiral can emulate what would be generic parameters of a container in a language with parametric polymorphism.<\/p>\n<p>With <code>elem_type<\/code> in hand, it becomes possible to implement map.<\/p>\n<pre>\/\/\/ Builds a new list whose elements are the results of applying the given function to each of the elements of the list.\n\/\/\/ (a List -&gt; b List) -&gt; a List -&gt; List b\ninl rec map f l = \n    inl t = elem_type l\n    inl loop = function\n        | x,xs -&gt; cons (f x) (map f xs)\n        | () -&gt; empty t\n    if box_is l then loop l\n    else join loop l : List t\ninl l = init.static 3 id |&gt; map ((*) 2) |&gt; dyn\n()\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(0L, (Rec0Case1(Tuple1(2L, (Rec0Case1(Tuple1(4L, Rec0Case0)))))))))\n<\/pre>\n<p>The static version of map works fine now.<\/p>\n<p>Here is how the non-static version looks like.<\/p>\n<pre>init 3 id |&gt; map ((*) 2) |&gt; dyn\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_1((var_0: int64)): Rec0 =\n    let (var_1: bool) = (var_0 &lt; 3L)\n    if var_1 then\n        let (var_2: int64) = (var_0 + 1L)\n        let (var_3: Rec0) = method_1((var_2: int64))\n        (Rec0Case1(Tuple1(var_0, var_3)))\n    else\n        Rec0Case0\nand method_2((var_0: Rec0)): Rec0 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        Rec0Case0\n    | Rec0Case1(var_1) -&gt;\n        let (var_2: int64) = var_1.mem_0\n        let (var_3: Rec0) = var_1.mem_1\n        let (var_4: int64) = (2L * var_2)\n        let (var_5: Rec0) = method_2((var_3: Rec0))\n        (Rec0Case1(Tuple1(var_4, var_5)))\nlet (var_0: int64) = 0L\nlet (var_1: Rec0) = method_1((var_0: int64))\nmethod_2((var_1: Rec0))\n<\/pre>\n<p>This does not demonstrate Spiral's true power. The function can map over lists that are partially static.<\/p>\n<pre>inl l = dyn (singleton 3) |&gt; cons 2 |&gt; cons 1 |&gt; cons 0\nmap ((*) 2) l |&gt; dyn\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_1((var_0: Rec0)): Rec0 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        Rec0Case0\n    | Rec0Case1(var_1) -&gt;\n        let (var_2: int64) = var_1.mem_0\n        let (var_3: Rec0) = var_1.mem_1\n        let (var_4: int64) = (2L * var_2)\n        let (var_5: Rec0) = method_1((var_3: Rec0))\n        (Rec0Case1(Tuple1(var_4, var_5)))\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(3L, Rec0Case0)))\nlet (var_1: Rec0) = method_1((var_0: Rec0))\n(Rec0Case1(Tuple1(0L, (Rec0Case1(Tuple1(2L, (Rec0Case1(Tuple1(4L, var_1)))))))))\n<\/pre>\n<p>The first 3 elements are done at compile time, and the rest is done at runtime.<\/p>\n<p>With the map done, <code>foldl<\/code> and <code>foldr<\/code> are straightforward enough.<\/p>\n<pre>\/\/\/ Applies a function f to each element of the collection, threading an accumulator argument through the computation. \n\/\/\/ The fold function takes the second argument, and applies the function f to it and the first element of the list. \n\/\/\/ Then, it feeds this result into the function f along with the second element, and so on. It returns the final result. \n\/\/\/ If the input function is f and the elements are i0...iN, then this function computes f (... (f s i0) i1 ...) iN.\n\/\/\/ (s -&gt; a -&gt; s) -&gt; s -&gt; a List -&gt; s\ninl rec foldl f s l = \n    inl loop = function\n        | x, xs -&gt; foldl f (f s x) xs\n        | () -&gt; s\n    if box_is l then loop l\n    else join loop l : s\ninl l = dyn (singleton 3) |&gt; cons 2 |&gt; cons 1 |&gt; cons 0\nfoldl (+) (dyn 0) l\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_1((var_0: Rec0), (var_1: int64)): int64 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        var_1\n    | Rec0Case1(var_2) -&gt;\n        let (var_3: int64) = var_2.mem_0\n        let (var_4: Rec0) = var_2.mem_1\n        let (var_5: int64) = (var_1 + var_3)\n        method_1((var_4: Rec0), (var_5: int64))\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(3L, Rec0Case0)))\nlet (var_1: int64) = 0L\nlet (var_2: int64) = (var_1 + 1L)\nlet (var_3: int64) = (var_2 + 2L)\nmethod_1((var_0: Rec0), (var_3: int64))\n<\/pre>\n<p>One thing the example above demonstrates is that Spiral does require the user to know whether compile time or runtime execution is being targeted. The above fragment is not ideal since it would be better to sum the static part of the list and then <code>dyn<\/code> the state rather than do so at the beginning.<\/p>\n<pre>\/\/\/ Applies a function to each element of the collection, threading an accumulator argument through the computation. \n\/\/\/ If the input function is f and the elements are i0...iN, then this function computes f i0 (...(f iN s)).\n\/\/\/ (a -&gt; s -&gt; s) -&gt; a List -&gt; s -&gt; s\ninl rec foldr f l s = \n    inl loop = function\n        | x, xs -&gt; f x (foldr f xs s)\n        | () -&gt; s\n    if box_is l then loop l\n    else join loop l : s\ninl l = dyn (singleton 3) |&gt; cons 2 |&gt; cons 1 |&gt; cons 0\nfoldr (+) l (dyn 0)\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_1((var_0: Rec0), (var_1: int64)): int64 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        var_1\n    | Rec0Case1(var_2) -&gt;\n        let (var_3: int64) = var_2.mem_0\n        let (var_4: Rec0) = var_2.mem_1\n        let (var_5: int64) = method_1((var_4: Rec0), (var_1: int64))\n        (var_3 + var_5)\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(3L, Rec0Case0)))\nlet (var_1: int64) = 0L\nlet (var_2: int64) = method_1((var_0: Rec0), (var_1: int64))\nlet (var_3: int64) = (2L + var_2)\n(1L + var_3)\n<\/pre>\n<p>The next are <code>head<\/code> and <code>tail<\/code>.<\/p>\n<pre>open Option\n\/\/\/ Returns the first element of the list.\n\/\/\/ a List -&gt; a Option\ninl head l =\n    inl t = elem_type l\n    match l with\n    | x, xs -&gt; some x\n    | () -&gt; none t\n\/\/\/ Returns the list without the first element.\n\/\/\/ a List -&gt; a List Option\ninl tail l =\n    inl t = elem_type l\n    match l with\n    | x, xs -&gt; some xs\n    | () -&gt; none (List t)\n<\/pre>\n<p>As the above are straightforward so there is no need to run them. That being said, it would be interesting to know how it might be possible to implement them in continuation passing style for greater efficiency.<\/p>\n<pre>\/\/\/ Returns the first element of the list.\n\/\/\/ a List -&gt; {some=(a -&gt; a) none=(a type -&gt; a)} -&gt; a\ninl head' l {some none} =\n    inl t = elem_type l\n    match l with\n    | x, xs -&gt; some x\n    | () -&gt; none t\n\/\/\/ Returns the list without the first element.\n\/\/\/ a List -&gt; {some=(a List -&gt; a List) none=(a List type -&gt; a List)} -&gt; a List\ninl tail' l {some none} =\n    inl t = elem_type l\n    match l with\n    | x, xs -&gt; some xs\n    | () -&gt; none (List t)\ninl l = dyn (singleton 3)\ntail' l {\n    some = id\n    none = inl x -&gt; failwith x \"The list is empty.\"\n    }\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(3L, Rec0Case0)))\nmatch var_0 with\n| Rec0Case0 -&gt;\n    (failwith \"The list is empty.\")\n| Rec0Case1(var_1) -&gt;\n    let (var_3: int64) = var_1.mem_0\n    var_1.mem_1\n<\/pre>\n<p>The best is left for <code>last<\/code>.<\/p>\n<pre>\/\/\/ Returns the last element of the list.\n\/\/\/ a List -&gt; a Option\ninl last l =\n    inl t = elem_type l\n    foldl (inl _ x -&gt; some x) (none t) l\ninl l = dyn (singleton 3) |&gt; cons 2 |&gt; cons 1 |&gt; cons 0\nlast l \n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nand Union2 =\n    | Union2Case0 of Tuple3\n    | Union2Case1\nand Tuple3 =\n    struct\n    val mem_0: int64\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\nlet rec method_1((var_0: Rec0)): Union2 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        (Union2Case0(Tuple3(2L)))\n    | Rec0Case1(var_1) -&gt;\n        let (var_2: int64) = var_1.mem_0\n        let (var_3: Rec0) = var_1.mem_1\n        method_2((var_3: Rec0), (var_2: int64))\nand method_2((var_0: Rec0), (var_1: int64)): Union2 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        (Union2Case0(Tuple3(var_1)))\n    | Rec0Case1(var_2) -&gt;\n        let (var_3: int64) = var_2.mem_0\n        let (var_4: Rec0) = var_2.mem_1\n        method_2((var_4: Rec0), (var_3: int64))\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(3L, Rec0Case0)))\nmethod_1((var_0: Rec0))\n<\/pre>\n<p>The above way of specializing it is close to ideal. It would be better had <code>method_1<\/code> been inlined, but this is a decent showing. As can be seen, the option type is staged and only the int inside is passed through until it is time to return from the function at which point the instantiation happens. In F#, this way of doing <code>last<\/code> would be grossly inefficient as a new option would be instantiated at each step. Very few languages allow passing of literals across call boundaries due to the uncertainty whether the optimizer will diverge. Spiral achieves its efficiency by making dealing with the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Halting_problem\" rel=\"nofollow\">halting problem<\/a> the user's responsibility. This is not a bad strategy - as the halting problem is NP Hard, other compilers' optimizers have no choice but to rely on fallible heuristics whereas the user has to determine whether the program will terminate anyway and has no say in what the black box is deciding. It is not so in Spiral.<\/p>\n<p>The essence of Spiral is to convert the termination proofs implicitly and informally present in the program into polymorphism.<\/p>\n<p>Here is how it would be done in CPS for that last bit of efficiency.<\/p>\n<pre>\/\/\/ Returns the last element of the list.\n\/\/\/ a List -&gt; {some=(a -&gt; a) none=(a type -&gt; a)} -&gt; a\ninl rec last' l {some none} =\n    inl t = elem_type l\n    inl loop = function\n        | x, xs -&gt; last' xs {some none=some x}\n        | () -&gt; none t\n    if box_is l then loop l\n    else join loop l : none t\ninl l = dyn (singleton 3)\nlast' l {\n    some = inl x _ -&gt; x\n    none = inl t -&gt; failwith t \"The list is empty.\"\n    }\n<\/pre>\n<pre>type Rec0 =\n    | Rec0Case0\n    | Rec0Case1 of Tuple1\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Rec0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet rec method_1((var_0: Rec0)): int64 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        (failwith \"The list is empty.\")\n    | Rec0Case1(var_1) -&gt;\n        let (var_3: int64) = var_1.mem_0\n        let (var_4: Rec0) = var_1.mem_1\n        method_2((var_4: Rec0), (var_3: int64))\nand method_2((var_0: Rec0), (var_1: int64)): int64 =\n    match var_0 with\n    | Rec0Case0 -&gt;\n        var_1\n    | Rec0Case1(var_2) -&gt;\n        let (var_3: int64) = var_2.mem_0\n        let (var_4: Rec0) = var_2.mem_1\n        method_2((var_4: Rec0), (var_3: int64))\nlet (var_0: Rec0) = (Rec0Case1(Tuple1(3L, Rec0Case0)))\nmethod_1((var_0: Rec0))\n<\/pre>\n<p>It can no longer be implemented in terms of fold, but otherwise is rather simple.<\/p>\n<p>Note that <code>head'<\/code>, <code>tail'<\/code> and <code>last'<\/code> are just more generic versions of the non-CPS versions. Assuming the 3 original functions were missing, here is how they might be implemented in terms of CPS'd ones.<\/p>\n<pre>\/\/\/ Returns the first element of the list.\n\/\/\/ a List -&gt; a Option\ninl head l = head' l {some none}\n\/\/\/ Returns the list without the first element.\n\/\/\/ a List -&gt; a List Option\ninl tail l = tail' l {some none}\n\/\/\/ Returns the last element of the list.\n\/\/\/ a List -&gt; a Option\ninl last l = last' l {some=const &lt;&lt; some; none}\n<\/pre>\n<p>Here <code>const<\/code> is simply <code>inl x _ -&gt; x<\/code> as was used inside <code>some<\/code> of the previous example. The continuation passing style is great for writing generic code in Spiral as it meshes well with its typing scheme. The monadic computations that will be shown in the following chapters are just syntax sugar over CPS.<\/p>\n<p>The capacity to make specialized functions from generic one like the above is an important factor in ensuring code correctness. Eliminating code duplication and ensuring single responsibility is possible without performance impact in Spiral.<\/p>\n<pre>\/\/\/ Returns a new list that contains the elements of the first list followed by elements of the second.\n\/\/\/ a List -&gt; a List -&gt; a List\ninl append = foldr cons\n\/\/\/ Returns a new list that contains the elements of each list in order.\n\/\/\/ a List List -&gt; a List\ninl concat l &amp; !elem_type !elem_type t = foldr append l (empty t)\n{List empty cons init map foldl foldr singleton head' tail' last' head tail last append concat} |&gt; stack\n<\/pre>\n<p>With this, the new List module is done.<\/p>\n<h4><a id=\"user-content-warning-on-combining-union-types-partial-active-patterns-and-join-points\" class=\"anchor\" aria-hidden=\"true\" href=\"#warning-on-combining-union-types-partial-active-patterns-and-join-points\"><\/a>Warning on combining union types, partial active patterns and join points<\/h4>\n<p>Union types in Spiral are an example of a well designed feature with some implementation issues. Union types work, partial active patterns work and join points work, but right now they are a pick two out of three kind of deal. The reason for this is related to how Spiral will aggressively rewrite even variables outside of its intended scope.<\/p>\n<pre>inl ab = box (.A \\\/ .B)\ninl x = dyn (ab .A, ab .A, ab .A)\nmatch x with\n| .A, .A, _ -&gt; 1\n| .A, .B, .B -&gt; 2\n| _, _, .A -&gt; 3\n| _ -&gt; 4   \n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\nlet (var_0: Union0) = Union0Case0\nlet (var_1: Union0) = Union0Case0\nlet (var_2: Union0) = Union0Case0\nmatch var_0 with\n| Union0Case0 -&gt;\n    match var_1 with\n    | Union0Case0 -&gt;\n        1L\n    | Union0Case1 -&gt;\n        match var_2 with\n        | Union0Case0 -&gt;\n            3L\n        | Union0Case1 -&gt;\n            2L\n| Union0Case1 -&gt;\n    match var_2 with\n    | Union0Case0 -&gt;\n        3L\n    | Union0Case1 -&gt;\n        4L\n<\/pre>\n<p>The above compiles nicely, but suppose a partial active pattern with a join point is inserted in the middle.<\/p>\n<pre>inl ab = box (.A \\\/ .B)\ninl x = dyn (ab .A, ab .A, ab .A)\ninl pat arg on_fail on_succ = join on_fail ()\nmatch x with\n| .A, .A, _ -&gt; 1\n| @pat _ -&gt; -1\n| .A, .B, .B -&gt; 2\n| _, _, .A -&gt; 3\n| _ -&gt; 4    \n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\nlet rec method_0((var_0: Union0), (var_1: Union0), (var_2: Union0)): int64 =\n    match var_0 with\n    | Union0Case0 -&gt;\n        match var_1 with\n        | Union0Case0 -&gt;\n            match var_2 with\n            | Union0Case0 -&gt;\n                3L\n            | Union0Case1 -&gt;\n                4L\n        | Union0Case1 -&gt;\n            match var_2 with\n            | Union0Case0 -&gt;\n                3L\n            | Union0Case1 -&gt;\n                2L\n    | Union0Case1 -&gt;\n        match var_2 with\n        | Union0Case0 -&gt;\n            3L\n        | Union0Case1 -&gt;\n            4L\nlet (var_0: Union0) = Union0Case0\nlet (var_1: Union0) = Union0Case0\nlet (var_2: Union0) = Union0Case0\nmatch var_0 with\n| Union0Case0 -&gt;\n    match var_1 with\n    | Union0Case0 -&gt;\n        1L\n    | Union0Case1 -&gt;\n        method_0((var_0: Union0), (var_1: Union0), (var_2: Union0))\n| Union0Case1 -&gt;\n    method_0((var_0: Union0), (var_1: Union0), (var_2: Union0))\n<\/pre>\n<p>What is going on here is that the evaluator is forgetting that it already tested the variables and starts unboxing them again inside the join point. This is because join points throw away local left to right rewrite information.<\/p>\n<p>A workaround would be to put join points in the clause bodies.<\/p>\n<pre>inl ab = box (.A \\\/ .B)\ninl x = dyn (ab .A, ab .A, ab .A)\ninl pat arg on_fail on_succ = on_fail ()\nmatch x with\n| .A, .A, _ -&gt; join 1\n| @pat _ -&gt; join -1\n| .A, .B, .B -&gt; join 2\n| _, _, .A -&gt; join 3\n| _ -&gt; join 4\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\nlet rec method_0(): int64 =\n    1L\nand method_1(): int64 =\n    3L\nand method_2(): int64 =\n    2L\nand method_3(): int64 =\n    4L\nlet (var_0: Union0) = Union0Case0\nlet (var_1: Union0) = Union0Case0\nlet (var_2: Union0) = Union0Case0\nmatch var_0 with\n| Union0Case0 -&gt;\n    match var_1 with\n    | Union0Case0 -&gt;\n        method_0()\n    | Union0Case1 -&gt;\n        match var_2 with\n        | Union0Case0 -&gt;\n            method_1()\n        | Union0Case1 -&gt;\n            method_2()\n| Union0Case1 -&gt;\n    match var_2 with\n    | Union0Case0 -&gt;\n        method_1()\n    | Union0Case1 -&gt;\n        method_3()\n<\/pre>\n<p>Now the result is what one might want.<\/p>\n<p>In the future this might not be an issue and in fact, it might just get fixed as a natural process of making the compiler run faster. The way patterns work now is inefficient from a compilation speed perspective and there is room for improvement there. In fact, since pattern matching is so ubiquitous in Spiral, that would be the first thing one would want to optimize in order to speed up compilation.<\/p>\n<p>The way they work now though has the great combination of them being elegant, highly effective and simple to implement.<\/p>\n<p>It is difficult to imagine what could be added to Spiral for it to generate better code for runtime at this point. Spiral's one pass is THE optimization pass to optimize these patterns at runtime, but it does not have any capacity to optimize its own compilation.<\/p>\n<p>Right now Spiral's worst problem is its poor library support. The libraries are always in text and have to be parsed and prepassed from the scratch on every compilation. The way most languages solve that is by inventing an intermediate bytecode format, but without a doubt there exists a language design that would allow both the language to be fused to libraries, and to optimize both itself and the programs it is applied to. Without a doubt, such a language would significantly exceed Spiral in quality.<\/p>\n<p>One thing is for certain, such a language would be hard to write as a standard compiler in F#. A lot more infrastructure support would be necessary in order to support a fundamentally different approach towards compiler construction. Its own platform and a surgical compiler like <a href=\"https:\/\/github.com\/TiarkRompf\/lancet\">Lancet<\/a> would be a prerequisite. MLs are more suited towards writing interpreters than language towers.<\/p>\n<p><a href=\"https:\/\/racket-lang.org\/\" rel=\"nofollow\">Racket<\/a> has a superior ecosystem for writing such a language compared to the .NET, but it is the author's opinion that the parser in particular is not the best place to do compile time evaluation of functions and that syntax should not be used for abstraction - it should be used for ergonomics and should be consistent. Parsing should be a step to get rid of syntax for the rest of the passes.<\/p>\n<p>Macros do not make sense in dynamic languages as a tool for language creation. They are absolute insanity in static languages. Often when static languages reach the limit of their design they cram macros to do everything else - like performance optimizations for example, and tout them as a feature rather than an admission of failure in language design.<\/p>\n<p>Wanting macros in order to optimize performance will never happen in Spiral.<\/p>\n<h3><a id=\"user-content-4-continuation-passing-style-monadic-computation-and-parsing\" class=\"anchor\" aria-hidden=\"true\" href=\"#4-continuation-passing-style-monadic-computation-and-parsing\"><\/a>4: Continuation Passing Style, Monadic Computation and Parsing<\/h3>\n<p>Now that union types are out of the way, slowly the subject can move towards the more fun stuff that can be done with the language. CPS is a great way of writing highly abstract, generic and very fast code in Spiral and so the language has support for programming in such a style using monadic syntax. Modules are a significant aid as well for programming in CPS.<\/p>\n<p>This chapter will be short and won't go into depth of how monads work. Neither will it explain how parsers work. As both of those subjects are highly complex, it would take a lot of time to cover them. For parsers combinators in particular, the place place to learn how they work would be to start with the <a href=\"http:\/\/www.quanttec.com\/fparsec\/\" rel=\"nofollow\">FParsec documentation<\/a>.<\/p>\n<p>For monads in particular, it is best to study their specific instances. There is a large amount of tutorials online regarding them, most often in the context of the Haskell language. More closer to home, the author's understanding of them went through a dramatic improvement once he stopped trying to figure out what the higher kinded types are doing and simply focused on the functions themselves in terms of flow.<\/p>\n<pre>inl on_succ a = (a,())\ninl on_log x = ((),Tuple.singleton x)\ninl (&gt;&gt;=) (a,w) f = \/\/ The writer monad.\n    inl a',w' = f a\n    (a',Tuple.append w w')\ninl add x y = x + y |&gt; on_succ\ninm x = add 1 1\ninm _ = on_log x\ninm y = add 3 4\ninm _ = on_log y\ninm z = add 5 6\ninm _ = on_log z\non_succ (x+y+z) \/\/ Tuple2(20L, Tuple1(2L, 7L, 11L))\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nand Tuple1 =\n    struct\n    val mem_0: int64\n    val mem_1: Tuple0\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nTuple1(20L, Tuple0(2L, 7L, 11L))\n<\/pre>\n<p>What the <code>inm<\/code> keyword does is merely rewrite <code>inm x = f<\/code> to <code>f &gt;&gt;= inl x -&gt; ...<\/code>. Meaning the above example could have been done manually as...<\/p>\n<pre>add 1 1 &gt;&gt;= inl x -&gt;\non_log x &gt;&gt;= inl _ -&gt;\nadd 3 4 &gt;&gt;= inl y -&gt;\non_log y &gt;&gt;= inl _ -&gt;\nadd 5 6 &gt;&gt;= inl z -&gt;\non_log z &gt;&gt;= inl _ -&gt;\non_succ (x+y+z)\n<\/pre>\n<p>The writer monad pattern is notable as it is used to accumulate the backwards trace in the ML library, so it is worth keeping in mind. Despite the power of monads, most of the time they are used to encapsulate state.<\/p>\n<p>Apart from <code>inm<\/code> there is also <code>inb<\/code>.<\/p>\n<pre>inl f x ret =\n    Console.writeline x\n    ret()\n    Console.writeline \"done\"\ninb x = f \"hello\"\nConsole.writeline \"doing work\"\n<\/pre>\n<pre>System.Console.WriteLine(\"hello\")\nSystem.Console.WriteLine(\"doing work\")\nSystem.Console.WriteLine(\"done\")\n<\/pre>\n<p>A pattern similar to the above is used to emulate stack allocation of Cuda memory in the ML library.<\/p>\n<p>Anyway, here is a simple Spiral parser library written in CPS style just for the sake of making a benchmark. A more advanced version in monadic style can be found in the standard library.<\/p>\n<h4><a id=\"user-content-parsing-benchmark\" class=\"anchor\" aria-hidden=\"true\" href=\"#parsing-benchmark\"><\/a>Parsing Benchmark<\/h4>\n<pre>let example = \n    \"example\",[option;tuple;loops;extern_;console],\"Module description.\",\n    \"\"\"\n\/\/\/ Converts a string to a parser stream.\n\/\/\/ string -&gt; parser_stream\ninl string_stream str {idx on_succ on_fail} =\n    inl f idx = idx &gt;= 0 &amp;&amp; idx &lt; string_length str\n    inl branch cond = if cond then on_succ (str idx) else on_fail \"string index out of bounds\" \n    match idx with\n    | a, b -&gt; branch (f a &amp;&amp; f b)\n    | _ -&gt; branch (f idx)\n\/\/\/ Runs a parser given the string and the expected return type.\n\/\/\/ t type -&gt; string -&gt; t parser -&gt; t\ninl run ret_type str parser = \n    inl stream = dyn str |&gt; string_stream\n    inl d = {\n        stream\n        on_succ = inl x state -&gt; id x\n        on_fail = inl x state -&gt; failwith ret_type x\n        on_type = ret_type\n        }\n    parser d {pos=dyn 0}\n\/\/\/ Reads a char.\n\/\/\/ char parser\ninl stream_char {stream on_succ on_fail} {state with pos} =\n    stream {\n        idx = pos\n        on_succ = inl c -&gt; on_succ c {state with pos=pos+1}\n        on_fail = inl msg -&gt; on_fail msg state\n        }\ninl is_digit x = x &gt;= '0' &amp;&amp; x &lt;= '9'\ninl is_whitespace x = x = ' '\ninl is_newline x = x = '\\n' || x = '\\r'\n\/\/\/ Reads a digit.\n\/\/\/ char parser\ninl digit {stream on_succ on_fail} state =\n    stream_char {\n        stream on_fail\n        on_succ = inl x state' -&gt; \n            if is_digit x then on_succ x state'\n            else on_fail \"digit\" state\n        } state\ninl convert_type = fs [text: \"System.Convert\"]\ninl to_uint64 x = Extern.FS.StaticMethod convert_type .ToUInt64 x uint64\n\/\/\/ Reads an 64-bit integer parser.\n\/\/\/ uint64 parser\ninl puint64 {stream on_succ on_fail on_type} state =\n    inl error state = on_fail \"puint64\" state\n    inl rec loop i on_fail state =\n        digit {\n            stream\n            on_fail=inl _ state -&gt; on_fail i state\n            on_succ=inl c state -&gt;\n                inl max = 1844674407370955161u64 \/\/ max of uint64 \/ 10u64\n                if i &lt;= max then\n                    inl i' = i * 10u64 + to_uint64 c - to_uint64 '0'\n                    if i &lt; i' then join loop i' on_succ state\n                    else error state\n                else error state\n            } state\n        : on_type\n    loop (dyn 0u64) (inl _ state -&gt; error state) state\n\/\/\/ The skips an all the whitespaces (including newlines) before succeeding.\n\/\/\/ unit parser\nmet rec spaces {d with stream on_succ on_fail on_type} state =\n    stream_char {\n        stream\n        on_fail = inl _ state -&gt; on_succ () state\n        on_succ = inl c state' -&gt; \n            if is_whitespace c || is_newline c then spaces d state'\n            else on_succ () state\n        } state\n    : on_type\n\/\/\/ Runs the first and then the second parser before returning the result of the second parser.\n\/\/\/ a parser -&gt; b parser -&gt; b parser\ninl (&gt;&gt;.) a b {d with on_succ} state = a {d with on_succ = inl _ state -&gt; b d state} state\n\/\/\/ Runs the first and then the second parser before returning the result of the first parser.\n\/\/\/ a parser -&gt; b parser -&gt; a parser\ninl (.&gt;&gt;) a b {d with on_succ} state = \n    a {d with on_succ = inl a state -&gt; \n        b {d with on_succ = inl _ state -&gt; on_succ a state} state\n        } state\n\/\/\/ Runs the tuple of parsers before returning their result.\n\/\/\/ tuple parser\ninl rec tuple l {d with on_succ} state =\n    match l with\n    | x :: xs -&gt;\n        x {d with on_succ = inl x state -&gt;\n            tuple xs {d with on_succ = inl xs state -&gt;\n                on_succ (x :: xs) state\n                } state\n            } state\n    | () -&gt; on_succ () state\n\/\/\/ Parses an unsigned 64-bit integer and returns it after parsing whitespaces.\n\/\/\/ uint64 parser\ninl num = puint64 .&gt;&gt; spaces\nrun (uint64,uint64,uint64) \"123 456 789\" (tuple (num, num, num))\n    \"\"\"\n<\/pre>\n<p>As the output is 376 lines long, it won't be pasted. Here is a straightforward translation of the above to F#.<\/p>\n<pre>let example2 = \n    \/\/\/ Converts a string to a parser stream.\n    \/\/\/ string -&gt; parser_stream\n    \/\/\/ Note: The functionality of this functions needs to be pared down in F# due to lack of intensional polymorphism.\n    let string_stream (str: string) (idx, on_succ, on_fail) =\n        if idx &gt;= 0 &amp;&amp; idx &lt; str.Length then on_succ str.[idx] else on_fail \"string index out of bounds\" \n    \/\/\/ Runs a parser given the string and the expected return type.\n    \/\/\/ string -&gt; t parser -&gt; t\n    let run str parser = \n        let stream = string_stream str\n        let d = \n            stream\n            ,fun x state -&gt; id x\n            ,fun x state -&gt; failwith x\n        parser d 0\n    \/\/\/ Reads a char.\n    \/\/\/ char parser\n    let stream_char (stream, on_succ, on_fail) pos =\n        stream \n            (pos\n            ,fun c -&gt; on_succ c (pos+1)\n            ,fun msg -&gt; on_fail msg pos\n            )\n    let is_digit x = x &gt;= '0' &amp;&amp; x &lt;= '9'\n    let is_whitespace x = x = ' '\n    let is_newline x = x = '\\n' || x = '\\r'\n    \/\/\/ Reads a digit.\n    \/\/\/ char parser\n    let digit (stream, on_succ, on_fail) state =\n        stream_char ( \n            stream \n            ,fun x state' -&gt; \n                if is_digit x then on_succ x state'\n                else on_fail \"digit\" state\n            ,on_fail\n            ) state\n            \n    \/\/\/ Reads an 64-bit integer parser.\n    \/\/\/ uint64 parser\n    let puint64 (stream, on_succ, on_fail) state =\n        let error state = on_fail \"puint64\" state\n        let rec loop i on_fail state =\n            digit (\n                stream\n                ,fun c state -&gt;\n                    let max = 1844674407370955161UL \/\/ max of uint64 \/ 10u64\n                    if i &lt;= max then\n                        let i' = i * 10UL + System.Convert.ToUInt64 c - System.Convert.ToUInt64 '0'\n                        if i &lt; i' then loop i' on_succ state\n                        else error state\n                    else error state\n                ,fun _ state -&gt; on_fail i state\n                ) state\n        loop 0UL (fun _ state -&gt; error state) state\n    \/\/\/ The skips an all the whitespaces (including newlines) before succeeding.\n    \/\/\/ unit parser\n    let rec spaces (stream, on_succ, on_fail as d) state =\n        stream_char (\n            stream\n            ,fun c state' -&gt; \n                if is_whitespace c || is_newline c then spaces d state'\n                else on_succ () state\n            ,fun _ state -&gt; on_succ () state\n            ) state\n    \/\/\/ Runs the first and then the second parser before returning the result of the second parser.\n    \/\/\/ a parser -&gt; b parser -&gt; b parser\n    let (&gt;&gt;.) a b (stream,on_succ,on_fail as d) state = a (stream,(fun _ state -&gt; b d state), on_fail) state\n    \/\/\/ Runs the first and then the second parser before returning the result of the first parser.\n    \/\/\/ a parser -&gt; b parser -&gt; a parser\n    let (.&gt;&gt;) a b (stream,on_succ,on_fail) state = \n        a (stream, (fun a state -&gt; b (stream, (fun _ state -&gt; on_succ a state),on_fail) state), on_fail) state\n    \/\/\/ Runs the tuple of parsers before returning their result.\n    \/\/\/ tuple parser\n    \/\/\/ Note: This one is ugly. It is impossible to make a generic tuple without great type hackery in F#. \n    \/\/\/ Check out FParsec.Pipes library to see how that might be done.\n    let tuple3 (a,b,c) (stream,on_succ,on_fail) =\n        a (\n            stream \n            ,fun a -&gt;\n                b (\n                    stream\n                    ,fun b -&gt;\n                        c (\n                            stream\n                            ,fun c -&gt;\n                                on_succ (a,b,c)\n                            ,on_fail)\n                    ,on_fail)\n            ,on_fail)\n    \/\/\/ Parses an unsigned 64-bit integer and returns it after parsing whitespaces.\n    \/\/\/ uint64 parser\n    let num = puint64 .&gt;&gt; spaces\n    run \"123 456 789\" (tuple3 (num, num, num))\n<\/pre>\n<p>Here are the performance figures from testing the above two programs using the <a href=\"https:\/\/github.com\/dotnet\/BenchmarkDotNet\">BenchmarkDotNet<\/a> library . What the above program is doing is merely parsing 3 unsigned 64-bit ints and returning them in a tuple.<\/p>\n<pre>           Method |       Mean |     Error |    StdDev |\n----------------- |-----------:|----------:|----------:|\n FullySpecialized |   292.8 ns | 0.8448 ns | 0.7902 ns |\n           FSharp | 3,616.2 ns | 8.9547 ns | 8.3762 ns |\n<\/pre>\n<p>The Spiral parser is about 12x times faster. That is quite a nice improvement and roughly what one could expect. The interesting thing not noted in the benchmark is that once the Spiral's output has been compiled to F# code, it runs instantaneously while in F#'s case, there is a noticeable delay while the .NET JIT tries to optimize it. Obviously, in Spiral's case the JIT does not have much work left for it. Apart from register allocation, Spiral already does everything else and does a better job if it finishes compiling.<\/p>\n<p>If anything, the above understates the advantage that Spiral has over F#. Given how poor F# is at optimizing <a href=\"https:\/\/docs.microsoft.com\/en-us\/dotnet\/fsharp\/language-reference\/computation-expressions\" rel=\"nofollow\">monads<\/a>, for parsers written in a monadic style the author would not be surprised to see the gap widen by another 10x.<\/p>\n<p>Nevertheless, one point in F#'s favor are compile times. The parser being tested here is somewhat trivial, but for a more complex parser such as the one for the Spiral compiler, the author would not be surprised to see it go up into 100s of thousands of lines of code. Given that the IDE gets crushed by the weight of a 20k LOC parser, it has him wondering how would he even compile such a monster? As a rule of the thumb, Spiral's evaluator generates around 3k LOC per second currently.<\/p>\n<p>For that sake, rather than CPSing everything it would be important to do boxing. In the previous section on lists, it was shown how in Spiral, the CPS form of <code>head<\/code>,<code>tail<\/code> and <code>last<\/code> is just a more generic form of the one that uses the option type. Note that this is not the case in other functional languages as they lack Spiral's inlining guarantees.<\/p>\n<p>There are two choices for doing boxing in Spiral and they are good to know in order to cut down on exceesive specialization and making sure the evaluator does not diverge.<\/p>\n<ol><li>Do term casting.<\/li>\n<\/ol><pre>\/\/\/ Term casting for parsers\n\/\/\/ a type -&gt; a parser -&gt; a parser\ninl term_cast typ p {d with on_succ on_fail on_type} state =\n    inl term_cast_uncurried g a b = \/\/ This is to make sure only one closure is allocated.\n        inl g = term_cast (inl a, b -&gt; g a b) (a,b)\n        inl a b -&gt; g (a,b)\n    p {d with \n        on_succ = term_cast_uncurried on_succ typ state\n        on_fail = term_cast_uncurried on_fail string state\n        } state\ninl puint64 d state = join puint64 d state \/\/ Make sure that the unrolled outer loop is rolled in.\n\/\/\/ Parses an unsigned 64-bit integer and returns it after parsing whitespaces.\n\/\/\/ uint64 parser\ninl num = term_cast uint64 (puint64 .&gt;&gt; spaces)\nrun (uint64,uint64,uint64) \"123 456 789\" (tuple (num, num, num))\n<\/pre>\n<p>The amount of code generated drops down to 141 LOC with this. Unfortunately it does make the program slower by about 40%.<\/p>\n<pre>           Method |       Mean |     Error |    StdDev |\n----------------- |-----------:|----------:|----------:|\n       TermCasted |   406.6 ns | 1.2316 ns | 1.1520 ns |\n FullySpecialized |   292.8 ns | 0.8448 ns | 0.7902 ns |\n           FSharp | 3,616.2 ns | 8.9547 ns | 8.3762 ns |\n<\/pre>\n<ol start=\"2\"><li>Box using union types.<\/li>\n<\/ol><pre>inl ParserResult x state = .ParserSucc, x, state \\\/ .ParserFail, string, state\n\/\/\/ Union boxing for parsers\n\/\/\/ a type -&gt; a parser -&gt; a parser\ninl box typ p {d with on_succ on_fail} state = \n    inl on_type = ParserResult typ state\n    inl box = box on_type\n    p {d with\n        on_succ = inl x state -&gt; box (.ParserSucc, x, state)\n        on_fail = inl x state -&gt; box (.ParserFail, x, state)\n        on_type\n        } state\n    |&gt; function\n        | .ParserSucc, x, state -&gt; on_succ x state\n        | .ParserFail, x, state -&gt; on_fail x state\ninl puint64 d state = join puint64 d state \/\/ Make sure that the unrolled outer loop is rolled in.\n\/\/\/ Parses an unsigned 64-bit integer and returns it after parsing whitespaces.\n\/\/\/ uint64 parser\ninl num = box uint64 (puint64 .&gt;&gt; spaces)\nrun (uint64,uint64,uint64) \"123 456 789\" (tuple (num, num, num))\n<\/pre>\n<p>This actually improves the running time significantly.<\/p>\n<pre>           Method |       Mean |     Error |    StdDev |\n----------------- |-----------:|----------:|----------:|\n       TermCasted |   406.6 ns | 1.2316 ns | 1.1520 ns |\n             Boxy |   199.4 ns | 0.9976 ns | 0.9332 ns |\n FullySpecialized |   292.8 ns | 0.8448 ns | 0.7902 ns |\n           FSharp | 3,616.2 ns | 8.9547 ns | 8.3762 ns |\n<\/pre>\n<p>The boxy parser is by far the best variant. It comes out to only 170 LOCs and is 45% faster than the fully specialized version and 18x times than the F# version. This also demonstrates that code duplication does have a noticeable performance impact. The lines of code reduced would be much more dramatic for a larger parser thereby making the application of this technique a necessity in a serious library. Currently the <code>Parsing<\/code> module in the standard library is lacking in that regard and the above benchmark is actually the first time the author used union type boxing on parsers or did benchmarking of a Spiral program. The <code>Parsing<\/code> module was not intented to be a serious parsing library, but to drive the development of the language in a challening area.<\/p>\n<p>When the first version of it was created Spiral did not have modules nor monadic syntax nor good error messages, but it did have a lot of compiler bugs as if to make up for it. Spiral's modules were created in part because refactoring the parser was simply so painful during those days - it would take the author hours to fix the type errors that in F# would have taken him 10m. It is much better now thankfully.<\/p>\n<p>As the author has no intention of doing so and wants to do machine learning instead, for those interested in parsing Spiral is a very good language to experiment with in the context of staged functional programming.<\/p>\n<p>It will no doubt be a very productive language for such a purpose depending on how much weight is put on performance. If full weight is put on it then there is no doubt that Spiral would be orders of magnitute more productive at such a task than other languages.<\/p>\n<p>The reason for that is that it is not enough to judge merely by how long would it take to write a parser, but how long would it take to get it on par with Spiral in performance. It took the author ~5h to make the parser for this benchmark in Spiral and then maybe 20m to transcribe it to F# by hand. In order to get to the Boxy level of performance, how long would it have taken to specialize all of that by hand and test it? Days?<\/p>\n<p>Just how hard would such a fast handwritten C-style parser be to modify after that? It would be completely inflexible, so a decent guess is quite hard. It would also take a special kind of masochism to deliberately write code in such a style.<\/p>\n<p>ML styled languages still have some advantages over Spiral due to having type inference which is a great aid to refactoring and explorability, but C offshots can be completely replaced with no regret.<\/p>\n<p>The 4 parsers benchmarked in this section can be found in <a href=\"https:\/\/github.com\/mrakgr\/The-Spiral-Language\/tree\/master\/Benchmarking\">this folder<\/a> of the repo.<\/p>\n<h3><a id=\"user-content-5-tensors-and-structural-reflection\" class=\"anchor\" aria-hidden=\"true\" href=\"#5-tensors-and-structural-reflection\"><\/a>5: Tensors and Structural Reflection<\/h3>\n<p>The development of Spiral was driven by the need for a language with great capability for abstraction whose semantics would allow for it to be compiled to very fast code suitable for GPUs and the architectures coming down the line. During the early days of its development when it was intended as a Cuda backend for the ML library Spiral actually had built in arrays that would track variables at on the type level, but that tensors could be designed like the way they currently could be was beyond the imagination of its author and makes him glad that he decided to complete the language instead of leaving Spiral in a half finished state as a crappy ML library backend.<\/p>\n<p>Tensors in Spiral represent the crystallization of its power; they are the point at which all of its features flow together to create something that cannot be done in any other language.<\/p>\n<p>Unlike parsers of the previous chapter which were complicated - so complicated that they could not be explained, tensors are actually rather simple and intended for all ages.<\/p>\n<p>The implementation of <code>HostTensor<\/code> in the standard library is somewhat convoluted due to needing to be generic in order to be reusable on the Cuda side in addition to having a lot of functionality, so this chapter will follow the format of giving a few examples of its most salient features and then show how a simpler tensor can be derived from first principles in order to attain understanding of it.<\/p>\n<p>Tensors in Spiral can have an arbitrary number of dimensions, arbitrary types and arbitrary layouts in addition to supporting views. Indexing into them emulates the partial application of functions. <code>init<\/code> for them takes arguments in curried form which supports scope control.<\/p>\n<p><em>Note: The HostTensor module was greatly redesigned since this was last written so the examples won't compile anymore. This tutorial will be updated at some point to reflect this. The tensor does not longer support ranges, but instead it has been split into <code>Tensor<\/code> and <code>View<\/code>. <code>View<\/code> in particular supports both range views as shown here and named tree-based views which make certain optimization such as for the LSTM layer a lot safer and easier to implement.<\/em><\/p>\n<pre>let example = \n    \"example\",[option;tuple;loops;extern_;console;host_tensor],\"Module description.\",\n    \"\"\"\ninl ar = array_create string 3\nTuple.foldl (inl i x -&gt; ar i &lt;- x; i+1) 0 (\"zero\",\"one\",\"two\") |&gt; ignore\nHostTensor.init (3,5,{from=2; to=5}) (inl a -&gt;\n    inl x = ar a\n    string_format \"x is {0}\" x |&gt; Console.writeline\n    inl b c -&gt; x, b, c\n    )\n|&gt; HostTensor.show.all\n|&gt; Console.writeline\n    \"\"\"\n<\/pre>\n<p>The generated code won't be posted as the loops for printing the tensors and initializing it are long, but here is what comes out when the program has been ran.<\/p>\n<pre>x is zero\nx is one\nx is two\n[|\n    [|\n        [|[zero, 0, 2]; [zero, 0, 3]; [zero, 0, 4]; [zero, 0, 5]|]\n        [|[zero, 1, 2]; [zero, 1, 3]; [zero, 1, 4]; [zero, 1, 5]|]\n        [|[zero, 2, 2]; [zero, 2, 3]; [zero, 2, 4]; [zero, 2, 5]|]\n        [|[zero, 3, 2]; [zero, 3, 3]; [zero, 3, 4]; [zero, 3, 5]|]\n        [|[zero, 4, 2]; [zero, 4, 3]; [zero, 4, 4]; [zero, 4, 5]|]\n    |]\n    [|\n        [|[one, 0, 2]; [one, 0, 3]; [one, 0, 4]; [one, 0, 5]|]\n        [|[one, 1, 2]; [one, 1, 3]; [one, 1, 4]; [one, 1, 5]|]\n        [|[one, 2, 2]; [one, 2, 3]; [one, 2, 4]; [one, 2, 5]|]\n        [|[one, 3, 2]; [one, 3, 3]; [one, 3, 4]; [one, 3, 5]|]\n        [|[one, 4, 2]; [one, 4, 3]; [one, 4, 4]; [one, 4, 5]|]\n    |]\n    [|\n        [|[two, 0, 2]; [two, 0, 3]; [two, 0, 4]; [two, 0, 5]|]\n        [|[two, 1, 2]; [two, 1, 3]; [two, 1, 4]; [two, 1, 5]|]\n        [|[two, 2, 2]; [two, 2, 3]; [two, 2, 4]; [two, 2, 5]|]\n        [|[two, 3, 2]; [two, 3, 3]; [two, 3, 4]; [two, 3, 5]|]\n        [|[two, 4, 2]; [two, 4, 3]; [two, 4, 4]; [two, 4, 5]|]\n    |]\n|]\n<\/pre>\n<p>Having the arguments to <code>init<\/code> be partially applied rather than given all at once is what allow the outside array to be indexed only in the outer loop. Had it been otherwise, the indexing would have needed to be done inside the innermost loop. <code>init<\/code> gives loops for free. Important operations such as map, rotation and reduction can be implemented in terms of init on the host (CPU) side.<\/p>\n<p>Tensors themselves emulate partial application of functions. Here is how they can be indexed into.<\/p>\n<pre>inl ar = array_create string 3\nTuple.foldl (inl i x -&gt; ar i &lt;- x; i+1) 0 (\"zero\",\"one\",\"two\") |&gt; ignore\ninl tns = HostTensor.init (3,5,{from=2; to=5}) (inl a -&gt;\n    inl x = ar a\n    inl b c -&gt; x, b, c\n    )\ninl f x = x 0 2 .get\ntns 0 |&gt; f |&gt; Console.writeline\ntns 1 |&gt; f |&gt; Console.writeline\ntns 2 |&gt; f |&gt; Console.writeline\n<\/pre>\n<pre>[zero, 0, 2]\n[one, 0, 2]\n[two, 0, 2]\n<\/pre>\n<p>This is convenient for views. Views can take more than a single argument in which case they need to be passed as a tuple. Like application, views work on dimensions from left to right - from the outermost to the innermost.<\/p>\n<pre>inl ar = array_create string 3\nTuple.foldl (inl i x -&gt; ar i &lt;- x; i+1) 0 (\"zero\",\"one\",\"two\") |&gt; ignore\ninl tns = HostTensor.init (3,5,{from=2; to=5}) (inl a -&gt;\n    inl x = ar a\n    inl b c -&gt; x, b, c\n    )\ntns.view (1,{from=2;near_to=4})\n|&gt; HostTensor.show.all\n|&gt; Console.writeline\n<\/pre>\n<pre>[|\n    [|\n        [|[zero, 2, 2]; [zero, 2, 3]; [zero, 2, 4]; [zero, 2, 5]|]\n        [|[zero, 3, 2]; [zero, 3, 3]; [zero, 3, 4]; [zero, 3, 5]|]\n    |]\n|]\n<\/pre>\n<p>Both the tensor applications and views are done done immutably. Apart from join points which are memoized, Spiral's metalanguage is pure and there is no way of mutably updating tuples or modules.<\/p>\n<p>For the sake of demonstration, here is how <code>.set<\/code> works. It is very similar to <code>.get<\/code>.<\/p>\n<pre>inl tns = HostTensor.init ({from=2; near_to=5}) id\ninl modify f t = t .set (t .get |&gt; f)\ntns 2 |&gt; modify ((*) 2)\ntns 3 |&gt; modify ((+) 22)\ntns 4 |&gt; modify (const -11)\nHostTensor.show.all tns\n|&gt; Console.writeline\n<\/pre>\n<pre>[|4; 25; -11|]\n<\/pre>\n<p>In addition to <code>.view<\/code>, <code>.view_span<\/code> is useful.<\/p>\n<pre>inl tns_a = HostTensor.init ({from=2; near_to=10}) id\nHostTensor.show.all tns_a\n|&gt; string_format \"tns_a = {0}\"\n|&gt; Console.write\ninl tns_b = tns_a.view_span {from=0;near_to=2}\nHostTensor.show.all tns_b\n|&gt; string_format \"tns_b = {0}\"\n|&gt; Console.write\ninl tns_c = tns_a.view_span {from=2;near_to=4}\nHostTensor.show.all tns_c\n|&gt; string_format \"tns_c = {0}\"\n|&gt; Console.write\ntns_c 0 .get |&gt; Console.writeline\ntns_c 1 .get |&gt; Console.writeline\n\/\/tns_c 2 .get |&gt; Console.writeline \/\/ Would trigger the range check assertion at compile time.\n<\/pre>\n<pre>tns_a = [|2; 3; 4; 5; 6; 7; 8; 9|]\ntns_b = [|2; 3|]\ntns_c = [|4; 5|]\n4\n5\n<\/pre>\n<p><code>.view_span<\/code> is similar to view except starts the indexing from the beginning of the dimensions of the target tensor and rebases the dimensions so they start from 0. That means that <code>.view_span -1<\/code> would always be an out of bounds error as would all negative values of the index. Given how Spiral specializes join points, it is useful for avoiding code bloat as well.<\/p>\n<p>For the sake of machine learning, it is preferable to keep the tensor sizes as constants to get rid of as many assertion at compile time as possible. Note that in the last line, had it not been commented out the compiler would have raised a type error at compile time.<\/p>\n<p>Tensors have even more to offer. By default, their layout is that of tuple of arrays. Meaning a tuple of type <code>float32 * int64 * int64<\/code> would be represented using 3 arrays internally each for the separate elements of the tuple. The main motivation behind this design is to make it easy to pass them through language boundaries onto the Cuda side as unless the arrays were of primitive types, it would be difficult to align them in memory, but there are performance benefits as well to allowing such a representation.<\/p>\n<p>That varies from problem to problem, so it would be even better if it was easy to switch between representations at will.<\/p>\n<pre>inl tns_toa = HostTensor.init (5,5) (inl a b -&gt; a,b)\ninl tns_aot = HostTensor.init.aot (5,5) (inl a b -&gt; tns_toa a b .get)\n()\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\n...\nlet (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(25L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(25L))\nlet (var_2: int64) = 0L\nmethod_0((var_0: (int64 [])), (var_1: (int64 [])), (var_2: int64))\nlet (var_5: (Tuple0 [])) = Array.zeroCreate&lt;Tuple0&gt; (System.Convert.ToInt32(25L))\nlet (var_6: int64) = 0L\nmethod_2((var_0: (int64 [])), (var_1: (int64 [])), (var_5: (Tuple0 [])), (var_6: int64))\n<\/pre>\n<p>The full thing won't be posted as it is 100 LOC, but it should be imaginable that <code>method_0<\/code> is <code>init<\/code> for <code>tns_toa<\/code> and <code>method_2<\/code> is <code>init<\/code> for <code>tns_aot<\/code>. Spiral's tensors are the perfect abstraction where layouts are concerned and it is possible to mix and match <code>.aot<\/code> and <code>.toa<\/code> layout using the <code>zip<\/code> function and it will still work fine. For most usecases, the default tuple of arrays is sensible.<\/p>\n<p>Saying tuple of arrays does not cover it completely though. The tensors work fine with modules.<\/p>\n<pre>inl a = HostTensor.init (3,3) (inl a b -&gt; {a b})\na 2 2 .set {a=99}\nHostTensor.show a\n|&gt; Console.writeline\n<\/pre>\n<pre>[|\n    [|{a = 0; b = 0}; {a = 0; b = 1}; {a = 0; b = 2}|]\n    [|{a = 1; b = 0}; {a = 1; b = 1}; {a = 1; b = 2}|]\n    [|{a = 2; b = 0}; {a = 2; b = 1}; {a = 99; b = 2}|]\n|]\n<\/pre>\n<p>When tensor are in <code>toa<\/code> form they have the added feature of allowing the module fields to be mutated individually. This is not possible in general in Spiral as modules and tuples are immutable; if they were represented as <code>aot<\/code> or if the modules were tuples this would not be possible. What is going on is that when they are represented as separate entities this changes their semantics to reflect that and this is the correct behavior even in a functional language.<\/p>\n<p>This is the short tour of the tensors in Spiral. The next section will be on how they are implemented.<\/p>\n<h4><a id=\"user-content-under-the-hood\" class=\"anchor\" aria-hidden=\"true\" href=\"#under-the-hood\"><\/a>Under the Hood<\/h4>\n<p>As was demonstrated, there are two aspects of tensor polymorphism - one was that they have an arbitrary number of dimensions and the other was that are layout polymorphic. In a language with weaker type systems that would require creating specific tensor instances for both of those concerns, but Spiral can handle them naturally.<\/p>\n<p>Even better, the dimensionality of the tensor is really a separate concern from its layout and so the two subjects can be taken in as separate pieces.<\/p>\n<h5><a id=\"user-content-layout-polymorphism\" class=\"anchor\" aria-hidden=\"true\" href=\"#layout-polymorphism\"><\/a>Layout Polymorphism<\/h5>\n<p>Creating an array of <code>int64 * int64 * int64<\/code> can be done like this in Spiral.<\/p>\n<pre>array_create (int64,int64,int64) 8\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    val mem_2: int64\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nArray.zeroCreate&lt;Tuple0&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>The above is also known as <code>aot<\/code> or <code>array of tuples<\/code> form. The opposite of it, the <code>toa<\/code> or <code>tuple of arrays<\/code> form would be this.<\/p>\n<pre>inl ar = array_create int64 8, array_create int64 8, array_create int64 8\n()\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>Rather than copy pasting like, it would be better if the type were mapped directly.<\/p>\n<pre>inl ar = Tuple.map (inl x -&gt; array_create x 8) (int64,int64,int64)\n()\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>The same output as in the above example shows up. By now, the general principle of how Spiral's tensors achieve their layout polymorphism should be becoming clearer. Of course, the above is woefully incomplete. For example, if the tuple were nested then the <code>toa<\/code> layout would not be achieved.<\/p>\n<pre>inl ar = Tuple.map (inl x -&gt; array_create x 8) (int64,(int64,int64))\n()\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: int64\n    new(arg_mem_0, arg_mem_1) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1}\n    end\nlet (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (Tuple0 [])) = Array.zeroCreate&lt;Tuple0&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>What should be done is to write a function capable mapping over nested tuples.<\/p>\n<pre>inl rec toa_map f = function\n    | x :: x' -&gt; toa_map f x :: toa_map f x'\n    | () -&gt; ()\n    | x -&gt; f x\ninl ar = toa_map (inl x -&gt; array_create x 8) (int64,(int64,int64))\n()\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>Note the subtle difference between <code>toa_map<\/code> and a regular <code>map<\/code>. In <code>| x :: x' -&gt; toa_map f x :: toa_map f x'<\/code> the function recurses on <code>x<\/code> as well, not just on the tail.<\/p>\n<p>For modules, a little extra is needed in <code>toa_map<\/code>.<\/p>\n<pre>inl toa_map f =\n    inl rec loop = function\n        | x :: x' -&gt; loop x :: loop x'\n        | () -&gt; ()\n        | {} as x -&gt; module_map (const loop) x\n        | x -&gt; f x\n    loop\ninl ar = toa_map (inl x -&gt; array_create x 8) {x=int64; y=int64,int64}\n()\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>Before the function is complete there are two more things needed. That would be to stop the function from unboxing every union type and to stop it from recursing on every module. This last one is more of a convenience than necessity.<\/p>\n<pre>inl toa_map f =\n    inl rec loop = function\n        | x when caseable_box_is x -&gt; f x \/\/ This needs to be in the first position to prevent the unboxing from triggering.\n        | x :: x' -&gt; loop x :: loop x'\n        | () -&gt; ()\n        | {!toa_map_block} &amp; x -&gt; module_map (const loop) x\n        | x -&gt; f x\n    loop\ninl ar = toa_map (inl x -&gt; array_create x 8) {x=int64; y=int64,int64; o=Option.Option float32}\n()\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0 of Tuple1\n    | Union0Case1\nand Tuple1 =\n    struct\n    val mem_0: float32\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\nlet (var_0: (Union0 [])) = Array.zeroCreate&lt;Union0&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_3: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\n<\/pre>\n<p>With this as the basis, it is easy to make a <code>toa<\/code> version of array create.<\/p>\n<pre>inl toa_array_create typ size = toa_map (inl x -&gt; array_create x size) typ\n<\/pre>\n<p>Indexing into such an array is quite similar to creating it. It is just a straightforward application of <code>toa_map<\/code>.<\/p>\n<pre>inl toa_index ar idx = toa_map (inl ar -&gt; ar idx) ar\ninl ar = toa_array_create {x=int64; y=int64,int64; o=Option.Option float32} 8\ninl el = toa_index ar 0\n()\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0 of Tuple1\n    | Union0Case1\nand Tuple1 =\n    struct\n    val mem_0: float32\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\nlet (var_0: (Union0 [])) = Array.zeroCreate&lt;Union0&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_3: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_4: Union0) = var_0.[int32 0L]\nlet (var_5: int64) = var_1.[int32 0L]\nlet (var_6: int64) = var_2.[int32 0L]\nlet (var_7: int64) = var_3.[int32 0L]\n<\/pre>\n<p>Setting such an array is a bit more difficult. In order to do that, <code>toa_map2<\/code> would be needed first.<\/p>\n<pre>inl toa_map2 f a b =\n    inl rec loop = function\n        | x, y when caseable_box_is x || caseable_box_is y -&gt; f x y\n        | x :: x', y :: y' -&gt; loop (x,y) :: loop (x',y')\n        | (), () -&gt; ()\n        | (), _ | _, () -&gt; error_type \"Tuple dimensions do not match.\"\n        | {!toa_map_block} &amp; x, {!toa_map_block} &amp; y -&gt; module_map (inl k y -&gt; loop (x k, y)) y\n        | x, y -&gt; f x y\n    loop (a,b)\n<\/pre>\n<p>With this it is possible to implement <code>toa_set<\/code>.<\/p>\n<pre>inl toa_set ar idx v = toa_map2 (inl ar v -&gt; ar idx &lt;- v) ar v |&gt; ignore\ninl ar = toa_array_create {x=int64; y=int64,int64; o=Option.Option float32} 8\ntoa_set ar 0 {x=2; y=1,1; o=Option.some 2.2f32}\n()\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0 of Tuple1\n    | Union0Case1\nand Tuple1 =\n    struct\n    val mem_0: float32\n    new(arg_mem_0) = {mem_0 = arg_mem_0}\n    end\nlet (var_0: (Union0 [])) = Array.zeroCreate&lt;Union0&gt; (System.Convert.ToInt32(8L))\nlet (var_1: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_2: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nlet (var_3: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(8L))\nvar_0.[int32 0L] &lt;- (Union0Case0(Tuple1(2.200000f)))\nvar_1.[int32 0L] &lt;- 2L\nvar_2.[int32 0L] &lt;- 1L\nvar_3.[int32 0L] &lt;- 1L\n<\/pre>\n<p>There is an interesting design decision of whether to allow partial setting of fields in modules in <code>toa_map2<\/code>.<\/p>\n<pre>| {!toa_map_block} &amp; x, {!toa_map_block} &amp; y -&gt; module_map (inl k y -&gt; loop (x k, y)) y\n<\/pre>\n<p>The above particular line could have also been written like so.<\/p>\n<pre>| {!toa_map_block} &amp; x, {!toa_map_block} &amp; y -&gt; module_map (inl k x -&gt; loop (x, y k)) x\n<\/pre>\n<p>Doing it like that would have disallowed the partial module sets as the map would have gone through every field of the <code>toa<\/code> array instead of every field of the setter.<\/p>\n<p>In this section the way to make layout polymorphic arrays was described and is how Spiral's tensors attain such a capability. The ability to reflect on the structure of everything in the language at no runtime cost is a powerful one and solves a lot of the issues currently facing persons writing numerical libraries. It also covers an aspect of polymorphism that parametric languages do not touch.<\/p>\n<p>The next section will be on how to take what has been done so far and make the 1d <code>toa<\/code> array, a tensor capable of support arbitrary dimensions.<\/p>\n<h5><a id=\"user-content-dimensionality-polymorphism\" class=\"anchor\" aria-hidden=\"true\" href=\"#dimensionality-polymorphism\"><\/a>Dimensionality Polymorphism<\/h5>\n<p>In one of the previous chapters, it was mentioned that Spiral is not necessarily less typesafe than F#, and that in some cases the opposite is in fact the case. Whenever dynamism and therefore union types are needed, a step is made into an entirely different paradigm for languages with weaker type systems due to have to push typechecking to runtime.<\/p>\n<pre>List.take 5 [1;2;3] \/\/ An error at runtime.\n<\/pre>\n<pre>Tuple.take 5 (1,2,3) \/\/ An error at compile time.\n<\/pre>\n<p>This comparison might seem unfair since lists in F# are explicitly runtime constructs, but that is in fact the point. Spiral's parent language as it stands now does not have any substitute for Spiral's tuples at all and has absolutely no way of making tensors arbitrary in their dimensions at compile time. All it can do is support very specific instances of them.<\/p>\n<p>So back when the author was making the ML library in F#, he had <code>d2M<\/code> for a 2D GPU tensor, and <code>d4M<\/code> for 4d GPU tensors. Afterwards he found the need for a 3d tensor so he made <code>d3M<\/code> too and so on. There was the <code>Df<\/code>, a lazy scalar host tensor as well.<\/p>\n<p>Layout polymorphism? Forget that. The best what was possible was having them be generic in their type.<\/p>\n<p>Now, there is no doubt that making a very specific instance of a tensor (such as <code>d2M<\/code>) is easier than making a fully blown generic tensor, but making such a tensor type is definitely easier than having to make a specific instance for all the endless varieties of tensors. Making specific instances of the more generic type by hand gets tiresome really quickly. It is humiliating to have to use personal effort because the tool one is using is not good enough.<\/p>\n<h6><a id=\"user-content-design-of-the-tensor\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-of-the-tensor\"><\/a>Design of the Tensor<\/h6>\n<pre>{\n    bodies = { ar size offset toa_map_block } structure\n    dim\n}\n<\/pre>\n<p>The above is very similar to the type of the Spiral's <code>HostTensor<\/code> in pseudo-code. It is actually one of its previous designs that is easier to explain as it has a more uniform structure.<\/p>\n<p>In it, <code>dim<\/code> is just the dimension of the tensor and might be <code>(2,3,4)<\/code> for a 3d tensor, <code>(10,20)<\/code> for a 2d tensor or <code>()<\/code> for a scalar tensor for example.<\/p>\n<p><code>size<\/code> and <code>offset<\/code> inside the <code>bodies<\/code> are directly related to it. Elements of <code>offset<\/code> are always to be multiples of their related <code>size<\/code> element.<\/p>\n<p>Before the coding can start, some simple examples need to be gone though by hand so that it becomes clear what is trying to be done in the first place.<\/p>\n<ol><li>On tensor creation.<\/li>\n<\/ol><pre>\/\/\/ Creating a 1d tensor of type int64 and dim 10\ninl dim = 10\ninl ar = array_create int64 dim\ninl tns = {\n    bodies = {\n        ar\n        size = Tuple.singleton 1\n        offset = Tuple.singleton 0\n        toa_map_block = ()\n        }\n    dim = Tuple.singleton dim\n    }\n<\/pre>\n<pre>\/\/\/ Creating a 2d tensor of type int64 and dim 10, 10\ninl dim = 10, 10\ninl ar = array_create int64 (10 * 10)\ninl tns = {\n    bodies = {\n        ar\n        size = 10, 1\n        offset = 0, 0\n        toa_map_block = ()\n        }\n    dim\n    }\n<\/pre>\n<pre>\/\/\/ Creating a 3d tensor of type int64 and dim 10, 5, 5\ninl dim = 10, 5, 5\ninl ar = array_create int64 (10 * 5 * 5)\ninl tns = {\n    bodies = {\n        ar\n        size = 5*5, 5, 1\n        offset = 0, 0, 0\n        toa_map_block = ()\n        }\n    dim\n    }\n<\/pre>\n<p>By now some patterns should be coming out. <code>ar<\/code> is always inserted directly into the tensor body, <code>size<\/code> is just the rightwards <a href=\"https:\/\/stackoverflow.com\/questions\/23491216\/f-cumulative-product-of-an-array\" rel=\"nofollow\">scan product<\/a> of <code>dim<\/code>, <code>offset<\/code> is always <code>dim<\/code> mapped to 0, and <code>dim<\/code> is always itself. The only difference is 1d when <code>size<\/code>, <code>offset<\/code> and <code>dim<\/code> are wrapped in a tuple.<\/p>\n<pre>\/\/\/ Creating a nd tensor in the array of tuples layout.\ninl tensor_aot_create typ dim =\n    inl dim = match dim with _ :: _ -&gt; dim | x -&gt; x :: ()\n    inl array_size :: size = Tuple.scanr (*) dim 1\n    inl offset = Tuple.map (const 0) dim\n    {\n    bodies = {\n        ar = array_create typ array_size\n        size offset\n        toa_map_block=()\n        }\n    dim\n    }\n<\/pre>\n<p>Here is the pattern abstracted and codified. The tuple of arrays version is similar to the above.<\/p>\n<pre>\/\/\/ Creating a nd tensor in the tuple of arrays layout.\ninl tensor_toa_create typ dim =\n    inl dim = match dim with _ :: _ -&gt; dim | x -&gt; x :: ()\n    inl array_size :: size = Tuple.scanr (*) dim 1\n    inl offset = Tuple.map (const 0) dim\n    inl make_body typ = {\n        toa_map_block=()\n        ar = array_create typ array_size\n        size offset\n        }\n    {\n    bodies = toa_map make_body typ            \n    dim\n    }\n<\/pre>\n<p>The two functions have a very similar internal structure. It can be factored out like so.<\/p>\n<pre>\/\/\/ Creating a nd tensor\ninl tensor_create {dsc with typ dim} =\n    inl dim = match dim with _ :: _ -&gt; dim | x -&gt; x :: ()\n    inl array_size :: size = Tuple.scanr (*) dim 1\n    inl offset = Tuple.map (const 0) dim\n    inl make_body typ = {\n        toa_map_block=()\n        ar = array_create typ array_size\n        size offset\n        }\n    {\n    bodies = \n        match dsc with\n        | {layout=x} -&gt; \n            match x with\n            | .aot -&gt; make_body typ\n            | .toa -&gt; toa_map make_body typ\n        | _ -&gt; toa_map make_body typ\n    dim\n    }\n<\/pre>\n<p>The beautiful thing about this is that since all the sizes are known, Spiral can track them at compile time.<\/p>\n<pre>inl tns = tensor_create {typ=int64,string,float32; dim=10,5,5}\n()\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(250L))\nlet (var_1: (string [])) = Array.zeroCreate&lt;string&gt; (System.Convert.ToInt32(250L))\nlet (var_2: (float32 [])) = Array.zeroCreate&lt;float32&gt; (System.Convert.ToInt32(250L))\n<\/pre>\n<p>Unless the tensor is returned from a join point, or a dynamic if statement, or put into an array, those fields will never be instantiated due being tracked at the type level and the tensor will have the very minimal overhead at runtime.<\/p>\n<p>Spiral's <code>HostTensor<\/code> actually tracks lower and upper bounds in <code>dim<\/code> as well, but that won't be done for this tutorial.<\/p>\n<ol start=\"2\"><li>On application.<\/li>\n<\/ol><pre>{\nbodies = {\n    ar\n    size = 25, 5, 1\n    offset = 0, 0, 0\n    toa_map_block = ()\n    }\ndim = 10, 5, 5\n}\n<\/pre>\n<p>For the purpose of explanation, the 3d tensor from the previous example will be the starting point.<\/p>\n<p>How should applying 2 to the tensor transform it?<\/p>\n<p>It should be into this 2d tensor.<\/p>\n<pre>{\nbodies = {\n    ar\n    size = 5, 1\n    offset = 25*2 + 0, 0 \/\/ 50\n    toa_map_block = ()\n    }\ndim = 5, 5\n}\n<\/pre>\n<p>Applying 3 to the above should turn it into this.<\/p>\n<pre>{\nbodies = {\n    ar\n    size = 1 :: ()\n    offset = 25*2 + 5*3 + 0 :: () \/\/ 65\n    toa_map_block = ()\n    }\ndim = 5 :: ()\n}\n<\/pre>\n<p>Now that it is has been applied twice, the resulting tensor has gone from 3d to 1d. Once more and it will be scalar. Here is simulating the application of 4.<\/p>\n<pre>{\nbodies = {\n    ar\n    size = ()\n    offset = 25*2 + 5*3 + 1*4 \/\/ 69\n    toa_map_block = ()\n    }\ndim = ()\n}\n<\/pre>\n<p>Note how now the offset it a scalar and can be used to index into an array. The rules of tensors are quite simple, only 0d (scalar) tensors can be indexed and set and cannot be applied, while the opposite holds for tensors with higher number of dimensions.<\/p>\n<p>Also note that <code>dim<\/code> plays no role in calculating the top of the new offset.<\/p>\n<p>What was omitted in the above example is the boundary checking. If the value being applied was out of range that would have triggered an error.<\/p>\n<p>Here is the above intuition in code.<\/p>\n<pre>inl tensor_apply i {dim=d::dim bodies} =\n    assert (i &gt;= 0 &amp;&amp; i &lt; d) \"Tensor application out of bounds\"\n    {\n    dim\n    bodies = \n        toa_map (inl {d with size=s::size offset=o::offset} -&gt;\n            inl o = o + i * s\n            inl offset = \n                match offset with\n                | o' :: offset -&gt; o + o' :: offset\n                | () -&gt; o\n            {d with size offset}\n            ) bodies\n    }\n<\/pre>\n<p>The <code>| o' :: offset -&gt; o + o' :: offset<\/code> might be surprising, but that is needed because the offsets might not be zero and that should not be thrown away. That can happen in tensors whose inner dimensions were viewed.<\/p>\n<pre>inl tns = tensor_create {typ=int64,string,float32; dim=10,5,5}\ntensor_apply 2 tns\n|&gt; tensor_apply 3\n|&gt; tensor_apply 4\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(250L))\nlet (var_1: (string [])) = Array.zeroCreate&lt;string&gt; (System.Convert.ToInt32(250L))\nlet (var_2: (float32 [])) = Array.zeroCreate&lt;float32&gt; (System.Convert.ToInt32(250L))\n(Env4(Tuple3((Env0(var_0, 69L)), (Env1(var_1, 69L)), (Env2(var_2, 69L)))))\n<\/pre>\n<p>The types were cut out in the above generated code. As can be seen since scalar tensors have only their offset and an array that is what gets printed.<\/p>\n<pre>inl tns = \n    tensor_create {typ=int64,string,float32; dim=10,5,5}\n    |&gt; tensor_apply (dyn 2)\n    |&gt; tensor_apply 3\n    |&gt; tensor_apply 4\njoin tns\n<\/pre>\n<pre>let rec method_0((var_0: (int64 [])), (var_1: int64), (var_2: (string [])), (var_3: (float32 []))): Env0 =\n    (Env0(Tuple4((Env1(var_0, var_1)), (Env2(var_2, var_1)), (Env3(var_3, var_1)))))\nlet (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(250L))\nlet (var_1: (string [])) = Array.zeroCreate&lt;string&gt; (System.Convert.ToInt32(250L))\nlet (var_2: (float32 [])) = Array.zeroCreate&lt;float32&gt; (System.Convert.ToInt32(250L))\nlet (var_3: int64) = 2L\nlet (var_4: bool) = (var_3 &gt;= 0L)\nlet (var_6: bool) =\n    if var_4 then\n        (var_3 &lt; 10L)\n    else\n        false\nlet (var_7: bool) = (var_6 = false)\nif var_7 then\n    (failwith \"Tensor application out of bounds\")\nelse\n    ()\nlet (var_8: int64) = (var_3 * 25L) \/\/ 50\nlet (var_9: int64) = (var_8 + 15L) \/\/ 65\nlet (var_10: int64) = (var_9 + 4L) \/\/ 69\nmethod_0((var_0: (int64 [])), (var_10: int64), (var_1: (string [])), (var_2: (float32 [])))\n<\/pre>\n<p>In the above example the first application was <code>dyn<\/code>ed. This makes the assertion check necessary at runtime, but the line of note is the last one. Due to common subexpression rewriting, in the end all the bodies end up with the same index variable.<\/p>\n<p>This will be reflected when they are being passed through join points.<\/p>\n<pre>met f (a,b,c) = a+b+c\ninl x = dyn 3\nf (x,x,x)\n<\/pre>\n<pre>let rec method_0((var_0: int64)): int64 =\n    let (var_1: int64) = (var_0 + var_0)\n    (var_1 + var_0)\nlet (var_0: int64) = 3L\nmethod_0((var_0: int64))\n<\/pre>\n<p>The moral of that is - for efficiency tensors should not be returned from anything apart from inlineables and should not be stored into arrays. Spiral's natural style is towards continuation passing and (mostly) functional purity. Tuples and modules should be used to store arguments whenever possible and opaque structures should be avoided.<\/p>\n<ol start=\"3\"><li>On indexing and setting.<\/li>\n<\/ol><p>These two are easy since all the hard work has already been done by <code>tensor_apply<\/code>.<\/p>\n<pre>inl tensor_index {bodies dim=()} = toa_map (inl {ar offset} -&gt; ar offset) bodies\ninl tensor_set {bodies dim=()} v = toa_map2 (inl {ar offset} v -&gt; ar offset &lt;- v) bodies v |&gt; ignore\ninl tns = \n    tensor_create {typ=int64,string,float32; dim=10,5,5}\n    |&gt; tensor_apply 2\n    |&gt; tensor_apply 3\n    |&gt; tensor_apply 4\ntensor_set tns (1,\"asd\",3.3f32)\n<\/pre>\n<pre>let (var_0: (int64 [])) = Array.zeroCreate&lt;int64&gt; (System.Convert.ToInt32(250L))\nlet (var_1: (string [])) = Array.zeroCreate&lt;string&gt; (System.Convert.ToInt32(250L))\nlet (var_2: (float32 [])) = Array.zeroCreate&lt;float32&gt; (System.Convert.ToInt32(250L))\nvar_0.[int32 69L] &lt;- 1L\nvar_1.[int32 69L] &lt;- \"asd\"\nvar_2.[int32 69L] &lt;- 3.300000f\n<\/pre>\n<p>Here is the <code>aot<\/code> form for good measure.<\/p>\n<pre>inl tns = \n    tensor_create {layout=.aot; typ=int64,string,float32; dim=10,5,5}\n    |&gt; tensor_apply 2\n    |&gt; tensor_apply 3\n    |&gt; tensor_apply 4\ntensor_set tns (1,\"asd\",3.3f32)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: string\n    val mem_2: float32\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nlet (var_0: (Tuple0 [])) = Array.zeroCreate&lt;Tuple0&gt; (System.Convert.ToInt32(250L))\nvar_0.[int32 69L] &lt;- Tuple0(1L, \"asd\", 3.300000f)\n<\/pre>\n<h6><a id=\"user-content-the-tensor-facade\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-tensor-facade\"><\/a>The Tensor Facade<\/h6>\n<p>Now that is is possible to create, apply, index and set the tensors the thing that remains is to make it applicable directly. To that, what is needed is to make a facade. The only thing of note in the following is that on standard application the facade rewraps itself. The rest should be straightforward.<\/p>\n<pre>inl rec tensor_facade tns = function\n    | .get -&gt; tensor_get tns\n    | .set v -&gt; tensor_set tns v\n    | .(_) &amp; x -&gt; tns x\n    | x -&gt; tensor_apply x tns |&gt; tensor_facade\ninl tensor_create = tensor_create &gt;&gt; tensor_facade\ninl tns = tensor_create {layout=.aot; typ=int64,string,float32; dim=10,5,5} \ntns 2 3 4 .set (1,\"asd\",3.3f32)\n<\/pre>\n<pre>type Tuple0 =\n    struct\n    val mem_0: int64\n    val mem_1: string\n    val mem_2: float32\n    new(arg_mem_0, arg_mem_1, arg_mem_2) = {mem_0 = arg_mem_0; mem_1 = arg_mem_1; mem_2 = arg_mem_2}\n    end\nlet (var_0: (Tuple0 [])) = Array.zeroCreate&lt;Tuple0&gt; (System.Convert.ToInt32(250L))\nvar_0.[int32 69L] &lt;- Tuple0(1L, \"asd\", 3.300000f)\n<\/pre>\n<h4><a id=\"user-content-closing-comments\" class=\"anchor\" aria-hidden=\"true\" href=\"#closing-comments\"><\/a>Closing Comments<\/h4>\n<p>Views on tensors are similar to apply and just push the offsets without reducing the dimensionality of the tensor. It won't be covered in the tutorial and the interested should just look into the standard library implementation of them for specific details.<\/p>\n<p>There is an interesting programming lesson that the author (re)experienced while making the tensor tutorial. Near the beginning there was a comment that the tensor design for the tutorial is more uniform than the one in the standard library. It is actually even more than that. The way it has been designed here is in fact how the author remembered the tensor. He knew it was not like this in the standard library, but when he looked he was actually surprised at how it was made indicating that he in fact forgot about it.<\/p>\n<p>It is not the first time it happened that a piece poorly fit into memory to him and won't be the last. Memory mismatches are a sure sign that a particular piece of software should be redesigned. Having been left alone for a while, all but the most salient features of a program faded from memory indicating that in fact the rest are worthless and should be removed. Once he is done with the documentation that will surely be done.<\/p>\n<p>With this all that has needed to be said in order to understand tensors has been said, but a digression needs to be made to highlight just how great they really are.<\/p>\n<p>The tensor tutorial should have been rather clear and straightforward and unless something has badly gone wrong, those reading this chapter should have a clear picture of their essence.<\/p>\n<p>The number of languages in which a tensor can be implemented in such a manner can at the time of writing, 2 days to 2018, be literally counted on one finger. Being able to implement tensors like this is what essentially convinced the author that now that he has Spiral, it might be a good idea challenge the other deep learning frameworks for supremacy even though they have big corporate sponsorship.<\/p>\n<p>Even in language like Julia which is made for numeric computation such generic functionality would require heavy use of macros. The kind of metaprogramming Julia allows could be a lot worse.<\/p>\n<p>Take PyTorch for instance, and go to the <a href=\"http:\/\/pytorch.org\/blog\/\" rel=\"nofollow\">tour of its internals<\/a>. It is essentially a tour of poor programming practices: using C macros for everything including making tensors type generic, the absolutely fearsome <code>static PyTypeObject py_FloatType<\/code> which is badly in need of modules or at least SML records, the friction between different components that just jumps out.<\/p>\n<p>From what has been heard of PyTorch in action by the author, there has been nothing but praise.<\/p>\n<p>Nevertheless, if the best tools for the job in 2018 for making tensors generic are the 1972 C macros then probably something went wrong somewhere and not just with PyTorch specifically.<\/p>\n<p>Whether it be composability or performance or lack of safety, those kinds of problems exist due to the weak type systems. But it is not like type systems have to be just about solving constraints, nor do they have to be segregated from the rest of the compiler passes. Lisps had the great idea of integrating parsing with the rest of compilation passes. There is nothing preventing similar to be done with a type system.<\/p>\n<p>Once that fusion is done, a piece of the power that is released can be seen in this chapter - a properly done tensor type.<\/p>\n<h3><a id=\"user-content-6-the-cuda-backend\" class=\"anchor\" aria-hidden=\"true\" href=\"#6-the-cuda-backend\"><\/a>6: The Cuda Backend<\/h3>\n<p>This section covers how the Cuda backend works in its entirety and is not mandatory for understanding how to program in Spiral. It can be skipped over. It does not cover GPU programming, but instead goes into some depth on what is going on under the hood when Spiral does GPU compilation.<\/p>\n<h4><a id=\"user-content-intro-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#intro-1\"><\/a>Intro<\/h4>\n<p>After a successful compilation of a Spiral program, at the top of the file there is a <code>let cuda_kernel = ...<\/code> statement. When no GPU code is run, that results in an empty string, but otherwise there will be code there.<\/p>\n<pre>let kernel1 =\n    \"kernel1\",[cuda_modules],\"Does the map kernel work?\",\n    \"\"\"\n\/\/\/ Initializes all the Cuda parts\ninb s = CudaModules (1024*1024) \/\/ The allocator takes 1Mb of memory from the heap.\n\/\/\/ Creates a host tensor with the given generator function.\ninl h = HostTensor.init 32 (inl x -&gt; x + 1) \n\/\/\/ Loads the tensor on the GPU based on the host tensor\ninl a1 = s.CudaTensor.from_host_tensor h\n\/\/\/ Makes a tensor of the same type and dimensions as `a1` and zeroes it.\ninl o1 = s.CudaTensor.zero_like a1\n\/\/\/ Calls the map operation. `a1` is the input and `o1` is the output.\ns.CudaKernel.map' (inl a _ -&gt; a * 2) a1 o1\n\/\/\/ Transfers the tensor back to host.\ninl a2 = s.CudaTensor.to_host_tensor o1\n\/\/\/ Zips the two tensors and prints them out.\nHostTensor.zip (h,a2) |&gt; HostTensor.show |&gt; Console.writeline\n    \"\"\"\n<\/pre>\n<pre>let cuda_kernels = \"\"\"\n#include \"cub\/cub.cuh\"\nextern \"C\" {\n    __global__ void method_23(long long int * var_0, long long int * var_1);\n    __device__ char method_24(long long int * var_0);\n    \n    __global__ void method_23(long long int * var_0, long long int * var_1) {\n        long long int var_2 = threadIdx.x;\n        long long int var_3 = blockIdx.x;\n        long long int var_4 = (128 * var_3);\n        long long int var_5 = (var_2 + var_4);\n        long long int var_6[1];\n        var_6[0] = var_5;\n        while (method_24(var_6)) {\n            long long int var_8 = var_6[0];\n            char var_9 = (var_8 &gt;= 0);\n            char var_11;\n            if (var_9) {\n                var_11 = (var_8 &lt; 32);\n            } else {\n                var_11 = 0;\n            }\n            char var_12 = (var_11 == 0);\n            if (var_12) {\n                \/\/ \"Argument out of bounds.\"\n            } else {\n            }\n            char var_14;\n            if (var_9) {\n                var_14 = (var_8 &lt; 32);\n            } else {\n                var_14 = 0;\n            }\n            char var_15 = (var_14 == 0);\n            if (var_15) {\n                \/\/ \"Argument out of bounds.\"\n            } else {\n            }\n            long long int var_16 = var_0[var_8];\n            long long int var_17 = var_1[var_8];\n            long long int var_18 = (var_16 * 2); \/\/ The actual work is done on this line.\n            var_1[var_8] = var_18;\n            long long int var_19 = (var_8 + 128);\n            var_6[0] = var_19;\n        }\n        long long int var_20 = var_6[0];\n    }\n    __device__ char method_24(long long int * var_0) {\n        long long int var_1 = var_0[0];\n        return (var_1 &lt; 32);\n    }\n}\n\"\"\"\n<\/pre>\n<p>What is shown above is a map kernel that multiplies all the elements of a tensor by 2. In the kernel there are a bunch of inactive range checks that can be turned on with a compiler switch.<\/p>\n<p>Here is the output of the above program at runtime.<\/p>\n<pre>[|[1, 2]; [2, 4]; [3, 6]; [4, 8]; [5, 10]; [6, 12]; [7, 14]; [8, 16]; [9, 18]; [10, 20]; [11, 22]; [12, 24]; [13, 26]; [14, 28]; [15, 30]; [16, 32]; [17, 34]; [18, 36]; [19, 38]; [20, 40]; [21, 42]; [22, 44]; [23, 46]; [24, 48]; [25, 50]; [26, 52]; [27, 54]; [28, 56]; [29, 58]; [30, 60]; [31, 62]; [32, 64]|]\n<\/pre>\n<h4><a id=\"user-content-how-cuda-kernels-are-compiled\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-cuda-kernels-are-compiled\"><\/a>How Cuda Kernels Are Compiled<\/h4>\n<p>At runtime, the program takes everything in the <code>cuda_kernels<\/code> variable and writes them to disk into the <code>cuda_kernels.cu<\/code> file which is located in the same place as the executable. It also creates a little batch script called <code>nvcc_router.bat<\/code>. Here are its contents. The paths are those provided by the <code>cfg<\/code> argument to the Spiral compiler directly.<\/p>\n<pre>SETLOCAL\nCALL \"C:\/Program Files (x86)\/Microsoft Visual Studio\/2017\/Community\\VC\/Auxiliary\/Build\/vcvarsall.bat\" x64 -vcvars_ver=14.11\nSET PATH=%PATH%;\"C:\/Program Files (x86)\/Microsoft Visual Studio\/2017\/Community\\VC\/Tools\/MSVC\/14.11.25503\/bin\/Hostx64\/x64\"\n\"C:\/Program Files\/NVIDIA GPU Computing Toolkit\/CUDA\/v9.0\\bin\/nvcc.exe\" -gencode=arch=compute_52,code=\\\"sm_52,compute_52\\\" --use-local-env --cl-version 2017 -I\"C:\/Program Files\/NVIDIA GPU Computing Toolkit\/CUDA\/v9.0\\include\" -I\"C:\/cub-1.7.4\" -I\"C:\/Program Files (x86)\/Microsoft Visual Studio\/2017\/Community\\VC\/Tools\/MSVC\/14.11.25503\/include\" --keep-dir \"C:\\Users\\Marko\\Source\\Repos\\The Spiral Language\\Temporary\\bin\\Release\" -maxrregcount=0  --machine 64 -ptx -cudart static  -o \"C:\\Users\\Marko\\Source\\Repos\\The Spiral Language\\Temporary\\bin\\Release\\cuda_kernels.ptx\" \"C:\\Users\\Marko\\Source\\Repos\\The Spiral Language\\Temporary\\bin\\Release\\cuda_kernels.cu\"\n<\/pre>\n<p>What this does is do some setting up and then calls NVCC (The Nvidia Cuda compiler) from the command line.<\/p>\n<p>It compiles the <code>cuda_kernels.cu<\/code> into <code>cuda_kernels.ptx<\/code> which is the LLVM IR assembly Cuda natively uses. It is not the actual assembly for the GPU, but an intermediate representation that various Cuda API functions use. Here is the output of the map kernel example just for show.<\/p>\n<pre>\/\/\n\/\/ Generated by NVIDIA NVVM Compiler\n\/\/\n\/\/ Compiler Build ID: CL-22781540\n\/\/ Cuda compilation tools, release 9.0, V9.0.176\n\/\/ Based on LLVM 3.4svn\n\/\/\n.version 6.0\n.target sm_52\n.address_size 64\n\t\/\/ .globl\tmethod_23\n.global .align 1 .b8 _ZN69_INTERNAL_47_tmpxft_000171e4_00000000_7_cuda_kernels_cpp1_ii_b5c879af6thrust6system6detail10sequential3seqE[1];\n.visible .entry method_23(\n\t.param .u64 method_23_param_0,\n\t.param .u64 method_23_param_1\n)\n{\n\t.reg .pred \t%p&lt;3&gt;;\n\t.reg .b32 \t%r&lt;3&gt;;\n\t.reg .b64 \t%rd&lt;25&gt;;\n\tld.param.u64 \t%rd10, [method_23_param_0];\n\tld.param.u64 \t%rd11, [method_23_param_1];\n\tmov.u32 \t%r1, %tid.x;\n\tcvt.u64.u32\t%rd12, %r1;\n\tmov.u32 \t%r2, %ctaid.x;\n\tmul.wide.u32 \t%rd13, %r2, 128;\n\tadd.s64 \t%rd24, %rd13, %rd12;\n\tsetp.gt.s64\t%p1, %rd24, 31;\n\t@%p1 bra \tBB0_3;\n\tcvta.to.global.u64 \t%rd14, %rd11;\n\tcvta.to.global.u64 \t%rd15, %rd10;\n\tadd.s64 \t%rd18, %rd13, %rd12;\n\tshl.b64 \t%rd19, %rd18, 3;\n\tadd.s64 \t%rd23, %rd15, %rd19;\n\tadd.s64 \t%rd22, %rd14, %rd19;\nBB0_2:\n\tld.global.u64 \t%rd20, [%rd23];\n\tshl.b64 \t%rd21, %rd20, 1;\n\tst.global.u64 \t[%rd22], %rd21;\n\tadd.s64 \t%rd23, %rd23, 1024;\n\tadd.s64 \t%rd22, %rd22, 1024;\n\tadd.s64 \t%rd24, %rd24, 128;\n\tsetp.lt.s64\t%p2, %rd24, 32;\n\t@%p2 bra \tBB0_2;\nBB0_3:\n\tret;\n}\n\t\/\/ .globl\t_ZN3cub11EmptyKernelIvEEvv\n.visible .entry _ZN3cub11EmptyKernelIvEEvv(\n)\n{\n\tret;\n}\n<\/pre>\n<p>As can be seen, the useless range checks got eliminated. And the multiply by two is converted into a left shift.<\/p>\n<pre>\tld.global.u64 \t%rd20, [%rd23];\n\tshl.b64 \t%rd21, %rd20, 1;\n\tst.global.u64 \t[%rd22], %rd21;\n<\/pre>\n<p>This is the meat of the loop where it loads from global memory, multiplies by 2 and then stores after that. The rest is setting up the kernel and implementing the loop.<\/p>\n<p>What can be done with the <code>.ptx<\/code> file is load the kernels inside them using the <a href=\"http:\/\/docs.nvidia.com\/cuda\/cuda-driver-api\/group__CUDA__MODULE.html#group__CUDA__MODULE\" rel=\"nofollow\">Cuda API functions<\/a> accessed through the <a href=\"https:\/\/github.com\/kunzmi\/managedCuda\">ManagedCuda wrapper library<\/a>.<\/p>\n<h4><a id=\"user-content-tour-of-the-standard-library-cuda-module\" class=\"anchor\" aria-hidden=\"true\" href=\"#tour-of-the-standard-library-cuda-module\"><\/a>Tour Of The Standard Library Cuda Module<\/h4>\n<p>The <code>Cuda<\/code> module is where the context is initialized and where the <code>run<\/code> function that actually launches the kernels resides. The kernel compilation to <code>.ptx<\/code> happens at runtime. It will be shown here in its entirety, but there is no need to think too deeply about what is going on. It is garden variety plumbing that can be summed in a sentence or two for each part.<\/p>\n<pre>let cuda =\n    (\n    \"Cuda\",[loops;console;array;host_tensor;extern_;object],\"The Cuda module.\",\n    \"\"\"\ninl ret -&gt; \n    open Extern\n    open Console\n    inl cuda_kernels = FS.Constant.cuda_kernels string\n    inl cuda_constant a t = !MacroCuda(t,[text: a])\n    inl cuda_constant_int constant () = cuda_constant constant int64\n    inl __blockDimX = cuda_constant_int \"blockDim.x\"\n    inl __blockDimY = cuda_constant_int \"blockDim.y\"\n    inl __blockDimZ = cuda_constant_int \"blockDim.z\"\n    inl __gridDimX = cuda_constant_int \"gridDim.x\"\n    inl __gridDimY = cuda_constant_int \"gridDim.y\"\n    inl __gridDimZ = cuda_constant_int \"gridDim.z\"\n    inl cuda_toolkit_path = @PathCuda\n    inl visual_studio_path = @PathVS2017\n    inl cub_path = @PathCub\n    inl env_type = fs [text: \"System.Environment\"]\n    inl context_type = fs [text: \"ManagedCuda.CudaContext\"]\n    use context = FS.Constructor context_type false\n    FS.Method context .Synchronize() ()\n<\/pre>\n<p>The only piece of functionality that happens at runtime are the last two lines where the Cuda context is created and then synchronized. The rest are merely definitions set up for later.<\/p>\n<pre>    inl compile_kernel_using_nvcc_bat_router (kernels_dir: string) =\n        inl path_type = fs [text: \"System.IO.Path\"]\n        inl combine x = FS.StaticMethod path_type .Combine x string\n    \n        inl file_type = fs [text: \"System.IO.File\"]\n        inl stream_type = fs [text: \"System.IO.Stream\"]\n        inl streamwriter_type = fs [text: \"System.IO.StreamWriter\"]\n        inl process_start_info_type = fs [text: \"System.Diagnostics.ProcessStartInfo\"]\n    \n        inl nvcc_router_path = combine (kernels_dir,\"nvcc_router.bat\")\n        inl procStartInfo = FS.Constructor process_start_info_type ()\n        FS.Method procStartInfo .set_RedirectStandardOutput true ()\n        FS.Method procStartInfo .set_RedirectStandardError true ()\n        FS.Method procStartInfo .set_UseShellExecute false ()\n        FS.Method procStartInfo .set_FileName nvcc_router_path ()\n        inl process_type = fs [text: \"System.Diagnostics.Process\"]\n        use process = FS.Constructor process_type ()\n        FS.Method process .set_StartInfo procStartInfo ()\n        inl print_to_standard_output = \n            closure_of (inl args -&gt; FS.Method args .get_Data() string |&gt; writeline) \n                (fs [text: \"System.Diagnostics.DataReceivedEventArgs\"] =&gt; ())\n        FS.Method process .\"OutputDataReceived.Add\" print_to_standard_output ()\n        FS.Method process .\"ErrorDataReceived.Add\" print_to_standard_output ()\n        inl concat = string_concat \"\"\n        inl (+) a b = concat (a, b)\n        \/\/\/ Puts quotes around the string.\n        inl quote x = (\"\\\"\",x,\"\\\"\")\n        inl vc_vars_args = \" x64 -vcvars_ver=14.11\"\n        inl quoted_vs_path_to_vcvars = combine(visual_studio_path, \"VC\/Auxiliary\/Build\/vcvarsall.bat\") |&gt; quote\n        inl quoted_vs_path_to_cl = combine(visual_studio_path, \"VC\/Tools\/MSVC\/14.11.25503\/bin\/Hostx64\/x64\") |&gt; quote\n        inl quoted_cuda_toolkit_path_to_include = combine(cuda_toolkit_path,\"include\") |&gt; quote\n        inl quoted_vc_path_to_include = combine(visual_studio_path, \"VC\/Tools\/MSVC\/14.11.25503\/include\") |&gt; quote\n        inl quoted_nvcc_path = combine(cuda_toolkit_path,@\"bin\/nvcc.exe\") |&gt; quote\n        inl quoted_cub_path_to_include = cub_path |&gt; quote\n        inl quoted_kernels_dir = kernels_dir |&gt; quote\n        inl target_path = combine(kernels_dir,\"cuda_kernels.ptx\")\n        inl quoted_target_path = target_path |&gt; quote\n        inl input_path = combine(kernels_dir,\"cuda_kernels.cu\")\n        inl quoted_input_path = input_path |&gt; quote\n        if FS.StaticMethod file_type .Exists input_path bool then FS.StaticMethod file_type .Delete input_path ()\n        FS.StaticMethod file_type .WriteAllText(input_path,cuda_kernels) ()\n   \n        inl _ = \n            if FS.StaticMethod file_type .Exists nvcc_router_path bool then FS.StaticMethod file_type .Delete nvcc_router_path ()\n            inl filestream_type = fs [text: \"System.IO.FileStream\"]\n            use nvcc_router_file = FS.StaticMethod file_type .OpenWrite(nvcc_router_path) filestream_type\n            use nvcc_router_stream = FS.Constructor streamwriter_type nvcc_router_file\n            inl write_to_batch = concat &gt;&gt; inl x -&gt; FS.Method nvcc_router_stream .WriteLine x ()\n            \"SETLOCAL\" |&gt; write_to_batch\n            (\"CALL \", quoted_vs_path_to_vcvars, vc_vars_args) |&gt; write_to_batch\n            (\"SET PATH=%PATH%;\", quoted_vs_path_to_cl) |&gt; write_to_batch\n            (\n            quoted_nvcc_path, \" -gencode=arch=compute_52,code=\\\\\\\"sm_52,compute_52\\\\\\\" --use-local-env --cl-version 2017\",\n            \" -I\", quoted_cuda_toolkit_path_to_include,\n            \" -I\", quoted_cub_path_to_include,\n            \" -I\", quoted_vc_path_to_include,\n            \" --keep-dir \",quoted_kernels_dir,\n            \" -maxrregcount=0  --machine 64 -ptx -cudart static  -o \",quoted_target_path,\" \",quoted_input_path\n            ) |&gt; write_to_batch\n        inl stopwatch_type = fs [text: \"System.Diagnostics.Stopwatch\"]\n        inl timer = FS.StaticMethod stopwatch_type .StartNew () stopwatch_type\n        if FS.Method process .Start() bool = false then failwith () \"NVCC failed to run.\"\n        FS.Method process .BeginOutputReadLine() ()\n        FS.Method process .BeginErrorReadLine() ()\n        FS.Method process .WaitForExit() ()\n        inl exit_code = FS.Method process .get_ExitCode() int32\n        assert (exit_code = 0i32) (\"NVCC failed compilation.\", exit_code)\n    \n        inl elapsed = FS.Method timer .get_Elapsed() (fs [text: \"System.TimeSpan\"])\n        !MacroFs((),[text: \"printfn \\\"The time it took to compile the Cuda kernels is: %A\\\" \"; arg: elapsed])\n        FS.Method context .LoadModulePTX target_path (fs [text: \"ManagedCuda.BasicTypes.CUmodule\"])\n    inl current_directory = FS.StaticMethod env_type .get_CurrentDirectory() string\n    inl modules = compile_kernel_using_nvcc_bat_router current_directory\n    writeline (string_concat \"\" (\"Compiled the kernels into the following directory: \", current_directory))\n<\/pre>\n<p>Here is the part the actually compiles the modules. It creates that batch script and sets up the <code>Process<\/code> object which it then uses to launch the script from the command line. The above code fragment is messy due to heavy use of macros, but it is straightforward.<\/p>\n<p>After the above is executed successfully, a <code>CUmodule<\/code> object is bound to the <code>modules<\/code> variable which holds all the kernels in binary format. This is used in the <code>run<\/code> function.<\/p>\n<pre>    inl dim3 = function\n        | {x y z} as m -&gt; m\n        | x,y,z -&gt; {x=x: int64; y=y: int64; z=z: int64}\n        | x,y -&gt; {x=x: int64; y=y: int64; z=1}\n        | x -&gt; {x=x: int64; y=1; z=1}\n    inl SizeT_type = fs [text: \"ManagedCuda.BasicTypes.SizeT\"]\n    inl CUdeviceptr_type = fs [text: \"ManagedCuda.BasicTypes.CUdeviceptr\"]\n    inl SizeT = FS.Constructor SizeT_type\n    inl CUdeviceptr = FS.Constructor CUdeviceptr_type &lt;&lt; SizeT\n    met run s {blockDim gridDim kernel} =\n        inl blockDim = dim3 blockDim\n        inl gridDim = dim3 gridDim\n        inl to_obj_ar args =\n            inl ty = fs [text: \"System.Object\"] |&gt; array\n            !MacroFs(ty,[fs_array_args: args; text: \": \"; type: ty])\n        inl kernel =\n            inl map_to_op_if_not_static {x y z} (x', y', z') = \n                inl f x x' = if lit_is x then const x else x' \n                f x x', f y y', f z z'\n            inl x,y,z = map_to_op_if_not_static blockDim (__blockDimX,__blockDimY,__blockDimZ)\n            inl x',y',z' = map_to_op_if_not_static gridDim (__gridDimX,__gridDimY,__gridDimZ)\n            inl _ -&gt; \/\/ This convoluted way of swaping non-literals for ops is so they do not get called outside of the kernel.\n                inl blockDim = {x=x(); y=y(); z=z()}\n                inl gridDim = {x=x'(); y=y'(); z=z'()}\n                kernel blockDim gridDim\n        inl join_point_entry_cuda x = !JoinPointEntryCuda(x())\n        inl method_name, args = join_point_entry_cuda kernel\n        \n        inl dim3 {x y z} = Tuple.map (to uint32) (x,y,z) |&gt; FS.Constructor (fs [text: \"ManagedCuda.VectorTypes.dim3\"])\n    \n        inl kernel_type = fs [text: \"ManagedCuda.CudaKernel\"]\n        inl cuda_kernel = FS.Constructor kernel_type (method_name,modules,s.data.context)\n        FS.Method cuda_kernel .set_GridDimensions(dim3 gridDim) ()\n        FS.Method cuda_kernel .set_BlockDimensions(dim3 blockDim) ()\n        match s.data.stream with\n        | () -&gt; FS.Method cuda_kernel .Run(to_obj_ar args) float32\n        | stream -&gt; FS.Method cuda_kernel .RunAsync(stream.extract,to_obj_ar args) ()\n<\/pre>\n<p>Before moving forward, here is how the above function is used in practice.<\/p>\n<pre>let tutorial1 =\n    \"tutorial1\",[cuda_modules],\"The placeholder for the tutorial examples\",\n    \"\"\"\ninb s = CudaModules (1024*1024) \ninl fact to = Loops.for {from=2; to state=dyn 1; body=inl {state i} -&gt; state * i}\ns.run {\n    blockDim=1\n    gridDim=1\n    kernel=inl blockDim gridDim -&gt; \n        fact 3 |&gt; ignore\n    }\n    \"\"\"\n<\/pre>\n<p>This would result in the factorial function as recursive GPU code.<\/p>\n<pre>let cuda_kernels = \"\"\"\n#include \"cub\/cub.cuh\"\nextern \"C\" {\n    __global__ void method_10();\n    __device__ long long int method_11(long long int var_0, long long int var_1);\n    \n    __global__ void method_10() {\n        long long int var_0 = 1;\n        long long int var_1 = 2;\n        long long int var_2 = method_11(var_0, var_1);\n    }\n    __device__ long long int method_11(long long int var_0, long long int var_1) {\n        char var_2 = (var_1 &lt;= 3);\n        if (var_2) {\n            long long int var_3 = (var_0 * var_1);\n            long long int var_4 = (var_1 + 1);\n            return method_11(var_3, var_4);\n        } else {\n            return var_0;\n        }\n    }\n}\n\"\"\"\n<\/pre>\n<p>In the standard library the following form would be used instead.<\/p>\n<pre>    kernel=cuda\n        fact 3 |&gt; ignore\n<\/pre>\n<p><code>cuda<\/code> is just shorthand for <code>inl blockDim gridDim -&gt;<\/code> built into the parser.<\/p>\n<p>There are points of interest that need to be explained for <code>run<\/code> to be fully understood.<\/p>\n<pre>            inl kernel =\n                inl map_to_op_if_not_static {x y z} (x', y', z') = \n                    inl f x x' = if lit_is x then const x else x' \n                    f x x', f y y', f z z'\n                inl x,y,z = map_to_op_if_not_static blockDim (__blockDimX,__blockDimY,__blockDimZ)\n                inl x',y',z' = map_to_op_if_not_static gridDim (__gridDimX,__gridDimY,__gridDimZ)\n                inl _ -&gt; \/\/ This convoluted way of swaping non-literals for ops is so they do not get called outside of the kernel.\n                    inl blockDim = {x=x(); y=y(); z=z()}\n                    inl gridDim = {x=x'(); y=y'(); z=z'()}\n                    kernel blockDim gridDim\n<\/pre>\n<p>What the above does is make sure that if the block and grid dimension sizes are known at compile time, that they are also passed into the kernel as literals. Otherwise they are used as intrinsics directly.<\/p>\n<p>In Cuda C code when <code>blockDim.x<\/code> is used directly, that is not a compile time constant, but a runtime variable. Spiral goes to an extra length in order to preserve information and intrinsics for substitutes literals whenever possible. The above code implements that.<\/p>\n<p>This is actually necessary for interop with the Cuda Unbound library which takes in block and grid dimensions as template parameters and requires them to be literals.<\/p>\n<pre>            inl join_point_entry_cuda x = !JoinPointEntryCuda(x())\n            inl method_name, args = join_point_entry_cuda kernel\n<\/pre>\n<p>The part directly after that is where the kernel gets executed. Cuda has a join point special to it and that is invoked using <code>!JoinPointEntryCuda(x())<\/code>. If this was a standard join point then that would be enough to call the function, but in Cuda's case things are a bit more complicated. Because all the calls have to go through the <code>ManagedCuda<\/code> library and then through the Cuda API, it would be very difficult to bake that call into the language directly.<\/p>\n<p>Instead what the Cuda join point does is compile the function and returns the method name and the free variables passed into the join point. The method name is used to extract the kernel and the free variables are passed as arguments to the <code>Run<\/code> and <code>RunAsync<\/code> method.<\/p>\n<pre>            inl dim3 {x y z} = Tuple.map (to uint32) (x,y,z) |&gt; FS.Constructor (fs [text: \"ManagedCuda.VectorTypes.dim3\"])\n    \n            inl kernel_type = fs [text: \"ManagedCuda.CudaKernel\"]\n            inl cuda_kernel = FS.Constructor kernel_type (method_name,modules,context)\n            FS.Method cuda_kernel .set_GridDimensions(dim3 gridDim) ()\n            FS.Method cuda_kernel .set_BlockDimensions(dim3 blockDim) ()\n            match stream with\n            | () -&gt; FS.Method cuda_kernel .Run(to_obj_ar args) float32\n            | stream -&gt; FS.Method cuda_kernel .RunAsync(stream.extract,to_obj_ar args) ()\n<\/pre>\n<p>Here is where the kernel is actually loaded.<\/p>\n<pre>inl cuda_kernel = FS.Constructor kernel_type (method_name,modules,context)\n<\/pre>\n<p>More specifically, on this line. It loads the kernel of <code>method_name<\/code> and uses the <code>modules<\/code> objects that holds all the compiled kernels from the previous step and the Cuda <code>context<\/code>.<\/p>\n<pre>            match stream with\n            | () -&gt; FS.Method cuda_kernel .Run(to_obj_ar args) float32\n            | stream -&gt; FS.Method cuda_kernel .RunAsync(stream.extract,to_obj_ar args) ()\n<\/pre>\n<p>Here is where the kernel is invoked synchronously or asynchronously depending on whether a stream was passed in.<\/p>\n<p>This covers everything needed to know in order to understand the Cuda backend. This is the Cuda backend in nearly its entirety.<\/p>\n<h4><a id=\"user-content-why-spiral-was-created\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-spiral-was-created\"><\/a>Why Spiral Was Created<\/h4>\n<p>The above is simple, but unless the language supports it then it is impossible to build it as a part of a library. Being able to do the above is one of the major motivators for the creation of Spiral. It simply saves such an enormous amount of work.<\/p>\n<h5><a id=\"user-content-how-it-used-to-be-done\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-it-used-to-be-done\"><\/a>How It Used To Be Done<\/h5>\n<p>The Spiral language originates from ML library of the same name done by author during the 2016 period.<\/p>\n<p>Here is how the map kernel used to look there.<\/p>\n<pre>\/\/\/ o &lt;- f(x)\ntype DeviceUnaryTransformModule(op: string, unique_name : string) = \n    let kernel_name = \"Map1Kernel\"+unique_name\n    let kernel_code = \n        [|\"\n        \/\/Kernel code:\n        extern \\\"C\\\" {\n            typedef float floatType;\n            __device__ inline floatType op(floatType x)\n            {\n                return \";op;\"\n            }\n        \n            \/\/ Device code\n            __global__ void \";kernel_name;\"(const floatType* A, floatType* O, const int N)\n            {\n                int i = blockDim.x * blockIdx.x + threadIdx.x;\n                const int stride = blockDim.x * gridDim.x;\n                while (i &lt; N)\n                {\n                    O[i] = op(A[i]);\n                    i += stride;\n                }\n            }\n        }\n        \" |] |&gt; String.concat \"\"\n    member val Kernel = load_kernel_nvrtc kernel_code kernel_name\n    member inline t.A\n            (str: CudaStream,\n                (ext_x: ^a -&gt; CUdeviceptr, x: ^a),\n                (ext_o: ^a -&gt; CUdeviceptr, o: ^a)) =\n        GuardSizes2(x,o)\n        let n = Size x\n        map_launcher(str, t.Kernel, n, [|ext_x x; ext_o o; n|])\n<\/pre>\n<p>It was used in two places.<\/p>\n<pre>let squareModule = lazy new DeviceUnaryTransformModule(\"x*x;\",\"Square\")\nlet logModule = lazy new DeviceUnaryTransformModule(\"logf(x);\",\"Log\")\n<\/pre>\n<p>Text macros are used as operations because there is not much choice apart from that for getting some generic kernel functionality. It was not much generic functionality though. Suppose I want to pass in two arguments into the map kernel instead of one.<\/p>\n<pre>\/\/\/ o &lt;- f(x,y)\ntype DeviceBinaryTransformModule(op: string, unique_name) = \n    let kernel_name = \"Map2Kernel\" + unique_name\n    let kernel_code = \n        [|\"\n        \/\/Kernel code:\n        extern \\\"C\\\" {\n            typedef float floatType;\n            __device__ inline floatType op(floatType x, floatType y)\n            {\n                return \";op;\"\n            }\n        \n            \/\/ Device code\n            __global__ void \";kernel_name;\"(const floatType* A, const floatType* B, floatType* O, const int N)\n            {\n                int i = blockDim.x * blockIdx.x + threadIdx.x;\n                const int stride = blockDim.x * gridDim.x;\n                while (i &lt; N)\n                {\n                    O[i] = op(A[i],B[i]);\n                    i += stride;\n                }\n            }\n        }\n        \" |] |&gt; String.concat \"\"\n    \n    member val Kernel = load_kernel_nvrtc kernel_code kernel_name\n    member inline t.A\n            (str: CudaStream,\n                (ext_x: ^a -&gt; CUdeviceptr, x: ^a),\n                (ext_y: ^a -&gt; CUdeviceptr, y: ^a),\n                (ext_o: ^a -&gt; CUdeviceptr, o: ^a)) =\n        GuardSizes3(x,y,o)\n        let n = Size x\n        map_launcher(str, t.Kernel, n, [|ext_x x; ext_y y; ext_o o; n|])\n<\/pre>\n<p>This one was used only for pointwise multiplication.<\/p>\n<pre>let hadamaradMultiplicationModule = lazy new DeviceBinaryTransformModule(\"x*y;\", \"HadMult\")\n<\/pre>\n<p>There were many more of such kernels and after 6 variation on the basic map, the author lost drive to more because once he had those he could implement the rest in terms of composition.<\/p>\n<p>As an example of what that means, here is how binary cross entropy was implemented in the previous library.<\/p>\n<pre>let inline cross_entropy_cost (target: ^a) (activations: ^a) = context {\n    let lt = target\n    let! ll = log_ activations\n    let! rt = scalar_matrix_add 1.0f -1.0f target\n    let! rl = scalar_matrix_add 1.0f -1.0f activations &gt;&gt;= log_\n    return! linear_layer_hadmult [|lt, ll; rt, rl|] \n            &gt;&gt;= sum\n            &gt;&gt;= scale (-1.0f \/ float32 (cols target))\n    }\n<\/pre>\n<p>The author could not write something like...<\/p>\n<pre>    inl cross_entropy = error {\n        fwd = inl x, y -&gt; -(y * log x + (one - y) * log (one - x))\n        bck = \n            inl (x, y) _ -&gt; (x - y) \/ (x * (one - x))\n            ,inl (x, y) _ -&gt; log (one - x) - log x\n        }\n<\/pre>\n<p>...but he did have individual operations for log and scalar matrix addition and Hadamarad multiplication so he could piece the required operation together.<\/p>\n<p>The difference between doing it directly and indirectly is the 6 intermediate allocations that the old library required to perform the same thing. Since GPUs are memory bound, that would make for a vast difference in performance.<\/p>\n<p>In the old library there simply was not a middle ground between writing all the kernels by hand and having to compose very small pieces in an inefficient manner. But the issues with the old arrangement did not stop there.<\/p>\n<p>Even if one is resolved to do the kernels by hand, some operations when done in composite require tracking a very large number of variables. Imagine the horror of having to deal with well over a dozen <code>float *<\/code> variables differentiated only by their name inside a single kernel and having no boundary or type checks to speak off. The author has had issues with swapping variables around by accident when there are just two of them of the same type in the same function.<\/p>\n<p>Mentally tracking over a dozen pointers to a tensor, their offsets and sizes with only their names to differentiate them would be simply impossible. And the author quickly realized that he could not take responsibility for such code in old library.<\/p>\n<p>Machine learning code is the worst in terms of debugging difficulty. Back when he first started, the author did appreciate dimensionality checking and had errors with some matrices being incorrectly transposed. There was one particular example where he hit 96% on Mnist despite the network propagating gradients in the wrong places.<\/p>\n<p>It very possible for mistakes to go unnoticed because only half the network gets updated or updated incorrectly, but the network still appears to work fine.<\/p>\n<p>Hence more than anywhere else, being able to reason about all aspects of code is of vital importance in a machine learning context. In fact, it is absolutely important everywhere.<\/p>\n<p>There is also one other point worth noting. All the operation in the old library have the <code>lazy<\/code> prefixed behind them. The reason for that is that they are compiled individually and the author found that NVRTC required around 0.5s to compile a single operation. With 20&gt; operations in the library that made for some massive compile times if all of them are compiled every time. NVCC is not much better. It requires 0.7s to compile an empty file and about 2.1s for a 3k file, so compiling all the kernels in single batch is important for speed's sake.<\/p>\n<p>This is also something that would be impossible to do without the support of a language. The assembling of information is in fact the dictionary definition for the world 'compilation'.<\/p>\n<h3><a id=\"user-content-7-object-orientation\" class=\"anchor\" aria-hidden=\"true\" href=\"#7-object-orientation\"><\/a>7: Object Orientation<\/h3>\n<p>Spiral is a functional language at its core, but it is good at doing OO despite not having specific features for it built in like other statically typed languages. Since OO is needed for passing around contexts in various key places, this chapter will be an overview of how it is done.<\/p>\n<h4><a class=\"anchor\" aria-hidden=\"true\" href=\"#motivation\"><\/a>Motivation<\/h4>\n<p>A significant amount of complexity in the Spiral's ML library comes from needing to manage Cuda memory directly. For a that a design pattern is needed to deal with it in the absence of garbage collection. Meaning all the data needs to be packed into some kind of object and passed around.<\/p>\n<p>In functional languages reader, state and writer monads are commonly used for this sort of dependency injection. An argument could be made that the reader pattern is a variant on the OO pattern except with the <code>self<\/code> argument on the opposite end.<\/p>\n<pre>inl f self a b = ... \/\/ OO pattern\ninl f a b self = ... \/\/ reader pattern\n<\/pre>\n<p>For some parts of the library it is much simpler to pass context in the first place and not bother with composition. The OO pattern shines there. Monads have the disadvantage of requiring all the code to be a part of the same monadic workflow. Some languages like Haskell go through great lengths to make that ergonomic, but Spiral is not such a language and it would be preferable to avoid using monads unless they are needed.<\/p>\n<p>It is easy to switch between the above patterns anyway.<\/p>\n<p>One other advantage is that OO breaks the usual top down ordering of functions.<\/p>\n<pre>inl a x = ...\ninl b x = ...\ninl c x = ...\n<\/pre>\n<p>In the above fragment, <code>a<\/code> cannot refer to <code>b<\/code> or <code>c<\/code> as they come after it, and <code>b<\/code> cannot refer to <code>c<\/code>, but it is possible to get around that using OO in Spiral which can be useful.<\/p>\n<p>The OO pattern can also be used for easy immutable updates. This is not something OO is known for.<\/p>\n<h4><a id=\"user-content-the-object\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-object\"><\/a>The Object<\/h4>\n<pre>\/\/\/ Converts the argument (usually a module) to the object form.\ninl obj s x = s x s\n<\/pre>\n<p><code>obj<\/code> is defined in <code>Core<\/code>. In <code>s x s<\/code> the <code>s x<\/code> part selects the method from <code>s<\/code> and then <code>s<\/code> is passed into that method.<\/p>\n<pre>let object =\n    (\n    \"Object\",[],\"The Object module.\",\n    \"\"\"\n{\ndata' = {}\ndata = inl {data'} -&gt; data'\ndata_add = inl s v -&gt; {s with data'=module_foldl (inl name s v -&gt; module_add name v s) (indiv self) v |&gt; heap} |&gt; obj\nmember_add = inl s -&gt; module_foldl (inl name s v -&gt; module_add name (inl s -&gt; v (obj s)) s) s &gt;&gt; obj\nmodule_add = inl s name v -&gt; module_add name (inl s name -&gt; v name (obj s)) s |&gt; obj\nunwrap = id\n} \n|&gt; obj\n    \"\"\") |&gt; module_\n<\/pre>\n<p>The base Object in Spiral is implemented like this. It continually wraps its members with the <code>obj<\/code> function in order to emulate OO access in standard statically typed OO languages. Maybe at some point remove functions will be added.<\/p>\n<p>The main thing to keep in mind when considering <code>data_add<\/code>, <code>member_add<\/code>, <code>module_add<\/code> is that the first argument to them is not an object, but an unwrapped module.<\/p>\n<pre>module_add name (inl s -&gt; v (obj s)) s\n<\/pre>\n<p>Hence when the module <code>s<\/code> is being passed into the member, it is always wrapped into an object. What would happen if the fragment was written like this...<\/p>\n<pre>module_add name v s\n<\/pre>\n<p>...is that then the unwrapped <code>s<\/code> would get passed into the member. The reason why <code>data_add<\/code> does not wrap <code>v<\/code> inside the module is to allow data fields to be accessed directly.<\/p>\n<p><code>data_add<\/code>, <code>member_add<\/code>, <code>module_add<\/code> always return an object so their return is piped to <code>obj<\/code>.<\/p>\n<pre>data' = {}\ndata = inl {data'} -&gt; data'\n<\/pre>\n<p>The actual data is stored in the <code>data'<\/code> field rather than <code>data<\/code> because trying to access an object will pass itself to its argument as the first move and result in a type error. Hence the only way to access the data directly in that case would be to call <code>unwrap<\/code> first or do it like it has been done and make a member that reroutes to it.<\/p>\n<p><code>data_add<\/code> in particular also makes sure that the <code>data'<\/code> is stored on the heap at all times rather than passed around as individual arguments through join points. It makes the generated code look decently nicer, speeds up compilation and improves runtime performance.<\/p>\n<p>Here is an example of it in use:<\/p>\n<pre>    Object\n        .member_add {run}\n        .data_add {context stream=()}\n    |&gt; ret\n<\/pre>\n<p><code>module_add<\/code> can be used similarly.<\/p>\n<pre>s.module_add .CudaTensor methods\n<\/pre>\n<p>Members can be accessed directly like...<\/p>\n<pre>s.allocate\n<\/pre>\n<p>Modules on the other hand require two type literals as arguments.<\/p>\n<pre>s.RegionMem.allocate\n<\/pre>\n<p>Data member are prefixed with <code>data<\/code>.<\/p>\n<pre>s.data.context\n<\/pre>\n<h3><a id=\"user-content-8-gpu-programming-basics\" class=\"anchor\" aria-hidden=\"true\" href=\"#8-gpu-programming-basics\"><\/a>8: GPU Programming Basics<\/h3>\n<p>GPU programming is well worth learning for those interested in drawing out as most performance as is possible from the machine, and lessons from it transfer over into CPU programming. At some point Spiral will drop Cuda and switch to supporting neural computing architectures, and when that happens the lesson learning in the GPU arena can be expected to transfer. The reason for that will be because those architectures will be similar to GPUs except with a lot more local memory.<\/p>\n<p>Knowing how to program effectively will never go out of date.<\/p>\n<p>This chapter will cover a few select kernels from the <code>CudaKernel<\/code> module found in the <code>Learning<\/code> project. There are more than will be covered in this chapter and more yet will be made in the future, but these can be counted on to illustrate how Spiral does Cuda programming.<\/p>\n<h4><a id=\"user-content-map\" class=\"anchor\" aria-hidden=\"true\" href=\"#map\"><\/a>map<\/h4>\n<pre>met map' w f in out = \n    inl in, out = zip in, zip out\n    inl to_dev_tensor = w.CudaTensor.to_dev_tensor\n    assert (in.dim = out.dim) \"The input and output dimensions must be equal.\"\n    inl in = flatten in |&gt; to_dev_tensor\n    inl out = flatten out |&gt; to_dev_tensor\n    inl in_a :: () = in.dim\n    \n    inl blockDim = 128\n    inl gridDim = min 64 (divup (s in_a) blockDim)\n    w.run {\n        blockDim gridDim\n        kernel = cuda \/\/ Lexical scoping rocks.\n            grid_for {blockDim gridDim} .x in_a {body=inl {i} -&gt;\n                inl out = out i\n                inl in = in i\n                out .set (f in.get out.get)\n                }\n        }\n<\/pre>\n<p>It actually more of a pain to set up the arguments before passing them into the kernel than writing the kernel itself. In fact, the actual kernel could be shortened to one line without much trouble.<\/p>\n<pre>cuda grid_for {blockDim gridDim} .x in_a {body=inl {i} -&gt; out i .set (f (in i .get) (out i .get))}\n<\/pre>\n<p>Lexical scoping just buys so much in this example that it is amazing. It scoops up both of the input and output tensors and allows me to pass in <code>in_a<\/code> as the loop dimension in one move. Considering how wide range of a functionality Spiral's tensors have compared to arrays that makes writing kernels significantly easier.<\/p>\n<p>The above example is simple as the tensors are <code>flatten<\/code>ed to 1d before being shipped off into the kernel, but other kernels will make use of the tensor's full functionality.<\/p>\n<p>Most of the functionality in the above fragment should be obvious from reading it, but <code>to_dev_tensor<\/code> and <code>divup<\/code> require an explanation.<\/p>\n<pre>inl divup a b = (a-1)\/b+1 \/\/ Integer division with rounding up. (a+b-1)\/b is another variant on this.\n<\/pre>\n<p>So <code>divup 9 5 = 2<\/code> and <code>divup 10 5 = 2<\/code>, but <code>divup 11 5 = 3<\/code>.<\/p>\n<p><code>to_dev_tensor<\/code> on the other hand does some transformation on the tensor under the hood. <code>CudaTensor<\/code>s have a reference to a pointer, but the device tensors need to be raw pointers. So <code>to_dev_tensor<\/code> strips the reference in order to allow the tensor to cross the language boundary. In addition to that, it also adds the offset to the pointer and replaces it with 0.<\/p>\n<p>In .NET land it is not possible to move pointers around - array pointers must always point at the beginning of it, but in C they can point to anywhere. By adding the offset to the pointer and replacing it with 0, that allows the offset to be passed through the join point as a literal and not manifest as an argument.<\/p>\n<p>In Spiral, some types which are not blittable like strings and chars can be passed onto the Cuda side at compile time as literals. This hold true for all types fully known at compile time.<\/p>\n<p>Apart from that the context is <code>w<\/code> in the <code>CudaKernel<\/code> module because <code>s<\/code> is taken up by the span <code>{near_to from} -&gt; near_to - from<\/code> function.<\/p>\n<p>The following example was already shown in the Cuda backend chapter, but here it is as a review.<\/p>\n<pre>let kernel1 =\n    \"kernel1\",[cuda_modules],\"Does the map kernel work?\",\n    \"\"\"\n\/\/\/ Initializes all the Cuda parts\ninb s = CudaModules (1024*1024) \/\/ The allocator takes 1Mb of memory from the heap.\n\/\/\/ Creates a host tensor with the given generator function.\ninl h = HostTensor.init 32 (inl x -&gt; x + 1) \n\/\/\/ Loads the tensor on the GPU based on the host tensor\ninl a1 = s.CudaTensor.from_host_tensor h\n\/\/\/ Makes a tensor of the same type and dimensions as `a1` and zeroes it.\ninl o1 = s.CudaTensor.zero_like a1\n\/\/\/ Calls the map operation. `a1` is the input and `o1` is the output.\ns.CudaKernel.map' (inl a _ -&gt; a * 2) a1 o1\n\/\/\/ Transfers the tensor back to host.\ninl a2 = s.CudaTensor.to_host_tensor o1\n\/\/\/ Zips the two tensors and prints them out.\nHostTensor.zip (h,a2) |&gt; HostTensor.show |&gt; Console.writeline\n    \"\"\"\n<\/pre>\n<p>The kind of code this fragment produces was described in the backend chapter, so it won't be covered here.<\/p>\n<p><code>map'<\/code> as shown here has the notable issue of needing to have its output feed to it. It is possible to abstract that away.<\/p>\n<pre>inl map w f in =\n    indiv join\n        inl in = zip in\n        inl out = w.CudaTensor.create {dim=in.dim; elem_type=type f in.elem_type}\n        map' w (inl in _ -&gt; f in) in out\n        stack out\n<\/pre>\n<p>All the functions apart from <code>map_redo_map<\/code> in the <code>CudaKernel<\/code> module have a variant that automatically allocates the output provided to them.<\/p>\n<p>The reason why map is written like the above instead of like...<\/p>\n<pre>inl map w f in =\n    inl in = zip in\n    inl out = w.CudaTensor.create {dim=in.dim; elem_type=type f in.elem_type}\n    map' w (inl in _ -&gt; f in) in out\n    out\n<\/pre>\n<p>...is because it was found out that the variant that uses join points and layout types has significantly better compile times. This will be covered in more detail later.<\/p>\n<p>What <code>map<\/code> does is allocate the output tensor based on the input dimension and the inferred output type and then passes that to <code>map'<\/code>. It also wraps around the input function so it throws away the output argument given to it.<\/p>\n<p>All auxiliary kernel variants do those and only those things.<\/p>\n<h4><a id=\"user-content-map_redo_map\" class=\"anchor\" aria-hidden=\"true\" href=\"#map_redo_map\"><\/a>map_redo_map<\/h4>\n<p>Zips, flattens the tensor to 1d, maps it, reduces it and then maps the output scalar tensor. This particular kernel is used for cost functions in the ML library.<\/p>\n<pre>\/\/\/ Zips, flattens the tensor to 1d, maps, reduces it and maps it.\n\/\/\/ Map is optional. Allocates a temporary tensor for the intermediary results.\ninl map_redo_map w {d with redo neutral_elem} in =\n    indiv join\n        inl to_dev_tensor = w.CudaTensor.to_dev_tensor\n   \n        inl run {map_out map_in blockDim gridDim} (!to_dev_tensor in) =\n            inl in_a :: () = in.dim\n            inl out = w.CudaTensor.create {elem_type=type map_in in.elem_type |&gt; map_out; dim=gridDim}\n            inl out' = to_dev_tensor out\n            w.run {\n                blockDim gridDim\n                kernel = cuda \n                    inl x = \n                        grid_for {blockDim gridDim} .x in_a {state=dyn neutral_elem; body=inl {state i} -&gt; redo state (map_in (in i .get)) }\n                        |&gt; cub_block_reduce {blockDim redo} |&gt; map_out\n                    if threadIdx.x = 0 then out' blockIdx.x .set x\n                }\n            out\n        inl in = zip in |&gt; flatten\n        inl map_in = match d with {map_in} -&gt; map_in | _ -&gt; id\n        inl map_out = match d with {map_out} -&gt; map_out | _ -&gt; id\n        inl in_a :: () = in.dim\n        inl span = s in_a\n        inl blockDim = lit_min span 1024\n        inl gridDim = 1 \/\/lit_min 64 (divup span blockDim)\n        inl r = \n            if gridDim = 1 then\n                run {map_out map_in blockDim gridDim} in\n            else\n                run {map_out=id; map_in blockDim gridDim} in\n                |&gt; run {map_out map_in=id; blockDim=gridDim; gridDim=1}\n        r 0 |&gt; stack\n<\/pre>\n<p>The actual Cuda part of the kernel is this...<\/p>\n<pre>inl x = \n    grid_for {blockDim gridDim} .x in_a {state=dyn neutral_elem; body=inl {state i} -&gt; redo state (map_in (in i .get)) }\n    |&gt; cub_block_reduce {blockDim redo} |&gt; map_out\nif threadIdx.x = 0 then out' blockIdx.x .set x\n<\/pre>\n<p>The kernel can be described in 3 steps:<\/p>\n<ol><li>The blocks iterate over the global memory and perform the maps and reductions as they go along. This means that if there is one block of 1024 iterating over a 1024*128 array, it will perform 128 reductions before moving to the next step.<\/li>\n<li>A block reduction. Whenever possible, Spiral offloads work to the Cuda Unbound library as writing Cuda kernels can be tricky.<\/li>\n<li>When that is done, only the first thread has the result of the block reduce. It writes that result to global memory after mapping it.<\/li>\n<\/ol><p>If the <code>gridDim<\/code> is 1, then the actually kernel needs to be run only once, but otherwise it must be done twice in order to perform the inter block reduction.<\/p>\n<h5><a id=\"user-content-flatten\" class=\"anchor\" aria-hidden=\"true\" href=\"#flatten\"><\/a>flatten<\/h5>\n<pre>        inl in = zip in |&gt; flatten\n<\/pre>\n<p>Like in the map, the input is zipped and then flattened. Since flatten was made after the Tensor chapter was made, it will be covered here. Here it is in full from the <code>HostTensor<\/code> module.<\/p>\n<pre>\/\/\/ Flattens the tensor to a single dimension.\ninl flatten tns =\n    match tns.dim with\n    | () -&gt; tns\n    | !(Tuple.map span) dim -&gt;\n        tns .set_dim (product dim)\n            .update_body (inl {d with size} -&gt;\n                Tuple.zip (dim,size)\n                |&gt; Tuple.reducel (inl d,s d',s' -&gt;\n                    assert (s = d' * s') \"The tensor must be contiguous in order to be flattened.\"\n                    d*s, s'\n                    )\n                |&gt; inl _,s -&gt; {d with size=s :: ()}\n                )\n<\/pre>\n<p>The dimension of the output tensor is set to the product of its spans, but in order for a tensor to be capable of being flattened it must be contiguous.<\/p>\n<pre>                |&gt; Tuple.reducel (inl d,s d',s' -&gt;\n                    assert (s = d' * s') \"The tensor must be contiguous in order to be flattened.\"\n                    d*s, s'\n                    )\n<\/pre>\n<p>In order for that to hold true the size of the outer dimension must equal the span of the inner dimension times its size.<\/p>\n<p>What that means is this - suppose there is a 2d tensor of dimensions (10,10) and size (10,1). That tensor could then be flattened to a 1d tensor of dimension 100 and size 1. But suppose a view of the tensor was taken along the inner dimension.<\/p>\n<pre>tns.view_span (inl a,b -&gt; a, 5)\n<\/pre>\n<p>Now the resulting tensor would have dimensions of (10,5) and size of (10,1).<\/p>\n<p>Since 10 &lt;&gt; 5 * 1 that means that the tensor is not contiguous. This also means that tensors that have been rotated cannot be flattened. Viewing the outermost dimension would be fine though.<\/p>\n<p>This covers flattening.<\/p>\n<pre>        inl map_in = match d with {map_in} -&gt; map_in | _ -&gt; id\n        inl map_out = match d with {map_out} -&gt; map_out | _ -&gt; id\n<\/pre>\n<p>Since <code>map_in<\/code> and <code>map_out<\/code> are optional, they are replaced with the identity function if they are missing inside the kernel.<\/p>\n<pre>        inl in_a :: () = in.dim\n        inl span = s in_a\n        inl blockDim = lit_min span 1024\n        inl gridDim = 1 \/\/lit_min 64 (divup span blockDim)\n<\/pre>\n<p>Here the <code>gridDim<\/code> is just set to 1 because the author was lazy and decided to leave the job of fiddling with the launch parameters for later.<\/p>\n<pre>\/\/\/ The template for lit_min and lit_max.\ninl lit_comp op a b =\n    if lit_is a &amp;&amp; lit_is b then op a b\n    elif lit_is a then a\n    elif lit_is b then b\n    else error_type \"a or b needs to be a literal\"\n\/\/\/ Returns the compile time expressible maximum of the two expressions.\ninl lit_max = lit_comp max\n\/\/\/ Returns the compile time expressible minimum of the two expressions.\ninl lit_min = lit_comp min\n<\/pre>\n<p><code>lit_min<\/code> is defined in <code>Core<\/code> like this. A standard min function would return the minimum of the two arguments. <code>lit_min<\/code> on the other hand does that only between two literals. If one of the arguments is not a literal, then it just returns the literal. If both are not literals then that is a type error.<\/p>\n<pre>        inl r = \n            if gridDim = 1 then\n                run {map_out map_in blockDim gridDim} in\n            else\n                run {map_out=id; map_in blockDim gridDim} in\n                |&gt; run {map_out map_in=id; blockDim=gridDim; gridDim=1}\n        r 0 |&gt; stack\n<\/pre>\n<p>Here is where the kernel is launched either once or twice depending on how many reductions are needed and then returned.<\/p>\n<pre>        inl run {map_out map_in blockDim gridDim} (!to_dev_tensor in) =\n            inl in_a :: () = in.dim\n            inl out = w.CudaTensor.create {elem_type=type map_in in.elem_type |&gt; map_out; dim=gridDim}\n            inl out' = to_dev_tensor out\n<\/pre>\n<p>Here the <code>run<\/code> infers the type and creates the output before passing it into the kernel.<\/p>\n<p>As can be seen the scaffolding for the actual kernels has a lot of details and needs to be covered a few times.<\/p>\n<h5><a id=\"user-content-cuda-loops\" class=\"anchor\" aria-hidden=\"true\" href=\"#cuda-loops\"><\/a>Cuda Loops<\/h5>\n<pre>let kernel2 =\n    \"kernel2\",[cuda_modules],\"Does the map_redo_map kernel work?\",\n    \"\"\"\ninb s = CudaModules (1024*1024)\ninl h = HostTensor.init 1024 ((+) 1)\ninl a1 = s.CudaTensor.from_host_tensor h\ns.CudaKernel.map_redo_map {neutral_elem=0; redo=(+)} a1\n|&gt; s.CudaTensor.print \/\/ 524800\n    \"\"\"\n<\/pre>\n<p>The above sums all the numbers between [1..1024].<\/p>\n<pre>#include \"cub\/cub.cuh\"\nextern \"C\" {\n    __global__ void method_21(long long int * var_0, long long int * var_1);\n    __device__ char method_22(long long int * var_0, long long int * var_1);\n    \n    __global__ void method_21(long long int * var_0, long long int * var_1) {\n        long long int var_2 = threadIdx.x;\n        long long int var_3 = blockIdx.x;\n        long long int var_4 = (1024 * var_3);\n        long long int var_5 = (var_2 + var_4);\n        long long int var_6 = 0;\n        long long int var_7[1];\n        long long int var_8[1];\n        var_7[0] = var_5;\n        var_8[0] = var_6;\n        while (method_22(var_7, var_8)) {\n            long long int var_10 = var_7[0];\n            long long int var_11 = var_8[0];\n            char var_12 = (var_10 &gt;= 0);\n            char var_14;\n            if (var_12) {\n                var_14 = (var_10 &lt; 1024);\n            } else {\n                var_14 = 0;\n            }\n            char var_15 = (var_14 == 0);\n            if (var_15) {\n                \/\/ \"Argument out of bounds.\"\n            } else {\n            }\n            long long int var_16 = var_0[var_10];\n            long long int var_17 = (var_11 + var_16);\n            long long int var_18 = (var_10 + 1024);\n            var_7[0] = var_18;\n            var_8[0] = var_17;\n        }\n        long long int var_19 = var_7[0];\n        long long int var_20 = var_8[0];\n        long long int var_21 = cub::BlockReduce&lt;long long int,1024,cub::BLOCK_REDUCE_WARP_REDUCTIONS,1,1&gt;().Sum(var_20);\n        long long int var_22 = threadIdx.x;\n        char var_23 = (var_22 == 0);\n        if (var_23) {\n            long long int var_24 = blockIdx.x;\n            char var_25 = (var_24 &gt;= 0);\n            char var_27;\n            if (var_25) {\n                var_27 = (var_24 &lt; 1);\n            } else {\n                var_27 = 0;\n            }\n            char var_28 = (var_27 == 0);\n            if (var_28) {\n                \/\/ \"Argument out of bounds.\"\n            } else {\n            }\n            var_1[var_24] = var_21;\n        } else {\n        }\n    }\n    __device__ char method_22(long long int * var_0, long long int * var_1) {\n        long long int var_2 = var_0[0];\n        long long int var_3 = var_1[0];\n        return (var_2 &lt; 1024);\n    }\n}\n<\/pre>\n<p>The produced kernel is of course a decent bit longer than 3 lines. The resulting output in whole is 745 LOC, so it can be pasted here due to its size. As this is a good opportunity to do so, Cuda loops will be covered here.<\/p>\n<p>Originally these kernels used the tail recursive loops covered in an earlier chapter. It seems like a good idea at the time since the Cuda compiler appeared to be able to perform the tail recursive optimization. Nevertheless, as the author has suspected might happen since literally nobody but him would ever try compiling Cuda loops in such a manner, when tail recursive loops are combined with tuple returns, NVCC will outright produce incorrect code that will lead to local write errors. Based on the reply to the bug report he sent, the NVCC does not support tail recursion optimization and presumably the bug won't be fixed.<\/p>\n<p>Hence just on the Cuda side, the language supports standard imperative loops, more specifically the while loop.<\/p>\n<pre>inl whilecd {cond state body} =\n    inl r = HostTensor.create {\n        array_create=array_create_cuda_local \n        elem_type=state \n        dim=()\n        }\n    r .set state\n    \/\/\/ Note: While must have a join point around it.\n    !While((join cond r.get), (r.set &lt;| body r.get))\n    r .get\n<\/pre>\n<p>As can be seen, Spiral's tensors can be used with Cuda local arrays without any changes even in tuple of array form.<\/p>\n<pre>    inl r = HostTensor.create {\n        array_create=array_create_cuda_local \n        elem_type=state \n        dim=()\n        }\n<\/pre>\n<p>The above fragment corresponds to this one in the code...<\/p>\n<pre>        long long int var_7[1];\n        long long int var_8[1];\n<\/pre>\n<p>Going forward.<\/p>\n<pre>r .set state\n<\/pre>\n<p>Is...<\/p>\n<pre>        var_7[0] = var_5;\n        var_8[0] = var_6;\n<\/pre>\n<p>The...<\/p>\n<pre>!While((join cond r.get), (r.set &lt;| body r.get))\n<\/pre>\n<p>...is everything in the while loop. Note that since C requires the conditional a single expression the easiest way to achieve that is to stick it in a join point. Compiler tends to split expressions into statements during the code generation phase which would spill out of the conditional unless it was lifted into a function.<\/p>\n<pre>r .get\n<\/pre>\n<p>Corresponds to after the while loop.<\/p>\n<pre>        long long int var_19 = var_7[0];\n        long long int var_20 = var_8[0];\n<\/pre>\n<p>Of the two variables in the tensor, one is a loop counter and the other is the actual state.<\/p>\n<pre>inl forcd {d with from body} =\n    inl finally =\n        match d with\n        | {finally} -&gt; finally\n        | _ -&gt; id\n    inl check =\n        match d with\n        | {near_to} from -&gt; from &lt; near_to \n        | {to} from -&gt; from &lt;= to\n        | {down_to} from -&gt; from &gt;= down_to\n        | {near_down_to} from -&gt; from &gt; near_down_to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` needs be present.\"\n    inl by =\n        match d with\n        | {by} -&gt; by\n        | {to | near_to} -&gt; 1\n        | {down_to | near_down_to} -&gt; -1\n    inl to =\n        match d with\n        | {(to ^ near_to ^ down_to ^ near_down_to)=to} -&gt; to\n        | _ -&gt; error_type \"Only one of `to`,`near_to`,`down_to`,`near_down_to` is allowed.\"\n    inl state = \n        match d with\n        | {state} -&gt; state\n        | _ -&gt; ()\n    inl state = {from state}\n    whilecd {\n        state\n        cond = inl {from state} -&gt; check from\n        body = inl {from state} -&gt; {state=body {state i=from}; from=from+by}\n        } .state\n    |&gt; finally\n<\/pre>\n<p>Most of this should be familiar from the loops chapter. Making a for loop from a while loop should not be much of a challenge by this point. It is a fairly straightforward extension of the concept.<\/p>\n<p>From the for loop comes the Cuda specialized <code>grid_for<\/code> loop which makes it straightforward to iterate over a tensor in a block-strided fashion.<\/p>\n<pre>inl grid_for_template {iteration_mode} {blockDim gridDim} axis dim =\n    inl from = threadIdx axis + blockDim axis * blockIdx axis - dim.from\n    inl by = gridDim axis * blockDim axis\n    inl near_to = dim.near_to\n    match iteration_mode with\n    | .items_per_thread {d with body} -&gt;\n        inl span = s dim\n        inl items_per_thread = divup span by\n        forcd {d with from=0;near_to=items_per_thread; body=inl {state i=item} -&gt;\n            inl i = from + by * item\n            inl num_valid = span - by * item\n            if i &lt; near_to then body {span num_valid item state i} else state\n            }\n    | .std d -&gt; forcd {d with from by near_to}\ninl grid_for_items = grid_for_template {iteration_mode=.items_per_thread}\ninl grid_for = grid_for_template {iteration_mode=.std}\n<\/pre>\n<p>The <code>axis<\/code> argument can only be <code>.x<\/code>, <code>.y<\/code> or <code>.z<\/code>. In Cuda examples online, they often compulsively write out...<\/p>\n<pre>int stride = gridDim.x * blockDim.x;\nfor {int i = threadIdx.x + blockDim.x * blockIdx.x; i &lt; N; i += stride} {...}\n<\/pre>\n<p>Or something to that effect. <code>grid_for<\/code> is inteded to abstract that away so as to make the code more concise and reduce the chance of error. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Structured_programming\" rel=\"nofollow\">Structured programming<\/a> started the trend of using loops and if statements, but Cuda itself firmly remains stuck in the late 50s in terms of what it offers inside the kernel. <code>grid_for<\/code> is a slight, but a notable improvement over using the raw for loop.<\/p>\n<p>The standard <code>grid_for<\/code> that has been shown functions like a normal loop, but the <code>grid_for_items<\/code> has a special purpose to it.<\/p>\n<p>To understand why it is needed here is a quiz. Where are these variables allocated?<\/p>\n<pre>int a;\nint b[2];\nint c[var_11];\n<\/pre>\n<p><code>a<\/code> and <code>b<\/code> are held in registers. This is despite <code>b<\/code> being an array. On the other hand since <code>c<\/code>'s size is not statically known, it is allocated in global memory or most likely the cache.<\/p>\n<p>Whether <code>b<\/code> is allocated in registers or not also depends on how it is indexed.<\/p>\n<pre>b[0]; \/\/ safe\nb[1]; \/\/ safe\nb[var_12]; \/\/ better allocate it in global memory after all\n<\/pre>\n<p>So the Cuda array's type depends on how it is used and the compiler will not give any indication of what is going on under the hood the programmer.<\/p>\n<p>Furthermore, one other important thing to keep in mind is that C compilers, and especially Cuda compilers are specialized for optimizing loops. If the loop boundaries are statically known, which they usually are if one is programming in Spiral, the Cuda compiler will strongly attempt to unroll it.<\/p>\n<p>If it is inspected with <code>print_static<\/code> then <code>item<\/code> will show up as a runtime variable, but Cuda will know the difference after unrolling and will be able to determine at compile time what the indices into the array are.<\/p>\n<p>This is important because variables held in registers can be accessed orders of magnitude more quickly than in main memory, and this particular trick is used in reduce + broadcast kernels which are needed to implement softmax error and layer normalization. It is also used in the sampler.<\/p>\n<h4><a id=\"user-content-d2_replicate_map\" class=\"anchor\" aria-hidden=\"true\" href=\"#d2_replicate_map\"><\/a>d2_replicate_map<\/h4>\n<p>This function replicates the first 1d tensor so it matches the 2d second. It is used for adding biases to the result of a matrix multiply.<\/p>\n<pre>\/\/\/ Replicates the 1d `in` and maps it along the outer dimension as determined by `in'`.\nmet d2_replicate_map' w f in in' out =\n    inl to_dev_tensor = w.CudaTensor.to_dev_tensor\n    \n    inl in, in', out = zip in, zip in', zip out\n    inl dim_in :: () = in.dim\n    inl dim_in'_a, dim_in'_b = in'.dim\n    assert (dim_in = dim_in'_b) \"Input's dimension must equal the second input's inner dimension.\"\n    assert (in'.dim = out.dim) \"Second input must have the same dimension as the output.\"\n    inl blockDimX = min warp_size (s dim_in)\n    inl blockDimY = min 32 (s dim_in'_a)\n    inl gridDim = min 64 (divup (s dim_in) blockDimX)\n    inl in = to_dev_tensor in\n    inl in' = to_dev_tensor in'\n    inl out = to_dev_tensor out\n    w.run {\n        gridDim\n        blockDim=blockDimX,blockDimY\n        kernel = cuda \n            inl grid_for = grid_for {gridDim blockDim}\n            grid_for .x dim_in'_b {body=inl {i} -&gt;\n                inl in = in i .get\n                inl in' j = in' j i\n                inl out j = out j i\n                grid_for .y dim_in'_a {body=inl {i} -&gt;\n                    inl in', out = in' i, out i\n                    out.set (f in in'.get out.get)\n                    }\n                }\n        }\n<\/pre>\n<p>The first half of the kernel is the standard fare - zips, assertions for dimension sizes, calls into <code>to_dev_tensor<\/code>, and determination of the block and grid dimensions. The actual kernel part is a bit more interesting than last time.<\/p>\n<pre>            inl grid_for = grid_for {gridDim blockDim}\n            grid_for .x dim_in'_b {body=inl {i} -&gt;\n                inl in = in i .get\n                inl in' j = in' j i\n                inl out j = out j i\n                grid_for .y dim_in'_a {body=inl {i} -&gt;\n                    inl in', out = in' i, out i\n                    out.set (f in in'.get out.get)\n                    }\n                }\n<\/pre>\n<p>On the very first line there is an example of partial application on the loop function. <code>grid_for<\/code> is applied with the module holding the block and grid dimensions. Then it is used twice. First it used to iterate over the innermost dimension along the <code>x<\/code> axis.<\/p>\n<pre>                inl in = in i .get\n                inl in' j = in' j i\n                inl out j = out j i\n<\/pre>\n<p>Inside the loop, the 1d tensor is indexed into and bound to <code>in<\/code>. But <code>in'<\/code> and <code>out<\/code> are wrapped around instead which has the effect of implicitly rotating them. It is the same as saying: \"apply the innermost dimension with <code>i<\/code> later.\"<\/p>\n<pre>                grid_for .y dim_in'_a {body=inl {i} -&gt;\n                    inl in', out = in' i, out i\n                    out.set (f in in'.get out.get)\n                    }\n<\/pre>\n<p>And that is what happens in the inner loop. The same <code>in<\/code> gets passed into <code>f<\/code> <code>dim_in'_a<\/code> number of times which has the effect of replicating it. Of course, often in Cuda programming it is not desirable to just replicate a tensor along some dimension as that would be wasteful. Instead the true benefit of having flexible kernel function is that they allow fusion which minimizes the number of intermediaries. As Cuda kernels are memory bound, that has an extreme positive effect on performance.<\/p>\n<p>Compilers themselves are utterly incapable of performing such optimizations on their own, so the responsibility for it falls on the user.<\/p>\n<p>That being said, the above kernel is a somewhat trivial example of this as only replicate and a map are fused, but it is possible to go significantly further.<\/p>\n<p>Here is the auxiliary for it.<\/p>\n<pre>inl d2_replicate_map w f in in' =\n    indiv join \n        inl in = zip in\n        inl in' =\n            match in' with\n            | by : int64 -&gt; \n                inl dim_in :: () = in.dim\n                HostTensor.create {elem_type=(); dim=by,dim_in}\n            | in' -&gt; zip in'\n        inl out = w.CudaTensor.create {elem_type=type f in.elem_type in'.elem_type; dim=in'.dim}\n        d2_replicate_map' w (inl a b _ -&gt; f a b) in in' out\n        stack out\n<\/pre>\n<p><code>in'<\/code> can also be a scalar value to indicate the size of the outermost dimension. Otherwise it is standard fare, it infers the type of the output, allocates it and passes it to the kernel, and strips the output from the mapping function.<\/p>\n<h5><a class=\"anchor\" aria-hidden=\"true\" href=\"#example\"><\/a>Example<\/h5>\n<pre>let kernel3 =\n    \"kernel3\",[cuda_modules],\"Does the d2_replicate_map kernel work?\",\n    \"\"\"\ninb s = CudaModules (1024*1024)\ninl inner_size = 8\ninl outer_size = 8\ninl h = HostTensor.init inner_size (const 123)\ninl h' = HostTensor.init (outer_size,inner_size) (inl a b -&gt; a,b)\ninl a1 = s.CudaTensor.from_host_tensor h\ninl a2 = s.CudaTensor.from_host_tensor h'\ninl o1 = s.CudaKernel.d2_replicate_map (inl a b -&gt; a, b) a1 a2\nTuple.iter s.CudaTensor.print (a1,a2,o1)\n    \"\"\"\n<\/pre>\n<pre>[|123; 123; 123; 123; 123; 123; 123; 123|]\n[|\n    [|[0, 0]; [0, 1]; [0, 2]; [0, 3]; [0, 4]; [0, 5]; [0, 6]; [0, 7]|]\n    [|[1, 0]; [1, 1]; [1, 2]; [1, 3]; [1, 4]; [1, 5]; [1, 6]; [1, 7]|]\n    [|[2, 0]; [2, 1]; [2, 2]; [2, 3]; [2, 4]; [2, 5]; [2, 6]; [2, 7]|]\n    [|[3, 0]; [3, 1]; [3, 2]; [3, 3]; [3, 4]; [3, 5]; [3, 6]; [3, 7]|]\n    [|[4, 0]; [4, 1]; [4, 2]; [4, 3]; [4, 4]; [4, 5]; [4, 6]; [4, 7]|]\n    [|[5, 0]; [5, 1]; [5, 2]; [5, 3]; [5, 4]; [5, 5]; [5, 6]; [5, 7]|]\n    [|[6, 0]; [6, 1]; [6, 2]; [6, 3]; [6, 4]; [6, 5]; [6, 6]; [6, 7]|]\n    [|[7, 0]; [7, 1]; [7, 2]; [7, 3]; [7, 4]; [7, 5]; [7, 6]; [7, 7]|]\n|]\n[|\n    [|[123, [0, 0]]; [123, [0, 1]]; [123, [0, 2]]; [123, [0, 3]]; [123, [0, 4]]; [123, [0, 5]]; [123, [0, 6]]; [123, [0, 7]]|]\n    [|[123, [1, 0]]; [123, [1, 1]]; [123, [1, 2]]; [123, [1, 3]]; [123, [1, 4]]; [123, [1, 5]]; [123, [1, 6]]; [123, [1, 7]]|]\n    [|[123, [2, 0]]; [123, [2, 1]]; [123, [2, 2]]; [123, [2, 3]]; [123, [2, 4]]; [123, [2, 5]]; [123, [2, 6]]; [123, [2, 7]]|]\n    [|[123, [3, 0]]; [123, [3, 1]]; [123, [3, 2]]; [123, [3, 3]]; [123, [3, 4]]; [123, [3, 5]]; [123, [3, 6]]; [123, [3, 7]]|]\n    [|[123, [4, 0]]; [123, [4, 1]]; [123, [4, 2]]; [123, [4, 3]]; [123, [4, 4]]; [123, [4, 5]]; [123, [4, 6]]; [123, [4, 7]]|]\n    [|[123, [5, 0]]; [123, [5, 1]]; [123, [5, 2]]; [123, [5, 3]]; [123, [5, 4]]; [123, [5, 5]]; [123, [5, 6]]; [123, [5, 7]]|]\n    [|[123, [6, 0]]; [123, [6, 1]]; [123, [6, 2]]; [123, [6, 3]]; [123, [6, 4]]; [123, [6, 5]]; [123, [6, 6]]; [123, [6, 7]]|]\n    [|[123, [7, 0]]; [123, [7, 1]]; [123, [7, 2]]; [123, [7, 3]]; [123, [7, 4]]; [123, [7, 5]]; [123, [7, 6]]; [123, [7, 7]]|]\n|]\n<\/pre>\n<h4><a id=\"user-content-mapi_d2_redo_map\" class=\"anchor\" aria-hidden=\"true\" href=\"#mapi_d2_redo_map\"><\/a>mapi_d2_redo_map<\/h4>\n<p>The inverse of the <code>d2_replicate<\/code> kernel. It reduces the outermost dimension of <code>in<\/code> - in other words it turns a tensor of dimensions <code>(a,b)<\/code> into <code>(b)<\/code>. It is equivalent to a transpose and then a reduce over the innermost dimension. An alternative to using this kernel would be to blast the target tensor with atomics in a regular map. That would work for some cases like in backwards step of <code>add_bias<\/code> at the cost of making training non-deterministic.<\/p>\n<pre>\/\/\/ Maps the two inputs and then reduces the first's outer dimension.\nmet mapi_d2_redo_map' w {d with redo neutral_elem} in in' out =\n    inl to_dev_tensor = w.CudaTensor.to_dev_tensor\n    \n    inl in, in', out = zip in, zip in', zip out\n    inl dim_in_a, dim_in_b = in.dim\n    inl dim_in' :: () = in'.dim\n    assert (dim_in' = dim_in_b) \"Input's inner dimension must equal the output's dimension.\"\n    assert (in'.dim = out.dim) \"Input and output's dimensions must be equal.\"\n    inl blockDimX = lit_min warp_size (s dim_in')\n    inl blockDimY = lit_min 32 (s dim_in_a)\n    inl gridDim = min 64 (divup (s dim_in') blockDimX)\n    inl in = to_dev_tensor in\n    inl in' = to_dev_tensor in'\n    inl out = to_dev_tensor out\n    inl map_out = match d with {map_out} -&gt; map_out | _ a _ _ -&gt; a\n    w.run {\n        gridDim\n        blockDim=blockDimX,blockDimY\n        kernel = cuda \n            inl grid_for = grid_for {blockDim gridDim}\n            grid_for .x dim_in_b {body=inl {i} -&gt;\n                inl in j = in j i\n                inl in' = in' i .get\n                inl out = out i\n                inl finally result = out.set (map_out result out.get)\n                inl state = \n                    grid_for .y dim_in_a {state=dyn neutral_elem; body=inl {state i=j} -&gt; \n                        inl in = in j\n                        match d with\n                        | {map_in} -&gt; redo state (map_in in.get in') \n                        | {mapi_in} -&gt; redo state (mapi_in i j in.get in') \n                        | _ -&gt; redo state in.get\n                        }\n                        \n                if blockDim.y &gt; 1 then\n                    inl near_to = blockDim.y\n                    inl ar = \n                        HostTensor.create {\n                            array_create=array_create_cuda_shared\n                            elem_type=state\n                            dim={from=1; near_to}, blockDim.x\n                            }\n                        |&gt; inl ar i -&gt; ar i threadIdx.x\n                    whilecd {\n                        state={near_to state}\n                        cond=inl {near_to} -&gt; near_to &gt;= 2\n                        body=inl {near_to state} -&gt;\n                            inl by = near_to\/2 \/\/ It might be worth trying `max 1 (near_to\/3)`\n                            if threadIdx.y &lt; near_to &amp;&amp; threadIdx.y &gt;= by then ar threadIdx.y .set state\n                            syncthreads()\n                            {\n                            near_to=by \n                            state=\n                                if threadIdx.y &lt; by then\n                                    forcd {from=threadIdx.y+by; by near_to state \n                                        body=inl {state i} -&gt; redo state (ar i .get)\n                                        }\n                                else\n                                    state\n                            }\n                        }\n                    |&gt; inl {state} -&gt; if threadIdx.y = 0 then finally state\n                else\n                    finally state\n            }\n        } |&gt; ignore\n<\/pre>\n<p>Because the Cub does not have a transposed reduction, the author had to implement it on his own.<\/p>\n<p>In order to understand what the kernel is doing, first it is required to visualize it. Imagine a 96x32 tensor and one block of size 32x32.<\/p>\n<p>What that block does is covers [0,31]x[0,31] and loads from main memory, does the reduction in place with the neutral element, slides to [32,63]x[0,31] and loads from main memory, does the reduction in place with the previous state, slides to [64,95]x[0,31] and loads from main memory, and does the final inplace reduction.<\/p>\n<pre>                inl state = \n                    grid_for .y dim_in_a {state=dyn neutral_elem; body=inl {state i=j} -&gt; \n                        inl in = in j\n                        match d with\n                        | {map_in} -&gt; redo state (map_in in.get in') \n                        | {mapi_in} -&gt; redo state (mapi_in i j in.get in') \n                        | _ -&gt; redo state in.get\n                        }\n<\/pre>\n<p>What was described above corresponds to this piece of code.<\/p>\n<p>There are now 32x32 threads in a block, each having their own <code>state<\/code> which after blockwise reduction will be reduced to 1x32 and then stored into the output tensor.<\/p>\n<pre>                if blockDim.y &gt; 1 then\n                    ...\n                else\n                    finally state\n<\/pre>\n<p>Obviously if the number of outermost threads is 1 then the tensor is already reduced and can be outputted.<\/p>\n<p>Otherwise a block wide reduction is performed.<\/p>\n<pre>                    inl near_to = blockDim.y\n                    inl ar = \n                        HostTensor.create {\n                            array_create=array_create_cuda_shared\n                            elem_type=state\n                            dim={from=1; near_to}, blockDim.x\n                            }\n                        |&gt; inl ar i -&gt; ar i threadIdx.x\n<\/pre>\n<p>It starts by defining a shared memory tensor. Immediately after creating it wrapped with a function that applies it with <code>threadIdx.x<\/code> along the innermost dimension. This simple technique of functional programming is greatly useful in facilitating reasoning.<\/p>\n<p>Whereas previously it was needed to reason about a 2d tensor, it now becomes effectively a 1d tensor and that other dimension can be assumed taken care of and left out of mind.<\/p>\n<p>Describing how to reduce a 1d tensor is not too complicated. Imagine the goal is to reduce a 1d tensor of size 8. At creation the tensor is uninitialized, so it will be described as thus.<\/p>\n<pre>01234567\n________\n<\/pre>\n<p>To reduce it in shared memory, the right half of the threads would need to write their states into it.<\/p>\n<pre>01234567\n____aaaa\n<\/pre>\n<p>Then a block synchronization is done and the left half of the threads (<code>0123<\/code>) reads from the tensor. They then performs the reduction with the <code>state<\/code> in their registers.<\/p>\n<pre>0123\n____\n<\/pre>\n<p>No need to synchronize here. <code>23<\/code> store their states into shared memory.<\/p>\n<pre>0123\n__bb\n<\/pre>\n<p>Then they synchronize. After that <code>01<\/code> read from <code>23<\/code> and perform the reduction with the state currently in their memory.<\/p>\n<pre>01\n__\n<\/pre>\n<p>The process is then repeated again. <code>1<\/code> stores...<\/p>\n<pre>01\n_c\n<\/pre>\n<p>...a block synchronization happens, and then <code>0<\/code> reads it and does the final reduction.<\/p>\n<pre>                    whilecd {\n                        state={near_to state}\n                        cond=inl {near_to} -&gt; near_to &gt;= 2\n                        body=inl {near_to state} -&gt;\n                            inl by = near_to\/2 \/\/ It might be worth trying `max 1 (near_to\/3)`\n                            if threadIdx.y &lt; near_to &amp;&amp; threadIdx.y &gt;= by then ar threadIdx.y .set state\n                            syncthreads()\n                            {\n                            near_to=by \n                            state=\n                                if threadIdx.y &lt; by then\n                                    forcd {from=threadIdx.y+by; by near_to state \n                                        body=inl {state i} -&gt; redo state (ar i .get)\n                                        }\n                                else\n                                    state\n                            }\n                        }\n<\/pre>\n<p>What was described above is what essentially happens in the code above. The action only happens along the <code>y<\/code> axis which makes reasoning easy. Tensor as an abstraction and stateful loops provide a definite benefit in terms reasoning. If the above was Cuda C code, then the programmer would be forced to reason about offsets, indices, mutation and various other things.<\/p>\n<pre>|&gt; inl {state} -&gt; if threadIdx.y = 0 then finally state\n<\/pre>\n<p>After the reduction is done, the <code>state<\/code> which is the final result is shipped of to be outputted to <code>finally<\/code>.<\/p>\n<pre>inl finally result = out.set (map_out result out.get)\n<\/pre>\n<p>It was bound to the environment around 40 lines ago.<\/p>\n<p>Before the explanation is done there also one subtle point that needs to be made about shared memory array sizing.<\/p>\n<pre>HostTensor.create {\n    array_create=array_create_cuda_shared\n    elem_type=state\n    dim={from=1; near_to}, blockDim.x\n    }\n<\/pre>\n<p>Based on the example shown...<\/p>\n<pre>01\n_c\n<\/pre>\n<p>0 never gets used so it should be clear why it was chosen that the outer dimension should start from 1. It is a nice way of preserving a pinch of memory.<\/p>\n<pre>01234567\n_cbbaaaa\n<\/pre>\n<p>But on closer examination the actual access pattern is like this.<\/p>\n<p>Would it not be possible to reuse more memory so as to require only 4 slots instead of 7?<\/p>\n<pre>01234567\n____aaaa\n____bb\n____c\n<\/pre>\n<p>Something like this? It would be possible, but it would increase the overall latency of the scheme because another call to synchronize would be needed between operations. Cuda makes very little guarantees about order of execution, so without an extra synchronize at the end what might happen is the <code>bb<\/code>s might be written before <code>aaaa<\/code>s finish reading.<\/p>\n<p>It is easier to just use an extra tad of memory to not have to deal with data races.<\/p>\n<h5><a id=\"user-content-example-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#example-1\"><\/a>Example<\/h5>\n<pre>let kernel4 =\n    \"kernel4\",[cuda_modules],\"Does the mapi_d2_redo_map' kernel work?\",\n    \"\"\"\ninb s = CudaModules (1024*1024)\ninl inner_size = 10\ninl outer_size = 4\ninl h = HostTensor.init (outer_size,inner_size) (inl _ x -&gt; x)\ninl h' = HostTensor.init inner_size id\ninl a1 = s.CudaTensor.from_host_tensor h\ninl a2 = s.CudaTensor.from_host_tensor h'\ninl o = \n    s.CudaKernel.mapi_d2_redo_map {\n        map_in=(+)\n        neutral_elem=0; redo=(+)\n        } a1 a2\nTuple.iter s.CudaTensor.print (a1,a2,o)\n    \"\"\"\n<\/pre>\n<pre>[|\n    [|0; 1; 2; 3; 4; 5; 6; 7; 8; 9|]\n    [|0; 1; 2; 3; 4; 5; 6; 7; 8; 9|]\n    [|0; 1; 2; 3; 4; 5; 6; 7; 8; 9|]\n    [|0; 1; 2; 3; 4; 5; 6; 7; 8; 9|]\n|]\n[|0; 1; 2; 3; 4; 5; 6; 7; 8; 9|]\n[|0; 8; 16; 24; 32; 40; 48; 56; 64; 72|]\n<\/pre>\n<p>It should be noted that the reduction kernel is also capable of replicating one of its arguments during the first map phase.<\/p>\n<h4><a id=\"user-content-mapi_d1_seq_broadcast\" class=\"anchor\" aria-hidden=\"true\" href=\"#mapi_d1_seq_broadcast\"><\/a>mapi_d1_seq_broadcast<\/h4>\n<p>This kernel does repeat reduction over items in registers along the inner dimension. It seems like a beast at first, but is actually simpler than the previous one since there is no need to understand how reduction works.<\/p>\n<pre>\/\/ Repeatedly reduces along the inner dimension and then maps the result of that reductions over the input in the previous step.\nmet mapi_d1_seq_broadcast' w {d with seq} in out = \n    inl to_dev_tensor = w.CudaTensor.to_dev_tensor\n    \n    inl in, out = zip in, zip out\n    inl dim_in_a, dim_in_b = in.dim\n    assert (in.dim = out.dim) \"The input and the output dimensions need to be equal\"\n    inl num_valid = s dim_in_b\n    inl items_per_thread, blockDim =\n        assert (lit_is num_valid) \"The inner dimension of the input to this kernel must be known at compile time.\"\n        if num_valid &lt;= 1024 then 1, num_valid\n        else divup num_valid 256, 256\n    inl gridDimY = min 64 (s dim_in_a)\n    inl in = to_dev_tensor in\n    inl out = to_dev_tensor out\n<\/pre>\n<p>This kernel will be unique so far in that it actually takes advantage of Spiral's significant ability to do local memory allocation. As the elements of the inner dimension must be known at compile time the kernel with raise a type error if that dimension is not known at compile time.<\/p>\n<pre>    inl items_per_thread, blockDim =\n        assert (lit_is num_valid) \"The inner dimension of the input to this kernel must be known at compile time.\"\n        if num_valid &lt;= 1024 then 1, num_valid\n        else divup num_valid 256, 256\n<\/pre>\n<p>As usual, the number 256 is picked off the cuff and probably different problems will have different ideal best values, but for ML purposes this setting should suffice for the time being.<\/p>\n<p>The actual kernel will be explained piece by piece as it is too much to take in at once.<\/p>\n<pre>    w.run {\n        blockDim\n        gridDim=1,gridDimY\n        kernel = cuda \n            inl dims = {blockDim gridDim}\n            grid_for dims .y dim_in_a {body=inl {i=j} -&gt;\n                inl in, out = in j, out j\n<\/pre>\n<p>Since kernel is in essence d1 reduction it begins by iterating over the outer dimension.<\/p>\n<pre>                \/\/\/ Creates the tensor of items.\n                inl create_items elem_type = HostTensor.create {\n                    array_create = array_create_cuda_local\n                    layout=.aot\n                    elem_type\n                    dim=items_per_thread\n                    }\n<\/pre>\n<p>It is immediately followed by this function declaration. Notice that it is done in array of tuples format which is not the default for Spiral's tensors. The reason for that is that the Cub functions which accept arrays cannot accept Spiral's tensors in their default form.<\/p>\n<pre>inl inner_loop = grid_for_items dims .x dim_in_b\n<\/pre>\n<p>Just like <code>create_items<\/code> is often used, so is iterating over the inner dimension so the <code>grid_for_items<\/code> is partially applied for that purpose.<\/p>\n<p>The way this kernel is used is that it does the initial <code>map_in<\/code> and then iterates over a tuple of <code>{map_in redo map_out}<\/code> executing each in turn.<\/p>\n<pre>                inl items = \n                    inl map = \n                        match d with\n                        | {map_in} i -&gt; map_in (in i .get)\n                        | {mapi_in} i -&gt; mapi_in j i (in i .get)\n                        | _ i -&gt; in i .get\n                    inl items = create_items (type map (dyn 0))\n                    inner_loop {body=inl {item i} -&gt; items item .set (map i)}\n                    items\n<\/pre>\n<p>Having to match on whether the mapping function is <code>map_in<\/code> or <code>mapi_in<\/code> or not there at all adds a lot of noise for the reader.<\/p>\n<p>But what it does is create a tensor in register memory, iterates over the input here and sets the tensor to the result of mapping the input over the function. It is a map operation just as the name implies.<\/p>\n<p>The <code>items<\/code> at the end is the tensor that has gone through the first phase the kernel.<\/p>\n<pre>                inl rec seq_loop items (d :: d') =\n                    match d with\n                    | {map_in} -&gt; \n                        inl items' = create_items (type map_in items.elem_type)\n                        inner_loop {body=inl {item} -&gt; items item .get |&gt; map_in |&gt; items' item .set}\n                        items'.bodies.ar\n                    | {mapi_in} -&gt;\n                        inl items' = create_items (type mapi_in j i items.elem_type)\n                        inner_loop {body=inl {item i} -&gt; items item .get |&gt; mapi_in j i |&gt; items' item .set}\n                        items'.bodies.ar\n                    | _ -&gt; items.bodies.ar\n                    |&gt; inl x -&gt; \n<\/pre>\n<p>This is the function that iterates through the sequence. First, if the <code>map_in<\/code> or <code>mapi_in<\/code> are present they get are applied to every element in the items tensor.<\/p>\n<pre>                        inl block_reduce redo = \n                            inl d = {blockDim redo}\n                            if num_valid % blockDim.x = 0 then cub_block_reduce d\n                            else cub_block_reduce {d with num_valid} \n                        match d with\n                        | {redo} -&gt; block_reduce redo x |&gt; broadcast_zero\n                        | {redo'} -&gt; block_reduce redo' x\n                    |&gt; inl x -&gt;\n<\/pre>\n<p>Then comes reduction. If the number of items does not exactly align with the block size, then a guarded reduction method is selected to make sure it does not read past the boundary. Also after the reduction is done there are two choices:<\/p>\n<ol><li><code>redo<\/code> does the reduction and the broadcasts the element to every thread in the block using shared memory.<\/li>\n<\/ol><pre>inl broadcast_zero x =\n    inl ar = array_create_cuda_shared x 1\n    if threadIdx.x = 0 then ar 0 &lt;- x\n    syncthreads()\n    ar 0\n<\/pre>\n<p>The first thread which holds the reduced elements first writes to a shared memory location and then all the elements read from it after synchronizing.<\/p>\n<ol start=\"2\"><li><code>redo'<\/code> functions much like the above, but without the broadcast.<\/li>\n<\/ol><p>After that comes the final step. There are also two branching paths at this point.<\/p>\n<pre>                        match d' with\n                        | () -&gt; \n                            inner_loop {body=inl {item i} -&gt;\n                                inl out = out i\n                                match d with\n                                | {map_out} -&gt; map_out (items item .get) x (out .get)\n                                | {mapi_out} -&gt; mapi_out j i (items item .get) x (out .get)\n                                |&gt; out .set\n                                }\n<\/pre>\n<p>If the current element in the sequence is last, then the result of applying <code>map_out<\/code> or <code>mapi_out<\/code> to each element of <code>items<\/code> is feed directly into the output tensor.<\/p>\n<pre>                        | _ -&gt;\n                            match d with\n                            | {map_out} -&gt; \n                                inl items' = create_items type map_out items.elem_type x\n                                inner_loop {body=inl {item} -&gt; map_out (items item .get) x |&gt; items' item .set}\n                                seq_loop items' d'\n                            | {mapi_out} -&gt; \n                                inl items' = create_items type mapi_out (dyn 0) (dyn 0) items.elem_type x\n                                inner_loop {body=inl {i item} -&gt; mapi_out j i (items item .get) x |&gt; items' item .set}\n                                seq_loop items' d'\n<\/pre>\n<p>Otherwise, it needs to create a new intermediate tensor and then sets the result of applying <code>map_out<\/code> or <code>mapi_out<\/code> to each element of the <code>items<\/code> tensor to it. Then it passes that as input to the next operation in the sequence.<\/p>\n<h5><a id=\"user-content-the-softmax-activation\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-softmax-activation\"><\/a>The Softmax Activation<\/h5>\n<p>The softmax activation will be covered here and the more complex layer normalization activation will be covered in the following chapter.<\/p>\n<pre>inl softmax x = exp x \/ replicate (sum (exp x))\n<\/pre>\n<p>The above is softmax in pseudo-code. <code>x<\/code> is a vector type.<\/p>\n<p>Alternatively, it would be better to start with this...<\/p>\n<pre>inl softmax x = \n    inl z = exp x\n    z \/ replicate (sum z)\n<\/pre>\n<p><code>sum<\/code> sums up all the values of a vector into a scalar, and replicate unfolds the scalar into a vector each holding the same values.<\/p>\n<p>Furthermore, for numerical stability the following trick is often employed in softmax.<\/p>\n<pre>inl softmax x = \n    inl max_x = replicate (max x)\n    inl z = exp (x - max_x)\n    z \/ replicate (sum z)\n<\/pre>\n<p>This has no effect on the way gradients are propagated during the backwards step. Derivative of  <code>x - max_x<\/code> with respect to <code>x<\/code> is just 1. <code>max_x<\/code> is considered to be a constant. Shifting the input to softmax by a constant does not affect the result during stable regimes, but if the inputs are unstable then they will trend towards 0 rather that towards infinity.<\/p>\n<p>Going forward, it is important to keep in mind that <code>max<\/code> and <code>sum<\/code> are reduce operations, and <code>exp<\/code> is a map operation. The kernel described in this chapter makes it possible to succinctly implement both softmax's forward and backward phases.<\/p>\n<p>Here is the forward phase.<\/p>\n<pre>inl z =\n    s.CudaKernel.mapi_d1_seq_broadcast {\n        seq = \n            {\n            redo=max \/\/ max x\n            map_out=inl x replicate_max_x -&gt; exp (x - replicate_max_x)\n            }\n            ,\n            {\n            redo=(+) \/\/ sum z\n            map_out=inl z replicate_sum_z -&gt; z \/ replicate_sum_z\n            }\n        } (primal x)\n<\/pre>\n<p>The second argument in <code>map_out<\/code> is implicitly replicated. That is the softmax forward.<\/p>\n<p>Assuming the output is present the backward step can be done in one step.<\/p>\n<pre>inl softmax x = exp x \/ replicate (sum (exp x))\n<\/pre>\n<p>For ease of rewriting the sum will be factored out...<\/p>\n<pre>inl softmax x = \n    inl s = replicate (sum (exp x))\n    exp x \/ s\n<\/pre>\n<p>There are <a href=\"https:\/\/eli.thegreenplace.net\/2016\/the-softmax-function-and-its-derivative\/\" rel=\"nofollow\">tutorials<\/a> that show how to take the derivative analytically, but for show it will be done here by emulating how automatic differentiation would take the trace through the above. The analysis will have the benefit of not needing to branch on the value of the index used.<\/p>\n<p>Here are a few simple rules.<\/p>\n<pre>o = a \/ b\nda += do \/ b\ndb += -do * da \/ (b * b)\n<\/pre>\n<p>This is the AD rule for propagating adjoints (gradients) through division.<\/p>\n<pre>o = replicate x\ndx += sum do\n<\/pre>\n<p>This is the rule for propagating gradients through <code>replicate<\/code>. It is a sumation.<\/p>\n<pre>o = sum x\ndx += replicate do\n<\/pre>\n<p>The rule for propagating gradients through summation is the opposite of the previous one.<\/p>\n<pre>o = exp x\ndx += do * exp x\n<\/pre>\n<p>With this last rule, we are all set to trace the derivative of the softmax with respect to its input.<\/p>\n<pre>inl v = exp x\ninl s = sum v\ninl r = replicate s\ninl z = v \/ r\n<\/pre>\n<p>This is just the softmax split into individual pieces similarly to how the compiler would do let insertion.<\/p>\n<pre>inl z = v \/ r\n<\/pre>\n<p>Starting with this, the gradients are propagated through <code>v<\/code> and <code>r<\/code>.<\/p>\n<pre>\/\/ inl z = v \/ r\ndv += dz \/ r\ndr += -dz * v \/ (r * r)\n<\/pre>\n<p>AD is sensitive to the ordering of operations. The backwards pass needs to be done exactly in the opposite way of forward pass. This is why a tape is frequently used for it.<\/p>\n<pre>\/\/inl r = replicate s\nds += sum dr\n<\/pre>\n<p>...<\/p>\n<pre>\/\/inl s = sum v\ndv += replicate ds\n<\/pre>\n<p>...<\/p>\n<pre>\/\/ inl v = exp x\ndx += dv * v\n<\/pre>\n<p>Here are all the steps on the same page.<\/p>\n<pre>dv += dz \/ r\ndr += -dz * v \/ (r * r)\nds += sum dr\ndv += replicate ds\ndx += dv * v\n<\/pre>\n<p>A naive AD implementation would do the above in 5 different steps. The goal of making Spiral was to allow fusion of the above steps into a single one. To start things off <code>dr<\/code> will be folded into <code>ds<\/code> which will then be folded into <code>dv<\/code>.<\/p>\n<pre>dv += dz \/ r\ndv += replicate (sum (-dz * v \/ (r * r)))\ndx += dv * v\n<\/pre>\n<p>Both of <code>dv<\/code>s can now be folded into <code>dx<\/code>.<\/p>\n<pre>dx += (dz \/ r + replicate (sum (-dz * v \/ (r * r)))) * v\n<\/pre>\n<p>Since <code>z = v \/ r<\/code> the above can be simplified further.<\/p>\n<pre>dx += (dz \/ r + replicate (sum (-dz * z \/ r))) * v\n<\/pre>\n<p>A valid operation here would be <code>sum (a * replicate b) = sum a \/ b<\/code>. <code>r<\/code> is <code>replicate s<\/code>. It is possible to take it out a level.<\/p>\n<pre>dx += (dz \/ r + replicate (sum (-dz * z) \/ s)) * v\n<\/pre>\n<p>A valid operation here would be <code>replicate (a * b) = replicate a * replicate b<\/code>.<\/p>\n<pre>dx += (dz \/ r + replicate (sum (-dz * z)) \/ replicate s) * v\n<\/pre>\n<p><code>r = replicate s<\/code> as per original definition.<\/p>\n<pre>dx += (dz \/ r + replicate (sum (-dz * z)) \/ r) * v\n<\/pre>\n<p>The above can be simplified using the distributive law.<\/p>\n<pre>dx += (dz + replicate (sum (-dz * z))) * v \/ r\n<\/pre>\n<p>Since <code>z = v \/ r<\/code> the above can be simplified further.<\/p>\n<pre>dx += (dz + replicate (sum (-dz * z))) * z\n<\/pre>\n<p>Since the <code>-<\/code> is a constant it is possible to factor it out.<\/p>\n<pre>dx += (dz - replicate (sum (dz * z))) * z\n<\/pre>\n<p>In this simplified form, all the steps can easily be implemented in a single kernel. <code>er<\/code> is <code>replicate (sum (dz * z))<\/code><\/p>\n<pre>s.CudaKernel.mapi_d1_seq_broadcast' {\n    seq = \n        {\n        map_in=inl z,dz -&gt; z*dz\n        redo=(+)\n        map_out=inl (z,dz) er dx -&gt; dx + (dz - er) * z\n        }\n    } (primal z, adjoint z) (adjoint x)\n<\/pre>\n<h3><a id=\"user-content-9-deep-learning-basics\" class=\"anchor\" aria-hidden=\"true\" href=\"#9-deep-learning-basics\"><\/a>9: Deep Learning Basics<\/h3>\n<p>At the time of writing this it is almost April 2018.<\/p>\n<p>Currently, the two dominant deep learning frameworks are without exception Tensorflow by Google and PyTorch by Facebook. Both are written in C++ and expose a Python front end from which the users access it. It is by no means unusual that big corporate sponsored frameworks would be widely used, but it is interesting to look at the deep learning ecosystems in other languages, or specifically the lack of such ecosystems. What is precisely interesting is that they are non-existent and most of the attempts at making frameworks have petered out.<\/p>\n<p>In a broader sense, what ML capabilities non-Python ecosystems have tend to be on the shallower end of learning which does not need the extreme GPU capabilities Spiral has.<\/p>\n<p>Deep learning has been popular for half a decade at least now, and had it been possible to make a great DL library in say F# (.NET), Scala (JVM), Haskell or Racket, then one can be sure that this would already have happened. The reason for that is that languages without first class staging have very deep flaws that preclude them from having effective Cuda backends and GPU abstractions. There is very little languages constructed in the classical manner whether they be static or dynamic offer in the making of a deep learning library.<\/p>\n<p>Statically typed languages have type systems that are simply too weak to be useful and dynamically typed languages are simply too inefficient. And Lisp macros are not a substitute for first class staged functions and join points.<\/p>\n<p>Hence it is easy to predict that great libraries will not get made in any of the aforementioned languages nor in any of the mainstream languages that aren't C++. The effort is simply too great and too strenuous in them. At best, .NET for example might get bindings for Tensorflow or CNTK if it has not already.<\/p>\n<p>And though Tensorflow, CNTK and PyTorch are written in C++ that is hardly a beaming recommendation for the language. C++ is widely known enough that its quality as a language can stand for itself.<\/p>\n<p>Making a machine learning library is hard enough to crush most langauges, but Spiral can make it a lot easier and the hows of that is what will be covered in this chapter. That being said, this is a language tutorial and not a deep learning tutorial so what won't be covered is to how to actual use deep learning to do interesting things. Instead it will be about library construction.<\/p>\n<h4><a id=\"user-content-primitives\" class=\"anchor\" aria-hidden=\"true\" href=\"#primitives\"><\/a>Primitives<\/h4>\n<p>At the time of writing the Spiral <code>Learning<\/code> library is quite small, but it has some advanced capabilities so this is the ideal point to introduce it before it gets any bigger. The one limitation it has on it is that it is restricted to first order <a href=\"https:\/\/en.wikipedia.org\/wiki\/Automatic_differentiation\" rel=\"nofollow\">automatic differentiation<\/a>. It would be possible to build a library that supports higher order AD in Spiral, but decision to support higher order AD is not a trivial one to make.<\/p>\n<p>It would add very non-negligible maintenance burden on it, and it possibly make it difficult to provide effective optimization. Higher order AD does not have a particularly good track record on practical use in deep learning anyway.<\/p>\n<p>For those not familiar with it, higher order AD would essentially allow adaptive learning methods and possibly metalearning methods like MAML, but while first order AD takes only a single pass per datapoint, higher order AD would require <code>n<\/code> passes per datapoint where the <code>n<\/code> is the number of parameters in the network. And approximated higher order methods require the whole dataset to be operated at once and do not allow minibatching.<\/p>\n<p>Spiral's ML library is intended to be of great practical use. It needs to be simple to extend and understand. It needs to be both highly flexible and performant.<\/p>\n<p>So the only dual number it uses internally is of the primal and adjoint variety.<\/p>\n<pre>inl float -&gt;\n    \/\/ #Primitives\n    inl zero = to float 0\n    inl one = to float 1\n    inl two = to float 2\n    inl infinity =\n        match float with\n        | _: float32 -&gt; infinityf32\n        | _: float64 -&gt; infinityf64\n    inl primal = function {primal} | primal -&gt; primal\n    inl adjoint = function {adjoint} -&gt; adjoint | _ -&gt; ()\n    inl primals = Struct.map primal\n    inl adjoints = Struct.map adjoint\n    inl on_non_nil B ret =\n        match B with\n        | () -&gt; ()\n        | B -&gt; ret B\n    inl dr s primal = {primal adjoint=s.CudaTensor.zero_like primal; block=()}\n<\/pre>\n<p>The library takes the type of the floating point number it is operating on as its first argument. This is generally used to convert literals to <code>float32<\/code> values at compile time and such since Spiral does not allow implicit conversions. <code>zero<\/code>, <code>one<\/code>, <code>two<\/code> and <code>infinity<\/code> are just some constants set up for convenience.<\/p>\n<p><code>Struct.map<\/code> is just the <code>toa_map<\/code> from the tensor chapter. Since functions that iterate over arbitrary types are so often used in Spiral, they have their own <code>Struct<\/code> module now.<\/p>\n<p><code>dr<\/code> takes in context and the primal and creates a dual with the adjoint set to zero.<\/p>\n<p><code>on_non_nil<\/code> is just a utility function.<\/p>\n<h5><a id=\"user-content-map-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#map-1\"><\/a>map<\/h5>\n<p>Map is useful for activation functions and pointwise operations like addition and Hadamarad multiplication.<\/p>\n<pre>    inl choose_adjoints in bck =\n        Struct.choose2 (function\n            | {primal adjoint} bck -&gt; {adjoint bck block=()}\n            | _ _ -&gt; ()) in bck\n        |&gt; inl x -&gt; Struct.map (inl x -&gt; x.adjoint) x, Struct.map (inl x -&gt; x.bck) x\n            \n    inl map {fwd bck} in s =\n        inl primal = primals in\n        inl out = s.CudaKernel.map fwd primal |&gt; dr s\n        inl adjoint, bck = choose_adjoints in bck\n        out, inl _ -&gt; join\n            inl bck (in, out) = Struct.map2 (inl bck -&gt; bck (in, out)) bck\n            s.CudaKernel.map' bck (primal, {out without block}) adjoint\n<\/pre>\n<p>In order to be flexible, while <code>fwd<\/code> is always a single function, <code>bck<\/code> is intended to be the same as the number of inputs. The reason for that is that not all variables will be dual numbers and hence the gradients won't be propagated to them. Example of such variables would the inputs and the labels for the cost function.<\/p>\n<p>Here is an example of how the <code>map<\/code> primitive can be used to implement activations.<\/p>\n<pre>    inl activation d = map {d with bck = Struct.map (inl bck (in, out) adjoint -&gt; adjoint + out.adjoint * (self in out.primal)) self}\n    inl sigmoid_fwd x = one \/ (one + exp -x)\n    inl sigmoid_bck out = out * (one - out)\n    inl sigmoid = activation {\n        fwd = sigmoid_fwd\n        bck = inl _ -&gt; sigmoid_bck\n        }\n<\/pre>\n<p>The <code>activation<\/code> is there to make sure that in the backward step the adjoint is added to and that the adjoint of <code>out<\/code> multiplies whatever flows out from the function.<\/p>\n<p>The above could also be implemented like.<\/p>\n<pre>inl sigmoid = map {\n    fwd = inl x -&gt; one \/ (one + exp -x)\n    bck = inl (in,out) adjoint -&gt; adjoint + out.adjoint * (out.primal * (one - out.primal))\n    }\n<\/pre>\n<p>Though pretty much always, it is instead preferable to abstract what is possible in order to compress the code size, increase readability and minimize the chance of error. In this case this piece of advice only refers to the <code>activation<\/code> function. <code>sigmoid_fwd<\/code> and <code>sigmoid_bck<\/code> have been factored out due to being used in multiple places.<\/p>\n<p>The following piece deserves a mention as it is the way AD is done.<\/p>\n<pre>        out, inl _ -&gt; \n            inl bck (in, out) = Struct.map2 (inl bck -&gt; bck (in, out)) bck\n            s.CudaKernel.map' bck (primal, {out without block}) adjoint\n<\/pre>\n<p>All the primitive functions return the output and the function to run in order to trigger the backward step. At the time the output is made, the adjoint of it is always zero, but after the forward pass is done the adjoint at the top of tape is set to 1 and propagated downwards and eventually through that output.<\/p>\n<p>As a short tutorial, remember the various rules used to take the derivatives of softmax. Here are the rules for division once again. It is the same for the matrix and the scalar case if one assumes the operators are overloaded.<\/p>\n<pre>c = a \/ b\nda += dc \/ b\ndb += -dc * a \/ (b * b)\n<\/pre>\n<p>Here is how this could be implemented in code in Spiral.<\/p>\n<pre>inl div a b =\n    inl primal = primal a * primal b\n    inl adjoint = ref zero\n    {primal adjoint block=()}, inl _ -&gt;\n        adjoint a := adjoint a + adjoint \/ primal b\n        adjoint b := adjoint b - adjoint * primal a \/ (primal b * primal b)\n<\/pre>\n<p>The above assumes that <code>a<\/code> and <code>b<\/code> are scalars rather than tensors. Regardless, the structure is quite similar to the <code>map<\/code> which is intended to be a generic interface for such functions.<\/p>\n<p>Being able to support aggressive inlining, Spiral is well suited for making AD libraries. That being said, this particular aspect of it - the ability to always inline the backward step is not particularly important for deep learning. Deep learning deals with large batched operations - a single matrix multiply is already on the factor of thousands scalar operations and what would take extreme optimization heroics by the compiler is generally put in by hand instead for such batched operations.<\/p>\n<p>What the <code>map<\/code> really offers is easy fusion of map operations.<\/p>\n<pre>    inl hadmult = activation {\n        fwd = inl a,b -&gt; a*b\n        bck = (inl (_,x) _ -&gt; x), (inl (x,_) _ -&gt; x)\n        }\n<\/pre>\n<p>Here is the standard Hadamarad multiplication. Should an operation like <code>a*b + c*d<\/code> be needed then it would be trivial to make a fresh activation.<\/p>\n<pre>    inl hadmult2 = activation {\n        fwd = inl a,b,c,d -&gt; a*b + c*d\n        bck = \n            inl (_,b,_,_) _ -&gt; b\n            ,inl (a,_,_,_) _ -&gt; a\n            ,inl (_,_,_,d) _ -&gt; d\n            ,inl (_,_,c,_) _ -&gt; c\n        }\n<\/pre>\n<p>Admittedly, this example a bit unergonomic, but it is a vast improvement over having to write a separate function in Cuda for this sort of thing. It would not be difficult to make a more generic Hadamarad multiplication that looks something like this.<\/p>\n<pre>    inl hadmult' = activation' {\n        fwd = Tuple.map (inl (a,b) -&gt; a*b) &gt;&gt; Tuple.foldl (+) zero\n        bck = inl x _ -&gt; Tuple.map (inl (a,b) -&gt; (b,a)) x\n        }\n<\/pre>\n<p>The exercise for this will be left to the reader.<\/p>\n<p>Being able to do such generic operations so directly is wonderful. In one sweep it is possible to cover an infinite space of possible map operations. And somewhat paradoxically, it is easier to prove generic code correct than it is specific instances of it. With this capability there is no need to worry whether one of the numerous float pointers are passed in the right order, dumb copy paste errors or array boundary violations which Cuda will not check for.<\/p>\n<p>Spiral's ML library is intended to cover the middle ground between what mainstream frameworks offer and what naive AD implementations offer.<\/p>\n<h5><a id=\"user-content-map_redo_map-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#map_redo_map-1\"><\/a>map_redo_map<\/h5>\n<p>This one is used for the cost functions.<\/p>\n<pre>    \/\/\/ Does not return a `dr` unlike the rest. This is an optimization in order to avoid having to call too many useless kernels that \n    \/\/\/ just to set the adjoint to 1. The current library is intended for a narrow purpose.\n    inl map_redo_map {fwd bck} in s =\n        inl primal = primals in\n        inl out = s.CudaKernel.map_redo_map fwd primal\n        inl adjoint, bck = choose_adjoints in bck\n        out, inl _ -&gt; join\n            inl out = s.CudaTensor.to_dev_tensor out\n            inl bck in = Struct.map2 (inl bck -&gt; bck (in, out.get)) bck\n            s.CudaKernel.map' bck primal adjoint\n<\/pre>\n<p>If the previous section was understood, then this one should not give any trouble. Here are the examples of two cost functions done using it and the generic template for them.<\/p>\n<pre>    inl error {fwd bck} label input s = \n        inl batch_size = primal input .span_outer |&gt; to float\n        inl div_by_minibatch_size x = x \/ batch_size\n        map_redo_map {\n            fwd = {\n                map_in = fwd\n                redo = (+)\n                neutral_elem = zero\n                map_out = div_by_minibatch_size\n                }\n            \/\/\/ The adjoint in error is always assumed to be one.\n            bck = Struct.map (inl bck (in, out) adjoint -&gt; adjoint + div_by_minibatch_size (bck in out)) bck\n            } (input, label) s\n    inl square = error {\n        fwd = inl (x,y) -&gt; (y - x) * (y - x)\n        bck = \n            inl (x,y) _ -&gt; two * (x - y)\n            ,inl (x,y) _ -&gt; -(two * (x - y))\n        }\n    inl cross_entropy = error {\n        fwd = inl x, y -&gt; -(y * log x + (one - y) * log (one - x))\n        bck = \n            inl (x, y) _ -&gt; (x - y) \/ (x * (one - x))\n            ,inl (x, y) _ -&gt; log (one - x) - log x\n        }\n<\/pre>\n<p>One design decision that needs to be highlighted is that the adjoint for the cost function is not seeded with one. Instead it is implicitly assumed to be that in the backward step. This is so recurrent nets do not have to make numerous redundant calls just to set the cost to one. One thing the <code>error<\/code> does as a service is automatically divide by the batch size.<\/p>\n<p>Apart from the two above, the library also has the softmax cross entropy which has a more elaborate implementation due to how its backward step works so it won't be shown here. As a matter of fast, that particular implementation is not fully fused yet and that bit of work will be left for the future.<\/p>\n<p>There is one more primitive to be covered before the tutorial can proceed to the next step.<\/p>\n<h5><a id=\"user-content-d2_replicate_map-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#d2_replicate_map-1\"><\/a>d2_replicate_map<\/h5>\n<pre>    inl d2_replicate_map {fwd bck={bck_in bck_in'}} in in' s =\n        inl primal, adjoint = primals in, adjoints in\n        inl primal', adjoint' = primals in', adjoints in'\n        inl out = s.CudaKernel.d2_replicate_map fwd primal primal' |&gt; dr s\n        out, inl _ -&gt; join\n            inl out = {out without block}\n            s.CudaKernel.mapi_d2_redo_map' bck_in (primal', out) primal adjoint\n            s.CudaKernel.d2_replicate_map' bck_in' primal (primal', out) adjoint'\n<\/pre>\n<p>This is is used for adding bias terms. It can also be used to implement multiplicative integration in a very easy manner.<\/p>\n<p>In order to actually be suitable for implementing activations of that sort, a more focused helper method is derived from it.<\/p>\n<pre>    inl d2_replicate_activation {fwd bck_in bck_in'} in =\n        inl neutral_elem = Struct.map (const zero) in\n        inl bck = {\n            bck_in={\n                map_in=inl (in', out) in -&gt; Struct.map ((*) out.adjoint) (bck_in in in' out.primal)\n                neutral_elem redo=Struct.map2 (+)\n                map_out=Struct.map2 (+)\n                }\n            bck_in'=inl in (in', out) -&gt; Struct.map2 (inl x adjoint -&gt; adjoint + out.adjoint*x) (bck_in' in in' out.primal)\n            }\n        d2_replicate_map { fwd bck } in\n<\/pre>\n<p>This one does the important part of multiplying the backwards returns by the output adjoint and adds to the input adjoint. It is similar to the standard <code>activation<\/code> shown in the previous section.<\/p>\n<p>An example of it in use will be shown in the multiplicative integration section.<\/p>\n<h4><a id=\"user-content-optimizers\" class=\"anchor\" aria-hidden=\"true\" href=\"#optimizers\"><\/a>Optimizers<\/h4>\n<p>Only these two are here for the time being.<\/p>\n<pre>    inl sgd learning_rate s {primal adjoint} = \n        s.CudaKernel.map' (inl _ P, A -&gt; P - learning_rate * A, zero) primal.empty (primal, adjoint)\n    inl clipped_sgd max learning_rate s {primal adjoint} = \n        s.CudaKernel.map' (inl _ P, A -&gt; \n            inl A = if A &lt; -max then -max elif A &gt; max then max else A\n            P - learning_rate * A, zero\n            ) primal.empty (primal, adjoint)\n<\/pre>\n<p>It is possible to do a frightening amount of stuff for deep learning with just a simple map. It covers all of the first order optimizers for one. At any rate, the implementation of optimizers is trivialized by the generic map and extending the library with Adam and RMSProp or whatever else would not be a problem.<\/p>\n<p>One thing that the library does not do, but should is to fuse all the map calls into a single one during the optimization phase. That would not be hard in Spiral and will be subject of future work.<\/p>\n<h4><a id=\"user-content-operations\" class=\"anchor\" aria-hidden=\"true\" href=\"#operations\"><\/a>Operations<\/h4>\n<pre>    inl apply_bck bck bck' _ = bck'(); bck()\n    inl (&gt;&gt;=) a b s =\n        inl a,a_bck = a s\n        inl b,b_bck = b a s\n        b, apply_bck a_bck b_bck\n    inl succ x _ = x, const ()\n<\/pre>\n<p>These are used to combine AD operations in a monadic manner. <code>&gt;&gt;=<\/code> is just a specific instance of a writer monad here.<\/p>\n<p>In essence, it relives the burden of having to deal with backwards step functions explicitly in a pure functional fashion. Remember that <code>inm a = b<\/code> is just syntax sugar for <code>b &gt;&gt;= inl a -&gt; ...<\/code><\/p>\n<pre>                    inm f = matmultb ((i,input.f),(h,state.f)) bias.f &gt;&gt;= sigmoid\n                    inm i' = matmultb ((i,input.i),(h,state.i)) bias.i &gt;&gt;= sigmoid\n                    inm o = matmultb ((i,input.o),(h,state.o)) bias.o &gt;&gt;= sigmoid\n                    inm c' = matmultb ((i,input.c),(h,state.c)) bias.c &gt;&gt;= tanh\n                    inm c =\n                        inm x1 = hadmult (f,c)\n                        inm x2 = hadmult (i',c')\n                        add (x1, x2)\n                    inm h' = tanh c\n                    inm h = hadmult (o, h')\n                    succ (h, (h, c))\n<\/pre>\n<p>Here is how the LSTM is implemented for example. Monadic computation saves a decent bit of boilerplate here. That being said, this is an example of a heavily unoptimized implementation of a LSTM. An optimized example would have only one or two steps rather than...21.<\/p>\n<h4><a id=\"user-content-layers\" class=\"anchor\" aria-hidden=\"true\" href=\"#layers\"><\/a>Layers<\/h4>\n<h5><a id=\"user-content-multiplicative-integration-rnn\" class=\"anchor\" aria-hidden=\"true\" href=\"#multiplicative-integration-rnn\"><\/a>Multiplicative Integration RNN<\/h5>\n<p>Finally in this section it becomes possible to give an example of the <code>d2_replicate<\/code> activation. This RNN variant is from the paper <a href=\"https:\/\/arxiv.org\/abs\/1606.06630\" rel=\"nofollow\">On Multiplicative Integration with Recurrent Neural Networks<\/a>.<\/p>\n<p>In 2016 the author was impressed by the novel notion of replacing addition with Hadamarad multiplication and decided there that this is something he wanted to use since getting a good improvement for such a simple change seemed like a worthwhile tradeoff. It was not until just recently however that he managed to implement it properly. It was a long and arduous journey to get it into this form.<\/p>\n<p>A short summary - the standard activation for a vanilla RNN is <code>f(Wx + Uh + b)<\/code>. Multiplicative integration replaces that <code>+<\/code> with <code>f(Wx .* Uh + b)<\/code>. It is possible to generalize that further by adding a bunch of bias terms so it becomes <code>f((Wx + b1) .* (Uh + b2) + b)<\/code>. This simplifies to <code>f(b1 .* Wx .* Uh + b2 .* Uh + b3 .* Wx + b4)<\/code>. The sheer amount of terms to be added and multiplied makes it challenging to write a custom Cuda kernel for. The backward step would double the number of variables to 12 as well due needing to pass adjoints. Not to mention, the biases need to be replicated as well.<\/p>\n<p>The sheer effort that would be needed to implement what would otherwise be a trivial use of <code>map<\/code> is one of the main catalysts for the author abandoning the old F# library and creating Spiral.<\/p>\n<p>So here it is presented, the multiplicative integration RNN in Spiral.<\/p>\n<pre>    \/\/\/ The multiplicative integration RNN.\n    inl mi size sublayer = \n        recurrent \n            {\n            size sublayer\n            weights = inl s -&gt;\n                open Initializer\n                {\n                input = tanh (sublayer.size, size) s |&gt; dr s\n                state = tanh (size, size) s |&gt; dr s\n                b1 = bias s size one\n                b2 = bias s size (to float 0.5)\n                b3 = bias s size (to float 0.5)\n                b4 = bias0 s size\n                } |&gt; heap\n            apply = inl {b1 b2 b3 b4 input state} s i -&gt;\n                match s with\n                | () -&gt;\n                    inm i = matmult (i, input)\n                    d2_replicate_activation {\n                        fwd=inl (b3,b4) i -&gt; b3*i + b4 |&gt; tanh_fwd\n                        bck_in=inl (b3,b4) i out -&gt; (i, one) |&gt; Tuple.map ((*) (tanh_bck out))\n                        bck_in'=inl (b3,b4) i out -&gt; b3 * tanh_bck out\n                        } (b3,b4) i\n                | _ -&gt;\n                    inm i = matmult (i, input)\n                    inm s = matmult (s, state)\n                    d2_replicate_activation {\n                        fwd=inl (b1,b2,b3,b4) (i,s) -&gt; b1*i*s + b2*s + b3*i + b4 |&gt; tanh_fwd\n                        bck_in=inl (b1,b2,b3,b4) (i,s) out -&gt; (i*s, s, i, one) |&gt; Tuple.map ((*) (tanh_bck out))\n                        bck_in'=inl (b1,b2,b3,b4) (i,s) out -&gt; (b1*s+b3, b1*i+b2) |&gt; Tuple.map ((*) (tanh_bck out))\n                        } (b1,b2,b3,b4) (i,s)\n                &gt;&gt;= inl x -&gt; succ (x,x)\n            }\n<\/pre>\n<p>The <code>tanh<\/code> activation is fused into the linear part. This did in fact make the network run a decent bit faster. It takes 4.6s per epoch on the <code>mini_shakespeare.txt<\/code> dataset. In contrast the unoptimized LSTM takes over 18s. It also trains a decent bit better than it, in terms of training error.<\/p>\n<p>It is quite nice, when presented in this form. The matrix multiplications are still unoptimized in the ML library, so this is probably not yet its full performance.<\/p>\n<h5><a id=\"user-content-map--layer-norm--relu\" class=\"anchor\" aria-hidden=\"true\" href=\"#map--layer-norm--relu\"><\/a>Map + Layer Norm + Relu<\/h5>\n<p>Before this is presented, layer norm will be shown on its own. This particular implementation is without the affine factor and acts like an activation instead. It is from the <a href=\"https:\/\/arxiv.org\/abs\/1611.04520\" rel=\"nofollow\">Normalizing the Normalizers<\/a> paper. It works great with Relus even in RNNs, and not at all with sigmoid activations.<\/p>\n<p>Layer norm (pseudo-code):<\/p>\n<pre>inl o x -&gt; \/\/ o is some constant\n    inl v = replicate (mean x)\n    v \/ sqrt (o*o + replicate (mean (v*v)))\n<\/pre>\n<pre>    inl layer_norm =\n        inl fwd o i s =\n            inl n = (primal i).dim |&gt; snd |&gt; HostTensor.span |&gt; to float\n            s.CudaKernel.mapi_d1_seq_broadcast {\n                seq = \n                    {\n                    redo=(+)\n                    map_out=inl i sum -&gt; i - sum \/ n\n                    }\n                    ,\n                    {\n                    map_in=inl v -&gt; v*v\n                    redo=(+)\n                    map_out=inl v vv -&gt; v \/ sqrt (o*o + vv \/ n)\n                    }\n                } (primal i)\n        inl bck o r i s =\n            inl n = (primal i).dim |&gt; snd |&gt; HostTensor.span |&gt; to float\n            s.CudaKernel.mapi_d1_seq_broadcast' {\n                seq = \n                    {\n                    map_in=inl dr,i -&gt; i\n                    redo=(+)\n                    map_out=inl dr,i sum -&gt; dr, i - sum \/ n\n                    }\n                    ,\n                    {\n                    map_in=inl dr,v -&gt; v*v\n                    redo=(+)\n                    map_out=inl dr,v vv -&gt; dr,v,sqrt (o*o + vv \/ n)\n                    }\n                    ,\n                    {\n                    map_in=inl dr,v,norm -&gt; -dr * v \/ (norm * norm)\n                    redo=(+)\n                    map_out=inl dr,v,norm bot -&gt; \n                        inl top = dr \/ norm\n                        inl bot = (bot * v) \/ (norm * n)\n                        top + bot\n                    }\n                    ,\n                    {\n                    redo=(+)\n                    map_out=inl dv dv_sum adjoint -&gt; adjoint + dv - dv_sum \/ n\n                    }\n                } (adjoint r, primal i) (adjoint i)\n        inl o i s -&gt;\n            inl r = fwd o i s |&gt; dr s\n            r, inl _ -&gt; bck o r i s\n<\/pre>\n<p>The <code>mapi_d1_seq_broadcast<\/code> is finally used to its full effect here. The backwards steps won't be elaborated like in the softmax section. As can be seen from the above on the backwards pass the LN is actually recalculated from the input. This is good as the cost of recalculation is pretty much nothing compared to having to do extra reads and writes from intermediaries.<\/p>\n<p>Immediately after getting the above to work, the author fused the <code>layer_norm<\/code> with the <code>relu<\/code> and found that it gave a decent speedup in the RNN. The Relu activation is really trivial to add so that specific variant of LN won't be pasted here to prevent it from bloating the size of the tutorial any further.<\/p>\n<p>But after that step he fused it with multiplicative integration. That came out as something really great. At this moment in time there isn't a single thing in the library that more exemplifies Spiral's approach than it. It also demonstrates the real benefit of <code>mapi<\/code> over the standard <code>map<\/code>.<\/p>\n<p>The function is relatively big, so it will be done in chunks.<\/p>\n<pre>    \/\/\/ Map + layer normalization + relu\n    inl map_ln_relu {fwd bck_in bck_in'} =\n        inl n bias i = \n            inl a,b = Struct.map (inl o -&gt; {o without block}) i |&gt; HostTensor.zip |&gt; inl x -&gt; x.dim\n            Struct.iter (inl {primal adjoint} -&gt;\n                inl f x = \n                    inl b' :: () = x.dim\n                    assert (b' = b) \"The bias has to have a dimension equal to the input\"\n                f primal; f adjoint\n                ) bias\n            HostTensor.span b |&gt; to float\n<\/pre>\n<p><code>n<\/code> returns the span of the innermost dimension. Despite that it seems more complicated, but all it is doing there are boundary checks.<\/p>\n<pre>        inl fwd' o b i w =\n            inl n = n b i\n            inl b = Struct.map (inl {primal} -&gt; w.CudaTensor.to_dev_tensor primal) b\n            w.CudaKernel.mapi_d1_seq_broadcast {\n                mapi_in=inl j i' i -&gt; Struct.map (inl x -&gt; x i' .get) b |&gt; inl b -&gt; fwd b i\n                seq = \n                    {\n                    redo=(+)\n                    map_out=inl i sum -&gt; i - sum \/ n\n                    }\n                    ,\n                    {\n                    map_in=inl v -&gt; v*v\n                    redo=(+)\n                    map_out=inl v vv -&gt; v \/ sqrt (o*o + vv \/ n) |&gt; relu_fwd\n                    }\n                } (Struct.map primal i)\n<\/pre>\n<p>The brilliant part of this is how instead of creating a separate <code>seq_broadcast<\/code> in order to replicate the biases, they are instead captured in lexical scope and passed to <code>fwd<\/code>. What the above does is maps the input, and then passes it to layer norm and then relu.<\/p>\n<p>The backward step is more involved, but fairly similar to the standard layer norm.<\/p>\n<pre>        inl bck' o r b i w =\n            inl n = n b i\n            inl b_primals = Struct.map (inl {primal} -&gt; w.CudaTensor.to_dev_tensor primal) b\n            inl b_adjoints = Struct.map (inl {adjoint} -&gt; w.CudaTensor.to_dev_tensor adjoint) b\n            w.CudaKernel.mapi_d1_seq_broadcast' {\n                mapi_in=inl j i' (dr,i) -&gt; \n                    inl b_primals = Struct.map (inl x -&gt; x i' .get) b_primals\n                    stack {b_primals i}, dr, fwd b_primals i\n<\/pre>\n<p><code>mapi_in<\/code> allows a great deal of extensibility in the kernel. It allows the kernel to index into tensor out of its scope. Of course, what really happens is that they get dragged through the join point and indexed into there, but this is a good way of describing it in actual code.<\/p>\n<pre>                seq = \n                    {\n                    map_in=inl bis,dr,i -&gt; i\n                    redo=(+)\n                    map_out=inl bis,dr,i sum -&gt; bis, dr, i - sum \/ n\n                    }\n                    ,\n                    {\n                    map_in=inl bis,dr,v -&gt; v*v\n                    redo=(+)\n                    map_out=inl bis,dr,v vv -&gt; bis,(if v &gt; zero then dr else zero),v,sqrt (o*o + vv \/ n)\n                    }\n                    ,\n                    {\n                    map_in=inl bis,dr,v,norm -&gt; -dr * v \/ (norm * norm)\n                    redo=(+)\n                    map_out=inl bis,dr,v,norm bot -&gt; \n                        inl top = dr \/ norm\n                        inl bot = (bot * v) \/ (norm * n)\n                        bis,top + bot\n                    }\n                    ,\n                    {\n                    map_in=snd\n                    redo=(+)\n                    mapi_out=inl _ i' {b_primals i},dv dv_sum is_adjoints -&gt; \n                        inl dx = dv - dv_sum \/ n\n                        bck_in b_primals i\n                        |&gt; Struct.map ((*) dx)\n                        \/\/ Note: The atomics make training non-deterministic.\n                        |&gt; Struct.iter2 (inl a -&gt; atomic_add (a i')) b_adjoints\n                        bck_in' b_primals i\n                        |&gt; Struct.map ((*) dx)\n                        |&gt; Struct.map2 (+) is_adjoints\n                    }\n                } (adjoint r, Struct.map primal i) (Struct.map adjoint i)\n        inl o b i s -&gt;\n            inl r = fwd' o b i s |&gt; dr s\n            r, inl _ -&gt; bck' o r b i s\n<\/pre>\n<p>Propagating the adjoints into the biases is achieved through atomics. This is actual the only flaw in the whole kernel. Whereas before the training would be fully deterministic, this causes wide swings from run to run. The author saw the MIRNN range from 160 to 180 on the first epoch. With the biases frozen it gets 163 every time which makes it look like a joke, but the author intends to continue believing in their theoretical advantages.<\/p>\n<p>What the above function allows is using LN+Relu with any kind of map operation whether it be MI or something else, in either feedforward or recurrent networks. As a result the layer with the fully fused activation looks identical to the standard one.<\/p>\n<pre>    inl miln o size sublayer = \n        recurrent \n            {\n            size sublayer\n            weights = inl s -&gt;\n                open Initializer\n                {\n                input = relu (sublayer.size, size) s |&gt; dr s\n                state = relu (size, size) s |&gt; dr s\n                b1 = bias s size one\n                b2 = bias s size (to float 0.5)\n                b3 = bias s size (to float 0.5)\n                b4 = bias0 s size\n                } |&gt; heap\n            apply = inl {b1 b2 b3 b4 input state} s i -&gt;\n                match s with\n                | () -&gt;\n                    inm i = matmult (i, input)\n                    map_ln_relu {\n                        fwd=inl (b3,b4) i -&gt; b3*i + b4 \n                        bck_in=inl (b3,b4) i -&gt; (i, one) \n                        bck_in'=inl (b3,b4) i -&gt; b3 \n                        } o (b3,b4) i\n                | _ -&gt;\n                    inm i = matmult (i, input)\n                    inm s = matmult (s, state)\n                    map_ln_relu {\n                        fwd=inl (b1,b2,b3,b4) (i,s) -&gt; b1*i*s + b2*s + b3*i + b4\n                        bck_in=inl (b1,b2,b3,b4) (i,s) -&gt; (i*s, s, i, one) \n                        bck_in'=inl (b1,b2,b3,b4) (i,s) -&gt; (b1*s+b3, b1*i+b2)\n                        } o (b1,b2,b3,b4) (i,s)\n                &gt;&gt;= inl x -&gt; succ (x,x)\n            }\n<\/pre>\n<p>Since it takes only one instead of two backward steps, it takes 4.3s per epoch versus 4.6s for the standard <code>mi<\/code> RNN without layer norm. All those complicated steps shown above take basically nothing - the real overhead is in memory movement to and from global memory.<\/p>\n<p>Implementing layer norm was one of the other great catalysts that drove the author to create Spiral. After 1.5 years of work, it is possible to present this piece - the fused map + layer norm + relu activation. It cannot be found anywhere else.<\/p>\n<p>With the first part of the grand quest complete, the Spiral ML library finally exceeds the old one in scope.<\/p>\n<h4><a id=\"user-content-layer-combinators\" class=\"anchor\" aria-hidden=\"true\" href=\"#layer-combinators\"><\/a>Layer Combinators<\/h4>\n<p>The primary purpose of layers is to make it easy to initialize the weights and the biases. The secondary purpose is to allow their parallel evaluation.<\/p>\n<p>Basic layer structure is this:<\/p>\n<pre>    inl gid _ = .(to string !GID())\n    inl layer d = {d with gid=gid(); block=()}\n<\/pre>\n<p>Every layer has a <code>layer_type<\/code> field and a <code>gid<\/code> field. The <code>gid<\/code> field holds a unique integer literal field which the evaluator uses to distinguish between the layers and cache their results so parts of the graph aren't executed more than once.<\/p>\n<pre>    inl input name size = layer {\n        layer_type = .input\n        name\n        size\n        }\n<\/pre>\n<p>The input layer is rather simple. It has a <code>name<\/code> and a <code>size<\/code> field.<\/p>\n<pre>    inl stateful layer_type {weights apply size sublayer} = \n        layer {\n            layer_type\n            size\n            sublayer\n            weights\n            apply\n            }\n    inl feedforward = stateful .feedforward\n    inl recurrent = stateful .recurrent\n<\/pre>\n<p>Feedforward and recurrent layers aren't much different from each other. Only their <code>layer_type<\/code> field differs. There are other layer types as well apart from the three shown above.<\/p>\n<p>Layer combinators are used to actually run and initialize the layers. And those combinators are derived from the <code>layer_map<\/code> and <code>layer_map_fold<\/code>.<\/p>\n<pre>    inl layer_map f network =\n        inl rec layer_map r = function\n            | {x with layer_type gid} -&gt;\n                match r with\n                | {$gid=x} -&gt; x, r\n                | _ -&gt;\n                    inl sublayer, r =\n                        match x with\n                        | {sublayer} -&gt; layer_map r sublayer\n                        | _ -&gt; (), r\n                    inl x = f {x with sublayer}\n                    x, {r with $gid=x}\n            | x :: x' -&gt; \n                inl x, r = layer_map r x\n                inl x', r = layer_map r x'\n                x :: x', r\n            | () -&gt; (), r\n            | {} as x -&gt;\n                module_foldl (inl k (m,r) x -&gt;\n                    inl x,r = layer_map r x\n                    module_add k x m, r\n                    ) ({},r) x\n            | x -&gt; error_type (\"Expected a layer. Got\", x)\n        layer_map {} network |&gt; fst\n<\/pre>\n<p><code>$<\/code> is the new injection pattern. <code>$<\/code> allows both matching and construction.<\/p>\n<pre>inl x = .a\n{$x=5} \/\/ {a=5}\n<\/pre>\n<p><code>layer_map<\/code> does bottom up mapping while caching the results. It is different than a standard map in that is allows the layers to access the previous result through the <code>sublayer<\/code> field. <code>r<\/code> is the module it uses to cache internal state.<\/p>\n<p><code>layer_map_fold<\/code> is similar except it also allows passing of state. This is useful for running recurrent networks.<\/p>\n<pre>    inl init s network = \n        layer_map (function\n            | {x with weights} -&gt; {x with weights = const (weights s)}\n            | x -&gt; x\n            ) network\n<\/pre>\n<p>This is how <code>init<\/code> is made. If the layer has a <code>weights<\/code> field it is passed in the context <code>s<\/code> and that returns the network weights after the function has been evaluated. Though rather than just setting weights to that, it wraps it in a function.<\/p>\n<p>This has the property that if the layer is reinitialized later, those parts that have already been initialized are not affected. The reason why this is done is that the net might have a body, but also branches that might need to be initialized separately.<\/p>\n<pre>    inl init_parallel s network = \n        layer_map (function\n            | {stream} | {layer_type=.input | .parallel} as x -&gt; x\n            | {x with weights} -&gt; {x with weights = const (weights s); stream=s.RegionStream.allocate.data.stream}\n            | x -&gt; {x with stream=s.RegionStream.allocate.data.stream}\n            ) network\n<\/pre>\n<p><code>init_parallel<\/code> extends on the standard <code>init<\/code> by also allocating a Cuda stream and storing it into the layer which can be used to execute layers in parallel.<\/p>\n<pre>    inl run x d s =\n        layer_map_fold (inl {x with layer_type gid} d -&gt;\n            match layer_type with\n            | .input -&gt; d.input x.name, d\n            | .stateless -&gt;\n                inl value, bck = indiv join\n                    inl a, b = x.apply x.sublayer s\n                    stack (a, term_cast b ())\n                value, {d with bck = apply_bck self bck}\n            | .non_differentiable -&gt;\n                inl value = indiv join x.apply x.sublayer s |&gt; stack\n                value, d\n            | .parallel -&gt; x.sublayer, d\n            | .feedforward -&gt;\n                inl value, bck = indiv join\n                    inl a, b = x.apply (x.weights()) x.sublayer s\n                    stack (a, term_cast b ())\n                value, {d with bck = apply_bck self bck}\n            | .recurrent -&gt;\n                inl state = match d.state with {$gid=state} -&gt; state | _ -&gt; ()\n                inl (value, state), bck = indiv join\n                    inl a, b = x.apply (x.weights()) state x.sublayer s\n                    stack (a, term_cast b ())\n                value, {d with bck = apply_bck self bck; state = {self with $gid=state}}\n            ) x d\n<\/pre>\n<p>For every layer, this function applies it and takes care to store the backward steps. It also term casts them so the compiler does not diverge for recurrent nets. Note that the results in the bottom layer are stored in the <code>sublayer<\/code> field as noted previously. The state on the other hand is passed through <code>d<\/code>.<\/p>\n<p>The way to imagine the graph being executed is to think of values flowing upwards from <code>input<\/code> layers through the <code>sublayer<\/code> nodes all the way up to the single node at the top. Inputs are passed vie <code>d<\/code> to the input fields and are accessed by name.<\/p>\n<h5><a id=\"user-content-a-note-on-compilation-times\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-note-on-compilation-times\"><\/a>A Note On Compilation Times<\/h5>\n<pre>There is some funny stuff with `indiv join` and `stack` going on. Throughout the source code of the library there will be such patterns used throughout. They have only a single purpose - to optimize compile times.\nWhen the author first made the LSTM work he received a shock in that it would take 4s to compile and adding another layer would add 2s to compilation. Considering that the nets he intends to train will be deep, the realization just how slow Spiral is made him lose his nerve. He first idea was to maybe rewrite the language in OCaml. He already considered it once before to do it Racket, but that fell through.\nThe second idea to spread join points and layout types around more readily. As it turns out that turned out to be greatly effective at crushing the compile times to around 1s and a negligible increase for each layer added. What was surprising to him not just how effective join points were at reducing compilation times, but that some of the functions which had no effect at runtime (like the kernel auxiliaries) and only did assertions and shape checking had a large impact on compile times as well.\nThe author thinks it might be possible to optimize compile times further, but he is at a loss as to how to do it past this point. He would welcome reviews from others on this part.\nSome of the things he tried like flattening the AST that he thought would have benefit made absolutely no difference to performance. Things like optimizing the way `Op`s are represented would severely undermine the ergonomics of the compiler. Parser could be sped up significantly, but it is not a overhead. Actually it is, but it is not the one that is bothersome.\nThe partial evaluator is an enigma - reason tells that it is already fast for the kinds of work it is doing based on a rough comparison with other compilers. It is probably even faster than most other functional languages simply due to how simple it is. On the other hand, if the kinds of things it is doing can't be made any faster that bodes ill for not just Spiral, but for the future of programming languages.\nTo put it like this - if Spiral was 10x slower than it was now, then it would be much less useful. If it was 100x slower then there would be no point in using it. This value estimate goes into the other direction as well. Spiral would be a lot more valuable if it were 10x faster. Performance simply matters everywhere and always - compilers are not the exception.\nThere might be some novel kinds of abstraction in the future that will be out of reach because the compilers are too sluggish to deal with them in practice.\nThe author is pessimistic on the future of software development because of this. There simply does not seem to be any replacement for the immutable data structures Spiral uses internally. They make it flexible and powerful, and they do not make it slow, but they place a ceiling on its performance.\nIt is just as well that he is interested in machine learning. Compared to classical programming, optimization methods have almost limitless potential to improve from here.\n<\/pre>\n<pre>    \/\/\/ The wavefront iteration optimization.\n    \/\/\/ Requires the non-input layers to have preallocated streams.\n    inl run_parallel x d s =\n        layer_map_fold (inl {x with layer_type gid} d -&gt;\n            match layer_type with\n            | .input -&gt; {value=d.input x.name; stream=s.data.stream; block=()}, d\n            | .parallel -&gt; x.sublayer, d\n            | _ -&gt;\n                inl stream = x.stream\n                inl s = s.data_add {stream}\n                inl values = Struct.map (inl {value} -&gt; value) x.sublayer\n                inl streams = \n                    Struct.choose (function\n                        | {stream=x} -&gt; stream.wait_on x; x\n                        | _ -&gt; ()) x.sublayer\n                inl wait_bck b =\n                    inl b _ =\n                        b ()\n                        Struct.iter (inl x -&gt; x.wait_on stream) streams\n                    term_cast b ()\n                match layer_type with\n                | .stateless -&gt;\n                    inl value, bck = indiv join\n                        inl a, b = x.apply values s\n                        stack (a, wait_bck b)\n                    {value stream block=()}, {d with bck = apply_bck self bck}\n                | .non_differentiable -&gt;\n                    inl value = indiv join x.apply values s |&gt; stack\n                    {value stream block=()}, d\n                | .feedforward -&gt;\n                    inl value, bck = indiv join\n                        inl a, b = x.apply (x.weights()) values s\n                        stack (a, wait_bck b)\n                    {value stream block=()}, {d with bck = apply_bck self bck}\n                | .recurrent -&gt;\n                    inl state = match d.state with {$gid=state} -&gt; state | _ -&gt; ()\n                    inl (value, state), bck = indiv join\n                        inl a, b = x.apply (x.weights()) state values s\n                        stack (a, wait_bck b)\n                    {value stream block=()}, {d with bck = apply_bck self bck; state = {self with $gid=state}}\n                ) x d\n        |&gt; inl x, d -&gt; Struct.map (inl {value} -&gt; value) x, d\n<\/pre>\n<p>The wavefront iteration is from this <a href=\"https:\/\/devblogs.nvidia.com\/optimizing-recurrent-neural-networks-cudnn-5\/\" rel=\"nofollow\">NVidia blog post<\/a>.<\/p>\n<p><code>run_parallel<\/code> is similar to <code>run<\/code>, but takes advantage of streams to perform the wavefront iteration optimization. The benefit of it is nowhere near as described in blog post - it is more like 20% in practice, but it is a good 20%. Unless there is a reason not to, <code>run_parallel<\/code> should be used instead of <code>run<\/code>.<\/p>\n<p>The author also tried to stream gemms as per that post, but got absolutely horrible results. For one, it is quite more difficult to use streams inside the layer instead of between them like in the above function.<\/p>\n<p>The first attempt to do is was to allocate them dynamically in order to avoid passing them around, but as it turns out, streams are insanely expensive to allocate so that plan fell through. The second plan was to preallocate and them pass them in, but that would just be too nasty - even if it worked it could potentially lead to data races for some kinds of operations so it was scrapped. It would have made the code too convoluted to bear with.<\/p>\n<p>The problem with the Cublas gemm is that they are simply too inflexible, not that they need to be streamed. Resolving that will be future work.<\/p>\n<p>There was an episode a month or so back when the author really wanted GC for Cuda memory and seriously considered rewriting Spiral in something else. Eventually, after trying out the above optimizations and thought about Cuda streams in more depth, he realize the problem was not just Cuda memory. Even if were possible to GC it somehow, GC cleanup and allocation might cause inadverted memory sharing in the presence of streams.<\/p>\n<p>Streams are probably the most poorly designed of all the Cuda features; not only do they block potential GC, they are so hard to reason about that they are barely usable. Preallocating them and using them for the wavefront iteration like this is probably their one blessed use. In all the other cases, actually fusing the kernels would be a far better choice.<\/p>\n<p>Within the confines of the ML library there is no problem in using them for the wavefront iteration, but if memory sections are shared with something else, there might be ways of causing memory corruption. In general, that should not be a worry though.<\/p>\n<h4><a id=\"user-content-loops-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#loops-1\"><\/a>Loops<\/h4>\n<p>The loops in this context refers to the functions that run the net once. Here they are for feedforward nets.<\/p>\n<pre>        inl train {d with network} =\n            inl rec loop c cost' = \n                function\n                | .unwrap -&gt; cost' \/ to float64 c\n                | input s {on_fail on_succ} -&gt;\n                    inl cost, {bck} = run network {input state = {}; bck=const ()} s\n                    inl cost' = cost' + to float64 (s.CudaTensor.get cost)\n                    inl state = loop (c+1) cost'\n                    if nan_is cost' then on_fail state\n                    else\n                        match d with\n                        | {optimizer} -&gt;\n                            bck()\n                            optimize network optimizer s\n                        | _ -&gt; ()\n                        on_succ state\n            loop (dyn 0) (dyn 0.0)\n<\/pre>\n<p>The loop has to do a bunch of things while it is running like abort early on nans in the cost, keep track of costs, call the backwards step and the optimizer. This is something that takes care of that.<\/p>\n<pre>        inl test {d with network} =\n            inl rec loop c cost' accuracy' accuracy_max' = \n                function\n                | .unwrap -&gt; cost' \/ to float64 c, accuracy', accuracy_max'\n                | input s {on_fail on_succ} -&gt;\n                    inl (cost, {value max}), {bck} = run network {input state = {}; bck=const ()} s\n                    inl cost' = cost' + to float64 (s.CudaTensor.get cost)\n                    inl accuracy' = accuracy' + s.CudaTensor.get value\n                    inl accuracy_max' = accuracy_max' + max\n                    inl state = loop (c+1) cost' accuracy' accuracy_max'\n                    if nan_is cost' then on_fail state\n                    else\n                        match d with\n                        | {optimizer} -&gt;\n                            bck()\n                            optimize network optimizer s\n                        | _ -&gt; ()\n                        on_succ state\n            loop (dyn 0) (dyn 0.0) (dyn 0) (dyn 0)\n<\/pre>\n<p>This one takes care of accuracy and the max accuracy.<\/p>\n<p>Both of <code>train<\/code> and <code>test<\/code> hold their state internally and are unwrapped at the end of the epoch. <code>train<\/code> and <code>test<\/code> are passed in as body for the <code>for<\/code> function.<\/p>\n<pre>    inl outer data =\n        Struct.foldl (inl s x -&gt;\n            match s with\n            | () -&gt; fst x.dim\n            | s -&gt; assert (s = fst x.dim) \"The data tensors need to have the same outer dimension.\"; s\n            ) () data\n    inl for {data body} s =\n        inl {from near_to} = outer data\n           \n        Loops.for' {\n            from near_to\n            state=body\n            body=inl {next state i} -&gt;\n                inl data = Struct.map (inl x -&gt; x i) data\n                s.refresh\n                inl s = s.RegionMem.create\n                state data s {\n                    on_fail=inl state -&gt;\n                        s.RegionMem.clear\n                        state.unwrap\n                    on_succ=inl state -&gt;\n                        s.RegionMem.clear\n                        next state\n                    }\n            finally=inl state -&gt; state.unwrap\n            }\n<\/pre>\n<p>Until now, no mention whatsoever has been made about how the library manages memory, so this is a good time to do it.<\/p>\n<pre>inl s = s.RegionMem.create\n<\/pre>\n<p>The context has <code>RegionMem<\/code> and <code>RegionStream<\/code>. When <code>.create<\/code> is passed into them a new region is created. That does not actually allocate any memory or streams, but instead allocates a new region. Under the hood, the region is a resizable array that holds all the references allocated through it.<\/p>\n<pre>s.RegionMem.clear\n<\/pre>\n<p>When this is called, all the references in that region have their count decremented by 1 and if that count is 0, they are set to null. For feedforward nets, that means the allocation always get cleared since their references are never shared with other regions.<\/p>\n<p>That does not actually dispose of the memory however.<\/p>\n<pre>s.refresh\n<\/pre>\n<p>It is this thing that does so. It filters out all the references set to null and rebuilds the free cells. It can be thought of as the equivalent of doing a GC run. This system is nowhere as good as actual GC in terms user ease or flexibility, but it should suffice for neural network needs. It can be improved to support things like very short term allocation without having to clean up everything, but that will be future work.<\/p>\n<pre>                state data s {\n                    on_fail=inl state -&gt;\n                        s.RegionMem.clear\n                        state.unwrap\n                    on_succ=inl state -&gt;\n                        s.RegionMem.clear\n                        next state\n                    }\n<\/pre>\n<p>Note that <code>state<\/code> here is either <code>train<\/code> or <code>test<\/code>. The <code>for<\/code> loop shown is a bit unusual in that the mapping function for it is also the state. The reason for that is because it is the easiest way of initializing the state. Rather being split into two, everything <code>train<\/code> and <code>test<\/code> need are right there where they should be.<\/p>\n<p>Doing it like that is a straightforward OO pattern in Spiral, but would be untypable in other static functional languages.<\/p>\n<p>The recurrent networks have their own <code>train<\/code> and will have <code>test<\/code>, but they are beyond the scope of this tutorial. They are more complicated, but have exactly the same purpose. Both feedforward and recurrent nets have their own dedicated gradient checking function which are useful for testing additions to the library.<\/p>\n<p>This covers all the important parts of the library.<\/p>\n<p>Fundamentally, ML libraries are not something to be worked on forever which is the unfortunate situation author found himself in. Much like parser combinators resolve the need for external parser generators, so can a ML library be fashioned in a similar vein. The greatest feature Spiral offers to the user is that it easily makes it possible to craft pieces like <code>map_ln_relu<\/code>. It is in turn made out of flexible pieces like <code>seq_broadcast<\/code>. They free the user of the drudgery of low level C programming and allow more a direct expression of desire. They also enable a vast amount of code reuse. Furthermore, Spiral's approach to programming allows the library an absolute level of integration with the language. This will be greatly useful when the time comes to make use of it as a module in a reinforcement learning context.<\/p>\n<p><code>map_ln_relu<\/code> is just the tip of the iceberg and this can be considered the first unfinished release of the library. But compared to all the work that was needed to make both the language and get the library to this point, the rest should be much easier. The great up front expense has been paid in full.<\/p>\n<h4><a id=\"user-content-example---feedforward-net-on-mnist\" class=\"anchor\" aria-hidden=\"true\" href=\"#example---feedforward-net-on-mnist\"><\/a>Example - Feedforward Net On Mnist<\/h4>\n<p>The examples are just to give a feel for how the code looks like in practice.<\/p>\n<pre>let learning9 =\n    \"learning9\",[cuda_modules;learning;mnist],\"Does the full training work with Mnist?\",\n    \"\"\"\ninb s = CudaModules (1024*1024*1024)\ninl float = float32\nopen Learning float\ninl minibatch_size = 128\ninl { test_images test_labels train_images train_labels} =\n    inl mnist_path = @\"C:\\ML Datasets\\Mnist\"\n    Mnist.load_mnist_tensors mnist_path\n    |&gt; s.CudaTensor.from_host_tensors\n    |&gt; module_map (inl _ x -&gt; x.round_split' minibatch_size)\n<\/pre>\n<p>This is a handy loader just for Mnist. Tensors have the <code>round_split<\/code> and <code>round_split'<\/code> methods in them that split the outermost dimension to <code>(minibatch_size,rest)<\/code> and <code>(rest,minibatch_size)<\/code>. So the shape of the tensors after the above would be <code>(rest,128,784)<\/code> after the above operation with the <code>rest<\/code> being statically known.<\/p>\n<pre>inl input_size = 784\ninl hidden_size = 10\ninl network = \n    open Feedforward.Layer\n    inl label = input .label hidden_size\n    inl network =\n        input .input input_size \n        |&gt; ln 0.0f32 256\n        |&gt; linear hidden_size \n        |&gt; init s\n    inl train = error Error.softmax_cross_entropy label network\n    inl test = parallel (train, accuracy label network)\n    {train test}\n<\/pre>\n<p>The layers can be easily piped using the <code>|&gt;<\/code> operator. There is no danger of failing to connect layers like this and no need to write unit tests for this purpose like some other libraries require the user to do.<\/p>\n<pre>Loops.for' {from=0; near_to=10;body=inl {next} -&gt; \n    open Feedforward.Pass\n    open Body\n    inl cost =\n        for {\n            data={input=train_images; label=train_labels}\n            body=train {\n                network=network.train\n                optimizer=Optimizer.sgd 0.3f32\n                }\n            } s\n    string_format \"Training: {0}\" cost |&gt; Console.writeline\n    if nan_is cost then\n        Console.writeline \"Training diverged. Aborting...\"\n    else\n        inl cost, ac, max_ac =\n            for {\n                data={input=test_images; label=test_labels}\n                body=test { network=network.test }\n                } s \n        string_format \"Testing: {0}({1}\/{2})\" (cost, ac, max_ac) |&gt; Console.writeline\n        next ()\n    }\n<\/pre>\n<p>Prepare the dataset, define the network and then run it. This is the part that does the last one. The really salient parts are those where it calls the <code>Learning<\/code> <code>for<\/code> function. The rest is just standard IO and checking for Nans.<\/p>\n<pre>        for {\n            data={input=train_images; label=train_labels}\n            body=train {\n                network=network.train\n                optimizer=Optimizer.sgd 0.3f32\n                }\n            } s\n<\/pre>\n<p>This is succinct enough that it actually does not need any further comment. One thing that would be good to mention is that since Spiral loves statically known dimensions, the <code>split_round<\/code> functions tend to throw away the excess in order to keep the tensor regular. So the test set rather than having 10k examples will have 9984 examples.<\/p>\n<pre>Training: 0.309259804475129\nTesting: 0.157295000082694(9493\/9984)\nTraining: 0.114737567477899\nTesting: 0.111421216703139(9622\/9984)\nTraining: 0.0818591211150345\nTesting: 0.0956207461153658(9671\/9984)\nTraining: 0.0638285737313553\nTesting: 0.0879654984933157(9693\/9984)\nTraining: 0.0516892505001723\nTesting: 0.0810565133280575(9715\/9984)\nTraining: 0.042687166776731\nTesting: 0.0776575743668498(9724\/9984)\nTraining: 0.0355717467236667\nTesting: 0.075778437509703(9734\/9984)\nTraining: 0.0297816385700105\nTesting: 0.0741259951931007(9747\/9984)\nTraining: 0.025186768730975\nTesting: 0.0730099095739066(9750\/9984)\nTraining: 0.0212900042135873\nTesting: 0.0725844820121076(9749\/9984)\n<\/pre>\n<p>Layer norm + relu works rather well as an activation given how they hit 95% in the first epoch. This is almost on par with batch norm. The Mnist test is useful for ensuring that new additions to the library are implemented correctly.<\/p>\n<h4><a id=\"user-content-example---recurrent-net-on-tiny_shakespeare\" class=\"anchor\" aria-hidden=\"true\" href=\"#example---recurrent-net-on-tiny_shakespeare\"><\/a>Example - Recurrent Net on tiny_shakespeare<\/h4>\n<p>Prepare the dataset.<\/p>\n<pre>let learning11 =\n    \"learning11\",[cuda_modules;timer;learning],\"Does the full training + sampling work with the char-RNN?\",\n    \"\"\"\ninb s = CudaModules (1024*1024*1024)\ninl float = float32\nopen Learning float\ninl size = {\n    seq = 1115394\n    minibatch = 64\n    step = 64\n    hot = 128\n    }\n\/\/ I got this dataset from Karpathy.\ninl path = @\"C:\\ML Datasets\\TinyShakespeare\\tiny_shakespeare.txt\"\ninl data = \n    macro.fs (array char) [text: \"System.IO.File.ReadAllText\"; args: path; text: \".ToCharArray()\"]\n    |&gt; Array.map (inl x -&gt; \n        inl x = to int64 x\n        assert (x &lt; size.hot) \"The inputs need to be in the [0,127] range.\"\n        to uint8 x\n        )\n    |&gt; HostTensor.array_as_tensor\n    |&gt; HostTensor.assert_size size.seq\n    |&gt; s.CudaTensor.from_host_tensor\n    |&gt; inl data -&gt; data.round_split size.minibatch\ninl minibatch,seq = data.dim\ninl input =\n    inl data = s.CudaTensor.to_dev_tensor data \n    s.CudaKernel\n        .init {dim=seq,minibatch} (inl seq minibatch -&gt;\n            data minibatch seq .get\n            )\n<\/pre>\n<p>This last part where the init is called transposes the dataset to the sequence of minibatches from a minibatch of sequences.<\/p>\n<pre>inl label = input.view_span (const {from=1}) .round_split' size.step\ninl input = input.view_span (inl x :: _ -&gt; x-1) .round_split' size.step\ninl data = {input label}\n<\/pre>\n<p>The labels are just the inputs shifted one to the left. Then they are split so that the final dimensions of the tensor are <code>(rest,64,64)<\/code> where the middle 64 is the number of steps and the innermost 64 is the minibatch size.<\/p>\n<p>Define the network.<\/p>\n<pre>inl network = \n    open Recurrent.Layer\n    \n    inl label = input .label 1 |&gt; encode.one_hot size.hot\n    inl input = input .input 1 |&gt; encode.one_hot size.hot\n    inl body =\n        input\n        |&gt; miln 0.05f32 128\n        |&gt; Feedforward.Layer.linear size.hot\n        |&gt; init_parallel s\n    inl train = \n        error Error.softmax_cross_entropy label body\n        |&gt; init_parallel s\n    \n    {train body}\n<\/pre>\n<p>The <code>encode.one_hot<\/code> is a layer that does encoding into the one-hot vector.<\/p>\n<p>Then here is the part that runs the network. It is similar to the feedforward case.<\/p>\n<pre>inb _ = Timer.timeit \"whole loop\"\nLoops.for' {from=0; near_to=5; body=inl {next i} -&gt; \n    open Recurrent.Pass\n    open Body\n    inl cost = \n        Timer.timeit (string_format \"iteration {0}\" i)\n        &lt;| inl _ -&gt;\n            for {\n                data\n                body=train {\n                    network=network.train\n                    optimizer=Optimizer.sgd 0.01f32\n                    }\n                } s\n    Console.writeline \"----\"\n    sample 0.6f32 2048 network.body '\\n' s\n    string_format \"Training: {0}\" cost |&gt; Console.writeline\n    if nan_is cost then\n        Console.writeline \"Training diverged. Aborting...\"\n    else\n        next ()\n    }\n<\/pre>\n<p>The key differences are - there is timing code, there is the <code>sample 0.6f32 2048 network.body '\\n' s<\/code> function in the middle and there is no test pass.<\/p>\n<p>Here is a sample run.<\/p>\n<pre>Starting timing for: whole loop\nStarting timing for: iteration 0\nThe time was 00:00:04.6501455 for: iteration 0\n----\nSample:\nwing somest gooteded, and, singe,\nThe to betith in sumese the pearated.\nWhis in sallers,\ufffdD you and(ed ansting the if the thou hinser, storsuld,\nAnd of in coment not in whatss,\nInd thin bevers,ing in drain,\nThe trust,\nAnd the sake thisher, and the that the deake wast,\nAnd all of marst the thatpen and at cond waiching to thessirserent,\ufffd:\nSill I here.\nThand the be thered lows,\nI and you cand I and tringher,\nBre this the me that\nThenens\nAnded I thingent then, sursted the peave were the whingerew the mane\nThe be, coeld me the wave, the cour:\nI the thend bete the deact, sird you make to the at wither.\nThery oreded mors beno, ripe comering thise dore siandsure,\nAs::\nWhat now the misest wearse\nThoughty,\nThouping the the ceelles ourte.\nThe farthersest now\nAnd at is wither, and that hath I with in you diantaces, the wast\nAnd all in,\nAnd the and ast yourd the the ceeptingers\nI his of the siand thend that to way, bytherest theard the not to the is thet the wime, themell the thisser:\nThat thees,\nAnd the wantersed be,\nInd asenoon,Theres,\nHereest to hard wers of in the of with yishinging this the mastereds,\nAnd cakere thes in brony and of forderselt:\nThen to gout, store torate the that?\nMELIR:\nThe will of fand the is's porto the thet save wed nothtont of thenciint, ind,\nTo me,\nAnd is with in and thenesand and himing wish not strogh\nAnd wither,\nAnd to dearons and m| mair beast a the whinges and the whath, as tomes, of my then the the that wis;\nBing, I good, make the stoll ot I me siet stome there sught.\nThenetront,\nAnd the to wime betore an thou lodser:\nAnd she for the this in your to meang't, sien and hourituteround.\nLARA:\nI werainert\nThertedE\nI laveres?\nTo lading inest the thend\nYould, will wionds\nAnd the hearf in goed ther,\nBut our of wach, the wagk\n'ot pith and not to dearse the bute.\nSIINO:\nAnd his this fore thisest note,\nHear'd ather's and mask?\nKIIIIO:\nThend and the wout.y this, fave and lathers, buster.\nSIO:\nAnd pare, Hinged in the couthar yourstast the and it lingels, lard whe sith histound on ond thourd in\n-----\nTraining: 170.018245865317\nStarting timing for: iteration 1\nThe time was 00:00:04.6143145 for: iteration 1\n----\nSample:\nNow therenowerad's forthy all the distarion theeld and and and the blainise is the trues\nSire, on to be the onound our lord and betion singed lord.\nPUCES:\nThenet is lord of the courteday,\nThe and halood a mone of the coundey countien words thou with is thou all toome sonerites ourself your with the and to there and tongues enday, of not came the of apoing to fallow and that how's it that to muding bove to beting the look of singanes\nWell to beastien madys, his as to to the devend and thenese lird is bother of refore, of our to apour.\nThat bunder,\nTo fars to their the parioule.\nOUCELIO:\nOre with it distlation of with so, and earded to him a worce whing the gooke to sourt that to the sontle upon the biend that be and it that with tayer hport and and to that of the to then the foold porn the po that is wither the with thenenoloue one to sonery come are parding the know lead in land you dirst the rad the tood comes have yet hare and the sire and stard to seariens and that revererow(\nAnd of then I mare\nin then the to somered his are the prother&lt;ing thou with thy lonted, and bothing thou may horeferst\nI come to the to and that am their will then deariour and bray be and betrows thou toole afere to that and and turation you way, and the stole will sindares0er:\nI saye, and his beanied thou asear,\nAnd are of too hishers\nMy the beis to hinging the wirl wasting the dayes, then the known the sanded, and to in not ture the wortle thou to the word.\nunder forterald thee this thought as pake dainice lot hast to theight hather,\nBelood is ase the mahe tonest to\nSo thou hathous tain to there astand them of is the downatterse: moy bearsice and thou sould to thes, to are thou and not to sunatard dearce?\nPOUDARO:\nAnd took, to come beating the your a may they with the eagen,\nTo to my that in fore to thereford,\nAnd mare the west in to thisker. I upon the power this downiald be songert as as to th|inus to piet more nother theee and and to to to thou thou and therede then of asse,A:\nTo to drood,\nNobe, becain I told the will thou how\n-----\nTraining: 130.864040150362\nStarting timing for: iteration 2\nThe time was 00:00:04.5930354 for: iteration 2\n----\nSample:\nridown of, and seeping the with of this is loved that bold that padyed and your the bedang kauch of bess thought to thou live,\nThe king, thou hast, thou sire of the till this thoughtders the everied,\nFor to shall my to to him the constance of toot and that therefore and sonronion\nThou gince,\nMard, and bard,\nAnd dastertart the surseding the sware.\nPETEN:\nYour love, the sentle,\nAnd but all the can to feres say is conferty this to but with that again, fallon that stome arm not I are the sealt forting the fored,\nThe tonely, them thou sine now that to the sinter, m|inger down the surtion less.\nFirst toous are to mare Hartery forst, that soorerage, bard letten,\nHave of in are with of the provedious the lord:\nI other in the romaster him, art of will thy soke sound.\nPOUMENTIO:\nDose:\nAnd thou will wister all seed fare\nAnd souls to to dost that toome all\nsleast us this the lick, and to doingsy head!\nANTIO:\nComence are that as the make told bear,\nWhat hand the deal your losk to my lord!\nPETCANDNO:\nWard, of the might and fortly manrrngness,\nThat hawe of butise Buty are mone my loves;\nHere, make the vant;\nAnd hpard, the will the good, fortows and the with to the word and not hearing of the to fartor.\nALIUS:\nMy that I landly comenion would what every not to lord;\nThese with though his ever will with father, and but with all stiding, Edward,\nAnd down: father, with now the encition:\nSack met your hand of have is then in the thought your hearts,\nAs my of mander stearself oftle manion the king they well.\nKING RUCHARD III:\nWeres the ling my fands you on the great and hpards the sile, in their comest the say as as a ment,\nThe looks of soncease sir thou beangled botter'd and then throst the stant forkented to the will of hath hpard that is beit though a seem the stand the sakes of the day hearted,\nWould with and the with this the west be to the doughtere you gime hath the contence.\nKING IO:\nHerecome,\nThe fard the cance ases to the word's wears,\nHend with then the erterver:\nFore fallon and the peanion'sted.\nGORINGLAN:\nSome th\n-----\nTraining: 120.591899198644\nStarting timing for: iteration 3\nThe time was 00:00:04.6190544 for: iteration 3\n----\nSample:\nrest worther that lork\nTo the seaven the pard that branch and where bray and his with the word mamallue that of the tower fast in the forsed but of is like his to the bust m|ine thy forther a thou stard thou are come to soes and but in that to with thou coot man spord and never of the world ressibing be father for the fall would were to the some stain, truch sorford,\nTo be so this to the wards this say the to buny of be thy sir, strut thou falled to and to be the king.\nARCISTEN:\nSir, whose fair, his day, are my last in this this but to to thou lord.\nowe the it.\nMare's pardous in the confort of your have with thou to be by the backers, that thou is this litter to the can poor your fail toome heart:\nAnd the stant should to the king of his dread of thy of the of the caunted say stay the rawell, and take thou how it that shall thou strenhon our king\nWhow a sing the to me sold abeary ware, cale to my pown.\nDERENTER:\nWhere should of the counter be toot, the hand, be prain so entrant thou carn, it the might with the purilit the will to the sucting mistruck at is to the seaver mone too lood not diest the ere his warther of the well my lord, this counst of a son, beden the fourtion of her to adiend, thoo both of his be proms the with at before of the fail the kingl| the surse thereing be chook of this of dowers seaster is sporisher returm hus a to blousing a that to denself,H:\nHow is but the shall m|ot come the horst both as of the strank we wife ble?\nAnd the it resign: and thy she do think not sirtus as the world hd holl me sower the gate a farle the surve say that drantly show, and thousence\nRomeos.\nPETROLIO:\nThy may not some is that with tpost it that well the complain the rest mine forturess in to the to the should not mucked hpard unterding thy love is this strants of so confort did make of my lord hand the lord, lost of the shall to say the man a my lord:\nAnd you to the coursely man all thou gran murse to the way to the king a brotce.\nGLOUCESTER:\nMy fall you go, I say good that for and a sind to of man if yo\n-----\nTraining: 114.656255665947\nStarting timing for: iteration 4\nThe time was 00:00:04.6133752 for: iteration 4\n----\nSample:\nthe king\nAnd rothy of thou lord, my trient our hath young rest lpettard tporcious are and most a was sould hlast tpose more his tords:\nThe darther are to thou farsing in\nSir, their command a manis me npard tpough for.\nQUEEN ELIZALUSNE:\nYou son,\nAnd thou vard is fires thou worst tporst that the man well in that at hland master's son, with in his man this live.\nPOLARUS:\nBut life, word!\nMENENIUS:\nWhat where or your bear me tpought for man are th|in tpose of word and npard0ent, the deping the strunch and of happy the long widims\nThere of your hast of did say, and but but is consignds husbard own beseefords of seeding, thou sound of that ment,\nAnd thou fortion thou man the courte, good of this with that lands to him, and are bring then Seevous not son; and one of the doth the stand it hand days\nWhat my mardiomer:\nAnd my the down that did birdured were best soining do with of therefore that kind which more thou what with you all for to than my one of minder('s sirst and with him mone to drand.\n\ufffdeloss better I purstian;\nAnd name.\nTo make are unto thy hand that shall be so.\nThisineding thou day dood of what to him such mand his time; us bispited hlard and say is the blatten are in holy sings.\nHENRY CANIE:\nThat with thou wpand with mistrous that words missenself!\nCORINA:\nGoother fdow you are thou have be come to to the wprown had tpought is meation, my life, monaster the call the ronds so such it, and surved the mast that fake that would thou re.\nKING RICHARD III:\nBath the parton but rather constains, my exery as be in must,\nFor of ma,\nthis shall sir, be the promes us would and as this is to comes\nAnd m|orant tooblain of know m|halst whose of must hd parton it it is to have lady, and how so king send thou are sear to this reason; a word to comes with and tpough, I diveress\nAnd the sirrle sir;\nWhy, like the sont.\nSecond to tpatter, our farding is they but grace the part as some of cannot say, the falt our thou not thou lord.\nROMEO:\nAnd to took in seaven of friar it thou propost stands thou am for from the last in\n-----\nTraining: 110.771023666157\nThe time was 00:00:25.4366961 for: whole loop\n<\/pre>\n<p>The net does not take long to start spouting Shakespearean sounding gibberish. Better nets and bigger datasets would improve performance no doubt. But this run is just intended to be equivalent of Mnist for recurrent nets. It would be possible to shave 0.3s from the runtime of each epoch by preprocessing the dataset more, but this way is the most convenient for when sampling is done.<\/p>\n<p>Based on looking at the above, it is clear even a tiny recurrent net with only 128 hidden units and a single layer is capable of learning a great deal of the structure of the dataset which is what the test intended to demonstrate.<\/p>\n<p>The author is definitely curious how a net with multiple layers and dilated connections would learn. That will definitely be subject of future work.<\/p>\n<h4><a id=\"user-content-future-work\" class=\"anchor\" aria-hidden=\"true\" href=\"#future-work\"><\/a>Future Work<\/h4>\n<p>Some of the loose ends currently in the library that need to be worked on when the author gets back to it.<\/p>\n<p>Getting to this point is a great relief to the author because the problem is no longer making the language which could take over a year, nor the huge arduous task of catching up and exceeding the old library. Rather, the problem is that the he has nothing to use it on. So the next task on his agenda will be to work on games for the agents. And what sort of better games than the ones where dumb people are willing to wager both money and status against potentially superhuman opponents.<\/p>\n<p>The future sure looks bright. The plan is simple; while much can't be expected with a simple one layer recurrent net like the one demonstrated on the char-RNN, it is a starting point and the continual process of constantly improving on that should yield bounty. It certainly beats the reality of pounding the pavement for the last 2.5 years.<\/p>\n<p>The following items are roughly in the order of priority. They describe the potential main avenues of improvement once the time is ripe.<\/p>\n<h5><a id=\"user-content-generic-matrix-multiplication\" class=\"anchor\" aria-hidden=\"true\" href=\"#generic-matrix-multiplication\"><\/a>Generic Matrix Multiplication<\/h5>\n<p>Usually the good parts of the library would want to be highlighted, but the quality of this function is poor. Nonetheless matrix multiplication is absolutely indispensable so it needs to be covered.<\/p>\n<p>Before it is shown, here is a short primer on how AD is done through matrix multiplication. <code>.T<\/code> represents matrix transpose.<\/p>\n<pre>C = A * B\ndA += C * B.T\ndB += A.T * C\n<\/pre>\n<p>Here is the function that implements that. It also adds the bias term.<\/p>\n<pre>    inl matmultb l bias s = \n        inl l =\n            match l with\n            | () -&gt; error_type \"The first argument must not be empty.\"\n            | (_,_) :: _ -&gt; l\n            | _ :: _ -&gt; l :: ()\n        inl C = \n            Tuple.foldl (inl C (A,B) -&gt;\n                match C with\n                | () -&gt; s.CudaBlas.gemm .nT .nT one (primal A) (primal B) |&gt; dr s\n                | C -&gt; s.CudaBlas.gemm' .nT .nT one (primal A) (primal B) one (primal C); C\n                ) () l\n        match bias with\n        | () -&gt; ()\n        | _ -&gt; s.CudaKernel.d2_replicate_map' (inl a b _ -&gt; a+b) (primal bias) (primal C) (primal C)\n        C, inl _ -&gt; join\n            inl C' = adjoint C\n            inl l =\n                Tuple.iter (inl A, B -&gt; \n                    on_non_nil (adjoint A) (inl A -&gt; s.CudaBlas.gemm' .nT .T one C' (primal B) one A)\n                    on_non_nil (adjoint B) (inl B -&gt; s.CudaBlas.gemm' .T .nT one (primal A) C' one B)\n                    ) l\n            match bias with\n            | () -&gt; ()\n            | _ -&gt; on_non_nil (adjoint bias) (inl bias -&gt; s.CudaKernel.mapi_d2_redo_map' {map_in=const;neutral_elem=zero;redo=(+);map_out=(+)} C' bias.empty bias)\n<\/pre>\n<p>As a rough example of how it would be used in pseudo-code:<\/p>\n<pre>\/\/ Equivalent of C = x*W + h*U + b\ninl C = matmultb ((x,W),(h,U)) b \/\/ for vanilla recurrent networks\n<\/pre>\n<pre>\/\/ Equivalent of C = x*W + b\ninl C = matmultb (x,W) b \/\/ for feedforward networks\n<\/pre>\n<p>So the function does have a bit flexibility in that it can accept an arbitrary number of terms to multiply and it does optimize memory usage a little. A naive AD implementation would create intermediaries for every multiplication, but that is all that it offers. In terms of functionality offered, this function is identical to the one found in the old ML library written in F#.<\/p>\n<p>Here is the first pain point:<\/p>\n<pre>                    on_non_nil (adjoint A) (inl A -&gt; s.CudaBlas.gemm' .nT .T one C' (primal B) one A)\n                    on_non_nil (adjoint B) (inl B -&gt; s.CudaBlas.gemm' .T .nT one (primal A) C' one B)\n<\/pre>\n<p>Two gemm calls in the backward steps. This is a problem because it does not at all take advantage that <code>C<\/code> is shared. Had the two calls been fused, that could have been taken advantage of for a significant boost in performance.<\/p>\n<p>Pain point two is this:<\/p>\n<pre>            Tuple.foldl (inl C (A,B) -&gt;\n                match C with\n                | () -&gt; s.CudaBlas.gemm .nT .nT one (primal A) (primal B) |&gt; dr s\n                | C -&gt; s.CudaBlas.gemm' .nT .nT one (primal A) (primal B) one (primal C); C\n                ) () l\n<\/pre>\n<p>This is not a problem when there is only a single pair of matrices to multiply, but in recurrent nets there tends to be two pairs. This function is wasteful because it would not take advantage of any potential sharing in the inputs, but it also would need to write multiple times to the output.<\/p>\n<p>There is a missed fusion opportunity here. This one is actually much more severe than it appears at first glance.<\/p>\n<pre>        match bias with\n        | () -&gt; ()\n        | _ -&gt; s.CudaKernel.d2_replicate_map' (inl a b _ -&gt; a+b) (primal bias) (primal C) (primal C)\n<\/pre>\n<p>Adding bias during the linear part seems sensible at first glance, but the author has started to reconsider this design choice as a throwback to the old ML library where he never even considered fusing the bias and the activation. It depends on the complexity of the activation though, but since matrix multiply is a n^3 operation a strong case can be made that biases should not be dealt with in the linear part.<\/p>\n<p>There is one more point to be made - the age of shallow nets is absolutely over. The ideal network is the one that connects to many previous layers in order to allow easy gradient propagation. That means many convolutional or matrix operations done in parallel that sum up at the end.<\/p>\n<p>Vertically speaking that would be the <a href=\"https:\/\/arxiv.org\/abs\/1608.06993\" rel=\"nofollow\">DenseNet<\/a> architecture.<\/p>\n<p>In the paper they represent the connectivity pattern by concatenating all the past inputs.<\/p>\n<pre>C = W * concat (a,b,c...)\n<\/pre>\n<p>This is actually a common optimization in <a href=\"https:\/\/devblogs.nvidia.com\/optimizing-recurrent-neural-networks-cudnn-5\/\" rel=\"nofollow\">LSTMs for example<\/a>, to turn multiple smaller matrix multiplies into a single big one. It is possible to make an alternative representation of the above operation.<\/p>\n<pre>C = sum(Wa*a,Wb*b,Wc*c...)\n<\/pre>\n<p>In other words, the operation can be done by using a large number of small matrix multiplications (or convolution operations.) And such a sufficiently fast and flexible kernel can be written in Spiral.<\/p>\n<p>There are also indications that this would be incredibly important in (quasi)recurrent networks.<\/p>\n<p><a href=\"https:\/\/openreview.net\/forum?id=B1DmUzWAW\" rel=\"nofollow\">SNAIL<\/a> and <a href=\"https:\/\/arxiv.org\/abs\/1803.01271\" rel=\"nofollow\">temporal convolution<\/a> architectures have especially good training properties due to ability to easily propagate gradients. Those general architectures would also benefit greatly for having flexible and fast matrix multiply and convolutional kernels that are not offered by the Cuda libraries.<\/p>\n<p>This makes the case for why ML libraries should offer their own matrix multiply primitive - based on the understanding and the knowledge accumulated during the past few years, it seems increasingly credible that the ideal nets are those that do a large number of small operations with numerous residual connections that stretch deep both into time and space.<\/p>\n<p>As the author does not need such networks quite yet, he is content to leave that work for the future.<\/p>\n<h5><a id=\"user-content-fuse-the-weight-updates\" class=\"anchor\" aria-hidden=\"true\" href=\"#fuse-the-weight-updates\"><\/a>Fuse The Weight Updates<\/h5>\n<p>The way it is currently structured, the weights are updated one by one. If there are 100 weights, then the map kernel must be called 100 times, once per weight.<\/p>\n<p>It should be possible to fuse all those calls into a single kernel.<\/p>\n<p>It would improve performance and be a cool thing to do that the other libraries can't touch.<\/p>\n<h5><a id=\"user-content-metalearning-via-optimization\" class=\"anchor\" aria-hidden=\"true\" href=\"#metalearning-via-optimization\"><\/a>Metalearning Via Optimization<\/h5>\n<p>While MAML is beyond the scope of the <code>Learning<\/code> library, the recent <a href=\"https:\/\/blog.openai.com\/reptile\/\" rel=\"nofollow\">Reptile<\/a> might be worth looking into. At the moment there are model based approaches to metalearning which use RNNs, and then there are optimization based approaches like Reptile. Right now the author does not understand how those two would mix together and would be interested in finding out.<\/p>\n<p>Hopefully, in the following months papers will come out that try to bridge that gap. The results will be interesting either way.<\/p>\n<h5><a id=\"user-content-dilated-rnns\" class=\"anchor\" aria-hidden=\"true\" href=\"#dilated-rnns\"><\/a>Dilated RNNs<\/h5>\n<p><a href=\"https:\/\/arxiv.org\/abs\/1710.02224\" rel=\"nofollow\">Dilated Recurrent Neural Networks<\/a> is one architecture that needs implementation. Very recently, like in the past two years DenseNets and temporal convolutions have emerged as an effective way to train feedforward nets and quasi-recurrent nets. Dilated connections are their recurrent counterpart.<\/p>\n<p>This might be rude to say about it, but LSTMs need ditching. While the internal state mechanism it uses provides a crucial insight in how to stabilize the training RNNs for long sequences, it still does mean that gradient descent can do the right thing for deep sequences.<\/p>\n<p>LSTM trains better than a vanilla RNN. It has more feature than it. Its defining equation is definitely bigger. But from a different perspective, LSTM does not improve the generality of a standard RNN, but rather it restricts it. Batch norm, layer norm, activation functions, residual connections...all the significant improvements over the base network are in essence constraints.<\/p>\n<p>Dilated connections by restricting the amount of work the net can do increase its power and shorten the gradient path. They should be the starting point for deep recurrent networks.<\/p>\n<p>Dilated connections will require the generic gemm and redoing the loop bodies in the library.<\/p>\n<h5><a id=\"user-content-dense-connections\" class=\"anchor\" aria-hidden=\"true\" href=\"#dense-connections\"><\/a>Dense Connections<\/h5>\n<p>Originally introduced in the <a href=\"https:\/\/arxiv.org\/abs\/1608.06993\" rel=\"nofollow\">Densely Connected Convolutional Networks<\/a>, the dense connections are a general method of significantly easing the training the networks in the vertical direction. They can be applied to both feedforward and recurrent networks and the author is interested in their application towards the later.<\/p>\n<p>Given how much improvement they demonstrate for deep networks, the author's view is that they are integral component of all modern networks and they should be a permanent fixture in any kind of net.<\/p>\n<p>Dense connections will require the generic gemm and redoing the loop bodies in the library.<\/p>\n<h5><a id=\"user-content-deep-layer-compilation-optimizations\" class=\"anchor\" aria-hidden=\"true\" href=\"#deep-layer-compilation-optimizations\"><\/a>Deep Layer Compilation Optimizations<\/h5>\n<p>In particular, deep layer of over 10 will require some modifications to how kernels are compiled. Instead of passing in arguments directly at some point it will be necessary to pass them in array in order to preserve the user's sanity. That will require implementing asynchronous memory copies from host to device among others. Without this, kernels will need to be specialized for every single argument change. For nets with 50 layers, it would not be good if there were 50 different variations of gemm.<\/p>\n<p>The priority of this item depends on the kinds of networks being compiled. Right now it is completely unnecessary, but eventually it will be indispensable.<\/p>\n<h5><a id=\"user-content-generic-convolutions\" class=\"anchor\" aria-hidden=\"true\" href=\"#generic-convolutions\"><\/a>Generic Convolutions<\/h5>\n<p>This one should be done at some point as well for the sake of completeness, though the CuDNN one will suffice in the meantime. Probably there will be new kinds of operations in the future. There is no way that matrix multiply and convolutions are the only two possible choices. Hence mastering convolutions should be on the roadmap at some point. Doing that would also have the benefit of completely breaking the dependency on the CuDNN library.<\/p>\n<p>Divisive normalization should also be implemented at some point and that can be expressed in terms of the generic convolution operator.<\/p>\n<h5><a id=\"user-content-1d-block-scan-and-reduce\" class=\"anchor\" aria-hidden=\"true\" href=\"#1d-block-scan-and-reduce\"><\/a>1D Block Scan and Reduce<\/h5>\n<p>The author already did the transposed versions of them, but in order to take advantage of Spiral's default toa representation, it might be good to break the dependency on Cub for the 1d case. C++ is really limited in various ways, and much like the majority of statically typed languages falls flat when it comes to defining new datatypes. It is not a high priority item and more like a homework assignment suitable for newcomers to the language should they feel like it.<\/p>\n<h5><a id=\"user-content-softmax-cost\" class=\"anchor\" aria-hidden=\"true\" href=\"#softmax-cost\"><\/a>Softmax Cost<\/h5>\n<p>It would be better to recalculate the input much like in layer norm. Right now it is implemented inefficiently. The author is learning the language he created as he goes along much like anyone else would. He did not think through the current implementation properly.<\/p>\n<h5><a id=\"user-content-checkpointing\" class=\"anchor\" aria-hidden=\"true\" href=\"#checkpointing\"><\/a>Checkpointing<\/h5>\n<p>The ability to save arbitrary types to disk and load them will get added to the language and the standard library. Checkpointing for the ML library will succinctly get built on top of that. This is something that will be necessary when training times get longer.<\/p>\n<h5><a id=\"user-content-improve-the-allocator\" class=\"anchor\" aria-hidden=\"true\" href=\"#improve-the-allocator\"><\/a>Improve The Allocator<\/h5>\n<p>Allocator should be more capable of handling short term allocations. This can be done by taking the idea from the previous stack based allocator and incorporating them into the current one.<\/p>\n<h5><a id=\"user-content-plastic-neurons\" class=\"anchor\" aria-hidden=\"true\" href=\"#plastic-neurons\"><\/a>Plastic Neurons<\/h5>\n<p>The results in the <a href=\"https:\/\/arxiv.org\/abs\/1804.02464\" rel=\"nofollow\">Differentiable Plasticity<\/a> are amazing and have completely convinced the author that these kinds of neurons are an essential architectural feature for what I want to do. Adding them to the library would be trivial with a generic gemm kernel, which just puts more weight on implementing it.<\/p>\n<p>The author expected there would be novel architectural discoveries necessitating the generic gemm beyond dense connections, but he was surprised how quickly it happened.<\/p>\n<h2><a id=\"user-content-user-guide-the-spiral-power\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-guide-the-spiral-power\"><\/a>User Guide: The Spiral Power<\/h2>\n<p>While the tutorial was meant to be a light introduction to the language, the user guide is intended to be an in depth guide not just into the language, but into the workings of the compiler. It is not intended to be exhaustive, but to cover the core language features and their implementation for the sake of making the user more confident in traversing the Spiral's source should that be needed.<\/p>\n<p>As it currently stands, the compiler in its entirety is at about 4.1k LOC. And while the parser and the codegens have their charm points, the part of interest which is the partial evaluator for Spiral is only 1.6k LOC. Out of that 1.6k probably around 1000 lines are vital to its functioning, meaning as the language continues to grow and more operatives get added, the essence of it should remain intact.<\/p>\n<p>It took the author 14 months to get the language up to this point, but with it as a reference it would be closer to 1 month long student project in terms of difficulty if a sensible tool like F# is used instead of C++.<\/p>\n<p>The first 9 chapters document the language's core. The rest can be read in any order.<\/p>\n<h3><a id=\"user-content-1-data-structures-abstraction-and-destructuring\" class=\"anchor\" aria-hidden=\"true\" href=\"#1-data-structures-abstraction-and-destructuring\"><\/a>1: Data Structures, Abstraction and Destructuring<\/h3>\n<pre>and TypedExpr =\n    \/\/ Data structures\n    | TyT of Ty\n    | TyV of TyTag \/\/ int * Ty\n    | TyBox of TypedExpr * Ty\n    | TyList of TypedExpr list\n    | TyMap of EnvTerm * MapType\n    | TyLit of Value\n    \/\/ Operations\n    | TyLet of TyTag * TypedExpr * TypedExpr * Ty * Trace\n    | TyState of TypedExpr * TypedExpr * Ty * Trace\n    | TyOp of Op * TypedExpr list * Ty\n    | TyJoinPoint of JoinPointKey * JoinPointType * Arguments * Ty\n<\/pre>\n<p>Nearly everything that can be done is Spiral is by manipulating the above 6 data structures. Chop off the first 3 and Spiral would be a pure dynamic functional language. <code>TyBox<\/code> represents a staged union type. <code>TyMap<\/code> is reused for both functions and modules. Modules are in essence functions without a body, they can be thought of as first class environments. In <code>TyMap<\/code> <code>EnvTerm<\/code> is almost the immutable map and can be thought of as that for all intents and purposes for now.<\/p>\n<p>The dogma of information flows in Spiral is to let as much of it through forward. Inlineables always preserve all information and unlike most other languages, Spiral allows the exact structures, that is <code>TypedExpr<\/code>s themselves, to be propagated forward through specialization boundaries. In Spiral those are join points, but in most other languages those are standard functions.<\/p>\n<p>In order to for there to be a separation between compile time and runtime, abstraction is necessary.<\/p>\n<p>Abstraction is the process of turning a non-variable (<code>TyBox<\/code>,<code>TyList<\/code>,<code>TyMap<\/code>,<code>TyLit<\/code>) into an abstract variable. There are 3 ways of doing that: join point returns, and dynamic if branch and dynamic case returns.<\/p>\n<p>When that is done, apart from what is preserved in the common subexpression dictionary, all information about the structure apart from its type is thrown away and it is replaced with a <code>TyV<\/code>. <code>TyV<\/code> is just a tag and a type representing a variable.<\/p>\n<p>Here are a few examples just to make sure the lesson sticks. It is rough typing the following out by hand. Keeping track of information is why compilers exist. The more a language allows the user to reason locally, the better it is at doing its job.<\/p>\n<pre>TyLit (TyInt64 1) -&gt; TyV(100, PrimT Int64T)\nTyBox (TyT(LitT (LitString \"None\")),UnionT {LitT (LitString \"None\"); ListT [LitT (LitString Some); PrimT Int64T]}) -&gt; TyV(101,UnionT {LitT (LitString \"None\"); ListT [LitT (LitString Some); PrimT Int64T]})\nTyList [TyV(2,PrimT Float32T); TyLit (LitString \"Hello\"); TyV(3,PrimT CharT)] -&gt; TyV(102,ListT [PrimT Float32T; PrimT StringT; PrimT CharT])\n<\/pre>\n<p>Destructuring is what usually comes after abstraction as it is used so often in the language. It is not the opposite of abstraction - it can't be since abstraction is the process throwing away information in a principled manner. Once that information is lost it can't be recovered.<\/p>\n<p>Destructuring is always done at bindings and at list and map creation and before join points entries, so fully abstracted variables will never appear in the environment when the printed via <code>print_static<\/code>.<\/p>\n<p>Here is how the process roughly works.<\/p>\n<pre>TyV(100, PrimT Int64T) -&gt; TyV(100, PrimT Int64T) \/\/ no change\nTyV(101,UnionT {LitT (LitString \"None\"); ListT [LitT (LitString Some); PrimT Int64T]}) -&gt; TyV(101,UnionT {LitT (LitString \"None\"); ListT [LitT (LitString Some); PrimT Int64T]}) \/\/ no change\nTyV(102,ListT [PrimT Float32T; PrimT StringT; PrimT CharT]) -&gt; TyList [TyV(103,PrimT Float32T); TyV(104,PrimT StringT); TyV(105,PrimT CharT)]\n<\/pre>\n<p>Spiral is crazy about turning variables inside out whenever it can. This has the effect of flattening them when they are passed through join points and makes it easy to support having partially static maps and lists in the language. Note that this does not change their type in any way. In the above examples the left and the right sides have completely equal types.<\/p>\n<p><code>destructure<\/code> is probably the single most important function in the language as without it everything else would be impossible. It is the first hurdle to overcome when making a language with first class staging.<\/p>\n<p>Note the the way F# for example or other functional languages do destructuring is not the same. There is a notable difference between the philosophy of Spiral and the rest of the pack. Spiral follows the dogma of <code>inline first and optimize later<\/code> while generally high level languages inherit the Lisp philosophy of <code>heap allocate first and optimize later<\/code>.<\/p>\n<p>That means in F# for example pattern matching is necessary to destructure a tuple since otherwise they would be packed. This is done at runtime. Here is a look into how F# does it based on its AST. The following examples uses its <a href=\"https:\/\/docs.microsoft.com\/en-us\/dotnet\/fsharp\/language-reference\/code-quotations\" rel=\"nofollow\">code quotation<\/a> feature.<\/p>\n<pre>let x = 1,2,3\nlet t =\n    &lt;@\n    let a,b,c = x\n    ()\n    @&gt;\n<\/pre>\n<pre>val x : int * int * int = (1, 2, 3)\nval t : Quotations.Expr&lt;unit&gt; =\n  Let (c, TupleGet (PropertyGet (None, x, []), 2),\n     Let (b, TupleGet (PropertyGet (None, x, []), 1),\n          Let (a, TupleGet (PropertyGet (None, x, []), 0), Value (&lt;null&gt;))))\n<\/pre>\n<p>Most compilers worth their salt would optimize this away, but Spiral's optimizations are actually a part of its semantics thereby making them guarantees. Replacing optimizations with guarantees and making it a part of the type system is important for getting predictable and predictably good performance.<\/p>\n<p>In Spiral something like a simple binding would mean something completely different.<\/p>\n<pre>inl a,b,c = x\n<\/pre>\n<p>There is no way that in the above example that <code>x<\/code> would not have already been destructured, hence what that tuple pattern does is tries to unbox a union type, but otherwise does absolutely nothing at runtime. It adds a few names to the environment and that is it. <code>destructure<\/code> would be called and then it would do nothing once it saw <code>TyList<\/code> as input.<\/p>\n<pre>inl a,b,c = join 1,2,3\n<\/pre>\n<p>Now if it was something like the above then it would be doing destructuring. It would go roughly <code>1,2,3 -&gt; var [int64; int64; int64]<\/code> in the join point and then <code>var [int64; int64; int64] -&gt; [var int64; var int64; var int64]<\/code> on destructuring.<\/p>\n<p>Keeping the environments fully destructured at all times is important for than just efficiency at runtime, making specialization coherent, supporting partially static structures and making other operations easier - it also solves one of the great language interop problems. \"How are arguments to be passed through join points to other languages?\"<\/p>\n<p>It is impossible to pass tuples of primitives to the Cuda side in their default form. The reason for that is that C might decide to layout the struct differently than the .NET JIT does and there is no way to access the layout information directly.<\/p>\n<p>By flattening structures out of the way completely at runtime, interop becomes a much easier task. That is one of the main motivations for supporting tuple of array tensors as the default as well.<\/p>\n<p>The time it took for the author to write the <code>destructure<\/code> function can be measured in many months.<\/p>\n<pre>let rec destructure (d: LangEnv) (r: TypedExpr): TypedExpr = \n    let inline destructure r = destructure d r\n    let inline cse_eval on_succ on_fail r = \n        match Map.tryFind r !d.cse_env with\n        | Some x -&gt; on_succ x\n        | None -&gt; on_fail r\n    let inline cse_recurse r = cse_eval destructure id r\n    let inline let_insert_cse r = \n        cse_eval \n            cse_recurse\n            (fun r -&gt;\n                let x = make_tyv_and_push_typed_expr d r\n                cse_add d r x\n                x)\n            r\n    let index_tuple_args r tuple_types = \n        let unpack (s,i as state) typ = \n            if is_unit typ then (tyt typ |&gt; destructure) :: s, i\n            else (destructure &lt;| TyOp(ListIndex,[r;TyLit &lt;| LitInt64 (int64 i)],typ)) :: s, i + 1\n        List.fold unpack ([],0) tuple_types\n        |&gt; fst |&gt; List.rev\n    let env_unseal r x =\n        let unseal (m,i as state) (k: string) typ = \n            if is_unit typ then Map.add k (tyt typ |&gt; destructure) m, i\n            else\n                let r = TyOp(MapGetField,[r; tyv(i,typ)], typ) |&gt; destructure \n                Map.add k r m, i + 1\n        Map.fold unseal (Map.empty,0) x |&gt; fst\n    let inline destructure_var r map_vvt map_funt =\n        match get_type r with\n        | ListT tuple_types -&gt; tyvv(map_vvt tuple_types)\n        | MapT (env,t) -&gt; tymap(map_funt env, t)\n        | _ -&gt; cse_recurse r\n    \n    match r with\n    | TyMap _ | TyList _ | TyLit _ -&gt; r\n    | TyBox _ -&gt; cse_recurse r\n    | TyT _ -&gt; destructure_var r (List.map (tyt &gt;&gt; destructure)) (Map.map (fun _ -&gt; (tyt &gt;&gt; destructure)) &gt;&gt; Env)\n    | TyV _ -&gt; destructure_var r (index_tuple_args r) (env_unseal r &gt;&gt; Env)\n    | TyOp _ -&gt; let_insert_cse r\n    | TyJoinPoint _ | TyLet _ | TyState _ -&gt; on_type_er (trace d) \"Compiler error: This should never appear in destructure. It should go directly into d.seq.\"\n<\/pre>\n<p>What was covered so far is essentially the <code>TyV _<\/code> case. <code>TyT _<\/code> case works very similarly to that except there is no need to make any work for the code generator in the later stage. Understanding the specifics of the function is not necessary at this point. What is desired is to show that there is a direct mapping between what was talked about and code.<\/p>\n<p>If one can understand the above function and how join points work, that would be enough to understand 90% of Spiral so it a goal to strive for.<\/p>\n<p>Based on what was discussed up to now, the first time reader of the user guide's model should roughly be as if the above function was written like this. The following is <code>destructure<\/code> without common subexpression elimination.<\/p>\n<pre>let rec destructure (d: LangEnv) (r: TypedExpr): TypedExpr = \n    let inline destructure r = destructure d r\n    let index_tuple_args r tuple_types = \n        let unpack (s,i as state) typ = \n            if is_unit typ then (tyt typ |&gt; destructure) :: s, i\n            else (destructure &lt;| TyOp(ListIndex,[r;TyLit &lt;| LitInt64 (int64 i)],typ)) :: s, i + 1\n        List.fold unpack ([],0) tuple_types\n        |&gt; fst |&gt; List.rev\n    let env_unseal r x =\n        let unseal (m,i as state) (k: string) typ = \n            if is_unit typ then Map.add k (tyt typ |&gt; destructure) m, i\n            else\n                let r = TyOp(MapGetField,[r; tyv(i,typ)], typ) |&gt; destructure \n                Map.add k r m, i + 1\n        Map.fold unseal (Map.empty,0) x |&gt; fst\n    let inline destructure_var r map_vvt map_funt =\n        match get_type r with\n        | ListT tuple_types -&gt; tyvv(map_vvt tuple_types)\n        | MapT (env,t) -&gt; tymap(map_funt env, t)\n        | _ -&gt; cse_recurse r\n    \n    match r with\n    | TyMap _ | TyList _ | TyLit _ -&gt; r\n    | TyBox _ -&gt; make_tyv_and_push_typed_expr d r\n    | TyT _ -&gt; destructure_var r (List.map (tyt &gt;&gt; destructure)) (Map.map (fun _ -&gt; (tyt &gt;&gt; destructure)) &gt;&gt; Env)\n    | TyV _ -&gt; destructure_var r (index_tuple_args r) (env_unseal r &gt;&gt; Env)\n    | TyOp _ -&gt; make_tyv_and_push_typed_expr d r\n    | TyJoinPoint _ | TyLet _ | TyState _ -&gt; on_type_er (trace d) \"Compiler error: This should never appear in destructure. It should go directly into d.seq.\"\n<\/pre>\n<p><code>destructure_var<\/code> can be completely thought of as a recursive map.<\/p>\n<p>Suppose something like <code>var [int64; int64]<\/code> bound to <code>var_12<\/code> is destructured.<\/p>\n<p>Inside <code>destructure<\/code> the type of the variable is iterated over.<\/p>\n<p>The first element is converted to <code>list_index var_12 0<\/code>. Then a recursive call is made on that. What happens is that then the <code>TyOp _<\/code> case gets hit. <code>make_tyv_and_push_typed_expr<\/code> then converts that into a <code>TyV<\/code> which then gets returned.<\/p>\n<p>The second element is converted to <code>list_index var_12 1<\/code>. Then a recursive call is made on that. What happens is that then the <code>TyOp _<\/code> case gets hit. <code>make_tyv_and_push_typed_expr<\/code> then converts that into a <code>TyV<\/code> which then gets returned.<\/p>\n<p>The reason why this is done recursively is because lists might have sublists and such. Without that only a single level would get destructured which would be poor.<\/p>\n<p>The reason why first converting to a <code>TyOp(ListIdex,[var_12,0])<\/code> and such is needed is for the code generation pass. As can be inferred from that, it stands to reason that <code>make_tyv_and_push_typed_expr<\/code> is doing something to pass that information along. What is that function is doing is commonly known as let insertion.<\/p>\n<h3><a id=\"user-content-2-let-insertion-and-common-subexpression-elimination\" class=\"anchor\" aria-hidden=\"true\" href=\"#2-let-insertion-and-common-subexpression-elimination\"><\/a>2: Let Insertion and Common Subexpression Elimination<\/h3>\n<p>When the code is generated, it tends to be a series of <code>let<\/code> statements punctuated by an expression.<\/p>\n<pre>let var_0 = ...\nlet var_1 = ...\nlet var_2 = ...\n...\n<\/pre>\n<p>Based on that information it be guessed that during evaluation a list of let statements is maintained in the environment that gets added every time <code>make_tyv_and_push_typed_expr<\/code> gets called. That is in fact exactly what happens, but the way the function is implemented might seem convoluted at first.<\/p>\n<pre>let inline make_tyv_and_push_typed_expr_template (even_if_unit: bool) (d: LangEnv) (ty_exp: TypedExpr): TypedExpr =\n    let ty = get_type ty_exp\n    if is_unit ty then\n        if even_if_unit then\n            let seq = !d.seq\n            let trace = trace d\n            d.seq := fun rest -&gt; TyState(ty_exp,rest,get_type rest,trace) |&gt; seq\n        tyt ty\n    else\n        let v = make_tyv_ty d ty\n        let seq = !d.seq\n        let trace = trace d\n        d.seq := fun rest -&gt; TyLet(v,ty_exp,rest,get_type rest,trace) |&gt; seq\n        tyv v\n    \nlet make_tyv_and_push_typed_expr_even_if_unit d ty_exp = make_tyv_and_push_typed_expr_template true d ty_exp\nlet make_tyv_and_push_typed_expr d ty_exp = make_tyv_and_push_typed_expr_template false d ty_exp\n<\/pre>\n<p>There are actually two variants of the function packed into a single one based on whether the type of the expression is a unit. Using some partial evaluation by hand, it should be possible to clarify the function by splitting them into two.<\/p>\n<pre>let make_tyv_and_push_typed_expr_even_if_unit (d: LangEnv) (ty_exp: TypedExpr): TypedExpr = \/\/make_tyv_and_push_typed_expr_template true d ty_exp\n    let ty = get_type ty_exp\n    let seq = !d.seq\n    let trace = trace d\n    \n    if is_unit ty then\n        d.seq := fun rest -&gt; TyState(ty_exp,rest,get_type rest,trace) |&gt; seq\n        tyt ty\n    else\n        let v = make_tyv_ty d ty\n        d.seq := fun rest -&gt; TyLet(v,ty_exp,rest,get_type rest,trace) |&gt; seq\n        tyv v\nlet make_tyv_and_push_typed_expr (d: LangEnv) (ty_exp: TypedExpr): TypedExpr = \/\/make_tyv_and_push_typed_expr_template false d ty_exp\n    let ty = get_type ty_exp\n    if is_unit ty then\n        tyt ty\n    else\n        let v = make_tyv_ty d ty\n        let seq = !d.seq\n        let trace = trace d\n        d.seq := fun rest -&gt; TyLet(v,ty_exp,rest,get_type rest,trace) |&gt; seq\n        tyv v\n<\/pre>\n<p>With this rewrite hopefully the distinction between the two functions should be clearer.<\/p>\n<p><code>d.seq<\/code> is in fact a list, but it is implemented as a closure of type <code>(TypedExpr -&gt; TypedExpr) ref<\/code>. To explain the reason for that, a look at the full <code>TypedExpr<\/code> type is needed once again.<\/p>\n<pre>and TypedExpr =\n    \/\/ Data structures\n    | TyT of Ty\n    | TyV of TyTag\n    | TyBox of TypedExpr * Ty\n    | TyList of TypedExpr list\n    | TyMap of EnvTerm * MapType\n    | TyLit of Value\n    \/\/ Operations\n    | TyLet of TyTag * TypedExpr * TypedExpr * Ty * Trace\n    | TyState of TypedExpr * TypedExpr * Ty * Trace\n    | TyOp of Op * TypedExpr list * Ty\n    | TyJoinPoint of JoinPointKey * JoinPointType * Arguments * Ty\n<\/pre>\n<p>The way it is implemented like that is because of <code>TyLet<\/code>. Using a list directly would be annoying because <code>TyLet<\/code> is a list itself and it would be troublesome to add to end if the arraignment was different.<\/p>\n<p>So the solution is to just do some lambda magic to do that instead and emulate an immutable stack that way in a statically safe manner. The same goes for <code>TyState<\/code> which is like <code>TyLet<\/code> except for side effecting operations and not new bindings.<\/p>\n<p><code>make_tyv_and_push_typed_expr<\/code> and <code>make_tyv_and_push_typed_expr_even_if_unit<\/code> are used for non side effecting and side effecting operations respectively. Destructuring that was shown in the previous chapter is an example of the former, for example it should not happen that runtime unit types such as type level literals should ever be instantiated. As those get printed as \"\" during code generation trying to do so would give an error when the F# or Cuda compilers get the code.<\/p>\n<p>As a part of general mechanism of supporting the first class types in Spiral, in both let insertion and destructuring any operations that would produce a unit type gets converted to a unwrapped (naked) type. Side effecting or potentially side effecting operations such as join points therefore must call <code>make_tyv_and_push_typed_expr_even_if_unit<\/code> before <code>destructure<\/code> gets to it otherwise it will cause errors. For example if <code>make_tyv_and_push_typed_expr_even_if_unit<\/code> was not called inside join points then those with unit type as return would not get printed at all.<\/p>\n<p>Even worse, repeating join points in local scope would get rewritten to the same variable. This would not be a problem for pure functions, but for side effecting ones it would lead to erroneous results.<\/p>\n<p>The way (CSE) common subexpression elimination works is simple in Spiral.<\/p>\n<pre>inl q = x + 3\n...\ninl w = x + 3\n...\n<\/pre>\n<p>All CSE happens in <code>destructure<\/code>. What happens is that the call to <code>destructure<\/code> gets <code>TyOp(Add,[TyV(5,PrimT Int64T); TyLit (LitInt64 3L)])<\/code> or similar as the argument. It hits the <code>TyOp<\/code> branch and then performs a check against the CSE dictionary with is a standard map of type <code>Map&lt;TypedExpr,TypedExpr&gt; ref<\/code>.<\/p>\n<p>The first time it happens in <code>inl q = x + 3<\/code>, the key is not present in the dictionary so what it does is calls <code>make_tyv_and_push_typed_expr<\/code> with the argument. It gets back something like <code>TyV(9,PrimT Int64T)<\/code>. That gets bound to the key and added to the dictionary like <code>Map.add (TyOp(Add,[TyV(5,PrimT Int64T); TyLit (LitInt64 3L)])) (TyV(9,PrimT Int64T))<\/code>.<\/p>\n<p>The actual code is a tad more complicated.<\/p>\n<pre>let cse_add' d r x = let e = !d.cse_env in if r &lt;&gt; x then Map.add r x e else e\nlet cse_add d r x = d.cse_env := cse_add' d r x\n<\/pre>\n<p>There is a check to make sure that the variable being added to the map is not itself because that could lead to divergence, but otherwise the intent should be straightforward.<\/p>\n<p>Once the new key and its value in the map, the program proceeds and <code>inl w = x + 3<\/code> is hit eventually. As the arguments flows through <code>destructure<\/code> the check for <code>TyOp(Add,[TyV(5,PrimT Int64T); TyLit (LitInt64 3L)])<\/code> and this time the key is present dictionary. Since it is present, instead calling <code>make_tyv_and_push_typed_expr<\/code> to perform let insertion, the expression just gets rewritten.<\/p>\n<p>Note that this happens recursively. Here is the full <code>destructure<\/code> once again.<\/p>\n<pre>let rec destructure (d: LangEnv) (r: TypedExpr): TypedExpr = \n    let inline destructure r = destructure d r\n    let inline cse_eval on_succ on_fail r = \n        match Map.tryFind r !d.cse_env with\n        | Some x -&gt; on_succ x\n        | None -&gt; on_fail r\n    let inline cse_recurse r = cse_eval destructure id r\n    let inline let_insert_cse r = \n        cse_eval \n            cse_recurse\n            (fun r -&gt;\n                let x = make_tyv_and_push_typed_expr d r |&gt; destructure\n                cse_add d r x\n                x)\n            r\n    let index_tuple_args r tuple_types = \n        let unpack (s,i as state) typ = \n            if is_unit typ then (tyt typ |&gt; destructure) :: s, i\n            else (destructure &lt;| TyOp(ListIndex,[r;TyLit &lt;| LitInt64 (int64 i)],typ)) :: s, i + 1\n        List.fold unpack ([],0) tuple_types\n        |&gt; fst |&gt; List.rev\n    let env_unseal r x =\n        let unseal (m,i as state) (k: string) typ = \n            if is_unit typ then Map.add k (tyt typ |&gt; destructure) m, i\n            else\n                let r = TyOp(MapGetField,[r; tyv(i,typ)], typ) |&gt; destructure \n                Map.add k r m, i + 1\n        Map.fold unseal (Map.empty,0) x |&gt; fst\n    let inline destructure_var r map_vvt map_funt =\n        match get_type r with\n        | ListT tuple_types -&gt; tyvv(map_vvt tuple_types)\n        | MapT (env,t) -&gt; tymap(map_funt env, t)\n        | _ -&gt; cse_recurse r\n    \n    match r with\n    | TyMap _ | TyList _ | TyLit _ -&gt; r\n    | TyBox _ -&gt; cse_recurse r\n    | TyT _ -&gt; destructure_var r (List.map (tyt &gt;&gt; destructure)) (Map.map (fun _ -&gt; (tyt &gt;&gt; destructure)) &gt;&gt; Env)\n    | TyV _ -&gt; destructure_var r (index_tuple_args r) (env_unseal r &gt;&gt; Env)\n    | TyOp _ -&gt; let_insert_cse r\n    | TyJoinPoint _ | TyLet _ | TyState _ -&gt; on_type_er (trace d) \"Compiler error: This should never appear in destructure. It should go directly into d.seq.\"\n<\/pre>\n<p>And here it is pared down so that the CSE using parts are highlighted.<\/p>\n<pre>let rec destructure (d: LangEnv) (r: TypedExpr): TypedExpr = \n    let inline destructure r = destructure d r\n    let inline cse_eval on_succ on_fail r = \n        match Map.tryFind r !d.cse_env with\n        | Some x -&gt; on_succ x\n        | None -&gt; on_fail r\n    let inline cse_recurse r = \n        cse_eval \n            destructure \/\/ on_succ\n            id \/\/ on_fail\n            r\n    let inline let_insert_cse r = \n        cse_eval \n            cse_recurse \/\/ on_succ\n            (fun r -&gt; \/\/ on_fail\n                let x = make_tyv_and_push_typed_expr d r |&gt; destructure\n                cse_add d r x\n                x)\n            r\n    \n    match r with\n    | TyMap _ | TyList _ | TyLit _ -&gt; r\n    | TyBox _ -&gt; cse_recurse r\n    | TyOp _ -&gt; let_insert_cse r\n    | TyJoinPoint _ | TyLet _ | TyState _ -&gt; on_type_er (trace d) \"Compiler error: This should never appear in destructure. It should go directly into d.seq.\"\n<\/pre>\n<p>It is possible to simplify it even further.<\/p>\n<pre>let rec destructure (d: LangEnv) (r: TypedExpr): TypedExpr = \n    let inline destructure r = destructure d r\n    let cse_recurse r = \n        match Map.tryFind r !d.cse_env with\n        | Some x -&gt; destructure x\n        | None -&gt; r\n    let let_insert_cse r = \n        match Map.tryFind r !d.cse_env with\n        | Some x -&gt; cse_recurse x\n        | None -&gt; \n            let x = make_tyv_and_push_typed_expr d r |&gt; destructure\n            cse_add d r x\n            x\n    \n    match r with\n    | TyMap _ | TyList _ | TyLit _ -&gt; r\n    | TyBox _ -&gt; cse_recurse r\n    | TyOp _ -&gt; let_insert_cse r\n    | TyJoinPoint _ | TyLet _ | TyState _ -&gt; on_type_er (trace d) \"Compiler error: This should never appear in destructure. It should go directly into d.seq.\"\n<\/pre>\n<p>This should hopefully be clear. <code>cse_recurse<\/code> does not call itself recursively. It does call <code>destructure<\/code> instead, but for all intents and purposes the function can thought of calling itself recursively.<\/p>\n<p>Suppose that the dictionary is like the following (in pseudo-code):<\/p>\n<pre>TyOp(Add,[TyV(5,PrimT Int64T); TyLit (LitInt64 3L)]) =&gt; TyV(9,PrimT Int64T)\nTyV(9,PrimT Int64T) =&gt; TyV(12,PrimT Int64T)\n<\/pre>\n<p>So <code>x+3<\/code> will first get rewritten to <code>TyV(9,PrimT Int64T)<\/code>. Then <code>destructure<\/code> will be called again.<\/p>\n<p>The <code>| TyV _ -&gt; destructure_var r (index_tuple_args r) (env_unseal r &gt;&gt; Env)<\/code> case will get hit. After that happens it will attempt to pattern match on this.<\/p>\n<pre>let inline destructure_var r map_vvt map_funt =\n    match get_type r with\n    | ListT tuple_types -&gt; tyvv(map_vvt tuple_types)\n    | MapT (env,t) -&gt; tymap(map_funt env, t)\n    | _ -&gt; cse_recurse r\n<\/pre>\n<p>It is not <code>ListT<\/code> nor <code>MapT<\/code> so <code>cse_recurse<\/code> in the last branch gets called. Then <code>TyV(9,PrimT Int64T)<\/code> would get rewritten to <code>TyV(12,PrimT Int64T)<\/code>.<\/p>\n<p><code>let_insert_cse<\/code> is just the same as <code>cse_recurse<\/code> apart from the extra step of doing let insertion.<\/p>\n<p>With this the let insertion and destructuring have been covered in full. The functions related to them serve as focal points of the entire language and with the understanding of them the paths to understanding of everything else become unblocked. In programming in general, programs tend to be easier to write than to read.<\/p>\n<p>Those two functions on the other hand are definitely on the exact opposite of the spectrum as they literally took months and months of refinement to make.<\/p>\n<p>CSE rewriting has some additional uses which will be covered in the following chapter.<\/p>\n<h3><a id=\"user-content-3-the-if-statement\" class=\"anchor\" aria-hidden=\"true\" href=\"#3-the-if-statement\"><\/a>3: The If Statement<\/h3>\n<p>Spiral originally started out with the dynamic <code>if<\/code> as is found in most statically typed languages, but as an experiment the author decided to make the static the default one. After two month of use, he noted that not even once had he used the other version and when he ran into a bug with the dynamic <code>if<\/code> that cemented its removal for good from the language.<\/p>\n<p>Here is the static <code>if<\/code> in full.<\/p>\n<pre>let if_static (d: LangEnv) (cond: Expr) (tr: Expr) (fl: Expr): TypedExpr =\n    match tev d cond with\n    | TyLit (LitBool true) -&gt; tev d tr\n    | TyLit (LitBool false) -&gt; tev d fl\n    | TyType (PrimT BoolT) as cond -&gt;\n        let b x = cse_add' d cond (TyLit &lt;| LitBool x)\n        let tr = tev_assume (b true) d tr\n        let fl = tev_assume (b false) d fl\n        let type_tr, type_fl = get_type tr, get_type fl\n        if type_tr = type_fl then\n            match tr, fl with\n            | TyLit (LitBool true), TyLit (LitBool false) -&gt; cond\n            | _ when tr = fl -&gt; tr\n            | _ -&gt; TyOp(IfStatic,[cond;tr;fl],type_tr) |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n        else on_type_er (trace d) &lt;| sprintf \"Types in branches of If do not match.\\nGot: %s and %s\" (show_ty type_tr) (show_ty type_fl)\n    | TyType cond -&gt; on_type_er (trace d) &lt;| sprintf \"Expected a bool in conditional.\\nGot: %s\" (show_ty cond)\n<\/pre>\n<p>The first 2 cases should be straightforward.<\/p>\n<pre>let if_static (d: LangEnv) (cond: Expr) (tr: Expr) (fl: Expr): TypedExpr =\n    match tev d cond with\n    | TyLit (LitBool true) -&gt; tev d tr\n    | TyLit (LitBool false) -&gt; tev d fl\n<\/pre>\n<p>At first the conditional is evaluated and if it turns out to be a literal only one of the branches is evaluated.<\/p>\n<pre>    | TyType (PrimT BoolT) as cond -&gt;\n        let b x = cse_add' d cond (TyLit &lt;| LitBool x)\n        let tr = tev_assume (b true) d tr\n        let fl = tev_assume (b false) d fl\n<\/pre>\n<p>Here the two branches are evaluated with an little extra twist. Suppose an example like the following.<\/p>\n<pre>if x &gt; 0 then\n    if x &gt; 0 then\n        123\n    else\n        \"qwe\"\nelse\n    456\n<\/pre>\n<p>In the outer if the conditional <code>x &gt; 0<\/code> will evaluate to an abstract variable and in <code>if_static<\/code> the third branch will get hit.<\/p>\n<p>Then once the conditional of the inner if is evaluated, what it will find in the CSE dictionary is that the <code>x &gt; 0<\/code> can be rewritten to <code>true<\/code>.<\/p>\n<pre>        let b x = cse_add' d cond (TyLit &lt;| LitBool x)\n        let tr = tev_assume (b true) d tr\n<\/pre>\n<p>The above two lines are responsible for this.<\/p>\n<p>Here is how <code>tev_assume<\/code> is implemented.<\/p>\n<pre>\/\/ #Type directed partial evaluation\nlet rec expr_peval (d: LangEnv) (expr: Expr): TypedExpr =\n    let inline tev d expr = expr_peval d expr\n    let inline apply_seq d x = !d.seq x\n    let inline tev_seq d expr = let d = {d with seq=ref id; cse_env=ref !d.cse_env} in tev d expr |&gt; apply_seq d\n    let inline tev_assume cse_env d expr = let d = {d with seq=ref id; cse_env=ref cse_env} in tev d expr |&gt; apply_seq d\n<\/pre>\n<p><code>tev<\/code> itself just calls the main evaluation functions. The <code>expr_eval<\/code> function is one of the core passes of Spiral as that is where all the partial evaluation happens. The rest are parsing and code generation.<\/p>\n<p>As <code>tev_seq<\/code> and <code>tev_assume<\/code> are very similar, it would be easier to start with an explanation of <code>tev_seq<\/code>.<\/p>\n<p>In Spiral there are constructs such as if statements, case expressions and join points which require their own scope.<\/p>\n<p>On entry therefore, the let statements that are being carried in the environment need to be cleared instead of being carried into the branch. <code>seq=ref id<\/code> is what represents this. If it was a standard list, it would just be set to <code>[]<\/code>, but here it is set to the identity function.<\/p>\n<p>Then there is the CSE dictionary that also needs to be handled. Since Spiral is lexically scoped it is fine if the expressions inside the local scope get rewritten to expressions that came before it, but what should not happen after the scope is exited for expressions to get rewritten with the ones that are now out of scope.<\/p>\n<pre>if c then \n    inl q = x + 2\n    ...\nelse\n    ...\ninl e = x + 2 \/\/ Do not want this to be rewritten to `q` inside the if statement.\n<\/pre>\n<p>If <code>cse_env<\/code> was a standard .NET mutable dictionary it would need to be copied at this point so that invariant is ensured, but since it is an F# immutable map, what should just be done is replacing the reference to it.<\/p>\n<p><code>tev_assume<\/code> has similar intent to <code>tev_seq<\/code> except the map for <code>cse_env<\/code> is taken as an argument; it is a more generic version of <code>tev_seq<\/code>.<\/p>\n<p>With this it has been established that the two functions just evaluate an expression in its own scope.<\/p>\n<p><code>tev d expr |&gt; apply_seq d<\/code><\/p>\n<p>After that evaluation is done, <code>apply_seq<\/code> is called. <code>tev<\/code> by itself only returns the last expression in the scope. After the call to <code>tev<\/code> the rest of the expressions can be found in <code>d.seq<\/code> where they have been let inserted.<\/p>\n<pre>let inline tev_seq (d: LangEnv) (expr: TypedExpr): TypedExpr = let d = {d with seq=ref id; cse_env=ref !d.cse_env} in tev d expr |&gt; apply_seq d\n<\/pre>\n<p>Appending the last expression to it is quite easy. The last expression just need to be applied to <code>d.seq<\/code> and the result will be all the statements in scope neatly arranged. If the <code>d.seq<\/code> is empty, then it would be the identity function and so applying an expression to it would return the same thing; there wouldn't be any empty statements. This is why <code>(TypedExpr -&gt; TypedExpr) ref<\/code> representation is so convenient.<\/p>\n<pre>| TyType (PrimT BoolT) as cond -&gt;\n    let b x = cse_add' d cond (TyLit &lt;| LitBool x)\n    let tr = tev_assume (b true) d tr\n    let fl = tev_assume (b false) d fl\n    let type_tr, type_fl = get_type tr, get_type fl\n    if type_tr = type_fl then\n        match tr, fl with\n        | TyLit (LitBool true), TyLit (LitBool false) -&gt; cond\n        | _ when tr = fl -&gt; tr\n        | _ -&gt; TyOp(IfStatic,[cond;tr;fl],type_tr) |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n    else on_type_er (trace d) &lt;| sprintf \"Types in branches of If do not match.\\nGot: %s and %s\" (show_ty type_tr) (show_ty type_fl)\n| TyType cond -&gt; on_type_er (trace d) &lt;| sprintf \"Expected a bool in conditional.\\nGot: %s\" (show_ty cond)\n<\/pre>\n<p>Hopefully what the first 4 lines are doing now should now be clear.<\/p>\n<pre>let type_tr, type_fl = get_type tr, get_type fl\n<\/pre>\n<p><code>get_type<\/code> is important so it will get its own treatment later, but its essence is in entirety described by its type which is <code>TypedExpr -&gt; Ty<\/code>. That is, it gets the type of the expression.<\/p>\n<pre>if type_tr = type_fl then\n<\/pre>\n<p>Since the conditional is dynamic, the branches need to be compared for equality. Then a tad optimization is done after the check.<\/p>\n<pre>match tr, fl with\n| TyLit (LitBool true), TyLit (LitBool false) -&gt; cond\n| _ when tr = fl -&gt; tr\n| _ -&gt; TyOp(IfStatic,[cond;tr;fl],type_tr) |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n<\/pre>\n<p>The first case does trigger sometimes such as in structural equality for example. The second case just checks if both branches are equal. In that case obviously the if statement is unnecessary.<\/p>\n<p>In the third branch is where the if statement gets let inserted. Note how <code>make_tyv_and_push_typed_expr_even_if_unit<\/code> needs to be called here. If that did not happen and the return was let to be grabbed by <code>destructure<\/code> it would not get printed if it had unit return and would be added to the CSE dictionary there.<\/p>\n<p>That is not the desired behavior.<\/p>\n<p>It is the desired behavior when destructuring unit types from a tuple, or with pure built-in arithmetic expressions, but not with if statements.<\/p>\n<pre>on_type_er (trace d) &lt;| sprintf \"Types in branches of If do not match.\\nGot: %s and %s\" (show_ty type_tr) (show_ty type_fl)\n<\/pre>\n<p>If the types of the branches do not match the above will raise a type error. Here is how it is implemented<\/p>\n<pre>let on_type_er trace message = TypeError(trace,message) |&gt; raise\n<\/pre>\n<p>The exception just takes the trace and the messages and then gets raised.<\/p>\n<h4><a id=\"user-content-raising-type-errors\" class=\"anchor\" aria-hidden=\"true\" href=\"#raising-type-errors\"><\/a>Raising Type Errors<\/h4>\n<p>It is possible to raise a type error directly in Spiral.<\/p>\n<pre>if c then\n    ...\nelse\n    error_type \"Some error.\"\n<\/pre>\n<p>If the conditional <code>c<\/code> is dynamic then the type error will always get triggered so this would not be very useful, but plenty of times functions of such form will get used with static conditionals. Plenty of times has <code>assert<\/code> been used in the tutorials.<\/p>\n<p>Here is how the <code>assert<\/code> is implemented in the core library. There is a slightly more sophisticated version in <code>Extern<\/code> calls <code>show<\/code> as well.<\/p>\n<pre>inl assert c msg = \n    inl raise = \n        if lit_is c then error_type\n        else failwith unit\n    \n    if c = false then raise msg\n<\/pre>\n<p>Here is how it is implemented in the partial evaluator.<\/p>\n<pre>| ErrorType,[a] -&gt; tev d a |&gt; fun a -&gt; on_type_er (trace d) &lt;| sprintf \"%s\" (show_typedexpr a)\n<\/pre>\n<p>So it is a simple <code>tev<\/code> and then the <code>on_type_er<\/code> gets called on the result of that. It is not particularly sophisticated, but then it does not have to be.<\/p>\n<h3><a id=\"user-content-4-boxing-of-union-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#4-boxing-of-union-types\"><\/a>4: Boxing of Union Types<\/h3>\n<p>Some languages can afford to be minimalist like Lisp, but Spiral is not that sort of language. It depends on a large number of first class features working seamlessly together. Without union types and term casting it would be harder to ensure convergence and the language would be poorer as a result. The lack of such features in a language with first class staging like Spiral would immediately become obvious and the need for them would arise on its own.<\/p>\n<p>The way boxing works is simple. If the argument being boxed is a subset of the type, then it gets wrapped in a <code>TyBox<\/code> otherwise it is a type error.<\/p>\n<p>Any complexity in the following function is just from needing to handle the various edge cases.<\/p>\n<pre>let type_box (d: LangEnv) (typec: Expr) (args: Expr): TypedExpr =\n    let typec &amp; TyType ty, args = tev2 d typec args\n    let substitute_ty = function\n        | TyBox(x,_) -&gt; tybox(x,ty)\n        | x -&gt; tybox(x,ty)\n    let (|TyRecUnion|_|) = function\n        | UnionT ty' -&gt; Some ty'\n        | RecT key -&gt; Some (set_field (rect_unbox d key))\n        | _ -&gt; None\n    \n    match ty, get_type args with\n    | x, r when x = r -&gt; args\n    | TyRecUnion ty', UnionT ty_args when Set.isSubset ty_args ty' -&gt;\n        let lam = inl' [\"typec\"; \"args\"] (op(Case,[v \"args\"; type_box (v \"typec\") (v \"args\")])) |&gt; inner_compile\n        apply d (apply d lam typec) args\n    | TyRecUnion ty', x when Set.contains x ty' -&gt; substitute_ty args\n    | _ -&gt; on_type_er (trace d) &lt;| sprintf \"Type constructor application failed. %s does not intersect %s.\" (show_ty ty) (get_type args |&gt; show_ty)\n<\/pre>\n<p>To start things off, here is how <code>tev2<\/code> is implemented.<\/p>\n<pre>let inline tev2 d a b = tev d a, tev d b\nlet inline tev3 d a b c = tev d a, tev d b, tev d c\nlet inline tev4 d a b c d' = tev d a, tev d b, tev d c, tev d d'\n<\/pre>\n<p>Not a big deal.<\/p>\n<pre>let typec &amp; TyType ty, args = tev2 d typec args\nlet substitute_ty = function\n    | TyBox(x,_) -&gt; tybox(x,ty)\n    | x -&gt; tybox(x,ty)\n<\/pre>\n<p><code>substitute_ty<\/code> ty is fairly straightforward. The reason why it exists is because the argument being boxed might already by a <code>TyBox<\/code> in which case the box is just replaced.<\/p>\n<pre>let (|TyRecUnion|_|) = function\n    | UnionT ty' -&gt; Some ty'\n    | RecT key -&gt; Some (set_field (rect_unbox d key))\n    | _ -&gt; None\n<\/pre>\n<p>Since the type being boxed to must be a union type, the above active pattern is what is responsible for that. If the type being boxed is a straightforward union type, then it just returns it. But if the type being boxed is a recursive type then it is unrolled a single level and treated like a union type.<\/p>\n<pre>match ty, get_type args with\n| x, r when x = r -&gt; args\n| TyRecUnion ty', UnionT ty_args when Set.isSubset ty_args ty' -&gt;\n    let lam = inl' [\"typec\"; \"args\"] (op(Case,[v \"args\"; type_box (v \"typec\") (v \"args\")])) |&gt; inner_compile\n    apply d (apply d lam typec) args\n| TyRecUnion ty', x when Set.contains x ty' -&gt; substitute_ty args\n| _ -&gt; on_type_er (trace d) &lt;| sprintf \"Type constructor application failed. %s does not intersect %s.\" (show_ty ty) (get_type args |&gt; show_ty)\n<\/pre>\n<p>The first case is straightforward. It is true that types can only be boxed into union types, but there is no need to throw an error on <code>box int64 1<\/code> for example. In that case returning the original should be fine.<\/p>\n<p>The second case is a bit less straightforward and deals with the case when a union type is being boxed with a larger union type as the target. Note the call to <code>Set.isSubset<\/code>. The description for it says <code>Evaluated to 'true' if all the elements of the first set are in the second.<\/code> What will happen in this case is that all the elements will get unboxed and then boxed into the new type.<\/p>\n<p>The third case is straightforward and just does the substitution.<\/p>\n<p>The fourth case is straightforward and just raises a type error.<\/p>\n<h3><a id=\"user-content-5-unboxing-of-union-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#5-unboxing-of-union-types\"><\/a>5: Unboxing of Union Types<\/h3>\n<p>Originally CSE was added to Spiral in order to propagate unboxing information. In the following function that handles the entirety of unboxing, it will be shown how that works.<\/p>\n<pre>        let case_ d v case =\n            let inline assume d v x branch = tev_assume (cse_add' d v x) d branch\n            match tev d v with\n            | a &amp; TyBox(b,_) -&gt; \n                cse_add d a b\n                let r = tev d case\n                cse_remove d a\n                r\n            | (TyV(_, t &amp; (UnionT _ | RecT _)) | TyT(t &amp; (UnionT _ | RecT _))) as v -&gt;\n                let make_up_vars_for_ty (l: Ty list): TypedExpr list = List.map (make_up_vars_for_ty d) l\n                let map_to_cases (l: TypedExpr list): (TypedExpr * TypedExpr) list = List.map (fun x -&gt; x, assume d v x case) l\n                            \n                match case_type d t |&gt; make_up_vars_for_ty |&gt; map_to_cases with\n                | (_, TyType p) :: cases as cases' -&gt; \n                    if List.forall (fun (_, TyType x) -&gt; x = p) cases then \n                        TyOp(Case,v :: List.collect (fun (a,b) -&gt; [a;b]) cases', p) \n                        |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n                    else \n                        let l = List.map (snd &gt;&gt; get_type) cases'\n                        on_type_er (trace d) &lt;| sprintf \"All the cases in pattern matching clause with dynamic data must have the same type.\\nGot: %s\" (listt l |&gt; show_ty)\n                | _ -&gt; failwith \"There should always be at least one clause here.\"\n            | _ -&gt; tev d case\n<\/pre>\n<p>Near the very top there is <code>tev_assume<\/code> that was also used in <code>if_static<\/code>.<\/p>\n<pre>        match tev d v with\n        | a &amp; TyBox(b,_) -&gt; \n            d.cse_env := cse_add' d a b\n            let r = tev d case\n            d.cse_env := cse_remove' d a\n            r\n<\/pre>\n<p>The first case is straightforward. If the type is a <code>TyBox _<\/code>, that is a staged boxed type instead of being fully boxed then the box is thrown away. <code>cse_env<\/code> is used to rewrite the variable <code>a<\/code> to what is inside it <code>b<\/code> as it goes to evaluate case. When if finishes evaluation the rewrite is removed and the result is returned.<\/p>\n<pre>        let rec case_type d args_ty =\n            let union_case = function\n                | UnionT l -&gt; Set.toList l\n                | _ -&gt; [args_ty]\n            match args_ty with\n            | RecT key -&gt; union_case (rect_unbox d key)\n            | x -&gt; union_case x\n...\n            | (TyV(_, t &amp; (UnionT _ | RecT _)) | TyT(t &amp; (UnionT _ | RecT _))) as v -&gt;\n                let make_up_vars_for_ty (l: Ty list): TypedExpr list = List.map (make_up_vars_for_ty d) l\n                let map_to_cases (l: TypedExpr list): (TypedExpr * TypedExpr) list = List.map (fun x -&gt; x, assume d v x case) l\n                            \n                match case_type d t |&gt; make_up_vars_for_ty |&gt; map_to_cases with\n<\/pre>\n<p>The 4 lines here make the meat of the function. First the boxed type is taken apart. If it is a union type then its member are returned in a list. If it is a recursive type then it is unrolled a level into a union type and then its members are returned into a list.<\/p>\n<p>Here is an example in pseudo-code.<\/p>\n<pre>union {.Some, int64 | .None} -&gt; [.Some, int64; .None]\n<\/pre>\n<p>Since raw types can't be in the case expression what is needed is to assign them variables. <code>make_up_vars_for_ty<\/code> is what does this.<\/p>\n<p>Here is an example in pseudo-code.<\/p>\n<pre>.Some, int64 -&gt; .Some, var int64\n.None -&gt; .None\n<\/pre>\n<p>Then those two members are mapped to their case bodies in <code>map_to_cases<\/code>.<\/p>\n<p>This part might be a bit confusing. There aren't any case bodies strictly speaking; there is only one <code>case<\/code>. But that one <code>case<\/code> is in fact enough to cover all the pattern matching needs due to the magic of intensional polymorphism and first class staging. It will be shown later how the pattern compiler works and hopefully it will be clear why this works after that. As a short summary, the patterns are all CPS'd and the static ifs suffice to cover the entirety of the pattern matching needs.<\/p>\n<pre>                | (_, TyType p) :: cases as cases' -&gt; \n                    if List.forall (fun (_, TyType x) -&gt; x = p) cases then \n                        TyOp(Case,v :: List.collect (fun (a,b) -&gt; [a;b]) cases', p) \n                        |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n                    else \n                        let l = List.map (snd &gt;&gt; get_type) cases'\n                        on_type_er (trace d) &lt;| sprintf \"All the cases in pattern matching clause with dynamic data must have the same type.\\nGot: %s\" (listt l |&gt; show_ty)\n                | _ -&gt; failwith \"There should always be at least one clause here.\"\n<\/pre>\n<p>By this point, all the important work has been done and the rest is error checking and let insertion for the code generator in the next phase of compilation.<\/p>\n<pre>            | _ -&gt; tev d case\n<\/pre>\n<p>If the argument is not a caseable (union or recursive) type then the unboxing is skipped. The partial evaluation just proceeds as normal.<\/p>\n<p>This is the entirety of unboxing.<\/p>\n<pre>let inline assume d v x branch = tev_assume (cse_add' d v x) d branch\n<\/pre>\n<p>As a recap, consider how in the tutorial chapters it was warned against using join points in tandem with partial active patterns and union types. By now it should be clear why this is so. When a join point is entered, the rewriting information in <code>cse_env<\/code> is thrown away. Since <code>cse_env<\/code> is used to propagate unboxing information as shown in the above function, that is where the difficulty arises.<\/p>\n<p>As it stands, unboxing in Spiral is fairly primitive. ML compilers go a lot further in optimizing pattern matching expressions than Spiral and in its current form, Spiral is not really suitable for writing runtime interpreters as its unboxing facilities are not developed enough.<\/p>\n<p>For example, it has structural equality, but...<\/p>\n<pre>inl ty = .A \\\/ .B \\\/ .C \\\/ .E\ninl a = box ty .A |&gt; dyn\ninl b = box ty .B |&gt; dyn\na = b\n<\/pre>\n<pre>type Union0 =\n    | Union0Case0\n    | Union0Case1\n    | Union0Case2\n    | Union0Case3\nlet rec method_0((var_0: Union0), (var_1: Union0)): bool =\n    match var_0 with\n    | Union0Case0 -&gt;\n        match var_1 with\n        | Union0Case0 -&gt;\n            true\n        | Union0Case1 -&gt;\n            false\n        | Union0Case2 -&gt;\n            false\n        | Union0Case3 -&gt;\n            false\n    | Union0Case1 -&gt;\n        match var_1 with\n        | Union0Case0 -&gt;\n            false\n        | Union0Case1 -&gt;\n            true\n        | Union0Case2 -&gt;\n            false\n        | Union0Case3 -&gt;\n            false\n    | Union0Case2 -&gt;\n        match var_1 with\n        | Union0Case0 -&gt;\n            false\n        | Union0Case1 -&gt;\n            false\n        | Union0Case2 -&gt;\n            true\n        | Union0Case3 -&gt;\n            false\n    | Union0Case3 -&gt;\n        match var_1 with\n        | Union0Case0 -&gt;\n            false\n        | Union0Case1 -&gt;\n            false\n        | Union0Case2 -&gt;\n            false\n        | Union0Case3 -&gt;\n            true\nlet (var_0: Union0) = Union0Case0\nlet (var_1: Union0) = Union0Case1\nmethod_0((var_0: Union0), (var_1: Union0))\n<\/pre>\n<p>This is n^2 compilation for structural equality for fully boxed types. It could be dealt with by introducing a different kind of case, but it is not a priority by any means at the moment.<\/p>\n<p>For <code>Result<\/code> and <code>Option<\/code> types, what Spiral has currently has is sufficient. The rest can be done by propagating types at compile time.<\/p>\n<h3><a id=\"user-content-6-join-points\" class=\"anchor\" aria-hidden=\"true\" href=\"#6-join-points\"><\/a>6: Join Points<\/h3>\n<p>Along with <code>inl<\/code>ineables, join points are Spiral's most iconic feature and what make it the crystallization of the staged functional programming style.<\/p>\n<p>Currently there are 4 kinds of join points in total and this section will cover the vanilla one. The rest are quite similar to it though. The join point for term casting cannot be placed directly, instead <code>term_cast<\/code> does it.<\/p>\n<p>Join points have a long history in Spiral. When work first started on Spiral the language originally had only inlineables, but some way of handling recursion had been necessary. So then the second thing that was added were <code>met<\/code>hods. They weren't join points at first, but a different kind of function that memoizes its result using a global dictionary.<\/p>\n<p>There wasn't a single great point where methods suddenly evolved into join points. It took the author a considerable amount of refactoring work in order to separate the concept of functions and where they join at the conceptual level. He had never seen such a feature in any other language up to that point so it was quite hard for him to internalize them and how they fit into the language.<\/p>\n<p>Regardless, join points are such an integral part of the whole staged functional programming experiences that language with staging, but no join points can be considered as simply toys and disregarded based on that fact.<\/p>\n<p>There are 2 aspects of join points that need to be kept in mind in order to understand them.<\/p>\n<ol><li>Unused variable elimination.<\/li>\n<\/ol><pre>inl a _ = ...\ninl b _ = ...a... \/\/ `a` should be in `b`'s lexical environment when the function is called\ninl c _ = ...a...b... \/\/ `a` and `b` should be in b's lexical environment when the function is called\n...\n<\/pre>\n<pre>inl a _ = ...\ninl b _ = ... \/\/ empty env\ninl c _ = ... \/\/ empty env\n...\n<\/pre>\n<p>Spiral always does unused variable filtering at function boundaries and that is necessary to guarantee predicable convergence for join points.<\/p>\n<p>If the environment always had everything it passed inside the join point, then those sucked in unused arguments would get specialized to as well which would not be good. This kind of optimization would not be necessary in dynamic languages, but is necessary in Spiral.<\/p>\n<p>Strictly speaking, Spiral does not do immediate filtration. It tends to keep a list of unused variables and delay the filtration until the last possible moment, but the semantics of immediate and lazy unused argument filtration are intended to be the same. It is just intended to be a performance optimization.<\/p>\n<ol start=\"2\"><li>Renaming.<\/li>\n<\/ol><p>If the filtered arguments into join points are something like <code>{a=var_3; b=var_6; c=var_10}<\/code> they should be renamed to <code>{a=var_0; b=var_1; c=var_2}<\/code> before the expression can be memoized. This is simple enough to do and took the author around 2.5 month in order to understand properly and that is despite thinking about it the whole time.<\/p>\n<p>In practice, Spiral's renamer also collects the join point arguments along the way so the traversal is not repeated. It also collects the renamed arguments as well.<\/p>\n<pre>    let rec renamer_apply_env' (r: EnvRenamer) (C x) = Map.map (fun k -&gt; renamer_apply_typedexpr' r) x\n    and renamer_apply_typedexpr' ({memo=memo; renamer=renamer; ref_call_args=call_args; ref_method_pars=method_pars} as r) e =\n        let inline f e = renamer_apply_typedexpr' r e\n        let inline rename (n,t as k) =\n            match renamer.TryGetValue n with\n            | true, v -&gt; v,t\n            | false, _ -&gt;\n                let n' = renamer.Count\n                renamer.Add(n,n')\n                call_args := k :: !call_args \n                let k' = n', t\n                method_pars := k' :: !method_pars\n                k'\n        match memo.TryGetValue e with\n        | true, e -&gt; e\n        | false, _ -&gt;\n            match e with\n            | TyT _ -&gt; e\n            | TyBox (n,t) -&gt; tybox(f n,t)\n            | TyList l -&gt; tyvv(List.map f l)\n            | TyMap(l,t) -&gt; tymap(renamer_apply_env' r l |&gt; consify_env_term, t)\n            | TyV (n,t as k) -&gt;\n                let n', _ as k' = rename k\n                if n' = n then e else tyv k'\n            | TyLit _ -&gt; e\n            | TyJoinPoint _ | TyOp _ | TyState _ | TyLet _ -&gt; failwith \"Only data structures in the env can be renamed.\"\n            |&gt; fun x -&gt; memo.[e] &lt;- x; x\n<\/pre>\n<p>The above is in fact a map over <code>TypedExpr<\/code> that just does memoization using the <code>memo<\/code>. Memoization inside the renamer is an important optimization when doing renaming since the standard and core library function will repeat themselves significantly in nested functions.<\/p>\n<p>There isn't too much need to dwell on the above function. It is merely big, but not particularly complicated and does what it is name says.<\/p>\n<p>With that out of the way, here is the actual join point.<\/p>\n<pre>let join_point_method (d: LangEnv) (expr: Expr): TypedExpr = \n    let {call_args=call_arguments; method_pars=method_parameters; renamer'=renamer}, renamed_env = renamer_apply_env d.env\n    let length = renamer.Count\n    let join_point_key: Node&lt;Expr * EnvTerm&gt; = nodify_memo_key (expr, renamed_env)\n    \n    let ret_ty = \n        let d = {d with env=renamed_env; ltag=ref length}\n        let join_point_dict: Dictionary&lt;_,_&gt; = join_point_dict_method\n        match join_point_dict.TryGetValue join_point_key with\n        | false, _ -&gt;\n            join_point_dict.[join_point_key] &lt;- JoinPointInEvaluation ()\n            let typed_expr = tev_method d expr\n            join_point_dict.[join_point_key] &lt;- JoinPointDone (method_parameters, typed_expr)\n            typed_expr\n        | true, JoinPointInEvaluation _ -&gt; \n            tev_rec d expr\n        | true, JoinPointDone (used_vars, typed_expr) -&gt; \n            typed_expr\n        |&gt; get_type\n    ty_join_point join_point_key JoinPointMethod call_arguments ret_ty \n    |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n<\/pre>\n<p>The <code>let join_point_key: Node&lt;Expr * EnvTerm&gt; = nodify_memo_key (expr, renamed_env)<\/code> line is a performance optimization so dictionary keys can be compared quickly. The actual key is <code>Expr * EnvTerm<\/code>, but a wrapper is put around it that ensures fast comparison.<\/p>\n<p>Here is the <code>Node<\/code> type.<\/p>\n<pre>type Node&lt;'a&gt;(expr:'a, symbol:int) = \n    member x.Expression = expr\n    member x.Symbol = symbol\n    override x.ToString() = sprintf \"&lt;tag %i&gt;\" symbol\n    override x.GetHashCode() = symbol\n    override x.Equals(y) = \n        match y with \n        | :? Node&lt;'a&gt; as y -&gt; symbol = y.Symbol\n        | _ -&gt; failwith \"Invalid equality for Node.\"\n    interface IComparable with\n        member x.CompareTo(y) = \n            match y with\n            | :? Node&lt;'a&gt; as y -&gt; compare symbol y.Symbol\n            | _ -&gt; failwith \"Invalid comparison for Node.\"\n<\/pre>\n<p>What <code>Node<\/code> does is allows two arbitrary expressions to be assigned an unique <code>symbol<\/code> and be compared using that.<\/p>\n<p>After renaming and making the <code>join_point_key<\/code>, the body of the join point is what comes.<\/p>\n<pre>let ret_ty = \n    let d = {d with env=renamed_env; ltag=ref length}\n<\/pre>\n<p>Now the environment is assigned the renamed environment and the tag reference which is incremented every time a new variable is made is set to one past the highest argument in scope.<\/p>\n<pre>let join_point_dict: Dictionary&lt;_,_&gt; = join_point_dict_method\nmatch join_point_dict.TryGetValue join_point_key with\n| false, _ -&gt;\n    join_point_dict.[join_point_key] &lt;- JoinPointInEvaluation ()\n    let typed_expr = tev_method d expr\n    join_point_dict.[join_point_key] &lt;- JoinPointDone (method_parameters, typed_expr)\n    typed_expr\n<\/pre>\n<p>This is the heart of the join point. In the first case which is hit when the join point is entered the first time, it sets the dictionary to indicate that the join point is being evaluated in case it is ever called recursively.<\/p>\n<p>Then it calls <code>tev_method<\/code>. That particular method is similar to <code>tev_seq<\/code>.<\/p>\n<pre>let inline tev_seq d expr = let d = {d with seq=ref id; cse_env=ref !d.cse_env} in tev d expr |&gt; apply_seq d\nlet inline tev_assume cse_env d expr = let d = {d with seq=ref id; cse_env=ref cse_env} in tev d expr |&gt; apply_seq d\nlet inline tev_method d expr = let d = {d with seq=ref id; cse_env=ref Map.empty} in tev d expr |&gt; apply_seq d\nlet inline tev_rec d expr = let d = {d with seq=ref id; cse_env=ref Map.empty; rbeh=AnnotationReturn} in tev d expr |&gt; apply_seq d\n<\/pre>\n<p>As can be seen when entering a new scope, all 4 functions set the <code>seq<\/code> to identity. <code>tev_seq<\/code> and <code>tev_assume<\/code> when entering a new scope make a copy of the <code>cse_env<\/code>. <code>tev_method<\/code> and <code>tev_rec<\/code> which are join point functions outright set it to empty.<\/p>\n<p><code>tev_method<\/code> finishes running and returns the body of the join point, the join point is set to finished and the <code>typed_expr<\/code> is returned from the branch.<\/p>\n<pre>| true, JoinPointInEvaluation _ -&gt; \n    tev_rec d expr\n<\/pre>\n<p>This particular case can only ever happen during recursion. If it hits it means the function has been called recursively. In order for it to not diverge it needs to get the return type.<\/p>\n<p>The <code>tev_rec<\/code> is exactly like <code>tev_method<\/code> except for setting the return behavior to <code>rbeh=AnnotationReturn<\/code>.<\/p>\n<p>There is only a single place where <code>rbeh<\/code> ever comes into play in the language. In the type annotation function.<\/p>\n<pre>let type_annot d a b =\n    match d.rbeh with\n    | AnnotationReturn -&gt; tev_annot {d with rbeh=AnnotationDive} b\n    | AnnotationDive -&gt;\n        let a, b = tev d a, tev_annot d b\n        let ta, tb = get_type a, get_type b\n        if ta = tb then a else on_type_er (trace d) &lt;| sprintf \"Type annotation mismatch.\\n%s &lt;&gt; %s\" (show_ty ta) (show_ty tb)\n<\/pre>\n<p>When <code>rbeh<\/code> is set to annotation return then it won't evaluate the body, but will just evaluate the right side of the annotation instead.<\/p>\n<pre>let inline tev_seq d expr = let d = {d with seq=ref id; cse_env=ref !d.cse_env} in tev d expr |&gt; apply_seq d\nlet inline tev_annot d expr = let d = {d with seq=ref id; cse_env=ref !d.cse_env} in tev d expr\n<\/pre>\n<p><code>tev_annot<\/code> is identical to <code>tev_seq<\/code> except it throws away the intermediate statements and only returns the final expression.<\/p>\n<pre>        | true, JoinPointDone (used_vars, typed_expr) -&gt; \n            typed_expr\n<\/pre>\n<p>The last case does no evaluation and just returns the body.<\/p>\n<pre>        |&gt; get_type\n    ty_join_point join_point_key JoinPointMethod call_arguments ret_ty \n    |&gt; make_tyv_and_push_typed_expr_even_if_unit d\n<\/pre>\n<p>This is the final stretch. The returned body gets converted into the return type <code>ret_ty<\/code> and then a join point is made and let inserted.<\/p>\n<pre>let ty_join_point key jp_type args ret_type = TyJoinPoint(key,jp_type,args,ret_type)\n<\/pre>\n<p>It might be more readable to make the join point directly, but a lot of the functions in the Spiral compiler that essentially act as the identity might have been been doing memoization in the past, so the author prefers to leave them alone in case inspiration hits him in order to make redesign easier. At any rate, the final part of the join point is extremely straightforward.<\/p>\n<p>Before this chapter is done, a note is needed regarding the join point dictionaries.<\/p>\n<pre>\/\/ #Main\nlet spiral_peval (settings: CompilerSettings) (Module(N(module_name,_,_,_)) as module_main) = \n    let join_point_dict_method = d0()\n    let join_point_dict_closure = d0()\n    let join_point_dict_type = d0()\n    let join_point_dict_cuda = d0()\n<\/pre>\n<p>They are initialized at the very start of compilation and hash via structural equality. Because passing state would be annoying otherwise and since they are only ever added to, they are present during the entire compilation. Parsing and prepass does not touch them, but the code generator reads from them.<\/p>\n<p>Apart from the types and the core library, the entire compiler is a part of a single 3200 line function. The author didn't like stepping into that at first, but once he did so he found that the arrangement works surprisingly well and has had no trouble with it since then.<\/p>\n<h3><a id=\"user-content-7-the-prepass-and-unused-variable-filtering\" class=\"anchor\" aria-hidden=\"true\" href=\"#7-the-prepass-and-unused-variable-filtering\"><\/a>7: The Prepass and Unused Variable Filtering<\/h3>\n<p>The prepass works before partial evaluation, but after parsing. Its goal is to convert <code>Function<\/code>s to <code>FunctionFilt<\/code>s which have the set of used variables in their bodies. It also calls <code>pattern_compile<\/code> in order to compile the <code>Pattern<\/code>s into <code>Expr<\/code>s.<\/p>\n<pre>and Expr = \n    | V of Node&lt;string&gt;\n    | Lit of Node&lt;Value&gt;\n    | Pattern of Node&lt;Pattern&gt;\n    | Function of Node&lt;FunctionCore&gt;\n    | FunctionFilt of Node&lt;Set&lt;string&gt; * Node&lt;FunctionCore&gt;&gt;\n    | VV of Node&lt;Expr list&gt;\n    | Op of Node&lt;Op * Expr list&gt;\n    | ExprPos of Pos&lt;Expr&gt;\n<\/pre>\n<p>The way it works is rather simple. Prepass is a rightwards fold + map over <code>Expr<\/code>. It also does memoization.<\/p>\n<pre>    and expr_prepass e =\n        e |&gt; memoize prepass_dict (fun e -&gt;\n            let inline f e = expr_prepass e\n            match e with\n            | V (N n) -&gt; Set.singleton n, e\n            | Op(N(op',l)) -&gt;\n                let l,l' = List.map f l |&gt; List.unzip\n                Set.unionMany l, op(op',l')\n            | VV (N l) -&gt; \n                let l,l' = List.map f l |&gt; List.unzip\n                Set.unionMany l, vv l'\n            | FunctionFilt(N (vars,N(name,body))) -&gt;\n                Set.remove name vars, e\n            | Function(N(name,body)) -&gt;\n                let vars,body = f body\n                Set.remove name vars, func_filt(vars,nodify_func(name,body))\n            | Lit _ -&gt; Set.empty, e\n            | Pattern pat -&gt; pattern_compile pat\n            | ExprPos p -&gt; \n                let vars, body = f p.Expression\n                vars, expr_pos p.Pos body\n            )\n<\/pre>\n<p>In a nutshell, if it find a <code>V _<\/code> then it adds it to the set. If it finds a function it removes variable with the name of its binding from the set. It does it top to bottom so the thing works.<\/p>\n<p>With this in mind, now it can be revealed what <code>EnvTerm<\/code> is.<\/p>\n<pre>and EnvTerm = \n| EnvConsed of ConsedNode&lt;Map&lt;string, TypedExpr&gt;&gt;\n| Env of Map&lt;string, TypedExpr&gt;\n| EnvUnfiltered of Map&lt;string, TypedExpr&gt; * Set&lt;string&gt;\n<\/pre>\n<p>In the previous versions of Spiral, the unused arguments were filtered every time <code>FunctionFilt<\/code> was reached, but the author decided he did not want the entire environment to be iterated over every time so he decided to add some laziness to the mix.<\/p>\n<p>Every time <code>EnvTerm<\/code> is accessed it is either converted to <code>Env<\/code> if it is unfiltered or a check is made first against the used variable set. <code>EnvConsed<\/code> is similar to <code>Env<\/code>. The environment is <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hash_consing\" rel=\"nofollow\">hash consed<\/a> in the renamer as a performance optimization. To see how that works check out <code>SpiralHashConsing.fs<\/code>. For the understanding of language semantics understanding hash consing is not particularly important. It might be worth reevaluating whether hash consing is worth having since the evaluator has went through several iteration since that was useful. It might turn out that reference equality in the renamer is all that is needed.<\/p>\n<pre>let c = function\n| Env env -&gt; env\n| EnvUnfiltered (env,used_vars) -&gt; Map.filter (fun k _ -&gt; used_vars.Contains k) env\n| EnvConsed env -&gt; env.node\nlet (|C|) x = c x\n<\/pre>\n<p>Whenever <code>C env<\/code> is seen as a part of the pattern, this is what happens. If it is unfiltered, then it gets filtered otherwise it is just extracted.<\/p>\n<p>This might seem complicated, but the type system takes care that no mistakes are made when dealing with environments. It was really trivial to move from immediate to lazy filtering.<\/p>\n<h3><a id=\"user-content-8-pattern-compilation\" class=\"anchor\" aria-hidden=\"true\" href=\"#8-pattern-compilation\"><\/a>8: Pattern Compilation<\/h3>\n<pre>let pattern_dict = d0()\nlet rec pattern_compile (pat: Node&lt;_&gt;) = \n    pat |&gt; memoize pattern_dict (fun pat -&gt;\n        let node = pat.Symbol\n        let pat = pat.Expression\n        let new_pat_var =\n            let mutable i = 0\n            let get_pattern_tag () = \n                let x = i\n                i &lt;- i + 1\n                x\n            fun () -&gt; sprintf \" pat_var_%i_%i\" node (get_pattern_tag())\n        let rec pattern_compile (arg: Expr) (pat: Pattern) (on_succ: Expr) (on_fail: Expr): Expr =\n            let inline cp arg pat on_succ on_fail = pattern_compile arg pat on_succ on_fail\n            let pat_tuple_helper l =\n                List.foldBack (fun pat (c,s,on_succ) -&gt; \n                    let arg = new_pat_var()\n                    c + 1, arg :: s,cp (v arg) pat on_succ on_fail) l (0,[],on_succ)\n        \n            match pat with\n            | PatClauses l -&gt; List.foldBack (fun (pat, exp) on_fail -&gt; cp arg pat exp on_fail) l on_fail\n            | PatE -&gt; on_succ\n            | PatVar x -&gt; l x arg on_succ\n            | PatTypeEq (exp,typ) -&gt;\n                let on_succ = cp arg exp on_succ on_fail\n                if_static (eq_type arg typ) on_succ on_fail\n                |&gt; case arg\n            | PatTuple l -&gt; \n                let count, args, on_succ = pat_tuple_helper l\n                list_taken_cps count arg on_fail (inl' args on_succ) |&gt; case arg\n            | PatCons l -&gt; \n                let count, args, on_succ = pat_tuple_helper l\n                list_taken_tail_cps count arg on_fail (inl' args on_succ) |&gt; case arg\n            | PatActive (a,b) -&gt;\n                let pat_var = new_pat_var()\n                l pat_var (ap (v a) arg) (cp (v pat_var) b on_succ on_fail)\n            | PatPartActive (a,pat) -&gt; \n                let pat_var = new_pat_var()\n                let on_succ = inl pat_var (cp (v pat_var) pat on_succ on_fail)\n                let on_fail = inl \"\" on_fail\n                ap' (v a) [arg; on_fail; on_succ]\n            | PatOr l -&gt; List.foldBack (fun pat on_fail -&gt; cp arg pat on_succ on_fail) l on_fail\n            | PatAnd l -&gt; List.foldBack (fun pat on_succ -&gt; cp arg pat on_succ on_fail) l on_succ\n            | PatNot p -&gt; cp arg p on_fail on_succ \/\/ switches the on_fail and on_succ arguments\n            | PatXor l -&gt;\n                let state_var = new_pat_var()\n                let state_var' = v state_var\n                let bool x = lit &lt;| LitBool x\n                let rec just_one = function\n                    | x :: xs -&gt; \n                        let xs = just_one xs\n                        inl state_var \n                            (cp arg x \n                                (if_static state_var' on_fail (ap xs (bool true))) \/\/ true case\n                                (ap xs state_var')) \/\/ false case\n                    | [] -&gt; inl state_var (if_static state_var' on_succ on_fail)\n                ap (just_one l) (bool false)\n            | PatLit x -&gt; \n                let x = lit x\n                let on_succ = if_static (eq arg x) on_succ on_fail\n                if_static (eq_type arg x) on_succ on_fail |&gt; case arg\n            | PatTypeLit x -&gt; \n                if_static (eq_type arg (type_lit_lift x)) on_succ on_fail \n                |&gt; case arg\n            | PatTypeLitBind x -&gt; \n                if_static (type_lit_is arg) (l x (type_lit_cast arg) on_succ) on_fail \n                |&gt; case arg\n            | PatWhen (p, e) -&gt; cp arg p (if_static e on_succ on_fail) on_fail\n            | PatModuleIs p -&gt; module_is_cps arg on_fail (cp arg p on_succ on_fail) |&gt; case arg\n            | PatModuleMember name -&gt; module_member_cps arg name on_fail (inl name on_succ) |&gt; case arg\n            | PatModuleRebind(name,b) -&gt; \n                let arg' = new_pat_var()    \n                module_member_cps arg name on_fail (inl arg' (cp (v arg') b on_succ on_fail)) \n                |&gt; case arg\n            | PatPos p -&gt; expr_pos p.Pos (cp arg p.Expression on_succ on_fail)\n            | PatTypeTermFunction(a,b) -&gt; \n                let va, vb = new_pat_var(), new_pat_var()\n                term_fun_dom_range_cps arg on_fail \n                &lt;| inl' [va; vb] (cp (v va) a (cp (v vb) b on_succ on_fail) on_fail)\n                \n        let main_arg = new_pat_var()\n        let arg = v main_arg\n                \n        let pattern_compile_def_on_succ = op(ErrorPatClause,[])\n        let pattern_compile_def_on_fail = op(ErrorPatMiss,[arg])\n        inl main_arg (pattern_compile arg pat pattern_compile_def_on_succ pattern_compile_def_on_fail) |&gt; expr_prepass\n        )\n<\/pre>\n<p>The pattern compilation function is a rather large function so it is worth going on it in turn. Its purpose is to turn a <code>Pattern<\/code> into an <code>Expr<\/code>ession for easier compilation down the road.<\/p>\n<p>As a brief history of its development, this way of compiling or rather dealing with patterns came only after 4.5 months into Spiral's development and it represents one of the big breakthroughs that gives Spiral the expressiveness that it has now.<\/p>\n<p>Before that Spiral had pattern matching and had so from the start, but instead of compiling <code>Pattern<\/code>s to <code>Expr<\/code>s what it did was try to evaluate them directly in the evaluator. That made dealing with them really, really difficult. Essentially, having another interpreted language directly in the evaluator made the patterns very non-composable and isolated from the rest of the language and during those times as a result Spiral could only deal with simple patterns that could be dealt completely at compile time. Without compiling them it would have been very difficult to implement partially static patterns such as for union types.<\/p>\n<p>Also not doing compilation would necessitate that the prepass also take care of propagating variables through the <code>Pattern<\/code> which made it much more complicated.<\/p>\n<p>Dealing with control flow in <code>Pattern<\/code>s is especially difficult when they are interpreted.<\/p>\n<p>What <code>pattern_compile<\/code> does is not just convert them to <code>Expr<\/code>, but while it is doing that it also CPS's them.<\/p>\n<p>It is worth going through the function top to bottom.<\/p>\n<pre>let pattern_dict = d0()\nlet rec pattern_compile (pat: Node&lt;_&gt;) = \n    pat |&gt; memoize pattern_dict (fun pat -&gt;\n        let node = pat.Symbol\n        let pat = pat.Expression\n        let new_pat_var =\n            let mutable i = 0\n            let get_pattern_tag () = \n                let x = i\n                i &lt;- i + 1\n                x\n            fun () -&gt; sprintf \" pat_var_%i_%i\" node (get_pattern_tag())\n<\/pre>\n<p>This is its entry. In order to take a precaution that a pattern is only compiled once, it is memoized. It slips the authors mind whether this is actually necessary, but here it is.<\/p>\n<p><code>new_pat_var<\/code> is just a tagger function that returns fresh names.<\/p>\n<pre>        let rec pattern_compile (arg: Expr) (pat: Pattern) (on_succ: Expr) (on_fail: Expr): Expr =\n            let inline cp arg pat on_succ on_fail = pattern_compile arg pat on_succ on_fail\n<\/pre>\n<p>If there is a programing problem, the continuation passing style probably has a solution for it. Note that using the <code>on_fail on_succ<\/code> order is preferable to <code>on_succ on_fail<\/code>, but the author got careless this time. This would be fixable, but the type system will not be as helpful as usual because both are of type <code>Expr<\/code> so refactoring is dangerous here. Modules were added in Spiral in order to make deciding function argument order easier and take care of situations like this. The modules are also there for extensibility. In F# since the type system is always on hand it is better to rely on it instead, though the author does hope that Spiral will inspire F# to improve its record syntax.<\/p>\n<p>Instead of diving directly into the function, it would be worth to start near the beginning just before it is called.<\/p>\n<pre>            let main_arg = new_pat_var()\n            let arg = v main_arg\n                    \n            let pattern_compile_def_on_succ = op(ErrorPatClause,[])\n            let pattern_compile_def_on_fail = op(ErrorPatMiss,[arg])\n            inl main_arg (pattern_compile arg pat pattern_compile_def_on_succ pattern_compile_def_on_fail) |&gt; expr_prepass\n<\/pre>\n<p>Always, the pattern is wrapped in an inlineable, and its main argument is passed into the function. By default, the on_succ and on_fail are errors which will trigger if the evaluator ever reaches them.<\/p>\n<p><code>PatClauses<\/code> is always the first to trigger and is never found inside another pattern.<\/p>\n<pre>| PatClauses of (Pattern * Expr) list\n<\/pre>\n<pre>            match pat with\n            | PatClauses l -&gt; List.foldBack (fun (pat, exp) on_fail -&gt; cp arg pat exp on_fail) l on_fail\n<\/pre>\n<p>Compilation of the pattern is a rightwards fold, meaning it proceeds from the end to the beginning.<\/p>\n<p><code>PatClauses<\/code> could be simplified by turning it into a combination of the <code>PatAnd<\/code> and the <code>PatOr<\/code> patterns. The way patterns work is as follows - if the pattern is successful, in the later stage the evaluator will call <code>on_succ<\/code>. Hence why <code>exp<\/code> replaces the default <code>on_succ<\/code> in <code>PatClauses<\/code>.<\/p>\n<p>On the other hand if the evaluator fail to find a match, it will call <code>on_fail<\/code>. It can be visualized like so:<\/p>\n<pre>pat1 pat2 pat3 -&gt; ...\npat4 -&gt; ...\npat5 -&gt; ...\n<\/pre>\n<p>Evaluator always starts at <code>pat1<\/code> and then moves to <code>pat2<\/code> if is successful. If it is a failure it goes to <code>pat4<\/code> and works from there.<\/p>\n<p>That the <code>pattern_compile<\/code> does is through the magic of CPS give the evaluator a choice of simply having the continuations there for it to call. It compiles the patterns into such a form that the desired behavior happens.<\/p>\n<pre>| PatE -&gt; on_succ\n<\/pre>\n<p>Suppose a pattern like <code>inl _ -&gt; body<\/code> directly in the language. There is only one clause and only <code>PatE<\/code> here.<\/p>\n<p>So based on what is known it could be expected that it would get compiled as something like <code>inl main_arg -&gt; on_succ<\/code>. <code>on_succ<\/code> is just the body here and so it works.<\/p>\n<pre>| PatVar x -&gt; l x arg on_succ\n<\/pre>\n<p><code>PatVar<\/code> is similar, but it it the arg must be bound to <code>x<\/code> first.<\/p>\n<p>So <code>inl x -&gt; body<\/code> would be compiled to something like (in pseudo-code):<\/p>\n<pre>inl pat_var_0_0 -&gt; \n    inl x = pat_var_0_0\n    body\n<\/pre>\n<p>Here is how <code>l<\/code>et is implemented in the language.<\/p>\n<pre>let l v b e = ap (inl v e) b\n<\/pre>\n<p>Let statements can easily be desugared into function abstraction and then application. In HM style languages there are some advantages to having separate let statements for the sake of generalization, but not in Spiral so it choses the easiest route. It made debugging a pain since it turns the program inside out and hence unprintable in raw form, but that does not matter anymore.<\/p>\n<h4><a id=\"user-content-type-equality-pattern\" class=\"anchor\" aria-hidden=\"true\" href=\"#type-equality-pattern\"><\/a>Type Equality Pattern<\/h4>\n<pre>            | PatTypeEq (exp,typ) -&gt;\n                let on_succ = cp arg exp on_succ on_fail\n                if_static (eq_type arg typ) on_succ on_fail\n                |&gt; case arg\n<\/pre>\n<p>Type equality is the <code>:<\/code> operator when on the pattern side. How this works could be written directly in the language itself as the following.<\/p>\n<pre>function\n| _: int64 -&gt; body1\n| _: float32 -&gt; body2\n| _: string -&gt; body3\n<\/pre>\n<pre>inl pat_var_1_0 -&gt;\n    if eq_type pat_var_1_0 int64 then body1\n    elif eq_type pat_var_1_0 float32 then body2\n    elif eq_type pat_var_1_0 string then body3\n    else on_fail \/\/ pattern miss\n<\/pre>\n<h4><a id=\"user-content-tuples\" class=\"anchor\" aria-hidden=\"true\" href=\"#tuples\"><\/a>Tuples<\/h4>\n<pre>let list_taken_cps count arg on_fail on_succ = op(ListTakeNCPS,[lit (LitInt32 count);arg;on_fail;on_succ])\nlet list_taken_tail_cps count arg on_fail on_succ = op(ListTakeNTailCPS,[lit (LitInt32 (count-1));arg;on_fail;on_succ])\n...\n            let pat_tuple_helper l =\n                List.foldBack (fun pat (c,s,on_succ) -&gt; \n                    let arg = new_pat_var()\n                    c + 1, arg :: s,cp (v arg) pat on_succ on_fail) l (0,[],on_succ)\n...\n            | PatTuple l -&gt; \n                let count, args, on_succ = pat_tuple_helper l\n                list_taken_cps count arg on_fail (inl' args on_succ) |&gt; case arg\n<\/pre>\n<p><code>pat_tuple_helper<\/code> takes the count of the pattern, makes up new pattern variables for each of the pattern arguments and also recursively maps the inner pattern.<\/p>\n<p>Here is an example of how this would work. The <code>|&gt; case<\/code> will be pretended not to exist for the sake of brevity.<\/p>\n<pre>inl a,b,c -&gt; body\n<\/pre>\n<pre>inl pat_var_2_0 -&gt; \/\/ pat_var_2_0 is main_arg\n    !ListTakeNCPS(\n        3i32\n        ,pat_var_2_0 \n        ,inl pat_var_2_1 pat_var_2_2 pat_var_2_3 -&gt; \n            inl a = pat_var_2_1\n            inl b = pat_var_2_2\n            inl c = pat_var_2_3\n            body\n        ,on_fail \/\/ pattern miss\n        )\n<\/pre>\n<p>As you can see in the above, the pattern compiler is a bit stupid and is doing redundant bindings. This has no runtime impact as <code>inl<\/code>ineables get inlined period, but does cause slowdown at compile time.<\/p>\n<p>Here an example with multiple patterns.<\/p>\n<pre>function\n| a,b,c -&gt; body1\n| q,w -&gt; body2\n<\/pre>\n<pre>inl pat_var_3_0 -&gt; \/\/ pat_var_3_0 is main_arg\n    !ListTakeNCPS(\n        3i32\n        ,pat_var_3_0\n        ,inl pat_var_3_1 pat_var_3_2 pat_var_3_3 -&gt; \n            inl a = pat_var_3_1\n            inl b = pat_var_3_2\n            inl c = pat_var_3_3\n            body1\n        ,!ListTakeNCPS(\n            2i32\n            ,pat_var_3_0\n            ,inl pat_var_3_4 pat_var_3_5 -&gt; \n                inl q = pat_var_3_4\n                inl w = pat_var_3_5\n                body2\n            ,on_fail \/\/ pattern_miss\n            )\n        )\n<\/pre>\n<p>Here is how <code>ListTakeN<\/code> operation is implemented in the evaluator.<\/p>\n<pre>let inline list_taken_template op_name loop d a arg on_fail on_succ = \n    match tev d a with\n    | TyLitIndex c -&gt; \n        match tev d arg with\n        | TyList args -&gt; loop [] (c,args)\n        | _ -&gt; tev d on_fail\n    | x -&gt; on_type_er (trace d) \"Expected an int literal as the first input to %s.\\nGot: %s\" op_name (show_typedexpr x)\nlet list_taken d a arg on_fail on_succ =\n    let rec loop args = function\n        | 0,[] -&gt; List.foldBack (fun arg on_succ -&gt; apply d on_succ arg) args (tev d on_succ)\n        | _,[] | 0, _ -&gt; tev d on_fail\n        | c,x :: x' -&gt; loop (x :: args) (c-1,x')\n    list_taken_template \"ListTakeN\" loop d a arg on_fail on_succ\n<\/pre>\n<p>Figuring this out will take some specialization by hand.<\/p>\n<pre>let list_taken (d: LangEnv) (a: Expr) (arg: Expr) (on_fail: Expr) (on_succ: Expr) = \n    match tev d a with\n    | TyLitIndex c -&gt; \n        match tev d arg with\n        | TyList args -&gt; \n            let rec loop args = function\n                | 0,[] -&gt; List.foldBack (fun arg on_succ -&gt; apply d on_succ arg) args (tev d on_succ)\n                | _,[] | 0, _ -&gt; tev d on_fail\n                | c,x :: x' -&gt; loop (x :: args) (c-1,x')\n            loop [] (c,args)\n        | _ -&gt; tev d on_fail\n    | x -&gt; on_type_er (trace d) \"Expected an int literal as the first input to ListTakeN.\\nGot: %s\" (show_typedexpr x)\n<\/pre>\n<p>The recursive loop is a bit convoluted so it is worth going through step by step. The loop can be split into two phases:<\/p>\n<ol><li>Making sure that the tuple has exactly the specified number of arguments.<\/li>\n<\/ol><pre>                | _,[] | 0, _ -&gt; tev d on_fail\n                | c,x :: x' -&gt; loop (x :: args) (c-1,x')\n<\/pre>\n<p>That is ensured by this particular segment.<\/p>\n<ol start=\"2\"><li>Assuming the number arguments in the tuple is correct, then they are all applied to the <code>on_succ<\/code> in correct order.<\/li>\n<\/ol><pre>                | 0,[] -&gt; List.foldBack (fun arg on_succ -&gt; apply d on_succ arg) args (tev d on_succ)\n<\/pre>\n<p>Why does the loop reverse the argument list before applying it? Absolutely no reason.<\/p>\n<p>If this were the tutorials, the author would quietly fix this, but since the user guide is meant for power users, it would be worth showing how the compiler can be improved in action. Here is the ammended function.<\/p>\n<pre>let list_taken (d: LangEnv) (a: Expr) (arg: Expr) (on_fail: Expr) (on_succ: Expr) = \n    match tev d a with\n    | TyLitIndex c -&gt; \n        match tev d arg with\n        | TyList args -&gt; \n            let rec loop = function\n                | 0,[] -&gt; List.fold (fun on_succ arg -&gt; apply d on_succ arg) (tev d on_succ) args\n                | _,[] | 0, _ -&gt; tev d on_fail\n                | c,_ :: x' -&gt; loop (c-1,x')\n            loop (c,args)\n        | _ -&gt; tev d on_fail\n    | x -&gt; on_type_er (trace d) \"Expected an int literal as the first input to ListTakeN.\\nGot: %s\" (show_typedexpr x)\n<\/pre>\n<p>This actually sped up the compiler by a few % since the tuple pattern is so frequently used.<\/p>\n<p>Moving on to the cons pattern.<\/p>\n<pre>            | PatCons l -&gt; \n                let count, args, on_succ = pat_tuple_helper l\n                list_taken_tail_cps count arg on_fail (inl' args on_succ) |&gt; case arg\n<\/pre>\n<p>It is quite similar to the standard tuple pattern.<\/p>\n<p>So is how it is implemented in the evaluator.<\/p>\n<pre>let list_taken_tail d a arg on_fail on_succ = \n    match tev d a with\n    | TyLitIndex c -&gt; \n        match tev d arg with\n        | TyList args -&gt; \n            let rec loop args = function\n                | 0,x' -&gt; List.foldBack (fun arg on_succ -&gt; apply d on_succ arg) (tyvv x' :: args) (tev d on_succ)\n                | _,[] -&gt; tev d on_fail\n                | c,x :: x' -&gt; loop (x :: args) (c-1,x')\n            loop [] (c,args)\n        | _ -&gt; tev d on_fail\n    | x -&gt; on_type_er (trace d) \"Expected an int literal as the first input to ListTakeNTail.\\nGot: %s\" (show_typedexpr x)\n<\/pre>\n<p>Note that in the cons pattern now collecting the leftover arguments has a definite purpose.<\/p>\n<h4><a id=\"user-content-active-patterns\" class=\"anchor\" aria-hidden=\"true\" href=\"#active-patterns\"><\/a>Active Patterns<\/h4>\n<pre>            | PatActive (a,b) -&gt;\n                let pat_var = new_pat_var()\n                l pat_var (ap (v a) arg) (cp (v pat_var) b on_succ on_fail)\n<\/pre>\n<p>Here is an example how this works.<\/p>\n<pre>inl !dyn x -&gt; body\n<\/pre>\n<pre>inl pat_var_4_0 -&gt;\n    inl pat_var_1 = dyn pat_var_4_0\n    inl x = pat_var_1\n    body\n<\/pre>\n<p>Here is the partial active pattern.<\/p>\n<pre>            | PatPartActive (a,pat) -&gt; \n                let pat_var = new_pat_var()\n                let on_succ = inl pat_var (cp (v pat_var) pat on_succ on_fail)\n                let on_fail = inl \"\" on_fail\n                ap' (v a) [arg; on_fail; on_succ]\n<\/pre>\n<p>Here is an example of its compilation in action.<\/p>\n<pre>inl f arg on_fail on_succ = on_succ arg\ninl @f x -&gt; body\n<\/pre>\n<pre>inl pat_var_5_0 -&gt;\n    inl on_succ pat_var_5_1 = \n        inl x = pat_var_5_1\n        body\n    inl on_fail _ = on_fail \/\/ pattern miss\n    f pat_var_5_0 on_fail on_succ\n<\/pre>\n<p>Spiral's ability to propagate information deeply makes it easy to compile active patterns to CPS'd control flow.<\/p>\n<h4><a id=\"user-content-boolean-patterns\" class=\"anchor\" aria-hidden=\"true\" href=\"#boolean-patterns\"><\/a>Boolean Patterns<\/h4>\n<pre>            | PatOr l -&gt; List.foldBack (fun pat on_fail -&gt; cp arg pat on_succ on_fail) l on_fail\n            | PatAnd l -&gt; List.foldBack (fun pat on_succ -&gt; cp arg pat on_succ on_fail) l on_succ\n            | PatNot p -&gt; cp arg p on_fail on_succ \/\/ switches the on_fail and on_succ arguments\n<\/pre>\n<p><code>PatOr<\/code> links the patterns through the <code>on_fail<\/code>. <code>PatAnd<\/code> links the arguments through the <code>on_succ<\/code>. <code>PatNot<\/code> just switches the arguments.<\/p>\n<p>The examples won't be provided for the above 3, the users can rest assured that they are doing their job.<\/p>\n<pre>            | PatXor l -&gt;\n                let state_var = new_pat_var()\n                let state_var' = v state_var\n                let bool x = lit &lt;| LitBool x\n                let rec just_one = function\n                    | x :: xs -&gt; \n                        let xs = just_one xs\n                        inl state_var \n                            (cp arg x \n                                (if_static state_var' on_fail (ap xs (bool true))) \/\/ true case\n                                (ap xs state_var')) \/\/ false case\n                    | [] -&gt; inl state_var (if_static state_var' on_succ on_fail)\n                ap (just_one l) (bool false)\n<\/pre>\n<p><code>PatXor<\/code> is definitely the most complex of all the patterns. Currently it can only be placed inside the module pattern. It is also misleadingly named as unlike the boolean pattern, it cannot be used to flip back and forth.<\/p>\n<p>For <code>pat0 ^ pat1 ... ^ patn<\/code> it ensures that only one of the patterns triggers. And is false when there are no patterns.<\/p>\n<p>It is a bit annoying to sketch out what it does hand by hand, so as an alternative consider this example written in F#.<\/p>\n<pre>let rec just_one state = function\n    | true :: xs -&gt; if state then false else just_one true xs\n    | false :: xs -&gt; just_one state xs\n    | [] -&gt; state\njust_one false [] \/\/ false\njust_one false [true;false] \/\/ true\njust_one false [false;false] \/\/ false\njust_one false [false;true;false] \/\/ true\njust_one false [false;false;true] \/\/ true\njust_one false [true;false;true] \/\/ false\n<\/pre>\n<p>What the <code>PatXor<\/code> does is translate the above into execution flow for the evaluator. It is just one example of how compilation makes composing languages much easier than interpretation. This is also an argument in favor of powerful type systems such as Spiral's because they make composition much easier. After a certain point this translates into a permanent advantage beyond what even dynamic languages offer due to the ability to propagate information through language boundaries.<\/p>\n<h4><a id=\"user-content-literal-patterns\" class=\"anchor\" aria-hidden=\"true\" href=\"#literal-patterns\"><\/a>Literal Patterns<\/h4>\n<pre>            | PatLit x -&gt; \n                let x = lit x\n                let on_succ = if_static (eq arg x) on_succ on_fail\n                if_static (eq_type arg x) on_succ on_fail |&gt; case arg\n<\/pre>\n<p>At some point in the future these might get their own dedicated CPS operations just like tuples, but it is interesting to study them the way they are right now. If the later this changes in the source there is no need to be surprised. The meaning of the pattern won't change.<\/p>\n<p>What is happening above is that first a type equality check is made (otherwise the equality could throw a type error) and then the equality is tested for. Here is how that would compile.<\/p>\n<pre>function\n| 1 -&gt; body1\n| 2 -&gt; body2\n<\/pre>\n<pre>inl pat_var_6_0 -&gt;\n    if eq_type pat_var_6_0 1 then\n        if pat_var_6_0 = 1 then body1\n        else \n            if eq_type pat_var_6_0 2 then\n                if pat_var_6_0 = 2 then body2\n                else on_fail \/\/ pattern miss\n            else on_fail \/\/ pattern miss\n    elif eq_type pat_var_6_0 2 then\n        if pat_var_6_0 = 2 then body2\n        else on_fail \/\/ pattern miss\n    else on_fail \/\/ pattern miss\n<\/pre>\n<p>The above demonstrates why would it be good to replace the operation with a single CPS'd operation.<\/p>\n<pre>            | PatTypeLit x -&gt; \n                if_static (eq_type arg (type_lit_lift x)) on_succ on_fail \n                |&gt; case arg\n<\/pre>\n<p>Since type literals are types they are compared using <code>eq_type<\/code> rather than <code>=<\/code>. Trying to do so with <code>=<\/code> would cause a type error.<\/p>\n<pre>            | PatTypeLitBind x -&gt; \n                if_static (type_lit_is arg) (l x (type_lit_cast arg) on_succ) on_fail \n                |&gt; case arg\n<\/pre>\n<p>The language has specialized operations in order to cast type literals to the term level like it is possible with functions. The full list of <code>Op<\/code>eratives will be in the reference.<\/p>\n<pre>let type_lit_cast x = (TypeLitCast,[x]) |&gt; op\n<\/pre>\n<p>Here is how <code>type_lit_cast<\/code> is implemented. The following is how it is done in the evaluator.<\/p>\n<pre>let type_lit_cast d a =\n    match tev d a with\n    | TyT (LitT x) -&gt; TyLit x\n    | _ -&gt; on_type_er (trace d) \"Expected a literal in type literal cast.\"\n<\/pre>\n<p>Type literals are always expected to be naked types.<\/p>\n<h4><a id=\"user-content-module-patterns\" class=\"anchor\" aria-hidden=\"true\" href=\"#module-patterns\"><\/a>Module Patterns<\/h4>\n<pre>            | PatModuleIs p -&gt; module_is_cps arg on_fail (cp arg p on_succ on_fail) |&gt; case arg\n            | PatModuleMember name -&gt; module_member_cps arg name on_fail (inl name on_succ) |&gt; case arg\n            | PatModuleRebind(name,b) -&gt; \n                let arg' = new_pat_var()    \n                module_member_cps arg name on_fail (inl arg' (cp (v arg') b on_succ on_fail)) \n                |&gt; case arg\n<\/pre>\n<p>The patterns remaining are nearly over. Explicit examples won't be provided for the above as it should be straightforward by now. As a note, <code>PatModuleIs<\/code> is only created when the module pattern is empty.<\/p>\n<p>Here is how <code>module_is_cps<\/code> is implemented in the evaluator.<\/p>\n<pre>let module_is_cps d a on_fail on_succ =\n    match tev d a with\n    | M(_,_,MapTypeModule) -&gt; tev d on_succ\n    | _ -&gt; tev d on_fail\n<\/pre>\n<p>Here is the other one.<\/p>\n<pre>let module_member_cps d a b on_fail on_succ =\n    match tev2 d a b with\n    | M(layout,env_term,MapTypeModule) &amp; recf, b -&gt; \n        match b with\n        | TyLit (LitString n) -&gt; \n            let inline unpack k = v_find env_term n (fun () -&gt; tev d on_fail) k\n            match layout with\n            | None -&gt; unpack (apply d (tev d on_succ))\n            | _ -&gt; unpack (apply d (tev d on_succ) &lt;&lt; layout_boxed_unseal d recf)\n        | x -&gt; on_type_er (trace d) &lt;| sprintf \"Expecting a string as the second argument to ModuleMemberCPS.\\nGot: %s\" (show_typedexpr x)\n    | x,_ -&gt; tev d on_fail\n<\/pre>\n<p>Whatever complexity is here, it is because the module pattern has to also work on layout types. Layout types need to be explicitly unsealed instead of reached into directly.<\/p>\n<p>Here is how <code>v_find<\/code> is implemented.<\/p>\n<pre>let inline v_find env x on_fail on_succ = \n    let run env = \n        match Map.tryFind x env with\n        | Some v -&gt; on_succ v\n        | None -&gt; on_fail()\n    match env with\n    | Env env -&gt; run env\n    | EnvConsed env -&gt; run env.node\n    | EnvUnfiltered (env, used_vars) -&gt; if used_vars.Contains x then run env else on_fail()\n<\/pre>\n<p>It roughly as one would expect. It takes an <code>env<\/code> and two continuations and runs them depending on whether it finds the member in the dictionary.<\/p>\n<p>With that in mind, here is <code>unpack<\/code> again with the type of <code>k<\/code> added.<\/p>\n<pre>let inline unpack (k: TypedExpr -&gt; TypedExpr) = v_find env_term n (fun () -&gt; tev d on_fail) k\n<\/pre>\n<p><code>k<\/code> is merely the <code>on_succ<\/code> argument.<\/p>\n<p>If the code is unclear then what the suggested course of action is to try inlining the code by hand and seeing if starts making sense.<\/p>\n<h4><a id=\"user-content-when-pattern\" class=\"anchor\" aria-hidden=\"true\" href=\"#when-pattern\"><\/a>When Pattern<\/h4>\n<pre>            | PatWhen (p, e) -&gt; cp arg p (if_static e on_succ on_fail) on_fail\n<\/pre>\n<p>This one is straightforward. The author had more difficulty parsing it than implementing it. What it does is first tries the left pattern side and makes the <code>on_succ<\/code> argument the when check.<\/p>\n<h4><a id=\"user-content-term-function-type-pattern\" class=\"anchor\" aria-hidden=\"true\" href=\"#term-function-type-pattern\"><\/a>Term Function Type Pattern<\/h4>\n<pre>            | PatTypeTermFunction(a,b) -&gt; \n                let va, vb = new_pat_var(), new_pat_var()\n                term_fun_dom_range_cps arg on_fail \n                &lt;| inl' [va; vb] (cp (v va) a (cp (v vb) b on_succ on_fail) on_fail)\n<\/pre>\n<p>This the <code>=&gt;<\/code> pattern that was covered during the implementation of <code>closure_of<\/code>. Here is how it is implemented in the evaluator.<\/p>\n<pre>        let term_fun_dom_range_cps d x on_fail on_succ =\n            match tev d x with\n            | TyType(TermFunctionT (a,b)) -&gt; \n                let on_succ = tev d on_succ    \n                apply d (apply d on_succ (tyt a)) (tyt b)\n            | x -&gt; tev d on_fail\n<\/pre>\n<p>It a pretty straightforward mapping. It extracts the types and applies them to the <code>on_succ<\/code> continuation otherwise it calls <code>on_fail<\/code>.<\/p>\n<h4><a id=\"user-content-the-pos-pattern\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-pos-pattern\"><\/a>The Pos Pattern<\/h4>\n<pre>| PatPos p -&gt; expr_pos p.Pos (cp arg p.Expression on_succ on_fail)\n<\/pre>\n<p>This one is not placed by the user but by the parser. This pattern is converted into the <code>ExprPos<\/code> node.<\/p>\n<p>Here is how it looks like in the evaluator.<\/p>\n<pre>let inline add_trace (d: LangEnv) x = {d with trace = x :: d.trace}\n...\n| ExprPos p -&gt; tev (add_trace d p.Pos) p.Expression\n<\/pre>\n<p>The trace is of course what the user gets when he makes a type error.<\/p>\n<h4><a id=\"user-content-a-note-on-case\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-note-on-case\"><\/a>A Note On Case<\/h4>\n<p>For the purpose of simplification whenever <code>|&gt; case arg<\/code> has appeared, it has been ignored.<\/p>\n<pre>let case arg case = (Case,[arg;case]) |&gt; op\n<\/pre>\n<p>What it does has been covered in the <code>Unboxing of Union Types<\/code> chapter of the user guide.<\/p>\n<h3><a id=\"user-content-9-layout-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#9-layout-types\"><\/a>9: Layout Types<\/h3>\n<p>Layout types came relatively late in Spiral's development, about six months in. The reason why that is remarkable is because the author thought for a long time on how to do them and eventually concluded that they were impossible. The reason for that is because having them would mean having to essentially capture chunks of scope and how could something like that possibly be done?<\/p>\n<p>Quite simply.<\/p>\n<p>First, it is time to take the lid off the list of all the <code>Ty<\/code>pes in Spiral. All of these have been demonstrated so far.<\/p>\n<pre>and Ty =\n    | PrimT of PrimitiveType\n    | ListT of Ty list\n    | LitT of Value\n    | MapT of EnvTy * MapType \/\/ function or module\n    | LayoutT of LayoutType * EnvTerm * MapType\n    | TermFunctionT of Ty * Ty\n    | UnionT of Set&lt;Ty&gt;\n    | RecT of JoinPointKey\n    | ArrayT of ArrayType * Ty\n    | DotNetTypeT of TypedExpr \/\/ macro\n    | CudaTypeT of TypedExpr \/\/ macro\n<\/pre>\n<p>Note the similarities between <code>MapT<\/code> and <code>LayoutT<\/code>.<\/p>\n<p>The type of <code>EnvTy<\/code> is <code>Map&lt;string,Ty&gt;<\/code>. While <code>EnvTerm<\/code> can be thought of as <code>Map&lt;string,TypedExpr&gt;<\/code>. <code>LayoutType<\/code> is rather simple.<\/p>\n<pre>type LayoutType =\n    | LayoutStack\n    | LayoutPackedStack\n    | LayoutHeap\n    | LayoutHeapMutable\n<\/pre>\n<p>It is just some extra data denoting what kind of type it is. Functionally though, <code>MapT<\/code> and <code>LayoutT<\/code> can be thought of as duals of each other. <code>LayoutT<\/code> is just a map with a bit more information propagated in it.<\/p>\n<p>Here is how the types are created.<\/p>\n<pre>let layout_to_none' d = function\n    | TyMap _ as a -&gt; a\n    | TyType(LayoutT(_,env,t)) as a -&gt; tymap(layout_env_term_unseal d a env,t)\n    | x -&gt; on_type_er (trace d) &lt;| sprintf \"Cannot turn the argument into a non-layout type. Got: %s\" (show_typedexpr x)\nlet layout_to_none d a = layout_to_none' d (tev d a)\nlet rec layoutify (layout: LayoutType) (d: LangEnv) = function\n    | TyMap(C env,t) -&gt;\n        let {renamer'=r}, env' = renamer_apply_envc env\n        if r.Count = 0 then LayoutT(layout,env',t) |&gt; tyt\n        else TyOp(layout_to_op layout,[a],LayoutT(layout,env',t)) |&gt; destructure d\n    | TyType(LayoutT(layout',_,_)) as a -&gt;\n        if layout &lt;&gt; layout' then layout_to_none' d a |&gt; layoutify layout d else a\n    | x -&gt; on_type_er (trace d) &lt;| sprintf \"Cannot turn the argument into a layout type. Got: %s\" (show_typedexpr x)\nlet layout_to layout d a = layoutify layout d (tev d a)\n<\/pre>\n<p>The above could be summarized as - run the renamer and then plug the resulting environment into the <code>LayoutT<\/code>. This is exactly what the join points do for their keys.<\/p>\n<p>If one can understand renaming then one can understand layout types.<\/p>\n<p>The only remaining thing is to show how the can be unsealed.<\/p>\n<pre>let rec layout_boxed_unseal d recf x =\n    let inline f x = layout_boxed_unseal d recf x\n    match x with\n    | TyV _ as v -&gt; TyOp(MapGetField,[recf;v],get_type v) |&gt; destructure d\n    | TyList l -&gt; tyvv (List.map f l)\n    | TyBox(a,b) -&gt; tybox (f a, b)\n    | TyMap(env, b) -&gt; tymap (layout_env_term_unseal d recf env, b)\n    | x -&gt; x\n        \nand layout_env_term_unseal d recf (C env) = Map.map (fun _ -&gt; layout_boxed_unseal d recf) env |&gt; Env\n<\/pre>\n<p><code>recf<\/code> is in the above two functions is just the variable being unsealed.<\/p>\n<p>Unsealing layout types is a standard map over <code>TypedExpr<\/code>. Note that if the unsealing is done twice on the same variable in local scope, CSE will prevent duplicate work from being done.<\/p>\n<pre>TyOp(MapGetField,[recf;v],get_type v) |&gt; destructure d\n<\/pre>\n<p>This is due to <code>destructure<\/code>. It will send the <code>TyOp<\/code>s to the CSE dictionary and will remember if they tried to be unsealed again.<\/p>\n<p>This is what is needed to understand layout types. With this all the features that are core to the full Spiral experience have been covered in depth.<\/p>\n<p><code>layout_boxed_unseal<\/code> and <code>layout_env_term_unseal<\/code> are used in various places throughout the evaluator when opening the layout types is needed. This won't be covered in the user guide.<\/p>\n<p>If you've managed to get this far, by now you are reasonably familiar with not just the language, but with its internals as well.<\/p>\n<p>Past this if you would like to know more, looking into the source is a reasonable option. Spiral is not a huge overbearing monolith of a language with hundreds of thousands of lines of code. The evaluator itself is at the time of writing 1.6k lines long. It is not commented, but a significant amount of effort has gone into refactoring it and you can be sure that if there is a feature in the language, you can find it in just one place in the evaluator.<\/p>\n<p>Well, apart from the <code>-<\/code> and <code>.<\/code> operators in the parsing stage. The author hates those things.<\/p>\n<p>If that is not enough, just ask him.<\/p>\n<h3><a id=\"user-content-10-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#10-macros\"><\/a>10: Macros<\/h3>\n<p>In Spiral macros can be treated as types and used to instantiate types which makes them unique in a statically typed language. They are definitely not needed to understand Spiral, but they should be in the user guide nonetheless.<\/p>\n<pre>let inline codegen_macro' show_typedexpr codegen print_type x = \n    let strb = StringBuilder()\n    let inline append (x: string) = strb.Append x |&gt; ignore\n    let (|LS|) = function\n            | TyLit (LitString x) | TypeString x -&gt; x\n            | _ -&gt; failwithf \"Iter's first three arguments must be strings.\"\n    let er x = failwithf \"Unknown argument in macro. Got: %s\" (show_typedexpr x)\n    let rec f = function\n        | TyList [TypeString \"text\"; LS x] -&gt; append x\n        | TyList [TypeString \"arg\"; x] -&gt; append (codegen x)\n        | TyList [TypeString \"args\"; TyTuple l] -&gt; append \"(\"; List.map codegen l |&gt; String.concat \", \" |&gt; append; append \")\"\n        | TyList [TypeString \"fs_array_args\"; TyTuple l] -&gt; append \"[|\"; List.map codegen l |&gt; String.concat \"; \" |&gt; append; append \"|]\"\n        | TyList [TypeString \"type\"; TyType x] -&gt; append (print_type x)\n        | TyList [TypeString \"types\"; TyTuple l] -&gt; append \"&lt;\"; List.map (get_type &gt;&gt; print_type) l |&gt; String.concat \", \" |&gt; append; append \"&gt;\" \n        | TyList [TypeString \"iter\"; TyList [LS begin_;LS sep;LS end_;ops]] -&gt;\n                append begin_\n                match ops with\n                | TyList (x :: xs) -&gt; f x; List.iter (fun x -&gt; append sep; f x) xs\n                | TyList [] -&gt; ()\n                | x -&gt; er x\n                append end_\n        | x -&gt; er x\n    match x with\n    | TyList x -&gt; List.iter f x\n    | x -&gt; er x\n    strb.ToString()\n<\/pre>\n<p>This function definitely looks like it came from the code generation phase and that would be correct. <code>show_typedexpr<\/code> is just there for printing errors. <code>codegen<\/code> for printing <code>TypedExpr<\/code>s and <code>print_type<\/code> for printing <code>Ty<\/code>pes are provided by the evaluator and the code generators separately and that allows macros to be printed in one way inside errors and in another way in code while retaining the same structure underneath.<\/p>\n<p>Macros are an ad-hoc feature there just for the sake of language interoperability and while needed, they should not be considered core to the language nor should be used for abstraction. They should be redone into something more principled.<\/p>\n<p>As it is their functionality just arose from needs of various sorts.<\/p>\n<p>List of operations:<\/p>\n<ol><li><code>text<\/code><\/li>\n<\/ol><p>Takes in a literal or a type string literal and prints it.<\/p>\n<ol start=\"2\"><li><code>arg<\/code><\/li>\n<\/ol><p>Takes in an argument and prints it.<\/p>\n<ol start=\"2\"><li><code>args<\/code><\/li>\n<\/ol><p>Takes in an argument or a tuple of arguments and prints them between parentheses and separated by commas.<\/p>\n<ol start=\"3\"><li><code>fs_array_args<\/code><\/li>\n<\/ol><p>Takes in an argument or a tuple of arguments and prints them between <code>[|<\/code> and <code>|]<\/code> and separated by semicolons.<\/p>\n<ol start=\"4\"><li><code>type<\/code><\/li>\n<\/ol><p>Takes in an argument and prints its type.<\/p>\n<ol start=\"5\"><li><code>types<\/code><\/li>\n<\/ol><p>Takes in an argument or a tuple of arguments and prints their type them between <code>&lt;<\/code> and <code>&gt;<\/code> and separated by commas.<\/p>\n<ol start=\"6\"><li><code>iter<\/code><\/li>\n<\/ol><p>It takes in the opener, separator and closer as strings or type level strings and a tuple of macro operations and executes them while printing the separator in-between them and the opener and the closer at the beginning and the end respectively.<\/p>\n<h4><a id=\"user-content-parser-macros\" class=\"anchor\" aria-hidden=\"true\" href=\"#parser-macros\"><\/a>Parser Macros<\/h4>\n<p>There are only 3 of them.<\/p>\n<pre>let case_parser_macro expr = \n    inbuilt_op_core '@' &gt;&gt;= fun a -&gt;\n        match a with\n        | \"PathCuda\" -&gt; settings.path_cuda90 |&gt; LitString |&gt; lit |&gt; preturn\n        | \"PathCub\" -&gt; settings.path_cub |&gt; LitString |&gt; lit |&gt; preturn\n        | \"PathVS2017\" -&gt; settings.path_vs2017 |&gt; LitString |&gt; lit |&gt; preturn\n        | a -&gt; failFatally &lt;| sprintf \"%s is not a valid parser macro.\" a\n<\/pre>\n<p>They are invoked like <code>inl path_cuda = @PathCuda<\/code> and so on. They are just there as a safeguard in case join point dictionaries ever get floated up a level. If that happens changing the compiler setting might give incorrect results if the paths were inserted during the partial evaluation stage since join points would not recognize the changes.<\/p>\n<h3><a id=\"user-content-11-operatives-and-the-core-library\" class=\"anchor\" aria-hidden=\"true\" href=\"#11-operatives-and-the-core-library\"><\/a>11: Operatives and the Core Library<\/h3>\n<p>Invoking operatives is the main way the user communicates with the partial evaluator. In Spiral most operations are virtualized and compile to calls to core library functions.<\/p>\n<p>Here is how it can be done.<\/p>\n<pre>inl (+) a b = !Add(a,b)\n<\/pre>\n<p>The <code>Add<\/code> is an <code>Op<\/code>erative and the <code>!<\/code> is the operative invocation unary operator. At the time of writing there are 93 <code>Op<\/code>eratives in total. A full list of them can be found with the rest of the types in <code>SpiralTypes.fs<\/code>.<\/p>\n<p>The core library can be found in <code>SpiralCoreLib.fs<\/code>. It is opened automatically every time the program is compiled.<\/p>\n<p>Here is a copy of it for easy perusal.<\/p>\n<pre>module Spiral.CoreLib\nlet module_ = Types.module_\nlet core =\n    (\n    \"Core\",[],\"The Core module.\",\n    \"\"\"\n\/\/\/ Lifts a literal to the type level.\ninl type_lit_lift x = !TypeLitCreate(x)\n\/\/\/ Raises a type error.\ninl error_type x = !ErrorType(x)\n\/\/\/ Prints an expression at compile time.\ninl print_static x = !PrintStatic(x)\n\/\/\/ Pushes the expression to runtime.\ninl dyn x = !Dynamize(x)\n\/\/\/ Creates a term function with the given two types.\ninl (=&gt;) a b = !TermFunctionTypeCreate(a,b)\n\/\/\/ Splits the union or recursive type into a tuple.\ninl split x = !TypeSplit(x)\n\/\/\/ Boxes a type.\ninl box a b = !TypeBox(a,b)\n\/\/\/ Converts module or a function to a stack layout type.\ninl stack x = !LayoutToStack(x)\n\/\/\/ Converts module or a function to a packed stack layout type.\ninl packed_stack x = !LayoutToPackedStack(x)\n\/\/\/ Converts module or a function to a heap layout type.\ninl heap x = !LayoutToHeap(x)\n\/\/\/ Converts module or a function to a mutable heap layout type.\ninl heapm x = !LayoutToHeapMutable(x)\n\/\/\/ Converts a layout type to a module or a function.\ninl indiv x = !LayoutToNone(x)\n\/\/\/ The type of a boolean.\ninl bool = type true \n\/\/\/ The type of a int64.\ninl int64 = type 0i64 \n\/\/\/ The type of a int32.\ninl int32 = type 0i32 \n\/\/\/ The type of a int16.\ninl int16 = type 0i16 \n\/\/\/ The type of a int8.\ninl int8 = type 0i8 \n\/\/\/ The type of a uint64.\ninl uint64 = type 0u64 \n\/\/\/ The type of a uint32.\ninl uint32 = type 0u32 \n\/\/\/ The type of a uint16.\ninl uint16 = type 0u16 \n\/\/\/ The type of a uint8.\ninl uint8 = type 0u8\n\/\/\/ The type of a float64.\ninl float64 = type 0f64\n\/\/\/ The type of a float32.\ninl float32 = type 0f32\n\/\/\/ The type of a string.\ninl string = type \"\"\n\/\/\/ The type of a char.\ninl char = type ' '\n\/\/\/ The type of a empty tuple.\ninl unit = type ()\n\/\/\/ Casts a type literal to the term level.\ninl type_lit_cast x = !TypeLitCast(x)\n\/\/\/ Returns whether the expression is a type literal as a bool.\ninl type_lit_is x = !TypeLitIs(x)\n\/\/\/ Cast a function to the term level.\ninl term_cast to from = !TermCast(to,from)\n\/\/\/ Does unchecked conversion for primitives.\ninl unsafe_convert to from = !UnsafeConvert(to,from) \n\/\/\/ Unary negation.\ninl negate x = !Neg(x)\n\/\/\/ Evaluates an expression and throws away the result.\ninl ignore x = ()\n\/\/\/ Returns an expression after evaluating it.\ninl id x = x\n\/\/\/ Throws away the second argument and returns the first.\ninl const x _ = x\n\/\/\/ Creates a reference.\ninl ref x = !ReferenceCreate(x)\n\/\/\/ Creates an array with the given type and the size.\ninl array_create typ size = !ArrayCreate(size,typ)\n\/\/\/ Returns the length of an array. Not applicable to Cuda arrays.\ninl array_length ar = !ArrayLength(ar)\n\/\/\/ Partial active pattern. In `on_succ` it also passes the type of the array as a type string.\ninl array_is x on_fail on_succ = !ArrayIs(x,on_fail,on_succ)\n\/\/\/ Type of an array with the given type.\ninl array t = type (array_create t 1)\n\/\/\/ Binary addition.\ninl (+) a b = !Add(a,b)\n\/\/\/ Binary subtraction.\ninl (-) a b = !Sub(a,b)\n\/\/\/ Binary multiplication.\ninl (*) a b = !Mult(a,b)\n\/\/\/ Binary division.\ninl (\/) a b = !Div(a,b)\n\/\/\/ Binary modulo.\ninl (%) a b = !Mod(a,b)\n\/\/\/ Applies the first argument to the second.\ninl (|&gt;) a b = b a\n\/\/\/ Applies the second argument to the first.\ninl (&lt;|) a b = a b\n\/\/\/ Applies the third argument to the first and then the result of that to the second.\ninl (&gt;&gt;) a b x = b (a x)\n\/\/\/ Applies the third argument to the second and then the result of that to the first.\ninl (&lt;&lt;) a b x = a (b x)\n\/\/\/ Binary less-than-or-equals.\ninl (&lt;=) a b = !LTE(a,b)\n\/\/\/ Binary less-than.\ninl (&lt;) a b = !LT(a,b)\n\/\/\/ Binary equals.\ninl (=) a b = !EQ(a,b)\n\/\/\/ Binary unequals.\ninl (&lt;&gt;) a b = !NEQ(a,b)\n\/\/\/ Binary greater-than.\ninl (&gt;) a b = !GT(a,b)\n\/\/\/ Binary greater-than-or-equals.\ninl (&gt;=) a b = !GTE(a,b)\n\/\/\/ Bitwise and.\ninl (&amp;&amp;&amp;) a b = !BitwiseAnd(a,b)\n\/\/\/ Bitwise or.\ninl (|||) a b = !BitwiseOr(a,b)\n\/\/\/ Bitwise xor.\ninl (^^^) a b = !BitwiseXor(a,b)\n\/\/\/ Tuple cons.\ninl (::) a b = !ListCons(a,b)\n\/\/\/ Shift left.\ninl (&lt;&lt;&lt;) a b = !ShiftLeft(a,b)\n\/\/\/ Shift right.\ninl (&gt;&gt;&gt;) a b = !ShiftRight(a,b)\n\/\/\/ Gets the first elements of a tuple.\ninl fst x :: _ = x\n\/\/\/ Gets the second element of a tuple.\ninl snd _ :: x :: _ = x\n\/\/\/ Unary negation.\ninl not x = x = false\n\/\/\/ Returns the length of a string.\ninl string_length x = !StringLength(x)\n\/\/\/ The .NET String.Format function.\n\/\/\/ https:\/\/msdn.microsoft.com\/en-us\/library\/system.string.format(v=vs.110).aspx\n\/\/\/ When its arguments are literals, it evaluates at compile time.\ninl string_format a b = !StringFormat(a,b)\n\/\/\/ The F# String.concat function\n\/\/\/ https:\/\/msdn.microsoft.com\/en-us\/visualfsharpdocs\/conceptual\/string.concat-function-%5Bfsharp%5D\n\/\/\/ When its arguments are literals, it evaluates at compile time.\ninl string_concat a b = !StringConcat(a,b)\n\/\/\/ Returns boolean whether the expression is a literal.\ninl lit_is x = !LitIs(x)\n\/\/\/ Returns boolean whether the expression is a box (but not an union type.)\ninl box_is x = !BoxIs(x)\n\/\/\/ Returns boolean whether the expression is a union or a recursive type (excluding boxes.)\ninl caseable_is x = !CaseableIs(x)\n\/\/\/ Returns boolean whether the expression is a union or a recursive type.\ninl caseable_box_is x = !CaseableBoxIs(x)\n\/\/\/ Raises an exception at runtime.\ninl failwith typ msg = !FailWith(typ,msg)\n\/\/\/ Asserts an expression. If the conditional is a literal it raises a type error instead.\ninl assert c msg = \n    inl raise = \n        if lit_is c then error_type\n        else failwith unit\n    \n    if c = false then raise msg\n\/\/\/ Returns the maximum of the two expressions.\ninl max a b = if a &gt; b then a else b\n\/\/\/ Returns the minimum of the two expressions.\ninl min a b = if a &gt; b then b else a\n\/\/\/ Returns boolean whether the two expressions are equal in their types.\ninl eq_type a b = !EqType(a,b)\n\/\/\/ Returns the values of a module in a tuple.\ninl module_values x = !ModuleValues(x)\n\/\/\/ Maps over a module.\n\/\/\/ (string type_lit -&gt; a -&gt; b) -&gt; a module -&gt; b module\ninl module_map f a = !ModuleMap(f,a)\n\/\/\/ Filters a module at compile time.\n\/\/\/ (string type_lit -&gt; a -&gt; bool) -&gt; a module -&gt; a module\ninl module_filter f a = !ModuleFilter(f,a)\n\/\/\/ Folds over a module left to right.\n\/\/\/ (string type_lit -&gt; state -&gt; a -&gt; state) -&gt; state -&gt; a module -&gt; state\ninl module_foldl f s a = !ModuleFoldL(f,s,a)\n\/\/\/ Folds over a module right to left.\n\/\/\/ (string type_lit -&gt; a -&gt; state -&gt; state) -&gt; a module -&gt; state -&gt; state\ninl module_foldr f a s = !ModuleFoldR(f,s,a)\n\/\/\/ Returns boolean whether the module has a member.\n\/\/\/ a module -&gt; string type_lit -&gt; bool\ninl module_has_member m x = !ModuleHasMember(m,x)\n\/\/\/ Unsafe upcast. Unlike the F# compiler, Spiral won't check its correctness.\ninl (:&gt;) a b = !UnsafeUpcastTo(b,a)\n\/\/\/ Unsafe downcast. Unlike the F# compiler, Spiral won't check its correctness.\ninl (:?&gt;) a b = !UnsafeDowncastTo(b,a)\n\/\/\/ Structural polymorphic equality for every type in the language (apart from functions.)\ninl (=) a b =\n    inl prim_eq = (=)\n    inl rec (=) a b =\n        inl body = function\n            | .(a), .(b) -&gt; a = b\n            | a :: as', b :: bs -&gt; a = b &amp;&amp; as' = bs\n            | {} &amp; a, {} &amp; b -&gt; module_values a = module_values b\n            | (), () -&gt; true\n\t\t\t| a, () -&gt; false \/\/ Just in case `b` has not been `Case`d up to this point. This can happen for `(int64, int64) \\\/ int64` kind of types.\n            | a, b when eq_type a b -&gt; prim_eq a b \/\/ This repeat eq_type check is because unboxed union types might lead to variables of different types to be compared.\n            | _ -&gt; false\n        if caseable_is a &amp;&amp; caseable_is b then join (body (a, b) : bool)\n        else body (a, b)\n    if eq_type a b then a = b\n    else error_type (\"Trying to compare variables of two different types. Got:\",a,b)\n\/\/\/ Returns the size a type.\n\/\/\/ type -&gt; int64\ninl sizeof x = !SizeOf(x)\n\/\/\/ Creates a .NET type from a macro.\ninl fs x = !DotNetTypeCreate(x)\n\/\/\/ Creates a Cuda type from a macro.\ninl cd x = !CudaTypeCreate(x)\n\/\/\/ Natural Logarithm.\ninl log x = !Log(x)\n\/\/\/ Exponent.\ninl exp x = !Exp(x)\n\/\/\/ Hyperbolic tangent. \ninl tanh x = !Tanh(x)\n\/\/\/ Macros.\ninl macro = {\n    \/\/\/ F# macro.\n    fs = inl typ expr -&gt; !MacroFs(typ,expr)\n    \/\/\/ Cuda macro.\n    cd = inl typ expr -&gt; !MacroCuda(typ,expr)\n    }\n{type_lit_lift error_type print_static dyn (=&gt;) cd fs log exp tanh array_create array_length array_is array\n split box stack packed_stack heap heapm indiv bool int64 int32 int16 int8 uint64 uint32 uint16 uint8 float64 float32\n string char unit type_lit_cast type_lit_is term_cast unsafe_convert negate ignore id const ref (+) (-) (*) (\/) (%)\n (|&gt;) (&lt;|) (&gt;&gt;) (&lt;&lt;) (&lt;=) (&lt;) (=) (&lt;&gt;) (&gt;) (&gt;=) (&amp;&amp;&amp;) (|||) (^^^) (::) (&lt;&lt;&lt;) (&gt;&gt;&gt;) fst snd not macro\n string_length lit_is box_is failwith assert max min eq_type module_values caseable_is caseable_box_is (:&gt;)\n (:?&gt;) (=) module_map module_filter module_foldl module_foldr module_has_member sizeof string_format string_concat} |&gt; stack\n    \"\"\") |&gt; module_\n<\/pre>\n<h4><a id=\"user-content-virtualization\" class=\"anchor\" aria-hidden=\"true\" href=\"#virtualization\"><\/a>Virtualization<\/h4>\n<p>Most operations in Spiral are virtualized, meaning they can be overridden.<\/p>\n<pre>inl f m =\n    open m\n    (1 + 2) * (3 + 4)\nf {}\n<\/pre>\n<pre>21L\n<\/pre>\n<p>If <code>+<\/code> and <code>*<\/code> overriden, the meaning of the expression can change completely.<\/p>\n<pre>inl f m =\n    open m\n    (1 + 2) * (3 + 4)\nf {\n    (+) = inl a b -&gt; string_format \"({0} + {1})\" (a,b)\n    (*) = inl a b -&gt; string_format \"({0} * {1})\" (a,b)\n    }\n<\/pre>\n<pre>\"((1 + 2) * (3 + 4))\"\n<\/pre>\n<p>Currently <code>if<\/code>,<code>&amp;&amp;<\/code>, <code>||<\/code> and <code>.<\/code> are keywords, but there is no reason why they couldn't be virtualized.<\/p>\n<p>The keyword <code>use<\/code> can be overridden by putting parentheses around it. This is how <code>use<\/code> is implemented in <code>Extern<\/code>.<\/p>\n<pre>inl (use) a b =\n    inl r = b a\n    FS.Method a .Dispose() unit\n    r\n<\/pre>\n<h3><a id=\"user-content-known-bugs\" class=\"anchor\" aria-hidden=\"true\" href=\"#known-bugs\"><\/a>Known Bugs<\/h3>\n<ol><li>\n<p>Passing <code>nan<\/code>s through join points causes the compiler to diverge. Since join points use structural equality for memoization and <code>nan = nan<\/code> is always <code>false<\/code>. That means that they cannot be compared and will always result in a new entry to their dictionary at join point boundaries. Note to numerical standards designers - don't do this again! Comparing <code>nan<\/code>s should be invalid, but returning <code>false<\/code> is not the answer.<\/p>\n<\/li>\n<li>\n<p><code>.NET<\/code> will happily marshal non <a href=\"https:\/\/docs.microsoft.com\/en-us\/dotnet\/framework\/interop\/blittable-and-non-blittable-types\" rel=\"nofollow\"><code>blittable<\/code><\/a> types past language boundaries corrupting memory and writing past the ends of the arrays. Spiral will give a type error at join points, but it is still possible to corrupt memory using transfer functions which aren't doing proper checking.<\/p>\n<\/li>\n<li>\n<p>This one happens rarely, but the way pattern matching is compiled can lead to surprises. Not only is Spiral overzealous in unboxing union types, but sometimes that same process can cause unintentional shadowing.<\/p>\n<\/li>\n<\/ol><pre>inl from = 1\nmatch {from = 2} with\n| {from=from'} -&gt; print_static {from from'}\n<\/pre>\n<pre>{from=lit 1i64; from'=lit 2i64}\n<\/pre>\n<p>This is what would be expected.<\/p>\n<pre>inl from = 1\nmatch {from = 2} with\n| {from near_to} -&gt; () \/\/ This branch binds the from in the module and shadows the one in scope.\n| {from=from'} -&gt; print_static {from from'}\n<\/pre>\n<pre>{from=lit 2i64; from'=lit 2i64}\n<\/pre>\n<p>Unfortunately there is no good way of fixing this. It is one of the consequences of how Spiral's patterns are compiled and there is no good alternative (as far as the author knows) to how it is done now.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:55+0000","updated_at":"2018-11-07T16:22:55+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":376,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/6266635?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":123,"title":"hyperbolic geometry in HyperRogue","url":"http:\/\/zenorogue.blogspot.com\/2012\/03\/hyperbolic-geometry-in-hyperbolic-rogue.html","content":"<b>NOTE: This post was originally published on March 18, 2012. It has been updated many times:\n2015 to reflect some new findings,\nSep 2016 to include the numerous new hyperbolic features,\nOct 2018 with even more features and links.<\/b>\n<p>The main goal of the HyperRogue project is to see how hyperbolic geometry can be used to achieve great game design (and it succeeds in this goal -- see <a href=\"https:\/\/www.rockpapershotgun.com\/2017\/09\/28\/hyperrogue-non-euclidean-roguelike\/\">e.g. this review on Rock Paper Shotgun<\/a>).\n(This is not the only goal, though -- HyperRogue is probably the most feature-rich engine for non-Euclidean geometry, that can be used for a plethora of visualization, research, and educational purposes; see <a href=\"http:\/\/www.roguetemple.com\/z\/hyper\/rogueviz.php\">RogueViz<\/a>.)\nThe purpose of this article is to explain what's the deal.\nIt is intended for mathematicians interested in hyperbolic geometry who would like a summary of HyperRogue, and for roguelike players who would like to know what hyperbolic geometry is -- some of them think that <a href=\"http:\/\/www.roguetemple.com\/z\/hyper.php\">HyperRogue<\/a> takes place on a 3D sphere, or that you will return to the same location if you move in a particular direction for a long time, which is not accurate at all, so I explain it here in more detail.<\/p><p>HyperRogue is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Roguelike\">roguelike<\/a> computer game. Roguelikes are a genre of computer games that have evolved separately from the mainstream -- even though they are commonly viewed as a subgenre of RPGs, they are in many senses closer to board games, like Chess where you have just one piece and the \"chessboard\" is randomly generated by the computer, than to other computer games. (Recently the elements of roguelikes, such as procedural generation and the challenge of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Permadeath\">permadeath<\/a>, are added to more and more computer games which do not really have much in common with true roguelikes, causing even more confusion.) In case of HyperRogue, the \"chessboard\" is an infinite hyperbolic plane. You can read more about it on <a href=\"http:\/\/en.wikipedia.org\/wiki\/Hyperbolic_geometry#Models_of_the_hyperbolic_plane\">Wikipedia<\/a>. This is not a sphere, but rather a hyperboloid (but also not in a normal space). The 3D object you see when you activate the \"eye distance\" option is this hyperboloid.<\/p><p>Hyperbolic geometry seems to be underused in art. M.C. Escher has been using it in some of his work (like <i>Circle Limit<\/i>). I have seen <a href=\"http:\/\/www.geometrygames.org\/HyperbolicGames\/\">Maze, Pool, and Sudoku on a hyperbolic plane<\/a>, but here the map is looped, and thus IMO less interesting than the real infinite world of HyperRogue. The original concept of Hyperbolic Rogue got quite popular despite being an extremely simple game, so I think this is a great area to explore.<\/p><p>Normally the game presents the world in the so called Poincar\u00e9 disc model (also used by the art mentioned above). The closer you are to the edge of the disc, the bigger distances are (five pixels close to the edge is a much bigger distance than five pixels in the center). On the Euclidean plane, the boundary of a circle of radius <i>r<\/i> is linear in <i>r<\/i>. On the hyperbolic plane it is exponential. This means that it is very hard to return to a place where you have already been in HyperRogue. You would have to return almost the same way.<\/p><p>Triangles in hyperbolic geometry have angles which sum to less than 180 degrees. For example, if you take two hexagons and one heptagons next to each other on the HyperRogue map, and construct a triangle with vertices in centers of these cells, the sum is (360\/6) + (360\/6) + (360\/7) = 171.4 degrees, so there is a \"defect\" of 60\/7 = 8.6 degrees. The defect is proportional to the area of the triangle. (Similar thing happens on a sphere: a triangle with a vertex on a pole and two on the equator has 90+90+90 = 180+90 degrees, which is \u03c0\/2 more than 180 degrees, and the total area of the sphere is 4\u03c0, which corresponds to eight of our triangles.) The visible in-game effect of this is that if you go from A to B, from B to C and from C back to A, the screen will (probably) appear rotated, making it easier to get lost.<\/p><p>This is the map structure I have found most suitable for a roguelike (or other similar grid-based game). Three hexagons would sum up to 180 degrees and we would get a plane. Three heptagons in each corner would be possible, but they would be too big (each of them would take the space of one heptagon in HyperRogue plus 1\/3 of each hexagon surrounding it).<\/p><p>Now how does this relate to the actual gameplay? As already mentioned, an important result of using the hyperbolic plane is that you almost never get into a place where you have been before. Some of the areas make other aspects of the hyperbolic geometry important, and some are just for fun.<\/p><p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/loem.png\">\n<img alt=\"Land of Eternal Motion\" src=\"http:\/\/roguetemple.com\/z\/hyper\/loem-thumb.png\" \/><\/a>\n<\/div>\n<b>Land of Eternal Motion::<\/b>When you are running away from monsters on a straight line in the Land of Eternal Motion, monsters cannot follow you on the same line, they have to use an equidistant curve instead. Since these curves are longer in hyperbolic geometry, it is quite easy to run away. \n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/hunting.png\">\n<img alt=\"Hunting Grounds\" src=\"http:\/\/roguetemple.com\/z\/hyper\/hunting-thumb.png\" \/><\/a>\n<\/div>\n<b>Hunting Grounds::<\/b>\nIn lands other than Land of Eternal Motion, when you are running away from a monster, it will have to chase you on exactly the same path as you, and if there are several monsters, they will have to make a queue. In most roguelikes, the basic strategy is that, when attacked by a group of monsters, you should run into a corridor, and fight them one by one. In HyperRogue, this effect can be achieved in an empty space.\n<br \/>\nAt low treasure counts, Hunting Grounds pits you against several dogs. Use the effect above to queue them up -- that would not be possible in an Euclidean hex grid (where they would pursue you forever), or in a spherical tesselation (where three dogs would eventually get you). \n<p>At higher treasure counts you also get ambushed by dogs, apparently from \"all\" directions. However, in the hyperbolic plane, there are more directions than it appears, so you can still escape!\n<\/p><p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/crossroad.png\">\n<img alt=\"Ultraparallel lines in the Crossroads\" src=\"http:\/\/roguetemple.com\/z\/hyper\/crossroad-thumb.png\" \/><\/a>\n<\/div>\n<b>Crossroads (and walls in general):<\/b> walls separating different areas are straight lines, and they do not cross. You cannot have that many non-crossing straight lines in a Euclidean world.<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/mirrorhall.png\">\n<img alt=\"Isometries in the Land of Mirrors\" src=\"http:\/\/roguetemple.com\/z\/hyper\/mirrorhall-thumb.png\" \/><\/a>\n<\/div>\n<b>Land of Mirrors:<\/b> mirages would be very nice to play in a Euclidean world (see <a href=\"http:\/\/caravelgames.com\/Articles\/Games.html\">DROD<\/a>'s mimics for something very similar), my point here was to check how this would work in a hyperbolic plane. The result seems to be that they are mostly annoying (they are supposed to stay parallel to you, but parallel lines diverge too quickly).\n<br \/>\nIn later versions, the Land of Mirrors has been replaced with <b>Hall of Mirrors<\/b> -- this adds mirrored walls that the Mimics can reflect off; this makes the land interesting in hyperbolic plane too :)\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/caves.png\">\n<img alt=\"Tree-like maze in the Living Caves\" src=\"http:\/\/roguetemple.com\/z\/hyper\/caves-thumb.png\" \/><\/a>\n<\/div>\n<b>Living Caves:<\/b> the cave has a tree-like structure: you go into a tunnel, sometimes you find forks, and you can be almost sure that both of the new tunnels do not end, and in fact, they quickly fork again. And loops are extremely unlikely. In an Euclidean world, this would not work. It would be at least a bit different.<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/lab.png\">\n<img alt=\"hyperbolic percolation in the Alchemist Lab\" src=\"http:\/\/roguetemple.com\/z\/hyper\/lab-thumb.png\" \/><\/a>\n<\/div>\n<b>Alchemist's Lab:<\/b> a bit like the Living Caves, the hyperbolic nature of the map means that you can usually get as far as you want in roughly any direction, no matter whether what color you are on. From the game <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hex_(board_game)\">Hex<\/a> we know that, in Euclidean geometry, if there is a red connection between top and bottom the screen, there is no blue connection between left and right. While this is still true in hyperbolic geometry, it does not feel like that, because we have many more directions there. This has ties to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Percolation_theory\">the percolation theory<\/a>.<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/icyland.png\">\n<img alt=\"Heat equation in the Icy Lands\" src=\"http:\/\/roguetemple.com\/z\/hyper\/icyland-thumb.png\" \/><\/a>\n<\/div>\n<b>Icy Lands:<\/b> that's a minor difference, but heat disperses more quickly here than it would on a Euclidean plane (actually the model in the original tech demo was a bit wrong).<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/shots\/camelot.png\">\n<img alt=\"hyperbolic circles in the Camelot\" src=\"http:\/\/roguetemple.com\/z\/shots\/camelot-thumb.png\" \/><\/a>\n<\/div>\n<b>Camelot:<\/b> In Camelot, you have to find the Holy Grail, which is in the center of the Round Table. The Round Table is a circle of radius 28. The number of cells in a circle of radius r is roughly 3 to the power of r\/2; in particular, there are 31659398 floor tiles inside the table, and 22860754 tiles of the table. Thus, when you start on the table and start moving in a completely random direction, you will run straight into the Holy Grail with probability only 1\/31659398 -- if you choose the direction based on the shape of the table, the probability is of course higher, but still, after a small number of steps you will usually end up back at the Table. You need to invent some strategy to precisely take you to the center.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/hive.png\">\n<img alt=\"army movement in the Hive\" src=\"http:\/\/roguetemple.com\/z\/shots\/hive_thumb.png\" \/><\/a>\n<\/div>\nThe <b>Hive<\/b> is an attempt at depicting how a war in hyperbolic plane would look. In our world, an army has a first long row of soldiers, then the second row, and so on. This way, the army moves quickly, and many soldiers can fight in parallel. However, this does not work in the world of HyperRogue: one column of soldiers may move on the straight line between points A and B is the shortest route, but if we have another column, it will have to move on an equidistant curve, which will be much longer. Therefore, if the army wants to move quickly, it will have just a small number of columns.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/shots\/temple.png\">\n<img alt=\"horocycles in the Temple of Cthulhu\" src=\"http:\/\/roguetemple.com\/z\/shots\/temple-thumb.png\" \/><\/a>\n<\/div>\n<b>Horocycles:<\/b> In the Hive and Camelot, you could see how big circles look in the hyperbolic plane. While in the Euclidean world, big circles resemble straight lines more and more, the situation in the world of HyperRogue is different -- they always look curved. If you are standing on a circle, it is easy to tell that this is not a straight line -- but locally, a fragment of a circle of radius, say, 10 tiles will look almost exactly the same as a fragment of a circle of radius, say, 100 tiles. A horocycle is what we get in the limit -- something like a circle of infinite radius. The <b>Temple of Cthulhu<\/b> is an infinite sequence of horocycles, where you have to go deeper and deeper to get more points, and <b>Whirlpool<\/b> is a <a href=\"http:\/\/roguetemple.com\/z\/hyper\/whirlpool.png\">horocycle which draws you toward its infinitely distant center<\/a>. \n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/Caribbean-sea.png\">\n<img alt=\"unmarked horocycles in the Caribbean\" src=\"http:\/\/roguetemple.com\/z\/hyper\/Caribbean-sea-thumb.png\" \/><\/a>\n<\/div>\n<b>Caribbean<\/b> is unmarked, and shows that, just as in Camelot, it may be hard to find the direction towards the center of the horocycle -- you have to use compasses to lead you there.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/ocean.png\">\n<img alt=\"equidistants in the Ocean\" src=\"http:\/\/roguetemple.com\/z\/hyper\/ocean-thumb.png\" \/><\/a>\n<\/div>\nA Great Wall is a straight line, and the coast is an equidistant curve at some distance from this great wall (the distance changes with the tides). As you can see, it is hard to tell the difference between a large circle, horocycle, or a distant equidistant -- so if you are in the Ocean, it is not immediate to tell whether the waves you see are because of the nearby coast, or because of the nearby Whirlpool. See <a href=\"http:\/\/www.roguetemple.com\/z\/hyper\/curves.php\">this article by Fulgur14<\/a> to read more about circles, horocycles, and equidistant curves.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/haunted.png\">\n<img alt=\"get lost in the Haunted Woods\" src=\"http:\/\/roguetemple.com\/z\/hyper\/haunted-thumb.png\" \/><\/a>\n<\/div>\nThe <b>Haunted Woods<\/b> is a land from where there is just one exit -- it is bounded by a single curve. For further confusion, this curve may look like a circle or horocycle, but it is not -- it is an equidistant curve -- while moving randomly inside a horocycle quickly gets you to its edge, in the Haunted Woods, if you cross the straight line and go randomly on the other side, you will likely be lost forever in the Haunted Woods, as you have to return to the location where you have been before, and this is virtually impossible, due to the exponential growth. The <b>Yendor Quest<\/b> is based on the same principle -- you have to go 100 steps to find the Key, then return. There is a marker which shows your way to the Key, but you have to return on your own, which is virtually impossible based only on luck or memory.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/princess.png\">\n<img alt=\"rescue the Princess!\" src=\"http:\/\/roguetemple.com\/z\/hyper\/princess-thumb.png\" \/><\/a>\n<\/div>\nIn the <b>Princess Challenge<\/b>, a mouse asks you to reach the Princess, locked somewhere in the Palace. She is 100 steps away, so you would not find here without help -- luckily, the mouse will lead you. The screenshot above shots the cell where the Princess is locked -- you have to press the green button in bottom left to open the doors; when you first see the princess, you are typically on the other side, and have to go around all these closed doors. This exhibits interesting properties of hyperbolic mazes -- there are obstacles on the way, and sometimes you will have to do significant detours to go around them, but if you know what you are doing, you will eventually reach the Princess with probability of (almost) 1. This is related to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Branching_process\">branching processes<\/a>.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/crossroads3.png\">\n<img alt=\"equidistants in the Ocean\" src=\"http:\/\/roguetemple.com\/z\/hyper\/crossroads3-thumb.png\" \/><\/a>\n<\/div>\nThe <b>Crossroads III<\/b>, as well as <b>the Elemental Planes<\/b> later in the game, show straight lines which cross at straight angles.\n<div style=\"text-align: center;\">\n<p><a href=\"http:\/\/roguetemple.com\/z\/hyper\/ivorytower.png\">\n<img alt=\"equidistants in the Ivory Tower\" src=\"http:\/\/roguetemple.com\/z\/hyper\/ivorytower-thumb.png\" \/><\/a>\n<\/p><\/div>\nThe <b>Ivory Tower<\/b> shows you how gravity could work in the hyperbolic world. Gravity is directed towards a straight line, thus making the horizontal levels into equidistant curves. Interestingly, while in the most of the game it is very hard to return to the location you have been to, in these lands, you simply go with the gravity to exit, and you turn up at almost exactly the same location you started. This is because, as usual, levels grow exponentially: moving 243 cells on the level 20 corresponds to moving just 243\/3<sup>(20-10)\/2<\/sup> = 1 cell on level 10. \n<div style=\"text-align: center;\">\n<p><a href=\"http:\/\/roguetemple.com\/z\/hyper\/yendorian.png\">\n<img alt=\"equidistants in the Yendorian Forest\" src=\"http:\/\/roguetemple.com\/z\/hyper\/yendorian-thumb.png\" \/><\/a><\/p><\/div>\nThe <b>Yendorian Forest<\/b> is another land in gravity. We already had straight lines that cross, and that are ultraparallel (diverging). Branches of the same tree in the Yendorian Forest are straight lines that diverge when you go upwards, but get closer and closer when you go towards the base level.\n<div style=\"text-align: center;\">\n<p><a href=\"http:\/\/roguetemple.com\/z\/hyper\/wine.png\">\n<img alt=\"regular pattern in the Vineyard\" src=\"http:\/\/roguetemple.com\/z\/shots\/wine_thumb.png\" \/><\/a>\n<\/p><\/div>\nThere are regular patterns in Euclidean plane, which generally look like grids. We also have regular patterns in the hyperbolic plane, but they look more like infinite trees. Such regular patterns appear in the <b>Vineyard<\/b>, <b>Land of Power<\/b>, <b>Zebra<\/b>, <b>Windy Plains<\/b>, <b>Emerald Mine<\/b>, and <b>Palace<\/b>. The picture above shows the Vineyard; see the <a href=\"http:\/\/www.roguetemple.com\/z\/hyper\/gallery.php\">Gallery of Lands<\/a> for the screenshots of the remaining regular lands.\n<div style=\"text-align: center;\">\n<p><a href=\"http:\/\/roguetemple.com\/z\/hyper\/burial.png\">\n<img alt=\"curvature in the Burial Grounds\" src=\"http:\/\/roguetemple.com\/z\/hyper\/burial-thumb.png\" \/><\/a>\n<\/p><\/div>\nIn the <b>Burial Grounds<\/b>, you have to fight enemies and excavate treasures, using a magical energy sword. This sword stays at the same relative angle to you when you move. However, after you go in a closed loop, the sword will usually be rotated, because of how angles work in the hyperbolic plane! This land obviously makes no sense in the Euclidean mode.\n<div style=\"text-align: center;\">\n<p><a href=\"http:\/\/roguetemple.com\/z\/hyper\/galapagos.png\">\n<img alt=\"all the species of turtles in Gal\u00e1pagos\" src=\"http:\/\/roguetemple.com\/z\/hyper\/galapagos-thumb.png\" \/><\/a>\n<\/p><\/div>\nIn <b>Gal\u00e1pagos<\/b>, each tortoise has 21 binary properties, thus leading to 2<sup>21<\/sup> = 2097152 subspecies. Tortoises at the given spot are usually exactly the same, but as you travel, the environmental properties change smoothly, and so do the properties of Tortoises. You have found a (random) baby tortoise, and you have to find a tortoise which exactly matches the baby. This would not be feasible in the Euclidean plane -- even if you exactly knew where the correct tortoise are, you would still have to go 1000 steps to reach it, but since you don't know, you would have to wander blindly. In the hyperbolic plane, the correct tortoise is always actually quite close, because of the exponential growth; furthermore, the simple method of always going towards the direction where more properties agree works. The coloring of tiles in the Galapagos (based on the number of factors which agree), as well as the height function used to generate the <b>Dragon Chasms<\/b>, is related to a generalization of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Wiener_process\">the Wiener process<\/a> to the hyperbolic plane domain -- there are similar generalizations to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Diamond-square_algorithm#Midpoint_displacement_algorithm\">Euclidean plane domains<\/a>, but for hyperbolic plane this is mathematically much more elegant.\n<div style=\"text-align: center;\">\n<p><a href=\"http:\/\/roguetemple.com\/z\/hyper\/loem-gp20\">\n<img alt=\"Land of Eternal Motion in GP(2,0)\" src=\"http:\/\/roguetemple.com\/z\/hyper\/loem-gp20-thumb.png\" \/><\/a>\n<\/p><\/div>\nYou can also change the geometry in HyperRogue. The first alternate geometry was, of course, the Euclidean mode -- see <a href=\"http:\/\/zenorogue.blogspot.com\/2015\/02\/euclidean-rogue.html\">this old post about it<\/a>. In 2018 the big part of work on HyperRogue was adding more and more geometries to experiment with. Disable bitruncation for stronger hyperbolic effects; apply the Goldberg-Coxeter construction to weaken them (which may make you lose to dogs in LoEM or Hunting Grounds -- even though the world is globally hyperbolic, there are parts which are flat); play on the sphere; use irregular, or mostly arbitrary Archimedean tilings; and so on. See <a href=\"http:\/\/www.roguetemple.com\/z\/hyper\/geoms.php\">this article<\/a> for all the possibilities.\n<p><\/p><div style=\"text-align: center;\">\n<a href=\"http:\/\/roguetemple.com\/z\/hyper\/hypersian-rug.png\">\n<img alt=\"Hypersian rug model\" src=\"http:\/\/roguetemple.com\/z\/hyper\/hypersian-rug-thumb.png\" \/><\/a>\n<\/div>\nThe Poincar\u00e9 disc model is just one of many possible models (\"maps\") that can be used to display the hyperbolic plane. HyperRogue also lets the players to play with other models. In the band model, you can see that your route through the game is almost a straight line; in the half-plane model, horocycles are horizontal and branches of Yendorian Forest are (almost) vertical; in the Klein model, straight lines look straight; the Hypersian rug model lets you play on something similar to the <a href=\"https:\/\/www.google.com\/search?q=hyperbolic+crochet&amp;tbm=isch\">hyperbolic crochet<\/a>, and the paper model creator even lets you to print it. See <a href=\"http:\/\/www.roguetemple.com\/z\/hyper\/model.php\">this article<\/a> for more details.\n<div>\n<\/div>","created_at":"2018-11-07T16:22:55+0000","updated_at":"2018-11-07T16:22:55+0000","published_at":"2012-03-18T00:00:00+0000","published_by":["Zeno Rogue"],"annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":14,"domain_name":"zenorogue.blogspot.com","preview_picture":"https:\/\/lh5.googleusercontent.com\/proxy\/F2V_sRZuACz7qdeRJmydKcqBNjuoz98R6IwTOlpvWtHkW-GV0niaJz7g1MsiQpFCmI2xJ3qGCTRs4PDq4TG9QY0=w1200-h630-p-k-no-nu","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":124,"title":"The James Mickens Collection \u2013 Daniel Compton","url":"https:\/\/danielcompton.net\/2015\/06\/11\/james-mickens-collection","content":"The James Mickens Collection \u2013 Daniel Compton\n\n<section id=\"posts\"><article><header>\n<p>\n\u2022\n<a class=\"permalink\" title=\"Permalink\" href=\"https:\/\/danielcompton.net\/2015\/06\/11\/james-mickens-collection\">\u221e<\/a>\n<\/p>\n<\/header><p><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/\">James Mickens<\/a> works at Microsoft Research. Amongst his more serious work, he has written some hilarious papers for Usenix, and given some funny talks. I\u2019ve collected a selection of my favourite quotes here, but you should really read and watch them all from start to finish.<\/p><p><strong>Videos:<\/strong><\/p><ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=D5xh0ZIEUOE\">Life is Terrible: Let\u2019s Talk About the Web<\/a> - The craziness that is JavaScript<\/li>\n<li><a href=\"https:\/\/vimeo.com\/95066828\">Computers are a Sadness, I am the Cure<\/a> - My all time favourite James Mickens comedy, featuring NoSQL Bane.<\/li>\n<\/ul><p><strong>Usenix papers:<\/strong><\/p><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/thenightwatch.pdf\">The Night Watch<\/a><\/h3><blockquote>\n<p>When you debug systems code, there are no high-level debates about font choices and the best kind of turquoise, because this is the Old Testament, an angry and monochromatic world, and it doesn\u2019t matter whether your Arial is Bold or Condensed when people are covered in boils and pestilence and Egyptian pharaoh oppression. HCI people discover bugs by receiving a concerned email from their therapist. Systems people discover bugs by waking up and discovering that their first-born children are missing and \u201cETIMEDOUT \u201d has been written in blood on the wall.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/thesaddestmoment.pdf\">The Saddest Moment<\/a><\/h3><blockquote>\n<p>Listen, regardless of which Byzantine fault tolerance protocol you pick, Twitter will still have fewer than two nines of availability. As it turns out, Ted the Poorly Paid Datacenter Operator will not send 15 cryptographically signed messages before he accidentally spills coffee on the air conditioning unit and then overwrites your tape backups with bootleg recordings of Nickelback. Ted will just do these things and then go home, because that\u2019s what Ted does. His extensive home collection of \u201cThundercats\u201d cartoons will not watch itself. Ted is needed, and Ted will heed the call of duty.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/nestofhornets.pdf\">Mobile Computing Research Is a Hornet\u2019s Nest of Deception and Chicanery<\/a><\/h3><blockquote>\n<p>So, your phone just starts doing stuff, all the stuff that it knows how to do, and it\u2019s just going nuts, and your apps are closing and opening and talking to the cloud and configuring themselves in unnatural ways, and your phone starts vibrating and rumbling with its little rumble pack, and it will gently sing like a tiny hummingbird of hate, and you\u2019ll look at the touchscreen, and you\u2019ll see that things are happening, my god, there are so many happenings, and you\u2019ll try to flip the phone over and take out the battery, because now you just want to kill it and move to Kansas and start over, but the back panel of the phone is attached by a molecule-sized screw that requires a special type of screwdriver that only Merlin possesses, and Merlin isn\u2019t nearby, and your phone is still rumbling, and by this point, you can understand the rumble, it\u2019s a twin language that you and your phone invented, and the phone is rumbling, and it\u2019s saying that it\u2019s far from done, that it has so much more that it wants to do, that there are so many of your frenemies that it wants to \u201caccidentally\u201d call and then leave you to deal with the social ramifications, and your phone, it buzzes, and you think that you see it smiling, and you begin to realize that land-line telephones were actually a pretty good idea.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/thisworldofours.pdf\">This World of Ours<\/a><\/h3><blockquote>\n<p>My point is that security people need to get their priorities straight. The \u201cthreat model\u201d section of a security paper resembles the script for a telenovela that was written by a paranoid schizophrenic: there are elaborate narratives and grand conspiracy theories, and there are heroes and villains with fantastic (yet oddly constrained) powers that necessitate a grinding battle of emotional and technical attrition. In the real world, threat models are much simpler. Basically, you\u2019re either dealing with Mossad or not-Mossad. If your adversary is not-Mossad, then you\u2019ll probably be fine if you pick a good password and don\u2019t respond to emails from <a href=\"https:\/\/danielcompton.net\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"0d4e65486c5d687e795d4c64435d643c3c7e4d7b647f787e206f6c7e666879236f6477237f7823\">[email\u00a0protected]<\/a> If your adversary is the Mossad, YOU\u2019RE GONNA DIE AND THERE\u2019S NOTHING THAT YOU CAN DO ABOUT IT. The Mossad is not intimidated by the fact that you employ https:\/\/. If the Mossad wants your data, they\u2019re going to use a drone to replace your cellphone with a piece of uranium that\u2019s shaped like a cellphone, and when you die of tumors filled with tumors, they\u2019re going to hold a press conference and say \u201cIt wasn\u2019t us\u201d as they wear t-shirts that say \u201cIT WAS DEFINITELY US\u201d.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/ToWashItAllAway.pdf\">To Wash It All Away<\/a><\/h3><blockquote>\n<p>The fourth and seventh errors represent uncaught JavaScript exceptions. In a rational universe, a single uncaught exception would terminate a program, and if a program continued to execute after throwing such an exception, we would know that Ragnarok is here and Odin is not happy. In the browser world, ignoring uncaught exceptions is called \u201cWednesday, and all days not called \u2018Wednesday.\u2019\u201d The JavaScript event loop is quite impervious to conventional notions of software reliability, so if an event handler throws an exception, the event loop will literally pretend like nothing happened and keep running. This ludicrous momentum continues even if, in the case of the seventh error, the Web page tries to call init() on an object that has no init() method. You should feel uncomfortable that a Web page can disagree with itself about the existence of initialization routines, but the page is still allowed to do things with things. Such a dramatic mismatch of expectations would be unacceptable in any other context. You would be sad if you went to the hospital to have your appendix removed, and the surgeon opened you up, and she said, \u201cI DIDN\u2019T EXPECT YOUR LIVER TO HAVE GILLS,\u201d and then she proceeded with her original surgical plan, despite the fact that you\u2019re apparently a mer-person. Being a mer-person should have non-ignorable ramifications in the material universe. Similarly, if a Web page thinks than an object should be initialized, but the object has no initialization method, the browser shouldn\u2019t laugh about it and then proceed under the assumption that the rest of the page is agnostic about whether its objects are composed of folly.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/theslowwinter.pdf\">The Slow Winter<\/a><\/h3><blockquote>\n<p>Perhaps the processor could run multiple copies of each program, comparing the results to detect errors? Perhaps a new video codec could tolerate persistently hateful levels of hardware error? All of these techniques could be implemented. However, John slowly realized that these solutions were just things that he could do, and inventing \u201ca thing that you could do\u201d is a low bar for human achievement. If I were walking past your house and I saw that it was on fire, I <em>could<\/em> try to put out the fire by finding a dingo and then teaching it how to speak Spanish. That\u2019s certainly a thing that I could do. However, when you arrived at your erstwhile house and found a pile of heirloom ashes, me, and a dingo with a chewed-up Rosetta Stone box, you would be less than pleased, despite my protestations that negative scientific results are useful and I had just proven that Spanish-illiterate dingoes cannot extinguish fires using mind power.<\/p>\n<\/blockquote><\/article><\/section>","created_at":"2018-11-07T16:22:55+0000","updated_at":"2018-11-07T16:22:55+0000","published_at":"2015-06-11T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"danielcompton.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":125,"title":"The James Mickens Collection \u2013 Daniel Compton","url":"https:\/\/danielcompton.net\/2015\/06\/11\/james-mickens-collection","content":"The James Mickens Collection \u2013 Daniel Compton\n\n<section><article><header>\n<p>\n\u2022\n<a class=\"permalink\" title=\"Permalink\" href=\"https:\/\/danielcompton.net\/2015\/06\/11\/james-mickens-collection\">\u221e<\/a>\n<\/p>\n<\/header><p><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/\">James Mickens<\/a> works at Microsoft Research. Amongst his more serious work, he has written some hilarious papers for Usenix, and given some funny talks. I\u2019ve collected a selection of my favourite quotes here, but you should really read and watch them all from start to finish.<\/p><p><strong>Videos:<\/strong><\/p><ul><li><a href=\"https:\/\/www.youtube.com\/watch?v=D5xh0ZIEUOE\">Life is Terrible: Let\u2019s Talk About the Web<\/a> - The craziness that is JavaScript<\/li>\n<li><a href=\"https:\/\/vimeo.com\/95066828\">Computers are a Sadness, I am the Cure<\/a> - My all time favourite James Mickens comedy, featuring NoSQL Bane.<\/li>\n<\/ul><p><strong>Usenix papers:<\/strong><\/p><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/thenightwatch.pdf\">The Night Watch<\/a><\/h3><blockquote>\n<p>When you debug systems code, there are no high-level debates about font choices and the best kind of turquoise, because this is the Old Testament, an angry and monochromatic world, and it doesn\u2019t matter whether your Arial is Bold or Condensed when people are covered in boils and pestilence and Egyptian pharaoh oppression. HCI people discover bugs by receiving a concerned email from their therapist. Systems people discover bugs by waking up and discovering that their first-born children are missing and \u201cETIMEDOUT \u201d has been written in blood on the wall.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/thesaddestmoment.pdf\">The Saddest Moment<\/a><\/h3><blockquote>\n<p>Listen, regardless of which Byzantine fault tolerance protocol you pick, Twitter will still have fewer than two nines of availability. As it turns out, Ted the Poorly Paid Datacenter Operator will not send 15 cryptographically signed messages before he accidentally spills coffee on the air conditioning unit and then overwrites your tape backups with bootleg recordings of Nickelback. Ted will just do these things and then go home, because that\u2019s what Ted does. His extensive home collection of \u201cThundercats\u201d cartoons will not watch itself. Ted is needed, and Ted will heed the call of duty.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/nestofhornets.pdf\">Mobile Computing Research Is a Hornet\u2019s Nest of Deception and Chicanery<\/a><\/h3><blockquote>\n<p>So, your phone just starts doing stuff, all the stuff that it knows how to do, and it\u2019s just going nuts, and your apps are closing and opening and talking to the cloud and configuring themselves in unnatural ways, and your phone starts vibrating and rumbling with its little rumble pack, and it will gently sing like a tiny hummingbird of hate, and you\u2019ll look at the touchscreen, and you\u2019ll see that things are happening, my god, there are so many happenings, and you\u2019ll try to flip the phone over and take out the battery, because now you just want to kill it and move to Kansas and start over, but the back panel of the phone is attached by a molecule-sized screw that requires a special type of screwdriver that only Merlin possesses, and Merlin isn\u2019t nearby, and your phone is still rumbling, and by this point, you can understand the rumble, it\u2019s a twin language that you and your phone invented, and the phone is rumbling, and it\u2019s saying that it\u2019s far from done, that it has so much more that it wants to do, that there are so many of your frenemies that it wants to \u201caccidentally\u201d call and then leave you to deal with the social ramifications, and your phone, it buzzes, and you think that you see it smiling, and you begin to realize that land-line telephones were actually a pretty good idea.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/thisworldofours.pdf\">This World of Ours<\/a><\/h3><blockquote>\n<p>My point is that security people need to get their priorities straight. The \u201cthreat model\u201d section of a security paper resembles the script for a telenovela that was written by a paranoid schizophrenic: there are elaborate narratives and grand conspiracy theories, and there are heroes and villains with fantastic (yet oddly constrained) powers that necessitate a grinding battle of emotional and technical attrition. In the real world, threat models are much simpler. Basically, you\u2019re either dealing with Mossad or not-Mossad. If your adversary is not-Mossad, then you\u2019ll probably be fine if you pick a good password and don\u2019t respond to emails from <a href=\"https:\/\/danielcompton.net\/cdn-cgi\/l\/email-protection\" class=\"__cf_email__\" data-cfemail=\"3b78537e5a6b5e484f6b7a52756b520a0a487b4d52494e4816595a48505e4f1559524115494e15\">[email\u00a0protected]<\/a> If your adversary is the Mossad, YOU\u2019RE GONNA DIE AND THERE\u2019S NOTHING THAT YOU CAN DO ABOUT IT. The Mossad is not intimidated by the fact that you employ https:\/\/. If the Mossad wants your data, they\u2019re going to use a drone to replace your cellphone with a piece of uranium that\u2019s shaped like a cellphone, and when you die of tumors filled with tumors, they\u2019re going to hold a press conference and say \u201cIt wasn\u2019t us\u201d as they wear t-shirts that say \u201cIT WAS DEFINITELY US\u201d.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/ToWashItAllAway.pdf\">To Wash It All Away<\/a><\/h3><blockquote>\n<p>The fourth and seventh errors represent uncaught JavaScript exceptions. In a rational universe, a single uncaught exception would terminate a program, and if a program continued to execute after throwing such an exception, we would know that Ragnarok is here and Odin is not happy. In the browser world, ignoring uncaught exceptions is called \u201cWednesday, and all days not called \u2018Wednesday.\u2019\u201d The JavaScript event loop is quite impervious to conventional notions of software reliability, so if an event handler throws an exception, the event loop will literally pretend like nothing happened and keep running. This ludicrous momentum continues even if, in the case of the seventh error, the Web page tries to call init() on an object that has no init() method. You should feel uncomfortable that a Web page can disagree with itself about the existence of initialization routines, but the page is still allowed to do things with things. Such a dramatic mismatch of expectations would be unacceptable in any other context. You would be sad if you went to the hospital to have your appendix removed, and the surgeon opened you up, and she said, \u201cI DIDN\u2019T EXPECT YOUR LIVER TO HAVE GILLS,\u201d and then she proceeded with her original surgical plan, despite the fact that you\u2019re apparently a mer-person. Being a mer-person should have non-ignorable ramifications in the material universe. Similarly, if a Web page thinks than an object should be initialized, but the object has no initialization method, the browser shouldn\u2019t laugh about it and then proceed under the assumption that the rest of the page is agnostic about whether its objects are composed of folly.<\/p>\n<\/blockquote><h3><a href=\"http:\/\/research.microsoft.com\/en-us\/people\/mickens\/theslowwinter.pdf\">The Slow Winter<\/a><\/h3><blockquote>\n<p>Perhaps the processor could run multiple copies of each program, comparing the results to detect errors? Perhaps a new video codec could tolerate persistently hateful levels of hardware error? All of these techniques could be implemented. However, John slowly realized that these solutions were just things that he could do, and inventing \u201ca thing that you could do\u201d is a low bar for human achievement. If I were walking past your house and I saw that it was on fire, I <em>could<\/em> try to put out the fire by finding a dingo and then teaching it how to speak Spanish. That\u2019s certainly a thing that I could do. However, when you arrived at your erstwhile house and found a pile of heirloom ashes, me, and a dingo with a chewed-up Rosetta Stone box, you would be less than pleased, despite my protestations that negative scientific results are useful and I had just proven that Spanish-illiterate dingoes cannot extinguish fires using mind power.<\/p>\n<\/blockquote><\/article><\/section>","created_at":"2018-11-07T16:22:55+0000","updated_at":"2018-11-07T16:22:55+0000","published_at":"2015-06-11T00:00:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"danielcompton.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":121,"title":"Talks by Alan Kay - Viewpoints Intelligent Archive","url":"https:\/\/tinlizzie.org\/IA\/index.php\/Talks_by_Alan_Kay","content":"<p>From Viewpoints Intelligent Archive<\/p><p>Jump to:\t\t\t\t\t<a href=\"#mw-head\">navigation<\/a>, \t\t\t\t\t<a href=\"#p-search\">search<\/a><\/p><p><\/p><table class=\"wikitable sortable\"><tr><th>Title<\/th>\n<th>Date<\/th>\n<th>Venue<\/th>\n<th>Type<\/th>\n<th>VPRI Record<\/th>\n<th>Status\n<\/th><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Computer_Applications:_A_Dynamic_Medium_for_Creative_Thought_by_Alan_Kay_(1972)\" title=\"Computer Applications: A Dynamic Medium for Creative Thought by Alan Kay (1972)\">Computer Applications: A Dynamic Medium for Creative Thought by Alan Kay<\/a><\/td>\n<td>000000001972-01-01-00001972<\/td>\n<td>Talk<\/td>\n<td>VPRI-0893<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_TED_(1982)\" title=\"Alan Kay at TED (1982)\">Alan Kay at TED 1982<\/a><\/td>\n<td>000000001982-01-01-00001982<\/td>\n<td>Talk<\/td>\n<td>VPRI-0009<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Joint_House_Congressional_Committee_Hearing_(1982)\" title=\"Joint House Congressional Committee Hearing (1982)\">Joint House Congressional Committee Hearing<\/a><\/td>\n<td>000000001982-07-15-0000July 15, 1982<\/td>\n<td>Washington DC<\/td>\n<td>Hearing<\/td>\n<td>VPRI-0530<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Interview_(1982)\" title=\"Alan Kay Interview (1982)\">Alan Kay Interview<\/a><\/td>\n<td>000000001982-11-01-0000November 1982<\/td>\n<td>Interview<\/td>\n<td>VPRI-0164<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/LINC_Twentieth_Anniversary_Symposium_(1983)\" title=\"LINC Twentieth Anniversary Symposium (1983)\">LINC Twentieth Anniversary Symposium<\/a><\/td>\n<td>000000001983-11-30-0000November 30, 1983<\/td>\n<td>Talk<\/td>\n<td>VPRI-0227<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Processors_in_Programming_Languages\" title=\"Processors in Programming Languages\">Processors in Programming Languages<\/a><\/td>\n<td>1983} (approx.)<\/td>\n<td>Talk<\/td>\n<td>VPRI-0234<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_MacFest_(1985)\" title=\"Alan Kay at MacFest (1985)\">Alan Kay at MacFest<\/a><\/td>\n<td>000000001985-05-20-0000May 20, 1985<\/td>\n<td>Talk<\/td>\n<td>VPRI-0012<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/The_Dynabook_--_Past,_Present_and_Future_(1986)\" title=\"The Dynabook -- Past, Present and Future (1986)\">The Dynabook -- Past, Present and Future<\/a><\/td>\n<td>000000001986-01-09-0000January 9, 1986<\/td>\n<td>Palo Alto, CA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0671<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_talk_at_Coors_Television_(late_1986_or_early_1987)\" title=\"Alan Kay talk at Coors Television (late 1986 or early 1987)\">A talk at Coors Television by Alan Kay<\/a><\/td>\n<td>000000001986-12-01-0000December 1986 (late 1986 or early 1987)<\/td>\n<td>Talk<\/td>\n<td>VPRI-0035<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/A_talk_at_User_Group_University_(1987)\" title=\"A talk at User Group University (1987)\">User Group University talk<\/a><\/td>\n<td>000000001987-01-01-00001987<\/td>\n<td>Talk<\/td>\n<td>VPRI-0102<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Pacific_Northwest_Bell_Technology_Forum\" title=\"Alan Kay at Pacific Northwest Bell Technology Forum\">Alan Kay at Pacific Northwest Bell Technology Forum<\/a><\/td>\n<td>000000001987-10-03-0000October 3, 1987<\/td>\n<td>Talk<\/td>\n<td>VPRI-0141<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Doing_with_Images_Makes_Symbols_(1987)\" title=\"Doing with Images Makes Symbols (1987)\">Doing with Images Makes Symbols<\/a><\/td>\n<td>000000001987-10-27-0000October 27, 1987<\/td>\n<td>Talk<\/td>\n<td>VPRI-0394<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/A_Vision_of_New_Age_with_Alan_Kay_(1988)\" title=\"A Vision of New Age with Alan Kay (1988)\">A Vision of New Age with Alan Kay<\/a><\/td>\n<td>000000001988-01-01-00001988<\/td>\n<td>California, USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0224<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Future_School_Now_(1988)\" title=\"Future School Now (1988)\">Future School Now (1988)<\/a><\/td>\n<td>000000001988-05-05-0000May 5, 1988<\/td>\n<td>Interview<\/td>\n<td>VPRI-0895<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/A_Conversation_with_Alan_Kay_(1989)\" title=\"A Conversation with Alan Kay (1989)\">A Conversation with Alan Kay<\/a><\/td>\n<td>000000001989-06-28-0000June 28, 1989<\/td>\n<td> <\/td>\n<td>Interview<\/td>\n<td>VPRI-0470<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/BCS_Mac_Awards_Banquet_and_MacTech_Wizards_Meeting_(1989)\" title=\"BCS Mac Awards Banquet and MacTech Wizards Meeting (1989)\">BCS Mac Awards Banquet and MacTech Wizards Meeting<\/a><\/td>\n<td>000000001989-08-10-0000August 10, 1989<\/td>\n<td>Talk<\/td>\n<td>VPRI-0357<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Seminar_with_Alan_Kay_on_Object_Oriented_Programming\" title=\"Seminar with Alan Kay on Object Oriented Programming\">Seminar with Alan Kay on Object Oriented Programming<\/a><\/td>\n<td>000000001989-01-01-00001989 (late 80's)<\/td>\n<td> <\/td>\n<td>Talk<\/td>\n<td>VPRI-0246<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_talk_at_BSO_MultiMedia_Group\" title=\"Alan Kay talk at BSO MultiMedia Group\">Alan Kay talk at BSO MultiMedia Group<\/a><\/td>\n<td>000000001989-01-01-00001989<\/td>\n<td>Holland<\/td>\n<td>Talk<\/td>\n<td>VPRI-0351<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Directions_in_Object-Oriented_Programming_From_Actions_to_Agents\" title=\"Directions in Object-Oriented Programming From Actions to Agents\">Directions in Object-Oriented Programming From Actions to Agents<\/a><\/td>\n<td>1989} (approx. maybe 1988)<\/td>\n<td>Talk<\/td>\n<td>VPRI-0219<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_MIT_Media_Lab_5th_Anniversary_Event_(1990)\" title=\"Alan Kay at MIT Media Lab 5th Anniversary Event (1990)\">Alan Kay at MIT Media Lab 5th Anniversary Event<\/a><\/td>\n<td>000000001990-10-01-0000October 1, 1990<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0316<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Portable_Portrait:_ALAN_KAY_(1990)\" title=\"Portable Portrait: ALAN KAY (1990)\">Portable Portrait: ALAN KAY (1990)<\/a><\/td>\n<td>000000001990-01-01-00001990<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Machine_That_Changed_The_World,_The;_Interview_with_Alan_Kay,_(1990)\" title=\"Machine That Changed The World, The; Interview with Alan Kay, (1990)\">Machine That Changed The World, The; Interview with Alan Kay, (1990)<\/a><\/td>\n<td>000000001990-01-01-00001990<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/A_concept_video_and_Alan_Kay_talk_at_WWDC_%2790_(VPRI_358)\" title=\"A concept video and Alan Kay talk at WWDC '90 (VPRI 358)\">A concept video and Alan Kay talk at WWDC '90 (VPRI 358)<\/a><\/td>\n<td>000000001990-01-01-00001990<\/td>\n<td>San Jose, CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0358<\/td>\n<td>Annotated\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Interview_(1990)\" title=\"Alan Kay Interview (1990)\">Alan Kay Interview (1990)<\/a><\/td>\n<td>000000001990-01-01-00001990<\/td>\n<td>Boston, MA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/An_Interview_with_Dr._Alan_Kay_(1991)\" title=\"An Interview with Dr. Alan Kay (1991)\">An Interview with Dr. Alan Kay<\/a><\/td>\n<td>000000001991-01-01-00001991<\/td>\n<td>Interview<\/td>\n<td>VPRI-0146<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_IRC_Conference_(1991)\" title=\"Alan Kay at IRC Conference (1991)\">Alan Kay at IRC Conference<\/a><\/td>\n<td>000000001991-02-04-0000February 4, 1991<\/td>\n<td>Talk<\/td>\n<td>VPRI-0320<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/End-User_Programming_by_Alan_Kay_(1991)\" title=\"End-User Programming by Alan Kay (1991)\">End-User Programming<\/a><\/td>\n<td>000000001991-03-21-0000March 21, 1991<\/td>\n<td>Talk<\/td>\n<td>VPRI-0147<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_the_Getty_Conference_(1991)\" title=\"Alan Kay at the Getty Conference (1991)\">Alan Kay at the Getty Conference<\/a><\/td>\n<td>000000001991-04-26-0000April 26, 1991<\/td>\n<td>Talk<\/td>\n<td>VPRI-0142, 0205, 0283<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Defining_Perspective,_INEL_Computing_Symposium_(1991)\" title=\"Defining Perspective, INEL Computing Symposium (1991)\">Defining Perspective, INEL Computing Symposium<\/a><\/td>\n<td>000000001991-09-10-0000September 10, 1991<\/td>\n<td>Idaho<\/td>\n<td>Talk<\/td>\n<td>VPRI-0189<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Kids_%26_COMPUTERS:_What%27s_a_parent_to_do%3F_Panel_Session_(1991)\" title=\"Kids &amp; COMPUTERS: What's a parent to do? Panel Session (1991)\">Kids &amp; COMPUTERS: What's a parent to do? Panel Session<\/a><\/td>\n<td>000000001991-09-21-0000September 21, 1991<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Panel<\/td>\n<td>VPRI-0443<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Kids_%26_COMPUTERS:_What%27s_a_parent_to_do%3F_(1991)\" title=\"Kids &amp; COMPUTERS: What's a parent to do? (1991)\">Kids &amp; COMPUTERS: What's a parent to do?<\/a><\/td>\n<td>000000001991-09-21-0000September 21, 1991<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0441<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Computerland_(1991)\" title=\"Alan Kay at Computerland (1991)\">Alan Kay at Computerland<\/a><\/td>\n<td>000000001991-10-01-0000October 1991<\/td>\n<td>Talk<\/td>\n<td>VPRI-0629<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_National_School_Board_Association_Meeting_(1992)\" title=\"Alan Kay at National School Board Association Meeting (1992)\">Alan Kay at National School Board Association Meeting<\/a><\/td>\n<td>000000001992-01-01-00001992<\/td>\n<td>Talk<\/td>\n<td>VPRI-0628<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_UCLA_CS_Department_-_Distinguished_Lecturer_Series_(1993)\" title=\"Alan Kay at UCLA CS Department - Distinguished Lecturer Series (1993)\">Alan Kay at UCLA CS Department - Distinguished Lecturer Series (1993)<\/a><\/td>\n<td>000000001993-04-13-0000April 13, 1993<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI 0630<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Xerox_PARC_(1993)\" title=\"Alan Kay at Xerox PARC (1993)\">Alan Kay at Xerox PARC (1993)<\/a><\/td>\n<td>000000001993-09-23-0000September 23, 1993<\/td>\n<td>Palo Alto, CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0437<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Art_Institute_of_Chicago\" title=\"Alan Kay at Art Institute of Chicago\">Alan Kay at Art Institute of Chicago<\/a><\/td>\n<td>000000001993-01-01-00001993 (approx.)<\/td>\n<td>Chicago, IL USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0624<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_World_Economic_Forum,_Davos,_Switzerland_(1994)\" title=\"Alan Kay at World Economic Forum, Davos, Switzerland (1994)\">Alan Kay at 1994 World Economic Forum, Davos, Switzerland<\/a><\/td>\n<td>000000001994-01-01-00001994<\/td>\n<td>Davos, Switzerland<\/td>\n<td>Panel<\/td>\n<td>VPRI-0631<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Delivering_the_Future_(1994)\" title=\"Delivering the Future (1994)\">Delivering the Future<\/a><\/td>\n<td>000000001994-03-02-0000March 2, 1994<\/td>\n<td>Talk<\/td>\n<td>VPRI-0622<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Super_Highway_Summit_at_UCLA\" title=\"Super Highway Summit at UCLA\">Super Highway Summit at UCLA<\/a><\/td>\n<td>000000001994-01-01-00001994 (approx.)<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Panel<\/td>\n<td>VPRI-0625<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Vannevar_Bush_Symposium_(1995)\" title=\"Alan Kay at Vannevar Bush Symposium (1995)\">Alan Kay at Vannevar Bush Symposium (1995)<\/a><\/td>\n<td>000000001995-01-01-00001995<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Technology_for_Learning_(1995)\" title=\"Technology for Learning (1995)\">Technology for Learning (1995)<\/a><\/td>\n<td>000000001995-06-27-0000June 27, 1995<\/td>\n<td>Talk<\/td>\n<td>VPRI-0820<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Technology_in_Education_House_Committee_Part_1_(1995)\" title=\"Technology in Education House Committee Part 1 (1995)\">Technology in Education House Committee<\/a><\/td>\n<td>000000001995-10-12-0000October 12, 1995<\/td>\n<td>Washington DC<\/td>\n<td>Hearing<\/td>\n<td>VPRI-0614-1<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_The_Association_of_American_Medical_Colleges_(1996)\" title=\"Alan Kay at The Association of American Medical Colleges (1996)\">Alan Kay at The Association of American Medical Colleges<\/a><\/td>\n<td>000000001996-11-10-0000November 10, 1996<\/td>\n<td>Talk<\/td>\n<td>VPRI-0697<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/International_Forum_on_Multimedia\" title=\"International Forum on Multimedia\">International Forum on Multimedia<\/a><\/td>\n<td>000000001997-01-01-00001997<\/td>\n<td>Talk<\/td>\n<td>VPRI-0742<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Stanford_EE_380_class_(1997)\" title=\"Alan Kay at Stanford EE 380 class (1997)\">Alan Kay at Stanford EE 380 class<\/a><\/td>\n<td>000000001997-06-04-0000June 4, 1997<\/td>\n<td>Palo Alto, CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0796<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_OOPSLA_1997:_The_Computer_Revolution_has_not_Happened_Yet\" title=\"Alan Kay at OOPSLA 1997: The Computer Revolution has not Happened Yet\">Alan Kay at OOPSLA 1997: The Computer Revolution has not Happened Yet<\/a><\/td>\n<td>000000001997-10-01-0000October 1997<\/td>\n<td>Atlanta, GA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0719<\/td>\n<td>Transcript\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Education_in_the_Digital_Age_(1998)\" title=\"Education in the Digital Age (1998)\">Education in the Digital Age<\/a><\/td>\n<td>000000001998-01-01-00001998<\/td>\n<td>Interview<\/td>\n<td>VPRI-0673<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_MIT-EECS_1998_Fall_Semester_Colloquium\" title=\"Alan Kay at MIT-EECS 1998 Fall Semester Colloquium\">Alan Kay at MIT-EECS 1998 Fall Semester Colloquium Series<\/a><\/td>\n<td>000000001998-01-01-00001998<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0834<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Interview_on_toys_for_education_(1999)\" title=\"Alan Kay Interview on toys for education (1999)\">Alan Kay Interview on toys for education (1999)<\/a><\/td>\n<td>000000001999-03-05-0000March 5, 1999<\/td>\n<td>Interview<\/td>\n<td>VPRI-0913<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Tale_of_Two_Schools:_Technology_in_the_Classroom_(1999)\" title=\"Tale of Two Schools: Technology in the Classroom (1999)\">Tale of Two Schools: Technology in the Classroom (1999)<\/a><\/td>\n<td>000000001999-06-04-0000June 4, 1999<\/td>\n<td>TV<\/td>\n<td>VPRI-0747<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Vantive_World_(1999)\" title=\"Vantive World (1999)\">Vantive World (1999)<\/a><\/td>\n<td>000000001999-01-01-00001999<\/td>\n<td>Talk<\/td>\n<td>VPRI-0827<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Camden_Technology_Conference_(2000)\" title=\"Camden Technology Conference (2000)\">Alan Kay and Don Lewis at Camden Technology Conference<\/a><\/td>\n<td>000000002000-01-04-0000January 4, 2000<\/td>\n<td>Talk<\/td>\n<td>VPRI-0714-1<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Squeak_Presentation_(2000)\" title=\"Alan Kay Squeak Presentation (2000)\">Alan Kay Squeak Presentation (2000)<\/a><\/td>\n<td>000000002000-06-09-0000June 9, 2000<\/td>\n<td>Glendale CA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0872<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Talk_at_Open_School_(2001)\" title=\"Alan Kay Talk at Open School (2001)\">Alan Kay Talk at Open School<\/a><\/td>\n<td>000000002001-01-01-00002001<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0989<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Squeak_Demo_at_ACM1_(2001)\" title=\"Alan Kay Squeak Demo at ACM1 (2001)\">Alan Kay at ACM1<\/a><\/td>\n<td>000000002001-01-01-00002001<\/td>\n<td>San Jose, CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-0654<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_SD%26M_conference_on_Graphical_User_Interface_(2001)\" title=\"Alan Kay at SD&amp;M conference on Graphical User Interface (2001)\">Alan Kay at SD&amp;M conference on Graphical User Interface<\/a><\/td>\n<td>000000002001-01-01-00002001<\/td>\n<td>Germany<\/td>\n<td>Talk<\/td>\n<td>VPRI-0713<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_4th_Annual_Lynford_Lecture_(2001)\" title=\"Alan Kay at 4th Annual Lynford Lecture (2001)\">Alan Kay at 4th Annual Lynford Lecture (2001)<\/a><\/td>\n<td>000000002001-10-01-0000October 2001<\/td>\n<td>New York, NY USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Presentation_at_HP_Tech_Con_(2003)\" title=\"Alan Kay Presentation at HP Tech Con (2003)\">Alan Kay Presentation at HP Tech Con (2003)<\/a><\/td>\n<td>000000002003-01-01-00002003<\/td>\n<td>Talk<\/td>\n<td>VPRI-0726<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Etech_2003\" title=\"Alan Kay at Etech 2003\">Alan Kay at Etech 2003<\/a><\/td>\n<td>000000002003-01-01-00002003<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Marshall_McLuhan_Lecture_(2003)\" title=\"Alan Kay at Marshall McLuhan Lecture (2003)\">Alan Kay at Marshall McLuhan Lecture (2003)<\/a><\/td>\n<td>000000002003-03-14-0000March 14, 2003<\/td>\n<td>Talk<\/td>\n<td>VPRI-0688<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Back_to_the_Future_of_Software_Development_(2003)\" title=\"Back to the Future of Software Development (2003)\">Back to the Future of Software Development by Alan Kay and David Smith<\/a><\/td>\n<td>000000002003-04-23-0000April 23, 2003<\/td>\n<td>Mountain View, CA USA<\/td>\n<td>Talk<\/td>\n<td> VPRI-0785<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Turing_Award_Lecture_(2004)\" title=\"Alan Kay Turing Award Lecture (2004)\">Alan Kay Turing Award Lecture (2004)<\/a><\/td>\n<td>000000002004-10-26-0000October 26, 2004<\/td>\n<td>Vancouver, BC Canada<\/td>\n<td>Talk<\/td>\n<td>VPRI-0761<\/td>\n<td>Annotated\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay,_%2705_Columbia_College_Commencement\" title=\"Alan Kay, '05 Columbia College Commencement\">Alan Kay, '05 Columbia College Commencement<\/a><\/td>\n<td>000000002005-01-01-00002005<\/td>\n<td>Chicago, IL USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Kyoto_Prize_Morning_TV_Show_Interview_(2005)\" title=\"Kyoto Prize Morning TV Show Interview (2005)\">Kyoto Prize Morning TV Show Interview (2005)<\/a><\/td>\n<td>000000002005-02-28-0000February 28, 2005<\/td>\n<td>San Diego, CA USA<\/td>\n<td>TV<\/td>\n<td>VPRI-0762<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Kyoto_Prize_Laureate_Symposium_Talk_at_SDSU\" title=\"Kyoto Prize Laureate Symposium Talk at SDSU\">Kyoto Prize Laureate Symposium Talk at SDSU<\/a><\/td>\n<td>000000002005-03-03-0000March 3, 2005<\/td>\n<td>San Diego CA USA<\/td>\n<td>Talk<\/td>\n<td>VPRI-1005<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay%27s_IFERS_address_at_the_American_Film_Institute\" title=\"Alan Kay's IFERS address at the American Film Institute\">Alan Kay's IFERS address at the American Film Institute<\/a><\/td>\n<td>000000002006-11-04-0000November 4, 2006<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Talk_at_MobilFest_2007\" title=\"Alan Kay Talk at MobilFest 2007\">Alan Kay Talk at MobilFest 2007<\/a><\/td>\n<td>000000002007-11-21-0000November 21, 2007<\/td>\n<td>Brazil<\/td>\n<td>Talk<\/td>\n<td>Transcript\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_:_July_2007_:_A_Conversation_with_CMU_Faculty_%26_Students\" title=\"Alan Kay : July 2007 : A Conversation with CMU Faculty &amp; Students\">Alan Kay : July 2007 : A Conversation with CMU Faculty &amp; Students<\/a><\/td>\n<td>000000002007-01-01-00002007<\/td>\n<td>Pittsburg, PA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay:_A_powerful_idea_about_teaching_ideas_at_TED_(2007)\" title=\"Alan Kay: A powerful idea about teaching ideas at TED (2007)\">Alan Kay: A powerful idea about teaching ideas at TED (2007)<\/a><\/td>\n<td>000000002007-03-01-0000March 2007<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay:_Beyond_the_Printing_Press_(2008)\" title=\"Alan Kay: Beyond the Printing Press (2008)\">Alan Kay: Beyond the Printing Press (2008)<\/a><\/td>\n<td>000000002008-05-20-0000May 20, 2008<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Talk_at_40th_Anniversary_of_Mother_of_All_Demos_(2008)\" title=\"Alan Kay Talk at 40th Anniversary of Mother of All Demos (2008)\">Alan Kay Talk at 40th Anniversary of Mother of All Demos (2008)<\/a><\/td>\n<td>000000002008-12-09-0000December 9, 2008<\/td>\n<td>San Francisco<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/An_Interview_with_Alan_Kay_by_Martin_Wasserman\" title=\"An Interview with Alan Kay by Martin Wasserman\">An Interview with Alan Kay by Martin Wasserman<\/a><\/td>\n<td>000000002008-12-09-0000December 9, 2008<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Honorary_Doctorate_Speech_at_Kyoto_University\" title=\"Honorary Doctorate Speech at Kyoto University\">Honorary Doctorate Speech at Kyoto University<\/a><\/td>\n<td>000000002009-01-20-0000January 20, 2009<\/td>\n<td> Kyoto, Japan<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay:_Normal_Considered_Harmful_(2009)\" title=\"Alan Kay: Normal Considered Harmful (2009)\">Alan Kay: Normal Considered Harmful (2009)<\/a><\/td>\n<td>000000002009-10-22-0000October 22, 2009<\/td>\n<td>Urbana-Champaign, IL<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/How_Complex_is_%22Personal_Computing%22%3F_(2009)\" title=\"How Complex is &quot;Personal Computing&quot;? (2009)\">How Complex is \"Personal Computing\"? (2009)<\/a><\/td>\n<td>000000002009-10-22-0000October 22, 2009<\/td>\n<td>Urbana-Champaign, IL USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay:_Big_Ideas_are_Sometimes_Powerful_Ideas\" title=\"Alan Kay: Big Ideas are Sometimes Powerful Ideas\">Alan Kay: Big Ideas are Sometimes Powerful Ideas<\/a><\/td>\n<td>000000002009-12-07-0000December 7, 2009<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_CES_2010\" title=\"Alan Kay at CES 2010\">Alan Kay at CES 2010<\/a><\/td>\n<td>000000002010-01-01-0000January 2010<\/td>\n<td>Las Vegas, NV USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Interview_at_eLearning_Symposium_(2010)\" title=\"Alan Kay Interview at eLearning Symposium (2010)\">Alan Kay Interview at eLearning Symposium (2010)<\/a><\/td>\n<td>000000002010-06-02-0000June 2, 2010<\/td>\n<td>Fairfax, VA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Talk_at_eLearning_Symposium_(2010)\" title=\"Alan Kay Talk at eLearning Symposium (2010)\">Alan Kay Talk at eLearning Symposium (2010)<\/a><\/td>\n<td>000000002010-06-02-0000June 2, 2010<\/td>\n<td>Fairfax, VA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_CS1_talk_at_UCLA_(2010)\" title=\"Alan Kay CS1 talk at UCLA (2010)\">Alan Kay CS1 talk at UCLA (2010)<\/a><\/td>\n<td>000000002010-10-01-0000October 2010<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_scrii_Global_Conference_(2011)\" title=\"Alan Kay at scrii Global Conference (2011)\">Alan Kay at scrii Global Conference (2011)<\/a><\/td>\n<td>000000002011-03-31-0000March 31, 2011<\/td>\n<td>San Jose, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay:_Programming_and_Scaling\" title=\"Alan Kay: Programming and Scaling\">Alan Kay: Programming and Scaling<\/a><\/td>\n<td>000000002011-07-21-0000July 21, 2011<\/td>\n<td>Potsdam, Germany<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Brick_Laying_or_Arch_Inventing_at_ECOOP_2011\" title=\"Brick Laying or Arch Inventing at ECOOP 2011\">Brick Laying or Arch Inventing at ECOOP 2011<\/a><\/td>\n<td>000000002011-07-28-0000July 28, 2011<\/td>\n<td>Lancaster, UK<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Rethinking_Design,_Risk,_and_Software_(2012)\" title=\"Rethinking Design, Risk, and Software (2012)\">Rethinking Design, Risk, and Software  at COFES<\/a><\/td>\n<td>000000002012-04-13-0000April 13, 2012<\/td>\n<td>Scottsdale, AZ, USA<\/td>\n<td>Talk<\/td>\n<td>Annotated\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_the_NITLE_Summit_(2012)\" title=\"Alan Kay at the NITLE Summit (2012)\">Alan Kay at the NITLE Summit (2012)<\/a><\/td>\n<td>000000002012-04-15-0000April 15, 2012<\/td>\n<td>Arlington, VA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_talk_at_DePaul_University_(2012)\" title=\"Alan Kay talk at DePaul University (2012)\">Alan Kay talk at DePaul  University (2012)<\/a><\/td>\n<td>000000002012-06-10-0000June 10, 2012<\/td>\n<td>Chicago, IL USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Extracting_Energy_from_the_Turing_Tarpit_(2012)\" title=\"Extracting Energy from the Turing Tarpit (2012)\">Extracting Energy from the Turing Tarpit (2012)<\/a><\/td>\n<td>000000002012-06-15-0000June 15, 2012<\/td>\n<td>San Francisco, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/An_Interview_with_Computing_Pioneer_Alan_Kay\" title=\"An Interview with Computing Pioneer Alan Kay\">An Interview with Computing Pioneer Alan Kay<\/a><\/td>\n<td>000000002012-07-01-0000July 2012<\/td>\n<td>Paderborn, Germany<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_SCIx_Keynote_Presentation_(2012)\" title=\"Alan Kay SCIx Keynote Presentation (2012)\">Alan Kay  SCIx Keynote Presentation (2012)<\/a><\/td>\n<td>000000002012-11-13-0000November 13, 2012<\/td>\n<td>Salt Lake City, UT USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Interview_by_Dave_Marvit_(2013)\" title=\"Alan Kay Interview by Dave Marvit (2013)\">Alan Kay Interview by Dave Marvit (2013)<\/a><\/td>\n<td>000000002013-01-23-0000January 23, 2013<\/td>\n<td>Mountain View CA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Keynote_at_NATF_2013_Part_1\" title=\"Alan Kay Keynote at NATF 2013 Part 1\">Alan Kay Keynote at NATF 2013 Part 1<\/a><\/td>\n<td>000000002013-01-24-0000January 24, 2013<\/td>\n<td>Mountain View, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Keynote_at_NATF_2013_Part_2\" title=\"Alan Kay Keynote at NATF 2013 Part 2\">Alan Kay Keynote at NATF 2013 Part 2<\/a><\/td>\n<td>000000002013-01-24-0000January 24, 2013<\/td>\n<td>Mountain View, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Technology_and_Transformation:_Vishal_Sikka_and_Alan_Kay_in_Conversation_with_Paul_Saffo\" title=\"Technology and Transformation: Vishal Sikka and Alan Kay in Conversation with Paul Saffo\">Technology and Transformation: Vishal Sikka and Alan Kay in Conversation with Paul Saffo<\/a><\/td>\n<td>000000002013-03-26-0000March 26, 2013<\/td>\n<td>Panel<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay:_Putting_Turing_to_Work_at_Heigelberg_Laureate_Forum_(2013)\" title=\"Alan Kay: Putting Turing to Work at Heigelberg Laureate Forum (2013)\">Alan Kay: Putting Turing to Work at Heigelberg Laureate Forum (2013)<\/a><\/td>\n<td>000000002013-09-27-0000September 27, 2013<\/td>\n<td>Heidelberg, Germany<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Qualcomm:_Is_it_really_%22Complex%22%3F_Or_did_we_just_make_it_%22Complicated%22%3F\" title=\"Alan Kay at Qualcomm: Is it really &quot;Complex&quot;? Or did we just make it &quot;Complicated&quot;?\">Alan Kay at Qualcomm: Is it really \"Complex\"? Or did we just make it \"Complicated\"?<\/a><\/td>\n<td>000000002013-10-30-0000October 30, 2013<\/td>\n<td>San Diego, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_-_Programming_Languages_%26_Programming\" title=\"Alan Kay - Programming Languages &amp; Programming\">Alan Kay - Programming Languages &amp; Programming<\/a><\/td>\n<td>000000002013-01-01-00002013<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay%27s_tribute_to_Ted_Nelson_at_%22Intertwingled%22_Fest_(2014)\" title=\"Alan Kay's tribute to Ted Nelson at &quot;Intertwingled&quot; Fest (2014)\">Alan Kay's tribute to Ted Nelson at \"Intertwingled\" Fest (2014)<\/a><\/td>\n<td>000000002014-01-01-00002014<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_LampsonFest_(2014)\" title=\"Alan Kay at LampsonFest (2014)\">Alan Kay at LampsonFest (2014)<\/a><\/td>\n<td>000000002014-02-21-0000February 21, 2014<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Founder_School_Session:_The_Future_Doesn%27t_Have_to_Be_Incremental_(2014)\" title=\"Alan Kay at Founder School Session: The Future Doesn't Have to Be Incremental (2014)\">Alan Kay at Founder School Session: The Future Doesn't Have to Be Incremental (2014)<\/a><\/td>\n<td>000000002014-04-03-0000April 3, 2014<\/td>\n<td>San Francisco, CA USA<\/td>\n<td>Talk<\/td>\n<td>Transcript\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Interview_with_Alan_Kay_by_Gardner_Campbell\" title=\"Interview with Alan Kay by Gardner Campbell\">Interview with Alan Kay by Gardner Campbell<\/a><\/td>\n<td>000000002014-06-20-0000June 20, 2014<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay,_2015:_Power_of_Simplicity\" title=\"Alan Kay, 2015: Power of Simplicity\">Alan Kay, 2015: Power of Simplicity<\/a><\/td>\n<td>000000002015-01-01-00002015<\/td>\n<td>Menlo Park, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_SAP_(2015)\" title=\"Alan Kay at SAP (2015)\">Alan Kay at SAP (2015)<\/a><\/td>\n<td>000000002015-01-01-00002015<\/td>\n<td>Palo Alto, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_-_Inventing_the_Future_Part_1_(2015)\" title=\"Alan Kay - Inventing the Future Part 1 (2015)\">Alan Kay - Inventing the Future Part 1 (2015)<\/a><\/td>\n<td>000000002015-01-01-00002015<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_-_Inventing_the_Future_Part_2_(2015)\" title=\"Alan Kay - Inventing the Future Part 2 (2015)\">Alan Kay - Inventing the Future Part 2 (2015)<\/a><\/td>\n<td>000000002015-01-01-00002015<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_STRAP_2015_-_Our_Most_Important_Revolution_part_(1\/2)\" title=\"Alan Kay at STRAP 2015 - Our Most Important Revolution part (1\/2)\">Alan Kay at STRAP 2015 - Our Most Important Revolution part (1\/2)<\/a><\/td>\n<td>000000002015-02-05-0000February 5, 2015<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_STRAP_2015_-_Our_Most_Important_Revolution_part_(2\/2)\" title=\"Alan Kay at STRAP 2015 - Our Most Important Revolution part (2\/2)\">Alan Kay at STRAP 2015 - Our Most Important Revolution part (2\/2)<\/a><\/td>\n<td>000000002015-02-05-0000February 5, 2015<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_Aspen_Institute\/Kennedy_Center_Arts_Summit_%22Science_and_Technology_as_Art%22_(2015)\" title=\"Alan Kay at Aspen Institute\/Kennedy Center Arts Summit &quot;Science and Technology as Art&quot; (2015)\">Alan Kay at Aspen Institute\/Kennedy Center Arts Summit \"Science and Technology as Art\" (2015)<\/a><\/td>\n<td>000000002015-05-16-0000May 16, 2015<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Rethinking_CS_Education\" title=\"Rethinking CS Education\"> Alan Kay, CrossRoads 2015<\/a><\/td>\n<td>000000002015-01-01-00002015<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_the_Media_Lab_30th_anniversary_(2015)\" title=\"Alan Kay at the Media Lab 30th anniversary (2015)\">Alan Kay at the Media Lab 30th anniversary (2015)<\/a><\/td>\n<td>000000002015-10-30-0000October 30, 2015<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Joe_Armstrong_%26_Alan_Kay_-_Joe_Armstrong_interviews_Alan_Kay\" title=\"Joe Armstrong &amp; Alan Kay - Joe Armstrong interviews Alan Kay\">Joe Armstrong &amp; Alan Kay - Joe Armstrong interviews Alan Kay<\/a><\/td>\n<td>000000002016-01-01-00002016<\/td>\n<td>London, UK<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/CHI_2016_Plenary:_Alan_Kay_in_conversation_with_Vishal_Sikka\" title=\"CHI 2016 Plenary: Alan Kay in conversation with Vishal Sikka\">CHI 2016 Plenary: Alan Kay in conversation with Vishal Sikka<\/a><\/td>\n<td>000000002016-03-12-0000March 12, 2016<\/td>\n<td>San Jose, CA USA<\/td>\n<td>Interview<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_UCLA_(2016)\" title=\"Alan Kay at UCLA (2016)\">Alan Kay at UCLA<\/a><\/td>\n<td>000000002016-04-07-0000April 7, 2016<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_CresstCon_(2016)\" title=\"Alan Kay at CresstCon (2016)\">Alan Kay at CresstCon (2016)<\/a><\/td>\n<td>000000002016-09-21-0000September 21, 2016<\/td>\n<td>Los Angeles, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_at_the_Thinking_about_Thinking_about_Seymour_Event_(2017)\" title=\"Alan Kay at the Thinking about Thinking about Seymour Event (2017)\">Alan Kay at the Thinking about Thinking about Seymour Event (2017)<\/a><\/td>\n<td>000000002017-01-27-0000January 27, 2017<\/td>\n<td>Cambridge, MA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/CrossRoads_2017\" title=\"CrossRoads 2017\"> Alan Kay keynote - Education That Takes Us To The 22nd Century<\/a><\/td>\n<td>000000002017-05-25-0000May 25, 2017<\/td>\n<td>San Francisco, CA USA<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Starship_Congress_2017:_Alan_Kay,_%22Preparations_for_Communicating_with_Aliens%22_(2017)\" title=\"Starship Congress 2017: Alan Kay, &quot;Preparations for Communicating with Aliens&quot; (2017)\">Starship Congress 2017: Alan Kay, \"Preparations for Communicating with Aliens\" (2017)<\/a><\/td>\n<td>000000002017-08-08-0000August 8, 2017<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_Lecture_-_CCSMFAIXD_IxD_Evolution_Course\" title=\"Alan Kay Lecture - CCSMFAIXD IxD Evolution Course\">Alan Kay Lecture - CCSMFAIXD IxD Evolution Course<\/a><\/td>\n<td>000000002017-10-09-0000October 9, 2017<\/td>\n<td>Detroit<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><tr><td><a href=\"https:\/\/tinlizzie.org\/IA\/index.php\/Alan_Kay_2018_Lindberg-King_Lecture\" title=\"Alan Kay 2018 Lindberg-King Lecture\">Alan Kay 2018 Lindberg-King Lecture<\/a><\/td>\n<td>000000002018-09-23-0000September 23, 2018<\/td>\n<td>Washington DC<\/td>\n<td>Talk<\/td>\n<td>\n<\/td><\/tr><\/table><p>Retrieved from \"<a dir=\"ltr\" href=\"http:\/\/tinlizzie.org\/IA\/index.php?title=Talks_by_Alan_Kay&amp;oldid=3041\">http:\/\/tinlizzie.org\/IA\/index.php?title=Talks_by_Alan_Kay&amp;oldid=3041<\/a>\"<\/p>","created_at":"2018-11-07T16:22:54+0000","updated_at":"2018-11-07T16:22:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"tinlizzie.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":120,"title":"jbelien's diary | Why use OpenStreetMap instead of Google Maps ? | OpenStreetMap","url":"https:\/\/www.openstreetmap.org\/user\/jbelien\/diary\/44356","content":"<p>Cet article est \u00e9galement disponible en fran\u00e7ais : <a href=\"http:\/\/www.openstreetmap.org\/user\/jbelien\/diary\/44358\" rel=\"nofollow\">http:\/\/www.openstreetmap.org\/user\/jbelien\/diary\/44358<\/a><\/p><hr \/><p><em>Article written for <a href=\"http:\/\/2018.opensummerofcode.be\/\" rel=\"nofollow\">open Summer of code 2018<\/a><\/em><\/p><hr \/><p>First of all, there is absolutely not doubt about the fact Google Maps is a good product and the Google Maps tools (API, JavaScript library, ...) work quite well and are easy to setup <strong>but<\/strong> ...<\/p><h2>Google is in charge<\/h2><p>When Google Maps started more than 10 years ago, everything was completely free to use. These last few months that has changed drastically.<br \/>The first decision made by Google was to limit the free number of requests per day. It didn't really impact the small users that were still under the limit but some big players decided to switch to other solution like OpenStreetMap (for instance Foursquare and Pinterest).<br \/>The latest change (June 2018) has much more impact : Google decided to reduce the limit of free requests (25000 map display per day to 28000 per month - that's around 1000 map display per day, so 25 times less) but also made it mandatory to give your credit card number even if you do not go over the free limit !<\/p><p>Impact on Belgian Prime Minister website :<\/p><p><img src=\"https:\/\/jbelien.be\/assets\/images\/2018-07-14-switch-to-osm-1.png\" alt=\"\" \/><\/p><p>Money is definitely an issue but worst, Google can decide what's displayed or hidden or highlighted and how it is displayed. What if Google decide to \"hide\" or \"highlight\" certain kind of shops for instance ; even worse, what if they decide to draw the border between countries wherever they want. All of this could quickly lead to ethical issues !<\/p><p>Long story short, you have absolutely <strong>no<\/strong> control about this, Google can change the term of use of their service when they want how they want and it will most probably not be in your advantage !<\/p><h2>Why OpenStreetMap ?<\/h2><blockquote>\n<p>OpenStreetMap is a free, editable map of the whole world that is being built by volunteers largely from scratch and released with an open-content license.<\/p>\n<p>The OpenStreetMap License allows free (or almost free) access to our map images and all of our underlying map data. The project aims to promote new and interesting uses of this data.<\/p>\n<p>-- Source: <a href=\"https:\/\/wiki.openstreetmap.org\/wiki\/About_OpenStreetMap\" rel=\"nofollow\">https:\/\/wiki.openstreetmap.org\/wiki\/About_OpenStreetMap<\/a><\/p>\n<\/blockquote><p>OpenStreetMap use is free as in <em>free beer<\/em> but also as in <em>free speech<\/em> !<br \/>Everyone can contribute and everyone can use it ! <a href=\"https:\/\/www.openstreetmap.org\/copyright\" rel=\"nofollow\">OpenStreetMap is open data<\/a> !\nWhether you're a developer, a scientist, an activist, a cartographer, ..., you can use OpenStreetMap database to fulfil your need.<\/p><h3>You have access to the data<\/h3><p>Google Maps doesn't give you access to the data behind the map (and that's the case for most proprietary solutions) but you have full access to the OpenStreetMap database !<\/p><p>Access to the data means you can build a map that suits <strong>you<\/strong> and not depend on decisions made by the map provider.<\/p><p><a href=\"http:\/\/openwhatevermap.xyz\/\" rel=\"nofollow\"><img src=\"https:\/\/jbelien.be\/assets\/images\/2018-07-14-switch-to-osm-2.png\" alt=\"\" \/><\/a><\/p><p>If you need some information, just download it ! Let's say you need all the buildings from Brussels for a research project, or you want to find the nearest ATM, or you want all the information from the country you're in vacation and you need it offline, ... ; those are all simple queries ! OpenStreetMap database is open !<\/p><p>OpenStreetMap covers the whole world with the same mapping rules so it's really easy to switch from local to global !<\/p><p>If you see that something is wrong in the map, just fix it, or notify the community and we'll fix it ! Fixing the map will not only benefit one big player but everyone in the world. OpenStreetMap is made by citizens just like you.<\/p><h3>Power of the community<\/h3><p>Thanks to hundreds of thousand contributors all over the world OpenStreetMap is probably the most up-to-date and complete map of the world.<\/p><p>When the city center of Brussels became officially accessible for pedestrian only, the update was immediately made in OpenStreetMap the day it became official ; same thing for the recent mobility changes in Ghent. Google Maps of course also made the change but it took a few weeks, sometimes months ... and that can lead to some issues (navigation, ...).<\/p><p>Third world countries are also a great example of the power of the community : humanitarian organization (like Doctors Without Borders, Red Cross, ...) often need maps to be able to prepare and coordinate their actions on the field. It usually costs a lot to buy maps from those regions (if there is a map available) so they made the switch to OpenStreetMap. They call us on a regular basis to ask the community to map part of the world where they have to take actions.<\/p><p>Example of difference between OpenStreetMap and Google Maps in Kathmandu :\n<a href=\"https:\/\/tools.geofabrik.de\/mc\/#14\/27.5092\/85.3607&amp;num=2&amp;mt0=mapnik&amp;mt1=google-map\" rel=\"nofollow\"><img src=\"https:\/\/jbelien.be\/assets\/images\/2018-07-14-switch-to-osm-3.png\" alt=\"\" \/><\/a><\/p><h3>That's nice but let's get technical ...<\/h3><p>Many libraries and tools are already using OpenStreetMap data.<\/p><p>You're looking for a <strong>baselayer<\/strong>, there are so much options available depending on your needs (standard basemap, basemap focused on cycling, on accessibility, ...). Have a look at the list available on the wiki : <a href=\"https:\/\/wiki.openstreetmap.org\/wiki\/Tiles#Servers\" rel=\"nofollow\">https:\/\/wiki.openstreetmap.org\/wiki\/Tiles#Servers<\/a> and if you want to play with Vector Tiles, have a look at <a href=\"https:\/\/openmaptiles.org\/\" rel=\"nofollow\">https:\/\/openmaptiles.org\/<\/a>.<br \/>OpenStreetMap Belgium also provides its own baselayers covering Belgium and surroundings : <a href=\"https:\/\/tile.osm.be\/\" rel=\"nofollow\">https:\/\/tile.osm.be\/<\/a>.<\/p><p>You're looking for a <strong>geocoding API<\/strong>, have a look at Nominatim : <a href=\"https:\/\/nominatim.openstreetmap.org\/\" rel=\"nofollow\">https:\/\/nominatim.openstreetmap.org\/<\/a>.<\/p><p>You're looking for a <strong>routing algorithm<\/strong>, check OSRM (<a href=\"http:\/\/project-osrm.org\/\" rel=\"nofollow\">http:\/\/project-osrm.org\/<\/a>), GraphHopper (<a href=\"https:\/\/www.graphhopper.com\/\" rel=\"nofollow\">https:\/\/www.graphhopper.com\/<\/a>), Itinero (<a href=\"http:\/\/www.itinero.tech\/\" rel=\"nofollow\">http:\/\/www.itinero.tech\/<\/a>), ...<\/p><p>Everything that's possible with Google Maps is possible with OpenStreetMap and (almost) as easy to use and set up !<\/p><p>You can find more information on <a href=\"https:\/\/switch2osm.org\/\" rel=\"nofollow\">https:\/\/switch2osm.org\/<\/a> !<\/p><p>You can also have a look at all the great libraries made by Mapbox : <a href=\"https:\/\/www.mapbox.com\/help\/how-web-apps-work\/\" rel=\"nofollow\">https:\/\/www.mapbox.com\/help\/how-web-apps-work\/<\/a><br \/>That's probably the easiest way to switch from Google Maps to OpenStreetMap : Mapbox libraries and API allow you to have all the function provided by Google Maps (basemap - you can even style your own -, directions, geocoding, ...) and are all based on OpenStreetMap data. Most of Mapbox products are <a href=\"https:\/\/github.com\/mapbox\" rel=\"nofollow\">open-source<\/a> but some services and functions come with a downside, you'll have to pay if you go over the request limits.<\/p><hr \/><p><strong>Some more reading (and sources used for this article) :<\/strong><\/p><p>In English:<\/p><ul><li><a href=\"https:\/\/switch2osm.org\/why-switch\/\" rel=\"nofollow\">https:\/\/switch2osm.org\/why-switch\/<\/a><\/li>\n<li><a href=\"https:\/\/wiki.openstreetmap.org\/wiki\/Why_OpenStreetMap%3F\" rel=\"nofollow\">https:\/\/wiki.openstreetmap.org\/wiki\/Why_OpenStreetMap%3F<\/a><\/li>\n<li><a href=\"http:\/\/geoawesomeness.com\/why-would-you-use-openstreetmap-if-there-is-google-maps\/\" rel=\"nofollow\">http:\/\/geoawesomeness.com\/why-would-you-use-openstreetmap-if-there-is-google-maps\/<\/a> (by <a href=\"https:\/\/twitter.com\/abuczkowski\" rel=\"nofollow\">A. Buczkowski<\/a>)<\/li>\n<li><a href=\"https:\/\/www.citylab.com\/design\/2016\/08\/google-maps-areas-of-interest\/493670\/\" rel=\"nofollow\">https:\/\/www.citylab.com\/design\/2016\/08\/google-maps-areas-of-interest\/493670\/<\/a> (by <a href=\"https:\/\/twitter.com\/mslaurabliss\" rel=\"nofollow\">L. Bliss<\/a>)<\/li>\n<\/ul><p>In French :<\/p><ul><li><a href=\"https:\/\/medium.com\/@cq94\/dont-be-evil-until-95f2e8dfaaad\" rel=\"nofollow\">https:\/\/medium.com\/@cq94\/dont-be-evil-until-95f2e8dfaaad<\/a> (by <a href=\"https:\/\/twitter.com\/cq94\" rel=\"nofollow\">C. Quest<\/a> from OpenStreetMap France)<\/li>\n<li><a href=\"https:\/\/medium.com\/@cq94\/tarifs-google-maps-pour-les-nuls-d47c1833c242\" rel=\"nofollow\">https:\/\/medium.com\/@cq94\/tarifs-google-maps-pour-les-nuls-d47c1833c242<\/a> (by <a href=\"https:\/\/twitter.com\/cq94\" rel=\"nofollow\">C. Quest<\/a> from OpenStreetMap France)<\/li>\n<\/ul>","created_at":"2018-11-07T16:22:53+0000","updated_at":"2018-11-07T16:22:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"www.openstreetmap.org","preview_picture":"https:\/\/www.openstreetmap.org\/assets\/osm_logo_256-cde84d7490f0863c7a0b0d0a420834ebd467c1214318167d0f9a39f25a44d6bd.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":118,"title":"Stripe - Online payment processing for internet businesses","url":"https:\/\/stripe.com\/","content":"<section id=\"always-improving\"><h2 class=\"common-Uppercase\">\n          \n          Always improving\n        <\/h2><p class=\"common-BodyText\">\n          Stripe is an always-improving toolchain that gains new features every month. Our world-class engineering team constantly iterates upon every facet of the Stripe stack. And from Apple Pay to Alipay, building on Stripe means you get early access to the latest technologies.\n        <\/p><a class=\"common-BodyText common-Link common-Link--arrow\" href=\"https:\/\/stripe.com\/us\/payments\">\n          Learn about Stripe's products\n        <\/a><\/section><section id=\"global-scale\"><h2 class=\"common-Uppercase\">\n          \n          Global scale\n        <\/h2><p class=\"common-BodyText\">\n          We help power millions of businesses in 100+ countries and across nearly every industry. Headquartered in San Francisco, Stripe has 9 global offices and hundreds of people working to help transform how modern businesses are built and run.\n        <\/p><a class=\"common-BodyText common-Link common-Link--arrow\" href=\"https:\/\/stripe.com\/about\">\n          More about us\n        <\/a><\/section>","created_at":"2018-11-07T16:22:51+0000","updated_at":"2018-11-07T16:22:51+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"stripe.com","preview_picture":"https:\/\/stripe.com\/img\/v3\/home\/social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":119,"title":"The Haskell School of Music \u2013 From Signals to Symphonies (2014) [pdf]","url":"https:\/\/news.ycombinator.com\/item?id=17517285","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D17517285\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"17517285\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"http:\/\/haskell.cs.yale.edu\/wp-content\/uploads\/2015\/03\/HSoM.pdf\" class=\"storylink\">The Haskell School of Music \u2013 From Signals to Symphonies (2014) [pdf]<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=yale.edu\">yale.edu<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        129 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=alokrai\" class=\"hnuser\">alokrai<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17517285\">3 months ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=17517285&amp;goto=item%3Fid%3D17517285\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=The%20Haskell%20School%20of%20Music%20%E2%80%93%20From%20Signals%20to%20Symphonies&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=The%20Haskell%20School%20of%20Music%20%E2%80%93%20From%20Signals%20to%20Symphonies\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=17517285&amp;auth=4b47bb97b9fcf0abedae43b8098839fd409f4ac1\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=17517285\">10\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:22:51+0000","updated_at":"2018-11-07T16:22:51+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":114,"title":"mozilla\/mentat","url":"https:\/\/github.com\/mozilla\/mentat","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/mozilla\/mentat\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/77a2cca871da10e040dce4a6f15ca503c038f649\/68747470733a2f2f7472617669732d63692e6f72672f6d6f7a696c6c612f6d656e7461742e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/mozilla\/mentat.svg?branch=master\" \/><\/a><\/p>\n<p><strong>Project Mentat is <a href=\"https:\/\/mail.mozilla.org\/pipermail\/firefox-dev\/2018-September\/006780.html\" rel=\"nofollow\">no longer being developed or actively maintained by Mozilla<\/a>.<\/strong>  This repository will be marked read-only in the near future.  You are, of course, welcome to fork the repository and use the existing code.<\/p>\n<p>Project Mentat is a persistent, embedded knowledge base. It draws heavily on <a href=\"https:\/\/github.com\/tonsky\/datascript\">DataScript<\/a> and <a href=\"http:\/\/datomic.com\" rel=\"nofollow\">Datomic<\/a>.<\/p>\n<p>Mentat is implemented in Rust.<\/p>\n<p>The first version of Project Mentat, named Datomish, <a href=\"https:\/\/github.com\/mozilla\/mentat\/tree\/clojure\">was written in ClojureScript<\/a>, targeting both Node (on top of <code>promise_sqlite<\/code>) and Firefox (on top of <code>Sqlite.jsm<\/code>). It also worked in pure Clojure on the JVM on top of <code>jdbc-sqlite<\/code>. The name was changed to avoid confusion with <a href=\"http:\/\/datomic.com\" rel=\"nofollow\">Datomic<\/a>.<\/p>\n<p>The Rust implementation gives us a smaller compiled output, better performance, more type safety, better tooling, and easier deployment into Firefox and mobile platforms.<\/p>\n<p><a href=\"https:\/\/mozilla.github.io\/mentat\" rel=\"nofollow\">Documentation<\/a><\/p>\n<hr \/><h2><a id=\"user-content-motivation\" class=\"anchor\" aria-hidden=\"true\" href=\"#motivation\"><\/a>Motivation<\/h2>\n<p>Mentat is intended to be a flexible relational (not key-value, not document-oriented) store that makes it easy to describe, grow, and reuse your domain schema.<\/p>\n<p>By abstracting away the storage schema, and by exposing change listeners outside the database (not via triggers), we hope to make domain schemas stable, and allow both the data store itself and embedding applications to use better architectures, meeting performance goals in a way that allows future evolution.<\/p>\n<h2><a id=\"user-content-data-storage-is-hard\" class=\"anchor\" aria-hidden=\"true\" href=\"#data-storage-is-hard\"><\/a>Data storage is hard<\/h2>\n<p>We've observed that data storage is a particular area of difficulty for software development teams:<\/p>\n<ul><li>\n<p>It's hard to define storage schemas well. A developer must:<\/p>\n<ul><li>Model their domain entities and relationships.<\/li>\n<li>Encode that model <em>efficiently<\/em> and <em>correctly<\/em> using the features available in the database.<\/li>\n<li>Plan for future extensions and performance tuning.<\/li>\n<\/ul><p>In a SQL database, the same schema definition defines everything from high-level domain relationships through to numeric field sizes in the same smear of keywords. It's difficult for someone unfamiliar with the domain to determine from such a schema what's a domain fact and what's an implementation concession \u2014 are all part numbers always 16 characters long, or are we trying to save space? \u2014 or, indeed, whether a missing constraint is deliberate or a bug.<\/p>\n<p>The developer must think about foreign key constraints, compound uniqueness, and nullability. They must consider indexing, synchronizing, and stable identifiers. Most developers simply don't do enough work in SQL to get all of these things right. Storage thus becomes the specialty of a few individuals.<\/p>\n<p>Which one of these is correct?<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>{:db\/id          :person\/email\n :db\/valueType   :db.type\/string\n :db\/cardinality :db.cardinality\/many     ; People can have multiple email addresses.\n :db\/unique      :db.unique\/identity      ; For our purposes, each email identifies one person.\n :db\/index       true}                    ; We want fast lookups by email.\n{:db\/id          :person\/friend\n :db\/valueType   :db.type\/ref\n :db\/cardinality :db.cardinality\/many}    ; People can have many friends.<\/pre><\/div>\n<div class=\"highlight highlight-source-sql\"><pre>CREATE TABLE people (\n  id INTEGER PRIMARY KEY,  -- Bug: because of the primary key, each person can have no more than 1 email.\n  email VARCHAR(64),       -- Bug?: no NOT NULL, so a person can have no email.\n                           -- Bug: nobody will ever have a long email address, right?\n);\nCREATE TABLE friendships (\n  FOREIGN KEY person REFERENCES people(id),  -- Bug?: no indexing, so lookups by friend or person will be slow.\n  FOREIGN KEY friend REFERENCES people(id),  -- Bug: no compound uniqueness constraint, so we can have dupe friendships.\n);<\/pre><\/div>\n<p>They both have limitations \u2014 the Mentat schema allows only for an open world (it's possible to declare friendships with people whose email isn't known), and requires validation code to enforce email string correctness \u2014 but we think that even such a tiny SQL example is harder to understand and obscures important domain decisions.<\/p>\n<\/li>\n<li>\n<p>Queries are intimately tied to structural storage choices. That not only hides the declarative domain-level meaning of the query \u2014 it's hard to tell what a query is trying to do when it's a 100-line mess of subqueries and <code>LEFT OUTER JOIN<\/code>s \u2014 but it also means a simple structural schema change requires auditing <em>every query<\/em> for correctness.<\/p>\n<\/li>\n<li>\n<p>Developers often capture less event-shaped than they perhaps should, simply because their initial requirements don't warrant it. It's quite common to later want to <a href=\"https:\/\/bugzilla.mozilla.org\/show_bug.cgi?id=1341939\" rel=\"nofollow\">know when a fact was recorded<\/a>, or <em>in which order<\/em> two facts were recorded (particularly for migrations), or on which device an event took place\u2026 or even that a fact was <em>ever<\/em> recorded and then deleted.<\/p>\n<\/li>\n<li>\n<p>Common queries are hard. Storing values only once, upserts, complicated joins, and group-wise maxima are all difficult for non-expert developers to get right.<\/p>\n<\/li>\n<li>\n<p>It's hard to evolve storage schemas. Writing a robust SQL schema migration is hard, particularly if a bad migration has ever escaped into the wild! Teams learn to fear and avoid schema changes, and eventually they ship a table called <code>metadata<\/code>, with three <code>TEXT<\/code> columns, so they never have to write a migration again. That decision pushes storage complexity into application code. (Or they start storing unversioned JSON blobs in the database\u2026)<\/p>\n<\/li>\n<li>\n<p>It's hard to share storage with another component, let alone share <em>data<\/em> with another component. Conway's Law applies: your software system will often grow to have one database per team.<\/p>\n<\/li>\n<li>\n<p>It's hard to build efficient storage and querying architectures. Materialized views require knowledge of triggers, or the implementation of bottleneck APIs. <em>Ad hoc<\/em> caches are often wrong, are almost never formally designed (do you want a write-back, write-through, or write-around cache? Do you know the difference?), and often aren't reusable. The average developer, faced with a SQL database, has little choice but to build a simple table that tries to meet every need.<\/p>\n<\/li>\n<\/ul><h2><a id=\"user-content-comparison-to-datascript\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-datascript\"><\/a>Comparison to DataScript<\/h2>\n<p>DataScript asks the question: \"What if creating a database were as cheap as creating a Hashmap?\"<\/p>\n<p>Mentat is not interested in that. Instead, it's strongly interested in persistence and performance, with very little interest in immutable databases\/databases as values or throwaway use.<\/p>\n<p>One might say that Mentat's question is: \"What if an SQLite database could store arbitrary relations, for arbitrary consumers, without them having to coordinate an up-front storage-level schema?\"<\/p>\n<p>(Note that <a href=\"http:\/\/martinfowler.com\/articles\/schemaless\/\" rel=\"nofollow\">domain-level schemas are very valuable<\/a>.)<\/p>\n<p>Another possible question would be: \"What if we could bake some of the concepts of <a href=\"http:\/\/www.baeldung.com\/cqrs-event-sourced-architecture-resources\" rel=\"nofollow\">CQRS and event sourcing<\/a> into a persistent relational store, such that the transaction log itself were of value to queries?\"<\/p>\n<p>Some thought has been given to how databases as values \u2014 long-term references to a snapshot of the store at an instant in time \u2014 could work in this model. It's not impossible; it simply has different performance characteristics.<\/p>\n<p>Just like DataScript, Mentat speaks Datalog for querying and takes additions and retractions as input to a transaction.<\/p>\n<p>Unlike DataScript, Mentat exposes free-text indexing, thanks to SQLite.<\/p>\n<h2><a id=\"user-content-comparison-to-datomic\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-datomic\"><\/a>Comparison to Datomic<\/h2>\n<p>Datomic is a server-side, enterprise-grade data storage system. Datomic has a beautiful conceptual model. It's intended to be backed by a storage cluster, in which it keeps index chunks forever. Index chunks are replicated to peers, allowing it to run queries at the edges. Writes are serialized through a transactor.<\/p>\n<p>Many of these design decisions are inapplicable to deployed desktop software; indeed, the use of multiple JVM processes makes Datomic's use in a small desktop app, or a mobile device, prohibitive.<\/p>\n<p>Mentat was designed for embedding, initially in an experimental Electron app (<a href=\"https:\/\/github.com\/mozilla\/tofino\">Tofino<\/a>). It is less concerned with exposing consistent database states outside transaction boundaries, because that's less important here, and dropping some of these requirements allows us to leverage SQLite itself.<\/p>\n<h2><a id=\"user-content-comparison-to-sqlite\" class=\"anchor\" aria-hidden=\"true\" href=\"#comparison-to-sqlite\"><\/a>Comparison to SQLite<\/h2>\n<p>SQLite is a traditional SQL database in most respects: schemas conflate semantic, structural, and datatype concerns, as described above; the main interface with the database is human-first textual queries; sparse and graph-structured data are 'unnatural', if not always inefficient; experimenting with and evolving data models are error-prone and complicated activities; and so on.<\/p>\n<p>Mentat aims to offer many of the advantages of SQLite \u2014 single-file use, embeddability, and good performance \u2014\u00a0while building a more relaxed, reusable, and expressive data model on top.<\/p>\n<hr \/><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>Please note that this project is released with a Contributor Code of Conduct.\nBy participating in this project you agree to abide by its terms.<\/p>\n<p>See <a href=\"https:\/\/github.com\/mozilla\/mentat\/blob\/master\/CONTRIBUTING.md\">CONTRIBUTING.md<\/a> for further notes.<\/p>\n<p>This project is very new, so we'll probably revise these guidelines. Please\ncomment on an issue before putting significant effort in if you'd like to\ncontribute.<\/p>\n<hr \/><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#building\"><\/a>Building<\/h2>\n<p>You first need to clone the project.  To build and test the project, we are using <a href=\"https:\/\/crates.io\/install\" rel=\"nofollow\">Cargo<\/a>.<\/p>\n<p>To build all of the crates in the project use:<\/p>\n<pre>cargo build\n<\/pre>\n<p>To run tests use:<\/p>\n<pre># Run tests for everything.\ncargo test --all\n# Run tests for just the query-algebrizer folder (specify the crate, not the folder),\n# printing debug output.\ncargo test -p mentat_query_algebrizer -- --nocapture\n<\/pre>\n<p>For most <code>cargo<\/code> commands you can pass the <code>-p<\/code> argument to run the command just on that package. So, <code>cargo build -p mentat_query_algebrizer<\/code> will build just the \"query-algebrizer\" folder.<\/p>\n<h2><a id=\"user-content-what-are-all-of-these-crates\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-are-all-of-these-crates\"><\/a>What are all of these crates?<\/h2>\n<p>We use multiple sub-crates for Mentat for four reasons:<\/p>\n<ol><li>To improve incremental build times.<\/li>\n<li>To encourage encapsulation; writing <code>extern crate<\/code> feels worse than just <code>use mod<\/code>.<\/li>\n<li>To simplify the creation of targets that don't use certain features: <em>e.g.<\/em>, a build with no syncing, or with no query system.<\/li>\n<li>To allow for reuse (<em>e.g.<\/em>, the EDN parser is essentially a separate library).<\/li>\n<\/ol><p>So what are they?<\/p>\n<h3><a id=\"user-content-building-blocks\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-blocks\"><\/a>Building blocks<\/h3>\n<h4><a id=\"user-content-edn\" class=\"anchor\" aria-hidden=\"true\" href=\"#edn\"><\/a><code>edn<\/code><\/h4>\n<p>Our EDN parser. It uses <code>rust-peg<\/code> to parse <a href=\"https:\/\/github.com\/edn-format\/edn\">EDN<\/a>, which is Clojure\/Datomic's richer alternative to JSON. <code>edn<\/code>'s dependencies are all either for representing rich values (<code>chrono<\/code>, <code>uuid<\/code>, <code>ordered-float<\/code>) or for parsing (<code>serde<\/code>, <code>peg<\/code>).<\/p>\n<p>In addition, this crate turns a stream of EDN values into a representation suitable to be transacted.<\/p>\n<h4><a id=\"user-content-mentat_core\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_core\"><\/a><code>mentat_core<\/code><\/h4>\n<p>This is the lowest-level Mentat crate. It collects together the following things:<\/p>\n<ul><li>Fundamental domain-specific data structures like <code>ValueType<\/code> and <code>TypedValue<\/code>.<\/li>\n<li>Fundamental SQL-related linkages like <code>SQLValueType<\/code>. These encode the mapping between Mentat's types and values and their representation in our SQLite format.<\/li>\n<li>Conversion to and from EDN types (<em>e.g.<\/em>, <code>edn::Keyword<\/code> to <code>TypedValue::Keyword<\/code>).<\/li>\n<li>Common utilities (some in the <code>util<\/code> module, and others that should be moved there or broken out) like <code>Either<\/code>, <code>InternSet<\/code>, and <code>RcCounter<\/code>.<\/li>\n<li>Reusable lazy namespaced keywords (<em>e.g.<\/em>, <code>DB_TYPE_DOUBLE<\/code>) that are used by <code>mentat_db<\/code> and EDN serialization of core structs.<\/li>\n<\/ul><h3><a id=\"user-content-types\" class=\"anchor\" aria-hidden=\"true\" href=\"#types\"><\/a>Types<\/h3>\n<h4><a id=\"user-content-mentat_query\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_query\"><\/a><code>mentat_query<\/code><\/h4>\n<p>This crate defines the structs and enums that are the output of the query parser and used by the translator and algebrizer. <code>SrcVar<\/code>, <code>NonIntegerConstant<\/code>, <code>FnArg<\/code>\u2026 these all live here.<\/p>\n<h4><a id=\"user-content-mentat_query_sql\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_query_sql\"><\/a><code>mentat_query_sql<\/code><\/h4>\n<p>Similarly, this crate defines an abstract representation of a SQL query as understood by Mentat. This bridges between Mentat's types (<em>e.g.<\/em>, <code>TypedValue<\/code>) and SQL concepts (<code>ColumnOrExpression<\/code>, <code>GroupBy<\/code>). It's produced by the algebrizer and consumed by the translator.<\/p>\n<h3><a id=\"user-content-query-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#query-processing\"><\/a>Query processing<\/h3>\n<h4><a id=\"user-content-mentat_query_algebrizer\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_query_algebrizer\"><\/a><code>mentat_query_algebrizer<\/code><\/h4>\n<p>This is the biggest piece of the query engine. It takes a parsed query, which at this point is <em>independent of a database<\/em>, and combines it with the current state of the schema and data. This involves translating keywords into attributes, abstract values into concrete values with a known type, and producing an <code>AlgebraicQuery<\/code>, which is a representation of how a query's Datalog semantics can be satisfied as SQL table joins and constraints over Mentat's SQL schema. An algebrized query is tightly coupled with both the disk schema and the vocabulary present in the store when the work is done.<\/p>\n<h4><a id=\"user-content-mentat_query_projector\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_query_projector\"><\/a><code>mentat_query_projector<\/code><\/h4>\n<p>A Datalog query <em>projects<\/em> some of the variables in the query into data structures in the output. This crate takes an algebrized query and a projection list and figures out how to get values out of the running SQL query and into the right format for the consumer.<\/p>\n<h4><a id=\"user-content-mentat_query_translator\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_query_translator\"><\/a><code>mentat_query_translator<\/code><\/h4>\n<p>This crate works with all of the above to turn the output of the algebrizer and projector into the data structures defined in <code>mentat_query_sql<\/code>.<\/p>\n<h4><a id=\"user-content-mentat_sql\" class=\"anchor\" aria-hidden=\"true\" href=\"#mentat_sql\"><\/a><code>mentat_sql<\/code><\/h4>\n<p>This simple crate turns those data structures into SQL text and bindings that can later be executed by <code>rusqlite<\/code>.<\/p>\n<h3><a id=\"user-content-the-data-layer-mentat_db\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-data-layer-mentat_db\"><\/a>The data layer: <code>mentat_db<\/code><\/h3>\n<p>This is a big one: it implements the core storage logic on top of SQLite. This crate is responsible for bootstrapping new databases, transacting new data, maintaining the attribute cache, and building and updating in-memory representations of the storage schema.<\/p>\n<h3><a id=\"user-content-the-main-crate\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-main-crate\"><\/a>The main crate<\/h3>\n<p>The top-level main crate of Mentat assembles these component crates into something useful. It wraps up a connection to a database file and the associated metadata into a <code>Store<\/code>, and encapsulates an in-progress transaction (<code>InProgress<\/code>). It provides modules for programmatically writing (<code>entity_builder.rs<\/code>) and managing vocabulary (<code>vocabulary.rs<\/code>).<\/p>\n<h3><a id=\"user-content-syncing\" class=\"anchor\" aria-hidden=\"true\" href=\"#syncing\"><\/a>Syncing<\/h3>\n<p>Sync code lives, for <a href=\"https:\/\/engineering.linkedin.com\/distributed-systems\/log-what-every-software-engineer-should-know-about-real-time-datas-unifying\" rel=\"nofollow\">referential reasons<\/a>, in a crate named <code>tolstoy<\/code>. This code is a work in progress; current state is a proof-of-concept implementation which largely relies on the internal transactor to make progress in most cases and comes with a basic support for timelines. See <a href=\"https:\/\/github.com\/mozilla\/mentat\/tree\/master\/tolstoy\/README.md\">Tolstoy's documentation<\/a> for details.<\/p>\n<h3><a id=\"user-content-the-command-line-interface\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-command-line-interface\"><\/a>The command-line interface<\/h3>\n<p>This is under <code>tools\/cli<\/code>. It's essentially an external consumer of the main <code>mentat<\/code> crate. This code is ugly, but it mostly works.<\/p>\n<hr \/><h2><a id=\"user-content-sqlite-dependencies\" class=\"anchor\" aria-hidden=\"true\" href=\"#sqlite-dependencies\"><\/a>SQLite dependencies<\/h2>\n<p>Mentat uses partial indices, which are available in SQLite 3.8.0 and higher. It relies on correlation between aggregate and non-aggregate columns in the output, which was added in SQLite 3.7.11.<\/p>\n<p>It also uses FTS4, which is <a href=\"http:\/\/www.sqlite.org\/fts3.html#section_2\" rel=\"nofollow\">a compile time option<\/a>.<\/p>\n<p>By default, Mentat specifies the <code>\"bundled\"<\/code> feature for <code>rusqlite<\/code>, which uses a relatively recent\nversion of SQLite. If you want to link against the system version of SQLite, omit <code>\"bundled_sqlite3\"<\/code>\nfrom Mentat's features.<\/p>\n<div class=\"highlight highlight-source-toml\"><pre>[dependencies.mentat]\nversion = \"0.6\"\n# System sqlite is known to be new.\ndefault-features = false<\/pre><\/div>\n<hr \/><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>Project Mentat is currently licensed under the Apache License v2.0. See the <code>LICENSE<\/code> file for details.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:50+0000","updated_at":"2018-11-07T16:22:50+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/131524?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":115,"title":"Jasper - A flexible and powerful issue reader for GitHub","url":"https:\/\/jasperapp.io\/","content":"<div class=\"overview\" id=\"overview\"><p>\n    You will be able to read, track and discover many active issues very flexibly by using Jasper in GitHub.\n    It is enabled by a powerful function called \"stream\" of Jasper.\n  <\/p><p class=\"stream-sample\">\n    example\n    org:electron is:issue label:bug\n  <\/p><\/div><div class=\"content\"><table class=\"compare-table\"><caption>A tracking condition comparison Jasper(<a href=\"https:\/\/jasperapp.io\/doc.html#stream\">streams<\/a>) and GitHub(<a href=\"https:\/\/help.github.com\/articles\/about-noble claion-emails\/\">email<\/a>)<\/caption>\n    <thead><tr><td>Tracking conditions<\/td><td>Jasper<\/td><td>GitHub<\/td><\/tr><\/thead><tbody><tr><td>Repositories <br \/>belongs to organizations or users<\/td><td>Any repositories<\/td><td>Any repositories<\/td><\/tr><tr><td>State <br \/>open, closed, approved and request changed<\/td><td>Any state<\/td><td>Only related to myself<\/td><\/tr><tr><td>Involves <br \/>assignees, authors, mentions, and comments<\/td><td>Any users<\/td><td>Only related to myself<\/td><\/tr><tr><td>Team Mentions <br \/>accessible teams<\/td><td>Any teams<\/td><td>Only related to my teams<\/td><\/tr><tr><td>Organizations <br \/>organizations and users<\/td><td>Any organizations<\/td><td>-<\/td><\/tr><tr><td>Labels <br \/>bug, enhancement, invalid and more<\/td><td>Any labels<\/td><td>-<\/td><\/tr><tr><td>Milestones <br \/>v0.1.0 and v1.0.0 and v2.0.0-alpha and more<\/td><td>Any milestones<\/td><td>-<\/td><\/tr><tr><td>Language <br \/>JavaScript, Ruby, Go and more<\/td><td>Any language<\/td><td>-<\/td><\/tr><tr><td>Type <br \/>issue and\/or pull request<\/td><td>Any types<\/td><td>-<\/td><\/tr><tr><td>Egosurfing <br \/>own libraries and like tools and more<\/td><td>Any words<\/td><td>-<\/td><\/tr><\/tbody><\/table><div class=\"users\"><p>Hundreds of developers are using Jasper in a business\/OSS activity. <br \/>And it is also used by several people in GitHub inc.<\/p><\/div><\/div>","created_at":"2018-11-07T16:22:50+0000","updated_at":"2018-11-07T16:22:50+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"jasperapp.io","preview_picture":"https:\/\/jasperapp.io\/image\/logo_full.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":116,"title":"Muzoti - Algorithmic Composition of Music","url":"https:\/\/www.muzoti.com\/","content":"<img class=\"silhouette\" src=\"https:\/\/www.muzoti.com\/static\/img\/mozart.png\" alt=\"image\" \/><p>\n\t\t\t\t\tMuzoti is the fruit of a 27 year personal obsession into what makes music sound, well... musical.\n\t\t\t\t<\/p><p>\n\t\t\t\t\tA music score can, to some degree, be reduced to a sequence of numbers, so why is it that one sequence sounds good, yet another sequence of randomly generated notes sounds terrible? What inherent structure in a musical pattern that makes it sound good.\n\t\t\t\t<\/p><p>\n\t\t\t\t\tThis is the question that I longed to answer. As the years went by I constantly tested my theories. I wrote programs that generated 'music' according the organising principles that I observed. The more musical the result, the more sure I became that I was on the right track...\n\t\t\t\t<\/p><p>\n\t\t\t\t\tEventually it came down to this: if our brain has evolved to predict patterns in space and time, and rewards us when we get it right, what would happen if you turned it on it's head? What would happen if you tried to maximally stimulate the brain's reward system by showing it generated patterns that alternate between repeating and varying material, not just on one, but on many temporal scales? If those patterns were patterns of notes, would that sound like music?  \n\t\t\t\t<\/p><p>\n\t\t\t\t\tMuzoti is an exploration of that thesis, and poses a fundamental question. Is music, or at least a substantial aspect of it, simply a by product of the way the brain has evolved, rather than a purely cultural phenomenon?\n\t\t\t\t<\/p><p>\n\t\t\t\t\tUnlike other generative music attempts, we're not using machine learning. Instead we are developing a solid basis for generative musical intelligence based on what we believe are universal organising principles of music and life. Our algorithms are style agnostic. Without such a solid framework, machine learning can only extract style information based on pre-comosed works. We believe machine learning, while valuable, should come second to a holistic theory of what music is and why it exists.\n\t\t\t\t<\/p>","created_at":"2018-11-07T16:22:50+0000","updated_at":"2018-11-07T16:22:50+0000","annotations":[],"mimetype":"text\/html","reading_time":1,"domain_name":"www.muzoti.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["keyboard"],"is_public":false,"id":117,"title":"ErgoDox EZ: An Incredible Mechanical Ergonomic Keyboard","url":"https:\/\/ergodox-ez.com\/","content":"<p>The ErgoDox EZ, in my opinion, is the most versatile keyboard on the market. It can do anything you need it to do, and if it\u2019s not right by default, you have unlimited options to create the perfect solution for your individual needs. <a href=\"https:\/\/thenextweb.com\/insider\/2016\/03\/03\/why-this-300-ergonomic-keyboard-might-actually-be-worth-the-money\/\" class=\"link\" target=\"_blank\">Read the full review<\/a><\/p>","created_at":"2018-11-07T16:22:50+0000","updated_at":"2019-01-21T21:54:58+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"ergodox-ez.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":109,"title":"Parinfer - simpler Lisp editing","url":"https:\/\/shaunlebron.github.io\/parinfer\/","content":"<div class=\"wrapper only-mobile\"><section><p><a href=\"http:\/\/twitter.com\/shaunlebron\">@shaunlebron<\/a> |\n    <a href=\"https:\/\/github.com\/shaunlebron\/parinfer\">GitHub<\/a><\/p><p><i class=\"fa fa-desktop fa-lg\">\n<strong>Best viewed<\/strong> on wide desktop browsers.  Sorry for the inconvenience.<\/i><\/p><\/section><\/div><div class=\"wrapper\"><section><p class=\"subtitle\">\nLet's simplify the way we write Lisp...\n<\/p><p><i class=\"fa fa-keyboard-o fa-lg\">\n<strong>Try it!<\/strong> Interrupt the animations below to try it for yourself. Click outside to restore it.<\/i><\/p><div><p><strong>Parinfer's \"Indent Mode\"<\/strong> rearranges parens when you change indentation:<\/p><\/div><div><p><strong>Insert or delete a line<\/strong> without rearranging parens:<\/p><\/div><div><p><strong>Comment a line<\/strong> without rearranging parens:<\/p><\/div><div><p><strong>Basic Paredit without hotkeys<\/strong> thanks to simple paren inference. (wrap, splice, slurp, barf)<\/p><\/div><div><p><strong>Preserve indentation<\/strong> when editing in \"Paren Mode\":<\/p><\/div><\/section><\/div><div class=\"wrapper fold\"><section><p>\nParinfer is a proof-of-concept editor mode for Lisp\nprogramming languages.  It simplifies the way we write Lisp by auto-adjusting\nparens when indentation changes and vice versa. The hope is to make basic\nLisp-editing easier for newcomers and experts alike, while still allowing\nexisting plugins like Paredit to satisfy the need for more advanced operations.\n<\/p><ul><li>Source code: <a href=\"http:\/\/github.com\/shaunlebron\/parinfer\">http:\/\/github.com\/shaunlebron\/parinfer<\/a><br \/><\/li><li>Created by: <a href=\"http:\/\/twitter.com\/shaunlebron\">@shaunlebron<\/a><br \/><\/li><li>Download (work in progress): <a href=\"#editor-plugins\">Editor Plugins<\/a>\n<\/li><\/ul><p><i class=\"fa fa-book fa-lg\">\n<strong>Read on<\/strong> for a full exploration of the motivations, rules, and effects of Parinfer.<\/i><\/p><\/section><\/div><div class=\"wrapper\"><section id=\"uncovering-lisp\"><h3>Uncovering Lisp<\/h3><p>\nMost programming languages have several syntax rules.  Lisp has one:\neverything is a list.  The first element is a function name, and the rest are\nits arguments.  Thus, the language is simply a collection of compile- and\nrun-time functions, trivially extensible.\n<\/p><div class=\"two-col\"><div class=\"col\"><p><strong>C Style<\/strong> expressions:<\/p><textarea id=\"code-c-expr\" rows=\"10\" cols=\"50\">\nfoo(1, 2, 3);\nconst a = 1;\nif (a == 1) foo(); else bar();\nconst b = (a == 1) ? 2 : 3;\n<\/textarea><\/div><div class=\"col\"><p><strong>Lisp Style<\/strong> expressions:<\/p><textarea id=\"code-lisp-expr\" rows=\"10\" cols=\"50\">\n(foo 1 2 3)\n(def a 1)\n(if (= a 1) (foo) (bar))\n(def b (if (= a 1) 2 3))\n<\/textarea><\/div><\/div><p>\nBut parentheses in Lisp are infamous for bunching together at the end of long\nexpressions. This indentation\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Indent_style#Lisp_style\">convention<\/a>\ncan be jarring at first if you are used to curly braces in other languages\nbeing on their own lines:\n<\/p><div class=\"two-col\"><div class=\"col\"><p><strong>C Style<\/strong> indentation (unusual in Lisp):<\/p><textarea id=\"code-c-indent\" rows=\"10\" cols=\"50\">\n(defn foo [a b]\n  (let [\n     x (+ a b)\n    ]\n    (println \"The sum is\" x)\n  )\n)\n<\/textarea><\/div><div class=\"col\"><p><strong>Lisp Style<\/strong> indentation:<\/p><textarea id=\"code-lisp-indent\" rows=\"10\" cols=\"50\">\n(defn foo [a b]\n  (let [x (+ a b)]\n    (println \"The sum is\" x)))\n<\/textarea><\/div><\/div><p>\nThe idea behind this convention is to make every line inform with content\nrather than just parens. Readability is helped by employing a Python-like\nindentation style.  This achieves a sort of balance\u2014 Indentation allows\nyou to <em>skim<\/em> while the parens allow you to <em>inspect<\/em>:<\/p><div class=\"two-col\"><div class=\"col\"><p><strong>Skim<\/strong> by focusing on indentation<\/p><textarea id=\"code-skim\" rows=\"10\" cols=\"50\">\n(defn foo [a b]\n  (let [x (+ a b)]\n    (println \"The sum is\" x)))\n<\/textarea><\/div><div class=\"col\"><p><strong>Inspect<\/strong> parens with your cursor when needed<\/p><textarea id=\"code-inspect\" rows=\"10\" cols=\"50\">\n(defn foo [a b]\n  (let [x (+ a b)]\n    (println \"The sum is\" x)))\n<\/textarea><\/div><\/div><p>\nThough both <em>perspectives<\/em> are visible at once, we must focus on one at\na time.  A\n<a href=\"https:\/\/twitter.com\/paulg\/status\/588511912331055104\">LEGO analogy<\/a>\nhelps here. Imagine each list in the previous example as a LEGO block\nstacked over its parent. Checking the sides to see the layers below is like\nchecking the parens at the end of a line.\n<\/p><div class=\"two-col\"><div class=\"col\"><p><strong>Indentation<\/strong> implies nesting.<\/p><img class=\"lego-img\" src=\"https:\/\/shaunlebron.github.io\/parinfer\/img\/lego-indent.png\" alt=\"image\" \/><\/div><div class=\"col\"><p><strong>Tilting<\/strong> clarifies nesting (close-parens shown).<\/p><img class=\"lego-img\" src=\"https:\/\/shaunlebron.github.io\/parinfer\/img\/lego-parens.png\" alt=\"image\" \/><\/div><\/div><p>\nThis is a physical analog to the way we read Lisp code.  Now let's look at the\nspace of tooling solutions that we use for <em>writing<\/em> Lisp code and\nspecifically how Parinfer can help.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"tools-for-writing-lisp\"><h3>Tools for Writing Lisp<\/h3><p>\nIn the previous section, we saw how Parens and Indentation provide two\nperspectives for reading Lisp code.  Unfortunately, this incurs some redundancy\nwhen <em>writing<\/em> Lisp since we must edit both for every change, ensuring\nthat one will correctly imply the other.\n<\/p><p>\nThere are existing tools to help with this.  To best represent how Parinfer\ncompares to them, we will represent this complex space of tooling in a way that\ncan be visually compared below.\n<\/p><p><i class=\"fa fa-gear fa-lg\">\n<strong>The black gear<\/strong> represents the current action we are taking.<\/i><\/p><p>\n<strong>The menial and default<\/strong> way to edit Lisp is to manually ensure\nour parens are balanced after inserting or removing them.  Then, we adjust the\nindentation to match it in kind.  This back-and-forth happens for most editing\ntasks.\n<\/p><p>\n<strong>Existing tools automate<\/strong> some of these editing tasks.\nFor example, <em>Paredit<\/em> forces you to transform or add parens in a\nbalanced way through special hotkeys.  And <em>Auto-indent<\/em> allows you to\nauto-correct indentation of selected lines when desired.  This automates\nthe tasks, but the back-and-forth actions are still manually triggered.<\/p><p>\n<strong>Parinfer is a new tool to <em>combine and simplify<\/em><\/strong> this\ntype of automation by naturally keeping Parens and Indentation in lockstep.  It\nformally <em>infers<\/em> changes to one based on the other.  The back-and-forth\nactions have been reduced with special modes, which we will explore next.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"writing-with-parinfer\"><h3>Writing with Parinfer<\/h3><p>\nAs we saw with the previous visualization, Parinfer will infer some changes to\nkeep Parens and Indentation inline with one another.  To keep this inference\nsimple and predictable, we have the user explicitly choose the degree of\nfreedom that <em>they<\/em> want full control of, while relinquishing some\ncontrol of the other to <em>Parinfer<\/em>.\n<\/p><p>\nThus, Parinfer consists of two modes:\n<\/p><ol><li>\n<strong>Indent Mode<\/strong> gives you full control of indentation, while\nParinfer corrects parens.<br \/><\/li>\n<li>\n<strong>Paren Mode<\/strong> gives you full control of parens, while Parinfer\ncorrects indentation.<br \/><\/li>\n<\/ol><p><i class=\"fa fa-lightbulb-o fa-lg\">\n<strong>To make this easier<\/strong>, Indent Mode is default. Paren Mode can be an advanced option.<\/i><\/p><p>\nSome noteworthy unintended use-cases which we will explore later:\n<\/p><ul><li>Indent Mode allows Paredit-like features without hotkeys.<\/li>\n  <li>Paren Mode can be used to fix incorrectly indented files before using with Indent Mode.<\/li>\n<\/ul><\/section><\/div><div class=\"wrapper\"><section id=\"mathematical-foundation\"><h3>Mathematical Foundation<\/h3><p><i class=\"fa fa-rocket fa-lg\">\n<strong>Feel free to <a href=\"#indent-mode\">skip this<\/a><\/strong> if you have no interest in the math that makes Parinfer possible.<\/i><\/p><p>\nThe foundation of Parinfer relies on a few somewhat formalized definitions and\nproperties:\n<\/p><div class=\"math-page mathjax\"><p>\nWe wish to define properties that each Parinfer mode should exhibit.\n<\/p><p>\nWith the following definitions:\n<\/p><ul><li>$I$ = (Indent Mode) function accepting code text and returning new text.<\/li>\n  <li>$P$ = (Paren Mode) function accepting code text and returning new text.<\/li>\n  <li>$R$ = (Reader) function accepting code text and returning its AST data.<\/li>\n<\/ul><p>\nThe following should be true:\n<\/p><ul><li>let $y = P(x)$\n        <p>$\\leftarrow$ run Paren-Mode on $x$ to get $y$<\/p>\n    <ul><li>$\\Longrightarrow R(x) = R(y)$\n        <p>$\\leftarrow$ Paren-Mode should never change the AST<\/p>\n      <\/li><li>$\\Longrightarrow P(y) = y$\n        <p>$\\leftarrow$ Paren-Mode should be idempotent<\/p>\n      <\/li><li>$\\Longrightarrow I(y) = y$\n        <p>$\\leftarrow$ Indent-Mode should never change the result of Paren-Mode<\/p>\n    <\/li><\/ul><\/li><\/ul><ul><li>let $z = I(x)$\n        <p>$\\leftarrow$ run Indent-Mode on $x$ to get $z$<\/p>\n    <ul><li>$\\mathrel{\\rlap{\\hskip .5em\/}}\\Longrightarrow R(x) = R(z)$\n        <p>$\\leftarrow$ Indent-Mode may change the AST (by design)<\/p>\n      <\/li><li>$\\Longrightarrow I(z) = z$\n        <p>$\\leftarrow$ Indent-Mode should be idempotent<\/p>\n    <\/li><\/ul><\/li><\/ul><\/div><p class=\"mathjax\">\nThe actual operations performed by the modes rely on a formal definition\nof what it means for Lisp code to be \"correctly formatted\".  We establish an\ninvariant\u2014 something that must be true for every line of code.  From that,\nParinfer corrects indentation or parens simply by choosing correct values for\n$i_n$ or $p_n$ (defined later) to satisfy this invariant:\n<\/p><ul class=\"mathjax\"><li><strong>Indent Mode<\/strong> forces correct values of $p_n$<\/li>\n  <li><strong>Paren Mode<\/strong> forces correct values of $i_n$<\/li>\n<\/ul><p>\nThe following is a concise reference (not a guide) to establishing this\ninvariant that Parinfer's modes rely on.\n<\/p><div class=\"math-page mathjax\"><p>\nWe wish to define necessary conditions for determining if a given file of Lisp\ncode is \"correctly formatted\".\n<\/p><p>\nWe start with a clarification that we only consider non-empty lines (i.e. lines\nthat have at least one non-whitespace, non-comment token).  Thus, all following\nreferences to line number $n$ will refer to the $n$th non-empty line.\n<\/p><p>\nTo proceed, we define the following:\n<\/p><ul><li>\n$t_n$ = (tokens) index of non-whitespace, non-comment tokens at line $n$\n<\/li>\n<li>\n$i_n$ = (indentation) x-position of $t_n[0]$\n<\/li>\n<li>\n$p_n$ = (parens) number of close-parens at the end of $t_n$\n<\/li>\n<li>\n$s_n$ = (stack) index of x-positions of all open-parens unclosed before $t_n[-p_n]$\n<\/li>\n<\/ul><p>\nNext, we define a function which determines if a line's indentation is valid:\n<\/p><ul><li>$f(i,p,s) = ( x_\\text{min} \\lt i \\leq x_\\text{max} )$\n    <p>$\\leftarrow$ verifies indentation is within a threshold defined below<\/p>\n  <\/li>\n<\/ul><p>\nwhere:\n<\/p><ul><li>$x_\\text{min} = \\begin{cases} s[-p-1] &amp; \\text{if } |s| &gt; p, \\\\ -1 &amp; \\text{otherwise} \\end{cases}$\n    <p>$\\leftarrow$ x-position of open-paren of parent list if it exists<\/p>\n  <\/li>\n  <li>$x_\\text{max} = \\begin{cases} s[-p] &amp; \\text{if } p \\gt 0, \\\\ \\infty &amp; \\text{otherwise} \\end{cases}$\n    <p>$\\leftarrow$ x-position of open-paren of previous sibling if it is a list<\/p>\n  <\/li>\n<\/ul><p>\nFinally, we define that a file of Lisp code is \"correctly formatted\" if:\n<\/p><ol><li>for every line $n &gt; 0$:\n    <ul><li>$t_n[0] \\ne $ \")\", and\n        <p>$\\leftarrow$ no line may start with a close-paren<\/p>\n      <\/li>\n      <li>$f(i_n, p_{n-1}, s_{n-1})$ is true\n        <p>$\\leftarrow$ indentation threshold that we defined above<\/p>\n      <\/li>\n    <\/ul><\/li>\n  <li>and for the last line $N$\n    <ul><li>$|s_N| = p_N$\n        <p>$\\leftarrow$ all open-parens must be closed at the last line<\/p>\n      <\/li>\n    <\/ul><\/li>\n<\/ol><\/div><p>\nThese rules are necessary and sufficient for determining when indentation is\nwhat we consider \"unambiguous\", but they are not sufficient in determining if\ncode is \"pretty\". For example:\n<\/p><ul><li>single-line files are okay<\/li>\n<li>\"tab\" lengths are flexible (one space, two spaces, etc)<\/li>\n<li>indentation of sibling lines are not required to be aligned<\/li>\n<\/ul><p>\nFormal descriptions of the actual operations performed by the modes are\npending, but informal ones follow in their respective sections below.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"indent-mode\"><h2>Indent Mode<\/h2><p>\nIndent Mode gives you full control of indentation, while Parinfer corrects or\ninserts close-parens where appropriate.  Specifically, it only touches the\ngroups of close-parens at the end of each line.  As a visual cue, we\n<em>slightly dim<\/em> these parens to signify their inferred nature.\n<\/p><p><i class=\"fa fa-keyboard-o fa-lg\">\n<strong>Try it!<\/strong> Interrupt the animations below to try it for yourself. Click outside to restore it.<\/i><\/p><div><p><strong>Indent<\/strong> to influence the structure of your code:<\/p><textarea id=\"code-indent\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Indent further<\/strong> to reach different thresholds:<\/p><textarea id=\"code-indent-far\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Indent multiple lines<\/strong> to see its effect:<\/p><textarea id=\"code-indent-multi\" rows=\"10\" cols=\"50\"><\/textarea><\/div><p>\nYou can select multiple lines and adjust their indentation the standard way using\nthe controls below.  If you are familiar with Paredit, these operations are\nroughly equivalent to those listed.\n<\/p><table class=\"paredit-table\"><tr><th>Controls<\/th>\n<th>Description<\/th>\n<th>Paredit equivalent<\/th>\n<\/tr><tr><td><kbd class=\"light\">Tab<\/kbd><\/td>\n<td>indent line(s)<\/td>\n<td>slurp line(s) down<\/td>\n<\/tr><tr><td><kbd class=\"light\">Shift<\/kbd> + <kbd class=\"light\">Tab<\/kbd><\/td>\n<td>dedent line(s)<\/td>\n<td>barf line(s) down<\/td>\n<\/tr><\/table><\/section><\/div><div class=\"wrapper\"><section id=\"interesting-consequences\"><h3>Interesting Consequences<\/h3><div><p><strong>Insert or delete a line<\/strong> without rearranging parens:<\/p><textarea id=\"code-line\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Comment a line<\/strong> without rearranging parens:<\/p><\/div><\/section><\/div><div class=\"wrapper\"><section id=\"indent-how-it-works\"><h3>How it works<\/h3><p>\nWe perform the following steps to rearrange close-parens based on indentation.<br \/>We will refer to these later as rules #1, #2, #3 and #4.\n<\/p><ol><li> remove all unmatched close-parens (for housekeeping)\n<\/li><li> remove all close-parens at the start and end of each line\n<\/li><li> keep all close-parens <em>inside<\/em> each line\n<\/li><li> for every resulting unmatched open-paren:\n  <ul><li> insert a matching close-paren at the end of its line or its last non-empty indented line\n  <\/li><\/ul><\/li><\/ol><p><i class=\"fa fa-keyboard-o fa-lg\">\n<strong>Try it!<\/strong> Edit the code below on the left to see how parens are inferred on the right.<\/i><\/p><div class=\"two-col\"><div class=\"col\"><p><strong>Input:<\/strong> close-parens are removed or kept.<\/p><textarea id=\"code-indent-input\" rows=\"10\" cols=\"50\">\n(defn on-click []\n  (swap! s update-in [:x] inc)\n  )\n(defn component\n  \"my docstring.\"\n  [messages]\n  [:div]\n   (for [[i msg] messages]\n     [:div msg])\n  )\n<\/textarea><\/div><div class=\"col\"><p><strong>Output:<\/strong> close-parens are inserted.<\/p><textarea id=\"code-indent-output\" rows=\"10\" cols=\"50\"><\/textarea><\/div><\/div><p>\nThis is the gist of what's happening.  There are more steps performed, but we\nwill just explore their effects in the next section.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"paredit-emerges\"><h3>Paredit emerges<\/h3><p>\nYou should be aware that the steps in the previous section have a side effect on\nwhat you type.  Interestingly, these effects translate into four of the main\n<a href=\"https:\/\/shaunlebron.github.io\/parinfer\/\">Paredit<\/a>\n operations.\n<\/p><table class=\"paredit-table light-keys\"><tr><th align=\"right\">Cause<\/th>\n<th align=\"center\">Effect<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><tr><td align=\"right\">Insert <kbd>(<\/kbd><\/td>\n<td align=\"center\">Wrap<\/td>\n<td align=\"left\">\ninserts a matching <kbd>)<\/kbd> as far as it can\n<p>i.e. \"wraps\" all possible elements to the right of your cursor<\/p>\n<\/td>\n<\/tr><tr><td align=\"right\">Insert <kbd>)<\/kbd><\/td>\n<td align=\"center\">Barf<\/td>\n<td align=\"left\">\nremoves the original <kbd>)<\/kbd> when inserted inside a matching pair\n<p>i.e. the current list \"barfs\" out all elements to the right of your cursor<\/p>\n<\/td>\n<\/tr><tr><td align=\"right\">Delete <kbd>(<\/kbd><\/td>\n<td align=\"center\">Splice<\/td>\n<td align=\"left\">\nremoves the matching <kbd>)<\/kbd>\n<p>i.e. \"splices\" the current list into its parent (or simply \"unwraps\" it)<\/p>\n<\/td>\n<\/tr><tr><td align=\"right\">Delete <kbd>)<\/kbd><\/td>\n<td align=\"center\">Slurp<\/td>\n<td align=\"left\">\ninserts another <kbd>)<\/kbd> as far as it can\n<p>i.e. the current list \"slurps\" all elements to the right of your cursor<\/p>\n<\/td>\n<\/tr><\/table><p>\nWe illustrate these operations in the following examples.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"inserting-parens\"><h3>Inserting Parens<\/h3><div><p><strong>Wrap<\/strong> by inserting an open-paren. It will auto-close as far as it can, due to rule #4.<\/p><textarea id=\"code-wrap\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Barf<\/strong> by inserting a close-paren before another.\nNotice the original is removed, due to rule #1.<\/p><textarea id=\"code-barf\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div class=\"caption\"><p><i class=\"fa fa-question-circle\">\nWhy can't I insert a close-paren in certain places?<\/i><\/p><p>Its corresponding open-paren must be there first. (see rule #1)<\/p><\/div><\/section><\/div><div class=\"wrapper\"><section id=\"deleting-parens\"><h3>Deleting Parens<\/h3><div><p><strong>Splice<\/strong> by removing an open-paren. Its corresponding close-paren is removed, due to rule #1.<\/p><textarea id=\"code-splice\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Slurp<\/strong> by deleting a close-paren inside a line. It is replaced further down, due to rule #4.<\/p><textarea id=\"code-slurp\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div class=\"caption\"><p><i class=\"fa fa-question-circle\">\nWhy can't I delete a close-paren in certain places?<\/i><\/p><p>You cannot delete an inferred close-paren. It is replaced as soon as you delete it. (see rule #4)<\/p><\/div><\/section><\/div><div class=\"wrapper\"><section id=\"knowing-when-parens-move-in-indent-mode\"><h3>Knowing When Parens Move<\/h3><p>\nAs a courtesy, <em>Indent Mode<\/em> will not move your parens until you are done typing\nin front of them.  Just move your cursor away when you're done.  A helpful\nanalogy might be to think of your cursor as a <em>paperweight<\/em> that keeps your\nparens from blowing away.\n<\/p><div><p><strong>Paren displaced<\/strong> when your cursor moves to another line. (displaced due to indentation)<\/p><textarea id=\"code-displaced\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Paren not displaced<\/strong> since you were given the chance to block it. (paren not at end of line)<\/p><textarea id=\"code-not-displaced\" rows=\"10\" cols=\"50\"><\/textarea><\/div><\/section><\/div><div class=\"wrapper\"><section id=\"inserting-quotes\"><h3>Inserting Quotes<\/h3><p>\n<em>Parinfer<\/em> cannot infer anything about quote positions like it can with parens.\nSo it doesn't try to do anything special with them, other than abandon\nprocessing if imbalanced quotes are detected.\n<\/p><div><p><strong>Quote<\/strong> insertion allows temporary paren imbalances until quote is closed:<\/p><textarea id=\"code-string\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div class=\"warning\"><p><i class=\"fa fa-lg fa-warning\">\nWARNING: Always make sure quotes are balanced when inside comments!<\/i><\/p><p>If there is an unclosed quote before a comment, which itself contains\nimbalanced quotes, they will balance each other out and fool <em>Parinfer<\/em> into\nthinking it is okay for processing.<\/p><\/div><div class=\"two-col\"><div class=\"col\"><p>\n<strong>BAD:<\/strong> An unclosed string in a comment can cause corrupted strings.<\/p><textarea id=\"code-warn-bad\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div class=\"col\"><p><i class=\"fa fa-lg fa-check green\">\n<strong>GOOD<\/strong>: Balance the quotes in the comment to prevent the problem.<\/i><\/p><textarea id=\"code-warn-good\" rows=\"10\" cols=\"50\"><\/textarea><\/div><\/div><\/section><\/div><div class=\"wrapper\"><section id=\"paren-mode\"><h2>Paren Mode<\/h2><p>\nParen Mode gives you full control of parens, while Parinfer corrects indentation.\nYou can still adjust indentation, but you won't be able indent\/dedent past certain\nboundaries set by parens on previous lines.  As a courtesy, this mode also\nmaintains relative indentation of child elements when their parent expressions\nshift.\n<\/p><p>\nHere are some things that cannot be done in Indent Mode:\n<\/p><div><p><strong>Tune indentation<\/strong> without worrying about crossing a paren boundary:<\/p><textarea id=\"code-paren-tune\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Avoid fracturing<\/strong> a multi-line expression when pushing its open-paren forward:<\/p><textarea id=\"code-paren-frac\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Indentation is maintained<\/strong> when parens shift, ensuring reversible operations:<\/p><\/div><div><p><strong>Nested expressions<\/strong> become automatically indented (temporary imbalances allowed):<\/p><textarea id=\"code-paren-wrap\" rows=\"10\" cols=\"50\"><\/textarea><\/div><\/section><\/div><div class=\"wrapper\"><section id=\"paren-how-it-works\"><h3>How it works<\/h3><p>\nParen Mode performs the following steps:\n<\/p><ol><li> Move close-parens at the start of a line to the end of the previous non-empty line.\n<\/li><li> Clamp the indentation of a line to the following range:\n  <ul><li> min: x-position of the parent open-paren (if it exists)\n    <\/li><li> max: x-position of the open-paren belonging to the previous non-empty line's last close-paren\n  <\/li><\/ul><\/li><li> Child elements of moved expressions should maintain their original relative indentation to them.\n<\/li><li> Cancel processing if there are any unmatched parens.\n<\/li><\/ol><\/section><\/div><div class=\"wrapper\"><section id=\"switching-modes\"><h3>Switching Modes<\/h3><p>\nIf there are paren imbalances in Paren Mode, the code is not processed, and you\nare prevented from switching to Indent Mode.  This safely quarantines the\nimbalances that could be misinterpreted in Indent Mode.  Thus, Paren Mode gives\nyou an environment to fix them, after which the code is automatically formatted\nand ready for Indent Mode should you choose to switch.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"fixing-existing-files\"><h3>Fixing existing files<\/h3><p>\nWe must take parens literally when opening an existing file. Incidentally,\nParen Mode is perfect for this job, so we preprocess existing files with it\nbefore allowing the switch to Indent Mode.\n<\/p><p><i class=\"fa fa-keyboard-o fa-lg\">\n<strong>Try it!<\/strong> Edit the code below on the left to see how it is formatted on the right.<\/i><\/p><div class=\"two-col\"><div class=\"col\"><p><strong>Input:<\/strong> Any existing file. Must be correctly balanced, but can be incorrectly formatted.<\/p><textarea id=\"code-paren-input\" rows=\"10\" cols=\"50\">\n(defn foo\n  ([a]\n    (foo a 1))\n  ([a b]\n    (let [sum (+ a b)\n          prod (* a b)\n          result { ; gather vals\n            :sum sum\n            :prod prod\n          }]\n      result)\n    ; TODO: something\n    ))\n<\/textarea><\/div><div class=\"col\"><p><strong>Output:<\/strong> Made ready for Parinfer by correcting indentation and moving close-parens.<\/p><textarea id=\"code-paren-output\" rows=\"10\" cols=\"50\"><\/textarea><\/div><\/div><p>\nNotice that this process is NOT an invasive pretty-printer.  Newline characters\nare never added or removed.  It preserves as much as it can of the original\ncode, only moving close-parens and changing indentation.\n<\/p><p>\n<em>Parinfer<\/em> should remain in Paren Mode if the file cannot be processed, due\nto the reasons stated in the previous section.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"knowing-when-parens-move-in-paren-mode\"><h3>Knowing When Parens Move<\/h3><p>\nAs a courtesy, <em>Paren Mode<\/em> will not move your parens while your cursor\nis behind them. Just move your cursor away when you're done.  A helpful\nanalogy might be to think of your cursor as a <em>paperweight<\/em> that keeps\nyour parens from blowing away.\n<\/p><div><p><strong>Parens displaced<\/strong> only after they are balanced.<\/p><textarea id=\"code-displaced-after-balance\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Parens not displaced<\/strong> if the cursor is behind it (allowing you to type).<\/p><textarea id=\"code-not-displaced-on-enter\" rows=\"10\" cols=\"50\"><\/textarea><\/div><div><p><strong>Parens displaced<\/strong> if the cursor is no longer behind them.<\/p><textarea id=\"code-displaced-after-cursor-leaves\" rows=\"10\" cols=\"50\"><\/textarea><\/div><\/section><\/div><div class=\"wrapper\"><h2>Conclusions<\/h2><\/div><div class=\"wrapper\"><section id=\"benefits\"><h3>Benefits<\/h3><p>\nInferring parentheses based on indentation seems to lead to simpler editing\nmechanics for Lisp code.  It leads to a system that keeps our code formatted\nwell. And it allows us to use paredit-like features without hotkeys.\n<\/p><p>\nI think the biggest win is its potential to quell fear of managing end-of-line\nparens by enforcing a direct driving relationship with indentation.\n<\/p><p>\nAnd just like <em>Paredit<\/em> it maintains paren-balanced code.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"downsides\"><h3>Downsides<\/h3><p>\nThe rules for what happens when inserting\/deleting parens must be learned.\nAlso, the case necessitating a \"Paren Mode\" comes at the cost of forcing the\nuser to understand when and how to switch editing modes.\n<\/p><p>\nAlso, the preprocessor step performed when opening files will cause more\nformatting-related changes in your commit history when collaborating with\nothers not using <em>Parinfer<\/em>.\n<\/p><\/section><\/div><div class=\"wrapper\"><section id=\"final-thoughts\"><h3>Final Thoughts<\/h3><p>\nRegardless of how we choose to edit our Lisp code, there seems to always be a\nbalancing act between maintaining the simplicity of how we interact with the\neditor and accepting some editor complexity to gain automation over these\npowerful but numerous parens.\n<\/p><p>\nBuilding the interactive examples for this page has allowed me to explore how\nwell Parinfer can play this balancing act, but only in a demo environment.\nThe <em>real test<\/em> will come once it becomes available to major editors.\nSee <a href=\"#editor-plugins\">editor plugins<\/a> for progress.\n<\/p><\/section><\/div><div class=\"wrapper\"><h2>Appendix<\/h2><\/div><div class=\"wrapper\"><section id=\"source-code\"><h3>Source Code<\/h3><p>\nThe text formatting code is intended to be editor-agnostic.  It is implemented\nin straightforward, imperative JavaScript, optimized for speed and designed\nto be easy to port to other languages. There is a test suite which is also\ndesigned to be easy to port with all test cases represented in JSON files.\n<\/p><p>\nThe editor demos on this site are created in CodeMirror with hooks to apply our\nformatters and update cursor position.  Source code for both the library and site\nare available on github:\n<\/p><p>\n<i class=\"fa fa-lg fa-code-fork\"> <a href=\"http:\/\/github.com\/shaunlebron\/parinfer\">http:\/\/github.com\/shaunlebron\/parinfer<\/a>\n<\/i><\/p><\/section><\/div><div class=\"wrapper\"><section id=\"editor-plugins\"><h3>Editor Plugins<\/h3><p>\nParinfer is still in early development. Several people have started integrating\nit into code editors at various stages of development.\n<\/p><p>\n<strong>Works in Progress:<\/strong>\n<\/p><ul><li>\n    <a href=\"https:\/\/github.com\/oakmac\/atom-parinfer\">atom-parinfer<\/a>\n    for <a href=\"https:\/\/atom.io\/\">Atom<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/eraserhd\/parinfer-rust\">parinfer-rust<\/a>\n    for <a href=\"https:\/\/www.vim.org\/\">Vim<\/a> and\n    <a href=\"https:\/\/neovim.io\/\">neovim<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/snoe\/nvim-parinfer.js\">nvim-parinfer.js<\/a>\n    for <a href=\"https:\/\/neovim.io\/\">neovim<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/bhurlow\/vim-parinfer\">vim-parinfer<\/a>\n    for <a href=\"http:\/\/www.vim.org\/\">Vim<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/Microsoft\/vscode-parinfer\">vscode-parinfer<\/a>\n    for <a href=\"https:\/\/code.visualstudio.com\/\">Visual Studio Code<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/oakmac\/sublime-text-parinfer\">sublime-text-parinfer<\/a>\n    for <a href=\"http:\/\/www.sublimetext.com\/\">Sublime Text<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/DogLooksGood\/parinfer-mode\">parinfer-mode<\/a>\n    for <a href=\"https:\/\/www.gnu.org\/software\/emacs\/\">Emacs<\/a>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/syl20bnr\/spacemacs\/issues\/5574\">Parinfer layer<\/a>\n    for <a href=\"http:\/\/spacemacs.org\/\">Spacemacs<\/a>\n    - <em>in progress<\/em>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/shaunlebron\/parinfer-codemirror\">parinfer-codemirror<\/a>\n    for <a href=\"https:\/\/codemirror.net\/\">CodeMirror<\/a>\n    - <em>used for this website!<\/em>\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/mauricioszabo\/lt_parinfer\">lt_parinfer<\/a>\n    for <a href=\"http:\/\/lighttable.com\/\">Light Table<\/a>\n  <\/li>\n  <li>\n    native support in <a href=\"https:\/\/github.com\/oakes\/Nightcode\">Nightcode<\/a>\n  <\/li>\n<\/ul><p>\nParinfer is available for some REPL environments as well:\n<\/p><ul><li>\n    <a href=\"https:\/\/github.com\/mfikes\/replete\">Replete<\/a> for iOS\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/binaryage\/dirac\">Dirac DevTools<\/a> for Google Chrome\n  <\/li>\n  <li>\n    <a href=\"http:\/\/jaredforsyth.com\/reepl\/\">Reepl<\/a> for the browser\n  <\/li>\n<\/ul><p>\n<a href=\"https:\/\/github.com\/shaunlebron\/parinfer\/issues\/new?title=new%20plugin\">Let me know<\/a>\nif you're working on a plugin, or\n<a href=\"https:\/\/github.com\/shaunlebron\/parinfer\/wiki\/Make-a-Plugin\">check the wiki<\/a>\nfor extra guidance.  Thanks!<\/p><\/section><\/div><div class=\"wrapper\"><section><h3>Acknowledgements<\/h3><ul><li>\n    <a href=\"https:\/\/elpa.gnu.org\/packages\/adjust-parens.html\">adjust-parens<\/a>\n    is an earlier idea for adjusting close-parens based on indentation in Emacs.\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/Malabarba\/aggressive-indent-mode\">aggressive-indent-mode<\/a>\n    re-indents code after every keystroke in Emacs, to remove the back-and-forth actions mentioned\n    in the <a href=\"#tools-for-writing-lisp\">Tools for Writing Lisp<\/a> section.\n    I believe Parinfer's Paren Mode is less aggressive in that it keeps\n    indentation within thresholds rather than hard limits (more on this in a\n    future section).\n  <\/li>\n  <li>\n    <a href=\"http:\/\/haml.info\">Haml<\/a>,\n    <a href=\"http:\/\/slim-lang.com\/\">Slim<\/a>,\n    <a href=\"http:\/\/jade-lang.com\/\">Jade<\/a> \u2014\n    comparing these indented-templating languages to Clojure's Hiccups is what\n    made me realize Lisp close-parens could be inferred from indentation.\n  <\/li>\n  <li>\n    Thanks to\n    <a href=\"https:\/\/github.com\/boxed\">Anders Hovm\u00f6ller<\/a> and\n    <a href=\"https:\/\/github.com\/darwin\">Antonin Hildebrand<\/a> for humoring me on an\n    <a href=\"https:\/\/github.com\/boxed\/indent-clj\/issues\/1\">early crazy idea<\/a>\n    and connecting me to some ideas about indentation and structural editors.\n  <\/li>\n  <li>\n    <a href=\"http:\/\/readable.sourceforge.net\/\">sweet-expressions<\/a>,\n    <a href=\"http:\/\/srfi.schemers.org\/srfi-49\/srfi-49.html\">i-expressions<\/a>,\n    <a href=\"https:\/\/github.com\/boxed\/indent-clj\">indent-clj<\/a>, et al\n    explored how parentheses can be inferred (but hidden).\n  <\/li>\n  <li>\n    <a href=\"http:\/\/learnyouahaskell.com\/higher-order-functions#function-application\">Haskell's $ operator<\/a>\n    infers a closing paren after all expressions to the right, even after\n    subsequent indented expressions, which is basically what Indent Mode in\n    Parinfer does.\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/Cirru\/sepal.clj\">Cirru Sepal in Clojure<\/a>\n    has an interesting approach to inferring Clojure parens from indentation\n    and other syntax sugar\u2014 <code>$<\/code>, <code>$ []<\/code>, <code>$\n    {}<\/code> and <code>,<\/code>.\n  <\/li>\n  <li>\n    It's worth mentioning\n    <a href=\"https:\/\/github.com\/mkremins\/flense\">Flense<\/a>\n    and\n    <a href=\"https:\/\/github.com\/darwin\/plastic\">Plastic<\/a>\n    as tools with ambitious structural editing concepts for Clojure text.\n  <\/li>\n  <li>\n    <a href=\"http:\/\/www.greenfoot.org\/frames\/\">Frame-Based Editing<\/a> and\n    <a href=\"https:\/\/scratch.mit.edu\/\">Scratch<\/a> remove parens (or curly braces)\n    using blocks (like the LEGO metaphor mentioned).\n  <\/li>\n  <li>\n    <a href=\"http:\/\/danmidwood.com\/content\/2014\/11\/21\/animated-paredit.html\">Paredit animations<\/a>\n    inspired the editor animations here.\n  <\/li>\n  <li>\n    <a href=\"https:\/\/github.com\/bbatsov\/clojure-style-guide#source-code-layout--organization\">The Clojure Style Guide<\/a>\n    helped me confirm some indentation conventions when formalizing Parinfer.\n  <\/li>\n  <li>\n    <a href=\"https:\/\/twitter.com\/escherize\">Bryan Maass<\/a>\n    explained to me how Lisp is like Lego\u2014 everything being the same\n    shape so they can snap together any way you see fit, while other syntaxes\n    have different shaped pieces which only fit together in special ways.\n  <\/li>\n  <li>\n    Thanks to\n    <a href=\"http:\/\/chrisoakman.com\/\">Chris Oakman<\/a>\n    for proof-reading this page, pushing me to make an efficient implementation\n    of Indent Mode, and starting the\n    <a href=\"https:\/\/github.com\/oakmac\/atom-parinfer\">atom-parinfer<\/a>\n    plugin.\n  <\/li>\n  <li>\n    Thanks to my friends at PROS for looking at an early draft of this page.\n  <\/li>\n  <li>\n    Thanks to the Clojure\/ClojureScript community for making me fall in love\n    with Lisp.\n  <\/li>\n  <li>\n    When John Carmack adopted Lisp (Racket) for Oculus Rift scripting,\n    it gave me a little more momentum to finish writing this!\n  <\/li>\n  <li>\n    I used the really cool <a href=\"https:\/\/github.com\/liabru\/gears-d3-js\">gears.d3.js<\/a>\n    library for the animated gears here.\n  <\/li>\n  <li>\n    I built Parinfer as an augmentation to the\n    <a href=\"https:\/\/codemirror.net\/\">CodeMirror<\/a>\n    editor, which I really enjoyed working with and studying to learn about\n    editors.\n  <\/li>\n<\/ul><\/section><\/div><section><div id=\"xkcd\"><a href=\"https:\/\/xkcd.com\/312\/\"><img src=\"https:\/\/shaunlebron.github.io\/parinfer\/img\/xkcd-frost.png\" alt=\"image\" \/><\/a><\/div><\/section>","created_at":"2018-11-07T16:22:49+0000","updated_at":"2018-11-07T16:22:49+0000","annotations":[],"mimetype":"text\/html","reading_time":17,"domain_name":"shaunlebron.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":110,"title":"","url":"https:\/\/user-images.githubusercontent.com\/1019692\/34446795-1","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:22:49+0000","updated_at":"2018-11-07T16:22:49+0000","annotations":[],"reading_time":0,"domain_name":"user-images.githubusercontent.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":111,"title":"bellecp\/fast-p","url":"https:\/\/github.com\/bellecp\/fast-p","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Quickly find and open a pdf among a collection of thousands of unsorted pdfs through fzf (fuzzy finder)<\/p>\n<ul><li><a href=\"#installation-on-unix-or-linux-based-systems\">Installation on Linux<\/a><\/li>\n<li><a href=\"#installation-on-osx-with-homebrew\">Installation on OSX<\/a><\/li>\n<li><a href=\"#usage\">Usage<\/a><\/li>\n<li><a href=\"#how-to-clear-the-cache\">How to clear the cache?<\/a><\/li>\n<li><a href=\"#see-it-in-action\">See it in action<\/a><\/li>\n<li><a href=\"#is-the-historical-bash-code-still-available\">Is the historical bash code still available?<\/a><\/li>\n<\/ul><ol><li>\n<p><strong>Requirements.<\/strong> Make sure the following requirements are satisfied:<\/p>\n<ul><li>install <code>pdftotext<\/code>. This comes with the texlive distribution on linux,\nOn ubuntu, <code>sudo apt-get install poppler-utils<\/code> .<\/li>\n<li>install <code>fzf<\/code>: <a href=\"https:\/\/github.com\/junegunn\/fzf\">https:\/\/github.com\/junegunn\/fzf<\/a><\/li>\n<li>install <code>GNU grep<\/code>,  <code>ag<\/code> (silver searcher).<\/li>\n<\/ul><\/li>\n<li>\n<p><strong>Install binary<\/strong>. Do either one of the two steps below:<\/p>\n<ul><li><strong>Compile from source with <code>go<\/code> and <code>go get<\/code>.<\/strong>\nWith a working <code>golang<\/code> installation, do\n<code>go get github.com\/bellecp\/fast-p<\/code>\nIt will fetch the code and its dependencies,\ncompile and create an executable <code>fast-p<\/code> in the <code>\/bin<\/code> folder of your go\ninstallation, typically <code>~\/go\/bin<\/code>. Make sure the command <code>fast-p<\/code> can be\nfound (for instance, add <code>~\/go\/bin<\/code> to your <code>$PATH<\/code>.)<\/li>\n<li>Or: <strong>Use the precompiled binary for your architecture.<\/strong> Download the binary that corresponds to your\narchitecture at <a href=\"https:\/\/github.com\/bellecp\/fast-p\/releases\">https:\/\/github.com\/bellecp\/fast-p\/releases<\/a> and make sure that\nthe command <code>fast-p<\/code> can be found. For instance,\nput the binary file <code>fast-p<\/code> in <code>~\/custom\/bin<\/code> and add <code>export  PATH=~\/custom\/bin:$PATH<\/code> to your <code>.bashrc<\/code>.<\/li>\n<\/ul><\/li>\n<li>\n<p><strong>Tweak your .bashrc<\/strong>. Add the following code to your <code>.bashrc<\/code><\/p>\n<\/li>\n<\/ol><pre>p () {\n    open=xdg-open   # this will open pdf file withthe default PDF viewer on KDE, xfce, LXDE and perhaps on other desktops.\n    ag -U -g \".pdf$\" \\\n    | fast-p \\\n    | fzf --read0 --reverse -e -d $'\\t'  \\\n        --preview-window down:80% --preview '\n            v=$(echo {q} | tr \" \" \"|\"); \n            echo -e {1}\"\\n\"{2} | grep -E \"^|$v\" -i --color=always;\n        ' \\\n    | cut -z -f 1 -d $'\\t' | tr -d '\\n' | xargs -r --null $open &gt; \/dev\/null 2&gt; \/dev\/null\n}\n<\/pre>\n<ul><li>You may replace <code>ag -U -g \".pdf$\"<\/code> with another command that returns a list of pdf files.<\/li>\n<li>You may replace <code>open=...<\/code> by your favorite PDF viewer, for instance <code>open=evince<\/code> or <code>open=okular<\/code>.<\/li>\n<\/ul><ol><li>Install <a href=\"https:\/\/brew.sh\/\" rel=\"nofollow\">homebrew<\/a> and  <strong>run<\/strong><\/li>\n<\/ol><pre>brew install bellecp\/fast-p\/fast-pdf-finder\n<\/pre>\n<p><em>The above brew formula is experimental.\nPlease report any issues\/suggestions\/feedback at <a href=\"https:\/\/github.com\/bellecp\/fast-p\/issues\/11\">https:\/\/github.com\/bellecp\/fast-p\/issues\/11<\/a><\/em><\/p>\n<ol start=\"2\"><li><strong>Tweak your .bashrc<\/strong>. Add the following code to your <code>.bashrc<\/code><\/li>\n<\/ol><pre>p () {\n    local open\n    open=open   # on OSX, \"open\" opens a pdf in preview\n    ag -U -g \".pdf$\" \\\n    | fast-p \\\n    | fzf --read0 --reverse -e -d $'\\t'  \\\n        --preview-window down:80% --preview '\n            v=$(echo {q} | gtr \" \" \"|\"); \n            echo -e {1}\"\\n\"{2} | ggrep -E \"^|$v\" -i --color=always;\n        ' \\\n    | gcut -z -f 1 -d $'\\t' | gtr -d '\\n' | gxargs -r --null $open &gt; \/dev\/null 2&gt; \/dev\/null\n}\n<\/pre>\n<ul><li>You may replace <code>ag -U -g \".pdf$\"<\/code> with another command that returns a list of pdf files.<\/li>\n<li>You may replace <code>open=...<\/code> by your favorite PDF viewer, for instance <code>open=evince<\/code> or <code>open=okular<\/code>.<\/li>\n<\/ul><p><strong>Remark:<\/strong> On OSX, we use the command line tools <code>gcut<\/code>, <code>gxargs<\/code>, <code>ggrep<\/code>, <code>gtr<\/code> which are the GNU versions\nof the tools <code>cut<\/code>, <code>xargs<\/code>, <code>grep<\/code>, <code>tr<\/code>. This way, we avoid the specifics of the versions of these tools pre-installed on OSX,\nand the same <code>.bashrc<\/code> code can be used for both OSX and GNU Linux systems.<\/p>\n<p>Use the command <code>p<\/code> to browse among the PDF files in the current directory and its subdirectories.<\/p>\n<p>The first run of the command will take some time to cache the text extracted from each pdf. Further runs of the command will be much faster since the text extraction will only apply to new pdfs.<\/p>\n<p>To clear the cache (which contains text extracted from PDF), you may safely remove the file\n<code>~\/.cache\/fast-p-pdftotext-output\/fast-p_cached_pdftotext_output.db<\/code>, or<br \/><code>~\/.cache\/fast-p_cached_pdftotext_output.db<\/code> in older version.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/user-images.githubusercontent.com\/1019692\/34446795-12229072-ecac-11e7-856a-ec0df0de60ae.gif\"><img src=\"https:\/\/user-images.githubusercontent.com\/1019692\/34446795-12229072-ecac-11e7-856a-ec0df0de60ae.gif\" alt=\"illustration of the p command\" \/><\/a><\/p>\n<p>Yes, see <a href=\"https:\/\/github.com\/bellecp\/fast-p\/blob\/master\/p\">https:\/\/github.com\/bellecp\/fast-p\/blob\/master\/p<\/a> but using the go binary as explained above is recommended for speed and interoperability.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:49+0000","updated_at":"2018-11-07T16:22:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/1019692?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":112,"title":"Zotero | Your personal research assistant","url":"https:\/\/www.zotero.org\/","content":"<section class=\"section-md\"><div class=\"row align-items-center\"><div class=\"feature-col col-md-6 order-md-first col-xl-5 offset-xl-1 left\"><h2 class=\"display-2\">Collect with a click.<\/h2><p>Zotero is the only software that automatically senses research on the web. Need an article from JSTOR or a preprint from arXiv.org? A news story from the New York Times or a book from a library? Zotero has you covered, everywhere.<\/p><\/div><\/div><\/section><section class=\"section-md\"><div class=\"row align-items-center\"><div class=\"feature-col col-md-6 col-xl-5 offset-xl-1 text-center left\"><img class=\"img-fluid\" src=\"https:\/\/www.zotero.org\/static\/images\/home\/organize.svg\" width=\"444\" height=\"335\" alt=\"image\" \/><\/div><div class=\"feature-col col-md-6 col-xl-5 right\"><h2 class=\"display-2\">Organize your way.<\/h2><p>Zotero helps you organize your research any way you want. You can sort items into collections and tag them with keywords. Or create saved searches that automatically fill with relevant materials as you work.<\/p><\/div><\/div><\/section><section class=\"section-md\"><div class=\"row align-items-center\"><div class=\"feature-col col-md-6 col-xl-5 text-center right\"><img class=\"img-fluid\" src=\"https:\/\/www.zotero.org\/static\/images\/home\/cite.svg\" width=\"444\" height=\"370\" alt=\"image\" \/><\/div><div class=\"feature-col col-md-6 order-md-first col-xl-5 offset-xl-1 left\"><h2 class=\"display-2\">Cite in style.<\/h2><p>Zotero instantly creates references and bibliographies for any text editor, and directly inside Word, LibreOffice, and Google Docs. With support for over 9,000 citation styles, you can format your work to match any style guide or publication.<\/p><\/div><\/div><\/section><section class=\"section-md\"><div class=\"row align-items-center\"><div class=\"feature-col col-md-6 col-xl-5 offset-xl-1 text-center left\"><img src=\"https:\/\/www.zotero.org\/static\/images\/home\/sync-md.svg\" alt=\"image\" \/><\/div><div class=\"feature-col col-md-6 col-xl-5 right\"><h2 class=\"display-2\">Stay in sync.<\/h2><p>Zotero can optionally synchronize your data across devices, keeping your notes, files, and bibliographic records seamlessly up to date. If you decide to sync, you can also always access your research from any web browser.<\/p><\/div><\/div><\/section><section class=\"section-md\"><div class=\"row align-items-center\"><div class=\"feature-col col-md-6 col-xl-5 text-center right\"><img class=\"img-fluid\" src=\"https:\/\/www.zotero.org\/static\/images\/home\/collaborate.svg\" width=\"444\" height=\"331\" alt=\"image\" \/><\/div><div class=\"feature-col col-md-6 order-md-first col-xl-5 offset-xl-1 left\"><h2 class=\"display-2\">Collaborate freely.<\/h2><p>Zotero lets you co-write a paper with a colleague, distribute course materials to students, or build a collaborative bibliography. You can share a Zotero library with as many people you like, at no cost.<\/p><\/div><\/div><\/section><section class=\"section-md\"><div class=\"row align-items-center\"><div class=\"feature-col col-md-6 col-xl-5 offset-xl-1 text-center left\"><img class=\"img-fluid\" src=\"https:\/\/www.zotero.org\/static\/images\/home\/privacy.svg\" width=\"444\" height=\"336\" alt=\"image\" \/><\/div><div class=\"feature-col col-md-6 col-xl-5 right\"><h2 class=\"display-2\">Rest easy.<\/h2><p>Zotero is <a href=\"https:\/\/github.com\/zotero\">open source<\/a> and developed by an independent, nonprofit organization that has no financial interest in your private information. With Zotero, you always stay in control of your own data.<\/p><\/div><\/div><\/section>","created_at":"2018-11-07T16:22:49+0000","updated_at":"2018-11-07T16:22:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"www.zotero.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":113,"title":"M: The Classic Intelligent Composing and Performing System | Cycling '74","url":"https:\/\/cycling74.com\/products\/m","content":"<p class=\"c74-box\" data-reactid=\"160\">M's main screen is designed to be a musical control panel that lets you do many things simultaneously while giving you a visual overview of all of the program's operations.<\/p><p class=\"c74-box\" data-reactid=\"161\">There are six windows in the main screen. Each window contains a particular grouping of controls, and the controls in all of the windows are interrelated as different parts of the same musical process. All of M's main screen windows are always active so that you can use any control whenever you like.<\/p><div class=\"c74-image c74-image--bg-light c74-image--padding-true\" data-reactid=\"162\"><div class=\"c74-image-progressive-wrapper\" data-reactid=\"163\"><div data-reactid=\"164\"><div class=\"placeholder\" data-reactid=\"165\"><noscript data-reactid=\"166\"><img height=\"353\" src=\"https:\/\/1cyjknyddcx62agyb002-web-assets.s3.amazonaws.com\/products\/m1.gif\" width=\"540\" alt=\"image\" \/><\/noscript><\/div><\/div><\/div><\/div><p class=\"c74-box\" data-reactid=\"167\">The <strong data-reactid=\"169\">Patterns Window<\/strong> contains four rows of controls for creating and managing Patterns. A <strong data-reactid=\"171\">Pattern<\/strong>, in M, is a collection of notes that can be transformed in a variety of ways. The Patterns Window also contains some controls for managing Voices. A <strong data-reactid=\"173\">Voice<\/strong> in M is a \"path\" through the program that begins with a Pattern.<\/p><p class=\"c74-box\" data-reactid=\"175\">The <strong data-reactid=\"177\">Conducting Window<\/strong> (Untitled in the picture above) contains controls that allow you to affect the operations of the whole program. You can change certain aspects of the program's operation in the <strong data-reactid=\"179\">Conducting Grid<\/strong>; you can start and stop the music, change the tempo, and do lots of other things that we'll describe later in this manual.<\/p><p class=\"c74-box\" data-reactid=\"181\">The <strong data-reactid=\"183\">Variables Window<\/strong> contains controls that allow you to transform M's Voices. A <strong data-reactid=\"185\">Variable<\/strong> in M is a category of transformation. In the Variables Window, you can change note ordering, transposition, texture, and density; you can introduce rhythmic \"feel\"; and you can set MIDI velocity ranges. You can also choose between different groups of Patterns that you've created.<\/p><p class=\"c74-box\" data-reactid=\"187\">The <strong data-reactid=\"189\">Cyclic Variables Window<\/strong> contains controls that allow you to establish cyclic variation for note durations, legato-staccato articulations, and accents.<\/p><p class=\"c74-box\" data-reactid=\"191\">The <strong data-reactid=\"193\">Midi Window<\/strong> contains controls that allow you to direct your Voices' outputs to different MIDI channels and to send program (patch) change numbers to your synthesizers.<\/p><p class=\"c74-box\" data-reactid=\"195\">The <strong data-reactid=\"197\">Snapshot Window<\/strong> (with the Camera icon) contains controls that allow you to store and recall groups of other screen controls. You can also record and play back sequences of Snapshots, which we call <strong data-reactid=\"199\">Slideshows<\/strong>.<\/p>","created_at":"2018-11-07T16:22:49+0000","updated_at":"2018-11-07T16:22:49+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"cycling74.com","preview_picture":"https:\/\/cycling74-web-assets.s3.amazonaws.com\/images\/photos\/max-with-bens-studio.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":107,"title":"nakkaya\/ferret","url":"https:\/\/github.com\/nakkaya\/ferret","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/ferret-lang.org\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/371124674157a7dc9d3061aead533f1f2a4b8914\/68747470733a2f2f62616467652e667572792e696f2f67682f6e616b6b6179612532466665727265742e737667\" alt=\"Current Release\" data-canonical-src=\"https:\/\/badge.fury.io\/gh\/nakkaya%2Fferret.svg\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/nakkaya\/ferret\/builds\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/554563124b029954e8dd0039873747d1f16cc94c\/68747470733a2f2f7472617669732d63692e6f72672f6e616b6b6179612f6665727265742e7376673f6272616e63683d6d6173746572\" alt=\"CI Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/nakkaya\/ferret.svg?branch=master\" \/><\/a>\n<a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/0727c45c2ed8efd1ffb16bc55f0c8246f4f2b631\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f4c6963656e73652d425344253230322d2d436c617573652d6f72616e67652e737667\"><img src=\"https:\/\/camo.githubusercontent.com\/0727c45c2ed8efd1ffb16bc55f0c8246f4f2b631\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f4c6963656e73652d425344253230322d2d436c617573652d6f72616e67652e737667\" alt=\"BSD 2 Clause License\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/License-BSD%202--Clause-orange.svg\" \/><\/a><\/p>\n<p>Ferret is a free software lisp implementation designed to be used in\nreal time embedded control systems. Ferret lisp compiles down to self\ncontained <em>C++11<\/em>. Generated code is portable between any Operating\nSystem and\/or Microcontroller that supports a <em>C++11<\/em> compliant\ncompiler. It has been verified to run on architectures ranging from\nembedded systems with as little as <em>2KB of RAM<\/em> to general purpose\ncomputers running Linux\/Mac OS X\/Windows.<\/p>\n<p>This repository contains the Ferret compiler. For more information about Ferret,\nincluding downloads and documentation for the latest release, check\nout <a href=\"https:\/\/ferret-lang.org\" rel=\"nofollow\">Ferret's website<\/a><\/p>\n<h2><a id=\"user-content-general-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#general-information\"><\/a>General Information<\/h2>\n<h2><a id=\"user-content-quick-start\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-start\"><\/a>Quick Start<\/h2>\n<p>Download latest Ferret release,<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>wget https:\/\/ferret-lang.org\/builds\/ferret.jar<\/pre><\/div>\n<p>A program that sums the first 5 positive numbers.<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>;;; lazy-sum.clj\n(defn positive-numbers\n  ([]\n   (positive-numbers 1))\n  ([n]\n   (cons n (lazy-seq (positive-numbers (inc n))))))\n(println (-&gt;&gt; (positive-numbers)\n              (take 5)\n              (apply +)))<\/pre><\/div>\n<p>Compile to binary using,<\/p>\n<div class=\"highlight highlight-source-shell\"><pre>$ java -jar ferret.jar -i lazy-sum.clj\n$ g++ -std=c++11 -pthread lazy-sum.cpp\n$ .\/a.out<\/pre><\/div>\n<\/article>","created_at":"2018-11-07T16:22:48+0000","updated_at":"2018-11-07T16:22:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/88499?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":108,"title":"hello-rust\/hello","url":"https:\/\/github.com\/hello-rust\/hello\/blob\/master\/README.md","content":"<p>A little command-line tool which helps reduce manual work for content creators.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#why\"><\/a>Why?<\/h3>\n<p>Social media is hard. On one side, you would like to keep all platforms updated\nand share your creations with the world. On the other hand, maintaining all of\nthem takes away a substantial amount of time from actually creating content.<\/p>\n<p>You might say that this is not an important problem to solve - and I agree to\nsome extent. But having a side-project next to work is a tough thing to handle,\nand every little step that can be automated saves some time for more important\nthings like... uhm... friends and family?<\/p>\n<p>You might say that this is cheating - and you might be right. Submitting an\narticle is a soulless work of honor that should be done manually as a sign of\nrespect (yo) to your audience. But guess what: most of your audience doesn't\ncare about your link on social media, it cares about your content.<\/p>\n<p>You might say there are similar tools - and that's true. There's\n<a href=\"https:\/\/buffer.com\" rel=\"nofollow\">Buffer<\/a>, which can do everything that this tool can, but\nbetter. Nevertheless, the goal is to build a free tool under a permissive\nlicense that is easy to hack on.<\/p>\n<p>Therefore I created <strong>hello<\/strong>.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h3>\n<pre>cargo install hello-rs\n<\/pre>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h3>\n<p>Create a <code>.env<\/code> file with your credentials. See <code>.env_dist<\/code> for an example.<\/p>\n<p>After that, run <code>hello -h<\/code> to get started.<\/p>\n<p>For example, to submit a link to Reddit, run<\/p>\n<pre>hello reddit subredditnamehere \"Give up\" \"https:\/\/www.youtube.com\/watch?v=DLzxrzFCyOs\"\n<\/pre>\n<h3><a id=\"user-content-currently-supported-platforms\" class=\"anchor\" aria-hidden=\"true\" href=\"#currently-supported-platforms\"><\/a>Currently supported platforms<\/h3>\n<ul class=\"contains-task-list\"><li class=\"task-list-item\"> Reddit<\/li>\n<li class=\"task-list-item\"> Twitter<\/li>\n<li class=\"task-list-item\"> HackerNews<\/li>\n<li class=\"task-list-item\"> Slack<\/li>\n<li class=\"task-list-item\"> Patreon (tough one, as the API does not support publishing yet)<\/li>\n<li class=\"task-list-item\"> Meetup.com<\/li>\n<li class=\"task-list-item\"> Discourse (tough one, as only admins can get an API key)<\/li>\n<li class=\"task-list-item\"> Bring your own!<\/li>\n<\/ul><h3><a id=\"user-content-how-to-get-your-credentials\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-get-your-credentials\"><\/a>How to get your credentials<\/h3>\n<h5><a id=\"user-content-reddit\" class=\"anchor\" aria-hidden=\"true\" href=\"#reddit\"><\/a>Reddit<\/h5>\n<p>Create an app at reddit.com\/prefs\/apps.\nAfter that, add your credentials to <code>.env<\/code>.\nThen you can run <code>hello<\/code> like so:<\/p>\n<pre>hello reddit yoursubredditname \"I gave up\" \"https:\/\/www.youtube.com\/watch?v=DLzxrzFCyOs\"\n<\/pre>\n<h5><a id=\"user-content-twitter\" class=\"anchor\" aria-hidden=\"true\" href=\"#twitter\"><\/a>Twitter<\/h5>\n<p>Make an app for yourself at apps.twitter.com and add the credentials to the\n<code>.env<\/code> file. On first execution of <code>hello<\/code>,\nyou will get an access token for your app via OAuth.<\/p>\n<p>Then run the following command to send a tweet:<\/p>\n<pre>hello twitter \"Hello! https:\/\/github.com\/hello-rust\/hello\"\n<\/pre>\n<p>Follow the instructions on the screen to save that for all subsequent requests.<\/p>\n<h5><a id=\"user-content-slack\" class=\"anchor\" aria-hidden=\"true\" href=\"#slack\"><\/a>Slack<\/h5>\n<p>Create an app at <a href=\"https:\/\/api.slack.com\/apps\" rel=\"nofollow\">https:\/\/api.slack.com\/apps<\/a>.\nThe app name doesn't matter. Choose the workspace that you want to send messages\nto.<\/p>\n<p>Click on \"create\". You will be redirected to the configuration page where you\nclick on \"permissions\".<\/p>\n<p>Look for \"scopes\" and select one of the following:<\/p>\n<ul><li><code>chat:write:user<\/code> if you want <code>hello<\/code> to send messages using your Slack username.<\/li>\n<li><code>chat:write:bot<\/code> if you want <code>hello<\/code> to send messages using the app username.<\/li>\n<\/ul><p>At the top, you should see your <code>OAuth<\/code> token, which you have to store in the\n<code>.env<\/code> file of <code>hello<\/code>. If you don't see your token, you might have to click on\n\"request approval\" to be allowed to install the app into the selected workspace.<\/p>\n<p>That's all you need. Save everything.<\/p>\n<p>Run the following command to send a message to <code>your-channel-name<\/code>:<\/p>\n<pre>hello slack \"your-channel-name\" \"Hello! https:\/\/github.com\/hello-rust\/hello\"\n<\/pre>\n<h4><a id=\"user-content-hackernews\" class=\"anchor\" aria-hidden=\"true\" href=\"#hackernews\"><\/a>HackerNews<\/h4>\n<p>Since HackerNews doesn't have an API for submitting links, we have to be\ncreative.\nWe use the awesome <a href=\"https:\/\/github.com\/jonhoo\/fantoccini\">fantoccini<\/a> to control\na <a href=\"https:\/\/github.com\/Fyrd\/caniuse\/issues\/2757#issuecomment-304529217\">WebDriver<\/a>\ncompatible browser.<\/p>\n<ol><li>Install geckodriver by fetching the <a href=\"https:\/\/github.com\/mozilla\/geckodriver\/releases\">latest build from their release page<\/a>.<\/li>\n<li>Add your HN credentials to <code>.env<\/code>.<\/li>\n<\/ol><p>Example:<\/p>\n<pre>hello hn \"Show HN: Hello, a CLI tool for managing social media\" https:\/\/github.com\/hello-rust\/hello\n<\/pre>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h3>\n<p>This tool was made possible by the excellent patrons of <a href=\"https:\/\/github.com\/hello-rust\/show\">\"Hello\nRust!\"<\/a>, a show about the Rust programming\nlanguage. With their help, this tool is made available to the public under dual\nMIT\/Apache license.<\/p>\n<p><a href=\"https:\/\/www.patreon.com\/bePatron?c=1568097\" rel=\"nofollow\">Become a patron<\/a> now to support\nfuture work and send pull requests for supporting other platforms.<\/p>","created_at":"2018-11-07T16:22:48+0000","updated_at":"2018-11-07T16:22:48+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/36301855?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":105,"title":"To Play Gin Rummy","url":"http:\/\/www.streetdirectory.com\/etoday\/to-play-gin-rummy-palafe.html","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:22:47+0000","updated_at":"2018-11-07T16:22:47+0000","annotations":[],"reading_time":0,"domain_name":"www.streetdirectory.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":106,"title":"Fortune Telling Playing Cards","url":"https:\/\/cafeastrology.com\/fortunetellingcards.html","content":"<p>Fortune Telling Playing Cards<\/p><p style=\"text-align: justify;\"><img border=\"0\" src=\"https:\/\/cafeastrology.com\/aces.jpg\" width=\"127\" height=\"255\" alt=\"image\" style=\"float: right; margin-left: 4px; margin-right: 4px; margin-top: 2px; margin-bottom: 2px;\" \/>As a teenager, I came across an old fortune-telling magazine that featured a divination method using a regular deck of playing cards. This method appealed to me because the cards were related to Astrology. In fact, the four suits were associated with the four elements (Fire, Earth, Air, and Water).\u00a0 <\/p><p style=\"text-align: justify;\">Long before people used cards to play games, they were used to foretell the future. In fact, playing cards were invented for divination purposes. <\/p><p style=\"text-align: justify;\">Although there are different ways to approach cartomancy, we will show you a simple method and lay out the basic meanings attached to each playing card below. <\/p><p style=\"text-align: justify;\">Once you know the basic meanings of the playing cards, you can choose a spread that is right for you. In fact, you can even choose to select only one card out of a deck to get a fast and easy \u201canswer\u201d to your problem, if you so desire. <\/p><p style=\"text-align: justify;\">First things first, the suits (Hearts, Clubs, Spades, and Diamonds) are associated with the four elements, as follows: <\/p><table border=\"1\" style=\"width: 27%; background-color: #B0C0A0;\"><tbody><tr><td style=\"width: 100%;\">\n<p style=\"text-align: center;\"><br \/>Hearts \u2013 Water<\/p>\n<p style=\"text-align: center;\">Clubs \u2013 Fire<\/p>\n<p style=\"text-align: center;\">Diamonds \u2013 Earth<\/p>\n<p style=\"text-align: center;\">Spades \u2013 Air<\/p>\n<\/td>\n<\/tr><\/tbody><\/table><table border=\"1\" style=\"width: 100%; background-color: #B0C0A0;\"><tbody><tr><td style=\"width: 100%;\">\n<p style=\"text-align: center;\"><b>What Each Playing Card Means:<\/b><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><p><b>Ace of Hearts <\/b>Love and happiness. The home, a love letter. This card is a particularly favorable card that indicates troubles and problems lifting.<\/p><p><b>King of Hearts <\/b>A fair-haired man with a good nature; or a man with Water signs predominating in his chart. Fair, helpful advice. Affectionate, caring man. This man helps you out without much talk. His actions reveal his kindness and concern.<\/p><p><b><img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/queenofhearts.gif\" width=\"71\" height=\"97\" alt=\"image\" style=\"float: right; margin-left: 4px; margin-right: 4px; margin-top: 2px; margin-bottom: 2px;\" \/>Queen of Hearts <\/b>A fair-haired woman with a good nature; or a woman with Water signs predominating in her chart. Kind advice. Affectionate, caring woman. Sometimes, this card can indicate the mother or a mother figure.<\/p><p><b>Jack of Hearts <\/b>A warm-hearted friend. A fair-haired youth; or a young person with Water signs predominating in his or her chart. Often this points to a younger admirer.<\/p><p><b>10 of Hearts <\/b>Good luck, success. This is an important card that suggests good fortune after difficulty.<\/p><p><b>9 of Hearts <\/b>The card of wishes. A wish\/dream fulfilled. Look to the card just preceding this one to determine what the querent desires.<\/p><p><b>8 of Hearts <\/b>Unexpected gift or visit; an invitation to a party.<\/p><p><b>7 of Hearts <\/b>Someone whose interest in you is unreliable; someone with fickle affections for you. This card can indicate lovesickness.<\/p><p><b>6 of Hearts <\/b>A sudden wave of good luck. Someone takes care of you, takes warm interest in you.<\/p><p><b>5 of Hearts <\/b>Jealousy; some ill-will from people around you.<\/p><p><b>4 of Hearts <\/b>Travel, change of home or business.<\/p><p><b>3 of Hearts <\/b>Love and happiness when the entire spread is generally favorable. In a difficult spread, this can indicate emotional problems and an inability to decide who to love.<\/p><p><b>2 of Hearts <\/b>A warm partnership or engagement. This is a very favorable card that indicates strength and support coming from a partner.<\/p><p><b>Ace of Clubs <\/b>Wealth, prosperity, unexpected money\/gain. However, in a difficult spread, this money may disappear almost as quickly as it appears.<\/p><p><b>King of Clubs <\/b>Dark-haired, kind-hearted man; or a man with Fire predominating in his chart. A generous, spirited man.<\/p><p><b><img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/queenofclubs.gif\" width=\"71\" height=\"97\" alt=\"image\" style=\"float: right; margin-left: 4px; margin-right: 4px; margin-top: 2px; margin-bottom: 2px;\" \/>Queen of Clubs <\/b>Dark-haired, confident woman; or a woman with Fire predominating in her chart. She may give you good advice.<\/p><p><b>Jack of Clubs <\/b>A dark-haired or fiery youth. Popular youth who is good-hearted and playful. Can also indicate an admirer.<\/p><p><b>10 of Clubs <\/b>Business success. Good luck with money. A trip taken now may result in a new friend or love interest.<\/p><p><b>9 of Clubs <\/b>Achievement; sometimes a wealthy marriage or a sudden windfall.<\/p><p><b>8 of Clubs <\/b>Work\/business problems that may have to do with jealousy. This is generally thought to be quite unfavorable.<\/p><p><b>7 of Clubs <\/b>Business\u00a0success, although there may be problems in love. A change in business that may have been expected or earned, such as a promotion.<\/p><p><b>6 of Clubs <\/b>Financial aid or success.<\/p><p><b>5 of Clubs <\/b>New friendships, alliances are made.<\/p><p><b>4 of Clubs <\/b>Beware of dishonesty or deceit; avoid blind acceptance of others at this time.<\/p><p><b>3 of Clubs <\/b>Love and happiness; successful marriage; a favorable long-term proposition. A second chance, particularly in an economical sense.<\/p><p><b>2 of Clubs <\/b>Obstacles to success; malicious gossip.<\/p><p><b>Ace of Spades <\/b>Misfortune; sometimes associated with death or, more often, a difficult ending.<\/p><p><b>King of Spades <\/b>Dark-haired man; or a man with Air predominating in his chart. An ambitious man, perhaps self-serving.<\/p><p><b><img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/queenofspades.gif\" width=\"71\" height=\"97\" alt=\"image\" style=\"float: right; margin-left: 4px; margin-right: 4px; margin-top: 2px; margin-bottom: 2px;\" \/>Queen of Spades <\/b>Widowed or divorced woman; or a woman with Air predominating in her chart.<\/p><p><b>Jack of Spades <\/b>A youth who is hostile or jealous.<\/p><p><b>10 of Spades <\/b>Worry; bad news.<\/p><p><b>9 of Spades <\/b>Illness, accident, bad luck. The querent is at his\/her personal low.<\/p><p><b>8 of Spades <\/b>Temptation, misfortune, danger, upsets.<\/p><p><b>7 of Spades <\/b>Advice that is best not taken; loss. There is some obstacle to success, and this indicates that obstacles may be coming from within the querent.<\/p><p><b>6 of Spades <\/b>Small changes and improvements.<\/p><p><b>5 of Spades <\/b>Opposition and obstacles that are temporary; a blessing in disguise. Sometimes indicates a negative or depressed person.<\/p><p><b>4 of Spades <\/b>Small worries, problems. Financial difficulties, personal lows.<\/p><p><b>3 of Spades <\/b>Breaks in relationships. Sometimes indicates that a third person is breaking into a relationship somehow.<\/p><p><b>2 of Spades <\/b>Breaks in relationships; deceit. A break in an important process in the querent\u2019s life. If the question concerns a particular romantic interest, this is considered a warning card \u2013 infidelity or separation is quite likely.<\/p><p><b>Ace of Diamonds <\/b>Change; a message, often about money, and usually good news.<\/p><p><b>King of Diamonds <\/b>Fair-haired or graying man, or a man with Earth predominating in his chart. A man of authority, status, or influence.<\/p><p><b><img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/queenofdiamonds.gif\" width=\"71\" height=\"97\" alt=\"image\" style=\"float: right; margin-left: 4px; margin-right: 4px; margin-top: 2px; margin-bottom: 2px;\" \/>Queen of Diamonds <\/b>Fair-haired woman; or a woman with Earth predominating in her chart. A gossip.<\/p><p><b>Jack of Diamonds <\/b>A youth, possibly in uniform. Or, a jealous person who may be unreliable. A person who brings news, generally negative, but relatively minor.<\/p><p><b>10 of Diamonds <\/b>A change in financial status, often for the better.<\/p><p><b>9 of Diamonds <\/b>A new business deal; travel; restlessness; a change of residence.<\/p><p><b>8 of Diamonds <\/b>New job; change in job situation. The young or the old may find love on a trip.<\/p><p><b>7 of Diamonds <\/b>An argument concerning finances, or on the job. Generally expected to be resolved happily.<\/p><p><b>6 of Diamonds <\/b>Relationship problems, arguments. Separation.<\/p><p><b>5 of Diamonds <\/b>Happiness and success. A change for the better. A birth, or good news for a child. A good time to start new projects.<\/p><p><b>4 of Diamonds <\/b>Financial upswing; an older person may give good advice.<\/p><p><b>3 of Diamonds <\/b>A legal letter. Be tactful with others in order to avoid disputes.<\/p><p><b>2 of Diamonds <\/b>A business partnership; a change in relationship; gossip.<\/p><table border=\"1\" style=\"width: 100%; background-color: #B0C0A0;\"><tbody><tr><td style=\"width: 100%;\">\n<p style=\"text-align: center;\"><b>The Spreads<\/b><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><p style=\"text-align: justify;\">The basic idea for laying out a spread is to shuffle the cards as you think of your question or problem, and then draw cards in a particular order and lay them out on the table before you. You can make your own \u201cfortune telling\u201d session as simple or as complex as you please:\u00a0<\/p><ul><li>\nDraw only one card for a simple and fast answer to a burning question.\u00a0\n<\/li>\n<li>\nA common spread involves drawing three cards \u2014 the first represents the past, the second shows the situation in the present, and the third reveals the outcome or future of the matter.\n<\/li>\n<li>\nSome will do a larger spread when looking for a general outlook for a person. For this method, draw three cards for each of these categories: Yourself, Your Family, Your Friends, What You Expect, What You Don\u2019t Expect, and The Outcome.\n<\/li>\n<li>\nUse your own categories, tailored to your question.\n<\/li>\n<\/ul><table border=\"1\" style=\"width: 100%; background-color: #B0C0A0;\"><tbody><tr><td style=\"width: 100%;\">\n<p style=\"text-align: center;\"><b>Interesting Notes About Playing Cards<\/b><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><p>There are thirteen cards in each of the four suits, corresponding to the 13 lunar months in a year. Fifty-two cards correspond to fifty-two weeks in a year.<\/p><p>Hearts correspond to the element of water, and signify love, friendship, happiness, and domestic concerns. Generally speaking, they are \u201chappy\u201d cards.<\/p><p>Clubs correspond to the element of fire, and generally signify business, ambition, and achievement issues. They generally indicate success and happiness.<\/p><p>Diamonds correspond to the element of earth, and generally signify career, and money issues.<\/p><p>Spades correspond to the element of air, and generally signify gossip, challenges, messages, and upsets.<\/p><p>Try our <b><a href=\"https:\/\/cafeastrology.com\/loveoracle.html\">Love Oracle<\/a><\/b>, which is based on fortune telling with playing cards.<\/p><table border=\"0\" cellpadding=\"9\" style=\"width: 600px; border-spacing: 9px;\"><tbody><tr><td align=\"center\" valign=\"top\">\n<p style=\"text-align: center;\"><a href=\"https:\/\/cafeastrology.com\/articles\/lovesexastrology\/\"><img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/ad\/amour4.png\" width=\"190\" height=\"189\" class=\"aligncenter\" alt=\"image\" \/><\/a><\/p>\n<p style=\"text-align: center;\"> <a href=\"https:\/\/cafeastrology.com\/articles\/lovesexastrology\/\">THE ASTROLOGY OF LOVE &amp; ATTRACTION:<br \/>ARTICLES &amp; GUIDES<\/a><\/p>\n<\/td>\n<td align=\"center\" valign=\"top\">\n<p><a href=\"https:\/\/cafeastrology.com\/compatibilityreports\/\"><img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/ad\/romanticadside.png\" width=\"190\" height=\"190\" class=\"aligncenter\" alt=\"image\" \/><\/a><\/p>\n<p><a href=\"https:\/\/cafeastrology.com\/compatibilityreports\/\">ROMANTIC COMPATIBILITY REPORTS<\/a><\/p>\n<\/td>\n<\/tr><\/tbody><\/table><hr \/><h4>Suggested Further Reading on Cartomancy:<\/h4><p><a href=\"http:\/\/www.amazon.com\/exec\/obidos\/redirect?path=ASIN\/965494040X&amp;link_code=as2&amp;camp=1789&amp;tag=cafeastrology-20&amp;creative=9325\">Playing Cards: Predicting Your Future (Astrology Complete Guide)<\/a><img src=\"http:\/\/www.assoc-amazon.com\/e\/ir?t=cafeastrology-20&amp;l=as2&amp;o=1&amp;a=965494040X\" width=\"1\" height=\"1\" border=\"0\" alt=\"\" \/><\/p><p>This book interprets each playing card\u2019s significance and offers a variety of fortune telling spreads.<\/p><p style=\"text-align: center;\">\u00a0<\/p><p><a href=\"https:\/\/cafeastrology.com\/futureforecastreports\/\"> <img border=\"0\" src=\"https:\/\/cafeastrology.com\/images\/whatdestinyawaitsyou.gif\" width=\"468\" height=\"60\" alt=\"image\" \/><\/a><\/p><p style=\"text-align: center;\">\u00a0<\/p>","created_at":"2018-11-07T16:22:47+0000","updated_at":"2018-11-07T16:22:47+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":7,"domain_name":"cafeastrology.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":103,"title":"Learning From Terminals to Design the Future of User Interfaces \u2014 Brandur Leach","url":"https:\/\/brandur.org\/interfaces","content":"<p>I was recently called out on Twitter for claiming that\nElectron-based Slack, with three teams configured,\nregularly takes 30+ seconds to load. They claimed that I\nwas either committing gross hyperbole, or the victim of\nsome localized problem. I responded by sending over a video\nof me opening Slack and loading each of my teams in\nsuccession. It was 45 seconds long. <em>My<\/em> claim is that this\nsort of loading time isn\u2019t unusual at all. It\u2019s just that\nwe\u2019re all used to it.<\/p><figure><p>\n    <\/p>\n  <figcaption>This is a video of me waiting for Slack\n    configured with three teams to fully load. It's 45\n    seconds long.<\/figcaption><\/figure><p>Modern applications and interfaces frustrate me. In today\u2019s\nworld every one of us has the awesome power of the greatest\ncomputers in human history in our pockets and at our desks.\nThe computational capacity at our finger tips would have\nbeen unimaginable even to the most audacious thinkers of\nthirty years ago.<\/p><p>These powerful devices should be propelling our workflows\nforward with us gangly humans left barely able to keep up,\nand yet, almost without exception we wait for our computers\ninstead of the other way around. We\u2019re conditioned\nourselves to think that waiting 30+ seconds for an app to\nload, or interrupting our workflow to watch a half second\nanimations a thousand times a day, are perfectly normal.<\/p><figure><p>\n    <img src=\"https:\/\/brandur.org\/assets\/interfaces\/yahoo-1995.jpg\" srcset=\"\/assets\/interfaces\/yahoo-1995@2x.jpg 2x, \/assets\/interfaces\/yahoo-1995.jpg 1x\" class=\"overflowing\" alt=\"image\" \/><\/p>\n  <figcaption>Yahoo circa 1995.<\/figcaption><\/figure><p>Somewhere around the late 90s or early 00s we made the\ndecision to jump ship from desktop apps and start writing\nthe lion\u2019s share of new software for the web. This was\nlargely for pragmatic reasons: the infrastructure to talk\nto a remote server became possible for the first time, good\ncross platform UI frameworks had always been elusive beasts\n[1], and desktop development frameworks were intimidating\ncompared to more approachable languages like Perl and PHP.<\/p><p>The other reason was cosmetic: HTML and CSS gave developers\ntotal visual control over what their interfaces looked\nlike, allowing them to brand them and build experiences\nthat were pixel-perfect according to their own ends. This\nseemed like a big improvement over more limiting desktop\ndevelopment, but it led us to the world we have today\nwhere every interface is a different size and shape, and\nthe common display conventions that we used to have to aid\nwith usability have become distant memories of the past.<\/p><p>Today, web apps are still being hailed as the future. With\nthe possible exception of mobile, most software companies\nare building their products for the web, and even when\nthey\u2019re not, web technology is considered a reasonable\nalternative for the desktop. Vocal groups proclaim that\nElectron-based apps convey huge benefits compared to\ntraditional options in productivity and flexibility, and\nare the way forward for all desktop software.<\/p><p>I\u2019m not on a mission to demean this technology, but as it\u2019s\ncontinually augmented with ever more unwieldy retrofits,\nthere\u2019s a widening disparity between what we can build with\nit compared to the best-written native apps. Software on\nthe web today takes too long to load, depends too heavily\non synchronous calls to slow networks, overemphasizes\nvisual gimmickry, and lacks the refinement that allows\nmastery by more experienced users to gain huge leverage for\nproductivity\u2019s sake.<\/p><p>In 2007, after releasing the iPhone, Steve Jobs told\ndevelopers that they could all write apps for the iPhone\n<em>today<\/em> \u2026 as long as they did it in HTML5. To his credit,\nhe reversed his position inside a year after realizing how\ncompromised the web experience was compared to native\noptions.<\/p><p>In 2012, Mark Zuckerberg ignited JavaScript proponents\neverywhere after announcing that Facebook\u2019s biggest mobile\nmistake was focusing on HTML5. Meanwhile, consumers\neverywhere celebrated as they were given a native app that\nwas far faster and more responsive.<\/p><p>Every one of us knows that when it comes to a smartphone,\nwe\u2019d use a native app over an in-browser HTML5 any day of\nthe week. Yet when it comes to the desktop, we\u2019re still\nusing Gmail, Reddit, Trello, and JIRA. Computers and\nnetworks tend to be fast enough that this software is \u201cgood\nenough\u201d. Tellingly though, we avoid this software whenever\nbetter options are available, like with our terminals and\ntext editors.<\/p><p>Web technology isn\u2019t conducive to fast and efficient UIs,\nbut that\u2019s not the only problem we\u2019re facing. Somewhere\nalong the way UX designers became addicted to catchy, but\nsuperfluous, interface effects.<\/p><p>Think of all the animations that an average user sits\nthrough in a day: switching between spaces in Mac OS,\n1Password\u2019s unlock, waiting for iOS to show the SpringBoard\nafter hitting the home button, entering full screen from a\nMac OS app, or switching between tabs in mobile Safari.<\/p><figure><p>\n    <\/p>\n  <figcaption>1Password's unlock animation. The stuttering\n    isn't a problem with the video on this page; it's\n    actually how the animation looks.<\/figcaption><\/figure><figure><p>\n    <\/p>\n  <figcaption>OS X Spaces, introduced in Leopard. A\n    nominally useful feature, but the mandatory animations\n    make them slow and unwieldy.<\/figcaption><\/figure><p>I liked every one of them the first time. The next five\nthousand times were less impressive. And the same goes for\nall the flourishes in this class \u2013 they look great in\nscreenshots and demos, but don\u2019t advance our ability to be\nproductive; in fact, they do the opposite.<\/p><figure><p><a href=\"https:\/\/brandur.org\/assets\/interfaces\/cmd-tab@2x.png\"><img src=\"https:\/\/brandur.org\/assets\/interfaces\/cmd-tab.png\" srcset=\"\/assets\/interfaces\/cmd-tab@2x.png 2x, \/assets\/interfaces\/cmd-tab.png 1x\" class=\"overflowing\" alt=\"image\" \/><\/a><\/p>\n  <figcaption>Will Cmd + Tab be the next victim of overzealous animation?<\/figcaption><\/figure><p>I live in fear that one day Apple will realize that they\u2019ve\nleft a gaping hole in their UX strategy and that task\nswitches from Cmd + Tab should be animated. Multiply that\nanimation\u2019s length by the average number of task switches\nper day by the number of users by their cost per second,\nand you\u2019d be able to see that millions of dollars a year in\nglobal productivity has evaporated overnight.<\/p><p>Animations are a particularly egregious visual gimmick, but\nthere are others: whitespace so extravagant that only a\nminute amount of content can fit on the screen, overly\nlarge font sizes, submenus where a row of links would do\njust as well, unlabeled icons that look neat but leave\ntheir users guessing as to what they do, fixed headers that\nobscure content. The list goes on.<\/p><figure><p>\n    <\/p>\n  <figcaption>Contrary to any \"modern\" interfaces, a\n    terminal is fast and responsive. There are no animations\n    or other superfluous visual baggage.<\/figcaption><\/figure><p>Many of us developers are building web applications for\nother people while simultaneously eschewing them ourselves\nas much as we possibly can. While our users move at glacial\nspeeds through pages on the web, we\u2019re sitting in terminal\nenvironments that aren\u2019t just fast, but come with the\npromise of incredible advancements in productivity to\nanyone willing to spend the time to master them.<\/p><p>Here\u2019s why I like using terminals and terminal programs:<\/p><ul><li><p>Startup\/loading time is negligible.<\/p><\/li>\n<li><p>Time to transition between different screens is\ninstantaneous (no animations in sight).<\/p><\/li>\n<li><p>Interface elements are limited, but uniform.<\/p><\/li>\n<li><p>The learning curve is steep, but rewarding. They\u2019re\noptimized for the experienced user rather than the first\ntimer. Given that successfully onboarded users may spend\ntens of thousands of hours in the UI over the course of\ntheir lifetimes, this is just good sense.<\/p><\/li>\n<li><p>Composability: I\u2019m far from a zealot singing the praises\nof the Unix philosophy, but <em>most<\/em> terminal apps produce\noutput that I can process in some way to get into another\nprogram. It could be way better, but it\u2019s leaps and\nbounds over what I can do on the desktop. Even copying\ntext out of a modern web app can be a tricky proposition\nif HTML elements aren\u2019t nested optimally.<\/p><\/li>\n<\/ul><figure><p>\n    <\/p>\n  <figcaption>Modern UIs have next to zero composability.\n    Even copying text can be a tricky proposition.\n    <\/figcaption><\/figure><p>If you ask a web designer about the elements of practical\ndesign in interfaces today (I say <em>practical<\/em> to\ndisambiguate from vague design tenets like <a href=\"https:\/\/www.vitsoe.com\/us\/about\/good-design\">Dieter Rams\u2019\nten principles of good design<\/a>), they\u2019d talk\nto you about text legibility, intuitiveness, and\nwhitespace. I\u2019d argue that we\u2019re optimizing for the wrong\nthings. UIs that are pretty and friendly are nice to have,\nbut the true values of a good interface should be speed\nand efficiency to make their users as productive as\npossible.<\/p><p>Let\u2019s dig into it by looking at the aspirational interface\nconcept from a great movie: <em>Minority Report<\/em>. <a href=\"https:\/\/www.youtube.com\/watch?v=PJqbivkm0Ms\">Here\u2019s a\nvideo<\/a> of it in action.<\/p><figure><p>\n    <a href=\"https:\/\/www.youtube.com\/watch?v=PJqbivkm0Ms\">\n      <img src=\"https:\/\/brandur.org\/assets\/interfaces\/minority-report.jpg\" srcset=\"\/assets\/interfaces\/minority-report@2x.jpg 2x, \/assets\/interfaces\/minority-report.jpg 1x\" class=\"overflowing\" alt=\"image\" \/><\/a>\n  <\/p>\n  <figcaption>A futuristic and unrealistic concept interface:\n    the computer waits on the human instead of the human\n    waiting on the computer.<\/figcaption><\/figure><p>I think we can all agree that the interface of this\nprospective future is incredible and desirable, but if we\ndrill into it, what\u2019s its most amazing aspect?<\/p><p>Years ago, I might have said that it was the wafer thin\nscreens. Or the incredible touch technology. But we have\nboth of those things now! In fact, what we have today is\n<em>better<\/em>; we can display more than two colors on screen!\nFar superior to anything they seem to have in Philip K.\nDick\u2019s dystopian future.<\/p><p>Today, by far the most amazing aspect is that it\u2019s an\ninterface that\u2019s keeping up to its user. Instead of waiting\non the computer to think about some text completion, show\nhim an animation because he\u2019s switching an app, or start up\na program, it\u2019s keeping up with everything he tells it do\nin real time. The computer waits on the human rather than\nthe other way around. Besides terminals and a few other\npieces of fringe technology, modern UIs don\u2019t even come\nclose to a future this fantastic.<\/p><p>A successful interface isn\u2019t one that looks good in a still\nscreenshot, it\u2019s one that maximizes our productivity and\nlets us <em>keep moving<\/em>. Legibility and whitespace are great,\nbut they\u2019re of vanishing unimportance compared to speed and\nresponsiveness.<\/p><p>Neither a terminal nor today\u2019s web apps are what the future\nshould look like, but the terminal is closer.<\/p><p>Unfortunately, terminals also <em>suck<\/em>. Although better than\nthe alternative in many ways, they\u2019ve failed to keep up\nwith any advancements from the last thirty odd years.\nHere\u2019s a few places where terminals could stand to be\ninspired by web technology:<\/p><ul><li><p>Rich media elements: images, videos, tabulated results,\netc. The terminal has needed an answer to these since\n1985, but still doesn\u2019t have one.<\/p><\/li>\n<li><p>Fonts. Monospace is the best family of fonts for\nprogramming, but is objectively terrible for reading. We\nshould be able to mix fonts within a single terminal\ninterface for optimal legibility.<\/p><\/li>\n<li><p>Whitespace and line-height: used in moderation, these do\nhelp make UI elements more distinctive and text more\nlegible.<\/p><\/li>\n<\/ul><p>Terminals also need a lot of other things before they\u2019re\never going to a plausible interface replacement for most\npeople. UI elements that aren\u2019t built around ASCII bar\ncharacters for example.<\/p><p>We need a reboot. We need toolkits that produce interfaces\nthat are fast, consistent, bug free, and composable <em>by\ndefault<\/em> so that good interfaces aren\u2019t just something\nproduced by the best developer\/designers in the world, but\ncould be reasonably expected from even junior people in the\nindustry.<\/p><p>We should be honest with ourselves and call out design\nanti-patterns that promote flashiness at the expense of\nefficiency.<\/p><p>We should stop babying our users and try to raise beginners\nand the less technical to the bar of modern day power users\nrather than produce software that\u2019s designed for the lowest\ncommon denominator. We need more applications like Vim,\nEmacs, and Irssi that push their users to improve and pay\nhuge dividends to those who are willing to make the effort,\nand we need to train people to use them.<\/p><p>We should build networked applications that cache content\nand make network fetches asynchronously to remote APIs so\nthat humans aren\u2019t waiting for data to come back over the\nwire while they\u2019re working.<\/p><p>There\u2019s a future out there where our software makes\neverything from filing a bug to paying off your credit card\nfast and efficient, but the path that we\u2019re on today isn\u2019t\nit.<\/p><div id=\"footnotes\"><p><sup id=\"footnote-1\"><a href=\"#footnote-1-source\">1<\/a><\/sup> Fans of Qt (and maybe others) will vehemently disagree\n    that there\u2019s never been a good cross platform UI\n    library. I\u2019d argue that SDKs like Qt were never quite\n    accessible enough and never produced good enough\n    results to be suitable for universal adoption.<\/p><\/div>","created_at":"2018-11-07T16:22:46+0000","updated_at":"2018-11-07T16:22:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":10,"domain_name":"brandur.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":104,"title":"Browserless","url":"https:\/\/www.browserless.io\/blog\/2018\/08\/01\/running-an-indie-business.html","content":"<main id=\"main\" class=\"main main-1\"><div class=\"main-content\"><div class=\"container\"><div class=\"row d-flex align-items-center\"><div class=\"col-lg-6 offset-lg-0 col-md-10 offset-md-1 col-sm-10 offset-sm-1 text-center text-lg-left mb-5 mb-lg-0\"><div class=\"pr-0 pr-xl-5 pr-lg-0\"><p class=\"main-content__lead lead mb-4\">Stop spending time running and scaling Chrome, and more time on your application.<\/p><div class=\"row d-none d-lg-flex\"><div class=\"col-lg-6 col-md-6\"><ul class=\"features-list list-unstyled m-0\"><li class=\"features-list__item d-flex align-items-center\"><i class=\"features-list__item--icon mr-2\">Emoji's and fonts just work<\/i><\/li><li class=\"features-list__item d-flex align-items-center\"><i class=\"features-list__item--icon mr-2\">Zero-setup in your app<\/i><\/li><li class=\"features-list__item d-flex align-items-center\"><i class=\"features-list__item--icon mr-2\">Add or remove instances<\/i><\/li><\/ul><\/div><\/div><\/div><\/div><div class=\"col-lg-6 text-center text-lg-right position-relative\"><div class=\"code_frame__3LBQB\"><pre>\/\/ Replace puppeteer.launch with puppeteer.connect\nconst browser = await puppeteer.connect({\n  browserWSEndpoint: 'wss:\/\/chrome.browserless.io\/'\n});\n\/\/ Everything else stays the same\nconst page = await browser.newPage();\nawait page.goto('https:\/\/example.com\/');\nawait page.screenshot({ path: 'screenshot.png' });\nbrowser.close();<\/pre><\/div><\/div><\/div><\/div><\/div><\/main><section id=\"pricing\" class=\"pricing pricing-1\"><div class=\"container\"><div class=\"row\"><div class=\"col-lg-4\"><div class=\"pricing-card pricing-card\"><h5 class=\"pricing-card__title mb-4\">Small<\/h5><h3 class=\"pricing-card__value\">$30month<\/h3><ul class=\"pricing-card__list list-unstyled mb-0 pb-1\"><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>No maximum API or session calls<\/strong><i data-feather=\"help-circle\" data-toggle=\"tooltip\" data-placement=\"top\" title=\"Tooltip on top\"><\/i><\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>10<\/strong> running browsers<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Email support<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Static sites with images<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Dynamic sites with interactivity<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Large sites with webgl and video<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Live\/Chat support<\/i><\/li><\/ul><p><a href=\"https:\/\/www.browserless.io\/sign-up?size=small\" class=\"btn btn--blue btn--lg\">Get Started<\/a><\/p><\/div><\/div><div class=\"col-lg-4\"><div class=\"pricing-card pricing-card-highlighted\"><h6 class=\"text-uppercase mb-0\">Popular Option<\/h6><h5 class=\"pricing-card__title mb-4\">Medium<\/h5><h3 class=\"pricing-card__value\">$50month<\/h3><ul class=\"pricing-card__list list-unstyled mb-0 pb-1\"><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>No maximum API or session calls<\/strong><i data-feather=\"help-circle\" data-toggle=\"tooltip\" data-placement=\"top\" title=\"Tooltip on top\"><\/i><\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>15<\/strong> running browsers<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Email support<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Static sites with images<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Dynamic sites with interactivity<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Large sites with webgl and video<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__false mr-2\">Live\/Chat support<\/i><\/li><\/ul><p><a href=\"https:\/\/www.browserless.io\/sign-up?size=medium\" class=\"btn btn--green btn--lg\">Get Started<\/a><\/p><\/div><\/div><div class=\"col-lg-4\"><div class=\"pricing-card\"><h5 class=\"pricing-card__title mb-4\">Large<\/h5><h3 class=\"pricing-card__value\">$100month<\/h3><ul class=\"pricing-card__list list-unstyled mb-0 pb-1\"><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>No maximum API or session calls<\/strong><i data-feather=\"help-circle\" data-toggle=\"tooltip\" data-placement=\"top\" title=\"Tooltip on top\"><\/i><\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\"><strong>20<\/strong> running browsers<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Email support<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Static sites with images<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Dynamic sites with interactivity<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Large sites with webgl and video<\/i><\/li><li class=\"d-flex align-items-center\"><i class=\"pricing-icon pricing-icon__true mr-2\">Live\/Chat support<\/i><\/li><\/ul><p><a href=\"https:\/\/www.browserless.io\/sign-up?size=large\" class=\"btn btn--blue btn--lg\">Get Started<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"container\"><h5><a href=\"https:\/\/www.browserless.io\/commercial-license\">Click here<\/a> to purchase a commercial license.<\/h5><\/div><\/section><section class=\"features-1 section section--padding__wmargin\"><div class=\"container\"><div class=\"row\"><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>Turn up the speed<\/h5><p class=\"pr-4\">Increase your app's performance by not having to manage Chrome or other browsers. Our smallest plan gives you access to 10 browser instances.<\/p><\/div><\/div><\/div><\/div><div class=\"row\"><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>No time restrictions<\/h5><p class=\"pr-4\">Sessions can run as long as you want, and even keep the browser open indefinitely.<\/p><\/div><\/div><\/div><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>No software or packages to install<\/h5><p class=\"pr-4\">Stop trying to get Chrome running in lambda, or fonts rendering properly and just use browserless.<\/p><\/div><\/div><\/div><div class=\"col-lg-4\"><div class=\"feature d-flex\"><div><h5>Crucial notifications and metrics<\/h5><p class=\"pr-4\">Your account page shows crucial information like sessions and queues plus includes email notifications.<\/p><\/div><\/div><\/div><\/div><\/div><\/section><section id=\"integrations\" class=\"section section--padding section--gradient__1\"><div class=\"container\"><div class=\"row\"><div class=\"preamble preamble--light col-lg-8 offset-lg-2\"><h3 class=\"title\">How it works<\/h3><p class=\"lead\">browserless takes care of all the dependencies, sandboxing, and management of the web browser. You can connect remotely and drive it with open-source libraries, use our pre-built REST APIs, or write and run your own functions with webhooks.<\/p><\/div><\/div><div class=\"row d-flex align-items-center\"><div class=\"col-lg-5 order-2 order-lg-1 mt-5 mt-lg-0\"><ul class=\"features list-unstyled mb-0\"><li class=\"d-flex active\" data-example=\"puppeteer\"><div class=\"ml-4\"><h5>Puppeteer Connect<\/h5><p>If you're using puppeteer's launch method, simply change it to connect and specify it to use browserless.<\/p><\/div><\/li><li class=\"d-flex\" data-example=\"webdriver\"><div class=\"ml-4\"><h5>Webdriver<\/h5><p>Use libraries like selenium with our new \/webdriver route, which implements the webdriver protocol<\/p><\/div><\/li><li class=\"d-flex\" data-example=\"rest\"><div class=\"ml-4\"><h5>REST API<\/h5><p>Use our pre-made API's like \/screenshot or \/pdf to do common tasks, or write your own!<\/p><\/div><\/li><\/ul><\/div><div class=\"col-lg-7 order-1 order-lg-2\"><div class=\"code_frame__3LBQB\"><pre>\/\/ From inside your Node application\nimport puppeteer from 'puppeteer';\n\/\/ Replace puppeteer.launch with puppeteer.connect\nconst browser = await puppeteer.connect({\n  browserWSEndpoint: 'wss:\/\/chrome.browserless.io'\n});\n\/\/ The rest of your script remains the same\nconst page = await browser.newPage();\nawait page.goto('https:\/\/example.com\/');\nawait page.screenshot({ path: 'screenshot.png' });\npage.close();<\/pre><\/div><\/div><\/div><\/div><\/section><section id=\"clients\" class=\"clients-1 section section--padding\"><div class=\"container section--padding__bottom\"><div class=\"row\"><div class=\"preamble col-lg-8 offset-lg-2\"><h3 class=\"title\">Words from our users<\/h3><p class=\"lead\">With over 10 million sessions executed, here's what users have to say.<\/p><\/div><\/div><\/div><\/section><section id=\"questions\" class=\"section section--padding\"><div class=\"container\"><div class=\"row\"><h3 class=\"title\">Frequently Asked Questions<\/h3><\/div><div class=\"row\"><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>I'm not a developer, what's this for?<\/h5><p>You can use browserless to do <i>automatically<\/i> what you would normally do <i>manually<\/i>. For instance you can use it to take pictures of your web-site (for social media), or generate nice documents for your client's dashboards.<\/p><p>Getting Chrome, or any browser, to run on most web servers is pretty challenging as well. There's more than 20 additional packages you'll need to install, a way to manage the browser so it doesn't crash your servers, and a way to isolate your sessions so it's a fresh browser every time. We take care of all of that plus add notifications and a dashboard for you.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>Is there a way to check how my workers are doing?<\/h5><p>Yup! We have a GraphQL API that allows you to check your workers health prior to sending a request or connecting. You can use the online GraphiQL editor to see all the <a href=\"https:\/\/api.browserless.io\/graphql?query=%7B%0A%20%20pressure(apiToken%3A%20%22INSERT-YOUR-TOKEN-HERE%22)%20%7B%0A%20%20%20%20recentlyRejected%0A%20%20%20%20isAvailable%0A%20%20%20%20queued%0A%20%20%20%20running%0A%20%20%20%20date%0A%20%20%7D%0A%7D\">fields here<\/a>. We also have an admin dashboard that shows important statistics like rejected or queued request counts.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>Google Cloud supports Chrome, why use browserless?<\/h5><p>Cost is the first reason, as running a single browser can cost up to $42.77 a month at full utilization with Functions, and $44.38 using App Engine with their recommended settings. If you were to run 10 concurrently on Functions (which is what our $30 a month plan supports), your cost would be $427 a month. It's not uncommon to have cases where the browser won't close properly, meaning your functions over-run and end up costing more.<\/p><p>We also have numerous features over most cloud-providers, including: human-only support, numerous REST API's to do common actions, push-button upgrades, and (coming soon) support for other browsers plus Selenium web-drivers.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>I'm not using Node, can I still use this service?<\/h5><p>Even if you're not using NodeJS or puppeteer, you can still make use of our function API that allows you to send in a body of code to run for you. This allows you to keep your stack the same without making changes. You can read more about the <a href=\"https:\/\/docs.browserless.io\/blog\/2018\/05\/26\/metered-functions.html#browserless-functions\">function endpoint here<\/a>.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>What libraries do you support?<\/h5><p>Currently we support only puppeteer, but are looking to expand and integrate other libraries soon. If you can't use puppeteer due to the nature of your project you can use the <a href=\"https:\/\/docs.browserless.io\/blog\/2018\/05\/26\/metered-functions.html#browserless-functions\">function endpoint here<\/a>.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>How secure is my account?<\/h5><p>We don't store any sensitive information in browserless whatsoever. We use Stripe's payment API, so we never store credit card or other personal information, and we securely store your password with a one-way hashing algorithm. Your email is only used to contact you about updates to the service, and we don't sell or expose it to third-parties.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>How fast is the service?<\/h5><p>It generally takes about 5 minutes to generate the workers when you sign-up. When connecting via puppeteer.connect, this generally happens in less than 50 milliseconds. On average a screenshot capture happens in about a second for static pages, and PDFs take roughly 2 seconds.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>How reliable is the service?<\/h5><p>See for yourself on the <a href=\"https:\/\/status.browserless.io\/\">status page<\/a>. Since each account gets its own dedicated instances there's no over-crowding or congestion.<\/p><\/div><div class=\"col-lg-4 col-md-6 mt-5 mt-md-0\"><h5>Does this support proxies?<\/h5><p>Yes, we support using third-party proxies. See our documentation on <a href=\"https:\/\/docs.browserless.io\/docs\/chrome.html#launching-with-a-proxy\">how to use proxies<\/a>.<\/p><\/div><\/div><\/div><\/section>","created_at":"2018-11-07T16:22:46+0000","updated_at":"2018-11-07T16:22:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"www.browserless.io","preview_picture":"https:\/\/browserless.io\/assets\/browserless-social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":101,"title":"Building a Kickass Portfolio | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=17671490","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673352\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Zyst\" class=\"hnuser\">Zyst<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675631\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cerberusss\" class=\"hnuser\">cerberusss<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678364\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=spiritcat\" class=\"hnuser\">spiritcat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676972\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=finaliteration\" class=\"hnuser\">finaliteration<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677408\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gitgud\" class=\"hnuser\">gitgud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676840\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JoshuaRLi\" class=\"hnuser\">JoshuaRLi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677839\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676118\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=matallo\" class=\"hnuser\">matallo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674747\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676102\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dramatize\" class=\"hnuser\">Dramatize<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676355\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676431\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=streety\" class=\"hnuser\">streety<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676725\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17680999\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rorykoehler\" class=\"hnuser\">rorykoehler<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677276\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=streety\" class=\"hnuser\">streety<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678961\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hypertexthero\" class=\"hnuser\">hypertexthero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675279\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676031\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hypertexthero\" class=\"hnuser\">hypertexthero<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676612\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674338\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676151\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676392\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sizzle\" class=\"hnuser\">sizzle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674958\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676388\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=s73v3r_\" class=\"hnuser\">s73v3r_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673511\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rozgo\" class=\"hnuser\">rozgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673773\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gjstein\" class=\"hnuser\">gjstein<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674565\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681042\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=B0N35\" class=\"hnuser\">B0N35<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17692950\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JoshuaRLi\" class=\"hnuser\">JoshuaRLi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677834\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17680988\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JoshuaRLi\" class=\"hnuser\">JoshuaRLi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17689517\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nikaoto\" class=\"hnuser\">Nikaoto<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17691923\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stir\" class=\"hnuser\">stir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676323\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jwilbs\" class=\"hnuser\">jwilbs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678434\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codeafin\" class=\"hnuser\">codeafin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673957\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674120\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=always_good\" class=\"hnuser\">always_good<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675916\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dyarosla\" class=\"hnuser\">dyarosla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673544\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwmenow_0140\" class=\"hnuser\">throwmenow_0140<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675550\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676222\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwmenow_0140\" class=\"hnuser\">throwmenow_0140<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677916\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=trevor-e\" class=\"hnuser\">trevor-e<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676890\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681023\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ihattendorf\" class=\"hnuser\">ihattendorf<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676916\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Brajeshwar\" class=\"hnuser\">Brajeshwar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677026\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=benjvmin\" class=\"hnuser\">benjvmin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676121\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676249\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=benjvmin\" class=\"hnuser\">benjvmin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676448\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gitgud\" class=\"hnuser\">gitgud<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676407\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=benjvmin\" class=\"hnuser\">benjvmin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676447\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bbaumgar\" class=\"hnuser\">bbaumgar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673639\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=egypturnash\" class=\"hnuser\">egypturnash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674241\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676242\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vinceguidry\" class=\"hnuser\">vinceguidry<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674026\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bbaumgar\" class=\"hnuser\">bbaumgar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674454\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kotrunga\" class=\"hnuser\">kotrunga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673840\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bbaumgar\" class=\"hnuser\">bbaumgar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674442\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nightski\" class=\"hnuser\">nightski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674928\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=s73v3r_\" class=\"hnuser\">s73v3r_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674778\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Hasz\" class=\"hnuser\">Hasz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674158\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=neogodless\" class=\"hnuser\">neogodless<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674286\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=osiutino\" class=\"hnuser\">osiutino<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677378\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jamestimmins\" class=\"hnuser\">jamestimmins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673725\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bbaumgar\" class=\"hnuser\">bbaumgar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674465\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=WilliamEdward\" class=\"hnuser\">WilliamEdward<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673801\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kmgr\" class=\"hnuser\">kmgr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675290\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mblode\" class=\"hnuser\">mblode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677402\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Yahivin\" class=\"hnuser\">Yahivin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673939\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674051\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=antcas\" class=\"hnuser\">antcas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677582\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681015\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jamestimmins\" class=\"hnuser\">jamestimmins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678019\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kotrunga\" class=\"hnuser\">kotrunga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673820\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673857\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kotrunga\" class=\"hnuser\">kotrunga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673952\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pera\" class=\"hnuser\">pera<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674054\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kotrunga\" class=\"hnuser\">kotrunga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674086\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JoshuaRLi\" class=\"hnuser\">JoshuaRLi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17695736\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gerogerke\" class=\"hnuser\">gerogerke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17680020\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681008\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nfriedly\" class=\"hnuser\">nfriedly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674723\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brycedorn\" class=\"hnuser\">brycedorn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675970\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mixmastamyk\" class=\"hnuser\">mixmastamyk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17680329\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675993\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ddoolin\" class=\"hnuser\">ddoolin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676385\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=droxey\" class=\"hnuser\">droxey<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674822\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nikofeyn\" class=\"hnuser\">nikofeyn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675375\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icebraining\" class=\"hnuser\">icebraining<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676250\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sizzle\" class=\"hnuser\">sizzle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674945\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676270\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=maaaats\" class=\"hnuser\">maaaats<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675652\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AlphaWeaver\" class=\"hnuser\">AlphaWeaver<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676592\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hnarayanan\" class=\"hnuser\">hnarayanan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17682619\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=foxhop\" class=\"hnuser\">foxhop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676577\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bastijn\" class=\"hnuser\">bastijn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675017\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thebradbain\" class=\"hnuser\">thebradbain<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677007\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=NewHatMatt\" class=\"hnuser\">NewHatMatt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674315\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676319\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=prince005\" class=\"hnuser\">prince005<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678061\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681047\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nsomaru\" class=\"hnuser\">nsomaru<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678077\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=prince005\" class=\"hnuser\">prince005<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678263\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moeamaya\" class=\"hnuser\">moeamaya<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674240\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674717\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ivanech\" class=\"hnuser\">ivanech<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675679\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676299\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ivanech\" class=\"hnuser\">ivanech<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676628\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677323\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stir\" class=\"hnuser\">stir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676469\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ivanech\" class=\"hnuser\">ivanech<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676610\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Naracion\" class=\"hnuser\">Naracion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678239\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681055\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Naracion\" class=\"hnuser\">Naracion<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17717210\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=whostolemyhat\" class=\"hnuser\">whostolemyhat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17679643\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17680986\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ChrisLomont\" class=\"hnuser\">ChrisLomont<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675326\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dcposch\" class=\"hnuser\">dcposch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677563\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JoshuaRLi\" class=\"hnuser\">JoshuaRLi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17695764\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tekknolagi\" class=\"hnuser\">tekknolagi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674222\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ckuhl\" class=\"hnuser\">ckuhl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681525\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=whicks\" class=\"hnuser\">whicks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675477\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=krtkush\" class=\"hnuser\">krtkush<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678127\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681059\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=petrocrat\" class=\"hnuser\">petrocrat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675353\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=canadianwriter\" class=\"hnuser\">canadianwriter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675267\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=canadianwriter\" class=\"hnuser\">canadianwriter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675382\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jonathanhefner\" class=\"hnuser\">jonathanhefner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676815\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=acconrad\" class=\"hnuser\">acconrad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673580\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dyarosla\" class=\"hnuser\">dyarosla<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673812\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=acconrad\" class=\"hnuser\">acconrad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675275\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673813\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=azhenley\" class=\"hnuser\">azhenley<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673695\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673767\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=onirom\" class=\"hnuser\">onirom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675030\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pjungwir\" class=\"hnuser\">pjungwir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677632\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=n_\" class=\"hnuser\">n_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676847\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sincerely\" class=\"hnuser\">sincerely<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677018\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=n_\" class=\"hnuser\">n_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677762\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rectangletangle\" class=\"hnuser\">rectangletangle<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673378\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwmenow_0140\" class=\"hnuser\">throwmenow_0140<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675643\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dramatize\" class=\"hnuser\">Dramatize<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676014\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stir\" class=\"hnuser\">stir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676405\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676039\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lgregg\" class=\"hnuser\">lgregg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676148\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Dramatize\" class=\"hnuser\">Dramatize<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676337\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=apankrat\" class=\"hnuser\">apankrat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675501\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Jac21\" class=\"hnuser\">Jac21<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681338\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gdubs\" class=\"hnuser\">gdubs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676368\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevejohnson\" class=\"hnuser\">stevejohnson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674841\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=praveenscience\" class=\"hnuser\">praveenscience<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17682350\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nfriend\" class=\"hnuser\">nfriend<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674034\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hnarayanan\" class=\"hnuser\">hnarayanan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675209\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=refrigerator\" class=\"hnuser\">refrigerator<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673994\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674670\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CogitoCogito\" class=\"hnuser\">CogitoCogito<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17682205\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RpFLCL\" class=\"hnuser\">RpFLCL<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674845\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=xPaw\" class=\"hnuser\">xPaw<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674300\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Zezima\" class=\"hnuser\">Zezima<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675580\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=heiswayi\" class=\"hnuser\">heiswayi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17685549\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ericliuche\" class=\"hnuser\">ericliuche<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675226\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jalcine\" class=\"hnuser\">jalcine<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674796\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676340\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=unhingingdog\" class=\"hnuser\">unhingingdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676175\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676234\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=unhingingdog\" class=\"hnuser\">unhingingdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676327\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stolson\" class=\"hnuser\">stolson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674347\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namanyayg\" class=\"hnuser\">namanyayg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674418\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bdcravens\" class=\"hnuser\">bdcravens<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674617\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namanyayg\" class=\"hnuser\">namanyayg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674662\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gbuk2013\" class=\"hnuser\">gbuk2013<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674561\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=palerdot\" class=\"hnuser\">palerdot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17680326\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=boogadaboogada\" class=\"hnuser\">boogadaboogada<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674383\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cjcenizal\" class=\"hnuser\">cjcenizal<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673882\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rezashirazian\" class=\"hnuser\">rezashirazian<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675411\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=volkk\" class=\"hnuser\">volkk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673426\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=B0N35\" class=\"hnuser\">B0N35<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17692935\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pSub_\" class=\"hnuser\">pSub_<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674066\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andrew_wc_brown\" class=\"hnuser\">andrew_wc_brown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676072\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wild_preference\" class=\"hnuser\">wild_preference<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677040\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ohmatt\" class=\"hnuser\">ohmatt<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676638\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=majormajor\" class=\"hnuser\">majormajor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676173\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=alexpetralia\" class=\"hnuser\">alexpetralia<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676108\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ace_of_spades\" class=\"hnuser\">ace_of_spades<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676273\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=r3bl\" class=\"hnuser\">r3bl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678296\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676693\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=philipmjohnson\" class=\"hnuser\">philipmjohnson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676719\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=spyhi\" class=\"hnuser\">spyhi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676752\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rainonmoon\" class=\"hnuser\">rainonmoon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676779\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ngold\" class=\"hnuser\">ngold<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677279\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wongma\" class=\"hnuser\">wongma<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677540\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mendelk\" class=\"hnuser\">mendelk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674125\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gumoro\" class=\"hnuser\">gumoro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674892\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lanzaa\" class=\"hnuser\">Lanzaa<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675124\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mendelk\" class=\"hnuser\">mendelk<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675229\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pletnes\" class=\"hnuser\">pletnes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674876\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pmiller2\" class=\"hnuser\">pmiller2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676218\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676706\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=vxNsr\" class=\"hnuser\">vxNsr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677244\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678737\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=firefoxd\" class=\"hnuser\">firefoxd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677788\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=J-dawg\" class=\"hnuser\">J-dawg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674199\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_8ks0\" class=\"hnuser\">_8ks0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674610\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=abledon\" class=\"hnuser\">abledon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674525\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=J-dawg\" class=\"hnuser\">J-dawg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674690\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Jarwain\" class=\"hnuser\">Jarwain<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17689081\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=namanyayg\" class=\"hnuser\">namanyayg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674499\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kickscondor\" class=\"hnuser\">kickscondor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676353\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pixelbash\" class=\"hnuser\">pixelbash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674271\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lawn\" class=\"hnuser\">lawn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674957\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fogleman\" class=\"hnuser\">fogleman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674791\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kotrunga\" class=\"hnuser\">kotrunga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675096\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ghostbrainalpha\" class=\"hnuser\">ghostbrainalpha<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17672579\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dudul\" class=\"hnuser\">dudul<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674907\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Sytten\" class=\"hnuser\">Sytten<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676914\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gbear605\" class=\"hnuser\">gbear605<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673698\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pknerd\" class=\"hnuser\">pknerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674755\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=r3bl\" class=\"hnuser\">r3bl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678972\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676722\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pknerd\" class=\"hnuser\">pknerd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17686075\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=deepaksurti\" class=\"hnuser\">deepaksurti<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17679621\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rozgo\" class=\"hnuser\">rozgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673636\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=servercobra\" class=\"hnuser\">servercobra<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673173\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676751\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wicker\" class=\"hnuser\">wicker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674930\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kelvin0\" class=\"hnuser\">kelvin0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674699\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nfriedly\" class=\"hnuser\">nfriedly<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674906\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=h4b4n3r0\" class=\"hnuser\">h4b4n3r0<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676598\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Baeocystin\" class=\"hnuser\">Baeocystin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677113\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676739\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eafkuor\" class=\"hnuser\">eafkuor<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674784\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=petepete\" class=\"hnuser\">petepete<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17672828\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677096\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=petepete\" class=\"hnuser\">petepete<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17678173\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kayman\" class=\"hnuser\">kayman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677651\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KennyCason\" class=\"hnuser\">KennyCason<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17675655\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676731\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KennyCason\" class=\"hnuser\">KennyCason<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677447\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nixpulvis\" class=\"hnuser\">nixpulvis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673680\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fvcproductions\" class=\"hnuser\">fvcproductions<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17681493\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kagerjay\" class=\"hnuser\">Kagerjay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674149\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=asimpletune\" class=\"hnuser\">asimpletune<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674410\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=pypanta\" class=\"hnuser\">pypanta<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17677712\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=terkel\" class=\"hnuser\">terkel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674185\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=throwawayaway12\" class=\"hnuser\">throwawayaway12<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674519\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=edhu2017\" class=\"hnuser\">edhu2017<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674678\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MikeKusold\" class=\"hnuser\">MikeKusold<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674656\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aspit\" class=\"hnuser\">aspit<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17676756\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lbotos\" class=\"hnuser\">lbotos<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17674549\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kermittd\" class=\"hnuser\">kermittd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17673506\">3 months ago<\/a><\/p><br \/><\/td><\/tr><\/table><table border=\"0\"><tr><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/td><td class=\"title\"><a href=\"https:\/\/news.ycombinator.com\/item?id=17671490&amp;p=2\" class=\"morelink\" rel=\"nofollow\">More<\/a><\/td><\/tr><\/table>","created_at":"2018-11-07T16:22:45+0000","updated_at":"2018-11-07T16:22:45+0000","annotations":[],"mimetype":"text\/html","reading_time":5,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":102,"title":"Literary Clock Made From E-reader","url":"http:\/\/www.instructables.com\/id\/Literary-Clock-Made-From-E-reader\/","content":"<section id=\"step1\" class=\"step\" data-stepid=\"S7T68NUJI4U6PYZ\"><h2 class=\"step-title\">Step 1: Tools and Materials<\/h2><noscript><img alt=\"Picture of Tools and Materials\" src=\"https:\/\/cdn.instructables.com\/FIS\/OD3L\/JK4UJIDT\/FISOD3LJK4UJIDT.LARGE.jpg\" width=\"4912\" height=\"3264\" \/><\/noscript><div class=\"step-body\"><p>Really the only thing needed is an e-reader (and a USB cable to connect to it). For this project, a Kindle was donated to me by a <a href=\"https:\/\/twitter.com\/mfrissen\" rel=\"nofollow\">friend<\/a>. It is a Kindle 3 WiFi (nicknamed K3, or K3W). You will find many second-hand earlier models like that on eBay for instance.<\/p><p>You'll need a computer (any operating system), with an SSH client like vSSH and an sFTP client like Filezilla installed (both are free). It helps to have a bit of experience with Linux, because that is what the Kindle runs on.<\/p><p>To have the Kindle stand upright in our cupboard, I made a stand from concrete. If you want to do the same, you'll need a food container in a shape you like, cling film, styrofoam, cement, hot glue or double sided tape, and a bucket (to mix the cement).<\/p><\/div><\/section><section id=\"step2\" class=\"step\" data-stepid=\"S3OFWMMJI4U08CW\"><h2 class=\"step-title\">Step 2: Jailbreaking the Kindle<\/h2><noscript><img alt=\"Picture of Jailbreaking the Kindle\" src=\"https:\/\/cdn.instructables.com\/FEN\/F93D\/JK4UKY5I\/FENF93DJK4UKY5I.LARGE.jpg\" width=\"3046\" height=\"1804\" \/><\/noscript><div class=\"step-body\"><p>In order to change the Kindle into a clock, we need to get into the system files. In order to do that, we need to open it up through a process called 'jailbreaking' (don't worry, it's not illegal if it's your property). An explanation to jailbreak the Kindle and a zip file with the necessary files can be found <a href=\"https:\/\/wiki.mobileread.com\/wiki\/Kindle_Screen_Saver_Hack_for_all_2.x,_3.x_%26_4.x_Kindles\" rel=\"nofollow\">here<\/a>. Also see <a href=\"https:\/\/www.mobileread.com\/forums\/showthread.php?t=88004\" rel=\"nofollow\">this overview<\/a> of all available software custom software. Find out which Kindle model you have on <a href=\"https:\/\/wiki.mobileread.com\/wiki\/Kindle_Serial_Numbers\" rel=\"nofollow\">this page<\/a>.<\/p><p>For this project, you only need to install the jailbreak hack and the <a href=\"https:\/\/www.mobileread.com\/forums\/showthread.php?t=88004\" rel=\"nofollow\">usbnet hack<\/a>, not the screen saver hack. USBNetwork will grant you remote shell access to your Kindle, either over USB or WiFi. What you will need, if you want to use the keyboard's keys, is the <a href=\"https:\/\/www.mobileread.com\/forums\/showthread.php?t=97636\" rel=\"nofollow\">Launchpad<\/a> hack.<\/p><p><em>Warning: I read this can potentially ruin your Kindle. Follow the instructions. Jailbreak at your own risk.<\/em><\/p><p>If you connect the Kindle to your computer, it will show up as an USB drive.  <\/p><p>Basically, all you need to do is put Update_jailbreak_0.13.N_***_install.bin (where *** is your Kindle version, in my case 'k3w') in the root folder of the Kindle when it is connected to your computer. <\/p><p>From the README file in the zip file: \"Now, eject &amp; unplug your Kindle, and go to *[HOME] -&gt; [MENU] &gt; Settings -&gt; [MENU] &gt; Update Your Kindle*. It should be quick.\u201d (note: that's two times clicking the menu button).<\/p><p>Then do the same for the USBNet and Launchpad files. You should now be able to log in to the device using SSH. On the Kindle, connect to the WiFi network. One way to find out its IP address is by logging into your WiFi router and looking it up there. Username is 'root', and the default root password for your model can be <a href=\"https:\/\/www.sven.de\/kindle\/?#\" rel=\"nofollow\">calculated<\/a>.<\/p><p>Then install Python on the Kindle, again using the <a href=\"https:\/\/www.mobileread.com\/forums\/showthread.php?t=88004\" rel=\"nofollow\">files<\/a> on the excellent Mobileread.com forum (thanks VoltaX2 in the comments below).<\/p><\/div><\/section><section id=\"step3\" class=\"step\" data-stepid=\"SH1K3G0JI4U6O4S\"><h2 class=\"step-title\">Step 3: Making an Image for Every Single Minute of the Day<\/h2><noscript><img alt=\"Picture of Making an Image for Every Single Minute of the Day\" src=\"https:\/\/cdn.instructables.com\/FQE\/1T2D\/JK4UKTCK\/FQE1T2DJK4UKTCK.LARGE.jpg\" width=\"4709\" height=\"3210\" \/><\/noscript><div class=\"step-body\"><p>There are 1,440 minutes in a day. Compiling a list with quotes for each and every one of them from different literary works is a massive undertaking. Big relief: others already did that for us.<\/p><p>In 2011, newspaper The Guardian asked its readers to submit quotes from books which mention times. They wanted to build an installation for a literary festival. So they have two versions of a list on their website (<a href=\"https:\/\/www.theguardian.com\/books\/table\/2011\/apr\/21\/literary-clock?CMP=twt_gu\" rel=\"nofollow\">1<\/a>, <a href=\"https:\/\/www.guardian.co.uk\/books\/booksblog\/2011\/apr\/15\/christian-marclay-the-clock-literature\" rel=\"nofollow\">2<\/a>).<\/p><p>I combined the two lists, cleaned them up, added a few times I found myself, and turned them into one CSV file.<\/p><p>Unfortunately the list does not cover all minutes of the day. I worked around this by using some quotes more than once, for instance if it can be used both in the AM and PM. More vague time indications can be used around a certain time, so this quote from Catcher in the Rye is used at 9.58AM: \"I didn't sleep too long, because I think it was only around ten o'clock when I woke up ... \"<\/p><p>Even with this pleasant list, two things took me an unreasonable amount of time. I needed to turn every single quotation from the list into an image. I wanted to make them fit nicely to the screen, so the font would be as large as possible for each quotation. <\/p><p>While scaling a text box to a certain height and width is easy to do manually in most photo editing software, it would have been an immense amount of work to create them one by one. Creating a script to do it for me however proved to be quite the task as well. In PHP (I used that programming language because it has nice functions to deal with text) I wrote a recursive function to find the best fit for each quotation, long or short. For each line, the script creates two PNG images, one with and one without metadata.<\/p><p>It uses the <a href=\"http:\/\/libertine-fonts.org\/show-me\/\" rel=\"nofollow\">Libertine font<\/a>, which I like because of its stylish look, because it is very complete (numbers, punctuation, diacritics) and because it's open source.<\/p><p>The other thing that took me a long time is identifying all time mentions in the quotations, because I wanted to write them in bold text. That makes the clock easier to use, especially when a quote is quite long. The problem is that in books, an impressive variation of time descriptions is used. It can be anything from '6.00 p.m.' or '18:11:00' to '0600h', 'around six o'clock', just 'at six', or 'twenty-eight minutes past eleven'. I made a script to try and find most of these variations, did the ones it couldn't find myself, and added them to the csv file. <\/p><p>If you want to make your own Kindle clock, you may use my scripts (find them attached below), but you can also just download all the resulting images.<\/p><\/div><aside class=\"downloads\"><h4 class=\"sr-only\">Attachments<\/h4><\/aside><\/section><section id=\"step4\" class=\"step\" data-stepid=\"SQ6ZAK1JI4U6QHB\"><h2 class=\"step-title\">Step 4: Starting and Stopping the Clock<\/h2><noscript><img alt=\"Picture of Starting and Stopping the Clock\" src=\"https:\/\/cdn.instructables.com\/FBP\/YNJJ\/JK4UKT6K\/FBPYNJJJK4UKT6K.LARGE.jpg\" width=\"4583\" height=\"3264\" \/><\/noscript><div class=\"step-body\"><p>I wanted to be able to start my literary clock by pressing the shortcut Shift+C on the small keyboard of the e-reader. Pressing it again stops the clock and turns the clock into a normal e-reader again.<\/p><p>First, create this folder: \/mnt\/us\/timelit and then put the scripts I attached below in there.<\/p><p>The images (see previous step) go into \/mnt\/us\/timelit\/images and \/mnt\/us\/timelit\/images\/metadata\/ <\/p><p>When you install the Launchpad hack, the folder \/mnt\/us\/launchpad is created. Create a new file there called startClock.ini and put this text in there:<\/p><pre>[Actions]\nC = !sh \/mnt\/us\/timelit\/startstopClock.sh &amp;<\/pre><p>That creates the shortcut Shift+C. If we press that, the bash-script startstopClock.sh starts. It stops the Kindle framework (the normal user interface), prevents the Kindle from going into power save mode and creates a small file (\/mnt\/us\/timelit\/clockisticking) to indicate the clock has started.<\/p><p>If the user presses Shift+C again and the clockisticking file is already there, startstopClock.sh will remove it and restart the Kindle.<\/p><p>startstopClock.sh also executes another script, showMetadata.sh, to enable the keystrokes that will show the metadata (using the command \/usr\/bin\/waitforkey). If the user pushes the 'next page' button on the sides of the Kindle, it will check if the clock is ticking and if it is, will show the same image as currently is shown (which file that is, is saved in the clockisticking file) but then with title and author at the bottom.<\/p><p>Changing the time on the display every minute is done by adding this line to \/etc\/crontab\/root:<\/p><pre>* * * * * sh \/mnt\/us\/timelit\/timelit.sh<\/pre><p>and then restart crontab like this: \/etc\/init.d\/cron restart<\/p><p>Every time it is run, timelit.sh checks if the 'clockisticking' file is created. If it is, timelit.sh proceeds to show the image for the current minute.<\/p><p><em>Note: you'll probably want to change the timezone in timelit.sh where it says 'TZ=CEST'.<\/em><\/p><\/div><aside class=\"downloads\"><h4 class=\"sr-only\">Attachments<\/h4><\/aside><\/section><section id=\"step5\" class=\"step\" data-stepid=\"S9F8QMLJI4U6Q86\"><h2 class=\"step-title\">Step 5: Making a Stand<\/h2><noscript><img alt=\"Picture of Making a Stand\" src=\"https:\/\/cdn.instructables.com\/F4Z\/VDNX\/JK4UJIE9\/F4ZVDNXJK4UJIE9.LARGE.jpg\" width=\"3264\" height=\"4912\" \/><img alt=\"Picture of Making a Stand\" src=\"https:\/\/cdn.instructables.com\/FBE\/UT8O\/JK4UJIEB\/FBEUT8OJK4UJIEB.LARGE.jpg\" width=\"4912\" height=\"3264\" \/><img alt=\"Picture of Making a Stand\" src=\"https:\/\/cdn.instructables.com\/FBP\/IV2V\/JK4UJII5\/FBPIV2VJK4UJII5.LARGE.jpg\" width=\"4912\" height=\"3264\" \/><img alt=\"Picture of Making a Stand\" src=\"https:\/\/cdn.instructables.com\/FCX\/X96U\/JK4UJII4\/FCXX96UJK4UJII4.LARGE.jpg\" width=\"4912\" height=\"3264\" \/><img alt=\"Picture of Making a Stand\" src=\"https:\/\/cdn.instructables.com\/FEK\/92ET\/JK4UJIFN\/FEK92ETJK4UJIFN.LARGE.jpg\" width=\"4912\" height=\"3264\" \/><img alt=\"Picture of Making a Stand\" src=\"https:\/\/cdn.instructables.com\/FCL\/Y99F\/JK4UJIFO\/FCLY99FJK4UJIFO.LARGE.jpg\" width=\"4912\" height=\"3264\" \/><\/noscript><div class=\"step-body\"><p>I was inspired by <a href=\"https:\/\/www.instructables.com\/id\/DIY-CONCRETE-Concrete-iPad-Stand\/\" rel=\"nofollow\">other<\/a> Instructables to make a concrete stand for my Kindle clock. I could also have made something out of wood (or even <a href=\"https:\/\/www.instructables.com\/id\/Book-to-Tablet-Stand\/\" rel=\"nofollow\">a book<\/a>), but I liked to try cement because I never did before and also because I thought the grey color would go nicely with the e-reader.<\/p><p>I cut a piece of styrofoam the size of the e-reader, plus a little extra for the USB cable to go in. I wrapped it in cling film and a bit of clear tape, so the cement would come off easily afterwards. I taped it to the bottom of the food container using double-sided tape.<\/p><p>Then I mixed enough cement to fill the food container to about 5 centimeters (2\") deep. I'm not sure, but I may not have used enough water, because the cement was less pourable than I had expected. I definitely should follow the <a href=\"https:\/\/www.instructables.com\/class\/Concrete-Class\/\" rel=\"nofollow\">concrete class<\/a> before my next try :-) <\/p><p>I put the cement in the container using a garden shovel, tamped it a bit, and then let it dry for two days.  <\/p><p>The next time I will try for a smoother surface by first sifting the cement to get rid of the small rocks, adding a bit more water and spend more time sanding the result. Then I will also make a small recess in the base so the USB cable goes to the back of the stand. This can be done using a straw. <\/p><\/div><\/section><section id=\"step6\" class=\"step\" data-stepid=\"S6GQGC8JI4U6R7G\"><h2 class=\"step-title\">Step 6: Further Ideas<\/h2><noscript><img alt=\"Picture of Further Ideas\" src=\"https:\/\/cdn.instructables.com\/FDU\/BVI2\/JK4UKSRR\/FDUBVI2JK4UKSRR.LARGE.jpg\" width=\"4584\" height=\"3198\" \/><\/noscript><div class=\"step-body\"><p>The literary clock looks really nice, and the quiz part works well. My girlfriend now and then checks to see from which book a quotation is from (she usually guesses correctly :). The stand did not come out quite how I hoped, but I'm looking forward to trying making a better one.<\/p><p>I will probably also add a lamp, either clamped on the device or incorporated into the new base. When the clock sits in the cupboard, sometimes it is a little too dark to be able to tell the time. <\/p><p>Instead of getting power for a lamp separately, one could power a lamp using <a href=\"https:\/\/www.instructables.com\/id\/Kindle-3-DIY-Light\/\" rel=\"nofollow\">power from the hinge slot<\/a> in the Kindle. Two slots are there for Kindle cases that have a lamp built in. You'd have to open the Kindle and do some soldering, or make your own metal clamps, but that would be sweet. One could even connect a <a href=\"https:\/\/www.youtube.com\/watch?v=Hz-aZBEX-JA\" rel=\"nofollow\">light sensor<\/a>, so the lamp will only switch on when it's getting dark.<\/p><p>Extra features I hope to get round to are<\/p><ul><li>having the clock stop between 1am and 6am, to save power    <\/li><li>turn of wifi for the same reason, but turning it on daily for a couple of minutes to synchronise the system clock    <\/li><li>showing the percentage of the current minute that has passed as small blocks at the bottom, just like the Kindle indicates the progress the reader is making in a book    <\/li><li>show a warning when the Kindle's battery is running out<\/li><\/ul><p>(these last two could be done by overlaying small images on the larger image using the Kindle's <a href=\"https:\/\/wiki.mobileread.com\/wiki\/Eips\" rel=\"nofollow\">eips<\/a> command, see my scripts for examples).<\/p><p>Other possible ideas are<\/p><ul><li>using keys on the Kindle to set the time    <\/li><li>show a default image when the clock starts and\/or when no image is found    <\/li><li>using a shortcut (shift-Q for instance) to toggle quiz mode    <\/li><li>have the Big Ben sound chime at the top of the hour (only during the day), as the Kindle has a nice speaker built in. Other sounds could be the sound of slamming a book shut or turning pages or even reading out a  quote.<\/li><\/ul><p>I hope you like the idea and this Instructable. Let me know if you have any questions or suggestions! <\/p><\/div><\/section><div class=\"contest-entries finalized-contests\"><div class=\"contest-entry\"><img alt=\"Clocks Contest\" src=\"https:\/\/cdn.instructables.com\/F1G\/H4TQ\/JHKTU6V6\/F1GH4TQJHKTU6V6.SMALL.jpg\" \/><p>Second Prize in the <br \/><a href=\"http:\/\/www.instructables.com\/contest\/clocks2018\/\">Clocks Contest<\/a><\/p><\/div><\/div><section id=\"imadeits\" class=\"imadeits\"><h2>Share<\/h2><\/section><section class=\"recommendations\"><h2>Recommendations<\/h2><ul class=\"list-unstyled projects\"><li><article><a href=\"http:\/\/www.instructables.com\/id\/Dovetailed-Arduino-Humidity-Sensor\/\"><img alt=\"Dovetailed Arduino Humidity Sensor\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F0X\/9LNX\/JNULC1BS\/F0X9LNXJNULC1BS.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><h3><a href=\"http:\/\/www.instructables.com\/id\/Dovetailed-Arduino-Humidity-Sensor\/\">Dovetailed Arduino Humidity Sensor<\/a><\/h3><footer><p>by <a href=\"http:\/\/www.instructables.com\/member\/WickedMakers\/\">WickedMakers<\/a> in\n                        \n                            <a href=\"http:\/\/www.instructables.com\/technology\/\">Technology<\/a><\/p><\/footer><\/div><\/article><\/li><li><article><a href=\"http:\/\/www.instructables.com\/id\/Custom-Design-Vertical-Turntable\/\"><img alt=\"Custom Design Vertical Turntable\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FT7\/GSNY\/JNULAL67\/FT7GSNYJNULAL67.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><h3><a href=\"http:\/\/www.instructables.com\/id\/Custom-Design-Vertical-Turntable\/\">Custom Design Vertical Turntable<\/a><\/h3><footer><p>by <a href=\"http:\/\/www.instructables.com\/member\/Joppe%20Spaans\/\">Joppe Spaans<\/a> in\n                        \n                            <a href=\"http:\/\/www.instructables.com\/technology\/audio\/\">Audio<\/a><\/p><\/footer><\/div><\/article><\/li><li><article><a href=\"http:\/\/www.instructables.com\/id\/MIDI-Step-Interface-1\/\"><img alt=\"MIDI Step Interface\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F1N\/8G5F\/JNULDZPR\/F1N8G5FJNULDZPR.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><h3><a href=\"http:\/\/www.instructables.com\/id\/MIDI-Step-Interface-1\/\">MIDI Step Interface<\/a><\/h3><footer><p>by <a href=\"http:\/\/www.instructables.com\/member\/KILOMX\/\">KILOMX<\/a> in\n                        \n                            <a href=\"http:\/\/www.instructables.com\/technology\/audio\/\">Audio<\/a><\/p><\/footer><\/div><\/article><\/li><li><article><a href=\"http:\/\/www.instructables.com\/class\/3D-CAM-and-CNC-Class\/\"><img alt=\"3D CAM and CNC Class\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FNR\/PT8J\/JDYUQECL\/FNRPT8JJDYUQECL.RECTANGLE1.jpg?auto=webp\" \/><\/a><div class=\"caption\"><img alt=\"3D CAM and CNC Class\" class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FYG\/LHT1\/IV0AKFRL\/FYGLHT1IV0AKFRL.SQUARE2.jpg?auto=webp\" \/><h3><a href=\"http:\/\/www.instructables.com\/id\/3D-CAM-and-CNC-Class\/\">3D CAM and CNC Class<\/a><\/h3><footer><p>1,111 Enrolled<\/p><\/footer><\/div><\/article><\/li><\/ul><ul class=\"list-unstyled contests\"><li><h3 class=\"sr-only\">Plastics Contest<\/h3><a href=\"http:\/\/www.instructables.com\/contest\/plastics2018\/\"><img class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FLR\/AT0U\/JMKVA284\/FLRAT0UJMKVA284.MEDIUM.jpg?auto=webp\" alt=\"Plastics Contest\" \/><\/a><\/li><li><h3 class=\"sr-only\">Halloween Contest 2018<\/h3><a href=\"http:\/\/www.instructables.com\/contest\/halloween18\/\"><img class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/FSA\/Q4G4\/JMQKX2XX\/FSAQ4G4JMQKX2XX.MEDIUM.jpg?auto=webp\" alt=\"Halloween Contest 2018\" \/><\/a><\/li><li><h3 class=\"sr-only\">Optics Contest<\/h3><a href=\"http:\/\/www.instructables.com\/contest\/optics\/\"><img class=\"lazy-img-scroll\" src=\"https:\/\/cdn.instructables.com\/F4H\/U1WQ\/JMKV2FGF\/F4HU1WQJMKV2FGF.MEDIUM.jpg?auto=webp\" alt=\"Optics Contest\" \/><\/a><\/li><\/ul><\/section>","created_at":"2018-11-07T16:22:45+0000","updated_at":"2018-11-07T16:22:45+0000","published_at":"2018-07-29T00:00:00+0000","published_by":["Instructables"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"www.instructables.com","preview_picture":"https:\/\/cdn.instructables.com\/ORIG\/FP5\/6IAA\/JK4ULNCJ\/FP56IAAJK4ULNCJ.jpg?width=2100","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":100,"title":"adius\/awesome-sheet-music","url":"https:\/\/github.com\/adius\/awesome-sheet-music\/blob\/master\/README.md","content":"<p>A curated list of awesome tools to create, edit and display sheet music.<\/p>\n<h2><a id=\"user-content-music-notation-software\" class=\"anchor\" aria-hidden=\"true\" href=\"#music-notation-software\"><\/a>Music Notation Software<\/h2>\n<ul><li><a href=\"http:\/\/capella.de\" rel=\"nofollow\">capella<\/a> (commercial) - Create complete scores.<\/li>\n<li><a href=\"http:\/\/finalemusic.com\" rel=\"nofollow\">finale<\/a> (commercial) - Compose, arrange, notate and print\nengraver-quality sheet music.<\/li>\n<li><a href=\"http:\/\/lilypond.org\" rel=\"nofollow\">lilypond<\/a> - Program and file format for music engraving.<\/li>\n<li><a href=\"http:\/\/musescore.org\" rel=\"nofollow\">musescore<\/a> - Create, play and print sheet music.<\/li>\n<li><a href=\"http:\/\/opusmodus.com\" rel=\"nofollow\">opusmodus<\/a> (commercial) - Software for music composition.<\/li>\n<li><a href=\"http:\/\/sibelius.com\" rel=\"nofollow\">sibelius<\/a> (commercial) - Create and share scores.<\/li>\n<\/ul><h2><a id=\"user-content-archives\" class=\"anchor\" aria-hidden=\"true\" href=\"#archives\"><\/a>Archives<\/h2>\n<ul><li><a href=\"http:\/\/flutetunes.com\" rel=\"nofollow\">flutetunes.com<\/a> - Free sheet music for the flute.<\/li>\n<li><a href=\"http:\/\/imslp.org\" rel=\"nofollow\">imslp.org<\/a> - Sharing the world\u2019s public domain music.<\/li>\n<li><a href=\"https:\/\/github.com\/cellist\/Lilypond-Sheet-Music\">Lilypond-Sheet-Music<\/a> - Cello music typeset with LilyPond.<\/li>\n<li><a href=\"http:\/\/musescore.org\" rel=\"nofollow\">musescore<\/a> - Share your sheet music and comment on others.<\/li>\n<li><a href=\"http:\/\/www.mutopiaproject.org\" rel=\"nofollow\">mutopia<\/a> - Sheet music editions of classical music for free.\nRepository available at <a href=\"https:\/\/github.com\/chrissawer\/The-Mutopia-Project\">github.com mutopia<\/a>.<\/li>\n<li><a href=\"https:\/\/github.com\/veltzer\/openbook\">openbook<\/a> - Open LilyPond real book for jazz musicians.<\/li>\n<\/ul><h2><a id=\"user-content-software-libraries\" class=\"anchor\" aria-hidden=\"true\" href=\"#software-libraries\"><\/a>Software Libraries<\/h2>\n<ul><li><a href=\"http:\/\/alphatab.net\" rel=\"nofollow\">alphatab<\/a> - Cross platform music notation and\nguitar tablature rendering library.<\/li>\n<li><a href=\"https:\/\/github.com\/starenka\/chordata\">ChorData<\/a> - Dummy chordbook for mandolin, ukulele and guitar.<\/li>\n<li><a href=\"https:\/\/github.com\/adius\/lilynode\">lilynode<\/a> - Node.js wrapper for lilypond.<\/li>\n<li><a href=\"https:\/\/github.com\/adius\/lilyware\">lilyware<\/a> - Node.js middleware for serving rendered lilypond files.<\/li>\n<li><a href=\"https:\/\/github.com\/aspiers\/ly2video\">ly2video<\/a> - Generating videos from LilyPond projects.<\/li>\n<li><a href=\"https:\/\/github.com\/cuthbertLab\/music21\">music21<\/a> - Toolkit for computational musicology.<\/li>\n<li><a href=\"https:\/\/github.com\/PencilCode\/musical.js\">musical.js<\/a> - A sequencing WebAudio synthesizer\nthat supports ABC notation.<\/li>\n<li><a href=\"https:\/\/github.com\/rism-ch\/verovio\">verovio<\/a> - An library and toolkit for\nengraving MEI music notation into SVG.<\/li>\n<li><a href=\"https:\/\/github.com\/btwael\/zazate.js\">zazate.js<\/a> - Music theory and notation library\nfor javascript and node.js.<\/li>\n<\/ul><h2><a id=\"user-content-websites\" class=\"anchor\" aria-hidden=\"true\" href=\"#websites\"><\/a>Websites<\/h2>\n<ul><li><a href=\"https:\/\/chromatik.com\" rel=\"nofollow\">chromatik<\/a> (commercial) - Play along to sheet music\nof previously unreleased tunes.<\/li>\n<li><a href=\"https:\/\/flat.io\" rel=\"nofollow\">flat.io<\/a> - The online music score editor for your compositions.<\/li>\n<li><a href=\"https:\/\/getinstinct.com\" rel=\"nofollow\">getinstinct<\/a> - Guitar lessons that listen as you play along.<\/li>\n<li><a href=\"http:\/\/lilybin.com\" rel=\"nofollow\">lilybin<\/a> - Web-based LilyPond editor and [github] project.<\/li>\n<li><a href=\"http:\/\/my.vexflow.com\" rel=\"nofollow\">my.vexflow<\/a> - Publish content with music notation, guitar tablature\nand chord diagrams without the need for special tools.<\/li>\n<li><a href=\"http:\/\/noteflight.com\" rel=\"nofollow\">noteflight<\/a> (commercial) - Browser based music notation creator and\nmusic composition community.<\/li>\n<li><a href=\"https:\/\/playgroundsessions.com\" rel=\"nofollow\">playgroundsessions<\/a> (commercial) - Learn how to play the piano with\ninteractive lessons featuring your favorite songs.<\/li>\n<li><a href=\"http:\/\/soundslice.com\" rel=\"nofollow\">soundslice<\/a> (commercial) - Learn music better with interactive\nnotation and tabs.<\/li>\n<li><a href=\"http:\/\/synthesiagame.com\" rel=\"nofollow\">synthesiagame<\/a> (commercial) - Learn how to play the piano\nusing falling notes.<\/li>\n<li><a href=\"https:\/\/github.com\/tiredpixel\/tunefl\">tunefl<\/a> - Freely generate and share mini lilypond scores with ease.<\/li>\n<li><a href=\"https:\/\/utab.com\" rel=\"nofollow\">utab<\/a> - Synced Chords &amp; Lyrics created by professional musicians.<\/li>\n<\/ul><h2><a id=\"user-content-resources\" class=\"anchor\" aria-hidden=\"true\" href=\"#resources\"><\/a>Resources<\/h2>\n<ul><li><a href=\"http:\/\/freemusiced.org\/free-sheet-music-links.html\" rel=\"nofollow\">freemusiced<\/a> - Fingering charts, scale sheets,\nfree sheet music and links to resources on the internet.<\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_Online_Digital_Musical_Document_Libraries\" rel=\"nofollow\">List of Online Digital Musical Document Libraries<\/a> -\nWikipedia list of online archives.<\/li>\n<\/ul><h2><a id=\"user-content-omr-optical-music-recognition\" class=\"anchor\" aria-hidden=\"true\" href=\"#omr-optical-music-recognition\"><\/a>OMR (Optical Music Recognition)<\/h2>\n<ul><li><a href=\"http:\/\/audiveris.org\" rel=\"nofollow\">Audiveris<\/a> - Optical Music Recognition software\nwhich processes the image of a music sheet\nto automatically provide symbolic music information in MusicXML.<\/li>\n<li><a href=\"https:\/\/github.com\/DDMAL\/Rodan\">Rodan<\/a> - Web-based document recognition system.<\/li>\n<\/ul><h2><a id=\"user-content-research\" class=\"anchor\" aria-hidden=\"true\" href=\"#research\"><\/a>Research<\/h2>\n<ul><li><a href=\"http:\/\/ddmal.music.mcgill.ca\" rel=\"nofollow\">Distributed digital music archives and libraries lab<\/a> -\nDevelopment and evaluation of practices, frameworks and tools\nfor the design and construction of worldwide distributed\ndigital music archives and libraries.<\/li>\n<\/ul>","created_at":"2018-11-07T16:22:42+0000","updated_at":"2018-11-07T16:22:42+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/969983?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":99,"title":"How Does Mastodon Work? - Kev Quirk","url":"https:\/\/kevq.uk\/how-does-mastodon-work\/","content":"<p><img class=\"aligncenter size-full wp-image-387\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/mastodon-header.png\" alt=\"\" width=\"790\" height=\"450\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-header.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-header-300x171.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-header-768x437.png 768w\" \/><\/p><p>I\u2019ve mentioned <a href=\"https:\/\/joinmastodon.org\">Mastodon<\/a> on a number of occasions on this blog. It is the only social media platform that I use, but for a new user it can be fairly confusing, as it doesn\u2019t work like other social media sites.<\/p>\n<p>A new member of my Mastodon instance, <a href=\"https:\/\/fosstodon.org\">Fosstodon<\/a>, wrote their first post stating that they\u2019re not really sure how it all works on Mastodon. Being the dutiful admin that I am, I pinged them back to let them know that I would find decent guide an post a link. To my surprise, I couldn\u2019t find a decent guide anywhere, so I decided to write one.<\/p>\n<p>I\u2019m going to try and cover all of the basics of Mastodon in this post, as well as the details of how it all works. By the end of this post, you will hopefully be a Mastodon expert. So, settle in and get a coffee, as this is going to be a long one I think.<\/p>\n<h2>Toot Toot<\/h2>\n<p>Let\u2019s start with the basics; Mastodon basically works like Twitter, but with a few differences:<\/p>\n<ul><li>A \u201cTweet\u201d on Mastodon is called a \u201cToot\u201d and you have a 500 character limit by default.<\/li>\n<li>You can set the privacy of a Toot to be public (default), unlisted, followers-only, or direct.<\/li>\n<li>You can also @mention other users, as well as add media, links and hashtags to your Toots.<\/li>\n<\/ul><p><img class=\"aligncenter size-full wp-image-389\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/toot-privacy.png\" alt=\"\" width=\"790\" height=\"450\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/toot-privacy.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/toot-privacy-300x171.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/toot-privacy-768x437.png 768w\" \/><\/p>\n<h3>Reply<\/h3>\n<p>You can, of course, reply to any Toot that you can see \u2013 after all what\u2019s the point in having a social network if you can\u2019t have a conversation.<\/p>\n<h3>Boost<\/h3>\n<p>If you come across a Toot that you like, you can <em>Boost <\/em>that Toot. This basically re-shares that Toot to your followers. However, unlike Twitter, you cannot add your own commentary to a Boosted Toot. This is by design so that a person\u2019s Toot is only Boosted in a way that gets the message they intended across.<\/p>\n<p>It\u2019s basically a way of preventing people being trolled via Boosting. E.g. you can\u2019t add \u201c<em>This Kev Quirk guy is a complete moron<\/em>\u201d to one of my Toots that you have Boosted (although many would probably like to).<\/p>\n<h3>Favourite<\/h3>\n<p>You can also Favourite a Toot. Which basically means that you support or agree with a Toot. It\u2019s the same as a Facebook like, or a Twitter heart.<\/p>\n<p>All Toots have a set of icons below them that allow you reply, boost or favourite. The icons look like this:<\/p>\n<p><img class=\"aligncenter size-full wp-image-408\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/mastodon-toot-icons.png\" alt=\"\" width=\"790\" height=\"250\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-toot-icons.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-toot-icons-300x95.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-toot-icons-768x243.png 768w\" \/><\/p>\n<h2>The Fediverse<\/h2>\n<p>If you read about Mastodon, you will often hear about <em>The Fediverse, <\/em>or <em>Federation <\/em>in general. This is a completely unique concept to Mastodon when compared to more mainstream social networks, like Twitter and Facebook. So many new users are likely to find this confusing. Let\u2019s clear things up, shall we?<\/p>\n<p>When you sign in to Twitter, you have a single timeline that is made up of the Tweets from people you follow. Mastodon is different, as it has three timelines \u2013 <em>Home<\/em>, <em>Local<\/em> and <em>Federated<\/em>.<\/p>\n<p>Each of these timelines has a different function that I will explain later on in this post, but for now let\u2019s look at how the Mastodon network works; the different timelines should then make more sense. Here\u2019s a quick video that introduces the concept of Mastodon\u2019s federation:<\/p>\n<div class=\"youtube_embed_wrapper alia_embed_wrapper\"><iframe width=\"880\" height=\"495\" src=\"https:\/\/www.youtube.com\/embed\/IPSbNdBmWKE?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe><\/div>\n<p>To explain Mastodon federation a little better, I\u2019m going to use email as an analogy through this explanation. Hopefully this will make things easier to digest.<\/p>\n<p>The Mastodon network is made up of individual servers, called <em>Instances<\/em>. If we use our email analogy; think of Mastodon as email as a whole. So if Mastodon is \u201cemail\u201d, then an Instance would be an email provider. For example, Gmail, Fastmail, or Office 365. They\u2019re all completely different servers that are run by completely different companies or people, but Gmail, Fastmail and Office 365 can all send email to one another.<\/p>\n<p>Mastodon is the same. My instance, Fosstodon, is run my myself and my friend Mike. Yet we can Toot with thousands of other Mastodon servers around the world that we do not run.<\/p>\n<p>If you want to Tweet someone on Twitter, you @mention them. It\u2019s the same on Mastodon. If I want to Toot to my friend Mike, I can simply add his handle (@mike) to my Toot. But what if I want to Toot someone else on another Mastodon instance? Well, I would use @handle@instance.name to mention them.<\/p>\n<p>For example, I follow the great <a href=\"https:\/\/www.youtube.com\/channel\/UCAPR27YUyxmgwm3Wc2WSHLw\">Chris Were<\/a> on Mastodon, but he\u2019s on another instance called <em>LinuxRocks.online<\/em>. So if I want to mention him in a Toot, I would add <em>@chriswere@linuxrocks.online <\/em>to my Toot. If Chris wanted to mention me, he would add <em>@kev@fosstodon.org <\/em>to his Toot.<\/p>\n<p><img class=\"aligncenter size-full wp-image-396\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/chriswere-toot.png\" alt=\"\" width=\"790\" height=\"280\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/chriswere-toot.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/chriswere-toot-300x106.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/chriswere-toot-768x272.png 768w\" \/><\/p>\n<p>Don\u2019t worry about having to remember people\u2019s handles and Instance names though, as Mastodon has an auto-complete feature that helps to populate a person\u2019s handle\/Instance once you start typing an @mention.<\/p>\n<p><img class=\"aligncenter size-full wp-image-424\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/mastodon-auto-complete.png\" alt=\"\" width=\"790\" height=\"380\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-auto-complete.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-auto-complete-300x144.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-auto-complete-768x369.png 768w\" \/><\/p>\n<h2>How Do Instances Connect?<\/h2>\n<p>If someone decides to start a new instance, how does the rest of the <em>Fediverse <\/em>know that they\u2019re there and to start communicating with that them? Well, this is where the community and <em>federation<\/em> comes in.<\/p>\n<p>There are thousands of Instances all over the world, and all Instances are connected by their users. So let\u2019s go back to our email analogy. If I\u2019m on Fastmail, I may not be aware of Gmail and all the users that are using Gmail. However, if I email someone with a Gmail email address, my Fastmail server then becomes aware of Gmail, so I then know where Gmail is and how to communicate with their servers.<\/p>\n<p>It\u2019s the same in Mastodon; Instance A may not be aware of Instance B, but if a user on Instance A follows a user on Instance B, Instance A then knows that Instance B exists and they will start communicating with one another. Furthermore, instance A may not have known that Instance C existed, but B did. So via Instance B, Instance A is now aware of Instance C too!<\/p>\n<p><img class=\"aligncenter size-full wp-image-400\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/instance-a-b.png\" alt=\"\" width=\"790\" height=\"250\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-a-b.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-a-b-300x95.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-a-b-768x243.png 768w\" \/><\/p>\n<p><img class=\"aligncenter size-full wp-image-401\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/instance-b-c.png\" alt=\"\" width=\"790\" height=\"400\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-b-c.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-b-c-300x152.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-b-c-768x389.png 768w\" \/><\/p>\n<p><img class=\"aligncenter size-full wp-image-402\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/instance-a-b-c.png\" alt=\"\" width=\"790\" height=\"400\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-a-b-c.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-a-b-c-300x152.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/instance-a-b-c-768x389.png 768w\" \/><\/p>\n<p>This process goes on and on, which causes a snowball effect of exponential growth of the <em>Fediverse<\/em>.<\/p>\n<p><img class=\"aligncenter size-full wp-image-398\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/fediverse-numbers.png\" alt=\"\" width=\"790\" height=\"205\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/fediverse-numbers.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/fediverse-numbers-300x78.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/fediverse-numbers-768x199.png 768w\" \/><\/p>\n<h2>The Timelines<\/h2>\n<p>Ok, now we know how the <em>Fediverse <\/em>works, lets go back to our three timelines and take a look at what they do:<\/p>\n<ul><li><strong>The Home Timeline<\/strong> \u2013 This is a simple one. It\u2019s basically the same as your Twitter timeline \u2013 it\u2019s the Toots from all the people you follow from across the <em>Fediverse.<\/em><\/li>\n<li><strong>The Local Timeline <\/strong>\u2013 This is all of the Toots from your Instance only. So whether you follow someone or not, you will see <em>all <\/em>of the activity that\u2019s going on within your Instance. This is a great way of finding new people to follow locally.<\/li>\n<li><strong>The Federated Timeline <\/strong>\u2013 This is a timeline of all Toots from all Instances across the <em>Fediverse<\/em>. Again, this is a great way of finding new people to follow as it literally contains Toots from thousands of people, but be warned \u2013 the Federated timeline can contain posts that some users may find offensive.<\/li>\n<\/ul><h2>Moderation Tools<\/h2>\n<p>If you do find something you don\u2019t want to see in Mastodon, there are a number of ways to moderate what you see. You can even report Toots to the staff of an Instance; they can then review the report and take appropriate action.<\/p>\n<p>You can see all moderation options using the three dots icon below any Toot. Within the menu, you have a number of moderation options:<\/p>\n<ul><li><strong>Mute <\/strong>\u2013 This will stop you seeing any of the Toots from that particular account.<\/li>\n<li><strong>Block <\/strong>\u2013 This will prevent that account from being able to interact with you. It will also Mute their Toots.<\/li>\n<li><strong>Report <\/strong>\u2013 This will send a report to the staff of an Instance. They should then review the report and take appropriate action. This can be no action, a warning to the user, or banning them from the Instance.<\/li>\n<\/ul><p><img class=\"aligncenter size-full wp-image-412\" src=\"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/mastodon-moderation.png\" alt=\"\" width=\"790\" height=\"445\" srcset=\"https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-moderation.png 790w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-moderation-300x169.png 300w, https:\/\/kevq.uk\/wp\/wp-content\/uploads\/2018\/08\/mastodon-moderation-768x433.png 768w\" \/><\/p>\n<p><em>Don\u2019t worry Alex, I\u2019m not going to block you! \ud83d\ude42<\/em><\/p>\n<h2>Why Mastodon?<\/h2>\n<p>So you now know the basics of how Mastodon works, and how to use it. But why would you want to use it over something like Twitter or Facebook, and why do I use it? Here are <strong><em>some <\/em><\/strong>of the reasons why\u2026<\/p>\n<h3>It\u2019s Open Source<\/h3>\n<p>It\u2019s open source so what, right? Well, no actually. Open source means that anyone can see the code that makes up Mastodon. You can\u2019t do this on many other social networks, so you have no idea what the site is doing in the background.<\/p>\n<p>Of course, many of Mastodon\u2019s users aren\u2019t software developers so they won\u2019t know what Mastodon\u2019s source code does (myself included), but there are hundreds, or even thousands of developers out there that have seen the code for Mastodon. If it was doing anything nefarious, trust me, we would know.<\/p>\n<p>Being open source gives us the reassurance that our data is not being harvested, or that we\u2019re not being spied on. Plus, if the main developer of Mastodon, Gargron, decides to hang up his keyboard, someone else can take his place and fork the project\u2019s source code.<\/p>\n<h3>Chronological Timeline<\/h3>\n<p>This may seem like a small thing, but for me it\u2019s one of the most important features of Mastodon. You see, Facebook and Twitter have \u201cclever\u201d algorithms that are designed to change the content\u2019s order on every refresh. This is designed to give the perception of new content and keep you on the site longer.<\/p>\n<p>That isn\u2019t the case with Mastodon. On Mastodon all three of the timelines are <strong>always <\/strong>chronologically displayed. So you can easily see what people have Tooted from when you were last online. This means it isn\u2019t specifically designed to take over your life (although it does, because it\u2019s awesome!) It just allows you to see what has been going on since you were last online, then go about the rest of your day.<\/p>\n<h3>No Adverts Or Tracking<\/h3>\n<p>That\u2019s right, NO ADVERTS! No sponsored posts, no <em>\u201cwe think you might like this\u201d <\/em>and no tracking! Mastodon is designed to bring people together, not make money. It\u2019s that simple.<\/p>\n<p>Having said that, most Instances do have Patreon pages, as running a Mastodon instance can get expensive. So if you use Mastodon, and enjoy it, I would urge you to contribute whatever you can to your Instance of choice. In my experience, the money is often used to improve the project.<\/p>\n<p>On Fosstodon, we use any donations over and above our ongoing fees to make donations to various open source projects. Which projects we donate to is decided by those who donate.<\/p>\n<h3>No Single Owner<\/h3>\n<p>Because Mastodon uses a collection of Instances, you\u2019re not at the beck and call of one site owner. If you don\u2019t like the direction an Instance is taking, you can pack your virtual bags and go. Mastodon even has an import\/export tool that allows you to migrate the people you follow from one account\/instance to another.<\/p>\n<h3>Interest Specific, But Not Really<\/h3>\n<p>The beauty of having different Instances is that many of them are underpinned specific to one interest. For example my instance, Fosstodon, is specific to Free and Open Source Software (FOSS), hence the name. However, we encourage the users of Fosstodon to talk about all sorts of interesting things, not just open source software.<\/p>\n<p>So although all the members of our Instance have a common interest (FOSS), there are lots of topics that get discussed regularly. There are also many generalist Instances that you can join if you wish. These are more like Twitter in that they don\u2019t have a specific interest underpinning them.<\/p>\n<p>Remember, you can also follow people who have different interests on other Instances, so your home feed can be extremely diverse \u2013 I know mine is!<\/p>\n<h2>Finding An Instance<\/h2>\n<p>There are thousands of Mastodon Instances across the <em>Fediverse<\/em>, so how do you find the right one? Well, finding the right one is very difficult, so here are some recommendations:<\/p>\n<ul><li><a href=\"https:\/\/fosstodon.org\">Fosstodon<\/a>\u2013 Selfish plug. Linux\/FOSS focused Instance.<\/li>\n<li><a href=\"https:\/\/linuxrocks.online\">Linux Rocks<\/a> \u2013 Another great Linux\/FOSS focused Instance.<\/li>\n<li><a href=\"https:\/\/floss.social\">Floss Social<\/a> \u2013 A third Linux\/FOSS Instance (is it obvious what I\u2019m interested in?)<\/li>\n<li><a href=\"https:\/\/mastodon.social\">Mastodon Social<\/a> \u2013 The original Mastodon Instance. It\u2019s a general instance that is run by Mastodon\u2019s creator.<\/li>\n<li><a href=\"https:\/\/mastodon.technology\">Mastodon Technology<\/a> \u2013 A generalist Instance for anyone interested in technology.<\/li>\n<\/ul><p>If none of the instances above work for you, you can take a look at <a href=\"https:\/\/instances.social\">Instances.Social<\/a> which is basically a search engine for Mastodon Instances. However, I\u2019ve always found the site awkward to use as it doesn\u2019t have many filtering options.<\/p>\n<h2>Mobile Apps<\/h2>\n<p>There are a number of mobile apps available for Mastodon. I have tried many of them for both Android and Apple iOS. My recommendations for the best Mastodon apps would be <a href=\"https:\/\/play.google.com\/store\/apps\/details?id=com.keylesspalace.tusky&amp;hl=nl\">Tusky<\/a> for Android and <a href=\"https:\/\/itunes.apple.com\/us\/app\/tootdon-for-mastodon\/id1282283934\">Tootdon<\/a> for iOS.<\/p>\n<p>However, if you want to check out all the other apps that are available, including desktop apps, <a href=\"https:\/\/github.com\/tootsuite\/documentation\/blob\/master\/Using-Mastodon\/Apps.md\">take a look here<\/a>.<\/p>\n<h2>Final Thoughts<\/h2>\n<p>Mastodon is awesome. It\u2019s full of friendly, interesting people. There are very few trolls that I have come across, no tracking and no adverts. It\u2019s just a great place to be.<\/p>\n<p>Hopefully this guide will help get you started with Mastodon, but if you have any other questions that this guide doesn\u2019t cover, feel free to leave a comment below and I\u2019ll do my best to answer.<\/p>\n<p>Finally, if you\u2019re already on Mastodon and want to follow me, you can do so <a href=\"https:\/\/fosstodon.org\/@kev\">here<\/a>.<\/p>","created_at":"2018-11-07T16:22:41+0000","updated_at":"2018-11-07T16:22:41+0000","published_at":"2018-08-04T09:25:24+0000","annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":11,"domain_name":"kevq.uk","preview_picture":"https:\/\/kevq.uk\/wp-content\/uploads\/2018\/08\/mastodon-header.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":96,"title":"SDF Public Access UNIX System","url":"https:\/\/sdf.org\/?join=","content":"<b>Linux\/UNIX<\/b> users can type 'ssh new@sdf.org' at their shell prompts.\n<p>For Microsoft Windows we highly recommend the free SSH client\n<a target=\"new\" href=\"http:\/\/www.chiark.greenend.org.uk\/~sgtatham\/putty\/download.html\">\nputty.exe<\/a>.<br \/><\/p><p>If you have any questions or cannot figure out how to use <b>SSH<\/b>, live help<br \/>is available on IRC via <b>irc.sdf.org<\/b> in the <b>#helpdesk<\/b> channel.<\/p><form action=\"https:\/\/www.paypal.com\/cgi-bin\/webscr\" method=\"post\">\n<div><em>Please be sure fill in the description<br \/>with your login and membership option.<\/em>\n<\/div><\/form>","created_at":"2018-11-07T16:22:40+0000","updated_at":"2018-11-07T16:22:40+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"sdf.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":97,"title":"Blog - Daniel Miessler","url":"https:\/\/danielmiessler.com\/blog\/","content":"<div>\n<p class=\"entry-meta\">Created: September 4, 2018 <time datetime=\"2018-09-04T15:40:27+00:00\" itemprop=\"datePublished\"><\/time><\/p><\/div><div class=\"entry-content\" itemprop=\"text\"><a class=\"entry-image-link\" href=\"https:\/\/danielmiessler.com\/blog\/unsupervised-learning-no-141\/\" aria-hidden=\"true\"><img width=\"150\" height=\"150\" src=\"https:\/\/danielmiessler.com\/images\/ul-icon-150x150.png\" class=\"post-image entry-image\" alt=\"Unsupervised Learning: No. 141\" itemprop=\"image\" srcset=\"https:\/\/danielmiessler.com\/images\/ul-icon-150x150.png 150w, https:\/\/danielmiessler.com\/images\/ul-icon-100x100.png 100w\" \/><\/a><p>Subscribe here to get this in your inbox every week. Security News \ud83d\ude2e\u00a0The governments of the \u201cFive Eyes\u201d (US, Australia, UK, Canada, and New Zealand) have asked the world\u2019s largest tech companies to build encryption backdoors into their systems. But more than just asking, they also said, \u201dShould governments continue to encounter impediments to lawful access to information\u2026<\/p><\/div>","created_at":"2018-11-07T16:22:40+0000","updated_at":"2018-11-07T16:22:40+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"danielmiessler.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":98,"title":"ActivityPub","url":"https:\/\/www.w3.org\/TR\/activitypub\/","content":"ActivityPub\n    <section id=\"Overview\"><h2 id=\"x1-overview\">1. Overview<\/h2><p>ActivityPub provides two layers:<\/p><ul><li>\n          <b>A server to server federation protocol<\/b>\n          (so decentralized websites can share information)\n        <\/li>\n        <li>\n          <b>A client to server protocol<\/b>\n          (so users, including real-world users, bots, and other automated processes,\n          can communicate with ActivityPub using their accounts on servers,\n          from a phone or desktop or web application or whatever)\n        <\/li>\n      <\/ul><p>\n        ActivityPub implementations can implement just one of these things or\n        both of them.\n        However, once you've implemented one, it isn't too many steps to\n        implement the other, and there are a lot of benefits to both (making\n        your website part of the decentralized social web, and being able to\n        use clients and client libraries that work across a wide variety of\n        social websites).\n      <\/p><p>\n        In ActivityPub, a user is represented by \"<a href=\"#actors\">actors<\/a>\"\n        via the user's accounts on servers.\n        User's accounts on different servers correspond to different actors.\n        Every Actor has:\n      <\/p><ul><li><b>An <code>inbox<\/code>:<\/b>  How they get messages from the world<\/li>\n        <li><b>An <code>outbox<\/code>:<\/b> How they send messages to others<\/li>\n      <\/ul><p>\n        <img src=\"https:\/\/www.w3.org\/TR\/activitypub\/illustration\/tutorial-1.png\" alt=\"Actor with inbox and outbox\" \/><\/p><p>\n        These are endpoints, or really, just URLs which are listed in the\n        ActivityPub actor's ActivityStreams description.\n        (More on ActivityStreams later).\n      <\/p><p>\n        Here's an example of the record of our friend Alyssa P. Hacker:\n      <\/p><div class=\"example\"><p>Example 1<\/p><pre class=\"highlight json hljs\" aria-busy=\"false\" aria-live=\"polite\">{\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n \"type\": \"Person\",\n \"id\": \"https:\/\/social.example\/alyssa\/\",\n \"name\": \"Alyssa P. Hacker\",\n \"preferredUsername\": \"alyssa\",\n \"summary\": \"Lisp enthusiast hailing from MIT\",\n \"inbox\": \"https:\/\/social.example\/alyssa\/inbox\/\",\n \"outbox\": \"https:\/\/social.example\/alyssa\/outbox\/\",\n \"followers\": \"https:\/\/social.example\/alyssa\/followers\/\",\n \"following\": \"https:\/\/social.example\/alyssa\/following\/\",\n \"liked\": \"https:\/\/social.example\/alyssa\/liked\/\"}<\/pre><\/div><p>\n        ActivityPub uses [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] for its vocabulary.\n        This is pretty great because ActivityStreams includes all the common\n        terms you need to represent all the activities and content flowing\n        around a social network.\n        It's likely that ActivityStreams already includes all the vocabulary\n        you need, but even if it doesn't, ActivityStreams can be extended\n        via [<cite><a class=\"bibref\" href=\"#bib-JSON-LD\">JSON-LD<\/a><\/cite>].\n        If you know what JSON-LD is, you can take advantage of the cool linked\n        data approaches provided by JSON-LD.\n        If you don't, don't worry, JSON-LD documents and ActivityStreams can be\n        understood as plain old simple JSON.\n        (If you're going to add extensions, that's the point at which JSON-LD\n        really helps you out).\n      <\/p><p>\n        So, okay.\n        Alyssa wants to talk to her friends, and her friends want to talk to\n        her!\n        Luckily these \"inbox\" and \"outbox\" things can help us out.\n        They both behave differently for GET and POST.\n        Let's see how that works:\n      <\/p><p>\n        <img src=\"https:\/\/www.w3.org\/TR\/activitypub\/illustration\/tutorial-2.png\" alt=\"Actor with messages flowing from rest of world to inbox and from outbox to rest of world\" \/><\/p><p>\n        Hey nice, so just as a recap:\n      <\/p><ul><li>\n          You can POST to someone's inbox to send them a message\n          (server-to-server \/ federation only... this <em>is<\/em> federation!)\n        <\/li>\n        <li>\n          You can GET from your inbox to read your latest messages\n          (client-to-server; this is like reading your social\n          network stream)\n        <\/li>\n        <li>\n          You can POST to your outbox to send messages to the world\n          (client-to-server)\n        <\/li>\n        <li>\n          You can GET from someone's outbox to see what messages they've\n          posted (or at least the ones you're authorized to see).\n          (client-to-server and\/or server-to-server)\n        <\/li>\n      <\/ul><p>\n        Of course, if that last one (GET'ing from someone's outbox) was the\n        only way to see what people have sent, this wouldn't be a very\n        efficient federation protocol!\n        Indeed, federation happens usually by servers posting messages sent by\n        actors to actors on other servers' inboxes.\n      <\/p><p>\n        Let's see an example!\n        Let's say Alyssa wants to catch up with her friend, Ben Bitdiddle.\n        She lent him a book recently and she wants to make sure he returns it\n        to her.\n        Here's the message she composes, as an ActivityStreams object:\n      <\/p><div class=\"example\"><p>Example 2<\/p><pre class=\"highlight json hljs\" aria-busy=\"false\" aria-live=\"polite\">{\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n \"type\": \"Note\",\n \"to\": [\"https:\/\/chatty.example\/ben\/\"],\n \"attributedTo\": \"https:\/\/social.example\/alyssa\/\",\n \"content\": \"Say, did you finish reading that book I lent you?\"}<\/pre><\/div><p>\n        This is a note addressed to Ben.\n        She POSTs it to her outbox.\n      <\/p><p>\n        <img src=\"https:\/\/www.w3.org\/TR\/activitypub\/illustration\/tutorial-3.png\" alt=\"Actor posting message to outbox\" \/><\/p><p>\n        Since this is a non-activity object, the server recognizes that this is\n        an object being newly created, and does the courtesy of wrapping it in\n        a Create activity.\n        (Activities sent around in ActivityPub generally follow the pattern of\n        some activity by some actor being taken on some object.\n        In this case the activity is a Create of a Note object, posted by a\n        Person).\n      <\/p><div class=\"example\"><p>Example 3<\/p><pre class=\"highlight json hljs\" aria-busy=\"false\" aria-live=\"polite\">{\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n \"type\": \"Create\",\n \"id\": \"https:\/\/social.example\/alyssa\/posts\/a29a6843-9feb-4c74-a7f7-081b9c9201d3\",\n \"to\": [\"https:\/\/chatty.example\/ben\/\"],\n \"actor\": \"https:\/\/social.example\/alyssa\/\",\n \"object\": {\"type\": \"Note\",\n            \"id\": \"https:\/\/social.example\/alyssa\/posts\/49e2d03d-b53a-4c4c-a95c-94a6abf45a19\",\n            \"attributedTo\": \"https:\/\/social.example\/alyssa\/\",\n            \"to\": [\"https:\/\/chatty.example\/ben\/\"],\n            \"content\": \"Say, did you finish reading that book I lent you?\"}}<\/pre><\/div><p>\n        Alyssa's server looks up Ben's ActivityStreams actor object, finds his\n        inbox endpoint, and POSTs her object to his inbox.\n      <\/p><p>\n        <img src=\"https:\/\/www.w3.org\/TR\/activitypub\/illustration\/tutorial-4.png\" alt=\"Server posting to remote actor's inbox\" \/><\/p><p>\n        Technically these are two separate steps... one is client to server\n        communication, and one is server to server communication (federation).\n        But, since we're using them both in this example, we can abstractly\n        think of this as being a streamlined submission from outbox to inbox:\n      <\/p><p>\n        <img src=\"https:\/\/www.w3.org\/TR\/activitypub\/illustration\/tutorial-5.png\" alt=\"Note flowing from one actor's outbox to other actor's inbox\" \/><\/p><p>\n        Cool!\n        A while later, Alyssa checks what new messages she's gotten.\n        Her phone polls her inbox via GET, and amongst a bunch of cat videos\n        posted by friends and photos of her nephew posted by her sister, she\n        sees the following:\n      <\/p><div class=\"example\"><p>Example 4<\/p><pre class=\"highlight json hljs\" aria-busy=\"false\" aria-live=\"polite\">{\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n \"type\": \"Create\",\n \"id\": \"https:\/\/chatty.example\/ben\/p\/51086\",\n \"to\": [\"https:\/\/social.example\/alyssa\/\"],\n \"actor\": \"https:\/\/chatty.example\/ben\/\",\n \"object\": {\"type\": \"Note\",\n            \"id\": \"https:\/\/chatty.example\/ben\/p\/51085\",\n            \"attributedTo\": \"https:\/\/chatty.example\/ben\/\",\n            \"to\": [\"https:\/\/social.example\/alyssa\/\"],\n            \"inReplyTo\": \"https:\/\/social.example\/alyssa\/posts\/49e2d03d-b53a-4c4c-a95c-94a6abf45a19\",\n            \"content\": \"&lt;p&gt;Argh, yeah, sorry, I'll get it back to you tomorrow.&lt;\/p&gt;\n                        &lt;p&gt;I was reviewing the section on register machines,\n                           since it's been a while since I wrote one.&lt;\/p&gt;\"}}<\/pre><\/div><p>Alyssa is relieved, and likes Ben's post:<\/p><div class=\"example\"><p>Example 5<\/p><pre class=\"highlight json hljs\" aria-busy=\"false\" aria-live=\"polite\">{\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n \"type\": \"Like\",\n \"id\": \"https:\/\/social.example\/alyssa\/posts\/5312e10e-5110-42e5-a09b-934882b3ecec\",\n \"to\": [\"https:\/\/chatty.example\/ben\/\"],\n \"actor\": \"https:\/\/social.example\/alyssa\/\",\n \"object\": \"https:\/\/chatty.example\/ben\/p\/51086\"}<\/pre><\/div><p>\n        She POSTs this message to her outbox.\n        (Since it's an activity, her server knows it doesn't need to wrap it in\n        a Create object).\n      <\/p><p>\n        Feeling happy about things, she decides to post a public message to her\n        followers.\n        Soon the following message is blasted to all the members of her\n        followers collection, and since it has the special Public group\n        addressed, is generally readable by anyone.\n      <\/p><div class=\"example\"><p>Example 6<\/p><pre class=\"highlight json hljs\" aria-busy=\"false\" aria-live=\"polite\">{\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n \"type\": \"Create\",\n \"id\": \"https:\/\/social.example\/alyssa\/posts\/9282e9cc-14d0-42b3-a758-d6aeca6c876b\",\n \"to\": [\"https:\/\/social.example\/alyssa\/followers\/\",\n        \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"],\n \"actor\": \"https:\/\/social.example\/alyssa\/\",\n \"object\": {\"type\": \"Note\",\n            \"id\": \"https:\/\/social.example\/alyssa\/posts\/d18c55d4-8a63-4181-9745-4e6cf7938fa1\",\n            \"attributedTo\": \"https:\/\/social.example\/alyssa\/\",\n            \"to\": [\"https:\/\/social.example\/alyssa\/followers\/\",\n                   \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"],\n            \"content\": \"Lending books to friends is nice.  Getting them back is even nicer! :)\"}}<\/pre><\/div><\/section><section id=\"conformance\"><h2 id=\"x2-conformance\">2. Conformance<\/h2><p>\n  As well as sections marked as non-normative, all authoring guidelines, diagrams, examples,\n  and notes in this specification are non-normative. Everything else in this specification is\n  normative.\n<\/p><p id=\"respecRFC2119\">The key words  <em class=\"rfc2119\">MAY<\/em>, <em class=\"rfc2119\">MUST<\/em>, <em class=\"rfc2119\">MUST NOT<\/em>, <em class=\"rfc2119\">SHOULD<\/em>, and <em class=\"rfc2119\">SHOULD NOT<\/em> are \n  to be interpreted as described in [<cite><a class=\"bibref\" href=\"#bib-RFC2119\">RFC2119<\/a><\/cite>].\n<\/p><section id=\"specification-profiles\"><h3 id=\"x2-1-specification-profiles\">2.1 Specification Profiles<\/h3><p>\n          This specification defines two closely related and interacting\n          protocols:\n        <\/p><dl><dt>A client to server protocol, or \"Social API\"<\/dt>\n          <dd>\n            This protocol permits a client to act <i>on behalf<\/i> of a user.\n            For example, this protocol is used by a mobile phone application to\n            interact with a social stream of the user's actor.\n          <\/dd>\n          <dt>A server to server protocol, or \"Federation Protocol\"<\/dt>\n          <dd>\n            This protocol is used to distribute activities between actors on\n            different servers, tying them into the same social graph.\n          <\/dd>\n        <\/dl><p>\n          The ActivityPub specification is designed so that once either of\n          these protocols are implemented, supporting the other is of very\n          little additional effort.\n          However, servers may still implement one without the other.\n          This gives three conformance classes:\n        <\/p><dl><dt>ActivityPub conformant Client<\/dt>\n          <dd>\n            This designation applies to any implementation of the entirety of the\n            client portion of the client to server protocol.\n          <\/dd>\n          <dt>ActivityPub conformant Server<\/dt>\n          <dd>\n            This designation applies to any implementation of the entirety of the\n            server portion of the client to server protocol.\n          <\/dd>\n          <dt>ActivityPub conformant Federated Server<\/dt>\n          <dd>\n            This designation applies to any implementation of the entirety of\n            the federation protocols.\n          <\/dd>\n        <\/dl><p>\n          It is called out whenever a portion of the specification only applies\n          to implementation of the federation protocol.\n          In addition, whenever requirements are specified, it is called out\n          whether they apply to the client or server (for the client-to-server\n          protocol) or whether referring to a sending or receiving server in\n          the server-to-server protocol.\n        <\/p><\/section><\/section><section id=\"obj\"><h2 id=\"x3-objects\">3. Objects<\/h2><p>\n        Objects are the core concept around which both [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] and\n        ActivityPub are built.\n        Objects are often wrapped in Activities and are contained in streams of\n        Collections, which are themselves subclasses of Objects.\n        See the [<cite><a class=\"bibref\" href=\"#bib-Activity-Vocabulary\">Activity-Vocabulary<\/a><\/cite>] document, particularly the\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#types\">Core Classes<\/a>;\n        ActivityPub follows the mapping of this vocabulary very closely.\n      <\/p><p>\n        ActivityPub defines some terms in addition to those provided by\n        ActivityStreams.\n        These terms are provided in the ActivityPub\n        <a href=\"http:\/\/www.w3.org\/TR\/json-ld\/#the-context\">JSON-LD context<\/a>\n        at\n        <code>https:\/\/www.w3.org\/ns\/activitystreams<\/code>.\n        Implementers <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> include the ActivityPub context in their\n        object definitions.\n        Implementers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> include additional context as appropriate.\n      <\/p><p>\n        ActivityPub shares the same\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/#urls\">\n          URI \/ IRI conventions as in ActivityStreams<\/a>.\n      <\/p><p>\n        Servers <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> validate the content they receive to avoid content\n        spoofing attacks.\n        (A server should do something at least as robust as checking that\n        the object appears as received at its origin, but mechanisms\n        such as checking signatures would be better if available).\n        No particular mechanism for verification is authoritatively specified by\n        this document, but please see <a href=\"#security-considerations\">Security \n        Considerations<\/a> for some suggestions and good practices.\n      <\/p><section id=\"obj-id\"><h3 id=\"x3-1-object-identifiers\">3.1 Object Identifiers<\/h3><p>\n          All Objects in [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] should have unique global\n          identifiers.\n          ActivityPub extends this requirement; all objects distributed by the\n          ActivityPub protocol <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> have unique global identifiers, unless they\n          are intentionally transient (short lived activities that are not\n          intended to be able to be looked up, such as some kinds of chat\n          messages or game notifications).\n          These identifiers must fall into one of the following groups:\n        <\/p><ol><li>\n            Publicly dereferencable URIs, such as HTTPS URIs, with their\n            authority belonging to that of their originating server.\n            (Publicly facing content <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> use HTTPS URIs).\n          <\/li>\n          <li>\n            An ID explicitly specified as the JSON <code>null<\/code> object,\n            which implies an anonymous object (a part of its parent context)\n          <\/li>\n        <\/ol><p>\n          Identifiers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be provided for activities posted in server to\n          server communication, unless the activity is intentionally transient.\n          However, for client to server communication, a server receiving an\n          object posted to the outbox with no specified <code>id<\/code> <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          allocate an object ID in the actor's namespace and attach it to the\n          posted object.\n        <\/p><p>All objects have the following properties:<\/p><dl><dt>id<\/dt>\n          <dd>\n            The object's unique global identifier (unless the object is transient,\n            in which case the <code>id<\/code> <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be omitted).\n          <\/dd>\n          <dt>type<\/dt>\n          <dd>\n            The type of the object.\n          <\/dd>\n        <\/dl><\/section><section id=\"retrieving-objects\"><h3 id=\"x3-2-retrieving-objects\">3.2 Retrieving objects<\/h3><p>\n          The HTTP GET method may be dereferenced against an object's\n          <code>id<\/code> property to retrieve the activity.\n          Servers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> use HTTP content negotiation as defined in [<cite><a class=\"bibref\" href=\"#bib-RFC7231\">RFC7231<\/a><\/cite>] to\n          select the type of data to return in response to a request,\n          but <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> present the ActivityStreams object representation\n          in response to\n          <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>,\n          and <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> also present the ActivityStreams representation in\n          response to <code>application\/activity+json<\/code> as well.\n          The client <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> specify an <code>Accept<\/code> header with the\n          <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>\n          media type in order to retrieve the activity.\n        <\/p><p>\n          Servers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> implement other behavior for requests which do not comply\n          with the above requirement.\n          (For example, servers may implement additional legacy protocols, or\n          may use the same URI for both HTML and ActivityStreams\n          representations of a resource).\n        <\/p><p>\n          Servers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> require authorization as specified in\n          <a href=\"#authorization\" class=\"sec-ref\">B.1 Authentication and Authorization<\/a>, and may additionally implement their\n          own authorization rules.\n          Servers <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> fail requests which do not pass their authorization\n          checks with the appropriate HTTP error code, or the 403 Forbidden\n          error code where the existence of the object is considered private.\n          An origin server which does not wish to disclose the existence of\n          a private target <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> instead respond with a status code of\n          404 Not Found.\n        <\/p><\/section><section id=\"source-property\"><h3 id=\"x3-3-the-source-property\">3.3 The source property<\/h3><p>\n          In addition to all the properties defined by the\n\t  [<cite><a class=\"bibref\" href=\"#bib-Activity-Vocabulary\">Activity-Vocabulary<\/a><\/cite>], ActivityPub extends the <code>Object<\/code> by\n\t  supplying the <code>source<\/code> property.\n          The <code>source<\/code> property is intended to convey some\n          sort of source from which the <code>content<\/code> markup\n          was derived, as a form of provenance, or to support future\n          editing by clients.\n          In general, clients do the conversion from <code>source<\/code>\n          to <code>content<\/code>, not the other way around.\n        <\/p><p>\n          The value of <code>source<\/code> is itself an object\n          which uses its own <code>content<\/code> and <code>mediaType<\/code>\n          fields to supply source information.\n        <\/p><div class=\"example\"><p>Example 8<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": [\"https:\/\/www.w3.org\/ns\/activitystreams\",\n               {\"@language\": \"en\"}],\n  \"type\": \"Note\",\n  \"id\": \"http:\/\/postparty.example\/p\/2415\",\n  \"content\": \"&lt;p&gt;I &lt;em&gt;really&lt;\/em&gt; like strawberries!&lt;\/p&gt;\",\n  \"source\": {\n    \"content\": \"I *really* like strawberries!\",\n    \"mediaType\": \"text\/markdown\"}\n}<\/pre><\/div><div class=\"note\"><p>Note: What to do when clients can't meaningfully handle a mediaType?<\/p><div class=\"\"><p>\n            In general, it's best to let a user edit their original post\n            in the same source format they originally composed it in.\n            But not all clients can reliably provide a nice interface for\n            all source types, and since clients are expected to do the\n            conversion from <code>source<\/code> to <code>content<\/code>,\n            some clients may work with a media type that another client\n            does not know how to work with.\n            While a client could feasibly provide the <code>content<\/code>\n            markup to be edited and ignore the source, this means that the\n            user will lose the more desirable form of the original\n            <code>source<\/code> in any future revisions.\n            A client doing so should thus provide a minimally obtrusive warning\n            cautioning that the original source format is not understood and is\n            thus being ignored.\n          <\/p><p>\n            For example, Alyssa P. Hacker likes to post to her ActivityPub\n            powered blog via an Emacs client she has written, leveraging\n            <a href=\"http:\/\/orgmode.org\/\">Org mode<\/a>.\n            Later she switches to editing on her phone's client, which\n            has no idea what <code>text\/x-org<\/code> is or how to render\n            it to HTML, so it provides a text box to edit the original\n            <code>content<\/code> instead.\n            A helpful warning displays above the edit area saying,\n            \"This was originally written in another markup language we don't\n            know how to handle.  If you edit, you'll lose your original\n            source!\"\n            Alyssa decides the small typo fix isn't worth losing her nice\n            org-mode markup and decides to make the update when she gets\n            home.\n          <\/p><\/div><\/div><\/section><\/section><section id=\"actors\"><h2 id=\"x4-actors\">4. Actors<\/h2><p>\n        ActivityPub actors are generally one of the\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#actor-types\">\n          ActivityStreams Actor Types<\/a>,\n        but they don't have to be. For example, a \n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-profile\">\n          Profile<\/a> object\n        might be used as an actor, or a type from an ActivityStreams extension.\n        Actors are <a href=\"#retrieving-objects\">retrieved<\/a> like any other\n        Object in ActivityPub.\n        Like other ActivityStreams objects, actors have an <code>id<\/code>,\n        which is a URI.\n        When entered directly into a user interface (for example on a login\n        form), it is desirable to support simplified naming.\n        For this purpose, ID normalization <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> be performed as follows:\n      <\/p><ol><li>\n          If the entered ID is a valid URI, then it is to be used directly.\n        <\/li>\n        <li>\n          If it appears that the user neglected to add a scheme for a URI that\n          would otherwise be considered valid, such as\n          <code>example.org\/alice\/<\/code>, clients <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> attempt to provide\n          a default scheme, preferably <code>https<\/code>.\n        <\/li>\n        <li>\n          Otherwise, the entered value should be considered invalid.\n        <\/li>\n      <\/ol><p>\n        Once the actor's URI has been identified, it should be dereferenced.\n      <\/p><div class=\"note\"><p>Note<\/p><p>ActivityPub does not dictate a specific relationship between\n        \"users\" and Actors; many configurations are possible.\n        There may be multiple human users or organizations controlling an\n        Actor, or likewise one human or organization may control multiple\n        Actors. Similarly, an Actor may represent a piece of software,\n        like a bot, or an automated process.\n        More detailed \"user\" modelling, for example linking together of Actors which\n        are controlled by the same entity, or allowing one Actor to be presented\n        through multiple alternate profiles or aspects, are at the discretion\n        of the implementation.<\/p><\/div><section id=\"actor-objects\"><h3 id=\"x4-1-actor-objects\">4.1 <i>Actor<\/i> objects<\/h3><p>\n          Actor objects <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> have, in addition to the properties mandated by\n          <a href=\"#obj-id\" class=\"sec-ref\">3.1 Object Identifiers<\/a>, the following properties:\n        <\/p><dl><dt id=\"inbox-property\">inbox<\/dt>\n          <dd>\n            A reference to an [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>]\n            <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n              <code>OrderedCollection<\/code><\/a>\n            comprised of all the messages received by the actor; see\n            <a href=\"#inbox\" class=\"sec-ref\">5.2 Inbox<\/a>.\n          <\/dd>\n          <dt id=\"outbox-property\">outbox<\/dt>\n          <dd>\n            An [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>]\n            <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n              <code>OrderedCollection<\/code><\/a>\n            comprised of all the messages produced by the actor; see\n            <a href=\"#outbox\" class=\"sec-ref\">5.1 Outbox<\/a>.\n          <\/dd>\n        <\/dl><p>\n          Implementations <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>, in addition, provide the following\n          properties:\n        <\/p><dl><dt id=\"following-property\">following<\/dt>\n          <dd>\n            A link to an [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] collection of the actors that\n            this actor is following; see <a href=\"#following\" class=\"sec-ref\">5.4 Following Collection<\/a>\n          <\/dd>\n          <dt id=\"followers-property\">followers<\/dt>\n          <dd>\n            A link to an [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] collection of the actors that\n            follow this actor; see <a href=\"#followers\" class=\"sec-ref\">5.3 Followers Collection<\/a>.\n          <\/dd>\n        <\/dl><p>\n          Implementations <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> provide the following properties:\n        <\/p><dl><dt id=\"liked-property\">liked<\/dt>\n          <dd>\n            A link to an [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] collection of objects this\n            actor has liked; see <a href=\"#liked\" class=\"sec-ref\">5.5 Liked Collection<\/a>.\n          <\/dd>\n        <\/dl><div class=\"example\"><p>Example 9<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": [\"https:\/\/www.w3.org\/ns\/activitystreams\",\n               {\"@language\": \"ja\"}],\n  \"type\": \"Person\",\n  \"id\": \"https:\/\/kenzoishii.example.com\/\",\n  \"following\": \"https:\/\/kenzoishii.example.com\/following.json\",\n  \"followers\": \"https:\/\/kenzoishii.example.com\/followers.json\",\n  \"liked\": \"https:\/\/kenzoishii.example.com\/liked.json\",\n  \"inbox\": \"https:\/\/kenzoishii.example.com\/inbox.json\",\n  \"outbox\": \"https:\/\/kenzoishii.example.com\/feed.json\",\n  \"preferredUsername\": \"kenzoishii\",\n  \"name\": \"\u77f3\u4e95\u5065\u8535\",\n  \"summary\": \"\u3053\u306e\u65b9\u306f\u305f\u3060\u306e\u4f8b\u3067\u3059\",\n  \"icon\": [\n    \"https:\/\/kenzoishii.example.com\/image\/165987aklre4\"\n  ]\n}<\/pre><\/div><p>Implementations <em class=\"rfc2119\" title=\"MAY\">MAY<\/em>, in addition, provide the following properties:<\/p><dl><dt id=\"streams-property\">streams<\/dt>\n          <dd>\n            A list of supplementary Collections which may be of interest.\n          <\/dd>\n          <dt id=\"preferredUsername\">preferredUsername<\/dt>\n          <dd>\n            A short username which may be used to refer to the actor, with no\n            uniqueness guarantees.\n          <\/dd>\n          <dt id=\"endpoints\">endpoints<\/dt>\n          <dd>\n            A json object which maps additional (typically server\/domain-wide)\n            endpoints which may be useful either for this actor or someone\n            referencing this actor.\n            This mapping may be nested inside the actor document as the value\n            or may be a link to a JSON-LD document with these properties.\n          <\/dd>\n        <\/dl><p>\n          The <code>endpoints<\/code> mapping <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> include the following\n          properties:\n        <\/p><dl><dt id=\"proxyUrl\">proxyUrl<\/dt>\n          <dd>\n            Endpoint URI so this actor's clients may access remote\n            ActivityStreams objects which require authentication to access.\n            To use this endpoint, the client posts an\n            <code>x-www-form-urlencoded<\/code> <code>id<\/code> parameter\n            with the value being the <code>id<\/code> of the requested\n            ActivityStreams object.\n          <\/dd>\n          \n          <dd>\n            If OAuth 2.0 bearer tokens [<cite><a class=\"bibref\" href=\"#bib-RFC6749\">RFC6749<\/a><\/cite>] [<cite><a class=\"bibref\" href=\"#bib-RFC6750\">RFC6750<\/a><\/cite>] are being used\n            for authenticating\n            <a href=\"#client-to-server-interactions\">client to server\n              interactions<\/a>,\n            this endpoint specifies a URI at which a browser-authenticated user\n            may obtain a new authorization grant.\n          <\/dd>\n          <dt id=\"oauthTokenEndpoint\">oauthTokenEndpoint<\/dt>\n          <dd>\n            If OAuth 2.0 bearer tokens [<cite><a class=\"bibref\" href=\"#bib-RFC6749\">RFC6749<\/a><\/cite>] [<cite><a class=\"bibref\" href=\"#bib-RFC6750\">RFC6750<\/a><\/cite>] are being used\n            for authenticating\n            <a href=\"#client-to-server-interactions\">client to server\n              interactions<\/a>,\n            this endpoint specifies a URI at which a client may acquire an\n\t    access token.\n          <\/dd>\n          <dt id=\"provideClientKey\">provideClientKey<\/dt>\n          <dd>\n            If Linked Data Signatures and HTTP Signatures are being used for\n            authentication and authorization, this endpoint specifies a URI at\n            which browser-authenticated users may authorize a client's public\n            key for\n            <a href=\"#client-to-server-interactions\">client to server\n              interactions<\/a>.\n          <\/dd>\n          <dt id=\"signClientKey\">signClientKey<\/dt>\n          <dd>\n            If Linked Data Signatures and HTTP Signatures are being used for\n            authentication and authorization, this endpoint specifies a URI at\n            which a client key may be signed by the actor's key for a time\n            window to act on behalf of the actor in interacting with foreign\n            servers.\n          <\/dd>\n          \n          <dd>\n            An optional endpoint\n            <a href=\"#shared-inbox-delivery\">\n              used for wide delivery of publicly addressed activities\n              and activities sent to followers<\/a>.\n            <code>sharedInbox<\/code> endpoints <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> also be publicly\n            readable <code>OrderedCollection<\/code> objects containing\n            objects addressed to the <a href=\"#public-addressing\">Public<\/a>\n            special collection.\n            Reading from the <code>sharedInbox<\/code> endpoint <em class=\"rfc2119\" title=\"MUST NOT\">MUST NOT<\/em> present\n            objects which are not addressed to the <code>Public<\/code>\n            endpoint.\n          <\/dd>\n        <\/dl><div class=\"note\"><p>Note<\/p><div class=\"\" id=\"as2-actor-properties\"><p>\n            As the upstream vocabulary for ActivityPub, any applicable\n            [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] property may be used on ActivityPub Actors.\n            Some ActivityStreams properties are particularly worth highlighting\n            to demonstrate how they are used in ActivityPub implementations.\n          <\/p><dl><dt id=\"url-property\">url<\/dt>\n            <dd>\n              A link to the actor's \"profile web page\", if not equal to the\n              value of <code>id<\/code>.\n            <\/dd>\n            <dt id=\"name-property\">name<\/dt>\n            <dd>\n              The preferred \"nickname\" or \"display name\" of the actor.\n            <\/dd>\n            <dt id=\"summary-property\">summary<\/dt>\n            <dd>A quick summary or bio by the user about themselves.<\/dd>\n            <dt id=\"icon-property\">icon<\/dt>\n            <dd>\n              A link to an image or an Image object which represents the user's\n              profile picture (this may be a thumbnail).\n            <\/dd>\n          <\/dl><\/div><\/div><\/section><\/section><section id=\"collections\"><h2 id=\"x5-collections\">5. Collections<\/h2><p>\n        [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] defines the collection concept; ActivityPub\n        defines several collections with special behavior.\n        Note that ActivityPub makes use of\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/#paging\">\n          ActivityStreams paging<\/a>\n        to traverse large sets of objects.\n      <\/p><p>\n        Note that some of these collections are specified to be of type\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n          <code>OrderedCollection<\/code><\/a>\n        specifically, while others are permitted to be either a\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-collection\">\n          <code>Collection<\/code><\/a>\n        or an\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n          <code>OrderedCollection<\/code><\/a>.\n        An\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n          <code>OrderedCollection<\/code><\/a>\n        <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be presented consistently in reverse chronological order.\n      <\/p><div class=\"note\"><p>Note<\/p><p class=\"\">\n        What property is used to determine the reverse chronological order\n        is intentionally left as an implementation detail.\n        For example, many SQL-style databases use an incrementing integer\n        as an identifier, which can be reasonably used for handling\n        insertion order in most cases.\n        In other databases, an insertion time timestamp may be preferred.\n        What is used isn't important, but the ordering of elements must\n        remain intact, with newer items first.\n        A property which changes regularly, such a \"last updated\" timestamp,\n        should not be used.\n      <\/p><\/div><section id=\"outbox\"><h3 id=\"x5-1-outbox\">5.1 Outbox<\/h3><p>\n          The outbox is discovered through the <code>outbox<\/code>\n          property of an <a href=\"#actors\">actor's<\/a> profile.\n          The <code>outbox<\/code> <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be an\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n            <code>OrderedCollection<\/code><\/a>.\n        <\/p><p>\n          The outbox stream contains activities the user has\n          published, subject to the ability of the requestor to retrieve the\n          activity (that is, the contents of the outbox are filtered by the\n          permissions of the person reading it).\n          If a user submits a request without\n          <a href=\"#authorization\">Authorization<\/a> the server should\n          respond with all of the <a href=\"#public-addressing\">Public<\/a>\n          posts.\n          This could potentially be all relevant objects published by the\n          user, though the number of available items is left to the\n          discretion of those implementing and deploying the server.\n        <\/p><p>\n          The outbox accepts HTTP POST requests, with behaviour described in\n          <a href=\"#client-to-server-interactions\">Client to Server\n            Interactions<\/a>.\n        <\/p><\/section><section id=\"inbox\"><h3 id=\"x5-2-inbox\">5.2 Inbox<\/h3><p>\n          The inbox is discovered through the <code>inbox<\/code>\n          property of an <a href=\"#actors\">actor's<\/a> profile.\n          The <code>inbox<\/code> <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be an\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n            <code>OrderedCollection<\/code><\/a>.\n        <\/p><p>\n          The inbox stream contains all activities received by the\n          actor.\n          The server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> filter content according to the requester's\n          permission.\n          In general, the owner of an inbox is likely to be able to access\n          all of their inbox contents. Depending on access control, some\n          other content may be public, whereas other content may require\n          authentication for non-owner users, if they can access the inbox\n          at all.\n        <\/p><p>\n          The server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> perform de-duplication of activities returned by\n          the inbox. Duplication can occur if an activity is addressed both\n          to an actor's followers, and a specific\n          actor who also follows the recipient actor, and the server has failed\n          to de-duplicate the recipients list.\n          Such deduplication <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be performed by comparing the\n          <code>id<\/code> of the activities and dropping any activities\n          already seen.\n        <\/p><p>\n          The inboxes of actors on federated servers accepts HTTP POST requests,\n          with behaviour described in <a href=\"#delivery\">Delivery<\/a>.\n          Non-federated servers <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> return a 405 Method Not Allowed upon\n          receipt of a POST request.\n        <\/p><\/section><section id=\"followers\"><h3 id=\"x5-3-followers-collection\">5.3 Followers Collection<\/h3><p>\n          Every <a href=\"#actors\">actor<\/a> <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> have a <code>followers<\/code>\n          collection.\n          This is a list of everyone who has sent a\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-follow\">Follow<\/a>\n          activity for the actor, added as a\n          <a href=\"#follow-activity-outbox\">side effect<\/a>.\n          This is where one would find a list of all the actors that are\n          following the actor.\n          The <code>followers<\/code> collection <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be either an\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n            <code>OrderedCollection<\/code><\/a>\n          or a\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-collection\">\n            <code>Collection<\/code><\/a>\n          and <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be filtered on privileges of an authenticated user\n          or as appropriate when no authentication is given.\n        <\/p><div class=\"note\"><p>Note: Default for notification targeting<\/p><p class=\"\">\n          The follow activity generally is a request to see the objects an actor\n          creates. This makes the Followers collection an appropriate default\n          target for <a href=\"#delivery\">delivery<\/a> of notifications.\n        <\/p><\/div><\/section><section id=\"following\"><h3 id=\"x5-4-following-collection\">5.4 Following Collection<\/h3><p>\n          Every actor <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> have a <code>following<\/code> collection.\n          This is a list of everybody that the actor has followed, added as a\n          <a href=\"#follow-activity-outbox\">side effect<\/a>.\n          The <code>following<\/code> collection <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be either an\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n            <code>OrderedCollection<\/code><\/a>\n          or a\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-collection\">\n            <code>Collection<\/code><\/a>\n          and <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be filtered on privileges of an authenticated user\n          or as appropriate when no authentication is given.\n        <\/p><\/section><section id=\"liked\"><h3 id=\"x5-5-liked-collection\">5.5 Liked Collection<\/h3><p>\n          Every actor <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> have a <code>liked<\/code> collection.\n          This is a list of every object from all of the actor's <code>Like<\/code>\n          activities, added as a <a href=\"#like-activity-outbox\">side effect<\/a>.\n          The <code>liked<\/code> collection <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be either an\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n            <code>OrderedCollection<\/code><\/a>\n          or a\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-collection\">\n            <code>Collection<\/code><\/a>\n          and <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be filtered on privileges of an authenticated user\n          or as appropriate when no authentication is given.\n        <\/p><\/section><section id=\"public-addressing\"><h3 id=\"x5-6-public-addressing\">5.6 Public Addressing<\/h3><p>\n          In addition to [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] collections and objects,\n          Activities may additionally be addressed to the special \"public\"\n          collection, with the identifier\n          <code>https:\/\/www.w3.org\/ns\/activitystreams#Public<\/code>.\n          For example:\n        <\/p><div class=\"example\"><p>Example 10<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n  \"id\": \"https:\/\/www.w3.org\/ns\/activitystreams#Public\",\n  \"type\": \"Collection\"\n}<\/pre><\/div><p>\n          Activities addressed to this special URI shall be accessible to all\n          users, without authentication.\n          Implementations <em class=\"rfc2119\" title=\"MUST NOT\">MUST NOT<\/em> deliver to the \"public\" special collection;\n          it is not capable of receiving actual activities.\n          However, actors <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> have a\n          <a href=\"#sharedInbox\"><code>sharedInbox<\/code><\/a>\n          endpoint which is available for efficient shared delivery of public\n          posts (as well as posts to followers-only); see\n          <a href=\"#shared-inbox-delivery\" class=\"sec-ref\">7.1.3 Shared Inbox Delivery<\/a>.\n        <\/p><div class=\"note\"><p>Note<\/p><div class=\"\"><p>\n            Compacting an ActivityStreams object using the ActivityStreams\n            JSON-LD context might result in\n            <code>https:\/\/www.w3.org\/ns\/activitystreams#Public<\/code>\n            being represented as simply <code>Public<\/code> or <code>as:Public<\/code>\n            which are valid representations of the Public collection.\n            Implementations which treat ActivityStreams objects as simply\n            JSON rather than converting an incoming activity over to a\n            local context using JSON-LD tooling should be aware of this\n            and should be prepared to accept all three representations.\n          <\/p><\/div><\/div><\/section><section id=\"likes\"><h3 id=\"x5-7-likes-collection\">5.7 Likes Collection<\/h3><p>\n          Every object <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> have a <code>likes<\/code> collection.\n          This is a list of all <code>Like<\/code> activities with this object\n          as the <code>object<\/code> property, added as a\n          <a href=\"#like-activity-inbox\">side effect<\/a>.\n          The <code>likes<\/code> collection <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be either an\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-orderedcollection\">\n            <code>OrderedCollection<\/code><\/a>\n          or a\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#dfn-collection\">\n            <code>Collection<\/code><\/a>\n          and <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be filtered on privileges of an authenticated user\n          or as appropriate when no authentication is given.\n        <\/p><div class=\"note\"><p>Note<\/p><div class=\"\"><p>\n            Care should be taken to not confuse the the\n            <a href=\"#likes\"><code>likes<\/code><\/a>\n            collection with the similarly named but different\n            <a href=\"#liked\"><code>liked<\/code><\/a> collection.\n            In sum:\n          <\/p><ul><li>\n              <b>liked:<\/b>\n              Specifically a property of actors.\n              This is a collection of <code>Like<\/code> activities performed\n              <i>by the actor<\/i>,\n              added to the collection as a\n              <a href=\"#like-activity-outbox\">\n                side effect of delivery to the outbox<\/a>.\n            <\/li>\n            <li>\n              <b>likes:<\/b>\n              May be a property of any object.\n              This is a collection of <code>Like<\/code> activities referencing\n              this object,\n              added to the collection as a\n              <a href=\"#like-activity-inbox\">\n                side effect of delivery to the inbox<\/a>.\n            <\/li>\n          <\/ul><\/div><\/div><\/section><\/section><section id=\"client-to-server-interactions\"><h2 id=\"x6-client-to-server-interactions\">6. Client to Server Interactions<\/h2><p>\n        Activities as defined by [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] are the core mechanism\n        for creating, modifying and sharing content within the social graph.\n      <\/p><p>\n        Client to server interaction takes place through clients posting\n        Activities to an actor's <a href=\"#outbox\">outbox<\/a>.\n        To do this, clients <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> discover the URL of the actor's outbox from\n\ttheir <a href=\"#actor-objects\">profile<\/a> and then <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> make an HTTP\n        <code>POST<\/code> request to this URL with the Content-Type of\n        <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>.\n        Servers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> interpret a Content-Type or Accept header of\n        <code>application\/activity+json<\/code> as equivalent\n        to <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>\n        for client-to-server interactions.\n        The request <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be authenticated with the credentials of the user to\n        whom the outbox belongs.\n        The body of the <code>POST<\/code> request <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> contain a single\n        Activity (which <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> contain embedded objects), or a single non-Activity\n        object which\n        <a href=\"#object-without-create\">will be wrapped in a Create activity\n          by the server<\/a>.\n      <\/p><div class=\"example\"><p>Example 11: Submitting an Activity to the Outbox<\/p><pre class=\"hljs http\" aria-busy=\"false\" aria-live=\"polite\">POST \/outbox\/ HTTP\/1.1\nHost: dustycloud.org\nAuthorization: Bearer XXXXXXXXXXX\nContent-Type: application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"\n{\n  \"@context\": [\"https:\/\/www.w3.org\/ns\/activitystreams\",\n               {\"@language\": \"en\"}],\n  \"type\": \"Like\",\n  \"actor\": \"https:\/\/dustycloud.org\/chris\/\",\n  \"name\": \"Chris liked 'Minimal ActivityPub update client'\",\n  \"object\": \"https:\/\/rhiaro.co.uk\/2016\/05\/minimal-activitypub\",\n  \"to\": [\"https:\/\/rhiaro.co.uk\/#amy\",\n         \"https:\/\/dustycloud.org\/followers\",\n         \"https:\/\/rhiaro.co.uk\/followers\/\"],\n  \"cc\": \"https:\/\/e14n.com\/evan\"\n}<\/pre><\/div><p>\n        If an Activity is submitted with a value in the <code>id<\/code>\n        property, servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> ignore this and generate a new <code>id<\/code>\n        for the Activity.\n        Servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> return a <code>201 Created<\/code> HTTP code, and unless\n        the activity is transient, <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> include the new <code>id<\/code> in the\n        <code>Location<\/code> header.\n      <\/p><div class=\"example\"><p>Example 12: Outbox response to submitted Activity<\/p><pre class=\"hljs http\" aria-busy=\"false\" aria-live=\"polite\">HTTP\/1.1 201 Created\nLocation: https:\/\/dustycloud.org\/likes\/345<\/pre><\/div><p id=\"remove-bto-bcc-before-delivery\">\n        The server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> remove the <code>bto<\/code> and\/or <code>bcc<\/code>\n        properties, if they exist, from the ActivityStreams object before\n        delivery, but <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> utilize the addressing originally stored\n        on the <code>bto<\/code> \/ <code>bcc<\/code> properties for determining\n        recipients in <a href=\"#delivery\">delivery<\/a>.\n      <\/p><p>\n        The server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> then add this new Activity to the\n        <a href=\"#outbox\">outbox<\/a> collection.\n        Depending on the type of Activity, servers may then be required to\n        carry out further side effects.\n        (However, there is no guarantee that time the Activity may appear in\n        the outbox.\n        The Activity might appear after a delay or disappear at any period).\n        These are described per individual Activity below.\n      <\/p><p>\n        Attempts to submit objects to servers not implementing client to server\n        support <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> result in a <code>405 Method Not Allowed<\/code> response.\n      <\/p><p>\n        HTTP caching mechanisms [<cite><a class=\"bibref\" href=\"#bib-RFC7234\">RFC7234<\/a><\/cite>] <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> be respected when\n        appropriate, both in clients receiving responses from servers as well\n        as servers sending responses to clients.\n      <\/p><section id=\"client-addressing\"><h3 id=\"x6-1-client-addressing\">6.1 Client Addressing<\/h3><p>\n          <strong>Clients<\/strong> are responsible for addressing new Activities\n          appropriately.\n          To some extent, this is dependent upon the particular client\n          implementation, but clients must be aware that the server will only\n          forward new Activities to addressees in the <code>to<\/code>,\n          <code>bto<\/code>, <code>cc<\/code>, <code>bcc<\/code>, and\n          <code>audience<\/code> fields.\n        <\/p><p>\n          The <a href=\"#followers\">Followers Collection<\/a> and\/or the\n          <a href=\"#public-addressing\">Public Collection<\/a> are good\n          choices for the default addressing of new Activities.\n        <\/p><p>\n          Clients <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> look at any objects attached to the new Activity via the\n          <code>object<\/code>, <code>target<\/code>, <code>inReplyTo<\/code> and\/or\n          <code>tag<\/code> fields, retrieve <em>their<\/em> <code>actor<\/code> or\n          <code>attributedTo<\/code> properties, and <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> also retrieve their addressing\n          properties, and add these to the <code>to<\/code> or <code>cc<\/code>\n          fields of the new Activity being created.\n          Clients <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> recurse through attached objects, but if doing so, <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          set a limit for this recursion.\n          (Note that this does not suggest that the client should \"unpack\"\n          collections of actors being addressed as individual recipients).\n        <\/p><p>\n          Clients <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> give the user the chance to amend this addressing in the\n          UI.\n        <\/p><p>\n          For example, when Chris likes the following article by Amy:\n        <\/p><div class=\"example\"><p>Example 13: An Article<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": [\"https:\/\/www.w3.org\/ns\/activitystreams\",\n               {\"@language\": \"en-GB\"}],\n  \"id\": \"https:\/\/rhiaro.co.uk\/2016\/05\/minimal-activitypub\",\n  \"type\": \"Article\",\n  \"name\": \"Minimal ActivityPub update client\",\n  \"content\": \"Today I finished morph, a client for posting ActivityStreams2...\",\n  \"attributedTo\": \"https:\/\/rhiaro.co.uk\/#amy\",\n  \"to\": \"https:\/\/rhiaro.co.uk\/followers\/\",\n  \"cc\": \"https:\/\/e14n.com\/evan\"\n}<\/pre><\/div><p>the like is generated by the client as:<\/p><div class=\"example\"><p>Example 14: A Like of the Article<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": [\"https:\/\/www.w3.org\/ns\/activitystreams\",\n               {\"@language\": \"en\"}],\n  \"type\": \"Like\",\n  \"actor\": \"https:\/\/dustycloud.org\/chris\/\",\n  \"summary\": \"Chris liked 'Minimal ActivityPub update client'\",\n  \"object\": \"https:\/\/rhiaro.co.uk\/2016\/05\/minimal-activitypub\",\n  \"to\": [\"https:\/\/rhiaro.co.uk\/#amy\",\n         \"https:\/\/dustycloud.org\/followers\",\n         \"https:\/\/rhiaro.co.uk\/followers\/\"],\n  \"cc\": \"https:\/\/e14n.com\/evan\"\n}<\/pre><\/div><p>The receiving outbox can then perform <a href=\"#delivery\">delivery<\/a>\n        to not only the followers of Chris (the liker), but also to Amy, and Amy's\n        followers and Evan, both of whom received the original article.<\/p><p>\n          Clients submitting the following activities to an <code>outbox<\/code>\n          <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> provide the <code>object<\/code> property in the activity:\n          <code>Create<\/code>, <code>Update<\/code>, <code>Delete<\/code>,\n          <code>Follow<\/code>, <code>Add<\/code>, <code>Remove<\/code>,\n          <code>Like<\/code>, <code>Block<\/code>, <code>Undo<\/code>.\n          Additionally, clients submitting the following activities to an outbox\n          <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> also provide the <code>target<\/code> property:\n          <code>Add<\/code>, <code>Remove<\/code>.\n        <\/p><\/section><section id=\"create-activity-outbox\"><h3 id=\"x6-2-create-activity\">6.2 Create Activity<\/h3><p>\n          The <code>Create<\/code> activity is used when posting a new object.\n          This has the side effect that the object embedded within the Activity\n          (in the <code>object<\/code> property) is created.\n        <\/p><p>\n          When a <code>Create<\/code> activity is posted, the <code>actor<\/code>\n          of the activity <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> be copied onto the <code>object<\/code>'s\n          <code>attributedTo<\/code> field.\n        <\/p><p>\n          A mismatch between addressing of the Create activity and its\n          <code>object<\/code> is likely to lead to confusion.\n          As such, a server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> copy any recipients of the Create activity\n          to its <code>object<\/code> upon initial distribution, and likewise\n          with copying recipients from the <code>object<\/code> to the wrapping\n          Create activity.\n          Note that it is acceptable for the <code>object<\/code>'s addressing\n          to be changed later without changing the <code>Create<\/code>'s\n          addressing (for example via an <code>Update<\/code> activity).\n        <\/p><section id=\"object-without-create\"><h4 id=\"x6-2-1-object-creation-without-a-create-activity\">6.2.1 Object creation without a Create Activity<\/h4><p>\n          For client to server posting, it is possible to submit an object for\n          creation without a surrounding activity.\n          The server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> accept a valid [<cite><a class=\"bibref\" href=\"#bib-ActivityStreams\">ActivityStreams<\/a><\/cite>] object that\n          isn't a subtype of <code>Activity<\/code> in the POST request to the\n          outbox.\n          The server then <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> attach this object as the <code>object<\/code>\n          of a <a href=\"#create-activity-outbox\">Create Activity<\/a>.\n          For non-transient objects, the server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> attach an\n          <code>id<\/code> to both the wrapping <code>Create<\/code> and its\n          wrapped <code>Object<\/code>.\n        <\/p><div class=\"note\"><p>Note<\/p><p class=\"\">\n          The <code>Location<\/code> value returned by the server should be the URL of\n          the new Create activity (rather than the object).\n        <\/p><\/div><p>\n          Any <code>to<\/code>, <code>bto<\/code>, <code>cc<\/code>, <code>bcc<\/code>,\n          and <code>audience<\/code> properties specified on the object <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be\n          copied over to the new Create activity by the server.\n        <\/p><div class=\"example\"><p>Example 15: Object with audience targeting<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n  \"type\": \"Note\",\n  \"content\": \"This is a note\",\n  \"published\": \"2015-02-10T15:04:55Z\",\n  \"to\": [\"https:\/\/example.org\/~john\/\"],\n  \"cc\": [\"https:\/\/example.com\/~erik\/followers\",\n         \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"]\n}<\/pre><\/div>The above example could be converted to this:<div class=\"example\"><p>Example 16: Create Activity wrapper generated by the server<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n  \"type\": \"Create\",\n  \"id\": \"https:\/\/example.net\/~mallory\/87374\",\n  \"actor\": \"https:\/\/example.net\/~mallory\",\n  \"object\": {\n    \"id\": \"https:\/\/example.com\/~mallory\/note\/72\",\n    \"type\": \"Note\",\n    \"attributedTo\": \"https:\/\/example.net\/~mallory\",\n    \"content\": \"This is a note\",\n    \"published\": \"2015-02-10T15:04:55Z\",\n    \"to\": [\"https:\/\/example.org\/~john\/\"],\n    \"cc\": [\"https:\/\/example.com\/~erik\/followers\",\n           \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"]\n  },\n  \"published\": \"2015-02-10T15:04:55Z\",\n  \"to\": [\"https:\/\/example.org\/~john\/\"],\n  \"cc\": [\"https:\/\/example.com\/~erik\/followers\",\n         \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"]\n}<\/pre><\/div><\/section><\/section><section id=\"delete-activity-outbox\"><h3 id=\"x6-4-delete-activity\">6.4 Delete Activity<\/h3><p>\n          The <code>Delete<\/code> activity is used to delete an already\n          existing object.\n          The side effect of this is that the server <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> replace the\n          <code>object<\/code> with a <code>Tombstone<\/code> of the object\n          that will be displayed in activities which reference the deleted\n          object.\n          If the deleted object is requested the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> respond with\n          either the HTTP 410 Gone status code if a <code>Tombstone<\/code>\n          object is presented as the response body, otherwise respond with a\n          HTTP 404 Not Found.\n        <\/p><p>A deleted object:<\/p><div class=\"example\"><p>Example 17<\/p><pre class=\"hljs json\" aria-busy=\"false\" aria-live=\"polite\">{\n  \"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n  \"id\": \"https:\/\/example.com\/~alice\/note\/72\",\n  \"type\": \"Tombstone\",\n  \"published\": \"2015-02-10T15:04:55Z\",\n  \"updated\": \"2015-02-10T15:04:55Z\",\n  \"deleted\": \"2015-02-10T15:04:55Z\"\n}<\/pre><\/div><\/section><section id=\"follow-activity-outbox\"><h3 id=\"x6-5-follow-activity\">6.5 Follow Activity<\/h3><p>\n          The <code>Follow<\/code> activity is used to subscribe to the\n          activities of another actor.\n        <\/p><p>\n          The side effect of receiving this in an <strong>outbox<\/strong>\n          is that the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> add the <code>object<\/code> to the\n          <code>actor<\/code>'s <a href=\"#following\"><code>following<\/code> Collection<\/a>\n          when and only if an <code>Accept<\/code> activity is subsequently\n          received with this <code>Follow<\/code> activity as its object.\n        <\/p><\/section><section id=\"add-activity-outbox\"><h3 id=\"x6-6-add-activity\">6.6 Add Activity<\/h3><p>\n          Upon receipt of an <code>Add<\/code> activity into the\n          <strong>outbox<\/strong>, the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          add the <code>object<\/code> to the collection specified in the\n          <code>target<\/code> property, unless:\n        <\/p><ul><li>\n            the <code>target<\/code> is not owned by the receiving\n            server, and thus they are not authorized to update it.\n          <\/li>\n          <li>\n            the <code>object<\/code> is not allowed to be added to the\n            <code>target<\/code> collection for some other reason, at the\n            receiving server's discretion.\n          <\/li>\n        <\/ul><\/section><section id=\"remove-activity-outbox\"><h3 id=\"x6-7-remove-activity\">6.7 Remove Activity<\/h3><p>\n          Upon receipt of a <code>Remove<\/code> activity into the\n          <strong>outbox<\/strong>, the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          remove the <code>object<\/code> from the collection specified in the\n          <code>target<\/code> property, unless:\n        <\/p><ul><li>\n            the <code>target<\/code> is not owned by the receiving server, and\n            thus they are not authorized to update it.\n          <\/li>\n          <li>\n            the <code>object<\/code> is not allowed to be removed from the\n            <code>target<\/code> collection for some other reason, at the\n            receiving server's discretion.\n          <\/li>\n        <\/ul><\/section><section id=\"like-activity-outbox\"><h3 id=\"x6-8-like-activity\">6.8 Like Activity<\/h3><p>\n          The <code>Like<\/code> activity indicates the <code>actor<\/code> likes\n          the <code>object<\/code>.\n        <\/p><p>\n          The side effect of receiving this in an <strong>outbox<\/strong>\n          is that the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> add the <code>object<\/code> to the\n          <code>actor<\/code>'s <a href=\"#liked\"><code>liked<\/code> Collection<\/a>.\n        <\/p><\/section><section id=\"block-activity-outbox\"><h3 id=\"x6-9-block-activity\">6.9 Block Activity<\/h3><p>\n          The <code>Block<\/code> activity is used to indicate that the posting\n          actor does not want another actor (defined in the <code>object<\/code>\n          property) to be able to interact with objects posted by the actor\n          posting the <code>Block<\/code> activity.\n          The server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> prevent the blocked user from interacting with any object\n          posted by the actor.\n        <\/p><p>\n          Servers <em class=\"rfc2119\" title=\"SHOULD NOT\">SHOULD NOT<\/em> deliver Block Activities to their <code>object<\/code>.\n        <\/p><\/section><section id=\"undo-activity-outbox\"><h3 id=\"x6-10-undo-activity\">6.10 Undo Activity<\/h3><p>\n          The <code>Undo<\/code> activity is used to undo a previous activity.\n          See the Activity Vocabulary documentation on\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#inverse\">\n            Inverse Activities and \"Undo\"<\/a>.\n          For example, <code>Undo<\/code> may be used to undo a previous\n          <code>Like<\/code>, <code>Follow<\/code>, or <code>Block<\/code>.\n          The undo activity and the activity being undone <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> both have the\n          same actor.\n          Side effects should be undone, to the extent possible.\n          For example, if undoing a <code>Like<\/code>, any counter that had\n          been incremented previously should be decremented appropriately.\n        <\/p><p>\n          There are some exceptions where there is an existing and explicit\n          \"inverse activity\" which should be used instead.\n          <code>Create<\/code> based activities should instead use\n          <code>Delete<\/code>, and <code>Add<\/code> activities should use\n          <code>Remove<\/code>.\n        <\/p><\/section><section id=\"client-to-server-outbox-delivery\"><h3 id=\"x6-11-delivery\">6.11 Delivery<\/h3><p>\n          Federated servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> perform delivery on all Activities posted to the\n          <strong>outbox<\/strong> according to <a href=\"#outbox-delivery\">\n          outbox delivery<\/a>.\n        <\/p><\/section><section id=\"uploading-media\" class=\"informative\"><h3 id=\"x6-12-uploading-media\">6.12 Uploading Media<\/h3><p><em>This section is non-normative.<\/em><\/p><p>\n          Servers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> support uploading document types to be referenced in\n          activites, such as images, video or other binary data, but the\n          precise mechanism is out of scope for this version of ActivityPub.\n          The Social Web Community Group is refining the protocol in the\n          <a href=\"https:\/\/www.w3.org\/wiki\/SocialCG\/ActivityPub\/MediaUpload\">ActivityPub Media Upload report<\/a>.\n        <\/p><\/section><\/section><section id=\"server-to-server-interactions\"><h2 id=\"x7-server-to-server-interactions\">7. Server to Server Interactions<\/h2><p>\n        Servers communicate with other servers and propagate information across\n        the social graph by posting activities to actors'\n        <a href=\"#inbox\">inbox<\/a> endpoints.\n        An Activity sent over the network <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> have an <code>id<\/code>,\n        unless it is intended to be transient (in which case it <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> omit the\n        <code>id<\/code>).\n      <\/p><p>\n        <code>POST<\/code> requests (eg. to the inbox) <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> be made with a Content-Type of\n        <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>\n        and <code>GET<\/code> requests (see also <a href=\"#retrieving-objects\" class=\"sec-ref\">3.2 Retrieving objects<\/a>) \n        with an Accept header of \n        <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>.\n        Servers <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> interpret a Content-Type or Accept\n        header of <code>application\/activity+json<\/code> as equivalent\n        to <code>application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"<\/code>\n        for server-to-server interactions.\n      <\/p><p>\n        In order to propagate updates throughout the social graph, Activities\n        are sent to the appropriate recipients.\n        First, these recipients are determined through following the\n        appropriate links between objects until you reach an actor, and then\n        the Activity is inserted into the actor's <em>inbox<\/em>\n        (<a href=\"#delivery\">delivery<\/a>).\n        This allows recipient servers to:\n      <\/p><ul><li>\n          conduct any side effects related to the Activity (for example,\n          notification that an actor has liked an object is used to update the\n          object's like count);\n        <\/li>\n        <li>\n          deliver the Activity to recipients of the original object, to ensure\n          updates are propagated to the whole social graph (see <a href=\"#inbox-forwarding\">\n          inbox delivery<\/a>).\n        <\/li>\n      <\/ul><p>Delivery is usually triggered by, for example:<\/p><ul><li>\n          an Activity being created in an actor's <a href=\"#outbox\">outbox<\/a>\n          with their <a href=\"#followers\">Followers Collection<\/a> as the\n          recipient.\n        <\/li>\n        <li>\n          an Activity being created in an actor's <a href=\"#outbox\">outbox<\/a>\n          with directly addressed recipients.\n        <\/li>\n        <li>\n          an Activity being created in an actors's <a href=\"#outbox\">outbox<\/a>\n          with user-curated collections as recipients.\n        <\/li>\n        <li>\n          an Activity being created in an actor's <a href=\"#outbox\">outbox<\/a> or\n          <a href=\"#inbox\">inbox<\/a> which references another object.\n        <\/li>\n      <\/ul><p>\n        Servers performing delivery to the <code>inbox<\/code> or\n        <code>sharedInbox<\/code> properties of actors on other servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em>\n        provide the <code>object<\/code> property in the activity:\n        <code>Create<\/code>, <code>Update<\/code>, <code>Delete<\/code>,\n        <code>Follow<\/code>, <code>Add<\/code>, <code>Remove<\/code>,\n        <code>Like<\/code>, <code>Block<\/code>, <code>Undo<\/code>.\n        Additionally, servers performing server to server delivery of the\n        following activities <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> also provide the <code>target<\/code>\n        property: <code>Add<\/code>, <code>Remove<\/code>.\n      <\/p><p>\n        HTTP caching mechanisms [<cite><a class=\"bibref\" href=\"#bib-RFC7234\">RFC7234<\/a><\/cite>] <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> be respected when\n        appropriate, both when receiving responses from other servers as well\n        as sending responses to other servers.\n      <\/p><section id=\"delivery\"><h3 id=\"x7-1-delivery\">7.1 Delivery<\/h3><p><i>\n          The following is required by federated servers communicating with\n          other federated servers only.\n        <\/i><\/p><p>\n          An activity is delivered to its targets (which are\n          <a href=\"#actors\">actors<\/a>) by first looking up the targets'\n          inboxes and then posting the activity to those inboxes.\n          Targets for delivery are determined by checking the\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#audienceTargeting\">\n            ActivityStreams audience targeting<\/a>;\n          namely, the <code>to<\/code>, <code>bto<\/code>, <code>cc<\/code>,\n          <code>bcc<\/code>, and <code>audience<\/code> fields of the activity.\n        <\/p><p>\n          The <a href=\"#inbox\">inbox<\/a> is determined by first\n          <a href=\"#retrieving-objects\">\n            retrieving the target actor's JSON-LD representation<\/a>\n          and then looking up the <code>inbox<\/code> property.\n          If a recipient is a <code>Collection<\/code> or <code>OrderedCollection<\/code>,\n          then the server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> dereference the collection (with the user's\n          credentials) and discover inboxes for each item in the collection.\n          Servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> limit the number of layers of indirections through\n          collections which will be performed, which <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be one.\n        <\/p><p>\n          Servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> de-duplicate the final recipient list. Servers <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> also\n          exclude actors from the list which are the same as the <code>actor<\/code>\n          of the Activity being notified about. That is, actors shouldn't have their\n          own activities delivered to themselves.\n        <\/p><div class=\"note\"><p>Note: Silent and private activities<\/p><p class=\"\">\n          What to do when there are no recipients specified is not defined,\n          however it's recommended that if no recipients are specified the\n          object remains completely private and access controls restrict the\n          access to object.\n          If the object is just sent to the \"public\" collection the object is\n          not delivered to any actors but is publicly viewable in the actor's\n          outbox.\n        <\/p><\/div><p>\n          An HTTP POST request (with authorization of the submitting user) is\n          then made to the <a href=\"#inbox\">inbox<\/a>, with the Activity as\n          the body of the request.\n          This Activity is added by the receiver as an <code>item<\/code> in the\n          <a href=\"#inbox\">inbox<\/a> OrderedCollection.\n          Attempts to deliver to an inbox on a non-federated server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          result in a <code>405 Method Not Allowed<\/code> response.\n        <\/p><p>\n          For federated servers performing delivery to a third party server,\n          delivery <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> be performed asynchronously, and <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> additionally\n          retry delivery to recipients if it fails due to network error.\n        <\/p><p>\n          <strong>Note:<\/strong> Activities being distributed between actors on\n          the same origin may use any internal mechanism, and are not required\n          to use HTTP.\n        <\/p><div class=\"note\"><p>Note: Relationship to Linked Data Notifications<\/p><p class=\"\" id=\"ldn-relationship\">\n          While it is not required reading to understand this specification,\n          it is worth noting that ActivityPub's targeting and delivery\n          mechanism overlaps with the\n          <a href=\"https:\/\/www.w3.org\/TR\/ldn\/\">Linked Data Notifications<\/a>\n          specification, and the two specifications may interoperably\n          combined.\n          In particular, the <code>inbox<\/code> property is the same between\n          ActivityPub and Linked Data Notifications, and the targeting\n          and delivery systems described in this document are supported\n          by Linked Data Notifications.\n          In addition to JSON-LD compacted ActivityStreams documents, Linked\n          Data Notifications also supports a number of RDF serializations\n          which are not required for ActivityPub implementations.\n          However, ActivityPub implementations which wish to be more broadly\n          compatible with Linked Data Notifications implementations may wish to\n          support other RDF representations.\n        <\/p><\/div><section id=\"outbox-delivery\"><h4 id=\"x7-1-1-outbox-delivery-requirements-for-server-to-server\">7.1.1 Outbox Delivery Requirements for Server to Server<\/h4><p>\n            When objects are received in the <a href=\"#outbox\">outbox<\/a>\n            (for servers which support both\n            <a href=\"#client-to-server-interactions\">Client to Server interactions<\/a>\n            and\n            <a href=\"#server-to-server-interactions\">Server to Server Interactions<\/a>),\n            the server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> target and deliver to:\n          <\/p><ul><li>\n              The <code>to<\/code>, <code>bto<\/code>, <code>cc<\/code>,\n              <code>bcc<\/code> or <code>audience<\/code> fields if their values\n              are individuals or Collections owned by the actor.\n            <\/li>\n          <\/ul><p>\n            These fields will have been <a href=\"#client-addressing\">\n            populated appropriately by the client<\/a> which posted the Activity\n            to the outbox.\n          <\/p><\/section><section id=\"inbox-forwarding\"><h4 id=\"x7-1-2-forwarding-from-inbox\">7.1.2 Forwarding from Inbox<\/h4><div class=\"note\"><p>Note: Forwarding to avoid the ghost replies problem<\/p><div class=\"\"><p>\n              The following section is to mitigate the \"ghost replies\" problem\n              which occasionally causes problems on federated networks.\n              This problem is best demonstrated with an example.\n            <\/p><p>\n              Alyssa makes a post about her having successfully presented a\n              paper at a conference and sends it to her followers collection,\n              which includes her friend Ben.\n              Ben replies to Alyssa's message congratulating her and includes\n              her followers collection on the recipients.\n              However, Ben has no access to see the members of Alyssa's\n              followers collection, so his server does not forward his messages\n              to their inbox.\n              Without the following mechanism, if Alyssa were then to reply to\n              Ben, her followers would see Alyssa replying to Ben without having\n              ever seen Ben interacting.\n              This would be very confusing!\n            <\/p><\/div><\/div><p>\n            When Activities are received in the <a href=\"#inbox\">inbox<\/a>, the\n            server needs to forward these to recipients that the origin was unable\n            to deliver them to. To do this, the server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> target and\n            <a href=\"#delivery\">deliver<\/a>\n            to the values of <code>to<\/code>, <code>cc<\/code>, and\/or <code>audience<\/code>\n            if and only if all of the following are true:\n          <\/p><ul><li>This is the first time the server has seen this Activity.<\/li>\n            <li>\n              The values of <code>to<\/code>, <code>cc<\/code>, and\/or\n              <code>audience<\/code> contain a Collection owned by the server.\n            <\/li>\n            <li>\n              The values of <code>inReplyTo<\/code>, <code>object<\/code>,\n              <code>target<\/code> and\/or <code>tag<\/code> are objects owned by\n              the server.\n              The server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> recurse through these values to look for linked objects\n              owned by the server, and <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> set a maximum limit for recursion (ie. the\n              point at which the thread is so deep the recipients followers may not mind\n              if they are no longer getting updates that don't directly involve the\n              recipient).\n              The server <em class=\"rfc2119\" title=\"MUST\">MUST<\/em> only target the values of <code>to<\/code>,\n              <code>cc<\/code>, and\/or <code>audience<\/code>\n              on the original object being forwarded, and not pick up any new\n              addressees whilst recursing through the linked objects\n              (in case these addressees were purposefully amended by or via the client).\n            <\/li>\n          <\/ul><p>\n            The server <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> filter its delivery targets according to\n            implementation-specific rules (for example, spam filtering).\n          <\/p><\/section><\/section><section id=\"create-activity-inbox\"><h3 id=\"x7-2-create-activity\">7.2 Create Activity<\/h3><p>\n          Receiving a <code>Create<\/code> activity in an <code>inbox<\/code> has\n          surprisingly few side effects; the activity should appear in the\n          actor's <code>inbox<\/code> and it is likely that the server will want\n          to locally store a representation of this activity and its accompanying\n          object.\n          However, this mostly happens in general with processing activities\n          delivered to an <code>inbox<\/code> anyway.\n        <\/p><\/section><section id=\"delete-activity-inbox\"><h3 id=\"x7-4-delete-activity\">7.4 Delete Activity<\/h3><p>\n          The side effect of receiving this is that (assuming the\n          <code>object<\/code> is owned by the sending actor \/ server) the\n          server receiving the delete activity <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> remove its representation\n          of the <code>object<\/code> with the same <code>id<\/code>, and <em class=\"rfc2119\" title=\"MAY\">MAY<\/em>\n          replace that representation with a <code>Tombstone<\/code> object.\n        <\/p><p>\n          (Note that after an activity has been transmitted from an origin\n          server to a remote server, there is nothing in the ActivityPub\n          protocol that can <em>enforce<\/em> remote deletion of an object's\n          representation).\n        <\/p><\/section><section id=\"follow-activity-inbox\"><h3 id=\"x7-5-follow-activity\">7.5 Follow Activity<\/h3><p>\n          The side effect of receiving this in an <strong>inbox<\/strong> is\n          that the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> generate either an <code>Accept<\/code> or\n          <code>Reject<\/code> activity with the Follow as the\n          <code>object<\/code> and deliver it to the <code>actor<\/code> of the\n          Follow.\n          The <code>Accept<\/code> or <code>Reject<\/code> <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be generated\n          automatically, or <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> be the result of user input (possibly after\n          some delay in which the user reviews).\n          Servers <em class=\"rfc2119\" title=\"MAY\">MAY<\/em> choose to not explicitly send a <code>Reject<\/code>\n          in response to a <code>Follow<\/code>, though implementors ought to\n          be aware that the server sending the request could be left in an\n          intermediate state.\n          For example, a server might not send a <code>Reject<\/code> to protect\n          a user's privacy.\n        <\/p><p>\n          In the case of receiving an <code>Accept<\/code> referencing this\n          <code>Follow<\/code> as the object, the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> add the\n          <code>actor<\/code> to the object actor's\n          <a href=\"#followers\">Followers Collection<\/a>.\n          In the case of a <code>Reject<\/code>, the server <em class=\"rfc2119\" title=\"MUST NOT\">MUST NOT<\/em> add the\n          actor to the object actor's\n          <a href=\"#followers\">Followers Collection<\/a>.\n        <\/p><div class=\"note\"><p>Note<\/p><div class=\"\"><p>\n            Sometimes a successful <code>Follow<\/code> subscription may\n            occur but at some future point delivery to the follower fails\n            for an extended period of time.\n            Implementations should be aware that there is no guarantee\n            that actors on the network will remain reachable and should\n            implement accordingly.\n            For instance, if attempting to deliver to an actor for perhaps\n            six months while the follower remains unreachable, it is\n            reasonable that the delivering server remove the subscriber\n            from the <code>followers<\/code> list.\n            Timeframes and behavior for dealing with unreachable actors are\n            left to the discretion of the delivering server.\n          <\/p><\/div><\/div><\/section><section id=\"accept-activity-inbox\"><h3 id=\"x7-6-accept-activity\">7.6 Accept Activity<\/h3><p>\n          The side effect of receiving this in an <strong>inbox<\/strong> is\n          determined by the type of the <code>object<\/code> received,\n          and it is possible to accept types not described in this document\n          (for example, an <code>Offer<\/code>).\n        <\/p><p>\n          If the <code>object<\/code> of an <code>Accept<\/code> received to an\n          <strong>inbox<\/strong> is a <code>Follow<\/code> activity\n          previously sent by the receiver, the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> add the\n          <code>actor<\/code> to the receiver's\n          <a href=\"#following\">Following Collection<\/a>.\n        <\/p><\/section><section id=\"reject-activity-inbox\"><h3 id=\"x7-7-reject-activity\">7.7 Reject Activity<\/h3><p>\n          The side effect of receiving this in an <strong>inbox<\/strong> is\n          determined by the type of the <code>object<\/code> received,\n          and it is possible to reject types not described in this document\n          (for example, an <code>Offer<\/code>).\n        <\/p><p>\n          If the <code>object<\/code> of a <code>Reject<\/code> received to an\n          <strong>inbox<\/strong> is a <code>Follow<\/code> activity\n          previously sent by the receiver, this means the recipient did not\n          approve the <code>Follow<\/code> request. The server <em class=\"rfc2119\" title=\"MUST NOT\">MUST NOT<\/em> add the\n          <code>actor<\/code> to the receiver's\n          <a href=\"#following\">Following Collection<\/a>.\n        <\/p><\/section><section id=\"add-activity-inbox\"><h3 id=\"x7-8-add-activity\">7.8 Add Activity<\/h3><p>\n          Upon receipt of an <code>Add<\/code> activity into the\n          <strong>inbox<\/strong>, the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          add the <code>object<\/code> to the collection specified in the\n          <code>target<\/code> property, unless:\n        <\/p><ul><li>\n            the <code>target<\/code> is not owned by the receiving server, and\n            thus they can't update it.\n          <\/li>\n          <li>\n            the <code>object<\/code> is not allowed to be added to the\n            <code>target<\/code> collection for some other reason, at the\n            receiver's discretion.\n          <\/li>\n        <\/ul><\/section><section id=\"remove-activity-inbox\"><h3 id=\"x7-9-remove-activity\">7.9 Remove Activity<\/h3><p>\n          Upon receipt of a <code>Remove<\/code> activity into the\n          <strong>inbox<\/strong>, the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em>\n          remove the <code>object<\/code> from the collection specified in the\n          <code>target<\/code> property, unless:\n        <\/p><ul><li>\n            the <code>target<\/code> is not owned by the receiving server, and\n            thus they can't update it.\n          <\/li>\n          <li>\n            the <code>object<\/code> is not allowed to be removed to the\n            <code>target<\/code> collection for some other reason, at the\n            receiver's discretion.\n          <\/li>\n        <\/ul><\/section><section id=\"like-activity-inbox\"><h3 id=\"x7-10-like-activity\">7.10 Like Activity<\/h3><p>\n          The side effect of receiving this in an <strong>inbox<\/strong> is\n          that the server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> increment the object's count of likes by\n          adding the received activity to the\n          <a href=\"#likes\"><code>likes<\/code><\/a> collection\n          if this collection is present.\n        <\/p><\/section><section id=\"announce-activity-inbox\"><h3 id=\"x7-11-announce-activity-sharing\">7.11 Announce Activity (sharing)<\/h3><p>\n          Upon receipt of an <code>Announce<\/code> activity in an\n          <strong>inbox<\/strong>, a server <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> increment the object's count\n          of shares by adding the received activity to the\n          <a href=\"#shares\"><code>shares<\/code><\/a> collection\n          if this collection is present.\n        <\/p><div class=\"note\"><p>Note<\/p><div class=\"\">The<code>Announce<\/code>activity is effectively what is known as\n          \"sharing\", \"reposting\", or \"boosting\" in other social networks.<\/div><\/div><\/section><section id=\"undo-activity-inbox\"><h3 id=\"x7-12-undo-activity\">7.12 Undo Activity<\/h3><p>\n          The <code>Undo<\/code> activity is used to undo the side effects\n          of previous activities.\n          See the ActivityStreams documentation on\n          <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/#inverse\">\n            Inverse Activities and \"Undo\"<\/a>.\n          The scope and restrictions of the <code>Undo<\/code> activity\n          are the same as for\n          <a href=\"#undo-activity-outbox\">\n            the Undo activity in the context of client to server\n            interactions<\/a>,\n          but applied to a federated context.\n        <\/p><\/section><\/section><section id=\"i18n-concerns\" class=\"appendix informative\"><h2 id=\"a-internationalization\">A. Internationalization<\/h2><p><em>This section is non-normative.<\/em><\/p><p>\n        Building an international base of users is important in a federated\n        network.\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/#naturalLanguageValues\">\n          ActivityStreams provides tooling for internationalization of content<\/a>,\n        which should be used whenever possible.\n        However, it can be difficult for implementations to determine which\n        <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/#defaultlangcontext\">\n          <code>@language<\/code> property<\/a>\n        to provide for user-submitted content.\n        The\n        <a href=\"https:\/\/www.w3.org\/International\/\"><abbr title=\"World Wide Web Consortium\">W3C<\/abbr> Internationalization group<\/a>\n        provides some\n        <a href=\"https:\/\/www.w3.org\/International\/wiki\/LanguageDetection\">\n          guidance on language detection<\/a>.\n      <\/p><\/section><section id=\"security-considerations\" class=\"appendix informative\"><h2 id=\"b-security-considerations\">B. Security Considerations<\/h2><p><em>This section is non-normative.<\/em><\/p><section id=\"security-verification\"><h3 id=\"b-2-verification\">B.2 Verification<\/h3><p>\n          Servers should not trust client submitted content, and\n          federated servers also should not trust content received from a server\n          other than the content's origin without some form of verification.\n        <\/p><p>\n          Servers should be careful to verify that new content is really posted\n          by the actor that claims to be posting it, and that the actor has\n          permission to update the resources it claims to.\n          See also\n          <a href=\"#obj\" class=\"sec-ref\">3. Objects<\/a> and <a href=\"#authorization\" class=\"sec-ref\">B.1 Authentication and Authorization<\/a>.\n        <\/p><\/section><section id=\"security-localhost\"><h3 id=\"b-3-accessing-localhost-uris\">B.3 Accessing localhost URIs<\/h3><p>\n          It is often convenient while developing to test against a process\n          running on localhost.\n          However, permitting requests to localhost in a production client or\n          server instance can be dangerous.\n          Making requests to URIs on localhost which do not require\n          authorization may unintentionally access or modify resources assumed\n          to be protected to be usable by localhost-only.\n        <\/p><p>\n          If your ActivityPub server or client does permit making requests to\n          localhost URIs for development purposes, consider making it a\n          configuration option which defaults to off.\n        <\/p><\/section><section id=\"security-uri-schemes\"><h3 id=\"b-4-uri-schemes\">B.4 URI Schemes<\/h3><p>\n          There are many types of URIs aside from just <code>http<\/code> and\n          <code>https<\/code>.\n          Some libraries which handle fetching requests at various URI schemes\n          may try to be smart and reference schemes which may be undesirable,\n          such as <code>file<\/code>.\n          Client and server authors should carefully check how their libraries\n          handle requests, and potentially whitelist only certain safe URI\n          types, such as <code>http<\/code> and <code>https<\/code>.\n        <\/p><\/section><section id=\"security-recursive-objects\"><h3 id=\"b-5-recursive-objects\">B.5 Recursive Objects<\/h3><p>\n          Servers should set a limit on how deep to recurse while resolving objects,\n          or otherwise specially handle ActivityStreams objects with recursive\n          references.\n\t  Failure to properly do so may result in denial-of-service security\n\t  vulnerabilities.\n        <\/p><\/section><section id=\"security-spam\"><h3 id=\"b-6-spam\">B.6 Spam<\/h3><p>\n          Spam is a problem in any network, perhaps especially so in federated\n          networks.\n          While no specific mechanism for combating spam is provided in\n          ActivityPub, it is recommended that servers filter incoming content\n          both by local untrusted users and any remote users through some\n          sort of spam filter.\n        <\/p><\/section><section id=\"security-federation-dos\"><h3 id=\"b-7-federation-denial-of-service\">B.7 Federation denial-of-service<\/h3><p>\n            Servers should implement protections against denial-of-service\n            attacks from other, federated servers.\n            This can be done using, for example, some kind of ratelimiting\n            mechanism.\n            Servers should be especially careful to implement this protection\n            around activities that involve side effects.\n            Servers <em class=\"rfc2119\" title=\"SHOULD\">SHOULD<\/em> also take care not to overload servers with\n            submissions, for example by using an exponential backoff strategy.\n          <\/p><\/section><section id=\"security-c2s-ratelimiting\"><h3 id=\"b-8-client-to-server-ratelimiting\">B.8 Client-to-server ratelimiting<\/h3><p>\n            Servers should ratelimit API client submissions.\n            This serves two purposes:\n            <\/p><ol><li>\n                It prevents malicious clients from conducting\n                denial-of-service attacks on the server.\n              <\/li>\n              <li>\n                It ensures that the server will not distribute so many\n                activities that it triggers another\n                server's <a href=\"#security-federation-dos\">denial-of-service\n                protections<\/a>.\n              <\/li>\n            <\/ol><\/section><section id=\"security-c2s-response-dos\"><h3 id=\"b-9-client-to-server-response-denial-of-service\">B.9 Client-to-server response denial-of-service<\/h3><p>\n          In order to prevent a client from being overloaded by oversized\n          Collections, servers should take care to limit the size of Collection \n          pages they return to clients.\n          Clients should still be prepared to limit the size of\n          responses they are willing to handle in case they connect to malicious or\n          compromised servers, for example by timing out and generating\n          an error.\n        <\/p><\/section><section id=\"security-sanitizing-content\"><h3 id=\"b-10-sanitizing-content\">B.10 Sanitizing Content<\/h3><p>\n          Any activity field being rendered for browsers (or other rich\n          text enabled applications) should take care to sanitize fields\n          containing markup to prevent cross site scripting attacks.\n        <\/p><\/section><section id=\"security-not-displaying-bto-bcc\"><h3 id=\"b-11-not-displaying-bto-and-bcc-properties\">B.11 Not displaying bto and bcc properties<\/h3><p>\n          <code>bto<\/code> and <code>bcc<\/code> already\n          <a href=\"#remove-bto-bcc-before-delivery\">must be removed for delivery<\/a>,\n          but servers are free to decide how to represent the object\n          in their own storage systems.\n          However, since <code>bto<\/code> and <code>bcc<\/code> are only intended\n          to be known\/seen by the original author of the object\/activity,\n          servers should omit these properties during display as well.\n        <\/p><\/section><\/section><section class=\"appendix informative\" id=\"acknowledgements\"><h2 id=\"c-acknowledgements\">C. Acknowledgements<\/h2><p><em>This section is non-normative.<\/em><\/p><p>\n        This specification comes from years of hard work and experience by a\n        number of communities exploring the space of federation on the web.\n        In particular, much of this specification is informed by\n        <a href=\"https:\/\/www.w3.org\/community\/ostatus\/wiki\/Main_Page\">OStatus<\/a>\n        and the\n        <a href=\"https:\/\/github.com\/pump-io\/pump.io\/blob\/master\/API.md\">Pump API<\/a>,\n        as pioneered by StatusNet (now GNU Social) and Pump.io.\n        Both of those initiatives were the product of many developers' hard work,\n        but more than anyone, Evan Prodromou has been a constant leader in this\n        space, and it is unlikely that ActivityPub would exist in something\n        resembling its current state without his hard work.\n      <\/p><p>\n        Erin Shepherd built the initial version of this specification, borrowed\n        from the ideas in the\n        <a href=\"https:\/\/github.com\/pump-io\/pump.io\/blob\/master\/API.md\">Pump API<\/a>\n        document, mostly as a complete rewrite of text, but sharing most of\n        the primary ideas while switching from ActivityStreams 1 to\n        ActivityStreams 2.\n      <\/p><p>\n        Jessica Tallon and Christopher Lemmer Webber took over as editors when\n        the standard moved to the <abbr title=\"World Wide Web Consortium\">W3C<\/abbr> Social Working Group and did the majority\n        of transition from Erin Shepherd's document to its current state as\n        ActivityPub.\n        Much of the document was rewritten and reorganized under the long feedback\n        process of the Social Working Group.\n      <\/p><p>\n        ActivityPub has been shaped by the careful input of many members in the\n        <abbr title=\"World Wide Web Consortium\">W3C<\/abbr> Social Working Group.\n        ActivityPub especially owes a great debt to Amy Guy, who has done more\n        than anyone to map the ideas across the different Social Working Group\n        documents through her work on [<cite><a class=\"bibref\" href=\"#bib-Social-Web-Protocols\">Social-Web-Protocols<\/a><\/cite>].\n        Amy also laid out the foundations for a significant refactoring of the\n        ActivityPub spec while sprinting for four days with Christopher Allan\n        Webber.\n        These revisions lead to cleaner separation between the client to server\n        and server components, along with clarity about ActivityPub's relationship\n        to [<cite><a class=\"bibref\" href=\"#bib-LDN\">LDN<\/a><\/cite>], among many other improvements.\n        Special thanks also goes to Benjamin Goering for putting together the\n        implementation report template.\n        We also thank mray for producing the spectacular tutorial illustrations\n        (which are licensed under the same license as the rest of this\n        document).\n      <\/p><p>\n        Many people also helped ActivityPub along through careful review.\n        In particular, thanks to:\n        Aaron Parecki,\n        AJ Jordan,\n        Benjamin Goering,\n        Caleb Langeslag,\n        Elsa Balderrama,\n        elf Pavlik,\n        Eugen Rochko,\n        Erik Wilde,\n        Jason Robinson,\n        Manu Sporny,\n        Michael Vogel,\n        Mike Macgirvin,\n        nightpool,\n        Puck Meerburg,\n        Sandro Hawke,\n        Sarven Capadisli,\n        Tantek \u00c7elik, and\n        Yuri Volkov.\n      <\/p><p>\n        This document is dedicated to all citizens of planet Earth.\n        You deserve freedom of communication; we hope we have contributed in\n        some part, however small, towards that goal and right.\n      <\/p><\/section><section class=\"appendix\"><h2 id=\"d-references\">D. References<\/h2><section id=\"normative-references\"><h3 id=\"d-1-normative-references\">D.1 Normative references<\/h3><dl class=\"bibliography\"><dt id=\"bib-Activity-Vocabulary\">[Activity-Vocabulary]<\/dt><dd><a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/\"><cite>Activity Vocabulary<\/cite><\/a>. J. Snell. ActivityStreams Working Group. Editors Draft. URL: <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/\">https:\/\/www.w3.org\/TR\/activitystreams-vocabulary\/<\/a>\n<\/dd><dt id=\"bib-ActivityStreams\">[ActivityStreams]<\/dt><dd><a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/\"><cite>Activity Streams 2.0<\/cite><\/a>. J. Snell. ActivityStreams Working Group. Editors Draft. URL: <a href=\"https:\/\/www.w3.org\/TR\/activitystreams-core\/\">https:\/\/www.w3.org\/TR\/activitystreams-core\/<\/a>\n<\/dd><dt id=\"bib-JSON-LD\">[JSON-LD]<\/dt><dd><a href=\"https:\/\/www.w3.org\/TR\/json-ld\/\"><cite>JSON-LD 1.0<\/cite><\/a>. Manu Sporny; Gregg Kellogg; Markus Lanthaler. W3C. 16 January 2014. W3C Recommendation. URL: <a href=\"https:\/\/www.w3.org\/TR\/json-ld\/\">https:\/\/www.w3.org\/TR\/json-ld\/<\/a>\n<\/dd><dt id=\"bib-LDN\">[LDN]<\/dt><dd><a href=\"https:\/\/www.w3.org\/TR\/ldn\/\"><cite>Linked Data Notifications<\/cite><\/a>. Sarven Capadisli; Amy Guy. W3C. 2 May 2017. W3C Recommendation. URL: <a href=\"https:\/\/www.w3.org\/TR\/ldn\/\">https:\/\/www.w3.org\/TR\/ldn\/<\/a>\n<\/dd><dt id=\"bib-RFC2119\">[RFC2119]<\/dt><dd><a href=\"https:\/\/tools.ietf.org\/html\/rfc2119\"><cite>Key words for use in RFCs to Indicate Requirement Levels<\/cite><\/a>. S. Bradner. IETF. March 1997. Best Current Practice. URL: <a href=\"https:\/\/tools.ietf.org\/html\/rfc2119\">https:\/\/tools.ietf.org\/html\/rfc2119<\/a>\n<\/dd><dt id=\"bib-RFC7231\">[RFC7231]<\/dt><dd><a href=\"https:\/\/tools.ietf.org\/html\/rfc7231\"><cite>Hypertext Transfer Protocol (HTTP\/1.1): Semantics and Content<\/cite><\/a>. R. Fielding, Ed.; J. Reschke, Ed.. IETF. June 2014. Proposed Standard. URL: <a href=\"https:\/\/tools.ietf.org\/html\/rfc7231\">https:\/\/tools.ietf.org\/html\/rfc7231<\/a>\n<\/dd><dt id=\"bib-RFC7234\">[RFC7234]<\/dt><dd><a href=\"https:\/\/tools.ietf.org\/html\/rfc7234\"><cite>Hypertext Transfer Protocol (HTTP\/1.1): Caching<\/cite><\/a>. R. Fielding, Ed.; M. Nottingham, Ed.; J. Reschke, Ed.. IETF. June 2014. Proposed Standard. URL: <a href=\"https:\/\/tools.ietf.org\/html\/rfc7234\">https:\/\/tools.ietf.org\/html\/rfc7234<\/a>\n<\/dd><\/dl><\/section><section id=\"informative-references\"><h3 id=\"d-2-informative-references\">D.2 Informative references<\/h3><dl class=\"bibliography\"><dt id=\"bib-Micropub\">[Micropub]<\/dt><dd><a href=\"https:\/\/www.w3.org\/TR\/micropub\/\"><cite>Micropub<\/cite><\/a>. Aaron Parecki. W3C. 23 May 2017. W3C Recommendation. URL: <a href=\"https:\/\/www.w3.org\/TR\/micropub\/\">https:\/\/www.w3.org\/TR\/micropub\/<\/a>\n<\/dd><dt id=\"bib-RFC6749\">[RFC6749]<\/dt><dd><a href=\"https:\/\/tools.ietf.org\/html\/rfc6749\"><cite>The OAuth 2.0 Authorization Framework<\/cite><\/a>. D. Hardt, Ed.. IETF. October 2012. Proposed Standard. URL: <a href=\"https:\/\/tools.ietf.org\/html\/rfc6749\">https:\/\/tools.ietf.org\/html\/rfc6749<\/a>\n<\/dd><dt id=\"bib-RFC6750\">[RFC6750]<\/dt><dd><a href=\"https:\/\/tools.ietf.org\/html\/rfc6750\"><cite>The OAuth 2.0 Authorization Framework: Bearer Token Usage<\/cite><\/a>. M. Jones; D. Hardt. IETF. October 2012. Proposed Standard. URL: <a href=\"https:\/\/tools.ietf.org\/html\/rfc6750\">https:\/\/tools.ietf.org\/html\/rfc6750<\/a>\n<\/dd><dd><a href=\"https:\/\/www.w3.org\/TR\/social-web-protocols\/\"><cite>Social Web Protocols<\/cite><\/a>. Amy Guy. W3C. 25 December 2017. W3C Note. URL: <a href=\"https:\/\/www.w3.org\/TR\/social-web-protocols\/\">https:\/\/www.w3.org\/TR\/social-web-protocols\/<\/a>\n<\/dd><\/dl><\/section><\/section><p role=\"navigation\" id=\"back-to-top\"><a href=\"#toc\"><abbr title=\"Back to Top\">\u2191<\/abbr><\/a><\/p>","created_at":"2018-11-07T16:22:40+0000","updated_at":"2018-11-07T16:22:40+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":47,"domain_name":"www.w3.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":94,"title":"C* Music Player","url":"https:\/\/cmus.github.io\/","content":"<h2 id=\"home\">cmus is a small, <em>fast<\/em> and powerful console music player for Unix-like operating systems.<a href=\"https:\/\/cmus.github.io\/cmus-2.4.3-osx.png\">\n        <img class=\"screenshot\" alt=\"cmus screenshot\" title=\"cmus' tree view\" src=\"https:\/\/cmus.github.io\/cmus-2.4.3-osx.png\" \/><\/a>Features<\/h2><div class=\"chapter\"><h3>Plugins<\/h3><ul><li><b>Input<\/b>: Ogg Vorbis, MP3, FLAC, Opus, Musepack, WavPack, WAV, AAC, MP4, audio CD,\n          and everything supported by ffmpeg (WMA, APE, MKA, TTA, SHN, ...) and\n          libmodplug<\/li>\n          <li><b>Output<\/b>: PulseAudio, ALSA, OSS, JACK, RoarAudio, CoreAudio, libao, sndio (OpenBSD), aRts, Sun, and\n          WaveOut (Windows)<\/li>\n        <\/ul><h3>Playing<\/h3><ul><li>Gapless playback<\/li>\n          <li>ReplayGain support<\/li>\n          <li>MP3 and Ogg streaming (SHOUTcast\/Icecast)<\/li>\n          <li>Play queue<\/li>\n          <li>Optional playback resume on startup<\/li>\n        <\/ul><h3>Interface<\/h3><ul><li>Instant startup, even with thousands of tracks<\/li>\n          <li>Powerful playlist filters \/ live filtering<\/li>\n          <li>Easy to use directory browser<\/li>\n          <li>Customizable colors<\/li>\n          <li>Dynamic keybindings. You can bind a key to any command, <code>:seek +1m<\/code> for example<\/li>\n          <li>Vi \/ less style search mode<\/li>\n          <li>Vi style command mode with tab completion<\/li>\n        <\/ul><h3>Misc<\/h3><ul><li>CUE sheets support<\/li>\n          <li>Excellent compilations handling<\/li>\n          <li>Uses Unicode internally for all string handling<\/li>\n          <li>Supports audio scrobbling (to e.g. <a href=\"http:\/\/last.fm\">Last.fm<\/a> or\n              <a href=\"http:\/\/libre.fm\">Libre.fm<\/a>) etc. via\n              <a href=\"https:\/\/github.com\/cmus\/cmus\/wiki\/status-display-programs\">\n              status display programs<\/a>\n          <\/li>\n          <li>Can run external commands for the currently selected files (tag-editor for example)<\/li>\n          <li>Can be controlled via UNIX socket using <code>cmus-remote<\/code> command<\/li>\n          <li>Known to work on Linux, OS X, FreeBSD, NetBSD, OpenBSD and Cygwin<\/li>\n        <\/ul><\/div><h2 id=\"documentation\">Documentation<\/h2><div class=\"chapter\"><h3>Installation<\/h3>In case your favourite package manager provides an up to date version of\n      cmus (see<a href=\"https:\/\/github.com\/cmus\/cmus\/wiki\/cmus-packages\">list of cmus packages<\/a>),\n      installation is straightforward, e.g.:<ul><li><b>Ubuntu<\/b>: <code>sudo apt-get install cmus<\/code><\/li>\n        <li><b>OS X<\/b>: <code>brew install cmus<\/code><\/li>\n      <\/ul>Otherwise, please consider compiling from source \u2014 it is not hard at all.<h3>Usage<\/h3>cmus comes with a great reference manual. If you have it installed, try<br \/><a href=\"https:\/\/github.com\/cmus\/cmus\/blob\/master\/Doc\/cmus-tutorial.txt\"><code class=\"man\">man cmus-tutorial<\/code><\/a>,<a href=\"https:\/\/github.com\/cmus\/cmus\/blob\/master\/Doc\/cmus.txt\"><code class=\"man\">man cmus<\/code><\/a>and<a href=\"https:\/\/github.com\/cmus\/cmus\/blob\/master\/Doc\/cmus-remote.txt\"><code class=\"man\">man cmus-remote<\/code><\/a>.<h3>Online Documentation<\/h3><ul><li><a href=\"https:\/\/github.com\/cmus\/cmus\/wiki\/\">cmus wiki<\/a><\/li>\n        <li><a href=\"https:\/\/github.com\/cmus\/cmus\/wiki\/reviews\">Articles about cmus<\/a> (part of the wiki)<\/li>\n      <\/ul><\/div><h2 id=\"download\">Download<\/h2><div class=\"chapter\"><h3>Stable Releases<\/h3><ul><li><a href=\"https:\/\/github.com\/cmus\/cmus\/archive\/v2.7.1.tar.gz\">cmus-2.7.1.tar.gz<\/a>, released 13.07.2015<br \/>Cmus has gained support for conditionals in format strings, sndio\n\t  output plugin (OpenBSD), VTX input plugin (ZX Spectrum, Atari,\n\t  Amstrad CPC), and optional mouse support.\n\t  More detailed list of changes can be found in the\n\t  <a href=\"https:\/\/github.com\/cmus\/cmus\/releases\/tag\/v2.7.0\">release notes<\/a>.<\/li>\n        <li><a href=\"https:\/\/github.com\/cmus\/cmus\/archive\/v2.6.2.tar.gz\">cmus-2.6.2.tar.gz<\/a>, released 29.06.2015<br \/>Major new features are <a href=\"https:\/\/i.imgur.com\/DdCUqCy.png\">display of all tracks when artist is selected in tree view<\/a>,\n\t  JACK output and Opus input plugins. Please see\n\t  <a href=\"https:\/\/github.com\/cmus\/cmus\/releases\/tag\/v2.6.0\">release notes<\/a> for a more detailed list of changes since 2.5.x.<\/li>\n        <li><a href=\"https:\/\/github.com\/cmus\/cmus\/archive\/v2.5.1.tar.gz\">cmus-2.5.1.tar.gz<\/a>, released 14.05.2014<br \/>Since 2.4.x., cmus has gained CUE sheets support, cdio input plugin, new\n          \u00abzenburn\u00bb color scheme, as well as tons of minor enhancements and\n          new features \u2014 detailed release notes can be found\n\t  <a href=\"http:\/\/article.gmane.org\/gmane.comp.audio.cmus.devel\/799\">here<\/a>.<\/li>\n        <li><a href=\"https:\/\/sourceforge.net\/projects\/cmus\/files\/cmus-v2.4.3.tar.bz2\/download\">cmus-v2.4.3.tar.bz2<\/a>, released 30.11.2011<br \/>Compared to 2.3.x series, this release features tons of\n\t  improvements, with most notable ones being live library\/playlist\n\t  filtering and playback resume support \u2014 please see the full release\n\t  notes <a href=\"https:\/\/sourceforge.net\/mailarchive\/message.php?msg_id=27403242\">here<\/a>.<\/li>\n        <li><a href=\"https:\/\/sourceforge.net\/projects\/cmus\/files\/cmus-v2.3.5.tar.bz2\/download\">cmus-v2.3.5.tar.bz2<\/a>, released 10.04.2011<br \/>Among other things, this release (compared to 2.2.0) features gapless MP3 playback,\n\t  native PulseAudio support, faster startup, and much improved\n\t  compilations handling.<\/li>\n        <li><a href=\"https:\/\/sourceforge.net\/projects\/cmus\/files\/cmus-v2.2.0.tar.bz2\/download\">cmus-v2.2.0.tar.bz2<\/a>, released 27.07.2007<br \/>When trees were tall, and grass was green...<\/li>\n      <\/ul><p>More information on older releases can be found on <a href=\"http:\/\/freshmeat.net\/projects\/cmus\/releases\">cmus' freshmeat page<\/a>.<\/p><h3>Bleeding Edge<\/h3><p>If you want to help cmus development and be among the first to try the latest and greatest features, use the <em>pu<\/em> (proposed updates) branch, e.g.:<\/p><pre class=\"block\">git clone -b pu <a href=\"https:\/\/github.com\/cmus\/cmus\">https:\/\/github.com\/cmus\/cmus.git<\/a>\ncd cmus\n.\/configure prefix=$HOME\/cmus\nmake install<\/pre><p>You may want to add the installation directory to your PATH, e.g.:<\/p><pre class=\"block\">echo 'export PATH=$HOME\/cmus\/bin:$PATH' &gt;&gt; ~\/.bashrc ; . ~\/.bashrc<\/pre><p>Since the branch is recreated, you need to use this command to update the source tree (instead of git pull):<\/p><pre class=\"block\">git remote update &amp;&amp; git reset --hard origin\/pu<\/pre><p>Don't hesitate to submit an <a href=\"https:\/\/github.com\/cmus\/cmus\/issues\">issue<\/a> if you have any problems, suggestions or bug reports.<\/p><\/div><h2>Development<\/h2><div class=\"chapter\"><h3>GitHub page<\/h3><p>\n        <a href=\"https:\/\/github.com\/cmus\/cmus\">https:\/\/github.com\/cmus\/cmus<\/a>\n      <\/p><h3>Mailing List<\/h3><p>The mailing list now serves as an archive for old releases and issues. Please use the github <a href=\"https:\/\/github.com\/cmus\/cmus\/issues\">issues<\/a> page for any problems, suggestions, or bug reports.<\/p><p>\n        <a href=\"https:\/\/lists.sourceforge.net\/lists\/listinfo\/cmus-devel\">cmus-devel@lists.sourceforge.net<\/a>\n      <\/p><p>\n        <a href=\"http:\/\/dir.gmane.org\/gmane.comp.audio.cmus.devel\">Gmane mirror<\/a>\n      <\/p><h3>Dependencies<\/h3><ul><li><a href=\"http:\/\/www.gnu.org\/software\/ncurses\/ncurses.html\">ncurses<\/a> (ncursesw recommended)<\/li>\n      <\/ul><h3>Optional Dependencies<\/h3><ul><li><a href=\"http:\/\/pulseaudio.org\/\">libpulse<\/a> (PulseAudio)<\/li>\n        <li><a href=\"http:\/\/www.alsa-project.org\/\">alsa-lib<\/a> (ALSA)<\/li>\n        <li><a href=\"http:\/\/www.opensound.com\/oss.html\">OSS<\/a> (OSS)<\/li>\n        <li><a href=\"http:\/\/roaraudio.keep-cool.org\/\">libroar<\/a> (RoarAudio)<\/li>\n        <li><a href=\"http:\/\/www.xiph.org\/ao\/\">libao<\/a> (libao)<\/li>\n        <li><a href=\"http:\/\/www.arts-project.org\/\">aRts<\/a> (aRts)<\/li>\n        <li><a href=\"http:\/\/www.gnu.org\/software\/libcdio\/\">libcdio<\/a> (cdio plugin)<\/li>\n        <li><a href=\"http:\/\/libcddb.sourceforge.net\/\">libcddb<\/a> (cdio plugin)<\/li>\n        <li><a href=\"http:\/\/flac.sourceforge.net\/\">FLAC<\/a> (.flac)<\/li>\n        <li><a href=\"https:\/\/www.opus-codec.org\/\">Opus<\/a> (.opus)<\/li>\n        <li><a href=\"http:\/\/www.vorbis.com\/\">libvorbis<\/a> (.ogg)<\/li>\n        <li><a href=\"http:\/\/www.musepack.net\/\">libmpcdec<\/a> (.mpc, .mpp, .mp+)<\/li>\n        <li><a href=\"http:\/\/www.wavpack.com\/\">wavpack<\/a> (.wv)<\/li>\n        <li><a href=\"http:\/\/www.underbit.com\/products\/mad\/\">libmad<\/a> (.mp3)<\/li>\n        <li><a href=\"http:\/\/ffmpeg.org\/\">libavcodec<\/a> (.wma, .ape, .mka, .tta, .shn, ...)<\/li>\n        <li><a href=\"http:\/\/www.audiocoding.com\/\">FAAD<\/a> (.aac)<\/li>\n        <li><a href=\"http:\/\/mpeg4ip.sourceforge.net\/\">mp4v2, FAAD<\/a> (.mp4, .m4a, .m4b)<\/li>\n        <li><a href=\"http:\/\/modplug-xmms.sourceforge.net\/\">libmodplug<\/a> (.mod, .s3m, ...)<\/li>\n        <li><a href=\"http:\/\/mikmod.raphnet.net\/\">libmikmod<\/a> (.mod, .s3m, ...)<\/li>\n      <\/ul><\/div>","created_at":"2018-11-07T16:22:33+0000","updated_at":"2018-11-07T16:22:33+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"cmus.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":95,"title":"No title found","url":"http:\/\/oysttyer.github.io\/docs\/userGuide.html","content":"<pre>     .#*^#=.                \n     %'.,`.#`               \n   ;',. .\/#`                                                   \n  ({.`,` #\/                 \n  `&amp; ,` %,~=*'\"*=~=-.,      \n   \\`=_\/'.``  -  `'.  *\\.                                      \n    (%.  -  -      \u02cb-. `&amp;   \n    `&amp;`  ~     @      . #                                       \n     `\\`. `    .....\u02ca  %'   \n       `^~._.,,,.-+=~*'                                            \n            ````'''         \n  _      _             _     \n \/.\\ | ||_  |_ |_ | | \/_\\|'`                                 \n \\_\/ \\_| _|'|_'|_ \\_| \\_ |                                   \n       |            |       <\/pre><ul><li><a href=\"#start\">Getting started<\/a>\n<ul><li><a href=\"#getcode\">Getting the code<\/a><\/li>\n<li><a href=\"#apikey\">Generating your API key<\/a><\/li>\n<li><a href=\"#switching\">Switching from TTYtter<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#usage\">Basic usage<\/a>\n<ul><li><a href=\"#commands\">Commands<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#command-line\">Command Line Options<\/a><\/li>\n<li><a href=\"#configuration\">Configuration<\/a><\/li>\n<li><a href=\"#contributing\">Feedback and contributing<\/a><\/li>\n<\/ul><h2 id=\"start\">\nGetting started\n<\/h2>\n<p>Welcome to <a href=\"http:\/\/oysttyer.github.io\/\">oysttyer<\/a>, a console-based Twitter client written in Perl. This guide describes the basic use of oysttyer. If anything is unclear, incorrect, or missing, please <a href=\"#contributing\">let us know<\/a>.<\/p>\n<h3 id=\"getcode\">\nGetting the code\n<\/h3>\n<p>oysttyer is a single Perl script. You will, of course, need a Perl interpreter and also cURL (included by default on most Linux and Unix distributions). You may also want <a href=\"http:\/\/search.cpan.org\/~ckaiser\/Term-ReadLine-TTYtter-1.4\/\">Term::Readline::TTYtter<\/a> in order to take advantage of some of the more colorful things oysttyer can do.<\/p>\n<p>There are two options for getting oysttyer: a \u201cstable\u201d release version and the latest development version. In practice, there is not much difference between the two since new releases happen quickly after bugs are fixed or features are added and development changes are (we hope!) not going to break anything.<\/p>\n<p>To get the stable release, go to the <a href=\"https:\/\/github.com\/oysttyer\/oysttyer\">GitHub repository<\/a> and click \u201cReleases\u201d download the most recent release and extract it.<\/p>\n<p>To get the latest development version oysttyer, you can clone the <a href=\"https:\/\/github.com\/oysttyer\/oysttyer\">GitHub repository<\/a> or simply download the raw oysttyer.pl file.<\/p>\n<h3 id=\"apikey\">\nGenerating your API key\n<\/h3>\n<p>Before you can get started, you should generate an API key. The key built in to oysttyer is frequently blocked by Twitter, so using your own key is adviseable.<\/p>\n<ol type=\"1\"><li>Log in to the <a href=\"https:\/\/apps.twitter.com\/\">Twitter Application Management page<\/a> with your Twitter user name and password<\/li>\n<li>Click <em>Create New App<\/em><\/li>\n<li>Fill out the new app form:\n<ul><li>Name \u2013 The globally unique name for your application. We suggest something like \u201coysttyer (YourCoolTwitterHandle)\u201d<\/li>\n<li>Description \u2013 Please use \u201cAn interactive console text-based command-line Twitter client written in Perl\u201d<\/li>\n<li>Website \u2013 Please use \u201chttp:\/\/oysttyer.github.io\/\u201d<\/li>\n<li>Callback URL \u2013 (leave blank)<\/li>\n<\/ul><\/li>\n<li>Check the box to agree to the developer agreement and click <em>Create Your Twitter Application<\/em><\/li>\n<li>Update your application to be able to send and receive direct messages\n<ul><li>Click the <em>Permissions<\/em> tab<\/li>\n<li>Select <em>Read, Write and Access direct messages<\/em><\/li>\n<li>Click <em>Update application<\/em><\/li>\n<\/ul><\/li>\n<li>Get your keys\n<ul><li>Click the <em>Key and Access Tokens<\/em> tab<\/li>\n<li>Note the <em>Consumer Key (API Key)<\/em> and <em>Consumer Secret (API Secret)<\/em>.<\/li>\n<\/ul><\/li>\n<\/ol><p>Now that you have your keys, you can authorize oysttyer. Your key and secret can be passed on the command line or in your <a href=\"#configuration\">configuration file<\/a>. An example of using it on the command line is:<\/p>\n<pre>perl .\/oysttyer.pl -oauthkey=nZdXCxVtQTsFJaybAr9OjApd0 -oauthsecret=ujN4ilao663PyDDgObkk0N6EIbAxxxG7MGoK2Yq7nzLLfOxZjE<\/pre>\n<p>Similarly, the configuration file settings are <code>oauthkey<\/code> and <code>oauthsecret<\/code> for the Consumer Key and Consumer Secret respectively.<\/p>\n<p>A third option is to generate your token and secret from the Application Management website and write that directly to the <code>.oysttyerkey<\/code> file:<\/p>\n<pre>ck=X&amp;cs=X&amp;at=YOUR_ACCESS_TOKEN&amp;ats=YOUR_ACCESS_TOKEN_SECRET<\/pre>\n<h3 id=\"switching\">\nSwitching from TTYTter\n<\/h3>\n<p>If you previously used TTYtter, you should find oysttyer very familiar. A few steps are necessary to make the switch:<\/p>\n<ol type=\"1\"><li>You have to re-authorise (you can\u2019t use your .ttytterkey) as we have a new API key<\/li>\n<li>Move\/rename your <code>.ttytterc<\/code> file to <code>.oysttyerrc<\/code><\/li>\n<li>If you use the ttytteristas pref it is now called oysttyeristas<\/li>\n<li>Read the <a href=\"https:\/\/github.com\/oysttyer\/oysttyer\/blob\/master\/CHANGELOG.markdown\">Changelog<\/a> to see what\u2019s new since TTYtter 2.1<\/li>\n<\/ol><h2 id=\"usage\">\nBasic usage\n<\/h2>\n<p>To start using oysttyer, run <code>perl .\/oysttyer.pl<\/code> in the directory where you\u2019ve downloaded osytter.pl. Soon tweets start spewing forth. Each tweet is displayed with an menu code (internal to oysttyer) and the username. For example:<\/p>\n<pre>h8&gt; &lt;FunnelFiasco&gt; I just really like writing documentation.<\/pre>\n<p>Menu codes are typically a letter followed by a single-digit number, except direct messages which prepend a <code>d<\/code> and search results (including threads, <code>\/again<\/code>, etc) which prepend a <code>z<\/code>.<\/p>\n<p>The username may also be prepended with special characters that give you more information about a tweet.<\/p>\n<table id=\"decorators\"><tr><th>\nDecorator\n<\/th>\n<th>\nMeaning\n<\/th>\n<\/tr><tr><td>\n@\n<\/td>\n<td>\nTweet is in reply to another tweet\n<\/td>\n<\/tr><tr><td>\n\u201c\n<\/td>\n<td>\nTweet quotes the tweet below\n<\/td>\n<\/tr><tr><td>\n\u2191\n<\/td>\n<td>\nTweet was quoted by the tweet above&lt;\/td\n<\/td><\/tr><tr><td>\n%\n<\/td>\n<td>\nTweet was retweeted into your timeline\n<\/td>\n<\/tr><tr><td>\n+\n<\/td>\n<td>\nTweet has location information\n<\/td>\n<\/tr><\/table><p>Ready to tweet? Just type your tweet and hit enter. But simply broadcasting is no fun (even if it\u2019s your personal #brand). Interaction is what makes twitter enjoyable. The <a href=\"#commands\">commands<\/a> section below describes how you can interact.<\/p>\n<h3 id=\"commands\">\nCommands\n<\/h3>\n<p>oysttyer provides many commands, all of which begin with <code>\/<\/code>. For the sake of your followers, anything that starts with <code>\/<\/code> is treated as a command and oysttyer will let you know if it doesn\u2019t recognize it. If you want to start a tweet with <code>\/<\/code>, double it (<code>\/\/<\/code>). The tables below describe most of the oysttyer commands, in hopefully-helpful sections:<\/p>\n<ul><li><a href=\"#commands-oysttyer\">oysttyer commands<\/a><\/li>\n<li><a href=\"#commands-tweet\">tweet commands<\/a><\/li>\n<li><a href=\"#commands-user\">user commands<\/a><\/li>\n<li><a href=\"#commands-list\">list commands<\/a><\/li>\n<\/ul><p>Commands that return multiple responses default to 20, but that can be adjusted by specifying +<em>N<\/em> after the command name. For example:<\/p>\n<pre>\/dmagain +5<\/pre>\n<p>will only give you your last 5 direct messages, instead of the last 20.<\/p>\n<h4 id=\"commands-oysttyer\">\noysttyer commands\n<\/h4>\n<table><tr><th>\nCommand\n<\/th>\n<th>\nDescription\n<\/th>\n<\/tr><tr><td>\n\/! <em>shell command<\/em>\n<\/td>\n<td>\nRun <em>shell command<\/em> in a subprocess\n<\/td>\n<\/tr><tr><td>\n\/add <em>key<\/em> <em>value<\/em>\n<\/td>\n<td>\nAdd <em>value<\/em> to <em>key<\/em> (see the <a href=\"#runtime\">runtime configuration section<\/a> for more information)\n<\/td>\n<\/tr><tr><td>\n\/cls\n<\/td>\n<td>\nClear the screen\n<\/td>\n<\/tr><tr><td>\n\/del <em>key<\/em> <em>value<\/em>\n<\/td>\n<td>\nRemove <em>value<\/em> from <em>key<\/em> (see the <a href=\"#runtime\">runtime configuration section<\/a> for more information)\n<\/td>\n<\/tr><tr><td>\n\/help\n<\/td>\n<td>\nDisplay help\n<\/td>\n<\/tr><tr><td>\n\/history\n<\/td>\n<td>\nDisplay your command history\n<\/td>\n<\/tr><tr><td>\n\/pop <em>key<\/em>\n<\/td>\n<td>\nPop the first value off of <em>key<\/em> (see the <a href=\"#runtime\">runtime configuration section<\/a> for more information)\n<\/td>\n<\/tr><tr><td>\n\/print <em>key<\/em>\n<\/td>\n<td>\nDisplay all settings (or the value of <em>key<\/em>, if specified)\n<\/td>\n<\/tr><tr><td>\n\/push <em>key<\/em> <em>value<\/em>\n<\/td>\n<td>\nPush <em>value<\/em> onto the stack for <em>key<\/em> (see the <a href=\"#runtime\">runtime configuration section<\/a> for more information)\n<\/td>\n<\/tr><tr><td>\n\/ruler\n<\/td>\n<td>\nPrint a \u201cruler\u201d of 140 characters\n<\/td>\n<\/tr><tr><td>\n\/set <em>key<\/em> <em>value<\/em>\n<\/td>\n<td>\nSet <em>key<\/em> to <em>value<\/em> (see the <a href=\"#runtime\">runtime configuration section<\/a> for more information)\n<\/td>\n<\/tr><tr><td>\n\/unset <em>key<\/em>\n<\/td>\n<td>\nUnsets <em>key<\/em> (see the <a href=\"#runtime\">runtime configuration section<\/a> for more information)\n<\/td>\n<\/tr><tr><td>\n\/vcheck\n<\/td>\n<td>\nCheck to see if you\u2019re running the latest version\n<\/td>\n<\/tr><tr><td>\n\/quit\n<\/td>\n<td>\nExit oysttyer\n<\/td>\n<\/tr><\/table><table><tr><th>\nCommand\n<\/th>\n<th>\nDescription\n<\/th>\n<\/tr><tr><td>\n\/re <em>menu code<\/em>\n<\/td>\n<td>\nReply to tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/ra <em>menu code<\/em>\n<\/td>\n<td>\nReply to all users in tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/rt <em>menu code<\/em> <em>message<\/em>\n<\/td>\n<td>\nRetweet tweet <em>menu code<\/em>. If a message is specified, quote tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><\/tr><tr><td>\n\/like <em>menu code<\/em>\n<\/td>\n<td>\nLike tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/unlike <em>menu code<\/em>\n<\/td>\n<td>\nRemove like for tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/lrt <em>menu code<\/em>\n<\/td>\n<td>\nLike and retweet tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/thread <em>menu code<\/em>\n<\/td>\n<td>\nDisplays the thread for tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/again <em>username<\/em>\n<\/td>\n<td>\nDisplays the recent tweets for <em>username<\/em> (if none is specified, shows recent tweets in your timeline again)\n<\/td>\n<\/tr><tr><td>\n\/deletelast\n<\/td>\n<td>\nDeletes your last tweet\n<\/td>\n<\/tr><tr><td>\n\/delete <em>menu code<\/em>\n<\/td>\n<td>\nDeletes tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/dump <em>menu code<\/em>\n<\/td>\n<td>\nDisplay all data for tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/dm <em>username<\/em> <em>message<\/em>\n<\/td>\n<td>\nSend a direct message to <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/dmsent\n<\/td>\n<td>\nDisplay direct messages you have sent\n<\/td>\n<\/tr><tr><td>\n\/dmagain\n<\/td>\n<td>\nDisplay direct messages sent you you\n<\/td>\n<\/tr><tr><td>\n\/dmrefresh\n<\/td>\n<td>\nCheck for new direct messages\n<\/td>\n<\/tr><tr><td>\n\/entities <em>menu code<\/em>\n<\/td>\n<td>\nDisplay the expanded t.co links in tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/search <em>keyword<\/em>\n<\/td>\n<td>\nSearch public tweets for <em>keyword<\/em>\n<\/td>\n<\/tr><tr><td>\n\/refresh\n<\/td>\n<td>\nCheck for new tweets\n<\/td>\n<\/tr><tr><td>\n\/rtsof <em>menu code<\/em>\n<\/td>\n<td>\nList accounts that retweeted tweet <em>menu code<\/em>\n<\/td>\n<\/tr><tr><td>\n\/rtsofme\n<\/td>\n<td>\nList your tweets that others have retweeted\n<\/td>\n<\/tr><tr><td>\n\/track <em>keywords<\/em>\n<\/td>\n<td>\nSearch for <em>keywords<\/em> in addition to your normal timeline\n<\/td>\n<\/tr><tr><td>\n\/tron <em>keywords<\/em>\n<\/td>\n<td>\nAdd additional keywords to <code>\/track<\/code>\n<\/td>\n<\/tr><tr><td>\n\/troff <em>keywords<\/em>\n<\/td>\n<td>\nStop <code>\/track<\/code>ing <em>keywords<\/em>\n<\/td>\n<\/tr><tr><td>\n\/trends\n<\/td>\n<td>\nList the current trending topics\n<\/td>\n<\/tr><tr><td>\n\/url <em>menu code<\/em>\n<\/td>\n<td>\nDisplay the URLs in tweet <em>menu code<\/em> and open in the application defined by the <code>urlopen<\/code> option\n<\/td>\n<\/tr><tr><td>\n\/web <em>menu code<\/em>\n<\/td>\n<td>\nDisplay the URL to tweet <em>menu code<\/em>\n<\/td>\n<\/tr><\/table><h4 id=\"commands-user\">\nUser commands\n<\/h4>\n<table><tr><th>\nCommand\n<\/th>\n<th>\nDescription\n<\/th>\n<\/tr><tr><td>\n\/follow <em>username<\/em>\n<\/td><td>\nStart following <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/unfollow <em>username<\/em>\n<\/td>\n<td>\nStop following <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/leave <em>username<\/em>\n<\/td>\n<td>\nAlias for <code>\/unfollow<\/code>\n<\/td>\n<\/tr><tr><td>\n\/block <em>username<\/em>\n<\/td>\n<td>\nBlock <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/unblock <em>username<\/em>\n<\/td>\n<td>\nUn-block <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/mute <em>username<\/em>\n<\/td>\n<td>\nMute <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/unmute <em>username<\/em>\n<\/td>\n<td>\nUn-mute <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/whois <em>username<\/em>\n<\/td>\n<td>\nDisplay information for <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/doesfollow <em>username<\/em> <em>seconduser<\/em>\n<\/td>\n<td>\nCheck if a user follows another (if a second account is not specified, it checks if the user follows you)\n<\/td>\n<\/tr><tr><td>\n\/friends <em>username<\/em>\n<\/td>\n<td>\nGet a list of the user\u2019s friends (if an account is not specified, it lists your friends)\n<\/td>\n<\/tr><tr><td>\n\/followers <em>username<\/em>\n<\/td>\n<td>\nGet a list of the user\u2019s followers (if an account is not specified, it lists your followers)\n<\/td>\n<\/tr><\/table><h4 id=\"commands-list\">\nList commands\n<\/h4>\n<table><tr><td>\n\/withlist <em>listname<\/em> <b>create<\/b> <b>(private|public)<\/b> <em>text<\/em>\n<\/td><td>\nCreate your own list <em>listname<\/em>, and make it either private or public, (<em>text<\/em> : an optional description of the created list)\n<\/td>\n<\/tr><tr><td>\n\/withlist <em>listname<\/em> <b>add<\/b> <em>username<\/em>\n<\/td><td>\nAdd user <em>username<\/em> to your own list <em>listname<\/em>\n<\/td>\n<\/tr><tr><td>\n\/withlist <em>listname<\/em> <b>delete<\/b> <em>username<\/em>\n<\/td><td>\nAdd user <em>username<\/em> to your own list <em>listname<\/em>\n<\/td>\n<\/tr><tr><td>\n\/withlist <em>listname<\/em> <b>list<\/b>\n<\/td><td>\nShow who\u2019s on your list <em>listname<\/em>\n<\/td>\n<\/tr><tr><td>\n\/withlis, \/withl, \/with, \/wl\n<\/td><td>\nAliases for <code>\/withlist<\/code>\n<\/td>\n<\/tr><tr><td>\n\/lists\n<\/td><td>\nDisplays your lists\n<\/td>\n<\/tr><tr><td>\n\/lists <em>username<\/em>\n<\/td><td>\nDisplays lists of the specified user <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/lists <em>username\/listname<\/em>\n<\/td><td>\nDisplays who\u2019s on the specified list of the specified user <em>username<\/em>\n<\/td>\n<\/tr><tr><td>\n\/lfollow <em>listname<\/em>\n<\/td><td>\nStart following <em>listname<\/em> This only marks you as a follower, it does not actually add the list to your timeline\n<\/td>\n<\/tr><tr><td>\n\/lleave <em>listname<\/em>\n<\/td><td>\nStop following <em>listname<\/em>\n<\/td>\n<\/tr><tr><td>\n\/liston <em>listname<\/em>\n<\/td><td>\nSets lists that are automatically queried and mixed with your timeline (You do not have to be following a list to add it to your timeline)\n<\/td>\n<\/tr><tr><td>\n\/listoff <em>listname<\/em>\n<\/td><td>\nTurn off showing tweets from that list in your timeline\n<\/td>\n<\/tr><tr><td>\n\/followers <em>listname<\/em>\n<\/td>\n<td>\nDisplays users who are marked as following the specified list\n<\/td>\n<\/tr><\/table><h2 id=\"command-line\">\nCommand Line Options\n<\/h2>\n<p>Oysttyer inherits many commands from ttytter. The full list of commands supported by ttytter is <a href=\"http:\/\/www.floodgap.com\/software\/ttytter\/copts.html\">documented here<\/a> but some of the more useful ones are:<\/p>\n<table><tr><th>\nCommand\n<\/th>\n<th>\nDescription\n<\/th>\n<th>\nExample\n<\/th>\n<\/tr><tr><td>\n-rc\n<\/td>\n<td>\nLocation of config file\n<\/td>\n<td>\n-keyf=oysttyerrc\n<\/td>\n<\/tr><tr><td>\n-keyf\n<\/td>\n<td>\nLocation of Key file\n<\/td>\n<td>\n-keyf=mykeyfile\n<\/td>\n<\/tr><tr><td>\n-status\n<\/td>\n<td>\nTweet directly from the command line\n<\/td>\n<td>\n-status=\u201cI\u2019m on a boat\u201d\n<\/td>\n<\/tr><tr><td>\n-verbose\n<\/td>\n<td>\nGive more debugging information\n<\/td>\n<td>\n-verbose\n<\/td>\n<\/tr><tr><td>\n-location\n<\/td>\n<td>\nSet your location - Note: you must enable tweet locations on the web interface\n<\/td>\n<td>\n-location\n<\/td>\n<\/tr><tr><td>\n-lat\n<\/td>\n<td>\nlatitude in decimal\n<\/td>\n<td>\n-lat=-43.55522919\n<\/td>\n<\/tr><tr><td>\n-long\n<\/td>\n<td>\nlongitude in decimal\n<\/td>\n<td>\n-long=172.4193726\n<\/td>\n<\/tr><tr><td>\n-script\n<\/td>\n<td>\nForces mode designed for being called via a script file, with minimal output\n<\/td>\n<td>\n-script\n<\/td>\n<\/tr><tr><td>\n-separator <em>string<\/em>\n<\/td>\n<td>\nSeparate refreshes by <em>string<\/em> to mark your place\n<\/td>\n<td>\n-separator=\u201cX\u2014x\u2014X\u2014x\u2014X\u201d\n<\/td>\n<\/tr><tr><td>\n-doublespace\n<\/td>\n<td>\nPrint an extra space after each tweet\n<\/td>\n<td>\n-doublespace\n<\/td>\n<\/tr><\/table><h2 id=\"configuration\">\nConfiguration\n<\/h2>\n<p>You can set many options using command line options:<\/p>\n<pre>$ perl .\/oysttyer -ansi=1 -newline=1 -vcheck<\/pre>\n<p>but these can also be set in a configuration file. The standard file is <code>~\/.oysttyerrc<\/code>, but another one can be chosen using the <code>-f<\/code> command line option.<\/p>\n<p>The same options above can be written in the configuration file like this<\/p>\n<pre>ansi=1\nnewline=1\nvcheck=1<\/pre>\n<p>Use <code>#<\/code> to denote a comment.<\/p>\n<p>For security and portability, it\u2019s recommended that your Twitter credentials be kept in a separate file, <code>~\/.oysttyerkey<\/code>. This file should be <code>chmod 600<\/code> (user read+write).<\/p>\n<h2 id=\"contributing\">\nContributing\n<\/h2>\n<p>Development is done via GitHub, the <a href=\"https:\/\/github.com\/oysttyer\/oysttyer\/blob\/master\/CONTRIBUTING.markdown\">CONTRIBUTING<\/a> document has details.<\/p>","created_at":"2018-11-07T16:22:33+0000","updated_at":"2018-11-07T16:22:33+0000","annotations":[],"mimetype":"text\/html","reading_time":8,"domain_name":"oysttyer.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":91,"title":"MOO home page","url":"https:\/\/www.moo.mud.org\/","content":"<h2>Overview<\/h2><p>LambdaMOO is both the name of a MUD and the name\nof MUD server software.   This page is about the server software, not\nthe particular MUD.<\/p><p>This site contains both this high level links page as well as the\n<a href=\"https:\/\/www.moo.mud.org\/mcp\/\">MCP home page<\/a>.\n<\/p><h2>Server source<\/h2><p>The <a href=\"http:\/\/sourceforge.net\/projects\/lambdamoo\/\">LambdaMOO sourceforge project<\/a> has the CVS\nrepository and <a href=\"http:\/\/sourceforge.net\/project\/showfiles.php?group_id=3692\">releases<\/a> of LambdaMOO.<\/p><h2>Databases<\/h2><p>A MUD running under LambdaMOO is implemented mostly in its\ndatabase.   There are several starting points for building a MUD with\nLambdaMOO called \"cores\".  Here are a few commonly used ones:<\/p><ul><li>LambdaCore - available from <a href=\"ftp:\/\/ftp.lambda.moo.mud.org\/pub\/MOO\/\">ftp:\/\/ftp.lambda.moo.mud.org\/pub\/MOO\/<\/a><\/li>\n    <li>JHCore - available from <a href=\"http:\/\/jhcore.sourceforge.net\/\">http:\/\/jhcore.sourceforge.net\/<\/a><\/li>\n    <li>enCore - available from <a href=\"http:\/\/lingua.utdallas.edu\/encore\/\">http:\/\/lingua.utdallas.edu\/encore\/<\/a><\/li>\n  <\/ul><h2>Further information\nThe\n    <a href=\"ftp:\/\/ftp.research.att.com\/dist\/eostrom\/MOO\/html\/ProgrammersManual_toc.html\">\n    LambdaMOO Programmer's Manual<\/a> details server\n    functionality.\n  The <a href=\"https:\/\/www.moo.mud.org\/moo-faq\/\">MOO-Cows FAQ<\/a>\n    and the\n    <a href=\"http:\/\/www.the-b.org\/moo\/new-archwiz-faq.html\">Newbie\n    Archwiz FAQ<\/a> answer commonly asked questions about the\n    server.\n  The <a href=\"http:\/\/www.hayseed.net\/MOO\/\">Lost Library of\n    MOO<\/a> contains links to quite a few useful documents about\n    MOOs.\n<\/h2>","created_at":"2018-11-07T16:22:32+0000","updated_at":"2018-11-07T16:22:32+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.moo.mud.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":92,"title":"How to implement a basic ActivityPub server","url":"https:\/\/blog.joinmastodon.org\/2018\/06\/how-to-implement-a-basic-activitypub-server\/","content":"<p>Today we\u2019ll be looking at how to connect the protocols powering Mastodon in the simplest way possible to enter the federated network. We will use static files, standard command-line tools, and some simple Ruby scripting, although the functionality should be easily adaptable to other programming languages.<\/p><p>First, what\u2019s the end goal of this exercise? We want to send a Mastodon user a message from our own, non-Mastodon server.<\/p>\n<p>So what are the ingredients required? The message itself will be formatted with ActivityPub, and it must be attributed to an ActivityPub actor. The actor must be discoverable via Webfinger, and the delivery itself must be cryptographically signed by the actor.<\/p>\n<h3 id=\"the-actor\">The actor<\/h3>\n<p>The actor is a publicly accessible JSON-LD document answering the question \u201cwho\u201d. JSON-LD itself is a quite complicated beast, but luckily for our purposes we can treat it as simple JSON with a <code>@context<\/code> attribute. Here is what an actor document could look like:<\/p>\n<div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n\t\"@context\": [\n\t\t\"https:\/\/www.w3.org\/ns\/activitystreams\",\n\t\t\"https:\/\/w3id.org\/security\/v1\"\n\t],\n\t\"id\": \"https:\/\/my-example.com\/actor\",\n\t\"type\": \"Person\",\n\t\"preferredUsername\": \"alice\",\n\t\"inbox\": \"https:\/\/my-example.com\/inbox\",\n\t\"publicKey\": {\n\t\t\"id\": \"https:\/\/my-example.com\/actor#main-key\",\n\t\t\"owner\": \"https:\/\/my-example.com\/actor\",\n\t\t\"publicKeyPem\": \"-----BEGIN PUBLIC KEY-----...-----END PUBLIC KEY-----\"\n\t}\n}<\/pre><\/div>\n<p>The <code>id<\/code> must be the URL of the document (it\u2019s a self-reference), and all URLs should be using HTTPS. You need to include an <code>inbox<\/code> even if you don\u2019t plan on receiving messages in response, because for legacy purposes Mastodon doesn\u2019t acknowledge inbox-less actors as compatible.<\/p>\n<p>The most complicated part of this document is the <code>publicKey<\/code> as it involves cryptography. The <code>id<\/code> will in this case refer to the actor itself, with a fragment (the part after <code>#<\/code>) to identify it\u2013this is because we are not going to host the key in a separate document (although we could). The <code>owner<\/code> must be the actor\u2019s <code>id<\/code>. Now to the hard part: You\u2019ll need to generate an RSA keypair.<\/p>\n<p>You can do this using OpenSSL:<\/p>\n<pre>openssl genrsa -out private.pem 2048\nopenssl rsa -in private.pem -outform PEM -pubout -out public.pem\n<\/pre>\n<p>The contents of the <code>public.pem<\/code> file is what you would put into the <code>publicKeyPem<\/code> property. However, JSON does not support verbatim line-breaks in strings, so you would first need to replace line-breaks with <code>\\n<\/code> instead.<\/p>\n<h3 id=\"webfinger\">Webfinger<\/h3>\n<p>What is Webfinger? It is what allows us to ask a website, \u201cDo you have a user with this username?\u201d and receive resource links in response. Implementing this in our case is really simple, since we\u2019re not messing with any databases and can hardcode what we want.<\/p>\n<p>The Webfinger endpoint is always under <code>\/.well-known\/webfinger<\/code>, and it receives queries such as <code>\/.well-known\/webfinger?resource=acct:bob@my-example.com<\/code>. Well, in our case we can cheat, and just make it a static file:<\/p>\n<div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n\t\"subject\": \"acct:alice@my-example.com\",\n\t\"links\": [\n\t\t{\n\t\t\t\"rel\": \"self\",\n\t\t\t\"type\": \"application\/activity+json\",\n\t\t\t\"href\": \"https:\/\/my-example.com\/actor\"\n\t\t}\n\t]\n}<\/pre><\/div>\n<p>The <code>subject<\/code> property here consists of the username (same as <code>preferredUsername<\/code> earlier) and the domain you\u2019re hosting on. This is how your actor will be stored on other Mastodon servers and how people will be able to mention it in toots. Only one link is required in the Webfinger response, and it\u2019s the link to the actor document.<\/p>\n<p>After this is uploaded to your webhost and available under your domain with a valid SSL certificate, you could already look up your actor from another Mastodon by entering <code>alice@my-example.com<\/code> into the search bar. Although it\u2019ll look quite barren.<\/p>\n<h3 id=\"the-message\">The message<\/h3>\n<p>ActivityPub messages practically consist of two parts, the message itself (the object) and a wrapper that communicates what\u2019s happening with the message (the activity). In our case, it\u2019s going to be a <code>Create<\/code> activity. Let\u2019s say \u201cHello world\u201d in response to my toot about writing this blog post:<\/p>\n<p>Here is how the document could look:<\/p>\n<div class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n\t\"@context\": \"https:\/\/www.w3.org\/ns\/activitystreams\",\n\t\"id\": \"https:\/\/my-example.com\/create-hello-world\",\n\t\"type\": \"Create\",\n\t\"actor\": \"https:\/\/my-example.com\/actor\",\n\t\"object\": {\n\t\t\"id\": \"https:\/\/my-example.com\/hello-world\",\n\t\t\"type\": \"Note\",\n\t\t\"published\": \"2018-06-23T17:17:11Z\",\n\t\t\"attributedTo\": \"https:\/\/my-example.com\/actor\",\n\t\t\"inReplyTo\": \"https:\/\/mastodon.social\/@Gargron\/100254678717223630\",\n\t\t\"content\": \"&lt;p&gt;Hello world&lt;\/p&gt;\",\n\t\t\"to\": \"https:\/\/www.w3.org\/ns\/activitystreams#Public\"\n\t}\n}<\/pre><\/div>\n<p>With the <code>inReplyTo<\/code> property we\u2019re chaining our message to a parent. The <code>content<\/code> property may contain HTML, although of course it will be sanitized by the receiving servers according to their needs \u2014 different implementations may find use for a different set of markup. Mastodon will only keep <code>p<\/code>, <code>br<\/code>, <code>a<\/code> and <code>span<\/code> tags. With the <code>to<\/code> property we are defining who should be able to view our message, in this case it\u2019s a special value to mean \u201ceveryone\u201d.<\/p>\n<p>For our purposes, we don\u2019t actually need to host this document publicly, although ideally both the activity and the object would be separately available under their respective <code>id<\/code>. Let\u2019s just save it under <code>create-hello-world.json<\/code> because we\u2019ll need it later.<\/p>\n<p>So the next question is, how do we send this document over, where do we send it, and how will Mastodon be able to trust it?<\/p>\n<h3 id=\"http-signatures\">HTTP signatures<\/h3>\n<p>To deliver our message, we will use POST it to the inbox of the person we are replying to (in this case, me). That inbox is <code>https:\/\/mastodon.social\/inbox<\/code>. But a simple POST will not do, for how would anyone know it comes from the real @alice@my-example.com and not literally anyone else? For that purpose, we need a HTTP signature. It\u2019s a HTTP header signed by the RSA keypair that we generated earlier, and that\u2019s associated with our actor.<\/p>\n<p>HTTP signatures is one of those things that are much easier to do with actual code instead of manually. The signature looks like this:<\/p>\n<pre>Signature: keyId=\"https:\/\/my-example.com\/actor#main-key\",headers=\"(request-target) host date\",signature=\"...\"\n<\/pre>\n<p>The <code>keyId<\/code> refers to public key of our actor, the <code>header<\/code> lists the headers that are used for building the signature, and then finally, the <code>signature<\/code> string itself. The order of the headers must be the same in plain-text and within the to-be-signed string, and header names are always lowercase. The <code>(request-target)<\/code> is a special, fake header that pins down the HTTP method and the path of the destination.<\/p>\n<p>The to-be-signed string would look something like this:<\/p>\n<pre>(request-target): post \/inbox\nhost: mastodon.social\ndate: Sun, 06 Nov 1994 08:49:37 GMT\n<\/pre>\n<p>Mind that there is only a \u00b130 seconds time window when that signature would be considered valid, which is a big reason why it\u2019s quite difficult to do manually. Anyway, assuming we\u2019ve got the valid date in there, we now need to build a signed string out of it. Let\u2019s put it all together:<\/p>\n<div class=\"highlight\"><pre class=\"language-ruby\" data-lang=\"ruby\">require 'http'\nrequire 'openssl'\ndocument      = File.read('create-hello-world.json')\ndate          = Time.now.utc.httpdate\nkeypair       = OpenSSL::PKey::RSA.new(File.read('private.pem'))\nsigned_string = \"(request-target): post \/inbox\\nhost: mastodon.social\\ndate: #{date}\"\nsignature     = Base64.strict_encode64(keypair.sign(OpenSSL::Digest::SHA256.new, signed_string))\nheader        = 'keyId=\"https:\/\/my-example.com\/actor\",headers=\"(request-target) host date\",signature=\"' + signature + '\"'\nHTTP.headers({ 'Host': 'mastodon.social', 'Date': date, 'Signature': header })\n    .post('https:\/\/mastodon.social\/inbox', body: document)<\/pre><\/div>\n<p>Let\u2019s save it as <code>deliver.rb<\/code>. I am using the HTTP.rb gem here, so you\u2019ll need to have that installed (<code>gem install http<\/code>). Finally, run the file with <code>ruby deliver.rb<\/code>, and your message should appear as a reply on my toot!<\/p>\n<h3>Conclusion<\/h3>\n<p>We have covered how to create a discoverable ActivityPub actor and how to send replies to other people. But there is a lot we haven\u2019t covered: How to follow and be followed (it requires a working inbox), how to have a prettier profile, how to support document forwarding with LD-Signatures, and more. If there is demand, I will write more in-depth tutorials!<\/p>\n<p>Read more on:<\/p>\n<ul><li><a href=\"https:\/\/www.w3.org\/TR\/activitypub\/\">ActivityPub<\/a><\/li>\n<li><a href=\"https:\/\/tools.ietf.org\/html\/rfc7033\">Webfinger<\/a><\/li>\n<li><a href=\"https:\/\/tools.ietf.org\/html\/draft-cavage-http-signatures-10\">HTTP Signatures<\/a><\/li>\n<\/ul><p>\n            \u2014\n              \n                \n                Eugen Rochko\n              \n            ,\n            <\/p>","created_at":"2018-11-07T16:22:32+0000","updated_at":"2018-11-07T16:22:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"blog.joinmastodon.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":93,"title":"jdenen\/mastodon.el","url":"https:\/\/github.com\/jdenen\/mastodon.el","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a href=\"https:\/\/melpa.org\/#\/mastodon\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/581876012863a065e8e0895728d66525182a6ccb\/68747470733a2f2f6d656c70612e6f72672f7061636b616765732f6d6173746f646f6e2d62616467652e737667\" alt=\"https:\/\/melpa.org\/packages\/mastodon-badge.svg\" data-canonical-src=\"https:\/\/melpa.org\/packages\/mastodon-badge.svg\" \/><\/a><\/p>\n<p><a href=\"https:\/\/travis-ci.org\/jdenen\/mastodon.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/08431639a62142df7181c388c73517a71c336917\/68747470733a2f2f7472617669732d63692e6f72672f6a64656e656e2f6d6173746f646f6e2e656c2e7376673f6272616e63683d6d6173746572\" alt=\"https:\/\/travis-ci.org\/jdenen\/mastodon.el.svg?branch=master\" data-canonical-src=\"https:\/\/travis-ci.org\/jdenen\/mastodon.el.svg?branch=master\" \/><\/a>\n  <a href=\"http:\/\/waffle.io\/jdenen\/mastodon.el\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/dab5aa08a01755b5ad47f534132c18ca41c30d23\/68747470733a2f2f62616467652e776166666c652e696f2f6a64656e656e2f6d6173746f646f6e2e656c2e706e673f6c6162656c3d696e25323070726f6772657373267469746c653d496e25323050726f6772657373\" alt=\"https:\/\/badge.waffle.io\/jdenen\/mastodon.el.png?label=in%20progress&amp;title=In%20Progress\" data-canonical-src=\"https:\/\/badge.waffle.io\/jdenen\/mastodon.el.png?label=in%20progress&amp;title=In%20Progress\" \/><\/a><\/p>\n<p>Emacs client for <a href=\"https:\/\/github.com\/tootsuite\/mastodon\">Mastodon<\/a><\/p>\n<p><a href=\"http:\/\/spacemacs.org\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/8f5ee1d9b7278ed2c63f1f89e9a583ec59b52b6b\/68747470733a2f2f63646e2e7261776769742e636f6d2f73796c3230626e722f73706163656d6163732f343432643032353737396461326636326663383663323038323730333639373731346462363531342f6173736574732f73706163656d6163732d62616467652e737667\" alt=\"https:\/\/cdn.rawgit.com\/syl20bnr\/spacemacs\/442d025779da2f62fc86c2082703697714db6514\/assets\/spacemacs-badge.svg\" data-canonical-src=\"https:\/\/cdn.rawgit.com\/syl20bnr\/spacemacs\/442d025779da2f62fc86c2082703697714db6514\/assets\/spacemacs-badge.svg\" \/><\/a><\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>Clone this repository and add the lisp directory to your load path.\n  Then, require it and go.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(add-to-list 'load-path \"\/path\/to\/mastodon.el\/lisp\")\n(require 'mastodon)<\/pre><\/div>\n<p>Or, with <code>use-package<\/code>:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(use-package mastodon\n  :load-path \"\/path\/to\/mastodon.el\/lisp\")<\/pre><\/div>\n<h3><a id=\"user-content-melpa\" class=\"anchor\" aria-hidden=\"true\" href=\"#melpa\"><\/a>MELPA<\/h3>\n<p>Add <code>MELPA<\/code> to your archives:<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(require 'package)\n(add-to-list 'package-archives\n             '(\"melpa\" . \"http:\/\/melpa.org\/packages\/\") t)<\/pre><\/div>\n<p>Update and install:<\/p>\n<p><code>M-x package-refresh-contents RET<\/code><\/p>\n<p><code>M-x package-install RET mastodon RET<\/code><\/p>\n<h3><a id=\"user-content-emoji\" class=\"anchor\" aria-hidden=\"true\" href=\"#emoji\"><\/a>Emoji<\/h3>\n<p><code>mastodon-mode<\/code> will enable <a href=\"https:\/\/github.com\/iqbalansari\/emacs-emojify\">Emojify<\/a> if it is loaded in your Emacs environment, so\n  there\u2019s no need to write your own hook anymore. <code>emojify-mode<\/code> is not required.<\/p>\n<h3><a id=\"user-content-discover\" class=\"anchor\" aria-hidden=\"true\" href=\"#discover\"><\/a>Discover<\/h3>\n<p><code>mastodon-mode<\/code> will provide a context menu for its keybindings if <a href=\"https:\/\/github.com\/mickeynp\/discover.el\">Discover<\/a> is\n  installed. It is not required.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<h3><a id=\"user-content-2-factor-auth\" class=\"anchor\" aria-hidden=\"true\" href=\"#2-factor-auth\"><\/a>2 Factor Auth<\/h3>\n<p>2FA is not supported yet. It is in the <a href=\"https:\/\/github.com\/jdenen\/mastodon.el\/milestone\/2\">plans<\/a> for the <code>1.0.0<\/code> release.<\/p>\n<p>If you have 2FA enabled and try to use mastodon.el, your Emacs client will hang until you `C-g` your way out.<\/p>\n<h3><a id=\"user-content-instance\" class=\"anchor\" aria-hidden=\"true\" href=\"#instance\"><\/a>Instance<\/h3>\n<p>Set <code>mastodon-instance-url<\/code> in your <code>.emacs<\/code> or <code>customize<\/code>. Defaults to the <a href=\"https:\/\/mastodon.social\" rel=\"nofollow\">flagship<\/a>.<\/p>\n<div class=\"highlight highlight-source-emacs-lisp\"><pre>(setq mastodon-instance-url \"https:\/\/my.instance.url\")<\/pre><\/div>\n<p>There is an option to have your user credentials (email address and password) saved to disk so you don\u2019t have to re-enter them on every restart.\n  The default is not to do this because if not properly configured it would save these unencrypted which is not a good default to have.\n  Customize the variable <code>mastodon-auth-source-file<\/code> if you want to enable this feature.<\/p>\n<h3><a id=\"user-content-timelines\" class=\"anchor\" aria-hidden=\"true\" href=\"#timelines\"><\/a>Timelines<\/h3>\n<p><code>M-x mastodon<\/code><\/p>\n<p>Opens a <code>*mastodon-home*<\/code> buffer in the major mode so you can see toots. You will be prompted for email and password. The app registration process will take place if your <code>mastodon-token-file<\/code> does not contain <code>:client_id<\/code> and <code>:client_secret<\/code>.<\/p>\n<h4><a class=\"anchor\" aria-hidden=\"true\" href=\"#keybindings\"><\/a>Keybindings<\/h4>\n<table><tbody><tr><th>Key<\/th><th>Action<\/th><\/tr><tr><td><i>In-buffer navigation<\/i><\/td><\/tr><tr><td><code>&lt;space&gt;<\/code><\/td><td>Scroll up (i.e. move down to older items)<\/td><\/tr><tr><td><code>&lt;delete&gt;<\/code><\/td><td>Scroll down (i.e. move up to newer items)<\/td><\/tr><tr><td><code>&lt;<\/code><\/td><td>Move to beginning of buffer<\/td><\/tr><tr><td><code>&gt;<\/code><\/td><td>Move to end of buffer<\/td><\/tr><tr><td><code>j<\/code><\/td><td>Go to next item (toot, notification)<\/td><\/tr><tr><td><code>k<\/code><\/td><td>Go to previous item (toot, notification)<\/td><\/tr><tr><td><code>&lt;tab&gt;<\/code> \/ <code>h<\/code><\/td><td>Go to the next interesting thing that has an action<\/td><\/tr><tr><td><code>&lt;S-tab&gt;<\/code> \/ <code>l<\/code><\/td><td>Go to the previous interesting thing that has an action<\/td><\/tr><tr><td><i>In-buffer actions<\/i><\/td><\/tr><tr><td><code>?<\/code><\/td><td>Open context menu (if <code>discover<\/code> is available)<\/td><\/tr><tr><td><code>c<\/code><\/td><td>Toggle the visibility of sensitive text (if there is text with a content warning)<\/td><\/tr><tr><td><code>b<\/code><\/td><td>Boost toot under <code>point<\/code><\/td><\/tr><tr><td><code>f<\/code><\/td><td>Favourite toot under <code>point<\/code><\/td><\/tr><tr><td><code>r<\/code><\/td><td>Reply to toot under <code>point<\/code><\/td><\/tr><tr><td><code>&lt;return&gt;<\/code> \/ <code>&lt;mouse-2&gt;<\/code><\/td><td>Perform action for the thing under point (or under mouse for <code>&lt;mouse-2&gt;<\/code>) if any<\/td><\/tr><tr><td><code>n<\/code><\/td><td>Compose a new toot<\/td><\/tr><tr><td><i>Switching to other buffers<\/i><\/td><\/tr><tr><td><code>A<\/code><\/td><td>Open author profile of toot under <code>point<\/code><\/td><\/tr><tr><td><code>N<\/code><\/td><td>Open buffer with notifications<\/td><\/tr><tr><td><code>F<\/code><\/td><td>Open federated timeline<\/td><\/tr><tr><td><code>H<\/code><\/td><td>Open home timeline<\/td><\/tr><tr><td><code>L<\/code><\/td><td>Open local timeline<\/td><\/tr><tr><td><code>N<\/code><\/td><td>Open notifications timeline<\/td><\/tr><tr><td><code>t<\/code><\/td><td>Open thread buffer for toot under <code>point<\/code><\/td><\/tr><tr><td><code>T<\/code><\/td><td>Prompt for tag and open its timeline<\/td><\/tr><tr><td><code>U<\/code><\/td><td>Open profile of user attached to toot under <code>point<\/code><\/td><\/tr><tr><td><i>Quitting<\/i><\/td><\/tr><tr><td><code>q<\/code><\/td><td>Quit mastodon buffer, leave window open<\/td><\/tr><tr><td><code>Q<\/code><\/td><td>Quit mastodon buffer and kill window<\/td><\/tr><\/tbody><\/table><h4><a id=\"user-content-legend\" class=\"anchor\" aria-hidden=\"true\" href=\"#legend\"><\/a>Legend<\/h4>\n<table><tbody><tr><th>Marker<\/th><th>Meaning<\/th><\/tr><tr><td><code>(B)<\/code><\/td><td>I boosted this toot.<\/td><\/tr><tr><td><code>(F)<\/code><\/td><td>I favourited this toot.<\/td><\/tr><\/tbody><\/table><h3><a id=\"user-content-toot-toot\" class=\"anchor\" aria-hidden=\"true\" href=\"#toot-toot\"><\/a>Toot toot<\/h3>\n<p><code>M-x mastodon-toot<\/code><\/p>\n<p>Pops a new buffer\/window with a <code>mastodon-toot<\/code> minor mode. Enter the\n  contents of your toot here. <code>C-c C-c<\/code> sends the toot. <code>C-c C-k<\/code> cancels.\n  Both actions kill the buffer and window.<\/p>\n<p>If you have not previously authenticated, you will be prompted for your\n  account email and password. <b>NOTE<\/b>: Email and password are NOT stored by mastodon.el.<\/p>\n<p>Authentication stores your access token in the <code>mastodon-auth--token<\/code>\n  variable. It is not stored on your filesystem, so you will have to\n  re-authenticate when you close\/reopen Emacs.<\/p>\n<h4><a id=\"user-content-keybindings-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#keybindings-1\"><\/a>Keybindings<\/h4>\n<table><tbody><tr><th>Key<\/th><th>Action<\/th><\/tr><tr><td><code>C-c C-c<\/code><\/td><td>Send toot<\/td><\/tr><tr><td><code>C-c C-k<\/code><\/td><td>Cancel toot<\/td><\/tr><tr><td><code>C-c C-w<\/code><\/td><td>Add content warning<\/td><\/tr><\/tbody><\/table><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#roadmap\"><\/a>Roadmap<\/h2>\n<p><a href=\"https:\/\/github.com\/jdenen\/mastodon.el\/milestone\/1\">Here<\/a> are the features I plan to implement before putting mastodon.el on MELPA.<\/p>\n<p><a href=\"https:\/\/github.com\/jdenen\/mastodon.el\/milestone\/2\">Here<\/a> are the plans I have for the <code>1.0.0<\/code> release.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>PRs, issues, and feature requests are very welcome!<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h3>\n<ol><li>Create an <a href=\"https:\/\/github.com\/jdenen\/mastodon.el\/issues\">issue<\/a> detailing the feature you\u2019d like to add.<\/li>\n  <li>Fork the repository and create a branch off of <code>develop<\/code>.<\/li>\n  <li>Create a pull request referencing the issue created in step 1.<\/li>\n<\/ol><h3><a id=\"user-content-fixes\" class=\"anchor\" aria-hidden=\"true\" href=\"#fixes\"><\/a>Fixes<\/h3>\n<ol><li>In an <a href=\"https:\/\/github.com\/jdenen\/mastodon.el\/issues\">issue<\/a>, let me know that you\u2019re working to fix it.<\/li>\n  <li>Fork the repository and create a branch off of <code>develop<\/code>.<\/li>\n  <li>Create a pull request referencing the issue from step 1.<\/li>\n<\/ol><h2><a id=\"user-content-connect\" class=\"anchor\" aria-hidden=\"true\" href=\"#connect\"><\/a>Connect<\/h2>\n<p>If you want to get in touch with me, give me a <a href=\"https:\/\/mastodon.social\/@johnson\" rel=\"nofollow\">toot<\/a> or leave an <a href=\"https:\/\/github.com\/jdenen\/mastodon.el\/issues\">issue<\/a>.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:32+0000","updated_at":"2018-11-07T16:22:32+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/4671695?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":88,"title":"Cardboard Children: Star Wars Rebellion Part One","url":"https:\/\/www.rockpapershotgun.com\/2016\/03\/29\/star-wars-rebellion-review\/","content":"<p>Hello youse.<\/p><p>Star Wars: Rebellion is a two-player board game from Fantasy Flight Games, re-telling the story of the original Star Wars trilogy in about four hours. The game can also be played with more players, in teams, but the best way to play seems to be head-to-head. It\u2019s the Empire vs The Rebel Alliance, across the entire Star Wars galaxy, and it encompasses strategy, tactics, combat, deception and good old fashioned trash talking. It\u2019s an absolute slam dunk of an experience.<\/p>\n<p>STAR WARS: REBELLION<\/p>\n<p>We just shot a segment for the BBC TV show videoGaiden that features Star Wars: Rebellion as part of a broader conversation about modern board games. We shot the segment inside a planetarium \u2013 a grand setting for one of the most grand games I\u2019ve ever played.<\/p>\n<p>This is a remarkable production, a premium product \u2013 expensive, but with a great deal of value in the box. The game comes with a great number of miniatures representing the iconic Star Wars troops and ships \u2013 Stormtroopers, AT-ATs, X-Wings, Y-Wings, even the Star Destroyers and the Death Star itself. The board comes in two large segments, creating a massive galaxy to play across. And every planet is essential.<\/p>\n<p>Every planet is essential to the experience, because here\u2019s the coolest hook of this Star Wars game \u2013 the Rebel player is hidden. The planet that holds the Rebel base is secret, and the Imperial player has to probe and expand across the galaxy looking for the base\u2019s secret location. While this is going on, the Rebel player is sabotaging Imperial plans and developments, and indulging in activities that increase Rebel sympathy. <\/p>\n<p>The Imperial player wins by finding that Rebel base and levelling it. The Rebel player wins by surviving and increasing Rebel sympathies until the Empire is overthrown. It\u2019s a beautiful, exciting, dynamic setup.<\/p>\n<p><img src=\"https:\/\/www.rockpapershotgun.com\/images\/16\/mar\/rebel2.jpg\" alt=\"\" \/><\/p>\n<p>By far the most impressive element of this board game is how streamlined it is. While there is incredible production quality and lots of great stuff inside the box, there isn\u2019t too much stuff. This is not one of those epic Fantasy Flight Games products of old, with a million chits and tokens to track and represent a million things. This design is super-clean, and any fuss or admin is kept to an absolute minimum, elevating the gameplay experience (and the storytelling experience) to the very top. <\/p>\n<p>Oh, I\u2019m excited. I\u2019m excited. So here\u2019s how it works. It\u2019s not even difficult to explain. <\/p>\n<p>At the start of the game, the Rebel player chooses a planet for her base. She then takes a \u201cProbe Deck\u201d full of cards representing all those planets on the board, and she removes the card that features her chosen base location. She keeps this. As the game progresses, and the Empire\u2019s probes do their work, and special Imperial missions are completed, the Imperial player will draw cards from this deck, eliminating planets from the search. It\u2019s beautiful. There is a mission card that the Empire player can activate that simply collects intel about the Rebel base, and this spits out more Probe Deck cards depending on how many units are at the Rebel base. So if the Rebel base is heavily populated, there\u2019s more chance of word leaking out about where it is. How cool is that? How logical and narratively strong is that? And that\u2019s just one card.<\/p>\n<p>Cards, missions and leaders. These things are the key to the game. Each player starts the game with a small pool of leaders, the leaders affiliated with the two sides at the very start of A New Hope. These leaders are used to activate mission cards and advance troops from space sector to space sector. When a leader is \u201csent on a mission\u201d, an opposing leader can be sent to oppose it, and there is a skill stat roll-off to see if the mission succeeds. It\u2019s very simple, very neat.<\/p>\n<p>It becomes clear very quickly that choosing when and how to use leaders is crucial to success in the game. Send too many leaders on missions and you can\u2019t move troops out. Send the wrong leader on the wrong mission and you might lack the perfect leader to oppose one of your enemy\u2019s moves. It makes even that basic leader allocation phase exciting.<\/p>\n<p>But the true excitement is in how you can use those decisions, and those mission cards, to lay traps. You can send a leader on a mission just to bring an opposing leader into action. Then you can enact another mission that affects that leader. What do I mean? Well, as an example, during my very first game I sucked a Rebel leader into a diplomacy situation then sent Boba Fett to capture her while she was there. So, in a story sense, leaders and missions interact with each other, creating the thrilling moments you expect from the Star Wars universe.<\/p>\n<p><img src=\"https:\/\/www.rockpapershotgun.com\/images\/16\/mar\/rebel1.jpg\" alt=\"\" \/><\/p>\n<p>Combat is key to planetary conquest or liberation, and conquest and liberation is key to production. The Empire player can use missions and influence to turn a planet bad, but can also send down Stormtroopers to oppress even a Rebel-sympathising planet, making the people work for them. The Rebels have to win planets over to their side and clear Imperial forces from planets that are being subjugated. It\u2019s so fucking good. And you want your influence token on those planets, because these planets will be building new troops for each player, spitting troops onto a production queue. Yes, you can build new Star Destroyers and watch with wide eyes as they get closer to being finished with each passing round.<\/p>\n<p>In a nutshell, that\u2019s the game. Send leaders on missions, send leaders into battle. Take over planets, or liberate them to make them your friend. Create new troops. Go again.<\/p>\n<p>But there\u2019s much more to talk about, and next week we\u2019ll dig into the small details of this game, the stuff that makes Star Wars: Rebellion a true stand-out. We\u2019ll talk about sabotage, and recruitment, and story. We\u2019ll talk about individual cards that make you pump your fist in joy. And if the BBC allow, we\u2019ll show you a little bit of video of the game in action.<\/p>\n<p>Be here next week for Part II.<\/p>","created_at":"2018-11-07T16:22:31+0000","updated_at":"2018-11-07T16:22:31+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":5,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/www.rockpapershotgun.com\/images\/16\/mar\/rebelheader.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":89,"title":"umpox\/TinyEditor","url":"https:\/\/github.com\/umpox\/TinyEditor","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Paste the following code into your browser address bar:<\/p>\n<pre>data:text\/html,&lt;body oninput=\"i.srcdoc=h.value+'&lt;style&gt;'+c.value+'&lt;\/style&gt;&lt;script&gt;'+j.value+'&lt;\/script&gt;'\"&gt;&lt;style&gt;textarea,iframe{width:100%;height:50%}body{margin:0}textarea{width:33.33%;font-size:18}&lt;\/style&gt;&lt;textarea placeholder=HTML id=h&gt;&lt;\/textarea&gt;&lt;textarea placeholder=CSS id=c&gt;&lt;\/textarea&gt;&lt;textarea placeholder=JS id=j&gt;&lt;\/textarea&gt;&lt;iframe id=i&gt;\n<\/pre>\n<p>If you want to read the code, check out <a href=\"https:\/\/github.com\/umpox\/TinyEditor\/blob\/master\/index.html\">index.html<\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:22:31+0000","updated_at":"2018-11-07T16:22:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/9516420?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":90,"title":"umpox\/TinyEditor","url":"https:\/\/github.com\/umpox\/TinyEditor","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Paste the following code into your browser address bar:<\/p>\n<pre>data:text\/html,&lt;body oninput=\"i.srcdoc=h.value+'&lt;style&gt;'+c.value+'&lt;\/style&gt;&lt;script&gt;'+j.value+'&lt;\/script&gt;'\"&gt;&lt;style&gt;textarea,iframe{width:100%;height:50%}body{margin:0}textarea{width:33.33%;font-size:18}&lt;\/style&gt;&lt;textarea placeholder=HTML id=h&gt;&lt;\/textarea&gt;&lt;textarea placeholder=CSS id=c&gt;&lt;\/textarea&gt;&lt;textarea placeholder=JS id=j&gt;&lt;\/textarea&gt;&lt;iframe id=i&gt;\n<\/pre>\n<p>If you want to read the code, check out <a href=\"https:\/\/github.com\/umpox\/TinyEditor\/blob\/master\/index.html\">index.html<\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:22:31+0000","updated_at":"2018-11-07T16:22:31+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/9516420?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":85,"title":"Tutorial Introduction to Guile","url":"https:\/\/www.gnu.org\/software\/guile\/docs\/guile-tut\/tutorial.html","content":"Tutorial Introduction to Guile\n<div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Fundamentals\">Fundamentals<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#dir\">(dir)<\/a><\/div>\n<h2 class=\"unnumbered\">Tutorial Introduction to Guile<\/h2>\n   <p>This document is an introduction into using Guile, the GNU extension language\nand system.  Guile is a dialect of the Scheme programming language, and I will\nassume you're at least confident about the very basics of Scheme or LISP in\ngeneral.  However, Guile also allows application developers to integrate it\ninto their code to provide scripting capability to it and extend the main\nGuile language with primitives specific to the application being extended. \nExactly this is what I want to give a brief introduction to.\n   <\/p><p>If you want to develop and test the code presented here on-the-fly for yourself\nwhile working through the text (it is for sure a good idea and even better if\nyou want to play around with certain stuff not literally presented inbetween),\nyou need of course to install GNU Guile (including the appropriate development\npackage containing header files, if necessary for your system) as well as\nGnuplot, and have a UNIX-like\nenvironment and some C compiler (just use gcc).  I worked on a GNU\/Linux\nmachine with Guile 1.8.7, Gnuplot 4.2 and gcc 4.3.1, but any recent Guile and\nGnuplot should do it.\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Tortoise\">Tortoise<\/a>,\nPrevious:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Top\">Top<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Top\">Top<\/a><\/div>\n<h2 class=\"chapter\">1 The Fundamentals<\/h2>\n<p>Guile is an implementation and dialect of the Scheme programming language\n(Scheme is a dialect of LISP itself).  It is intended to be used as the\nextension language of the GNU project.\n   <\/p><p>That means that Guile is designed as a library you can include into your own\nproject and make the interpreter run code within it; additionally, you can\nprovide special procedures in this Guile environment that interface to the\ncore of your application.  This way, Scheme scripts written by a user can\nmanipulate stuff \u201cwithin\u201d your application.\n   <\/p><p>The point of this all is to provide an easy way for you to make your application\n<em>extensible<\/em> and <em>scriptable<\/em>.  For instance, if you need some sort of\nconfiguration files or even real scripting support for your application,\nyou can use\nGuile instead of rolling your own \u201csmall\u201d configuration or scripting\nlanguage and interpreter; it already exists, so don't waste your time on yet\nanother one but instead keep working on the new and exciting parts of your\nproject!  Having Guile as some \u201cuniversal\u201d scripting language (at least that's\nwhat the Guile people would like it to be, but I think it's suited to this\ngoal) also means that a user does not have to learn different languages for\neach application she wants to configure or script.  Instead she can do so on\nall her favourite software with just learning Scheme.  Besides, at least my\npersonal opinion is that Scheme is very nice and fun to program in, and\nvery well suited to small pieces of code like scripts.\n   <\/p><p>And the very best:  The current development on Guile allows the core Guile\n\u201cplatform\u201d to run not only Scheme code but support other languages as well\non top of Guile.  So you can just integrate Guile as scripting interpreter\ninto your code, and have it also run scripts in ECMAScript, Emacs Lisp or\nany other language that Guile will implement in the future (like Perl,\nTcl, Python or others) \u2013 and you don't have to think a single instant about\nthis!\n   <\/p><p>In concrete terms, installing Guile on your system provides basically two\nthings for you:\n   <\/p><p>First, you can use the <samp>guile<\/samp> command-line utility as a Scheme\ninterpreter to write programs (or as a table-top calculator\nif you want to).\n   <\/p><p>Second, use the <kbd>libguile<\/kbd> programming library to run applications\nusing Guile as scripting extension; or write your own that can access the\ninterpreter and make use of it to script your code.\n\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Guiling\">Guiling<\/a>,\nPrevious:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Fundamentals\">Fundamentals<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Top\">Top<\/a><\/div>\n<h2 class=\"chapter\">2 The Tortoise<\/h2>\n<blockquote><div>\n\u201cWhy did you call him Tortoise, if he wasn't one?\u201d Alice asked.\n   <p>\u201cWe called him Tortoise because he taught us,\u201d said the Mock Turtle angrily.\n   <\/p><p>(Lewis Carroll, Alice's Adventures in Wonderland)\n<\/p><\/div><\/blockquote>\n   <p>As the project to demonstrate Guile, we're going to develop a very simple\n\u201cTortoise\u201d graphics program.  It will use Gnuplot for graphics output, but\nshould be easy to adapt to any other graphics systems.\n   <\/p><p>It will produce graphics output by assuming that there's a tortoise sitting in\nthe middle of the screen; this tortoise is able to perform some basic\ninstructions given by the user.\n   <\/p><p>You can ask it to turn left by a certain number of degrees (or right by giving\na negative number), or you can intstruct it to walk a certain number of steps\nforward. \nIt has got a pen and you can ask it to carry it either in its paws or behind\nits ear, so that when it moves it will leave a mark on the ground or not.\n   <\/p><p>Finally, if you got yourself and the poor tortoise completely confused, you\ncan ask it to walk away to a fresh (empty) ground, sit in the middle and\nface right, just as at the very start.\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#First-Test\">First Test<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Tortoise\">Tortoise<\/a><\/div>\n<h3 class=\"section\">2.1 The Core System<\/h3>\n<p>Let's finally start and implement the core program that will keep track of the\ntortoise, its movements and the graphics output.  This will be implemented in\nC.  As mentioned before, I'm going to use Gnuplot for the graphics output\n(at first I wanted to use the Gtk+ toolkit, but there were some problems with\nthis approach I will come back to later).\n   <\/p><p>The idea is to start a Gnuplot process in the background and send it commands\nto draw the lines we want on the screen over a pipe.  Here's the code for a\nprogram that will start up a Gnuplot process and keep connected to it\nvia a pipe ready to get further plotting commands:\n<\/p><pre class=\"verbatim\">\/* Simple backend for a Logo like tortoise drawer.  *\/\n#include &lt;stdlib.h&gt;\n#include &lt;stdio.h&gt;\n#include &lt;unistd.h&gt;\nstatic const int WIDTH = 10;\nstatic const int HEIGHT = 10;\nstatic FILE*\nstart_gnuplot ()\n{\n  FILE* output;\n  int pipes[2];\n  pid_t pid;\n  pipe (pipes);\n  pid = fork ();\n  if (!pid)\n    {\n      dup2 (pipes[0], STDIN_FILENO);\n      execlp (\"gnuplot\", NULL);\n      return; \/* Not reached.  *\/\n    }\n  output = fdopen (pipes[1], \"w\");\n  fprintf (output, \"set multiplot\\n\");\n  fprintf (output, \"set parametric\\n\");\n  fprintf (output, \"set xrange [-%d:%d]\\n\", WIDTH, WIDTH);\n  fprintf (output, \"set yrange [-%d:%d]\\n\", HEIGHT, HEIGHT);\n  fprintf (output, \"set size ratio -1\\n\");\n  fprintf (output, \"unset xtics\\n\");\n  fprintf (output, \"unset ytics\\n\");\n  fflush (output);\n  return output;\n}\nstatic FILE* global_output;\nint\nmain (int argc, char* argv[])\n{\n  global_output = start_gnuplot ();\n  return EXIT_SUCCESS;\n}\n<\/pre>\n   <p>Note that here we're not doing any error checks on the system routines; you\nshouldn't be doing this yourself, but it keeps this code as simple as possible. \nAs that so far has nothing to do with what you want to read about in this\ntutorial, I think this should be the best way to go.\n   <\/p><p>What we're doing here is starting a Gnuplot process with the <var>start_gnuplot<\/var>\nroutine and opening a pipe that can feed commands to it into\n<var>global_output<\/var>, so that we are later able to plot lines.\n   <\/p><p>Gnuplot is started with a fixed coordinate range (-10 to 10 in both x and y\ndirections).  We're going to use parametric mode so we won't get any problems\ndrawing vertical lines, and multiplot mode in order to allow building the\ngraphics incrementally by adding single lines each at a time.\n   <\/p><p>Now, this code adds a routine for plotting a line from (<var>x1<\/var>, <var>y1<\/var>)\nto (<var>x2<\/var>, <var>y2<\/var>):\n<\/p><pre class=\"verbatim\">static void\ndraw_line (FILE* output, double x1, double y1, double x2, double y2)\n{\n  fprintf (output, \"plot [0:1] %f + %f * t, %f + %f * t notitle\\n\",\n           x1, x2 - x1, y1, y2 - y1);\n  fflush (output);\n}\n<\/pre>\n   <p>(You may want to read up on parametric plotting in Gnuplot or on parametric\nequations for lines if you're not sure what's going on here.  Or just believe\nme that this will do what we want for the moment.)\n   <\/p><p>Finally, we can write the routines that will control the tortoise; here's some\ntrigonometry involved, and you will need to <code>#include &lt;math.h&gt;<\/code>:\n<\/p><pre class=\"verbatim\">static double x, y;\nstatic double direction;\nstatic int pendown;\nstatic void\ntortoise_reset ()\n{\n  x = y = 0.0;\n  direction = 0.0;\n  pendown = 1;\n  fprintf (global_output, \"clear\\n\");\n  fflush (global_output);\n}\nstatic void\ntortoise_pendown ()\n{\n  pendown = 1;\n}\nstatic void\ntortoise_penup ()\n{\n  pendown = 0;\n}\nstatic void\ntortoise_turn (double degrees)\n{\n  direction += M_PI \/ 180.0 * degrees;\n}\nstatic void\ntortoise_move (double length)\n{\n  double newX, newY;\n  newX = x + length * cos (direction);\n  newY = y + length * sin (direction);\n  if (pendown)\n    draw_line (global_output, x, y, newX, newY);\n  x = newX;\n  y = newY;\n}\n<\/pre>\n   <p>That's it, just add a <code>tortoise_reset ();<\/code> call to the <var>main<\/var> routine\nafter starting Gnuplot, so that the tortoise starts at a well-defined position.\n<\/p><div class=\"node\"><hr \/>Previous:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Backend\">Backend<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Tortoise\">Tortoise<\/a><\/div>\n<h3 class=\"section\">2.2 Testing it out<\/h3>\n<p>Of course you want to try it now, don't you?  At least I'm burning with\nexcitement...  So, we're going to give the tortoise actually some instructions,\ndirectly via C code in the <var>main<\/var> routine:\n<\/p><pre class=\"verbatim\">{\n  int i;\n  tortoise_pendown (); \/* This is unnecessary, but makes it clearer.  *\/\n  for (i = 1; i &lt;= 4; ++i)\n    {\n      tortoise_move (3.0);\n      tortoise_turn (90.0);\n    }\n}\n<\/pre>\n   <p>As a side-note:  This program leaves the Gnuplot process alive when terminating;\nwe could send it a <samp>quit<\/samp> command before finishing, but later on we\nwon't be able to do so and instead of cooking up anything more complicated,\nI'll just leave it like that.  It basically works this way and even has the\nadvantage that the Gnuplot window stays open until you close it, despite the\nfact that our tortoise-program has already finished.  And if you get worried\nabout the processes, just do a <samp>killall gnuplot<\/samp> afterwards...\n   <\/p><p>I don't know about you, but I myself like using Makefiles; so save the full\ncode we've worked out to <samp>tortoise.c<\/samp> and this as <samp>Makefile<\/samp>:\n<\/p><pre class=\"verbatim\"># Basic Makefile for the tortoise package.\nCFLAGS =\nLIBS =\n.PHONY: clean build run\nbuild: tortoise\nclean:\n\trm -f tortoise tortoise.o\nrun: tortoise\n\t.\/tortoise\ntortoise: tortoise.o\n\tgcc $&lt; -o $@ $(LIBS)\ntortoise.o: tortoise.c\n\tgcc -c $&lt; -o $@ $(CFLAGS)\n<\/pre>\n   <p>Of course, <var>CFLAGS<\/var> and <var>LIBS<\/var> are quite useless at the moment, but\nthey will get some use later, so I just included them right away.  Now,\ndoing a <samp>make run<\/samp> should compile and run your code and open up a\nGnuplot window with the resulting graphics; for me it looks like this:\n   <\/p><div class=\"block-image\"><img src=\"https:\/\/www.gnu.org\/software\/guile\/docs\/guile-tut\/plot1.png\" alt=\"plot1.png\" \/><\/div>\n   <p>Congratulations, you've just done your first tortoise-graphics!  Still, we\ndon't want to fix and compile the instructions directly with our C code as\nwe just did; instead, the user should be able to interactively control the\ntortoise.\n\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Further\">Further<\/a>,\nPrevious:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Tortoise\">Tortoise<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Top\">Top<\/a><\/div>\n<h2 class=\"chapter\">3 Becoming Guiled<\/h2>\n<p>This is where Guile comes into play.  Without it, we could write some additional\ncode that reads, parses and interprets commands to the tortoise, calling the\nappropriate tortoise primitives as already defined.\n   <\/p><p>Sounds not too bad?  Well, probably it isn't.  But in the C code we also used a\nloop to draw the square \u2013 otherwise we would have had to repeat the same two\nlines (movement and turn) four times.  And what if we wanted to draw a polygon\nwith, say, 1024 vertices?  I hope you are convinced now that to make our tool\nreally useful, we also needed to implement some means of looping.\n   <\/p><p>Hm, while we're at it:  Our tortoise-package can also be used to build up\na Koch's curve quite elegantly using recursion (try a quick internet search if\nyou can't wait until I'll come back to it in more detail later) \u2013 so want to\nimplement some sort of procedures that can recursively call themselves?\n   <\/p><p>All in all, we were probably going to implement a full-fledged tortoise\nprogramming language \u2013 but think about it:  Has this\nanything to do with our original application and goal?  I hate to spoil you\nall the fun (writing programming languages can be quite fun), but no, it hasn't.\n   <\/p><p>Luckily, we need not go through all this effort; Guile already provides a\nperfectly good way of making our tortoise programmable.  Scheme should be\nquite a good language to perform all we want (loops, procedures, recursion)\n\u2013 and much more. \nWe just need to link our tortoise to the Guile-world, which will be the topic\nof the next sections.\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Second-Test\">Second Test<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Guiling\">Guiling<\/a><\/div>\n<h3 class=\"section\">3.1 Adding Guile<\/h3>\n<p>First, let's add Guile to our code; you need to <code>#include &lt;libguile.h&gt;<\/code>\nas the \u201cmaster header\u201d.  In our <var>main<\/var> routine, take out again the\ntemporary test code we used to produce the first output, and replace it\ninstead by these lines:\n<\/p><pre class=\"verbatim\">scm_with_guile (&amp;register_functions, NULL);\nscm_shell (argc, argv);\n<\/pre>\n   <p><var>scm_with_guile<\/var> may be a bit weird at first glance; what it does is\ncalling the provided routine (<var>register_functions<\/var> \u2013 we'll create it\nin a few moments).  Additionally, it also \u201cactivates\u201d Guile during this\ncall, so that\nother Guile library functions can be called from within\n<var>register_functions<\/var>.  Don't worry about what exactly goes on there, you\njust need to remember that this indirection is necessary if you want to work\nwith Guile.\n   <\/p><p>(As a side-note:  There's also a <var>scm_init_guile<\/var> method that\ndoes this \u201cactivation\u201d directly and without this peculiar call-back; it may\nseem more reasonable and easy-to-understand to you, and you can well use that\none instead.  However, Guile recommends using <var>scm_with_guile<\/var> for better\nportability where possible.)\n   <\/p><p>The call to <var>scm_shell<\/var> actually runs the Guile REPL (Read-Evaluate-Print\nLoop, that shell-like command prompt for interactive Scheme evaluation you\nalso get when starting up <samp>guile<\/samp> directly) \u2013 the idea is to do all\nthe initialization we need for our application, and finally run the REPL,\nwhere the user can then enter her commands in Scheme code.  Actually, this\nroutine does a little more:  It also processes the command-line arguments given\nto our tortoise-program in the way <samp>guile<\/samp> handles its arguments;\nfor instance, you can then do <samp>tortoise -s foobar.scm<\/samp> and have your\ncommands read from <samp>foobar.scm<\/samp>.  That's why we need to pass it\n<var>argc<\/var> and <var>argv<\/var>, in case you were interested.\n   <\/p><p><var>scm_shell<\/var> does not return, but rather keep on until the user closes\nthe REPL; then our tortoise-program will also finish as a whole. \nDo you remember that I still owe you an explanation why using the Gtk+ toolkit\nfor graphics would have been more complicated than Gnuplot?  The reason I\ndecided against Gtk+ is that for the Gtk+ toolkit you do something similar:\nAfter your program's initialization, you call out to a Gtk+ main loop that\nwill process incoming events and also never fully return control to your\nprogram \u2013 but in this case, we can't run <em>both<\/em> Gtk+'s main loop and\nthe Guile REPL; at least not without writing a multi-threaded application. \nThat would have been again some unnecessary complication; I hope you\nagree with me now...  But if not, just go ahead and try converting our\npackage to Gtk+!  I guess that's an interesting exercise.\n   <\/p><p>Finally, we still have to create <var>register_functions<\/var>:\n<\/p><pre class=\"verbatim\">static void*\nregister_functions (void* data)\n{\n  return NULL;\n}\n<\/pre>\n   <p>You're right, it's not really interesting or even useful for now, but we'll fill\nit in later.  Notice the two <code>void*<\/code>?  The argument <var>data<\/var> gets\npassed whatever we want from <var>scm_with_guile<\/var> \u2013 that's why there's\n<code>NULL<\/code> as second argument in the call, but we could make it a pointer to\nan integer or even a large struct in case we needed to pass some data to\n<var>register_functions<\/var>.  In turn, the return value of\n<var>register_functions<\/var>, which is also <code>NULL<\/code> in this case but could be\nanything as complex as you want, is returned from <var>scm_with_guile<\/var> \u2013\nsomething we also don't need.\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Register-Functions\">Register Functions<\/a>,\nPrevious:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Init-Guile\">Init Guile<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Guiling\">Guiling<\/a><\/div>\n<h3 class=\"section\">3.2 Compiling with Guile<\/h3>\n<p>Once again, we'll try to compile and run what we have so far.  To do so, we\nhave to tell the compiler where to look for <samp>libguile.h<\/samp> as well as the\nlinker to include <kbd>libguile<\/kbd>.  You can work out the appropriate flags\nfor yourself, but the utility <samp>guile-config<\/samp> that comes with Guile\ncan do it for you.\n   <\/p><p>So all we need to do is update the Makefile like this (and now the variables\nget some meaning!):\n<\/p><pre class=\"verbatim\">CFLAGS = `guile-config compile`\nLIBS = `guile-config link`\n<\/pre>\n   <p>With these adaptions, you should be able to compile and run the program again. \nThis time, the Gnuplot window that pops up should just be empty (and stay so),\nbut on the command-line a Guile REPL will be available just as if you had\nstarted <samp>guile<\/samp> yourself.\n   <\/p><p>You can do some Scheme programming, but so far nothing to use our little\ntortoise at all:\n<\/p><pre class=\"verbatim\">guile&gt; (define (foobar a b) (+ a b))\nguile&gt; (foobar (* 2 3) 4)\n10\nguile&gt; (map foobar '(1 2 3) '(4 5 6))\n(5 7 9)\nguile&gt; (tortoise-reset)\nBacktrace:\nIn current input:\n   1: 0* (tortoise-reset)\n&lt;unnamed port&gt;:1:1: In expression (tortoise-reset):\n&lt;unnamed port&gt;:1:1: Unbound variable: tortoise-reset\nABORT: (unbound-variable)\n<\/pre>\n   <p>That's quite nice, isn't it?  Still, without control over the tortoise, it does\nnot make sense at all; you can just use <samp>guile<\/samp> directly without all\nthe effort we went through if you want to program in Scheme.  But we can fix\nthis easily.\n<\/p><div class=\"node\"><hr \/>Next:\u00a0<a rel=\"next\" accesskey=\"n\" href=\"#Third-Test\">Third Test<\/a>,\nPrevious:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Second-Test\">Second Test<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Guiling\">Guiling<\/a><\/div>\n<h3 class=\"section\">3.3 Adding the Tortoise<\/h3>\n<p>As I've promised, we'll integrate the tortoise into the Guile environment now. \nWe can simply tell Guile to make our tortoise-procedures available from\nScheme with <var>scm_c_define_gsubr<\/var>; these need to be called from within\nthe activated \u201cGuile mode\u201d, so we add the calls to <var>register_functions<\/var>:\n<\/p><pre class=\"verbatim\">scm_c_define_gsubr (\"tortoise-reset\", 0, 0, 0, &amp;tortoise_reset);\nscm_c_define_gsubr (\"tortoise-penup\", 0, 0, 0, &amp;tortoise_penup);\nscm_c_define_gsubr (\"tortoise-pendown\", 0, 0, 0, &amp;tortoise_pendown);\nscm_c_define_gsubr (\"tortoise-turn\", 1, 0, 0, &amp;tortoise_turn);\nscm_c_define_gsubr (\"tortoise-move\", 1, 0, 0, &amp;tortoise_move);\n<\/pre>\n   <p>The first arguments are the Scheme-names for the procedures we're going to\nregister, the last arguments pointers to the C functions that implement them. \nThe numbers inbetween define the number of arguments the procedures take,\nnamely required, optional and whether there's a rest-list argument nor not \u2013\nin our case, there are neither optional arguments nor rest-lists, but\n<samp>tortoise-turn<\/samp> and <samp>tortoise-move<\/samp> take one required\nargument (the angle or distance, respectively).\n   <\/p><p>Unfortunatly, this is only one half of the changes needed; Guile represents\nall Scheme values with the <kbd>SCM<\/kbd> type, and thus both the return values and\nargument-types of the C procedures need to be <kbd>SCM<\/kbd>'s.  In order to get the\n<code>double<\/code> values out, we need some Guile API functions; our\ntortoise procedures become:\n<\/p><pre class=\"verbatim\">static SCM\ntortoise_reset ()\n{\n  x = y = 0.0;\n  direction = 0.0;\n  pendown = 1;\n  fprintf (global_output, \"clear\\n\");\n  fflush (global_output);\n  return SCM_UNSPECIFIED;\n}\nstatic SCM\ntortoise_pendown ()\n{\n  SCM result = scm_from_bool (pendown);\n  pendown = 1;\n  return result;\n}\nstatic SCM\ntortoise_penup ()\n{\n  SCM result = scm_from_bool (pendown);\n  pendown = 0;\n  return result;\n}\nstatic SCM\ntortoise_turn (SCM degrees)\n{\n  const double value = scm_to_double (degrees);\n  direction += M_PI \/ 180.0 * value;\n  return scm_from_double (direction * 180.0 \/ M_PI);\n}\nstatic SCM\ntortoise_move (SCM length)\n{\n  const double value = scm_to_double (length);\n  double newX, newY;\n  newX = x + value * cos (direction);\n  newY = y + value * sin (direction);\n  if (pendown)\n    draw_line (global_output, x, y, newX, newY);\n  x = newX;\n  y = newY;\n  return scm_list_2 (scm_from_double (x), scm_from_double (y));\n}\n<\/pre>\n   <p>You've already guessed it, <var>scm_to_double<\/var> gets the value \u201cout of\u201d a\nScheme number (if it is a number at all, but we won't care about this and\nassume we're only fed correct arguments).\n   <\/p><p>You also noticed for sure that I made our procedures return some \u201creasonable\u201d\nvalues, using some other Guile API stuff:  <code>SCM_UNSPECIFIED<\/code> means just\nthat there's no return value (like <code>void<\/code> in C).  If you define a\nprocedure that has no meaningful result, you at least need to return\n<code>SCM_UNSPECIFIED<\/code> from it to make Guile happy (and assert this fact).\n   <\/p><p>The <samp>penup<\/samp>\nand <samp>pendown<\/samp> commands return (as a Scheme boolean, created by\n<var>scm_from_bool<\/var>) the state of the pen before the requested change.\n   <\/p><p><samp>turn<\/samp> and <samp>move<\/samp> return the new direction and coordinate\n(as a Scheme list with two entries), respectively \u2013 of course, you already\nknow what <var>scm_from_double<\/var> and <var>scm_list_2<\/var> do, right?\n   <\/p><p>Congratulations, now we've managed to build the whole tortoise package with\nGuile!  And left is all the fun working with it...\n<\/p><div class=\"node\"><hr \/>Previous:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Register-Functions\">Register Functions<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Guiling\">Guiling<\/a><\/div>\n<h3 class=\"section\">3.4 Having Fun with our Tortoise<\/h3>\n<p>Recompile the program (should be easy thanks to our neat Makefile) and run it. \nOnce again, you should get the Guile (Tortoise!) REPL, but now our procedures\nshould really be available.  Try out all of the functionality with a session\nlike this:\n<\/p><pre class=\"verbatim\">guile&gt; (tortoise-move 1)\n(1.0 0.0)\nguile&gt; (tortoise-turn 90)\n90.0\nguile&gt; (tortoise-penup)\n#t\nguile&gt; (tortoise-move 5)\n(1.0 5.0)\nguile&gt; (tortoise-pendown)\n#f\nguile&gt; (tortoise-move 1)\n(1.0 6.0)\nguile&gt; (tortoise-reset)\nguile&gt; (tortoise-move (sqrt 2))\n(1.4142135623731 0.0)\nguile&gt; (quit)\n<\/pre>\n   <p>(I won't include the expected graphics output, just see for yourself...)  I\nhope everything works for you, too.  If not, it's the perfect time to go back\nto the code and try to find what's going wrong...  Otherwise, let's have some\nfun and do real programming with our tortoise!\n   <\/p><p>For instance, we can automate the task of drawing polygons with a Scheme\nfunction; try out this code:\n<\/p><pre class=\"lisp\">     (define (draw-polygon! circumference vertices)\n       (let ((side (\/ circumference vertices))\n             (angle (\/ 360 vertices)))\n         (let iterate ((i 1))\n           (if (&lt;= i vertices)\n             (begin\n               (tortoise-move side)\n               (tortoise-turn angle)\n               (iterate (1+ i)))))))\n     \n     (draw-polygon! 16 4)\n     \n     (tortoise-penup)\n     (tortoise-move 1)\n     (tortoise-turn 30)\n     (tortoise-pendown)\n     (draw-polygon! 12 3)\n     \n     (tortoise-penup)\n     (tortoise-move -2)\n     (tortoise-turn -100)\n     (tortoise-pendown)\n     (draw-polygon! 10 64)\n<\/pre>\n   <p>This is what I get, but try for yourself:\n   <\/p><div class=\"block-image\"><img src=\"https:\/\/www.gnu.org\/software\/guile\/docs\/guile-tut\/plot2.png\" alt=\"plot2.png\" \/><\/div>\n   <p>Remember that I promised you something about Koch's curve?  Now we're going to\nconstruct it, using a recursive function.  See for instance\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/Koch_curve\">http:\/\/en.wikipedia.org\/wiki\/Koch_curve<\/a> for details about what Koch's\ncurve is.  Anyways, here comes the code; try to find out how it works, if you\nwant:\n<\/p><pre class=\"lisp\">     (define (koch-line length depth)\n       (if (zero? depth)\n         (tortoise-move length)\n         (let ((sub-length (\/ length 3))\n               (sub-depth (1- depth)))\n           (for-each (lambda (angle)\n                       (koch-line sub-length sub-depth)\n                       (tortoise-turn angle))\n                     '(60 -120 60 0)))))\n     \n     (define (snowflake length depth sign)\n       (let iterate ((i 1))\n         (if (&lt;= i 3)\n           (begin\n             (koch-line length depth)\n             (tortoise-turn (* sign -120))\n             (iterate (1+ i))))))\n     \n     (snowflake 8 3 1)\n     (tortoise-turn 180)\n     (snowflake 8 3 -1)\n<\/pre>\n   <p>Are you impressed?  I hope at least a little so...  Unfortunatly, the\nGnuplot-over-a-pipe approach is quite slow (at least on my system), so you\ncould probably have been even more impressed if the graphics would have been\nthere more quickly; but anyways, at least I like this very much!  Regarding the\nScheme code, I won't comment it further, as not to spoil all your fun\nthinking about it.  I encourage you to do so, I like this recursive\nconstruction very much.  Oh, and here's my resulting plot:\n   <\/p><div class=\"block-image\"><img src=\"https:\/\/www.gnu.org\/software\/guile\/docs\/guile-tut\/plot3.png\" alt=\"plot3.png\" \/><\/div>\n   <p>Looks quite interesting, doesn't it?  By the way, this is also for me the first\ntime I constructed the \u201cinverse\u201d snowflake with \u201cnegative sign\u201d \u2013 I don't\nknow if that's done in general, but I just wanted to know the result, so here\nit is (the left half of the image, which got constructed by the second call).\n\n<\/p><div class=\"node\"><hr \/>Previous:\u00a0<a rel=\"previous\" accesskey=\"p\" href=\"#Guiling\">Guiling<\/a>,\nUp:\u00a0<a rel=\"up\" accesskey=\"u\" href=\"#Top\">Top<\/a><\/div>\n<h2 class=\"chapter\">4 Next Steps<\/h2>\n<p>I hope this introduction so far gave you an idea about what Guile can do for\nyou and how it basically works.  For what I needed with my first Guile project,\nthe stuff so far was nearly enough; however, that was of course only the tip\nof the iceberg, and there's <em>a lot<\/em> of things you could still\nlearn.\n   <\/p><p>For instance, there are of course a whole lot of other functions in the\nGuile library to work with Scheme objects; nearly any primitive that is\naccessible from Scheme can be used from within a C procedure, also \u2013 there's\neven a routine to capture continuations!\n   <\/p><p>Starting up a REPL as we did is nice, but if you just want to run a user's\nconfiguration script at start-up of your application and then continue with\nyour code, there are also routines that evaluate single Scheme commands or\neven run a file and return afterwards.\n   <\/p><p>And if you decided that one tortoise is not enough, you could introduce some\nsort of \u201ctortoise object\u201d that can be created and passed to the various\nroutines.  This would allow a user to control as many tortoises in parallel as\nshe wants \u2013 such\na user-defined Scheme object is called a <dfn>SMOB<\/dfn> in Guile and is of\ncourse also easy to realize.\n   <\/p><p>If you ever really need any of the stuff I mentioned or are just curious about\nwhat else you can do, I strongly suggest you browse the\n<a href=\"http:\/\/www.gnu.org\/software\/guile\/manual\/\">Guile Reference Manual<\/a>. \nSurely you can find all you need (and much more) there!  And if you need\nfurther help, try asking on the\n<a href=\"http:\/\/www.gnu.org\/software\/guile\/mail\/mail.html\">Mailing Lists<\/a>.\n   <\/p><p>I want to thank David Drysdale for his own \u201cTutorial Introduction to Guile\u201d,\nupon which this document is largly based (in fact, mostly a rewrite of with\nsome updates and \u2013 as I see it, of course \u2013 improvements).\n   <\/p><p>Finally, I hope my write-up was useful to you; in case you have any comments,\nsuggestions, tips or just want to drop me some note whatsoever, I look forward\nto receiving a message.  You can contact me as Daniel Kraft,\n<a href=\"mailto:d@domob.eu\">d@domob.eu<\/a>.  Good luck with Guile and hopefully you also enjoyed\nreading through this tutorial a little bit!\n   <\/p><div class=\"contents\"><h2>Table of Contents<\/h2><ul><li><a name=\"toc_Top\" href=\"#Top\" id=\"toc_Top\">Tutorial Introduction to Guile<\/a>\n<\/li><li><a name=\"toc_Fundamentals\" href=\"#Fundamentals\" id=\"toc_Fundamentals\">1 The Fundamentals<\/a>\n<\/li><li><a name=\"toc_Tortoise\" href=\"#Tortoise\" id=\"toc_Tortoise\">2 The Tortoise<\/a>\n<ul><li><a href=\"#Backend\">2.1 The Core System<\/a>\n<\/li><li><a href=\"#First-Test\">2.2 Testing it out<\/a>\n<\/li><\/ul><\/li><li><a name=\"toc_Guiling\" href=\"#Guiling\" id=\"toc_Guiling\">3 Becoming Guiled<\/a>\n<ul><li><a href=\"#Init-Guile\">3.1 Adding Guile<\/a>\n<\/li><li><a href=\"#Second-Test\">3.2 Compiling with Guile<\/a>\n<\/li><li><a href=\"#Register-Functions\">3.3 Adding the Tortoise<\/a>\n<\/li><li><a href=\"#Third-Test\">3.4 Having Fun with our Tortoise<\/a>\n<\/li><\/ul><\/li><li><a name=\"toc_Further\" href=\"#Further\" id=\"toc_Further\">4 Next Steps<\/a>\n<\/li><\/ul><\/div>","created_at":"2018-11-07T16:22:30+0000","updated_at":"2018-11-07T16:22:30+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":22,"domain_name":"www.gnu.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":86,"title":"Adding Custom Endpoints | REST API Handbook","url":"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/adding-custom-endpoints\/","content":"<div id=\"secondary\" class=\"widget-area\" role=\"complementary\">\n\t\t\n\t\t<div id=\"secondary-content\">\n\t\t\t\t\t\t<aside id=\"search-5\" class=\"widget widget_search\"><div class=\"search-section section clear hide-if-js\">\n\n\n\t\n\t<form role=\"search\" method=\"get\" class=\"searchform searchform-handbook\" action=\"https:\/\/developer.wordpress.org\/rest-api\/\">\n\t\t<div><label for=\"search-field\" class=\"screen-reader-text\">Search for:<\/label>\n\t\t<input type=\"text\" id=\"search-field\" class=\"search-field\" placeholder=\"Search \u2026\" value=\"\" name=\"s\" \/><button class=\"button button-primary button-search\"><i class=\"dashicons dashicons-search\">Search plugins<\/i><\/button>\n\t\n\t<\/div><\/form>\n\n<\/div>\n<\/aside><aside id=\"nav_menu-4\" class=\"widget widget_nav_menu\"><h2 class=\"widget-title\">Chapters<\/h2><div class=\"menu-rest-api-table-of-contents-container\"><ul id=\"menu-rest-api-table-of-contents\" class=\"menu\"><li id=\"menu-item-22855\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-22855\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/\">REST API Handbook<\/a><\/li>\n<li id=\"menu-item-23109\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-has-children menu-item-23109\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/\">Reference<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-23136\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23136\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/posts\/\">Posts<\/a><\/li>\n\t<li id=\"menu-item-23139\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23139\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/post-revisions\/\">Post Revisions<\/a><\/li>\n\t<li id=\"menu-item-23132\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23132\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/categories\/\">Categories<\/a><\/li>\n\t<li id=\"menu-item-23142\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23142\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/tags\/\">Tags<\/a><\/li>\n\t<li id=\"menu-item-23135\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23135\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/pages\/\">Pages<\/a><\/li>\n\t<li id=\"menu-item-23137\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23137\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/comments\/\">Comments<\/a><\/li>\n\t<li id=\"menu-item-23147\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23147\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/taxonomies\/\">Taxonomies<\/a><\/li>\n\t<li id=\"menu-item-23134\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23134\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/media\/\">Media<\/a><\/li>\n\t<li id=\"menu-item-23130\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23130\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/users\/\">Users<\/a><\/li>\n\t<li id=\"menu-item-23133\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23133\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/post-types\/\">Post Types<\/a><\/li>\n\t<li id=\"menu-item-23138\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23138\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/post-statuses\/\">Post Statuses<\/a><\/li>\n\t<li id=\"menu-item-23131\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23131\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/reference\/settings\/\">Settings<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-25309\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-has-children menu-item-25309\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/\">Using the REST API<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-23174\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23174\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/global-parameters\/\">Global Parameters<\/a><\/li>\n\t<li id=\"menu-item-23084\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23084\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/pagination\/\">Pagination<\/a><\/li>\n\t<li id=\"menu-item-23182\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23182\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/linking-and-embedding\/\">Linking and Embedding<\/a><\/li>\n\t<li id=\"menu-item-23087\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23087\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/discovery\/\">Discovery<\/a><\/li>\n\t<li id=\"menu-item-22864\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-22864\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/authentication\/\">Authentication<\/a><\/li>\n\t<li id=\"menu-item-25345\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-25345\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/frequently-asked-questions\/\">Frequently Asked Questions<\/a><\/li>\n\t<li id=\"menu-item-24929\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-24929\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/backbone-javascript-client\/\">Backbone.js Client<\/a><\/li>\n\t<li id=\"menu-item-25310\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-25310\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/client-libraries\/\">Client Libraries<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-25319\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook current-rest-api-handbook-ancestor current-menu-ancestor current-menu-parent current-rest-api-handbook-parent menu-item-has-children menu-item-25319\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/\">Extending the REST API<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-25323\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-25323\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/modifying-responses\/\">Modifying Responses<\/a><\/li>\n\t<li id=\"menu-item-25397\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook current-menu-item menu-item-25397\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/adding-custom-endpoints\/\">Adding Custom Endpoints<\/a><\/li>\n\t<li id=\"menu-item-23096\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23096\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/adding-rest-api-support-for-custom-content-types\/\">Custom Content Types<\/a><\/li>\n\t<li id=\"menu-item-22859\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-22859\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/schema\/\">Schema<\/a><\/li>\n\t<li id=\"menu-item-23187\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-23187\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/glossary\/\">Glossary<\/a><\/li>\n\t<li id=\"menu-item-22862\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-22862\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/routes-and-endpoints\/\">Routes and Endpoints<\/a><\/li>\n\t<li id=\"menu-item-22861\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-22861\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/controller-classes\/\">Controller Classes<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-25439\" class=\"menu-item menu-item-type-post_type menu-item-object-rest-api-handbook menu-item-25439\"><a href=\"https:\/\/developer.wordpress.org\/rest-api\/changelog\/\">Changelog<\/a><\/li>\n<\/ul><\/div><\/aside><\/div>\n\t<\/div>\n\n\n\t\t<main id=\"primary\" class=\"site-main post-25388 rest-api-handbook type-rest-api-handbook status-publish hentry type-handbook\" role=\"main\"><div class=\"breadcrumb-trail breadcrumbs\" itemprop=\"breadcrumb\">\n\t\t\tBrowse: \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\" title=\"WordPress Developer Resources\" rel=\"home\">Home<\/a> \/ \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\/rest-api\/\">REST API Handbook<\/a> \/ \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/\" title=\"Extending the REST API\">Extending the REST API<\/a> \/ Adding Custom Endpoints\n\t\t<\/div>\n\t\t\n\t\t\t\n<h1>Adding Custom Endpoints <a class=\"github-edit\" href=\"https:\/\/github.com\/WP-API\/docs\/edit\/master\/extending-the-rest-api\/adding-custom-endpoints.md\"><img src=\"https:\/\/developer.wordpress.org\/wp-content\/plugins\/wporg-cli\/assets\/images\/github-mark.svg\" alt=\"image\" \/> Edit<\/a><\/h1>\n\n\n<div class=\"table-of-contents\"><h2>Topics<\/h2><ul class=\"items\"><li><a href=\"#bare-basics\">Bare Basics<\/a>\n<ul><li><a href=\"#namespacing\">Namespacing<\/a><\/li>\n<li><a href=\"#arguments\">Arguments<\/a><\/li>\n<li><a href=\"#return-value\">Return Value<\/a><\/li>\n<li><a href=\"#permissions-callback\">Permissions Callback<\/a>\n<\/li><\/ul><\/li>\n<li><a href=\"#the-controller-pattern\">The Controller Pattern<\/a><\/li>\n<li><a href=\"#examples\">Examples<\/a><\/li><\/ul><\/div>\n<p>The WordPress REST API is more than just a set of default routes. It is also a tool for creating custom routes and endpoints. The WordPress front-end provides a default set of URL mappings, but the tools used to create them (e.g. the Rewrites API, as well as the query classes: <code>WP_Query<\/code>, <code>WP_User<\/code>, etc) are also available for creating your own URL mappings, or custom queries.<\/p>\n<p>This document details how to create a totally custom route, with its own endpoints. We\u2019ll first work through a short example, then expand it out to the full controller pattern as used internally.<\/p>\n<h2 class=\"toc-heading\" id=\"bare-basics\" tabindex=\"-1\">Bare Basics <a href=\"#bare-basics\" class=\"anchor\">#Bare Basics<\/a><\/h2>\n<p>So you want to add custom endpoints to the API? Fantastic! Let\u2019s get started with a simple example.<br \/>\nLet\u2019s start with a simple function that looks like this:<\/p>\n<section class=\"source-content\"><div class=\"source-code-container\">\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\n\/**\n * Grab latest post title by an author!\n *\n * @param array $data Options for the function.\n * @return string|null Post title for the latest,\u2028 * or null if none.\n *\/\nfunction my_awesome_func( $data ) {\n  $posts = get_posts( array(\n    'author' =&gt; $data['id'],\n  ) );\n\n  if ( empty( $posts ) ) {\n    return null;\n  }\n\n  return $posts[0]-&gt;post_title;\n}\n<\/pre>\n<\/div>\n<p class=\"source-code-links\"><a href=\"#\" class=\"show-complete-source\">Expand full source code<\/a><a href=\"#\" class=\"less-complete-source\">Collapse full source code<\/a><\/p>\n<\/section><p>To make this available via the API, we need to register a route. This tells the API to respond to a given request with our function. We do this through a function called <code>register_rest_route<\/code>, which should be called in a callback on <code>rest_api_init<\/code> to avoid doing extra work when the API isn\u2019t loaded.<\/p>\n<p>We need to pass in three things to <code>register_rest_route<\/code>: the namespace, the route we want, and the options. We\u2019ll come back to the namespace in a bit, but for now, let\u2019s pick <code>myplugin\/v1<\/code>. We\u2019re going to have the route match anything with <code>\/author\/{id}<\/code>, where <code>{id}<\/code> is an integer.<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\nadd_action( 'rest_api_init', function () {\n  register_rest_route( 'myplugin\/v1', '\/author\/(?P&lt;id&gt;\\d+)', array(\n    'methods' =&gt; 'GET',\n    'callback' =&gt; 'my_awesome_func',\n  ) );\n} );\n<\/pre>\n<p>Right now, we\u2019re only registering the one endpoint for the route. The term \u201croute\u201d refers to the URL, whereas \u201cendpoint\u201d refers to the function behind it that corresponds to a method <em>and<\/em> a URL (for more information, see the <a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/glossary\/\">Glossary<\/a>).<\/p>\n<p>For example, if your site domain is <code>example.com<\/code> and you\u2019ve kept the API path of <code>wp-json<\/code>, then the full URL would be <code>http:\/\/example.com\/wp-json\/myplugin\/v1\/author\/(?P\\d+)<\/code>.<\/p>\n<p>Each route can have any number of endpoints, and for each endpoint, you can define the HTTP methods allowed, a callback function for responding to the request and a permissions callback for creating custom permissions. In addition you can define allowed fields in the request and for each field specify a default value, a sanitization callback, a validation callback, and whether the field is required.<\/p>\n<h3 class=\"toc-heading\" id=\"namespacing\" tabindex=\"-1\">Namespacing <a href=\"#namespacing\" class=\"anchor\">#Namespacing<\/a><\/h3>\n<p>Namespaces are the first part of the URL for the endpoint. They should be used as a vendor\/package prefix to prevent clashes between custom routes. Namespaces allows for two plugins to add a route of the same name, with different functionality.<\/p>\n<p>Namespaces in general should follow the pattern of <code>vendor\/v1<\/code>, where <code>vendor<\/code> is typically your plugin or theme slug, and <code>v1<\/code> represents the first version of the API. If you ever need to break compatibility with new endpoints, you can then bump this to <code>v2<\/code>.<\/p>\n<p>The above scenario, two routes with the same name, from two different plugins, requires all vendors to use a unique namespace. Failing to do so is analogous to a failure to use a vendor function prefix, class prefix and\/ or class namespace in a theme or plugin, which is <strong>very very bad<\/strong>.<\/p>\n<p>An added benefit of using namespaces is that clients can detect support for your custom API. The API index lists out the available namespaces on a site:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n{\n  \"name\": \"WordPress Site\",\n  \"description\": \"Just another WordPress site\",\n  \"url\": \"http:\/\/example.com\/\",\n  \"namespaces\": [\n    \"wp\/v2\",\n    \"vendor\/v1\",\n    \"myplugin\/v1\",\n    \"myplugin\/v2\",\n  ]\n}\n<\/pre>\n<p>If a client wants to check that your API exists on a site, they can check against this list. (For more information, see the <a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/discovery\/\">Discovery guide<\/a>.)<\/p>\n<p class=\"toc-jump\"><a href=\"#top\">Top \u2191<\/a><\/p><h3 class=\"toc-heading\" id=\"arguments\" tabindex=\"-1\">Arguments <a href=\"#arguments\" class=\"anchor\">#Arguments<\/a><\/h3>\n<p>By default, routes receive all arguments passed in from the request. These are merged into a single set of parameters, then added to the Request object, which is passed in as the first parameter to your endpoint:<\/p>\n<section class=\"source-content\"><div class=\"source-code-container\">\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\nfunction my_awesome_func( WP_REST_Request $request ) {\n  \/\/ You can access parameters via direct array access on the object:\n  $param = $request['some_param'];\n\n  \/\/ Or via the helper method:\n  $param = $request-&gt;get_param( 'some_param' );\n\n  \/\/ You can get the combined, merged set of parameters:\n  $parameters = $request-&gt;get_params();\n\n  \/\/ The individual sets of parameters are also available, if needed:\n  $parameters = $request-&gt;get_url_params();\n  $parameters = $request-&gt;get_query_params();\n  $parameters = $request-&gt;get_body_params();\n  $parameters = $request-&gt;get_json_params();\n  $parameters = $request-&gt;get_default_params();\n\n  \/\/ Uploads aren't merged in, but can be accessed separately:\n  $parameters = $request-&gt;get_file_params();\n}\n<\/pre>\n<\/div>\n<p class=\"source-code-links\"><a href=\"#\" class=\"show-complete-source\">Expand full source code<\/a><a href=\"#\" class=\"less-complete-source\">Collapse full source code<\/a><\/p>\n<\/section><p>(To find out exactly how parameters are merged, check the source of <code>WP_REST_Request::get_parameter_order()<\/code>; the basic order is body, query, URL, then defaults.)<\/p>\n<p>Normally, you\u2019ll get every parameter brought in unaltered. However, you can register your arguments when registering your route, which allows you to run sanitization and validation on these.<\/p>\n<p>If the request has the <code>Content-type: application\/json<\/code> header set and valid JSON in the body, <code>get_json_params()<\/code> will return the parsed JSON body as an associative array.<\/p>\n<p>Arguments are defined as a map in the key <code>args<\/code> for each endpoint (next to your <code>callback<\/code> option). This map uses the name of the argument of the key, with the value being a map of options for that argument. This array can contain a key for <code>default<\/code>, <code>required<\/code>, <code>sanitize_callback<\/code> and <code>validate_callback<\/code>.<\/p>\n<ul><li><code>default<\/code>: Used as the default value for the argument, if none is supplied.<\/li>\n<li><code>required<\/code>: If defined as true, and no value is passed for that argument, an error will be returned. No effect if a default value is set, as the argument will always have a value.<\/li>\n<li><code>validate_callback<\/code>: Used to pass a function that will be passed the value of the argument. That function should return true if the value is valid, and false if not.<\/li>\n<li><code>sanitize_callback<\/code>: Used to pass a function that is used to sanitize the value of the argument before passing it to the main callback.<\/li>\n<\/ul><p>Using <code>sanitize_callback<\/code> and <code>validate_callback<\/code> allows the main callback to act only to process the request, and prepare data to be returned using the <code>WP_REST_Response<\/code> class. By using these two callbacks, you will be able to safely assume your inputs are valid and safe when processing.<\/p>\n<p>Taking our previous example, we can ensure the parameter passed in is always a number:<\/p>\n<section class=\"source-content\"><div class=\"source-code-container\">\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\nadd_action( 'rest_api_init', function () {\n  register_rest_route( 'myplugin\/v1', '\/author\/(?P&lt;id&gt;\\d+)', array(\n    'methods' =&gt; 'GET',\n    'callback' =&gt; 'my_awesome_func',\n    'args' =&gt; array(\n      'id' =&gt; array(\n        'validate_callback' =&gt; function($param, $request, $key) {\n          return is_numeric( $param );\n        }\n      ),\n    ),\n  ) );\n} );\n<\/pre>\n<\/div>\n<p class=\"source-code-links\"><a href=\"#\" class=\"show-complete-source\">Expand full source code<\/a><a href=\"#\" class=\"less-complete-source\">Collapse full source code<\/a><\/p>\n<\/section><p>You could also pass in a function name to <code>validate_callback<\/code>, but passing certain functions like <code>is_numeric<\/code> directly will not only throw a warning about having extra parameters passed to it, but will also return <code>NULL<\/code> causing the callback function to be called with invalid data. We hope to <a href=\"https:\/\/core.trac.wordpress.org\/ticket\/34659\">eventually solve this problem in WordPress core<\/a>.<\/p>\n<p>We could also use something like <code>'sanitize_callback' =&gt; 'absint'<\/code> instead, but validation will throw an error, allowing clients to understand what they\u2019re doing wrong. Sanitization is useful when you would rather change the data being input rather than throwing an error (such as invalid HTML).<\/p>\n<p class=\"toc-jump\"><a href=\"#top\">Top \u2191<\/a><\/p><h3 class=\"toc-heading\" id=\"return-value\" tabindex=\"-1\">Return Value <a href=\"#return-value\" class=\"anchor\">#Return Value<\/a><\/h3>\n<p>After your callback is called, the return value is then converted to JSON, and returned to the client. This allows you to return basically any form of data. In our example above, we\u2019re returning either a string or null, which are automatically handled by the API and converted to JSON.<\/p>\n<p>Like any other WordPress function, you can also return a <code>WP_Error<\/code> instance. This error information will be passed along to the client, along with a 500 Internal Service Error status code. You can further customise your error by setting the <code>status<\/code> option in the <code>WP_Error<\/code> instance data to a code, such as <code>400<\/code> for bad input data.<br \/>\nTaking our example from before, we can now return an error instance:<\/p>\n<section class=\"source-content\"><div class=\"source-code-container\">\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\n\/**\n * Grab latest post title by an author!\n *\n * @param array $data Options for the function.\n * @return string|null Post title for the latest,\n * or null if none.\n *\/\nfunction my_awesome_func( $data ) {\n  $posts = get_posts( array(\n    'author' =&gt; $data['id'],\n  ) );\n\n  if ( empty( $posts ) ) {\n    return new WP_Error( 'no_author', 'Invalid author', array( 'status' =&gt; 404 ) );\n  }\n\n  return $posts[0]-&gt;post_title;\n}\n<\/pre>\n<\/div>\n<p class=\"source-code-links\"><a href=\"#\" class=\"show-complete-source\">Expand full source code<\/a><a href=\"#\" class=\"less-complete-source\">Collapse full source code<\/a><\/p>\n<\/section><p>When an author doesn\u2019t have any posts belonging to them, this will return a 404 Not Found error to the client:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\nHTTP\/1.1 404 Not Found\n\n[{\n   \"code\": \"no_author\",\n   \"message\": \"Invalid author\",\n   \"data\": { \"status\": 404 }\n}]\n<\/pre>\n<p>For more advanced usage, you can return a <code>WP_REST_Response<\/code> object. This object \u201cwraps\u201d normal body data, but allows you to return a custom status code, or custom headers. You can also <a href=\"https:\/\/developer.wordpress.org\/rest-api\/using-the-rest-api\/linking-and-embedding\/\">add links to your response<\/a>. The quickest way to use this is via the constructor:<\/p>\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\n$data = array( 'some', 'response', 'data' );\n\n\/\/ Create the response object\n$response = new WP_REST_Response( $data );\n\n\/\/ Add a custom status code\n$response-&gt;set_status( 201 );\n\n\/\/ Add a custom header\n$response-&gt;header( 'Location', 'http:\/\/example.com\/' );\n<\/pre>\n<p>When wrapping existing callbacks, you should always use <code>rest_ensure_response()<\/code> on the return value. This will take raw data returned from an endpoint and automatically turn it into a <code>WP_REST_Response<\/code> for you. (Note that <code>WP_Error<\/code> is not converted to a <code>WP_REST_Response<\/code> to allow proper error handling.)<\/p>\n<p class=\"toc-jump\"><a href=\"#top\">Top \u2191<\/a><\/p><h3 class=\"toc-heading\" id=\"permissions-callback\" tabindex=\"-1\">Permissions Callback <a href=\"#permissions-callback\" class=\"anchor\">#Permissions Callback<\/a><\/h3>\n<p>You can also register a permissions callback for the endpoint. This is a function that checks if the user can perform the action (reading, updating, etc) before the real callback is called. This allows the API to tell the client what actions they can perform on a given URL without needing to attempt the request first.<\/p>\n<p>This callback can be registered as <code>permission_callback<\/code>, again in the endpoint options next to your <code>callback<\/code> option. This callback should return a boolean or a <code>WP_Error<\/code> instance. If this function returns true, the response will be processed. If it returns false, a default error message will be returned and the request will not proceed with processing. If it returns a <code>WP_Error<\/code>, that error will be returned to the client.<\/p>\n<p>The permissions callback is run after remote authentication, which sets the current user. This means you can use <code>current_user_can<\/code> to check if the user that has been authenticated has the appropriate capability for the action, or any other check based on current user ID. Where possible, you should always use <code>current_user_can<\/code>; instead of checking if the user is logged in (authentication), check whether they can perform the action (authorization).<\/p>\n<p>Continuing with our previous example, we can make it so that only editors or above can view this author data. We can check a number of different capabilities here, but the best is <code>edit_others_posts<\/code>, which is really the core of what an editor is. To do this, we just need a callback here:<\/p>\n<section class=\"source-content\"><div class=\"source-code-container\">\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\nadd_action( 'rest_api_init', function () {\n  register_rest_route( 'myplugin\/v1', '\/author\/(?P&lt;id&gt;\\d+)', array(\n    'methods' =&gt; 'GET',\n    'callback' =&gt; 'my_awesome_func',\n    'args' =&gt; array(\n      'id' =&gt; array(\n        'validate_callback' =&gt; 'is_numeric'\n      ),\n    ),\n    'permission_callback' =&gt; function () {\n      return current_user_can( 'edit_others_posts' );\n    }\n  ) );\n} );\n<\/pre>\n<\/div>\n<p class=\"source-code-links\"><a href=\"#\" class=\"show-complete-source\">Expand full source code<\/a><a href=\"#\" class=\"less-complete-source\">Collapse full source code<\/a><\/p>\n<\/section><p>Note that the permission callback also receives the Request object as the first parameter, so you can do checks based on request arguments if you need to.<\/p>\n<p class=\"toc-jump\"><a href=\"#top\">Top \u2191<\/a><\/p><h2 class=\"toc-heading\" id=\"the-controller-pattern\" tabindex=\"-1\">The Controller Pattern <a href=\"#the-controller-pattern\" class=\"anchor\">#The Controller Pattern<\/a><\/h2>\n<p>The controller pattern is a best practice for working with complex endpoints with the API.<\/p>\n<p>It is recommended that you read <a href=\"https:\/\/developer.wordpress.org\/rest-api\/extending-the-rest-api\/controller-classes\/#extending-internal-classes\">\u201cExtending Internal Classes\u201d<\/a> before reading this section. Doing so will familiarize you with the patterns used by the default routes, which is the best practice. While it is not required that the class you use to process your request extends the <code>WP_REST_Controller<\/code> class or a class that extends it, doing so allows you to inherit work done in those classes. In addition, you can rest assured that you\u2019re following best practices based on the controller methods you\u2019re using.<\/p>\n<p>At their core, controllers are nothing more than a set of commonly named methods to match up with REST conventions, along with some handy helpers. Controllers register their routes in a <code>register_routes<\/code> method, respond to requests with <code>get_items<\/code>, <code>get_item<\/code>, <code>create_item<\/code>, <code>update_item<\/code> and <code>delete_item<\/code>, and have similarly named permission check methods. Following this pattern will ensure you don\u2019t miss any steps or functionality in your endpoints.<\/p>\n<p>To use controllers, you first need to subclass the base controller. This gives you a base set of methods, ready for you to add your own behaviour into.<\/p>\n<p>Once we\u2019ve subclassed the controller, we need to instantiate the class to get it working. This should be done inside of a callback hooked into <code>rest_api_init<\/code>, which ensures we only instantiate the class when we need to. The normal controller pattern is to call <code>$controller-&gt;register_routes()<\/code> inside this callback, where the class can then register its endpoints.<\/p>\n<p class=\"toc-jump\"><a href=\"#top\">Top \u2191<\/a><\/p><h2 class=\"toc-heading\" id=\"examples\" tabindex=\"-1\">Examples <a href=\"#examples\" class=\"anchor\">#Examples<\/a><\/h2>\n<p>The following is a \u201cstarter\u201d custom route:<\/p>\n<section class=\"source-content\"><div class=\"source-code-container\">\n<pre class=\"brush: php; title: ; notranslate\" title=\"\">\n&lt;?php\n\nclass Slug_Custom_Route extends WP_REST_Controller {\n\n  \/**\n   * Register the routes for the objects of the controller.\n   *\/\n  public function register_routes() {\n    $version = '1';\n    $namespace = 'vendor\/v' . $version;\n    $base = 'route';\n    register_rest_route( $namespace, '\/' . $base, array(\n      array(\n        'methods'             =&gt; WP_REST_Server::READABLE,\n        'callback'            =&gt; array( $this, 'get_items' ),\n        'permission_callback' =&gt; array( $this, 'get_items_permissions_check' ),\n        'args'                =&gt; array(\n\n        ),\n      ),\n      array(\n        'methods'             =&gt; WP_REST_Server::CREATABLE,\n        'callback'            =&gt; array( $this, 'create_item' ),\n        'permission_callback' =&gt; array( $this, 'create_item_permissions_check' ),\n        'args'                =&gt; $this-&gt;get_endpoint_args_for_item_schema( true ),\n      ),\n    ) );\n    register_rest_route( $namespace, '\/' . $base . '\/(?P&lt;id&gt;[\\d]+)', array(\n      array(\n        'methods'             =&gt; WP_REST_Server::READABLE,\n        'callback'            =&gt; array( $this, 'get_item' ),\n        'permission_callback' =&gt; array( $this, 'get_item_permissions_check' ),\n        'args'                =&gt; array(\n          'context' =&gt; array(\n            'default' =&gt; 'view',\n          ),\n        ),\n      ),\n      array(\n        'methods'             =&gt; WP_REST_Server::EDITABLE,\n        'callback'            =&gt; array( $this, 'update_item' ),\n        'permission_callback' =&gt; array( $this, 'update_item_permissions_check' ),\n        'args'                =&gt; $this-&gt;get_endpoint_args_for_item_schema( false ),\n      ),\n      array(\n        'methods'             =&gt; WP_REST_Server::DELETABLE,\n        'callback'            =&gt; array( $this, 'delete_item' ),\n        'permission_callback' =&gt; array( $this, 'delete_item_permissions_check' ),\n        'args'                =&gt; array(\n          'force' =&gt; array(\n            'default' =&gt; false,\n          ),\n        ),\n      ),\n    ) );\n    register_rest_route( $namespace, '\/' . $base . '\/schema', array(\n      'methods'  =&gt; WP_REST_Server::READABLE,\n      'callback' =&gt; array( $this, 'get_public_item_schema' ),\n    ) );\n  }\n\n  \/**\n   * Get a collection of items\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|WP_REST_Response\n   *\/\n  public function get_items( $request ) {\n    $items = array(); \/\/do a query, call another class, etc\n    $data = array();\n    foreach( $items as $item ) {\n      $itemdata = $this-&gt;prepare_item_for_response( $item, $request );\n      $data[] = $this-&gt;prepare_response_for_collection( $itemdata );\n    }\n\n    return new WP_REST_Response( $data, 200 );\n  }\n\n  \/**\n   * Get one item from the collection\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|WP_REST_Response\n   *\/\n  public function get_item( $request ) {\n    \/\/get parameters from request\n    $params = $request-&gt;get_params();\n    $item = array();\/\/do a query, call another class, etc\n    $data = $this-&gt;prepare_item_for_response( $item, $request );\n\n    \/\/return a response or error based on some conditional\n    if ( 1 == 1 ) {\n      return new WP_REST_Response( $data, 200 );\n    } else {\n      return new WP_Error( 'code', __( 'message', 'text-domain' ) );\n    }\n  }\n\n  \/**\n   * Create one item from the collection\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|WP_REST_Request\n   *\/\n  public function create_item( $request ) {\n    $item = $this-&gt;prepare_item_for_database( $request );\n\n    if ( function_exists( 'slug_some_function_to_create_item' ) ) {\n      $data = slug_some_function_to_create_item( $item );\n      if ( is_array( $data ) ) {\n        return new WP_REST_Response( $data, 200 );\n      }\n    }\n\n    return new WP_Error( 'cant-create', __( 'message', 'text-domain' ), array( 'status' =&gt; 500 ) );\n  }\n\n  \/**\n   * Update one item from the collection\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|WP_REST_Request\n   *\/\n  public function update_item( $request ) {\n    $item = $this-&gt;prepare_item_for_database( $request );\n\n    if ( function_exists( 'slug_some_function_to_update_item' ) ) {\n      $data = slug_some_function_to_update_item( $item );\n      if ( is_array( $data ) ) {\n        return new WP_REST_Response( $data, 200 );\n      }\n    }\n\n    return new WP_Error( 'cant-update', __( 'message', 'text-domain' ), array( 'status' =&gt; 500 ) );\n  }\n\n  \/**\n   * Delete one item from the collection\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|WP_REST_Request\n   *\/\n  public function delete_item( $request ) {\n    $item = $this-&gt;prepare_item_for_database( $request );\n\n    if ( function_exists( 'slug_some_function_to_delete_item' ) ) {\n      $deleted = slug_some_function_to_delete_item( $item );\n      if ( $deleted ) {\n        return new WP_REST_Response( true, 200 );\n      }\n    }\n\n    return new WP_Error( 'cant-delete', __( 'message', 'text-domain' ), array( 'status' =&gt; 500 ) );\n  }\n\n  \/**\n   * Check if a given request has access to get items\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|bool\n   *\/\n  public function get_items_permissions_check( $request ) {\n    \/\/return true; &lt;--use to make readable by all\n    return current_user_can( 'edit_something' );\n  }\n\n  \/**\n   * Check if a given request has access to get a specific item\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|bool\n   *\/\n  public function get_item_permissions_check( $request ) {\n    return $this-&gt;get_items_permissions_check( $request );\n  }\n\n  \/**\n   * Check if a given request has access to create items\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|bool\n   *\/\n  public function create_item_permissions_check( $request ) {\n    return current_user_can( 'edit_something' );\n  }\n\n  \/**\n   * Check if a given request has access to update a specific item\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|bool\n   *\/\n  public function update_item_permissions_check( $request ) {\n    return $this-&gt;create_item_permissions_check( $request );\n  }\n\n  \/**\n   * Check if a given request has access to delete a specific item\n   *\n   * @param WP_REST_Request $request Full data about the request.\n   * @return WP_Error|bool\n   *\/\n  public function delete_item_permissions_check( $request ) {\n    return $this-&gt;create_item_permissions_check( $request );\n  }\n\n  \/**\n   * Prepare the item for create or update operation\n   *\n   * @param WP_REST_Request $request Request object\n   * @return WP_Error|object $prepared_item\n   *\/\n  protected function prepare_item_for_database( $request ) {\n    return array();\n  }\n\n  \/**\n   * Prepare the item for the REST response\n   *\n   * @param mixed $item WordPress representation of the item.\n   * @param WP_REST_Request $request Request object.\n   * @return mixed\n   *\/\n  public function prepare_item_for_response( $item, $request ) {\n    return array();\n  }\n\n  \/**\n   * Get the query params for collections\n   *\n   * @return array\n   *\/\n  public function get_collection_params() {\n    return array(\n      'page'     =&gt; array(\n        'description'       =&gt; 'Current page of the collection.',\n        'type'              =&gt; 'integer',\n        'default'           =&gt; 1,\n        'sanitize_callback' =&gt; 'absint',\n      ),\n      'per_page' =&gt; array(\n        'description'       =&gt; 'Maximum number of items to be returned in result set.',\n        'type'              =&gt; 'integer',\n        'default'           =&gt; 10,\n        'sanitize_callback' =&gt; 'absint',\n      ),\n      'search'   =&gt; array(\n        'description'       =&gt; 'Limit results to those matching a string.',\n        'type'              =&gt; 'string',\n        'sanitize_callback' =&gt; 'sanitize_text_field',\n      ),\n    );\n  }\n}\n<\/pre>\n<\/div>\n<p class=\"source-code-links\"><a href=\"#\" class=\"show-complete-source\">Expand full source code<\/a><a href=\"#\" class=\"less-complete-source\">Collapse full source code<\/a><\/p>\n<\/section><div class=\"bottom-of-entry\">\u00a0<\/div>\n\n\t\t\t\n\t\t<\/main>","created_at":"2018-11-07T16:22:30+0000","updated_at":"2018-11-07T16:22:30+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":16,"domain_name":"developer.wordpress.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":87,"title":"Introduction to Plugin Development | Plugin Developer Handbook","url":"https:\/\/developer.wordpress.org\/plugins\/intro\/","content":"<div class=\"widget-area\" role=\"complementary\">\n\t\t\n\t\t<div>\n\t\t\t\t\t\t<aside id=\"search-4\" class=\"widget widget_search\"><div class=\"search-section section clear hide-if-js\">\n\n\n\t\n\t<form role=\"search\" method=\"get\" class=\"searchform searchform-handbook\" action=\"https:\/\/developer.wordpress.org\/plugins\/\">\n\t\t<div><label for=\"search-field\" class=\"screen-reader-text\">Search for:<\/label>\n\t\t<input type=\"text\" class=\"search-field\" placeholder=\"Search \u2026\" value=\"\" name=\"s\" \/><button class=\"button button-primary button-search\"><i class=\"dashicons dashicons-search\">Search plugins<\/i><\/button>\n\t\n\t<\/div><\/form>\n\n<\/div>\n<\/aside><aside id=\"nav_menu-3\" class=\"widget widget_nav_menu\"><h2 class=\"widget-title\">Chapters<\/h2><div class=\"menu-plugin-table-of-contents-container\"><ul id=\"menu-plugin-table-of-contents\" class=\"menu\"><li id=\"menu-item-5791\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook current-menu-item menu-item-has-children menu-item-5791\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/intro\/\">Introduction to Plugin Development<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-5792\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-5792\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/intro\/what-is-a-plugin\/\">What is a Plugin?<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-10905\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-10905\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/\">Plugin Basics<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-10916\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10916\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/header-requirements\/\">Header Requirements<\/a><\/li>\n\t<li id=\"menu-item-10915\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10915\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/including-a-software-license\/\">Including a Software License<\/a><\/li>\n\t<li id=\"menu-item-10927\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10927\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/activation-deactivation-hooks\/\">Activation \/ Deactivation Hooks<\/a><\/li>\n\t<li id=\"menu-item-10926\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10926\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/uninstall-methods\/\">Uninstall Methods<\/a><\/li>\n\t<li id=\"menu-item-10925\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10925\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/best-practices\/\">Best Practices<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-10935\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-10935\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/security\/\">Plugin Security<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11232\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11232\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/security\/checking-user-capabilities\/\">Checking User Capabilities<\/a><\/li>\n\t<li id=\"menu-item-11231\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11231\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/security\/data-validation\/\">Data Validation<\/a><\/li>\n\t<li id=\"menu-item-11230\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11230\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/security\/securing-input\/\">Securing Input<\/a><\/li>\n\t<li id=\"menu-item-11572\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11572\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/security\/securing-output\/\">Securing Output<\/a><\/li>\n\t<li id=\"menu-item-11229\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11229\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/security\/nonces\/\">Nonces<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-10939\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-10939\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/hooks\/\">Hooks<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-10951\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10951\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/hooks\/actions\/\">Actions<\/a><\/li>\n\t<li id=\"menu-item-10950\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10950\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/hooks\/filters\/\">Filters<\/a><\/li>\n\t<li id=\"menu-item-10949\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10949\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/hooks\/custom-hooks\/\">Custom Hooks<\/a><\/li>\n\t<li id=\"menu-item-10948\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10948\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/hooks\/advanced-topics\/\">Advanced Topics<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-37669\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-37669\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/privacy\/\">Privacy<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-37670\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-37670\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/privacy\/suggesting-text-for-the-site-privacy-policy\/\">Policy Content<\/a><\/li>\n\t<li id=\"menu-item-37671\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-37671\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/privacy\/adding-the-personal-data-exporter-to-your-plugin\/\">Personal Data Export<\/a><\/li>\n\t<li id=\"menu-item-37672\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-37672\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/privacy\/adding-the-personal-data-eraser-to-your-plugin\/\">Personal Data Erasure<\/a><\/li>\n\t<li id=\"menu-item-37673\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-37673\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/privacy\/privacy-related-options-hooks-and-capabilities\/\">Advanced Topics<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-10963\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-10963\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/administration-menus\/\">Administration Menus<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-10965\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10965\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/administration-menus\/top-level-menus\/\">Top-Level Menus<\/a><\/li>\n\t<li id=\"menu-item-10964\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10964\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/administration-menus\/sub-menus\/\">Sub-Menus<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-10995\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-10995\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/shortcodes\/\">Shortcodes<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-10999\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10999\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/shortcodes\/basic-shortcodes\/\">Basic Shortcodes<\/a><\/li>\n\t<li id=\"menu-item-10998\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10998\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/shortcodes\/enclosing-shortcodes\/\">Enclosing Shortcodes<\/a><\/li>\n\t<li id=\"menu-item-10997\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10997\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/shortcodes\/shortcodes-with-parameters\/\">Shortcodes with Parameters<\/a><\/li>\n\t<li id=\"menu-item-10996\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10996\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/shortcodes\/tinymce-enhanced-shortcodes\/\">TinyMCE Enhanced Shortcodes<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11038\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11038\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/settings\/\">Settings<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11041\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11041\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/settings\/settings-api\/\">Settings API<\/a><\/li>\n\t<li id=\"menu-item-11043\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11043\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/settings\/using-settings-api\/\">Using Settings API<\/a><\/li>\n\t<li id=\"menu-item-11042\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11042\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/settings\/options-api\/\">Options API<\/a><\/li>\n\t<li id=\"menu-item-11040\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11040\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/settings\/custom-settings-page\/\">Custom Settings Page<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11054\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11054\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/metadata\/\">Metadata<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11057\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11057\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/metadata\/managing-post-metadata\/\">Managing Post Metadata<\/a><\/li>\n\t<li id=\"menu-item-11056\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11056\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/metadata\/custom-meta-boxes\/\">Custom Meta Boxes<\/a><\/li>\n\t<li id=\"menu-item-11055\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11055\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/metadata\/rendering-post-metadata\/\">Rendering Post Metadata<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11075\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11075\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/post-types\/\">Custom Post Types<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11078\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11078\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/post-types\/registering-custom-post-types\/\">Registering Custom Post Types<\/a><\/li>\n\t<li id=\"menu-item-11077\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11077\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/post-types\/working-with-custom-post-types\/\">Working with Custom Post Types<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-15127\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-15127\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/taxonomies\/\">Taxonomies<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11076\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11076\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/taxonomies\/working-with-custom-taxonomies\/\">Working with Custom Taxonomies<\/a><\/li>\n\t<li id=\"menu-item-15128\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-15128\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/taxonomies\/split-terms-wp-4-2\/\">Split Terms (WP 4.2+)<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11095\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11095\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/users\/\">Users<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11098\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11098\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/users\/working-with-users\/\">Working with Users<\/a><\/li>\n\t<li id=\"menu-item-11097\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11097\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/users\/working-with-user-metadata\/\">Working with User Metadata<\/a><\/li>\n\t<li id=\"menu-item-11096\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11096\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/users\/roles-and-capabilities\/\">Roles and Capabilities<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11094\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11094\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/http-api\/\">HTTP API<\/a><\/li>\n<li id=\"menu-item-11155\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11155\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/javascript\/\">JavaScript, Ajax, &amp; jQuery<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11662\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11662\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/javascript\/jquery\/\">jQuery<\/a><\/li>\n\t<li id=\"menu-item-11661\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11661\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/javascript\/ajax\/\">Ajax<\/a><\/li>\n\t<li id=\"menu-item-11156\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11156\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/javascript\/enqueuing\/\">Server Side PHP and Enqueuing<\/a><\/li>\n\t<li id=\"menu-item-25267\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-25267\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/javascript\/heartbeat-api\/\">Heartbeat API<\/a><\/li>\n\t<li id=\"menu-item-11663\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11663\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/javascript\/summary\/\">Summary<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11168\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11168\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/cron\/\">Cron<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11172\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11172\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/cron\/understanding-wp-cron-scheduling\/\">Understanding WP-Cron Scheduling<\/a><\/li>\n\t<li id=\"menu-item-11171\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11171\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/cron\/scheduling-wp-cron-events\/\">Scheduling WP Cron Events<\/a><\/li>\n\t<li id=\"menu-item-11170\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11170\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/cron\/hooking-into-the-system-task-scheduler\/\">Hooking WP-Cron Into the System Task Scheduler<\/a><\/li>\n\t<li id=\"menu-item-11169\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11169\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/cron\/simple-testing\/\">Simple Testing of WP-Cron<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11182\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11182\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/internationalization\/\">Internationalization<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11183\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11183\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/internationalization\/localization\/\">Localization<\/a><\/li>\n\t<li id=\"menu-item-11184\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11184\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/internationalization\/how-to-internationalize-your-plugin\/\">How to Internationalize Your Plugin<\/a><\/li>\n\t<li id=\"menu-item-15321\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-15321\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/internationalization\/security\/\">Internationalization Security<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11194\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11194\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/wordpress-org\/\">The Plugin Directory<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-15320\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-15320\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/wordpress-org\/detailed-plugin-guidelines\/\">Detailed Plugin Guidelines<\/a><\/li>\n\t<li id=\"menu-item-11197\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11197\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/wordpress-org\/planning-your-plugin\/\">Planning Plugins<\/a><\/li>\n\t<li id=\"menu-item-15317\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-15317\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/wordpress-org\/how-to-use-subversion\/\">Using Subversion<\/a><\/li>\n\t<li id=\"menu-item-37546\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-37546\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/wordpress-org\/alerts-and-warnings\/\">Alerts and Warnings<\/a><\/li>\n\t<li id=\"menu-item-15319\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-15319\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/wordpress-org\/plugin-developer-faq\/\">Developer FAQ<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-11204\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-has-children menu-item-11204\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/developer-tools\/\">Developer Tools<\/a>\n<ul class=\"sub-menu\"><li id=\"menu-item-11206\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11206\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/developer-tools\/debug-bar-and-add-ons\/\">Debug Bar and Add-Ons<\/a><\/li>\n\t<li id=\"menu-item-11205\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-11205\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/developer-tools\/helper-plugins\/\">Helper Plugins<\/a><\/li>\n<\/ul><\/li>\n<li id=\"menu-item-10906\" class=\"menu-item menu-item-type-post_type menu-item-object-plugin-handbook menu-item-10906\"><a href=\"https:\/\/developer.wordpress.org\/plugins\/credits\/\">Credits<\/a><\/li>\n<\/ul><\/div><\/aside><\/div>\n\t<\/div>\n\n\n\t\t<main class=\"site-main post-5787 plugin-handbook type-plugin-handbook status-publish hentry type-handbook\" role=\"main\"><div class=\"breadcrumb-trail breadcrumbs\" itemprop=\"breadcrumb\">\n\t\t\tBrowse: \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\" title=\"WordPress Developer Resources\" rel=\"home\">Home<\/a> \/ \n\t\t\t<a href=\"https:\/\/developer.wordpress.org\/plugins\/\">Plugin Handbook<\/a> \/ Introduction to Plugin Development\n\t\t<\/div>\n\t\t\n\t\t\t\n<h1>Introduction to Plugin Development<\/h1>\n\n\n<p>Welcome to the Plugin Developer Handbook. Whether you\u2019re writing your first plugin or your fiftieth, we hope this resource helps you write the best plugin possible.<\/p>\n<p>The Plugin Developer Handbook covers a variety of topics \u2014 everything from what should be in the plugin header, to security best practices, to tools you can use to build your plugin. It\u2019s also a work in progress \u2014 if you find something missing or incomplete, please edit and make it better.<\/p>\n<p>There are three major components to WordPress:<\/p>\n<ul><li>core<\/li>\n<li>themes<\/li>\n<li>plugins<\/li>\n<\/ul><p>This handbook is about plugins and how they interact with WordPress. It will help you understand how they work and how to create your own.<\/p>\n<h2>Why We Make Plugins<\/h2>\n<p>If there\u2019s one cardinal rule in WordPress development, it\u2019s this: <strong>Don\u2019t touch WordPress core<\/strong>. This means that you don\u2019t edit core WordPress files to add functionality to your site. This is because, when WordPress updates to a new version, it overwrites all the core\u00a0files. Any functionality you want to add should therefore be added through plugins using approved WordPress APIs.<\/p>\n<p>WordPress plugins can be as simple or as complicated as you need them to be, depending on what you want to do. The simplest plugin is a single PHP file. The <a title=\"Hello Dolly Plugin\" href=\"https:\/\/wordpress.org\/plugins\/hello-dolly\/\">Hello Dolly<\/a> plugin is an example of such a plugin. The plugin PHP file just needs a <a href=\"https:\/\/developer.wordpress.org\/plugins\/the-basics\/header-requirements\/\">Plugin Header<\/a>, a couple of PHP functions, and some <a href=\"https:\/\/developer.wordpress.org\/plugins\/hooks\/\">hooks<\/a> to attach your functions to.<\/p>\n<p>Plugins allow you to greatly extend the functionality of WordPress without touching WordPress core itself.<\/p>\n\n\n<div class=\"bottom-of-entry\">\u00a0<\/div>\n\n\t\t\t\n\t\t<\/main>","created_at":"2018-11-07T16:22:30+0000","updated_at":"2018-11-07T16:22:30+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"developer.wordpress.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["php"],"is_public":false,"id":80,"title":"PHP: The Right Way","url":"https:\/\/phptherightway.com\/","content":"<section class=\"chapter\" id=\"welcome\"><p>There\u2019s a lot of outdated information on the Web that leads new PHP users astray,\npropagating bad practices and insecure code. <em>PHP: The Right Way<\/em> is an easy-to-read,\nquick reference for PHP popular coding standards, links to authoritative tutorials\naround the Web and what the contributors consider to be best practices at the present\ntime.<\/p><p><em>There is no canonical way to use PHP<\/em>. This website aims to introduce new PHP\ndevelopers to some topics which they may not discover until it is too late, and aims\nto give seasoned pros some fresh ideas on those topics they\u2019ve been doing for years\nwithout ever reconsidering. This website will also not tell you which tools to use, but\ninstead offer suggestions for multiple options, when possible explaining the differences\nin approach and use-case.<\/p><p>This is a living document and will continue to be updated with more helpful information\nand examples as they become available.<\/p><h2 id=\"translations\">Translations<\/h2><p><em>PHP: The Right Way<\/em> is translated into many different languages:<\/p><h2 id=\"book\">Book<\/h2><p>The most recent version of <em>PHP: The Right Way<\/em> is also available in PDF, EPUB and MOBI formats. <a href=\"https:\/\/leanpub.com\/phptherightway\">Go to Leanpub<\/a><\/p><h2 id=\"how-to-contribute\">How to Contribute<\/h2><p>Help make this website the best resource for new PHP programmers! <a href=\"https:\/\/github.com\/codeguy\/php-the-right-way\/tree\/gh-pages\">Contribute on GitHub<\/a><\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"use_the_current_stable_version\"><h2 id=\"use_the_current_stable_version_title\">Use the Current Stable Version (7.2)<\/h2><p>If you are getting started with PHP, start with the current stable release of <a href=\"http:\/\/php.net\/downloads.php\">PHP 7.2<\/a>. PHP 7.x adds many <a href=\"#language_highlights\">new features<\/a> over the older 5.x versions. The engine has been largely re-written, and PHP is now even quicker than older versions.<\/p><p>You should try to upgrade to the latest stable version quickly - PHP 5.6 <a href=\"http:\/\/php.net\/supported-versions.php\">will not receive security updates beyond 2018<\/a>.  Upgrading is easy, as there are not many <a href=\"http:\/\/php.net\/manual\/migration72.incompatible.php\">backwards compatibility breaks<\/a>. If you are not sure which version a function or feature is in, you can check the PHP documentation on the <a href=\"http:\/\/php.net\/manual\/\">php.net<\/a> website.<\/p><\/section><section class=\"chapter\" id=\"builtin_web_server\"><h2 id=\"builtin_web_server_title\">Built-in web server<\/h2><p>With PHP 5.4 or newer, you can start learning PHP without installing and configuring a full-fledged web server.\nTo start the server, run the following command from your terminal in your project\u2019s web root:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&gt; php -S localhost:8000<\/pre><\/figure><ul><li><a href=\"http:\/\/php.net\/features.commandline.webserver\">Learn about the built-in, command line web server<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"mac_setup\"><h2 id=\"mac_setup_title\">Mac Setup<\/h2><p>macOS comes prepackaged with PHP but it is normally a little behind the latest stable release. There are multiple ways to install the latest PHP version on macOS.<\/p><h3 id=\"install-php-via-homebrew\">Install PHP via Homebrew<\/h3><p><a href=\"https:\/\/brew.sh\/\">Homebrew<\/a> is a package manager for macOS that helps you easily install PHP and various extensions. The Homebrew core repository provides \u201cformulae\u201d for PHP 5.6, 7.0, 7.1, and 7.2. Install the latest version with this command:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>brew install php@7.2\n<\/pre><\/div><\/div><p>You can switch between Homebrew PHP versions by modifying your <code class=\"highlighter-rouge\">PATH<\/code> variable. Alternatively, you can use <a href=\"https:\/\/github.com\/philcook\/brew-php-switcher\">brew-php-switcher<\/a> to switch PHP versions automatically.<\/p><h3 id=\"install-php-via-macports\">Install PHP via Macports<\/h3><p>The <a href=\"https:\/\/www.macports.org\/install.php\">MacPorts<\/a> Project is an open-source community initiative to design an\neasy-to-use system for compiling, installing, and upgrading either\ncommand-line, X11 or Aqua based open-source software on the OS X operating\nsystem.<\/p><p>MacPorts supports pre-compiled binaries, so you don\u2019t need to recompile every\ndependency from the source tarball files, it saves your life if you don\u2019t\nhave any package installed on your system.<\/p><p>At this point, you can install <code class=\"highlighter-rouge\">php54<\/code>, <code class=\"highlighter-rouge\">php55<\/code>, <code class=\"highlighter-rouge\">php56<\/code>, <code class=\"highlighter-rouge\">php70<\/code> or <code class=\"highlighter-rouge\">php71<\/code> using the <code class=\"highlighter-rouge\">port install<\/code> command, for example:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>sudo port install php56\nsudo port install php71\n<\/pre><\/div><\/div><p>And you can run <code class=\"highlighter-rouge\">select<\/code> command to switch your active PHP:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>sudo port select --set php php71\n<\/pre><\/div><\/div><h3 id=\"install-php-via-phpbrew\">Install PHP via phpbrew<\/h3><p><a href=\"https:\/\/github.com\/phpbrew\/phpbrew\">phpbrew<\/a> is a tool for installing and managing multiple PHP versions. This can be really useful if two different\napplications\/projects require different versions of PHP, and you are not using virtual machines.<\/p><h3 id=\"install-php-via-liips-binary-installer\">Install PHP via Liip\u2019s binary installer<\/h3><p>Another popular option is <a href=\"https:\/\/php-osx.liip.ch\/\">php-osx.liip.ch<\/a> which provides one liner installation methods for versions 5.3 through 7.3.\nIt doesn\u2019t overwrite the PHP binaries installed by Apple, but installs everything in a separate location (\/usr\/local\/php5).<\/p><h3 id=\"compile-from-source\">Compile from Source<\/h3><p>Another option that gives you control over the version of PHP you install, is to <a href=\"https:\/\/secure.php.net\/install.macosx.compile\">compile it yourself<\/a>.\nIn that case be sure to have installed either <a href=\"https:\/\/github.com\/kennethreitz\/osx-gcc-installer\">Xcode<\/a> or Apple\u2019s substitute\n<a href=\"https:\/\/developer.apple.com\/downloads\">\u201cCommand Line Tools for XCode\u201d<\/a> downloadable from Apple\u2019s Mac Developer Center.<\/p><h3 id=\"all-in-one-installers\">All-in-One Installers<\/h3><p>The solutions listed above mainly handle PHP itself, and do not supply things like Apache, Nginx or a SQL server.\n\u201cAll-in-one\u201d solutions such as <a href=\"https:\/\/www.mamp.info\/en\/downloads\/\">MAMP<\/a> and <a href=\"https:\/\/www.apachefriends.org\/index.html\">XAMPP<\/a> will install these other bits of software for\nyou and tie them all together, but ease of setup comes with a trade-off of flexibility.<\/p><\/section><section class=\"chapter\" id=\"windows_setup\"><h2 id=\"windows_setup_title\">Windows Setup<\/h2><p>You can download the binaries from <a href=\"http:\/\/windows.php.net\/download\/\">windows.php.net\/download<\/a>. After the extraction of PHP, it is recommended to set the <a href=\"http:\/\/www.windows-commandline.com\/set-path-command-line\/\">PATH<\/a> to the root of your PHP folder (where php.exe is located) so you can execute PHP from anywhere.<\/p><p>For learning and local development, you can use the built in webserver with PHP 5.4+ so you don\u2019t need to worry about\nconfiguring it. If you would like an \u201call-in-one\u201d which includes a full-blown webserver and MySQL too then tools such\nas the <a href=\"https:\/\/www.microsoft.com\/web\/downloads\/platform.aspx\">Web Platform Installer<\/a>, <a href=\"http:\/\/www.apachefriends.org\/en\/xampp.html\">XAMPP<\/a>, <a href=\"http:\/\/www.easyphp.org\/\">EasyPHP<\/a>, <a href=\"http:\/\/open-server.ru\/\">OpenServer<\/a> and <a href=\"http:\/\/www.wampserver.com\/en\/\">WAMP<\/a> will\nhelp get a Windows development environment up and running fast. That said, these tools will be a little different from\nproduction so be careful of environment differences if you are working on Windows and deploying to Linux.<\/p><p>If you need to run your production system on Windows, then IIS7 will give you the most stable and best performance. You\ncan use <a href=\"http:\/\/phpmanager.codeplex.com\/\">phpmanager<\/a> (a GUI plugin for IIS7) to make configuring and managing PHP simple. IIS7 comes with\nFastCGI built in and ready to go, you just need to configure PHP as a handler. For support and additional resources\nthere is a <a href=\"http:\/\/php.iis.net\/\">dedicated area on iis.net<\/a> for PHP.<\/p><p>Generally running your application on different environment in development and production can lead to strange bugs popping up when you go\nlive. If you are developing on Windows and deploying to Linux (or anything non-Windows) then you should consider using a <a href=\"https:\/\/phptherightway.com\/#virtualization_title\">Virtual Machine<\/a>.<\/p><p>Chris Tankersley has a very helpful blog post on what tools he uses to do <a href=\"http:\/\/ctankersley.com\/2016\/11\/13\/developing-on-windows-2016\/\">PHP development using Windows<\/a>.<\/p><\/section><section class=\"chapter\" id=\"common_directory_structure\"><h2 id=\"common_directory_structure_title\">Common Directory structure<\/h2><p>A common question among those starting out with writing programs for the web is, \u201cwhere do I put my stuff?\u201d Over the years, this answer has consistently been \u201cwhere the <code class=\"highlighter-rouge\">DocumentRoot<\/code> is.\u201d Although this answer is not complete, it\u2019s a great place to start.<\/p><p>For security reasons, configuration files should not be accessible by a site\u2019s visitors; therefore, public scripts are kept in a public directory and private configurations and data are kept outside of that directory.<\/p><p>For each team, CMS, or framework one works in, a standard directory structure is used by each of those entities. However, if one is starting a project alone, knowing which filesystem structure to use can be daunting.<\/p><p><a href=\"https:\/\/twitter.com\/pmjones\">Paul M. Jones<\/a> has done some fantastic research into common practices of tens of thousands of github projects in the realm of PHP. He has compiled a standard file and directory structure, the <a href=\"https:\/\/github.com\/php-pds\/skeleton\">Standard PHP Package Skeleton<\/a>, based on this research. In this directory structure, <code class=\"highlighter-rouge\">DocumentRoot<\/code> should point to <code class=\"highlighter-rouge\">public\/<\/code>, unit tests should be in the <code class=\"highlighter-rouge\">tests\/<\/code> directory, and third party libraries, as installed by <a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">composer<\/a>, belong in the <code class=\"highlighter-rouge\">vendor\/<\/code> directory. For other files and directories, abiding by the <a href=\"https:\/\/github.com\/php-pds\/skeleton\">Standard PHP Package Skeleton<\/a> will make the most sense to contributors of a project.<\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"code_style_guide\"><p>The PHP community is large and diverse, composed of innumerable libraries, frameworks, and components. It is common for\nPHP developers to choose several of these and combine them into a single project. It is important that PHP code adhere\n(as close as possible) to a common code style to make it easy for developers to mix and match various libraries for\ntheir projects.<\/p><p>The <a href=\"https:\/\/www.php-fig.org\/\">Framework Interop Group<\/a> has proposed and approved a series of style recommendations. Not all of them related\nto code-style, but those that do are <a href=\"https:\/\/www.php-fig.org\/psr\/psr-1\/\">PSR-1<\/a>, <a href=\"https:\/\/www.php-fig.org\/psr\/psr-2\/\">PSR-2<\/a> and <a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">PSR-4<\/a>. These\nrecommendations are merely a set of rules that many projects like Drupal, Zend, Symfony, Laravel, CakePHP, phpBB, AWS SDK,\nFuelPHP, Lithium, etc are adopting. You can use them for your own projects, or continue to use your own\npersonal style.<\/p><p>Ideally, you should write PHP code that adheres to a known standard. This could be any combination of PSRs, or one\nof the coding standards made by PEAR or Zend. This means other developers can easily read and work with your code, and\napplications that implement the components can have consistency even when working with lots of third-party code.<\/p><ul><li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-1\/\">Read about PSR-1<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-2\/\">Read about PSR-2<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">Read about PSR-4<\/a><\/li>\n  <li><a href=\"https:\/\/pear.php.net\/manual\/en\/standards.php\">Read about PEAR Coding Standards<\/a><\/li>\n  <li><a href=\"https:\/\/symfony.com\/doc\/current\/contributing\/code\/standards.html\">Read about Symfony Coding Standards<\/a><\/li>\n<\/ul><p>You can use <a href=\"https:\/\/pear.php.net\/package\/PHP_CodeSniffer\/\">PHP_CodeSniffer<\/a> to check code against any one of these recommendations, and plugins for text\neditors like <a href=\"https:\/\/github.com\/benmatselby\/sublime-phpcs\">Sublime Text<\/a> to be given real-time feedback.<\/p><p>You can fix the code layout automatically by using one of the following tools:<\/p><ul><li>One is the <a href=\"https:\/\/cs.sensiolabs.org\/\">PHP Coding Standards Fixer<\/a> which has a very well tested codebase.<\/li>\n  <li>Also, the <a href=\"https:\/\/github.com\/squizlabs\/PHP_CodeSniffer\/wiki\/Fixing-Errors-Automatically\">PHP Code Beautifier and Fixer<\/a> tool which is included with PHP_CodeSniffer can be used to adjust your code accordingly.<\/li>\n<\/ul><p>And you can run phpcs manually from shell:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>phpcs -sw --standard=PSR2 file.php\n<\/pre><\/div><\/div><p>It will show errors and describe how to fix them.\nIt can also be helpful to include this command in a git hook.\nThat way, branches which contain violations against the chosen standard cannot enter the repository until those\nviolations have been fixed.<\/p><p>If you have PHP_CodeSniffer, then you can fix the code layout problems reported by it, automatically, with the\n<a href=\"https:\/\/github.com\/squizlabs\/PHP_CodeSniffer\/wiki\/Fixing-Errors-Automatically\">PHP Code Beautifier and Fixer<\/a>.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>phpcbf -w --standard=PSR2 file.php\n<\/pre><\/div><\/div><p>Another option is to use the <a href=\"https:\/\/cs.sensiolabs.org\/\">PHP Coding Standards Fixer<\/a>.\nIt will show which kind of errors the code structure had before it fixed them.<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>php-cs-fixer fix -v --level=psr2 file.php\n<\/pre><\/div><\/div><p>English is preferred for all symbol names and code infrastructure. Comments may be written in any language easily\nreadable by all current and future parties who may be working on the codebase.<\/p><p>Finally, a good supplementary resource for writing clean PHP code is <a href=\"https:\/\/github.com\/jupeter\/clean-code-php\">Clean Code PHP<\/a>.<\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"programming_paradigms\"><h2 id=\"programming_paradigms_title\">Programming Paradigms<\/h2><p>PHP is a flexible, dynamic language that supports a variety of programming techniques. It has evolved dramatically over\nthe years, notably adding a solid object-oriented model in PHP 5.0 (2004), anonymous functions and namespaces in\nPHP 5.3 (2009), and traits in PHP 5.4 (2012).<\/p><h3 id=\"object-oriented-programming\">Object-oriented Programming<\/h3><p>PHP has a very complete set of object-oriented programming features including support for classes, abstract classes,\ninterfaces, inheritance, constructors, cloning, exceptions, and more.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.oop5\">Read about Object-oriented PHP<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/language.oop5.traits\">Read about Traits<\/a><\/li>\n<\/ul><p>PHP supports first-class functions, meaning that a function can be assigned to a variable. Both user-defined and\nbuilt-in functions can be referenced by a variable and invoked dynamically. Functions can be passed as arguments to\nother functions (a feature called <em>Higher-order Functions<\/em>) and functions can return other functions.<\/p><p>Recursion, a feature that allows a function to call itself, is supported by the language, but most PHP code\nis focused on iteration.<\/p><p>New anonymous functions (with support for closures) are present since PHP 5.3 (2009).<\/p><p>PHP 5.4 added the ability to bind closures to an object\u2019s scope and also improved support for callables such that they\ncan be used interchangeably with anonymous functions in almost all cases.<\/p><ul><li>Continue reading on <a href=\"https:\/\/phptherightway.com\/pages\/Functional-Programming.html\">Functional Programming in PHP<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/functions.anonymous\">Read about Anonymous Functions<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/class.closure\">Read about the Closure class<\/a><\/li>\n  <li><a href=\"https:\/\/wiki.php.net\/rfc\/closures\">More details in the Closures RFC<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/language.types.callable\">Read about Callables<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.call-user-func-array\">Read about dynamically invoking functions with <code class=\"highlighter-rouge\">call_user_func_array()<\/code><\/a><\/li>\n<\/ul><p>PHP supports various forms of meta-programming through mechanisms like the Reflection API and Magic Methods. There are\nmany Magic Methods available like <code class=\"highlighter-rouge\">__get()<\/code>, <code class=\"highlighter-rouge\">__set()<\/code>, <code class=\"highlighter-rouge\">__clone()<\/code>, <code class=\"highlighter-rouge\">__toString()<\/code>, <code class=\"highlighter-rouge\">__invoke()<\/code>, etc. that allow\ndevelopers to hook into class behavior. Ruby developers often say that PHP is lacking <code class=\"highlighter-rouge\">method_missing<\/code>, but it is\navailable as <code class=\"highlighter-rouge\">__call()<\/code> and <code class=\"highlighter-rouge\">__callStatic()<\/code>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.oop5.magic\">Read about Magic Methods<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/intro.reflection\">Read about Reflection<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/language.oop5.overloading\">Read about Overloading<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"namespaces\"><h2 id=\"namespaces_title\">Namespaces<\/h2><p>As mentioned above, the PHP community has a lot of developers creating lots of code. This means that one library\u2019s PHP\ncode might use the same class name as another. When both libraries are used in the same namespace, they collide\nand cause trouble.<\/p><p><em>Namespaces<\/em> solve this problem. As described in the PHP reference manual, namespaces may be compared to operating\nsystem directories that <em>namespace<\/em> files; two files with the same name may co-exist in separate directories. Likewise,\ntwo PHP classes with the same name may co-exist in separate PHP namespaces. It\u2019s as simple as that.<\/p><p>It is important for you to namespace your code so that it may be used by other developers without fear of colliding\nwith other libraries.<\/p><p>One recommended way to use namespaces is outlined in <a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">PSR-4<\/a>, which aims to provide a standard file, class and\nnamespace convention to allow plug-and-play code.<\/p><p>In October 2014 the PHP-FIG deprecated the previous autoloading standard: <a href=\"https:\/\/www.php-fig.org\/psr\/psr-0\/\">PSR-0<\/a>. Both PSR-0 and PSR-4 are still perfectly usable.  The latter requires PHP 5.3, so many PHP 5.2-only projects implement PSR-0.<\/p><p>If you\u2019re going to use an autoloader standard for a new application or package, look into PSR-4.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.namespaces\">Read about Namespaces<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-0\/\">Read about PSR-0<\/a><\/li>\n  <li><a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">Read about PSR-4<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"standard_php_library\"><h2 id=\"standard_php_library_title\">Standard PHP Library<\/h2><p>The Standard PHP Library (SPL) is packaged with PHP and provides a collection of classes and interfaces. It is made up\nprimarily of commonly needed datastructure classes (stack, queue, heap, and so on), and iterators which can traverse\nover these datastructures or your own classes which implement SPL interfaces.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.spl\">Read about the SPL<\/a><\/li>\n  <li><a href=\"https:\/\/www.lynda.com\/PHP-tutorials\/Up-Running-Standard-PHP-Library\/175038-2.html\">SPL video course on Lynda.com(Paid)<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"command_line_interface\"><h2 id=\"command_line_interface_title\">Command Line Interface<\/h2><p>PHP was created to write web applications, but is also useful for scripting command line interface (CLI) programs.\nCommand line PHP programs can help automate common tasks like testing, deployment, and application administration.<\/p><p>CLI PHP programs are powerful because you can use your app\u2019s code directly without having to create and secure a web\nGUI for it. Just be sure <strong>not<\/strong> to put your CLI PHP scripts in your public web root!<\/p><p>Try running PHP from your command line:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&gt; php -i<\/pre><\/figure><p>The <code class=\"highlighter-rouge\">-i<\/code> option will print your PHP configuration just like the <a href=\"https:\/\/secure.php.net\/function.phpinfo\"><code class=\"highlighter-rouge\">phpinfo()<\/code><\/a> function.<\/p><p>The <code class=\"highlighter-rouge\">-a<\/code> option provides an interactive shell, similar to ruby\u2019s IRB or python\u2019s interactive shell. There are a number\nof other useful <a href=\"https:\/\/secure.php.net\/features.commandline.options\">command line options<\/a>, too.<\/p><p>Let\u2019s write a simple \u201cHello, $name\u201d CLI program. To try it out, create a file named <code class=\"highlighter-rouge\">hello.php<\/code>, as below.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nif ($argc !== 2) {\n    echo \"Usage: php hello.php [name].\\n\";\n    exit(1);\n}\n$name = $argv[1];\necho \"Hello, $name\\n\";<\/pre><\/figure><p>PHP sets up two special variables based on the arguments your script is run with. <a href=\"https:\/\/secure.php.net\/reserved.variables.argc\"><code class=\"highlighter-rouge\">$argc<\/code><\/a> is an integer\nvariable containing the argument <em>count<\/em> and <a href=\"https:\/\/secure.php.net\/reserved.variables.argv\"><code class=\"highlighter-rouge\">$argv<\/code><\/a> is an array variable containing each argument\u2019s <em>value<\/em>.\nThe first argument is always the name of your PHP script file, in this case <code class=\"highlighter-rouge\">hello.php<\/code>.<\/p><p>The <code class=\"highlighter-rouge\">exit()<\/code> expression is used with a non-zero number to let the shell know that the command failed. Commonly used\nexit codes can be found <a href=\"https:\/\/www.gsp.com\/cgi-bin\/man.cgi?section=3&amp;topic=sysexits\">here<\/a>.<\/p><p>To run our script, above, from the command line:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&gt; php hello.php\nUsage: php hello.php [name]\n&gt; php hello.php world\nHello, world<\/pre><\/figure><ul><li><a href=\"https:\/\/secure.php.net\/features.commandline.options\">Learn about running PHP from the command line<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"xdebug\"><h2 id=\"xdebug_title\">Xdebug<\/h2><p>One of the most useful tools in software development is a proper debugger. It allows you to trace the execution of your\ncode and monitor the contents of the stack. Xdebug, PHP\u2019s debugger, can be utilized by various IDEs to provide\nBreakpoints and stack inspection. It can also allow tools like PHPUnit and KCacheGrind to perform code coverage\nanalysis and code profiling.<\/p><p>If you find yourself in a bind, willing to resort to <code class=\"highlighter-rouge\">var_dump()<\/code>\/<code class=\"highlighter-rouge\">print_r()<\/code>, and you still can\u2019t find the solution -\nmaybe you need to use the debugger.<\/p><p><a href=\"https:\/\/xdebug.org\/docs\/install\">Installing Xdebug<\/a> can be tricky, but one of its most important features is \u201cRemote Debugging\u201d - if\nyou develop code locally and then test it inside a VM or on another server, Remote Debugging is the feature that you\nwill want to enable right away.<\/p><p>Traditionally, you will modify your Apache VHost or .htaccess file with these values:<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">php_value xdebug.remote_host 192.168.?.?\nphp_value xdebug.remote_port 9000<\/pre><\/figure><p>The \u201cremote host\u201d and \u201cremote port\u201d will correspond to your local computer and the port that you configure your IDE to\nlisten on. Then it\u2019s just a matter of putting your IDE into \u201clisten for connections\u201d mode, and loading the URL:<\/p><div class=\"highlighter-rouge\"><div class=\"highlight\"><pre>http:\/\/your-website.example.com\/index.php?XDEBUG_SESSION_START=1\n<\/pre><\/div><\/div><p>Your IDE will now intercept the current state as the script executes, allowing you to set breakpoints and probe the\nvalues in memory.<\/p><p>Graphical debuggers make it very easy to step through code, inspect variables, and eval code against the live runtime.\nMany IDE\u2019s have built-in or plugin-based support for graphical debugging with Xdebug. MacGDBp is a free, open-source,\nstand-alone Xdebug GUI for Mac.<\/p><ul><li><a href=\"https:\/\/xdebug.org\/docs\/\">Learn more about Xdebug<\/a><\/li>\n  <li><a href=\"https:\/\/www.bluestatic.org\/software\/macgdbp\/\">Learn more about MacGDBp<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"dependency_management\"><p>There are a ton of PHP libraries, frameworks, and components to choose from. Your project will likely use \nseveral of them \u2014 these are project dependencies. Until recently, PHP did not have a good way to manage\nthese project dependencies. Even if you managed them manually, you still had to worry about autoloaders.\nThat is no longer an issue.<\/p><p>Currently there are two major package management systems for PHP - <a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">Composer<\/a> and <a href=\"https:\/\/phptherightway.com\/#pear\">PEAR<\/a>. Composer is currently\nthe most popular package manager for PHP, however for a long time PEAR was the primary package manager in use.\nKnowing PEAR\u2019s history is a good idea, since you may still find references to it even if you never use it.<\/p><\/section><section class=\"chapter\" id=\"composer_and_packagist\"><h2 id=\"composer_and_packagist_title\">Composer and Packagist<\/h2><p>Composer is the recommended dependency manager for PHP. List your project\u2019s dependencies in a <code class=\"highlighter-rouge\">composer.json<\/code> file and,\nwith a few simple commands, Composer will automatically download your project\u2019s dependencies and setup autoloading for\nyou. Composer is analogous to NPM in the node.js world, or Bundler in the Ruby world.<\/p><p>There is a plethora of PHP libraries that are compatible with Composer and ready to be used in your project. These\n\u201cpackages\u201d are listed on <a href=\"https:\/\/packagist.org\/\">Packagist<\/a>, the official repository for Composer-compatible PHP libraries.<\/p><h3 id=\"how-to-install-composer\">How to Install Composer<\/h3><p>The safest way to download composer is by <a href=\"https:\/\/getcomposer.org\/download\/\">following the official instructions<\/a>.\nThis will verify the installer is not corrupt or tampered with.\nThe installer installs a <code class=\"highlighter-rouge\">composer.phar<\/code> binary in your <em>current working directory<\/em>.<\/p><p>We recommend installing Composer <em>globally<\/em> (e.g. a single copy in <code class=\"highlighter-rouge\">\/usr\/local\/bin<\/code>). To do so, run this command next:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">mv composer.phar \/usr\/local\/bin\/composer<\/pre><\/figure><p><strong>Note:<\/strong> If the above fails due to permissions, prefix with <code class=\"highlighter-rouge\">sudo<\/code>.<\/p><p>To run a locally installed Composer you\u2019d use <code class=\"highlighter-rouge\">php composer.phar<\/code>, globally it\u2019s simply <code class=\"highlighter-rouge\">composer<\/code>.<\/p><h4 id=\"installing-on-windows\">Installing on Windows<\/h4><p>For Windows users the easiest way to get up and running is to use the <a href=\"https:\/\/getcomposer.org\/Composer-Setup.exe\">ComposerSetup<\/a> installer, which\nperforms a global install and sets up your <code class=\"highlighter-rouge\">$PATH<\/code> so that you can just call <code class=\"highlighter-rouge\">composer<\/code> from any\ndirectory in your command line.<\/p><h3 id=\"how-to-define-and-install-dependencies\">How to Define and Install Dependencies<\/h3><p>Composer keeps track of your project\u2019s dependencies in a file called <code class=\"highlighter-rouge\">composer.json<\/code>. You can manage it\nby hand if you like, or use Composer itself. The <code class=\"highlighter-rouge\">composer require<\/code> command adds a project dependency\nand if you don\u2019t have a <code class=\"highlighter-rouge\">composer.json<\/code> file, one will be created. Here\u2019s an example that adds <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a>\nas a dependency of your project.<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">composer require twig\/twig:^2.0<\/pre><\/figure><p>Alternatively, the <code class=\"highlighter-rouge\">composer init<\/code> command will guide you through creating a full <code class=\"highlighter-rouge\">composer.json<\/code> file\nfor your project. Either way, once you\u2019ve created your <code class=\"highlighter-rouge\">composer.json<\/code> file you can tell Composer to\ndownload and install your dependencies into the <code class=\"highlighter-rouge\">vendor\/<\/code> directory. This also applies to projects\nyou\u2019ve downloaded that already provide a <code class=\"highlighter-rouge\">composer.json<\/code> file:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">composer install<\/pre><\/figure><p>Next, add this line to your application\u2019s primary PHP file; this will tell PHP to use Composer\u2019s\nautoloader for your project dependencies.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nrequire 'vendor\/autoload.php';<\/pre><\/figure><p>Now you can use your project dependencies, and they\u2019ll be autoloaded on demand.<\/p><h3 id=\"updating-your-dependencies\">Updating your dependencies<\/h3><p>Composer creates a file called <code class=\"highlighter-rouge\">composer.lock<\/code> which stores the exact version of each package it\ndownloaded when you first ran <code class=\"highlighter-rouge\">composer install<\/code>. If you share your project with others,\nensure the <code class=\"highlighter-rouge\">composer.lock<\/code> file is included, so that when they run <code class=\"highlighter-rouge\">composer install<\/code> they\u2019ll\nget the same versions as you.  To update your dependencies, run <code class=\"highlighter-rouge\">composer update<\/code>. Don\u2019t use\n<code class=\"highlighter-rouge\">composer update<\/code> when deploying, only <code class=\"highlighter-rouge\">composer install<\/code>, otherwise you may end up with different\npackage versions on production.<\/p><p>This is most useful when you define your version requirements flexibly. For instance, a version\nrequirement of <code class=\"highlighter-rouge\">~1.8<\/code> means \u201canything newer than <code class=\"highlighter-rouge\">1.8.0<\/code>, but less than <code class=\"highlighter-rouge\">2.0.x-dev<\/code>\u201d. You can also use\nthe <code class=\"highlighter-rouge\">*<\/code> wildcard as in <code class=\"highlighter-rouge\">1.8.*<\/code>. Now Composer\u2019s <code class=\"highlighter-rouge\">composer update<\/code> command will upgrade all your\ndependencies to the newest version that fits the restrictions you define.<\/p><p>To receive notifications about new version releases you can sign up for <a href=\"https:\/\/www.versioneye.com\/\">VersionEye<\/a>, a web service\nthat can monitor your GitHub and BitBucket accounts for <code class=\"highlighter-rouge\">composer.json<\/code> files and send emails with new\npackage releases.<\/p><h3 id=\"checking-your-dependencies-for-security-issues\">Checking your dependencies for security issues<\/h3><p>The <a href=\"https:\/\/security.sensiolabs.org\/\">Security Advisories Checker<\/a> is a web service and a command-line tool, both will examine your <code class=\"highlighter-rouge\">composer.lock<\/code>\nfile and tell you if you need to update any of your dependencies.<\/p><h3 id=\"handling-global-dependencies-with-composer\">Handling global dependencies with Composer<\/h3><p>Composer can also handle global dependencies and their binaries. Usage is straight-forward, all you need\nto do is prefix your command with <code class=\"highlighter-rouge\">global<\/code>. If for example you wanted to install PHPUnit and have it\navailable globally, you\u2019d run the following command:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">composer global require phpunit\/phpunit<\/pre><\/figure><p>This will create a <code class=\"highlighter-rouge\">~\/.composer<\/code> folder where your global dependencies reside. To have the installed\npackages\u2019 binaries available everywhere, you\u2019d then add the <code class=\"highlighter-rouge\">~\/.composer\/vendor\/bin<\/code> folder to your\n<code class=\"highlighter-rouge\">$PATH<\/code> variable.<\/p><ul><li><a href=\"https:\/\/getcomposer.org\/doc\/00-intro.md\">Learn about Composer<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"pear\"><h2 id=\"pear_title\">PEAR<\/h2><p>A veteran package manager that some PHP developers enjoy is <a href=\"https:\/\/pear.php.net\/\">PEAR<\/a>. It behaves similarly to Composer,\nbut has some notable differences.<\/p><p>PEAR requires each package to have a specific structure, which means that the author of the package must prepare it for\nusage with PEAR. Using a project which was not prepared to work with PEAR is not possible.<\/p><p>PEAR installs packages globally, which means after installing them once they are available to all projects on that\nserver. This can be good if many projects rely on the same package with the same version but might lead to problems if\nversion conflicts between two projects arise.<\/p><h3 id=\"how-to-install-pear\">How to install PEAR<\/h3><p>You can install PEAR by downloading the <code class=\"highlighter-rouge\">.phar<\/code> installer and executing it. The PEAR documentation has\ndetailed <a href=\"https:\/\/pear.php.net\/manual\/installation.getting.php\">install instructions<\/a> for every operating system.<\/p><p>If you are using Linux, you can also have a look at your distribution package manager. Debian and Ubuntu,\nfor example, have an apt <code class=\"highlighter-rouge\">php-pear<\/code> package.<\/p><h3 id=\"how-to-install-a-package\">How to install a package<\/h3><p>If the package is listed on the <a href=\"https:\/\/pear.php.net\/packages.php\">PEAR packages list<\/a>, you can install it by specifying the official name:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">pear install foo<\/pre><\/figure><p>If the package is hosted on another channel, you need to <code class=\"highlighter-rouge\">discover<\/code> the channel first and also specify it when\ninstalling. See the <a href=\"https:\/\/pear.php.net\/manual\/guide.users.commandline.channels.php\">Using channel docs<\/a> for more information on this topic.<\/p><ul><li><a href=\"https:\/\/pear.php.net\/\">Learn about PEAR<\/a><\/li>\n<\/ul><h3 id=\"handling-pear-dependencies-with-composer\">Handling PEAR dependencies with Composer<\/h3><p>If you are already using <a href=\"https:\/\/phptherightway.com\/#composer_and_packagist\">Composer<\/a> and you would like to install some PEAR code too, you can use Composer to\nhandle your PEAR dependencies. This example will install code from <code class=\"highlighter-rouge\">pear2.php.net<\/code>:<\/p><figure class=\"highlight\"><pre class=\"language-json\" data-lang=\"json\">{\n    \"repositories\": [\n        {\n            \"type\": \"pear\",\n            \"url\": \"https:\/\/pear2.php.net\"\n        }\n    ],\n    \"require\": {\n        \"pear-pear2\/PEAR2_Text_Markdown\": \"*\",\n        \"pear-pear2\/PEAR2_HTTP_Request\": \"*\"\n    }\n}<\/pre><\/figure><p>The first section <code class=\"highlighter-rouge\">\"repositories\"<\/code> will be used to let Composer know it should \u201cinitialize\u201d (or \u201cdiscover\u201d in PEAR\nterminology) the pear repo. Then the <code class=\"highlighter-rouge\">require<\/code> section will prefix the package name like this:<\/p><blockquote>\n  <p>pear-channel\/Package<\/p>\n<\/blockquote><p>The \u201cpear\u201d prefix is hardcoded to avoid any conflicts, as a pear channel could be the same as another packages vendor\nname for example, then the channel short name (or full URL) can be used to reference which channel the package is in.<\/p><p>When this code is installed it will be available in your vendor directory and automatically available through the\nComposer autoloader:<\/p><blockquote>\n  <p>vendor\/pear-pear2.php.net\/PEAR2_HTTP_Request\/pear2\/HTTP\/Request.php<\/p>\n<\/blockquote><p>To use this PEAR package simply reference it like so:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$request = new pear2\\HTTP\\Request();<\/pre><\/figure><ul><li><a href=\"https:\/\/getcomposer.org\/doc\/05-repositories.md#pear\">Learn more about using PEAR with Composer<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"the_basics\"><h2 id=\"the_basics_title\">The Basics<\/h2><p>PHP is a vast language that allows coders of all levels the ability to produce code not only quickly, but efficiently.\nHowever, while advancing through the language, we often forget the basics that we first learnt (or overlooked) in favor\nof short cuts and\/or bad habits. To help combat this common issue, this section is aimed at reminding coders of the\nbasic coding practices within PHP.<\/p><\/section><section class=\"chapter\" id=\"date_and_time\"><p>PHP has a class named DateTime to help you when reading, writing, comparing or calculating with date and time. There\nare many date and time related functions in PHP besides DateTime, but it provides nice object-oriented interface to\nmost common uses. It can handle time zones, but that is outside this short introduction.<\/p><p>To start working with DateTime, convert raw date and time string to an object with <code class=\"highlighter-rouge\">createFromFormat()<\/code> factory method\nor do <code class=\"highlighter-rouge\">new DateTime<\/code> to get the current date and time. Use <code class=\"highlighter-rouge\">format()<\/code> method to convert DateTime back to a string for\noutput.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$raw = '22. 11. 1968';\n$start = DateTime::createFromFormat('d. m. Y', $raw);\necho 'Start date: ' . $start-&gt;format('Y-m-d') . \"\\n\";<\/pre><\/figure><p>Calculating with DateTime is possible with the DateInterval class. DateTime has methods like <code class=\"highlighter-rouge\">add()<\/code> and <code class=\"highlighter-rouge\">sub()<\/code> that\ntake a DateInterval as an argument. Do not write code that expect same number of seconds in every day, both daylight\nsaving and timezone alterations will break that assumption. Use date intervals instead. To calculate date difference\nuse the <code class=\"highlighter-rouge\">diff()<\/code> method. It will return new DateInterval, which is super easy to display.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ create a copy of $start and add one month and 6 days\n$end = clone $start;\n$end-&gt;add(new DateInterval('P1M6D'));\n$diff = $end-&gt;diff($start);\necho 'Difference: ' . $diff-&gt;format('%m month, %d days (total: %a days)') . \"\\n\";\n\/\/ Difference: 1 month, 6 days (total: 37 days)<\/pre><\/figure><p>On DateTime objects you can use standard comparison:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nif ($start &lt; $end) {\n    echo \"Start is before the end!\\n\";\n}<\/pre><\/figure><p>One last example to demonstrate the DatePeriod class. It is used to iterate over recurring events. It can take two\nDateTime objects, start and end, and the interval for which it will return all events in between.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ output all thursdays between $start and $end\n$periodInterval = DateInterval::createFromDateString('first thursday');\n$periodIterator = new DatePeriod($start, $periodInterval, $end, DatePeriod::EXCLUDE_START_DATE);\nforeach ($periodIterator as $date) {\n    \/\/ output each date in the period\n    echo $date-&gt;format('Y-m-d') . ' ';\n}<\/pre><\/figure><p>A popular PHP API extension is <a href=\"https:\/\/carbon.nesbot.com\/\">Carbon<\/a>. It inherits everything in the DateTime class, so involves minimal code alterations, but extra features include Localization support, further ways to add, subtract and format a DateTime object, plus a means to test your code by simulating a date and time of your choosing.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.datetime\">Read about DateTime<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.date\">Read about date formatting<\/a> (accepted date format string options)<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"design_patterns\"><p>When you are building your application it is helpful to use common patterns in your code and common patterns for the\noverall structure of your project. Using common patterns is helpful because it makes it much easier to manage your code\nand lets other developers quickly understand how everything fits together.<\/p><p>If you use a framework then most of the higher level code and project structure will be based on that framework, so a\nlot of the pattern decisions are made for you. But it is still up to you to pick out the best patterns to follow in the\ncode you build on top of the framework. If, on the other hand, you are not using a framework to build your application\nthen you have to find the patterns that best suit the type and size of application that you\u2019re building.<\/p><p>You can learn more about PHP design patterns and see working examples at:<\/p><p><a href=\"https:\/\/designpatternsphp.readthedocs.io\/\">https:\/\/designpatternsphp.readthedocs.io\/<\/a><\/p><\/section><section class=\"chapter\" id=\"php_and_utf8\"><h2 id=\"php_and_utf8_title\">Working with UTF-8<\/h2><p><em>This section was originally written by <a href=\"https:\/\/alexcabal.com\/\">Alex Cabal<\/a> over at\n<a href=\"https:\/\/phpbestpractices.org\/#utf-8\">PHP Best Practices<\/a> and has been used as the basis for our own UTF-8 advice<\/em>.<\/p><h3 id=\"theres-no-one-liner-be-careful-detailed-and-consistent\">There\u2019s no one-liner. Be careful, detailed, and consistent.<\/h3><p>Right now PHP does not support Unicode at a low level. There are ways to ensure that UTF-8 strings are processed OK,\nbut it\u2019s not easy, and it requires digging in to almost all levels of the web app, from HTML to SQL to PHP. We\u2019ll aim\nfor a brief, practical summary.<\/p><h3 id=\"utf-8-at-the-php-level\">UTF-8 at the PHP level<\/h3><p>The basic string operations, like concatenating two strings and assigning strings to variables, don\u2019t need anything\nspecial for UTF-8. However, most string functions, like <code class=\"highlighter-rouge\">strpos()<\/code> and <code class=\"highlighter-rouge\">strlen()<\/code>, do need special consideration. These\nfunctions often have an <code class=\"highlighter-rouge\">mb_*<\/code> counterpart: for example, <code class=\"highlighter-rouge\">mb_strpos()<\/code> and <code class=\"highlighter-rouge\">mb_strlen()<\/code>. These <code class=\"highlighter-rouge\">mb_*<\/code> strings are made\navailable to you via the <a href=\"https:\/\/secure.php.net\/book.mbstring\">Multibyte String Extension<\/a>, and are specifically designed to operate on Unicode strings.<\/p><p>You must use the <code class=\"highlighter-rouge\">mb_*<\/code> functions whenever you operate on a Unicode string. For example, if you use <code class=\"highlighter-rouge\">substr()<\/code> on a\nUTF-8 string, there\u2019s a good chance the result will include some garbled half-characters. The correct function to use\nwould be the multibyte counterpart, <code class=\"highlighter-rouge\">mb_substr()<\/code>.<\/p><p>The hard part is remembering to use the <code class=\"highlighter-rouge\">mb_*<\/code> functions at all times. If you forget even just once, your Unicode\nstring has a chance of being garbled during further processing.<\/p><p>Not all string functions have an <code class=\"highlighter-rouge\">mb_*<\/code> counterpart. If there isn\u2019t one for what you want to do, then you might be out\nof luck.<\/p><p>You should use the <code class=\"highlighter-rouge\">mb_internal_encoding()<\/code> function at the top of every PHP script you write (or at the top of your\nglobal include script), and the <code class=\"highlighter-rouge\">mb_http_output()<\/code> function right after it if your script is outputting to a browser.\nExplicitly defining the encoding of your strings in every script will save you a lot of headaches down the road.<\/p><p>Additionally, many PHP functions that operate on strings have an optional parameter letting you specify the character\nencoding. You should always explicitly indicate UTF-8 when given the option. For example, <code class=\"highlighter-rouge\">htmlentities()<\/code> has an\noption for character encoding, and you should always specify UTF-8 if dealing with such strings. Note that as of PHP 5.4.0, UTF-8 is the default encoding for <code class=\"highlighter-rouge\">htmlentities()<\/code> and <code class=\"highlighter-rouge\">htmlspecialchars()<\/code>.<\/p><p>Finally, If you are building a distributed application and cannot be certain that the <code class=\"highlighter-rouge\">mbstring<\/code> extension will be\nenabled, then consider using the <a href=\"https:\/\/packagist.org\/packages\/patchwork\/utf8\">patchwork\/utf8<\/a> Composer package. This will use <code class=\"highlighter-rouge\">mbstring<\/code> if it is available, and\nfall back to non UTF-8 functions if not.<\/p><h3 id=\"utf-8-at-the-database-level\">UTF-8 at the Database level<\/h3><p>If your PHP script accesses MySQL, there\u2019s a chance your strings could be stored as non-UTF-8 strings in the database\neven if you follow all of the precautions above.<\/p><p>To make sure your strings go from PHP to MySQL as UTF-8, make sure your database and tables are all set to the\n<code class=\"highlighter-rouge\">utf8mb4<\/code> character set and collation, and that you use the <code class=\"highlighter-rouge\">utf8mb4<\/code> character set in the PDO connection string. See\nexample code below. This is <em>critically important<\/em>.<\/p><p>Note that you must use the <code class=\"highlighter-rouge\">utf8mb4<\/code> character set for complete UTF-8 support, not the <code class=\"highlighter-rouge\">utf8<\/code> character set! See\nFurther Reading for why.<\/p><h3 id=\"utf-8-at-the-browser-level\">UTF-8 at the browser level<\/h3><p>Use the <code class=\"highlighter-rouge\">mb_http_output()<\/code> function to ensure that your PHP script outputs UTF-8 strings to your browser.<\/p><p>The browser will then need to be told by the HTTP response that this page should be considered as UTF-8. Today, it is common to set the character set in the HTTP response header like this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nheader('Content-Type: text\/html; charset=UTF-8')<\/pre><\/figure><p>The historic approach to doing that was to include the <a href=\"http:\/\/htmlpurifier.org\/docs\/enduser-utf8.html\">charset <code class=\"highlighter-rouge\">&lt;meta&gt;<\/code> tag<\/a> in your page\u2019s <code class=\"highlighter-rouge\">&lt;head&gt;<\/code> tag.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ Tell PHP that we're using UTF-8 strings until the end of the script\nmb_internal_encoding('UTF-8');\n\/\/ Tell PHP that we'll be outputting UTF-8 to the browser\nmb_http_output('UTF-8');\n\/\/ Our UTF-8 test string\n$string = '\u00cal s\u00edla erin l\u00fb e-govaned v\u00een.';\n\/\/ Transform the string in some way with a multibyte function\n\/\/ Note how we cut the string at a non-Ascii character for demonstration purposes\n$string = mb_substr($string, 0, 15);\n\/\/ Connect to a database to store the transformed string\n\/\/ See the PDO example in this document for more information\n\/\/ Note the `charset=utf8mb4` in the Data Source Name (DSN)\n$link = new PDO(\n    'mysql:host=your-hostname;dbname=your-db;charset=utf8mb4',\n    'your-username',\n    'your-password',\n    array(\n        PDO::ATTR_ERRMODE =&gt; PDO::ERRMODE_EXCEPTION,\n        PDO::ATTR_PERSISTENT =&gt; false\n    )\n);\n\/\/ Store our transformed string as UTF-8 in our database\n\/\/ Your DB and tables are in the utf8mb4 character set and collation, right?\n$handle = $link-&gt;prepare('insert into ElvishSentences (Id, Body) values (?, ?)');\n$handle-&gt;bindValue(1, 1, PDO::PARAM_INT);\n$handle-&gt;bindValue(2, $string);\n$handle-&gt;execute();\n\/\/ Retrieve the string we just stored to prove it was stored correctly\n$handle = $link-&gt;prepare('select * from ElvishSentences where Id = ?');\n$handle-&gt;bindValue(1, 1, PDO::PARAM_INT);\n$handle-&gt;execute();\n\/\/ Store the result into an object that we'll output later in our HTML\n$result = $handle-&gt;fetchAll(\\PDO::FETCH_OBJ);\nheader('Content-Type: text\/html; charset=UTF-8');\n?&gt;&lt;!doctype html&gt;\n&lt;html&gt;\n    &lt;head&gt;\n        &lt;meta charset=\"UTF-8\"&gt;\n        &lt;title&gt;UTF-8 test page&lt;\/title&gt;\n    &lt;\/head&gt;\n    &lt;body&gt;\n        &lt;?php\n        foreach($result as $row){\n            print($row-&gt;Body);  \/\/ This should correctly output our transformed UTF-8 string to the browser\n        }\n        ?&gt;\n    &lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/figure><h3 id=\"further-reading\">Further reading<\/h3><ul><li><a href=\"https:\/\/secure.php.net\/language.operators.string\">PHP Manual: String Operations<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.strings\">PHP Manual: String Functions<\/a>\n    <ul><li><a href=\"https:\/\/secure.php.net\/function.strpos\"><code class=\"highlighter-rouge\">strpos()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.strlen\"><code class=\"highlighter-rouge\">strlen()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.substr\"><code class=\"highlighter-rouge\">substr()<\/code><\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.mbstring\">PHP Manual: Multibyte String Functions<\/a>\n    <ul><li><a href=\"https:\/\/secure.php.net\/function.mb-strpos\"><code class=\"highlighter-rouge\">mb_strpos()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-strlen\"><code class=\"highlighter-rouge\">mb_strlen()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-substr\"><code class=\"highlighter-rouge\">mb_substr()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-internal-encoding\"><code class=\"highlighter-rouge\">mb_internal_encoding()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.mb-http-output\"><code class=\"highlighter-rouge\">mb_http_output()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.htmlentities\"><code class=\"highlighter-rouge\">htmlentities()<\/code><\/a><\/li>\n      <li><a href=\"https:\/\/secure.php.net\/function.htmlspecialchars\"><code class=\"highlighter-rouge\">htmlspecialchars()<\/code><\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"https:\/\/stackoverflow.com\/questions\/571694\/what-factors-make-php-unicode-incompatible\">Stack Overflow: What factors make PHP Unicode-incompatible?<\/a><\/li>\n  <li><a href=\"https:\/\/stackoverflow.com\/questions\/140728\/best-practices-in-php-and-mysql-with-international-strings\">Stack Overflow: Best practices in PHP and MySQL with international strings<\/a><\/li>\n  <li><a href=\"https:\/\/mathiasbynens.be\/notes\/mysql-utf8mb4\">How to support full Unicode in MySQL databases<\/a><\/li>\n  <li><a href=\"https:\/\/www.sitepoint.com\/bringing-unicode-to-php-with-portable-utf8\/\">Bringing Unicode to PHP with Portable UTF-8<\/a><\/li>\n  <li><a href=\"https:\/\/stackoverflow.com\/questions\/8218230\/php-domdocument-loadhtml-not-encoding-utf-8-correctly\">Stack Overflow: DOMDocument loadHTML does not encode UTF-8 correctly<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"i18n_l10n\"><h2 id=\"i18n_l10n_title\">Internationalization (i18n) and Localization (l10n)<\/h2><p><em>Disclaimer for newcomers: i18n and l10n are numeronyms, a kind of abbreviation where numbers are used to shorten\nwords - in our case, internationalization becomes i18n and localization, l10n.<\/em><\/p><p>First of all, we need to define those two similar concepts and other related things:<\/p><ul><li><strong>Internationalization<\/strong> is when you organize your code so it can be adapted to different languages or regions\nwithout refactorings. This action is usually done once - preferably, at the beginning of the project, or else you will\nprobably need some huge changes in the source!<\/li>\n  <li><strong>Localization<\/strong> happens when you adapt the interface (mainly) by translating contents, based on the i18n work done\nbefore. It usually is done every time a new language or region needs support and is updated when new interface pieces\nare added, as they need to be available in all supported languages.<\/li>\n  <li><strong>Pluralization<\/strong> defines the rules required between distinct languages to interoperate strings containing numbers and \ncounters. For instance, in English when you have only one item, it is singular, and anything different from that is \ncalled plural; plural in this language is indicated by adding an S after some words, and sometimes changes parts of it.\nIn other languages, such as Russian or Serbian, there are two plural forms in addition to the singular - you may even\nfind languages with a total of four, five or six forms, such as Slovenian, Irish or Arabic.<\/li>\n<\/ul><h2 id=\"common-ways-to-implement\">Common ways to implement<\/h2><p>The easiest way to internationalize PHP software is by using array files and using those strings in templates, such as\n<code class=\"highlighter-rouge\">&lt;h1&gt;&lt;?=$TRANS['title_about_page']?&gt;&lt;\/h1&gt;<\/code>. This way is, however, hardly recommended for serious projects, as it poses\nsome maintenance issues along the road - some might appear in the very beginning, such as pluralization. So, please,\ndon\u2019t try this if your project will contain more than a couple of pages.<\/p><p>The most classic way and often taken as reference for i18n and l10n is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gettext\">Unix tool called <code class=\"highlighter-rouge\">gettext<\/code><\/a>. It dates\nback to 1995 and is still a complete implementation for translating software. It is easy enough to get running, while\nstill sporting powerful supporting tools. It is about Gettext we will be talking here. Also, to help you not get messy\nover the command-line, we will be presenting a great GUI application that can be used to easily update your l10n source<\/p><h3 id=\"other-tools\">Other tools<\/h3><p>There are common libraries used that support Gettext and other implementations of i18n. Some of them may seem easier to\ninstall or sport additional features or i18n file formats. In this document, we focus on the tools provided with the\nPHP core, but here we list others for completion:<\/p><ul><li><a href=\"https:\/\/github.com\/auraphp\/Aura.Intl\">aura\/intl<\/a>: Provides internationalization (I18N) tools, specifically package-oriented per-locale message\ntranslation. It uses array formats for message. Does not provide a message extractor, but does provide advanced\nmessage formatting via the <code class=\"highlighter-rouge\">intl<\/code> extension (including pluralized messages).<\/li>\n  <li><a href=\"https:\/\/github.com\/oscarotero\/Gettext\">oscarotero\/Gettext<\/a>: Gettext support with an OO interface; includes improved helper functions, powerful\nextractors for several file formats (some of them not supported natively by the <code class=\"highlighter-rouge\">gettext<\/code> command), and can also export\nto other formats besides <code class=\"highlighter-rouge\">.mo\/.po<\/code> files. Can be useful if you need to integrate your translation files into other\nparts of the system, like a JavaScript interface.<\/li>\n  <li><a href=\"https:\/\/symfony.com\/doc\/current\/components\/translation.html\">symfony\/translation<\/a>: supports a lot of different formats, but recommends using verbose XLIFF\u2019s. Doesn\u2019t\ninclude helper functions nor a built-in extractor, but supports placeholders using <code class=\"highlighter-rouge\">strtr()<\/code> internally.<\/li>\n  <li><a href=\"https:\/\/docs.zendframework.com\/zend-i18n\/translation\">zend\/i18n<\/a>: supports array and INI files, or Gettext formats. Implements a caching layer to save you from\nreading the filesystem every time. It also includes view helpers, and locale-aware input filters and validators.\nHowever, it has no message extractor.<\/li>\n<\/ul><p>Other frameworks also include i18n modules, but those are not available outside of their codebases:<\/p><ul><li><a href=\"https:\/\/laravel.com\/docs\/master\/localization\">Laravel<\/a> supports basic array files, has no automatic extractor but includes a <code class=\"highlighter-rouge\">@lang<\/code> helper for template files.<\/li>\n  <li><a href=\"https:\/\/www.yiiframework.com\/doc\/guide\/2.0\/en\/tutorial-i18n\">Yii<\/a> supports array, Gettext, and database-based translation, and includes a messages extractor. It is backed by the\n<a href=\"https:\/\/secure.php.net\/manual\/intro.intl.php\"><code class=\"highlighter-rouge\">Intl<\/code><\/a> extension, available since PHP 5.3, and based on the <a href=\"http:\/\/www.icu-project.org\">ICU project<\/a>; this enables Yii to run powerful\nreplacements, like spelling out numbers, formatting dates, times, intervals, currency, and ordinals.<\/li>\n<\/ul><p>If you decide to go for one of the libraries that provide no extractors, you may want to use the gettext formats, so\nyou can use the original gettext toolchain (including Poedit) as described in the rest of the chapter.<\/p><h2 id=\"gettext\">Gettext<\/h2><h3 id=\"installation\">Installation<\/h3><p>You might need to install Gettext and the related PHP library by using your package manager, like <code class=\"highlighter-rouge\">apt-get<\/code> or <code class=\"highlighter-rouge\">yum<\/code>.\nAfter installed, enable it by adding <code class=\"highlighter-rouge\">extension=gettext.so<\/code> (Linux\/Unix) or <code class=\"highlighter-rouge\">extension=php_gettext.dll<\/code> (Windows) to\nyour <code class=\"highlighter-rouge\">php.ini<\/code>.<\/p><p>Here we will also be using <a href=\"https:\/\/poedit.net\">Poedit<\/a> to create translation files. You will probably find it in your system\u2019s package\nmanager; it is available for Unix, Mac, and Windows, and can be <a href=\"https:\/\/poedit.net\/download\">downloaded for free on their website<\/a>\nas well.<\/p><h3>Structure<\/h3><h4 id=\"types-of-files\">Types of files<\/h4><p>There are three files you usually deal with while working with gettext. The main ones are PO (Portable Object) and\nMO (Machine Object) files, the first being a list of readable \u201ctranslated objects\u201d and the second, the corresponding\nbinary to be interpreted by gettext when doing localization. There\u2019s also a POT (Template) file, which simply contains\nall existing keys from your source files, and can be used as a guide to generate and update all PO files. Those template\nfiles are not mandatory: depending on the tool you are using to do l10n, you can go just fine with only PO\/MO files.\nYou will always have one pair of PO\/MO files per language and region, but only one POT per domain.<\/p><h3 id=\"domains\">Domains<\/h3><p>There are some cases, in big projects, where you might need to separate translations when the same words convey \ndifferent meaning given a context. In those cases, you split them into different <em>domains<\/em>. They are, basically, named\ngroups of POT\/PO\/MO files, where the filename is the said <em>translation domain<\/em>. Small and medium-sized projects usually,\nfor simplicity, use only one domain; its name is arbitrary, but we will be using \u201cmain\u201d for our code samples.\nIn <a href=\"https:\/\/symfony.com\/doc\/current\/components\/translation.html\">Symfony<\/a> projects, for example, domains are used to separate the translation for validation messages.<\/p><h4 id=\"locale-code\">Locale code<\/h4><p>A locale is simply a code that identifies one version of a language. It is defined following the <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_ISO_639-1_codes\">ISO 639-1<\/a> and \n<a href=\"https:\/\/en.wikipedia.org\/wiki\/ISO_3166-1_alpha-2\">ISO 3166-1 alpha-2<\/a> specs: two lower-case letters for the language, optionally followed by an underline and two\nupper-case letters identifying the country or regional code. For <a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html#Rare-Language-Codes\">rare languages<\/a>, three letters are used.<\/p><p>For some speakers, the country part may seem redundant. In fact, some languages have dialects in different\ncountries, such as Austrian German (<code class=\"highlighter-rouge\">de_AT<\/code>) or Brazilian Portuguese (<code class=\"highlighter-rouge\">pt_BR<\/code>). The second part is used to distinguish\nbetween those dialects - when it is not present, it is taken as a \u201cgeneric\u201d or \u201chybrid\u201d version of the language.<\/p><h3 id=\"directory-structure\">Directory structure<\/h3><p>To use Gettext, we will need to adhere to a specific structure of folders. First, you will need to select an arbitrary\nroot for your l10n files in your source repository. Inside it, you will have a folder for each needed locale, and a\nfixed <code class=\"highlighter-rouge\">LC_MESSAGES<\/code> folder that will contain all your PO\/MO pairs. Example:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">&lt;project root&gt;\n \u251c\u2500 src\/\n \u251c\u2500 templates\/\n \u2514\u2500 locales\/\n    \u251c\u2500 forum.pot\n    \u251c\u2500 site.pot\n    \u251c\u2500 de\/\n    \u2502  \u2514\u2500 LC_MESSAGES\/\n    \u2502     \u251c\u2500 forum.mo\n    \u2502     \u251c\u2500 forum.po\n    \u2502     \u251c\u2500 site.mo\n    \u2502     \u2514\u2500 site.po\n    \u251c\u2500 es_ES\/\n    \u2502  \u2514\u2500 LC_MESSAGES\/\n    \u2502     \u2514\u2500 ...\n    \u251c\u2500 fr\/\n    \u2502  \u2514\u2500 ...\n    \u251c\u2500 pt_BR\/\n    \u2502  \u2514\u2500 ...\n    \u2514\u2500 pt_PT\/\n       \u2514\u2500 ...<\/pre><\/figure><h3 id=\"plural-forms\">Plural forms<\/h3><p>As we said in the introduction, different languages might sport different plural rules. However, gettext saves us from\nthis trouble once again. When creating a new <code class=\"highlighter-rouge\">.po<\/code> file, you will have to declare the <a href=\"http:\/\/docs.translatehouse.org\/projects\/localization-guide\/en\/latest\/l10n\/pluralforms.html\">plural rules<\/a> for that\nlanguage, and translated pieces that are plural-sensitive will have a different form for each of those rules. When\ncalling Gettext in code, you will have to specify the number related to the sentence, and it will work out the correct\nform to use - even using string substitution if needed.<\/p><p>Plural rules include the number of plurals available and a boolean test with <code class=\"highlighter-rouge\">n<\/code> that would define in which rule the\ngiven number falls (starting the count with 0). For example:<\/p><ul><li>Japanese: <code class=\"highlighter-rouge\">nplurals=1; plural=0<\/code> - only one rule<\/li>\n  <li>English: <code class=\"highlighter-rouge\">nplurals=2; plural=(n != 1);<\/code> - two rules, first if N is one, second rule otherwise<\/li>\n  <li>Brazilian Portuguese: <code class=\"highlighter-rouge\">nplurals=2; plural=(n &gt; 1);<\/code> - two rules, second if N is bigger than one, first otherwise<\/li>\n<\/ul><p>Now that you understood the basis of how plural rules works - and if you didn\u2019t, please look at a deeper explanation\non the <a href=\"https:\/\/lingohub.com\/blog\/2013\/07\/php-internationalization-with-gettext-tutorial\/#Plurals\">LingoHub tutorial<\/a> -, you might want to copy the ones you need from a <a href=\"http:\/\/docs.translatehouse.org\/projects\/localization-guide\/en\/latest\/l10n\/pluralforms.html\">list<\/a> instead\nof writing them by hand.<\/p><p>When calling out Gettext to do localization on sentences with counters, you will have to give him the\nrelated number as well. Gettext will work out what rule should be in effect and use the correct localized version.\nYou will need to include in the <code class=\"highlighter-rouge\">.po<\/code> file a different sentence for each plural rule defined.<\/p><h3 id=\"sample-implementation\">Sample implementation<\/h3><p>After all that theory, let\u2019s get a little practical. Here\u2019s an excerpt of a <code class=\"highlighter-rouge\">.po<\/code> file - don\u2019t mind with its format,\nbut with the overall content instead; you will learn how to edit it easily later:<\/p><figure class=\"highlight\"><pre class=\"language-po\" data-lang=\"po\">msgid \"\"\nmsgstr \"\"\n\"Language: pt_BR\\n\"\n\"Content-Type: text\/plain; charset=UTF-8\\n\"\n\"Plural-Forms: nplurals=2; plural=(n &gt; 1);\\n\"\nmsgid \"We are now translating some strings\"\nmsgstr \"N\u00f3s estamos traduzindo algumas strings agora\"\nmsgid \"Hello %1$s! Your last visit was on %2$s\"\nmsgstr \"Ol\u00e1 %1$s! Sua \u00faltima visita foi em %2$s\"\nmsgid \"Only one unread message\"\nmsgid_plural \"%d unread messages\"\nmsgstr[0] \"S\u00f3 uma mensagem n\u00e3o lida\"\nmsgstr[1] \"%d mensagens n\u00e3o lidas\"<\/pre><\/figure><p>The first section works like a header, having the <code class=\"highlighter-rouge\">msgid<\/code> and <code class=\"highlighter-rouge\">msgstr<\/code> especially empty. It describes the file encoding,\nplural forms and other things that are less relevant.\nThe second section translates a simple string from English to\nBrazilian Portuguese, and the third does the same, but leveraging string replacement from <a href=\"https:\/\/secure.php.net\/manual\/function.sprintf.php\"><code class=\"highlighter-rouge\">sprintf<\/code><\/a> so the\ntranslation may contain the user name and visit date.\nThe last section is a sample of pluralization forms, displaying\nthe singular and plural version as <code class=\"highlighter-rouge\">msgid<\/code> in English and their corresponding translations as <code class=\"highlighter-rouge\">msgstr<\/code> 0 and 1\n(following the number given by the plural rule). There, string replacement is used as well so the number can be seen\ndirectly in the sentence, by using <code class=\"highlighter-rouge\">%d<\/code>. The plural forms always have two <code class=\"highlighter-rouge\">msgid<\/code> (singular and plural), so it is\nadvised not to use a complex language as the source of translation.<\/p><h3 id=\"discussion-on-l10n-keys\">Discussion on l10n keys<\/h3><p>As you might have noticed, we are using as source ID the actual sentence in English. That <code class=\"highlighter-rouge\">msgid<\/code> is the same used\nthroughout all your <code class=\"highlighter-rouge\">.po<\/code> files, meaning other languages will have the same format and the same <code class=\"highlighter-rouge\">msgid<\/code> fields but\ntranslated <code class=\"highlighter-rouge\">msgstr<\/code> lines.<\/p><p>Talking about translation keys, there are two main \u201cschools\u201d here:<\/p><ol><li><em><code class=\"highlighter-rouge\">msgid<\/code> as a real sentence<\/em>.\n The main advantages are:\n    <ul><li>if there are pieces of the software untranslated in any given language, the key displayed will still maintain some\n meaning. Example: if you happen to translate by heart from English to Spanish but need help to translate to French,\n you might publish the new page with missing French sentences, and parts of the website would be displayed in English\n instead;<\/li>\n      <li>it is much easier for the translator to understand what\u2019s going on and do a proper translation based on the\n <code class=\"highlighter-rouge\">msgid<\/code>;<\/li>\n      <li>it gives you \u201cfree\u201d l10n for one language - the source one;<\/li>\n      <li>The only disadvantage: if you need to change the actual text, you would need to replace the same <code class=\"highlighter-rouge\">msgid<\/code>\n across several language files.<\/li>\n    <\/ul><\/li>\n  <li><em><code class=\"highlighter-rouge\">msgid<\/code> as a unique, structured key<\/em>.\nIt would describe the sentence role in the application in a structured way, including the template or part where the\nstring is located instead of its content.\n    <ul><li>it is a great way to have the code organized, separating the text content from the template logic.<\/li>\n      <li>however, that could bring problems to the translator that would miss the context. A source language file would be\n needed as a basis for other translations. Example: the developer would ideally have an <code class=\"highlighter-rouge\">en.po<\/code> file, that\n translators would read to understand what to write in <code class=\"highlighter-rouge\">fr.po<\/code> for instance.<\/li>\n      <li>missing translations would display meaningless keys on screen (<code class=\"highlighter-rouge\">top_menu.welcome<\/code> instead of <code class=\"highlighter-rouge\">Hello there, User!<\/code>\n on the said untranslated French page). That is good it as would force translation to be complete before publishing -\n however, bad as translation issues would be remarkably awful in the interface. Some libraries, though, include an\n option to specify a given language as \u201cfallback\u201d, having a similar behavior as the other approach.<\/li>\n    <\/ul><\/li>\n<\/ol><p>The <a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html\">Gettext manual<\/a> favors the first approach as, in general, it is easier for translators and users in\ncase of trouble. That is how we will be working here as well. However, the <a href=\"https:\/\/symfony.com\/doc\/current\/components\/translation\/usage.html#creating-translations\">Symfony documentation<\/a> favors\nkeyword-based translation, to allow for independent changes of all translations without affecting templates as well.<\/p><h3 id=\"everyday-usage\">Everyday usage<\/h3><p>In a typical application, you would use some Gettext functions while writing static text in your pages. Those sentences\nwould then appear in <code class=\"highlighter-rouge\">.po<\/code> files, get translated, compiled into <code class=\"highlighter-rouge\">.mo<\/code> files and then, used by Gettext when rendering\nthe actual interface. Given that, let\u2019s tie together what we have discussed so far in a step-by-step example:<\/p><h4 id=\"1-a-sample-template-file-including-some-different-gettext-calls\">1. A sample template file, including some different gettext calls<\/h4><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php include 'i18n_setup.php' ?&gt;\n&lt;div id=\"header\"&gt;\n    &lt;h1&gt;&lt;?=sprintf(gettext('Welcome, %s!'), $name)?&gt;&lt;\/h1&gt;\n    &lt;!-- code indented this way only for legibility --&gt;\n    &lt;?php if ($unread): ?&gt;\n        &lt;h2&gt;&lt;?=sprintf(\n            ngettext('Only one unread message',\n                     '%d unread messages',\n                     $unread),\n            $unread)?&gt;\n        &lt;\/h2&gt;\n    &lt;?php endif ?&gt;\n&lt;\/div&gt;\n&lt;h1&gt;&lt;?=gettext('Introduction')?&gt;&lt;\/h1&gt;\n&lt;p&gt;&lt;?=gettext('We\\'re now translating some strings')?&gt;&lt;\/p&gt;<\/pre><\/figure><ul><li><a href=\"https:\/\/secure.php.net\/manual\/function.gettext.php\"><code class=\"highlighter-rouge\">gettext()<\/code><\/a> simply translates a <code class=\"highlighter-rouge\">msgid<\/code> into its corresponding <code class=\"highlighter-rouge\">msgstr<\/code> for a given language. There\u2019s also\nthe shorthand function <code class=\"highlighter-rouge\">_()<\/code> that works the same way;<\/li>\n  <li><a href=\"https:\/\/secure.php.net\/manual\/function.ngettext.php\"><code class=\"highlighter-rouge\">ngettext()<\/code><\/a> does the same but with plural rules;<\/li>\n  <li>there\u2019s also <a href=\"https:\/\/secure.php.net\/manual\/function.dgettext.php\"><code class=\"highlighter-rouge\">dgettext()<\/code><\/a> and <a href=\"https:\/\/secure.php.net\/manual\/function.dngettext.php\"><code class=\"highlighter-rouge\">dngettext()<\/code><\/a>, that allows you to override the domain for a single\ncall. More on domain configuration in the next example.<\/li>\n<\/ul><h4 id=\"2-a-sample-setup-file-i18n_setupphp-as-used-above-selecting-the-correct-locale-and-configuring-gettext\">2. A sample setup file (<code class=\"highlighter-rouge\">i18n_setup.php<\/code> as used above), selecting the correct locale and configuring Gettext<\/h4><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/**\n * Verifies if the given $locale is supported in the project\n * @param string $locale\n * @return bool\n *\/\nfunction valid($locale) {\n   return in_array($locale, ['en_US', 'en', 'pt_BR', 'pt', 'es_ES', 'es']);\n}\n\/\/setting the source\/default locale, for informational purposes\n$lang = 'en_US';\nif (isset($_GET['lang']) &amp;&amp; valid($_GET['lang'])) {\n    \/\/ the locale can be changed through the query-string\n    $lang = $_GET['lang'];    \/\/you should sanitize this!\n    setcookie('lang', $lang); \/\/it's stored in a cookie so it can be reused\n} elseif (isset($_COOKIE['lang']) &amp;&amp; valid($_COOKIE['lang'])) {\n    \/\/ if the cookie is present instead, let's just keep it\n    $lang = $_COOKIE['lang']; \/\/you should sanitize this!\n} elseif (isset($_SERVER['HTTP_ACCEPT_LANGUAGE'])) {\n    \/\/ default: look for the languages the browser says the user accepts\n    $langs = explode(',', $_SERVER['HTTP_ACCEPT_LANGUAGE']);\n    array_walk($langs, function (&amp;$lang) { $lang = strtr(strtok($lang, ';'), ['-' =&gt; '_']); });\n    foreach ($langs as $browser_lang) {\n        if (valid($browser_lang)) {\n            $lang = $browser_lang;\n            break;\n        }\n    }\n}\n\/\/ here we define the global system locale given the found language\nputenv(\"LANG=$lang\");\n\/\/ this might be useful for date functions (LC_TIME) or money formatting (LC_MONETARY), for instance\nsetlocale(LC_ALL, $lang);\n\/\/ this will make Gettext look for ..\/locales\/&lt;lang&gt;\/LC_MESSAGES\/main.mo\nbindtextdomain('main', '..\/locales');\n\/\/ indicates in what encoding the file should be read\nbind_textdomain_codeset('main', 'UTF-8');\n\/\/ if your application has additional domains, as cited before, you should bind them here as well\nbindtextdomain('forum', '..\/locales');\nbind_textdomain_codeset('forum', 'UTF-8');\n\/\/ here we indicate the default domain the gettext() calls will respond to\ntextdomain('main');\n\/\/ this would look for the string in forum.mo instead of main.mo\n\/\/ echo dgettext('forum', 'Welcome back!');\n?&gt;<\/pre><\/figure><h4 id=\"3-preparing-translation-for-the-first-run\">3. Preparing translation for the first run<\/h4><p>One of the great advantages Gettext has over custom framework i18n packages is its extensive and powerful file format.\n\u201cOh man, that\u2019s quite hard to understand and edit by hand, a simple array would be easier!\u201d Make no mistake,\napplications like <a href=\"https:\/\/poedit.net\">Poedit<\/a> are here to help - <em>a lot<\/em>. You can get the program from <a href=\"https:\/\/poedit.net\/download\">their website<\/a>,\nit\u2019s free and available for all platforms. It\u2019s a pretty easy tool to get used to, and a very powerful one at the same\ntime - using all features Gettext has available. This guide is based on PoEdit 1.8.<\/p><p>In the first run, you should select \u201cFile &gt; New\u2026\u201d from the menu. You\u2019ll be asked straight ahead for the language:\nhere you can select\/filter the language you want to translate to, or use that format we mentioned before, such as\n<code class=\"highlighter-rouge\">en_US<\/code> or <code class=\"highlighter-rouge\">pt_BR<\/code>.<\/p><p>Now, save the file - using that directory structure we mentioned as well. Then you should click \u201cExtract from sources\u201d,\nand here you\u2019ll configure various settings for the extraction and translation tasks. You\u2019ll be able to find all those\nlater through \u201cCatalog &gt; Properties\u201d:<\/p><ul><li>Source paths: here you must include all folders from the project where <code class=\"highlighter-rouge\">gettext()<\/code> (and siblings) are called - this\nis usually your templates\/views folder(s). This is the only mandatory setting;<\/li>\n  <li>Translation properties:\n    <ul><li>Project name and version, Team and Team\u2019s email address: useful information that goes in the .po file header;<\/li>\n      <li>Plural forms: here go those rules we mentioned before - there\u2019s a link in there with samples as well. You can\n  leave it with the default option most of the time, as PoEdit already includes a handy database of plural rules for\n  many languages.<\/li>\n      <li>Charsets: UTF-8, preferably;<\/li>\n      <li>Source code charset: set here the charset used by your codebase - probably UTF-8 as well, right?<\/li>\n    <\/ul><\/li>\n  <li>Source keywords: The underlying software knows how <code class=\"highlighter-rouge\">gettext()<\/code> and similar function calls look like in several\nprogramming languages, but you might as well create your own translation functions. It will be here you\u2019ll add those\nother methods. This will be discussed later in the \u201cTips\u201d section.<\/li>\n<\/ul><p>After setting those points it will run a scan through your source files to find all the localization calls. After every\nscan PoEdit will display a summary of what was found and what was removed from the source files. New entries will fed\nempty into the translation table, and you\u2019ll start typing in the localized versions of those strings. Save it and a .mo\nfile will be (re)compiled into the same folder and ta-dah: your project is internationalized.<\/p><h4 id=\"4-translating-strings\">4. Translating strings<\/h4><p>As you may have noticed before, there are two main types of localized strings: simple ones and those with plural\nforms. The first ones have simply two boxes: source and localized string. The source string cannot be modified as\nGettext\/Poedit do not include the powers to alter your source files - you should change the source itself and rescan\nthe files. Tip: you may right-click a translation line and it will hint you with the source files and lines where that\nstring is being used.\nOn the other hand, plural form strings include two boxes to show the two source strings, and tabs so you can configure\nthe different final forms.<\/p><p>Whenever you change your sources and need to update the translations, just hit Refresh and Poedit will rescan the code,\nremoving non-existent entries, merging the ones that changed and adding new ones. It may also try to guess some\ntranslations, based on other ones you did. Those guesses and the changed entries will receive a \u201cFuzzy\u201d marker,\nindicating it needs review, appearing golden in the list. It is also useful if you have a translation team and someone\ntries to write something they are not sure about: just mark Fuzzy, and someone else will review later.<\/p><p>Finally, it is advised to leave \u201cView &gt; Untranslated entries first\u201d marked, as it will help you <em>a lot<\/em> to not forget\nany entry. From that menu, you can also open parts of the UI that allow you to leave contextual information for\ntranslators if needed.<\/p><h3 id=\"tips--tricks\">Tips &amp; Tricks<\/h3><h4 id=\"possible-caching-issues\">Possible caching issues<\/h4><p>If you are running PHP as a module on Apache (<code class=\"highlighter-rouge\">mod_php<\/code>), you might face issues with the <code class=\"highlighter-rouge\">.mo<\/code> file being cached. It\nhappens the first time it is read, and then, to update it, you might need to restart the server. On Nginx and PHP5 it\nusually takes only a couple of page refreshes to refresh the translation cache, and on PHP7 it is rarely needed.<\/p><p>As preferred by many people, it is easier to use <code class=\"highlighter-rouge\">_()<\/code> instead of <code class=\"highlighter-rouge\">gettext()<\/code>. Many custom i18n libraries from\nframeworks use something similar to <code class=\"highlighter-rouge\">t()<\/code> as well, to make translated code shorter. However, that is the only function\nthat sports a shortcut. You might want to add in your project some others, such as <code class=\"highlighter-rouge\">__()<\/code> or <code class=\"highlighter-rouge\">_n()<\/code> for <code class=\"highlighter-rouge\">ngettext()<\/code>,\nor maybe a fancy <code class=\"highlighter-rouge\">_r()<\/code> that would join <code class=\"highlighter-rouge\">gettext()<\/code> and <code class=\"highlighter-rouge\">sprintf()<\/code> calls. Other libraries, such as\n<a href=\"https:\/\/github.com\/oscarotero\/Gettext\">oscarotero\u2019s Gettext<\/a> also provide helper functions like these.<\/p><p>In those cases, you\u2019ll need to instruct the Gettext utility on how to extract the strings from those new functions.\nDon\u2019t be afraid; it is very easy. It is just a field in the <code class=\"highlighter-rouge\">.po<\/code> file, or a Settings screen on Poedit. In the editor,\nthat option is inside \u201cCatalog &gt; Properties &gt; Source keywords\u201d. Remember: Gettext already knows the default functions\nfor many languages, so don\u2019t be afraid if that list seems empty. You need to include there the specifications of those\nnew functions, following <a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html#Language-specific-options\">a specific format<\/a>:<\/p><ul><li>if you create something like <code class=\"highlighter-rouge\">t()<\/code> that simply returns the translation for a string, you can specify it as <code class=\"highlighter-rouge\">t<\/code>.\nGettext will know the only function argument is the string to be translated;<\/li>\n  <li>if the function has more than one argument, you can specify in which one the first string is - and if needed, the\nplural form as well. For instance, if we call our function like this: <code class=\"highlighter-rouge\">__('one user', '%d users', $number)<\/code>, the\nspecification would be <code class=\"highlighter-rouge\">__:1,2<\/code>, meaning the first form is the first argument, and the second form is the second\nargument. If your number comes as the first argument instead, the spec would be <code class=\"highlighter-rouge\">__:2,3<\/code>, indicating the first form is\nthe second argument, and so on.<\/li>\n<\/ul><p>After including those new rules in the <code class=\"highlighter-rouge\">.po<\/code> file, a new scan will bring in your new strings just as easy as before.<\/p><h3 id=\"references\">References<\/h3><ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Internationalization_and_localization\">Wikipedia: i18n and l10n<\/a><\/li>\n  <li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gettext\">Wikipedia: Gettext<\/a><\/li>\n  <li><a href=\"https:\/\/lingohub.com\/blog\/2013\/07\/php-internationalization-with-gettext-tutorial\/\">LingoHub: PHP internationalization with gettext tutorial<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/manual\/book.gettext.php\">PHP Manual: Gettext<\/a><\/li>\n  <li><a href=\"https:\/\/www.gnu.org\/software\/gettext\/manual\/gettext.html\">Gettext Manual<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"dependency_injection\"><p>From <a href=\"https:\/\/wikipedia.org\/wiki\/Dependency_injection\">Wikipedia<\/a>:<\/p><blockquote>\n  <p>Dependency injection is a software design pattern that allows the removal of hard-coded dependencies and makes it\npossible to change them, whether at run-time or compile-time.<\/p>\n<\/blockquote><p>This quote makes the concept sound much more complicated than it actually is. Dependency Injection is providing a\ncomponent with its dependencies either through constructor injection, method calls or the setting of properties. It is\nthat simple.<\/p><\/section><section class=\"chapter\" id=\"basic_concept\"><h2 id=\"basic_concept_title\">Basic Concept<\/h2><p>We can demonstrate the concept with a simple, yet naive example.<\/p><p>Here we have a <code class=\"highlighter-rouge\">Database<\/code> class that requires an adapter to speak to the database. We instantiate the adapter in the\nconstructor and create a hard dependency. This makes testing difficult and means the <code class=\"highlighter-rouge\">Database<\/code> class is very tightly\ncoupled to the adapter.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nnamespace Database;\nclass Database\n{\n    protected $adapter;\n    public function __construct()\n    {\n        $this-&gt;adapter = new MySqlAdapter;\n    }\n}\nclass MysqlAdapter {}<\/pre><\/figure><p>This code can be refactored to use Dependency Injection and therefore loosen the dependency.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nnamespace Database;\nclass Database\n{\n    protected $adapter;\n    public function __construct(MySqlAdapter $adapter)\n    {\n        $this-&gt;adapter = $adapter;\n    }\n}\nclass MysqlAdapter {}<\/pre><\/figure><p>Now we are giving the <code class=\"highlighter-rouge\">Database<\/code> class its dependency rather than creating it itself. We could even create a method\nthat would accept an argument of the dependency and set it that way, or if the <code class=\"highlighter-rouge\">$adapter<\/code> property was <code class=\"highlighter-rouge\">public<\/code> we\ncould set it directly.<\/p><\/section><section class=\"chapter\" id=\"complex_problem\"><h2 id=\"complex_problem_title\">Complex Problem<\/h2><p>If you have ever read about Dependency Injection then you have probably seen the terms <em>\u201cInversion of Control\u201d<\/em> or\n<em>\u201cDependency Inversion Principle\u201d<\/em>. These are the complex problems that Dependency Injection solves.<\/p><h3 id=\"inversion-of-control\">Inversion of Control<\/h3><p>Inversion of Control is as it says, \u201cinverting the control\u201d of a system by keeping organizational control entirely\nseparate from our objects. In terms of Dependency Injection, this means loosening our dependencies by controlling and\ninstantiating them elsewhere in the system.<\/p><p>For years, PHP frameworks have been achieving Inversion of Control, however, the question became, which part of control\nare we inverting, and where to? For example, MVC frameworks would generally provide a super object or base controller\nthat other controllers must extend to gain access to its dependencies. This <strong>is<\/strong> Inversion of Control, however,\ninstead of loosening dependencies, this method simply moved them.<\/p><p>Dependency Injection allows us to more elegantly solve this problem by only injecting the dependencies we need, when we\nneed them, without the need for any hard coded dependencies at all.<\/p><h3 id=\"solid\">S.O.L.I.D.<\/h3><h4 id=\"single-responsibility-principle\">Single Responsibility Principle<\/h4><p>The Single Responsibility Principle is about actors and high-level architecture. It states that \u201cA class should have\nonly one reason to change.\u201d This means that every class should <em>only<\/em> have responsibility over a single part of the\nfunctionality provided by the software. The largest benefit of this approach is that it enables improved code\n<em>reusability<\/em>. By designing our class to do just one thing, we can use (or re-use) it in any other program without\nchanging it.<\/p><h4 id=\"openclosed-principle\">Open\/Closed Principle<\/h4><p>The Open\/Closed Principle is about class design and feature extensions. It states that \u201cSoftware entities (classes,\nmodules, functions, etc.) should be open for extension, but closed for modification.\u201d This means that we should design\nour modules, classes and functions in a way that when a new functionality is needed, we should not modify our existing\ncode but rather write new code that will be used by existing code. Practically speaking, this means that we should write\nclasses that implement and adhere to <em>interfaces<\/em>, then type-hint against those interfaces instead of specific classes.<\/p><p>The largest benefit of this approach is that we can very easily extend our code with support for something new without\nhaving to modify existing code, meaning that we can reduce QA time, and the risk for negative impact to the application\nis substantially reduced. We can deploy new code, faster, and with more confidence.<\/p><h4 id=\"liskov-substitution-principle\">Liskov Substitution Principle<\/h4><p>The Liskov Substitution Principle is about subtyping and inheritance. It states that \u201cChild classes should never break\nthe parent class\u2019 type definitions.\u201d Or, in Robert C. Martin\u2019s words, \u201cSubtypes must be substitutable for their base\ntypes.\u201d<\/p><p>For example, if we have a <code class=\"highlighter-rouge\">FileInterface<\/code> interface which defines an <code class=\"highlighter-rouge\">embed()<\/code> method, and we have <code class=\"highlighter-rouge\">Audio<\/code> and <code class=\"highlighter-rouge\">Video<\/code>\nclasses which both implement the <code class=\"highlighter-rouge\">embed()<\/code> method, then we can expect that the usage of the <code class=\"highlighter-rouge\">embed()<\/code> method will always\ndo the thing that we intend. If we later create a <code class=\"highlighter-rouge\">PDF<\/code> class or a <code class=\"highlighter-rouge\">Gist<\/code> class which implement the <code class=\"highlighter-rouge\">FileInterface<\/code>\ninterface, we will already know and understand what the <code class=\"highlighter-rouge\">embed()<\/code> method will do. The largest benefit of this approach\nis that we have the ability to build flexible and easily-configurable programs, because when we change one object of a\ntype (e.g., <code class=\"highlighter-rouge\">FileInterface<\/code>) to another we don\u2019t need to change anything else in our program.<\/p><h4 id=\"interface-segregation-principle\">Interface Segregation Principle<\/h4><p>The Interface Segregation Principle (ISP) is about <em>business-logic-to-clients<\/em> communication. It states that \u201cNo client\nshould be forced to depend on methods it does not use.\u201d This means that instead of having a single monolithic interface\nthat all conforming classes need to implement, we should instead provide a set of smaller, concept-specific interfaces\nthat a conforming class implements one or more of.<\/p><p>For example, a <code class=\"highlighter-rouge\">Car<\/code> or <code class=\"highlighter-rouge\">Bus<\/code> class would be interested in a <code class=\"highlighter-rouge\">steeringWheel()<\/code> method, but a <code class=\"highlighter-rouge\">Motorcycle<\/code> or <code class=\"highlighter-rouge\">Tricycle<\/code>\nclass would not. Conversely, a <code class=\"highlighter-rouge\">Motorcycle<\/code> or <code class=\"highlighter-rouge\">Tricycle<\/code> class would be interested in a <code class=\"highlighter-rouge\">handlebars()<\/code> method, but a\n<code class=\"highlighter-rouge\">Car<\/code> or <code class=\"highlighter-rouge\">Bus<\/code> class would not. There is no need to have all of these types of vehicles implement support for both\n<code class=\"highlighter-rouge\">steeringWheel()<\/code> as well as <code class=\"highlighter-rouge\">handlebars()<\/code>, so we should break-apart the source interface.<\/p><h4 id=\"dependency-inversion-principle\">Dependency Inversion Principle<\/h4><p>The Dependency Inversion Principle is about removing hard-links between discrete classes so that new functionality can\nbe leveraged by passing a different class. It states that one should <em>\u201cDepend on Abstractions. Do not depend on\nconcretions.\u201d<\/em>. Put simply, this means our dependencies should be interfaces\/contracts or abstract classes rather than\nconcrete implementations. We can easily refactor the above example to follow this principle.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nnamespace Database;\nclass Database\n{\n    protected $adapter;\n    public function __construct(AdapterInterface $adapter)\n    {\n        $this-&gt;adapter = $adapter;\n    }\n}\ninterface AdapterInterface {}\nclass MysqlAdapter implements AdapterInterface {}<\/pre><\/figure><p>There are several benefits to the <code class=\"highlighter-rouge\">Database<\/code> class now depending on an interface rather than a concretion.<\/p><p>Consider that we are working in a team and the adapter is being worked on by a colleague. In our first example, we\nwould have to wait for said colleague to finish the adapter before we could properly mock it for our unit tests. Now\nthat the dependency is an interface\/contract we can happily mock that interface knowing that our colleague will build\nthe adapter based on that contract.<\/p><p>An even bigger benefit to this method is that our code is now much more scalable. If a year down the line we decide\nthat we want to migrate to a different type of database, we can write an adapter that implements the original interface\nand injects that instead, no more refactoring would be required as we can ensure that the adapter follows the contract\nset by the interface.<\/p><\/section><section class=\"chapter\" id=\"containers\"><h2 id=\"containers_title\">Containers<\/h2><p>The first thing you should understand about Dependency Injection Containers is that they are not the same thing as\nDependency Injection. A container is a convenience utility that helps us implement Dependency Injection, however, they\ncan be and often are misused to implement an anti-pattern, Service Location. Injecting a DI container as a Service\nLocator in to your classes arguably creates a harder dependency on the container than the dependency you are replacing.\nIt also makes your code much less transparent and ultimately harder to test.<\/p><p>Most modern frameworks have their own Dependency Injection Container that allows you to wire your dependencies together\nthrough configuration. What this means in practice is that you can write application code that is as clean and de-\ncoupled as the framework it is built on.<\/p><\/section><section class=\"chapter\" id=\"further_reading\"><h2 id=\"further_reading_title\">Further Reading<\/h2><ul><li><a href=\"http:\/\/ralphschindler.com\/2011\/05\/18\/learning-about-dependency-injection-and-php\">Learning about Dependency Injection and PHP<\/a><\/li>\n  <li><a href=\"http:\/\/fabien.potencier.org\/article\/11\/what-is-dependency-injection\">What is Dependency Injection?<\/a><\/li>\n  <li><a href=\"https:\/\/mwop.net\/blog\/260-Dependency-Injection-An-analogy.html\">Dependency Injection: An analogy<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/dependency-injection-huh--net-26903\">Dependency Injection: Huh?<\/a><\/li>\n  <li><a href=\"https:\/\/medium.com\/philipobenito\/dependency-injection-as-a-tool-for-testing-902c21c147f1\">Dependency Injection as a tool for testing<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"databases\"><p>Many times your PHP code will use a database to persist information. You have a few options to connect and interact\nwith your database. The recommended option <strong>until PHP 5.1.0<\/strong> was to use native drivers such as <a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a>, <a href=\"https:\/\/secure.php.net\/pgsql\">pgsql<\/a>,\n<a href=\"https:\/\/secure.php.net\/mssql\">mssql<\/a>, etc.<\/p><p>Native drivers are great if you are only using <em>one<\/em> database in your application, but if, for example, you are using\nMySQL and a little bit of MSSQL, or you need to connect to an Oracle database, then you will not be able to use the\nsame drivers. You\u2019ll need to learn a brand new API for each database \u2014 and that can get silly.<\/p><\/section><section class=\"chapter\" id=\"mysql_extension\"><h2 id=\"mysql_extension_title\">MySQL Extension<\/h2><p>The <a href=\"https:\/\/secure.php.net\/mysqli\">mysql<\/a> extension for PHP is incredibly old and has been superseded by two other extensions:<\/p><ul><li><a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/pdo\">pdo<\/a><\/li>\n<\/ul><p>Not only did development stop long ago on <a href=\"https:\/\/secure.php.net\/mysqli\">mysql<\/a>, but it was <a href=\"https:\/\/secure.php.net\/migration55.deprecated\">deprecated as of PHP 5.5.0<\/a>, and <strong>has been <a href=\"https:\/\/secure.php.net\/manual\/migration70.removed-exts-sapis.php\">officially removed in PHP 7.0<\/a><\/strong>.<\/p><p>To save digging into your <code class=\"highlighter-rouge\">php.ini<\/code> settings to see which module you are using, one option is to search for <code class=\"highlighter-rouge\">mysql_*<\/code>\nin your editor of choice. If any functions such as <code class=\"highlighter-rouge\">mysql_connect()<\/code> and <code class=\"highlighter-rouge\">mysql_query()<\/code> show up, then <code class=\"highlighter-rouge\">mysql<\/code> is\nin use.<\/p><p>Even if you are not using PHP 7.x yet, failing to consider this upgrade as soon as possible will lead to greater\nhardship when the PHP 7.x upgrade does come about. The best option is to replace mysql usage with <a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a> or <a href=\"https:\/\/secure.php.net\/pdo\">PDO<\/a> in\nyour applications within your own development schedules so you won\u2019t be rushed later on.<\/p><p><strong>If you are upgrading from <a href=\"https:\/\/secure.php.net\/mysqli\">mysql<\/a> to <a href=\"https:\/\/secure.php.net\/mysqli\">mysqli<\/a>, beware lazy upgrade guides that suggest you can simply find and replace <code class=\"highlighter-rouge\">mysql_*<\/code> with <code class=\"highlighter-rouge\">mysqli_*<\/code>. Not only is that a gross oversimplification, it misses out on the advantages that mysqli provides, such as parameter binding, which is also offered in <a href=\"https:\/\/secure.php.net\/pdo\">PDO<\/a>.<\/strong><\/p><ul><li><a href=\"https:\/\/secure.php.net\/mysqlinfo.api.choosing\">PHP: Choosing an API for MySQL<\/a><\/li>\n  <li><a href=\"http:\/\/wiki.hashphp.org\/PDO_Tutorial_for_MySQL_Developers\">PDO Tutorial for MySQL Developers<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"pdo_extension\"><h2 id=\"pdo_extension_title\">PDO Extension<\/h2><p><a href=\"https:\/\/secure.php.net\/pdo\">PDO<\/a> is a database connection abstraction library \u2014 built into PHP since 5.1.0 \u2014 that provides a common\ninterface to talk with many different databases. For example, you can use basically identical code to interface with\nMySQL or SQLite:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ PDO + MySQL\n$pdo = new PDO('mysql:host=example.com;dbname=database', 'user', 'password');\n$statement = $pdo-&gt;query(\"SELECT some_field FROM some_table\");\n$row = $statement-&gt;fetch(PDO::FETCH_ASSOC);\necho htmlentities($row['some_field']);\n\/\/ PDO + SQLite\n$pdo = new PDO('sqlite:\/path\/db\/foo.sqlite');\n$statement = $pdo-&gt;query(\"SELECT some_field FROM some_table\");\n$row = $statement-&gt;fetch(PDO::FETCH_ASSOC);\necho htmlentities($row['some_field']);<\/pre><\/figure><p>PDO will not translate your SQL queries or emulate missing features; it is purely for connecting to multiple types of\ndatabase with the same API.<\/p><p>More importantly, <code class=\"highlighter-rouge\">PDO<\/code> allows you to safely inject foreign input (e.g. IDs) into your SQL queries without worrying\nabout database SQL injection attacks.\nThis is possible using PDO statements and bound parameters.<\/p><p>Let\u2019s assume a PHP script receives a numeric ID as a query parameter. This ID should be used to fetch a user record\nfrom a database. This is the <code class=\"highlighter-rouge\">wrong<\/code> way to do this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$pdo = new PDO('sqlite:\/path\/db\/users.db');\n$pdo-&gt;query(\"SELECT name FROM users WHERE id = \" . $_GET['id']); \/\/ &lt;-- NO!<\/pre><\/figure><p>This is terrible code. You are inserting a raw query parameter into a SQL query. This will get you hacked in a\nheartbeat, using a practice called <a href=\"http:\/\/wiki.hashphp.org\/Validation\">SQL Injection<\/a>. Just imagine if a hacker passes in an inventive <code class=\"highlighter-rouge\">id<\/code> parameter by\ncalling a URL like <code class=\"highlighter-rouge\">http:\/\/domain.com\/?id=1%3BDELETE+FROM+users<\/code>. This will set the <code class=\"highlighter-rouge\">$_GET['id']<\/code> variable to <code class=\"highlighter-rouge\">1;DELETE\nFROM users<\/code> which will delete all of your users! Instead, you should sanitize the ID input using PDO bound parameters.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$pdo = new PDO('sqlite:\/path\/db\/users.db');\n$stmt = $pdo-&gt;prepare('SELECT name FROM users WHERE id = :id');\n$id = filter_input(INPUT_GET, 'id', FILTER_SANITIZE_NUMBER_INT); \/\/ &lt;-- filter your data first (see [Data Filtering](#data_filtering)), especially important for INSERT, UPDATE, etc.\n$stmt-&gt;bindParam(':id', $id, PDO::PARAM_INT); \/\/ &lt;-- Automatically sanitized for SQL by PDO\n$stmt-&gt;execute();<\/pre><\/figure><p>This is correct code. It uses a bound parameter on a PDO statement. This escapes the foreign input ID before it is\nintroduced to the database preventing potential SQL injection attacks.<\/p><p>For writes, such as INSERT or UPDATE, it\u2019s especially critical to still <a href=\"#data_filtering\">filter your data<\/a> first and sanitize it for other things (removal of HTML tags, JavaScript, etc).  PDO will only sanitize it for SQL, not for your application.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.pdo\">Learn about PDO<\/a><\/li>\n<\/ul><p>You should also be aware that database connections use up resources and it was not unheard-of to have resources\nexhausted if connections were not implicitly closed, however this was more common in other languages. Using PDO you can\nimplicitly close the connection by destroying the object by ensuring all remaining references to it are deleted, i.e.\nset to NULL. If you don\u2019t do this explicitly, PHP will automatically close the connection when your script ends -\nunless of course you are using persistent connections.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/pdo.connections\">Learn about PDO connections<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"databases_interacting\"><h2 id=\"databases_interacting_title\">Interacting with Databases<\/h2><p>When developers first start to learn PHP, they often end up mixing their database interaction up with their\npresentation logic, using code that might look like this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;ul&gt;\n&lt;?php\nforeach ($db-&gt;query('SELECT * FROM table') as $row) {\n    echo \"&lt;li&gt;\".$row['field1'].\" - \".$row['field1'].\"&lt;\/li&gt;\";\n}\n?&gt;\n&lt;\/ul&gt;<\/pre><\/figure><p>This is bad practice for all sorts of reasons, mainly that it\u2019s hard to debug, hard to test, hard to read and it is\ngoing to output a lot of fields if you don\u2019t put a limit on there.<\/p><p>While there are many other solutions to doing this - depending on if you prefer <a href=\"https:\/\/phptherightway.com\/#object-oriented-programming\">OOP<\/a> or\n<a href=\"https:\/\/phptherightway.com\/#functional-programming\">functional programming<\/a> - there must be some element of separation.<\/p><p>Consider the most basic step:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nfunction getAllFoos($db) {\n    return $db-&gt;query('SELECT * FROM table');\n}\n$results = getAllFoos($db);\nforeach ($results as $row) {\n    echo \"&lt;li&gt;\".$row['field1'].\" - \".$row['field1'].\"&lt;\/li&gt;\"; \/\/ BAD!!\n}<\/pre><\/figure><p>That is a good start. Put those two items in two different files and you\u2019ve got some clean separation.<\/p><p>Create a class to place that method in and you have a \u201cModel\u201d. Create a simple <code class=\"highlighter-rouge\">.php<\/code> file to put the presentation\nlogic in and you have a \u201cView\u201d, which is very nearly <a href=\"https:\/\/code.tutsplus.com\/tutorials\/mvc-for-noobs--net-10488\">MVC<\/a> - a common OOP architecture for most\n<a href=\"https:\/\/phptherightway.com\/#frameworks\">frameworks<\/a>.<\/p><p><strong>foo.php<\/strong><\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$db = new PDO('mysql:host=localhost;dbname=testdb;charset=utf8mb4', 'username', 'password');\n\/\/ Make your model available\ninclude 'models\/FooModel.php';\n\/\/ Create an instance\n$fooModel = new FooModel($db);\n\/\/ Get the list of Foos\n$fooList = $fooModel-&gt;getAllFoos();\n\/\/ Show the view\ninclude 'views\/foo-list.php';<\/pre><\/figure><p><strong>models\/FooModel.php<\/strong><\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nclass FooModel\n{\n    protected $db;\n    public function __construct(PDO $db)\n    {\n        $this-&gt;db = $db;\n    }\n    public function getAllFoos() {\n        return $this-&gt;db-&gt;query('SELECT * FROM table');\n    }\n}<\/pre><\/figure><p><strong>views\/foo-list.php<\/strong><\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php foreach ($fooList as $row): ?&gt;\n    &lt;li&gt;&lt;?= $row['field1'] ?&gt; - &lt;?= $row['field1'] ?&gt;&lt;\/li&gt;\n&lt;?php endforeach ?&gt;<\/pre><\/figure><p>This is essentially the same as what most modern frameworks are doing, albeit a little more manual. You might not\nneed to do all of that every time, but mixing together too much presentation logic and database interaction can be a\nreal problem if you ever want to <a href=\"https:\/\/phptherightway.com\/#unit-testing\">unit-test<\/a> your application.<\/p><p><a href=\"https:\/\/phpbridge.org\/docs\/\">PHPBridge<\/a> has a great resource called <a href=\"https:\/\/phpbridge.org\/intro-to-php\/creating_a_data_class\">Creating a Data Class<\/a> which covers a very similar topic, and is great for\ndevelopers just getting used to the concept of interacting with databases.<\/p><\/section><section class=\"chapter\" id=\"databases_abstraction_layers\"><h2 id=\"databases_abstraction_layers_title\">Abstraction Layers<\/h2><p>Many frameworks provide their own abstraction layer which may or may not sit on top of <a href=\"https:\/\/secure.php.net\/book.pdo\">PDO<\/a>. These will often\nemulate features for one database system that is missing from another by wrapping your queries in PHP methods, giving\nyou actual database abstraction instead of just the connection abstraction that PDO provides. This will of course add a\nlittle overhead, but if you are building a portable application that needs to work with MySQL, PostgreSQL and SQLite\nthen a little overhead will be worth it the sake of code cleanliness.<\/p><p>Some abstraction layers have been built using the <a href=\"https:\/\/www.php-fig.org\/psr\/psr-0\/\">PSR-0<\/a> or <a href=\"https:\/\/www.php-fig.org\/psr\/psr-4\/\">PSR-4<\/a> namespace standards so can be\ninstalled in any application you like:<\/p><ul><li><a href=\"https:\/\/github.com\/auraphp\/Aura.Sql\">Aura SQL<\/a><\/li>\n  <li><a href=\"https:\/\/www.doctrine-project.org\/projects\/dbal.html\">Doctrine2 DBAL<\/a><\/li>\n  <li><a href=\"http:\/\/propelorm.org\/\">Propel<\/a><\/li>\n  <li><a href=\"https:\/\/packages.zendframework.com\/docs\/latest\/manual\/en\/index.html#zendframework\/zend-db\">Zend-db<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"templating\"><p>Templates provide a convenient way of separating your controller and domain logic from your presentation logic.\nTemplates typically contain the HTML of your application, but may also be used for other formats, such as XML.\nTemplates are often referred to as \u201cviews\u201d, which make up <strong>part of<\/strong> the second component of the\n<a href=\"https:\/\/phptherightway.com\/pages\/Design-Patterns.html#model-view-controller\">model\u2013view\u2013controller<\/a> (MVC) software architecture pattern.<\/p><\/section><section class=\"chapter\" id=\"templating_benefits\"><h2 id=\"templating_benefits_title\">Benefits<\/h2><p>The main benefit to using templates is the clear separation they create between the presentation logic and the rest of\nyour application. Templates have the sole responsibility of displaying formatted content. They are not responsible for\ndata lookup, persistence or other more complex tasks. This leads to cleaner, more readable code which is especially\nhelpful in a team environment where developers work on the server-side code (controllers, models) and designers work on\nthe client-side code (markup).<\/p><p>Templates also improve the organization of presentation code. Templates are typically placed in a \u201cviews\u201d folder, each\ndefined within a single file. This approach encourages code reuse where larger blocks of code are broken into smaller,\nreusable pieces, often called partials. For example, your site header and footer can each be defined as templates,\nwhich are then included before and after each page template.<\/p><p>Finally, depending on the library you use, templates can offer more security by automatically escaping user-generated\ncontent. Some libraries even offer sand-boxing, where template designers are only given access to white-listed\nvariables and functions.<\/p><\/section><section class=\"chapter\" id=\"plain_php_templates\"><h2 id=\"plain_php_templates_title\">Plain PHP Templates<\/h2><p>Plain PHP templates are simply templates that use native PHP code. They are a natural choice since PHP is actually a\ntemplate language itself. That simply means that you can combine PHP code within other code, like HTML. This is\nbeneficial to PHP developers as there is no new syntax to learn, they know the functions available to them, and their\ncode editors already have PHP syntax highlighting and auto-completion built-in. Further, plain PHP templates tend to be\nvery fast as no compiling stage is required.<\/p><p>Every modern PHP framework employs some kind of template system, most of which use plain PHP by default. Outside of\nframeworks, libraries like <a href=\"http:\/\/platesphp.com\/\">Plates<\/a> or <a href=\"https:\/\/github.com\/auraphp\/Aura.View\">Aura.View<\/a> make working with plain PHP templates easier by\noffering modern template functionality such as inheritance, layouts and extensions.<\/p><h3 id=\"simple-example-of-a-plain-php-template\">Simple example of a plain PHP template<\/h3><p>Using the <a href=\"http:\/\/platesphp.com\/\">Plates<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php \/\/ user_profile.php ?&gt;\n&lt;?php $this-&gt;insert('header', ['title' =&gt; 'User Profile']) ?&gt;\n&lt;h1&gt;User Profile&lt;\/h1&gt;\n&lt;p&gt;Hello, &lt;?=$this-&gt;escape($name)?&gt;&lt;\/p&gt;\n&lt;?php $this-&gt;insert('footer') ?&gt;<\/pre><\/figure><h3 id=\"example-of-plain-php-templates-using-inheritance\">Example of plain PHP templates using inheritance<\/h3><p>Using the <a href=\"http:\/\/platesphp.com\/\">Plates<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php \/\/ template.php ?&gt;\n&lt;html&gt;\n&lt;head&gt;\n    &lt;title&gt;&lt;?=$title?&gt;&lt;\/title&gt;\n&lt;\/head&gt;\n&lt;body&gt;\n&lt;main&gt;\n    &lt;?=$this-&gt;section('content')?&gt;\n&lt;\/main&gt;\n&lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/figure><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php \/\/ user_profile.php ?&gt;\n&lt;?php $this-&gt;layout('template', ['title' =&gt; 'User Profile']) ?&gt;\n&lt;h1&gt;User Profile&lt;\/h1&gt;\n&lt;p&gt;Hello, &lt;?=$this-&gt;escape($name)?&gt;&lt;\/p&gt;<\/pre><\/figure><\/section><section class=\"chapter\" id=\"compiled_templates\"><h2 id=\"compiled_templates_title\">Compiled Templates<\/h2><p>While PHP has evolved into a mature, object oriented language, it <a href=\"http:\/\/fabien.potencier.org\/article\/34\/templating-engines-in-php\">hasn\u2019t improved much<\/a> as\na templating language. Compiled templates, like <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a>, <a href=\"https:\/\/github.com\/box\/brainy\">Brainy<\/a>, or <a href=\"https:\/\/www.smarty.net\/\">Smarty<\/a>*, fill this void by offering a new syntax that has\nbeen geared specifically to templating. From automatic escaping, to inheritance and simplified control structures,\ncompiled templates are designed to be easier to write, cleaner to read and safer to use. Compiled templates can even be\nshared across different languages, <a href=\"https:\/\/mustache.github.io\/\">Mustache<\/a> being a good example of this. Since these templates must be compiled\nthere is a slight performance hit, however this is very minimal when proper caching is used.<\/p><p><em>*While Smarty offers automatic escaping, this feature is NOT enabled by default.<\/em><\/p><h3 id=\"simple-example-of-a-compiled-template\">Simple example of a compiled template<\/h3><p>Using the <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-html-jinja\" data-lang=\"html+jinja\">{% include 'header.html' with {'title': 'User Profile'} %}\n&lt;h1&gt;User Profile&lt;\/h1&gt;\n&lt;p&gt;Hello, {{ name }}&lt;\/p&gt;\n{% include 'footer.html' %}<\/pre><\/figure><h3 id=\"example-of-compiled-templates-using-inheritance\">Example of compiled templates using inheritance<\/h3><p>Using the <a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a> library.<\/p><figure class=\"highlight\"><pre class=\"language-html-jinja\" data-lang=\"html+jinja\">\/\/ template.html\n&lt;html&gt;\n&lt;head&gt;\n    &lt;title&gt;{% block title %}{% endblock %}&lt;\/title&gt;\n&lt;\/head&gt;\n&lt;body&gt;\n&lt;main&gt;\n    {% block content %}{% endblock %}\n&lt;\/main&gt;\n&lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/figure><figure class=\"highlight\"><pre class=\"language-html-jinja\" data-lang=\"html+jinja\">\/\/ user_profile.html\n{% extends \"template.html\" %}\n{% block title %}User Profile{% endblock %}\n{% block content %}\n    &lt;h1&gt;User Profile&lt;\/h1&gt;\n    &lt;p&gt;Hello, {{ name }}&lt;\/p&gt;\n{% endblock %}<\/pre><\/figure><\/section><section class=\"chapter\" id=\"templating_further_reading\"><h2 id=\"templating_further_reading_title\">Further Reading<\/h2><h3 id=\"articles--tutorials\">Articles &amp; Tutorials<\/h3><ul><li><a href=\"http:\/\/fabien.potencier.org\/article\/34\/templating-engines-in-php\">Templating Engines in PHP<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/an-introduction-to-views-templating-in-codeigniter--net-25648\">An Introduction to Views &amp; Templating in CodeIgniter<\/a><\/li>\n  <li><a href=\"https:\/\/www.smashingmagazine.com\/2011\/10\/getting-started-with-php-templating\/\">Getting Started With PHP Templating<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/roll-your-own-templating-system-in-php--net-16596\">Roll Your Own Templating System in PHP<\/a><\/li>\n  <li><a href=\"https:\/\/laracasts.com\/series\/laravel-from-scratch\/episodes\/7\">Master Pages<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/working-with-templates-in-symfony-2--cms-21172\">Working With Templates in Symfony 2<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/box\/brainy\/wiki\/Writing-Safe-Templates\">Writing Safer Templates<\/a><\/li>\n<\/ul><h3 id=\"libraries\">Libraries<\/h3><ul><li><a href=\"https:\/\/github.com\/auraphp\/Aura.View\">Aura.View<\/a> <em>(native)<\/em><\/li>\n  <li><a href=\"https:\/\/laravel.com\/docs\/blade\">Blade<\/a> <em>(compiled, framework specific)<\/em><\/li>\n  <li><a href=\"https:\/\/github.com\/box\/brainy\">Brainy<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"http:\/\/dwoo.org\/\">Dwoo<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/github.com\/nette\/latte\">Latte<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/github.com\/bobthecow\/mustache.php\">Mustache<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/phptal.org\/\">PHPTAL<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"http:\/\/platesphp.com\/\">Plates<\/a> <em>(native)<\/em><\/li>\n  <li><a href=\"https:\/\/www.smarty.net\/\">Smarty<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/twig.symfony.com\/\">Twig<\/a> <em>(compiled)<\/em><\/li>\n  <li><a href=\"https:\/\/framework.zend.com\/manual\/2.3\/en\/modules\/zend.view.quick-start.html\">Zend\\View<\/a> <em>(native, framework specific)<\/em><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"errors\"><h2 id=\"errors_title\">Errors<\/h2><p>In many \u201cexception-heavy\u201d programming languages, whenever anything goes wrong an exception will be thrown. This is\ncertainly a viable way to do things, but PHP is an \u201cexception-light\u201d programming language. While it does have\nexceptions and more of the core is starting to use them when working with objects, most of PHP itself will try to keep\nprocessing regardless of what happens, unless a fatal error occurs.<\/p><p>For example:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">$ php -a\nphp &gt; echo $foo;\nNotice: Undefined variable: foo in php shell code on line 1<\/pre><\/figure><p>This is only a notice error, and PHP will happily carry on. This can be confusing for those coming from\n\u201cexception-heavy\u201d languages, because referencing a missing variable in Python for example will throw an exception:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">$ python\n&gt;&gt;&gt; print foo\nTraceback (most recent call last):\n  File \"&lt;stdin&gt;\", line 1, in &lt;module&gt;\nNameError: name 'foo' is not defined<\/pre><\/figure><p>The only real difference is that Python will freak out over any small thing, so that developers can be super sure any\npotential issue or edge-case is caught, whereas PHP will keep on processing unless something extreme happens, at which\npoint it will throw an error and report it.<\/p><h3 id=\"error-severity\">Error Severity<\/h3><p>PHP has several levels of error severity. The three most common types of messages are errors, notices and warnings.\nThese have different levels of severity; <code class=\"highlighter-rouge\">E_ERROR<\/code>, <code class=\"highlighter-rouge\">E_NOTICE<\/code>, and <code class=\"highlighter-rouge\">E_WARNING<\/code>. Errors are fatal run-time errors and\nare usually caused by faults in your code and need to be fixed as they\u2019ll cause PHP to stop executing. Notices are\nadvisory messages caused by code that may or may not cause problems during the execution of the script, execution is\nnot halted. Warnings are non-fatal errors, execution of the script will not be halted.<\/p><p>Another type of error message reported at compile time are <code class=\"highlighter-rouge\">E_STRICT<\/code> messages. These messages are used to suggest\nchanges to your code to help ensure best interoperability and forward compatibility with upcoming versions of PHP.<\/p><h3 id=\"changing-phps-error-reporting-behaviour\">Changing PHP\u2019s Error Reporting Behaviour<\/h3><p>Error Reporting can be changed by using PHP settings and\/or PHP function calls. Using the built in PHP function\n<code class=\"highlighter-rouge\">error_reporting()<\/code> you can set the level of errors for the duration of the script execution by passing one of the\npredefined error level constants, meaning if you only want to see Errors and Warnings - but not Notices - then you can\nconfigure that:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nerror_reporting(E_ERROR | E_WARNING);<\/pre><\/figure><p>You can also control whether or not errors are displayed to the screen (good for development) or hidden, and logged\n(good for production). For more information on this check out the <a href=\"https:\/\/phptherightway.com\/#error_reporting\">Error Reporting<\/a> section.<\/p><h3 id=\"inline-error-suppression\">Inline Error Suppression<\/h3><p>You can also tell PHP to suppress specific errors with the Error Control Operator <code class=\"highlighter-rouge\">@<\/code>. You put this operator at the\nbeginning of an expression, and any error that\u2019s a direct result of the expression is silenced.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\necho @$foo['bar'];<\/pre><\/figure><p>This will output <code class=\"highlighter-rouge\">$foo['bar']<\/code> if it exists, but will simply return a null and print nothing if the variable <code class=\"highlighter-rouge\">$foo<\/code> or\n<code class=\"highlighter-rouge\">'bar'<\/code> key does not exist. Without the error control operator, this expression could create a <code class=\"highlighter-rouge\">PHP Notice: Undefined\nvariable: foo<\/code> or <code class=\"highlighter-rouge\">PHP Notice: Undefined index: bar<\/code> error.<\/p><p>This might seem like a good idea, but there are a few undesirable tradeoffs. PHP handles expressions using an <code class=\"highlighter-rouge\">@<\/code> in a\nless performant way than expressions without an <code class=\"highlighter-rouge\">@<\/code>. Premature optimization may be the root of all programming\narguments, but if performance is particularly important for your application\/library it\u2019s important to understand the\nerror control operator\u2019s performance implications.<\/p><p>Secondly, the error control operator <strong>completely<\/strong> swallows the error. The error is not displayed, and the error is\nnot sent to the error log. Also, stock\/production PHP systems have no way to turn off the error control operator. While\nyou may be correct that the error you\u2019re seeing is harmless, a different, less harmless error will be just as silent.<\/p><p>If there\u2019s a way to avoid the error suppression operator, you should consider it. For example, our code above could be\nrewritten like this:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/Null Coalescing Operator\necho $foo['bar'] ?? '';<\/pre><\/figure><p>One instance where error suppression might make sense is where <code class=\"highlighter-rouge\">fopen()<\/code> fails to find a file to load. You could check\nfor the existence of the file before you try to load it, but if the file is deleted after the check and before the\n<code class=\"highlighter-rouge\">fopen()<\/code> (which might sound impossible, but it can happen) then <code class=\"highlighter-rouge\">fopen()<\/code> will return false <em>and<\/em> throw an error. This\nis potentially something PHP should resolve, but is one case where error suppression might seem like the only valid\nsolution.<\/p><p>Earlier we mentioned there\u2019s no way in a stock PHP system to turn off the error control operator. However, <a href=\"https:\/\/xdebug.org\/docs\/basic\">Xdebug<\/a> has\nan <code class=\"highlighter-rouge\">xdebug.scream<\/code> ini setting which will disable the error control operator. You can set this via your <code class=\"highlighter-rouge\">php.ini<\/code> file\nwith the following.<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">xdebug.scream = On<\/pre><\/figure><p>You can also set this value at runtime with the <code class=\"highlighter-rouge\">ini_set<\/code> function<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nini_set('xdebug.scream', '1')<\/pre><\/figure><p>The \u201c<a href=\"https:\/\/secure.php.net\/book.scream\">Scream<\/a>\u201d PHP extension offers similar functionality to Xdebug\u2019s, although Scream\u2019s ini setting is named\n<code class=\"highlighter-rouge\">scream.enabled<\/code>.<\/p><p>This is most useful when you\u2019re debugging code and suspect an informative error is suppressed. Use scream with care,\nand as a temporary debugging tool. There\u2019s lots of PHP library code that may not work with the error control operator\ndisabled.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.operators.errorcontrol\">Error Control Operators<\/a><\/li>\n  <li><a href=\"https:\/\/www.sitepoint.com\/\">SitePoint<\/a><\/li>\n  <li><a href=\"https:\/\/xdebug.org\/docs\/basic\">Xdebug<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/book.scream\">Scream<\/a><\/li>\n<\/ul><h3 id=\"errorexception\">ErrorException<\/h3><p>PHP is perfectly capable of being an \u201cexception-heavy\u201d programming language, and only requires a few lines of code to\nmake the switch. Basically you can throw your \u201cerrors\u201d as \u201cexceptions\u201d using the <code class=\"highlighter-rouge\">ErrorException<\/code> class, which extends\nthe <code class=\"highlighter-rouge\">Exception<\/code> class.<\/p><p>This is a common practice implemented by a large number of modern frameworks such as Symfony and Laravel. In debug\nmode <em>(or dev mode)<\/em> both of these frameworks will display a nice and clean <em>stack trace<\/em>.<\/p><p>There are also some packages available for better error and exception handling and reporting. Like <a href=\"https:\/\/filp.github.io\/whoops\/\">Whoops!<\/a>, which comes with the default installation of Laravel and can be used in any framework as well.<\/p><p>By throwing errors as exceptions in development you can handle them better than the usual result, and if you see an\nexception during development you can wrap it in a catch statement with specific instructions on how to handle the\nsituation. Each exception you catch instantly makes your application that little bit more robust.<\/p><p>More information on this and details on how to use <code class=\"highlighter-rouge\">ErrorException<\/code> with error handling can be found at\n<a href=\"https:\/\/secure.php.net\/class.errorexception\">ErrorException Class<\/a>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.operators.errorcontrol\">Error Control Operators<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.constants\">Predefined Constants for Error Handling<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.error-reporting\"><code class=\"highlighter-rouge\">error_reporting()<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/phptherightway.com\/#error_reporting\">Reporting<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"exceptions\"><h2 id=\"exceptions_title\">Exceptions<\/h2><p>Exceptions are a standard part of most popular programming languages, but they are often overlooked by PHP programmers.\nLanguages like Ruby are extremely Exception heavy, so whenever something goes wrong such as a HTTP request failing, or\na DB query goes wrong, or even if an image asset could not be found, Ruby (or the gems being used) will throw an\nexception to the screen meaning you instantly know there is a mistake.<\/p><p>PHP itself is fairly lax with this, and a call to <code class=\"highlighter-rouge\">file_get_contents()<\/code> will usually just get you a <code class=\"highlighter-rouge\">FALSE<\/code> and a\nwarning.\nMany older PHP frameworks like CodeIgniter will just return a false, log a message to their proprietary logs and maybe\nlet you use a method like <code class=\"highlighter-rouge\">$this-&gt;upload-&gt;get_error()<\/code> to see what went wrong. The problem here is that you have to go\nlooking for a mistake and check the docs to see what the error method is for this class, instead of having it made\nextremely obvious.<\/p><p>Another problem is when classes automatically throw an error to the screen and exit the process. When you do this you\nstop another developer from being able to dynamically handle that error. Exceptions should be thrown to make a\ndeveloper aware of an error; they then can choose how to handle this. E.g.:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n$email = new Fuel\\Email;\n$email-&gt;subject('My Subject');\n$email-&gt;body('How the heck are you?');\n$email-&gt;to('guy@example.com', 'Some Guy');\ntry\n{\n    $email-&gt;send();\n}\ncatch(Fuel\\Email\\ValidationFailedException $e)\n{\n    \/\/ The validation failed\n}\ncatch(Fuel\\Email\\SendingFailedException $e)\n{\n    \/\/ The driver could not send the email\n}\nfinally\n{\n    \/\/ Executed regardless of whether an exception has been thrown, and before normal execution resumes\n}<\/pre><\/figure><h3 id=\"spl-exceptions\">SPL Exceptions<\/h3><p>The generic <code class=\"highlighter-rouge\">Exception<\/code> class provides very little debugging context for the developer; however, to remedy this, it is\npossible to create a specialized <code class=\"highlighter-rouge\">Exception<\/code> type by sub-classing the generic <code class=\"highlighter-rouge\">Exception<\/code> class:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nclass ValidationException extends Exception {}<\/pre><\/figure><p>This means you can add multiple catch blocks and handle different Exceptions differently. This can lead to the\ncreation of a <em>lot<\/em> of custom Exceptions, some of which could have been avoided using the SPL Exceptions\nprovided in the <a href=\"https:\/\/phptherightway.com\/#standard_php_library\">SPL extension<\/a>.<\/p><p>If for example you use the <code class=\"highlighter-rouge\">__call()<\/code> Magic Method and an invalid method is requested then instead of throwing a\nstandard Exception which is vague, or creating a custom Exception just for that, you could just\n<code class=\"highlighter-rouge\">throw new BadMethodCallException;<\/code>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/language.exceptions\">Read about Exceptions<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/spl.exceptions\">Read about SPL Exceptions<\/a><\/li>\n  <li><a href=\"https:\/\/www.brandonsavage.net\/exceptional-php-nesting-exceptions-in-php\/\">Nesting Exceptions In PHP<\/a><\/li>\n  <li><a href=\"http:\/\/ralphschindler.com\/2010\/09\/15\/exception-best-practices-in-php-5-3\">Exception Best Practices in PHP 5.3<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"security\"><p>The best resource I\u2019ve found on PHP security is <a href=\"https:\/\/paragonie.com\/blog\/2017\/12\/2018-guide-building-secure-php-software\">The 2018 Guide to Building Secure PHP Software<\/a> by\n<a href=\"https:\/\/paragonie.com\/\">Paragon Initiative<\/a>.<\/p><\/section><section class=\"chapter\" id=\"web_application_security\"><h2 id=\"web_application_security_title\">Web Application Security<\/h2><p>It is very important for every PHP developer to learn <a href=\"https:\/\/paragonie.com\/blog\/2015\/08\/gentle-introduction-application-security\">the basics of web application security<\/a>, which can be broken\ndown into a handful of broad topics:<\/p><ol><li>Code-data separation.\n    <ul><li>When data is executed as code, you get SQL Injection, Cross-Site Scripting, Local\/Remote File Inclusion, etc.<\/li>\n      <li>When code is printed as data, you get information leaks (source code disclosure or, in the case of C programs,\nenough information to bypass <a href=\"http:\/\/searchsecurity.techtarget.com\/definition\/address-space-layout-randomization-ASLR\">ASLR<\/a>).<\/li>\n    <\/ul><\/li>\n  <li>Application logic.\n    <ul><li>Missing authentication or authorization controls.<\/li>\n      <li>Input validation.<\/li>\n    <\/ul><\/li>\n  <li>Operating environment.\n    <ul><li>PHP versions.<\/li>\n      <li>Third party libraries.<\/li>\n      <li>The operating system.<\/li>\n    <\/ul><\/li>\n  <li>Cryptography weaknesses.\n    <ul><li><a href=\"https:\/\/paragonie.com\/blog\/2016\/01\/on-design-and-implementation-stealth-backdoor-for-web-applications\">Weak random numbers<\/a>.<\/li>\n      <li><a href=\"https:\/\/paragonie.com\/blog\/2015\/05\/using-encryption-and-authentication-correctly\">Chosen-ciphertext attacks<\/a>.<\/li>\n      <li><a href=\"http:\/\/blog.ircmaxell.com\/2014\/11\/its-all-about-time.html\">Side-channel information leaks<\/a>.<\/li>\n    <\/ul><\/li>\n<\/ol><p>There are bad people ready and willing to exploit your web application. It is important that you take necessary\nprecautions to harden your web application\u2019s security. Luckily, the fine folks at\n<a href=\"https:\/\/www.owasp.org\/\">The Open Web Application Security Project<\/a> (OWASP) have compiled a comprehensive list of known security issues and\nmethods to protect yourself against them. This is a must read for the security-conscious developer. <a href=\"https:\/\/phpsecurity.readthedocs.io\/en\/latest\/index.html\">Survive The Deep End: PHP Security<\/a> by Padraic Brady is also another good web application security guide for PHP.<\/p><ul><li><a href=\"https:\/\/www.owasp.org\/index.php\/Guide_Table_of_Contents\">Read the OWASP Security Guide<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"password_hashing\"><h2 id=\"password_hashing_title\">Password Hashing<\/h2><p>Eventually everyone builds a PHP application that relies on user login. Usernames and passwords are stored in a\ndatabase and later used to authenticate users upon login.<\/p><p>It is important that you properly <a href=\"https:\/\/wikipedia.org\/wiki\/Cryptographic_hash_function\"><em>hash<\/em><\/a> passwords before storing them. Hashing and encrypting are <a href=\"https:\/\/paragonie.com\/blog\/2015\/08\/you-wouldnt-base64-a-password-cryptography-decoded\">two very different things<\/a>\nthat often get confused.<\/p><p>Hashing is an irreversible, one-way function. This produces a fixed-length string that cannot be feasibly reversed.\nThis means you can compare a hash against another to determine if they both came from the same source string, but you\ncannot determine the original string. If passwords are not hashed and your database is accessed by an unauthorized\nthird-party, all user accounts are now compromised.<\/p><p>Unlike hashing, encryption is reversible (provided you have the key). Encryption is useful in other areas, but is a poor\nstrategy for securely storing passwords.<\/p><p>Passwords should also be individually <a href=\"https:\/\/wikipedia.org\/wiki\/Salt_(cryptography)\"><em>salted<\/em><\/a> by adding a random string to each password before hashing. This prevents dictionary attacks and the use of \u201crainbow tables\u201d (a reverse list of crytographic hashes for common passwords.)<\/p><p>Hashing and salting are vital as often users use the same password for multiple services and password quality can be poor.<\/p><p>Additionally, you should use <a href=\"https:\/\/paragonie.com\/blog\/2016\/02\/how-safely-store-password-in-2016\">a specialized <em>password hashing<\/em> algoithm<\/a> rather than fast, general-purpose\ncryptographic hash function (e.g. SHA256). The short list of acceptable password hashing algorithms (as of June 2018)\nto use are:<\/p><ul><li>Argon2 (available in PHP 7.2 and newer)<\/li>\n  <li>Scrypt<\/li>\n  <li><strong>Bcrypt<\/strong> (PHP provides this one for you; see below)<\/li>\n  <li>PBKDF2 with HMAC-SHA256 or HMAC-SHA512<\/li>\n<\/ul><p>Fortunately, nowadays PHP makes this easy.<\/p><p><strong>Hashing passwords with <code class=\"highlighter-rouge\">password_hash<\/code><\/strong><\/p><p>In PHP 5.5 <code class=\"highlighter-rouge\">password_hash()<\/code> was introduced. At this time it is using BCrypt, the strongest algorithm currently\nsupported by PHP. It will be updated in the future to support more algorithms as needed though. The <code class=\"highlighter-rouge\">password_compat<\/code>\nlibrary was created to provide forward compatibility for PHP &gt;= 5.3.7.<\/p><p>Below we hash a string, and then check the hash against a new string. Because our two source strings are different\n(\u2018secret-password\u2019 vs. \u2018bad-password\u2019) this login will fail.<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\nrequire 'password.php';\n$passwordHash = password_hash('secret-password', PASSWORD_DEFAULT);\nif (password_verify('bad-password', $passwordHash)) {\n    \/\/ Correct Password\n} else {\n    \/\/ Wrong password\n}<\/pre><\/figure><p><code class=\"highlighter-rouge\">password_hash()<\/code> takes care of password salting for you. The salt is stored, along with the algorithm and \u201ccost\u201d, as part of the hash.  <code class=\"highlighter-rouge\">password_verify()<\/code> extracts this to determine how to check the password, so you don\u2019t need a separate database field to store your salts.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/function.password-hash\">Learn about <code class=\"highlighter-rouge\">password_hash()<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/ircmaxell\/password_compat\"><code class=\"highlighter-rouge\">password_compat<\/code> for PHP &gt;= 5.3.7 &amp;&amp; &lt; 5.5<\/a><\/li>\n  <li><a href=\"https:\/\/wikipedia.org\/wiki\/Cryptographic_hash_function\">Learn about hashing in regards to cryptography<\/a><\/li>\n  <li><a href=\"https:\/\/wikipedia.org\/wiki\/Salt_(cryptography)\">Learn about salts<\/a><\/li>\n  <li><a href=\"https:\/\/wiki.php.net\/rfc\/password_hash\">PHP <code class=\"highlighter-rouge\">password_hash()<\/code> RFC<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"data_filtering\"><h2 id=\"data_filtering_title\">Data Filtering<\/h2><p>Never ever (ever) trust foreign input introduced to your PHP code. Always sanitize and validate foreign input before\nusing it in code. The <code class=\"highlighter-rouge\">filter_var()<\/code> and <code class=\"highlighter-rouge\">filter_input()<\/code> functions can sanitize text and validate text formats (e.g.\nemail addresses).<\/p><p>Foreign input can be anything: <code class=\"highlighter-rouge\">$_GET<\/code> and <code class=\"highlighter-rouge\">$_POST<\/code> form input data, some values in the <code class=\"highlighter-rouge\">$_SERVER<\/code> superglobal, and the\nHTTP request body via <code class=\"highlighter-rouge\">fopen('php:\/\/input', 'r')<\/code>. Remember, foreign input is not limited to form data submitted by the\nuser. Uploaded and downloaded files, session values, cookie data, and data from third-party web services are foreign\ninput, too.<\/p><p>While foreign data can be stored, combined, and accessed later, it is still foreign input. Every time you process,\noutput, concatenate, or include data in your code, ask yourself if the data is filtered properly and can it be trusted.<\/p><p>Data may be <em>filtered<\/em> differently based on its purpose. For example, when unfiltered foreign input is passed into HTML\npage output, it can execute HTML and JavaScript on your site! This is known as Cross-Site Scripting (XSS) and can be a\nvery dangerous attack. One way to avoid XSS is to sanitize all user-generated data before outputting it to your page by\nremoving HTML tags with the <code class=\"highlighter-rouge\">strip_tags()<\/code> function or escaping characters with special meaning into their respective\nHTML entities with the <code class=\"highlighter-rouge\">htmlentities()<\/code> or <code class=\"highlighter-rouge\">htmlspecialchars()<\/code> functions.<\/p><p>Another example is passing options to be executed on the command line. This can be extremely dangerous (and is usually\na bad idea), but you can use the built-in <code class=\"highlighter-rouge\">escapeshellarg()<\/code> function to sanitize the executed command\u2019s arguments.<\/p><p>One last example is accepting foreign input to determine a file to load from the filesystem. This can be exploited by\nchanging the filename to a file path. You need to remove <code class=\"highlighter-rouge\">\"\/\"<\/code>, <code class=\"highlighter-rouge\">\"..\/\"<\/code>, <a href=\"https:\/\/secure.php.net\/security.filesystem.nullbytes\">null bytes<\/a>, or other characters from the\nfile path so it can\u2019t load hidden, non-public, or sensitive files.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.filter\">Learn about data filtering<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.filter-var\">Learn about <code class=\"highlighter-rouge\">filter_var<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/function.filter-input\">Learn about <code class=\"highlighter-rouge\">filter_input<\/code><\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/security.filesystem.nullbytes\">Learn about handling null bytes<\/a><\/li>\n<\/ul><h3 id=\"sanitization\">Sanitization<\/h3><p>Sanitization removes (or escapes) illegal or unsafe characters from foreign input.<\/p><p>For example, you should sanitize foreign input before including the input in HTML or inserting it into a raw SQL query.\nWhen you use bound parameters with <a href=\"#databases\">PDO<\/a>, it will sanitize the input for you.<\/p><p>Sometimes it is required to allow some safe HTML tags in the input when including it in the HTML page. This is very\nhard to do and many avoid it by using other more restricted formatting like Markdown or BBCode, although whitelisting\nlibraries like <a href=\"http:\/\/htmlpurifier.org\/\">HTML Purifier<\/a> exists for this reason.<\/p><p><a href=\"https:\/\/secure.php.net\/filter.filters.sanitize\">See Sanitization Filters<\/a><\/p><h3 id=\"unserialization\">Unserialization<\/h3><p>It is dangerous to <code class=\"highlighter-rouge\">unserialize()<\/code> data from users or other untrusted sources.  Doing so can allow malicious users to instantiate objects (with user-defined properties) whose destructors will be executed, <strong>even if the objects themselves aren\u2019t used<\/strong>.  You should therefore avoid unserializing untrusted data.<\/p><p>If you absolutely must unserialize data from untrusted sources, use PHP 7\u2019s <a href=\"https:\/\/secure.php.net\/manual\/function.unserialize.php\"><code class=\"highlighter-rouge\">allowed_classes<\/code><\/a> option to restrict which object types are allowed to be unserialized.<\/p><h3 id=\"validation\">Validation<\/h3><p>Validation ensures that foreign input is what you expect. For example, you may want to validate an email address, a\nphone number, or age when processing a registration submission.<\/p><p><a href=\"https:\/\/secure.php.net\/filter.filters.validate\">See Validation Filters<\/a><\/p><\/section><section class=\"chapter\" id=\"configuration_files\"><h2 id=\"configuration_files_title\">Configuration Files<\/h2><p>When creating configuration files for your applications, best practices recommend that one of the following methods be\nfollowed:<\/p><ul><li>It is recommended that you store your configuration information where it cannot be accessed directly and pulled in\nvia the file system.<\/li>\n  <li>If you must store your configuration files in the document root, name the files with a <code class=\"highlighter-rouge\">.php<\/code> extension. This ensures\nthat, even if the script is accessed directly, it will not be output as plain text.<\/li>\n  <li>Information in configuration files should be protected accordingly, either through encryption or group\/user file\nsystem permissions.<\/li>\n  <li>It is a good idea to ensure that you do not commit configuration files containing sensitive information e.g. passwords or API tokens to source control.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"register_globals\"><h2 id=\"register_globals_title\">Register Globals<\/h2><p><strong>NOTE:<\/strong> As of PHP 5.4.0 the <code class=\"highlighter-rouge\">register_globals<\/code> setting has been removed and can no longer be used. This is only\nincluded as a warning for anyone in the process of upgrading a legacy application.<\/p><p>When enabled, the <code class=\"highlighter-rouge\">register_globals<\/code> configuration setting that makes several types of variables (including ones from\n<code class=\"highlighter-rouge\">$_POST<\/code>, <code class=\"highlighter-rouge\">$_GET<\/code> and <code class=\"highlighter-rouge\">$_REQUEST<\/code>) available in the global scope of your application. This can easily lead to security\nissues as your application cannot effectively tell where the data is coming from.<\/p><p>For example: <code class=\"highlighter-rouge\">$_GET['foo']<\/code> would be available via <code class=\"highlighter-rouge\">$foo<\/code>, which can override variables that have not been declared.\nIf you are using PHP &lt; 5.4.0 <strong>make sure<\/strong> that <code class=\"highlighter-rouge\">register_globals<\/code> is <strong>off<\/strong>.<\/p><ul><li><a href=\"https:\/\/secure.php.net\/security.globals\">Register_globals in the PHP manual<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"error_reporting\"><h2 id=\"error_reporting_title\">Error Reporting<\/h2><p>Error logging can be useful in finding the problem spots in your application, but it can also expose information about\nthe structure of your application to the outside world. To effectively protect your application from issues that could\nbe caused by the output of these messages, you need to configure your server differently in development versus\nproduction (live).<\/p><h3 id=\"development\">Development<\/h3><p>To show every possible error during <strong>development<\/strong>, configure the following settings in your <code class=\"highlighter-rouge\">php.ini<\/code>:<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">display_errors = On\ndisplay_startup_errors = On\nerror_reporting = -1\nlog_errors = On<\/pre><\/figure><blockquote>\n  <p>Passing in the value <code class=\"highlighter-rouge\">-1<\/code> will show every possible error, even when new levels and constants are added in future PHP\nversions. The <code class=\"highlighter-rouge\">E_ALL<\/code> constant also behaves this way as of PHP 5.4. -\n<a href=\"https:\/\/secure.php.net\/function.error-reporting\">php.net<\/a><\/p>\n<\/blockquote><p>The <code class=\"highlighter-rouge\">E_STRICT<\/code> error level constant was introduced in 5.3.0 and is not part of <code class=\"highlighter-rouge\">E_ALL<\/code>, however it became part of\n<code class=\"highlighter-rouge\">E_ALL<\/code> in 5.4.0. What does this mean? In terms of reporting every possible error in version 5.3 it means you must\nuse either <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL | E_STRICT<\/code>.<\/p><p><strong>Reporting every possible error by PHP version<\/strong><\/p><ul><li>&lt; 5.3 <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL<\/code><\/li>\n  <li>\u00a0 5.3 <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL | E_STRICT<\/code><\/li>\n  <li>&gt; 5.3 <code class=\"highlighter-rouge\">-1<\/code> or <code class=\"highlighter-rouge\">E_ALL<\/code><\/li>\n<\/ul><h3 id=\"production\">Production<\/h3><p>To hide errors on your <strong>production<\/strong> environment, configure your <code class=\"highlighter-rouge\">php.ini<\/code> as:<\/p><figure class=\"highlight\"><pre class=\"language-ini\" data-lang=\"ini\">display_errors = Off\ndisplay_startup_errors = Off\nerror_reporting = E_ALL\nlog_errors = On<\/pre><\/figure><p>With these settings in production, errors will still be logged to the error logs for the web server, but will not be\nshown to the user. For more information on these settings, see the PHP manual:<\/p><ul><li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.error-reporting\">error_reporting<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.display-errors\">display_errors<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.display-startup-errors\">display_startup_errors<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/errorfunc.configuration#ini.log-errors\">log_errors<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"testing\"><p>Writing automated tests for your PHP code is considered a best practice and can lead to well-built applications.\nAutomated tests are a great tool for making sure your application does not break when you are making changes or adding\nnew functionality and should not be ignored.<\/p><p>There are several different types of testing tools (or frameworks) available for PHP, which use different approaches -\nall of which are trying to avoid manual testing and the need for large Quality Assurance teams, just to make sure\nrecent changes didn\u2019t break existing functionality.<\/p><\/section><section class=\"chapter\" id=\"test_driven_development\"><h2 id=\"test_driven_development_title\">Test Driven Development<\/h2><p>From <a href=\"https:\/\/wikipedia.org\/wiki\/Test-driven_development\">Wikipedia<\/a>:<\/p><blockquote>\n  <p>Test-driven development (TDD) is a software development process that relies on the repetition of a very short\ndevelopment cycle: first the developer writes a failing automated test case that defines a desired improvement or new\nfunction, then produces code to pass that test and finally refactors the new code to acceptable standards. Kent Beck,\nwho is credited with having developed or \u2018rediscovered\u2019 the technique, stated in 2003 that TDD encourages simple\ndesigns and inspires confidence.<\/p>\n<\/blockquote><p>There are several different types of testing that you can do for your application:<\/p><h3>Unit Testing<\/h3><p>Unit Testing is a programming approach to ensure functions, classes and methods are working as expected, from the point\nyou build them all the way through the development cycle. By checking values going in and out of various functions and\nmethods, you can make sure the internal logic is working correctly. By using Dependency Injection and building \u201cmock\u201d\nclasses and stubs you can verify that dependencies are correctly used for even better test coverage.<\/p><p>When you create a class or function you should create a unit test for each behavior it must have. At a very basic level\nyou should make sure it errors if you send it bad arguments and make sure it works if you send it valid arguments. This\nwill help ensure that when you make changes to this class or function later on in the development cycle that the old\nfunctionality continues to work as expected. The only alternative to this would be <code class=\"highlighter-rouge\">var_dump()<\/code> in a test.php, which is\nno way to build an application - large or small.<\/p><p>The other use for unit tests is contributing to open source. If you can write a test that shows broken functionality\n(i.e. fails), then fix it, and show the test passing, patches are much more likely to be accepted. If you run a project\nwhich accepts pull requests then you should suggest this as a requirement.<\/p><p><a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a> is the de-facto testing framework for writing unit tests for PHP applications, but there\nare several alternatives<\/p><ul><li><a href=\"https:\/\/github.com\/atoum\/atoum\">atoum<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/crysalead\/kahlan\">Kahlan<\/a><\/li>\n  <li><a href=\"https:\/\/peridot-php.github.io\/\">Peridot<\/a><\/li>\n  <li><a href=\"http:\/\/simpletest.org\">SimpleTest<\/a><\/li>\n<\/ul><h3 id=\"integration-testing\">Integration Testing<\/h3><p>From <a href=\"https:\/\/wikipedia.org\/wiki\/Integration_testing\">Wikipedia<\/a>:<\/p><blockquote>\n  <p>Integration testing (sometimes called Integration and Testing, abbreviated \u201cI&amp;T\u201d) is the phase in software testing in\nwhich individual software modules are combined and tested as a group. It occurs after unit testing and before\nvalidation testing. Integration testing takes as its input modules that have been unit tested, groups them in larger\naggregates, applies tests defined in an integration test plan to those aggregates, and delivers as its output the\nintegrated system ready for system testing.<\/p>\n<\/blockquote><p>Many of the same tools that can be used for unit testing can be used for integration testing as many of the same\nprinciples are used.<\/p><p>Sometimes also known as acceptance testing, functional testing consists of using tools to create automated tests that\nactually use your application instead of just verifying that individual units of code are behaving correctly and that\nindividual units can speak to each other correctly. These tools typically work using real data and simulating actual\nusers of the application.<\/p><ul><li><a href=\"https:\/\/docs.seleniumhq.org\/\">Selenium<\/a><\/li>\n  <li><a href=\"http:\/\/mink.behat.org\/\">Mink<\/a><\/li>\n  <li><a href=\"https:\/\/codeception.com\/\">Codeception<\/a> is a full-stack testing framework that includes acceptance testing tools<\/li>\n  <li><a href=\"https:\/\/datasift.github.io\/storyplayer\/\">Storyplayer<\/a> is a full-stack testing framework that includes support for creating and destroying test environments on demand<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"behavior_driven_development\"><h2 id=\"behavior_driven_development_title\">Behavior Driven Development<\/h2><p>There are two different types of Behavior-Driven Development (BDD): SpecBDD and StoryBDD. SpecBDD focuses on technical\nbehavior of code, while StoryBDD focuses on business or feature behaviors or interactions. PHP has frameworks for both\ntypes of BDD.<\/p><p>With StoryBDD, you write human-readable stories that describe the behavior of your application. These stories can then\nbe run as actual tests against your application. The framework used in PHP applications for StoryBDD is <a href=\"http:\/\/behat.org\/\">Behat<\/a>, which\nis inspired by Ruby\u2019s <a href=\"https:\/\/cucumber.io\/\">Cucumber<\/a> project and implements the Gherkin DSL for describing feature behavior.<\/p><p>With SpecBDD, you write specifications that describe how your actual code should behave. Instead of testing a function\nor method, you are describing how that function or method should behave. PHP offers the <a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a> framework for this\npurpose. This framework is inspired by the <a href=\"https:\/\/rspec.info\/\">RSpec project<\/a> for Ruby.<\/p><h3 id=\"bdd-links\">BDD Links<\/h3><ul><li><a href=\"http:\/\/behat.org\/\">Behat<\/a>, the StoryBDD framework for PHP, inspired by Ruby\u2019s <a href=\"https:\/\/cucumber.io\/\">Cucumber<\/a> project;<\/li>\n  <li><a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a>, the SpecBDD framework for PHP, inspired by Ruby\u2019s <a href=\"https:\/\/rspec.info\/\">RSpec<\/a> project;<\/li>\n  <li><a href=\"https:\/\/codeception.com\/\">Codeception<\/a> is a full-stack testing framework that uses BDD principles.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"complementary_testing_tools\"><h2 id=\"complementary_testing_tools_title\">Complementary Testing Tools<\/h2><p>Besides individual testing and behavior driven frameworks, there are also a number of generic frameworks and helper\nlibraries useful for any preferred approach taken.<\/p><ul><li><a href=\"https:\/\/www.seleniumhq.org\/\">Selenium<\/a> is a browser automation tool which can be <a href=\"https:\/\/github.com\/giorgiosironi\/phpunit-selenium\/\">integrated with PHPUnit<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/padraic\/mockery\">Mockery<\/a> is a Mock Object Framework which can be integrated with <a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a> or <a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/phpspec\/prophecy\">Prophecy<\/a> is a highly opinionated yet very powerful and flexible PHP object mocking framework. It\u2019s integrated with\n<a href=\"https:\/\/www.phpspec.net\/\">PHPSpec<\/a> and can be used with <a href=\"https:\/\/phpunit.de\/\">PHPUnit<\/a>.<\/li>\n  <li><a href=\"https:\/\/github.com\/php-mock\/php-mock\">php-mock<\/a>  is a library to help to mock PHP native functions.<\/li>\n  <li><a href=\"https:\/\/github.com\/infection\/infection\">Infection<\/a> is a PHP implementation of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Mutation_testing\">Mutation Testing<\/a> to help to measure the effectiveness of your tests.<\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"servers_and_deployment\"><p>PHP applications can be deployed and run on production web servers in a number of ways.<\/p><\/section><section class=\"chapter\" id=\"platform_as_a_service\"><h2 id=\"platform_as_a_service_title\">Platform as a Service (PaaS)<\/h2><p>PaaS provides the system and network architecture necessary to run PHP applications on the web. This means little to no\nconfiguration for launching PHP applications or PHP frameworks.<\/p><p>Recently PaaS has become a popular method for deploying, hosting, and scaling PHP applications of all sizes. You can\nfind a list of <a href=\"#php_paas_providers\">PHP PaaS \u201cPlatform as a Service\u201d providers<\/a> in our <a href=\"#resources\">resources section<\/a>.<\/p><\/section><section class=\"chapter\" id=\"virtual_or_dedicated_servers\"><h2 id=\"virtual_or_dedicated_servers_title\">Virtual or Dedicated Servers<\/h2><p>If you are comfortable with systems administration, or are interested in learning it, virtual or dedicated servers give\nyou complete control of your application\u2019s production environment.<\/p><h3 id=\"nginx-and-php-fpm\">nginx and PHP-FPM<\/h3><p>PHP, via PHP\u2019s built-in FastCGI Process Manager (FPM), pairs really nicely with <a href=\"https:\/\/nginx.org\/\">nginx<\/a>, which is a lightweight,\nhigh-performance web server. It uses less memory than Apache and can better handle more concurrent requests. This is\nespecially important on virtual servers that don\u2019t have much memory to spare.<\/p><ul><li><a href=\"https:\/\/nginx.org\/\">Read more on nginx<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/install.fpm\">Read more on PHP-FPM<\/a><\/li>\n  <li><a href=\"https:\/\/nealpoole.com\/blog\/2011\/04\/setting-up-php-fastcgi-and-nginx-dont-trust-the-tutorials-check-your-configuration\/\">Read more on setting up nginx and PHP-FPM securely<\/a><\/li>\n<\/ul><h3 id=\"apache-and-php\">Apache and PHP<\/h3><p>PHP and Apache have a long history together. Apache is wildly configurable and has many available\n<a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/\">modules<\/a> to extend functionality. It is a popular choice for shared servers and an easy setup for PHP\nframeworks and open source apps like WordPress. Unfortunately, Apache uses more resources than nginx by default and\ncannot handle as many visitors at the same time.<\/p><p>Apache has several possible configurations for running PHP. The most common and easiest to setup is the <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/prefork.html\">prefork MPM<\/a>\nwith mod_php5. While it isn\u2019t the most memory efficient, it is the simplest to get working and to use. This is probably\nthe best choice if you don\u2019t want to dig too deeply into the server administration aspects. Note that if you use\nmod_php5 you MUST use the prefork MPM.<\/p><p>Alternatively, if you want to squeeze more performance and stability out of Apache then you can take advantage of the\nsame FPM system as nginx and run the <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/worker.html\">worker MPM<\/a> or <a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/event.html\">event MPM<\/a> with mod_fastcgi or mod_fcgid. This configuration will\nbe significantly more memory efficient and much faster but it is more work to set up.<\/p><p>If you are running Apache 2.4 or later, you can use <a href=\"https:\/\/httpd.apache.org\/docs\/current\/mod\/mod_proxy_fcgi.html\">mod_proxy_fcgi<\/a> to get great performance that is easy to setup.<\/p><ul><li><a href=\"https:\/\/httpd.apache.org\/\">Read more on Apache<\/a><\/li>\n  <li><a href=\"https:\/\/httpd.apache.org\/docs\/2.4\/mod\/mpm_common.html\">Read more on Multi-Processing Modules<\/a><\/li>\n  <li><a href=\"https:\/\/blogs.oracle.com\/opal\/entry\/php_fpm_fastcgi_process_manager\">Read more on mod_fastcgi<\/a><\/li>\n  <li><a href=\"denied:hhttps:\/\/httpd.apache.org\/mod_fcgid\/\">Read more on mod_fcgid<\/a><\/li>\n  <li><a href=\"https:\/\/httpd.apache.org\/docs\/current\/mod\/mod_proxy_fcgi.html\">Read more on mod_proxy_fcgi<\/a><\/li>\n  <li><a href=\"https:\/\/serversforhackers.com\/video\/apache-and-php-fpm\">Read more on setting up Apache and PHP-FPM with mod_proxy_fcgi<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"building_and_deploying_your_application\"><h2 id=\"building_and_deploying_your_application_title\">Building and Deploying your Application<\/h2><p>If you find yourself doing manual database schema changes or running your tests manually before updating your files\n(manually), think twice! With every additional manual task needed to deploy a new version of your app, the chances for\npotentially fatal mistakes increase. Whether you\u2019re dealing with a simple update, a comprehensive build process or even\na continuous integration strategy, <a href=\"https:\/\/wikipedia.org\/wiki\/Build_automation\">build automation<\/a> is your friend.<\/p><p>Among the tasks you might want to automate are:<\/p><ul><li>Dependency management<\/li>\n  <li>Compilation, minification of your assets<\/li>\n  <li>Running tests<\/li>\n  <li>Creation of documentation<\/li>\n  <li>Packaging<\/li>\n  <li>Deployment<\/li>\n<\/ul><h3 id=\"deployment-tools\">Deployment Tools<\/h3><p>Deployment tools can be described as a collection of scripts that handle common tasks of software deployment. The deployment tool is not a part of your software, it acts on your software from \u2018outside\u2019.<\/p><p>There are many open source tools available to help you with build automation and deployment, some are written in PHP others aren\u2019t. This shouldn\u2019t hold you back from using them, if they\u2019re better suited for the specific job. Here are a few examples:<\/p><p><a href=\"https:\/\/www.phing.info\/\">Phing<\/a> can control your packaging, deployment or testing process from within a XML build file. Phing (which is based on <a href=\"https:\/\/ant.apache.org\/\">Apache Ant<\/a>) provides a rich set of tasks usually needed to install or update a web application and can be extended with additional custom tasks, written in PHP. It\u2019s a solid and robust tool and has been around for a long time, however the tool could be perceived as a bit old fashioned because of the way it deals with configuration (XML files).<\/p><p><a href=\"http:\/\/capistranorb.com\/\">Capistrano<\/a> is a system for <em>intermediate-to-advanced programmers<\/em> to execute commands in a structured, repeatable way on one or more remote machines. It is pre-configured for deploying Ruby on Rails applications, however you can successfully deploy PHP systems with it. Successful use of Capistrano depends on a working knowledge of Ruby and Rake.<\/p><p><a href=\"https:\/\/ansistrano.com\">Ansistrano<\/a> is a couple of Ansible roles to easily manage the deployment process (deploy and rollback) for scripting applications such as PHP, Python and Ruby. It\u2019s an Ansible port for <a href=\"http:\/\/capistranorb.com\/\">Capistrano<\/a>. It\u2019s been used by quite a lot of PHP companies already.<\/p><p><a href=\"http:\/\/rocketeer.autopergamene.eu\/\">Rocketeer<\/a> gets its inspiration and philosophy from the Laravel framework. Its goal is to be fast, elegant and easy to use with smart defaults. It features multiple servers, multiple stages, atomic deploys and deployment can be performed in parallel. Everything in the tool can be hot swapped or extended, and everything is written in PHP.<\/p><p><a href=\"https:\/\/deployer.org\/\">Deployer<\/a> is a deployment tool written in PHP. It\u2019s simple and functional. Features include running tasks in parallel, atomic deployment and keeping consistency between servers. Recipes of common tasks for Symfony, Laravel, Zend Framework and Yii are available. Younes Rafie\u2019s article  <a href=\"https:\/\/www.sitepoint.com\/deploying-php-applications-with-deployer\/\">Easy Deployment of PHP Applications with Deployer<\/a> is a great tutorial for deploying your application with the tool.<\/p><p><a href=\"https:\/\/www.magephp.com\/\">Magallanes<\/a> is another tool written in PHP with simple configuration done in YAML files. It has support for multiple servers and environments, atomic deployment, and has some built in tasks that you can leverage for common tools and frameworks.<\/p><h4>Further reading:<\/h4><ul><li><a href=\"https:\/\/code.tutsplus.com\/tutorials\/automate-your-projects-with-apache-ant--net-18595\">Automate your project with Apache Ant<\/a><\/li>\n  <li><a href=\"https:\/\/deployingphpapplications.com\/\">Deploying PHP Applications<\/a> - paid book on best practices and tools for PHP deployment.<\/li>\n<\/ul><h3 id=\"server-provisioning\">Server Provisioning<\/h3><p>Managing and configuring servers can be a daunting task when faced with many servers. There are tools for dealing with this so you can automate your infrastructure to make sure you have the right servers and that they\u2019re configured properly. They often integrate with the larger cloud hosting providers (Amazon Web Services, Heroku, DigitalOcean, etc) for managing instances, which makes scaling an application a lot easier.<\/p><p><a href=\"https:\/\/www.ansible.com\/\">Ansible<\/a> is a tool that manages your infrastructure through YAML files. It\u2019s simple to get started with and can manage complex and large scale applications. There is an API for managing cloud instances and it can manage them through a dynamic inventory using certain tools.<\/p><p><a href=\"https:\/\/puppet.com\/\">Puppet<\/a> is a tool that has its own language and file types for managing servers and configurations. It can be used in a master\/client setup or it can be used in a \u201cmaster-less\u201d mode. In the master\/client mode the clients will poll the central master(s) for new configuration on set intervals and update itself if necessary. In the master-less mode you can push changes to your nodes.<\/p><p><a href=\"https:\/\/www.chef.io\/\">Chef<\/a> is a powerful Ruby based system integration framework that you can build your whole server environment or virtual boxes with. It integrates well with Amazon Web Services through their service called OpsWorks.<\/p><h4 id=\"further-reading-1\">Further reading:<\/h4><ul><li><a href=\"https:\/\/serversforhackers.com\/an-ansible-tutorial\">An Ansible Tutorial<\/a><\/li>\n  <li><a href=\"https:\/\/leanpub.com\/ansible-for-devops\">Ansible for DevOps<\/a> - paid book on everything Ansible<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/ansible-for-aws\">Ansible for AWS<\/a> - paid book on integrating Ansible and Amazon Web Services<\/li>\n  <li><a href=\"http:\/\/www.jasongrimes.org\/2012\/06\/managing-lamp-environments-with-chef-vagrant-and-ec2-1-of-3\/\">Three part blog series about deploying a LAMP application with Chef, Vagrant, and EC2<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/chef-cookbooks\/php\">Chef Cookbook which installs and configures PHP and the PEAR package management system<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/playlist?list=PL11cZfNdwNyPnZA9D1MbVqldGuOWqbumZ\">Chef video tutorial series<\/a><\/li>\n<\/ul><h3 id=\"continuous-integration\">Continuous Integration<\/h3><blockquote>\n  <p>Continuous Integration is a software development practice where members of a team integrate their work frequently,\nusually each person integrates at least daily \u2014 leading to multiple integrations per day. Many teams find that this\napproach leads to significantly reduced integration problems and allows a team to develop cohesive software more\nrapidly.<\/p>\n<\/blockquote><p><em>\u2013 Martin Fowler<\/em><\/p><p>There are different ways to implement continuous integration for PHP. <a href=\"https:\/\/travis-ci.org\/\">Travis CI<\/a> has done a great job of\nmaking continuous integration a reality even for small projects. Travis CI is a hosted continuous integration service\nfor the open source community. It is integrated with GitHub and offers first class support for many languages including\nPHP.<\/p><h4 id=\"further-reading-2\">Further reading:<\/h4><ul><li><a href=\"https:\/\/jenkins.io\/\">Continuous Integration with Jenkins<\/a><\/li>\n  <li><a href=\"https:\/\/www.phptesting.org\/\">Continuous Integration with PHPCI<\/a><\/li>\n  <li><a href=\"https:\/\/www.jetbrains.com\/teamcity\/\">Continuous Integration with Teamcity<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"virtualization\"><p>Running your application on different environments in development and production can lead to strange bugs popping up\nwhen you go live. It\u2019s also tricky to keep different development environments up to date with the same version for all\nlibraries used when working with a team of developers.<\/p><p>If you are developing on Windows and deploying to Linux (or anything non-Windows) or are developing in a team, you\nshould consider using a virtual machine. This sounds tricky, but besides the widely known virtualization environments\nlike VMware or VirtualBox, there are additional tools that may help you setting up a virtual environment in a few easy\nsteps.<\/p><\/section><section class=\"chapter\" id=\"vagrant\"><h2 id=\"vagrant_title\">Vagrant<\/h2><p><a href=\"https:\/\/www.vagrantup.com\/\">Vagrant<\/a> helps you build your virtual boxes on top of the known virtual environments and will configure these\nenvironments based on a single configuration file. These boxes can be set up manually, or you can use \u201cprovisioning\u201d\nsoftware such as <a href=\"https:\/\/puppet.com\/\">Puppet<\/a> or <a href=\"https:\/\/www.chef.io\/\">Chef<\/a> to do this for you. Provisioning the base box is a great way to ensure that\nmultiple boxes are set up in an identical fashion and removes the need for you to maintain complicated \u201cset up\u201d\ncommand lists. You can also \u201cdestroy\u201d your base box and recreate it without many manual steps, making it easy to create\na \u201cfresh\u201d installation.<\/p><p>Vagrant creates folders for sharing your code between your host and your virtual machine, which means that you can\ncreate and edit your files on your host machine and then run the code inside your virtual machine.<\/p><h3 id=\"a-little-help\">A little help<\/h3><p>If you need a little help to start using Vagrant there are some services that might be useful:<\/p><ul><li><a href=\"http:\/\/rove.io\/\">Rove<\/a>: service that allows you to pre-generate typical Vagrant builds, PHP among the options. The provisioning is\nmade with Chef.<\/li>\n  <li><a href=\"https:\/\/puphpet.com\/\">Puphpet<\/a>: simple GUI to set up virtual machines for PHP development. <strong>Heavily focused in PHP<\/strong>. Besides local VMs,\nit can be used to deploy to cloud services as well. The provisioning is made with Puppet.<\/li>\n  <li><a href=\"https:\/\/www.getprotobox.com\/\">Protobox<\/a>: is a layer on top of vagrant and a web GUI to setup virtual machines for web development. A single YAML\ndocument controls everything that is installed on the virtual machine.<\/li>\n  <li><a href=\"http:\/\/phansible.com\/\">Phansible<\/a>: provides an easy to use interface that helps you generate Ansible Playbooks for PHP based projects.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"docker\"><h2 id=\"docker_title\">Docker<\/h2><p><a href=\"https:\/\/www.docker.com\/\">Docker<\/a> - a lightweight alternative to a full virtual machine - is so called because it\u2019s all about \u201ccontainers\u201d.  A container is a building block which, in the simplest case, does one specific job, e.g. running a web server.  An \u201cimage\u201d is the package you use to build the container - Docker has a repository full of them.<\/p><p>A typical LAMP application might have three containers: a web server, a PHP-FPM process and MySQL. As with shared folders in Vagrant, you can leave your application files where they are and tell Docker where to find them.<\/p><p>You can generate containers from the command line (see example below) or, for ease of maintenance, build a <code class=\"highlighter-rouge\">docker-compose.yml<\/code> file for your project specifying which to create and how they communicate with one another.<\/p><p>Docker may help if you\u2019re developing multiple websites and want the separation that comes from installing each on it\u2019s own virtual machine, but don\u2019t have the necessary disk space or the time to keep everything up to date.  It\u2019s efficient: the installation and downloads are quicker, you only need to store one copy of each image however often it\u2019s used, containers need less RAM and share the same OS kernel, so you can have more servers running simultaneously, and it takes a matter of seconds to stop and start them, no need to wait for a full server boot.<\/p><h3 id=\"example-running-your-php-applications-in-docker\">Example: Running your PHP Applications in Docker<\/h3><p>After <a href=\"https:\/\/docs.docker.com\/install\/\">installing docker<\/a> on your machine, you can start a web server with one command.\nThe following will download a fully functional Apache installation with the latest PHP version, map <code class=\"highlighter-rouge\">\/path\/to\/your\/php\/files<\/code> to the document root, which you can view at <code class=\"highlighter-rouge\">http:\/\/localhost:8080<\/code>:<\/p><figure class=\"highlight\"><pre class=\"language-console\" data-lang=\"console\">docker run -d --name my-php-webserver -p 8080:80 -v \/path\/to\/your\/php\/files:\/var\/www\/html\/ php:apache<\/pre><\/figure><p>This will initialize and launch your container. <code class=\"highlighter-rouge\">-d<\/code> makes it runs in the background. To stop and start it, simply run <code class=\"highlighter-rouge\">docker stop my-php-webserver<\/code> and <code class=\"highlighter-rouge\">docker start my-php-webserver<\/code> (the other parameters are not needed again).<\/p><p>The command above shows a quick way to run a basic server.  There\u2019s much more you can do (and thousands of pre-built images in the <a href=\"https:\/\/hub.docker.com\/\">Docker Hub<\/a>). Take time to learn the terminology and read the <a href=\"https:\/\/docs.docker.com\/\">Docker User Guide<\/a> to get the most from it, and don\u2019t run random code you\u2019ve downloaded without checking it\u2019s safe \u2013 unofficial images may not have the latest security patches. If in doubt, stick to the <a href=\"https:\/\/hub.docker.com\/explore\/\">official repositiories<\/a>.<\/p><p>The <a href=\"https:\/\/phpdocker.io\/generator\">PHPDocker.io<\/a> site will auto-generate all the files you need for a fully-featured LAMP\/LEMP stack, including your choice of PHP version and extensions.<\/p><ul><li><a href=\"https:\/\/www.docker.com\/\">Docker Website<\/a><\/li>\n  <li><a href=\"https:\/\/docs.docker.com\/install\/\">Docker Installation<\/a><\/li>\n  <li><a href=\"https:\/\/docs.docker.com\/\">Docker User Guide<\/a><\/li>\n  <li><a href=\"https:\/\/hub.docker.com\/\">Docker Hub<\/a><\/li>\n  <li><a href=\"https:\/\/hub.docker.com\/explore\/\">Docker Hub - official images<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"caching\"><p>PHP is pretty quick by itself, but bottlenecks can arise when you make remote connections, load files, etc. \nThankfully, there are various tools available to speed up certain parts of your application, or reduce the number of times these various time-consuming tasks need to run.<\/p><\/section><section class=\"chapter\" id=\"opcode_cache\"><h2 id=\"opcode_cache_title\">Opcode Cache<\/h2><p>When a PHP file is executed, it must first be compiled into <a href=\"https:\/\/secure.php.net\/manual\/internals2.opcodes.php\">opcodes<\/a> (machine language instructions for the CPU). If the source code is unchanged, the opcodes will be the same, so this compilation step becomes a waste of CPU resources.<\/p><p>An opcode cache prevents redundant compilation by storing opcodes in memory and reusing them on successive calls. It will typically check signature or modification time of the file first, in case there have been any changes.<\/p><p>It\u2019s likely an opcode cache will make a significant speed improvement to your application.  Since PHP 5.5 there is one built in - <a href=\"https:\/\/secure.php.net\/book.opcache\">Zend OPcache<\/a>. Depending on your PHP package\/distribution, it\u2019s usually turned on by default -\u00a0check <a href=\"https:\/\/secure.php.net\/manual\/opcache.configuration.php#ini.opcache.enable\">opcache.enable<\/a> and the output of <code class=\"highlighter-rouge\">phpinfo()<\/code> to make sure. For earlier versions there\u2019s a PECL extension.<\/p><p>Read more about opcode caches:<\/p><ul><li><a href=\"https:\/\/secure.php.net\/book.opcache\">Zend OPcache<\/a> (bundled with PHP since 5.5)<\/li>\n  <li>Zend OPcache (formerly known as Zend Optimizer+) is now <a href=\"https:\/\/github.com\/zendtech\/ZendOptimizerPlus\">open source<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/book.apc\">APC<\/a> - PHP 5.4 and earlier<\/li>\n  <li><a href=\"https:\/\/xcache.lighttpd.net\/\">XCache<\/a><\/li>\n  <li><a href=\"https:\/\/www.iis.net\/downloads\/microsoft\/wincache-extension\">WinCache<\/a> (extension for MS Windows Server)<\/li>\n  <li><a href=\"https:\/\/wikipedia.org\/wiki\/List_of_PHP_accelerators\">list of PHP accelerators on Wikipedia<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"object_caching\"><h2 id=\"object_caching_title\">Object Caching<\/h2><p>There are times when it can be beneficial to cache individual objects in your code, such as with data that is expensive\nto get or database calls where the result is unlikely to change. You can use object caching software to hold these\npieces of data in memory for extremely fast access later on. If you save these items to a data store after you retrieve\nthem, then pull them directly from the cache for following requests, you can gain a significant improvement in\nperformance as well as reduce the load on your database servers.<\/p><p>Many of the popular bytecode caching solutions let you cache custom data as well, so there\u2019s even more reason to take\nadvantage of them. APCu, XCache, and WinCache all provide APIs to save data from your PHP code to their memory cache.<\/p><p>The most commonly used memory object caching systems are APCu and memcached. APCu is an excellent choice for object\ncaching, it includes a simple API for adding your own data to its memory cache and is very easy to setup and use. The\none real limitation of APCu is that it is tied to the server it\u2019s installed on. Memcached on the other hand is\ninstalled as a separate service and can be accessed across the network, meaning that you can store objects in a\nhyper-fast data store in a central location and many different systems can pull from it.<\/p><p>Note that when running PHP as a (Fast-)CGI application inside your webserver, every PHP process will have its own cache,\ni.e. APCu data is not shared between your worker processes. In these cases, you might want to consider using memcached\ninstead, as it\u2019s not tied to the PHP processes.<\/p><p>In a networked configuration APCu will usually outperform memcached in terms of access speed, but memcached will be\nable to scale up faster and further. If you do not expect to have multiple servers running your application, or do not\nneed the extra features that memcached offers then APCu is probably your best choice for object caching.<\/p><p>Example logic using APCu:<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/\/ check if there is data saved as 'expensive_data' in cache\n$data = apc_fetch('expensive_data');\nif ($data === false) {\n    \/\/ data is not in cache; save result of expensive call for later use\n    apc_add('expensive_data', $data = get_expensive_data());\n}\nprint_r($data);<\/pre><\/figure><p>Note that prior to PHP 5.5, APC provides both an object cache and a bytecode cache. APCu is a project to bring APC\u2019s\nobject cache to PHP 5.5+, since PHP now has a built-in bytecode cache (OPcache).<\/p><ul><li><a href=\"https:\/\/github.com\/krakjoe\/apcu\">APCu<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.apc\">APC Functions<\/a><\/li>\n  <li><a href=\"https:\/\/memcached.org\/\">Memcached<\/a><\/li>\n  <li><a href=\"https:\/\/redis.io\/\">Redis<\/a><\/li>\n  <li><a href=\"https:\/\/xcache.lighttpd.net\/wiki\/XcacheApi\">XCache APIs<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/ref.wincache\">WinCache Functions<\/a><\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"phpdoc\"><h2 id=\"phpdoc_title\">PHPDoc<\/h2><p>PHPDoc is an informal standard for commenting PHP code. There are a <em>lot<\/em> of different <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/index.html\">tags<\/a> available. The full list\nof tags and examples can be found at the <a href=\"https:\/\/docs.phpdoc.org\/index.html\">PHPDoc manual<\/a>.<\/p><p>Below is an example of how you might document a class with a few methods;<\/p><figure class=\"highlight\"><pre class=\"language-php\" data-lang=\"php\">&lt;?php\n\/**\n * @author A Name &lt;a.name@example.com&gt;\n * @link http:\/\/www.phpdoc.org\/docs\/latest\/index.html\n *\/\nclass DateTimeHelper\n{\n    \/**\n     * @param mixed $anything Anything that we can convert to a \\DateTime object\n     *\n     * @throws \\InvalidArgumentException\n     *\n     * @return \\DateTime\n     *\/\n    public function dateTimeFromAnything($anything)\n    {\n        $type = gettype($anything);\n        switch ($type) {\n            \/\/ Some code that tries to return a \\DateTime object\n        }\n        throw new \\InvalidArgumentException(\n            \"Failed Converting param of type '{$type}' to DateTime object\"\n        );\n    }\n    \/**\n     * @param mixed $date Anything that we can convert to a \\DateTime object\n     *\n     * @return void\n     *\/\n    public function printISO8601Date($date)\n    {\n        echo $this-&gt;dateTimeFromAnything($date)-&gt;format('c');\n    }\n    \/**\n     * @param mixed $date Anything that we can convert to a \\DateTime object\n     *\/\n    public function printRFC2822Date($date)\n    {\n        echo $this-&gt;dateTimeFromAnything($date)-&gt;format('r');\n    }\n}<\/pre><\/figure><p>The documentation for the class as a whole has the <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/author.html\">@author<\/a> tag and a <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/link.html\">@link<\/a> tag. The <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/author.html\">@author<\/a> tag is used to\ndocument the author of the code and can be repeated for documenting several authors. The <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/link.html\">@link<\/a> tag is used to link to a website indicating a relationship between the website and the code.<\/p><p>Inside the class, the first method has a <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/param.html\">@param<\/a> tag documenting the type, name and description of the parameter\nbeing passed to the method. Additionally it has the <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/return.html\">@return<\/a> and <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/throws.html\">@throws<\/a> tags for documenting the return type, and any exceptions that could be thrown respectively.<\/p><p>The second and third methods are very similar and have a single <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/param.html\">@param<\/a> tag as did the first method. The important\ndifference between the second and third methods\u2019 doc block is the inclusion\/exclusion of the <a href=\"https:\/\/docs.phpdoc.org\/references\/phpdoc\/tags\/return.html\">@return<\/a> tag.\n<code class=\"highlighter-rouge\">@return void<\/code> explicitly informs us that there is no return; historically omitting the <code class=\"highlighter-rouge\">@return void<\/code> statement also results in the same (no return) action.<\/p><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"from_the_source\"><h2 id=\"from_the_source_title\">From the Source<\/h2><ul><li><a href=\"https:\/\/secure.php.net\/\">PHP Website<\/a><\/li>\n  <li><a href=\"https:\/\/secure.php.net\/docs.php\">PHP Documentation<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"people_to_follow\"><h2 id=\"people_to_follow_title\">People to Follow<\/h2><p>It\u2019s difficult to find interesting and knowledgeable PHP\ncommunity members when you are first starting out. You can\nfind an abbreviated list of PHP community members to get you started at:<\/p><ul><li><a href=\"https:\/\/www.ogprogrammer.com\/2017\/06\/28\/how-to-get-connected-with-the-php-community\/\">https:\/\/www.ogprogrammer.com\/2017\/06\/28\/how-to-get-connected-with-the-php-community\/<\/a><\/li>\n  <li><a href=\"https:\/\/twitter.com\/CalEvans\/lists\/phpeople\">https:\/\/twitter.com\/CalEvans\/lists\/phpeople<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"mentoring\"><h2 id=\"mentoring_title\">Mentoring<\/h2><ul><li><a href=\"https:\/\/php-mentoring.org\/\">php-mentoring.org<\/a> - Formal, peer to peer mentoring in the PHP community.<\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"php_paas_providers\"><h2 id=\"php_paas_providers_title\">PHP PaaS Providers<\/h2><ul><li><a href=\"https:\/\/www.ctl.io\/appfog\/\">AppFog<\/a><\/li>\n  <li><a href=\"https:\/\/aws.amazon.com\/elasticbeanstalk\/\">AWS Elastic Beanstalk<\/a><\/li>\n  <li><a href=\"https:\/\/www.cloudways.com\/\">Cloudways<\/a><\/li>\n  <li><a href=\"https:\/\/www.engineyard.com\/features\">Engine Yard Cloud<\/a><\/li>\n  <li><a href=\"https:\/\/www.fortrabbit.com\/\">fortrabbit<\/a><\/li>\n  <li><a href=\"https:\/\/cloud.google.com\/appengine\/docs\/php\/\">Google App Engine<\/a><\/li>\n  <li><a href=\"https:\/\/devcenter.heroku.com\/categories\/php-support\">Heroku<\/a><\/li>\n  <li><a href=\"https:\/\/console.bluemix.net\/docs\/runtimes\/php\/getting-started.html#getting_started\">IBM Cloud<\/a><\/li>\n  <li><a href=\"https:\/\/jelastic.com\/\">Jelastic<\/a><\/li>\n  <li><a href=\"https:\/\/azure.microsoft.com\/\">Microsoft Azure<\/a><\/li>\n  <li><a href=\"https:\/\/nanobox.io\/\">Nanobox<\/a><\/li>\n  <li><a href=\"https:\/\/run.pivotal.io\/\">Pivotal Web Services<\/a><\/li>\n  <li><a href=\"https:\/\/platform.sh\/\">Platform.sh<\/a><\/li>\n  <li><a href=\"https:\/\/www.openshift.com\/\">Red Hat OpenShift<\/a><\/li>\n<\/ul><p>To see which versions these PaaS hosts are running, head over to <a href=\"http:\/\/phpversions.info\/paas-hosting\/\">PHP Versions<\/a>.<\/p><\/section><section class=\"chapter\" id=\"frameworks\"><h2 id=\"frameworks_title\">Frameworks<\/h2><p>Rather than re-invent the wheel, many PHP developers use frameworks to build out web applications. Frameworks abstract\naway many of the low-level concerns and provide helpful, easy-to-use interfaces to complete common tasks.<\/p><p>You do not need to use a framework for every project. Sometimes plain PHP is the right way to go, but if you do need a\nframework then there are three main types available:<\/p><ul><li>Micro Frameworks<\/li>\n  <li>Full-Stack Frameworks<\/li>\n  <li>Component Frameworks<\/li>\n<\/ul><p>Micro-frameworks are essentially a wrapper to route a HTTP request to a callback, controller, method, etc as quickly as\npossible, and sometimes come with a few extra libraries to assist development such as basic database wrappers and the\nlike. They are prominently used to build remote HTTP services.<\/p><p>Many frameworks add a considerable number of features on top of what is available in a micro-framework; these are\ncalled Full-Stack Frameworks. These often come bundled with ORMs, Authentication packages, etc.<\/p><p>Component-based frameworks are collections of specialized and single-purpose libraries. Disparate component-based\nframeworks can be used together to make a micro- or full-stack framework.<\/p><\/section><section class=\"chapter\" id=\"components\"><h2 id=\"components_title\">Components<\/h2><p>As mentioned above \u201cComponents\u201d are another approach to the common goal of creating, distributing and implementing\nshared code. Various component repositories exist, the main two of which are:<\/p><p>Both of these repositories have command line tools associated with them to help the installation and upgrade processes,\nand have been explained in more detail in the <a href=\"https:\/\/phptherightway.com\/#dependency_management\">Dependency Management<\/a> section.<\/p><p>There are also component-based frameworks and component-vendors that offer no framework at all. These projects provide\nanother source of packages which ideally have little to no dependencies on other packages, or specific frameworks.<\/p><p>For example, you can use the <a href=\"https:\/\/github.com\/fuelphp\/validation\">FuelPHP Validation package<\/a>, without needing to use the FuelPHP framework itself.<\/p><ul><li><a href=\"http:\/\/auraphp.com\/framework\/\">Aura<\/a><\/li>\n  <li>CakePHP Components\n    <ul><li><a href=\"https:\/\/github.com\/cakephp\/collection\">Collection<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/database\">Database<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/datasource\">Datasource<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/event\">Event<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/i18n\">I18n<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/cakephp\/orm\">ORM<\/a><\/li>\n    <\/ul><\/li>\n  <li><a href=\"https:\/\/github.com\/fuelphp\">FuelPHP<\/a><\/li>\n  <li><a href=\"https:\/\/github.com\/hoaproject\">Hoa Project<\/a><\/li>\n  <li><a href=\"https:\/\/symfony.com\/doc\/current\/components\/index.html\">Symfony Components<\/a><\/li>\n  <li><a href=\"https:\/\/thephpleague.com\/\">The League of Extraordinary Packages<\/a><\/li>\n  <li>Laravel\u2019s Illuminate Components\n    <ul><li><a href=\"https:\/\/github.com\/illuminate\/container\">IoC Container<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/illuminate\/database\">Eloquent ORM<\/a><\/li>\n      <li><a href=\"https:\/\/github.com\/illuminate\/queue\">Queue<\/a><\/li>\n    <\/ul><\/li>\n<\/ul><p><em>Laravel\u2019s <a href=\"https:\/\/github.com\/illuminate\">Illuminate components<\/a> will become better decoupled from the Laravel framework. For now, only the\ncomponents best decoupled from the Laravel framework are listed above.<\/em><\/p><\/section><section class=\"chapter\" id=\"other_resources\"><h2 id=\"other_resources_title\">Other Useful Resources<\/h2><h3 id=\"cheatsheets\">Cheatsheets<\/h3><ul><li><a href=\"http:\/\/phpcheatsheets.com\/\">PHP Cheatsheets<\/a> - for variable comparisons, arithmetics and variable testing in various\nPHP versions<\/li>\n  <li><a href=\"https:\/\/www.owasp.org\/index.php\/PHP_Security_Cheat_Sheet\">PHP Security Cheatsheet<\/a><\/li>\n<\/ul><h3 id=\"more-best-practices\">More best practices<\/h3><ul><li><a href=\"https:\/\/phpbestpractices.org\/\">PHP Best Practices<\/a><\/li>\n  <li><a href=\"https:\/\/www.airpair.com\/php\/posts\/best-practices-for-modern-php-development\">Best practices for Modern PHP Development<\/a><\/li>\n  <li><a href=\"https:\/\/kinsta.com\/blog\/php-versions\/\">Why You Should Be Using Supported PHP Versions<\/a><\/li>\n<\/ul><h3 id=\"news-around-the-php-and-web-development-communities\">News around the PHP and web development communities<\/h3><p>You can subscribe to weekly newsletters to keep yourself informed on new libraries, latest news, events and general\nannouncements, as well as additional resources being published every now and then:<\/p><ul><li><a href=\"http:\/\/www.phpweekly.com\">PHP Weekly<\/a><\/li>\n  <li><a href=\"https:\/\/javascriptweekly.com\/\">JavaScript Weekly<\/a><\/li>\n  <li><a href=\"https:\/\/frontendfoc.us\/\">Frontend Focus<\/a><\/li>\n  <li><a href=\"https:\/\/mobiledevweekly.com\/\">Mobile Web Weekly<\/a><\/li>\n<\/ul><p>There are also Weeklies on other platforms you might be interested in; here\u2019s <a href=\"https:\/\/github.com\/jondot\/awesome-weekly\">a list of some<\/a>.<\/p><h3 id=\"php-universe\">PHP universe<\/h3><ul><li><a href=\"https:\/\/blog.phpdeveloper.org\/\">PHP Developer blog<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"videos\"><h2>Video Tutorials<\/h2><h3 id=\"youtube-channels\">YouTube Channels<\/h3><ul><li><a href=\"https:\/\/www.youtube.com\/user\/phpacademy\">PHP Academy<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/user\/thenewboston\">The New Boston<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/user\/businessgeek\">Sherif Ramadan<\/a><\/li>\n  <li><a href=\"https:\/\/www.youtube.com\/user\/LevelUpTuts\">Level Up Tuts<\/a><\/li>\n<\/ul><h3 id=\"paid-videos\">Paid Videos<\/h3><ul><li><a href=\"https:\/\/teamtreehouse.com\/library\/php-standards-and-best-practices\">Standards and Best practices<\/a><\/li>\n  <li><a href=\"https:\/\/www.pluralsight.com\/search?q=php\">PHP Training on Pluralsight<\/a><\/li>\n  <li><a href=\"https:\/\/www.lynda.com\/search?q=php\">PHP Training on Lynda.com<\/a><\/li>\n  <li><a href=\"https:\/\/code.tutsplus.com\/categories\/php\/courses\">PHP Training on Tutsplus<\/a><\/li>\n  <li><a href=\"https:\/\/laracasts.com\/\">Laracasts<\/a><\/li>\n<\/ul><\/section><section class=\"chapter\" id=\"books\"><h2 id=\"books_title\">Books<\/h2><p>There are many PHP books; sadly some are now quite old and no longer accurate. In particular, avoid books on \u201cPHP 6\u201d, a version that will now never exist. The next major release of PHP after 5.6 was \u201cPHP 7\u201d, <a href=\"https:\/\/wiki.php.net\/rfc\/php6\">partly because of this<\/a>.<\/p><p>This section aims to be a living document for recommended books on PHP development in general. If you would like your\nbook to be added, send a PR and it will be reviewed for relevancy.<\/p><h3 id=\"free-books\">Free Books<\/h3><ul><li><a href=\"http:\/\/daylerees.com\/php-pandas\/\">PHP Pandas<\/a> - Aims to teach everyone how to be a web developer.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/phptherightway\/\">PHP The Right Way<\/a> - This website is available as a book completely for free.<\/li>\n  <li><a href=\"https:\/\/paragonie.com\/book\/pecl-libsodium\">Using Libsodium in PHP Projects<\/a> - Guide to using Libsodium PHP extension\nfor modern, secure, and fast cryptography.<\/li>\n<\/ul><h3 id=\"paid-books\">Paid Books<\/h3><ul><li><a href=\"https:\/\/apisyouwonthate.com\/\">Build APIs You Won\u2019t Hate<\/a> - Everyone and their dog wants an API,\nso you should probably learn how to build them.<\/li>\n  <li><a href=\"http:\/\/shop.oreilly.com\/product\/0636920033868.do\">Modern PHP<\/a> - covers modern PHP features, best practices, testing, tuning, deployment and setting up a dev environment.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/buildingsecurephpapps\">Building Secure PHP Apps<\/a> - Learn the security basics that a senior\ndeveloper usually acquires over years of experience, all condensed down into one quick and easy handbook<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/mlaphp\">Modernizing Legacy Applications In PHP<\/a> - Get your code under control in a series of\nsmall, specific steps<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/securingphp-coreconcepts\">Securing PHP: Core Concepts<\/a> - A guide to some of the most common\nsecurity terms and provides some examples of them in every day PHP<\/li>\n  <li><a href=\"http:\/\/www.scalingphpbook.com\/\">Scaling PHP<\/a> - Stop playing sysadmin and get back to coding<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/signalingphp\">Signaling PHP<\/a> - PCNLT signals are a great help when writing PHP scripts that\nrun from the command line.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/minimumviabletests\">Minimum Viable Tests<\/a> - Long-time PHP testing evangelist Chris Hartjes goes over what he feels is the minimum you need to know to get started.<\/li>\n  <li><a href=\"https:\/\/leanpub.com\/ddd-in-php\">Domain-Driven Design in PHP<\/a> - See real examples written in PHP showcasing Domain-Driven Design Architectural Styles (Hexagonal Architecture, CQRS or Event Sourcing), Tactical Design Patterns, and Bounded Context Integration.<\/li>\n<\/ul><\/section><div class=\"back-to-top\"><p><a href=\"#top\" class=\"top\">Back to Top<\/a><\/p><\/div><section class=\"chapter\" id=\"user_groups\"><h2 id=\"user_groups_title\">PHP User Groups<\/h2><p>If you live in a larger city, odds are there\u2019s a PHP user group nearby. You can easily find your local PUG at\n<a href=\"https:\/\/php.ug\/\">PHP.ug<\/a>. Alternate sources might be <a href=\"http:\/\/www.meetup.com\/find\/\">Meetup.com<\/a> or a search for <code class=\"highlighter-rouge\">php user group near me<\/code>\nusing your favorite search engine (i.e. <a href=\"https:\/\/www.google.com\/search?q=php+user+group+near+me\">Google<\/a>). If you live in a smaller town, there may not be a\nlocal PUG; if that\u2019s the case, start one!<\/p><p>Special mention should be made of two global user groups: <a href=\"https:\/\/nomadphp.com\/\">NomadPHP<\/a> and <a href=\"http:\/\/phpwomen.org\/\">PHPWomen<\/a>. <a href=\"https:\/\/nomadphp.com\/\">NomadPHP<\/a> offers twice monthly\nonline user group meetings with presentations by some of the top speakers in the PHP community.\n<a href=\"http:\/\/phpwomen.org\/\">PHPWomen<\/a> is a non-exclusive user group originally targeted towards the women in the PHP world. Membership is open to\neveryone who supports a more diverse community. PHPWomen provide a network for support, mentorship and education, and\ngenerally promote the creating of a \u201cfemale friendly\u201d and professional atmosphere.<\/p><p><a href=\"https:\/\/wiki.php.net\/usergroups\">Read about User Groups on the PHP Wiki<\/a><\/p><\/section><section class=\"chapter\" id=\"conferences\"><h2 id=\"conferences_title\">PHP Conferences<\/h2><p>The PHP community also hosts larger regional and national conferences in many countries around the world. Well-known\nmembers of the PHP community usually speak at these larger events, so it\u2019s a great opportunity to learn directly from\nindustry leaders.<\/p><p><a href=\"https:\/\/secure.php.net\/conferences\/index.php\">Find a PHP Conference<\/a><\/p><\/section><section class=\"chapter\" id=\"elephpants\"><h2 id=\"elephpants_title\">ElePHPants<\/h2><p><a href=\"https:\/\/secure.php.net\/elephpant.php\">ElePHPant<\/a> is that beautiful mascot of the PHP project with elephant in their design. It was originally designed for the PHP project in 1998 by <a href=\"http:\/\/www.elroubio.net\/\">Vincent Pontier<\/a> - spiritual father of thousands of elePHPants around the world and 10 years later adorable plush elephant toy came to birth as well. Now elePHPants are present at many PHP conferences and with many PHP developers at their computers for fun and inspiration.<\/p><p><a href=\"https:\/\/7php.com\/elephpant\/\">Interview with Vincent Pontier<\/a><\/p><\/section>","created_at":"2018-11-07T16:22:29+0000","updated_at":"2019-04-14T16:36:50+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":108,"domain_name":"phptherightway.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":81,"title":"Let\u2019s all go back to Tumblr","url":"https:\/\/theoutline.com\/post\/5811\/why-tumblr-is-better-than-twitter-and-we-should-bring-it-back","content":"<p>Maybe you\u2019re familiar with the many reasons why people hate Twitter. It\u2019s a time sink, it enables the spread of fake news and bigotry, it encourages white hot anger predicated on context collapse, it\u2019s allowed mildly clever people to build careers on a foundation of bullshit, and so on. The memes are good, and the animal GIFs too (and the political solidarity and the righteous signal boosting and all of the other legitimately positive things, please don\u2019t tell me about them) but increasingly there is a sense \u2014 or rather, a range of senses \u2014 that much of this fussing around is not exactly healthy, for ourselves or society or anyone else.<\/p><p>This is all true to a degree, but lately I\u2019ve been thinking about Twitter\u2019s badness less in the context of the website itself, and more in the context of another website I used to frequent all the time and no longer do: Tumblr. Do you remember Tumblr? Of course, because the site is still objectively popular: According <a href=\"https:\/\/www.alexa.com\/siteinfo\/tumblr.com\">to Alexa\u2019s metrics<\/a>, it\u2019s nearly one of the 50 most popular websites in the world. But you are not there;  our president is not there; I am not there. Not anymore.<\/p><p>What made Tumblr so great? Again, I think about it in comparison with Twitter, and have pulled out a few reasons. On Twitter, the shortness of the posts forces the development of a few recognizable personality types: posters who almost exclusively tell jokes (a medium better suited to brevity) and posters who <em>don\u2019t<\/em> tell jokes, which means they\u2019re being almost exclusively serious, a mentality often conflated with brevity. But a great many serious things cannot be articulated in 140 or 280 characters, and so you get the <a href=\"https:\/\/theoutline.com\/post\/5542\/bring-back-irony\">most easily reducible<\/a> \u2014 and therefore least accurate \u2014 version of the feeling or idea. Either that, or you get lengthy Twitter threads, which are ponderous and boring.<\/p><p>But on Tumblr, people could go on for at long as they needed to, a valuable tool for posters who could actually justify it. (And I use the past tense here in the context of my own experience; if you\u2019re still doing this, bless you and yours.) Posts could be as short as necessary, but you could also find a historical deep dive, an interesting think piece (and not the kind derisively referred to as \u201ctakes\u201d), a photo essay, or simply just a nice blog about someone\u2019s life. The whole impetus behind following people on social media is, \u201cHey, I like this person\u2019s brain, and am open to spending more time with it.\u201d Tumblrs delivered the full, unrestrained range of someone\u2019s head \u2014 funny, serious, and everything else.<\/p><p>There were a lot of people talking at all different lengths about the subjects they wanted, which enabled another major appeal: the community. Here is how it went. You started off on Tumblr by following some people you knew of, whether personally or not, and slowly ventured deeper in the world as you lurked other blogs, or discovered new people reblogged by people you already liked. Pretty soon, you could be a small part of a <a href=\"https:\/\/theoutline.com\/post\/5065\/tumblr-asexual-history-bloggers-ace-historical?zd=1&amp;zi=odextmyw\">thriving ecosystem<\/a> of posters, discoursing and joking with each other.<\/p><p>Again, this is how the internet works; it\u2019s how blogging platforms before Tumblr worked; it\u2019s even how Twitter works. But insularity was encouraged because of Tumblr\u2019s boxed-in, user-unfriendly design. You couldn\u2019t easily search for text, and there was no easy way to track a post\u2019s movement through the blogosphere, so communities grew slowly. Compare, for example, how posts go viral on Twitter. On Twitter, <a href=\"https:\/\/twitter.com\/jeremypgordon\/status\/1028344679279013889\">I could tweet<\/a> something facile and snarky like \u201cAmazing to consider the Oscars have contorted themselves for a decade because of blowback for not nominating THE DARK KNIGHT, which is respectfully one of the dumbest movies to ever be taken ultra seriously by people,\u201d and immediately blow up.<\/p><p>There will be some thoughtful replies, but almost all of them will come from people I\u2019m not friendly with. I\u2019ll have no interest in debating them because I have no idea where they\u2019re coming from, ideologically or rhetorically. Everyone who sees the tweet will see the replies, too, experiencing all sides of many arguments all at once, with no curation or indication of which is worth following. There will be nothing familiar or contextual about the conversation, just a bunch of strangers arguing with each other.<\/p><blockquote class=\"pullquote pullquote--inline\" data-hash=\"116958866\"><div>Sometimes a overfamiliar bubble is more conducive to well-intentioned discourse, instead of a dozen randos screaming \u201cwhat\u2019s wrong with you, you Nazi and\/or anti-Nazi fuck.\u201d<\/div><\/blockquote><p>On Tumblr, if the same post went viral (or better yet, a well-reasoned essay on why <em>The Dark Knight<\/em> is a perfectly entertaining, visually arresting action blockbuster with the emotional depth of a puddle), all I\u2019d see would be generic \u201cthis post was reblogged\u201d notifications with a smidgen of attached text. The only detailed reactions I\u2019d read would be the ones on blogs I was already following. Meanwhile, if people had enough of an issue to talk about it, they\u2019d be reblogging it on their blogs, for the purpose of talking to <em>their<\/em> community \u2014 a much more achievable and worthwhile pursuit than trying to craft the argument that would appeal to everybody in the world all at once. I\u2019d never see that; my day would go on. We\u2019d be in our little worlds, feeling out the kinks of the potential arguments on our own schedule.<\/p><p>Did this create overfamiliar bubbles prone to social regulation by the standards of the more sensitive members? Yes, of course. But sometimes a overfamiliar bubble is more conducive to well-intentioned discourse, instead of a dozen randos screaming \u201cwhat\u2019s wrong with you, you Nazi and\/or anti-Nazi fuck.\u201d Sure, you\u2019d witness some truly embarrassing back-and-forths predicated on bad faith and lack of context. But it happened less, because the relative slowness of Tumblr meant people took longer to reply, which means that it took people longer to think about what they wanted to say. Even better, at some point one party would lose interest and the world would move on as usual.<\/p><p>The sum of this was thoughtful, personalized content (for lack of a better word) consumed at a slower, more natural pace where eventually everyone would lose interest in being mean to each other. Also, memes and animal GIFs and political solidarity and all of the good things. Wow! Doesn\u2019t that sound great? Wouldn\u2019t Twitter be better if it was more like that? It was like that, at least for a while; I participated in at least three distinct, naturally occurring social scenes encompassing parts of college and half of my twenties. I met a <em>lot<\/em> of people for the first time after meeting them on Tumblr, and almost all of them turned out to be actually cool. Whether or not I enjoyed someone\u2019s Tumblr after following them for a while was, in fact, an immensely accurate prediction for how I\u2019d enjoy them in real life.<\/p><p>It\u2019s nearly impossible for me to conceive of Tumblr being as big as it still is, because almost nobody I know uses it anymore. Everyone I knew aged out of it, lost interest in chronicling their personal lives or got jobs writing the kinds of blogs they used to write for free. As denizens of scenes peeled off for different pastures (Twitter, post-graduation life, parenthood), that sense of community became difficult to find in new forms. Tumblr, after all, is still user-unfriendly. There is no easy way to find a new set of blogs without doing a lot of manual clicking around, and the wide range of Tumblr types (such as those almost exclusively devoted to social justice, or fandoms) made it difficult to stumble upon the exact thing I wanted.<\/p><p>This is damning for a social network, and every time I\u2019ve tried to \u201cget back into\u201d Tumblr in the last year, it\u2019s like hanging around a ghost town,  and it just drives me back to the whole depressing Twitter cycle, where at least people are still talking, even if it\u2019s mostly in the form of yelling.<\/p><p>This tension of insularity is at least partly the company\u2019s fault, and a big part of why Tumblr <a href=\"https:\/\/www.emarketer.com\/Article\/Tumblr-Struggles-Grow-Its-User-Base\/1013596\">never meaningfully grew<\/a> or monetized after its initial boom period at the turn of the decade. It remains popular in the sense that people use it, but it\u2019s just\u2026 around, no longer accessibly special in a way that demands our attention. The userbase isn\u2019t as depleted <a href=\"https:\/\/theoutline.com\/post\/4137\/myspace-tom-mark-zuckerberg-facebook?zd=1&amp;zi=wyowceyg\">as Myspace<\/a>, but it remains much farther from the conversation about the future of digital media than ever seemed possible when it was first acquired by Yahoo, in a 2013 deal now universally <a href=\"https:\/\/digiday.com\/marketing\/tumblr-is-neglected-by-marketers\/\">regarded as a failure<\/a>. (Founder David Karp <a href=\"https:\/\/theoutline.com\/post\/2526\/david-karp-daddy-tumblr-resignation?zd=1&amp;zi=znbpgqxr\">has long since gone<\/a>, presumably to enjoy his hundreds of millions of dollars. Hey David, send me some.)<\/p><p>The ethos of Tumblr is more easily recognizable in a platform like Tinyletter, where people craft small batch blogs for a curated following, the downside being that they\u2019re entirely siloed in their own worlds with no chance of outside interaction. But considering how hectic and intrusive the modern internet can feel, this isolation feels like an asset, not a bug. Snail mail might never make a comeback, but the pleasures of one-on-one communication are evergreen.<\/p><p>Tumblr\u2019s irrelevance in the digital economy is a problem if you invested in the company, but not so much if you\u2019re a user who never drifted away. The platform remains full of the potential it once had, theoretically. So why not come back? Why don\u2019t we all go back? I\u2019ve tried, and I still haven\u2019t had much luck finding a new rhythm; if you have any good Tumblrs worth following, let me know. I\u2019d love to give it another shot.<\/p><p>Hey you! We want to know what you think about The Outline (and you can win some cool swag too). We know you love to answer questions, so <a href=\"https:\/\/theoutline.typeform.com\/to\/t0ykjM?source=post-footer-5811\" target=\"_blank\" rel=\"noopener\">take our 5 minute survey<\/a>.<\/p>","created_at":"2018-11-07T16:22:29+0000","updated_at":"2018-11-07T16:22:29+0000","published_at":"2018-08-14T15:30:00+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"theoutline.com","preview_picture":"https:\/\/outline-prod.imgix.net\/20180814-0Rr2xTUiNgxPiIcrpM4m?auto=format&q=60&w=1280&s=2f77caca29493789410121eba6d586dc","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":82,"title":"Functional Composition - Chris Ford - YouTube","url":"https:\/\/www.youtube.com\/watch?v=Mfsnlbd-4xQ","content":"<div id=\"body-container\"><div id=\"page-container\"><div id=\"page\" class=\"watch        video-Mfsnlbd-4xQ clearfix\"><div id=\"player\" class=\"content-alignment       watch-small\" role=\"complementary\"><div id=\"player-mole-container\"><div id=\"watch-queue-mole\" class=\"video-mole mole-collapsed hid\"><div id=\"watch-queue\" class=\"watch-playlist player-height\"><div class=\"hid\"><p>Watch Queue<\/p><p>Queue<\/p><p>__count__\/__total__<\/p><\/div><\/div><\/div><\/div><\/div><div id=\"content\" class=\"content-alignment\" role=\"main\"><div id=\"watch7-container\" class=\"\"><div id=\"watch7-main-container\"><div id=\"watch7-main\" class=\"clearfix\"><div id=\"watch7-content\" class=\"watch-main-col\" itemscope=\"itemscope\" itemid=\"\" itemtype=\"http:\/\/schema.org\/VideoObject\"><div id=\"watch-action-panels\" class=\"watch-action-panels yt-uix-button-panel hid yt-card yt-card-has-padding\"><div id=\"action-panel-transcript\" class=\"action-panel-content hid\"><div id=\"watch-actions-transcript-loading\"><div class=\"action-panel-loading\"><p class=\"yt-spinner\">\n        \n    \nLoading...\n    \n  <\/p><\/div><\/div><div id=\"watch-actions-transcript\" class=\"hid\"><h2 class=\"yt-card-title\">\nTranscript\n    <\/h2><div id=\"watch-transcript-container\" class=\"yt-scrollbar\"><p>The interactive transcript could not be loaded.<\/p><\/div><\/div><\/div><div id=\"action-panel-report\" class=\"action-panel-content hid\" data-auth-required=\"true\"><div class=\"action-panel-loading\"><p class=\"yt-spinner\">\n        \n    \nLoading...\n    \n  <\/p><\/div><\/div><div id=\"action-panel-rental-required\" class=\"action-panel-content hid\"><p><strong>Rating is available when the video has been rented.<\/strong><\/p><\/div><div id=\"action-panel-error\" class=\"action-panel-content hid\"><p>This feature is not available right now. Please try again later.<\/p><\/div><\/div><div id=\"action-panel-details\" class=\"action-panel-content yt-uix-expander yt-card yt-card-has-padding yt-uix-expander-collapsed\"><div id=\"watch-description\" class=\"yt-uix-button-panel\"><div id=\"watch-description-content\"><div id=\"watch-description-clip\"><div id=\"watch-description-text\" class=\"\"><p id=\"eow-description\" class=\"\">Music theory is one of the most naturally elegant and functional domains. It's a perfect fit for Clojure, which is why the Overtone audio environment makes hacking sound so much fun.<\/p><p>Chris will start with the basic building block of sound, the sine wave, and will gradually accumulate abstractions culminating in a canon by Johann Sebastian Bach.<\/p><p>By the end you should agree that anyone who's a developer is a musician too - even if they don't know it yet.<\/p><\/div><\/div><\/div><\/div><\/div><div id=\"watch-discussion\" class=\"branded-page-box yt-card\"><div class=\"action-panel-loading\"><p class=\"yt-spinner\">\n        \n    \nLoading...\n    \n  <\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>  \n      <div class=\"yt-dialog hid\" id=\"feed-privacy-lb\"><div class=\"yt-dialog-base\"><div class=\"yt-dialog-fg\" role=\"dialog\"><div class=\"yt-dialog-fg-content\"><div class=\"yt-dialog-loading\"><div class=\"yt-dialog-waiting-content\"><p class=\"yt-spinner\">\n        \n    \nLoading...\n    \n  <\/p><\/div><\/div><div class=\"yt-dialog-working\"><div class=\"yt-dialog-working-bubble\"><div class=\"yt-dialog-waiting-content\"><p class=\"yt-spinner\">\n        \n    \n        Working...\n    \n  <\/p><\/div><\/div><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:22:29+0000","updated_at":"2018-11-07T16:22:29+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/Mfsnlbd-4xQ\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":83,"title":"What Did Ada Lovelace's Program Actually Do?","url":"https:\/\/twobithistory.org\/2018\/08\/18\/ada-lovelace-note-g.html","content":"<p>The story of Microsoft\u2019s founding is one of the most famous episodes in\ncomputing history. In 1975, Paul Allen flew out to Albuquerque to demonstrate\nthe BASIC interpreter that he and Bill Gates had written for the Altair\nmicrocomputer. Because neither of them had a working Altair, Allen and Gates\ntested their interpreter using an emulator that they wrote and ran on Harvard\u2019s\ncomputer system. The emulator was based on nothing more than the published\nspecifications for the Intel 8080 processor. When Allen finally ran their\ninterpreter on a real Altair\u2014in front of the person he and Gates hoped would\nbuy their software\u2014he had no idea if it would work. But it did. The next month,\nAllen and Gates officially founded their new company.<\/p><p>Over a century before Allen and Gates wrote their BASIC interpreter, Ada\nLovelace wrote and published a computer program. She, too, wrote a program for\na computer that had only been described to her. But her program, unlike the\nMicrosoft BASIC interpreter, was never run, because the computer she was\ntargeting was never built.<\/p><p>Lovelace\u2019s program is often called the world\u2019s first computer program. Not\neveryone agrees that it should be called that. Lovelace\u2019s legacy, it turns out,\nis one of computing history\u2019s most hotly debated subjects. Walter Isaacson has\nwritten that the dispute about the extent and merit of her contributions\nconstitutes a \u201cminor academic specialty.\u201d<sup><a href=\"#fn:1\" class=\"footnote\">1<\/a><\/sup> Inevitably, the fact that\nLovelace was a woman has made this dispute a charged one. Historians have cited\nall kinds of primary evidence to argue that the credit given to Lovelace is\neither appropriate or undeserved. But they seem to spend less time explaining\nthe technical details of her published writing, which is unfortunate, because\nthe technical details are the most fascinating part of the story. Who wouldn\u2019t\nwant to know exactly how a program written in 1843 was supposed to work?<\/p><p>In fairness, Lovelace\u2019s program is not easy to explain to the layperson without\nsome hand-waving. It\u2019s the intricacies of her program, though, that make it so\nremarkable. Whether or not she ought to be known as \u201cthe first programmer,\u201d her\nprogram was specified with a degree of rigor that far surpassed anything that\ncame before. She thought carefully about how operations could be organized into\ngroups that could be repeated, thereby inventing the loop. She realized how\nimportant it was to track the state of variables as they changed, introducing a\nnotation to illustrate those changes. As a programmer myself, I\u2019m startled to\nsee how much of what Lovelace was doing resembles the experience of writing\nsoftware today.<\/p><p>So let\u2019s take a closer look at Lovelace\u2019s program. She designed it to calculate\nthe Bernoulli numbers. To understand what those are, we have to go back a\ncouple millennia to the genesis of one of mathematics\u2019 oldest problems.<\/p><h2 id=\"sums-of-powers\">Sums of Powers<\/h2><p>The Pythagoreans lived on the shores of the Mediterranean and worshiped\nnumbers. One of their pastimes was making triangles out of pebbles.<\/p><p><img src=\"https:\/\/twobithistory.org\/images\/triangular_numbers1.png\" alt=\"\" \/><\/p><p>One pebble followed by a row of two pebbles makes a triangle\ncontaining three pebbles. Add another row of three pebbles and you get a\ntriangle containing six pebbles. You can continue like this, each time adding a\nrow with one more pebble in it than the previous row. A triangle with six rows\ncontains 21 pebbles. But how many pebbles does a triangle with 423 rows\ncontain?<\/p><p>What the Pythagoreans were looking for was a way to calculate the following\nwithout doing all the addition:<\/p><p>They eventually realized that, if you place two triangles of the same size up\nagainst each other so that they form a rectangle, you can find the area of the\nrectangle and divide by two to get the number of pebbles in each of the\ntriangles:<\/p><p><img src=\"https:\/\/twobithistory.org\/images\/triangular_numbers2.png\" alt=\"\" \/><\/p><p>Archimedes later explored a similar problem. He was interested in the following\nseries:<\/p><p>You might visualize this series by imagining a stack of progressively larger\nsquares (made out of tiny cubes), one on top of the other, forming a pyramid.\nArchimedes wanted to know if there was an easy way to tell how many cubes would\nbe needed to construct a pyramid with, say, 423 levels. He recorded a solution\nthat also permits a geometrical interpretation.<sup><a href=\"#fn:2\" class=\"footnote\">2<\/a><\/sup><\/p><p>Three pyramids can be fit together to form a rectangular prism with a tiny,\none-cube-high extrusion at one end. That little extrusion happens to be a\ntriangle that obeys the same rules that the Pythagoreans used to make their\npebble triangles. (<a href=\"https:\/\/www.youtube.com\/watch?v=aXbT37IlyZQ\">This video<\/a>\nmight be a more helpful explanation of what I mean.) So the volume of the whole\nshape is given by the following equation:<\/p><p>By substituting the Pythagorean equation for the sum of the first <em>n<\/em>\nintegers and doing some algebra, you get this:<\/p><p>In 499, the Indian mathematician and astronomer, Aryabhata, published a work\nknown as the <em>Aryabhatiya<\/em>, which included a formula for calculating the sum of\ncubes:<\/p><p>A formula for the sum of the first <em>n<\/em> positive integers raised to the fourth\npower wasn\u2019t published for another 500 years.<sup><a href=\"#fn:3\" class=\"footnote\">3<\/a><\/sup><\/p><p>You might be wondering at this point if there is a general method for finding\nthe sum of the first <em>n<\/em> integers raised to the <em>kth<\/em> power.\nMathematicians were wondering too. Johann Faulhaber, a German mathematician and\nslightly kooky numerologist, was able to calculate formulas for sums of\nintegers up to the 17th power, which he published in 1631. But this may have\ntaken him years and he did not state a general solution.<sup><a href=\"#fn:4\" class=\"footnote\">4<\/a><\/sup> Blaise Pascal\nfinally outlined a general method in 1665, though it depended on first knowing\nhow to calculate the sum of integers raised to every lesser power.<sup><a href=\"#fn:5\" class=\"footnote\">5<\/a><\/sup> To\ncalculate the sum of the first <em>n<\/em> positive integers raised to the sixth power,\nfor example, you would first have to know how to calculate the sum of the first\n<em>n<\/em> positive integers raised to the fifth power.<\/p><p>A more practical general solution was stated in the posthumously published work\nof Swiss mathematician Jakob Bernoulli, who died in 1705. Bernoulli began by\nderiving the formulas for calculating the sums of the first <em>n<\/em> positive\nintegers to the first, second, and third powers.<sup><a href=\"#fn:6\" class=\"footnote\">6<\/a><\/sup> These he gave in\npolynomial form, so they looked like the below:<\/p><p>Using Pascal\u2019s Triangle, Bernoulli realized that these polynomials followed a\npredictable pattern. Essentially, Bernoulli broke the coefficients of each\nterm down into two factors, one of which he could determine using Pascal\u2019s\nTriangle and the other which he could derive from the interesting property that\nall the coefficients in the polynomial seemed to always add to one. Figuring\nout the exponent that should be attached to each term was no problem,\nbecause that also followed a predictable pattern. The factor of each\ncoefficient that had to be calculated using the sums-to-one rule formed a\nsequence that became known as the Bernoulli numbers.<\/p><p>Bernoulli\u2019s discovery did not mean that it was now trivial to calculate the sum\nof the first positive <em>n<\/em> integers to any given power. In order to calculate\nthe sum of the first positive <em>n<\/em> integers raised to the <em>kth<\/em> power,\nyou would need to know every Bernoulli number up to the <em>kth<\/em> Bernoulli\nnumber. Each Bernoulli number could only be calculated if the previous\nBernoulli numbers were known. But calculating a long series of Bernoulli\nnumbers was significantly easier than deriving each sum of powers formula in\nturn, so Bernoulli\u2019s discovery was a big advance for mathematics.<\/p><h2 id=\"babbage\">Babbage<\/h2><p>Charles Babbage was born in 1791, nearly a century after Bernoulli died. I\u2019ve\nalways had some vague idea that Babbage designed but did not build a mechanical\ncomputer. But I\u2019ve never entirely understood how that computer was supposed to\nwork. The basic ideas, as it happens, are not that difficult to grasp, which is\ngood news. Lovelace\u2019s program was designed to run on one of Babbage\u2019s machines,\nso we need to take another quick detour here to talk about how those machines\nworked.<\/p><p>Babbage designed two separate mechanical computing machines. His first machine\nwas called the Difference Engine. Before the invention of the pocket\ncalculator, people relied on logarithmic tables to calculate the product of\nlarge numbers. (There is a good <a href=\"https:\/\/youtu.be\/VRzH4xB0GdM\">Numberphile\nvideo<\/a> on how this was done.) Large logarithmic\ntables are not difficult to create, at least conceptually, but the sheer number\nof calculations that need to be done in order to create them meant that in\nBabbage\u2019s time they often contained errors. Babbage, frustrated by this, sought\nto create a machine that could tabulate logarithms mechanically and therefore\nwithout error.<sup><a href=\"#fn:7\" class=\"footnote\">7<\/a><\/sup><\/p><p>The Difference Engine was not a computer, because all it did was add and\nsubtract. It took advantage of a method devised by the French mathematician\nGaspard de Prony that broke the process of tabulating logarithms down into\nsmall steps.<sup><a href=\"#fn:8\" class=\"footnote\">8<\/a><\/sup> These small steps involved only addition and subtraction,\nmeaning that a small army of people without any special mathematical aptitude\nor training could be employed to produce a table. De Prony\u2019s method, known as\nthe method of divided differences, could be used to tabulate any polynomial.\nPolynomials, in turn, could be used to approximate logarithmic and\ntrigonometric functions.<\/p><p>To get a sense of how this process worked, consider the following simple\npolynomial function:<\/p><p>The method of divided differences involves finding the difference between each\nsuccessive value of <em>y<\/em> for different values of <em>x<\/em>. The differences\nbetween these differences are then found, and possibly the differences between\nthose next differences themselves, until a constant difference appears. These\ndifferences can then be used to get the next value of the polynomial simply by\nadding.<\/p><p>Because the above polynomial is only a second-degree polynomial, we are able to\nfind the constant difference after only two columns of differences:<\/p><p><\/p><table><thead><tr><th>x<\/th>\n        <th>y<\/th>\n        <th>Diff 1<\/th>\n        <th>Diff 2<\/th>\n      <\/tr><\/thead><tbody><tr><td>1<\/td>\n        <td>2<\/td>\n        <td>\u00a0<\/td>\n        <td>\u00a0<\/td>\n      <\/tr><tr><td>2<\/td>\n        <td>5<\/td>\n        <td>3<\/td>\n        <td>\u00a0<\/td>\n      <\/tr><tr><td>3<\/td>\n        <td>10<\/td>\n        <td>5<\/td>\n        <td>2<\/td>\n      <\/tr><tr><td>4<\/td>\n        <td>17<\/td>\n        <td>7<\/td>\n        <td>2<\/td>\n      <\/tr><tr><td>5<\/td>\n        <td>?<\/td>\n        <td>?<\/td>\n        <td>2<\/td>\n      <\/tr><tr><td>\u2026<\/td>\n        <td>\u2026<\/td>\n        <td>\u2026<\/td>\n        <td>\u2026<\/td>\n      <\/tr><\/tbody><\/table><p>Now, since we know that the constant difference is 2, we can find the value of\n<em>y<\/em> when <em>x<\/em> is 5 through addition only. If we add 2 to 7, the last entry in\nthe \u201cDiff 1\u201d column, we get 9. If we add 9 to 17, the last entry in the <em>y<\/em>\ncolumn, we get 26, our answer.<\/p><p>Babbage\u2019s Difference Engine had, for each difference column in a table like the\none above, a physical column of gears. Each gear was a decimal digit and one\nwhole column was a decimal number. The Difference Engine had eight columns of\ngears, so it could tabulate a polynomial up to the seventh degree. The columns\nwere initially set with values matching an early row in the difference table,\nworked out ahead of time. A human operator would then turn a crank shaft,\ncausing the constant difference to propagate through the machine as the value\nstored on each column was added to the next.<\/p><p>Babbage was able to build a small section of the Difference Engine and use it\nto demonstrate his ideas at parties.<sup><a href=\"#fn:9\" class=\"footnote\">9<\/a><\/sup> But even after spending an amount of\npublic money equal to the cost of two large warships, he never built the entire\nmachine.<sup><a href=\"#fn:10\" class=\"footnote\">10<\/a><\/sup> Babbage could not find anyone in the early 1800s that could make\nthe number of gears he needed with sufficient accuracy. A working Difference\nEngine would not be built until the 1990s, after the advent of precision\nmachining. There is <a href=\"https:\/\/www.youtube.com\/watch?v=BlbQsKpq3Ak\">a great video on\nYouTube<\/a> demonstrating a working\nDifference Engine on loan to the Computer History Museum in Mountain View,\nwhich is worth watching even just to listen to the marvelous sounds the machine\nmakes while it runs.<\/p><p>Babbage eventually lost interest in the Difference Engine when he realized that\na much more powerful and flexible machine could be built. His Analytical Engine\nwas the machine that we know today as Babbage\u2019s mechanical computer. The\nAnalytical Engine was based on the same columns of gears used in the Difference\nEngine, but whereas the Difference Engine only had eight columns, the\nAnalytical Engine was supposed to have many hundreds more. The Analytical\nEngine could be programmed using punched cards like a Jacquard Loom and could\nmultiply and divide as well as add and subtract. In order to perform one of\nthese operations, a section of the machine called the \u201cmill\u201d would rearrange\nitself into the appropriate configuration, read the operands off of other\ncolumns used for data storage, and then write the result back to another\ncolumn.<\/p><p>Babbage called his new machine the Analytical Engine because it was powerful\nenough to do something resembling mathematical analysis. The Difference Engine\ncould tabulate a polynomial, but the Analytical Engine would be able to\ncalculate, for example, the coefficients of the polynomial expansion of another\nexpression. It was an amazing machine, but the British government wisely\ndeclined to fund its construction. So Babbage went abroad to Italy to try to\ndrum up support for his idea.<\/p><h2 id=\"notes-by-the-translator\">Notes by The Translator<\/h2><p>In Turin, Babbage met Italian engineer and future prime minister Luigi\nMenabrea. He persuaded Menabrea to write an outline of what the Analytical\nEngine could accomplish. In 1842, Menabrea published a paper on the topic in\nFrench. The following year, Lovelace published a translation of Menabrea\u2019s\npaper into English.<\/p><p>Lovelace, then known as Ada Byron, first met Babbage at a party in 1833, when\nshe was 17 and he was 41.<sup><a href=\"#fn:11\" class=\"footnote\">11<\/a><\/sup> Lovelace was fascinated with Babbage\u2019s\nDifference Engine. She could also understand how it worked, because she had\nbeen extensively tutored in mathematics throughout her childhood. Her mother,\nAnnabella Milbanke, had decided that a solid grounding in mathematics would\nward off the wild, romantic sensibility that possessed Lovelace\u2019s father, Lord\nByron, the famous poet. After meeting in 1833, Lovelace and Babbage remained a\npart of the same social circle and wrote to each other frequently.<\/p><p>Ada Byron married William King in 1835. King later became the Earl of Lovelace,\nmaking Ada the Countess of Lovelace. Even after having three children, she\ncontinued her education in mathematics, employing Augustus de Morgan, who\ndiscovered De Morgan\u2019s laws, as her tutor.<sup><a href=\"#fn:12\" class=\"footnote\">12<\/a><\/sup> Lovelace saw the potential of\nBabbage\u2019s Analytical Machine immediately and was eager to work with him to\npromote the idea. A friend suggested that she translate Menabrea\u2019s paper for an\nEnglish audience.<sup><a href=\"#fn:13\" class=\"footnote\">13<\/a><\/sup><\/p><p>Menabrea\u2019s paper gave a brief overview of how the Difference Engine worked,\nthen showed how the Analytical Engine would be a far superior machine. The\nAnalytical Engine would be so powerful that it could \u201cform the product of two\nnumbers, each containing twenty figures, in <em>three minutes<\/em>\u201d<sup><a href=\"#fn:14\" class=\"footnote\">14<\/a><\/sup> (emphasis in\nthe original). Menabrea gave further examples of the machine\u2019s capabilities,\ndemonstrating how it could solve a simple system of linear equations and\nexpand the product of two binomial expressions. In both cases, Menabrea\nprovided what Lovelace called \u201cdiagrams of development,\u201d which listed the\nsequence of operations that would need to be performed to calculate the correct\nanswer.<sup><a href=\"#fn:15\" class=\"footnote\">15<\/a><\/sup> These were programs in the same sense that Lovelace\u2019s own program\nwas a program and they were originally published the year before. But as we\nwill see, Menabrea\u2019s programs were only simple examples of what was possible.\nAll of them were trivial in the sense that they did not require any kind of\nbranching or looping.<\/p><p>Lovelace appended a series of notes to her translation of Menabrea\u2019s paper that\ntogether ran much longer than the original work. It was here that she made her\nmajor contributions to computing. In Note A, which Lovelace attached to\nMenabrea\u2019s initial description of the Analytical Engine, Lovelace explained at\nsome length and often in lyrical language the promise of a machine that could\nperform arbitrary mathematical operations. She foresaw that a machine like the\nAnalytical Engine wasn\u2019t just limited to numbers and could in fact act on\nany objects \u201cwhose mutual fundamental relations could be expressed by those of\nthe abstract science of operations, and which should be also susceptible of\nadaptations to the action of the operating notation and mechanism of the\nengine.\u201d<sup><a href=\"#fn:16\" class=\"footnote\">16<\/a><\/sup> She added that the machine might one day, for example, compose\nmusic. This insight was all the more remarkable given that Menabrea saw the\nAnalytical Engine primarily as a tool for automating \u201clong and arid\ncomputation,\u201d which would free up the intellectual capacities of brilliant\nscientists for more advanced thinking.<sup><a href=\"#fn:17\" class=\"footnote\">17<\/a><\/sup> The miraculous foresight that\nLovelace demonstrated in Note A is one major reason that she is celebrated\ntoday.<\/p><p>The other famous note is Note G. Lovelace begins Note G by arguing that,\ndespite its impressive powers, the Analytical Machine cannot really be said to\n\u201cthink.\u201d This part of Note G is what Alan Turing would later refer to as \u201cLady\nLovelace\u2019s Objection.\u201d Nevertheless, Lovelace continues, the machine can do\nextraordinary things. To illustrate its ability to handle even more complex\nproblems, Lovelace provides her program calculating the Bernoulli numbers.<\/p><p>The full program, in the expanded \u201cdiagram of development\u201d format that Lovelace\nexplains in Note D, can be seen\n<a href=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/c\/cf\/Diagram_for_the_computation_of_Bernoulli_numbers.jpg\">here<\/a>.\nThe program is essentially a list of operations, specified using the usual\nmathematical symbols. It doesn\u2019t appear that Babbage or Lovelace got as far as\ndeveloping anything like a set of op codes for the Analytical Engine.<\/p><p>Though Lovelace was describing a method for computing the entire sequence of\nBernoulli numbers up to some limit, the program she provided only illustrated\none step of that process. Her program calculated a number that she called B7,\nwhich modern mathematicians know as the eighth Bernoulli number. Her program\nthus sought to solve the following equation:<\/p><p>In the above, each term represents a coefficient in the polynomial formula for\nthe sum of integers to a particular power. Here that power is eight, since the\neighth Bernoulli number first appears in the formula for the sum of positive\nintegers to the eighth power. The B and A numbers represent the two kinds of\nfactors that Bernoulli discovered. B1 through B7 are all different Bernoulli\nnumbers, indexed according to Lovelace\u2019s indexing. A0 through A5 represent the\nfactors of the coefficients that Bernoulli could calculate using Pascal\u2019s\nTriangle. The values of A0, A1, A3, and A5 appear below. Here <em>n<\/em> represents\nthe index of the Bernoulli number in the sequence of odd-numbered Bernoulli\nnumbers starting with the first. Lovelace\u2019s program used <em>n<\/em> = 4.<\/p><p>I\u2019ve created a\n<a href=\"https:\/\/gist.github.com\/sinclairtarget\/ad18ac65d277e453da5f479d6ccfc20e\">translation<\/a>\nof Lovelace\u2019s program into C, which may be easier to follow. Lovelace\u2019s program\nfirst calculates A0 and the product B1A1. It then enters a loop that repeats\ntwice to calculate B3A3 and B5A5, since those are formed according to an\nidentical pattern. After each product is calculated, it is added with all the\nprevious products, so that by the end of the program the full sum has been\nobtained.<\/p><p>Obviously the C translation is not an exact recreation of Lovelace\u2019s program.\nIt declares variables on the stack, for example, whereas Lovelace\u2019s variables\nwere more like registers. But it makes obvious the parts of Lovelace\u2019s program\nthat were so prescient. The C program contains two <code class=\"highlighter-rouge\">while<\/code> loops, one nested\ninside the other. Lovelace\u2019s program did not have <code class=\"highlighter-rouge\">while<\/code> loops exactly, but\nshe made groups of operations and in the text of her note specified when they\nshould repeat. The variable <code class=\"highlighter-rouge\">v10<\/code>, in the original program and in the C\ntranslation, functions as a counter variable that decrements with each loop, a\nconstruct any programmer would be familiar with. In fact, aside from the\nprofusion of variables with unhelpful names, the C translation of Lovelace\u2019s\nprogram doesn\u2019t look that alien at all.<\/p><p>The other thing worth mentioning quickly is that translating Lovelace\u2019s program\ninto C was not that difficult, thanks to the detail present in her diagram.\nUnlike Menabrea\u2019s tables, her table includes a column labeled \u201cIndication of\nchange in the value on any Variable,\u201d which makes it much easier to follow the\nmutation of state throughout the program. She adds a superscript index here to\neach variable to indicate the successive values they hold. A superscript of\ntwo, for example, means that the value being used here is the second value that\nhas been assigned to the variable since the beginning of the program.<\/p><h2 id=\"the-first-programmer\">The First Programmer?<\/h2><p>After I had translated Lovelace\u2019s program into C, I was able to run it on my\nown computer. To my frustration, I kept getting the wrong result. After some\ndebugging, I finally realized that the problem wasn\u2019t the code that I had\nwritten. The bug was in the original!<\/p><p>In her \u201cdiagram of development,\u201d Lovelace gives the fourth operation as\n<code class=\"highlighter-rouge\">v5 \/ v4<\/code>. But the correct ordering here is <code class=\"highlighter-rouge\">v4 \/ v5<\/code>. This may well have been\na typesetting error and not an error in the program that Lovelace devised. All\nthe same, this must be the oldest bug in computing. I marveled that, for ten\nminutes or so, unknowingly, I had wrestled with this first ever bug.<\/p><p>Jim Randall, another blogger that has <a href=\"https:\/\/enigmaticcode.wordpress.com\/tag\/bernoulli-numbers\/\">translated Lovelace\u2019s program into\nPython<\/a>, has noted\nthis division bug and two other issues. What does it say about Ada Lovelace\nthat her published program contains minor bugs? Perhaps it shows that she\nwas attempting to write not just a demonstration but a real program. After all,\ncan you really be writing anything more than toy programs if you aren\u2019t also\nwriting lots of bugs?<\/p><p>One Wikipedia article calls Lovelace the first to publish a \u201ccomplex\nprogram.\u201d<sup><a href=\"#fn:18\" class=\"footnote\">18<\/a><\/sup> Maybe that\u2019s the right way to think about Lovelace\u2019\naccomplishment. Menabrea published \u201cdiagrams of development\u201d in his paper a\nyear before Lovelace published her translation. Babbage also wrote more than\ntwenty programs that he never published.<sup><a href=\"#fn:19\" class=\"footnote\">19<\/a><\/sup> So it\u2019s not quite accurate to say\nthat Lovelace wrote or published the first program, though there\u2019s always room\nto quibble about what exactly constitutes a \u201cprogram.\u201d Even so, Lovelace\u2019s\nprogram was miles ahead of anything else that had been published before. The\nlongest program that Menabrea presented was 11 operations long and contained no\nloops or branches; Lovelace\u2019s program contains 25 operations and a nested loop\n(and thus branching). Menabrea wrote the following toward the end of his paper:<\/p><blockquote>\n  <p>When once the engine shall have been constructed, the difficulty will be\nreduced to the making of the cards; but as these are merely the translation\nof algebraic formulae, it will, by means of some simple notation, be easy to\nconsign the execution of them to a workman.<sup><a href=\"#fn:20\" class=\"footnote\">20<\/a><\/sup><\/p>\n<\/blockquote><p>Neither Babbage nor Menabrea were especially interested in applying the\nAnalytical Engine to problems beyond the immediate mathematical challenges that\nfirst drove Babbage to construct calculating machines. Lovelace saw that the\nAnalytical Engine was capable of much more than Babbage or Menabrea could\nimagine. Lovelace also grasped that \u201cthe making of the cards\u201d would not be a\nmere afterthought and that it could be done well or done poorly. This is hard\nto appreciate without understanding her program from Note G and seeing for\noneself the care she put into designing it. But having done that, you might\nagree that Lovelace, even if she was not the very first programmer, was the\nfirst programmer to deserve the title.<\/p><p><em>\nIf you enjoyed this post, more like it come out every two weeks! Follow\n<a href=\"https:\/\/twitter.com\/TwoBitHistory\">\n  @TwoBitHistory\n<\/a> on Twitter or subscribe to the\n<a href=\"https:\/\/twobithistory.org\/feed.xml\">\n  RSS feed\n<\/a>\nto make sure you know when a new post is out.\n<\/em><\/p><p><em>Previously on TwoBitHistory\u2026<\/em><\/p><div class=\"footnotes\"><ol><li>\n      <p>Walter Isaacson, The Innovators (New York: Simon &amp; Schuster Paperbacks, 2015), 25.\u00a0<a href=\"#fnref:1\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Janet Beery, \u201cSums of Powers of Positive Integers: Archimedes,\u201d Mathematical Association of America, accessed August 18, 2018, <a href=\"https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-archimedes-287-212-bce-greece-italy\">https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-archimedes-287-212-bce-greece-italy<\/a>.\u00a0<a href=\"#fnref:2\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Janet Beery, \u201cSums of Powers of Positive Integers: Abu Ali Al-Hasan Ibn Al-Hasan Ibn Al-Haytham,\u201d Mathematical Association of America, accessed August 18, 2018, <a href=\"https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-abu-ali-al-hasan-ibn-al-hasan-ibn-al-haytham-965-1039-egypt\">https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-abu-ali-al-hasan-ibn-al-hasan-ibn-al-haytham-965-1039-egypt<\/a>.\u00a0<a href=\"#fnref:3\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Janet Beery, \u201cSums of Powers of Positive Integers: Conclusion,\u201d Mathematical Association of America, accessed August 18, 2018, <a href=\"https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-conclusion\">https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-conclusion<\/a>.\u00a0<a href=\"#fnref:4\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Janet Beery, \u201cSums of Powers of Positive Integers: Blaise Pascal,\u201d Mathematical Association of America, accessed August 18, 2018, <a href=\"https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-blaise-pascal-1623-1662-france\">https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-blaise-pascal-1623-1662-france<\/a>.\u00a0<a href=\"#fnref:5\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Janet Beery, \u201cSums of Powers of Positive Integers: Jakob Bernoulli,\u201d Mathematical Association of America, accessed August 18, 2018, <a href=\"https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-jakob-bernoulli-1654-1705-switzerland\">https:\/\/www.maa.org\/press\/periodicals\/convergence\/sums-of-powers-of-positive-integers-jakob-bernoulli-1654-1705-switzerland<\/a>.\u00a0<a href=\"#fnref:6\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 19.\u00a0<a href=\"#fnref:7\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 20.\u00a0<a href=\"#fnref:8\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Robert Scoble, \u201cA Demo of Charles Babbage\u2019s Difference Engine,\u201d YouTube, June 17, 2010, accessed August 18, 2018, <a href=\"https:\/\/www.youtube.com\/watch?v=BlbQsKpq3Ak&amp;feature=youtu.be&amp;t=7m37s\">https:\/\/www.youtube.com\/watch?v=BlbQsKpq3Ak&amp;feature=youtu.be&amp;t=7m37s<\/a>.\u00a0<a href=\"#fnref:9\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 22.\u00a0<a href=\"#fnref:10\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 8.\u00a0<a href=\"#fnref:11\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 17.\u00a0<a href=\"#fnref:12\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 25.\u00a0<a href=\"#fnref:13\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Luigi Menabrea, \u201cSketch of the Analytical Engine,\u201d Scientific Memoirs 3 (1843): 686, accessed August 18, 2018, <a href=\"https:\/\/books.google.com\/books\/about\/Scientific_Memoirs_Selected_from_the_Tra.html?id=qsY-AAAAYAAJ\">https:\/\/books.google.com\/books\/about\/Scientific_Memoirs_Selected_from_the_Tra.html?id=qsY-AAAAYAAJ<\/a>.\u00a0<a href=\"#fnref:14\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Ada Lovelace, \u201cTranslator\u2019s Notes to M. Menabrea\u2019s Memoir,\u201d Scientific Memoirs 3 (1843): 712, accessed August 18, 2018, <a href=\"https:\/\/books.google.com\/books\/about\/Scientific_Memoirs_Selected_from_the_Tra.html?id=qsY-AAAAYAAJ\">https:\/\/books.google.com\/books\/about\/Scientific_Memoirs_Selected_from_the_Tra.html?id=qsY-AAAAYAAJ<\/a>.\u00a0<a href=\"#fnref:15\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Lovelace, 694.\u00a0<a href=\"#fnref:16\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Menabrea, 690.\u00a0<a href=\"#fnref:17\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>\u201cBernoulli number,\u201d Wikipedia, accessed August 18, 2018, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bernoulli_number\">https:\/\/en.wikipedia.org\/wiki\/Bernoulli_number<\/a>.\u00a0<a href=\"#fnref:18\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Isaacson, 29.\u00a0<a href=\"#fnref:19\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li>\n      <p>Menabrea, 689.\u00a0<a href=\"#fnref:20\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>","created_at":"2018-11-07T16:22:29+0000","updated_at":"2018-11-07T16:22:29+0000","annotations":[],"mimetype":"text\/html","reading_time":21,"domain_name":"twobithistory.org","preview_picture":"https:\/\/twobithistory.org\/images\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":84,"title":"Salary Negotiation: Make More Money, Be More Valued","url":"https:\/\/www.kalzumeus.com\/2012\/01\/23\/salary-negotiation\/","content":"<p>[<strong>Editor\u2019s note<\/strong>: At nearly 7,000 words, you probably don\u2019t want to try reading this on an iDevice. \u00a0Bookmark it and come back later.]<\/p><p>Imagine something a wee bit outside your comfort zone. \u00a0Nothing scandalous: just something you don\u2019t do often, don\u2019t particularly enjoy, and slightly more challenging than \u201ctotally trivial.\u201d \u00a0Maybe reciting poetry while simultaneously standing on one foot.<\/p><p>If I told you I would pay you a hundred thousand dollars if you did five minutes of poetry recital while standing on one foot, would you do it? \u00a0It\u2019s an absurd image, but play it straight. \u00a0There is no hidden gotcha here. \u00a0You won\u2019t be videotaped. \u00a0Your friends will never see you make a fool of yourself. \u00a0The revolution will not be YouTubed. \u00a0The offer is exactly as simple as you think it is: poetry, foot, $100,000.<\/p><p>Would you read poetry for me?<\/p><p><strong>Of course you would.<\/strong> \u00a0You\u2019d be screamingly stupid not to. \u00a0In fact, not only would you read poetry, you\u2019d probably take a poetry class to make sure you did it right, or go to the gym to verify \u201cYep, sure enough, I can stand on one foot. \u00a0Phew. \u00a0Pass me the Shakespeare.\u201d \u00a0If you couldn\u2019t stand on one foot, you\u2019d fix that, because you know that is much easier than other things you routinely accomplish and you suddenly have a hundred thousand wonderful reasons to learn it, too.<\/p><p>What if you were talking about this at dinner with your friends, and one of them said \u201cOh, no, I\u2019d never do that. \u00a0I just don\u2019t do poetry. \u00a0I\u2019m an engineer. \u00a0And besides, my father told me that people who stand on one foot look silly. \u00a0And what do I need $100,000 for anyhow?\u201d \u00a0You would not clap them on the back and say \u201cDamn straight! \u00a0Man, poets, always trying to tempt virtuous engineers into their weird poetry-spouting flamingo-standing ways.\u201d \u00a0You\u2019d say \u201cDude, it\u2019s five minutes. \u00a0Heck, I\u2019ll help you practice.\u201d<\/p><p>This is pretty much how I feel every time I talk to my engineering friends about salary negotiation. \u00a0We overwhelmingly suck at it. \u00a0We have turned sucking at it into a <em>perverse badge of virtue<\/em>. \u00a0We make no affirmative efforts to un-suck ourselves and, to the extent we read about it at all, we read bad advice and repeat it, pretending that this makes us wise.<\/p><p>Dude, it\u2019s five minutes. \u00a0Let\u2019s un-suck your negotiation.<\/p><p>(New to the blog? \u00a0Hiya. \u00a0I generally write as an engineer for engineers. \u00a0Non-engineers can benefit from many of the same techniques, though the hiring market isn\u2019t nearly as in your favor at the moment as it is for engineers in most major US metro areas.)<\/p><h2 id=\"why-negotiation-matters\">Why Negotiation Matters<\/h2><p>Your salary negotiation \u2014 which routinely takes less than 5 minutes to conclude \u2014 has an outsized influence on what your compensation is. \u00a0Compensation can include money or things which are more-or-less fungible replacements for money, but it can also include interesting things which you value from \u201cmore time with your family\u201d to \u201copportunities to do tasks which you find fulfilling\u201d to \u201cperks which make a meaningful difference in your day-to-day quality of life.\u201d \u00a0That makes your negotiation <strong>five very important minutes<\/strong>. \u00a0You generally can\u2019t do a totally bang up job on any five minutes of work this year and have your boss give you an extra $5,000. \u00a0<strong>You can trivially pick up $5,000 in salary negotiations just by sucking <em>less<\/em>.<\/strong><\/p><p>Since salaries are shockingly durable over time, particularly if one is not good at negotiating, you can expect a $5,000 increase in salary to compound with your standard annual read-the-HR-chart-percent raise, cause a similar increase in your 401k contribution (which also compounds), and establish a higher peg for any further jobs you take (if you\u2019re unsavvy and allow these other jobs access to your prior salary history, at any rate). \u00a0Accordingly, over the next ten years, the value of $5,000 a year extra salary is close to $100k gross, and the value of $15,000 a year extra (very achievable if you\u2019re e.g. a young engineer who doesn\u2019t realize that the hiring market is <em><strong>on fire<\/strong><\/em> right now, even outside of tech epicenters like Silicon Valley) is over $100k even net of taxes.<\/p><h2 id=\"shifting-your-mindset-to-embrace-negotiation\">Shifting Your Mindset To Embrace Negotiation<\/h2><p>We\u2019ll discuss tactical advice in a moment, but let\u2019s talk about the psychology of negotiation first. \u00a0I think that middle class Americans are socialized from a very young age to view negotiation as something that is vaguely disreputable and engaged in only by poor people. \u00a0Think of the associations you have with the word \u201chaggling\u201d: do you think of a successful young professional talking about thousands of dollars in a brightly lit office? \u00a0No, you probably think of an old woman arguing over a trivial sum of money in a dirty flea market.<\/p><p>If I were a little more paranoid and a little more Marxist, I\u2019d honestly think that you\u2019re so misinformed about reality that that is almost prima facie evidence of a conspiracy to keep you in the dark about this, to the advantage of people who a) you won\u2019t negotiate with and b) who will feel absolutely no compunctions about negotiating with you. \u00a0Principally, this will be your employers. \u00a0People say that your house is the biggest purchase you\u2019ll ever make, but it won\u2019t be the most consequential negotiation. \u00a0If you\u2019re sane only about 25% or so of your gross income is subject to the results of real estate negotiations. \u00a0Close to 100% is subject to the results of salary negotiations. \u00a0Thus, your salary negotiations are probably going to be the most important financial decisions you will ever make. \u00a0We socialize middle class Americans to go into them unprepared, demotivated, and fearful of success.<\/p><p>The reality is that <strong>rich, successful people negotiate<\/strong>. \u00a0(This is one important way in which they get \u2014 and stay \u2014 rich.) \u00a0It is an all-day-every-day thing in much of the business world, which is where most rich people get their money.<\/p><p>Salary negotiations are very\u00a0asymmetrical. \u00a0\u00a0Companies know this and routinely exploit it. \u00a0Job seekers don\u2019t, perhaps because they think doing so would be unfair and the word \u201cexploit\u201d makes them acutely uncomfortable. \u00a0So we often default by pretending that the employer is evaluating the negotiation like we would. \u00a0This is not true, and acting like it is true will harm both your interests and the interests of your future employer.<\/p><p>For example, many people\u2019s mental model of employment is that an employee with a $60,000 a year salary costs about $60,000 a year to hire. \u00a0If they negotiate $65,000 instead, that\u2019s $5,000 extra which has to come from\u2026 somewhere. \u00a0If the negotiation breaks down, then that is $60,000 saved. \u00a0<strong>This mental model is broken<\/strong>.<\/p><p>First, get into the habit of seeing employees like employers see them: in terms of fully-loaded costs. \u00a0To hire someone you need to pay for their salary, true, but you also have taxes, a benefits package, employer contributions to retirement, healthcare, that free soda your HR department loves mentioning in the job ads, and what have you. \u00a0(Trivia: for a US employer of professionals, the largest component after salary is usually healthcare, followed by payroll taxes.) \u00a0The fully-loaded costs of employees are much higher than their salary: exactly how much higher depends on your locality\u2019s laws, your benefits package, and a bunch of other HR administrivia, but a reasonable guesstimate is between 150% and 200% of their salary.<\/p><p>The fully loaded cost of an engineer receiving market salaries these days in California or New York is close to $20,000 a month. \u00a0It is \u201conly\u201d $10,000 a month if they\u2019re receiving a heavily below-market salary, such as if they\u2019re working for a startup. \u00a0If you have a kid brother who majored in Flemish Dance and got a modest full-time job at a non-profit, his fully-loaded cost is still probably $4,000 a month or more.<\/p><p>This is a roundabout way of telling you that companies are not sensitive to small differences in employee wages because employees are so darned expensive anyhow. \u00a0You see $5,000 and think \u201cHoly cow, even after taxes that\u2019s a whole new vacation. \u00a0Five thousand dollars. \u00a0Five thousand dollars. \u00a0It would be so very, very greedy of me to ask for five thousand whole dollars.\u201d \u00a0The HR department sees $5,000 and thinks \u201cMeh, even after we kick in the extra taxes, that is only about 3% of their fully-loaded cost for this year anyhow, or seven hundredths of one percent of that team\u2019s hiring budget. \u00a0I wonder if the cafeteria has carrot cake today?\u201d<\/p><p><strong>Virtually any amount of money available to you personally is mouse droppings to your prospective employer<\/strong>. \u00a0They will not feel offended if you ask for it. \u00a0(I received a comment that this is untrue for startups by someone today. \u00a0For a funded startup which has enough engineers to warrant a foosball table, the company payroll is well north of $100,000 a month. \u00a0Making a new hire is a big commitment, but they still have a lot of flexibility on the \u00a0details because the details do not shave months off of their runway.)<\/p><p>We\u2019ve been talking about your employer as an abstraction, but in the instant case you\u2019re talking to an actual person. \u00a0Let\u2019s call him Bob. \u00a0It is Bob\u2019s job to get you signed with the company as cheaply as possible, but Bob is not super motivated to do so, because <strong>Bob is not spending Bob\u2019s money to hire you<\/strong>. \u00a0Bob is spending Bob\u2019s <em>budget<\/em>. \u00a0Bob generally does not get large performance incentives for shaving money off of his hiring budget: you get a new Macbook if you convince Bob to give you $5k extra, but Bob gets (if he is anomalously lucky) a dinner at TGIFridays if he convinces you to take $5k less. \u00a0In fact, there are many organizations (and Bobs) for whom power, status, and money come from asking for <em>more<\/em> budget every year. \u00a0If you turn out to be on the expensive side, that is great for Bob, because a) he manages a high-powered peon so he must be a high-powered manager and b) this will help Bob get more budget next quarter. \u00a0So if you\u2019re worried about what Bob will think of your moral character, or you want to compensate Bob because you feel you owe him for this job opportunity, do Bob a solid and negotiate in a spirited fashion with him.<\/p><p>You don\u2019t owe Bob for giving you this job opportunity, by the way. \u00a0Internalize this: everyone in this discussion is a businessman. \u00a0(Some might call themselves \u201cregular employees,\u201d which just means they\u2019re businessmen with self-confidence issues and poor business skills.) \u00a0If the deal makes economic sense, it will happen. \u00a0If it doesn\u2019t, firm handshakes will be exchanged, non-specific promises will be uttered, and everyone will forget about this discussion in a matter of hours. \u00a0You will not be blackballed for negotiating. \u00a0Bob couldn\u2019t care less and, even if he did care, he has better things to do with his time than worry about a candidate he didn\u2019t hire. \u00a0Bob is working through a list of a dozen people right now, and his manager Dave is being such a hard case about that project\u2019s schedule, and he\u2019s not sure he can make his daughter\u2019s piano recital, and the cafeteria\u2019s carrot cake was a little dry. \u00a0Bob is <strong>far, far less invested in this negotiation than you are<\/strong>.<\/p><h2 id=\"your-negotiation-started-before-you-applied-to-this-job\">Your Negotiation Started Before You Applied To This Job<\/h2><p>Your negotiation doesn\u2019t happen in a vacuum. \u00a0Generic career advice is a little outside the scope of this post (though I\u2019ve previously written <a href=\"https:\/\/www.kalzumeus.com\/2011\/10\/28\/dont-call-yourself-a-programmer\/\">a bit with engineers in mind<\/a> that folks from many walks of life tell me was useful), but to make a long story short, many people think job searches go something like this:<\/p><ol><li>See ad for job on Monster.com<\/li>\n  <li>Send in a resume.<\/li>\n  <li>Get an interview.<\/li>\n  <li>Get asked for salary requirements.<\/li>\n  <li>Get offered your salary requirement plus 5%.<\/li>\n  <li>Try to negotiate that offer, if you can bring yourself to.<\/li>\n<\/ol><p>This is an effective strategy for job searching if you enjoy alternating bouts of being unemployed, being poorly compensated, and then treated like a disposable peon. \u00a0(I served three years as a disposable peon in a Japanese megacorp and might be projecting a tad bit here. \u00a0Regardless, my loss is your gain.)<\/p><p>You will have much, much better results if your job search looks something more like:<\/p><ol><li>(Optional but recommended) Establish a reputation in your field as someone who delivers measurable results vis-a-vis improving revenue or reducing costs.<\/li>\n  <li>Have a hiring manager talk with you, specifically, about an opening that they want you, specifically, to fill.<\/li>\n  <li>Talk informally (and then possibly formally) and come to the conclusion that this would be a great thing if both sides could come to a mutually fulfilling offer.<\/li>\n  <li>Let them take a stab at what that mutually fulfilling offer would look like.<\/li>\n  <li>Suggest ways that they could improve it such that the path is cleared for you doing that voodoo that you do so well to improve their revenue and\/or reduce their costs.<\/li>\n  <li>(Optional) Give the guy hiring you a resume to send to HR, for their records. \u00a0Nobody will read it, because <strong>resumes are an institution created to mean that no one has to read resumes<\/strong>. \u00a0Since no one will read it, we put it in the process where it literally doesn\u2019t matter whether it happens or not, because if you had your job offer contingent on a document that everyone knows no one reads, that would be pretty effing stupid now wouldn\u2019t it.<\/li>\n<\/ol><p>You might think that desirable jobs at well-managed companies (Google, Microsoft, hot startup FooWithTheWhat.ly, etc) have layers and layers of bureaucratic scar tissue (a <a href=\"http:\/\/37signals.com\/svn\/archives2\/dont_scar_on_the_first_cut.php\">great image from 37Signals<\/a>) to ensure that their hiring will conform to established processes and that offers will not be given to candidates sourced by using informal networks and interpersonal connections. \u00a0If you believe this, you have a dangerously incomplete mental model of how the world operates. \u00a0I have a specific recommendation for you to make that model more complete: start talking to people who actually work for those companies and who have hiring authority. \u00a0Virtually no company has a hiring process which is accurately explained by blog posts about the company. \u00a0No company anywhere has a hiring process which is accurately explained by their own documents about how the hiring process works.<\/p><p>I won\u2019t give names, but all of the following are companies you\u2019ve heard of:<\/p><ul><li>Ironclad non-compete with an IP assignment provision of major multinational\u2026 struck from the contract with four sentences of discussion.<\/li>\n  <li>Major popular tech employer offered desirable employee $X as a salary because \u201cit was the max the HR department allows for that position.\u201d \u00a0He got hired that week at $2X. \u00a0All parties \u2014 hiring organization, HR, and employee \u2014 think they pulled one over on the other participants.<\/li>\n  <li><del>Funny story goes here<\/del>. I now can\u2019t tell you the funny story, because literally two hours before publication someone emailed me for advice about a situation that he believes is incredibly unjust at his company, and it is exactly the funny story to the letter. \u00a0Now if I tell you the funny story he might think \u201cDang, I write Patrick in confidence and it ends up on the blog.\u201d \u00a0So, no funny story today. \u00a0Suffice it to say that in my old age I treat Dilbert less as farce and more as documentary.<\/li>\n  <li>\u201cWe can\u2019t hire engineers fast enough through our standard processes so, meh, I guess we\u2019ll circumvent them by just tossing $1 million per employee at whomever they currently work for. \u00a0Who cares, it isn\u2019t <em>my<\/em> million.\u201d<\/li>\n<\/ul><h2 id=\"when-does-a-salary-negotiation-happen\">When Does A Salary Negotiation Happen?<\/h2><p>Only negotiate salary after you have agreement in principle from someone with hiring authority that, if a mutually acceptable compensation can be agreed upon, you will be hired.<\/p><p>This is really, really important because it has direct implications for your negotiating strategy. \u00a0First, the company is going to spend a lot of time and effort on getting you to the point of agreement-in-principle. \u00a0Pretend you\u2019ve gone through six rounds of interviews. \u00a0(You probably won\u2019t if you get hired on informal networks, because all barriers vanish when important people want a deal to get done, but let me give some advice to someone a little less well-situated.) \u00a0Do some quick mental math on what that actually cost the company, with reference to \u201cone man-month of an engineer\u2019s time costs $20k\u201d like we discussed earlier. \u00a0You\u2019ll quickly reach the conclusion that the company has spent thousands of dollars <em>just talking to you<\/em>, and that doesn\u2019t even count the thousands they spent deciding to talk to you instead of whoever isn\u2019t in the room right now. \u00a0Walking away from the negotiation means that they lose all that investment. \u00a0(Yeah, sunk cost fallacy and all, but since people predictably act in this fashion you should, well, predict that they will act in this fashion.) \u00a0They really <strong>want<\/strong>\u00a0to reach an agreement with you.<\/p><p>The second implication is that the inner serf worrying \u201cIf I even attempt to negotiate this, the deal will fall through\u201d is worrying for nothing. \u00a0They\u2019ve got thousands invested in this discussion by this point. \u00a0They <strong>want<\/strong> you. \u00a0The absolute worst outcome of negotiating an offer in good faith is that you will get exactly the contents of that offer. \u00a0Let me say that again for emphasis: <strong>negotiating never makes (worthwhile) offers worse<\/strong>. \u00a0This means you need what political scientists call a commitment strategy: you always, as a matter of policy, negotiate all offers. \u00a0(In this wide world I\u2019m sure you can find a company who still makes exploding offers, where you get one yay-or-nay and then the offer is gone. \u00a0You have a simple recourse to them: refuse them and deal with people who are willing to be professionals. \u00a0You\u2019re not a peasant. \u00a0Don\u2019t act like one.)<\/p><p>This also means <strong>you do not start negotiating until you already have a Yes-If<\/strong>. \u00a0(Yes-If we agree on terms.) \u00a0Do not start negotiating from No-But. \u00a0(No-But we might hire you anyway if you\u2019re really, really effing cheap.) \u00a0You don\u2019t want to work for a No-But for the same reasons that <a href=\"http:\/\/www.joelonsoftware.com\/articles\/GuerrillaInterviewing3.html\">smart employers hate hiring candidates who are a No-But<\/a> (No-But maybe if not on my team, etc). \u00a0If they\u2019re leaning to not hiring you, you will compromise excessively on negotiation to get them to hire you. \u00a0Compromising excessively is not the point of the exercise. \u00a0It is a seller\u2019s market for talent right now: sell to someone who is happy to buy.<\/p><p>This means that <strong>any discussion of compensation prior to hearing Yes-If is premature<\/strong>. \u00a0If you\u2019re still at the job interview and you\u2019re talking price\u00a0<strong>you are doing something wrong<\/strong>. \u00a0(Read the room: it is entirely possible that you came for a job interview, finished it, and proceeded directly to a salary negotiation. \u00a0That\u2019s probably suboptimal, but it is OK. \u00a0Just don\u2019t give the employer the option of having the schedule be job interview, salary negotiation, and back to job interview if they discover that you have a spine.) \u00a0The ideal resolution to the job interview is for both sides to be optimistic about the arrangement, and then you close with a warm handshake and \u201cI look forward to receiving your offer by, oh, would tomorrow be enough time for you to run the numbers?\u201d<\/p><p>You then have a high likelihood of doing your salary negotiation over email, which is likely to your advantage versus doing it in real time. \u00a0Email gives you arbitrary time to prepare your responses. \u00a0Especially for engineers, you are likely less disadvantaged by email than you are by having an experienced negotiator talking to you.<\/p><h2 id=\"the-first-rule-is-what-everyone-tells-you-it-is-never-give-a-number-first\">The First Rule Is What Everyone Tells You It Is: Never Give A Number First<\/h2><p>Every handbook on negotiation and every blog post will tell you not to give a number first. \u00a0This advice is almost always right. \u00a0It is so right, you have to construct crazy hypotheticals to find edge cases where it would not be right.<\/p><p>For example, if your previous salary was set during the dot-com bubble and you are negotiating after the bubble popped, you might mention it to anchor your price higher such that the step down will be less severe than it would be if you engaged in free negotiations unencumbered by the bubbilicious history. \u00a0Does this sound vaguely disreputable to you? \u00a0Good. \u00a0This vaguely disreputable abuse of history is what every employer asking for salary history, salary range, or desired salary is doing. \u00a0They are all using your previous anomalously low salary \u2014 a salary which did not reflect your true market worth, because you were young or inexperienced or unskilled at negotiation or working at a different firm or in another line of work entirely \u2014 to justify paying you an anomalously low salary in the future.<\/p><p><strong>Never give a number.<\/strong><\/p><p>\u201cBut Patrick,\u201d you cry. \u00a0\u201cI don\u2019t want to be <em>difficult<\/em>.\u201d \u00a0You\u2019re not being difficult. \u00a0You\u2019re not doing anything immoral. \u00a0You\u2019re not being unprofessional. \u00a0They\u2019re businessmen, sometimes they don\u2019t have all the information they would love to have prior to making a decision. \u00a0They\u2019ll deal.<\/p><p>They <em>already<\/em> deal with every employee that they\u2019ve ever had who was not a doormat at negotiations, <strong>which includes essentially all of the employees they really value<\/strong>. \u00a0Ramit Sethi (more on him later) introduced me to a concept that he calls Competence Triggers: basically, if you have to judge someone\u2019s skill based on a series of brief interactions, you\u2019re going to pattern match their behavior against other people who you like. \u00a0When people with hiring authority think of winners, they think of <em>people like them<\/em> who live and breathe this business thing. \u00a0They negotiate things as a matter of course: that is a major portion of the value they bring to the company. \u00a0Volunteering a number when asked says the same thing to people with hiring authority that flunking FizzBuzz says to an engineer: this person may be a wonderful snowflake in other regards, but on the thing I care about, they\u2019re catastrophically incompetent. \u00a0It will also cause them to retroactively question competencies they\u2019d previously credited you with.<\/p><p>I have literally heard that feedback, in so many words, from folks with whom I\u2019ve had successful business dealings. \u00a0(A funny in hindsight story: I cost myself five figures with a single email. \u00a0The particulars are boring, but suffice it to say I fairly recently made a wet-behind-the-ears-engineer error in quoting a client. \u00a0He noticed. \u00a0So did my bank statement. \u00a0My bank statement kept quiet, but the client opined that it made him think less of me until we actually got to work together.)<\/p><p>So anyhow, you may well hear reasons why you should give a number.<\/p><p><strong>Objection<\/strong>: \u201cI really need a number to move the process forward.\u201d<\/p><p><strong>What you should think<\/strong>: \u201cYou\u2019re lying to me to attempt to get me to compromise my negotiating position.\u201d<\/p><p><strong>What you should say<\/strong>: \u201cI\u2019m more concerned at the moment with talking to you about discovering whether we\u2019re a mutual fit. \u00a0If we\u2019re a great fit, then I can be flexible on the numbers with you and you can be flexible on the numbers with me. \u00a0If we\u2019re not a great fit, then the numbers are ultimately irrelevant, because your company only hires A players and I only work at roles I would be an A player at.\u201d<\/p><p>(Don\u2019t talk like that normally? \u00a0Fine then, talk like yourself, but say substantially the same things. \u00a0Engineers overestimate how different we really are from business people: we say \u201c10x engineer,\u201d they say \u201cA player,\u201d but at the end of the day we believe that there are vast differences in productivity between workers. \u00a0OK, gut check: is this something we actually believe to be true or just something we wish for? \u00a0If it is actually your best guess about the state of reality, that has immediate news-you-can-use implications about how you should conduct your life.)<\/p><p><strong>Objection<\/strong>: \u201cThis form needs a number.\u201d<\/p><p><strong>What you should think<\/strong>: \u201cYou\u2019re lying to me to attempt to get me to compromise my negotiating position.\u201d<\/p><p><strong>What you should say<\/strong>: \u201cGive me git access and I\u2019ll fix it in a jiffy! \u00a0<em>both people laugh<\/em> No, seriously, speaking, I\u2019m more concerned at the moment with discovering whether we\u2019re a mutual fit\u2026 \u00a0Oh, it\u2019s physically impossible? \u00a0Put in $1 then to get the ball rolling, and we\u2019ll circle back to this later.\u201d<\/p><p><strong>Objection<\/strong>: \u201cWe want to figure out whether you\u2019re an appropriate candidate for the position.\u201d<\/p><p><strong>What you should think<\/strong>: \u201cYou\u2019re lying to me to attempt to get me to compromise my negotiating position.\u201d<\/p><p><strong>What you should say<\/strong>: \u201cIt\u2019s so important to me that this is a good mutual fit for us. \u00a0Let\u2019s talk about why I\u2019m a great fit for this position: I know you\u2019re concerned about $FILL_IN_THE_BLANK. \u00a0In addition to my previous successes doing it, I have some great ideas for what I\u2019d do about that if I was working at your company. \u00a0Would you like to drill into those or is there another job area you\u2019re more concerned about to start with?\u201d<\/p><p><strong>Objection<\/strong>: \u201cI\u2019m sorry, great try at a dodge there, but I just can\u2019t go forward without a number.\u201d<\/p><p><strong>What you should think<\/strong>: \u201cYou\u2019re lying to me to attempt to get me to compromise my negotiating position.\u201d<\/p><p><strong>What you should say (if you\u2019re an engineer): \u201c<\/strong>Well, you know, I would hate to have to walk away from the negotiation over this. \u00a0Working with your company looked it would have been such a wonderful opportunity. \u00a0I hear the hiring market is super-tight right now, would you like me to introduce you to other candidates? \u00a0Maybe we can shave a couple of months off of you filling this position.\u201d<\/p><p><strong>What you should say (if you\u2019re not an engineer)<\/strong>: \u201cDamn, I guess I should have studied engineering.\u201d<\/p><p><strong>What you should say (if you\u2019re a little put out by that comment)<\/strong>: \u201cWell, you know, salary is only one component of the total compensation package. \u00a0In terms of total compensation, we\u2019re probably looking at something like $FILL_IN_NUMBER_HERE.\u201d \u00a0(Suggested calculation: take the package value from your last company and add 5~10%. \u00a0If you don\u2019t know how to calculate the value of your compensation package, <em>learn that<\/em>, but as a rough guesstimate salary + 30 ~ 50% for full-time employees in professional roles and the multiplier tends to scale up as your base salary scales up.)<\/p><p>P.S. I double majored in making things and making things up. \u00a0The joking comes from a place of love. \u00a0OK, love and\u00a0schadenfreude, in solution with each other.<\/p><h2 id=\"listen-to-what-people-tell-you-repeat-it-back-to-them\">Listen To What People Tell You. \u00a0Repeat It Back To Them.<\/h2><p>Properly run negotiations are not jockeying contests, they\u2019re persuasive\u00a0exercises. \u00a0(We\u2019ll give the company a pass on the \u201cwhat\u2019s your number?\u201d question because it is an established social ritual that they get one free pass at screwing you. \u00a0You still don\u2019t have to cooperate with it, though.) \u00a0You know what people find persuasive? \u00a0Their own words. \u00a0People love their own words. \u00a0When you talk to them, you should use their own words. \u00a0Seriously, watch the eyes light up.<\/p><p>Did the solicitation for the job say \u201cWe are seeking someone with strong skills at scaling traffic in a fast-moving environment\u201d? \u00a0Pick out the key words. \u00a0Scaling traffic. \u00a0Fast-moving environment. \u00a0\u201cScaling traffic\u201d doesn\u2019t sound like how I\u2019d phrase it if I were writing or speaking for myself, but if you\u2019ve just described your need to me as scaling traffic, by golly I will tell you how great I am at scaling traffic. \u00a0Reinterpret or rephrase the (true!) bits of your own story such that it fits the narrative framework which they have conveniently told you that they are going to respond to. \u00a0Did you previously work at a small business which was unencumbered by lots of process? \u00a0Sounds like a fast-moving environment, right? \u00a0Call it <em>exactly<\/em> that, then.<\/p><p><strong>Micro-tip<\/strong>: Take notes during job interviews and salary negotiations. \u00a0It\u2019s easy: go to the convenience store before the job interview, buy a writing instrument and a $1 notebook, jot down occasional notes when appropriate.<\/p><p><em>Can I do that?!<\/em>\u00a0 Of course you can. \u00a0Do you know anyone who you\u2019ve ever thought \u201cMan, I thought they were competent, but then it turned out they had a notebook so I had to write them off?\u201d \u00a0No. \u00a0Taking notes says \u201cI\u2019m attentive and detail-oriented and I care about what you say.\u201d \u00a0(Make sure you can take notes without playing with your pen or otherwise appearing to fidget.) \u00a0In terms of specific things that should get your pen moving, among others, I would focus on specific words they use and concerns they have so that you can come back to them later in the conversation. \u00a0Numbers are another good thing to hit the notebook, because numbers should only ever trend in a direction of \u201cBetter to you,\u201d so you don\u2019t want to do something stupid like saying \u201cSo how many days of vacation was that again?\u201d and let a 24 suddenly become a 20. \u00a0(You might think \u201cI\u2019m going to write down the offer so I have proof of it for later.\u201d \u00a0Get offers written, that goes hopefully without saying, but get it written\u00a0<em>by them<\/em>\u00a0and\/or follow-up the discussion with an email recapping the important points and asking if you understood them correctly. \u00a0Your notes will not convince their HR\u00a0apparatus\u00a0to honor the agreement in event of a retroactive miscommunication, but an email from their decisionmaker likely will.)<\/p><p>People say the\u00a0damnedest\u00a0things. \u00a0For example, someone might spontaneously volunteer during a job interview that they\u2019ve been interviewing for the position for six months. \u00a0(None of my clients would ever say that, of course, but then again one would hope none of their consultants would chop five figures off their own invoice with an email.) \u00a0If they say the position has been open for six months, take a note of that. \u00a0During the salary negotiation, if they have a pricing objection, one of your first responses should be \u201cI appreciate that this is a little more money than you might have been thinking about, but this is an opportunity to get this position filled without <em>delaying your business by another six months<\/em>. \u00a0What is the value of that six months of execution to you?\u201d \u00a0(Conversely, don\u2019t say stupid things during job interviews such as \u201cI need this job because\u2026\u201d \u00a0You never need a job. \u00a0Being needy means that the party who is not needy has automatic leverage over you: your <a href=\"http:\/\/en.wikipedia.org\/wiki\/Best_alternative_to_a_negotiated_agreement\">BATNA<\/a> to the negotiation is very poor. \u00a0Instead of being needy, aim for \u201cI\u2019m enthusiastic about the opportunity with working with you, assuming we can come to mutually satisfactory terms.\u201d)<\/p><p><strong>Micro-tip<\/strong>: Notice how often I say \u201cWe\u201d and variations on \u201cmutual win.\u201d \u00a0Those work pretty well. \u00a0The only thing better than \u201cWe\u201d is \u201cYou\u201d (and variants), because people care a heck of a lot more about their problems than about your problems. \u00a0(This advice is stolen shamelessly from <a href=\"http:\/\/www.amazon.com\/How-Win-Friends-Influence-People\/dp\/1439167346\">Dale Carnegie<\/a>.) \u00a0This means that a) you should talk about their problems, concerns, and wishes and b) you should guard against your own natural tendency to bring up irrelevant things like your own problems, which typically will not help you sell the decisionmaker on adopting the mutual win you\u2019re proposing. \u00a0Similarly, I generally try to phrase things positively rather than score debating points. \u00a0(\u201cYou just said X, but that was contradicted by your earlier statement Y, which means\u2026\u201d wins debating points but does not win friends and influence people. \u00a0You might try something like \u201cGood good, but taking into account your earlier concerns about Y\u2026\u201d)<\/p><p>Many people will tell you that you should familiarize yourself with the approximate salary range for the position in your region. \u00a0This advice is easy to act on (go to a salary aggregation site, guess what \u201cthe position\u201d is, pray that this gives you a better number than rand(40000,120000)), but it leaves a lot to be desired. \u00a0It is 2012. \u00a0Facebook and LinkedIn exist. \u00a0You should, before any job interview, have intimate knowledge of the target company. \u00a0Prospective peers within the company are one obvious way to get it. \u00a0So are ex-employees, folks who\u2019ve had dealings with them professionally, etc. \u00a0Key things you want to learn:<\/p><ul><li>What do they value?<\/li>\n  <li>Who do they value within the company? \u00a0(Roles? \u00a0Titles? \u00a0Groups?)<\/li>\n  <li>What does the career path look like for successful people within the company?<\/li>\n  <li>Roughly speaking, how generous are they with regard to axes that you care about?<\/li>\n  <li>Do they have any compensation levers which are anomalously easy to operate? \u00a0(For example, if you asked around, you might hear a few people say that a particular firm pushes back modestly on out-of-band increases in salary but they\u2019ll give in-the-money option grants like candy.)<\/li>\n  <li>All the fuzzy stuff: what\u2019s the corporate culture like? \u00a0Yadda yadda.<\/li>\n<\/ul><p>You can even bring a lot of these questions to the job interview, which is (again) prior to the negotiation. \u00a0(Maybe not \u201cSo are you guys tightwads?\u201d but culture-esque questions like \u201cWhat are the projects this company thinks are really key to its future and how would a motivated person go about getting on them?\u201d are both a) totally fair game and b) will win you brownie points just for asking. \u00a0Similarly, a lot of employees will, out of company loyalty, attempt to sell you on taking the job with the company by trading you very useful information.)<\/p><p>The more you know, the more options you have when doing negotiation, because you\u2019ll have more things and more motivational things which you can offer in exchange for things you want. \u00a0It will also help you avoid making mistakes like e.g. getting into a rigid classification system where the classification you\u2019re aiming at will make forward advancement towards your goals very difficult. \u00a0(Example: there are some companies where Product and QA are run like separate fiefdoms which haven\u2019t forgotten the most recent war, and in those companies getting hired as an engineer may not be a career enhancing move if you like making things for a living. \u00a0There are other companies where people cross-function in those responsibilities all the time and applying for a job advertising as \u201cSupport Engineer\u201d makes lateral moves onto customer-facing projects trivial. \u00a0You can find which one you\u2019re applying to by taking any engineer out for coffee.)<\/p><p>There was a <a href=\"http:\/\/rachelbythebay.com\/w\/2012\/01\/21\/notfree\/\">post<\/a> recently on Hacker News about someone\u2019s experience with a job offer from Google. \u00a0They wanted more money. \u00a0The recruiters offered to think it over, and came back with the reply that Google\u2019s food benefit was worth a significant amount of money, with a calculation to back it up. \u00a0That is a pretty brilliant reply. \u00a0Google\u2019s food benefit is about as old as the company. \u00a0Approximately all people wanting to work at Google are aware of its existence. \u00a0However, the explicit calculation of what it is worth is new, so if you bring up that calculation, by implication you\u2019re offering newly found value to the negotiation. \u00a0This successfully convinces people that they didn\u2019t really need that extra money. \u00a0It is so successful at this that Google recruiters apparently have this entire interaction scripted, since multiple people report having the exact same experience.<\/p><p>You should steal this tactic. \u00a0You are an expert in your own skill set, life story, and (ideally) value you can create for the company. \u00a0However, the person you are talking to is not. \u00a0If they ever resist about something which you want, consider reaching into the treasure chest that they are buying mostly blind and revealing one of the many glittering jewels inside. \u00a0They are going to get them all anyhow if they buy the chest, but each one you bring out decreases the perceived risk of buying it and therefor increases its perceived value.<\/p><p><strong>Company<\/strong>: We can\u2019t see our way to $88,000.<\/p><p><strong>Applicant<\/strong>: Well, I know you do a significant amount of business with your online store. \u00a0At my last company, I increased sales by 3% by $YADDA_YADDA. \u00a0What would a 1% increase in sales be worth to you?<\/p><p><strong>Company<\/strong>: Well, I don\u2019t have that figure in front of me, but\u2026<\/p><p><strong>Applicant<\/strong>: Would it be safe to say \u201cmillions of dollars\u201d?<\/p><p><strong>Company<\/strong>: That sounds about right, yeah.<\/p><p><strong>Applicant<\/strong>: Great, I can\u2019t wait to get started. \u00a0Getting me that extra $4,000 would make this a much easier decision. \u00a0Considering that this is conceivably worth millions to you, we\u2019d be silly not to do business with each other.<\/p><p><strong>Company<\/strong>: I\u2019ll see what I can do.<\/p><p><strong>Applicant<\/strong>: Let me help give you some options! \u00a0[See below.]<\/p><p>(This hypothetical applicant is doing well on the negotiation but apparently needs to do more research on what conversion optimization specialists can get away with charging these days. \u00a0Here, let me help: six figure salary with all the usual perks as an employee, \u201csenior engineer project rates\u201d through \u201cyou might not believe me if I told you\u201d as a consultant.)<\/p><p>Anyhow, simply by bringing attention to something which was hopefully already in bold print on their resume, they just increased their perceived value to the company, thus justifying the company moving a lever which (again) the company isn\u2019t really sensitive to at the end of the day.<\/p><h2 id=\"you-have-a-multi-dimensional-preference-set-use-it\">You Have A Multi-Dimensional Preference Set. \u00a0Use It.<\/h2><p>Don\u2019t overly focus on your salary number. \u00a0It is important (of course), but there are many parts of your compensation package, and many more things that you value. \u00a0Should you and the other party reach an impasse on any part of it, offer to table that part of the discussion (to be returned to later) and bring up a different topic. \u00a0You can then trade improvements for concessions (or apparent concessions) on the earlier topic.<\/p><p><strong>Employer<\/strong>: \u201cWe were thinking $80,000.\u201d<\/p><p><strong>Applicant<\/strong>: \u201c$80,000 is interesting (*) but not quite where we need to be to get this done. \u00a0Do you have any flexibility on that number?\u201d<\/p><p><strong>Employer<\/strong>: \u201cI think I can convince HR to approve $84,000 but that is the best I can do.\u201d<\/p><p><strong>Applicant<\/strong>: \u201cI appreciate that. \u00a0$84,000, huh. \u00a0Well, it isn\u2019t quite what I had in mind, but the right package offer could make that attractive. \u00a0How much vacation comes with the package?\u201d<\/p><p><strong>Employer<\/strong>: \u201c20 days a year.\u201d<\/p><p><strong>Applicant<\/strong>: \u201cIf you could do 24 days a year, I could compromise on $84,000.\u201d<\/p><p><strong>Employer<\/strong>: \u201cI think I can do that.\u201d<\/p><p>For those keeping score at home: the applicant never gives up anything but the employer will walk away feeling he got a good deal.<\/p><p><strong>* Micro-tip<\/strong>: \u201cInteresting\u201d is a wonderful word: it is positive and non-commital at the same time. \u00a0If they tell you a number, tell them it is an \u201cinteresting\u201d number, not a \u201cwonderful\u201d number.<\/p><p>Hoping around the offer also helps you defuse common negotiating tactics like \u201cI have to go to $EXTERNAL_AUTHORITY to get approval of that.\u201d \u00a0(This is in the negotiation playbook, because it works well: it injects an automatic delay in the process, and gives you a scapegoat for refusing a request while not being guilty of the refusal yourself. \u00a0You should strongly consider having an $EXTERNAL_AUTHORITY of your own. \u00a0Significant others work well. \u00a0Note that in the US your would-be employer is legally prohibited from breathing about the subject of your marital status, so something like \u201cWe\u2019ll have to talk that over\u201d or \u201cThat sounds reasonable, but I\u2019ll have to run it by the family\u201d has the dual virtues of being a socially acceptable reason to delay any major decision while also being equally available to unattached young\u2019uns. \u00a0I talk shop with my family all the time. \u00a0I\u2019ll certainly continue discussing employment with my family after it includes my\u00a0fiancee, too.)<\/p><p>Anyhow, say your decisionmaker says that approving deviations from the company\u2019s salary structure is outside of his discretion and those evil ogres in HR will likely deny his request. \u00a0That\u2019s fine. \u00a0Express sympathy with him, because he just said he wants to give you more but can\u2019t, then refocus the discussion on things which are within his personal authority. \u00a0(Vacation days, work hours, project assignments, travel opportunities, professional development opportunities, and the like are good areas to probe at.) \u00a0You can then use the unspent \u201cYou wanted to do something nice for me\u201d obligation which he just acknowledged on one of the things which he has authority to grant you.<\/p><h2 id=\"for-your-further-perusal\">For Your Further Perusal<\/h2><p>I\u2019m deeply indebted to a few buddies of mine, principally Thomas at <a href=\"http:\/\/www.matasano.com\">Matasano<\/a> and <a href=\"http:\/\/www.iwillteachyoutoberich.com\">Ramit Sethi<\/a>, for teaching me to be less of a doormat in terms of negotiation. \u00a0Thomas has forgotten more than I\u2019ll ever know about doing negotiations with clients. \u00a0Check out <a href=\"https:\/\/hn.algolia.com\">Hacker News search<\/a> with [tptacek negotiation] for some good advice, or (if you\u2019re in Chicago) take him out to coffee.<\/p><p>Some years after I wrote this article, Josh Doody, one of my buddies, wrote <a href=\"https:\/\/www.amazon.com\/Fearless-Salary-Negotiation-step-step\/dp\/0692568689\">the literal book<\/a> on salary negotiation. If you learn best from books, I recommend it. If you\u2019d prefer more personalized advice, <a href=\"https:\/\/fearlesssalarynegotiation.com\/salary-negotiation-guide\/\">he has that available<\/a>, too.<\/p>","created_at":"2018-11-07T16:22:29+0000","updated_at":"2018-11-07T16:22:29+0000","annotations":[],"mimetype":"text\/html","reading_time":35,"domain_name":"www.kalzumeus.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":79,"title":"ahungry\/geben","url":"https:\/\/github.com\/ahungry\/geben","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>Geben is a software package that interfaces Emacs to DBGp protocol\nwith which you can debug running scripts interactively. At this point,\nthe DBGp protocol is supported in several scripting languages with\nhelp of custom extensions.<\/p>\n<ul><li>PHP with Xdebug 2.*<\/li>\n<li>Perl, Python, Ruby and Tcl with Komodo Debugger Extensions<\/li>\n<\/ul><p><a href=\"https:\/\/melpa.org\/#\/geben\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a90e4f042b138eb45ba955769349514b9fa32d24\/68747470733a2f2f6d656c70612e6f72672f7061636b616765732f676562656e2d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"https:\/\/melpa.org\/packages\/geben-badge.svg\" \/><\/a><\/p>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#geben\">Geben<\/a>\n<ul><li><a href=\"#features\">Features<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#installation\">Installation<\/a>\n<ul><li><a href=\"#php\">PHP<\/a>\n<ul><li><a href=\"#command-line\">Command Line<\/a><\/li>\n<li><a href=\"#web\">Web<\/a><\/li>\n<li><a href=\"#vms\">VMs<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#perl\">Perl<\/a><\/li>\n<li><a href=\"#python\">Python<\/a><\/li>\n<li><a href=\"#ruby\">Ruby<\/a><\/li>\n<li><a href=\"#tcl\">Tcl<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#usage\">Usage<\/a><\/li>\n<li><a href=\"#about\">About<\/a>\n<ul><li><a href=\"#known-issues\">Known Issues<\/a>\n<\/li>\n<li><a href=\"#support\">Support<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><\/li>\n<\/ul><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<ul><li>continuation commands: run\/stop\/step-in\/step-over\/step-out<\/li>\n<li>set\/unset\/listing breakpoints<\/li>\n<li>expression evaluation<\/li>\n<li>STDOUT\/STDERR redirection<\/li>\n<li>backtrace listing<\/li>\n<li>variable inspection<\/li>\n<\/ul><p>We recommend installing geben via melpa: <code>M-x package-install geben<\/code>.<\/p>\n<p>Alternatively clone this repository and add it to your <code>load-path<\/code>.<\/p>\n<h2><a id=\"user-content-php\" class=\"anchor\" aria-hidden=\"true\" href=\"#php\"><\/a>PHP<\/h2>\n<p>To debug PHP scripts, you will need to install PHP, <a href=\"http:\/\/xdebug.org\" rel=\"nofollow\">Xdebug<\/a> and optionally a web server.  Please visit their official sites to get packages and instructions of installation and configuration:<\/p>\n<h3><a id=\"user-content-command-line\" class=\"anchor\" aria-hidden=\"true\" href=\"#command-line\"><\/a>Command Line<\/h3>\n<p>To begin testing this out, create a simple PHP file (\/tmp\/gebenTest.php) as such:<\/p>\n<div class=\"highlight highlight-text-html-php\"><pre>&lt;?php\n$x = 1;\necho $x;\n$y = 2;\necho $y;<\/pre><\/div>\n<p>You can now hop over to emacs, and enable the geben listener (default port 9000) by running <code>M-x geben<\/code>.<\/p>\n<p>At this point, you need to run the script with the xdebug configuration set up to point to this listener.<\/p>\n<p>Run this to invoke your script with xdebug dbgp enabled (or add to your CLI php.ini file to ensure it runs with these settings every time):<\/p>\n<div class=\"highlight highlight-source-shell\"><pre> php -d xdebug.remote_enable=on \\\n     -d xdebug.remote_host=127.0.0.1 \\\n     -d xdebug.remote_port=9000 \\\n     -d xdebug.remote_handler=dbgp \\\n     -d xdebug.idekey=geben \\\n     -d xdebug.remote_autostart=On \\\n     \/tmp\/gebenTest.php<\/pre><\/div>\n<p>You will notice the script does not complete, it pauses in your emacs session.  Press <code>SPC<\/code> to step forward a line at a time.<\/p>\n<h3><a id=\"user-content-web\" class=\"anchor\" aria-hidden=\"true\" href=\"#web\"><\/a>Web<\/h3>\n<ol><li>\n<p>Run Emacs.<\/p>\n<\/li>\n<li>\n<p>Start geben <code>M-x geben<\/code><\/p>\n<\/li>\n<li>\n<p>Access to the target PHP script page with any browser.\nYou may need to add a query parameter <code>XDEBUG_SESSION_START<\/code> if you configured Xdebug to require manual trigger to start a remote debugging session.\ne.g. <a href=\"http:\/\/www.example.com\/test.php?XDEBUG_SESSION_START=1\" rel=\"nofollow\">http:\/\/www.example.com\/test.php?XDEBUG_SESSION_START=1<\/a><\/p>\n<\/li>\n<li>\n<p>Soon the server and Geben establish a debugging session\nconnection. Then Emacs loads the script source code of the entry\npage in a buffer.<\/p>\n<\/li>\n<\/ol><h3><a id=\"user-content-vms\" class=\"anchor\" aria-hidden=\"true\" href=\"#vms\"><\/a>VMs<\/h3>\n<ol><li>\n<p>Add a mapping from files on the host to files on the vm\/docker image with <code>M-x customize-variable geben-path-mappings<\/code>\nor set in emacs config with <code>(setq geben-path-mappings '((\"&lt;project base on host&gt;\" \"&lt;project base on vm&gt;\"))<\/code><\/p>\n<\/li>\n<li>\n<p>Check those mappings with <code>C-h v geben-path-mappings<\/code>. Beware that the mappings only show up after geben was required.<\/p>\n<\/li>\n<li>\n<p>Open a file you are interested in setting breakpoints on the host machine. Find the line and issue <code>M-x geben-add-current-line-to-predefined-breakpoints<\/code>.<\/p>\n<\/li>\n<li>\n<p>Check that a breakpoint has been set with <code>C-h v geben-predefined-breakpoints<\/code><\/p>\n<\/li>\n<li>\n<p>Follow process in <a href=\"#web\">web<\/a><\/p>\n<\/li>\n<li>\n<p>To set new breakpoints in an active geben buffer, you can use <code>b<\/code> or <code>M-x geben-set-breakpoint-line<\/code>. To set a new breakpoint outside an active geben buffer, open the file on the host system and use <code>M-x geben-add-current-line-to-predefined-breakpoints<\/code> again.<\/p>\n<\/li>\n<\/ol><h2><a id=\"user-content-perl\" class=\"anchor\" aria-hidden=\"true\" href=\"#perl\"><\/a>Perl<\/h2>\n<p><a href=\"https:\/\/github.com\/ahungry\/geben\/issues\/28\">Documentation missing<\/a><\/p>\n<h2><a id=\"user-content-python\" class=\"anchor\" aria-hidden=\"true\" href=\"#python\"><\/a>Python<\/h2>\n<p><a href=\"https:\/\/github.com\/ahungry\/geben\/issues\/28\">Documentation missing<\/a><\/p>\n<h2><a id=\"user-content-ruby\" class=\"anchor\" aria-hidden=\"true\" href=\"#ruby\"><\/a>Ruby<\/h2>\n<p><a href=\"https:\/\/github.com\/ahungry\/geben\/issues\/28\">Documentation missing<\/a><\/p>\n<h2><a id=\"user-content-tcl\" class=\"anchor\" aria-hidden=\"true\" href=\"#tcl\"><\/a>Tcl<\/h2>\n<p><a href=\"https:\/\/github.com\/ahungry\/geben\/issues\/28\">Documentation missing<\/a><\/p>\n<h2><a id=\"user-content-nodejs\" class=\"anchor\" aria-hidden=\"true\" href=\"#nodejs\"><\/a>Node.js<\/h2>\n<p>Even though Node.js ha(s|d) some dbgp protocol support via the <a href=\"https:\/\/www.npmjs.com\/package\/komodo-debug\" rel=\"nofollow\">komodo-debug<\/a> that extension seems to be unmainted and succeded by the newer debug implementations from 7.x forward and geben is not supporting it.\nWe recommend <a href=\"https:\/\/github.com\/NicolasPetton\/jade\">jade<\/a> to debug node applications with Emacs.<\/p>\n<p>You can control the debugger with several keys.<\/p>\n<pre>     - spc     step into\/step over\n     - i       step into\n     - o       step over\n     - r       step out\n     - g       run\n     - c       run to cursor\n     - e       eval php expression\n     - b       set a breakpoint at a line (see: defcustom geben-predefined-breakpoints)\n     - B       set a breakpoint interactively\n     - u       unset a breakpoint at a line\n     - U       clear all breakpoints\n     - C-c b  display breakpoint list\n     - &gt;       set redirection mode\n     - C-u t  change redirection mode\n     - d       display backtrace\n     - t       display backtrace\n     - v       display context variables\n     - C-c f  visit script file\n     - w       where\n     - q       stop\n<\/pre>\n<p>When you hit any unbound key of <code>geben-mode<\/code>, Geben will ask you to\nedit the original script file. Say yes and Geben will attempts to\nload the script file via <code>TRAMP<\/code>.<\/p>\n<p>To quit Geben: <code>M-x geben-end<\/code><\/p>\n<h2><a id=\"user-content-php-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#php-1\"><\/a>PHP<\/h2>\n<ul><li>\n<p>Xdebug does not support STDERR command feature so that STDERR\nredirection feature does not work expectedly.<\/p>\n<\/li>\n<li>\n<p>Xdebug does not implement <code>dbgp:<\/code> scheme feature so that with\n<code>step-in<\/code> command into a lambda function (you can create it with\n<code>create_function<\/code> in PHP) or mocked function the cursor position is located at\nan invalid line. Geben could handle this more gracefully.<\/p>\n<\/li>\n<li>\n<p>Xdebug may tell invalid line number on breaking by <code>return<\/code> type\nbreakpoint. To this case Geben indicates the cursor at the top of\nthe file in where the current breakpoint exists.<\/p>\n<\/li>\n<li>\n<p>Xdebug unexpectedly breaks on returning from class\/instance method\nif there is a <code>call<\/code> type breakpoint to the method.<\/p>\n<\/li>\n<li>\n<p>Conditional breakpoints currently cannot be deleted<\/p>\n<\/li>\n<\/ul><p>We always need your support - bug reports, feature requests,\nand code\/documents\/design contributions.<\/p>\n<p>To submit one or more of them, please file an issue here or email\nMatthew Carter <a href=\"mailto:m@ahungry.com\">m@ahungry.com<\/a>.<\/p>\n<p>GPLv3<\/p>\n<\/article>","created_at":"2018-11-07T16:22:28+0000","updated_at":"2018-11-07T16:22:28+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/10437774?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":76,"title":"A Road to Common Lisp \/ Steve Losh","url":"http:\/\/stevelosh.com\/blog\/2018\/08\/a-road-to-common-lisp\/","content":"<p>I\u2019ve gotten a bunch of emails asking for advice on how to learn Common Lisp in\nthe present day.  I decided to write down all the advice I\u2019ve been giving\nthrough email and social media posts in the hopes that someone might find it\nuseful.<\/p><p>One disclaimer up front: this is <em>a<\/em> road to Common Lisp, not <em>the<\/em> road to\nCommon Lisp.  It\u2019s what I followed (without some of the dead ends) and has\na <em>lot<\/em> of my personal opinions baked in, but it is by no means the only way to\nlearn the language.<\/p><p>This post has been translated into\n<a href=\"https:\/\/gist.github.com\/y2q-actionman\/49d7587912b2786eb68643afde6ca192\">Japanese<\/a>.\nI can\u2019t vouch for the accuracy of any translations.<\/p><h2 id=\"context\">Context<\/h2><p>I think it\u2019s important to have a sense of where Common Lisp came from and what\nkind of a language it is before you start learning it.  There are some things\nthat will seem very strange if you\u2019re coming straight from modern languages,\nbut will make more sense if you\u2019ve got a bit of background context.<\/p><h3 id=\"history\">History<\/h3><p>Common Lisp has a long, deep history.  I\u2019m not going to try to cover it all here\n\u2014 if you\u2019re interested you should check out some of the following (in roughly\nincreasing order of detail):<\/p><ul><li>Wikipedia\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lisp_(programming_language)#History\">History of Lisp<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Common_Lisp#History\">History of Common Lisp<\/a>.<\/li>\n<li>The <a href=\"http:\/\/www.gigamonkeys.com\/book\/introduction-why-lisp.html#where-it-began\">Where it Began section in Practical Common Lisp<\/a>.<\/li>\n<li>The <a href=\"https:\/\/www.cs.cmu.edu\/Groups\/\/AI\/lang\/lisp\/faq\/lisp_2.faq\">History: Where did Lisp come from?<\/a> section of the comp.lang.lisp FAQ.<\/li>\n<li><a href=\"http:\/\/www.nhplace.com\/kent\/Papers\/cl-untold-story.html\">Common Lisp: the Untold Story<\/a> by Kent Pitman.<\/li>\n<li><a href=\"https:\/\/www.dreamsongs.com\/Files\/HOPL2-Uncut.pdf\">The Evolution of Lisp<\/a> by Guy Steele and Richard Gabriel.<\/li>\n<\/ul><p>I realize you probably won\u2019t want to read all of the links above immediately, so\nhere\u2019s a whirlwind tour of sixty years of Lisp.<\/p><p>Lisp began in the late 1950\u2019s.  It was invented by John McCarthy at MIT.<\/p><p>Over the next twenty or so years various versions and dialects of Lisp grew and\nflourished.  Some of the more notable dialects were Maclisp, BBN Lisp\/Interlisp,\nFranz Lisp, Spice Lisp, and Lisp Machine Lisp.  There were others too.  The\npoint is that there were a <em>lot<\/em> of different implementations, all growing,\nchanging, and trying out different things.<\/p><p>(Scheme also originated in this time frame, but took a very different route and\ndiverged from the path we\u2019re looking at.  I won\u2019t cover Scheme in this post.)<\/p><p>In the early 1980s people decided that having a whole slew of\nmutually-incompatible dialects of Lisp might be not be ideal.  An effort was\nmade to take these different languages that had grown organically and produce\none common language that would satisfy the needs of everyone (or at least\na reasonable subset of \u201ceveryone\u201d).  In 1984 the first edition of Guy Steele\u2019s\n<a href=\"https:\/\/www.cs.cmu.edu\/Groups\/AI\/html\/cltl\/cltl2.html\">Common Lisp: the Language<\/a> was published.<\/p><p>If you do some math you\u2019ll see that at the time the book was published Lisp had\naround twenty-five years of real-world use, experimentation, experience, and\nhistory to draw upon.  Even so, the book alone didn\u2019t quite satisfy everyone and\nin 1986 a committee (X3J13) was formed to produce an ANSI specification for\nCommon Lisp.<\/p><p>While the committee worked on the standardization process, in 1990 the second\nedition of Common Lisp: the Language was published.  This was more\ncomprehensive and contained some of the things the committee was working on\n(see the comp.lang.lisp FAQ linked above for more on this).  At this point the\nLisp family of languages had over thirty years of experience and history to\ndraw upon.  For comparison: Python (a \u201cmodern\u201d language many people think of as\nalso being \u201ckind of old\u201d) <a href=\"https:\/\/en.wikipedia.org\/wiki\/History_of_Python#Early_history\">was released<\/a> for the first time the\nfollowing year.<\/p><p>In 1992 the X3J13 committee published the first draft of the new Common Lisp\nANSI standard for public review (see Pitman\u2019s paper).  The draft was approved in\n1994 and the approved specification was finally published in 1995.  At this\npoint Lisp was over thirty-five years old.  The first version of Ruby <a href=\"https:\/\/en.wikipedia.org\/wiki\/Ruby_(programming_language)#First_publication\">was\nreleased<\/a> in December of that year.<\/p><p>That\u2019s the end of the history lesson.  There has not been another revision of\nthe ANSI specification of Common Lisp.  The version published in 1995 is the one\nthat is still used today \u2014 if you see something calling itself \u201can\nimplementation of Common Lisp\u201d today, that is the specification it\u2019s referring\nto.<\/p><h3 id=\"consequences\">Consequences<\/h3><p>I wanted to give you a quick overview of the history of Common Lisp because I\nwant you to know what you\u2019re getting yourself into.  I want you to realize that\nCommon Lisp is a stable, large, practical, extensible, ugly language.\nUnderstanding these characteristics will make a lot of things make more sense\nas you learn the language, and I want to talk a little bit more about each of\nthem before I start offering recommendations.<\/p><h4 id=\"escaping-the-hamster-wheel-of-backwards-incompatibility\">Escaping the Hamster Wheel of Backwards Incompatibility<\/h4><p>If you\u2019re coming from other languages, you\u2019re probably used to things breaking\nwhen you \u201cupgrade\u201d your language implementation and\/or libraries.  If you want\nto run Ruby code you wrote ten years ago on the latest version of Ruby, it\u2019s\nprobably going to take some effort to update it.  My current day job is in Scala,\nand if a library\u2019s last activity is more than 2 or 3 years old on Github I just\nassume it won\u2019t work without a significant amount of screwing around on my part.\nThe Hamster Wheel of Backwards Incompatibility we deal with every day is a fact\nof life in most modern languages, though some are certainly better than others.<\/p><p>If you learn Common Lisp, this is usually not the case.  In the next section of\nthis post I\u2019ll be recommending a book written in 1990.  You can run its code,\nunchanged, in a Common Lisp implementation released last month.  After years of\njogging on the Hamster Wheel of Backwards Incompatibility I cannot tell you how\nmuch of a <em>relief<\/em> it is to be able to write code and reasonably expect it to\nstill work in twenty years.<\/p><p>Of course, this is only the case for the language itself \u2014 if you depend on any\nlibraries there\u2019s always the chance they might break when you update them.  But\nI\u2019ve found the stability of the core language is contagious, and overall the\nCommon Lisp community seems fairly good about maintaining backwards\ncompatibility.<\/p><p>I\u2019ll be honest though: there are exceptions.  As you learn the language and\nstart using libraries you\u2019ll start noticing some library authors who don\u2019t\nbother to document and preserve stable APIs for their libraries, and if\nstaying off the Hamster Wheel is important to you you\u2019ll learn to avoid relying\non code written by those people as much as possible.<\/p><h4 id=\"practicality-begets-purity\">Practicality Begets Purity<\/h4><p>Another thing to understand about Common Lisp is that it\u2019s a large, practical\nlanguage.  The second edition of Common Lisp: the Language (usually abbreviated\nas \u201cCLtL2\u201d by Common Lisp programmers) is 971 pages long, not including the\npreface, references, or index.  You can get a surprising amount done by writing\npure Common Lisp without much extra support.<\/p><p>When programming applications in Common Lisp people will often depend on\na small(ish) number of stable libraries, and library writers often try to\nminimize dependencies by utilizing as much of the core language as possible.\nI try to stick to fewer than ten or so dependencies for my applications and no\nmore than two or three for my libraries (preferably zero, if possible), but I\u2019m\nprobably a bit more conservative than most folks.  I <em>really<\/em> don\u2019t like the\nHamster Wheel.<\/p><p>It\u2019s also worth noting that since Common Lisp has been around and stable for so\nlong, it has <em>libraries<\/em> older and more stable than many programming languages.\nFor example: Bordeaux Threads (the de-facto threading library for Common Lisp)\nwas first proposed in 2004 and released soon after (2006 at the latest but\npossibly earlier, it\u2019s hard to tell because so many links are dead now), which\nmakes it about fourteen years old.  So yes, threading is handled by a library,\nbut I\u2019m not worried about it breaking my code in the next decade or two.<\/p><p>My advice is this: as you learn Common Lisp and look for libraries, try to\nsuppress the voice in the back of your head that says \u201cThis project was last\nupdated six years ago?  That\u2019s probably abandoned and broken.\u201d  The stability of\nCommon Lisp means that sometimes libraries can just be <em>done<\/em>, not <em>abandoned<\/em>,\nso don\u2019t dismiss them out of hand.<\/p><h4 id=\"extensibility\">Extensibility<\/h4><p>Part of Common Lisp\u2019s practicality comes from its extensibility.  No one has\nbeen clamoring for a new version of the specification that adds features\nbecause Common Lisp\u2019s extensibility allows users to add new features to the\nlanguage as plain old libraries, without having to alter the core language.\nMacros are what might come to mind when you hear \u201cLisp extensibility\u201d, and of\ncourse that\u2019s part of it.  Macros allow users to write libraries that would\nneed to be core language features in other languages.<\/p><p>Common Lisp doesn\u2019t include string interpolation.  You want it?  No problem, you\ndon\u2019t have to wait for <a href=\"https:\/\/docs.scala-lang.org\/overviews\/core\/string-interpolation.html\">Scala\n2.10<\/a> or\n<a href=\"https:\/\/www.python.org\/dev\/peps\/pep-0498\/\">Python 3.6<\/a>, just <a href=\"https:\/\/edicl.github.io\/cl-interpol\/\">use\na library<\/a>.<\/p><p>Want to try some nondeterministic programming without any boilerplate?  <a href=\"https:\/\/nikodemus.github.io\/screamer\/\">Grab\na library<\/a>.<\/p><p>Pattern matching syntax can make for some really beautiful, readable code.\nCommon Lisp doesn\u2019t include it, but of course <a href=\"https:\/\/github.com\/guicho271828\/trivia\/wiki\/What-is-pattern-matching%3F-Benefits%3F\">there\u2019s a library<\/a>.<\/p><p>Enjoying algebraic data types in Haskell or Scala?  Here\u2019s your\n<a href=\"https:\/\/github.com\/tarballs-are-good\/cl-algebraic-data-type\">library<\/a>.<\/p><p>All of these libraries rely on macros to make using them feel seamless.  Of\ncourse you could <em>do<\/em> all of that without macros, but you\u2019ve have to add a layer\nof boilerplate to manage evaluation.  This:<\/p><pre>(match foo\n  '(list x y z) (lambda (x y z) (+ x y z))\n  '(vector x y) (lambda (x y) (- x y)))\n<\/pre><p>just doesn\u2019t flow off the fingers like:<\/p><pre>(match foo\n  ((list x y z) (+ x y z))\n  ((vector x y) (- x y)))\n<\/pre><p>No one\u2019s up in arms trying to get a new revision of the Common Lisp standard to\nadd pattern matching because you can write it as a library and get 90% or more\nof what you\u2019ve get if it were built in.  The language gives you enough power to\nextend it in a way that feels like the extension was there from the beginning.<\/p><p>Having things that are core features in other languages be provided by libraries\nmight seem at odds with the previous section about minimizing dependencies, and\nto some extent that\u2019s true.  But I think there\u2019s a happy medium where you can\nwrite stable libraries in the core language and then depend on a small number of\nthose libraries in your applications to add exactly the features you need for\nany particular problem.<\/p><h4 id=\"power\">Power<\/h4><p>Macros are one of the things that make Lisp so extensible, because they let you\ntransform arbitrary code into other arbitrary code.  This is true for macros in\nlanguages like C too, but Common Lisp macros are different because they\u2019re <em>part\nof the language<\/em>.<\/p><p>In C you have a layer of macros on top, written in a preprocessor macro\nlanguage.  The macro layer and the language layer are separate from each other,\nwith the macro layer providing one one extra level of abstractive power (which,\ndon\u2019t get me wrong, is certainly useful).<\/p><p>In Common Lisp, you write macros <em>in Common Lisp itself<\/em>.  You can then use\nthose macros to write functions, and use those functions to write more macros.\nInstead of two stratified layers it\u2019s a <em>feedback loop<\/em> of abstractive power.<\/p><p>But macros aren\u2019t the only thing about Common Lisp that make it so practical and\nextensible.  Something people often don\u2019t realize is that while Common Lisp is\nan extremely high-level language thanks to macros, it also has plenty of\nlow-level facilities as part of the language.  It\u2019s never going to be as\nlow-level as something like C, Rust, or Forth, but you might be surprised at\nsome of the things that the ANSI spec includes.<\/p><p>Want to see the assembly code a particular function compiles down to?\n<a href=\"http:\/\/clhs.lisp.se\/Body\/f_disass.htm\"><code>DISASSEMBLE<\/code><\/a> it!<\/p><p>Want to stack-allocate something to avoid some garbage collection?  X3J13\n<a href=\"http:\/\/clhs.lisp.se\/Body\/d_dynami.htm\">thought of that<\/a>.<\/p><p>Need arrays of unboxed floats to ship to a graphics card?  <a href=\"http:\/\/clhs.lisp.se\/Body\/15_ab.htm\">The standard allows\nfor that<\/a>.<\/p><p>Think <code>GOTO<\/code> should be considered helpful, not harmful?  Well, okay, we\u2019re all\nadults here.  <a href=\"http:\/\/clhs.lisp.se\/Body\/s_tagbod.htm\">Good luck<\/a>, try not to shoot your foot off.<\/p><p>Need to do unsigned 8-bit arithmetic in your Game Boy emulator, but would prefer\nit to compile down to just a machine instruction or two?  <a href=\"https:\/\/pdfs.semanticscholar.org\/7089\/0eff0e93aba49174a9346731f4bf9225706d.pdf\">It\u2019s\npossible<\/a>.<\/p><p>Not all Common Lisp implementations actually perform all these optimizations,\nbut the designers of Common Lisp had the foresight to include the language\nfeatures needed to support them.  You can write vanilla Common Lisp as defined\nby the standard and trust that it will run everywhere, and implementations that\n<em>do<\/em> support these kinds of things will take advantage of the optimization\nopportunities.<\/p><p>This combination of supporting extremely high-level programming with macros and\na reasonable amount of low-level optimization mean that even though the\nspecification is over twenty years old, it\u2019s still a good solid base to build on\ntoday.  The thirty years of experience and history the designers were drawing\nfrom allowed them to create a very practical language that has survived for\ndecades.<\/p><h4 id=\"ugliness\">Ugliness<\/h4><p>It\u2019s also important to realize that while Common Lisp might be very practical,\nthe need to accommodate existing users and dialects means that there are plenty\nof ugly parts.  If you buy a paper copy of the second edition of Common Lisp:\nthe Language and look up \u201ckludges\u201d in the index you\u2019ll find this:<\/p><p><a href=\"http:\/\/stevelosh.com\/media\/images\/blog\/2018\/07\/lisp-kludge.jpeg\"><img src=\"http:\/\/stevelosh.com\/media\/images\/blog\/2018\/07\/lisp-kludge.jpeg\" alt=\"Photo of a page of CLtL2's Index, listing &quot;kludges&quot; as pages 1 to 971\" \/><\/a><\/p><p>Common Lisp is not a beautiful crystal of programming language design.  It\u2019s\na scruffy workshop with a big pegboard wall of tools, a thin layer of sawdust on\nthe floor, a filing cabinet in the office with a couple of drawers that open\nperpendicular to the rest, and there\u2019s a weird looking saw with <code>RPLACD<\/code> written\non the side sitting off in a corner where no one\u2019s touched it for twenty years.<\/p><p>This historical baggage is a price paid to ensure Common Lisp had a future.  It\nmade it practical for people using the old dialects to actually adopt Common\nLisp with a reasonable amount of effort.  If the designers had tried to make it\nperfect and beautiful this could have made it too different to port\nimplementations and code to and might have resulted in the language being\nignored, instead of being adopted and embraced.<\/p><p>If all of this hasn\u2019t scared you away from the language, let\u2019s talk about how\nyou can learn it in 2018.<\/p><p>If you search around on the internet for Common Lisp tutorials and guides,\nyou\u2019re not going to find as much as you might expect.  This is because a lot of\nCommon Lisp reference material was created before or during the infancy of the\ninternet.  There are a <em>lot<\/em> of books about Common Lisp out there.  Some are\nbetter than others.  I\u2019ll recommend the ones I think are the best, but don\u2019t\nhesitate to browse around and find others.<\/p><h3 id=\"get-a-lisp\">Get a Lisp<\/h3><p>To get started with Common Lisp you\u2019ll need to install a Common Lisp\nimplementation.  Common Lisp is an ANSI specification, so there are multiple\nimplementations of it, which gives you choices.  There are a bunch of options,\nbut I\u2019ll make it simple for you:<\/p><ul><li>If you\u2019re using MacOS and want a single GUI app you can download from the App\nStore, choose <a href=\"https:\/\/ccl.clozure.com\/\">ClozureCL<\/a> (often abbreviated \u201cCCL\u201d).<\/li>\n<li>Otherwise, choose <a href=\"http:\/\/www.sbcl.org\/\">SBCL<\/a>.<\/li>\n<\/ul><p>That\u2019s Clozure with a Z.  Clojure is something entirely different that just\nhappens to have a confusingly similar name.<\/p><p>You might also hear of something called CLISP, which sounds like it might be\nwhat you want.  It\u2019s not.  CLISP is just another implementation, but it hasn\u2019t\nhad a release in eight years (even though development is still ongoing in its\nsource repos!) and it\u2019s not as commonly used as CCL or SBCL, so it\u2019ll be harder\nto find help if you have questions about the installation, etc.<\/p><p>You might also hear about something called Roswell.  Don\u2019t use Roswell, you don\u2019t\nneed it (yet (or at all)).<\/p><p>Just install SBCL or CCL for now, you can explore the other options once you\u2019ve\ngot your bearings a bit better.<\/p><h3 id=\"pick-an-editor\">Pick an Editor<\/h3><p>You might hear people tell you that you <em>must<\/em> learn Emacs before learning\nCommon Lisp.  They\u2019re wrong.  You can get started learning the language just\nfine in whatever text editor you\u2019re comfortable in.<\/p><p>If you don\u2019t have a preference, CCL itself comes bundled with a text editor on\nMacOS.  That one will work just fine to start.<\/p><p>Emacs, Vim, Sublime Text, Atom, whatever, for now it doesn\u2019t matter.  As long as\nit can balance parentheses, highlight comments and strings, and autoindent Lisp\ncode that\u2019s all you need to start.  Worry about shaving the editor yak once\nyou\u2019re more comfortable in the language.<\/p><h3 id=\"hello-lisp\">Hello, Lisp<\/h3><p>To check that you\u2019ve got everything set up properly, make a <code>hello.lisp<\/code> file\nwith the following contents:<\/p><div class=\"highlight\"><pre class=\"language-lisp\" data-lang=\"lisp\">(defun hello ()\n  (write-line \"What is your name?\")\n  (let ((name (read-line)))\n    (format t \"Hello, ~A.~%\" name)))<\/pre><\/div><p>Don\u2019t worry about what this means yet, it\u2019s just a check that everything\u2019s\nworking properly.<\/p><p>Open an SBCL or CCL REPL (<a href=\"https:\/\/en.wikipedia.org\/wiki\/Read%E2%80%93eval%E2%80%93print_loop\">Read\/Eval\/Print\nLoop<\/a>) and\nload the file by entering <code>(load \"hello.lisp\")<\/code>, then call the function and make\nsure it works.  It should look something like this if you picked SBCL:<\/p><pre>$ sbcl\n* (load \"hello.lisp\")\nT\n* (hello)\nWhat is your name?\nSteve\nHello, Steve.\nNIL\n*\n<\/pre><p>Or if you chose CCL but still want to use the command line, rather than the\nMacOS app (the command line program might be annoyingly named <code>ccl64<\/code> if you\u2019re\non a 64-bit system):<\/p><pre>$ ccl64\nClozure Common Lisp Version ...\n? (load \"hello.lisp\")\n#P\"\/home\/sjl\/Desktop\/hello.lisp\"\n? (hello)\nWhat is your name?\nSteve\nHello, Steve.\nNIL\n?\n<\/pre><p>If your arrow keys and backspace don\u2019t work in the REPL, use <a href=\"https:\/\/github.com\/hanslub42\/rlwrap\"><code>rlwrap<\/code><\/a>\nto fix that.  <code>rlwrap sbcl<\/code> will give you a non-miserable REPL.  <code>rlwrap<\/code> is\na handy tool to have in your toolbox anyway.<\/p><p>The best book I\u2019ve found for getting started in Common Lisp is <a href=\"https:\/\/www.cs.cmu.edu\/~dst\/LispBook\/\">Common Lisp:\nA Gentle Introduction to Symbolic Computation<\/a>.  This book really\ndoes strive to be gentle.  Even if you\u2019ve programmed before I\u2019d still recommend\nstarting here because it eases you into the language.<\/p><p>The 1990 edition is available free from the site, and there\u2019s a 2013 reprint\nwhich fixes some minor errors in the 1990 version.  If you can afford it I\u2019d\nrecommend buying the 2013 edition, but the 1990 version will also do fine.<\/p><p>Go through the book and <em>do all the exercises<\/em>.  This will take a while, and\nis mainly meant to get you started overcoming some of the main obstacles to\nbeing comfortable in Common Lisp, such as:<\/p><ul><li>How am I ever going to remember all these weird function names?<\/li>\n<li>Why do people use strings so rarely?<\/li>\n<li>When do I need the god damn quotation mark?<\/li>\n<\/ul><p>If you find the book is moving too slow, just skim forward a bit.  Skimming is\na very useful skill to practice as a programmer.  I think it\u2019s better for\nauthors to err on the side of explaining too much when writing books and\ndocumentation \u2014 expert readers should be comfortable skimming if you explain too\n<em>much<\/em>, but new users will be stuck wallowing in confusion if you\u2019re too terse.\nCreating hours of newbie misery and confusion to save a few flicks of an\nexpert\u2019s scroll wheel is a poor tradeoff to make.<\/p><p>You should also join the <code>#clschool<\/code> channel on the Freenode IRC network so you\ncan ask questions if you get stuck.  For the most part people there are friendly\nand helpful, though I\u2019ll warn you in advance that there\u2019s at least one person\nwho can sometimes be abrasive.  There\u2019s also a <code>#clnoobs<\/code> channel, but that was\nmostly abandoned during the latest wave of Freenode spam because no one had ops\nto help combat the spam.<\/p><p>If IRC isn\u2019t your thing there\u2019s also a <a href=\"https:\/\/discord.gg\/tffeu2x\">Discord\nserver<\/a> that some of us hang out in.  Join the\n<code>#common-lisp<\/code> channel there and we\u2019ll be happy to help you.<\/p><h3 id=\"getting-practical\">Getting Practical<\/h3><p>Once you\u2019ve finished that book the next one you should attack is <a href=\"http:\/\/www.gigamonkeys.com\/book\/\">Practical\nCommon Lisp<\/a>.  You can get a paper copy if you want, but the full book\nis available on the site for free.<\/p><p>You can skip the editor\/programming environment part because the environment it\nrecommends (Lisp in a Box) is abandoned and no longer works.  Just keep using\nthe programming environment you\u2019re comfortable with for now.<\/p><p>Unfortunately the book doesn\u2019t include exercises.  If you <em>really<\/em> want to get\nthe most out of it you can type in all the code as you\u2019re reading it and poke at\nit, but if you\u2019ve already done the exercises in the previous book it\u2019s probably\nsafe to just sit down and read the book carefully.  Don\u2019t read more than\na chapter or two a day.  It will take a while for your brain to digest all the\ninformation.<\/p><p>Make sure you understand everything as you go through the book.  Don\u2019t be afraid\nto ask questions on IRC or Discord (or email me if you want, I don\u2019t mind) if\nsomething\u2019s not clear.<\/p><p>You should also begin to get comfortable looking up things in <a href=\"http:\/\/www.lispworks.com\/documentation\/lw70\/CLHS\/Front\/Contents.htm\">the Common Lisp\nlanguage specification<\/a> itself.  It\u2019s the ultimate manual for Common Lisp.\nIt can be pretty dense at points, but can answer many questions you might have\nif you read it slowly and carefully.  You can either use the index page to find\nwhat you\u2019re looking for or just search on Google for \u201cclhs whatever\u201d (CLHS\nstands for \u201cCommon Lisp HyperSpec\u201d, which is the hyperlinked, HTML version of\nthe spec).  If you already use the Dash app for MacOS, it has the Common Lisp\nspec available.<\/p><p>(Some people will tell you to learn the language by just reading the spec.  That\u2019s\nridiculous \u2014 it\u2019s like trying to learn French by reading a dictionary.  It\u2019s\na useful tool to have, but not the only one you\u2019ll need.)<\/p><h3 id=\"make-something\">Make Something<\/h3><p>Once you\u2019ve got those two books under your belt and some practice using the\nspec, it\u2019s time to make something without someone holding your hand.  It doesn\u2019t\nhave to be anything big or special, the goal is to just write some Lisp without\nhaving the answer on the next page.<\/p><p>If you need some ideas:<\/p><ul><li>Do some <a href=\"https:\/\/projecteuler.net\/\">Project Euler<\/a> problems.<\/li>\n<li>Do some <a href=\"https:\/\/adventofcode.com\/\">Advent of Code<\/a> exercises.<\/li>\n<li>Make a <a href=\"https:\/\/twitter.com\/git_commands\">stupid Twitter bot<\/a>.<\/li>\n<li>Make a personal calendar program that records your appointments, checks the\nweather forecast the day of, etc.<\/li>\n<li>Use <a href=\"https:\/\/github.com\/vydd\/sketch\">Sketch<\/a> to implement the stuff in some <a href=\"https:\/\/www.youtube.com\/user\/codingmath\/videos\">Coding Math videos<\/a>.<\/li>\n<\/ul><p>It doesn\u2019t really matter what you make, just make <em>something<\/em> on your own.<\/p><h3 id=\"lisp-as-a-system\">Lisp as a System<\/h3><p>At this point it\u2019s time to take your Common Lisp skills up a notch.  Up until\nnow I\u2019ve told you to just use any text editor because it\u2019s more important to\nget you some experience with the language, but now it\u2019s time to dive into the\ndeep end.<\/p><p>In most languages the development process looks something like this:<\/p><ol><li>Edit some code in the project with an editor.<\/li>\n<li>Compile the project (some languages skip this step).<\/li>\n<li>Run the project (or the tests).<\/li>\n<li>Observe the output (in the console, a browser, etc).<\/li>\n<li>Go to 1.<\/li>\n<\/ol><p>This is not how most Common Lisp users interact with the language.  In Common\nLisp, the development cycle looks more like this:<\/p><ol><li>Start a Lisp process.<\/li>\n<li>Load the project.<\/li>\n<li>Edit some code with your editor.<\/li>\n<li>Tell the running process to compile <em>only the code you edited<\/em>.<\/li>\n<li>Interact with the changed code in the process via the REPL, an HTTP request, etc.<\/li>\n<li>Observe the output (in the console, a browser, etc).<\/li>\n<li>Go to 3.<\/li>\n<\/ol><p>When you embrace the Lisp way of working you\u2019ll rarely recompile and reload an\nentire project.  Usually you\u2019ll write a function (or a macro, or parameter, or\nwhatever), compile <em>just that function<\/em>, maybe poke at it in the REPL a bit, and\nthen move on to the next function.  This has some advantages over the\ntraditional compile-everything-then-run approach.<\/p><p>First: compiling a small chunk of code is fast.  I just timed compiling a few of\nthe larger functions in one of my projects and they took around 50-80\nmicroseconds.  You don\u2019t have to wait for the compiler, so your\nconcentration\/thought process never has time to wander.<\/p><p>Another advantage is that when you get back the results of your compilation (and\nrunning), any errors or warnings you receive are almost certainly related to the\nfew lines of code you just compiled.  If you compile a ten-line function, run\nit, and get a division by zero error you can immediately focus in on the ten\nlines you just compiled and think about what changed.<\/p><p>Because the Lisp process is always running, as soon as you compile a function\nit\u2019s ready to be used in the REPL.  You can throw some arbitrary data at it and\ninspect the results to see how it behaves in isolation before you build more\nthings on top of it.  This cycle of making a function, compiling it, poking at\nit to make sure it\u2019s working as expected, and moving on happens <em>constantly<\/em>.<\/p><p>In contrast, when working in languages like Scala or Python I almost never find\nmyself writing one single function and compiling or running the project\nimmediately.  Spinning up the compiler or running the unit tests takes at\n<em>least<\/em> a second or two (or sometimes <em>minutes<\/em> in Scala, unfortunately) so to\navoid having a constant stream of gaps in my thought I end up writing a bunch of\nfunctions at once, and then I run the project or tests once I know they have\na chance of working.<\/p><p>But then when I get back an error I have much more surface area to check,\nbecause I\u2019ve added a lot of new code!  So now I have to track down a problem\nthat might be in something I wrote four minutes ago, whereas in Lisp I would\nonly have to ever look at the code I wrote in the last few seconds.<\/p><p>I\u2019ve started using IntelliJ with Scala to help make this a bit less painful.  It\ndoes help with the compile times because it recompiles things on the fly, but it\ndoesn\u2019t solve the rest of the problem.  I can write a Scala function in IntelliJ\nand it will be compiled immediately, but I can\u2019t <em>interact<\/em> with it immediately\nlike I can in Common Lisp.<\/p><p>When you work in this style with Common Lisp I think you\u2019ll really grow to love\nit.  Writing in other languages will begin to feel like shipping your code off\nto the DMV and getting it back a week later with a page full of red ink\nsomewhere in the hundred forms you filled out.  Writing in Common Lisp feels\nlike interacting with a living, breathing organism, or like <a href=\"https:\/\/www.reddit.com\/r\/lisp\/comments\/4oo1cp\/common_lisp_for_clojure_programmer\/d4eec68\/\">teaching things to\nan eager assistant<\/a>.<\/p><p>This philosophy of Lisp being not just a programming <em>language<\/em> but a living,\nbreathing programming <a href=\"https:\/\/www.dreamsongs.com\/Files\/Incommensurability.pdf\"><em>system<\/em><\/a> goes beyond just the short\nfeedback loop and interactive REPL, too.<\/p><p>As an example: imagine you\u2019re making a video game and have a bug somewhere in\nyour damage calculation that will occasionally cause a division by zero.  Now\nlet\u2019s say you\u2019re working on the code for a particular quest.  You\u2019ll start the\ngame, load a save file at the beginning of the quest, and start going through\nthe steps.  All of a sudden, in the middle of killing the final monster for the\nquest, you hit the damage bug!  In traditional languages, one of two things\nmight happen:<\/p><ol><li>The game crashes, and you get a stack trace and maybe a core dump.<\/li>\n<li>You\u2019ve wrapped a <code>try<\/code> block around the main game loop that logs a stack\ntrace and ignores errors and allows the game to continue.<\/li>\n<\/ol><p>Case 1 is pretty bad.  You\u2019ve got to try to track down the bug from a snapshot\nof what things looked like at the time (the stack trace and core dump).  And\neven if you manage to fix it, now you\u2019ve got to redo all that playing to get\nback to testing your quest code that you were originally working on.<\/p><p>Case 2 is bad, in a different way.  If you just ignore errors all the time, the\ngame might now be in a weird state.  You also might lose some critical context\nthat\u2019s necessary to debug the problem, unless you\u2019re also saving a core dump\n(but I don\u2019t know of many people who save a core dump on every exception).<\/p><p>In Common Lisp you can certainly choose to panic on or ignore errors, but\nthere\u2019s a better way to work.  When an error is signaled in Common Lisp, it\ndoesn\u2019t unwind the stack.  The Lisp process will pause execution at that point\nand open a window in your editor showing you the stack trace.  Your warrior\u2019s\nsword is hovering over the monster, waiting for you.  At this point you can\ncommunicate with the running process at the REPL to see what\u2019s going on.  You\ncan examine variables in the stack, or even run any arbitrary code you want.<\/p><p>Once you figure out the problem (\u201cOh, I see, the <code>calculate-armor-percentage<\/code>\nfunction returns <code>0<\/code> if a shielding spell ran out during the same frame\u201d) you\ncan fix the code, recompile the problematic function, and <em>restart the execution\nof that function (or any other one!) in the call stack<\/em>!  Your warrior\u2019s sword\nlands, and you move back to what you were doing before.<\/p><p>You don\u2019t have to track down the bug from just a stack trace, like a detective\ntrying to piece together what happened by the <a href=\"https:\/\/www.usenix.org\/system\/files\/1311_05-08_mickens.pdf\">blood stains<\/a> on\nthe wall.  You can examine the crime <em>as it\u2019s happening<\/em> and intervene to save\nthe victim.  It\u2019s like if you could run your code in a debugger with\na breakpoint at every single line that only activates if something goes wrong!<\/p><p>Maybe you don\u2019t make video games, sure, but this process can be useful in all\nkinds of contexts.  Maybe you\u2019re writing a web app that talks to an API\nsomewhere, and are debugging a request that fails between two calls to the API,\ne.g.  between \u201ccreate widget <code>foo<\/code>\u201d and \u201cadd <code>foo<\/code> to widget list <code>bar<\/code>\u201d.\nInstead of just aborting the request, logging a stack trace, and now leaving\nthings in a possibly weird state (<code>foo<\/code> having been created without being in the\nexpected <code>bar<\/code> list), you can fix the problem and allow the request to finish\nproperly.<\/p><p>Of course this won\u2019t always work.  If you\u2019ve got a big function that does some\nside effects and then crashes, restarting execution of the function would make\nthe side effects happen again.  But if you divide up your functions well (<a href=\"https:\/\/groups.google.com\/forum\/message\/raw?msg=comp.lang.lisp\/9SKZ5YJUmBg\/Fj05OZQomzIJ\">one\nfunction to a function!<\/a>) this case is pretty rare.  And even when it\ndoes happen, it just means you\u2019re back in the same situation you\u2019re in <em>by\ndefault<\/em> with other languages!<\/p><p>Support for this style of interactive development doesn\u2019t just come from some\nfancy editor plugins \u2014 it\u2019s baked into the bones of the language.  For example:\nthe standard specifies a method named\n<a href=\"http:\/\/clhs.lisp.se\/Body\/f_upda_1.htm\"><code>update-instance-for-redefined-class<\/code><\/a>\nthat lets you customize what happens to objects when their class is redefined!\nThis isn\u2019t something you\u2019ll use all the time, but something like Sketch (a\nCommon Lisp equivalent of Java\u2019s Processing library) <a href=\"https:\/\/github.com\/vydd\/sketch\/blob\/19fe20502d26fc75752e710dfe5106ed13345c5a\/src\/sketch.lisp#L118-L121\">uses it<\/a> to\nautomatically update the running sketch when you redefine its class.\nDynamically updating running code in a safe, consistent way doesn\u2019t require any\ndark magic in Common Lisp because it\u2019s the expected, usual way to work.<\/p><p>So how do you actually <em>get<\/em> this wonderful interactive experience?  The bad\nnews is that you\u2019re going to need to shave the editor yak.  You really only have\ntwo choices here:<\/p><ul><li>Emacs with <a href=\"https:\/\/common-lisp.net\/project\/slime\/\">SLIME<\/a> or <a href=\"https:\/\/github.com\/joaotavora\/sly\">Sly<\/a>.<\/li>\n<li>Vim (or Neovim) with <a href=\"https:\/\/github.com\/l04m33\/vlime\">Vlime<\/a> or <a href=\"https:\/\/github.com\/kovisoft\/slimv\">Slimv<\/a>.<\/li>\n<\/ul><p>I wish this weren\u2019t the case, but those are really only the realistic options\ntoday (aside from the editing environments for the (expensive) commercial\nLisps).<\/p><p>If you\u2019re like me and already have Vim burned too deeply into your fingers to\never get it out, I\u2019d recommend Vim with Vlime.  It will give you 80% of the\nexperience you\u2019ll get with Emacs.<\/p><p>Otherwise go with Emacs.  You might want to look into\n<a href=\"https:\/\/portacle.github.io\/\">Portacle<\/a>, which bundles Emacs and SLIME and\na bunch of other things together, or you might want to have a go at setting up\nEmacs and SLIME or Sly yourself.  I can\u2019t really give you much advice on the\nEmacs side of things because I haven\u2019t had much experience with it, so you\u2019ll\nneed to do a bit of research here.<\/p><p>Whatever you choose, spend some time setting up your editor and environment of\nchoice.  This will be a lot of fiddly metawork, but will pay off handsomely as\nyou continue working in Lisp.<\/p><p>On a side note: if anyone is interested in making a Common Lisp <a href=\"https:\/\/langserver.org\/\">LSP<\/a>\nlanguage server, I think it would be a hugely useful contribution to the\ncommunity.  Having an LSP server would mean you could get a much nicer\nprogramming experience in many editors out of the box, which would help new\npeople quite a lot.<\/p><p>I think you could piggyback on top of Swank to do a lot of the language-side\nstuff, and it would mostly be a matter of implementing the LSP interface.  If\nthis sounds interesting to you, please let me know \u2014 I\u2019d be willing to help.\nI\u2019ve done some work at my day job making a Scala LSP language server that uses\nIntelliJ as a backend, so I have at least some idea of how that sausage gets\nmade.  I just don\u2019t have the time or motivation to do an entire LSP server for\nCommon Lisp all by myself.<\/p><h3 id=\"learning-paradigms\">Learning Paradigms<\/h3><p>At this point you should have a pretty good handle on the basics of Common Lisp,\nand have set up one of the more powerful development environments.  Your next\ngoals should be to learn how to write idiomatic Common Lisp and to get some\npractice using your fancy new environment.<\/p><p>I think the perfect book for both of these is <a href=\"https:\/\/github.com\/norvig\/paip-lisp\">Paradigms of Artificial\nIntelligence Programming<\/a>, often abbreviated as PAIP.  The book was\nrecently made available for free as a PDF, or you can buy a used paper copy if\nyou prefer.<\/p><p>This book was written in 1992 so it\u2019s not about the hyped up AI fields you\u2019ve\nbeen hearing about in the news like machine learning \u2014 instead it\u2019s a tour of\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Symbolic_artificial_intelligence\">Good Old-Fashioned AI<\/a>.  Even if you\u2019re not particularly interested in\nthis kind of AI, the book is a great example of how to write Common Lisp code.<\/p><p>One thing I really love about this book is that almost all the functions in it\nhave docstrings.  If you look at most other programming books they omit the\ndocumentation strings, presumably for space reasons and because they feel the\nsurrounding text is documentation enough.  But writing helpful docstrings is an\nart in and of itself, and I think books that omit them train readers that \u201cgood\ncode omits docstrings\u201d, which is a bad habit to get into.<\/p><p>The book contains <em>plenty<\/em> of exercises, conveniently categorized by how\ndifficult or involved they are:<\/p><ul><li><code>S<\/code> for \u201cseconds\u201d.<\/li>\n<li><code>M<\/code> for \u201cminutes\u201d.<\/li>\n<li><code>H<\/code> for \u201chours\u201d.<\/li>\n<li><code>D<\/code> for \u201cdays\u201d.<\/li>\n<\/ul><p>This is a very good idea which more books should steal.  Do all of the <code>S<\/code> and\n<code>M<\/code> exercises, and try your hand at at least a few of the <code>H<\/code> ones.  If a <code>D<\/code>\nsounds particularly interesting don\u2019t be afraid to spend some time on it\n\u2014 really digging into a problem is exactly what you need at this point in your\nLisp journey.<\/p><h3 id=\"switch-things-up\">Switch Things Up<\/h3><p>Now that you\u2019re comfortable in Common Lisp and your programming environment,\nit\u2019s time to push yourself out of your comfort zone again.  At the beginning I\nhad you choose either SBCL or CCL.  Now I want you to install whichever one you\ndidn\u2019t originally choose and make sure all the code you\u2019ve written so far runs\nin it.<\/p><p>This may seem a bit like running in place, but making sure your code runs in\nmore than one implementation will keep you honest.  It will force you to write\nportable code that doesn\u2019t rely on anything implementation-specific that might\nchange in the next decade or two.  And you might even discover that you like\nthis other implementation better than the original \u2014 maybe CCL\u2019s super-fast\ncompile times make you smile, or SBCL\u2019s strong type inference catches more of\nyour bugs.<\/p><p>Go through all the code you\u2019ve written so far and make sure it all runs in the\nnew implementation.  You might also want to take this opportunity to refactor or\nrewrite some of it \u2014 you\u2019ve learned a lot since you first started, so your\nearliest Common Lisp code will probably look pretty rough to you now.<\/p><h3 id=\"recipes-for-success\">Recipes for Success<\/h3><p>The final technical book I\u2019ll recommend to every aspiring Lisp programmer is\n<a href=\"http:\/\/weitz.de\/cl-recipes\/\">Common Lisp Recipes<\/a>, sometimes abbreviated as CLR.  Unlike most of\nthe other books I\u2019ve recommended so far this one is relatively recent: it was\npublished in 2015.  It\u2019s not free, but I think it\u2019s well worth the money it\ncosts.<\/p><p>The book is written by the author of several very heavily used Common Lisp\nlibraries.  It\u2019s a bit of a grab bag of topics (which is why I think you need\na decent amount of Lisp under your belt before you tackle it) but it\u2019s a very\nwell-written grab bag that will teach you a lot of things you won\u2019t find in\nother books.<\/p><h3 id=\"final-patterns\">Final Patterns<\/h3><p>If you\u2019ve gotten this far you\u2019re pretty invested in Common Lisp, and I want to\nrecommend one not-strictly-technical book that I think you\u2019ll really enjoy:\nPatterns of Software by Richard Gabriel.  It\u2019s available as a PDF on <a href=\"https:\/\/www.dreamsongs.com\/Books.html\">the\nauthor\u2019s site<\/a>, and you can still find used print copies online if you\nprefer.<\/p><p>This is <em>not<\/em> the \u201cGang of Four\u201d\/\u201cDesign Patterns\u201d book that you might have\nalready read or heard about, but is a set of essays on a variety of\nloosely-related topics.  It\u2019s the best book I\u2019ve read so far this year.  I don\u2019t\nwant to spoil anything in it for you, so I\u2019ll just say that I think you\u2019ll find\nit well worth your time.<\/p><h2 id=\"where-to-go-from-here\">Where to Go From Here<\/h2><p>If you made it through all the books and activities in the previous section:\ncongratulations, you\u2019re off to a great start!  Now that you\u2019ve got a decent\nhandle on the core language you can explore in many different directions,\ndepending on your interests.<\/p><h3 id=\"macros\">Macros<\/h3><p>If you want to learn the secrets of macros, you\u2019ll probably want to read and\nwork through <a href=\"http:\/\/www.paulgraham.com\/onlisp.html\">On Lisp<\/a> and <a href=\"https:\/\/letoverlambda.com\/\">Let Over Lambda<\/a> (in that order).<\/p><p>I\u2019ll say that you should take both books (<em>especially<\/em> the latter) with a large\ngrain of salt.  A lot of Common Lisp users don\u2019t agree with all of the arguments\nand style in these books, but I think they can still provide plenty of value if\nyou read them with a critical mind.<\/p><h3 id=\"object-oriented-programming-with-clos\">Object-Oriented Programming with CLOS<\/h3><p>Common Lisp has some very sophisticated support for Object-Oriented Programming\nthrough CLOS.  If you\u2019re like me and have bad memories of OOP from working in\na Java cube farm, I\u2019d urge you to give CLOS a fair chance to change your mind.<\/p><p>Start with <a href=\"https:\/\/www.amazon.com\/Object-Oriented-Programming-COMMON-LISP-Programmers\/dp\/0201175894\">Object-Oriented Programming in COMMON LISP: A Programmer\u2019s Guide to\nCLOS<\/a>.  It\u2019s a wonderfully-written, short and to-the-point book that will\ngive you a good overview of how CLOS is intended to be used.<\/p><p>If you really want to bend your mind, try <a href=\"https:\/\/www.amazon.com\/Art-Metaobject-Protocol-Gregor-Kiczales\/dp\/0262610744\">The Art of the Metaobject\nProtocol<\/a> (usually abbreviated as AMOP).  This book will probably take you\na couple of tries to get through.  Read it until you hit a mental wall, go work\non other things for a couple of months, and come back and try again.  Repeat\nthat process as many times as necessary.<\/p><h3 id=\"low-level-programming\">Low-Level Programming<\/h3><p>Low-level programming can mean a lot of different things, so I\u2019ll just mention\none possibility here.<\/p><p>If you\u2019re interested in writing emulators for old computers, I wrote <a href=\"http:\/\/stevelosh.com\/blog\/2016\/12\/chip8-cpu\/\">a series\nof posts<\/a> on making a <a href=\"https:\/\/en.wikipedia.org\/wiki\/CHIP-8\">CHIP-8<\/a> emulator in Common Lisp.\n<a href=\"https:\/\/github.com\/kingcons\/cl-6502\">cl-6502<\/a> is an emulator for the processor used in the NES (and lots of other\nthings) and has a really nice <a href=\"http:\/\/redlinernotes.com\/docs\/cl-6502.pdf\">literate programming<\/a> version\nthat\u2019s wonderful to read through.<\/p><h3 id=\"web-development\">Web Development<\/h3><p>Unfortunately I don\u2019t have too many suggestions for web development in Common\nLisp.  I\u2019ve made a conscious effort to avoid web development in the past five or\nso years, because it seems like the Hamster Wheel of Backwards Incompatibility\nhas become more of a Hamster Centrifuge in that field.<\/p><p>There is a <code>#lispweb<\/code> channel on Freenode and a <code>#webdev<\/code> channel in the Lisp\nDiscord, so if you have questions you could start by asking there.  Those\nchannels are a bit less populated than the other Lisp channels, so don\u2019t expect\nan answer immediately.<\/p><h3 id=\"game-development\">Game Development<\/h3><p>Common Lisp has a small but enthusiastic community of people who like making\ngames.  There\u2019s a <code>#lispgames<\/code> channel on Freenode and a <code>#gamedev<\/code> channel on\nthe Lisp Discord that you should join if you\u2019re interested.<\/p><p><a href=\"http:\/\/landoflisp.com\/\">Land of Lisp<\/a> is a fun book to go through.  The coding\nstyle in the book has some\u2026 \u201ceccentricities\u201d, which is why I don\u2019t recommend\nit as a first book on Lisp (e.g. using <code>ash<\/code> instead of <code>truncate<\/code> or <code>floor<\/code>\nfor integer division), but if you know the language and just want to get started\nmaking some simple games I think you\u2019ll enjoy working through it.<\/p><p>If you want an excuse to make a game in Lisp in a week, the Lisp Game Jam is\nsomething you can join.  It\u2019s usually held once or twice each year, so you\u2019ll\nhave to search around (or ask in <code>#lispgames<\/code>) to find out when the next one is.<\/p><p>Lisp doesn\u2019t have any engine as full-featured as Unity, but several people are\ncurrently working on making 3D game engines.  Ask around to see what people are\nusing these days.  Unfortunately a 3D game engine will generally need to\ninterface with the OS to render images and produce audio, and so can\u2019t be\nwritten in pure Common Lisp.  This means that some running on the Hamster Wheel\nof Backwards Incompatibility will be necessary to keep up with OS changes (e.g.\n<a href=\"https:\/\/www.macrumors.com\/2018\/06\/05\/apple-deprecates-opengl-opencl-gaming\/\">Apple deprecating OpenGL<\/a>).<\/p><p>If you\u2019re interested in old-school ASCII\/tile-based games, I\u2019ve personally done\nsome work with using <a href=\"https:\/\/github.com\/HiTECNOLOGYs\/cl-charms\">ncurses<\/a> and <a href=\"http:\/\/foo.wyrd.name\/en:bearlibterminal\">bearlibterminal<\/a> in Common Lisp.\nThere\u2019s something really fun about making a game people can play over telnet!\nFeel free to get in touch with me if you\u2019re interested in that kind of stuff and\nwant to know more.<\/p><h3 id=\"window-management\">Window Management<\/h3><p>If you\u2019re running Linux and like tinkering with your desktop environment,\n<a href=\"https:\/\/stumpwm.github.io\/\">StumpWM<\/a> is an X window manager written in Common Lisp.  I\u2019ve just recently\nswitched back to Linux so I\u2019ve only been using it for about two months, but it\u2019s\nreally pleasant to be able to customize my working environment with Common Lisp.<\/p><p>StumpWM has a small but friendly community \u2014 if you\u2019re looking for a non-trivial\nopen source Common Lisp project to contribute to, StumpWM would be a great\nchoice.<\/p><h3 id=\"unit-testing\">Unit Testing<\/h3><p>If you\u2019re coming from a modern language, especially one with a lot of\ntest-driven development advocates, you might be surprised at the lack of an\nemphasis on unit testing in Common Lisp.  I think one reason for this is that in\nsome languages a unit test is the simplest way to actually <em>run<\/em> a function,\nbut Lisp\u2019s interactive style of development gives you an even easier\nalternative: just <em>run the function<\/em> in the REPL!<\/p><p>Despite the lack of heavy unit testing in the community, there are almost as\nmany unit testing <em>frameworks<\/em> as there are Common Lisp programmers!  This is\nprobably because making a unit testing framework is so easy with a few macros.\nI love <a href=\"https:\/\/github.com\/lmj\/1am\">1am<\/a>, but there are <em>plenty<\/em> more to choose from.<\/p><p>Whichever one you choose, please make sure to be a good citizen and create\na separate ASDF system for your unit tests, so people can use your library\nwithout having to load Yet Another Testing Framework.<\/p><h3 id=\"more-implementations\">More Implementations<\/h3><p>I had you use SBCL and CCL because those are the most popular free Common Lisp\nimplementations today, but they aren\u2019t the only actively-developed ones out\nthere.  There\u2019s plenty of others you might want to explore:<\/p><ul><li><a href=\"https:\/\/common-lisp.net\/project\/armedbear\/\">ABCL<\/a> runs on the JVM.<\/li>\n<li><a href=\"https:\/\/common-lisp.net\/project\/ecl\/main.html\">ECL<\/a> can be embedded in a C program, and can translate Common Lisp code to C code.<\/li>\n<li><a href=\"https:\/\/github.com\/clasp-developers\/clasp\">CLASP<\/a> is still under development, but is an implementation designed to be easy to interoperate with C++.<\/li>\n<li><a href=\"http:\/\/www.lispworks.com\/\">Lispworks<\/a> and <a href=\"https:\/\/franz.com\/products\/allegro-common-lisp\/\">Allegro CL<\/a> are commercial implementations with a lot of extra features and support, but are not free.<\/li>\n<\/ul><p>(I omitted CLISP because I\u2019m mad at them for choosing a name that confuses the\nheck out of new people.  Hey, I warned you this post would contain Opinions\u2122.)<\/p><p>I tend to use SBCL for my own projects, but I make sure the units tests for all\nmy libraries run in SBCL, CCL, ABCL, and ECL.  This keeps me honest and gives me\na reasonable degree of confidence that I\u2019m writing portable code.<\/p><h2 id=\"modern-common-lisp\">Modern Common Lisp<\/h2><p>Common Lisp is old and stable, but that doesn\u2019t mean it\u2019s stagnant.  The\nlanguage gives you plenty of power to build on, and before I wrap this up I want\ngo over a couple of recent developments in the Common Lisp world that the older\nbooks you\u2019ve been learning from don\u2019t talk about.  I also want to clarify some\nthings that often trip up new people.<\/p><h3 id=\"structure\">Structure<\/h3><p>Common Lisp\u2019s terminology for various parts of projects is often confusing to\nnew people because it\u2019s old and uses a lot of words that we use now (like\n\u201cpackage\u201d) to mean subtly different things than people mean today. Things get\neasier once you internalize what Common Lisp means by the terms.<\/p><p>(Side note: I posted a quick-and-dirty version of this section as a <a href=\"https:\/\/lobste.rs\/s\/fwhuz5\/my_lisp_journey_1_getting_started_with#c_ebhvzq\">comment<\/a>\non Lobste.rs while I was waiting for a plane \u2014 this section of the post is an\nexpanded version of that comment.)<\/p><h4 id=\"packages\">Packages<\/h4><p>We often see questions in IRC and Discord that look something like: \u201cHow do I\nexport a class from a package\u201d?  Questions worded like this are a sign of a\nvery common misunderstanding about what packages in Common Lisp <em>actually are<\/em>.<\/p><p><strong>A package in Common Lisp is a container for symbols<\/strong>.  That\u2019s it.  They\u2019re\na way to group related names (symbols) together so you don\u2019t have to do the\nmiserable prefixing of every name with <code>mylibrary-...<\/code> like you need to do in\nEmacs Lisp or C to avoid name clashes.<\/p><p>You don\u2019t export a class from a package, you export a <em>symbol<\/em>.  You don\u2019t\nimport a function, you import the <em>symbol<\/em> it\u2019s attached to.  This sounds\npedantic, but is important to keep clear in your head as you start using the\npackage system.  If you\u2019re not clear on what exactly a symbol <em>is<\/em>, I wrote\na <a href=\"http:\/\/stevelosh.com\/blog\/2016\/06\/symbolic-computation\/\">separate post<\/a> just about symbols which you might find helpful.<\/p><p>Another major tripping point for new people is the relationship between packages\nand files.  Or, rather: the completely <em>lack<\/em> of any relationship in Common\nLisp.<\/p><p>In many languages like Python, Java, or Clojure, a file\u2019s package and its\nlocation on the hard drive are tied together.  For example: when you say <code>import\nfoo.bar.baz<\/code> in Python, Python will look for a <code>baz.py<\/code> file inside the\n<code>foo\/bar\/<\/code> directory (it\u2019s a little more complicated than this, but that doesn\u2019t\nmatter for this example).<\/p><p>In Common Lisp, this is not the case.  <strong>Files and packages are completely\nunrelated in Common Lisp.<\/strong>  You can have many files that all work in the same\npackage, or one file that switches between many packages, or even create or\nmodify packages at runtime.<\/p><p>This gives you the flexibility to work however you want.  For example: in my\nprocedural art library <a href=\"https:\/\/github.com\/sjl\/flax\">Flax<\/a> most of the packages are each used in one\nspecific file, much like you would do in modern languages.  But the\n<code>flax.drawing<\/code> package contains not only a drawing protocol but also several\nimplementations of that protocol (PNG, SVG, etc), and so I split the code into\n<a href=\"https:\/\/github.com\/sjl\/flax\/tree\/master\/src\/drawing\">a series of separate files<\/a>, each one dealing with how to draw\na single format (plus one for the protocol itself).<\/p><p>I could have created separate packages for each implementation and set up the\nimports\/exports between them, but I didn\u2019t feel like the extra boilerplate was\nworth it.  Common Lisp is flexible enough to let you make such choices.<\/p><p>So if files and packages aren\u2019t related, the next question is: how does Common\nLisp know where to <em>find<\/em> anything on disk when it comes time to load the code?<\/p><h4 id=\"systems\">Systems<\/h4><p>A system in Common Lisp is a collection of serveral things:<\/p><ul><li>Some code.<\/li>\n<li>A description of how to load that code.<\/li>\n<li>A list of other systems this system depends on, which need to be loaded prior\nto loading this one.<\/li>\n<li>Some metadata like author, license, version, homepage, etc.<\/li>\n<\/ul><p>The Common Lisp language itself has no knowledge of systems.  If you look at\n<a href=\"https:\/\/www.cs.cmu.edu\/Groups\/AI\/html\/cltl\/clm\/node120.html\">section 11.9<\/a> of CLtL2 you\u2019ll see that it was imagined that\neach author would write their own custom file to load their code.  But since\nCommon Lisp gives you the power to abstract almost anything, people eventually\nabstracted the process of loading Common Lisp code.<\/p><p><a href=\"https:\/\/common-lisp.net\/project\/asdf\/\">ASDF<\/a> is a Common Lisp library bundled with most modern implementations\nwhich handles defining and loading systems.  The name ASDF stands for \u201cAnother\nSystem Definition Facility\u201d, so as you might guess there have been several\nother such libraries.  ASDF is the one everyone uses today.<\/p><p>ASDF standardizes the process of defining a system into something like this:<\/p><ul><li>The system definition(s) for a project called <code>foo<\/code> would be in a file named <code>foo.asd<\/code>.<\/li>\n<li>Each system is defined with a <code>(defsystem ...)<\/code> form inside this file.<\/li>\n<\/ul><p>We\u2019ll talk more about what a \u201cproject\u201d is shortly.  Note the extension of the\nfile is <code>asd<\/code>, not <code>asdf<\/code>, which is a little confusing, but was probably chosen\nto work in environments with three-letter-extension limits.<\/p><p>The <a href=\"https:\/\/common-lisp.net\/project\/asdf\/#documentation\">ASDF manual<\/a> is the definitive resource for the syntax and semantics of\n<code>defsystem<\/code>, but can be a little heavy to read if you\u2019re just getting started.\nAnother way to get started is to read some <code>.asd<\/code> files of some small-to-medium\nsized open source projects and see how they handle things.<\/p><p>Systems and packages are orthogonal in Common Lisp.  Some systems (like small\nlibraries) will define exactly one package.  Some systems will define multiple\npackages.  Rarely a system might not define any new packages, but will use or\nadd to an existing one.<\/p><p>For example:<\/p><ul><li>My directed graph library <a href=\"https:\/\/github.com\/sjl\/cl-digraph\">cl-digraph<\/a> contains a system called <code>cl-digraph<\/code>.<\/li>\n<li>That system has a description of how to load the code, which lives in the\n<a href=\"https:\/\/github.com\/sjl\/cl-digraph\/blob\/master\/cl-digraph.asd\"><code>cl-digraph.asd<\/code><\/a> file.<\/li>\n<li>One of the files specified for loading is <a href=\"https:\/\/github.com\/sjl\/cl-digraph\/blob\/master\/package.lisp\"><code>package.lisp<\/code><\/a>,\nwhich creates a package called <code>digraph<\/code>.<\/li>\n<\/ul><p>Even though ASDF standardizes some aspects of system definition, it still gives\nyou plenty of flexibility.  As you read projects by different authors you\u2019ll\nencounter different ways of organizing systems \u2014 this can be a little\noverwhelming at first, but it means you can organize a system in the way that\nworks <em>best for that system<\/em>, which is really nice once you\u2019ve got some\nexperience under your belt.<\/p><p>One example of this is how people define packages for their systems.  There are\na couple of common ways to do this you\u2019ll see in the wild:<\/p><ul><li>A single <code>package.lisp<\/code> file which contains all the definitions for all the\npackages in the project, and gets loaded before all other files.  This is the\nstrategy I usually prefer.<\/li>\n<li>Each file defines its package at the top of the file, much like you would in\nClojure or other modern languages.  Care is taken in the system definition to\nload the files in the correct order so that each package is defined before it\nis ever used.<\/li>\n<\/ul><p>To review: a system is a collection of code and a description of how to load it,\na list of its dependencies, and some metadata.  Now let\u2019s move up one level\nhigher to the final layer of structure you need to know about.<\/p><h4 id=\"projects\">Projects<\/h4><p>A project in Common Lisp is not an official term defined anywhere that I know\nof, but is a word that\u2019s generally used to mean something like a library,\na framework, an application, etc.<\/p><p>A project will usually define at least one system, because systems are where\nyou describe how to load the code, and if a project didn\u2019t define a system how\nwould you know how to load its code?  My string-wrapping library <a href=\"https:\/\/github.com\/sjl\/bobbin\">Bobbin<\/a> is\na project that defines <em>two<\/em> systems:<\/p><ul><li><p>The <code>bobbin<\/code> system contains the actual data structure and API.  It has no\ndependencies.<\/p><\/li>\n<li><p>The <code>bobbin\/test<\/code> system contains the unit tests.  It depends on the <code>bobbin<\/code>\nsystem (because that\u2019s the code it\u2019s going to test) and the <code>1am<\/code> system (a\nunit test framework).  I made this a separate system because it allows users\nto load the main code without also having to load the unit testing framework\nif they\u2019re not going to be running the tests.<\/p><\/li>\n<\/ul><p>Both of these systems are defined in the <a href=\"https:\/\/github.com\/sjl\/bobbin\/blob\/master\/bobbin.asd\"><code>bobbin.asd<\/code> file<\/a>.  ASDF\n<a href=\"https:\/\/common-lisp.net\/project\/asdf\/asdf.html#index-find_002dsystem\">treats systems with a forward slash in their name specially<\/a> and\nknows to look for them in the <code>asd<\/code> file named with the text before the slash.<\/p><p>We saw how Common Lisp has no concept of a system \u2014 that concept comes from\nASDF.  Similarly, ASDF has no concept of the internet or of reaching out to\nsomewhere to download things.  ASDF assumes you have somehow acquired the\nsystems you want to load and stored them on your hard drive, perhaps by sending\na check to an address and receiving a copy of the code on floppy disk, as many\nof my old Lisp books offer in their final pages.<\/p><p><a href=\"https:\/\/www.quicklisp.org\/beta\/\">Quicklisp<\/a> is another library that works on top of ASDF to provide the\n\u201cdownload projects from the internet automatically if necessary\u201d functionality\nthat people expect in the modern world.  So when you say <code>(ql:quickload\n:bobbin)<\/code> you\u2019re asking Quicklisp to download Bobbin (and any dependencies) if\nnecessary, and then hand it off to ASDF to actually load the code of the\n<code>bobbin<\/code> system.<\/p><p>Unlike ASDF, Quicklisp is relatively new in the Common Lisp world (it\u2019s only\nabout eight years old) and so is not bundled with any modern Lisp\nimplementations that I know of, which is why you need to install it separately.<\/p><h4 id=\"recap\">Recap<\/h4><p>Here\u2019s a quick recap of the different layers of project structure you\u2019ll\nencounter in Common Lisp.  Jot these down on a post it note you can refer to as\nyou\u2019re learning.<\/p><ul><li><strong>Files<\/strong> are files on your hard drive.<\/li>\n<li><strong>Packages<\/strong> are containers of symbols. They are orthogonal to files.<\/li>\n<li><strong>Systems<\/strong> are collections of code, instructions on how to load that code,\ndependency lists, and metadata. They are orthogonal to packages.<\/li>\n<li><strong>Projects<\/strong> are high-level collections of\u2026 \u201cstuff\u201d such as code,\ndocumentation, maybe some image assets, etc.  They are (mostly) orthogonal to\nsystems (are you seeing a trend here?).<\/li>\n<li>Common Lisp itself knows about files and packages.<\/li>\n<li>ASDF adds systems.<\/li>\n<li>Quicklisp adds the internet.<\/li>\n<\/ul><h3 id=\"common-libraries\">Common Libraries<\/h3><p>Common Lisp doesn\u2019t have as <em>large<\/em> of a community as some newer languages, but\nit still has a lot of libraries because it\u2019s had a community for a longer time.\nThe stability of the core language means that many libraries written in portable\nCommon Lisp ten or fifteen years ago can still run just fine today.<\/p><p>In this final section I\u2019ll give you a quick overview of some of the more popular\nlibraries you might run into as you learn the language.  You don\u2019t have to use\nall of them, but it\u2019s helpful to have some idea of what\u2019s available.<\/p><h4 id=\"alexandria\">Alexandria<\/h4><p><a href=\"https:\/\/common-lisp.net\/project\/alexandria\/\">Alexandria<\/a> is one of the most popular Common Lisp libraries (the\nname is a pun on the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Library_of_Alexandria\">Library of Alexandria<\/a>), and it\u2019s\na collection of all kinds of useful little utility functions like\n<code>read-file-into-byte-vector<\/code> and <code>map-permutations<\/code>.<\/p><p>There are a <em>lot<\/em> of utility libraries for Common Lisp around \u2014 one rite of\npassage is building up your own personal utility library over time \u2014 but\nAlexandria is the most popular one.  Most projects with any dependencies at all\nwill eventually end up with Alexandria in the dependency graph somewhere.<\/p><h4 id=\"bordeaux-threads\">Bordeaux Threads<\/h4><p><a href=\"https:\/\/common-lisp.net\/project\/bordeaux-threads\/\">Bordeaux Threads<\/a> was mentioned earlier.  Threads aren\u2019t part of the Common\nLisp standard, but most implementations provide their own custom interface for\nworking with them.  Bordeaux Threads wraps all these implementation-specific\ninterfaces and provides an API so you can write threaded code that will work\nportably.<\/p><p>If you\u2019re looking for something like Java\u2019s <code>new Thread(() -&gt; foo()).start()<\/code>,\nthis is what you want.<\/p><h4 id=\"cffi\">CFFI<\/h4><p><a href=\"https:\/\/common-lisp.net\/project\/cffi\/\">CFFI<\/a> is a foreign-function interface library that lets you load\nC libraries (e.g. <code>foo.dylib<\/code> or <code>foo.so<\/code>) and call the functions in them.  It\nworks by wrapping implementation-specific interfaces, because this isn\u2019t part of\nthe Common Lisp standard.<\/p><p>Unfortunately it has the same name as Python\u2019s FFI library, so if you\u2019re\nsearching for documentation make sure you\u2019re looking at the right version.<\/p><h4 id=\"cl-ppcre\">CL-PPCRE<\/h4><p><a href=\"https:\/\/edicl.github.io\/cl-ppcre\/\">CL-PPCRE<\/a> is an implementation of Perl-compatible regular expressions.\nIf you\u2019re looking to use regular expressions in Common Lisp, this is what you\nwant.<\/p><h4 id=\"drakma\">Drakma<\/h4><p><a href=\"https:\/\/edicl.github.io\/drakma\/\">Drakma<\/a> is an HTTP client.  If you need to make an HTTP request, this is what\nyou want.  There are other HTTP clients around, but Drakma is commonly used and\nis fine for almost anything you might need.<\/p><h4 id=\"iterate\">Iterate<\/h4><p><a href=\"https:\/\/common-lisp.net\/project\/iterate\/\">Iterate<\/a> is a replacement for the <code>loop<\/code> macro.  It works similarly,\nbut has a more Lispy syntax and a well-defined API for extending it with new\niteration constructs.  I really like it myself, but beware: if you get used\nto <code>iterate<\/code> going back to vanilla <code>loop<\/code> will feel painful.<\/p><p><a href=\"https:\/\/common-lisp.net\/project\/local-time\/\">local-time<\/a> is a library for working with time and dates in Common Lisp.  The\nstandard has some basic support for times built in, but if you want to do much\ncalculation with times (including timezones) this is probably what you want.  If\nyou\u2019re looking for something like <a href=\"http:\/\/www.joda.org\/joda-time\/\">Joda Time<\/a> in Common Lisp, this is\nas close as you\u2019re going to get.<\/p><h4 id=\"lparallel\">lparallel<\/h4><p><a href=\"https:\/\/lparallel.org\/\">lparallel<\/a> is a library that builds on top of Bordeaux Threads to make common\nparallel processing operations much easier.  Think of it as <a href=\"https:\/\/www.gnu.org\/software\/parallel\/\">GNU Parallel<\/a> for\nLisp, with a few extra features (e.g. channels and tasks).<\/p><p>For example: if you\u2019ve got a big vector you\u2019re mapping over with\n<code>(map 'vector #'work some-vector)<\/code> you can split it into chunks and\nrun in multiple threads by changing it to\n<code>(lparallel:pmap 'vector #'work some-vector)<\/code>.<\/p><h4 id=\"named-readtables\">Named Readtables<\/h4><p><a href=\"https:\/\/github.com\/melisgl\/named-readtables\">Named readtables<\/a> is a library that adds namespaces for\nreadtables.<\/p><p>One painful part of the standard is that reader macros are added and removed to\nthe global readtable on the fly, so if you load multiple systems that define the\nsame reader macros things can get messy.  Named readtables adds some much-needed\nhygiene to that process.  If you\u2019re working with reader macros at all you\nabsolutely want to use this.<\/p><h4 id=\"roswell\">Roswell<\/h4><p><a href=\"https:\/\/github.com\/roswell\/roswell\">Roswell<\/a> is a couple of things rolled into one.  It\u2019s a C program that\nhandles installing and running multiple different Common Lisp implementations\n(kind of like <a href=\"https:\/\/github.com\/creationix\/nvm\">NVM<\/a> or\n<a href=\"https:\/\/rvm.io\/\">rvm<\/a>), and it also provides a unified way to write small shell\nscripts in Common Lisp and compile them into binaries.<\/p><p>I used Roswell for a little over a year, but I eventually stopped and now\nI don\u2019t think it\u2019s worth the trouble, for a couple of reasons.<\/p><p>First: if you write portable code you generally don\u2019t need to worry running\na particular version of an implementation, because Common Lisp is so stable.\nI usually just install the latest version of each implementation I use with\na package manager or by building from source.<\/p><p>Second: after using it for a while I found that Roswell was always very brittle\nto upgrade, and whenever things broke it would spew an almost JVM-sized stack\ntrace without a decent error message.<\/p><p>For me, the negatives outweighed the positives.  I\u2019d recommend simply using the\nlatest version of the implementations you care about and writing portable code.\nFor the compiling-into-binaries functionality I\u2019d recommend using your\nimplementation\u2019s built-in support for this, or using UIOP\u2019s wrapper around that,\nor using a separate library like <a href=\"https:\/\/shinmera.github.io\/deploy\/\">Deploy<\/a>.<\/p><p>Of course your mileage might vary.  If you find yourself <em>really<\/em> needing to run\nspecific versions of specific Common Lisp implementations in rapid succession,\nyou should look into Roswell.<\/p><h4 id=\"series\">SERIES<\/h4><p><a href=\"https:\/\/www.cliki.net\/Series\">SERIES<\/a> was almost included in Common Lisp (it\u2019s in <a href=\"https:\/\/www.cs.cmu.edu\/Groups\/AI\/html\/cltl\/clm\/node347.html\">Appendix A of\nCLtL2<\/a>), but didn\u2019t quite make it.  It\u2019s a library for writing\nfunctional code that looks like the traditional <code>map<\/code> and <code>filter<\/code> and <code>reduce<\/code>\noperations but which compiles down to efficient loops.<\/p><p>If you\u2019re looking for Clojure\u2019s transducers in Common Lisp, this is what you\nwant.<\/p><h4 id=\"st-json\">st-json<\/h4><p>JSON support in Common Lisp is a god damn mess.  There are <a href=\"https:\/\/sites.google.com\/site\/sabraonthehill\/home\/json-libraries\">an absurd number of\nJSON libraries<\/a> and I don\u2019t really <em>like<\/em> any of them.<\/p><p>For me, the most important quality I need in a JSON library is an unambiguous,\none-to-one mapping of types.  For example: some libraries will deserialize JSON\narrays as Lisp lists, and JSON <code>true<\/code>\/<code>false<\/code> as <code>t<\/code>\/<code>nil<\/code>.  But this means <code>[]<\/code> and\n<code>false<\/code> both deserialize to <code>nil<\/code>, so you can\u2019t reliably round trip anything!<\/p><p>I\u2019ve settled on using <a href=\"https:\/\/marijnhaverbeke.nl\/st-json\/\">st-json<\/a> and wrapping it up to be a little more\nergonomic with some glue code.  It\u2019s not the fastest solution out there, but\nit works for my needs.  There are plenty of other options out there, so if you\nhave different needs than me you should look into them.<\/p><h4 id=\"usocket\">usocket<\/h4><p><a href=\"https:\/\/common-lisp.net\/project\/usocket\/\">usocket<\/a> is a library for networking sockets.  Sockets and networking aren\u2019t\npart of the Common Lisp standard, but most implementations provide a custom\ninterface for working with them.  usocket wraps the implementation-specific\ninterfaces and provides an API so you can write networking code portably.<\/p><p>If you want to make Lisp listen on a port and read streams of bytes from\nclients, or want to connect to a port and send raw bytes to it, this is what you\nwant.<\/p><h2 id=\"good-luck\">Good Luck!<\/h2><p>I hope this whirlwind tour was useful.  Common Lisp is an old, deep language.\nIt\u2019s not something you can learn in a month, but if you\u2019re willing to spend the\ntime it will reward careful study.<\/p><p>Feel free to email me or pop into IRC or Discord if you have questions.<\/p><p>Good luck!<\/p>","created_at":"2018-11-07T16:22:27+0000","updated_at":"2018-11-07T16:22:27+0000","annotations":[],"mimetype":"text\/html","reading_time":53,"domain_name":"stevelosh.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":77,"title":"Hashlife","url":"https:\/\/en.wikipedia.org\/wiki\/Hashlife","content":"<a class=\"mw-jump-link\" href=\"#mw-head\">Jump to navigation<\/a>\n\t\t\t\t<a class=\"mw-jump-link\" href=\"#p-search\">Jump to search<\/a>\n\t\t\t\t<div id=\"mw-content-text\" lang=\"en\" dir=\"ltr\" class=\"mw-content-ltr\"><div class=\"mw-parser-output\"><div class=\"thumb tright\"><div class=\"thumbinner\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/File:Turing_Machine_in_Golly.png\" class=\"image\"><img alt=\"\" src=\"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/05\/Turing_Machine_in_Golly.png\/379px-Turing_Machine_in_Golly.png\" width=\"379\" height=\"212\" class=\"thumbimage\" srcset=\"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/05\/Turing_Machine_in_Golly.png\/569px-Turing_Machine_in_Golly.png 1.5x, \/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/05\/Turing_Machine_in_Golly.png\/758px-Turing_Machine_in_Golly.png 2x\" data-file-width=\"1256\" data-file-height=\"704\" \/><\/a>  <div class=\"thumbcaption\">The 6,366,548,773,467,669,985,195,496,000 (6 octillionth) generation of a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Turing_machine\" title=\"Turing machine\">Turing machine<\/a> in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Conway%27s_Game_of_Life\" title=\"Conway's Game of Life\">Life<\/a> computed in less than 30 seconds on an <a href=\"https:\/\/en.wikipedia.org\/wiki\/Intel_Core\" title=\"Intel Core\">Intel Core<\/a> Duo 2GHz CPU using hashlife in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Golly_(program)\" title=\"Golly (program)\">Golly<\/a>. Computed by detecting a repeating cycle in the pattern, and skipping ahead to any requested generation.<\/div><\/div><\/div>\n<p><b>Hashlife<\/b> is a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Memoization\" title=\"Memoization\">memoized<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Algorithm\" title=\"Algorithm\">algorithm<\/a> for computing the long-term fate of a given starting configuration in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Conway%27s_Game_of_Life\" title=\"Conway's Game of Life\">Conway's Game of Life<\/a> and related <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cellular_automaton\" title=\"Cellular automaton\">cellular automata<\/a>, much more quickly than would be possible using alternative algorithms that simulate each time step of each cell of the automaton. The algorithm was first described by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bill_Gosper\" title=\"Bill Gosper\">Bill Gosper<\/a> in the early 1980s while he was engaged in research at the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Xerox_Palo_Alto_Research_Center\" class=\"mw-redirect\" title=\"Xerox Palo Alto Research Center\">Xerox Palo Alto Research Center<\/a>. Hashlife was originally implemented on <a href=\"https:\/\/en.wikipedia.org\/wiki\/Symbolics\" title=\"Symbolics\">Symbolics<\/a> <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lisp_machine\" title=\"Lisp machine\">Lisp machines<\/a> with the aid of the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Flavors_(programming_language)\" title=\"Flavors (programming language)\">Flavors<\/a> extension.\n<\/p>\n\n\n<h2>Hashlife[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=1\" title=\"Edit section: Hashlife\">edit<\/a>]<\/h2>\n<p>Hashlife is designed to exploit large amounts of spatial and temporal <a href=\"https:\/\/en.wikipedia.org\/wiki\/Redundancy_(information_theory)\" title=\"Redundancy (information theory)\">redundancy<\/a> in most Life rules. For example, in <a href=\"https:\/\/en.wikipedia.org\/wiki\/Conway%27s_Game_of_Life\" title=\"Conway's Game of Life\">Conway's Life<\/a>, many seemingly random patterns end up as collections of simple <a href=\"https:\/\/en.wikipedia.org\/wiki\/Still_life_(CA)\" class=\"mw-redirect\" title=\"Still life (CA)\">still lifes<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Oscillator_(CA)\" class=\"mw-redirect\" title=\"Oscillator (CA)\">oscillators<\/a>.\n<\/p>\n<h3>Representation[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=2\" title=\"Edit section: Representation\">edit<\/a>]<\/h3>\n<p>The field is typically treated as a theoretically <a href=\"https:\/\/en.wikipedia.org\/wiki\/Infinity\" title=\"Infinity\">infinite<\/a> grid, with the <a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Pattern_(CA)&amp;action=edit&amp;redlink=1\" class=\"new\" title=\"Pattern (CA) (page does not exist)\">pattern<\/a> in question centered near the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Origin_(mathematics)\" title=\"Origin (mathematics)\">origin<\/a>. A <a href=\"https:\/\/en.wikipedia.org\/wiki\/Quadtree\" title=\"Quadtree\">quadtree<\/a> is used to represent the field. Given a square of 2<sup>2<i>k<\/i><\/sup> cells, 2<sup><i>k<\/i><\/sup> on a side, at the <i>k<\/i>th level of the tree, the hash table stores the 2<sup><i>k<\/i>-1<\/sup>-by-2<sup><i>k<\/i>-1<\/sup> square of cells in the center, 2<sup><i>k<\/i>-2<\/sup> generations in the future.  For example, for a 4x4 square it stores the 2x2 center, one generation forward; and for an 8x8 square it stores the 4x4 center, two generations forward.\n<\/p>\n<h3>Hashing[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=3\" title=\"Edit section: Hashing\">edit<\/a>]<\/h3>\n<p>While a quadtree trivially has far more <a href=\"https:\/\/en.wikipedia.org\/wiki\/Computational_overhead\" class=\"mw-redirect\" title=\"Computational overhead\">overhead<\/a> than other simpler representations (such as using a <a href=\"https:\/\/en.wikipedia.org\/wiki\/Matrix_(mathematics)\" title=\"Matrix (mathematics)\">matrix<\/a> of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Bit\" title=\"Bit\">bits<\/a>), it allows for various optimizations. As the name suggests, the algorithm uses <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hash_table\" title=\"Hash table\">hash tables<\/a> to store the nodes of the quadtree. Many subpatterns in the tree are usually identical to each other; for example the pattern being studied may contain many copies of the same <a href=\"https:\/\/en.wikipedia.org\/wiki\/Spaceship_(CA)\" class=\"mw-redirect\" title=\"Spaceship (CA)\">spaceship<\/a>, or even large swathes of empty space. These subpatterns will all <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hash_function\" title=\"Hash function\">hash<\/a> to the same position in the hash table, and thus many copies of the same subpattern can be stored using the same hash table entry. In addition, these subpatterns only need to be evaluated once, not once per copy as in other Life algorithms.\n<\/p><p>This itself leads to significant improvements in resource requirements; for example a generation of the various <a href=\"https:\/\/en.wikipedia.org\/wiki\/Breeder_(CA)\" class=\"mw-redirect\" title=\"Breeder (CA)\">breeders<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Spacefiller_(CA)\" class=\"mw-redirect\" title=\"Spacefiller (CA)\">spacefillers<\/a>, which grow at <a href=\"https:\/\/en.wikipedia.org\/wiki\/Polynomial\" title=\"Polynomial\">polynomial<\/a> speeds, can be evaluated in Hashlife using <a href=\"https:\/\/en.wikipedia.org\/wiki\/Logarithmic_growth\" title=\"Logarithmic growth\">logarithmic<\/a> space and time.\n<\/p>\n<h3>Superspeed and caching[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=4\" title=\"Edit section: Superspeed and caching\">edit<\/a>]<\/h3>\n<p>A further speedup for many patterns can be achieved by evolving different nodes at different speeds. For example, one could compute twice the number of generations forward for a node at the (<i>k<\/i>+1)-th level compared to one at the <i>k<\/i>th. For sparse or repetitive patterns such as the classical <a href=\"https:\/\/en.wikipedia.org\/wiki\/Gun_(CA)\" class=\"mw-redirect\" title=\"Gun (CA)\">glider gun<\/a>, this can result in tremendous speedups, allowing one to compute <i>bigger<\/i> patterns at <i>higher<\/i> generations <i>faster<\/i>, sometimes <a href=\"https:\/\/en.wikipedia.org\/wiki\/Exponential_growth\" title=\"Exponential growth\">exponentially<\/a>. To take full advantage of this feature, subpatterns from past generations should be <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cache_(computing)\" title=\"Cache (computing)\">saved<\/a> as well.\n<\/p><p>Since different patterns are allowed to run at different speeds, some implementations, like Gosper's own <code>hlife<\/code> program, do not have an interactive display, but simply compute a preset end result for a starting pattern, usually run from the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Command_line\" class=\"mw-redirect\" title=\"Command line\">command line<\/a>. More recent programs such as <a href=\"https:\/\/en.wikipedia.org\/wiki\/Golly_(program)\" title=\"Golly (program)\">Golly<\/a>, however, have a graphical interface that can drive a Hashlife-based engine.\n<\/p><p>The typical behavior of a Hashlife program on a conducive pattern is as follows: first the algorithm runs slower compared to other algorithms because of the constant overhead associated with <a href=\"https:\/\/en.wikipedia.org\/wiki\/Hash_function\" title=\"Hash function\">hashing<\/a> and building the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Quadtree\" title=\"Quadtree\">tree<\/a>; but later, enough data will be gathered and its speed will increase tremendously \u2013 the rapid increase in speed is often described as \"<a href=\"https:\/\/en.wikipedia.org\/wiki\/Exponential_growth\" title=\"Exponential growth\">exploding<\/a>\".\n<\/p>\n<h2>Drawbacks[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=5\" title=\"Edit section: Drawbacks\">edit<\/a>]<\/h2>\n<p>Like many <a href=\"https:\/\/en.wikipedia.org\/wiki\/Memoization\" title=\"Memoization\">memoized<\/a> codes, Hashlife can consume significantly more <a href=\"https:\/\/en.wikipedia.org\/wiki\/Computer_memory\" title=\"Computer memory\">memory<\/a> than other algorithms, especially on moderate-sized patterns with a lot of entropy, or which contain subpatterns poorly aligned to the bounds of the quadtree nodes (i.e. power-of-two sizes); the cache is a vulnerable component. It can also consume more time than other algorithms on these patterns. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Golly_(program)\" title=\"Golly (program)\">Golly<\/a>, among other Life simulators, has options for toggling between Hashlife and conventional algorithms.\n<\/p><p>Hashlife is also significantly more complex to <a href=\"https:\/\/en.wikipedia.org\/wiki\/Computer_programming\" title=\"Computer programming\">implement<\/a>. For example, it needs a dedicated <a href=\"https:\/\/en.wikipedia.org\/wiki\/Garbage_collection_(computer_science)\" title=\"Garbage collection (computer science)\">garbage collector<\/a> to remove unused nodes from the cache.\n<\/p>\n<h2>See also[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=6\" title=\"Edit section: See also\">edit<\/a>]<\/h2>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Optimization_(computer_science)\" class=\"mw-redirect\" title=\"Optimization (computer science)\">Optimization (computer science)<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Purely_functional_data_structure\" title=\"Purely functional data structure\">Purely functional data structure<\/a>, of which the hashed quadtree is one<\/li><\/ul><h2>References[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=7\" title=\"Edit section: References\">edit<\/a>]<\/h2>\n<ul><li><cite class=\"citation journal\">Gosper, Bill (1984). \"Exploiting Regularities in Large Cellular Spaces\". <i>Physica D<\/i>. Elsevier. <b>10<\/b> (1\u20132): 75\u201380. <a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_object_identifier\" title=\"Digital object identifier\">doi<\/a>:<a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/doi.org\/10.1016%2F0167-2789%2884%2990251-3\">10.1016\/0167-2789(84)90251-3<\/a>.<\/cite><\/li><\/ul><h2>External links[<a href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;action=edit&amp;section=8\" title=\"Edit section: External links\">edit<\/a>]<\/h2>\n<ul><li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ericweisstein.com\/encyclopedias\/life\/HashLife.html\">HashLife from Eric Weisstein's Treasure Trove of Life<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/tomas.rokicki.com\/hlife\/\">Tomas Rokicki's implementation of hashlife<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.argentum.freeserve.co.uk\/lex_h.htm#hashlife\">Entry<\/a> in the <a rel=\"nofollow\" class=\"external text\" href=\"https:\/\/web.archive.org\/web\/20160712171724\/http:\/\/www.argentum.freeserve.co.uk\/lex_home.htm\">Life Lexicon<\/a><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/www.ddj.com\/dept\/ai\/184406478\">Explanation of the algorithm<\/a> from <i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Dr._Dobb%27s_Journal\" title=\"Dr. Dobb's Journal\">Dr. Dobb's Journal<\/a><\/i><\/li>\n<li><a rel=\"nofollow\" class=\"external text\" href=\"http:\/\/jennyhasahat.github.io\/hashlife.html\">Gosper's Algorithm (Hashlife) Explained<\/a><\/li><\/ul><div role=\"navigation\" class=\"navbox\" aria-labelledby=\"Conway&amp;#039;s_Game_of_Life_and_related_cellular_automata\"><table class=\"nowraplinks collapsible collapsed navbox-inner\"><tbody><tr><th scope=\"col\" class=\"navbox-title\" colspan=\"2\"><div class=\"plainlinks hlist navbar mini\"><ul><li class=\"nv-view\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template:Conway%27s_Game_of_Life\" title=\"Template:Conway's Game of Life\"><abbr title=\"View this template\">v<\/abbr><\/a><\/li><li class=\"nv-talk\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Template_talk:Conway%27s_Game_of_Life\" title=\"Template talk:Conway's Game of Life\"><abbr title=\"Discuss this template\">t<\/abbr><\/a><\/li><li class=\"nv-edit\"><a class=\"external text\" href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Template:Conway%27s_Game_of_Life&amp;action=edit\"><abbr title=\"Edit this template\">e<\/abbr><\/a><\/li><\/ul><\/div><div id=\"Conway&amp;#039;s_Game_of_Life_and_related_cellular_automata\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Conway%27s_Game_of_Life\" title=\"Conway's Game of Life\">Conway's Game of Life<\/a> and related <a href=\"https:\/\/en.wikipedia.org\/wiki\/Cellular_automaton\" title=\"Cellular automaton\">cellular automata<\/a><\/div><\/th><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Structures<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Breeder_(cellular_automaton)\" title=\"Breeder (cellular automaton)\">Breeder<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Garden_of_Eden_(cellular_automaton)\" title=\"Garden of Eden (cellular automaton)\">Garden of Eden<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Glider_(Conway%27s_Life)\" title=\"Glider (Conway's Life)\">Glider<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Gun_(cellular_automaton)\" title=\"Gun (cellular automaton)\">Gun<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Methuselah_(cellular_automaton)\" title=\"Methuselah (cellular automaton)\">Methuselah<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Oscillator_(cellular_automaton)\" title=\"Oscillator (cellular automaton)\">Oscillator<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Puffer_train\" title=\"Puffer train\">Puffer train<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Rake_(cellular_automaton)\" title=\"Rake (cellular automaton)\">Rake<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Reflector_(cellular_automaton)\" title=\"Reflector (cellular automaton)\">Reflector<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Replicator_(cellular_automaton)\" title=\"Replicator (cellular automaton)\">Replicator<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Sawtooth_(cellular_automaton)\" title=\"Sawtooth (cellular automaton)\">Sawtooth<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Spacefiller\" title=\"Spacefiller\">Spacefiller<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Spaceship_(cellular_automaton)\" title=\"Spaceship (cellular automaton)\">Spaceship<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Spark_(cellular_automaton)\" title=\"Spark (cellular automaton)\">Spark<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Still_life_(cellular_automaton)\" title=\"Still life (cellular automaton)\">Still life<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\"><a href=\"https:\/\/en.wikipedia.org\/wiki\/Life-like_cellular_automaton\" title=\"Life-like cellular automaton\">Life variants<\/a><\/th><td class=\"navbox-list navbox-even hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Day_and_Night_(cellular_automaton)\" title=\"Day and Night (cellular automaton)\">Day and Night<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Highlife_(cellular_automaton)\" title=\"Highlife (cellular automaton)\">Highlife<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Life_without_Death\" title=\"Life without Death\">Life without Death<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Seeds_(cellular_automaton)\" title=\"Seeds (cellular automaton)\">Seeds<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Concepts<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Moore_neighborhood\" title=\"Moore neighborhood\">Moore neighborhood<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Speed_of_light_(cellular_automaton)\" title=\"Speed of light (cellular automaton)\">Speed of light<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Von_Neumann_neighborhood\" title=\"Von Neumann neighborhood\">Von Neumann neighborhood<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Implementations<\/th><td class=\"navbox-list navbox-even hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Golly_(program)\" title=\"Golly (program)\">Golly<\/a><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Life_Genesis\" class=\"mw-redirect\" title=\"Life Genesis\">Life Genesis<\/a><\/i><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Mirek%27s_Cellebration\" title=\"Mirek's Cellebration\">Mirek's Cellebration<\/a><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Video_Life\" title=\"Video Life\">Video Life<\/a><\/i><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Key people<\/th><td class=\"navbox-list navbox-odd hlist\"><div>\n<ul><li><a href=\"https:\/\/en.wikipedia.org\/wiki\/John_Horton_Conway\" title=\"John Horton Conway\">John Conway<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Martin_Gardner\" title=\"Martin Gardner\">Martin Gardner<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Bill_Gosper\" title=\"Bill Gosper\">Bill Gosper<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Richard_K._Guy\" title=\"Richard K. Guy\">Richard Guy<\/a><\/li><\/ul><\/div><\/td><\/tr><tr><th scope=\"row\" class=\"navbox-group\">Popular culture<\/th><td class=\"navbox-list navbox-even hlist\"><div>\n<ul><li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Bloom_(novel)\" title=\"Bloom (novel)\">Bloom<\/a><\/i><\/li>\n<li><i><a href=\"https:\/\/en.wikipedia.org\/wiki\/Wake_(Sawyer_novel)\" title=\"Wake (Sawyer novel)\">Wake<\/a><\/i><\/li><\/ul><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n\n\n\n\n\n<\/div><noscript><img src=\"https:\/\/en.wikipedia.org\/wiki\/Special:CentralAutoLogin\/start?type=1x1\" alt=\"\" title=\"\" width=\"1\" height=\"1\" \/><\/noscript><\/div>\t\t\t\t\t<div class=\"printfooter\">\n\t\t\t\t\t\tRetrieved from \"<a dir=\"ltr\" href=\"https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;oldid=856628049\">https:\/\/en.wikipedia.org\/w\/index.php?title=Hashlife&amp;oldid=856628049<\/a>\"\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<div class=\"visualClear\">\n\t\t\t\t\t\t\t<\/div>","created_at":"2018-11-07T16:22:27+0000","updated_at":"2018-11-07T16:22:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"en.wikipedia.org","preview_picture":"https:\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/0\/05\/Turing_Machine_in_Golly.png\/1200px-Turing_Machine_in_Golly.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":78,"title":"Wilfred\/deadgrep","url":"https:\/\/github.com\/Wilfred\/deadgrep","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p>Deadgrep is the fast, beautiful text search that your Emacs\ndeserves.<\/p>\n<p><a href=\"https:\/\/travis-ci.org\/Wilfred\/deadgrep\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/cb4135d7f8e55026c123df3b30a3b363b99d1d05\/68747470733a2f2f7472617669732d63692e6f72672f57696c667265642f64656164677265702e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/Wilfred\/deadgrep.svg?branch=master\" \/><\/a> <a href=\"https:\/\/coveralls.io\/github\/Wilfred\/deadgrep?branch=master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f31904c3007034dbf2ccbe6307b7cdd43220c8db\/68747470733a2f2f636f766572616c6c732e696f2f7265706f732f6769746875622f57696c667265642f64656164677265702f62616467652e7376673f6272616e63683d6d6173746572\" alt=\"Coverage Status\" data-canonical-src=\"https:\/\/coveralls.io\/repos\/github\/Wilfred\/deadgrep\/badge.svg?branch=master\" \/><\/a> <a href=\"http:\/\/www.melpa.org\/#\/deadgrep\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/39d30856cf6449da6b63162d4d4d64c94dd49393\/687474703a2f2f7777772e6d656c70612e6f72672f7061636b616765732f64656164677265702d62616467652e737667\" alt=\"MELPA\" data-canonical-src=\"http:\/\/www.melpa.org\/packages\/deadgrep-badge.svg\" \/><\/a><\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_screenshot.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_screenshot.png\" alt=\"screenshot\" \/><\/a><\/p>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#usage\">Usage<\/a>\n<ul><li><a href=\"#installation\">Installation<\/a><\/li>\n<li><a href=\"#keybindings\">Keybindings<\/a><\/li>\n<li><a href=\"#minibuffer\">Minibuffer<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#features\">Features<\/a>\n<ul><li><a href=\"#super-fast\">Super Fast<\/a><\/li>\n<li><a href=\"#context-aware\">Context Aware<\/a><\/li>\n<li><a href=\"#easy-filtering\">Easy Filtering<\/a><\/li>\n<li><a href=\"#polish\">Polish<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#alternative-projects\">Alternative Projects<\/a><\/li>\n<\/ul><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h3>\n<ol><li>\n<p>Install <a href=\"https:\/\/github.com\/BurntSushi\/ripgrep\">ripgrep<\/a>. You need\nversion 0.8 or later.<\/p>\n<\/li>\n<li>\n<p>Install deadgrep from MELPA (do <code>M-x package-install<\/code>\n<kbd>RET<\/kbd> <code>deadgrep<\/code> within Emacs 25.1 or later),\nor copy it into your <code>~\/.emacs.d<\/code>.<\/p>\n<\/li>\n<li>\n<p>(Optional) add deadgrep to a convenient shortcut. I like F5.<\/p>\n<\/li>\n<\/ol><div class=\"highlight highlight-source-emacs-lisp\"><pre>(global-set-key (kbd \"&lt;f5&gt;\") #'deadgrep)<\/pre><\/div>\n<h3><a id=\"user-content-keybindings\" class=\"anchor\" aria-hidden=\"true\" href=\"#keybindings\"><\/a>Keybindings<\/h3>\n<table><thead><tr><th>Key<\/th>\n<th>Action<\/th>\n<\/tr><\/thead><tbody><tr><td><kbd>RET<\/kbd><\/td>\n<td>Visit the result, file or push button at point<\/td>\n<\/tr><tr><td><kbd>o<\/kbd><\/td>\n<td>Visit the result in another window<\/td>\n<\/tr><tr><td><kbd>n<\/kbd> and <kbd>p<\/kbd><\/td>\n<td>Move between results or buttons<\/td>\n<\/tr><tr><td><kbd>g<\/kbd><\/td>\n<td>Re-run the search<\/td>\n<\/tr><tr><td><kbd>TAB<\/kbd><\/td>\n<td>Expand\/collapse results for a file<\/td>\n<\/tr><tr><td><kbd>C-c<\/kbd> <kbd>C-k<\/kbd><\/td>\n<td>Stop a running search<\/td>\n<\/tr><tr><td><kbd>C-u<\/kbd><\/td>\n<td>A prefix argument prevents searches from starting automatically.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-minibuffer\" class=\"anchor\" aria-hidden=\"true\" href=\"#minibuffer\"><\/a>Minibuffer<\/h3>\n<p>You use the minibuffer to enter a new search term.<\/p>\n<p>You can also reuse a previous search term with <kbd>M-p<\/kbd> in the\nminibuffer. To edit the default search term, use <kbd>M-n<\/kbd>.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h2>\n<h3><a id=\"user-content-super-fast\" class=\"anchor\" aria-hidden=\"true\" href=\"#super-fast\"><\/a>Super Fast<\/h3>\n<p>Deadgrep uses ripgrep for extremely fast text searches.<\/p>\n<p>If you change your search settings, deadgrep will immediately re-run\nyour search.<\/p>\n<h3><a id=\"user-content-context-aware\" class=\"anchor\" aria-hidden=\"true\" href=\"#context-aware\"><\/a>Context Aware<\/h3>\n<p>Deadgrep works hard to minimise your keystrokes.<\/p>\n<p><strong>Search term<\/strong>: If the region is active, deadgrep uses that.<\/p>\n<p><strong>Search directory<\/strong>: If your current file is in a git project,\ndeadgrep uses that for your search directory.<\/p>\n<p><strong>Regexp and case sensitivity options<\/strong>: Deadgrep re-uses whatever\nsettings you used in your last search.<\/p>\n<p><strong>Globbing and file types<\/strong>: Deadgrep suggests file types and globbing\noptions that match the file you started the search from.<\/p>\n<p><strong>Host<\/strong>: If you're editing a remote file with Tramp, deadgrep will\nsearch the remote machine. This requires <code>rg<\/code> to be installed on that\nmachine.<\/p>\n<h3><a id=\"user-content-easy-filtering\" class=\"anchor\" aria-hidden=\"true\" href=\"#easy-filtering\"><\/a>Easy Filtering<\/h3>\n<p>Didn't get the results you wanted? It's easy to change the search\nterm, search type, or search directory, directly from the results\nbuffer. Just push the relevant button.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_filters.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_filters.png\" alt=\"screenshot\" \/><\/a><\/p>\n<h3><a id=\"user-content-polish\" class=\"anchor\" aria-hidden=\"true\" href=\"#polish\"><\/a>Polish<\/h3>\n<p>Deadgrep uses spinners to give you feedback on whether your search has\nfinished.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_spinner.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_spinner.png\" alt=\"screenshot\" \/><\/a><\/p>\n<p>It highlights regexp syntax according to the syntax accepted by <code>rg<\/code>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_highlight.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_highlight.png\" alt=\"screenshot\" \/><\/a><\/p>\n<p>When navigating to a line that matched, the relevant part of the line\nis temporarily highlighted.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_highlight_relevant.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_highlight_relevant.png\" alt=\"screenshot\" \/><\/a><\/p>\n<p>You can collapse and expand files with <code>TAB<\/code>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_collapsed.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_collapsed.png\" alt=\"screenshot\" \/><\/a><\/p>\n<p>Deadgrep handles minified files robustly.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/deadgrep_truncated.png\"><img src=\"https:\/\/github.com\/Wilfred\/deadgrep\/raw\/master\/docs\/deadgrep_truncated.png\" alt=\"screenshot\" \/><\/a><\/p>\n<p>You can always jump to exactly the position that point is on, even\nwhen searching files that contain tab characters. You can also\nnavigate to the file itself from the file headings.<\/p>\n<h2><a id=\"user-content-alternative-projects\" class=\"anchor\" aria-hidden=\"true\" href=\"#alternative-projects\"><\/a>Alternative Projects<\/h2>\n<p>I believe that deadgrep is the best tool for doing Emacs text\nsearches, but there are some other great tools out there. See\n<a href=\"https:\/\/github.com\/Wilfred\/deadgrep\/blob\/master\/docs\/ALTERNATIVES.md\">ALTERNATIVES<\/a> for a discussion.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:27+0000","updated_at":"2018-11-07T16:22:27+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/70800?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":73,"title":"1200x825, 9.7inch E-Ink display HAT for Raspberry Pi, IT8951 controller, USB\/SPI\/I80\/I2C interface","url":"https:\/\/www.waveshare.com\/9.7inch-e-paper-hat.htm","content":"<strong>SKU: <\/strong>15195<br \/><strong>Part Number: <\/strong>9.7inch e-Paper HAT<br \/><strong>Brand: <\/strong>Waveshare<div class=\"price-review\"><div class=\"add-to-box choose_add-to-box\"><p>\n    \n                            \n     \n                            \n        \n                                        \n                                            $169.99                                    \n                            \n    \n            \n                <label for=\"qty\">Qty:<\/label>\n        \n        \n        \n                <\/p><\/div><\/div><div class=\"add-to-box choose_add-to-box\"><p><strong>Select Version: <\/strong><\/p><div class=\"clearfix p-choose-wrap\"><div class=\"li choose-color-shouji p-choose\"><div class=\"dd\"><div class=\"item selected\"><a href=\"#none;\" title=\"1200x825, 9.7inch E-Ink display HAT for Raspberry Pi\">&#13;\n                                            <img width=\"60\" height=\"60\" class=\"img_size\" data-img=\"1\" src=\"https:\/\/www.waveshare.com\/media\/catalog\/product\/cache\/1\/thumbnail\/75x75\/9df78eab33525d08d6e5fb8d27136e95\/9\/_\/9.7inch-e-paper-hat-1.jpg\" alt=\"1200x825, 9.7inch E-Ink display HAT for Raspberry Pi\" \/><\/a><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:22:26+0000","updated_at":"2018-11-07T16:22:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.waveshare.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":74,"title":"The Federation - a statistics hub","url":"https:\/\/the-federation.info\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:22:26+0000","updated_at":"2018-11-07T16:22:26+0000","annotations":[],"reading_time":0,"domain_name":"the-federation.info","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":75,"title":"Inklet Trackpad Tablet for creativity on the go","url":"https:\/\/tenonedesign.com\/inklet.php","content":"<h2>How do I install and run Inklet?<\/h2><p><b>Step 1:<\/b> When it\u2019s done downloading, drag it to your Applications folder.<br \/><b>Step 2:<\/b> Launch Inklet. You will see a black droplet on the toolbar in the upper right-hand corner or your screen. Inklet will run in the background while you go about your business elsewhere on your laptop.<br \/><\/p><h2>How do I use Inklet with Photoshop or another drawing program?<\/h2><p><b>Step 1:<\/b> Open a drawing program such as Photoshop.<br \/><b>Step 2:<\/b> When you're ready to start Inklet, click on the droplet icon and select 'Start Inklet,' or use the hotkey (the default hotkey is control+option+i). Draw with your Pogo on the trackpad. If you don't have a Pogo, you can change the drawing tool to 'finger' in the Inklet drawing preferences.<br \/><b>Step 3:<\/b> Move your Pogo lightly around on the trackpad to allow the cursor to hover without inking. Increase pressure to draw. heavier pressure will result in a thicker stroke.<\/p><h2>I\u2019m drawing, but nothing is showing up?!<\/h2><p>Make sure you have a drawing program open under Inklet.\n\t\t\t\t Compatible programs are listed a few questions down.\n\t\t\t\t Also, make sure you have a pen tool selected in your program, and you\u2019re not trying to draw with white on a white background.\n\t\t\t\t That never works.<\/p><h2>How do I register Inklet?<\/h2><p>You will be given an activation key in your purchase confirmation email.\n\t\t\t\tActivating with this key will remove the demo message in Inklet.\n\t\t\t\tClick the black inklet icon in your top right toolbar, and select 'Preferences...'\n\t\t\t\tNow, select the Register tab, and paste your key into the text field.\n\t\t\t\tIf you don't see your confirmation email, make sure to check your spam folder.<\/p><h2>Can I draw without a Pogo?<\/h2><p>Yes, just change the drawing tool to 'finger' in the Inklet drawing preferences.\n\t\t\t\t It\u2019s not as easy, but it\u2019s way possible.<\/p><h2>Can I use Inklet with an external Magic Trackpad?<\/h2><p>Yes. If an internal trackpad is not available, Inklet will automatically detect and use a connected Magic Trackpad.\n\t\t\t\tInklet works best on internal trackpads, as pressure sensitivity is limited on the Magic Trackpad.<\/p><h2>How might I uninstall Inklet?<\/h2><p>Drag Inklet from your Applications folder into the trash<\/p><p>If you don\u2019t see an answer to your question here, <a href=\"https:\/\/tenonedesign.com\/about.php\">let us know!<\/a><\/p>","created_at":"2018-11-07T16:22:26+0000","updated_at":"2018-11-07T16:22:26+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"tenonedesign.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":72,"title":"1200x825, 9.7inch E-Ink display HAT for Raspberry Pi, IT8951 controller, USB\/SPI\/I80\/I2C interface","url":"https:\/\/www.waveshare.com\/9.7inch-e-paper-hat.htm","content":"<strong>SKU: <\/strong>15195<br \/><strong>Part Number: <\/strong>9.7inch e-Paper HAT<br \/><strong>Brand: <\/strong>Waveshare<div class=\"price-review\"><div class=\"add-to-box choose_add-to-box\"><p>\n    \n                            \n     \n                            \n        \n                                        \n                                            $169.99                                    \n                            \n    \n            \n                <label for=\"qty\">Qty:<\/label>\n        \n        \n        \n                <\/p><\/div><\/div><div class=\"add-to-box choose_add-to-box\"><p><strong>Select Version: <\/strong><\/p><div class=\"clearfix p-choose-wrap\" id=\"choose\"><div class=\"li choose-color-shouji p-choose\" id=\"choose-color\"><div class=\"dd\"><div class=\"item selected\"><a href=\"#none;\" title=\"1200x825, 9.7inch E-Ink display HAT for Raspberry Pi\">&#13;\n                                            <img width=\"60\" height=\"60\" class=\"img_size\" data-img=\"1\" src=\"https:\/\/www.waveshare.com\/media\/catalog\/product\/cache\/1\/thumbnail\/75x75\/9df78eab33525d08d6e5fb8d27136e95\/9\/_\/9.7inch-e-paper-hat-1.jpg\" alt=\"1200x825, 9.7inch E-Ink display HAT for Raspberry Pi\" \/><\/a><\/div><\/div><\/div><\/div><\/div>","created_at":"2018-11-07T16:22:24+0000","updated_at":"2018-11-07T16:22:24+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"www.waveshare.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":69,"title":"A tour of JavaScript timers on the web","url":"https:\/\/nolanlawson.com\/2018\/09\/01\/a-tour-of-javascript-timers-on-the-web\/","content":"<p>Pop quiz: what is the difference between these JavaScript timers?<\/p><ul><li>Promises<\/li>\n<li><code>setTimeout<\/code><\/li>\n<li><code>setInterval<\/code><\/li>\n<li><code>setImmediate<\/code><\/li>\n<li><code>requestAnimationFrame<\/code><\/li>\n<li><code>requestIdleCallback<\/code><\/li>\n<\/ul><p>More specifically, if you queue up all of these timers at once, do you have any idea which order they\u2019ll fire in?<\/p><p>If not, you\u2019re probably not alone. I\u2019ve been doing JavaScript and web programming for years, I\u2019ve worked for a browser vendor for two of those years, and it\u2019s only recently that I really came to understand all these timers and how they play together.<\/p><p>In this post, I\u2019m going to give a high-level overview of how these timers work, and when you might want to use them. I\u2019ll also cover the <a href=\"http:\/\/lodash.com\/\">Lodash<\/a> functions <code>debounce()<\/code> and <code>throttle()<\/code>, because I find them useful as well.<\/p><h2>Promises and microtasks<\/h2><p>Let\u2019s get this one out of the way first, because it\u2019s probably the simplest. A <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Global_Objects\/Promise\">Promise<\/a> callback is also called a \u201cmicrotask,\u201d and it runs at the same frequency as <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/MutationObserver\">MutationObserver<\/a> callbacks. Assuming <a href=\"https:\/\/github.com\/whatwg\/html\/commit\/9d7cf125f960e6bb8d9b7c9456595f505f2e9d4b\">queueMicrotask()<\/a> ever makes it out of spec-land and into browser-land, it will also be the same thing.<\/p><p><a href=\"https:\/\/pouchdb.com\/2015\/05\/18\/we-have-a-problem-with-promises.html\">I\u2019ve already written a lot about promises<\/a>. One quick misconception about promises that\u2019s worth covering, though, is that they don\u2019t give the browser a chance to breathe. Just because you\u2019re queuing up an asynchronous callback, that doesn\u2019t mean that the browser can render, or process input, or do any of the stuff we want browsers to do.<\/p><p>For example, let\u2019s say we have a function that blocks the main thread for 1 second:<\/p><pre class=\"brush: jscript; title: ; notranslate\" title=\"\">\nfunction block() {\n  var start = Date.now()\n  while (Date.now() - start &lt; 1000) { \/* wheee *\/ }\n}\n<\/pre><p>If we were to queue up a bunch of microtasks to call this function:<\/p><pre class=\"brush: jscript; title: ; notranslate\" title=\"\">\nfor (var i = 0; i &lt; 100; i++) {\n  Promise.resolve().then(block)\n}\n<\/pre><p>This would block the browser for about 100 seconds. It\u2019s basically the same as if we had done:<\/p><pre class=\"brush: jscript; title: ; notranslate\" title=\"\">\nfor (var i = 0; i &lt; 100; i++) {\n  block()\n}\n<\/pre><p>Microtasks execute immediately after any synchronous execution is complete. There\u2019s no chance to fit in any work between the two. So if you think you can break up a long-running task by separating it into microtasks, then it won\u2019t do what you think it\u2019s doing.<\/p><h2>setTimeout and setInterval<\/h2><p>These two are cousins: <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/WindowOrWorkerGlobalScope\/setTimeout\"><code>setTimeout<\/code><\/a> queues a task to run in <em>x<\/em> number of milliseconds, whereas <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/WindowOrWorkerGlobalScope\/setInterval\"><code>setInterval<\/code><\/a> queues a recurring task to run every <em>x<\/em> milliseconds.<\/p><p>The thing is\u2026 browsers don\u2019t really respect that milliseconds thing. You see, historically, web developers have abused <code>setTimeout<\/code>. <em>A lot.<\/em> To the point where browsers have had to add mitigations for <code>setTimeout(\/* ... *\/, 0)<\/code> to avoid locking up the browser\u2019s main thread, because a lot of websites tended to throw around <code>setTimeout(0)<\/code> like confetti.<\/p><p>This is the reason that a lot of the tricks in <a href=\"https:\/\/www.crashmybrowser.com\/\">crashmybrowser.com<\/a> don\u2019t work anymore, such as queuing up a <code>setTimeout<\/code> that calls two more <code>setTimeout<\/code>s, which call two more <code>setTimeout<\/code>s, etc. I covered a few of these mitigations from the Edge side of things in <a href=\"https:\/\/blogs.windows.com\/msedgedev\/2017\/06\/01\/input-responsiveness-event-loop-microsoft-edge\/\">\u201cImproving input responsiveness in Microsoft Edge\u201d<\/a>.<\/p><p>Broadly speaking, a <code>setTimeout(0)<\/code> doesn\u2019t really run in zero milliseconds. Usually, it runs in 4. Sometimes, it may run in 16 (this is what Edge does when it\u2019s on battery power, for instance). Sometimes it may be clamped to 1 second (e.g., <a href=\"https:\/\/github.com\/WICG\/interventions\/issues\/5\">when running in a background tab<\/a>). These are the sorts of tricks that browsers have had to invent to prevent runaway web pages from chewing up your CPU doing useless <code>setTimeout<\/code> work.<\/p><p>So that said, <code>setTimeout<\/code> does allow the browser to run some work before the callback fires (unlike microtasks). But if your goal is to allow input or rendering to run before the callback, <code>setTimeout<\/code> is usually not the best choice because it only incidentally allows those things to happen. Nowadays, there are better browser APIs that can hook more directly into the browser\u2019s rendering system.<\/p><h2>setImmediate<\/h2><p>Before moving on to those \u201cbetter browser APIs,\u201d it\u2019s worth mentioning this thing. <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Window\/setImmediate\"><code>setImmediate<\/code><\/a> is, for lack of a better word \u2026 weird. If you look it up on <a href=\"https:\/\/caniuse.com\/#feat=setimmediate\">caniuse.com<\/a>, you\u2019ll see that only Microsoft browsers support it. And yet <a href=\"https:\/\/nodejs.org\/api\/timers.html#timers_setimmediate_callback_args\">it also exists in Node.js<\/a>, and has lots of <a href=\"https:\/\/www.npmjs.com\/search?q=setimmediate\">\u201cpolyfills\u201d on npm<\/a>. What the heck is this thing?<\/p><p><code>setImmediate<\/code> was originally proposed by Microsoft to get around the problems with <code>setTimeout<\/code> described above. Basically, <code>setTimeout<\/code> had been abused, and so the thinking was that we can create a new thing to allow <code>setImmediate(0)<\/code> to actually be <code>setImmediate(0)<\/code> and not this funky \u201cclamped to 4ms\u201d thing. You can see <a href=\"https:\/\/lists.w3.org\/Archives\/Public\/public-web-perf\/2011Jul\/0012.html\">some discussion about it from Jason Weber back in 2011<\/a>.<\/p><p>Unfortunately, <code>setImmediate<\/code> was only ever adopted by IE and Edge. Part of the reason it\u2019s still in use is that it has a sort of superpower in IE, where it allows input events like keyboard and mouseclicks to \u201cjump the queue\u201d and fire before the <code>setImmediate<\/code> callback is executed, whereas IE doesn\u2019t have the same magic for <code>setTimeout<\/code>. (Edge eventually fixed this, as detailed in the previously-mentioned post.)<\/p><p>Also, the fact that <code>setImmediate<\/code> exists in Node means that a lot of \u201cNode-polyfilled\u201d code is using it in the browser without really knowing what it does. It doesn\u2019t help that the differences between Node\u2019s <code>setImmediate<\/code> and <code>process.nextTick<\/code> are very confusing, and even <a href=\"https:\/\/nodejs.org\/en\/docs\/guides\/event-loop-timers-and-nexttick\/\">the official Node docs<\/a> say the names should really be reversed. (For the purposes of this blog post though, I\u2019m going to focus on the browser rather than Node because I\u2019m not a Node expert.)<\/p><p>Bottom line: use <code>setImmediate<\/code> if you know what you\u2019re doing and you\u2019re trying to optimize input performance for IE. If not, then just don\u2019t bother. (Or only use it in Node.)<\/p><h2>requestAnimationFrame<\/h2><p>Now we get to the most important <code>setTimeout<\/code> replacement, a timer that actually hooks into the browser\u2019s rendering loop. By the way, if you don\u2019t know how the browser event loops works, I strongly recommend <a href=\"https:\/\/youtu.be\/cCOL7MC4Pl0\">this talk by Jake Archibald<\/a>. Go watch it, I\u2019ll wait.<\/p><p>Okay, now that you\u2019re back, <code>requestAnimationFrame<\/code> basically works like this: it\u2019s sort of like a <code>setTimeout<\/code>, except instead of waiting for some unpredictable amount of time (4 milliseconds, 16 milliseconds, 1 second, etc.), it executes <em>before the browser\u2019s next style\/layout calculation step<\/em>. Now, as Jake points out in his talk, there is a minor wrinkle in that it actually executes <em>after<\/em> this step in Safari, IE, and Edge &lt;18, but let's ignore that for now since it's usually not an important detail.<\/p><p>The way I think of <code>requestAnimationFrame<\/code> is this: whenever I want to do some work that I know is going to modify the browser's style or layout \u2013 for instance, changing CSS properties or starting up an animation \u2013 I stick it in a <code>requestAnimationFrame<\/code> (abbreviated to <code>rAF<\/code> from here on out). This ensures a few things:<\/p><ol><li>I'm less likely to <a href=\"https:\/\/developers.google.com\/web\/fundamentals\/performance\/rendering\/avoid-large-complex-layouts-and-layout-thrashing\">layout thrash<\/a>, because all of the changes to the DOM are being queued up and coordinated.<\/li>\n<li>My code will naturally adapt to the performance characteristics of the browser. For instance, if it's a low-cost device that is struggling to render some DOM elements, <code>rAF<\/code> will naturally slow down from the usual 16.7ms intervals (on 60 Hertz screens) and thus it won't bog down the machine in the same way that running a lot of <code>setTimeout<\/code>s or <code>setInterval<\/code>s might.<\/li>\n<\/ol><p>This is why animation libraries that don't rely on CSS transitions or keyframes, such as <a href=\"https:\/\/greensock.com\/\">GreenSock<\/a> or <a href=\"https:\/\/github.com\/chenglou\/react-motion\">React Motion<\/a>, will typically make their changes in a <code>rAF<\/code> callback. If you're animating an element between <code>opacity: 0<\/code> and <code>opacity: 1<\/code>, there's no sense in queuing up a billion callbacks to animate every possible intermediate state, including <code>opacity: 0.0000001<\/code> and <code>opacity: 0.9999999<\/code>.<\/p><p>Instead, you're better off just using <code>rAF<\/code> to let the browser tell you how many frames you're able to paint during a given period of time, and calculate the \"tween\" for that particular frame. That way, slow devices naturally end up with a slower framerate, and faster devices end up with a faster framerate, which wouldn't necessarily be true if you used something like <code>setTimeout<\/code>, which operates independently of the browser's rendering speed.<\/p><h2>requestIdleCallback<\/h2><p><code>rAF<\/code> is probably the most useful timer in the toolkit, but <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Background_Tasks_API\"><code>requestIdleCallback<\/code><\/a> is worth talking about as well. The <a href=\"https:\/\/caniuse.com\/#feat=requestidlecallback\">browser support isn't great<\/a>, but there's a <a href=\"https:\/\/www.npmjs.com\/package\/requestidlecallback\">polyfill that works just fine<\/a> (and it uses <code>rAF<\/code> under the hood).<\/p><p>In many ways <code>rAF<\/code> is similar to <code>requestIdleCallback<\/code>. (I'll abbreviate it to <code>rIC<\/code> from now on. Starting to sound like a pair of troublemakers from West Side Story, huh? \"There go Rick and Raff, up to no good!\")<\/p><p>Like <code>rAF<\/code>, <code>rIC<\/code> will naturally adapt to the browser's performance characteristics: if the device is under heavy load, <code>rIC<\/code> may be delayed. The difference is that <code>rIC<\/code> fires on the browser \"idle\" state, i.e. when the browser has decided it doesn't have any tasks, microtasks, or input events to process, and you're free to do some work. It also gives you a \"deadline\" to track how much of your budget you're using, which is a nice feature.<\/p><p>Dan Abramov has <a href=\"https:\/\/youtu.be\/v6iR3Zk4oDY\">a good talk from JSConf Iceland 2018<\/a> where he shows how you might use <code>rIC<\/code>. In the talk, he has a webapp that calls <code>rIC<\/code> for every keyboard event while the user is typing, and then it updates the rendered state inside of the callback. This is great because a fast typist can cause many <code>keydown<\/code>\/<code>keyup<\/code> events to fire very quickly, but you don't necessarily want to update the rendered state of the page for every keypress.<\/p><p>Another good example of this is a \u201cremaining character count\u201d indicator on Twitter or Mastodon. I use <code>rIC<\/code> for this in <a href=\"https:\/\/pinafore.social\">Pinafore<\/a>, because I don't really care if the indicator updates for every single key that I type. If I'm typing quickly, it's better to prioritize input responsiveness so that I don't lose my sense of flow.<\/p><div data-shortcode=\"caption\" id=\"attachment_6665\" class=\"wp-caption aligncenter\"><img data-attachment-id=\"6665\" data-permalink=\"https:\/\/nolanlawson.com\/2018\/09\/01\/a-tour-of-javascript-timers-on-the-web\/screenshot-2018-09-01-13-59-21\/\" data-orig-file=\"https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=570&amp;h=224\" data-orig-size=\"590,232\" data-comments-opened=\"1\" data-image-meta=\"{&quot;aperture&quot;:&quot;0&quot;,&quot;credit&quot;:&quot;&quot;,&quot;camera&quot;:&quot;&quot;,&quot;caption&quot;:&quot;&quot;,&quot;created_timestamp&quot;:&quot;0&quot;,&quot;copyright&quot;:&quot;&quot;,&quot;focal_length&quot;:&quot;0&quot;,&quot;iso&quot;:&quot;0&quot;,&quot;shutter_speed&quot;:&quot;0&quot;,&quot;title&quot;:&quot;&quot;,&quot;orientation&quot;:&quot;0&quot;}\" data-image-title=\"Screenshot of Pinafore with some text entered in the text box and a digit counter showing the number of remaining characters\" data-image-description=\"\" data-medium-file=\"https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=570&amp;h=224?w=300\" data-large-file=\"https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=570&amp;h=224?w=570\" src=\"https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=570&amp;h=224\" alt=\"Screenshot of Pinafore with some text entered in the text box and a digit counter showing the number of remaining characters\" width=\"570\" height=\"224\" class=\"size-full wp-image-6665\" srcset=\"https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=570&amp;h=224 570w, https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=150&amp;h=59 150w, https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png?w=300&amp;h=118 300w, https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png 590w\" \/><p class=\"wp-caption-text\">In Pinafore, the little horizontal bar and the \u201ccharacters remaining\u201d indicator update as you type.<\/p><\/div><p>One thing I\u2019ve noticed about <code>rIC<\/code>, though, is that it\u2019s a little finicky in Chrome. In Firefox it seems to fire whenever I would, intuitively, think that the browser is \u201cidle\u201d and ready to run some code. (Same goes for the polyfill.) In mobile Chrome for Android, though, I\u2019ve noticed that whenever I scroll with touch scrolling, it might delay <code>rIC<\/code> for several seconds even after I\u2019m done touching the screen and the browser is doing absolutely nothing. (I suspect the issue I\u2019m seeing is <a href=\"https:\/\/bugs.chromium.org\/p\/chromium\/issues\/detail?id=811451\">this one<\/a>.)<\/p><p><em><strong>Update:<\/strong> Alex Russell from the Chrome team <a href=\"https:\/\/toot.cafe\/@slightlyoff\/100655566963584982\">informs me<\/a> that this is a known issue and should be fixed soon!<\/em><\/p><p>In any case, <code>rIC<\/code> is another great tool to add to the tool chest. I tend to think of it this way: use <code>rAF<\/code> for critical rendering work, use <code>rIC<\/code> for non-critical work.<\/p><h2>debounce and throttle<\/h2><p>These two functions aren\u2019t built in to the browser, but they\u2019re so useful that they\u2019re worth calling out on their own. If you aren\u2019t familiar with them, there\u2019s <a href=\"https:\/\/css-tricks.com\/debouncing-throttling-explained-examples\/\">a good breakdown in CSS Tricks<\/a>.<\/p><p>My standard use for <a href=\"https:\/\/lodash.com\/docs\/4.17.10#debounce\"><code>debounce<\/code><\/a> is inside of a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Events\/resize\"><code>resize<\/code><\/a> callback. When the user is resizing their browser window, there\u2019s no point in updating the layout for every <code>resize<\/code> callback, because it fires too frequently. Instead, you can <code>debounce<\/code> for a few hundred milliseconds, which will ensure that the callback eventually fires once the user is done fiddling with their window size.<\/p><p><a href=\"https:\/\/lodash.com\/docs\/4.17.10#throttle\"><code>throttle<\/code><\/a>, on the other hand, is something I use much more liberally. For instance, a good use case is inside of a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/GlobalEventHandlers\/onscroll\"><code>scroll<\/code><\/a> event. Once again, it\u2019s usually senseless to try to update the rendered state of the app for every <code>scroll<\/code> callback, because it fires too frequently (and the frequency can vary from browser to browser and from input method to input method\u2026 ugh). Using <code>throttle<\/code> normalizes this behavior, and ensures that it only fires every <em>x<\/em> number of milliseconds. You can also tweak Lodash\u2019s <code>throttle<\/code> (or <code>debounce<\/code>) function to fire at the start of the delay, at the end, both, or neither.<\/p><p>In contrast, I wouldn\u2019t use <code>debounce<\/code> for the scrolling scenario, because I don\u2019t want the UI to only update after the user has explicitly stopped scrolling. That can get annoying, or even confusing, because the user might get frustrated and try to keep scrolling in order to update the UI state (e.g. in an infinite-scrolling list). <code>throttle<\/code> is better in this case, because it doesn\u2019t wait for the <code>scroll<\/code> event to stop firing.<\/p><p><code>throttle<\/code> is a function I use all over the place for all kinds of user input, and even for some regularly-scheduled tasks like IndexedDB cleanups. It\u2019s extremely useful. Maybe it should just be baked into the browser some day!<\/p><h2>Conclusion<\/h2><p>So that\u2019s my whirlwind tour of the various timer functions available in the browser, and how you might use them. I probably missed a few, because there are certainly some exotic ones out there (<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/API\/Window\/postMessage\"><code>postMessage<\/code><\/a> or <a href=\"https:\/\/developers.google.com\/web\/updates\/2018\/07\/page-lifecycle-api\">lifecycle events<\/a>, anyone?). But hopefully this at least provides a good overview of how I think about JavaScript timers on the web.<\/p>","created_at":"2018-11-07T16:22:23+0000","updated_at":"2018-11-07T16:22:23+0000","published_at":"2018-09-01T22:10:42+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":11,"domain_name":"nolanlawson.com","preview_picture":"https:\/\/nolanwlawson.files.wordpress.com\/2018\/09\/screenshot-2018-09-01-13-59-21.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":70,"title":"Research Literature Management with Emacs","url":"https:\/\/www.anand-iyer.com\/blog\/2017\/research-literature-management-with-emacs.html","content":"<p>A major activity in the daily routine of a graduate student is reading a lot of\nliterature. Obviously, this results in a huge pile of papers to manage over time\nand hence it is unsurprising that we seek out a literature management solution\natleast once during our graduate life. Fortunately, there\nare\n<a href=\"http:\/\/www.mendeley.com\">several<\/a> <a href=\"http:\/\/www.readcube.com\">excellent<\/a> <a href=\"http:\/\/www.zotero.org\">solutions<\/a> today\nthat addresses this problem. I\u2019ve tried many of them, but finally settled\non something that works for me. This post describes my current workflow;\nhopefully it will be useful for someone else too.<\/p><p>At a high-level, my literature management system is really simple. I keep a\nbibliography file to which I add papers that I find interesting. For each entry\nin this file, I link two things: the PDF of the paper, and notes on the paper.\nWith this setting, I can quickly search for an entry, find the PDF and\/or notes\nfor the matching entry. All this is made possible by Emacs, with the help of a\nfew packages. Let me explain this in a bit of details.<\/p><h3 id=\"basic-setup\">Basic Setup<\/h3><p>The first thing to do is to set up a few basic things. As I mentioned, I have\none bibliography file where I want new entries to go. I also need a folder to\nstore all the PDFs and a way to store notes on each of these entries. For this\ntask, I leverage the <a href=\"https:\/\/github.com\/jkitchin\/org-ref\"><strong>org-ref<\/strong><\/a> package.\nHere\u2019s my org-ref setup:<\/p><figure class=\"highlight\"><pre class=\"language-lisp\" data-lang=\"lisp\">(setq org-ref-bibliography-notes \"~\/org\/ref\/notes.org\"\n      org-ref-default-bibliography '(\"~\/org\/ref\/master.bib\")\n      org-ref-pdf-directory \"~\/org\/ref\/pdfs\/\")<\/pre><\/figure><p>The code should be self-explanatory: bibliography entries are stored in a file\nnamed <code>master.bib<\/code>, while PDF files corresponding to the entries go in the\n<code>pdfs<\/code> folder. I chose to store my notes in a single org file <code>notes.org<\/code>. Note\nthat it is entirely possible to have separate note files for every entry in your\nbibliography file. My <code>org<\/code> folder is synced with a cloud storage service.<\/p><h3 id=\"adding-entries\">Adding Entries<\/h3><p>When I come across a paper that I\u2019m interested in reading, I add it to\n<code>master.bib<\/code>. Most of the time, I can <em>simply drag-and-drop the paper link or\nDOI<\/em> to the bibliography file and let org-ref take care of creating the entry\nand downloading the PDF. Otherwise, I simply copy the bibtex entry and then run\n<code>doi-utils-get-bibtex-entry-pdf<\/code> to get the PDF.<\/p><p>Once I decide to read (and make notes on) a paper that is in my bibliography\nfile, I then proceed to make an entry for it in my <code>notes.org<\/code> file. I do this in\neither of two ways: if I have the <code>master.bib<\/code> file open, I can simply go to the\nentry I want to read, and then use the <code>org-ref-open-bibtex-notes<\/code> command. As\nthe name suggests, this command opens the corresponding notes entry in the\n<code>org-ref-bibliography-notes<\/code> file, creating a new entry if it can\u2019t find one. If\nI don\u2019t have the <code>master.bib<\/code> file open, I use the excellent <code>helm-bibtex<\/code>\npackage to search through my bibliography files. When I find the entry I\u2019m\nlooking for, I open notes (or create it) with helm\u2019s available\nactions on the selected entry. For this to work, we\nneed to let <code>helm-bibtex<\/code> know where the notes and PDF files are:<\/p><figure class=\"highlight\"><pre class=\"language-lisp\" data-lang=\"lisp\">(setq bibtex-completion-bibliography \"~\/org\/ref\/master.bib\")\n      bibtex-completion-library-path \"~\/org\/ref\/pdfs\"\n      bibtex-completion-notes-path \"~\/org\/ref\/notes.org\") <\/pre><\/figure><p>You can customize the fields in the entry created in <code>notes.org<\/code> by both\n<code>org-ref<\/code> and <code>helm-bibtex<\/code> to your liking.<\/p><h3 id=\"taking-notes\">Taking Notes<\/h3><p>Now that I have an entry in my <code>notes.org<\/code> file for the paper that I want to\nread, the next step is to take notes as I read. You can simply type down your\nthoughts below the entry in <code>notes.org<\/code>, but I like seeing the notes\nside-by-side with the page on which it was made. For this purpose, I use\nthe <a href=\"https:\/\/github.com\/rudolfochrist\/interleave\"><strong>interleave<\/strong><\/a> package. To\nuse the <code>interleave<\/code> package, all I need to do is to add the <code>INTERLEAVE_PDF<\/code>\nproperty to the <code>PROPERTIES<\/code> section of the entry in my <code>notes.org<\/code> file. Then,\nto open the PDF, I place the cursor on the title, and invoke <code>interleave-mode<\/code>\ncommand. This will display the PDF side-by-side with my notes, where I can\nnavigate, see or add notes on any page.<\/p><p>The combination of <code>org-ref<\/code>, <code>helm-bibtex<\/code> and <code>interleave<\/code> gives me great\nflexibility. I can easily search for entries in my bibliography files using\n<code>helm-bibtex<\/code>. From there, I can open the paper and\/or its notes. Since I use\nEmacs to write papers, <code>org-ref<\/code> makes it easy to insert citations, look up\nnotes and download papers as I write. Having the notes in plain text lets me\nsearch through them easily. Finally, as a bonus, <code>org<\/code> mode lets me easily\nexport my notes in several formats.<\/p><p>To summarize, having the literature management system tightly coupled with my paper\nwriting environment (Emacs with AUCTeX) worked quite well for me.<\/p>","created_at":"2018-11-07T16:22:23+0000","updated_at":"2018-11-07T16:22:23+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"www.anand-iyer.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":71,"title":"Shinmera\/qtools","url":"https:\/\/github.com\/Shinmera\/qtools\/blob\/master\/README.md","content":"<h2><a id=\"user-content-about-qtools-\" class=\"anchor\" aria-hidden=\"true\" href=\"#about-qtools-\"><\/a>About Qtools <a href=\"https:\/\/travis-ci.org\/Shinmera\/qtools\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/24e4613b5b23f1293ad992ce8ff0e6b34645a630\/68747470733a2f2f7472617669732d63692e6f72672f5368696e6d6572612f71746f6f6c732e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/Shinmera\/qtools.svg?branch=master\" style=\"float: right;\" \/><\/a><\/h2>\n<p>Qtools is a collection of utilities to help development with CommonQt. There are a lot of things in there, from name mapping over garbage handling to widget class definition. Some tools are straightforward, others are quite complex. I'll try to explain everything as best I can.<\/p>\n<h2><a id=\"user-content-fundamentals\" class=\"anchor\" aria-hidden=\"true\" href=\"#fundamentals\"><\/a>Fundamentals<\/h2>\n<p>There are three layers working beneath Qtools, of which you need to know about in order to understand how to use the various facilities Qtools offers.<\/p>\n<h3><a id=\"user-content-qt\" class=\"anchor\" aria-hidden=\"true\" href=\"#qt\"><\/a><a href=\"http:\/\/qt-project.org\/doc\/qt-4.8\/\" rel=\"nofollow\">Qt<\/a><\/h3>\n<p>Qt is a gigantic toolkit that mainly concerns itself with GUI creation. It is cross-platform and written in C++. While its size and large user-base means that it is very mature and usable for pretty much any kind of GUI, the fact that it's written in C++ makes things complicated. The usual way to interoperate with C++ libraries and projects is to create a C API wrapper.<\/p>\n<h3><a id=\"user-content-smoke\" class=\"anchor\" aria-hidden=\"true\" href=\"#smoke\"><\/a><a href=\"https:\/\/techbase.kde.org\/Development\/Languages\/Smoke\" rel=\"nofollow\">Smoke<\/a><\/h3>\n<p>And this is what Smoke does. It generates a C wrapper so that other languages can make use of Qt through C-FFI. Smoke is divided up into a bunch of modules, each being associated with a part of Qt, such as qtcore, qtgui, qtopengl, phonon, etc.<\/p>\n<h3><a id=\"user-content-commonqt\" class=\"anchor\" aria-hidden=\"true\" href=\"#commonqt\"><\/a><a href=\"http:\/\/common-lisp.net\/project\/commonqt\/\" rel=\"nofollow\">CommonQt<\/a><\/h3>\n<p>The heavy lifting and ground-work that is required to interface with Smoke (and thus with Qt) is done by CommonQt. By itself, CommonQt offers a complete framework to create Qt GUIs from Common Lisp out.<\/p>\n<h3><a id=\"user-content-qtools\" class=\"anchor\" aria-hidden=\"true\" href=\"#qtools\"><\/a><a href=\"https:\/\/shinmera.github.io\/qtools\/\" rel=\"nofollow\">Qtools<\/a><\/h3>\n<p>Unfortunately, working with CommonQt itself is a bit awkward. While it offers everything you need, the way you have to make use of it is sub-par. Qtools attempts to help with this by offering another layer over CommonQt to smooth everything out. However, since you might not like certain parts of the Qtools ecosystem, it should be possible for you to only use the features you like, rather than being forced to use everything. So, you can always mix and match \"plain\" CommonQt and Qtools extensions.<\/p>\n<h2><a id=\"user-content-getting-started\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-started\"><\/a>Getting Started<\/h2>\n<h3><a id=\"user-content-qt-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#qt-1\"><\/a>Qt<\/h3>\n<p>In case you already know Qt, you may skip this part and go straight to the next section. I'll briefly explain some of the core Qt concepts and how they fit together. However, you should still read the <a href=\"http:\/\/doc.qt.io\/qt-4.8\/\" rel=\"nofollow\">official Qt documentation<\/a> for anything that's left unclear.<\/p>\n<p>Since Qt is a C++ library it is heavily focused around classes and methods. At the top of the class hierarchy sits <a href=\"http:\/\/doc.qt.io\/qt-4.8\/qobject.html\" rel=\"nofollow\">QObject<\/a>. Almost all classes in Qt inherit from this. The more important one for us however is the <a href=\"http:\/\/doc.qt.io\/qt-4.8\/qwidget.html\" rel=\"nofollow\">QWidget<\/a> class, which already gives you everything you need to get started. It can act as a window or be contained in other widgets. It has all the sizing mechanics and is properly hooked into the event system that allows you to catch when keys are pressed, the window is resized, and so forth.<\/p>\n<p>Usually you create your new widgets by subclassing something like QWidget and then overriding methods to do what you want or simply attach other widgets to it and connect them up so that they can communicate to each other via the signal system. Classes have slots and signals. Signals are a form of method signature that describe a list of arguments and their type. Slots are the corresponding methods that perform actions when a signal reaches them. Wiring things up functions by connecting a signal and a source instance to a slot and a target instance. When then a signal is signalled onto the source instance, the target instance's slot method is called.<\/p>\n<p>And that's pretty much all of the core mechanics that you need to know about. For everything else, you usually refer to the documentation page for the specific class involved and see what kind of methods or signals you can use to do what you want. Seriously, the docs are good. I use them constantly while developing.<\/p>\n<h3><a id=\"user-content-qtools-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#qtools-1\"><\/a>Qtools<\/h3>\n<p>Before getting started with explaining the details on the various parts of Qtools, I'll go through a basic project setup of a project using it.<\/p>\n<p>First you will want an ASDF system to define your project. In its dependencies it should contain <code>:qtools<\/code> and the smoke modules you require, usually just <code>:qtcore<\/code> and <code>:qtgui<\/code>.<\/p>\n<pre>(asdf:defsystem qtools-intro\n  ...\n  :depends-on (:qtools :qtcore :qtgui))\n<\/pre>\n<p>If you want to just try things out on the REPL real quick, just <code>(ql:quickload '(qtools qtcore qtgui))<\/code> should suffice.<\/p>\n<p>Then of course you'll usually want a package to put all your stuff in. Instead of <code>:use #:cl<\/code> you will most likely want to <code>:use #:cl+qt<\/code>. This package combines the symbols of the common-lisp base package, the commonqt package, and qtools. Thus all the symbols and functions you usually need for development are already included.<\/p>\n<pre>(defpackage #:qtools-intro\n  (:use #:cl+qt)\n  (:export #:main))\n<\/pre>\n<p>CommonQt, and Qtools itself, require a few extensions to the standard reader syntax. For this reason you will want to change the readtable using <code>named-readtables:in-readtable<\/code>, which <code>cl+qt<\/code> includes. Readtable changes are on a <em>per-file<\/em> basis, so you need both an <code>in-package<\/code> and an <code>in-readtable<\/code> call on every file.<\/p>\n<pre>(in-package #:qtools-intro)\n(in-readtable :qtools)\n<\/pre>\n<p>This sets up everything you need to get started writing an actual GUI. So let's do that as well. In Qt things are organised as widgets. Windows are widgets, buttons are widgets, text fields are widgets, etc. Qtools mirrors this.<\/p>\n<pre>(define-widget main-window (QWidget)\n  ())\n<\/pre>\n<p>This <code>define-widget<\/code> form is syntactically equivalent to <code>defclass<\/code>. The only change is that the first argument in the superclass list is the Qt class to inherit from. Now we'll want to add some things to display in the widget.<\/p>\n<pre>(define-subwidget (main-window name) (q+:make-qlineedit main-window)\n  (setf (q+:placeholder-text name) \"Your name please.\"))\n(define-subwidget (main-window go) (q+:make-qpushbutton \"Go!\" main-window))\n<\/pre>\n<p>This adds a QLineEdit widget called <code>name<\/code> to the <code>main-window<\/code> and sets its placeholder text to <code>\"Your name please.\"<\/code>. The second form adds a button called <code>go<\/code> with a label of <code>\"Go!\"<\/code>. Simple stuff. The body of the <code>define-subwidget<\/code> form can contain any number of statements. By default, the symbols of all the other subwidgets and slots defined prior to the <code>define-subwidget<\/code> form are bound to their corresponding values. This is useful if you for example need to define a layout, as we will do now.<\/p>\n<pre>(define-subwidget (main-window layout) (q+:make-qhboxlayout main-window)\n  (q+:add-widget layout name)\n  (q+:add-widget layout go))\n<\/pre>\n<p>This sets up the displaying part of our GUI, but so far we haven't made it react to anything yet. Reacting to events in Qt happens through signals and slots. Slots are functions that receive signals, and signals are event carriers.<\/p>\n<pre>(define-signal (main-window name-set) (string))\n(define-slot (main-window go) ()\n  (declare (connected go (pressed)))\n  (declare (connected name (return-pressed)))\n  (signal! main-window (name-set string) (q+:text name)))\n(define-slot (main-window name-set) ((new-name string))\n  (declare (connected main-window (name-set string)))\n  (q+:qmessagebox-information main-window \"Greetings\" (format NIL \"Good day to you, ~a!\" new-name)))\n<\/pre>\n<p>We're doing things a bit roundabout here to illustrate creating signals. <code>define-signal<\/code> introduces a new signal called <code>name-set<\/code> that takes a single <code>string<\/code> as argument. We then define a new slot that is connected to the <code>go<\/code> button's <code>pressed<\/code> signal (which has no arguments) as well as the <code>name<\/code> field's <code>return-pressed<\/code>. We then simply fetch the current text of our <code>name<\/code> field and send it out again with our custom signal. The second slot catches this signal again and uses it to display a message box.<\/p>\n<p>You can now try and see what kind of magical wonders you have created by using <code>with-main-window<\/code> to launch everything: <code>(with-main-window (window 'main-window))<\/code>. This will block until your window is closed, as Qt needs to capture the thread to handle events.<\/p>\n<p>And that's that. The only thing we didn't take a look at here is <code>define-override<\/code>, which allows you to define override functions for your Qt classes. So, if you for example want to manually draw onto a widget you can override its <code>paintEvent<\/code> method using this.<\/p>\n<pre>(define-override (main-window paint-event) (event)\n  (declare (ignore event))\n  (with-finalizing ((painter (q+:make-qpainter main-window)))\n    (q+:fill-rect painter (q+:rect main-window) (q+:qt.white))))\n<\/pre>\n<p>This'll make the background of our window completely white. The important thing to note here is the <code>with-finalizing<\/code>. What has been rather well hidden from you so far, is that as Qt is a C++ framework, you will have to do manual memory management. Qtools makes this a lot easier by offering macros and automations to take most of it off of your hands. For example, all the sub widgets we defined are automatically deleted once the main window is.<\/p>\n<p>A general note about developing with Qtools\/CommonQt: While custom function bodies such as from qt-slots, overrides, initializers, and finalizers reside on the CL side and can thus be redefined at any time and take effect immediately, adding or removing qt-slots and overrides will not affect already created instances. This is to say, if you run your application and recompile your override, the effect will be visible immediately. But if you add a new slot, override, or signal, the existing instance will not have them. This is due to the fact that these things need to be tied to the C++ class, which will not update existing instances when it is changed, like CLOS usually does. This means that if you change a running widget by adding new components, you need to recreate or restart it to see the effects.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#deployment\"><\/a>Deployment<\/h2>\n<p>So, you've built a neat little Qt application and you would like to deploy it and ship it to people that aren't developers. For this, you will want to dump a binary and bundle it together with the necessary shared libraries. Qtools will take care of this for you, to the point where it becomes very trivial to do.<\/p>\n<p>You will need an ASDF system that compiles your program and add the following options to your system definition:<\/p>\n<pre>:defsystem-depends-on (:qtools)\n:build-operation \"qt-program-op\"\n:build-pathname \"executable-name\"\n:entry-point \"my-package:my-launch-function\"\n<\/pre>\n<p>You can change the string of <code>:build-pathname<\/code> to whatever pathname-name you'd like your executable to have. You should change the stirng of <code>:entry-point<\/code> to be a designator for an external symbol that is either your main window class name, or a function that handles the launching of your GUI.<\/p>\n<p>Once you have updated your ASD, you should launch a clean instance of your implementation from the command line -- make sure not to load slime or anything else that creates threads. Then simply invoke <code>(asdf:operate :build-op :system-name :force T)<\/code>. For sbcl, that would be:<\/p>\n<pre>sbcl --eval \"(asdf:operate :build-op :system-name :force T)\"\n<\/pre>\n<p>This will compile your system, gather some info on it, copy the necessary shared libraries to the deployment folder, and finally dump an image with your system ready to go. It should put it all into a new <code>bin<\/code> folder within your project's root. If everything went right, you should be able to just launch the executable within and be greeted with your nice GUI. You can then just ZIP up the build folder and ship that-- it should contain all of the necessary shared library dependencies to run outside of your development environment.<\/p>\n<p>As an example for a working system that can be deployed, have a look at <a href=\"https:\/\/github.com\/shinmera\/halftone\">halftone<\/a> or <a href=\"https:\/\/github.com\/shirakumo\/cl-gamepad\">cl-gamepad-visualizer<\/a>.<\/p>\n<p>For more customisation and other details of the deployment system, see the underlying library <a href=\"https:\/\/shinmera.github.io\/deploy\/\" rel=\"nofollow\">Deploy<\/a>.<\/p>\n<p>If you need to debug the built binary on Windows, the default behaviour is a bit annoying, as it causes Windows to completely prevent all terminal output. To circumvent this, you must either run it under mintty, or deploy it with the feature <code>:qtools-deploy-console<\/code> set before the build-op.<\/p>\n<h2><a id=\"user-content-qtools-components\" class=\"anchor\" aria-hidden=\"true\" href=\"#qtools-components\"><\/a>Qtools Components<\/h2>\n<h3><a id=\"user-content-name-conversion\" class=\"anchor\" aria-hidden=\"true\" href=\"#name-conversion\"><\/a>Name Conversion<\/h3>\n<p>Common Lisp and C++ follow different naming conventions for classes, variables, methods, and types. Qtools offers a couple of functions to attempt to translate between the two. Finding a Qt class can be done with <code>find-qt-class-name<\/code>. It searches a static table of known Qt classes (<code>*qt-class-map*<\/code>) by first stripping all dashes from the name, and then case-insensitively finding a matching name.<\/p>\n<p>Translating method names can be done through <code>to-method-name<\/code>, which translates dashes to mean that the next character should be in uppercase. So <code>foo-bar<\/code> is <code>fooBar<\/code>. In case you pass it a string instead of a symbol however, it will not do any translation. This is important to take care of edge-cases, where this primitive translation would prohibit using a certain name. If you need a method signature instead of just a name, there are <code>specified-type-method-name<\/code> that takes a name and a list of type specifiers, and <code>determined-type-method-name<\/code>, which attempts to determine the type of the arguments, rather than requiring type specifiers directly.<\/p>\n<p>Types are translated using <code>qt-type-of<\/code>, <code>eqt-type-of<\/code>, <code>qt-type-for<\/code>, and <code>to-type-name<\/code>. Where <code>qt-type-for<\/code> translates a type specifier and <code>qt-type-of<\/code> tries to determine the type through a value. Reversing from a Qt type specifier to a CL type is possible with <code>cl-type-for<\/code> and <code>ecl-type-for<\/code>.<\/p>\n<h3><a id=\"user-content-object-handling\" class=\"anchor\" aria-hidden=\"true\" href=\"#object-handling\"><\/a>Object Handling<\/h3>\n<p>In C++ there usually is no garbage collector, so you need to carefully clean up yourself or you'll create a memory leak. In Common Lisp we're used to the luxury of not having to worry about this, thanks to garbage collection. Sadly, we cannot use the garbage collector to also take care of C++ objects, as they live in a different world not governed by us. So, garbage collection of Qt objects is still our own worry.<\/p>\n<p>To make this all just a smidgeon easier, Qtools introduces a system of <em>finalizables<\/em>, the central point of which is the <code>finalize<\/code> generic function. This function should take care of all necessary cleanup when an object is no longer needed. For <code>qobject<\/code>s, this means running eventual cleanup (through <code>finalize-using-class<\/code>) and then being <code>delete<\/code>d, thus properly removed from memory. However, <code>finalize<\/code> can not only be used for Qt objects, but for anything else as well. Especially interesting here are <code>finalizable<\/code> objects, whose slots can specify an additional argument <code>:finalized<\/code>, which dictates whether <code>finalize<\/code> is run on the slot's value when the class instance is <code>finalize<\/code>d.<\/p>\n<p>Often times we only need Qt instances for a certain lexical context of code. For this case, <code>with-finalizing<\/code> (and <code>with-finalizing*<\/code>) can be used, which is a counterpart to <code>let<\/code> that calls <code>finalize<\/code> on all its bindings once the form exits. This will take care of most cases. For cases where the instance may escape, or has to stay bound in a closure, there's <code>with-gc-finalized<\/code>. This wraps the value of each binding in a <code>gc-finalized<\/code> object. Using such a container, we can use Common Lisp's garbage collector to keep track of the references and, once the object is garbage collected it calls <code>finalize<\/code> on its inner value, ensuring proper removal. The <code>with-gc-finalized<\/code> uses a <code>symbol-macrolet<\/code> to ensure that you don't have to worry about the boxing. However, if you manually use <code>gc-finalized<\/code> objects, there's the <code>make-gc-finalized<\/code> and <code>unbox<\/code> functions (with corresponding reader macros <code>#&gt;<\/code> and <code>#&lt;<\/code> in the <code>:qtools<\/code> read-table).<\/p>\n<p>While this kind of system takes care of a lot of cases, it's still not perfect and it may happen that you accidentally create a memory leak somewhere. I wish there was an easier way, but alas, life is difficult. You can try to solve this kind of situation by debugging Qtools and keeping track of which objects get finalized and which don't.<\/p>\n<p>Another occasional task is to copy an object instance. Qtools offers a <code>copy<\/code> and <code>copy-using-class<\/code> methods which handle proper copying for a couple of Qt objects, but sadly by far not all. If you want to use the copying system for a class that isn't handled by Qtools by default, you can define your own using <code>define-copy-method<\/code>.<\/p>\n<p>Qtools will handle the printing of QObject instances as well and provides you with a mechanism of defining print methods for a qclass, just like you normally would for CL classes by <code>print-object<\/code>. Simply use <code>define-print-method<\/code>.<\/p>\n<h3><a id=\"user-content-widgets\" class=\"anchor\" aria-hidden=\"true\" href=\"#widgets\"><\/a>Widgets<\/h3>\n<p>Qt deals with widgets. As such, making everything associated with them simple and easy to use should be a primary objective. Qtools' widget system attempts to do exactly that. The central part to this is the <code>define-widget<\/code> macro. This expands to a <code>defclass<\/code> with the following effects: It sets <code>widget-class<\/code> as the metaclass, sets the first item of the superclass list as the qt-superclass, and injects <code>widget<\/code> as a superclass, if it isn't one already. This means that essentially you can use anything you could in a standard <code>defclass<\/code> without having to worry about the necessary default options.<\/p>\n<p>However, just with a single <code>define-widget<\/code> you won't get far ahead. You still need to use CommonQt's way of declaring slots, signals, and overrides, which is quite cumbersome. You can do that of course, but there is a more convenient method, which is to use <code>define-slot<\/code>, <code>define-signal<\/code>, and <code>define-override<\/code>. These essentially translate to class options, albeit in a detached way. Each of these define statements is of the following syntax:<\/p>\n<pre>(define-* (widget-class name) arglist &amp;body body)\n<\/pre>\n<p>Some of them take optional extra arguments in the name-list, such as a method-name in the case of <code>define-slot<\/code> and <code>define-override<\/code>. In the case of <code>define-signal<\/code> the body is discarded. However, even with these extensions things are rather cumbersome: You need to manually define slots for each of the widgets you want to use inside your widget class, and define their behaviour in an <code>initialize-instance<\/code> function. This is unwieldy, which is why Qtools also adds <code>define-initializer<\/code>, <code>define-finalizer<\/code>, and <code>define-subwidget<\/code>. The first two follow in signature to the above minus the <code>arglist<\/code> and do what you might expect them to do: handle initialization and finalization. The initializer and finalizer forms take an optional priority argument in their name-list. The higher, the sooner. <code>define-subwidget<\/code> on the other hand looks like this:<\/p>\n<pre>(define-subwidget (widget-class name) initform &amp;body body)\n<\/pre>\n<p>And its effects are two-fold. First, it adds a finalized slot to the widget-class called name. Then, it adds an initializer (with priority 10) that sets the slot-value to the value of <code>initform<\/code> and then evaluates the <code>body<\/code> forms. This by itself takes care of the repetitive slot and initializer definition, but without an extra helper called <code>with-slots-bound<\/code>, it would still be annoying to write functions, as you would have to reference widgets using <code>slot-value<\/code> everywhere. <code>with-slots-bound<\/code> is like <code>with-slots<\/code>, but it binds <em>all<\/em> direct class-slot values to their respective slot names. Every <code>define-*<\/code> function's body is automatically wrapped in a <code>with-slots-bound<\/code>, to make this convenience possible.<\/p>\n<p>If you do not like this behaviour, due to potential symbol clashes and general confusion that might arise from the implicit action, you can instead use <code>cl+qt:defmethod<\/code> and <code>declare<\/code> statements. This is actually what all the <code>define-*<\/code> (with the exception of <code>define-signal<\/code>) expand to: A <code>cl+qt:defmethod<\/code> with an appropriate declaration inserted into the body. The <code>cl+qt:defmethod<\/code> behaves exaclty like <code>cl:defmethod<\/code> with the exception of allowing the handling of custom declaration forms. Qtools defines the following declarations: <code>slot<\/code>, <code>override<\/code>, <code>initializer<\/code>, and <code>finalizer<\/code>. In the case of a <code>slot<\/code> definition, an extra declaration called <code>connected<\/code> is also available. The effects of the declarations are as you might expect, and have the following signatures:<\/p>\n<pre>(slot slot-name args)\n(connected slot-name (signal-name &amp;rest args))\n(override &amp;optional method-name)\n(initializer &amp;optional (priority 0))\n(finalizer &amp;optional (priority 0))\n<\/pre>\n<p>The user may define additional declarations using <code>define-method-declaration<\/code>.<\/p>\n<p>At this point it is useful to note about the general startup sequence of a widget. Once <code>make-instance<\/code> is called, it calls to <code>initialize-instance<\/code>. Qtools defines a primary method on this specialized on the widget class. It then immediately calls the next method (<code>call-next-method<\/code>), followed by <code>construct<\/code> which should call <code>qt:new<\/code>, instantiating the C++ parts. Following that are the initializers in the order of their priority (highest goes first). This means that if you (for some reason) define an <code>:after<\/code> method on <code>initialize-instance<\/code> for your own widget, it will be run after all initializers have completed. For finalization it is the same: all finalizers are run before your own <code>finalize<\/code> method is run. However, if you define a <code>:before<\/code> method on <code>finalize<\/code> for your class, it will be run before the <code>:finalized<\/code> slots are finalized, otherwise all primary and <code>:after<\/code> methods should not access <code>:finalized<\/code> slots anymore. If you absolutely do need to do things before normal initialization or finalization, you can define an <code>:around<\/code> method.<\/p>\n<p>One last widget-related definition form is <code>define-menu<\/code>, which is a very convenient way of specifying menus:<\/p>\n<pre>(define-menu (my-widget File)\n  (:item (\"Open...\" (ctrl o))\n    (open-file))\n  (:menu recent-files)\n  (:separator)\n  (:item (\"Save\" (ctrl s))\n    (save-file))\n  (:item (\"Save As...\" (ctrl alt s))\n    (save-file NIL))\n  (:menu \"Export\"\n    (:item \"PNG\" (save-file NIL \"png\")))\n  (:separator)\n  (:item (\"Quit\" (ctrl q))\n    (#_close widget)))\n<\/pre>\n<p>Out of the box, it supports <code>:item<\/code>, <code>:menu<\/code>, and <code>:separator<\/code> components. The item takes a name, which can be either a string for a label, a list of string and keyboard mnemonic, or a symbol indicating the class slot that contains the item widget, and it takes a body of forms to execute if the item is triggered. Menus take a name as a string and a body of components to contain, or a symbol indicating the slot that contains the menu widget. New components can be added with <code>define-menu-content-type<\/code>.<\/p>\n<p>If your widget has a Qt superclass to whose constructor you would like to pass certain arguments, you can do so by adding a method on <code>construct<\/code> that calls <code>qt:new<\/code> with the arguments you need. You can also use a shorthand to do the same by adding a <code>(:constructor ..)<\/code> option to your widget definition form. This expands to a <code>defmethod<\/code> form that binds all slots (<code>with-all-slots-bound<\/code>) and otherwise just contains a call to <code>qt:new<\/code> supplying the arguments of the option form verbatim. In effect this means that any unquoted symbol in the option denotes a slot reference. Note that the call to <code>construct<\/code> happens before the widget initializers are run. Thus you will not be able to use slot values that are set by subwidgets or initializers.<\/p>\n<p>As a final touch, Qtools offers macros for connecting slots and emitting signals. These translate to CommonQt's <code>emit-signal<\/code> and <code>connect<\/code> functions, and thus just offer a bit of syntactic sugar. They're called <code>signal!<\/code> and <code>connect!<\/code>. Their use is simple enough:<\/p>\n<pre> (connect! input (text-edited string) widget (text-changed string))\n (signal! input (text-edited string) \"Eyyyy\")\n<\/pre>\n<p>An experimental variant for the adventurous is <code>generic-signal<\/code>, which attempts to determine the argument types by their values at run-time. It does therefore not require specifying the type explicitly, but might instead screw up and choose a wrong type for the signal and thus fail to emit.<\/p>\n<h3><a id=\"user-content-q\" class=\"anchor\" aria-hidden=\"true\" href=\"#q\"><\/a>Q+<\/h3>\n<p>By default with CommonQt, calling Qt methods happens with the <code>#_<\/code> reader macro. This requires you to follow the proper case of the class and method names. Having this kind of mixture of conventions in the code is a bit jarring. While Qtools offers solutions to deal with the discrepancies of defining your own classes and widgets using the various <code>define-*<\/code> macros, Q+ fixes the method calling discrepancy. In order to use Q+ you have a choice of either using the <code>q+<\/code> macro, or using the <code>:qtools<\/code> read-table. Using the <code>q+<\/code> macro directly an example translates like this:<\/p>\n<pre>(let ((widget (#_new QWidget)))\n  (#_setWindowTitle widget \"Hello!\")\n  (#_show widget)\n  (#_exec *qapplication*))\n(let ((widget (q+ make-qwidget)))\n  (q+ set-window-title widget \"Hello!\")\n  (q+ show widget)\n  (q+ exec *qapplication*))\n<\/pre>\n<p>And the same using the readtable:<\/p>\n<pre>(let ((widget (q+:make-qwidget)))\n  (q+:set-window-title widget \"Hello!\")\n  (q+:show widget)\n  (q+:exec *qapplication*))\n<\/pre>\n<p>The difference is minimal in the typed code. However, the second approach will give you the convenience of letting the editor display the possible arguments and a docstring linking to the Qt methods. The second example is read by the Common Lisp reader to the first example. There is therefore no code difference in how the two work. If you use the <code>cl+qt<\/code> package, you can also take advantage of an extended <code>setf<\/code> macro. Using it, the second line would look like so:<\/p>\n<pre>(setf (q+ window-title widget) \"Hello!\")\n<\/pre>\n<p>Some of the setter functions require multiple values to be set at once. The updated <code>setf<\/code> can also deal with that:<\/p>\n<pre>(setf (q+ window painter) (values 0 0 100 100))\n<\/pre>\n<p>The <code>setf<\/code> has extra support for <code>q+<\/code>, but is otherwise identical to <code>cl:setf<\/code> and actually expands to that for all other places.<\/p>\n<p>In order to access enum values, you simply use the class name followed by a dot and the enum name. Constructors are the class name prefixed with \"make\". Static functions are the class name, a dash, and the method name in the standard translation scheme.<\/p>\n<pre>(q+:qt.blue)\n(q+:make-qpushbutton \"Foo!\")\n(q+:qmessagebox-information parent \"!\" \"hello!\")\n<\/pre>\n<p>For the specific arguments, names, and everything else, refer to the <a href=\"http:\/\/qt-project.org\/doc\/qt-4.8\/\" rel=\"nofollow\">Qt4.8<\/a> documentation. It's very good, trust me. The only thing you need to be aware of is the name conversion rules that Q+ uses to determine the proper Lisp symbol to use:<\/p>\n<ul><li>Method <code>QImage::constScanLine<\/code> =&gt; <code>q+:const-scan-line<\/code><\/li>\n<li>Method <code>Foo::set_widget<\/code> =&gt; <code>q+:set_widget<\/code><\/li>\n<li>Constant <code>QImage::InvertRgb<\/code> =&gt; <code>q+:qimage.invert-rgb<\/code><\/li>\n<li>Constant <code>QImage::Format_ARGB32_Premultiplied<\/code> =&gt; <code>q+:qimage.format_argb32_premultiplied<\/code><\/li>\n<li>Static Method <code>QFileDialog::getExistingDirectory<\/code> =&gt; <code>q+:qfiledialog-get-existing-directory<\/code><\/li>\n<li>Constructor <code>QImage::QImage<\/code> =&gt; <code>q+:make-qimage<\/code><\/li>\n<\/ul><p>For Q+ to work seamlessly in conjunction with ASDF systems and compiling\/loading code, you have to make sure that the smoke modules are set up correctly.<\/p>\n<p><code>q+<\/code> and the reader extension dynamically compile wrapper functions for the Qt methods you access. You can, however, also precompile all possible methods for the currently active set of smoke modules. To do this, you can either <code>:depends-on (:q+)<\/code> or compile a source file using <code>write-everything-to-file<\/code> and include it in your ASDF system. If you choose this approach, you will not need to switch the readtable or use the <code>q+<\/code> macro, as the package will be available fully populated.<\/p>\n<h3><a id=\"user-content-smoke-modules\" class=\"anchor\" aria-hidden=\"true\" href=\"#smoke-modules\"><\/a>Smoke Modules<\/h3>\n<p>In order to be able to use the various parts of Qt, the corresponding smoke modules need to be loaded. By default CommonQt loads <code>qtcore<\/code> and <code>qtgui<\/code> when <code>make-qapplication<\/code> is called. However, if you want to use, say, the OpenGL parts you'll also need <code>qtopengl<\/code>.<\/p>\n<p>Qtools provides ASDF systems for all the different smoke modules. That way, you can simply push the modules you want into your project's ASDF system dependencies and it'll ensure that the modules are available at compile and load time. Having the modules loaded at both times is especially important for Q+ to work properly. An example system making use of this would look like<\/p>\n<pre>(asdf:defsystem foo\n  ...\n  :depends-on (:qtcore :qtgui))\n<\/pre>\n<p>For a list of available smoke modules, see <code>*smoke-modules*<\/code>.<\/p>\n<h3><a id=\"user-content-fast-calling\" class=\"anchor\" aria-hidden=\"true\" href=\"#fast-calling\"><\/a>Fast Calling<\/h3>\n<p>In some cases it is vital to minimise overhead to a method call. Since usually dispatch is dynamic at run-time and un\/marshalling of arguments and return values is involved, the standard way of calling methods might be too slow. In that case, you can use Qtools' fast-call mechanism to perform a method call as quickly as possible. At a cost, naturally.<\/p>\n<pre>(let ((size (q+:make-qsize 100 100)))\n  (fast-call (set-height QSize int NIL) size 200)\n  size)\n<\/pre>\n<p>Fast-calling is possible if you know the exact method signature to call and are willing to translate the arguments to their proper required CFFI types. In the case of passing QObjects, you must pass their pointer. You can get that pointer by <code>qt::qobject-pointer<\/code>. The last argument in the method-descriptor is the return value type, which you can always set to NIL if you don't care about the return value.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<p>A couple of example applications using Qtools can be found in the <a href=\"https:\/\/github.com\/Shinmera\/qtools\/tree\/master\/examples\/\">examples\/<\/a> folder: <code>qtools-evaluator<\/code>, <code>qtools-titter<\/code>, <code>qtools-melody<\/code>, and <code>qtools-opengl<\/code>. Each of them can be loaded by their name, and launched using the <code>main<\/code> function from their package.<\/p>\n<h2><a id=\"user-content-extending-qtools\" class=\"anchor\" aria-hidden=\"true\" href=\"#extending-qtools\"><\/a>Extending Qtools<\/h2>\n<h3><a id=\"user-content-copying-and-finalizing\" class=\"anchor\" aria-hidden=\"true\" href=\"#copying-and-finalizing\"><\/a>Copying and Finalizing<\/h3>\n<p>In order to account for your own objects and operations you can extend the <code>copy<\/code> and <code>finalize<\/code> functions by using <code>define-copy-method<\/code> and <code>define-finalize-method<\/code> or <code>defmethod<\/code> directly. The two define macros bring the convenience of automatically resolving to a Qt class (and thus using <code>copy\/finalize-using-class<\/code>) if possible, making it all look a bit cleaner.<\/p>\n<pre>(define-copy-method (instance QPixmap)\n  \"Creates a new QPixmap using QPixmap::copy (deep copy).\"\n  (#_copy instance (#_rect instance)))\n<\/pre>\n<p>Since copying and finalizing are operations associated with a certain amount of ambiguity, it is advisable to always write documentation strings for your <code>copy<\/code>\/<code>finalize<\/code> methods. That way users can get a better idea of what will happen by reading about it using <code>describe-copy-method<\/code> and <code>describe-finalize-method<\/code> respectively.<\/p>\n<h3><a id=\"user-content-adding-defmethod-declarations\" class=\"anchor\" aria-hidden=\"true\" href=\"#adding-defmethod-declarations\"><\/a>Adding <code>defmethod<\/code> declarations<\/h3>\n<p>Using <code>define-method-declaration<\/code> you can add your own processing to method declarations. Your function should extract the necessary information from its declaration arguments and the <code>*method*<\/code> form. Each method declaration processing function should return a single form (like a macro) to be put before the resulting <code>defmethod<\/code>. The existing declaration processors are really short:<\/p>\n<pre>(define-method-declaration override (&amp;optional name)\n  (let ((slot (qtools:to-method-name (or name (form-fiddle:lambda-name *method*)))))\n    (with-widget-class (widget-class)\n      `(set-widget-class-option ',widget-class :override '(,slot ,name)))))\n<\/pre>\n<p>In this example we're using <a href=\"http:\/\/github.com\/Shinmera\/form-fiddle\">form-fiddle<\/a> to parse the method form. Using a library like that to ensure proper destructuring is important, as otherwise it's easy to accidentally butcher the method form, or get the wrong information.<\/p>\n<h3><a id=\"user-content-extending-the-menu-definition\" class=\"anchor\" aria-hidden=\"true\" href=\"#extending-the-menu-definition\"><\/a>Extending the menu definition<\/h3>\n<p>The menu definition form allows for arbitrary content types, so you may add new ones yourself by using <code>define-menu-content-type<\/code>. Each content type definition can return two values: an initform and a side-form. The initform will be put into the initialization function for the menu and thus evaluated when the widget is created. The side-form is put alongside the initializer definition and thus evaluated during compilation. If your menu type needs to modify the widget class in some way, that should be done through the side-forms. If it needs to connect signals, add items, or perform similar actions that involve Qt, that should go into the initform. You can call the expansion of other component types using <code>build-menu-content<\/code>. During the time your content-type function is run, <code>*widget*<\/code> is bound to the class-name of the widget and during initialization it is bound to the actual widget instance.<\/p>\n<h2><a id=\"user-content-qtools-concepts\" class=\"anchor\" aria-hidden=\"true\" href=\"#qtools-concepts\"><\/a>Qtools Concepts<\/h2>\n<p>Qtools has grown to be a large library with a lot of rather complicated concepts. I will try to describe them here, in order to retain some information in non-code form and make things clearer to the average user. It is not necessary to read and understand this section to use Qtools, but it may be useful to be aware of the underlying ideas and functionality that make Qtools work.<\/p>\n<h3><a id=\"user-content-finalizables\" class=\"anchor\" aria-hidden=\"true\" href=\"#finalizables\"><\/a>Finalizables<\/h3>\n<p>A finalizable is implemented using two classes, one serving as the metaclass and the other as a superclass. The metaclass is required in order to allow a custom slot type that supports the <code>:finalized<\/code> argument. The superclass is necessary in order to allow methods such as <code>finalize<\/code> to operate on instances of the finalizable classes. The handling of the finalized slots is done through a general method on <code>finalize<\/code> that scans through the slots of the class instance and then calls <code>finalize<\/code> on each slot for which the definition is set to be <code>:finalized<\/code>. This makes finalization of class slots automatic and convenient.<\/p>\n<p>Since finalizables don't add any metaclass properties, there is no need to manually calculate inheritance order. However, as with all custom slot definitions, the slot properties must be copied over from the direct-slot instance to the effective slot. In the case of <code>finalizable-class<\/code> this happens in <code>compute-effective-slot-definition<\/code>.<\/p>\n<p>In order to support finalizing of Qt class instances that don't have a CL class equivalent, the <code>finalize<\/code> method is extended wit ha <code>finalize-using-class<\/code> that is dispatched to using the Qt class and the instance of a <code>qobject<\/code> instance is passed to <code>finalize<\/code>.<\/p>\n<h3><a id=\"user-content-widgets-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#widgets-1\"><\/a>Widgets<\/h3>\n<p>As with finalizables, the widget is implemented using two classes, the <code>widget-class<\/code> metaclass and the <code>widget<\/code> superclass. These both inherit from the finalizable equivalents. The main crux of the <code>widget-class<\/code> lies in its <code>widget-class-direct-options<\/code> and <code>widget-class-extern-options<\/code>. The direct options are the options that are passed to a <code>re\/<\/code> <code>initialize-instance<\/code> (and thus also to <code>defclass<\/code>). They're caught in the appropriately specialised methods and then stored on the class. The effect of this is that we can fully recompute the class definition at any time, and potentially add or remove options without influencing the original <code>defclass<\/code> statement. This is where the extern-options come in.<\/p>\n<p>Using <code>set-widget-class-option<\/code> options can be added to the class definition dynamically at any point in the program. This function then adds the option to the class' extern-options and then calls <code>reinitialize-instance<\/code>, which in turn causes the class to get effectively redefined outside of its <code>defclass<\/code> form. This redefinition also allows us to change CommonQt class options. Using this we can create forms outside of the original <code>defclass<\/code> that act as if they were actually options in the <code>defclass<\/code> form.<\/p>\n<p>Qtools effectively only provides two forms that do this: <code>define-signal<\/code> and <code>defmethod<\/code>. The <code>define-signal<\/code> is relatively straightforward and simply expands to a class option set to add a new signal option. The <code>defmethod<\/code> is an extensible machine in itself.<\/p>\n<p>What's special about the <code>cl+qt:defmethod<\/code> is that it inspects the declaration forms in the method body. It then checks for each declaration form whether a handler function exist and if so, calls that function. Such a <code>method-declaration<\/code> function can then return forms to be put into the macroexpansion of the <code>cl+qt:defmethod<\/code>, before the resulting <code>cl:defmethod<\/code>. The processed declaration is then left out of the <code>cl:defmethod<\/code>form as it is assumed that it isn't a standard common lisp declaration. However, the declaration function also has the ability to change the contents of the <code>cl:defmethod<\/code> form itself, by manipulating <code>*method*<\/code>. This allows the declaration to output special handling for the method body, for example.<\/p>\n<p>This kind of extensible declaration mechanism is necessary both to allow further evolving of Qtools in the future as well as adaptation by users. It also offers a very \"native-like\" way of specifying external effects of a method. Qtools uses this construct then to allow definition of slots, overrides, finalizers, and initializers.<\/p>\n<p>In case the user doesn't appreciate the <code>defmethod<\/code> way, Qtools then provides <code>define-*<\/code> alternative functions that simply wrap over <code>defmethod<\/code>, establish some default bindings, and take care of naming and specialising the method.<\/p>\n<p>The <code>define-subwidget<\/code> deserves special attention here, as it does more work than the rest. A large part of defining widgets is adding sub-components to it, which is a task that usually involves a lot of repetition or awkward function sharing: Setting up a slot to hold the instance, defining or using an initializer to set it up. <code>define-subwidget<\/code> makes this both distributed and simple by both taking care of setting up an appropriate initializer function, and automatically adding the slot to the class using, again, <code>set-widget-class-option<\/code>. This slot is also always automatically set to be <code>finalized<\/code> in order to ensure that all widgets are properly cleaned up when the GUI is no longer needed.<\/p>\n<p>A minor problem regarding this approach is the same problem that appears with all of CL's definition forms. While developing incrementally, merely removing the definition form from the source file, will not actually remove it from the image. This can trip developers up, as definitions will still be active later. In this case it means having widgets still sticking around, or initializers running, etc. For this purpose there are corresponding <code>remove-*<\/code> functions to all the <code>define-*<\/code> functions to allow easy removal. This part cannot possibly be automated due to the nature of Common Lisp, but it is at least simple to correct should the need arise.<\/p>\n<h3><a id=\"user-content-q-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#q-1\"><\/a>Q+<\/h3>\n<p>CommonQt's way of dealing with method calls is the simple and most direct way of doing it. The first possible alternative to remove the need for typing method names in their corresponding case would be to simply introduce a different reader macro that automatically translates a <code>example-function<\/code>-like name into <code>exampleFunction<\/code> as is done already in other parts of Qtools. However, this has two downsides, the first being that there does not exist a 1:1 mapping of methods anywhere. The dynamic computation of the function name means that there isn't a full correspondence table anywhere. The second downside is that methods are still not passable as first-class objects.<\/p>\n<p>The way to solve this is to generate actual CL wrapper functions to the method calls. This allows us to use them as first-class objects, have at least compile-time argument number checking, and have a linkage of wrapper name to Qt method by listing it in the docstring.<\/p>\n<p>In order to achieve this, there are two possible choices. First, the entirety of all possible wrappers can be computed once, and then subsequently loaded into the image and used directly. However, this creates two new problems. Computing all wrappers, compiling them, and loading them, takes a long time and subsequently litters the image with thousands of functions and symbols that won't ever be used in the program. Then, we have a problem with the smoke modules, as we need to know which modules will be used in a potential application ahead of time, load them all, and then generate the wrappers. We cannot generate wrappers for each module separately, as the methods from different classes share the same wrapper functions. This means that whenever a different set of modules is needed, the wrappers need to be regenerated, recompiled, and reloaded. A lot of time and space goes to waste with this. However, this approach also has an advantage: As all functions are always available, it is easy to develop with. Arguments and docstrings will be readily available through the editor. Qtools offers this approach through <code>q+-compile-and-load<\/code>.<\/p>\n<p>The second approach is to dynamically only compile what is needed. That way, the image only ever contains wrappers for function that are actually called (at some point). However, this complicates things a lot. When a function is compiled that calls such a function, it doesn't exist yet. Even worse, when the form is read, the symbol for the function does not exist yet and isn't external! In order to catch this problem, a modified reader macro is necessary. This reader macro will detect when a call to a wrapper function is made, and instead transform it into a macro call that then sees to it that the wrapper will be created. Modifying the reader in such a way is a heavy change, and should only be used sparingly, however there is no alternative here. Qtools does not force you to use this reader extension, you can always just use the macro directly.<\/p>\n<p>However, dynamic compilation complications don't end there. Since we never dump the function to a file, it only ever exists in the environment it was compiled in. That means, if you compile a function that then dynamically generates the wrapper function, the wrapper won't be available anymore at load time. Qtools solves this issue with a trick. The <code>q+<\/code> macro expands to a <code>load-time-value<\/code> form that then generates the wrapper function. That way, the wrapper will always be available at load- and execution time, while posing no overhead to the execution time, as it will return a value to that won't impact anything.<\/p>\n<p>Function referencing gets the same problems as function calling, so the Qtools readtable also contains an overridden <code>#'<\/code> reader macro to handle that. In the case of a wrapper call, it expands to <code>Q+FUN<\/code> which in turn expands to a <code>load-time-value<\/code> form that generates and returns the function object.<\/p>\n<p>As is probably obvious by now, Qtools also implements the second approach. Therefore, the choice as a user is yours: You can statically precompile everything and use it directly, or you an use the dynamic on the fly compilation using either a readtable expansion, or a simple macro. The concepts to make this all possible are rather complex, but the actual function wrapper compilations are quite straight-forward. The system is currently not suited for extension, but I see no need to allow that as the kind of Qt methods that can exist are fixed and Q+ should handle all that are relevant.<\/p>\n<p>Q+ does one last thing to fix the \"issue\" of having setters instead of being able to use <code>setf<\/code>. For this purpose it has an extended <code>setf<\/code> macro that checks if a place is a function in the Q+ package. if so, this place\/value pair is instead expanded to a call to the appropriate wrapper function with the function name transformed. However, that's not the only reason to do this. The second is that some setters require multiple arguments to be set at once. Usually, <code>cl:setf<\/code> allows for cases like these by permitting setf expanders to accept multiple values. However, the number of values is fixed, and there's no way to dynamically know how many values where passed. Since Q+ needs to dispatch based on the number of values, this is not a viable approach. Therefore, with <code>cl+qt:setf<\/code> a <code>(values ..)<\/code> value form is specially treated and its arguments are inlined into the wrapper call. This also means that it isn't possible to use multiple values of a returning function as the values to a setter call, however while that is an inconsistency, I don't think it will be a big issue. If it turns out to be problematic in the later run, this will have to be changed to a dynamic analysis at run-time, which is an overhead I wanted to avoid.<\/p>\n<h2><a id=\"user-content-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#support\"><\/a>Support<\/h2>\n<p>Currently the following implementations are tested and supported by Qtools:<\/p>\n<ul><li><a href=\"http:\/\/www.sbcl.org\/\" rel=\"nofollow\">SBCL<\/a><\/li>\n<li><a href=\"http:\/\/ccl.clozure.com\/\" rel=\"nofollow\">CCL<\/a><\/li>\n<li><a href=\"https:\/\/common-lisp.net\/project\/ecl\/\" rel=\"nofollow\">ECL<\/a><\/li>\n<\/ul><p>You can run it on x86_64 Linux, OS X, and Windows, where SBCL on Linux is the recommended platform.<\/p>\n<p>It may or may not work more or less smoothly on other implementations and platforms depending on MOP and CommonQt support and general implementation quirks.<\/p>","created_at":"2018-11-07T16:22:23+0000","updated_at":"2018-11-07T16:22:23+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":36,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/4833204?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":68,"title":"Medieval Fantasy City Generator | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=17925666","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DanAndersen\" class=\"hnuser\">DanAndersen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934226\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=arethuza\" class=\"hnuser\">arethuza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934980\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Maultasche\" class=\"hnuser\">Maultasche<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938094\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=desertrider12\" class=\"hnuser\">desertrider12<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935721\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smogcutter\" class=\"hnuser\">smogcutter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936371\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=iagooar\" class=\"hnuser\">iagooar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936313\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kiliancs\" class=\"hnuser\">kiliancs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937415\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=franciscop\" class=\"hnuser\">franciscop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936168\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DonHopkins\" class=\"hnuser\">DonHopkins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936519\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thom\" class=\"hnuser\">thom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933456\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianbicking\" class=\"hnuser\">ianbicking<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934824\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=siavosh\" class=\"hnuser\">siavosh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936352\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933995\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=setr\" class=\"hnuser\">setr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937567\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938272\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=setr\" class=\"hnuser\">setr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938469\">60 days ago<\/a><\/p><br \/><div class=\"comment\">&gt;I\u2019ve also wondered about reduced scale in open world games<p>This used to be a popular strategy, my favorite being lego island 2. You don\u2019t need anything as grand an excuse as a neet pc or detective to allow for it: just take away the fucking car.<\/p><p>&gt;It should allow the game designer to treat a procedural world like he were a location scout. They can then build sets within that world where story elements can take place, locations modified as required and the remaining 99% left as a living world that swallows up the player<\/p><p>This is how I believe open world games are usually done now, if not actually generated, then done mentally (imagine a town by a river, add a simple crisis and let things play themselves out. Not necessarily likely events, and certainly not realistic, but sufficient).<\/p><p>Most open world games do have a decent number of setpieces to find. The problem, I think, is thats <i>all they are<\/i>. Oblivion, skyrim, witcher, etc all make the same mistake: \u201cyou can climb any mountain\u201d <i>and thats where it stops<\/i>. They are open <i>only<\/i> in the sense that you can travel in a non-linear path, in an non-linear order of the story. But they forget that the main interesting facet of games as a medium is <i>interaction<\/i>. That the player has <i>autonomy<\/i>.<\/p><p>They\u2019re merely open in story, but mechanically they\u2019re extremely linear. They don\u2019t provide <i>open interactions<\/i>. It doesn\u2019t matter what backdrop you give it, or how well designed and populated things are, if <i>I can\u2019t interact with them<\/i>.<\/p><p>Roguelikes (note the <i>k<\/i>), simulation-games and dwarf fortress (and from tabletops: adnd and its kin) are my favorite in this regard: you can interact with everything, in a somewhat limited and controlled fashion, with the potential for interesting, <i>unexpected<\/i> results. They essentially forgo any real plot or story in favor of letting player autonomy act as its own source of interest. The games and its internal interactions are mechanically complex, while the world around it is significantly less so.<\/p><p>But thus, my opinion is the opposite of yours: the advancements in technology surrounding games, outside of perhaps AR\/VR, currently and in the near-future, offers <i>nothing new<\/i>. The same style of games have already existed, and will continue on independent of these advancements. The only difference will be their graphical complexity, and more likely than not, each advance in graphics will come at the cost of the game\u2019s design, in general. We\u2019re in for nothing new.\n              <\/p><\/div><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17939618\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=staticfloat\" class=\"hnuser\">staticfloat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17942278\">59 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smogcutter\" class=\"hnuser\">smogcutter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936427\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936488\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AnIdiotOnTheNet\" class=\"hnuser\">AnIdiotOnTheNet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933507\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=indigochill\" class=\"hnuser\">indigochill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933648\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Retric\" class=\"hnuser\">Retric<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933794\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Ntrails\" class=\"hnuser\">Ntrails<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933924\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AnIdiotOnTheNet\" class=\"hnuser\">AnIdiotOnTheNet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937461\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dstick\" class=\"hnuser\">dstick<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933582\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swift532\" class=\"hnuser\">swift532<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933731\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thedirt0115\" class=\"hnuser\">thedirt0115<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936332\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sandworm101\" class=\"hnuser\">sandworm101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935733\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mywittyname\" class=\"hnuser\">mywittyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936621\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Merg\" class=\"hnuser\">Merg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935129\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moomin\" class=\"hnuser\">moomin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933869\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=doytch\" class=\"hnuser\">doytch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934316\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=whamlastxmas\" class=\"hnuser\">whamlastxmas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935617\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thom\" class=\"hnuser\">thom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938423\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gameswithgo\" class=\"hnuser\">gameswithgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933910\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moomin\" class=\"hnuser\">moomin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936756\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LargeWu\" class=\"hnuser\">LargeWu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934448\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fomite\" class=\"hnuser\">Fomite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936368\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=virtualritz\" class=\"hnuser\">virtualritz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934042\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jpmoral\" class=\"hnuser\">jpmoral<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933899\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gameswithgo\" class=\"hnuser\">gameswithgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933907\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jpmoral\" class=\"hnuser\">jpmoral<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934118\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mikepurvis\" class=\"hnuser\">mikepurvis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934577\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wlesieutre\" class=\"hnuser\">wlesieutre<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934775\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cissou\" class=\"hnuser\">cissou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17931922\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LoSboccacc\" class=\"hnuser\">LoSboccacc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932202\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933354\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_Microft\" class=\"hnuser\">_Microft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933006\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=srt19170\" class=\"hnuser\">srt19170<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17951923\">58 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_Microft\" class=\"hnuser\">_Microft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17958513\">57 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Animats\" class=\"hnuser\">Animats<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932455\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=presscast\" class=\"hnuser\">presscast<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932478\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=duiker101\" class=\"hnuser\">duiker101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932506\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blackbrokkoli\" class=\"hnuser\">blackbrokkoli<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933953\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jsilence\" class=\"hnuser\">jsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936809\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lbriner\" class=\"hnuser\">lbriner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932806\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gonzo41\" class=\"hnuser\">gonzo41<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932928\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MayeulC\" class=\"hnuser\">MayeulC<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933866\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=logfromblammo\" class=\"hnuser\">logfromblammo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935270\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=markatkinson\" class=\"hnuser\">markatkinson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933343\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ivanhoe\" class=\"hnuser\">ivanhoe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933033\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=131012\" class=\"hnuser\">131012<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933272\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=baud147258\" class=\"hnuser\">baud147258<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933786\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sonnyblarney\" class=\"hnuser\">sonnyblarney<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933477\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933811\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rileyphone\" class=\"hnuser\">rileyphone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934433\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gnode\" class=\"hnuser\">gnode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934414\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anonytrary\" class=\"hnuser\">anonytrary<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932294\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gnode\" class=\"hnuser\">gnode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934508\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Svoka\" class=\"hnuser\">Svoka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933847\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=samstave\" class=\"hnuser\">samstave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937416\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=uberman\" class=\"hnuser\">uberman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933803\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GistNoesis\" class=\"hnuser\">GistNoesis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933080\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=toomanybeersies\" class=\"hnuser\">toomanybeersies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17931944\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dang\" class=\"hnuser\">dang<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932159\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=n4r9\" class=\"hnuser\">n4r9<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933055\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ez15972\" class=\"hnuser\">ez15972<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932255\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloydde\" class=\"hnuser\">lloydde<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932196\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jschuur\" class=\"hnuser\">jschuur<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932625\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrec\" class=\"hnuser\">mrec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933944\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932607\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shusson\" class=\"hnuser\">shusson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932612\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933805\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mywittyname\" class=\"hnuser\">mywittyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936764\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937572\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhaps0dy\" class=\"hnuser\">rhaps0dy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933206\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kagerjay\" class=\"hnuser\">Kagerjay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934477\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikb\" class=\"hnuser\">erikb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933650\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=leanthonyrn\" class=\"hnuser\">leanthonyrn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934056\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikb\" class=\"hnuser\">erikb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936372\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fomite\" class=\"hnuser\">Fomite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936400\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikb\" class=\"hnuser\">erikb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17944669\">59 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevoski\" class=\"hnuser\">stevoski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932950\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fapjacks\" class=\"hnuser\">fapjacks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17939312\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brandonjm\" class=\"hnuser\">brandonjm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932029\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DonHopkins\" class=\"hnuser\">DonHopkins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936294\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crashdown\" class=\"hnuser\">crashdown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933025\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shezi\" class=\"hnuser\">shezi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932439\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ahje\" class=\"hnuser\">ahje<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932501\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KirinDave\" class=\"hnuser\">KirinDave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936270\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skvj\" class=\"hnuser\">skvj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936748\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=midasz\" class=\"hnuser\">midasz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932448\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bencollier49\" class=\"hnuser\">bencollier49<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932674\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932868\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stankot\" class=\"hnuser\">stankot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932764\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wackro\" class=\"hnuser\">wackro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932930\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stankot\" class=\"hnuser\">stankot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934044\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=geokon\" class=\"hnuser\">geokon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933021\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SilverSlash\" class=\"hnuser\">SilverSlash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17945244\">59 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kagerjay\" class=\"hnuser\">Kagerjay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934465\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ralphael\" class=\"hnuser\">ralphael<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932632\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=selimthegrim\" class=\"hnuser\">selimthegrim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935650\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w0ns88\" class=\"hnuser\">w0ns88<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932707\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=suttan\" class=\"hnuser\">suttan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932795\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jgh\" class=\"hnuser\">jgh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932919\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:22:22+0000","updated_at":"2018-11-07T16:22:22+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":67,"title":"Medieval Fantasy City Generator | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=17925666","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DanAndersen\" class=\"hnuser\">DanAndersen<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934226\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=arethuza\" class=\"hnuser\">arethuza<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934980\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Maultasche\" class=\"hnuser\">Maultasche<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938094\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=desertrider12\" class=\"hnuser\">desertrider12<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935721\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smogcutter\" class=\"hnuser\">smogcutter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936371\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=iagooar\" class=\"hnuser\">iagooar<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936313\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kiliancs\" class=\"hnuser\">kiliancs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937415\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=franciscop\" class=\"hnuser\">franciscop<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936168\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DonHopkins\" class=\"hnuser\">DonHopkins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936519\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thom\" class=\"hnuser\">thom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933456\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ianbicking\" class=\"hnuser\">ianbicking<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934824\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=siavosh\" class=\"hnuser\">siavosh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936352\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933995\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=setr\" class=\"hnuser\">setr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937567\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938272\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=setr\" class=\"hnuser\">setr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938469\">60 days ago<\/a><\/p><br \/><div class=\"comment\">&gt;I\u2019ve also wondered about reduced scale in open world games<p>This used to be a popular strategy, my favorite being lego island 2. You don\u2019t need anything as grand an excuse as a neet pc or detective to allow for it: just take away the fucking car.<\/p><p>&gt;It should allow the game designer to treat a procedural world like he were a location scout. They can then build sets within that world where story elements can take place, locations modified as required and the remaining 99% left as a living world that swallows up the player<\/p><p>This is how I believe open world games are usually done now, if not actually generated, then done mentally (imagine a town by a river, add a simple crisis and let things play themselves out. Not necessarily likely events, and certainly not realistic, but sufficient).<\/p><p>Most open world games do have a decent number of setpieces to find. The problem, I think, is thats <i>all they are<\/i>. Oblivion, skyrim, witcher, etc all make the same mistake: \u201cyou can climb any mountain\u201d <i>and thats where it stops<\/i>. They are open <i>only<\/i> in the sense that you can travel in a non-linear path, in an non-linear order of the story. But they forget that the main interesting facet of games as a medium is <i>interaction<\/i>. That the player has <i>autonomy<\/i>.<\/p><p>They\u2019re merely open in story, but mechanically they\u2019re extremely linear. They don\u2019t provide <i>open interactions<\/i>. It doesn\u2019t matter what backdrop you give it, or how well designed and populated things are, if <i>I can\u2019t interact with them<\/i>.<\/p><p>Roguelikes (note the <i>k<\/i>), simulation-games and dwarf fortress (and from tabletops: adnd and its kin) are my favorite in this regard: you can interact with everything, in a somewhat limited and controlled fashion, with the potential for interesting, <i>unexpected<\/i> results. They essentially forgo any real plot or story in favor of letting player autonomy act as its own source of interest. The games and its internal interactions are mechanically complex, while the world around it is significantly less so.<\/p><p>But thus, my opinion is the opposite of yours: the advancements in technology surrounding games, outside of perhaps AR\/VR, currently and in the near-future, offers <i>nothing new<\/i>. The same style of games have already existed, and will continue on independent of these advancements. The only difference will be their graphical complexity, and more likely than not, each advance in graphics will come at the cost of the game\u2019s design, in general. We\u2019re in for nothing new.\n              <\/p><\/div><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17939618\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=staticfloat\" class=\"hnuser\">staticfloat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17942278\">59 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smogcutter\" class=\"hnuser\">smogcutter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936427\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=knolan\" class=\"hnuser\">knolan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936488\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AnIdiotOnTheNet\" class=\"hnuser\">AnIdiotOnTheNet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933507\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=indigochill\" class=\"hnuser\">indigochill<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933648\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Retric\" class=\"hnuser\">Retric<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933794\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Ntrails\" class=\"hnuser\">Ntrails<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933924\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AnIdiotOnTheNet\" class=\"hnuser\">AnIdiotOnTheNet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937461\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dstick\" class=\"hnuser\">dstick<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933582\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=swift532\" class=\"hnuser\">swift532<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933731\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thedirt0115\" class=\"hnuser\">thedirt0115<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936332\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sandworm101\" class=\"hnuser\">sandworm101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935733\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mywittyname\" class=\"hnuser\">mywittyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936621\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Merg\" class=\"hnuser\">Merg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935129\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moomin\" class=\"hnuser\">moomin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933869\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=doytch\" class=\"hnuser\">doytch<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934316\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=whamlastxmas\" class=\"hnuser\">whamlastxmas<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935617\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thom\" class=\"hnuser\">thom<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17938423\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gameswithgo\" class=\"hnuser\">gameswithgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933910\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=moomin\" class=\"hnuser\">moomin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936756\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LargeWu\" class=\"hnuser\">LargeWu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934448\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fomite\" class=\"hnuser\">Fomite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936368\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=virtualritz\" class=\"hnuser\">virtualritz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934042\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jpmoral\" class=\"hnuser\">jpmoral<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933899\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gameswithgo\" class=\"hnuser\">gameswithgo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933907\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jpmoral\" class=\"hnuser\">jpmoral<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934118\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mikepurvis\" class=\"hnuser\">mikepurvis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934577\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wlesieutre\" class=\"hnuser\">wlesieutre<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934775\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cissou\" class=\"hnuser\">cissou<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17931922\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=LoSboccacc\" class=\"hnuser\">LoSboccacc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932202\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=codetrotter\" class=\"hnuser\">codetrotter<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933354\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_Microft\" class=\"hnuser\">_Microft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933006\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=srt19170\" class=\"hnuser\">srt19170<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17951923\">58 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=_Microft\" class=\"hnuser\">_Microft<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17958513\">57 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Animats\" class=\"hnuser\">Animats<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932455\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=presscast\" class=\"hnuser\">presscast<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932478\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=duiker101\" class=\"hnuser\">duiker101<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932506\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blackbrokkoli\" class=\"hnuser\">blackbrokkoli<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933953\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jsilence\" class=\"hnuser\">jsilence<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936809\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lbriner\" class=\"hnuser\">lbriner<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932806\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gonzo41\" class=\"hnuser\">gonzo41<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932928\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=MayeulC\" class=\"hnuser\">MayeulC<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933866\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=logfromblammo\" class=\"hnuser\">logfromblammo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935270\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=markatkinson\" class=\"hnuser\">markatkinson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933343\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ivanhoe\" class=\"hnuser\">ivanhoe<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933033\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=131012\" class=\"hnuser\">131012<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933272\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=baud147258\" class=\"hnuser\">baud147258<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933786\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sonnyblarney\" class=\"hnuser\">sonnyblarney<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933477\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933811\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rileyphone\" class=\"hnuser\">rileyphone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934433\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gnode\" class=\"hnuser\">gnode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934414\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=anonytrary\" class=\"hnuser\">anonytrary<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932294\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gnode\" class=\"hnuser\">gnode<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934508\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Svoka\" class=\"hnuser\">Svoka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933847\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=samstave\" class=\"hnuser\">samstave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937416\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=uberman\" class=\"hnuser\">uberman<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933803\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=GistNoesis\" class=\"hnuser\">GistNoesis<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933080\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=toomanybeersies\" class=\"hnuser\">toomanybeersies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17931944\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dang\" class=\"hnuser\">dang<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932159\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=n4r9\" class=\"hnuser\">n4r9<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933055\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ez15972\" class=\"hnuser\">ez15972<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932255\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=lloydde\" class=\"hnuser\">lloydde<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932196\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jschuur\" class=\"hnuser\">jschuur<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932625\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrec\" class=\"hnuser\">mrec<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933944\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932607\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shusson\" class=\"hnuser\">shusson<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932612\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933805\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mywittyname\" class=\"hnuser\">mywittyname<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936764\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=CalRobert\" class=\"hnuser\">CalRobert<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17937572\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rhaps0dy\" class=\"hnuser\">rhaps0dy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933206\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kagerjay\" class=\"hnuser\">Kagerjay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934477\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikb\" class=\"hnuser\">erikb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933650\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=leanthonyrn\" class=\"hnuser\">leanthonyrn<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934056\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikb\" class=\"hnuser\">erikb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936372\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Fomite\" class=\"hnuser\">Fomite<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936400\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=erikb\" class=\"hnuser\">erikb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17944669\">59 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stevoski\" class=\"hnuser\">stevoski<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932950\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fapjacks\" class=\"hnuser\">fapjacks<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17939312\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=brandonjm\" class=\"hnuser\">brandonjm<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932029\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=DonHopkins\" class=\"hnuser\">DonHopkins<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936294\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crashdown\" class=\"hnuser\">crashdown<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933025\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=shezi\" class=\"hnuser\">shezi<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932439\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ahje\" class=\"hnuser\">ahje<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932501\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=KirinDave\" class=\"hnuser\">KirinDave<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936270\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=skvj\" class=\"hnuser\">skvj<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17936748\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=midasz\" class=\"hnuser\">midasz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932448\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bencollier49\" class=\"hnuser\">bencollier49<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932674\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BerislavLopac\" class=\"hnuser\">BerislavLopac<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932868\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stankot\" class=\"hnuser\">stankot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932764\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wackro\" class=\"hnuser\">wackro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932930\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stankot\" class=\"hnuser\">stankot<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934044\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=geokon\" class=\"hnuser\">geokon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17933021\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SilverSlash\" class=\"hnuser\">SilverSlash<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17945244\">59 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Kagerjay\" class=\"hnuser\">Kagerjay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17934465\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ralphael\" class=\"hnuser\">ralphael<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932632\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=selimthegrim\" class=\"hnuser\">selimthegrim<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17935650\">60 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=w0ns88\" class=\"hnuser\">w0ns88<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932707\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=suttan\" class=\"hnuser\">suttan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932795\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jgh\" class=\"hnuser\">jgh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17932919\">61 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:22:21+0000","updated_at":"2018-11-07T16:22:21+0000","annotations":[],"mimetype":"text\/html","reading_time":4,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":66,"title":"cjbarber\/ToolsOfTheTrade","url":"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<blockquote>\n<p>Tools of The Trade, from <a href=\"https:\/\/news.ycombinator.com\" rel=\"nofollow\">Hacker News<\/a>.<\/p>\n<\/blockquote>\n<h5><a id=\"user-content-background\" class=\"anchor\" aria-hidden=\"true\" href=\"#background\"><\/a>Background<\/h5>\n<p>In 2010, Joshua Schachter, the founder of Delicious, <a href=\"https:\/\/news.ycombinator.com\/item?id=1769910\" rel=\"nofollow\">posted the following<\/a> on Hacker News:<\/p>\n<blockquote>\n<p>When I first started delicious, we had to host most of the services ourselves. CVS, mail, mailing lists, etc etc etc.<\/p>\n<p>These days, lots of that stuff is available as SaaS. What are the tools and services people use instead of hosting their own?<\/p>\n<p>(I'm not talking about actual production services like EC2 and Heroku and whatnot. We can go over this in another thread.)<\/p>\n<\/blockquote>\n<p>In 2013, Sharjeel Qureshi, <a href=\"https:\/\/news.ycombinator.com\/item?id=5235137\" rel=\"nofollow\">posted the following<\/a>:<\/p>\n<blockquote>\n<p>Few years ago, Joshua Schachter started this thread on HN for discussing hosted useful services: <a href=\"https:\/\/news.ycombinator.com\/item?id=1769910\" rel=\"nofollow\">https:\/\/news.ycombinator.com\/item?id=1769910<\/a><\/p>\n<p>The contribution in thread introduced many interesting SaaS services which can immensely help in deploying services as well as development.<\/p>\n<p>It's been three years since then. What do we have today?<\/p>\n<\/blockquote>\n<p>Many thanks to the big contributors to the previous threads, including garrettdimon, espeed, netshade, and cmadan, and many more that I haven't named.<\/p>\n<h5><a id=\"user-content-now\" class=\"anchor\" aria-hidden=\"true\" href=\"#now\"><\/a>Now<\/h5>\n<p>I've collected more data from Hacker News, AngelList and Quora, to make the 2015 and beyond version. This list also includes self-hosted as well as hosted services.<\/p>\n<p>It's on GitHub for a reason! Please submit pull requests.<\/p>\n<h2><a id=\"user-content-supported-with-\ufe0f-by-instabug\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-with-\ufe0f-by-instabug\"><\/a>Supported with \u2764\ufe0f by <a href=\"https:\/\/instabug.com\/?utm_source=toolsofthetrade&amp;utm_medium=spon&amp;utm_content=header\" rel=\"nofollow\">Instabug<\/a><\/h2>\n<p><a href=\"https:\/\/instabug.com\/?utm_source=toolsofthetrade&amp;utm_medium=spon&amp;utm_content=banner\" rel=\"nofollow\"><img src=\"https:\/\/user-images.githubusercontent.com\/9888943\/46256666-08746700-c4ae-11e8-8c64-406bbbfe2e0f.png\" alt=\"instabug\" \/><\/a><\/p>\n<blockquote>\n<p>Instabug will enable you to easily collect feedback from your beta testers and users. The SDK will automatically attach all the details you would need to debug and fix bugs and crashes faster. Join the top apps in the world that rely on Instabug to enhance their app quality and iterate faster. <a href=\"https:\/\/instabug.com\/?utm_source=toolsofthetrade&amp;utm_medium=spon&amp;utm_content=learnmore\" rel=\"nofollow\">learn more.<\/a>\ud83d\ude80<a href=\"https:\/\/instabug.com\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/cb3c1dfbdecd9b7411d57861eba9286374ce6638\/68747470733a2f2f696e7374616275672d67612e61707073706f742e636f6d2f55412d34313938323038382d362f6769746875622f546f6f6c734f6654686554726164653f706978656c\" alt=\"\" data-canonical-src=\"https:\/\/instabug-ga.appspot.com\/UA-41982088-6\/github\/ToolsOfTheTrade?pixel\" \/><\/a><\/p>\n<\/blockquote>\n<h2><a id=\"user-content-contents\" class=\"anchor\" aria-hidden=\"true\" href=\"#contents\"><\/a>Contents<\/h2>\n<ul><li><a href=\"#accountinginvoicing\">Accounting\/Invoicing<\/a><\/li>\n<li><a href=\"#api-builder\">API Builder<\/a><\/li>\n<li><a href=\"#app-developer-tools\">App Developer Tools<\/a><\/li>\n<li><a href=\"#application-distribution\">Application Distribution<\/a><\/li>\n<li><a href=\"#application-performance\">Application Performance<\/a><\/li>\n<li><a href=\"#aws-tools\">AWS Tools<\/a><\/li>\n<li><a href=\"#backend-aas\">Backend-aaS<\/a><\/li>\n<li><a href=\"#banking\">Banking<\/a><\/li>\n<li><a href=\"#billing--payment-processing\">Billing &amp; Payment Processing<\/a><\/li>\n<li><a href=\"#browseremail-testing\">Browser\/Email Testing<\/a><\/li>\n<li><a href=\"#bugissue-tracking\">Bug\/Issue Tracking<\/a><\/li>\n<li><a href=\"#business--traffic-analytics\">Business &amp; Traffic Analytics<\/a><\/li>\n<li><a href=\"#business-cards-and-print-material\">Business Cards and Print Material<\/a><\/li>\n<li><a href=\"#community-tools\">Community Tools<\/a><\/li>\n<li><a href=\"#content-creationinfographics\">Content Creation\/Infographics<\/a><\/li>\n<li><a href=\"#continuous-integrationcode-quality\">Continuous Integration\/Code Quality<\/a><\/li>\n<li><a href=\"#conversion-optimization--ab-testing\">Conversion Optimization &amp; A\/B Testing<\/a><\/li>\n<li><a href=\"#crmsales-tools\">CRM\/Sales Tools<\/a><\/li>\n<li><a href=\"#customer-feedback\">Customer Feedback<\/a><\/li>\n<li><a href=\"#customer-supporthelp-desks\">Customer Support\/Help Desks<\/a><\/li>\n<li><a href=\"#dashboards\">Dashboards<\/a><\/li>\n<li><a href=\"#data\">Data<\/a><\/li>\n<li><a href=\"#database-aas\">Database-aaS<\/a><\/li>\n<li><a href=\"#database\">Database<\/a><\/li>\n<li><a href=\"#deployment\">Deployment<\/a><\/li>\n<li><a href=\"#design-collaboration\">Design Collaboration<\/a><\/li>\n<li><a href=\"#designers\">Designers<\/a><\/li>\n<li><a href=\"#dns\">DNS<\/a><\/li>\n<li><a href=\"#documentation\">Documentation<\/a><\/li>\n<li><a href=\"#email-collectionlanding-page-apps\">Email Collection\/Landing Page Apps<\/a><\/li>\n<li><a href=\"#email-marketing\">Email Marketing<\/a><\/li>\n<li><a href=\"#engineering-metrics\">Engineering Metrics<\/a><\/li>\n<li><a href=\"#errorexception-handling\">Error\/Exception Handling<\/a><\/li>\n<li><a href=\"#forms--surveys\">Forms \/ Surveys<\/a><\/li>\n<li><a href=\"#fundraising--investor-relations\">Fundraising \/ Investor Relations<\/a><\/li>\n<li><a href=\"#group-communicationchat-tools\">Group Communication\/Chat Tools<\/a><\/li>\n<li><a href=\"#heroku-tools\">Heroku Tools<\/a><\/li>\n<li><a href=\"#hr\">HR<\/a><\/li>\n<li><a href=\"#identity-verification\">Identity Verification<\/a><\/li>\n<li><a href=\"#income-analytics\">Income Analytics<\/a><\/li>\n<li><a href=\"#knowledge-trackingwiki\">Knowledge Tracking\/Wiki<\/a><\/li>\n<li><a href=\"#load-testing\">Load Testing<\/a><\/li>\n<li><a href=\"#localization--internationalization\">Localization &amp; Internationalization<\/a><\/li>\n<li><a href=\"#log-monitoring\">Log Monitoring<\/a><\/li>\n<li><a href=\"#naming\">Naming<\/a><\/li>\n<li><a href=\"#notes\">Notes<\/a><\/li>\n<li><a href=\"#offsite-backups\">Offsite Backups<\/a><\/li>\n<li><a href=\"#ops-alerts-and-scheduling\">Ops Alerts and Scheduling<\/a><\/li>\n<li><a href=\"#other-apis\">Other APIs<\/a><\/li>\n<li><a href=\"#paas\">PaaS<\/a><\/li>\n<li><a href=\"#password-management\">Password Management<\/a><\/li>\n<li><a href=\"#payments-billing--downloads\">Payments, Billing &amp; Downloads<\/a><\/li>\n<li><a href=\"#payroll\">Payroll<\/a><\/li>\n<li><a href=\"#personal-machine-backups\">Personal Machine Backups<\/a><\/li>\n<li><a href=\"#personal-productivity\">Personal Productivity<\/a><\/li>\n<li><a href=\"#phonepbxsms\">Phone\/PBX\/SMS<\/a><\/li>\n<li><a href=\"#planning--project-management\">Planning &amp; Project Management<\/a><\/li>\n<li><a href=\"#presentations--slides\">Presentations \/ Slides<\/a><\/li>\n<li><a href=\"#privacy-policy-terms--conditions-legal-documents\">Privacy Policy, Terms &amp; Conditions, Legal Documents<\/a><\/li>\n<li><a href=\"#prototypingmockups\">Prototyping\/Mockups<\/a><\/li>\n<li><a href=\"#remote-collaboration\">Remote Collaboration<\/a><\/li>\n<li><a href=\"#remote-workers\">Remote Workers<\/a><\/li>\n<li><a href=\"#search\">Search<\/a><\/li>\n<li><a href=\"#security\">Security<\/a><\/li>\n<li><a href=\"#seo-tools\">SEO Tools<\/a><\/li>\n<li><a href=\"#server-monitoring\">Server Monitoring<\/a><\/li>\n<li><a href=\"#shipping\">Shipping<\/a><\/li>\n<li><a href=\"#site-search\">Site Search<\/a><\/li>\n<li><a href=\"#social-media-marketing\">Social Media Marketing<\/a><\/li>\n<li><a href=\"#source-code-hosting\">Source Code Hosting<\/a><\/li>\n<li><a href=\"#sources-of-clicksad-platforms\">Sources of Clicks\/Ad Platforms<\/a><\/li>\n<li><a href=\"#space-rental\">Space Rental<\/a><\/li>\n<li><a href=\"#status-blogsuser-alerts\">Status Blogs\/User Alerts<\/a><\/li>\n<li><a href=\"#storage\">Storage<\/a><\/li>\n<li><a href=\"#system-monitoring\">System Monitoring<\/a><\/li>\n<li><a href=\"#task-scheduling\">Task Scheduling<\/a><\/li>\n<li><a href=\"#time-tracking\">Time Tracking<\/a><\/li>\n<li><a href=\"#transactional-email\">Transactional Email<\/a><\/li>\n<li><a href=\"#user-feedback\">User Feedback<\/a><\/li>\n<li><a href=\"#user-management\">User Management<\/a><\/li>\n<li><a href=\"#user-testing\">User Testing<\/a><\/li>\n<li><a href=\"#video-hosting\">Video Hosting<\/a><\/li>\n<li><a href=\"#vps\">VPS<\/a><\/li>\n<li><a href=\"#websockets-aas\">WebSockets-aaS<\/a><\/li>\n<\/ul><h3><a id=\"user-content-identity-verification\" class=\"anchor\" aria-hidden=\"true\" href=\"#identity-verification\"><\/a>Identity Verification<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/cognitohq.com\" rel=\"nofollow\">Cognito<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getcognito\" rel=\"nofollow\">@getcognito<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Frictionless, modern identity verification that starts with just a phone number.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/blockscore.com\" rel=\"nofollow\">BlockScore<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getblockscore\" rel=\"nofollow\">@getblockscore<\/a><\/td>\n<td align=\"left\">$1\/verification - $5\/verification<\/td>\n<td align=\"left\">Individual ID verifications and company verifications.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/onfido.com\" rel=\"nofollow\">Onfido<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Onfido\" rel=\"nofollow\">@Onfido<\/a><\/td>\n<td align=\"left\">$2\/verification<\/td>\n<td align=\"left\">Individual ID verifications, Document Checks and Facial recognition.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-browseremail-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#browseremail-testing\"><\/a>Browser\/Email Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.browserstack.com\" rel=\"nofollow\">BrowserStack<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/browserstack\" rel=\"nofollow\">@browserstack<\/a><\/td>\n<td align=\"left\">$35\/mo - $359\/mo<\/td>\n<td align=\"left\">Live, web-based browser testing.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/litmus.com\" rel=\"nofollow\">Litmus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/litmusapp\" rel=\"nofollow\">@litmusapp<\/a><\/td>\n<td align=\"left\">$79\/mo - $399\/mo<\/td>\n<td align=\"left\">Lets you preview your campaigns across 30+ real email clients and devices in minutes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/saucelabs.com\" rel=\"nofollow\">Sauce Labs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/saucelabs\" rel=\"nofollow\">@saucelabs<\/a><\/td>\n<td align=\"left\">$12\/mo - $149\/mo<\/td>\n<td align=\"left\">Run your web and mobile app tests across hundreds of real browsers and platforms instantly.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.emailonacid.com\" rel=\"nofollow\">EmailOnAcid<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/EmailonAcid\" rel=\"nofollow\">@emailonacid<\/a><\/td>\n<td align=\"left\">$45\/mo - $295\/mo<\/td>\n<td align=\"left\">Quickly ensure that your emails don't look like crap in any email client.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.rainforestqa.com\" rel=\"nofollow\">Rainforest QA<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/rainforestqa\" rel=\"nofollow\">@rainforestqa<\/a><\/td>\n<td align=\"left\">$500\/mo - $2000\/mo<\/td>\n<td align=\"left\">Insanely simple integration testing.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/debugmail.io\" rel=\"nofollow\">DebugMail<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Easy to use fake mail (SMTP) server for developers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mailosaur.com\" rel=\"nofollow\">Mailosaur<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mailosaur\" rel=\"nofollow\">@mailosaur<\/a><\/td>\n<td align=\"left\">$19\/mo - $199\/mo<\/td>\n<td align=\"left\">Email, SMS and Spam testing with virtual SMTP servers for business-grade usage.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mailtrap.io\" rel=\"nofollow\">Mailtrap<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Mailtrap\" rel=\"nofollow\">@Mailtrap<\/a><\/td>\n<td align=\"left\">Free - $24.99\/mo<\/td>\n<td align=\"left\">A fake SMTP server for development and staging environments along with a REST API for testing.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-bugissue-tracking\" class=\"anchor\" aria-hidden=\"true\" href=\"#bugissue-tracking\"><\/a>Bug\/Issue Tracking<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/bitbucket.org\" rel=\"nofollow\">BitBucket Issues<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bitbucket\" rel=\"nofollow\">@bitbucket<\/a><\/td>\n<td align=\"left\">$10\/mo - $200\/mo, Unlimited private code repositories<\/td>\n<td align=\"left\">Host, manage, and share Git and Mercurial repositories in the cloud. Free, unlimited private repositories for up to 5 developers give teams the flexibility to grow and code without restrictions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/bugherd.com\" rel=\"nofollow\">BugHerd<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bugherd\" rel=\"nofollow\">@bugherd<\/a><\/td>\n<td align=\"left\">$29\/mo - $180\/mo<\/td>\n<td align=\"left\">Turn client feedback into actionable tasks. BugHerd lets you quickly see, at a glance, how your project is going and what everyone is working on. The task board lets you keep team members in sync by assigning and scheduling tasks with a simple drag and drop.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/bugify.com\" rel=\"nofollow\">Bugify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bugify\" rel=\"nofollow\">@bugify<\/a><\/td>\n<td align=\"left\">$59.00<\/td>\n<td align=\"left\">Self hosted issue management system. One-time payment. Written in PHP.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\">GitHub Issues<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/GitHub\" rel=\"nofollow\">@GitHub<\/a><\/td>\n<td align=\"left\">$7\/mo - $50\/mo<\/td>\n<td align=\"left\">Build software better, together. GitHub is the largest code host on the planet with over 13.2 million repositories. Large or small, every repository comes with the same powerful tools. These tools are open to the community for public projects and secure for private projects.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/about.gitlab.com\" rel=\"nofollow\">GitLab Issues<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gitlabhq\" rel=\"nofollow\">@gitlabhq<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">GitLab is open source software to collaborate on code that is used by more than 100,000 organisations worldwide. Unlimited private repositories on GitLab.com or host your own instance. Enterprise Edition includes deep LDAP support.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/huboard.com\" rel=\"nofollow\">Huboard<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/huboard\" rel=\"nofollow\">@huboard<\/a><\/td>\n<td align=\"left\">$7\/mo - $24\/mo<\/td>\n<td align=\"left\">GitHub issues made awesome Instant project management for GitHub repositories<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/jira\" rel=\"nofollow\">JIRA<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/JIRA\" rel=\"nofollow\">@JIRA<\/a><\/td>\n<td align=\"left\">$10\/mo hosted - $10\/yr self-hosted<\/td>\n<td align=\"left\">JIRA is the tracker for teams planning and building great products. Thousands of teams choose JIRA to capture and organize issues, assign work, and follow team activity. At your desk or on the go with the new mobile interface, JIRA helps your team get the job done.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/lighthouseapp.com\" rel=\"nofollow\">Lighthouse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/lighthouseapp\" rel=\"nofollow\">@lighthouseapp<\/a><\/td>\n<td align=\"left\">$25\/mo - $100\/mo<\/td>\n<td align=\"left\">Whether you're a large company or a small bootstrapped team, Lighthouse is the perfect ticket tracking solution. Collaborate effortlessly on projects. Whether you\u2019re a team of 5 or studio of 50, Lighthouse will help you keep track of your project development with ease.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pinitto.me\" rel=\"nofollow\">Pinitto.me<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Post It Notes on a virtual corkboard (OSS)<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sifterapp.com\" rel=\"nofollow\">Sifter<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sifterapp\" rel=\"nofollow\">@sifterapp<\/a><\/td>\n<td align=\"left\">$29\/mo - $149\/mo<\/td>\n<td align=\"left\">Less configuring. More doing. We've put in the time researching bug tracking to help create the simplest possible workflow for you to get work done. From time-to-time we even blog about some of our ideas around this optimal bug tracking process\u2026<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/usersnap.com\" rel=\"nofollow\">Usersnap<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/usersnap\" rel=\"nofollow\">@usersnap<\/a><\/td>\n<td align=\"left\">$19\/mo-$99\/mo<\/td>\n<td align=\"left\">Usersnap is visual bug reporting for anyone working on web projects. Get visual feedback and precious browser information with every bug report to reproduce and fix them even faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/instabug.com\" rel=\"nofollow\">Instabug<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/instabug\" rel=\"nofollow\">@instabug<\/a><\/td>\n<td align=\"left\">$49\/mo-$349\/mo<\/td>\n<td align=\"left\">Instabug is comprehensive bug reporting and in-app feedback tool for mobile apps. Receive detailed feedback attached with every bug report sent by your testers or users to debug  and build better apps, faster.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-planning--project-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#planning--project-management\"><\/a>Planning &amp; Project Management<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.aha.io\" rel=\"nofollow\">Aha!<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/aha_io\" rel=\"nofollow\">@aha_io<\/a><\/td>\n<td align=\"left\">$69\/mo, Ask about startup plan<\/td>\n<td align=\"left\">The new way to create brilliant product strategy and visual roadmaps.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sprint.ly\" rel=\"nofollow\">Sprintly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sprintly\" rel=\"nofollow\">@sprintly<\/a><\/td>\n<td align=\"left\">$49\/mo- $399\/mo<\/td>\n<td align=\"left\">Don't ask how projects are going. Watch how they're going in real-time. Use our elegant interface to prioritize, tag, manage, estimate, and measure your software developers' progress in real-time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/podio.com\" rel=\"nofollow\">Podio<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Podio\" rel=\"nofollow\">@Podio<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Teamwork made easy. News and views from the team behind Podio - changing the way the world works since 2009.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getflow.com\" rel=\"nofollow\">Flow<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/flowapp\" rel=\"nofollow\">@flowapp<\/a><\/td>\n<td align=\"left\">$19\/mo - $249\/mo<\/td>\n<td align=\"left\">Stop managing projects from your inbox. Flow is a collaborative task management app for the web and iPhone.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/basecamp.com\" rel=\"nofollow\">Basecamp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/37signals\" rel=\"nofollow\">@37signals<\/a><\/td>\n<td align=\"left\">$20\/mo - $150\/mo<\/td>\n<td align=\"left\">The official account for Basecamp\u00ae. Helping Basecamp customers every Mon-Fri 9am-6pm CT! Basecamp to help organize the store design, develop fixtures, and manage craftspeople. Primarily through word-of-mouth alone, Basecamp has become the world\u2019s #1 project management tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.apollohq.com\" rel=\"nofollow\">Apollo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/applicomhq\" rel=\"nofollow\">@applicomhq<\/a><\/td>\n<td align=\"left\">$23\/mo - $148\/mo<\/td>\n<td align=\"left\">Integrated Project and Contact Management Done Right Apollo is project and contact management done right. Using Apollo, you will realise that it's built to help you get things done, quickly and efficiently. With Apollo, you will always know where your projects, your contacts and your life are at and you will feel on top of everything \u2014 regardless of how hectic your schedule is.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pivotaltracker.com\" rel=\"nofollow\">Pivotal Tracker<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/pivotaltracker\" rel=\"nofollow\">@pivotaltracker<\/a><\/td>\n<td align=\"left\">$7\/mo - $175\/mo<\/td>\n<td align=\"left\">BUILD BETTER SOFTWARE FASTER. Break your project down into bite-sized stories, which get your product closer to the business goal. Use points to estimate each story\u2019s relative complexity and prioritize it in the backlog.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/asana.com\" rel=\"nofollow\">Asana<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/asana\" rel=\"nofollow\">@asana<\/a><\/td>\n<td align=\"left\">$50\/mo - $800\/mo<\/td>\n<td align=\"left\">Teamwork without email. Asana is our go-to for prioritizing projects, keeping up w\/orders &amp; staying on top of a growing to-do list<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/weekplan.net\" rel=\"nofollow\">WeekPlan<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/weekplan\" rel=\"nofollow\">@weekplan<\/a><\/td>\n<td align=\"left\">$7\/mo - $19\/mo<\/td>\n<td align=\"left\">Time management inspired by the \"7 habits of highly effective people\" Features: goals of the week, week view and quadrant matrix, pomodoro timer, shared workspaces, etc...<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/trello.com\" rel=\"nofollow\">Trello<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/trello\" rel=\"nofollow\">@trello<\/a><\/td>\n<td align=\"left\">$5\/mo<\/td>\n<td align=\"left\">Organize anything, together Trello is the fastest, easiest way to organize anything, from your day-to-day work, to a favorite side project, to your greatest life plans.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.blossom.co\" rel=\"nofollow\">Blossom<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/blossom_io\" rel=\"nofollow\">@blossom_io<\/a><\/td>\n<td align=\"left\">$19\/mo - $149\/mo<\/td>\n<td align=\"left\">Agile Project Management. Blossom gives each member of the team clear overview about who\u2019s doing what &amp; why and at the same time it helps you to focus on what matters most. With Blossom you can efficiently manage your whole development process in one place, built with simplicity in mind. Blossom is based on the principles ofKanban, a way of working that emphasizes iterative delivery cycles and continuously improves the workflow of your team or organization.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.redmine.org\" rel=\"nofollow\">Redmine<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Flexible project management web application. Written using Ruby on Rails framework, it is cross-platform and cross-database.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/jira\/agile\" rel=\"nofollow\">JIRA Agile<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/JIRA\" rel=\"nofollow\">@jira<\/a><\/td>\n<td align=\"left\">$10\/mo - $30\/mo<\/td>\n<td align=\"left\">Dream big, work smart, deliver fast. Makers of @JIRA, @Confluence, @Bitbucket and more. Software to plan, collaborate, code, and support. Built for teams<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tomsplanner.com\" rel=\"nofollow\">Tom's Planner<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tomsplanner\" rel=\"nofollow\">@tomsplanner<\/a><\/td>\n<td align=\"left\">$9\/mo - $19\/mo<\/td>\n<td align=\"left\">Tom's Planner is online Gantt chart software that allows anyone to create, collaborate and share Gantt Charts online with drag and drop simplicity. It's web based, extremely intuitive and easy-to-use.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/leankit.com\" rel=\"nofollow\">LeanKit<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/leankit\" rel=\"nofollow\">@leankit<\/a><\/td>\n<td align=\"left\">$15\/mo - $19\/mo<\/td>\n<td align=\"left\">Instant Project visibility. In LeanKit, you map your organization\u2019s processes onto virtual whiteboards. On each board the process steps are represented as vertical and horizontal lanes. Cards represent work items, which team members update and move from across the board as they complete their share of the work. Rather than having to ask for status reports, managers and customers can just look at the board. Board updates are visible in seconds around the globe and e-mail alerts and RSS feeds are available, so you and your team can take immediate action to resolve issues before they turn into serious problems.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.breeze.pm\" rel=\"nofollow\">Breeze<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/BreezeTeam\" rel=\"nofollow\">@BreezeTeam<\/a><\/td>\n<td align=\"left\">$29\/mo - $129\/mo<\/td>\n<td align=\"left\">Organize and track everything. Breeze shows you what's being worked on, who's working on what, where things are in the workflow and how much time it took.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/cushionapp.com\/\" rel=\"nofollow\">Cushion<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/cushionapp\" rel=\"nofollow\">@cushionapp<\/a><\/td>\n<td align=\"left\">$8\/mo - $48\/mo<\/td>\n<td align=\"left\">Piece of mind for freelancers. We built Cushion as a team of freelancers who always had trouble planning and managing our year\u2014constantly taking on too much work and never thinking to take a break.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/clickup.com\" rel=\"nofollow\">Clickup<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/clickup_app\" rel=\"nofollow\">@clickup_app<\/a><\/td>\n<td align=\"left\">$0 - $5\/mo<\/td>\n<td align=\"left\">ClickUp's #1 rated productivity software is making more productive projects with a beautifully designed and intuitive platform.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-time-tracking\" class=\"anchor\" aria-hidden=\"true\" href=\"#time-tracking\"><\/a>Time Tracking<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/toggl.com\" rel=\"nofollow\">Toggl<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/toggl\" rel=\"nofollow\">@toggl<\/a><\/td>\n<td align=\"left\">Free - $59\/mo<\/td>\n<td align=\"left\">Toggl\u2019s time tracker is built for speed and ease of use. Time logging with Toggl is so simple that you\u2019ll actually use it.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/clockify.me\" rel=\"nofollow\">Clockify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/clockify\" rel=\"nofollow\">@clockify<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Clockify is the only 100% Free time tracking software that lets you and your team track time with one click. It works like Toggl but offers unlimited features and unlimited users.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hubstaff.com\" rel=\"nofollow\">Hubstaff<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Hubstaff\" rel=\"nofollow\">@hubstaff<\/a><\/td>\n<td align=\"left\">Free - $9\/mo<\/td>\n<td align=\"left\">Hubstaff is time tracking software designed to make remote team management more effective and efficient. You just have to sign up, download our intuitive desktop app and push the start button to begin tracking time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tickspot.com\/\" rel=\"nofollow\">Tickspot<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tickspot\" rel=\"nofollow\">@tickspot<\/a><\/td>\n<td align=\"left\">Free - $149\/mo<\/td>\n<td align=\"left\">Straightforward time tracking software to help your team run more profitable projects. Whether you prefer iOS, Android, the Apple Watch, or your desktop computer, Tick is the easiest way to track your time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.kimai.org\/\" rel=\"nofollow\">Kimai<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kimai_org\" rel=\"nofollow\">@kimai_org<\/a><\/td>\n<td align=\"left\">self-hosted<\/td>\n<td align=\"left\">Kimai is a Free open source timetracker. It tracks work time and prints out a summary of your activities on demand. Yearly, monthly, daily, by customer, by project, by action, etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/samg\/timetrap\">Timetrap<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">simple open source command line time tracking<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-app-developer-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#app-developer-tools\"><\/a>App Developer Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.appannie.com\" rel=\"nofollow\">App Annie<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/appannie\/\" rel=\"nofollow\">@appannie\/<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Make better decisions with our app store data. We are the leader for app store analytics, app rankings, and market intelligence.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/appfigures.com\" rel=\"nofollow\">App Figures<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/appfigures\" rel=\"nofollow\">@appfigures<\/a><\/td>\n<td align=\"left\">$9\/mo<\/td>\n<td align=\"left\">A beautiful app tracking platform for app developers and publishers.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-localization--internationalization\" class=\"anchor\" aria-hidden=\"true\" href=\"#localization--internationalization\"><\/a>Localization &amp; Internationalization<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/localizejs.com\" rel=\"nofollow\">Localize.js<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$25\/mo - $150\/mo<\/td>\n<td align=\"left\">Website translation with just a few lines of code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gengo.com\" rel=\"nofollow\">Gengo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0.06 - $0.17\/word<\/td>\n<td align=\"left\">Human translation API<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/translate.google.com\/manager\/website\/\" rel=\"nofollow\">Google Website Translator<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Make your website instantly available in 90+ languages<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-business--traffic-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#business--traffic-analytics\"><\/a>Business &amp; Traffic Analytics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kissmetricshq.com\" rel=\"nofollow\">KISSmetrics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kissmetrics\/\" rel=\"nofollow\">@kissmetrics\/<\/a><\/td>\n<td align=\"left\">$150\/mo - $500\/mo<\/td>\n<td align=\"left\">KISSmetrics tells you who\u2019s doing it. Every last piece gets connected to a real person. All of it. It doesn\u2019t matter if people bounce around between different browsers and devices. Or even if it takes them 6 months to come back. You\u2019ll see what real people do.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.localytics.com\" rel=\"nofollow\">Localytics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/localytics\/\" rel=\"nofollow\">@localytics\/<\/a><\/td>\n<td align=\"left\">Free up to 10k MAUs, $200\/mo - $2700\/mo above that<\/td>\n<td align=\"left\">Find out what works in your mobile or web app. Do more of it. All in one place. Advanced analytics provide data and insight to help you build more successful apps. Integrated Marketing helps you easily engage and acquire more customers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mixpanel.com\" rel=\"nofollow\">Mixpanel<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mixpanel\/\" rel=\"nofollow\">@mixpanel\/<\/a><\/td>\n<td align=\"left\">$150\/mo - $2000\/mo<\/td>\n<td align=\"left\">Actions speak louder than page views. For years, companies have pushed page views as a primary measure of success. Page view counts are popular because they are easy to report, but ultimately cannot tell you how engaged your visitors are. Mixpanel lets you measure what customers do in your app by reporting actions, not page views.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/amplitude.com\" rel=\"nofollow\">Amplitude<\/a><\/td>\n<td align=\"left\">$299\/mo<\/td>\n<td align=\"left\">Mobile Analytics for decision makers<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/snowplowanalytics.com\" rel=\"nofollow\">Snowplow<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SnowPlowData\" rel=\"nofollow\">@SnowPlowData<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Snowplow is the most powerful, flexible, scalable web analytics platform in the world. Snowplow enables analysts to perform a wide variety of both simple and sophisticated analytics on your web analytics data.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/segment.com\" rel=\"nofollow\">Segment<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$29\/mo - $349\/mo<\/td>\n<td align=\"left\">The right way to manage your tools. The idea is simple: one pipeline for all your data. Send data to any third-party tool with a single integration.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/clicky.com\" rel=\"nofollow\">Clicky<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/clicky\" rel=\"nofollow\">@clicky<\/a><\/td>\n<td align=\"left\">$9.99\/mo - $19.99\/mo<\/td>\n<td align=\"left\">Real Time Web Analytics. Clicky lets you see every visitor and every action they take on your web site, with the option to attach custom data to visitors, such as usernames or email addresses. Analyze each visitor individually and see their full history.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.google.com\/analytics\/\" rel=\"nofollow\">Google Analytics<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Google Analytics lets you measure your advertising ROI as well as track your Flash, video, and social networking sites and applications.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/matomo.org\" rel=\"nofollow\">Matomo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/matomo_org\" rel=\"nofollow\">@matomo_org<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Whether you are an individual blogger, a small business, or a large corporation, Matomo helps you gain valuable insights to help your business or readership grow.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/chartio.com\" rel=\"nofollow\">Chartio<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chartio\" rel=\"nofollow\">@chartio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Visualize and explore your data with Chartio. Create interactive charts and perfect dashboards through an intuitive drag and drop interface. Switch from basic tables to sophisticated data visualizations in a single click. Powerful filters let you slice and dice your data, and you can drill down into most charts without configuring a thing.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/chartbeat.com\" rel=\"nofollow\">Chartbeat<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chartbeat\" rel=\"nofollow\">@chartbeat<\/a><\/td>\n<td align=\"left\">$9.95\/mo - $49.95\/mo<\/td>\n<td align=\"left\">Build a loyal and valuable audience for your site. Chartbeat's real-time traffic and audience-behavior data shows you who's on your site and how they're engaging with your content right now \u2014 so you can take action on what matters when it matters.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/calq.io\" rel=\"nofollow\">Calq<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/CalqAnalytics\" rel=\"nofollow\">@CalqAnalytics<\/a><\/td>\n<td align=\"left\">$0 - $2500\/mo<\/td>\n<td align=\"left\">Advanced custom analytics for mobile and web applications. Calq is an analytics platform that measures user actions rather than page views. An action can be anything a user does: reviewing a product, playing a level on a mobile game, making a purchase on your site, anything. Calq's ability to work with custom events AND custom data is what raises it above more traditional analytics platforms.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.gosquared.com\" rel=\"nofollow\">GoSquared<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/GoSquared\" rel=\"nofollow\">@gosquared<\/a><\/td>\n<td align=\"left\">\u00a321.60 - \u00a3396\/mo<\/td>\n<td align=\"left\">Easy to use real-time web analytics.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.improvely.com\" rel=\"nofollow\">Improvely<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/improvelycom\" rel=\"nofollow\">@improvelycom<\/a><\/td>\n<td align=\"left\">$29 - $899\/mo<\/td>\n<td align=\"left\">Conversion tracking and click fraud monitoring platform. The easiest way to track the performance of marketing campaigns and monitor them for click fraud.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/keen.io\" rel=\"nofollow\">Keen IO<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/keen_io\" rel=\"nofollow\">@keen_io<\/a><\/td>\n<td align=\"left\">$0 - $2000+\/mo<\/td>\n<td align=\"left\">Custom analytics shouldn't be a pain in the backend. Keen IO's powerful APIs do the heavy lifting for you, so you can gather all the data you want and start getting the answers you need.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/heapanalytics.com\" rel=\"nofollow\">Heap Analytics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/heap\" rel=\"nofollow\">@heap<\/a><\/td>\n<td align=\"left\">0 - $599+<\/td>\n<td align=\"left\">Instant, retroactive analytics for web and iOS. No code required.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/get.gaug.es\" rel=\"nofollow\">Gauges<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gaugesapp\" rel=\"nofollow\">@GaugesApp<\/a><\/td>\n<td align=\"left\">$6-$48\/mo<\/td>\n<td align=\"left\">Gauges provides real time web analytics such as how many people visit your site, where they come from, and where they go.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getWisdom.io\" rel=\"nofollow\">Wisdom<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free - $2000+\/mo<\/td>\n<td align=\"left\">Session Replay. Wisdom is the most accurate live visitor session recorder service available. Focusing only on session replay, Wisdom reconstructs a virtual desktop screen for every visitor, across all tabs, to capture the true feel of every visitor's experience.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-conversion-optimization--ab-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#conversion-optimization--ab-testing\"><\/a>Conversion Optimization &amp; A\/B Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.optimizely.com\" rel=\"nofollow\">Optimizely<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/optimizely\" rel=\"nofollow\">@optimizely<\/a><\/td>\n<td align=\"left\">$17\/mo - $359\/mo<\/td>\n<td align=\"left\">A\/B testing you'll actually use. Track engagement, clicks, conversions, sign ups, or anything else that matters to you and your business. Optimizely's custom goal tracking provides an endless range of measurable actions that you can define. Just tell Optimizely what to measure, and we will do the rest.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/vwo.com\" rel=\"nofollow\">Visual Website Optimizer<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wingify\" rel=\"nofollow\">@wingify<\/a><\/td>\n<td align=\"left\">$49\/mo - $129\/mo<\/td>\n<td align=\"left\">Increase your website sales and conversions. Using Visual Website Optimizer, they A\/B test different versions of their website and landing pages to find out which one works best. Made for marketers, our tool is incredibly easy to use, and doesn't need IT resources.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.eyequant.com\" rel=\"nofollow\">EyeQuant<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/eyequant\" rel=\"nofollow\">@eyequant<\/a><\/td>\n<td align=\"left\">$199\/mo - $999\/mo<\/td>\n<td align=\"left\">Instantly understand what your visitors will see and miss in their first seconds on your site, and improve your conversions. Analyse live sites or mockups within seconds, no code required.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/optimize.google.com\" rel=\"nofollow\">Optimize by Google<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Optimize allows you to test variants of web pages and see how they perform against an objective that you specify.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-user-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-management\"><\/a>User Management<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/developer.okta.com\/\" rel=\"nofollow\">Okta<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/OktaDev\" rel=\"nofollow\">@OktaDev<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Okta adds authentication, authorization, and user management to your web or mobile app within minutes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/auth0.com\" rel=\"nofollow\">Auth0<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/c2id.net\" rel=\"nofollow\">Connect2id<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/connect2id\" rel=\"nofollow\">@Connect2id<\/a><\/td>\n<td align=\"left\">\u20ac299\/mo - \u20ac999\/mo<\/td>\n<td align=\"left\">Hosted Connect2id server \u2013 a flexible, secure and certified OpenID Connect\/OAuth 2.0 identity provider.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-user-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-testing\"><\/a>User Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/silverbackapp.com\" rel=\"nofollow\">Silverback 2.0<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/silverbackapp\" rel=\"nofollow\">@silverbackapp<\/a><\/td>\n<td align=\"left\">69.95<\/td>\n<td align=\"left\">Guerrilla usability testing software for designers and developers. Silverback makes it easy, quick and cheap for everyone to perform guerrilla usability tests with no setup and no expense, using hardware already in your Mac.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hotjar.com\" rel=\"nofollow\">HotJar<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/hotjar\" rel=\"nofollow\">@hotjar<\/a><\/td>\n<td align=\"left\">Free - $29\/mo (personal)<\/td>\n<td align=\"left\">Records videos and collects heatmaps of your site visitor actions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getWisdom.io\" rel=\"nofollow\">Wisdom<\/a><\/td>\n<td align=\"left\">Free - $2000+\/mo<\/td>\n<td align=\"left\">Session Replay<\/td>\n<td align=\"left\">Wisdom is the most accurate live visitor session recorder service available. Focusing only on session replay, Wisdom reconstructs a virtual desktop screen for every visitor, across all tabs, to capture the true feel of every visitor's experience.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-hr\" class=\"anchor\" aria-hidden=\"true\" href=\"#hr\"><\/a>HR<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.workday.com\" rel=\"nofollow\">Workday<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/workday\" rel=\"nofollow\">@workday<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Workday works the way people work\u2014collaboratively, on the go, and in real-time. Explore the product previews below to learn how Workday can change the way you work. With powerful business applications and a user experience that's unmatched in enterprise software, Workday gives you everything you need to transform your business.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.lever.co\" rel=\"nofollow\">Lever<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/lever\" rel=\"nofollow\">@lever<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">A modern web app for hiring. Leverage your entire company \u2013 interviewers, managers, and recruiters \u2013 to source, vet, and close.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zenefits.com\" rel=\"nofollow\">Zenefits<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zenefits\" rel=\"nofollow\">@zenefits<\/a><\/td>\n<td align=\"left\">$0\/mo<\/td>\n<td align=\"left\">The #1 All-In-One HR Platform. Payroll. Benefits. Time. Compliance. All online, all in one place.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.testdome.com\/\" rel=\"nofollow\">TestDome<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/TestDome\" rel=\"nofollow\">@TestDome<\/a><\/td>\n<td align=\"left\">$8\/candidate - $20\/candidate<\/td>\n<td align=\"left\">Automated testing of programming skills, ask candidates to write real code before calling them for an interview.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hackerrank.com\/\" rel=\"nofollow\">HackerRank<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/hackerrank\" rel=\"nofollow\">@hackerrank<\/a><\/td>\n<td align=\"left\">paid<\/td>\n<td align=\"left\">End-to-end technical recruiting platform for hiring engineers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/peopledoc.com\" rel=\"nofollow\">PeopleDoc<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PeopleDoc_Inc\" rel=\"nofollow\">@PeopleDoc_Inc<\/a><\/td>\n<td align=\"left\">\n<\/td><td align=\"left\">HR Service delivery platform designed to simplify complex HR operations, compliance and can easily serve employees anytime, anywhere.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.bamboohr.com\/\" rel=\"nofollow\">BambooHR<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bamboohr\" rel=\"nofollow\">@bamboohr<\/a><\/td>\n<td align=\"left\">paid<\/td>\n<td align=\"left\">HR software with heart.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/peopledoc.com\" rel=\"nofollow\">PeopleDoc<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PeopleDoc_Inc\" rel=\"nofollow\">@PeopleDoc_Inc<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">HR Service delivery platform designed to simplify complex HR operations, compliance and can easily serve employees anytime, anywhere.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.bamboohr.com\" rel=\"nofollow\">BambooHR<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bamboohr\" rel=\"nofollow\">@bamboohr<\/a><\/td>\n<td align=\"left\">paid<\/td>\n<td align=\"left\">HR software with heart.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hiringplan.io\" rel=\"nofollow\">HiringPlan<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">HiringPlan is a headcount planning tool for startups with built-in market data. It helps founders adopt a structured compensation plan, pay people fairly and preserve cash and equity runway.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/tldroptions.io\" rel=\"nofollow\">TLDROptions<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">A tool for helping employees understand the potential value of their stock options.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-payroll\" class=\"anchor\" aria-hidden=\"true\" href=\"#payroll\"><\/a>Payroll<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/gusto.com\" rel=\"nofollow\">Gusto<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gustohq\" rel=\"nofollow\">@gustohq<\/a><\/td>\n<td align=\"left\">$29\/mo + $6\/user<\/td>\n<td align=\"left\">Payroll and benefits that put people first, easy setup, automated tax filings and thoughtful support.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/wagepoint.com\" rel=\"nofollow\">WagePoint<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wagepoint\" rel=\"nofollow\">@wagepoint<\/a><\/td>\n<td align=\"left\">$20 + $2\/employy<\/td>\n<td align=\"left\">The Simple, Fast &amp; Friendly way to pay your employees.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-continuous-integrationcode-quality\" class=\"anchor\" aria-hidden=\"true\" href=\"#continuous-integrationcode-quality\"><\/a>Continuous Integration\/Code Quality<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/travis-ci.org\" rel=\"nofollow\">Travis<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/travisci\" rel=\"nofollow\">@travisci<\/a><\/td>\n<td align=\"left\">Free - $489\/mo<\/td>\n<td align=\"left\">Hi I\u2019m Travis CI, a hosted continuous integration service for open source and private projects: <a href=\"https:\/\/travis-ci.com\/\" rel=\"nofollow\">travis-ci.com<\/a>  System status updates: <a href=\"https:\/\/twitter.com\/traviscistatus\" rel=\"nofollow\">@traviscistatus<\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.appveyor.com\" rel=\"nofollow\">AppVeyor<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/appveyor\" rel=\"nofollow\">@appveyor<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">AppVeyor automates building, testing and deployment of .NET applications, helping your team to focus on delivering great apps.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/codeship.com\" rel=\"nofollow\">Codeship<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/codeship\" rel=\"nofollow\">@codeship<\/a><\/td>\n<td align=\"left\">Start with 100 builds\/mo Free, Unlimited plans start at 49$<\/td>\n<td align=\"left\">Continuous Delivery as a service, start testing and deploying your code immediately<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/circleci.com\" rel=\"nofollow\">Circle<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/circleci\" rel=\"nofollow\">@circleci<\/a><\/td>\n<td align=\"left\">$19\/mo - $269\/mo<\/td>\n<td align=\"left\">Ship better code, faster. Easy, fast, continuous integration and deployment for web apps.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/nevercode.io\" rel=\"nofollow\">Nevercode<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/nevercodehq\" rel=\"nofollow\">@nevercodehq<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/houndci.com\" rel=\"nofollow\">Hound<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/houndci\" rel=\"nofollow\">@houndci<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Automated Code Review. Take care of pesky code reviews with a trusty Hound<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/codeclimate.com\" rel=\"nofollow\">CodeClimate<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/codeclimate\" rel=\"nofollow\">@codeclimate<\/a><\/td>\n<td align=\"left\">$0\/mo - $399\/mo<\/td>\n<td align=\"left\">Automated  Code Review. Code Climate hosted software metrics help you ship quality Ruby and JavaScript code faster. Get control of your technical debt with real time static analysis of your code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.codacy.com\" rel=\"nofollow\">Codacy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0 - $150\/mo<\/td>\n<td align=\"left\">Automated Code Review. Continuous Static Analysis designed to complement your unit tests. Similar to CodeClimate.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/codecov.io\" rel=\"nofollow\">Codecov<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0 - $5\/mo<\/td>\n<td align=\"left\">Hosted Code Coverage. Code coverage reporting done right.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/semaphoreci.com\" rel=\"nofollow\">Semaphore<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/semaphoreci\" rel=\"nofollow\">@semaphoreci<\/a><\/td>\n<td align=\"left\">$14\/mo - $899\/mo<\/td>\n<td align=\"left\">Create an Amazing Workflow. Semaphore assumes that your private or open source project is on GitHub. There are no new dependencies, hooks or SSH keys to manage. It works without any change in source code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.solanolabs.com\" rel=\"nofollow\">Solano CI<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/solanolabs\" rel=\"nofollow\">@SolanoLabs<\/a><\/td>\n<td align=\"left\">$15\/mo - $100\/mo<\/td>\n<td align=\"left\">Faster Continuous Integration and Deployment with patented auto-parallelization. Solano CI sets up Continuous Integration in minutes, Frees you from managing a build server, and lets you deploy software 10x - 80x faster by running tests in parallel safely and automatically. It also lets you use our massively scalable environment even before you push to CI. Seamlessly integrates into existing workflows. Free 14-day trial, no credit card required. Formerly loved as tddium.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/jenkins.io\" rel=\"nofollow\">Jenkins<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/jenkinsci\" rel=\"nofollow\">@jenkinsci<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Jenkins provides continuous integration services for software development. It is a server-based system that supports SCM tools including AccuRev, CVS, Subversion, Git, Mercurial, Perforce, Clearcase and RTC, and can execute Apache Ant and Apache Maven based projects as well as arbitrary shell scripts and Windows batch commands. Released under the MIT License, Jenkins is Free software.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/bamboo\" rel=\"nofollow\">Bamboo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/atlassian\" rel=\"nofollow\">@atlassian<\/a><\/td>\n<td align=\"left\">$10\/mo - $1000\/mo<\/td>\n<td align=\"left\">Bamboo does more than just run builds and tests. It connects issues, commits, test results, and deploys so the whole picture is available to your entire product team \u2013 from project managers, to devs and testers, to sys admins.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/buildkite.com\" rel=\"nofollow\">Buildkite (Buildbox)<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/buildkite\" rel=\"nofollow\">@buildkite<\/a><\/td>\n<td align=\"left\">$15\/dev\/mo<\/td>\n<td align=\"left\">Semi-hosted continuous integration and deployment. Buildkite uses your own infrastructure to run builds so you can test any language or run any deployment scripts. You can run as many parallel agents (and builds) as you want.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/crucible\" rel=\"nofollow\">Crucible<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/atlassian\" rel=\"nofollow\">@atlassian<\/a><\/td>\n<td align=\"left\">$10\/mo - $8000\/mo<\/td>\n<td align=\"left\">Code review system. Review code, discuss changes, share knowledge, and identify defects with Crucible's flexible review workflow. It's code review made easy for Git, Subversion, CVS, Perforce, and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/coveralls.io\" rel=\"nofollow\">Coveralls<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/coverallsapp\" rel=\"nofollow\">@coverallsapp<\/a><\/td>\n<td align=\"left\">$0\/mo - $50\/mo<\/td>\n<td align=\"left\">Coveralls works with your continuous integration server to give you test coverage history and statistics. It integrates with any langauge and is Free for open source.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/about.testributor.com\" rel=\"nofollow\">Testributor<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/testributor\" rel=\"nofollow\">@testributor<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Testributor is an open source Continuous Integration platform. A hosted version is available for Free, both for open source and private projects.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.wercker.com\" rel=\"nofollow\">Wercker<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wercker\" rel=\"nofollow\">@wercker<\/a><\/td>\n<td align=\"left\">$0\/mo - $350\/mo<\/td>\n<td align=\"left\">Wercker is a Docker-Native CI\/CD Automation platform for Kubernetes &amp; Microservice Deployments.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/monkeytest.it\" rel=\"nofollow\">Monkey Test It<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/monkeytestit\" rel=\"nofollow\">@monkeytestit<\/a><\/td>\n<td align=\"left\">$0\/mo - $199\/mo<\/td>\n<td align=\"left\">Automated website checking tool to catch bugs on your website (broken links, missing images, JS errors, etc) before your users do. Integrates with Slack, most CI systems, webooks, and supports built-in scheduling.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/concourse-ci.org\" rel=\"nofollow\">Concourse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/concourseci\" rel=\"nofollow\">@concourseci<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Concourse is an open-source continuous thing-doer. Built on the simple mechanics of resources, tasks, and jobs, Concourse presents a general approach to automation that makes it great for CI\/CD.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-dashboards\" class=\"anchor\" aria-hidden=\"true\" href=\"#dashboards\"><\/a>Dashboards<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.geckoboard.com\" rel=\"nofollow\">Geckoboard<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/geckoboard\" rel=\"nofollow\">@geckoboard<\/a><\/td>\n<td align=\"left\">$17\/mo - $899\/mo<\/td>\n<td align=\"left\">Meet Geckoboard. It's Your Key Data, In One Place. Geckoboard monitors your business\u2019s vital signs \u2013 don\u2019t wait, see it live on a business dashboard as it happens. Focus on what matters and react faster to important events.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.telemetrytv.com\" rel=\"nofollow\">Telemetry<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/telemetrytv\" rel=\"nofollow\">@telemetrytv<\/a><\/td>\n<td align=\"left\">$9\/mo - $749\/mo<\/td>\n<td align=\"left\">Build realtime dashboard with powerful visualizations that look beautiful on big screen TVs, desktop computers, mobile devices, and embedded systems\u2014all using a simple REST API that works with all modern languages.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/dashing.io\" rel=\"nofollow\">Dashing<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Dashing is a Sinatra based framework that lets you build beautiful dashboards.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.klipfolio.com\" rel=\"nofollow\">Klipfolio<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/klipfolio\" rel=\"nofollow\">@klipfolio<\/a><\/td>\n<td align=\"left\">$5\/user\/mo - $20\/user\/mo<\/td>\n<td align=\"left\">Meet Your Business Dashboard. Connect to any data service to bring your key numbers together on one dashboard. Assign your data to visualizations to show the story behind the numbers. Cultivate a data-driven culture by sharing dashboards with everyone on your team.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/grafana.com\" rel=\"nofollow\">Grafana<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/grafana\" rel=\"nofollow\">@grafana<\/a><\/td>\n<td align=\"left\">$0 - $90\/mo (+9\/user\/mo)<\/td>\n<td align=\"left\">The analytics platform for all your metrics. Grafana allows you to query, visualize, alert on and understand your metrics no matter where they are stored. Create, explore, and share dashboards with your team and foster a data driven culture.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/redash.io\" rel=\"nofollow\">Redash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Self-hosted or $49\/mo - 450\/mo<\/td>\n<td align=\"left\">\"Redash has support for querying multiple databases, including: Redshift, Google BigQuery, PostgreSQL, MySQL, Graphite, Presto, Google Spreadsheets, Cloudera Impala, Hive and custom scripts.\"<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cyfe.com\" rel=\"nofollow\">Cyfe<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Cyfe\" rel=\"nofollow\">@cyfe<\/a><\/td>\n<td align=\"left\">$0 - $29\/mo<\/td>\n<td align=\"left\">Cyfe offers all-in-one web dashboards for your clients to help them monitor and analyze data scattered across all their online services like Google Analytics, Salesforce, AdSense, MailChimp, Amazon, Facebook, WordPress, Twitter and more from one single location in real-time.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-errorexception-handling\" class=\"anchor\" aria-hidden=\"true\" href=\"#errorexception-handling\"><\/a>Error\/Exception Handling<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/catchjs.com\" rel=\"nofollow\">CatchJS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$49 - $499<\/td>\n<td align=\"left\">JavaScript error reporting for web apps. Can take screenshots and collect click trails, so reproducing bugs becomes easy.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/try.crashlytics.com\" rel=\"nofollow\">Crashlytics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/crashlytics\" rel=\"nofollow\">@crashlytics<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Crash reports and grouping for easier analysis. Basic analytics and reports. iOS &amp; Android<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sentry.io\/welcome\/\" rel=\"nofollow\">Sentry<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getsentry\" rel=\"nofollow\">@getsentry<\/a><\/td>\n<td align=\"left\">Open Source &amp; $24\/mo - $199\/mo<\/td>\n<td align=\"left\">Sentry notifies you when your users experience errors. Know immediately when things happen in your application. Engage users before they have a chance to report a problem.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.honeybadger.io\" rel=\"nofollow\">HoneyBadger<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/honeybadgerapp\" rel=\"nofollow\">@honeybadgerapp<\/a><\/td>\n<td align=\"left\">$39\/mo - $249\/mo<\/td>\n<td align=\"left\">Exception, uptime, and performance monitoring for Ruby. It tells you about errors, downtime and performance issues as they happen. And it gives you the tools you need to fix them ...without burying you in data. Without silly rate limits or per-server fees.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.bugsnag.com\" rel=\"nofollow\">BugSnag<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/bugsnag\" rel=\"nofollow\">@bugsnag<\/a><\/td>\n<td align=\"left\">$29\/mo - $249\/mo<\/td>\n<td align=\"left\">Automatic, full-stack error monitoring. Web app monitoring for Rails, PHP, Node.js, Java, and every other leading platform.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/raygun.com\" rel=\"nofollow\">Raygun<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/raygunio\" rel=\"nofollow\">@raygunio<\/a><\/td>\n<td align=\"left\">$14\/mo - $199\/mo<\/td>\n<td align=\"left\">Exceptional Error Tracking. Your software faults get automatically sent to the Raygun service and analysis begins immediately. Raygun intelligently groups your errors so you're dealing with root causes, not every single error instance!<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/airbrake.io\" rel=\"nofollow\">Airbrake<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/airbrake\" rel=\"nofollow\">@airbrake<\/a><\/td>\n<td align=\"left\">$49 - $249\/mo<\/td>\n<td align=\"left\">No More Searching Log Files Capture and Track Your Application's Exceptions in 3 Minutes. Airbrake is the leading exception reporting service, currently providing error tracking for 50,000 applications with support for 18 programming languages.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/rollbar.com\" rel=\"nofollow\">Rollbar<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/rollbar\" rel=\"nofollow\">@rollbar<\/a><\/td>\n<td align=\"left\">$12\/mo - $1249\/mo<\/td>\n<td align=\"left\">Take control of your errors. Rollbar is platform-agnostic and can accept data from anything that can speak HTTP and JSON. You can use our official libraries for Ruby, Python, PHP, Node.js, JavaScript, Android, iOS, or Flash, or roll your own with our API.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/errorception.com\" rel=\"nofollow\">Errorception<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/errorception\" rel=\"nofollow\">@errorception<\/a><\/td>\n<td align=\"left\">$5\/mo - $59\/mo<\/td>\n<td align=\"left\">Painless JavaScript Error Tracking. Errorception is a simple and painless way to find out about JavaScript errors, as they occur in your users' browsers. All you need to do is insert a script tag on your page, and you will start recording errors as they happen in real-time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/errbit.com\" rel=\"nofollow\">Errbit<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">OSS<\/td>\n<td align=\"left\">The open source error catcher that's Airbrake API compliant.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.overops.com\" rel=\"nofollow\">OverOps<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/overopshq\" rel=\"nofollow\">@overopshq<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">God Mode in Production Code for java and scala applications.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-application-distribution\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-distribution\"><\/a>Application Distribution<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.hockeyapp.net\" rel=\"nofollow\">HockeyApp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/hockeyapp\" rel=\"nofollow\">@hockeyapp<\/a><\/td>\n<td align=\"left\">Free - $129\/mo depending on number of apps and number of owners<\/td>\n<td align=\"left\">Distribution of iOS, Android, Windows Phone and Mac OS apps. Includes analytics, user feedback and crash reports.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/setapp.com\" rel=\"nofollow\">Setapp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/setapp\" rel=\"nofollow\">@setapp<\/a><\/td>\n<td align=\"left\">$9.99\/mo<\/td>\n<td align=\"left\">Setapp is a single subscription with a huge collection of quality apps for macOS<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-log-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#log-monitoring\"><\/a>Log Monitoring<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.fluentd.org\" rel=\"nofollow\">Fluentd<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/fluentd\" rel=\"nofollow\">@fluentd<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Set Up Once, Collect More. Fluentd is an open source data collector designed for processing data streams. 150+ plugins instantly enable you to store the data for Log Management, Big Data Analytics, etc<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/cloudera\/flume\">Flume<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.graylog.org\" rel=\"nofollow\">Graylog<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/graylog2\" rel=\"nofollow\">@graylog2<\/a><\/td>\n<td align=\"left\">Field-tested open source data analytics system used and trusted all around the world. Search your logs, create charts, send reports and be alerted when something happens. All running on the existing JVM in your datacenter.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/logentries.com\" rel=\"nofollow\">LogEntries<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/logentries\" rel=\"nofollow\">@logentries<\/a><\/td>\n<td align=\"left\">$16\/mo - $245\/mo<\/td>\n<td align=\"left\">Log Management &amp; Analytics Made Easy. Logentries provides an easy-to-use cloud service for log management and analytics.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.loggly.com\" rel=\"nofollow\">Loggly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/loggly\" rel=\"nofollow\">@loggly<\/a><\/td>\n<td align=\"left\">$49\/mo - $349\/mo<\/td>\n<td align=\"left\">Solve operational problems faster. Loggly helps cloud-centric organizations\u2014organizations that build and manage cloud-facing applications\u2014to solve operational problems faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.elastic.co\/products\/logstash\" rel=\"nofollow\">Logstash<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/logstash\" rel=\"nofollow\">@logstash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Ship logs from any source, parse them, get the right timestamp, index them, and search them. Logstash is a tool for managing events and logs. You can use it to collect logs, parse them, and store them for later use (like, for searching). Speaking of searching, Logstash comes with a web interface for searching and drilling into all of your logs.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/papertrailapp.com\" rel=\"nofollow\">Papertrail<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/papertrailapp\" rel=\"nofollow\">@papertrailapp<\/a><\/td>\n<td align=\"left\">$7\/mo - $230\/mo<\/td>\n<td align=\"left\">Frustration-Free log management. Get started in seconds. Use Papertrail's time-saving log tools, flexible system groups, team-wide access, long-term archives, charts and analytics exports, monitoring webhooks, and 45-second setup.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/stackify.com\" rel=\"nofollow\">Stackify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Stackify\" rel=\"nofollow\">@Stackify<\/a><\/td>\n<td align=\"left\">$15\/mo<\/td>\n<td align=\"left\">Connecting the dots for you. Stackify focuses on application health, magnifying critical insights for developers, operations, and support teams.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/etsy\/statsd\/\">statsd<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sumologic.com\" rel=\"nofollow\">Sumo Logic<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SumoLogic\" rel=\"nofollow\">@SumoLogic<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Log Data is Big Data. Once enabled by the administrator, the new data will be searchable in the customer account. Sumo Logic  provides an Application for Data Volume for out-of\u2013the-box dashboards and searches that deliver a comprehensive view of data usage volume by category, collector, source name, and hosts.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-application-performance\" class=\"anchor\" aria-hidden=\"true\" href=\"#application-performance\"><\/a>Application Performance<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.appneta.com\" rel=\"nofollow\">AppNeta<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/AppNeta\" rel=\"nofollow\">@AppNeta<\/a><\/td>\n<td align=\"left\">Free to $119 \/ mo<\/td>\n<td align=\"left\">Full-stack application monitoring for web apps. Get visibility into code, network, and end user, especially for polyglot and service-oriented applications, by looking at transactions, errors, browser metrics, host metrics, and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dripstat.com\" rel=\"nofollow\">DripStat<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dripstat\" rel=\"nofollow\">@DripStat<\/a><\/td>\n<td align=\"left\">$20\/mo<\/td>\n<td align=\"left\">Application monitoring for Java. Next gen Java APM. Full visibility across your stack.  Specificially designed for Java. Supports SQL databases, MongoDB and Cassandra.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/newrelic.com\" rel=\"nofollow\">New Relic<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/NewRelic\" rel=\"nofollow\">@NewRelic<\/a><\/td>\n<td align=\"left\">$149\/mo<\/td>\n<td align=\"left\">Application monitoring for all your web apps. It\u2019s about gaining actionable, real-time business insights from the billions of metrics your software is producing, including user click streams, mobile activity, end user experiences and transactions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/appsignal.com\" rel=\"nofollow\">AppSignal<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/AppSignal\" rel=\"nofollow\">@AppSignal<\/a><\/td>\n<td align=\"left\">$49\/mo - $259\/mo<\/td>\n<td align=\"left\">Better monitoring for your Rails applications. Get detailled statistics on your site's performance with mean and 90th percentile measurements.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/instrumentalapp.com\" rel=\"nofollow\">Instrumental<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/instrumental\" rel=\"nofollow\">@instrumental<\/a><\/td>\n<td align=\"left\">$150\/mo - $750\/mo<\/td>\n<td align=\"left\">Monitor Your App in Realtime. Instrumental\u2019s made to monitor metrics at a ridiculously high scale. At rates of 500,000 metrics per second and higher, it doesn\u2019t even break a sweat<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atatus.com\" rel=\"nofollow\">Atatus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/atatusapp\" rel=\"nofollow\">@atatusapp<\/a><\/td>\n<td align=\"left\">$12\/mo - $159\/mo<\/td>\n<td align=\"left\">Simple JavaScript Error Tracking. Atatus is a simple error tracking and uptime monitoring system. Add two lines of code and get alerted on any errors that occurs in your application in realtime.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-load-testing\" class=\"anchor\" aria-hidden=\"true\" href=\"#load-testing\"><\/a>Load Testing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/github.com\/newsapps\/beeswithmachineguns\">Bees with Machine Guns!<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/flood.io\" rel=\"nofollow\">Flood.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/flood_io\" rel=\"nofollow\">@flood_io<\/a><\/td>\n<td align=\"left\">Free to $399\/mo<\/td>\n<td align=\"left\">Auto setup and results summaries\/graphs of JMeter and Gatling load tests.  Can easily scale to 100K+ reqs\/min.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.neustar.biz\/services\/web-performance\/load-testing\" rel=\"nofollow\">Neustar Website Load Testing<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Neustar\" rel=\"nofollow\">@Neustar<\/a><\/td>\n<td align=\"left\">$80\/mo<\/td>\n<td align=\"left\">Tackle performance problems such as bandwidth limitations, error rates exceeding thresholds, server PU limitations and much more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/loader.io\" rel=\"nofollow\">Loader.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free to 100.00$ \/ mo<\/td>\n<td align=\"left\">Loader.io is a Free load testing service that allows you to stress test your web-apps\/apis with thousands of concurrent connections.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/locust.io\" rel=\"nofollow\">Locust.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/locustio\" rel=\"nofollow\">@locustio<\/a><\/td>\n<td align=\"left\">Open Source<\/td>\n<td align=\"left\">Self-hosted load testing tool written in Python, with the tests themselves written in Python code<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/k6.io\" rel=\"nofollow\">k6.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/k6_io\" rel=\"nofollow\">@k6_io<\/a><\/td>\n<td align=\"left\">Free, Open Source<\/td>\n<td align=\"left\">k6 is a developer centric open source load testing tool for testing the performance of your backend infrastructure. It\u2019s built with Go and JavaScript to integrate well into your development workflow, so you can stay on top of performance without fuzz.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/loadimpact.com\/\" rel=\"nofollow\">loadimpact.com<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/loadimpact\" rel=\"nofollow\">@loadimpact<\/a><\/td>\n<td align=\"left\">$99.00+<\/td>\n<td align=\"left\">A SaaS performance testing solution for modern DevOps teams, built on top of the open source k6.io project.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-server-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#server-monitoring\"><\/a>Server Monitoring<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.serverdensity.com\" rel=\"nofollow\">Server Density<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/serverdensity\" rel=\"nofollow\">@serverdensity<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Premium hosted website and server monitoring tool. All your activity syncs in real time - from starting new instances to upgrading or deleting old ones. Work wherever you want - through web, mobile, API or directly with your provider. Everything stays in sync.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.datadoghq.com\" rel=\"nofollow\">Datadog<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/datadoghq\" rel=\"nofollow\">@datadoghq<\/a><\/td>\n<td align=\"left\">$0\/mo - $15\/host\/mo<\/td>\n<td align=\"left\">Datadog is a monitoring service for IT, Operations and Development teams who write and run applications at scale, and want to turn the massive amounts of data produced by their apps, tools and services into actionable insight.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.circonus.com\" rel=\"nofollow\">Circonus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/circonus\" rel=\"nofollow\">@circonus<\/a><\/td>\n<td align=\"left\">$15\/host\/mo - $25\/host\/mo<\/td>\n<td align=\"left\">Circonus combines multiple monitoring, alerting, event reporting, and analytical tools into one unified solution. Use any data, in any application, from any system, and visualize it in real-time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/truesightpulse.bmc.com\" rel=\"nofollow\">TrueSight Pulse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/truesightpulse\" rel=\"nofollow\">@truesightpulse<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Real-time visibility into cloud and server infrastructure<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.librato.com\" rel=\"nofollow\">Librato<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Librato\" rel=\"nofollow\">@Librato<\/a><\/td>\n<td align=\"left\">$0.05\/metric\/mo to $0.30\/metric\/mo<\/td>\n<td align=\"left\">Librato provides a complete solution for monitoring and understanding the metrics that impact your business at all levels of the stack.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/scoutapp.com\/\" rel=\"nofollow\">Scout<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/prometheus.io\" rel=\"nofollow\">Prometheus<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.site24x7.com\" rel=\"nofollow\">Site24x7<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Site24x7\" rel=\"nofollow\">@site24x7<\/a><\/td>\n<td align=\"left\">$9\/mo starter - $225\/enterprise<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/uptimerobot.com\" rel=\"nofollow\">Uptime Robot<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/uptimerobot\" rel=\"nofollow\">@uptimerobot<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free basic HTTP\/HTTPS website monitoring<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.binarycanary.com\" rel=\"nofollow\">BinaryCanary<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-customer-supporthelp-desks\" class=\"anchor\" aria-hidden=\"true\" href=\"#customer-supporthelp-desks\"><\/a>Customer Support\/Help Desks<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.desk.com\" rel=\"nofollow\">Desk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/desk\" rel=\"nofollow\">@desk<\/a><\/td>\n<td align=\"left\">$3\/mo - $50\/mo<\/td>\n<td align=\"left\">Deliver Customer Service That Wows. Desk.com creates the leading customer service application, which helps fast-growing companies deliver outstanding customer support. Desk.com's intuitive user interface and powerful features make solving customers' problems more efficient for the entire company. Plus, Desk.com is the only Customer Service Application backed by Salesforce, providing easy integration with other Salesforce services and robust security. There are thousands of companies using Desk.com as their help desk software application, from household names like Square and Instagram to the burrito shop down the street. Give it a try for Free.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.helpscout.net\" rel=\"nofollow\">HelpScout<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/helpscout\" rel=\"nofollow\">@helpscout<\/a><\/td>\n<td align=\"left\">$15\/mo<\/td>\n<td align=\"left\">Scalable customer support, no help desk headaches. Based on conditions you specify, Help Scout automatically performs one or more actions.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zendesk.com\" rel=\"nofollow\">ZenDesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zendesk\" rel=\"nofollow\">@zendesk<\/a><\/td>\n<td align=\"left\">$1\/mo - $195\/mo<\/td>\n<td align=\"left\">Relationships between businesses and their customers can be hard. Better customer service starts with better communication Zendesk brings all your customer conversations into one place.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.groovehq.com\" rel=\"nofollow\">Groove<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/groove\" rel=\"nofollow\">@groove<\/a><\/td>\n<td align=\"left\">$15\/ mo<\/td>\n<td align=\"left\">Everything you need to deliver awesome, personal support to every customer. Your customers will never know that you\u2019re using a helpdesk. To them, your messages look like regular emails. It feels like personal support, and it helps you build deeper relationships with your customers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.intercom.com\" rel=\"nofollow\">Intercom<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/intercom\" rel=\"nofollow\">@intercom<\/a><\/td>\n<td align=\"left\">$49\/mo - $449\/mo<\/td>\n<td align=\"left\">The easiest way to see and talk to your users. Intercom is a single platform where you can see in real-time who is using your product and send personalized messages to the right users at the right time based on their behavior.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/tenderapp.com\" rel=\"nofollow\">Tender<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tenderapp\" rel=\"nofollow\">@tenderapp<\/a><\/td>\n<td align=\"left\">$9\/mo - $99\/mo<\/td>\n<td align=\"left\">Better, Simpler, Customer Support Software. Support your customers in the open! With public forums, you can offer a public space to your users to discuss common issues and get feedback, while still keeping certain categories private (billing, orders, ...). Power users can subscribe to categories and new discussions, and help out other customers.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.enchant.com\" rel=\"nofollow\">Enchant<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/enchanthq\" rel=\"nofollow\">@enchanthq<\/a><\/td>\n<td align=\"left\">$9\/ mo<\/td>\n<td align=\"left\">It's like gmail on steroids! Enchant is a powerful helpdesk that helps your team deliver awesome support to each and every customer. To your customers, it's just email. They will never see a ticket number and will never have to log into anything!<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/freshdesk.com\" rel=\"nofollow\">Freshdesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/freshdesk\" rel=\"nofollow\">@freshdesk<\/a><\/td>\n<td align=\"left\">$16\/mo - $70\/mo<\/td>\n<td align=\"left\">Everything you need to deliver Exceptional Customer Support. Freshdesk keeps you from running behind issues blindly and gets your customer support issues under control.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/userdeck.com\" rel=\"nofollow\">UserDeck<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/user_deck\" rel=\"nofollow\">@user_deck<\/a><\/td>\n<td align=\"left\">$0 - $25\/mo<\/td>\n<td align=\"left\">Customer support software that embeds into your existing website.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sirportly.com\" rel=\"nofollow\">Sirportly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sirportly\" rel=\"nofollow\">@sirportly<\/a><\/td>\n<td align=\"left\">\u00a30 - \u00a315\/mo<\/td>\n<td align=\"left\">Grow your business and provide world class customer support. Simply setup your helpdesk in less than a few minutes. Integrate with your other software tools and take advantage of the automated rules and macros to scale your customer support to a new level, become more professional and customer focused and turn your customers into raving fans with Sirportly.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.olark.com\/\" rel=\"nofollow\">Olark<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/snapengage.com\/\" rel=\"nofollow\">SnapEngage<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Custom pricing per integration, or from $81\/month<\/td>\n<td align=\"left\">Highly customised live chat integrations<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getsatisfaction.com\/corp\/\" rel=\"nofollow\">Get Satisfaction!<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Customer communities for social support, social marketing &amp; customer feedback - online community software. Get Satisfaction is the leading customer engagement platform that helps companies build better relationships with their customers and prospects, through the best online customer community.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.reamaze.com\" rel=\"nofollow\">Reamaze<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/reamaze\" rel=\"nofollow\">@reamaze<\/a><\/td>\n<td align=\"left\">$15\/mo<\/td>\n<td align=\"left\">Lightweight, Lightspeed Help Desk. Email, Social, Branded, Integrated. Reamaze provides your team with helpdesk functionality that integrates with your application, as well as integrations with popular 3rd party tools in your workflow.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.jitbit.com\/helpdesk\/\" rel=\"nofollow\">Jitbit Helpdesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/jitbithelpdesk\" rel=\"nofollow\">@jitbithelpdesk<\/a><\/td>\n<td align=\"left\">$29\/mo - $199\/mo<\/td>\n<td align=\"left\">A help desk app that actually makes your work easier, not harder. Comes in both hosted and on-premise versions. Very well designed and easy to use. Has all the must-have features and doesn't get in your way.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.drift.com\" rel=\"nofollow\">Drift<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/drift\" rel=\"nofollow\">@drift<\/a><\/td>\n<td align=\"left\">Free for &lt; 100 contacts, paid from $49\/mo<\/td>\n<td align=\"left\">Stop wasting your website traffic. Sales-oriented live chat and in-app messaging, with chatbot automation.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zammad.com\" rel=\"nofollow\">Zammad<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zammadhq\" rel=\"nofollow\">@zammadhq<\/a><\/td>\n<td align=\"left\">$0\/mo - $24\/mo<\/td>\n<td align=\"left\">Zammad is a award winning open source helpdesk\/customer support system with a beautiful UI and many features to manage customer communication via several channels like telephone, facebook, twitter, chat and e-mails. It is distributed under the GNU AFFERO General Public License (AGPL).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.supporthero.io\" rel=\"nofollow\">Support Hero<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/supportheroapp\" rel=\"nofollow\">@supportheroapp<\/a><\/td>\n<td align=\"left\">$49\/mo - $199\/mo<\/td>\n<td align=\"left\">Knowledge base and tutorial management platform with smart analytics. Support Hero is a self-service knowledge base that helps companies educate their clients. Helping them to create and deliver a game-changing user education experience that increases retention and engagement, while reducing customer relationship costs.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-transactional-email\" class=\"anchor\" aria-hidden=\"true\" href=\"#transactional-email\"><\/a>Transactional Email<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/postmarkapp.com\" rel=\"nofollow\">Postmark<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/postmarkapp\" rel=\"nofollow\">@postmarkapp<\/a><\/td>\n<td align=\"left\">$1.50<\/td>\n<td align=\"left\">Email delivery for web apps \u2013 done right. Postmark removes the headaches of delivering and parsing transactional email for webapps with minimal setup time and zero maintenance. We have years of experience getting email to the inbox, so you can work and rest easier.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.mailgun.com\" rel=\"nofollow\">MailGun<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Mail_Gun\" rel=\"nofollow\">@Mail_Gun<\/a><\/td>\n<td align=\"left\">$20.00<\/td>\n<td align=\"left\">The Email Service For Developers. Easy SMTP integration and a simple, RESTful API abstracts away the messy details of email. Scale quickly, whether you need to send 10 or 10 million emails.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/ses\" rel=\"nofollow\">Amazon Simple Email Service<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/AWSSupport\" rel=\"nofollow\">@AWSSupport<\/a><\/td>\n<td align=\"left\">$0 for the first 62,000 emails you send each month<\/td>\n<td align=\"left\">Flexible, affordable, and highly-scalable email sending and receiving platform for businesses and developers<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sendgrid.com\" rel=\"nofollow\">SendGrid<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SendGrid\" rel=\"nofollow\">@SendGrid<\/a><\/td>\n<td align=\"left\">$9.95\/mo - $399.95<\/td>\n<td align=\"left\">Email Delivery. Simplified. SendGrid delivers billions of emails for companies of all sizes every month. Select the package that best fits with your sending volume, set-up your account, and let SendGrid take care of the rest!<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.critsend.com\" rel=\"nofollow\">CritSend<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/critsend\" rel=\"nofollow\">@critsend<\/a><\/td>\n<td align=\"left\">$50\/mo - $3000\/mo<\/td>\n<td align=\"left\">The Best SMTP Relay for Developers. Use the most reliable infrastructure for your transactional and bulk emails. It only takes 5 minutes to setup Critsend and start enjoying fast delivery time and automatic scalability.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/postageapp.com\" rel=\"nofollow\">Postage<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/postagebird\" rel=\"nofollow\">@postagebird<\/a><\/td>\n<td align=\"left\">$9\/mo - $399\/mo<\/td>\n<td align=\"left\">The easier way to send email from web apps. Postage app helps design, send, and analyze emails within minutes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sendwithus.com\" rel=\"nofollow\">Sendwithus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/send_with_us\" rel=\"nofollow\">@sendwithus<\/a><\/td>\n<td align=\"left\">Free 'Hacker' plan of 1000 messages\/mo<\/td>\n<td align=\"left\">Transactional email A\/B testing and drip campaigns<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sparkpost.com\" rel=\"nofollow\">SparkPost<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/SparkPost\" rel=\"nofollow\">@SparkPost<\/a><\/td>\n<td align=\"left\">Free - $474\/mo<\/td>\n<td align=\"left\">The most performant email delivery service \u2014 whether your app or website sends hundreds, or billions, of messages.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-other-apis\" class=\"anchor\" aria-hidden=\"true\" href=\"#other-apis\"><\/a>Other APIs<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.filestack.com\" rel=\"nofollow\">Filestack<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/FileStack\" rel=\"nofollow\">@FileStack<\/a><\/td>\n<td align=\"left\">$0\/mo - $49\/mo<\/td>\n<td align=\"left\">Easy, Powerful File Uploads. Connect, Store, and Process any file from anywhere on the Internet<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/openexchangerates.org\" rel=\"nofollow\">Open Exchange Rates<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$12\/mo - $97\/mo<\/td>\n<td align=\"left\">Real-time exchange rates &amp; currency conversion JSON API. A simple and easy-to-integrate exchange rates API in JSON format, with HTTPS and JSONP support, with examples, guides and full documentation.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/formapi.io\" rel=\"nofollow\">FormAPI<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/form_api\" rel=\"nofollow\">@form_api<\/a><\/td>\n<td align=\"left\">$49\/mo - $249\/mo<\/td>\n<td align=\"left\">PDF Service for Developers. Makes it easy for programmers to fill out and sign PDF documents. Saves a lot of development time if you need to generate a lot of contracts, invoices, etc.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-site-search\" class=\"anchor\" aria-hidden=\"true\" href=\"#site-search\"><\/a>Site Search<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.elastic.co\/products\/elasticsearch\" rel=\"nofollow\">Elasticsearch<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/elasticsearch\" rel=\"nofollow\">@elasticsearch<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">an end-to-end search and analytics platform. infinitely versatile. By combining the massively popular Elasticsearch, Logstash and Kibana we have created an end-to-end stack that delivers actionable insights in real-time from almost any type of structured and unstructured data source. Built and supported by the engineers behind each of these open source products, the Elasticsearch ELK stack makes searching and analyzing data easier than ever before.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/swiftype.com\" rel=\"nofollow\">Swiftype<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/swiftype\" rel=\"nofollow\">@Swiftype<\/a><\/td>\n<td align=\"left\">Free - $250\/mo, Enterprise Plans available<\/td>\n<td align=\"left\">Swiftype is a hosted website search service, available as a web crawler or as an API integration. API clients are available for major frameworks and languages, plugins are available for major third party platforms.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.algolia.com\" rel=\"nofollow\">Algolia<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Algolia\" rel=\"nofollow\">@Algolia<\/a><\/td>\n<td align=\"left\">$49\/mo - $449\/mo<\/td>\n<td align=\"left\">Build Realtime Search. Algolia is a fully hosted search service, available as a REST API. API clients are also available for all major frameworks, platforms and languages.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/lucene.apache.org\/solr\/\" rel=\"nofollow\">Apache Solr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">An open source search platform, based and co-released with Apache Lucene.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/cloudsearch\/\" rel=\"nofollow\">Amazon Cloudsearch<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Search SaaS<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-email-marketing\" class=\"anchor\" aria-hidden=\"true\" href=\"#email-marketing\"><\/a>Email Marketing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.mailcharts.com\" rel=\"nofollow\">MailCharts<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mailcharts\" rel=\"nofollow\">@mailcharts<\/a><\/td>\n<td align=\"left\">$30\/mo<\/td>\n<td align=\"left\">Track &amp; understand how your competitors use email marketing. MailCharts tracks over a thousand companies, bringing you actionable insights to help you improve your email marketing strategy, make data-driven decisions and gain design and content inspiration.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/customer.io\" rel=\"nofollow\">Customer.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$50\/mo - $1250\/Mo<\/td>\n<td align=\"left\">Send email based on what people do or don't do in your app. Customer.io lets you send newsletters to segments of customers using data from your site.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getvero.com\" rel=\"nofollow\">Vero<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/veroapp\" rel=\"nofollow\">@veroapp<\/a><\/td>\n<td align=\"left\">$99\/mo<\/td>\n<td align=\"left\">Send emails based on what your customers do. Vero makes it easy to create segments of customers based on the attributes you capture (e.g. age, location, gender) and the actions your customers take (e.g. logged in, used feature x, checked out, etc.).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mailchimp.com\" rel=\"nofollow\">Mailchimp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Mailchimp\" rel=\"nofollow\">@Mailchimp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.campaignmonitor.com\" rel=\"nofollow\">Campaign Monitor<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/CampaignMonitor\" rel=\"nofollow\">@CampaignMonitor<\/a><\/td>\n<td align=\"left\">$9\/mo - $699\/mo<\/td>\n<td align=\"left\">CAMPUnbounce Feature Tour<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sendy.co\" rel=\"nofollow\">Sendy<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getSendy\" rel=\"nofollow\">@getsendy<\/a><\/td>\n<td align=\"left\">$59 one time fee, host yourselfor use a hosted sendy instance from a variety of providers.<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/image-charts.com\" rel=\"nofollow\">Image-Charts<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/imagecharts\" rel=\"nofollow\">@imagecharts<\/a><\/td>\n<td align=\"left\">Free - $49\/mo, self-hosted plan available<\/td>\n<td align=\"left\">Include animated charts as image into emails, no server-side rendering, 1 URL = 1 chart, compatible with Google Image Charts.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.drip.com\" rel=\"nofollow\">Drip<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getdrip\" rel=\"nofollow\">@getdrip<\/a><\/td>\n<td align=\"left\">Free for &lt; 100 subscribers, paid plans from $49\/mo<\/td>\n<td align=\"left\">Advanced marketing automation with workflows, drip campaigns, conversion tracking etc..<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.mailerlite.com\" rel=\"nofollow\">MailerLite<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mailerlite\" rel=\"nofollow\">@mailerlite<\/a><\/td>\n<td align=\"left\">Free - $140\/mo<\/td>\n<td align=\"left\">Free plan offers a lot of features<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/markethero.io\/\" rel=\"nofollow\">MarketHero<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$99+\/mo<\/td>\n<td align=\"left\">Autoresponder \/ email marketing tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/emailoctopus.com\" rel=\"nofollow\">EmailOctopus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/emailoctopus\" rel=\"nofollow\">@emailoctopus<\/a><\/td>\n<td align=\"left\">Free up to 2,500 subscribers, paid plans from $19\/mo<\/td>\n<td align=\"left\">Low cost email marketing using Amazon SES.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-email-collectionlanding-page-apps\" class=\"anchor\" aria-hidden=\"true\" href=\"#email-collectionlanding-page-apps\"><\/a>Email Collection\/Landing Page Apps<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.launchrock.com\" rel=\"nofollow\">Launchrock<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/launchrock\" rel=\"nofollow\">@launchrock<\/a><\/td>\n<td align=\"left\">$49\/mo - $199\/mo<\/td>\n<td align=\"left\">Even if you know how to code a web page with HTML, you'll love how much faster it is with our landing page builder. Load up your logo and graphic assets, set up a few base colors from your brand palette and publish away.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/unbounce.com\" rel=\"nofollow\">Unbounce<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/unbounce\" rel=\"nofollow\">@unbounce<\/a><\/td>\n<td align=\"left\">$49\/mo - $199\/mo<\/td>\n<td align=\"left\">Unbounce empowers marketers to act independently from technical teams, improving their efficiency and their ability to generate sales. Produce high-converting landing pages without dealing with I.T. bottlenecks. See how Unbounce can enhance your campaigns and maximize your marketing spend.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.leadpages.net\/\" rel=\"nofollow\">LeadPages<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$25\/mo - $199\/mo<\/td>\n<td align=\"left\">Generate leads and increase revenue using the industry-leading landing page builder with accompanying suite of lead generation and opt-in tools.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/instapage.com\/\" rel=\"nofollow\">Instapge<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$29\/mo - $127\/mo<\/td>\n<td align=\"left\">A landing page solution for optimizing your ad spend. Build, Integrate, Collaborate &amp; Optimize.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/kickofflabs.com\" rel=\"nofollow\">KickoffLabs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kickofflabs\" rel=\"nofollow\">@kickofflabs<\/a><\/td>\n<td align=\"left\">$29\/mo - $99\/mo<\/td>\n<td align=\"left\">Stop building landing pages. Start building smarter campaigns. Send customers to a tailored page that speaks to them. Keep them engaged with signup-forms, newsletters and \u2013 best of all \u2013 a very cool, unique REWARD system for customer referrals.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.prefinery.com\" rel=\"nofollow\">Prefinery<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/prefinery\" rel=\"nofollow\">@prefinery<\/a><\/td>\n<td align=\"left\">$19\/mo-  $399\/mo<\/td>\n<td align=\"left\">Stress-Free Beta Invitation and Management. Prefinery is beta invitations &amp; management for serious product launches. You need more than just a landing page -- outsource your beta to Prefinery!<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-crmsales-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#crmsales-tools\"><\/a>CRM\/Sales Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kloudless.com\" rel=\"nofollow\">Kloudless<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kloudless\" rel=\"nofollow\">@kloudless<\/a><\/td>\n<td align=\"left\">Free - $500\/mo<\/td>\n<td align=\"left\">Kloudless allows you to integrate with several CRM solutions via a single &amp; comprehensive RESTful API. The Kloudless CRM Unified API allows you to integrate with Salesforce, SugarCRM, Microsoft Dynamics, Zoho, Oracle Sales Cloud, HubSpot, Pipeliner, and more with a single code-base. Kloudless also maintains all the integrations so you can focus on building awesome solutions faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.salesforce.com\" rel=\"nofollow\">Salesforce<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/salesforce\" rel=\"nofollow\">@salesforce<\/a><\/td>\n<td align=\"left\">TRANSFORM CONSUMERS INTO CUSTOMERS ACROSS ALL DIGITAL CHANNELS. The ExactTarget Marketing Cloud, built on the Salesforce1 Platform, allows marketers to create 1:1 campaigns like never before. So you can combine traditional digital channels like email, mobile, social, and the web with any conceivable IP-addressable product to turn consumers into customers.<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.salesforceiq.com\" rel=\"nofollow\">SalesforceIQ<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/salesforceiq\" rel=\"nofollow\">@salesforceiq<\/a><\/td>\n<td align=\"left\">\u00a317-85\/user\/mo<\/td>\n<td align=\"left\">Helps businesses build stronger relationships using Relationship Intelligence.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sugarcrm.com\" rel=\"nofollow\">SugarCRM<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sugarcrm\" rel=\"nofollow\">@sugarcrm<\/a><\/td>\n<td align=\"left\">$35\/mo - $150\/mo<\/td>\n<td align=\"left\">Sugar's open, flexible platform easily solves real business problems. From automating sales, marketing and customer support to creating a custom CRM application, we've got you covered. By placing the individual at the center of its solution, SugarCRM puts the \"i\" in CRM and empowers every customer-facing individual to create extraordinary customer relationships.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.insightly.com\" rel=\"nofollow\">Insight.ly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/insightly\" rel=\"nofollow\">@insightly<\/a><\/td>\n<td align=\"left\">$7\/mo<\/td>\n<td align=\"left\">Small Business CRM. Manage contacts, organizations, partners, vendors and suppliers. See everything from background, email history, events, projects or opportunities<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/close.io\" rel=\"nofollow\">Close.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/closeio\" rel=\"nofollow\">@closeio<\/a><\/td>\n<td align=\"left\">$59\/mo - $299\/mo<\/td>\n<td align=\"left\">Sales Communication Platform. Close.io automatically logs sent and received emails with your leads. You can send\/receive emails painlessly from Close.io. Or, by entering your IMAP and SMTP mail settings, Close.io can track emails that you send from Gmail or any email client.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.streak.com\" rel=\"nofollow\">Streak<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/streak\" rel=\"nofollow\">@streak<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Manage email support inside Gmail. Streak lets you keep track of all your deals right from your inbox. We let you group emails from the same customer together into one view and push that customer through your pipeline. When a new email comes in, you'll have all the context you need.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getbase.com\" rel=\"nofollow\">Base<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getbase\" rel=\"nofollow\">@getbase<\/a><\/td>\n<td align=\"left\">$15\/mo - $125\/mo<\/td>\n<td align=\"left\">The Sales and CRM Software Your Team Will Actually Use. Base is designed to significantly boost your team\u2019s sales productivity and give you the core sales tools you need to grow your business. Your leads come from a variety of sources. Lead management in Base helps you organize leads and assign them to the right sales reps so they can be followed up on and qualified quickly. After being qualified, convert a lead in Base and all of the contact information you have about your lead is transferred your new customer contact card. Optionally, you can specify follow up tasks and even create a deal at the same time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pipedrive.com\" rel=\"nofollow\">Pipedrive<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$9\/mo<\/td>\n<td align=\"left\">Pipeline software that gets you organized. Pipedrive is built for salespeople who need to put in serious effort to turn leads into sales. It helps to organize the work and spend less time on admin.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.contactually.com\" rel=\"nofollow\">Contactually<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/contactually\" rel=\"nofollow\">@Contactually<\/a><\/td>\n<td align=\"left\">$17.99\/mo - $99.99\/mo<\/td>\n<td align=\"left\">Maximize your network ROI. More referrals. More repeat business. Contactually helps businesses follow up with the right people, at the right time, to maximize relationship ROI.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-social-media-marketing\" class=\"anchor\" aria-hidden=\"true\" href=\"#social-media-marketing\"><\/a>Social Media Marketing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/buffer.com\" rel=\"nofollow\">Buffer<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/buffer\" rel=\"nofollow\">@buffer<\/a><\/td>\n<td align=\"left\">$10\/mo<\/td>\n<td align=\"left\">Buffer is the easiest way to publish on social media. Buffer helps you share to Twitter, Facebook and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hootsuite.com\" rel=\"nofollow\">HootSuite<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/HootSuite\" rel=\"nofollow\">@HootSuite<\/a><\/td>\n<td align=\"left\">$8.99\/mo<\/td>\n<td align=\"left\">Add speed and agility to your social media strategy. From one dashboard you\u2019ll schedule Tweets and Facebook posts, monitor conversations, and more. When you need to prove your social ROI, quickly create and customize ready-to-present analytics reports.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/claim.io\/welcome\" rel=\"nofollow\">Claim.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/claimio\" rel=\"nofollow\">@claimio<\/a><\/td>\n<td align=\"left\">$189\/mo - $499\/mo<\/td>\n<td align=\"left\">Owning your name on 300 Social media sites not only makes it easier for people to find you or your business online - it also works as a \"social media identity insurance\", protecting you from name squatting and identity fraud, minimizing risk to your brand.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.salesforce.com\/products\/marketing-cloud\/overview\/\" rel=\"nofollow\">Exacttarget Marketing Cloud\/Buddy Media<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/marketingcloud\" rel=\"nofollow\">@marketingcloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Bringing you closer to your social customers with an integrated social content solution. Run integrated campaigns across Facebook, Twitter, YouTube and your websites Increase fans, followers and advocates. Publish engaging and interactive social apps. Easily create landing pages and microsites and extend social to your websites. Understand engagement trends, demographics, conversions and business metrics with powerful analytics.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sproutsocial.com\" rel=\"nofollow\">Sprout Social<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sproutsocial\" rel=\"nofollow\">@sproutsocial<\/a><\/td>\n<td align=\"left\">$59\/mo - $1500\/mo<\/td>\n<td align=\"left\">Powerful Social Media Software. Sprout lets you post messages on Facebook, Twitter, Google+ and LinkedIn simultaneously from one easy-to-use message composition tool. Shorten links, attach photos, target your audience on Facebook, customize your posts and much more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/f5bot.com\" rel=\"nofollow\">F5Bot<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">F5Bot is Free social media keyword monitoring. F5Bot will send you an email when your startup, your product, or your competitors are mentioned on Hacker News or Reddit.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-naming\" class=\"anchor\" aria-hidden=\"true\" href=\"#naming\"><\/a>Naming<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.trademarkia.com\" rel=\"nofollow\">Trademarkia<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/trademarkia\" rel=\"nofollow\">@trademarkia<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Trademarkia is one of the largest trademark search engines in the world. LegalForce's network of licensed patent attorneys and agents have filed hundreds of patent applications for companies of every size.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.namerobot.com\" rel=\"nofollow\">NameRobot<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/namerobotEN\" rel=\"nofollow\">@namerobotEN<\/a>]<\/td>\n<td align=\"left\">0$ - 300$\/mo<\/td>\n<td align=\"left\">Find, create and check the name for your project. NameRobot offers everything you need to create suitable naming ideas in a short time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/whois.domaintools.com\" rel=\"nofollow\">DomainTools Whois Lookup<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/DomainTools\" rel=\"nofollow\">@DomainTools<\/a><\/td>\n<td align=\"left\">Free - $99\/mo<\/td>\n<td align=\"left\">Go beyond ordinary Whois to discover the people or organizations behind a domain name or IP address.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-space-rental\" class=\"anchor\" aria-hidden=\"true\" href=\"#space-rental\"><\/a>Space Rental<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/42floors.com\" rel=\"nofollow\">42Floors<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/42floors\" rel=\"nofollow\">@42floors<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The Best Place to Find Office &amp; Commercial Space Rentals. We're gathering listing data from everyone in the market. Even including off-market listings that landlords haven't yet posted on 42Floors.com or anywhere else.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/liquidspace.com\" rel=\"nofollow\">Liquidspace<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/LiquidSpace\" rel=\"nofollow\">@LiquidSpace<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Optimize Real Estate for Your Enterprise. Whether planning ahead or booking on the fly, find and reserve a great place to work. Rent professional conference and meeting rooms, private offices, or coworking spaces daily or hourly.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pivotdesk.com\" rel=\"nofollow\">PivotDesk<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PivotDesk\" rel=\"nofollow\">@PivotDesk<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">An office sharing marketplace that helps startups that need space find host companies that have excess space.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-community-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#community-tools\"><\/a>Community Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.discourse.org\" rel=\"nofollow\">Discourse<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/discourse\" rel=\"nofollow\">@discourse<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Ready for a new discussion platform? Discussion software is a group of people interested in a common topic who are willing to type paragraphs to each other on a web page.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-personal-productivity\" class=\"anchor\" aria-hidden=\"true\" href=\"#personal-productivity\"><\/a>Personal Productivity<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/www.tomato.es\/\" rel=\"nofollow\">Tomatoes<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/tomatoesapp\" rel=\"nofollow\">@tomatoesapp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Track your time and be productive with the Pomodoro Technique. Tomatoes is a \"pomodoro tracker\", a Pomodoro technique\u00ae driven time tracker. Track your time using 25 minutes slots called \"pomodoros\".<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.rescuetime.com\" rel=\"nofollow\">RescueTime<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/rescuetime\" rel=\"nofollow\">@rescuetime<\/a><\/td>\n<td align=\"left\">Free - $9\/mo<\/td>\n<td align=\"left\">RescueTime gives you an accurate picture of how you spend your time to help you become more productive every day.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/qotoqot.com\/qbserve\/\" rel=\"nofollow\">Qbserve<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Qbserve\" rel=\"nofollow\">@Qbserve<\/a><\/td>\n<td align=\"left\">$40 one-time<\/td>\n<td align=\"left\">Qbserve is a time tracker for Mac that does whatever RescueTime can but also tracks project time automatically, generates invoices, and stores tracked data locally.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/timingapp.com\/\" rel=\"nofollow\">Timing<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/TimingApp\" rel=\"nofollow\">@TimingApp<\/a><\/td>\n<td align=\"left\">$29 - $79<\/td>\n<td align=\"left\">Automatic time and productivity tracking for Mac. Helps you stay on track with your work and ensures no billable hours get lost (if you are billing hourly).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/fman.io\" rel=\"nofollow\">fman<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/m_herrmann\" rel=\"nofollow\">@m_herrmann<\/a><\/td>\n<td align=\"left\">$14.00<\/td>\n<td align=\"left\">Manage and transfer your files with ease. For Windows, Mac and Linux.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/wakatime.com\" rel=\"nofollow\">WakaTime<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/WakaTime\" rel=\"nofollow\">@WakaTime<\/a><\/td>\n<td align=\"left\">Free - $9\/mo<\/td>\n<td align=\"left\">Automatic productivity metrics using text editor plugins. Goals, leaderboards, GitHub integration, automatic project\/language\/branch detection.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-prototypingmockups\" class=\"anchor\" aria-hidden=\"true\" href=\"#prototypingmockups\"><\/a>Prototyping\/Mockups<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/creately.com\" rel=\"nofollow\">Creately<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/creately\" rel=\"nofollow\">@creately<\/a><\/td>\n<td align=\"left\">Free - $750\/mo<\/td>\n<td align=\"left\">Web based diagramming tool for fast easy diagrams. Supports flowcharts, mock-ups, wire-frames, mind maps, organizational charts, network diagrams, AWS diagrams, UML diagrams and many other diagram types. Real-time collaboration plus innovative productivity features to create diagrams 3 times faster.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.apple.com\/keynote\/\" rel=\"nofollow\">Keynote<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$19.99<\/td>\n<td align=\"left\">Easily create gorgeous presentations with the all-new Keynote, featuring powerful yet easy-to-use tools and dazzling effects that will make you a very hard act to follow. Also checkout the <a href=\"https:\/\/keynotopia.com\" rel=\"nofollow\">Keynotopia Themes<\/a> to get all the common UI elements for iOS, Android etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.omnigroup.com\/omniGraffle\" rel=\"nofollow\">OmniGraffle<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/omniGraffle\" rel=\"nofollow\">@omniGraffle<\/a><\/td>\n<td align=\"left\">$99.99<\/td>\n<td align=\"left\">Diagramming Worth a Thousand Words. Need a diagram, process chart, quick page-layout, website wireframe or graphic design? OmniGraffle can help you make eye-popping graphic documents quickly by keeping lines connected to shapes even when they're moved, providing powerful styling tools, and magically organizing diagrams with just one click. Whether you need a quick<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/moqups.com\" rel=\"nofollow\">moqups<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/moqups\" rel=\"nofollow\">@moqups<\/a><\/td>\n<td align=\"left\">$9\/mo -  $39\/mo<\/td>\n<td align=\"left\">Moqups is a nifty HTML5 App used to create wireframes, mockups or UI concepts, prototypes depending on how you like to call them. The most stunning HTML5 app for creating resolution-independent SVG mockups &amp; wireframes for your next project.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/balsamiq.com\" rel=\"nofollow\">Balsamiq<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/balsamiq\" rel=\"nofollow\">@balsamiq<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Life's too short for bad software! Balsamiq Mockups is a rapid wireframing tool that helps you Work Faster &amp; Smarter. It reproduces the experience of sketching on a whiteboard, but using a computer.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/proto.io\" rel=\"nofollow\">Proto.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/protoio\" rel=\"nofollow\">@protoio<\/a><\/td>\n<td align=\"left\">$24\/mo - $199\/mo<\/td>\n<td align=\"left\">Silly-fast mobile prototyping. Build high-fidelity fully interactive mobile app prototypes in minutes. Prototypes can be viewed on browser or device giving a real experience to the user how the app will look like and behave. Multiple devices like smart phones and tablets\/pads are supported including iPhone, iPad and Android devices.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.invisionapp.com\" rel=\"nofollow\">invision<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/InVisionApp\" rel=\"nofollow\">@InVisionApp<\/a><\/td>\n<td align=\"left\">$0\/mo - $100+\/mo<\/td>\n<td align=\"left\">Free Web &amp; Mobile (iOS, Android) Prototyping and UI Mockup Tool. Transform your Web &amp; Mobile (iOS, Android) designs into clickable, interactive Prototypes and Mockups. Share and Collaborate on them with others.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.sketchapp.com\" rel=\"nofollow\">Sketch<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/sketchapp\" rel=\"nofollow\">@sketchapp<\/a><\/td>\n<td align=\"left\">$99.00<\/td>\n<td align=\"left\">Professional digital design for Mac. Sketch gives you the power, flexibility and speed you always wanted in a lightweight and easy-to-use package. Finally you can focus on what you do best: Design.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.figma.com\" rel=\"nofollow\">Figma<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/figmadesign\" rel=\"nofollow\">@figmadesign<\/a><\/td>\n<td align=\"left\">Free - $12\/editor\/mo<\/td>\n<td align=\"left\">Design and prototyping tool, with a focus on collaboration and teams.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.animaapp.com\" rel=\"nofollow\">Anima<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/animaapp\" rel=\"nofollow\">@animaapp<\/a><\/td>\n<td align=\"left\">$0 - $39\/mo<\/td>\n<td align=\"left\">Sketch to Code Toolkit. Anima converts web design to code. Make your Sketch design responsive, interactive and alive, and export to HTML\/JS\/CSS.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/framer.com\" rel=\"nofollow\">Framer<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/framer\" rel=\"nofollow\">@framer<\/a><\/td>\n<td align=\"left\">Free Trial - $12\/mo<\/td>\n<td align=\"left\">Pixel-perfect, responsive designs. Truly high-fidelity prototypes. Seamless collaboration. Framer powers the product teams at Dropbox, Pinterest, Twitter, and thousands more.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-content-creationinfographics\" class=\"anchor\" aria-hidden=\"true\" href=\"#content-creationinfographics\"><\/a>Content Creation\/Infographics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/visual.ly\" rel=\"nofollow\">Visual.ly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Visually\" rel=\"nofollow\">@Visually<\/a><\/td>\n<td align=\"left\">$195\/mo - $994\/mo<\/td>\n<td align=\"left\">ORIGINAL VISUAL CONTENT FOR BRANDS. We only work with the best creative talent available. Thousands of designers, journalists, animators and developers are standing by to help you achieve your goals and take your project to the next level.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.canva.com\" rel=\"nofollow\">Canva<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/canva\" rel=\"nofollow\">@canva<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Design great social media images with text and graphics for Free or a few bucks depending on the images you select<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-customer-feedback\" class=\"anchor\" aria-hidden=\"true\" href=\"#customer-feedback\"><\/a>Customer Feedback<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.pickfu.com\" rel=\"nofollow\">PickFu<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/pickfu\" rel=\"nofollow\">@pickfu<\/a><\/td>\n<td align=\"left\">$20\/mo - $299\/mo<\/td>\n<td align=\"left\">REAL CONSUMER FEEDBACK IN MINUTES. PickFu is a tool that provides instant, unbiased and insightful public opinion on questions that you care about.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.promoter.io\" rel=\"nofollow\">Promoter.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/promoter_io\" rel=\"nofollow\">@Promoter_io<\/a><\/td>\n<td align=\"left\">$50\/mo - $500\/mo<\/td>\n<td align=\"left\">Helps companies quickly gain predictive customer intelligence &amp; insights driven by NPS (Net Promoter) to increase revenue and reduce churn.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-data\" class=\"anchor\" aria-hidden=\"true\" href=\"#data\"><\/a>Data<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.factual.com\" rel=\"nofollow\">Factual<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Factual\" rel=\"nofollow\">@Factual<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">GLOBAL DATA. LOCAL CONTEXT. Factual\u2019s location platform enriches mobile location signals with definitive global data, enabling personalized and contextually relevant mobile experiences. Built from billions of inputs, the data is constantly updated by Factual\u2019s real-time data stack.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-database\" class=\"anchor\" aria-hidden=\"true\" href=\"#database\"><\/a>Database<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.citusdata.com\" rel=\"nofollow\">Citus<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Citus scales out PostgreSQL through sharding, replication, and query parallelization. So you can focus on your application, not worry about your database. Citus is available as open source, enterprise software, and fully managed database as a service.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/bulbflow.org\" rel=\"nofollow\">Bulbs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Python persistence framework for graph databases.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.datomic.com\" rel=\"nofollow\">Datomic<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/datomic_team\" rel=\"nofollow\">@datomic_team<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The fully transactional, cloud-ready, immutable database. Immutable data means strong consistency combined with horizontal read scalability, plus built-in caching. Datomic is a distributed database designed to enable scalable, flexible and intelligent applications, running on next-generation cloud architectures.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/tinkerpop.apache.org\" rel=\"nofollow\">Tinkerpop<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Open source software products in the graph space.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.vertabelo.com\" rel=\"nofollow\">Vertabelo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Vertabelo\" rel=\"nofollow\">@vertabelo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Web-based tool for database design. Vertabelo allows you to visually design database models for PostgreSQL, MySQL, Oracle, SQL Server, SQLite, and IBM DB2. You can import the existing database structure from SQL, XML, or using reverse engineering tool. After you design a model, you can generate SQL script or ready-to-use code for various ORMs (Propel, jOOQ, SQLAlchemy, or <a href=\"http:\/\/mobile-orm.vertabelo.com\/\" rel=\"nofollow\">Vertabelo Mobile ORM<\/a>).<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/tableplus.io\" rel=\"nofollow\">TablePlus<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/TablePlus\" rel=\"nofollow\">@TablePlus<\/a><\/td>\n<td align=\"left\">$0 - $49<\/td>\n<td align=\"left\">Modern, native client with intuitive GUI tools to create, access, query &amp; edit multiple databases: MySQL, PostgreSQL, SQLite, Microsoft SQL Server, Amazon Redshift, MariaDB, CockroachDB, Vertica, Cassandra, Oracle, and Redis.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dbngin.com\" rel=\"nofollow\">DBngin<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dbngin\" rel=\"nofollow\">@DBngin<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">An all-in-one tool for database version management. It allows you to set up a local database server of any versions within a click.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-accountinginvoicing\" class=\"anchor\" aria-hidden=\"true\" href=\"#accountinginvoicing\"><\/a>Accounting\/Invoicing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.getharvest.com\" rel=\"nofollow\">Harvest<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/harvest\" rel=\"nofollow\">@harvest<\/a><\/td>\n<td align=\"left\">$12\/mo - $99\/mo<\/td>\n<td align=\"left\">Time Tracking Made Easy. Time tracking is simple and lightning fast with Harvest. Set up takes seconds, and there\u2019s nothing to install. We\u2019ve simplified the timesheet and timesheets approval process so you can stay focused on work.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getballpark.com\" rel=\"nofollow\">Ballpark<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$12\/mo - $99\/mo<\/td>\n<td align=\"left\">Stop sending your clients ugly paper invoices. Go paperless today, with Ballpark. Our beautiful, web-based invoices and estimates make it easier than ever to get paid and discuss projects with your clients and colleagues.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paysimple.com\" rel=\"nofollow\">PaySimple<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PaySimple\" rel=\"nofollow\">@PaySimple<\/a><\/td>\n<td align=\"left\">$34.95\/mo<\/td>\n<td align=\"left\">Simplify how you bill and collect. PaySimple is an industry-leading provider of payment management solutions. PaySimple simplifies billing and collection processes by enabling you to bill, collect and deposit all of your payments automatically. Our customized, secure ASP solution includes auto-recurring billing, electronic check processing, direct debit and credit card processing at some of the lowest rates available.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.freshbooks.com\" rel=\"nofollow\">FreshBooks<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/freshbooks\" rel=\"nofollow\">@freshbooks<\/a><\/td>\n<td align=\"left\">$19.95\/mo - $39.95\/mo<\/td>\n<td align=\"left\">Accounting Made for You, the Non-Accountant. FreshBooks is simple and intuitive, so accounting isn't intimidating. Plus you can talk to a real, live person anytime you have a question, 9am-6pm EDT, Mon-Fri.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.Freeagent.com\" rel=\"nofollow\">FreeAgent<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Freeagent\" rel=\"nofollow\">@Freeagent<\/a><\/td>\n<td align=\"left\">US $20\/mo<\/td>\n<td align=\"left\">Accounting software trusted by over 35,000 Freelancers and small businesses. Reconcile money in and out of the business via your electronic bank statements and build monthly balance charts.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.blinksale.com\" rel=\"nofollow\">Blinksale<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/blinksale\" rel=\"nofollow\">@blinksale<\/a><\/td>\n<td align=\"left\">$15.00<\/td>\n<td align=\"left\">INTRODUCING BLINKSALE UNLIMITED. ONE PLAN. ONE PRICE. UNLIMITED EVERYTHING. Blinksale makes you look your best. With over a dozen invoice designs &amp; thank you notes, you\u2019re sure to put your best foot forward. Every time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/cashboardapp.com\" rel=\"nofollow\">Cashboard<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/cashboard\" rel=\"nofollow\">@cashboard<\/a><\/td>\n<td align=\"left\">$8.25\/mo - $250\/mo<\/td>\n<td align=\"left\">FreeLANCE TIME TRACKING &amp; INVOICE SOFTWARE TRUSTED BY THOUSANDS, WORLDWIDE. Cashboard is the tool we designed to remedy that situation. It works for our software consultancy and we think it\u2019ll work for you too. We launched Cashboard in Spring of 2007after many sleepless nights of hard work. It was the first solution to combine estimates, invoices, time tracking, and online payments into one tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paydirtapp.com\" rel=\"nofollow\">Paydirt<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/paydirtapp\" rel=\"nofollow\">@paydirtapp<\/a><\/td>\n<td align=\"left\">$8\/mo - $149\/mo<\/td>\n<td align=\"left\">Smart Time Tracking Easier Invoicing Online Payments. In Paydirt, you can start a timer from any page in one click. No fiddly menus. No navigating around. Just a start button for each task.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.indinero.com\" rel=\"nofollow\">inDinero<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/qbo.intuit.com\/qbo27\/login?webredir\" rel=\"nofollow\">QuickBooks Online<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.xero.com\" rel=\"nofollow\">Xero<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/Fast409A.io\" rel=\"nofollow\">Fast409A<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">$2000 (depending on startup funding stage)<\/td>\n<td align=\"left\">The Fast409A.io process blends the best of software and human expertise. Each 409A valuation takes an hour or less of your time to submit, is delivered in days not weeks, and is customised to the nuances of your individual startup.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/startuprunway.io\" rel=\"nofollow\">Runway<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Most startups die from running out of cash. Runway is an easy, visual way to plan and manage your startup's most precious resource, explore scenarios and track planned versus actual spending.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-privacy-policy-terms--conditions-legal-documents\" class=\"anchor\" aria-hidden=\"true\" href=\"#privacy-policy-terms--conditions-legal-documents\"><\/a>Privacy Policy, Terms &amp; Conditions, Legal Documents<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/www.iubenda.com\" rel=\"nofollow\">iubenda<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/iubenda\" rel=\"nofollow\">@iubenda<\/a><\/td>\n<td align=\"left\">Free - $27\/year - customization services<\/td>\n<td align=\"left\">The easiest way to generate a professional, customizable, self-updating privacy policy. Choose between 6 languages. Documents hosted and kept up to date. Backed by real lawyers. Additional assistance service with premium legal team for custom Privacy Policy and Terms &amp; Conditions.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-income-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#income-analytics\"><\/a>Income Analytics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/baremetrics.com\" rel=\"nofollow\">Baremetrics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Baremetrics\" rel=\"nofollow\">@Baremetrics<\/a><\/td>\n<td align=\"left\">$79\/mo - $249\/mo<\/td>\n<td align=\"left\">One click and you get dozens of valuable metrics from your Stripe account<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/getmetrics.io\" rel=\"nofollow\">getMetrics<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/getMEtrics\" rel=\"nofollow\">@getMetrics<\/a><\/td>\n<td align=\"left\">$25\/mo - $39\/mo<\/td>\n<td align=\"left\">Stripe metrics and notifications for Stripe events<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-payments-billing--downloads\" class=\"anchor\" aria-hidden=\"true\" href=\"#payments-billing--downloads\"><\/a>Payments, Billing &amp; Downloads<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.paypal.com\" rel=\"nofollow\">PayPal<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PayPal\" rel=\"nofollow\">@PayPal<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">PayPal is an international e-commerce business allowing payments and money transfers to be made through the Internet. Online money transfers serve as electronic alternatives to paying with traditional paper methods, such as checks and money orders<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gumroad.com\" rel=\"nofollow\">Gumroad<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/gumroad\" rel=\"nofollow\">@gumroad<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">See higher conversion, lower fees, and more customer control. Sell films directly to your viewers. Creating digital products is hard, selling them shouldn't be. We let you start selling downloads in seconds.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.fetchapp.com\" rel=\"nofollow\">FetchApp<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/fetchapp\" rel=\"nofollow\">@fetchapp<\/a><\/td>\n<td align=\"left\">$5\/mo - $500\/mo<\/td>\n<td align=\"left\">The Simpler way to Fetch. Simply put, FetchApp allows you to sell and digitally delivery downloadable goods<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.chargify.com\" rel=\"nofollow\">Chargify<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chargify\" rel=\"nofollow\">@chargify<\/a><\/td>\n<td align=\"left\">$459\/mo - $65\/mo<\/td>\n<td align=\"left\">Easily Manage Your Recurring Revenue Business. Customers sign up, make payments, use coupons, upgrade... You bill one-time &amp; recurring fees using whatever pricing model you need, charge cards, send invoices &amp; reminders, etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/recurly.com\" rel=\"nofollow\">Recurly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Recurly\" rel=\"nofollow\">@Recurly<\/a><\/td>\n<td align=\"left\">$99\/mo - $259\/mo<\/td>\n<td align=\"left\">Subscription Billing Automation. As the leading recurring billing platform, Recurly ensures setup is easy, integrations are quick, and our service scales with the needs of your business. With Recurly you'll be ready to accept payments and focus on growing your sales in no time.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.chargebee.com\" rel=\"nofollow\">ChargeBee<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chargebee\" rel=\"nofollow\">@chargebee<\/a><\/td>\n<td align=\"left\">$49\/mo - $249\/mo<\/td>\n<td align=\"left\">ChargeBee is an easy to use recurring billing and invoicing solution for online businesses<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/chargeover.com\" rel=\"nofollow\">ChargeOver<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/chargeover\" rel=\"nofollow\">@ChargeOver<\/a><\/td>\n<td align=\"left\">$65\/mo- $549\/mo<\/td>\n<td align=\"left\">ChargeOver is a full featured invoicing platform supporting payment plans, recurring billing, one-time billing, and all the major gateway backends.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-billing--payment-processing\" class=\"anchor\" aria-hidden=\"true\" href=\"#billing--payment-processing\"><\/a>Billing &amp; Payment Processing<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.braintreepayments.com\" rel=\"nofollow\">Braintree<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/braintree\" rel=\"nofollow\">@braintree<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Accept payments in your app or website. Braintree handles transactions for some of the fastest growing mobile companies like Uber, Airbnb, HotelTonight and Fab. With native, easy-to-follow SDKs for iOS, Android and Windows Phone you can quickly add native payments to your app.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.dwolla.com\" rel=\"nofollow\">Dwolla<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dwolla\" rel=\"nofollow\">@dwolla<\/a><\/td>\n<td align=\"left\">25\u00a2\/transaction<\/td>\n<td align=\"left\">The best way to move money. Dwolla, Inc. is an agent of Veridian Credit Union and all funds associated with your account in the Dwolla network are held in a pooled account at Veridian Credit Union. These funds are not eligible for individual insurance, and may not be eligible for share insurance by the National Credit Union Share Insurance Fund. Dwolla, Inc. is the operator of a software platform that communicates user instructions for funds transfers to Veridian Credit Union.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/stripe.com\" rel=\"nofollow\">Stripe<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/stripe\" rel=\"nofollow\">@stripe<\/a><\/td>\n<td align=\"left\">2.9% + 30\u00a2\/successful charge.<\/td>\n<td align=\"left\">Feature-packed payments. No need to design payment forms from scratch. Stripe Checkout offers a beautiful, customizable payment flow that works great across desktop and mobile. When you use Checkout, you\u2019re always up-to-date, with no extra code required.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pinpayments.com\" rel=\"nofollow\">Pin<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/pin_payments\" rel=\"nofollow\">@pin_payments<\/a><\/td>\n<td align=\"left\">2.9% + 30\u00a2\/successful charge.<\/td>\n<td align=\"left\">Payments, Rebooted. Accepting credit card payments from a global audience typically requires a merchant account. The process of establishing a merchant account for each currency can be too difficult and costly for small businesses.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.paymill.com\" rel=\"nofollow\">PayMill<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Paymill\" rel=\"nofollow\">@Paymill<\/a><\/td>\n<td align=\"left\">0.28 \u20ac - 0.25 \u20ac<\/td>\n<td align=\"left\">Online payments made easy. Make payments personal by customising the checkout according to the flow of your website<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.spreedly.com\" rel=\"nofollow\">Spreedly<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/spreedly\" rel=\"nofollow\">@spreedly<\/a><\/td>\n<td align=\"left\">$150\/mo - $1500\/mo<\/td>\n<td align=\"left\">Payments as a Platform. One of Spreedly's major benefits is reaching a large number of merchant accounts by working across multiple payment gateways. As a direct merchant you can transact globally but deposit funds in unique merchant accounts based on geographic or other business rules. As a SaaS platform you can support the unique merchant accounts of your individual customers. A payment gateway token is your way to indicate to us which unique merchant account this particular transaction will go against. Each unique merchant account = one unique payment gateway token.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/go.wepay.com\" rel=\"nofollow\">WePay<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/wepay\" rel=\"nofollow\">@wepay<\/a><\/td>\n<td align=\"left\">2.9% + 30\u00a2\/transaction.<\/td>\n<td align=\"left\">WePay is the first payments engine to offer platforms \u2014 marketplaces, crowdfunding, and business software\/tools \u2014 a way to own their customer experience while still shielding them from 100 percent of fraud and regulatory risk.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paddle.com\" rel=\"nofollow\">Paddle<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/PaddleHQ\" rel=\"nofollow\">@PaddleHQ<\/a><\/td>\n<td align=\"left\">5% + 50\u00a2\/transaction.<\/td>\n<td align=\"left\">Payment processing and fulfillment, specialized for desktop apps and SaaS subscription services. Handles VAT and invoicing for you, so your accounting becomes easier.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/fattmerchant.com\/\" rel=\"nofollow\">Fattmerchant<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/fattmerchant\" rel=\"nofollow\">@fattmerchant<\/a><\/td>\n<td align=\"left\">0% + interchange + $99+\/mo<\/td>\n<td align=\"left\">Payments unified.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-banking\" class=\"anchor\" aria-hidden=\"true\" href=\"#banking\"><\/a>Banking<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.simple.com\" rel=\"nofollow\">Simple<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/simple\" rel=\"nofollow\">@simple<\/a>)<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The way banking should be.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-phonepbxsms\" class=\"anchor\" aria-hidden=\"true\" href=\"#phonepbxsms\"><\/a>Phone\/PBX\/SMS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.plivo.com\/\" rel=\"nofollow\">Plivo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tropo.com\/\" rel=\"nofollow\">Tropo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.twilio.com\/\" rel=\"nofollow\">Twilio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.phonebooth.com\/\" rel=\"nofollow\">PhoneBooth<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.talkdesk.com\/\" rel=\"nofollow\">TalkDesk<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hellofax.com\/\" rel=\"nofollow\">HelloFax<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.dialpad.com\" rel=\"nofollow\">Dialpad<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dialpadHQ\" rel=\"nofollow\">@DialpadHQ<\/a><\/td>\n<td align=\"left\">$15\/user\/mo<\/td>\n<td align=\"left\">people-first phone and meeting system that includes internet fax and follows you where you work<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/smith.ai\/keypad\" rel=\"nofollow\">Keypad by Smith.ai<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/smithdotai\" rel=\"nofollow\">@smithdotai<\/a><\/td>\n<td align=\"left\">$25\/mo<\/td>\n<td align=\"left\">The most advanced &amp; affordable cloud phone system for small businesses.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-system-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#system-monitoring\"><\/a>System Monitoring<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.appneta.com\/\" rel=\"nofollow\">AppNeta (Tracelytics)<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/sensuapp.org\/\" rel=\"nofollow\">Sensu<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/riemann.io\" rel=\"nofollow\">Riemann<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.pingdom.com\/\" rel=\"nofollow\">Pingdom<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/uptimerobot.com\/\" rel=\"nofollow\">UptimeRobot<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/wheresitup.com\/\" rel=\"nofollow\">Where's It Up?<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Nagios<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Smokeping<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-search\" class=\"anchor\" aria-hidden=\"true\" href=\"#search\"><\/a>Search<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/bonsai.io\/\" rel=\"nofollow\">Bonsai<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.websolr.com\" rel=\"nofollow\">WebSolr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.searchify.com\/\" rel=\"nofollow\">Searchify<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.searchblox.com\" rel=\"nofollow\">SearchBlox<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-security\" class=\"anchor\" aria-hidden=\"true\" href=\"#security\"><\/a>Security<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/portswigger.net\/burp\" rel=\"nofollow\">Burp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/duo.com\/\" rel=\"nofollow\">DuoSecurity<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/authy.com\/\" rel=\"nofollow\">Authy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.anchorFree.com\" rel=\"nofollow\">AnchorFree<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/encrypt.me\" rel=\"nofollow\">Encrypt.me<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tinfoilsecurity.com\/\" rel=\"nofollow\">Tinfoil Security<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-shipping\" class=\"anchor\" aria-hidden=\"true\" href=\"#shipping\"><\/a>Shipping<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.shipwire.com\" rel=\"nofollow\">Shipwire<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.shyp.com\" rel=\"nofollow\">Shyp<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.getwhiplash.com\" rel=\"nofollow\">Whiplash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-user-feedback\" class=\"anchor\" aria-hidden=\"true\" href=\"#user-feedback\"><\/a>User Feedback<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.uservoice.com\/\" rel=\"nofollow\">Uservoice<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/userecho.com\/\" rel=\"nofollow\">Userecho<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-designers\" class=\"anchor\" aria-hidden=\"true\" href=\"#designers\"><\/a>Designers<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/dribbble.com\" rel=\"nofollow\">Dribbble<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/sortfolio.com\" rel=\"nofollow\">Sortfolio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.behance.net\" rel=\"nofollow\">Behance<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-notes\" class=\"anchor\" aria-hidden=\"true\" href=\"#notes\"><\/a>Notes<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.google.com\/docs\/about\/\" rel=\"nofollow\">Google Docs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/evernote.com\/\" rel=\"nofollow\">Evernote<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.google.com\/keep\/\" rel=\"nofollow\">Google Keep<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/workflowy.com\/\" rel=\"nofollow\">Workflowy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/quip.com\/\" rel=\"nofollow\">Quip<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/etherpad.org\/\" rel=\"nofollow\">Etherpad<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.kamihq.com\" rel=\"nofollow\">Kami<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">View, edit, annotate, and collaborate on documents on your browser<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.onenote.com\/\" rel=\"nofollow\">OneNote<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/msonenote\" rel=\"nofollow\">@msonenote<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">The digital note-taking app for your devices<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.taskade.com\" rel=\"nofollow\">Taskade<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/taskade\" rel=\"nofollow\">@taskade<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Real-time collaborative notes, checklists, and outlines for teams.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/nulis.io\" rel=\"nofollow\">Nulis<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Outliner \/ mind-mapping tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/writingstreak.io\" rel=\"nofollow\">WritingStreak<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Develop a writing habit, take daily notes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dynalist.io\/\" rel=\"nofollow\">Dynalist<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/dynalisthq\" rel=\"nofollow\">@dynalisthq<\/a><\/td>\n<td align=\"left\">Free - $8\/mo<\/td>\n<td align=\"left\">Outliner, note-taking app, checklists.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/notion.so\" rel=\"nofollow\">Notion<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/notionhq\" rel=\"nofollow\">@notionhq<\/a><\/td>\n<td align=\"left\">Free -  $8\/user\/mo<\/td>\n<td align=\"left\">Write, plan, collaborate, and get organized. Notion is all you need \u2014 in one tool.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.inkdrop.info\" rel=\"nofollow\">Inkdrop<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$4.99\/mo<\/td>\n<td align=\"left\">The Note-Taking App for Markdown Lovers.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-group-communicationchat-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#group-communicationchat-tools\"><\/a>Group Communication\/Chat Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/getkandan.com\" rel=\"nofollow\">Kandan<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.yammer.com\/\" rel=\"nofollow\">Yammer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/limechat.net\/\" rel=\"nofollow\">Limechat<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.flowdock.com\/\" rel=\"nofollow\">Flowdock<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free for teams up to 5 people<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.stride.com\" rel=\"nofollow\">Stride<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/slack.com\/\" rel=\"nofollow\">Slack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free - $12.5\/user<\/td>\n<td align=\"left\">Team collaboration and chat tool. Supports webhooks and has integrations with other tools.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.skype.com\" rel=\"nofollow\">Skype<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Video conference, chat and voice call VOIP application.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/hangouts.google.com\/\" rel=\"nofollow\">Google Hangouts<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Video conference and chat application by Google. It is browser based and supports multiple participants on the same video call.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.gotomeeting.com\" rel=\"nofollow\">GoToMeeting<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.irccloud.com\/\" rel=\"nofollow\">IRCCloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/buddycloud.com\" rel=\"nofollow\">Buddycloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gitter.im\" rel=\"nofollow\">Gitter<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Chat for open-source and private development teams, with strong Github integration features<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/appear.in\/\" rel=\"nofollow\">appear.in<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">one click video conversations<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.convo.com\/\" rel=\"nofollow\">Convo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.zoom.us\/\" rel=\"nofollow\">Zoom<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/telegram.org\/\" rel=\"nofollow\">Telegram<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/matrix.org\/\" rel=\"nofollow\">Matrix<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Decentralised Open Source chat protocol with <a href=\"https:\/\/matrix.org\/docs\/projects\/try-matrix-now.html#application-services\" rel=\"nofollow\">bridges to everywhere<\/a><\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-remote-collaboration\" class=\"anchor\" aria-hidden=\"true\" href=\"#remote-collaboration\"><\/a>Remote Collaboration<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.teamviewer.com\" rel=\"nofollow\">TeamViewer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.screenmailer.com\" rel=\"nofollow\">Screenmailer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#dns\"><\/a>DNS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/dyn.com\/dns\/\" rel=\"nofollow\">DynDNS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudflare.com\/\" rel=\"nofollow\">Cloudflare<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/route53\/\" rel=\"nofollow\">Amazon Route 53<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/dnsimple.com\/\" rel=\"nofollow\">DNSimple<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudns.net\/\" rel=\"nofollow\">ClouDNS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/Freedns.afraid.org\" rel=\"nofollow\">FreeDNS<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-status-blogsuser-alerts\" class=\"anchor\" aria-hidden=\"true\" href=\"#status-blogsuser-alerts\"><\/a>Status Blogs\/User Alerts<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.statuspage.io\/\" rel=\"nofollow\">StatusPage.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tumblr.com\/\" rel=\"nofollow\">Tumblr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.hellobar.com\" rel=\"nofollow\">HelloBar<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/status.io\" rel=\"nofollow\">Status.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/statusio\" rel=\"nofollow\">@statusio<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Hosted system status pages<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-forms--surveys\" class=\"anchor\" aria-hidden=\"true\" href=\"#forms--surveys\"><\/a>Forms \/ Surveys<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.wufoo.com\" rel=\"nofollow\">Wufoo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Google Forms<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.typeform.com\" rel=\"nofollow\">Typeform<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/typeform\" rel=\"nofollow\">@typeform<\/a><\/td>\n<td align=\"left\">$0\/mo - $25\/mo<\/td>\n<td align=\"left\">Ask Awesomely! Typeform makes asking questions easy, human &amp; beautiful. A user experience that makes your questions look &amp; feel great everywhere. Stimulated, inspired, excited, happy respondents boost completion rates. Gain insights with integrated analysis tools.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/qualaroo.com\" rel=\"nofollow\">Qualaroo<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/qualarooinc\" rel=\"nofollow\">@qualarooinc<\/a><\/td>\n<td align=\"left\">$63\/mo - 499\/mo<\/td>\n<td align=\"left\">Qualaroo website surveys uncover customer insights that lead to better business results.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/formcarry.com\" rel=\"nofollow\">Formcarry<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$0\/mo - $99\/mo<\/td>\n<td align=\"left\">Handle forms without a single line of back-end code.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/formkeep.com\" rel=\"nofollow\">FormKeep<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/formkeep\" rel=\"nofollow\">@formkeep<\/a><\/td>\n<td align=\"left\">$59\/mo - $199\/mo<\/td>\n<td align=\"left\">Premium Forms Backend service with built in spam protection, data integrations to over a 1000 services, data retention policies and more.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-source-code-hosting\" class=\"anchor\" aria-hidden=\"true\" href=\"#source-code-hosting\"><\/a>Source Code Hosting<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/bitbucket.org\/\" rel=\"nofollow\">BitBucket<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.codebasehq.com\/\" rel=\"nofollow\">Codebase<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/\">GitHub<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/unfuddle.com\/\" rel=\"nofollow\">Unfuddle<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/about.gitlab.com\/\" rel=\"nofollow\">GitLab<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/launchpad.net\/\" rel=\"nofollow\">Launchpad<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.tuxfamily.org\" rel=\"nofollow\">TuxFamily<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pypi.python.org\/pypi\/kforge\" rel=\"nofollow\">KForge<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.versionshelf.com\" rel=\"nofollow\">VersionShelf<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">30-day Free trial, <a href=\"https:\/\/www.versionshelf.com\/about_Free_plan\" rel=\"nofollow\">Free for students\/teachers<\/a>.<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.assembla.com\/home\" rel=\"nofollow\">Assembla<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/assembla\" rel=\"nofollow\">@assembla<\/a><\/td>\n<td align=\"left\">$24\/m - $99\/m<\/td>\n<td align=\"left\">Git hosting, all tightly integrated with your tasks. Browse files online, compare revisions, merge code and maintain a sustainable codebase.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-design-collaboration\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-collaboration\"><\/a>Design Collaboration<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.conceptshare.com\" rel=\"nofollow\">ConceptShare<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.framebench.com\" rel=\"nofollow\">Framebench<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/framebench\" rel=\"nofollow\">@framebench<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">The 'smooth' way to Share, Review &amp; Discuss any file online.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.notableapp.com\/\" rel=\"nofollow\">Notable<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-paas\" class=\"anchor\" aria-hidden=\"true\" href=\"#paas\"><\/a>PaaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.heroku.com\" rel=\"nofollow\">Heroku<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">1000 Free hours\/mo for verified accounts<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudfoundry.org\" rel=\"nofollow\">Cloud Foundry<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">host it yourself, or use one of MANY hosting services<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/pagodabox.io\" rel=\"nofollow\">Pagodabox<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.clever-cloud.com\" rel=\"nofollow\">Clever Cloud<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/cloud.google.com\/appengine\/docs\/\" rel=\"nofollow\">Google App Engine<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.openshift.com\" rel=\"nofollow\">OpenShift<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">OpenShift gives you all the tools you need to develop, host and scale your apps in the public or private cloud. Get started today.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.engineyard.com\" rel=\"nofollow\">Engine Yard<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.ctl.io\/appfog\/\" rel=\"nofollow\">AppFog<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/jelastic.com\/\" rel=\"nofollow\">Jelastic<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.cloudbees.com\" rel=\"nofollow\">CloudBees<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/azure.microsoft.com\" rel=\"nofollow\">Microsoft Azure<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Known for IaaS<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/aws.amazon.com\/elasticbeanstalk\/\" rel=\"nofollow\">Amazon Web Services<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Like Azure, it's more known for IaaS<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/deis.io\" rel=\"nofollow\">Deis<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/scalingo.com\" rel=\"nofollow\">Scalingo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">European PAAS<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/eldarion.cloud\/\" rel=\"nofollow\">Eldarion<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Service for open source Kelproject that they develop in Python and Go<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-vps\" class=\"anchor\" aria-hidden=\"true\" href=\"#vps\"><\/a>VPS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.linode.com\/\" rel=\"nofollow\">Linode<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.ramnode.com\/\" rel=\"nofollow\">Ramnode<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.digitalocean.com\/\" rel=\"nofollow\">DigitalOcean<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.vultr.com\/\" rel=\"nofollow\">Vultr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.ovh.com\" rel=\"nofollow\">OVH<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.scaleway.com\/\" rel=\"nofollow\">Scaleway<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">\n<\/td><\/tr><\/tbody><\/table><h3><a id=\"user-content-heroku-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#heroku-tools\"><\/a>Heroku Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/hirefire.io\/\" rel=\"nofollow\">Hirefire<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-aws-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#aws-tools\"><\/a>AWS Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/ylastic.com\/\" rel=\"nofollow\">Ylastic<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.skeddly.com\" rel=\"nofollow\">Skeddly<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.gorillastack.com\" rel=\"nofollow\">GorillaStack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Tool for automating processes inside AWS including Cost Optimisation, Backup and more!<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-database-aas\" class=\"anchor\" aria-hidden=\"true\" href=\"#database-aas\"><\/a>Database-aaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.humongous.io\" rel=\"nofollow\">HumongouS.io<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">HumongouS.io is a web-based user interface (GUI) for MongoDB.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/mlab.com\" rel=\"nofollow\">mLab<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.compose.com\" rel=\"nofollow\">Compose<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Compose is a fully-managed platform used by developers to deploy, host and scale databases (Elasticsearch and MongoDB.)<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/redislabs.com\" rel=\"nofollow\">RedisLabs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/RedisLabsInc\" rel=\"nofollow\">@RedisLabsInc<\/a><\/td>\n<td align=\"left\">Free - $338+\/mo<\/td>\n<td align=\"left\">RedisLabs offers fully-managed cloud service for hosting and running your redis or memcache datasets in a highly-available and scalable manner, with predictable and stable top performance.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-backend-aas\" class=\"anchor\" aria-hidden=\"true\" href=\"#backend-aas\"><\/a>Backend-aaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/firebase.google.com\" rel=\"nofollow\">Firebase<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/Firebase\" rel=\"nofollow\">@Firebase<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">A powerful API to store and sync data in realtime. Build realtime mobile and web apps in minutes using client-side code and the powerful Firebase API. When data changes, apps built with Firebase update instantly across every device -- web or mobile.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/hood.ie\/\" rel=\"nofollow\">Hoodie<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.baasbox.com\" rel=\"nofollow\">BaasBox<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/loopback.io\/\" rel=\"nofollow\">LoopBack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/paraio.com\" rel=\"nofollow\">Para<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/para_io\" rel=\"nofollow\">@Para_IO<\/a><\/td>\n<td align=\"left\">Free - $99\/mo<\/td>\n<td align=\"left\">A flexible and affordable backend API for busy developers.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-websockets-aas\" class=\"anchor\" aria-hidden=\"true\" href=\"#websockets-aas\"><\/a>WebSockets-aaS<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/pusher.com\" rel=\"nofollow\">Pusher<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.scaledrone.com\" rel=\"nofollow\">ScaleDrone<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-ops-alerts-and-scheduling\" class=\"anchor\" aria-hidden=\"true\" href=\"#ops-alerts-and-scheduling\"><\/a>Ops Alerts and Scheduling<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.pagerduty.com\/\" rel=\"nofollow\">PagerDuty<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.opsgenie.com\" rel=\"nofollow\">Opsgenie<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/opsgenie\" rel=\"nofollow\">@opsgenie<\/a><\/td>\n<td align=\"left\">$0 - $16 user\/mo<\/td>\n<td align=\"left\">We make alerts work for you. We provide the tools you need to design meaningful, actionable alerts and ensure the right people are notified.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.victorops.com\" rel=\"nofollow\">VictorOps<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$9 - $49 user\/month<\/td>\n<td align=\"left\">Incident Management to Support the DevOps Lifecycle, from First Alert to Post-Incident Review<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-video-hosting\" class=\"anchor\" aria-hidden=\"true\" href=\"#video-hosting\"><\/a>Video Hosting<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/wistia.com\" rel=\"nofollow\">Wistia<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/developer.jwplayer.com\" rel=\"nofollow\">JW Player<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Powerful suite of APIs and tools that enable video hosting, content delivery, playback, and data insights.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-knowledge-trackingwiki\" class=\"anchor\" aria-hidden=\"true\" href=\"#knowledge-trackingwiki\"><\/a>Knowledge Tracking\/Wiki<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/confluence\" rel=\"nofollow\">Confluence<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.atlassian.com\/software\/confluence\/questions\" rel=\"nofollow\">Confluence Questions (Q&amp;A for your team)<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">bitbucket's default Wiki<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">comes with own repo, markdown\/creole, much lighter than Confluence.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/slimwiki.com\" rel=\"nofollow\">SlimWiki<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/documize.com\" rel=\"nofollow\">Documize<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-offsite-backups\" class=\"anchor\" aria-hidden=\"true\" href=\"#offsite-backups\"><\/a>Offsite Backups<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"http:\/\/www.tarsnap.com\" rel=\"nofollow\">Tarsnap<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.rsync.net\/\" rel=\"nofollow\">Rsync.net <\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/spideroak.com\" rel=\"nofollow\">SpiderOak<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/spideroak\" rel=\"nofollow\">@spideroak<\/a><\/td>\n<td align=\"left\">First 2gb Free - $100\/100gb\/year (check <a href=\"https:\/\/twitter.com\/spideroak\" rel=\"nofollow\">twitter<\/a> for deals)<\/td>\n<td align=\"left\">Store. Sync. Share. Privately.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-personal-machine-backups\" class=\"anchor\" aria-hidden=\"true\" href=\"#personal-machine-backups\"><\/a>Personal Machine Backups<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.crashplan.com\" rel=\"nofollow\">Crashplan<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\">Arq + S3\/Glacier<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-remote-workers\" class=\"anchor\" aria-hidden=\"true\" href=\"#remote-workers\"><\/a>Remote Workers<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.upwork.com\" rel=\"nofollow\">oDesk<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.Freelancer.com\/\" rel=\"nofollow\">Freelancer<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/taskarmy.com\" rel=\"nofollow\">TaskArmy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/99designs.com\/\" rel=\"nofollow\">99Designs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.fiverr.com\" rel=\"nofollow\">Fiverr<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.remotelyawesomejobs.com\" rel=\"nofollow\">Remotely Awesome Jobs<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-deployment\" class=\"anchor\" aria-hidden=\"true\" href=\"#deployment\"><\/a>Deployment<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.chef.io\" rel=\"nofollow\">Chef<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.fabfile.org\" rel=\"nofollow\">Fabric<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/puppet.com\" rel=\"nofollow\">Puppet<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.ansible.com\" rel=\"nofollow\">Ansible<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.vagrantup.com\" rel=\"nofollow\">Vagrant<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/saltstack.com\/community\/\" rel=\"nofollow\">Salt<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/prelang.com\" rel=\"nofollow\">Prelang<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Automatic Rails Development<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ngrok.com\" rel=\"nofollow\">Ngrok<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-seo-tools\" class=\"anchor\" aria-hidden=\"true\" href=\"#seo-tools\"><\/a>SEO Tools<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.accuranker.com\/\" rel=\"nofollow\">AccuRanker<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ahrefs.com\" rel=\"nofollow\">Ahrefs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ahrefs\" rel=\"nofollow\">@ahrefs<\/a><\/td>\n<td align=\"left\">$79\/mo - $2500\/mo<\/td>\n<td align=\"left\">Ahrefs provide a complete digital marketing suite with tools for analyzing back links, analizing websites, rank tracking, content exploring and more.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/serpbook.com\" rel=\"nofollow\">SerpBook<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.woorank.com\" rel=\"nofollow\">WooRank<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/woorank\" rel=\"nofollow\">@woorank<\/a><\/td>\n<td align=\"left\">Free - $49\/mo<\/td>\n<td align=\"left\">WooRank analyzes your website for optimization best practices and shows how it ranks against your competition. Its real-time brandable reports, consisting of over 150 data-points, help you to instantly spot critical issues that impact traffic, usability and conversions. Sync your analytics, social accounts and keywords for even more robust tracking.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/moz.com\" rel=\"nofollow\">Moz<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/moz\" rel=\"nofollow\">@moz<\/a><\/td>\n<td align=\"left\">$99\/mo - $599\/mo<\/td>\n<td align=\"left\">Moz provides you all the tools you need to effectively do search engine optimization. On site optimization graders, competitor tracking, back link analysis, rank trakcing and many more features available to users.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/kwfinder.com\/\" rel=\"nofollow\">KWFinder<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/mangools_com\" rel=\"nofollow\">@mangools_com<\/a><\/td>\n<td align=\"left\">$30+\/mo<\/td>\n<td align=\"left\">Keyword research tool<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-api-builder\" class=\"anchor\" aria-hidden=\"true\" href=\"#api-builder\"><\/a>API Builder<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.getpostman.com\/\" rel=\"nofollow\">Postman<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/postmanclient\" rel=\"nofollow\">@postmanclient<\/a><\/td>\n<td align=\"left\">Free - $8\/mo<\/td>\n<td align=\"left\">Build, test, document and monitor your APIs faster. Used by 5 million developers and more than 100,000 companies.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/deployd.com\" rel=\"nofollow\">Deployd<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/deploydapp\" rel=\"nofollow\">@deploydapp<\/a><\/td>\n<td align=\"left\">Free (OSS)<\/td>\n<td align=\"left\">Design, build, and scale APIs for web and mobile apps in minutes instead of days.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/apiary.io\" rel=\"nofollow\">Apiary<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/apiaryio\" rel=\"nofollow\">@apiary<\/a><\/td>\n<td align=\"left\">Free - $99\/mo<\/td>\n<td align=\"left\">Powerful API Design Stack. Built for Developers. Work together to quickly design, prototype, document and test APIs.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-password-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#password-management\"><\/a>Password Management<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.lastpass.com\" rel=\"nofollow\">Lastpass<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/1password.com\" rel=\"nofollow\">1Password<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.passpack.com\" rel=\"nofollow\">Passpack<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.keepassx.org\" rel=\"nofollow\">KeePassX<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\"><a href=\"https:\/\/www.eff.org\/deeplinks\/2014\/07\/protecting-your-anonymity-how-sex-workers\" rel=\"nofollow\">Recommended by EFF<\/a><\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/keepassxc.org\/\" rel=\"nofollow\">KeePassXC<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/KeePassXC\" rel=\"nofollow\">@KeePassXC<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">further development of KeePassX<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.enpass.io\" rel=\"nofollow\">Enpass<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.dashlane.com\" rel=\"nofollow\">Dashlane<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">\n<\/td><\/tr><tr><td align=\"left\"><a href=\"https:\/\/bitwarden.com\/\" rel=\"nofollow\">Bitwarden<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free for personal use \/ Self-hosted<\/td>\n<td align=\"left\">Open Source Password Management Solutions<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-sources-of-clicksad-platforms\" class=\"anchor\" aria-hidden=\"true\" href=\"#sources-of-clicksad-platforms\"><\/a>Sources of Clicks\/Ad Platforms<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/neoreach.com\/\" rel=\"nofollow\">NeoReach<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.sitescout.com\/\" rel=\"nofollow\">SiteScout<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/advertise.bingads.microsoft.com\" rel=\"nofollow\">Bing Ads<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.taboola.com\/\" rel=\"nofollow\">Taboola<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.outbrain.com\/\" rel=\"nofollow\">Outbrain<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/www.google.com\/adwords\/\" rel=\"nofollow\">Google Adwords<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.facebook.com\/business\/\" rel=\"nofollow\">Facebook Advertising<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-storage\" class=\"anchor\" aria-hidden=\"true\" href=\"#storage\"><\/a>Storage<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kloudless.com\" rel=\"nofollow\">Kloudless<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/kloudless\" rel=\"nofollow\">@Kloudless<\/a><\/td>\n<td align=\"left\">Free - $500\/mo<\/td>\n<td align=\"left\">Kloudless provides developers with a single cloud storage API in the place of many<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/cloudbuddy.cloud\" rel=\"nofollow\">CloudBuddy<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">$1\/10GB\/month<\/td>\n<td align=\"left\">SFTP Online Storage<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-task-scheduling\" class=\"anchor\" aria-hidden=\"true\" href=\"#task-scheduling\"><\/a>Task Scheduling<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.easycron.com\" rel=\"nofollow\">EasyCron<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ifttt.com\" rel=\"nofollow\">IFTTT<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/zapier.com\" rel=\"nofollow\">Zapier<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zapier\" rel=\"nofollow\">Zapier<\/a><\/td>\n<td align=\"left\">$99\/mo - $15\/mo<\/td>\n<td align=\"left\">Superpowers to get your work done. A Zap is a link between two apps (a \"trigger\" and an \"action\"). Zaps run automatically in the background every few minutes to move and manage data on your behalf. Only live Zaps count against your limit \u2014 you can have as many paused and unfinished Zaps as you'd like. For example, one Zap might be \"Send me an SMS every time I get a new email\".<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.integromat.com\" rel=\"nofollow\">Integromat<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/integromat\" rel=\"nofollow\">@integromat<\/a><\/td>\n<td align=\"left\">Free for &lt; 1000 operations or 100 MB\/mo, paid from $9\/mo<\/td>\n<td align=\"left\">Automates your tasks &amp; connects to almost any app or service. Very powerful toolset with features like error handling, iterators, aggregators, routers, functions, data stores, etc.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/deadmanssnitch.com\" rel=\"nofollow\">Dead Man's Snitch<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/deadmanssnitch\" rel=\"nofollow\">@DeadMansSnitch<\/a><\/td>\n<td align=\"left\">Free - $49\/mo<\/td>\n<td align=\"left\">Monitoring for scheduled tasks (like cron) and service heartbeats. Makes it easy to know when and why a particular job failed.<\/td>\n<\/tr><\/tbody><\/table><h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/kapeli.com\/dash\" rel=\"nofollow\">Dash<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/zealdocs.org\" rel=\"nofollow\">Zeal<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/zealdocs\" rel=\"nofollow\">@zealdocs<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Offline documentation browser<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/devdocs.io\/\" rel=\"nofollow\">DevDocs<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/DevDocs\" rel=\"nofollow\">@DevDocs<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">DevDocs combines multiple API documentations in a fast, organized, and searchable interface. You don't have to use your mouse \u2014 use keyboard shortcuts. The search supports fuzzy matching. You can search using your browser's address bar. DevDocs is Free and open source.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-engineering-metrics\" class=\"anchor\" aria-hidden=\"true\" href=\"#engineering-metrics\"><\/a>Engineering Metrics<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.staticobject.com\" rel=\"nofollow\">Static Object<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/StaticObjectDev\" rel=\"nofollow\">@StaticObjectDev<\/a><\/td>\n<td align=\"left\">$295\/mo - $495\/mo<\/td>\n<td align=\"left\">Metrics that enhance transparency for engineering teams<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/gitprime.com\" rel=\"nofollow\">Gitprime<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/GitPrime\" rel=\"nofollow\">@GitPrime<\/a><\/td>\n<td align=\"left\">$749\/mo - $2,549\/mo<\/td>\n<td align=\"left\">Metrics in context for data-driven engineering leaders<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-business-cards-and-print-material\" class=\"anchor\" aria-hidden=\"true\" href=\"#business-cards-and-print-material\"><\/a>Business Cards and Print Material<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/www.moo.com\" rel=\"nofollow\">Moo<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/www.psprint.com\" rel=\"nofollow\">PSPrint<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"http:\/\/vistaprint.com\" rel=\"nofollow\">Vista Print<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-presentations--slides\" class=\"anchor\" aria-hidden=\"true\" href=\"#presentations--slides\"><\/a>Presentations \/ Slides<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/prezi.com\/\" rel=\"nofollow\">Prezi<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/gamell\/markpress\">markpress<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">HTML presentations from Markdown files<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/github.com\/hakimel\/reveal.js\">Reveal.js<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Flexible HTML presenation framework for creating beautiful slides<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/slides.com\/\" rel=\"nofollow\">Slides<\/a><\/td>\n<td align=\"left\">-<\/td>\n<td align=\"left\">Free - $20\/mo<\/td>\n<td align=\"left\">Online editor built with Reveal.js that includes a graphical interface<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-fundraising--investor-relations\" class=\"anchor\" aria-hidden=\"true\" href=\"#fundraising--investor-relations\"><\/a>Fundraising \/ Investor Relations<\/h3>\n<table><thead><tr><th align=\"left\">Service<\/th>\n<th align=\"left\">Twitter<\/th>\n<th align=\"left\">Pricing<\/th>\n<th align=\"left\">Description<\/th>\n<\/tr><\/thead><tbody><tr><td align=\"left\"><a href=\"https:\/\/captable.io\" rel=\"nofollow\">Captable.io<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/captable_io\" rel=\"nofollow\">@captable_io<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Captable.io is a free, fully-featured cap table management tool that enables step-by-step creation, easy collaboration &amp; sharing, convertible note &amp; option calculators, and powerful round &amp; exit modeling.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/startupdisclosure.io\" rel=\"nofollow\">Disclosure<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Helps startups fulfill investor update requests, so investors can manage and aggregate portfolio information. It replaces today's cumbersome email based processes.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/notegenie.io\" rel=\"nofollow\">NoteGenie<\/a>  \/ <a href=\"https:\/\/safegenie.io\" rel=\"nofollow\">SAFEGenie<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">A calculator for exploring the impact of convertible notes and SAFEs on founder ownership.<\/td>\n<\/tr><tr><td align=\"left\"><a href=\"https:\/\/ipo-ready.com\" rel=\"nofollow\">IPO Ready<\/a><\/td>\n<td align=\"left\"><a href=\"https:\/\/twitter.com\/ltse\" rel=\"nofollow\">@ltse<\/a><\/td>\n<td align=\"left\">Free<\/td>\n<td align=\"left\">Quickly assess your startup's IPO readiness. Learn what it takes to prepare for an IPO.<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-use\" class=\"anchor\" aria-hidden=\"true\" href=\"#use\"><\/a>Use<\/h2>\n<p>The best ways to use this list are:<\/p>\n<ul><li>by browing the <a href=\"#contents\">contents<\/a><\/li>\n<li>by using <kbd>command<\/kbd> + <kbd>F<\/kbd> to search the contents<\/li>\n<\/ul><p>This list also uses tags to help when searching the contents:<\/p>\n<ul><li><strong>Hosted?<\/strong> <code>Hosted<\/code>, <code>Self-hosted<\/code><\/li>\n<\/ul><h2><a id=\"user-content-see-also\" class=\"anchor\" aria-hidden=\"true\" href=\"#see-also\"><\/a>See also<\/h2>\n<ul><li><a href=\"https:\/\/github.com\/styfle\/awesome-online-ide\">Awesome Online IDEs<\/a> - A list of awesome online development environments<\/li>\n<li><a href=\"https:\/\/www.apify.com\/metamn\/2MS8r-api-https-github-com-cjbarber-toolsofthetrade\" rel=\"nofollow\">Data Extractor for Tools of The Trade<\/a> - Data extracted with Apify to JSON\/CSV\/XLS<\/li>\n<\/ul><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h2>\n<p><strong><a href=\"https:\/\/www.chrisbarber.co\" rel=\"nofollow\">Chris Barber<\/a><\/strong><\/p>\n<p><a href=\"https:\/\/github.com\/cjbarber\"><img src=\"https:\/\/camo.githubusercontent.com\/61c3d43199fc0b9a40b5a0817d4316f273598d89\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f666f6c6c6f776572732f636a6261726265722e7376673f7374796c653d736f6369616c266c6162656c3d466f6c6c6f77253230636a626172626572\" alt=\"GitHub\" data-canonical-src=\"https:\/\/img.shields.io\/github\/followers\/cjbarber.svg?style=social&amp;label=Follow%20cjbarber\" \/><\/a> <a href=\"https:\/\/twitter.com\/chrisbarber\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/d61fd278cca6002cb53a3b245074e18f5e9a6bc0\/68747470733a2f2f696d672e736869656c64732e696f2f747769747465722f666f6c6c6f772f63687269736261726265722e7376673f7374796c653d736f6369616c\" alt=\"Twitter\" data-canonical-src=\"https:\/\/img.shields.io\/twitter\/follow\/chrisbarber.svg?style=social\" \/><\/a><\/p>\n<p><strong><a href=\"https:\/\/davison.io\" rel=\"nofollow\">Craig Davison<\/a><\/strong><\/p>\n<p><a href=\"https:\/\/github.com\/davisonio\"><img src=\"https:\/\/camo.githubusercontent.com\/6b985f93d58bac679fa7f728b3eec92ea82f12ba\/68747470733a2f2f696d672e736869656c64732e696f2f6769746875622f666f6c6c6f776572732f64617669736f6e696f2e7376673f7374796c653d736f6369616c266c6162656c3d466f6c6c6f772532304064617669736f6e696f\" alt=\"GitHub\" data-canonical-src=\"https:\/\/img.shields.io\/github\/followers\/davisonio.svg?style=social&amp;label=Follow%20@davisonio\" \/><\/a> <a href=\"https:\/\/twitter.com\/davisonio\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/9dc94489441aea9763d5ef56641c172de7dbca2c\/68747470733a2f2f696d672e736869656c64732e696f2f747769747465722f666f6c6c6f772f64617669736f6e696f2e7376673f7374796c653d736f6369616c\" alt=\"Twitter\" data-canonical-src=\"https:\/\/img.shields.io\/twitter\/follow\/davisonio.svg?style=social\" \/><\/a><\/p>\n<p>...and <a href=\"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade\/graphs\/contributors\">contributors<\/a>.<\/p>\n<p><strong><a href=\"https:\/\/github.com\/cjbarber\/ToolsOfTheTrade\/blob\/master\/contributing.md\">Contributions are welcome!<\/a><\/strong><\/p>\n<\/article>","created_at":"2018-11-07T16:22:19+0000","updated_at":"2018-11-07T16:22:19+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":62,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/3237905?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":65,"title":"Kickball\/awesome-selfhosted","url":"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/README.md","content":"<p><a href=\"https:\/\/github.com\/sindresorhus\/awesome\"><img src=\"https:\/\/camo.githubusercontent.com\/13c4e50d88df7178ae1882a203ed57b641674f94\/68747470733a2f2f63646e2e7261776769742e636f6d2f73696e647265736f726875732f617765736f6d652f643733303566333864323966656437386661383536353265336136336531353464643865383832392f6d656469612f62616467652e737667\" alt=\"Awesome\" data-canonical-src=\"https:\/\/cdn.rawgit.com\/sindresorhus\/awesome\/d7305f38d29fed78fa85652e3a63e154dd8e8829\/media\/badge.svg\" \/><\/a><\/p>\n<p>Selfhosting is the process of locally hosting and managing applications instead of renting from SaaS providers.<\/p>\n<p>This is a list of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Free_software\" rel=\"nofollow\">Free<\/a> Software <a href=\"https:\/\/en.wikipedia.org\/wiki\/Network_service\" rel=\"nofollow\">network services<\/a> and <a href=\"https:\/\/en.wikipedia.org\/wiki\/Web_application\" rel=\"nofollow\">web applications<\/a> which can be hosted locally. Non-Free software is listed on the <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/non-free.md\">Non-Free<\/a> page.<\/p>\n<p>See <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/.github\/CONTRIBUTING.md\">Contributing<\/a>.<\/p>\n<hr \/><ul><li>List of Software\n<ul><li><a href=\"#analytics\">Analytics<\/a><\/li>\n<li><a href=\"#archiving-and-digital-preservation-dp\">Archiving and Digital Preservation (DP)<\/a><\/li>\n<li><a href=\"#automation\">Automation<\/a><\/li>\n<li><a href=\"#blogging-platforms\">Blogging Platforms<\/a><\/li>\n<li><a href=\"#bookmarks-and-link-sharing\">Bookmarks and Link Sharing<\/a><\/li>\n<li><a href=\"#calendaring-and-contacts-management\">Calendaring and Contacts Management<\/a><\/li>\n<li><a href=\"#communication-systems\">Communication systems<\/a>\n<ul><li><a href=\"#custom-communication-systems\">Custom communication systems<\/a><\/li>\n<li><a href=\"#email\">Email<\/a>\n<ul><li><a href=\"#complete-solutions\">Complete solutions<\/a><\/li>\n<li><a href=\"#mail-transfer-agents\">Mail Transfer Agents<\/a><\/li>\n<li><a href=\"#mail-delivery-agents\">Mail Delivery Agents<\/a><\/li>\n<li><a href=\"#mailing-lists-and-newsletters\">Mailing lists and newsletters<\/a><\/li>\n<li><a href=\"#webmail-clients\">Webmail clients<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#irc\">IRC<\/a><\/li>\n<li><a href=\"#sip\">SIP<\/a><\/li>\n<li><a href=\"#ipbx\">IPBX<\/a><\/li>\n<li><a href=\"#social-networks-and-forums\">Social Networks and Forums<\/a><\/li>\n<li><a href=\"#xmpp\">XMPP<\/a>\n<ul><li><a href=\"#xmpp-servers\">XMPP Servers<\/a><\/li>\n<li><a href=\"#xmpp-web-clients\">XMPP Web Clients<\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<li><a href=\"#conference-management\">Conference Management<\/a><\/li>\n<li><a href=\"#content-management-systems-cms\">Content Management Systems (CMS)<\/a>\n<ul><li><a href=\"#e-commerce\">E-commerce<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#dns\">DNS<\/a><\/li>\n<li><a href=\"#document-management\">Document Management<\/a><\/li>\n<li><a href=\"#e-books-and-integrated-library-systems-ils\">E-books and Integrated Library Systems (ILS)<\/a><\/li>\n<li><a href=\"#federated-identityauthentication\">Federated Identity\/Authentication<\/a><\/li>\n<li><a href=\"#feed-readers\">Feed Readers<\/a><\/li>\n<li><a href=\"#file-sharing-and-synchronization\">File Sharing and Synchronization<\/a>\n<ul><li><a href=\"#distributed-filesystems\">Distributed filesystems<\/a><\/li>\n<li><a href=\"#file-transfersynchronization\">File transfer\/synchronization<\/a><\/li>\n<li><a href=\"#peer-to-peer-filesharing\">Peer-to-peer filesharing<\/a><\/li>\n<li><a href=\"#object-storagefile-servers\">Object storage\/file servers<\/a><\/li>\n<li><a href=\"#single-clickdrag-n-drop-upload\">Single-click\/drag-n-drop upload<\/a><\/li>\n<li><a href=\"#web-based-file-managers\">Web based file managers<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#games\">Games<\/a><\/li>\n<li><a href=\"#gateways\">Gateways<\/a><\/li>\n<li><a href=\"#groupware\">Groupware<\/a><\/li>\n<li><a href=\"#human-resources-management-hrm\">Human Resources Management (HRM)<\/a><\/li>\n<li><a href=\"#internet-of-things-iot\">Internet of Things (IoT)<\/a><\/li>\n<li><a href=\"#learning-and-courses\">Learning and Courses<\/a><\/li>\n<li><a href=\"#maps-and-global-positioning-system-gps\">Maps and Global Positioning System (GPS)<\/a><\/li>\n<li><a href=\"#media-streaming\">Media Streaming<\/a>\n<ul><li><a href=\"#multimedia-streaming\">Multimedia Streaming<\/a><\/li>\n<li><a href=\"#audio-streaming\">Audio Streaming<\/a><\/li>\n<li><a href=\"#video-streaming\">Video Streaming<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#miscother\">Misc\/Other<\/a><\/li>\n<li><a href=\"#money-budgeting-and-management\">Money, Budgeting and Management<\/a><\/li>\n<li><a href=\"#monitoring\">Monitoring<\/a><\/li>\n<li><a href=\"#note-taking-and-editors\">Note-taking and Editors<\/a><\/li>\n<li><a href=\"#office-suites\">Office Suites<\/a><\/li>\n<li><a href=\"#password-managers\">Password Managers<\/a><\/li>\n<li><a href=\"#pastebins\">Pastebins<\/a><\/li>\n<li><a href=\"#personal-dashboards\">Personal Dashboards<\/a><\/li>\n<li><a href=\"#photo-and-video-galleries\">Photo and Video Galleries<\/a><\/li>\n<li><a href=\"#polls-and-events\">Polls and Events<\/a><\/li>\n<li><a href=\"#proxy\">Proxy<\/a><\/li>\n<li><a href=\"#read-it-later-lists\">Read it Later Lists<\/a><\/li>\n<li><a href=\"#resource-planning\">Resource Planning<\/a>\n<ul><li><a href=\"#enterprise-resource-planning\">Enterprise Resource Planning<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#search-engines\">Search Engines<\/a><\/li>\n<li><a href=\"#software-development\">Software Development<\/a>\n<ul><li><a href=\"#project-management\">Project Management<\/a><\/li>\n<li><a href=\"#bug-trackers\">Bug Trackers<\/a><\/li>\n<li><a href=\"#idetools\">IDE\/Tools<\/a><\/li>\n<li><a href=\"#continuous-integration\">Continuous Integration<\/a><\/li>\n<li><a href=\"#faas-serverless\">FaaS\/Serverless<\/a><\/li>\n<li><a href=\"#api-management\">API Management<\/a><\/li>\n<li><a href=\"#documentation-generators\">Documentation Generators<\/a><\/li>\n<li><a href=\"#localization\">Localization<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#static-site-generators\">Static site generators<\/a><\/li>\n<li><a href=\"#task-managementto-do-lists\">Task management\/To-do lists<\/a><\/li>\n<li><a href=\"#ticketing\">Ticketing<\/a><\/li>\n<li><a href=\"#url-shorteners\">URL Shorteners<\/a><\/li>\n<li><a href=\"#vpn\">VPN<\/a><\/li>\n<li><a href=\"#web-servers\">Web servers<\/a><\/li>\n<li><a href=\"#wikis\">Wikis<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#self-hosting-solutions\">Self-hosting Solutions<\/a><\/li>\n<li><a href=\"#list-of-licenses\">List of Licenses<\/a><\/li>\n<li><a href=\"#external-links\">External links<\/a><\/li>\n<li><a href=\"#contributing\">Contributing<\/a><\/li>\n<li><a href=\"#license\">License<\/a><\/li>\n<\/ul><hr \/><h2><a id=\"user-content-analytics\" class=\"anchor\" aria-hidden=\"true\" href=\"#analytics\"><\/a>Analytics<\/h2>\n<p>For personal analytics\/dashboards, see <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted#personal-dashboards\">Personal Dashboards<\/a><\/p>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em>Web Analytics<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.awstats.org\/\" rel=\"nofollow\">AWStats<\/a> - Generates web, streaming, ftp or mail server statistics graphically. (<a href=\"https:\/\/github.com\/eldy\/awstats\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/count.ly\" rel=\"nofollow\">Countly<\/a> - Real time mobile and web analytics, crash reporting and push notifications platform. (<a href=\"https:\/\/github.com\/countly\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/druid.io\/\" rel=\"nofollow\">Druid<\/a> - Distributed, column-oriented, real-time analytics data store. (<a href=\"https:\/\/github.com\/druid-io\/druid\/\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/usefathom.com\" rel=\"nofollow\">Fathom Analytics<\/a> - Simple &amp; trustworthy website analytics. (<a href=\"https:\/\/github.com\/usefathom\/fathom\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/goaccess.io\/\" rel=\"nofollow\">GoAccess<\/a> - Real-time web log analyzer and interactive viewer that runs in a terminal. (<a href=\"https:\/\/github.com\/allinurl\/goaccess\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/hastic.io\" rel=\"nofollow\">Hastic<\/a> - Hackable time series pattern recognition tool with UI for Grafana. (<a href=\"https:\/\/github.com\/hastic\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python\/Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/matomo.org\/\" rel=\"nofollow\">Matomo<\/a> - Leading open-source analytics platform that gives you more than just powerful analytics, formerly known as Piwik. (<a href=\"https:\/\/github.com\/matomo-org\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.openwebanalytics.com\/\" rel=\"nofollow\">Open Web Analytics<\/a> - Google Analytics and Piwik alternative. (<a href=\"https:\/\/github.com\/padams\/Open-Web-Analytics\/\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/rakam.io\/\" rel=\"nofollow\">Rakam<\/a> - Custom analytics platform that allows you to create your own analytics services. Integrate with any data source (web, mobile, IoT etc.), analyze data with SQL and create dashboards. (<a href=\"https:\/\/github.com\/rakam-io\/rakam\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/serposcope.serphacker.com\/\" rel=\"nofollow\">Serposcope<\/a> - Serposcope is a free and open-source rank tracker to monitor websites ranking in Google and improve your SEO performances. (<a href=\"https:\/\/github.com\/serphacker\/serposcope\">Source Code<\/a>) <code>MIT<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/snowplowanalytics.com\/\" rel=\"nofollow\">Snowplow<\/a> - Have every single event, from your websites, mobile apps, desktop applications and server-side systems, stored in your own data warehouse and available to action in real-time. (<a href=\"https:\/\/github.com\/snowplow\/\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Scala<\/code><\/li>\n<li><a href=\"https:\/\/suet.co\/\" rel=\"nofollow\">Suet<\/a> <code>\u26a0<\/code> - Detailed analytics and reporting for your Mailgun transactional emails. (<a href=\"https:\/\/github.com\/kehers\/suet\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><p><em>Business Intelligence<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.metabase.com\/\" rel=\"nofollow\">Metabase<\/a> - Simple Dashboarding and GUI Query tool, Nightly Emails and Slack Integration w\/ PostgreSQL, MySQL, Redshift and other DBs. (<a href=\"https:\/\/github.com\/metabase\/metabase\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Clojure<\/code><\/li>\n<li><a href=\"http:\/\/redash.io\" rel=\"nofollow\">Redash<\/a> - connect to over 18 types of databases (SQL and \"NoSQL\"), query your data, visualize it and create dashboards. Everything has a URL that can be shared. Slack and HipChat integration. (<a href=\"https:\/\/demo.redash.io\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/getredash\/redash\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/superset.apache.org\/\" rel=\"nofollow\">Superset<\/a> - Modern, enterprise-ready business intelligence web application. (<a href=\"https:\/\/github.com\/apache\/incubator-superset\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-archiving-and-digital-preservation-dp\" class=\"anchor\" aria-hidden=\"true\" href=\"#archiving-and-digital-preservation-dp\"><\/a>Archiving and Digital Preservation (DP)<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>Some <a href=\"#content-management-systems-cms\">Content Management System<\/a> solutions also feature archiving and digital preservation.<\/p>\n<ul><li><a href=\"https:\/\/www.accesstomemory.org\/\" rel=\"nofollow\">Access to Memory (AtoM)<\/a> - Web-based, open source application for standards-based archival description and access in a multilingual, multi-repository environment. (<a href=\"https:\/\/demo.accesstomemory.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/artefactual\/atom\">Source Code<\/a>) <code>PHP<\/code> <code>AGPL-3.0-only<\/code><\/li>\n<li><a href=\"https:\/\/www.archivematica.org\/\" rel=\"nofollow\">Archivematica<\/a> - Mature digital preservation system designed to maintain standards-based, long-term access to collections of digital objects. (<a href=\"http:\/\/sandbox.archivematica.org\/administration\/accounts\/login\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/artefactual\/archivematica\">Source Code<\/a>) <code>Python<\/code> <code>AGPL-3.0-only<\/code><\/li>\n<li><a href=\"https:\/\/archivesspace.org\/\" rel=\"nofollow\">ArchivesSpace<\/a> - Archives information management application for managing and providing Web access to archives, manuscripts and digital objects. (<a href=\"https:\/\/archivesspace.org\/application\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/archivesspace\/archivesspace\">Source Code<\/a>) <code>Ruby<\/code> <code>ECL-2.0<\/code><\/li>\n<li><a href=\"https:\/\/pirate.github.io\/bookmark-archiver\" rel=\"nofollow\">Bookmark Archiver<\/a> - Self-hosted \"wayback machine\" that creates HTML &amp; screenshot archives of sites from your bookmarks, browsing history, RSS feeds, or other sources. (<a href=\"https:\/\/archive.sweeting.me\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/pirate\/bookmark-archiver\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/collectiveaccess.org\/\" rel=\"nofollow\">Collective Access: Providence<\/a> - Highly configurable Web-based framework for management, description, and discovery of digital and physical collections supporting a variety of metadata standards, data types, and media formats. (<a href=\"https:\/\/github.com\/collectiveaccess\/providence\">Source Code<\/a>) <code>PHP<\/code> <code>GPL-3.0-only<\/code><\/li>\n<\/ul><h2><a id=\"user-content-automation\" class=\"anchor\" aria-hidden=\"true\" href=\"#automation\"><\/a>Automation<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/www.alltubedownload.net\" rel=\"nofollow\">Alltube<\/a> - Web interface for <a href=\"https:\/\/github.com\/rg3\/youtube-dl\">youtube-dl<\/a>, a program to download videos and audio from <a href=\"https:\/\/rg3.github.io\/youtube-dl\/supportedsites.html\" rel=\"nofollow\">more than 100 websites<\/a>. (<a href=\"http:\/\/www.alltubedownload.net\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Rudloff\/alltube\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/amiunique.org\/\" rel=\"nofollow\">AmIUnique<\/a> - Learn how identifiable you are on the Internet (browser fingerprinting tool). (<a href=\"https:\/\/github.com\/DIVERSIFY-project\/amiunique\">Source Code<\/a>) <code>MIT<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/muesli\/beehive\">Beehive<\/a> - Flexible event and agent system, which allows you to create your own agents that perform automated tasks triggered by events and filters. <code>AGPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.campaignchain.com\/\" rel=\"nofollow\">CampaignChain<\/a> <code>\u26a0<\/code> - A marketing hub allowing to schedule posts on social media (a la Hootsuite+Buffer) and combine them with newsletter campaigns. (<a href=\"https:\/\/github.com\/CampaignChain\/campaignchain\/tree\/1.0\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/couchpota.to\/\" rel=\"nofollow\">CouchPotato<\/a> - CouchPotato is an automatic Video Library Manager for Movies. Automatic torrent\/nzb searching, downloading, and processing at the qualities you want. (<a href=\"https:\/\/github.com\/CouchPotato\/CouchPotatoServer\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/guptachetan1997\/Episodes\">Episodes<\/a> - <code>\u26a0<\/code> Self Hosted TV show Episode tracker and recommender built using django, bootstrap4. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/cristoper\/feedmixer\">feedmixer<\/a> - FeedMixer is a WSGI (Python3) micro web service which takes a list of feed URLs and returns a new feed consisting of the most recent n entries from each given feed(Returns Atom, RSS, or JSON). (<a href=\"https:\/\/mretc.net\/feedmixer\/json?f=http:\/\/hnrss.org\/newest&amp;f=http:\/\/americancynic.net\/atom.xml&amp;n=1\" rel=\"nofollow\">Demo<\/a>) <code>WTFPL<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/fhem.de\/fhem.html\" rel=\"nofollow\">FHEM<\/a> - FHEM is used to automate common tasks in the household like switching lamps and heating. It can also be used to log events like temperature or power consumption. You can control it via web or smartphone frontends, telnet or TCP\/IP directly. (<a href=\"https:\/\/svn.fhem.de\/trac\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/gekko.wizb.it\/\" rel=\"nofollow\">Gekko<\/a> - Gekko is a Bitcoin TA trading and backtesting bot which support multiple exchanges and cryptocurrencies. (<a href=\"https:\/\/github.com\/askmike\/gekko\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/rembo10\/headphones\">Headphones<\/a> - Automated music downloader for NZB and Torrent, written in Python. It supports SABnzbd, NZBget, Transmission, \u00b5Torrent, Deluge and Blackhole. <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/healthchecks.io\/\" rel=\"nofollow\">Healthchecks<\/a> - Django app which listens for pings and sends alerts when pings are late. (<a href=\"https:\/\/github.com\/healthchecks\/healthchecks\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/home-assistant.io\/\" rel=\"nofollow\">Home Assistant<\/a> - Open-source home automation platform. (<a href=\"https:\/\/home-assistant.io\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/home-assistant\/home-assistant\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Binnette\/homebank-converter\">homebank-converter<\/a> - Web app to convert an export bank file to compatible Homebank csv. (<a href=\"http:\/\/binnette.github.io\/homebank-converter\/\" rel=\"nofollow\">Demo<\/a>) <code>AGPL-3.0<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/cantino\/huginn\">Huginn<\/a> - Allows you to build agents that monitor and act on your behalf. <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/http2pic.haschek.at\/\" rel=\"nofollow\">Http2pic<\/a> - Website screenshots\/renderer. It uses the wkhtmltox to render websites with various options. (<a href=\"https:\/\/github.com\/chrisiaut\/http2pic\">Source Code<\/a>) <code>Apache 2.0<\/code> <code>PHP\/Javascript<\/code><\/li>\n<li><a href=\"https:\/\/kibitzr.github.io\" rel=\"nofollow\">Kibitzr<\/a> - Lightweight personal web assistant with powerful integrations. (<a href=\"https:\/\/github.com\/kibitzr\/kibitzr\/\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/DobyTang\/LazyLibrarian\">LazyLibrarian<\/a> <code>\u26a0<\/code> - LazyLibrarian is a program to follow authors and grab metadata for all your digital reading needs. It uses a combination of Goodreads Librarything and optionally GoogleBooks as sources for author info and book info. <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/lidarr.audio\/\" rel=\"nofollow\">Lidarr<\/a> - Lidarr is a music collection manager for Usenet and BitTorrent users. (<a href=\"https:\/\/github.com\/Lidarr\/Lidarr\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pymedusa\/SickRage\">Medusa<\/a> - Automatic Video Library Manager for TV Shows. It watches for new episodes of your favorite shows, and when they are posted it does its magic. <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/nodered.org\/\" rel=\"nofollow\">Node RED<\/a> - Browser-based flow editor that helps you wiring hardware devices, APIs and online services to create IoT solutions. (<a href=\"https:\/\/github.com\/node-red\/node-red\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.openhab.org\" rel=\"nofollow\">openHAB<\/a> - Vendor and technology agnostic open source software for home automation. (<a href=\"https:\/\/github.com\/openhab\/openhab\">Source Code<\/a>) <code>EPL-1.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/taroved\/pol\">PolitePol<\/a> - Online tool for creation of RSS feeds for any web page. (<a href=\"http:\/\/politepol.com\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/poffer.gabinaureche.com\" rel=\"nofollow\">Poffer<\/a> <code>\u26a0<\/code> - Tool that makes it easier to share the content you like thanks to Pocket+Buffer. (<a href=\"https:\/\/github.com\/Zhouzi\/Poffer\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/pyload.net\/\" rel=\"nofollow\">pyLoad<\/a> - Lightweight, customizable and remotely manageable downloader for 1-click-hosting sites like rapidshare.com or uploaded.to. (<a href=\"https:\/\/github.com\/pyload\/pyload\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/radarr.video\/\" rel=\"nofollow\">Radarr<\/a> - Radarr is an independent fork of Sonarr reworked for automatically downloading movies via Usenet and BitTorrent, \u00e0 la Couchpotato. (<a href=\"https:\/\/github.com\/Radarr\/Radarr\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/sebsauvage\/rss-bridge\">RSS-Bridge<\/a> - rss-bridge is a PHP project capable of generating ATOM feeds for websites which don't have one. <code>Public domain<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/taophp\/rss-merger\">RSS Merger<\/a> - PHP script which will take multiple RSS \/ Atom feeds as input and merge them into a single RSS feed. <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/sickrage.github.io\/\" rel=\"nofollow\">SickRage<\/a> - SickRage is an automatic Video Library Manager for TV Shows. Automatic torrent\/nzb searching, downloading, and processing at the qualities you want. (<a href=\"https:\/\/github.com\/SickRage\/SickRage\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/sonarr.tv\/\" rel=\"nofollow\">Sonarr<\/a> - Automatic TV Shows downloader and manager for Usenet and BitTorrent. It can grab, sort and rename new episodes and automatically upgrade the quality of files already downloaded when a better quality format becomes available. (<a href=\"https:\/\/github.com\/Sonarr\/Sonarr\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/trigger-happy.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">TriggerHappy<\/a> - Open source clone of IFTTT, a bridge between your internet services. (<a href=\"https:\/\/github.com\/foxmask\/django-th\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ziahamza\/webui-aria2\">WebUI-aria2<\/a> - Interface to interact with the aria2 downloader. Very simple to use, just download and open index.html in any web browser. (<a href=\"http:\/\/ziahamza.github.io\/webui-aria2\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jchampemont\/WTFDYUM\">WTFDYUM<\/a> <code>\u26a0<\/code> - Why The Fuck Did You Unfollow Me - Find out who stops following you on Twitter. (<a href=\"https:\/\/www.wtfdyu.me\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/jchampemont\/WTFDYUM\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/carlos8f\/zenbot\">Zenbot 3<\/a> - Zenbot is a lightweight, extendable, artificially intelligent trading bot for Bitcoin, Ether, Litecoin, and more. <code>MIT<\/code> <code>Node.js<\/code><\/li>\n<\/ul><h2><a id=\"user-content-blogging-platforms\" class=\"anchor\" aria-hidden=\"true\" href=\"#blogging-platforms\"><\/a>Blogging Platforms<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See also <a href=\"#static-site-generators\">Static Site Generators<\/a>, <a href=\"#content-management-systems-cms\">Content Management Systems<\/a> and <a href=\"http:\/\/www.weblogmatrix.org\/\" rel=\"nofollow\">WeblogMatrix<\/a><\/p>\n<ul><li><a href=\"https:\/\/anchorcms.com\/\" rel=\"nofollow\">Anchor CMS<\/a> - Free, lightweight, faster-than-a-bullet, simple blogging system, made for art\u2013directed posts. (<a href=\"https:\/\/github.com\/anchorcms\/anchor-cms\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/antville.org\" rel=\"nofollow\">Antville<\/a> - Free, open source project aimed at the development of a high performance, feature rich weblog hosting software. (<a href=\"https:\/\/github.com\/antville\/antville\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/lehollandaisvolant.net\/blogotext\/\" rel=\"nofollow\">Blogotext<\/a> - Free blog-engine written in PHP and using SQLite. This offers you both an unmatched simplicity during installation and great performances. (<a href=\"https:\/\/github.com\/timovn\/blogotext\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.bludit.com\/\" rel=\"nofollow\">Bludit<\/a> <code>\u26a0<\/code> - Simple application to build a site or blog in seconds. Bludit uses flat-files (text files in JSON format) to store posts and pages. (<a href=\"https:\/\/demo.bludit.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/dignajar\/bludit\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/chyrplite.net\" rel=\"nofollow\">Chyrp Lite<\/a> - Extra-awesome, extra-lightweight blog engine. (<a href=\"https:\/\/github.com\/xenocrat\/chyrp-lite\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/dante-stories.herokuapp.com\/\" rel=\"nofollow\">Dante Stories<\/a> - A self hosted Medium platform built with Ruby on Rails. (<a href=\"https:\/\/github.com\/michelson\/dante-stories\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/dotclear.org\/\" rel=\"nofollow\">Dotclear<\/a> - Take control over your blog. (<a href=\"https:\/\/hg.dotclear.org\/dotclear\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/formtools.org\/\" rel=\"nofollow\">Formtools<\/a> - Powerful, flexible, free and open source PHP\/MySQL script to manage your forms and data. (<a href=\"https:\/\/github.com\/formtools\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/ghost.org\/\" rel=\"nofollow\">Ghost<\/a> - Just a blogging platform. (<a href=\"https:\/\/github.com\/TryGhost\/Ghost\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/hexo.io\/\" rel=\"nofollow\">Hexo<\/a> - Fast, simple and powerful blog framework, powered by Node.js. (<a href=\"https:\/\/github.com\/hexojs\/hexo\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/hotglue.me\/\" rel=\"nofollow\">Hotglue<\/a> - Freehand CMS which allows to construct websites directly in a web-browser. It uses flat files for storage and provides an intuitive GUI. (<a href=\"https:\/\/hotglue.me\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/k0a1a\/hotglue2\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.htmly.com\/\" rel=\"nofollow\">htmly<\/a> - Databaseless Blogging Platform (Flat-File Blog). (<a href=\"https:\/\/www.htmly.com\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/danpros\/htmly\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/withknown.com\/\" rel=\"nofollow\">Known<\/a> - Single website for all your content. (<a href=\"https:\/\/github.com\/idno\/idno\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/noddity.com\/\" rel=\"nofollow\">Noddity<\/a> - It's a blog, it's a wiki, it's a fast CMS. (<a href=\"https:\/\/github.com\/TehShrike\/noddity\">Source Code<\/a>) <code>WTFPL<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/plume-org.github.io\/Plume\/\" rel=\"nofollow\">Plume<\/a> - Federated blogging engine, based on ActivityPub. (<a href=\"https:\/\/baptiste.gelez.xyz\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Plume-org\/Plume\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/www.pluxml.org\/\" rel=\"nofollow\">PluXml<\/a> - XML-based blog\/CMS platform. (<a href=\"https:\/\/github.com\/pluxml\/PluXml\">Source Code<\/a>) <code>GPL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.postleaf.org\/\" rel=\"nofollow\">Postleaf<\/a> - Open source blogging platform with inline editing, handlebar templates, and a beautiful user interface. (<a href=\"https:\/\/github.com\/Postleaf\/postleaf\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/b3log.org\/\" rel=\"nofollow\">Solo<\/a> - Blogging system written in Java, feel free to create your or your team own blog. (<a href=\"http:\/\/88250.b3log.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/b3log\/solo\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/wordpress.org\/\" rel=\"nofollow\">WordPress<\/a> - Create a beautiful website or blog. (<a href=\"https:\/\/github.com\/WordPress\/WordPress\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-bookmarks-and-link-sharing\" class=\"anchor\" aria-hidden=\"true\" href=\"#bookmarks-and-link-sharing\"><\/a>Bookmarks and Link Sharing<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/bookieio\/Bookie\">Bookie<\/a> - Python based delicious.com replacement. <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/dyu\/bookmarks\">dyu\/bookmarks<\/a> - Single-threaded\/process bookmark app powered by leveldb and uWebSockets. Supports importing from Delicious and Chrome. (<a href=\"https:\/\/dyuproject.com\/bookmarks\/\" rel=\"nofollow\">Demo<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/docs.services.mozilla.com\/howtos\/run-fxa.html\" rel=\"nofollow\">Firefox Auth Server<\/a> - This project implements the core server-side API for Firefox Accounts. (<a href=\"https:\/\/github.com\/mozilla\/fxa-auth-server\/\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>Nodejs<\/code>\n<ul><li><a href=\"https:\/\/docs.services.mozilla.com\/howtos\/run-fxa.html\" rel=\"nofollow\">Firefox Content Server<\/a> - Static server that hosts Firefox Account sign up, sign in, email verification, etc. flows. (<a href=\"https:\/\/github.com\/mozilla\/fxa-content-server\/\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/docs.services.mozilla.com\/howtos\/run-sync-1.5.html\" rel=\"nofollow\">Firefox Sync Server<\/a> - Sync Firefox bookmarks, passwords, history, tabs, preferences. (<a href=\"https:\/\/github.com\/mozilla-services\/syncserver\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>Python<\/code><\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/geekmarks.dmitryfrank.com\/\" rel=\"nofollow\">Geekmarks<\/a> - Personal bookmarking service focused on speed and organization using hierarchical tags. (<a href=\"https:\/\/github.com\/dimonomid\/geekmarks\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/prologic\/golinks\">golinks<\/a> - Web application that allows you to create smart bookmarks, commands and aliases by pointing your web browser's default search engine at a running instance. Similar to bunny1 or yubnub. (<a href=\"https:\/\/search.mills.io\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/lobste.rs\" rel=\"nofollow\">Lobsters<\/a> - Run your own link aggregation site. (<a href=\"https:\/\/github.com\/jcs\/lobsters\">Source Code<\/a>) <code>BSD<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/nofussbm.herokuapp.com\/signup.html\" rel=\"nofollow\">No Fuss Bookmarks<\/a> - Very simple software and service to store bookmarks especially designed for hackers (that don't need fancy interfaces, but nice API). (<a href=\"https:\/\/github.com\/mapio\/nofussbm\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/getpinry.com\/\" rel=\"nofollow\">Pinry<\/a> - The tiling image board system for people who want to save, tag, and share images, videos, and webpages. (<a href=\"https:\/\/github.com\/pinry\/pinry\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/semanticscuttle.sourceforge.net\/\" rel=\"nofollow\">SemanticScuttle<\/a> - SemanticScuttle is a social bookmarking tool experimenting with features like structured tags and collaborative tag descriptions. (<a href=\"https:\/\/sourceforge.net\/p\/semanticscuttle\/code\/ci\/master\/tree\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/shaarli\/Shaarli\">Shaarli<\/a> - Personal, minimalist, super-fast, no-database bookmarking and link sharing platform. (<a href=\"https:\/\/demo.shaarli.org\" rel=\"nofollow\">Demo<\/a>) <code>Zlib<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/RadhiFadlillah\/shiori\">Shiori<\/a> - Simple bookmark manager built with Go. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/plainmade\/unmark\">unmark<\/a> - Open source to do app for links. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/bitbucket.org\/ymarks\/ymarks-server\" rel=\"nofollow\">ymarks<\/a> - Keep your browser's bookmarks synchronized without limiting yourself to one provider. <code>WTFPL<\/code> <code>C<\/code><\/li>\n<\/ul><h2><a id=\"user-content-calendaring-and-contacts-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#calendaring-and-contacts-management\"><\/a>Calendaring and Contacts Management<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>Some <a href=\"#groupware\">Groupware<\/a> solutions also feature calendar\/address book editing and synchronization.<\/p>\n<p>See <a href=\"https:\/\/en.wikipedia.org\/wiki\/Comparison_of_CalDAV_and_CardDAV_implementations\" rel=\"nofollow\">https:\/\/en.wikipedia.org\/wiki\/Comparison_of_CalDAV_and_CardDAV_implementations<\/a><\/p>\n<p><em>CalDAV or CardDAV servers<\/em><\/p>\n<ul><li><a href=\"http:\/\/sabre.io\/baikal\/\" rel=\"nofollow\">Ba\u00efkal<\/a> - Lightweight CalDAV and CardDAV server based on sabre\/dav. (<a href=\"https:\/\/github.com\/sabre-io\/Baikal\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.calendarserver.org\/\" rel=\"nofollow\">CalendarServer<\/a> - Apple, Inc.'s standards-compliant server implementing the CalDAV and CardDAV protocols shipped with macOS Server. (<a href=\"https:\/\/github.com\/apple\/ccs-calendarserver\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/keithp.com\/calypso\/\" rel=\"nofollow\">calypso<\/a> - Python-based CalDAV and CardDAV server, forked from Radicale. (<a href=\"https:\/\/keithp.com\/git\/calypso.git\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.davical.org\/\" rel=\"nofollow\">DAViCal<\/a> - Server for calendar sharing (CalDAV) that uses a PostgreSQL database as a data store. (<a href=\"https:\/\/gitlab.com\/davical-project\/davical\" rel=\"nofollow\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.etesync.com\" rel=\"nofollow\">EteSync Server<\/a> - End-to-end encrypted and journaled personal information server supporting calendar and contact data, offering its own clients. (<a href=\"https:\/\/github.com\/etesync\/server-skeleton\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python\/Django<\/code><\/li>\n<li><a href=\"http:\/\/radicale.org\/\" rel=\"nofollow\">Radicale<\/a> - Simple calendar and contact server with extremely low administrative overhead. (<a href=\"https:\/\/github.com\/Kozea\/Radicale\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/sabre.io\/\" rel=\"nofollow\">SabreDAV<\/a> - Open source CardDAV, CalDAV, and WebDAV framework and server. (<a href=\"https:\/\/github.com\/sabre-io\/dav\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.xandikos.org\/\" rel=\"nofollow\">Xandikos<\/a> - Open source CardDAV, CalDAV and WebDAV server with minimal administrative overhead, backed by a Git repository. (<a href=\"https:\/\/github.com\/jelmer\/xandikos\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<\/ul><p><em>CalDAV or CardDAV clients.<\/em><\/p>\n<ul><li><a href=\"http:\/\/agendav.org\/\" rel=\"nofollow\">AgenDAV<\/a> - Multilanguage CalDAV web client with a rich AJAX interface and shared calendars support. (<a href=\"https:\/\/github.com\/agendav\/agendav\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.davdroid.com\/\" rel=\"nofollow\">DAVDroid<\/a> - Open-source CalDAV\/CardDAV suite and sync app for Android. (<a href=\"https:\/\/gitlab.com\/bitfireAT\/davdroid\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/www.inf-it.com\/open-source\/clients\/infcloud\/\" rel=\"nofollow\">InfCloud<\/a> - Open source CalDAV\/CardDAV web client implementation. (<a href=\"https:\/\/www.inf-it.com\/infcloud\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/www.inf-it.com\/InfCloud_0.13.1.zip\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/www.etesync.com\/faq\/#web-client\" rel=\"nofollow\">EteSync Web<\/a> - EteSync's official Web-based client (i.e., their Web app). (<a href=\"https:\/\/client.etesync.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/etesync\/etesync-web\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>TypeScript<\/code><\/li>\n<\/ul><h2><a id=\"user-content-communication-systems\" class=\"anchor\" aria-hidden=\"true\" href=\"#communication-systems\"><\/a>Communication systems<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<h3><a id=\"user-content-custom-communication-systems\" class=\"anchor\" aria-hidden=\"true\" href=\"#custom-communication-systems\"><\/a>Custom communication systems<\/h3>\n<ul><li><a href=\"https:\/\/broid.ai\" rel=\"nofollow\">Broid<\/a> - Broid enables rich conversations on all messaging channels within a single schema integration using W3C standards. (<a href=\"https:\/\/www.broid.ai\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/broidHQ\/integrations\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/centrifugal\/centrifugo\">Centrifugo<\/a> - Language-agnostic real-time messaging (Websocket or SockJS) server. (<a href=\"https:\/\/github.com\/centrifugal\/centrifugo#demo\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/rafael-santiago\/cherry\">Cherry<\/a> - Tiny webchat server. <code>GPL-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/freenetproject.org\/index.html\" rel=\"nofollow\">Freenet<\/a> - Anonymously share files, browse and publish \"freesites\" (web sites accessible only through Freenet) and chat on forums. (<a href=\"https:\/\/github.com\/freenet\/fred\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/moose-team.github.io\/friends\/\" rel=\"nofollow\">Friends<\/a> - P2P chat powered by the web. (<a href=\"https:\/\/github.com\/moose-team\/friends\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/gnunet.org\/\" rel=\"nofollow\">GNUnet<\/a> - Free software framework for decentralized, peer-to-peer networking. (<a href=\"https:\/\/gnunet.org\/git\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/hawkpost.co\" rel=\"nofollow\">Hawkpost<\/a> - HawkPost is a web app that lets you create unique links that you can share with a person that desires to send you important information but doesn't know how to encrypt it. The message is encrypted in their browser and sent to your email address. (<a href=\"https:\/\/github.com\/whitesmith\/hawkpost\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/hubl.in\/\" rel=\"nofollow\">Hubl.in<\/a> - WebRTC powered video conference, chat and collaborative editor. (<a href=\"https:\/\/github.com\/linagora\/hublin\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/jitsi.org\/Projects\/JitsiMeet\" rel=\"nofollow\">Jitsi Meet<\/a> - Jitsi Meet is an OpenSource (MIT) WebRTC Javascript application that uses Jitsi Videobridge to provide high quality, scalable video conferences. (<a href=\"https:\/\/github.com\/jitsi\/jitsi-meet\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/jitsi.org\/Projects\/JitsiVideobridge\" rel=\"nofollow\">Jitsi Video Bridge<\/a> - WebRTC compatible Selective Forwarding Unit (SFU) that allows for multiuser video communication. (<a href=\"https:\/\/github.com\/jitsi\/jitsi-videobridge\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/getkandan.com\/\" rel=\"nofollow\">Kandan<\/a> - Kandan is an Open Source Alternative to HipChat. (<a href=\"https:\/\/github.com\/kandanapp\/kandan\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/php-kchat\/kchat\">KChat<\/a> - PHP Based Live Chat Aplication. <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/sdelements.github.io\/lets-chat\/\" rel=\"nofollow\">Lets-Chat<\/a> - Self hosted chat suite written in Node. (<a href=\"https:\/\/github.com\/sdelements\/lets-chat\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/livehelperchat.com\/\" rel=\"nofollow\">Live Helper Chat<\/a> - Live Support chat for your website. (<a href=\"https:\/\/github.com\/LiveHelperChat\/livehelperchat\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.mattermost.org\/\" rel=\"nofollow\">Mattermost<\/a> - Open-source, on-prem Slack-alternative. It can be integrated with <a href=\"https:\/\/about.gitlab.com\/\" rel=\"nofollow\">Gitlab<\/a>. (<a href=\"https:\/\/github.com\/mattermost\/platform\">Source Code<\/a>) <code>AGPL-3.0\/Apache<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/dystroy.org\/miaou\/login\" rel=\"nofollow\">MiAOU<\/a> - Multi-room persistent chat server. (<a href=\"https:\/\/github.com\/Canop\/miaou\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/mibew.org\" rel=\"nofollow\">Mibew<\/a> - Mibew Messenger is an open-source live support application written in PHP and MySQL. It enables one-on-one chat assistance in real-time directly from your website. (<a href=\"https:\/\/mibew.org\/demo2\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Mibew\/mibew\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/wiki.mumble.info\/wiki\/Main_Page\" rel=\"nofollow\">Mumble<\/a> - Low-latency, high quality voice\/text chat software. (<a href=\"https:\/\/github.com\/mumble-voip\/mumble\">Source Code<\/a>, <a href=\"https:\/\/wiki.mumble.info\/wiki\/3rd_Party_Applications\" rel=\"nofollow\">Clients<\/a>) <code>BSD<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/IgorAntun\/node-chat\">Node-Chat<\/a> - Not-so-basic open-source chat with admin features. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/pushjet.io\/\" rel=\"nofollow\">Pushjet<\/a> - Self hosted push notification service built with a simple API making sending notifications easy. (<a href=\"https:\/\/github.com\/Pushjet\/Pushjet-Server-Api\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/rallly.co\" rel=\"nofollow\">Rallly<\/a> - Rallly is a free collaborative scheduling service. (<a href=\"https:\/\/github.com\/lukevella\/Rallly\">Source Code<\/a>) <code>CC-BY-SA-4.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/retroshare.org\" rel=\"nofollow\">RetroShare<\/a> - Secured and decentralized communication system. Offers decentralized chat, forums, messaging, file transfer. (<a href=\"https:\/\/github.com\/RetroShare\/RetroShare\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>\u0421++<\/code><\/li>\n<li><a href=\"https:\/\/ring.cx\/\" rel=\"nofollow\">Ring<\/a> - Free software for universal communication which respects freedoms and privacy of its users. (<a href=\"https:\/\/github.com\/savoirfairelinux\/ring-project\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/rocket.chat\/\" rel=\"nofollow\">Rocket.Chat<\/a> - Teamchat solution similar to Gitter.im or Slack. (<a href=\"https:\/\/github.com\/RocketChat\/Rocket.Chat\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/spectrum.im\/\" rel=\"nofollow\">Spectrum 2<\/a> - Spectrum 2 is an open source instant messaging transport.  It allows users to chat together even when they are using different IM networks. (<a href=\"https:\/\/github.com\/hanzz\/spectrum2\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/www.spreed.me\/\" rel=\"nofollow\">Spreed<\/a> - WebRTC audio\/video calls, conferencing server, and web client. (<a href=\"https:\/\/github.com\/strukturag\/spreed-webrtc\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/matrix.org\/docs\/projects\/server\/synapse.html\" rel=\"nofollow\">Synapse<\/a> - Server for <a href=\"https:\/\/matrix.org\/\" rel=\"nofollow\">Matrix<\/a>, an open standard for decentralized persistent communication. (<a href=\"https:\/\/github.com\/matrix-org\/synapse\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code>\n<ul><li><a href=\"http:\/\/matrix.org\/docs\/projects\/client\/matrix-console.html\" rel=\"nofollow\">Matrix Console Web<\/a> - Web client meant to be a showcase of Matrix capabilities, and reference implementation of the Matrix standard. (<a href=\"https:\/\/github.com\/matrix-org\/matrix-angular-sdk\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/riot.im\" rel=\"nofollow\">RIOT<\/a> - Glossy Matrix web client with an emphasis on performance and usability. (<a href=\"https:\/\/github.com\/vector-im\/riot-web\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Javascript<\/code><\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/www.syndie.de\" rel=\"nofollow\">Syndie<\/a> - Syndie is a libre system for operating distributed forums. <code>CC0-1.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/typpo\/textbelt\">TextBelt<\/a> <code>\u26a0<\/code> - Outgoing SMS API that uses carrier-specific gateways to deliver your text messages for free, and without ads. <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/tox.chat\/\" rel=\"nofollow\">Tox<\/a> - Distributed, secure messenger with audio and video chat capabilities. (<a href=\"https:\/\/github.com\/irungentoo\/toxcore\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/blog.trailofbits.com\/2015\/12\/15\/self-hosted-video-chat-with-tuber\/\" rel=\"nofollow\">Tuber<\/a> - Peer-to-peer video chat that works. (<a href=\"https:\/\/github.com\/trailofbits\/tubertc\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/zeronet.io\/\" rel=\"nofollow\">ZeroNet<\/a> <code>\u26a0<\/code> - Open, free, and uncensorable websites, using Bitcoin cryptography and BitTorrent network. (<a href=\"https:\/\/github.com\/HelloZeroNet\/ZeroNet\">Source Code<\/a>) <code>GNU<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/zulip.org\" rel=\"nofollow\">Zulip<\/a> - Zulip is a powerful, open source group chat application. (<a href=\"https:\/\/github.com\/zulip\/zulip\">Source Code<\/a>) <code>Apache\/Other<\/code> <code>Python<\/code><\/li>\n<\/ul><h3><a id=\"user-content-email\" class=\"anchor\" aria-hidden=\"true\" href=\"#email\"><\/a>Email<\/h3>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<h4><a id=\"user-content-complete-solutions\" class=\"anchor\" aria-hidden=\"true\" href=\"#complete-solutions\"><\/a>Complete solutions<\/h4>\n<p><em>Simple deployment of a mail server, e.g. for inexperienced or impatient admins.<\/em><\/p>\n<ul><li><a href=\"https:\/\/github.com\/tomav\/docker-mailserver\">docker-mailserver<\/a> - Fullstack but simple mail server (smtp, imap, antispam, antivirus, etc.). Only configuration files, no SQL database. Keep it simple and versioned. Easy to deploy and upgrade. <code>MIT<\/code> <code>Docker<\/code><\/li>\n<li><a href=\"https:\/\/inboxen.org\" rel=\"nofollow\">Inboxen<\/a> - Inboxen is a service that provides you with an infinite number of unique inboxes. (<a href=\"https:\/\/github.com\/Inboxen\/Inboxen\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/progmaticltd\/homebox\">homebox<\/a> - A suite of Ansible scripts to deploy a fully functional mail server on Debian. Unobtrusive and automatic as much as possible, focusing on stability and security. <code>GPL-3.0<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"http:\/\/www.iredmail.org\/\" rel=\"nofollow\">iRedMail<\/a> - Full-featured mail server solution based on Postfix and Dovecot. (<a href=\"https:\/\/bitbucket.org\/zhb\/iredmail\/commits\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"https:\/\/mailcow.email\/\" rel=\"nofollow\">Mailcow<\/a> - Mail server suite based on Dovecot, Postfix and other open source software, that provides a modern Web UI for administration. (<a href=\"https:\/\/github.com\/andryyy\/mailcow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/mailu.io\/\" rel=\"nofollow\">Mailu<\/a> - Mailu is a simple yet full-featured mail server as a set of Docker images. (<a href=\"https:\/\/github.com\/Mailu\/Mailu\/wiki\/Demo-server\">Demo<\/a>, <a href=\"https:\/\/github.com\/Mailu\/Mailu\">Source Code<\/a>) <code>MIT<\/code> <code>Docker\/Python<\/code><\/li>\n<li><a href=\"https:\/\/mailinabox.email\/\" rel=\"nofollow\">Mail-in-a-Box<\/a> - Turns any Ubuntu server into a fully functional mail server with one command. (<a href=\"https:\/\/github.com\/mail-in-a-box\/mailinabox\">Source Code<\/a>) <code>CC0-1.0<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"http:\/\/modoboa.org\/en\/\" rel=\"nofollow\">Modoboa<\/a> - Modoboa is a mail hosting and management platform including a modern and simplified Web User Interface. (<a href=\"https:\/\/github.com\/tonioo\/modoboa\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.qmailtoaster.com\/\" rel=\"nofollow\">Qmailtoaster<\/a> - Stable, full-featured, easy-to-install mail server based on qmail. (<a href=\"https:\/\/github.com\/QMailToaster\/\">Source Code<\/a>) <code>Multiple<\/code> <code>Linux<\/code><\/li>\n<li><a href=\"https:\/\/gitlab.com\/simple-nixos-mailserver\/nixos-mailserver\" rel=\"nofollow\">Simple NixOS Mailserver<\/a> - A complete mailserver solution leveraging the Nix Ecosystem. <code>GPL-3.0<\/code> <code>Nix<\/code><\/li>\n<\/ul><h4><a id=\"user-content-mail-transfer-agents\" class=\"anchor\" aria-hidden=\"true\" href=\"#mail-transfer-agents\"><\/a>Mail Transfer Agents<\/h4>\n<p><em>MTAs \/ SMTP servers<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.courier-mta.org\/\" rel=\"nofollow\">Courier MTA<\/a> - Fast, scalable, enterprise mail\/groupware server providing ESMTP, IMAP, POP3, webmail, mailing list, basic web-based calendaring and scheduling services. (<a href=\"http:\/\/www.courier-mta.org\/repo.html\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/www.exim.org\/\" rel=\"nofollow\">Exim<\/a> - Message transfer agent (MTA) developed at the University of Cambridge. (<a href=\"http:\/\/git.exim.org\/exim.git\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/haraka.github.io\/\" rel=\"nofollow\">Haraka<\/a> - High-performance, pluginable SMTP server written in Javascript. (<a href=\"https:\/\/github.com\/haraka\/Haraka\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/mailcatcher.me\/\" rel=\"nofollow\">MailCatcher<\/a> - Ruby gem that deploys a simply SMTP MTA gateway that accepts all mail and displays in web interface. Useful for debugging or development. (<a href=\"https:\/\/github.com\/sj26\/mailcatcher\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/m242\/maildrop\">Maildrop<\/a> - Disposable email SMTP server, also useful for development. <code>MIT<\/code> <code>Scala<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/mailhog\/MailHog\">MailHog<\/a> - Small Golang executable which runs an SMTP MTA gateway that accepts all mail and displays in web interface. Useful for debugging or development. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/opensmtpd.org\/\" rel=\"nofollow\">OpenSMTPD<\/a> - Secure SMTP server implementation from the OpenBSD project. (<a href=\"http:\/\/cvsweb.openbsd.org\/cgi-bin\/cvsweb\/src\/usr.sbin\/smtpd\/\" rel=\"nofollow\">Source Code<\/a>) <code>ISC<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/www.postfix.org\/\" rel=\"nofollow\">Postfix<\/a> - Fast, easy to administer, and secure Sendmail replacement. <code>IPL-1.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/www.qmail.org\/top.html\" rel=\"nofollow\">Qmail<\/a> - Secure Sendmail replacement. (<a href=\"https:\/\/sources.debian.net\/src\/netqmail\/1.06-5\/\" rel=\"nofollow\">Source Code<\/a>) <code>CC0-1.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/www.sendmail.com\/sm\/open_source\/\" rel=\"nofollow\">Sendmail<\/a> - Message transfer agent (MTA). <code>Sendmail<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/slimta.org\" rel=\"nofollow\">Slimta<\/a> - Mail Transfer Library built on Python. (<a href=\"https:\/\/github.com\/slimta\/python-slimta\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<\/ul><h4><a id=\"user-content-mail-delivery-agents\" class=\"anchor\" aria-hidden=\"true\" href=\"#mail-delivery-agents\"><\/a>Mail Delivery Agents<\/h4>\n<p><em>MDAs - IMAP\/POP3 software<\/em><\/p>\n<ul><li><a href=\"http:\/\/cyrusimap.org\/\" rel=\"nofollow\">Cyrus IMAP\/POP3<\/a> - Intended to be run on sealed servers, where normal users are not permitted to log in. (<a href=\"https:\/\/github.com\/cyrusimap\/cyrus-imapd\">Source Code<\/a>) <code>BSD<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/www.dovecot.org\/\" rel=\"nofollow\">Dovecot<\/a> - IMAP and POP3 server written primarily with security in mind. (<a href=\"https:\/\/github.com\/dovecot\/core\">Source Code<\/a>) <code>MIT\/LGPL-2.1<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/www.mailpiler.org\/wiki\/start\" rel=\"nofollow\">Piler<\/a> - feature rich open source email archiving solution. (<a href=\"https:\/\/bitbucket.org\/jsuto\/piler\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<\/ul><h4><a id=\"user-content-mailing-lists-and-newsletters\" class=\"anchor\" aria-hidden=\"true\" href=\"#mailing-lists-and-newsletters\"><\/a>Mailing lists and Newsletters<\/h4>\n<p><em>Mailing lists servers and mass mailing software - one message to many recipients.<\/em><\/p>\n<ul><li><a href=\"http:\/\/dadamailproject.com\/\" rel=\"nofollow\">Dada Mail<\/a> - Web-based list management system that can be used for announcement lists and\/or discussion lists. (<a href=\"https:\/\/github.com\/justingit\/dada-mail\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/freeCodeCamp\/mail-for-good\">Mail For Good<\/a> <code>\u26a0<\/code> - Open source email campaign management tool for nonprofits. <code>BSD<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/www.gnu.org\/software\/mailman\/\" rel=\"nofollow\">Mailman<\/a> - The Gnu mailing list server. <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/mailtrain.org\/\" rel=\"nofollow\">Mailtrain<\/a> - self hosted newsletter application built on Node.js (v5+) and MySQL (v5.5+ or MariaDB). (<a href=\"https:\/\/github.com\/andris9\/mailtrain\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/mailyherald.org\/\" rel=\"nofollow\">MailyHerald<\/a> - Self-hosted Mailchimp alternative that you can easily integrate with your site. Helps you send and manage your application mailings. It support email marketing and conducting the daily stream of notifications you send to your users. (<a href=\"https:\/\/github.com\/Sology\/maily_herald\">Source Code<\/a>) <code>LGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.mautic.org\/\" rel=\"nofollow\">Mautic<\/a> - Mautic is marketing automation software (email, social and more). (<a href=\"https:\/\/github.com\/mautic\/mautic\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/phplist.org\" rel=\"nofollow\">phpList<\/a> - Newsletter and email marketing with advanced management of subscribers, bounces, and plugins. (<a href=\"https:\/\/github.com\/phpList\/\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/atech\/postal\">Postal<\/a> - Fully featured open source mail delivery platform for incoming and outgoing e-mail. <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/schleuder.nadir.org\/\" rel=\"nofollow\">Schleuder<\/a> - GPG-enabled mailing list manager with resending-capabilities. (<a href=\"https:\/\/0xacab.org\/schleuder\/schleuder\/tree\/master\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.sympa.org\/\" rel=\"nofollow\">Sympa<\/a> - Mailing list manager. <code>GPL-2.0<\/code> <code>Perl<\/code><\/li>\n<\/ul><h4><a id=\"user-content-webmail-clients\" class=\"anchor\" aria-hidden=\"true\" href=\"#webmail-clients\"><\/a>Webmail clients<\/h4>\n<ul><li><a href=\"http:\/\/cypht.org\/index.html\" rel=\"nofollow\">Cypht<\/a> - Feed reader for your email accounts. (<a href=\"https:\/\/github.com\/jasonmunro\/hm3\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/synox\/disposable-mailbox\">Disposable Mailbox<\/a> - Simple disposable mailbox web-app based on a catch-all IMAP mailbox. (<a href=\"https:\/\/bhadoomail.com\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.horde.org\/apps\/imp\/\" rel=\"nofollow\">IMP<\/a> - HORDE application that provides webmail access to IMAP and POP3 accounts. (<a href=\"http:\/\/demo.horde.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/www.horde.org\/download\/imp\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/mailcare.io\" rel=\"nofollow\">MailCare<\/a> - Open source disposable email address service. (<a href=\"https:\/\/github.com\/mailcare\/mailcare\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.mailpile.is\/\" rel=\"nofollow\">Mailpile<\/a> - Webmail client with search, filtering, encryption features and more. (<a href=\"https:\/\/github.com\/mailpile\/Mailpile\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.rainloop.net\/\" rel=\"nofollow\">RainLoop<\/a> - Simple, modern and fast webmail with IMAP\/SMTP Support and multi accounting. (<a href=\"http:\/\/demo.rainloop.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/RainLoop\/rainloop-webmail\">Source Code<\/a>). <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/roundcube.net\" rel=\"nofollow\">Roundcube<\/a> - Browser-based IMAP client with an application-like user interface. (<a href=\"https:\/\/github.com\/roundcube\/roundcubemail\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/squirrelmail.org\" rel=\"nofollow\">SquirrelMail<\/a> - Another browser-based IMAP client. (<a href=\"https:\/\/sourceforge.net\/p\/squirrelmail\/code\/HEAD\/tree\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.afterlogic.org\/webmail-lite\" rel=\"nofollow\">WebMail Lite<\/a> - Web-based IMAP Mail client. (<a href=\"https:\/\/github.com\/afterlogic\/webmail-lite\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h3><a id=\"user-content-irc\" class=\"anchor\" aria-hidden=\"true\" href=\"#irc\"><\/a>IRC<\/h3>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Internet_Relay_Chat\" rel=\"nofollow\">IRC<\/a> communication software<\/em><\/p>\n<ul><li><a href=\"http:\/\/convos.by\/\" rel=\"nofollow\">Convos<\/a> - Always online web IRC client. (<a href=\"http:\/\/demo.convos.by\" rel=\"nofollow\">Demo<\/a>, <a href=\"http:\/\/github.com\/nordaaker\/convos\">Source Code<\/a>) <code>Artistic-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/khlieng\/dispatch\">Dispatch<\/a> - A self-hosted web IRC client written in Go. (<a href=\"https:\/\/dispatch.khlieng.com\/connect\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/kiwiirc.com\/\" rel=\"nofollow\">Kiwi IRC<\/a> - A responsive web IRC client with theming support. (<a href=\"https:\/\/kiwiirc.com\/nextclient\/\" rel=\"nofollow\">Demo<\/a>), (<a href=\"https:\/\/github.com\/kiwiirc\/kiwiirc\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/thelounge.github.io\/\" rel=\"nofollow\">The Lounge<\/a> - A self-hosted web IRC client. (<a href=\"https:\/\/demo.thelounge.chat\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/thelounge\/lounge\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/quassel-irc.org\/\" rel=\"nofollow\">Quassel IRC<\/a> - distributed IRC client, meaning that one (or multiple) client(s) can attach to and detach from a central core. (<a href=\"https:\/\/github.com\/quassel\/quassel\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/robustirc.net\/\" rel=\"nofollow\">Robust IRC<\/a> - RobustIRC is IRC without netsplits. Distributed IRC server, based on RobustSession protocol. (<a href=\"https:\/\/github.com\/robustirc\/robustirc\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/weechat.org\/\" rel=\"nofollow\">Weechat<\/a> - Fast, light and extensible chat client. <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/wiki.znc.in\/ZNC\" rel=\"nofollow\">ZNC<\/a> - Advanced IRC bouncer. (<a href=\"https:\/\/github.com\/znc\/znc\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>C++<\/code><\/li>\n<\/ul><h3><a id=\"user-content-sip\" class=\"anchor\" aria-hidden=\"true\" href=\"#sip\"><\/a>SIP<\/h3>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Session_Initiation_Protocol\" rel=\"nofollow\">SIP<\/a> telephony software<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.asterisk.org\/\" rel=\"nofollow\">Asterisk<\/a> - Easy to use but advanced IP PBX system, VoIP gateway and conference server. <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/freeswitch.org\/\" rel=\"nofollow\">FreeSWITCH<\/a> - Scalable open source cross-platform telephony platform. (<a href=\"https:\/\/freeswitch.org\/stash\/projects\/FS\/repos\/freeswitch\/browse\" rel=\"nofollow\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/www.sipcapture.org\/\" rel=\"nofollow\">Homer<\/a> - Troubleshooting and monitoring VoIP calls. (<a href=\"https:\/\/github.com\/sipcapture\/homer\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Angular\/C<\/code><\/li>\n<li><a href=\"http:\/\/www.kamailio.org\/w\/\" rel=\"nofollow\">Kamailio<\/a> - Modular SIP server (registrar\/proxy\/router\/etc). (<a href=\"https:\/\/github.com\/kamailio\/kamailio\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/dev.guardianproject.info\/projects\/ostel\/wiki\/Server_Documentation\" rel=\"nofollow\">Ostel<\/a> - Secure SIP telephony setup with ZRTP encryption. <code>GPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/www.sip3.io\/\" rel=\"nofollow\">Tapir<\/a> - Troubleshooting and real-time monitoring of VoIP-based systems. (<a href=\"https:\/\/github.com\/sip3io\/tapir\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java\/Kotlin<\/code><\/li>\n<\/ul><h3><a id=\"user-content-ipbx\" class=\"anchor\" aria-hidden=\"true\" href=\"#ipbx\"><\/a>IPBX<\/h3>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em><a href=\"https:\/\/en.wikipedia.org\/wiki\/IP_PBX\" rel=\"nofollow\">IPBX<\/a> telephony software<\/em><\/p>\n<ul><li><a href=\"http:\/\/www.freepbx.org\" rel=\"nofollow\">Freepbx<\/a> - Web-based open source GUI that controls and manages Asterisk. (<a href=\"http:\/\/git.freepbx.org\/projects\/FREEPBX\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.fusionpbx.com\/\" rel=\"nofollow\">FusionPBX<\/a> - Open source project that provides a customizable and flexible web interface to the very powerful and highly scalable multi-platform voice switch called FreeSWITCH. (<a href=\"https:\/\/github.com\/fusionpbx\/fusionpbx\">Source Code<\/a>) <code>MPL-1.1<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/2600hz.org\/\" rel=\"nofollow\">Kazoo<\/a> - KAZOO is an open-source, highly scalable software platform designed to provide carrier-grade VoIP switch functions and features. (<a href=\"https:\/\/github.com\/2600hz\/KAZOO\">Source Code<\/a>) <code>MPL-1.1<\/code> <code>Erlang<\/code><\/li>\n<li><a href=\"http:\/\/wazo.community\/\" rel=\"nofollow\">Wazo<\/a> - Full-featured IPBX solution built atop Asterisk with integrated Web administration interface and REST-ful API. (<a href=\"https:\/\/github.com\/wazo-pbx\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python\/PHP<\/code><\/li>\n<\/ul><h3><a id=\"user-content-social-networks-and-forums\" class=\"anchor\" aria-hidden=\"true\" href=\"#social-networks-and-forums\"><\/a>Social Networks and Forums<\/h3>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/abilian\/abilian-sbe\">Abilian SBE<\/a> - Open Source Collaboration and Social Networking framework and platform. <code>LGPL-2.1<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.getanahita.com\/\" rel=\"nofollow\">Anahita<\/a> - Open Source Social Networking Framework and Platform. (<a href=\"https:\/\/github.com\/anahitasocial\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/bbpress.org\/\" rel=\"nofollow\">bbPress<\/a> - bbPress is forum software with a twist from the creators of WordPress. Easily setup discussion forums inside your WordPress.org powered site. (<a href=\"https:\/\/bbpress.trac.wordpress.org\/browser\/trunk\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/trybootcamp.vitorfs.com\" rel=\"nofollow\">Bootcamp<\/a> - Enterprise social network. (<a href=\"https:\/\/github.com\/vitorfs\/bootcamp\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/buddycloud.com\/\" rel=\"nofollow\">Buddycloud<\/a> - Tools, libraries, services and a community to build user-to-user, group and social messaging into your app. Saves time. Scales up. Supports you. (<a href=\"https:\/\/github.com\/buddycloud\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/buddypress.org\/about\/\" rel=\"nofollow\">BuddyPress<\/a> - Powerful plugin that takes your WordPress.org powered site beyond the blog with social-network features like user profiles, activity streams, user groups, and more. (<a href=\"https:\/\/buddypress.svn.wordpress.org\/trunk\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/daveajones\/cartulary\">cartulary<\/a> - RSS reader, readability tool, article archiver, microblogger, social graph manager and reading list manager. <code>CDDL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/diasporafoundation.org\/\" rel=\"nofollow\">diaspora*<\/a> - Distributed social networking server. (<a href=\"https:\/\/podupti.me\/go.php\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/diaspora\/diaspora\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/www.discourse.org\/\" rel=\"nofollow\">Discourse<\/a> - Advanced forum \/ community solution based on Ruby and JS. (<a href=\"https:\/\/try.discourse.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/discourse\/discourse\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/dyu\/comments\/\">dyu\/comments<\/a> - Real-time, markdown-enabled comment engine powered by leveldb. (<a href=\"https:\/\/dyu.github.io\/comments\/real-time\/\" rel=\"nofollow\">Demo<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/elgg.org\/\" rel=\"nofollow\">Elgg<\/a> - Powerful open source social networking engine. (<a href=\"https:\/\/github.com\/Elgg\/Elgg\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/flarum.org\" rel=\"nofollow\">Flarum<\/a> - Delightfully simple forums. Flarum is the next-generation forum software that makes online discussion fun again. (<a href=\"https:\/\/github.com\/flarum\/flarum\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/flaskbb.org\/\" rel=\"nofollow\">flaskbb<\/a> - FlaskBB is forum software written in Python using the microframework Flask. You can easily create new topics, posts and send other users private messages. It also includes basic administration and moderation tools. (<a href=\"https:\/\/github.com\/sh4nks\/flaskbb\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/fluxbb.org\/\" rel=\"nofollow\">FluxBB<\/a> - Fast, light, user-friendly forum software for your website. (<a href=\"https:\/\/github.com\/fluxbb\/fluxbb\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/friendi.ca\/\" rel=\"nofollow\">Friendica<\/a> - Social Communication Server. (<a href=\"https:\/\/github.com\/friendica\/friendica\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/gnu.io\/social\/\" rel=\"nofollow\">GNU social<\/a> - Social communication software for both public and private communications. (<a href=\"https:\/\/git.gnu.io\/gnu\/gnu-social\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/hubzilla.org\" rel=\"nofollow\">Hubzilla<\/a> - Decentralized identity, privacy, publishing, sharing, cloud storage, and communications\/social platform. (<a href=\"https:\/\/framagit.org\/hubzilla\/core\" rel=\"nofollow\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.humhub.org\/\" rel=\"nofollow\">HumHub<\/a> - Flexible kit for private social networks. (<a href=\"https:\/\/github.com\/humhub\/humhub\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/posativ.org\/isso\/\" rel=\"nofollow\">Isso<\/a> - Lightweight commenting server written in Python and Javascript. It aims to be a drop-in replacement for Disqus. (<a href=\"https:\/\/github.com\/posativ\/isso\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.loomio.org\/\" rel=\"nofollow\">Loomio<\/a> - Loomio is a collaborative decision-making tool that makes it easy for anyone to participate in decisions which affect them. (<a href=\"https:\/\/github.com\/loomio\/loomio\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/joinmastodon.org\/\" rel=\"nofollow\">Mastodon<\/a> - Federated microblogging server, an alternative to GNU social. (<a href=\"https:\/\/github.com\/tootsuite\/mastodon\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/movim.eu\/\" rel=\"nofollow\">Movim<\/a> - Modern, federated social network based on XMPP, with a fully featured group-chat, subscriptions and microblogging. (<a href=\"https:\/\/github.com\/movim\/movim\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.mybb.com\/\" rel=\"nofollow\">MyBB<\/a> - Free, extensible forum software package. (<a href=\"https:\/\/github.com\/mybb\/mybb\">Source Code<\/a>) <code>LGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/newebe.org\/\" rel=\"nofollow\">Newebe<\/a> - Distributed Social Network. (<a href=\"https:\/\/github.com\/gelnior\/newebe\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/nodebb.org\/\" rel=\"nofollow\">NodeBB<\/a> - Node.js based forum software built for the modern web. (<a href=\"https:\/\/github.com\/NodeBB\/NodeBB\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/pleroma.social\" rel=\"nofollow\">Pleroma<\/a> - Federated microblogging server, Mastodon, GNU social, &amp; ActivityPub compatible. (<a href=\"https:\/\/git.pleroma.social\/pleroma\/pleroma\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Elixir<\/code><\/li>\n<li><a href=\"http:\/\/www.goodoldweb.com\/\" rel=\"nofollow\">Orange Forum<\/a> - Orange Forum is an easy to deploy forum that has minimal dependencies and uses very little javascript. (<a href=\"https:\/\/groups.goodoldweb.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/s-gv\/orangeforum\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.opensource-socialnetwork.org\/\" rel=\"nofollow\">OSSN<\/a> - Open Source Social Network (OSSN) is a social networking software written in PHP. It allows you to make a social networking website and helps your members build social relationships, with people who share similar professional or personal interests. (<a href=\"https:\/\/github.com\/opensource-socialnetwork\/opensource-socialnetwork\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.oxwall.org\/\" rel=\"nofollow\">Oxwall<\/a> - Oxwall is used for a wide range of projects starting from family sites and custom social networks to collaboration tools and enterprise community solutions. (<a href=\"https:\/\/bitbucket.org\/oxwall\/public\" rel=\"nofollow\">Source Code<\/a>) <code>CPAL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.phpbb.com\/\" rel=\"nofollow\">phpBB<\/a> - Flat-forum bulletin board software solution that can be used to stay in touch with a group of people or can power your entire website. (<a href=\"https:\/\/github.com\/phpbb\/phpbb\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pixelpark\/ppnet\">PPnet<\/a> - Create and host your own social network. <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/pump.io\/\" rel=\"nofollow\">Pump.io<\/a> - Stream server that does most of what people really want from a social network. (<a href=\"https:\/\/github.com\/e14n\/pump.io\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.simplemachines.org\/\" rel=\"nofollow\">Simple Machines Forum<\/a> - Free, professional grade software package that allows you to set up your own online community within minutes. (<a href=\"https:\/\/github.com\/SimpleMachines\/SMF2.1\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/socialhome.network\" rel=\"nofollow\">Socialhome<\/a> - Federated and decentralized profile builder and social network engine. (<a href=\"https:\/\/socialhome.network\/public\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/jaywink\/socialhome\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/hacpai.com\/\" rel=\"nofollow\">Symphony<\/a> - Modern community (forum\/SNS\/blog) platform written in Java. (<a href=\"https:\/\/github.com\/b3log\/symphony\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/www.talkyard.io\/\" rel=\"nofollow\">Talkyard<\/a> - Create a community, where your users can suggest ideas and get questions answered. And have friendly open-ended discussions and chat (Slack\/StackOverflow\/Discourse\/Reddit\/Disqus hybrid). (<a href=\"https:\/\/www.talkyard.io\/forum\/latest\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/debiki\/talkyard\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Scala<\/code><\/li>\n<li><a href=\"http:\/\/www.telescopeapp.org\/\" rel=\"nofollow\">Telescope<\/a> - Open-source social news app built with Meteor. (<a href=\"http:\/\/demo.telescopeapp.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/TelescopeJS\/Telescope\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/tokumei.co\/\" rel=\"nofollow\">Tokumei<\/a> - Anonymous microblogging platform. (<a href=\"https:\/\/demo.tokumei.co\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/git.tokumei.co\/tokumei\/tokumei\" rel=\"nofollow\">Source Code<\/a>) <code>ISC<\/code> <code>rc<\/code><\/li>\n<li><a href=\"http:\/\/twister.net.co\/\" rel=\"nofollow\">twister<\/a> - Fully decentralized P2P microblogging platform leveraging the free software implementations of Bitcoin and BitTorrent protocols. (<a href=\"https:\/\/github.com\/miguelfreitas\/twister-core\">Source Code<\/a>) <code>MIT<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/vanillaforums.org\/\" rel=\"nofollow\">Vanilla Forums<\/a> - Simple and flexible forum software. (<a href=\"https:\/\/github.com\/vanilla\/vanilla\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h3><a id=\"user-content-xmpp\" class=\"anchor\" aria-hidden=\"true\" href=\"#xmpp\"><\/a>XMPP<\/h3>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em><a href=\"https:\/\/en.wikipedia.org\/wiki\/XMPP\" rel=\"nofollow\">Extensible Messaging and Presence Protocol<\/a> software<\/em><\/p>\n<h4><a id=\"user-content-xmpp-servers\" class=\"anchor\" aria-hidden=\"true\" href=\"#xmpp-servers\"><\/a>XMPP Servers<\/h4>\n<ul><li><a href=\"https:\/\/www.ejabberd.im\/\" rel=\"nofollow\">ejabberd<\/a> - XMPP instant messaging server. (<a href=\"https:\/\/github.com\/processone\/ejabberd\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Erlang<\/code><\/li>\n<li><a href=\"http:\/\/kontalk.org\/\" rel=\"nofollow\">Kontalk<\/a> - Kontalk is an Open Source Messenger, similar to WhatsApp (app for android only currently), including end-to-end encryption, server is based on Tigase XMPP Server. (<a href=\"https:\/\/github.com\/kontalk\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/metronome.im\/\" rel=\"nofollow\">Metronome IM<\/a> - Fork of Prosody IM. (<a href=\"https:\/\/github.com\/maranda\/metronome\/\">Source Code<\/a>) <code>MIT<\/code> <code>Lua<\/code><\/li>\n<li><a href=\"https:\/\/www.erlang-solutions.com\/products\/mongooseim.html\" rel=\"nofollow\">MongooseIM<\/a> - Mobile messaging platform with a focus on performance and scalability. (<a href=\"https:\/\/github.com\/esl\/MongooseIM\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Erlang<\/code><\/li>\n<li><a href=\"http:\/\/www.igniterealtime.org\/projects\/openfire\/\" rel=\"nofollow\">Openfire<\/a> - Real time collaboration (RTC) server. (<a href=\"https:\/\/github.com\/igniterealtime\/Openfire\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/prosody.im\/\" rel=\"nofollow\">Prosody IM<\/a> - Feature-rich and easy to configure XMPP server. (<a href=\"http:\/\/hg.prosody.im\/\" rel=\"nofollow\">Source Code<\/a>) <code>MIT<\/code> <code>Lua<\/code><\/li>\n<li><a href=\"http:\/\/www.tigase.net\/content\/tigase-xmpp-server\" rel=\"nofollow\">Tigase<\/a> - XMPP server implementation in Java. <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<\/ul><h4><a id=\"user-content-xmpp-web-clients\" class=\"anchor\" aria-hidden=\"true\" href=\"#xmpp-web-clients\"><\/a>XMPP Web Clients<\/h4>\n<ul><li><a href=\"http:\/\/candy-chat.github.io\/candy\/\" rel=\"nofollow\">Candy<\/a> - Multi user XMPP client written in Javascript. (<a href=\"https:\/\/github.com\/candy-chat\/candy\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/conversejs.org\/\" rel=\"nofollow\">Converse.js<\/a> - Free and open-source XMPP chat client in your browser. (<a href=\"https:\/\/github.com\/jcbrand\/converse.js\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/jsxc.org\" rel=\"nofollow\">JSXC<\/a> - Real-time XMPP web chat application with video calls, file transfer and encrypted communication. There are also versions for Nextcloud\/Owncloud and SOGo. (<a href=\"https:\/\/github.com\/jsxc\/jsxc\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/getkaiwa.com\/\" rel=\"nofollow\">Kaiwa<\/a> - Web based chat client in the style of common paid alternatives. (<a href=\"https:\/\/github.com\/digicoop\/kaiwa\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.salut-a-toi.org\/\" rel=\"nofollow\">Salut \u00e0 Toi<\/a> - Multipurpose, multi frontend, libre and decentralized communication tool. (<a href=\"http:\/\/repos.goffi.org\/sat\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code>\n<ul><li><a href=\"http:\/\/wiki.goffi.org\/wiki\/Libervia\/en\" rel=\"nofollow\">Libervia<\/a> - Web frontend from Salut \u00e0 Toi. (<a href=\"http:\/\/repos.goffi.org\/libervia\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-conference-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#conference-management\"><\/a>Conference Management<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/usecod.com\/\" rel=\"nofollow\">Conference Organizing Distribution (COD)<\/a> - Create conference and event websites built on top of Drupal. (<a href=\"http:\/\/cgit.drupalcode.org\/cod\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/frab.github.io\/frab\/\" rel=\"nofollow\">frab<\/a> - web-based conference planning and management system. It helps to collect submissions, to manage talks and speakers and to create a schedule. (<a href=\"https:\/\/github.com\/frab\/frab\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/pkp.sfu.ca\/ocs\/\" rel=\"nofollow\">Open Conference Systems (OCS)<\/a> - is a free Web publishing tool that will create a complete Web presence for your scholarly conference. (<a href=\"https:\/\/pkp.sfu.ca\/ocs\/ocs_demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/pkp\/ocs\">Source Code<\/a>) <code>GPL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/opencfp\/opencfp\">OpenCFP<\/a> - OpenCFP is a PHP-based conference talk submission system. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/openconferenceware.org\/\" rel=\"nofollow\">OpenConferenceWare<\/a> - An open source web application for supporting conference-like events. This customizable, general-purpose platform provides proposals, sessions, schedules, tracks, user profiles. (<a href=\"https:\/\/github.com\/osbridge\/openconferenceware\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/osem.io\/\" rel=\"nofollow\">osem<\/a> - Event management tailored to free Software conferences. (<a href=\"http:\/\/demo.osem.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/openSUSE\/osem\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/pretalx.org\" rel=\"nofollow\">pretalx<\/a> - Web-based event management, including running a Call for Papers, reviewing submissions, and scheduling talks. Exports and imports for various related tools. (<a href=\"https:\/\/github.com\/pretalx\/pretalx\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-content-management-systems-cms\" class=\"anchor\" aria-hidden=\"true\" href=\"#content-management-systems-cms\"><\/a>Content Management Systems (CMS)<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>CMS are a practical way to setup a website with many features. CMS often come with third party plugins, themes and functionality that is easy to add and customize to your needs. See also <a href=\"#blogging-platforms\">Blogging Platforms<\/a> and <a href=\"#static-site-generators\">Static Site Generators<\/a><\/p>\n<ul><li><a href=\"https:\/\/www.apiq.io\/\" rel=\"nofollow\">APIQ CMS<\/a> - Simple and powerful Ruby on Rails CMS for developers. (<a href=\"http:\/\/demo.apiq.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/apiqcms\/kms\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/apostrophecms.org\/\" rel=\"nofollow\">Apostrophe<\/a> - Node.js CMS with a focus on extensible in-context editing tools. (<a href=\"http:\/\/demo.apostrophecms.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/punkave\/apostrophe\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/backdropcms.org\/\" rel=\"nofollow\">Backdrop CMS<\/a> - The comprehensive CMS for small to medium sized businesses and non-profits. (<a href=\"https:\/\/github.com\/backdrop\/backdrop\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/bauncms.com\/\" rel=\"nofollow\">Baun<\/a> - Modern, lightweight, extensible CMS for PHP. (<a href=\"https:\/\/github.com\/BaunCMS\/Baun\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.bigtreecms.org\/\" rel=\"nofollow\">BigTree CMS<\/a> - Straightforward, well documented, and capable written with PHP and MySQL. (<a href=\"https:\/\/github.com\/bigtreecms\/BigTree-CMS\">Source Code<\/a>) <code>LGPL-2.1<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/bolt.cm\/\" rel=\"nofollow\">Bolt CMS<\/a> - Open source Content Management Tool, which strives to be as simple and straightforward as possible. (<a href=\"https:\/\/try.bolt.cm\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/bolt\/bolt\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.cmsmadesimple.org\/\" rel=\"nofollow\">CMS Made Simple<\/a> - Open source content management system, faster and easier management of website contents, scalable for small businesses to large corporations. (<a href=\"http:\/\/svn.cmsmadesimple.org\/svn\/cmsmadesimple\/trunk\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/getcockpit.com\" rel=\"nofollow\">Cockpit<\/a> - Simple Content Platform to manage any structured content. (<a href=\"https:\/\/github.com\/agentejo\/cockpit\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.concrete5.org\/\" rel=\"nofollow\">Concrete 5 CMS<\/a> - Open source content management system. (<a href=\"https:\/\/github.com\/concrete5\/concrete5\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.couchcms.com\/\" rel=\"nofollow\">CouchCMS<\/a> - Simple Open-Source CMS for designers. (<a href=\"https:\/\/github.com\/CouchCMS\/CouchCMS\">Source Code<\/a>) <code>CPAL-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/getdirectus.com\/\" rel=\"nofollow\">Directus<\/a> - Directus is a powerful and intuitive headless CMS for managing SQL databases with custom architectures. Built around a robust and extensible API, this decoupled content management framework is perfect for websites, apps, or multi-client projects. (<a href=\"https:\/\/github.com\/directus\/directus\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.drupal.org\/\" rel=\"nofollow\">Drupal<\/a> - Advanced open source content management platform. (<a href=\"http:\/\/cgit.drupalcode.org\/drupal\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.elabftw.net\" rel=\"nofollow\">eLabFTW<\/a> - Online lab notebook for research labs. Store experiments, use a database to find reagents or protocols, use trusted timestamping to legally timestamp an experiment, export as pdf or zip archive, share with collaborators\u2026. (<a href=\"https:\/\/demo.elabftw.net\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/elabftw\/elabftw\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/get-simple.info\/\" rel=\"nofollow\">GetSimple CMS<\/a> - The Simplest Content Management System. Ever. (<a href=\"https:\/\/github.com\/GetSimpleCMS\/GetSimpleCMS\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.impresspages.org\/\" rel=\"nofollow\">ImpressPages CMS<\/a> - Easy code meets easy admin. (<a href=\"https:\/\/www.impresspages.org\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/impresspages\/ImpressPages\">Source Code<\/a>) <code>GPL-3.0\/MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.joomla.org\/\" rel=\"nofollow\">Joomla!<\/a> - Advanced Content Management System (CMS). (<a href=\"https:\/\/github.com\/joomla\/joomla-cms\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/keystonejs.com\/\" rel=\"nofollow\">KeystoneJS<\/a> - CMS and Web Application Platform. (<a href=\"http:\/\/demo.keystonejs.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/keystonejs\/keystone\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/modx.com\/\" rel=\"nofollow\">MODX<\/a> - MODX is an advanced content management and publishing platform. The current version is called 'Revolution'. (<a href=\"https:\/\/github.com\/modxcms\/revolution\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.neos.io\" rel=\"nofollow\">Neos<\/a> - Neos or TYPO3 Neos (for version 1) is a modern, open source CMS. (<a href=\"https:\/\/git.typo3.org\/Packages\/TYPO3.Neos.git\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/noosfero.org\/\" rel=\"nofollow\">Noosfero<\/a> - Noosfero is a web platform for social and solidarity economy networks with blog, e-Portfolios, CMS, RSS, thematic discussion, events agenda and collective intelligence for solidarity economy in the same system. (<a href=\"https:\/\/gitlab.com\/noosfero\/noosfero\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/octobercms.com\/\" rel=\"nofollow\">october<\/a> - Free, open-source, self-hosted CMS platform. (<a href=\"https:\/\/github.com\/octobercms\/october\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/omeka.org\" rel=\"nofollow\">Omeka<\/a> - Create complex narratives and share rich collections, adhering to Dublin Core standards with Omeka on your server, designed for scholars, museums, libraries, archives, and enthusiasts. (<a href=\"http:\/\/omeka.org\/showcase\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/omeka\/Omeka\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/pagekit.com\/\" rel=\"nofollow\">Pagekit<\/a> - New modern CMS to create and share. (<a href=\"https:\/\/github.com\/pagekit\/pagekit\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/picocms.org\/\" rel=\"nofollow\">Pico<\/a> - Stupidly simple, blazing fast, flat file CMS. (<a href=\"https:\/\/github.com\/picocms\/Pico\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.pimcore.org\/\" rel=\"nofollow\">Pimcore<\/a> - Multi-Channel Experience and Engagement Management Platform. (<a href=\"https:\/\/github.com\/pimcore\/pimcore\">Source Code<\/a>) <code>BSD<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/plone.org\/\" rel=\"nofollow\">Plone<\/a> - Powerful open-source CMS system. (<a href=\"https:\/\/github.com\/plone\">Source Code<\/a>) <code>ZPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/processwire.com\/\" rel=\"nofollow\">ProcessWire<\/a> - ProcessWire is an open source content management system (CMS) and web application framework aimed at the needs of designers, developers and their clients. (<a href=\"https:\/\/github.com\/ryancramerdesign\/ProcessWire\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/propertywebbuilder.com\" rel=\"nofollow\">PropertyWebBuilder<\/a> - The ultimate Ruby on Rails engine for creating real estate websites. (<a href=\"https:\/\/propertywebbuilder.herokuapp.com\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/etewiah\/property_web_builder\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/publify\/publify\">Publify<\/a> - Simple but full featured web publishing software. <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.redaxo.org\" rel=\"nofollow\">REDAXO<\/a> - Simple, flexible and useful content management system (documentation only available in German). (<a href=\"https:\/\/github.com\/redaxo\/redaxo\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/redaxscript.com\" rel=\"nofollow\">Redaxscript<\/a> - Ultra lightweight CMS for MySQL, SQLite and PostgreSQL. (<a href=\"https:\/\/demo.redaxscript.com\/demo\/login\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/redaxmedia\/redaxscript\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.roadiz.io\/\" rel=\"nofollow\">Roadiz<\/a> -  Modern CMS based on a node system which can handle many types of services. (<a href=\"https:\/\/github.com\/roadiz\/roadiz\/\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.silverstripe.org\" rel=\"nofollow\">SilverStripe<\/a> - Easy to use CMS with powerful MVC framework underlying. (<a href=\"http:\/\/demo.silverstripe.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/silverstripe\">Source Code<\/a>) <code>BSD<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.sphido.org\/\" rel=\"nofollow\">Sphido<\/a> - Fast, lightweight, flat file CMS for PHP. (<a href=\"https:\/\/github.com\/sphido\/cms\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.spip.net\/fr\" rel=\"nofollow\">SPIP<\/a> - Publication system for the Internet aimed at collaborative work, multilingual environments, and simplicity of use for web authors. (<a href=\"https:\/\/core.spip.net\/projects\/spip\/repository\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.subrion.org\" rel=\"nofollow\">Subrion<\/a> - Subrion is a free open source content management system that allows you to build websites for any purpose. Yes, from blog to corporate mega portal. (<a href=\"http:\/\/demos.subrion.com\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/intelliants\/subrion\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/textpattern.com\/\" rel=\"nofollow\">Textpattern<\/a> - Flexible, elegant and easy-to-use CMS. (<a href=\"http:\/\/textpattern.co\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/textpattern\/textpattern\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/typo3.org\/\" rel=\"nofollow\">TYPO3<\/a> - Powerful and advanced CMS with a large community. (<a href=\"https:\/\/github.com\/TYPO3\/TYPO3.CMS\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/umbraco.com\/\" rel=\"nofollow\">Umbraco<\/a> - The friendly CMS. Free and open source with an amazing community. (<a href=\"https:\/\/github.com\/umbraco\/Umbraco-CMS\">Source Code<\/a>) <code>MIT<\/code> <code>.NET<\/code><\/li>\n<li><a href=\"https:\/\/wagtail.io\/\" rel=\"nofollow\">Wagtail<\/a> - Django content management system focused on flexibility and user experience. (<a href=\"https:\/\/github.com\/wagtail\/wagtail\">Source Code<\/a>) <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.wondercms.com\" rel=\"nofollow\">WonderCMS<\/a> - WonderCMS is the smallest flat file CMS since 2008. (<a href=\"https:\/\/www.wondercms.com\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/robiso\/wondercms\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/wordpress.org\/\" rel=\"nofollow\">WordPress<\/a> - The worlds most-used blogging and CMS engine. (<a href=\"https:\/\/github.com\/WordPress\/WordPress\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><p><em>Recipe management<\/em><\/p>\n<ul><li><a href=\"https:\/\/github.com\/open-eats\/OpenEats\">OpenEats<\/a> - Recipe management site that allows users to create, store, share and rate recipes, create grocery lists, and more. (<a href=\"https:\/\/open-eats.github.io\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<\/ul><h3><a id=\"user-content-e-commerce\" class=\"anchor\" aria-hidden=\"true\" href=\"#e-commerce\"><\/a>E-commerce<\/h3>\n<ul><li><a href=\"https:\/\/www.attendize.com\/\" rel=\"nofollow\">Attendize<\/a> - Ticket selling and event management platform. (<a href=\"https:\/\/github.com\/attendize\/attendize\">Source Code<\/a>) <code>AAL<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.coreshop.org\" rel=\"nofollow\">CoreShop<\/a> - CoreShop is a e-commerce plugin for Pimcore. (<a href=\"https:\/\/github.com\/coreshop\/CoreShop\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/drupalcommerce.org\" rel=\"nofollow\">Drupal Commerce<\/a> - Drupal Commerce is a popular e-commerce module for Drupal CMS, with support for dozens of payment, shipping, and shopping related modules. (<a href=\"https:\/\/github.com\/drupalcommerce\/commerce\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/magento.com\/\" rel=\"nofollow\">Magento<\/a> - Leading provider of open omnichannel innovation. (<a href=\"https:\/\/magento.com\/schedule-a-demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/magento\/magento2\">Source Code<\/a>) <code>OSL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/microweber.com\/\" rel=\"nofollow\">Microweber<\/a> - Drag and Drop CMS and online shop. (<a href=\"http:\/\/demo.microweber.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/microweber\/microweber\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.openbazaar.org\" rel=\"nofollow\">OpenBazaar<\/a> - Decentralized marketplace using cryptocurrency. (<a href=\"https:\/\/github.com\/openbazaar\/openbazaar-go\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.opencart.com\" rel=\"nofollow\">OpenCart<\/a> - Free open source shopping cart solution. (<a href=\"https:\/\/github.com\/opencart\/opencart\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/open-classifieds.com\/\" rel=\"nofollow\">Open Classifieds<\/a> - Free open-source, self-hosted CMS for classifieds sites. (<a href=\"https:\/\/github.com\/open-classifieds\/openclassifieds2\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.opensourcepos.org\/\" rel=\"nofollow\">Open Source POS<\/a> - Open Source Point of Sale is a web based point of sale system. (<a href=\"https:\/\/github.com\/opensourcepos\/opensourcepos\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/osclass.org\/\" rel=\"nofollow\">Osclass<\/a> - One-stop shop to building your own classifieds marketplace. (<a href=\"https:\/\/github.com\/osclass\/Osclass\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/oxidforge.org\" rel=\"nofollow\">OXID eShop<\/a> - OXID eShop is a flexible open source e-commerce software with a wide range of functionalities. (<a href=\"http:\/\/demoshop.oxid-esales.com\/community-edition\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/OXID-eSales\/oxideshop_ce\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/openfoodnetwork.org\/\" rel=\"nofollow\">Open Food Network<\/a> - Online marketplace for local food. It enables a network of independent online food stores that connect farmers and food hubs with individuals and local businesses. (<a href=\"https:\/\/github.com\/openfoodfoundation\/openfoodnetwork\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.prestashop.com\/\" rel=\"nofollow\">PrestaShop<\/a> - PrestaShop offers a free, open-source and fully scalable e-commerce solution. (<a href=\"http:\/\/demo.prestashop.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/PrestaShop\/PrestaShop\">Source Code<\/a>) <code>OSL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/pretix.eu\/\" rel=\"nofollow\">Pretix<\/a> - Django based ticket sales platform for events. (<a href=\"https:\/\/github.com\/pretix\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/reactioncommerce.com\/\" rel=\"nofollow\">Reaction Commerce<\/a> - Customizable, real-time reactive, JavaScript commerce platform. (<a href=\"https:\/\/github.com\/reactioncommerce\/reaction\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/getsaleor.com\/\" rel=\"nofollow\">Saleor<\/a> - Django based open-sourced e-commerce storefront. (<a href=\"https:\/\/demo.getsaleor.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/mirumee\/saleor\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.shuup.com\/\" rel=\"nofollow\">Shuup<\/a> - Django powered fully customizable open source e-commerce framework for small and large sites. (<a href=\"https:\/\/github.com\/shuup\/shuup\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/shopware.com\/community\/\" rel=\"nofollow\">Shopware Community Edition<\/a> - PHP based ppen source e-commerce software made in Germany. (<a href=\"https:\/\/www.shopwaredemo.de\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/shopware\/shopware\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/sylius.org\/\" rel=\"nofollow\">Sylius<\/a> - Symfony2 powered open source full-stack platform for eCommerce. (<a href=\"http:\/\/sylius.org\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Sylius\/Sylius\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/thelia.net\/\" rel=\"nofollow\">Thelia<\/a> - Thelia is an open source and flexible e-commerce solution. (<a href=\"http:\/\/demo.thelia.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/thelia\/thelia\">Source Code<\/a>) <code>LGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.woothemes.com\/woocommerce\/\" rel=\"nofollow\">WooCommerce<\/a> - WordPress based e-commerce solution. (<a href=\"https:\/\/github.com\/woothemes\/woocommerce\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-dns\" class=\"anchor\" aria-hidden=\"true\" href=\"#dns\"><\/a>DNS<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#dns\">https:\/\/github.com\/n1trux\/awesome-sysadmin#dns<\/a><\/p>\n<ul><li><a href=\"https:\/\/coredns.io\/\" rel=\"nofollow\">CoreDNS<\/a> - Plugin driven DNS Server with support for proxying to Google's DNS-over-HTTPS. (<a href=\"https:\/\/github.com\/coredns\/coredns\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.nsupdate.info\/\" rel=\"nofollow\">nsupdate.info<\/a> - nsupdate.info is a dynamic DNS service. (<a href=\"https:\/\/www.nsupdate.info\/account\/register\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/nsupdate-info\/nsupdate.info\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/spftoolbox.charlesabarnes.com\" rel=\"nofollow\">SPF Toolbox<\/a> - Application to look up DNS records such as SPF, MX, Whois, and more. (<a href=\"https:\/\/github.com\/bulbajackel\/SPFtoolbox\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-document-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#document-management\"><\/a>Document Management<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.casebox.org\" rel=\"nofollow\">CaseBox<\/a> - Manage all your organization's information in one system. (<a href=\"https:\/\/github.com\/KETSE\/casebox\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP\/Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Edraens\/EdPaper\">EdPaper<\/a> - PDF organizer with users management. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.mayan-edms.com\" rel=\"nofollow\">Mayan EDMS<\/a> - Free Open Source Electronic Document Management System. An electronic vault for your documents with preview generation, OCR, and automatic categorization among other features. (<a href=\"https:\/\/gitlab.com\/mayan-edms\/mayan-edms\" rel=\"nofollow\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/danielquinn\/paperless\">Paperless<\/a> - Scan, index, and archive all of your paper documents. <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-e-books-and-integrated-library-systems-ils\" class=\"anchor\" aria-hidden=\"true\" href=\"#e-books-and-integrated-library-systems-ils\"><\/a>E-books and Integrated Library Systems (ILS)<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em>Personal e-book management software.<\/em><\/p>\n<ul><li><a href=\"https:\/\/calibre-ebook.com\/\" rel=\"nofollow\">Calibre<\/a> - E-book library manager that can view, convert, and catalog e-books in most of the major e-book formats and provides a built-in Web server for remote clients. (<a href=\"https:\/\/calibre-ebook.com\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/launchpad.net\/calibre\" rel=\"nofollow\">Source code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code>\n<ul><li><a href=\"http:\/\/projekte.textmulch.de\/bicbucstriim\/\" rel=\"nofollow\">BicBucStriim<\/a> - Provides web-based access to your Calibre Library's e-book collection. (<a href=\"https:\/\/github.com\/rvolz\/BicBucStriim\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/janeczku\/calibre-web\">Calibre Web<\/a> - Web app providing a clean interface for browsing, reading and downloading eBooks using an existing Calibre database. (<a href=\"https:\/\/github.com\/janeczku\/calibre-web\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/blog.slucas.fr\/en\/oss\/calibre-opds-php-server\" rel=\"nofollow\">COPS<\/a> - Lightweight e-book server alternative to Calibre content server or <a href=\"https:\/\/calibre2opds.com\/2011\/11\/03\/new-site-for-calibre2opds\/\" rel=\"nofollow\">Calibre2OPDS<\/a>. (<a href=\"http:\/\/cops-demo.slucas.fr\/index.php\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/seblucas\/cops\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/getpolarized.io\/\" rel=\"nofollow\">Polar Bookshelf<\/a> - Polar is a personal knowledge repository for PDF and web content supporting incremental reading and document annotation. (<a href=\"https:\/\/github.com\/burtonator\/polar-bookshelf\">Source Code<\/a>) <code>GPL<\/code> <code>Javascript<\/code><\/li>\n<\/ul><\/li>\n<\/ul><p><em>Enterprise-class library management software.<\/em><\/p>\n<ul><li><a href=\"https:\/\/evergreen-ils.org\" rel=\"nofollow\">Evergreen<\/a> - Highly-scalable software for libraries that helps library patrons find library materials, and helps libraries manage, catalog, and circulate those materials. (<a href=\"https:\/\/github.com\/evergreen-library-system\/Evergreen\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PL\/pgSQL<\/code><\/li>\n<li><a href=\"https:\/\/koha-community.org\/\" rel=\"nofollow\">Koha<\/a> - Enterprise-class ILS with modules for acquisitions, circulation, cataloging, label printing, offline circulation for when Internet access is not available, and much more. (<a href=\"https:\/\/koha-community.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Koha-Community\/Koha\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Perl<\/code><\/li>\n<\/ul><h2><a id=\"user-content-federated-identityauthentication\" class=\"anchor\" aria-hidden=\"true\" href=\"#federated-identityauthentication\"><\/a>Federated Identity\/Authentication<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#identity-management\">https:\/\/github.com\/n1trux\/awesome-sysadmin#identity-management<\/a><\/p>\n<h2><a id=\"user-content-feed-readers\" class=\"anchor\" aria-hidden=\"true\" href=\"#feed-readers\"><\/a>Feed Readers<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.commafeed.com\/\" rel=\"nofollow\">CommaFeed<\/a> - Google Reader inspired self-hosted RSS reader. (<a href=\"https:\/\/github.com\/Athou\/commafeed\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/feedbin.com\/\" rel=\"nofollow\">Feedbin<\/a> - Simple, fast and nice looking RSS reader. (<a href=\"https:\/\/github.com\/feedbin\/feedbin\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/feedhq.org\/\" rel=\"nofollow\">FeedHQ<\/a> - FeedHQ is a web-based feed reader. (<a href=\"https:\/\/github.com\/feedhq\/feedhq\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/freshrss.org\/\" rel=\"nofollow\">FreshRSS<\/a> - Self-hostable RSS feed aggregator. (<a href=\"http:\/\/demo.freshrss.org\/i\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/FreshRSS\/FreshRSS\">Source Code<\/a>, <a href=\"https:\/\/github.com\/Alkarex\/EasyRSS\">Mobile app<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/1pxsolidblack.pl\/jarr-en.html\" rel=\"nofollow\">JARR<\/a> - JARR (Just Another RSS Reader) is a web-based news aggregator and reader (fork of Newspipe). (<a href=\"https:\/\/jarr.info\/login?next=%2F\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/jaesivsm\/JARR\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/tontof.net\/kriss\/feed\/\" rel=\"nofollow\">Kriss Feed<\/a> - Simple and smart (or stupid) feed reader. (<a href=\"http:\/\/tontof.net\/feed\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/tontof\/kriss_feed\/\">Source Code<\/a>) <code>CC0-1.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/leed.idleman.fr\" rel=\"nofollow\">Leed<\/a> - Leed (for Light Feed) is a Free and minimalist RSS aggregator. (<a href=\"https:\/\/github.com\/ldleman\/Leed\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/socketubs\/leselys\">Leselys<\/a> - Your very elegant RSS reader. <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/cubny\/lite-reader\">Lite-Reader<\/a> - Read your feeds on your own machine with a simple and lite application. (<a href=\"http:\/\/cubny.com\/lite-reader\/\" rel=\"nofollow\">Demo<\/a>) <code>BSD-3-Clause<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/moonmoon.org\/\" rel=\"nofollow\">Moonmoon<\/a> - simple feed agregator (planet like): it only aggregates feeds and spits them out in one single page. (<a href=\"https:\/\/github.com\/mauricesvay\/moonmoon\">Source Code<\/a>) <code>BSD<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/miniflux.net\/\" rel=\"nofollow\">Miniflux<\/a> - Miniflux 2 is a minimalist and open source news reader, written in Go and PostgreSQL. (<a href=\"https:\/\/github.com\/miniflux\/miniflux\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/www.newsblur.com\/\" rel=\"nofollow\">NewsBlur<\/a> - NewsBlur is a personal news reader that brings people together to talk about the world. A new sound of an old instrument. (<a href=\"https:\/\/github.com\/samuelclay\/NewsBlur\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/newspipe.io\/\" rel=\"nofollow\">Newspipe<\/a> - Newspipe is a web news aggregator and reader. (<a href=\"https:\/\/www.newspipe.org\/signup\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/newspipe\/newspipe\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/reader.nunux.org\/\" rel=\"nofollow\">Nunux Reader<\/a> - Simple, fast and reactive RSS reader. (<a href=\"https:\/\/github.com\/ncarlier\/nunux-reader\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/readerself.com\/\" rel=\"nofollow\">Reader-Self<\/a> - Self-hosted rss reader (php \/ mysql or sqlite) - Google Reader alternative. (<a href=\"https:\/\/github.com\/readerself\/readerself\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.allthingsrss.com\/rss2email\/\" rel=\"nofollow\">RSS2EMail<\/a> - Fetches RSS\/Atom-feeds and pushes new Content to any email-receiver, supports OPML. (<a href=\"https:\/\/github.com\/wking\/rss2email\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pietheinstrengholt\/rssmonster\">RSS Monster<\/a> - RSS Monster is an easy to use web-based RSS aggregator and reader compatible with the Fever API, created as an alternative for Google Reader. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/herrbischoff\/screaming-liquid-tiger\">Screaming Liquid Tiger<\/a> - Simple script to automatically generate valid RSS and Atom feeds from a list of media files in the same folder. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/selfoss.aditu.de\/\" rel=\"nofollow\">Selfoss<\/a> - New multipurpose rss reader, live stream, mashup, aggregation web application. (<a href=\"https:\/\/github.com\/SSilence\/selfoss\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/sismics.com\/reader\/\" rel=\"nofollow\">Sismics Reader<\/a> - Free and open source feeds reader, including all major Google Reader features. (<a href=\"https:\/\/www.sismics.com\/reader\/#!\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/sismics\/reader\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/swanson\/stringer\">Stringer<\/a> - Work-in-progress self-hosted, anti-social RSS reader. <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/fazalmajid\/temboz\">Temboz<\/a> - Two-column feed reader emphasizing filtering capabilities to manage information overload. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/tt-rss.org\/gitlab\/fox\/tt-rss\" rel=\"nofollow\">Tiny Tiny RSS<\/a> - Open source web-based news feed (RSS\/Atom) reader and aggregator. (<a href=\"http:\/\/framanews.org\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code>\n<ul><li><a href=\"http:\/\/gritttt-rss.nicolashoening.de\/\" rel=\"nofollow\">gritttt-rss<\/a> - More features for Tiny Tiny RSS. (<a href=\"https:\/\/github.com\/nhoening\/gritttt-rss\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/mboinet\/ttrss-mobile\">ttrss-mobile<\/a> - Mobile webapp for Tiny Tiny RSS. <code>AGPL-3.0<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/kucrut\/ttrss-reader\">ttrss-reader<\/a> - Light and responsive client for TTRSS. <code>GPL-2.0<\/code> <code>Javascript<\/code><\/li>\n<\/ul><\/li>\n<li><a href=\"https:\/\/getstream.io\/winds\/\" rel=\"nofollow\">Winds<\/a> <code>\u26a0<\/code> - Open source and beautiful RSS reader built using React\/Redux\/Sails\/Node and Stream. It showcases personalized feeds powered by the Stream API. (<a href=\"https:\/\/winds.getstream.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/GetStream\/Winds\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h2><a id=\"user-content-file-sharing-and-synchronization\" class=\"anchor\" aria-hidden=\"true\" href=\"#file-sharing-and-synchronization\"><\/a>File Sharing and Synchronization<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>Some <a href=\"#groupware\">Groupware<\/a> solutions also feature file sharing and synchronization.<\/p>\n<h4><a id=\"user-content-distributed-filesystems\" class=\"anchor\" aria-hidden=\"true\" href=\"#distributed-filesystems\"><\/a>Distributed filesystems<\/h4>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#distributed-filesystems\">https:\/\/github.com\/n1trux\/awesome-sysadmin#distributed-filesystems<\/a><\/p>\n<h4><a id=\"user-content-file-transfersynchronization\" class=\"anchor\" aria-hidden=\"true\" href=\"#file-transfersynchronization\"><\/a>File transfer\/synchronization<\/h4>\n<ul><li><a href=\"https:\/\/git-annex.branchable.com\/\" rel=\"nofollow\">Git Annex<\/a> - File synchronization between computers, servers, external drives. (<a href=\"https:\/\/git.joeyh.name\/index.cgi\/git-annex.git\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Haskell<\/code><\/li>\n<li><a href=\"https:\/\/kinto.readthedocs.org\" rel=\"nofollow\">Kinto<\/a> - Kinto is a minimalist JSON storage service with synchronisation and sharing abilities. (<a href=\"https:\/\/github.com\/Kinto\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/nextcloud.com\/\" rel=\"nofollow\">Nextcloud<\/a> - Access and share your files, calendars, contacts, mail and <a href=\"https:\/\/apps.nextcloud.com\/\" rel=\"nofollow\">more<\/a> from any device, on your terms. (<a href=\"https:\/\/demo.nextcloud.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/nextcloud\/server\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.openssh.com\/\" rel=\"nofollow\">OpenSSH\/SFTP<\/a> - Secure File Transfer Program. (<a href=\"http:\/\/cvsweb.openbsd.org\/cgi-bin\/cvsweb\/src\/usr.bin\/ssh\" rel=\"nofollow\">Source Code<\/a>) <code>BSD<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/owncloud.org\/\" rel=\"nofollow\">ownCloud<\/a> - All-in-one solution for saving, synchronizing, viewing, editing and sharing files, calendars, address books and more. (<a href=\"https:\/\/github.com\/owncloud\/core\">Source Code<\/a>, <a href=\"https:\/\/github.com\/owncloud\/core\/wiki\/Apps\">Clients<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/pydio.com\/\" rel=\"nofollow\">Pydio<\/a> - Turn any web server into a powerful file management system and an alternative to mainstream cloud storage providers. (<a href=\"https:\/\/github.com\/pydio\/pydio-core\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.samba.org\/\" rel=\"nofollow\">Samba<\/a> - Samba is the standard Windows interoperability suite of programs for Linux and Unix. It provides secure, stable and fast file and print services for all clients using the SMB\/CIFS protocol. (<a href=\"https:\/\/git.samba.org\/samba.git\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/www.seafile.com\/en\/home\/\" rel=\"nofollow\">Seafile<\/a> - File hosting and sharing solution primary for teams and organizations. (<a href=\"https:\/\/seacloud.cc\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/haiwen\/seafile\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/sparkleshare.org\/\" rel=\"nofollow\">SparkleShare<\/a> - Self hosted, instant, secure file sync. (<a href=\"https:\/\/github.com\/hbons\/SparkleShare\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/www.syncany.org\/\" rel=\"nofollow\">Syncany<\/a> - Secure file sync software for arbitrary storage backends, an open-source cloud storage and filesharing application. Securely synchronize your files to any kind of storage. <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/syncthing.net\/\" rel=\"nofollow\">Syncthing<\/a> - Syncthing is an open source peer-to-peer file synchronisation tool. (<a href=\"https:\/\/github.com\/syncthing\/syncthing\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.cis.upenn.edu\/~bcpierce\/unison\/\" rel=\"nofollow\">Unison<\/a> - Unison is a file-synchronization tool for OSX, Unix, and Windows. <code>GPL-3.0<\/code> <code>OCaml<\/code><\/li>\n<li><a href=\"http:\/\/z-push.org\/\" rel=\"nofollow\">Z-Push<\/a> - Implementation of Microsoft\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/ActiveSync\" rel=\"nofollow\">ActiveSync<\/a> protocol. (<a href=\"https:\/\/stash.z-hub.io\/projects\/ZP\/repos\/z-push\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h4><a id=\"user-content-peer-to-peer-filesharing\" class=\"anchor\" aria-hidden=\"true\" href=\"#peer-to-peer-filesharing\"><\/a>Peer-to-peer filesharing<\/h4>\n<ul><li><a href=\"https:\/\/webtorrent.io\/\" rel=\"nofollow\">bittorrent-tracker<\/a> - Simple, robust, BitTorrent tracker (client and server) implementation. (<a href=\"https:\/\/github.com\/feross\/bittorrent-tracker\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jpillora\/cloud-torrent\">cloud-torrent<\/a> - Torrent Web Client with HTTP retrievable or streamable downloaded files. <code>AGPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/datproject.org\" rel=\"nofollow\">Dat Project<\/a> - Powerful decentralized file sharing applications built from a large ecosystem of modules. (<a href=\"https:\/\/github.com\/datproject\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/file.pizza\/\" rel=\"nofollow\">FilePizza<\/a> - Peer-to-peer file transfers in your browser. (<a href=\"https:\/\/github.com\/kern\/filepizza\">Source Code<\/a>) <code>BSD<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/mozilla\/send\">Firefox Send<\/a> - A file sharing experiment which allows you to send encrypted files to other users. <code>MPL-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/feross\/instant.io\">instant.io<\/a> - Streaming file transfer over WebTorrent. (<a href=\"https:\/\/instant.io\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/boramalper\/magnetico\">Magnetico<\/a> - Magnetico is the first autonomous (self-hosted) BitTorrent DHT search engine suite that is designed for end-users. <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/sergiotapia\/magnetissimo\">Magnetissimo<\/a> - Search engine that indexes all popular torrent sites. <code>MIT<\/code> <code>Elixir<\/code><\/li>\n<li><a href=\"http:\/\/erdgeist.org\/arts\/software\/opentracker\/\" rel=\"nofollow\">Opentracker<\/a> - Open and free bittorrent tracker. It aims for minimal resource usage and is intended to run at your wlan router. (<a href=\"http:\/\/erdgeist.org\/gitweb\/opentracker\/\" rel=\"nofollow\">Source Code<\/a>) <code>Beerware<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/asapach\/peerflix-server\">peerflix-server<\/a> - Downloads torrent files and provides a direct link download or a direct link stream. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.qbittorrent.org\/\" rel=\"nofollow\">qBittorrent<\/a> - Free cross-platform bittorrent client with a feature rich Web UI for remote access. (<a href=\"https:\/\/github.com\/qbittorrent\/qBittorrent\/\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/swetorrentking\/rartracker\">rartracker<\/a> - Complete private bittorrent tracker. <code>WTFPL<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/reep.io\" rel=\"nofollow\">Reep<\/a> - In-browser peer-to-peer file transfer and streaming made easy. (<a href=\"https:\/\/github.com\/KodeKraftwerk\/reepio\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/transmissionbt.com\/\" rel=\"nofollow\">Transmission<\/a> - Fast, easy, Free Bittorrent client. (<a href=\"https:\/\/github.com\/transmission\/transmission\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<\/ul><h4><a id=\"user-content-object-storagefile-servers\" class=\"anchor\" aria-hidden=\"true\" href=\"#object-storagefile-servers\"><\/a>Object storage\/file servers<\/h4>\n<ul><li><a href=\"https:\/\/minio.io\/\" rel=\"nofollow\">Minio<\/a> - Minio is an open source object storage server compatible with Amazon S3 APIs. (<a href=\"https:\/\/github.com\/minio\/minio\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.zenko.io\/cloudserver\" rel=\"nofollow\">Zenko CloudServer<\/a> - Zenko CloudServer, an open-source Node.js implementation of a server handling the Amazon S3 protocol. (<a href=\"https:\/\/github.com\/scality\/S3\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h4><a id=\"user-content-single-clickdrag-n-drop-upload\" class=\"anchor\" aria-hidden=\"true\" href=\"#single-clickdrag-n-drop-upload\"><\/a>Single-click\/drag-n-drop upload<\/h4>\n<ul><li><a href=\"https:\/\/github.com\/broncowdd\/BoZoN\">BoZoN<\/a> - Minimalist Drag and drop file sharing app. <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/coquelicot.potager.org\/\" rel=\"nofollow\">Coquelicot<\/a> - Coquelicot is a \u201cone-click\u201d file sharing web application with a focus on protecting users\u2019 privacy. (<a href=\"https:\/\/coquelicot.potager.org\/gitweb\/?p=coquelicot.git\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/silverwind\/droppy\">droppy<\/a> - droppy is a self-hosted cloud server with an interface similar to desktop file managers and has capabilities to edit files on-the-fly as well as view and playback media directly in the browser. (<a href=\"http:\/\/droppy.silverwind.io\/\" rel=\"nofollow\">Demo<\/a>) <code>BSD<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/epoupon\/fileshelter\">FileShelter<\/a> - FileShelter is a self-hosted software that allows you to easily share files over the Internet. (<a href=\"http:\/\/fileshelter.demo.poupon.io\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/axeloz\/filesharing\">Files Sharing<\/a> - Open Source and self-hosted files sharing application based on unique and temporary links. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/imgbi\/img.bi\">img.bi<\/a> - img.bi is a secure image hosting. Images are encrypted using AES-256 with random key in browser before upload. <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ipfspics\/server\">ipfs.pics<\/a> - ipfs.pics is a distributed image hosting website. <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/gitlab.com\/mojo42\/Jirafeau\" rel=\"nofollow\">Jirafeau<\/a> - Jirafeau is a web site permitting to upload a file in a simple way and give an unique link to it. (<a href=\"http:\/\/jirafeau.net\/\" rel=\"nofollow\">Demo<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/andreimarcu\/linx-server\">linx<\/a> - File sharing application and pastebin with API, auto-expiry, deletion keys, and web seed support. (<a href=\"https:\/\/linx.li\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/git.framasoft.org\/luc\/lufi\" rel=\"nofollow\">lufi<\/a> - Let's Upload that FIle, client-side encrypted. (<a href=\"https:\/\/demo.lufi.io\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/git.framasoft.org\/luc\/lufi\/tree\/master\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ldidry\/lutim\">lutim<\/a> - Let's Upload That Image. <code>AGPL-3.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ethanal\/nimbus\">Nimbus<\/a> - Drag-and-drop shortlink file sharer for OS X. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/micahflee\/onionshare\">OnionShare<\/a> - Securely and anonymously share a file of any size. <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.pictshare.net\/\" rel=\"nofollow\">PictShare<\/a> - PictShare is a multi lingual, open source image hosting service with a simple resizing and upload API. (<a href=\"https:\/\/www.pictshare.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/chrisiaut\/pictshare\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/root-gg\/plik\">Plik<\/a> - Plik is a scalable and friendly temporary file upload system. (<a href=\"https:\/\/plik.root.gg\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Pomf\/Pomf\">Pomf<\/a> - Simple file uploading and sharing, source for the now shut down site Pomf.se. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.projectsend.org\/\" rel=\"nofollow\">ProjectSend<\/a> - Upload files and assign them to specific clients you create. Give access to those files to your clients. (<a href=\"https:\/\/github.com\/ignacionelson\/ProjectSend\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/psi-4ward\/psitransfer\">PsiTransfer<\/a> - Simple open source self-hosted file sharing solution with robust up-\/download-resume and password protection. <code>BSD<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/eikek\/sharry\">Sharry<\/a> - Share files easily over the internet between authenticated and anonymous users (both ways) with resumable up- and downloads. (<a href=\"https:\/\/sharrydemo.eknet.org\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>Scala\/Java<\/code><\/li>\n<li><a href=\"https:\/\/uguu.se\/\" rel=\"nofollow\">Uguu<\/a> - Stores files and deletes after X amount of time. (<a href=\"https:\/\/github.com\/nokonoko\/uguu\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Upload\/Up1\">Up1<\/a> - Client-side Encrypted Image Host. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/u.pste.pw\" rel=\"nofollow\">uPste<\/a> - Private file hosting application with an emphasis on serving technology communities. (<a href=\"https:\/\/github.com\/TheReverend403\/uPste\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.youtransfer.io\" rel=\"nofollow\">YouTransfer<\/a> - YouTransfer is a simple but elegant self-hosted file transfer and sharing solution. (<a href=\"http:\/\/demo.youtransfer.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/remie\/YouTransfer\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><p><em>Command-line file upload<\/em><\/p>\n<ul><li><a href=\"https:\/\/github.com\/dsx\/beauties\">Beauties<\/a> - Minimalist file sharing written in Go, to be used primarily from Unix shell (e.g. with curl). Can be built as a Debian package for easy install. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/transfer.sh\" rel=\"nofollow\">transfer.sh<\/a> - Easy file sharing from the command line. (<a href=\"https:\/\/github.com\/dutchcoders\/transfer.sh\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<\/ul><h4><a id=\"user-content-web-based-file-managers\" class=\"anchor\" aria-hidden=\"true\" href=\"#web-based-file-managers\"><\/a>Web based file managers<\/h4>\n<ul><li><a href=\"https:\/\/oupala.github.io\/apaxy\/\" rel=\"nofollow\">Apaxy<\/a> - Theme built to enhance the experience of browsing web directories, using the mod_autoindex Apache module and some CSS to override the default style of a directory listing. (<a href=\"https:\/\/github.com\/AdamWhitcroft\/Apaxy\">Source Code<\/a>) <code>Unlicense<\/code> <code>HTML<\/code><\/li>\n<li><a href=\"http:\/\/www.directorylister.com\/\" rel=\"nofollow\">DirectoryLister<\/a> - Simple PHP based directory lister that lists a directory and all it's sub-directories and allows you to navigate there within. (<a href=\"https:\/\/github.com\/DirectoryLister\/DirectoryLister\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/encode-explorer.siineiolekala.net\/\" rel=\"nofollow\">Encode Explorer<\/a> - Encode Explorer is a single page file browser, it is simple and functional. (<a href=\"http:\/\/encode-explorer.siineiolekala.net\/explorer\/index.php\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/marekrei\/encode-explorer\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/soyuka\/explorer\">explorer<\/a> - Highly-configurable directory listing made with nodejs. (<a href=\"https:\/\/github.com\/soyuka\/explorer\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/filebrowser.github.io\/\" rel=\"nofollow\">filemanager<\/a> - Web File Manager which can be used as a middleware or standalone app. (<a href=\"https:\/\/github.com\/filebrowser\/filebrowser\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go\/VueJS<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/xataz\/gobrowser\">goBrowser<\/a> - Simple http file browser. (<a href=\"https:\/\/github.com\/xataz\/gobrowser\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/larsjung.de\/h5ai\/\" rel=\"nofollow\">h5ai<\/a> - Modern file indexer for HTTP web servers with focus on your files. Directories are displayed in a appealing way and browsing them is enhanced by different views, a breadcrumb and a tree overview. (<a href=\"https:\/\/larsjung.de\/h5ai\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/lrsjng\/h5ai\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/misterunknown\/ifm\/\">IFM<\/a> - Single script file manager. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.resourcespace.com\" rel=\"nofollow\">ResourceSpace<\/a> - ResourceSpace open source digital asset management software is the simple, fast, and free way to organise your digital assets. (<a href=\"https:\/\/www.resourcespace.com\/trial\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/www.resourcespace.com\/svn\" rel=\"nofollow\">Source Code<\/a>) <code>Other<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/sprut.io\" rel=\"nofollow\">Sprut.io<\/a> - 2 panel file manager with drag and drop features, code editor, text search, hotkeys. (<a href=\"https:\/\/demo.sprut.io:9443\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/LTD-Beget\/sprutio\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/nebulade\/surfer\">Surfer<\/a> - Simple static file server with webui to manage files. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.tagspaces.org\/\" rel=\"nofollow\">TagSpaces<\/a> - TagSpaces is an offline, cross-platform file manager and organiser that also can function as a note taking app. The WebDAV version of the application can be installed on top of a WebDAV servers such as Nextcloud or ownCloud. (<a href=\"http:\/\/demo.tagspaces.org\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/tagspaces\/tagspaces\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Javascript<\/code><\/li>\n<\/ul><h2><a id=\"user-content-games\" class=\"anchor\" aria-hidden=\"true\" href=\"#games\"><\/a>Games<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/doublespeakgames\/adarkroom\">A Dark Room<\/a> - Minimalist text adventure game for your browser. (<a href=\"http:\/\/adarkroom.doublespeakgames.com\/\" rel=\"nofollow\">Demo<\/a>) <code>MPL-2.0<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/huytd\/agar.io-clone\">Agar.IO Clone<\/a> - Agar.io clone written with Socket.IO and HTML5 canvas. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jbreindel\/battlecraft\">battlecraft<\/a> - Fully distributed multiplayer browser game. (<a href=\"http:\/\/battlecraft.online\" rel=\"nofollow\">Demo<\/a>) <code>Apache-2.0<\/code> <code>Erlang<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ellisonleao\/clumsy-bird\">Clumsy Bird<\/a> - MelonJS port of the famous Flappy Bird Game. (<a href=\"http:\/\/www.ellison.rocks\/clumsy-bird\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/play.elevatorsaga.com\/\" rel=\"nofollow\">elevatorsaga<\/a> - The elevator programming game. (<a href=\"https:\/\/github.com\/magwo\/elevatorsaga\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Hextris\/hextris\">Hextris<\/a> - Fast paced HTML5 puzzle game inspired by Tetris. (<a href=\"http:\/\/hextris.io\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/lichess.org\/\" rel=\"nofollow\">Lila<\/a> - The forever free, adless and open source chess server powering lichess.org, with official iOS and Android client apps. (<a href=\"https:\/\/github.com\/ornicar\/lila\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Scala<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/abrenaut\/posio\/\">Posio<\/a> - Geography multiplayer game. (<a href=\"https:\/\/posio.abrenaut.com\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/sbpp.github.io\" rel=\"nofollow\">SourceBans++<\/a> - Admin, ban, and communication management system for games running on the Source engine. (<a href=\"https:\/\/github.com\/sbpp\/sourcebans-pp\">Source Code<\/a>) <code>CC-BY-SA-4.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/mpcovcd\/spyfall\">Spyfall<\/a> - Fan made web version of the Spyfall party game. (<a href=\"http:\/\/spyfall.crabhat.com\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>HTML\/Javascript<\/code><\/li>\n<li><a href=\"http:\/\/tournamentmango.com\/\" rel=\"nofollow\">TournamentMango<\/a> - TournamentMango is an open source tournament bracket and user management system. You can build an archive of players and keep track of all their scores over time as well as their regular characters, games, and aliases. (<a href=\"https:\/\/github.com\/seiyria\/tournamentmango\">Source Code<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<\/ul><h2><a id=\"user-content-gateways\" class=\"anchor\" aria-hidden=\"true\" href=\"#gateways\"><\/a>Gateways<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/liftoffsoftware.com\/Products\/GateOne\" rel=\"nofollow\">GateOne<\/a> - Gate One is an HTML5 web-based terminal emulator and SSH client. (<a href=\"https:\/\/github.com\/liftoff\/GateOne\/\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/guac-dev.org\/\" rel=\"nofollow\">Guacamole<\/a> - Guacamole is a clientless remote desktop gateway. It supports standard protocols like VNC and RDP. (<a href=\"https:\/\/github.com\/glyptodon\/\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java\/C<\/code><\/li>\n<li><a href=\"https:\/\/oneye-project.org\/\" rel=\"nofollow\">oneye<\/a> - Cloud software to access your data from everywhere with any browser. (<a href=\"https:\/\/wiki.oneye-project.org\/0.9:demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/oneye\/oneye\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/os.js.org\/\" rel=\"nofollow\">OS.js<\/a> - Desktop implementation for your browser with a fully-fledged window manager, Application APIs, GUI toolkits and filesystem abstraction. (<a href=\"https:\/\/demo.os-js.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/os-js\/OS.js\">Source Code<\/a>) <code>BSD<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h2><a id=\"user-content-groupware\" class=\"anchor\" aria-hidden=\"true\" href=\"#groupware\"><\/a>Groupware<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/www.citadel.org\/doku.php\" rel=\"nofollow\">Citadel<\/a> - Groupware including email, calendar\/scheduling, address books, forums, mailing lists, IM, wiki and blog engines, RSS aggregation and more. (<a href=\"http:\/\/www.citadel.org\/doku.php\/installation:source\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/cozy.io\" rel=\"nofollow\">Cozy Cloud<\/a> - Personal cloud where you can read your emails or manage and sync your contact, files or calendars, with an app store full of community contributions. (<a href=\"https:\/\/github.com\/cozy\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.egroupware.org\/\" rel=\"nofollow\">egroupware<\/a> - Software suite including calendars, address books, notepad, project management tools, client relationship management tools (CRM), knowledge management tools, a wiki and a CMS. (<a href=\"https:\/\/github.com\/EGroupware\/egroupware\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.horde.org\/\" rel=\"nofollow\">Horde<\/a> - The Horde Project is about creating high quality Open Source applications and libraries, based on PHP and the Horde Framework. (<a href=\"http:\/\/demo.horde.org\/login.php\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/horde\/horde\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.honestrepair.net\/index.php\/hrcloud2\/\" rel=\"nofollow\">HRCloud2<\/a> - Full-featured home hosted Cloud Drive, Personal Assistant, App Launcher, File Converter, Streamer, Share Tool and more. (<a href=\"https:\/\/github.com\/zelon88\/HRCloud2\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/kolab.org\/\" rel=\"nofollow\">Kolab<\/a> - Kolab community is a unified communication and collaboration system. (<a href=\"https:\/\/git.kolab.org\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0\/LGPL-2.1\/GPL-3.0<\/code> <code>C++\/Python\/PHP<\/code><\/li>\n<li><a href=\"https:\/\/kopano.com\/\" rel=\"nofollow\">Kopano<\/a> - Groupware suite including e-mail, calendars, tasks, todos and notes. Featuring a modern WebApp, DeskApp and mobile access over Z-Push\/ActiveSync. (<a href=\"http:\/\/demo.kopano.com\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/stash.kopano.io\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>C\/Python\/PHP<\/code><\/li>\n<li><a href=\"https:\/\/openmeetings.apache.org\/index.html\" rel=\"nofollow\">Openmeetings<\/a> - Openmeetings provides video conferencing, instant messaging, white board, collaborative document editing and other groupware tools using API functions of the Red5 Streaming Server for Remoting and Streaming. (<a href=\"https:\/\/openmeetings.apache.org\/scm.html\" rel=\"nofollow\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/sogo.nu\/\" rel=\"nofollow\">SOGo<\/a> - SOGo offers multiple ways to access the calendaring and messaging data. CalDAV, CardDAV, GroupDAV, as well as ActiveSync, including native Outlook compatibility and Web interface. (<a href=\"http:\/\/demo.sogo.nu\/SOGo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/inverse-inc\/sogo\">Source Code<\/a>) <code>LGPL-2.1<\/code> <code>Objective-C<\/code><\/li>\n<li><a href=\"http:\/\/www.suitecrm.com\/\" rel=\"nofollow\">SuiteCRM<\/a> - The award-winning, enterprise-class open source CRM. (<a href=\"https:\/\/github.com\/salesagility\/SuiteCRM\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.tine20.org\" rel=\"nofollow\">Tine 2.0<\/a> - Contacts, Calendar, Tasks, WebDAV, ActiveSync, VOIP, Mail-Client, CRM, Sales, Projects, Timetracker. (<a href=\"https:\/\/demo.tine20.net\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/packages.tine20.com\/maintenance\/source\/\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0\/Other<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.zimbra.com\/\" rel=\"nofollow\">Zimbra Collaboration<\/a> - Email, calendar, collaboration server with Web interface and lots of integrations. (<a href=\"https:\/\/github.com\/zimbra\">Source Code<\/a>) <code>GPL-2.0\/CPAL-1.0<\/code> <code>Java<\/code><\/li>\n<\/ul><h2><a id=\"user-content-human-resources-management-hrm\" class=\"anchor\" aria-hidden=\"true\" href=\"#human-resources-management-hrm\"><\/a>Human Resources Management (HRM)<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.admidio.org\/\" rel=\"nofollow\">admidio<\/a> - Admidio is a free open source user management system for websites of organizations and groups. The system has a flexible role model so that it\u2019s possible to reflect the structure and permissions of your organization. (<a href=\"https:\/\/www.admidio.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Admidio\/admidio\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/icehrm.com\/\" rel=\"nofollow\">IceHrm<\/a> - IceHrm employee management system allows companies to centralize confidential employee information. (<a href=\"https:\/\/icehrm.com\/demo.php\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/gamonoid\/icehrm\/\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.orangehrm.com\/\" rel=\"nofollow\">OrangeHRM<\/a> - OrangeHRM is a comprehensive HRM system that captures all the essential functionalities required for any enterprise. (<a href=\"https:\/\/sourceforge.net\/projects\/orangehrm\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.sentrifugo.com\/\" rel=\"nofollow\">Sentrifugo<\/a> - Sentrifugo is a HRM system that can be easily configured to meet your organizational needs. (<a href=\"https:\/\/github.com\/sapplica\/sentrifugo\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/timeoff.management\" rel=\"nofollow\">TimeOff.Management<\/a> - Simple yet powerful absence management software for small and medium size business. (<a href=\"https:\/\/app.timeoff.management\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/timeoff-management\/application\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h2><a id=\"user-content-internet-of-things-iot\" class=\"anchor\" aria-hidden=\"true\" href=\"#internet-of-things-iot\"><\/a>Internet Of Things (IoT)<\/h2>\n<ul><li><a href=\"http:\/\/iotdashboard.pythonanywhere.com\/en\/\" rel=\"nofollow\">Iotdashboard<\/a> - Fast Django server for IOT Devices. (<a href=\"https:\/\/github.com\/electrocoder\/iotdashboard\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>JavaScript<\/code><\/li>\n<li><a href=\"https:\/\/www.devicehive.com\/\" rel=\"nofollow\">DeviceHive<\/a> - Open Source IoT Plaform with a wide range of integration options. (<a href=\"https:\/\/playground.devicehive.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/devicehive\/devicehive-java-server\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/thingsboard.io\/\" rel=\"nofollow\">Thingsboard<\/a> - Open-source IoT Platform - Device management, data collection, processing and visualization. (<a href=\"https:\/\/demo.thingsboard.io\/signup\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/thingsboard\/thingsboard\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/thingspeak.com\/\" rel=\"nofollow\">Thingspeak<\/a> - Open source \u201cInternet of Things\u201d application and API to store and retrieve data from things using HTTP. (<a href=\"https:\/\/thingspeak.com\/channels\/public\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/iobridge\/thingspeak\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<\/ul><h2><a id=\"user-content-learning-and-courses\" class=\"anchor\" aria-hidden=\"true\" href=\"#learning-and-courses\"><\/a>Learning and Courses<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/www.canvaslms.com\/\" rel=\"nofollow\">Canvas LMS<\/a> - Canvas is the trusted, open-source learning management system (LMS) that is revolutionizing the way we educate. (<a href=\"https:\/\/canvas.instructure.com\/register\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/instructure\/canvas-lms\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/chamilo.org\/\" rel=\"nofollow\">Chamilo LMS<\/a> - Chamilo LMS allows you to create a virtual campus for the provision of online or semi-online training. (<a href=\"https:\/\/github.com\/chamilo\/chamilo-lms\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.edx.org\/\" rel=\"nofollow\">edX<\/a> - The Open edX platform is open-source code that powers <a href=\"https:\/\/www.edx.org\/\" rel=\"nofollow\">edX.org<\/a>. (<a href=\"https:\/\/github.com\/edx\/\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.ilias.de\" rel=\"nofollow\">ILIAS<\/a> - ILIAS is the Learning Management System that can cope with anything you throw at it. (<a href=\"http:\/\/demo.ilias.de\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ILIAS-eLearning\/ILIAS\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.lxhive.com\/\" rel=\"nofollow\">lxHive<\/a> - Open Source ExperienceAPI compliant Learning Record Store (LRS) - previously code-named TinCanAPI. (<a href=\"https:\/\/github.com\/Brightcookie\/lxHive\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/mahara.org\/\" rel=\"nofollow\">Mahara<\/a> - Open Source fully featured web application to build students electronic portfolio. (<a href=\"https:\/\/github.com\/MaharaProject\/mahara\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/moodle.org\/\" rel=\"nofollow\">Moodle<\/a> - Moodle is a learning and courses platform with one of the largest open source communities worldwide. (<a href=\"https:\/\/moodle.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/git.moodle.org\/gw\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.openeclass.org\/\" rel=\"nofollow\">Open eClass<\/a> - Open eClass is an advanced e-learning solution that can enhance the teaching and learning process. (<a href=\"http:\/\/demo.openeclass.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/gunet\/openeclass\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/documen.tician.de\/relate\/\" rel=\"nofollow\">RELATE<\/a> - RELATE is a web-based courseware package, includes features such as: flexible rules, statistics, multi-course support, class calendar. (<a href=\"https:\/\/github.com\/inducer\/relate\/\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.sakaiproject.org\/\" rel=\"nofollow\">Sakai<\/a> - The Sakai project provides a flexible and feature-rich environment for teaching, learning, research and other collaboration. (<a href=\"https:\/\/www.sakaiproject.org\/try-sakai\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/sakaiproject\/sakai\">Source Code<\/a>) <code>ECL-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/schooltool.org\/\" rel=\"nofollow\">SchoolTool<\/a> - SchoolTool is free administrative software for schools. It includes demographics, gradebook, attendance, calendaring, reporting and more for primary and secondary schools. (<a href=\"http:\/\/bazaar.launchpad.net\/~schooltool-owners\/schooltool\/2.8\/files\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-maps-and-global-positioning-system-gps\" class=\"anchor\" aria-hidden=\"true\" href=\"#maps-and-global-positioning-system-gps\"><\/a>Maps and Global Positioning System (GPS)<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/graphhopper.com\/\" rel=\"nofollow\">Graphhopper<\/a> - Fast routing library and server using OpenStreetMap. (<a href=\"https:\/\/github.com\/graphhopper\/graphhopper\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/MapBBCode\/share.mapbbcode.org\">MapBBCodeShare<\/a> - Tool for sharing custom OSM maps. Support for annotated markers, polygons, lines, multi-format import\/export, multiple layers, shortlinks. (<a href=\"http:\/\/share.mapbbcode.org\/\" rel=\"nofollow\">Demo<\/a>) <code>WTFPL\/Other<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.opengts.org\/\" rel=\"nofollow\">OpenGTS<\/a> - Entry-level fleet tracking system. Supports variety of tracking devices and protocols. Comes with rich web-interface and reporting features. (<a href=\"http:\/\/track.opengts.org\/track\/Track\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/sourceforge.net\/projects\/opengts\/files\/server-base\/\" rel=\"nofollow\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/www.openstreetmap.org\/\" rel=\"nofollow\">OpenStreetMap<\/a> - OpenStreetMap is a map of the world, created by people like you and free to use under an open license. (<a href=\"https:\/\/github.com\/openstreetmap\/openstreetmap-website\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/LINKIWI\/orion-web\">Orion<\/a> - Powerful OwnTracks API-compliant location data visualization frontend for the web. (<a href=\"https:\/\/linkiwi.github.io\/orion-web\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Python\/Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/owntracks\/recorder\">OwnTracks Recorder<\/a> <code>\u26a0<\/code> - Store and access data published by <a href=\"http:\/\/owntracks.org\/\" rel=\"nofollow\">OwnTracks<\/a> location tracking apps. <code>GPL-2.0<\/code> <code>C<\/code>\/<code>Lua<\/code><\/li>\n<li><a href=\"http:\/\/tileserver.readthedocs.io\/\" rel=\"nofollow\">TileServer GL<\/a> - Vector and raster maps with GL styles. Server side rendering by Mapbox GL Native. Map tile server for Mapbox GL JS, Android, iOS, Leaflet, OpenLayers, GIS via WMTS, etc. (<a href=\"https:\/\/github.com\/klokantech\/tileserver-gl\">Source Code<\/a>) <code>Multiple<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/klokantech\/tileserver-php\">TileServer PHP<\/a> - Serve map tiles from any PHP hosting. <code>BSD<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.traccar.org\/\" rel=\"nofollow\">Traccar<\/a> - Java application to track GPS positions. Supports loads of tracking devices and protocols, has an Android and iOS App. Has a web interface to view your trips. (<a href=\"http:\/\/demo.traccar.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/tananaev?tab=repositories\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/umap.openstreetmap.fr\/en\/\" rel=\"nofollow\">uMap<\/a> - Create maps with OpenStreetMap layers in a minute and embed them in your site. (<a href=\"https:\/\/github.com\/umap-project\/umap\">Source Code<\/a>) <code>WTFPL<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-media-streaming\" class=\"anchor\" aria-hidden=\"true\" href=\"#media-streaming\"><\/a>Media Streaming<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See also <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_streaming_media_systems\" rel=\"nofollow\">https:\/\/en.wikipedia.org\/wiki\/List_of_streaming_media_systems<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Comparison_of_streaming_media_systems\" rel=\"nofollow\">https:\/\/en.wikipedia.org\/wiki\/Comparison_of_streaming_media_systems<\/a><\/p>\n<h3><a id=\"user-content-multimedia-streaming\" class=\"anchor\" aria-hidden=\"true\" href=\"#multimedia-streaming\"><\/a>Multimedia Streaming<\/h3>\n<ul><li><a href=\"https:\/\/macosforge.github.io\/dss\/\" rel=\"nofollow\">Darwin Streaming Server<\/a> - High performance server for streaming QuickTime and MPEG-4 media over RTP and RTSP protocols. Originated as Apple\u2019s QTSS. (<a href=\"https:\/\/github.com\/macosforge\/dss\">Source Code<\/a>) <code>APSL-2.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"http:\/\/flumotion.net\/\" rel=\"nofollow\">Flumotion<\/a> - Streaming media server with intuitive graphical administration tools. (<a href=\"https:\/\/github.com\/Flumotion\/flumotion\">Source Code<\/a>) <code>LGPL-2.1<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ridhwaans\/homehost\">homehost<\/a> <code>\u26a0<\/code> - Self-hosted React + Redux app that streams your media collection (music, movies, books, podcasts, comics etc). <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.icecast.org\/\" rel=\"nofollow\">Icecast 2<\/a> - streaming audio\/video server which can be used to create an Internet radio station or a privately running jukebox and many things in between. (<a href=\"https:\/\/git.xiph.org\/?p=icecast-server.git;a=summary\" rel=\"nofollow\">Source Code<\/a>, <a href=\"http:\/\/www.icecast.org\/apps\/\" rel=\"nofollow\">Clients<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/mistserver.org\/\" rel=\"nofollow\">MistServer<\/a> - Streaming media server that works well in any streaming environment. (<a href=\"https:\/\/github.com\/DDVTECH\/mistserver\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"http:\/\/sourceforge.net\/projects\/minidlna\/\" rel=\"nofollow\">ReadyMedia<\/a> - Simple media server software, with the aim of being fully compliant with DLNA\/UPnP-AV clients. Formerly known as MiniDLNA. (<a href=\"https:\/\/sourceforge.net\/p\/minidlna\/git\/ci\/master\/tree\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/wiki.gnome.org\/action\/show\/Projects\/Rygel\" rel=\"nofollow\">Rygel<\/a> - Rygel is a UPnP AV MediaServer that allows you to easily share audio, video, and pictures. Media player software may use Rygel to become a MediaRenderer that may be controlled remotely by a UPnP or DLNA Controller. (<a href=\"https:\/\/gitlab.gnome.org\/GNOME\/rygel\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/u-wave.net\/\" rel=\"nofollow\">\u00fcWave<\/a> <code>\u26a0<\/code> - \u00fcWave is a self-hosted collaborative listening platform. Users take turns playing media\u2014songs, talks, gameplay videos, or anything else\u2014from a variety of media sources like YouTube and SoundCloud. (<a href=\"https:\/\/wlk.yt\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/u-wave\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h3><a id=\"user-content-audio-streaming\" class=\"anchor\" aria-hidden=\"true\" href=\"#audio-streaming\"><\/a>Audio Streaming<\/h3>\n<ul><li><a href=\"http:\/\/ampache.org\/\" rel=\"nofollow\">Ampache<\/a> - Web based audio\/video streaming application. (<a href=\"http:\/\/play.dogmazic.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ampache\/ampache\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/airsonic.github.io\/\" rel=\"nofollow\">Airsonic<\/a> - Open-source web-based media streamer and jukebox. A fork of Subsonic's last open-source release, before it switched licenses. (<a href=\"https:\/\/github.com\/airsonic\/airsonic\">Source Code<\/a>, <a href=\"https:\/\/airsonic.github.io\/docs\/apps\/\" rel=\"nofollow\">Clients<\/a>) <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/www.azuracast.com\/\" rel=\"nofollow\">AzuraCast<\/a> - A modern and accessible self-hosted web radio management suite. (<a href=\"https:\/\/github.com\/AzuraCast\/AzuraCast\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/beets.io\/\" rel=\"nofollow\">Beets<\/a> - Music library manager and MusicBrainz tagger (command-line and Web interface). (<a href=\"https:\/\/github.com\/beetbox\/beets\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.fomori.org\/cherrymusic\/\" rel=\"nofollow\">CherryMusic<\/a> - Minimalistic Web-Mediaplayer. (<a href=\"https:\/\/github.com\/devsnd\/cherrymusic\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jkbrzt\/cloudtunes\">cloudtunes<\/a> <code>\u26a0<\/code> - Web-based music player for the cloud. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/compactd\/compactd\">Compactd<\/a> - Remote music player that supports adding content. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/xouabita\/friends-radio\">FriendsRadio<\/a> <code>\u26a0<\/code> - Share music with your friends from Youtube and Soundcloud. (<a href=\"https:\/\/friends-radio.herokuapp.com\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/funkwhale.audio\/\" rel=\"nofollow\">Funkwhale<\/a> - A modern, web-based, convivial, multi-user and free music server. (<a href=\"https:\/\/demo.funkwhale.audio\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/code.eliotberriot.com\/funkwhale\/funkwhale\" rel=\"nofollow\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python\/Django<\/code><\/li>\n<li><a href=\"https:\/\/gnu.io\/fm\/\" rel=\"nofollow\">GNU FM<\/a> - Running music community websites, alternative to last.fm. (<a href=\"http:\/\/git.savannah.gnu.org\/cgit\/librefm.git\/\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/andrewrk\/groovebasin\">Groove Basin<\/a> - Music player server with a web-based user interface inspired by Amarok 1.4. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.icecast.org\/\" rel=\"nofollow\">Icecast 2<\/a> - streaming audio\/video server which can be used to create an Internet radio station or a privately running jukebox and many things in between. (<a href=\"https:\/\/git.xiph.org\/?p=icecast-server.git;a=summary\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/koel.phanan.net\/\" rel=\"nofollow\">koel<\/a> - Personal music streaming server that works. (<a href=\"https:\/\/github.com\/phanan\/koel\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/libretime.org\" rel=\"nofollow\">LibreTime<\/a> - Simple, open source platform that lets you broadcast streaming radio on the web (fork of <a href=\"https:\/\/github.com\/sourcefabric\/Airtime\">Airtime<\/a>). (<a href=\"https:\/\/github.com\/LibreTime\/libretime\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/mopidy.readthedocs.org\/\" rel=\"nofollow\">Mopidy<\/a> - Extensible music server. Offers a superset of the mpd API, as well as integration with 3rd party services like Spotify, SoundCloud etc. (<a href=\"https:\/\/github.com\/mopidy\/mopidy\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code>\n<ul><li><a href=\"https:\/\/github.com\/martijnboland\/moped\">Moped<\/a> - Responsive HTML5 + Javascript client for the Mopidy music server. <code>MIT<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pimusicbox\/mopidy-musicbox-webclient\">Mopidy MusicBox<\/a> - Web Client for Mopidy Music Server. <code>Apache-2.0<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Lesterpig\/mopidy-party\">Mopidy-Party<\/a> - Mopidy web extension designed for party! Let your guests manage the sound. <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<\/ul><\/li>\n<li><a href=\"http:\/\/www.musicpd.org\/\" rel=\"nofollow\">mpd<\/a> - Daemon to remotely play music, stream music, handle and organize playlists. Many clients available. (<a href=\"https:\/\/github.com\/MusicPlayerDaemon\/MPD\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C++<\/code>\n<ul><li><a href=\"http:\/\/www.ympd.org\/\" rel=\"nofollow\">ympd<\/a> - Standalone MPD Web GUI written in C, utilizing Websockets and Bootstrap\/JS. (<a href=\"https:\/\/github.com\/notandy\/ympd\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<\/ul><\/li>\n<li><a href=\"http:\/\/mstream.io\/\" rel=\"nofollow\">mStream<\/a> - Music streaming server with GUI management tools. Runs on Mac, Windows, and Linux. (<a href=\"https:\/\/darncoyotes.mstream.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/IrosTheBeggar\/mStream\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/henrik242\/musicbrowser\">Music Browser<\/a> - Browser and streamer for your music collection. It is runs on most operating systems, and is light enough to run flawlessly on NAS devices. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/musikcube.com\/\" rel=\"nofollow\">musikcube<\/a> - Streaming audio server with Linux\/macOS\/Windows\/Android clients. (<a href=\"https:\/\/github.com\/clangen\/musikcube\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/www.sonerezh.bzh\/\" rel=\"nofollow\">Sonerezh<\/a> - Self-hosted, web-based application for stream your music, everywhere. (<a href=\"https:\/\/www.sonerezh.bzh\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/sonerezh\/sonerezh\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/volumio.org\/\" rel=\"nofollow\">Volumio<\/a> - A free and open source linux distribution, designed and fine-tuned exclusively for music playback. (<a href=\"https:\/\/github.com\/volumio\/Volumio2\">Source Code<\/a>) <code>GPLv3<\/code> <code>NodeJS<\/code><\/li>\n<\/ul><h3><a id=\"user-content-video-streaming\" class=\"anchor\" aria-hidden=\"true\" href=\"#video-streaming\"><\/a>Video Streaming<\/h3>\n<ul><li><a href=\"https:\/\/packages.debian.org\/stable\/crtmpserver\" rel=\"nofollow\">crtmpserver<\/a> - High performance RTMP\/RTSP streaming server. <code>GPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/calzoneman\/sync\">CyTube<\/a> - CyTube is a web application providing media synchronization, chat, and more for an arbitrary number of channels. (<a href=\"http:\/\/cytu.be\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Hellowlol\/HTPC-Manager\">Hellowlol HTPC Manager fork<\/a> - Fully responsive interface to manage all your favorite media on your HTPC. (<a href=\"https:\/\/github.com\/styxit\/HTPC-Manager\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.icecast.org\/\" rel=\"nofollow\">Icecast 2<\/a> - streaming audio\/video server which can be used to create an Internet radio station or a privately running jukebox and many things in between. (<a href=\"https:\/\/git.xiph.org\/?p=icecast-server.git;a=summary\" rel=\"nofollow\">Source Code<\/a>, <a href=\"https:\/\/www.icecast.org\/apps\/\" rel=\"nofollow\">Clients<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/mistserver.org\/\" rel=\"nofollow\">MistServer<\/a> - Streaming media server that works well in any streaming environment. (<a href=\"https:\/\/github.com\/DDVTECH\/mistserver\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pastapojken\/Myflix\">Myflix<\/a> - Self-hosted, super lightweight Netflix alternative. <code>MIT<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"http:\/\/oddnetworks.com\" rel=\"nofollow\">Odd Networks<\/a> - Open source video management system (VMS) and API for collections and videos with supporting SDKs for Roku, Apple iOS\/tvOS, Google Android, and Amazon FireTV. (<a href=\"https:\/\/github.com\/oddnetworks\/oddworks\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/joinpeertube.org\/en\/\" rel=\"nofollow\">PeerTube<\/a> - Decentralized video streaming platform using P2P (BitTorrent) directly in the web browser. (<a href=\"https:\/\/github.com\/Chocobozzz\/PeerTube\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/datarhei.github.io\/restreamer\/\" rel=\"nofollow\">Restreamer<\/a> - Restreamer allows you to do h.264 real-time video streaming on your website without a streaming provider. (<a href=\"https:\/\/github.com\/datarhei\/restreamer\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs\/Docker<\/code><\/li>\n<li><a href=\"https:\/\/moeiscool.github.io\/Shinobi\/\" rel=\"nofollow\">Shinobi<\/a> - Open Source CCTV software written in Node with both IP and local camera support. (<a href=\"https:\/\/github.com\/moeiscool\/Shinobi\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/streamaserver.org\/\" rel=\"nofollow\">Streama<\/a> - Self hosted streaming media server. (<a href=\"https:\/\/github.com\/streamaserver\/streama\">Source Code<\/a>) <code>MIT<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/www.videolan.org\/\" rel=\"nofollow\">VideoLAN Client (VLC)<\/a> - Cross-platform multimedia player client and server supporting most multimedia files as well as DVDs, Audio CDs, VCDs, and various streaming protocols. (<a href=\"https:\/\/github.com\/videolan\/vlc\">Source Code<\/a>) <code>Multiple<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/www.zoneminder.com\/\" rel=\"nofollow\">Zoneminder<\/a> - Closed-circuit television (CCTV) software application which supports IP, USB and Analog cameras. (<a href=\"https:\/\/github.com\/ZoneMinder\/ZoneMinder\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-miscother\" class=\"anchor\" aria-hidden=\"true\" href=\"#miscother\"><\/a>Misc\/Other<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/demo.fouroneone.io\/\" rel=\"nofollow\">411<\/a> - Alert Management Web Application. (<a href=\"https:\/\/github.com\/etsy\/411\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Ardakilic\/alerthub\">AlertHub<\/a> <code>\u26a0<\/code> - AlertHub is a simple tool to get alerted from GitHub releases. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/anchr.io\" rel=\"nofollow\">Anchr<\/a> - Anchr is a toolbox for tiny tasks on the internet, including bookmark collections, URL shortening and (encrypted) image uploads. (<a href=\"https:\/\/github.com\/n1try\/anchr\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.anuko.com\/time_tracker\/index.htm\" rel=\"nofollow\">Anuko<\/a> - Anuko provides simple time and project tracking on a selfhosted basis. (<a href=\"https:\/\/timetracker.anuko.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/anuko\/timetracker\">Source Code<\/a>) <code>Other<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/asciiflow.com\/\" rel=\"nofollow\">asciiflow<\/a> - Flow Diagram Drawing Tool. (<a href=\"https:\/\/github.com\/lewish\/asciiflow2\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Java\/JavaScript<\/code><\/li>\n<li><a href=\"http:\/\/www.blynk.cc\/\" rel=\"nofollow\">blynk<\/a> - Platform with iOs and Android apps to control Arduino, ESP8266, Raspberry Pi and similar microcontroller boards over the Internet. (<a href=\"https:\/\/github.com\/blynkkk\/blynk-server\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/www.cups.org\/\" rel=\"nofollow\">CUPS<\/a> - The Common Unix Print System uses Internet Printing Protocol (IPP) to support printing to local and network printers. (<a href=\"https:\/\/github.com\/apple\/cups\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/domainmod.org\" rel=\"nofollow\">DomainMOD<\/a> - Application to manage your domains and other internet assets in a central location. DomainMOD includes a Data Warehouse framework that allows you to import your WHM\/cPanel web server data so that you can view, export, and report on your data. (<a href=\"https:\/\/demo.domainmod.org\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/domainmod\/domainmod\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/easybook-project.org\/\" rel=\"nofollow\">EasyBook Project<\/a> - Book publishing as easy as it should be. (<a href=\"https:\/\/github.com\/javiereguiluz\/easybook\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/devfake\/flox\">Flox<\/a> <code>\u26a0<\/code> - Self hosted movie, TV series and anime watch list with a 3-point rating system. Uses The Movie Database backend for information. (<a href=\"http:\/\/80.240.132.120\/flox-demo\/public\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/formspree.io\/\" rel=\"nofollow\">formspree<\/a> <code>\u26a0<\/code> - Just send your form to our URL and we'll forward it to your email. No PHP, Javascript or sign up required. (<a href=\"http:\/\/testformspree.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/formspree\/formspree\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/geneweb.tuxfamily.org\/wiki\/GeneWeb\" rel=\"nofollow\">GeneWeb<\/a> - GeneWeb is an open source genealogy software written in OCaml. It comes with a Web interface and can be used off-line or as a Web service. (<a href=\"https:\/\/demo.geneweb.tuxfamily.org\/gw7\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/geneweb\/geneweb\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>OCAML<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/howsecureismypassword\/hsimp\">How Secure Is My Password<\/a> - Rather than just saying a password is \"weak\" or \"strong\", How Secure is My Password? lets your users know how long it would take someone to crack their password. (<a href=\"https:\/\/howsecureismypassword.net\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/majodev\/google-webfonts-helper\">google-webfonts-helper<\/a> <code>\u26a0<\/code> - Hassle-Free Way to Self-Host Google Fonts. Get eot, ttf, svg, woff and woff2 files + CSS snippets. (<a href=\"https:\/\/google-webfonts-helper.herokuapp.com\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Algram\/ytdl-webserver\">ytdl-webserver<\/a> - Docker-ready webserver for downloading youtube videos. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.kimai.org\/\" rel=\"nofollow\">Kimai<\/a> - Simple time and project tracking. (<a href=\"http:\/\/www.kimai.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/kimai\/kimai\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/securestate\/king-phisher\">King Phisher<\/a> - King Phisher is a tool for testing and promoting user awareness by simulating real world phishing attacks. <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jlelse\/Maily-Form\">Maily Form<\/a> - Self-hosted service you can use to place forms on static sites. It uses nodemailer and you can host it with Docker. <code>GPL-3.0<\/code> <code>Nodejs\/Docker<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/drichard\/mindmaps\">Mindmaps<\/a> - Open source, offline capable, mind mapping application. (<a href=\"http:\/\/drichard.org\/mindmaps\/\" rel=\"nofollow\">Demo<\/a>) <code>AGPL-3.0<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/monicahq.com\/\" rel=\"nofollow\">Monica<\/a> - Personal relationship manager, and a new kind of CRM to organize interactions with your friends and family. (<a href=\"https:\/\/github.com\/monicahq\/monica\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/musical-artifacts.com\/\" rel=\"nofollow\">Musical Artifacts<\/a> - Helping to catalog, preserve and free the artifacts you need to produce music. (<a href=\"https:\/\/github.com\/lfzawacki\/musical-artifacts\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ondras\/my-mind\">My Mind<\/a> - Web application for creating and managing mind maps. (<a href=\"https:\/\/my-mind.github.io\/?url=examples%2Ffeatures.mymind\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/nnmm.nl\/\" rel=\"nofollow\">nnmm<\/a> - Super tiny pastebin\/url minifier \"microservice\". (<a href=\"https:\/\/github.com\/Mechazawa\/nnmm\">Source Code<\/a>) <code>Beerware<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/notica.us\" rel=\"nofollow\">Notica<\/a> - Lets you send browser notifications from your terminal to your desktop or phone. No installation or registration is required. (<a href=\"https:\/\/github.com\/tannercollin\/Notica\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/ombi.io\/\" rel=\"nofollow\">Ombi<\/a> - A content request system for Plex\/Emby, connects to SickRage, CouchPotato, Sonarr, with a growing feature set. (<a href=\"https:\/\/demo.ombi.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/tidusjar\/Ombi\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/revealjs.com\" rel=\"nofollow\">revealjs<\/a> - Framework for easily creating beautiful presentations using HTML. (<a href=\"https:\/\/revealjs.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/hakimel\/reveal.js\">Source Code<\/a>) <code>MIT<\/code> <code>JavaScript<\/code><\/li>\n<li><a href=\"http:\/\/sane-project.org\/\" rel=\"nofollow\">SANE Network Scanning<\/a> - Allow remote clients to access image acquisition devices (scanners) available on the local host. (<a href=\"http:\/\/www.sane-project.org\/cvs.html\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/swordbeta\/trello-burndown\">Trello Burndown<\/a> <code>\u26a0<\/code> - Easy to use SCRUM burndown chart for Trello boards. <code>MIT<\/code> <code>Go\/Docker<\/code><\/li>\n<li><a href=\"https:\/\/ulterius.io\" rel=\"nofollow\">Ulterius<\/a> - Ulterius is an open-source remote desktop software with lots of awesome functions. (<a href=\"https:\/\/rainway.io\/register\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Ulterius\/server\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"http:\/\/www.vimbadmin.net\/\" rel=\"nofollow\">ViMbAdmin<\/a> - Provides a web based virtual mailbox administration system to allow mail administrators to easily manage domains, mailboxes and aliases. (<a href=\"http:\/\/www.vimbadmin.net\/demo\/auth\/login\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/opensolutions\/ViMbAdmin\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/visualcaptcha.net\/\" rel=\"nofollow\">visualCaptcha<\/a> - Configurable captcha solution, focusing on accessibility and simplicity whilst maintaining security. (<a href=\"http:\/\/demo.visualcaptcha.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/emotionLoop\/visualCaptcha\">Source Code<\/a>) <code>MIT<\/code> <code>PHP\/Nodejs\/Ruby\/Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Finesse\/web-fonts-repository\">Web fonts repository<\/a> - A simple webfont hosting. Google Fonts alternative for your own fonts. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.webtrees.net\" rel=\"nofollow\">webtrees<\/a> - Webtrees is the web's leading on-line collaborative genealogy application. (<a href=\"https:\/\/dev.webtrees.net\/demo-stable\/index.php?ctype=gedcom&amp;ged=demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/fisharebest\/webtrees\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-money-budgeting-and-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#money-budgeting-and-management\"><\/a>Money, Budgeting and Management<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See also <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#it-asset-management\">https:\/\/github.com\/n1trux\/awesome-sysadmin#it-asset-management<\/a><\/p>\n<ul><li><a href=\"https:\/\/akaunting.com\/\" rel=\"nofollow\">Akaunting<\/a> - Akaunting is a free, online and open source accounting software designed for small businesses and freelancers. (<a href=\"https:\/\/github.com\/akaunting\/akaunting\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/arecker\/bennedetto\">Benedetto<\/a> - Bennedetto is a simple, turn-based budget management app. <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/paukiatwee\/budgetapp\/\">Budget App<\/a> - Budget App is an open source personal budgeting application. <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/www.dotledger.com\/\" rel=\"nofollow\">Dot Ledger<\/a> - Web-based personal finance management tool. (<a href=\"https:\/\/demo.dotledger.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/dotledger\/dotledger\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/www.economizzer.org\/\" rel=\"nofollow\">Economizzer<\/a> - An easy and secure system for you to manage your personal money and achieve your goals, and can be accessed by computer, tablet or smartphone. (<a href=\"https:\/\/github.com\/gugoan\/economizzer#live-demo\">Demo<\/a>, <a href=\"https:\/\/github.com\/gugoan\/economizzer\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/gaynetdinov\/ex_money\">ExMoney<\/a> - Self-hosted personal finance app. <code>ISC<\/code> <code>Elixir<\/code><\/li>\n<li><a href=\"https:\/\/firefly-iii.org\/\" rel=\"nofollow\">Firefly III<\/a> - Firefly III is a modern financial manager. It helps you to keep track of your money and make budget forecasts. It supports credit cards, has an advanced rule engine and can import data from many banks. It's powered by Laravel and requires PHP7.2. (<a href=\"https:\/\/demo.firefly-iii.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/firefly-iii\/firefly-iii\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/galette.eu\/dc\/\" rel=\"nofollow\">Galette<\/a> - Galette is a membership management web application towards non profit organizations. (<a href=\"http:\/\/git.tuxfamily.org\/galette\/galette.git\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/grr.devome.com\/?lang=en\" rel=\"nofollow\">GRR<\/a> - Assets management and booking for small\/medium companies. (<a href=\"https:\/\/github.com\/JeromeDevome\/GRR\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/hospitalrun.io\/\" rel=\"nofollow\">Hospital Run<\/a> - Hospital Run is offline enabled hospital management software. (<a href=\"https:\/\/github.com\/HospitalRun\/hospitalrun-server\">Source Code<\/a>, <a href=\"http:\/\/hospitalrun.io\/demo\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/inventaire.io\/welcome\" rel=\"nofollow\">Inventaire<\/a> - Collaborative resources mapper project, while yet only focused on exploring books mapping with wikidata and ISBNs. (<a href=\"https:\/\/github.com\/inventaire\/inventaire\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.invoiceninja.org\/\" rel=\"nofollow\">Invoice Ninja<\/a> - Powerful tool to invoice clients online. (<a href=\"https:\/\/app.invoiceninja.com\/invoices\/create\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/invoiceninja\/invoiceninja\">Source Code<\/a>) <code>AAL<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/invoiceplane.com\/\" rel=\"nofollow\">InvoicePlane<\/a> - Manage quotes, invoices, payments and customers for your small business. (<a href=\"https:\/\/demo.invoiceplane.com\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/InvoicePlane\/InvoicePlane\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/ihatemoney.org\/\" rel=\"nofollow\">IHateMoney<\/a> - Manage your shared expenses, easily. (<a href=\"https:\/\/ihatemoney.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/spiral-project\/ihatemoney\">Source Code<\/a>) <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/kresus.org\/\" rel=\"nofollow\">Kresus<\/a> - Open source personal finance manager. (<a href=\"https:\/\/kresus.org\/en\/demo.html\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/kresusapp\/kresus\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.partkeepr.org\" rel=\"nofollow\">PartKeepr<\/a> - PartKeepr is an electronic part inventory management software. It helps you to keep track of your available parts and assist you with re-ordering parts. (<a href=\"https:\/\/demo.partkeepr.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/partkeepr\/PartKeepr\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/silverstrike.org\/\" rel=\"nofollow\">SilverStrike<\/a> - Personal finance management made easy. (<a href=\"https:\/\/demo.silverstrike.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/agstrike\/silverstrike\">Source Code<\/a>) <code>MIT<\/code> <code>Python\/Django<\/code><\/li>\n<li><a href=\"https:\/\/dev.sigpipe.me\/dashie\/StockazNG\" rel=\"nofollow\">StockazNG<\/a> - Asset Management System. <code>MIT<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-monitoring\" class=\"anchor\" aria-hidden=\"true\" href=\"#monitoring\"><\/a>Monitoring<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#monitoring\">https:\/\/github.com\/n1trux\/awesome-sysadmin#monitoring<\/a> and <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#metric--metric-collection\">https:\/\/github.com\/n1trux\/awesome-sysadmin#metric--metric-collection<\/a><\/p>\n<h2><a id=\"user-content-note-taking-and-editors\" class=\"anchor\" aria-hidden=\"true\" href=\"#note-taking-and-editors\"><\/a>Note-taking and Editors<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em>See also <a href=\"#wikis\">Wikis<\/a><\/em><\/p>\n<ul><li><a href=\"https:\/\/bulletnotes.io\/\" rel=\"nofollow\">BulletNotes<\/a> - Workflowy \/ Dynalist clone with Kanban (Trello) and Calendar functionality. Organize everything. (<a href=\"https:\/\/gitlab.com\/NickBusey\/BulletNotes\" rel=\"nofollow\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/boostnote.io\/\" rel=\"nofollow\">Boostnote<\/a> - The note-taking app for programmers that focuses on markdown, snippets, and customizability. (<a href=\"https:\/\/github.com\/BoostIO\/Boostnote\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>JavaScript<\/code><\/li>\n<li><a href=\"https:\/\/demo.codimd.org\/\" rel=\"nofollow\">CodiMD<\/a> - Realtime collaborative markdown notes on all platforms, formerly HackMD CE. (<a href=\"https:\/\/github.com\/hackmdio\/codimd\/\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Node.js<\/code><\/li>\n<li><a href=\"http:\/\/dillinger.io\/\" rel=\"nofollow\">dillinger<\/a> - The last Markdown editor, ever. (<a href=\"https:\/\/github.com\/joemccann\/dillinger\/\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/draw.io\" rel=\"nofollow\">draw.io<\/a> - Diagram software for making flowcharts, process diagrams, org charts, UML, ER and network diagrams. (<a href=\"https:\/\/github.com\/jgraph\/drawio\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>JavaScript<\/code><\/li>\n<li><a href=\"https:\/\/laverna.cc\/\" rel=\"nofollow\">Laverna<\/a> - Laverna is a Javascript note taking application with Markdown editor and encryption support. Consider it like open source alternative to Evernote. (<a href=\"https:\/\/laverna.cc\/index.html#notes\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Laverna\/laverna\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/leanote.org\/\" rel=\"nofollow\">Leanote<\/a> - Leanote, Not Just A Notepad! Open source cloud notepad. (<a href=\"https:\/\/leanote.com\/note\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/leanote\/leanote\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/georgeosddev.github.com\/markdown-edit\/\">Markdown Edit<\/a> - Online markdown editor\/viewer. (<a href=\"https:\/\/github.com\/georgeOsdDev\/markdown-edit\">Source Code<\/a>) <code>MIT<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/meemo.minimal-space.de\/\" rel=\"nofollow\">Meemo<\/a> - Personal notes stream with Markdown support. (<a href=\"https:\/\/github.com\/nebulade\/meemo\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pereorga\/minimalist-web-notepad\">minimalist-web-notepad<\/a> - Minimalist notepad.cc clone. <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/n1try\/mininote\">MiniNote<\/a> - Simple Markdown note-taking app with persistence. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/SSilence\/notes\">notes<\/a> - Simple script for personal notes with a single page with a WYSIWYG Editor. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/FoxUSA\/OpenNote\">OpenNote<\/a> - OpenNote was built to be an open web-based alternative to Microsoft OneNote (T) and EverNote. (<a href=\"https:\/\/foxusa.github.io\/OpenNote\/OpenNote\/#\/folder\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"http:\/\/paperwork.rocks\" rel=\"nofollow\">Paperwork<\/a> - OpenSource note-taking and archiving alternative to Evernote, Microsoft OneNote and Google Keep. (<a href=\"https:\/\/github.com\/twostairs\/paperwork\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/novavex\/textpad\">savepad<\/a> - Minimalist notepad based on notepad.cc. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.sharelatex.com\/\" rel=\"nofollow\">ShareLaTex<\/a> - Web-based collaborative LaTeX editor. (<a href=\"https:\/\/github.com\/sharelatex\/sharelatex\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/standardnotes.org\" rel=\"nofollow\">Standard Notes<\/a> - Simple and private notes app. Protect your privacy while getting more done. That's Standard Notes. (<a href=\"https:\/\/app.standardnotes.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/standardnotes\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/domchristie.github.io\/to-markdown\/\" rel=\"nofollow\">to-markdown<\/a> - HTML to Markdown converter written in JavaScript. (<a href=\"https:\/\/github.com\/domchristie\/to-markdown\">Source Code<\/a>) <code>MIT<\/code> <code>JavaScript<\/code><\/li>\n<li><a href=\"https:\/\/turtl.it\/\" rel=\"nofollow\">Turtl<\/a> - Totally private personal database and note taking app. (<a href=\"https:\/\/github.com\/turtl\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>CommonLisp<\/code><\/li>\n<\/ul><h2><a id=\"user-content-office-suites\" class=\"anchor\" aria-hidden=\"true\" href=\"#office-suites\"><\/a>Office Suites<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/cryptpad.fr\/\" rel=\"nofollow\">CryptPad<\/a> - CryptPad is the zero knowledge realtime collaborative editor (rich-text, files, source-code, ...). (<a href=\"https:\/\/github.com\/xwiki-labs\/cryptpad\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/ethercalc.org\/\" rel=\"nofollow\">EtherCalc<\/a> - Web spreadsheet. (<a href=\"https:\/\/github.com\/audreyt\/ethercalc\">Source Code<\/a>) <code>CPAL-1.0\/Other<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/JohnMcLear\/draw\">EtherDraw<\/a> - Intuitive collaborative drawing web based tool. <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/etherpad.org\/\" rel=\"nofollow\">Etherpad<\/a> - Etherpad is a highly customizable Open Source online editor providing collaborative editing in really real-time. (<a href=\"https:\/\/demo.sandstorm.io\/appdemo\/h37dm17aa89yrd8zuqpdn36p6zntumtv08fjpu8a8zrte7q1cn60\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ether\/etherpad-lite\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/gobby\/gobby\/wiki\/Dedicated%20Server\">Infinoted<\/a> - Server for <a href=\"https:\/\/github.com\/gobby\/gobby\/wiki\">Gobby<\/a>, a multi-platform collaborative text editor. (<a href=\"https:\/\/github.com\/gobby\/gobby\">Source Code<\/a>) <code>MIT<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/helpcenter.onlyoffice.com\/faq\/server-opensource.aspx\" rel=\"nofollow\">ONLYOFFICE<\/a> - Office suite that enables you to manage documents, projects, team and customer relations in one place. (<a href=\"https:\/\/github.com\/ONLYOFFICE\/DocumentServer\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/PHPOffice\">PHPOffice<\/a> - PHPOffice contains libraries which permits to write and read files from most office suites. <code>LGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/webodf.org\/\" rel=\"nofollow\">WebODF<\/a> - Tools and libraries to view and edit Open Document Format (ODF) files. (<a href=\"https:\/\/github.com\/kogmbh\/WebODF\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>HTML5<\/code><\/li>\n<\/ul><h2><a id=\"user-content-password-managers\" class=\"anchor\" aria-hidden=\"true\" href=\"#password-managers\"><\/a>Password Managers<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/bitwarden.com\/\" rel=\"nofollow\">Bitwarden<\/a> - Password manager with webapp, browser extension, and mobile app. (<a href=\"https:\/\/github.com\/bitwarden\/core\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/keeweb.info\/\" rel=\"nofollow\">keeweb<\/a> - This webapp is a browser and desktop password manager compatible with KeePass databases. (<a href=\"https:\/\/github.com\/keeweb\/keeweb\">Source Code<\/a>) <code>MIT<\/code> <code>HTML5<\/code><\/li>\n<li><a href=\"https:\/\/www.passbolt.com\/\" rel=\"nofollow\">Passbolt<\/a> - Password manager dedicated for managing passwords in a collaborative way on any Web server, using a MySQL database backend. (<a href=\"https:\/\/github.com\/passbolt\/passbolt_api\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.syspass.org\/\" rel=\"nofollow\">sysPass<\/a> - Multiuser password management system. (<a href=\"http:\/\/demo.syspass.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/nuxsmin\/sysPass\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/teampass.net\/\" rel=\"nofollow\">Teampass<\/a> - Password manager dedicated for managing passwords in a collaborative way on any server Apache, MySQL and PHP. (<a href=\"https:\/\/github.com\/nilsteampassnet\/TeamPass\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-pastebins\" class=\"anchor\" aria-hidden=\"true\" href=\"#pastebins\"><\/a>Pastebins<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/sametmax\/0bin\">0bin<\/a> - Client side encrypted pastebin. <code>WTFPL<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/bepasty-server.readthedocs.io\/en\/latest\/\" rel=\"nofollow\">bepasty<\/a> - A pastebin for all kinds of files. (<a href=\"https:\/\/github.com\/bepasty\/bepasty-server\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/codervault\">CoderVault<\/a> - Open source, self-hosted, snippet manager. (<a href=\"https:\/\/github.com\/codervault\/codervault\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/cryptonote.me\/\" rel=\"nofollow\">cryptonote<\/a> - Simple open source web application that lets users encrypt and share messages that can only be read once. (<a href=\"https:\/\/github.com\/alainmeier\/cryptonote\">Source Code<\/a>) <code>BSD<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Edraens\/EdPaste\">EdPaste<\/a> - Self-hosted pastebin written in Laravel (PHP Framework). (<a href=\"https:\/\/paste.edra.ovh\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/solusipse\/fiche\">fiche<\/a> - Command line pastebin, all you need is netcat. (<a href=\"http:\/\/termbin.com\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/gitnet.fr\/deblan\/gist\" rel=\"nofollow\">GIST<\/a> - GIST is an open-source application to share code. (<a href=\"https:\/\/gist.deblan.org\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/hastebin.com\/about.md\" rel=\"nofollow\">hastebin<\/a> - Open source pastebin written in node.js. (<a href=\"http:\/\/hastebin.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/seejohnrun\/haste-server\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.modernpaste.com\" rel=\"nofollow\">Modern Paste<\/a> - Modern pastebin with a contemporary, minimalistic user interface backed by a robust feature set. (<a href=\"https:\/\/github.com\/LINKIWI\/modern-paste\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/search.cpan.org\/dist\/App-mojopaste\/\" rel=\"nofollow\">mojopaste<\/a> - Perl based pastebin. (<a href=\"http:\/\/p.thorsen.pm\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/jhthorsen\/app-mojopaste\">Source Code<\/a>) <code>Artistic-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/chmllr\/NoteHub\">NoteHub<\/a> - Free and Hassle-free Pastebin for Markdown Pages. Simple, clean, password provided, generated-short link. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/phpaste.sourceforge.io\/\" rel=\"nofollow\">Paste<\/a> - Paste is forked from the original source pastebin.com used before it was bought. (<a href=\"https:\/\/github.com\/jordansamuel\/PASTE\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/prologic\/pastebin\">pastebin<\/a> - Simple pastebin service with convenient api and CLI. (<a href=\"https:\/\/paste.mills.io\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ewhal\/Pastebin\">Pastebin<\/a> - Modern self-hosted pastebin service with a restful API. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ptpb\/pb\">pb<\/a> - Lightweight pastebin (and url shortener) built using flask. (<a href=\"https:\/\/ptpb.pw\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/bhanderson\/pbnh\">pbnh<\/a> - Pastebin inspired from project pb and hastebin, with an API and a SQL-based backend. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/pste.pw\" rel=\"nofollow\">Phaste<\/a> - Phalcon-based, MySQL pastebin application with privacy-respecting Piwik integration and syntax highlighting. (<a href=\"https:\/\/github.com\/FoxDev\/Phaste\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/privatebin.info\/\" rel=\"nofollow\">PrivateBin<\/a> - PrivateBin is a minimalist, opensource online pastebin\/discussion board where the server has zero knowledge of hosted data. (<a href=\"https:\/\/privatebin.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/PrivateBin\/PrivateBin\">Source Code<\/a>) <code>Zlib<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/phonicmouse\/SharpPaste\">SharpPaste<\/a> - Cross-platform C# pastebin with client-side AES-256 encryption that just works. (<a href=\"https:\/\/sharppaste.nl\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>C#\/NancyFX<\/code><\/li>\n<li><a href=\"https:\/\/snibox.github.io\/\" rel=\"nofollow\">Snibox<\/a> - Code snippets manager with attractive tag-oriented interface. (<a href=\"https:\/\/snibox-demo.herokuapp.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/snibox\/snibox\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/snipt.net\/\" rel=\"nofollow\">snipt<\/a> - Long-term memory for coders. Share and store code snippets. (<a href=\"https:\/\/github.com\/nicksergeant\/snipt\/\">Source Code<\/a>) <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/claudehohl\/Stikked\">Stikked<\/a> - Advanced and beautiful pastebin. (<a href=\"https:\/\/paste.scratchbook.ch\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/algolia\/sup3rS3cretMes5age\">Sup3rS3cretMes5age<\/a> - Very simple (to deploy and to use) secret message service using Hashicorp Vault as a secrets storage. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/andreineculau\/tastebin\">tastebin<\/a> - Updated version of hastebin plus additional features. <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h2><a id=\"user-content-personal-dashboards\" class=\"anchor\" aria-hidden=\"true\" href=\"#personal-dashboards\"><\/a>Personal Dashboards<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#metric--metric-collection\">https:\/\/github.com\/n1trux\/awesome-sysadmin#metric--metric-collection<\/a><\/p>\n<ul><li><a href=\"https:\/\/github.com\/cdubz\/babybuddy\">Baby Buddy<\/a> - Helps caregivers track baby sleep, feedings, diaper changes, and tummy time. (<a href=\"https:\/\/babybuddy.herokuapp.com\/\" rel=\"nofollow\">Demo<\/a>) <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/foxmask\/dj-diabetes\">Dj Diabetes<\/a> - My Glucose Manager - follow your daily health. <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/habitica.com\/\" rel=\"nofollow\">Habitica<\/a> - Habit tracker app which treats your goals like a Role Playing Game. Previously called HabitRPG. (<a href=\"https:\/\/github.com\/HabitRPG\/habitrpg\">Source Code<\/a>) <code>GPL-3.0\/CC-BY-NC-SA-3.0\/CC-BY-SA-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/tomershvueli\/homepage\">Homepage<\/a> - Simple, standalone, self-hosted PHP page that is your window to your server and the web. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/causefx\/iDashboard-PHP\">iDashboard-PHP<\/a> - HTPC Dashboard to load website services. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Swagielka\/simple-dash\">simple-dash<\/a> - A simple, fully responsive Dashboard to forward to the services of your choice. (<a href=\"http:\/\/simple-dash-demo.bplaced.net\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/allegro.tech\/tipboard\/\" rel=\"nofollow\">Tipboard<\/a> - In-house, tasty, local dashboarding system. (<a href=\"https:\/\/github.com\/allegro\/tipboard\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/wger.de\/\" rel=\"nofollow\">wger<\/a> - Web-based personal workout, fitness and weight logger\/tracker. It can also be used as a simple gym management utility and offers a full REST API as well. (<a href=\"https:\/\/wger.de\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/rolandgeider\/wger\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-photo-and-video-galleries\" class=\"anchor\" aria-hidden=\"true\" href=\"#photo-and-video-galleries\"><\/a>Photo and Video Galleries<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/chevereto.com\/free\" rel=\"nofollow\">Chevereto Free<\/a> - Powerful and fast image hosting script that allows you to create your very own full featured image hosting website in just minutes. (<a href=\"https:\/\/github.com\/Chevereto\/Chevereto-Free\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/coppermine-gallery.net\/\" rel=\"nofollow\">Coppermine<\/a> - Multilingual photo gallery that integrates with various bulletin boards. Includes upload approval and password protected albums. (<a href=\"http:\/\/coppermine-gallery.net\/demo\/cpg15x\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/coppermine-gallery\/cpg1.6.x\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/cumulusclips.org\/\" rel=\"nofollow\">CumulusClips<\/a> - Your own video sharing website with CumulusClips video sharing script. You can build a YouTube clone where users can upload, rate, comment on videos, and much more. (<a href=\"https:\/\/tube.devosi.org\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/benschwarz.github.io\/gallery-css\/\" rel=\"nofollow\">Gallery CSS<\/a> - Gallery.css is all CSS. Think: Simple, maintainable and understandable galleries without the use of Javascript. (<a href=\"https:\/\/github.com\/benschwarz\/gallery-css\">Source Code<\/a>) <code>MIT<\/code> <code>CSS<\/code><\/li>\n<li><a href=\"http:\/\/lychee.electerious.com\/\" rel=\"nofollow\">Lychee<\/a> - Open source grid and album based photo-management-system. (<a href=\"https:\/\/github.com\/electerious\/Lychee\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/mediadrop.video\/\" rel=\"nofollow\">MediaDrop<\/a> - Video, audio and podcast publication platform. (<a href=\"https:\/\/github.com\/mediadrop\/mediadrop\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/mediagoblin.org\" rel=\"nofollow\">Mediagoblin<\/a> - Free software media publishing platform that anyone can run. You can think of it as a decentralized alternative to Flickr, YouTube, SoundCloud, etc. (<a href=\"https:\/\/savannah.gnu.org\/projects\/mediagoblin\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/sebsauvage\/MinigalNano\">MinigalNano<\/a> - MinigalNano is a very simple image gallery. It adheres to the KISS principle and is very easy to install. MinigalNano does not have a web admin interface: You just upload your images in the photo folder on your server (using FTP, SFTP). <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/trebonius0\/Photato\">Photato<\/a> - Self-hosted photo gallery, accessible through a responsive WebUI. Directly uses and indexes a specific folder in the filesystem. <code>AGPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/blog.zx2c4.com\/567\" rel=\"nofollow\">Photofloat<\/a> - Web 2.0 Photo Gallery Done Right via Static JSON and Dynamic Javascript. (<a href=\"http:\/\/photos.jasondonenfeld.com\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/thibaud-rohmer\/PhotoLight\">PhotoLight<\/a> - The easiest photo gallery there is. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.photoshow-gallery.com\/\" rel=\"nofollow\">Photoshow<\/a> - Responsive Web Gallery. (<a href=\"https:\/\/github.com\/thibaud-rohmer\/PhotoShow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/piwigo.org\/\" rel=\"nofollow\">Piwigo<\/a> - Photo gallery software for the web, built by an active community of users and developers. (<a href=\"http:\/\/piwigo.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Piwigo\/Piwigo\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/blog.plumi.org\/\" rel=\"nofollow\">Plumi<\/a> - Create your own sophisticated video-sharing site. (<a href=\"https:\/\/github.com\/plumi\/plumi.app\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/saimn\/sigal\">sigal<\/a> - Yet another simple static gallery generator. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.ubergallery.net\" rel=\"nofollow\">UberGallery<\/a> - UberGallery is an easy to use, simple to manage, web photo gallery. UberGallery does not require a database and supports JPEG, GIF and PNG file types. Simply upload your images and UberGallery will automatically generate thumbnails and output HTML. (<a href=\"https:\/\/github.com\/UberGallery\/UberGallery\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/videobin.org\/code\" rel=\"nofollow\">Videobin<\/a> - Simple video upload and sharing service with transcoding. (<a href=\"http:\/\/videobin.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/r-w-x.org\/videobin.git\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.zenphoto.org\/\" rel=\"nofollow\">Zenphoto<\/a> - Open-source gallery and CMS project. (<a href=\"https:\/\/github.com\/zenphoto\/zenphoto\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-polls-and-events\" class=\"anchor\" aria-hidden=\"true\" href=\"#polls-and-events\"><\/a>Polls and Events<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/calagator.org\/\" rel=\"nofollow\">Calagator<\/a> - Event aggregator. (<a href=\"http:\/\/calagator.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/calagator\/calagator\">Source code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/chr15m\/Clerk\">Clerk<\/a> - Simple event logger to keep track of periodic events, habits, etc. as they occur. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/primelife.ercim.eu\/results\/opensource\/63-dudle\" rel=\"nofollow\">dudle<\/a> - Online scheduling application. (<a href=\"https:\/\/dudle.inf.tu-dresden.de\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/kellerben\/dudle\">Source code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/getfider.com\" rel=\"nofollow\">Fider<\/a> - Open source alternative to UserVoice for customer feedback. (<a href=\"https:\/\/demo.fider.io\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/getfider\/fider\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/framadate.org\/\" rel=\"nofollow\">Framadate<\/a> - Online service for planning an appointment or make a decision quickly and easily: Make a poll, Define dates or subjects to choose, Send the poll link to your friends or colleagues, Discuss and make a decision. (<a href=\"https:\/\/framadate.org\/aqg259dth55iuhwm\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/git.framasoft.org\/framasoft\/framadate\" rel=\"nofollow\">Source Code<\/a>) <code>CECILL-B<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/kyela.net\/\" rel=\"nofollow\">Ky\u00e9l\u00e0<\/a> - Participation polls for group events. (<a href=\"http:\/\/kyela.net\/55232734237c8\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/abienvenu\/Kyela\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.limesurvey.org\" rel=\"nofollow\">LimeSurvey<\/a> - Feature-rich Open Source web based polling software. Supports extensive survey logic. (<a href=\"https:\/\/demo.limesurvey.org\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/LimeSurvey\/LimeSurvey\">Source code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/nuages.domainepublic.net\/\" rel=\"nofollow\">Nuages<\/a> - Collaborative meeting poll system, similar to doodle or rdvz. (<a href=\"http:\/\/git.domainepublic.net\/?p=nuages.git;a=tree\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.phpback.org\" rel=\"nofollow\">PHPBack<\/a> - The open source feedback system. (<a href=\"http:\/\/www.phpback.org\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ivandiazwm\/phpback\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/tellform.com\" rel=\"nofollow\">TellForm<\/a> - Beautiful online form and survey creator. (<a href=\"https:\/\/tellform.com\/examples\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/whitef0x0\/tellform\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h2><a id=\"user-content-proxy\" class=\"anchor\" aria-hidden=\"true\" href=\"#proxy\"><\/a>Proxy<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/n1try\/http2-serverpush-proxy\">http2-serverpush-proxy<\/a> - Reverse proxy that helps to automatically utilize HTTP\/2.0's server push mechanism for static websites. Available as middleware and standalone application. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/code.kryo.se\/iodine\/\" rel=\"nofollow\">iodine<\/a> - IPv4 over DNS tunnel solution, enabling you to start up a socks5 proxy listener. (<a href=\"https:\/\/github.com\/yarrick\/iodine\">Source Code<\/a>) <code>ISC<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/thekvs\/microproxy\">microproxy<\/a> - lightweight non-caching HTTP\/HTTPS proxy server. (<a href=\"https:\/\/github.com\/thekvs\/microproxy\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/joshdick\/miniProxy\/\">miniProxy<\/a> - Simple web proxy written in PHP that can allow you to bypass Internet content filters, or to browse the internet anonymously. Only one php file. (<a href=\"https:\/\/github.com\/joshdick\/miniProxy\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/lehollandaisvolant.net\/tout\/oranjeproxy\/\" rel=\"nofollow\">Oranjeproxy<\/a> - Anonymizing web proxy. (<a href=\"https:\/\/github.com\/AmauryCarrade\/OranjeProxy\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.php-proxy.com\/\" rel=\"nofollow\">PHP-Proxy<\/a> - Web proxy script built specifically to be fast, easy to modify and to support video sites such as YouTube. (<a href=\"https:\/\/unblockvideos.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Athlon1600\/php-proxy-app\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.apsis.ch\/pound\/\" rel=\"nofollow\">Pound<\/a> - Light-weight reverse proxy and load balancer for HTTP\/HTTPS. (<a href=\"http:\/\/www.apsis.ch\/pound\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"http:\/\/www.privoxy.org\" rel=\"nofollow\">Privoxy<\/a> - Non-caching web proxy with advanced filtering capabilities for enhancing privacy, modifying web page data and HTTP headers, controlling access, and removing ads and other obnoxious Internet junk. <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/OptimalBits\/redbird\">Redbird<\/a> - A modern reverse proxy for node that includes cluster, HTTP2, LetsEncrypt, and Docker support. <code>BSD-2-Clause<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/www.squid-cache.org\/\" rel=\"nofollow\">Squid<\/a> - Caching proxy for the Web supporting HTTP, HTTPS, FTP, and more. It reduces bandwidth and improves response times by caching and reusing frequently-requested web pages. (<a href=\"https:\/\/code.launchpad.net\/squid\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/swiperproxy.github.io\/\" rel=\"nofollow\">Swiperproxy<\/a> - Lightning-fast, open source web proxy that is easy for you to run and customize. (<a href=\"https:\/\/github.com\/swiperproxy\/swiperproxy\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/banu.com\/tinyproxy\/\" rel=\"nofollow\">Tinyproxy<\/a> - Light-weight HTTP\/HTTPS proxy daemon. (<a href=\"https:\/\/github.com\/tinyproxy\/tinyproxy\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/traefik.io\/\" rel=\"nofollow\">Traefik<\/a> - Tr\u00e6f\u026ak is a modern HTTP reverse proxy and load balancer made to deploy microservices with ease. It supports several backends (Docker, Swarm, Mesos\/Marathon, \u2026) to manage its configuration automatically and dynamically. (<a href=\"https:\/\/github.com\/containous\/traefik\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<\/ul><h2><a id=\"user-content-read-it-later-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#read-it-later-lists\"><\/a>Read it Later Lists<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/keeper.nunux.org\/\" rel=\"nofollow\">Nunux Keeper<\/a> - Your personal content curation service. (<a href=\"https:\/\/github.com\/ncarlier\/nunux-keeper\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.wallabag.org\" rel=\"nofollow\">Wallabag<\/a> - Wallabag, formerly Poche, is a web application allowing you to save articles to read them later with improved readability. (<a href=\"http:\/\/demo.wallabag.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/wallabag\/wallabag\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-resource-planning\" class=\"anchor\" aria-hidden=\"true\" href=\"#resource-planning\"><\/a>Resource Planning<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"http:\/\/farmos.org\/\" rel=\"nofollow\">farmOS<\/a> - Web-based farm record keeping application. (<a href=\"https:\/\/github.com\/farmOS\/farmOS\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/grocy.info\/\" rel=\"nofollow\">Grocy<\/a> - grocy is a web-based selfhosted groceries management solution for your home, available in English, German, Italian and Norwegian. (<a href=\"https:\/\/demo-en.grocy.info\/stockoverview\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/berrnd\/grocy\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Tanibox\/tania-core\">Tania<\/a> - Tania is a free and open source farming management system for everyone. You can manage your areas, reservoirs, farm tasks, inventories, and the crop growing progress. <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<\/ul><h3><a id=\"user-content-enterprise-resource-planning\" class=\"anchor\" aria-hidden=\"true\" href=\"#enterprise-resource-planning\"><\/a>Enterprise Resource Planning<\/h3>\n<ul><li><a href=\"https:\/\/erpnext.com\" rel=\"nofollow\">ERPNext<\/a> - Free open source ERP system. (<a href=\"https:\/\/demo.erpnext.com\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/frappe\/erpnext\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/ledgersmb.org\/\" rel=\"nofollow\">LedgerSMB<\/a> - Integrated accounting and ERP system for small and midsize businesses, with double entry accounting, budgeting, invoicing, quotations, projects, orders and inventory management, shipping and more. (<a href=\"https:\/\/demo.cloud.efficito.com\/erp\/1.5\/login.pl\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ledgersmb\/LedgerSMB\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"http:\/\/odoo.com\" rel=\"nofollow\">Odoo<\/a> - Free open source ERP system. (<a href=\"https:\/\/demo.odoo.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/odoo\/odoo\">Source Code<\/a>) <code>LGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.tryton.org\/\" rel=\"nofollow\">Tryton<\/a> - Free open source business solution. (<a href=\"http:\/\/www.tryton.org\/download.html\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/hg.tryton.org\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<\/ul><h2><a id=\"user-content-search-engines\" class=\"anchor\" aria-hidden=\"true\" href=\"#search-engines\"><\/a>Search Engines<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/ambar.cloud\" rel=\"nofollow\">Ambar<\/a> - Document Search Engine (OCR, Store &amp; Search). (<a href=\"https:\/\/app.ambar.cloud\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/RD17\/ambar\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs\/Python<\/code><\/li>\n<li><a href=\"http:\/\/www.gigablast.com\/\" rel=\"nofollow\">Gigablast<\/a> - open source search engine. (<a href=\"http:\/\/www.gigablast.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/gigablast\/open-source-search-engine\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/beniz.github.io\/seeks\/\" rel=\"nofollow\">Seeks<\/a> - Web search proxy and collaborative distributed tool for websearch. (<a href=\"https:\/\/github.com\/beniz\/seeks\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/asciimoo.github.io\/searx\/\" rel=\"nofollow\">Searx<\/a> - Privacy-respecting, hackable metasearch engine. (<a href=\"https:\/\/searx.me\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/asciimoo\/searx\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/yacy.net\/en\/index.html\" rel=\"nofollow\">Yacy<\/a> - Peer based, decentralized search engine server. (<a href=\"http:\/\/search.yacy.net\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/yacy\/yacy_search_server\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Java<\/code><\/li>\n<\/ul><h2><a id=\"user-content-software-development\" class=\"anchor\" aria-hidden=\"true\" href=\"#software-development\"><\/a>Software Development<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<h3><a id=\"user-content-project-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#project-management\"><\/a>Project Management<\/h3>\n<p><em>See also <a href=\"#ticketing\">Ticketing<\/a>, <a href=\"#task-managementto-do-lists\">Task management\/To-do lists<\/a><\/em><\/p>\n<ul><li><a href=\"https:\/\/bloodhound.apache.org\/\" rel=\"nofollow\">Apache Bloodhound<\/a> - Manage software products, Keep track of features, tasks and bugs. <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/bonobogitserver.com\/\" rel=\"nofollow\">Bonobo Git Server<\/a> - Set up your own self hosted git server on IIS for Windows. Manage users and have full control over your repositories with a nice user friendly graphical interface. (<a href=\"https:\/\/github.com\/jakubgarfield\/Bonobo-Git-Server\">Source Code<\/a>) <code>MIT<\/code> <code>C#<\/code><\/li>\n<li><a href=\"https:\/\/www.fossil-scm.org\/index.html\/doc\/trunk\/www\/index.wiki\" rel=\"nofollow\">Fossil<\/a> - Distributed version control system featuring wiki and bug tracker. <code>BSD<\/code> <code>C<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/iluminar\/goodwork\">Goodwork<\/a> - Self hosted project management and collaboration tool powered by Laravel &amp; VueJS. (<a href=\"https:\/\/goodworkfor.life\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/gitblit.com\/\" rel=\"nofollow\">Gitblit<\/a> - Pure Java stack for managing, viewing, and serving Git repositories. (<a href=\"https:\/\/github.com\/gitblit\/gitblit\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/gitbucket.github.io\/gitbucket-news\/\" rel=\"nofollow\">gitbucket<\/a> - The easily installable GitHub clone powered by Scala. (<a href=\"https:\/\/github.com\/gitbucket\/gitbucket\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Scala\/Java<\/code><\/li>\n<li><a href=\"https:\/\/gitea.io\" rel=\"nofollow\">Gitea<\/a> - Community managed fork of Gogs, lightweight code hosting solution. (<a href=\"https:\/\/try.gitea.io\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/go-gitea\/gitea\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/gitlab.org\/\" rel=\"nofollow\">GitLab<\/a> - Self Hosted Git repository management, code reviews, issue tracking, activity feeds and wikis. (<a href=\"https:\/\/gitlab.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/gitlab.com\/gitlab-org\/gitlab-ce\" rel=\"nofollow\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/gitlist.org\/\" rel=\"nofollow\">Gitlist<\/a> - Web-based git repository browser - GitList allows you to browse repositories using your favorite browser, viewing files under different revisions, commit history and diffs. (<a href=\"https:\/\/github.com\/klaussilveira\/gitlist\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/gitolite.com\/gitolite\/index.html\" rel=\"nofollow\">Gitolite<\/a> - Gitolite allows you to setup git hosting on a central server, with fine-grained access control and many more powerful features. (<a href=\"https:\/\/github.com\/sitaramc\/gitolite\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"http:\/\/gitprep.yukikimoto.com\/\" rel=\"nofollow\">GitPrep<\/a> - Portable Github clone. (<a href=\"http:\/\/perlcodesample.sakura.ne.jp\/gitprep\/gitprep.cgi\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/yuki-kimoto\/gitprep\">Source Code<\/a>) <code>Artistic-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/alberthier\/git-webui\">Git WebUI<\/a> - Standalone web based user interface for git repositories. <code>Apache-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/gogs.io\/\" rel=\"nofollow\">Gogs<\/a> - Painless self-hosted Git Service written in Go. (<a href=\"https:\/\/try.gogs.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/gogits\/gogs\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/kallithea-scm.org\/\" rel=\"nofollow\">Kallithea<\/a> - Source code management system that supports two leading version control systems, Mercurial and Git, with a web interface. (<a href=\"https:\/\/kallithea-scm.org\/repos\/kallithea\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jonashaag\/klaus\/\">Klaus<\/a> - Simple, easy-to-set-up Git web viewer that Just Works. <code>ISC<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/lavagna.io\" rel=\"nofollow\">Lavagna<\/a> - Lavagna is an open-source issue\/project management tool designed for small teams. Lightweight, pure Java, easy to install, easy to use. (<a href=\"https:\/\/github.com\/digitalfondue\/lavagna\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/octobox.io\/\" rel=\"nofollow\">Octobox<\/a> <code>\u26a0<\/code> - Take back control of your GitHub Notifications. (<a href=\"https:\/\/github.com\/octobox\/octobox\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.openproject.org\" rel=\"nofollow\">OpenProject<\/a> - OpenProject is a web-based project management system. (<a href=\"https:\/\/github.com\/opf\/openproject\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/phabricator.org\/\" rel=\"nofollow\">Phabricator<\/a> - Collection of web applications that help build better software. (<a href=\"https:\/\/secure.phabricator.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/phacility\/phabricator\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.phproject.org\/\" rel=\"nofollow\">Phproject<\/a> - High performance full-featured project management system. (<a href=\"http:\/\/demo.phproject.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Alanaktion\/phproject\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.redmine.org\/\" rel=\"nofollow\">Redmine<\/a> - Redmine is a flexible project management web application. (<a href=\"http:\/\/demo.redmine.org\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/svn.redmine.org\/redmine\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/rhodecode.com\/\" rel=\"nofollow\">RhodeCode<\/a> - RhodeCode is an open source platform for software development teams. It unifies and simplifies repository management for Git, Subversion, and Mercurial. (<a href=\"https:\/\/try.rhodecode.com\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/code.rhodecode.com\/\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.scm-manager.org\/\" rel=\"nofollow\">SCM Manager<\/a> - The easiest way to share and manage your Git, Mercurial and Subversion repositories over http. (<a href=\"https:\/\/bitbucket.org\/sdorra\/scm-manager\/src\" rel=\"nofollow\">Source Code<\/a>) <code>BSD<\/code> <code>Java<\/code><\/li>\n<li><a href=\"https:\/\/taiga.io\/\" rel=\"nofollow\">Taiga<\/a> - Agile Project Management Tool based on the Kanban and Scrum methods. (<a href=\"https:\/\/github.com\/taigaio\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/www.thebuggenie.com\/\" rel=\"nofollow\">The Bug Genie<\/a> - The Bug Genie is a friendly project management and issue tracking tool. (<a href=\"https:\/\/github.com\/thebuggenie\/thebuggenie\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/trac.edgewall.org\/\" rel=\"nofollow\">Trac<\/a> - Trac is an enhanced wiki and issue tracking system for software development projects. <code>BSD<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/www.tuleap.org\/\" rel=\"nofollow\">Tuleap<\/a> - Tuleap is a libre suite to plan, track, code and collaborate on software projects. (<a href=\"https:\/\/tuleap.net\/plugins\/git\/tuleap\/tuleap\/stable?p=tuleap%2Fstable.git&amp;a=tree\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<\/ul><h3><a id=\"user-content-bug-trackers\" class=\"anchor\" aria-hidden=\"true\" href=\"#bug-trackers\"><\/a>Bug Trackers<\/h3>\n<p>See <strong><a href=\"#ticketing\">Ticketing<\/a><\/strong><\/p>\n<h3><a id=\"user-content-idetools\" class=\"anchor\" aria-hidden=\"true\" href=\"#idetools\"><\/a>IDE\/Tools<\/h3>\n<ul><li><a href=\"https:\/\/doc.bblf.sh\" rel=\"nofollow\">Babelfish<\/a> - Self-hosted server for source code parsing. It can parse any file, in any supported language, extract an Abstract Syntax Tree from it, and convert it to a Universal Abstract Syntax Tree which can enable further analysis and transformation. (<a href=\"https:\/\/github.com\/bblfsh\/\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/www.eclipse.org\/che\/\" rel=\"nofollow\">Eclipse Che<\/a> - Open source workspace server and cloud IDE. (<a href=\"https:\/\/github.com\/eclipse\/che\">Source Code<\/a>) <code>EPL-1.0<\/code> <code>Docker\/Java<\/code><\/li>\n<li><a href=\"https:\/\/icecoder.net\/\" rel=\"nofollow\">ICEcoder<\/a> - ICEcoder is a web IDE \/ browser based code editor, which allows you to develop websites directly within the web browser. (<a href=\"http:\/\/demo.icecoder.net\/ICEcoder\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/mattpass\/ICEcoder\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/jsbin.com\/\" rel=\"nofollow\">JS Bin<\/a> - Open source collaborative web development debugging tool. (<a href=\"https:\/\/github.com\/jsbin\/jsbin\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/api.judge0.com\" rel=\"nofollow\">Judge0 API<\/a> - Open source API to compile and run source code. (<a href=\"https:\/\/github.com\/judge0\/api\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"http:\/\/www.koding.com\/\" rel=\"nofollow\">Koding<\/a> - The simplest way to manage your entire Dev Infrastructure. (<a href=\"https:\/\/github.com\/koding\/koding\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/micro.glark.io\" rel=\"nofollow\">Microglark<\/a> - Hacky minimalistic remote pair programming editor. (<a href=\"http:\/\/micro.glark.io\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/Bluefinch\/microglark\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/regexr.com\/\" rel=\"nofollow\">Regexr<\/a> - RegExr is a HTML\/JS based tool for creating, testing, and learning about Regular Expressions. (<a href=\"https:\/\/github.com\/gskinner\/regexr\/\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/requestb.in\/\" rel=\"nofollow\">RequestBin<\/a> - Inspect HTTP requests. Debug webhooks. (<a href=\"https:\/\/github.com\/Runscope\/requestbin\">Source Code<\/a>) <code>MIT<\/code> <code>python<\/code><\/li>\n<li><a href=\"https:\/\/www.rstudio.com\/products\/rstudio\/#Server\" rel=\"nofollow\">RStudio Server<\/a> - Web browser based IDE for R. (<a href=\"https:\/\/github.com\/rstudio\/rstudio\/\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Java\/C++<\/code><\/li>\n<li><a href=\"http:\/\/aerokube.com\/selenoid\/latest\/\" rel=\"nofollow\">Selenoid<\/a> - Lightweight Selenium hub implementation launching browsers within Docker containers. (<a href=\"https:\/\/github.com\/aerokube\/selenoid\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/sourcegraph.com\" rel=\"nofollow\">sourcegraph<\/a> - Sourcegraph is a fast, open-source, fully-featured code search and navigation engine written in Go. (<a href=\"https:\/\/github.com\/sourcegraph\/sourcegraph\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/b3log\/wide\">Wide<\/a> - Web-based IDE for Teams using Go programming language\/Golang. (<a href=\"http:\/\/wide.b3log.org\/signup\" rel=\"nofollow\">Demo<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/zalando\/zalenium\">Zalenium<\/a> - Allows anyone to have a disposable and flexible Docker-based Selenium Grid infrastructure featuring video recording, live preview and online\/offline dashboards. <code>Apache-2.0<\/code> <code>Java\/Shell<\/code><\/li>\n<\/ul><h3><a id=\"user-content-continuous-integration\" class=\"anchor\" aria-hidden=\"true\" href=\"#continuous-integration\"><\/a>Continuous Integration<\/h3>\n<ul><li><a href=\"https:\/\/concourse.ci\/\" rel=\"nofollow\">Concourse<\/a> - Concourse is a CI tool that treats pipelines as first class objects and containerizes every step along the way. (<a href=\"https:\/\/ci.concourse.ci\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/concourse\/concourse\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/drone.io\/\" rel=\"nofollow\">drone<\/a> - Drone is a Continuous Delivery platform built on Docker, written in Go. (<a href=\"https:\/\/github.com\/drone\/drone\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/www.factor.io\/\" rel=\"nofollow\">Factor<\/a> - It's like IFTTT (if-this-then-that) for Dev and DevOps. (<a href=\"https:\/\/github.com\/factor-io\/factor\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/golangci.com\/\" rel=\"nofollow\">GolangCI<\/a> - Open Source automated code review service for Go integrated with GitHub pull requests. (<a href=\"https:\/\/github.com\/golangci\/golangci\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/jenkins-ci.org\/\" rel=\"nofollow\">Jenkins<\/a> - Continuous Integration Server. (<a href=\"https:\/\/github.com\/jenkinsci\/jenkins\/\">Source Code<\/a>) <code>MIT<\/code> <code>Java<\/code><\/li>\n<li><a href=\"http:\/\/laminar.ohwg.net\" rel=\"nofollow\">Laminar<\/a> - Fast, lightweight, simple and flexible Continuous Integration. (<a href=\"https:\/\/github.com\/ohwgiles\/laminar\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/php-censor\/php-censor\">PHP Censor<\/a> - Open source self-hosted continuous integration server for PHP projects. <code>BSD-2-Clause<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.phptesting.org\/\" rel=\"nofollow\">PHPCI<\/a> - Free and open source continuous integration specifically designed for PHP. (<a href=\"https:\/\/github.com\/block8\/phpci\">Source Code<\/a>) <code>BSD-2-Clause<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/strider-cd.github.io\/\" rel=\"nofollow\">Strider<\/a> - Open Source Continuous Deployment \/ Continuous Integration platform. (<a href=\"https:\/\/github.com\/Strider-CD\/strider\">Source Code<\/a>) <code>BSD<\/code> <code>Nodejs<\/code><\/li>\n<\/ul><h3><a id=\"user-content-faasserverless\" class=\"anchor\" aria-hidden=\"true\" href=\"#faasserverless\"><\/a>FaaS\/Serverless<\/h3>\n<p><em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Serverless_computing\" rel=\"nofollow\">Serverless computing on Wikipedia<\/a><\/em><\/p>\n<ul><li><a href=\"https:\/\/github.com\/metrue\/fx\">FX<\/a> - fx is a tool to help you do Function as a Service with painless on your own servers. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/iron-io\/functions\">IronFunctions<\/a> - The serverless microservices platform by <a href=\"https:\/\/www.iron.io\/\" rel=\"nofollow\">iron.io<\/a>. <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.openfaas.com\/\" rel=\"nofollow\">OpenFaaS<\/a> - Serverless Functions Made Simple for Docker &amp; Kubernetes. (<a href=\"https:\/\/github.com\/openfaas\/faas\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<\/ul><h3><a id=\"user-content-api-management\" class=\"anchor\" aria-hidden=\"true\" href=\"#api-management\"><\/a>API Management<\/h3>\n<ul><li><a href=\"https:\/\/www.dreamfactory.com\/\" rel=\"nofollow\">DreamFactory<\/a> - Turns any SQL\/NoSQL\/Structured data into Restful API. (<a href=\"https:\/\/github.com\/dreamfactorysoftware\/dreamfactory\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/LINKIWI\/endpoint\/\">Endpoint<\/a> - Super simple mock HTTP API endpoints that return static JSON data, for testing webhooks and client libraries in development. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/www.fusio-project.org\/\" rel=\"nofollow\">Fusio<\/a> - Open-source API management platform which helps to build and manage REST APIs. (<a href=\"http:\/\/fusio-project.org\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/apioo\/fusio\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jsoendermann\/hapttic\">Hapttic<\/a> - Simple HTTP server that forwards all requests to a shell script to handle webhooks you receive. <code>Apache-2.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/www.paraio.org\" rel=\"nofollow\">Para<\/a> - Flexible and modular backend framework\/server for object persistence, API development and authentication. (<a href=\"https:\/\/github.com\/erudika\/para\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>Java<\/code><\/li>\n<\/ul><h3><a id=\"user-content-documentation-generators\" class=\"anchor\" aria-hidden=\"true\" href=\"#documentation-generators\"><\/a>Documentation Generators<\/h3>\n<p>See also <a href=\"#static-site-generators\">Static site generators<\/a><\/p>\n<ul><li><a href=\"http:\/\/haldean.org\/\" rel=\"nofollow\">Docstore<\/a> - Static document hosting without any server-side processing, does not require you to recompile every time you change an article. Clone the repository and add articles in the text\/ directory to get started. (<a href=\"https:\/\/github.com\/haldean\/docstore\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"http:\/\/ricostacruz.com\/flatdoc\/\" rel=\"nofollow\">Flatdoc<\/a> - Small Javascript file that fetches Markdown files and renders them as full pages. <code>MIT<\/code> <code>Javascript<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/mil\/markdown-tree\">markdown-tree<\/a> - Serve a hierarchy \/ tree directory of markdown files. Use intended for small sites built in markdown. <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/docs.readthedocs.org\/en\/latest\/install.html\" rel=\"nofollow\">Read the Docs<\/a> - Host documentation, making it fully searchable and easy to find; import your docs using any major version control system, including Mercurial, Git, Subversion, and Bazaar. (<a href=\"https:\/\/readthedocs.org\/projects\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/rtfd\/readthedocs.org\">Source Code<\/a>) <code>MIT<\/code> <code>Python<\/code><\/li>\n<\/ul><h3><a id=\"user-content-localization\" class=\"anchor\" aria-hidden=\"true\" href=\"#localization\"><\/a>Localization<\/h3>\n<ul><li><a href=\"https:\/\/www.accent.reviews\/\" rel=\"nofollow\">Accent<\/a> - Open-source, self-hosted, developer-oriented translation tool. (<a href=\"https:\/\/github.com\/mirego\/accent\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Elixir<\/code><\/li>\n<li><a href=\"http:\/\/parrot-translate.github.io\/parrot.github.io\/\" rel=\"nofollow\">Parrot<\/a> - Self-hosted Localization Management Platform built with Go and Angular 2. (<a href=\"https:\/\/github.com\/parrot-translate\/parrot\">Source Code<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/pootle.translatehouse.org\" rel=\"nofollow\">Pootle<\/a> - Online translation and localization tool. (<a href=\"https:\/\/github.com\/translate\/pootle\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/weblate.org\" rel=\"nofollow\">Weblate<\/a> - Web-based translation tool with tight version control integration. (<a href=\"https:\/\/demo.weblate.org\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/WeblateOrg\/weblate\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/zanata.org\" rel=\"nofollow\">Zanata<\/a> - Web-based translation platform for translators, content creators and developers to manage localisation projects. (<a href=\"https:\/\/github.com\/zanata\/zanata-platform\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Java<\/code><\/li>\n<\/ul><h2><a id=\"user-content-static-site-generators\" class=\"anchor\" aria-hidden=\"true\" href=\"#static-site-generators\"><\/a>Static site generators<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/staticsitegenerators.net\" rel=\"nofollow\">https:\/\/staticsitegenerators.net<\/a> and <a href=\"https:\/\/www.staticgen.com\" rel=\"nofollow\">https:\/\/www.staticgen.com<\/a><\/p>\n<h2><a id=\"user-content-task-managementto-do-lists\" class=\"anchor\" aria-hidden=\"true\" href=\"#task-managementto-do-lists\"><\/a>Task management\/To-do lists<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em>See also <a href=\"#project-management\">Project Management<\/a> and <a href=\"#ticketing\">Ticketing<\/a>.<\/em><\/p>\n<ul><li><a href=\"https:\/\/github.com\/arshad\/crepido\">Crepido<\/a> - Create (kanban) boards to track users and projects from flat markdown files. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/kanboard.net\/\" rel=\"nofollow\">Kanboard<\/a> - Simple and open source visual task board. (<a href=\"https:\/\/github.com\/kanboard\/kanboard\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.mytinytodo.net\/\" rel=\"nofollow\">myTinyTodo<\/a> - Simple way to manage your todo list in AJAX style. Uses PHP, jQuery, SQLite\/MySQL. GTD compliant. (<a href=\"http:\/\/www.mytinytodo.net\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/bitbucket.org\/maxpozdeev\/mytinytodo\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/chr15m\/omgnata\">omgnata<\/a> - Mobile friendly zero-feature TODO list web app. Unix philosophy. <code>GPL-3.0<\/code> <code>ClojureScript<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/lgg\/tasks.php\">PHP Task\/Todo list<\/a> - Simple task\/todo list that uses a JSON text file for the tasks. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/restya.com\/board\/\" rel=\"nofollow\">Restyaboard<\/a> - Open source Trello-like kanban board. (<a href=\"http:\/\/restya.com\/board\/demo.html\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/RestyaPlatform\/board\">Source Code<\/a>) <code>OSL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/scrumblr.ca\/\" rel=\"nofollow\">scrumblr<\/a> - Collaborative Online Scrum Tool Using Websockets, Node.js, jQuery, and CSS3. (<a href=\"http:\/\/scrumblr.ca\/demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ocdtrekkie\/scrumblr\/tree\/sandstorm\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/taskboard.matthewross.me\/\" rel=\"nofollow\">TaskBoard<\/a> - Kanban-inspired app for keeping track of things that need to get done. (<a href=\"https:\/\/taskboard.matthewross.me\/demo.html\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/kiswa\/TaskBoard\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.taskfreak.com\/original\" rel=\"nofollow\">Taskfreak<\/a> - Simple but efficient web based task manager written in PHP. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/m1guelpf\/tasks\">Tasks<\/a> - Simple tasks and notes manager written in PHP, jQuery and Bootstrap using a custom flat file database. <code>MPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/thewhitetulip\/Tasks\">Tasks<\/a> - Kanban based to-do list manager written in Go. <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/RaymiiOrg\/tasks.php\">tasks.php<\/a> - Simple task\/todo list manager. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/taskwarrior.org\/\" rel=\"nofollow\">Taskwarrior<\/a> - Taskwarrior is Free and Open Source Software that manages your TODO list from your command line. It is flexible, fast, efficient, and unobtrusive. It does its job then gets out of your way. (<a href=\"https:\/\/taskwarrior.org\/download\/#git\" rel=\"nofollow\">Source Code<\/a>) <code>MIT<\/code> <code>C++<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/satrun77\/tinyissue\">Tinyissue<\/a> - Simple Issue Tracking for Teams. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/prologic\/todo\">todo<\/a> - Simple todo list manager. (<a href=\"https:\/\/todo.mills.io\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/www.getontracks.org\/\" rel=\"nofollow\">Tracks<\/a> - Web-based application to help you implement David Allen\u2019s <a href=\"https:\/\/en.wikipedia.org\/wiki\/Getting_Things_Done\" rel=\"nofollow\">Getting Things Done\u2122<\/a> methodology. (<a href=\"https:\/\/github.com\/TracksApp\/tracks\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.usevolition.com\" rel=\"nofollow\">Volition<\/a> - Opinionated open-source task management. (<a href=\"https:\/\/www.usevolition.com\/today\/new?guest=true\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/garrettqmartin8\/volition\">Source Code<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/wekan.github.io\/\" rel=\"nofollow\">Wekan<\/a> - Open-source Trello-like kanban. (<a href=\"https:\/\/oasis.sandstorm.io\/appdemo\/m86q05rdvj14yvn78ghaxynqz7u2svw6rnttptxx49g1785cdv1h\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/wekan\/wekan\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/wheatbin.com\/\" rel=\"nofollow\">Wheatbin<\/a> - Project management software that combines Kanban methodology with the Law of the Harvest. (<a href=\"https:\/\/github.com\/wheatbin\/wheatbin\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-ticketing\" class=\"anchor\" aria-hidden=\"true\" href=\"#ticketing\"><\/a>Ticketing<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p><em>See also <a href=\"#task-managementto-do-lists\">Task management\/To-do lists<\/a> and <a href=\"#project-management\">Project Management<\/a><\/em><\/p>\n<ul><li><a href=\"https:\/\/getbrimir.com\/\" rel=\"nofollow\">Brimir<\/a> - Simple and clean open-source ticket manager written in Ruby on Rails. (<a href=\"https:\/\/github.com\/ivaldi\/brimir\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.bugzilla.org\/\" rel=\"nofollow\">Bugzilla<\/a> - General-purpose bugtracker and testing tool originally developed and used by the Mozilla project. <code>MPL-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"http:\/\/bumpy-booby.derivoile.fr\/\" rel=\"nofollow\">Bumpy Booby<\/a> - Simple, responsive and highly customizable PHP bug tracking system. (<a href=\"https:\/\/github.com\/piero-la-lune\/Bumpy-Booby\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.cerberusweb.com\/\" rel=\"nofollow\">Cerb<\/a> - Group-based e-mail management project. (<a href=\"https:\/\/github.com\/wgm\/cerb\">Source Code<\/a>) <code>DPL<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/coordino.com\/\" rel=\"nofollow\">Coordino<\/a> - Open Source question and answer system on top of CakePHP. (<a href=\"https:\/\/github.com\/Datawalke\/Coordino\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/Taskulu\/deskulu\">Deskulu<\/a> - Opensource helpdesk and ticketing system based on Drupal 7. (<a href=\"https:\/\/help.taskulu.com\" rel=\"nofollow\">Demo<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/diamantedesk.com\/\" rel=\"nofollow\">DiamanteDesk<\/a> - DiamanteDesk is FREE Open Source easy-to-use help-desk solution. (<a href=\"http:\/\/diamantedesk.com\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/eltrino\/diamantedesk-application\">Source Code<\/a>) <code>OSL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.flyspray.org\/\" rel=\"nofollow\">Flyspray<\/a> - Uncomplicated, web-based bug tracking system. (<a href=\"https:\/\/github.com\/Flyspray\/flyspray\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/freescout-helpdesk\/freescout\">FreeScout<\/a> - Open source clone of Help Scout: email-based customer support application, help desk and shared mailbox. <code>AGPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/huboard\/huboard\">HuBoard<\/a> <code>\u26a0<\/code> - Instant project management for your GitHub issues (Connects directly GitHub API). <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.mantisbt.org\/\" rel=\"nofollow\">MantisBT<\/a> - Self hosted bug tracker, fits best for software development. (<a href=\"https:\/\/www.mantisbt.org\/bugs\/my_view_page.php\" rel=\"nofollow\">Demo<\/a>), (<a href=\"https:\/\/github.com\/mantisbt\/mantisbt\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/www.opensupports.com\/\" rel=\"nofollow\">OpenSupports<\/a> - Multi language ticket system with FAQ, role management, metrics and canned response features. (<a href=\"https:\/\/www.opensupports.com\/demo\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/opensupports\/opensupports\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/osticket.com\/\" rel=\"nofollow\">osTicket<\/a> - Manage, organize and archive all your support requests and responses in one place. (<a href=\"https:\/\/github.com\/osTicket\/osTicket\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.otrs.com\/\" rel=\"nofollow\">OTRS<\/a> - Trouble ticket system for assigning tickets to incoming queries and tracking further communications. (<a href=\"https:\/\/github.com\/OTRS\/otrs\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/www.bestpractical.com\/rt\/\" rel=\"nofollow\">Request Tracker<\/a> - An enterprise-grade issue tracking system. (<a href=\"https:\/\/github.com\/bestpractical\/rt\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/getsentry\/onpremise\">Sentry On-Premise<\/a> - A powerful error tracking platform with wide language support and a robust API. (<a href=\"https:\/\/github.com\/getsentry\/sentry\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python\/Django<\/code><\/li>\n<li><a href=\"http:\/\/sit-it.org\/\" rel=\"nofollow\">SIT<\/a> - SCM-agnostic, file-based, offline-first, immutable issue tracker. (<a href=\"https:\/\/github.com\/sit-it\/sit\">Source Code<\/a>) <code>MIT<\/code> <code>Apache-2.0<\/code> <code>Rust<\/code>.<\/li>\n<li><a href=\"http:\/\/www.thebuggenie.org\" rel=\"nofollow\">TheBugGenie<\/a> - Ticket system with extensive user rights system. (<a href=\"https:\/\/github.com\/thebuggenie\/thebuggenie\">Source Code<\/a>) <code>MPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/zammad.org\/\" rel=\"nofollow\">Zammad<\/a> - Easy to use but powerful open-source support and ticketing system. (<a href=\"https:\/\/github.com\/zammad\/zammad\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Ruby<\/code><\/li>\n<\/ul><h2><a id=\"user-content-url-shorteners\" class=\"anchor\" aria-hidden=\"true\" href=\"#url-shorteners\"><\/a>URL Shorteners<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/kutt.it\" rel=\"nofollow\">Kutt<\/a> - A modern URL shortener with support for custom domains. (<a href=\"https:\/\/github.com\/thedevs-network\/kutt\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/tommyku\/link-shortener-front-end\">Link-shortener-bot<\/a> <code>\u26a0<\/code> - URL shortener using a Telegram Bot. (<a href=\"http:\/\/t.me\/GiveMeShortLinkBot\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/linkrapp.com\" rel=\"nofollow\">Linkr<\/a> - Beautiful, fast URL shortening. (<a href=\"https:\/\/demo.linkrapp.com\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/LINKIWI\/linkr\">Source Code<\/a>) <code>MIT<\/code> <code>Python\/Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ldidry\/lstu\">Lstu<\/a> - <em>Let's SHorten That Url<\/em> - Lightweight URL shortener. <code>WTFPL<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ethanal\/nimbus\">Nimbus<\/a> - URL shortener and file sharer with a drag-and-drop OS X menu bar client and web interface. <code>MIT<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/project.polr.me\/\" rel=\"nofollow\">Polr<\/a> - Modern, minimalist, modular, and lightweight URL shortener. (<a href=\"https:\/\/github.com\/Cydrobolt\/polr\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/ziyasal\/reducio\">reduc.io<\/a> - URL shortener service written in Scala, using Akka-Http and Redis. <code>MIT<\/code> <code>Scala<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/sqozz\/schort\">schort<\/a> - No login, no javascript, just short links. (<a href=\"https:\/\/s.wx0.de\" rel=\"nofollow\">Demo<\/a>) <code>CC0-1.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/prologic\/shorturl\">shorturl<\/a> - Simple URL shortener with very tiny URLs. (<a href=\"https:\/\/url.mills.io\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/pips-\/shuri\">shuri<\/a> - SHort URI - Lighweight URL shortener. <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/cagataycali\/url-shortener\">url-shortener<\/a> <code>\u26a0<\/code> - Shitty url shortener, emoji and AI powered. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/yourls.org\/\" rel=\"nofollow\">YOURLS<\/a> - YOURLS is a set of PHP scripts that will allow you to run Your Own URL Shortener. Features include password protection, URL customization, bookmarklets, statistics, API, plugins, jsonp. (<a href=\"https:\/\/github.com\/YOURLS\/YOURLS\/\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<\/ul><h2><a id=\"user-content-vpn\" class=\"anchor\" aria-hidden=\"true\" href=\"#vpn\"><\/a>VPN<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#vpn\">https:\/\/github.com\/n1trux\/awesome-sysadmin#vpn<\/a><\/p>\n<h2><a id=\"user-content-web-servers\" class=\"anchor\" aria-hidden=\"true\" href=\"#web-servers\"><\/a>Web servers<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See <a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin#web\">https:\/\/github.com\/n1trux\/awesome-sysadmin#web<\/a><\/p>\n<h2><a id=\"user-content-wikis\" class=\"anchor\" aria-hidden=\"true\" href=\"#wikis\"><\/a>Wikis<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<p>See also <a href=\"#documentation-generators\">Documentation Generators<\/a>, <a href=\"http:\/\/www.wikimatrix.org\/\" rel=\"nofollow\">Wikimatrix<\/a>, <a href=\"http:\/\/wikiindex.org\/Category:Wiki_Engine\" rel=\"nofollow\">Wiki Engines on WikiIndex<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/List_of_wiki_software\" rel=\"nofollow\">List of wiki software on wikipedia<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Comparison_of_wiki_software\" rel=\"nofollow\">Comparison of wiki software on wikipedia<\/a>.<\/p>\n<ul><li><a href=\"https:\/\/www.bookstackapp.com\/\" rel=\"nofollow\">BookStack<\/a> - BookStack is a simple, self-hosted, easy-to-use platform for organizing and storing information. It allows for documentation to be stored in a book like fashion. (<a href=\"https:\/\/www.bookstackapp.com\/#demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/ssddanbrown\/BookStack\">Source Code<\/a>) <code>MIT<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/schollz\/cowyo\/\">Cowyo<\/a> - Cowyo is a feature-rich wiki for minimalists. (<a href=\"https:\/\/cowyo.com\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/django-wiki\/django-wiki\">django-wiki<\/a> - Wiki system with complex functionality for simple integration and a superb interface. Store your knowledge with style: Use django models. (<a href=\"https:\/\/demo.django.wiki\/\" rel=\"nofollow\">Demo<\/a>) <code>GPL-3.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/documize.com\" rel=\"nofollow\">Documize<\/a> - Modern Docs + Wiki software with built-in workflow, single binary executable, just bring MySQL\/Percona. (<a href=\"https:\/\/github.com\/documize\/community\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/www.dokuwiki.org\/DokuWiki\" rel=\"nofollow\">Dokuwiki<\/a> - Easy to use, lightweight, standards-compliant wiki engine with a simple syntax allowing reading the data outside the wiki. All data is stored in plain files, therefore no database is required. (<a href=\"https:\/\/github.com\/splitbrain\/dokuwiki\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/jgm\/gitit\">Gitit<\/a> - Wiki program that stores pages and uploaded files in a git repository, which can then be modified using the VCS command line tools or the wiki's web interface. <code>GPL-2.0<\/code> <code>Haskell<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/gollum\/gollum\">Gollum<\/a> - Simple, Git-powered wiki with a sweet API and local frontend. <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/claudioc\/jingo\">jingo<\/a> - Git based wiki engine written for node.js, with a decent design, a search capability and a good typography. (<a href=\"http:\/\/jingo.cica.li:6067\/wiki\/home\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/claudioc\/jingo\">Source Code<\/a>) <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/www.mediawiki.org\/wiki\/MediaWiki\" rel=\"nofollow\">Mediawiki<\/a> - MediaWiki is a free and open-source wiki software package written in PHP. It serves as the platform for Wikipedia and the other Wikimedia projects, used by hundreds of millions of people each month. (<a href=\"https:\/\/phabricator.wikimedia.org\/diffusion\/MW\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/moinmo.in\/\" rel=\"nofollow\">MoinMoin<\/a> - Advanced, easy to use and extensible WikiEngine with a large community of users. (<a href=\"http:\/\/hg.moinmo.in\/moin\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/minad\/olelo\">Olelo<\/a> - Olelo is a wiki that stores pages in a Git repository, supports many markup styles and has an extensible, hackable architecture. (<a href=\"http:\/\/gitwiki.org\/\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Ruby<\/code><\/li>\n<li><a href=\"https:\/\/www.getoutline.com\/\" rel=\"nofollow\">Outline<\/a> <code>\u26a0<\/code> - An open, extensible, wiki for your team built using React and Node.js. (<a href=\"https:\/\/github.com\/outline\/outline\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/sbrl\/Pepperminty-Wiki\/\">Pepperminty Wiki<\/a> - Complete markdown-powered wiki contained in a single PHP file. (<a href=\"https:\/\/starbeamrainbowlabs.com\/labs\/peppermint\/build\/\" rel=\"nofollow\">Demo<\/a>) <code>MPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/xy2z\/PineDocs\">PineDocs<\/a> - Simple, fast, customizable and lightweight site for browsing files. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.pmwiki.org\" rel=\"nofollow\">PmWiki<\/a> - Wiki-based system for collaborative creation and maintenance of websites. <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/raneto.com\/\" rel=\"nofollow\">Raneto<\/a> - Raneto is an open source Knowledgebase platform that uses static Markdown files to power your Knowledgebase. <code>MIT<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/realms.io\/\" rel=\"nofollow\">Realms<\/a> - Git-backed wiki inspired by Gollum. (<a href=\"https:\/\/github.com\/scragg0x\/realms-wiki\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Python<\/code><\/li>\n<li><a href=\"http:\/\/tiddlywiki.com\/\" rel=\"nofollow\">TiddlyWiki<\/a> - Reusable non-linear personal web notebook. (<a href=\"https:\/\/github.com\/Jermolene\/TiddlyWiki5\">Source Code<\/a>) <code>BSD<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"https:\/\/tiki.org\" rel=\"nofollow\">Tiki<\/a> - Wiki CMS Groupware with the most built-in features. (<a href=\"https:\/\/tiki.org\/Demo\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/sourceforge.net\/p\/tikiwiki\/code\/HEAD\/tree\/\" rel=\"nofollow\">Source Code<\/a>) <code>LGPL-2.1<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/twiki.org\/\" rel=\"nofollow\">TWiki<\/a> - TWiki is a Perl-based structured wiki application, typically used to run a collaboration platform, knowledge or document management system, a knowledge base, or team portal. (<a href=\"http:\/\/twiki.org\/cgi-bin\/view\/Sandbox\/WebHome\" rel=\"nofollow\">Demo<\/a>, <a href=\"http:\/\/svn.twiki.org\/svn\/twiki\/\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-1.0<\/code> <code>Perl<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/prologic\/wiki\">wiki<\/a> - Simple Markdown based wiki engine. (<a href=\"https:\/\/wiki.mills.io\" rel=\"nofollow\">Demo<\/a>) <code>MIT<\/code> <code>Go<\/code><\/li>\n<li><a href=\"https:\/\/wiki.js.org\/\" rel=\"nofollow\">Wiki.js<\/a> - Modern, lightweight and powerful wiki app built on NodeJS, Git and Markdown. (<a href=\"https:\/\/docs.requarks.io\/wiki\" rel=\"nofollow\">Demo<\/a>) <code>AGPL-3.0<\/code> <code>Nodejs<\/code><\/li>\n<li><a href=\"http:\/\/wikiss.tuxfamily.org\/\" rel=\"nofollow\">WiKiss<\/a> - Wiki, simple to use and install. (<a href=\"https:\/\/svnweb.tuxfamily.org\/listing.php?repname=wikiss\/svn&amp;path=%2F&amp;sc=0\" rel=\"nofollow\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"http:\/\/www.xwiki.org\" rel=\"nofollow\">XWiki<\/a> - Second generation wiki that allows the user to extend its functionalities with a powerful extension-based architecture. (<a href=\"http:\/\/playground.xwiki.org\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/xwiki\/xwiki-platform\">Source Code<\/a>) <code>LGPL-2.1<\/code> <code>Java<\/code><\/li>\n<\/ul><h2><a id=\"user-content-self-hosting-solutions\" class=\"anchor\" aria-hidden=\"true\" href=\"#self-hosting-solutions\"><\/a>Self-hosting Solutions<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/1backend\/1backend\">1Backend<\/a> - Self-host web apps, microservices and lambdas on your server. Advanced features enable service reuse and composition. <code>AGPL-3.0<\/code> <code>Go<\/code><\/li>\n<li><a href=\"http:\/\/dietpi.com\/\" rel=\"nofollow\">DietPi<\/a> - Minimal Debian OS optimized for single-board computers, which allows you to easily install and manage several services for selfhosting at home. (<a href=\"https:\/\/github.com\/Fourdee\/DietPi\">Source Code<\/a>) <code>GPL-2.0<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"https:\/\/dfabric.github.io\/DPlatform-Shell\/\" rel=\"nofollow\">DPlatform<\/a> - Deploy self-hosted apps easily: simple, bloat-free, independent installation. (<a href=\"https:\/\/github.com\/j8r\/DPlatform\">Source Code<\/a>) <code>MIT<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"https:\/\/freedombone.net\/\" rel=\"nofollow\">FreedomBone<\/a> - Home server configuration based on Debian. (<a href=\"https:\/\/code.freedombone.net\/bashrc\/freedombone\" rel=\"nofollow\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Shell<\/code><\/li>\n<li><a href=\"https:\/\/wiki.debian.org\/FreedomBox\" rel=\"nofollow\">FreedomBox<\/a> - Community project to develop, design and promote personal servers running free software for private, personal, communications. <code>GPL-3.0<\/code> <code>Python\/Other<\/code><\/li>\n<li><a href=\"https:\/\/www.freenas.org\/\" rel=\"nofollow\">FreeNAS<\/a> - Network-attached storage (NAS) software based on FreeBSD and the OpenZFS file system. Support for SMB, AFP, NFS, iSCSI, SSH, rsync and FTP\/TFTP protocols. Advanced features include full-disk encryption and plug-ins. (<a href=\"https:\/\/github.com\/freenas\/freenas\">Source Code<\/a>) <code>BSD-3-Clause<\/code> <code>Python\/Other<\/code><\/li>\n<li><a href=\"https:\/\/gitlab.com\/NickBusey\/HomelabOS\" rel=\"nofollow\">HomelabOS<\/a> - Your very own offline-first privacy-centric open-source data-center. <code>MIT<\/code> <code>Docker<\/code><\/li>\n<li><a href=\"http:\/\/www.openmediavault.org\/\" rel=\"nofollow\">OpenMediaVault<\/a> - OpenMediaVault is the next generation network attached storage (NAS) solution based on Debian Linux. It contains services like SSH, (S)FTP, SMB\/CIFS, DAAP media server, RSync, BitTorrent client and many more. (<a href=\"https:\/\/github.com\/openmediavault\/openmediavault\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>PHP<\/code><\/li>\n<li><a href=\"https:\/\/piratebox.cc\/start\" rel=\"nofollow\">Piratebox<\/a> - DIY anonymous offline file-sharing and communications system built with free software and inexpensive off-the-shelf hardware. (<a href=\"https:\/\/github.com\/PirateBox-Dev\">Source Code<\/a>). <code>GPL-3.0<\/code> <code>Python\/Other<\/code><\/li>\n<li><a href=\"http:\/\/puffin.rocks\/\" rel=\"nofollow\">Puffin<\/a> - Lightweight webapp catalog based on containers, with user interface \u00e0 la mobile app store. (<a href=\"http:\/\/puffin.rocks\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/puffinrocks\/puffin\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python\/Docker<\/code><\/li>\n<li><a href=\"https:\/\/sandstorm.io\/\" rel=\"nofollow\">Sandstorm<\/a> - Personal server for running self-hosted apps easily and securely. (<a href=\"https:\/\/demo.sandstorm.io\/\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/sandstorm-io\/sandstorm\">Source Code<\/a>) <code>Apache-2.0<\/code> <code>C++\/Other<\/code><\/li>\n<li><a href=\"https:\/\/github.com\/sovereign\/sovereign\">sovereign<\/a> - Set of Ansible playbooks to build and maintain your own private cloud: email, calendar, contacts, file sync, IRC bouncer, VPN, and more. <code>GPL-3.0<\/code> <code>YAML\/Other<\/code><\/li>\n<li><a href=\"https:\/\/syncloud.org\/\" rel=\"nofollow\">Syncloud<\/a> - Your own online file storage, social network or email server. (<a href=\"https:\/\/github.com\/syncloud\/platform\">Source Code<\/a>) <code>GPL-3.0<\/code> <code>Python\/Other<\/code><\/li>\n<li><a href=\"http:\/\/ubos.net\/\" rel=\"nofollow\">UBOS<\/a> - Linux distro that runs on indie boxes (personal servers and IoT devices). Single-command installation and management of apps - Jenkins, Mediawiki, Owncloud, WordPress, etc., and other <a href=\"http:\/\/ubos.net\/about\/\" rel=\"nofollow\">features<\/a>. <code>GPL-3.0<\/code> <code>Perl\/Other<\/code><\/li>\n<li><a href=\"http:\/\/WikiSuite.org\" rel=\"nofollow\">WikiSuite<\/a> - The most comprehensive and integrated Free \/ Libre \/ Open Source software suite ever developed. (<a href=\"http:\/\/wikisuite.org\/Source-Code\" rel=\"nofollow\">Source Code<\/a>) <code>Multiple<\/code> <code>ClearOS<\/code><\/li>\n<li><a href=\"https:\/\/yunohost.org\/\" rel=\"nofollow\">YunoHost<\/a> - Server operating system aiming to make self-hosting accessible to everyone. (<a href=\"https:\/\/yunohost.org\/#\/try\" rel=\"nofollow\">Demo<\/a>, <a href=\"https:\/\/github.com\/YunoHost\">Source Code<\/a>) <code>AGPL-3.0<\/code> <code>Python\/Other<\/code><\/li>\n<\/ul><hr \/><h2><a id=\"user-content-list-of-licenses\" class=\"anchor\" aria-hidden=\"true\" href=\"#list-of-licenses\"><\/a>List of Licenses<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><code>\u26a0 <\/code> - Depends on a third party network service<\/li>\n<li><code>AAL<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/AAL\" rel=\"nofollow\">Attribution Assurance License<\/a><\/li>\n<li><code>AGPL-3.0<\/code> - <a href=\"https:\/\/www.gnu.org\/licenses\/agpl-3.0\" rel=\"nofollow\">GNU Affero General Public License 3.0<\/a><\/li>\n<li><code>AGPL-3.0-only<\/code> - <a href=\"https:\/\/spdx.org\/licenses\/AGPL-3.0-only.html\" rel=\"nofollow\">GNU Affero General Public License 3.0 only<\/a><\/li>\n<li><code>Apache-2.0<\/code> - <a href=\"http:\/\/www.apache.org\/licenses\/\" rel=\"nofollow\">Apache, Version 2.0<\/a><\/li>\n<li><code>APSL-2.0<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/APSL-2.0\" rel=\"nofollow\">Apple Public Source License, Version 2.0<\/a><\/li>\n<li><code>Artistic-2.0<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/Artistic-2.0\" rel=\"nofollow\">Artistic License Version 2.0<\/a><\/li>\n<li><code>Beerware<\/code> - <a href=\"https:\/\/spdx.org\/licenses\/Beerware.html\" rel=\"nofollow\">Beerware License<\/a><\/li>\n<li><code>BSD-2-Clause<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/BSD-2-Clause\" rel=\"nofollow\">BSD 2-clause \"Simplified\"<\/a><\/li>\n<li><code>BSD-3-Clause<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/BSD-3-Clause\" rel=\"nofollow\">BSD 3-Clause \"New\" or \"Revised\"<\/a><\/li>\n<li><code>CC-BY-NC-SA-3.0<\/code> - <a href=\"https:\/\/creativecommons.org\/licenses\/by-nc-sa\/3.0\/\" rel=\"nofollow\">Creative Commons Attribution-NonCommercial-ShareAlike 3.0 International License<\/a><\/li>\n<li><code>CC-BY-SA-3.0<\/code> - <a href=\"https:\/\/creativecommons.org\/licenses\/by-sa\/3.0\/\" rel=\"nofollow\">Creative Commons Attribution-ShareAlike 3.0 International License<\/a><\/li>\n<li><code>CC-BY-SA-4.0<\/code> - <a href=\"https:\/\/creativecommons.org\/licenses\/by-sa\/4.0\/\" rel=\"nofollow\">Creative Commons Attribution-ShareAlike 4.0 International License<\/a><\/li>\n<li><code>CC0-1.0<\/code> - <a href=\"https:\/\/creativecommons.org\/about\/cc0\/\" rel=\"nofollow\">Public Domain<\/a><\/li>\n<li><code>CDDL-1.0<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/CDDL-1.0\" rel=\"nofollow\">Common Development and Distribution License<\/a><\/li>\n<li><code>CECILL-B<\/code> - <a href=\"http:\/\/www.cecill.info\/licences\/Licence_CeCILL-B_V1-en.txt\" rel=\"nofollow\">CEA CNRS INRIA Logiciel Libre<\/a><\/li>\n<li><code>CPAL-1.0<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/CPAL-1.0\" rel=\"nofollow\">Common Public Attribution License Version 1.0<\/a><\/li>\n<li><code>DPL<\/code> - <a href=\"https:\/\/cerb.ai\/license\/\" rel=\"nofollow\">Devblocks Public License 1.0<\/a><\/li>\n<li><code>ECL-2.0<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/ECL-2.0\" rel=\"nofollow\">Educational Community License, Version 2.0 <\/a><\/li>\n<li><code>EPL-1.0<\/code> - <a href=\"https:\/\/www.eclipse.org\/legal\/epl-v10.html\" rel=\"nofollow\">Eclipse Public License, Version 1.0<\/a><\/li>\n<li><code>GPL-1.0<\/code> - <a href=\"https:\/\/www.gnu.org\/licenses\/gpl-1.0\" rel=\"nofollow\">GNU General Public License<\/a><\/li>\n<li><code>GPL-2.0<\/code> - <a href=\"https:\/\/www.gnu.org\/licenses\/old-licenses\/gpl-2.0.en.html\" rel=\"nofollow\">GNU General Public License 2.0<\/a><\/li>\n<li><code>GPL-3.0<\/code> - <a href=\"http:\/\/www.gnu.org\/licenses\/gpl-3.0.en.html\" rel=\"nofollow\">GNU General Public License 3.0<\/a><\/li>\n<li><code>GPL-3.0-only<\/code> - <a href=\"https:\/\/spdx.org\/licenses\/GPL-3.0-only.html\" rel=\"nofollow\">GNU General Public License 3.0<\/a><\/li>\n<li><code>IPL-1.0<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/IPL-1.0\" rel=\"nofollow\">IBM Public License<\/a><\/li>\n<li><code>ISC<\/code> - <a href=\"https:\/\/www.isc.org\/downloads\/software-support-policy\/isc-license\/\" rel=\"nofollow\">Internet Systems Consortium License<\/a><\/li>\n<li><code>LGPL-2.1<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/LGPL-2.1\" rel=\"nofollow\">Lesser General Public License 2.1<\/a><\/li>\n<li><code>LGPL-3.0<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/LGPL-3.0\" rel=\"nofollow\">Lesser General Public License 3.0<\/a><\/li>\n<li><code>MIT<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/MIT\" rel=\"nofollow\">MIT License<\/a><\/li>\n<li><code>MPL-1.1<\/code> - <a href=\"https:\/\/www.mozilla.org\/media\/MPL\/1.1\/index.txt\" rel=\"nofollow\">Mozilla Public License Version 1.1<\/a><\/li>\n<li><code>MPL-2.0<\/code> - <a href=\"https:\/\/www.mozilla.org\/MPL\/2.0\/index.txt\" rel=\"nofollow\">Mozilla Public License<\/a><\/li>\n<li><code>Multiple<\/code> - Various different licenses, for different components of the project's software.<\/li>\n<li><code>OSL-3.0<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/osl-3.0.php\" rel=\"nofollow\">Open Software License 3.0<\/a><\/li>\n<li><code>Other<\/code> - Non-standard license, usually unique to the project itself.<\/li>\n<li><code>Sendmail<\/code> - <a href=\"https:\/\/www.sendmail.com\/pdfs\/open_source\/sendmail_license.pdf\" rel=\"nofollow\">Sendmail License<\/a><\/li>\n<li><code>Unlicense<\/code> - <a href=\"http:\/\/unlicense.org\/\" rel=\"nofollow\">The Unlicense<\/a><\/li>\n<li><code>WTFPL<\/code> - <a href=\"http:\/\/www.wtfpl.net\/about\/\" rel=\"nofollow\">Do What the Fuck You Want to Public License<\/a><\/li>\n<li><code>Zlib<\/code> - <a href=\"https:\/\/opensource.org\/licenses\/Zlib\" rel=\"nofollow\">Zlib\/libpng License<\/a><\/li>\n<li><code>ZPL-2.0<\/code> - <a href=\"http:\/\/opensource.org\/licenses\/ZPL-2.0\" rel=\"nofollow\">Zope Public License 2.0<\/a><\/li>\n<\/ul><hr \/><h2><a id=\"user-content-external-links\" class=\"anchor\" aria-hidden=\"true\" href=\"#external-links\"><\/a>External links<\/h2>\n<p><strong><a href=\"#\"><code>^        back to top        ^<\/code><\/a><\/strong><\/p>\n<ul><li><a href=\"https:\/\/github.com\/onurakpolat\/awesome-bigdata\">Awesome Big Data<\/a> - Curated list of awesome big data frameworks, resources and other awesomeness.<\/li>\n<li><a href=\"https:\/\/github.com\/awesomedata\/awesome-public-datasets\">Awesome Public Datasets<\/a> - List of high quality, topic-centric public data sources.<\/li>\n<li><a href=\"https:\/\/github.com\/n1trux\/awesome-sysadmin\">Awesome Sysadmin<\/a> - Curated list of amazingly awesome open source sysadmin resources.<\/li>\n<li>Lists of software aimed at privacy and decentralization in some form: <a href=\"https:\/\/prism-break.org\/en\/\" rel=\"nofollow\">PRISM Break<\/a>, <a href=\"https:\/\/www.privacytools.io\/\" rel=\"nofollow\">privacytools.io<\/a>, <a href=\"https:\/\/redecentralize.github.io\/alternative-internet\/\" rel=\"nofollow\">Alternative Internet<\/a>, <a href=\"http:\/\/libreprojects.net\/\" rel=\"nofollow\">Libre Projects<\/a><\/li>\n<li>Dynamic Domain Name services: <a href=\"https:\/\/freedns.afraid.org\/domain\/registry\/\" rel=\"nofollow\">Afraid.org<\/a>, <a href=\"https:\/\/pagekite.net\/\" rel=\"nofollow\">Pagekite<\/a><\/li>\n<li>Communities\/forums: <a href=\"https:\/\/www.reddit.com\/r\/selfhosted\" rel=\"nofollow\">\/r\/selfhosted<\/a>, <a href=\"https:\/\/indieweb.org\/\" rel=\"nofollow\">IndieWeb<\/a><\/li>\n<\/ul><hr \/><h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>Contributing guidelines can be found <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/.github\/CONTRIBUTING.md\">here<\/a>.<\/p>\n<h2><a id=\"user-content-authors\" class=\"anchor\" aria-hidden=\"true\" href=\"#authors\"><\/a>Authors<\/h2>\n<p>The list of authors can be found <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/AUTHORS.md\">here<\/a>.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>This list is under the <a href=\"https:\/\/github.com\/Kickball\/awesome-selfhosted\/blob\/master\/LICENSE\">Creative Commons Attribution-ShareAlike 3.0 Unported<\/a> License.<\/p>","created_at":"2018-11-07T16:22:18+0000","updated_at":"2018-11-07T16:22:18+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":90,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/2252004?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":64,"title":"shaarli\/Shaarli","url":"https:\/\/github.com\/shaarli\/Shaarli","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/shaarli\/Shaarli\/blob\/master\/doc\/md\/images\/doc-logo.png\"><img src=\"https:\/\/github.com\/shaarli\/Shaarli\/raw\/master\/doc\/md\/images\/doc-logo.png\" alt=\"Shaarli logo\" \/><\/a><\/p>\n<p>The personal, minimalist, super-fast, database free, bookmarking service.<\/p>\n<p><em>Do you want to share the links you discover?<\/em>\n<em>Shaarli is a minimalist link sharing service that you can install on your own server.<\/em>\n<em>It is designed to be personal (single-user), fast and handy.<\/em><\/p>\n<p><a href=\"https:\/\/github.com\/shaarli\/Shaarli\/releases\/tag\/v0.9.7\"><img src=\"https:\/\/camo.githubusercontent.com\/9044983daba4883937d6132cf39312714bf9764c\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f737461626c652d76302e392e372d626c75652e737667\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/stable-v0.9.7-blue.svg\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/shaarli\/Shaarli\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/4dbc0f482b5cff11c396bfcbe91882c619fbd8f7\/68747470733a2f2f696d672e736869656c64732e696f2f7472617669732f73686161726c692f53686161726c692f737461626c652e7376673f6c6162656c3d737461626c65\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/travis\/shaarli\/Shaarli\/stable.svg?label=stable\" \/><\/a>\n\u2022\n<a href=\"https:\/\/github.com\/shaarli\/Shaarli\/releases\/tag\/v0.10.2\"><img src=\"https:\/\/camo.githubusercontent.com\/0d3f95537b6e732937594a899c2937d2b1639213\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c61746573742d76302e31302e322d626c75652e737667\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/latest-v0.10.2-blue.svg\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/shaarli\/Shaarli\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5bfceb0b71208f22f972125804978f82dbb5c498\/68747470733a2f2f696d672e736869656c64732e696f2f7472617669732f73686161726c692f53686161726c692f6c61746573742e7376673f6c6162656c3d6c6174657374\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/travis\/shaarli\/Shaarli\/latest.svg?label=latest\" \/><\/a>\n\u2022\n<a href=\"https:\/\/github.com\/shaarli\/Shaarli\"><img src=\"https:\/\/camo.githubusercontent.com\/2ff8d2bcd365273463cff30ebc786deadda9c109\/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6d61737465722d76302e31302e782d626c75652e737667\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/badge\/master-v0.10.x-blue.svg\" \/><\/a>\n<a href=\"https:\/\/travis-ci.org\/shaarli\/Shaarli\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/f15f771b7f847768674327add71b31078640e2dc\/68747470733a2f2f696d672e736869656c64732e696f2f7472617669732f73686161726c692f53686161726c692e7376673f6c6162656c3d6d6173746572\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/travis\/shaarli\/Shaarli.svg?label=master\" \/><\/a><\/p>\n<p><a href=\"https:\/\/gitter.im\/shaarli\/Shaarli\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/da2edb525cde1455a622c58c0effc3a90b9a181c\/68747470733a2f2f6261646765732e6769747465722e696d2f4a6f696e253230436861742e737667\" alt=\"Join the chat at https:\/\/gitter.im\/shaarli\/Shaarli\" data-canonical-src=\"https:\/\/badges.gitter.im\/Join%20Chat.svg\" \/><\/a>\n<a href=\"https:\/\/www.bountysource.com\/teams\/shaarli\/issues\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/b5905ff4b14ab26b6e42272646373b86bcb281ca\/68747470733a2f2f7777772e626f756e7479736f757263652e636f6d2f62616467652f7465616d3f7465616d5f69643d3139353833267374796c653d626f756e746965735f7265636569766564\" alt=\"Bountysource\" data-canonical-src=\"https:\/\/www.bountysource.com\/badge\/team?team_id=19583&amp;style=bounties_received\" \/><\/a>\n<a href=\"https:\/\/hub.docker.com\/r\/shaarli\/shaarli\/\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/c1c2f7834426c784a937707250ec419a26f3800f\/68747470733a2f2f696d672e736869656c64732e696f2f646f636b65722f70756c6c732f73686161726c692f73686161726c692e737667\" alt=\"Docker repository\" data-canonical-src=\"https:\/\/img.shields.io\/docker\/pulls\/shaarli\/shaarli.svg\" \/><\/a><\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#demo\"><\/a>Demo<\/h3>\n<p>You can use this <a href=\"https:\/\/demo.shaarli.org\" rel=\"nofollow\">public demo instance of Shaarli<\/a>.\nIt runs the latest development version of Shaarli and is updated\/reset daily.<\/p>\n<p>Login: <code>demo<\/code>; Password: <code>demo<\/code><\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h3>\n<p>Shaarli is <a href=\"http:\/\/en.wikipedia.org\/wiki\/Free_software\" rel=\"nofollow\">Free Software<\/a>. See <a href=\"https:\/\/github.com\/shaarli\/Shaarli\/blob\/master\/COPYING\">COPYING<\/a> for a detail of the contributors and licenses for each individual component.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:17+0000","updated_at":"2018-11-07T16:22:17+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/8275533?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":63,"title":"Lisp for the Web","url":"https:\/\/www.adamtornhill.com\/articles\/lispweb.htm","content":"<p>\n          <img src=\"https:\/\/www.adamtornhill.com\/articles\/lispwebpics\/title_page.jpg\" width=\"234\" height=\"324\" alt=\"Lisp for the web - the book\" \/><\/p><p> The following article and code has been updated as the book\n      <a href=\"https:\/\/leanpub.com\/lispweb\">Lisp for the Web<\/a>. You\n      can get it for any price you want\n      at <a href=\"https:\/\/leanpub.com\/lispweb\">Leanpub<\/a>.\n      <\/p><p>\n        With his essay \n        <a href=\"http:\/\/www.paulgraham.com\/avg.html\" target=\"_top\">\n          Beating the Averages\n        <\/a>\n        , Paul Graham told the story of how his web start-up Viaweb outperformed its competitors by using Lisp. Lisp? Did I parse that correctly? That ancient language with all those scary parentheses? Yes, indeed! And with the goal of identifying its strengths and what they can do for us, I'll put Lisp to work developing a web application. In the process we'll find out how a 50 years old language can be so well-suited for modern web development and yes, it's related to all those parentheses.\n      <\/p><h2>\n        What to expect\n      <\/h2><p>\n        Starting from scratch, we'll develop a three-tier web application. I'll show how to:\n      <\/p><ul><li>\n          utilize powerful open source libraries for expressing dynamic HTML and JavaScript in Lisp,\n        <\/li>\n        <li>\n          develop a small, embedded domain specific language tailored for my application,\n        <\/li>\n        <li>\n          extend the typical development cycle by modifying code in a running system and execute code during compilation,\n        <\/li>\n        <li>\n          and finally migrate from data structures in memory to persistent objects using a third party database.\n        <\/li>\n      <\/ul><p>\n        I'll do this in a live system transparent to the users of the application. Because Lisp is so high-level, I'll be able to achieve everything in just around 70 lines of code.\n      <\/p><p>\n        This article will not teach you Common Lisp (for that purpose I recommend \n        <a href=\"https:\/\/www.adamtornhill.com\/reviews\/pcl.htm\" target=\"_top\">\n          Practical Common Lisp\n        <\/a>\n        ). Instead, I'll give a short overview of the language and try to explain the concepts as I introduce them, just enough to follow the code. The idea is to convey a feeling of how it is to develop in Lisp rather than focusing on the details.\n      <\/p><h2>\n        The Lisp story\n      <\/h2><p>\n        Lisp is actually a family of languages discovered by John McCarthy 50 years ago. The characteristic of Lisp is that Lisp code is made out of Lisp data structures with the practical implication that it is not only natural, but also highly effective to write programs that write programs. This feature has allowed Lisp to adapt over the years. For example, as object-oriented programming became popular, powerful object systems could be implemented in Lisp as libraries without any change to the core language. Later, the same proved to be true for aspect-oriented programming.\n      <\/p><p>\n        This idea is not only applicable to whole paradigms of programming. Its true strength lays in solving everyday problems. With Lisp, it's straightforward to build-up a domain specific language allowing us to program as close to the problem domain as our imagination allows. I'll illustrate the concept soon, but before we kick-off, let's look closer at the syntax of Lisp.\n      <\/p><h2>\n        Crash course in Lisp\n      <\/h2><p>\n        What Graham used for Viaweb was Common Lisp, an ANSI standardized language, which we'll use in this article too (the other main contender is Scheme, which is considered cleaner and more elegant, but with a much smaller library).\n      <\/p><p>\n        Common Lisp is a high-level interactive language that may be either interpreted or compiled. You interact with Lisp through its \n        <i>\n          top-level\n        <\/i>\n        . The top-level is basically a prompt. On my system it looks like this:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;\n      <\/pre><p>\n        Through the top-level, we can enter expressions and see the results (the values returned by the top-level are \n        <b>\n          highlighted\n        <\/b>\n        ):\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(+ 1 2 3)\n\t<b>6<\/b>\n      <\/pre><p>\n        As we see in the example, Lisp uses a prefix notation. A parenthesized expression is referred to as a \n        <i>\n          form\n        <\/i>\n        . When fed a form such as \n        \n          (+ 1 2 3)\n        \n        , Lisp generally treats the first element (+) as a function and the rest as arguments. The arguments are evaluated from left to right and may themselves be function calls:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(+ 1 2 (\/ 6 2))\n\t<b>6<\/b>\n      <\/pre><p>\n        We can <b>de<\/b>fine our own <b>fun<\/b>ctions with defun:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(defun say-hello (to)\n                    (format t \"Hello, ~a\" to))\n      <\/pre><p>\n        Here we're defining a function \n        \n          say-hello\n        \n        , taking one argument: \n        \n          to\n        \n        . The format function is used to print a greeting and resembles a \n        \n          printf\n        \n         on steroids. Its first argument is the output stream and here we're using \n        \n          t\n        \n         as a shorthand for standard output. The second argument is a string, which in our case contains an embedded directive \n        \n          ~a\n        \n         instructing \n        \n          format\n        \n         to consume one argument and output it in human-readable form. We can call our function like this:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(say-hello \"ACCU\")\n\t<b>Hello, ACCU<\/b>\n\t<b>NIL<\/b>\n      <\/pre><p>\n        The first line is the side-effect, printing \"Hello, ACCU\" and NIL is the return value from our function. By default, Common Lisp returns the value of the last expression. From here we can redefine say-hello to return its greeting instead:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(defun say-hello (to)\n                     (format nil \"Hello, ~a\" to))\n      <\/pre><p>\n        With nil as its destination, format simply returns its resulting string:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(say-hello \"ACCU\")\n\t<b>\"Hello, ACCU\"<\/b>\n      <\/pre><p>\n        Now we've gotten rid of the side-effect. Programming without side-effects is in the vein of functional programming, one of the paradigms supported by Lisp. Lisp is also dynamically typed. Thus, we can feed our function a number instead:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(say-hello 42)\n\t<b>\"Hello, 42\"<\/b>\n      <\/pre><p>\n        In Lisp, functions are first-class citizens. That means, we can create them just like any other object and we can pass them as arguments to other functions. Such functions taking functions as arguments are called \n        <i>\n          higher-order functions\n        <\/i>\n        . One example is \n        \n          mapcar\n        \n        . \n        \n          mapcar\n        \n         takes a function as its first argument and applies it subsequently to the elements of one or more given lists:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(mapcar #'say-hello (list \"ACCU\" 42 \"Adam\"))\n\t<b>(\"Hello, ACCU\" \"Hello, 42\" \"Hello, Adam\")<\/b>\n      <\/pre><p>\n        The funny #' is just a shortcut for getting at the function object. As you see above, \n        \n          mapcar\n        \n         collects the result of each function call into a list, which is its return value. This return value may of course serve as argument to yet another function:\n      <\/p><pre class=\"lispcode\">\n        CL-USER&gt;(sort (mapcar #'say-hello (list \"ACCU\" 42 \"Adam\")) #'string-lessp)\n\t<b>(\"Hello, 42\" \"Hello, ACCU\" \"Hello, Adam\")<\/b>\n      <\/pre><p>\n        Lisp itself isn't hard, although it may take some time to wrap ones mindset around the functional style of programming. As you see, Lisp expressions are best read inside-out. But the real secret to understanding Lisp syntax is to realize that the language doesn't have one; what we've been entering above is basically parse-trees, generated by compilers in other languages. And, as we'll see soon, exactly this feature makes it suitable for metaprogramming.\n      <\/p><h2>\n        The Brothers are History\n      <\/h2><p>\n        Remember the hot gaming discussions 20 years ago? \"Giana Sisters\" really was way better than \"Super Mario Bros\", wasn't it? We'll delegate the question to the wise crowd by developing a web application. Our web application will allow users to add and vote for their favourite retro games. A screenshot of the end result is provided in Figure 1 below.\n      <\/p><p>\n        <img src=\"https:\/\/www.adamtornhill.com\/articles\/lispwebpics\/Figure1.png\" alt=\"Retro Games front page\" width=\"800\" height=\"625\" \/><\/p><p>\n        From now on, I start to persist my Lisp code in textfiles instead of just entering expressions into the top-level. Further, I define a \n        <i>\n          package\n        <\/i>\n         for my code. Packages are similar to namespaces in C++ or Java's packages and helps to prevent name collisions (the main distinction is that packages in Common Lisp are first-class objects).\n      <\/p><pre class=\"lispcode\">\n        (defpackage :retro-games\n             (:use :cl :cl-who :hunchentoot :parenscript))\n      <\/pre><p>\n        The new package is named \n        \n          :retro-games\n        \n         and I also specify other packages that we'll use initially:\n      <\/p><ul><li>\n          CL is Common Lisp's standard package containing the whole language.\n        <\/li>\n        <li>\n          <a href=\"http:\/\/weitz.de\/cl-who\/\" target=\"_top\">\n            CL-WHO\n          <\/a>\n           is a library for converting Lisp expressions into XHTML.\n        <\/li>\n        <li>\n          <a href=\"http:\/\/weitz.de\/hunchentoot\/\" target=\"_top\">\n            Hunchentoot\n          <\/a>\n           is a web-server, written in Common Lisp itself, and provides a toolkit for building dynamic web sites.\n        <\/li>\n        <li>\n          <a href=\"http:\/\/common-lisp.net\/project\/parenscript\/\" target=\"_top\">\n            ParenScript\n          <\/a>\n           allows us to compile Lisp expressions into JavaScript. We'll use this for client-side validation.\n        <\/li>\n      <\/ul><p>\n        With my package definition in place, I'll put the rest of the code inside it by switching to the \n        \n          :retro-games\n        \n         package:\n      <\/p><pre class=\"lispcode\">\n        (in-package :retro-games)\n      <\/pre><p>\n        Most top levels indicate the current package in their prompt. On my system the prompt now looks like this:\n      <\/p><pre class=\"lispcode\">\n        RETRO-GAMES&gt;\n      <\/pre><h2>\n        Representing Games\n      <\/h2><p>\n        With the package in place, we can return to the problem. It seems to require some representation of a game and I'll choose to abstract it as a class:\n      <\/p><pre class=\"lispcode\">\n        (defclass game ()\n            ((name  :initarg  :name)\n             (votes :initform 0)))\n      <\/pre><p>\n        The expression above defines the class \n        \n          game\n        \n         without any user-specified superclasses, hence the empty list \n        \n          ()\n        \n         as second argument. A game has two \n        <i>\n          slots\n        <\/i>\n         (slots are similar to attributes or members in other languages): a \n        \n          name\n        \n         and the number of accumulated \n        \n          votes\n        \n        . To create a game object I invoke \n        \n          make-instance\n        \n         and passes it the name of the class to instantiate:\n      <\/p><pre class=\"lispcode\">\n        RETRO-GAMES&gt;(setf many-lost-hours (make-instance 'game :name \"Tetris\"))\n\t<b>#&lt;GAME @ #x7213da32&gt;<\/b>\n      <\/pre><p>\n        Because I specified an <b>init<\/b>ial <b>arg<\/b>ument in my definition of the \n        \n          name\n        \n         slot, I can pass this argument directly and initialize that slot to \"Tetris\". The \n        \n          votes\n        \n         slot doesn't have an initial argument. Instead I specify the code I want to run during instantiation to compute its initial value through \n        \n          :initform\n        \n        . In this case the code is trivial, as I only want to initialize the number of votes to zero. Further, I use \n        \n          setf\n        \n         to assign the object created by \n        \n          make-instance\n        \n         to the variable many-lost-hours.\n      <\/p><p>\n        Now that we got an instance of game we would like to do something with it. We could of course write code ourselves to access the slots. However, there's a more lispy way; \n        \n          defclass\n        \n         provides the possibility to automatically generate accessor functions for our slots:\n      <\/p><pre class=\"lispcode\">\n        (defclass game ()\n          ((name  <b>:reader   name<\/b> \n                  :initarg  :name)\n           (votes <b>:accessor votes<\/b> \n                  :initform 0)))\n      <\/pre><p>\n        The option \n        \n          :reader\n        \n         in the name slot will automatically create a read function and the option \n        \n          :accessor\n        \n         used for the \n        \n          votes\n        \n         slot will create both read and write functions. Lisp is pleasantly uniform in its syntax and these generated functions are invoked just like any other function:\n      <\/p><pre class=\"lispcode\">\n        RETRO-GAMES&gt;(name many-lost-hours)\n        <b>\"Tetris\"<\/b>\n        RETRO-GAMES&gt;(votes many-lost-hours)\n        <b>0<\/b>\n        RETRO-GAMES&gt;(incf (votes many-lost-hours))\n        <b>1<\/b>\n        RETRO-GAMES&gt;(votes many-lost-hours)\n        <b>1<\/b>\n      <\/pre><p>\n        The only new function here is \n        \n          incf\n        \n        , which when given one argument increases its value by one. We can encapsulate this mechanism in a method used to vote for the given game:\n      <\/p><pre class=\"lispcode\">\n        (defmethod vote-for (user-selected-game)\n\t   (incf (votes user-selected-game)))\n      <\/pre><p>\n        The top-level allows us to immediately try it out and vote for Tetris:\n      <\/p><pre class=\"lispcode\">\n        \n       RETRO-GAMES&gt;(votes many-lost-hours)\n       <b>1<\/b>\n       RETRO-GAMES&gt;(vote-for many-lost-hours)\n       <b>2<\/b>\n       RETRO-GAMES&gt;(votes many-lost-hours)\n       <b>2<\/b>\n      <\/pre><h2>\n        A prototypic back end\n      <\/h2><p>\n        Before we can jump into the joys of generating web pages, we need a back end for our application. Because Lisp makes it so easy to modify existing applications, it's common to start out really simple and let the design evolve as we learn more about the problem we're trying to solve. Thus, I'll start by using a list in memory as a simple, non-persistent storage.\n      <\/p><pre class=\"lispcode\">\n        (defvar *games* '())\n      <\/pre><p>\n        The expression above defines and initializes the global variable (actually the Lisp term is <i>special variable<\/i>) \n        \n          *games*\n        \n         to an empty list. The asterisks aren't part of the syntax; it's just a naming convention for globals. Lists may not be the most efficient data structure for all problems, but Common Lisp has great support for lists and they are easy to work with. Later we'll change to a real database and, with that in mind, I encapsulate the access to \n        \n          *games*\n        \n         in some small functions:\n      <\/p><pre class=\"lispcode\">\n        (defun game-from-name (name)\n\t  (find name *games* :test #'string-equal \n\t                     :key  #'name))\n      <\/pre><p>\n        Our first function \n        \n          game-from-name\n        \n         is implemented in terms of \n        \n          find\n        \n        . \n        \n          find\n        \n         takes an item and a sequence. Because we're comparing strings I tell find to use the function \n        \n          string-equal\n        \n         for comparison (remember, #' is a short cut to refer to a function). I also specify the key to compare. In this case, we're interested in comparing the value returned by the \n        \n          name\n        \n         function on each \n        \n          game\n        \n         object.\n      <\/p><p>\n        If there's no match \n        \n          find\n        \n         returns \n        \n          NIL\n        \n        , which evaluates to false in a boolean context. That means we can reuse \n        \n          game-from-name\n        \n         when we want to know if a game is stored in the \n        \n          *games*\n        \n         list. However, we want to be clear with our intent:\n      <\/p><pre class=\"lispcode\">\n        (defun game-stored? (game-name)\n\t (game-from-name game-name))\n      <\/pre><p>\n        As illustrated in Figure 1, we want to present the games sorted on popularity. Using Common Lisp's \n        \n          sort\n        \n         function this is pretty straightforward; we only have to take care, because for efficiency reasons \n        \n          sort\n        \n         is destructive. That is, \n        \n          sort\n        \n         is allowed to modify its argument.  We can preserve our \n        \n          *games*\n        \n         list by passing a copy to \n        \n          sort\n        \n        . I tell \n        \n          sort\n        \n         to return a list sorted in descending order based on the value returned by the \n        \n          votes\n        \n         function invoked on each game:\n      <\/p><pre class=\"lispcode\">\n        (defun games ()\n\t (sort (copy-list *games*) #'&gt; :key #'votes))\n      <\/pre><p>\n        So far the queries. Let's define one more utility for actually adding games to our storage:\n      <\/p><pre class=\"lispcode\">\n        (defun add-game (name)\n\t (unless (game-stored? name)\n\t   (push (make-instance 'game :name name) *games*)))\n      <\/pre><p>\n        \n          push\n        \n         is a modifying operation and it prepends the game instantiated by \n        \n          make-instance\n        \n         to the \n        \n          *games*\n        \n         list. Let's try it all out at the top level.\n      <\/p><pre class=\"lispcode\">\n        \n      RETRO-GAMES&gt;(games)\n      <b>NIL<\/b>\n      RETRO-GAMES&gt;(add-game \"Tetris\")\n      <b>(#&lt;GAME @ #x71b943c2&gt;)<\/b>\n      RETRO-GAMES&gt;(game-from-name \"Tetris\")\n      <b>#&lt;GAME @ #x71b943c2&gt;<\/b>\n      RETRO-GAMES&gt;(add-game \"Tetris\")\n      <b>NIL<\/b>\n      RETRO-GAMES&gt;(games)\n      <b>(#&lt;GAME @ #x71b943c2&gt;)<\/b>\n      RETRO-GAMES&gt;(mapcar #'name (games))\n      <b>(\"Tetris\")<\/b>\n      <\/pre><p>\n        The values returned to the top level may not look too informative. It's basically the printed representation of a game object. Common Lisp allows us to customize how an object shall be printed, but we will not go into the details. Instead, with this prototypic back end in place, we're prepared to enter the web.\n      <\/p><h2>\n        Generating HTML dynamically\n      <\/h2><p>\n        The first step in designing an embedded domain specific language is to find a Lisp representation of the target language. For HTML this is really simple as both HTML and Lisp are represented in tree structures, although Lisp is less verbose. Here's an example using the CL-WHO library:\n      <\/p><pre class=\"lispcode\">\n        \n       (with-html-output (*standard-output* nil :indent t)\n\t   (:html\n              (:head\n                 (:title \"Test page\"))\n              (:body\n                 (:p \"CL-WHO is really easy to use\"))))\n      <\/pre><p>\n        This code will expand into the following HTML, which is outputted to \n        \n          *standard-output*\n        \n        :\n      <\/p><pre class=\"lispcode\">\n        \n       &lt;html&gt;\n          &lt;head&gt;\n             &lt;title&gt;Test page &lt;\/title&gt;\n          &lt;\/head&gt;\n          &lt;body&gt;\n             &lt;p&gt; CL-WHO is really easy to use &lt;\/p&gt;\n          &lt;\/body&gt;\n       &lt;\/html&gt;\n      <\/pre><p>\n        CL-WHO also allows us to embed Lisp expressions, setting the scene for dynamic web pages.\n      <\/p><h2>\n        Macros: Fighting the evils of code duplication\n      <\/h2><p>\n        Although CL-WHO does provide a tighter representation than raw HTML we're still facing the potential risk of code duplication; the html, head, and body tags form a pattern that will recur on all pages. And it'll only get worse as we start to write strict and validating XHTML 1.0, where we have to include more tags and attributes and, of course, start every page with that funny DOCTYPE line.\n      <\/p><p>\n        Further, if you look at Figure 1 you'll notice that the retro games page has a header with a picture of that lovely Commodore (photo by Bill Bertram - thanks!) and a strap line. I want to be able to define that header once and have all my pages using it automatically. \nThe problem screams for a suitable abstraction and this is where Lisp differs from other languages. In Lisp, we can actually take on the role of a language designer and extend the language with our own syntax. The feature that allows this is macros. Syntactically, macros look like functions, but are entirely different beasts. Sure, just like functions macros take arguments. The difference is that the arguments to macros are source code, because macros are used by the compiler to generate code.\n      <\/p><p>\n        Macros can be a conceptual challenge as they erase the line between compile time and runtime. What macros do are expanding themselves into code that are actually compiled. In their expansion macros have access to the whole language, including other macros, and may call functions, create objects, etc.\n      <\/p><p>\n        So, let's put this amazing macro mechanism to work by defining a new syntactic construct, the \n        \n          standard-page\n        \n        . A \n        \n          standard-page\n        \n         will abstract away all XHTML boiler plate code and automatically generate the heading on each page. The macro will take two arguments. The first is the title of the page and the second the code defining the body of the actual web-page. Here's a simple usage example:\n      <\/p><pre class=\"lispcode\">\n        \n       (standard-page (:title \"Retro Games\")\n\t\t      (:h1 \"Top Retro Games\")\n\t\t      (:p \"We'll write the code later...\"))\n      <\/pre><p>\n        Much of the macro will be straightforward CL-WHO constructs. Using the backquote syntax(the ` character), we can specify a template for the code we want to generate:\n      <\/p><pre class=\"lispcode\">\n        \n       (defmacro standard-page ((&amp;key title) &amp;body body)\n\t `(with-html-output-to-string (*standard-output* nil :prologue t :indent t)\n\t   (:html :xmlns \"http:\/\/www.w3.org\/1999\/xhtml\"\n\t\t  :xml\\:lang \"en\" \n\t\t  :lang \"en\"\n\t\t  (:head \n\t\t   (:meta :http-equiv \"Content-Type\" \n\t\t\t  :content    \"text\/html;charset=utf-8\")\n\t\t   (:title <b>,title<\/b>)\n\t\t   (:link :type \"text\/css\" \n\t\t\t  :rel \"stylesheet\"\n\t\t\t  :href \"\/retro.css\"))\n\t\t  (:body \n\t\t   (:div :id \"header\" ; Retro games header\n\t\t\t (:img :src \"\/logo.jpg\" \n\t\t\t       :alt \"Commodore 64\" \n\t\t\t       :class \"logo\")\n\t\t\t (:span :class \"strapline\" \n\t\t\t\t\"Vote on your favourite Retro Game\"))\n\t\t   ,@<b>body<\/b>))))\n      <\/pre><p>\n        Within the backquoted expression we can use \n        \n          ,\n        \n         (comma) to evaluate an argument and \n        \n          ,@\n        \n         (comma-at) to evaluate and splice a list argument. Remember, the arguments to a macro are code. In this example the first argument \n        \n          title\n        \n         is bound to \"Retro Games\" and the second argument \n        \n          body\n        \n         contains the \n        \n          :h1\n        \n         and \n        \n          :p\n        \n         expressions wrapped-up in a list. In the macro definition, the code bound to these arguments is simply inserted on the proper places in our backquoted template code.\n      <\/p><p>\n        The power we get from macros become evident as we look at the generated code. The three lines in the usage example above expands into this (note that Lisp symbols are case-insensitive and thus usually presented in uppercase):\n      <\/p><pre class=\"lispcode\">\n        \n       (WITH-HTML-OUTPUT-TO-STRING (*STANDARD-OUTPUT* NIL :PROLOGUE T :INDENT T)  \n\t(:HTML :XMLNS \"http:\/\/www.w3.org\/1999\/xhtml\"\n\t       :|XML:LANG| \"en\" \n\t       :LANG \"en\"\n\t       (:HEAD\n\t\t(:META :HTTP-EQUIV \"Content-Type\" \n\t\t       :CONTENT \"text\/html;charset=utf-8\")\n\t\t<b>(:TITLE \"Retro Games\")<\/b>\n\t\t(:LINK :TYPE \"text\/css\" \n\t\t       :REL \"stylesheet\" \n\t\t       :HREF \"\/retro.css\"))\n\t       (:BODY\n\t\t(:DIV :ID \"header\"\n\t\t      (:IMG :SRC \"\/logo.jpg\" \n\t\t\t    :ALT \"Commodore 64\" \n\t\t\t    :CLASS \"logo\")\n\t\t      (:SPAN :CLASS \"strapline\" \n\t\t\t     \"Vote on your favourite Retro Game\"))\n\t\t<b>(:H1 \"Top Retro Games\")\n\t\t(:P \"We'll write the code later...\")<\/b>)))\n      <\/pre><p>\n        This is a big win; all this is code that we don't have to write. Now that we have a concise way to express web-pages with a uniform look, it's time to introduce Hunchentoot.\n      <\/p><h2>\n        More than an opera\n      <\/h2><p>\n        Named after a Zappa sci-fi opera, Edi Weitz's Hunchentoot is a full featured web-server written in Common Lisp. To launch Hunchentoot, we just invoke its \n        \n          start-server\n        \n         function:\n      <\/p><pre class=\"lispcode\">\n        RETRO-GAMES&gt;(start-server :port 8080)\n      <\/pre><p>\n        \n          start-server\n        \n         supports several arguments, but we're only interested in specifying a port other than the default port 80. And that's it - the server's up and running. We can test it by pointing a web browser to http:\/\/localhost:8080\/, which should display Hunchentoot's default page. To actually publish something, we have to provide Hunchentoot with a <i>handler<\/i>. In Hunchentoot all requests are dynamically dispatched to an associated handler and the framework contains several functions for defining dispatchers. The code below creates a dispatcher and adds it to Hunchentoot's dispatch table:\n      <\/p><pre class=\"lispcode\">\n        (push (create-prefix-dispatcher \"\/retro-games.htm\" 'retro-games) *dispatch-table*)\n      <\/pre><p>\n        The dispatcher will invoke the function, \n        \n          retro-games\n        \n        ,  whenever an URI request starts with \n        \n          \/retro-games.htm\n        \n        . Now we just have to define the \n        \n          retro-games\n        \n         function that generates the HTML:\n      <\/p><pre class=\"lispcode\">\n        \n       (defun retro-games ()\n\t (standard-page (:title \"Retro Games\")\n\t\t\t(:h1 \"Top Retro Games\")\n\t\t\t(:p \"We'll write the code later...\")))\n      <\/pre><p>\n        That's it - the retro games page is online. But I wouldn't be quick to celebrate; while we took care to abstract away repetitive patterns in \n        \n          standard-page\n        \n        , we've just run into another more subtle form of duplication. The problem is that every time we want to create a new page we have to explicitly create a dispatcher for our handle. Wouldn't it be nice if Lisp could do that automatically for us? Basically I want to define a function like this:\n      <\/p><pre class=\"lispcode\">\n        \n       (define-url-fn (retro-games)\n\t (standard-page (:title \"Retro Games\")\n\t\t\t(:h1 \"Top Retro Games\")\n\t\t\t(:p \"We'll write the code later...\")))\n      <\/pre><p>\n        and have Lisp to create a handler, associate it with a dispatcher and put it in the dispatch table as I compile the code. Guess what, using macros the syntax is ours. All we have to do is reformulate our wishes in a \n        \n          defmacro\n        \n        :\n      <\/p><pre class=\"lispcode\">\n        \n       (defmacro define-url-fn ((name) &amp;body body)\n\t `(progn\n\t    (defun ,<b>name<\/b> ()\n\t      ,@<b>body<\/b>)\n\t    (push (create-prefix-dispatcher ,(format nil \"\/~(~a~).htm\" <b>name<\/b>) ',<b>name<\/b>) *dispatch-table*)))\n      <\/pre><p>\n        Now our \"wish code\" above actually compiles and generates the following Lisp code (macro arguments <b>highlighted<\/b>):\n      <\/p><pre class=\"lispcode\">\n        \n       (PROGN \n\t(DEFUN <b>RETRO-GAMES ()\n\t  (STANDARD-PAGE (:TITLE \"Retro Games\")\n\t\t\t (:H1 \"Top Retro Games\")\n\t\t\t (:P \"We'll write the code later...\")))<\/b>\n\t(PUSH (CREATE-PREFIX-DISPATCHER \"\/<b>retro-games<\/b>.htm\" '<b>RETRO-GAMES<\/b>) *DISPATCH-TABLE*))\n      <\/pre><p>\n        There's a few interesting things about this macro:\n      <\/p><ol><li>\n          It illustrates that macros can take other macros as arguments. The Lisp compiler will continue to expand the macros  and \n          \n            standard-page\n          \n           will be expanded too, writing even more code for us.\n        <\/li>\n        <li>\n          Macros may execute code as they expand. The prefix string \"\/retro-games.htm\" is assembled with \n          \n            format\n          \n           during macro expansion time. By using comma, I evaluate the form and there's no trace of it in the generated code - just the resulting string.\n        <\/li>\n        <li>\n          A macro must expand into a single form, but we actually need two forms; a function definition and the code for creating a dispatcher. \n          \n            progn\n          \n           solves this problem by wrapping the forms in a single form and then evaluating them in order.\n        <\/li>\n      <\/ol><h2>\n        Putting it together\n      <\/h2><p>\n        Phew, that was a lot of Lisp in a short time. But using the abstractions we've created, we're able to throw together the application in no time. Let's code out the main page as it looks in Figure 1 above:\n      <\/p><pre class=\"lispcode\">\n        \n       (define-url-fn (retro-games)\n\t (standard-page (:title \"Top Retro Games\")\n\t\t\t(:h1 \"Vote on your all time favourite retro games!\")\n\t\t\t(:p \"Missing a game? Make it available for votes \" (:a :href \"new-game.htm\" \"here\"))\n\t\t\t(:h2 \"Current stand\")\n\t\t\t(:div :id \"chart\" ; <b>For CSS styling of links<\/b>\n\t\t\t      (:ol\n\t\t\t       (dolist (game (games))\n\t\t\t\t (htm  \n\t\t\t\t  (:li \n\t\t\t\t   (:a :href (format nil \"vote.htm?name=~a\" (name game)) \"Vote!\")\n\t\t\t\t   (fmt \"~A with ~d votes\" (name game) (votes game)))))))))\n      <\/pre><p>\n        Here we utilize our freshly developed embedded domain specific language for defining URL functions (\n        \n          define-url-fn\n        \n        ) and creating \n        \n          standard-pages\n        \n        . The following lines are straightforward XHTML generation, including a link to <i>new-game.htm<\/i>; a page we haven't specified yet. We will use some CSS to style the Vote! links to look and feel like buttons, which is why I wrap the list in a \n        \n          div\n        \n        -tag.\n      <\/p><p>\n        The first embedded Lisp code is \n        \n          dolist\n        \n        .  We use it to create each game item in the ordered HTML list. \n        \n          dolist\n        \n          works by iterating over a list, in this case the return value from the \n        \n          games\n        \n        -function, subsequently binding each element to the \n        \n          game\n        \n         variable. Using \n        \n          format\n        \n         and the access methods on the \n        \n          game\n        \n         object, I assemble the presentation and a destination for Vote!. Here's some sample HTML output from one session:\n      <\/p><pre class=\"lispcode\">\n        \n      &lt;div id='chart'&gt;\n         &lt;ol&gt;\n            &lt;li&gt;\n               &lt;a href='vote.htm?name=Super Mario Bros'&gt;Vote!&lt;\/a&gt; Super Mario Bros with 12 votes\n            &lt;\/li&gt;\n            &lt;li&gt;\n               &lt;a href='vote.htm?name=Last Ninja'&gt;Vote!&lt;\/a&gt; Last Ninja with 11 votes\n            &lt;\/li&gt;\n         &lt;\/ol&gt;\n      &lt;\/div&gt;\n      <\/pre><p>\n        As the user presses Vote! we'll get a request for \n        \n          vote.htm\n        \n         with the name of the game attached as a query parameter. Hunchentoot provides a \n        \n          parameter\n        \n         function that, just as you might expect, returns the value of the parameter named by the following string. We pass this value to our back end abstraction \n        \n          game-from-name\n        \n         and binds the result to a local variable with \n        \n          let\n        \n        :\n      <\/p><pre class=\"lispcode\">\n        \n       (define-url-fn (vote)\n\t (let ((game (game-from-name (parameter \"name\"))))\n\t   (if game\n\t       (vote-for game))\n\t   (redirect \"\/retro-games.htm\")))\n      <\/pre><p>\n        After a vote-for the requested game, Hunchentoot's \n        \n          redirect\n        \n         function takes the client to the updated chart.\n      <\/p><p>\n        <img src=\"https:\/\/www.adamtornhill.com\/articles\/lispwebpics\/Figure2.png\" alt=\"Retro Games add game page\" width=\"800\" height=\"627\" \/><\/p><p>\n        Now when we're able to vote we need some games to \n        \n          vote-for\n        \n        . In the code for the \n        \n          retro-games\n        \n         page above, I included a link to <i>new-game.htm<\/i>. That page is displayed in Figure 2. Basically it contains a HTML form with a text input for the game name:\n      <\/p><pre class=\"lispcode\">\n        \n       (define-url-fn (new-game)\n\t (standard-page (:title \"Add a new game\")\n\t\t\t(:h1 \"Add a new game to the chart\")\n\t\t\t(:form :action \"\/game-added.htm\" :method \"post\" \n\t\t\t       (:p \"What is the name of the game?\" (:br)\n\t\t\t\t   (:input :type \"text\"  \n\t\t\t\t\t   :name \"name\" \n\t\t\t\t\t   :class \"txt\"))\n\t\t\t       (:p (:input :type \"submit\" \n\t\t\t\t\t   :value \"Add\" \n\t\t\t\t\t   :class \"btn\")))))\n      <\/pre><p>\n        As the user submits the form, its data is sent to <i>game-added.htm<\/i>:\n      <\/p><pre class=\"lispcode\">\n        \n       (define-url-fn (game-added)\n\t (let ((name (parameter \"name\")))\n\t   (unless (or (null name) (zerop (length name)))\n\t     (add-game name))\n\t   (redirect \"\/retro-games.htm\")))\n      <\/pre><p>\n        The first line in our URL function should look familiar; just as in our \n        \n          vote\n        \n         function, we fetch the value of the name parameter and binds it to a local variable (\n        \n          name\n        \n        ). Here we have to guard against an empty name. After all, there's nothing forcing the user to write anything into the field before submitting the form (we'll see in a minute how to add client-side validation). If we get a valid name, we add it to our database through the \n        \n          add-game\n        \n         function.\n      <\/p><h2>\n        Expressing JavaScript in Lisp\n      <\/h2><p>\n        Say we want to ensure that the user at least typed something before submitting the form. Can we do that in Lisp? Yes, actually. We can write Lisp code that compiles into JavaScript and we use the  ParenScript library for the task.\n      <\/p><p>\n        Unobtrusive JavaScript is an important design principle and ParenScript supports that too. But in Lisp this becomes less of an issue; I'm not actually writing JavaScript, everything is Lisp. Thus I embed my event handler in the form:\n      <\/p><pre class=\"lispcode\">\n        \n      (:form :action \"\/game-added.htm\" :method \"post\" \n\t     :onsubmit \n\t     (ps-inline\n\t      (when (= name.value \"\")\n\t\t(alert \"Please enter a name.\")\n\t\t(return false)))\n      <\/pre><p>\n        This code will compile into the following mixture of HTML and JavaScript:\n      <\/p><pre class=\"lispcode\">\n        \n      &lt;form action='\/game-added.htm' method='post' \n         onsubmit='javascript:if (name.value == \"\") {\n            alert(\"Please enter a name.\");\n            return false;\n         }'&gt;\n      <\/pre><h2>\n        Persistent Objects\n      <\/h2><p>\n        Initially we kind of ducked the problem with persistence. To get things up and running as quickly as possible, we used a simple list in memory as \"database\". That's fine for prototyping but we still want to persist all added games in case we shutdown the server. Further, there are some potential threading issues with the current design. Hunchentoot is multithreaded and requests may come in different threads. We can solve all that by migrating to a thread-safe database. And with Lisp, design decisions like that are only a macro away; please meet Elephant!\n      <\/p><p>\n        <a href=\"http:\/\/common-lisp.net\/project\/elephant\/\" target=\"_top\">\n          Elephant\n        <\/a>\n         is a wickedly smart persistent object protocol and database. To actually store things on  disk, Elephant supports several back ends such as PostGres and SqlLite. In this example I'll use Berkeley DB, simply because it has the best performance with Elephant.\n      <\/p><p>\n        The first step is to open a store controller, which serves as a bridge between Lisp and the back end:\n      <\/p><pre class=\"lispcode\">\n        (open-store '(:BDB \"\/home\/adam\/temp\/gamedb\/\"))\n      <\/pre><p>\n        Here I just specify that we're using Berkely DB (\n        \n          :BDB\n        \n        ) and give a directory for the database files. Now, let's make some persistent objects. Have a look at our current \n        \n          game\n        \n         class again:\n      <\/p><pre class=\"lispcode\">\n        \n\t (defclass game ()\n\t   ((name :reader   name \n\t\t  :initarg :name)\n\t    (votes :accessor votes \n\t\t   :initform 0)))\n      <\/pre><p>\n        Elephant provides a convenient \n        \n          defpclass\n        \n         macro that creates persistent classes. The \n        \n          defpclass\n        \n         usage looks very similar to Common Lisp's \n        \n          defclass\n        \n        , but it adds some new features; we'll use \n        \n          :index\n        \n         to specify that we want our slots to be retrievable by slot values. I also add an <b>init<\/b>ial <b>arg<\/b>ument to \n        \n          votes\n        \n        , which I use later when transforming our old games into this persistent class:\n      <\/p><pre class=\"lispcode\">\n        \n\t (defpclass persistent-game ()\n\t   ((name :reader name \n\t\t  :initarg :name \n\t\t  :index t)\n\t    (votes :accessor votes \n\t\t   :initarg :votes \n\t\t   :initform 0 \n\t\t   :index t)))\n      <\/pre><p>\n        The Elephant abstraction is really clean; persistent objects are created just like any other object:\n      <\/p><pre class=\"lispcode\">\n        \n\t RETRO-GAMES&gt;(make-instance 'persistent-game :name \"Winter Games\")\n\t <b>#&lt;PERSISTENT-GAME oid:100&gt;<\/b>\n      <\/pre><p>\n        Elephant comes with a set of functions for easy retrieval. If we want all instances of our \n        \n          persistent-game\n        \n         class, it's a simple as this: \n      <\/p><pre class=\"lispcode\">\n        \n\t RETRO-GAMES&gt;(get-instances-by-class 'persistent-game)\n\t <b>(#&lt;PERSISTENT-GAME oid:100&gt;)<\/b>\n      <\/pre><p>\n        We can of course keep a reference to the returned list or, because we know we just instantiated a \n        \n          persistent-game\n        \n        , call a method on it directly:\n      <\/p><pre class=\"lispcode\">\n        \n\t RETRO-GAMES&gt;(name (first (get-instances-by-class 'persistent-game)))\n\t <b>\"Winter Games\"<\/b>\n      <\/pre><p>\n        We took care earlier to encapsulate the access to the back end and that pays off now. We just have to change those functions to use the Elephant API instead of working with our \n        \n          *games*\n        \n         list. The query functions are quit simple; because we indexed our \n        \n          name\n        \n         slot, we can use \n        \n          get-instance-by-value\n        \n         to get the matching persistent object:\n      <\/p><pre class=\"lispcode\">\n        \n\t  (defun game-from-name (name)\n\t    (get-instance-by-value 'persistent-game 'name name))\n      <\/pre><p>\n        Just like our initial implementation using \n        \n          find\n        \n        , \n        \n          get-instance-by-value\n        \n         returns \n        \n          NIL\n        \n         in case no object with the given name is stored. That means that we can keep \n        \n          game-stored?\n        \n         exactly as it is without any changes. But what about adding a new game? Well, we no longer need to maintain any references to the created objects. The database does that for us. But, we have to change \n        \n          add-game\n        \n         to make an instance of \n        \n          persistent-game\n        \n         instead of our old \n        \n          game\n        \n         class. And even though Elephant is thread-safe we have to ensure that the transactions are atomic. Elephant provides a nice \n        \n          with-transaction\n        \n         macro to solve this problem:\n      <\/p><pre class=\"lispcode\">\n        \n\t   (defun add-game (name)\n\t     (with-transaction ()\n\t\t\t(unless (game-stored? name)\n\t\t\t    (make-instance 'persistent-game :name name))))\n      <\/pre><p>\n        Just one final change before we can compile and deploy our new back end: the \n        \n          games\n        \n         function responsible for returning a list of all games sorted on popularity: \n      <\/p><pre class=\"lispcode\">\n        \n\t    (defun games ()\n\t      (nreverse (get-instances-by-range 'persistent-game 'votes nil nil)))\n      <\/pre><p>\n        \n          votes\n        \n         is an indexed slot, so we can use \n        \n          get-instances-by-range\n        \n         to retrieve a sorted list. The last two arguments are both \n        \n          nil\n        \n        , which will retrieve all stored games. The returned list will be sorted from lowest score to highest, so I apply \n        \n          nreverse\n        \n         to reverse the list (the n in \n        \n          nreverse\n        \n         indicates that it is a destructive function).\n      <\/p><h2>\n        Remembering the Games\n      <\/h2><p>\n        Obviously we want to keep all previously added games. After all, users shouldn't suffer because we decide to change the implementation. So, how do we transform existing games into persistent objects? The simplest way is to map over the \n        \n          *games*\n        \n         list and instantiate a \n        \n          persistent-game\n        \n         with the same slot values as the old games:\n      <\/p><pre class=\"lispcode\">\n        \n\t    RETRO-GAMES&gt;(mapcar\t<b>#'(lambda (old-game)\n\t\t\t\t     (make-instance 'persistent-game \n\t\t\t\t\t\t    :name (name old-game)\n\t\t\t\t\t\t    :votes (votes old-game)))<\/b> \n\t\t\t\t*games*)\n      <\/pre><p>\n        We could have defined a function for this task using \n        \n          defun\n        \n         but, because it is a one-off operation, I go with an anonymous function aka <i>lambda function<\/i> (see the <b>highlighted<\/b> code above). And that's it - all games have been moved into a persistent database. We can now set \n        \n          *games*\n        \n         to \n        \n          NIL\n        \n         (effectively making all old games available for garbage collection) and even make the \n        \n          *games*\n        \n         symbol history by removing it from the package:\n      <\/p><pre class=\"lispcode\">\n        \n\t    RETRO-GAMES&gt; (setf *games* nil)\n\t    <b>NIL<\/b>\n\t    RETRO-GAMES&gt; (unintern '*games*)\n\t    <b>T<\/b>\n      <\/pre><h2>\n        Outro\n      <\/h2><p>\n        This article has really just scratched the surface of what Lisp can do. Yet I hope that if you made it this far, you have seen that behind all those parenthesis there's a lot of power. With its macro system, Lisp can basically be what you want it to.\n      <\/p><p>\n        Due to the dynamic and interactive nature of Lisp it's a perfect fit for prototyping. And because Lisp programs are so easy to evolve, that prototype may end up as a full-blown product one day.\n      <\/p><h2>\n        References\n      <\/h2><p>\n        The full source code for the Retro Games application is available \n        <a href=\"https:\/\/www.adamtornhill.com\/code\/retro-games.lisp\" target=\"_top\">\n          here\n        <\/a>\n        .\n      <\/p><p>\n        Matthew Snyder wrote a sequel. His article is available here: \n        <a href=\"http:\/\/msnyder.info\/posts\/2011\/07\/lisp-for-the-web-part-ii\/\">\n          Lisp for the Web. Part II\n        <\/a>\n      <\/p>","created_at":"2018-11-07T16:22:16+0000","updated_at":"2018-11-07T16:22:16+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":27,"domain_name":"www.adamtornhill.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":62,"title":"Blogging with org-publish in Emacs","url":"https:\/\/news.ycombinator.com\/item?id=17940393","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D17940393\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"17940393\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/vicarie.in\/posts\/blogging-with-org.html\" class=\"storylink\">Blogging with org-publish in Emacs<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=vicarie.in\">vicarie.in<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        81 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=narendraj9\" class=\"hnuser\">narendraj9<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=17940393\">58 days ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=17940393&amp;goto=item%3Fid%3D17940393\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=Blogging%20with%20org-publish%20in%20Emacs&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=Blogging%20with%20org-publish%20in%20Emacs\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=17940393&amp;auth=f08cfbafd001a31cc65dfd141000f3b24912c198\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=17940393\">8\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:22:15+0000","updated_at":"2018-11-07T16:22:15+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":58,"title":"quelpa\/quelpa-use-package","url":"https:\/\/github.com\/quelpa\/quelpa-use-package\/blob\/master\/README.md","content":"<p><a href=\"https:\/\/framagit.org\/steckerhalter\/quelpa\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/e998f827325aea467cb1f137052327f98252c621\/68747470733a2f2f6672616d616769742e6f72672f737465636b657268616c7465722f7175656c70612f7261772f6d61737465722f6c6f676f2f7175656c70612d6c6f676f2d6836342e706e67\" data-canonical-src=\"https:\/\/framagit.org\/steckerhalter\/quelpa\/raw\/master\/logo\/quelpa-logo-h64.png\" alt=\"image\" style=\"float: right;\" \/><\/a><\/p>\n<p><strong>Note:<\/strong> repo is now on <a href=\"https:\/\/framagit.org\/steckerhalter\/quelpa-use-package\" rel=\"nofollow\">framagit.org<\/a>, github is just a mirror<\/p>\n<p>If you are using <a href=\"https:\/\/github.com\/jwiegley\/use-package\">use-package<\/a> (which can help to simplify your .emacs) you can use the <a href=\"https:\/\/framagit.org\/steckerhalter\/quelpa\" rel=\"nofollow\">quelpa<\/a> handler provided by <code>quelpa-use-package<\/code>.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p><strong>Requirements<\/strong>: Emacs 24.3<\/p>\n<p>Assuming you have bootstrapped <code>quelpa<\/code>, install <code>quelpa-use-package<\/code> (which installs <code>use-package<\/code> as a dependency) and require the library:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(quelpa\n '(quelpa-use-package\n   :fetcher git\n   :url \"https:\/\/framagit.org\/steckerhalter\/quelpa-use-package.git\"))\n(require 'quelpa-use-package)<\/pre><\/div>\n<h2><a id=\"user-content-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#usage\"><\/a>Usage<\/h2>\n<p>After that it is possible to call <code>use-package<\/code> with the <code>:quelpa<\/code> keyword:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>;; installs abc-mode with quelpa\n(use-package abc-mode :quelpa)\n;; does the same (`t' is optional)\n(use-package abc-mode :quelpa t)\n;; again... (if the package would have another name)\n(use-package abc-mode :quelpa abc-mode)\n;; passes upgrade parameter to quelpa\n(use-package abc-mode :quelpa (:upgrade t))\n;; uses the given recipe\n(use-package abc-mode\n  :quelpa (abc-mode :fetcher github :repo \"mkjunker\/abc-mode\"))\n;; recipe with plist arguments\n(use-package abc-mode\n  :quelpa ((abc-mode :fetcher github :repo \"mkjunker\/abc-mode\") :upgrade t))<\/pre><\/div>\n<h3><a id=\"user-content-using-quelpa-with-ensure\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-quelpa-with-ensure\"><\/a>Using quelpa with <code>:ensure<\/code><\/h3>\n<p>To make <code>:ensure t<\/code> use <code>quelpa<\/code> instead of <code>package.el<\/code> set the <code>use-package-ensure-function<\/code> in your init file:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(setq use-package-ensure-function 'quelpa)<\/pre><\/div>\n<p>After that:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(use-package abc-mode :ensure t)<\/pre><\/div>\n<p>will install <code>abc-mode<\/code> with <code>quelpa<\/code>.<\/p>\n<p>And if you enable <code>use-package-always-ensure<\/code>:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(setq use-package-always-ensure t)<\/pre><\/div>\n<p>then<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(use-package abc-mode)<\/pre><\/div>\n<p>will install <code>abc-mode<\/code> with <code>quelpa<\/code>.<\/p>\n<h3><a id=\"user-content-conditional-execution\" class=\"anchor\" aria-hidden=\"true\" href=\"#conditional-execution\"><\/a>Conditional execution<\/h3>\n<p>Note that the <code>:quelpa<\/code> keyword is inserted after <code>:if<\/code>, <code>:when<\/code>, <code>:unless<\/code> and <code>:requires<\/code> so that you can make the installation of a package depend on some requirement, for example:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(use-package magit-filenotify\n  :when (fboundp 'file-notify-add-watch)\n  :quelpa (magit-filenotify :fetcher github :repo \"magit\/magit-filenotify\")<\/pre><\/div>\n<p>In this case <code>magit-filenotify<\/code> is only installed if the function <code>file-notify-add-watch<\/code> is bound.<\/p>\n<p>Likewise you can use <code>:requires<\/code> to make the installation depend on a feature being available:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(use-package magit-filenotify\n  :requires filenotify\n  :quelpa (magit-filenotify :fetcher github :repo \"magit\/magit-filenotify\"))<\/pre><\/div>\n<h3><a id=\"user-content-overriding-use-package-always-ensure\" class=\"anchor\" aria-hidden=\"true\" href=\"#overriding-use-package-always-ensure\"><\/a>Overriding <code>use-package-always-ensure<\/code><\/h3>\n<p>To install some packages with quelpa but use <code>use-package-always-ensure<\/code> to install all others from an ELPA repo <code>:ensure<\/code> needs to be disabled if the <code>:quelpa<\/code> keyword is found.<\/p>\n<p><code>quelpa-use-package<\/code> provides an advice for this purpose which can be activated thus:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(quelpa-use-package-activate-advice)<\/pre><\/div>\n<p>To disable it again you can use:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(quelpa-use-package-deactivate-advice)<\/pre><\/div>","created_at":"2018-11-07T16:22:14+0000","updated_at":"2018-11-07T16:22:14+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/6677657?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":59,"title":"Generative Art Finds Its Prodigy","url":"https:\/\/www.artnome.com\/news\/2018\/8\/8\/generative-art-finds-its-prodigy","content":"<div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-3d937e108ff8de7866f0\"><div class=\"sqs-block-content\"><p>In our last article \"<a href=\"https:\/\/www.artnome.com\/news\/2018\/8\/8\/why-love-generative-art\">Why Love Generative Art<\/a>?\" we had a blast putting the genre into the context of modern art history. In this article we interview contemporary generative art prodigy (my words, not his) Manolo Gamboa Naon from Argentina.\u00a0<\/p><p>Manolo's work feels like it is the result of the entire contents of twentieth-century art and design being put into a blender.\u00a0 Once chopped down into its most essential geometry, Manolo then lovingly pieces it back together with algorithms and code to produce art that is simultaneously futuristic and nostalgic.\u00a0His work serves as a welcome (and needed) bridge into digital art and an antidote for those who see the genre as cold, mechanical, and discontinuous with the history of art.<\/p><p>We couldn't be more excited to share our interview with Manolo, his first to be published in English. But before we dive in, let's have some fun and deconstruct a few examples of his work. For me, seeing his work side by side with the masters of twentieth-century art highlights just how well it holds its own.\u00a0<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-3ca9629ee1dc9f2cee28\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;bbccclll&lt;\/em&gt; - &lt;a href=&quot;https:\/\/www.behance.net\/gallery\/66986781\/bbccclll&quot;&gt;Manolo Gamboa Naon&lt;\/a&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/66986781\/bbccclll&quot;&gt;,&amp;nbsp;June 20, 2018&lt;\/a&gt;&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0b\/1534180885985\/15abb866986781.5b29e4ec7e323.png\" alt=\"bbccclll  -  Manolo Gamboa Naon  ,&amp;nbsp;June 20, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"bbccclll  -  Manolo Gamboa Naon  ,&amp;nbsp;June 20, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0b\/1534180885985\/15abb866986781.5b29e4ec7e323.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a0b\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0b\/1534180885985\/15abb866986781.5b29e4ec7e323.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>bbccclll<\/em> - <a href=\"https:\/\/www.behance.net\/gallery\/66986781\/bbccclll\">Manolo Gamboa Naon<\/a><a href=\"https:\/\/www.behance.net\/gallery\/66986781\/bbccclll\">,\u00a0June 20, 2018<\/a><\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-5 span-5\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-52b7e81a318dc68bc787\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;Arrowhead Picture -&lt;\/em&gt; Wassily Kandinsky, 1923&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0d\/1534180885989\/Kandisnky2.jpg\" alt=\"Arrowhead Picture -  Wassily Kandinsky, 1923\" \/><\/noscript><img class=\"thumb-image\" alt=\"Arrowhead Picture -  Wassily Kandinsky, 1923\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0d\/1534180885989\/Kandisnky2.jpg\" data-image-dimensions=\"669x798\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a0d\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0d\/1534180885989\/Kandisnky2.jpg\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>Arrowhead Picture -<\/em> Wassily Kandinsky, 1923<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><div class=\"col sqs-col-7 span-7\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-c410e737571600f6232b\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;Composition 8&lt;\/em&gt; - Wassily Kandinsky, 1923&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0f\/1536417643831\/37.262_ph_web-1.jpg\" alt=\"Composition 8  - Wassily Kandinsky, 1923\" \/><\/noscript><img class=\"thumb-image\" alt=\"Composition 8  - Wassily Kandinsky, 1923\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0f\/1536417643831\/37.262_ph_web-1.jpg\" data-image-dimensions=\"1280x891\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a0f\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a0f\/1536417643831\/37.262_ph_web-1.jpg\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>Composition 8<\/em> - Wassily Kandinsky, 1923<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-9c0da9c0c7f8c9f4afad\"><div class=\"sqs-block-content\"><p>I see Wassily Kandisnky as an obvious artistic influence on Manolo. The two share a masterful use of color and composition and an interest in exploring spiritual and psychological effects of color and geometry. Nowhere is this more apparent than in Manolo's series of works titled <em>bbccclll<\/em>, which have all the rhythm and beauty of Kandinsky's early-1920s lyrical abstractions.\u00a0Kandinsky said of abstract painting that it is \"the most difficult\" of all the arts, noting:<\/p><blockquote><div>It demands that you know how to draw well, that you have a heightened sensitivity for composition and for color, and that you be a true poet. This last is essential.\u00a0<\/div><\/blockquote><p>Manolo's visual poetry checks all of these boxes and does it through code and pixels alone.\u00a0His poetry is most evident in the range of styles and emotions he can elicit from the most basic elements of geometry. For example, let's compare Manolo's Kandinsky-esque <em>bbccclll <\/em>with a work that feels closer to Sonia and Robert Delaunay, Manolo's <em>CUDA.<\/em>\u00a0We can quickly see how Manolo triggers a completely different range of emotions by shifting the color and placement of just two basic elements, the circle and the triangle.\u00a0<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-e24931f606b4aaf16efc\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/66472289\/CUDA&quot;&gt;&lt;em&gt;CUDA&lt;\/em&gt;&amp;nbsp; -&amp;nbsp;Manolo Gamboa Naon&lt;\/a&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/66472289\/CUDA&quot;&gt;,&amp;nbsp;June 5, 2018&lt;\/a&gt;&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a11\/1534180885996\/daf2fc66472289.5b17455677d71.png\" alt=\"CUDA &amp;nbsp; -&amp;nbsp;Manolo Gamboa Naon  ,&amp;nbsp;June 5, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"CUDA &amp;nbsp; -&amp;nbsp;Manolo Gamboa Naon  ,&amp;nbsp;June 5, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a11\/1534180885996\/daf2fc66472289.5b17455677d71.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a11\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a11\/1534180885996\/daf2fc66472289.5b17455677d71.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><a href=\"https:\/\/www.behance.net\/gallery\/66472289\/CUDA\"><em>CUDA<\/em>\u00a0 -\u00a0Manolo Gamboa Naon<\/a><a href=\"https:\/\/www.behance.net\/gallery\/66472289\/CUDA\">,\u00a0June 5, 2018<\/a><\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-6 span-6\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-0f0134a4d4155456f5a3\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;Circular Shapes Sun and Moon &lt;\/em&gt;- Robert Delaunay, 1912,1931&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a13\/1534180886005\/robert-delaunay-circular-shapes-sun-and-moon-1912-31_a-G-9230954-9761621.jpg\" alt=\"Circular Shapes Sun and Moon  - Robert Delaunay, 1912,1931\" \/><\/noscript><img class=\"thumb-image\" alt=\"Circular Shapes Sun and Moon  - Robert Delaunay, 1912,1931\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a13\/1534180886005\/robert-delaunay-circular-shapes-sun-and-moon-1912-31_a-G-9230954-9761621.jpg\" data-image-dimensions=\"473x473\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a13\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a13\/1534180886005\/robert-delaunay-circular-shapes-sun-and-moon-1912-31_a-G-9230954-9761621.jpg\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>Circular Shapes Sun and Moon <\/em>- Robert Delaunay, 1912,1931<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><div class=\"col sqs-col-6 span-6\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-fb9456386c6fb0c3fadb\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;Prismes \u00e9lectriques&lt;\/em&gt; - Sonia Delaunay, 1914&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a15\/1534180885999\/Sonia_Delaunay%2C_1914%2C_Prismes_e%CC%81lectriques%2C_oil_on_canvas%2C_250_x_250_cm%2C_Muse%CC%81e_National_d%27Art_Moderne.jpg\" alt=\"Prismes \u00e9lectriques  - Sonia Delaunay, 1914\" \/><\/noscript><img class=\"thumb-image\" alt=\"Prismes \u00e9lectriques  - Sonia Delaunay, 1914\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a15\/1534180885999\/Sonia_Delaunay%2C_1914%2C_Prismes_e%CC%81lectriques%2C_oil_on_canvas%2C_250_x_250_cm%2C_Muse%CC%81e_National_d%27Art_Moderne.jpg\" data-image-dimensions=\"620x621\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a15\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a15\/1534180885999\/Sonia_Delaunay%2C_1914%2C_Prismes_e%CC%81lectriques%2C_oil_on_canvas%2C_250_x_250_cm%2C_Muse%CC%81e_National_d%27Art_Moderne.jpg\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>Prismes \u00e9lectriques<\/em> - Sonia Delaunay, 1914<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-35816fe6f91b46928a69\"><div class=\"sqs-block-content\"><p>Indeed, Sonia Delaunay sounds as if she is referring to Manolo's sensitivity to color and prolific body of work when she said:\u00a0<\/p><blockquote><div>He who knows how to appreciate color relationships, the influence of one color on another, their contrasts and dissonances, is promised an infinitely diverse imagery.<\/div><\/blockquote><p>Another one of my favorite works by Manolo,\u00a0<em>ppllnnn<\/em>, has a really strong Max Ernst vibe for me. The highly detailed and organic texture of this work reminds me of similar textures that Ernst was able to generate by pioneering techniques like <a href=\"https:\/\/www.tate.org.uk\/art\/art-terms\/f\/frottage\">frottage<\/a> and <a href=\"https:\/\/www.tate.org.uk\/art\/art-terms\/d\/decalcomania\">decalcomania<\/a> to introduce complexity and randomness into his own works.<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-5d754451c1a4b085f6e6\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn&quot;&gt;ppllnnn -&amp;nbsp;&lt;\/a&gt;&lt;\/em&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn&quot;&gt;M&lt;\/a&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn&quot;&gt;anolo Gamboa Naon&lt;\/a&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn&quot;&gt;,&amp;nbsp;&lt;\/a&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn&quot;&gt;July 31, 2018&lt;\/a&gt;&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a17\/1534180886008\/6a1ad768554355.5b610bbfacb1c.png\" alt=\"ppllnnn -&amp;nbsp;   M  anolo Gamboa Naon  ,&amp;nbsp;  July 31, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"ppllnnn -&amp;nbsp;   M  anolo Gamboa Naon  ,&amp;nbsp;  July 31, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a17\/1534180886008\/6a1ad768554355.5b610bbfacb1c.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a17\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a17\/1534180886008\/6a1ad768554355.5b610bbfacb1c.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em><a href=\"https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn\">ppllnnn -\u00a0<\/a><\/em><a href=\"https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn\">M<\/a><a href=\"https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn\">anolo Gamboa Naon<\/a><a href=\"https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn\">,\u00a0<\/a><a href=\"https:\/\/www.behance.net\/gallery\/68554355\/ppllnnn\">July 31, 2018<\/a><\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-6 span-6\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-d5d0f75101530a652c6f\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;100,000 Doves&lt;\/em&gt; - Max Ernst, 1925&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a19\/1536442320951\/13-03-05.jpg\" alt=\"100,000 Doves  - Max Ernst, 1925\" \/><\/noscript><img class=\"thumb-image\" alt=\"100,000 Doves  - Max Ernst, 1925\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a19\/1536442320951\/13-03-05.jpg\" data-image-dimensions=\"728x586\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a19\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a19\/1536442320951\/13-03-05.jpg\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>100,000 Doves<\/em> - Max Ernst, 1925<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><div class=\"col sqs-col-6 span-6\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-c9ce6e3ec67c219b5158\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;The Gramineous Bicycle Garnished with Bells the Dappled Fire Damps and the Echinoderms Bending the Spine to Look for Caresses - &lt;\/em&gt;Max Ernst, 1921&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a1b\/1534180886011\/Ernst_One.jpg\" alt=\"The Gramineous Bicycle Garnished with Bells the Dappled Fire Damps and the Echinoderms Bending the Spine to Look for Caresses -  Max Ernst, 1921\" \/><\/noscript><img class=\"thumb-image\" alt=\"The Gramineous Bicycle Garnished with Bells the Dappled Fire Damps and the Echinoderms Bending the Spine to Look for Caresses -  Max Ernst, 1921\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a1b\/1534180886011\/Ernst_One.jpg\" data-image-dimensions=\"630x472\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a1b\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a1b\/1534180886011\/Ernst_One.jpg\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>The Gramineous Bicycle Garnished with Bells the Dappled Fire Damps and the Echinoderms Bending the Spine to Look for Caresses - <\/em>Max Ernst, 1921<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-441599667d920735563d\"><div class=\"sqs-block-content\"><p>Ernst, always open to surprise and chance, once said:\u00a0<\/p><blockquote><div>Painting is not for me either decorative amusement, or the plastic invention of felt reality; it must be every time: invention, discovery, revelation.<\/div><\/blockquote><p>As you will see in our interview, invention, discovery, and revelation are also at the core of Manolo's art-making process.<\/p><p>Before starting the interview, I'd like to \u00a0thank Artnome's brilliant digital collections analyst, Kaesha Freyaldenhoven,\u00a0who acted as our English-to-Spanish interpreter and later transcribed the interview, translating it into English. None of this would be possible without her enthusiastic assistance, and we very are lucky to have it.<\/p><h2>An Interview with Manolo Gamboa Naon<\/h2><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-57c1de455045da3d2747\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/66418627\/PPCCCC&quot;&gt;PPCCCC - Manolo Gamboa Naon,&amp;nbsp;June 4, 2018&lt;\/a&gt;&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a1d\/1536502761893\/5284ef66418627.5b15abe58ab28.png\" alt=\"PPCCCC - Manolo Gamboa Naon,&amp;nbsp;June 4, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"PPCCCC - Manolo Gamboa Naon,&amp;nbsp;June 4, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a1d\/1536502761893\/5284ef66418627.5b15abe58ab28.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451e70a6adfadf3f9a1d\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a1d\/1536502761893\/5284ef66418627.5b15abe58ab28.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><a href=\"https:\/\/www.behance.net\/gallery\/66418627\/PPCCCC\">PPCCCC - Manolo Gamboa Naon,\u00a0June 4, 2018<\/a><\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-ad4be37dcc0f0c8302dc\"><div class=\"sqs-block-content\"><p><strong>Jason Bailey (JB)<\/strong>:\u00a0A lot of generative artists either start as artists first or programmers first and then build the other skillset. Can you tell us a little about your background? How did you first end up making generative art?<\/p><p><strong>Manolo Gamboa Naon (M)<\/strong>:\u00a0I was young. I was thirteen. But, well, I think in that moment, I started making images but I didn\u2019t know what I was doing. I did not realize that there were other artists out there doing the same things I was doing. Only after many years and finding other artists, did I say, 'Wow! There are people doing things with Flash that I now appreciate in this moment.' I later switched to Processing.<\/p><p><strong>JB<\/strong>:\u00a0How long have you been using Processing?<\/p><p><strong>M<\/strong>: Seven years.<\/p><p><strong>JB<\/strong>:\u00a0How did you learn it?<\/p><p><strong>M<\/strong>:\u00a0I had an <a href=\"https:\/\/www.amazon.com\/Learning-Processing-Beginners-Programming-Interaction\/dp\/0123736021\">orange book<\/a> - <em>Schiffman<\/em> - during this time. But I also started studying design as a career. They encouraged us to learn Processing for a year. We had to learn how to program in the course.\u00a0But during this time, I was more interested in creating interactive things rather than design.<\/p><p><strong>JB<\/strong>:\u00a0I am often surprised how people misunderstand generative art. I had a professor who told me generative art would always be limited, as there is no such thing as an accident with a program. He believed accidents are where discovery happens. I disagreed. When I was making generative art, there were often surprises when I would run the code, and I would build and adjust the work based on those surprises. I am curious about your thoughts on the creative process in coding art. Is it a discovery process with trial and error and accidents and discovery like traditional art making? Or do you have the complete outcome in mind before you start? Or maybe neither?<\/p><p><strong>M<\/strong>:\u00a0Generally, one has an idea, makes a first draft of this idea, and then begins correcting. All of the time, when I come and create,\u00a0the most beautiful parts of the work are born from the errors. After a certain point, I believe that the maturity of my style was formed by making small errors because I was discovering as I went along.\u00a0From these errors, I take an idea and it stays. I learn how to manipulate from these errors.\u00a0The error is central to the work of generative artists apart from, obviously, the rules, and the rules become text that converts into an image.\u00a0It is impossible to have what you imagine become what you see.\u00a0The beginning is errors, errors, errors, errors. They are beautiful errors.<\/p><h3>Manolo works in series across themes. Below we look at <em>pppp <\/em>as a theme with several works that Manolo developed over time:<\/h3><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-a8f448c87107d4f49baa\"><div class=\"sqs-block-content\"><p><strong>JB<\/strong>:For me, your sense of color and pattern are what stands out the most. I feel like your choice of color palettes is very smart, and you evoke strong feelings through this alone. For example, you have some recent works that makes me nostalgic for the '80s, with shapes and colors that were very popular in that decade. Where do you get the inspiration for your color palette?<\/p><\/div><\/div><div class=\"row sqs-row\"><div class=\"col sqs-col-6 span-6\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-536038ab0fb158de96b9\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;NNN&lt;\/em&gt; -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a68\/1534180886045\/80s.png\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a68\/1534180886045\/80s.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a68\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a68\/1534180886045\/80s.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>NNN<\/em> -\u00a0Manolo Gamboa Naon,\u00a0March 5, 2018<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-a96a0e2dd3287cf3656b\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;NNN&lt;\/em&gt; -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6b\/1536503348976\/c237d362848723.5a9dd8b36c611.png\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6b\/1536503348976\/c237d362848723.5a9dd8b36c611.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a6b\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6b\/1536503348976\/c237d362848723.5a9dd8b36c611.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>NNN<\/em> -\u00a0Manolo Gamboa Naon,\u00a0March 5, 2018<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><div class=\"col sqs-col-6 span-6\"><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-9d917ef1cc855af5a52c\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;NNN&lt;\/em&gt; -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6d\/1536503127549\/26d33362848723.5a9dd8b36dab5.png\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6d\/1536503127549\/26d33362848723.5a9dd8b36dab5.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a6d\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6d\/1536503127549\/26d33362848723.5a9dd8b36dab5.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>NNN<\/em> -\u00a0Manolo Gamboa Naon,\u00a0March 5, 2018<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-6b487291c1e01bb1f62d\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;NNN&lt;\/em&gt; -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6f\/1536503306043\/5dd7fd62848723.5a9dd8b36c343.png\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"NNN  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;March 5, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6f\/1536503306043\/5dd7fd62848723.5a9dd8b36c343.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a6f\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a6f\/1536503306043\/5dd7fd62848723.5a9dd8b36c343.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>NNN<\/em> -\u00a0Manolo Gamboa Naon,\u00a0March 5, 2018<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-f60cda26c264545a9a3c\"><div class=\"sqs-block-content\"><p><strong>M<\/strong>:\u00a0Color is a problem in my life. Realistically, when I began making generative artwork, I realized that programmers - I mean, I don\u2019t want to generalize - but they do not give color a lot of importance. They do not have an intention. But in the past five or six years, I have been attempting to feel more comfortable using colors. Because it matters -\u00a0a lot. And now, sometimes I spend more time forming a color palette than programming.\u00a0My inspiration comes from looking around all the time. I look at a lot of things from design. Instagram, Twitter, all the time searching for references. A movie, an old newspaper. Inspiration comes from many places. In my work, I intend to evoke something -\u00a0from a time or of a certain quality.<\/p><p><strong>JB<\/strong>:\u00a0What are the biggest changes you have seen in the last decade in Processing and generative art?<\/p><p><strong>M<\/strong>:\u00a0During the Flash epoch, people were generating compelling visual art. Since then there have been changes, but I am not sure if the changes have been progressive.\u00a0I want to learn English so that I can start getting to know the world. I mean, there are many people who are interested in generative art on Twitter.\u00a0Sometimes on the forums I do not know what people are saying, but I do understand the code.<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-5e9cf67157a5113abd65\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;em&gt;MMGGK2&lt;\/em&gt; -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;August 23, 2018&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a71\/1536445161377\/\" alt=\"MMGGK2  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;August 23, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"MMGGK2  -&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;August 23, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a71\/1536445161377\/\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a71\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a71\/1536445161377\/\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><em>MMGGK2<\/em> -\u00a0Manolo Gamboa Naon,\u00a0August 23, 2018<\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-28db63aa809064dda84d\"><div class=\"sqs-block-content\"><p><strong>JB<\/strong>:\u00a0What are the new trends? How do you think things like AI, machine learning, and deep learning will impact the future of generative art?<\/p><p><strong>M<\/strong>:\u00a0I think there are many artists working in deep learning and artificial intelligence right now. There is a lot of great progress.\u00a0I prefer to work with the geometry, not so much on the side of AI or machine learning. It does not interest me that much.\u00a0Maybe one day I will start getting into it. But from my perspective, they are two separate paths. Maybe I\u2019m just old.\u00a0I am very impressed by the work. But I do not know if this is something that I have a desire to pursue.<\/p><p><strong>JB<\/strong>:\u00a0I know many generative artists,\u00a0particularly in the Processing community,\u00a0make the code for their projects open source for others to learn from. Is this something you do as well? Why or why not?<\/p><p><strong>M<\/strong>:\u00a0A few weeks ago I published all of the code for everything I have created since 2007. Absolutely there are some that don\u2019t work, but everything I have created on Processing is online.\u00a0For me, the code of others were huge learning tools, and it seems like it is a good thing to share these sorts of things.\u00a0A few people have asked me for my code and it would always take me a long time to respond individually, but in this form, I can pass along the information. Although I speak Spanish, the good thing about code is that I can read it in any language. Code transcends language, and to me, that is beautiful.<\/p><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-0300a35b4e5b2e84b660\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/62551541\/LLLLPP&quot;&gt;&lt;em&gt;LLLLPP&lt;\/em&gt;&amp;nbsp;-&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;February 26, 2018&lt;\/a&gt;&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a73\/1536508715264\/1f69d962551541.5a9446366690a.png\" alt=\"LLLLPP &amp;nbsp;-&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;February 26, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"LLLLPP &amp;nbsp;-&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;February 26, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a73\/1536508715264\/1f69d962551541.5a9446366690a.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a73\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a73\/1536508715264\/1f69d962551541.5a9446366690a.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><a href=\"https:\/\/www.behance.net\/gallery\/62551541\/LLLLPP\"><em>LLLLPP<\/em>\u00a0-\u00a0Manolo Gamboa Naon,\u00a0February 26, 2018<\/a><\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-e0d4fff7290a143578ca\"><div class=\"sqs-block-content\"><p><strong>JB<\/strong>:\u00a0How have you shown your work in the past? Do you only show it online because it is digital? Or have you shown it in galleries as well?<\/p><p><strong>M<\/strong>: Principally,\u00a0I share all of my work online. I really like the idea of people sharing online. My works have never been for printing. Really, I prefer to post them online and share the digital images in that way. I consume art through the internet, and I prefer that it stays there. That the works are viewed online is very important to me. I like the movement and allowing the work itself to live. Although there have been some artists in Buenos Aires who have printed their works and shared them in that way, this is not the route that I would like to take. I prefer to keep my works digital.<\/p><p><strong>JB<\/strong>:\u00a0Where do you get your inspiration?<\/p><p><strong>M<\/strong>:\u00a0Only after learning about the Argentinian scene of generative art did I become familiar with artists from the world. Artists who inspire me include:<\/p><ul><li><a href=\"http:\/\/ravenkwok.com\/\">Raven Kwok<\/a><\/li><li><a href=\"http:\/\/benfry.com\/\">Ben Fry<\/a><\/li><li><a href=\"http:\/\/reas.com\/\">Casey Reas<\/a><\/li><li><a href=\"https:\/\/joshuadavis.com\/\">Joshua Davis<\/a><\/li><li><a href=\"http:\/\/feltron.com\/index.html\">Nicholas Felton<\/a><\/li><li><a href=\"https:\/\/twitter.com\/paredros\">Ivan Ivanoff<\/a><\/li><\/ul><p><strong>JB<\/strong>:\u00a0Is there anything else you would like to share with the Artnome audience?<\/p><p><strong>M<\/strong>:\u00a0I feel that images are dying, they are disposable. Their lifetimes are very short, considering the quantity of production of images. But I also think that there is a moment of contemporary art on the rise.\u00a0Art is becoming a more central part of the culture.\u00a0It has always been for music, but now, the visual arts - design, for example - are occupying more important positions in the lives of the people. I love what is happening.\u00a0<\/p><p>There is also the destruction of the idea of an artistic genius. Well, sure there are people that are doing really good things, but it is not like it was before. Many different and talented artists are gaining recognition for their work.\u00a0I think it is important to see the destruction of the artistic genius to understand that there are many people who are capable of doing what they did.\u00a0<\/p><p>Communication between artists is also very important, with the help of the internet. Through the internet, I have been able to meet people who understand and appreciate what I am doing.<\/p><h2>Conclusion<\/h2><\/div><\/div><div class=\"sqs-block image-block sqs-block-image\" data-block-type=\"5\" id=\"block-d36b971aabb49b7ccf4b\"><div class=\"sqs-block-content\">\n  \n  \t\n      <div class=\"image-block-outer-wrapper layout-caption-below design-layout-inline combination-animation-none individual-animation-none individual-text-animation-none\">\n      \n        <div class=\"intrinsic\">\n          \n            <div class=\"image-block-wrapper   has-aspect-ratio\" data-description=\"&lt;p&gt;&lt;a href=&quot;https:\/\/www.behance.net\/gallery\/68418633\/gggggggg&quot;&gt;&lt;em&gt;gggggggg&lt;\/em&gt;&amp;nbsp;-&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;July 28, 2018&lt;\/a&gt;&lt;\/p&gt;\">\n              <noscript><img src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a75\/1536508967252\/27590f68418633.5b5c22de5b89b.png\" alt=\"gggggggg &amp;nbsp;-&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;July 28, 2018\" \/><\/noscript><img class=\"thumb-image\" alt=\"gggggggg &amp;nbsp;-&amp;nbsp;Manolo Gamboa Naon,&amp;nbsp;July 28, 2018\" data-image=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a75\/1536508967252\/27590f68418633.5b5c22de5b89b.png\" data-image-dimensions=\"1920x1920\" data-image-focal-point=\"0.5,0.5\" data-load=\"false\" data-image-id=\"5b95451f70a6adfadf3f9a75\" data-type=\"image\" src=\"https:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451f70a6adfadf3f9a75\/1536508967252\/27590f68418633.5b5c22de5b89b.png\" \/><\/div>\n          \n          \n          <div class=\"image-caption-wrapper\">\n            <div class=\"image-caption\"><p><a href=\"https:\/\/www.behance.net\/gallery\/68418633\/gggggggg\"><em>gggggggg<\/em>\u00a0-\u00a0Manolo Gamboa Naon,\u00a0July 28, 2018<\/a><\/p><\/div>\n          <\/div>\n          \n        <\/div>\n      \n      <\/div>\n    \n  \n<\/div><\/div><div class=\"sqs-block html-block sqs-block-html\" data-block-type=\"2\" id=\"block-ea1b186ecfe4ca8bba69\"><div class=\"sqs-block-content\"><p>One of the great joys of having a platform like Artnome to share my thoughts on the intersection of art and tech is the opportunity to introduce people to deserving artists whom they may not know yet. Manolo may not be a fan of the idea of the artist as \"genius,\"\u00a0but he certainly fits all of my personal criteria for the distinction. His sensitivity to color and composition, the speed at which he explores new concepts, and the volume of compelling work he produces puts him into a class of his own. I feel lucky that Manolo agreed to let us interview him for Artnome so we can all have a little more insight into his remarkable work. I'd encourage you all to follow him on Twitter <a href=\"https:\/\/twitter.com\/manoloidee\">@manoloide<\/a> and to check out the rest of his amazing work on <a href=\"https:\/\/www.behance.net\/manoloide\">Behance<\/a>.\u00a0<\/p><p>I'd like to again thank Kaesha Freyoldenhoven for her amazing job as interpreter and translator for this interview. As always, if you have questions or ideas for a post, feel free to reach out to jason@artnome.com. If you have not already, I recommend you sign up for our newsletter to keep up with news and new articles.<\/p><\/div><\/div><div class=\"sqs-block newsletter-block sqs-block-newsletter\" data-block-type=\"51\" id=\"block-yui_3_17_2_1_1536523819302_239813\"><div class=\"sqs-block-content\">\n<div class=\"newsletter-form-wrapper newsletter-form-wrapper--layoutFloat newsletter-form-wrapper--alignCenter\">\n  <form class=\"newsletter-form\" data-form-id=\"5b958263352f531880dc9386\" autocomplete=\"on\" method=\"post\" action=\"action\">\n    <header class=\"newsletter-form-header\">\n      \n    <\/header>\n    \n      <div class=\"newsletter-form-footnote\"><p>We respect your privacy.<\/p><\/div>\n    \n    <div class=\"hidden form-submission-text\">Thank you!<\/div>\n    \n  <\/form>\n<\/div>\n<\/div><\/div>","created_at":"2018-11-07T16:22:14+0000","updated_at":"2018-11-07T16:22:14+0000","published_at":"2018-09-09T00:00:00+0000","published_by":["Jason Bailey"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":11,"domain_name":"www.artnome.com","preview_picture":"http:\/\/static1.squarespace.com\/static\/59413d96e6f2e1c6837c7ecd\/t\/5b95451e70a6adfadf3f9a09\/1536503671665\/15fe1e62172035.5a87280d713e4.png?format=1000w","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":60,"title":"Pretty fontification of source code blocks","url":"https:\/\/orgmode.org\/worg\/org-contrib\/babel\/examples\/fontify-src-code-blocks.html","content":"<p>&#13;\nDefine these faces, in your <code>.emacs<\/code> file, <i>before<\/i> requiring Org (if you&#13;\ndon't have your own color theme):&#13;\n<\/p><div class=\"org-src-container\"><pre class=\"src src-emacs-lisp\">(defface org-block-begin-line&#13;\n  '((t (:underline \"#A7A6AA\" :foreground \"#008ED1\" :background \"#EAEAFF\")))&#13;\n  \"Face used for the line delimiting the begin of source blocks.\")&#13;\n&#13;\n(defface org-block-background&#13;\n  '((t (:background \"#FFFFEA\")))&#13;\n  \"Face used for the source block background.\")&#13;\n&#13;\n(defface org-block-end-line&#13;\n  '((t (:overline \"#A7A6AA\" :foreground \"#008ED1\" :background \"#EAEAFF\")))&#13;\n  \"Face used for the line delimiting the end of source blocks.\")&#13;\n<\/pre><\/div>","created_at":"2018-11-07T16:22:14+0000","updated_at":"2018-11-07T16:22:14+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":0,"domain_name":"orgmode.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":61,"title":"rejeep\/prodigy.el","url":"https:\/\/github.com\/rejeep\/prodigy.el","content":"<p>Manage external services from within Emacs<\/p><p>I came up with the idea when I got to work one Monday morning and\nbefore I could start working I had to manually start ten or so\nservices.<\/p><p>To get rid of this tedious work, I started working on this Emacs\nplugin, which provides a nice and simple GUI to manage services.<\/p><p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/rejeep\/prodigy.el\/blob\/master\/prodigy.png\"><img src=\"https:\/\/github.com\/rejeep\/prodigy.el\/raw\/master\/prodigy.png\" alt=\"Prodigy\" \/><\/a><\/p><h2>Installation<\/h2><p>Add <code>prodigy<\/code> to your <a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a> file:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(depends-on \"prodigy\")<\/pre><\/div><p>If you are on a Mac, disable nap mode for Emacs, otherwise requests\nwill be very slow when Emacs enters nap mode:<\/p><div class=\"highlight highlight-source-shell\"><pre>$ defaults write org.gnu.Emacs NSAppSleepDisabled -bool YES<\/pre><\/div><h2>Usage<\/h2><p>Start Prodigy with <code>M-x prodigy<\/code>. You should see a list of all defined\nservices.<\/p><h3>Services<\/h3><p>Services can be defined in a few different ways. See doc-string for\ninformation about available properties to specify: <code>M-x describe-variable RET prodigy-services<\/code>.<\/p><p>Properties that accepts a function as argument all get a property list\nas argument, for example:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service\n  :command (lambda (&amp;rest args)\n             (let ((service (plist-get args :service)))\n               ;; ...\n               )))<\/pre><\/div><p>You can also use the <code>prodigy-callback<\/code> macro to simplify the argument\nhandling.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service\n  :command (prodigy-callback (service)\n             ;; ...\n             ))<\/pre><\/div><p>Depending on property, the <code>args<\/code> list contain various properties.<\/p><h4>prodigy-define-service (<code>&amp;rest args<\/code>)<\/h4><p>Services can be defined using the function <code>prodigy-define-service<\/code>:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service :prop value ...)<\/pre><\/div><h4>prodigy-services<\/h4><p>Services can be defined by setting the variable <code>prodigy-services<\/code>:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(setq prodigy-services\n '((:prop value ...)\n   (:prop value ...)))<\/pre><\/div><h3>Viewing process output<\/h3><p>In the prodigy window, you can see a process' output with the <code>$<\/code> key.\nProcess output buffers use the <code>prodigy-view-mode<\/code> and do some special\npre-processing to the process output. The buffer output will be tailed\n(\u00e0 la <code>tail -f<\/code>) if the point is at the buffer end. There are two\nsignificant variables that influence process output:<\/p><ul><li><code>prodigy-output-filters<\/code> is a list of filters to apply to the\noutput (currently, the process will be ansi-colorized and <code>^M<\/code>\nliterals will be stripped). Filter functions should take a single\nargument, the <code>output<\/code> string, and should return a string.<\/li>\n<li><code>prodigy-process-on-output-hook<\/code> is a hook that runs on\nprocess output. Each function in the hook takes two arguments,\n<code>service<\/code> (the service data structure) and <code>output<\/code> (the service's\noutput).<\/li>\n<\/ul><h3>Tags<\/h3><p>Services can have any number of tags. Tags do not have to be\npredefined. If they are, the service will inherit all the tags\nproperties. Tags can also have tags. A service will inherit all tags\nrecursively.<\/p><p>See doc-string for information about available properties to specify:\n<code>M-x describe-variable RET prodigy-tags<\/code>.<\/p><h4>prodigy-define-tag (<code>&amp;rest args<\/code>)<\/h4><p>Tags can be defined using the function <code>prodigy-define-tag<\/code>:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-tag :prop value ...)<\/pre><\/div><h4>prodigy-tags<\/h4><p>Tags can be defined by setting the variable <code>prodigy-tags<\/code>:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(setq prodigy-tags\n '((:prop value ...)\n   (:prop value ...)))<\/pre><\/div><h3>Filters<\/h3><p>Filters is a way to show only specific services in the Prodigy\nbuffer. For example services with specific tag or with a name matching\na string.<\/p><h4>prodigy-add-filter (<code>&amp;rest args<\/code>)<\/h4><p>To add a filter, use <code>prodigy-add-filter<\/code>:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-add-filter :tag 'foo)\n(prodigy-add-filter :name \"bar\")<\/pre><\/div><h4>prodigy-filters<\/h4><p>You can also set the variable <code>prodigy-filters<\/code> directly:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(setq prodigy-filters\n      '((:tag foo)\n        (:name \"bar\")))<\/pre><\/div><h3>Status<\/h3><p>Each service is associated with a status. The built in statuses are:<\/p><ul><li><code>stopped<\/code> (default) - The process is not running.<\/li>\n<li><code>running<\/code> - The process is running. If the process status is <code>run<\/code>,\nthis status will be used.<\/li>\n<li><code>ready<\/code> - The process is \"actually\" ready. This will be set when the\nservice outputs a message that matches its <code>ready-message<\/code> property,\nor it can be set manually.<\/li>\n<li><code>stopping<\/code> - Set when a service is stopping.<\/li>\n<li><code>failed<\/code> - The process failed. A service has this status if:\n<ul><li>It is not started within <code>prodigy-start-tryouts<\/code> seconds.<\/li>\n<li>Or, it is not stopped within <code>prodigy-stop-tryouts<\/code> seconds.<\/li>\n<li>Or, if the process exit code is non zero.<\/li>\n<\/ul><\/li>\n<\/ul><p>The only way Prodigy has an idea of the service status, is to look at\nthe process status (note the difference between service and process\nstatus). The process status is however not always a very good\nindication of the service \"actual\" status. For example, it takes about\nfive seconds to start a Rails server, but the process status will be\n<code>run<\/code> almost instantly after started.<\/p><p>To improve the service status, there is a function called\n<code>prodigy-set-status<\/code>, that can change the status of a service. The\nfunction takes two arguments: The <code>service<\/code> and the <code>status-id<\/code>. The\nstatus id has to be one of the statuses in <code>prodigy-status-list<\/code>.<\/p><p>You can create your own status with the function\n<code>prodigy-define-status<\/code>. See doc-string for information about\navailable properties to specify: <code>M-x describe-variable RET prodigy-status-list<\/code>.<\/p><p>For more information, see status example below!<\/p><h2>Examples<\/h2><h3>Python Simple HTTP Server<\/h3><p>This service start a Python Simple HTTP Server on port <code>6001<\/code>. When\nstopping the service, the <code>sigkill<\/code> signal is used.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service\n  :name \"Python app\"\n  :command \"python\"\n  :args '(\"-m\" \"SimpleHTTPServer\" \"6001\")\n  :cwd \"\/path\/to\/my\/project\"\n  :tags '(work)\n  :stop-signal 'sigkill\n  :kill-process-buffer-on-stop t)<\/pre><\/div><h3>Nodemon Server<\/h3><p>This service starts a Nodemon server on port <code>6002<\/code>. The project is\nusing NVM (Node Version Manager), so before the process starts, NVM is\nset up.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service\n  :name \"Node app\"\n  :command \"nodemon\"\n  :cwd \"\/path\/to\/my\/project\"\n  :args '(\"app.coffee\")\n  :port 6002\n  :tags '(work node)\n  :init-async (lambda (done)\n                (nvm-use-for \"\/path\/to\/my\/project\" done)))<\/pre><\/div><h3>Sinatra Server<\/h3><p>This service starts a Sinatra server on port <code>6003<\/code>. The project is\nusing RVM (Ruby Version Manager), so before the process starts, RVM is\nset up.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service\n  :name \"Sinatra\"\n  :command \"server\"\n  :cwd \"\/path\/to\/my\/project\"\n  :path '(\"\/path\/to\/my\/project\/bin\")\n  :port 6003\n  :tags '(work ruby)\n  :init-async (lambda (done)\n                (rvm-activate-ruby-for \"\/path\/to\/my\/project\" done)))<\/pre><\/div><h3>Foreman<\/h3><p>Prodigy also works with <a href=\"http:\/\/theforeman.org\/\" rel=\"nofollow\">Foreman<\/a>.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-service\n  :name \"Foreman\"\n  :command \"foreman\"\n  :args '(\"start\")\n  :cwd \"\/path\/to\/my\/project\")<\/pre><\/div><h3>Tag inheritance<\/h3><p>If a service has a tag and that tag is defined (see\n<code>prodigy-define-tag<\/code>), the service inherits the tag properties. The\ninheritance is recursive, so if any of the service tags has tags\nitself, the service will inherit those tag properties as well.<\/p><p>This is best illustrated with an example. Rails can run with many\ndifferent servers. Each server indicate that it's ready with a\ndifferent log message. In the example code below, a tag is defined for\neach server and one for Rails that inherits all those servers.<\/p><p>That means that a service that is tagged with <code>rails<\/code>, will be set to\nready if it uses any of the three servers. But since there is a tag\nfor each server, a non Rails service that uses any of the servers can\nsimply use that tag.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-tag\n  :name 'thin\n  :ready-message \"Listening on 0\\\\.0\\\\.0\\\\.0:[0-9]+, CTRL\\\\+C to stop\")\n(prodigy-define-tag\n  :name 'webrick\n  :ready-message \"WEBrick::HTTPServer#start: pid=[0-9]+ port=[0-9]+\")\n(prodigy-define-tag\n  :name 'mongrel\n  :ready-message \"Ctrl-C to shutdown server\")\n(prodigy-define-tag\n  :name 'rails\n  :tags '(thin mongrel webrick))\n(prodigy-define-service\n  :name \"Rails Project\"\n  :command \"bundle\"\n  :args '(\"exec\" \"rails\" \"server\")\n  :cwd \"\/path\/to\/my\/project\"\n  :tags '(rails))\n(prodigy-define-service\n  :name \"Thin Project\"\n  :command \"bundle\"\n  :args '(\"exec\" \"thin\" \"start\")\n  :cwd \"\/path\/to\/my\/project\"\n  :tags '(thin))<\/pre><\/div><h3>Fine Tuning Status<\/h3><p>Prodigy can only look at the <em>process<\/em> status to determine the\n<em>service<\/em> status. To make status even more useful, you can set status\nmanually. Prodigy provides the function <code>prodigy-set-status<\/code> for\nthis. In this example, we create a tag <code>rails<\/code> that will set the\nstatus to <code>ready<\/code> when the server is actually ready.<\/p><p>The services that are tagged with <code>rails<\/code> will all inherit this.<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-tag\n  :name 'rails\n  :on-output (lambda (&amp;rest args)\n               (let ((output (plist-get args :output))\n                     (service (plist-get args :service)))\n                 (when (or (s-matches? \"Listening on 0\\.0\\.0\\.0:[0-9]+, CTRL\\\\+C to stop\" output)\n                           (s-matches? \"Ctrl-C to shutdown server\" output))\n                   (prodigy-set-status service 'ready)))))\n(prodigy-define-service\n  :name \"Rails\"\n  :command \"bundle\"\n  :args '(\"exec\" \"rails\" \"server\")\n  :cwd \"\/path\/to\/my\/project\"\n  :tags '(rvm rails))\n(prodigy-define-service\n  :name \"Sinatra\"\n  :command \"bundle\"\n  :args '(\"exec\" \"rackup\")\n  :cwd \"\/path\/to\/my\/project\"\n  :tags '(rvm rails))<\/pre><\/div><h2>Troubleshoot<\/h2><h3>Jekyll<\/h3><p>For some unknown reason, Jekyll fail with this error:<\/p><pre>error: invalid byte sequence in US-ASCII. Use --trace to view backtrace\n<\/pre><p>This can be solved by adding a <code>jekyll<\/code> tag, like this:<\/p><div class=\"highlight highlight-source-lisp\"><pre>(prodigy-define-tag\n  :name 'jekyll\n  :env '((\"LANG\" \"en_US.UTF-8\")\n         (\"LC_ALL\" \"en_US.UTF-8\")))<\/pre><\/div><p>Then tag your services with the <code>jekyll<\/code> tag.<\/p><h2>Changelog<\/h2><p>This is a short summary of changes between versions.<\/p><h3>v0.7.0<\/h3><ul><li>Tail process buffer if the point is at the end<\/li>\n<li>Misc bug fixes<\/li>\n<\/ul><h3>v0.6.0<\/h3><ul><li>Adding new function <code>prodigy-callback<\/code>, allowing for some syntactic\nsugar using property lambdas (that take plist as argument).<\/li>\n<li>For some properties that accepts a lambda as argument, the lambda\narguments is now a plist. <strong>Breaking change!<\/strong><\/li>\n<\/ul><h3>v0.5.0<\/h3><ul><li>New property <code>:ready-message<\/code> used to set the process status to\nready when matching output appears.<\/li>\n<li>Clear process buffer contents using <code>k<\/code><\/li>\n<\/ul><h3>v0.4.0<\/h3><ul><li>Do not require current working directory (see <code>:cwd<\/code> property).<\/li>\n<li>Improved test suite.<\/li>\n<li>Process output buffer has major mode <code>prodigy-view-mode<\/code>.<\/li>\n<li>Font lock automatically enabled in view mode.<\/li>\n<li>Truncate process output buffer (see variables\n<code>prodigy-view-buffer-maximum-size<\/code>,\n<code>prodigy-view-truncate-by-default<\/code> and <code>:truncate-output<\/code> property).<\/li>\n<li>Fix bug when restarting a service cleared buffer contents.<\/li>\n<li>Jump between services with status with <code>M-n<\/code> and <code>M-p<\/code>.<\/li>\n<\/ul><h3>v0.3.0<\/h3><ul><li>Tag inheritance (see <code>prodigy-define-tag<\/code>).<\/li>\n<li>On process output hook (see <code>:on-output<\/code> property).<\/li>\n<li>Improved status handling (see <code>prodigy-set-status<\/code>).<\/li>\n<li>Using <code>hl-line<\/code>.<\/li>\n<li>Updated status colors.<\/li>\n<li>Url property can be a list of urls (see <code>:url<\/code> property).<\/li>\n<li>Async restarts.<\/li>\n<li>Force kill process (using prefix argument to <code>prodigy-stop<\/code>).<\/li>\n<li>Service path can be string, list or lambda (see <code>:path<\/code> property).<\/li>\n<li>Service command can be string or lambda (see <code>:command<\/code> property).<\/li>\n<li>Service args can be list or lambda (see <code>:args<\/code> property).<\/li>\n<li>Hidden tags (see <code>:hide<\/code> property).<\/li>\n<\/ul><h2>Contribution<\/h2><p>Contribution is much welcome!<\/p><p>If your contribution is a bug fix, create your topic branch from\n<code>master<\/code>. If it is a new feature, check if there exists a WIP branch\n(<code>vMAJOR.MINOR-wip<\/code>). If it does, use that branch, otherwise use <code>master<\/code>.<\/p><p>Install <a href=\"https:\/\/github.com\/cask\/cask\">Cask<\/a> if you haven't\nalready, then:<\/p><pre>$ cd \/path\/to\/prodigy.el\n$ cask\n<\/pre><p>Run all tests with:<\/p><pre>$ make\n<\/pre>","created_at":"2018-11-07T16:22:14+0000","updated_at":"2019-02-07T04:44:29+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/20277?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":57,"title":"Seek A New Sin, by Osiris Saline","url":"https:\/\/osirissaline.bandcamp.com\/album\/seek-a-new-sin","content":"<div><h3 class=\"hd\">\n        \n            \n        \n        \n        Streaming + Download\n    <\/h3>\n    \n    \n        <p>Includes unlimited streaming via the free Bandcamp app, plus high-quality download in MP3, FLAC and more.<\/p>\n        \n\t\t\n  \n  \n        \n    \n    \n        <p>\n            Purchasable with gift card<\/p>\n    \n    <p><\/p><h4 class=\"ft compound-button main-button\">\n            \n                \u00a0\n            \n            \n            \n                \n                $11.99\n                            AUD\n                            \u00a0or more\n                \n            \n        <\/h4>\n    \n    \n    \n    \n    <h4 class=\"ft compound-button send-as-gift\">\n        \u00a0\n    <\/h4><\/div><div><h3 class=\"hd\">\n        \n            <a class=\"band-link\" data-bind=\"click: fullDiscographyBuyDialog\">\n              Full Digital Discography\n            <\/a>\n        \n        \n        \n    <\/h3>\n    \n    <div class=\"art compact\"><a class=\"more-button\" data-bind=\"click: expandCoverArt\">\n                <p>15<\/p>\n                                <p>releases<\/p>\n            <\/a><\/div>\n    <div class=\"bd\"><p>Get <strong>all 15 Osiris Saline releases<\/strong> available on Bandcamp and save <strong>20%<\/strong>.<\/p><p>Includes unlimited streaming via the free Bandcamp app, plus high-quality downloads of\n        \n            \n        \n        \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/tombs-of-love\">Tombs Of Love<\/a>,\n            \n            \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/out-of-sight\">Out Of Sight<\/a>,\n            \n            \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/tend-to-it\">Tend To It<\/a>,\n            \n            \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/nails\">Nails<\/a>,\n            \n            \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/lozenges-for-the-soul-instrumentals\">Lozenges For The Soul: Instrumentals<\/a>,\n            \n            \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/the-viable-alternative\">The Viable Alternative<\/a>,\n            \n            \n        \n            \n                <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/parahelion-ost\">Parahelion OST<\/a>,\n            \n            \n        \n            \n            \n                \n                    <a href=\"https:\/\/osirissaline.bandcamp.com\/album\/memories-from-a-dead-star\">Memories From A Dead Star<\/a>, \n                    \n                    \n                        and <a data-bind=\"click: moreTitles\">7 more<\/a>.\n                    \n                    \n                \n            \n        \n        \n            \n                , and , .\n            \n        \n        \n    <\/p><\/div>\n    \n        <p>\n            Purchasable with gift card<\/p>\n    \n    \n        <h4 class=\"ft compound-button main-button\">\n            \u00a0\n            \n            \n            $48.71 \n                    AUD\n                    or more\n            \n            \n            \n            (<strong>20% OFF<\/strong>)\n            \n        <\/h4>\n    \n    \n    \n    <h4 class=\"ft compound-button send-as-gift\">\n        \u00a0\n    <\/h4><\/div>","created_at":"2018-11-07T16:22:13+0000","updated_at":"2018-11-07T16:22:13+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"osirissaline.bandcamp.com","preview_picture":"https:\/\/f4.bcbits.com\/img\/a2365674209_5.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":55,"title":"Automating the Deployment of a Scalable WordPress Site","url":"https:\/\/www.digitalocean.com\/community\/tutorials\/automating-the-deployment-of-a-scalable-wordpress-site","content":"<p>In this guide, we will create and deploy a scalable WordPress instance consisting of a MySQL database server, a GlusterFS distributed filesystem, Nginx web servers, and an Nginx load balancer.  By using <a href=\"https:\/\/www.digitalocean.com\/company\/blog\/automating-application-deployments-with-user-data\/\">user-data<\/a> and <a href=\"https:\/\/www.digitalocean.com\/community\/tutorials\/an-introduction-to-droplet-metadata\">Droplet meta-data<\/a>, we will automate the deployment of our site.  Finally, we will provide a Ruby script which will automate this entire process and ease the creation of scalable Wordpress sites.  Through this tutorial, you will learn about the power and flexibility of user-data and Droplet meta-data in deploying services on DigitalOcean.<\/p><h2 id=\"step-one-&#x2014;-planning-our-deployment\">Step One \u2014 Planning our Deployment<\/h2><p>The deployment we create in this tutorial will consist of a single MySQL database server, multiple GlusterFS servers in a cluster, multiple Nginx web servers, and a single Nginx load balancer.<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/automate_wp_cluster\/wp-cluster.png\" alt=\"WordPress Deployment\" \/><\/p><p>Before we begin we should know:<\/p><ul><li>What size Droplet we will use for our MySQL server<\/li>\n<li>How many GlusterFS nodes we will create<\/li>\n<li>What size our GlusterFS nodes will be<\/li>\n<li>How many web server nodes we will need<\/li>\n<li>What size Droplets we will use for our web servers<\/li>\n<li>What size Droplet we will use for our load balancer<\/li>\n<li>The domain name we will use for our new site<\/li>\n<\/ul><p>We can add additional nodes or scale up the nodes we created if we need to later.  Once we have decided on these details, we can begin deploying our site.<\/p><h2 id=\"step-two-&#x2014;-deploying-mysql\">Step Two \u2014 Deploying MySQL<\/h2><p>We will begin by deploying our MySQL server.  To do this, we will create a default Ubuntu 14.04 x64 Droplet using the following user-data.<\/p><pre>#!\/bin\/bash\nexport DEBIAN_FRONTEND=noninteractive;\nexport PUBLIC_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/public\/0\/ipv4\/address)\nexport PRIVATE_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/private\/0\/ipv4\/address)\napt-get update;\napt-get -y install mysql-server;\nmysqladmin -u root create wordpress;\nmysqladmin -u root password \"mysql_password\";\nsed -i.bak \"s\/127.0.0.1\/$PRIVATE_IP\/g\" \/etc\/mysql\/my.cnf;\nservice mysql restart;\nmysql -uroot -pmysql_password -e \"CREATE USER 'wordpress'@'%' IDENTIFIED BY 'mysql_password'\";\nmysql -uroot -pmysql_password -e \"GRANT ALL PRIVILEGES ON wordpress.* TO 'wordpress'@'%'\";\n<\/pre><p>This user-data script will perform the following functions on our new Droplet:<\/p><p>First, we export a variable which tells <code>apt-get<\/code> that we are running in non-interactive mode to prevent it from prompting for any input when it installs packages.<\/p><pre>export DEBIAN_FRONTEND=noninteractive;\n<\/pre><p>Next, we use Droplet meta-data to get the Droplet's public and private IP addresses and assign them to variables:<\/p><pre>export PUBLIC_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/public\/0\/ipv4\/address)\nexport PRIVATE_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/private\/0\/ipv4\/address)\n<\/pre><pre>Note: Droplet Meta-Data is not available in NYC1, NYC2, and AMS1 at this time.\n<\/pre><p>We then use <code>apt<\/code> to install the MySQL server.<\/p><pre>apt-get update;\napt-get -y install mysql-server;\n<\/pre><p>Now we need to create a new database called <strong>wordpress<\/strong>.<\/p><pre>mysqladmin -u root create wordpress;\n<\/pre><p>Then we set a password for our MySQL root user.<\/p><pre>mysqladmin -u root password \"mysql_password\";\n<\/pre><p>Because our MySQL server will be accepting queries from our web servers, we need to have it listen on the private IP address rather than only on <code>localhost<\/code>. To do this, we will use <code>sed<\/code> to update the MySQL configuration file by doing a find and replace and then restart the service.<\/p><pre>sed -i.bak \"s\/127.0.0.1\/$PRIVATE_IP\/g\" \/etc\/mysql\/my.cnf;\nservice mysql restart;\n<\/pre><p>Finally, we will create a new MySQL user called <strong>wordpress<\/strong> and give it permission to access the <em>wordpress<\/em> database.<\/p><pre>mysql -uroot -pmysql_password -e \"CREATE USER 'wordpress'@'%' IDENTIFIED BY 'mysql_password'\";\nmysql -uroot -pmysql_password -e \"GRANT ALL PRIVILEGES ON wordpress.* TO 'wordpress'@'%'\";\n<\/pre><p>By deploying our new Droplet with this user-data script, we will have a configured MySQL server listening on its private IP address and with our configured database and user without ever logging in via SSH or the console.<\/p><h2 id=\"step-three-&#x2014;-deploying-glusterfs\">Step Three \u2014 Deploying GlusterFS<\/h2><p>Before deploying our GlusterFS cluster, we need to decide how many nodes we will deploy.  There are two variables that will go into this decision.  First, we need to decide how much space we need and then we need to decide on a replica setting to use.  The replica setting tells GlusterFS how many copies of any file to store.  For example, a replica setting of 2 will mean that every file is duplicated on at least 2 servers.  This will cut our available storage in half since we are keeping two copies of each file but will provide improved redundancy.  The number of GlusterFS nodes we create must be a multiple of our replica setting.  For a cluster with a replica setting of 2, we will need to create our nodes in a multiple of 2 (so 2, 4, 6, or 8 nodes would be acceptable).<\/p><p>For this example, we will deploy a 4 node GlusterFS cluster using a replica setting of 2.<\/p><p>For our first 3 nodes, we will use the following user-data script:<\/p><pre>#!\/bin\/bash\nexport DEBIAN_FRONTEND=noninteractive;\napt-get update;\napt-get install -y python-software-properties;\nadd-apt-repository -y ppa:gluster\/glusterfs-3.5;\napt-get update;\napt-get install -y glusterfs-server;\n<\/pre><p>Again, we first set the <code>DEBIAN_FRONTEND<\/code> variable so <code>apt<\/code> knows that we are running in non-interactive mode:<\/p><pre>export DEBIAN_FRONTEND=noninteractive;\n<\/pre><p>We then update our <code>apt<\/code> database and install <code>python-software-properties<\/code>, which is needed to add the <a href=\"https:\/\/help.launchpad.net\/Packaging\/PPA\">PPA<\/a> for GlusterFS.<\/p><pre>apt-get update;\napt-get install -y python-software-properties;\n<\/pre><p>Next we will add the GlusterFS PPA so we can grab our deb packages.<\/p><pre>add-apt-repository -y ppa:gluster\/glusterfs-3.5;\n<\/pre><p>Then we will update our <code>apt<\/code> database again and install glusterfs-server.<\/p><pre>apt-get install -y glusterfs-server;\n<\/pre><p>For our first three nodes, this is all we need to do.  Make a note of the private IP addresses assigned to each of these new Droplets as we will need them when creating our final GlusterFS node and creating our volume.<\/p><p>For our final node, we will use the following user-data script:<\/p><pre>#!\/bin\/bash\nexport DEBIAN_FRONTEND=noninteractive;\nexport PRIVATE_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/private\/0\/ipv4\/address)\napt-get update;\napt-get install -y python-software-properties;\nadd-apt-repository -y ppa:gluster\/glusterfs-3.5;\napt-get update;\napt-get install -y glusterfs-server;\nsleep 30;\ngluster peer probe node1_private_ip;\ngluster peer probe node2_private_ip;\ngluster peer probe node3_private_ip;\ngluster volume create file_store replica 2 transport tcp node1_private_ip:\/gluster node2_private_ip:\/gluster node3_private_ip:\/gluster $PRIVATE_IP:\/gluster force;\ngluster volume start file_store;\n<\/pre><pre>Note: If you do not want to enable replication you should not include the \"replica\" setting in your \"volume create\" command.\n<\/pre><p>The first section of this user-data script is pretty similar to the one we used on the other GlusterFS nodes, though we are assigning our new Droplet's private IP to the $PRIVATE_IP variable.  Once <code>glusterfs-server<\/code> is installed, though, we do some additional work. <\/p><p>First, our script will wait 30 seconds for the new glusterfs-server to start up and be available.<\/p><pre>sleep 30\n<\/pre><p>Then we probe the three GlusterFS Droplets we created earlier in order to add all four to a cluster.<\/p><pre>gluster peer probe node1_private_ip;\ngluster peer probe node2_private_ip;\ngluster peer probe node3_private_ip;\n<\/pre><p>Next we will create our GlusterFS volume named \"file<em>store\" with a replica setting of 2 and including all four of our nodes.  Since we wont know the IP address of our newest node yet we will use the $PRIVATE<\/em>IP variable for it.<\/p><pre>gluster volume create file_store replica 2 transport tcp node1_private_ip:\/gluster node2_private_ip:\/gluster node3_private_ip:\/gluster $PRIVATE_IP:\/gluster force;\n<\/pre><p>Finally, we will start the new volume to make it accessible from our clients:<\/p><pre>gluster volume start file_store;\n<\/pre><p>We now have a distributed filesystem where we can keep our WordPress files that will be accessible to all our web server nodes.<\/p><h2 id=\"step-four-&#x2014;-deploying-nginx-web-servers\">Step Four \u2014 Deploying Nginx Web Servers<\/h2><p>Now that we have a database server and a distributed filesystem all set, we can deploy our web servers.   We will use the following user-data script to deploy our first Nginx web server node and configure our WordPress installation inside our GlusterFS volume.<\/p><pre>#!\/bin\/bash\nexport DEBIAN_FRONTEND=noninteractive;\nexport PRIVATE_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/private\/0\/ipv4\/address)\napt-get update;\napt-get -y install nginx glusterfs-client php5-fpm php5-mysql;\nsed -i s\/\\;cgi\\.fix_pathinfo\\=1\/cgi\\.fix_pathinfo\\=0\/g \/etc\/php5\/fpm\/php.ini;\nmkdir \/gluster;\nmount -t glusterfs gluter_node_private_ip:\/file_store \/gluster;\necho \"gluster_node_private_ip:\/file_store \/gluster glusterfs defaults,_netdev 0 0\" &gt;&gt; \/etc\/fstab;\nmkdir \/gluster\/www;\nwget https:\/\/raw.githubusercontent.com\/ryanpq\/do-wpc\/master\/default -O \/etc\/nginx\/sites-enabled\/default;\nservice nginx restart;\n# Get Wordpress Files\nwget https:\/\/wordpress.org\/latest.tar.gz -O \/root\/wp.tar.gz;\ntar -zxf \/root\/wp.tar.gz -C \/root\/;\ncp -Rf \/root\/wordpress\/* \/gluster\/www\/.;\ncp \/gluster\/www\/wp-config-sample.php \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_NAME', 'database_name_here'\/'DB_NAME', 'wordpress'\/g\" \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_USER', 'username_here'\/'DB_USER', 'wordpress'\/g\" \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_PASSWORD', 'password_here'\/'DB_PASSWORD', 'mysql_password'\/g\" \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_HOST', 'localhost'\/'DB_HOST', 'mysql_private_ip'\/g\" \/gluster\/www\/wp-config.php;\nchown -Rf www-data:www-data \/gluster\/www;\n<\/pre><p>This script is a bit more complicated than our previous ones, so let's break it down step by step.<\/p><p>First, we will again set the <code>DEBIAN_FRONTEND<\/code> variable as we have in our previous scripts and populate our <code>$PRIVATE_IP<\/code> variable.<\/p><pre>export DEBIAN_FRONTEND=noninteractive;\nexport PRIVATE_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/private\/0\/ipv4\/address)\n<\/pre><p>Next, we will update our <code>apt<\/code> database and install Nginx, the glusterfs client, and the php libraries we will need.<\/p><pre>apt-get update;\napt-get -y install nginx glusterfs-client php5-fpm php5-mysql;\n<\/pre><p>Then we will use <code>sed<\/code>'s find and replace functionality to update our <code>php.ini<\/code> file and set the cgi.fixpathinfo variable to 0.<\/p><pre>sed -i s\/\\;cgi\\.fix_pathinfo\\=1\/cgi\\.fix_pathinfo\\=0\/g \/etc\/php5\/fpm\/php.ini;\n<\/pre><p>Now we'll create a folder called <code>\/gluster<\/code> in the root of our disk image and mount our GlusterFS volume there.  Then we will create an fstab entry so our GlusterFS volume is automatically mounted when the Droplet boots.<\/p><pre>mkdir \/gluster;\nmount -t glusterfs gluter_node_private_ip:\/file_store \/gluster;\necho \"gluster_node_private_ip:\/file_store \/gluster glusterfs defaults,_netdev 0 0\" &gt;&gt; \/etc\/fstab;\n<\/pre><p>Then we will create a folder called <code>www<\/code> in our GlusterFS volume.  This folder will act as our web root.<\/p><pre>mkdir \/gluster\/www;\n<\/pre><p>Next we will pull a new Nginx configuration file from a remote server.  This file will set our web root to <code>\/gluster\/www<\/code> and ensure Nginx is configured to use PHP.  You can view this configuration file <a href=\"https:\/\/raw.githubusercontent.com\/ryanpq\/do-wpc\/master\/default\">here<\/a>. Once we have replaced our Nginx configuration file we will restart the service for this change to take effect.<\/p><pre>wget https:\/\/raw.githubusercontent.com\/ryanpq\/do-wpc\/master\/default -O \/etc\/nginx\/sites-enabled\/default;\nservice nginx restart;\n<\/pre><p>Now we will grab a copy of the latest version of WordPress, extract it and copy its contents to our new web root.<\/p><pre>wget https:\/\/wordpress.org\/latest.tar.gz -O \/root\/wp.tar.gz;\ntar -zxf \/root\/wp.tar.gz -C \/root\/;\ncp -Rf \/root\/wordpress\/* \/gluster\/www\/.;\n<\/pre><p>Next, we will copy the sample WordPress configuration file to <code>wp-config.php<\/code>.<\/p><pre>cp \/gluster\/www\/wp-config-sample.php \/gluster\/www\/wp-config.php;\n<\/pre><p>And update its variables to match our new environment, again using <code>sed<\/code>'s find and replace function.<\/p><pre>sed -i \"s\/'DB_NAME', 'database_name_here'\/'DB_NAME', 'wordpress'\/g\" \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_USER', 'username_here'\/'DB_USER', 'wordpress'\/g\" \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_PASSWORD', 'password_here'\/'DB_PASSWORD', 'mysql_password'\/g\" \/gluster\/www\/wp-config.php;\nsed -i \"s\/'DB_HOST', 'localhost'\/'DB_HOST', 'mysql_private_ip'\/g\" \/gluster\/www\/wp-config.php;\n<\/pre><p>And finally, we will make sure that the files in our web root are owned by the user <strong>www-data<\/strong> which our Nginx process will be running as.<\/p><pre>chown -Rf www-data:www-data \/gluster\/www;\n<\/pre><p>We now have our first web server node all set up and ready to receive requests.<\/p><p>Since each of our web server nodes shares the same GlusterFS volume for storage, there are fewer steps for each additional node we create.  For additional nodes we will use the following user-data script:<\/p><pre>#!\/bin\/bash\nexport DEBIAN_FRONTEND=noninteractive;\nexport PRIVATE_IP=$(curl -s http:\/\/169.254.169.254\/metadata\/v1\/interfaces\/private\/0\/ipv4\/address)\napt-get update;\napt-get -y install nginx glusterfs-client php5-fpm php5-mysql;\nsed -i s\/\\;cgi\\.fix_pathinfo\\=1\/cgi\\.fix_pathinfo\\=0\/g \/etc\/php5\/fpm\/php.ini;\nmkdir \/gluster;\nmount -t glusterfs gluster_node_private_ip:\/file_store \/gluster;\necho \"gluster_node_private_ip:\/file_store \/gluster glusterfs defaults,_netdev 0 0\" &gt;&gt; \/etc\/fstab;\nmkdir \/gluster\/www;\nwget https:\/\/raw.githubusercontent.com\/ryanpq\/do-wpc\/master\/default -O \/etc\/nginx\/sites-enabled\/default;\nservice nginx restart;\n<\/pre><p>For our additional web nodes we will still be installing the same packages, mounting our GlusterFS volume and replacing our Nginx configuration file but we will not need to do any setup of our WordPress instance since we did this when creating our first node.<\/p><p>The final step in this deployment is to create our load balancer. We will be using another Nginx server for this purpose. To set up this node we will use the following user-data script:<\/p><pre>#!\/bin\/bash\nexport DEBIAN_FRONTEND=noninteractive;\napt-get update;\napt-get -y install nginx;\nlbconf=\"\nserver {\n    listen 80 default_server;\n    listen [::]:80 default_server ipv6only=on;\n    root \/usr\/share\/nginx\/html;\n    index index.php index.html index.htm;\n    location \/ {\n      proxy_pass  http:\/\/backend;\n      include proxy_params;\n    }\n}\nupstream backend  {\n    ip_hash;\n    server web_node_1_private_ip\n    server web_node_2_private_ip\n}\n\"\necho $lbconf &gt; \/etc\/nginx\/sites-enabled\/default;\nservice nginx restart;\n<\/pre><p>For our load balancer's user-data script, we will be building our Nginx configuration directly in the script.  We start off much as we did with our other Droplets by ensuring that <code>apt<\/code> knows we are running in non-interactive mode.<\/p><pre>export DEBIAN_FRONTEND=noninteractive;\n<\/pre><p>Then we will install Nginx:<\/p><pre>apt-get update;\napt-get -y install nginx;\n<\/pre><p>Next we will create our new Nginx configuration in a variable called <code>lbconf<\/code>. Adding an entry for each of our web servers in the <em>upstream backend<\/em> section.<\/p><pre>lbconf=\"\nserver {\n    listen 80 default_server;\n    listen [::]:80 default_server ipv6only=on;\n    root \/usr\/share\/nginx\/html;\n    index index.php index.html index.htm;\n    location \/ {\n      proxy_pass  http:\/\/backend;\n      include proxy_params;\n    }\n}\nupstream backend  {\n    ip_hash;\n    server web_node_1_private_ip\n    server web_node_2_private_ip\n}\n\"\n<\/pre><p>We will then write the <code>lbconf<\/code> variable to our Nginx configuration file, replacing its current content.<\/p><pre>echo $lbconf &gt; \/etc\/nginx\/sites-enabled\/default;\n<\/pre><p>And finally, we'll restart Nginx for this configuration to take effect.<\/p><pre>service nginx restart;\n<\/pre><h2 id=\"step-six-&#x2014;-setting-up-dns\">Step Six \u2014 Setting Up DNS<\/h2><p>Now, before we access our new WordPress site via the browser, we should set up a DNS entry for it.  We will do this via the control panel.<\/p><p>In the DigitalOcean control panel, click on <strong>DNS<\/strong>. In the <strong>Add Domain<\/strong> form, enter your domain name and select your load balancer Droplet from the drop-down menu, then click <strong>Create Domain<\/strong>.<\/p><p>In order to use the <strong>www<\/strong> subdomain for your site you will need to create another record in this new domain.<\/p><p>Click <strong>Add Record<\/strong> and choose the <strong>CNAME<\/strong> record type. In the name field enter <strong>www<\/strong> and in the hostname field enter <strong>@<\/strong>.  This will direct requests for the www subdomain to the same location as your main domain (your load balancer Droplet).<\/p><h2 id=\"step-seven-&#x2014;-configuring-wordpress\">Step Seven \u2014 Configuring WordPress<\/h2><p>Now that we have launched all our Droplets and configured our domain, we can access our new WordPress site by visiting our newly configured domain in a web browser.<\/p><p><img src=\"https:\/\/assets.digitalocean.com\/articles\/automate_wp_cluster\/wp-setup.png\" alt=\"Configure Wordpress\" \/><\/p><p>We will be prompted here to create a user account and give our new site a name.  Once we have done this our deployment is complete and we can begin using the new site.<\/p><h2 id=\"step-eight-&#x2014;-automating-the-process\">Step Eight \u2014 Automating the Process<\/h2><p>Now that we can create our WordPress deployment without ever needing to ssh into a Droplet, we can take things a step further and automate this process using the <a href=\"https:\/\/developers.digitalocean.com\/documentation\/v2\/\">DigitalOcean API<\/a>.  <\/p><p>A sample Ruby script has been created based on this tutorial which will prompt the user to provide the relevant details and then automatically deploy a new scablable WordPress instance. You can find this script <a href=\"https:\/\/github.com\/ryanpq\/do-wpc\">on GitHub<\/a>.<\/p><h2>Conclusion<\/h2><p>We now have a scalable WordPress deployment but there are additional steps we can take to ensure our new site is secure and stable.<\/p>","created_at":"2018-11-07T16:22:12+0000","updated_at":"2018-11-07T16:22:12+0000","published_at":"2015-04-13T18:10:45+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":14,"domain_name":"www.digitalocean.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":56,"title":"WordPress at Scale","url":"https:\/\/www.scalewp.io\/","content":"<div class=\"markdown-body markdown-content\">\n<div id=\"wpas\">\n<h2 id=\"gfm-1\">Can WordPress Scale?<\/h2>\n<h3 id=\"gfm-2\">Yes. But there are requirements:<\/h3>\n<p>A single WordPress can push hundreds of millions of pageviews a month; it can serve tens of thousands of concurrent logged-in users; and it can be lightning-fast the whole time. It is known. The question is not whether WordPress itself can scale, but whether or not your implementation is ready.<\/p>\n<p>This site contains best practices from teams with real-world experience in delivering WordPress at scale. They provide a solid foundation for planning a scalable implementation, both for development teams and site owners alike.<\/p>\n<div class=\"bulletbox bb-left\">\n<h3>Developers<\/h3>\n<p>We all stand on the shoulders of giants. Learn trusted architectural patterns for scalable website infrastructure, as well as the optimizations that nearly all big sites leverage.\n<\/p><\/div>\n<div class=\"bulletbox bb-right\">\n<h3>Site Owners<\/h3>\n<p>When and how to invest in scalability is one of the most important business decisions you'll make. Use this site to have productive discussions with your developers and vendors.\n<\/p><\/div>\n\n<p>This website will arm you with the knowledge you need to realize your vision for a highly scalable WordPress implementation. Get started with the links below.<\/p>\n<p><\/p>\n<div class=\"article-links\"><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/elastic-architecture\/\">Elastic Architecture | Horizontal Scalability Is the Scalability That Matters<\/a><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/page-caching\/\">Page Caching | The Key to Internet-Scale Traffic is Reverse-Proxy Caching<\/a><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/object-caching\/\">Object Caching | Persistent Object Caching Speeds Up Dynamic Pageviews<\/a><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/query-performance\/\">Query Performance | The Database Is the Ultimate Bottleneck<\/a><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/searching-for-scale\/\">Searching for Scale | Improve User Experience and Database Performance With a Search Index<\/a><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/a-real-world-scalable-architecture\/\">A Real-World Scalable Architecture | Scalable Architecture is More Than Just LAMP<\/a><a class=\"article-link\" href=\"https:\/\/www.scalewp.io\/development-and-workflow\/\">Development and Workflow | Measure Twice. Cut Once. Revert as Needed<\/a><\/div>\n<div id=\"social-links\">\n<div class=\"prefix\">\n\t\t\tShare\t\t<\/div>\n<div class=\"icon twitter\">\n\t\t\t<a href=\"https:\/\/twitter.com\/intent\/tweet?url=https%3A%2F%2Fwww.scalewp.io%2F&amp;text=WordPress+at+Scale%3A+How+to+run+a+WordPress+site+at+scale\" target=\"_blank\" title=\"Share on Twitter\"><p><\/p><\/a>\n\t\t<\/div>\n<div class=\"icon facebook\">\n\t\t\t<a href=\"https:\/\/www.facebook.com\/sharer\/sharer.php?u=https%3A%2F%2Fwww.scalewp.io%2F\" target=\"_blank\" title=\"Share on Facebook\"><p><\/p><\/a>\n\t\t<\/div>\n<div class=\"prefix\">\n\t\t\tContribute\t\t<\/div>\n<div class=\"icon github\">\n\t\t\t<a href=\"https:\/\/github.com\/pantheon-systems\/wordpress-at-scale\/edit\/master\/_pages\/scale.md\" target=\"_blank\" title=\"Fork on GitHub\"><p><\/p><\/a>\n\t\t<\/div>\n<\/div>\n<\/div>\n                        <\/div>","created_at":"2018-11-07T16:22:12+0000","updated_at":"2018-11-07T16:22:12+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":1,"domain_name":"www.scalewp.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":53,"title":"How to Use JSON Data Fields in MySQL Databases","url":"https:\/\/www.sitepoint.com\/use-json-data-fields-mysql-databases\/","content":"<p><img src=\"https:\/\/dab1nmslvvntp.cloudfront.net\/wp-content\/uploads\/2016\/04\/1461769463mysql.svg\" alt=\"JSON in MySQL\" width=\"40%\" class=\"alignright size-full wp-image-129277\" \/>In my article <a href=\"http:\/\/www.sitepoint.com\/sql-vs-nosql-differences\/\">SQL vs NoSQL: The Differences<\/a>, I mentioned the line between SQL and NoSQL databases has become increasingly blurred with each camp adopting features from the other. MySQL 5.7 InnoDB and PostgreSQL 9.4 databases both directly support JSON document types in a single field. In this article, we\u2019ll examine MySQL\u2019s JSON implementation in more detail.<\/p><p><em>(PostgreSQL supported JSON before version 9.4 and any database will accept JSON documents as a single string blob. However, MySQL and PostgreSQL now directly support validated JSON data in real key\/value pairs rather than a basic string.)<\/em><\/p><h2>Just Because You Can Store JSON \u2026<\/h2><p><em>\u2026 it doesn\u2019t follow you should.<\/em><\/p><div class=\"widget maestro maestro-content-type-html hide-for-tablet-SP hide-for-mobile-SP\" id=\"maestro-636\"><p>--ADVERTISEMENT--<\/p><\/div><p>Normalization is a technique used to optimize the database structure. The First Normal Form (1NF) rule governs that every column should hold a single value \u2014 which is broken by storing multi-value JSON documents.<\/p><p>If you have clear relational data requirements, use appropriate single-value fields. JSON should be used sparingly as a last resort. JSON value fields cannot be indexed, so avoid using it on columns which are updated or searched regularly. In addition, fewer client applications support JSON, and the technology is newer and possibly less stable than other types.<\/p><p>That said, there are good JSON use-cases for sparsely-populated data or custom attributes.<\/p><h2>Create a Table with a JSON Field<\/h2><p>Consider a shop selling books. A <strong>book<\/strong> table will have an ID, <abbr title=\"International Standard Book Number\">ISBN<\/abbr>, title, publisher, number of pages and other relational data which applies to all books. Presume we want to add any number of category tags to any book. We could achieve this in SQL using:<\/p><ol><li>a <strong>tag<\/strong> table which stored each tag name against a unique ID, and<\/li>\n<li>a <strong>tagmap<\/strong> table with many-to-many records mapping book IDs to tag IDs<\/li>\n<\/ol><p>It\u2019ll work, but it\u2019s cumbersome and considerable effort for a minor feature. Therefore, we\u2019ll define a <strong>tags<\/strong> JSON field in our MySQL database\u2019s <strong>book<\/strong> table:<\/p><pre class=\"language-sql\">CREATE TABLE `book` (&#13;\n  `id` mediumint(8) unsigned NOT NULL AUTO_INCREMENT,&#13;\n  `title` varchar(200) NOT NULL,&#13;\n  `tags` json DEFAULT NULL,&#13;\n  PRIMARY KEY (`id`)&#13;\n) ENGINE=InnoDB;<\/pre><p>Note that JSON columns cannot have a default value, be used as a primary key, be referenced as a foreign key or have an index. You can create <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/create-table.html#create-table-secondary-indexes-virtual-columns\">secondary indexes on generated virtual columns<\/a>, but it\u2019s possibly easier to retain an indexed value in a separate field.<\/p><h2>Adding JSON Data<\/h2><p>Whole JSON documents can be passed in <strong>INSERT<\/strong> or <strong>UPDATE<\/strong> statements. For example, our book tags can be passed as an array:<\/p><pre class=\"language-sql\">INSERT INTO `book` (`title`, `tags`)&#13;\nVALUES (&#13;\n  'ECMAScript 2015: A SitePoint Anthology',&#13;\n  '[\"JavaScript\", \"ES2015\", \"JSON\"]'&#13;\n);<\/pre><p>JSON can also be created with the:<\/p><ul><li><a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-creation-functions.html#function_json-array\">JSON_ARRAY() function<\/a> which creates arrays, e.g.\n<pre class=\"language-sql\">-- returns [1, 2, \"abc\"]:&#13;\nSELECT JSON_ARRAY(1, 2, 'abc');<\/pre>\n<\/li>\n<li><a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-creation-functions.html#function_json-object\">JSON_OBJECT() function<\/a> which creates objects, e.g.\n<pre class=\"language-sql\">-- returns {\"a\": 1, \"b\": 2}:&#13;\nSELECT JSON_OBJECT('a', 1, 'b', 2);<\/pre>\n<\/li>\n<li><a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-modification-functions.html#function_json-merge\">JSON_MERGE() function<\/a> to merge documents, e.g.\n<pre class=\"language-sql\">-- returns [\"a\", 1, {\"key\": \"value\"}]:&#13;\nSELECT JSON_MERGE('[\"a\", 1]', '{\"key\": \"value\"}');<\/pre>\n<\/li>\n<li>or you can <code>(CAST anyValue AS JSON)<\/code>.<\/li>\n<\/ul><p>The <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-attribute-functions.html#function_json-type\">JSON_TYPE() function<\/a> allows you to check JSON value types. It should return OBJECT, ARRAY or an error, e.g.<\/p><pre class=\"language-sql\">-- returns ARRAY:&#13;\nSELECT JSON_TYPE('[1, 2, \"abc\"]');&#13;\n&#13;\n-- returns OBJECT:&#13;\nSELECT JSON_TYPE('{\"a\": 1, \"b\": 2}');&#13;\n&#13;\n-- returns an error:&#13;\nSELECT JSON_TYPE('{\"a\": 1, \"b\": 2');<\/pre><p>Similarly, the <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-attribute-functions.html#function_json-valid\">JSON_VALID() function<\/a> returns 1 when the JSON is valid:<\/p><pre class=\"language-sql\">-- returns 1:&#13;\nSELECT JSON_TYPE('[1, 2, \"abc\"]');&#13;\n&#13;\n-- returns 1:&#13;\nSELECT JSON_TYPE('{\"a\": 1, \"b\": 2}');&#13;\n&#13;\n-- returns 0:&#13;\nSELECT JSON_TYPE('{\"a\": 1, \"b\": 2');<\/pre><p>Attempting to insert an invalid JSON document will raise an error, and the whole record will not be inserted\/updated.<\/p><h2>Searching JSON Data<\/h2><p>The <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-search-functions.html#function_json-contains\">JSON_CONTAINS() function<\/a> accepts the JSON document being searched and another to compare against. It returns 1 when a match is found, e.g.<\/p><pre class=\"language-sql\">-- all books with the 'JavaScript' tag:&#13;\nSELECT * FROM `book` &#13;\nWHERE JSON_CONTAINS(tags, '[\"JavaScript\"]');<\/pre><p>The similar <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-search-functions.html#function_json-search\">JSON_SEARCH() function<\/a> returns the path to the given match or NULL when there\u2019s no match. It is passed the JSON document being searched, <code>'one'<\/code> to find the first match or <code>'all'<\/code> to find all matches, and a search string, e.g.<\/p><pre class=\"language-sql\">-- all books with tags starting 'Java':&#13;\nSELECT * FROM `book` &#13;\nWHERE JSON_SEARCH(tags, 'one', 'Java%') IS NOT NULL;<\/pre><h2>JSON Paths<\/h2><p>A JSON path targets values and can be used to extract or modify parts of a document. The <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-search-functions.html#function_json-extract\">JSON_EXTRACT() function<\/a> demonstrates this by extracting one or more values:<\/p><pre class=\"language-sql\">-- returns \"SitePoint\":&#13;\nSELECT JSON_EXTRACT(&#13;\n  '{\"id\": 1, \"website\": \"SitePoint\"}', &#13;\n  '$.website'&#13;\n);<\/pre><p>All JSON path definitions start with a <code>$<\/code> followed by other selectors:<\/p><ul><li>a period followed by a name, e.g. <code>$.website<\/code><\/li>\n<li><code>[N]<\/code> where N is the position in a zero-indexed array<\/li>\n<li>a <code>.[*]<\/code> wildcard which evaluates all members of an object<\/li>\n<li>a <code>[*]<\/code> wildcard which evaluates all members of an array<\/li>\n<li>a <code>prefix**suffix<\/code> wildcard which evaluates to all paths that begin with the named prefix and end with the named suffix.<\/li>\n<\/ul><p>Referring to this JSON document:<\/p><pre class=\"language-sql\">{&#13;\n  \"a\": 1,&#13;\n  \"b\": 2,&#13;\n  \"c\": [3, 4],&#13;\n  \"d\": {&#13;\n    \"e\": 5,&#13;\n    \"f\": 6&#13;\n  }&#13;\n}<\/pre><p>the following paths:<\/p><ul><li><code>$.a<\/code> returns <code>1<\/code><\/li>\n<li><code>$.c<\/code> returns <code>[3, 4]<\/code><\/li>\n<li><code>$.c[1]<\/code> returns <code>4<\/code><\/li>\n<li><code>$.d.e<\/code> returns <code>5<\/code><\/li>\n<li><code>$**.e<\/code> returns <code>[5]<\/code><\/li>\n<\/ul><h2>Extracting JSON Paths in Queries<\/h2><p>We can extract the first tag from our <strong>book<\/strong> table using the query:<\/p><pre class=\"language-sql\">SELECT&#13;\n  name,&#13;\n  tags-&gt;\"$[0]\" AS `tag1`&#13;\nFROM `book`;<\/pre><p>For a more complex example, presume we have a <strong>user<\/strong> table with JSON profile data, e.g.<\/p><table><tr><th>id<\/th>\n<th>name<\/th>\n<th>profile<\/th>\n<\/tr><tr><td>1<\/td>\n<td>Craig<\/td>\n<td>\n{<br \/>\u201ctwitter\u201d: \u201c@craigbuckler\u201d,<br \/>\u201cfacebook\u201d: \u201ccraigbuckler\u201d,<br \/>\u201cgoogleplus\u201d: \u201ccraigbuckler\u201d<br \/>}\n<\/td>\n<\/tr><tr><td>2<\/td>\n<td>SitePoint<\/td>\n<td>\n{<br \/>\u201ctwitter\u201d: \u201c@sitepointdotcom\u201d<br \/>}\n<\/td>\n<\/tr><\/table><p>We can extract the Twitter name using a JSON path, e.g.<\/p><pre class=\"language-sql\">SELECT&#13;\n  name, profile-&gt;\"$.twitter\" AS `twitter`&#13;\nFROM `user`;<\/pre><p>We can also reference a JSON path in the WHERE clause to only return users with a Twitter account:<\/p><pre class=\"language-sql\">SELECT&#13;\n  name, profile-&gt;\"$.twitter\" AS `twitter`&#13;\nFROM `user`&#13;\nWHERE&#13;\n  profile-&gt;\"$.twitter\" IS NOT NULL;<\/pre><h2>Modifying Part of a JSON Document<\/h2><p>There are several <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-modification-functions.html\">MySQL functions to modify parts of a JSON document<\/a> using path notation. These include:<\/p><ul><li><code>JSON_SET(doc, path, val[, path, val]...)<\/code> \u2014<br \/>inserts or updates data in the document<\/li>\n<li><code>JSON_INSERT(doc, path, val[, path, val]...)<\/code> \u2014<br \/>inserts data into the document<\/li>\n<li><code>JSON_REPLACE(doc, path, val[, path, val]...)<\/code> \u2014<br \/>replaces data in the document<\/li>\n<li><code>JSON_MERGE(doc, doc[, doc]...)<\/code> \u2014<br \/>merges two or more documents<\/li>\n<li><code>JSON_ARRAY_APPEND(doc, path, val[, path, val]...)<\/code> \u2014<br \/>appends values to the end of an array<\/li>\n<li><code>JSON_ARRAY_INSERT(doc, path, val[, path, val]...)<\/code> \u2014<br \/>inserts an array within the document<\/li>\n<li><code>JSON_REMOVE(doc, path[, path]...)<\/code> \u2014<br \/>removes data from the document.<\/li>\n<\/ul><p>We can therefore add a \u2018programming\u2019 tag to any book which already has a \u2018JavaScript\u2019 tag:<\/p><pre class=\"language-sql\">UPDATE `book`&#13;\n  SET tags = JSON_MERGE(tags, '[\"technical\"]')&#13;\nWHERE&#13;\n  JSON_SEARCH(tags, 'one', 'technical') IS NULL AND&#13;\n  JSON_SEARCH(tags, 'one', 'JavaScript') IS NOT NULL;<\/pre><h2>Further Information<\/h2><p>The MySQL manual provides further information about the <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json.html\">JSON data type<\/a> and the <a href=\"https:\/\/dev.mysql.com\/doc\/refman\/en\/json-functions.html\">associated JSON functions<\/a>.<\/p><p>Again, I urge you not to use JSON unless it\u2019s absolutely necessary. You could emulate an entire document-oriented NoSQL database in MySQL, but it would negate many <a href=\"http:\/\/www.sitepoint.com\/sql-vs-nosql-choose\/\">benefits of SQL<\/a>, and you might as well switch to a real NoSQL system. That said, JSON data types could save some effort for more obscure data requirements within a SQL-based application.<\/p><div class=\"Article_authorBio l-mv4 t-bg-white m-border l-pa3\"><div class=\"l-d-f l-pt3\"><a class=\"l-s5 l-pa0 l-mr3 l-round l-o-h\" href=\"https:\/\/www.sitepoint.com\/author\/craig-buckler\/\">\n            <img src=\"https:\/\/secure.gravatar.com\/avatar\/439aeaff7de2bae365adc3eb4947b44d?s=96&amp;d=mm&amp;r=g\" width=\"96\" height=\"96\" alt=\"\" class=\"avatar avatar-96 wp-user-avatar wp-user-avatar-96 photo avatar-default\" \/><\/a><\/div><p>Craig is a freelance UK web consultant who built his first page for IE2.0 in 1995. Since that time he's been advocating standards, accessibility, and best-practice HTML5 techniques. He's created enterprise specifications, websites and online applications for companies and organisations including the UK Parliament, the European Parliament, the Department of Energy &amp; Climate Change, Microsoft, and more. He's written more than 1,000 articles for SitePoint and you can find him <a href=\"http:\/\/twitter.com\/craigbuckler\">@craigbuckler<\/a>.<\/p><\/div>","created_at":"2018-11-07T16:22:11+0000","updated_at":"2018-11-07T16:22:11+0000","published_at":"2016-04-29T17:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":6,"domain_name":"www.sitepoint.com","preview_picture":"https:\/\/www.sitepoint.com\/wp-content\/themes\/sitepoint\/assets\/images\/icon.web.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":54,"title":"digitalbazaar\/php-json-ld","url":"https:\/\/github.com\/digitalbazaar\/php-json-ld","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>[![Build Status][travis-ci-png]][travis-ci-site]\n[travis-ci-png]: <a href=\"https:\/\/travis-ci.org\/digitalbazaar\/php-json-ld.png?branch=master\" rel=\"nofollow\">https:\/\/travis-ci.org\/digitalbazaar\/php-json-ld.png?branch=master<\/a>\n[travis-ci-site]: <a href=\"https:\/\/travis-ci.org\/digitalbazaar\/php-json-ld\" rel=\"nofollow\">https:\/\/travis-ci.org\/digitalbazaar\/php-json-ld<\/a><\/p>\n<h2><a id=\"user-content-introduction\" class=\"anchor\" aria-hidden=\"true\" href=\"#introduction\"><\/a>Introduction<\/h2>\n<p>This library is an implementation of the <a href=\"http:\/\/json-ld.org\/\" rel=\"nofollow\">JSON-LD<\/a> specification in <a href=\"http:\/\/php.net\" rel=\"nofollow\">PHP<\/a>.<\/p>\n<p>JSON, as specified in <a href=\"http:\/\/tools.ietf.org\/html\/rfc7159\" rel=\"nofollow\">RFC7159<\/a>, is a simple language for representing\nobjects on the Web. Linked Data is a way of describing content across\ndifferent documents or Web sites. Web resources are described using\nIRIs, and typically are dereferencable entities that may be used to find\nmore information, creating a \"Web of Knowledge\". <a href=\"http:\/\/json-ld.org\/\" rel=\"nofollow\">JSON-LD<\/a> is intended\nto be a simple publishing method for expressing not only Linked Data in\nJSON, but for adding semantics to existing JSON.<\/p>\n<p>JSON-LD is designed as a light-weight syntax that can be used to express\nLinked Data. It is primarily intended to be a way to express Linked Data\nin JavaScript and other Web-based programming environments. It is also\nuseful when building interoperable Web Services and when storing Linked\nData in JSON-based document storage engines. It is practical and\ndesigned to be as simple as possible, utilizing the large number of JSON\nparsers and existing code that is in use today. It is designed to be\nable to express key-value pairs, RDF data, <a href=\"http:\/\/www.w3.org\/TR\/rdfa-core\/\" rel=\"nofollow\">RDFa<\/a> data,\n<a href=\"http:\/\/microformats.org\/\" rel=\"nofollow\">Microformats<\/a> data, and <a href=\"http:\/\/www.w3.org\/TR\/microdata\/\" rel=\"nofollow\">Microdata<\/a>. That is, it supports every\nmajor Web-based structured data model in use today.<\/p>\n<p>The syntax does not require many applications to change their JSON, but\neasily add meaning by adding context in a way that is either in-band or\nout-of-band. The syntax is designed to not disturb already deployed\nsystems running on JSON, but provide a smooth migration path from JSON\nto JSON with added semantics. Finally, the format is intended to be fast\nto parse, fast to generate, stream-based and document-based processing\ncompatible, and require a very small memory footprint in order to operate.<\/p>\n<h2><a id=\"user-content-quick-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#quick-examples\"><\/a>Quick Examples<\/h2>\n<div class=\"highlight highlight-text-html-php\"><pre>$doc = (object)array(\n  \"http:\/\/schema.org\/name\" =&gt; \"Manu Sporny\",\n  \"http:\/\/schema.org\/url\" =&gt; (object)array(\"@id\" =&gt; \"http:\/\/manu.sporny.org\/\"),\n  \"http:\/\/schema.org\/image\" =&gt; (object)array(\"@id\" =&gt; \"http:\/\/manu.sporny.org\/images\/manu.png\")\n);\n$context = (object)array(\n  \"name\" =&gt; \"http:\/\/schema.org\/name\",\n  \"homepage\" =&gt; (object)array(\"@id\" =&gt; \"http:\/\/schema.org\/url\", \"@type\" =&gt; \"@id\"),\n  \"image\" =&gt; (object)array(\"@id\" =&gt; \"http:\/\/schema.org\/image\", \"@type\" =&gt; \"@id\")\n);\n\/\/ compact a document according to a particular context\n\/\/ see: http:\/\/json-ld.org\/spec\/latest\/json-ld\/#compacted-document-form\n$compacted = jsonld_compact($doc, $context);\necho json_encode($compacted, JSON_PRETTY_PRINT | JSON_UNESCAPED_SLASHES);\n\/* Output:\n{\n  \"@context\": {...},\n  \"image\": \"http:\/\/manu.sporny.org\/images\/manu.png\",\n  \"homepage\": \"http:\/\/manu.sporny.org\/\",\n  \"name\": \"Manu Sporny\"\n}\n*\/\n\/\/ compact using URLs\njsonld_compact('http:\/\/example.org\/doc', 'http:\/\/example.org\/context');\n\/\/ expand a document, removing its context\n\/\/ see: http:\/\/json-ld.org\/spec\/latest\/json-ld\/#expanded-document-form\n$expanded = jsonld_expand($compacted) {\necho json_encode($expanded, JSON_PRETTY_PRINT | JSON_UNESCAPED_SLASHES);\n\/* Output:\n{\n  \"http:\/\/schema.org\/image\": [{\"@id\": \"http:\/\/manu.sporny.org\/images\/manu.png\"}],\n  \"http:\/\/schema.org\/name\": [{\"@value\": \"Manu Sporny\"}],\n  \"http:\/\/schema.org\/url\": [{\"@id\": \"http:\/\/manu.sporny.org\/\"}]\n}\n*\/\n\/\/ expand using URLs\njsonld_expand('http:\/\/example.org\/doc');\n\/\/ flatten a document\n\/\/ see: http:\/\/json-ld.org\/spec\/latest\/json-ld\/#flattened-document-form\n$flattened = jsonld_flatten($doc);\n\/\/ all deep-level trees flattened to the top-level\n\/\/ frame a document\n\/\/ see: http:\/\/json-ld.org\/spec\/latest\/json-ld-framing\/#introduction\n$framed = jsonld_frame($doc, $frame);\n\/\/ document transformed into a particular tree structure per the given frame\n\/\/ normalize a document using the RDF Dataset Normalization Algorithm\n\/\/ (URDNA2015), see: http:\/\/json-ld.github.io\/normalization\/spec\/\n$normalized = jsonld_normalize(\n  $doc, array('algorithm' =&gt; 'URDNA2015', 'format' =&gt; 'application\/nquads'));\n\/\/ normalized is a string that is a canonical representation of the document\n\/\/ that can be used for hashing, comparison, etc.\n\/\/ force HTTPS-only context loading:\n\/\/ use built-in secure document loader\njsonld_set_document_loader('jsonld_default_secure_document_loader');\n\/\/ set a default custom document loader\njsonld_set_document_loader('my_custom_doc_loader');\n\/\/ a custom loader that demonstrates using a simple in-memory mock for\n\/\/ certain contexts before falling back to the default loader\n\/\/ note: if you want to set this loader as the new default, you'll need to\n\/\/ store the previous default in another variable first and access that inside\n\/\/ the loader\nglobal $mocks;\n$mocks = array('http:\/\/example.com\/mycontext' =&gt; (object)array(\n  'hombre' =&gt; 'http:\/\/schema.org\/name'));\nfunction mock_load($url) {\n  global $jsonld_default_load_document, $mocks;\n  if(isset($mocks[$url])) {\n    \/\/ return a \"RemoteDocument\", it has these three properties:\n    return (object)array(\n      'contextUrl' =&gt; null,\n      'document' =&gt; $mocks[$url],\n      'documentUrl' =&gt; $url);\n  }\n  \/\/ use default loader\n  return call_user_func($jsonld_default_load_document, $url);\n}\n\/\/ use the mock loader for just this call, witout modifying the default one\n$compacted = jsonld_compact($foo, 'http:\/\/example.com\/mycontext', array(\n  'documentLoader' =&gt; 'mock_load'));\n\/\/ a custom loader that uses a simplistic in-memory cache (no invalidation)\nglobal $cache;\n$cache = array();\nfunction cache_load($url) {\n  global $jsonld_default_load_document, $cache;\n  if(isset($cache[$url])) {\n    return $cache[$url];\n  }\n  \/\/ use default loader\n  $doc = call_user_func($jsonld_default_load_document, $url);\n  $cache[$url] = $doc;\n  return $doc;\n}\n\/\/ use the cache loader for just this call, witout modifying the default one\n$compacted = jsonld_compact($foo, 'http:\/\/schema.org', array(\n  'documentLoader' =&gt; 'cache_load'));<\/pre><\/div>\n<h2><a id=\"user-content-commercial-support\" class=\"anchor\" aria-hidden=\"true\" href=\"#commercial-support\"><\/a>Commercial Support<\/h2>\n<p>Commercial support for this library is available upon request from\n<a href=\"http:\/\/digitalbazaar.com\/\" rel=\"nofollow\">Digital Bazaar<\/a>: <a href=\"mailto:support@digitalbazaar.com\">support@digitalbazaar.com<\/a><\/p>\n<h2><a id=\"user-content-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#source\"><\/a>Source<\/h2>\n<p>The source code for the PHP implementation of the JSON-LD API\nis available at:<\/p>\n<p><a href=\"http:\/\/github.com\/digitalbazaar\/php-json-ld\">http:\/\/github.com\/digitalbazaar\/php-json-ld<\/a><\/p>\n<h2><a id=\"user-content-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#tests\"><\/a>Tests<\/h2>\n<p>This library includes a sample testing utility which may be used to verify\nthat changes to the processor maintain the correct output.<\/p>\n<p>To run the sample tests you will need to get the test suite files by cloning\nthe <code>json-ld.org<\/code> and <code>normalization<\/code> repositories hosted on GitHub:<\/p>\n<p>Then run the PHPUnit test.php application and point it at the directories\ncontaining the tests:<\/p>\n<pre>phpunit --group json-ld.org test.php -d {PATH_TO_JSON_LD_ORG\/test-suite}\nphpunit --group normalization test.php -d {PATH_TO_NORMALIZATION\/tests}\n<\/pre>\n<\/article>","created_at":"2018-11-07T16:22:11+0000","updated_at":"2018-11-07T16:22:11+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/167436?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":51,"title":"On Generative Algorithms: Introduction \u00b7 inconvergent","url":"https:\/\/inconvergent.net\/generative\/","content":"<div class=\"text\"><h2>Sections<\/h2><ol><li><a href=\"https:\/\/inconvergent.net\/generative\/#introduction\">Introduction<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/hyphae\/\">Hyphae<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/trees\/\">Trees<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/linetrace\/\">Linetrace<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-line\/\">Differential Line<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-mesh\/\">Differential Mesh<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-mesh-3d\/\">Differential Mesh 3d<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/fractures\/\">Fractures<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/sand-spline\/\">Sand Spline<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-lattice\/\">Differential Lattice<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/sand-creatures\/\">Sand Creatures<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/sand-glyphs\/\">Sand Glyphs<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/not-a-conclusion\/\">Not a Conclusion<\/a><\/li>\n<\/ol><p><em>This essay was written between late 2015 and mid 2016.<\/em><\/p><p>I've always been fascinated with patterns. It doesn't really matter what\nkind of patterns; I've played with networks, leaves and leaf venation,\nbranches, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lichtenberg_figure\">lightning<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Boids\">flocking<\/a>, tracing outlines\nof shapes, river formation, rock sediments, landscapes, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Slime_mold\">slime mold<\/a>, lichens,\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Reaction%E2%80%93diffusion_system\">reaction-diffusion<\/a>,\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Cellular_automaton\">cellular\nautomaton<\/a>, some fractals, and a few other things. I think what I\nenjoy the most is how complex and intricate results you can get from a set\nof simple rules.<\/p><\/div><div class=\"text\"><p>Recently I've been particularly interested in biological patterns, and\ndifferential growth. My background is in numerical mathematics, not\nbiology. So I have limited knowledge of how biological systems actually\nwork. Even so, I've been experimenting quite a lot with recreating various\nbiological behaviours. Part of the challenge is to try to recreate some\npattern or behaviour with as few and as simple rules as possible.<\/p><p>Sometimes I manage to recreate the phenomenon I set out to create, and\nsometimes not. More often than not I get something interesting, even if it\nis not always what I expected.<\/p><h2>A Confession<\/h2><p>In the interest of disclosure I should point out that I sometimes base my\nwork on the ideas that I have seen done elsewhere by a number of other\ngenerative artists. Most notably <a href=\"http:\/\/complexification.net\">Jared Tarbell<\/a> and <a href=\"http:\/\/n-e-r-v-o-u-s.com\/\">Nervous System<\/a>. For instance the\nalgorithm I named <a href=\"https:\/\/github.com\/inconvergent\/orbitals\">Orbitals<\/a> (image\nbelow) is very heavily based on <a href=\"http:\/\/www.complexification.net\/gallery\/machines\/happyPlace\/\">Happy\nPlace<\/a> by Tarbell.<\/p><\/div><div class=\"text\"><p> I point this out because I have always found it hard to know where a\npiece of software\u2014and in my particular case, the images generated by that\nsoftware\u2014is <em>novel<\/em> enough to be viewed as a separate work.<\/p><\/div><div class=\"text\"><h2>Inconvergent<\/h2><p>I started working with generative algorithms when I was supposed to be\nstudying for my exams at university. For that reason I bought the domain\nname <a href=\"https:\/\/inconvergent.net\/\">inconvergent.net<\/a>, as a joke on me diverging from my\nstudies. The first thing I did was to copy several of Tarbell's algorithms\nusing Javascript\/Canvas. After a while I started getting a few other ideas\nof my own.<\/p><\/div><div class=\"text\"><p>Apart from Orbitals, the first algorithm I got working was <a href=\"https:\/\/inconvergent.net\/generative\/hyphae\">Hyphae<\/a>. It came to me when I was trying to recreate the\nbehaviour of <a href=\"http:\/\/n-e-r-v-o-u-s.com\/shop\/generativeProduct.php?code=99\">Hyphae<\/a>\nby Nervous System. At this point I hadn't realized how complicated that\nalgorithm actually is, and I spent way too much time trying and failing to\nget it anywhere near decent. I implemented it later, and you can read the\n<a href=\"http:\/\/algorithmicbotany.org\/papers\/venation.sig2005.pdf\">Siggraph\npaper<\/a> describing it if you are interested. It is a fascinating\nread!<\/p><h2>Code<\/h2><p>I make almost all of my code available at <a href=\"https:\/\/github.com\/inconvergent\">Github<\/a>. All the following\nsections have a link to the corresponding repository. Unfortunately not all\nof it is entirely well documented or up to date.<\/p><a class=\"next\" href=\"https:\/\/inconvergent.net\/generative\/hyphae\/\">Read More<\/a><\/div>","created_at":"2018-11-07T16:22:10+0000","updated_at":"2018-11-07T16:22:10+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"inconvergent.net","preview_picture":"https:\/\/inconvergent.net\/img\/generative\/head.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":52,"title":"On Generative Algorithms: Introduction \u00b7 inconvergent","url":"https:\/\/inconvergent.net\/generative\/","content":"<div class=\"text\"><h2>Sections<\/h2><ol><li><a href=\"https:\/\/inconvergent.net\/generative\/#introduction\">Introduction<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/hyphae\/\">Hyphae<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/trees\/\">Trees<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/linetrace\/\">Linetrace<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-line\/\">Differential Line<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-mesh\/\">Differential Mesh<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-mesh-3d\/\">Differential Mesh 3d<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/fractures\/\">Fractures<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/sand-spline\/\">Sand Spline<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/differential-lattice\/\">Differential Lattice<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/sand-creatures\/\">Sand Creatures<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/sand-glyphs\/\">Sand Glyphs<\/a><\/li>\n<li><a href=\"https:\/\/inconvergent.net\/generative\/not-a-conclusion\/\">Not a Conclusion<\/a><\/li>\n<\/ol><p><em>This essay was written between late 2015 and mid 2016.<\/em><\/p><p>I've always been fascinated with patterns. It doesn't really matter what\nkind of patterns; I've played with networks, leaves and leaf venation,\nbranches, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Lichtenberg_figure\">lightning<\/a>, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Boids\">flocking<\/a>, tracing outlines\nof shapes, river formation, rock sediments, landscapes, <a href=\"https:\/\/en.wikipedia.org\/wiki\/Slime_mold\">slime mold<\/a>, lichens,\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Reaction%E2%80%93diffusion_system\">reaction-diffusion<\/a>,\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Cellular_automaton\">cellular\nautomaton<\/a>, some fractals, and a few other things. I think what I\nenjoy the most is how complex and intricate results you can get from a set\nof simple rules.<\/p><\/div><div class=\"text\"><p>Recently I've been particularly interested in biological patterns, and\ndifferential growth. My background is in numerical mathematics, not\nbiology. So I have limited knowledge of how biological systems actually\nwork. Even so, I've been experimenting quite a lot with recreating various\nbiological behaviours. Part of the challenge is to try to recreate some\npattern or behaviour with as few and as simple rules as possible.<\/p><p>Sometimes I manage to recreate the phenomenon I set out to create, and\nsometimes not. More often than not I get something interesting, even if it\nis not always what I expected.<\/p><h2>A Confession<\/h2><p>In the interest of disclosure I should point out that I sometimes base my\nwork on the ideas that I have seen done elsewhere by a number of other\ngenerative artists. Most notably <a href=\"http:\/\/complexification.net\">Jared Tarbell<\/a> and <a href=\"http:\/\/n-e-r-v-o-u-s.com\/\">Nervous System<\/a>. For instance the\nalgorithm I named <a href=\"https:\/\/github.com\/inconvergent\/orbitals\">Orbitals<\/a> (image\nbelow) is very heavily based on <a href=\"http:\/\/www.complexification.net\/gallery\/machines\/happyPlace\/\">Happy\nPlace<\/a> by Tarbell.<\/p><\/div><div class=\"text\"><p> I point this out because I have always found it hard to know where a\npiece of software\u2014and in my particular case, the images generated by that\nsoftware\u2014is <em>novel<\/em> enough to be viewed as a separate work.<\/p><\/div><div class=\"text\"><h2>Inconvergent<\/h2><p>I started working with generative algorithms when I was supposed to be\nstudying for my exams at university. For that reason I bought the domain\nname <a href=\"https:\/\/inconvergent.net\/\">inconvergent.net<\/a>, as a joke on me diverging from my\nstudies. The first thing I did was to copy several of Tarbell's algorithms\nusing Javascript\/Canvas. After a while I started getting a few other ideas\nof my own.<\/p><\/div><div class=\"text\"><p>Apart from Orbitals, the first algorithm I got working was <a href=\"https:\/\/inconvergent.net\/generative\/hyphae\">Hyphae<\/a>. It came to me when I was trying to recreate the\nbehaviour of <a href=\"http:\/\/n-e-r-v-o-u-s.com\/shop\/generativeProduct.php?code=99\">Hyphae<\/a>\nby Nervous System. At this point I hadn't realized how complicated that\nalgorithm actually is, and I spent way too much time trying and failing to\nget it anywhere near decent. I implemented it later, and you can read the\n<a href=\"http:\/\/algorithmicbotany.org\/papers\/venation.sig2005.pdf\">Siggraph\npaper<\/a> describing it if you are interested. It is a fascinating\nread!<\/p><h2>Code<\/h2><p>I make almost all of my code available at <a href=\"https:\/\/github.com\/inconvergent\">Github<\/a>. All the following\nsections have a link to the corresponding repository. Unfortunately not all\nof it is entirely well documented or up to date.<\/p><a class=\"next\" href=\"https:\/\/inconvergent.net\/generative\/hyphae\/\">Read More<\/a><\/div>","created_at":"2018-11-07T16:22:10+0000","updated_at":"2018-11-07T16:22:10+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"inconvergent.net","preview_picture":"https:\/\/inconvergent.net\/img\/generative\/head.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":49,"title":"2.10.1.1 Initializing the Data Directory Manually Using mysqld","url":"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/data-directory-initialization-mysqld.html","content":"<div class=\"titlepage\"><div><h4 class=\"title\">2.10.1.1\u00a0Initializing the Data Directory Manually Using mysqld<\/h4><\/div><\/div><p>\n        This section describes how to initialize the data directory\n        using <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a>, the MySQL server.\n<\/p><div class=\"note\"><p>Note<\/p><p>\n          In MySQL 8.0, the default authentication plugin has changed\n          from <code class=\"literal\">mysql_native_password<\/code> to\n          <code class=\"literal\">caching_sha2_password<\/code>, and the\n          <code class=\"literal\">'root'@'localhost'<\/code> administrative account\n          uses <code class=\"literal\">caching_sha2_password<\/code> by default. If\n          you prefer that the <code class=\"literal\">root<\/code> account use the\n          previous default authentication plugin\n          (<code class=\"literal\">mysql_native_password<\/code>), see\n          <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/upgrading-from-previous-series.html#upgrade-caching-sha2-password-root-account\" title=\"caching_sha2_password and the root Administrative Account\">caching_sha2_password and the root Administrative Account<\/a>.\n<\/p><\/div><p>\n        The following instructions assume that your current location is\n        the MySQL installation directory, represented here by\n        <em class=\"replaceable\"><code>BASEDIR<\/code><\/em>:\n      <\/p><pre class=\"language-terminal\">shell&gt; cd <em class=\"replaceable\">BASEDIR<\/em><\/pre><p>\n        To initialize the data directory, invoke\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> with the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a> or\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a> option,\n        depending on whether you want the server to generate a random\n        initial password for the <code class=\"literal\">'root'@'localhost'<\/code>\n        account.\n      <\/p><p>\n        On Windows, use one of these commands:\n      <\/p><pre class=\"language-terminal\">C:\\&gt; bin\\mysqld --initialize --console\nC:\\&gt; bin\\mysqld --initialize-insecure --console<\/pre><p>\n        On Unix and Unix-like systems, it is important to make sure that\n        the database directories and files are owned by the\n        <code class=\"literal\">mysql<\/code> login account so that the server has\n        read and write access to them when you run it later. To ensure\n        this, start <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> from the system\n        <code class=\"literal\">root<\/code> account and include the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_user\"><code class=\"option\">--user<\/code><\/a> option as shown here:\n      <\/p><pre class=\"language-terminal\">shell&gt; bin\/mysqld --initialize --user=mysql\nshell&gt; bin\/mysqld --initialize-insecure --user=mysql<\/pre><p>\n        Otherwise, execute the program while logged in as\n        <code class=\"literal\">mysql<\/code>, in which case you can omit the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_user\"><code class=\"option\">--user<\/code><\/a> option from the command.\n      <\/p><p>\n        Regardless of platform, use\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a> for \u201csecure by\n        default\u201d installation (that is, including generation of a\n        random initial <code class=\"literal\">root<\/code> password). In this case,\n        the password is marked as expired and you will need to choose a\n        new one. With the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a> option, no\n        <code class=\"literal\">root<\/code> password is generated; it is assumed\n        that you will assign a password to the account in timely fashion\n        before putting the server into production use.\n      <\/p><p>\n        It might be necessary to specify other options such as\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_basedir\"><code class=\"option\">--basedir<\/code><\/a> or\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_datadir\"><code class=\"option\">--datadir<\/code><\/a> if\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> cannot identify the correct locations\n        for the installation directory or data directory. For example\n        (enter the command on one line):\n      <\/p><pre class=\"language-terminal\">shell&gt; bin\/mysqld --initialize --user=mysql\n         --basedir=\/opt\/mysql\/mysql\n         --datadir=\/opt\/mysql\/mysql\/data<\/pre><p>\n        Alternatively, put the relevant option settings in an option\n        file and pass the name of that file to\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a>. For Unix and Unix-like systems,\n        suppose that the option file name is\n        <code class=\"filename\">\/opt\/mysql\/mysql\/etc\/my.cnf<\/code>. Put these\n        lines in the file:\n      <\/p><pre class=\"language-ini\">[mysqld]\nbasedir=\/opt\/mysql\/mysql\ndatadir=\/opt\/mysql\/mysql\/data<\/pre><p>\n        Then invoke <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> as follows (enter the\n        command on a single line with the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/option-file-options.html#option_general_defaults-file\"><code class=\"option\">--defaults-file<\/code><\/a> option first):\n      <\/p><pre class=\"language-terminal\">shell&gt; bin\/mysqld --defaults-file=\/opt\/mysql\/mysql\/etc\/my.cnf\n         --initialize --user=mysql<\/pre><p>\n        On Windows, suppose that <code class=\"literal\">C:\\my.ini<\/code> contains\n        these lines:\n      <\/p><pre class=\"language-ini\">[mysqld]\nbasedir=C:\\\\Program Files\\\\MySQL\\\\MySQL Server 8.0\ndatadir=D:\\\\MySQLdata<\/pre><p>\n        Then invoke <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> as follows (the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/option-file-options.html#option_general_defaults-file\"><code class=\"option\">--defaults-file<\/code><\/a> option must be\n        first):\n      <\/p><pre class=\"language-terminal\">C:\\&gt; bin\/mysqld --defaults-file=C:\\my.ini --initialize --console<\/pre><p>\n        When invoked with the\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a> or\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a> option,\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> performs the following initialization\n        sequence.\n<\/p><div class=\"note\"><p>Note<\/p><p>\n          The server writes any messages to its standard error output.\n          This may be redirected to the error log, so look there if you\n          do not see the messages on your screen. For information about\n          the error log, including where it is located, see\n          <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/error-log.html\" title=\"5.4.2\u00a0The Error Log\">Section\u00a05.4.2, \u201cThe Error Log\u201d<\/a>.\n        <\/p><p>\n          On Windows, use the <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_console\"><code class=\"option\">--console<\/code><\/a>\n          option to direct messages to the console.\n<\/p><\/div><div class=\"orderedlist\"><ol class=\"orderedlist\" type=\"1\"><li class=\"listitem\"><p>\n            The server checks for the existence of the data directory as\n            follows:\n<\/p>\n<div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>\n                If no data directory exists, the server creates it.\n              <\/p><\/li><li class=\"listitem\"><p>\n                If a data directory exists but is not empty (that is, it\n                contains files or subdirectories), the server exits\n                after producing an error message:\n              <\/p><pre class=\"language-simple\">[ERROR] --initialize specified but the data directory exists. Aborting.<\/pre><p>\n                In this case, remove or rename the data directory and\n                try again.\n              <\/p><p>\n                An existing data directory is permitted to be nonempty\n                if every entry has a name that begins with a period\n                (<code class=\"literal\">.<\/code>).\n<\/p><\/li><\/ul><\/div>\n<\/li><li class=\"listitem\"><p>\n            Within the data directory, the server creates the\n            <code class=\"literal\">mysql<\/code> system database and its tables,\n            including the grant tables, server-side help tables, and\n            time zone tables. For a complete listing and description of\n            the grant tables, see <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/privilege-system.html\" title=\"6.2\u00a0The MySQL Access Privilege System\">Section\u00a06.2, \u201cThe MySQL Access Privilege System\u201d<\/a>.\n          <\/p><\/li><li class=\"listitem\"><p>\n            The server initializes the\n            <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/glossary.html#glos_system_tablespace\" title=\"system tablespace\">system\n            tablespace<\/a> and related data structures needed to\n            manage <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/innodb-storage-engine.html\" title=\"Chapter\u00a015\u00a0The InnoDB Storage Engine\"><code class=\"literal\">InnoDB<\/code><\/a> tables.\n<\/p>\n<div class=\"note\"><p>Note<\/p><p>\n              After <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a> sets up the\n              <code class=\"literal\">InnoDB<\/code>\n              <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/glossary.html#glos_system_tablespace\" title=\"system tablespace\">system\n              tablespace<\/a>, changes to some tablespace\n              characteristics require setting up a whole new\n              <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/glossary.html#glos_instance\" title=\"instance\">instance<\/a>. This\n              includes the file name of the first file in the system\n              tablespace and the number of undo logs. If you do not want\n              to use the default values, make sure that the settings for\n              the <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/innodb-parameters.html#sysvar_innodb_data_file_path\"><code class=\"literal\">innodb_data_file_path<\/code><\/a>\n              and <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/innodb-parameters.html#sysvar_innodb_log_file_size\"><code class=\"literal\">innodb_log_file_size<\/code><\/a>\n              configuration parameters are in place in the MySQL\n              <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/glossary.html#glos_configuration_file\" title=\"configuration file\">configuration\n              file<\/a> before running <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a>. Also\n              make sure to specify as necessary other parameters that\n              affect the creation and location of\n              <code class=\"literal\">InnoDB<\/code> files, such as\n              <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/innodb-parameters.html#sysvar_innodb_data_home_dir\"><code class=\"literal\">innodb_data_home_dir<\/code><\/a> and\n              <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/innodb-parameters.html#sysvar_innodb_log_group_home_dir\"><code class=\"literal\">innodb_log_group_home_dir<\/code><\/a>.\n            <\/p><p>\n              If those options are in your configuration file but that\n              file is not in a location that MySQL reads by default,\n              specify the file location using the\n              <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_defaults-extra-file\"><code class=\"option\">--defaults-extra-file<\/code><\/a>\n              option when you run <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysqld.html\" title=\"4.3.1\u00a0mysqld \u2014 The MySQL Server\"><strong>mysqld<\/strong><\/a>.\n<\/p><\/div>\n<\/li><li class=\"listitem\"><p>\n            The server creates a <code class=\"literal\">'root'@'localhost'<\/code>\n            superuser account and other reserved accounts (see\n            <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/reserved-users.html\" title=\"6.3.5\u00a0Reserved User Accounts\">Section\u00a06.3.5, \u201cReserved User Accounts\u201d<\/a>). Some reserved accounts\n            are locked and cannot be used by clients, but\n            <code class=\"literal\">'root'@'localhost'<\/code> is intended for\n            administrative use and you should assign it a password.\n          <\/p><p>\n            The server's action with respect to a password for the\n            <code class=\"literal\">'root'@'localhost'<\/code> account depends on how\n            you invoke it:\n<\/p>\n<div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>\n                With <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a> but not\n                <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a>,\n                the server generates a random password, marks it as\n                expired, and writes a message displaying the password:\n              <\/p><pre class=\"language-simple\">[Warning] A temporary password is generated for root@localhost:\niTag*AfrH5ej<\/pre><\/li><li class=\"listitem\"><p>\n                With\n                <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a>,\n                (either with or without\n                <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a> because\n                <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a>\n                implies <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a>),\n                the server does not generate a password or mark it\n                expired, and writes a warning message:\n              <\/p><pre class=\"language-simple\">Warning] root@localhost is created with an empty password ! Please\nconsider switching off the --initialize-insecure option.<\/pre><\/li><\/ul><\/div>\n<p>\n            See later in this section for instructions on assigning a\n            new <code class=\"literal\">'root'@'localhost'<\/code> password.\n          <\/p><\/li><li class=\"listitem\"><p>\n            The server populates the server-side help tables if content\n            is available (in the\n            <code class=\"filename\">fill_help_tables.sql<\/code> file). The server\n            does not populate the time zone tables; to do so, see\n            <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/time-zone-support.html\" title=\"5.1.13\u00a0MySQL Server Time Zone Support\">Section\u00a05.1.13, \u201cMySQL Server Time Zone Support\u201d<\/a>.\n          <\/p><\/li><li class=\"listitem\"><p>\n            If the <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_init-file\"><code class=\"option\">--init-file<\/code><\/a> option was\n            given to name a file of SQL statements, the server executes\n            the statements in the file. This option enables you to\n            perform custom bootstrapping sequences.\n          <\/p><p>\n            When the server operates in bootstrap mode, some\n            functionality is unavailable that limits the statements\n            permitted in the file. These include statements that relate\n            to account management (such as <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/create-user.html\" title=\"13.7.1.3\u00a0CREATE USER Syntax\"><code class=\"literal\">CREATE\n            USER<\/code><\/a> or <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/grant.html\" title=\"13.7.1.6\u00a0GRANT Syntax\"><code class=\"literal\">GRANT<\/code><\/a>),\n            replication, and global transaction identifiers.\n          <\/p><\/li><li class=\"listitem\"><p>\n            The server exits.\n<\/p><\/li><\/ol><\/div><p>\n        After you initialize the data directory by starting the server\n        with <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a> or\n        <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a>, start the\n        server normally (that is, without either of those options) and\n        assign the <code class=\"literal\">'root'@'localhost'<\/code> account a new\n        password:\n<\/p><div class=\"orderedlist\"><ol class=\"orderedlist\" type=\"1\"><li class=\"listitem\"><p>\n            Start the server. For instructions, see\n            <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/starting-server.html\" title=\"2.10.2\u00a0Starting the Server\">Section\u00a02.10.2, \u201cStarting the Server\u201d<\/a>.\n          <\/p><\/li><li class=\"listitem\"><p>\n            Connect to the server:\n<\/p>\n<div class=\"itemizedlist\"><ul class=\"itemizedlist\"><li class=\"listitem\"><p>\n                If you used <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize\"><code class=\"option\">--initialize<\/code><\/a>\n                but not\n                <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a> to\n                initialize the data directory, connect to the server as\n                <code class=\"literal\">root<\/code> using the random password that\n                the server generated during the initialization sequence:\n              <\/p><pre class=\"language-terminal\">shell&gt; mysql -u root -p\nEnter password: <em class=\"replaceable\">(enter the random root password here)<\/em><\/pre><p>\n                Look in the server error log if you do not know this\n                password.\n              <\/p><\/li><li class=\"listitem\"><p>\n                If you used\n                <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_initialize-insecure\"><code class=\"option\">--initialize-insecure<\/code><\/a> to\n                initialize the data directory, connect to the server as\n                <code class=\"literal\">root<\/code> without a password:\n              <\/p><pre class=\"language-terminal\">shell&gt; mysql -u root --skip-password<\/pre><\/li><\/ul><\/div>\n<\/li><li class=\"listitem\"><p>\n            After connecting, assign a new <code class=\"literal\">root<\/code>\n            password:\n          <\/p><pre class=\"language-sql\">mysql&gt; ALTER USER 'root'@'localhost' IDENTIFIED BY '<em class=\"replaceable\">new_password'<\/em>;<\/pre><\/li><\/ol><\/div><div class=\"note\"><p>Note<\/p><p>\n          An attempt to connect to the host <code class=\"literal\">127.0.0.1<\/code>\n          normally resolves to the <code class=\"literal\">localhost<\/code> account.\n          However, this fails if the server is run with the\n          <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_skip-name-resolve\"><code class=\"option\">--skip-name-resolve<\/code><\/a> option. If\n          you plan to do that, make sure that an account exists that can\n          accept a connection. For example, to be able to connect as\n          <code class=\"literal\">root<\/code> using\n          <code class=\"option\">--host=127.0.0.1<\/code> or\n          <code class=\"option\">--host=::1<\/code>, create these accounts:\n        <\/p><pre class=\"language-sql\">CREATE USER 'root'@'127.0.0.1' IDENTIFIED BY '<em class=\"replaceable\">root-password<\/em>';\nCREATE USER 'root'@'::1' IDENTIFIED BY '<em class=\"replaceable\">root-password<\/em>';<\/pre><p>\n          It is possible to put those statements in a file to be\n          executed by the <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/server-options.html#option_mysqld_init-file\"><code class=\"option\">--init-file<\/code><\/a>\n          option discussed previously.\n<\/p><\/div><div class=\"note\"><p>Note<\/p><p>\n          The data directory initialization sequence performed by the\n          server does not substitute for the actions performed by\n          <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysql-secure-installation.html\" title=\"4.4.2\u00a0mysql_secure_installation \u2014 Improve MySQL Installation Security\"><strong>mysql_secure_installation<\/strong><\/a> or\n          <a class=\"link\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysql-ssl-rsa-setup.html\" title=\"4.4.3\u00a0mysql_ssl_rsa_setup \u2014 Create SSL\/RSA Files\"><strong>mysql_ssl_rsa_setup<\/strong><\/a>. See\n          <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysql-secure-installation.html\" title=\"4.4.2\u00a0mysql_secure_installation \u2014 Improve MySQL Installation Security\">Section\u00a04.4.2, \u201c<strong>mysql_secure_installation<\/strong> \u2014 Improve MySQL Installation Security\u201d<\/a>, and\n          <a class=\"xref\" href=\"https:\/\/dev.mysql.com\/doc\/refman\/8.0\/en\/mysql-ssl-rsa-setup.html\" title=\"4.4.3\u00a0mysql_ssl_rsa_setup \u2014 Create SSL\/RSA Files\">Section\u00a04.4.3, \u201c<strong>mysql_ssl_rsa_setup<\/strong> \u2014 Create SSL\/RSA Files\u201d<\/a>.\n<\/p><\/div>","created_at":"2018-11-07T16:22:09+0000","updated_at":"2018-11-07T16:22:09+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"dev.mysql.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":50,"title":"Functional Core, Imperative Shell","url":"https:\/\/www.destroyallsoftware.com\/screencasts\/catalog\/functional-core-imperative-shell","content":"<p>\n        Purely functional code makes some things easier to understand: because values don't change, you can call functions and know that only their return value matters\u2014they don't change anything outside themselves. But this makes many real-world applications difficult: how do you write to a database, or to the screen?<\/p><p>In this screencast we look at one method for crossing this divide. We review a Twitter client whose core is functional: managing tweets, syncing timelines to incoming Twitter API data, remembering cursor positions within the tweet list, and rendering tweets to text for display. This functional core is surrounded by a shell of imperative code: it manipulates stdin, stdout, the database, and the network, all based on values produced by the functional core.<\/p><p>This design has many nice side effects. For example, testing the functional pieces is very easy, and it often naturally allows isolated testing with no test doubles. It also leads to an imperative shell with few conditionals, making reasoning about the program's state over time much easier.\n      <\/p>","created_at":"2018-11-07T16:22:09+0000","updated_at":"2018-11-07T16:22:09+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"www.destroyallsoftware.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":46,"title":"activitypub\/rwot\/even_more_distributed_activitypub.org \u00b7 master \u00b7 Christopher Allan Webber \/ talks","url":"https:\/\/gitlab.com\/dustyweb\/talks\/blob\/master\/activitypub\/rwot\/even_more_distributed_activitypub.org","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:22:08+0000","updated_at":"2018-11-07T16:22:08+0000","annotations":[],"reading_time":0,"domain_name":"gitlab.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":47,"title":"Improve Website Performance by Enabling Caching in Apache - Serverlab","url":"https:\/\/www.serverlab.ca\/tutorials\/linux\/web-servers-linux\/improve-website-performance-by-enabling-caching-in-apache\/","content":"<h2>Overview<\/h2>\n<p>Most web sites have large amounts of content that remains unchanged or is rarely modified after publication. And everytime it is requested from the web server, modified or not, it is reprocessed and transmitted to the client, unnecessarily consuming valuable system resources and network bandwidth. As your website or application gains in popularity, more demand we be put on your server to process content that is, although dynamically generated, static.<\/p>\n<p>This isn\u2019t an efficient use of resources and it will eventually end up costing you a lot of money, as you install more hardware to keep up with load. This is were caching comes in. Let\u2019s not waste CPU cycles or RAM processing previously accessed content that is unlikely to change within a determined amount of time. Instead, we\u2019ll serve pre-processed content, temporarily stored on the server or the\u00a0 web browsers cache. And only after content has been modified or a set amount of time has passed will we reprocess the requested content.<\/p>\n<h2>Configure Caching in Apache<\/h2>\n<p>Apache comes with three modules for caching content, one enables it and the remaining two determine where the cache store exists \u2013 on disk or in memory. Determining which module to use for the cache store depends on your available hardware resources and performance requirements.<\/p>\n<p>Serving from disk is slow but less expensive. Serving from memory is fast, but expensive \u2013 both in cost and in resource consumption. However, you can boost disk cache performances by placing it on SSDFLASH storage instead of conventional spinning disk.<\/p>\n<ol><li>Ensure the cache_module is being loaded by Apache, by verifying the following line exists in Apache\u2019s server configuration file, uncommented.\n<pre>LoadModule cache_module modules\/mod_cache.so<\/pre>\n<\/li>\n<li>For disk caching, ensure the disk_cache_module is being loaded by Apache. Look for the following line, uncommented.\n<pre>LoadModule disk_cache_module modules\/mod_disk_cache.so<\/pre>\n<\/li>\n<li>Add the following lines in to either the Apache server configuration file (for global) or inside of a VirtualHost directive (application localized).\n<pre>CacheEnable disk \/&#13;\nCacheRoot \/webapps\/cache\/app1&#13;\nCacheDefaultExpire 3600&#13;\nCacheDisable \/wp-admin<\/pre>\n<table><tbody><tr><th>CacheEnable disk \/<\/th>\n<td>Unable caching to disk for relative directory \u201c\/\u201d.<\/td>\n<\/tr><tr><th>Cacheroot<\/th>\n<td>Set the cache store directory, where all cached content will be saved.<\/td>\n<\/tr><tr><th>CacheDefaultExpire<\/th>\n<td>Set the default expire date, relative to the original request date, in seconds.<\/td>\n<\/tr><tr><th>CacheDisable<\/th>\n<td>Disables caching for relative paths following the option. Sensitive areas and those that shouldn\u2019t be cached should be added here.<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<\/ol><h2>Setting Content Expiration<\/h2>\n<p>Caching requires an expiration date for it to work. Without an expiration date on your content, the cache cannot determine if it is stale or not.<\/p>\n<h3>Use Apache\u2019s Mod_Expires<\/h3>\n<p>This module allows you to define expiration dates for your content, as a whole or individually based on type or matching string.<\/p>\n<ol><li>Ensure the module is being loaded into Apache. Open the server configuration file (httpd.conf in CentOS) and look for this line. Uncomment it, if it is commented out with a \u2018#\u2019.\n<pre>LoadModule expires_module modules\/mod_expires.so<\/pre>\n<\/li>\n<li>Add the following lines to either the Apache server configuration, a virtual host configuration, a directory directive, or .htaccess, depending on where you want your caching policy set.\n<pre>&lt;IfModule mod_expires.c&gt;&#13;\n        ExpiresActive On&#13;\n        ExpiresDefault \"access plus 1 day\"&#13;\n&lt;\/IfModule&gt;&lt;\/pre&gt;<\/pre>\n<table><tbody><tr><th>ExpiresActive\u00a0On<\/th>\n<td>Turns on mod expires.<\/td>\n<\/tr><tr><th>ExpiresDefault<\/th>\n<td>Set the default expiration date for all content.\u00a0<strong>Access plus 1 day<\/strong>\u00a0sets the expiration time to the content\u2019s access time + 1 day, meaning it will be cached for 24 hours.<\/td>\n<\/tr><\/tbody><\/table><\/li>\n<li>If you want to assign different expiration values to specific content types, you can use the\u00a0<strong>ExpiresByType<\/strong>\u00a0option, in addition to or without the\u00a0<strong>ExpiresDefault<\/strong>\u00a0option. Here are a few examples:\n<pre><dfn>&lt;IfModule mod_expires.c&gt;&#13;\n        ExpiresActive On&#13;\n        ExpiresDefault \"access plus 1 day\"<\/dfn>&#13;\n        ExpiresByType image\/jpg \"access plus 5 days\"&#13;\n        ExpiresByType image\/jpeg \"access plus 5 days\"&#13;\n        ExpiresByType image\/gif \"access plus 5 days\"&#13;\n        ExpiresByType image\/png \"access plus 5 days\"&#13;\n        ExpiresByType text\/css \"access plus 1 month\"&#13;\n        ExpiresByType application\/pdf \"access plus 1 month\"&#13;\n        ExpiresByType text\/x-javascript \"access plus 1 month\"&#13;\n        ExpiresByType application\/x-shockwave-flash \"access plus 1 month\"&#13;\n        ExpiresByType image\/x-icon \"access plus 1 year\"&#13;\n<dfn>&lt;\/IfModule&gt;<\/dfn><\/pre>\n<\/li>\n<\/ol><h3>Use HTTP Headers in the Web Application<\/h3>\n<p>Expiration and last modification dates can be defined by the web application using HTTP Header parameters. This gives content freshness control to the developers or the application. How you do this depends on your application.<\/p>\n<ul><li><strong>HTML<\/strong><br \/>\nUse the Meta tag and define the content age. In the example below, set the cache to private (requesting client only) and the maximum age to 1 hour (3600 seconds).<br \/><code>&lt;meta http-equiv=\"Cache-control\" content=\"private,max-age:3600\"&gt;<\/code><\/li>\n<li><strong>PHP<\/strong><br \/>\nUse the\u00a0<strong>Header()<\/strong>\u00a0function to set the content age in the documents header.\n<pre>header(\"Cache-Control: private, max-age=3600\");<\/pre>\n<\/li>\n<\/ul><h2>Cache Non-Expiring Content<\/h2>\n<p>Somtimes you need to cache content that doesn\u2019t have an expiration date set. As mentioned earlier, expiration dates are a requirement for the caching mechanism to work, by default. However, we can instruct Apache to append a default expiration date to contents that have no defined an expiration date, and cache it. The CacheIgnoreNoLastMod allows us to do this.<\/p>\n<ol><li>Add the\u00a0<strong>CacheIgnoreNoLastMod<\/strong>\u00a0option with a value of On to location where you have enabled caching \u2013 Apache\u2019s server configuration file or a Virtual Hosts configuration file.\n<pre><dfn>CacheEnable disk \/&#13;\nCacheRoot \/webapps\/cache\/app1&#13;\nCacheDefaultExpire 3600&#13;\nCacheDisable \/wp-admin<\/dfn>&#13;\nCacheIgnoreNoLastMod On<\/pre>\n<\/li>\n<\/ol><h2>Prevent Browsers from Caching Content<\/h2>\n<p>The problem with the above caching configurations is your application\u2019s contents are cached in two locations: on the client and on the server. If you update your content and want the user to see it immediately, they may not be able to if their locally cached version hasn\u2019t expired. You can shorten the expiration time, but then you may defeat the purpose of caching entirely.<\/p>\n<p>Take full control of caching, instead, by enabling it on the server only. Force the web browser to obtain the content from the server for every request, but serve preprocessed Java or PHP content, for example, unless the content has been modified. This ensures the user always sees the most recent version, while not wasting CPU cycles or RAM reprocessing unnecessarilly.<\/p>\n<ol><li>Add the\u00a0<strong>CacheIgnoreCacheControl<\/strong>\u00a0option to the location have caching enabled \u2013 Apache\u2019s server configuration file or a virtual host\u2019s configuration \u2013 with a value of On . This causes Apache to ignore browser content refresh requests. All content will be served from the server\u2019s cache, where possible, until it has expired.\n<pre><dfn>CacheEnable disk \/&#13;\nCacheRoot \/webapps\/cache\/app1&#13;\nCacheDefaultExpire 3600&#13;\nCacheDisable \/wp-admin&#13;\nCacheIgnoreNoLastMod On<\/dfn>&#13;\nCacheIgnoreCacheControl On<\/pre>\n<\/li>\n<li>Save your changes.<\/li>\n<li>Restart Apache to apply your changes.<\/li>\n<\/ol><h2>WordPress and Other Application Frameworks<\/h2>\n<p>Frameworks, like WordPress\u2019 and ones similar to CodeIgniter, route all content through Index.php. Depending on how you\u2019ve written your mod_rewrite rules, your cached content will display incorrectly. You may find that contents for one requested page, category, etc, show up in place of what was actually requested.<\/p>\n<p>To combat this, we need to ensure Apache\u2019s cache considers the parameters appended to the end of the index.php when content is cached and pulled. Modify the Rewrite rule for index.php, as seen in the example below.<\/p>\n<pre># WordPress Permalink rewrites&#13;\nRewriteBase \/&#13;\nRewriteRule ^index.php$ \u2013 [L]&#13;\nRewriteCond %{REQUEST_FILENAME} !-f&#13;\nRewriteCond %{REQUEST_FILENAME} !-d&#13;\n<strong>RewriteRule ^(.*)$ \/index.php\/$1 [L]<\/strong><\/pre>","created_at":"2018-11-07T16:22:08+0000","updated_at":"2018-11-07T16:22:08+0000","published_at":"2014-01-30T15:44:53+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":6,"domain_name":"www.serverlab.ca","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":48,"title":"Structure and Interpretation of Computer Programs, 2e: Top","url":"https:\/\/sarabander.github.io\/sicp\/html\/index.xhtml","content":"Structure and Interpretation of Computer Programs, 2e: Top\n<section><a href=\"#pagetop\" accesskey=\"t\">\u21e1<\/a><p><abbr>Second Edition<\/abbr><br \/>Unofficial Texinfo Format 2.andresraba6.6<\/p><strong>Harold Abelson and Gerald Jay Sussman with Julie Sussman, foreword by Alan J. Perlis<\/strong><p>\u00a91996 by The Massachusetts Institute of Technology\n<\/p><p>This work is licensed under a Creative Commons Attribution-ShareAlike 4.0\nInternational License (<a href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/4.0\/\"><abbr>CC BY-SA 4.0<\/abbr><\/a>).\nBased on a work at <a href=\"http:\/\/mitpress.mit.edu\/sicp\/\">mitpress.mit.edu<\/a>.\n<\/p><p>The <abbr>MIT<\/abbr> Press<br \/>Cambridge, Massachusetts<br \/>London, England\n<\/p><p>McGraw-Hill Book Company<br \/>New York, St. Louis, San Francisco,<br \/>Montreal, Toronto\n<\/p><p>This book is one of a series of texts written by faculty of the Electrical\nEngineering and Computer Science Department at the Massachusetts Institute of\nTechnology.  It was edited and produced by The <abbr>MIT<\/abbr> Press under a\njoint production-distribution arrangement with the McGraw-Hill Book Company.\n<\/p><p>Unofficial Texinfo Format <a href=\"http:\/\/sicpebook.wordpress.com\">2.andresraba6.6<\/a> (September 16, 2015),\nbased on <a href=\"http:\/\/www.neilvandyke.org\/sicp-texi\/\">2.neilvandyke4<\/a> (January 10, 2007).\n<\/p><div class=\"shortcontents\"><ul class=\"no-bullet\"><li><a id=\"stoc-Unofficial-Texinfo-Format\" href=\"#toc-Unofficial-Texinfo-Format\">Unofficial Texinfo Format<\/a><\/li>\n<li><a id=\"stoc-Dedication-1\" href=\"#toc-Dedication-1\">Dedication<\/a><\/li>\n<li><a id=\"stoc-Foreword-1\" href=\"#toc-Foreword-1\">Foreword<\/a><\/li>\n<li><a id=\"stoc-Preface-to-the-Second-Edition\" href=\"#toc-Preface-to-the-Second-Edition\">Preface to the Second Edition<\/a><\/li>\n<li><a id=\"stoc-Preface-to-the-First-Edition\" href=\"#toc-Preface-to-the-First-Edition\">Preface to the First Edition<\/a><\/li>\n<li><a id=\"stoc-Acknowledgments-1\" href=\"#toc-Acknowledgments-1\">Acknowledgments<\/a><\/li>\n<li><a id=\"stoc-Building-Abstractions-with-Procedures\" href=\"#toc-Building-Abstractions-with-Procedures\">1 Building Abstractions with Procedures<\/a><\/li>\n<li><a id=\"stoc-Building-Abstractions-with-Data\" href=\"#toc-Building-Abstractions-with-Data\">2 Building Abstractions with Data<\/a><\/li>\n<li><a id=\"stoc-Modularity_002c-Objects_002c-and-State\" href=\"#toc-Modularity_002c-Objects_002c-and-State\">3 Modularity, Objects, and State<\/a><\/li>\n<li>\n<\/li><li><a id=\"stoc-Computing-with-Register-Machines\" href=\"#toc-Computing-with-Register-Machines\">5 Computing with Register Machines<\/a><\/li>\n<li><a id=\"stoc-References-1\" href=\"#toc-References-1\">References<\/a><\/li>\n<li><a id=\"stoc-List-of-Exercises\" href=\"#toc-List-of-Exercises\">List of Exercises<\/a><\/li>\n<li><a id=\"stoc-List-of-Figures\" href=\"#toc-List-of-Figures\">List of Figures<\/a><\/li>\n<li><a id=\"stoc-Term-Index-1\" href=\"#toc-Term-Index-1\">Term Index<\/a><\/li>\n<li><a id=\"stoc-Colophon-1\" href=\"#toc-Colophon-1\">Colophon<\/a><\/li>\n<\/ul><\/div><\/section><a href=\"#pagebottom\" accesskey=\"b\">\u21e3<\/a>","created_at":"2018-11-07T16:22:08+0000","updated_at":"2018-11-07T16:22:08+0000","annotations":[],"mimetype":"application\/xhtml+xml","language":"en","reading_time":0,"domain_name":"sarabander.github.io","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":45,"title":"Brave a very 80s apocalyptic wasteland in Doom conversion Ashes 2063","url":"https:\/\/www.rockpapershotgun.com\/2018\/09\/24\/ashes-2063-post-apoc-doom-mod-launch\/","content":"<div><main><p>Leather-clad raiders with neon green mohawks roam irradiated monuments to capitalism, their acid-spitting lizard-dogs in tow. <a href=\"https:\/\/www.rockpapershotgun.com\/game\/ashes-2063\/\">Ashes 2063<\/a>, by modder Vostyok is about as 80s an apocalypse can get, even after filtering it through a 2018 iteration of a 90s game engine. It\u2019s a free post-apocalyptic FPS (or total conversion mod, if you\u2019re old-school) built on the ever-popular <a href=\"https:\/\/zdoom.org\/index\">GZDoom<\/a> engine.\u00a0<a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=47573\">Its first episode launched yesterday<\/a>.<\/p>\n<p>While 98% new, Ashes 2063 still requires some manner of Doom data file to fill the gaps. Fortunately, the (free) <a href=\"https:\/\/freedoom.github.io\/\">Freedoom<\/a> will suffice, if you don\u2019t own <a href=\"https:\/\/www.rockpapershotgun.com\/game\/doom-ii\/\">Doom 2<\/a>. Below, a launch trailer.<\/p>\n<p>The game is still in development, and has been in the works since at least early 2015, but this first episode and its nine levels should be plenty to chew on for now, out just in time for Doom\u2019s 25th anniversary. Ashes 2063 might be using Doom\u2019s engine, but it feels a lot more like Duke Nukem 3D, with inventory management, some NPC interaction, trading and even a motorcycle to putter around its larger maps on. Combat feels tight, especially in its ammo distribution, so when you\u2019re up close you might want to try just stabbing some bad dudes with your alt-fire knife attack.<\/p>\n<p>Ashes 2063 is one of the big projects I was hoping would release in time for Doom\u2019s 25th anniversary \u2013 the original shareware release happening on December 10th, 1993. The latest official entry in the series \u2013 <a href=\"https:\/\/www.rockpapershotgun.com\/game\/doom-eternal\/\">Doom Eternal<\/a> \u2013 won\u2019t be out until next year, so let\u2019s celebrate by keeping up with the community. There\u2019s been a ton of stuff released in 2018 alone, much of it by creators younger than the game itself. I wonder if Doom\u2019s 30th will be as momentous, and feature even more fresh blood?<\/p>\n<p>You can find Ashes 2063 <a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=47573\">here on the GZDoom forums<\/a>, although the episode one download is all the way down <a href=\"https:\/\/forum.zdoom.org\/viewtopic.php?f=19&amp;t=47573&amp;start=795#p1073682\">on Page 54 here<\/a>. You\u2019ll need the <a href=\"https:\/\/zdoom.org\/downloads\">GZDoom<\/a> engine and either Doom 2 or <a href=\"https:\/\/freedoom.github.io\/\">Freedoom<\/a>\u2018s WAD file (you\u2019ll probably want Freedoom: Phase 2) to fill in the blanks.<\/p>\n<p><strong>PSA:<\/strong> <em>Friends don\u2019t let friends play GZDoom mods with blurry textures. Follow <a href=\"https:\/\/www.youtube.com\/watch?v=UnhrPeNoQfA\">this little guide video<\/a> to get those pixels pin-sharp in just 30 seconds or your money back.<\/em><\/p>\n\t\t\t\t\t\t\t\t\t<\/main><\/div><div class=\"container-690\"><p class=\"disclaimer\">\n\t\t\t\t\tIf you click our links to online stores and make a purchase we may receive a few pennies. Find more information <a href=\"https:\/\/www.rockpapershotgun.com\/affiliate-linking-policy\/\">here<\/a>.\t\t\t\t<\/p><a href=\"https:\/\/www.rockpapershotgun.com\/support-us?utm_source=article-promo&amp;utm_medium=web&amp;utm_campaign=supporter\" class=\"support-us-promo\" rel=\"nofollow\">\n\t\t\n\t\t\t<img src=\"https:\/\/www.rockpapershotgun.com\/wp-content\/themes\/rps-2018\/images\/logo-inverted.svg\" alt=\"Rock Paper Shotgun\" \/>\n\t\t\n\t\t\n\t\t\t\n\t\t\t\tSupport RPS and get an ad-free site, extra articles, and free stuff!\t\t\t\n\t\t\tTell me more\n\t\t\n\t\t\n\t<\/a><\/div>","created_at":"2018-11-07T16:22:07+0000","updated_at":"2018-11-07T16:22:07+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":2,"domain_name":"www.rockpapershotgun.com","preview_picture":"https:\/\/assets.rockpapershotgun.com\/images\/2018\/09\/Screenshot_Doom_20180924_192942.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":44,"title":"mbsync - synchronize IMAP4 and Maildir mailboxes - Linux Man Pages (1)","url":"https:\/\/www.systutorials.com\/docs\/linux\/man\/docs\/linux\/man\/1-mbsync\/","content":"<div class=\"main-content\"><h2>NAME<\/h2>mbsync - synchronize IMAP4 and Maildir mailboxes<h2>SYNOPSIS<\/h2><b>mbsync<\/b>[<i>options<\/i>...] {{<i>channel<\/i>[<b>:<\/b><i>box<\/i>[{<b>,<\/b>|<b>\\n<\/b>}...]]|<i>group<\/i>} ...|<b>-a<\/b>}<h2>DESCRIPTION<\/h2><b>mbsync<\/b>is a command line application which synchronizes mailboxes;\ncurrently Maildir and IMAP4 mailboxes are supported.\nNew messages, message deletions and flag changes can be propagated both ways;\nthe operation set can be selected in a fine-grained manner.<br \/>Synchronization is based on unique message identifiers (UIDs), so no\nidentification conflicts can occur (as opposed to some other mail synchronizers).\nOTOH,<b>mbsync<\/b>is susceptible to UID validity changes (that<i>should<\/i>never happen, but see \"Compatibility\" in the README).\nSynchronization state is kept in one local text file per mailbox pair;\nmultiple replicas of a mailbox can be maintained.<h2>OPTIONS<\/h2><dl style=\"font-size: 85%;\"><dt><b>-c<\/b>, <b>--config<\/b> <i>file<\/i><\/dt><dd>\nRead configuration from <i>file<\/i>.\nBy default, the configuration is read from ~\/.mbsyncrc.\n<\/dd><dt><b>-a<\/b>, <b>--all<\/b><\/dt><dd>\nSelect all configured channels. Any channel\/group specifications on the command\nline are ignored.\n<\/dd><dt><b>-l<\/b>, <b>--list<\/b><\/dt><dd>\nDon't synchronize anything, but list all mailboxes in the selected channels\nand exit.\n<\/dd><dt><b>-C<\/b>[<b>m<\/b>][<b>s<\/b>], <b>--create<\/b>[<b>-master<\/b>|<b>-slave<\/b>]<\/dt><dd>\nOverride any <b>Create<\/b> options from the config file. See below.\n<\/dd><dt><b>-R<\/b>[<b>m<\/b>][<b>s<\/b>], <b>--remove<\/b>[<b>-master<\/b>|<b>-slave<\/b>]<\/dt><dd>\nOverride any <b>Remove<\/b> options from the config file. See below.\n<\/dd><dt><b>-X<\/b>[<b>m<\/b>][<b>s<\/b>], <b>--expunge<\/b>[<b>-master<\/b>|<b>-slave<\/b>]<\/dt><dd>\nOverride any <b>Expunge<\/b> options from the config file. See below.\n<\/dd><dt>{<b>-n<\/b>|<b>-N<\/b>|<b>-d<\/b>|<b>-f<\/b>|<b>-0<\/b>|<b>-F<\/b>},<br \/>\u00a0{<b>--new<\/b>|<b>--renew<\/b>|<b>--delete<\/b>|<b>--flags<\/b>|<b>--noop<\/b>|<b>--full<\/b>}<\/dt><dd>\n<\/dd><dt>{<b>-L<\/b>|<b>-H<\/b>}[<b>n<\/b>][<b>N<\/b>][<b>d<\/b>][<b>f<\/b>],<br \/>\u00a0{<b>--pull<\/b>|<b>--push<\/b>}[<b>-new<\/b>|<b>-renew<\/b>|<b>-delete<\/b>|<b>-flags<\/b>]<\/dt><dd>\nOverride any <b>Sync<\/b> options from the config file. See below.\n<\/dd><dt><b>-h<\/b>, <b>--help<\/b><\/dt><dd>\nDisplay a summary of command line options.\n<\/dd><dt><b>-v<\/b>, <b>--version<\/b><\/dt><dd>\nDisplay version information.\n<\/dd><dt><b>-V<\/b>, <b>--verbose<\/b><\/dt><dd>\nEnable <i>verbose<\/i> mode, which displays what is currently happening.\n<\/dd><dt><b>-D<\/b>[<b>C<\/b>][<b>m<\/b>][<b>M<\/b>][<b>n<\/b>|<b>N<\/b>][<b>s<\/b>]],<br \/>\u00a0<b>--debug<\/b>[<b>-crash<\/b>|<b>-maildir<\/b>|<b>-main<\/b>|<b>-net<\/b>|<b>-net-all<\/b>|<b>-sync<\/b>]<\/dt><dd>\nEnable debugging categories:\n<b>C<\/b>, <b>crash<\/b> - use built-in crash handler\n<br \/><b>m<\/b>, <b>maildir<\/b> - print maildir debug info\n<br \/><b>M<\/b>, <b>main<\/b> - print main debug info\n<br \/><b>n<\/b>, <b>net<\/b> - print network traffic (protocol only)\n<br \/><b>N<\/b>, <b>net-all<\/b> - print network traffic (including payloads)\n<br \/><b>s<\/b>, <b>sync<\/b> - print synchronization debug info\nAll categories except <b>crash<\/b> implictly enable <i>verbose<\/i> mode.\nWithout category specification, all categories except net-all are enabled.\n<\/dd><dt><b>-q<\/b>, <b>--quiet<\/b><\/dt><dd>\nSuppress progress counters (this is implicit if stdout is no TTY,\nor any debugging categories are enabled) and notices.\nIf specified twice, suppress warning messages as well.\n<\/dd><\/dl><h2>CONFIGURATION<\/h2>The configuration file is mandatory;<b>mbsync<\/b>will not run without it.\nLines starting with a hash mark (<b>#<\/b>) are comments and are ignored entirely.\nConfiguration items are keywords followed by one or more arguments;\narguments containing spaces must be enclosed in double quotes (<b>\"<\/b>),\nand literal double quotes and backslashes (<b>\\<\/b>) must be backslash-escaped.\nAll keywords (including those used as arguments) are case-insensitive.\nBash-like home directory expansion using the tilde (<b>~<\/b>) is supported\nin all options which represent local paths.\nThere are a few global options, the rest applies to particular sections.\nSections are started by a section keyword and are terminated by an empty line\nor end of file.\nEvery section defines an object with an identifier unique within that\nobject class.<p>\nThere are two basic object classes: Stores and Channels. A Store defines\na collection of mailboxes; basically a folder, either local or remote.\nA Channel connects two Stores, describing the way the two are synchronized.\n<br \/>There are two auxiliary object classes: Accounts and Groups. An Account\ndescribes the connection part of remote Stores, so a server connection can be\nshared between multiple Stores. A Group aggregates multiple Channels to\nsave typing on the command line.\n<\/p><p>\nFile system locations (in particular, <b>Path<\/b> and <b>Inbox<\/b>) use the\nStore's internal path separators, which may be slashes, periods, etc., or\neven combinations thereof.\n<br \/>Mailbox names, OTOH, always use canonical path separators, which are\nUnix-like forward slashes.\n<\/p><h3>All Stores<\/h3>These options can be used in all supported Store types.<br \/>In this context, the term \"remote\" describes the second Store within a Channel,\nand not necessarily a remote server.<br \/>The special mailbox<b>INBOX<\/b>exists in every Store; its physical location\nin the file system is Store type specific.<dl style=\"font-size: 85%;\"><dt><b>Path<\/b> <i>path<\/i><\/dt><dd>\nThe location of the Store in the (server's) file system.\nIf this is no absolute path, the reference point is Store type specific.\nThis string is prepended to the mailbox names addressed in this Store,\nbut is not considered part of them; this is important for <b>Patterns<\/b>\nin the Channels section.\nNote that you <b>must<\/b> append a slash if you want to specify an entire\ndirectory.\n(Default: none)\n<\/dd><dt><b>MaxSize<\/b> <i>size<\/i>[<b>k<\/b>|<b>m<\/b>][<b>b<\/b>]<\/dt><dd>\nMessages larger than that will not be propagated into this Store.\nThis is useful for weeding out messages with large attachments.\n<b>K<\/b> and <b>M<\/b> can be appended to the size to specify KiBytes resp.\nMeBytes instead of bytes. <b>B<\/b> is accepted but superfluous.\nIf <i>size<\/i> is 0, the maximum message size is <b>unlimited<\/b>.\n(Default: <i>0<\/i>)\n<\/dd><dt><b>MapInbox<\/b> <i>mailbox<\/i><\/dt><dd>\nCreate a virtual mailbox (relative to <b>Path<\/b>) which aliases\nthe <b>INBOX<\/b>. Makes sense in conjunction with <b>Patterns<\/b> in the\nChannels section, though with a Maildir slave, you probably want to\nplace <b>Inbox<\/b> under <b>Path<\/b> instead.\nThis virtual mailbox does not support subfolders.\n<\/dd><dt><b>Flatten<\/b> <i>delim<\/i><\/dt><dd>\nFlatten the hierarchy within this Store by substituting the canonical\nhierarchy delimiter <b>\/<\/b> with <i>delim<\/i>.\nThis can be useful when the MUA used to access the Store provides\nsuboptimal handling of hierarchical mailboxes, as is the case with\n<b>Mutt<\/b>.\nA common choice for the delimiter is <b>.<\/b>.\n<br \/>Note that flattened sub-folders of the <b>INBOX<\/b> always end up\nunder <b>Path<\/b>, including the \"INBOX<i>delim<\/i>\" prefix.\n<\/dd><dt><b>Trash<\/b> <i>mailbox<\/i><\/dt><dd>\nSpecifies a mailbox (relative to <b>Path<\/b>) to copy deleted messages to\nprior to expunging.\nSee <b>RECOMMENDATIONS<\/b> and <b>INHERENT PROBLEMS<\/b> below.\n(Default: none)\n<\/dd><dt><b>TrashNewOnly<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\nWhen trashing, copy only not yet propagated messages. This makes sense if the\nremote Store has a <b>Trash<\/b> as well (with <b>TrashNewOnly<\/b> <i>no<\/i>).\n(Default: <i>no<\/i>)\n<\/dd><dt><b>TrashRemoteNew<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\nWhen expunging the remote Store, copy not yet propagated messages to this\nStore's <b>Trash<\/b>. When using this, the remote Store does not need an own\n<b>Trash<\/b> at all, yet all messages are archived.\n(Default: <i>no<\/i>)\n<\/dd><\/dl><h3>Maildir Stores<\/h3>The reference point for relative<b>Path<\/b>s is the current working directory.<p>\nAs <b>mbsync<\/b> needs UIDs, but no standardized UID storage scheme exists for\nMaildir, <b>mbsync<\/b> supports two schemes, each with its pros and cons.\n<br \/>The <b>native<\/b> scheme is stolen from the latest Maildir patches to <b>c-client<\/b>\nand is therefore compatible with <b>pine<\/b>. The UID validity is stored in a\nfile named .uidvalidity; the UIDs are encoded in the file names of the messages.\n<br \/>The <b>alternative<\/b> scheme is based on the UID mapping used by <b>isync<\/b>\nversions 0.8 and 0.9.x. The invariant parts of the file names of the messages\nare used as keys into a Berkeley database named .isyncuidmap.db, which holds\nthe UID validity as well.\n<br \/>The <b>native<\/b> scheme is faster, more space efficient, endianness independent\nand \"human readable\", but will be disrupted if a message is copied from another\nmailbox without getting a new file name; this would result in duplicated UIDs\nsooner or later, which in turn results in a UID validity change, making\nsynchronization fail.\nThe <b>alternative<\/b> scheme would fail if a MUA changed a message's file name\nin a part <b>mbsync<\/b> considers invariant; this would be interpreted as a\nmessage deletion and a new message, resulting in unnecessary traffic.\n<br \/><b>Mutt<\/b> is known to work fine with both schemes.\n<br \/>Use <b>mdconvert<\/b> to convert mailboxes from one scheme to the other.\n<\/p><dl style=\"font-size: 85%;\"><dt><b>MaildirStore<\/b> <i>name<\/i><\/dt><dd>\nDefine the Maildir Store <i>name<\/i>, opening a section for its parameters.\n<\/dd><dt><b>AltMap<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\nUse the <b>alternative<\/b> UID storage scheme for mailboxes in this Store.\nThis does not affect mailboxes that do already have a UID storage scheme;\nuse <b>mdconvert<\/b> to change it.\n(Default: <i>no<\/i>)\n<\/dd><dt><b>Inbox<\/b> <i>path<\/i><\/dt><dd>\nThe location of the <b>INBOX<\/b>. This is <i>not<\/i> relative to <b>Path<\/b>,\nbut it is allowed to place the <b>INBOX<\/b> inside the <b>Path<\/b>.\n(Default: <i>~\/Maildir<\/i>)\n<\/dd><dt><b>InfoDelimiter<\/b> <i>delim<\/i><\/dt><dd>\nThe character used to delimit the info field from a message's basename.\nThe Maildir standard defines this to be the colon, but this is incompatible\nwith DOS\/Windows file systems.\n(Default: the value of <b>FieldDelimiter<\/b>)\n<\/dd><\/dl><h3>IMAP4 Accounts<\/h3><dl style=\"font-size: 85%;\"><dt><b>IMAPAccount<\/b> <i>name<\/i><\/dt><dd>\nDefine the IMAP4 Account <i>name<\/i>, opening a section for its parameters.\n<\/dd><dt><b>Host<\/b> <i>host<\/i><\/dt><dd>\nSpecify the DNS name or IP address of the IMAP server.\n<br \/>If <b>Tunnel<\/b> is used, this setting is needed only if <b>SSLType<\/b> is\nnot <i>None<\/i> and <b>CertificateFile<\/b> is not used,\nin which case the host name is used for certificate subject verification.\n<\/dd><dt><b>Port<\/b> <i>port<\/i><\/dt><dd>\nSpecify the TCP port number of the IMAP server.  (Default: 143 for IMAP,\n993 for IMAPS)\n<br \/>If <b>Tunnel<\/b> is used, this setting is ignored.\n<\/dd><dt><b>User<\/b> <i>username<\/i><\/dt><dd>\nSpecify the login name on the IMAP server.\n<\/dd><dt><b>Pass<\/b> <i>password<\/i><\/dt><dd>\nSpecify the password for <i>username<\/i> on the IMAP server.\nNote that this option is <b>NOT<\/b> required.\nIf neither a password nor a password command is specified in the\nconfiguration file, <b>mbsync<\/b> will prompt you for a password.\n<\/dd><dt><b>PassCmd<\/b> [<b>+<\/b>]<i>command<\/i><\/dt><dd>\nSpecify a shell command to obtain a password rather than specifying a\npassword directly. This allows you to use password files and agents.\nThe command must produce exactly one line on stdout; the trailing newline is\noptional.\nPrepend <b>+<\/b> to the command to indicate that it produces TTY output\n(e.g., a decryption password prompt); failure to do so will merely produce\nmessier output.\n<\/dd><dt><b>Tunnel<\/b> <i>command<\/i><\/dt><dd>\nSpecify a command to run to establish a connection rather than opening a TCP\nsocket.  This allows you to run an IMAP session over an SSH tunnel, for\nexample.\n<\/dd><dt><b>AuthMechs<\/b> <i>type<\/i> ...<\/dt><dd>\nThe list of acceptable authentication mechanisms.\nIn addition to the mechanisms listed in the SASL registry (link below),\nthe legacy IMAP <b>LOGIN<\/b> mechanism is known.\nThe wildcard <b>*<\/b> represents all mechanisms that are deemed secure\nenough for the current <b>SSLType<\/b> setting.\nThe actually used mechanism is the most secure choice from the intersection\nof this list, the list supplied by the server, and the installed SASL modules.\n(Default: <b>*<\/b>)\n<\/dd><dt><b>SSLType<\/b> {<i>None<\/i>|<i>STARTTLS<\/i>|<i>IMAPS<\/i>}<\/dt><dd>\nSelect the connection security\/encryption method:\n<br \/><i>None<\/i> - no security.\nThis is the default when <b>Tunnel<\/b> is set, as tunnels are usually secure.\n<br \/><i>STARTTLS<\/i> - security is established via the STARTTLS extension\nafter connecting the regular IMAP port 143. Most servers support this,\nso it is the default (unless a tunnel is used).\n<br \/><i>IMAPS<\/i> - security is established by starting SSL\/TLS negotiation\nright after connecting the secure IMAP port 993.\n<\/dd><dt><b>SSLVersions<\/b> [<i>SSLv2<\/i>] [<i>SSLv3<\/i>] [<i>TLSv1<\/i>] [<i>TLSv1.1<\/i>] [<i>TLSv1.2<\/i>]<\/dt><dd>\nSelect the acceptable SSL\/TLS versions.\nUse of SSLv2 is strongly discouraged for security reasons, but might be the\nonly option on some very old servers.\nGenerally, the newest TLS version is recommended, but as this confuses some\nservers, <b>TLSv1<\/b> is the default.\n<\/dd><dt><b>SystemCertificates<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\nWhether the system's default root cerificate store should be loaded.\n(Default: <i>yes<\/i>)\n<\/dd><dt><b>CertificateFile<\/b> <i>path<\/i><\/dt><dd>\nFile containing additional X.509 certificates used to verify server\nidentities. Directly matched peer certificates are always trusted,\nregardless of validity.\n<br \/>Note that the system's default certificate store is always used\n(unless <b>SystemCertificates<\/b> is disabled)\nand should not be specified here.\n<\/dd><dt><b>PipelineDepth<\/b> <i>depth<\/i><\/dt><dd>\nMaximum number of IMAP commands which can be simultaneously in flight.\nSetting this to <i>1<\/i> disables pipelining.\nThis is mostly a debugging only option.\n(Default: <i>unlimited<\/i>)\n<\/dd><\/dl><h3>IMAP Stores<\/h3>The reference point for relative<b>Path<\/b>s is whatever the server likes it\nto be; probably the user's $HOME or $HOME\/Mail on that server. The location\nof<b>INBOX<\/b>is up to the server as well and is usually irrelevant.<dl style=\"font-size: 85%;\"><dt><b>IMAPStore<\/b> <i>name<\/i><\/dt><dd>\nDefine the IMAP4 Store <i>name<\/i>, opening a section for its parameters.\n<\/dd><dt><b>Account<\/b> <i>account<\/i><\/dt><dd>\nSpecify which IMAP4 Account to use. Instead of defining an Account and\nreferencing it here, it is also possible to specify all the Account options\ndirectly in the Store's section - this makes sense if an Account is used for\none Store only anyway.\n<\/dd><dt><b>UseNamespace<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\nSelects whether the server's first \"personal\" NAMESPACE should be prefixed to\nmailbox names. Disabling this makes sense for some broken IMAP servers.\nThis option is meaningless if a <b>Path<\/b> was specified.\n(Default: <i>yes<\/i>)\n<\/dd><dt><b>PathDelimiter<\/b> <i>delim<\/i><\/dt><dd>\nSpecify the server's hierarchy delimiter.\n(Default: taken from the server's first \"personal\" NAMESPACE)\n<br \/>Do <b>NOT<\/b> abuse this to re-interpret the hierarchy.\nUse <b>Flatten<\/b> instead.\n<\/dd><\/dl><h3>Channels<\/h3><dl style=\"font-size: 85%;\"><dt><b>Channel<\/b> <i>name<\/i><\/dt><dd>\nDefine the Channel <i>name<\/i>, opening a section for its parameters.\n<\/dd><dt>{<b>Master<\/b>|<b>Slave<\/b>} <b>:<\/b><i>store<\/i><b>:<\/b>[<i>mailbox<\/i>]<\/dt><dd>\nSpecify the Master resp. Slave Store to be connected by this Channel.\nIf <b>Patterns<\/b> are specified, <i>mailbox<\/i> is interpreted as a\nprefix which is not matched against the patterns, and which is not\naffected by mailbox list overrides.\nOtherwise, if <i>mailbox<\/i> is omitted, <b>INBOX<\/b> is assumed.\n<\/dd><dt><b>Pattern<\/b>[<b>s<\/b>] [<b>!<\/b>]<i>pattern<\/i> ...<\/dt><dd>\nInstead of synchronizing only one mailbox pair, synchronize all mailboxes\nthat match the <i>pattern<\/i>(s). The mailbox names are the same on both\nMaster and Slave. Patterns are IMAP4 patterns, i.e., <b>*<\/b> matches anything\nand <b>%<\/b> matches anything up to the next hierarchy delimiter. Prepending\n<b>!<\/b> to a pattern makes it an exclusion. Multiple patterns can be specified\n(either by supplying multiple arguments or by using <b>Pattern<\/b> multiple\ntimes); later matches take precedence.\n<br \/>Note that <b>INBOX<\/b> is not matched by wildcards, unless it lives under\n<b>Path<\/b>.\n<br \/>The mailbox list selected by <b>Patterns<\/b> can be overridden by a mailbox\nlist in a channel reference (a <b>Group<\/b> specification or the command line).\n<br \/>Example: \"<b>Patterns<\/b>\u00a0<i>%\u00a0!Trash<\/i>\"\n<\/dd><dt><b>MaxSize<\/b> <i>size<\/i>[<b>k<\/b>|<b>m<\/b>][<b>b<\/b>]<\/dt><dd>\nAnalogous to the homonymous option in the Stores section, but applies equally\nto Master and Slave. Note that this actually modifies the Stores, so take care\nnot to provide conflicting settings if you use the Stores in multiple Channels.\n<\/dd><dt><b>MaxMessages<\/b> <i>count<\/i><\/dt><dd>\nSets the maximum number of messages to keep in each Slave mailbox.\nThis is useful for mailboxes where you keep a complete archive on the server,\nbut want to mirror only the last messages (for instance, for mailing lists).\nThe messages that were the first to arrive in the mailbox (independently of\nthe actual date of the message) will be deleted first.\nMessages that are flagged (marked as important) and (by default) unread\nmessages will not be automatically deleted.\nIf <i>count<\/i> is 0, the maximum number of messages is <b>unlimited<\/b>\n(Default: <i>0<\/i>).\n<\/dd><dt><b>ExpireUnread<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\nSelects whether unread messages should be affected by <b>MaxMessages<\/b>.\nNormally, unread messages are considered important and thus never expired.\nThis ensures that you never miss new messages even after an extended absence.\nHowever, if your archive contains large amounts of unread messages by design,\ntreating them as important would practically defeat <b>MaxMessages<\/b>. In this\ncase you need to enable this option.\n(Default: <i>no<\/i>).\n<\/dd><dt><b>Sync<\/b> {<i>None<\/i>|[<i>Pull<\/i>] [<i>Push<\/i>] [<i>New<\/i>] [<i>ReNew<\/i>] [<i>Delete<\/i>] [<i>Flags<\/i>]|<i>All<\/i>}<\/dt><dd>\nSelect the synchronization operation(s) to perform:\n<br \/><i>Pull<\/i> - propagate changes from Master to Slave.\n<br \/><i>Push<\/i> - propagate changes from Slave to Master.\n<br \/><i>New<\/i> - propagate newly appeared messages.\n<br \/><i>ReNew<\/i> - previously refused messages are re-evaluated for propagation.\nUseful after flagging affected messages in the source Store or enlarging\nMaxSize in the destination Store.\n<br \/><i>Delete<\/i> - propagate message deletions. This applies only to messages that\nare actually gone, i.e., were expunged. The affected messages in the remote\nStore are marked as deleted only, i.e., they won't be really deleted until\nthat Store is expunged.\n<br \/><i>Flags<\/i> - propagate flag changes. Note that Deleted\/Trashed is a flag as\nwell; this is particularly interesting if you use <b>mutt<\/b> with the\nmaildir_trash option.\n<br \/><i>All<\/i> (<b>--full<\/b> on the command line) - all of the above.\nThis is the global default.\n<br \/><i>None<\/i> (<b>--noop<\/b> on the command line) - don't propagate anything.\nUseful if you want to expunge only.\n<\/dd><dt><\/dt><dd>\n<i>Pull<\/i> and <i>Push<\/i> are direction flags, while <i>New<\/i>, <i>ReNew<\/i>,\n<i>Delete<\/i> and <i>Flags<\/i> are type flags. The two flag classes make up a\ntwo-dimensional matrix (a table). Its cells are the individual actions to\nperform. There are two styles of asserting the cells:\n<br \/>In the first style, the flags select entire rows\/colums in the matrix. Only\nthe cells which are selected both horizontally and vertically are asserted.\nSpecifying no flags from a class is like specifying all flags from this class.\nFor example, \"<b>Sync<\/b>\u00a0<i>Pull<\/i>\u00a0<i>New<\/i>\u00a0<i>Flags<\/i>\" will propagate\nnew messages and flag changes from the Master to the Slave,\n\"<b>Sync<\/b>\u00a0<i>New<\/i>\u00a0<i>Delete<\/i>\" will propagate message arrivals and\ndeletions both ways, and \"<b>Sync<\/b>\u00a0<i>Push<\/i>\" will propagate all changes\nfrom the Slave to the Master.\n<br \/>In the second style, direction flags are concatenated with type flags; every\ncompound flag immediately asserts a cell in the matrix. In addition to at least\none compound flag, the individual flags can be used as well, but as opposed to\nthe first style, they immediately assert all cells in their respective\nrow\/column. For example,\n\"<b>Sync<\/b>\u00a0<i>PullNew<\/i>\u00a0<i>PullDelete<\/i>\u00a0<i>Push<\/i>\" will propagate\nmessage arrivals and deletions from the Master to the Slave and any changes\nfrom the Slave to the Master.\nNote that it is not allowed to assert a cell in two ways, e.g.\n\"<b>Sync<\/b>\u00a0<i>PullNew<\/i>\u00a0<i>Pull<\/i>\" and\n\"<b>Sync<\/b>\u00a0<i>PullNew<\/i>\u00a0<i>Delete<\/i>\u00a0<i>Push<\/i>\" induce error messages.\n<\/dd><dt><b>Create<\/b> {<i>None<\/i>|<i>Master<\/i>|<i>Slave<\/i>|<i>Both<\/i>}<\/dt><dd>\nAutomatically create missing mailboxes [on the Master\/Slave].\nOtherwise print an error message and skip that mailbox pair if a mailbox\nand the corresponding sync state does not exist.\n(Global default: <i>None<\/i>)\n<\/dd><dt><b>Remove<\/b> {<i>None<\/i>|<i>Master<\/i>|<i>Slave<\/i>|<i>Both<\/i>}<\/dt><dd>\nPropagate mailbox deletions [to the Master\/Slave].\nOtherwise print an error message and skip that mailbox pair if a mailbox\ndoes not exist but the corresponding sync state does.\n<br \/>For MailDir mailboxes it is sufficient to delete the cur\/ subdirectory to\nmark them as deleted. This ensures compatibility with <b>SyncState *<\/b>.\n<br \/>Note that for safety, non-empty mailboxes are never deleted.\n<br \/>(Global default: <i>None<\/i>)\n<\/dd><dt><b>Expunge<\/b> {<i>None<\/i>|<i>Master<\/i>|<i>Slave<\/i>|<i>Both<\/i>}<\/dt><dd>\nPermanently remove all messages [on the Master\/Slave] marked for deletion.\nSee <b>RECOMMENDATIONS<\/b> below.\n(Global default: <i>None<\/i>)\n<\/dd><dt><b>CopyArrivalDate<\/b> {<i>yes<\/i>|<i>no<\/i>}<\/dt><dd>\nSelects whether their arrival time should be propagated together with\nthe messages.\nEnabling this makes sense in order to keep the time stamp based message\nsorting intact.\nNote that IMAP does not guarantee that the time stamp (termed <b>internal\ndate<\/b>) is actually the arrival time, but it is usually close enough.\n(Default: <i>no<\/i>)\n<\/dd><\/dl><p>\n<b>Sync<\/b>, <b>Create<\/b>, <b>Remove<\/b>, <b>Expunge<\/b>,\n<b>MaxMessages<\/b>, and <b>CopyArrivalDate<\/b>\ncan be used before any section for a global effect.\nThe global settings are overridden by Channel-specific options,\nwhich in turn are overridden by command line switches.\n<\/p><dl style=\"font-size: 85%;\"><dt><b>SyncState<\/b> {<b>*<\/b>|<i>path<\/i>}<\/dt><dd>\nSet the location of this Channel's synchronization state files. <b>*<\/b> means\nthat the state should be saved in a file named .mbsyncstate in the\nSlave mailbox itself; this has the advantage that you needn't to care for the\nstate file if you delete the mailbox, but it works only with Maildir mailboxes,\nobviously. Otherwise this is interpreted as a string to prepend to the Slave\nmailbox name to make up a complete path.\n<br \/>This option can be used outside any section for a global effect. In this case\nthe appended string is made up according to the pattern\n<b>:<\/b><i>master<\/i><b>:<\/b><i>master-box<\/i><b>_:<\/b><i>slave<\/i><b>:<\/b><i>slave-box<\/i>\n(see also <b>FieldDelimiter<\/b> below).\n<br \/>(Global default: <i>~\/.mbsync\/<\/i>).\n<\/dd><\/dl><h3>Groups<\/h3><dl style=\"font-size: 85%;\"><dt><b>Group<\/b> <i>name<\/i> [<i>channel<\/i>[<b>:<\/b><i>box<\/i>[<b>,<\/b>...]]] ...<\/dt><dd>\nDefine the Group <i>name<\/i>, opening a section for its parameters.\nNote that even though Groups have an own namespace, they will \"hide\" Channels\nwith the same name on the command line.\n<br \/>One or more Channels can be specified on the same line.\n<br \/>If you supply one or more <i>box<\/i>es to a <i>channel<\/i>, they will be used\ninstead of what is specified in the Channel's Patterns.\nThe same can be done on the command line, except that there newlines can be\nused as mailbox name separators as well.\n<\/dd><dt><b>Channel<\/b>[<b>s<\/b>] <i>channel<\/i>[<b>:<\/b><i>box<\/i>[<b>,<\/b>...]] ...<\/dt><dd>\nAdd the specified channels to the group. This option can be specified multiple\ntimes within a Group.\n<\/dd><\/dl><h3>Global Options<\/h3><dl style=\"font-size: 85%;\"><dt><b>FSync<\/b> <i>yes<\/i>|<i>no<\/i><\/dt><dd>\n<br \/>Selects whether <b>mbsync<\/b> performs forced flushing, which determines\nthe level of data safety after system crashes and power outages.\nDisabling it is reasonably safe for file systems which are mounted with\ndata=ordered mode.\nEnabling it is a wise choice for file systems mounted with data=writeback,\nin particular modern systems like ext4, btrfs and xfs. The performance impact\non older file systems may be disproportionate.\n(Default: <i>yes<\/i>)\n<\/dd><dt><b>FieldDelimiter<\/b> <i>delim<\/i><\/dt><dd>\nThe character to use to delimit fields in the string appended to a global\n<b>SyncState<\/b>.\n<b>mbsync<\/b> prefers to use the colon, but this is incompatible with\nDOS\/Windows file systems.\nThis option is meaningless for <b>SyncState<\/b> if the latter is <b>*<\/b>,\nobviously. However, it also determines the default of <b>InfoDelimiter<\/b>.\n(Global default: <i>;<\/i> on Windows, <i>:<\/i> everywhere else)\n<\/dd><dt><b>BufferLimit<\/b> <i>size<\/i>[<b>k<\/b>|<b>m<\/b>][<b>b<\/b>]<\/dt><dd>\nThe per-Channel, per-direction instantaneous memory usage above which\n<b>mbsync<\/b> will refrain from using more memory. Note that this is no\nabsolute limit, as even a single message can consume more memory than\nthis.\n(Default: <i>10M<\/i>)\n<\/dd><\/dl><h2>CONSOLE OUTPUT<\/h2>If<b>mbsync<\/b>'s output is connected to a console, it will print progress\ncounters by default. The output will look like this:<p>\nC: 1\/2  B: 3\/4  M: +13\/13 *23\/42 #0\/0  S: +0\/7 *0\/0 #0\/0\n<\/p><p>\nThis represents the cumulative progress over channels, boxes, and messages\naffected on master and slave, respectively.\nThe message counts represent added messages, messages with updated flags,\nand trashed messages, respectively.\nNo attempt is made to calculate the totals in advance, so they grow over\ntime as more information is gathered.\n<\/p><h2>RECOMMENDATIONS<\/h2>Make sure your IMAP server does not auto-expunge deleted messages - it is\nslow, and semantically somewhat questionable. Specifically, Gmail needs to\nbe configured not to do it.<p>\nBy default, <b>mbsync<\/b> will not delete any messages - deletions are\npropagated by marking the messages as deleted on the remote store.\nOnce you have verified that your setup works, you will typically want to\nset <b>Expunge<\/b> to <b>Both<\/b>, so that deletions become effective.\n<\/p><p>\n<b>mbsync<\/b>'s built-in trash functionality relies on <b>mbsync<\/b> doing\nthe expunging of deleted messages. This is the case when it propagates\ndeletions of previously propagated messages, and the trash is on the target\nstore (typically your IMAP server).\n<br \/>However, when you intend <b>mbsync<\/b> to trash messages which were not\npropagated yet, the MUA must mark the messages as deleted without expunging\nthem (e.g., <b>Mutt<\/b>'s <b>maildir_trash<\/b> option). Note that most\nmessages are propagated a long time before they are deleted, so this is a\ncorner case you probably do not want to optimize for. This also implies\nthat the <b>TrashNewOnly<\/b> and <b>TrashRemoteNew<\/b> options are typically\nnot very useful.\n<\/p><p>\nIf your server supports auto-trashing (as Gmail does), it is probably a\ngood idea to rely on that instead of <b>mbsync<\/b>'s trash functionality.\nIf you do that, and intend to synchronize the trash like other mailboxes,\nyou should not use <b>mbsync<\/b>'s <b>Trash<\/b> option at all.\n<\/p><h2>INHERENT PROBLEMS<\/h2>Changes done after<b>mbsync<\/b>has retrieved the message list will not be\nsynchronised until the next time<b>mbsync<\/b>is invoked.<p>\nUsing <b>Trash<\/b> on IMAP Stores without the UIDPLUS extension (notably,\nM$ Exchange up to at least 2010) bears a race condition: messages will be\nlost if they are marked as deleted after the message list was retrieved but\nbefore the mailbox is expunged.\nThere is no risk as long as the IMAP mailbox is accessed by only one client\n(including <b>mbsync<\/b>) at a time.\n<\/p><h2>FILES<\/h2><dl style=\"font-size: 85%;\"><dt><b>~\/.mbsyncrc<\/b>\n<\/dt><dd>\nDefault configuration file\n<\/dd><dt><b>~\/.mbsync\/<\/b>\n<\/dt><dd>\nDirectory containing synchronization state files\n<\/dd><\/dl><h2>SEE ALSO<\/h2><a href=\"https:\/\/www.systutorials.com\/docs\/linux\/man\/1-mdconvert\/\">mdconvert<\/a>(1),<a href=\"https:\/\/www.systutorials.com\/docs\/linux\/man\/1-isync\/\">isync<\/a>(1),<a href=\"https:\/\/www.systutorials.com\/docs\/linux\/man\/1-mutt\/\">mutt<\/a>(1),<a href=\"https:\/\/www.systutorials.com\/docs\/linux\/man\/5-maildir\/\">maildir<\/a>(5)<p>\nUp to date information on <b>mbsync<\/b> can be found at <a href=\"http:\/\/isync.sf.net\/\">http:\/\/isync.sf.net\/<\/a>\n<\/p><p>\nSASL mechanisms are listed at\n<a href=\"http:\/\/www.iana.org\/assignments\/sasl-mechanisms\/sasl-mechanisms.xhtml\">http:\/\/www.iana.org\/assignments\/sasl-mechanisms\/sasl-mechanisms.xhtml<\/a>\n<\/p><h2>AUTHORS<\/h2>Originally written by Michael R. Elkins,\nrewritten and currently maintained by Oswald Buddenhagen,\ncontributions by Theodore Y. Ts'o.<hr \/><\/div><p>Linux Man Pages Copyright Respective Owners. Site Copyright \u00a9 <a rel=\"nofollow\" href=\"https:\/\/www.systutorials.com\/\">SysTutorials<\/a>. All Rights Reserved.<\/p>","created_at":"2018-11-07T16:22:06+0000","updated_at":"2018-11-07T16:22:06+0000","annotations":[],"mimetype":"text\/html","reading_time":19,"domain_name":"www.systutorials.com","preview_picture":"http:\/\/www.systutorials.com\/docs\/linux\/man\/content\/tux.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":41,"title":"basecamp\/trix","url":"https:\/\/github.com\/basecamp\/trix","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><h3><a id=\"user-content-a-rich-text-editor-for-everyday-writing\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-rich-text-editor-for-everyday-writing\"><\/a>A Rich Text Editor for Everyday Writing<\/h3>\n<p><strong>Compose beautifully formatted text in your web application.<\/strong> Trix is a WYSIWYG editor for writing messages, comments, articles, and lists\u2014the simple documents most web apps are made of. It features a sophisticated document model, support for embedded attachments, and outputs terse and consistent HTML.<\/p>\n<p>Trix is an open-source project from <a href=\"https:\/\/basecamp.com\/\" rel=\"nofollow\">Basecamp<\/a>, the creators of <a href=\"http:\/\/rubyonrails.org\/\" rel=\"nofollow\">Ruby on Rails<\/a>. Millions of people trust their text to Basecamp, and we built Trix to give them the best possible editing experience. See Trix in action in the <a href=\"https:\/\/basecamp.com\/3-is-coming\" rel=\"nofollow\">all-new Basecamp 3<\/a>.<\/p>\n<h3><a id=\"user-content-different-by-design\" class=\"anchor\" aria-hidden=\"true\" href=\"#different-by-design\"><\/a>Different By Design<\/h3>\n<p>Most WYSIWYG editors are wrappers around HTML\u2019s <code>contenteditable<\/code> and <code>execCommand<\/code> APIs, designed by Microsoft to support live editing of web pages in Internet Explorer 5.5, and <a href=\"https:\/\/blog.whatwg.org\/the-road-to-html-5-contenteditable#history\" rel=\"nofollow\">eventually reverse-engineered<\/a> and copied by other browsers.<\/p>\n<p>Because these APIs were never fully specified or documented, and because WYSIWYG HTML editors are enormous in scope, each browser\u2019s implementation has its own set of bugs and quirks, and JavaScript developers are left to resolve the inconsistencies.<\/p>\n<p>Trix sidesteps these inconsistencies by treating <code>contenteditable<\/code> as an I\/O device: when input makes its way to the editor, Trix converts that input into an editing operation on its internal document model, then re-renders that document back into the editor. This gives Trix complete control over what happens after every keystroke, and avoids the need to use <code>execCommand<\/code> at all.<\/p>\n<h3><a id=\"user-content-built-for-the-modern-web\" class=\"anchor\" aria-hidden=\"true\" href=\"#built-for-the-modern-web\"><\/a>Built for the Modern Web<\/h3>\n<details><summary>Trix supports all evergreen, self-updating desktop and mobile browsers.<\/summary><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/1d4b77d1164fbcce16d6fd49052ea673816c268e\/68747470733a2f2f73332e616d617a6f6e6177732e636f6d2f747269782d6465706f742f746573742d7374617475732d696d616765732f747269782e737667\"><img src=\"https:\/\/camo.githubusercontent.com\/1d4b77d1164fbcce16d6fd49052ea673816c268e\/68747470733a2f2f73332e616d617a6f6e6177732e636f6d2f747269782d6465706f742f746573742d7374617475732d696d616765732f747269782e737667\" data-canonical-src=\"https:\/\/s3.amazonaws.com\/trix-depot\/test-status-images\/trix.svg\" alt=\"image\" \/><\/a><\/details><p>Trix is built with emerging web standards, notably <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/Web_Components\/Using_custom_elements\" rel=\"nofollow\">Custom Elements<\/a>, <a href=\"https:\/\/dom.spec.whatwg.org\/#mutation-observers\" rel=\"nofollow\">Mutation Observer<\/a>, and <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Global_Objects\/Promise\" rel=\"nofollow\">Promises<\/a>. Eventually we expect all browsers to implement these standards. In the meantime, Trix includes <a href=\"https:\/\/en.wikipedia.org\/wiki\/Polyfill_(programming)\" rel=\"nofollow\">polyfills<\/a> for missing functionality.<\/p>\n<p>Include the bundled <code>trix.css<\/code> and <code>trix.js<\/code> files in the <code>&lt;head&gt;<\/code> of your page.<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;head&gt;\n  \u2026\n  &lt;link rel=\"stylesheet\" type=\"text\/css\" href=\"trix.css\"&gt;\n  &lt;script type=\"text\/javascript\" src=\"trix.js\"&gt;&lt;\/script&gt;\n&lt;\/head&gt;<\/pre><\/div>\n<p><code>trix.css<\/code> includes default styles for the Trix toolbar, editor, and attachments. Skip this file if you prefer to define these styles yourself.<\/p>\n<p>To use your own polyfills, or to target only browsers that support all of the required standards, include <code>trix-core.js<\/code> instead.<\/p>\n<h2><a id=\"user-content-creating-an-editor\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-an-editor\"><\/a>Creating an Editor<\/h2>\n<p>Place an empty <code>&lt;trix-editor&gt;&lt;\/trix-editor&gt;<\/code> tag on the page. Trix will automatically insert a separate <code>&lt;trix-toolbar&gt;<\/code> before the editor.<\/p>\n<p>Like an HTML <code>&lt;textarea&gt;<\/code>, <code>&lt;trix-editor&gt;<\/code> accepts <code>autofocus<\/code> and <code>placeholder<\/code> attributes. Unlike a <code>&lt;textarea&gt;<\/code>, <code>&lt;trix-editor&gt;<\/code> automatically expands vertically to fit its contents.<\/p>\n<h2><a id=\"user-content-integrating-with-forms\" class=\"anchor\" aria-hidden=\"true\" href=\"#integrating-with-forms\"><\/a>Integrating With Forms<\/h2>\n<p>To submit the contents of a <code>&lt;trix-editor&gt;<\/code> with a form, first define a hidden input field in the form and assign it an <code>id<\/code>. Then reference that <code>id<\/code> in the editor\u2019s <code>input<\/code> attribute.<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;form \u2026&gt;\n  &lt;input id=\"x\" type=\"hidden\" name=\"content\"&gt;\n  &lt;trix-editor input=\"x\"&gt;&lt;\/trix-editor&gt;\n&lt;\/form&gt;<\/pre><\/div>\n<p>Trix will automatically update the value of the hidden input field with each change to the editor.<\/p>\n<h2><a id=\"user-content-populating-with-stored-content\" class=\"anchor\" aria-hidden=\"true\" href=\"#populating-with-stored-content\"><\/a>Populating With Stored Content<\/h2>\n<p>To populate a <code>&lt;trix-editor&gt;<\/code> with stored content, include that content in the associated input element\u2019s <code>value<\/code> attribute.<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;form \u2026&gt;\n  &lt;input id=\"x\" value=\"Editor content goes here\" type=\"hidden\" name=\"content\"&gt;\n  &lt;trix-editor input=\"x\"&gt;&lt;\/trix-editor&gt;\n&lt;\/form&gt;<\/pre><\/div>\n<p>Always use an associated input element to safely populate an editor. Trix won\u2019t load any HTML content inside a <code>&lt;trix-editor&gt;\u2026&lt;\/trix-editor&gt;<\/code> tag.<\/p>\n<h2><a id=\"user-content-styling-formatted-content\" class=\"anchor\" aria-hidden=\"true\" href=\"#styling-formatted-content\"><\/a>Styling Formatted Content<\/h2>\n<p>To ensure what you see when you edit is what you see when you save, use a CSS class name to scope styles for Trix formatted content. Apply this class name to your <code>&lt;trix-editor&gt;<\/code> element, and to a containing element when you render stored Trix content for display in your application.<\/p>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;trix-editor class=\"trix-content\"&gt;&lt;\/trix-editor&gt;<\/pre><\/div>\n<div class=\"highlight highlight-text-html-basic\"><pre>&lt;div class=\"trix-content\"&gt;Stored content here&lt;\/div&gt;<\/pre><\/div>\n<p>The default <code>trix.css<\/code> file includes styles for basic formatted content\u2014including bulleted and numbered lists, code blocks, and block quotes\u2014under the class name <code>trix-content<\/code>. We encourage you to use these styles as a starting point by copying them into your application\u2019s CSS with a different class name.<\/p>\n<h2><a id=\"user-content-storing-attached-files\" class=\"anchor\" aria-hidden=\"true\" href=\"#storing-attached-files\"><\/a>Storing Attached Files<\/h2>\n<p>Trix automatically accepts files dragged or pasted into an editor and inserts them as attachments in the document. Each attachment is considered <em>pending<\/em> until you store it remotely and provide Trix with a permanent URL.<\/p>\n<p>To store attachments, listen for the <code>trix-attachment-add<\/code> event. Upload the attached files with XMLHttpRequest yourself and set the attachment\u2019s URL attribute upon completion. See the <a href=\"https:\/\/trix-editor.org\/js\/attachments.js\" rel=\"nofollow\">attachment example<\/a> for detailed information.<\/p>\n<p>If you don\u2019t want to accept dropped or pasted files, call <code>preventDefault()<\/code> on the <code>trix-file-accept<\/code> event, which Trix dispatches just before the <code>trix-attachment-add<\/code> event.<\/p>\n<p>You can manipulate a Trix editor programmatically through the <code>Trix.Editor<\/code> interface, available on each <code>&lt;trix-editor&gt;<\/code> element through its <code>editor<\/code> property.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var element = document.querySelector(\"trix-editor\")\nelement.editor  \/\/ is a Trix.Editor instance<\/pre><\/div>\n<h2><a id=\"user-content-understanding-the-document-model\" class=\"anchor\" aria-hidden=\"true\" href=\"#understanding-the-document-model\"><\/a>Understanding the Document Model<\/h2>\n<p>The formatted content of a Trix editor is known as a <em>document<\/em>, and is represented as an instance of the <code>Trix.Document<\/code> class. To get the editor\u2019s current document, use the <code>editor.getDocument<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.getDocument()  \/\/ is a Trix.Document instance<\/pre><\/div>\n<p>You can convert a document to an unformatted JavaScript string with the <code>document.toString<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var document = element.editor.getDocument()\ndocument.toString()  \/\/ is a JavaScript string<\/pre><\/div>\n<h3><a id=\"user-content-immutability-and-equality\" class=\"anchor\" aria-hidden=\"true\" href=\"#immutability-and-equality\"><\/a>Immutability and Equality<\/h3>\n<p>Documents are immutable values. Each change you make in an editor replaces the previous document with a new document. Capturing a snapshot of the editor\u2019s content is as simple as keeping a reference to its document, since that document will never change over time. (This is how Trix implements undo.)<\/p>\n<p>To compare two documents for equality, use the <code>document.isEqualTo<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var document = element.editor.getDocument()\ndocument.isEqualTo(element.editor.getDocument())  \/\/ true<\/pre><\/div>\n<h2><a id=\"user-content-getting-and-setting-the-selection\" class=\"anchor\" aria-hidden=\"true\" href=\"#getting-and-setting-the-selection\"><\/a>Getting and Setting the Selection<\/h2>\n<p>Trix documents are structured as sequences of individually addressable characters. The index of one character in a document is called a <em>position<\/em>, and a start and end position together make up a <em>range<\/em>.<\/p>\n<p>To get the editor\u2019s current selection, use the <code>editor.getSelectedRange<\/code> method, which returns a two-element array containing the start and end positions.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.getSelectedRange()  \/\/ [0, 0]<\/pre><\/div>\n<p>You can set the editor\u2019s current selection by passing a range array to the <code>editor.setSelectedRange<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Select the first character in the document\nelement.editor.setSelectedRange([0, 1])<\/pre><\/div>\n<h3><a id=\"user-content-collapsed-selections\" class=\"anchor\" aria-hidden=\"true\" href=\"#collapsed-selections\"><\/a>Collapsed Selections<\/h3>\n<p>When the start and end positions of a range are equal, the range is said to be <em>collapsed<\/em>. In the editor, a collapsed selection appears as a blinking cursor rather than a highlighted span of text.<\/p>\n<p>For convenience, the following calls to <code>setSelectedRange<\/code> are equivalent when working with collapsed selections:<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.setSelectedRange(1)\nelement.editor.setSelectedRange([1])\nelement.editor.setSelectedRange([1, 1])<\/pre><\/div>\n<h3><a id=\"user-content-directional-movement\" class=\"anchor\" aria-hidden=\"true\" href=\"#directional-movement\"><\/a>Directional Movement<\/h3>\n<p>To programmatically move the cursor or selection through the document, call the <code>editor.moveCursorInDirection<\/code> or <code>editor.expandSelectionInDirection<\/code> methods with a <em>direction<\/em> argument. The direction can be either <code>\"forward\"<\/code> or <code>\"backward\"<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Move the cursor backward one character\nelement.editor.moveCursorInDirection(\"backward\")\n\/\/ Expand the end of the selection forward by one character\nelement.editor.expandSelectionInDirection(\"forward\")<\/pre><\/div>\n<h3><a id=\"user-content-converting-positions-to-pixel-offsets\" class=\"anchor\" aria-hidden=\"true\" href=\"#converting-positions-to-pixel-offsets\"><\/a>Converting Positions to Pixel Offsets<\/h3>\n<p>Sometimes you need to know the <em>x<\/em> and <em>y<\/em> coordinates of a character at a given position in the editor. For example, you might want to absolutely position a pop-up menu element below the editor\u2019s cursor.<\/p>\n<p>Call the <code>editor.getClientRectAtPosition<\/code> method with a position argument to get a <a href=\"https:\/\/drafts.fxtf.org\/geometry\/#DOMRect\" rel=\"nofollow\"><code>DOMRect<\/code><\/a> instance representing the left and top offsets, width, and height of the character at the given position.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var rect = element.editor.getClientRectAtPosition(0)\n[rect.left, rect.top]  \/\/ [17, 49]<\/pre><\/div>\n<h2><a id=\"user-content-inserting-and-deleting-text\" class=\"anchor\" aria-hidden=\"true\" href=\"#inserting-and-deleting-text\"><\/a>Inserting and Deleting Text<\/h2>\n<p>The editor interface provides methods for inserting, replacing, and deleting text at the current selection.<\/p>\n<p>To insert or replace text, begin by setting the selected range, then call one of the insertion methods below. Trix will first remove any selected text, then insert the new text at the start position of the selected range.<\/p>\n<h3><a id=\"user-content-inserting-plain-text\" class=\"anchor\" aria-hidden=\"true\" href=\"#inserting-plain-text\"><\/a>Inserting Plain Text<\/h3>\n<p>To insert unformatted text into the document, call the <code>editor.insertString<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Insert \u201cHello\u201d at the beginning of the document\nelement.editor.setSelectedRange([0, 0])\nelement.editor.insertString(\"Hello\")<\/pre><\/div>\n<h3><a id=\"user-content-inserting-html\" class=\"anchor\" aria-hidden=\"true\" href=\"#inserting-html\"><\/a>Inserting HTML<\/h3>\n<p>To insert HTML into the document, call the <code>editor.insertHTML<\/code> method. Trix will first convert the HTML into its internal document model. During this conversion, any formatting that cannot be represented in a Trix document will be lost.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Insert a bold \u201cHello\u201d at the beginning of the document\nelement.editor.setSelectedRange([0, 0])\nelement.editor.insertHTML(\"&lt;strong&gt;Hello&lt;\/strong&gt;\")<\/pre><\/div>\n<h3><a id=\"user-content-inserting-a-file\" class=\"anchor\" aria-hidden=\"true\" href=\"#inserting-a-file\"><\/a>Inserting a File<\/h3>\n<p>To insert a DOM <a href=\"http:\/\/www.w3.org\/TR\/FileAPI\/#file\" rel=\"nofollow\"><code>File<\/code><\/a> object into the document, call the <code>editor.insertFile<\/code> method. Trix will insert a pending attachment for the file as if you had dragged and dropped it onto the editor.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Insert the selected file from the first file input element\nvar file = document.querySelector(\"input[type=file]\").file\nelement.editor.insertFile(file)<\/pre><\/div>\n<h3><a id=\"user-content-inserting-a-content-attachment\" class=\"anchor\" aria-hidden=\"true\" href=\"#inserting-a-content-attachment\"><\/a>Inserting a Content Attachment<\/h3>\n<p>Content attachments are self-contained units of HTML that behave like files in the editor. They can be moved or removed, but not edited directly, and are represented by a single character position in the document model.<\/p>\n<p>To insert HTML as an attachment, create a <code>Trix.Attachment<\/code> with a <code>content<\/code> attribute and call the <code>editor.insertAttachment<\/code> method. The HTML inside a content attachment is not subject to Trix\u2019s document conversion rules and will be rendered as-is.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>var attachment = new Trix.Attachment({ content: '&lt;span class=\"mention\"&gt;@trix&lt;\/span&gt;' })\nelement.editor.insertAttachment(attachment)<\/pre><\/div>\n<h3><a id=\"user-content-inserting-a-line-break\" class=\"anchor\" aria-hidden=\"true\" href=\"#inserting-a-line-break\"><\/a>Inserting a Line Break<\/h3>\n<p>To insert a line break, call the <code>editor.insertLineBreak<\/code> method, which is functionally equivalent to pressing the return key.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Insert \u201cHello\\n\u201d\nelement.editor.insertString(\"Hello\")\nelement.editor.insertLineBreak()<\/pre><\/div>\n<h3><a id=\"user-content-deleting-text\" class=\"anchor\" aria-hidden=\"true\" href=\"#deleting-text\"><\/a>Deleting Text<\/h3>\n<p>If the current selection is collapsed, you can simulate deleting text before or after the cursor with the <code>editor.deleteInDirection<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ \u201cBackspace\u201d the first character in the document\nelement.editor.setSelectedRange([1, 1])\nelement.editor.deleteInDirection(\"backward\")\n\/\/ Delete the second character in the document\nelement.editor.setSelectedRange([1, 1])\nelement.editor.deleteInDirection(\"forward\")<\/pre><\/div>\n<p>To delete a range of text, first set the selected range, then call <code>editor.deleteInDirection<\/code> with either direction as the argument.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Delete the first five characters\nelement.editor.setSelectedRange([0, 4])\nelement.editor.deleteInDirection(\"forward\")<\/pre><\/div>\n<h2><a id=\"user-content-working-with-attributes-and-nesting\" class=\"anchor\" aria-hidden=\"true\" href=\"#working-with-attributes-and-nesting\"><\/a>Working With Attributes and Nesting<\/h2>\n<p>Trix represents formatting as sets of <em>attributes<\/em> applied across ranges of a document.<\/p>\n<p>By default, Trix supports the inline attributes <code>bold<\/code>, <code>italic<\/code>, <code>href<\/code>, and <code>strike<\/code>, and the block-level attributes <code>heading1<\/code>, <code>quote<\/code>, <code>code<\/code>, <code>bullet<\/code>, and <code>number<\/code>.<\/p>\n<h3><a id=\"user-content-applying-formatting\" class=\"anchor\" aria-hidden=\"true\" href=\"#applying-formatting\"><\/a>Applying Formatting<\/h3>\n<p>To apply formatting to the current selection, use the <code>editor.activateAttribute<\/code> method.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.insertString(\"Hello\")\nelement.editor.setSelectedRange([0, 5])\nelement.editor.activateAttribute(\"bold\")<\/pre><\/div>\n<p>To set the <code>href<\/code> attribute, pass a URL as the second argument to <code>editor.activateAttribute<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.insertString(\"Trix\")\nelement.editor.setSelectedRange([0, 4])\nelement.editor.activateAttribute(\"href\", \"https:\/\/trix-editor.org\/\")<\/pre><\/div>\n<h3><a id=\"user-content-removing-formatting\" class=\"anchor\" aria-hidden=\"true\" href=\"#removing-formatting\"><\/a>Removing Formatting<\/h3>\n<p>Use the <code>editor.deactivateAttribute<\/code> method to remove formatting from a selection.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.setSelectedRange([2, 4])\nelement.editor.deactivateAttribute(\"bold\")<\/pre><\/div>\n<h3><a id=\"user-content-formatting-with-a-collapsed-selection\" class=\"anchor\" aria-hidden=\"true\" href=\"#formatting-with-a-collapsed-selection\"><\/a>Formatting With a Collapsed Selection<\/h3>\n<p>If you activate or deactivate attributes when the selection is collapsed, your formatting changes will apply to the text inserted by any subsequent calls to <code>editor.insertString<\/code>.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.activateAttribute(\"italic\")\nelement.editor.insertString(\"This is italic\")<\/pre><\/div>\n<h3><a id=\"user-content-adjusting-the-nesting-level\" class=\"anchor\" aria-hidden=\"true\" href=\"#adjusting-the-nesting-level\"><\/a>Adjusting the Nesting Level<\/h3>\n<p>To adjust the nesting level of quotes, bulleted lists, or numbered lists, call the <code>editor.increaseNestingLevel<\/code> and <code>editor.decreaseNestingLevel<\/code> methods.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.activateAttribute(\"quote\")\nelement.editor.increaseNestingLevel()\nelement.editor.decreaseNestingLevel()<\/pre><\/div>\n<h2><a id=\"user-content-using-undo-and-redo\" class=\"anchor\" aria-hidden=\"true\" href=\"#using-undo-and-redo\"><\/a>Using Undo and Redo<\/h2>\n<p>Trix editors support unlimited undo and redo. Successive typing and formatting changes are consolidated together at five-second intervals; all other input changes are recorded individually in undo history.<\/p>\n<p>Call the <code>editor.undo<\/code> and <code>editor.redo<\/code> methods to perform an undo or redo operation.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.undo()\nelement.editor.redo()<\/pre><\/div>\n<p>Changes you make through the editor interface will not automatically record undo entries. You can save your own undo entries by calling the <code>editor.recordUndoEntry<\/code> method with a description argument.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>element.editor.insertString(\"Hello\")\nelement.editor.recordUndoEntry(\"Insert Text\")<\/pre><\/div>\n<h2><a id=\"user-content-loading-and-saving-editor-state\" class=\"anchor\" aria-hidden=\"true\" href=\"#loading-and-saving-editor-state\"><\/a>Loading and Saving Editor State<\/h2>\n<p>Serialize an editor\u2019s state with <code>JSON.stringify<\/code> and restore saved state with the <code>editor.loadJSON<\/code> method. The serialized state includes the document and current selection, but does not include undo history.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>\/\/ Save editor state to local storage\nlocalStorage[\"editorState\"] = JSON.stringify(element.editor)\n\/\/ Restore editor state from local storage\nelement.editor.loadJSON(JSON.parse(localStorage[\"editorState\"]))<\/pre><\/div>\n<h2><a id=\"user-content-observing-editor-changes\" class=\"anchor\" aria-hidden=\"true\" href=\"#observing-editor-changes\"><\/a>Observing Editor Changes<\/h2>\n<p>The <code>&lt;trix-editor&gt;<\/code> element emits several events which you can use to observe and respond to changes in editor state.<\/p>\n<ul><li>\n<p><code>trix-before-initialize<\/code> fires when the <code>&lt;trix-editor&gt;<\/code> element is attached to the DOM just before Trix installs its <code>editor<\/code> object.<\/p>\n<\/li>\n<li>\n<p><code>trix-initialize<\/code> fires when the <code>&lt;trix-editor&gt;<\/code> element is attached to the DOM and its <code>editor<\/code> object is ready for use.<\/p>\n<\/li>\n<li>\n<p><code>trix-change<\/code> fires whenever the editor\u2019s contents have changed.<\/p>\n<\/li>\n<li>\n<p><code>trix-selection-change<\/code> fires any time the selected range changes in the editor.<\/p>\n<\/li>\n<li>\n<p><code>trix-focus<\/code> and <code>trix-blur<\/code> fire when the editor gains or loses focus, respectively.<\/p>\n<\/li>\n<li>\n<p><code>trix-file-accept<\/code> fires when a file is dropped or inserted into the editor. You can access the DOM <code>File<\/code> object through the <code>file<\/code> property on the event. Call <code>preventDefault<\/code> on the event to prevent attaching the file to the document.<\/p>\n<\/li>\n<li>\n<p><code>trix-attachment-add<\/code> fires after an attachment is added to the document. You can access the Trix attachment object through the <code>attachment<\/code> property on the event. If the <code>attachment<\/code> object has a <code>file<\/code> property, you should store this file remotely and set the attachment\u2019s URL attribute. See the <a href=\"http:\/\/trix-editor.org\/js\/attachments.js\" rel=\"nofollow\">attachment example<\/a> for detailed information.<\/p>\n<\/li>\n<li>\n<p><code>trix-attachment-remove<\/code> fires when an attachment is removed from the document. You can access the Trix attachment object through the <code>attachment<\/code> property on the event. You may wish to use this event to clean up remotely stored files.<\/p>\n<\/li>\n<\/ul><p>Trix is open-source software, freely distributable under the terms of an <a href=\"https:\/\/github.com\/basecamp\/trix\/blob\/master\/LICENSE\">MIT-style license<\/a>. The <a href=\"https:\/\/github.com\/basecamp\/trix\">source code is hosted on GitHub<\/a>.<\/p>\n<p>We welcome contributions in the form of bug reports, pull requests, or thoughtful discussions in the <a href=\"https:\/\/github.com\/basecamp\/trix\/issues\">GitHub issue tracker<\/a>. Please see the <a href=\"https:\/\/github.com\/basecamp\/trix\/blob\/master\/CODE_OF_CONDUCT.md\">Code of Conduct<\/a> for our pledge to contributors.<\/p>\n<p>Trix was created by <a href=\"https:\/\/twitter.com\/javan\" rel=\"nofollow\">Javan Makhmali<\/a> and <a href=\"https:\/\/twitter.com\/sstephenson\" rel=\"nofollow\">Sam Stephenson<\/a>, with development sponsored by <a href=\"https:\/\/basecamp.com\/\" rel=\"nofollow\">Basecamp<\/a>.<\/p>\n<h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#building-from-source\"><\/a>Building From Source<\/h3>\n<p>Trix is written in <a href=\"https:\/\/github.com\/jashkenas\/coffeescript\">CoffeeScript<\/a> and compiled to JavaScript with <a href=\"https:\/\/github.com\/javan\/blade\">Blade<\/a>.<\/p>\n<p>From inside a checkout of the Trix Git repository, issue the following commands to build the distributable files in <code>dist\/<\/code>:<\/p>\n<pre>$ bin\/setup\n$ bin\/blade build\n<\/pre>\n<h3><a id=\"user-content-developing-in-browser\" class=\"anchor\" aria-hidden=\"true\" href=\"#developing-in-browser\"><\/a>Developing In-Browser<\/h3>\n<p>You can spawn a development web server to work on Trix in a more convenient fashion. Instead of manually rebuilding the source each time, just reload a page in your browser to see your changes.<\/p>\n<p>To develop in-browser, run <code>bin\/setup<\/code> and follow the displayed instructions.<\/p>\n<h3><a id=\"user-content-running-tests\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-tests\"><\/a>Running Tests<\/h3>\n<p>Make sure you\u2019re set up to build from source using the instructions above. Then run <code>bin\/blade runner<\/code> and visit the displayed URL to run the Trix test suite.<\/p>\n<h3><a id=\"user-content-pull-requests\" class=\"anchor\" aria-hidden=\"true\" href=\"#pull-requests\"><\/a>Pull Requests<\/h3>\n<p>Only commit changes to Trix\u2019s source (everything except the compiled files in <code>\/dist<\/code>) and leave the <a href=\"https:\/\/github.com\/basecamp\/trix\/blob\/master\/src\/trix\/VERSION\">VERSION<\/a> unchanged. We update both when publishing new <a href=\"https:\/\/github.com\/basecamp\/trix\/releases\">releases<\/a>. \u2764\ufe0f<\/p>\n<hr \/><p>\u00a9 2018 Basecamp, LLC.<\/p>\n<\/article>","created_at":"2018-11-07T16:22:05+0000","updated_at":"2018-11-07T16:22:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":12,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/13131?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":42,"title":"Tailwind CSS","url":"https:\/\/tailwindcss.com\/","content":"<div class=\"flex flex-col\"><div class=\"min-h-screen bg-pattern bg-center bg-grey-lighter border-t-6 border-tailwind-teal flex items-center justify-center leading-tight p-6 pb-16\"><div><div><p class=\"text-center\">\n          <a href=\"https:\/\/github.com\/tailwindcss\/tailwindcss\/releases\" class=\"text-lg text-grey hover:text-grey-dark font-semibold tracking-tight\">v0.7.2<\/a>\n        <\/p><\/div><p><a class=\"mt-6 sm:mt-0 mx-auto sm:mx-2 max-w-xs rounded-full text-center leading-none font-semibold block px-12 py-3 border-2 border-tailwind-teal bg-tailwind-teal text-white hover:border-tailwind-teal-light hover:bg-tailwind-teal-light hover:bg-tailwind-teal-light\" href=\"https:\/\/tailwindcss.com\/docs\/what-is-tailwind\">Learn More<\/a>\n        <a class=\"mt-6 sm:mt-0 mx-auto sm:mx-2 max-w-xs rounded-full text-center leading-none font-semibold block px-12 py-3 border-2 border-tailwind-teal text-tailwind-teal hover:text-white hover:border-tailwind-teal-light hover:bg-tailwind-teal-light hover:bg-tailwind-teal-light\" href=\"https:\/\/github.com\/tailwindcss\/tailwindcss\">GitHub<\/a><\/p><\/div><\/div><\/div><div class=\"px-6\"><p>A project by<\/p><div class=\"flex justify-center\"><div class=\"pb-4 flex justify-center flex-wrap max-w-md xl:max-w-full border-b-2 border-grey-light\"><div class=\"mb-8 w-64 md:flex-x-center\"><div class=\"flex items-center\"><img class=\"h-16 w-16 rounded-full\" src=\"https:\/\/res.cloudinary.com\/dchiba2d6\/image\/upload\/c_scale,q_auto,w_128\/v1531232470\/tailwindcss\/adam-profile.jpg\" alt=\"\" \/><p><a href=\"https:\/\/twitter.com\/adamwathan\" class=\"block hover:underline text-dark text-lg text-medium\">Adam Wathan<\/a>\n            <a href=\"https:\/\/twitter.com\/adamwathan\" class=\"block hover:underline text-tailwind-teal-dark text-sm\">@adamwathan<\/a><\/p><\/div><\/div><div class=\"mb-8 w-64 md:flex-x-center\"><div class=\"flex items-center\"><img class=\"h-16 w-16 rounded-full\" src=\"https:\/\/res.cloudinary.com\/dchiba2d6\/image\/upload\/c_scale,q_auto,w_128\/v1531232470\/tailwindcss\/reinink-profile.jpg\" alt=\"\" \/><p><a href=\"https:\/\/twitter.com\/reinink\" class=\"block hover:underline text-dark text-lg text-medium\">Jonathan Reinink<\/a>\n            <a href=\"https:\/\/twitter.com\/reinink\" class=\"block hover:underline text-tailwind-teal-dark text-sm\">@reinink<\/a><\/p><\/div><\/div><div class=\"mb-8 w-64 md:flex-x-center\"><div class=\"flex items-center\"><img class=\"h-16 w-16 rounded-full\" src=\"https:\/\/res.cloudinary.com\/dchiba2d6\/image\/upload\/c_scale,q_auto,w_128\/v1531232470\/tailwindcss\/hemphill-profile.jpg\" alt=\"image\" \/><p><a href=\"https:\/\/twitter.com\/davidhemphill\" class=\"block hover:underline text-dark text-lg text-medium\">David Hemphill<\/a>\n            <a href=\"https:\/\/twitter.com\/davidhemphill\" class=\"block hover:underline text-tailwind-teal-dark text-sm\">@davidhemphill<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"mt-10 mb-16 mx-auto flex flex-wrap justify-center text-grey-darker\"><a class=\"block flex items-center hover:text-tailwind-teal\" href=\"https:\/\/github.com\/tailwindcss\/tailwindcss\">\n      \n      GitHub\n    <\/a><a class=\"block ml-6 sm:ml-12 flex items-center hover:text-tailwind-teal\" href=\"https:\/\/twitter.com\/tailwindcss\">\n      \n      Twitter\n    <\/a><\/div><\/div>","created_at":"2018-11-07T16:22:05+0000","updated_at":"2018-11-07T16:22:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"tailwindcss.com","preview_picture":"https:\/\/tailwindcss.com\/img\/twitter-large-card.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":43,"title":"Tailwind CSS","url":"https:\/\/tailwindcss.com\/","content":"<div class=\"flex flex-col\"><div class=\"min-h-screen bg-pattern bg-center bg-grey-lighter border-t-6 border-tailwind-teal flex items-center justify-center leading-tight p-6 pb-16\"><div><div><p class=\"text-center\">\n          <a href=\"https:\/\/github.com\/tailwindcss\/tailwindcss\/releases\" class=\"text-lg text-grey hover:text-grey-dark font-semibold tracking-tight\">v0.7.2<\/a>\n        <\/p><\/div><p><a class=\"mt-6 sm:mt-0 mx-auto sm:mx-2 max-w-xs rounded-full text-center leading-none font-semibold block px-12 py-3 border-2 border-tailwind-teal bg-tailwind-teal text-white hover:border-tailwind-teal-light hover:bg-tailwind-teal-light hover:bg-tailwind-teal-light\" href=\"https:\/\/tailwindcss.com\/docs\/what-is-tailwind\">Learn More<\/a>\n        <a class=\"mt-6 sm:mt-0 mx-auto sm:mx-2 max-w-xs rounded-full text-center leading-none font-semibold block px-12 py-3 border-2 border-tailwind-teal text-tailwind-teal hover:text-white hover:border-tailwind-teal-light hover:bg-tailwind-teal-light hover:bg-tailwind-teal-light\" href=\"https:\/\/github.com\/tailwindcss\/tailwindcss\">GitHub<\/a><\/p><\/div><\/div><\/div><div class=\"px-6\"><p>A project by<\/p><div class=\"flex justify-center\"><div class=\"pb-4 flex justify-center flex-wrap max-w-md xl:max-w-full border-b-2 border-grey-light\"><div class=\"mb-8 w-64 md:flex-x-center\"><div class=\"flex items-center\"><img class=\"h-16 w-16 rounded-full\" src=\"https:\/\/res.cloudinary.com\/dchiba2d6\/image\/upload\/c_scale,q_auto,w_128\/v1531232470\/tailwindcss\/adam-profile.jpg\" alt=\"\" \/><p><a href=\"https:\/\/twitter.com\/adamwathan\" class=\"block hover:underline text-dark text-lg text-medium\">Adam Wathan<\/a>\n            <a href=\"https:\/\/twitter.com\/adamwathan\" class=\"block hover:underline text-tailwind-teal-dark text-sm\">@adamwathan<\/a><\/p><\/div><\/div><div class=\"mb-8 w-64 md:flex-x-center\"><div class=\"flex items-center\"><img class=\"h-16 w-16 rounded-full\" src=\"https:\/\/res.cloudinary.com\/dchiba2d6\/image\/upload\/c_scale,q_auto,w_128\/v1531232470\/tailwindcss\/reinink-profile.jpg\" alt=\"\" \/><p><a href=\"https:\/\/twitter.com\/reinink\" class=\"block hover:underline text-dark text-lg text-medium\">Jonathan Reinink<\/a>\n            <a href=\"https:\/\/twitter.com\/reinink\" class=\"block hover:underline text-tailwind-teal-dark text-sm\">@reinink<\/a><\/p><\/div><\/div><div class=\"mb-8 w-64 md:flex-x-center\"><div class=\"flex items-center\"><img class=\"h-16 w-16 rounded-full\" src=\"https:\/\/res.cloudinary.com\/dchiba2d6\/image\/upload\/c_scale,q_auto,w_128\/v1531232470\/tailwindcss\/hemphill-profile.jpg\" alt=\"image\" \/><p><a href=\"https:\/\/twitter.com\/davidhemphill\" class=\"block hover:underline text-dark text-lg text-medium\">David Hemphill<\/a>\n            <a href=\"https:\/\/twitter.com\/davidhemphill\" class=\"block hover:underline text-tailwind-teal-dark text-sm\">@davidhemphill<\/a><\/p><\/div><\/div><\/div><\/div><div class=\"mt-10 mb-16 mx-auto flex flex-wrap justify-center text-grey-darker\"><a class=\"block flex items-center hover:text-tailwind-teal\" href=\"https:\/\/github.com\/tailwindcss\/tailwindcss\">\n      \n      GitHub\n    <\/a><a class=\"block ml-6 sm:ml-12 flex items-center hover:text-tailwind-teal\" href=\"https:\/\/twitter.com\/tailwindcss\">\n      \n      Twitter\n    <\/a><\/div><\/div>","created_at":"2018-11-07T16:22:05+0000","updated_at":"2018-11-07T16:22:05+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"tailwindcss.com","preview_picture":"https:\/\/tailwindcss.com\/img\/twitter-large-card.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":40,"title":"fediverse-stacks \u2013 EtherCalc","url":"https:\/\/ethercalc.org\/fediverse-stacks","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:22:04+0000","updated_at":"2018-11-07T16:22:04+0000","annotations":[],"reading_time":0,"domain_name":"ethercalc.org","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":38,"title":"aparrish\/seaduck","url":"https:\/\/github.com\/aparrish\/seaduck","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>By <a href=\"https:\/\/www.decontextualize.com\/\" rel=\"nofollow\">Allison Parrish<\/a><\/p>\n<p>Sea Duck is a minimally opinionated rough-draft JavaScript framework for\nproducing narratives through simulation. All you have to do is define a list of\nnouns, a list of actions that happen to those nouns (thereby producing events),\nand a <a href=\"http:\/\/tracery.io\" rel=\"nofollow\">Tracery<\/a> grammar to turn those events into text.<\/p>\n<p>Note: This library uses a number of ES6 features\n(<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Global_Objects\/Map\" rel=\"nofollow\">maps<\/a>,\n<a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Statements\/function*\" rel=\"nofollow\">generator\nfunctions<\/a>)\nand I haven't used a transpiler or anything to provide backwards compatibility.\nInternet Explorer probably won't work (but Edge probably will?). Sorry!<\/p>\n<p>I made this framework primarily as a toy for experimentation and teaching,\nbecause I couldn't find any similar frameworks (available in a more-or-less\ncomplete form to the general public) that make it easy to get started with\nmaking narratives from simulated events. The idea was to facilitate quick\nprototypes for projects along the lines of Darius Kazemi's <a href=\"http:\/\/tinysubversions.com\/nanogenmo\/novel-2.pdf\" rel=\"nofollow\">Teens Wander Around\nA House<\/a> (<a href=\"https:\/\/github.com\/dariusk\/NaNoGenMo\/issues\/2\">NaNoGenMo thread\nhere<\/a>) As a consequence, this\nframework doesn't contain a built-in world model, or a system for solving\nconstraints or action planning (though you could implement any of those things\nyourself on top of the framework).<\/p>\n<p>Pull requests or suggestions for overall architecture are solicited.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>You can npm install from this git repo and <code>require('seaduck')<\/code> in your code:<\/p>\n<pre>npm install https:\/\/github.com\/aparrish\/seaduck --save\n<\/pre>\n<p>Or you can <a href=\"https:\/\/github.com\/aparrish\/seaduck\/blob\/master\/build\/seaduck-bundle.js\">download the bundle<\/a> (click \"Raw\" and\nthen <code>File &gt; Save as...<\/code> or equivalent in your browser) and add it to your\nproject with a <code>&lt;script&gt;<\/code> tag:<\/p>\n<pre>&lt;script src=\"seaduck-bundle.js\"&gt;&lt;\/script&gt;\n<\/pre>\n<p>... which will create a global <code>seaduck<\/code> variable available to the rest of your\nJavaScript code.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<p>You can find annotated examples in the <code>examples\/<\/code> folder. If you've downloaded\nthis repo, you can run them like so:<\/p>\n<pre>node examples\/01_basic.js\n<\/pre>\n<p>The examples are also available in adapted form on the <a href=\"https:\/\/editor.p5js.org\/\" rel=\"nofollow\">p5.js web\neditor<\/a>:<\/p>\n<ul><li><a href=\"https:\/\/editor.p5js.org\/allison.parrish\/sketches\/ByC_daWcX\" rel=\"nofollow\">Example 1: The Basics<\/a><\/li>\n<li><a href=\"https:\/\/editor.p5js.org\/allison.parrish\/sketches\/S15I3TbcQ\" rel=\"nofollow\">Example 2: With\ntags<\/a><\/li>\n<li><a href=\"https:\/\/editor.p5js.org\/allison.parrish\/sketches\/Hkrhppb9X\" rel=\"nofollow\">Example 3:\nRelations<\/a><\/li>\n<li><a href=\"https:\/\/editor.p5js.org\/allison.parrish\/sketches\/BJflApbcm\" rel=\"nofollow\">Example 4: Rooms and\nobjects<\/a><\/li>\n<\/ul><h2><a id=\"user-content-concepts-and-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#concepts-and-usage\"><\/a>Concepts and usage<\/h2>\n<p>The purpose of a Sea Duck narrative is to generate sentences from events that\ntranspire as nouns interact according to a set of rules. At each step of the\nsimulation, the state of the nouns (properties and relations) are checked, and\nthen potentially changed, generating narrative events in the process, which are\nrendered as text using Tracery grammars. A Sea Duck narrative consists of the\nfollowing components:<\/p>\n<ul><li>Nouns<\/li>\n<li>Actions<\/li>\n<li>Relations<\/li>\n<li>Events<\/li>\n<li>Discourse<\/li>\n<\/ul><p>Create Sea Duck narrative object like so:<\/p>\n<pre>let n = new seaduck.Narrative({\n    \"nouns\": [\n        ... list of noun objects ...\n    ],\n    \"actions\": [\n        ... list of action definitions ...\n    ],\n    \"initialize\": function() { ...code to execute on startup... },\n    \"traceryDiscourse\": {\n        ... tracery rules for each kind of event ...\n    }\n});\n<\/pre>\n<p>After creation, you can step through the simulation using <code>.step()<\/code>, which\nreturns a list of <code>StoryEvent<\/code> objects created during the simulation step; or\nyou can call <code>.stepAndRender()<\/code>, which collects events from <code>.step()<\/code> and\nrenders them using the associated Tracery grammar.<\/p>\n<h3><a id=\"user-content-nouns-and-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#nouns-and-properties\"><\/a>Nouns and properties<\/h3>\n<p>A <code>noun<\/code> is something that will participate in the narrative (a person, place,\nor thing). Here's an example noun:<\/p>\n<pre>{\n  \"name\": \"Joe\",\n  \"properties\": {\n    \"happiness\": 0,\n    \"hungry\": true\n  },\n  \"tags\": [\"person\"]\n}\n<\/pre>\n<p>The <code>name<\/code> attribute identifies the noun and is also the string that will be\nused to render it in the Tracery output; the <code>properties<\/code> attribute is an\nobject for storing values associated with this noun; and the <code>tags<\/code> attribute\nis a list of strings \"tagging\" this noun. (The tags are used to group objects\ninto categories for the purpose of relations and action matching; see below.)\nNouns must have <em>all<\/em> of these attributes defined.<\/p>\n<h3><a id=\"user-content-actions\" class=\"anchor\" aria-hidden=\"true\" href=\"#actions\"><\/a>Actions<\/h3>\n<p>An <code>action<\/code> is a series of checks to perform against nouns on each step, along\nwith code for what to do if those checks pass. A typical action is structured\nlike this:<\/p>\n<pre>{\n  \"name\": \"eat\",\n  \"match\": [\"#person\", \"#food\"],\n  \"when\": function(a, b) {\n    return a.properties.hungry &amp;&amp; !b.properties.eaten;\n  },\n  \"action\": function*(a, b) {\n    yield (new seaduck.StoryEvent(\"eat\", a, b));\n    a.properties.hungry = false;\n    b.properties.eaten = true;\n  }\n}\n<\/pre>\n<p>The value of the <code>when<\/code> attribute should be a function that takes one or two\n<code>noun<\/code>s as parameters and returns <code>true<\/code> or <code>false<\/code>. If this function returns\n<code>true<\/code>, then the action check succeeds, and the function associated with\nthe <code>action<\/code> attribute will be executed.<\/p>\n<p>Inside the <code>when<\/code> and <code>action<\/code> functions, <code>this<\/code> is bound to the Sea Duck\n<code>Narrative<\/code> object itself, to make it easy to call any of its methods. (Read\non for more information on the methods you can call.)<\/p>\n<p>The <code>name<\/code> attribute of an action is currently ignored (but might be used for\nsomething else in the future).<\/p>\n<h3><a id=\"user-content-matching\" class=\"anchor\" aria-hidden=\"true\" href=\"#matching\"><\/a>Matching<\/h3>\n<p>The <code>match<\/code> attribute of an <code>action<\/code> specifies which nouns or groups of nouns\nto will be given to the <code>when<\/code> and <code>action<\/code> functions. The list can contain\neither one or two items; if one item, then the <code>when<\/code> function will be called\nfor each matching noun(s) it turn. If two items, then the <code>when<\/code> function will\nbe called once for each pair of items from both matching lists, with the items\nfrom the corresponding lists being passed as the first and second parameters,\nrespectively. If the string in the list starts with <code>#<\/code>, then all nouns\nmatching the tag will be included; otherwise, only the noun whose name exactly\nmatches will be included.<\/p>\n<p>For example, the following <code>match<\/code> list:<\/p>\n<pre>\"match\": [\"#person\"]\n<\/pre>\n<p>... would cause the <code>when<\/code> function to be called once for every noun tagged\n\"person,\" while this <code>match<\/code> list:<\/p>\n<pre>\"match\": [\"#dog\", \"#cat\"]\n<\/pre>\n<p>... would cause the <code>when<\/code> function to be called for every possible pairing of\nnouns labelled <code>dog<\/code> and <code>cat<\/code>. Without the <code>#<\/code> sign, only the noun with\nexactly the matching name will be included, so:<\/p>\n<pre>\"match\": [\"Joe\", \"cupcake\"]\n<\/pre>\n<p>... would cause the <code>when<\/code> function to be called only once (with the noun named\n<code>Joe<\/code> and the noun named <code>cupcake<\/code> as parameters).<\/p>\n<h3><a id=\"user-content-yielding-actions\" class=\"anchor\" aria-hidden=\"true\" href=\"#yielding-actions\"><\/a>Yielding actions<\/h3>\n<p>Note that <em>updating parameters or manipulating relations<\/em> does not itself\ncreate narrative events. It's up to you to decide what changes to story state\n\"mean\" in terms of the events that they create.<\/p>\n<p>To that end, the <code>action<\/code> function should be a <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/JavaScript\/Reference\/Statements\/function*\" rel=\"nofollow\">generator\nfunction<\/a>\nthat <code>yield<\/code>s <code>StoryEvent<\/code> objects. (Generator functions are used so that\n<code>actions<\/code> can potentially generate multiple events; you need to put a <code>*<\/code> after\nthe word <code>function<\/code> in the function definition to make it a generator\nfunction.) The <code>StoryEvent<\/code> constructor takes at least one and up to three\nparameters, specifying the event's verb, the first noun, and the second noun.\nFor example, the following <code>action<\/code> function would yield three different\nevents:<\/p>\n<pre>\"action\": function*(a, b) {\n    yield new seaduck.StoryEvent(\"rain\"); \/\/ just verb\n    yield new seaduck.StoryEvent(\"sleep\", a); \/\/ verb + subject\n    yield new seaduck.StoryEvent(\"see\", a, b); \/\/ verb + subject + object\n}\n<\/pre>\n<h3><a id=\"user-content-discourse-rendering-events-with-tracery\" class=\"anchor\" aria-hidden=\"true\" href=\"#discourse-rendering-events-with-tracery\"><\/a>Discourse (rendering events with Tracery)<\/h3>\n<p>When \"rendering\" <code>StoryEvent<\/code>s as sentences, the name of the verb will be\nmatched to a Tracery rule. When rendering, the rules <code>nounA<\/code> and <code>nounB<\/code> will\nbe set to the corresponding nouns in the <code>StoryEvent<\/code>. For example, the\n<code>traceryDiscourse<\/code> section to render the events in the example above might look\nlike:<\/p>\n<pre>\"traceryDiscourse\": {\n    \"rain\": [\"It's raining.\"],\n    \"sleep\": [\"#nounA# falls asleep.\"],\n    \"see\": [\"#nounA# sees #nounB# out of the corner of their eye.\"]\n}\n<\/pre>\n<p>The properties of each noun are also available as rules in the Tracery grammar\nwith the names <code>nounA_&lt;propname&gt;<\/code> and <code>nounB_&lt;propname&gt;<\/code>, where <code>&lt;propname&gt;<\/code> is\nthe name of the property.<\/p>\n<h3><a id=\"user-content-relations\" class=\"anchor\" aria-hidden=\"true\" href=\"#relations\"><\/a>Relations<\/h3>\n<p>In Sea Duck, a relation is a named boolean value that relates one noun to\nanother. You can check for relations between nouns in a <code>when<\/code> function and\ncreate (or destroy) relations between nouns in an <code>action<\/code> function. For\nexample, if you wanted to create an <code>in_love<\/code> relation between two nouns in an\n<code>action<\/code> function, you might write:<\/p>\n<pre>this.relate(\"in_love\", a, b);\n<\/pre>\n<p>To check to see if two nouns are related to each other with a particular\nrelationship, use <code>this.isRelated(...)<\/code>:<\/p>\n<pre>if (this.isRelated(\"in_love\", a, b)) { ... }\n<\/pre>\n<p>The <code>this.reciprocal()<\/code> method creates a two-way relation (so that <code>a<\/code> is\n<code>in_love<\/code> with <code>b<\/code> and <code>b<\/code> is <code>in_love<\/code> with <code>a<\/code>):<\/p>\n<pre>this.reciprocal(\"in_love\", a, b);\n<\/pre>\n<p>You can erase a relation using <code>this.unrelate()<\/code> or <code>this.unreciprocal()<\/code>:<\/p>\n<pre>this.unrelate(\"in_love\", a, b); \/\/ a is no longer in_love with b\nthis.unreciprocal(\"in_love\", a, b); \/\/ a &amp; b are no longer in_love w\/each other\n<\/pre>\n<p>The function <code>this.allRelatedByTag(...)<\/code> function gives you a list of all\nobjects with the given tag that a noun is related to with the named relation.\nFor example, to find all nouns that <code>a<\/code> is <code>in_love<\/code> with tagged <code>person<\/code>:<\/p>\n<pre>this.allRelatedByTag(\"in_love\", a, \"person\") \/\/ array of related nouns\n<\/pre>\n<p>If you're pretty sure that there's only one related noun by tag, you can use\n<code>this.relatedByTag(...)<\/code> to only get the first result:<\/p>\n<pre>this.relatedByTag(\"in_love\", a, \"person\") \/\/ one noun object\n<\/pre>\n<h3><a id=\"user-content-initializing\" class=\"anchor\" aria-hidden=\"true\" href=\"#initializing\"><\/a>Initializing<\/h3>\n<p>The <code>initialize<\/code> attribute of the object passed to the <code>Narrative<\/code> constructor\nshould point to a generator function, which will be called when the <code>.step()<\/code>\nmethod is called before any actions are checked. This is a great place to put\nany code to initialize relations between objects or generate introductory\nstory events. For example, the following <code>initialize<\/code> function generates\nstory events for each <code>hungry<\/code> noun:<\/p>\n<pre>  \"initialize\": function*() {\n    for (let noun of this.getNounsByProperty(\"hungry\", true)) {\n      yield (new seaduck.StoryEvent(\"isHungry\", noun));\n    }\n  }\n<\/pre>\n<h3><a id=\"user-content-finding-noun-objects\" class=\"anchor\" aria-hidden=\"true\" href=\"#finding-noun-objects\"><\/a>Finding noun objects<\/h3>\n<p>Two handy methods for looking up noun objects:<\/p>\n<ul><li><code>this.getNounsByProperty(prop, val)<\/code> returns an array of noun objects for each\nnoun whose property named by <code>prop<\/code> is equal to <code>val<\/code>;<\/li>\n<li><code>this.noun(name)<\/code> returns the noun with the given name. (This is handy in\n<code>initialize<\/code> or advanced uses of <code>action<\/code> functions where you don't\nhave a reference to the object itself.)<\/li>\n<\/ul><h3><a id=\"user-content-when-narration-ends\" class=\"anchor\" aria-hidden=\"true\" href=\"#when-narration-ends\"><\/a>When narration ends<\/h3>\n<p>Narration ends when either of the following two conditions obtain:<\/p>\n<ul><li>No events are generated by any action<\/li>\n<li>The state of the narrative (i.e., noun properties and relations) did not\nchange from one step to the next<\/li>\n<\/ul><p>In both cases, Sea Duck inserts a special <code>StoryEvent<\/code> with the verb <code>_end<\/code> to\nindicate that the story is over. You can manually end a narration by yielding\nan event with this verb from an <code>action<\/code> function. Subsequent calls to\n<code>.step()<\/code> (and <code>.stepAndrender()<\/code>) will return empty arrays.<\/p>\n<p>A typical loop for stepping through the narration simulation might look like\nthis:<\/p>\n<pre>let n = seaduck.Narrative(...your narration specification here...);\nlet maxSteps = 100; \/\/ maximum number of steps to perform\nfor (let i = 0; i &lt; maxSteps; i++) {\n    let storyEvents = n.stepAndRender();\n    if (storyEvents.length &gt; 0) {\n        for (let ev of storyEvents) {\n            console.log(ev);\n        }\n    }\n    else {\n        break;\n    }\n}\n<\/pre>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<pre>Copyright 2018 Allison Parrish\nPermission is hereby granted, free of charge, to any person obtaining a copy of\nthis software and associated documentation files (the \"Software\"), to deal in\nthe Software without restriction, including without limitation the rights to\nuse, copy, modify, merge, publish, distribute, sublicense, and\/or sell copies\nof the Software, and to permit persons to whom the Software is furnished to do\nso, subject to the following conditions:\nThe above copyright notice and this permission notice shall be included in all\ncopies or substantial portions of the Software.\nTHE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\nSOFTWARE.\n<\/pre>\n<\/article>","created_at":"2018-11-07T16:22:03+0000","updated_at":"2018-11-07T16:22:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/125839?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":39,"title":"froggey\/Mezzano","url":"https:\/\/github.com\/froggey\/Mezzano","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/froggey\/Mezzano\/blob\/master\/doc\/screenshot1.png\"><img src=\"https:\/\/github.com\/froggey\/Mezzano\/raw\/master\/doc\/screenshot1.png\" alt=\"Screenshot\" \/><\/a><\/p>\n<h2><a id=\"user-content-pre-built-images\" class=\"anchor\" aria-hidden=\"true\" href=\"#pre-built-images\"><\/a>Pre-built images<\/h2>\n<p>Demo releases are available through <a href=\"https:\/\/github.com\/froggey\/Mezzano\/releases\">GitHub<\/a>.<\/p>\n<p>These releases are designed to be run in VirtualBox, though QEMU is also supported.\n2GB of RAM, a virtio-net NIC and an Intel HDA audio controller are recommended.<\/p>\n<h2><a id=\"user-content-building-from-source\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-from-source\"><\/a>Building from source<\/h2>\n<p>See the MBuild repo: (<a href=\"https:\/\/github.com\/froggey\/MBuild\">https:\/\/github.com\/froggey\/MBuild<\/a>)<\/p>\n<p>For help &amp; support, join #mezzano on Freenode (irc.freenode.net)<\/p>\n<h2><a id=\"user-content-major-changes-since-demo-3\" class=\"anchor\" aria-hidden=\"true\" href=\"#major-changes-since-demo-3\"><\/a>Major changes since Demo 3<\/h2>\n<ul><li>FAT32 support has been implemented by Bruno Cichon (ebrasca).<\/li>\n<li>McCLIM has been ported by fittestbits.<\/li>\n<li>Quicklisp has been ported by Peter S. Housel.<\/li>\n<li>Improved introspection tools: DISASSEMBLE and ED have been implemented.<\/li>\n<li>Generational collection has been added to the garbage collector.<\/li>\n<li>New SSA-based compiler backend, supporting unboxed value representations.<\/li>\n<li>Gray streams support has been overhauled.<\/li>\n<\/ul><h2><a id=\"user-content-major-changes-since-demo-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#major-changes-since-demo-2\"><\/a>Major changes since Demo 2<\/h2>\n<ul><li>Trentino, a media player, has been implemented by Eugene Zaikonnikov.<\/li>\n<li>Further improvements to conformance, stability and performance.<\/li>\n<li>The CLOS implementation follows the MOP much more closely.<\/li>\n<li>More traditional window management.<\/li>\n<li>Booting from CD\/USB on real hardware is now possible.<\/li>\n<li>Driver support for Intel HDA audio devices.<\/li>\n<li>VirtualBox guest (mouse &amp; display) integration.<\/li>\n<\/ul><h2><a id=\"user-content-major-changes-since-demo-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#major-changes-since-demo-1\"><\/a>Major changes since Demo 1<\/h2>\n<ul><li>Many improvements to conformance, stability and performance.<\/li>\n<li>The editor has been greatly improved, thanks to Burton Samograd.<\/li>\n<li>The system now functions correctly on computers with more than 1GB of RAM.<\/li>\n<li>The allocator and garbage collector now make much better use of available memory, with far fewer GC cycles occuring.<\/li>\n<li>(ROOM T) prints more detailed information about allocated objects.<\/li>\n<li>Transparency and premultiplied alpha support in the GUI.<\/li>\n<li>And more!<\/li>\n<\/ul><h2><a id=\"user-content-additional-information\" class=\"anchor\" aria-hidden=\"true\" href=\"#additional-information\"><\/a>Additional information<\/h2>\n<p>\"Hypothymis azurea - Kaeng Krachan\" by JJ Harrison (<a href=\"mailto:jjharrison89@facebook.com\">jjharrison89@facebook.com<\/a>)\n[CC BY-SA 3.0 (<a href=\"http:\/\/creativecommons.org\/licenses\/by-sa\/3.0\" rel=\"nofollow\">http:\/\/creativecommons.org\/licenses\/by-sa\/3.0<\/a>)], via Wikimedia Commons\n<a href=\"https:\/\/commons.wikimedia.org\/wiki\/File:Hypothymis_azurea_-_Kaeng_Krachan.jpg\" rel=\"nofollow\">https:\/\/commons.wikimedia.org\/wiki\/File:Hypothymis_azurea_-_Kaeng_Krachan.jpg<\/a><\/p>\n<p>\"Mandarin Pair\" by \u00a9 Francis C. Franklin \/ CC-BY-SA-3.0.\nLicensed under CC BY-SA 3.0 via Wikimedia Commons - <a href=\"http:\/\/commons.wikimedia.org\/wiki\/File:Mandarin_Pair.jpg\" rel=\"nofollow\">http:\/\/commons.wikimedia.org\/wiki\/File:Mandarin_Pair.jpg<\/a><\/p>\n<p>\"Handsome\" by Andy Morffew - <a href=\"https:\/\/www.flickr.com\/photos\/andymorffew\/19377769093\/in\/album-72157630893775092\/\" rel=\"nofollow\">https:\/\/www.flickr.com\/photos\/andymorffew\/19377769093\/in\/album-72157630893775092\/<\/a>\n[CC BY 2.0 (<a href=\"http:\/\/creativecommons.org\/licenses\/by\/2.0\" rel=\"nofollow\">http:\/\/creativecommons.org\/licenses\/by\/2.0<\/a>)]<\/p>\n<p>Includes Dejavu Fonts 2.37 (<a href=\"http:\/\/dejavu-fonts.org\/\" rel=\"nofollow\">http:\/\/dejavu-fonts.org\/<\/a>)<\/p>\n<p>Some icons from Icojam (<a href=\"http:\/\/www.icojam.com\" rel=\"nofollow\">http:\/\/www.icojam.com<\/a>)<\/p>\n<\/article>","created_at":"2018-11-07T16:22:03+0000","updated_at":"2018-11-07T16:22:03+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/72612?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":37,"title":"wordpress rest api custom endpoint don't serialize response at DuckDuckGo","url":"https:\/\/duckduckgo.com\/?ia=web&q=wordpress+rest+api+custom+endpoint+don%27t+serialize+response&t=ffab","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:22:02+0000","updated_at":"2018-11-07T16:22:02+0000","annotations":[],"reading_time":0,"domain_name":"duckduckgo.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":36,"title":"My Journey into Fractals | Hacker News","url":"https:\/\/news.ycombinator.com\/item?id=18146761","content":"<table border=\"0\" cellpadding=\"0\" style=\"border-spacing: 0px; width: 100%;\"><tr><td><a href=\"https:\/\/news.ycombinator.com\"><img src=\"https:\/\/news.ycombinator.com\/y18.gif\" width=\"18\" height=\"18\" alt=\"image\" \/><\/a><\/td>\n                  <td><b class=\"hnname\"><a href=\"https:\/\/news.ycombinator.com\/news\">Hacker News<\/a><\/b>\n              <a href=\"https:\/\/news.ycombinator.com\/newest\">new<\/a> | <a href=\"https:\/\/news.ycombinator.com\/newcomments\">comments<\/a> | <a href=\"https:\/\/news.ycombinator.com\/show\">show<\/a> | <a href=\"https:\/\/news.ycombinator.com\/ask\">ask<\/a> | <a href=\"https:\/\/news.ycombinator.com\/jobs\">jobs<\/a> | <a href=\"https:\/\/news.ycombinator.com\/submit\">submit<\/a>            <\/td><td>\n                              <a href=\"https:\/\/news.ycombinator.com\/login?goto=item%3Fid%3D18146761\">login<\/a>\n                          <\/td>\n              <\/tr><\/table><table class=\"fatitem\" border=\"0\"><tr class=\"athing\" id=\"18146761\"><td align=\"right\" valign=\"top\" class=\"title\">      <\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"title\"><a href=\"https:\/\/medium.com\/@bananaft\/my-journey-into-fractals-d25ebc6c4dc2\" class=\"storylink\">My Journey into Fractals<\/a> (<a href=\"https:\/\/news.ycombinator.com\/from?site=medium.com\">medium.com<\/a>)<\/td><\/tr><tr><td colspan=\"2\"><\/td><td class=\"subtext\">\n        366 points by <a href=\"https:\/\/news.ycombinator.com\/user?id=jypepin\" class=\"hnuser\">jypepin<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18146761\">33 days ago<\/a>  | <a href=\"https:\/\/news.ycombinator.com\/hide?id=18146761&amp;goto=item%3Fid%3D18146761\">hide<\/a> | <a href=\"https:\/\/hn.algolia.com\/?query=My%20Journey%20into%20Fractals&amp;sort=byDate&amp;dateRange=all&amp;type=story&amp;storyText=false&amp;prefix&amp;page=0\" class=\"hnpast\">past<\/a> | <a href=\"https:\/\/www.google.com\/search?q=My%20Journey%20into%20Fractals\">web<\/a> | <a href=\"https:\/\/news.ycombinator.com\/fave?id=18146761&amp;auth=4127bfbfc78c3a0d01a364060938782ce57fa624\">favorite<\/a> | <a href=\"https:\/\/news.ycombinator.com\/item?id=18146761\">38\u00a0comments<\/a>              <\/td><\/tr><\/table><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"10\" width=\"0\" alt=\"image\" \/><br \/><div style=\"text-align: center;\"><a href=\"https:\/\/news.ycombinator.com\/newsguidelines.html\">Guidelines<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/newsfaq.html\">FAQ<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Support<\/a>\n        | <a href=\"https:\/\/github.com\/HackerNews\/API\">API<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/security.html\">Security<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/lists\">Lists<\/a>\n        | <a href=\"https:\/\/news.ycombinator.com\/bookmarklet.html\" rel=\"nofollow\">Bookmarklet<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/legal\/\">Legal<\/a>\n        | <a href=\"http:\/\/www.ycombinator.com\/apply\/\">Apply to YC<\/a>\n        | <a href=\"mailto:hn@ycombinator.com\">Contact<\/a><form method=\"get\" action=\"\/\/hn.algolia.com\/\"><div>Search:\n          <\/div><\/form>\n            <\/div>","created_at":"2018-11-07T16:22:01+0000","updated_at":"2018-11-07T16:22:01+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":34,"title":"Apache vs Nginx: Practical Considerations","url":"https:\/\/www.digitalocean.com\/community\/tutorials\/apache-vs-nginx-practical-considerations","content":"<p>Apache and Nginx are the two most common open source web servers in the world.  Together, they are responsible for serving over 50% of traffic on the internet.  Both solutions are capable of handling diverse workloads and working with other software to provide a complete web stack.<\/p><p>While Apache and Nginx share many qualities, they should not be thought of as entirely interchangeable.  Each excels in its own way and it is important to understand the situations where you may need to reevaluate your web server of choice.  This article will be devoted to a discussion of how each server stacks up in various areas.<\/p><h2 id=\"general-overview\">General Overview<\/h2><p>Before we dive into the differences between Apache and Nginx, let's take a quick look at the background of these two projects and their general characteristics.<\/p><h3 id=\"apache\">Apache<\/h3><p>The Apache HTTP Server was created by Robert McCool in 1995 and has been developed under the direction of the Apache Software Foundation since 1999.  Since the HTTP web server is the foundation's original project and is by far their most popular piece of software, it is often referred to simply as \"Apache\".<\/p><p>The Apache web server has been the most popular server on the internet since 1996.  Because of this popularity, Apache benefits from great documentation and integrated support from other software projects.<\/p><p>Apache is often chosen by administrators for its flexibility, power, and widespread support.  It is extensible through a dynamically loadable module system and can process a large number of interpreted languages without connecting out to separate software.<\/p><h3 id=\"nginx\">Nginx<\/h3><p>In 2002, Igor Sysoev began work on Nginx as an answer to the C10K problem, which was a challenge for web servers to begin handling ten thousand concurrent connections as a requirement for the modern web.  The initial public release was made in 2004, meeting this goal by relying on an asynchronous, events-driven architecture.<\/p><p>Nginx has grown in popularity since its release due to its light-weight resource utilization and its ability to scale easily on minimal hardware.  Nginx excels at serving static content quickly and is designed to pass dynamic requests off to other software that is better suited for those purposes.<\/p><p>Nginx is often selected by administrators for its resource efficiency and responsiveness under load.  Advocates welcome Nginx's focus on core web server and proxy features.<\/p><h2 id=\"connection-handling-architecture\">Connection Handling Architecture<\/h2><p>One big difference between Apache and Nginx is the actual way that they handle connections and traffic.  This provides perhaps the most significant difference in the way that they respond to different traffic conditions.<\/p><h3>Apache<\/h3><p>Apache provides a variety of multi-processing modules (Apache calls these MPMs) that dictate how client requests are handled.  Basically, this allows administrators to swap out its connection handling architecture easily.  These are:<\/p><ul><li><strong>mpm_prefork<\/strong>: This processing module spawns processes with a single thread each to handle request.  Each child can handle a single connection at a time.  As long as the number of requests is fewer than the number of processes, this MPM is very fast.  However, performance degrades quickly after the requests surpass the number of processes, so this is not a good choice in many scenarios.  Each process has a significant impact on RAM consumption, so this MPM is difficult to scale effectively.  This may still be a good choice though if used in conjunction with other components that are not built with threads in mind.  For instance, PHP is not thread-safe, so this MPM is recommended as the only safe way of working with <code>mod_php<\/code>, the Apache module for processing these files.<\/li>\n<li><strong>mpm_worker<\/strong>: This module spawns processes that can each manage multiple threads.  Each of these threads can handle a single connection.  Threads are much more efficient than processes, which means that this MPM scales better than the prefork MPM.  Since there are more threads than processes, this also means that new connections can immediately take a free thread instead of having to wait for a free process.<\/li>\n<li><strong>mpm_event<\/strong>: This module is similar to the worker module in most situations, but is optimized to handle keep-alive connections.  When using the worker MPM, a connection will hold a thread regardless of whether a request is actively being made for as long as the connection is kept alive.  The event MPM handles keep alive connections by setting aside dedicated threads for handling keep alive connections and passing active requests off to other threads.  This keeps the module from getting bogged down by keep-alive requests, allowing for faster execution.  This was marked stable with the release of Apache 2.4.<\/li>\n<\/ul><p>As you can see, Apache provides a flexible architecture for choosing different connection and request handling algorithms.  The choices provided are mainly a function of the server's evolution and the increasing need for concurrency as the internet landscape has changed.<\/p><h3>Nginx<\/h3><p>Nginx came onto the scene after Apache, with more awareness of the concurrency problems that would face sites at scale.  Leveraging this knowledge, Nginx was designed from the ground up to use an asynchronous, non-blocking, event-driven connection handling algorithm.<\/p><p>Nginx spawns worker processes, each of which can handle thousands of connections.  The worker processes accomplish this by implementing a fast looping mechanism that continuously checks for and processes events.  Decoupling actual work from connections allows each worker to concern itself with a connection only when a new event has been triggered.<\/p><p>Each of the connections handled by the worker are placed within the event loop where they exist with other connections.  Within the loop, events are processed asynchronously, allowing work to be handled in a non-blocking manner.  When the connection closes, it is removed from the loop.  <\/p><p>This style of connection processing allows Nginx to scale incredibly far with limited resources.  Since the server is single-threaded and processes are not spawned to handle each new connection, the memory and CPU usage tends to stay relatively consistent, even at times of heavy load.<\/p><h2 id=\"static-vs-dynamic-content\">Static vs Dynamic Content<\/h2><p>In terms of real world use-cases, one of the most common comparisons between Apache and Nginx is the way in which each server handles requests for static and dynamic content.<\/p><h3>Apache<\/h3><p>Apache servers can handle static content using its conventional file-based methods.  The performance of these operations is mainly a function of the MPM methods described above.<\/p><p>Apache can also process dynamic content by embedding a processor of the language in question into each of its worker instances.  This allows it to execute dynamic content within the web server itself without having to rely on external components.  These dynamic processors can be enabled through the use of dynamically loadable modules.<\/p><p>Apache's ability to handle dynamic content internally means that configuration of dynamic processing tends to be simpler.  Communication does not need to be coordinated with an additional piece of software and modules can easily be swapped out if the content requirements change.<\/p><h3>Nginx<\/h3><p>Nginx does not have any ability to process dynamic content natively.  To handle PHP and other requests for dynamic content, Nginx must pass to an external processor for execution and wait for the rendered content to be sent back.  The results can then be relayed to the client.<\/p><p>For administrators, this means that communication must be configured between Nginx and the processor over one of the protocols Nginx knows how to speak (http, FastCGI, SCGI, uWSGI, memcache).  This can complicate things slightly, especially when trying to anticipate the number of connections to allow, as an additional connection will be used for each call to the processor.<\/p><p>However, this method has some advantages as well.  Since the dynamic interpreter is not embedded in the worker process, its overhead will only be present for dynamic content.  Static content can be served in a straight-forward manner and the interpreter will only be contacted when needed.  Apache can also function in this manner, but doing so removes the benefits in the previous section.<\/p><h2 id=\"distributed-vs-centralized-configuration\">Distributed vs Centralized Configuration<\/h2><p>For administrators, one of the most readily apparent differences between these two pieces of software is whether directory-level configuration is permitted within the content directories.<\/p><h3>Apache<\/h3><p>Apache includes an option to allow additional configuration on a per-directory basis by inspecting and interpreting directives in hidden files within the content directories themselves.  These files are known as <code>.htaccess<\/code> files.<\/p><p>Since these files reside within the content directories themselves, when handling a request, Apache checks each component of the path to the requested file for an <code>.htaccess<\/code> file and applies the directives found within.  This effectively allows decentralized configuration of the web server, which is often used for implementing URL rewrites, access restrictions, authorization and authentication, even caching policies.<\/p><p>While the above examples can all be configured in the main Apache configuration file, <code>.htaccess<\/code> files have some important advantages.  First, since these are interpreted each time they are found along a request path, they are implemented immediately without reloading the server.  Second, it makes it possible to allow non-privileged users to control certain aspects of their own web content without giving them control over the entire configuration file.<\/p><p>This provides an easy way for certain web software, like content management systems, to configure  their environment without providing access to the central configuration file.  This is also used by shared hosting providers to retain control of the main configuration while giving clients control over their specific directories.<\/p><h3>Nginx<\/h3><p>Nginx does not interpret <code>.htaccess<\/code> files, nor does it provide any mechanism for evaluating per-directory configuration outside of the main configuration file.  This may be less flexible than the Apache model, but it does have its own advantages.<\/p><p>The most notable improvement over the <code>.htaccess<\/code> system of directory-level configuration is increased performance.  For a typical Apache setup that may allow <code>.htaccess<\/code> in any directory, the server will check for these files in <em>each<\/em> of the parent directories leading up to the requested file, for each request.  If one or more <code>.htaccess<\/code> files are found during this search, they must be read and interpreted.  By not allowing directory overrides, Nginx can serve requests faster by<br \/>doing a single directory lookup and file read for each request (assuming that the file is found in the conventional directory structure).<\/p><p>Another advantage is security related.  Distributing directory-level configuration access also distributes the responsibility of security to individual users, who may not be trusted to handle this task well.  Ensuring that the administrator maintains control over the entire web server can prevent some security missteps that may occur when access is given to other parties.<\/p><p>Keep in mind that it is possible to turn off <code>.htaccess<\/code> interpretation in Apache if these concerns resonate with you.<\/p><h2 id=\"file-vs-uri-based-interpretation\">File vs URI-Based Interpretation<\/h2><p>How the web server interprets requests and maps them to actual resources on the system is another area where these two servers differ.<\/p><h3>Apache<\/h3><p>Apache provides the ability to interpret a request as a physical resource on the filesystem or as a URI location that may need a more abstract evaluation.  In general, for the former Apache uses <code>&lt;Directory&gt;<\/code> or <code>&lt;Files&gt;<\/code> blocks, while it utilizes <code>&lt;Location&gt;<\/code> blocks for more abstract resources.<\/p><p>Because Apache was designed from the ground up as a web server, the default is usually to interpret requests as filesystem resources.  It begins by taking the document root and appending the portion of the request following the host and port number to try to find an actual file.  Basically, the filesystem hierarchy is represented on the web as the available document tree.<\/p><p>Apache provides a number of alternatives for when the request does not match the underlying filesystem.  For instance, an <code>Alias<\/code> directive can be used to map to an alternative location.  Using <code>&lt;Location&gt;<\/code> blocks is a method of working with the URI itself instead of the filesystem.  There are also regular expression variants which can be used to apply configuration more flexibly throughout the filesystem.<\/p><p>While Apache has the ability to operate on both the underlying filesystem and the webspace, it leans heavily towards filesystem methods.  This can be seen in some of the design decisions, including the use of <code>.htaccess<\/code> files for per-directory configuration.  The <a href=\"http:\/\/httpd.apache.org\/docs\/2.4\/sections.html#whichwhen\">Apache docs<\/a> themselves warn against using URI-based blocks to restrict access when the request mirrors the underlying filesystem.<\/p><h3>Nginx<\/h3><p>Nginx was created to be both a web server and a proxy server.  Due to the architecture required for these two roles, it works primarily with URIs, translating to the filesystem when necessary.<\/p><p>This can be seen in some of the ways that Nginx configuration files are constructed and interpreted.Nginx does not provide a mechanism for specifying configuration for a filesystem directory and instead parses the URI itself.<\/p><p>For instance, the primary configuration blocks for Nginx are <code>server<\/code> and <code>location<\/code> blocks.  The <code>server<\/code> block interprets the host being requested, while the <code>location<\/code> blocks are responsible for matching portions of the URI that comes after the host and port.  At this point, the request is being interpreted as a URI, not as a location on the filesystem.<\/p><p>For static files, all requests eventually have to be mapped to a location on the filesystem. First, Nginx selects the server and location blocks that will handle the request and then combines the document root with the URI, adapting anything necessary according to the configuration specified.<\/p><p>This may seem similar, but parsing requests primarily as URIs instead of filesystem locations allows Nginx to more easily function in both web, mail, and proxy server roles.  Nginx is configured simply by laying out how to respond to different request patterns.  Nginx does not check the filesystem until it is ready to serve the request, which explains why it does not implement a form of <code>.htaccess<\/code> files.<\/p><h2 id=\"modules\">Modules<\/h2><p>Both Nginx and Apache are extensible through module systems, but the way that they work differ significantly.<\/p><h3>Apache<\/h3><p>Apache's module system allows you to dynamically load or unload modules to satisfy your needs during the course of running the server.  The Apache core is always present, while modules can be turned on or off, adding or removing additional functionality and hooking into the main server.<\/p><p>Apache uses this functionality for a large variety tasks.  Due to the maturity of the platform, there is an extensive library of modules available.  These can be used to alter some of the core functionality of the server, such as <code>mod_php<\/code>, which embeds a PHP interpreter into each running worker.<\/p><p>Modules are not limited to processing dynamic content, however.  Among other functions, they can be used used for rewriting URLs, authenticating clients, hardening the server, logging, caching, compression, proxying, rate limiting, and encrypting.  Dynamic modules can extend the core functionality considerably without much additional work.<\/p><h3>Nginx<\/h3><p>Nginx also implements a module system, but it is quite different from the Apache system.  In Nginx, modules are not dynamically loadable, so they must be selected and compiled into the core software.<\/p><p>For many users, this will make Nginx much less flexible.  This is especially true for users who are not comfortable maintaining their own compiled software outside of their distribution's conventional packaging system.  While distributions' packages tend to include the most commonly used modules, if you require a non-standard module, you will have to build the server from source yourself.<\/p><p>Nginx modules are still very useful though, and they allow you to dictate what you want out of your server by only including the functionality you intend to use.  Some users also may consider this more secure, as arbitrary components cannot be hooked into the server.  However, if your server is ever put in a position where this is possible, it is likely compromised already.<\/p><p>Nginx modules allow many of the same capabilities as Apache modules.  For instance, Nginx modules can provide proxying support, compression, rate limiting, logging, rewriting, geolocation, authentication, encryption, streaming, and mail functionality.<\/p><h2 id=\"support-compatibility-ecosystem-and-documentation\">Support, Compatibility, Ecosystem, and Documentation<\/h2><p>A major point to consider is what the actual process of getting up and running will be given the landscape of available help and support among other software.<\/p><h3>Apache<\/h3><p>Because Apache has been popular for so long, support for the server is fairly ubiquitous.  There is a large library of first- and third-party documentation available for the core server and for task-based scenarios involving hooking Apache up with other software.<\/p><p>Along with documentation, many tools and web projects include tools to bootstrap themselves within an Apache environment.  This may be included in the projects themselves, or in the packages maintained by your distribution's packaging team.<\/p><p>Apache, in general, will have more support from third-party projects simply because of its market share and the length of time it has been available.  Administrators are also somewhat more likely to have experience working with Apache not only due to its prevalence, but also because many people start off in shared-hosting scenarios which almost exclusively rely on Apache due to the <code>.htaccess<\/code> distributed management capabilities.<\/p><h3>Nginx<\/h3><p>Nginx is experiencing increased support as more users adopt it for its performance profile, but it still has some catching up to do in some key areas.<\/p><p>In the past, it was difficult to find comprehensive English-language documentation regarding Nginx due to the fact that most of the early development and documentation were in Russian.  As interest in the project grew, the documentation has been filled out and there are now plenty of administration resources on the Nginx site and through third parties.<\/p><p>In regards to third-party applications, support and documentation is becoming more readily available, and package maintainers are beginning, in some cases, to give choices between auto-configuring for Apache and Nginx.  Even without support, configuring Nginx to work with alternative software is usually straight-forward so long as the project itself documents its requirements (permissions, headers, etc).<\/p><h2 id=\"using-apache-and-nginx-together\">Using Apache and Nginx Together<\/h2><p>After going over the benefits and limitations of both Apache and Nginx, you may have a better idea of which server is more suited to your needs.  However, many users find that it is possible to leverage each server's strengths by using them together.<\/p><p>The conventional configuration for this partnership is to place Nginx in front of Apache as a reverse proxy.  This will allow Nginx to to handle all requests from clients.  This takes advantage of Nginx's fast processing speed and ability to handle large numbers of connections concurrently.<\/p><p>For static content, which Nginx excels at, the files will be served quickly and directly to the client.  For dynamic content, for instance PHP files, Nginx will proxy the request to Apache, which can then process the results and return the rendered page.  Nginx can then pass the content back to the client.<\/p><p>This setup works well for many people because it allows Nginx to function as a sorting machine.  It will handle all requests it can and pass on the ones that it has no native ability to serve.  By cutting down on the requests the Apache server is asked to handle, we can alleviate some of the blocking that occurs when an Apache process or thread is occupied.<\/p><p>This configuration also allows you to scale out by adding additional backend servers as necessary.  Nginx can be configured to pass to a pool of servers easily, increasing this configuration's resilience to failure and performance.<\/p><h2 id=\"conclusion\">Conclusion<\/h2><p>As you can see, both Apache and Nginx are powerful, flexible, and capable.  Deciding which server is best for you is largely a function of evaluating your specific requirements and testing with the patterns that you expect to see.<\/p><p>There are differences between these projects that have a very real impact on the raw performance, capabilities, and the implementation time necessary to get each solution up and running.  However, these usually are the result of a series of trade offs that should not be casually dismissed.  In the end, there is no one-size-fits-all web server, so use the solution that best aligns with your objectives.<\/p>","created_at":"2018-11-07T16:22:00+0000","updated_at":"2018-11-07T16:22:00+0000","published_at":"2015-01-28T15:33:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":16,"domain_name":"www.digitalocean.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":35,"title":"How To Configure Content Caching Using Apache Modules On A VPS","url":"https:\/\/www.digitalocean.com\/community\/tutorials\/how-to-configure-content-caching-using-apache-modules-on-a-vps","content":"<h3>What is Web Caching<\/h3><p>Web caching is a way of improving server performance by allowing commonly requested content to be stored in an easier to access location.  This allows the visitor to access the content faster instead of having to fetch the same data multiple times.<\/p><p>By effectively creating caching rules, content that is suitable for caching will be stored to conserve resources, while highly-dynamic content will be served normally.  In this guide, we will be discussing how to configure Apache using its caching modules.<\/p><p><strong>Note<\/strong>: This guide was written with Apache 2.2 in mind.  Changes to Apache 2.4 have lead to the replacement of some of the modules discussed in this guide.  Due to this consideration, not all of the steps recommended below will work on Apache 2.4 installations.<\/p><h2>An Introduction to Caching in Apache<\/h2><p>Apache has a number of different methods of caching content that is frequently accessed.  The two most common modules that enable this functionality are called \"mod_cache\" and \"mod_file_cache\".<\/p><h3>The mod_file_cache Module<\/h3><p>The mod_file_cache module is the simpler of the two caching mechanisms.  It works by caching content that:<\/p><ul><li>Is requested frequently<\/li>\n\t<li>Changes <em>very<\/em> infrequently<\/li>\n<\/ul><p>If these two requirements are met, then mod_file_cache may be useful.  It works by performing some of the file access operations on commonly used files when the server is started.<\/p><h3>The mod_cache Module<\/h3><p>The mod_cache module provides HTTP-aware caching schemes.  This means that the files will be cached according to an instruction specifying how long a page can be considered \"fresh\".<\/p><p>It performs these operations using either the \"mod_mem_cache\" module or the \"mod_disk_cache\" modules.  These are more complex caching models than mod_file_cache and are more useful in most circumstances.<\/p><h2>Using mod_file_cache with Apache<\/h2><p>The mod_file_cache module is useful to cache files that will not change for the life of the current Apache instance.  The techniques used with this module will cause any subsequent changes to not be applied until the server is restarted.<\/p><p>These caching mechanisms can only be used with normal files, so no dynamically generated content or files generated by special content handlers will work here.<\/p><p>The module provides two directives that are used to accomplish caching in different ways.<\/p><h3>MMapFile<\/h3><p>MMapFile is a directive used to create a list of files and then map those files into memory.  This is done only at server start up, so it is essential that none of the files set to use this type of caching are changed.<\/p><p>You can set up this type of caching in the server configuration file.  This is done by specifying files to be cached in memory in a space-separated list:<\/p><pre>MMapFile \/var\/www\/index.html \/var\/www\/otherfile.html var\/www\/static-image.jpg<\/pre><p>These files will be held in memory and served from there when the resource is requested.  If any of the files are changed, you need to restart the server.<\/p><h3>CacheFile<\/h3><p>This directive works by opening handles to the files listed.  It maintains a table of these open file descriptors and uses it to cut time it takes to open these files.<\/p><p>Again, changes to the file during operation of the server will not be recognized by the cache.  The original contents will continue to be served until the server is restarted.<\/p><p>This directive is used by specifying a space-separated list of files that should be cached with this method:<\/p><pre>CacheFile \/this\/file.html that\/file.html another\/file\/to\/server.html<\/pre><p>This will cause these files to be cached on server start.<\/p><h2>Using mod_cache with Apache<\/h2><p>The mod_cache module is a more flexible and powerful caching module.  It functions by implementing HTTP-aware caching of commonly accessed files.<\/p><p>While all caching mechanisms rely on serving files in some persistent state, mod_cache can handle changing content by configuring how long a file is valid for caching.<\/p><p>The module relies on two other modules to do the majority of the cache implementation.  These are \"mod_disk_cache\" and \"mod_mem_cache\".<\/p><p>The difference between these two is where the cache is kept, either on disk or in memory respectively.  These are stored and retrieved using URI based keys.  This is important to note as you can improve the caching of your site by turning on canonical naming.<\/p><p>This can be accomplished by putting this directive in the server configuration or virtual host definition:<\/p><pre>UseCanonicalName On<\/pre><h2>How to Configure Caching<\/h2><p>We will examine some common configuration directives and how they affect the functionality of the caching mechanisms.<\/p><p>If you look in the \"\/etc\/apache2\/mods-available\" directory, you can see some of the default configuration files for these modules.<\/p><h3>Configuring mod_mem_cache<\/h3><p>Let's look at the mod_mem_cache configuration:<\/p><pre>sudo nano \/etc\/apache2\/mods-available\/mem_cache.conf<\/pre><pre>&lt;IfModule mod_mem_cache.c&gt;&#13;\n\tCacheEnable mem \/&#13;\n\tMCacheSize 4096&#13;\n\tMCacheMaxObjectCount 100&#13;\n\tMCacheMinObjectSize 1&#13;\n\tMCacheMaxObjectSize 2048&#13;\n&lt;\/IfModule&gt;<\/pre>These directives are only read if the mod_mem_cache module is loaded.  This can be done by typing the following:<pre>sudo a2enmod mem_cache&#13;\nsudo service apache2 restart<\/pre>This will enable mod_mem_cache and also mod_cache.<pre>CacheEnable mem \/<\/pre><p>The \"CacheEnable mem \/\" line tells apache to create a memory cache for contents stored under \"\/\" (which is everything).<\/p><pre>MCacheSize 4096&#13;\nMCacheMaxObjectCount 100<\/pre><p>The next few lines describe the total size of the cache and the kinds of objects that will be stored.  The \"MCacheSize\" directive and the \"MCacheMaxOjectCount\" directive both describe the maximum size of the cache, first in terms of memory usage, and then in terms of the maximum amount of objects.<\/p><pre>MCacheMinObjectSize 1&#13;\nMCacheMaxObjectSize 2048<\/pre><p>The next two lines describe the kinds of data that will be cached, in terms of memory usage.  The default values specify that files between 1 byte and 2 kilobytes will be considered for caching.<\/p><h3>Configuring mod_disk_cache<\/h3><p>We can learn about a different set of directives by examining the mod_disk_cache configuration file:<\/p><pre>sudo nano \/etc\/apache2\/mods-available\/disk_cache.conf<\/pre><pre>&lt;IfModule mod_disk_cache.c&gt;&#13;\n\tCacheRoot \/var\/cache\/apache2\/mod_disk_cache&#13;\n\t#CacheEnable disk \/&#13;\n\tCacheDirLevels 5&#13;\n\tCacheDirLength 3&#13;\n&lt;\/IfModule&gt;<\/pre><p>This configuration is loaded if you enable the mod_disk_cache module, which can be done by typing:<\/p><pre>sudo a2enmod disk_cache&#13;\nsudo service apache2 restart<\/pre><p>This command will also enable mod_cache in order to work properly.&#13;\n&#13;\n<\/p><pre>CacheRoot \/var\/cache\/apache2\/mod_disk_cache&#13;\n#CacheEnable disk \/<\/pre><p>The \"CacheRoot\" directive specifies where the cached content will be kept.  The \"CacheEnable disk \/\" directive is disabled by default.  It is suggested that you enable this on a virtual host basis to get a better idea of what will be cached.<\/p><pre>CacheDirLevels 5&#13;\nCacheDirLength 3<\/pre><p>The other two directives determine the caching structure within the cache root.  Each cached element is hashed by the its URL, and then the hash is used as a filename and directory path.<\/p><p>The CacheDirLevel decides how many directories to create from the hash string and the CacheDirLength decides how many characters are in each directory name.<\/p><p>For example, if you have a file that hashes to \"abcdefghijklmnopqrstuvwxyz\", then a CacheDirLevel of 2 and a CacheDirLength of 4 would lead to this file being stored in:<\/p><pre>[path_of_cache_root]\/abcd\/efgh\/ijklmnopqrstuv<\/pre><p>Caching that is stored on disk can become large depending on the expiration dates of the content.  Apache includes a tool called \"htcacheclean\" to pare the cache down to a configured size.  This is outside the scope of this guide.<\/p><h2>Using CacheLock to Avoid Overwhelming the Backend<\/h2><p>A problem can arise on a busy server when a cached resource expires.<\/p><p>The cache that needs to be refreshed will have to refetch the file from the normal file resource.  During this time, if there are more requests for the file.  This can create a huge spike in requests to the backend server as the cached version is being refreshed.<\/p><p>To avoid this situation, it is possible to enable a lock file that indicates that the resource is being recached and that subsequent requests should not go to the backend, because the issue is being addressed.<\/p><p>This lock can prevent apache from trying to cache the same resource multiple times when first caching.  It also will serve the stale resource until the refreshed cache is complete.<\/p><p>Three directives are used to control CacheLock:<\/p><pre>CacheLock [ On | Off ]&#13;\nCacheLockMaxAge [time_in_seconds]&#13;\nCacheLockPath [\/path\/to\/lock\/directory]<\/pre><p>The first directive turns on the feature and the third directive establishes the directory where resource locks will be created.<\/p><p>The second directive, CacheLockMaxAge, is used to establish the longest time in seconds that a lock file will be considered valid.  This is important in case there is a failure or an abnormal delay in refreshing a resource.<\/p><h2>Conclusion<\/h2><p>Caching in Apache can be simple or involved depending on your needs.  While any kind of caching can improve your site performance, it is important to test your configurations to ensure that they are operating correctly.<\/p><p>It is also essential that you are familiar with the repercussions of improperly configured caching.  It is sometimes necessary to re-evaluate your security practices after implementing caching to ensure that private resources are not accidentally being cached for public consumption.<\/p><p>The apache user documentation has plenty of information about <a href=\"http:\/\/httpd.apache.org\/docs\/2.2\/caching.html\">how to configure caching<\/a> if you get stuck.  Even if you have a handle on the configuration, it is a helpful reference and good resource.<\/p><p>By Justin Ellingwood<\/p>","created_at":"2018-11-07T16:22:00+0000","updated_at":"2018-11-07T16:22:00+0000","published_at":"2013-08-16T19:40:38+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"www.digitalocean.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":31,"title":"heyscrumpy\/tiptap","url":"https:\/\/github.com\/heyscrumpy\/tiptap","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p>A renderless and extendable rich-text editor for <a href=\"https:\/\/github.com\/vuejs\/vue\">Vue.js<\/a><\/p>\n<p><a href=\"https:\/\/www.npmjs.com\/package\/tiptap\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/45640c7935966691ef69c2b4401a7ccb18dfc059\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f7469707461702e7376673f6c6162656c3d76657273696f6e\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/tiptap.svg?label=version\" \/><\/a>\n<a href=\"https:\/\/npmcharts.com\/compare\/tiptap?minimal=true\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/300517580b224509195daeda531d3a8951b65376\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f646d2f7469707461702e737667\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/dm\/tiptap.svg\" \/><\/a>\n<a href=\"https:\/\/www.npmjs.com\/package\/tiptap\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/adb05aba2e592bd062bd1f188bc172fe9bc7946e\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f6c2f7469707461702e737667\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/l\/tiptap.svg\" \/><\/a>\n<a href=\"https:\/\/www.npmjs.com\/package\/tiptap\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5a2f758524c9cbdd29a0c40689543051ed3c0ac1\/68747470733a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f7469707461702f646973742f7469707461702e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a6526636f6c6f72423d303030303030\" alt=\"\" data-canonical-src=\"https:\/\/img.badgesize.io\/https:\/\/unpkg.com\/tiptap\/dist\/tiptap.min.js?compression=gzip&amp;label=size&amp;colorB=000000\" \/><\/a><\/p>\n<h2><a id=\"user-content-why-i-built-tiptap\" class=\"anchor\" aria-hidden=\"true\" href=\"#why-i-built-tiptap\"><\/a>Why I built tiptap<\/h2>\n<p>I was looking for a text editor for <a href=\"https:\/\/github.com\/vuejs\/vue\">Vue.js<\/a> and found some solutions that didn't really satisfy me. The editor should be easy to extend and not based on old dependencies such as jQuery. For React there is already a great editor called <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\">Slate.js<\/a>, which impresses with its modularity. I came across <a href=\"https:\/\/github.com\/prosemirror\">Prosemirror<\/a> and decided to build on it. Prosemirror is a toolkit for building rich-text editors that are already in use at many well-known companies such as <em>Atlassian<\/em> or <em>New York Times<\/em>.<\/p>\n<h3><a id=\"user-content-what-means-renderless\" class=\"anchor\" aria-hidden=\"true\" href=\"#what-means-renderless\"><\/a>What means <code>renderless<\/code>?<\/h3>\n<p>With renderless components you'll have (almost) full control over markup and styling. I don't want to tell you what a menu should look like or where it should be rendered in the DOM. That's all up to you. There is also a <a href=\"https:\/\/adamwathan.me\/renderless-components-in-vuejs\/\" rel=\"nofollow\">good article<\/a> about renderless components by Adam Wathan.<\/p>\n<h3><a id=\"user-content-how-is-the-data-stored-under-the-hood\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-is-the-data-stored-under-the-hood\"><\/a>How is the data stored under the hood?<\/h3>\n<p>You can save your data as a raw <code>HTML<\/code> string or can get a <code>JSON<\/code>-serializable representation of your document. And of course, you can pass these two types back to the editor.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h2>\n<p>To check out some live examples, visit <a href=\"https:\/\/tiptap.scrumpy.io\/\" rel=\"nofollow\">tiptap.scrumpy.io<\/a>.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<pre>npm install tiptap\n<\/pre>\n<p>or<\/p>\n<pre>yarn add tiptap\n<\/pre>\n<h2><a id=\"user-content-basic-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-setup\"><\/a>Basic Setup<\/h2>\n<div class=\"highlight highlight-text-html-vue\"><pre>&lt;template&gt;\n  &lt;editor&gt;\n    &lt;!-- Add HTML to the scoped slot called `content` --&gt;\n    &lt;div slot=\"content\" slot-scope=\"props\"&gt;\n      &lt;p&gt;Hi, I'm just a boring paragraph&lt;\/p&gt;\n    &lt;\/div&gt;\n  &lt;\/editor&gt;\n&lt;\/template&gt;\n&lt;script&gt;\n\/\/ Import the editor\nimport { Editor } from 'tiptap'\nexport default {\n  components: {\n    Editor,\n  },\n}\n&lt;\/script&gt;<\/pre><\/div>\n<h2><a id=\"user-content-editor-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#editor-properties\"><\/a>Editor Properties<\/h2>\n<table><thead><tr><th><strong>Property<\/strong><\/th>\n<th align=\"center\"><strong>Type<\/strong><\/th>\n<th align=\"center\"><strong>Default<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><code>editable<\/code><\/td>\n<td align=\"center\"><code>Boolean<\/code><\/td>\n<td align=\"center\"><code>true<\/code><\/td>\n<td>When set to <code>false<\/code> the editor is read-only.<\/td>\n<\/tr><tr><td><code>doc<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>The editor state object used by Prosemirror. You can also pass HTML to the <code>content<\/code> slot. When used both, the <code>content<\/code> slot will be ignored.<\/td>\n<\/tr><tr><td><code>watchDoc<\/code><\/td>\n<td align=\"center\"><code>Boolean<\/code><\/td>\n<td align=\"center\"><code>true<\/code><\/td>\n<td>If set to <code>true<\/code> the content gets updated whenever <code>doc<\/code> changes.<\/td>\n<\/tr><tr><td><code>extensions<\/code><\/td>\n<td align=\"center\"><code>Array<\/code><\/td>\n<td align=\"center\"><code>[]<\/code><\/td>\n<td>A list of extensions used, by the editor. This can be <code>Nodes<\/code>, <code>Marks<\/code> or <code>Plugins<\/code>.<\/td>\n<\/tr><tr><td><code>@init<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>undefined<\/code><\/td>\n<td>This will return an Object with the current <code>state<\/code> and <code>view<\/code> of Prosemirror on init.<\/td>\n<\/tr><tr><td><code>@update<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>undefined<\/code><\/td>\n<td>This will return an Object with the current <code>state<\/code> of Prosemirror, a <code>getJSON()<\/code> and <code>getHTML()<\/code> function on every change.<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-scoped-slots\" class=\"anchor\" aria-hidden=\"true\" href=\"#scoped-slots\"><\/a>Scoped Slots<\/h2>\n<table><thead><tr><th><strong>Name<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><code>editor<\/code><\/td>\n<td>Here the content will be rendered.<\/td>\n<\/tr><tr><td><code>menubar<\/code><\/td>\n<td>Here a menu bar will be rendered.<\/td>\n<\/tr><tr><td><code>menububble<\/code><\/td>\n<td>Here a menu bubble will be rendered.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-slot-properties\" class=\"anchor\" aria-hidden=\"true\" href=\"#slot-properties\"><\/a>Slot Properties<\/h3>\n<p>The <code>menubar<\/code> and <code>menububble<\/code> slot will receive some properties.<\/p>\n<table><thead><tr><th><strong>Property<\/strong><\/th>\n<th align=\"center\"><strong>Type<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><code>nodes<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td>A list of available nodes with active state and command.<\/td>\n<\/tr><tr><td><code>marks<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td>A list of available marks with active state and command.<\/td>\n<\/tr><tr><td><code>focused<\/code><\/td>\n<td align=\"center\"><code>Boolean<\/code><\/td>\n<td>Whether the editor is focused.<\/td>\n<\/tr><tr><td><code>focus<\/code><\/td>\n<td align=\"center\"><code>Function<\/code><\/td>\n<td>A function to focus the editor.<\/td>\n<\/tr><\/tbody><\/table><h2><a id=\"user-content-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#extensions\"><\/a>Extensions<\/h2>\n<p>By default, the editor will only support paragraphs. Other nodes and marks are available as <strong>extensions<\/strong>. There is a package called <code>tiptap-extensions<\/code> with the most basic nodes, marks, and plugins.<\/p>\n<h3><a id=\"user-content-available-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#available-extensions\"><\/a>Available Extensions<\/h3>\n<div class=\"highlight highlight-text-html-vue\"><pre>&lt;template&gt;\n  &lt;editor :extensions=\"extensions\"&gt;\n    &lt;div slot=\"content\" slot-scope=\"props\"&gt;\n      &lt;h1&gt;Yay Headlines!&lt;\/h1&gt;\n      &lt;p&gt;All these &lt;strong&gt;cool tags&lt;\/strong&gt; are working now.&lt;\/p&gt;\n    &lt;\/div&gt;\n  &lt;\/editor&gt;\n&lt;\/template&gt;\n&lt;script&gt;\nimport { Editor } from 'tiptap'\nimport {\n  \/\/ Nodes\n  BlockquoteNode,\n  CodeBlockNode,\n  CodeBlockHighlightNode,\n  HardBreakNode,\n  HeadingNode,\n  ImageNode,\n  OrderedListNode,\n  BulletListNode,\n  ListItemNode,\n  TodoItemNode,\n  TodoListNode,\n  \/\/ Marks\n  BoldMark,\n  CodeMark,\n  ItalicMark,\n  LinkMark,\n  StrikeMark,\n  UnderlineMark,\n  \/\/ General Extensions\n  HistoryExtension,\n  PlaceholderExtension,\n} from 'tiptap-extensions'\nexport default {\n  components: {\n    Editor,\n  },\n  data() {\n    return {\n      extensions: [\n        new BlockquoteNode(),\n        new BulletListNode(),\n        new CodeBlockNode(),\n        new HardBreakNode(),\n        new HeadingNode({ maxLevel: 3 }),\n        new ImageNode(),\n        new ListItemNode(),\n        new OrderedListNode(),\n        new TodoItemNode(),\n        new TodoListNode(),\n        new BoldMark(),\n        new CodeMark(),\n        new ItalicMark(),\n        new LinkMark(),\n        new StrikeMark(),\n        new UnderlineMark(),\n        new HistoryExtension(),\n        new PlaceholderExtension(),\n      ],\n    }\n  },\n}\n&lt;\/script&gt;<\/pre><\/div>\n<h3><a id=\"user-content-create-custom-extensions\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-custom-extensions\"><\/a>Create Custom Extensions<\/h3>\n<p>The most powerful feature of tiptap is that you can create your own extensions. There are 3 types of extensions.<\/p>\n<table><thead><tr><th><strong>Type<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><code>Extension<\/code><\/td>\n<td>The most basic type. It's useful to register some <a href=\"https:\/\/prosemirror.net\/docs\/guide\/\" rel=\"nofollow\">Prosemirror plugins<\/a> or some input rules.<\/td>\n<\/tr><tr><td><code>Node<\/code><\/td>\n<td>Add a custom node. Nodes are block elements like a headline or a paragraph.<\/td>\n<\/tr><tr><td><code>Mark<\/code><\/td>\n<td>Add a custom mark. Marks are used to add extra styling or other information to inline content like a strong tag or links.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-extension-class\" class=\"anchor\" aria-hidden=\"true\" href=\"#extension-class\"><\/a>Extension Class<\/h3>\n<table><thead><tr><th><strong>Method<\/strong><\/th>\n<th align=\"center\"><strong>Type<\/strong><\/th>\n<th align=\"center\"><strong>Default<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><code>get name()<\/code><\/td>\n<td align=\"center\"><code>String<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define a name for your extension.<\/td>\n<\/tr><tr><td><code>get defaultOptions()<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>{}<\/code><\/td>\n<td>Define some default options. The options are available as <code>this.$options<\/code>.<\/td>\n<\/tr><tr><td><code>get plugins()<\/code><\/td>\n<td align=\"center\"><code>Array<\/code><\/td>\n<td align=\"center\"><code>[]<\/code><\/td>\n<td>Define a list of <a href=\"https:\/\/prosemirror.net\/docs\/guide\/\" rel=\"nofollow\">Prosemirror plugins<\/a>.<\/td>\n<\/tr><tr><td><code>keys({ schema })<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define some keybindings.<\/td>\n<\/tr><tr><td><code>inputRules({ schema })<\/code><\/td>\n<td align=\"center\"><code>Array<\/code><\/td>\n<td align=\"center\"><code>[]<\/code><\/td>\n<td>Define a list of input rules.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-nodemark-class\" class=\"anchor\" aria-hidden=\"true\" href=\"#nodemark-class\"><\/a>Node|Mark Class<\/h3>\n<table><thead><tr><th><strong>Method<\/strong><\/th>\n<th align=\"center\"><strong>Type<\/strong><\/th>\n<th align=\"center\"><strong>Default<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><code>get name()<\/code><\/td>\n<td align=\"center\"><code>String<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define a name for your node or mark.<\/td>\n<\/tr><tr><td><code>get defaultOptions()<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>{}<\/code><\/td>\n<td>Define some default options. The options are available as <code>this.$options<\/code>.<\/td>\n<\/tr><tr><td><code>get schema()<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define a <a href=\"https:\/\/prosemirror.net\/docs\/guide\/#schema\" rel=\"nofollow\">schema<\/a>.<\/td>\n<\/tr><tr><td><code>get view()<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define a node view as a vue component.<\/td>\n<\/tr><tr><td><code>keys({ type, schema })<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define some keybindings.<\/td>\n<\/tr><tr><td><code>command({ type, schema, attrs })<\/code><\/td>\n<td align=\"center\"><code>Object<\/code><\/td>\n<td align=\"center\"><code>null<\/code><\/td>\n<td>Define a command. This is used for menus to convert to this node or mark.<\/td>\n<\/tr><tr><td><code>inputRules({ type, schema })<\/code><\/td>\n<td align=\"center\"><code>Array<\/code><\/td>\n<td align=\"center\"><code>[]<\/code><\/td>\n<td>Define a list of input rules.<\/td>\n<\/tr><tr><td><code>get plugins()<\/code><\/td>\n<td align=\"center\"><code>Array<\/code><\/td>\n<td align=\"center\"><code>[]<\/code><\/td>\n<td>Define a list of <a href=\"https:\/\/prosemirror.net\/docs\/guide\/\" rel=\"nofollow\">Prosemirror plugins<\/a>.<\/td>\n<\/tr><\/tbody><\/table><h3><a id=\"user-content-create-a-node\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-a-node\"><\/a>Create a Node<\/h3>\n<p>Let's take a look at a real example. This is basically how the default <code>blockquote<\/code> node from <a href=\"https:\/\/www.npmjs.com\/package\/tiptap-extensions\" rel=\"nofollow\"><code>tiptap-extensions<\/code><\/a> looks like.<\/p>\n<div class=\"highlight highlight-source-js\"><pre>import { Node } from 'tiptap'\nimport { wrappingInputRule, setBlockType, wrapIn } from 'tiptap-commands'\nexport default class BlockquoteNode extends Node {\n  \/\/ choose a unique name\n  get name() {\n    return 'blockquote'\n  }\n  \/\/ the prosemirror schema object\n  \/\/ take a look at https:\/\/prosemirror.net\/docs\/guide\/#schema for a detailed explanation\n  get schema() {\n    return {\n      content: 'block+',\n      group: 'block',\n      defining: true,\n      draggable: false,\n      \/\/ define how the editor will detect your node from pasted HTML\n      \/\/ every blockquote tag will be converted to this blockquote node\n      parseDOM: [\n        { tag: 'blockquote' },\n      ],\n      \/\/ this is how this node will be rendered\n      \/\/ in this case a blockquote tag with a class called `awesome-blockquote` will be rendered\n      \/\/ the '0' stands for its text content inside\n      toDOM: () =&gt; ['blockquote', { class: 'awesome-blockquote' }, 0],\n    }\n  }\n  \/\/ this command will be called from menus to add a blockquote\n  \/\/ `type` is the prosemirror schema object for this blockquote\n  \/\/ `schema` is a collection of all registered nodes and marks\n  command({ type, schema }) {\n    return wrapIn(type)\n  }\n  \/\/ here you can register some shortcuts\n  \/\/ in this case you can create a blockquote with `ctrl` + `&gt;`\n  keys({ type }) {\n    return {\n      'Ctrl-&gt;': wrapIn(type),\n    }\n  }\n  \/\/ a blockquote will be created when you are on a new line and type `&gt;` followed by a space\n  inputRules({ type }) {\n    return [\n      wrappingInputRule(\/^\\s*&gt;\\s$\/, type),\n    ]\n  }\n}<\/pre><\/div>\n<h3><a id=\"user-content-create-a-node-as-a-vue-component\" class=\"anchor\" aria-hidden=\"true\" href=\"#create-a-node-as-a-vue-component\"><\/a>Create a Node as a Vue Component<\/h3>\n<p>The real power of the nodes comes in combination with Vue components. Let us build an iframe node, where you can change its URL (this can also be found in our <a href=\"https:\/\/github.com\/heyscrumpy\/tiptap\/tree\/master\/examples\/Components\/Routes\/Embeds\">examples<\/a>).<\/p>\n<div class=\"highlight highlight-source-js\"><pre>import { Node } from 'tiptap'\nexport default class IframeNode extends Node {\n  get name() {\n    return 'iframe'\n  }\n  get schema() {\n    return {\n      \/\/ here you have to specify all values that can be stored in this node\n      attrs: {\n        src: {\n          default: null,\n        },\n      },\n      group: 'block',\n      selectable: false,\n      \/\/ parseDOM and toDOM is still required to make copy and paste work\n      parseDOM: [{\n        tag: 'iframe',\n        getAttrs: dom =&gt; ({\n          src: dom.getAttribute('src'),\n        }),\n      }],\n      toDOM: node =&gt; ['iframe', {\n        src: node.attrs.src,\n        frameborder: 0,\n        allowfullscreen: 'true',\n      }],\n    }\n  }\n  \/\/ return a vue component\n  \/\/ this can be an object or an imported component\n  get view() {\n    return {\n      \/\/ there are some props available\n      \/\/ `node` is a Prosemirror Node Object\n      \/\/ `updateAttrs` is a function to update attributes defined in `schema`\n      \/\/ `editable` is the global editor prop whether the content can be edited\n      props: ['node', 'updateAttrs', 'editable'],\n      data() {\n        return {\n          \/\/ save the iframe src in a new variable because `this.node.attrs` is immutable\n          url: this.node.attrs.src,\n        }\n      },\n      methods: {\n        onChange(event) {\n          this.url = event.target.value\n          \/\/ update the iframe url\n          this.updateAttrs({\n            src: this.url,\n          })\n        },\n      },\n      template: `\n        &lt;div class=\"iframe\"&gt;\n          &lt;iframe class=\"iframe__embed\" :src=\"url\"&gt;&lt;\/iframe&gt;\n          &lt;input class=\"iframe__input\" type=\"text\" :value=\"url\" @input=\"onChange\" v-if=\"editable\" \/&gt;\n        &lt;\/div&gt;\n      `,\n    }\n  }\n}<\/pre><\/div>\n<h2><a id=\"user-content-building-a-menu\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-a-menu\"><\/a>Building a Menu<\/h2>\n<p>This is a basic example of building a custom menu. A more advanced menu can be found at the <a href=\"https:\/\/tiptap.scrumpy.io\" rel=\"nofollow\">examples page<\/a>.<\/p>\n<div class=\"highlight highlight-text-html-vue\"><pre>&lt;template&gt;\n  &lt;editor :extensions=\"extensions\"&gt;\n    &lt;div slot=\"menubar\" slot-scope=\"{ nodes, marks }\"&gt;\n      &lt;div v-if=\"nodes &amp;&amp; marks\"&gt;\n        &lt;button :class=\"{ 'is-active': nodes.heading.active({ level: 1 }) }\" @click=\"nodes.heading.command({ level: 1 })\"&gt;\n          H1\n        &lt;\/button&gt;\n        &lt;button :class=\"{ 'is-active': marks.bold.active() }\" @click=\"marks.bold.command()\"&gt;\n          Bold\n        &lt;\/button&gt;\n      &lt;\/div&gt;\n    &lt;\/div&gt;\n    &lt;div slot=\"content\" slot-scope=\"props\"&gt;\n      &lt;p&gt;This text can be made bold.&lt;\/p&gt;\n    &lt;\/div&gt;\n  &lt;\/editor&gt;\n&lt;\/template&gt;\n&lt;script&gt;\nimport { Editor } from 'tiptap'\nimport { HeadingNode, BoldMark } from 'tiptap-extensions'\nexport default {\n  components: {\n    Editor,\n  },\n  data() {\n    return {\n      extensions: [\n        new HeadingNode({ maxLevel: 3 }),\n        new BoldMark(),\n      ],\n    }\n  },\n}\n&lt;\/script&gt;<\/pre><\/div>\n<h2><a id=\"user-content-development-setup\" class=\"anchor\" aria-hidden=\"true\" href=\"#development-setup\"><\/a>Development Setup<\/h2>\n<p>Currently, only Yarn is supported for development because of a feature called workspaces we are using here.<\/p>\n<div class=\"highlight highlight-source-shell\"><pre># install dependencies\nyarn install\n# serve examples at localhost:3000\nyarn start\n# build dist files for packages\nyarn build:packages\n# build dist files for examples\nyarn build:examples<\/pre><\/div>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing<\/h2>\n<p>Please see <a href=\"https:\/\/github.com\/heyscrumpy\/tiptap\/blob\/master\/CONTRIBUTING.md\">CONTRIBUTING<\/a> for details.<\/p>\n<h2><a id=\"user-content-credits\" class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h2>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#license\"><\/a>License<\/h2>\n<p>The MIT License (MIT). Please see <a href=\"https:\/\/github.com\/heyscrumpy\/tiptap\/blob\/master\/LICENSE.md\">License File<\/a> for more information.<\/p>\n<\/article>","created_at":"2018-11-07T16:21:59+0000","updated_at":"2018-11-07T16:21:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":7,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/37333181?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":32,"title":"Fetching ActivityPub Feeds \u00b7 Gokberk Yaltirakli","url":"https:\/\/gkbrk.com\/2018\/06\/fetching-activitypub-feeds\/","content":"<p>Mastodon is a federated social network that uses the ActivityPub protocol to\nconnect separate communities into one large network. Both Mastodon and the\nActivityPub protocol are increasing in usage every day. Compared to formats like\nRSS, which are pull-based, ActivityPub is push-based. This means rather than\nyour followers downloading your feed regularly to check if you have shared\nanything, you send each follower (or each server as an optimization) the content\nyou shared.<\/p><p>While this decreases latency in your followers receiving your updates, it does\ncomplicate the implementation of readers. But fortunately, it is still possible\nto pull the feed of ActivityPub users. Just like the good old days.<\/p><p>In this article; we\u2019re going to start from a handle like <code>leo@niu.moe<\/code>, and end\nup with a feed of my latest posts.<\/p><h2>WebFinger<\/h2><p>First of all, let\u2019s look at how the fediverse knows how to find the ActivityPub\nendpoint for a given handle. The way this is done is quite similar to email.<\/p><p>To find the domain name, let\u2019s split the handle into the username and domain\nparts.<\/p><div class=\"highlight\"><pre>handle           = 'leo@niu.moe'\nusername, domain = handle.split('@')\n<\/pre><\/div><p>Next, we need to make a request to the domain\u2019s webfinger endpoint in order to\nfind more data about the account. This is done by performing a GET request to\n<code>\/.well-known\/webfinger<\/code>.<\/p><div class=\"highlight\"><pre>wf_url = 'https:\/\/{}\/.well-known\/webfinger'.format(domain)\nwf_par = {'resource': 'acct:{}'.format(handle)}\nwf_hdr = {'Accept': 'application\/jrd+json'}\n# Perform the request\nwf_resp = requests.get(wf_url, headers=wf_hdr, params=wf_par).json()\n<\/pre><\/div><p>Now we have our WebFinger response. We can filter this data in order to find the\ncorrect ActivityPub endpoint. We need to do this because webfinger can return a\nvariety of URLs, not just ActivityPub.<\/p><h2>Filtering the endpoints<\/h2><p>The response we get from WebFinger looks like this.<\/p><div class=\"highlight\"><pre>{\n  \"subject\": \"acct:leo@niu.moe\",\n  \"aliases\": [\n    \"https:\/\/niu.moe\/@leo\",\n    \"https:\/\/niu.moe\/users\/leo\"\n  ],\n  \"links\": [\n    {\n      \"rel\": \"http:\/\/webfinger.net\/rel\/profile-page\",\n      \"type\": \"text\/html\",\n      \"href\": \"https:\/\/niu.moe\/@leo\"\n    },\n    {\n      \"rel\": \"http:\/\/schemas.google.com\/g\/2010#updates-from\",\n      \"type\": \"application\/atom+xml\",\n      \"href\": \"https:\/\/niu.moe\/users\/leo.atom\"\n    },\n    {\n      \"rel\": \"self\",\n      \"type\": \"application\/activity+json\",\n      \"href\": \"https:\/\/niu.moe\/users\/leo\"\n    }\n  ]\n}\n<\/pre><\/div><p>Depending on the server, there might be more or less entries in the <em>links<\/em>\nkey. What we are intereted in is the URL with the type\n<code>application\/activity+json<\/code>. Let\u2019s go through the array and find the link URL\nwe\u2019re looking for.<\/p><div class=\"highlight\"><pre>matching = (link['href'] for link in wf_resp['links'] if link['type'] == 'application\/activity+json')\nuser_url = next(matching, None)\n<\/pre><\/div><h2>Fetching the feed link<\/h2><p>We can fetch our feed URL using <em>requests<\/em> like before. One detail to note here\nis the content type that we need to specify in order to get the data in the\nformat we want.<\/p><div class=\"highlight\"><pre>as_header = {'Accept': 'application\/ld+json; profile=\"https:\/\/www.w3.org\/ns\/activitystreams\"'}\nuser_json = requests.get(user_url, headers=as_header).json()\n<\/pre><\/div><p><em>user_json<\/em> is a dictionary that contains information about the user. This\ninformation includes the username, profile summary, profile picture and other\nURLs related to the user. One such URL is the <em>\u201cOutbox\u201d<\/em>, which is basically a\nfeed of whatever that user shares publicly.<\/p><p>This is the final URL we need to follow, and we will have the user feed.<\/p><div class=\"highlight\"><pre>feed_url  = user_json['outbox']\n<\/pre><\/div><p>In ActivityPub, the feed is an <em>OrderedCollection<\/em>. And those can be\npaginated. The first page can be empty, or have all the content. Or it can be\none event for each page. This is completely up to the implementation. In order\nto handle this transparently, let\u2019s write a generator that will fetch the next\npages when they are requested.<\/p><div class=\"highlight\"><pre>def parse_feed(url):\n    feed = requests.get(url, headers=as_header).json()\n    if 'orderedItems' in feed:\n        for item in feed['orderedItems']:\n            yield item\n    next_url = None\n    if 'first' in feed:\n        next_url = feed['first']\n    elif 'next' in feed:\n        next_url = feed['next']\n    if next_url:\n        for item in parse_feed(next_url):\n            yield item\n<\/pre><\/div><blockquote>\n<p>Now; for the purposes of a blog post and for writing simple feed parsers, this\ncode works with most servers. But this is not a fully spec-complient function\nfor grabbing all the pages of content. Technically <em>next<\/em> and <em>first<\/em> can be\nlists of events instead of other links, but I haven\u2019t come across that in the\nwild. It is probably a good idea to write your code to cover more edge cases\nwhen dealing with servers on the internet.<\/p>\n<\/blockquote><h2>Printing the first 10 posts<\/h2><p>The posts in ActivityPub contain HTML and while this is okay for web browsers,\nwe should strip the HTML tags before printing them to the terminal.<\/p><p>Here\u2019s how we can do that with the <em>BeautifulSoup<\/em> and <em>html<\/em> modules.<\/p><div class=\"highlight\"><pre>def clean_html(s):\n    text = BeautifulSoup(s, 'html.parser').get_text()\n    return html.unescape(text)\ni = 0\nfor item in parse_feed(feed_url):\n    try:\n        # Only new tweets\n        assert item['type'] == 'Create'\n        content = item['object']['content']\n        text = clean_html(content)\n        print(text)\n        i += 1\n    except:\n        continue\n    if i == 10:\n        break\n<\/pre><\/div><h2>Future Work<\/h2><p>Mastodon is not the only implementation of ActivityPub, and each implementation\ncan do things in slightly different ways. While writing code to interact with\nActivityPub servers, you should always consult the specification document.<\/p><h2>Useful Links<\/h2><ul><li><a href=\"https:\/\/www.w3.org\/TR\/2018\/REC-activitypub-20180123\/\">ActivityPub Spec (at the time of writing)<\/a><\/li>\n<li><a href=\"https:\/\/www.w3.org\/TR\/activitypub\/\">ActivityPub Spec (Latest)<\/a><\/li>\n<\/ul><hr \/><div class=\"pagination\"><p><i class=\"fa fa-chevron-left\">\n            <a href=\"https:\/\/gkbrk.com\/2018\/08\/mastodon-bot-in-common-lisp\/\">Mastodon Bot in Common Lisp<\/a><\/i><\/p><p><a href=\"https:\/\/gkbrk.com\/2018\/04\/generating-vanity-infohashes-for-torrents\/\">Generating Vanity Infohashes for Torrents<\/a>\n            <i class=\"fa fa-chevron-right\"><\/i><\/p><\/div><hr \/><h3>Comments BETA<\/h3><br \/>","created_at":"2018-11-07T16:21:59+0000","updated_at":"2018-11-07T16:21:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"gkbrk.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":33,"title":"emilengler\/sysget","url":"https:\/\/github.com\/emilengler\/sysget","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/emilengler\/sysget\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/822ba418e00bbf55e028496b46a6eb5dbb669aee\/68747470733a2f2f7472617669732d63692e6f72672f656d696c656e676c65722f7379736765742e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/emilengler\/sysget.svg?branch=master\" \/><\/a><\/p>\n<h3><a id=\"user-content-a-front-end-for-every-package-manager\" class=\"anchor\" aria-hidden=\"true\" href=\"#a-front-end-for-every-package-manager\"><\/a>A front-end for every package manager<br \/><\/h3>\n<p>sysget is a bridge that lets you use one syntax to every package manager on every unix-based operating system.<br \/>You probably all know the problem when you are on a new distro and don't know anything about the package manager. With sysget you just need to remember one syntax for every package manager.<br \/>The syntax is mostly same with apt-get so it should be easy to use. <br \/><\/p>\n<h3><a id=\"user-content-supported-package-managers\" class=\"anchor\" aria-hidden=\"true\" href=\"#supported-package-managers\"><\/a>Supported package managers:<\/h3>\n<ul><li>apt-get<\/li>\n<li>xbps<\/li>\n<li>dnf<\/li>\n<li>yum<\/li>\n<li>zypper<\/li>\n<li>eopkg<\/li>\n<li>pacman<\/li>\n<li>emerge<\/li>\n<li>pkg<\/li>\n<li>chromebrew<\/li>\n<li>homebrew<\/li>\n<li>nix<\/li>\n<li>snap<\/li>\n<li>npm<\/li>\n<li>flatpak (See notice)<\/li>\n<\/ul><h3><a id=\"user-content-features\" class=\"anchor\" aria-hidden=\"true\" href=\"#features\"><\/a>Features<\/h3>\n<ul><li>search for packages<\/li>\n<li>install packages<\/li>\n<li>remove packages<\/li>\n<li>remove orphans<\/li>\n<li>clear package manager cache<\/li>\n<li>update database<\/li>\n<li>upgrade system<\/li>\n<li>upgrade single package<\/li>\n<\/ul><h3><a id=\"user-content-how-to-install\" class=\"anchor\" aria-hidden=\"true\" href=\"#how-to-install\"><\/a>How to install<\/h3>\n<p><code>make &amp;&amp; sudo make install<\/code><br \/>No dependencies needed<\/p>\n<h3><a id=\"user-content-example\" class=\"anchor\" aria-hidden=\"true\" href=\"#example\"><\/a>Example<\/h3>\n<p>To search for a package<\/p>\n<pre>sysget search &lt;search query&gt;\n<\/pre>\n<p>To install a package<\/p>\n<pre>sysget install &lt;package name&gt;\n<\/pre>\n<p>To remove a package<\/p>\n<pre>sysget remove &lt;package name&gt;\n<\/pre>\n<p>To update the database<\/p>\n<pre>sysget update\n<\/pre>\n<p>To upgrade the system<\/p>\n<pre>sysget upgrade\n<\/pre>\n<p>To upgrade a specific package<\/p>\n<pre>sysget upgrade &lt;package name&gt;\n<\/pre>\n<p>To remove orphans<\/p>\n<pre>sysget autoremove\n<\/pre>\n<p>To clean the cache of the package manager<\/p>\n<pre>sysget clean\n<\/pre>\n<h3><a id=\"user-content-flatpak-notice\" class=\"anchor\" aria-hidden=\"true\" href=\"#flatpak-notice\"><\/a>Flatpak notice<\/h3>\n<p>Flatpak works different than other package managers.<br \/>If you want to install a package you need to provide the remote name or the URL of a .flatpakref<br \/><strong>For Example:<\/strong><br \/><\/p>\n<pre>sysget install flathub org.gimp.GIMP\n<\/pre>\n<p>or<\/p>\n<pre>sysget install https:\/\/flathub.org\/repo\/appstream\/org.gimp.GIMP.flatpakref\n<\/pre>\n<h6><a class=\"anchor\" aria-hidden=\"true\" href=\"#credits\"><\/a>Credits<\/h6>\n<p><a href=\"https:\/\/github.com\/termget\/termget\">TermGet<\/a><\/p>\n<\/article>","created_at":"2018-11-07T16:21:59+0000","updated_at":"2018-11-07T16:21:59+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":1,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/12272949?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":30,"title":"ianstormtaylor\/slate","url":"https:\/\/github.com\/ianstormtaylor\/slate","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p style=\"text-align: center;\">\n  <a href=\"#\"><img src=\"https:\/\/github.com\/ianstormtaylor\/slate\/raw\/master\/docs\/images\/banner.png\" alt=\"image\" \/><\/a>\n<\/p>\n<p style=\"text-align: center;\">\n  A <em>completely<\/em> customizable framework <br \/>for building rich text editors.\n<\/p>\n<p style=\"text-align: center;\">\n  <a href=\"#why\"><strong>Why?<\/strong><\/a> \u00b7\n  <a href=\"#principles\"><strong>Principles<\/strong><\/a> \u00b7\n  <a href=\"http:\/\/slatejs.org\" rel=\"nofollow\"><strong>Demo<\/strong><\/a> \u00b7\n  <a href=\"#examples\"><strong>Examples<\/strong><\/a> \u00b7\n  <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/docs\/general\/plugins.md\"><strong>Plugins<\/strong><\/a> \u00b7\n  <a href=\"http:\/\/docs.slatejs.org\" rel=\"nofollow\"><strong>Documentation<\/strong><\/a> \u00b7\n  <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/Contributing.md\"><strong>Contributing!<\/strong><\/a>\n<\/p>\n<p style=\"text-align: center;\">\n  <a href=\"http:\/\/slatejs.org\" rel=\"nofollow\"><img src=\"https:\/\/github.com\/ianstormtaylor\/slate\/raw\/master\/docs\/images\/preview.png\" alt=\"image\" \/><\/a>\n<\/p>\n<p style=\"text-align: center;\">\n  <a href=\"https:\/\/unpkg.com\/slate\/dist\/slate.min.js\" rel=\"nofollow\">\n    <img src=\"https:\/\/camo.githubusercontent.com\/ef0d60eaec5bd27f719512dd8365b1e32118975c\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652f646973742f736c6174652e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate\/dist\/slate.min.js?compression=gzip&amp;label=size\" alt=\"image\" \/><\/a>\n  <a href=\"https:\/\/slate-slack.herokuapp.com\" rel=\"nofollow\">\n    <img src=\"https:\/\/camo.githubusercontent.com\/27c7f6d9d83da7d417163a6ef87827dfcd792b4a\/68747470733a2f2f736c6174652d736c61636b2e6865726f6b756170702e636f6d2f62616467652e737667\" data-canonical-src=\"https:\/\/slate-slack.herokuapp.com\/badge.svg\" alt=\"image\" \/><\/a>\n  <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/packages\/slate\/package.json\">\n    <img src=\"https:\/\/camo.githubusercontent.com\/704a8d560849845ed1212cdebff23da2a7992190\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652e7376673f6d61784167653d33363030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate.svg?maxAge=3600&amp;label=version&amp;colorB=007ec6\" alt=\"image\" \/><\/a>\n<\/p>\n<p>Slate lets you build rich, intuitive editors like those in <a href=\"https:\/\/medium.com\/\" rel=\"nofollow\">Medium<\/a>, <a href=\"https:\/\/www.dropbox.com\/paper\" rel=\"nofollow\">Dropbox Paper<\/a> or <a href=\"https:\/\/www.google.com\/docs\/about\/\" rel=\"nofollow\">Google Docs<\/a>\u2014which are becoming table stakes for applications on the web\u2014without your codebase getting mired in complexity.<\/p>\n<p>It can do this because all of its logic is implemented with a series of plugins, so you aren't ever constrained by what <em>is<\/em> or <em>isn't<\/em> in \"core\". You can think of it like a pluggable implementation of <code>contenteditable<\/code> built on top of <a href=\"https:\/\/facebook.github.io\/react\/\" rel=\"nofollow\">React<\/a> and <a href=\"https:\/\/facebook.github.io\/immutable-js\/\" rel=\"nofollow\">Immutable<\/a>. It was inspired by libraries like <a href=\"https:\/\/facebook.github.io\/draft-js\/\" rel=\"nofollow\">Draft.js<\/a>, <a href=\"http:\/\/prosemirror.net\/\" rel=\"nofollow\">Prosemirror<\/a> and <a href=\"http:\/\/quilljs.com\/\" rel=\"nofollow\">Quill<\/a>.<\/p>\n<blockquote>\n<p>\ud83e\udd16<strong>Slate is currently in beta<\/strong>. Its core API is useable now, but you might need to pull request a fix or two for advanced use cases. Some of its APIs are not \"finalized\" and maybe change over time as we discover better solutions.<\/p>\n<\/blockquote>\n<br \/><h3><a id=\"user-content-why\" class=\"anchor\" aria-hidden=\"true\" href=\"#why\"><\/a>Why?<\/h3>\n<p>Why create Slate? Well... <em>(Beware: this section has a few of <a href=\"https:\/\/github.com\/ianstormtaylor\">my<\/a> opinions!)<\/em><\/p>\n<p>Before creating Slate, I tried a lot of the other rich text libraries out there\u2014<a href=\"https:\/\/facebook.github.io\/draft-js\/\" rel=\"nofollow\"><strong>Draft.js<\/strong><\/a>, <a href=\"http:\/\/prosemirror.net\/\" rel=\"nofollow\"><strong>Prosemirror<\/strong><\/a>, <a href=\"http:\/\/quilljs.com\/\" rel=\"nofollow\"><strong>Quill<\/strong><\/a>, etc. What I found was that while getting simple examples to work was easy enough, once you started trying to build something like <a href=\"https:\/\/medium.com\/\" rel=\"nofollow\">Medium<\/a>, <a href=\"https:\/\/www.dropbox.com\/paper\" rel=\"nofollow\">Dropbox Paper<\/a> or <a href=\"https:\/\/www.google.com\/docs\/about\/\" rel=\"nofollow\">Google Docs<\/a>, you ran into deeper issues...<\/p>\n<ul><li>\n<p><strong>The editor's \"schema\" was hardcoded and hard to customize.<\/strong> Things like bold and italic were supported out of the box, but what about comments, or embeds, or even more domain-specific needs?<\/p>\n<\/li>\n<li>\n<p><strong>Transforming the documents programmatically was very convoluted.<\/strong> Writing as a user may have worked, but making programmatic changes, which is critical for building advanced behaviors, was needlessly complex.<\/p>\n<\/li>\n<li>\n<p><strong>Serializing to HTML, Markdown, etc. seemed like an afterthought.<\/strong> Simple things like transforming a document to HTML or Markdown involved writing lots of boilerplate code, for what seemed like very common use cases.<\/p>\n<\/li>\n<li>\n<p><strong>Re-inventing the view layer seemed inefficient and limiting.<\/strong> Most editors rolled their own views, instead of using existing technologies like React, so you have to learn a whole new system with new \"gotchas\".<\/p>\n<\/li>\n<li>\n<p><strong>Collaborative editing wasn't designed for in advance.<\/strong> Often the editor's internal representation of data made it impossible to use to for a realtime, collaborative editing use case without basically rewriting the editor.<\/p>\n<\/li>\n<li>\n<p><strong>The repositories were monolithic, not small and reusable.<\/strong> The code bases for many of the editors often didn't expose the internal tooling that could have been re-used by developers, leading to having to reinvent the wheel.<\/p>\n<\/li>\n<li>\n<p><strong>Building complex, nested documents was impossible.<\/strong> Many editors were designed around simplistic \"flat\" documents, making things like tables, embeds and captions difficult to reason about and sometimes impossible.<\/p>\n<\/li>\n<\/ul><p>Of course not every editor exhibits all of these issues, but if you've tried using another editor you might have run into similar problems. To get around the limitations of their API's and achieve the user experience you're after, you have to resort to very hacky things. And some experiences are just plain impossible to achieve.<\/p>\n<p>If that sounds familiar, you might like Slate.<\/p>\n<p>Which brings me to how Slate solves all of that...<\/p>\n<br \/><h3><a id=\"user-content-principles\" class=\"anchor\" aria-hidden=\"true\" href=\"#principles\"><\/a>Principles<\/h3>\n<p>Slate tries to solve the question of \"<a href=\"#why\">Why?<\/a>\" with a few principles:<\/p>\n<ol><li>\n<p><strong>First-class plugins.<\/strong> The most important part of Slate is that plugins are first-class entities\u2014the core editor logic is even implemented as its own plugin. That means you can <em>completely<\/em> customize the editing experience, to build complex editors like Medium's or Dropbox's, without having to fight against the library's assumptions.<\/p>\n<\/li>\n<li>\n<p><strong>Schema-less core.<\/strong> Slate's core logic doesn't assume anything about the schema of the data you'll be editing, which means that there are no assumptions baked into the library that'll trip you up when you need to go beyond the most basic use cases.<\/p>\n<\/li>\n<li>\n<p><strong>Nested document model.<\/strong> The document model used for Slate is a nested, recursive tree, just like the DOM itself. This means that creating complex components like tables or nested block quotes are possible for advanced use cases. But it's also easy to keep it simple by only using a single level of hierarchy.<\/p>\n<\/li>\n<li>\n<p><strong>Parallel to the DOM.<\/strong> Slate's data model is based on the DOM\u2014the document is a nested tree, it uses selections and ranges, and it exposes all the standard event handlers. This means that advanced behaviors like tables or nested block quotes are possible. Pretty much anything you can do in the DOM, you can do in Slate.<\/p>\n<\/li>\n<li>\n<p><strong>Stateless views and immutable data.<\/strong> By using React and Immutable.js, the Slate editor is built in a stateless fashion using immutable data structures, which leads to much easier to reason about code, and a much easier time writing plugins.<\/p>\n<\/li>\n<li>\n<p><strong>Intuitive changes.<\/strong> Slate documents are edited using \"changes\", that are designed to be high-level and extremely intuitive to write and read, so that custom functionality is as expressive as possible. This greatly increases your ability to reason about your code.<\/p>\n<\/li>\n<li>\n<p><strong>Collaboration-ready data model.<\/strong> The data model Slate uses\u2014specifically how changes are applied to the document\u2014has been designed to allow for collaborative editing to be layered on top, so you won't need to rethink everything if you decide to make your editor collaborative.<\/p>\n<\/li>\n<li>\n<p><strong>Clear \"core\" boundaries.<\/strong> With a plugin-first architecture, and a schema-less core, it becomes a lot clearer where the boundary is between \"core\" and \"custom\", which means that the core experience doesn't get bogged down in edge cases.<\/p>\n<\/li>\n<\/ol><br \/><h3><a id=\"user-content-demo\" class=\"anchor\" aria-hidden=\"true\" href=\"#demo\"><\/a>Demo<\/h3>\n<p>Check out the <a href=\"http:\/\/slatejs.org\" rel=\"nofollow\"><strong>live demo<\/strong><\/a> of all of the examples!<\/p>\n<br \/><h3><a id=\"user-content-examples\" class=\"anchor\" aria-hidden=\"true\" href=\"#examples\"><\/a>Examples<\/h3>\n<p>To get a sense for how you might use Slate, check out a few of the examples:<\/p>\n<ul><li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/plain-text\"><strong>Plain text<\/strong><\/a> \u2014 showing the most basic case: a glorified <code>&lt;textarea&gt;<\/code>.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/rich-text\"><strong>Rich text<\/strong><\/a> \u2014 showing the features you'd expect from a basic editor.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/markdown-shortcuts\"><strong>Auto-markdown<\/strong><\/a> \u2014 showing how to add key handlers for Markdown-like shortcuts.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/links\"><strong>Links<\/strong><\/a> \u2014 showing how wrap text in inline nodes with associated data.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/images\"><strong>Images<\/strong><\/a> \u2014 showing how to use void (text-less) nodes to add images.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/hovering-menu\"><strong>Hovering menu<\/strong><\/a> \u2014 showing how a contextual hovering menu can be implemented.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/tables\"><strong>Tables<\/strong><\/a> \u2014 showing how to nest blocks to render more advanced components.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/paste-html\"><strong>Paste HTML<\/strong><\/a> \u2014 showing how to use an HTML serializer to handle pasted HTML.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\/code-highlighting\"><strong>Code Highlighting<\/strong><\/a> \u2014 showing how to use decorations to dynamically mark text.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/examples\"><strong>See all the examples...<\/strong><\/a><\/li>\n<\/ul><p>If you have an idea for an example that shows a common use case, pull request it!<\/p>\n<br \/><h3><a id=\"user-content-plugins\" class=\"anchor\" aria-hidden=\"true\" href=\"#plugins\"><\/a>Plugins<\/h3>\n<p>Slate encourages you to write small, reusable modules. Check out the public ones you can use in your project!<\/p>\n<ul><li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate-plugins\/tree\/master\/packages\/slate-auto-replace\"><code>slate-auto-replace<\/code><\/a> auto-replaces text as the user types. Useful for \"smart\" typography!<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate-plugins\/tree\/master\/packages\/slate-collapse-on-escape\"><code>slate-collapse-on-escape<\/code><\/a> simply collapses the selection when <code>escape<\/code> is pressed.<\/li>\n<li><a href=\"https:\/\/github.com\/GitbookIO\/slate-edit-code\"><code>slate-edit-code<\/code><\/a> adds code editing behavior like tab-to-indent, and enter-to-soft-break.<\/li>\n<li><a href=\"https:\/\/github.com\/GitbookIO\/slate-edit-list\"><code>slate-edit-list<\/code><\/a> adds rich, nested list editing behavior.<\/li>\n<li><a href=\"https:\/\/github.com\/GitbookIO\/slate-edit-table\"><code>slate-edit-table<\/code><\/a> adds complex table editing behavior!<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate-plugins\/tree\/master\/packages\/slate-paste-linkify\"><code>slate-paste-linkify<\/code><\/a> wraps the selected text in a link when a URL is pasted from the clipboard.<\/li>\n<li><a href=\"https:\/\/github.com\/GitbookIO\/slate-prism\"><code>slate-prism<\/code><\/a> highlights code blocks with <a href=\"http:\/\/prismjs.com\/\" rel=\"nofollow\">Prism.js<\/a>!<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate-plugins\/tree\/master\/packages\/slate-soft-break\"><code>slate-soft-break<\/code><\/a> adds a soft break when <code>enter<\/code> is pressed.<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate-plugins\/tree\/master\/packages\/slate-drop-or-paste-images\"><code>slate-drop-or-paste-images<\/code><\/a> lets users drop or paste images to insert them!<\/li>\n<li><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/docs\/general\/plugins.md\"><strong>See all the plugins...<\/strong><\/a><\/li>\n<\/ul><br \/><h3><a class=\"anchor\" aria-hidden=\"true\" href=\"#documentation\"><\/a>Documentation<\/h3>\n<p>If you're using Slate for the first time, check out the <a href=\"http:\/\/docs.slatejs.org\/walkthroughs\/installing-slate\" rel=\"nofollow\">Getting Started<\/a> walkthroughs and the <a href=\"http:\/\/docs.slatejs.org\/guides\" rel=\"nofollow\">Guides<\/a> to familiarize yourself with Slate's architecture and mental models. Once you've gotten familiar with those, you'll probably want to check out the full <a href=\"http:\/\/docs.slatejs.org\/slate-core\" rel=\"nofollow\">API Reference<\/a>.<\/p>\n<ul><li><a href=\"http:\/\/docs.slatejs.org\/walkthroughs\/installing-slate\" rel=\"nofollow\"><strong>Walkthroughs<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/docs.slatejs.org\/guides\" rel=\"nofollow\"><strong>Guides<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/docs.slatejs.org\/slate-core\" rel=\"nofollow\"><strong>Reference<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/docs.slatejs.org\/general\/faq\" rel=\"nofollow\"><strong>FAQ<\/strong><\/a><\/li>\n<li><a href=\"http:\/\/docs.slatejs.org\/general\/resources\" rel=\"nofollow\"><strong>Resources<\/strong><\/a><\/li>\n<\/ul><p>If even that's not enough, you can always <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/packages\/slate\">read the source itself<\/a>, which is heavily commented.<\/p>\n<p>There are also translations of the documentation into other languages:<\/p>\n<p>If you're maintaining a translation, feel free to pull request it here!<\/p>\n<br \/><h3><a id=\"user-content-packages\" class=\"anchor\" aria-hidden=\"true\" href=\"#packages\"><\/a>Packages<\/h3>\n<p>Slate's codebase is monorepo managed with <a href=\"https:\/\/lernajs.io\/\" rel=\"nofollow\">Lerna<\/a>. It consists of a handful of packages\u2014although you won't always use all of them. They are:<\/p>\n<table><thead><tr><th><strong>Package<\/strong><\/th>\n<th><strong>Version<\/strong><\/th>\n<th><strong>Size<\/strong><\/th>\n<th><strong>Description<\/strong><\/th>\n<\/tr><\/thead><tbody><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate\"><code>slate<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/9483bcf521e3eeac7c6237094483f180353791bc\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate\/dist\/slate.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/ef0d60eaec5bd27f719512dd8365b1e32118975c\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652f646973742f736c6174652e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate\/dist\/slate.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>Slate's core data model logic.<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-base64-serializer\"><code>slate-base64-serializer<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-base64-serializer\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/7c61af94f8219d5a862b5ff83a50d80c100fcbbf\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d6261736536342d73657269616c697a65722e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-base64-serializer.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-base64-serializer\/dist\/slate-base64-serializer.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/5fab75e45c421971201f4765fa6c7642d2279427\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d6261736536342d73657269616c697a65722f646973742f736c6174652d6261736536342d73657269616c697a65722e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-base64-serializer\/dist\/slate-base64-serializer.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>A Base64 string serializer for Slate documents.<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-html-serializer\"><code>slate-html-serializer<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-html-serializer\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/8583f29cfb97377e818f80cb31eecdae74608b54\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d68746d6c2d73657269616c697a65722e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-html-serializer.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-html-serializer\/dist\/slate-html-serializer.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/882ed7b942e3ae907d52a1f9f11404e3a893c187\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d68746d6c2d73657269616c697a65722f646973742f736c6174652d68746d6c2d73657269616c697a65722e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-html-serializer\/dist\/slate-html-serializer.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>An HTML serializer for Slate documents.<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-hyperscript\"><code>slate-hyperscript<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-hyperscript\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/257a4e29bd3f86fe12b5ed0e2f6aefb0da6360cc\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d68797065727363726970742e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-hyperscript.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-hyperscript\/dist\/slate-hyperscript.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/02631e7628f2cb4d7cf635a719412e94b4ea1ef4\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d68797065727363726970742f646973742f736c6174652d68797065727363726970742e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-hyperscript\/dist\/slate-hyperscript.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>A hyperscript tool to write JSX Slate documents!<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-plain-serializer\"><code>slate-plain-serializer<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-plain-serializer\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/eace82498ca12aab11092c31baeed403a4593381\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d706c61696e2d73657269616c697a65722e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-plain-serializer.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-plain-serializer\/dist\/slate-plain-serializer.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/ac2fd26137e36a68011f7605be6a50c04c23fb99\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d706c61696e2d73657269616c697a65722f646973742f736c6174652d706c61696e2d73657269616c697a65722e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-plain-serializer\/dist\/slate-plain-serializer.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>A plain text serializer for Slate documents.<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-prop-types\"><code>slate-prop-types<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-prop-types\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/d5d25f6501b5d097854320211ccfbb535eb6b895\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d70726f702d74797065732e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-prop-types.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-prop-types\/dist\/slate-prop-types.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/469a17bd88f528a87d455e3479cbca56a1321eb8\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d70726f702d74797065732f646973742f736c6174652d70726f702d74797065732e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-prop-types\/dist\/slate-prop-types.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>React prop types for checking Slate values.<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-react\"><code>slate-react<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-react\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/0efb210e990fd16f69b10249bc7e290b2d3eba92\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d72656163742e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-react.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-react\/dist\/slate-react.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/6825855d42a42971cc8ae019bebf7ea001894034\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d72656163742f646973742f736c6174652d72656163742e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-react\/dist\/slate-react.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>React components for rendering Slate editors.<\/td>\n<\/tr><tr><td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-hotkeys\"><code>slate-hotkeys<\/code><\/a><\/td>\n<td><a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/slate-hotkeys\/package.json\"><img src=\"https:\/\/camo.githubusercontent.com\/7f466977bb0854ef9c6c0d8d277e14859ae1165d\/68747470733a2f2f696d672e736869656c64732e696f2f6e706d2f762f736c6174652d686f746b6579732e7376673f6d61784167653d32353932303030266c6162656c3d76657273696f6e26636f6c6f72423d303037656336\" alt=\"\" data-canonical-src=\"https:\/\/img.shields.io\/npm\/v\/slate-hotkeys.svg?maxAge=2592000&amp;label=version&amp;colorB=007ec6\" \/><\/a><\/td>\n<td><a href=\"https:\/\/unpkg.com\/slate-hotkeys\/dist\/slate-hotkeys.min.js\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/03dd295e7d8c3af954eeb5962ecf0dc586335fa5\/687474703a2f2f696d672e626164676573697a652e696f2f68747470733a2f2f756e706b672e636f6d2f736c6174652d686f746b6579732f646973742f736c6174652d686f746b6579732e6d696e2e6a733f636f6d7072657373696f6e3d677a6970266c6162656c3d73697a65\" alt=\"\" data-canonical-src=\"http:\/\/img.badgesize.io\/https:\/\/unpkg.com\/slate-hotkeys\/dist\/slate-hotkeys.min.js?compression=gzip&amp;label=size\" \/><\/a><\/td>\n<td>Detect common keypresses in a platform-agnostic way<\/td>\n<\/tr><\/tbody><\/table><br \/><h3><a id=\"user-content-contributing\" class=\"anchor\" aria-hidden=\"true\" href=\"#contributing\"><\/a>Contributing!<\/h3>\n<p>All contributions are super welcome! Check out the <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/Contributing.md\">Contributing instructions<\/a> for more info!<\/p>\n<p>Slate is <a href=\"https:\/\/github.com\/ianstormtaylor\/slate\/blob\/master\/License.md\">MIT-licensed<\/a>.<\/p>\n<\/article>","created_at":"2018-11-07T16:21:58+0000","updated_at":"2018-11-07T16:21:58+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":6,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/311752?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":28,"title":"27 Top Landing Page Themes For WordPress: Boost Conversions With Ease","url":"https:\/\/bloggingwizard.com\/top-wordpress-landing-pages-sales-pages-themes\/","content":"<div class=\"feature-image columns small-12\"><img width=\"750\" height=\"300\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Best-Landing-Page-Themes-For-WordPress-Featured.png\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"Best Landing Page Themes For WordPress\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Best-Landing-Page-Themes-For-WordPress-Featured.png 750w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Best-Landing-Page-Themes-For-WordPress-Featured-320x128.png 320w\" \/><\/div>\n\t\t<p>A landing page, is a standalone page that is designed around a specific goal and free of distractions.<\/p>\n<p>Any page on your site could, in theory, be classed as a landing page because visitors land on different pages <a href=\"https:\/\/bloggingwizard.com\/landing-page-optimization\/\" target=\"_blank\" rel=\"noopener\"><em>after clicking on an ad, a link in an email, a call-to-action on social media, or many other sources<\/em><\/a>.<\/p>\n<p>Your primary goal is to get your visitors to take action.<\/p>\n<p>Perhaps you\u2019re looking to convert them into members, subscribers, or loyal customers. Or maybe you want them to buy your products or services.<\/p>\n<p>Whatever your objective is, you need a WordPress theme that supports that goal.<\/p>\n<p>But with thousands of WordPress themes available, how do you find one that specializes in <a href=\"https:\/\/bloggingwizard.com\/landing-page-optimization\/\" target=\"_blank\" rel=\"noopener\">landing pages<\/a>?<\/p>\n<p>It\u2019s easy. You grab your favorite beverage. Settle yourself comfortably. And check out the themes below.<\/p>\n<p>We\u2019ve done the legwork for you and narrowed the search to the best WordPress themes that have landing pages.<\/p>\n<p>What\u2019s more, we\u2019ve divided the themes into five sections so you can pinpoint the theme that suits your requirements best.<\/p>\n<div id=\"toc_container\" class=\"no_bullets\"><p class=\"toc_title\">Contents<\/p><ul class=\"toc_list\"><li><a href=\"#Our-top-picks-for-WordPress-landing-page-themes\">Our top picks for WordPress landing page themes<\/a><ul><li><a href=\"#Divi\">Divi<\/a><\/li><li><a href=\"#Extra\">Extra<\/a><\/li><li><a href=\"#Growtheme\">Growtheme<\/a><\/li><li><a href=\"#Rise\">Rise<\/a><\/li><\/ul><\/li><li><a href=\"#Themes-with-landing-page-templates\">Themes with landing page templates<\/a><ul><li><a href=\"#Biz-Lady\">Biz Lady<\/a><\/li><li><a href=\"#Hearten\">Hearten<\/a><\/li><li><a href=\"#Monochrome-Pro\">Monochrome Pro<\/a><\/li><li><a href=\"#Infinity-Pro\">Infinity Pro<\/a><\/li><li><a href=\"#Parallax-Pro\">Parallax Pro<\/a><\/li><li><a href=\"#Showcase-Pro\">Showcase Pro<\/a><\/li><li><a href=\"#Aspire-Pro\">Aspire Pro<\/a><\/li><li><a href=\"#Agent-Focused-Pro\">Agent Focused Pro<\/a><\/li><\/ul><\/li><li><a href=\"#Multi-purpose-themes\">Multi-purpose themes<\/a><ul><li><a href=\"#Leadinjection\">Leadinjection<\/a><\/li><li><a href=\"#Inbound\">Inbound<\/a><\/li><li><a href=\"#X-The-Theme\">X | The Theme<\/a><\/li><li><a href=\"#Marketing-Pro\">Marketing Pro<\/a><\/li><li><a href=\"#Digital-Agency\">Digital Agency<\/a><\/li><li><a href=\"#Stratus\">Stratus<\/a><\/li><li><a href=\"#Leadx\">Leadx<\/a><\/li><li><a href=\"#Urip\">Urip<\/a><\/li><\/ul><\/li><li><a href=\"#Coming-soon-landing-page-themes\">Coming soon landing page themes<\/a><ul><li><a href=\"#Landing\">Landing<\/a><\/li><li><a href=\"#CMP-Fifty\">CMP Fifty<\/a><\/li><\/ul><\/li><li><a href=\"#Free-themes\">Free themes<\/a><ul><li><a href=\"#Zerif-Lite\">Zerif Lite<\/a><\/li><li><a href=\"#OnePress\">OnePress<\/a><\/li><li><a href=\"#Bento\">Bento<\/a><\/li><li><a href=\"#Optimizer\">Optimizer<\/a><\/li><li><a href=\"#ResponsiveBoat\">ResponsiveBoat<\/a><\/li><\/ul><\/li><li><a href=\"#Conclusion\">Conclusion<\/a><\/li><\/ul><\/div>\n\n<h2>Our top picks for WordPress landing page themes<\/h2>\n<p>Let\u2019s start with our favorite landing page themes.<\/p>\n<p>These four are all versatile themes that you can use to create great-looking landing pages.<\/p>\n<p>Take a look and see what you think.<\/p>\n<h3>Divi<\/h3>\n<p><img class=\"size-full wp-image-29193 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Divi-Theme-WordPress-Landing-Page-Themes.jpg\" alt=\"Divi Theme\" width=\"550\" height=\"268\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Divi-Theme-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Divi-Theme-WordPress-Landing-Page-Themes-320x156.jpg 320w\" \/><\/p>\n<p>When you choose the Divi Theme, you also get the visual drag-and-drop Divi Page Builder. Using this powerful combination, you can create all types of landing pages.<\/p>\n<p><a class=\"thirstylink\" title=\"Divi\" href=\"https:\/\/bloggingwizard.com\/go\/divi\/\" target=\"_blank\" rel=\"nofollow noopener\">Divi<\/a>, from <strong>Elegant Themes<\/strong>, ships with over 20 pre-made page layouts that include templates for Contact, Portfolio, and Membership pages. These provide a quick way to get started as you can use the proven designs and swap the demo text to your content. Any changes you make can be saved to your Divi Library to save you time in the future.<\/p>\n<p>But you\u2019re not limited to pre-made layouts because the success of Divi lies in its versatility.<\/p>\n<p>You can create any landing page using the design elements like Testimonials, Calls To Action, Countdown Timers, and Pricing Tables.<\/p>\n<p>Using the WYSIWYG editor, you can drag and drop any element onto your page and then customize it to your liking. You can adjust fonts, colors, sizing and spacing and even apply custom CSS.<\/p>\n<p>Building your landing page is one thing; testing how effective it is another. That\u2019s where Divi Leads comes in.<\/p>\n<p><img class=\"size-full wp-image-29192 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Divi-Leads-WordPress-Landing-Page-Themes.png\" alt=\"Divi Leads\" width=\"550\" height=\"334\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Divi-Leads-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Divi-Leads-WordPress-Landing-Page-Themes-296x180.png 296w\" \/><\/p>\n<p><a href=\"https:\/\/www.elegantthemes.com\/gallery\/divi\/divi-leads\/\" target=\"_blank\" rel=\"noopener\">Divi Leads<\/a> allows you to split test different content and design options to see which converts best according to your defined goals. Using these insights, you\u2019ll be able to increase click-through rate and reduce bounce rate on your landing pages.<\/p>\n<p><strong>Price: <\/strong>$89\/year or $249\/Lifetime Access<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/divi\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Divi<\/a>\u00a0 <a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/divi\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Extra<\/h3>\n<p><img class=\"size-full wp-image-29195 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Extra-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Extra Theme\" width=\"550\" height=\"300\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Extra-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Extra-Theme-WordPress-Landing-Page-Themes-320x175.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/bloggingwizard.com\/go\/extra\/\" target=\"_blank\" rel=\"noopener\">Extra<\/a> also uses the Divi Page Builder, but it\u2019s more suitable for creating magazine-style blogs and online publications.<\/p>\n<p>In short: <em>Where <strong>Divi\u2019s<\/strong> primary focus is on the page, <strong>Extra\u2019s<\/strong> focus is on the post feed.<\/em><\/p>\n<p>But you\u2019ll still need landing pages even if you\u2019re running a blog. Using the Divi Builder, you can create any landing page. Plus Extra has three unique page building options.<\/p>\n<p><strong>Extra\u2019s Category Builder<\/strong> gives you complete control over your home and category pages. You can build unique pages on a per-category-basis, by adding elements to your page:<\/p>\n<p><img class=\"size-full wp-image-29194 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Extra-Divi-Builder-WordPress-Landing-Page-Themes.png\" alt=\"Extra Divi Builder\" width=\"550\" height=\"356\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Extra-Divi-Builder-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Extra-Divi-Builder-WordPress-Landing-Page-Themes-278x180.png 278w\" \/><\/p>\n<p><strong>Extra<\/strong> also features two other useful page builder options. <strong>Projects<\/strong> is ideal for designers and writers to showcase your portfolio, while <strong>Products <\/strong>(with built-in WooCommerce integration) is perfect to sell products to your readers.<\/p>\n<p><strong>Price: <\/strong>$89\/year or $249\/Lifetime Access<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/extra\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Extra<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/extra\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Growtheme<\/h3>\n<p><img class=\"size-full wp-image-29196 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Growtheme-WordPress-Landing-Page-Themes.png\" alt=\"Growtheme\" width=\"550\" height=\"273\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Growtheme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Growtheme-WordPress-Landing-Page-Themes-320x159.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/bloggingwizard.com\/go\/growtheme\/\" target=\"_blank\" rel=\"noopener\">Growtheme<\/a> is focused on turning your visitors into customers by building your email list.<\/p>\n<p>According to this <a href=\"https:\/\/www.getresponse.com\/resources\/reports\/state-of-email-marketing-by-industry-2016.html\" target=\"_blank\" rel=\"noopener\">GetResponse Survey<\/a>, email marketing is still the most effective marketing channel to grow your audience.<\/p>\n<p>With that in mind, Growtheme\u2019s developer, Jascha Brinkmann, created a WordPress theme with all the <a href=\"https:\/\/bloggingwizard.com\/wordpress-mailing-list-plugins-superior-list-building-power\/\" target=\"_blank\" rel=\"noopener\">list-building features<\/a> included \u2013 no need for multiple plugins and separate landing page products.<\/p>\n<p>You can add Feature Boxes, After-Post Opt-ins, Author Byline Opt-ins, Sticky Sidebar Widgets, and more to any post or page.<\/p>\n<p>Because everything is self-contained, you can apply your theme design to all the built-in landing pages, so you maintain a consistent look and feel.<\/p>\n<p>Growtheme includes an Upside Down Homepage, a Welcome Gate, a Basic Squeeze Page, an Opt-in Confirmation Page, and a Thank You Page, with more to be added soon.<\/p>\n<p><strong>Price: <\/strong>$197<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/growtheme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Growtheme<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/growtheme\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Rise<\/h3>\n<p><img class=\"size-full wp-image-29197 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Rise-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Rise Theme\" width=\"550\" height=\"363\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Rise-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Rise-Theme-WordPress-Landing-Page-Themes-273x180.png 273w\" \/><\/p>\n<p>Thrive Themes build <strong>conversion-focused themes<\/strong>, and <a href=\"https:\/\/bloggingwizard.com\/go\/thrive-rise\/\" target=\"_blank\" rel=\"noopener\">Rise<\/a> is no exception. It\u2019s the ideal blogging theme developed in collaboration with affiliate marketer Zac Johnson.<\/p>\n<p>Rise is a flexible and effective theme that comes with five color styles, extensive admin options, multiple widget areas, and loads of shortcodes to display lead-generating opt-ins.<\/p>\n<p>It includes a Blank Landing Page Template that preserves your site\u2019s overall branding. The distraction-free landing page template removes the sidebar, all footer widgets, and the top navigation elements.<\/p>\n<p>Aside from the blank template, there are also templates for two Lead Generation Pages, a Sales Page, a Download Page, and an Email Confirmation Page.<\/p>\n<p><strong>Price: <\/strong>$49<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/thrive-rise\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Rise<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/thrive-rise\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h2>Themes with landing page templates<\/h2>\n<p>In this section, we have a collection of themes that have their own landing page templates.<\/p>\n<p>To start with, here are two themes for the ladies.<\/p>\n<h3>Biz Lady<\/h3>\n<p><img class=\"size-full wp-image-29209 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Biz-lady-Theme-WordPress-Landing-Page-Themes.jpeg\" alt=\"Biz lady Theme\" width=\"550\" height=\"368\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Biz-lady-Theme-WordPress-Landing-Page-Themes.jpeg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Biz-lady-Theme-WordPress-Landing-Page-Themes-269x180.jpeg 269w\" \/><\/p>\n<p>Continuing with the feminine theme from APrettierWeb is <a class=\"thirstylink\" title=\"Biz Lady\" href=\"https:\/\/bloggingwizard.com\/go\/biz-lady-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Biz Lady<\/a>. This theme is ideal for online business owners such as coaches, service providers, and freelancers. The theme integrates with the LifterLMS and WooCommerce plugins so you can sell your online courses, services, and products.<\/p>\n<p>Biz Lady has page templates in various shapes and sizes: 750px wide, 1170px wide, and a Full-width page. Then there\u2019s a Page Sidebar, where you choose unique sidebar for pages; a Lead Page; and a Blank Page.<\/p>\n<p>Unlike Boss Lady, all the landing pages include the header and footer.<\/p>\n<p>You can see the theme in action on <a href=\"http:\/\/aprettierweb.com\/\" target=\"_blank\" rel=\"noopener\">A Prettier Web<\/a>.<\/p>\n<p><strong>Price: <\/strong>$79<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/biz-lady-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Biz Lady<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/biz-lady-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Hearten<\/h3>\n<p><img class=\"size-full wp-image-29212 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Heaten-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Heaten Theme\" width=\"550\" height=\"364\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Heaten-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Heaten-Theme-WordPress-Landing-Page-Themes-272x180.png 272w\" \/><\/p>\n<p>If you\u2019re a female business coach, then the <a class=\"thirstylink\" title=\"Hearten\" href=\"https:\/\/bloggingwizard.com\/go\/hearten-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Hearten<\/a> theme by Athena will suit you down to the ground.<\/p>\n<p><strong>Hearten<\/strong> was designed for professional female coaches who want to get their online biz up-and-running ASAP.<\/p>\n<p>The elegant, feminine design includes social media buttons to grow your audience, opt-in forms to build your list, and banners to promote your products and services.<\/p>\n<p>You can choose from two built-in Lead Page templates, either with or without the header and footer. Plus you can build fabulous pages without any coding knowledge thanks to the <a href=\"http:\/\/www.wpsuperstars.net\/elementor-review\/\" target=\"_blank\" rel=\"noopener\">Elementor<\/a> page builder.<\/p>\n<p><strong>Price: <\/strong>$59<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/hearten-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Hearten<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/hearten-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n        <div class=\"content-box note\">\n            Note\n            <div>\n                <p><strong>StudioPress Themes<\/strong><\/p>\n<p>All <a href=\"https:\/\/bloggingwizard.com\/go\/studiopress-themes\/\" target=\"_blank\" rel=\"noopener\">StudioPress themes<\/a> include a Landing Page Template \u2013 it\u2019s a basic page with nothing on it. This gives you the option of working with the theme or using a <a href=\"https:\/\/bloggingwizard.com\/top-wordpress-landing-page-plugins\/\" target=\"_blank\" rel=\"noopener\">page builder plugin<\/a> to create your landing pages.<\/p>\n<p>StudioPress themes have limited customization options to avoid them getting bloated and hindering performance. The \u2018StudioPress Way\u2019 is to add code snippets. But if you prefer using plugins rather than dabbling with code, then you could try using the Genesis Dambuster.<\/p>\n<p>The <a href=\"http:\/\/www.genesisdambuster.com\/\" target=\"_blank\" rel=\"noopener\">Genesis Dambuster<\/a> plugin integrates Genesis with Beaver Builder, so you don\u2019t have to add extra code for full-width pages or posts.<\/p>\n            <\/div>\n        <\/div>\n        \n<h3>Monochrome Pro<\/h3>\n<p><img class=\"size-full wp-image-29216 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Monochrome-Pro-Theme-WordPress-Landing-Page-Themes.jpg\" alt=\"Monochrome Pro Theme\" width=\"550\" height=\"239\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Monochrome-Pro-Theme-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Monochrome-Pro-Theme-WordPress-Landing-Page-Themes-320x139.jpg 320w\" \/><\/p>\n<p>The <a class=\"thirstylink\" title=\"Monochrome Pro Theme\" href=\"https:\/\/bloggingwizard.com\/go\/monochrome-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Monochrome Pro Theme<\/a> by StudioPress emphasizes the importance of white space and typography. If you\u2019re looking for a no-fuss minimalist design, then this theme is for you.<\/p>\n<p>Monochrome Pro is the epitome of distraction-free, but the Landing Page Template also removes the header and footer to leave you with a blank canvas for your content.<\/p>\n<p>There are also templates for Author, Contact and Pricing pages. And there are the usual StudioPress options for widget areas<\/p>\n<p>Selling products in Monochrome Pro (and other StudioPress Themes) is made easier with the built-in online store that is pre-styled for WooCommerce:<\/p>\n<p><img class=\"size-full wp-image-29215 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Monochrome-Pro-Shop-WordPress-Landing-Page-Themes.png\" alt=\"Monochrome Pro Shop\" width=\"550\" height=\"250\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Monochrome-Pro-Shop-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Monochrome-Pro-Shop-WordPress-Landing-Page-Themes-320x145.png 320w\" \/><\/p>\n<p>Use Monochrome Pro to communicate your message in a clear, direct manner.<\/p>\n<p><strong>Price: <\/strong>$129.95 (includes Genesis Framework)<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/monochrome-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Monochrome Pro<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/monochrome-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Infinity Pro<\/h3>\n<p><img class=\"size-full wp-image-29214 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Infinity-Pro-Theme-WordPress-Landing-Page-Themes.jpg\" alt=\"Infinity Pro Theme\" width=\"550\" height=\"270\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Infinity-Pro-Theme-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Infinity-Pro-Theme-WordPress-Landing-Page-Themes-320x157.jpg 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Infinity Pro\" href=\"https:\/\/bloggingwizard.com\/go\/infinity-theme\/\" target=\"_blank\" rel=\"nofollow noopener\"> Infinity Pro<\/a> is an elegant and responsive theme designed to position your online brand. There are plenty of customization options, including 13 widget areas, making it one of the most flexible themes from StudioPress.<\/p>\n<p>Aside from the standard Landing Page template there are also templates for Author, Contact, and Lead Capture pages.<\/p>\n<p>Infinity Pro also includes a Protected Content page that you can use for restricted membership areas on your site:<\/p>\n<p><img class=\"size-full wp-image-29213 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Infinity-Pro-Protected-Content-WordPress-Landing-Page-Themes.png\" alt=\"Infinity Pro Protected Content\" width=\"550\" height=\"312\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Infinity-Pro-Protected-Content-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Infinity-Pro-Protected-Content-WordPress-Landing-Page-Themes-317x180.png 317w\" \/><\/p>\n<p><strong>Price: <\/strong>$129.95 (includes Genesis Framework)<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/infinity-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Infinity<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/infinity-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Parallax Pro<\/h3>\n<p><img class=\"size-full wp-image-29233 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme.jpg\" alt=\"Parallax Pro Theme\" width=\"550\" height=\"270\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-320x157.jpg 320w\" \/><\/p>\n<p>The <a href=\"https:\/\/bloggingwizard.com\/go\/parallax-pro\/\" target=\"_blank\" rel=\"noopener\">Parallax Pro<\/a> theme is designed to tell a story. The theme includes five homepage widget areas that you can customize to fit around your brand message.<\/p>\n<p>Parallax Pro keeps your content flowing seamlessly as readers scroll from section to section down the page. As they reach the end of your story there\u2019s a CTA (Call To Action) to prompt their next step:<\/p>\n<p><img class=\"size-full wp-image-29231 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-CTA.png\" alt=\"Parallax Pro Theme CTA\" width=\"550\" height=\"275\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-CTA.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-CTA-320x160.png 320w\" \/><\/p>\n<p>Parallax Pro makes it easy to convert visitors into fans.<\/p>\n<p>As you\u2019d expect, the Landing Page template is plain and simple:<\/p>\n<p><img class=\"size-full wp-image-29232 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-Landing-Page.png\" alt=\"Parallax Pro Theme Landing Page\" width=\"490\" height=\"400\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-Landing-Page.png 490w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Parallax-Pro-Theme-Landing-Page-221x180.png 221w\" \/><\/p>\n<p>Elsewhere, this minimal black and white theme comes with your choice of accent color (blue, green, orange, pink or red) for content boxes.<\/p>\n<p><strong>Price: <\/strong>$99.95 (includes Genesis Framework)<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/parallax-pro\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Parallax Pro<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/parallax-pro\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Showcase Pro<\/h3>\n<p><img class=\"size-full wp-image-29235 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Showcase-Pro-Theme.jpg\" alt=\"Showcase Pro Theme\" width=\"550\" height=\"222\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Showcase-Pro-Theme.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Showcase-Pro-Theme-320x129.jpg 320w\" \/><\/p>\n<p><a href=\"https:\/\/bloggingwizard.com\/go\/showcase-pro\/\" target=\"_blank\" rel=\"noopener\">Showcase Pro<\/a> is designed to exhibit your best content and images with maximum impact in a clean, bright style.<\/p>\n<p>It comes with three homepage and four footer widget areas so you can display more information about you and your business.<\/p>\n<p>There are templates for a Landing Page, an About Page, and a Pricing Table:<\/p>\n<p><img class=\"size-full wp-image-29234 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Showcase-Pro-Pricing-Table.png\" alt=\"Showcase Pro Pricing Table\" width=\"550\" height=\"337\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Showcase-Pro-Pricing-Table.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Showcase-Pro-Pricing-Table-294x180.png 294w\" \/><\/p>\n<p>All in all, it\u2019s the perfect design to showcase and sell your products and services.<\/p>\n<p><strong>Price: <\/strong>$129.95 (includes Genesis Framework)<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/showcase-pro\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Showcase Pro<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/showcase-pro\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Aspire Pro<\/h3>\n<p><img class=\"size-full wp-image-533609 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Aspire-Pro-Landing-Page-Theme.jpg\" alt=\"Aspire Pro Landing Page Theme\" width=\"550\" height=\"350\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Aspire-Pro-Landing-Page-Theme.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Aspire-Pro-Landing-Page-Theme-283x180.jpg 283w\" \/><\/p>\n<p><a href=\"https:\/\/bloggingwizard.com\/go\/aspire-pro\/\" target=\"_blank\" rel=\"noopener\">Aspire Pro<\/a> is a popular StudioPress third-party theme designed by Appfinite. You can build your site with some customizable homepage and footer widget areas.<\/p>\n<p>There is a selection of pre-made page templates included with the theme: Default, Archives, Categories, and Blog.<\/p>\n<p>One significant difference is that Aspire Pro uses a wider Landing Page template (1200px vs 800px) compared to the other StudioPress themes but still retains a simple style.<\/p>\n<p>Aspire Pro also includes a Portfolio Page template so that you can link to your published content elsewhere on the web:<\/p>\n<p><img class=\"size-full wp-image-533610 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Aspire-Pro-Portfolio-Theme.png\" alt=\"Aspire Pro Portfolio Theme\" width=\"550\" height=\"321\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Aspire-Pro-Portfolio-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Aspire-Pro-Portfolio-Theme-308x180.png 308w\" \/><\/p>\n<p>If you\u2019re serious about your business, then the Aspire Pro theme is the way to go.<\/p>\n<p><strong>Price: <\/strong>$99.95 (includes Genesis Framework)<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/aspire-pro\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Aspire Pro<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/aspire-pro\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Agent Focused Pro<\/h3>\n<p><img class=\"size-full wp-image-533608 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Agent-Focused-Pro-Theme-Theme.jpg\" alt=\"Agent Focused Pro Theme Theme\" width=\"550\" height=\"295\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Agent-Focused-Pro-Theme-Theme.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Agent-Focused-Pro-Theme-Theme-320x172.jpg 320w\" \/><\/p>\n<p><em>Contrary to popular belief, this is not a theme for the intelligence agencies.<\/em><\/p>\n<p><a class=\"thirstylink\" title=\"Agent Focused Pro\" href=\"https:\/\/bloggingwizard.com\/go\/agent-focused-theme\/\" target=\"_blank\" rel=\"nofollow noopener\"> Agent Focused Pro<\/a> is a premium WordPress theme designed for both real estate professionals and their clients.<\/p>\n<p>It\u2019s perfect for showing off your personality, your brand, and your real estate listings. The homepage includes a search bar right under the featured image, so your clients can narrow their property selection.<\/p>\n<p>In addition to the standard page templates, Agent Focused Pro has a collection of Real Estate pages: Featured Listings, Single Listing, Communities, Single community, and Community Type:<\/p>\n<p><img class=\"size-full wp-image-533607 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Agent-Focused-Pro-Community-Type-Theme.png\" alt=\"Agent Focused Pro Community Type Theme\" width=\"478\" height=\"400\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Agent-Focused-Pro-Community-Type-Theme.png 478w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Agent-Focused-Pro-Community-Type-Theme-215x180.png 215w\" \/><\/p>\n<p>Put yourself and your properties on the market with the Agent Focused Pro theme.<\/p>\n<p><strong>Price: <\/strong>$99.95 (includes Genesis Framework)<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/agent-focused-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Agent Focused<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/agent-focused-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h2>Multi-purpose themes<\/h2>\n<p>Multi-purpose themes need to be handled with care. Sometimes they are overkill.<\/p>\n<p><em>What do we mean by that?<\/em><\/p>\n<p>Well, for starters, they usually have so many options that there\u2019s a steep learning curve. Plus most features aren\u2019t used and it can impact page loading times.<\/p>\n<p>In contrast, a more lightweight theme like those from StudioPress are faster and easier to use.<\/p>\n<p>The other consideration is theme lock-in.<\/p>\n<p>Many of these multi-purpose themes have built-in page builders, which is great for designing landing pages quickly, but not so good when you decide to change your theme in the future. Most of them leave a trail of shortcodes behind.<\/p>\n<p>Again, the alternative option is to use a more lightweight theme with an external <a href=\"https:\/\/bloggingwizard.com\/page-builder-plugins-for-wordpress\/\" target=\"_blank\" rel=\"noopener\">page builder plugin<\/a> like Beaver Builder, Elementor or Thrive Architect.<\/p>\n<p>So that\u2019s the issue with multi-purpose themes being overkill. They\u2019re great in some respects, but they\u2019re built with the aim of pleasing as many people as possible, which can have drawbacks.<\/p>\n<p>Enough said. Let\u2019s take a look at some multi-purpose themes.<\/p>\n<h3>Leadinjection<\/h3>\n<p><img class=\"size-full wp-image-29228 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Lead-Injection-Theme.png\" alt=\"Lead Injection Theme\" width=\"550\" height=\"280\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Lead-Injection-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Lead-Injection-Theme-320x163.png 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Leadinjection\" href=\"https:\/\/bloggingwizard.com\/go\/leadinjection-theme\/\" target=\"_blank\" rel=\"nofollow noopener\"> Leadinjection<\/a> is a WordPress landing page theme designed for professional marketers, business owners, and affiliates to launch landing pages within minutes.<\/p>\n<p>There are several pre-made <strong>landing page templates<\/strong> for Insurance, SEO Service, eBook, Online Course, Application, Conference, and Medical, with more coming soon:<\/p>\n<p><img class=\"size-full wp-image-29227 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Lead-Injection-Templates.png\" alt=\"Lead Injection Templates\" width=\"550\" height=\"328\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Lead-Injection-Templates.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Lead-Injection-Templates-302x180.png 302w\" \/><\/p>\n<p>The theme also comes with the <strong>Visual Composer<\/strong> drag-and-drop page builder plugin, plus the <strong>leadModal<\/strong> plugin for capturing leads.<\/p>\n<p><strong>Price: <\/strong>$39<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/leadinjection-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Leadinjection<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/leadinjection-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Inbound<\/h3>\n<p><img class=\"size-full wp-image-533611 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Inbound-Landing-Page-Theme.png\" alt=\"Inbound Landing Page Theme\" width=\"550\" height=\"300\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Inbound-Landing-Page-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Inbound-Landing-Page-Theme-320x175.png 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Inbound\" href=\"https:\/\/bloggingwizard.com\/go\/inbound-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Inbound<\/a> is a conversion-focused landing page theme for WordPress.<\/p>\n<p>It features a drag-and-drop visual page builder that lets you design pages without any coding skills. Plus there are no fixed layouts or column sizes so you can create any design you want.<\/p>\n<p>Inbound comes with a growing collection of pre-made landing page templates. For instance, there are templates for running an SEO Agency or selling your eBooks or Apps. Plus you can use multiple landing page designs on a single site.<\/p>\n<p><img class=\"size-full wp-image-29225 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Inbound-Templates.jpeg\" alt=\"Inbound Templates\" width=\"550\" height=\"300\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Inbound-Templates.jpeg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Inbound-Templates-320x175.jpeg 320w\" \/><\/p>\n<p><strong>Price: <\/strong>$49<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/inbound-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Inbound<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/inbound-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>X | The Theme<\/h3>\n<p><img class=\"size-full wp-image-533613 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/X-Theme-Landing-Page-Theme.png\" alt=\"X Theme Landing Page Theme\" width=\"550\" height=\"327\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/X-Theme-Landing-Page-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/X-Theme-Landing-Page-Theme-303x180.png 303w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"X-Theme\" href=\"https:\/\/bloggingwizard.com\/go\/x-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">X-Theme<\/a> from Themeco is billed as <strong><em>The Ultimate WordPress Theme<\/em><\/strong>. That\u2019s some claim, but it\u2019s certainly packed with <strong>eXtra<\/strong> features.<\/p>\n<p>For starters, <strong>X-Theme<\/strong> includes four unique designs inside one WordPress theme that are called \u201cStacks.\u201d They\u2019re like themes within a theme:<\/p>\n<ul><li><strong>Integrity<\/strong> works well for creatives and professionals.<\/li>\n<li><strong>Renew<\/strong> showcases your content in a clear, concise manner.<\/li>\n<li><strong>Icon<\/strong> offers a modern and minimal design.<\/li>\n<li><strong>Ethos <\/strong>is specifically for magazine and newspaper sites.<\/li>\n<\/ul><p><strong>X-Theme<\/strong> has also developed entire demo sites in different niches. For instance, they have Church, Band, Wedding, Agency, and Food Truck. These fully-designed sites come complete with all pages, images, and text content to get you started.<\/p>\n<ul><li><strong>Choose your Stack, pick your Demo, and you\u2019re good to go:<\/strong><\/li>\n<\/ul><p><img class=\"size-full wp-image-533612 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/X-Theme-Demo-Landing-Page-Theme.png\" alt=\"X Theme Demo Landing Page Theme\" width=\"444\" height=\"400\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/X-Theme-Demo-Landing-Page-Theme.png 444w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/X-Theme-Demo-Landing-Page-Theme-200x180.png 200w\" \/><\/p>\n<p>Of course, you don\u2019t have to use a Demo site.<\/p>\n<p><strong>X-Theme<\/strong> has numerous blank page templates so you can choose what combination you want when creating different landing pages. For example:<\/p>\n<ul><li>Header, Footer;<\/li>\n<li>Header, No Footer;<\/li>\n<li>No Header, No Footer.<\/li>\n<\/ul><p>To help create your pages, <strong>X-Theme<\/strong> includes a couple of page builder plugins. You can choose between <strong>Cornerstone<\/strong>, a 100% front-end page builder, or <strong>Visual Composer<\/strong>.<\/p>\n<p><strong>X-Theme<\/strong> also comes with dozens of premium plugins at no extra charge, including Snippet (Schema Tool), Typekit Integration, and Convert Plug.<\/p>\n<p><strong>Price: <\/strong>$55<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/x-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get X Theme<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/x-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Marketing Pro<\/h3>\n<p><img class=\"size-full wp-image-29230 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Marketing-Pro-Theme.png\" alt=\"Marketing Pro Theme\" width=\"550\" height=\"278\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Marketing-Pro-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Marketing-Pro-Theme-320x162.png 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Marketing Pro\" href=\"https:\/\/bloggingwizard.com\/go\/marketing-pro-theme\/\" target=\"_blank\" rel=\"nofollow noopener\"> Marketing Pro<\/a> is a multi-purpose WordPress theme designed specifically for digital marketers, SEO experts and business owners.<\/p>\n<p>The theme comes with nine high-converting ready-made templates:<\/p>\n<p><img class=\"size-full wp-image-29229 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Marketing-Pro-Templates.png\" alt=\"Marketing Pro Templates\" width=\"550\" height=\"360\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Marketing-Pro-Templates.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Marketing-Pro-Templates-275x180.png 275w\" \/><\/p>\n<p>The theme also includes the drag-and-drop page builder plugin <strong>Visual Composer<\/strong> to help you create your site.<\/p>\n<p>There\u2019s a flexible landing page layout to get you started, plus there are four opt-in designs to boost your conversions: Hero, Shortcode, Sidebar Widget, and After Blog Post.<\/p>\n<p><strong>Price: <\/strong>$49<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/marketing-pro-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Marketing Pro<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/marketing-pro-demo-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Digital Agency<\/h3>\n<p><img class=\"size-full wp-image-29257 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Digital-Agency-Theme-WordPress-Landing-Page-Themes.jpg\" alt=\"Digital Agency Theme\" width=\"550\" height=\"280\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Digital-Agency-Theme-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Digital-Agency-Theme-WordPress-Landing-Page-Themes-320x163.jpg 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Digital Agency\" href=\"https:\/\/bloggingwizard.com\/go\/digital-agency-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Digital Agency<\/a> is a multi-purpose WordPress theme built on the Bootstrap 3 Framework.<\/p>\n<p>There are three unique demo homepage variations to choose from, all geared towards SEO and Marketing Agencies.<\/p>\n<p><img class=\"size-full wp-image-29256 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Digital-Agency-Demo-WordPress-Landing-Page-Themes.jpeg\" alt=\"Digital Agency Demo\" width=\"550\" height=\"345\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Digital-Agency-Demo-WordPress-Landing-Page-Themes.jpeg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Digital-Agency-Demo-WordPress-Landing-Page-Themes-287x180.jpeg 287w\" \/><\/p>\n<p>The theme is easy to use and customize, whether you\u2019re a WordPress pro or a beginner. It includes the drag-and-drop Visual Composer page builder plugin, so you won\u2019t need any coding skills.<\/p>\n<p>Digital Agency comes with a selection of pre-made page templates so you can create pages like About, Services, and FAQ. Install the template, change the dummy content, and you\u2019re done.<\/p>\n<p>The theme also has some page layouts for Case Studies and Portfolio pages. These can be configured with 2, 3, or 4 columns.<\/p>\n<p><strong>Price: <\/strong>$59<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/digital-agency-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Digital Agency<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/digital-agency-demo-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Stratus<\/h3>\n<p><img class=\"size-full wp-image-29266 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Stratus Theme\" width=\"550\" height=\"277\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Theme-WordPress-Landing-Page-Themes-320x161.png 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Stratus\" href=\"https:\/\/bloggingwizard.com\/go\/stratus-theme\/\" target=\"_blank\" rel=\"nofollow noopener\"> Stratus<\/a> is a flexible WordPress theme, designed specifically for App and SaaS companies. It\u2019s also suitable for Startups and Product Development Companies who want to showcase their products and services<\/p>\n<p>Pages are easy to design with the built-in page builder. You choose from 20 different types of <strong>Content Blocks<\/strong> to create your page layout. For example:<\/p>\n<ul><li>Brands \u2013 to create a row of logos:<\/li>\n<\/ul><p><img class=\"size-full wp-image-29264 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Metabox-Brands-WordPress-Landing-Page-Themes.jpg\" alt=\"Stratus Metabox Brands\" width=\"550\" height=\"95\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Metabox-Brands-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Metabox-Brands-WordPress-Landing-Page-Themes-320x55.jpg 320w\" \/><\/p>\n<ul><li>Call to Action \u2013 to add a call to action statement with button:<\/li>\n<\/ul><p><img class=\"size-full wp-image-29265 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Metabox-CTA-WordPress-Landing-Page-Themes.jpg\" alt=\"Stratus Metabox CTA\" width=\"550\" height=\"89\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Metabox-CTA-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Metabox-CTA-WordPress-Landing-Page-Themes-320x52.jpg 320w\" \/><\/p>\n<p>Once you\u2019ve created your content, you can choose the look and feel of your pages with a boxed or a full layout effect:<\/p>\n<p><img class=\"size-full wp-image-29263 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Boxed-Full-Layout-WordPress-Landing-Page-Themes.jpeg\" alt=\"Stratus Boxed Full Layout\" width=\"550\" height=\"327\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Boxed-Full-Layout-WordPress-Landing-Page-Themes.jpeg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Stratus-Boxed-Full-Layout-WordPress-Landing-Page-Themes-303x180.jpeg 303w\" \/><\/p>\n<p><strong>Price: <\/strong>$59<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/stratus-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Stratus<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/stratus-demo-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Leadx<\/h3>\n<p><img class=\"size-full wp-image-29259 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Leadx-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Leadx Theme\" width=\"550\" height=\"277\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Leadx-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Leadx-Theme-WordPress-Landing-Page-Themes-320x161.png 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Leadx\" href=\"https:\/\/bloggingwizard.com\/go\/leadx-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Leadx<\/a> is a responsive Landing Page and Lead Marketing theme for WordPress.<\/p>\n<p>You can use Leadx to build a Product Landing Page or a Service Landing Page within minutes.<\/p>\n<p>And it also works fine for medical, insurance, software, app, startup, travel, marketing, and agency Landing Pages too.<\/p>\n<p>Leadx has created a number of demo landing pages for the different businesses:<\/p>\n<p><img class=\"size-full wp-image-29258 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Leadx-Demo-Pages-WordPress-Landing-Page-Themes.png\" alt=\"Leadx Demo Pages\" width=\"550\" height=\"183\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Leadx-Demo-Pages-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Leadx-Demo-Pages-WordPress-Landing-Page-Themes-320x106.png 320w\" \/><\/p>\n<p>You also get some pre-designed lead generation forms that integrate with the Contact Form 7 and Gravity Form plugins. You can gather orders and subscribers straight away.<\/p>\n<p>The Leadx theme, like many of the other themes, is powered by Visual Composer. And, to make it even more powerful, Leadx have added some custom features and shortcodes.<\/p>\n<p><strong>Price: <\/strong>$46<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/leadx-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Leadx<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/leadx-demo-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Urip<\/h3>\n<p><img class=\"size-full wp-image-29268 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Urip-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Urip Theme\" width=\"550\" height=\"277\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Urip-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Urip-Theme-WordPress-Landing-Page-Themes-320x161.png 320w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Urip\" href=\"https:\/\/bloggingwizard.com\/go\/urip-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Urip<\/a> is a multi-purpose WordPress Landing Page theme with a modern, clean, and professional design that works for many businesses.<\/p>\n<p>The developers, Ninetheme, have created five demo layouts:<\/p>\n<ul><li>Hero Subscribe \u2013 above the fold form that integrates with MailChimp<\/li>\n<li>Hero Form \u2013 a modern, eye-catching opt-in form<\/li>\n<li>Portfolio \u2013 to showcase your portfolio<\/li>\n<li>App \u2013 perfect for promoting your apps<\/li>\n<li>Event \u2013 to promote your event or conference<\/li>\n<\/ul><p><img class=\"size-full wp-image-29267 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Urip-Portfolio-Layout-WordPress-Landing-Page-Themes.png\" alt=\"Urip Portfolio Layout\" width=\"550\" height=\"336\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Urip-Portfolio-Layout-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Urip-Portfolio-Layout-WordPress-Landing-Page-Themes-295x180.png 295w\" \/><\/p>\n<p>Each layout has different features that you can pick and customize for your own custom design. Again, Visual Composer is included with the theme to help you create your pages.<\/p>\n<p><strong>Price: <\/strong>$44<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/urip-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Urip<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/urip-demo-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h2>Coming soon landing page themes<\/h2>\n<p>A Coming Soon Page is a great example of a landing page. Sometimes we\u2019re not quite ready to reveal our new product or service. But you can still get visitors to take action by joining the waiting list or following you on social media for updates.<\/p>\n<p>Here\u2019s a couple of landing page themes that can help.<\/p>\n<h3>Landing<\/h3>\n<p><img class=\"size-full wp-image-29287 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Theme-WordPress-Landing-Page-Themes.jpg\" alt=\"Landing Theme\" width=\"550\" height=\"376\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Theme-WordPress-Landing-Page-Themes.jpg 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Theme-WordPress-Landing-Page-Themes-263x180.jpg 263w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"Landing\" href=\"https:\/\/bloggingwizard.com\/go\/landing\/\" target=\"_blank\" rel=\"nofollow noopener\"> Landing<\/a> does exactly what it says on the label \u2013 it\u2019s a WordPress theme from Themify for making landing pages.<\/p>\n<p>The theme comes with 25+ pre-made landing page layouts, including a <strong>Coming Soon Page<\/strong>:<\/p>\n<p><img class=\"size-full wp-image-29285 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Coming-Soon-WordPress-Landing-Page-Themes.jpg\" alt=\"Landing Coming Soon\" width=\"300\" height=\"345\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Coming-Soon-WordPress-Landing-Page-Themes.jpg 300w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Coming-Soon-WordPress-Landing-Page-Themes-157x180.jpg 157w\" \/><\/p>\n<p>Other pre-made layouts include landing pages for agencies, apps, events, ebook, products, weddings, portfolios, shops, and marketers.<\/p>\n<p><img class=\"size-full wp-image-29286 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Layouts-WordPress-Landing-Page-Themes.png\" alt=\"Landing Layouts\" width=\"550\" height=\"354\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Layouts-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Layouts-WordPress-Landing-Page-Themes-280x180.png 280w\" \/><\/p>\n<p>Designing a landing page is fast and easy with the pre-designed layouts. You only have to<\/p>\n<p>load your chosen layout, edit the text and images, and you\u2019re done.<\/p>\n<p>But with Themify\u2019s drag-and-drop builder, you can design any landing page you like. Similar to the Divi Builder you can build pages in the backend or frontend. You can see your design changes with the Live Preview feature.<\/p>\n<p><img class=\"size-full wp-image-29288 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Themify-Builder-WordPress-Landing-Page-Themes.png\" alt=\"Landing Themify Builder\" width=\"550\" height=\"261\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Themify-Builder-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Landing-Themify-Builder-WordPress-Landing-Page-Themes-320x152.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/themify.me\/addons\" target=\"_blank\" rel=\"noopener\">Builder Addons<\/a> are the modules that extend the Themify Builder. For instance, the Coming Soon Page uses the Contact Addon. They can be purchased individually or as a bundle.<\/p>\n<p><strong>Price: <\/strong>$49 or $79 for Themify Club<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/landing\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Landing<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/landing-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>CMP Fifty<\/h3>\n<p><img class=\"size-full wp-image-29306 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/CMP-Fifty-WordPress-Landing-Page-Themes.png\" alt=\"CMP Fifty\" width=\"550\" height=\"363\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/CMP-Fifty-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/CMP-Fifty-WordPress-Landing-Page-Themes-273x180.png 273w\" \/><\/p>\n<p><a class=\"thirstylink\" title=\"CMP Fifty\" href=\"https:\/\/bloggingwizard.com\/go\/cmp-fifty-theme\/\" target=\"_blank\" rel=\"nofollow noopener\"> CMP Fifty<\/a> is a landing page theme that works alongside the free CMP \u2013 <a href=\"https:\/\/wordpress.org\/plugins\/cmp-coming-soon-maintenance\/\" target=\"_blank\" rel=\"noopener\">Coming Soon &amp; Maintenance Plugin<\/a>.<\/p>\n<p>We all need to carry out maintenance on our websites from time to time; e.g. updating themes and plugins, or working on a new homepage design.<\/p>\n<p>Rather than showing the standard \u201cTemporarily unavailable due to maintenance\u201d message, CMP Fifty lets you setup and display a custom landing page for your website visitors.<\/p>\n<p>There are two halves to the landing page.<\/p>\n<p>On one-half, you can add your message plus options for visitors to connect via social media and\/or subscribe for updates.<\/p>\n<p>On the other half, you can use images and photos from your own library or <a href=\"http:\/\/unsplash.com\" target=\"_blank\" rel=\"noopener\">Unsplash<\/a> that rotate on the slider.<\/p>\n<p><img class=\"size-full wp-image-29307 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/CMP-Fifty-Pages-WordPress-Landing-Page-Themes.png\" alt=\"CMP Fifty Pages\" width=\"550\" height=\"290\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/CMP-Fifty-Pages-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/CMP-Fifty-Pages-WordPress-Landing-Page-Themes-320x169.png 320w\" \/><\/p>\n<p>When you\u2019ve completed your work, you can reveal your updated site manually or via a timer.<\/p>\n<p><strong>Price: <\/strong>$10<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/cmp-fifty-theme\/\" target=\"_blank\" rel=\"nofollow noopener\">Get CMP Fifty<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/bloggingwizard.com\/go\/cmp-fifty-theme-demo\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h2>Free themes<\/h2>\n<p>In this final section, we\u2019ve rounded up <strong>five free themes<\/strong> from the WordPress Directory that are perfect for landing pages.<\/p>\n<h3>Zerif Lite<\/h3>\n<p><img class=\"size-full wp-image-29310 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Zerif-Lite-Theme.png\" alt=\"Zerif Lite Theme\" width=\"550\" height=\"275\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Zerif-Lite-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Zerif-Lite-Theme-320x160.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/wordpress.org\/themes\/zerif-lite\/\" target=\"_blank\" rel=\"noopener\">Zerif Lite<\/a> is a responsive one-page WordPress theme that\u2019s clean, modern, flat and minimal.<\/p>\n<p>It\u2019s perfect for both business and personal sites.<\/p>\n<p>Use the content blocks to build your landing page and showcase your team, products, clients, about info, testimonials, latest posts from the blog, contact form, plus additional calls to action.<\/p>\n<p>You can create memorable pages with the smooth parallax effect. And you can add smooth animations with the built-in content slider, too.<\/p>\n<p><img class=\"size-full wp-image-29309 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Zerif-Lite-Parallax.png\" alt=\"Zerif Lite Parallax\" width=\"550\" height=\"144\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Zerif-Lite-Parallax.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Zerif-Lite-Parallax-320x84.png 320w\" \/><\/p>\n<p>Zerif Lite is compatible with WooCommerce, so you can build a front page for your store in minutes.<\/p>\n<p><strong>Price: <\/strong>Free<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/wordpress.org\/themes\/zerif-lite\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Zerif Lite<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/themeisle.com\/demo\/?theme=Zerif%20Lite\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>OnePress<\/h3>\n<p><img class=\"size-full wp-image-29312 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/OnePress-Theme-WordPress-Landing-Page-Themes.png\" alt=\"OnePress Theme\" width=\"550\" height=\"254\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/OnePress-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/OnePress-Theme-WordPress-Landing-Page-Themes-320x148.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/wordpress.org\/themes\/onepress\/\" target=\"_blank\" rel=\"noopener\">OnePress<\/a> is a creative and flexible one-page WordPress theme from FameThemes.<\/p>\n<p>The elegant theme works well for businesses, freelancers, and digital agencies. It\u2019s excellent for gallery portfolios and product showcases:<\/p>\n<p><img class=\"size-full wp-image-29311 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/OnePress-Portfolio-WordPress-Landing-Page-Themes.png\" alt=\"OnePress Portfolio\" width=\"550\" height=\"145\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/OnePress-Portfolio-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/OnePress-Portfolio-WordPress-Landing-Page-Themes-320x84.png 320w\" \/><\/p>\n<p>Like Zerif Lite, you build your landing page using content blocks. You can have sections for your Team, Services, Blog, Contact, etc.<\/p>\n<p>And it\u2019s compatible with the latest version of WooCommerce, so you can build your storefront.<\/p>\n<p><strong>Price: <\/strong>Free<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/wordpress.org\/themes\/onepress\/\" target=\"_blank\" rel=\"nofollow noopener\">Get OnePress<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"http:\/\/www.famethemes.com\/preview\/?theme=OnePress\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Bento<\/h3>\n<p><img class=\"size-full wp-image-533614 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Bento-Theme.png\" alt=\"Bento Theme\" width=\"550\" height=\"306\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Bento-Theme.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Bento-Theme-320x178.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/wordpress.org\/themes\/bento\/\" target=\"_blank\" rel=\"noopener\">Bento<\/a> is a free multi-purpose WordPress theme suitable for a broad range of projects.<\/p>\n<p>For a free theme, it oozes premium quality design and features. It\u2019s mobile-friendly, retina-ready, optimized for speed, and implements SEO (search engine optimization) best practices.<\/p>\n<p>Some smaller features include native infinite scroll, fixed menu options, and specific settings for customizing the look of each page.<\/p>\n<p>Bento also comes with a link to install a free visual content builder, which allows you to create professional layouts without any coding knowledge.<\/p>\n<p>You can divide your content into rows and columns as required. Then add blocks \u2013 from simple text to image sliders, maps, videos and action buttons, as well as sidebar widgets \u2013 to build your pages.<\/p>\n<p>Bento also includes a single-page layout so you can create a section-based landing page.<\/p>\n<p><strong>Price:\u00a0<\/strong>Free<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/wordpress.org\/themes\/bento\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Bento<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"http:\/\/satoristudio.net\/bento\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>Optimizer<\/h3>\n<p><img class=\"size-full wp-image-29314 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Optimizer-Theme-WordPress-Landing-Page-Themes.png\" alt=\"Optimizer Theme\" width=\"550\" height=\"264\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Optimizer-Theme-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Optimizer-Theme-WordPress-Landing-Page-Themes-320x154.png 320w\" \/><\/p>\n<p><a href=\"https:\/\/wordpress.org\/themes\/optimizer\/\" target=\"_blank\" rel=\"noopener\">Optimizer<\/a> is an easy-to-customize multi-purpose WordPress theme.<\/p>\n<p>It\u2019s suitable for many types of website \u2013 small business, portfolio, photography, corporate, personal, eCommerce, and even single landing pages.<\/p>\n<p>You can create your pages live without touching any code. Use the drag-and-drop widgets to add a text block, video block, slider, call to action, maps and more with a few clicks.<\/p>\n<p>The theme comes with a choice of site layout: full-width or boxed. Other design options include an image slider, logo upload, 600+ Fonts, theme options, custom CSS field and color options.<\/p>\n<p>Optimizer is retina ready, translation ready, and WooCommerce-ready. Plus it\u2019s optimized for search engines and mobile devices.<\/p>\n<p><strong>Price: <\/strong>Free<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/wordpress.org\/themes\/optimizer\/\" target=\"_blank\" rel=\"nofollow noopener\">Get Optimizer<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/optimizerwp.com\/\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h3>ResponsiveBoat<\/h3>\n<p><img class=\"size-full wp-image-29316 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/ResponsiveBoat-WordPress-Landing-Page-Themes.png\" alt=\"ResponsiveBoat\" width=\"550\" height=\"330\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/ResponsiveBoat-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/ResponsiveBoat-WordPress-Landing-Page-Themes-300x180.png 300w\" \/><\/p>\n<p><a href=\"https:\/\/wordpress.org\/themes\/responsiveboat\/\" target=\"_blank\" rel=\"noopener\">ResponsiveBoat<\/a> is a free responsive WordPress theme with a friendly and colorful design. It\u2019s actually a child theme of Zerif Lite.<\/p>\n<p>The theme is suitable for web agencies, digital studios, product showcases, and business portfolios.<\/p>\n<p>ResponsiveBoat works as a one-page theme, and you can slide smoothly through the sections with the beautiful parallax effect.<\/p>\n<p><img class=\"size-full wp-image-29315 aligncenter\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/ResponsiveBoat-Parallax-WordPress-Landing-Page-Themes.png\" alt=\"ResponsiveBoat Parallax\" width=\"550\" height=\"267\" srcset=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/ResponsiveBoat-Parallax-WordPress-Landing-Page-Themes.png 550w, https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/ResponsiveBoat-Parallax-WordPress-Landing-Page-Themes-320x155.png 320w\" \/><\/p>\n<p>It also comes with a full-screen layout and a full-width homepage header, plus it integrates with WooCommerce.<\/p>\n<p>Overall, ResponsiveBoat has a modern and interactive appearance.<\/p>\n<p><strong>Price: <\/strong>Free<\/p>\n<p><a class=\"button medium orange\" href=\"https:\/\/wordpress.org\/themes\/responsiveboat\/\" target=\"_blank\" rel=\"nofollow noopener\">Get ResponsiveBoat<\/a>\u00a0\u00a0<a class=\"button medium orange\" href=\"https:\/\/themeisle.com\/demo\/?theme=ResponsiveBoat\" target=\"_blank\" rel=\"nofollow noopener\">View Demo<\/a><\/p>\n<h2>Conclusion<\/h2>\n<p>We\u2019ve shared a boat load of highly functional landing page themes that look great.<\/p>\n<p>Now it\u2019s time to evaluate your needs and choose a WordPress theme that fits them perfectly.<\/p>\n<p><strong>Not found a theme you like? Here are a few other theme roundups that might have what you need:<\/strong><\/p>\n<ul><li><a href=\"https:\/\/bloggingwizard.com\/wordpress-themes-for-bloggers\/\" target=\"_blank\" rel=\"noopener\">41 Fantastic WordPress Themes For Serious Bloggers<\/a><\/li>\n<li><a href=\"https:\/\/bloggingwizard.com\/wordpress-portfolio-themes\/\" target=\"_blank\" rel=\"noopener\">37 Great WordPress Portfolio Themes For Freelancers And Agencies<\/a><\/li>\n<li><a href=\"https:\/\/bloggingwizard.com\/free-wordpress-themes\/\" target=\"_blank\" rel=\"noopener\">49 Free WordPress Themes For Bloggers And Businesses<\/a><\/li>\n<\/ul><p><strong><em>Disclosure:<\/em><\/strong><em>\u00a0This post contains affiliate links. This means we may make a small commission if you make a purchase. This doesn\u2019t cost you any more but it does help us to continue publishing helpful content \u2013 thank you for your support!<\/em><\/p>\n<img class=\"no_pin swp_hidden_pin_image\" src=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Best-Landing-Page-Themes-For-WordPress-Pinterest.png\" data-pin-url=\"https:\/\/bloggingwizard.com\/top-wordpress-landing-pages-sales-pages-themes\/\" data-pin-media=\"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Best-Landing-Page-Themes-For-WordPress-Pinterest.png\" alt=\"\" data-pin-description=\"\" \/><div class=\"swp_social_panel swp_flat_fresh swp_default_custom_color swp_individual_full_color swp_other_custom_color scale-110 scale-full_width\" data-min-width=\"1100\" data-panel-position=\"below\" data-float=\"left\" data-float-mobile=\"bottom\" data-count=\"0\" data-float-color=\"#ffffff&#10;\"><div class=\"nc_tweetContainer swp_twitter\" data-network=\"twitter\"><a rel=\"nofollow noreferrer noopener\" target=\"_blank\" href=\"https:\/\/twitter.com\/intent\/tweet?text=27+Top+Landing+Page+Themes+For+WordPress%3A+Boost+Conversions+With+Ease&amp;url=https%3A%2F%2Fbloggingwizard.com%2Ftop-wordpress-landing-pages-sales-pages-themes%2F&amp;via=bloggingwizard\" data-link=\"https:\/\/twitter.com\/intent\/tweet?text=27+Top+Landing+Page+Themes+For+WordPress%3A+Boost+Conversions+With+Ease&amp;url=https%3A%2F%2Fbloggingwizard.com%2Ftop-wordpress-landing-pages-sales-pages-themes%2F&amp;via=bloggingwizard\" class=\"nc_tweet\"><i class=\"sw swp_twitter_icon\">Tweet<\/i><\/a><\/div><div class=\"nc_tweetContainer swp_pinterest\" data-network=\"pinterest\"><a rel=\"nofollow noreferrer noopener\" class=\"nc_tweet\" data-count=\"0\" data-link=\"https:\/\/pinterest.com\/pin\/create\/button\/?url=https:\/\/bloggingwizard.com\/top-wordpress-landing-pages-sales-pages-themes\/&amp;media=https%3A%2F%2Fbloggingwizard.com%2Fwp-content%2Fuploads%2F2017%2F06%2FBest-Landing-Page-Themes-For-WordPress-Pinterest.png&amp;description=27+Top+Landing+Page+Themes+For+WordPress%3A+Boost+Conversions+With+Ease+via+%40adamjc\"><i class=\"sw swp_pinterest_icon\">Pin<\/i><\/a><\/div><div class=\"nc_tweetContainer swp_facebook\" data-network=\"facebook\"><a rel=\"nofollow noreferrer noopener\" target=\"_blank\" href=\"https:\/\/www.facebook.com\/share.php?u=https%3A%2F%2Fbloggingwizard.com%2Ftop-wordpress-landing-pages-sales-pages-themes%2F\" data-link=\"https:\/\/www.facebook.com\/share.php?u=https%3A%2F%2Fbloggingwizard.com%2Ftop-wordpress-landing-pages-sales-pages-themes%2F\" class=\"nc_tweet\"><i class=\"sw swp_facebook_icon\">Share<\/i><\/a><\/div><div class=\"nc_tweetContainer total_shares total_sharesalt\">933 Shares<\/div><\/div><div class=\"swp-content-locator\">\n        <div class=\"about-the-author\">\n\n            <div class=\"image\"><img alt=\"\" src=\"https:\/\/secure.gravatar.com\/avatar\/61666654e804ff07f4876d62e200354d?s=120&amp;d=blank&amp;r=g\" srcset=\"https:\/\/secure.gravatar.com\/avatar\/61666654e804ff07f4876d62e200354d?s=240&amp;d=blank&amp;r=g 2x\" class=\"avatar avatar-120 photo\" height=\"120\" width=\"120\" \/><\/div>\n\n            <h3>David Hartshorne<\/h3>\n\n            <p>David Hartshorne is a <a href=\"https:\/\/azaharmedia.com\/\">freelance writer<\/a> working with business owners and marketing teams to create detailed, actionable content that resonates with their audience. When he\u2019s not writing about digital marketing and technology, you\u2019ll find him chilling with a thriller in Spain.<\/p>\n\n            <footer class=\"author-meta\"><\/footer><\/div>\n\n        \n                <div class=\"postoptin text-center\">\n                    <h3>Want to grow your blog 425% faster?<\/h3>\n                    <p class=\"lead\">Newsletter subscribers get FREE access to 15+ guides, templates &amp; checklists to accelerate your blog's growth.<\/p><a href=\"#afterpost-optin-inline\" data-open=\"afterpost-optin-inline\" class=\"button normal cta\">Yes, I'm In!<\/a>\n                        <div id=\"afterpost-optin-inline\" class=\"fast reveal\" data-reveal=\"\" data-animation-in=\"scale-in-up\" data-animation-out=\"scale-out-down\">\n                            <h2>Get Our Subscriber-Only Content &amp; Blogging Tips Newsletter<\/h2>\n                            <h3>Enter your email address below to get started:<\/h3><form action=\"\/top-wordpress-landing-pages-sales-pages-themes\/\" method=\"post\" class=\"form\"><div><input type=\"hidden\" name=\"FORM\" value=\"518113560\" \/><input type=\"email\" placeholder=\"Your email address...\" value=\"\" name=\"email\" class=\"required email\" required=\"required\" \/><button type=\"submit\" name=\"subscribe\">Awesome. Subscribe Now.<\/button><\/div><\/form><p><small>Includes Email Updates. Unsubscribe Anytime.<\/small><\/p>\n                            <button class=\"close-button\" data-close=\"\" aria-label=\"Close modal\" type=\"button\">\n                                \u00d7\n                            <\/button>\n                        <\/div><p><small>Includes Email Updates. Unsubscribe Anytime.<\/small><\/p>\n                <\/div>        \n\t\t\t<\/div>","created_at":"2018-11-07T16:21:57+0000","updated_at":"2018-11-07T16:21:57+0000","published_at":"2017-06-21T10:55:42+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":21,"domain_name":"bloggingwizard.com","preview_picture":"https:\/\/bloggingwizard.com\/wp-content\/uploads\/2017\/06\/Best-Landing-Page-Themes-For-WordPress-Social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":29,"title":"abo-abo\/hydra","url":"https:\/\/github.com\/abo-abo\/hydra","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/abo-abo\/hydra\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/606344d2efe212c85fcf83b2b3f9098c56d22110\/68747470733a2f2f7472617669732d63692e6f72672f61626f2d61626f2f68796472612e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/travis-ci.org\/abo-abo\/hydra.svg?branch=master\" \/><\/a><\/p>\n<p>This is a package for GNU Emacs that can be used to tie related commands into a family of short\nbindings with a common prefix - a Hydra.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/7188a57a097771ce80af534a4183c759717b54af\/687474703a2f2f6f72656d6163732e636f6d2f646f776e6c6f61642f48796472612e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/7188a57a097771ce80af534a4183c759717b54af\/687474703a2f2f6f72656d6163732e636f6d2f646f776e6c6f61642f48796472612e6a7067\" alt=\"hydra\" data-canonical-src=\"http:\/\/oremacs.com\/download\/Hydra.jpg\" \/><\/a><\/p>\n<h2><a id=\"user-content-description-for-poets\" class=\"anchor\" aria-hidden=\"true\" href=\"#description-for-poets\"><\/a>Description for Poets<\/h2>\n<p>Once you summon the Hydra through the prefixed binding (the body + any one head), all heads can be\ncalled in succession with only a short extension.<\/p>\n<p>The Hydra is vanquished once Hercules, any binding that isn't the Hydra's head, arrives.  Note that\nHercules, besides vanquishing the Hydra, will still serve his original purpose, calling his proper\ncommand.  This makes the Hydra very seamless, it's like a minor mode that disables itself\nauto-magically.<\/p>\n<h2><a id=\"user-content-description-for-pragmatics\" class=\"anchor\" aria-hidden=\"true\" href=\"#description-for-pragmatics\"><\/a>Description for Pragmatics<\/h2>\n<p>Imagine that you have bound <kbd>C-c j<\/kbd> and <kbd>C-c k<\/kbd> in your\nconfig.  You want to call <kbd>C-c j<\/kbd> and <kbd>C-c k<\/kbd> in some\n(arbitrary) sequence. Hydra allows you to:<\/p>\n<ul><li>\n<p>Bind your functions in a way that pressing <kbd>C-c jjkk3j5k<\/kbd> is\nequivalent to pressing <kbd>C-c j C-c j C-c k C-c k M-3 C-c j M-5 C-c\nk<\/kbd>. Any key other than <kbd>j<\/kbd> or <kbd>k<\/kbd> exits this state.<\/p>\n<\/li>\n<li>\n<p>Assign a custom hint to this group of functions, so that you know immediately\nafter pressing <kbd>C-c<\/kbd> that you can follow up with <kbd>j<\/kbd> or\n<kbd>k<\/kbd>.<\/p>\n<\/li>\n<\/ul><p>If you want to quickly understand the concept, see <a href=\"https:\/\/www.youtube.com\/watch?v=_qZliI1BKzI\" rel=\"nofollow\">the video demo<\/a>.<\/p>\n<p><strong>Table of Contents<\/strong><\/p>\n<ul><li><a href=\"#sample-hydras\">Sample Hydras<\/a>\n<ul><li><a href=\"#the-one-with-the-least-amount-of-code\">The one with the least amount of code<\/a><\/li>\n<li><a href=\"#the-impressive-looking-one\">The impressive-looking one<\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#community-wiki\">Community wiki<\/a><\/li>\n<li><a href=\"#the-rules-of-hydra-tics\">The Rules of Hydra-tics<\/a>\n<ul><li><a href=\"#hydra-awesome\"><code>hydra-awesome<\/code><\/a><\/li>\n<li><a href=\"#awesome-map-and-awesome-binding\"><code>awesome-map<\/code> and <code>awesome-binding<\/code><\/a><\/li>\n<li><a href=\"#awesome-plist\"><code>awesome-plist<\/code><\/a>\n<ul><li><a href=\"#pre-and-post\"><code>:pre<\/code> and <code>:post<\/code><\/a><\/li>\n<li><a href=\"#exit\"><code>:exit<\/code><\/a><\/li>\n<li><a href=\"#foreign-keys\"><code>:foreign-keys<\/code><\/a><\/li>\n<li><a href=\"#color\"><code>:color<\/code><\/a><\/li>\n<li><a href=\"#timeout\"><code>:timeout<\/code><\/a><\/li>\n<li><a href=\"#hint\"><code>:hint<\/code><\/a><\/li>\n<li><a href=\"#bind\"><code>:bind<\/code><\/a><\/li>\n<\/ul><\/li>\n<li><a href=\"#awesome-docstring\"><code>awesome-docstring<\/code><\/a><\/li>\n<li><a href=\"#awesome-head-1\"><code>awesome-head-1<\/code><\/a>\n<ul><li><a href=\"#head-binding\"><code>head-binding<\/code><\/a><\/li>\n<li><a href=\"#head-command\"><code>head-command<\/code><\/a><\/li>\n<li><a href=\"#head-hint\"><code>head-hint<\/code><\/a><\/li>\n<li><a href=\"#head-plist\"><code>head-plist<\/code><\/a><\/li>\n<\/ul><\/li>\n<\/ul><\/li>\n<\/ul><h2><a id=\"user-content-the-one-with-the-least-amount-of-code\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-one-with-the-least-amount-of-code\"><\/a>The one with the least amount of code<\/h2>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-zoom (global-map \"&lt;f2&gt;\")\n  \"zoom\"\n  (\"g\" text-scale-increase \"in\")\n  (\"l\" text-scale-decrease \"out\"))<\/pre><\/div>\n<p>With this simple code, you can:<\/p>\n<ul><li>Start zooming in with <kbd>&lt;f2&gt; g<\/kbd>.<\/li>\n<li>Continue to zoom in with <kbd>g<\/kbd>.<\/li>\n<li>Or zoom out with <kbd>l<\/kbd>.<\/li>\n<li>Zoom in five times at once with <kbd>5g<\/kbd>.<\/li>\n<li>Stop zooming with <em>any<\/em> key that isn't <kbd>g<\/kbd> or <kbd>l<\/kbd>.<\/li>\n<\/ul><p>For any Hydra:<\/p>\n<ul><li><code>digit-argument<\/code> can be called with <kbd>0<\/kbd>-<kbd>9<\/kbd>.<\/li>\n<li><code>negative-argument<\/code> can be called with <kbd>-<\/kbd>.<\/li>\n<li><code>universal-argument<\/code> can be called with <kbd>C-u<\/kbd>.<\/li>\n<\/ul><h2><a id=\"user-content-the-impressive-looking-one\" class=\"anchor\" aria-hidden=\"true\" href=\"#the-impressive-looking-one\"><\/a>The impressive-looking one<\/h2>\n<p>Here's the result of pressing <kbd>.<\/kbd> in the good-old Buffer menu:<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/c1df3628134ce20bc6c17fd6397fa7fd07461335\/687474703a2f2f6f72656d6163732e636f6d2f646f776e6c6f61642f68796472612d6275666665722d6d656e752e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/c1df3628134ce20bc6c17fd6397fa7fd07461335\/687474703a2f2f6f72656d6163732e636f6d2f646f776e6c6f61642f68796472612d6275666665722d6d656e752e706e67\" alt=\"hydra-buffer-menu\" data-canonical-src=\"http:\/\/oremacs.com\/download\/hydra-buffer-menu.png\" \/><\/a><\/p>\n<p>The code is large but very simple:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-buffer-menu (:color pink\n                             :hint nil)\n  \"\n^Mark^             ^Unmark^           ^Actions^          ^Search\n^^^^^^^^-----------------------------------------------------------------\n_m_: mark          _u_: unmark        _x_: execute       _R_: re-isearch\n_s_: save          _U_: unmark up     _b_: bury          _I_: isearch\n_d_: delete        ^ ^                _g_: refresh       _O_: multi-occur\n_D_: delete up     ^ ^                _T_: files only: % -28`Buffer-menu-files-only\n_~_: modified\n\"\n  (\"m\" Buffer-menu-mark)\n  (\"u\" Buffer-menu-unmark)\n  (\"U\" Buffer-menu-backup-unmark)\n  (\"d\" Buffer-menu-delete)\n  (\"D\" Buffer-menu-delete-backwards)\n  (\"s\" Buffer-menu-save)\n  (\"~\" Buffer-menu-not-modified)\n  (\"x\" Buffer-menu-execute)\n  (\"b\" Buffer-menu-bury)\n  (\"g\" revert-buffer)\n  (\"T\" Buffer-menu-toggle-files-only)\n  (\"O\" Buffer-menu-multi-occur :color blue)\n  (\"I\" Buffer-menu-isearch-buffers :color blue)\n  (\"R\" Buffer-menu-isearch-buffers-regexp :color blue)\n  (\"c\" nil \"cancel\")\n  (\"v\" Buffer-menu-select \"select\" :color blue)\n  (\"o\" Buffer-menu-other-window \"other-window\" :color blue)\n  (\"q\" quit-window \"quit\" :color blue))\n(define-key Buffer-menu-mode-map \".\" 'hydra-buffer-menu\/body)<\/pre><\/div>\n<p>Looking at the code, you can see <code>hydra-buffer-menu<\/code> as sort of a namespace construct that wraps\neach function that it's given in code that shows that hint and makes it easy to call the related\nfunctions. One additional function is created and returned as the result of <code>defhydra<\/code> -\n<code>hydra-buffer-menu\/body<\/code>.  This function does nothing except setting up the hint and the keymap, and\nis usually the entry point to complex hydras.<\/p>\n<p>To write your own hydras, you can:<\/p>\n<ul><li>Either modify an existing hydra to do what you want to do.<\/li>\n<li>Or read <a href=\"#the-rules-of-hydra-tics\">the rules<\/a>,\n<a href=\"https:\/\/github.com\/abo-abo\/hydra\/blob\/master\/hydra-examples.el\">the examples<\/a>,\nthe docstrings and comments in the source.<\/li>\n<\/ul><p>You can find some user created hydras and more documentation in the project's\n<a href=\"https:\/\/github.com\/abo-abo\/hydra\/wiki\/\">community wiki<\/a>. Feel free to add your\nown or edit the existing ones.<\/p>\n<p>Each hydra (take <code>awesome<\/code> as a prefix to make it more specific) looks like this:<\/p>\n<pre>(defhydra hydra-awesome (awesome-map awesome-binding awesome-plist)\n  awesome-docstring\n  awesome-head-1\n  awesome-head-2\n  awesome-head-3\n  ...)\n<\/pre>\n<h2><a id=\"user-content-hydra-awesome\" class=\"anchor\" aria-hidden=\"true\" href=\"#hydra-awesome\"><\/a><code>hydra-awesome<\/code><\/h2>\n<p>Each hydra needs a name, and this one is named <code>hydra-awesome<\/code>. You can name your hydras as you wish,\nbut I prefer to start each one with <code>hydra-<\/code>, because it acts as an additional namespace layer, for example:\n<code>hydra-zoom<\/code>, <code>hydra-helm<\/code>, <code>hydra-apropos<\/code> etc.<\/p>\n<p>If you name your hydra <code>hydra-awesome<\/code>, the return result of <code>defhydra<\/code> will be <code>hydra-awesome\/body<\/code>.<\/p>\n<p>Here's what <code>hydra-zoom\/body<\/code> looks like, if you're interested:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defun hydra-zoom\/body nil\n  \"Create a hydra with a \\\"&lt;f2&gt;\\\" body and the heads:\n\\\"g\\\":    `text-scale-increase',\n\\\"l\\\":    `text-scale-decrease'\nThe body can be accessed via `hydra-zoom\/body'.\"\n  (interactive)\n  (hydra-default-pre)\n  (when hydra-is-helpful\n    (if hydra-lv\n        (lv-message\n         (eval hydra-zoom\/hint))\n      (message\n       (eval hydra-zoom\/hint))))\n  (hydra-set-transient-map\n   hydra-zoom\/keymap\n   (lambda nil\n     (hydra-keyboard-quit)\n     nil)\n   nil)\n  (setq prefix-arg\n        current-prefix-arg))<\/pre><\/div>\n<h2><a id=\"user-content-awesome-map-and-awesome-binding\" class=\"anchor\" aria-hidden=\"true\" href=\"#awesome-map-and-awesome-binding\"><\/a><code>awesome-map<\/code> and <code>awesome-binding<\/code><\/h2>\n<p>This can be any keymap, for instance, <code>global-map<\/code> or <code>isearch-mode-map<\/code>.<\/p>\n<p>For this example:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-zoom (global-map \"&lt;f2&gt;\")\n  \"zoom\"\n  (\"g\" text-scale-increase \"in\")\n  (\"l\" text-scale-decrease \"out\"))<\/pre><\/div>\n<ul><li><code>awesome-map<\/code> is <code>global-map<\/code><\/li>\n<li><code>awesome-binding<\/code> is <code>\"&lt;f2&gt;\"<\/code><\/li>\n<\/ul><p>And here's the relevant generated code:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(unless (keymapp (lookup-key global-map (kbd \"&lt;f2&gt;\")))\n  (define-key global-map (kbd \"&lt;f2&gt;\") nil))\n(define-key global-map [f2 103]\n  (function hydra-zoom\/text-scale-increase))\n(define-key global-map [f2 108]\n  (function hydra-zoom\/text-scale-decrease))<\/pre><\/div>\n<p>As you see, <code>\"&lt;f2&gt;\"<\/code> is used as a prefix for <kbd>g<\/kbd> (char value 103) and <kbd>l<\/kbd>\n(char value 108).<\/p>\n<p>If you don't want to use a map right now, you can skip it like this:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-zoom (nil nil)\n  \"zoom\"\n  (\"g\" text-scale-increase \"in\")\n  (\"l\" text-scale-decrease \"out\"))<\/pre><\/div>\n<p>Or even simpler:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-zoom ()\n  \"zoom\"\n  (\"g\" text-scale-increase \"in\")\n  (\"l\" text-scale-decrease \"out\"))<\/pre><\/div>\n<p>But then you would have to bind <code>hydra-zoom\/text-scale-increase<\/code> and\n<code>hydra-zoom\/text-scale-decrease<\/code> yourself.<\/p>\n<h2><a id=\"user-content-awesome-plist\" class=\"anchor\" aria-hidden=\"true\" href=\"#awesome-plist\"><\/a><code>awesome-plist<\/code><\/h2>\n<p>You can read up on what a plist is in\n<a href=\"https:\/\/www.gnu.org\/software\/emacs\/manual\/html_node\/elisp\/Property-Lists.html\" rel=\"nofollow\">the Elisp manual<\/a>.<\/p>\n<p>You can use <code>awesome-plist<\/code> to modify the behavior of each head in some way.\nBelow is a list of each key.<\/p>\n<h3><a id=\"user-content-pre-and-post\" class=\"anchor\" aria-hidden=\"true\" href=\"#pre-and-post\"><\/a><code>:pre<\/code> and <code>:post<\/code><\/h3>\n<p>You can specify code that will be called before each head, and after the body. For example:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-vi (:pre (set-cursor-color \"#40e0d0\")\n                    :post (progn\n                            (set-cursor-color \"#ffffff\")\n                            (message\n                             \"Thank you, come again.\")))\n  \"vi\"\n  (\"l\" forward-char)\n  (\"h\" backward-char)\n  (\"j\" next-line)\n  (\"k\" previous-line)\n  (\"q\" nil \"quit\"))<\/pre><\/div>\n<p>Thanks to <code>:pre<\/code>, each time any head is called, the cursor color is changed.\nAnd when the hydra quits, the cursor color will be made black again with <code>:post<\/code>.<\/p>\n<h3><a id=\"user-content-exit\" class=\"anchor\" aria-hidden=\"true\" href=\"#exit\"><\/a><code>:exit<\/code><\/h3>\n<p>The <code>:exit<\/code> key is inherited by every head (they can override it) and influences what will happen\nafter executing head's command:<\/p>\n<ul><li><code>:exit nil<\/code> (the default) means that the hydra state will continue - you'll still see the hint and be able to use short bindings.<\/li>\n<li><code>:exit t<\/code> means that the hydra state will stop.<\/li>\n<\/ul><h3><a id=\"user-content-foreign-keys\" class=\"anchor\" aria-hidden=\"true\" href=\"#foreign-keys\"><\/a><code>:foreign-keys<\/code><\/h3>\n<p>The <code>:foreign-keys<\/code> key belongs to the body and decides what to do when a key is pressed that doesn't\nbelong to any head:<\/p>\n<ul><li><code>:foreign-keys nil<\/code> (the default) means that the hydra state will stop and the foreign key will\ndo whatever it was supposed to do if there was no hydra state.<\/li>\n<li><code>:foreign-keys warn<\/code> will not stop the hydra state, but instead will issue a warning without\nrunning the foreign key.<\/li>\n<li><code>:foreign-keys run<\/code> will not stop the hydra state, and try to run the foreign key.<\/li>\n<\/ul><h3><a id=\"user-content-color\" class=\"anchor\" aria-hidden=\"true\" href=\"#color\"><\/a><code>:color<\/code><\/h3>\n<p>The <code>:color<\/code> key is a shortcut. It aggregates <code>:exit<\/code> and <code>:foreign-keys<\/code> key in the following way:<\/p>\n<pre>| color    | toggle                     |\n|----------+----------------------------|\n| red      |                            |\n| blue     | :exit t                    |\n| amaranth | :foreign-keys warn         |\n| teal     | :foreign-keys warn :exit t |\n| pink     | :foreign-keys run          |\n<\/pre>\n<p>It's also a trick to make you instantly aware of the current hydra keys that you're about to press:\nthe keys will be highlighted with the appropriate color.<\/p>\n<h3><a id=\"user-content-timeout\" class=\"anchor\" aria-hidden=\"true\" href=\"#timeout\"><\/a><code>:timeout<\/code><\/h3>\n<p>The <code>:timeout<\/code> key starts a timer for the corresponding amount of seconds that disables the hydra.\nCalling any head will refresh the timer.<\/p>\n<h3><a id=\"user-content-hint\" class=\"anchor\" aria-hidden=\"true\" href=\"#hint\"><\/a><code>:hint<\/code><\/h3>\n<p>The <code>:hint<\/code> key will be inherited by each head. Each head is allowed to override it, of course.\nOne value that makes sense is <code>:hint nil<\/code>. See below for an explanation of head hint.<\/p>\n<h3><a id=\"user-content-bind\" class=\"anchor\" aria-hidden=\"true\" href=\"#bind\"><\/a><code>:bind<\/code><\/h3>\n<p>The <code>:bind<\/code> key provides a lambda to be used to bind each head.  This is quite advanced and rarely\nused, you're not likely to need it.  But if you would like to bind your heads with e.g. <code>bind-key<\/code>\ninstead of <code>define-key<\/code> you can use this option.<\/p>\n<p>The <code>:bind<\/code> key can be overridden by each head. This is useful if you want to have a few heads that\nare not bound outside the hydra.<\/p>\n<h3><a id=\"user-content-base-map\" class=\"anchor\" aria-hidden=\"true\" href=\"#base-map\"><\/a><code>:base-map<\/code><\/h3>\n<p>Use this option if you want to override <code>hydra-base-map<\/code> for the current hydra.<\/p>\n<h2><a id=\"user-content-awesome-docstring\" class=\"anchor\" aria-hidden=\"true\" href=\"#awesome-docstring\"><\/a><code>awesome-docstring<\/code><\/h2>\n<p>This can be a simple string used to build the final hydra hint.  However, if you start it with a\nnewline, the key-highlighting and Ruby-style string interpolation becomes enabled, as you can see in\n<code>hydra-buffer-menu<\/code> above.<\/p>\n<p>To highlight a key, just wrap it in underscores. Note that the key must belong to one of the heads.\nThe key will be highlighted with the color that is appropriate to the behavior of the key, i.e.  if\nthe key will make the hydra exit, the color will be blue.<\/p>\n<p>To insert an empty character, use <code>^<\/code>. The only use of this is to have your code aligned as\nnicely as the result.<\/p>\n<p>To insert a dynamic Elisp variable, use <code>%<\/code>` followed by the variable. Each time the variable\nchanges due to a head, the docstring will be updated. <code>format<\/code>-style width specifiers can be used.<\/p>\n<p>To insert a dynamic Elisp expression, use e.g. <code>%(length (dired-get-marked-files))<\/code>.  If a head will\nchange the amount of marked files, for example, it will be appropriately updated.<\/p>\n<p>If the result of the Elisp expression is a string and you don't want to quote it, use this form:\n<code>%s(shell-command-to-string \"du -hs\")<\/code>.<\/p>\n<h2><a id=\"user-content-awesome-head-1\" class=\"anchor\" aria-hidden=\"true\" href=\"#awesome-head-1\"><\/a><code>awesome-head-1<\/code><\/h2>\n<p>Each head looks like this:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(head-binding head-command head-hint head-plist)<\/pre><\/div>\n<p>For the head <code>(\"g\" text-scale-increase \"in\")<\/code>:<\/p>\n<ul><li><code>head-binding<\/code> is <code>\"g\"<\/code>.<\/li>\n<li><code>head-command<\/code> is <code>text-scale-increase<\/code>.<\/li>\n<li><code>head-hint<\/code> is <code>\"in\"<\/code>.<\/li>\n<li><code>head-plist<\/code> is <code>nil<\/code>.<\/li>\n<\/ul><h3><a id=\"user-content-head-binding\" class=\"anchor\" aria-hidden=\"true\" href=\"#head-binding\"><\/a><code>head-binding<\/code><\/h3>\n<p>The <code>head-binding<\/code> is a string that can be passed to <code>kbd<\/code>.<\/p>\n<h3><a id=\"user-content-head-command\" class=\"anchor\" aria-hidden=\"true\" href=\"#head-command\"><\/a><code>head-command<\/code><\/h3>\n<p>The <code>head-command<\/code> can be:<\/p>\n<ul><li>\n<p>command name, like <code>text-scale-increase<\/code>.<\/p>\n<\/li>\n<li>\n<p>a lambda, like<\/p>\n<pre>  (\"g\" (lambda ()\n         (interactive)\n         (let ((current-prefix-arg 4))\n           (call-interactively #'magit-status)))\n       \"git\")\n<\/pre>\n<\/li>\n<li>\n<p>nil, which exits the hydra.<\/p>\n<\/li>\n<li>\n<p>a single sexp, which will be wrapped in an interactive lambda.<\/p>\n<\/li>\n<\/ul><p>Here's an example of the last option:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-launcher (:color blue)\n   \"Launch\"\n   (\"h\" man \"man\")\n   (\"r\" (browse-url \"http:\/\/www.reddit.com\/r\/emacs\/\") \"reddit\")\n   (\"w\" (browse-url \"http:\/\/www.emacswiki.org\/\") \"emacswiki\")\n   (\"s\" shell \"shell\")\n   (\"q\" nil \"cancel\"))\n(global-set-key (kbd \"C-c r\") 'hydra-launcher\/body)<\/pre><\/div>\n<h3><a id=\"user-content-head-hint\" class=\"anchor\" aria-hidden=\"true\" href=\"#head-hint\"><\/a><code>head-hint<\/code><\/h3>\n<p>In case of a large body docstring, you usually don't want the head hint to show up, since\nyou've already documented it in the body docstring.\nYou can set the head hint to <code>nil<\/code> to do this.<\/p>\n<p>Example:<\/p>\n<div class=\"highlight highlight-source-lisp\"><pre>(defhydra hydra-zoom (global-map \"&lt;f2&gt;\")\n  \"\nPress _g_ to zoom in.\n\"\n  (\"g\" text-scale-increase nil)\n  (\"l\" text-scale-decrease \"out\"))<\/pre><\/div>\n<h3><a id=\"user-content-head-plist\" class=\"anchor\" aria-hidden=\"true\" href=\"#head-plist\"><\/a><code>head-plist<\/code><\/h3>\n<p>Here's a list of body keys that can be overridden in each head:<\/p>\n<ul><li><code>:exit<\/code><\/li>\n<li><code>:color<\/code><\/li>\n<li><code>:bind<\/code><\/li>\n<\/ul><\/article>","created_at":"2018-11-07T16:21:57+0000","updated_at":"2018-11-07T16:21:57+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":9,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/2937359?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":27,"title":"25+ Best Free Landing Page WordPress Themes 2018 - Colorlib","url":"https:\/\/colorlib.com\/wp\/free-landing-page-wordpress-themes\/","content":"<ul class=\"meta clr meta-with-title\"><li class=\"meta-date\"><time class=\"updated\" datetime=\"2018-10-30\" itemprop=\"datePublished\" pubdate=\"\">October 30, 2018<\/time><\/li><li class=\"meta-author\"><\/li><li class=\"meta-category\"><a href=\"https:\/\/colorlib.com\/wp\/category\/free-themes\/\" title=\"Free Themes\" class=\"term-2161\" data-wpel-link=\"internal\">Free Themes<\/a><\/li><li class=\"meta-comments comment-scroll\"><a href=\"https:\/\/colorlib.com\/wp\/free-landing-page-wordpress-themes\/#view_comments\" class=\"comments-link\" data-wpel-link=\"internal\">1 Comment<\/a><\/li><\/ul><div class=\"single-blog-content entry clr\" itemprop=\"text\"><div>   <ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-6200795649377457\" data-ad-slot=\"4162710921\">  <\/ins><\/div><p>The landing page is like a billboard for your product, it\u2019s capable of enticing or scaring your customers away, though in most cases you are looking to retain them. On the web, to capture the attention of a potential customer we typically rely on landing pages; specifically designed websites and\/or separate parts of a website that are created to promote a specific product, service, or anything related to customer acquisition. The landing page is where the visitor can for the first time, learn about your product(s) and how they are useful.<\/p><p>WordPress excels in this niche because it provides so much variety in terms of content alignment, and the integration of third-party tools that further help to create a unique and professional design experience. While no single theme can be called the best, we do recommend for webmasters to understand their own product, before deciding on a final theme. The free landing page themes for WordPress that you\u2019re about to explore, have been checked for their quality, list of features, and overall feel of design \u2014 to reflect modern industry standards, in particular things like responsive design, and a variety of homepage sections.<\/p><p>Also, we encourage that you visit our other roundup, which is about <a href=\"https:\/\/colorlib.com\/wp\/landing-page-wordpress-themes\/\" data-wpel-link=\"internal\">premium landing page themes for WordPress<\/a> \u2014 that list has further 20+ choices of themes, though the difference is that those themes come from premium developers, so the quality and ability to customize will go up significantly. That is not to say though, that the themes below are not of good quality, we always ensure that our roundups feature only the best of the best designs.<\/p><h2>Shapely<\/h2><p><img class=\"aligncenter size-full wp-image-95234 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"970\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme-300x243.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme-768x621.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme-1024x828.jpg 1024w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-95234\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme.jpg\" alt=\"\" width=\"1200\" height=\"970\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme-300x243.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme-768x621.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/shapely-wordpress-theme-1024x828.jpg 1024w\" \/><\/noscript>The business world is famous for its tendency to maximize efficiency, and prioritize quality in order to increase profits. Themes like Shapely have pledged to reflect those qualities, while promising to enhance your business page. A standard, one-page format was chosen, due to its practicality and ease with which it can be navigated. Shapely is ideal for any landing page, business, store, ecommerce, or freelancer website. Despite this niche appeal, it can accommodate almost any other field of activity.<p>Many of the most prolific and sought-after WordPress plugins will be compatible with your page: Yoast SEO, Google Analytics by Yoast, Contact Form 7, Gravity Forms, and Jetpack.\u00a0 It is also possible to expand your current business model, by introducing the WooCommerce plugin. Thanks to this innovative tool, anyone can craft a fully-functional online store in a matter of moments. Selling your high-quality merchandise has never been easier.<\/p><p>A smart businessman knows the field on which the game is being played. When it comes to the internet, those who manage to understand search engines will come out on top. Shapely is SEO-ready, and this search engine optimization makes it more likely for your site to appear on results listings.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow\" href=\"https:\/\/colorlib.com\/wp\/shapely\/\" data-wpel-link=\"internal\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/shapely\/\" data-wpel-link=\"external\">Demo<\/a><h2>Illdy<\/h2><p><img class=\"aligncenter size-full wp-image-84136 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"957\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme-300x239.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme-768x612.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme-1024x817.jpg 1024w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-84136\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme.jpg\" alt=\"\" width=\"1200\" height=\"957\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme-300x239.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme-768x612.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/illdy-free-business-landing-page-theme-1024x817.jpg 1024w\" \/><\/noscript><p>For the last theme of our best free landing page themes roundup, we will leave you with the Illdy theme \u2014 free WordPress theme for application developers, software enthusiasts, and anyone who runs a product-based business. From top to bottom, The App is built to withstand critique and demands of your visitors. From custom 404 pages, to custom page types. The App can tackle SEO issues without needing any additional plugins.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow\" href=\"https:\/\/colorlib.com\/wp\/themes\/illdy\/\" data-wpel-link=\"internal\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-3\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/illdy\/\" data-wpel-link=\"external\">Demo<\/a><h2>Antreas<\/h2><p><img class=\"aligncenter wp-image-193058 size-full jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"891\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-300x223.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-768x570.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-1024x760.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-600x446.jpg 600w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter wp-image-193058 size-full\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot.jpg\" alt=\"\" width=\"1200\" height=\"891\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-300x223.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-768x570.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-1024x760.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/antreas-screenshot-600x446.jpg 600w\" \/><\/noscript><p>Antreas is a swift and simple free WordPress business website theme. It\u2019s the platform you\u2019ve been waiting for to take your business online. Easy to use and very intuitive, Antreas makes the process painless and quick. WP Customizer makes editing your pages and sections a visual, coding-free task. Drop your favorite widgets into sidebars and sections. Add as many as you wish, thanks to Antreas\u2019s broad plugin compatibility. Convenient shortcodes aimed at business user needs are plentiful within Antreas. With them, guiding users to your physical location with Google Maps is effortless. You can also introduce your product or service line, pricing and features in the blink of an eye. Several layout and grid styles and hover animations keep users engaged and interested. There\u2019s finally no excuse not to have an elegant, functional website for your business. Get Antreas for free today!<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/out\/antreas\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-3\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/out\/antreas-demo\/\" data-wpel-link=\"external\">Demo<\/a><h2>Portum Material<\/h2><p><img class=\"aligncenter size-full wp-image-175762 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"900\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-768x576.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-1024x768.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-600x450.jpg 600w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-175762\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material.jpg\" alt=\"\" width=\"1200\" height=\"900\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-768x576.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-1024x768.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-material-600x450.jpg 600w\" \/><\/noscript><br \/><a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/portum-material\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-3\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/portum-material\/\" data-wpel-link=\"external\">Demo<\/a><h2>Pixova Lite<\/h2><p><img class=\"aligncenter size-full wp-image-147864 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"981\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme-300x245.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme-768x628.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme-1024x837.jpg 1024w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-147864\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme.jpg\" alt=\"\" width=\"1200\" height=\"981\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme-300x245.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme-768x628.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/pixova-lite-free-parallax-wordpress-theme-1024x837.jpg 1024w\" \/><\/noscript><p>Pixova Lite is a WordPress theme with a minimalistic and modern look. If you are a freelancer, artist or an e-commerce shop owner you should totally try it. It is very easy to use and free to download! Pixova Lite has several theme options to customize areas with your personal touch. Commercial transactions are done with the amazing WooCommerce. Pixova Lite is versatile and built for any type of task given. It comes with a responsive layout compatible with all screens and browsers. It is considered too to be particularly mobile friendly.<\/p><p>You will get it with unlimited colors and pre-built choices for headers and footers. Google Fonts and tons of icons sets are provided as well! Pixova Lite features a sidebar and smooth one-page scrolling for main design. Blogs and contact pages for readers or audience to give feedback are available. It also supports blogs and testimonies from happy clients. Translation files and WPML compatibility are included. Pixova Lite has a strong code made with validation that erases compatibility issues. It offers smooth Parallax sections to divide your one-page style with some edge. Get all these benefits for free with a single one-click install! Or go with its paid pro version too! Download Pixova Lite!<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow\" href=\"https:\/\/colorlib.com\/wp\/themes\/pixova-lite\/\" data-wpel-link=\"internal\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/pixova-lite\" data-wpel-link=\"external\">Demo<\/a><h2>Ascendant<\/h2><p><img class=\"aligncenter size-full wp-image-95226 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"990\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme-300x248.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme-768x634.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme-1024x845.jpg 1024w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-95226\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme.jpg\" alt=\"\" width=\"1200\" height=\"990\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme-300x248.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme-768x634.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/ascendant-simple-landing-page-theme-1024x845.jpg 1024w\" \/><\/noscript><p>Ascendant is a newly released multipurpose WordPress. It is made for all kinds of pages and setups. It is ideal for starters, small businesses or individuals looking to express themselves. However, it has 3 different versions to adjust to the client\u2019s needs. Start with the free version, and you will get convinced to step up. Ascendant has a layout equipped with endless styles and customizations. You will get countless color schemes, post pages, and numerous sidebars. It also features Google Fonts and WooCommerce plug-ins. Create a crafty modern looking website simple and easy with its featured shortcodes. Get this all browsers compatible theme now\u2026Get Ascendant!<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/cpothemes.com\/theme\/ascendant\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/cpothemes.com\/theme\/ascendant http:\/\/demos.cpothemes.com\/ascendant\" data-wpel-link=\"external\">Demo<\/a><h2>Portum One<\/h2><p><img class=\"aligncenter size-full wp-image-175765 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"900\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-768x576.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-1024x768.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-600x450.jpg 600w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-175765\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one.jpg\" alt=\"\" width=\"1200\" height=\"900\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-768x576.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-1024x768.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-one-600x450.jpg 600w\" \/><\/noscript><br \/><a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/portum-one\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-3\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/portum-one\/\" data-wpel-link=\"external\">Demo<\/a><h2>Unite<\/h2><p><img class=\"aligncenter size-full wp-image-94268 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min.jpg?is-pending-load=1\" alt=\"Home Unite Demo-min\" width=\"1024\" height=\"481\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min-300x141.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min-768x361.jpg 768w\" data-lazy-sizes=\"(max-width: 1024px) 100vw, 1024px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94268\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min.jpg\" alt=\"Home Unite Demo-min\" width=\"1024\" height=\"481\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min-300x141.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Home-Unite-Demo-min-768x361.jpg 768w\" \/><\/noscript><p>Unite is our personal WordPress theme that we\u2019re offering in our free themes library. While the main intention behind Unite is for this theme to be used on Wedding websites, we know for a fact that many of our theme users are using Unite as a landing page template for individual projects, and even products in some cases.<\/p><p>The page design is fully customizable, and if you\u2019re using a page builder such as Visual Composer \u2014 you\u2019ll have no trouble whatsoever to create a pleasant landing page website for your next big idea. Unite includes custom shortcodes, Bootstrap 3 design patterns, a mobile responsive design, and many ways for you to customize the theme for your own liking.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow\" href=\"https:\/\/colorlib.com\/wp\/unite\/\" data-wpel-link=\"internal\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/unite\/\" data-wpel-link=\"external\">Demo<\/a><h2>Portum Clean<\/h2><p><img class=\"aligncenter size-full wp-image-175764 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"900\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-768x576.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-1024x768.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-600x450.jpg 600w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-175764\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean.jpg\" alt=\"\" width=\"1200\" height=\"900\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-768x576.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-1024x768.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Portum-clean-600x450.jpg 600w\" \/><\/noscript><a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/portum-clean\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-3\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/colorlib.com\/portum-clean\/\" data-wpel-link=\"external\">Demo<\/a><h2>Panoramica<\/h2><p><img class=\"aligncenter size-full wp-image-94274 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1.jpg?is-pending-load=1\" alt=\"Panoramica Pro \u2013 Just another CPOThemes Collection site-min\" width=\"1024\" height=\"481\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1-300x141.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1-768x361.jpg 768w\" data-lazy-sizes=\"(max-width: 1024px) 100vw, 1024px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94274\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1.jpg\" alt=\"Panoramica Pro \u2013 Just another CPOThemes Collection site-min\" width=\"1024\" height=\"481\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1-300x141.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Panoramica-Pro-\u2013-Just-another-CPOThemes-Collection-site-min-1-768x361.jpg 768w\" \/><\/noscript><p>Panoramica is a landing page for photographers, for creative people, and those who appreciate a more earthy touch to website design. We say that not because Panoramica doesn\u2019t feel visually impactful, quite the contrary actually; the theme has a modern yet classical feel to it, with a boxed content layout that puts your portfolio at the center of the attention.<\/p><p>The list of features isn\u2019t too shabby either. Panormaica is sporting custom shortcodes, flexible typography options across all pages, it\u2019s packaged with demo content, you can easily translate the theme in other languages, the theme options are powered by a powerful options panel, you can choose from limitless color combinations, and if all else fails \u2014 there\u2019s a friendly support team waiting to answer your queries.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/cpothemes.com\/theme\/panoramica\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/demos.cpothemes.com\/panoramicapro\/\" data-wpel-link=\"external\">Demo<\/a><h2>Illustrious<\/h2><p><img class=\"aligncenter size-full wp-image-94269 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min.jpg?is-pending-load=1\" alt=\"Illustrious-min\" width=\"1024\" height=\"481\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min-300x141.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min-768x361.jpg 768w\" data-lazy-sizes=\"(max-width: 1024px) 100vw, 1024px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94269\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min.jpg\" alt=\"Illustrious-min\" width=\"1024\" height=\"481\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min-300x141.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Illustrious-min-768x361.jpg 768w\" \/><\/noscript><p>Similarly to the theme shown above, Illustrious is also a landing page theme for WordPress users, in particular those who specialize in photography, and other fields of interest where portfolio showcasing might be the number one priority. Everything in this beautiful theme seems to just work, from a transparent menu navigation, to a boxed layout slider with striking visual images.<\/p><p>If you are an agency, you can use the services components to outline your services, while the portfolio section is primarily designed for you to cover your expertise in the field. The portfolio item pages have their own design, making the design more personalized.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/cpothemes.com\/theme\/illustrious\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/demos.cpothemes.com\/illustriouspro\/\" data-wpel-link=\"external\">Demo<\/a><h2>MedZone<\/h2><p><img class=\"aligncenter size-full wp-image-126554 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme.jpg?is-pending-load=1\" alt=\"\" width=\"1200\" height=\"1006\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme-300x252.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme-768x644.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme-1024x858.jpg 1024w\" data-lazy-sizes=\"(max-width: 1200px) 100vw, 1200px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-126554\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme.jpg\" alt=\"\" width=\"1200\" height=\"1006\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme.jpg 1200w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme-300x252.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme-768x644.jpg 768w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/medzone-free-medical-wordpress-theme-1024x858.jpg 1024w\" \/><\/noscript><p>MedZone is a brand new premium quality WordPress theme. It has a very strong orientation to medical institutions and its variants. It is a one-page minimalistic design with smooth scrolling. Find tons of light and easy customizations oriented to doctors and services. Medzone has an appointment booking system, a log-in account system and contact pages. It also provides staff pages to all the details of those taking care of you. You will find many special services for admission, emergency online alert, and schedules.<\/p><p>MedZone gives you a flexible layout with sections for recent posts a large header and footer too. Colors and posts can be handled with advanced controllers. It is 100% responsive and retina ready, basically made for all browsers and devices. It even comes with a clean code that makes it run smoothly! MedZone is enhanced on speed and retina ready. You will get great image quality and fast loading. For commercial aspects, you will get WooCommerce plugin integration. MedZone provides awesome features for texts on several counts. Unlimited Google Fonts and WPML translations. Make your business greater and connected to people with this tool! Have a good look at it and use it right away! Use MedZone!<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/www.machothemes.com\/theme\/medzone-lite\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-3\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/www.demos.machothemes.com\/medzone-lite\/\" data-wpel-link=\"external\">Demo<\/a><h2>Book Landing Page<\/h2><p><img class=\"aligncenter size-full wp-image-94262 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min.jpg?is-pending-load=1\" alt=\"Book Landing Page-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94262\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min.jpg\" alt=\"Book Landing Page-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Book-Landing-Page-min-768x577.jpg 768w\" \/><\/noscript><p>Landing pages come in many different sizes and shapes, and quite often you will see that specific WordPress themes are built for a single purpose only, as is the case with this magnificent Book Landing Page \u2014 designed to be used by book authors and agencies that wish to highlight their books in their own special page design. Remember, with some modification you can use this theme on your existing WordPress site.<\/p><p>Rara Themes build high quality products over and over again, and this theme can be trusted to take care of your book publishing needs, at least for the web that is. One feature in particular we thought was clever, was the book preview element in the form of preview chapters. Give your potential book customers a chance to read the first few pages of your book to give them a taste of your writing style.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/book-landing-page\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/raratheme.com\/preview\/book-landing-page\/\" data-wpel-link=\"external\">Demo<\/a><h2>Construction Landing Page<\/h2><p><img class=\"aligncenter size-full wp-image-94265 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min.jpg?is-pending-load=1\" alt=\"Construction Landing Page-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94265\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min.jpg\" alt=\"Construction Landing Page-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Construction-Landing-Page-min-768x577.jpg 768w\" \/><\/noscript><p>Here is another stellar design work from Rara Themes, this time you\u2019re looking at a landing page theme that\u2019s built for construction businesses, including agencies and contractors. The design has a clean feeling to it, with eye catching content elements, and a brilliant parallax integration that will keep customers engaged and happy.<\/p><p>Immediately upon arriving on the homepage of Construction Landing Page theme, you\u2019ll notice a huge \u201cGet a Quote\u201d form attached to the right side of the header of the page, right below the beautiful menu. This element instantly lets your visitors know that you\u2019re accepting reservations and bookings using your website, which then allows them more freely browse the rest of your landing page to see if you are potentially a great fit for their needs.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/construction-landing-page\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/raratheme.com\/preview\/construction-landing-page\/\" data-wpel-link=\"external\">Demo<\/a><h2>Lawyer Landing Page<\/h2><p><img class=\"aligncenter size-full wp-image-94271 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min.jpg?is-pending-load=1\" alt=\"Lawyer Landing Page-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94271\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min.jpg\" alt=\"Lawyer Landing Page-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lawyer-Landing-Page-min-768x577.jpg 768w\" \/><\/noscript><p>We have to give Rara Themes a huge props, these last 3 themes (including this one) have all been from them, and as you can tell; they\u2019re very niche specific, but don\u2019t fail to impress you with custom features and tools that would suit those given niche situations. Lawyer Landing Page isn\u2019t an exception, and includes a similar quoting form that we saw above in the Construction theme. What better way to reuse design patterns, especially when they\u2019re relevant.<\/p><p>As someone who works in the law, perhaps you\u2019re an agency altogether, you will be looking to cover some common ground for your visitors. That includes your specialty of law areas, types of services you typically provide, highlight what your past customers have said about you, and of course, give your team members their own mention so that visitors know whom they are going to be working with in the case of using your law firm for their requirements.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/lawyer-landing-page\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/raratheme.com\/preview\/lawyer-landing-page\/\" data-wpel-link=\"external\">Demo<\/a><h2>Variant Landing Page<\/h2><p><img class=\"aligncenter size-full wp-image-94278 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min.jpg?is-pending-load=1\" alt=\"Variant Landing Page-min\" width=\"1024\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min-768x576.jpg 768w\" data-lazy-sizes=\"(max-width: 1024px) 100vw, 1024px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94278\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min.jpg\" alt=\"Variant Landing Page-min\" width=\"1024\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Variant-Landing-Page-min-768x576.jpg 768w\" \/><\/noscript><p>Variant is a WordPress landing page theme with a fully responsive design. It\u2019s optimized for the latest version of WordPress, and is based on WordPress Customizer for its design flexibility. Variant has its own massive documentation file explaining how the theme works, and the design itself can be used on any number of sites without any restrictions.<\/p><p>If you want to capture leads, and increase your conversion rates with a professional design; look no further than what Variant has to offer. The theme takes only a few minutes to install and setup, and the user interface is based on minimal standards so that you can quickly achieve your goals.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/variant-landing-page\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/variant-landing-page\/\" data-wpel-link=\"external\">Demo<\/a><h2>Layla<\/h2><p><img class=\"aligncenter size-full wp-image-94280 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla.jpg?is-pending-load=1\" alt=\"Layla\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94280\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla.jpg\" alt=\"Layla\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Layla-768x577.jpg 768w\" \/><\/noscript><p>Layla is a new landing page theme, with a lovely design, a multi-purpose functionality, and lots of character that is going to define the user experience for your visitors. It\u2019s ready to help you increase your eCommerce sales with a seamless WooCommerce integration, likewise it is powered by a pristine homepage design; ready to take on even the most difficult of design critics. The blog design has its own layout allocated, guaranteeing that in the case of you wanting to publish content regularly \u2014 the posts will look and feel great.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/layla\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/layla\/\" data-wpel-link=\"external\">Demo<\/a><h2>Courtyard<\/h2><p><img class=\"aligncenter size-full wp-image-94266 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min.jpg?is-pending-load=1\" alt=\"Courtyard-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94266\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min.jpg\" alt=\"Courtyard-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Courtyard-min-768x577.jpg 768w\" \/><\/noscript><p>Precise Themes are known for the clean and minimal approach to designing WordPress themes. Courtyard is a landing page template specifically designed for Hotels, Resorts, and Guesthouses. It\u2019s fully ready to take on your visitors, and eventually onboard them as your new customers. Using the full-width hero slider on the homepage, you can create a strong first impression for new visitors to your platform. Using a featured called \u201cRoom Slider\u201d you can create custom room listings and highlight their individual features, and pricing, of course.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/courtyard\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/demo.precisethemes.com\/courtyard\/\" data-wpel-link=\"external\">Demo<\/a><h2>Metrolo<\/h2><p><img class=\"aligncenter size-full wp-image-94282 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo.jpg?is-pending-load=1\" alt=\"Metrolo\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94282\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo.jpg\" alt=\"Metrolo\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metrolo-768x577.jpg 768w\" \/><\/noscript><p>Metrolo might feel old-school from the first glance, but more often than not it is these \u201cold school\u201d designs that most visitors find easy to navigate, and generally easy to understand. It\u2019s a flexible WordPress landing page template, ideal for professionals and small business owners alike. With powerful post and page customization options, anyone can quickly become a master of their own design. Metrolo is most commonly seen on business websites related to travel, construction, photography, graphics design, education, weddings, restaurants, and similar service businesses.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/metrolo\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/demo.wphoot.com\/metrolo\/\" data-wpel-link=\"external\">Demo<\/a><h2>BB Mobile Application<\/h2><p><img class=\"aligncenter size-full wp-image-94261 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min.jpg?is-pending-load=1\" alt=\"\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94261\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min.jpg\" alt=\"\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/BB-Mobile-Application-min-768x577.jpg 768w\" \/><\/noscript><p>It\u2019s pretty clear that mobile apps aren\u2019t going to go anywhere, not in the near future. As a matter of fact, the demand for mobile apps will continue to grow, and so will the demand for high-quality WordPress theme designs that can act as landing pages for mobile apps. This is why we are putting in this roundup the BB Mobile App theme. It\u2019s a responsive WP theme, powered by a clean design and powerful features, the exact recipe you need for explaining what your mobile app is for, and how it can benefit the person using it. The design of BB Mobile feels unforced, effortless to navigate, and clearly shows that quite a bit of thinking has gone into the process of creating the finalized version of such a beautiful theme.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/bb-mobile-application\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/www.themeshopy.com\/bb-mobile-application-theme\/\" data-wpel-link=\"external\">Demo<\/a><h2>Lighthouse<\/h2><p><img class=\"aligncenter size-full wp-image-94272 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min.jpg?is-pending-load=1\" alt=\"Lighthouse-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94272\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min.jpg\" alt=\"Lighthouse-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Lighthouse-min-768x577.jpg 768w\" \/><\/noscript><p>Lighthouse wasn\u2019t created only to serve as a landing page, this is a multi-purpose theme that\u2019s focused around improving your search engine rankings, especially on sites like Google. Each page, each element, everything in the Lighthouse theme design has a meaning, and carefully designed schema markup so that search engines can easily interpret that meaning, and display it on search engine result pages.<\/p><p>While writers and bloggers are getting a kick out of the Lighthouse design full speed, other industries seen using Lighthouse theme are travel, magazines, affiliate sites, startups, and product launch sites. Lighthouse as a theme is heavily widgetized, creating a way for you to add custom content to many different parts of your website.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/lighthouse\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/lighthouse\/\" data-wpel-link=\"external\">Demo<\/a><h2>Winsome<\/h2><p><img class=\"aligncenter size-full wp-image-94279 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1.jpg?is-pending-load=1\" alt=\"Winsome-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94279\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1.jpg\" alt=\"Winsome-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Winsome-min-1-768x577.jpg 768w\" \/><\/noscript><p>Winsome is easily going to win over your attention, as soon as you lay your eyes on that captivating multi-page design, you know you\u2019re in for a pleasant WP theme adventure. Let\u2019s talk about Winsome\u2019s features, both Free and Pro versions provide users a reasonable amount of tools for maximizing the authenticity of your website. Theme Options is for customizing the little small details about your website, for fonts and colors you have further separate options panels. Winsome has been tested to work across tablets, and popular smartphone devices that are powered by Android and iOS. Winsome\u2019s multi-page design has a tremendous blog design, one worth checking out just for the sake of seeing how creative these guys are.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/winsome\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/promenadethemes.com\/demo\/winsome\/multi-page\/\" data-wpel-link=\"external\">Demo<\/a><h2>Marvel<\/h2><p><img class=\"aligncenter size-full wp-image-94281 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1.jpg?is-pending-load=1\" alt=\"Marvel\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94281\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1.jpg\" alt=\"Marvel\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Marvel-1-768x577.jpg 768w\" \/><\/noscript><p>Marvel is a child-theme for the Safreen theme, both are ultra flexible and ultra customizable WordPress themes for landing page situations. The Marvel themes relies heavily on unique functionality that\u2019s structured using jQuery, CSS3, and HTML5. With smart features, industry level design, and a friendly developer who\u2019s happy to answer your questions; Marvel has the potential to be your household theme for personal and creative projects much the same.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/marvel\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/marvel\/\" data-wpel-link=\"external\">Demo<\/a><h2>Business Elite<\/h2><p><img class=\"aligncenter size-full wp-image-94263 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min.jpg?is-pending-load=1\" alt=\"Business Elite-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94263\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min.jpg\" alt=\"Business Elite-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Business-Elite-min-768x577.jpg 768w\" \/><\/noscript><p>Business Elite is a flagship theme from Web Dorado. Whenever you see a theme developer use the word \u2018Elite\u2019 in their work, you can pretty much expect phenomenal quality, no questions asked. The fact that Business Elite has 95% 5-star ratings should help to reassure of the fact. As a landing page theme, Business Elite is oriented around a search engine friendly web design, but likewise to be a unique theme; which is achieved through animation effects, powerful sliders, custom fonts, and compatibility with leading WordPress plugins.<\/p><p>It happens occasionally, you find a really beautiful WordPress theme, but as it turns out it doesn\u2019t work with your favorite plugins, and that can be a total bummer! Business Elite tells you beforehand, which plugins it is compatible with, so you can decide for yourself. Generally, the Web Dorado developers have a good reputation around the WP industry, so it\u2019s worth trusting them.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/business-elite\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/themedemo.web-dorado.com\/theme-businesselite\/\" data-wpel-link=\"external\">Demo<\/a><h2>Metlux<\/h2><p><img class=\"aligncenter size-full wp-image-94273 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min.jpg?is-pending-load=1\" alt=\"Metlux-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94273\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min.jpg\" alt=\"Metlux-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Metlux-min-768x577.jpg 768w\" \/><\/noscript><p>Metlux has a wicked responsive design, with unique features like fixed navigation (with animated effects), the CSS codebase sits on SASS; allowing you to make more meaningful design changes, it\u2019s integrated with custom-built plugins for higher functionality, the markup of the design is semantic-based for better organization, there are unique shortcodes for things like buttons, and you can create as many full-scale header images as required.<\/p><p>Sure, many WordPress themes will reuse the same principles, over and over again.. it\u2019s part of the business, but Metlux definitely brings something extra with it, even simple widgets like \u201cteam members\u201d are so uniquely designed, they require a second look just to appreciate their creative beauty.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/metlux\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/oceanwebthemes.com\/themes\/metlux\/\" data-wpel-link=\"external\">Demo<\/a><h2>Prolific<\/h2><p><img class=\"aligncenter size-full wp-image-94283 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific.jpg?is-pending-load=1\" alt=\"Prolific\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94283\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific.jpg\" alt=\"Prolific\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Prolific-768x577.jpg 768w\" \/><\/noscript><p>Prolific defines itself by its own name, and thus we know that it\u2019s going to be a multi-purpose theme. General and most widely used purpose for Prolific has been small businesses, corporate companies, education sector, and eCommerce users. Built to serve business websites from the ground up, Prolific is capable of bringing many different ways of promoting your business in the digital world.<\/p><p>The unique homepage sections are created to cover the must-have essentials of business oriented websites, including Team Members, Testimonials, Blog, Featured Information, Contact Us, Client Highlights, and even some Animation effects are throw into the mix! Every section are managed with widgets. You can add the section as much as you want and order the position on front page.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/prolific\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/www.demo.acmethemes.com\/prolific\/\" data-wpel-link=\"external\">Demo<\/a><h2>Unifield<\/h2><p><img class=\"aligncenter size-full wp-image-94277 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min.jpg?is-pending-load=1\" alt=\"Unifield-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94277\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min.jpg\" alt=\"Unifield-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Unifield-min-768x577.jpg 768w\" \/><\/noscript><p>Unifield helps your business to keep things simple, it\u2019s a corporate business theme for professional WordPress users. This theme has seen use across industries like health, finance, and personal projects. Unifield theme is fully compatible with the latest version of WordPress, including plugins. It possesses a beautifully and professionally organized design. The developers structured it to support the WooCommerce plugin for eCommerce stores. All you have to do is install WooCommerce, click activate and whola.. your digital store is already driving sales! You can manage its typography through the Customizer options panel, where you can choose from hundreds of free fonts as provided by Google.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/unifield\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/gracethemes.com\/demo\/unifield\/\" data-wpel-link=\"external\">Demo<\/a><h2>Esol<\/h2><p><img class=\"aligncenter size-full wp-image-94267 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min.jpg?is-pending-load=1\" alt=\"Esol-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94267\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min.jpg\" alt=\"Esol-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Esol-min-768x577.jpg 768w\" \/><\/noscript><p>Esol has a universal parallax based design. It\u2019s a multi-purpose landing page theme for professionals, business owners, and agencies that need a lot of room for discussion. Also, a lot of bloggers have used Esol to sell their own products. Furthermore, it can bridge the gap between a homepage, blog, and shopping cart pages. The developers at Asia Themes created and maintained this lovely theme. These guys have solid knowledge of trending WordPress topics, and thus integrate them into their themes with absolute ease.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/esol\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/asiathemes.asia\/preview\/esol\/\" data-wpel-link=\"external\">Demo<\/a><h2>Teletype<\/h2><p><img class=\"aligncenter size-full wp-image-94276 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min.jpg?is-pending-load=1\" alt=\"Teletype-min\" width=\"1023\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min-768x577.jpg 768w\" data-lazy-sizes=\"(max-width: 1023px) 100vw, 1023px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94276\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min.jpg\" alt=\"Teletype-min\" width=\"1023\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min.jpg 1023w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Teletype-min-768x577.jpg 768w\" \/><\/noscript><p>If you need an authentic way to express themselves on the web, then Teletype is perfect for you. Teletype is a Bootstrap based WordPress theme for bloggers, writers, and creative artists. It\u2019s still a landing page though, because the homepage is section-based and has numerous landing page elements. But as far as we can tell, Teletype would best work as a landing page theme for creatives and freelancers. You should also be keen to promote your services and\/or products using your website. Moreover, the developers included content sections for the front page template. Furthermore, you can use the Customizers to manage and sort them. Theme comes with frontpage element as custom widget Icon+Text block.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/teletype\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"http:\/\/demo.dinevthemes.com\/teletype\/\" data-wpel-link=\"external\">Demo<\/a><h2>Clarina<\/h2><p><img class=\"aligncenter size-full wp-image-94264 jetpack-lazy-image\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1.jpg?is-pending-load=1\" alt=\"Clarina-min\" width=\"1024\" height=\"768\" data-lazy-srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1-768x576.jpg 768w\" data-lazy-sizes=\"(max-width: 1024px) 100vw, 1024px\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><\/p><noscript><img class=\"aligncenter size-full wp-image-94264\" src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1.jpg\" alt=\"Clarina-min\" width=\"1024\" height=\"768\" srcset=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1.jpg 1024w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1-300x225.jpg 300w, https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/Clarina-min-1-768x576.jpg 768w\" \/><\/noscript><p>If diversity combined with professionalism is what you seek, then look no further than Clarina; free, modern, elegant, and overall beautiful WordPress theme for landing page needs. Users of Clarina are typically putting this theme to work for their mobile application, product showcase, and agency portfolio needs. The developers also equipped this theme with right widgets and features to make those sites work in any kind of scenario. Moreover, the strong back-end of Llorix One theme, guarantees that Clarina is going to perform with immaculate quality. It will perform well not only in website speed metrics, but also search engine rankings performance.<\/p> <a class=\"nectar-button medium accent-color\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/wordpress.org\/themes\/clarina\/\" data-wpel-link=\"external\">More info \/ Download<\/a> <a class=\"nectar-button medium extra-color-2\" target=\"_blank\" rel=\"nofollow external\" href=\"https:\/\/demo.themeisle.com\/clarina\/\" data-wpel-link=\"external\">Demo<\/a><p class=\"footer-notice\">Thank you for visiting Colorlib and reading this article! We highly appreciate it! Now you might want to learn how to <a href=\"https:\/\/colorlib.com\/wp\/how-to-make-a-website\/\" data-wpel-link=\"internal\">make a website<\/a> using our free <a href=\"https:\/\/colorlib.com\/wp\/themes\/\" data-wpel-link=\"internal\">WordPress themes<\/a>.<\/p><p>  <ins class=\"adsbygoogle\" data-ad-client=\"ca-pub-6200795649377457\" data-ad-slot=\"1487373328\" data-ad-format=\"auto\"> <\/ins><\/p><\/div><div class=\"post-tags clr\"><a href=\"https:\/\/colorlib.com\/wp\/tag\/landing-page\/\" rel=\"tag\" data-wpel-link=\"internal\">Landing Page<\/a><a href=\"https:\/\/colorlib.com\/wp\/tag\/marketing\/\" rel=\"tag\" data-wpel-link=\"internal\">Marketing<\/a><a href=\"https:\/\/colorlib.com\/wp\/tag\/one-page\/\" rel=\"tag\" data-wpel-link=\"internal\">One Page<\/a><a href=\"https:\/\/colorlib.com\/wp\/tag\/parallax\/\" rel=\"tag\" data-wpel-link=\"internal\">Parallax<\/a><\/div><div class=\"wpex-social-share position-horizontal style-flat\" data-source=\"https%3A%2F%2Fcolorlib.com%2Fwp%2F\" data-url=\"https%3A%2F%2Fcolorlib.com%2Fwp%2Ffree-landing-page-wordpress-themes\" data-title=\"25+ Best Free Landing Page WordPress Themes 2018\" data-image=\"https%3A%2F%2Fcolorlib.com%2Fwp%2Fwp-content%2Fuploads%2Fsites%2F2%2Ffree-landing-page-wordpress-themes.jpg\" data-twitter-handle=\"colorlib\" data-summary=\"The%20landing%20page%20is%20like%20a%20billboard%20for%20your%20product%2C%20it%27s%20capable%20of%20enticing%20or%20scaring%20your%20customers%20away%2C%20though%20in%20most%20cases%20you%20are%20looking%20to%20retain%20them.%20On\" data-specs=\"menubar=no,toolbar=no,resizable=yes,scrollbars=yes,height=600,width=600\"><div class=\"theme-heading  social-share-title\">Share This<\/div><ul class=\"clr\"><li class=\"wpex-facebook\"><a href=\"#\" role=\"button\" data-wpel-link=\"internal\">Share<\/a><\/li><li class=\"wpex-twitter\"><a href=\"#\" role=\"button\" data-wpel-link=\"internal\">Tweet<\/a><\/li><li class=\"wpex-pinterest\"><a href=\"#\" role=\"button\" data-wpel-link=\"internal\">Pin It<\/a><\/li><li class=\"wpex-linkedin\"><a href=\"#\" role=\"button\" data-wpel-link=\"internal\">Share<\/a><\/li><\/ul><\/div><section class=\"author-bio clr\"><div class=\"author-bio-avatar\"> <a href=\"https:\/\/colorlib.com\/wp\/author\/alexivanovs\/\" title=\"Visit Author Page\" data-wpel-link=\"internal\"> <img alt=\"\" src=\"https:\/\/secure.gravatar.com\/avatar\/05d0f38ea274afc31bdb7a79e29dd60b?s=74&amp;d=mm&amp;r=g\" class=\"avatar avatar-74 photo jetpack-lazy-image\" height=\"74\" width=\"74\" srcset=\"denied:data:image\/gif;base64, R0lGODlhAQABAIAAAAAAAP\/\/\/yH5BAEAAAAALAAAAAABAAEAAAIBRAA7\" \/><img alt=\"\" src=\"https:\/\/secure.gravatar.com\/avatar\/05d0f38ea274afc31bdb7a79e29dd60b?s=74&amp;d=mm&amp;r=g\" srcset=\"https:\/\/secure.gravatar.com\/avatar\/05d0f38ea274afc31bdb7a79e29dd60b?s=148&amp;d=mm&amp;r=g 2x\" class=\"avatar avatar-74 photo\" height=\"74\" width=\"74\" \/><\/a><\/div><div class=\"author-bio-content clr\"><h4 class=\"author-bio-title\"> <a href=\"https:\/\/colorlib.com\/wp\/author\/alexivanovs\/\" title=\"Visit Author Page\" data-wpel-link=\"internal\">Alex Ivanovs<\/a><\/h4><div class=\"author-bio-description clr\"><p>Alex is a freelance writer with more than 10 years of experience in design, development, and small business. His work has been featured in publications like Entrepreneur, Huffington Post, TheNextWeb, and others.<\/p><\/div><\/div><\/section><div class=\"related-posts clr\"><div class=\"theme-heading  related-posts-title\">Related Posts<\/div><div class=\"wpex-row clr\"><article class=\"related-post clr nr-col span_1_of_3 col-1 post-115073 post type-post status-publish format-standard has-post-thumbnail hentry category-free-themes tag-cafeteria tag-coffee-shop tag-free-themes tag-restaurant tag-wordpress-themes entry has-media\"><figure class=\"related-post-figure clr\"><a href=\"https:\/\/colorlib.com\/wp\/free-restaurant-wordpress-theme\/\" title=\"20+ Free and Stunning Restaurant WordPress Themes\" rel=\"bookmark\" class=\"related-post-thumb\" data-wpel-link=\"internal\"> <img src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/free-restaurant-wordpress-themes-506x409.jpg\" data-no-retina=\"\" width=\"506\" height=\"409\" alt=\"image\" \/><\/a><\/figure><div class=\"related-post-content clr\"><h4 class=\"related-post-title entry-title\"> <a href=\"https:\/\/colorlib.com\/wp\/free-restaurant-wordpress-theme\/\" rel=\"bookmark\" data-wpel-link=\"internal\">20+ Free and Stunning Restaurant WordPress Themes<\/a><\/h4><div class=\"related-post-excerpt clr\"><p>Being in the restaurant niche business is a complicated thing. This is mainly because there\u2026<\/p><\/div><\/div><\/article><article class=\"related-post clr nr-col span_1_of_3 col-2 post-13810 post type-post status-publish format-standard has-post-thumbnail hentry category-free-themes tag-business tag-corporate tag-free-themes tag-full-screen tag-portfolio tag-restaurant tag-templatemonster entry has-media\"><figure class=\"related-post-figure clr\"><a href=\"https:\/\/colorlib.com\/wp\/top-20-wordpress-themes-for-free\/\" title=\"Premium Themes That Don\u2019t Cost You a Cent: Top 20 WordPress Themes For Free\" rel=\"bookmark\" class=\"related-post-thumb\" data-wpel-link=\"internal\"> <img src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/top-free-wordpress-themes-506x288.jpg\" alt=\"Free WordPress Themes\" data-no-retina=\"\" width=\"506\" height=\"288\" \/><\/a><\/figure><div class=\"related-post-content clr\"><h4 class=\"related-post-title entry-title\"> <a href=\"https:\/\/colorlib.com\/wp\/top-20-wordpress-themes-for-free\/\" rel=\"bookmark\" data-wpel-link=\"internal\">Premium Themes That Don\u2019t Cost You a Cent: Top 20 WordPress Themes For Free<\/a><\/h4><div class=\"related-post-excerpt clr\"><p>For a passion-driven professional or rapidly growing enterprise, having a webpage online is the best\u2026<\/p><\/div><\/div><\/article><article class=\"related-post clr nr-col span_1_of_3 col-3 post-116484 post type-post status-publish format-standard has-post-thumbnail hentry category-free-themes tag-charity tag-church tag-church-themes tag-free-themes tag-non-profit tag-wordpress-themes entry has-media\"><figure class=\"related-post-figure clr\"><a href=\"https:\/\/colorlib.com\/wp\/free-charity-wordpress-themes\/\" title=\"15 Best Free Charity WordPress Themes 2018\" rel=\"bookmark\" class=\"related-post-thumb\" data-wpel-link=\"internal\"> <img src=\"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/free-charity-wordpress-themes-506x400.jpg\" data-no-retina=\"\" width=\"506\" height=\"400\" alt=\"image\" \/><\/a><\/figure><div class=\"related-post-content clr\"><h4 class=\"related-post-title entry-title\"> <a href=\"https:\/\/colorlib.com\/wp\/free-charity-wordpress-themes\/\" rel=\"bookmark\" data-wpel-link=\"internal\">15 Best Free Charity WordPress Themes 2018<\/a><\/h4><div class=\"related-post-excerpt clr\"><p>It is common knowledge how hard it was set up a beautiful website for your\u2026<\/p><\/div><\/div><\/article><\/div><\/div>","created_at":"2018-11-07T16:21:56+0000","updated_at":"2018-11-07T16:21:56+0000","published_at":"2018-10-30T06:48:42+0000","published_by":["Alex Ivanovs"],"annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":19,"domain_name":"colorlib.com","preview_picture":"https:\/\/cdn.colorlib.com\/wp\/wp-content\/uploads\/sites\/2\/free-landing-page-wordpress-themes.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":25,"title":"Networking for Nerds","url":"http:\/\/benjaminreinhardt.com\/networking-for-nerds\/","content":"<p>\u201cNetworking.\u201d What a word. When I left grad school in 2015, networking was for slick-haired salesmen, former jocks, and social parasites in general. Real heroes built better mousetraps and the world beat a path to their doors. But then I looked at what real successes actually do - they lead a balancing act between connecting and building. Though Edison, Einstein, Musk (and anybody you may have heard of) have done real work, their projects would have been stillborn or failed if they didn\u2019t know the right people: partners, employees, backers, supporters.<\/p><p>Always the slave to the scientific method, I proceeded with a \u201cnetworking might be useful\u201d as a hypothesis and did many experiments. Some were successful, some less so. In the process I developed a set of tactics that unleashed many opportunities - a stint as an EIR at Susa Ventures, speaking at conferences in Brazil and SXSW, and getting a company off the ground. (I like to think that) these were all built on a door of real value, but tactical network building provided the key.<\/p><p>Here are just a few of the specific tactics I\u2019ve developed. At the end of the day, all of them boil down to one thing - make it easy for the other person - to remember you, to help, and to meet.<\/p><h3 id=\"tactic-1-the-projection-model\">Tactic 1: The Projection Model<\/h3><p>How do some people seem to wind up with advocates everywhere? They create a little mental model of themselves that makes it easy for everybody they meet.<\/p><p>Everybody has a \u201cprojection\u201d - a small representation - of each person they know living in their heads. You know that friend who\u2019s \u201cthe badass interface designer who is looking to get into machine learning and consults?\u201d She has an amazing projection and makes it easy for opportunities to find her. A projection can be distinct - \u201cBen is looking for X, Y, Z and can help with A, B, C.\u201d Or it can be fuzzy - \u201cBen\u2019s that guy with curly hair I think who I met at that meetup and we talked about a bunch of interesting stuff.\u201d<\/p><p>If it\u2019s distinct, whenever X comes up in conversation, you will instantly come to mind, and you\u2019ll start getting messages like \u201cwe were having a discussion about X and would love to bring you on as an advisor.\u201d<\/p><p>In my experience, there\u2019s a direct correlation between how well I create a distinct projection and people\u2019s ability to help. When I was \u201cthe AR\/VR guy who\u2019s interested in giving talks\u201d it led to a new friend introing me to an organization in Brazil that was doing a course on future tech, cascading to lecturing in Sao Paulo, making new friends, and opening up several collaborations.<\/p><p>The trick is that projections are small and can only hold a few pieces of information (rule of thumb - about three.)  In order to maximize the effectiveness of your projections, figure out which things you want to be associated with in people\u2019s heads and be excited about them. Nothing is more memorable than distinct excitement.  Similarly, try to figure out the things that drive the whomever you\u2019re talking to and incorporate them into your mental representation of that person.  The best way to encourage people to keep you in mind is to keep them in mind. (NOTE:  Many people take this to the extreme and straight up ask the question: \u201chow can I help?\u201d This is exactly the sentiment you should have, but it makes your counterpart feel like you weren\u2019t actually listening )<\/p><h4 id=\"tell-a-story\">Tell a Story:<\/h4><p>How do you build a strong projection? A strong story. Stories stick in people\u2019s heads better than anything else. Make sure you have a good one. All you need is a short narrative with at least one compelling character (you), a captivating theme, and some kind of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Story_arc\">narrative arc<\/a>.  Think about your life as a series of towns connected by a road, where the towns are the relevant parts of your background connected by your theme. Frankly, most people won\u2019t remember the towns at all. They\u2019ll just viscerally remember whether the theme was compellingly supported.<\/p><p>Telling a story doesn\u2019t mean you should lie or exaggerate - people can tell. The piece that can be uncomfortable for us nerds is making meaning out of what are empirically disconnected events.<\/p><h3 id=\"tactic-2-know-your-ask\">Tactic 2: Know your ask<\/h3><h4 id=\"make-it-easy-to-help\">Make it Easy to Help<\/h4><p>At the end of the day, everybody has so much going on in their own heads that it\u2019s a miracle if you can exist there at all. How do you maximize the chance that you\u2019ll get a precious slice of brain space? Use stories and have clear asks.<\/p><h4 id=\"honing-your-ask\">Honing your Ask<\/h4><p>Reality check: people are going to remember at most three things about you. Make sure one of them is your <strong>ask<\/strong>- the most impactful thing they can do to help you. To maximize your ask, make it short, clear, and a single step. Otherwise, it will be harder to remember and execute. Making asks is often uncomfortable for smart successful people - you want to get by on your own and be so awesome that people will figure out how to help you on your own. This is not going to happen. On the other hand, if you have a clear actionable ask, people are eager to jump on it.<\/p><p>Years ago, whenever I would travel somewhere, I would passively post \u201cI\u2019m going to be in town, who wants to hang out?\u201d Maybe one or two people would respond if I was lucky. Now I actively ask friends who have lived there \u201cwho are the best people you know there? Can you put me in touch with them?\u201d The result? Friendships and adventures.<\/p><p>Some good examples of asks:<\/p><ul><li>\n    <p>\u201cI\u2019m looking to get in touch with people who are director level or above in the elder care industry\u201d<\/p>\n  <\/li>\n  <li>\n    <p>\u201cI need a new apartment in San Francisco under $2200, preferably in the mission.\u201d<\/p>\n  <\/li>\n  <li>\n    <p>\u201cI\u2019m looking for customer-facing jobs in robotics that take advantage of my unique combination of technical skills and people skills\u201d<\/p>\n  <\/li>\n<\/ul><p>Why are they good?<\/p><p><strong>Single Action to Execute -<\/strong> connect with people, or a link on the internet.<\/p><p><strong>Detailed -<\/strong> they don\u2019t force you to ask clarifying questions.<\/p><h4 id=\"when-to-ask\">When to ask<\/h4><p>Timing on asks is subtle. In the limit, you obviously don\u2019t want to walk up to a stranger and say \u201chey, I\u2019m looking for a job.\u201d Nor do you want to hold off on telling someone what you need until they\u2019ve invited you to their wedding. Instead of thinking about time, think about interest. Once someone is interested in you, they\u2019re likely to remember how they can help, regardless of how long they\u2019ve known you. I\u2019ve gotten to that point within a minute - when I mentioned I was doing research on elder care his eyes lit up and told me \u201cOh! I have a sister who\u2019s worked in that field for twenty years.\u201d (interest)  \u201cGreat! Would she be interested in chatting?\u201d (ask.)<\/p><h3 id=\"tactic-3-pre-meeting-motions\">Tactic 3: Pre-Meeting Motions<\/h3><p>Since relationships are built through 1-1 interactions, you can\u2019t build meaningful relationships without 1-1 meetings. It took me a long time to realize that there are tons of things you can do to make a meeting great before it even starts.<\/p><h4 id=\"scheduling\">Scheduling<\/h4><p>Let\u2019s assume you don\u2019t have a personal assistant to schedule your meetings.  Let\u2019s also assume you don\u2019t have enough leverage that you can get away with texting someone \u201cI\u2019m hanging out at Ritual Coffee for the next hour. Come find me.\u201d Here are some tactical do\u2019s and do not\u2019s that make scheduling simple.<\/p><p><strong>Do:<\/strong> Say what you want to talk about during the meeting in the email. It makes it easy for them to say yes and it gives you both a little reminder when you go back and check the email right before the meeting.<\/p><p><strong>Don\u2019t:<\/strong> Use the term \u201cI\u2019d like to pick your brain about \u2026\u201d It screams \u201cI\u2019m not willing to do the legwork to come up with specific questions and topics.\u201d<\/p><p><strong>Do:<\/strong> Give as much information as possible - <em>\u201cI\u2019m based in the Mission and I\u2019m in SOMA on Thursday afternoons.\u201d<\/em><\/p><p><strong>Don\u2019t:<\/strong> Be vague or pretend to be busy when you\u2019re not. One of my favorite lines to avoid vagueness is: \u201cI\u2019m flexible on time and place but like to offer a concrete suggestion\u2026\u201d<\/p><p><img src=\"http:\/\/benjaminreinhardt.com\/pictures\/posts\/2017-12-26-networking-for-nerds\/bad-example.png\" alt=\"image\" \/><\/p><p><strong>Do:<\/strong> Anchor the conversation with 2-3 options around both time and place. Windows of time are totally fine. *\u201cWould you like to meet over coffee at Stable Cafe between 2-4pm on Thursday or Friday?\u201d *<\/p><p><strong>Don\u2019t:<\/strong> Ask \u201cwhen is best for you?\u201d or say \u201cwe should grab coffee!\u201d without an anchor time. It seems nice because you give them all the choice, but you\u2019re also putting all the work on their desk and gives them no information about your preferences.\n<img src=\"http:\/\/benjaminreinhardt.com\/pictures\/posts\/2017-12-26-networking-for-nerds\/anchoring.png\" alt=\"image\" \/><\/p><p><strong>Calendar Invite -<\/strong> It might seem obvious, but many people still screw up the concluding act leading up to a meeting - a calendar event. Busy folks live by their calendars, so make it easy for them to remember the meeting - where\/when it\u2019s happening, who it\u2019s with, and possibly what it\u2019s about. Here\u2019s a little sample from my calendar:<\/p><p><img src=\"http:\/\/benjaminreinhardt.com\/pictures\/posts\/2017-12-26-networking-for-nerds\/calendar-invite.png\" alt=\"image\" \/><\/p><p>It could be even better if I had titled it \u201cBen&lt;&gt;Jim discuss AI Strategies\u201d or put the subject in the notes section.<\/p><p>Do your homework before a meeting. Stalk your counterpart on the internet a bit. Read their twitter\/blog, skim their LinkedIn, Github, what have you. The whole goal is to understand a bit more about who they are, what they\u2019ve done, and what excites them. It\u2019s also just flattering when someone pulls out a relevant reference to your web page in a conversation.<\/p><h3 id=\"make-it-easy\">Make it easy<\/h3><p>At the end of the day, all of this boils down to ways to make it easy - for people to keep you in mind, to help you, and to engage with you. If you keep that heuristic in mind, it will be a snap to internalize these tactics and come up with ones of your own.<\/p><p>If you enjoyed this post, please share it with your friends and if you have any questions, please get in touch.<\/p>","created_at":"2018-11-07T16:21:54+0000","updated_at":"2018-11-07T16:21:54+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":8,"domain_name":"benjaminreinhardt.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":26,"title":"Howdy, Cow Pokes: Original Star Wars Trilogy Imagined As Spaghetti Western Movie Posters (I Love Spaghetti FYI)","url":"https:\/\/geekologie.com\/2012\/08\/howdy-cow-poke-original-star-wars-trilog.php","content":"<p><img alt=\"star-wars-western-1.jpg\" src=\"http:\/\/www.geekologie.com\/2012\/08\/20\/star-wars-western-1.jpg\" width=\"640\" height=\"960\" class=\"mt-image-none\" \/><\/p><p>This is a series of posters by artist <a href=\"http:\/\/timothyandersonart-printstore.blogspot.com\/\">Tim Anderson<\/a> (links to personal website where you can buy prints, $40) imagining<a href=\"http:\/\/blastr.com\/2012\/08\/3-awesome-star-wars-poste.php\"> the first three Star Wars movies as spaghetti westerns<\/a> by Italian director Sergio Leone.  Would Leone have had enough decency to not make any prequels?   Probably.  *spins gun on finger, accidentally shoots Jar Jar in the face*  Woopsie doopsie!  \"That wasn't an accident, was it?\" <em> Is a tauntaun penis ever warm to the touch?<\/em><\/p><p>Hit the jump for the other two.<\/p><p><img alt=\"star-wars-western-2.jpg\" src=\"http:\/\/www.geekologie.com\/2012\/08\/20\/star-wars-western-2.jpg\" width=\"640\" height=\"960\" class=\"mt-image-none\" \/><\/p><p><img alt=\"star-wars-western-3.jpg\" src=\"http:\/\/www.geekologie.com\/2012\/08\/20\/star-wars-western-3.jpg\" width=\"640\" height=\"960\" class=\"mt-image-none\" \/><\/p><p>Thanks to XtraFeely, who has like, more nerve endings than a regular person or something.<\/p><p><a href=\"https:\/\/facebook.com\/sharer\/sharer.php?text=Howdy, Cow Pokes: Original Star Wars Trilogy Imagined As Spaghetti Western Movie Posters (I Love Spaghetti FYI)&amp;u=http:\/\/geekologie.com\/2012\/08\/howdy-cow-poke-original-star-wars-trilog.php\" class=\"fb-circle\">Facebook<\/a>\n        \n        <a href=\"http:\/\/www.pinterest.com\/pin\/create\/button\/?description=Howdy, Cow Pokes: Original Star Wars Trilogy Imagined As Spaghetti Western Movie Posters (I Love Spaghetti FYI)&amp;media=http:\/\/geekologie.com\/2012\/08\/20\/star-wars-western-1.jpg&amp;url=http:\/\/geekologie.com\/2012\/08\/howdy-cow-poke-original-star-wars-trilog.php\" class=\"pinterest-circle\">Pinterest<\/a><\/p><p>Read More: , , , , , , , , , <\/p>","created_at":"2018-11-07T16:21:54+0000","updated_at":"2018-11-07T16:21:54+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"geekologie.com","preview_picture":"http:\/\/geekologie.com\/2012\/08\/20\/star-wars-western-3.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":23,"title":"How to Get Things Done When You Don't Feel Like It","url":"https:\/\/queue.acm.org\/detail.cfm?id=3280677","content":"<p>\n<a href=\"https:\/\/queue.acm.org\/app\/\" target=\"_new\"><img src=\"https:\/\/queue.acm.org\/app\/2018_09-10_lrg.png\" height=\"120\" alt=\"September\/October 2018 issue of acmqueue\" \/><\/a>\n<b><a href=\"https:\/\/queue.acm.org\/app\/\">The September\/October issue of acmqueue is out now<\/a><\/b>\n<\/p><p><a href=\"https:\/\/cdn.coverstand.com\/3rd_pty\/acm\/login.html?&amp;btx_i=537163\">Subscribers and ACM Professional members login here<\/a>\n<br style=\"clear: both;\" \/><\/p><hr \/><br \/><h3><a href=\"https:\/\/queue.acm.org\/listing.cfm?typefilter=Thesoftsideofsoftware&amp;sort=publication_date&amp;order=desc&amp;qc_type=Thesoftsideofsoftware&amp;article_type=&amp;item_topic=all&amp;filter_type=topic&amp;page_title=The%20Soft%20Side%20of%20Software&amp;filter=all\">The Soft Side of Software<\/a><\/h3><h4><a href=\"https:\/\/queue.acm.org\/listing.cfm?item_topic=Business\/Management&amp;qc_type=theme_list&amp;filter=Business\/Management&amp;page_title=Business\/Management&amp;order=desc\">Business\/Management<\/a><\/h4><p>\n\u00a0\n<a href=\"https:\/\/dl.acm.org\/ft_gateway.cfm?id=3280677&amp;ftid=2004312&amp;dwn=1\">\n<img src=\"https:\/\/queue.acm.org\/img\/icon_pdf.png\" alt=\"Download PDF version of this article\" \/>\nPDF\n<\/a>\n<\/p><label>September 18, 2018<br \/><b><a class=\"descriptor\" href=\"https:\/\/queue.acm.org\/issuedetail.cfm?issue=3277539\">Volume 16, issue 4 <\/a><\/b><\/label><h3>Kate Matsudaira<\/h3><p>Have you ever come into work, sat down at your computer to begin a project, opened your editor, and then just stared at the screen? This happens to me all the time, so I understand your struggle.<\/p><p>Even if you love your job, you don't always feel like doing it every day. There are so many factors that influence your ability to show up to work with enthusiasm and then work hard all day long. <\/p><p>External events can take priority in your mind\u2014family struggles, a breakup, a sick pet\u2014and make it hard to focus. Then, of course, there are the struggles at work that can make it hard to feel motivated. Getting a bad review can knock you off course. Likewise, if you work really hard on a project and your manager doesn't seem to value it at all, you might wonder why you are working so hard. <\/p><p>Other times you have to work on tasks you don't enjoy (for me that is writing lots of tests, or documentation) or projects that aren't challenging. If your work is uninteresting or if a task you've been assigned seems beneath your ability, finding your motivation can be hard.<\/p><p>So, what do you do? Many people turn to procrastination or ignoring the task\u2014but that only postpones the inevitable. You can try to talk your way out of the assignment, and maybe your manager will support you, but at some point the work needs to get done. <\/p><p>If you want to be successful, then it serves you better to rise to the occasion no matter what. That means learning how to push through challenges and deliver valuable results.<\/p><p>Since this happens to me quite often, I have captured five of my best strategies for turning out amazing work even when I don't feel like it.<\/p><h4>Gamify your process<\/h4><p>Dealing with a really big project used to hold me back. If the project had lots of tasks I didn't know how to do or that seemed really difficult, I resisted even starting because I was so overwhelmed by the scope.<\/p><p>Of course, this meant I procrastinated until only the minimum amount of time remained to complete the project. Then I would end up working crazy long hours, and sometimes I ended up with code that \"worked\" but was in no way ready for prime time (e.g., a few bugs, not enough coverage of edge cases, minimal testing, working only in my dev environment because I couldn't make it work on staging, etc.). This was super stressful and usually meant my work wasn't as good as it could have been if I had only started earlier.<\/p><p>This was one of the biggest obstacles early in my career: I had a hard time getting started.<\/p><p>I discovered that if I made the process of getting started easier, those first few steps on a daunting project became more tenable. Once I took a few steps, it was so much easier to keep going.<\/p><p>My solution was to approach a project by turning it into as many tiny steps as possible. That way I could get a few really easy wins under my belt. For example, each step would be a task such as \"Search for ______ on Google\" or \"Have a conversation with ______.\" Crossing things off your to-do list gives your brain a happy little dopamine hit, even if the tasks are tiny\u2014it keeps your motivation up and your excuses down.<\/p><p>Try breaking your next project into the smallest increments you possibly can. Each step should be really small (I try for tasks that take 15 minutes or less) and really easy to accomplish, so that you can get a win!<\/p><p>You have to overcome inertia. Little wins add up and make it easier to do that.<\/p><h4>Reserve calendar time for every project<\/h4><p>Set aside time on your calendar specifically for working on a task you're having trouble starting. Treat it as seriously as you would any other appointment. You must show up and you must work on that project.<\/p><p>Reserve an amount of time that is realistic for making progress\u2014at least 30 minutes to an hour. This strategy is key for busy people or managers. If you don't schedule the time to do meaningful strategic work, your time will fill up with tactical tasks. <\/p><p>And what if you don't feel like working on the task at the appointed time? Set a timer when you're starting work. Set it for 10 minutes and tell yourself you have to work only until the timer goes off.<\/p><p>Start working on the list of tiny steps you have created for yourself: google something; set up your project; send one email; review one document.<\/p><p>Almost always, taking one or two of these tiny steps will get your brain working, and it will be easier to keep going. You'll do one task, cross it off the list, and then do another. Your timer for 10 minutes will go off, and you'll just keep going because now you're engaged with the project.<\/p><p>If you're really not engaged with it after 10 minutes (though this rarely happens to me), then let yourself take a break. But block off another chunk of time on your calendar to come back to it soon. <\/p><h4>Get other people involved<\/h4><p>Sometimes the best way to get yourself to do something is to make yourself accountable to another person.<\/p><p>According to a study by the American Society of Training and Development,<sup>1<\/sup> people who commit to someone else have a 65 percent chance of accomplishing the goals they set. That number goes up to 95 percent if you commit to a specific accountability appointment with that person.<\/p><p>Our brains are wired not to want to let down other people. If someone invests in you by agreeing to help you accomplish your goal, you are driven to do your part by a desire to live up to that commitment.<\/p><p>There are a few ways to do this:<\/p><p>\u2022 Set deadlines with your manager for when certain aspects of the project will be complete, and schedule regular check-ins on status.<\/p><p>\u2022 Ask for help on a part of the project. With the help of another person to reduce your workload, you can get other parts of the project done. Set a time to meet with your helper to combine your results.<\/p><p>\u2022 Make a recurring date with a peer to work together. For example, if you are both tasked with running a series of tedious tests that you both would rather put off, set a time to sit together and get them done.<\/p><p>\u2022 Embrace the scrum part of Agile and have daily standups with your teammates.<\/p><p>Delegating work can be especially helpful when you have a really big project in front of you. Sometimes the scope of a project is so overwhelming that it is hard to get started; if you can solicit help from your team to tackle some of the project, then you can focus your efforts on a more manageable workload.<\/p><h4>Talk about it<\/h4><p>Externalizing problems can make them a lot easier to deal with. Things tend to get blown out of proportion in our minds, especially when we are stressed about them.<\/p><p>I can't tell you how many times I've started talking to someone about how stressed I am about a project\u2014like I don't have any ideas for an article, or it's so hard I have no clue how I will solve it\u2014that by the time I'm done talking, I actually come away full of inspiration. Other times, I am just so stressed about what could go wrong (or what is going wrong) that I fast become overwhelmed.<\/p><p>Scientific studies have shown that talking about feelings out loud actually decreases stress and the bad feelings we're experiencing. Brain imaging done at UCLA<sup>2<\/sup> demonstrated that when a person was shown a picture of an angry face, the amygdala became more active. This is the part of the brain responsible for activating the body's \"alarm\" system\u2014it lets you know that you have something to fear and kicks your body into action to deal with that threat.<\/p><p>When the study participants were able to name what they saw, however, the simple act of putting the feeling they saw into words caused the amygdala activity to decrease. Not only that, but each participant's right ventrolateral prefrontal cortex then became activated. Other studies have indicated that this is the area of the brain associated with processing emotion and putting words to emotional experiences.<\/p><p>So, talking about your big project might be just the thing to help you get started.<\/p><p>Plus, if you talk to smart friends or mentors, they might have suggestions for how you can start or experiences to share about how they did something similar. You can become more relaxed and smarter at the same time.<\/p><h4>Practice the art of \"precrastination\"<\/h4><p>Do you ever have trouble working from home because you get distracted by unwashed dishes in the sink or laundry that needs to be folded? You have probably been told that you are a procrastinator, but, in fact, you might be just the opposite.<\/p><p>I used to be a master procrastinator. I would find any excuse to keep from starting work, or even thinking about it. As I learned again and again, procrastination is a bad thing. It comes from a fear of getting started, so you actively keep yourself from making progress by doing things that keep your mind off of what you have to do.<\/p><p>But there is something called \"precrastination,\" and it is actually really good for you.<\/p><p>As you're working on a project, your brain needs to take breaks\u2014not just to recharge, but also to form new connections and create new ideas. That's why getting up to wash the dishes, fold the laundry, take a shower, take a walk, or any other low-key activity that allows you to let your mind wander for a while can be really good for your productivity overall.<\/p><p>When you do something that feels satisfying, your brain releases dopamine (just as it does when you cross an item off a to-do list\u2014because it feels good!). So, when you take a walk midway through your work session, your brain gets a hit of dopamine.<\/p><p>That dopamine triggers the parts of your brain associated with creativity and gets them working. That's when those magical aha! moments happen, because your brain is sending energy to the areas that help you make connections and see things in new ways.<\/p><p>Next time you are stuck on a project you don't want to start, try doing something that you know will be satisfying. You just might have a bright idea while you're rinsing off your dishes, and that will make you excited to run over to your computer and get to work.<\/p><h4>References<\/h4><p>1. Oppong, T. 2017. The accountability effect: a simple way to achieve your goals and boost your performance. The Mission (January 16); <a href=\"https:\/\/medium.com\/the-mission\/the-accountability-effect-a-simple-way-to-achieve-your-goals-and-boost-your-performance-8a07c76ef53a\">https:\/\/medium.com\/the-mission\/the-accountability-effect-a-simple-way-to-achieve-your-goals-and-boost-your-performance-8a07c76ef53a<\/a>.<\/p><p>2. Wolpert, S. 2007. Putting feelings into words produces therapeutic effects in the brain; UCLA neuroimaging study supports ancient Buddhist teachings. UCLA Newsroom (June 21); <a href=\"http:\/\/newsroom.ucla.edu\/releases\/Putting-Feelings-Into-Words-Produces-8047\">http:\/\/newsroom.ucla.edu\/releases\/Putting-Feelings-Into-Words-Produces-8047<\/a>.<\/p><h4>Related articles<\/h4><p> <b><a href=\"https:\/\/queue.acm.org\/detail.cfm?id=1122687\">People and Process<\/a><\/b> <br \/>Minimizing the pain of business process change <br \/>James Champy <br \/><a href=\"https:\/\/queue.acm.org\/detail.cfm?id=1122687\">https:\/\/queue.acm.org\/detail.cfm?id=1122687<\/a> <\/p><p> <b><a href=\"https:\/\/queue.acm.org\/detail.cfm?id=2996549\">Fresh Starts<\/a><\/b> <br \/>Just because you have been doing it the same way doesn't mean you are doing it the right way. <br \/>Kate Matsudaira <br \/><a href=\"https:\/\/queue.acm.org\/detail.cfm?id=2996549\">https:\/\/queue.acm.org\/detail.cfm?id=2996549<\/a> <\/p><p> <b><a href=\"https:\/\/queue.acm.org\/detail.cfm?id=966731\">IM, Not IP (Information Pollution)<\/a><\/b> <br \/>A steady dose of realtime interruptions is toxic to anyone's health. <br \/>Jakob Nielsen <br \/><a href=\"https:\/\/queue.acm.org\/detail.cfm?id=966731\">https:\/\/queue.acm.org\/detail.cfm?id=966731<\/a> <\/p><p><b>Kate Matsudaira<\/b> is an experienced technology leader. She worked in big companies such as Microsoft and Amazon and three successful startups (Decide acquired by eBay, Moz, and Delve Networks acquired by Limelight) before starting her own company, Popforms (https:\/\/popforms.com\/), which was acquired by Safari Books. Having spent her early career as a software engineer, she is deeply technical and has done leading work on distributed systems, cloud computing, and mobile. She has experience managing entire product teams and research scientists and has built her own profitable business. She is a published author, keynote speaker, and has been honored with awards such as Seattle's Top 40 under 40. She sits on the board of\u00a0<i>acmqueue<\/i>\u00a0and maintains a personal blog at katemats.com.<\/p><p>Copyright \u00a9 2018 held by owner\/author. Publication rights licensed to ACM.<\/p><p>\n<img class=\"floatLeft\" src=\"https:\/\/queue.acm.org\/img\/q%20stamp_small.jpg\" width=\"26\" height=\"45\" alt=\"acmqueue\" \/><\/p><p><em>Originally published in Queue vol. 16, no. 4<\/em>\u2014\n<br \/>see this item in the <a href=\"https:\/\/portal.acm.org\/citation.cfm?id=3280677\">ACM Digital Library<\/a>\n<\/p><br \/><hr style=\"border-style: none; border: 0; background-color: gray; color: gray; size: 1px;\" \/>Related:<p>\nBridget Kromhout - <a href=\"https:\/\/queue.acm.org\/detail.cfm?id=3185224\"><b>Containers Will Not Fix Your Broken Culture (and Other Hard Truths)<\/b><\/a>\n<br \/>Complex socio-technical systems are hard; film at 11.\n<\/p><br \/><hr style=\"border-style: none; border: 0; background-color: gray; color: gray; size: 1px;\" \/><h3>Comments<\/h3><p>(newest first)<\/p>You explained many things and help with many things. Here is what my problem looks like. I have things to do, and I know how to do them. Doing these tasks give me no dopamin, its like repeating myself. I always end up having meny good stuff done but the besic tasks still waiting. Such as, I do a realy cool website with many new and useful functionality. But at the end I get to do the basic CRUD stuff. I take a month for project but than I procrastinate on CRUD stuff for 2 months<hr style=\"size: 1px; border-style: none; border: 0; background-color: gray; color: gray; width: 160px;\" \/>Very good, insightful, and helpful ideas. Thanks for posting.<hr style=\"size: 1px; border-style: none; border: 0; background-color: gray; color: gray; width: 160px;\" \/>Breaking your project up into a thousand tiny realistic goals is one of the best ideas ever! We sabotage ourselves when we try to conceptualise the project all as one big mountain when really one step at a time is how we actually climb it!<hr style=\"size: 1px; border-style: none; border: 0; background-color: gray; color: gray; width: 160px;\" \/>Very practical, pragmatic and meaningful. Certainly rings a bell ... and it amazes me that the solution can be this simple yet, I feel it is true.<hr style=\"size: 1px; border-style: none; border: 0; background-color: gray; color: gray; width: 160px;\" \/><form id=\"edit_form\" action=\"postcomment.cfm\" method=\"post\"><div>\nLeave this field empty <\/div><\/form><form action=\"postcomment.cfm\" method=\"post\">\n<p>Post a Comment:<\/p>\n<div><br \/><label for=\"body_text\">Comment: (Required - 4,000 character limit - HTML syntax is <strong>not<\/strong> allowed and <b>will be removed<\/b>)<\/label>\n<br \/><textarea name=\"body_text\" rows=\"8\" id=\"body_text\" cols=\"80\"><\/textarea><br \/><\/div><\/form><hr style=\"border-style: none; border: 0; background-color: gray; color: gray; size: 1px;\" \/><p>\n<a href=\"#\"><img src=\"https:\/\/queue.acm.org\/img\/logo_acm.gif\" alt=\"image\" \/><\/a>\n<br \/>\u00a9 2018 ACM, Inc. All Rights Reserved.\n<\/p>","created_at":"2018-11-07T16:21:53+0000","updated_at":"2018-11-07T16:21:53+0000","annotations":[],"mimetype":"text\/html","reading_time":11,"domain_name":"queue.acm.org","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":24,"title":"Inter UI font family","url":"https:\/\/rsms.me\/inter\/\/","content":"<p>\n  Inter UI started out in late 2016 as an experiment to build a perfectly\n  pixel-fitting font at a specific small size (11px.) The idea was that\n  by crafting a font in a particular way, with a particular coordinate system\n  (Units Per EM), and for a particular target rasterization size (11), it would\n  be possible to get the best of both sharpness and readability.\n  <\/p><p>\n  However after a few months of using an early version of Inter UI, it dawned\n  on everyone exposed to the test that this approach had some serious real-world\n  problems. Most notably that it was really hard to read longer text. Because of\n  the pixel-aligning nature of that approach, the font took an almost\n  <a href=\"https:\/\/www.figma.com\/file\/HPqDViSCB8fAWuxaV2ousFMv\">mono-spaced appearance<\/a>,\n  making it really easy to read numbers, punctuation and very short\n  words, but eye-straining to read anything longer.\n  <\/p><p>\n  The project was rebooted with a different approach, sticking with the\n  specific UPM, but crafting glyphs and kerning in a way that made for\n  more variation in the rhythm and smoother vertical and horizontal stems.\n  As Inter UI was being developed, it was tested on an internal version of\n  <a href=\"https:\/\/www.figma.com\/\">Figma<\/a>\u2014where the author of Inter UI works as a designer\u2014and slowly improved upon based on experience and feedback.\n  <\/p><p>\n  Inter UI works great for Latin text and pretty well for \n  Cyrillic. There's still some work to be done on Cyrillic and\n  <a href=\"https:\/\/github.com\/rsms\/inter\/blob\/master\/CONTRIBUTING.md\">contributions are warmly welcomed<\/a>. The playground contains <a href=\"https:\/\/rsms.me\/inter\/lab\/?sample=Body%20text%201&amp;size=16\">a lot of samples<\/a>, including some common <a href=\"https:\/\/rsms.me\/inter\/lab\/?sample=Kerning%20body%20multi-lang&amp;size=16\">non English-language words in the playground.<\/a>\n  <\/p><p>\n  Please refer to the <a href=\"https:\/\/rsms.me\/inter\/glyphs\/\">glyph repertoire<\/a>\n  for an overview of currently-available glyphs and their quality.\n  <\/p><ul class=\"faq\"><li class=\"q\" id=\"faq-using-features\">\n      How do I enable and disable font features?\n    <\/li>\n    <li class=\"a\">\n      In web browsers you'll want to use\n      <a href=\"https:\/\/developer.mozilla.org\/en-US\/docs\/Web\/CSS\/font-feature-settings\"><q>font-feature-settings<\/q><\/a>.\n      In Figma you can access features via the\n      <a href=\"https:\/\/help.figma.com\/text\/opentype-features\">Advanced Typography panel.<\/a>\n      In Illustrator, Photoshop and friends, you can access features via the\n      <a href=\"https:\/\/helpx.adobe.com\/illustrator\/using\/special-characters.html#opentype_panel_overview\">Characters and OpenType panels.<\/a>\n      Sketch doesn't provide a UI for configuring font features, but there's\n      <a href=\"https:\/\/sketchtalk.io\/discussion\/comment\/1478\/#Comment_1478\">a workaround using macOS's native font UI.<\/a>\n    <\/li>\n    <li class=\"q\" id=\"faq-unhinted-vs-hinted\">\n      What is the difference between \"unhinted\" and \"hinted\" font files?\n    <\/li>\n    <li class=\"a\">\n      The font files in the \"hinted\" folders have additional data in them\n      for assisting\n      <a href=\"https:\/\/en.wikipedia.org\/wiki\/ClearType\">ClearType<\/a>,\n      the text rasterizer used by Microsoft Windows (and some Linux distributions.)\n      You want to use the \"hinted\" fonts only if you are targeting Windows users\n      <em>and<\/em> prefer the different look of these \"hinted\" fonts.\n      Additionally, hinting data makes the font files larger, so if you are\n      using Inter UI on websites, the extra size of these files is another\n      consideration to make.\n      <a href=\"https:\/\/www.typotheque.com\/articles\/hinting\">\n      This article explains hinting at a greater length.<\/a>\n    <\/li>\n    <li class=\"q\" id=\"faq-cdn\">\n      How reliable are the fonts served from rsms.me\/inter? Is it on a CDN?\n    <\/li>\n    <li class=\"a\">\n      rsms.me\/inter is backed by GitHub's server network and distributed\n      globally on the CloudFlare CDN, making usage of\n      <q>https:\/\/rsms.me\/inter\/inter-ui.css<\/q> and associated font\n      files very reliable and fast throughout the world. \n    <\/li>\n    <li class=\"q\" id=\"faq-contribute\">\n      Can I help with improving Inter UI?\n    <\/li>\n    <li class=\"a\">\n      Yes you can! Inter UI is an open-source project, meaning the source\n      code\u2014or \"source design\" if you will\u2014that is used to build the font files\n      <a href=\"https:\/\/github.com\/rsms\/inter\">are freely available<\/a> to improve upon.\n      Font making requires a fair bit of technical work and\n      depending on what you'd like to do, some things might be more fun\n      depending on your technical skills.\n      The <a href=\"https:\/\/github.com\/rsms\/inter\/blob\/master\/CONTRIBUTING.md\">\"Contributing\" document<\/a> is a great place to start. The document outlines where\n      you can have the biggest impact, how things are setup and how to get\n      started.\n    <\/li>\n    \n    <li class=\"a\">\n      Inter UI was developed in an a private, internal git repository\n      starting in November 2016, prior to being published on August 22, 2017.\n      Between November 2016 and August 2017, there were\n      2\u2006990\u2006150 line edits made across 247 versions.\n      The reason the public GitHub repository does not reflect this is the\n      fact that the project was initially only internal at the company where\n      the author works and had some sensitive information \"checked in\",\n      like AWS server details and internal author identity in\n      all commit messages. Maybe one day we can write an elaborate git\n      filter-branch program and convert the filter the old repository to make\n      it public, but what would be the point of that?\u00a0 :\u2014)\n    <\/li>\n    <li class=\"q\" id=\"faq-contact\">\n      I've made a cool thing that uses Inter UI, can I share it with you?\n      \u00a0or<\/li>\n    <li class=\"q\">\n      I have a different question\n    <\/li>\n    <li class=\"a\">\n      Reach out on <a href=\"https:\/\/twitter.com\/rsms\">Twitter (@rsms)<\/a> or over <a href=\"https:\/\/rsms.me\/cdn-cgi\/l\/email-protection#f98b988a948c8ab997968d909697d78a9c\">email<\/a>\n    <\/li>\n  <\/ul>","created_at":"2018-11-07T16:21:53+0000","updated_at":"2018-11-07T16:21:53+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"rsms.me","preview_picture":"https:\/\/rsms.me\/inter\/res\/poster.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":17,"title":"Steli Efti explains how to price a product, sell it to customers, and build a sales team.","url":"https:\/\/stripe.com\/atlas\/guides\/ama-steli-efti","content":"<p>Each week, we invite an industry leader to host a Q&amp;A session with the Stripe Atlas Community. Steli Efti, cofounder and CEO of Close.io, recently answered questions from founders in Stripe\u00a0Atlas.<\/p><p><strong>What cold email subject lines lead to the highest open\u00a0rates?<\/strong><\/p><p>First, a caveat. I\u2019ve seen really high open rates, in the 80 percent range, but that doesn\u2019t mean that those emails were ultimately successful. One of the best subject lines I ever received from a salesperson was the subject line \u201cReally disappointed\u2026\u201d and when I saw that, I instantly thought: \u201cWhat did we mess up? What did I do wrong? What happened here?\u201d It was the first email I opened after a long vacation. I had hundreds and hundreds of emails, but that one stood\u00a0out.<\/p><p>The subject line was \u201cVery disappointed\u2026\u201d and then the body of the email continued, \u201c\u2026that we haven\u2019t been able to connect. I called you multiple times because our web development shop\u2026\u201d I thought: \u201cMother effer, you got me to open the email\u201d, and then I deleted it, because if the first thing that starts our relationship is you tricking me into opening your email, that\u2019s not going to make me trust you and want to\u00a0continue.<\/p><p>I could make the subject line \u201cI have your parents in my basement\u201d and it would get a very high open rate. Even people whose parents aren\u2019t alive anymore would open that email. But that doesn\u2019t mean it\u2019s going to translate into a customer\u00a0relationship.<\/p><p>That said, a healthy benchmark to aim for\u2014if you don\u2019t use any trickery\u2014is somewhere between 30-40%. If you can get that in outbound emails, that\u2019s very, very good. If your open rate is below 15%, then you\u2019re in trouble, and I would focus on improving your subject line before doing anything\u00a0else.<\/p><p><strong>What are the characteristics of great cold\u00a0emails?<\/strong><\/p><p>A great cold email clearly and concisely answers the right questions at the right time, and it ends with a single call to\u00a0action.<\/p><p>The subject line both piques somebody\u2019s curiosity and promises something that the body of the email will deliver. If I\u2019m receiving your email, I read the subject line and I think: \u201cHuh, what is this about?\u201d So I open it. Your body text needs to immediately answer the questions in my mind: Who is this? What do you want from me? Of all of the people in the world, why did you reach out to\u00a0me?<\/p><p>If you answer those questions well, each one of your sentences makes me think: \u201cOkay, this is who this person is, this is what they do, and this is why they think this is important to\u00a0me.\u201d<\/p><p>You need to end your cold email with a single, specific call to action\u2014just one. It cannot say: \u201cRead this email, do this thing, sign up for a free trial, click the calendar, and also talk to these three people.\u201d No one will ever do that. Just end it with one\u00a0thing.<\/p><p>So in sum: a curiosity-provoking subject line, a question-answering body text, and a single CTA. That\u2019s the flow of a great cold\u00a0email.<\/p><p><strong>If you cold call prospects, what\u2019s the best way to actually start the\u00a0conversation?<\/strong><\/p><p>It\u2019s similar to cold emailing. The most important thing is to put yourself in the shoes of the person receiving the call. When somebody picks up the phone and they don\u2019t know the number, they immediately think: \u201cWho is this person? And are they trying to sell me something?\u201d You need to quickly answer these two questions. State who you are, what you do, and why you\u2019re\u00a0calling.<\/p><p>When it comes to cold calling, almost more important than what you say is how you say it. Your tonality\u2014your voice\u2014will create a certain feeling in the person that receives the call. If you sound overly nervous, you could say the greatest thing on Earth and I\u2019m going to want to hang up. If you sound overly aggressive, you could say, \u201cI\u2019m about to wire you a million dollars\u201d, and I\u2019m going to hang up. It doesn\u2019t matter how great the words are if the way you deliver them is bad. You could say something terrible, deliver it well, and you\u2019ll get people to keep listening\u2014and they won\u2019t even know\u00a0why!<\/p><p>If you want to sound good on the call, you need to put yourself in a good state of mind. You want to be excited to be talking to them, because, if you\u2019re not excited, why should they be excited to listen to you? You want to be clear and concise in your language\u2014not too slow but not too fast. If you\u2019re nervous, afraid, or anxious, it\u2019ll sound weird and I\u2019ll want to hang\u00a0up.<\/p><p>So: Be in a good state of mind, say hello, and explain who you are and why you\u2019re\u00a0calling.<\/p><p><strong>What is the average response rate, per medium, for a B2B SaaS\u00a0company?<\/strong><\/p><p>When it comes to emails, you should have a minimum open rate of 20-30%. Your response rate among those people should be in the 10-20%\u00a0range.<\/p><p>When it comes to cold calls, you should have a minimum answering rate of about 15%. Your leads-to-something rate among people who answer the phone should be about\u00a015%.<\/p><p>And again, those numbers are the lower bounds. If you are lower than that, stop doing what you\u2019re doing because fundamentally it won\u2019t work out. The math just won\u2019t add up. If you\u2019re in that state, something about what you\u2019re doing\u2014your subject line or how you start the conversation\u2014is fundamentally wrong. You need to fix it before you\u00a0continue.<\/p><p><strong>Do you have any advice for founders trying to price their\u00a0product?<\/strong><\/p><p><a href=\"https:\/\/stripe.com\/atlas\/guides\/business-of-saas\">Charge more<\/a>. If you\u2019re an early stage founder building a B2B business, my number one piece of advice is to charge\u00a0more.<\/p><p>Most founders try to price their product so cheaply that price will never be an objection. That\u2019s not the right way to price your\u00a0product.<\/p><p>The sweet spot in pricing is when you\u2019re at a point\u00a0where:<\/p><ul><li>\n    <p>20% of the people you talk to say: \u201cThis is insanely expensive! I would never buy this. This is\u00a0outrageous!\u201d<\/p>\n  <\/li>\n  <li>\n    <p>20% of people are like: \u201cThis is kind of cheap, I would even pay more for\u00a0this.\u201d<\/p>\n  <\/li>\n  <li>\n    <p>60% of people really think about whether your product is worth the price. And then they conclude: \u201cYeah, if it was cheaper that would be nice. But I do think the value is much higher than the\u00a0price.\u201d<\/p>\n  <\/li>\n<\/ul><p>That\u2019s the pricing sweet\u00a0spot.<\/p><p><strong>How can you create urgency at the end of a web\u00a0demo?<\/strong><\/p><p>There are two\u00a0ways.<\/p><p>The first way is quite common but I\u2019m not a big fan of it. \u201cHey, we\u2019re running a promotion! If you buy before X amount of time you get Y better price, or you get Z more seats, or you get an upgraded product for the lower package, or whatever.\u201d Some kind of a promotion that tells them: \u201cHey, if you make a decision sooner rather than later there\u2019s going to be a big lasting benefit for you.\u201d But I dislike this approach. I think it\u2019s\u00a0gimmicky.<\/p><p>The best way to create urgency is to understand what is urgent in this person\u2019s professional life and then tie your product to that thing. Can your product help them accomplish this really important\u00a0thing?<\/p><p>For example, if the biggest priority that we have right now in our business is to hit our quota for numbers, and you have a marketing tool and you talk to me about how your marketing tool will help me get more likes on Twitter, there\u2019s going to be a disconnect between my priorities and what you\u2019re telling me. But if you could credibly explain to me how your marketing tool will help me hit my numbers? Now you have my attention. Try to figure out what\u2019s urgent for them, and then connect your product to that\u00a0thing.<\/p><p>You can learn more about this in <a href=\"https:\/\/close.io\/resources\/demos\/\">my free\u00a0ebook<\/a>.<\/p><p><strong>When there is human interaction in the sales process, what things should always\u00a0happen?<\/strong><\/p><p>The most important thing is qualification. You\u2019re going to have the most success if you act like an expert doctor. First you spend a lot of time trying to deeply understand who is in front of you\u2014how they think and feel, what the problem really is. Once you have a diagnosis for what\u2019s going on, you need to ask yourself, can our medicine help this person? And if the answer is yes, then you\u2019re not debating, discussing, or asking. You\u2019re\u00a0prescribing.<\/p><p>At that point, you go: \u201cI know what to do here. Here\u2019s what you have, here\u2019s our medicine, here are the alternatives, and here\u2019s why ours is the best. You\u2019re going to start taking the medicine today, and you\u2019re going to take two per day for 14\u00a0days.\u201d<\/p><p>If you do that, it\u2019s not a debate anymore. I made the decision for you because I\u2019m an expert and I want you to succeed. But in order to get to this point, the first thing I need to do is truly be convinced that our product is the best for\u00a0you.<\/p><p>A lot of people don\u2019t take that step. They think: \u201cWell, I have a product, I need to try to sell it to everybody.\u201d So when they start interacting with customers, they instantly jump into selling mode: \u201cLet me tell you about my medicine. My medicine is really great. Here\u2019s everything that\u2019s great about it. And also it\u2019s really cheap.\u201d But we haven\u2019t had a conversation about who the person is, what pain they have, and what solutions they\u2019ve tried\u00a0before.<\/p><p>You need to spend time truly understanding them. It creates trust and credibility, so at the moment you tell them \u201cI know how to fix this\u2014my medicine\u201d, they\u2019re going to believe\u00a0you.<\/p><p>The next thing I\u2019ll highlight is you need to follow up and follow through. You need to champion the relationship. The ball is never in their court. It\u2019s always your job to help them and guide them to keep progressing to a\u00a0decision.<\/p><p><strong>Should you charge enterprise customers for\u00a0pilots?<\/strong><\/p><p>Yes. You want paid pilots. When it comes to enterprise, you want to uncover real buying intent as fast as possible. Even if the product is totally buggy or not finished yet, you should ask for a paid\u00a0pilot.<\/p><p>A lot of times, people who are insecure about their product will offer it for free as a way to feel more comfortable, as a way to offer the customer something that\u2019s \u201cfair\u201d. I would argue strongly against\u00a0that.<\/p><p>If you\u2019re inclined to do that, don\u2019t. Instead, ask them for money, tell them it\u2019s completely refundable, and then don\u2019t under any circumstance spend that money. Put it in a separate bank account. It\u2019s not revenue until the customer has stayed for six months and says that they are happy with everything\u2014then you can touch the\u00a0money.<\/p><p>This has the same effect as giving your product away for free\u2014there\u2019s zero risk for the customer\u2014but by doing this you\u2019ll weed out bad customers and you\u2019ll learn how to get customers to pay\u00a0you.<\/p><p>In the enterprise world, if you\u2019re not putting a price tag on your product, it\u2019s not going to be valued. A lot of times people think I\u2019m going to start by not asking for money and then it\u2019ll organically lead to asking for money. That\u2019s not true. You have to charge enterprise customers, no matter how early it is. If you don\u2019t, a lot of people are going to be friendly and give you pleasant feedback. \u201cOh, new technology, of course I want to see this!\u201d It\u2019s even going to feel like you\u2019re accomplishing things. But you\u2019ll be wasting your\u00a0time.<\/p><p><strong>Any tips for building and scaling a sales\u00a0team?<\/strong><\/p><p>There are four\u00a0things.<\/p><p>In the beginning, you\u2019ll sell in a non-scalable way, and that\u2019s totally fine. But eventually you have to start selling in ways that you could teach somebody else to do. That\u2019s step\u00a0one.<\/p><p>Step two. Once you\u2019ve had some success selling in a teachable, repeatable way, hire junior salespeople, ideally in small teams of two or three. Transition from doing it yourself to teaching a few other people how to get similar\u00a0results.<\/p><p>The reason why it\u2019s important to hire two or three people at a time is that if you just hire one person and they don\u2019t perform well, you\u2019re not going to know if it\u2019s you teaching them poorly, if it\u2019s them, or if it\u2019s the market. But if you have two or three people, it\u2019s very unlikely that they\u2019re all going to perform the same way. One person will do much better than the other two, which will give you a lot of insight into what kind of person you should\u00a0hire.<\/p><p>The third step is to hire a Junior Sales Manager. This is ideally somebody that is a year or two ahead of you\u2014someone who was at a startup at your stage two years ago and they became a manager there and helped grow the\u00a0team.<\/p><p>You hire this person because they\u2019ve seen the future! They\u2019ve done it before. At this point you\u2019ll have a little bit of a process, and they\u2019ll instantly know how to improve it. Once they scale the team to 15-20 people and the team has become so big that you need to have multiple Sales Managers, that\u2019s the time to find a VP or a Senior Sales Manager, somebody who is really experienced in scaling things. They\u2019ll know how to open more sales offices, improve the compensation structure, put together a training and career plan for the company, and hire other senior\u00a0people.<\/p><p>You want to go from doing it yourself to hiring junior people to hiring the junior manager to hiring the senior manager. You can learn more in <a href=\"https:\/\/close.io\/resources\/the-sales-hiring-playbook\/\">my free\u00a0ebook<\/a>.<\/p><p><strong>Any tips for selling\u00a0internationally?<\/strong><\/p><p>Start by focusing on international customers who already speak your language, so that you don\u2019t have to speak a foreign\u00a0language.<\/p><p>For us, 40% of our customers are international, but we didn\u2019t do anything for that. Our product is still just in English and we sell it in US dollars. We actually make it hard for international customers to purchase, but they still do because our product is strong enough, good enough, valuable enough for people to want to buy\u00a0it.<\/p><p>If you\u2019re just starting to sell internationally, I would <em>not<\/em> do a lot of over-optimization to support local languages and currencies. It\u2019s too complex to handle at the\u00a0beginning.<\/p><p>That said, if you\u2019re early on your journey and you <em>do<\/em> want to do local optimization, I\u2019d suggest doing the simplest thing you can do: Take your website, and create a \/germany landing page that routes based on IP address. It could still be in your native language, and it might just have a few copy changes, but you\u2019ll quickly learn whether you should further invest in\u00a0localization.<\/p><p><strong>Any parting\u00a0thoughts?<\/strong><\/p><p>Don\u2019t think of yourself as somebody without sales experience. You have sales\u00a0experience!<\/p><p>Sales is nothing more than results-driven communication. Any time you communicate with someone and you have an end goal, you are selling. Any time someone does that to you, they are selling to you and you\u2019re\u00a0buying.<\/p><p>Selling is just a type of communication. You\u2019re already doing it all of the time. So don\u2019t doubt yourself. Just do\u00a0it.<\/p><aside class=\"collect-email-guide\">\n  <\/aside><a href=\"https:\/\/stripe.com\/atlas\/guides\" class=\"common-Link common-Link--arrowL\">Return to guides<\/a>","created_at":"2018-11-07T16:21:52+0000","updated_at":"2018-11-07T16:21:52+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":13,"domain_name":"stripe.com","preview_picture":"https:\/\/stripe.com\/img\/v3\/atlas\/guide\/social.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":18,"title":"","url":"https:\/\/www.pmkelly.net\/publications\/thesis.pdf","created_at":"2018-11-07T16:21:52+0000","updated_at":"2018-11-07T16:21:52+0000","annotations":[],"reading_time":0},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":19,"title":"","url":"https:\/\/www.pmkelly.net\/publications\/thesis.pdf","created_at":"2018-11-07T16:21:52+0000","updated_at":"2018-11-07T16:21:52+0000","annotations":[],"reading_time":0},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":20,"title":"PDF","url":"https:\/\/www.pmkelly.net\/publications\/wage2008.pdf","content":"Lambda Calculus as a Work ow Model<br \/>\nPeter M. Kelly, Paul D. Coddington, and Andrew L. Wendelborn School of Computer Science<br \/>\nUniversity of Adelaide<br \/>\nSouth Australia 5005, Australia<br \/>\n{pmk,paulc,andrew}@cs.adelaide.edu.au<br \/>\nAbstract<br \/>\nData-oriented work ows are often used in scienti c ap-<br \/>\nplications for executing a set of dependent tasks across mul- <br \/>\ntiple computers. We discuss how these can be modeled us-<br \/>\ning lambda calculus, and how ideas from functional pro-<br \/>\ngramming are applicable in the design of work ows. Such<br \/>\nan approach avoids the restrictions often found in work ow<br \/>\nlanguages, permitting the implementation of complex appli- <br \/>\ncation logic and data manipulation. This paper explains why lambda calculus is an appropri-<br \/>\nate model for work ow representation, and how a suitably <br \/>\nef cient implementation can provide a wide range of ca-<br \/>\npabilities to developers. The presented approach also per-<br \/>\nmits high-level work ow features to be implemented at user <br \/>\nlevel, in terms of a small set of low-level primitives provided <br \/>\nby the language implementation. <br \/>\n1 Introduction<br \/>\nWork ow systems [17] have emerged in recent years as<br \/>\ntools for building distributed applications involving the co- <br \/>\nordination of software components at different sites. In sci- <br \/>\nenti c  elds, these are generally based on a data-oriented<br \/>\nmodel, where a series of side effect free operations are per-<br \/>\nformed on a collection of input data to produce a result.<br \/>\nEach operation is realised as a taskthat is executed on a re-<br \/>\nmote computer, and invoked by the work ow engine over <br \/>\nthe network. Tasks are generally either jobs submitted via <br \/>\nbatch queuing systems such as Condor [15] or Globus [8],<br \/>\nor invocations of RPC-style services using protocols such<br \/>\nas SOAP. Examples of work ow systems include DAGMan<br \/>\n[15], Chimera [9], Taverna [6], and Kepler [5].<br \/>\nAmodel of computation speci es the way in which the<br \/>\ntask invocation is carried out at an abstract level. Usually,<br \/>\nthis is based on a set of data dependencies between tasks,<br \/>\nso that a given task only executes once all of its inputs are<br \/>\navailable, and the outputs are sent to other tasks. These can be expressed as a directed acyclic graph (DAG), which<br \/>\ncan be edited graphically or textually. This model is some- <br \/>\ntimes extended with additional features for loops, condi- <br \/>\ntional tests, and nested work ows.<br \/>\nMany of these systems are based on models of com-<br \/>\nputation which only support a limited set of programming<br \/>\nconstructs. Typically, these models are either not Turing<br \/>\ncomplete, or do not support  ne grained computation. This<br \/>\nmeans that only very simple programming logic can be im-<br \/>\nplemented in the work ow language, and any complex work<br \/>\nmust be carried out by external jobs or services written in <br \/>\nmore powerful programming languages. This is a prob- <br \/>\nlem for complex work ows, because it sometimes means<br \/>\nthat a developer must switch to a different language to per-<br \/>\nform certain actions, such as data manipulation or interme-<br \/>\ndiate computation on the values exchanged between differ-<br \/>\nent tasks.<br \/>\nThis issue can be addressed by using a more  exible<br \/>\nmodel of computation which is Turing complete and can <br \/>\nbe implemented in such a way that both high-level and low- <br \/>\nlevel programming is equally well supported. It is prefer-<br \/>\nable that such a model maintains the advantages of existing<br \/>\nwork ow models, such as explicit data dependencies, lack<br \/>\nof side effects, and a level of abstraction above that of main-<br \/>\nstream imperative programming languages. In this paper, <br \/>\nwe discuss the suitability of lambda calculus for expressing <br \/>\nwork ows, and how it can meet these requirements. <br \/>\n1.1 Lambda calculus<br \/>\nLambda calculus [2] is an abstract model of computation<br \/>\nwhich is the theoretical foundation of functional program- <br \/>\nming. It speci es a notation in which functions are de ned as lambda abstractions , consisting of a set of arguments and<br \/>\na body, such as the following:<br \/>\n( a.  b.  c.b (a c c))<br \/>\nWhen such a function is applied to arguments, its body<br \/>\nis instantiated by replacing all variable references with the<br \/>\n1<br \/>\n<br \/>\nsupplied arguments. Evaluation proceeds via a sequence of <br \/>\nreductions, each of which transforms the expression into a<br \/>\nversion that is closer to its normal form, or result value:<br \/>\n(  a.  b.  c.b (a c c)) + sqrt 8<br \/>\n=&gt; sqrt (+ 8 8)<br \/>\n=&gt; 4<br \/>\nAlthough the basic lambda calculus model does not in-<br \/>\nclude any built-in operations or data types, any language<br \/>\nimplementation based on the model will provide some set<br \/>\nof primitives. These may be coarse grained, such as oper- <br \/>\nations to invoke web services with data values representing <br \/>\nXML trees, or  ne grained, such as arithmetic operations<br \/>\nand numeric values. The set of primitives is an implemen-<br \/>\ntation choice which is independent of the basic model of<br \/>\nprogram representation and evaluation.<br \/>\nThe most common technique for evaluating lambda cal-<br \/>\nculus is graph reduction [11]. In this model, the program is<br \/>\nrepresented in memory as a graph, much like a syntax tree<br \/>\nproduced by a parser. Evaluation proceeds by traversing the <br \/>\ngraph to locate expressions that can be reduced, and replac- <br \/>\ning the relevant graph nodes with the result obtained from<br \/>\nevaluation of the expression. This graph representation has<br \/>\nsimilarities with DAG-based work ows, as we shall see in<br \/>\nsection 3.<br \/>\nAn important property of the lambda calculus is that the<br \/>\nresult of evaluation is independent of the order in which the<br \/>\nreductions are performed. This is known as the Church-<br \/>\nRosser theorem [3], and is what enables separate parts of <br \/>\nthe graph to be safely reduced in parallel. In the context of <br \/>\nwork ows, this means invoking multiple tasks at the same<br \/>\ntime. This relies on the assumption that all tasks are side<br \/>\neffect free, which is typically the case for scienti c work-<br \/>\n ows. <br \/>\n1.2 Motivation<br \/>\nThere are several advantages to using lambda calculus as<br \/>\na model for expressing work ows. Firstly, text-based rep- <br \/>\nresentations scale better than graphical views, which can be<br \/>\nan issue for complex work ows. Text editors are easily ca- <br \/>\npable of handling large  les with hundreds of lines of code, <br \/>\nwhile graphical work ow editors generally do not provide<br \/>\nan interface that deals with such complexity easily. Expres-<br \/>\nsions are also a more compact syntax than separate node<br \/>\nand edge declarations.<br \/>\nSecondly, the use of lambda calculus as a computational<br \/>\nmodel lends itself to the application of theory developed<br \/>\nin the context of functional programming. Common tasks<br \/>\nsuch as iteration, tail recursion, and conditional expressions <br \/>\ncan easily be expressed with this model, as can data struc- <br \/>\nture manipulation.<br \/>\nFinally, we can apply implementation techniques from\t<br \/>\nFigure 1. Task dependency DAG<br \/>\nthe functional programming community to execute work- <br \/>\n ows more ef ciently by compiling them into executable<br \/>\ncode, which is particularly important for complex work-<br \/>\n ows including data transformation and application logic. <br \/>\n2 Graph types Task dependency graphs, sometimes referred to as di-<br \/>\nrected acyclic graphs (DAGs), contain nodes specifying <br \/>\ntasks, and edges representing data dependencies. The tasks<br \/>\nare executed in such an order that if task A depends on task<br \/>\nB, then B will be executed  rst, with its output data being<br \/>\ntransferred to A. Each task can be viewed as a \u201cfunction\u201d <br \/>\ntaking a set of input values and producing an output value. <br \/>\nIn implementation terms, function evaluation corresponds<br \/>\nto invoking a computation on a remote resource. This model<br \/>\nis typically referred to as data ow.<br \/>\nFigure 1 shows an example work ow. Task A depends<br \/>\non B and C, which both depend on D. Thus the execution<br \/>\norder will be D  rst, then B and C (either sequentially or<br \/>\nconcurrently), and  nally A. The \u201cresult\u201d of the work ow<br \/>\nwill be the output data generated by A.<br \/>\nAn alternative model is a graph of application nodes. In<br \/>\nthis model, a task (or equivalently, function call) is modeled<br \/>\nas a sequence of applicationnodes, one per input. The last<br \/>\nlink in the chain is a reference to a function (which may be<br \/>\nexecuted remotely), and each of the input parameters are<br \/>\nalso graphs. Figure 2 shows the above work ow in this<br \/>\nmodel, with application nodes represented by the character@.<br \/>\nThe key differences with this model are that nodes can<br \/>\ncorrespond to data values, and functions are treated as val-<br \/>\nues. Functions can thus be passed around as parameters and<br \/>\nreturned as results from so-called higher orderfunctions.<br \/>\nThis property enables a wide range of useful techniques that<br \/>\nare common in functional programming, and is also useful<br \/>\nfor modeling abstract work ows, as described in Section4.7.<br \/>\n2<br \/>\n<br \/>\nFigure 2. Function application graph<br \/>\n3 Work ow representation There are several common ways of representing the in-<br \/>\nformation that speci es a work ow structure. One that is <br \/>\ncommon in many work ow systems is a graphical represen-<br \/>\ntation, in which task nodes are represented as shapes, with <br \/>\nlines and arrows indicating dependencies between them. <br \/>\nThis representation is aimed at end users who utilise a<br \/>\ngraphical editor to construct their work ow. Figure 1 is a<br \/>\ntypical example of this representation.<br \/>\nEven if a graphical editor is used, a text-based format<br \/>\nis also present for storing the  le on disk. An obvious se-<br \/>\nrialisation of the graphical view is to represent the graph <br \/>\nby a series of statements de ning the set of nodes, along <br \/>\nwith another set de ning the edges. The statements may be<br \/>\nrepresented either as XML elements, or using some other<br \/>\nsyntax. An example of the graph from Figure 1 represented<br \/>\nin this way is the following: <br \/>\nnode A <br \/>\nnode B<br \/>\nnode C<br \/>\nnode D<br \/>\nedge B -&gt; A <br \/>\nedge C -&gt; A <br \/>\nedge D -&gt; B<br \/>\nedge D -&gt; C<br \/>\nAn alternative way of representing a graph is as an ex-<br \/>\npression , in which each node is speci ed by writing its<br \/>\nname, followed by a list of parenthesised sub-graphs that <br \/>\npoint to it. In fact, this corresponds exactly to the syntax <br \/>\nused by the lambda calculus. The same graph expressed in<br \/>\nthis manner would be as follows: <br \/>\nA (B D) (C D)\t<br \/>\nFigure 3. Graph requiring a let expression<br \/>\nThis representation is more concise, and closer to the<br \/>\nway in which programmers normally write expressions. <br \/>\nHowever, it does not properly handle tasks whose output <br \/>\nis sent to more than one destination, as is the case with D. <br \/>\nThe syntax is still acceptable in this example, since D takes<br \/>\nno inputs. However, tasks which do take inputs, such as H<br \/>\nin Figure 3, need a better solution. This problem can be<br \/>\nsolved using a lambda abstraction, by binding a symbol to <br \/>\nthe sub-graph, and referencing it where necessary: (  x.E (F x) (G x)) (H I J)<br \/>\nIt is straightforward to extend the syntax of lambda cal-<br \/>\nculus with letexpressions, which can be transformed at<br \/>\nparse time into the above format, e.g: <br \/>\nlet x = (H I J) in (E (F x) (G x))<br \/>\nEach of these graphical and textual formats is capable<br \/>\nof representing the same information, and transformation <br \/>\nfrom one to another is straightforward. A work ow engine<br \/>\nthat uses lambda calculus as its input language can execute<br \/>\nwork ows that have been exported from graphical editors, <br \/>\nor translated from other DAG-based work ow languages.<br \/>\nConverting these work ows into lambda calculus does<br \/>\nnot alter their high-level nature. We have still described <br \/>\nnothing about what each of the tasks actually does, how the<br \/>\nwork ow engine causes them to be invoked, or the way in<br \/>\nwhich data is passed between them. These are still details<br \/>\nwhich we consider to be abstracted away; the only differ-<br \/>\nence is the notation in which we have expressed the work- <br \/>\n ows. In Section 4, we will discuss how these functions <br \/>\nmay be implemented in concrete terms. <br \/>\n4 Modeling work ow features<br \/>\nA model of computation speci es the abstract nature of<br \/>\ntasks and how they are used together during execution, but <br \/>\ndoes not specify how those tasks are actually implemented.<br \/>\nAny work ow system must provide facilities for invoking a<br \/>\n3<br \/>\n<br \/>\ntask on a remote machine. In most cases, this invocation is <br \/>\nprovided as a primitive within the work ow language, and<br \/>\nimplemented explicitly within the execution engine. This<br \/>\nis unavoidable for work ow languages that are not Turing<br \/>\ncomplete, because the low-level interaction with the remote <br \/>\nhost often requires language facilities that are only present <br \/>\nin more powerful languages.Because lambda calculus is capable of expressing arbi-<br \/>\ntrarily complex computations, an ef cient implementation<br \/>\nthat provides a few basic data types and low-level opera-<br \/>\ntions can enable task invocation to be implemented in terms<br \/>\nof the work ow language itself. Instead of coarse grained<br \/>\nprimitives for launching tasks,  ne grained primitives for <br \/>\nlow-level operations like establishing TCP connections and <br \/>\nwriting binary data may be used to build up higher level ab-<br \/>\nstractions for task invocation. This approach is analogous<br \/>\nto a micro kernel based operating system, in which all high-<br \/>\nlevel functionality is implemented at user level on top of<br \/>\nbasic primitives provided by the runtime environment ex- <br \/>\nposed by the kernel.<br \/>\nThe work ows discussed in the previous section rep-<br \/>\nresented each task as a function application, such as (B<br \/>\nD) , which is a call to the task B, supplying the output of<br \/>\nD. These functions were not de ned, but assumed to be <br \/>\nmapped by the work ow engine into remote tasks. This<br \/>\nmapping can be achieved by making each function a built-<br \/>\nin primitive, or de ning those functions in the work ow <br \/>\nlanguage as expressions which call lower level networking <br \/>\nprimitives to invoke the remote computation. In this section,<br \/>\nwe describe how this latter approach can be used to provide<br \/>\nmuch of the functionality that is normally implemented as<br \/>\nprimitives by existing work ow engines. <br \/>\n4.1 Primitives<br \/>\nIn order to implement high-level work ow functionality<br \/>\nin terms of low-level operations, it is necessary for the lan- <br \/>\nguage implementation to provide a small number of basic <br \/>\nprimitives. These should include capabilities for manipulat- <br \/>\ning data at the byte level, performing arithmetic operations,<br \/>\nand manipulating data structures. Table 1 lists a possible set<br \/>\nof operations. The built-in data types should include at least<br \/>\nnumeric values, cons pairs, and nil values.<br \/>\nWhile the basic lambda calculus can theoretically be<br \/>\nused to express any computation, primitives like these are<br \/>\nnecessary in order to achieve reasonable performance. Ad- <br \/>\nditionally, many optimisations can be made within the lan- <br \/>\nguage implementation to gain ef ciency, such as storing<br \/>\ncons lists internally as arrays, and compiling expressions<br \/>\ninto executable code. As we are interested primarily in the<br \/>\ncomputational model, we will not go into further details of<br \/>\nthese optimisations here.<br \/>\nMany abstractions can be built up from these primitives\t+, -, *, \/, %\tArithmetic functions\t<br \/>\n==, !=, &lt;, &lt;=, &gt;, &gt;=\tNumeric comparison\t<br \/>\nand, or, not, if\tLogical &amp; conditional\t<br \/>\ncons, head, tail\tList operations\t<br \/>\nconnect\tNetwork connections\t<br \/>\nTable 1. Primitive operations<br \/>\nthat can be used to support high-level work ow functional- <br \/>\nity. Cons lists can be used to support a wide range of data <br \/>\nstructures, such as strings (lists of character codes), binary <br \/>\ndata (lists of byte values), and XML trees (lists of element<br \/>\nobjects containing strings and other elements). Conditionals<br \/>\nare supported by the iffunction; iteration can be achieved<br \/>\nusing tail recursion, and map,  lter, and reduction opera-<br \/>\ntions can be de ned in terms of the built-in list operations. <br \/>\nThese techniques are well known in the functional program- <br \/>\nming literature. The connect primitive deserves special attention. We<br \/>\nde ne it as a built-in function taking three parameters: host,<br \/>\nport, and output stream. The  rst two are used when estab- <br \/>\nlishing a connection, while the third is a cons list supplied <br \/>\nby the program, which the runtime system writes out to the<br \/>\nconnection. The result of a call to the connectfunction<br \/>\nis itself a cons list, which can be read from lazily to obtain<br \/>\ndata from the connection. It can be treated as a string by in-<br \/>\nterpreting each value in the list as a character code, which is <br \/>\nuseful for implementing text-based protocols such as HTTP. <br \/>\n4.2 Job submission<br \/>\nA batch queuing system accepts a request from a client<br \/>\nspecifying details of a job to run, and then noti es the client <br \/>\nonce the job has completed. The way in which the net-<br \/>\nwork interaction occurs depends on the protocol used by<br \/>\njob submission mechanism. In some cases, separate net-<br \/>\nwork connections may be used to submit the request and <br \/>\nthen later retrieve the result; in others, a single connection <br \/>\nmay be kept open until the job completes, and the results<br \/>\nsent back directly. The logic to handle this interaction may<br \/>\nbe implemented as a set of support routines provided by the<br \/>\nwork ow system, using lambda calculus combined with the <br \/>\nprimitives given above. As a simple example, consider a job queue that accepts<br \/>\nsubmissions via HTTP GET requests, and sends the job <br \/>\noutput directly back to the client in the HTTP response.<br \/>\nA script on the web server named runjob.placcepts a<br \/>\nprogram parameter specifying the name of the executable<br \/>\n le, and an argsparameter containing the command line<br \/>\narguments to be passed to the program. The code in Figure<br \/>\n4 invokes this script by sending it a request with the query<br \/>\nstring formed based on supplied program name and argu-<br \/>\nment list. This code assumes separately de ned routines<br \/>\n4<br \/>\n<br \/>\nsubmit_job = ( host.  prog.  args.<br \/>\n(parse_response\t<br \/>\n(connect host 80\t<br \/>\n(++ \"GET \/runjob.pl?program=\"<br \/>\n(++ (urlencode prog)<br \/>\n(++ \"&amp;args=\"<br \/>\n(++ (urlencode (join \" \" args))\t<br \/>\n\" HTTP\/1.0<br \/>\r\n<br \/>\r\n\")))))))\t<br \/>\nFigure 4. Support routine for job submission<br \/>\nare available for string concatenation, list joins, parameter <br \/>\nencoding, and HTTP response parsing. It also assumes that<br \/>\nthe parser converts strings into cons lists of characters. Individual tasks can be be implemented as functions<br \/>\nwhich call this routine, supplying concrete details about the<br \/>\njob to be executed. For example, a task A that takes two<br \/>\nstring arguments could be de ned as follows:<br \/>\nA = (  x.  y.submit_job \"hydra\"<br \/>\n\"\/home\/pmk\/myprog\" (cons x (cons ynil)))<br \/>\nThis task may then be called from a work ow, in this<br \/>\ncase taking as input the results of tasks B and C:<br \/>\nA (B D) (C D) <br \/>\nImplementing tasks in such a manner preserves our abil-<br \/>\nity to deal with them at a high level. The work ow speci - <br \/>\ncation remains the same as the example given in Section 3; <br \/>\nhere we have gone into one more level of detail to specify <br \/>\nhow it is realised in actual terms. The concrete implementa-<br \/>\ntions may be arbitrarily complex, involving all of the neces-<br \/>\nsary mechanisms to invoke a job, including support actions<br \/>\nsuch as staging data in or out of the remote host, and moni-<br \/>\ntoring for task completion. <br \/>\n4.3 Services<br \/>\nServices can be accessed in a similar manner to the<br \/>\nabove. RPC mechanisms are generally implemented by <br \/>\nhaving the client establish a connection to the server, send <br \/>\nit a request containing the operation name and input param-<br \/>\neters, and then having the server send back the result. For<br \/>\nexample, web services utilise the HTTP request\/response<br \/>\nmechanism and encode the parameters and results using <br \/>\nXML and SOAP.<br \/>\nA work ow task corresponding to a web service invoca-<br \/>\ntion can be de ned as a function which makes calls to sup- <br \/>\nport routines that submit a SOAP request to a web server. <br \/>\nDepending on how the work ow system is implemented, <br \/>\nthe values passed between tasks could be represented as<br \/>\nplain strings, or more complex data structures such as XML<br \/>\ntrees. It is the responsibility of the support routines to con-\t<br \/>\npostreq = (  host.  path.  req.<br \/>\n(connect host 80\t<br \/>\n(++ \"POST \"<br \/>\n(++ path<br \/>\n(++ \" HTTP\/1.0<br \/>\r\n\"<br \/>\n(++ \"Content-Type: text\/xml<br \/>\r\n\"<br \/>\n(++ \"Content-Length: \"<br \/>\n(++ (numtostring (len req))<br \/>\n(++ \"<br \/>\r\n<br \/>\r\n\" req)))))))))\t<br \/>\nsoapcall = (  host.  path.  body.<br \/>\n(parse_response\t<br \/>\n(postreq host path\t<br \/>\n(++ \"\"<br \/>\n(++ body\t<br \/>\n\"\")))))\t<br \/>\nFigure 5. Support routines for web services<br \/>\nvert between these representations and the on-the-wire for- mat. As an example, the code in Figure 5 implements a sim-<br \/>\npli ed version of SOAP over HTTP. The postreqfunc-<br \/>\ntion accepts a host and URL path, as well as a request body.<br \/>\nIt makes a HTTP POST request containing the appropri-<br \/>\nate headers. The soapcallfunction accepts the host and<br \/>\npath, as well as the arguments to be passed to the service,<br \/>\nwhich are assumed to be already encoded in XML. It wraps<br \/>\nthese in a SOAP envelope and posts the request to the server.<br \/>\nThe response is then parsed to extract the result value.<br \/>\nConsider a work ow task B that calls a stock<br \/>\nquote service. It takes a single parameter specify-<br \/>\ning the symbol name, and submits a SOAP request to<br \/>\nhttp:\/\/stockexchange\/getquote using the above routines:<br \/>\nB = (  sym.soapcall \"stockexchange\"<br \/>\n\"\/getquote\" (++ \"\" (++ sym\"\"))) <br \/>\n4.4 Shims<br \/>\nA common problem encountered when developing<br \/>\nwork ows is that tasks sometimes use different input and <br \/>\noutput formats. These formats may contain the same infor-<br \/>\nmation but with a different syntax, or similar information <br \/>\nthat requires both syntactic and semantic transformations to <br \/>\nachieve compatibility. When such conversion is needed, ad-<br \/>\nditional components must be added to the work ow to per-<br \/>\nform the necessary conversion. These are known as shims<br \/>\nor adapters [7, 10].<br \/>\nThe need for shims is due to the lack of support in many<br \/>\nwork ow languages for data manipulation. Instead of be-<br \/>\ning able to access  elds of an object or perform basic string<br \/>\n5<br \/>\n<br \/>\nmanipulation using built-in language constructs or APIs, a <br \/>\ndeveloper must create separate components or services to<br \/>\nperform these tasks. Features like these, which are com-<br \/>\nmon in regular programming languages, involve signi cant<br \/>\neffort and complexity in work ow languages. While this <br \/>\ncost is sometimes worth paying in order to get the bene- <br \/>\n ts provided by work ow languages, there are many cases<br \/>\nwhere the tradeoff is questionable.Of course, these problems could be avoided by writing<br \/>\nthe work ow entirely in a language like Java or C, but this<br \/>\nwould involve giving up other bene ts like automatic par- <br \/>\nallelisation and fault tolerance that are generally provided <br \/>\nby work ow languages. Instead, it is preferable to use a<br \/>\nlanguage which meets both types of requirements. Lambda<br \/>\ncalculus, combined with basic primitives such as those sug-<br \/>\ngested in Section 4.1, meets this criterion. <br \/>\n4.5 Data parallelism<br \/>\nWork ows can often bene t from data parallelism,<br \/>\nwhere each item in a list of values is processed separately, <br \/>\npotentially on different computers. This sort of process- <br \/>\ning is common in task farming middleware, but is not well<br \/>\nsupported by DAG-based work ow languages. The reason<br \/>\nis that the latter typically assume a  xed number of tasks,<br \/>\nthus requiring sequential iteration to process lists [14]. A<br \/>\nsolution is to use the mapconstruct common in functional<br \/>\nprogramming languages, which can easily be expressed in<br \/>\nlambda calculus using the cons,head , andtail primi-<br \/>\ntives: <br \/>\nmap = (  f.  lst.<br \/>\nif lst (cons (f (head lst))<br \/>\n(map f (tail lst)))<br \/>\nnil)<br \/>\nSuch a call can be automatically parallelised by the run-<br \/>\ntime system in the case of strict evaluation, but may re- <br \/>\nquire explicit annotations [16] in the case of lazy evalua-<br \/>\ntion. Other operations like filterandreduce can be<br \/>\nimplemented in a manner similar to the above. <br \/>\n4.6 Control structures<br \/>\nSome work ow engines provide limited ability to control<br \/>\nthe  ow of execution based on data computed during the <br \/>\nwork ow. These constructs are standard features present<br \/>\nin all major programming languages, and their implemen-<br \/>\ntation in functional programming languages is well known. <br \/>\nFor example, conditionals are implemented using the if<br \/>\nfunction, which takes a boolean expression as well as true <br \/>\nand false branches. When evaluated, it evaluates the condi-<br \/>\ntional parameter and returns either the second or third argu-<br \/>\nment, depending on the result. Sub-work ows can be modeled as functions. Wherever<br \/>\nthey are used, each input link corresponds to an actual pa-<br \/>\nrameter, and the output link is the result of the function call.<br \/>\nMultiple outputs can be handled by wrapping them in lists. <br \/>\nIteration can be modeled using tail recursion, which can be <br \/>\nperformed in constant space by most functional language<br \/>\nimplementations. Each iteration is simply another call to<br \/>\nthe same function but with updated values passed as param-<br \/>\neters. A portion of the work ow that needs to be executed<br \/>\nmultiple times for different inputs would be expressed as an <br \/>\napplication of the mapfunction to a lambda abstraction, as<br \/>\ndescribed in Section 4.5. <br \/>\n4.7 Abstract work ows<br \/>\nSome work ow engines support the concept of abstract<br \/>\nwork ows , which are work ow speci cations that do not<br \/>\nexplicitly specify which resources are to be used. Concrete<br \/>\nwork ows are constructed by instantiating an abstract work-<br \/>\n ow with a speci c set of resources. In lambda calculus, ab- <br \/>\nstract work ows can be modeled as higher order functions<br \/>\nwhich take parameters specifying the concrete implementa-<br \/>\ntions of tasks. For example, the work ow from Section 3 <br \/>\ncan be parameterised as follows: (  A.  B.  C.  D.A (B D) (C D))<br \/>\nThis function can then be applied to a set of function<br \/>\nparameters which implement tasks A-D. If none of the pa- <br \/>\nrameters contain any free variables, it is a concrete work-<br \/>\n ow which can be directly executed. Otherwise, it is a \u201cless<br \/>\nabstract\u201d work ow that contains some implementation de-<br \/>\ntails but is still parameterised. For example, the following<br \/>\nabstract work ow speci es each task as a web service call <br \/>\nto a speci c operation, but is parameterised by the service URLs: <br \/>\nAWF = (  url1.  url2.  url3.  url4.<br \/>\n(  A.  B.  C.  D.A (B D) (C D))<br \/>\n(  x.y.wscall url1 \"a\" ...)<br \/>\n(  x.wscall url2 \"b\" ...)<br \/>\n(  x.wscall url3 \"c\" ...)<br \/>\n(wscall url4 \"d\" ...))<br \/>\nThis abstract speci cation can then be instantiated into a<br \/>\nconcrete work ow by applying the function to a set of pa-<br \/>\nrameters specifying a speci c set of services to be accessed: <br \/>\n(AWF \"http:\/\/a.org\/analyse\" \"http:\/\/b.org\/filter\" \"http:\/\/c.org\/process\" \"http:\/\/d.org\/query\")<br \/>\nOne application of the abstract work ow concept is to<br \/>\nprovide quality of service (QoS) mechanisms, whereby ser-<br \/>\nvices are chosen at runtime based on certain requirements.<br \/>\n6<br \/>\n<br \/>\nFor example, a user of the above work ow may want to <br \/>\nspecify that they want to use the cheapest service available<br \/>\nthat implements A, the fastest available versions of services<br \/>\nB and C, and the most reliable version of service D. Assum-<br \/>\ning the work ow engine provides functions to determine the <br \/>\nbest choice in each case, this could be expressed as follows: <br \/>\n(AWF (find_cheapest \"a\")(find_fastest \"b\") <br \/>\n(find_fastest \"c\")<br \/>\n(find_most_reliable \"d\"))<br \/>\nAbstract work ows de ned using these techniques are a<br \/>\n exible way of achieving reuse. A scientist may develop<br \/>\na work ow parameterised by service addresses, and then <br \/>\nrun it in their local environment by supplying the necessary <br \/>\nset of URLs. The abstract work ow could subsequently be<br \/>\nshared with other scientists, who may run it with the local<br \/>\nversions of those services hosted at their own institution, or<br \/>\nwith a different set of input data. Such usage is equivalent<br \/>\nto a script which uses a con guration  le or command-line <br \/>\nparameters, rather than using hard-coded information.<br \/>\nThe ability to express abstract work ows in this man-<br \/>\nner does not require explicit support from the work ow en- <br \/>\ngine. It comes about as a natural consequence of the in- <br \/>\nherent support for higher order functions that is present in <br \/>\nlambda calculus. It is one of the many examples of how<br \/>\npowerful functionality can be built up from a minimal set<br \/>\nof language constructs, instead of extending a monolithicinfrastructure. <br \/>\n5 Implementation<br \/>\nAlthough the lambda calculus model is suitable for rep-<br \/>\nresenting work ows, additional requirements must be met<br \/>\nfor these representations to be executedin a manner that<br \/>\nachieves the advantages typically associated with work ow <br \/>\nengines. This combination of representation and execution<br \/>\nrequirements together de nes the semantics and pragmatics<br \/>\nof the computational environment in which work ows are <br \/>\nexecuted. Such an implementation effectively constitutes <br \/>\nan implementation of a functional programming language.<br \/>\nParallel evaluation of multiple sub-expressions must be<br \/>\nprovided in order for more than one task to be in progress at <br \/>\nthe same time. Techniques such as parallel graph reduction <br \/>\n[1] may be used to achieve this, based either on strict or lazy<br \/>\nevaluation. In the latter case, strictness analysis [4] and\/or<br \/>\nmanual annotations [16] can be used to indicate which ex-<br \/>\npressions may be safely reduced in parallel.<br \/>\nNetwork connections should be supported in such a man-<br \/>\nner that they can be mapped onto data streams, and prefer-<br \/>\nably exposed using a cons list abstraction. This enables the<br \/>\ndata sent to or received from a service to be treated as data, and for all of the programming techniques that are appli-<br \/>\ncable to list processing to be used in the context of data<br \/>\nstreams. This is particularly useful when combined with<br \/>\nlazy evaluation, because it enables results from tasks to be <br \/>\nprocessed incrementally.<br \/>\nIdeally, some measure of fault tolerance should be pro-<br \/>\nvided, so that if a remote job or service fails, then it can be <br \/>\nretried on another host. The side effect free nature of the<br \/>\nlambda calculus model means that it is safe to re-execute a<br \/>\ntask, even if it was already part way through execution at<br \/>\nthe time that it failed. This property is relied on by many<br \/>\ndistributed computing systems to provide correctness even <br \/>\nin the face of partial failures.<br \/>\nFinally, the execution engine should be ef cient enough<br \/>\nto support  ne grained computation. This is important due <br \/>\nto the fact that low-level primitives are used by remote task<br \/>\ninvocation mechanisms, and also to support data manip-<br \/>\nulation and intermediate computation on the data values<br \/>\nproduced and consumed by external tasks. Techniques for<br \/>\nimplementing functional languages ef ciently are given in [11].<br \/>\nIn [13], we presented an example of a work ow engine<br \/>\nthat implements many of these ideas. It supports paral- <br \/>\nlelism, native code compilation, and an array-based cons list<br \/>\nrepresentation for ef cient stream processing. We have also<br \/>\nimplemented an XSLT compiler designed for web service-<br \/>\nbased work ows [12], using the lambda calculus as an in-<br \/>\ntermediate target. Our current research explores the appli- <br \/>\ncation of these ideas. In many ways, the ideas put forward <br \/>\nin this paper are a reduction of our experiences developing<br \/>\nand applying the system reported in [12] and [13].<br \/>\nThe issues addressed in this paper are solely related to<br \/>\nthe modeling and execution of work ows. The development<br \/>\nof visual programming environments and higher level work-<br \/>\n ow languages are separate concerns that can be addressed<br \/>\nwithin the overall framework we have presented here. Edi- <br \/>\ntors or compilers which output lambda calculus can utilise <br \/>\nan execution engine meeting the above requirements, while<br \/>\nexposing a different (but compatible) interface or syntax to<br \/>\ndevelopers. <br \/>\n6 Conclusion<br \/>\nLambda calculus is a simple, abstract model of compu-<br \/>\ntation. It is independent of the granularity of operations, <br \/>\nand can thus be applied to coarse grained work ows as well <br \/>\nas  ne grained computation. The side effect free nature of <br \/>\nthe model and incorporation of explicit data dependencies<br \/>\nenables techniques such as parallel graph reduction to be<br \/>\nused to coordinate the concurrent invocation of operations.<br \/>\nWhen applied to work ows, this enables multiple remote<br \/>\noperations to be in progress at the same time.<br \/>\n7<br \/>\n<br \/>\nPrevious research on functional programming has pro-<br \/>\nduced techniques for ef ciently executing languages based <br \/>\non the lambda calculus model. These techniques can be<br \/>\nused to build work ow enactment engines in such a way<br \/>\nthat supports features necessary for work ows, such as par- <br \/>\nallelism, but also permits arbitrarily complex,  ne grained <br \/>\ncomputation to be performed within the work ow language<br \/>\nitself. This  ne grained computation can be used to im-<br \/>\nplement support functionality such as network protocols for<br \/>\nlaunching remote tasks, and transforming data between the<br \/>\nrepresentations used by different tasks.<br \/>\nThe lambda calculus model provides a solid founda-<br \/>\ntion for functional programming and, by extension, data-<br \/>\noriented work ows. We have shown how common work-<br \/>\n ow constructs can be expressed in terms of lambda calcu-<br \/>\nlus, both at an abstract level and with regards to concrete<br \/>\nimplementation details. This approach to designing work- <br \/>\n ow engines achieves power through simplicity, and per- <br \/>\nmits the requirements of concurrency and distribution to be<br \/>\nmet without sacri cing expressiveness. <br \/>\nReferences<br \/>\n[1] Lennart Augustsson and Thomas Johnsson. Parallel graph reduction with the (v , g)-machine. In FPCA<br \/>\n\u201989: Proceedings of the fourth international confer- <br \/>\nence on Functional programming languages and com- <br \/>\nputer architecture , pages 202\u2013213, New York, NY,<br \/>\nUSA, 1989. ACM Press.<br \/>\n[2] Alonzo Church. A set of postulates for the founda- tion of logic. The Annals of Mathematics, 2nd Ser. ,<br \/>\n33(2):346\u2013366, April 1932.<br \/>\n[3] Alonzo Church and J. Barkley Rosser. Some prop- erties of conversion. Transactions of the American<br \/>\nMathematical Society , 39(3):472\u2013482, May 1936.<br \/>\n[4] Chris Clack and Simon L. Peyton-Jones. Strictness analysis \u2013 a practical approach. In J. P. Jouannaud,<br \/>\neditor, Conference on Functional Programming and<br \/>\nComputer Architecture , number 201 in Lecture Notes<br \/>\nin Computer Science, pages 35\u201339, Berlin, 1985. <br \/>\nSpringer-Verlag.<br \/>\n[5] Bertram Ludascher et al. Scienti c work ow manage- ment and the Kepler system. Concurrency and Com-<br \/>\nputation: Practice &amp; Experience , 18(10):1039\u20131065,<br \/>\n2006.<br \/>\n[6] Tom Oinn et al. Taverna: Lessons in creating a work ow environment for the life sciences. Con-<br \/>\ncurrency and Computation: Practice and Experience ,<br \/>\n18(10):1067\u20131100, August 2006. Special Issue on <br \/>\nGrid Work ow. [7] U Radetzki et al. Adapters, shims, and glue\u2013service<br \/>\ninteroperability for in silico experiments. Bioinfor-<br \/>\nmatics , 22(9):1137\u201343, May 2006.<br \/>\n[8] Ian Foster and Carl Kesselman. Globus: A metacom- puting infrastructure toolkit. The International Jour-<br \/>\nnal of Supercomputer Applications and High Perfor-<br \/>\nmance Computing , 11(2):115\u2013128, Summer 1997.<br \/>\n[9] Ian Foster, Jens Vockler, Michael Wilde, and Yong Zhao. Chimera: A virtual data system for represent-<br \/>\ning, querying, and automating data derivation. In 14th<br \/>\nInternational Conference on Scienti c and Statistical<br \/>\nDatabase Management (SSDBM\u201902) , 2002.<br \/>\n[10] Duncan Hull, Robert Stevens, Phillip Lord, Chris Wroe, and Carole Goble. Treating shimantic web syn- <br \/>\ndrome with ontologies. In First Advanced Knowl-<br \/>\nedge Technologies workshop on Semantic Web Ser- <br \/>\nvices (AKT-SWS04) . KMi, The Open University, Mil-<br \/>\nton Keynes, UK, 2004. (See Workshop proceedings<br \/>\nCEUR-WS.org (ISSN:16130073) Volume 122 - AKT- SWS04).<br \/>\n[11] Simon L. Peyton Jones. The Implementation of Func-<br \/>\ntional Programming Languages . Prentice Hall, 1987.<br \/>\n[12] Peter M. Kelly, Paul D. Coddington, and Andrew L. Wendelborn. Distributed, parallel web service orches- <br \/>\ntration using XSLT. In 1st IEEE International Confer-<br \/>\nence on e-Science and Grid Computing , Melbourne,<br \/>\nAustralia, December 2005.<br \/>\n[13] Peter M. Kelly, Paul D. Coddington, and Andrew L. Wendelborn. A distributed virtual machine for paral- <br \/>\nlel graph reduction. In 8th International Conference<br \/>\non Parallel and Distributed Computing Applications <br \/>\nand Technologies (PDCAT \u201907) , Adelaide, Australia,<br \/>\nDecember 2007.<br \/>\n[14] B. Lud\u00e4scher and I. Altintas. On providing declara- tive design and programming constructs for scienti c<br \/>\nwork ows based on process networks. Technical Note<br \/>\nSciDAC-SPA-TN-2003-01, 2003.<br \/>\n[15] Douglas Thain, Todd Tannenbaum, and Miron Livny. Distributed computing in practice: The Condor expe-rience. Concurrency and Computation: Practice and<br \/>\nExperience , 2004.<br \/>\n[16] P. W. Trinder, K. Hammond, H.-W. Loidl, and S. L. Peyton Jones. Algorithm + strategy = paral-lelism. J. Funct. Program. , 8(1):23\u201360, 1998.<br \/>\n[17] Jia Yu and Rajkumar Buyya. A taxonomy of work ow management systems for grid computing. Journal of<br \/>\nGrid Computing , 3(3\u20134):171\u2013200, September 2005.<br \/>\n8","created_at":"2018-11-07T16:21:52+0000","updated_at":"2018-11-07T16:21:52+0000","published_at":"2008-03-05T12:12:54+0000","annotations":[],"mimetype":"application\/pdf","reading_time":29,"domain_name":"www.pmkelly.net","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":21,"title":"How Lisp Became God's Own Programming Language","url":"https:\/\/twobithistory.org\/2018\/10\/14\/lisp.html","content":"<p>When programmers discuss the relative merits of different programming\nlanguages, they often talk about them in prosaic terms as if they were so many\ntools in a tool belt\u2014one might be more appropriate for systems programming,\nanother might be more appropriate for gluing together other programs to\naccomplish some ad hoc task. This is as it should be. Languages have different\nstrengths and claiming that a language is better than other languages without\nreference to a specific use case only invites an unproductive and vitriolic\ndebate. \n<\/p><p>But there is one language that seems to inspire a peculiar universal reverence:\nLisp. Keyboard crusaders that would otherwise pounce on anyone daring to\nsuggest that some language is better than any other will concede that Lisp is\non another level. Lisp transcends the utilitarian criteria used to judge other\nlanguages, because the median programmer has never used Lisp to build anything\npractical and probably never will, yet the reverence for Lisp runs so deep that\nLisp is often ascribed mystical properties. Everyone\u2019s favorite webcomic,\n<em>xkcd<\/em>, has depicted Lisp this way at least twice: In <a href=\"https:\/\/xkcd.com\/224\/\">one\ncomic<\/a>, a character reaches some sort of Lisp\nenlightenment, which appears to allow him to comprehend the fundamental\nstructure of the universe. In <a href=\"https:\/\/xkcd.com\/297\/\">another comic<\/a>, a robed,\nsenescent programmer hands a stack of parentheses to his padawan, saying that\nthe parentheses are \u201celegant weapons for a more civilized age,\u201d suggesting that\nLisp has all the occult power of the Force.<\/p><p>Another great example is Bob Kanefsky\u2019s parody of a\nsong called \u201cGod Lives on Terra.\u201d His parody, written in the mid-1990s and\ncalled \u201cEternal Flame\u201d, describes how God must have created the world using\nLisp. The following is an excerpt, but the full set of lyrics can be found in\nthe <a href=\"https:\/\/www.gnu.org\/fun\/jokes\/eternal-flame.en.html\">GNU Humor\nCollection<\/a>:<\/p><blockquote>\n  <p><em>For God wrote in Lisp code<br \/>When he filled the leaves with green.<br \/>The fractal flowers and recursive roots:<br \/>The most lovely hack I\u2019ve seen.<br \/>And when I ponder snowflakes,<br \/>never finding two the same,<br \/>I know God likes a language<br \/>with its own four-letter name.<\/em><\/p>\n<\/blockquote><p>I can only speak for myself, I suppose, but I think this \u201cLisp Is Arcane Magic\u201d\ncultural meme is the most bizarre and fascinating thing ever. Lisp was\nconcocted in the ivory tower as a tool for artificial intelligence research, so\nit was always going to be unfamiliar and maybe even a bit mysterious to the\nprogramming laity. But programmers now <a href=\"https:\/\/www.reddit.com\/r\/ProgrammerHumor\/comments\/5c14o6\/xkcd_lisp\/d9szjnc\/\">urge each other to \u201ctry Lisp before you\ndie\u201d<\/a>\nas if it were some kind of mind-expanding psychedelic. They do this even though\nLisp is now the second-oldest programming language in widespread use, younger\nonly than Fortran, and even then by just one year.<sup id=\"fnref:1\"><a href=\"#fn:1\" class=\"footnote\">1<\/a><\/sup> Imagine if your job were\nto promote some new programming language on behalf of the organization or team\nthat created it. Wouldn\u2019t it be great if you could convince everyone that your\nnew language had divine powers? But how would you even do that? How does a\nprogramming language come to be known as a font of hidden knowledge?<\/p><p>How did Lisp get to be this way?<\/p><p><img src=\"https:\/\/twobithistory.org\/images\/byte_lisp.jpg\" alt=\"Byte Magazine Cover, August, 1979.\" \/><em>The cover of Byte Magazine, August, 1979.<\/em><\/p><h2 id=\"theory-a-the-axiomatic-language\">Theory A: The Axiomatic Language<\/h2><p>John McCarthy, Lisp\u2019s creator, did not originally intend for Lisp to be an\nelegant distillation of the principles of computation. But, after one or two\nfortunate insights and a series of refinements, that\u2019s what Lisp became. Paul\nGraham\u2014we will talk about him some more later\u2014has written that, with Lisp,\nMcCarthy \u201cdid for programming something like what Euclid did for geometry.\u201d<sup id=\"fnref:2\"><a href=\"#fn:2\" class=\"footnote\">2<\/a><\/sup>\nPeople might see a deeper meaning in Lisp because McCarthy built Lisp out of\nparts so fundamental that it is hard to say whether he invented it or\ndiscovered it.<\/p><p>McCarthy began thinking about creating a language during the 1956 Darthmouth\nSummer Research Project on Artificial Intelligence. The Summer Research Project\nwas in effect an ongoing, multi-week academic conference, the very first in the\nfield of artificial intelligence. McCarthy, then an assistant professor of\nMathematics at Dartmouth, had actually coined the term \u201cartificial\nintelligence\u201d when he proposed the event.<sup id=\"fnref:3\"><a href=\"#fn:3\" class=\"footnote\">3<\/a><\/sup> About ten or so people attended\nthe conference for its entire duration.<sup id=\"fnref:4\"><a href=\"#fn:4\" class=\"footnote\">4<\/a><\/sup> Among them were Allen Newell and\nHerbert Simon, two researchers affiliated with the RAND Corporation and\nCarnegie Mellon that had just designed a language called IPL.<\/p><p>Newell and Simon had been trying to build a system capable of generating proofs\nin propositional calculus. They realized that it would be hard to do this while\nworking at the level of the computer\u2019s native instruction set, so they decided\nto create a language\u2014or, as they called it, a \u201cpseudo-code\u201d\u2014that would help\nthem more naturally express the workings of their \u201cLogic Theory Machine.\u201d<sup id=\"fnref:5\"><a href=\"#fn:5\" class=\"footnote\">5<\/a><\/sup>\nTheir language, called IPL for \u201cInformation Processing Language\u201d, was more of a\nhigh-level assembly dialect then a programming language in the sense we mean\ntoday. Newell and Simon, perhaps referring to Fortran, noted that other\n\u201cpseudo-codes\u201d then in development were \u201cpreoccupied\u201d with representing\nequations in standard mathematical notation.<sup id=\"fnref:6\"><a href=\"#fn:6\" class=\"footnote\">6<\/a><\/sup> Their language focused instead\non representing sentences in propositional calculus as lists of symbolic\nexpressions. Programs in IPL would basically leverage a series of\nassembly-language macros to manipulate and evaluate expressions within one or\nmore of these lists.<\/p><p>McCarthy thought that having algebraic expressions in a language,\nFortran-style, would be useful. So he didn\u2019t like IPL very much.<sup id=\"fnref:7\"><a href=\"#fn:7\" class=\"footnote\">7<\/a><\/sup> But he\nthought that symbolic lists were a good way to model problems in artificial\nintelligence, particularly problems involving deduction. This was the germ of\nMcCarthy\u2019s desire to create an algebraic list processing language, a language\nthat would resemble Fortran but also be able to process symbolic lists like\nIPL.<\/p><p>Of course, Lisp today does not resemble Fortran. Over the next few years,\nMcCarthy\u2019s ideas about what an ideal list processing language should look like\nevolved. His ideas began to change in 1957, when he started writing routines\nfor a chess-playing program in Fortran. The prolonged exposure to Fortran\nconvinced McCarthy that there were several infelicities in its design, chief\namong them the awkward <code class=\"highlighter-rouge\">IF<\/code> statement.<sup id=\"fnref:8\"><a href=\"#fn:8\" class=\"footnote\">8<\/a><\/sup> McCarthy invented an alternative,\nthe \u201ctrue\u201d conditional expression, which returns sub-expression A if the\nsupplied test succeeds and sub-expression B if the supplied test fails and\nwhich <em>also<\/em> only evaluates the sub-expression that actually gets returned.\nDuring the summer of 1958, when McCarthy worked to design a program that could\nperform differentiation, he realized that his \u201ctrue\u201d conditional expression\nmade writing recursive functions easier and more natural.<sup id=\"fnref:9\"><a href=\"#fn:9\" class=\"footnote\">9<\/a><\/sup> The\ndifferentiation problem also prompted McCarthy to devise the <em>maplist<\/em>\nfunction, which takes another function as an argument and applies it to all the\nelements in a list.<sup id=\"fnref:10\"><a href=\"#fn:10\" class=\"footnote\">10<\/a><\/sup> This was useful for differentiating sums of\narbitrarily many terms.<\/p><p>None of these things could be expressed in Fortran, so, in the fall of 1958,\nMcCarthy set some students to work implementing Lisp. Since McCarthy was now an\nassistant professor at MIT, these were all MIT students. As McCarthy and his\nstudents translated his ideas into running code, they made changes that further\nsimplified the language. The biggest change involved Lisp\u2019s syntax. McCarthy\nhad originally intended for the language to include something called\n\u201cM-expressions,\u201d which would be a layer of syntactic sugar that made Lisp\u2019s\nsyntax resemble Fortran\u2019s. Though M-expressions could be translated to\nS-expressions\u2014the basic lists enclosed by parentheses that Lisp is known for\u2014\nS-expressions were really a low-level representation meant for the machine. The\nonly problem was that McCarthy had been denoting M-expressions using square\nbrackets, and the IBM 026 keypunch that McCarthy\u2019s team used at MIT did not\nhave any square bracket keys on its keyboard.<sup id=\"fnref:11\"><a href=\"#fn:11\" class=\"footnote\">11<\/a><\/sup> So the Lisp team stuck with\nS-expressions, using them to represent not just lists of data but function\napplications too. McCarthy and his students also made a few other\nsimplifications, including a switch to prefix notation and a memory model\nchange that meant the language only had one real type.<sup id=\"fnref:12\"><a href=\"#fn:12\" class=\"footnote\">12<\/a><\/sup><\/p><p>In 1960, McCarthy published his famous paper on Lisp called \u201cRecursive\nFunctions of Symbolic Expressions and Their Computation by Machine.\u201d By that\ntime, the language had been pared down to such a degree that McCarthy realized\nhe had the makings of \u201can elegant mathematical system\u201d and not just another\nprogramming language.<sup id=\"fnref:13\"><a href=\"#fn:13\" class=\"footnote\">13<\/a><\/sup> He later wrote that the many simplifications that\nhad been made to Lisp turned it \u201cinto a way of describing computable functions\nmuch neater than the Turing machines or the general recursive definitions used\nin recursive function theory.\u201d<sup id=\"fnref:14\"><a href=\"#fn:14\" class=\"footnote\">14<\/a><\/sup> In his paper, he therefore presented Lisp\nboth as a working programming language and as a formalism for studying the\nbehavior of recursive functions.<\/p><p>McCarthy explained Lisp to his readers by building it up out of only a very\nsmall collection of rules. Paul Graham later retraced McCarthy\u2019s steps, using\nmore readable language, in his essay <a href=\"http:\/\/languagelog.ldc.upenn.edu\/myl\/llog\/jmc.pdf\">\u201cThe Roots of\nLisp\u201d<\/a>. Graham is able to\nexplain Lisp using only seven primitive operators, two different notations for\nfunctions, and a half-dozen higher-level functions defined in terms of the\nprimitive operators. That Lisp can be specified by such a small sequence of\nbasic rules no doubt contributes to its mystique. Graham has called McCarthy\u2019s\npaper an attempt to \u201caxiomatize computation.\u201d<sup id=\"fnref:15\"><a href=\"#fn:15\" class=\"footnote\">15<\/a><\/sup> I think that is a great way\nto think about Lisp\u2019s appeal. Whereas other languages have clearly artificial\nconstructs denoted by reserved words like <code class=\"highlighter-rouge\">while<\/code> or <code class=\"highlighter-rouge\">typedef<\/code> or <code class=\"highlighter-rouge\">public\nstatic void<\/code>, Lisp\u2019s design almost seems entailed by the very logic of\ncomputing. This quality and Lisp\u2019s original connection to a field as esoteric\nas \u201crecursive function theory\u201d should make it no surprise that Lisp has so much\nprestige today.<\/p><h2 id=\"theory-b-machine-of-the-future\">Theory B: Machine of the Future<\/h2><p>Two decades after its creation, Lisp had become, according to the famous\n<a href=\"https:\/\/en.wikipedia.org\/wiki\/Jargon_File\"><em>Hacker\u2019s Dictionary<\/em><\/a>, the \u201cmother\ntongue\u201d of artificial intelligence research. Early on, Lisp spread quickly,\nprobably because its regular syntax made implementing it on new machines\nrelatively straightforward. Later, researchers would keep using it because of\nhow well it handled symbolic expressions, important in an era when so much of\nartificial intelligence was symbolic. Lisp was used in seminal artificial\nintelligence projects like the <a href=\"https:\/\/hci.stanford.edu\/winograd\/shrdlu\/\">SHRDLU natural language\nprogram<\/a>, the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Macsyma\">Macsyma algebra\nsystem<\/a>, and the <a href=\"https:\/\/en.wikipedia.org\/wiki\/ACL2\">ACL2 logic\nsystem<\/a>.<\/p><p>By the mid-1970s, though, artificial intelligence researchers were running out\nof computer power. The PDP-10, in particular\u2014everyone\u2019s favorite machine for\nartificial intelligence work\u2014had an 18-bit address space that increasingly was\ninsufficient for Lisp AI programs.<sup id=\"fnref:16\"><a href=\"#fn:16\" class=\"footnote\">16<\/a><\/sup> Many AI programs were also supposed to\nbe interactive, and making a demanding interactive program perform well on a\ntime-sharing system was challenging. The solution, originally proposed by Peter\nDeutsch at MIT, was to engineer a computer specifically designed to run Lisp\nprograms. These Lisp machines, as I described in <a href=\"https:\/\/twobithistory.org\/2018\/09\/30\/chaosnet.html\">my last post on Chaosnet<\/a>, would give each user a dedicated\nprocessor optimized for Lisp. They would also eventually come with development\nenvironments written entirely in Lisp for hardcore Lisp programmers. Lisp\nmachines, devised in an awkward moment at the tail of the minicomputer era but\nbefore the full flowering of the microcomputer revolution, were\nhigh-performance personal computers for the programming elite.<\/p><p>For a while, it seemed as if Lisp machines would be the wave of the future.\nSeveral companies sprang into existence and raced to commercialize the\ntechnology. The most successful of these companies was called Symbolics,\nfounded by veterans of the MIT AI Lab. Throughout the 1980s, Symbolics produced\na line of computers known as the 3600 series, which were popular in the AI\nfield and in industries requiring high-powered computing. The 3600 series\ncomputers featured large screens, bit-mapped graphics, a mouse interface, and\n<a href=\"https:\/\/youtu.be\/gV5obrYaogU?t=201\">powerful graphics and animation software<\/a>.\nThese were impressive machines that enabled impressive programs. For example,\nBob Culley, who worked in robotics research and contacted me via Twitter, was\nable to implement and visualize a path-finding algorithm on a Symbolics 3650 \nin 1985. He explained to me that bit-mapped graphics and object-oriented\nprogramming (available on Lisp machines via <a href=\"https:\/\/en.wikipedia.org\/wiki\/Flavors_(programming_language)\">the Flavors\nextension<\/a>) were\nvery new in the 1980s. Symbolics was the cutting edge.<\/p><p><img src=\"https:\/\/twobithistory.org\/images\/symbolics.jpg\" alt=\"Bob Culley's path-finding program.\" \/><em>Bob Culley\u2019s path-finding program.<\/em><\/p><p>As a result, Symbolics machines were outrageously expensive. The Symbolics 3600\ncost $110,000 in 1983.<sup id=\"fnref:16:1\"><a href=\"#fn:16\" class=\"footnote\">16<\/a><\/sup> So most people could only marvel at the power of\nLisp machines and the wizardry of their Lisp-writing operators from afar. But\nmarvel they did. <em>Byte Magazine<\/em> featured Lisp and Lisp machines several times\nfrom 1979 through to the end of the 1980s. In the August, 1979 issue, a special\non Lisp, the magazine\u2019s editor raved about the new machines being developed at\nMIT with \u201cgobs of memory\u201d and \u201can advanced operating system.\u201d<sup id=\"fnref:17\"><a href=\"#fn:17\" class=\"footnote\">17<\/a><\/sup> He thought\nthey sounded so promising that they would make the two prior years\u2014which saw\nthe launch of the Apple II, the Commodore PET, and the TRS-80\u2014look boring by\ncomparison. A half decade later, in 1985, a <em>Byte Magazine<\/em> contributor\ndescribed writing Lisp programs for the \u201csophisticated, superpowerful Symbolics\n3670\u201d and urged his audience to learn Lisp, claiming it was both \u201cthe language\nof choice for most people working in AI\u201d and soon to be a general-purpose\nprogramming language as well.<sup id=\"fnref:18\"><a href=\"#fn:18\" class=\"footnote\">18<\/a><\/sup><\/p><p>I asked Paul McJones, who has done lots of Lisp <a href=\"http:\/\/www.softwarepreservation.org\/projects\/LISP\/\">preservation\nwork<\/a> for the Computer\nHistory Museum in Mountain View, about when people first began talking about\nLisp as if it were a gift from higher-dimensional beings. He said that the\ninherent properties of the language no doubt had a lot to do with it, but he\nalso said that the close association between Lisp and the powerful artificial\nintelligence applications of the 1960s and 1970s probably contributed too. When\nLisp machines became available for purchase in the 1980s, a few more people\noutside of places like MIT and Stanford were exposed to Lisp\u2019s power and the\nlegend grew. Today, Lisp machines and Symbolics are little remembered, but they\nhelped keep the mystique of Lisp alive through to the late 1980s.<\/p><h2 id=\"theory-c-learn-to-program\">Theory C: Learn to Program<\/h2><p>In 1985, MIT professors Harold Abelson and Gerald Sussman, along with Sussman\u2019s\nwife, Julie Sussman, published a textbook called <em>Structure and Interpretation\nof Computer Programs<\/em>. The textbook introduced readers to programming using the\nlanguage Scheme, a dialect of Lisp. It was used to teach MIT\u2019s introductory\nprogramming class for two decades. My hunch is that SICP (as the title is\ncommonly abbreviated) about doubled Lisp\u2019s \u201cmystique factor.\u201d SICP took Lisp\nand showed how it could be used to illustrate deep, almost philosophical\nconcepts in the art of computer programming. Those concepts were general enough\nthat any language could have been used, but SICP\u2019s authors chose Lisp. As a\nresult, Lisp\u2019s reputation was augmented by the notoriety of this bizarre and\nbrilliant book, which has intrigued generations of programmers (and also become\n<a href=\"https:\/\/knowyourmeme.com\/forums\/meme-research\/topics\/47038-structure-and-interpretation-of-computer-programs-hugeass-image-dump-for-evidence\">a very strange\nmeme<\/a>).\nLisp had always been \u201cMcCarthy\u2019s elegant formalism\u201d; now it was also \u201cthat\nlanguage that teaches you the hidden secrets of programming.\u201d<\/p><p>It\u2019s worth dwelling for a while on how weird SICP really is, because I think\nthe book\u2019s weirdness and Lisp\u2019s weirdness get conflated today. The weirdness\nstarts with the book\u2019s cover. It depicts a wizard or alchemist approaching a\ntable, prepared to perform some sort of sorcery. In one hand he holds a set of\ncalipers or a compass, in the other he holds a globe inscribed with the words\n\u201ceval\u201d and \u201capply.\u201d A woman opposite him gestures at the table; in the\nbackground, the Greek letter lambda floats in mid-air, radiating light.<\/p><p><img src=\"https:\/\/twobithistory.org\/images\/sicp.jpg\" alt=\"The cover art for SICP.\" \/><em>The cover art for SICP.<\/em><\/p><p>Honestly, what is going on here? Why does the table have animal feet? Why is\nthe woman gesturing at the table? What is the significance of the inkwell? Are\nwe supposed to conclude that the wizard has unlocked the hidden mysteries of\nthe universe, and that those mysteries consist of the \u201ceval\/apply\u201d loop and\nthe Lambda Calculus? It would seem so. This image alone must have done an\nenormous amount to shape how people talk about Lisp today.<\/p><p>But the text of the book itself is often just as weird. SICP is unlike most\nother computer science textbooks that you have ever read. Its authors explain\nin the foreword to the book that the book is not merely about how to program in\nLisp\u2014it is instead about \u201cthree foci of phenomena: the human mind, collections\nof computer programs, and the computer.\u201d<sup id=\"fnref:19\"><a href=\"#fn:19\" class=\"footnote\">19<\/a><\/sup> Later, they elaborate, describing\ntheir conviction that programming shouldn\u2019t be considered a discipline of\ncomputer science but instead should be considered a new notation for \u201cprocedural\nepistemology.\u201d<sup id=\"fnref:20\"><a href=\"#fn:20\" class=\"footnote\">20<\/a><\/sup> Programs are a new way of structuring thought that only\nincidentally get fed into computers. The first chapter of the book gives a\nbrief tour of Lisp, but most of the book after that point is about much more\nabstract concepts. There is a discussion of different programming paradigms, a\ndiscussion of the nature of \u201ctime\u201d and \u201cidentity\u201d in object-oriented systems,\nand at one point a discussion of how synchronization problems may arise because\nof fundamental constraints on communication that play a role akin to the fixed\nspeed of light in the theory of relativity.<sup id=\"fnref:21\"><a href=\"#fn:21\" class=\"footnote\">21<\/a><\/sup> It\u2019s heady stuff.<\/p><p>All this isn\u2019t to say that the book is bad. It\u2019s a wonderful book. It discusses\nimportant programming concepts at a higher level than anything else I have\nread, concepts that I had long wondered about but didn\u2019t quite have the\nlanguage to describe. It\u2019s impressive that an introductory programming textbook\ncan move so quickly to describing the fundamental shortfalls of object-oriented\nprogramming and the benefits of functional languages that minimize mutable\nstate. It\u2019s mind-blowing that this then turns into a discussion of how a stream\nparadigm, perhaps something like today\u2019s\n<a href=\"https:\/\/rxjs-dev.firebaseapp.com\/\">RxJS<\/a>, can give you the best of both\nworlds. SICP distills the essence of high-level program design in a way\nreminiscent of McCarthy\u2019s original Lisp paper. The first thing you want to do\nafter reading it is get your programmer friends to read it; if they look it\nup, see the cover, but then don\u2019t read it, all they take away is that some\nmysterious, fundamental \u201ceval\/apply\u201d thing gives magicians special powers over\ntables with animal feet. I would be deeply impressed in their shoes too.<\/p><p>But maybe SICP\u2019s most important contribution was to elevate Lisp from\ncurious oddity to pedagogical must-have. Well before SICP, people told each\nother to learn Lisp as a way of getting better at programming. The 1979 Lisp\nissue of <em>Byte Magazine<\/em> is testament to that fact. The same editor that raved\nabout MIT\u2019s new Lisp machines also explained that the language was worth\nlearning because it \u201crepresents a different point of view from which to analyze\nproblems.\u201d<sup id=\"fnref:22\"><a href=\"#fn:22\" class=\"footnote\">22<\/a><\/sup> But SICP presented Lisp as more than just a foil for other\nlanguages; SICP used Lisp as an <em>introductory<\/em> language, implicitly making the\nargument that Lisp is the best language in which to grasp the fundamentals of\ncomputer programming. When programmers today tell each other to try Lisp before\nthey die, they arguably do so in large part because of SICP. After all, the\nlanguage <a href=\"https:\/\/en.wikipedia.org\/wiki\/Brainfuck\">Brainfuck<\/a> presumably offers\n\u201ca different point of view from which to analyze problems.\u201d But people learn\nLisp instead because they know that, for twenty years or so, the Lisp point of\nview was thought to be so useful that MIT taught Lisp to undergraduates before\nanything else.<\/p><h2 id=\"lisp-comes-back\">Lisp Comes Back<\/h2><p>The same year that SICP was released, Bjarne Stroustrup published the first\nedition of <em>The C++ Programming Language<\/em>, which brought object-oriented\nprogramming to the masses. A few years later, the market for Lisp machines\ncollapsed and the AI winter began. For the next decade and change, C++ and then\nJava would be the languages of the future and Lisp would be left out in the\ncold.<\/p><p>It is of course impossible to pinpoint when people started getting excited\nabout Lisp again. But that may have happened after Paul Graham, Y-Combinator\nco-founder and Hacker News creator, published a series of influential essays\npushing Lisp as the best language for startups. In his essay <a href=\"http:\/\/www.paulgraham.com\/avg.html\">\u201cBeating the\nAverages,\u201d<\/a> for example, Graham argued that\nLisp macros simply made Lisp more powerful than other languages. He claimed\nthat using Lisp at his own startup, Viaweb, helped him develop features faster\nthan his competitors were able to. <a href=\"https:\/\/web.archive.org\/web\/20061004035628\/http:\/\/wiki.alu.org\/Chris-Perkins\">Some programmers at\nleast<\/a>\nwere persuaded. But the vast majority of programmers did not switch to Lisp.<\/p><p>What happened instead is that more and more Lisp-y features have been\nincorporated into everyone\u2019s favorite programming languages. Python got list\ncomprehensions. C# got Linq. Ruby got\u2026 well, Ruby <a href=\"http:\/\/www.randomhacks.net\/2005\/12\/03\/why-ruby-is-an-acceptable-lisp\/\">is a\nLisp<\/a>.\nAs Graham noted even back in 2001, \u201cthe default language, embodied in a\nsuccession of popular languages, has gradually evolved toward Lisp.\u201d<sup id=\"fnref:23\"><a href=\"#fn:23\" class=\"footnote\">23<\/a><\/sup> \nThough other languages are gradually becoming like Lisp, Lisp itself somehow\nmanages to retain its special reputation as that mysterious language that few\npeople understand but everybody should learn. In 1980, on the occasion of\nLisp\u2019s 20th anniversary, McCarthy wrote that Lisp had survived as long as it\nhad because it occupied \u201csome kind of approximate local optimum in the space of\nprogramming languages.\u201d<sup id=\"fnref:24\"><a href=\"#fn:24\" class=\"footnote\">24<\/a><\/sup> That understates Lisp\u2019s real influence. Lisp\nhasn\u2019t survived for over half a century because programmers have begrudgingly\nconceded that it is the best tool for the job decade after decade; in fact, it\nhas survived even though most programmers do not use it at all. Thanks to its\norigins and use in artificial intelligence research and perhaps also the legacy\nof SICP, Lisp continues to fascinate people. Until we can imagine God creating\nthe world with some newer language, Lisp isn\u2019t going anywhere.<\/p><p><em>\nIf you enjoyed this post, more like it come out every two weeks! Follow\n<a href=\"https:\/\/twitter.com\/TwoBitHistory\">\n  @TwoBitHistory\n<\/a> on Twitter or subscribe to the\n<a href=\"https:\/\/twobithistory.org\/feed.xml\">\n  RSS feed\n<\/a>\nto make sure you know when a new post is out.\n<\/em><\/p><p><em>Previously on TwoBitHistory\u2026<\/em><\/p><div class=\"footnotes\"><ol><li id=\"fn:1\">\n      <p>John McCarthy, \u201cHistory of Lisp\u201d, 14, Stanford University, February 12, 1979, accessed October 14, 2018, <a href=\"http:\/\/jmc.stanford.edu\/articles\/lisp\/lisp.pdf\">http:\/\/jmc.stanford.edu\/articles\/lisp\/lisp.pdf<\/a>.\u00a0<a href=\"#fnref:1\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:2\">\n      <p>Paul Graham, \u201cThe Roots of Lisp\u201d, 1, January 18, 2002, accessed October 14, 2018, <a href=\"http:\/\/languagelog.ldc.upenn.edu\/myl\/llog\/jmc.pdf\">http:\/\/languagelog.ldc.upenn.edu\/myl\/llog\/jmc.pdf<\/a>.\u00a0<a href=\"#fnref:2\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:3\">\n      <p>Martin Childs, \u201cJohn McCarthy: Computer scientist known as the father of AI\u201d, The Independent, November 1, 2011, accessed on October 14, 2018, <a href=\"https:\/\/www.independent.co.uk\/news\/obituaries\/john-mccarthy-computer-scientist-known-as-the-father-of-ai-6255307.html\">https:\/\/www.independent.co.uk\/news\/obituaries\/john-mccarthy-computer-scientist-known-as-the-father-of-ai-6255307.html<\/a>.\u00a0<a href=\"#fnref:3\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:4\">\n      <p>Lisp Bulletin History. <a href=\"http:\/\/www.artinfo-musinfo.org\/scans\/lb\/lb3f.pdf\">http:\/\/www.artinfo-musinfo.org\/scans\/lb\/lb3f.pdf<\/a>\u00a0<a href=\"#fnref:4\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:5\">\n      <p>Allen Newell and Herbert Simon, \u201cCurrent Developments in Complex Information Processing,\u201d 19, May 1, 1956, accessed on October 14, 2018, <a href=\"http:\/\/bitsavers.org\/pdf\/rand\/ipl\/P-850_Current_Developments_In_Complex_Information_Processing_May56.pdf\">http:\/\/bitsavers.org\/pdf\/rand\/ipl\/P-850_Current_Developments_In_Complex_Information_Processing_May56.pdf<\/a>.\u00a0<a href=\"#fnref:5\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:6\">\n      <p>ibid.\u00a0<a href=\"#fnref:6\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:7\">\n      <p>Herbert Stoyan, \u201cLisp History\u201d, 43, Lisp Bulletin #3, December 1979, accessed on October 14, 2018, <a href=\"http:\/\/www.artinfo-musinfo.org\/scans\/lb\/lb3f.pdf\">http:\/\/www.artinfo-musinfo.org\/scans\/lb\/lb3f.pdf<\/a>\u00a0<a href=\"#fnref:7\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:8\">\n      <p>McCarthy, \u201cHistory of Lisp\u201d, 5.\u00a0<a href=\"#fnref:8\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:9\">\n      <p>ibid.\u00a0<a href=\"#fnref:9\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:10\">\n      <p>McCarthy \u201cHistory of Lisp\u201d, 6.\u00a0<a href=\"#fnref:10\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:11\">\n      <p>Stoyan, \u201cLisp History\u201d, 45\u00a0<a href=\"#fnref:11\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:12\">\n      <p>McCarthy, \u201cHistory of Lisp\u201d, 8.\u00a0<a href=\"#fnref:12\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:13\">\n      <p>McCarthy, \u201cHistory of Lisp\u201d, 2.\u00a0<a href=\"#fnref:13\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:14\">\n      <p>McCarthy, \u201cHistory of Lisp\u201d, 8.\u00a0<a href=\"#fnref:14\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:15\">\n      <p>Graham, \u201cThe Roots of Lisp\u201d, 11.\u00a0<a href=\"#fnref:15\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:16\">\n      <p>Guy Steele and Richard Gabriel, \u201cThe Evolution of Lisp\u201d, 22, History of Programming Languages 2, 1993, accessed on October 14, 2018, <a href=\"http:\/\/www.dreamsongs.com\/Files\/HOPL2-Uncut.pdf\">http:\/\/www.dreamsongs.com\/Files\/HOPL2-Uncut.pdf<\/a>.\u00a0<a href=\"#fnref:16\" class=\"reversefootnote\">\u21a9<\/a>\u00a0<a href=\"#fnref:16:1\" class=\"reversefootnote\">\u21a9<sup>2<\/sup><\/a><\/p>\n    <\/li>\n    <li id=\"fn:17\">\n      <p>Carl Helmers, \u201cEditorial\u201d, Byte Magazine, 154, August 1979, accessed on October 14, 2018, <a href=\"https:\/\/archive.org\/details\/byte-magazine-1979-08\/page\/n153\">https:\/\/archive.org\/details\/byte-magazine-1979-08\/page\/n153<\/a>.\u00a0<a href=\"#fnref:17\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:18\">\n      <p>Patrick Winston, \u201cThe Lisp Revolution\u201d, 209, April 1985, accessed on October 14, 2018, <a href=\"https:\/\/archive.org\/details\/byte-magazine-1985-04\/page\/n207\">https:\/\/archive.org\/details\/byte-magazine-1985-04\/page\/n207<\/a>.\u00a0<a href=\"#fnref:18\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:19\">\n      <p>Harold Abelson, Gerald Jay. Sussman, and Julie Sussman, Structure and Interpretation of Computer Programs (Cambridge, Mass: MIT Press, 2010), xiii.\u00a0<a href=\"#fnref:19\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:20\">\n      <p>Abelson, xxiii.\u00a0<a href=\"#fnref:20\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:21\">\n      <p>Abelson, 428.\u00a0<a href=\"#fnref:21\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:22\">\n      <p>Helmers, 7.\u00a0<a href=\"#fnref:22\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:23\">\n      <p>Paul Graham, \u201cWhat Made Lisp Different\u201d, December 2001, accessed on October 14, 2018, <a href=\"http:\/\/www.paulgraham.com\/diff.html\">http:\/\/www.paulgraham.com\/diff.html<\/a>.\u00a0<a href=\"#fnref:23\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n    <li id=\"fn:24\">\n      <p>John McCarthy, \u201cLisp\u2014Notes on its past and future\u201d, 3, Stanford University, 1980, accessed on October 14, 2018, <a href=\"http:\/\/jmc.stanford.edu\/articles\/lisp20th\/lisp20th.pdf\">http:\/\/jmc.stanford.edu\/articles\/lisp20th\/lisp20th.pdf<\/a>.\u00a0<a href=\"#fnref:24\" class=\"reversefootnote\">\u21a9<\/a><\/p>\n    <\/li>\n  <\/ol><\/div>","created_at":"2018-11-07T16:21:52+0000","updated_at":"2018-11-07T16:21:52+0000","annotations":[],"mimetype":"text\/html","reading_time":20,"domain_name":"twobithistory.org","preview_picture":"https:\/\/twobithistory.org\/images\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":22,"title":"13 Games in \u2264 13kB of JavaScript \u00b7 js13kGames 2018","url":"https:\/\/blog.github.com\/2018-10-05-js13kgames-highlights-2018\/","content":"<p>As the seventh annual <a href=\"http:\/\/js13kgames.com\/\">js13kGames competition<\/a> comes to a close, a grand total of 274 games were submitted. Even more impressive, each one was created in a single month, using less than 13 kB.<\/p><p>We rounded up a few of our favorites featuring a number of different styles and genres. From dark shooters and pixelated beat \u2018em ups to perplexing puzzle and platform games\u2014enjoy some downtime this weekend and play them all (or fork and hack on them with your own customizations)!<\/p><h2 id=\"underrun\">UNDERRUN<\/h2><p><a href=\"https:\/\/js13kgames.com\/entries\/underrun\">UNDERRUN<\/a> is a twin, stick shooter \u201cin 256 shades of brown,\u201d using webGL from <a href=\"https:\/\/github.com\/phoboslab\">@phoboslab<\/a>. In this game, you must defend yourself from predators while figuring out how to restore power to fix all system failures. Sounds simple enough, right? See for yourself when you play this highly-addictive shooter (and enjoy the haunting music). Read more about how the game was created in the <a href=\"https:\/\/phoboslab.org\/log\/2018\/09\/underrun-making-of\">retrospective<\/a>.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/underrun\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/phoboslab\/underrun\">View source <\/a><\/p><h2 id=\"envisionator\">Envisionator<\/h2><p><a href=\"https:\/\/github.com\/DennisBengs\">@DennisBengs<\/a> created the challenging puzzle game, <a href=\"https:\/\/js13kgames.com\/entries\/envisionator\">Envisionator<\/a>. The goal of the game is to escape a building on lockdown by giving a robot commands. What\u2019s the catch? The robot needs you to give it each and every direction, step by step\u2014one false move, and\u2026well, you\u2019ll see! Play Envisionater to see if you can escape.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/envisionator\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/DennisBengs\/envisionator\">View source<\/a><\/p><h2 id=\"onoff\">ONOFF<\/h2><p>Things aren\u2019t as black and white as they appear in <a href=\"https:\/\/js13kgames.com\/entries\/onoff\">ONOFF<\/a>. Dodge spikes, jump over pits, and toggle between dimensions. Think you can overcome each level of traps? You\u2019re in for a treat with this mind-boggling, fast-paced platformer from <a href=\"https:\/\/github.com\/starzonmyarmz\">@starzonmyarmz<\/a> and <a href=\"https:\/\/github.com\/braddunbar\">@braddunbar<\/a>. Play it to see what we mean!<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/onoff\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/starzonmyarmz\/js13k-2018\">View source<\/a><\/p><h2 id=\"the-chroma-incident\">The Chroma Incident<\/h2><p><a href=\"https:\/\/js13kgames.com\/entries\/the-chroma-incident\">The Chroma Incident<\/a> by <a href=\"https:\/\/github.com\/Rybar\">@Rybar<\/a> is also a twin, stick shooter but with a few more colors than UNDERRUN. The problem is the color\u2019s been stolen by the Achromats, and it\u2019s up to you to bring it back. Shoot your way through areas to reclaim those colors\u2014give it a go!<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/the-chroma-incident\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"http:\/\/github.com\/Rybar\/js13k2018\/\">View source<\/a><\/p><h2 id=\"the-matr13k\">The Matr13k<\/h2><p>Get nostalgic and relive some of the intense fight scenes with Neo from The Matrix. Use the arrow keys, <kbd>S<\/kbd> to kick and <kbd>D<\/kbd> punch your way through this JavaScript matrix from <a href=\"https:\/\/github.com\/agar3s\">@agar3s<\/a>. Can you find a way to the end of the rabbit hole before it\u2019s too late? Play The <a href=\"https:\/\/js13kgames.com\/entries\/the-matr13k\">Matri13k<\/a> and test your combat skills.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/the-matr13k\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/agar3s\/js13k2018\/\">View source<\/a><\/p><h2 id=\"1024-moves\">1024 moves<\/h2><p>Not to be confused with <a href=\"https:\/\/github.com\/gabrielecirulli\/2048\">2048<\/a>(!), <a href=\"https:\/\/js13kgames.com\/entries\/1024-moves\">1024 Moves<\/a> is a polished puzzle game from <a href=\"https:\/\/github.com\/GregPeck\">@GregPeck<\/a>. Get the ball, and avoid the holes\u2014what\u2019s the catch? See if you can solve the entire game in less than 1,024 moves. Play and test your problem-solving skills.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/1024-moves\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/GregPeck\/1024-moves\">View source<\/a><\/p><h2 id=\"geoquiz-2\">Geoquiz 2<\/h2><p>Think you know a little bit about world geography? Or are you lost with even the simplest of directions? Prove how much of a geography all-star you are by playing <a href=\"https:\/\/js13kgames.com\/entries\/geoquiz-2\">Geoquiz2<\/a>\u2014or brush up on your worldly knowledge. You can even read about how <a href=\"https:\/\/github.com\/xem\">@xem<\/a> made the game in the <a href=\"https:\/\/xem.github.io\/articles\/#js13k18\">GeoQuiz2 retrospective<\/a>.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/geoquiz-2\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/xem\/geoquiz2\">View source<\/a><\/p><h2 id=\"spacecraft\">Spacecraft<\/h2><p><a href=\"https:\/\/github.com\/tricsi\">@tricsi<\/a>\u2019s <a href=\"https:\/\/js13kgames.com\/entries\/spacecraft\">Spacecraft<\/a> challenges you to collect as many data tokens as possible from the planets and moons of the Solar System. It\u2019s easy\u2014until gravity accelerates your ship, and you have to avoid obstacles along the way in, \u201cspace, the final frontier.\u201d How far can you go before your probe goes offline? The only way to find out is to play on.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/spacecraft\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/tricsi\/spacecraft\">View source<\/a><\/p><h2 id=\"off-the-line\">Off the line<\/h2><p>How are your gaming reflexes? You\u2019ll quickly find out when you jump <a href=\"https:\/\/js13kgames.com\/entries\/off-the-line\">Off the Line<\/a> to collect coins in this arcade tapper from <a href=\"https:\/\/github.com\/regularkid\">@regularkid<\/a>. Take your time to figure out the best way to collect coins, or go crazy with a timed, ultra difficult ULTRA MEGA MODE (if you\u2019re feeling lucky). Play it and see how many coins you can collect.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/off-the-line\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/regularkid\/offtheline\">View source<\/a><\/p><h2 id=\"exo\">Exo<\/h2><p>You are the commander of a long-forgotten expedition to a distant star, and there are forces out to get you. Survive waves upon wave of enemies in <a href=\"https:\/\/js13kgames.com\/entries\/exo\">Exo<\/a>, a space-based tower defence game brought to you by <a href=\"https:\/\/github.com\/scorp200\">@scorp200<\/a>. Play Exo to unravel the story, arm your base, and reclaim your expedition.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/exo\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/scorp200\/td13k\">View source<\/a><\/p><h2 id=\"everyones-sky\">Everyone\u2019s Sky<\/h2><p>You are in control of your destiny in this space-based exploration game. Will you fight for the good of all or make enemies by being evil? Forge alliances, study star systems, fight against enemy combatants, and more in <a href=\"https:\/\/js13kgames.com\/entries\/everyones-sky\">Everyone\u2019s Sky<\/a> from <a href=\"https:\/\/github.com\/remvst\">@remvst<\/a>.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/everyones-sky\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/remvst\/everyones-sky\">View source<\/a><\/p><h2 id=\"submersible-warship-2063\">Submersible Warship 2063<\/h2><p>In <a href=\"https:\/\/github.com\/herebefrogs\">@herebefrogs<\/a>\u2019s <a href=\"https:\/\/js13kgames.com\/entries\/submersible-warship-2063\">Submersible Warship 2063<\/a>, enemy submarines are invading, fast. Make strategic use of your sonar to identify targets and evade torpedoes. Can you beat them before they beat you? Stay off enemy radar, and fight on by playing Submersible Warship 2063.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/submersible-warship-2063\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/herebefrogs\/submersible-warship-2063\">View source<\/a><\/p><h2 id=\"re-wire\">Re-wire<\/h2><p>If you enjoy playing high-stakes puzzles, <a href=\"https:\/\/js13kgames.com\/entries\/re-wire\">Re-wire<\/a> was made for you. Bring the system back online by rewiring power nodes, but watch out for the traps! This game from <a href=\"https:\/\/github.com\/JMankopf\">@JMankopf<\/a> will have you\u2026 wired to it for hours.<\/p><p><a href=\"https:\/\/js13kgames.com\/games\/re-wire\/index.html\"><strong>\u25ba Play in your browser<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/JMankopf\/js13k-rewire\">View source<\/a><\/p><p>This was such a difficult list to narrow down, as we enjoyed playing all of the JS13K entries. There are hundreds more to discover including a <a href=\"https:\/\/js13kgames.com\/entries\/bellwoods\">procedurally generated art game<\/a>, an <a href=\"https:\/\/js13kgames.com\/entries\/audio-dash\">audio surfing game<\/a>, and even a <a href=\"https:\/\/js13kgames.com\/entries\/js13k-battlegrounds\">13kB Battle Royale game<\/a>\u2014watch out PUBG and Fortnite!<\/p><p><a href=\"http:\/\/js13kgames.com\/#winners\">View this year\u2019s list of games from 13kGames<\/a><\/p><p>Do you have a favorite, a high score, or a fork of your own to share? We\u2019d love to hear about it! Let us know with the <a href=\"https:\/\/twitter.com\/hashtag\/js13k\"><code class=\"highlighter-rouge\">#js13k<\/code> hashtag<\/a>.<\/p><p>Thank you to everyone who participated, all of the judges, and especially <a href=\"https:\/\/github.com\/end3r\">@end3r<\/a> for running this great competition every year. Until next time! &lt;3<\/p>","created_at":"2018-11-07T16:21:52+0000","updated_at":"2018-11-07T16:21:52+0000","published_at":"2018-10-05T00:00:00+0000","published_by":["leereilly"],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":4,"domain_name":"blog.github.com","preview_picture":"https:\/\/user-images.githubusercontent.com\/121322\/46445380-75486500-c72b-11e8-9720-aed86a8ba44e.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":16,"title":"You Suck at Excel with Joel Spolsky","url":"http:\/\/www.youtube.com\/oembed?format=xml&url=https:\/\/www.youtube.com\/watch?v=0nbkaYsR94c","content":"<iframe id=\"video\" width=\"480\" height=\"270\" src=\"https:\/\/www.youtube.com\/embed\/0nbkaYsR94c?feature=oembed\" frameborder=\"0\" allowfullscreen=\"allowfullscreen\">[embedded content]<\/iframe>","created_at":"2018-11-07T16:21:51+0000","updated_at":"2018-11-07T16:21:51+0000","annotations":[],"mimetype":"text\/xml","reading_time":0,"domain_name":"www.youtube.com","preview_picture":"https:\/\/i.ytimg.com\/vi\/0nbkaYsR94c\/maxresdefault.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":14,"title":"Getting Started with GNUs","url":"https:\/\/www.bateman.io\/post\/setup-gnus\/","content":"<p>Emacs is quite the beast. It really isn\u2019t an editor so much as a operating system, and a pretty good one. It brings the promise of lisp machines onto every modern operating system and makes it quite easy to build a fantastic experience across systems.\n<\/p><p>Obviously, one of the more integral functions of such an operating system is email, and here, Emacs is a embarassment of riches. There are many avalible email clients in Emacs. Having tried Wunderlust and a few others, I found they didn\u2019t have much advantages over using a normal client other than org-mode integration. However, there is one that stands out from the rest: GNUs.<\/p>\n<p>Not only a competitent email client, GNUs provides a fantastic way to get through large RSS Feeds and mailing lists quickly, clearly designed by people who are ingesting a lot of information on the regular. It does this by providing a clean interface with almost no required user interation while providing some of the nice feature of a GUI. The simple nature of GNUs interface lets you quickly get through some of the more intense RSS feeds and mailing list. For example, now that I use GNUs, I simply skim all of the submissions to Hacker News right from the fire hse and it still takes less time than reading the first couple pages used to. But thats only one of the many improvements I have seen to productivity.<\/p>\n<p>One of the benifits I didn\u2019t anticipate is that when one uses w3m to render your emails, every image-only marketing email doesn\u2019t take any of your attention and you only get the gist from every email of a similar, text friendly nature. This leaves only the important mail left to actually read and sort through, meaning you spend less time looking at worthless marketing information and more time doing something else!<\/p>\n<p>Now here comes the hard part: GNUs configuration is documented in a very Emacs-y way. Many Emacs users much more versed than I will chuckle at that statement. What I mean is that the Emacs community is very intelligent but isn\u2019t super good at collessing information into one place and ensuring that information stays alive. This means some of the better resources are now just 404 responses and what is left is wiki articles with a ton of broken links. When I started learning Emacs, I was afraid it was because the community was dying. But what I have found since is that Emacs has never been doing better than now, but that most people either use a prepackaged set of extensions and configurations or are very tinkery and build whatever they can\u2019t have out of the box. This means that most of the documentation is right in the code and there aren\u2019t a lot of clear examples outside of random .emacs files on github that say \u201cwarning I don\u2019t know if this actually works.\u201d<\/p>\n<p>However, after sifting through a ton of crap, I have built a fairly effective GNUs configuration for my usage. With that said, email is different for everyone and I don\u2019t promise to have the one-true-config for GNUs that will be everything for everyone.<\/p>\n<p>The first thing I have learned, is that nnimap in GNUs is terrible. From what I have gathered myself, I think its a pretty bad idea to try and use it. It makes everything very slow to start and isn\u2019t what elisp is good for. But, there are tools that work with GNUs that are very good that for talking to IMAP servers. The one I decided on is mbsync (used to be known as isync). mbsync provides good support for IMAP and gmail specfic IMAP. It isn\u2019t perfect, but it is the fastest maildir compatible IMAP client and is also pretty reliable (something that I didn\u2019t fully appreciate as a nessesary feature during my short stint with the bug ridden offlineimap).<\/p>\n<p>Getting mbsync is non-obvious but simple on debian. One just installs isync and this installs mbsync and a legacy wrapper for it to act as isync. From there, you will want to download your email. This isn\u2019t instantanious but only took me 45 minutes or so with 50,000 emails but YMMV. Your primary hold up will likely be that Googles IMAP servers get upset if you make too many requests too quickly and will stop giving you data for a period of time. This doesn\u2019t happen regularly, but did happen to me multiple times when setting up my accounts. A good starting .mbsyncrc (living in your $HOME directory) is as follows:<\/p>\n<pre># Not gmail account\nIMAPAccount personal\nHost imap.not-google.com\nuser &lt;USERNAME&gt;\npass &lt;PASSWORD&gt;\nPipelineDepth 50\nIMAPStore personal-remote\nAccount personal\nMaildirStore personal-local\nPath ~\/.mail\/personal\nInbox ~\/.mail\/personal\/Inbox\nSubFolders Verbatim\nChannel personal\nMaster :personal-remote:\nSlave :personal-local:\nPatterns \"INBOX\" \"Other\" \"Drafts\"\nCreate Both\nSyncState *\n# Gmail Account\nIMAPAccount gmail\nHost imap.gmail.com\nuser &lt;USERNAME&gt;\npass &lt;PASSWORD&gt;\nSSLType IMAPS\nCertificateFile ~\/.cert\/imap.gmail.com.pem\nAuthMechs LOGIN\nPipelineDepth 50\nIMAPStore gmail-remote\nAccount gmail\nMaildirStore gmail-local\nPath ~\/.mail\/gmail\nInbox ~\/.mail\/gmail\/Inbox\nSubFolders Verbatim\nChannel gmail\nMaster :gmail-remote:\nSlave :gmail-local:\nPatterns *\nCreate Both\nSyncState *\n# More accounts ....\n<\/pre>\n<p>However, if you choose to use this as a long-term solution I recommend getting mbsync to get your password out of a .gpg (or a custom commandline password manager, which, btw, is probably better for ease of use), which it supports pretty simply. Just replace the pass line for each email with:<\/p>\n<pre>PassCmd \"gpg2 -q --for-your-eyes-only --no-tty -d ~\/.mailpass.gpg\"\n<\/pre>\n<p>Be sure to direct it to the proper file for each password. I haven\u2019t figured out yet how to put them all into a single .authinfo.gpg yet though, so if you find that out, please email me!<\/p>\n<p>From here, you should just be able to run:<\/p>\n<pre>mbsync -a\n<\/pre>\n<p>This should download all your mail from each account. You probably want to set up a recurring task with systemd to run this every 5 minutes or so. There is a great Arch Wiki page <a href=\"https:\/\/wiki.archlinux.org\/index.php\/isync#Automatic_synchronization\">here<\/a> describing it<\/p>\n<p>From here, we have our mail (or atleast most of it) local on our machine. Now we can start getting into the nitty gritty of reading it from GNUs. Like all great *nix utilities, it requires its own config file in your $HOME directory called .gnus.el.<\/p>\n<p>Because of the simplicity of how we set up our maildirs, the configuration isn\u2019t too bad. To get GNUs reading our email, all we need in our .gnus.el file is the following:<\/p>\n<pre>(setq user-mail-address    \"example@not-google.com\"\n      user-full-name       \"Your Name\")\n(setq gnus-select-method '(nnnil \"\"))\n(setq gnus-secondary-select-methods\n      '((nntp \"news.gwene.org\")\n        (nnmaildir \"personal\"\n\t\t   (directory \"~\/.mail\/personal\/\"))\n\t(nnmaildir \"gmail\"\n\t\t   (directory \"~\/.mail\/gmail\/\"))\n\t))\n(setq send-mail-function 'smtpmail-send-it)\n(setq smtpmail-smtp-server \"smtp.not-google.com\")\n(setq smtpmail-smtp-service 587)\n(setq mm-text-html-renderer 'gnus-w3m)\n(setq gnus-inhibit-images nil)\n<\/pre>\n<p>You can see that all we are doing is pointing GNUs to those folders and telling them they are maildirs. There is a certain beauty in the fact that there is a expert tool doing its job and all your mail client has to do is focus on providing email for you.<\/p>\n<p>You can add as many emails as you would like by the way to this set up, I personally have added 6 and they all work fine.<\/p>\n<p>Now that we have this file, we can run M-x gnus in Emacs and after a short loading period, it should come up with your email boxes as you have configured them!<\/p>\n<p>Now that that is all in order, you should be able to read your email in a simplistic fashion in gnus with org-mode links. Go to the first mailbox with unread mail, hit enter twice and then read that email and hit n to see the next email. If you reach the end of a mailbox, just hit n again and you will be taken to the next mailbox.<\/p>\n<p>This is awesome and fast, which is everything we wanted from email in emacs. What alludes to the power of GNUs is that this is the most basic getting started configuration, we can do better.<\/p>\n<p>One thing that I am personally stumped with is setting up is a mythical notmuch search functionality. Many have spoke of it on mailing-lists but the primary source config is a 404 these days unfortunately. With that said, I have found <a href=\"https:\/\/gist.github.com\/lchausmann\/9483876\">this gist<\/a> which has fantastic information but requires that All Mail be downloaded to work with Gmail which I have yet to figure out. Again, if you find anything, feel free to email me!<\/p>\n<p>From here, you probably want to add some RSS feeds. I like <a href=\"http:\/\/gwene.org\/\">http:\/\/gwene.org\/<\/a> for this task. We already set it up as as newsgroup provider when we setup our email, so all you need to do is open emacs (which lands you at what is called the Groups page) and from the Groups page, press ^. You should be given a list of newsgroups\/RSS feeds provided by Gwene. To check if a RSS feed is provided by them, go to their website and try and add it. If it is already added, they will give you the name of the newsgroup, which you can just search the group list server buffer (the one we opened with ^) for it and hit u when your cursor is on that line to subscribe to it. In addition, you can type U to bring up a prompt so you don\u2019t have to search the buffer for said group.<\/p>\n<p>From here, you should have a pretty slick, org-compatible setup that gets you through your email! If you have any questions or ideas, just email me at sam[at]bateman[dot]io!<\/p>","created_at":"2018-11-07T16:21:50+0000","updated_at":"2018-11-07T16:21:50+0000","published_at":"2018-06-18T23:52:15+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":8,"domain_name":"www.bateman.io","preview_picture":"https:\/\/source.unsplash.com\/category\/technology\/1600x900","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":15,"title":"What Is Pleroma?","url":"https:\/\/blog.soykaf.com\/post\/what-is-pleroma\/","content":"<p>In my last few posts I talked a bit about <a href=\"https:\/\/pleroma.social\">Pleroma<\/a>. Since then, we had a lot of questions about what Pleroma is, how it is different from GNU Social and Mastodon, why you should use it and so on. This post will be an introduction to Pleroma, so read on if you are interested.<\/p><h2 id=\"what-is-pleroma\">What is Pleroma?<\/h2>\n<p>Pleroma is a microblogging server software that can federate (= exchange messages with) other servers that support the same federation standards (OStatus and ActivityPub). What that means is that you can host a server for yourself or your friends and stay in control of your online identity, but still exchange messages with people on larger servers. Pleroma will federate with all servers that implement either OStatus or ActivityPub, like GNU Social, Friendica, Hubzilla and Mastodon.<\/p>\n<h2 id=\"how-does-it-work\">How does it work?<\/h2>\n<p>Here\u2019s a screenshot of the main interface.<\/p>\n<p><img src=\"https:\/\/blog.soykaf.com\/img\/pleroma-1.png\" class=\"fit\" alt=\"image\" \/><\/p>\n<p>When you are not logged in, you can see the Public Timeline and The Whole Known Network. The Public Timeline is every post that was made on that server. If you are looking for new people, it\u2019s often a good idea to look through the public timeline of other Pleroma servers to see what people there are writing. It will also help you figure out if a server is a good one to join.<\/p>\n<p>The Whole Known Network contains all local posts PLUS all the posts from other servers that reached this one. This means that it will contain all the posts from users that users on that server are following or communicating with. Looking through the whole known network will give you a good overview of the fediverse (our term for the network of servers).<\/p>\n<p><img src=\"https:\/\/blog.soykaf.com\/img\/pleroma-3.png\" class=\"fit\" alt=\"image\" \/><\/p>\n<p>When you are logged in, you have access to a few more features. You can post new messages, of course, which will be sent to your followers. Pleroma has a default maximum of 5000 characters per message, but this is just a suggestion. A server can set it to any value.<\/p>\n<p>Your main timeline is the one that contains all the people your are following. You can do the usual functions like favorite, repeat and reply from the timelines. You also get a Notifications section, which will contain notifications about posts you have been mentioned in or posts of yours that have been shared or favorited by other people.<\/p>\n<p>Pleroma also features a little chat in the sidebar. This is a real time chat that does not federate (for now). Every user on your instance is a member of this chatroom.<\/p>\n<p><img src=\"https:\/\/blog.soykaf.com\/img\/pleroma-4.png\" class=\"fit\" alt=\"image\" \/><\/p>\n<p>When you click on the little [+] symbol on the upper right, you will open an inline conversation view. This gives you a quick overview into conversation threads. You can do all the usual functions from inside that conversation view as well.<\/p>\n<p>Those are the basic features. I\u2019d encourage you to take a look at a Pleroma server to find out what else there is to see, like themes, custom backgrounds, attachment hiding and so on. Here are some servers that you could check out:<\/p>\n<ul><li><a href=\"https:\/\/pleroma.soykaf.com\">https:\/\/pleroma.soykaf.com<\/a><\/li>\n<li><a href=\"https:\/\/shigusegubu.club\">https:\/\/shigusegubu.club<\/a><\/li>\n<li><a href=\"https:\/\/social.sakamoto.gq\">https:\/\/social.sakamoto.gq<\/a><\/li>\n<li><a href=\"https:\/\/pleroma.iro-iro.xyz\">https:\/\/pleroma.iro-iro.xyz<\/a><\/li>\n<li><a href=\"https:\/\/social.netzkombinat.su\">https:\/\/social.netzkombinat.su<\/a><\/li>\n<\/ul><h2 id=\"how-is-it-different-from-mastodon\">How is it different from Mastodon?<\/h2>\n<p>If you are currently using Mastodon, you are probably interested in the differences between Mastodon and Pleroma. Here are some of the big ones.<\/p>\n<h3 id=\"lower-system-requirements\">Lower system requirements.<\/h3>\n<p>Pleroma can run well on a Raspberry Pi or a $2.50 Vultr instance. This makes it affordable to host for single user instances. You can still run a hundred users or so on instances this small, though, so it also works well for bigger instances.<\/p>\n<h3 id=\"less-moving-parts\">Less moving parts.<\/h3>\n<p>Pleroma is built on a lot less technology than Mastodon. To run a Mastodon instance, you need Rails, PostgreSQL, Redis, Sidekiq, NodeJS and - if you want search - ElasticSearch. For Pleroma, you only need Elixir and PostgreSQL, while still getting all the features. This simplifies installation and makes maintenance somewhat easier.<\/p>\n<h3 id=\"internally-activitypub\">Internally ActivityPub.<\/h3>\n<p>Mastodon is, internally, based around the concept of \u201cStatuses\u201d. It uses ActivityPub to federate, but it has no internal concept of Activities, it just generates \u2018fake\u2019 Activities for message exchange.<\/p>\n<p>Pleroma is ActivityPub even in its internal data structures. Activities are actually saved as JSON in the database, so the external and internal representations are the same.<\/p>\n<p>Why should you care? Because it makes it easier for Pleroma to add new Activity types. Adding a new Activity type in Pleroma involves no database changes, you just need to add some rules how to handle them. This will help us to add new features like groups or polls in the future.<\/p>\n<h3 id=\"compatible-with-mastodon-clients-including-the-mastodon-frontend\">Compatible with Mastodon clients (including the Mastodon frontend).<\/h3>\n<p>This is not really a difference, but Pleroma implements the Client-Server APIs of both GNU Social and Mastodon, so you can use most of their clients with Pleroma, too. For Mastodon clients, this includes Twidere, Tusky, Mastalab, Tootdon and many others.<\/p>\n<p>The biggest surprise for many is that Pleroma includes the Mastodon UI, too. As the Mastodon UI is mostly just another API client, it will run fine on Pleroma. As you can see in the screenshot, you can use it quite well on any Pleroma instance. We bundle it with Pleroma, so you don\u2019t need to install anything new.<\/p>\n<p><img class=\"fit\" src=\"https:\/\/blog.soykaf.com\/img\/pleroma-2.png\" alt=\"image\" \/><\/p>\n<p>I hope you learned some new things about Pleroma in this post. If you want to find out more, I\u2019d suggest just making an account on one of the instances I linked above and try it out a bit. You can also take a look at the <a href=\"https:\/\/pleroma.social\">homepage<\/a>, which also has links to the code, or chat with us at #pleroma on freenode.<\/p>\n<p>If you want to install your own instance, you can take a look at Jeff\u2019s post <a href=\"https:\/\/i2p.rocks\/blog\/the-magical-world-of-pleroma-setting-up-your-instance.html\">The magical world of Pleroma, setting up your instance<\/a>, where he describes the process step-by-step.<\/p>\n<p>Either way, have fun, and see you on the fediverse! :)<\/p>\n<p><img class=\"fit\" src=\"https:\/\/blog.soykaf.com\/img\/pleroma-5.png\" alt=\"image\" \/><\/p>","created_at":"2018-11-07T16:21:50+0000","updated_at":"2018-11-07T16:21:50+0000","published_at":"2018-03-07T10:28:10+0000","annotations":[],"mimetype":"text\/html","language":"en_US","reading_time":4,"domain_name":"blog.soykaf.com","preview_picture":"https:\/\/blog.soykaf.com\/img\/pleroma-5.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":12,"title":"Index","url":"https:\/\/www.sylvia-ritter.com\/index\/","content":"wallabag can't retrieve contents for this article. Please <a href=\"http:\/\/doc.wallabag.org\/en\/user\/errors_during_fetching.html#how-can-i-help-to-fix-that\">troubleshoot this issue<\/a>.\n","created_at":"2018-11-07T16:21:49+0000","updated_at":"2018-11-07T16:21:49+0000","annotations":[],"reading_time":0,"domain_name":"www.sylvia-ritter.com","http_status":"Unread"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":13,"title":"Rich Link Preview","url":"https:\/\/richpreview.com\/?url=qa.getpterotype.com","content":"Rich Link Preview\n\t\n\t\t<noscript><p class=\"chromeframe\">Your browser does not support Javascript<\/p><\/noscript>\n\t\t\n\t\t\n\t\t \n\t\t<section id=\"results\"><div class=\"container\"><img src=\"https:\/\/richpreview.com\/img\/ico-bad.svg\" alt=\"image\" \/><h2 class=\"textCenter mainTextColor textNoMargin\">error 10<\/h2><h2 class=\"textCenter\">To preserve resources, we need to verify that you are a human. <a href=\"#\" data-toggle=\"modal\" data-target=\"#captchaModal\">Please try again.<\/a><\/h2><\/div><\/section><div class=\"modal\" id=\"captchaModal\"><div class=\"modal-body\"><div class=\"dialog\"><div id=\"checkCaptchaForm\" class=\"captcha-zone\"><div class=\"zone\"><h2 class=\"textCenter textNoMargin\">this is just a human question<\/h2><div class=\"bg-progressbar m-bg-color\"><h3 class=\"textCenter textNoMargin\">error 10<\/h3><p class=\"textCenter textNoMargin\">\n\t\t\t\t\t\t\t\t\t\t\tThis specific URL was requested before in the past 15 minutes.\n\t\t\t\t\t\t\t\t\t\t\tPlease enter the captcha to confirm that you are a human.\n\t\t\t\t\t\t\t\t\t\t<\/p><\/div><\/div><\/div><\/div><\/div><\/div> \n\t\t\t\t\t\n\t\t<footer id=\"copyrightContent\"><div class=\"container\"><div class=\"zone\"><div class=\"rt-sm-8\"><p>\u00a9 2018 by richpreview.com - <a href=\"https:\/\/richpreview.com\/privacy\/\" title=\"Privacy statement\" rel=\"nofollow\">Privacy<\/a><\/p><\/div><div class=\"rt-sm-4\"><p class=\"rightZone\"><a href=\"https:\/\/richardstoolbox.com\/?url=qa.getpterotype.com\"><img id=\"logo-rt\" src=\"https:\/\/richpreview.com\/img\/logo-rt.svg\" alt=\"Richard's Toolbox\" title=\"Richard's Toolbox\" \/><em>Richard's<\/em> <strong>Toolbox<\/strong><\/a><\/p><\/div><\/div><\/div> \n\t\t<\/footer><p><\/p><h2 class=\"textCenter\">Please wait while checking<\/h2>\n\t\t\t<i class=\"ri ri-spinner ri-spin centerZone\"> \n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t<\/i>","created_at":"2018-11-07T16:21:49+0000","updated_at":"2018-11-07T16:21:49+0000","annotations":[],"mimetype":"text\/html","reading_time":0,"domain_name":"richpreview.com","preview_picture":"https:\/\/richpreview.com\/richpreview.png","http_status":"404"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":11,"title":"What is the revenue generation model for DuckDuckGo?","url":"https:\/\/news.ycombinator.com\/item?id=18320413","content":"<table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stephengillie\" class=\"hnuser\">stephengillie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320530\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andy_ppp\" class=\"hnuser\">andy_ppp<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320663\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=thelittleone\" class=\"hnuser\">thelittleone<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320785\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Nebbers\" class=\"hnuser\">Nebbers<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321100\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=the_duke\" class=\"hnuser\">the_duke<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321359\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=smbullet\" class=\"hnuser\">smbullet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321510\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=technimad\" class=\"hnuser\">technimad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322141\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ge0rg\" class=\"hnuser\">ge0rg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322556\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Too\" class=\"hnuser\">Too<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325271\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"280\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=confounded\" class=\"hnuser\">confounded<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18323897\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Arcsech\" class=\"hnuser\">Arcsech<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18324719\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=enriquto\" class=\"hnuser\">enriquto<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18336968\">8 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=weaksauce\" class=\"hnuser\">weaksauce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321952\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hiccuphippo\" class=\"hnuser\">hiccuphippo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18327238\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rovr138\" class=\"hnuser\">rovr138<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18327847\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=philipswood\" class=\"hnuser\">philipswood<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321628\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Strongerpass\" class=\"hnuser\">Strongerpass<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321906\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=rovr138\" class=\"hnuser\">rovr138<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18327871\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=eddie-spaghetti\" class=\"hnuser\">eddie-spaghetti<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322159\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=isaiahg\" class=\"hnuser\">isaiahg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322575\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320878\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=avip\" class=\"hnuser\">avip<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320993\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321009\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yeutterg\" class=\"hnuser\">yeutterg<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321178\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"240\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321269\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=avip\" class=\"hnuser\">avip<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321675\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=weberc2\" class=\"hnuser\">weberc2<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322531\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=powerslacker\" class=\"hnuser\">powerslacker<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322734\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yellowapple\" class=\"hnuser\">yellowapple<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321262\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=weaksauce\" class=\"hnuser\">weaksauce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321965\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yellowapple\" class=\"hnuser\">yellowapple<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325648\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=icotyl\" class=\"hnuser\">icotyl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321355\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Secretmapper\" class=\"hnuser\">Secretmapper<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321515\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tobr\" class=\"hnuser\">tobr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322760\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mountain_Skies\" class=\"hnuser\">Mountain_Skies<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18323133\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=djsumdog\" class=\"hnuser\">djsumdog<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18350679\">6 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gmueckl\" class=\"hnuser\">gmueckl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321436\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cbluth\" class=\"hnuser\">cbluth<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320724\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrhappyunhappy\" class=\"hnuser\">mrhappyunhappy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321063\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BonesJustice\" class=\"hnuser\">BonesJustice<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321093\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=amelius\" class=\"hnuser\">amelius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321238\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrhappyunhappy\" class=\"hnuser\">mrhappyunhappy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325610\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jake_the_third\" class=\"hnuser\">jake_the_third<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321281\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=barbecue_sauce\" class=\"hnuser\">barbecue_sauce<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321204\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrhappyunhappy\" class=\"hnuser\">mrhappyunhappy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325408\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fileeditview\" class=\"hnuser\">fileeditview<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321207\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mononokay\" class=\"hnuser\">Mononokay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321272\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jkaplowitz\" class=\"hnuser\">jkaplowitz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321660\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fileeditview\" class=\"hnuser\">fileeditview<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321410\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JCSato\" class=\"hnuser\">JCSato<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321205\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=new_guy\" class=\"hnuser\">new_guy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321112\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=andyidsinga\" class=\"hnuser\">andyidsinga<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321142\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=aylmao\" class=\"hnuser\">aylmao<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322300\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Technetium_Hat\" class=\"hnuser\">Technetium_Hat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18328538\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AznHisoka\" class=\"hnuser\">AznHisoka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321005\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bkanber\" class=\"hnuser\">bkanber<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321128\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dpcan\" class=\"hnuser\">dpcan<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320690\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cletus\" class=\"hnuser\">cletus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320782\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=robin_reala\" class=\"hnuser\">robin_reala<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320861\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=paulcole\" class=\"hnuser\">paulcole<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320882\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=bowlich\" class=\"hnuser\">bowlich<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322331\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stephengillie\" class=\"hnuser\">stephengillie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320825\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cletus\" class=\"hnuser\">cletus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322063\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=efdee\" class=\"hnuser\">efdee<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322289\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sanderjd\" class=\"hnuser\">sanderjd<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321121\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"200\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=stephengillie\" class=\"hnuser\">stephengillie<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321225\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jimmy1\" class=\"hnuser\">jimmy1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320706\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Ibethewalrus\" class=\"hnuser\">Ibethewalrus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18324971\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jimmy1\" class=\"hnuser\">jimmy1<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18346469\">7 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=daveFNbuck\" class=\"hnuser\">daveFNbuck<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320841\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=clubm8\" class=\"hnuser\">clubm8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320726\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=cgriswald\" class=\"hnuser\">cgriswald<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320807\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=hakfoo\" class=\"hnuser\">hakfoo<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18324082\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=woah\" class=\"hnuser\">woah<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320852\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Marsymars\" class=\"hnuser\">Marsymars<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321210\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chiefalchemist\" class=\"hnuser\">chiefalchemist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320799\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gaius\" class=\"hnuser\">gaius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321177\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=diggernet\" class=\"hnuser\">diggernet<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321843\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jvzr\" class=\"hnuser\">jvzr<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18327483\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=chiefalchemist\" class=\"hnuser\">chiefalchemist<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321198\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=izacus\" class=\"hnuser\">izacus<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321226\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gaius\" class=\"hnuser\">gaius<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325824\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickpsecurity\" class=\"hnuser\">nickpsecurity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321211\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mojuba\" class=\"hnuser\">mojuba<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321137\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gmueckl\" class=\"hnuser\">gmueckl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321470\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mojuba\" class=\"hnuser\">mojuba<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321615\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gmueckl\" class=\"hnuser\">gmueckl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321737\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=blackhaz\" class=\"hnuser\">blackhaz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321165\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nickpsecurity\" class=\"hnuser\">nickpsecurity<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321242\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JCSato\" class=\"hnuser\">JCSato<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321219\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=43920\" class=\"hnuser\">43920<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321323\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=gmueckl\" class=\"hnuser\">gmueckl<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321484\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=detaro\" class=\"hnuser\">detaro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321930\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=JCSato\" class=\"hnuser\">JCSato<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321999\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=we_love_slon\" class=\"hnuser\">we_love_slon<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322735\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=losvedir\" class=\"hnuser\">losvedir<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322002\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=piyush_soni\" class=\"hnuser\">piyush_soni<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18322358\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=isostatic\" class=\"hnuser\">isostatic<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325267\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=piyush_soni\" class=\"hnuser\">piyush_soni<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325445\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kmlx\" class=\"hnuser\">kmlx<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325693\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=piyush_soni\" class=\"hnuser\">piyush_soni<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18324620\">9 days ago<\/a>  [flagged]<\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=kmitz\" class=\"hnuser\">kmitz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18323190\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mr_toad\" class=\"hnuser\">mr_toad<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18323964\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dna_polymerase\" class=\"hnuser\">dna_polymerase<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320789\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=nunobrito\" class=\"hnuser\">nunobrito<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320985\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wslh\" class=\"hnuser\">wslh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321120\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dna_polymerase\" class=\"hnuser\">dna_polymerase<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321636\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=wslh\" class=\"hnuser\">wslh<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18323172\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=RobertDeNiro\" class=\"hnuser\">RobertDeNiro<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320838\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dna_polymerase\" class=\"hnuser\">dna_polymerase<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320869\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mrhappyunhappy\" class=\"hnuser\">mrhappyunhappy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321132\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dmurray\" class=\"hnuser\">dmurray<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321016\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BoorishBears\" class=\"hnuser\">BoorishBears<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320576\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=clubm8\" class=\"hnuser\">clubm8<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320733\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BoorishBears\" class=\"hnuser\">BoorishBears<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18335285\">8 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SiempreViernes\" class=\"hnuser\">SiempreViernes<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320641\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=BoorishBears\" class=\"hnuser\">BoorishBears<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18335292\">8 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=__initbrian__\" class=\"hnuser\">__initbrian__<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325894\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lurkars\" class=\"hnuser\">Lurkars<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321609\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=__initbrian__\" class=\"hnuser\">__initbrian__<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18325904\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Lurkars\" class=\"hnuser\">Lurkars<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18334808\">8 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=ing33k\" class=\"hnuser\">ing33k<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321267\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=jancsika\" class=\"hnuser\">jancsika<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321098\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321122\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320618\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=catacombs\" class=\"hnuser\">catacombs<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320693\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=tokyodude\" class=\"hnuser\">tokyodude<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320988\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SharkBl4de\" class=\"hnuser\">SharkBl4de<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321187\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dgacmu\" class=\"hnuser\">dgacmu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320721\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=crispyporkbites\" class=\"hnuser\">crispyporkbites<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320752\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=dgacmu\" class=\"hnuser\">dgacmu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320813\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=SquareWheel\" class=\"hnuser\">SquareWheel<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321034\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320640\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sveme\" class=\"hnuser\">sveme<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320761\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320864\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=scribu\" class=\"hnuser\">scribu<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320703\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320862\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321001\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=mauricioc\" class=\"hnuser\">mauricioc<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321051\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fucking_tragedy\" class=\"hnuser\">fucking_tragedy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321097\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321221\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fucking_tragedy\" class=\"hnuser\">fucking_tragedy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321307\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"80\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321077\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"120\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Mononokay\" class=\"hnuser\">Mononokay<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321308\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"160\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=utopcell\" class=\"hnuser\">utopcell<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321336\">9 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=yladiz\" class=\"hnuser\">yladiz<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321030\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sireat\" class=\"hnuser\">sireat<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321075\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=koboll\" class=\"hnuser\">koboll<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321082\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=Rinzler89\" class=\"hnuser\">Rinzler89<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321040\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=AznHisoka\" class=\"hnuser\">AznHisoka<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321015\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=fucking_tragedy\" class=\"hnuser\">fucking_tragedy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321050\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=StanislavPetrov\" class=\"hnuser\">StanislavPetrov<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321026\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"0\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"14\" alt=\"image\" \/><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=the_other_guy\" class=\"hnuser\">the_other_guy<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18320885\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>\n            <table border=\"0\"><tr><td class=\"ind\"><img src=\"https:\/\/news.ycombinator.com\/s.gif\" height=\"1\" width=\"40\" alt=\"image\" \/><\/td><td valign=\"top\" class=\"votelinks\"><div style=\"text-align: center;\"><\/div><\/td><td class=\"default\"><p><a href=\"https:\/\/news.ycombinator.com\/user?id=sctb\" class=\"hnuser\">sctb<\/a> <a href=\"https:\/\/news.ycombinator.com\/item?id=18321134\">10 days ago<\/a><\/p><br \/><\/td><\/tr><\/table>","created_at":"2018-11-07T16:21:47+0000","updated_at":"2018-11-07T16:21:47+0000","annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"news.ycombinator.com","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":10,"title":"facundoolano\/advenjure","url":"https:\/\/github.com\/facundoolano\/advenjure","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\">\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/github.com\/facundoolano\/advenjure\/blob\/master\/example.gif\"><img src=\"https:\/\/github.com\/facundoolano\/advenjure\/raw\/master\/example.gif\" alt=\"Example game\" \/><\/a><\/p>\n<p>Advenjure is a text adventure (or interactive fiction) game engine. I wrote it as an excuse to learn Clojure.\nSome of its distinctive features are:<\/p>\n<ul><li>Target the terminal and the browser with the same codebase.<\/li>\n<li>Unix-like prompt with smart tab completion and command history<\/li>\n<li>Customizable verbs and plugins.<\/li>\n<li>LucasArts inspired dialog trees.<\/li>\n<\/ul><h2><a id=\"user-content-example-game\" class=\"anchor\" aria-hidden=\"true\" href=\"#example-game\"><\/a>Example game<\/h2>\n<p>You can see the code for a working example in the <a href=\"https:\/\/github.com\/facundoolano\/advenjure-example\">advenjure-example<\/a> repository and play it online <a href=\"https:\/\/facundoolano.github.io\/advenjure\" rel=\"nofollow\">here<\/a>.<\/p>\n<p>For a fully fledged game see <a href=\"https:\/\/github.com\/facundoolano\/house-taken-over\">House Taken Over<\/a>.<\/p>\n<h2><a id=\"user-content-installation\" class=\"anchor\" aria-hidden=\"true\" href=\"#installation\"><\/a>Installation<\/h2>\n<p>Add the following to your project map as a dependency:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>[advenjure \"0.9.0\"]<\/pre><\/div>\n<h2><a id=\"user-content-basic-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#basic-usage\"><\/a>Basic Usage<\/h2>\n<h3><a id=\"user-content-creating-items\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-items\"><\/a>Creating items<\/h3>\n<p>Text adventures consist mainly of moving around rooms and interacting with items\nthrough verb commands such as GO, LOOK, TAKE, etc.<\/p>\n<p>Items are represented by maps in advenjure. the <code>advenjure.items\/make<\/code>\nfunction takes a name, a description and a set of key-value pairs to customize behavior.\nFor example:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[advenjure.items :as items])\n(def magazine (items\/make \"magazine\"\n                          \"The cover read 'Sports Almanac 1950-2000'\"\n                          :take true\n                          :read \"Oh l\u00e0 l\u00e0? Oh l\u00e0 l\u00e0!?\"))<\/pre><\/div>\n<p>That will define a magazine item that can be taken (put in the player's inventory)\nand can be read (which differs from looking at it).<\/p>\n<p>You can provide a vector of names so the player can refer to it by one of its synonyms;\nthe first name in the vector will be considered its canonical name:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def magazine (items\/make [\"magazine\" \"sports magazine\" \"newspaper\"]\n                          \"The cover read 'Sports Almanac 1950-2000'\"\n                          :take true\n                          :read \"Oh l\u00e0 l\u00e0? Oh l\u00e0 l\u00e0!?\"))<\/pre><\/div>\n<p>Like <code>:take<\/code> and <code>:read<\/code> there are keywords for the other actions\n(<code>:look-at<\/code>, <code>:open<\/code>, <code>:close<\/code>, <code>:unlock<\/code>, etc.).<\/p>\n<p>A special kind of items are those that can contain other items:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def magazine (items\/make \"bag\" :items #{magazine} :closed true))<\/pre><\/div>\n<p>The bag contains the magazine, but since it's <code>:closed<\/code> the player needs to open it\nbefore being able to look inside it and take its contents. Note that marking an\nobject as <code>:closed<\/code> also implies that OPEN and CLOSE verbs can be applied to it\n(i.e. it means <code>:open true, :close true<\/code>).<\/p>\n<h3><a id=\"user-content-creating-rooms\" class=\"anchor\" aria-hidden=\"true\" href=\"#creating-rooms\"><\/a>Creating rooms<\/h3>\n<p>Once you've created a bunch of items, you'll need to put them in a room (if not directly\ninto the player's inventory). Rooms are also maps and also have an\n<code>advenjure.rooms\/make<\/code> function to build them:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[advenjure.rooms :as rooms])\n(def bedroom (rooms\/make \"Bedroom\"\n                         \"A smelling bedroom.\"\n                         :initial-description \"I woke up in a smelling little bedroom, without windows.\"))<\/pre><\/div>\n<p>Note that rooms can have only one name. <code>:initial-description<\/code> is an optional attribute\nto define how the player will describe a room the first time he visits it,\nusually with a more verbose description. If <code>:initial-description<\/code> is not defined,\nand whenever the LOOK AROUND command is entered, the regular description will be used.<\/p>\n<p>To add items to a room use <code>advenjure.rooms\/add-item<\/code>:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def bedroom (-&gt; (rooms\/make \"Bedroom\"\n                             \"A smelling bedroom.\"\n                             :initial-description \"I woke up in a smelling little bedroom, without windows.\")\n                 (rooms\/add-item (items\/make \"bed\" \"It was the bed I slept in.\"))\n                 (rooms\/add-item magazine \"On the floor was a sports magazine.\")))<\/pre><\/div>\n<p>The second parameter is an optional room-specific description of the item. It will be used\nto mention the item while describing the room (as opposed of the default <code>a &lt;item&gt; is here.<\/code>).<\/p>\n<h3><a id=\"user-content-building-a-room-map\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-a-room-map\"><\/a>Building a room map<\/h3>\n<p>Once you have some rooms, you need to connect them to build a room map, which is\nnothing but a plain clojure hash map. First map the room to some id keyword,\nthen connect the rooms using the <code>advenjure.rooms\/connect<\/code> function:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(def room-map (-&gt; {:bedroom bedroom\n                   :living living\n                   :outside outside}\n                  (rooms\/connect :bedroom :north :living)\n                  (rooms\/connect :living :east :outside)))<\/pre><\/div>\n<p>An alternative function, <code>advenjure.rooms\/one-way-connect<\/code>, allows connecting the\nrooms just in one direction.<\/p>\n<h3><a id=\"user-content-building-and-running-a-game\" class=\"anchor\" aria-hidden=\"true\" href=\"#building-and-running-a-game\"><\/a>Building and running a game<\/h3>\n<p>The next building block is the game map itself, which contains the room map,\nthe player's inventory and a pointer to the current room. <code>advenjure.game\/make<\/code>\nhelps to build it:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(require '[advenjure.game :as game])\n(game\/make room-map :bedroom)<\/pre><\/div>\n<p>The room keyword defines what room the player will be in when the game starts.\nIf you want to start off the game with some items in the player's inventory,\njust pass them in a set as the third argument.<\/p>\n<p>Lastly, the <code>advenjure.game\/run<\/code> takes a game state map, a boolean function\nto tell if the game has finished and an optional string to print before it starts.\nPutting it all together in a <code>-main<\/code> function:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>(defn -main\n  \"Build and run the game.\"\n  [&amp; args]\n  (let [game-state (game\/make room-map :bedroom)\n        finished? #(= (:current-room %) :outside)]\n    (game\/run game-state finished? :start-message \"Welcome to the advenjure!\")))<\/pre><\/div>\n<p>The game flows by taking the initial game state map, prompting the user for a command,\napplying the command to produce a new game state and repeat the process until the\n<code>finished?<\/code> condition is met, which, in the example above means entering the\n<code>:outside<\/code> room.<\/p>\n<h2><a id=\"user-content-advanced-usage\" class=\"anchor\" aria-hidden=\"true\" href=\"#advanced-usage\"><\/a>Advanced Usage<\/h2>\n<p>There is a number of advanced features available in the engine:<\/p>\n<ul><li>Overriding messages: use custom messages for a given action on a room or item.<\/li>\n<li>Pre conditions: function hook to define whether an action can be performed.<\/li>\n<li>Post conditions: function hook to customize how the game state is modified after an action is performed.<\/li>\n<li>Dialogs: interactive dialogs with 'character' items, in the style of the LucasArts graphic adventures.<\/li>\n<li>Text customization and internationalization.<\/li>\n<li>Custom verbs\/commands.<\/li>\n<li>Plugin hooks to customize behavior without modifying the library.<\/li>\n<\/ul><p>I'm waiting for the APIs to stabilize (and get a lot of free time) before fully documenting all those features,\nbut I'd be happy to write something up if you need help with something specific, just file an issue!<\/p>\n<h2><a id=\"user-content-run-on-the-browser\" class=\"anchor\" aria-hidden=\"true\" href=\"#run-on-the-browser\"><\/a>Run on the browser<\/h2>\n<p>The codebase is prepared to run both in the terminal with Clojure and the browser with ClojureScript.\nAn example configuration, using lein-cljsbuild would be:<\/p>\n<div class=\"highlight highlight-source-clojure\"><pre>:cljsbuild\n    {:builds\n     {:main {:source-paths [\"src\"]\n             :compiler {:output-to \"main.js\"\n                        :main example.core\n                        :optimizations :simple\n                        :pretty-print false\n                        :optimize-constants true\n                        :static-fns true}}}<\/pre><\/div>\n<p>Then the command <code>lein cljsbuild once<\/code> will output a <code>main.js<\/code> file that can be included in any web page to run the game.\nThe HTML should have a <code>#terminal<\/code> div and include the <a href=\"https:\/\/github.com\/facundoolano\/advenjure-example\/blob\/master\/resources\/jquery.terminal-0.11.7.css\">jQuery Terminal CSS<\/a> to properly render the terminal.<\/p>\n<p>The current limitations of the ClojureScript version of the library are:<\/p>\n<ul><li>No internationalization support (since clojure-gettext does not support ClojureScript).<\/li>\n<li>Can use up to <code>:simple<\/code> optimizations (not <code>:advanced<\/code>), since <a href=\"https:\/\/github.com\/clojure\/clojurescript\/wiki\/Optional-Self-hosting\">ClojureScript self-hosting<\/a> is required for some of the advanced features.<\/li>\n<\/ul><p>See the <a href=\"https:\/\/github.com\/facundoolano\/advenjure-example\">advenjure-example<\/a> for a game that targets both the terminal and the browser.<\/p>\n<h2><a class=\"anchor\" aria-hidden=\"true\" href=\"#roadmap\"><\/a>Roadmap<\/h2>\n<p><a href=\"https:\/\/github.com\/facundoolano\/advenjure\/milestones\">Check the milestones<\/a>.<\/p>\n<\/article>","created_at":"2018-11-07T16:21:46+0000","updated_at":"2018-11-07T16:21:46+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":5,"domain_name":"github.com","preview_picture":"https:\/\/avatars2.githubusercontent.com\/u\/1040941?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":9,"title":"Federated Sites \u2013 switching.social","url":"https:\/\/switching.social\/federated-sites\/","content":"<p>Most alternative social media are \u201cfederated\u201d, or decentralised. This means that instead of being on one large site (like Facebook or Twitter are), federated sites are split across many smaller sites that are linked together. People who sign up on one small site can make friends with people on other small sites, because they\u2019re part of the same federated network. From the user\u2019s point of view, it\u2019s just like using one large site.<\/p>\n<p>This may sound complicated, but all of us use a federated system every day: our telephones. There are thousands of phone companies around the world all linked together. You can call any phone worldwide from any other phone because federation allows them to be part of one global telephone network.<\/p>\n<p>Connected small sites tend to be more ethical because small sites\u2019 users have more choice: if a small site misbehaves, its users can change to a different small site on the same network, just like they would change to a new phone company if their old phone company lets them down.<\/p>\n<p>Federated networks also tend to be better for privacy, because the data on them is spread across many small sites instead of concentrated on one big site. There is no way for any single site to sell everyone\u2019s data, because they do not have (and never will have) everyone\u2019s data.<\/p>\n<p>The federated social network Mastodon <a href=\"https:\/\/www.youtube.com\/watch?v=IPSbNdBmWKE\" target=\"_blank\" rel=\"noopener\">has quite a good video<\/a> explaining how decentralisation works.<\/p>\n<p>(Federations also let anyone set up their own part of the network, which many people and organisations do. It can be quite technically demanding to do this entirely on your own. A much easier option is to use a managed hosting service like <a href=\"https:\/\/masto.host\/\" target=\"_blank\" rel=\"noopener\">masto.host<\/a>, where you control the site but the service manages all the dull technical stuff.)<\/p>\n<p><a href=\"https:\/\/switching.social\/\">back to the front page<\/a><\/p>","created_at":"2018-11-07T16:21:45+0000","updated_at":"2018-11-07T16:21:45+0000","annotations":[],"mimetype":"text\/html","language":"en_GB","reading_time":1,"domain_name":"switching.social","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":7,"title":"mntmn\/interim","url":"https:\/\/github.com\/mntmn\/interim\/blob\/master\/README.md","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/b45606befa813b1e31b56e3b54204ece3b536394\/687474703a2f2f64756d702e6d6e746d6e2e636f6d2f626a6f732e706e67\"><img src=\"https:\/\/camo.githubusercontent.com\/b45606befa813b1e31b56e3b54204ece3b536394\/687474703a2f2f64756d702e6d6e746d6e2e636f6d2f626a6f732e706e67\" alt=\"Interim OS Bomberjacket Logo\" data-canonical-src=\"http:\/\/dump.mntmn.com\/bjos.png\" \/><\/a><\/p>\n<h2><a id=\"user-content-interim\" class=\"anchor\" aria-hidden=\"true\" href=\"#interim\"><\/a>Interim<\/h2>\n<p>Interim OS is a radical new operating system with a focus on minimalism. It steals conceptually from Lisp machines (language-based kernel) and Plan 9 (everything is a file system). It boots to a JITting Lisp-like REPL and offers the programmer\/user the system's resources as filesystems.<\/p>\n<p>Interim runs on:<\/p>\n<ul><li>Raspberry Pi 2 (Broadcom VideoCore4\/ARMv7, Bare Metal)<\/li>\n<li>ARM5+ Linux (Hosted)<\/li>\n<li>Intel\/AMD x64 Linux, Mac OS X (Hosted)<\/li>\n<li>Intel\/AMD x86 Linux, Windows (Hosted, win32 .exe cross compiled from Linux)<\/li>\n<li>Motorola 68020+ Amiga OS 3.x+ (Hosted)<\/li>\n<\/ul><p>Detailed docs are in <em>docs\/index.html<\/em>.<\/p>\n<h2><a id=\"user-content-contact--faq--get-involved\" class=\"anchor\" aria-hidden=\"true\" href=\"#contact--faq--get-involved\"><\/a>Contact \/ FAQ \/ Get Involved<\/h2>\n<p>The official IRC channel for development and usage questions is #interim-os on freenode.<\/p>\n<p>If you find bugs, please file an issue in the <a href=\"https:\/\/github.com\/mntmn\/interim\/issues\">Github issue tracker<\/a>.<\/p>\n<h2><a id=\"user-content-architecture\" class=\"anchor\" aria-hidden=\"true\" href=\"#architecture\"><\/a>Architecture<\/h2>\n<ul><li>Layer 2: Bitmapped Terminal \/ Editor<\/li>\n<li>Layer 1: Sledge Lisp JIT Compiler<\/li>\n<li>Layer 0: Platform Interface (Startup code and filesystems)<\/li>\n<\/ul><h2><a id=\"user-content-design-choices\" class=\"anchor\" aria-hidden=\"true\" href=\"#design-choices\"><\/a>Design Choices<\/h2>\n<ul><li>The shell is the editor is the REPL is the language is the compiler.<\/li>\n<li>Namespacing allows sandboxing and network transparency.<\/li>\n<li>You need only one text encoding: UTF-8.<\/li>\n<li>All facilities must be navigable by keyboard.<\/li>\n<\/ul><h2><a id=\"user-content-building\" class=\"anchor\" aria-hidden=\"true\" href=\"#building\"><\/a>Building<\/h2>\n<ol><li>Get the dependencies. For a (Linux-)hosted version, this is only GCC and SDL1 or SDL2 if you want windowed graphics. Sledge can also use \/dev\/fb0 instead if available on your platform. If you want to cross-compile for an ARM platform, get <em>arm-none-eabi-gcc<\/em> and <em>libnewlib-arm-none-eabi<\/em>.<\/li>\n<\/ol><p>1b. For a hosted version on Mac OS X, you will need GNU binutils, which can be obtained via homebrew: <code>brew install binutils<\/code>. You can also get SDL2 via homebrew.<\/p>\n<ol start=\"2\"><li>\n<p>To build the hosted variant, cd to <code>sledge<\/code> and <code>.\/build_x64.sh<\/code>.<\/p>\n<\/li>\n<li>\n<p>To cross-compile for bare metal, use <code>.\/rpi2-build.sh<\/code>. This will create build\/interim-arm.elf.<\/p>\n<\/li>\n<\/ol><h2><a id=\"user-content-running-hosted\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-hosted\"><\/a>Running (Hosted)<\/h2>\n<pre>cd sledge\n.\/sledge\n<\/pre>\n<p>To launch the graphical REPL\/Shell:<\/p>\n<pre>.\/sledge &lt; tests\/boot2.l\n<\/pre>\n<h2><a id=\"user-content-running-raspberry-pi-2\" class=\"anchor\" aria-hidden=\"true\" href=\"#running-raspberry-pi-2\"><\/a>Running (Raspberry Pi 2)<\/h2>\n<p>Prepare a bootable SD card with the usual FAT partition that has the Pi-specific boot blobs in it and copy <code>kernel.img<\/code> into it. You can recycle any other Raspberry OS distribution, i.e. Raspbian for this. Just replace the kernel7.img and delete cmdline.txt. Keyboard input is currently only over UART, so you will probably want to connect a UART-&gt;USB cable to another computer and use it to control Interim.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/50a511ffd26bbe7d5e12aa34edd65ac07768c952\/687474703a2f2f64756d702e6d6e746d6e2e636f6d2f696e746572696d2d70617065722f696c6c757374726174696f6e732f696e746572696d2d706963747572652e6a70673f7761743d31\"><img src=\"https:\/\/camo.githubusercontent.com\/50a511ffd26bbe7d5e12aa34edd65ac07768c952\/687474703a2f2f64756d702e6d6e746d6e2e636f6d2f696e746572696d2d70617065722f696c6c757374726174696f6e732f696e746572696d2d706963747572652e6a70673f7761743d31\" alt=\"Interim OS Screenshot\" data-canonical-src=\"http:\/\/dump.mntmn.com\/interim-paper\/illustrations\/interim-picture.jpg?wat=1\" \/><\/a><\/p>\n<h2><a id=\"user-content-roadmap\" class=\"anchor\" aria-hidden=\"true\" href=\"#roadmap\"><\/a>Roadmap<\/h2>\n<ul><li>IEEE 754 double-precision floating point numbers.<\/li>\n<li>Vectors.<\/li>\n<li>Native string maps (dictionaries) based on the existing sd_hash.<\/li>\n<li>JavaScript to Interim Lisp compiler.<\/li>\n<\/ul><h2><a id=\"user-content-licenses\" class=\"anchor\" aria-hidden=\"true\" href=\"#licenses\"><\/a>Licenses<\/h2>\n<p>Interim OS: GNU GPLv3 or later, (C) 2015-2016 Lukas F. Hartmann \/ @mntmn<\/p>\n<p>Interim OS is free software: you can redistribute it and\/or modify\nit under the terms of the GNU General Public License as published by\nthe Free Software Foundation, either version 3 of the License, or\n(at your option) any later version.\nThis program is distributed in the hope that it will be useful,\nbut WITHOUT ANY WARRANTY; without even the implied warranty of\nMERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\nGNU General Public License for more details.\nYou should have received a copy of the GNU General Public License\nalong with this program.  If not, see <a href=\"http:\/\/www.gnu.org\/licenses\/\" rel=\"nofollow\">http:\/\/www.gnu.org\/licenses\/<\/a>.<\/p>\n<ul><li>newlib: GNU GPLv2, Maintained by Corinna Vinschen, Jeff Johnston \/ Red Hat Inc.<\/li>\n<li>devices\/rpi2\/uspi: GNU GPLv3, Copyright (C) 2014-2015  R. Stange <a href=\"mailto:rsta2@o2online.de\">rsta2@o2online.de<\/a><\/li>\n<li>devices\/rpi2\/rpi-boot: Copyright (C) 2013 by John Cronin <a href=\"mailto:jncronin@tysos.org\">jncronin@tysos.org<\/a><\/li>\n<li>devices\/rpi2\/mmu.c for Raspberry Pi 2 adapted from code Copyright (C) 2015 by Goswin von Brederlow, GNU GPLv3+<\/li>\n<li>devices\/rpi2\/fat FAT file system module Copyright by (C) 2013 by ChaN, Free Software \/ Custom License<\/li>\n<li>GNU Unifont: GNU GPLv2+, originally by Roman Czyborra<\/li>\n<li>devices\/imx233 IMX233 Port contains code from U-Boot, GPLv2+, written by various engineers. See license headers in source files.<\/li>\n<\/ul><p>Codecs<\/p>\n<ul><li>minimp3: GNU LGPL, Copyright (C) 2001, 2002 Fabrice Bellard, (C) 2007 Martin J. Fiedler<\/li>\n<li>uPNG: zlib License, Copyright (C) 2005-2010 Lode Vandevenne, (C) 2010 Sean Middleditch<\/li>\n<li>NanoJPEG: KeyJ's Research License, Copyright (C) 2014 Martin J. Fiedler<\/li>\n<\/ul>","created_at":"2018-11-07T16:21:44+0000","updated_at":"2018-11-07T16:21:44+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"github.com","preview_picture":"https:\/\/avatars1.githubusercontent.com\/u\/150458?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":8,"title":"Open Graph protocol","url":"http:\/\/ogp.me\/","content":"<h2>The <a href=\"http:\/\/ogp.me\/\">Open Graph protocol<\/a> enables any web page to become a\nrich object in a social graph. For instance, this is used on Facebook to allow\nany web page to have the same functionality as any other object on Facebook.While many different technologies and schemas exist and could be combined\ntogether, there isn't a single technology which provides enough information to\nrichly represent any web page within the social graph. The Open Graph protocol\nbuilds on these existing technologies and gives developers one thing to\nimplement. Developer simplicity is a key goal of the Open Graph protocol which\nhas informed many of\n<a href=\"http:\/\/www.scribd.com\/doc\/30715288\/The-Open-Graph-Protocol-Design-Decisions\">the technical design decisions<\/a>.To turn your web pages into graph objects, you need to add basic metadata to\nyour page. We've based the initial version of the protocol on\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/RDFa\">RDFa<\/a> which means that you'll place\nadditional <code>&lt;meta&gt;<\/code> tags in the <code>&lt;head&gt;<\/code> of your web page. The four required\nproperties for every page are:<code>og:title<\/code> - The title of your object as it should appear within the graph,\ne.g., \"The Rock\".\n<code>og:type<\/code> - The <a href=\"#types\">type<\/a> of your object, e.g., \"video.movie\".  Depending on\nthe type you specify, other properties may also be required.\n<code>og:image<\/code> - An image URL which should represent your object within the\ngraph.\n<code>og:url<\/code> - The canonical URL of your object that will be used as its\npermanent ID in the graph, e.g., \"http:\/\/www.imdb.com\/title\/tt0117500\/\".\nAs an example, the following is the Open Graph protocol markup for <a href=\"http:\/\/www.imdb.com\/title\/tt0117500\/\">The Rock on\nIMDB<\/a>:&lt;html prefix=\"og: http:\/\/ogp.me\/ns#\"&gt;\n&lt;head&gt;\n&lt;title&gt;The Rock (1996)&lt;\/title&gt;\n&lt;meta property=\"og:title\" content=\"The Rock\" \/&gt;\n&lt;meta property=\"og:type\" content=\"video.movie\" \/&gt;\n&lt;meta property=\"og:url\" content=\"http:\/\/www.imdb.com\/title\/tt0117500\/\" \/&gt;\n&lt;meta property=\"og:image\" content=\"http:\/\/ia.media-imdb.com\/images\/rock.jpg\" \/&gt;\n...\n&lt;\/head&gt;\n...\n&lt;\/html&gt;\n<a id=\"optional\" href=\"#optional\">Optional Metadata<\/a>The following properties are optional for any object and are generally\nrecommended:<code>og:audio<\/code> - A URL to an audio file to accompany this object.\n<code>og:description<\/code> - A one to two sentence description of your object.\n<code>og:determiner<\/code> - The word that appears before this object's title\nin a sentence. An <a href=\"#enum\">enum<\/a> of (a, an, the, \"\", auto). If <code>auto<\/code> is \nchosen, the consumer of your data should chose between \"a\" or \"an\".\nDefault is \"\" (blank).\n<code>og:locale<\/code> - The locale these tags are marked up in.\nOf the format <code>language_TERRITORY<\/code>. Default is <code>en_US<\/code>.\n<code>og:locale:alternate<\/code> - An <a href=\"#array\">array<\/a> of other locales this page is \navailable in.\n<code>og:site_name<\/code> - If your object is part of a larger web site, the name which\nshould be displayed for the overall site. e.g., \"IMDb\".\n<code>og:video<\/code> - A URL to a video file that complements this object.\nFor example (line-break solely for display purposes):&lt;meta property=\"og:audio\" content=\"http:\/\/example.com\/bond\/theme.mp3\" \/&gt;\n&lt;meta property=\"og:description\" \n  content=\"Sean Connery found fame and fortune as the\n           suave, sophisticated British agent, James Bond.\" \/&gt;\n&lt;meta property=\"og:determiner\" content=\"the\" \/&gt;\n&lt;meta property=\"og:locale\" content=\"en_GB\" \/&gt;\n&lt;meta property=\"og:locale:alternate\" content=\"fr_FR\" \/&gt;\n&lt;meta property=\"og:locale:alternate\" content=\"es_ES\" \/&gt;\n&lt;meta property=\"og:site_name\" content=\"IMDb\" \/&gt;\n&lt;meta property=\"og:video\" content=\"http:\/\/example.com\/bond\/trailer.swf\" \/&gt;\nThe RDF schema (in <a href=\"http:\/\/en.wikipedia.org\/wiki\/Turtle_(syntax)\">Turtle<\/a>) \ncan be found at <a href=\"http:\/\/ogp.me\/ns\/ogp.me.ttl\">ogp.me\/ns<\/a>.Some properties can have extra metadata attached to them.\nThese are specified in the same way as other metadata with <code>property<\/code> and\n<code>content<\/code>, but the <code>property<\/code> will have extra <code>:<\/code>.The <code>og:image<\/code> property has some optional structured properties:<code>og:image:url<\/code> - Identical to <code>og:image<\/code>.\n<code>og:image:secure_url<\/code> - An alternate url to use if the webpage requires\nHTTPS.\n<code>og:image:type<\/code> - A <a href=\"http:\/\/en.wikipedia.org\/wiki\/Internet_media_type\">MIME type<\/a> for this image.\n<code>og:image:width<\/code> - The number of pixels wide.\n<code>og:image:height<\/code> - The number of pixels high.\n<code>og:image:alt<\/code> - A description of what is in the image (not a caption). If the page specifies an og:image it should specify <code>og:image:alt<\/code>.\nA full image example:&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/ogp.jpg\" \/&gt;\n&lt;meta property=\"og:image:secure_url\" content=\"https:\/\/secure.example.com\/ogp.jpg\" \/&gt;\n&lt;meta property=\"og:image:type\" content=\"image\/jpeg\" \/&gt;\n&lt;meta property=\"og:image:width\" content=\"400\" \/&gt;\n&lt;meta property=\"og:image:height\" content=\"300\" \/&gt;\n&lt;meta property=\"og:image:alt\" content=\"A shiny red apple with a bite taken out\" \/&gt;\nThe <code>og:video<\/code> tag has the identical tags as <code>og:image<\/code>. Here is an example:&lt;meta property=\"og:video\" content=\"http:\/\/example.com\/movie.swf\" \/&gt;\n&lt;meta property=\"og:video:secure_url\" content=\"https:\/\/secure.example.com\/movie.swf\" \/&gt;\n&lt;meta property=\"og:video:type\" content=\"application\/x-shockwave-flash\" \/&gt;\n&lt;meta property=\"og:video:width\" content=\"400\" \/&gt;\n&lt;meta property=\"og:video:height\" content=\"300\" \/&gt;\nThe <code>og:audio<\/code> tag only has the first 3 properties available\n(since size doesn't make sense for sound):&lt;meta property=\"og:audio\" content=\"http:\/\/example.com\/sound.mp3\" \/&gt;\n&lt;meta property=\"og:audio:secure_url\" content=\"https:\/\/secure.example.com\/sound.mp3\" \/&gt;\n&lt;meta property=\"og:audio:type\" content=\"audio\/mpeg\" \/&gt;\nIf a tag can have multiple values, just put multiple versions of the same\n<code>&lt;meta&gt;<\/code> tag on your page. The first tag (from top to bottom) is given\npreference during conflicts.&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/rock.jpg\" \/&gt;\n&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/rock2.jpg\" \/&gt;\nPut structured properties after you declare their root tag. Whenever\nanother root element is parsed, that structured property\nis considered to be done and another one is started.For example:&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/rock.jpg\" \/&gt;\n&lt;meta property=\"og:image:width\" content=\"300\" \/&gt;\n&lt;meta property=\"og:image:height\" content=\"300\" \/&gt;\n&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/rock2.jpg\" \/&gt;\n&lt;meta property=\"og:image\" content=\"http:\/\/example.com\/rock3.jpg\" \/&gt;\n&lt;meta property=\"og:image:height\" content=\"1000\" \/&gt;\nmeans there are 3 images on this page, the first image is <code>300x300<\/code>, the middle\none has unspecified dimensions, and the last one is <code>1000<\/code>px tall.In order for your object to be represented within the graph, you need to\nspecify its type. This is done using the <code>og:type<\/code> property:&lt;meta property=\"og:type\" content=\"website\" \/&gt;\nWhen the community agrees on the schema for a type, it is added to the list\nof global types. All other objects in the type system are\n<a href=\"http:\/\/en.wikipedia.org\/wiki\/CURIE\">CURIEs<\/a> of the form&lt;head prefix=\"my_namespace: http:\/\/example.com\/ns#\"&gt;\n&lt;meta property=\"og:type\" content=\"my_namespace:my_type\" \/&gt;\nThe global types are grouped into verticals. Each vertical has its\nown namespace. The <code>og:type<\/code> values for a namespace are always prefixed with\nthe namespace and then a period.\nThis is to reduce confusion with user-defined namespaced types which always\nhave colons in them.<a id=\"type_music\" href=\"#type_music\">Music<\/a><code>og:type<\/code> values:<a name=\"type_music.song\" href=\"#type_music.song\"><code>music.song<\/code><\/a><code>music:duration<\/code> - <a href=\"#integer\">integer<\/a> &gt;=1 - The song's length in seconds.\n<code>music:album<\/code> - <a href=\"#type_music.album\">music.album<\/a> <a href=\"#array\">array<\/a> -\nThe album this song is from.\n<code>music:album:disc<\/code> - <a href=\"#integer\">integer<\/a> &gt;=1 -\nWhich disc of the album this song is on.\n<code>music:album:track<\/code> - <a href=\"#integer\">integer<\/a> &gt;=1 -\nWhich track this song is.\n<code>music:musician<\/code> - <a href=\"#type_profile\">profile<\/a> <a href=\"#array\">array<\/a> -\nThe musician that made this song.\n<a name=\"type_music.album\" href=\"#type_music.album\"><code>music.album<\/code><\/a><code>music:song<\/code> - <a href=\"#type_music.song\">music.song<\/a> - The song on this album.\n<code>music:song:disc<\/code> - <a href=\"#integer\">integer<\/a> &gt;=1 -\nThe same as <code>music:album:disc<\/code> but in reverse.\n<code>music:song:track<\/code> - <a href=\"#integer\">integer<\/a> &gt;=1 -\nThe same as <code>music:album:track<\/code> but in reverse.\n<code>music:musician<\/code> - <a href=\"#type_profile\">profile<\/a> -\nThe musician that made this song.\n<code>music:release_date<\/code> - <a href=\"#datetime\">datetime<\/a> - \nThe date the album was released.\n<a name=\"type_music.playlist\" href=\"#type_music.playlist\"><code>music.playlist<\/code><\/a><code>music:song<\/code> - Identical to the ones on <a href=\"#type_music.album\">music.album<\/a>\n<code>music:song:disc<\/code>\n<code>music:song:track<\/code>\n<code>music:creator<\/code> - <a href=\"#type_profile\">profile<\/a> - The creator of this playlist.\n<a name=\"type_music.radio_station\" href=\"#type_music.radio_station\"><code>music.radio_station<\/code><\/a><code>music:creator<\/code> - <a href=\"#type_profile\">profile<\/a> - The creator of this station.\n<a id=\"type_video\" href=\"#type_video\">Video<\/a><code>og:type<\/code> values:<a name=\"type_video.movie\" href=\"#type_video.movie\"><code>video.movie<\/code><\/a><code>video:actor<\/code> - <a href=\"#type_profile\">profile<\/a> <a href=\"#array\">array<\/a> -\nActors in the movie.\n<code>video:actor:role<\/code> - <a href=\"#string\">string<\/a> - The role they played.\n<code>video:director<\/code> - <a href=\"#type_profile\">profile<\/a> <a href=\"#array\">array<\/a> -\nDirectors of the movie.\n<code>video:writer<\/code> - <a href=\"#type_profile\">profile<\/a> <a href=\"#array\">array<\/a> -\nWriters of the movie.\n<code>video:duration<\/code> - <a href=\"#integer\">integer<\/a> &gt;=1 - \nThe movie's length in seconds.\n<code>video:release_date<\/code> - <a href=\"#datetime\">datetime<\/a> - \nThe date the movie was released.\n<code>video:tag<\/code> - <a href=\"#string\">string<\/a> <a href=\"#array\">array<\/a> -\nTag words associated with this movie.\n<a name=\"type_video.episode\" href=\"#type_video.episode\"><code>video.episode<\/code><\/a><code>video:actor<\/code> - Identical to <a href=\"#type_video.movie\">video.movie<\/a>\n<code>video:actor:role<\/code>\n<code>video:director<\/code>\n<code>video:writer<\/code>\n<code>video:duration<\/code>\n<code>video:release_date<\/code>\n<code>video:tag<\/code>\n<code>video:series<\/code> - <a href=\"#type_video.tv_show\">video.tv_show<\/a> -\nWhich series this episode belongs to.\n<a name=\"type_video.tv_show\" href=\"#type_video.tv_show\"><code>video.tv_show<\/code><\/a>A multi-episode TV show.\nThe metadata is identical to <a href=\"#type_video.movie\">video.movie<\/a>.<a name=\"type_video.other\" href=\"#type_video.other\"><code>video.other<\/code><\/a>A video that doesn't belong in any other category.\nThe metadata is identical to <a href=\"#type_video.movie\">video.movie<\/a>.<a id=\"no_vertical\" href=\"#no_vertical\">No Vertical<\/a>These are globally defined objects that just don't fit into a vertical but\nyet are broadly used and agreed upon.<code>og:type<\/code> values:<a name=\"type_article\" href=\"#type_article\" id=\"type_article\"><code>article<\/code><\/a> - Namespace URI: <a href=\"http:\/\/ogp.me\/ns\/article\"><code>http:\/\/ogp.me\/ns\/article#<\/code><\/a><code>article:published_time<\/code> - <a href=\"#datetime\">datetime<\/a> - \nWhen the article was first published.\n<code>article:modified_time<\/code> - <a href=\"#datetime\">datetime<\/a> - \nWhen the article was last changed.\n<code>article:expiration_time<\/code> - <a href=\"#datetime\">datetime<\/a> - \nWhen the article is out of date after.\n<code>article:author<\/code> - <a href=\"#type_profile\">profile<\/a> <a href=\"#array\">array<\/a> -\nWriters of the article.\n<code>article:section<\/code> - <a href=\"#string\">string<\/a> - A high-level section name. E.g. Technology\n<code>article:tag<\/code> - <a href=\"#string\">string<\/a> <a href=\"#array\">array<\/a> -\nTag words associated with this article.\n<a name=\"type_book\" href=\"#type_book\" id=\"type_book\"><code>book<\/code><\/a> - Namespace URI: <a href=\"http:\/\/ogp.me\/ns\/book\"><code>http:\/\/ogp.me\/ns\/book#<\/code><\/a><code>book:author<\/code> - <a href=\"#type_profile\">profile<\/a> <a href=\"#array\">array<\/a> - Who wrote this book.\n<code>book:isbn<\/code> - <a href=\"#string\">string<\/a> -\nThe <a href=\"http:\/\/en.wikipedia.org\/wiki\/International_Standard_Book_Number\">ISBN<\/a>\n<code>book:release_date<\/code> - <a href=\"#datetime\">datetime<\/a> - The date the book was released.\n<code>book:tag<\/code> - <a href=\"#string\">string<\/a> <a href=\"#array\">array<\/a> -\nTag words associated with this book.\n<a name=\"type_profile\" href=\"#type_profile\" id=\"type_profile\"><code>profile<\/code><\/a> - Namespace URI: <a href=\"http:\/\/ogp.me\/ns\/profile\"><code>http:\/\/ogp.me\/ns\/profile#<\/code><\/a><code>profile:first_name<\/code> - <a href=\"#string\">string<\/a> - A name normally given to an individual by a parent or self-chosen.\n<code>profile:last_name<\/code> - <a href=\"#string\">string<\/a> - A name inherited from a family or marriage and by which the individual is commonly known.\n<code>profile:username<\/code> - <a href=\"#string\">string<\/a> - A short unique string to identify them.\n<code>profile:gender<\/code> - <a href=\"#enum\">enum<\/a>(male, female) - Their gender.\n<a name=\"type_website\" href=\"#type_website\" id=\"type_website\"><code>website<\/code><\/a> - Namespace URI: <a href=\"http:\/\/ogp.me\/ns\/website\"><code>http:\/\/ogp.me\/ns\/website#<\/code><\/a>No additional properties other than the basic ones.\nAny non-marked up webpage should be treated as <code>og:type<\/code> website.The following types are used when defining attributes in Open Graph protocol.<b>Type<\/b>\n  <b>Description<\/b>\n  <b>Literals<\/b>\n<a name=\"bool\" href=\"#bool\" id=\"bool\">Boolean<\/a>\n  A Boolean represents a true or false value\n  true, false, 1, 0\n<a name=\"datetime\" href=\"#datetime\" id=\"datetime\">DateTime<\/a>\n  A DateTime represents a temporal value composed of a date\n    (year, month, day) and an optional time component (hours, minutes)\n  <a href=\"http:\/\/en.wikipedia.org\/wiki\/ISO_8601\">ISO 8601<\/a>\n<a name=\"enum\" href=\"#enum\" id=\"enum\">Enum<\/a>\n  A type consisting of bounded set of constant string values \n  (enumeration members).\n  A string value that is a member of the enumeration\n<a name=\"float\" href=\"#float\" id=\"float\">Float<\/a>\n  A 64-bit signed floating point number\n  All literals that conform to the following formats:1.234<br \/>-1.234<br \/>1.2e3<br \/>-1.2e3<br \/>7E-10<br \/>\n<a name=\"integer\" href=\"#integer\" id=\"integer\">Integer<\/a>\n  A 32-bit signed integer. In many languages integers over 32-bits become \n    floats, so we limit Open Graph protocol for easy multi-language use.\n  All literals that conform to the following formats:1234<br \/>-123<br \/>\n<a name=\"string\" href=\"#string\" id=\"string\">String<\/a>\n  A sequence of Unicode characters\n  All literals composed of Unicode characters with no escape characters\n<a name=\"url\" href=\"#url\" id=\"url\">URL<\/a>\n  A sequence of Unicode characters that identify an Internet resource.\n  All valid URLs that utilize the http:\/\/ or https:\/\/ protocols\nYou can discuss the Open Graph Protocol in\n<a href=\"https:\/\/www.facebook.com\/groups\/opengraph\/\">the Facebook group<\/a> or on \n<a href=\"http:\/\/groups.google.com\/group\/open-graph-protocol\">the developer mailing list<\/a>.\nIt is currently being consumed by Facebook\n(<a href=\"http:\/\/developers.facebook.com\/docs\/opengraph\/\">see their documentation<\/a>), Google (<a href=\"https:\/\/developers.google.com\/+\/web\/+1button\/#plus-snippet\">see their documentation<\/a>), and\n<a href=\"http:\/\/developer.mixi.co.jp\/en\/connect\/mixi_plugin\/mixi_check\/spec_mixi_check\/\">mixi<\/a>.\nIt is being published by IMDb, Microsoft, NHL, Posterous, Rotten Tomatoes,\nTIME, Yelp, and many many others.The open source community has developed a number of parsers and publishing\ntools. Let the Facebook group know if you've built something awesome too!<a href=\"http:\/\/developers.facebook.com\/tools\/debug\/\">Facebook Object Debugger<\/a> -\nFacebook's official parser and debugger\n<a href=\"http:\/\/www.google.com\/webmasters\/tools\/richsnippets\">Google Rich Snippets Testing Tool<\/a> - Open Graph protocol support in specific verticals and Search Engines.\n<a href=\"https:\/\/github.com\/niallkennedy\/open-graph-protocol-tools\">PHP Validator and Markup Generator<\/a> -  OGP 2011 input validator and markup generator in PHP5 objects\n<a href=\"http:\/\/github.com\/scottmac\/opengraph\">PHP Consumer<\/a> -\na small library for accessing of Open Graph Protocol data in PHP\n<a href=\"http:\/\/buzzword.org.uk\/2010\/opengraph\/#php\">OpenGraphNode in PHP<\/a> -\na simple parser for PHP\n<a href=\"http:\/\/pypi.python.org\/pypi\/PyOpenGraph\">PyOpenGraph<\/a> -\na library written in Python for parsing Open Graph protocol\ninformation from web sites\n<a href=\"http:\/\/github.com\/intridea\/opengraph\">OpenGraph Ruby<\/a> -\nRuby Gem which parses web pages and extracts Open Graph protocol markup\n<a href=\"http:\/\/github.com\/callumj\/opengraph-java\">OpenGraph for Java<\/a> -\nsmall Java class used to represent the Open Graph protocol\n<a href=\"http:\/\/search.cpan.org\/~tobyink\/RDF-RDFa-Parser\/lib\/RDF\/RDFa\/Parser.pm\">RDF::RDFa::Parser<\/a> -\nPerl RDFa parser which understands the Open Graph protocol\n<a href=\"http:\/\/wordpress.org\/plugins\/facebook\/\">WordPress plugin<\/a> -\nFacebook's official WordPress plugin, which adds Open Graph metadata to WordPress powered sites. \n<a href=\"http:\/\/wordpress.org\/plugins\/wp-facebook-open-graph-protocol\/\">Alternate WordPress OGP plugin<\/a> -\n  A simple lightweight WordPress plugin which adds Open Graph metadata to WordPress powered sites.<\/h2>","created_at":"2018-11-07T16:21:44+0000","updated_at":"2018-11-07T16:21:44+0000","published_at":"2017-07-27T00:00:00+0000","annotations":[],"mimetype":"text\/html","reading_time":10,"domain_name":"ogp.me","preview_picture":"http:\/\/ogp.me\/logo.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":4,"title":"How to include the WordPress media selector in your plugin | Jeroen Sormani","url":"https:\/\/jeroensormani.com\/how-to-include-the-wordpress-media-selector-in-your-plugin\/","content":"<div class=\"entry-featured-image\"><img width=\"1500\" height=\"840\" src=\"https:\/\/jeroensormani.com\/content\/uploads\/2015\/06\/use-media-selector-in-your-plugin.jpg\" class=\"attachment-post-thumbnail size-post-thumbnail wp-post-image\" alt=\"\" srcset=\"https:\/\/jeroensormani.com\/content\/uploads\/2015\/06\/use-media-selector-in-your-plugin.jpg 1500w, https:\/\/jeroensormani.com\/content\/uploads\/2015\/06\/use-media-selector-in-your-plugin-300x168.jpg 300w, https:\/\/jeroensormani.com\/content\/uploads\/2015\/06\/use-media-selector-in-your-plugin-1024x573.jpg 1024w\" \/><\/div><p>Recently I had to include a image selector in one of the plugins I was developing. Going for the pure WordPress look and feel, I really wanted to use the image selector everyone knows from the \u2018Add Media\u2019 button when you modify a post. To accomplish this, you\u2019d need to go through a couple of steps I will lay out here.<\/p><h3 id=\"create-a-settings-page\">Create a settings page<\/h3><p>You may already have a settings page ready where you want to include the media selector on, if so, you can skip this step and go to the next one. Use the following code to add a new submenu page to the \u2018Settings\u2019 menu.<br \/><\/p><p>What this does is add a \u2018Media selector\u2019 link to the settings page,\u00a0and uses the \u2018media_selector_settings_page_callback\u2019 function to display the setting fields.<\/p><p>And just to be clear, this will work, but is anything but the best way of adding a settings page. Please use the Settings API for that, and if you want to make it easy for yourself, generate a settings page with the <a href=\"http:\/\/wpsettingsapi.jeroensormani.com\/\">WP Settings API tool<\/a>.<\/p><h3 id=\"add-the-settings-fields\">Add the settings fields<\/h3><p>In order for the media selector to fully work, you need at least a upload button and a input field. The input field can be a text field, but also a hidden field. In this post we\u2019re inserting the post ID of the attachment that is being selected\/uploaded, so I chose for a\u00a0hidden field.<\/p><h3 id=\"include-the-java-scripts\">Include\u00a0the (java)scripts<\/h3><p>Right now you should have a page that you can visit, and see a button and a empty image. When clicking the button, nothing will happen as we still need to include some scripts. WordPress has made this very easy for us, and you can use the following code for that<\/p><p>For the sake of this post I will print the scripts directly in the footer, but PLEASE! <a href=\"https:\/\/codex.wordpress.org\/Plugin_API\/Action_Reference\/admin_enqueue_scripts\">enqueue your javascript<\/a> in your plugin. The following script will do a couple of things;<\/p><ol><li>On the \u2018#upload_image_button\u2019 button click, it will open the media selector<\/li>\n<li>Set the image preview \u2018src\u2019 so the preview will be immediately visible<\/li>\n<li>Set the attachment ID in the hidden input field for us to save later<\/li>\n<\/ol><h3 id=\"save-the-selected-image\">Save the selected image<\/h3><p>I just realised I was missing a &lt;form&gt; tag and a save button on my settings page, so I added that below. Also adding a bit of code to save the attachment ID, and pre-filling the preview and input field when they are saved.<\/p><h3 id=\"done-\">Done.<\/h3><p>That should be it, you will now have a settings page, with a official WordPress media selector allowing you to select and save the attachment ID. I want to say again that the code is used to give a example on how to implement the WordPress media selector, not on how to create a proper settings page, enqueue scripts or anything.<\/p><p>Here\u2019s a snippet with my entire test plugin;<\/p><h3 id=\"thank-you\">Thank you<\/h3><p>A big and special thank you to <strong><a href=\"http:\/\/mikejolley.com\/2012\/12\/using-the-new-wordpress-3-5-media-uploader-in-plugins\/\">Mike Jolley<\/a><\/strong>\u00a0where I initially learned this from. The Javascript in the post comes from him.<\/p>","created_at":"2018-11-07T16:21:43+0000","updated_at":"2018-11-07T16:21:43+0000","published_at":"2015-09-25T09:34:06+0000","annotations":[],"mimetype":"text\/html","reading_time":2,"domain_name":"jeroensormani.com","preview_picture":"https:\/\/jeroensormani.com\/content\/uploads\/2015\/06\/use-media-selector-in-your-plugin.jpg","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["emacs","mu4e"],"is_public":false,"id":5,"title":"Sending HTML Mail with mu4e","url":"https:\/\/matt.hackinghistory.ca\/2016\/11\/18\/sending-html-mail-with-mu4e\/","content":"<p> John Kitchin <a href=\"http:\/\/kitchingroup.cheme.cmu.edu\/blog\/2016\/10\/29\/Sending-html-emails-from-org-mode-with-org-mime\/\">has a terrific post<\/a> detailing some configuration\/improvements to mu4e that make it easier to send html mail.  This original post is really great, but I ran into quite a bit of trouble following this advice. <\/p>\n<p> He uses a cool feature of mu4e,  <code>org-mu4e-compose-org-mode<\/code>, which toggles the major mode of  the message buffer between org-mode (when you\u2019re in the message body) and mu4e-compose-mode (when you\u2019re in the headers area).  With a couple of custom functions, it\u2019s easy to convert the org text to html and send a mime-multipart email from Emacs, which is quite convenient.  If you add <code>org-mu4e-compose-org-mode<\/code> as a hook to <code>mu4e-compose-mode<\/code>, you can compose in html by default, which is great. <\/p>\n<p> Unfortunately, <code>org-mu4e-compose-org-mode<\/code> is <b>deprecated<\/b> on account of its instability, and while John doesn\u2019t have any problems with it, for me it was unworkable. It turns out that this \u201cmode\u201d isn\u2019t really a standard emacs mode at all \u2013 instead, it\u2019s a sly workaround that trickily adds an internal function to the \u2019post-command-hook in the draft buffer and switches major modes based on position.  This is a neat hack, but since the function <i>invokes the major modes directly<\/i>, setting thefunction as a hook to  <code>mu4e-compose-mode<\/code> leads to some funky, inadvertent looping effects. On my machine, for some reason, those effects send the underlying <code>message-send<\/code> function crazy, and instead of sending directly, I get this amazingly annoying question: <\/p>\n<blockquote>\n<p> Already sent message via mail; resend? (y or n) y <\/p>\n<\/blockquote>\n<p> On its own, that is already annoying; but worse, the sent message doesn\u2019t get saved to my Sent folder. Instead, it\u2019s lost completely. <\/p>\n<p> Anyway, after fruitless hours of paying around with this, I realized that the problem could be fixed by adding a new hook to the mu4e compose functions (rather than to the compose <i>mode<\/i>).  I\u2019ve <a href=\"https:\/\/github.com\/djcb\/mu\/pull\/952\">submitted those changes as a pull request<\/a> and hopefully they will be accepted; if not, though, feel free to navigate back to <a href=\"https:\/\/github.com\/titaniumbones\/mu\/tree\/org-mu4e-compose-stabilization\">my branch<\/a> and pull\/install mu from there.  With those small changes, I now have frictionless html email working very quickly within emacs, using this small bit of code.  It is at least 90% stolen: <\/p>\n<div class=\"org-src-container\">\n<pre class=\"src src-emacs-lisp\">;; this is stolen from John but it didn't work for me until I\n;; made those changes to mu4e-compose.el\n(defun htmlize-and-send ()\n  \"When in an org-mu4e-compose-org-mode message, htmlize and send it.\"\n  (interactive)\n  (when (member 'org~mu4e-mime-switch-headers-or-body post-command-hook)\n    (org-mime-htmlize)\n    (org-mu4e-compose-org-mode)\n    (mu4e-compose-mode)\n    (message-send-and-exit)))\n\n\n;; This overloads the amazing C-c C-c commands in org-mode with one more function\n;; namely the htmlize-and-send, above.\n(add-hook 'org-ctrl-c-ctrl-c-hook 'htmlize-and-send t)\n\n;; Originally, I set the `mu4e-compose-mode-hook' here, but\n;; this new hook works much, much better for me.  \n(add-hook 'mu4e-compose-post-hook\n          (defun do-compose-stuff ()\n            \"My settings for message composition.\"\n            (org-mu4e-compose-org-mode)))\n<\/pre>\n<\/div>\n<p> It feels great to have gotten this far. There are still some small things I\u2019d like to be able to improve; and I think I would like to add a few wrapper functions and keybindings to my setup, but for now I\u2019m pretty efficient. <\/p>\n<p> Still missing: <\/p>\n<ul class=\"org-ul\"><li>a better HTML viewing interface! right now, html mails render as mostly text \u2013 it would be nice to have a rendered html message by default in emacs. This is an issue several times a day when I get promotional emails from organizations I work with \u2013 usually the html part is really important. I can access these in the browser but it\u2019s comparatively awkward.<\/li>\n<li>a way to forward these html emails to someone intact \u2013 right now, the html parts are discarded.  No idea how hard it would be to do this.<\/li>\n<\/ul><p> Thanks John and Dirk-Jan for these great tools! <\/p>","created_at":"2018-11-07T16:21:43+0000","updated_at":"2018-11-12T18:38:02+0000","published_by":["matt"],"annotations":[],"mimetype":"text\/html","reading_time":3,"domain_name":"matt.hackinghistory.ca","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":6,"title":"My Self-Hosted Lifestyle \u2014 jackyalcin\u00e9","url":"https:\/\/jacky.wtf\/weblog\/my-self-hosted-lifestyle\/","content":"<main role=\"group\" class=\"e-content\" itemprop=\"articleBody\" data-track-content=\"\" data-content-name=\"blog-post\" data-content-piece=\"My Self-Hosted Lifestyle\"><p>For some time now, I\u2019ve moved a lot of stuff that\u2019s part of my personal\u00a0routine\ninto my personal server. <a href=\"https:\/\/jacky.wtf\/weblog\/new-infra\/\">After rebuilding Kandor<\/a>, my personal\u00a0server\u2019s\nlayout; I took to adding a few services that I\u2019d find\u00a0useful.<\/p><h2 id=\"nextcloud\">Nextcloud<\/h2>\n<p>I use <a href=\"https:\/\/nextcloud.org\">NextCloud<\/a> to handle a lot of things that Google\u2019s cloud file\u00a0suite\nwould offer. I don\u2019t have collaborative documents - I don\u2019t really use\u00a0that\nanymore. Nor do I handle my e-mail; Google handles it well. I <a href=\"https:\/\/www.jeffkee.com\/self-hosted-emails-are-dying-out-and-why-they-should\/\">don\u2019t agree\u00a0that\ncentralized email services are better<\/a> since it goes against what email\u00a0is\nmeant to\u00a0be.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_090748-d859a71d56da8de4d872c7cc5dca2a1b70998f119b77b64bf2fa9a9ed9edcf46.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>The login screen is mundane, nothing really special. I have my instance set\u00a0up\nwith a few more steps for security reasons so I can use it comfortably\u00a0from\nanywhere I\u2019d like. Once I\u2019m in, I can see the stuff I got lying\u00a0around.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_091013-eaeabc8d48fefd7075d267e86a31c1fe74403c6b1ba399d20cfa03dfd128bfe5.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>This reflects the layout I use in my top-most directory on my host\u00a0machine.\nI link a few of the folders over so it\u2019s a seamless experience for me\u00a0whenever\nI manipulate files on my system. New screenshots? Uploaded. Downloaded a new\u00a0PDF\nor ePub for reading on my tablet? Synced in a few moments to NextCloud\u00a0and\ndown to my\u00a0device.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_091225-73136bd03815fca66421b1456fc9616a55dd57ab1e18f98afa50de63357a35dd.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>A lot of the tasks and actions one takes can be reviewed in the \u201cActivity\u201d\u00a0app\nwindow. Since it\u2019s only me on this instance, this feed tends to be a soft\u00a0audit\nlog for my personal use. It\u2019s nice to see and check in - especially if I\u2019m\u00a0too\nlazy to open the app and see what\u2019s synced; I occasionally just pop this\u00a0open.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_091435-80bbc8d6492d5a59724522fe618b6bd6f931bb04cfd35b0f52ef5fafb86ae533.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>One of the applications within NextCloud I like using a lot is the \u201cNotes\u201d\u00a0app.\nIt\u2019s ironic since I really only use it for viewing stuff on the Web. If I\u2019m\u00a0out\nor not in the mood to use my automatic note taking setup in NeoVim, I pop\u00a0this\nopen and begin typing away! Having it automatically handle things like lists\u00a0and\nheadings makes it a daily driver for\u00a0me.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_091743-4cafbbb9945472ec01f33f8b5822594308de5c52ccfa6fbf91f64f088da949a4.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>Out of all of the services within NextCloud that I use the most, the\u00a0news\nreading application is definitely <em>way<\/em> up there. Not only can I\u00a0handle\nRSS and ATOM feeds in here; I can handle my podcast subscriptions.\u00a0This\nmakes it very easy for to just pop some in when I\u2019m on my way to the\u00a0gym,\ncooking or just working on the computer and don\u2019t want music in my\u00a0ears.<\/p>\n<p>All of these things from NextCloud as well as its ability to extend\u00a0are\nthings that keep me tethered to the service. I used to use OwnCloud\u00a0prior\nto the forking. NextCloud has managed to keep the quality of service\u00a0and\nsoftware high without compromising on\u00a0performance.<\/p>\n<h2 id=\"wallabag\">Wallabag<\/h2>\n<p><a href=\"https:\/\/www.wallabag.org\/en\/\">Wallabag<\/a> is a super-simple, \u201cdo one thing right\u201d kind of\u00a0application.\nWith a few browser extensions, it ingests URIs (link to blogs or what\u00a0have\nyou) and turns them into more simpler reading formats for later use.\u00a0This\nis something I\u2019ve done a lot but didn\u2019t want to send out to other\u00a0services\nto track and manipulate how I use and read information on the\u00a0Web.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_092344-66d1d183a86934591407f13ca6378ff278d1deb7bed799e1c903e57ce0fdbf36.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>There\u2019s not much more to say about this application. On Android, namely\u00a0my\ntablet, I can read through my entries like flicking through pages\u00a0of\na newspaper when I\u2019m at home, on the go or just looking to pass\u00a0time.<\/p>\n<h2 id=\"peertube\">PeerTube<\/h2>\n<p><a href=\"http:\/\/joinpeertube.org\/en\/\">PeerTube<\/a> is a project funded by the same people behind <a href=\"https:\/\/www.wallabag.org\/en\/\">Wallabag<\/a>.\nI\u2019ve been using it recently in a very exploratory fashion. The bulk of\u00a0my\nuploads are recordings from my live streams or imported videos\u00a0from\nelsewhere on the Internet that I\u2019d like to\u00a0keep.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_093456-e2956a5ff8d37b88eef88b90477d4193b0755bcff1784856357ed738ed7f58ff.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<p>I have some <em>sort<\/em> of curation going on. I guess? For now, I\u2019ve\u00a0been\nopting to keep a lot of the videos in the vein of imported from\u00a0YouTube\nwith crediting or something I\u2019ve crafted\u00a0myself.<\/p>\n<p><img class=\"db w-100 h-auto\" src=\"https:\/\/jacky.wtf\/assets\/posts\/Screenshot_20181027_093550-c87c700bdde787e048076b8f27cab90bc4ffcae45cd6ee95c8cd8e1b40d314a3.png\" crossorigin=\"anonymous\" alt=\"image\" \/><\/p>\n<h2 id=\"thats-all\">That\u2019s\u00a0All!<\/h2>\n<p>I\u2019m aiming to make this post the landing page for a <code class=\"highlighter-rouge\">\/hosted<\/code> page\u00a0or\nsomething when I get enough time and handling around things. I try to\u00a0help\npromote self-hosted solutions or just anything that isn\u2019t\u00a0heavily\ncentralized to promote safety and diversity of the toolkits we\u00a0have\navailable on the\u00a0Web.<\/p><\/main>","created_at":"2018-11-07T16:21:43+0000","updated_at":"2018-11-07T16:21:43+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":3,"domain_name":"jacky.wtf","preview_picture":"https:\/\/jacky.wtf\/assets\/196x196-a66b5e433f41dccff47073bb72df31ca88c9163467a3ef48618d115e1f4cee0c.png","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":["css","html","web","icons"],"is_public":false,"id":3,"title":"akveo\/eva-icons","url":"https:\/\/github.com\/akveo\/eva-icons","content":"<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/camo.githubusercontent.com\/e8ce30f25955aef4382bc58e9a0f6225ae9fc1a0\/68747470733a2f2f692e696d6775722e636f6d2f6f446d71446b452e6a7067\"><img src=\"https:\/\/camo.githubusercontent.com\/e8ce30f25955aef4382bc58e9a0f6225ae9fc1a0\/68747470733a2f2f692e696d6775722e636f6d2f6f446d71446b452e6a7067\" data-canonical-src=\"https:\/\/i.imgur.com\/oDmqDkE.jpg\" alt=\"image\" \/><\/a><\/p><p><strong>Eva Icons<\/strong> is a pack of more than 480 beautifully crafted Open Source icons for common actions and items. Additionally Eva Icons supports 4 animation types: <code>zoom<\/code>, <code>pulse<\/code>, <code>shake<\/code> and <code>flip<\/code>.  Download on desktop to use them in your digital products for Web, iOS and Android. Icons are provided in two visual types: <code>Fill<\/code> and <code>Outline<\/code> and in serveral formats, including <code>PNG<\/code>, <code>SVG<\/code>, <code>font<\/code>, <code>Sketch<\/code>, etc.<\/p><h2>Download<\/h2><ul><li><a href=\"https:\/\/akveo.github.io\/eva-icons\/eva-icons.zip\" rel=\"nofollow\">Complete Eva Icons Package<\/a><\/li>\n<li><a href=\"https:\/\/www.npmjs.com\/package\/eva-icons\" rel=\"nofollow\">NPM Package<\/a><\/li>\n<li>you also can download icons one by one in <code>PNG<\/code> and <code>SVG<\/code> formats from <a href=\"https:\/\/akveo.github.io\/eva-icons\/\" rel=\"nofollow\">Eva Icons Website<\/a>.<\/li>\n<\/ul><h3>CDN<\/h3><p>Load from CDN in your project:<\/p><div class=\"highlight highlight-text-html-basic\"><pre>&lt;script src=\"https:\/\/unpkg.com\/eva-icons\"&gt;&lt;\/script&gt;<\/pre><\/div><p>After including the script, <code>eva<\/code> will be available as a global variable.<\/p><h3>NPM<\/h3><ul><li>Install the package:<\/li>\n<\/ul><pre>npm i eva-icons\n<\/pre><ul><li>Include it to your page:<\/li>\n<\/ul><div class=\"highlight highlight-text-html-basic\"><pre>&lt;script src=\"path\/to\/dist\/eva-icons.js\"&gt;&lt;\/script&gt;<\/pre><\/div><ul><li>Or require the package (may vary depending on your build system):<\/li>\n<\/ul><div class=\"highlight highlight-source-js\"><pre>const eva = require('eva-icons');<\/pre><\/div><div class=\"highlight highlight-source-js\"><pre>import * as eva from 'eva-icons';<\/pre><\/div><h2>How to use<\/h2><h3>JavaScript<\/h3><ul><li>Add the <code>data-eva<\/code> attribute with the icon name to an element:<\/li>\n<\/ul><div class=\"highlight highlight-text-html-basic\"><pre>&lt;i data-eva=\"github\"&gt;&lt;\/i&gt;<\/pre><\/div><ul><li>Call <code>eva.replace();<\/code> to replace all elements with the <code>data-eva<\/code> data attribute with SVG elements. You can also pass some additional parameters to the <code>replace<\/code> method to modify the <code>replace<\/code> function behavior.<\/li>\n<\/ul><div class=\"highlight highlight-text-html-basic\"><pre>&lt;!DOCTYPE html&gt;\n&lt;html lang=\"en\"&gt;\n  &lt;title&gt;&lt;\/title&gt;\n  &lt;script src=\"https:\/\/unpkg.com\/eva-icons\"&gt;&lt;\/script&gt;\n  &lt;body&gt;\n  \n    &lt;i data-eva=\"github\"&gt;&lt;\/i&gt;\n    &lt;script&gt;\n      eva.replace()\n    &lt;\/script&gt;\n  &lt;\/body&gt;\n&lt;\/html&gt;<\/pre><\/div><p><em>Thanks to Feather Icons for the build process inspiration.<\/em><\/p><h3>Fonts<\/h3><p>Eva Icons are also available as a Web Font.<\/p><ul><li>Include the font css into your page:<\/li>\n<\/ul><div class=\"highlight highlight-text-html-basic\"><pre>&lt;link href=\"path\/to\/style\/eva-icons.css\"&gt;<\/pre><\/div><ul><li>Add <code>eva<\/code> and <code>eva-icon<\/code> classes to an element:<\/li>\n<\/ul><div class=\"highlight highlight-text-html-basic\"><pre>&lt;i class=\"eva eva-github\"&gt;&lt;\/i&gt;<\/pre><\/div><p>We recommend using SVG icons due to better rendering and performance capabilities, <a href=\"https:\/\/css-tricks.com\/icon-fonts-vs-svg\/\" rel=\"nofollow\">more details<\/a>.<\/p><h2>Documentation<\/h2><h3><code>eva.replace({ ... })<\/code><\/h3><p>Replaces all elements that have a <code>data-eva<\/code> attribute with SVG markup corresponding to the element's <code>data-eva<\/code> attribute value.<\/p><h2>License<\/h2><p><a href=\"https:\/\/github.com\/akveo\/eva-icons\/blob\/master\/LICENSE.txt\">MIT<\/a> license.<\/p><h2>More from Akveo<\/h2><ul><li><a href=\"https:\/\/github.com\/akveo\/nebular\">Nebular<\/a> - Angular Component, Auth and Security<\/li>\n<li><a href=\"https:\/\/github.com\/akveo\/ngx-admin\">ngx-admin<\/a> - the best Angular admin template<\/li>\n<\/ul><h2>How can I support the developers?<\/h2><ul><li>Star our GitHub repo \u2b50\ufe0f<\/li>\n<li>Create pull requests, submit bugs, suggest new features or documentation updates \ud83d\udd27<\/li>\n<li>Follow us on <a href=\"https:\/\/twitter.com\/akveo_inc\" rel=\"nofollow\">Twitter<\/a> \ud83d\udc3e<\/li>\n<li>Like our page on <a href=\"https:\/\/www.facebook.com\/akveo\/\" rel=\"nofollow\">Facebook<\/a> \ud83d\udc4d<\/li>\n<\/ul><h2>From Developers<\/h2><p>Made with \u2764\ufe0f by <a href=\"http:\/\/akveo.com?utm_source=github&amp;utm_medium=nebular_readme\" rel=\"nofollow\">Akveo team<\/a>. Follow us on <a href=\"https:\/\/twitter.com\/akveo_inc\" rel=\"nofollow\">Twitter<\/a> to get the latest news first!\nWe're always happy to receive your feedback!<\/p>","created_at":"2018-11-07T16:21:42+0000","updated_at":"2018-11-12T15:20:38+0000","published_by":[""],"annotations":[],"mimetype":"text\/html","language":"en","reading_time":2,"domain_name":"github.com","preview_picture":"https:\/\/avatars3.githubusercontent.com\/u\/11921694?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":1,"title":"wallabag\/wallabag","url":"https:\/\/github.com\/wallabag\/wallabag","content":"<article class=\"markdown-body entry-content\" itemprop=\"text\"><p><a href=\"https:\/\/travis-ci.org\/wallabag\/wallabag\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/678ae0b6ed76a31216eed883721071278281cf1d\/68747470733a2f2f6170692e7472617669732d63692e6f72672f77616c6c616261672f77616c6c616261672e7376673f6272616e63683d6d6173746572\" alt=\"Build Status\" data-canonical-src=\"https:\/\/api.travis-ci.org\/wallabag\/wallabag.svg?branch=master\" \/><\/a>\n<a href=\"https:\/\/scrutinizer-ci.com\/g\/wallabag\/wallabag\/?branch=master\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/51b9e9a418c82619beb66f037da0ee0fac19da8c\/68747470733a2f2f7363727574696e697a65722d63692e636f6d2f672f77616c6c616261672f77616c6c616261672f6261646765732f7175616c6974792d73636f72652e706e673f623d6d6173746572\" alt=\"Scrutinizer Code Quality\" data-canonical-src=\"https:\/\/scrutinizer-ci.com\/g\/wallabag\/wallabag\/badges\/quality-score.png?b=master\" \/><\/a>\n<a href=\"https:\/\/gitter.im\/wallabag\/wallabag\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/7d924c483d0d9dff10763921aea2038e660e1252\/68747470733a2f2f6261646765732e6769747465722e696d2f67697474657248512f6769747465722e737667\" alt=\"Gitter\" data-canonical-src=\"https:\/\/badges.gitter.im\/gitterHQ\/gitter.svg\" \/><\/a><\/p>\n<p>wallabag is a self-hostable PHP application allowing you to not miss any content anymore.\nClick, save and read it when you can. It extracts content so that you can read it when you have time.<\/p>\n<p>More information on our website: <a href=\"https:\/\/wallabag.org\" rel=\"nofollow\">wallabag.org<\/a>.<\/p>\n<p>If you do not have your own server, consider <a href=\"https:\/\/wallabag.it\" rel=\"nofollow\">the wallabag.it hosting solution<\/a>.<\/p>\n<p><a target=\"_blank\" rel=\"noopener noreferrer\" href=\"https:\/\/raw.githubusercontent.com\/wallabag\/logo\/master\/_default\/typo-horizontal\/png\/sm\/logo-typo-horizontal-black-no-bg-no-border-sm.png\"><img src=\"https:\/\/raw.githubusercontent.com\/wallabag\/logo\/master\/_default\/typo-horizontal\/png\/sm\/logo-typo-horizontal-black-no-bg-no-border-sm.png\" alt=\"wallabag logo\" \/><\/a><\/p>\n<p>Please read <a href=\"https:\/\/doc.wallabag.org\/en\/admin\/installation\/requirements.html\" rel=\"nofollow\">the documentation to see the wallabag requirements<\/a>.<\/p>\n<p>Then you can install wallabag by executing the following commands:<\/p>\n<pre>git clone https:\/\/github.com\/wallabag\/wallabag.git\ncd wallabag &amp;&amp; make install \n<\/pre>\n<p>Now, <a href=\"https:\/\/doc.wallabag.org\/en\/admin\/installation\/virtualhosts.html\" rel=\"nofollow\">configure a virtual host<\/a> to use your wallabag.<\/p>\n<p><a href=\"https:\/\/install-app.yunohost.org\/?app=wallabag2\" rel=\"nofollow\"><img src=\"https:\/\/camo.githubusercontent.com\/a657193184c779d657ef3f1e2834590d6615a950\/68747470733a2f2f696e7374616c6c2d6170702e79756e6f686f73742e6f72672f696e7374616c6c2d776974682d79756e6f686f73742e706e67\" alt=\"Install Wallabag with YunoHost\" data-canonical-src=\"https:\/\/install-app.yunohost.org\/install-with-yunohost.png\" \/><\/a><\/p>\n<p>Wallabag app for <a href=\"https:\/\/yunohost.org\" rel=\"nofollow\">YunoHost<\/a>. See <a href=\"https:\/\/github.com\/YunoHost-Apps\/wallabag2_ynh\">here<\/a><\/p>\n<p>Copyright \u00a9 2013-2018 Nicolas L\u0153uillet <a href=\"mailto:nicolas@loeuillet.org\">nicolas@loeuillet.org<\/a>\nThis work is free. You can redistribute it and\/or modify it under the\nterms of the MIT License. See the COPYING file for more details.<\/p>\n<\/article>","created_at":"2018-11-07T16:21:41+0000","updated_at":"2018-11-07T16:21:41+0000","annotations":[],"mimetype":"text\/html","language":"en","reading_time":0,"domain_name":"github.com","preview_picture":"https:\/\/avatars0.githubusercontent.com\/u\/4143872?s=400&v=4","http_status":"200"},{"is_archived":1,"is_starred":0,"tags":[],"is_public":false,"id":2,"title":"Using mu4e \u00b7 Brool","url":"http:\/\/www.brool.com\/post\/using-mu4e\/","content":"<p>January 19, 2013\u00a0|\u00a0\n  \n  <a href=\"http:\/\/www.brool.com\/tags\/coding\" class=\"tag-link\">coding<\/a>\n  \n  <a href=\"http:\/\/www.brool.com\/tags\/linux\" class=\"tag-link\">linux<\/a><\/p><p>A thread on\n<a href=\"http:\/\/www.reddit.com\/r\/emacs\/comments\/16ju6p\/what_package_do_people_use_to_read_imap_mail_now\/\">Reddit<\/a>\nnudged me to try the Emacs mail readers again. I have tried a bunch of\nthem (Wanderlust, Gnus, mew) but found them either too idiosyncratic or\npainful to use, and for the most part just read my e-mail by logging\ninto Google directly.<\/p><p>However, I just installed mu \/ mu4e and it works really well \u2014 well\nenough that I prefer reading e-mail in Emacs now, as opposed to just\nlogging into Google mail and reading it there. I ended up archiving\nabout 10 years of mail (approximately 10k of messages) off Gmail and mu\ncan re-index it in seconds.<\/p><h2 id=\"installing-offlineimap\">Installing offlineimap<\/h2><p>I used offlineimap to sync everything with Gmail. <a href=\"http:\/\/www.djcbsoftware.nl\/code\/mu\/mu4e\/Gmail-configuration.html#Gmail-configuration\">Steal this\nconfiguration<\/a>\nand put it in ~\/.offlineimaprc (or whatever \u2014 but don\u2019t name it\n.offlineimap, because that will cause errors). For my laptop I added:<\/p><div class=\"highlight\"><pre class=\"language-bash\" data-lang=\"bash\">maxage = 5\nautorefresh = 15\n<\/pre><\/div><p>\u2026 to keep only 5 days of mail and to automatically update it every 15\nminutes.<\/p><p>Updated 2016-12-21: Well, scratch that. Unless you want to allow\nunsafe applications to access your data, you\u2019re going to have to use Oauth.<\/p><p>The instructions are in the comments of <a href=\"https:\/\/github.com\/OfflineIMAP\/offlineimap\/blob\/master\/offlineimap.conf\">this configuration file<\/a>, but for posterity:<\/p><ol><li>Go to the <a href=\"https:\/\/console.developers.google.com\/project\">Google Developer Console<\/a><\/li>\n<li>Create a new project<\/li>\n<li>On left bar, select Credentials<\/li>\n<li>Go to OAuth consent screen and fill in the minimum necessary to submit the form<\/li>\n<li>Go to Credentials tab and select \u201cOauth Client ID\u201d<\/li>\n<li>Choose \u201cother\u201d application type and enter an appropriate name<\/li>\n<\/ol><p>So that gives you the <em>client ID<\/em> and the <em>client secret<\/em>.  Wait, you\u2019re not done. You need a refresh token.<\/p><ol><li>Clone out <a href=\"https:\/\/github.com\/google\/gmail-oauth2-tools\">https:\/\/github.com\/google\/gmail-oauth2-tools<\/a>.<\/li>\n<li>Run <code>python\/oauth2.py --generate_oauth2_token --client_id=yourclient --client_secret=yoursecret<\/code><\/li>\n<li>This will have you go to your browser to get a verification token that you then enter. That will give you a refresh token.<\/li>\n<\/ol><p>In your <code>.offlineimaprc<\/code>, you\u2019ll have something like:<\/p><pre>[Repository Personal_Remote]\n...\noauth2_client_id = longstringofdigitsandnumbers.apps.googleusercontent.com\noauth2_client_secret = longstringofdigitsandnumbers\noauth2_request_url = https:\/\/accounts.google.com\/o\/oauth2\/token\noauth2_refresh_token = 1\/longstringofdigitsandnumbers\n<\/pre><h2 id=\"installing-mu4e\">Installing mu4e<\/h2><p>The version of mu4e available in Ubuntu 12.04LTS is really old and\nthrows spurious errors, so you\u2019ll need to build your own. <a href=\"https:\/\/github.com\/djcb\/mu\">Get mu from\ngithub<\/a> and then rebuild. I needed a few\nextra libraries in my installation.<\/p><div class=\"highlight\"><pre class=\"language-bash\" data-lang=\"bash\">sudo apt-get install libglib2.0-dev libgmime-2.6-dev libxapian-dev\nsudo apt-get install texinfo\nautoreconf -i\n<\/pre><\/div><h2 id=\"installing-into-emacs\">Installing into Emacs<\/h2><p>Change into mu4e, do a make \/ sudo make install. I used pretty much just\nthe <a href=\"http:\/\/www.djcbsoftware.nl\/code\/mu\/mu4e\/Longer-configuration.html\">longer configuration in the\ndocumentation<\/a>,\nalthough I put everything into a deferred load (code below), and there\nare the following caveats:<\/p><ul><li>Viewing a message in a browser is handy, so I took the code from the\n<a href=\"http:\/\/www.emacswiki.org\/emacs\/mu4e\">Emacs Wiki<\/a>.<\/li>\n<li>I needed to override mu4e-html2text-command because about 10% of\nHTML messages weren\u2019t rendering.<\/li>\n<li>The\n<a href=\"http:\/\/www.djcbsoftware.nl\/code\/mu\/mu4e\/Viewing-images-inline.html#Viewing-images-inline\">documentation<\/a>\nsays to set mu4e-view-show-images to show images, but it\u2019s actually\nmu4e-show-images.<\/li>\n<\/ul><div class=\"highlight\"><pre class=\"language-common-lisp\" data-lang=\"common-lisp\">(defun email () \n  (interactive)\n  (when (not (featurep 'mu4e))\n    (add-to-list 'load-path \"\/usr\/local\/share\/emacs\/site-lisp\/mu4e\/\")\n    (require 'mu4e)\n    (require 'org-mu4e)\n    ;; defaults\n    (setq mu4e-maildir \"~\/Maildir\")\n    (setq mu4e-drafts-folder \"\/gmail\/[Gmail].Drafts\")\n    (setq mu4e-sent-folder   \"\/gmail\/[Gmail].Sent Mail\")\n    (setq mu4e-trash-folder  \"\/gmail\/[Gmail].Trash\")\n    ;; don't save message to Sent Messages, Gmail\/IMAP takes care of this\n    (setq mu4e-sent-messages-behavior 'delete)\n    ;; setup some handy shortcuts\n    ;; you can quickly switch to your Inbox -- press ``ji''\n    ;; then, when you want archive some messages, move them to\n    ;; the 'All Mail' folder by pressing ``ma''.\n    (setq mu4e-maildir-shortcuts\n          '( (\"\/gmail\/INBOX\"               . ?i)\n             (\"\/gmail\/[Gmail].IMPORTANT\"   . ?!)\n             ;; (\"\/gmail\/[Gmail].Sent Mail\"   . ?s)\n             ;; (\"\/gmail\/[Gmail].Trash\"       . ?t)\n             (\"\/gmail\/[Gmail].All Mail\"    . ?a)))\n    ;; allow for updating mail using 'U' in the main view:\n    ;; I have this running in the background anyway\n    ;; (setq mu4e-get-mail-command \"offlineimap\")\n    ;; something about ourselves\n    (setq\n     user-mail-address \"user@domain\"\n     user-full-name  \"username\"\n     message-signature nil)\n    ;; sending mail -- replace USERNAME with your gmail username\n    ;; also, make sure the gnutls command line utils are installed\n    ;; package 'gnutls-bin' in Debian\/Ubuntu\n    (require 'smtpmail)\n    (if (string&lt; emacs-version \"24\")\n        (setq message-send-mail-function 'smtpmail-send-it\n              starttls-use-gnutls t\n              smtpmail-starttls-credentials '((\"smtp.gmail.com\" 587 nil nil))\n              smtpmail-auth-credentials\n              '((\"smtp.gmail.com\" 587 \"user@domain\" nil))\n              smtpmail-default-smtp-server \"smtp.gmail.com\"\n              smtpmail-smtp-server \"smtp.gmail.com\"\n              smtpmail-smtp-service 587)\n      ;; alternatively, for emacs-24 you can use:\n      (setq message-send-mail-function 'smtpmail-send-it\n            smtpmail-stream-type 'starttls\n            smtpmail-default-smtp-server \"smtp.gmail.com\"\n            smtpmail-smtp-server \"smtp.gmail.com\"\n            smtpmail-smtp-service 587)\n      )\n    ;; don't keep message buffers around\n    (setq message-kill-buffer-on-exit t)\n    ;; show images\n    (setq mu4e-show-images t)\n    ;; use imagemagick, if available\n    (when (fboundp 'imagemagick-register-types)\n      (imagemagick-register-types))\n    ;;; message view action\n    (defun mu4e-msgv-action-view-in-browser (msg)\n      \"View the body of the message in a web browser.\"\n      (interactive)\n      (let ((html (mu4e-msg-field (mu4e-message-at-point t) :body-html))\n            (tmpfile (format \"%s\/%d.html\" temporary-file-directory (random))))\n        (unless html (error \"No html part for this message\"))\n        (with-temp-file tmpfile\n          (insert\n           \"\"\n           \"\"\n           html))\n        (browse-url (concat \"file:\/\/\" tmpfile))))\n    (add-to-list 'mu4e-view-actions\n                 '(\"View in browser\" . mu4e-msgv-action-view-in-browser) t)\n    ;; convert org mode to HTML automatically\n    (setq org-mu4e-convert-to-html t)\n    ;; need this to convert some e-mails properly\n    (setq mu4e-html2text-command \"html2text -utf8 -width 72\")\n  )\n  (mu4e)\n)\n(defalias 'org-mail 'org-mu4e-compose-org-mode)\n<\/pre><\/div><h2 id=\"handy-keys-in-mu4e\">Handy Keys in mu4e<\/h2><ul><li>F \u2014 to forward<\/li>\n<li>C \u2014 to compose<\/li>\n<li>O \u2014 opens an attachment in detail view<\/li>\n<li>aV \u2014 views mail in HTML mode<\/li>\n<li>ji, j!, ja \u2014 jump to mailbox. (See mu4e-maildir-shortcuts above).<\/li>\n<li>mi, m!, ma \u2014 move message to mailbox<\/li>\n<li>V \u2014 turn on\/off duplicates (handy if you\u2019re looking at All Mail)<\/li>\n<li>W \u2014 turn on\/off related messages<\/li>\n<li>\/ \u2014 narrow the current view via search<\/li>\n<li>?, d, m \u2014 mark for unread \/ delete \/ move<\/li>\n<li>x \u2014 execute deferred marks<\/li>\n<\/ul><h2 id=\"org-mode\">Org Mode<\/h2><p>One of the neat things about mu4e is the fact that you can send messages\nbuilt in org-mode, although the current implementation is a little\nflaky. Unless you do things the right way you\u2019ll often get \u201cError 70,\u201d\nand a plain unadorned e-mail will be sent out. Unfortunately, my friends\nand associates are not yet at that advanced stage of development that\nthey\u2019ll read org-mode format directly.<\/p><p>To use org-mode via compose, you:<\/p><ul><li>Compose a message<\/li>\n<li>Run M-x org-mu4e-compose-org-mode (hah, note my alias above)<\/li>\n<li>And, finally, and most important: you must send the e-mail via \\^C\n\\^S <em>only when the cursor is in the header<\/em>. If you try to send it\nwhile the cursor is in the body of the e-mail, it won\u2019t work.<\/li>\n<\/ul><p>Yeah, not ideal. But hopefully fixable and it isn\u2019t too hard in\npractice; only a small percentage of the e-mails I send out are rich\ntext.<\/p><p>You may want to change the default formatting options; I found that\nhaving a Table of Contents in my e-mails was a bit too pretentious, and\nthe formatting for code blocks left too much white space. Apply the\nfollowing diff:<\/p><div class=\"highlight\"><pre class=\"language-diff\" data-lang=\"diff\">diff --git a\/mu4e\/org-mu4e.el b\/mu4e\/org-mu4e.el\nindex 44d5ea1..6697486 100644\n--- a\/mu4e\/org-mu4e.el\n+++ b\/mu4e\/org-mu4e.el\n@@ -170,7 +170,9 @@ and images in a multipart\/related part.\"\n            ;; because we probably don't want to export a huge style file\n            (org-export-htmlize-output-type 'inline-css)\n            ;; makes the replies with \"&gt;\"s look nicer\n-           (org-export-preserve-breaks t)\n+           (org-export-preserve-breaks nil)\n+            ;; turn off table of contents\n+            (org-export-with-toc nil)\n            ;; dvipng for inline latex because MathJax doesn't work in mail\n            (org-export-with-LaTeX-fragments 'dvipng)\n            ;; to hold attachments for inline html images\n<\/pre><\/div><h3 id=\"org-mode-example\">Org Mode Example<\/h3><p>And, okay, to be honest it\u2019s not entirely useful in the day to day, but\nit is pretty neat that if you need to you can send an e-mail via:<\/p><pre>Remember we were talking about functions of the form $f(x) = x^2$ the other day?  I wanted to show you what it looked like.\n#+name: python-data\n#+BEGIN_SRC python :exports both :results replace\nreturn [ (ix, ix*ix) for ix in xrange(0,11) ]\n#+END_SRC\nIt's hard to visualize this, though, so it helps to look at the graph.\n#+BEGIN_SRC R :var dat=python-data :exports both :results output graphics :file graph.png\nlibrary(ggplot2)\nggplot(data=dat, aes(x=V1, y=V2))+geom_point()\n#+END_SRC\n<\/pre><p>\u2026 to get<\/p><p><img src=\"http:\/\/images.brool.com\/blog\/technology\/org-email.png\" alt=\"\" \/><\/p><p><i>Comments are moderated whenever I remember that I have a blog.<\/i><\/p><p><a id=\"article-comment-add\" href=\"denied:javascript:lanyon_reply('using-mu4e',%20'0')\">Add a comment<\/a><\/p><p><a class=\"pagination-item newer\" href=\"http:\/\/www.brool.com\/post\/living-with-a-leaf\/\">&lt;&lt; Living with a Leaf<\/a>\n  \n  \n    <a class=\"pagination-item older\" href=\"http:\/\/www.brool.com\/post\/speeding-up-flyspell-region\/\">Speeding Up flyspell-region &gt;&gt;<\/a><\/p>","created_at":"2018-11-07T16:21:41+0000","updated_at":"2018-11-07T16:21:41+0000","annotations":[],"mimetype":"text\/html","reading_time":7,"domain_name":"www.brool.com","http_status":"200"}]